Splitting Forces in FRPR Pretensioned Concrete

Splitting Forces in FRPR Pretensioned Concrete
SP 138-1
Splitting Forces in FRPR
Pretensioned Concrete
by W.R. de Sitter and R.A. Vonk
Synopsis: In concrete pretensioned with non-metallic Fiber
Reinforced Plastic Reinforcement (FRPR) the Hoyer effect leads
to high splitting stresses due to confinement of radial deformations of bars or strands in the transfer zone. Incompatible
linear temperature expansion can aggravate the splitting stresses. Bond in the transfer zone is heavily influenced by the
confined radial expansion as demonstrated by tests with Arapree
bars in light weight concrete. Very short transfer lengths (80
mm) have been measured. Three calculation approaches for splitting stresses are presented; the Elasto Plastic approach, the
Concrete Deformation approach and the Fracture Energy approach.
The Elasto Plastic model has been checked using a discrete
element model including tensile softening of concrete. The
presented formula are confirmed by a few tests on Arapree
pretensioned prims.
Keywords: Fiber reinforced plastics; lightweight concrete; pretensioning;
splitting stresses; stress transfer
1
2
de Sitter and Vonk
W. Reinold de Sitter
Hollandsche Beton Groep nv director R&D
Eindhoven University of Technology
parttime professor for Concrete Structures with the
Faculty of Architecture and Building Science
Rene A. Vonk
Eindhoven University of Technology
Completed his PhD-thesis on softening of concrete and
received his doctor's title in 1992 at the
Faculty of Architecture and Building Science
INTRODUCTION
The transfer of prestress in pretensioned concrete leads to
tensile splitting stresses in the anchorage zone. These can be
attributed to two causes (fig. 1).
1. The Hoyer effect; release of pretension causes radial expansion of the prestress material due to it's Poisson coefficient.
This leads to radial compressive stresses resulting in increased bond in the anchorage zone.
2. Transfer of bond stresses along a prestress bar or strand in
the anchorage zone introduces tensile stresses in the tangential direction in the concrete.
The Hoyer effect is restrained by the concrete leading to
radial stresses Pr normal to the interface of the bar/strand
with the concrete. A high level op Pr increases the bond stresses and the angle ~. Therefore both causes of splitting forces
are interrelated. In section 3 it will be shown that even in
the case of smooth surface texture and very low bond the Hoyer
effect may lead to splitting. At the extreme end of the anchorage zone the shear stress r 1r must be zero because rr 1 is zero
at the free concrete surface. In this area the Hoyer effect
will dominate the splitting forces. The material properties
make FRPR sensitive to the Hoyer effect.
The coefficient of radial thermal expansion ar of FRPR is
governed by the (more or less) linear thermal expansion of the
bonding resin. In many types of FRPR the value of ar is 5 to 8
times the va 1ue for concrete. An increase in temperature then
will lead to tangential splitting stresses in the concrete. The
combination of thermal stresses with stresses caused by the
Hoyer effect and prestress force transfer did lead to severe
longitudinal cracking in thin pretensioned elements (fig.2).
Acknowledgement
This research program is supported by the European Community through the Brite-Euram progrnm.
FRP Reinforcement
3
THE HOYER EFFECT COMBINED WITH TEMPERATURE EXPANSION
The E-modulus E1 in the longitudinal direction of FRPR is
governed by the E-modu l us E1r of the fibers. The E-modu l i vary
from 80 000 to 140 000 MPA for aramid fibers and 4000 to 6000
MPa for resins. If the contribution of the resin is neglected:
El
=
Ar E1r
/Al
The longitudinal strain £1ro and the radial contraction t:ro in
prestensioned state:
The radial contract ion is governed by the poisson coefficient
!Ires of the resin if we neglect the influence of the radial
contraction of the fibers.
After release of pretension the expansion of the resin in the
longitudinal direction is restrained by the concrete and by the
fibers.
Transfer zone:
(1)
and
(2)
After the transfer zone radial expansion due to release of
pretension does not play a role but longitudinal deformation of
the resin is restrained by the fibers as well as the surrounding concrete. Hence, the apparent modulus varies between
EreJl-vres) and EreJ(l-vres-2v2reJ
After the transfer zone:
£ri = +( 1+11 re,) (ares -a.c) f. T
(3)
After the transfer:
Pri = £riEreJ(l-llres-211 2reJ
(4)
These expressions may be refined to include combined moduli of
fibers and resin according to volumetric laws. However even a
small amount of enclosed air may have a large influence on composite moduli in the transverse direction. Therefore composite
moduli must be measured on representative samples; they can
only be approximated from the moduli of the constituent materials.
4
de Sitter and Vonk
Comparing express ion ( 2) and ( 4) it follows that, depending on
the relative relation between the Hoyer stress and the temperature induced stresses craking will start either in the anchorage zone or at some distance from the free end.
PULL-OUT AND PUSH-IN BOND STRENGTH
In pretensioned concrete bond in the transfer zone is identical
with bond as observed in a push-in test. However in the ultimate load situation bond in the anchorage zone is more analogous
with a pull-out condition.
The influence of transverse expansion and contract ion on bond
has been investigated by comparison of bond in pull-out and
push-in tests (fig.3) [1]. These tests were too few to permit
firm quantative conclusions but they indicate significant
differences in bond. In the push-in tests arapree bars (nominal: 8 mm) with embedded sand were tensioned to values varying
between 23.64 and 26.46 kN. The circular samples of lightweight
concrete were cast around these bars with cover 21 to 35 mm.
After 3 days wet curing the tensil_e splitting strength fct was
1.6 to 2.5 MPa. Then the force F1 1n the upper part of the bar
was gradually released. The reaction forces were measured as a
check on (F 2-F 1 ) as well as the slip of the bar with respect to
the l.w. concrete. In the pull-out tests the bars were stressed
to a force 2.4 - 6.6 kN. Then the force F2 in the lower half of
the bar was increased. The mean bond stress 1 was evaluated
taking into account the embedded length of 35 mm, the actual
circumference of the bar and the peak value of ( F2-F 1 ). The
maximum observed values rmax averaged over 35 mm are listed in
Table 1. Visible cracking and crack pattern were noted.
As expected the push-in tests showed more cracking than the
pull-out tests. In series 3 (the only non cracked push-in test)
the average push-in bond stress was 11.07 MPa as compared to
6.74 MPa in the case of the corresponding pull-out test.
TRANSFER LENGTH AND SPLITTING
Transfer length and splitting were further investigated [2] in
tests on 300 mm long prisms pretensioned with 5.7 mm Arapree
bars at a nominal jacking force of 17 kN. Four types of surface
texture were included, varying from perfectly smooth to coarse
embedded sand. Prisms had sides of 30, 35, 40 and 45 mm.
In table 2 the occurence of cracking is summarized. The 40*40
prism showed cracks (fig. 4) after an interval varying between
1 minute to 3 months after testing. Typical strains measured at
the concrete surface are given in fig. 5. The considerable
spread is attributed to local differences in concrete E-modulus
due to aggregate distribution.
FRP Reinforcement
5
The modulus E1t of the fibers is 120 000 MPa and the cross
section of 100 000 fibers in the 5.7 mm bar is 11.1 mm 2 •
£10
17000/ (11.1 *120000) =1. 28%
£~
-v
Pr;
0.0045*4000/(1-0.35) = 28MPa
£w
= -0.35*1.276%=-0.45%
CALCULATIONS
Cracking due to het Hoyer effect has been simulated by Van Gils
[2] using the computer program UDEC. The effect of splitting
stresses in two dimensional cross-sections of square and circular prisms were investigated. Only the results for a circular
cross-section are discussed here, as they differ only slightly
from the results for a square cross-section and can be used for
comparison with simple design formulas. Three discrete radial
cracks were modelled in the simulations, since the tests almost
always showed failure due to the formation of three cracks.
Conclusions:
a] Embedded sand in the surface is essential to transfer of
pretension. However, if some sand is embedded the amount and
coarseness of the sand has not a dominant influence on the
transfer length, which varies between 60 and 120 mm.
(In [3] a transfer length of 135 mm in the case of 20*2,6
mm 2 flat Arapree strips was reported).
b] Prisms with smooth bars show a very reduced transfer of
pretension but they crack nonetheless. Thus showing the
predominant influence of the Hoyer effect on cracking.
This is supported further by the facts that (1) three is the
smallest number of cracks for which a kinematically possible
failure mechanism is created (fig. 7), (2) the number of cracks
is limited to three to five by the large aggregates (d~. = 8 mm)
and ( 3) three cracks result in the lowest peak stress and the
most brittle post-peak behaviour. Since concrete cracking is
not perfectly brittle, a stress-crack-opening relation with
linear softening up to ou = 0.03 mm was used in the simulations
(Fig. 6b, [4]).
Fig. 8 gives the residual pressure in the arapree as as function of the initial pressure. It shows that failure changes from
ductile to brittle when the diameter of the concrete prism increases. Failure takes place due to the growth of the radial
cracks from the inside to the outside. The maximum residual
pressure is found when the crack front reaches the outside of
the prism.
6
de Sitter and Vonk
Due to the relatively small size of the prisms, the average
opening of the cracks is still small when this happens, which
means that the crack stresses have not entered far in the
softening branch of fig. 6b and still have a significant value
over the total length of the cracks.
The simulations indicate that an elastoplastic model can give
good predictions of the maximum allowable initial pressure in
the arapree. For (dc/d) 2 >>1 the theory of elasticity yields:
Eres(l
+u)
)Pr
fc(l-ur.J
This relation is shown in fig. 8 as the tangent in the ong1n.
According to the theory of plasticity, the maximum residual
pressure is:
Pr = (dc-d)fcJd
(5.2)
An estimate of the maximum allowable pressure in the arapree is
found by calculating the intersection point between the elastic
and the plastic line, which gives
prlmax
E ( 1 +u )
=2c(l+
res
c
)
fct
(5.3)
Ec( 1-ur.J
The result for fct = 3.4 MPa, ENs = 4000 MPa and ~ = 30 000 MPa
is showin fig. 8.
According to an energy approach, failure can only take place
when the initial elastic deformation energy in the arapree:
Eea
1T(l-ure,}
4
p2 rl d2
(5.4)
Eres
is at least equal
radial cracks
to the energy needed to create the three
(5.5)
This results in
12c
Prlmax = (
Eres
G,c )'>
1Td2 (l-ures )
(5.6)
FRP Reinforcement
7
The fracture energy Grc can be determined in tests and is equal
to 0.5 fct5~ in the simulations with the computer program UDEC
(see fig. 6o)
In a third approach the minimum expansion ~d of the arapree
upon relief of the prestress, necessary to find a crack opening
5" for the three cracks, can be calculated. Because the concrete
pieces move as rigid blocks, as they are stress-free before and
after the relief of the prestress, this calculation results in
the simple relation.
2!lr 1
=
M
5
=-"
(5. 7)
13
According to equation 2 this expansion corresponds to an initial pressure in the arapree:
Eres
Primax = --,(1-_-U-.,.-
(5.8)
res
The results of the three approaches are compared in fig 9 for d
= 5.7 mm. The fact that the curves for the energy approach and
the deformation approach give higher allowable pressures for
small diameters of the concrete prism than the elastoplastic
approach indicates the ductile character of the failure for
those diameters. For larger diameter a more brittle failure is
predicted by the models. This is all in accordance with the
results of the computer simulations. The preferable design
curve is that of the elastoplastic approach because it guarantees a residual stress which is close to the maximum. It will
result in the highest shear-stress transfer and thus in the shortest transfer length of the prestress in the anchorage zone.
Attention should be given to the required safety factor as the
failure of the concrete prisms is brittle for large diameters
and a delayed crack growth is observed in the tests of section
4.
Fig. 9 gives also a comparison of the model predictions and the
test results of section 4. It shows that the test results agree
with the elastoplastic model. The deformation approach is found
to be very conservative. One additional comparison with test
results can be made. Concrete prisms with a square cross-section of 25x25 mm 2 made of a special high-strength Densit mortar
(fq = 11MPa, f~ = 180 MPa, Ec = 65000 MPa) were also pretensionea with a force of 17 kN(p 1 = 27.5 MPa). The elastoplastic
model, the energy model and the deformation model predict an
allowable initial stress p1 of 41.5, 33.9 and 18.7 MPa, respectively.
8
de Sitter and Vonk
The successful production of the prisms agrees with the first
two models. The prisms cracked when they were submerged in
water of 60°C. According to the equations in section 2, the
rise in temperature increased the initial pressure in the
arapree to 42.3 MPa, which is more than the elastoplastic model
and the energy model allow.
Conclusions
Three approaches to calculate the critical initial confined
radial pressure Prlmax have been presented; the deformation
criterium, the elasto plastic criterium and the fracture energy
criterium. Based on a very limited number of tests and practical experience the following preliminary conclusions are
drawn.
1) In FRPR pretensioned concrete the confined radial deformations (Hoyer effect) are critical for splitting forces.
2) The critical confined pressure can be calculated using the
elasto plastic criterium (5.3)
3) A safer value for Pr 1 ma~ may be obtained by comparison of the
elasto plastic criter1um and the fracture energy criterium
(5.6) and taking the smallest value.
4) Due to interaction with the Hoyer effect there is a significant difference between bond in push-in tests and bond in
pull-out tests. Hence transfer lengths should either be
directly measured or based on measurements of bond in pushin tests.
NEW DEVELOPMENTS
AKZO and HBG have developed a compressible coating for FRPR (as
well as epoxy coated strands) in order to reduce radial stresses due to the Hoyer effect and incompatible temperature expansion coefficients. Short duration tests have shown that this
coating solves the problem. There is an influence on bond but
transfer lengths remain significantly less than for steel
strands. Long duration tests are in progress. If these tests
are successful a major obstacle to the use of FRPR in thin
pretensioned concrete elements will have been removed. A patent
application has been submitted.
REFERENCES
[1] Sarneel, R.G.M.; 1 "Het aanhechtingsgedrag van Arapree aan
lichtbeton" deel 1 & 2 ( in Dutch) MSc. report may 1990
Eindhoven University of Technology.
[2] Gils, A.W.M. van; "Overdrachtslengte en het optreden van
splijtscheuren bij Arapree <jJ 5 mm" deel 1 & 2 (in Dutch)
MSc. report july 1991 Eindhoven University of Technology.
FRP Reinforcement
9
[3] Egas, M.; Sitter, W.R. de; "Tests on Arapree Pretensioned
Sleepers" International Symposium on precast Railway Sleepers;
Madrid april 1991 Monograficos 8. Collegia de Ingenieros de
Caminos, Canales y Puertos
[4] Vonk, R.A.; "Softening of Concrete loaded in Compression"
Ph.D. Thesis june 1992 Eindhoven University of Technology.
SYMBOLS
cover ; mm
bar diameter ; mm
concrete cylinder diameter mm
tensile spl.strength; MPa
pressure; MPa
radius of bar; mm
outer radius of cylinder; mm
area ; mm2
Young's moduljus; MPa
deformation energy; Nmm/mm
force ; N
fracture energy; Nmm/mm
thermal expansion; 1/aC
crack opening ; mm
strain
angle
Poisson coefficient
shear stress; MPa
normal stress; MPa
temp. difference ; ac
radial expansion; mm
indices:
concrete
fiber
resin
res
radial direction
longitudinal
tangential
initial
10
de Sitter and Vonk
TABLE 1 -BOND IN PULL-OUT/PUSH-IN CONDITIONS
Pull-Out
1
se
ries
2
Push-in
3
5
4
1
2
3
4
cover
21
28
35
35
35
21
28
35
35
crakcs
yes
no
no
no
no
yes
yes
no
yes
mm
8.4
4.77
6.74
7.89
7.20
6.89
6.25
11.07
8.78
MPa
fct
2.35
2.37
1.61
2.41
2.30
2.10
2.45
1.61
2.41
MPa
F
3.94
2.47
5.85
6.59
0
20.4
20.18
13.9
18.1
kN
F2
11.33
6.67
11.78
13.53
6.34
26.64
25.68
23.64
2582
kN
r
mAV
TABLE 2- OCCURRENCE OF CRACKING
texture
coarse
normal
fine
smooth
30
X
30 mm 2
yes
yes
yes
yes
35
X
35 mm 2
yes
yes
ves
yes
40 x 40 mm 2
later
later
no
later
c:f::E/ •
I
1,. .....,.
!...... !
I
.·· .
. ··-occ···.
..f... ......
cb
·· .... ,
ai
_________ :
Fig. 1-Hoyer effect
45x 45 mm 2
no
no
no
no
FRP Reinforcement
Figure 2
Fig. 3-Pusb-in and pull-out tests
11
12
de Sitter and Vonk
Figure 4
0.3
0.25
0.2
0
.g
..2.
s::::
·~
'Iii
0.15
0.1
.
.:·
'
• .......... ~- •••
...••• l.
'::../:,..:-':
'
' '"''"'/'f
; /,;..
/,
0.05
.,., .. ,............ ,.............,.............,..........; ............ ; .............,........ ..
...
0
(0.05) 0
20
40
60
BO
100
120
140
1BO
1BO
200
220
240
280
280
location [mm]
coarse 16.7 kN nonnal15.9 kN ftna17.1 kN
----
·······
--
smooth 15.7 kN
Fig. 5.....;....Compressive strain in 45 x 45 mm2 300 mm prisms
17 kN nominal pretension
300
FRP Reinforcement
....,
8!. ---- fct
~
t:l
a)
gj
Q)
~
~
'iii
1::
.s
crack opening cS [mm]
Figure 6
Fig. 7-3 cracks
= minimum degrees of freedom
13
14
de Sitter and Vonk
35
A
30
l!r -
/
Elastoplastic model
-t;.
25
ro
0..
~
20
15
Q
10
5
0
20
10
0
Pri
30
40
50
[MPa]
Fig. 8-UDEC calculations
50 ,-------------------------------,
Elastoplastic model
40
Energy model
ro
0..
2
Deformation model
30
X
ro
E_ 20
...
10
/
/
o
Cracked prism
•
Uncracked prism
I
I
0
0
10
20
30
40
50
de [mm]
Fig. 9-Model predictions and test results
60
SP 138-2
Investigation of S-2
Glass/Epoxy
Strands in Concrete
by R. Sen, D. Mariscal, and M. Shahawy
Synopsis:
A comprehensive durability study of S-2 glass/epoxy pretensioned
beams exposed to wet/dry cycles in 15% salt solution indicated a complete
loss of its effectiveness within 3-9 months of exposure. This paper presents
results of subsequent follow-up investigations to identify the cause of this
deterioration and also to examine practical measures that could be used to
prevent its occurrence.
The analysis of the test results suggests that the most likely cause of
failure was diffusion of hydroxyl ions from the concrete pore solution
through the Shell Epon 9310 resin. This is supported by SEM micrographs of the failed beam.
While these conclusions are only valid for the Shell Epon 9310 resin
examined, diffusion is also likely to be a characteristic of other types of
resins, e.g. vinylesters or polyesters. This makes long term protection of
glass fibers in concrete problematic.
Keywords: Diffusion; durability; epoxy resins; fiber reinforced plastics; glass
fibers; pretensioning; rupture; salt water; scanning electron microscope;
stresses
15
16
Sen, Mariscal, and Shahawy
ACI Member Rajan Sen is an Associate Professor of Civil Engineering and
Mechanics at the University of South Florida, Tampa. He is the author
and co-author of over seventy technical publications and was the co-editor
of the first ASCE proceedings on the application of advanced composites
for Civil Engineering Structures.
Daniel Mariscal is a Structural Engineer with Zurn Balcke-Durr, Tampa,
FL. He holds a B.S.C.E from Pontifical Catholic University of Peru and
a M.S.C.E. in Civil Engineering from the University of South Florida,
Tampa.
ACI Member Mohsen Shahawy is the Chief Structural Analyst, Florida
Department of Transportation, Tallahassee, FL.
He has authored
numerous technical papers. Dr Shahawy is a member of ACI Committee
440, FRP Bar and Tendon Reinforcement and Transportation Research
Board's Committee A2C07 on Structural Fiber Reinforced Plastics.
INTRODUCTION
In recent years, fiberglass composites have been increasingly
considered for pretensioning, as reinforcement, as two and three
dimensional grids and for strengthening concrete beams ( 1-4 ). These
applications all share one important characteristic, namely, the fiberglass
composite is directly in contact with concrete. Because glass fibers are very
vulnerable in an alkaline concrete environment (5), an implicit assumption
is that the resin system used can provide a protective barrier that assures
the integrity of the fiberglass composite. The validity of this hypothesis had
not been rigorously investigated until quite recently.
In 1989, the Florida Department of Transportation (FOOT), in
cooperation with the US Department of Transportation, sponsored a 26
month study on the feasibility of using fiberglass pretensioned piles in a
marine environment (6). As part of this study, tests were undertaken to
determine the durability of identical pre-cracked and uncracked S-2
glass/epoxy pretensioned beams subjected to simulated tidal cycles in 15%
salt solution (see Figure 1).
In the tests, durability was indirectly assessed from changes in
ultimate capacity of the S-2 glass/epoxy pretensioned beams. This
indicated a rapid and substantial loss in load carrying capacity with
FRP Reinforcement
17
exposure. For the pre-cracked beams, this reduction occurred after only
3-9 months of exposure (see Figure 2). It took somewhat longer, between
12-18 months, for the uncracked beams (see Figure 3). No commensurate
reduction was observed in the unexposed control beams tested over the
same 20 month period. Complete details are presented elsewhere (7).
Inspection of Figures 2-3 shows that the recorded failure loads in
the exposed S-2 glass/epoxy pretensioned beams approached that for
unreinforced beams. Since no slip was recorded, this meant that prolonged
exposure had led to a complete destruction of the S-2 glass.
This paper describes investigations undertaken to determine why the
epoxy was unable to protect the S-2 glass fibers and what measures are
needed if this problem is to be solved. For simplicity, all subsequent
references to 'fiberglass' imply the 3/8 in. (9.5 mm) S-2 glass/epoxy strands
used in the testing.
OBJECTIVES OF STUDY
The primary objectives of this investigation were as follows:
1.
To identify the most probable cause of deterioration of the
fiberglass strands.
2.
To corroborate the results of ultimate load tests with
scanning electron microscope examination of fiberglass
samples.
3.
To identify measures that can enhance the durability of glass
fiber composites in a concrete environment.
RESEARCH SIGNIFICANCE
Knowledge of the mechanism by which S-2 glass fibers were
destroyed is essential for devising appropriate protective measures. Such
measures can greatly extend the scope of application of glass fiber
composites in concrete. The investigations described may also be adapted
to evaluate the durability of other fiber reinforced plastics in alternative
applications.
18
Sen, Mariscal, and Shahawy
CAUSE OF FAILURE
Damage to the fiberglass strands could be due to stress corrosion or
glass/alkali reaction. The most likely cause of failure may be determined
by a careful review of experimental data from the durability tests relating
to stress levels in the fiberglass strands and the alkalinity of the concrete
useJ.
Materials
Fiberglass -- The S-2 glass used in the durability tests is a lower
cost version of the military grade S glass fiber. It is a magnesia-aluminasilicate glass with mechanical properties superior to E glass, used in
demonstration structures constructed in Europe (8).
Its chemical
composition (9) and mechanical properties (10) are summarized in Tables
1-2.
An epoxy resin, Shell Epon 9310, was selected for the pultrusion of
the S-2 glass rods since the amine cured resins are among the best systems
for resisting the effects of an alkaline and saline environment. The
properties of the Shell Epon 9310 epoxy resin (11) are listed in Table 3.
Seven 0.118 in. (3 mm) diameter S-2 glass/epoxy rods were held
together by plastic ties to make up the 3/8 in. (9.5 mm) strands (effective
area = 0.0702 in 2 (45.3 mm 2 )), used for pretensioning the specimens (12).
Two such strands were used in each test specimen.
Concrete -- The concrete used for the durability specimens was an
FDOT approved mix using Type III rapid hardening cement, requiring a
water/cement ratio of 0.41. Typically, the sodium oxide equivalent of this
cement is below 0.6% (see Table 4) making it a low alkali cement
according to ASTM C 150 classification. In liquid form, its pH value was
estimated to be 12.5 (13). This pH is in the 12.5-13.5 range for pore fluid
in normal concrete and represents a caustic or highly alkaline solution (14).
Stress Corrosion
Stress corrosion refers to the characteristic property of fiber
reinforced plastics in which the failure strength under long term sustained
loads in a chemical environment is lower than its short term tensile
strength. In air, this phenomenon is referred to as 'stress rupture'.
FRP Reinforcement
19
The ultimate tensile strength, frgu• and elastic modulus, Erg• of the
fiberglass strands used in the durability study were 273.5 ksi (1.89 GPa) and
9.16 x 106 psi (63.1 GPa) respectively (12). The target jacking force used
for pretensioning each strand was 10,000 lbs ( 44.5 kN) or 0.52frgu· Prior to
release, the average force in the fiberglass strands was measured to be
approximately 8,000 lbs (35.6 kN) or 0.41frgu·
The effective prestress at the end of the tests can be estimated as
0.38frgu assuming losses were 7.1% of the stress before release, i.e. the same
as that measured in a parallel creep study using identical fiberglass strands
jacked to the same level in specimens having the same dimensions (15).
The results of stress rupture tests on S glass/epoxy composites
conducted over a 10 year period at the Lawrence Livermore Laboratory,
CA (16) showed that only 2% of the test specimens failed at sustained
stresses of 0.4frgu while none failed at stresses of 0.35frgu· Since S glass and
S-2 glass have comparable properties, the stress rupture characteristics of
the fiberglass strands used in the durability study will also be similar.
While stress corrosion characteristics will be worse than stress
rupture ones, the combination of relatively low stress levels in the
fiberglass strands coupled with its deterioration within 3-9 months (see
Figure 2) suggest that stress corrosion could not have been the principal
cause of failure. Further evidence to support this is provided by the results
of scanning electron micrographs presented later. This shows the
deterioration to be as pronounced in the submerged region (point #1 in
Figure 2) located 6 in. (150 mm) from the end as that at mid-span (point
#5) even though its sustained stress level was about 40% lower (the
transfer length was measured to be about 10 in).
Glass/Alkali Reaction
Concrete has microscopic voids whose size and distribution depend
on numerous factors such as water/cement ratio, mode of vibration, or
curing (14). Because of the alkalinity of the concrete used, capillary water
present in these voids from the wet/dry cycles used in the durability tests
would also be highly alkaline.
Silica, the largest constituent of S-2 glass (see Table 1), and alkalis
react chemically (17). Hydroxyl ions from the alkaline pore solution can
dissolve the basic silicon-oxygen-silicon structural network of S-2 glass as:
-Si - 0 - Si - + OH--+ -Si - OH + Sio- (in solution)
(1)
20
Sen, Mariscal, and Shahawy
This dissolution of silica is accompanied by a rapid and severe strength
loss.
If concrete's pore solution could penetrate the epoxy resin protecting
the S-2 glass fibers, the ensuing silica/glass reaction would lead to a
dramatic loss in strength over a relatively short period of time as observed
in the testing.
Silica/alkali reaction can also explain why the lesser stressed S-2
glass fibers in the submerged end suffered damage. Thus, the most likely
cause of premature failure was a breakdown in the epoxy barrier protecting
the S-2 glass fibers in concrete.
BREAKDOWN OF EPOXY BARRIER
Ingress of hydroxyl ions from the concrete pore solution through the
epoxy resin could be attributed to the following:
1.
Damage to the fiberglass strand surface during fabrication of
specimens or cracking of the epoxy resin during
pre tensioning.
2.
Diffusion through the epoxy resin.
Resin Damage
If the fiberglass strands were damaged during fabrication, such
damage could be expected to occur in a random fashion and not be
confined to the critical mid-span section in all the specimens.
Furthermore, a 'damage theory' cannot explain why the pre-cracked beams
failed earlier than the uncracked ones. At best, surface damage could have
been a contributory factor.
The strain in the epoxy, corresponding to the maximum jacking load
of 10,000 lbs. (44.5 kN) was 1.56% [10,000/(0.0702)/9.16 x 106 ]. This is
well below the 4% strain limit for Shell Epon 9310 quoted by its producers
(see Table 3). While localized cracking is a possibility, its likely random
distribution cannot explain the preferential failure of the pre-cracked
specimens.
FRP Reinforcement
21
Diffusion
Since neither damage nor cracking can reconcile the experimental
data, the most probable cause of ingress of hydroxyl ions is due to diffusion
through the epoxy resin.
Inspection of the properties of the Shell Epon Resin in Table 3
shows that it had a moisture absorption rate of 1.5% of its weight after 14
days immersion in 200°F water. At the 70°F temperature of the laboratory
(18), it could be expected to have a smaller, though significant, diffusion
rate.
Diffusion can explain why the pre-cracked specimens failed earlier
compared to the uncracked specimens shown in Figures 2 and 3. The midspan location of the cracks, provided a direct passageway for the alkaline
pore solution to reach the fiberglass strands at the critical section. It can
also explain why the fiberglass strands in the submerged part of the beams
experienced more damage - the diffusion was greater here because the
capillary voids were never empty.
SCANNING ELECTRON MICROSCOPE EXAMINATION
To corroborate the results of the ultimate load tests and to
determine the extent of degradation of glass fibers, scanning electron
microscope (SEM) examinations were carried out. In view of limitations
of cost, only one fiberglass beam was examined. This uncracked beam,
FG-W2, had displayed a classic corrosion failure mode after 18 months
exposure to wet/dry cycles.
Representative samples were taken from FG-W2 at locations
corresponding to the permanently wet, 'splash zone' and permanently dry
zones, i.e. points #1, #5 and #9 in Figures 2-3. To obtain a measure of
the damage, if any, a stressed fiberglass strand exposed only to laboratory
conditions was used as control. This was one of two strands that were
tensioned during the fabrication of the durability specimens on December
28, 1989, but not used subsequently because of failure of the other strand
(6).
All samples were professionally prepared for the SEM examination
to a half micron final polish (see Figure 4). They were subsequently
sputter coated with gold paladium and stored in a dust free vacuum
chamber prior to their examination at the Surface Analysis Laboratory,
University of South Florida, Tampa. Since only a limited number of
22
Sen, Mariscal, and Shahawy
samples were examined, the findings presented are necessarily preliminary.
Nevertheless, they provide valuable information on the relative state of
identically prepared samples taken from exposed and un-exposed segments
of failed beams.
Results
Permanently Dry Location -- Figure 5 compares micrographs for the
control (unexposed) specimen with that for the sample taken from the dry
end of the beam (point #9). Inspection of this micrograph shows that the
relative state of the glass fibers is comparable with little evidence of
degradation of the fibers. This corroborates the test results where no
significant reduction in ultimate strength was observed for the control
beams (see Figures 2 and 3).
Splash Zone-- Figure 6 compares two micrographs for a sample at
the splash zone (point #5). Inspection of this micrograph shows that there
was much greater deterioration and degradation of the glass fibers located
nearer the concrete interface. By contrast, there is less dissolution of the
fibers though more cracking at fibers located farther away.
Permanently Wet Location -- Figure 7 presents two micrographs for
the sample that was permanently submerged (point # 1). As for the
sample taken from the splash zone, there is considerable evidence of
deterioration and degradation particularly in those fibers located close to
the concrete interface. The extent of degradation for this sample appears
to be greater in comparison to the sample at the splash zone (Figure 6).
Discussion of Results
Overall, the micrographs corroborate the results of the ultimate load
tests. There is little damage at the dry end (Figure 5) but much greater
damage in the submerged (Figure 7) and splash zone locations (Figure 6).
The extent of the degradation is greater near the concrete interface than
near the middle of the fiber. Such observations are consistent with the
conclusion reached earlier that diffusion by the epoxy resin was largely
responsible for the destruction of the fiberglass. The damage to the S-2
glass fibers observed is commensurate with silica's dissolution in alkalis.
FRP Reinforcement
23
IMPROVING DURABILITY OF FIBERGLASS IN CONCRETE
Since diffusion of the alkaline pore solution by the epoxy resin was
identified as the primary cause of failure of the fiberglass, it might be
argued that the use of an impermeable epoxy would be the ideal solution
for improving durability. From a practical standpoint, however, such a
resin is likely to be prohibitively expensive and may be impossible to use
in a pultrusion process. More importantly, the experience of epoxy coating
of reinforcement bars suggests that reliance on a single method of
protection is unwise in the long run.
In view of this, a comprehensive protective system must be devised.
At its core, the alkali resistance of the glass should be increased while
simultaneously reducing alkalinity and porosity of the concrete. (The
simplest and possibly the most trouble-free solution would be use of nonalkaline cement. Unfortunately, it may not be possible to produce it
economically).
Whereas S-2 glass fibers were used in this study because of their
superior mechanical properties, in future applications greater emphasis
should be placed on durability. Careful consideration should be given to
the use of alkali resistant (AR) glass (19), which is more durable in
concrete. The resin system selected should be the least permeable,
commensurate with costs. The possibility of using additional protection
such as circumferential wrapping with a polymeric yarn (20) should be
considered. This will help prevent surface damage and also provide an
additional barrier against alkali attack.
The permeability of the concrete should be reduced to increase
durability. This may be achieved by using condensed silica fume whose
fine particle size allows voids in the cement paste or between the
aggregates and cement paste to be filled. The cement with the least
alkalinity should be used in construction. As noted earlier, the cement
used in this study qualifies as a low alkali cement but its pH of 12.5 proved
too high. In addition, procedures successfully used for improving durability
of glass fiber reinforced concrete (21), should be carefully investigated
since they could well apply for pultruded specimens.
Whatever system of protection is adopted must be carefully
evaluated using both accelerated and real time tests. Only when tests
provide complete assurance of durability can glass fiber composites be used
in applications requiring direct contact with concrete.
24
Sen, Mariscal, and Shahawy
CONCLUSIONS
The following conclusions can be drawn:
1.
The primary cause of deterioration of S-2 glass/epoxy strands was
diffusion of alkaline pore solution through the Shell Epon 9310
resin (see Table 3). The low stress level (estimated to be 0.38frgu at
mid-span and 0.23frgu at point #1) and the relatively short time
needed for deterioration (3-9 months) make stress corrosion
improbable. Other possible causes such as surface damage or resin
cracking are also implausible since they cannot explain a very
important result from the durability testing, namely, that the precracked beams failed much earlier than the uncracked ones. Thus,
their contribution, to the destruction of fiberglass, if any, was minor.
2.
The results of the scanning electron microscope investigation
corroborate the findings from the durability testing. There was little
damage to glass fibers in the dry segment (see Figure 5) but much
greater damage in the wet or wet/dry segments (Figure 6 and 7).
Damage was also greater nearer the concrete interface than away
from it.
3.
It is highly improbable that the Shell Epon 9310 resin used in the
study is the only resin that absorbs moisture by diffusion. Less
expensive polyesters and vinylesters can be expected to show
comparable diffusion characteristics. In view of this, the durability
of all glass fiber composites in concrete must be carefully
investigated. Several measures are proposed that may eventually
lead to a satisfactory solution. Until such time, however, direct
contact of glass fiber composites with wet concrete should be
avoided.
ACKNOWLEDGEMENTS
The investigation reported in this paper was carried out with the
financial support of a research grant from the US and Florida Department
of Transportation to the University of South Florida, Tampa, FL.
Additional financial support was provided by the College of Engineering,
Department of Civil Engineering and the Division of Sponsored Research
at the University of South Florida. However, the opinions, findings and
conclusions expressed in this paper are those of the writers and not
necessarily those of the Florida or US Department of Transportation.
FRP Reinforcement
25
REFERENCES
1.
Iyer, S. and Kumaraswamy, C. (1988) Performance Evaluation of
Glass Fiber Composite Cable for Prestressing Concrete Units. 33rd
International SAMPE Symposium 33, Anaheim, CA.
2.
Faza, S.S. and GangaRao, Hota (1991) Bending Response of
Beams Reinforced with FRP Rebars for Varying Concrete
Strengths, in Advanced Composite Materials in Civil Engineering (Ed.
S. Iyer and R. Sen), ASCE Specialty Conference, Las Vegas, NV.
3.
Goodspeed, C., Schmeckpeper, E., Gross, T., Henry, R., Yost, J. and
M. Zhang (1991) Cyclical Testing of Concrete Beams Reinforced
with Fiber Reinforced Plastic (FRP) Grids, in Advanced Composite
Materials in Civil Engineering (Ed. S. Iyer and R. Sen), ASCE
Specialty Conference, Las Vegas, NV.
4.
Saadatmanesh, H. and Ehsani, M. (1992) R/C Beams Strengthened
with GFRP Plates: Experimental Study, ACI Spring Convention,
Washington, D.C. March.
5.
Bentur, A. and Mindess, S. (1990) Fiber Reinforced Cementitious
Composites, Elsevier Science Publishers, New York, New York.
6.
Sen, R., Issa, M. and Mariscal, D. (1992) Feasibility of Fiberglass
Final report
Pretensioned Piles in a Marine Environment.
submitted to the Florida and US Department of Transportation,
August, pp. 318.
7.
Sen, R., Mariscal, D. and Shahawy, M. (1993) Durability of
Fiberglass Pretensioned Beams. To appear in ACI Structural
Journal.
8.
Iyer, S. and Sen, R. (1991) Advanced Composite Materials in Civil
Engineering ASCE Specialty Conference, Las Vegas, NV.
9.
Greenwood, M.E. (1991) Private Communication with R. Sen.
10.
S-2 Glass® Fiber (1990) Enhanced Properties for Demanding
Applications, Publication N2 15-PL-16154, Owens-Corning Fiberglas
Corporation, Ohio.
11.
Shell Chemical Company (1990), Technical Bulletin SC:712:90,
Epon Resin 9310 Houston, Texas.
26
Sen, Mariscal, and Shahawy
12.
Iyer, S. (1991) Final Report on Fiberglass Cable Testing, Rapid
City, SD, December, 64 pp.
13.
Florida Mining and Materials Corporation, Laboratory Test
Certificate for Type III Cement (1992), Brooksville, FL
14.
Mehta, P.K. (1986) Concrete - Structure, Properties and Materials,
Prentice-Hall, New Jersey.
15.
Issa, M. and Amer, A. (1992) Prestress Losses in Fiberglass
Prestressed Concrete Columns. Paper under review.
16.
Glaser, R.E., Moore, R.L. and Chiao, T.T. (1983) Life Estimation
of an S-Glass/Epoxy Composite under Sustained Tensile Loading,
Composites technology Review, Vol. 5, No. 1, pp. 21-26.
17.
Iler, R. (1979) The Chemistry of Silica, John Wiley & Sons, New
York, NY.
18.
Mariscal, D. (1991) "Strength and Durability of Fiberglass
Pretensioned Elements". MSCE Thesis submitted to Department
of Civil Engineering & Mechanics, University of South Florida,
Tampa, FL, December, pp. 301.
19.
Majumdar, A.J. and Ryder, J.F. (1968) Glass Technology, Vol 9,
No 3, June, pp. 78-84.
20.
Zoch, P., Kimura, H., Iwasaki, T. and Heym, M. (1990) Carbon
Fiber Composite Cables - A New Class of Prestressing Members.
Presented at the 1990 TRB Meeting, Washington, D.C.
21.
Hayashi, M., Sato, S. and Fujii, H. (1985) Some ways to improve
durability of GFRC. Proceedings - Durability of Glass Fiber
Reinforced Concrete Symposium, PCI, Chicago, IL, November.
NOTATION
1-I.T.
L.T.
Erg
frgu
High tide
Low tide
Elastic modulus of S-2 glass/epoxy strand
Ultimate tensile strength of S-2 glass/epoxy strand
FRP Reinforcement
27
Table 1 Approximate Chemical Composition of S-2 and E Glass Fibers (9)
Component
(wt%)
Si0 2
S-2 Glass
E Glass
69
55
19
CaO
Alz0 3
21
6
Bz03
MgO
2.7
0.5
1.2
Nap, K 20
Other 1
14
7.3
4.3
Fez0 3 , Zr 20 3 , Ti0 2 , ZnO
Table 2
Mechanical Properties for S-2 and E Glass Fibers (10)
Property
ASTM
Standard
E Glass
S-2 Glass
02101
500-550
250
665-700
350
02343
10-10.5
11.8
12.5-13
12.9
02343
4.5-4.9
5.4-5.8
Tensile Strength, ksi at
72' F (Single Filament)
1000' F
Tensile Modulus, x10 6 psi at
72'F
1000' F
Failure Strain, %
1 psi
1 ksi
=
=
6.895 kPa
6.895 MPa
28
Sen, Mariscal, and Shahawy
Table 3
Properties of Shell Epon® Resin 9310 System ( 11)
Property
Density, grfcc
Moisture
Absorption 2,
% wt, 14 days
System A 1
System B
1.18
1.20
1.5
Tensile Strength, ksi at
73'F
300'F
11
Tensile Modulus, ksi at
73'F
300'F
450
453
205
4
4.0
5.2
11
3.8
Failure Strain, % at
73' F
300' F
1 psi = 6.895 kPa
1 ksi = 6.895 MPa
Curing schedules (System A similar to that used for fiberglass rods)
Sample immersed in 200' F water
Table 4
Chemical Composition of Cement (by weight) (13)
Component
wt%
Si0 2
20.42
CaO
63.35
Al 20 3
5.15
Fe 20 3
4.2
MgO
0.74
so 3
3.36
Nap Equiv.
0.51
Others 1
2.03
Others includes insoluble resins and L.O.I.
FRP Reinforcement
29
Fig. 1-Tidal simulation set up
3.5 . - - - - - - - . - - - - - - - - - . , - - - - - - - . - - - - - - - - ,
For 2400 J.l£ concrete strain or actual failure
Pre-cracked Fiberglass Beams
3.0
Iii'
c.
g
2.5
----------o
"0
cU
0
...J
2.0
Q)
1.5
-
- - - - - - ____ -cr ____ Control (Unexposed)
0
#9
E
:;:::;
5
till"~~
I
cU
8ft.
•
t=H.T.
L.T.
1.0
Unreinforced capacity (average)
0.5
0.0
0
L __ _ _ _ _...J....._ _ _ _ ___J..._ _ _ _ _ _J....__ _ _ _ __J
-0
5
10
15
20
Exposure Time (Months)
Fig. 2-Ultimate strength change with exposure for precracked
fiberglass beams
30
Sen, Mariscal, and Shahawy
3.5 r - - - - - - , - - - - - - , - - - - - - - . - - - - - - - - - ,
For 2400 IJ.C concrete strain or actual failure
3.0
(i)
c.
2.5
"C
ca
2.0
-
1.5
s
1.0
g
Un-cracked Fiberglass Beams
- - - - - - -<:r - - __ ~~n!:o~ (Unexposed)
------o
0
_J
Q)
ca
.§
#9
t 111-~-~-H.T.
~
0.5
L.T .
8ft.
Unreinforced capacity (average)
0.0 L - - - - - - - - ' - - - - - - - " - - - - - - - L - - - - - - - '
20
5
10
15
0
Exposure Time (Months)
Fig. 3-Ultimate strength change with exposure for uncracked
fiberglass beams
Fig. 4-SEM specimens- control, #1, #5, #9
FRP Reinforcement
Fig. 5-Unexposed control specimen
Fig. 5-Always dry specimen
31
32
Sen, Mariscal, and Shahawy
Fig. 6-Splash zone specimen away from concrete interface
Fig. 6-Splash zone specimen near concrete interface
FRP Reinforcement
Fig. 7-Submerged specimen away from concrete interface
Fig. 7-Submerged specimen near concrete interface
33
SP 138-3
Characteristics of
Aramid FRP Rods
by K. Mukae, S. Kumagai, H. Nakai,
and H. Asai
Synopsis: As part of their research into the use of aramid (brand
name:Technora®) FRP rods as tensioning material for prestressed concrete or as
concrete reinforcement the authors conducted tension test, creep test and fatigue
test on the material. In the assessment of the strength characteristics of the
material, the anchoring method has a serious influence on the evaluation. In these
tests,we appraise the <uamid FRP rods with a bond type anchoring method which
has a high performance.
The tests showed that the tensile strength ofTechnora® aramid FRP rods has a
normal distribution, with values equal to or above those obtained for steel
reinforcement bars. The creep failure curve showed a logarithmic relationship
with the time axis; the experimental equation is show here. From the results of a
fatigue test with 2 mill ion loading cycles, it was ascertained that the fatigue
strength of the material is adequate for practical use.
Keywords: Anchorage (structural); bond (concrete to reinforcement); creep
properties; fatigue tests; fiber reinforced plastics; prestressed concrete;
tensile strength; tests
35
36
Mukae et al
Kunihiro Mukae is a chief research engineer at the Institute of Technology &
Development, Sumitomo Construction Co.,Ltd., Japan. He has been studying
concrete materials and prestressed concrete structures, and is currently
researching durability of concrete members and prestressed concrete structures
with fiber reinforced plastics.
Shin'ichiro Kumagai is a manager at the Civil Engineering Division, Sumitomo
Construction Co.,Ltd., Japan. He has been involved in various fields of research
and development, focussing chiefly on reinforced concrete and prestressed
concrete structures. He is currently researching practical applications of fiber
reinforced plastics.
Hiroshi Nakai is a manager at the Institute of Technology & Development,
Sumitomo Construction Co.,Ltd., Japan. His research interests are focussed on
the behavior of reinforced concrete, and on prestressed concrete structures with
fiber reinforced plastics.
Hiroshi Asai is a senior engineer at the Institute of Technology & Development,
Sumitomo Construction Co.,Ltd.,Japan. His research interests are the
characteristics of fiber reinforced plastics and anchorings.
I. INTRODUCTION
Research is in progress in the construction industry into the use of FRP (fiber
reinforced plastics) rods with high tensile strength as tendons for prestressed
concrete or as reinforcement material for concrete. In comparison to steel FRP
rods are highly resistant to corrosion in a saline environment. For this reason,
they are expected to prove very useful in harsh environments such as marine
environments.
For FRP rods to be used for tendons in prestressed concrete or for ground
anchors, etc., a suitable anchoring method for FRP rods has to be developed.
The authors have developed a bond type of anchoring for aramid FRP (AFRP)
tendons, and have applied this anchoring with AFRP tendons to actual
structures. II ,2]
The present paper describes tension test, creep test and fatigue test carried out on
AFRP tendons with a bond type anchoring. and applies the results of these tests
to a consideration of tensile stress standards for AFRP tendons.
FRP Reinforcement
37
2. CHARACTERISTICS OF AFRP TENDONS
The AFRP rods used in the present tests were developed in jointly with Teijin
Co.,Ltd. The rods consisted of bundles of aramid fibers (brand
name:Technora®), which were impregnated with vinyl ester resin and hardened
after being formed into rods by the pultrusion method. As shown in Fig. I, the
rods consisted of a <j>6mm base, around which were wound three types of aramid
fiber. First, a winding fiber was wound around the rod to deform its surface, and
on top of this were wound a longitudinal securing fiber and a round securing
fiber, which were fixed with resin. With these three types of winding fiber, the
AFRP rods are assured of a good bond in concrete or mortar.
The rods were anchored using a bond type anchoring. This bond type anchoring
device consisted of a single AFRP rod inserted into a 16mm inside diameter steel
tube, and fixed by the injection of shrinkage compensating mortar. Water/cement
and admixture ratio in the shrinkage compensating mortar was 31%, and the C:S
(cement+admixture: fine aggregate) ratio was 1: I. Bolts were added to the outer
surface of the steel tube, and these were secured with nuts.
Various anchoring methods for FRP tendons have been discussed by researchers,
but a completely practical anchoring method has not yet been found.[3,4] Even
with the bond type anchoring method used in the tests discussed here, it was not
possible to attain I 00% of the theoretical strength of the rods. Therefore, the test
results given here are the results of tests of AFRP tendons including their
anchoring.
3. TENSION TEST
3.1 Test Method
The device used for tension test is shown in Fig.2. The test sample consisted of a
single <j>6mm diameter rod of total length 3.8m ,test length 3.0m. A hydraulic jack
was used for the loading device. The loading speed was 196 MPa/min., and the
sample was brought to breaking point by monotonous loading. The tensile load
was measured using a load cell, and the deformation of the tendon was calculated
from the values observed using electrical displacement gauges CD & @ in the
diagram. These displacement gauges were removed when the tension of the
sample had reached around 1400MPa. The test was conducted at room
temperature.
38
Mukae et al
The theoretical tensile strength and elastic coefficient of the AFRP tendon is
obtained from the following pair of equations:
(I)
(2)
where:
cr = tensile strength
E =elastic coefficient
V =mixing ratio
subscript t =total FRP
subscript f = fiber
subscript m = resin
The values obtained from the above equations for elastic coefficient and tensile
strength were Et = 48,600MPa, and crt= 2,260MPa respectively.
3.2 Test Results
Tension test was carried out on a total of 54 samples, each of which was made to
fail by rupture of the rod. In 78% of the samples, the failure point was in the
vicinity of the anchorage. This is considered to be caused by the concentration of
stress in the rod in the vicinity of the anchorage. The histogram for the tensile
strength obtained in this test is shown in Fig.3, and the normal probability
distribution is shown in Fig.4. It can be said that the distribution of the tensile
strength of AFRP is very near to the normal distribution. The average tensile
strength was crav=l,885MPa, or 84% of the theoretical tensile strength. The
standard deviation was 45MPa. AFRP tendons were found to possess
approximately the same tensile strength as PC steel strands.
As can be seen from Fig.5, the graph of the stress-strain relationship shows a
tendency to dip around 0.2 craV' and to bulge around 0.5-0.8 crav· However, this
tendency is very slight, and on the whole it may said that the graph follows a
straight line. The average elastic coefficient calculated from the values observed
between 0.02 crav and 0.8 crav was 53,000MPa.
FRP Reinforcement
39
3.3 Setting of Standard Values
Standard values for FRP tendons were calculated from the results of the static
tension test according to the following equation:
(3)
where:
<Ju =standard values for FRP tendons
<Jav =average values obtained from tension test
K =constant
<Jsd = standard deviation in tension test
Constant K here is a value for determining reliability. The standard tensile
strength of FRP tendons was taken to be the value obtained at the limit of
reliability (99.87%, when K=3). The standard tensile strength thus obtained ( <Ju)
of I ,750MPa was 77% of the theoretical tensile strength.
4. CREEP TEST
4.1 Test Method
This test was perfonned on a plumb lever type creep testing device, as shown in
Fig.6. Since an AFRP rod has an elastic coefficient approximately 1/4 of that of
steel, the elongation on loading is proportionately greater. The creep testing
device was therefore arranged so that its lower support could be moved to
accommodate the deformation of the tendon during loading. The total length of
the test sample was 0.90m, and the test length 0.30m, as shown in Fig.7. The
test was conducted in a room maintained at a constant temperature of20± 3°C.
4.2 Test Results
When AFRP tendons are exposed for long periods to loads smaller than their
tensile strength, creep rupture occurs. The relationship between load stress and
creep rupture time is shown in Fig.8. The rupture time varies with the load, the
time elapsing before rupture being inversely proportional to the size of the load
stress. When time is shown on a logarithmic scale, the graph of the relationship
40
Mukae et al
between load stress and time becomes a nearly straight line. The experimental
equation for this is shown below (equation 4):
cr = 1670- 102 log t
(4)
where:
cr =load stress (MPa)
t =creep failure time (hours)
This experimental equation was derived from observations made of the sample at
creep stress levels of between I ,390MPa and I ,770MPa. Assuming this
relationship holds even at creep stress levels below this range, the stress level at
which creep failure occurs in I 00 years time is estimated at I ,060MPa, or 61%
of the standard tensile strength.
5. FATIGUE TEST
5.1 Test Method
Fatigue test was conducted on samples of identical length to those used for the
creep test shown in Fig.7. The test was conducted on a mobile fatigue testing
device with maximum load ±20tf, as shown in Fig.9. A standard loading speed
of 6Hz was adapted and the test was performed at varying stress double
amplitudes of 94%, 73% and 55% of the average tensile strength crav (Table I).
The S-N curve of steel samples exhibits bending when the number of cycles
reaches the range of 106-107, and there is a fatigue limit where rupture does not
occur at numbers of cycles above this range. In macro-molecular materials, the
fatigue limit is not as clear as it is with steel. In the present test, tension test was
conducted to measure the tensile strength of samples which did not rupture even
after 2 million cycles.
5.2 Test Results
The S-N curve for AFRP tendons is shown in Fig. 10. In series 2 and 3, there is
a tendency for the number of load cycles to decrease as the stress double
amplitude increases. In series 2, the samples can withstand 2 million cycles at
astress double amplitude of around 600MPa. The equivalent figure for series 3
was 520MPa.
FRP Reinforcement
41
In the case of series 2, at stress double amplitude of 608MPa a decrease in
strength to 94% of the standard tensile strength was observed after 2 million
cycles. As shown in Fig. II, other samples also showed a similar decrease in
tensile strength at high stress double amplitude. It is assumed that fatigue failure
does not occur after a particular number of cycles, but that the overall
performance of AFRP tendons falls off gradually until failure occurs. The stress
double amplitude at which tensile strength fell below the standard level after 2
million cycles was 380MPa for series 2, and SSOMPa for series 3.
We shall now consider series 1. This series exhibits a different tendency to those
of series 2 and 3, in that the number of cycles to fatigue failure decreases in
proportion to the decrease in stress double amplitude. Fig.12 shows the results of
the fatigue test plotted with average stress [( crmax +crmin)/2] on the vertical axis,
against time to failure on the horizontal axis. The straight line in the graph
represents the creep failure curve obtained from the creep test. The samples in
series 1 are clustered around the creep failure curve, and these appear to have
suffered creep failure.
6. INVESTIGATION OF STANDARD TENSILE STRENGTHS
The standard tensile strength of AFRP tendons should be set with reference to the
static tensile strength, but also to the creep and fatigue characteristics. In AFRP
tendons, relaxation start at the time of prestressing. The stress levels of AFRP
tendons at the time of prestressing are set at a level of stress (below I 060 MPa)
where creep failure due to stress reduction caused by relaxation will not occur for
I 00 years time. The authors have derived the following expe1imental equation
from the test of the relaxation occurring in AFRP tendons (equation 5)[5]:
cr1 = crL (0.95 - 0.031 log t)
(5)
where:
crt= Tendon stress after t hours
crL = c · E
c =strain
E = elastic coefficient
t =time (hours)
The tension at the time of prestressing such that the tension of the AFRP tendons
in I 00 years time= I 060MPa is calculated according to this formula as 1380MPa.
This tensile strength is approximately 80% of the standard tensile strength.
42
Mukae et al
The fatigue strength of AFRP tendons is 380MPa on condition that maximum
stress ( crmax) is 1387MPa which is equivalent to prestressing stress. This value
rises to 550MPa on condition that crmax is 1040MPa which is equivalent to the
residual stress after I 00 years. This fatigue strength is superior even to that of
steel tendons (200-300MPa). It seems, also, that within the limits of practical
stress variations, falling off of tensile strength due to fatigue is not a significant
problem.
As a result of the creep and fatigue tests conducted, it was deemed appropriate to
set the standard tensile strength of AFRP tendons at 1750MPa. The tension
applied to the tendons at the time of prestressing may be approximately 80% of
the standard tensile strength.
7. CONCLUSION
Within the limitations of the tension, creep and fatigue tests conducted on Q>6mm
Technora® AFRP tendons at room temperature, the following conclusions may
be drawn:
(I) The average tensile strength of AFRP tendons is 1890MPa, equivalent to or
greater than that of PC steel strands.
(2) An equation was derived to express the logarithmic relationship observed
between the maintained stress and the creep failure time of AFRP tendons.
(3) The range of stress double amplitude between which AFRP tendons do not
rupture after 2 million load cycles is 500-600MPa. The stress double amplitude
range within which performance does not decrease under a maximum load of
1387MPa is 380MPa.
(4) As a result of the static tension, creep and fatigue tests conducted on samples
of AFRP tendons, the standard tensile strength was set at 1750MPa, and the
tension at prestressing at ~ 80% of the standard tensile strength.
8. REFERENCES
[I] Mizutani et a!., "The construction of a post-tension bridge with aramid
FRPs," 46th annual seminar of the Civil Engineering Institute of Japan, Sept.,
1990, V, pp.262-263.
[2] Mashiko et al., "The construction of a pretension bridge with aramid FRPs,"
46th annual seminar of the Civil Engineering Institute of Japan, Sept., 1990, V,
pp.264-265.
[3] Kobayashi et al.,"Prestressed Concrete using Fiber Reinforced Plastics
FRP Reinforcement
43
Tendon," Journal of Prestressed Concrete, Japan, Vol.30, No.5, Sept.-Oct.,
1988, pp.l9-26.
[4] Kosaka et al , "The anchoring of FRP rod for tendon with silent nonexplosive demolition agent," 44th annual seminar of the Civil Engineering
Institute of Japan, Oct.., 1989, V, pp.320-321.
[5] Asai et al., "Stress relaxation characteristics of aramid FRP rods," 46th
annual seminar of the Civil Engineering Institute of Japan, Sept., 1990, V,
pp.230-231.
TABLE 1 -FATIGUE TEST CONDITIONS
Maximum
Series
(MPa)
stress
crmax /crav
(%)
Stress double
amplitude
crmax -crmin (MPa)
Number
of
samples
I
1735
94
86
~
174
10
2
1387
73
139
~
694
17
3
1040
55
347
~
867
14
44
Mukae et al
Fiber
Round Securin
Fiber
Longitudinal Secturing Fiber
Fig. 1-Spiral wound rod (¢6 mm)
Displacement Gauges
AFRP
Tendon
2500
3000
3800
Fig. 2-Static tension testing device
Center Hole Jack
Anchoring
FRP Reinforcement
1750
1800
1850
1900
1950
2000
Stress (MPa)
Fig. 3-Histogram
II I I
, , ;. !
1950
: :r·:_
"2
:
Q.,
:
.
. «9
'
6
Vl
Vl
I .0cW~
1900
'
.
:,:
·:
1950
:
.
'
..
1900
..
1850
~-
~
ell
0
.
1850
<nS
~~
~
~:515 g~
Normal Probability
(%)
Fig. 4-Normal probability distribution
45
46
Mukae et al
2000~------~------~------~------~
(Eiastic Coefficient=53,000MPa~ ... ···
-....... -... --. --- .. --. -. -.. -. . .--- -... . . -.- -.------ -- .... .... -.-~
1500
i
,.
'6!
·····················-<-·····-··········---~~:
~1000
/:
[/J
···········~
500
/
0~
0
'
.......
______._______._______
~------~
2
Strain (%)
Fig. 5-Stress-strain curve
Fig. 6-Creep testing device
3
4
FRP Reinforcement
Bond Type Anchoring
AFRP Rod
/
1:
300mm
lillilillillillill
J :: J
Cross
300rnm
:1
Section
AFRP Rod
Fig. ?-Structure of test samples
2000~--~--~--~--~--~--~--~--~
1800
~
0..
6
0
Creep Rupture
e
Non Rupture
00.0
1600
·o ,
0\)0
"'"'
~
etv
:o
cil 1400
•.
1200
___.__~
104
1000~--~--~--~--~--~--_.
0.01
0.1
10
100 1000
Time (h)
Fig. 8-Creep rupture curve
47
48
Mukae et al
Fig. 9-Fatigue testing device
~:o,
·a'2 1000
Serise I
Serise 2
~----------------------
1:1
~
E
tl
100~----~------~----~----~------~
1000
Number of cycles to failure (cycles)
Fig. 10-S-N cutves
FRP Reinforcement
2000
~
~
~
c::
Q)
b
1800
C/l
~
·;;;
c::
Q)
f-o
1600
1400 L_....:::l::::::::::~L-~:C::::::::::~.l-~__J
0
200
400
800
1000
600
Stress Range (MPa)
(a max-a min)
Fig. 11-Tensile strength after 2 million loading cycles
versus stress double amplitude
2000
1800
;f
1600
N
1400
6
>.
.s
E 1200
t)
+
K
o:s
E
1000
t)
'-"
400~--~--~--._
0.01
0.1
10
__.___.___.___
100
Time
1000
I04
._~
105
(h)
Fig. 12-Average stress versus rupture curve
106
49
SP 138-4
Bond Characteristics of FRP
Rod and Concrete after
Freezing and Thawing
Deterioration
by M. Mashima and K. Iwamoto
Synopsis: Recently, a non-metallic reinforcement is
developed using new synthetic fiber, such as carbon,
aramid and vinylon fiber in Japan. The fiber is made
into a FRP rod. This material has advanced properties, for example, corrosion free, light weight and
high strength, and are expected to apply for the
practical structures. However, it is important to
study engineering properties and design method in
many fields theoretically and experimentally. In
present paper, the bond characteristics are discussed
because the expansion coefficient of non metallic
fiber is different from conventional concrete. The
results from the pull-out tests are, ( 1 )the bond
strength of FRP rod is ensured for the concrete
structures, and (2)the deterioration of bond property
is not appeared in CFRP, GFRP and VFRP however a
little reduction is observed at AFRP rod.
Keywords: Bond strength; bond stress; carbon; fiber reinforced plastics;
fibers; freeze thaw durability; glass; pullout tests
51
52
Mashima and Iwamoto
ACI Member,
is a Associate
Professor in the Civil Engineering Department at the
Osaka City University, Japan. His research interests
are in cementitious composite materials using synthetic fiber and durability of concrete structure.
Dr Mitsuyasu Mashima,
Mr Kaoru Iwamoto is a chief engineer in Development
Department of Kinki Concrete Industries Co., Ltd. He
has been associated with concrete reinforced by
chopped-fiber and/or fiber-tendon.
INTRODUCTION
The development of science and technology makes
rapidly progress, and the fruits
of innovation
appear in each academic field.
A high advanced and
new functional material has been under development to
use its superior characteristics such as high
strength, high elasticity, light weight and high
durability. There is a great interest for the civil
engineers to apply new material ~or practical construction works[l,2]. For example, a continuous fiber
reinforcement is manufactured by binding a highly
strong fiber with resin. This reinforcement, a FRP
rod is could be substituted for rebar and/or PC
tendon.
While the requirement for the infrastructure
becomes gradually high and severe, the criterion of
selection is changing from a cost supremacy to an
additional valuable function for the construction
material. The FRP rod discussed in present study is
one of typical functional material due to not only
high strength but also corrosion free, non-magnetic
property and lightweight feature.
All of structures consist of various structural
materials, and it is necessary for material to maintain safety during performance period. There is no
doubt that the FRP rod as a new material has a superior characteristic in relative short term. However
the durability against various kinds of deterioration
is not discussed sufficiently as not to pass enough
time to ensure after development.
The concrete structure always receives various
FRP Reinforcement
53
kinds of deterioration under the natural environment,
physically and chemically. The structure applied new
material is also under similar environment. From the
view point of weather condition, this study aims at
the influence of freezing and thawing action on bond
characteristics between concrete and FRP reinforcement.
RESEARCH SIGNIFICANCE
It is not necessary to take account of the
temperature stress in the designing of conventional
steel reinforced concrete structure, because the
coefficient of thermal expansion is almost similar
between rebar and concrete. However, the expansive
coefficient of FRP rod is different from concrete and
steel as shown in Table 1. And, the expansion coefficient even comes to negative in case of aramid FRP
rod. In addition, the resin as binding material has
very large expansion coefficient. Therefore, the
thermal stress acts at the interfaces between FRP rod
and concrete by heat change and affect the interactive properties.
Therefore, it is important to study an influence
of deterioration by a heat cycle in case of designing
concrete structure. This study investigates an influence of low temperature cycle under freezing, and
provides the useful information of bond property.
EXPERIMENTALS
Experimental outline
The object of present experimental study is to
obtain the change of bond characteristics by repeating freeze and thaw action. The bond strength is
measured by the pull-out test after a prescribed
freezing and thawing cycles in advance. The freezing
and thawing are carried out at 200 cycles or more to
confirm the durability. The curing period of specimen
is for 14 days after concrete placement and then a
freeze and thaw test is begun. Though the specimen
progress curing while freezing and thawing, the
influence is not taken account.
54
Mashima and Iwamoto
Bond Test
There are many factors influencing on bond
strength such as position and direction of rebar[3],
surface texture [ 4], embedded length [ 5], lateral
reinforcement for concrete[6] and testing method. One
of typical method on bond behavior is a pull-out
test though there are many testing methods proposed.
A compressive stress is acting in concrete in case of
pull-out testing method, and a stress state is different from the actual concrete structure. However,
the pull-out method
has the following advantages;
( 1) the dimension of specimen is small, ( 2) the
testing method is simple, and (3) the conventional
universal loading machine can be used for the pullout test without special equipment. Though there is
a little problem for a purpose of measuring the
actual bond strength, the pull-out test is convenient
for relative comparison like present study. The test
method of bond strength is specified as one of materials testing method in CPllO, ASTM and RILEM. However, there are some differences in details such as
specimen dimension, reinforcing method of concrete
and bond length[7].
The specimen is a cube of lOxlOxlOcm in present
experiment and FRP rod is embedded at the center of
specimen cube, as shown in Figure 1. The bond length
is four times of FRP rod diameter at the opposite end
of loading side. The rod is debonded with concrete at
loading side in order to reduce the influence of
stress disturbance on bond behavior by introducing
the load. There is no lateral reinforcement for
concrete from splitting failure. The testing method
of bond strength has not been standardized yet in
Japan, bu+ the method described above is based upon a
draft submitted to discussions for standardization.
The bond strength is calculated from following equation.
T=( P/4nD 2 )xa,
here;a=300/fc,
and
fc is compressive strength of concrete.
The pull-out load is not able to introduce into
FRP rod with the mechanical chuck like test of rebar
because FRP rod receives the damage for monofilament.
A coupler using the screw cut steel pipe is installed
to the FRP rod end, and then the load is introduced
through the tie-rod. A rubber mat is configurated at
the loading surface of specimen in order to distrib-
FRP Reinforcement
ute the load to specimen uniformly,
Figure 2.
55
as shown in
Freezing and Thawing Test
A testing machine based on ASTM C 666 specification gives repetition of freezing and thawing under
water. The specimen after freezing and thawing
cycles is used for the bond test specimen described
above directly. However the prism specimen of
10xl0x40cm is used to control the test temperature of
specimen as the temperature is need to measure at the
core of prism specimen.
PREPARATION OF SPECIMEN
Many FRP rods begin to be available in the
market. This is due to the variation of a fiber type,
a binding material and a manufacture processing
method. Among the variation of FRP, the following FRP
rods are used for experiment as shown in Table 2. The
typical engineering properties are listed in Table 2.
The subject of concrete class is for general
purpose structures. Then the objective slump ~s 7.5cm
and specified concrete strength is 280 kgf/cm . Table
3 shows the mixing proportion of concrete according
to design condition.
In the bond testing method specified, the position of rebar is recommended to be horizontal at
concrete casting. However, the position of FRP reinforcement is kept vertical in present experiment.
This is because of the prospective of low bond
strength for FRP rod and the attempt to reduce the
influence of bleeding. As the placing surface becomes
loading side in this case, the evenness of placing
surface affects on the strength characteristics. The
rubber mat has been installed at loading test in
order to avoid this phenomenon.
The test specimen is submitted to freezing and
thawing machine after curing for 14 days in water of
20 degrees in centigrade.
56
Mashima and Iwamoto
RESULTS AND DISCUSSION
Bond Characteristics
The bond characteristics of various FRP rod are
compared in Figure 3. Figure shows the relationship
between bond stress and slip displacement at free
end. This is the results before no freezing and
thawing action. Generally speaking, the rod of high
bond strength has small slip displacement. The smaller bond strength, the larger displacement at free
end. If the deformed processing of surface is appropriate, t~e bond strength is able to obtain almost
100 kgf/cm or more.
The surface ribs of deformed bar is caused by
mechanical interlocking resistance and consequently
the rebar indicates a sufficient bond strength[8].
However, There are many cases of splitting failure of
concrete during the pull-out test for rebar. The
debonding is suddenly occurred at small slip displacement. The rebar after bond test is shown in
Figure 4.
A sand coated carbon FRP indicates the highest
bond strength among experiments. The bond strength
is larger than rebar but slip displacement at free
end is the smallest. Two failure patterns are observed as shown in Figure 5. One is that the debonding occurs at the interface between CFRP rod and sand
layer, and sand leaves in concrete after slip out.
Other case is that the debonding occurs at the interface of concrete and sand layer. In later case, sand
coated rod of CFRP is pull out with mortar on the rod
surface. Former case seems to be larger than later
case, but the remarkable difference is not appeared
on the bond strength. The bond test result is relatively scattered as a results of ununiformly coating
of sand.
A coiled aramid FRP appears both high bond resistance and large slip displacement. The final
failure is occurred by pulling out of FRP rod with
the large damage of FRP. This indicates unevenness
processing of surface is working efficiently so much.
There are some cases that coiled fiber is separated
from rod though it is not completely. As the bond
function is not reducing so much, the frictional
resistance continues until the entirely slip out of
rod from concrete and consequently slip displacement
becomes large. Figure 6 shows the coiled type of AFRP
FRP Reinforcement
57
after bond test.
A glass fiber FRP rod also indicates the failure
by pulling out after appearing a large slip displacement. However, the lost of bond function is catastrophic due to the completely debonding of coiled
fiber from rod surface as shown in Figure 7. There
are a few case of debonding between concrete and rod
without separation of fiber coiled. The bond strength
is very low in this case.
A vinylon fiber FRP is also of coiled type. VFRP
rod appears similar behavior of GFRP rod, but slip
displacement is very small. As recognized in Figure
8, coiled vinylon fiber separate clearly from rod.
The excellent affinity of vinylon fiber does not
almost seem to be contributing to bond by the processing using resin.
There are many cases that the failure occurs by
the separation of fiber coiled from FRP. In case of
coiled type FRP, the lost of bond function appears
suddenly. This seems to be adhesive failure of resin
used. Considering these failure type, it is essential
that the selection of adhesion and the processing is
important factor for the bond characteristics. The
simultaneous processing of coiling is desirable at
the time of rod manufacturing.
The unevenness of surface is relatively rough in
a braided aramid FRP according to visual appearance.
However the bond strength is not so large. This is
due to the difference on surface unevenness from the
rib of rebar, and the unevenness structure of braided
rod is not to generate bearing resistance like rebar.
The bond strength seems to be influenced by the
longitudinal deformed convex in the braided AFRP.
However, rough leads to large frictional resistance
which is one of bond resistance. As a result, slip
displacement is large. This is recognized from the
fact that mortar attaches on the surface of braided
AFRP, shown in Figure 9. Final failure of bond test
is characteristic that concrete is splitting after
observation of large slip displacement. Therefore,
debonding of braided AFRP is catastrophic.
A strand carbon fiber FRP has very large unevenness on the surface, and it seems to have mechanical
bonding with concrete. However, the bond strength is
the smallest in the experiments, as shown in Figure
3. The structure if rod is twist and the surface of
strand CFRP is relatively smooth in micro structure
shown in Figure 10. This leads to low frictional
58
Mashima and Iwamoto
resistance. In addition, the mechanical bond resistance is also not so large due to the small longitudinal unevenness by strand process. Though the debonding occurs at early stage, the slippage progresses
without the decrease of bonding load caused by strand
structure until pull out.
Freezing and Thawing Deterioration
Figures 11 to 17 show a change of the bond
strength with the progress of freezing and thawing
cycles. In these figures, the result indicates the
relative strength for the specimen of no freezing and
thaw deterioration. This might be convenient to
compare the change of durability here.
The change of bond strength in rebar is shown in
Figure 11. No change seems to be recognized in the
bond strength between rebar and concrete according to
the results up to 200 freeze and thaw cycles though
there are wide scattering. In the case of rebar test,
there are some cases that the crack occurs during
freezing and thawing. In case of cracking, bond
strength becomes to be largely reduced. Therefore,
the crack occurrence of concrete causes large scattering of results.
Figures 12 and 13 show that the bond strength
both of glass and vinylon FRP seems not to be reduced
up to 300 cycles of freezing and thawing.
As for the coiled AFRP rod in Figure 14, a
little reduction of bond strength is recognized
according to the progressing cycles of freezing and
thawing. The crack is occurred in the specimen by
freezing and thawing over more than 150 cycles. Then
the distinct reduction of bond strength is observed
for braided AFRP at 600 cycles of freezing and thawing as shown in Figure 15.
The bond strength of the stand type of carbon
fiber FRP is small compared with other fibers. However Figure 16 shows that the bond strength is not
influenced so much by freezing and thawing. A crack
occurrence of concrete is
larger than other FRPs.
The strand CFRP might have no influence by freezing
and thawing deterioration in Figure 17.
FRP Reinforcement
59
CONCLUSION
The results obtained from present experimental
study are summarized as follows;
(1) The bond strength of an FRP rod of coiled type is
influenced by manufacture processing and materials
very much. Simultaneous processing is recommended at
the time of rod production using the same binding
resin.
(2) Unevenness of rod surface and processing method
influences to the bond strength.
(3) The unevenness processing of rod surface is
relatively
favorable condition, and the bond
strength of almost 100kgf/cm 2 can be obtained.
(4) Glass, vinylon and carbon FRP is not influenced
to the bond strength so much by freezing and thawing
deterioration.
(5) Aramid fiber FRP, both of braided and coiled
types, reduce the bond strength gradually with
progress of freezing and thawing.
REFERENCE
[1] K.W. Neale & P. Labossiere(Ed): Advanced Composite Materials in Bridges and Structures, Canadian
Society foe Civil Engineer, 1992, 705p
[2] JSCE Sub-committee on Continuous Fiber Reinforcing Materials(Ed): Proc. on Continuous Fiber Reinforcing Materials to Concrete Structures, Japan
Society of Civil Engineers, 1992, 314p
[3] Isteg-Stahl-Geselschaft: Erlauterungen zur Verwendung von Rippen-Torstahl, 2nd Edt., 1961
[4] C. Bach: Der Widerstand einbetonierten Eisen
gegen Gleiten in seiner Abhangigkeit von der Lange
der Eiseneinlage, Zeitshrift des Vereines Deutscher
Ingenieur, 1911, p.859
[5] P.H. Ferguson & J.N.Thompson: Development Length
for Large High Strength Reinforcing Bars, ACI Journal, Vol.62, No.8, 1965, pp933-951
[6] K. Ogura & T. Kameta: Study on Splitting Bond
Strength of Deformed-Bar, Proc of Annual Conference
of Architectural Institute of Japan, 1972, pp987-990
[7] K. Okada & H. Muguruma(Ed): Handbook on Concrete
Engineering, Asakura Shoten, 1981, pp.416-418
[8] L.A. Lutz & Gergley: Mechanism of Bond and Slip
of Deformed Bar in Concrete. ACI Journal, Vol.64,
No.ll, 1967, pp.711-721
60
Mashima and Iwamoto
TABLE 1 - THERMAL EXPANSION COEFFICIENT
Type
Expansion cgefficient
(lOxlO- /C)
Steel
Concrete
CFRP
GFRP
AFRP
VFRP
12
10
0.6 - 1.0
9 - 10
-6 - -2
4.4
TABLE 2 - FRP RODS USED AND ENGINEERING PROPERTIES
Type
Fiber
used
CFRP carbon( PAN)
CFRP carbon( PAN)
GFRP
AR glass
aramid
AFRP
aramid
AFRP
VFRP
vinylon
rebar
steel
Surface
processing
Tensile
Young's elong. fibe.
volum(
streng2h modulu~
(%)
(kgf/mm ) (kgf/mm ) (%)
sand coated
strand
coiled
coiled
braided
coiled
deformed
162
180
134
190
160
78
45
12,700
14,000
4,400
5,400
6,600
3,700
21,000
1.5
1.6
3.0
3.6
2.2
3.0
20-
TABLE 3 - MIXING PROPORTION OF CONCRETE
Max. size
of Agg.
(mm)
Slump
Air
W/C
s/a
(em)
(%)
(%)
(%)
20
7.5
5.0
51
52
Unit Weight(kg/m3)
w
c
s
G
171
335
888
840
Ad.*
0.067
*AE-Water Reducing Agent
53
64
65
65
65
72
FRP Reinforcement
61
reinforcement
0
0
~
I
o:::q--1.:.:.:.:~~.:::::.:.:.::+. - - - - - -
0
40
100
Fig. 1-Dimension of test specimen (unit: mm)
Spherical bed
Sp
Fastening nut
Fig. 2-Schematic view of pull-out test
62
Mashima and Iwamoto
300.---~----~-----------,-----,----.
CFRP(sand coated)
250
/X Rebar
_/({sp I itt i ng}
//1/
.t
/
C/l
C/l
~
;
co
1
GFRP
~-·:~.
---·················j··--.. --------........... j •.•
·:.-.:_,.. ,_
AFRP (fiber co i I ed)
/::·~-:~-=-~,rv""~-
.i/
! /1:. VFRP
......
{/)
-c
j
150
.-------------·
. ·"·' •••
100
I l'
/Y--~---,.
..
/1 ........
50
QL-----------------------------~--~
0
0.5
1.5
SI i p Di sp I acemen t
2
2.5
(mm)
Fig. 3-Bond stress-slip displacement at free end
3
FRP Reinforcement
Fig. 4-Rebar after pull-out test
l[tfltllHfi\fifllflliP.ff!T~fr~~tifHI flfll Ill HH!I ~J!Hfl! Ulllllll !Ill IIIIi 111111 i l!.!
"'· __,_, ..._,. .:,.1:1·:~~::}
_"____ "\,;·i·:·::-;iiJ!i
a
- -:1... ~,.~-
, ~
Fig. 5-CFRP rod (sand coated) after pull-out test
63
64
Mashima and Iwamoto
Fig. 6-AFRP rod (fiber coiled) after pull-out test
ll.lfllf!
'I"·
I;!
Fig. 7-GFRP rod after pull-out test
FRP Reinforcement
Fig. 8-VFRP rod after pull-out test
:t..
~-
...._.
,,
~
~
'"'-
''"~"" ·~<"'.:
Fig. 9-AFRP rod (braided) after pull-out test
65
66
Mashima and Iwamoto
Fig. 10-AFRP rod (strand) after pull-out test
150
Rebar
CD
0
~
c:r:
0
-
.s::.
Ol
c:
Q)
.....
GJ
0
0
8 CD
8
0
0
0
(j)
0
0 0
0
0
'U
c:
0
0
[(l
0
0
50
100
150
200
250
300
Freezing and Thawing Cycles
Fig. 11-Bond strength versus freezing and thawing cycles
FRP Reinforcement
150
GFRP
0
0
~
a:
()
100
0
-
.s::
CD
O'l
cQ)
....
Ci5
50
"U
c
0
[()
0
0
100
50
150
200
250
300
Freezing and Thawing Cycles
Fig. 12-Bond strength versus freezing and thawing cycles
150
VFRP
° 80
0
~
a: 1001 B 0 o.
.s::
a,
8
~·
8 8
§
i~
<r
c
Q)
....
Ci5
"0
c
50
0
[()
0
0
50
100
150
200
250
300
Freezing and Thawing Cycles
Fig. 13-Bond strength versus freezing and thawing cycles
67
68
Mashima and Iwamoto
150
AFRP(fiber coiled)
Do
0
~
a: 100(( D o
..r::.
c:
Q)
.....
Ci5
"0
c:
0
8
0,
0 0
0 0
0
§ 0
Q
0
0
50
0
8
0
0
[()
0
0
50
0
100
150
200
250
0
()
0
()
300
Freezing and Thawing Cycles
Fig. 14-Bond strength versus freezing and thawing cycles
150
AFRP(braided)
0
~
a:
-
100°
()
..r::.
©
8
0
0
0
0
Ol
c:
Q)
.....
(j)
"0
0
0
@
0
0
50
c:
0
8
0
0
0
()
0
0
0
[()
0
0
100
200
300
400
500
0
()
()
8
600
Freezing and Thawing Cycles
Fig. 15-Bond strength versus freezing and thawing cycles
FRP Reinforcement
Freezing and Thawing Cycles
Fig. 16-Bond strength versus freezing and thawing cycles
150
CFRP(sand coat e'd)
0
~
0:
..c
-rn
c
0
100
0
0
0
0
~
U5
"'0
c
0
50
[l)
0
0
100
200
300
400
500
600
Freezing and Thawing Cycles
Fig. 17-Bond strength versus freezing and thawing cycles
69
SP 138-5
Evaluation of FRP as
Reinforcement for
Concrete Bridges
by A.H. Rahman, D.A. Taylor,
and C. Y. Kingsley
.fu'.!!!!Psis: A comprehensive research programme to investigate the suitability of a
fibre-reinforced plastic (FRP) to reinforce concrete is described. The investigation
focuses on highway bridge decks and barrier walls. In determining the research
needs, careful consideration has been given to the loads and environment to which
highway bridges are subjected in northern North America. Short-term tension,
creep, fatigue and durability tests are being carried out on FRP specimens in the
first phase of the 3-phase programme. Tests completed so far indicate a small yet
noticeable change in strength and stiffness of the FRP with change in temperature,
small creep strain rates computed after 175 days of sustained loading, and
satisfactory fatigue behaviour under a tensile load cycling between 10% and 30%
of the tensile strength.
Keywords: Bridge decks; bridges (structures); creep properties; durability;
fatigue (materials); fiber reinforced plastics; reinforced concrete; tensile
strength
71
72
Rahman, Taylor, and Kingsley
Dr. Habib Rahman is a Research Associate in the Institute for Research in
Construction, National Research Council Canada. He received his doctorate in
1986 from Carleton University, Ottawa. For the past three years, he has been
conducting research on rehabilitation of structures, particularly on the evaluation of
deteriorated concrete slabs and FRP as reinforcement for concrete.
Dr. Don Taylor is a Senior Research Officer in the Institute for Research in
Construction, National Research Council Canada. He graduated from the
University of Toronto and received his Ph.D from Cambridge University in 1971.
Since 1973 he has conducted research on snow loads on roof, progressive collapse,
strength of window glass, and recently on the performance of FRP bars in concrete.
Mr. Charles Kingsley is a Technical Officer in the Institute for Research
in Construction, National Research Council Canada. Since graduating in Civil
Engineering from the University of Ottawa in 1991, he has been conducting tests
on deteriorated concrete slabs and FRP reinforcement for concrete.
INTRODUCTION
Corrosion of reinforcing steel is causing havoc in many concrete structures.
Concrete bridges in northern North America are among the most severely affected
structures. The measures developed so far to mitigate the problem have met with
varying degrees of success.
An effective solution would be an alternative reinforcement that does not
corrode. Being highly corrosion-resistant, fibre-reinforced plastic (FRP) has the
potential to be such an alternative.
FRPs are known to be not only highly corrosion-resistant, but also generally
much stronger in tension than steel. However, bridges in northern North America
offer a tough set of tests for FRP. The bridges are exposed to a harsh environment
characterized by a wide range of temperature and humidity, freeze-thaw action and
deicing salt. They are subjected to cyclic live load and impact in addition to
sustained dead load. Therefore, properties such as creep, fatigue and durability in
addition to bond and stiffness need careful evaluation before applying FRP to
bridges.
This paper describes a research programme now in progress to investigate a
proprietary fibre-reinforced plastic (FRP) for reinforcing concrete structures,
particularly bridge decks and barrier walls. Results obtained so far on the tensile
properties are also presented.
FRP Reinforcement
73
BACKGROUND
The well-known principal characteristics ofFRPs are excellent resistance to
corrosion, high tensile strength and low modulus of elasticity. Together, the last
two features make FRPs attractive as prestressing tendons. But in ordinary
reinforced concrete, a low modulus of elasticity of the reinforcement may
significantly increase deflection and cracking, particularly in large-span flexural
members such as building slabs [1]. In bridge decks, however, this may not be a
serious weakness since the slab spans are small and significant membrane action
can develop.
A review of available literature was canied out to determine research needs
for FRP as a reinforcing material in concrete bridges. Tensile properties of FRP
[2,3] and flexural behaviour of FRP-reinforced beams and strip slabs
[4,5,6,7,8,9,10] have been evaluated mostly under short-term and monotonic loads.
Fatigue and durability of FRP and FRP-reinforced concrete have been investigated
to a limited extent [10,11,12]. More study of creep, fatigue and durability
properties is required to establish the applicability of FRP to bridges.
RESEARCH OBJECTIVE AND SCHEME
A research programme was initiated with the general objective of
determining whether a proprietary FRP, NEFMAC, is suitable to reinforce concrete
bridge decks and barrier walls in Canada and the northem United States. In bridge
decks, FRP has to function under a temperature range of -30 to 50 oc with frequent
freeze-thaw action, in a saline-alkaline environment, and under cyclic traffic loads
for a desirable life-time of 50 years. FRP used to reinforce barrier walls must be
able to withstand vehicular impact loads under the same environmental conditions.
A 3-phase research programme has been developed to investigate the
suitability of NEFMAC as a reinforcement for concrete bridges. In Phase 1, the
basic mechanical and durability properties are to be determined by laboratory
testing. In Phase 2, finite element analysis of NEFMAC-reinforced bridge deck
slabs will be carried out to study various design parameters. Model testing of
bridge decks and barrier walls reinforced with NEFMAC will also be carried out
under simulated loading and environmental conditions. A field evaluation is
proposed in Phase 3, in which the performance of a NEFMAC-reinforced bridge
deck and a barrier wall will be monitored for at least 5 years; design criteria and
guidelines will also be developed in that phase.
THE MATERIAL
NEFMAC is a FRP produced as a two- or three-dimensional grid. Fig. 1
illustrates the construction of a 2-D grid. A characteristic feature of the
construction is that bond to concrete of any bar is achieved through mechanical
anchorage provided by the cross-bars. By varying the size of the bars and grid
74
Rahman, Taylor, and Kingsley
spacing, a wide range of demand of cross sectional area can be met. The fibres used
are mainly E-glass, carbon and aramid, and the matrix is a vinyl ester resin.
In this investigation, two types of2-D NEFMAC are being evaluated: Type
CG, reinforced with a mixture of glass and carbon fibres, and Type C, with carbon
fibres only. The total fibre content in either type is 36-38 percent by volume; in
Type CG, 74% of the fibres is glass and the rest is carbon.
TESTS OF BASIC PROPERTIES
Phase 1 of the research programme is currently in progress. A brief
description of the Phase-1 tests together with results obtained to date is given in the
following.
The primary objective of the Phase-1 tests is to determine the usable limit of
NEFMAC's strength for the intended applications. NEFMAC has high tensile
strength under short-term loading. Pullout tests show that embedment of only two
grid joints is adequate to develop the full strength of the bars [13]. However, the
maximum stress it can be subjected to under sustained, cyclic or impact loads under
the exposure conditions stated earlier is not yet known.
Thus short-term tension, creep, fatigue, and durability tests are being
conducted to evaluate the performance of NEFMAC under the anticipated
conditions of loading and environment. The tests are being carried out directly on
NEFMAC bars without encasing them in concrete so that unambiguous inferences
can be drawn about their performance. Number of specimens is five for each
short-term tension test and three for all other tests.
Short-term tension tests
Their purpose is to determine
i) ultimate strength, modulus of elasticity, and elongation under
short-term tension
ii) influence of ambient temperature in the range of -30 octo 50 oc on the
above properties
iii) influence of the number of bars in the specimen (width of the grid),
varying from 1 through 5, on the ultimate tensile strength.
In addition to determining the influence of ambient temperature and number
of bars in the specimen, the tests will provide a reference set of tensile properties
useful for evaluating the influence of sustained and cyclic tensile loads, and of
exposure to chemicals and UV radiation.
The tests are conducted under constant strain rate to reach failure within 2
to 3 minutes. Strains are measured by strain gauges and/or a LVDT attached to the
specimen by clip-on brackets. An environmental chamber is used to condition and ·
test the specimens under temperatures different from the room temperature.
FRP Reinforcement
75
Long-term tension tests
These are being carried out to determine
i) time-deformation (creep) behaviour under different levels of sustained
load
ii) effect of sustained load on tensile strength
iii) influence of elevated temperature (50 °C) on the behaviour under
sustained load.
Single-bar specimens are loaded to 10, 25, 40, 60 and 75 percent of the
short-term tensile strength for a duration of 10,000 hours (417 days).
Some FRPs not only creep but also lose strength under sustained load. The
specimens that remain unbroken after being under sustained load for 417 days will
be subjected to short-term tension test to evaluate their residual strengths. In one
set of tests, the specimens are being maintained at a temperature of 50 "C to
determine whether elevated temperature influences the behaviour under sustained
load.
Tensile fatigue tests
These are being carried out to determine
i) the relationship of load range to number of load cycles endured
ii) effect of cyclic loading on tensile strength
iii) effect of low temperature (-30 aq on the fatigue behaviour.
Tension-tension fatigue tests are being carried out on single-bar specimens
under 6 load ranges: 10-30, 10-40, 10-50, 30-50, 30-60, and 30-70 percent of the
tensile strength. Some of the lower load ranges may appear unusually low, but
stress in FRPs like NEFMAC in actual use may have to be restricted to rather low
limits considering their low modulus of elasticity and tendency to fail in a brittle
manner.
The loading cycle is repeated at a rate of 5 Hz up to 4 million times or until
the specimen fails, wbichever occurs earlier. A major bridge (Highway Class A,
OHBDC [14]) is designed to carry about 2 million trucks in its design life-time of
50 years. For a bridge deck, this is equivalent to 4 million axle load repetitions.
Specimens enduring 4 million cycles are tested in tension to determine the loss of
strength. One set of tests is being carried out with the specimens maintained at a
temperature of -30 acto determine the effect of low temperature on the endurance
limit.
Durability tests
These tests are being carried out to determine the effects of the following on
the short-term tensile strength:
i) saline, alkaline and saline-alkaline environment
76
Rahman, Taylor, and Kingsley
ii) freeze-thaw action
iii) exposure to ultra-violet rays.
Single-bar specimens are being exposed to saline and alkaline solutions
separately as well as to a combined solution. Some specimens exposed to
saline-alkaline solution are loaded in tension up to 60 percent of the ultimate
strength. Under the freeze-thaw tests, specimens are undergoing approximately 5
cycles of freezing and thawing per day. For UV exposure, a xenon-arc
weatherometer provides accelerated exposure.
The effects of these exposure conditions will be determined by evaluating
tensile strengths at intervals of 120, 240 and 360 days. Since bond ofNEFMAC to
concrete depends on the mechanical anchorage provided by the intersecting bars,
the effect of freeze-thaw action on the anchoring strength of the bar will also be
determined.
Other tests
The coefficients of thermal expansion in the temperature range of -30 octo
50 oc are being determined. Also being evaluated is the transverse shearing
strength of the NEFMAC bars as reinforcing bars are subjected to significant dowel
action across flexural or diagonal cracks in a bridge deck slab.
TEST RESULTS
A grip has been developed for testing single-bar specimens of NEFMAC in
tension, creep and fatigue (Fig. 2). It is made of a thick-walled intemally-threaded
pipe of 20 mm intemal diameter. The ends of the test bar are cast in epoxy resin
within the pipe. Bar specimens are cut out from the 2-D grid so that a 3 mm length
of the cross-bars protrudes from each side of the test bar. These protrusions were
found to be vital for successful anchorage. A 125 mm (2 grid junctions)
embedment was found adequate for the Type-CO bars in all tests except fatigue
which required 225 mm (3 grid junctions). For the Type-C bars, all tests required
225 mm embedment.
Short-term tension tests
These tests have been carried out on the Type-CO specimens at -30, -15, 0,
20, 35 and 50 °C. The results are shown in Table 1. It is seen that the coefficient
of variation, a measure of scatter in the test results, is remarkably low.
For room temperature tests, all five specimens were instrumented to
determine modulus of elasticity and elongation. For tests at the other temperatures,
only two specimens were instrumented for modulus of elasticity only. The moduli
of elasticity were computed from strain gauge readings and elongations from
LVDT readings over a gauge length of 100 mm. Four strain gauges were used in
tests at room temperature. Since strains at the centre of the grid were found
identical to those at the intersection, only the central gauges were used in tests at
FRP Reinforcement
77
other temperatures. Locations of strain gauges and the LVDT are shown in Fig. 3.
Volumetric method was used to determine the cross-sectional area of the
bars because of their irregular laminar construction. The average cross-sectional
area of 30 samples of the Type-CG bars is 109 mm2 , which was used to compute
the strength and modulus of elasticity shown in Table 1.
The effect of temperature on ultimate tensile strength and modulus of
elasticity is shown in Fig. 4. Except for the strength at 35 oc and the modulus of
elasticity at -15 °C, values of both properties decrease with increasing temperature.
It appears that the Type-CG NEFMAC loses about 9.5% of its strength and 2.5% of
its stiffness at 20 oc when the temperature rises to 50 °C. On the other hand, it gains
in strength by about 11% and in stiffness by 2.5% when temperature falls to -30 °C.
In Fig. 5, the load is plotted against the elongation recorded by the LVDT
during a typical test at room temperature. Although the actual trend of the falling
branch may not be exactly as shown because the brackets holding the LVDT were
disturbed by breaking of fibres, it is seen that the specimen does not break
completely after the peak load is reached but continues to elongate under falling
load as fibres break. While this post-peak elongation capacity provides no
significant benefit under slowly applied loads, a favourable failure behaviour under
impact loads can be expected.
Long-term tension (creep) tests
Single bar specimens are subjected to sustained loads with the help of dead
weights magnified by combined lever and pulley action. Two strain gauges are
mounted at the middle location (Fig. 3) of each specimen. Additionally, a LVDT
has been attached to one of every set of three specimens as a back-up. A
data-acquisition system has been recording the temperature and all strain gauge and
LVDT readings, initially every hour and subsequently once every day.
Fig. 6 shows typical time-strain plots for two specimens under loads equal
to 60% of their respective ultimate strengths at room temperature. The strains
represent the average of the two strain gauge readings. Both plots indicate mostly
sporadic but abrupt jumps in strains followed by gradual declines. An explanation
of such a behaviour is yet to be found.
Table 2 shows the strains occurring immediately after the loads were
applied and the computed creep strain rates at three time intervals for all specimens.
The creep strain rate at any time was computed as the slope of the best-fitting
straight line through the time-strain data points recorded up to that time. It is
observed that some of these rates are negative. While an exact explanation is yet to
be found, instrument error is suspected. Regardless of their signs, the creep strain
rates are very small. For example, under the 40% loading at room temperature, the
rate (algebraic average of three) at 175 days is 1.3 microstrains/day for Type-CG
and 0.79 microstrains/day for Type-C specimens. These rates amount to a 175-day
creep strain of about 2.7% of the initial strain for the Type-CG and 1.8% for the
Type-C. The creep strain rate is observed to increase generally with increasing
78
Rahman, Taylor, and Kingsley
level of sustained load. A general decline of the creep strain rate with time is also
observed.
Further analysis of the creep test data will be performed after the scheduled
10 000 hour (417 days) duration of the tests. Tension tests will also be performed
at that time on all the specimens remaining intact to determine if the sustained load
affected their tensile strengths.
Tensile fatigue tests
So far, fatigue tests at room temperature and under a load range of 10-30%
of the tensile strength have been carried out for both Type-CO and Type-C
specimens. All specimens (3 of each type) endured at least 4 million cycles without
failure except one Type-CO that failed at 862 000 cycles. An additional Type-CO
specimen was tested and found to last 4 million cycles without failure.
All specimens surviving the fatigue test were tested in tension to determine
their residual tensile strengths. Type-CO was found to retain 92% and Type-C 93%
of their respective original tensile strengths.
Coefficient of thermal expansion
The longitudinal coefficient of thermal expansion for the Type-CO
specimens was determined for four temperature intervals in the range of -30 octo
50 °C. Initially, the specimens were conditioned in a chamber at a temperature of
-30 oc for an hour. The temperature was then raised to -9 oc and maintained for
another hour followed by recording the change in length using L VDTs. The
procedure was repeated for the remaining temperature steps.
Fig. 7 shows the plots of thermal strain with temperature. It is observed that
the coefficient of thermal expansion, indicated by the slope of the lines, varies
almost linearly with temperature in the given range. The mean value of the
coefficient of thermal expansion is computed as 8.39 x 10·6 /°C from the slopes of
the best fitting straight lines.
CONCLUDING REMARKS
Before FRP materials are applied to bridges as reinforcement for concrete,
creep, fatigue and durability properties of the materials must be evaluated. Very
little is known so far about these properties. Under the present research
programme, these properties are being investigated in a comprehensive manner.
The following observations are summarized from the tests completed so far:
L The ultimate tensile strength and modulus of elasticity of Type-CO
NEFMAC decrease by 9.5% and 2.5%, respectively when temperature rises to 50
oc from room temperature. On the other hand, strength increases by 11% and
modulus of elasticity by 2.5% when the temperature falls to -30 oc.
2. Creep deformation in NEFMAC is very small. Under a tensile load
FRP Reinforcement
79
equal to 40% of the tensile strength sustained for 175 days, the Type-CO and the
Type-C undergo creep strains which are 2.7% and 1.8% of their respective initial
strains.
3. Under tension-tension fatigue load cycling between 10 and 30% of the
tensile strength and applied at a frequency of 5 Hz, both types of NEFMAC can
endure 4 million cycles. After enduring 4 million cycles, the Type-CO retained
92% and Type-C 93% of their respective original tensile strengths.
4. The coefficient of thermal expansion of the Type-CO is reasonably
COnStant in the temperature range Of -30 °C tO 50 °C and is about 8.39 X 10"6 fOC.
REFERENCES
1.
Rahman, A.H. and Taylor, D.A., Deflections of FRP-Reinforced Slabs-A
Finite Element Study, Proceedings of the First International Conference on
Advanced Composite Materials in Bridges and Structures, Sherbrooke, 1992,
pp. 607-616.
2.
Pleimann, L.G., Strength, Modulus of Elasticity, and Bond of Deformed FRP
Rods, Proceedings of the ASCE specialty conference Advanced Composites
Materials in Civil Engineering Structures, Las Vegas, 1991, pp. 99-110.
3.
Porter, M.L. and Barnes, B.A., Tensile Testing of Glass Fibre Composite Rod,
Proceedings of the ASCE Specialty Conference on Advanced Composites
Materials in Civil Engineering Structures, Las Vegas, 1991, pp. 123-131.
4.
Nawy, E.G., Neuwerth, G.E. and Phillips, C.J., Behaviour of Fibre-Glass
Reinforced Concrete Beams, Journal of the Structural Division, ASCE, Vol.
97, No. ST9, 1971, pp. 2203-2215.
5.
Nawy, E.G. and Neuwerth, G.E., Fibre-Glass Reinforced Concrete Slabs and
Beams, Journal of the Structural Division, ASCE, Vol. 103, No. ST2, 1977,
pp. 421-440.
6.
Bank, L.C., Xi, Z. and Mosallam, A.S., Experimental Study of FRP Grating
Reinforced Concrete Slabs, Proceedings of the ASCE Specialty Conference on
Advanced Composites Materials in Civil Engineering Structures, Las Vegas,
1991,pp.lll-122.
7.
Faza, S.S. and GangaRao, V.S., Bending Response of Beams Reinforced with
FRP Rebars for Varying Concrete Strengths, Proceedings of the ASCE
Specialty Conference on Advanced Composites Materials in Civil Engineering
Structures, Las Vegas, 1991, pp. 262-270.
8.
Satoh, K., Kodama, K. and Ohki, H., A Study on the Bending Behaviour of
Repaired Reinforced Concrete Beams using Fibre Reinforced Plastic ( FRP)
and Polymer Mortar, Proceedings of the ACI International Conference on
Evaluation and Rehabilitation of Concrete Structures and Innovations in
Design, Hong Kong, 1991, pp. 1017-1032.
80
9.
Rahman, Taylor, and Kingsley
Saadatmanesh, H. and Ehsani, M.R., Fibre Composite Bar for Reinforced
Concrete Construction, Journal of Composite Materials, Vol. 25, 1991 pp.
188-203.
10. Ozawa, K., Sekizima, K. and Okamura, H., Flexural Fatigue Behaviour of
Concrete Beam with FRP Reinforcement, Transactions of the Japan Concrete
Institute, Vol. 9, 1987 pp. 289-296.
11. Goodspeed, C., Schmeckpeper, E., Gross, T., Henry, R., Yost, J. and Zhang,
M., Cyclical Testing of Concrete Beams Reinforced with Fibre Reinforced
Plastic ( FRP) Grids, Proceedings of the ASCE Specialty Conference on
Advanced Composites Materials in Civil Engineering Structures, Las Vegas,
1991, pp. 278-287.
12. Chaallal, 0., Houde, J., Benmokrane, B. and AYtcin, P.C., Use of a New
Glass-Fibre Rod as Reinforcement for Concrete Structures, Proceedings of the
ACI International Conference on Evaluation and Rehabilitation of Concrete
Structures and Innovations in Design, Hong Kong, 1991, pp. 515-528.
13. Nefcom Corporation, NEFMAC-Technical Leaflet 1 (undated).
14. Ministry of Transportation Ontario, Ontario Highway Bridge Design Code,
1983.
ACKNOWLEDGMENT
This research was carried out in collaboration with Autocon Equipment Inc.
and the Industrial Research Assistance Programme of National Research Council
Canada. The authors also acknowledge the collaboration of Dr. Michael Lacasse
and Mr. Jim Margeson of IRC's Materials Laboratory in setting up some of the
durability tests.
FRP Reinforcement
TABLE 1 -TENSILE PROPERTIES OF TYPE-CG NEFMAC
Test
temperature,
Specimen
No.
Ultimate
strength,
MPa
Modulus of Elongation
elasticity,
at ult. load,
GPa
percent
-30
1
2
3
4
5
849
888
884 882
916 (2.5)
871
*
*
*
871
881
833 864
849 (2.3)
886
42.7
45.4
-15
1
2
3
4
5
869
849
800 843
845 (2.7)
853
41.8
42.3
0
1
2
3
4
5
20
1
2
3
4
5
837
808
794 797
826 (5.0)
722
41.7
40.9
42.2 41.4
42.0 (1.6)
40.4
2.10
2.40
1.95 2.07
2.05 (9.0)
1.85
35
1
2
3
4
5
830
833
826 823
830 (1.8)
794
*
*
*
*
*
*
*
*
oc
1
2
3
4
5
42.5
*
*
*
*
*
44.1
*
*
*
*
*
42.1
*
*
*
*
*
41.9
43.1
*
*
*
*
*
*
40.8
40.8
40.8
679
*
706
*
50
730 722
40.4
*
759 (3.7)
40.7
733
40.0
* not scheduled to be determined
Figures in bold face indicate mean values
Figures in parentheses indicate coefficient of variation, %
*
*
*
*
*
81
00
TABLE 2- CREEP STRAIN RATE
Test
temp.,
Load',
oc
%
Type-CG
Specimen
Nominal
Actual load ! Initial
kN
%' I micro! strain
i
No.
I
:
9.0
'
10.4
Type-C
Creep strain rate,
microstrain I day
-4.41
-7.33
2270
0.60
0.30
2410
2.86
1.80
I
------
3
~---
----~----
' _____ . __ l
- - - - - - ________ j_
I
I
25
2
I
24.6
21.4
I
Actual load ! Initial
kN
'
i
strain
50 days
i
2026
0.10
0.22
I
2014
-0.67
-0.51
-0.27
!
2012
-0.75
-0.54
-0.29
i
I
I
1.15
I
I
5261
5.23
3.51
5030
-0.27
0.26
0.10
----
-·
5030
-3.65
---.
-0.64
-0.09
-~·
I
20
40
2
41.0
35.6
i
!
6.5
I
9.6
------
1.95
I
3
~~~~
-
·- ·------i
I
60
2
3
I
75
2
3
I
I
52.5
60.4
I
8790
3.07
1.52
8308
7.05
3.60 I
8422 '
3.92
11902f 13.08
16.42
12418
13500_,2
I
16.1
23.7
1.78
2.80
I
15211
12.61
7.61
66.6 i 76.6
15363
:
I
15572
25.44
10.79
I
I
11.46
1.62
2.22 '
!
7.32 :
18.98 2
:.
0.48
0.90
27.82 2
.
I
i
i
!
I
I
!
3.98
1.51
I.IO
0.56
4702
0.95
II 0.07
I
0.23
i
5.85 I
3.41
4509
I
2.75
0.84
1.33
6983
I
0.09
-0.77
-0.75
1.59
0.71
0.38
2.57
2.75
7.44
3.06
1.48
58.6 ! 10858
6.80
4.24
2.33
10934
-4.69
-3.23
-0.71
! 71.6
I
!
5.29
I
48.6
0.30
110936
4.35
5.90
0.13 I
37.6 I 6933
! 7150
'
39.8
100 days 175 days
I
4461
I
25.5
11.592
percent of ultimate load capacity @ 20oC; 86.9 kN for Type-CG, 67.9 kN for Type-C
values from one strain gauge only
* specimen pulled out of grip
1
2
microstrain I day
' -2.39
I
3
-----
Creep strain rate,
%' ! micro-
50 days · I 00 days i 175 days
2185
2
10
N
13267
4.82
0.02
0.56
13548
13.61
*
*
*
*
*
*
FRP Reinforcement
100 mm x 100 mm grid
_JUUUUUL
JDDDDDC
[email protected]~
DOC
Ill
nnrI
Section a-a
Plan of 2-D NEFMAC grid
Fig. 1-Details of 2-D NEFMAC grid construction
Threaded rod to machine grip
'
A
lntemally threaded
pipe of 20 mm int. dia.
E
E
~
Section A-A
Vertical section through grip
Fig. 2-Details of grip for tension, creep, and fatigue tests
83
84
Rahman, Taylor, and Kingsley
2 additional strain gauges
in tests at room temperature
2 strain gauges (on opposite faces)
in tests at all temperatures
I
Fig. 3-Strain gauge and LVDT locations on NEFMAC specimen
45
900
- - a - Tensile strength
ell
0..
::;s
..d
44
---t:r-- Modulus of
850
0
elasticity
'5!:I'n
;6
43
~
:g
Vl
Vl
~
Vl
ell
0..
ell
800
03
..._
!:I
42
1l
0
Vl
::l
'"3
~
.§ 750
"0
0
41
§
700
::;s
40
-40
-30
-20
-10
0
10
Temperature,
20
30
40
50
60
oc
Fig. 4-Effect of temperature on tensile properties of Type-CG NEFMAC
FRP Reinforcement
85
100
90
80
70
fJ
60
-d'
(1j
50
0
....l
40
30
20
10
0
0
2
3
4
5
Extension, mm
Fig. 5-Typical load-elongation curve for Type-CG NEFMAC
14000
12000
10000
c:
·c;:;
b
~
r-
8000
"'.....
0
u
~
6000
---TypeC
4000
---TypeCG
2000
0
0
20
40
60
80
100
120
140
160
180
Time, days
Fig. 6-Time-strain behavior at room temperature under a sustained
load of 60 percent of the ultimate tensile strength
86
Rahman, Taylor, and Kingsley
1000
900
800
- D - - sample I
700
------fr----
c::
'til
600
---<>---- sample 3
"'0....
500
sam pIe 2
.b
u
::§ 400
300
200
100
0
-40
-20
0
Temperature,
20
40
ac
Fig. 7-Thermal expansion of Type-CG NEFMAC
60
SP 138-6
Creep Rupture Behavior of
FRP Elements for Prestressed
Concrete - Phenomenon, Results
and Forecast Models
by H. Budelmann and F.S. Rostasy
Synopsis: FRP tensile elements exhibit - when subjected to a
high axial tensile stress - the so-called creep rupture phenomenon. Thereby, the time of endurance until fracture dependent
on the level of the permanent stress is the relation to be experimentally derived. The creep rupture phenomenon principally
exists for all structural materials. Experiments, however,
prove that for prestressing steel it is of no practical relevance: the usual permanent steel stresses which are in the
range of 75% of its characteristic tensile strength can be
borne indefinitely without fracture or strength loss. This is
however not the case for FRP, whose stress rupture behaviour is
also markedly influenced by the micro-environment around the
element and dependent on the type of fiber and matrix employed.
The paper presents an outline of the results known so far, of
the experimental techniques, of the methods of statistical evaluation and the forecast of the long-term behaviour of specific
FRP elements. It is shown that the characteristic stress rupture line is the essential basis for the derivation of the admissible permanent prestress of FRP tensile elements.
Keywords: Creep properties; fiber reinforced plastics; fracture properties;
models; prestressed concrete; rupture; statistical analysis; strength; tensile
strength; tensile stress
87
88
Budelmann and Rostasy
Harald Budelmann, born 1952, studies of civil engineering at
the Technical University Braunschweig, Ph.D. 1986, senior research engineer at the TU Braunschweig until 1991, today professor of structural materials at the University of Kassel,
Germany.
Ferdinand S. Rostasy, born 1932, studies of civil engineering
at the University of Stuttgart; Dr.-Ing. 1958; practical work
and research until 1976, from then on professor of structural
materials at Braunschweig, Germany.
INTRODUCTION
Tensile elements used for the prestressing of concrete members
are primarily subjected to long-term static stresses. Hence,
the static long-term strength must be known. FRP elements exhibit the phenomenon of creep rupture when subjected to a constant tensile force. The endurance time is a function of the
stress level, the micro-environment around the element and the
temperature. For the stipulation of the admissible permanent
tensile stress of FRP elements the phenomena of creep rupture
and of strength retention must be taken into consideration.
The problem of delayed failure of, especially, FRP md-compos i tes has been dealt with for many years by the aerospace/
mechanical industries. Today, the mechanisms of the time- and
stress-dependent damage and of the delayed failure are partly
understood. With the help of micromechanical approaches and numerical procedures a viable prediction can be achieved; for example (11, 12).
It is not the aim of this paper to discuss the phenomenon of
creep rupture of FRP fundamentally or to improve the prediction
models known so far. Based on the requirements of structural
engineering, proposals for the necessary test work and for the
derivation of the permissible prestressing force from test results are presented. An outline will be given on test results
known so far, on experimental techniques, on the evaluation of
the tests and on the prediction of the endurance.
FRP tensile elements for prestressing consist of unidirectional, endless glass, aramid or carbon fibers embedded in a polymeric matrix resin. A review of the properties of FRP concerning their use for structures is given in ( 1). Another paper of
this symposium (2) contains further information on FRP materials and their application so far.
FRP Reinforcement
89
PHENOMENON OF CREEP RUPTURE
The creep rupture phenomenon pri nc i pall y exists for all s truetural materials. However, for prestressing steel it is of no
practical relevance. The usual permanent steel stresses in the
range of 75% of the characteristic tensile strength can be
endured indefinitely without fracture or strength loss.
This is however not the case for FRP. A FRP tensile element
subjected to a constant force Fcl in a specific environment,
with
( 1)
may fail after a time t as shown in Fig. 1, if the force F 1
is high enough. Feu is tthe relevant short-time static tensire
rupture force of the element considered.
Prior to failure at tu, fiber fractures and microcracks in the
polymer matrix followed by debonding will occur, rapidly
growing in the final phase. The time-dependence of the total
strain mirrors the accumulation of damage prior to the failure
of the tensile element.
The higher the sustained tensile force of a FRP element in relation to the short-time static rupture force, the shorter the
endurance time tu until failure will be. Therefore, for the
stipulation of Uie permissible prestress the presumptive service life span ts of the structure has to be decided. The relevant resistance of the FRP material against long-term static
stress is the characteristic tensile rupture force Fclk (tsl at
the end of service life.
For the prediction of endurance, extensive testing is necessary. The aim of the tests is the determination of the characteristic endurance line Fclk (tu), depicted schematically in
Fig. 2. Since the endurance lime t will scatter strongly (coefficients of variation of 50 to 1Ltlo% have been observed), a
sufficient number of creep rupture tests on distinctliy different force levels Fcli must be performed. By statistical evaluation of the test results the mean endurance line Fcl (tul
(survival probability of 50%) or the characteristic enl:mrance
F lk (t) (e.g. survival probability of 95 %) can be determine~. Wit~ the characteristic endurance line the characteristic
rupture force Fclk (tsl can be predicted.
As mentioned above, creep rupture is accompanied by a cumu l ative damage, commencing some time prior to failure. For the ultimate limit state design the virgin tensile strength of the
FRP should be available during service life span ts without any
significant loss. Also this demand has to be consiaered for the
stipulation of the permissible prestress. If in a creep rupture
90
Budelmann and Rostasy
test the sustained force Fcl is removed prior to creep rupture
at the time te, the residual tensile rupture force Fer can be
determined (s. Fig. 2).
EXPERIMENTAL TECHNIQUES AND TEST RESULTS
Experimental Techniques
The set-up for creep rupture tests with FRP bars must meet the
following requirements:
1. A long-term static tensile force nearly up to the short-time
rupture force of the FRP bar must be produced and kept constant.
2. A single bar anchorage with a high efficiency both
short-time loading and long-term loading is needed.
for
3. A substantial part of the bar tested should be exposed to
the relevant environment.
4. The time dependent deformation of the bar should be measured
along a free length, in contact with the medium of environment.
The environment is of great influence on the creep rupture behaviour of the most FRP materials (3). The FRP suited for preand post-tensioning of concrete members will not be damaged by
ions from usual acidic solutions. But more import is the influence of alkaline solutions with pH~ 12.5, e.g. fresh concrete
or the pore solution of hardened concrete. By testing the FRP
in cementitious solution a potential weakening effect of moisture is simultaneously incorporated.
Several types of test set-up in air have been developed in the
past, e.g. (4, 5). A very useful dead load lever system has
been deve 1oped at the TU De 1ft ( 6) for creep tests with a rami d
bars in an alkaline solution.
A principle sketch of the test set-up recently realized at the
TU Braunschweig is given in Fig. 3. Fig. 4 shows the complete
test rig. The load is kept constant by a steel spring package.
A climate vessel, sealed against the bar with a rubber bellow,
contains the alkaline solution. The anchorage devices are placed outside the liquid. For the deformation measurement of the
bar in the solution a LVDT is used, placed above the liquid.
Creep tests with various FRP materia 1 s within an EC- founded
joint research project are under way (7).
FRP Reinforcement
91
Test Results
Some years ago numerous creep rupture tests have been performed
for various epoxy- or polyester-impregnated bundles of glass
and aramid fibers, e.g. (4, 8). Most of the tests were executed
in air at 20 ·c. The results cannot be applied to other environmental conditions.
In Germany comprehensive tests for GFRP bars in air at 20 ·c
have been performed (5). The mean curves and the range of scatter of test results are drawn in Fig. 5.
Arapree strips, consisting of aramid fibers in epoxy resin, are
mainly used for pretensioning. Therefore, long-term loaded AFRP
elements were subjected to alkaline solution (6). It should be
noted, that the synthetic alkaline solution used (sat. Ca(OH)zsolution + 0,4 n KOH-solution) is more aggressive to the aram1d
fibers than the normal concrete pore solution. In Fig. 6 test
results are depicted. The evalution of the creep rupture lines
drawn in Fig. 6 will be dealt with in chapter 4.
Systematic creep rupture tests with CFRP in a relevant environment are not known. AFRP are less sensitive to an alkaline medium than GFRP; CFRP show the least sensitivity. Today's state
of experimental knowledge on the creep rupture behaviour of FRP
is insufficient.
LONG-TERM STRENGTH AND STRENGTH RETENTION
Prediction of Endurance
The service life time ts of a structure will certainly ~pan several decades; e.g. 50 years which correspond to 5 · 10 hours.
Long-term creep rupture tests rarely exceed a duration of ten
to twenty thousand hours. Hence reliable prediction becomes necessary. In the literature several theoretical models are reported.
By application of the fracture mechanics concept for the timeand stress-dependent growth of damage the endurance can be expressed as (9):
l
~~
Feu
[s
t
u] - ~
(2)
to
With t 0 ~ l h t in hours; 8 is a flaw size parameter relevant
for the describe~ material and element; n is a material parame-
92
Budelmann and Rostasy
ter. Both, Band n must be derived on basis of tests. Eq.(2)
represents a straight line in a double ln-plot. Of course,
Eq.(2) is a deterministic formulation. For the analysis of reliability the scatter of Fcl, Feu and tu must be taken into account.
In Fig. 7 the evaluation of Eq. (2) is shown for a set of material parameter B and n, which were derived from tests with GFRP
bars (10). Both, the creep rupture force Fcl and the strength
retention force Fer are normalized by the mean short-term rupture force Fem of the GFRP material . The example shows, that a
perman5nt force of 0,65 Fern would lead to failure after about
5 . 10 hours of loading.
Strength Retention
The strength retention is the residual tensile strength Fer of
an element after removal of the sustained loading at the time
t , prior to creep rupture. Basing on Eq.(2) and with a suitaBle damage criterion in (9) an analytical expression for the
strength retention was derived:
( 3)
For
the
example
given
in
Fig.
7 -
0,65 Fern' leading to failure after tu
=
a
per~anent
force of
hours - the rewith this formula. Da-
5 · 10
sidual strength Fer has been ev~uated
mage developes rap1dly if te > 10 hours.
For a stipulated service life time t
and for a prescribed
strength retention (e.g. Fer/Fern~ 0,95T the required endurance
time tu ~ ts can be estimated:
( 4)
Then, with the help of Eq. (2) the greatest possible Fcl to guarantee the required endurance can be found.
If a constant action Scl to a structure is assumed, this action
must be less than Fcl· Deliberations of this kind may serve as
basis for the development of the design resistance.
FRP Reinforcement
93
Statistical Evaluation
Creep rupture tests show, that the endurance times tui of the
specific force Fcli scatter considerably. Generally, for the
statistical evaluation several density functions h(tui) are
suited, e.g. Gaussian, Weibull etc. The result of the evaluation are characteristic endurance times tuki on the chosen
force levels Fcli· Using Eq.(2) the characteristic endurance
line can be expressed:
nk
(5)
In Fig. 6, both the mean creep rupture line and the one with a
survival probability of 95% were evaluated and drawn according
to Eq.(2), resp. Eq.(5).
Certainly, also other formulations can be chosen (4).
For the stipulation of the permissible force on basis of partial safety factors, the coefficient of variation vel of Fclk
must be known. Since it hardly can be derived from creep rupture test results, an approximative method has been presented
in (3). Assuming a ln-normal distribution of the endurance
times on a force level and a parallel position of Fclk (tuk)
and Fclm(tuml to each other in a double-ln-plot (as shown by
tests), the coefficient of variation v 1 of the long-term
strength can be deduced from the known coefficient of variation
of the endurance time vt:
(6)
with k the factor of the t-distribution. With this approximative transformation of the coefficient of variation, still to
be verified by experiments, the influence of the variability of
the endurance times on the variability of the long-term
strength can be investigated.
Necessary Test Work
The creep rupture behaviour of FRP bars is important for the
stipulation of the admissible prestress. For the prediction of
endurance, extensive testing in necessary.
94
Budelmann and Rostasy
The characteristic short-term tensile strength Fck must serve
as a basis for creep rupture tests. Creep rupture behaviour
should be investigated on at least three distinctly different
force levels Fcli' e.g.: 0,8/0,7/ 0,6 Fck' see Fig. 2. The
force is maintained constant until break occurs, the fracture
times tui are being r~corded. The range of endurance times tu
should cover up to 10 hours. Experience show that at least 1u
to 15 results tui per force level Fcli are necessary.
The time dependent strain development should be man ito red in
order to get information on the damage process. In addition a
number of creep rupture tests should be finished prior to damage and the residual strength should be determined.
Of course, the relevant environmental condition of the bar in
the structure has to be considered in the tests.
CONCLUSIONS
Long-term static tensile stresses reduce the tensile strength
of FRP and may lead to creep rupture. For the stipulation of
the admissible prestress of FRP elements the phenomenon of
creep rupture and strength retention is decisive. The creep
rupture behaviour of FRP must be investigated by tests. By
means of statistical evaluation of the test results and with a
prediction model the characteristic tensile rupture force at
the presumptive end of service life can be estimated. An example of this procedure is given. Besides, a glance has been
thrown at the creep rupture phenomenon, at experimental techniques and test results.
LITERATURE
(1)
Rostasy, F.S.; Budelmann, H.; Hankers, Ch.: Faserverbundwerkstoffe im Stahl beton- und Spannbetonbau. Beton- und
Stahlbetonbau 87(1992), Heft 5 und 6.
(2)
Rostasy, F.S.: FRP Tensile Elements for Prestressed Concrete Structures - State of the Art, Potentials and Limits. ACI Int. Symposium on FRP Reinforcement for Concrete Structures, Invited Paper, Session III, March 1993,
Vancouver.
(3)
FIP Commission on Prestressing Materials and Systems:
High Strength Fiber Composite Tensile Elements for Structural Concrete. State-of-Art-Report. July 1992, unpublished.
FRP Reinforcement
95
(4)
Chiao, T.T.; Wells, J.E.; Moore, R.L.; Hamstad, M.A.:
Stress-Rupture Behaviour of Strands of an Organic Fiber/
Epoxy Matrix. 3rd Conf. on Composite Materials: Testing
and Design, ASTM STP 546, 1974, pp. 209/224.
(5)
Rehm, G.; Franke, L.; Patzak, M.: Zur Frage der Krafteinleitung in kunstharzgebundene Glasfaserstabe. DAfStb Heft
304, 1979, pp. 19/43.
(6)
den Uijl, J.A.: Mechanical Properties of Arapree. Part 4:
Creep and Stress-rupture. Report 25-87-31, TU Delft,
1991.
(7)
Fiber Composite Elements and Techniques as Non-Metallic
Reinforcement of Concrete. EC- Research Project, BREU 91
0515, started 1992.
(8)
Phoenix, S.L.; Wu, E.M.: Statistics for the Time-Dependent Failure of Kevlar-49/Epoxy Composites: Micromechanical Modelling and Data Interpretation. Research Report
UCCR-53365, Lawrence-Livermore Laboratory, University of
California, 1983.
(9)
Franke, L.: Schadensakkumulation und Restfestigkeit im
Licht der Bruchmechanik. Festschrift G. Rehm: Fortschritte im konstruktiven Ingenieurbau. Verlag W. Ernst
und Sohn, Berlin, 1984, pp. 187/197.
(10)
Rehm, G.; Schlottke, B.: Ubertragbarkeit von Werkstoffkennwerten bei Glasfaser-Harz-Verbundstaben. Mitt. des
Inst. f. Werkstoffe, Universitat Stuttgart, 1987/3.
(11)
Dillard, D.A.; Morris, D. H.; Brinson, H.F.: Predicting
Viscoelastic Response and Delayed Failures in General Laminated Composites. ASTM 6. Conf. on Composite Materials:
Testing and Design, Phoenix, 1981.
(12)
Beckwith, S.W.: Viscoelastic Characterization of a Nonlinear Glass/Epoxy Composite Using Micromechanics Theory.
Annual Meeting of JANNAF, San Francisco, 1975.
96
Budelmann and RosUtsy
T = canst.
environment
= canst. =E
0
creep
rupture
~.+
T
OL----------------------tLu__________
t
Fig. 1-Creep rupture phenomenon under sustained loading
fc.~l
(l n)
short - time strength
creep rupture
h (ln tul
• creep rupture
o termination of long- term test
"' residual tensile strength
t, tu
Fig. 2-Creep rupture and strength retention
(ln)
FRP Reinforcement
~1:-flH-tt-lt--t-11-- Anchorage Device
!1-11-11-+-11-- Vessel Suspension
Fig. 3-Principle sketch of the creep rupture test set-up
97
98
Budelmann and Rostasy
Fig. 4-Creep rupture test set-up of TV Braunschweig
FRP Reinforcement
~m
GFRP
T= 20°C, air
1,0,~
......
~:,...
--;;;;;:: T_s_c_a_tt_er_ra_ng~e------i
o.a
1
0,6
O,L.
-
vt =0,64, polyester matrix
vt = 0, 64, expoxy matrix - - - - 1
various cross- sections
0,2
10"
in h
Fig. 5-Creep rupture of GFRP
Fct 1 , 0 . . - - - - - - - - - - - - - - - - - - ARAPREE f 100K ; 1,5 x 20 mm
Fcm
Ac= 30mrrf; vf =37%
0,9
alkaline solution; 20°(; pH 13
(ln)
fck= 30kN; fcm= 36 kN
0,8
• broken
o not broken
0,7
•
0,6
10
10 4
10 5
106
t(ln} in [ h J
Fig. 6-Creep rupture of Arapree in alkaline solution
99
100
Budelmann and Rostasy
1,0
0,9
0,8
'-1 L.t'
E
LLU
"0
§
0,7
-~ L.t'
E
LLU
F.*
_..£1
0,6
Fcm
= const.
GFRP; v1 ::.: 0.65
20°C. air
n=35
B::: 7.06
0,5
L.,__ _...J..__ ___,___ ____.__ ____.JL__ _.L.,_____J
100
10 5
5 ·10 5
lnt,t[h)
Fig. ?-Calculated creep rupture line and line of
strength retention (example)
SP 138-7
Properties of Fiber Reinforced
Plastic Rods for
Prestressing Tendons
by T. Uomoto and H. Hodhod
Synopsis:The mechanical behavior and tensile strength of three kinds of FRP
rods was investigated experimentally. For each material, three different fiber
volume fractions were tested in a.xial tension. The stress-strain relationships
and strength distributions were obtained. The results were correlated to the
behavior a.nd strength of the basic strengthening elements, fibers, a.s determined expeimenta.lly. It yeilded the possibility of predicting rods modulus,
but not strength, from those of the fibers. The strength distributions showed
a. shift tha.t is not, generally, proportional to rods fiber content. Investigation of this phenomenon, through stress analysis a.t the grips a.nd thorough
inspection of failed rods, assured the change of rods fa.ilnre modes for different fibers content. The effect of grips could lea.d to one of two shear ia.ilure
modes instead of tension mode. Therefore, a.n apparent strength reduction
is observed. Appropriate design of the gripping system is needed, in view of
rods properties, in order to get the best performance of the rods.
Keywords: Fiber reinforced plastics; prestressing steels; shear strength;
strength analysis; stress analysis; stress-strain relationships; tensile strength
101
102
Uomoto and Hodhod
ACI member Taketo UOMOTO is a. professor a.t the Institute of Industrial Science, University of Tokyo, Tokyo, Ja.pa.n. His current research interests include non destructive test methods, utilization of indus trial wastes,
corrosion a.nd corrosion protection of RC structures. He is a. member of ACI
Committee 440.
Hosam HODHOD is a.n assistant lecturer a.t the faculty of engineering, Cairo University, Giza., Egypt. He received his B.Sc. a.nd M.Sc. degrees
from Cairo University, a.nd received the doctoral degree from the university of
Tokyo. His research interests include fibrous composite materials, numerical
a.nd statistical approaches for structural problems a.nd simulation of materials
behavior.
INTRODUCTION
Concrete structures, in the coming century, ha.ve to keep up with structures
complexity, aggressive environments a.nd employment of high technologies.
One of the most important elements tha.t governs the performance a.nd durability of concrete structures is the reinforcement. Along the ma.ny decades of
concrete application in structures, steel wa.s the only reinforcement. However,
its corrosion wa.s figured a.s one of the major sources of structures deterioration. Therefore, there is a. need for durable kind of reinforcement. High
technology applications in different fields requires non-metallic reinforcement
in order to ha.ve a.ccura.te a.nd efficient operation (a.s in some transportation
a.nd medical applications). The repair a.nd rehabilitation of concrete structures, too, requires lighter reinforcement particularly for critical or hard-toaccess locations. All these requirements, in addition to high tensile strength,
recommended fi her reinforced plastics (FRP) a.s a.n ideal alternative to steel
reinforcement.
As the application of this ma.teria.l in concrete structures is a. recent
one, a.n extensive research is needed for achieving a. feasible utilization of
it. In addition to economical considerations, there a.re ma.ny parameters to
investigate. The parameters are reinforcement geometry, fibers/matrix combination (constituents materials), fi hers trajectories and fiber volume fraction
(V1 ). The chosen parameters, for characterizing FRP reinforcement, in this
study are fibers material a.nd volume fraction. The rods a.re tested in static
tension which is the fundamental ca.se of loading encountered in practice.
This study aims a.t introducing the behavior of FRP reinforcement a.nd its
FRP Reinforcement
103
variation, statistically, according to different reinforcement systems and iiber
contents. In view of these results, and by comparing iibers properties, the
factors affecting rods behavior in practice are clariiied.
MATERIALS AND TESTING SYSTEM
As mentioned above, there are many parameters for deiining FRP material. Hence, it seems appropriate to deiine the reinforcement studied in this
research so that all the conclusions become limited to this type.
The rods studied herein are circular in section and axially reinforced
with continuous parallel ii bers. Fibers diameters are, in average, about 10
micrometers. Three iibers materials are used: PAN carbon iibers, Technora
aramid iibers and S-glass ii bers. Fibers binding matrix is of the family of
epoxy acrylate resins and has a brand name of Ripoxy. The rods have plain,
smooth, surface and are gripped using the system shown in Figure 1, which
was developed by Kobayashi (1). The grips consist of two conical steel wedges
coniining the rod end and contained in steel sleeve of inner conical surface
and outer cylindrical surface. Testing machine is Autograph (displacement
controlled) with 10 ton capacity. The grips are connected to the machine by
steel thin walled cylinder (the third element in Figure 1). In practice, this
cylinder is not needed as the sleeve flange rests on concrete surface.
RODS SPECIMENS AND RESULTS
The tested rods were 6 mm in diameter and 40 em in length. To ensure
sufficient gripping, the ends were coated with a mixture of a liquid adhesive
and iron powder to make the surface, inside the grips, rough. Three iiber
volume fractions were studied: 0.45, 0.55 and 0.66. One hundred specimens
per material per volume fraction were tested. Stroke rate was 5 mm/min, for
aramid FRP (AFRP) and glass FRP (GFRP) rods, and 2 mm/min for carbon
FRP (CFRP) rods. This accounts for large stiffness of CFRP rods so that
the same loading rate would be obtained for all rods. Tensile strengths were
calculated for all the cases, but only few of them were checked for the stressstrain relationships. The reproducibility of the relations did not necessitate
more measurements. The strains were measured as the average reading of
pairs of resistance strain gages attached to the rods at the center. Typical
stress-strain relationships, normalized by V1 are shown in Figure 2. Strength
cumulative distributions are shown in Figure 3.
104
Uomoto and Hodhod
FillERS BEHAVIOR
As FRP rods are composed mainly of continuous fibers, that are much
stronger and stiffer than the binding matrix, fi hers behavior is expected to
be a useful mean of rods behavior evaluation. Single fibers tests were conducted on the three materials: aramid, carbon and glass. The specimens were
prepared according to the Japanese industrial standards (JIS R7601/1986).
The standard testpiece is shown in Figure 4(a). Single fiber is attached to a
rectangular frame of card paper. The frame contains longitudinal opening of
25 mm length, across which the fiber is pasted. The frame is mounted in the
testing machine and, just before loading, is cut on both sides of the fibers as
shown in Figure 4(b). The load is transferred to the fibers through the frame
at the ends. Then, The fiber carries solely the total load along the gage length.
Fibers stress-strain relationships are shown in Figure 5, they are typical for all fibers. Fibers diameters and number of tested fibers in addition to mean values for tensile strength, Young's modulus and failure strains
are given in Table 1. Strength distributions were obtained and are shown
in Figure 6. Data was fitted to Weibull distribution in the form f(x) =
.,'!xm-l exp[-(!")"'], where f(x) is the probability density function and m
and a are constants. Constants values are given in Figure 6 for all cases.
More details on these experiments can be found elsewhere (2).
DISCUSSION
Stress-Strain Relationships
Obviously, rods stress-strain relationships are practically linear. However, some curvature is observed for the case of AFRP rods, in forms material
softening followed by hardening till failure. The same curvature is observed
in the case of aramid fi hers itself. This curvature can be interpreted by recalling the polymeric structure of aramid fi hers. Polymers are composed of
chains of the repeated monomer. Generally, these chains are folded in the
thickness of the fi her. For high strength aramid polymers, there is no evidence on the complete folding but the chains are, still, not perfectly straight.
Up on loading, the chains are straightened and, therefore, fi her response is
becoming softer due to this type of plastic behavior. With increasing load
level, the percentage of straightened chains increases and the response becomes more elastic. Therefore, the stress- strain relationship becomes stiffer.
As the straightened chains cannot be folded again, upon unloading, the descending branch of stress-strain relationship in cyclic loading should maintain
the maximum slope attained in the loading part. Upon reloading the slope
FRP Reinforcement
105
is maintained until the highest load level, in the previous cycle, is reached.
Then, the slope increases gradually, again. Figure 7 shows the cyclic loading
of single aramid fiber as obtained experimentally. It confirms the above explanation. Similar results were obtained for AFRP rods. too.
One common rule for predicting both stress-strain relationships and
strength of composite materials is the law of mixtures (3). Axial Young's
modulus, of rods, can be obtained as the weighted sum of those of constituents
moduli. Usually, values of such matrix properties are very small compared
with those of fibers. Hence, one can neglect the matrix in the calculations.
Therefore, rods Young's modulus Ec, according to the law of mixtures, becomes
(1)
A model that employs the measurements of fibers moduli in axial tension for predicting rod modulus (4) gave the value of E 1 as the mean of all
measurements. The calculated rods response, according to equation (1 ), is
plotted on the experimentally obtained relationships in Figure 8. The good
agreement assures the validity of the law of mixtures for predicting rods response in axial direction. However, the difference, between experimental and
calculated responses, is relatively big for the case of GFRP. This is attributed
to matrix response that is ignored in equation (1). Another factor, is the difference in Poisson's contraction of glass fi hers and the binding matrix (5).
This difference leads to lateral tensile stresses in the matrix, when matrix
contraction is larger than fiber contraction. These stresses increase the stiffness of the matrix in the axial direction, than that obtained in uniaxial tensile
testing. The cases of other fibers (aramid and carbon) did not show lower
fiber contraction than that of the matrix (5), and hence such big difference
was not observed.
Strength Distributions
For AFRP rods, the strength distributions for different fiber contents
are almost equally spaced. This implies that the strength increases proportional to fibers volume fraction, as one may intuitively expects. The cases
of CFRP and GFRP rods, however, are not similar. For GFRP, the shift of
strength distribution, at v, = 0.55, from that at Vt = 0.45 is proportional to
However, the case of
= 0.66 does not show the expected increase of
strength.
For CFRP rods, all cases show that strength increase is not proportional to
Some interesting distribution was observed for the case of V1 = 0.66.
This distribution includes strength values as small as the low values obtained
for V1 = 0.55.
v,.
v,.
v,
106
Uomoto and Hodhod
The failure shape of the specimens was irregular in most of the cases ( clear
failure surface could not be identified). However, observations, during testing, showed that failure initiated, in the cases of low strength rods, near the
grips. This indicates the effect of grips on failure mechanism and strength
distribution.
The grips exert lateral pressure on the rods and this induces additional
axial stresses (6). Therefore, a reduction in rods strength is expected. In
order to determine the actual failure shape and material strength, special
specimens were tested (7). Rods with reduced sections at the center were
prepared and tested in axial tension. Different volume fractions in the range
0.45-0.66 were tested. The failure shape was irregular for the case of AFRP
but occurred at the center for both CFRP and GFRP rods. The normalized
mean strength obtained for all the cases is shown in Table 2 together with
those for the full rods and fibers.
The difference between actual strength (for the rods with reduced sections) and fibers strength is obvious for the cases of AFRP and GFRP. If the
law of mixture, in a form similar to equation (1), is applied, the normalized
strength should be equal to fibers strength. This indicates the invalidity of
applying law of mixtures for predicting FRP rods strength. Comparing the
figures in Table 2, one can reach some conclusions. The case of AFRP is not
affected by the grips and exhibits strengths similar to the specimens with reduced sections. It should be mentioned that the irregular failure shape could
be attributed to the interaction between fi hers and binding matrix that results in fiber bundle-like behavior for the rods.
The case of GFRP rods is similar to AFRP only up to Vi = 0.55. Then,
large deviation from the actual strength is observed. Thorough investigation
of rods ends showed that the ends are damaged and slippage of rods ends
from the grips took place. In view of the lateral pressure from the grips,
the failure of rods by interlaminar shear stresses was identified as a possible
failure mode if sufficient lateral pressure is applied (5). The lateral pressure
at the grips is proportional to the axial load in the rods that is proportional
to Vr. Therefore, this behavior was observed mainly at high values of Vr.
In case of CFRP rods, all the cases show strengths lower than the actual ones. Perhaps the closest values are observed for the case of V1 = 0.45,
where the failure shape was similar to the ideal one, in case of reduced sections specimens. The other cases, however, did not show any damage of the
ends or slippage from the grips. Nevertheless, the mode of failure has obviously changed. A detailed analysis of this case showed that the rods failed
by shearing the fibers (cross-laminar shear failure) due to the applied lateral
pressure at the grips. The geometry of the wedges (being circular in section)
caused the magnification of this effect in some cases rather than others and,
hence, resulted in the two-mode distribution observed for Vr = 0.66. In case
FRP Reinforcement
107
of Vt = 0.55, axial loads are not as high as the case of Vt = 0.66 so that this
effect could not cause two-mode distribution but only red need strength.
Based on the above facts, the design of grips shape, and perhaps material, becomes very important for getting the full benefit of the FRP rods
strength. It, also, becomes material dependent as the effect varies from one
kind of FRP rods to the other.
Another interesting feature can be found in Table 2. That is, the
strength of the fi hers at the standard length is not an appropriate measure
for evaluating, even on qualitative basis, the strength of FRP rods. This is
due to the fact that the fi hers behave, inside the composite, as an integrated
continu urn. The stress concentration and propagation is dependent on other
factors like fibers transfer length and fiber/matrix bond. The situation becomes more complicated when the fi hers have a length effect to be identified.
This interprets why glass fi hers, that possess mean strength about 70% of
carbon fibers strength, exhibits strength comparable to carbon fibers, in the
composite.
CONCLUSIONS
The performance ofFRP rods under axial tension was characterized. Three
different fibers materials aligned axially in circular rods were used for the
study. Three different fiber volume fractions were tested. Strength distributions for all the cases were obtained. The stress-strain relationships for
all rods were determined and found linear in all the cases. Though practically linear, AFRP rods showed some curvature that could be attributed to
the polymeric structure of the material. Comparison with the fibers shows
that these relationships can be deduced from fibers stress-strain relationships
according to the law of mixtures. This law, however, proved to be misleading in predicting FRP rods strength from fibers strength at standard length.
Other parameters are needed for evaluating composite strength. These include fiber/ matrix interaction characteristics.
The strength of FRP rods gripped using the employed gripping system
is affected to different extents by the lateral pressure at the grips. This is
based on rods materials and volume fraction. High volume fractions are affected more due to their high axial loads and the consequently high lateral
pressure from the grips. This pressure induces additional axial tensile stresses
and lateral compressive and shear stresses. The latter could lead to splitting
of rods ends and subsequent slippage and apparent low strength. The compressive stresses could lead to other mode of failures like cross-laminar shear
failure and cause apparently low strength. These secondary failure modes,
108
Uomoto and Hodhod
that become dominant for high fiber volume fractions, call for the redesign of
end grips for better performance and efficient utilization of FRP rods.
REFERENCES
1. Kobayashi, K., "Anchors for Fiber Reinforced Plastic Tendons for Prestressed concrete", Seiken leaflet, no. 158, 1987.(In Japanese)
2. Uomoto, T. and Hodhod, H., "Properties of Fiber reinforced Plastic
Rods for Prestressing Tendons of Concrete (2): Behaviour of Fibers for
FRP Rods Under Tensile Loading", Seisan kenkyu: Journal of I.I.S.
University of Tokyo, vol.43, no.3,1991, pp.19-22.
3. Parratt,N.J ., "Fiber Reinforced Materials Technology", Van Nostrand
Reinhold Company, London, 1972.
4. Hodhod, H. and Uomoto, T., "Experimental Model for Ideal Tensile
Failure of FRP Rods", Proceedings of the 13th. annual meeting of JCI,
June 1991, pp.975-980.
5. Hodhod, H.,"Employment of Constituents Properties in Evaluation and
Interpretation of FRP rods Mechanical Properties", Doctoral Dissertation submitted to the Department of Civil Engineering, University of
Tokyo, Sept. 1992.
6. Barton, M.,"The Circular Cylinder with Band of Uniform Pressure on a
Finite Length of the Surface", J .Appl. Mech., Vol.8, Sept.1941, pp.A97A104.
7. Hodhod, H. and Uomoto, T ., "Evaluation of FRP Rods Tensile Strength
Using Monte Carlo Simulation" ,JSCE symposi urn on FRP reinforcement in concrete structures, April 1992, pp ..
8. Hodhod, H. and Uomoto, T., "Effect of State of Stress at the grips and
Matrix Properties on Tensile Strength of CFRP Rods", Proceedings of
Japan Society for Civil Engineers, August 1992.
FRP Reinforcement
109
TABLE 1 -MEAN PROPERTIES OF FRP RODS REINFORCING FIBERS
Ma.teria.l
Diameter
(pm)
Number of Tested
Fibers
Tensile Strength
(MPa.)
Young's modulus
(GPa.)
Ara.mid
Ca.rbon
12
7
13
154
150
155
3800
3300
2500
85
223
86
Gla.ss
Fa.il ure Strain
(%)
4.4
1.4
2.9
TABLE 2 - NORMALIZED TENSILE STRENGTH OF FRP RODS AND
FIBERS STRENGTH
Material Fibers
AFRP
CFRP
GFRP
3800
3300
2500
Tensile Strength (1t1Pa)/ V1
Full Rods
Rods with
v, = 0.45 v, = 0.55 v, = 0.66 Reduced Sections
3000
3015
3030
3150
2730
2400
2200
3100
3015
2700
3050
3000
Fig. 1-Grips for FRP rods testing
110
Uomoto and Hodhod
~
4000
...........
3000
I::!
----~
~
'-
2000
'"'.l
'"'.l
1000
m;;"
>
;,;'' ::
~
,,,,
,..,.:...
(Jj
\H
0
0
2.0
Strain
4.0
6.0
(o/o)
Fig. 2-Normalized stress-strain relationships for FRP rods
(T-J = 0.66)
FRP Reinforcement
111
1.0
IAFRPI
.~!:ii~~~
····~·-() . !j>.. ;~~
s---o ·.··v:.·········.· ~45·
<
f
•(
IGFRPI
0
~~~~~~~~~~~~~
800
1200
1600
2000
Strength (MPa)
Fig. 3-Strength distributions of FRP rods at different
J.f
112
Uomoto and Hodhod
Upper Chuck
,...
Punch for 3mm Bolt
':'
/ 1'\
a
a
~
Single Fiber
11'1
N
'-... k:'
Adhesive
0
,..
Lower Chuck
l~m:l
.,
20mm
(a)
(b)
Fig. 4-Standard fiber testpiece before and during loading
~
Q..
5000
......,.......................·.,.......... ,......................f.·
4000
:
.....:....................
l
- ;.~·~EP~-~t~:,:.--
~
{'.)
{'.)
3000
2000
~
a..
V)
1000
'
0
0
1.0
2.0
3.0
Strain
4.0
5.0
(o/o)
Fig. 5-Fibers stress-strain relationships
6.0
FRP Reinforcement
0.20
0.15
0.10
0.05
0.15
0.10
0.05
0
L-~~~~~~~~~--_J
0
2000
4000
Strength (MPa)
Fig. 6-Fibers tensile strength distributions
6000
113
114
Uomoto and Hodhod
5000
-
Aramid Fibers
4000
Cyclic toa<'iing
':::!
~
~
~
~
...
.....
3000
2000
~
V)
1000
0
0
1.0
2.0
3.0
Strain
(o/o)
Fig. 7-Cyclic loading of aramid fibers
4.0
5.0
FRP Reinforcement
115
3000
2000
1000
·'" ~- Calculated
i:.......Expe:dmen1:'
0
-
3000
~
~
~
.....
.....
~
I
2000
I
1000
'"~- · CalculatE!d
~Experiment
lo...
.....
tr.l
0
3000
0
2.0
4.0
6.0
0
2.0
4.0
6.0
2000
1000
0
Strain
(%)
Fig. 8-Prediction of FRP rods stress-strain relationships (Jj- = 0.66)
SP 138-8
Evaluation Items and Methods
of FRP Reinforcement as
Structural Elements
by S. Mochizuki, Y. Matsuzaki,
and M. Sugita
"Synopsis:"
In Japan FRP reinforcement has been given attention as the substitute structural
material for steel bar of reinforced concrete element. Architectural Institute of Japan (AU for abbreviation) set up the Committee of Continuous Fiber Composite
Material to deal with the need for the use of FRP in near future in 1988. The committee has performed research activity for three years with composition of structural
working group and material one and published report in three volumes.
This paper concerns evaluating items and methods of FRP and FRP reinforced concrete elements of AIJ from the structural viewpoints. To evaluate new material
which is beyond the Building Standards Code such as FRP is very difficult. It is
certainly dangerous to evaluate new material according to the conventional standards and the result of present researchs of the related structures. However it is
impossible to evaluate new material in complete disregard of the above two points.
Therefore the evaluation of this paper may be said to be with reference to the concept of conventional standards and the tendency of present researchs of reinforced
concrete structure in Japan and the same time, with suggestions to promote the future use of FRP and FRP reinforced concrete elements.
Keywords: Anchorage (structural); bonding; fiber reinforced plastics; joints
(junctions); mechanical properties; prestressed concrete; reinforced
concrete; structural members
117
118
Mochizuki, Matsuzaki, and Sugita
S. Mochizuki is a professor in the Department of Architecture, Musashi Institute of
Technology, Tokyo, Japan. He graduated and got his MEn and DEn from Waseda
University, Tokyo, Japan. His research activities cover a wide range of topics related to concrete structures and technology such as precast concrete structure, earthquake shear wall, finite element analysis and design procedures.
Y. Matsuzaki is a professor in the Department of Architecture, Science University
of Tokyo, Tokyo, Japan. He graduated and got his MEn and DEn from Tokyo
Institute of Technology, Tokyo, Japan. His research activities also cover a wide
range related to concrete structures and technology such as precast concrete structure, high strength concrete and continuous fiber composite materials.
M. Sugita is a General Manager of the Technology Division, Shimizu Corporation.
He graduated and got his MD from Waseda University, Tokyo in I 966. He has been
engaging in research of concrete structures with new materials. A member of ACI
440 Committee.
INTRODUCTION
In Japan, Fiber Reinforced Plastic (FRP) reinforcement has attracted attention as a
substitute for steel reinforcement in steel-reinforced concrete elements. In order to
cope with future needs related to the use of FRP, Architectural Institute of Japan
(AIJ) established the Committee of Continuous Fiber Composite Materials in I 988.
Based on the results of research activities by the Committee's Structural Working
Group, this paper summarizes the evaluation items, methods of FRP reinforcement
and FRP-reinforced concrete elements. From hereinafter, both FRP reinforcement
and FRP-reinforced concrete elements will be referred to as FRP. The results of this
study form the basis of the Design Guidelines for Continuous Fiber-Reinforced
Concrete, a national project that is currently proceeding under the Japanese government and in cooperation with private and academic organizations.
According to the research done by the Industrial Structure Council, Ministry of International Trade and Industry (Japan), Japan's new-materials market scale, which
was 500 billion yen in I 980, has been predicted to reach 5.4 trillion yen by the year
2000. After the two oil crises of I 973 and I 979, during I 980's new materials began
to be developed rapidly in various industrial fields. For example composite hybrid
materials was developed to provide compound functions, such as compactness and
lightness even to architectural structure.
One of the most important things in developing new materials is to establish evaluated items and to create usage for new materials. Particularly for new materials
whose characteristics far exceed the scope of the existing materials, it is dangerous
to make judgements based on the existing items and methods; thorough consideration of this fact is thus required.
This paper is the result of three years' research of the Structural Working Group of
the Committee of Continuous Fiber Composite Materials (hereinafter referred to as
the "Committee of Composite Materials") of Architectural Institute of Japan (AU).
In addition to providing an overview of the study results it is important to develop
the correct attitude and policy to serve as a basis for establishing evaluation items
FRP Reinforcement
119
and methods. Thus, the basic policy of FRP evaluation has been described here
before the FRP evaluation items and methods, which themselves have been limited
to FRP reinforcement and FRP-reinforced concrete elements. It is not yet possible
to include FRP-reinforced concrete architecture as an object. While some chapters
specifically discuss the bond, anchors, and joints as the appropriate characteristics
for reinforcement, a chapter is discussed on prestressed concrete elements, for FRP
reinforced concrete elements.
RESEARCH SIGNIFICANCE
As the FRP evaluation items and methods discussed in this paper are based on the
limited data currently available, they naturally are not sufficient. However, they will
likely contribute greatly to the guidelines for future FRP development and its usage.
PRINCIPLES OF EVALUATION
Profile of the Committee on Composite Materials
In the studies of the new materials with characteristics correlating to materials and
structures (i.e., studies spread over several domains), the thinking of the researchers
can be better understood when an overview is provided before the study contents.
The one given here includes the establishment of the Committee.
The Committee on Composite Materials was established after the Committee of
Concrete Composite Materials Using Continuous Fibers (hereinafter "CCC Committee"). The CCC Committee was assigned to AIJ in May 1988 in a three-year
project to survey and research for preparations of evaluation items and methods for
new materials made of continuous fibers (FRP). The CCC Committee consists of 24
member companies, including general contractors, textile makers, iron and steel
makers, and other major companies related to new materials. Thus, to perform the
study assigned by the CCC Committee, AIJ established the Committee on Composite Materials (chairman: Koichi Kishitani). This was established as part of the study
to verify the use of FRP-reinforced concrete as a future structural element. For the
research, the Materials Work Group and the Structural Work Group were assigned
to serve under the Committee on Composite Materials.
The Materials Work Group and the Structural Work Group issue reports annually.
The report of 1988 has covered the [then-present] situation and problems of FRP
reinforcement of concrete structural elements. The report of 1989 summarizes the
evaluation items necessary for applying FRP reinforcement to concrete structural
elements, and the methods to evaluate the [then-present] situation . The report of
1990 also discusses the evaluation methods of FRP reinforcement and FRP-reinforced concrete, other problems to be solved and future subjects. This paper mainly
summarizes the report that the Structural Working Group prepared in 1990.
120
Mochizuki, Matsuzaki, and Sugita
Principles of Evaluation
Fig. 1 shows an example of the material characteristics of fibers used for FRP reinforcement. The figure indicates that the tensile strength of high-strength carbon fiber HTCF, for example, is about 14,000kgf/cm>, or roughly 4 times the yield
strength of 3,500kgf/cm2 of the SD35 steel reinforcement. In contrast, Young's
modulus of the aramid was as low as 1/10 that of steel reinforcement. In addition,
the fiber itself is elastic and does not exhibit any toughness. As FRP reinforcement
is a composite material made from continuous fibers of carbon, polyamide, glass,
etc., and has a matrix of epoxy, polyester resin, or cement, its characteristics are all
the more complicated. In terms of mechanical characteristics, such as high strength,
elasticity, and toughness, which are required of steel-reinforced concrete structural
elements, fiber properties can be said to exceed the scope of the mechanical characteristics of steel reinforcement. Thus, in evaluating FRP reinforcement and FRP
reinforced concrete, it is highly dangerous to apply the design criteria of the existing
conventional steel-reinforced concrete structures.
Article 38 of Japan's Construction Standards Law stipulates the following concerning special materials, such as FRP reinforcement and FRP-reinforced concrete elements: "Architecture that employs special architectural materials or construction
methods must be approved by the Minister of Construction as being equal or superior to those provided for in the Construction Standards Law." In terms of this
clause, it cannot be denied that the standards for the existing steel-reinforced concrete structures are the closest to those for FRP-reinforced concrete structures. Today, in Japan, the propagation of FRP-reinforced concrete structures cannot be realized without the approval by the Minister of Construction. Thus, it is a contradiction to evaluate such special materials as FRP-reinforced concrete.
Considering this inconsistency, the Structural Work Group is basically working on
evaluation criteria for steel-reinforced concrete structures, for the following reasons:
1) At present, it is nearly impossible to establish an independent system to evaluate
FRP reinforcement and FRP-reinforced concrete elements.
2) Many studies and results of questionnaires given to companies, which are used as
data for preparing evaluation criteria, are based on those of steel-reinforced structures.
3) Evaluation criteria are conventional in their nature, so they should not be too
different from the existing (similar) evaluation criteria. Whereas the design system
for steel-reinforced concrete has already a history of more than 100 years, it has
been Jess than 10 years since FRP reinforcement was first considered as a structural
element to replace reinforcement.
It was judged impossible to establish an independent evaluation system for FRP
reinforcement and FRP-reinforced concrete elements, based on the results obtained
after considering these points. The studies and questionnaire results from companies, which form a major part of the evaluation criteria data, are related to development studies. These development studies are based on the steel-reinforced concrete
design system. A compliance with the evaluation criteria of steel-reinforced concrete structures was the result of acknowledging these facts. Evaluation criteria of
structural materials are deeply related to architectural safety performance. Architectural safety performance should be evaluated based on social consensus and should
be continuous. Based on the previously mentioned design philosophy, it was
judged that evaluation criteria should be continuous.
FRP Reinforcement
121
MATERIAL PROPERTIES OF FRP REINFORCEMENT
Definition of FRP Reinforcement
Shapes of the major FRP reinforcements used at present include continuous fiber
reinforcement (round bars, deformed bar, twisted strips, combined strips, grids,
mesh, fabric) and continuous-fiber elements (plates and others). The fibers used are
carbon, glass, aramid, and vinylon. Resin and cement matrix were also used. Marks
should be attached to indicate the type of fiber, type of matrix, shape, tensile
strength, and surface treatment. This is to ensure the quality of FRP reinforcement
when it is used as a structural material. The following are examples of some proposed markings.
CFRD3920: Carbon fiber, resin type, deformed bar, tensile strength = 3920N/mm 2
GFRM 1960: Glass fiber, resin type, mesh, tensile strength = 1960N/mm 2
To determine the strength of FRP reinforcement, a unified nominal diameter or unified nominal cross-sectional area is necessary. There are two concepts related to the
cross-sectional area: one is to derive the total cross-sectional area of the fiber, and
the other is to derive a cross-sectional area that includes the matrix as well. In Japan,
the latter is adopted. For example, for combined strip reinforcement, the cross-sectional area is calculated from the value obtained by dividing the actual fiber area by
the fiber volume ratio. A method to work out the nominal cross-sectional area from
the weight and density per unit of length of FRP reinforcement has been proposed.
The fiber-mix ratio is expressed in terms of the fiber volume ratio and is determined
from the weight, cross-sectional area ratio, via impregnation or sulfuric acid extraction methods. However the necessary unification hasn't yet to take place.
Tensile Property
The tensile strength of FRP reinforcement is affected not only by the tensile
strength, volume ratio, size, and cross-sectional area of the fiber, but also by the
bond performance of fibers and matrix. With certain FRP reinforcements, for example, the tensile strength will drop when the diameter is increased. As the diameter is increased, the number of fibers will also increase, but the bond performance
of the fibers and matrix will decrease. However, the tensile strength fluctuation
coefficient is 2-7%, which indicates a small-scale strength of dispersion.
Because FRP reinforcement does not have a definite yield point, the standard
strength of FRP reinforcement, corresponding to that of steel-reinforced concrete
structures, must be evaluated in terms of tensile strength, giving consideration to
dispersion. It was proposed that the standard strength ofFRP reinforcement is taken
as 70% of the value obtained from subtracting 3 times the basic deviation s from the
tensile strength. The decrease coefficient of 70% was determined, corresponding to
that of the steel reinforcement.
F = 0.7 (stu- 3s).
Where F : standard strength
stu : tensile strength
s : standard deviation
122
Mochizuki, Matsuzaki, and Sugita
Three types of stiffness are used to evaluate FRP reinforcement stiffness: 1)
Young's modulus P/Ae, 2) axial stiffness P/e, and 3) displacement stiffness P/d. In
this case, the definition of load P, strain e, and level of elongation d are important.
For load, 1/3 or 2/3 of tensile strength is used. The cross-sectional area, as described
above, is worked out in terms of nominal cross-sectional area.
Details of the methods used in the tensile tests to evaluate tensile characteristics
were prescribed by the Material Working Group and have been omitted here.
The above refers to the tensile characteristics of FRP reinforcement of straight parts,
but the bent parts have characteristics appropriate to the composite material. On the
bent part, for example, uneven fiber sag results from production, and because the
fibers have no yield zone, they fracture violently when stress is applied, causing
tensile strength of the bent parts to drop to 30-70% of that of the straight parts.
Other Characteristics
There are few studies on the compression characteristics of FRP reinforcement.
Although a formula based on the mix rule has been proposed as an equation of
compression strength, it is difficult to determine the compression strength acting on
fibers; thus, other methods of evaluation is desired.
According to the report, the fact that FRP reinforcement is a nonisotropic material
means it may rupture explosively when used in related elements and when transverse cracking occurs: Unlike steel reinforcement, it has no dowel effect. However,
lack of related test results means that the shear strength cannot yet be evaluated.
BOND AND ANCHOR OF FRP REINFORCEMENT
Bond
At present, bond is not evaluated differently for steel reinforcement when they are
used. Important subjects of evaluation include long-term performance, repeated
loading, and large diameter. Another issue of consideration is how the bond performance of FRP reinforcement depends on the surface shape deformity and matrix
performance. Because many factors affect the bond performance of FRP reinforcement, it is impossible to apply the results of certain bond tests as standard test methods. Thus, evaluation of bond performance in design should be done separately for
each element and use. A few test methods available for evaluating typical bond
status are described next.
Standard Bond Pulling Test
This test method evaluates the critical bond performance under sufficient confinement. Therefore the bond performance must be exhibited to 100% of its ability and
fracture mode should be determined by pulling the FRP reinforcement. Fig. 2 is an
example of an experiment in which steel reinforcement was used. This test method
was done by pulling off the FRP reinforcement that was embedded in a concrete
FRP Reinforcement
123
block that had a large enough cross-section. The side length of the block should be
at least 12-16 times the diameter of the FRP reinforcement and the bond length
about 3 times the nominal diameter. The quantity of slipping are measured and the
load are applied to the free end, thus, allowing a load-slip curve to be derived.
Split Bond Pulling Test
This test is to evaluate the performance of FRP reinforcement when it is used as
main reinforcement of the element subjected to shear force. The performance is
measured at the time when the periphery is split, therefore destroying the bond performance. Fig. 3 shows a cantilever test and a Schmidt Thro test similar to the steel
reinforcement tests. The bond length is necessary to be at least 15 times the nominal
diameter of the FRP reinforcement and the minim urn coverage about 1.5 times considering the foregone experimental results. The bond split strength and slip displacement at the loading end are measured.
Anchor
Almost experiments on performance evaluation of anchor are the experiments of
anchor equipment structures for post-tensional in prestressed concrete structures.
For an anchor system, using something other than anchor equipment, end bending,
node application, intersecting bars systems are available. For the end bending system, the strength of bent part is only 30-70% that of the straight one. This makes the
development of sound bend details an important subject. Fig. 4 shows an example of
the bending anchor test. The strength of bent part differs with the processing
method, shape of fiber reinforcement, and type of fiber used. The strength of bent
part is determined by the bending rigidity of inside part of bending processing, as
well as the FRP reinforcement elongation capacity.
Grid type reinforcement with intersecting bars is used to secure bond and anchoring
force by bearing force of the intersecting elements. In this case as well, the evaluation should be done via tests that can reproduce stress status in elements and during
usage in elements. The evaluation items on the anchor of grid type reinforcement
are the spacing and numbers and the projecting length of the intersecting bars. At
the grid intersections, the transfer of shear force from the bearings force of the intersecting bars must be ensured.
STRUCTURAL PROPERTIES OF FRP REINFORCED CONCRETE
Axial Force
Because FRP reinforcement is expensive, its use in compression bars is not economical. However, its use in tensional bars, by leaving compression to concrete, has
been considered. A future use to be considered is its application to column elements,
but at present, only basic experiments have been done on the confinement effect for
application to column hoop reinforcement. Thus, evaluation methods cannot yet be
summarized.
124
Mochizuki, Matsuzaki, and Sugita
Two evaluation items, the axial reinforcement effect and the lateral confinement
effect, are discussed for applying FRP reinforcement to columns. Axial reinforcement effects were scarcely attained within 1.2-2.4% of axial reinforcement ratio.
This is due to the low rigidity of FRP reinforcement.2) The items used to evaluate
the lateral confinement effect were (1) quantity of lateral reinforcement that can
produce lateral confinement, (2) rupture of hoop comers, and (3) the axial deformation limit. The quantity of lateral reinforcement that can produce lateral confinement is considered to be larger than steel reinforcement due to the low rigidity of
FRP reinforcement. However, the lateral confinement effect of FRP reinforcement
is the same as conventional steel in the 0.8-3.6% range of lateral reinforcement
ratio.2) Rupture at hoop corners is not a problem with axial strength alone. But
generally for FRP reinforcement as a whole, rupture conditions at hoop corners
must be clarified. The deformation limits based on each rupture mode are necessary
for evaluating lateral reinforcement. The axial deformation limit is determined by
rupture at hoop comers; experiment results show the limit to be 6% for polyamide
fiber reinforcement and 2-3% for carbon fiber reinforcement.
Bending Moment
As is the case for steel-reinforced concrete elements, the evaluation of bending reinforcement in elements with FRP reinforcement is also classified into 2 categories of
1) cross-sectional analysis and 2) element characteristics. Fig. 5 shows a test piece
from a pure bending experiment 20cm wide, 30cm thick, and a span of 240cm. Fig.
6 shows that the strain distribution is nearly linear,3) and that the supposition for
plain retention is satisfied. Fig. 7 exhibits a comparison of the analysis done by
assuming the moment-curvature relationship and the experiments3). The stressstrain relations of FRP reinforcement was determined from material test results.
The stress-strain relations of concrete was determined by assuming that these can be
expressed with the e function and ultimate strain of 0.35%. Although the analysis
values were slightly lower than the experiment values, they still nearly coincided.
Thus, the evaluation of cross-sectional analysis of the bending of FRP-reinforced
concrete elements is said to be similar to that of steel-reinforced concrete elements.
As for the evaluation of element characteristics, the load-deflection relationship
with the tensile reinforcement ratio p, as a parameter is shown in Fig. 8. The relationship between stiffness after bending crack K and the product of tensile reinforcement ratio pt and tensile reinforcement stiffness E is shown in Fig. 9.3) In Fig.
8, the other specimens than those marked with an "x" (which indicates the main
reinforcing bars were ruptured) were ruptured when concrete was crushed. Therefore, it can be said that bending toughness can be attained even when the FRP reinforcement with elastic rupture is used as main reinforcement. Also, as shown in Fig.
9, the flexural rigidity after bending crack can also be controlled by the tensile
reinforcement ratio and its stiffness value. Therefore, on the whole, the design of
FRP reinforced concrete elements for bending moment can be done similarly to how
it is applied to steel-reinforced concrete elements. If the rupture mode is based on
the expectation that the compression concrete will be crushed and the current FRP
reinforcement has low stiffness, flexural rigidity after bending cracking will be considerably low.
FRP Reinforcement
125
Shear Force
The following describes the shear behavior of beam elements using FRP reinforcement as shear and main reinforcements. In comparison with the conventional steelreinforced concrete elements, evaluation was based on the shear crack width/shear
stiffness and the maximum shear strength/rupture mode. For beams using FRP reinforcement to shear reinforcement, shear crack width will exceed that of steel-reinforced concrete beams.4) Greater shear crack width have been reported for beams
using FRP reinforcement to main reinforcement than in the case of steel reinforcement. 5) This is the result of the drop in shear deformation confinement capacity
from low rigidity of FRP reinforcement.
Compared to steel reinforcement, the maximum shear strength is low in either case
when FRP reinforcement is used for main or shear reinforcement. 6) When FRP
reinforcement is used to main reinforcement, bearing shear force of the compression concrete is decreased by the low rigidity of FRP reinforcement. Thus, it is
considered possible to resolve the problem by making the axial stiffness equal to
that of steel-reinforced concrete. When FRP reinforcement is used to shear reinforcement, the maximum shear strength is derived from the rupture at the bent part
of shear reinforcement, or at the intersections of the main reinforcement and shear
reinforcement. At present, it is necessary to confirm the strength of the reinforcing
materials in experiments and to reflect the results in the designs.
STRUCTURAL PROPERTIES OF FRP PRESTRESSED CONCRETE ELEMENTS
Effective Prestressing
One of the evaluated criteria of the prestressed concrete elements is the effective
prestressing of the tension rods. The following are points necessary for evaluating
the effectiveness of prestressing: (1) the introduced prestress force and (2) the prestress efficiency ratio. In the "Standard for Structural Design and Construction of
Prestressed Concrete Structures" of AU, the introduced prestress force is equal to or
less than 75% of the standard tensile strength and 85% of the standard yield
strength of the PC steel rods. Even though there is no yield strength in FRP reinforcement, each type of this reinforcement still should be evaluated. At present, the
introduced prestress force is about 60% ofFRP reinforcement tensile strength and is
considered adequate.
The effective prestress force can be determined by considering ( 1) loss from friction
between tension rod and sheath, (2) the setup loss of anchor equipment, and (3) the
losses due to relaxation of tension rods and concrete creep and drying shrinkage.
The loss from friction between tension rod and sheaths differs greatly with the combination of the type of tension rod and the type of sheath. For example, the apparent
friction coefficient of a steel sheath with PC steel strands is 0.11, and the steel
sheath with carbon FRP reinforcement is 0.07. Namely, the value is smaller when
carbon FRP reinforcement is used as the tension rod. 7) Thus, at present, the friction coefficient is basically established by performing experiments for each combination of tension rod and sheath. In general the setup loss of anchor equipment is
said to be small when compared to the value of the setup quantity. Due to the small
elastic modulus of FRP reinforcement. The setup quantity for the wedge type is
1.4mm and is relatively small in value, so it is the principle design for wedge an-
126
Mochizuki, Matsuzaki, and Sugita
chorage to calculate the set loss of each anchor equipment. When FRP reinforcement is used in tension rods, it presents both an advantage and a disadvantage for the
drop in prestress: The former is low elastic modulus and the latter, large relaxation.
The relaxation ratio at lOOhrs is reported to be 3% for carbon FRP reinforcement
and 12% for polyamide FRP reinforcementS) At present, the FRP reinforcement
relaxation value, elastic modulus, the concrete creep factor and quantity of drying
shrinkage are generally determined first; then the prestress efficiency ratio should
be worked out based on the former.
Bending and Shear Characteristics
For PC elements using FRP reinforcement in the tension rods, bending crack load
and bending ultimate strength with a good bond can be calculated similar to the
cases in which the PC steel rods were used as the tension rods. However, if the FRP
reinforcement bond is insufficient or when there is no bond, bending ultimate
strength must be determined by calculating the stress increment of the tension elements considering bond deterioration or lack of bond. For PC elements using FRP
reinforcement in the tension rods, flexural rigidity before bending crack can be obtained in the same way as a conventional case of PC steel rods used as tension rods.
When the bond is good, flexural rigidity after bending crack of PC elements with
FRP reinforcement as tension rods can be obtained from effective moment of inertia.9) At present, however, it is calculated from the moment-curvature relationship
using the stress-strain relationship of concrete and FRP reinforcement. When the
bond is insufficient or when there is no bond, flexural rigidity after bending crack
must be worked out by adding either a bond deterioration factor, or eliminating
bond from the calculations.
Shear performance of PC elements using FRP reinforcement as tension rods is basically the same when PC steel rods are used in the PC elements. However, as experiment data on shear fracture are not yet available, the principle is to set the same
allowable shear force of PC elements using FRP reinforcement in their tension rods
as that of steel reinforced concrete elements disregarding the effect of prestressing.
Anchor Equipment
The currently used anchor systems can be classified into the wedge type and the
bond type with a cement material filled in between the sleeves. Generally, tension
rods are either anchored by wedges, or bond to the sleeves, which are then anchored
to the anchor plates. Thus, the final anchorage has steel plates, like conventional
anchor equipment. Anchor equipment varies with FRP reinforcement, and performance is confirmed in experiments. Performance must ensure 95% or more of the
tensile strength of FRP reinforcement.
FRP Reinforcement
127
CONCLUSION
I. To ensure the performance of FRP reinforcement as structural material, it is necessary to define symbols and nominal diameter or nominal cross-sectional area.
Another crucial subject is the definition of standard strength and rigidity, which will
be used as a basis of the mechanical properties of FRP reinforcement.
2. No standard test methods are available for evaluating FRP reinforcement bond
and anchors. It is thus preferable that bond performance for design be evaluated for
each use and point of use. Fig. 3 shows the examples of bond experiments.
3. The performance for bending moment and shear force of FRP-reinforced concrete elements can basically be evaluated in the same way as those of steel-reinforced concrete elements. However, to estimate the toughness of the beams that are
subjected to bending moment, the concrete on the compression side must be
crushed. When FRP reinforcement is used as a main reinforcement, the maximum
shear force and rigidity after cracking is low compared to when it is used as a shear
reinforcement.
4. In terms of FRP reinforcement utilization, expectations are the highest for the
tension rods of prestressed structures. However, to determine prestressing, the friction coefficient and relaxation must be worked out in experiments on combinations
of actual tension rods and sheaths.
References
1. Maruyama, T.; Honma, M.; and Okamura, H., "Experimental Study on the Tensile Strength at the Bending Part of Fiber Reinforced Plastic Rods," Proceedings of
the Japan Concrete, Vol.12, No.I, 1990, pp.I025-1030.
2. Okamoto, T.; Endo, K.; Tanigaki, M.; Ishibashi, K.; and Watanabe, K., "Study on
Braided Aramid Fiber Rods (Part 13. Centrally Loaded Compression Tests of
Square-confined Concrete Columns)," Summaries of Technical Papers of Annual
Meeting Architectural Institute of Japan, Structure II, 1990, pp.993-994.
3. Umebayashi, K.; Matsuzaki, I.; Nakano, K.; and Kawai, K., "Experimental Study
on Flexural Behavior of Concrete Beams Reinforced with New Reinforcing Materials," Summaries of Technical Papers of Annual Meeting Architectural Institute of
Japan, Structure II, 1990, pp.855-856.
4. Mikami, H.; Kato, M.; Takeuchi, H.; and Tamura, T., "Flexural and Shear Behaviors of RC Beams Reinforced with Braided FRP Rods in Spiral Shape, "Proceedings
of the Japan Concrete Institute, Vol.ll, No.I, 1989, pp.813-818.
5. Tottori, S.; Terada, T.; Wakui, H.; and Miyata, S., "Study on Behavior of Shear
Failure of Concrete Beams Reinforced with FRP Bars, "Proceedings of the Japan
Concrete Institute, Vol.lO, No.3, 1988, pp.541-546.
6. Saito, H.; Tsuji, Y.; Sekijima, K.; and Ogawa, H., "Flexural and Shear Behaviors
of concrete Beams Reinforced with FRP," Proceedings of the Japan Concrete Institute, Vol.lO, No.3, 1988, pp.547-552.
7. Sakai, H.; Yagi, K.; Koga, M.; and Kawamoto, Y., "The Application of the Tendons made of CFRP Rods with Post-Tensioning Method to Edge Girder of the Slab
Type Bridge-- Construction of the Bridge 'Bachigawa Minami Bashi' --,"Concrete
128
Mochizuki, Matsuzaki, and Sugita
Journal, Vol.28, No. II, 1990, pp.14-24.
8. Katawaki, K.; and Nishigaki, I., "Studies on Application of Fiber Reinforced
Plastics for Prestressed Concrete, Japan, Vol.30, No.5, September to October, 1988.
9. Kitta, T.; Ikeda, H.; Honda, B.; and Ohhara, H., "Improvement of Bending and
Shear Properties of Concrete Beam Reinforced by FRP Rods, "Proceedings of the
Japan Concrete Institute, Vol.l2, No.1, 1991, pp.1087 -1092.
I 5 0
!lTC F
"'"'
.!::"'
"'
I 0 0
E
E
AF
"""
IITCF/AF
IIMCF/AF
IIMCF/GF
.!<!
t:>
5 (
bnr
SD35
c (%) strain
Fig. 1-Tensile properties of composite materials
,..
st eel pipe
d,=Z5n
I
unbonded
T
L
I
1
1
+
bond.
reg JOn
WSG~
I
t
unbonded
I
f- rub ber
tt t tt t tot t t~ :~~::
l
=t
one
ease
8. (
301. 1
318. 5
Fig. 2-Example of bond experiment
8.(
FRP Reinforcement
129
Fig. 3-Examples of bond experiment (a) Cantilever type (b) Schmidt Thro's
:0:200
FRP rcinforcemen
'
I
I
FHP reinforcement
'
~J
C\
d
·~
I
I
I
~
I
"0
~
Main reinforcement
:;;: 100
Nl
j
I
~Sd
:0: 9dtD
:;, IOOtl~diD
J
7
~in relnforccacnl
I
I
100
1oo
200
Fig. 4-Example of anchor experiment
~ ::
10
35
...
.... '" ...
DD I-N)VflfEJ~
:::o
I
L1Q!U
L1Q!U
300
I
·~
100
I -P/2
Fig. 5-Specimen
I -P/2
100
I
300
130
Mochizuki, Matsuzaki, and Sugita
Fig. 6-Strain distribution curves
20
16
bent
(Lon·cs)
-
Experiment
•••• Analyse
12
Fig. ?-Relationship between moment and curvature
FRP Reinforcement
25
Load
(ton)
20
"'
::jJ'
~I
i'
~
•
10
~\
'\)~
~ ;<~'~>'bY'"
15
Jo••.
/ . · ·n.
--·
reinforccocnt
·-\-.
i' -~--1 --,
,n,l\1\ - - - - - - 'b\:,.. .... : •• . •. . . .•
t.-----,--1<
.
-71
:v~·-
l:) / . ~-' ~-~~~
~;0· ,'' r-t-~·'L
j
5
X Fracture of
/-...........____
/..,"
~
../
/
·~'
/
.
/./
'
_,""'
Dcflccilon (mo)
- .....r_ ----
Fig. 8-Relationship between load and deflection
,~:;/
K(kg/em)
100
-
. >:\ 'b
O:AF
6:CF(nraidod)
'i7:CI(Spiral)
\]
/
50
t-
.A. /
IS
i ,.
e .Y
O:GF
Whi te:Noroal
concrete
Dlack:lligh strength
concrete
p, · E (lO'kg/co')
0,
IZ.
0.11
0.21
0.31
Fig. 9-Relationship between K and p 1 • E
131
SP 138-9
Mechanical Properties of
Composite Beams by FRP
by E. Sueoka, K. Yasuoka, 0. Kiyomiya,
M. Yamada, M. Shikamori
Synopsis:
For better durability and more rapid construction
work of reinforced concrete marine structures, the authors
designed composite beams that consist of channel-shaped precast
concrete forms reinforced with FRP and inner reinforced concrete. In this study, static flexure tests and shear tests were
conducted, in order to study the basic mechanical properties and
to estimate design methods of these types of beams. The flexural capacity coincided well with the values calculated by a
conventional design method. The shear capacity was able to be
calculated on the safe side by 20-30% by using the Niwa equation. No dislocation occurred between the form and the inner
concrete up to the level of the design load. However, dislocations were observed between them at shear stresses of more than
22 kgf /cm2, and therefore shear connectors (studs) were needed
for their prevention.
Keywords: Beams (supports); composite construction; cracking (fracturing);
fiber reinforced plastics; flexural strength; mechanical properties; precast
concrete; shear properties; studs
133
134
Sueoka et al
E. Sueoka is a research engineer of the technical research
institute at Toyo Construction Co., Ltd. He has been conducting
research on marine concrete for eight years.
K. Yasuoka is a manager of design department at Toyo Construction Co., Ltd. He has been closely associated with design and
performance of port facilities.
Dr. 0. Kiyomiya is a chief of structural mechanics laboratory at
Port and Harbour Research Institute, Ministry of Transport,
Japan.
He has been researching on structural mechanics for
nineteen years.
M. Yamada is a member of structural mechanics laboratory at Port
and Harbour Research Institute, Ministry of Transport, Japan.
He has been researching structural mechanics of marine concrete
structures.
M. Shikamori is the director at Coastal Development Institute of
Technology, Japan. He has been associated with port and coastal
engineering.
INTRODUCTION
Corrosion protection is one of the major issues for marine
reinforced concrete structures. Durable structural members can
be fabricated using fiber reinforced plastics (FRP), which are
much more resistant to corrosion than steel. Since marine
construction work is characterized by changeable marine weather,
rapid and simple construction by means of precast members is
beneficial. To meet these requirements, the authors proposed a
composite beam which consists of reinforced concrete filled in a
channel-shaped permanent form which is made of precast concrete
reinforced with FRP grids. The durability of this composite
beam is ensured by arranging FRP in the precast form whose cover
is small and by arranging steel in the inner concrete over which
the cover is sufficient for protecting the steel from corrosion.
The form and the inner concrete are bonded mechanically to
resist external force.
Static flexure tests and shear tests were conducted, in
order to study the basic mechanical properties and to estimate
design methods of this type of beam. In this paper, results of
these tests and the applicability of design methods to the
composite beams were described.
FRP Reinforcement
135
OUTLINE OF LOADING TEST
Specimens
Table 1 shows the types of specimens, and Fig. 1 shows the
shapes and dimensions of the specimens and arrangement of the
reinforcing grids and bars.
The specimens consist of channelshaped precast concrete forms reinforced with FRP grids, and
inner reinforced concrete.
The specified design concrete
strength of the form, f 28 is 240 kgf/cm2.
The cross-sectional
shape of the specimens is rectangular. In order to study the
influence of the bond between the form and the inner concrete on
the overall mechanical properties, two types of composite beams
were fabricated.
In No. 3 and No. 6 specimens, steel studs
(dia. 16 mm x 150 mm) are embedded at 20 em intervals across the
interface between the form and the inner concrete as shear
connectors. The other specimens have no studs. The inner
surfaces of the forms are wire-brushed to expose the coarse
aggregate. The FRP in the forms is an epoxy resin reinforced
with carbon fibers (CFRP). CFRP has a higher modulus of elasticity and better resistance to chemicals than other FRP. The
properties of the FRP are shown in Table 2. Table 3 shows the
strength of the concretes.
Loading Method
Third-point loading (shear-span ratio: 3.15) and centerpoint loading (shear-span ratio: 1.30) were applied for flexure
and shear testing, respectively. After unloading at the initial
crack and at approximately 50% of the breaking load, the static
loading was monotonically continued using a hydraulic jack with
a capacity of 200 tf, to collapse of the beams. Displacements
of the beams, strains of FRP, steel, and concrete, crack widths,
and dislocations between the form and the inner concrete were
measured at the points as shown in Fig. 2. The measurements of
the dislocations were taken using a two-directional transducer
set on the interface on the top side of the beams.
TEST RESULTS
Flexure Tests
The flexural crack patterns and the load-displacement curves
are shown in Figs. 3 and 4, respectively. The crack spacing was
approximately 20 em, which corresponds to the FRP spacing.
After chipping of concrete, 1 t was observed that all of the
136
Sueoka et al
cracks started on the forms and extended into the inner concrete. After the initial crack developed, the stiffness of the
beams decreased. After reinforcing steel yielded, the displacements of the beams and the strain in the steel and FRP continued
to increase. However, no rupture of FRP was observed. Finally
the concrete near the loading point was compressed to collapse.
Ductilities of the composite beams were obtained by calculating
the ratios of the displacements at failure to the displacement
at yield of steel. The ductility of the beam with studs was
5.4, and that of the beam without studs was 6.0. By the use of
both FRP and steel in these composite beams, higher ductility
was ensured.
Ductile failure was observed in either of the
beams.
The measured displacements of the composite beams were
compared with the displacements until the yield of steel calculated using the effective moment of inertia (Eq. 1) in consideration of the cracks, as specified in the Standard Speci
fication(1).
The measured displacements were slightly greater
than the calculated values.
Mcrd )'
(
fe= [( Mdmax
fu+
1-
(
Mcrd )')
Mdmax
fer
]
~[g
( 1)
where
effective moment of inertia
limit moment causing flexural cracking
at cross section
maximum design moment
moment of inertia of gross cross section
moment of inertia of cracked section
transformed to concrete
Fig. 5 shows the relationship between the flexural crack
widths and the strains of FRP, and Fig. 6 shows the relationship
between the flexural crack widths and the strains of steel. Eq.
2 below, specified in the Standard Specification(2) for calculating flexural crack widths, was also plotted in both figures.
The influence of creep and drying shrinkage of concrete on crack
w= k
/4 C +0.7 (cs- ¢)I
-jf;-
( 2 )
where
k
c
Cs
¢
an
Es
constant to take into account the influence of bond
characteristics of steel, which may be set equal to
1.0 for deformed bars, 1.3 for plain bars and prestressing steel
concrete cover (em)
center-to-center distance of steel bars (em)
diameter of steel bars (ern)
stress increment in steel
modulus of elasticity of reinforcement
FRP Reinforcement
137
widths were not taken into account in this equation.The measured
flexural crack widths substantially coincided with the values
calculated using the strains and concrete cover of FRP.
In
other words, it was understood that when calculating the crack
widths of composite beams which contain FRP externally and steel
internally, they could be calculated by using only the FRP
strains in Eq. 2, provided that FRP and steel are sufficiently
apart (13 em in this case). When both FRP and steel strains
were used, the calculated values were much larger than the
measured values.
Fig. 7 shows the strain distributions in the center cross
section, and Fig. 8 shows the dislocations between the forms and
the inner concrete. From the two figures, it was made clear
that the composite beams with and without studs nearly satisfied
Bernoulli's assumption.
No significant dislocation at the
interfaces between the forms and the inner concrete were observed until the collapse. Consequently, a sufficient bond was
obtained between the forms and the inner concrete, up to the
level of the design load, by the exposure of coarse aggregate
before placing the inner concrete.
Shear Tests
Figs. 9 and 10 show the crack patterns of shear specimens
and the load-displacement curves, respectively.
While the
applied load was small, flexural cracks were observed in the
centers of the specimens. After shear cracks occurred, a tied
arch mechanism was observed.
Finally the concrete forming the
arch was compressed to collapse, but neither rupture of FRP nor
yield of the steel stirrups was observed.
Fig. 11 shows the dislocations between the forms and the
inner concrete. While the specimens with studs exhibited no
appreciable dislocation, behaving as one integral body, those
without studs gradually exhibited splits between the form and
the inner concrete when they approached the end points. Splits
up to approximately 2 mm wide were observed at failure. However, even those specimens without studs showed sufficient bond up
to 80 tf. The split occurred at a shear stress of 22 kgf /cm2,
which is much greater than 10 kgf/cm2, the value reported in the
tests in which the permanent precast forms were used only as the
bottom of the beams(3). In other words, sufficient bond against
shear is obtained between the form and the inner concrete by
exposing the coarse aggregate on the inner surfaces of the form.
Placing studs across the interfaces secures the bond further.
138
Sueoka et al
Comparison of Test Results with Calculated Results
Table 4 shows the test results and calculated results of
flexural specimens. The calculation of mechanical properties of
the materials is based on the data given in Tables 2 and 3. The
initial crack load was calculated using flexure test data for
the concrete for the forms with the gross cross section. The
measured initial crack loads were approximately 50% of the
calculated values. The cause is considered to be stress concentration on the FRP grid intersections, as reported in the experiments by Tsuji et al(4).
The flexural capacity and the ultimate strain of FRP were
calculated using the balance equation of forces with an equivalent stress block specified in the Standard Specification(5).
The authors assumed the breaking load of the flexural specimens
at a strain of a compressive edge of concrete of 0.0035. The
yield load of steel bars was calculated using a balance of
forces with neutral axis and yield stress of steel bars. Here
the form and the inner concrete were assumed to be integral, and
the effective concrete strength was calculated using the area
ratios of the form and the inner concrete. The test results of
the flexural capacity agreed well with the calculated values.
Table 5 shows the shear test results and calculated results.
The shear capacity and the ultimate strain of FRP were calculated using the Niwa equations(6), Eqs. 3 and 4, respectively, on
the assumption of a tied arch mechanism.
The effect of the
stirrups on the shear capacity was disregarded, because it is
considered to be marginal in beams with such a small shear-span
ratio.
Since the test results of the shear capacity were
slightly greater than the calculated results, it was found that
the beams of small shear-strain ratios with steel and FRP reinV=K fc' '/'bwd (l +v'pw) /
T=V/tana
[1 +(a/d)']
( 3)
( 4)
where
V
T
K
Pw
fc'
a/d
bw
r
d
shear capacity
tensile force of steel bars
constant (=0.53)
ratio of longitudinal reinforcement
(steel + FRP) area to area of web concrete
compressive strength of concrete
shear-span ratio (=a)
web width of member
length of bearing board
effective depth
FRP Reinforcement
139
forcement can be designed on the safe side by 20 - 30% by adopting the tied arch mechanism design method. The measured ultimate strains of FRP coincided relatively well with the calculated values, but they were smaller than the values calculated from
the measured shear capacity (approx. 15,000 x 10-6). This is
attributable mainly to the anchor effect of the grid FRP(7).
CONCLUSION
1. In the flexure tests, the composite specimens failed by
concrete compression after the yield of steel bars. No rupture
of FRP was observed. The flexural capacity coincided well with
the values calculated by a conventional design method.
2. In the shear tests, the composite specimens failed by
concrete compression near the load-applying point. As in the
flexure tests, no rupture of FRP was observed. The shear capacity was able to be calculated on the safe side by 20 - 30% by
using the Niwa equations with a tied arch mechanism.
3. No dislocation occurred between the form and the inner
concrete up to the level of the design load for flexure test.
Therefore it was confirmed that sufficient bond was able to be
obtained by exposing the coarse aggregate on the inner surfaces
of the forms and studs. However, dislocations were observed in
the specimens without studs across the interfaces at shear
stresses of more than 22 kgf/cm2.
4. Most of the flexural cracks started from the FRP grid
intersections in the bottom side of the forms.
The initial
crack loads were approximately 50% of the values calculated from
the flexural strength of concrete with a gross cross section.
The crack widths was able to be calculated on the safe side by
substituting plain bars with an equivalent cross section for FRP
in the method specified in the Standard Specification and by
disregarding the reinforcing steel in the inner concrete.
5. The elongation at failure of the FRP used was 1.8%. The
measurements of the strain suggest that different properties are
obtained by the use of conventional FRP with smaller elongations. Designers should pay attention to the mechanical properties of FRP, especially elongation.
ACKNOWLEDGEMENT
This study was conducted as a part of a joint study by the
"Committee for the Study of Research and Development of Marine
Structures Using Fiber Materials (FRP)", organized under the
Coastal Development Institute of Technology. The authors express their thanks to Mr. Hirota, T., Chairman, and all of the
other members of the Committee for their assistance and advice.
140
Sueoka et al
To ensure increased durability of this beam, we are considering the introduction of prestress force to the extent that no
crack occurs under dead load in the future. And if the studs
are to be used, it is necessary to consider their durability to
corrosion.
In this study we have had results of only a few experiments,
it is necessary to accumulate more data.
REFERENCES
(1) Japan Society of Civil Engineers: Standard Specification for
Design and Construction of Concrete Structures 1991
(Design), pp. 92-93
(2) Japan Society of Civil Engineers: Standard Specification for
Design and Construction of Concrete Structures 1991
(Design), pp. 85-86
(3) Hirota, T., Ohtsuki, N., Moriwake, A., Habuchi, T., "A Study
on Flexural Behavior of Beams Using Precast Concrete Slabs
as Permanent Forms Reinforced with FRP", Transaction of the
Japan Concrete Institute, Vol. 13-2, 1991, pp. 789-794.
(4) Tsuji, Y., Sekijima, K., Nakajima, N., Saito, H., "Mechanical Behaviors of Concrete Beam Reinforced with Grid Shape
FRP and Effects of Chemical Prestress 2", Concrete Research
and Technology, Japan Concrete Institute, Vol. 2, No.1,
1991, pp. 85-94.
(5) Japan Society of Civil Engineers: Standard Specification for
Design and Construction of Concrete Structures 1991
(Design), pp. 51-54
(6) Niwa, J.,
based on
Institute
1983, pp.
"Design Equation for Shear Capacity of Deep Beams
FEM Analysis",
Proceedings of Japan Concrete
2nd Colloquium on Shear Analysis of RC Structures,
119-126.
(7) Hirota, T., Ohtsuki, N., Naito, H., Hamasaki, K., "An Experimental Study on the Flexural Behavior of Concrete Beams
Reinforced with Grid-Shaped FRP Bars", Transaction of the
Japan Concrete Institute, Vol. 13-2, 1991, pp. 795-800.
FRP Reinforcement
141
TABLE 1 - TYPES OF SPECIMENS
Specimen
No.
1
2
3
4
5
6
Precast
form
Placed
Placed
Placed
Placed
Placed
Placed
Structure of specimen
Inner reinforced
Studs
concrete
-
Expected
failure mode
Flexural failure
Flexural failure
Flexural failure
Shear failure
Shear failure
Shear failure
-
Placed
Placed
-
Arranged
-
-
Placed
Placed
-
Arranged
TABLE 2 - CHARACTERISTICS OF FRP
FRP size
(Diameter)
ccc-
1 0
1 2
1 6
Elongation
Young's modulus
(kgf/cm')
(%)
(kgf/cm')
1. 6 8
1. 5 8
1. 5 9
1. 8
1. 9
1. 8
9. 8 2
8. 2 4
9. 1 4
Nominal cross
sectional area
(mm ')
Tensile strength
3 9. 6
5 5. 4
1 0 2 . 8
X
10 4
X
10'
TABLE 3 -MEASURED CONCRETE STRENGTH
Specimen
No.
1
2
3
4
5
6
Compressive strength
(kgf/cm')
Precast
Inner
form
concrete
4
4
4
4
4
4
3
4
4
6
3
3
5
2
2
5
8
8
-
2 9 3
2 9 3
-
2 7 8
2 7 8
Flexural strength
(kgf/cm')
Precast
Inner
form
concrete
5 0. 0
-
5
5
5
4
4
2 .
2 .
2 .
7.
7.
4
4
1
1
1
4 4. 5
4 4 . 5
-
4 7. 4
4 7 . 4
142
Sueoka et al
TABLE 4 - TEST RESULTS AND CALCULATED RESULTS
OF FLEXURAL SPECIMENS
Specimen
No.
1
2
3
lni tial
cracking load
(ton f)
Test
Cal.
4. 6
9. 2
6. 0
15. 7
15. 7
7. 5
Ul tiaate
Yield load of
Flexural
capacity
strain of FRP
steel bars
(X 10.')
(ton f)
(ton f)
Test C'al.
Test
Cal.
Test
Cal.
25. 7 27. 9 17400
14100
15100
26. 0
17100
53. 6 52. 6
22.3
51. 6 52. 6 13200
26. 0
22. 3
17100
Test:Test results, Cal.
Fall ure type
Concrete crushing
Concrete crushing
Concrete crushing
:Calculated results
TABLE 5 - TEST RESULTS AND CALCULATED RESULTS
OF SHEAR SPECIMENS
Specimen
No.
4
5
6
Shear
Ultimate
Capacity
Strain of FRP
(X 10- 6 )
(ton f)
Test
Cal.
Test
Ca I.
52. 2
44. 7
6000
6100
12 3. 0 101. 8
12000
11500
131. 0 101. 8
13500
11500
Test:Test results, Cal.
Failure type
Concrete crushing
Concrete crushing
Concrete crushing
:Calculated results
FRP Reinforcement
143
[Cross Section]
PO
f-L-.-
0
~
0
-
0
0
·~
0
0
I
0
'--'-----
~
~
1m
-
12i:
=
:;;
~
lo 11
I
I
:;;
T
!!.22
-
11-
0
~
0
=
lo LJ
I
ill
DIJ
.LL
0
~
\k.l
I
-
[email protected] 00-~00
I~
~ 16
I
~
4WI UU 4UU
I
[Side]
(Precast channel foro)
c
..
.
(Inner reinforced concrete)
125~
I. 200
12501
[Bottom]
(Precast channel fon)
i ii
ll]flilitfJJ&IiliEiffi:S?OLtrElfll
(Inner reinforced concrete)
Flexural specilen(No. 3)
~~•.~••~·•M~m~J
Shear specilen(No. 6)
Fig. !-Dimensions and bar arrangement of specimens
144
Sueoka et al
·~ [:.·l·l·l·i:···~6:[:• :• 1F·I···I·IIj ·~ [:JifEiiJ:j
fb
~
1moo·J. 200
3. 3oo
~~
81200.,, 600
1. too
Shear specloen
Flexural specloen
<D
Displacement transducer
(2) "shaped displaceoent transducer
<ID Two-directional d!splaceoent transducer
@ Strain gauge for concrete <ID Strain gauge for FRP
<ID Strain gauge for steel
Fig. 2-Arrangement of sensors
- - : Cracks
·---- :
F R
r
Grid
0 : intersections
1~:~~-J
6
6
Fig. 3-Flexural crack patterns on flexural specimens
If
FRP Reinforcement
145
50.------,-----.------,-----.-----,
I
I
--+----j_---mr,-~
50
I
,.... 40 ----+--
.8"'
-
30
10
o~~~--4--+~r==r~==T=~~
0
5
10 15 20 25 30 35 40 45 50
Displacement(mm)
Fig. 4--Load-displacement curves of flexural specimens
4-.0
Plain
I
:
steel bar
-----;-----T---
3.5
e
5 3.0
..c:
2.5
"0
2.0
....
"'
-"'
(.)
"'....
u
1
I
:
! .--:/Y"-_.,.,1
-
----~----
1
I
1.5
1.0
0
2
4
I
I
I
1
I
I
I
'
---~----f-~-
0.5
0
steel bar
-~----T----
I~
5
8 1 0 1 2 1 4 1 5 1 8 20
Strain of FRP ( x 10- 3 )
Fig. 5--Relationship between flexural crack width and strain of FRP
~
e
5 0.75
~ O.g)
"'
-"'
(.)
~ 0.25
u
0~~~+--+--+--+--+--+--+--+~
0
2
4
5
8
1 0 1 2 1 4 1 5 1 8 20
. Strain of steel ( x 10- 4 )
Fig. 6--Relationship between flexural crack width and strain of steel
146
Sueoka et al
0 2 4 5 8
500
"E
E
'-.../
...,
..c::
""
i No. 2
5t
() 1 Ot
• 15t
20t
<) 25t
0
400
I
I
<>
300
I
I
I
I
--~--,--,--
--1--...l-
I
I
200
Q)
100
24
20
15
12 8
Strain
4
0
cx1o-4 )
0 2 4- 5 8
500
"E
E
'-'
...,
..c::
""
5t
() 1 Ot
• 15t
20t
<) 25t
0
400
'
.
------
<>
300
200
Q)
:0::
I
I
I
I
I
I
!
I
--r---r-
-
__ L __ ,_
100
24-
20
15
12 8
Strain
4-
0
cx1 o-4 )
Fig. 7-Strain distribution in the center cross section
50
-
I
I
I
I
I
I
I No.2
-~--L-~ __ L_~--~--L-~--1
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
50
~
c::
0
...... 4-0
-+--~-~--~-~--~--~-~--
1
I
I
I
I
I
I
I
--~--L_J __ L_~--J--L-~--1
I
I
I
I
I
I
I
I
I
I
I
I
30-
'0
__,"'0
20
-~--------------------------.
1
I
10
-~
~:Transverse dislocation
•:Longitudinal dislocation
I
4-0 50 80 1 00 1 20 1 4-0 1 50 1 80
2 mm)
Dislocation
(X1
o-
50
-
~
c::
0
'0
__,"'0
:--:--:--:--:--:--:-:-I
50
-
4-0 -
·~
20
10
I
I
I
I
I No.3
~--~---~-~--~--1- -~--~-~---
1
30
I
_--~
1
I
I
I
I
__ I__ J __ L
I
I
I
I
I
I
I
I
I
I
I
I
__ I___ ( __ L
I
I
-~--I
--11 O:Transverse dislocation J
--if •
I
:Longitudinal dislocation
i
i
i
i
i
i
i
0~~-+--+-~--~--+--4--~--+-~
-50 20 4-0 50 8010012014-0150180
2 mm)
Dislocation
(X1
o-
Fig. 8--Dislocation between precast form and
inner concrete of flexural specimens
FRP Reinforcement
I'''& I
1.,.~ I
1"''1fl1b
zs
zs
zs
zs
zs
zs
I
Fig. 9-Crack pattern on shear specimens
14-0
I
120 ------1----.
..... 80
"'0
I
I
-----~------L----1
I
0
....:>
=""-oL-.L"'-1
I
c:
'0
I
- - - - - _j -
-;;::- 100
-~---
50
---:------
4-0
------:----
20
I
I
- - - - t - - - - - -I----
~-- - - -
0: No. 4
2:.:
No.
5
~------+~------~~--~==O=F:=N=o=·=6~
0-F0
5
10
15
20
Disp!acement(mm)
Fig. 10-Load-displacement curves of shear specimens
147
148
Sueoka et al
140,-,--,--.--.---.--.--.---.--.-~
120
~100
0
..... 80
"0
~
--'
50
40
20
0~~-+--~--+-~~-+--~--+-~--~
-50 20 40 50 80 100 120 140 150 180
2 mm)
Dislocation (X1
o-
140
,-,-----,---,---,--,~-,--~--.------,--~
I
120
~100
0
--'
I
I
I
I
I
1
I
I
I
I
I
I
I
__ L __ L_J __ L-~--~--L-~-1
I
I
I
I
I
I
I
80
--t--~-~--t-~--+--~-~--
50
-~--------------------
40
-1
"0
~
I
~--~-~--~-~--~--~-~--
1
1
20
I
I
I
I
I
I
I
[]:Transverse dislocation
II:Longitudinal dislocation
-~--r-,--r-~--~--r-~--
1
I
I
I
I
I
1
I
0~~-+--~--+-~---+--4---+--4--~
-5 0
20 40 50 80 100 120 140 150 1 80
2 mm)
Dislocation (X1
o-
Fig. 11--Dislocation between precast form and
inner concrete of shear specimens
SP 138-10
Aramid Fabric as a
Reinforcement for Concrete
by J. Kasperkiewicz and H.W. Reinhardt
Synopsis:
Several researchers have used aramid fibres as
prestressing tendons and as chopped fibres in concrete
due to their high strength and chemical resistance. In
this investigation, an aramid fabric is used as bending
reinforcement in concrete beams. Cracking, bending
capacity and ductility are determined experimentally.
It is shown how effective the aramid fabric can be if
applied in several layers as tensile reinforcement.
Keywords: Beams (supports); bending; cracking (fracturing); ductility; fabric;
fiber reinforced plastics; prestressing; reinforced concrete
149
150
Kasperkiewicz and Reinhardt
Dr. Janusz Kasperkiewicz, Assoc. Prof., Institute of
Fundamental Technological Research, Polish Academy of
Sciences, Warsaw, Poland. BS degrees in Civil Engineering, Mathematics and Physics. Main interests: structure and fracture of concrete, fibrous concrete and
other brittle matrix composite materials, stereology
Hans W. Reinhardt holds the chair for Building Materials at Stuttgart University and is managing director of
FMPA Otto-Graf-Institute at Stuttgart, Germany. He is
member of ACI Committee 446 - Fracture Mechanics.
ARAMID FIBRES IN CONCRETE
"Aramid fibres" is a term describing shortly a whole
group of man-made, organic polyamides. They are almost
fully crystalline and strong in the longitudinal direction, while characterized by weaker secondary bonds in
the transverse direction. Representatives of the group
are such products as Kevlar of DuPont, Technora of
Teijin, Arenka of Enka, Twaron of AKZO.
During recent years, various civil engineering applications of the aramid fibres embedded in a polymer matrix
were succesfully proposed as a prestressing reinforcement for concrete [ 1 to 4) . According to Shigeyuki
Akiham et al. [5], aramid fibres are also excellent
reinforcing fibres for cement matrices.
Aragrid, the product which is the topic of the present
investigation, is a square woven net approximately
10x10 mm made of threads composet of thin aramid fibres
with a width of a single thread being about 1-2 mm
(Fig. 1).
The threads are bundles of tiny filaments of 0,01 mm
diameter. To increase their durability and to protect
the filaments against mechanical damage the threads are
covered with protective plastic sizing.
As strength and size are concerned these fibres are
similar to asbestos. However, they are not carcinogenic. The tensile strength of some aramid fibres exceeds
high quality steel. The density of aramid is about 1.45
gjcm3 , its ultimate strain is between 2% and 5%. The
fibres have a very high stiffness which corresponds
roughly to a Young's modulus of 40 - 150 GPa.
Some technical data tables show tensile strengths of
2500-4000 MPa and a modulus of elasticity of 130-150
GPa.
FRP Reinforcement
151
Aramid fibres have basically good chemical resistance.
For example, Kevlar 49 was reported to show no degradation after 0.5 h boiling in 1% hydrochlorid acid or in
1% caustic soda. However, severe chemical environment
like 3 months immersion in HCl (pH 1), and in Na(OH) 2
(pH 13) resulted in about 80% strength loss. In the
intermediate stage, i. e. between pH 4 and pH 8 the
strength loss is expected to be about 20%. It is known
that uv light affects the stability of aramid. If the
fabric is embedded in concrete this is no problem any
more.
PREPARATION OF THE TESTS
Two types of aramid fabric were delivered for the pilot
tests. They are referred to in what follows as small
mesh (SM) and large mesh (LM) types. The SM and LM
types were different in the spacing of the main threads
of the net: about 10 mm or 13 mm respectively. The
fabric is delivered in long sheets, about 1 m wide, of
straw colour, of which the samples for testing can
easily be cut out using scissors.
Bare samples of the fabric were tested in direct tension in both directions: in the main direction and the
perpendicular direction. Since the number of threads
per unit length is different in warp and weft direction
it is preferred to use the carrying capacity of a single thread as a practical strength parameter instead of
the strength per cross-section.
The pilot test series on fabric non-embedded in concrete showed a carrying capacity per thread of 0.618
kN. A second series performed on material from another
delivery showed 0.779 kN per thread (Coeff. of Var.:
5.2%). The ultimate strain was found to be 1.5-2.3%.
The modulus of elasticity of the material cannot be
determined accurately because the exact cross-section
of the threads is unknown. A very rough evaluation of
the actual cross-section of the threads led to a modulus between 30 and 70 GPa.
The composition of 1 m3 fresh concrete
Table 1.
is given
in
The fresh concrete had a density of 2340 kgjm3 , with an
air content of 2.2%. After 28 days the cube compressive
strength was 44.5 MPa and the splitting strength was
4.9 MPa.
The beams were cast in steel forms of 500 x 100 x 100
mm, with the fabric ribbons placed at the bottom (see
Fig. 1, right). The concrete mix was sieved to get a
152
Kasperkiewicz and Reinhardt
small amount of mortar with a maximum grain size of 4
mm. A 5 mm thick layer of the mortar was poured at the
bottom of the form before a layer of fabric was placed
using additional steel sheets to fix the fabric in the
correct position, see Fig. 2. Then the remaining part
of the concrete was cast on the top of the fabric.
Altogether, 18 beams were prepared either with no reinforcements, with reinforcements of LM type fabric, or
reinforcement of SM type fabric. Two beams of smaller
depth were also made.
After positioning the fabric and filling the rema1n1ng
part of the form, the specimens have been vibrated and
stored for 7 days in a fog room, and 21 days in the
laboratory (2o·c and 60% RH). Standard strength tests
have been performed at 28 days, while the beams have
been loaded at the age of about 30 days.
THE TESTS: LOADING PROCEDURE AND OBSERVATIONS
All the specimens were loaded in pure bending. Force
and deflection readings were obtained via an X-Y plotter (F-o diagramm). On the side surfaces of the beams
pairs of grips - k-1 through q-r - for dial gauges have
been glued as shown in Fig. 3.
Each pair of grips were 100 mm apart, and about 6 mm
from the bottom or top surface of the beam. Deflections
of the beam were measured at mid-span. The width of the
cracks was measured by a low power, hand-held microscope with a scale, after stopping the loading procedure.
After placing the beams in the testing machine they
were loaded monotonically, with short breakes at every
multiple of 100 kN. At 100 kN, 200 kN, etc., readings
of horizontal elongation were taken over the four horizontal bases (k-1 through q-r; see Fig. 3).
The appearance of the first crack was accompanied by a
distinct, sudden decrease of the bending force, observed as a jump on the X-Y plot. At this moment the crack
position was identified and its width was measured
using the microscope. Thereupon the specimen loading
continued.
The range of the inductive gauge (LVDT) was about 5 mm,
and the plotter had to be disconnected before the specimen failed. The fabric was still effective long after
the first crack appeared. Large deflections at the
centre of the cracked beam were observed.
In two cases the unloading-reloading loops were created
(see Fig. 4) to see how stable the rupture process is
FRP Reinforcement
153
(specimens G and H).
RESULTS
The results from the measurements are presented in
Table 2. The equivalent stress values have been calculated as the maximum tensile stress in an uncracked
specimen of similar geometry. Comparison of the maximum
equivalent tensile stress values obtained with all the
tested specimens is shown in Fig. 5.
Most of the beams showed a similar behaviour. The first
part of the force-displacement plot was linear, i. e.
the specimen behaved linear elastic. Beyond the linear
response, the first crack could be expected. However,
no visible crack could be observed at that moment. When
the displacement was increased further a crack became
visible and started to propagate until it was stopped
by the action of the fabric. Simultaneously, the load
dropped to a certain value. These two stress values are
referred to as the first maximum and first minimum
stress, respectively. As in a composite beam, the fabric reinforcement carried now the tensile force while
the compressive force is supplied by the concrete. The
load could be increased until the full bending capacity
was reached. This was accompanied by a large crack
opening and vertical displacement and finally by the
fracture of the longitudinal threads of the fabric.
Table 2 gives this second maximum stress and the maximum maximum stress which is equal either to the first
or to the second maximum depending on the amount of
reinforcement.
The crack width value of Table 2 concerns the width of
the crack identified as soon as possible after the
first steep decrease of the bending force value. Practically it corresponds to the first minimum of the
bending force.
In the last two columns of Table 2 the observed number
of cracks and - if applicable - the spacing of the
cracks are given. Specimen 'I' was broken in shear: the
cracks appeared exactly under the points of application
of the bending force.
Fig. 5 shows a comparison of observed maximum equivalent bending stresses for batch I and II. The stresses
are calculated by dividing the bending moment by the
section modulus. The small letters refer to batch I
whereas the capital letters refer to batch II. Averaging the results of batch I and II the bending strength of the plain concrete beams was 4.94 MPa with 2% a
coefficient of variation, the strength of the beams
154
Kasperkiewicz and Reinhardt
with one layer of fabric was 4.96 MPa (37% c.v.) and
with two layers 7.38 MPa (15%).
ANALYSIS OF THE RESULTS
Fig. 6 to 8 show the individual results while, in Fig.
9, averaged diagrams are compared which were obtained
for specimens with no fabric, 1 layer of fabric, and
with 2 layers of fabric, respectively. At the small
number of specimens tested it was impossible to observe
any significant difference in the behaviour of specimens with large and small mesh fabric. Therefore, they
are treated here as one common group.
It was not known in advance whether placing a relatively smooth flat layer of fabric in concrete would create
a discontinuity surface of harmful mechanical effect.
It appeared not to be the case. Single fabric layer
reinforced beams were at least as strong as plain concrete beams and the interface between fabric and concrete did not appear as a discontinuity. The flexural
toughness is drastically increased instead.
One single layer did not increase the flexural strength
over the strength of plain concrete. Two layers of
fabric caused an increase of the carrying capacity by
50%. The flexural toughness index Is as definded in
ASTM Recommendation C-1018 (6] increased due to the
fabric: 1 layer of fabric caused a flexural toughness
index Is of about 2.0, 2 layers resulted in about 5.0
while it is 1.0 for plain concrete beams.
To reach the reinforcing effect of two layers of fabric
four 5 mm thick steel bars with yield strength of 220
MPa would have to be used . This comparison does not
take account of the crack arresting effectiveness of
both solutions, which in case of the aramid fabric is
expected to be more favourable than for steel bars.
During the test it was possible to decrease the rate of
loading near the peak value from 50 to 20 or even 15%
of the total testing machine stroke per second. At this
loading rate some strain-softening effect could be seen
even in case of plain concrete specimens. This rate
difference is probably only of secondary importance for
the testing of the fabric reinforced beams.
Although the measurements were realized using very
unsophisticated equipment it could be clearly observed
that, after the appearance of the first crack, the
fabric became 'fully activated' when the crack openend
about 0.2 mm. The relation between a certain critical
crack width and the number of fabric layers used is
FRP Reinforcement
155
subject of future studies.
An analytical experiment can be performed using the
arbitrary assumption that every single Aramid thread
which is composed of many filaments can carry the same
maximum force of 0.823 kN. The loading capacity of the
fabric reinforced beam should then depend directly on
the number of the longitudinal threads placed in the
tensile zone. This number varied in the present tests
together with the cross-section depth of the beam.
Comparing the calculated fictious bending strength with
the measured experimental results, a correlation coefficient of, r = 0.712 was found (Fig. 10).
This analysis indicates that the fabric can be treated
not as a 20 •textile' but rather as a bundle of longitudinal threads in the principal direction, separated
by perpendicular threads of no mechanical importance.
The transverse threads have only a technological
function, i. e. providing the even spacing of the principal threads.
Figs. 7 to 9 show a rather stable force deflection
behaviour which was especially clear in case of specimens 'G' and 'H' in Fig. 8. In a nonlinear structural
analysis, this behaviour would be modelled as elasticplastic or even rigid plastic which would enable the
determination of plastic hinges and yield lines.
CONCLUSIONS
Placing a single layer of aramid fabric changed the
flexural toughness·of the tested beams considerably. It
did not increase the carrying capacity compared to
plain concrete
beams. An increase of the carrying
capacity of 30 to 50% can be expected when using 2
layers of such a fabric. The conclusion concerns so far
only the geometry of the tested beams, and not any
other geometries.
The carrying capacity of a single aramid thread is
about 0.8 kN. It seems appropriate to treat the strip
of the fabric not as a two dimensional textile, but
rather as a collection of individual, longitudinal
threads of equal strength. However, in case of a slab,
the biaxial reinforcement should be activated. An even
more promising improvement of the strength of beams can
be expected by casting not 2 but 3 layers of the aramid
fabric. Further tests of aramid fabric reinforcement
should be carried out in order to get a more comprehensive understanding.
156
Kasperkiewicz and Reinhardt
ACKNOWLEDGEMENT
The authors are indebted to Dr. F. Paul for his substantial contribution to the experiments. The first
author acknowledges gratefully the grant of DAAD (German Exchange Service) which enabled him to stay at
Stuttgart University as a Visiting Professor.
REFERENCES
[1]
Burgoyne, c.J.
(ed.) Engineering application of
Parafil ropes. Imperial College of Science and
Technology. London, January 1988, 90 pp
[2]
Gerritse, A., Werner, J., Groenewegen, L.
Longterm properties of Arapree. IABSE Symp. Durability
of Structures. Lisbon, 1989
[3]
Gerritse, A., Werner, J.
First application of
Arapree. In "Fibre reinforced cements and concretes: Recent Developements", ed. R.N. Swamy, B.
Barr, Elsevier Appl. Science, London and New York,
1989, pp 50-59
[4]
Reinhardt, H.W., Gerritse, A., Werner, J.
Arapree: a new prestressing material going into practice. FIP notes Nr. 4, 1991, pp 15-18
[ 5]
Shigeyuki Akiham et al.
'Experimental study on
aramid fibre reinforced cement composites "AFRC".
Mechanical properties of AFRC with short fibres.
III Int. Symp. on Developements in fibre reinforced cement and concrete, FRC86, vel. I, eds. R.N.
Swamy, R. L. Wagstaffe, D.R. Oakley, Sheffield
1986, pp 125-131
[6]
ASTM - C 1018 - Standard test method for flexural
thoughness and first-crack strength of fibre reinforced concrete (using beams with third-point
loading) . Annual Book of ASTM Standards, Vol
04.02, 1985, pp 637-44
TABLE 1 - CONCRETE COMPOSITION (1m 3)
Portland Cement (PZ 35
Aggregate 0-16 rnm
Proportion of fraction
Proportion of fraction
Proportion of fraction
Water
F)
0-2 rnm
2-8 rnm
8-16 rnm
305 kg
1858 kg
35 %
35 %
30 %
177 kg (wjc=0.58)
TABLE 2- OBSERVED MAXIMUM EQUIVALENT BENDING STRESSES
Cross-section
h
b
[mm]
[mm]
No. of
fabric
layers
1st
max.
[MPa]
Stress values
1st
2nd
min.
max.
[MPa]
[MPa]
Max.
max.
[MPa]
Crack
width
No. of
cracks
Crack
spacing
[mm]
[mm]
Batch I.
a
b
c
d
e
f
g
h
i
100
100
100
100
100
100
100
100
44
100
100
100
100
100
100
100
100
100
0
0
0
1
1
2
2
2
1
4.85
5.02
5.10
4.79
4.49
4.63
4.43
5.39
4.72
-
-
1.44
1.48
3.39
3.44
3.74
3.93
-
-
2.39
6.68
6.05
7.40
4.72
4.85
5.02
5.10
4.79
4.49
6.68
6.05
7.40
4.72
-
0.42
0.35
0.22
0.15
0.04
0.02
1
1
1
1
1
1
2
2
1
-
90
60
-
Batch II.
A 100
B 100
c 100
D 100
E 100
F 100
G 100
H 100
I
48
100
100
100
100
100
100
100
100
100
0
0
0
1
1
2
2
2
1
4.82
4.97
4.86
5.25
4.87
4.67
4.33
5.61
4.72
-
2.26
3.22
3.18
2.74
3.87
2.83
-
3.83
4.35
7.55
7.31
9.31
5.66
4.82
4.97
4.86
5.25
4.87
7.55
7.31
9.31
5.66
-
0.24
0.08
0.12
0.19
0.03
0.16
1
1
1
1
1
2
1
3
2
-
20-30
(shear)
158
Kasperkiewicz and Reinhardt
Fig. 1-Aragrid -
schematic
Fig. 1-Aragrid in reinforced beams
f, g, h after testing
fabric
concrete
~ 5mm
Fig. 2-Mold with aramid fabric
FRP Reinforcement
159
F
micr.
k,l
k(or----t-,(~)
m(q)
;'
;
v
7i
435
i
n(r)
t
m,n
~
~
Fig. 3-Loading and measuring set-up
Equivalent stress [MPa]
a~--------------------------~
7~------------~~~rrl~~----1
s~------~~~~~--~+--rr---1
5-HII--~-+f-4-:JII!+T#------,.Y--+--1f-----i
4~-+1----/-tr--F-+---r--t£----lf--t-----1
3-t+-+r"--i-----T'---t--1-t'------+---1'-------1
2~--+r--~-7-++---~~~------1
1-11--+-1--1-~;,___,r-+----,~..,...::..------i
0-f------.-~--r-----T""---t---r---4
0
1
2
3
4
5
6.
Deflections [mm]
Fig. 4-Loading and unloading cutves - specimen G and H
160
Kasperkiewicz and Reinhardt
10
Maximum equivalent stresslMPa}_
9
8
7
6
6
4
3
2
Ab ~ c Cd De E
a
0
0
0
0
1
1
Fg
f
2
2
I
H i
Gh
2
Number of layers
Fig. 5-0bserved maximum equivalent bending stress; small letters: Batch I,
capital letters: Batch II
8 Eqivalent stress [M Pa]
7
j:t
5
4
~
3
\
2
\...
1
0
0
1
2
3
I
I
4
5
6
Deflections [mm]
Fig. 6---Equivalent stress-deflection diagrams for beams
a, b, c, A. B, C (plain concrete)
FRP Reinforcement
Eqivalent stress [MPa]
8~----------------------------~
7~--------------------------~~
6~----------------------------~
5-Hr-------------~~--+-------~
4-Hffi-------~C---£---~~~~----~
3~~~~~~~----=9~~~~
2~~~~--------~+-----~----~
1~------------------~~~~--~
0~----~---T----~---;----~--~
0
Deflections [mm]
Fig. ?-Equivalent stress-deflection diagrams for beams
d, e, i, D, E, I (one layer of fabric)
Eqivalent stress [MPa]
8~----------------------------~
74-----------~~~~~~~---4
6,_--~~~~~~~~~~~-4
5~~~~~~~~--~--~~----~
4~~~~~--------~--~~----~
3~~~------------~--~~--~
2~--------------------~~----~
1~--------------------~~----~
04---~~--~----~~~----~--~
0
1
2
3
4
5
6
Deflections [mm]
Fig. 8-Equivalent stress-deflection diagrams for beams
f, g, h, F, G, H (two layers of fabric), unloading/reloading
loops omitted
161
162
Kasperkiewicz and Reinhardt
s
one layer
2
plain concrete
0+------+------+-----~----~~----;-----~
s
2
0
5
4
8
o [mm]
Fig. 9-Generalization of all obtained results.
Equivalent stress versus central deflection
Obeerved
10
_JL
II
8
~-
7
6
2
~
v
v
_...,- ~
4
3
~
_...,-
6
~
~
r • 0.712
0
0
2
3
4
6
6
7
8
II
10
Calculated
assumptions :
·JeW)
t
(1/3)h
(1/9)h
r- (Sh/9)·6
(2/3)h
~
M•rxnxF
o=
c-6mm
61\Vb~
Fig. 10-Analysis with single thread effectiveness for the
equivalent bending strength, observed versus calculated values
SP 138-11
Experimental Study on
Tensile Strength of
Bent Portion of FRP Rods
by T. Maruyama, M. Honma,
and H. Okamura
Synopsis: In this study, the authors introduced bends into FRP rods and
then after embedding them in concrete, applied loads to investigate the
tensile strength of the bent portions. The results show that the FRP rods
ruptured at the bend, and that tensile strength decreases as the curvature
of the bend increases. They also indicate that the tensile strength of the
bend varies with the strength of concrete, the fiber type and the method
by which the rods are manufactured.
Keywords: Bending; fiber reinforced plastics; high strength concretes; shear
properties; strength; tensile strength
163
164
Maruyama, Honma, and Okamura
1.
INTRODUCTION
For FRP rods to be used as concrete reinforcement, they must be
bent in the same way as hooks and stirrups on the anchor end of steel
reinforcement if conventional structural specifications are to be observed.
Information on the tensile strength of the bent portion (bend strength) is
therefore required if FRP rods are to be used as reinforcement in practical
applications. Previous studies indicate that when used as shear reinforcement in beams, FRP rods have failed at the bent portion (1). It is also
reported that tensile strength drops as bend radius decreases (3). These
studies were mainly concerned with hand-formed FRP grid reinforcement, and no reports exist concerning the bend strength of FRP rods with
a 50% - 60% fiber content that have been automatically produced on a
braiding machine or the like.
In this paper, the authors investigated the bend strength of concreteembedded FRP rods by conducting tensile tests on typical carbon or
aramid FRP rods for three different bend radii and two different strengths
of concrete.
2.
2.1
OUTLINE OF EXPERIMENT
FRP Rod Bending Process
The FRP rods used in the experiments were 7-strand CFRP (Carbon
Fiber Reinforced Plastic) rods, pultrusion CFRP rods, and braided AFRP
(Aramid Fiber Reinforced Plastic) rods. Steel bar reinforcement was also
used for comparison purposes. The properties of the rods are given in
Table 1. AFRP rods are characterized by being weaker and having a
smaller elastic modulus, but also having a greater elongation capacity,
than CFRP rods.
The external appearance of the rods is shown in Fig.l. Strand CFRP
is made by twisting six strands around a single central strand; pultrusion
CFRP is made by wrapping rovings spirally around the rod; braided AFRP
is made by weaving several pre-pregs (resin-impregnated continuous
fiber ravings and mat) into a braid.
As shown in Table 2, three internal radii for the FRP rod bends were
chosen: Smm, 15mm and 25mm. In the case of strand and braided rods,
braided pre-pregs were first bent over metal bars of the required radius
and the epoxy resin matrix was then heat hardened. This process causes
FRP Reinforcement
165
the bend cross section to collapse and flatten, inducing slack in the element
wires on the inside of the bend.
With the pultrusion CFRP rods, the pre-pregs were heat-hardened
after being bent over semi-circular grooved metal molds having radii
almost equal to the rod diameters. As a result, these rods retained their
circular cross section through the bend.
2.2
Specimens and Testing Method
The shapes and dimensions of the specimens are shown in Fig. 2.
The embedded length of the FRP rods was SOmm along the tension axis
and 1SOmm at right angles to it. The transmission of tensile load was
monitored by a strain gauge fitted 70mm from the beginning of the bend.
The concrete mix proportions are shown in Table 3, high early strength
portland cement being used. Two strengths of concrete were used: high
strength concrete ~about SO N I mm 2 ) and ultra high strength concrete
(about 100 N /mm ). The specimens were moist-cured for about a week,
after which they were stored in air until the experiment. As shown in Fig.
2 and Photo 1, in the tests specimens were fastened by nuts at the grip, and
tension applied using a 1SO kN hydraulic jack. A SO kN load cell was used
to measure the loading. The load was normally applied statically. Loading
was repeated several times on a number of the specimens.
3.
3.1
RESULTS AND OBSERVATIONS
Rupture
In the case of high strength concrete (SO N/mm 2), rupture in the
FRP rods occurred mostly on the loading side at the point where the
bending process had been initiated, as Photo 2 shows. Though yet to be
confirmed, it is thought that the fibers tear gradually, those on the inside
of the bend going first. In the strand CFRP rods, it appears that rupture
started with one of the seven strands tearing, the tensile force then transferring to the other strands in order until they all tore. With the pultrusion
CFRP rods, once some of the fibers started to tear the rupture advanced
very rapidly and occurred almost instantaneously. In the case of the AFRP
rods, the fibers appear to have torn more gradually due to their greater
elongation capacity.
166
Maruyama, Honma, and Okamura
With ultra high strength concrete (100 N/mm 2), the way and the
positions at which the rods ruptured were practically the same as with 50
N I mm2 concrete. In the case of steel bar reinforcement, rupture occurred
in the straight portion outside the concrete for all the bars, producing the
same results as in tension tests.
3.2
Tensile Strength of Bent Portion
The results of tests usin~50 N/mm 2 concrete are shown in Table 4,
and of those using 100 N/mm concrete in Table 5. The curves in Fig. 3
show the ratio of the tensile strength of the straight portion and of the bend
for different bend radii. With all types ofFRP rod, bend strength decreased
as the radius became smaller. While the way in which the strength decreased varied with the rod type, the decrease was generally linear over a
range of curvature of 0 to 0.05 mm-1, while for larger curvatures the rate
flattened out. With pultrusion CFRP rods (Fig. 3(a)), although the amount
of data is not altogether sufficient, a sharp decrease in bend tensile strength
was observed with a 25mm bend radius but no significant changes were
observed with 15mm and 5mm radii. Bend strength, for example, was
about 60% that of straight-line strength for 25mm and 15mm radii, and
about 55% for 5mm. These results suggest that pultrusion CFRP rods tend
to be affected by the bending process for the following reasons: the fibers
along the inside of the bend do not bear the tensile force uniformly due to
the slackness induced; the dense fiber arrangement; the carbon fibers have
a low elongation capacity and are highly brittle when subject to non-axial
stress.
For strand CFRP rods using 50 N/mm 2 concrete (Fig. 3(b)), bend
strength decreases hyperbolically as curvature increases: to about 65% of
straight-line strength for a 25mm radius, 60% for 15mm and 50% for 5mm
radii. Although of similar size for the 5mm radius, the decreases in
strength for 25mm and 15mm radii are smaller than with pultrusion CFRP
rods. This is thought to be because the strands have a low bending rigidity,
because gaps in the cross section allow the seven strands to move as the
rod deforms, and because rupture takes place gradually and not instantaneously. The same general trends were observed with 100 N/mm 2 concrete.
Fig. 3(c) shows tensile strength curves for braided AFRP rods. The
tensile strength of these rods also decreases hyperbolically, though to a
lesser degree than with CFRP rods, dropping to between 60% and 80% of
straight-line strength for a 25mm radius, between 60% and 65% for 15mm
and between 45% and 55% for 5mm radii. This is thought to be because
aramid fibers have a greater capacity of elongation, they have a smaller
FRP Reinforcement
167
modulus of elasticity, and their braided structure gives them a lower
bending rigidity and makes them deform more easily. Further, the rupture
is thought to develop more gradually because AFRP rods have a relatively
large Poisson ratio, around 0.6.
It is known that repeated stress loading causes fatigue rupture at
the bent portion of steel reinforcement stirrups, so more data relating to
FRP rod tensile strength needs to be gathered and examined thoroughly.
3.3
Influence of Concrete Strength
Fig. 4 shows the relationship between rod tensile strength and
concrete strength. Pultrusion CFRP rods (Fig. 4(a)) behaved much the
same with both 50 N/mm2 and 100 Njmm2 concrete. For strand CFRP
rods (Fig. 4(b)), the use of 100 N/mm concrete realized greater tensile
strengths for all bend radii, the increase being around 200 N I mm 2 or 15%
in terms of bend strength. With AFRP rods (Fig. 4(c)), the influence of
concrete strength became more pronounced as bend radius decreased.
Though unconfirmed, the reason concrete strength had so little
effect on bend strength in the case of pultrusion CFRP rods is thought to
be because the tensile force caused the ravings wrapped around the rod
for deformation purposes to pull away and lose adhesion with the concrete.
With the strand CFRP and AFRP rods, meanwhile, the better adhesion with the concrete caused a decrease in the amount of tensile force
transmitted to the bent portion of the rods, thus demonstrating the effect
of concrete strength on bend strength.
3.4
Transmission of Tensile Forces Through Bent Portion
Fig. 5 shows a typical relationship between the strain in the loading
side of the rod and the tensile force strain transmitted through the bent
portion. Tensile force transmissibility differs with the type of rod and
bending method, but increases in the order pultrusion CFRP rods, strand
CFRP rods, braided AFRP rods. The following overall trends were observed: the larger the bend radius, the greater the tensile force transmitted;
and the lower the concrete strength, the greater the tensile force transmitted.
168
Maruyama, Honma, and Okamura
For pultrusion CFRP rods (Fig. S(a)), the almost circular cross
section and smooth inner surface of the bent portion means low adhesion
with the concrete, so that tensile force is transmitted to the anchored side
at an early stage. In the case of the 5mm bend radius there was virtually
no transmission of tensile force to the anchored side; even at a load of 27
kN just before rupture, the strain on the anchored side was a mere 480!!,
compared with 6700!! on the loading side, giving a transmissibility of
around 7%. Transmission of tensile force gradually increased with bend
radius. In the case of a 25mm radius, transmissibility just before rupture
reached about 22%. For 100 N I mm 2 concrete, the slope of the strain curve
steepens and transmissibility declines further. These observations suggest
that in the case of a small bend radius, tensile force is retained by contact
pressure between the FRP rod and concrete at the bent portion.
In the case of strand CFRP rods (Fig. S(b)), the load at which
transmission of tensile force to the anchored side starts is not significantly
affected by bend radius, and there were no great changes in residual strain
on the anchored side. Transmissibility in the final stages was between 26%
and 28% for the Smm radius, between 26% and 28% for the 15mm and
around 36% for the 25mm radii. The following are thought to the reasons
why transmissibility is so much larger in strand than pultrusion CFRP
rods: adhesive strength is smaller, at about 5 N I mm 2; inter-strand spaces
allow individual strands to move relatively to each other; the greater
deforming capacity of the rod enables it to transmit the tensile force more
linearly. With 100 Nlmm 2 concrete, the slope of the strain curve steepens
and transmissibility decreases.
With braided AFRP rods [Fig. S(c)], for bend radii of 15mm and
over, the strain curves remain similar regardless of concrete strength. This
indicates that transmission of tensile force to the anchored side occurred
at an early stage. In the end, while transmissibility was only about 10% for
a Smm radius, it increased sharply to about 46% for 15mm and to 73% for
25mm radii. The following are considered to be some of the reasons such
a large transmission of tensile force occurs in AFRP rods: the rod surface
is smoother; adhesive power is smaller; the greater elongation capacity of
aramid fibers, and smaller bending rigidity. For 100 Nlmm 2 strength
concrete, the trends observed above are almost the same and the effect of
concrete strength seems to be only slight.
4.
SUMMARY
As a result of tensile strength tests on the bent portion of CFRP and
AFRP rods with bend radii of 5mm, 15mm and 25mm respectively, the
following have been verified within the limits of the experiments:
FRP Reinforcement
169
CFRP and AFRP rods all ruptured at the bent portion, the location
of the rupture being the beginning of the bend on the loading side. No
rupturing at the bend occurred with steel-bar reinforcement.
(1)
(2)
The tensile strength of the bent portion of FRP rods tended to
decrease hyperbolically as the curvature of the processed portion increased. In the case of strand CFRP rods, for example, strength was about
65% of straight-line tensile strength for a 25mm radius bend, about 60%
for a 15mm bend and about 50% for a 5mm bend.
(3)
The rate at which tensile strength decreased varied with bend
radius, fiber type, and bending method employed.
(4)
The differences in bend strength seen between 50 N/mm 2 and 100
N/mm2 concrete varied to some degree with the type of rod, though the
higher strength concrete did produce higher bend strengths.
(5)
The amount of tensile force transmitted to the anchored side
through the bent portion increased with the bend radius, but the degree
of increase varied with the type of FRP rod, bending method, etc. The
increase was largest in AFRP rods, followed by strand-CFRP and then
pultrusion-CFRP rods, in that order.
REFERENCES
(1)
Ozawa, K., Sekisima, K., and Okamura, H., "Fatigue Properties of
FRP Reinforced Concrete Beam in Bending," Annual Report of JCI, No.9,
1987, pp.269-274.
(2)
Wakui, H., et al, "Shear Resisting Behavior of PC Beams Using FRP
as Tendons and Spiral Hoops," Annual Report of JCI, No. 11, 1989, pp.835838.
(3)
Miyata, S., et al, "Experimental Study on Tensile Strength of FRP
Bent Bar," Transactions of The Japan Concrete Institute, Vol. 11, 1989,
pp.185-192.
(4)
Okamura, H., Farghary, S.A., and Ueda, T., "Behaviors of Reinforced Concrete Beams with Stirrups Failing in Shear under Fatigue
Loading," Proc. of JSCE, No.308, Apr.l981, pp.109-122
170
Maruyama, Honma, and Okamura
TABLE 1 - FRP ROD PROPERTIES
Type of FRP Rods
CFRP
strands
CFRP
pultrusion
AFRP
braided
Diameter(mm)
7.5
6.0
8.0
6.0
Sectional Area (mm2)
30.4
28.3
50.0
28.3
2,090 '
1,740 '
1,370'
Tensile Strength (N/mm 2)
(N/mm 2)
Elastic Modulas
Maximum Strain (%)
137,300
144,200
62,900
1.6
1.2
2.2
' Breaking Point
" Yield Point
TABLE 2- FRP ROD BENDS
Radius r(mm) Curvature 1/r (mm- 1)
5
0.200
15
0.067
25
0.040
~
Steel
bar
510"
207,900
--
TABLE 3 - MIX PROPORTIONS AND COMPRESSIVE STRENGTHS
Concrete Max.Size of Slump Water Sand
Unit Weight (kg/m 3 )
Strength Aggregate
Cement Aggregate
(em) Ratio Ratio
(N/mm 2 )
(mm)
Water Cement Silica Sand Gravel Admix.
50
20
9
0.45
0.44
162
360
--
804
1031
4
100
20
11
0.31
0.42
124
400
80
760
1053
12
Concrete
Strength
(N/mm 2 )
Compressive Strength (N/mm 2 )
Standard Curing at 14 Days
(Ave.)
Same Curing as Specimen
(Ave.)
50
48.7
48.9
48.5
(47.9)
44.6
42.7
55.8
(49.2)
100
87.6
89.7
91.4
(89.5)
101.4
100.0
95.4
(98.9)
172
Maruyama, Honma, and Okamura
TABLE 4 - FRP ROD BEND STRENGTH (50 N/mm 2 )
Type of FRP Rods
r~5
CFRP
Strands
CFRP
Pultrusion
AFRP
Braided
Steel
Bar
Load(N)
31,685
26,458
32,332
14,523
Strength(N/mm 2 )
1,045
941
647
510
0.499
0.540
0.470
1.000
Load(N)
36,137
30,390
45,424
--
Strength (N/mm 2 )
1,245
1,078
912
--
0.597
0.620
0.661
--
Ratio
r~15
Ratio
r~25
Load(N)
41,835
--
52,318
--
Strength(N/mm 2 )
1,373
--
1,049
--
0.659
--
0.762
--
Ratio
TABLE 5 - FRP ROD BEND STRENGTH (100 N/mm 2)
Type of FRP Rods
r~5
CFRP
Strands
CFRP
Pultrusion
AFRP
Braided
Load(N)
36,618
26,193
43,375
Strength(N/mm 2 )
1,206
932
873
0.577
0.534
0.632
Load(N)
44,169
24,536
48,719
Strength(N/mm 2 )
1,451
873
971
0.696
0.500
0.710
Load(N)
48,043
29,420
55,270
Strength(N/mm2 )
1,579
1,040
1,108
0.757
0.600
0.805
Ratio
r~15
Ratio
r~25
Ratio
FRP Reinforcement
CFRP
strands
CFRP
pul t rusion
~braided
~AFRP
Fig. 1-Types of FRP rod
Grip
FRP Rod
Strain Gage(Loading Side)
Strain Gage
(Anchoring Side)
200
50
Anchor
+-
50
70
80
100
400rrun
Fig. 2-Specimen dimensions and loading systems
173
174
Maruyama, Honma, and Okamura
(c)AFRP
Braided
(b)CFRP
Strands
(a)CFRP
Pultrusion
1.0
0.8
0-
0.6
-- I
-·-
I
0.4
Concrete Strength
o---o 100N/mm2
•-- -e 50N/mm 2
25 15
0
25 15
25 15
5
Radius of Bent Portions (mm)
5
0.1
0 2 0
0.1
0.1
0.2 0
Carvature of Bent Portions (mm-1)
0.2
0
Fig. 3-Relationship between tensile strength and bend radius
2,000
~
N
§
......_
z
~
1,500
bO
(f)
1,000
•
0
~--·-·-6
QJ
I
......
.....
(f)
cQJ
E-<
~:. --~--_,)
v
..c:
.....
c
QJ
I.<
.....
(b)CFRP
Strands
(a)CFRP
Pultrusion
... --
/
t:::r-'
----·
o-
.-(
o---,. ...... -'-
I
I
yo'
o---o r=25
500
(c)AFRP
Braided
r=15
l'r-.-!::. r=5
50
I
100
I
50
100
50
100
Concrete Strength (N/mm 2 )
Fig. 4--Relationship between bend and concrete strengths
FRP Reinforcement
30
20
3,20010
800
1 '600
60
QJ
'0
......
(J'J
~
~.GOO
10
gp
....:!
'20
....:!
......
'0
<ll
0
z
QJ
rl
20 ......
4. 800 -
(/)
i:l
QJ
QJ
.r::
E-<
+-'
0
i:l
0
1,500
i:l
......
3.200
<ll
~ 18,000 ~ _r_:~:RP ~~rnided)l~----~~~~~::c=--_-)t-_..-.. -... ~- 50
-L
..
.·
12.000
6, 000
1/
-~----~~-- //1
I/ /
~
/[ ..
/~/ /.
-~!!!
;{! .·i.-!
;_:..
! l ;:
1- -
__:~"~
t----
40
cr.- 20
,]_
Concreto Sl:n"!ngt,,
~-- ~ON/unn2
!: l ~
-------- 100N/nunl
0 ~r~J--~-G~,.~o~o~o~~J~2.~D~O~D~~~--~--~--~-JO
:n
n
G.
noo
12. ooo
Strain on the Anchoring Side (*10- 6)
Fig. 5--Loading side strain and strain transmitted through bend
175
176
Maruyama, Honma, and Okamura
Photo 1-Tensile loading methoa
Photo 2-FRP rod rupture
SP 138-12
Research and Development
of Grid Shaped FRP
Reinforcement
by T. Fujisaki, T. Nakatsuji,
and M. Sugita
Synopsis: The grid shaped FRP reinforcement has been developed
to prevent the deterioration of concrete structures owing to
the corrosion of reinforcement. This reinforcement is made
from high-strength continuous fibers impregnated with resin
and formed into a grid shape to ensure bond with concrete. At
the time when this development has been carried out, joint
R&D with some universities as well as a technological
development project in our company has been done to clarify
fundamental properties of this reinforcement and structural
behavior of reinforced concrete members with it. Applications
of this reinforcement to actual structures started on civil
engineering structures such as tunnels, LPG tanks, and so on.
In Japan, applications of Advanced Composite Materials to
building structures require the approval of the Minister of
Construction. Therefore, to apply this reinforcement in
precast concrete curtain walls, heat resistance test of it,
fire resistance test of panels using it and so on were
conducted to obtain the approval of the Minister of
Construction. This is the first time that FRP reinforcement
was used in building structures in Japan. Hereafter we
attempt to apply this reinforcement in box frame type
reinforced concrete structures.
Keywords: Composite materials; deterioration; fiber reinforced plastics; fire
resistance; flexural strength; reinforcing materials; specific heat
177
178
Fujisaki, Nakatsuji, and Sugita
Tadashi Fujisaki is a research engineer at the Technology
Division, Shimizu Corporation. He graduated from Tokyo
Institute of Technology in 1977. His research interest is in
the area of FRP reinforcement and its application in building
structures.
Teruyuki Nakatsuji is a manager at the Technology Division,
Shimizu Corporation. He graduated and received his MD (1973)
and DD (1982) from Osaka University. He has been engaging in
research of structures using new materials.
Minoru Sugita is a general manager at the Technology Division,
Shimizu Corporation. He graduated and received his MD from
Waseda University in 1966. He has been taken part in many
development projects on structures using new materials. He is
a member of ACI 440 Committee.
INTRODUCTION
Recently in Japan, reinforced concrete structures have
sometimes been seriously deteriorated by the corrosion of
reinforcing steel bars. Over the past few years, a
considerable number of researches have been made on FRP
reinforcement to improve the durability of concrete
structures against this deterioration. In this current, we
have developed a grid shaped FRP reinforcement for concrete
structures (1).
The grid shaped FRP reinforcement is made from highstrength continuous fibers (carbon, glass, and others)
primarily impregnated with vinyl ester resin. It is formed
into a grid shape by a filament winding process, therefore
bond with concrete for it is the result of mechanical action
at cross points of the grid. This is the first feature of
this FRP reinforcement. The second feature is that prearrangement by integral molding at FRP factories improves the
work productivity .
This paper first provides an overview of research,
development, and application of the grid shaped FRP
reinforcement. And then describes the heat resistance of it,
fire resistance and flexural behavior of concrete panels
using it as reinforcement, on the basis of the results of the
tests conducted for the purpose of applying it to precast
concrete curtain walls. Further introduces an application
example to precast concrete curtain walls.
FRP Reinforcement
179
OVERVIEW OF RESEARCH, DEVELOPMENT AND APPLICATION
Research and Development
Table 1 is an overview of research and development (R&D)
of the grid shaped FRP reinforcement to the present. R&D was
begun in 1984. In the beginning, fundamental R&D was done for
a technological development project in our company. Later,
joint R&D was done with some universities.
As fundamental R&D for a technological development
project in our company began, we clarified tensile properties
of this reinforcement (See next section "Tensile Properties"
), and conducted studies related to the lap splice joint
mechanism (the required length of lap splice joint) by
bending tests ( 1). These results proved that this
reinforcement is effective as reinforcement for concrete
structures.
In joint R&D done with some universities, we first
clarified shear and compressive properties -- fundamental
mechanical characteristics -- and then determined the
mechanism by anchoring to concrete. It became clear that shear
strength is about 50% of tensile strength within the limits
of that experiment, and that anchorage capacity greater than
tensile strength can be secured by developing this
reinforcement more than 2 cross points into concrete within
the limits of that experiment (1). Additionally, the chemical
resistance of this reinforcement was examined by a relaxation
test. Judging from the test results, this reinforcement is
stable under alkaline and acid environments (2). Moreover, to
determine the behavior of concrete with this reinforcement in
fires or at high temperature, heat resistance of this
reinforcement was also clarified (See next section "Heat
Resistance" ) .
We then tested, in cooperation with universities to
verify applicability of this reinforcement to concrete
structures. For wall panel development, in-plane bending-shear
tests under reversed cyclic lateral loading and bending tests
under out-of plane loading were done, and structural
characteristics was clarified (3, 4, and see next section "
Flexural Behavior" ). For slab panel development, beamanchoring methods, flexural behavior, and long-term deflection
were comfirmed. To improve flexural behavior, we studied the
possibility of using the pretension prestress method (4).
Also, wall and slab panel fire resistance was studied (see
next section ·Fire Resistance" ) . Furthermore, for
applications to columns and beams, we confirmed flexural,
shear, and fatigue behavior and so on.
180
Fujisaki, Nakatsuji, and Sugita
Application
Application of the grid shaped FRP reinforcement to
actual structures started on civil engineering structures in
1986. It has been widely used as a reinforcement for shotcrete
in the New Austrian Tunnelling Method and in the slope
protection, and as a reinforcement for tunnel lining (1, 5).
Fig. 1 shows an example of tunnel lining.
Lately, this reinforcement was adopted as a reinforcement
for shotcrete of landslide protection of liquid petroleum gas
inground storage tank, as shown in Fig. 2. The setting rate
became much higher than that of conventional welded wire
fabrics, because of this FRP reinforcement's light weight and
large size ( 4).
Fig. 3 Shows a installing work of a fender concrete plate
using this reinforcement to prevent the corrosion of
reinforcement by the salt attack of sea water (4).
Fig. 4 shows a prestressed concrete pedestrian bridge.
This reinforcement was used as tendons at bottom and as
reinforcement at top of this slab. Five years have passed
since this bridge was built, with no problems resulting thus
far (4).
PRECAST CONCRETE CURTAIN WALL
Tensile Properties
Fig. 5 shows the tension testing method of the grid
shaped FRP reinforcement. Both ends of the specimen are
strengthened with steel pipes, inner parts of which are
filled up epoxy resin.
Fig. 6 shows the load-strain relationships of three types
of this reinforcements. The tensile strength of them is 1.2
times that of a deformed steel bar (SD345) of the same bar
number. As only glass fibers or carbon fibers are used in this
reiforcement, the relationship between load and strain is
linear until rupture, as is the case with G8 and C8 in Fig.
6. However, by combining two kinds of fibers, the
relationship between load and strain becomes non-linear due
to shear flow failures in the longitudinal reinforcement, as
is the case with H8 in Fig. 6.
Heat Resistance
In order to examine the heat resistance of the grid
shaped FRP reinforcement (H8) in the case of suffering high
temperature at fire, we conducted two types of heating
tension tests; one is the tension test under heating, the
FRP Reinforcement
181
other is that after cooling to room temperature.
Fig. 7 shows the heating tension testing method of this
reinforcement. Both ends of the specimen are strengthened
with resin mortar.
The test procedure of the former is the following; 1)
increase the temperature to test one at a rate of 5~ per
minute; 2) then, hold there until the tension test is over;
3) start the tension test 10 minutes later when the
temperature reaches the test one. And, the test procedure of
the latter is the following; 1) increase the temperature to
test one at a rate of 5 oc per minute; 2) then, hold there for
20 minutes; 3) decrease the temperature to room one
naturally; 4) conduct the tension test at room temperature.
The results of the tests are shown in Fig. 8, expressed
in terms of the ratio of the heated tensile strength to the
unheated tensile strength. In the results of the tension test
under heating, this ratio declines gradually from room
temperature to 100~, and remains about 60% of the unheated
tensile strength at around 100 to 250 oc. In the results of
the test after cooling to room temperature, this ratio returns
to 100. That is to say, the tensile strength after heat
treatment is the same as the unheated one.
Consequently, by designing the concrete structures with
the grid shaped FRP reinforcement on condition that the
tensile strength of this reinforcement under heating will
decrease, it is possible to use this reinforcement in
concrete structures.
Fire Resistance
To apply the precast concrete panels with the grid shaped
FRP reinforcement to curtain walls, we conducted fire
resistance tests for 30 minutes, as specified in Japanese
Industrial Standard (6).
Fig. 9 shows a test specimen. The panels used are 3600mm
high, 1500mm wide, and 150mm thick. The reinforcements used
are H8 (@100mm sq. spacing), G8 ( n shape) and H8 (bar type).
Specimen consists of two panels which are separated along
their width by a joint (25mm). These panels are composed the
size of 3600mm high, 3025mm wide, and 150mm thick by steel
beems and fasteners at the back surface. The compressive
strength of lightweight concrete used in the panels is 288 to
317 kgf/cm 2 , and the water content in percentage of total
weight is 5 to 8 %.
Fig. 10 shows the relationships
between temperature and heating time. The temperature of
heating surface was increased along the heating temperature
curve. The results of the heating test and the impact test
dropping a weight at the heating surface are as follows; 1)
the maximum temperature of back surface of lightweight
concrete is no more than 260~ which is the flash point of
wood; 2) the maximum temperature of reinforcement at the
heating surface side is about 300 oc , we can estimate that the
182
Fujisaki, N akatsuji, and Sugita
tensile strength of this reinforcement at this temperature
will maintain about 60% of that at room temperature (See
scetion "Heat Resistance" above); 3) the maximum out-of-plane
deflection at the center of this specimen is 14mm; 4) the
maximum crack width is 0.55mm, comparatively narrow; 5) at
the impact test, only a little hollow is produced.
These results shows that it is not necessary to consider
the structural capacity as well as the fire resistance. So we
judged that the concrete panels with the grid shaped FRP
reinforcement are succeeded in the fire resistance test.
On these grounds, we have come to the conclusion that it
is possible to use the grid shaped FRP reinforcement in
precast concrete curtain walls.
Frexural Behavior
To apply precast concrete panels with the grid shaped FRP
reinforcement to curtain walls, we conducted a bending test
to confirm the wind resistance.
Three specimens are prepared for this test; one is the
same panel with Fig. 9; next is the concrete panel after the
fire resistance test which was descrived in the preceding
section; last is the concrete panel with Fig. 9, but FRP
reinforcements are replaced with round steel bars, 6¢, and
this panel is designed as a spandrel typed curtain wall in a
50m-high building. The compressive strength of lightweight
concrete used in the panels is 316 to 338kgf/cm 2 •
The load is applied to the specimen by four points
bending with bearing span of 3000mm and loading span of
1000mm.
The load-deflection relationships of specimens are shown
in Fig. 11. The following results are obtained; 1) the
flexural cracking load exceeds the design load; 2) a flexural
compressive failure occurs in the specimens with the grid
shaped FRP reinforcement; 3) the maximum bending strength of
the precast concrete panels declines little, even after fire
resistance test; 4) the elastic theory asuuming the total
cross-sectional area before flexural cracking and that
neglecting tensile stress of the concrete after flexural
cracking are applicable to calculate a initial stiffness (Ig)
, a post-cracking stiffness (Icr), and a maximum strength (Pu)
. 5)addionally, we clarified through both test and analysis
using Navier hypothesis that maximum crack width under the
design load is 0.3mm (the allowable crack with of curtain
walls) or less, even when there is flexural cracking.
On the flexural behavior, also, we have come to the
conclusion that it is possible to use the grid shaped FRP
reinforcement in precast concrete curtain walls.
FRP Reinforcement
183
Application
The concrete curtain walls of a building near the
seashore are affected considerably by sea breeze, so their
reiforcement may corrode by salt attack. To prevent this
corrosion, precast concrete curtain walls with the grid
shaped FRP reinforcement were used at a building in Yokohama.
In Japan, application of advanced composite materials, such
as this reinforcement, to building structures requires the
approval of the Minister of Construction; the curtain walls of
this building is the first for which such approval was
obtained.
This is a 47m-high building, and has spandrel typed
curtain walls, 2100mm high, 3275mm wide and 150mm thick. So,
referring to the aforementioned results of the bending test,
H8 (@100mm sq. spacing, 4 sides are H13.) and G8 (D shape)
were used.
Fig 12 shows the placement work of this reinforcement at
a precast concrete factory. As this reinforcement, formed in
the dimensions of curtain wall, is light weight, the
placement work time became fairly shorter than that of steel
bars.
Fig 17 shows the installing work of precast concrete
curtain walls with this reinforcement at the construction
site. The installing work similar to that of steel bars were
conducted. Three years have passed since this building was
built, with no problems resulting thus far.
CONCLUSIONS
The above is an overview of research, development and
application of the grid shaped FRP reinforcement, a
description of tests for application to precast concrete
cutain walls, and an introduction of actual application to
precast concrete cutain walls.
In civil egineering structures, this reinforcement is
applied many structures such as tunnels, oil tanks, slopes,
bridges and port facilities. However, in building structures,
this reinforcement is applied only free access floor tiles
and curtain walls. Hereafter we attempt to apply this
reinforcement in box frame type reinforced concrete
structures.
Requirements for increased durability and function in
cocrete structures will continue to boost the need for
materials like the grid shaped FRP reinforcement in future.
The authors hope that this report will be the trigger that
increases such need.
184
Fujisaki, N akatsuji, and Sugita
ACKNOWLEDGEMENTS
Research, development, and application were done joinly
with Asahi Glass Matex Co., Ltd.; the authors deeply
appreciate the valuable instructions and attentive
cooperation.
Joint R&D was respectively done with the following
universities; the University of Tokyo, Tokyo Institute of
Technology, Science University of Tokyo, Musashi Institute of
Technology, Fukui University, Gunma University, Osaka
University, Doshisha University. We wish to express our
gratitude to many joint members for valuable advice.
REFERENCES
1. Fujisaki, T.; Sekijima, K.; Matsuzaki, Y.; and Okamura, H.,
"New Material for Reinforced Concrete in Place of Reinforced
Steel Bars," IABSE Symposium, Paris-Versailles, 1987.
2. Hayashi, K.; Sekine, K.; Sekijirna, K.; and Nakatsuji, T., "
Application of FRP Grid Reinforcement for Concrete and Soil,"
46th Annual Conference, Composite Institute, The Society of
the Plastics Industry, Inc., 1991.
3. Fujisaki, T; and Kobayashi, K., "Application of New Fiber
Reinforced Composite Material (NFM) to Shear Wall of Concrete
Structures," Proceedings of Ninth World Conference on
Earthquake Engineering, Tokyo, 1988.
4. Sekijima, K.; and Hiraga, H., "Fiber Reinforced Plastics
Grid Reinforcement for Concrete Structures," IABSE Symposium,
Brussels, 1990.
5. Ikeda, K.; Sekijima, K.; and Okamura, H., "New Materials
for Tunnel Supports,· IABSE 13th Congress, Helsinki, 1988.
6.Japanese Standards Association, "Japanese Industrial
Standard A 1304 -- Method of Fire Resistance Test for
Structural Parts of Buildings,· 1975.
TABLE 1- OVERVIEW OF RESEARCH AND DEVELOPMENT
I - ·---
Purpose
-~
l.=
____ _
I
Examination
1Clarification of FUndamental Properties of FRP Reinforcement 1
H Clarification
of Mechanical Properties
H Clarification
of Bond Mechanism
H Clarification
J-c
of L\lrabili ty
1
Lj Clarification of Heat Resistance
I Proof
H
of the Applicability to Concrete Structures
Application to Eartilquake Resisting Wall
Precast Wall Structure
Application to Wall Panel
H. Application to
H
-j Application to Slab Panel
Lj Application to Beam and Column
1-
fI
f-I-
J-c
1-
1-
Tensile Strength of Fibers
Tensile Strength
Canpressi ve Strength
Shear Strength
Strength of Cross Point
Lap Splice Joint
Anchorage to Concrete
Chemical Resistance under Constant Tensile I.::eforrnation
I.::eterioration of Tensile Strength under Spa Atlroshere
Creep Fracture
Tensile Strength under/after Heating/Cooling
Behavior under Reversed Cyclic Lateral Loading
Shear Resistance of Vertical Joint
Flexural Behavior
Fire Resistance
Anchorage to Beam
Time-I.::ependent I.::eflection
Flexural Behavior
Fire Resistance
Effectiveness of Prestressing
Effectiveness of Main and Shear Reinforcement
Confined Effect by Lateral Reinforcement
Fatigue Strength
Flexural Behavior after Sustained Loading
Flexural Behavior, Reinforced with Exposed FRP in Air
Effectiveness of Chemical Prestressing
-
~
:::0
""0
:::0
(!)
s·
0'
f6
'"1
3(!)
::I
.....
--
00
VI
186
Fujisaki, Nakatsuji, and Sugita
Fig. 1-FRP reinforcement for tunnel lining
Fig. 2-FRP reinforcement for shotcrete
FRP Reinforcement
Fig. 3-Installing work of fender plate using FRP reinforcement
Fig. 4-Prestressed concrete pedestrian bridge using
FRP reinforcement for tendons
187
188
Fujisaki, Nakatsuji, and Sugita
grip frame
'
:: '
~~,,
0
0
(V')
specimen
steel pipe
Fig. 5-Tension testing method
0 ~0----~----~2-----3L---~4----~s----~6
Strain (%)
Fig. 6-Load-strain relationships
FRP Reinforcement
~ gripframe
~
,-------, ;;
I
, temperature
,---~~~binet
I
I
I
r---_j''L---,
I
I
I
I
I
I
I
specimen ,
I
I
I
I
I
I
I
I
I
0
0
("")
I
I
I
mortar
Fig. 7-Heating tension testing method
§ 100
.-
0
·..-i
.j..l
•
•
•
0
r:::
QJ
.j..l
0
~ 75
..c:
t'r:::o
.bUl
0
QJ
0
0
0
0
50
r:::
0
·..-i
Ul
H8
@ 25
e-
under heating
• after cooling to room temperature
0
I
0
0
50
I
100
I
150
I
200
250
Temperature ("C)
Fig. 8-Heat resistance test results
189
190
Fujisaki, Nakatsuji, and Sugita
___25300T
1500
~------::::...:_:.
'I
•I
.,
I
I,
I
I
I'
I
[::!:H8 (bar type)
-~--;-
i--
---~-
I
I
I
I I
e
I
I I
I
I_ I
1
I
'--- ,..
hd·+
R'-----9
'
-
-,I
thermocouple
shape)
8 (II
I
!
I
I
I I
I I
I
I
0
0
\0
('f')
-+
I
n~'
~
steel fastener
I
I
., 300.,
900
I
I
I
I
I
•
1-
0
0
-r--
\0
!:;.
I
I
N
H8 (@lOOmm sq.
[\ deflection
spacing) -.impact
1.1
I
I
I
------'--
'
I
_,__ -
•
. . -
.
:
'
'
steel beam
/[±]
11±1
I
'
Fig. 9-Specimen for fire resistance test
FRP Reinforcement
~
~1000~----~------~~------~~----4-----_,
~
i
I
,heating temperature curve
~
1~
15o
hrng sute
t----#c.__~·-
I
I
500 t-r-1-----t-1--- reinforcement -----'---------l
(heating
surface
side)
I
,
.
concrete (75mm deep)
reinforcement
250 !f------A---""""""'..,.---t--r'-- ( back surface side)
I back ~urface
0~~~~=:)150
Time (min.)
Fig. 10-Temperature-time relationships
c 15 .------;-----,-------.------,
0
.!-)
~
after fire
resistance
~ 10~--------.--------.---~~.+--------1
without fire
reinforced concrete
~~~------­
resistance
test
~o
--
I
load
100
150
200
Deflection (mm)
Fig. 11-Load-deflection relationships
191
192
Fujisaki, Nakatsuji, and Sugita
Fig. 12-Placement of FRP reinforcement
Fig. 13-lnstalling work of curtain walls using FRP reinforcement
SP 138-13
Lateral Confinement of
Concrete Using
FRP Reinforcement
by A. Nanni, M.S. Norris,
and N .M. Bradford
Synopsis:
Lateral confinement of concrete members by means of spirally wrapping
fiber-reinforced-plastic (FRP) composites onto the concrete surface may
increase compressive strength and ultimate strain (pseudo-ductility). It may also
provide a mechanism for shear resistance, and inhibit longitudinal steel
reinforcement buckling. Lateral confinement of concrete members as a
strengthening/repair technology is expected to have an impact in the
rehabilitation/renovation of buildings and infrastructure. Structures that have
been damaged, or need to comply with new code requirements, or are subjected
to more severe usage are the primary targets. In this project, an experimental
and analytical study of concrete strengthened with FRP lateral confinement is
conducted using compression cylinders (300 and 600 mm in length) and 1/4
scale column-type specimens. The latter specimens have a circular cross section
and given longitudinal/transverse steel reinforcement characteristics.
Column-type specimens are subjected to cyclic flexure with and without axial
compression. When an aramid FRP tape is used as the lateral reinforcement, the
variables are tape area and spiral pitch. In the case of filament winding with
glass fiber, the thickness of the FRP shell is varied. The limited experimental
results obtained at this stage of the research program indicate that lateral
confinement significantly increases compressive strength and pseudo-ductility
under uniaxial compression.
Keywords: Bend tests; compression tests; concretes; fiber reinforced plastics:
fibers; glass fibers; lateral confinement; loading tests; reinforcing materials;
repairs; strengthening; wrapping
193
194
Nanni, Norris, and Bradford
ACI member Antonio Nanni is an associate professor in the Department of
Architectural Engineering at the Pennsylvania State University. His research
interests are in concrete materials and structures. He is Chairman of ACI
Committee 440, FRP Reinforcement.
ACI member Michael S. Norris is a graduate research assistant in the
Department of Architectural Engineering at the Pennsylvania State University.
He received his BArch from the University of Idaho in 1991. His research
interests are in new structural materials.
Nick M. Bradford is a graduate research assistant in the Department of
Architectural Engineering at the Pennsylvania State University where he
received his BAE in 1992. His research interests are in structural retrofitting
using new construction materials.
INTRODUCTION
Strengthening/Repair by Lateral Confinement
The aging of the U.S. infrastructure and buildings inventory, the better
understanding of natural phenomena (e.g., earthquakes) with the consequent
upgrading of building code requirements, and the need for improved structural
performance (due to more sophisticated uses or heavier loads) are among the
major reasons challenging the construction industry for the development of new
strengthening/repair technologies. Traditionally, the U.S. approach has given
preference to the demolition/reconstruction option. In a more immediate past,
due to both economical and cultural reasons, there has been a shift in attitude,
with greater attention being devoted to the strengthening/repair option. The
events following the 1989 Lorna Prieta earthquake have epitomized this
tendency. It is therefore necessary for researchers, practitioners, and contractors
to devote attention to this technical area and to develop economically sound and
safe technologies. When considering strengthening/repair type of work, the
following should be considered: a) predominance of labor and shut-down costs
as opposed to material costs; b) time and site constraints; c) durability; and d)
difficulty in selection, design and effectiveness evaluation. Research and
development efforts in new technologies should address the issues of automation
and formulation of design/prediction algorithms.
Lateral confinement of concrete members by means of spirally wrapping an
FRP tape or a resin-impregnated strand (filament winding) onto the concrete
surface is a potential strengthening/repair technique. Lateral confinement may
increase compressive strength and ultimate strain (pseudo-ductility), provide a
mechanism for shear resistance, and inhibit longitudinal steel reinforcement
buckling. For lateral confinement, the advantages of FRP composites over steel
reinforcement or jacketing are: larger contact area and low profile, no corrosion,
flexibility, possibility of pretensioning (even with prismatic cross sections), and
FRP Reinforcement
195
ease of automatic installation. In this project, an experimental and analytical
study of this strengthening/repair technology is conducted using conventional
compression cylinders, double-length compression cylinders, and l/4 scale
column-type reinforced concrete (RC) specimens with different
longitudinal/transverse steel reinforcement characteristics. The latter specimens
are subjected to cyclic flexure with and without axial compression. The lateral
FRP reinforcement consists of a continuous flattened tube made of braided
aramid fiber in one case, and, in the other, of a continuous glass strand placed by
a filament winding machine. The effects of different areas and spiral pitches for
the tape, and thickness of the FRP shell for filament winding are investigated.
Previous Experiences
Japan-- In the mid eighties, Ohbayashi Co. and Mitsubishi Kasei Co.
developed the concept of strengthening and retrofitting existing RC structures
using carbon fiber strands and mats (CFRP). Three types of structures were
targeted: building columns (Katsumata et al. 1987; Katsumata et al. 1988),
bridge columns (Kobatake et al. 1990), and chimneys (Katsumata and Kimura
1990). According to their method, CFRP strands impregnated with resins are
spirally wound onto the surface of an existing RC member. In the case of bridge
columns and chimneys, CFRP mats may be adhered first to the concrete surface
in the longitudinal direction of the structural member so that flexural strength is
also enhanced. The primary function of the spirally wound strand is to improve
shear capacity and ductility of the member. Experimental work to evaluate the
potential of this method and the development of the first winding machine has
been undertaken at the Technical Research Institute of Ohbayashi Co. This
research shows that the benefits of the strengthening/repair method are
remarkable. Improvements in strength of 1.5 times and maximum deformation
ability up to four times greater than that of the original member were recorded
using zero-pretension winding (Katsumata et al. 1987). Both circular and
prismatic cross section elements without conventional steel hoop or spiral
reinforcement were investigated. Specimens were not subjected to axial
compression, only shear and bending moment were applied. Tests have shown
that the low strain capacity of carbon fiber and its brittleness (even when epoxy
impregnated) are a limiting factor. For prismatic elements, corners needed to be
beveled prior to fiber winding (Kobatake et al., 1989).
U.S.A. -- In the area of structural rehabilitation, the firm of Fyfe Associates is
proposing to retrofit bridge columns with glass/epoxy FRP jackets (Fyfe Jan.
1992). The initial work was supported by the California Department of
Transportation with the objective of developing a method to enhance flexural
and shear performance in the critical regions of bridge columns (Priestley et al.
Dec. 1991 ). Experimental results on columns subjected to combined axial
compression and lateral loads, have demonstrated that jackets made of glass
FRP are effective in inhibiting shear failure and forcing the development of
ductile flexural modes of inelastic deformation.
196
Nanni, Norris, and Bradford
EXPERIMENTAL PROGRAM
Materials
The following is a description and a characterization of the materials being
used in this research project.
Concrete and Steel -- All specimens were made with ready-mix concrete
having the characteristics shown in Table 1. The longitudinal steel
reinforcement for column-type specimens was deformed wire, size D8 (8.1 mm
in diameter), and 516 MPa yield point. Welded wire mesh (2 mm wire diameter
and 25 by 25 mm wire spacing) was used in lieu of transverse ties for some
column-type specimens.
Braided Aramid FRP Tapes -- Tapes consisting of a collapsed tubular section
(made of braided aramid fiber and impregnated with epoxy resin) were used for
the first type of lateral reinforcement. The aramid fiber was Kevlar 49 (grade
6000 denier). The epoxy resin was Dow Chemical DER 330 combined with the
hardener Ancamide 506 by Pacific Anchor Chemical (mixing ratio 100 to 55
parts). Three tape sizes were used and were identified as: K24, K48, and K64.
The corresponding fiber-only areas were: 10.8, 21.7 and 28.9 mm2. The
nominal cross section areas after impregnation and flattening were: 19, 39, and
52 mm2. The nominal transverse side of each tape after flattening was: 10, 13,
and 18 mm.
The amount of pretension in the tape during wrapping has a significant effect
on the mechanical properties of the braided tape itself. This is because braided
strands run at an angle with respect to the longitudinal axis of the tape, and the
presence of pretension during resin hardening increases the overall stiffness of
the tape. In order to evaluate the effect of pretensioning, an epoxy-impregnated
tape (size K64) was wound around a rectangular hollow steel pipe. The pipe
was longitudinally split into two halves, so that jacks inside the pipe could be
activated to apply the desired level of tension to the tape after it was wound.
The jack pressure was maintained until the tape had completely cured. After
resin hardening, two types of tape samples were obtained by cutting the
rectangular-shaped spiral: straight samples (A type) and comer samples (B type,
corner diameter= 62 mm). For the straight samples (A type), conventional
uniaxial tensile testing was conducted after fabricating molded anchors at the
ends of the specimen. For the corner samples (B type), one leg of the specimen
was fully embedded into concrete (to provide anchorage) while the other leg was
pulled to failure by a jack. The area bonded in concrete extended to past the end
of the corner so that no bending was applied to the tape during the test. The
results of 30 tests for the straight section and 15 tests for the corner section are
given in Table 2. The specimens were subdivided into three groups depending
on the tensioning force in the tape during resin hardening (i.e., 0.0, 0.2 and 1.0
kN). From an analysis of the results given in the table, it appears that
pretensioning affects the ultimate capacity and stiffness of the tape. However, a
low level of pretension (0.2 kN) provides adequate efficiency. Considering that
the average strength of the mechanically manufactured rod of the same geometry
is 68.87 kN, the efficiency of the tape pretensioned at 0.2 kN is 87 percent.
Corner sections, as expected, are relatively weaker. It was concluded that for
FRP Reinforcement
197
lateral reinforcement, a pretensioning force of 0.2 kN was easily attainable and,
at the same time, was sufficient to ensure acceptable strength and stiffness of the
tape.
The nominal load capacity of the three tape sizes pretensioned at 0.2 kN
(straight sample) was 21, 42, and 56 kN.
Glass Filament Winding -- Twelve concrete samples 600-mm in length were
shipped to a fabricator for the application of lateral confinement by filament
winding. In this case, the FRP shell was made of E-Glass strands impregnated
with a polyester resin. The strand were helically wound around the specimen
(mounted on a rotating mandrel). The winding angle was 65 degrees. The
thickness of the FRP shell was 0.6, 1.2, 2.4 and 3.6 mm, respectively. Two
specimen per each shell type were fabricated using a bonding agent at the
concrete-FRP interface. For the 14-mm shell, two additional cases were
considered: alternative bonding agent, and no bonding. Prior to compression
testing, the 600-mm cylinders were saw-cut in halves.
Specimen Fabrication
Specimen Type -- Three groups of concrete specimens, all with same
diameter (150-mm), were used: Group A -conventional compression cylinders
(300 mm long); Group B - double-length compression cylinders (600 mm long);
and Group C- column-type specimens (1525 mm long). Groups A and B were
used exclusively for uniaxial compression tests. The longer length of Group B
was needed for two reasons: first, to facilitate filament winding (a 25-mm
diameter pipe was inserted along the axis of the cylinder to allow placement in
the rotating mandrel of the winding machine); and, second, to allow for the
placement of longitudinal steel reinforcement as for the specimens of Group C
(see Figure I). Specimens of Group C simulate a 1/4 scale circular column and
were intended for cyclic flexural test with and without axial compression. All
specimens in this group were reinforced with six D8-size longitudinal rods (see
Figure 1). Some specimens also had transverse reinforcement in the form of a
welded wire mesh cage. Specimens were cast with a sleeve along the central
axis to receive a postensioning rod for the application of the axial compression.
The sleeve diameter was 19 or 32 mm, depending on the rod size.
Tape Wrapping-- In the first wrapping technique, the braided aramid FRP
tape was applied to the specimen in the university laboratory. The aramid tape
was dipped into the epoxy resin bath for approximately one minute and rumpled
by hand to remove air trapped in the interstices of the tape. The impregnated
tape was passed through a die made of silicone rubber to remove excess resin,
and wound around a steel pipe. This intermediate step allowed the operator to
maintain the desired pretension of 0.2 kN in the tape during winding onto the
concrete cylinder. Prewinding also allowed a better control on the tape pitch.
Three pitches of 0, 25, and 50 mm were used for the three tape sizes. After
wrapping, the FRP was allowed to cure for at least seven days at room
temperature.
Filament Winding-- Filament winding was executed at the industrial plant of
an FRP manufacturer specializing in pressure pipe and tank production.
198
Nanni, Norris, and Bradford
Experimental Results
Uniaxial Compression-- As of now, 20 specimens of Group A have been
tested in uniaxial compression. The results of these tests are summarized in the
stress-strain diagrams of Figures 2, 3, and 4. Each figure corresponds to a
specific tape size and each curve is the average response of two specimens (the
curve relative to plain concrete is the average of six specimens). Figure 2 shows
that the presence of the K24 tape has no effect when wrapped at a 50-mm pitch
(K24-2). The photograph in Figure 5 shows that the typical failure mode for one
specimen of sub-group K24-2 is still of the shear type. At 25 mm pitch (K24-1 ),
the pseudo-ductility of the specimen is increased without increasing ultimate
strength. At 0 mm pitch (K24-0), the maximum capacity of the machine (1.3
MN) was reached without taking the specimen to failure. The increase in both
strength and stiffness occurring in this case is evident. The photograph in Figure
6 shows the condition of one of the two K24-0 specimens after the conclusion of
the test. No sign of FRP deterioration is visible.
The diagram of Figure 3, relative to the K48 tape, shows that at 50 mm pitch
(K48-2) there is no significant contribution from the lateral confinement. For
the case of 25 mm pitch (K48-l ), again the maximum capacity of the testing
machine was reached prior to failure. A photograph of one specimen of this
series is shown in Figure 7. Extensive cracking and bulging of the fractured
concrete can be seen between the tape spirals. The K48-0 specimens are to be
tested with a different machine having a capacity of 4.4 MN.
The case of K64 tape is presented in Figure 4. This time the 50 mm pitch
(K64-2) shows a considerable increase in pseudo-ductility with no strength
improvement (a photograph of a failed specimen is given in Figure 8). The
specimens with 25 mm pitch (K64-I) could not be failed. The K64-0 specimens
were not tested. Considering the stress-strain curves of all the sub-t:,rroups tested,
the slope of the second branch of the curve, when it exists, is clearly
proportional to the area of lateral reinforcement as suggested by other
researchers (Katsumata eta!. 1988)
Specimens which were not failed will be re-tested with the higher capacity
machine together with those of sub-groups K48-0, k64-0, and those of Group B.
Flexure-- As of now, four specimens have been tested in flexure without
axial compression. All four specimens were wrapped with aramid FRP tape at
25-mm pitch. The first sample used K24 tape (C04). Two samples, one with
steel shear reinforcement and one without, used K48 tape (C07 and ClO). The
fourth sample used K64 tape (COl). The testing set-up is shown in the sketch of
Figure 9. The specimen was simply supported over a span of 1372 mm and
loaded at mid-span. Full load cycles (push and pull) were applied below the
cracking strength (half-cycles 1 to 4), in the elastic-cracked region (half-cycles 5
to 10), and in the plastic-cracked region (half-cycles 11 to 16). The following
load cycles (half-cycles 17 and higher) attempted to be at peak-load and
post-peak load levels. The analytical interaction diagram for the RC cross
section without considering the lateral confinement is presented in Figure 10.
The diagram shows that the nominal moment capacity at zero axial force is 8.81
kNm, which corresponds to a load of 28.9 kN.
FRP Reinforcement
199
The experimental load-deflection diagrams for the four specimens as tested
are shown in Figures I 1, 12, 13, and 14, following the order assigned by the
specimen code (i.e., CO I, C04, C07, and C 10). Since only a selection of
hysteresis loops is reported in the diagrams, the number shown in the proximity
of each curve indicates the half-cycle number (odd value for the "push" direction
and even value for the "pull" direction). The peak loads were +34.4 and -34.3
kN for specimen C-Ol, +38.7 and -31.6 for specimen C-04, +34.8 and -30.1 for
specimen C-07, and +31.2 and -31.1 for specimen C-10. The experimental
strength of all specimens is higher than the predicted value. Apart from
specimen C-04 that failed immediately after peaking, the hysteresis loops are
very similar among specimens. This indicates that the contribution of the lateral
confinement is not significant. This was expected because failure is
tension-controlled. The crack pattern was very similar in all specimens. All
cracks, including those in proximity of the supports, were vertical (flexural
type). After starting from the tensile zone of the specimen during one
half-cycle, almost every crack propagated through the entire cross section at the
inversion of the load.
DISCUSSION
Effect of Confinement
In uncracked and microcracked concrete under compression, axial strain due
to loading can be converted to transverse strain using the Poisson's ratio. The
effect of confinement on concrete is directly related to its ability to restrain this
transverse strain. As the transverse strain increases, the amount of pressure
exerted by the confinement increases. The resulting biaxial state of stress is
beneficial to the cracking capacity and strength of concrete (Kupfer eta!. 1969,
Newman and Newman 1972). Since the confining stress is dependant upon the
transverse strain, the confinement is said to be passive in nature (Ahmad and
Shah 1982). The benefits of the biaxial state of stress can be observed, for
example, in the stress-strain curve relative to sub-group k24-0 (Figure 2), as it
departs from the plain concrete curve in the range 25 to 45 MPa.
Independent of the amount of lateral confinement, all specimens so far tested
show a bend-over point at approximately the strain level which corresponds to
the maximum capacity of the plain concrete. After this point, concrete
undergoes extensive cracking and the relationship between axial and transverse
strain cannot be accurately determined using the Poisson's ratio. The
confinement carries an increasingly greater portion of the load (as an analogy,
the lateral confinement has become a pressure pipe holding liquified material).
During this stage, the specimen shows a flat or a rising curve depending on the
amount of lateral confinement. The slope of the curve also depends on the
stress-strain curve of the confining material. If steel tubing is used and reaches
its yielding point, then a gradual loss of load capacity in the post-cracking zone
is experienced (Chai eta!. I 991, Mander eta!. 1988). With a sufficient amount
of FRP as the confining material (linear-elastic behavior till failure), the
specimen experiences a slow increase in load capacity which may continue until
the FRP wrapping has achieved its ultimate strength. Experimental results
obtained by Bazant eta!. (1986) on concrete cylinders encased in a pressure
200
Nanni, Norris, and Bradford
vessel with 150-mm walls show that specimens with this confinement could
carry loads in excess of 2,000 MPa.
With reference to the interaction diagram of Figure I 0, it is hoped to
demonstrate that lateral confinement alters the behavior of a column-type
specimen, when concrete crushing controls the failure mode. The next set of
experiments, conducted under cyclic flexure with an axial compression of 200
kN, should demonstrate this point.
Analytical Model
Prior to the development of design procedures for concrete members with
lateral FRP confinement, an analytical model must be established to predict the
performance of the confined concrete. The model must account for different
material types (concrete and FRP), volume and distribution of confinement, and
wrapping techniques. The model could be used in both the areas of uniaxial
compression and compression-controlled flexure. Concrete failure has been
modeled following the Mohr-Coulomb theory (failure occurs along the critical
shear plane of the specimen). In the case of laterally confined concrete, it
appears possible to consider a combination of this approach (to predict the first
part of the curve) and the pressure-pipe analogy (to predict the second part of the
curve). Work is being conducted in this area and will be reported at a later date.
CONCLUSIONS
Only preliminary conclusions can be drawn at this stage of the research
project. Lateral confinement in the form of braided, "field-produced" aramid
FRP tape has been used. The loss of efficiency with respect to the
"factory-produced" tape form is relatively modest, provided that some
pretension is present in the tape at the time of winding. There is experimental
evidence to demonstrate that the compressive strength and pseudo-ductility of
concrete is increased by the presence of lateral FRP confinement. Experimental
work will continue in order to establish the upper values of this enhancement,
and to evaluate a second type of FRP confinement produced according to the
method of filament winding. On the analytical side, a model based on a
combination of the Mohr-Coulomb theory and the pressure pipe-analogy is
being developed. Testing on column-type specimen will continue with the more
interesting case of specimen subjected to axial compression and flexure. For
specimens subjected to flexure and with tension-controlled failure, the effect of
lateral confinement is not significant.
ACKNOWLEDGMENTS
The support of the National Science Foundation under Grant No.
MSS-91 08598 is gratefully acknowledged. Mitsui Construction Co., Tokyo,
Japan, has provided materials and financial support. Dow Chemical and Pacific
Anchor Chemical have provided materials.
FRP Reinforcement
201
REFERENCES
Ahmad, S.H and S.P. Shah (1982). "Stress-Strain Curves of Concrete Confined
by Spiral Reinforcement," ACI Journal, Vol.79, No.6, pp 484-490.
Bazant, Z.P., Bishop, F.C. and Chang, T.P. (1986). "Confined Compression
Tests of Cement Paste and Concrete up to 300 ksi," ACI Journal, Vol.83, No.4,
pp. 553-560.
Chai, Y.H, Priestley, M.J.N. and Seible, F. (1991). "Seismic Retrofit of Circular
Bridge Columns For Enhanced Flexural Performance," ACI Structural Journal,
Vol. 88, No.5, pp. 572-584.
Fyfe, E. (Jan. 1992). "Strengthening Bridge Columns with Composites,"
Transportation Research Board, 71 Annual Meeting, Washington, D.C.
Katsumata H., Kimura K. (May 1990). "Applications of Retrofit Method with
Carbon Fiber for Existing Reinforced Concrete Structures," The 22nd Joint
UJNR Panel Meeting, U.S.- Japan Workshop, Gaithersburg, MD, pp. 1-28.
Katsumata H., Kobatake Y., and Takeda, T. (1987). "A Study on The
Strengthening with Carbon Fiber for Earthquake-Resistance Capacity of
Existing Reinforced Concrete Columns," Proceedings of the Seminar on Repair
and Retrofit of Structures, Workshop on Repair and Retrofit of Existing
Structures, U.S.-Japan Panel on Wind an Seismic Effects, UJNR, pp. 18-1 to
18-23.
Katsumata, H., Kobatake, Y., and Takeda, T. (1988). "A Study on Strengthening
with Carbon Fiber for Earthquake-Resistant Capacity of Existing Reinforced
Concrete Columns," Proceedings of 9WCEE, VII, pp. 517-522.
Kobatake, Y., Katsumata. H., and Okajima, T. (1990). "Retrofitting with Carbon
Fiber for Seismic Capacity of Existing Reinforced Concrete Bridge Columns,"
Proceedings of the 45th Annual Conference of the Japan Society of (in
Japanese)
Kupfer, H., Hilsdorf, H.K. and Rusch, H. (1969). "Behavior of Concrete Under
Biaxial Stresses," ACI Journal, Vol. 66, pp. 656-66.
Mander, J.B, Priestley, M.J.N. and Park, R. (1988). "Theoretical Stress-Strain
Model For Confined Concrete," J. Struct. Engrg., ASCE, Vol. 114, No.8, pp.
1805-1826.
Newman, K. and Newman, J .B. ( 1972). "Failure Theories and Design Criteria
for Plain Concrete," Part 2, Solid Mechanics and Engineering Design.
Wiley-Interscience, New York, pp. 83/1-83/33.
Priestley, M.J.N., Fyfe, E., and Seible F. (Dec. 1991). "Retrofitting Bridge
Columns with Fiberglass/Epoxy Composite Jackets," Proceedings, Caltrans
Research Seminar, Sacramento, CA, 8 pp.
202
Nanni, Norris, and Bradford
TABLE 1 -MATRIX PROPORTIONS AND PROPERTIES
Portland Cement Type I
(kg/m3)
Fly-ash Class F
(kg/m3)
Water
(kg/m3)
Coarse Aggregate (9 mm MAS)* (kg/m3)
Fine Aggregate*
(kg/m3)
Water Reducer
(1/100 kg of cement)
Slump
Unit Weight
Air Content
I
I
*
0. 8
(mm)
100
(kg/m3) 2268
(%)
28-day Compressive Strength
Standard Deviation (6 samples)
Note:
327
53
168
950
823
= Saturated
(MPa)
(MPa)
1. 3
35.6
2. 5
Surface Dry
TABLE 2 -RESULTS OF TENSILE TESTS FOR BRAIDED ARAMID FRP
TAPE (K64)
I
I
I
I
I
I
Specimen
Type
I
lA
0.0
I
I
I
type!straight
avg.
(C.V.)
typelcorner
avg.
(C.V.)
I
IB
Pu
(kN)
I
I 54.3
I (12. 8>
I
I 38.5
I (4. 8)
Pretension Force (kN)
0.2
Elastic
Modulus
(GPa)
Pu
(kN)
60.4
(24. 6)
59.8
(7. 7)
N/A
46.2
(7. 5)
1.0
Elastic
Modulus
(GPa)
(kN)
Elastic
Modulus
(GPa)
62.2
(10. 0)
61.9
(9. 9)
60.6
(6.8)
48.3
(3. 6)
N/A
N/A
Pu
FRP Reinforcement
--'
'
,-----·
203
---
·---
Longitudinal~·- •• ,
DB
Reinforcement
(Group C)
''
\
\
'\
'
:Steel Pipe
.
C)~--~------r,',
',\
19 or 32mm
~
8
151.7mm ·- · - - - --------- --f----- ----~--- -~~ -~:~e_r_ - - · - - \
........ ___ .:.::~---- -/
::' (Groups B &
'
'
I
I
\
I
\
\
I
I
J
I
I
Welded Wire M e s h / '
Transverse Reinforcement
''
...... __
_...... -
Fig. 1-Typical cross-section for specimens of Groups Band C
80
70
60
~50
e"" 40
rJl
~
30
20
10
0
0
--- Plain Concrete
0.005
STRAIN (mm!mm)
-ir-
K24- 2
-+- K24-1
0.01
O.ot5
...... K24-0
Fig. 2-Uniaxial compression stress-strain diagram (Group A, K24 tape)
204
Nanni, Norris, and Bradford
~r-------------,-------------r-----------~
Not Failed
~50 r-------~~----r---------------~------~t---~
6
~
: t--------#---+-----+----+---------l
0.005
0.01
0.015
SIRAIN (mm/mm)
-----Plain Concrete
-il-
K48- 2
--+-- K48-1
Fig. 3--Uniaxial compression stress-strain diagram (Group A, K48 tape)
~
Not Faile
70
60
';' 50
=-
1:
20
10
0.005
0.01
0.015
SIRAIN(mm)
----- Plain Concrete
-il-
K64- 2
--+-- K64-1
Fig. 4-Uniaxial compression stress-strain diagram (Group A, K64 tape)
FRP Reinforcement
205
Fig. 5-Typical failure of specimen type K24-2
Fig. 6-Specimen type K24-0 after loading to 1.3 MN (machine capacity)
206
Nanni, Norris, and Bradford
Fig. 7-Specimen type K48-1 after loading to 1.3 MN (machine capacity)
Fig. 8-Typical failure of specimen type K64-2
FRP Reinforcement
207
Fig. 9-Flexural test set-up for Group C (no axial force)
,c- .
.... ...
4·
...
'
~,
\\
\
I
'Z'Z-
0
---
,., .....
/
-*
,
-----
,., .....
---
/
..........
,_8,.,
.....
.....
,/'
t?J.r..
Mt>MEoH
Fig. 10-Interaction diagram for reinforced concrete section (Group C)
208
Nanni, Norris, and Bradford
40
30
20
z
10
~
Ci
<C
Pull
Pusb
0
g
-10
-20
-30
-40
-40
-30
-20
-10
0
10
DEFLECTION (mm)
20
40
30
Fig. 11-Load-deflection hysteresis loops for specimen COl (K64 at 25 mm)
40
21
30
20
z
~
~
10
Pull
Pusb
0
g
-10
-20
-30
22
-40
-40
-30
-20
-10
0
10
DEFLECTION (mm)
20
30
40
Fig. 12-Load-deflection hysteresis loops for specimen C04 (K24 at 25 mm)
FRP Reinforcement
209
40
17
30
20
10
z
~
Pusb
Pull
0
~
3
-10
-20
-30
18
-40
-40
-30
-20
-10
0
10
DEFLECTION (mm)
40
30
20
Fig. 13-Load-deflection hysteresis loops for specimen C07 (K48 at 25 mm)
40
17
19
30
20
10
z
~
Pusb
Pull
s
0
0
-10
-20
-30
18
-40
-40
-30
-20
-10
0
10
DEFLECTION (mm)
20
30
40
Fig. 14-Load-deflection hysteresis loops for specimen ClO (K48 at 25 mm
and without steel shear reinforcement)
SP 138-14
Application of ThreeDimensional Fabric Reinforced
Concrete to Building Panels
by H. Nakagawa, M. Kobayashi,
T. Suenaga, T. Ouchi,
S. Watanabe, and K. Satoyama
Synopsis: The three-dimensional fabric studied as a reinforcement for concrete is a stereo-fabric made of rovings of fibers, woven into three directions
and impregnated with epoxy resin. Fiber material, number of filaments, and
distance between rovings can be varied easily. Efficient production is also
possible since three-dimensional weaving, resin impregnation, and hardening
can all be done by an automatic weaving machine.
The authors investigated the flexural and fire resistance behaviors of
three-dimensional fabric reinforced concrete (30-FRC) toward applying the
material to building panels. The fibers studied were carbon fiber and aramid
fiber, and the matrix was vinylon short-fiher reinforced concrete. The results
demonstrate that 30-FRC panels have sufficient flexural strength and rigidity to
withstand design wind loads, and the fire resistance of 60 minutes was achieved.
30-FRC panels have been used for curtain walls, parapets, partition
walls, louvers, etc., and installations amount to 7,000 m2 •
Keywords: Carbon; curtain walls; durability; epoxy resins; fabric; fiber
reinforced concretes; fibers; fire resistance; flexural tests; panels; specific
heat; wind resistance
211
----
212
Nakagawa et al
H. Nakagawa is a Research Engineer for Kajima Technical Research Institute.
His research interests arc in the behavior and the modeling of fiber reinforced
concrete members.
M. Kobayashi is a Assistant Manager for Kajima Technical Research Institute.
He has been actively involved with the composite structures and the structural
behaviors for over 25 years.
T. Suenaga is a Senior Research Engineer for Kajima Technical Research Institute. His research interests are in the behavior and the durability of fiber reinforced concrete members.
T. Ouchi is a Senior Research Engineer for Kajima Technical Research Institute. His research interests are in the fire safety engineering design on buildings. especially the structural fire safety design.
S. Watanabe is a Senior Research Engineer for Kajima Technical Research
Institute. His research interests arc in the structural behavior of fiber reinforced
concrete members.
K. Satoyama is a Research Engineer for Kajima Technical Research Institute.
His research interests are in the behavior and the fire resistance of fiber reinforced concrete members.
INTRODUCTION
For many years, the deterioration of reinforced concrete has been a major
concern of the construction industry in Japan. One cause of deterioration is the
corrosion of steel bars, a result of salt damage. It is now an issue of the utmost
import to establish countermeasures against this problem for buildings in waterfront areas.
Laborsavings in construction work have also become a subject of interest,
due to Japan's labor shortage and desire to reduce working hours and construction costs.
Pressure arising from these factors has accelerated the increase in
size, decrease in weight, and prefabrication of construction members.
Under
these circumstances, strong, light and durable FRP bars made of high-performance fiber nwings in which resin is impregnated and hardened arc regarded as a
prospective replacement material for steel reinforcing bars (1).
Three-dimensional fabric (3-D fabric) is made by weaving fiber nwings
in three directions. For 3D-FRC, 3-D fabric is impregnated with epoxy resin
and filled with cement matrix (see Fig. 1). By changing the fiber material, the
number of filaments, and the distance between ravings, fiber reinforcement
appropriate to various stress conditions can be produced. Moreover, the
mechanical bond strength between fibers and matrix is high because of the
checkered pattern structure of the ravings (2).
FRP Reinforcement
213
This report describes the material properties of 3D-FRC members, their
flexural behavior and thermal durability, and the structural and fire resistance
behaviors of 3D-FRC planks to be used for parapets and curtain walls.
PROPERTIES OF COMPOSITE
Four-point flexural tests were conducted on vinylon short-fiber reinforced concrete (VFRC) planks with and without 3-D fabric reinforcement
using carbon fiber made from polyacrylonitrile (PAN-type carbon fiber) or
aramid fiber, and the effects of the presence of 3-D fabrics and short vinylon
fibers on flexural characteristics were examined. The thermal durability of
flexural strength was also examined by testing specimens soaked in 80°C water
or autoclaved at 180°C. The 3-D fabrics used were made of PAN-type carbon
fiber, aramid fiber, vinylon fiber, or AR-glass fiber. The physical properties of
the fibers and the tensile stress-strain curves of FRP rods composing 3-D fabric
In Fig. 2, the tensile stress of FRP rod is
are shown in Table 1 and Fig. 2.
calculated on the basis of the net area of fibers (no including the area of epoxy
resin). The tensile stress-strain relationships arc approximately linear up to the
ultimate tensile strength. The tensile strengths and modulus of elasticity of
FRP rods are 70% to 80% of those of filaments. It is considered that the 20%
to 30% of filaments arc not aligned under equal tension.
Flexural Behavior
The specifications of the 3-D fabrics and the mix proportions of VFRC
arc listed in Table 2 and Table 3.
PAN-type carbon fiber and aramid fiber
were used for the X and Y axes of 3-D fabrics. The Z axis fibers of all specimens were PAN-type carbon fibers in consideration for easy manufacturing.
Low shrinkage cement was used to prevent cracking and deformation of the
planks due to drying shrinkage, and micro balloons, fine (under 100 ,urn of particle size) hollow glass balls made from pozzolana, were used to reduce weight.
Specimens were steam cured at a peak temperature of 40°C for 12 hours after
placing. Loading tests were performed using four-point loading with equal
spacing, with a cross-head speed of 1.0 mm/min (sec Fig. 3).
Load-deflection curves for CF48 and AF16 specimens are shown in
Fig. 4. They reveal the reinforcing effects of short vinyl on fibers mixed in the
matrix. The more the vinylon fiber content is raised, the larger is the increase
of modulus of rupture (MOR) and flexural rigidity in the second region. In the
same way, the vinylon fiber content is raised, the smaller is the drop in strength
after crack occurrence. The curves for specimens without short vinylon fibers
have an apparent yielding range after the first crack point. This is because the
main stress on matrix of extreme tension fiber is released and shifts to the tension reinforcements of 3-D fabric to make the fibers elongate.
In the case of
specimens with short vinylon fibers, on the other hand, the apparent yielding
range did not appear because the matrix could withstand the stress even after
crack occurrence by using the short fibers as a bridge. Fig. 5 shows the loaddeflection curves for specimens reinforced with different 3-D fabrics and a
1.0% short vinylon fiber content.
The flexural characteristics of 3D-FRC
214
Nakagawa et al
depend on the properties of the matrix up to the initial crack occurrence, and on
the material and fiber content used in the 3-D fabric in the range beyond.
Influential factors are: rupture elongation of fiber for maximum deflection; and
tensile strength, modulus of elasticity and fiber content for MOR and flexural
rigidity.
Thermal Durability
The specifications of 3-D fabrics and the mix proportions of the mortar
matrix of specimens tested for thermal durability are listed in Table 4 and
Table 5.
PAN-type carbon fiber, aramid fiber, vinylon fiber, and AR-glass
fiber were used for the X axis of 3-D fabrics.
The matrix was river sand
mortar using high-early-strength Portland cement. Specimens were steam
cured at a peak temperature of 60°C for 5 hours after placing. Some specimens
were soaked in 80°C water for the designated period (up to 30 days), while
others were cured hy autoclaving at 180°C from 1 to 5 times (5 hours per session). All specimens were then dried indoors for 2 to 3 days before being
subjected to flexural testing.
Loading tests were performed under four-point
loading with equal spacing and a cross-head speed of 1.0 mm/min (sec Fig. 6).
Fig. 7 shows the relation between MOR and days soaked in 80°C water,
and Fig. 8 the relation between MOR and number of autoclave sessions. While
the MOR of specimens with AR-glass fiber decreased slightly with longer
periods of soaking in hot water, the MORs of other specimens arc decreased
little up to 30 days. Also, the rate at which the MOR of specimens with ARglass fiber changed was extremely low compared with that of conventional GRC
made with short AR-glass fibers (3). The MORs of specimens autoclavcd at
180°C all declined after the first session. This is assumed to he a result of the
deterioration of epoxy/matrix bond strength. After subsequent sessions, MORs
of carbon fiber and aramid fiber specimens were relatively unchanged, those of
AR-glass fiber specimens fell gradually, and vinylon fiber specimens declined
radically as thermal deterioration destroyed the reinforcing capacity of these
fibers.
APPLICATION I : PARAPET PANEL
The exterior of the Suidohashi Building of Tokyo Dental College (sec
Fig. 9) was composed of steel-stud-framed ( 4) granite-faced curtain walls.
The panels of curtain walls were made of concrete reinforced with short carbon
fibers.
However, the steel-stud-framed panels were not suited to the parapet
because of the corrosion of the steel stud frames hy exposed to weathering.
This is the reason the durable 3D-FRC panels were used.
Fig. 10 shows the
geometry of a standard panel. The nwings used for 3-D fabrics were aramid
fibers (24KF) for the X and Yaxes and carbon fibers (12KF) for the Z axis. in
consideration of possible trouble concerning radio-wave interference.
YFRC
(Yr=l.25%) was used for the matrix. The 3-D fabrics (panel type) were arranged in a row with the maximum width of 600 mm in the direction of X axis
and hound to the 3-D fabrics (he am type) and the joint sheets (2- D fabrics).
The weight of a standard panel was about one ton, and the panels covered an
FRP Reinforcement
215
area of 600m 2 •
Wind Resistance --Test
---Wind resistance tests were conducted on specimens with and without
granite facing. The size of the specimens was the same as that of the largest
panels used on the Suidobashi Building, 3573 mm in width and 2133 mm in
height. All other specifications were the same as those of the standard panel.
The design wind loads for this case were large at 5.0 kN/m 2 on the negative
pressure side and 3.3 kN/m 2 on the positive pressure side. The specimens were
pressurized according to the pressurization procedure shown in Fig. 11, i.e.,
repeated twice from the negative design load to the positive one, pulsated in a 10
second cycle for 5 minutes in the negative side, repeated once up to 1.5 times as
large as the design load, and lastly pressurized to the maximum load of testing
machine (8.8 kN/m 2).
Test results on the displacement and strain of specimens are summarized
in Table 6.
The wind load-displacement and wind load-strain relationships,
both as measured at the center of upper rib where maximum displacement occurred, are shown in Fig. 12. Displacement and strain of the specimens showed
elastic relationships with variation of wind load, indicating that 3D-FRC panels
Other notable results are as follows:
display stable deformation behavior.
displacement at the center of upper rib was approximately twice that at the lower
rib because of differences in shape and arrangement of fasteners(pin support in
the upper, fixed in the lower); displacement of specimens with granite facing
was 2 to 3 times larger than that of specimens without granite facing, because
the granite depends greatly on the flexural rigidity of 3D-FRC panel.
APPLICATION II : CURTAIN WALL
The 3D-FRC was first applied on a large scale (1500 m2) to the 23-story
"Sea Fort Square" building now under construction in the waterfront area of
Shinagawa, Tokyo (see Fig. 13). A unit of this tile-finished curtain wall is
composed of 3D-FRC panel and steel stud frame, both being connected with
bolts at intervals of 600 mm. The original design specified aluminum panels
for this curtain wall. However, due to concerns over possible salt-induced
corrosion of aluminum, and because a significantly heavier wall material such as
conventional precast concrete could not be used to be against the load-bearing
capacity of the steel skeleton, light and durable 3D-FRC was adopted.
Fig. 14
shows the geometry of the panels. The rovings used for the 3-D fabrics were
PAN-type carbon fibers (24KF in X and Y axes and 12KF in Z axis), and joint
rods made of 120KF carbon fibers were used to bind two 3-D fabrics in case of
beam panels. Flexural tests and fire resistance tests were performed to confirm
the material's safety and to obtain the approval of the relevant authorities.
216
Nakagawa et al
Flexural Test
Flexural tests were conducted on tile-finished column panels.
The
column panels were fastened to the steel structure by rocking lower fasteners and
sliding upper fasteners, while the beam panels were supported by the column
panels through the upper fasteners and binding frames.
Thus, the column
panels bore part of the wind load on the windows and the beam panels. As the
test loads were modeled on the actual distributed wind loads and the concentrated loads from the beam panels, the specimens were loaded by inverse symmetry
cyclic loading. The test procedure is illustrated in Fig. 15. There were two
specimens and load directions were opposite each other, one being positive
pressure and the other negative pressure. The deflection of the specimens, the
slippage between the 3D-FRC panel and the steel frame, and surface strain of
the specimens were measured.
Test results at the design load, the initial crack, the flexural yield, and
the maximum flexural strength are summarized in Table 7 and load-deflection
curves are shown in Fig. 16.
No damage such as crack or tile separation was
apparent up to the design load. The deflection of the specimens increased elastically with increasing load, and the deflection ratios (deflection/span) at the
design load were sufficiently smaller than the design specification of 1/300,
being 1/1170 for positive pressure and l/7HO for negative pressure. The initial
cracking loads were 150% (positive pressure) and 110% (negative pressure) of
the design load, and crack widths were very narrow at less than !1.05 mm. The
yielding loads of the steel frame were 370% (positive pressure) and 240%
(negative pressure) of the design load. And the maximum flexural strength
were 570% (positive pressure) and 370% (negative pressure) of the design load.
Deflections at maximum flexural strength were extremely large: 51.5 mm for
positive pressure and 50.9 mm for negative pressure, demonstrating that the 3DFRC panels arc extremely tough, and there were no tile separations at this point.
Fire Resistance Test
The fire resistance tests of 60 minutes were conducted on one specimen
heated from the exterior face and another heated from the interior face, according to the JIS A 1304 standard fire test procedure (5). The specimens (see
Fig. 17) were modified from column panels to suit the size of the furnace and
were not finished with tile; all other specifications were the same as for the
panels to he used on the building. It was a primary consideration that thermal
expansion of the steel frame might cause a panel heated from the steel frame side
to crack, so, the holes in the steel frame for the panel connectors were oversized.
The furnace was held at the standard furnace temperature, and the temperature
and deformation of the specimens were measured.
The maximum temperature and deformation at each measuring point are
listed in Table H, and temperature histories arc shown in Fig. 1H. The maximum temperatures on the unexposed surface of the specimens were 103°C when
heated from the exterior face, and 11 0°C when heated from the interior face,
well below the 260°C, which is the maximum allowable temperature specified
hy regulations of JIS A 1304. The maximum temperatures at the 3-D fabric on
FRP Reinforcement
217
heated side were 516°C when heated from the exterior face and 349°C when
heated from the interior face, while those at the 3-D fabric on the hack side were
152°C and 1100C. Considering the results of tensile strength tests of 3-D
fabrics after heating to 400°C, shown in Fig. 19, the hack side Hwings of 3-D
fabrics had no damage, although the heated-side nwings apparently deteriorated. The maximum deformation appeared 10 minutes after the tests were
started, and was 28 mm under heating from the interior face.
Most of the
deformation was probably caused by deflection of the steel frame due to the
temperature differential between the outside and the inside of the steel frame.
There were no deformations, fractures, or cracks which would weaken fire
resistance, and the fire resistance of 60 minutes was achieved.
CONCLUSIONS
The material properties, flexural behavior and thermal durability of 3DFRC were examined. Both structural and fire resistance behaviors were also
investigated in order to use 3D-FRC panels for buildings.
The main conclusions arc as follows:
1)
The flexural characteristics of 3D-FRC members depend on the propertics of the matrix up to the initial crack occurrence, and on the material
and content of fibers used in the 3-D fabrics in the range beyond.
2)
3D-FRC made with carbon fibers or aramid fibers deteriorates very little
after being soaked in 80°C water or autoclavcd at 180°C.
3)
3D-FRC panels used for the parapet of the Suidohashi Building of Tokyo
Dental College had enough flexural strength and rigidity to withstand
design wind loads.
4)
3D-FRC curtain walls applied to the building of Sea Fort Square had
enough flexural strength and the fire resistance of 60 minutes.
3D-FRC panels have been used for curtain walls, parapets, partition
walls, louvers, etc., and installations amount to 7,000 m 2 . These panels are
expected to be used in a wide range of applications.
ACKNOWLEDGMENTS
The authors are grateful to Dr. Akihama, FRC Corporation, for his
valuable advice in this research effort. The authors would also like to thank
Arisawa Mfg. Co., Ltd. for supplying the 3-D fabrics.
218
Nakagawa et al
REFERENCES
1. Fujisaki, T., Sekijima, K., Matsuzaki, Y. and Okamura, H.: New Material for
Reinforced Concrete in Place of Reinforcing Steel Bars, IABSE Symposium
in Paris-Versailles, 1987, pp. 413-418.
2. Fukuta, K., Kitano, T., Nagatsuka, Y., Aoki, E., Funahashi, M., Shirai, S.,
Anahara, M. and Murayama, K.: Composite Materials Reinforced with Three
Dimensional Fabric, Progress of Japan-Sweden Research Cooperation in
Composite Materials, 1988, pp. 14-17.
3. Akihama, S., Suenaga, T., Tanaka, M. and Hayashi, M.: Properties of GFRC
with Low Alkaline Cement, Fiber Reinforced Concrete Symposium ACI,
Fall Convention in Baltimore, 1986, SP105-ll, pp. 189-209.
4. Akihama, S., Ogawa, K., Suenaga, T.. Uchida, 1., Fujii, H. and Hayashi, M.:
Development of New GFRC Cladding, 6th Biennial Congress of the GRCA
in Edinburgh, 1987, pp. 41-52.
5. Method of Fire Resistance Test for Structural Parts of Buildings, Japanese
Industrial Standard, A 1304, 1975.
TABLE 1 - PHYSICAL PROPERTIES OF FIBERS
Fiber type
Diameter
(11m)
Specific
gravity
Tensile
strength
(GPa)
Modulus of
elasticity
(GPa)
Elongation
(%)
PAN-type
carbon fiber
7.0
1.8
3.63
235
1.5
Aramid fiber
12.4
1.4
3.04
73.5
4.4
Vinylon fiber
14.2
1.3
1.47
36.3
6.6
AR- glass fiber
13.5
2.7
1.47
73.5
2.0
FRP Reinforcement
219
TABLE 2 - SPECIFICATIONS OF 3-D FABRICS
(FLEXURAL BEHAVIOR TEST)
Specimen
No.
X, Y axes
CF48
PAN-type
carbon fiber
CF72
Pitch of rovings
(mm)
Number of
filaments
Fiber type
Aramid fiber
AF24
X, Y axes
Z axis
12KF
30
20
48KF
0.54
0.79
72KF
16KF
AF16
Z axis
Tension
reinforcement
ratio Pt (%)
0.56
PAN-CF
12KF
24KF
30
20
0.82
KF : 1000 f1laments
TABLE 3 -MIX PROPORTIONS OF VFRC
W/C
(%)
S/C
47.3
0.13
Fiber content
v, (vol.%)
Cement
Aggregate
Admixture
0.0
Low shrinkage
cement
1.0
Micro balloon Methylcellulose
1.5
TABLE 4 - SPECIFICATIONS OF 3-D FABRICS
(THERMAL DURABILITY TEST)
Specimen
No.
Fiber type
CF72
Pitch of rovi ngs
(mm)
Number of
filaments
X, y axes
Z axis
Tension
reinforcement
ratio Pt (%)
12KF
20
30
0.45
PAN-CF
24KF
PAN-CF
12KF
20
30
0.47
18KF
PAN-CF
24KF
PAN-CF
12KF
20
30
0.45
19KF
PAN-CF
24KF
PAN-CF
12KF
20
30
0.45
X axis
Y axis
PAN-type
carbon fiber
72KF
24KF
AF24
Aramid
fiber
24KF
VF18
Vi nylon
fiber
GF19
AR-glass
fiber
z
axis
KF : 1000 filaments
TABLE 5 - MIX PROPORTION OF MORTAR MATRIX
W/C
(%)
S/C
Cement
Aggregate
Admixture
45
2.2
High- early- strength
Portland cement
River sand
Water reducing
admixture
TABLE 6- RESULTS OF WIND RESISTANCE TESTS
~
Specimen without granite facing
Specimen with granite facing
Displacement
Strain
Displacement
( Interior)
ratio
face
Displacement
Strain
Displacement
(Exterior)
ratio
face
Strain
( Interior)
face
Positive
design load
Upper rib
Plank
Lower rib
0.48mm
0.32mm
0.20mm
1/4580
1/6880
1/11000
59 IJ
46 IJ
49 IJ
0.94mm
0.87mm
0.59mm
1/2340
1/2530
1/3730
-68 IJ
-64 IJ
-61 IJ
83 IJ
41 IJ
58 IJ
Negative
design load
Upper rib
Plank
Lower rib
0.79mm
0.56mm
0.53mm
1/2780
1/3930
1/4150
-87 IJ
-64 IJ
-81 IJ
1.78mm
1.55mm
1.01mm
1/1240
1/1420
1/2180
103 IJ
103 IJ
99 IJ
-154 IJ
-65 IJ
-1151J
Maximum
positive
load
Upper rib
Plank
Lower rib
0.93mm
0.81mm
0.33mm
1/2370
1/2720
1/6670
184 J..l
130 IJ
141 IJ
2.09mm
2.15mm
1.31mm
1/1050
1/1020
1/1680
- 160 IJ
-172 IJ
-154 IJ
223 IJ
127 IJ
179 IJ
Maximum
negative
load
Upper rib
Plank
Lower rib
1.46mm
1.16mm
1.09mm
1/1510
1/1900
1/2020
-196 IJ
-191 J..l
-221 IJ
3.24mm
2.88mm
1.84mm
1/680
1/760
1/1200
202 IJ
194 J..l
186 IJ
-288 IJ
-113 IJ
-222 IJ
Displacement ratio
Displacement/Span
~
TABLE 7- FLEXURAL TEST RESULTS
\
Initial
rigidity
(kN/mm)
Design load
Load
(kN)
Initial crack
Deflection Deflection
(mm)
ratio
1
Positive
pressure
8.3
19.6
2.4
--
Negative
pressure
9.4
29.4
3.6
--
1170
1
780
Flexural yield
Maximum flexural
strength
Load
(kN)
Deflection
(mm)
Load
(kN)
Deflection
(mm)
Load
(kN)
Deflection
(mm)
31.4
4.1
76.5
6.1
107.9
51.5
33.3
4.2
70.6
10.5
110.8
50.9
Deflectron ratro - Deflectron/Span
TABLE 8- FIRE RESISTANCE TEST RESULTS
~
Heated from
the exterior face
Heated from
the interior face
Maximum
Time
(min)
Exposed surface
937.5
60
971.7
60
Unexposed
surface
103.0
69.5
110.2
103
Heated- side
83.1
84
952.7
60
Back side
56.5
101.5
864.6
60.5
Heated- side
515.5
61
348.6
62.5
Back side
151.7
73
109.9
77
Center of 3D- FRC panel
3.6
12
28.0
10
3D- FRC panel
Temperature
Maximum
Time
(min)
Steel frame
("C)
3-D fabric
Deformation
(mm)
I»
FRP Reinforcement
223
Fig. !-Conceptual drawing of 3-D fabric
3.0r--------------------,
PAN-type
carbon fiber
5.0
Strain
£
Fig. 2-Tensile stress-strain curves of FRP rods
(%)
224
Nakagawa et al
lp
,, 'R'
1.0
I
I I I I II I I I II 1 II I i l l ! I l l I I r I I I I I 1 I I I I I
'R'
-
400
400
I
400
N
I
I=;~~~
N
I too I ~
==:,-
3-D fabric (360 x 20 x 1 ,380)
00
(')0
@"
y
Lx
::::;;:;'-0
to~l_ _ _ _ _ _ _ _ ___::@_;_3o'-----------'-illto N
.
1,400
.
Fig. 3-Geometry of flexural test specimen
..
~
z
20 0..
5
~
0
0..
.,
II)
!!
';j
iii
10
.
5
)(
u::
0
0
0 100
100 100 0
Deflection 5 (mm)
0
0 1 00
1 00
100 200
200
Deflection 5 (mm)
Fig. 4-Load-deflection curves (CF48, AF16)
200
FRP Reinforcement
!
(/)
iii
10
.,~
u:
o~------t,----~,~ono------7-,s~o------~200
Deflection IS (mm)
Fig. 5-Load-deflection curves (Jj-
= 1.0 percent)
p
111£1 II II
1501
200
I
I
11111111: II II 111111111£
I
~:
I
200
II~ II II}]
-lsol w
Fig. 6-Geometry of durability test specimen
PAN-type
carbon fiber
Aramid fiber
AR- glass fiber
Vinylon fiber
Umit of l)rOC)Of'tionelity
(mNn of ell)
Fig. 7-Relation between MOR and days soaked in 80 C water
225
226
Nakagawa et al
-PAN-type
carbon fiber
AR- glass fiber
0
Number of autoclaving (180"C x 5hours)
Fig. 8-Relation between MOR and number of autoclave sessions
Fig. 9-Suidobashi Building, Tokyo Dental College
FRP Reinforcement
Fig. tO-Geometry of standard panel (Suidobashi Building)
8.8kN/m'
=2~1~#'_ ·~t ~t'=~
a..
i
'""""' .
1
Negative
CJ
-10
Static load
Fig. 11-Wind load test pressurization procedure
227
228
Nakagawa et al
,...----,----,---,i-,J
load (8.8kN/m'
!'E Max
-------·-----------·---·---
__, -·-·t"-
'
'
!?;S
---r----:·--·--:·;:5
2
i
4
i 6 (mm)
...........J _____ j ____ _
I
i
I
, ' I
--- -5-+--f-'
······-·-··J···~·-·--··-~·-·_j_
.
.
i
(Specimen with granite facing)
(Specimen without granite facing)
(Wind load vs. displacement)
I
! ..............1....
-··--···+
..... !
.................i .......--..1·········+···
-400
-200
...... l
-l
Design load
(-5.0kN/m')
------T------r·---- ·.
i
1
---t--j- ..
(Specimen with granite facing)
(Specimen without granite facing)
(Wind load vs. strain)
Fig. 12-Relation between wind load and displacement and strain
FRP Reinforcement
Fig. 13-Sea Fort Square
a
I
BOO 25I
a·
z,;m
~
!BOOk
2.~
a
::!!.
B-8'
section
A-A' section
BOO I
Fig. 14-Geometry of curtain wall panel (Sea Fort Square)
229
230
Nakagawa et al
700
2.100
700
J1'!
BOO
IQO
I
Fig. 15-Flexural test procedure (Sea Fort Square)
z
12Sr---~----r---,---~
.><
~ 100
z
6
12Sr----r----r---,---~
a..100
]
]
so
0
(Positive pressure)
(Negative pressure)
Fig. 16--Load-deflection curves
d
100
FRP Reinforcement
231
:;::::::·::-::·::::·::;:
r-----,
I
I
I
b.
:
I
I
o:
0
I
D.:
I
I
I
I
o:
0
I
I
I
I
IU
b.
0
0
b.
I
L-----J
'-....
__
ot
11:
II!
11:
II!
.,:
:!
L-----J I!
11:
Ill
:
L~.::;:.::.-:.::..::.:-.~J
i.!-:.::.-:.::.;;.::.::.::•.1
Exterior face
••!
Measuring points
]db[
]0
Interior face
I
940
J5
940
2,870
940
0
b.
:Temperature
: Deformation
I
Fig. 17---Geometry of fire resistance test specimen
Time (min)
Fig. 18-Temperature histories (heated from the interior surface)
232
Nakagawa et al
,....., 4.0 r - - - - - , - - - - - - - , - - - - - , - - - - - - - , - - - ,
co
a..
(!)
.........
..r:.
....Cl
1:
....~
3.0
(/)
~
·u;
1:
~
2.01---------~--------~--------+-------~--~
100
200
400
Temperature (°C)
Fig. 19-Relation between tensile strength (after heating)
and maximum temperature for 3-D fabrics
SP 138-15
Strengthening of Reinforced
Concrete Chimneys, Columns
and Beams with Carbon Fiber
Reinforced Plastics
by C. Ballinger, T. Maeda,
and T. Hoshijima
Synopsis:
This paper presents information on the development, and actual
use, of carbon fiber reinforced plastic (CFRP) to strengthen
reinforced concrete chimneys, bridge piers, and beams in Japan;
bridge beams in Switzerland; and ongoing structural research and
use of fiber reinforced plastic (FRP) composite materials to
strengthen such structures in the U.S. The concept and equipment
for strengthening existing reinforced concrete chimneys, by
wrapping them with carbon fiber reinforced plastic materials,
began in Japan.
The procedure permitted earthquake damaged
chimneys to be repaired without taking them out of service.
Research in Switzerland has led to the use of adhesively bonded
sheets of carbon fiber reinforced plastic laminates to strengthen
existing bridges. This concept is an extension of extensive use
of bonded stee 1 p1ates to strengthen many types of structures
throughout Europe. Research, development and some use of these
techniques has been done in the U.S.
Keywords: Bridge piers; bridges (structures); carbon; chimneys; columns
(supports); composite materials; concretes; fiber reinforced plastics;
floors; glass fibers; strengthening; wrapping
233
234
Ballinger, Maeda, and Hoshijima
ACI fellow Craig A. Ballinger is President of Craig Ballinger &
Associates, Vienna, Virginia. He is retired from the Federal
Highway Administration, where he managed the FHWA's R&D program on
structural FRP composites. He is a member of ACI Committees 215,
408, 437, 440 and 503; and Chairman of the Transportation Research
Board's Committee A2C07 on Structural Fiber Reinforced Plastics.
Toshikatsu Maeda is Vice President of Mitsubishi Kasei America,
Inc., Menlo Park, California.
Tokitaro Hoshijima is a Senior Manager, Carbon Products Division,
Mitsubishi Kasei Corporation, Tokyo, Japan.
INTRODUCTION
Deterioration of reinforced concrete due to corrosion of the
reinforcing steel and environmental effects on the concrete is a
long-term but almost inevitable phenomenon.
Inadequacies and
errors in design produce structures that are relatively weak.
Excessive in-service loading caused by wind, earthquakes, live or
dead loads may crack and weaken even new structures.
Taken
together, deter i oration and excessive loading are the cause of
loss of function of hundreds of thousands of reinforced concrete
structures all over the world.
Although this situation is well known, prediction of the rate
and/or degree of deterioration and weakness is a function of how
well the design code matches the loading and environmental
conditions. The need, or priority, for repairing or retro-fit
strengthening the structures is often a function of the degree of
perceived weakness, and the financial resources available to the
owner.
There is no question that there is not enough money
available to demolish and rebuild all deteriorated and weak
reinforced concrete structures. Thus, improved, cost effective
and reliable techniques and materials are urgently needed to
restore or upgrade such structures, so that they may have a long
useful service life.
Moreover, it is essential that at least lifeline structures
(highways, pipelines, power liner, etc.) and those that pose a
significant hazard to the public if they would collapse must
either be structurally strengthened or demolished.
Such
structures include: bridges, buildings and tall chimneys.
Until recently, the most common method of strengthening was
to install circumferential reinforced concrete or steel jackets
around circular sections and (in some cases) to adhesively bond
steel plates to the sides and underside of flat beams. Although
FRP Reinforcement
235
these concepts are genera 11 y quite effective they do have the
following drawbacks:
Reinforced Concrete Jackets
Increases weight and cross-sectional area of structure
may influence seismic response and/or clearance for
function.
Additional work is necessary to remove existing
concrete cover over steel reinforcement and to attach
the additional steel reinforcing.
Steel Plate Jackets
• Difficult and expensive to install; individual plates
are heavy and they must be welded together to form a
continuous jacket.
Since steel jackets cannot be constructed to fit in
direct contact with the existing concrete shape, the
annular space between the steel plate jacket and the
concrete must be filled with a structural concrete or
epoxy-type resin that is capable of transferring
structural loads. This process is labor-intensive and
very difficult to completely fill the annular space.
Although less than with reinforced concrete jackets,
the increases weight and cross-sectional area of
structure may influence seismic response and/or
clearance for function.
Adhesively Bonded Steel Plates
Somewhat difficult and expensive to install; because
the individual plates are heavy and they must be
welded together to form a continuous integral plate.
Weight of plates creates significant stresses on the
adhesive
Thickness of plates may restrict clearance for
function.
DEVELOPMENT OF FIBER REINFORCED PLASTIC STRENGTHENING CONCEPTS
Developments in Japan
The need to strengthen structures is particularly critical in
Japan. where many earthquakes occur every year. Some of them are
quite 1arge and they cause extensive damage to many types of
structures. Structures designed on the basis of the old Building
Standards Act and the old design specifications for highway
bridges do not have sufficient strength to withstand a major
earthquake.
In 1984 the Japan Concrete Engineering Society published the
"Earthquake Retrofitting of Existing Reinforced Concrete
Structures," which recommended use of reinforced concrete and
steel jackets to strengthen structures. It was subsequently found
that these methods are expensive and difficult to install. for the
reasons cited above.
236
Ballinger, Maeda, and Hoshijima
Research and deve 1opment work on new and more economi ca 1
techniques for strengthening reinforced concrete structures began
several years ago in Japan. The Japan Highway Public Cooperation
(JH) and the Carbon Fiber Retrofitting System (CRS) study group
(Obayashi Corporation and Mitsubishi Kasei, and others). The
focus of this program was on use of carbon fiber reinforced
plastic prepreg and carbon fiber strand (CRS). These materials
have the following merits:
High tensile strength, 2450 Mpa (355 ksiJ
High elastic modulus, 236 GPa (34.1 x 10 psi)
Very low weight/stiffness ratio (1/5th that of steel)
Negligible increase in dimensions of the existing
structural member
Excellent durability
Not subject to corrosion
Retrofitting work is much easier and faster than with
steel and reinforced concrete jackets and bonded repairs.
With regard to disadvantages, or negative attributes, their
magnitude and effect are often a matter of degree. Nonetheless,
the following factors should be recognized:
Ultraviolet (UV) degradation of the resin; that can be
minimized with a surface gel coat and/or with additives to
the resin.
Moisture permeability of the resin and degradation of the
resin properties. This is a function of the type of resin
and the degree of cure.
Influence of ambient temperature. Changes in temperature
will affect resin properties, especially thermoplastics.
Flammability. Although properly formulated resins may
burn, they do not support combustion.
Smoke and toxicity. Although smoke from burning resins
may be toxic the level of hazard is relatively low,
because FRP repairs are normally in the open or the amount
of material is small.
The research and development work of this group was directed
to three major types of reinforced concrete members: bridge
columns, tall chimneys, and beams and floors.
The following
describes the R&D work and how the techniques have been used to
date.
Bridge Columns -- Bridge column designs based on the old
Japanese specifications do not have sufficient strength against
earthquakes:
the main vertical (flexural) reinforcement is
usually terminated at mid-height and the lateral circumferential
confining reinforcement is inadequate to provide the necessary
shear strength and ductility.
The Japan Road Associ at ion's
"Design Specifications for Highway Bridges" now contain improved
criteria for seismic design of bridges. (1)
In essence these
specifications indicate that in order to avoid brittle failure,
bridge columns should meet the following criteria:
1. The shear strength shou 1d be higher than the f lexura 1
strength.
FRP Reinforcement
237
2. The lateral resistance should be maintained when the
column is subjected to large deformations.
3. The lateral resistance should be determined by the
flexural strength of the bottom portion of the column.
The new retrofit concept developed by the CRS study group is
shown in Figure 1. It involves application of sheets of carbon
fiber reinforced plastic (CFRP) in the longitudinal (vertical)
direction and CFRP sheets or strands in the transverse
(circumferential) direction, to provide the necessary flexural and
shear strength and ductility to meet the current seismic design
criteria.
The basic design concept for the CRS method is
ductility-oriented.
The structural test program involved testing five 1/3 scale
specimens that were based on the columns used in a bridge on the
Tomei Highway. The dimensions and reinforcing bar arrangements
for the specimens are shown in Figure 2. The retrofit options
studied are shown in Table 1. The tensile strength and elastic
modulus for the carbon fiber materials used for the specimens are
shown in Table 2.
For specimens No. 1, 2 and 5 the main
reinforcement was terminated at the height of 90 em from the
foundation, at which point the amount of longitudinal
reinforcement was reduced to 1/2 of the reinforcement in the lower
portion of the column.
For specimens No. 3 and 4 the main
reinforcement was not terminated, to study the performance of
retrofitting the bottom part of the column. The percentage of
longitudinal and transverse reinforcement was 1.2% and 0.047%,
respectively.
For flexural strengthening, the moment distribution for the
specimens is shown in Figure 3.
The amount of carbon fiber
flexural strengthening was calculated so that the ultimate
flexural force of the main bar termination should be higher than
the ultimate flexural force of the bottom portion. It was also
assumed that deformation of the strengthened portion should be
relatively small. The flexural design was based on the following:
1. The shortest delta-M moment was calculated by:
delta-M = Mut - Mr
(1)
where: delta-M = shortest moment; Mu~ = The maximum moment
which acts at the main-bar-terminat1on portion until the
bottom portion reaches the ultimate flexural state; Mr =
Maximum flexural moment resistance.
2. The amount of flexural strengthening was calculated by the
following equation, converting from the amount of steel
reinforcing to that of carbon fibers:
AcF
=
As( 0 s/ 0 cF)
(2)
Where: AcF = Required amount of carbon fibers; A =
Required amount of steel reinforcing bars, correspon~ing
to delta-M; a~= Yield strength of steel reinforcing bars;
acF = Carbon tiber seismic design strength.
The seismic design strength used for carbon fibers was 1640
Mpa (237 ksi), which is based on the assured tensile strength of
238
Ballinger, Maeda, and Hoshijima
2450 Mpa (355 ksi) and a 1.5 safety factor. The stress-strain
relationship or carbon fiber is shown in Figure 4. The stressstrain relationship for concrete and steel reinforcing bars was
the same as the current code criteria.
However the ultimate
compressive strain for concrete confined by transverse carbon
fiber materials was found to be 0.0040 rather than the Japanese
code criteria of 0.0035.
The amount of shear strengthening was calculated, based on the
following criteria:
1. Bridge columns should not fail in shear.
2. Shear force acting in the columns should be resisted only
by the retrofitting carbon fibers.
3. The amount of carbon fibers can be converted from the
amount of reinforcing steel bars, using equation 2.
4. The contribution of shear reinforcement to the shear
capacity is the same as required by the current seismic
design criteria.
V5 = PwUWY.bd/1.15
(3)
Where: V5 = Contribution of shear reinforcement to shear
capacity; Pw = Percentage of shear reinforcement; a =
Yield strength of shear reinforcement; b = Column width;
d = Effective depth of column.
The loading apparatus is shown in Figure 5. For specimens ~o.
1 and 4 a constant axial force (axial stress = 588 N/mm ),
equivalent to the structural dead load, and lateral cyclic loads
were applied. Only lateral loads were applied to specimen No. 5.
The typical lateral loading pattern is shown in Figure 6; which
indicates the maximum lateral displacement of 6 delta y, that is
6 times the displacement necessary to cause yield of the vertical
reinforcing bars.
The test results are summarized in Table 3 and the load
displacement relationships are shown in Figure 7 (a) to (f). The
definitions used for evaluation are:
1. Cracking - development of first crack
2. Yielding - The longitudinal reinforcement begins to yield
in tension
3. Ultimate - concrete was heavily damaged.
- carbon fibers began to fracture critically
displacement reached maximum of loading
apparatus, #4.
4. Ductility Factor (J.L)
ultimate displacement/yield
displacement, (8 0 /&y)
Analysis of the test results provides guidance for designing
the retrofit zone for bridge columns with terminated longitudinal
reinforcement, as shown in Figure 8. At first the acting moment
is defined as the acting maximum moment unt i 1 the base of the
column reaches the ultimate flexural state.
The retrofit is
applied up to the point at which the acting moment is 90% of the
moment resistance of the main-bar-terminated zone before
retrofit.(2,3)
FRP Reinforcement
239
The following conclusions were drawn from the research study:
1. The f1 exura 1 strength of co 1umns is increased by
adhesively bonding carbon fiber sheets in the longitudinal
(vertical) direction.
2. The shear strength and ductility are improved by wrapping
carbon fiber strands and/or sheets in the transverse
(circumferential) direction.
3. Under adequate combination of vert i ca 1 and hor i zonta 1
reinforcement with carbon fiber materials, the earthquake
capacity of bridge columns with terminated primary
reinforcement can be improved. The potential failure zone
moves from the zone of terminated reinforcing to the base
of the co 1umn, and the duct i1 i ty of the co 1umn base is
improved.
Chimneys -- Over the years earthquakes have caused severe
damage to thousands of reinforced concrete chimneys in Japan. For
example, the Fukui earthquake (1948) and the Miyagiken Oki
earthquake (1978) caused the collapse of a majority of the
chimneys in Japan. Many of the surviving chimneys were severely
cracked. Following those earthquakes it was recognized that the
majority of the collapsed chimneys broke at approximately 2/3 of
their height.
Recognition of this fact subsequently led to
development and use of the bridge column retrofitting concept to
strengthen both damaged and undamaged chimneys.
In order to develop an economical procedure for retrofitting
very tall chimneys it was necessary to develop automated equipment
The chimney
for applying the carbon fiber prepreg material.
wrapping machine and process are shown in Figure 9. Use of the
machine does not require the use of either heavy cranes or
scaffolding, and it may be used in areas with limited space around
the chimney. Of primary importance is the fact that this retrofit
concept does not require restricting or stopping use of the
chimney.(4)
The
1.
2.
3.
following steps are used for the retrofit procedure:
Remove personnel ladder and lightning conductors.
Repair severely damaged areas of concrete on the surface.
Bond carbon fiber reinforced plastic (CFRP) sheets in the
vertical direction and then wind CFRP sheets or strands
around the chimney.
4. Apply coating for environmental protection
5. Reinstall the personnel ladder and lightning conductors.
Repair of Beams and Floors-- A limited amount of research has
been conducted by the CRS study group on flexural strengthening
reinforced concrete beams by bonding carbon fiber materials to the
tensile surface.
One prestressed concrete highway bridge in Japan has been
repaired with bonded CFRP materials.
The floor of an apartment was cracked, with cracks as wide
240
Ballinger, Maeda, and Hoshijima
as 4 em (1.6 in).
It was felt that deflection and
vibration would increase the cracking.
Although such cracks are frequently repaired by injection of
an epoxy resin, it was felt that for this case the epoxy wou 1d
leak through the floor into the apartment below. The repair was
done by bonding carbon fiber reinforced plastic (CFRP) over the
cracks to prevent crack growth and corrosion of the existing steel
reinforcing bars. The cracks were repaired by bonding 50 em (19.7
in) wide material per 1 m (3.3 ft) of crack length. The CFRP tape
extended 30 em (11.8 in) on each side of the cracks. The tensile
strength of the CFRP tape was 175 Kg/em of width.
Developments in Switzerland
Over the past 10 years research has been conducted at the
Swiss Federal Institute for Research and Testing (EMPA) on
strengthening beams with adhesively bonded carbon fiber laminates
(cured CFRP plates). (5) This was essentially an extension of
considerable research and over 20 years of actual use of bonded
steel plates to strengthen reinforced concrete beams, floor, and
walls throughout Europe.(6-9)
In 1989 the 228m (748ft) multispan continuous prestressed
concrete box girder Ibach bridge, in Lucerne County, was damaged
when a prestressing strand in a 39m (128 ft) span was accidently
cut.
The damage caused "posting" of the bridge, for heavy
military convoys.
The bridge was repaired in 1991 by bonding 2 mm (0.079 in)
thick by 150 mm (5.9 in) wide strips of CFRP laminate to the
underside of the damaged beam. The CFRP laminate had a tensile
strength ot 1900 Mpa (275 ksi) and an elastic modulus of 129 Gpa
(18.6 x 10 psi).
With regard to the economics of using high priced CFRP
materials, it should be noted that the weight/stiffness ratio for
carbon fiber is 1/5 of that for steel, and the strength of carbon
fiber is considerably greater than common construction-grade
steel. For the Ibach bridge the repairs were made with 6.2 kg
(13.7 lb) of CFRP laminates rather than 175 kg (386 lb) of steel
plates, a 28.2:1 reduction in weight. Moreover, the repairs were
quickly made at night with a lightweight mobile man-lift platform
rather than with more heavily rated (and more expensive) lifting
equipment.
Developments in the United States
Although the U.S. Federal Highway Administration and several
State highway departments have expressed interest in FRP
FRP Reinforcement
241
materials, research and development and actual use of such
materials is advancing quite slowly. This is primarily the result
of very limited private industry interest in funding the
development of the necessary FRP materials and application
technology. (10)
Bridge Columns -- The only use of structural FRP materials for
highway bridges in the U.S. is in California, where some bridge
columns are being wrapped with glass and aramid fiber reinforced
plastic, to provide increased strength against earthquakes. This
retrofit work is the result of recent bridge failures near San
Francisco and subsequent structural analyses that indicated that
many bridge columns (piers) did not meet current seismic design
criteria.
Selection of the materials and development of the technique
was done at the University of California at San Diego. Use of FRP
materials is an alternate to the commonly accepted jacketing with
structural steel plates and the subsequent filling of the annular
space between the column and the jacket with a cement-based
grout. ( 11)
Bu i 1dings -- Although FRP materia 1s have not been used to
repair buildings in the U.S., some buildings have been repaired
with bonded steel plates, in a manner similar to that commonly
used throughout Europe. Use of bonded steel plates for buildings
is probably due to the fact that building owners are more
concerned with the installed 1ife-cyc le cost effectiveness of
repairs vs. the cost to reconstruct. Use of bonded steel plates
shou 1d 1ead the way to use of bonded carbon fiber reinforced
plastic (CFRP) materials.
CONCLUSIONS
The foregoing has clearly shown that it is technically
possible to strengthen reinforced concrete structures with carbon
fiber reinforced plastic (CFRP) materials. There is no question
that the cost of such materials is high and their use involves
materials and technologies that are not well known to civil
engineers. Nonetheless the results of extensive R&D work and use
of these materials both overseas and within the U.S. has shown
that their use is not only viable but also economically
justifiable. With regard to the question of using of glass fiber
reinforced p1ast i c ( GFRP) rather than carbon fiber reinforced
plastic (CFRP), the answer is in the cost vs. structural
properties relationships. One evaluation factor is the installed
cost to carry a pound of load. Although glass fiber materials are
much less expensive than carbon (or graphite) fiber materials, the
effective strength of GFRP is less than 2/3 that of CFRP and the
elastic (E) modulus is about 1/6 of that of CFRP and steel.
There is no question that at least some of the thousands of
242
Ballinger, Maeda, and Hoshijima
damaged and/or substandard highway bridges could effectively be
repaired with CFRP materials.
Similarly, such materials and
techniques could also be used to repair a wide variety of types of
buildings that have been damaged because of corrosion or
overloads, or to retrofit structures to prevent possible damage
due to earthquakes. The overriding question is not the cost per
pound of the materials, but rather the installed cost
effectiveness of the repair work. Although CFRP materials and
installation techniques may not be completely developed such
materials have many attributes that are superior to currently used
materials, and continued work to develop this technology is
warranted.
ACKNOWLEDGEMENTS
Information for this paper was obtained from several sources,
including those cited in the References.
REFERENCES
1. JRA, "Design Specifications for Highway Bridges," (Part 5 on
Seismic Design), Japan Road Association, 1990.
2. Katsumata, H., Matsuda, T., et. al., "Earthquake-Resistant
Capacity of Reinforced Concrete Bridge Columns Retrofitted
with Carbon Fibers," unpublished paper by Obayashi Corporation
and the Japan Highway Public Corporation, 1992.
3. Higashida, N., Kobatake, Y., et. al., "Retrofit of Reinforced
Bridge Columns with Carbon Fibers," Proceedings of the 45th
annual meeting, Japan Civil Engineering Association, 1990.
4. Katsumata, H., Vagi, K., "Applications of Retrofit Method with
Carbon Fiber for Existing Reinforced Concrete Structures," by
Obayashi Corporation and Mitsubishi Kasei Corporation,
Presented a 22nd Joint UJNR Panel Meeting on Repair and
Retrofit of Existing Structures," at National Institute for
Standards and Testing (NIST), Gaithersburg, MD, 1990.
5. Meier, U. and Kaiser, H. , "Strengthening of Structures with
CFRP Laminates," Specialty Conference on Advanced Composite
Materials in Civil Engineering Structures, Las Vegas, Nevada,
American Society of Civil Engineers, New York, January 1991.
6. MacDonald, M.D., "The Flexural Behavior of Concrete Beams with
Bonded External Reinforcement," Department of the Environment,
Department of Transport,Supplementary Report 415, Transport
Road Research Laboratory, Crowthorne, Berkshire, U.K., 1978.
7. Raithby, K.D., "External Strengthening of Concrete Bridges
FRP Reinforcement
243
with Bonded Steel Plates," Department of the Environment,
Department of Transport, Supplementary Report 612, Transport
and Road Research Laboratory, Crowthorne, Berkshire, U.K.,
1980.
8. Mays, G.C., "Structural Applications of Adhesives in Civil
Engineering," Materials Science and Technology, Vol. 1,
November 1985.
9. Swamy, R.N., Jones, R., and Bloxham, J.W., "Structural
Behavior of Reinforced Concrete Beams Strengthened by EpoxyBonded Steel Plates," The Structural Engineer, Vol 65A, No.
2, February 1987.
10. Klaiber, F.W., Dunker, K.F., Wipf, T.J., and Sanders, W.W.
Jr., "Methods of Strengthening Existing Highway Bridges,"
National Cooperative Highway Research Program, Report 293,
Transportation Research Board, National Research Council ,
Washington, D.C., September 1987.
11. Seqad, "Flexural Test of High Strength Fiber Retrofitted
Columns," Fyfe Associates, Del Mar, California, Report #91-04,
May, 1991.
TABLE 1 - DETAILS OF TEST SPECIMENS
Speci.en
Hullber
Position
Included
Included
Not Included
lot In~~~~
of
Roooe of '' ·
Ten~lnatton
R~trofit
Nothing
F • S
s
s
Ter~~inati'
1
Port·on
450
450
-;30
-,00
"l!.:i
~
c,
'"
Lf..i.l><ll
ortJ.Qn_
J2lli>L
"
-600
2 sheets
2 sheets
:~
2;h;;ts
2-;h;.t
1 sheet stund
~
:::::
TABLE 2 - MATERIALS PROPERTIES OF CARBON FIBERS
Specimen
Numbers
2, 4
3
5
Tensile
Strength
Moa
2,880
2,860
2 850
Young's
Modulus
Goa
246
226
236
244
Ballinger, Maeda, and Hoshijima
TABLE 3 -TEST RESULTS
Specimen
Nullber
c, .. ""'
2.6
2.8
1.9
2.2
6.7
27.4
28.4
24.5
29.4
34.3
1
2
3
4
5
...,;.
'1d
,.d(Kol loho.l•l
odlkNI loho.l-1
101.9
123.5
127.4
116.6
98.0
20.0
27.6
26.2
22.0
22.3
,.d(Kol
112.7
136.2
136.2
151.9
137.2
'"'
t\O<te
I Di<O.IMI
61.7
,.d(Kol
112.7
131.3
142.1
151.9
114.7
55.3
52.1
244.2
63.8
"'
I Oi<o.IMI
61.7
113.0
131.0
244.2
111.7
Ductility
factor
Failure
-
Failure
Po~it1on
3.0
4.1
5.0
11.1
5.0
terain.
bottoa
botttMt
bottoa
bott011
Bending
Shearin
Bending
--
Bending
CF Sht.cls or S1r.111ds
Fig. 1-Carbon fiber retrofitting technique
-
~
..,._
-
:-----,.--
0
0
;;
;;
-
-
...
...
0
~
Ar
0
0
!A
016
r;:
nr
~
-
~
D 10
-
~
D 16
~
"'
M
,o
0
c1Jl.
0
0
D.lQ
l..\
...
,o
llr
DIO
"'
~~~
·-·
~
48
DI
48
~
0
0
__!lli!
~
0
0
-
-.
0·8 scclion
--
Fig. 2-Bridge column test specimens
0
~
~·~
"'"'
FRP Reinforcement
Yielding Momer.: nt
Termin:-.liOI'\ Portion My
8-o
8....
srir•
column
Acting
Mo:-nent :1.1
Termination Ponion M'Jt
N
Termination Portion
Maximum
Mom em
Flexural Moment(tf· m)
Fig. 3-Moment distribution for test specimens
.. Er
--~----?r------~
ord:Design Stress at Carbon F>ber (kgl/em1)
" - - - - - j O ld
or:Streu of Carbon Fiber (kgf/cm')
Ef:Young's Mod~ Ius of CMbon Fiber (kgf/cm')
Er=2.S•I06
tr:Strain of Carbon Fiber
Fig. 4-Stress-strain relationship for carbon fibers
Orid c Pier
Reaction
0
Wall
~
Test noor
1100
Fig. 5-Loading apparatus for tests
245
246
Ballinger, Maeda, and Hoshijima
18
DisplaccmcOJ control
17
16
15
12 I J
14
Crack
Uy
JOy
40y
SJy
6oy
Fig. 6---Typical loading pattern for tests
p (tl)
IS
.,,,
z~,
]$,
tOO
t~O (~)
l
-II
(d) No.4 Specimen
(a) No. I Specimen
p (til
tl
16, J6,
•d.
100
!'50
61-1
(b) No.2 Specimc::
(e) No.5A Specimen
(c) No.3 Specimen
(I) No,5B Spcc.:uucn
Fig. 7-Typical load-displacement relationship for tests
FRP Reinforcement
CF Rcinrorccd
in Shear
1. . . . . . . . .,,~~
Fig. 8-Carbon fiber retrofit region
247
248
Ballinger, Maeda, and Hoshijima
Temporary removal of
gangway ladder and
lightning conductors
Painting of airplane
markings
Fig. 9-Chimney wrapping machine and process
SP 138-16
Behavior of Externally
Confined Concrete Columns
by H. Saadatmanesh, M.R. Ehsani,
and M.W. Li
Synopsis: Fiber composites have become increasingly popular in the civil
engineering community in recent years. The primary area of research and
development of fiber composites in the concrete industry has been related to
fiber-reinforced-plastic (FRP) rebars. This paper presents a different application
of fiber composites in concrete structures; namely, confinement of concrete
columns with fiber composite wraps for improved ductility and seismic
performance. The confinement is accomplished by wrapping thin, high-strength
fiber composite belts around the columns. The belts are made very thin to
result in sufficient flexibility for them to be wrapped around the circular as well
as rectangular columns. The belts can be wrapped around the column in
individual rings or in a continuous spiral. The ends of the belts can be
mechanically coupled or the belts could be epoxy bonded to the column. The
confinement provided by the belts results in significant increase in the crushing
strain of concrete well above that of unconfined concrete. This will improve the
overall strength and ductility of the column. The paper presents the results of
an analytical study and an ongoing experimental study of concrete columns
externally confined with fiber composite wraps.
Keywords: Bridges (structures); carbon; columns (supports); composite
materials; confined concrete; ductility; earthquake resistant structures; fiber
reinforced plastics; glass; shear strength; strengthening; wrapping
249
250
Saadatmanesh, Ehsani, and Li
ACI member Hamid Saadatmanesh is Assistant Professor of Civil
Engineering and Engineering Mechanics at the University ofArizona in Tucson.
His research interests include rehabilitation and strengthening of structures and
the application of fiber composites in civil engineering structures. He is the
Secretary of ACI Committee 440, FRP Tendon and Reinforcements.
Mohammad R. Ehsani is Associate Professor of Civil Engineering at the
University of Arizona. He is a Member of ACI Committee 408 where he chairs
a Subcommittee on Bond of FRP Rebars. As a member of Committee 440 he
chairs the Subcommittee on State-of-the-Art Report. Dr. Ehsani is a registered
professional engineer in Arizona and California.
Mu- Wen Li is a Graduate Research Assistant at the Department of Civil
Engineering and Engineering Mechanics at the University ofArizona in Tucson.
His research interests include seismic design of structures and application of
fiber composites in civil engineering structures.
INTRODUCTION
Bridge failures in the recent earthquakes in California have attracted the
attention of the bridge engineering community to the substandard detailing of
bridges for seismic forces. The structural inadequacies in many of the bridge
columns constructed before the new seismic design provisions were adapted
include:
Shear Strength
Many columns in older bridges contain inadequate transverse
reinforcement. A typical detail usually consists of No. 3 or No. 4 transverse
hoops at 300 mm (12 in) spacing, regardless of the column cross sectional
dimensions. This inadequate lateral reinforcement results in columns with low
ductility and poor energy absorption characteristics.
Low Flexural Ductility
In many older bridge columns the ends of the hoops were often simply
lapped in the cover concrete for anchorage. As a result, the amount of
confinement provided by the hoops will become ineffective after the ultimate
concrete compression strain has been reached and the cover concrete has
FRP Reinforcement
251
spalled.
Inadequate Lap Length
Prior to the new seismic design specifications, the column longitudinal
reinforcement used to be spliced with starter bars extending from the foundation
with a lap length of 20 times the bar diameter. For larger size bars, this will
not be enough to develop the yield strength of the bars.
Many researchers have shown that increasing the confinement of concrete
in the potential plastic hinge regions will increase the failure strain of concrete
and therefore the overall ductility. For this reason, retrofit methods typically
utilize schemes for increasing the confining forces either in the potential plastic
hinge regions or over the entire column.
In this paper a new technique for seismic strengthening of concrete
columns is presented. This technique involves wrapping thin, high-strength
composite belts around the column. The belts can be wrapped in individual
rings or in a continuous spiral as shown in Figure 1. The ends of the belts can
be mechanically buckled or the entire belt can be epoxy bonded to the column.
Additional advantages can be gained at the time of strengthening by laterally
prestressing the column. The prestressing can be accomplished by wrapping the
belts under tension around the column. Lateral prestressing of columns will
delay crack formation and improve serviceability. The following is a summary
of the benefits of this strengthening technique.
Increased Ductility
As a result of the confinement provided by the belts the concrete will fail
at a larger strain than if unconfined. Depending on the degree of confinement
the failure strain of concrete can be increased by a factor of six or more.
Increased Strength
The lateral pressure exerted by the belts will increase the compressive
strength of the concrete in both the core and shell regions, resulting in higher
axial load carrying capacity. The lateral confinement provided by the belts will
also provide additional support against buckling of the longitudinal bars.
252
Saadatmanesh, Ehsani, and Li
Circular and Square Sections
The flexibility of the belts allows them to be wrapped around circular as
well as rectangular columns.
Low Maintenance
Because of their resistance to electrochemical deterioration, fiber
composites are not affected by salt spray, moisture and other aggressive
environmental factors; therefore, no corrosion protection or painting will be
necessary.
However, ultraviolet light can adversely affect some fiber
composites. This problem can be eliminated by providing a protective coating
for the straps during the manufacturing process.
Light Weight
The low density of composites (typically less than one-fifth that of steel)
will significantly simplify the construction procedure and will reduce cost.
Temporary vs. Permanent
The proposed method will cause no disturbance to the integrity of the
existing structures; i.e., no anchor bolts, dowels, etc., will be required. As a
result, this method can be used as a permanent or temporary solution. For
example, in the case of mechaniclly coupled belts, if at a later time more
effective alternatives are developed, the belts can be easily removed.
Aesthetics
The belts are very thin, therefore, they will not alter the appearance of
the structure. If desired, a thin layer of concrete can be applied to cover the
straps.
In the following sections analytical models that quantify the gain in
strength and ductility of retrofitted columns as well as an ongoing experimental
study at the University of Arizona are discussed.
FRP Reinforcement
253
ANALYTICAL MODELS AND PARAMETRIC STUDY
Analytical models were developed to calculate the ultimate moment,
curvature at failure, and the corresponding axial load for circular and
rectangular column cross sections confined with fiber composite belts. The
models were used to generate axial load-moment-curvature interaction diagrams
for columns with various design parameters. The following assumptions were
made in the analysis.
1. Linear strain distribution through full depth of cross section;
2.
3.
4.
5.
6.
Small deformations;
No creep and shrinkage deformations;
No shear deformation;
No tensile strength for concrete;
Complete composite action between confining composite materials
and concrete column, i.e., no slip;
7. No confinement contribution from original stirrups; and
8. Uniform confinement at comers and along sides of rectangular cross
sections. It is noted that the confinement will be less effective for
rectangular cross sections as compared with circular cross sections.
Experimental studies will be needed to determine the exact level of
confinement in rectangular sections. Here, however, for illustration
purposes, it is assumed the same confinement is provided for circular
sections and rectangular sections.
The stress-strain curves for confined and unconfined concrete for circular
sections in compression proposed by Mander et al. [l] were used in the
analysis. The stress-strain model is illustrated in Figure 2, and is based on an
equation suggested by Popovics [2]. An approach similar to the one used by
Mander, Priestly and Park [3] was adopted to determine the effective lateral
confining pressure for use in the equation of stress-strain curve proposed by
Mander et al. The effective lateral confining pressure is given by
(1)
where
(2)
f1 = lateral pressure from transverse reinforcement, in this case, the composite
belt; ~ = confinement effectiveness coefficient; Ae = area of effectively
confined concrete; and Ace = effective area of concrete enclosed by composite
belts.
The lateral confining pressure may be found by considering the free-body
254
Saadatmanesh, Ehsani, and Li
diagram as shown in Figure 3. Under the combined effects of axial
compression and column flexure, the compression zone dilates as failure
approaches. The outward expansion of the core concrete is restrained by the
composite belt placing the belt in circumferential tension and the concrete in
radial compression. From the equilibrium of forces the lateral confining
pressure in the composite belt can be obtained as
(3)
where p,, = ratio of volume of composite belt to the volume of confining
concrete core; and fu, = tensile strength of composite belt.
Assuming that an arching action in the form of a second-degree parabola
with an initial slope of 45 o occurs in the clear area between successive belts, the
area of effectively confined concrete core at midway between the belts is
calculated from
(4)
where s' = clear vertical spacing between belts; and d, = diameter of column.
The effective area of concrete core enclosed by the belt is given by
(5)
where Pee = ratio of area of longitudinal reinforcement to area of core section;
and Ac = area of concrete enclosed by composite belt. From equations 2, 4 and
5 the confinement effectiveness coefficient for circular cross sections can be
obtained as
(6)
A similar approach was used to derive a relationship for the confinement
effectiveness coefficient for rectangular cross sections given by
(7)
where b and h are the width and height of the rectangle.
The stress-strain relationship of the composite belt was assumed to be
linear elastic to failure. Two types of composite belts were used in the
parametric study: E-glass fiber reinforced and carbon fiber reinforced belts.
FRP Reinforcement
255
The mechanical properties of these belts were obtained from the data provided
by the manufacturer [4]. The modulus of elasticity and tensile strength of glass
and carbon fiber belts were 48.2 GPa (7,000 ksi), 1103 MPa (160 ksi); and 172
GPa (25,000 ksi) and 2862 MPa (415 ksi), respectively.
Figures 4a and 4b show the stress strain curves of unconfined concrete
and concrete confined with 5 mm (0.2 in), 10 mm (0.4 in) and 15 mm (0.6 in)
thick composite belts made of E-glass and carbon fibers, respectively. The
stress-strain curves shown were developed according to the models discussed
above for a 1524 mm (60 in) diameter circular column.
The analytical models described above were used to quantify the gain in
strength and ductility of retrofitted rectangular and circular columns.
An incremental deformation technique and the strain compatibility
method were used to generate the axial load-moment-curvature relationships for
circular and rectangular cross sections. Figure 5 shows the cross section of the
circular column. The unconfined concrete compressive strength was 34.4 MPa
(5,000 psi); Grade 60 steel rebars were used as longitudinal reinforcement. The
column was fully confined with 10 mm (0.4 in) thick composite belts; in other
words, the vertical spacing between belts was zero.
Figure 6 shows the axial load-moment-curvature relationship as well as
the ductility factor given by, 4>J4>Y' versus axial load ratio, P/Po> where 4>u =
curvature at failure; tPy = curvature at first yield of longitudinal bars, P = axial
load, and Po - ultimate axial load of unconfined column. For comparison, the
diagrams for the column before strengthening, in other words, unconfined
column is shown with dotted lines. The curves for the retrofitted column are
shown for two types of belts. The solid line represents the curves for carbon
fiber belts and the dashed lines represents the curves for E-glass belts. As can
be seen from Figures 6(a) and (b), the strength and ductility of the column have
significantly increased compared with the values before strengthening. The
carbon fiber belt has resulted in higher strength and ductility gain as compared
with values for theE-glass belt. Carbon fiber belts, however, are significantly
more expensive than E-glass belts and therefore might not be very economical
for this type of application.
Figures 7(a) and (b) show the relationship between the ductility factor,
4>J4>Y' and thickness of the strap for both types of belts and three different
spacings between the belts; that is, s' = 0 mm, s' = 152 mm (6 in), and s'
= 304 mm (12 in). As can be seen from these figures, the ductility factor
increases as the thickness increases, however, the rate of this increase, indicated
by the slope of the curves, decreases as the spacing between the belts increases.
Figure 8 shows the cross section and the reinforcement details of the
rectangular column used in this study. The concrete compressive strength was
34.4 MPa (5,000 psi). The thickness of the composite belt was 10 mm (0.4 in)
256
Saadatmanesh, Ehsani, and Li
and it fully confined the column; in other words, the vertical spacing between
the straps was zero. Figures 9(a) and 9(b) show the axial load-momentcurvature diagrams and the ductility factor versus axial load ratio for unconfined
column and column confined with E-glass and carbon fiber belts, respectively.
The curves belonging to unconfined column are shown with dotted lines; the
curves belonging to column retrofitted with E-glass and carbon fiber belts are
shown with dashed and solid lines, respectively. Similar behavior, like that for
circular column, was observed for this column. Significant increases in
strength, and ductility indicated by large curvature at failure, were observed for
the retrofitted column as compared with the ductility and strength of the
unconfined column. As was the case for the circular column, carbon fiber belts
resulted in greater increases in ductility and strength compared with the values
for the column strengthened with E-glass belts.
ONGOING EXPERIMENTAL STUDY
A comprehensive testing program of concrete columns retrofitted with
fiber composite belts is currently underway at the University of Arizona. A
number of 1,4-scale models of prototype circular and rectangular columns will
be tested. The columns will be retrofitted with E-glass belts. In addition to the
retrofitted columns, control column specimens will be tested before retrofitting.
The test results of the retrofitted columns and control column specimens will be
compared to experimentally verify the gain in strength and ductility of retrofitted
columns. All columns will be tested under a constant axial load and reversed
inelastic cyclic loads. The hysteresis loops of the retrofitted columns will be
compared with that of control, unretrofitted column specimens to quantify the
gain in strength and ductility.
Figure 10 shows the dimensions and
reinforcement details of the circular columns to be retrofitted and tested.
A unique testing system was also developed by the authors to examine
an effective coupling technique for the composite belts. The system included
a 304 mm (12 in) high concrete column stub configured in two semi circle with
an opening inside to allow placement of a jack. Figure 11 shows a schematic
of the testing system. The two semi circles are placed against one another to
form a complete circle with an opening inside. The composite belt is wrapped
around the column stub and is either mechanically or chemically coupled. A
hydraulic jack is then placed inside the opening to allow introduction of an
expansive force inside the column core to simulate the dilation of the concrete
core under actual loading conditions. Figure 12 shows a picture of the column
stub used in the experimental study.
Various coupling techniques including epoxy bonding and mechanical
coupling of the belt's ends have been experimentally examined. Both systems
performed successfully with no failure in the end connections until the failure
of the belt itself. Figure 13 shows a picture of the mechanical coupler which
FRP Reinforcement
257
was successfully used in the study. The lower pin in the buckle will be placed
through loops at the ends of the belts to connect the belt ends to the buckle.
The top pins shown on the buckle allow free rotation and therefore elimination
of any moment that could be developed in the buckle during tensioning. The
high-strength steel rebar shown connects the two ends of the buckle together.
Tensioning of this rebar will allow tensioning of the belts during strengthening.
CONCLUSIONS
The analytical and experimental studies performed thus far on the
behavior of concrete columns strengthened with fiber composite belts indicate
that this strengthening method can be used to effectively and economically
increase the strength and ductility of concrete columns in the potential plastic
hinge regions or over the entire height of the column.
The stress-strain models for concrete confined with fiber composite belts
for both circular and rectangular columns indicate significant increases in the
stress and strain at failure and therefore the overall ductility; although E-glass
has a larger elongation at failure than carbon fiber, carbon fiber has a larger
energy absorbing capacity indicated by its larger area under the stress strain
curve; and the ductility factor increases linearly with the increase in thickness
of the belts, but the rate of increase in ductility factor decreases as the spacing
between the belts increases.
ACKNOWLEDGEMENT
This study was sponsored by the National Science Foundation Grant NO.
MSS-9022667, Dr. John B. Scalzi program director. The support of the
National Science Foundation is Greatly appreciated.
REFERENCES
1.
2.
3.
4.
Mander, J.B., M.J.N. Priestly and R. Park, "Seismic Design of Bridge
Piers," Research Report No. 84-2, 1984, University of Canterbury, New
Zealand.
Popovics, S., "A Numerical Approach to the Complete Stress-Strain
Curves for Concrete," Cement and Concrete Research, Vol. 3, No. 5,
1973, pp. 583-599.
Mander, J.B., M.J.N. Priestley and R. Park, "Theoretical Stress-Strain
Model for Confined Concrete," Journal of Structural Engineering,
ASCE, Vol. 114, No.8, August 1988, pp.1804-1826.
"FIBERITE" Materials Handbook, Imperial Chemical Company (ICI),
March, 1989.
258
Saadatmanesh, Ehsani, and Li
Fig. 1-Concrete columns strengthened with fiber composite belts
Confined
Concrete
~
I
f'c c
Strap
Fracture
u)
CJ)
~
(f)
Unconfined
Concrete
f'c o
£co
2£co
£cc
£cu
Strain, £c
Fig. 2-Stress-strain model proposed for unconfined and confined concrete
FRP Reinforcement
Strap
ds
fu s
fu s
Fig. 3-Confining action of strap
UnconrLned
.. s·;,;;,; .....
--,_--.-iii-,;,.---
····.:··~··
1~0.0
t:""='j'5"";;,'-
'0'
120.0
:!
110.0
60.0
::::--·
e:·-·
30.0
0.0-f----....,.-----r-----,
0.000
0.00!1
0.010
0.015
Stroin, £,
(o)
1!10.0
'0'
120.0
~
90.0
a.
..:
~·-
60.0
.~-·
--
JO.O
0.0-f----....,.-----r-----,
0.000
0.005
0.010
0.015
Strain,£,
(b)
Fig. 4-Stress-strain model of unconfined and confined
concrete for circular column (a) E-glass (b) carbon fiber
259
260
Saadatmanesh, Ehsani, and Li
50.!1 mm Covor
Ao " -16,-164 mm z
Fig. 5-Cross section of circular column
~ no
.
,•' =·· ········· ...
' 4l
,'' ...:
.,,.'
0.010
0.056
0.042
0.1111
.·
u
0.014
QJlVAl\JRE (ym)
Ill
20.4
l7l
34.0
'U
IIOIOT (kN-m)
(a)
50.0
.(
~
40.0
oi
~
u
JO.O
~
>....
20.0
i=
10.0
.
::J
u
:::l
0
0.0
0.0
0.4
0.8
1.2
1.6
2.0
P/Po
(b)
Fig. 6-Axial load-moment cmvature diagram and ductility
factors versus axial load ratio of rectangular column
FRP Reinforcement
261
.. 0.0..............................
40.0
• - 152.1 --.-.-:
·:ro1:a -.;.--
JO.O
.(
t}
20.0
-:;....·-----
10.0
::;.......~-:..:.::;.·:.-
0.0
0.0
4.0
8.0
12.0
16.0
20.0
16.0
20.0
t (mm)
(a)
,
0.0-.
......·..................
• - 152.1 -
40.0
-·;·-: ·3a1:a-.;.·-
JO.O
.(
t}
20.0
......
.... ··
~
10.0
0.0
0.0
4.0
8.0
12.0
t (mm)
(b)
Fig. 7-Ductility versus thickness for circular column ifco' = 34.45 MPa
(5,000 psi), P/P0 = 0.1) (a) E-glass (b) carbon fiber
h • 1,029 mm
E
.,E
"!.
•
a
Sllrrup
50.8 mm Cover
Fig. 8-Cross section of rectangular column
262
Saadatmanesh, Ehsani, and Li
0.110
0.0118
O.Mii
0.044
0.1122
2.1
ll
WNA!UlE (\h'n)
J.l
1.8
6.0
•ld
UOMENT {lcN-m)
(a)
100.0
~
it
ri
0
80.0
Unconf~ned
~T:9~~~~~~~~~
60.0
Cor bon
1-
u
L!
40.0
>-
1-
:::1
t
20.0
0
0.0
:::l
·-0.0
0.4
0.8
1.2
1.6
2.0
P/Po
(b)
Fig. 9-Axial load - Moment-curvature diagram and ductility
factor versus axial load ratio of rectangular column
FRP Reinforcement
263
Reinforcement Details
For Column Specimen
....
Vertical
"~
-
Load
=
f-.....li·
f--Horizo ntal
....CD
1--4
No.4
No.4
1--14
~
.....
'
0
II
.,p,
..:l
-
4
I'
·.... "'\.
9
@
Hoops
@
H
Ce n te r -
.__.No.6
42"
3. 0"
Center
len gi tudinal
Bars
No.6 r1 Bars
Plus No.6
Straight Bars
'
~
@
Hoops
Gage Wir e Hoops
3.5" Cen ter
/
llo
No.4
Hoops
No.4
/
_L ~
~
Squa re
Load
-
.
1"-No.6 L.J
Bars
-I
Fig. 10-Reinforcement details of test columns
Bars
264
Saadatmanesh, Ehsani, and Li
TEST
FRP
SETUP
BELT
~
=
TEST
I
T ABLE
"I
.,,,,.,
'\
~
I
,,,,,
STRONG
Fig. 11-Column stub for testing of composite belts
FLOOR
FRP Reinforcement
Fig. 12-Column stub with composite belt
Fig. 13-Mechanical lock for buckling of composite belts
265
SP 138-17
Flexural Behavior of
Cementitious Composites
Reinforced by Pitch-Based
High Modulus
Continuous Carbon Fibers
by T. Yamada, K. Yamada,
and K. Kubomura
Synopsis:
The Potential of using pitch-based high modulus carbon
fiber was investigated as a reinforcement in
cementitious composites for structural reinforced
concrete (RC) members. For this purpose, effects of
carbon fiber mechanical properties on the mechanical
properties of carbon fiber reinforced cementitious
composites were studied through the three-point
flexural test by using several pitch-based high modulus
carbon fiber rods of varying fiber moduli and
strengths. For the specimens with a fiber volume
fraction larger than the critical volume fraction, the
flexural strength is found to exceed the mortar matrix
strength and is linearly proportional to the sum of all
rod strengths, and the flexural modulus after matrix
cracking is found to also be linearly proportional to
the sum of all rod stiffnesses.
Keywords: Carbon; composite materials; fibers; flexural strength; flexural
tests; mechanical properties; reinforced concrete
267
268
Yamada, Yamada, and Kubomura
Takashi Yamada received his BSc in chemical engineering
from Keio University in 1986. He is a researcher with
Chemicals of Nippon Steel Corporation. His research
interest covers carbon fiber surface treatment and
carbon fiber reinforced concrete.
Kanji Yamada received his BSc in structural engineering
from Fukui University in 1970. He is a senior
researcher with Chemicals of Nippon Steel Corporation.
His research interest is in mechanical properties of
carbon fiber reinforced concrete.
Kenji Kubomura received his PhD in structural dynamics
from the Dept. of Aero. & Astro. of Massachusetts
Institute of Technology in 1978. He is responsible to
the development of pitch-based carbon fiber and its
applications at Chemicals of Nippon Steel.
INTRODUCTION
In the early 1970's, carbon fibers have found large
volume applications in sporting goods such as golf
shafts and fishing rods. At the some time, much of the
theoretical work on polymeric composites was carried
out in the aerospace industry to exploit high specific
strength and modulus for aerospace applications.
Presently, close to 10,000 tons of PAN based carbon
fiber is produced annually for various applications
which could not be conceived of at the time of
invention in 1960.
Around 1972, the effects of carbon fiber reinforcement
on cementitious composites were reported[1,2,3] and the
potential of using it with cementitious materials
became apparent. Carbon fiber reinforced cement(CFRC)
is a composite material possessing excellent properties
such as high specific strength, high specific modulus,
and good corrosion resistance. Mechanical properties of
CFRC are dependent on the properties of carbon fibers,
volume fraction of the fibers, fiber reinforcing
methods, and interfacial bond properties between the
fibers and cementitious matrix.
Initially in 1986, carbon fibers in short staple form
were applied to curtain wall panels of a high rise
building in Japan, and since then they have been used
in many other buildings. Also, structural elements such
as precast curtain wall panels and floor panels were
fabricated using short staple carbon fibers. When a
FRP Reinforcement
269
small amount of fiber is used as reinforcement, CFRC
has significantly higher flexural strength and
toughness than plain cementitious materials [1-5].
In reinforced concrete members, steel bars and meshes
are used to get higher flexural strength and toughness
after matrix cracking. Also, there is a strong
potential of using carbon fiber reinforced bars and
meshes in reinforced concrete members. Several
investigations were carried out to understand the
effects of replacing steels by carbon fibers[6,7].
However, the effects of fiber property differences are
not fully investigated in these works, although many
carbon fibers of different properties have been
developed recently. Since fiber strength and modulus
may significantly alter the post-cracking performance
of RC members, the investigation of their effects is
necessary to make better use of the fibers.
High modulus carbon fibers are less expensively
manufactured from mesophased pitch by orienting
graphite crystallites along the fiber axis by simply
rising the heat-treatment temperature above those of
the polyacryilnitrile(PAN)-precursor. Using different
fibers, modulus and strength of the fiber reinforced
plastic bars are made different from those of steel
bars, dependent on the fiber tensile modulus, strength
and volume fraction. When these bars are used in the
reinforced concrete members, the post-cracking
performance could be much different from that of steel
bars.
In the present work, several pitch-based carbon fibers
of varying tensile moduli and strengths were used to
study the effects of modulus and strength of carbon
fiber reinforced bars on
the flexural performance of
CFRC of low reinforcement ratios through three-point
bending tests. Since the properties of carbon fiber
reinforced bars influence strongly the CFRC performance
after cementitious matrix cracking, the relations
between the fiber properties and flexural properties of
CFRC on post-cracking were investigated.
One of the most important parameters controlling the
CFRC performance is the bonding between the carbon
fiber rods and cement matrix .. However, to focus our
investigation on the fiber property differences, one
surface treatment of the rod surface was chosen and
applied to all the rods used in the tests, which was
developed in our laboratory with epoxy emulsion
containing colloidal Sio 2 particles (10 nm). This
demonstrated good bonding[8,9].
270
Yamada, Yamada, and Kubomura
Experimental
Test
Specimens
Four mesophase pitch-based carbon fibers (3, 000
filaments) of four different tensile moduli (A, B, C, D)
were used in the tests whose tensile properties are
shown in Table 1. All fibers were oxidation treated and
unsized. From these carbon fibers, unidirectional CFRP
rods were fabricated to have a fiber volume fraction of
about 50 percent and a diameter of 1. 1 mm. Their
tensile properties are also shown in Table 1. The
matrix was a water-in-oil type epoxy emulsion
containing fine Sio 2 (10 nm), which was developed in
our laboratory. This epoxy emulsion is a mixture of
EPICOUTE 828, a hardening agent (amine), and some
surface-active agents.
Unidirectional CFRC flexural test specimens of 25 mm by
40 mm cross section were fabricated by placing from one
to five rods, with equal spacing, at the lowest part of
each specimen with a covering depth of from 0.02 mm to
0. 2 mm as shown in Figure 1. The covering depth was
initially targeted at 0. 02 mm, but it was scattered
between 0.02 to 0.2 mm due to fabrication difficulties.
In the present experiment, the volume fraction is
defined as a value of the cross sectional area of all
the continuous carbon fibers used in the rods divided
by the cross-sectional area of the specimen. Thus, the
volume fractions of continuous fibers are from 0.021 to
0.126 percent depending upon the number of rods placed.
In Table 2, the cementitious matrix composition of test
specimens is shown. The basic mix ingredients in CFRC
were ordinary portland cement, fine aggregates with
particle size ranging from 4 to 32 microns, ceramic
spheres as lightweight aggregates with particle size
ranging from 24 to 96 microns, superplasticizer, and
antifoaming agent. This matrix was mixed with a volume
fraction of 2 percent of short staples of isotropic
carbon fiber G of tensile modulus 49 GPa, tensile
strength 960 MPa and specific gravity 1. 66. Physical
and chemical properties of the cementitious matrix are
shown in Table 3. All specimens were cured in water as
usual for 28 days.
Three-Point
Flexural
Test
Analysis
Figure 1 shows the dimensions of the unidirectional
CFRC test specimen (25x40x500 mm) . The specimens were
tested by three point loading on a span of 400 mm, with
FRP Reinforcement
271
a displacement transducer attached to each specimen at
the load-point to measure the mid-span deflections.
Flexural loading was displacement-controlled with a
quasi-static deflection rate of 1/400 times the span
length per minute in the standard room temperature.
Results
and
Discussion
To characterize the mechanical properties of CFRC
reinforced with CFRP rods, maximum flexural strengths
and flexural moduli after the matrix cracking, and the
deflections associated with the maximum load were
analyzed.
Flexural
Strength
For case where the specimen fails in the fiber tensile
fracture mode, the bending moment at failure, Mf, can
be given by the following equation;
T
( 1)
where T is the tensile strength of total CFRP rods
placed in the specimen, which is the product of the
cross-sectional area of all the rods placed in the
specimen, the fiber volume fraction in the rods and the
fiber tensile strength. The distance, lc, which is a
constant for each type of specimen, is the distance
between the center of the rod and the place at which an
equivalent compressive load equal to T is applied. The
assumption that equation (1) can hold for CFRP rod
reinforced concrete is drawn from the fact [10] that
equation ( 1) holds for steel reinforced concrete when
it fails in the steel tensile mode.
For the case where the amount of reinforcement fiber is
smaller, the failure moment defined by equation (1)
becomes less than the moment, Mm, at which the matrix
cracking occurs. In Figure 2, this is illustrated,
where for volume fractions less than the .critical
volume fraction, Vf(crit)' the maximum moments which
are associated with matrix cracking are constant, and
for volume fractions greater than Vf(crit)' the maximum
moments which are associated with fiber tensile failure
are given by equation (1). Vf (crit) is defined at the
intersection of straight lines of equation (1) and of
the matrix cracking moment, Mm.
272
Yamada, Yamada, and Kubomura
In Figure 3, typical flexural moment-deflection curves,
and K, are shown for specimens reinforced by the
fiber D of a volume fraction much smaller than
Vf (crit) (0. 02 percent), and much greater than
Vf(crit) (0.11 percent), respectively. In general,
specimens carry a lesser or greater load after matrix
cracking depending on the amount of reinforcement
fiber. When the reinforcement fiber volume fraction is
smaller than Vf(crit)' first, matrix cracking occurs at
the first point X, and then fiber failure occurs at
point Y with a smaller moment than that at the matrix
cracking. On the other hand, when the reinforcement
fiber volume fraction is larger than Vf (crit), first,
the matrix cracking occurs at point X', and then, the
fiber failure follows at point Y' with a larger moment
than that at the matrix cracking. In order that the
specimen flexural failure moment exceeds the matrix
flexural failure moment, the fiber volume fraction
needs to be larger than Vf(crit).
J
Analogous to the flexural strength of steel bars, the
flexural strength, crc, is defined by the following
relation:
Me
3Pl
a=-=-c
z
2bh 2
(2)
where Me is the maximum bending moment, Z is the
section modulus, P is the maximum applied load, 1 is
the span length, b is the specimen width and h is the
specimen thickness.
In Figures 4, 5, 6 and 7, flexural strengths are
calculated by equation (2) and plotted against fiber
volume fractions for specimens reinforced with fiber A,
B, C and D rods, respectively.
In each of Figures 4, 5, 6 and 7, a linear line is
drawn from the origin
to flexural strengths at fiber
volume fractions higher than about 0.08 percent by the
least square method.
In Table 4, critical volume fractions Vf (crit), are
obtained as the fiber volume fraction at the
intersection of the linear line and the matrix flexural
strength(11.4 MPa), and lc's are calculated from the
slope of the linear line and T through equation 3, for
specimens, A, B, C and D and shown in the same table.
FRP Reinforcement
273
The lc is given by the following equation;
a
·b·h
=----'-c_
__
1
c
2
6T
(3}
where, crc and Tare given by Figures 4, 5, 6 and 7. For
the specimens tested in the present study, the values
of lc fall in the range of 80 to 95 percent of the
specimen thickness. It should be noted that the
flexural strength of specimens having a higher volume
fraction
(greater than 0.12 percent)
could be
calculated by the lc values shown in the table.
Flexural strengths of the majority of the specimens
having a volume fraction less than Vf(crit) are higher
than the matrix strength for the fibers A and B, and
are lower for the fibers C and D. It is thought that
the flexural strength below Vf(crit) is larger than the
matrix flexural strength. However, there are several
data contradicting this. This may be associated with
fabrication problems.
In Figure 8, the flexural strengths and the total rod
strengths, T, are plotted with open squares for the
specimens having a volume fraction less than Vf(crit)
and
with solid squares for the specimens having a
volume fraction greater than vf(crit).
Then, a straight line is drawn from the origin to solid
squares with the least square method. For the values of
T greater than 2000 Newtons, where the matrix flexural
strength and the straight line intersect, solid squares
cluster to the line close enough for stating that the
flexural strength and the total rod strength are
related through equation (1).
Flexural
Modulus
In Figure 9, flexural moduli after the matrix cracking
are plotted against volume fractions for specimens
reinforced with the fibers A and D of the fiber volume
fractions 0.043 through 0.126 percents. The flexural
modulus, E, is calculated by the following equation;
lla·l 2
E=---
6h. llc5
( 4)
where llcr is the flexural stress increment for the
deflection increment, Ll0(=1.5 mm), between deflections
274
Yamada, Yamada, and Kubomura
1.5 mm and 3.0 mm where the tensile load across the
cracks is thought to be carried by the rods. The
specimens reinforced with the fiber D (solid squares)
show a higher flexural modulus than that of the
specimen reinforced with the fiber A(open squares) when
they are compared at the same volume fraction. This is
because (1) the lc's of all specimens are close to each
other, which suggests that micro crack openings or
crack lengths are close regardless of types of
reinforcing fibers for a given deflection, (2) the
modulus of the uncracked part is not much influenced by
the amount of reinforcing fibers because the amount of
fibers is too small, and (3) the tensile load is
transmitted across the cracks by only the fibers.
Therefore, the modulus difference between different
reinforcing fibers after the matrix cracking is
associated with the fibers tensile modulus.
Also, in Figure 9, straight lines, A, B, C and D are
drawn for specimens reinforced with the fibers, A, B, C
and D, respectively by the linear curve fit for each
group of data, although data points for fibers B and C
are not included in the figure.
For fiber volume
fractions larger than 0.06 percent in the figure, the
flexural modulus increases with the increasing fiber
modulus, and for those less than 0. 06 percent, the
flexural modulus dependence on the fiber modulus is not
apparent. It is interesting to note that the flexural
modulus dependence on the fiber modulus is apparent in
the range of fiber volume fraction where the equation
(1) holds clearly.
Though all moment deflection curves are not shown in
this paper, several curves for specimens of the volume
fraction less than 0.03 percent did not show any load
increase after the matrix cracking. When the fiber
volume fraction exceeded 0.03 percent, all the curves
showed a range of increasing load for increasing
deflection after the matrix cracking.
In Figure 10, the flexural moduli after matrix cracking
and the total rod stiffnesses, which are defined as the
product of the cross-sectional area of all the rods
placed in the specimen, the fiber volume fraction in
the rod and the fiber tensile modulus, are plotted
along with a straight line
drawn by the linear curve
fit. It appears that the flexural modulus after the
matrix cracking is linearly related to the total rod
stiffness. This linear relation between the flexural
modulus and the total rod stiffness after the matrix
cracking is equivalent to the linear relation between
the flexural strength and total rod strength.
FRP Reinforcement
Allowable
275
Deflection
In Figure 11, flexural stress versus deflection curves
are shown for specimens of the fibers B, C, and D, with
fiber volume fractions, 0.111, 0.111 and 0.106 percent,
respectively, along with a curve for the specimen
without reinforcing rod. Also, the largest crack widths
are shown up to 0.2 rnrn which is the generally accepted
allowable maximum crack in reinforced cementitious
matrix, which is almost independent from the type of
reinforcing fibers. It should be noted that since
cracks are detected by eyes and measured by loupe(xlO),
the measurement confidence is relatively low. In Figure
12, typical crack patterns are also shown for the
fractured specimens of the fibers B, C, and D, with
fiber volume fractions, 0.111, 0.111 and 0.106 percent,
respectively, along with the distances between cracks.
It appears that the cracks are spaced almost equally
both for specimens of fibers B and C, although the
numbers of cracks are different(four cracks for the
fiber B specimen and seven cracks for the fiber C
specimen) . It is also interesting to note that the
number of cracks increases as the failure deflection
increases.
To investigate relations between fiber tensile modulus
and flexural deflection at the maximum load,
deflections at the maximum load are plotted against
volume fractions for the fiber B and D specimens in
Figure 13 because the fiber strengths of these fibers
are almost the same, along with those values for the
matrix. The flexural deflections of the fiber B
specimens for fiber volume fractions larger than 0.08,
which is close to the critical volume fractions of
these specimens, are larger than those of the fiber D
specimens.
Discussion
It appears that the flexural strength of CFRC specimens
is associated with the reinforcing rod total strength,
T, which is the product of the cross-sectional area of
all rods placed in the specimen, the fiber volume
fraction in the rod and the fiber tensile strength,
when the volume fraction is greater than specimen's
critical volume fraction Vf(crit). Once lc is obtained
from equation (1), the maximum flexural strength of the
CFRC, Mf, could be calculated for the specimen with the
given fiber strength and the amount of fiber larger
than the critical fiber volume fraction.
276
Yamada, Yamada, and Kubomura
If the fiber and matrix bonding is weak and the rods
are pulled out after the matrix cracking, then it is
conceivable that lc, which is somehow related to the
distance between the center of rod and the neutral
axis, is altered. Among the present experiments, in the
specimens of the fibers B and C shown in Figure 10,
several small cracks were observed after failure. This
suggests that the tensile strain in the matrix is
released by these small cracks because the rod matrix
bonding is probably strong enough. If the bonding is
weak, then the tensile strain can be released by the
rod pull out and further opening of the initial crack,
conceivably resulting in a changing lc. Though an
experiment was not conducted, it can also be thought
that the matrix compressive failure could make equation
(1) invalid when the flexural strength, Mf, becomes
large and the compressive stress exceeds the matrix
compressive strength.
From Table 4, it appears that the lc 's are almost
constant independent of fiber type. When the lc is
independent, Mf is a linear function of the total rod
tensile strength with the slope being lc regardless of
fiber type. Figure 8 supports this argument. The
strength of total reinforcing rods(the total rod
strength), T, is expressed by the following equation;
(5)
where A is the cross section area of all the rods, Sf
is the reinforcing fiber strength, Vf is the fiber
volume fraction and k is the fiber strength translation
factor. From Figure 8, the critical fiber strength
T (crit) can be defined at the intersection of the
linear line in the figure and the matrix flexural
strength as shown in the figure. Comparing this figure
with Figures 4, 5, 6 and 7, the critical fiber volume
fraction Vf(crit) is related to the critical total rods
strength, T(crit)' by the following relation;
T(crit)
=
k • A • Sf • Vf(crit)
When the reinforcement fiber tensile strength,
increases, the vf(crit) decreases.
(6)
Sf,
The flexural modulus after matrix cracking is linearly
related to the total rod stiffness after matrix
cracking. In the deflection range(l.5 mm to 3 mm) where
FRP Reinforcement
277
all flexural moduli were calculated, only one large
crack opening was observed for each specimen and the
opening widths at a specific deflection in this range
were almost the same regardless of fiber type.
Therefore, the length of the rods bridging the crack
opening is independent from types of reinforcing
fibers. Further, the matrix and rod bonding strength
can be considered almost independent also from fiber
type because all rods were fabricated from the same
resin system and the surface was completely covered
with the same matrix resin. This same bonding strength
among different fibers implies that the length of rod
de-bonding at the bridging is also independent from
fiber type. This leads to the same length of the rod
detaching at the crack. If this is the case, then, to
get the same deflection, the detached part of the rod
must be elongated with the same amount, resulting in a
linear relation between the deflection and the total
rod stiffness because lc is almost independent from
the reinforcing fiber type. This relation is shown in
Figure 10.
Though it is expected that the failure flexural
deflection increases for fiber volume fractions larger
than Vf(crit)' when the fiber failure strain increases,
the relation among the fibers, A, B, C and D are not
so. It is possible that the experiments were conducted
for relatively low volume fractions and the validity of
equation ( 1) is somehow obscure in these volume
fraction ranges.
For steel reinforced concrete the Japanese RC
standard [ 11] states that the steel volume fraction
must be greater than or equal to 0.2 percent. Similar
to this standard, carbon fiber reinforced concrete
must have a critical volume fraction, Vf, which is
defined by equation ( 6) . For the present pitch based
carbon fibers, the value is about 0.08.
Conclusions
Replicated tests were conducted on cementitious
composites reinforced by mesophase pitch-based high
modulus continuous carbon fibers to study the relations
among the fiber tensile modulus and strength and the
specimen flexural modulus, strength and deflection. The
results indicate the following:
1. The flexural strength of the specimens is linearly
related to the total rods strength above a critical
278
Yamada, Yamada, and Kubomura
volume fraction.
2. The flexural modulus of the specimens after matrix
cracking is related to the product of the total rod
stiffness.
3. For fibers with the same tensile strength, the
failure flexural deflections may increase when the
fiber modulus is decreased for fiber volume fractions
larger than the critical finer volume fraction.
References
1. Ali, M.A., Majumdar, A.J.,and Rayment, D.L., Carbon
Fiber Reinforcement of Cement. Cement and Concrete
Research, vol.2, No.2, 201-212pp. 1972
2. Waller, J.A. Carbon Fiber Reinforcement of Cement
Composites. ACI Publication sp44-8. Fiber reinforced
concrete, pp.l43-161. 1974
3. Sarkar, S., and Bailey, M. B. Structural properties
of carbon fiber reinforced cement. RILEM Symposium.
pp . 3 61 - 3 7 1 . 1 9 7 5
4. Larson, B.K.,Drzal, L.T., and Sorousian, P. Carbon
fiber-cement adhesion in carbon fiber reinforced cement
composites. Composites. vol.21,No.3, pp.205-215. 1990
5.
Pariz S.,
Mohamad N.,
and Abdulrahman A.
Statistical Variations in the Mechanical Properties of
Carbon Fiber Reinforced Cement Composites. ACI
Materials Journal, March-April 1992
6. Nakagawa, H., Suenaga, T., Akihama, S. Mechanical
Properties of Three-Dimensional Fabric Reinforced
Concretes and Their Application to Building. Proc. 1st
Japan International SAMPE Symposium. pp.l587-1592. 1989
7. Ohno, S., Yonezawa, T., Inoue, I. Mechanical
Properties of Cement Composites Reinforced with Newly
Developed 3-Demensional Fabric Carbon Fibers. Summaries
of Technical Papers of Annual Meeting Architectural
Institute of Japan. pp.697-698. 1991
8. USP 4,902,537
9. Yamada, T., Yamada, K., Hayashi, R., Herai, T.
Adhesion at the Interface between Carbon Fiber and
Cementitious Matrix,
36th International
SAMPE
Symposium, pp.362-371, 1991 April
FRP Reinforcement
279
10. Edited by Architectural Institute of Japan
for Ultimate Strength Design
Structures", pp.36-39, 1987
of Reinforced
"Data
Concrete
11. Japan Society of Civil Engineers "Design Standard
of Reinforced Concrete Structures",
pp.47-48, 1986
TABLE 1 -TENSILE PROPERTIES OF CARBON FIBERS
Pitch Based CF
A
B
c
D
Fiber Tensile Modulus, GPa
Fiber Tensile Strength, MPa
Fiber Elongation, %
Density, g/cm 3
201
2832
1.41
1.95
298
3518
1.18
2.03
383
3675
0.96
2.03
492
3469
0.70
2.10
Rod Tensile Modulus, GPa
Rod Tensile Strength, MPa
Rod Elongation, %
Rod Diamter, mm
65.0
778.9
1.20
1.0
84.7
849.5
1.00
1.0
108.3
883.4
0.82
1.0
134.1
803.8
0.60
1.0
TABLE 2 - CEMENTITIOUS MATRIX COMBINATION
Portoland Cement
Fine Aggregate
55%
II%
Lightweight Aggregate Isotropic Carbon Fibre G
1.8%
4.3%
Water
27.4%
TABLE 3 - MECHANICAL AND CHEMICAL PROPERTIES
OF CEMENTITIOUS MATRIX
Olemical of
portland cement,
Si02
CaO
~rcent
21.4
60.8
Flexural Strength ,MPa
Fe20 3 MgO
2.86
2.17
Al203
5.84
Flexural Modulus ,GPa
Density ,ycm 3
13.73
1.68
Strain at Failure,xlO -6
Mechanical
11.4
2000
280
Yamada, Yamada, and Kubomura
TABLE 4 -
Vf(crit)
AND
OF THE UNIDIRECTIONAL CFRC
lc
Carlxm Fibers
Vf (critical), %
lc, mm
A
B
c
D
~.~~2
22.~
~.06~
~.064
2~.~
~.~fil
212
2Jj
CF-ROD
t ______ j_o o2-o 2
·r· . nun
/l
.;..____ 40 nun -----..l. ./
~H
I
I
:d/1
l.Onun/min
r/r
-._•
I
...:.-d.:-:
:
: d/2
:~d
p
14----------------·· 400 mm --------------.....
Fig. !-Dimensions of unidirectional CFRC test specimens
(25 mm x 40 mm x 500 mm)
FRP Reinforcement
281
cQ):
E
0
:E
v 1< vf(crit)
v t > vf(crit)
Vf( crit)
Volume Fraction
Fig. 2-Illustration of relation between flexural moment and volume fraction
12
10
.
E
(.)
8
z
-
~
r:::
Cl)
E
0
:E
6
4
2
0
0
2
3
4
5
Deflection , mm
Fig. 3-Flexural moment versus deflection curves (fiber D)
282
Yamada, Yamada, and Kubomura
24
a.
20
-...
-...
16
<U
:!!
J:
Cl
c:::
12
en
8
Q)
l-Y
....
1:1
----- ---------c
><
Q)
u:::
4
v. v
0
0.00
0.02
[2
/
!>"'"----
V'
----- ------
/
cti
::l
c
c
c
/
0.04
0.06 0.08
0.10
0.12
0.14
Volume Fraction,%
Fig. 4-Flexural strength versus volume fraction (fiber A)
24
<U
20
:!!
s:£
16
a.
-...
-...
Cl
c:::
12
en
8
Q)
c
p
----- -----
cti
::l
><
u:::
Q)
4
v
0
0.00
y
v
jV
c
----- ----- ----- ------y v
c
./
0.02
0.04 0.06
0.08
0.10
0.12
0.14
Volume Fraction, %
Fig. 5-Flexural strength versus volume fraction (fiber B)
FRP Reinforcement
24
[]
cu
c..
~
./ ~
-
16
8
9
"b
><
u.
4
/
0.02
0.04
.r:
Cl
c:
Q)
....
en
«i
....
:I
12
Q)
""
/
20
----- -----
L
0
0.00
[]/
~/
~----
,;a
.............. ...............
------
[]
0.06
0.08
0.10
0.12
0.14
Volume Fraction, %
Fig. 6-Fiexural strength versus volume fraction (fiber C)
24
./~
cu
c..
~
-
.r:.
Cl
c:
20
16
-
12
:I
4
Q)
....
en
~
----- ----
/[]~
8
«i
....
><
Q)
u:::
:~
[]
/0
~
----- ---- ...
..................
/.
0
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Volume Fraction,%
Fig. 7-Fiexural strength versus volume fraction (fiber D)
283
284
Yamada, Yamada, and Kubomura
1000
0
2000
3000
4000
5000
T,N
Fig. 8-Fiexural strength versus total rod strength
0
~._~~~~~~~~~~~~~_J
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Volume Fraction, %
Fig. 9-Fiexural modulus versus volume fraction
FRP Reinforcement
285
5
...
4
y
as
Q.
"
ui
/
3
'0
0
..... " ll
[
::::J
::::J
:E
a;
/
c I
2
A
~k
c
...
c
::::J
><
....
Q)
iL
0
~
0
t
/ll
~
...
c
c
c
c
200
100
300
400
500
600
Total Rod Stiffness, kN
Fig. 10-Flexural modulus versus total rod stiffness
CRACK O.lOmm
30
TRACK 0.20mm
'''
20
''
'
'''
FiberD
0.111%
10
0
~~~~~~~~~~------~~~----~
0
2
4
6
8
Deflection, mm
Fig. 11-Flexural stress-deflection curves (fiber B, C, D, and matrix)
286
Yamada, Yamada, and Kubomura
FiberC
FiberB
I
0
I
G,
0
I''•
l1
II
(:\1 I
0
I
I
1111
I
I
I
1111
I
I
0
: 36 ;i1i!si 2sl
~38 is~i
! !!
I
'
1111
I
I
::m.&: :
!
FiberD
0
(
(mm)
0
0
Fig. 12-Crack pattern of the specimens (fiber B, C, D)
6
5
E
E
4
c:0
3
Cl)
2
u
;;:::
~
~
Cl)
0
0
"
0.00 0.02
.
1:1
c
c
•
c
I
8
•
B
•
•
.c
•
•
0.04 0.06 0.08 0.10
c
c
X
8
D
Matrix
0.12 0.14
Volume Fraction, %
Fig. 13-Deflection at maximum load versus volume fraction
(fiber B, D, and matrix)
SP 138-18
Control of Cracking by Use
of Carbon Fiber Net as
Reinforcement for Concrete
by T. Makizumi, Y. Sakamoto,
and S. Okada
Synopsis: This study investigates the cracking control effectiveness of Carbon
Fiber Net (CFN) reinforcement , a two-dimensional grid consisting of sets of
continuous carbon fibers, for the case of flexural cracking. Prestressed concrete
sheet piles reinforced with and without CFN reinforcement were tested in
bending and the crack widths were examined. The CFN used in these tests had a
element spacing of 20 mm in each direction; each element consisted of three
strands each of 18,000 (18K) filaments. The netting was located in the
specimens at a concrete cover of 3mm. Since CFN could be set near the concrete
surface and the transverse strands of CFN play an important roll in resisting the
applied tensile force, the crack widths were controlled effectively. A model for
the prediction of the crack width in concrete reinforced with CFN, considering
this resisting mechanism, is proposed in this paper. Good agreement with the
calculated results and the experimental data is obtained.
Keywords: Bending; carbon; cover; cracking (fracturing); fiber reinforced
concretes; prestressed concrete
287
288
Makizumi, Sakamoto, and Okada
ACI member Tatsunori MAKIZUMI is an associate professor of civil
engineering at Kyushu University, Japan. He obtained a doctoral degree from
Kyushu University in 1983. His interests include continuous fiber reinforced
concrete, lightweight concrete and drying shrinkage cracking of concrete.
Yoshifumi SAKAMOTO is a professor of civil engineering at Kyushu
University, Japan. He is specializing in the development of cement materials and
admixtures .
Shin'ichiro OKADA is a research engineer of carbon materials at Osaka Gas
Co., LTD. He received a BS degree in mechanical engineering from University
of Kobe, Japan in 1987.
INTRODUCTION
Carbon fiber has a great resistance to corrosion so that it is possible to
reinforce concrete at the surface of concrete using carbon fiber fabrics, such as
Carbon Fiber Net (CFN). Surface reinforcement appears to be an effective way
to control crack width in concrete. Since the thickness of the concrete cover is a
major variable in the crack control equations for beams, the smaller the concrete
cover, the smaller should be the crack width.
Conventional steel reinforcing bars need sufficient concrete cover to prevent
bar corrosion and, therefore, cannot be placed near the surface of concrete. This
required amount of concrete cover cannot be disregarded in designing the
reinforcement for the crack control in relatively thin concrete members. For
instance, the prestressed concrete sheet pile standardized by JIS A 5326 needs a
fair amount of conventional reinforcement; the ratio of reinforcement area to
gross concrete area is about 1%, only for the control of cracking.
In this test, the effect of the surface reinforcement using CFN on the control of
crack width was examined. Prestressed concrete sheet piles of 8 em thickness,
reinforced with and without CFN, were tested in bending. The thickness of
concrete cover was 3mm for CFN and 22mm for the conventional reinforcing
bars, respectively. As a result. the crack width of concrete reinforced with CFN
was one-half of that of the conventional reinforced concrete, though the cross
sectional area of CFN was about one-seventh of that of the conventional
reinforcing bars.
From these points of view, surface reinforcement using CFN appears to be an
efficient and reasonable way for crack control in concrete and one of the most
useful applications of FRP to concrete reinforcement.
CARBON FIRER NET REINFORCEMENT (CFN)
Carbon fiber net is in the form of a layer of continuous mesh fabricated from an
assembly of continuous fiber elements. Fiber strands are perpendicularly
interwoven alternately three times so that the each element of the net consists of
FRP Reinforcement
289
three strands. Their intersections are rigidly connected by epoxy resin. Also the
whole of the net is formed by epoxy resin with matrix content of about 40%. Fig.
1 shows an example of CFN. Characteristics of CFN are as follows:
(1) Two-dimensional reinforcement CFN is available for use in planar concrete
structures, such as a slab, wall, panel, shell or tank (1).
(2) Due to the resistance of the transverse strand against tensile force, positive
anchorage of the reinforcement to the concrete is obtained within a short
length.
(3) It is possible to reinforce concrete at the surface without providing any
corrosion protection.
(4) Mesh spacing, and the number of filaments each way, are easily varied
according to the application and the needs of economization.
(5) No assembling of reinforcement at the site is necessary.
TEST I'ROCEHlJRES
The PC sheet pile tested has a thickness of 8 em, width of 50 em and length of
2 m. It is prestressed by pretension method at the age of 2 days using four PC
wires (type SWPR 7 A 9.3). According to JIS, the cracking width at the surface
is limited to less than 0.05 mm at a bending moment of 0.78 tonf·m so that four
reinforcing bars are actually provided to control cracking.
Six sets of prestressed concrete sheet piles were cast and there were two
specimens per set. The details of the test specimens are shown in Fig. 2.
Specimens SIC, S2C and S3C were cast using CFN. Specimens Sl and SIC
were cast with conventional Dl3 bars, which are actually used for the crack
control reinforcement in the prestressed concrete sheet pile. Specimens S2 and
S2C were cast with conventional D6 bars and specimens S3 and S3C without
conventional reinforcing bars. It is noted that the concrete cover is different
between the conventional reinforcing bars and CFN. In this test, CFN was set in
the specimen at a concrete cover of 3 mm.
The CFN was fabricated at element spacings of 20 mm in each direction and
each element consisted of three strands of 18K (18,000) filaments each. It is
made from Pitch-type continuous fibers having a diameter of 1 x 10-5m, tensile
strength of 200 kgf/mm2 and elastic modulus of 18 tonf/mm2. The cross-section
area of a single CFN layer (width of 50 em) in each direction is 0.353 cm2. This
cross-section area is smaller than that of the conventional reinforcing bars. The
cross sectional areas of two Dl3 bars and two D6 bars are 2.534 cm2 and 0.634
cm2. respectively.
Concerning the mix proportion of the concrete, the water/cement ratio was
26%, maximum size of aggregate was 20 mm and slump was 5 em. At the time
of making the specimens, a CFN layer was placed on the bottom form without
any spacers, since the density of CFN is less than that of concrete and a thin
cover can be secured easily. Placement of the reinforcement and casting of the
concrete are shown in Fig. 3. Zooming in on the CFN layer before concrete
vibrating is shown in Fig. 4. After casting, concrete was compacted by an outer
vibrator and cured by steam to get a strength of 300 kgf/cm2 at the age of 2 days
for prestressing.
Bending tests were conducted by the use of simple beams with third point
loading at the span length of 1 m. In order to measure crack width easily, the
290
Makizumi, Sakamoto, and Okada
specimen was set upside down and load was applied upward using an oil
pressure jack, as shown in Fig. 5. The crack widths were measured at both sides
and the middle of the tensile face at the center of span, and nearby both supports
using a microscopic crack meter of forty magnifications. Fig. 6. shows the
measurement in progress.
RESULTS AND DISCUSSION
The crack width observed versus applied bending moment is shown in Fig. 7.
The crack widths in the specimens using CFN are much less than those in the
specimens without CFN. Though the cross-sectional area of CFN is about oneseventh of that when using conventional reinforcing bars, CFN reduces the crack
width by one-half in the case of using Dl3 bars comparing specimen SIC with
specimen Sl. Furthermore, when not using conventional reinforcing bars as in
specimen S3C. the crack width is less than 0.05 mm which is the limiting width
at the critical cracking moment 0.78 tonf.m. From these results, it is recognized
that CFN reinforcement is quite effective for controlling crack width.
The decrease in crack width when using CFN reinforcement. as shown in Fig.
7, is mainly due to placing the reinforcement near the tension face of concrete,
since the thickness of the concrete cover is a major variable in the crack control
equations for beams. For typical examples, the crack control recommendations
proposed in the JSCE (Japan Society of Civil Engineers) Code (2), the ACI 31889 Code and the European Model Code for Concrete Structures are such that:
JSCE Code; W = k (4c + 0. 7( s- db )j·fs
E,
(2)
ACl Code; W = 2.20 .{, ',jd,.A
E,
CEB Code; W
=
0.7 f (3c + 0.05 db
E
PR
(1)
J
(3)
where
W : mean crack width
k : bond coefficient, 1.0 for ribbed bars, 1.3 for round bars
c : clear concrete cover
s : spacing of reinforcement
db : diameter of the reinforcement
.{,: stress of the reinforcement at the cracked section
B: ratio of distance between neutral axis and tension face to distance between
neutral axis and centroid of the reinforcement
d,.: thickness of cover from tension face to center of the reinforcement
A: area of concrete symmetric with reinforcement divided by number of
reinforcing bars
pR: = AR/ A , AR is the cross sectional area of the reinforcement
FRP Reinforcement
291
Fig. 8 compares the crack widths in CFN reinforced concrete calculated by
each equation and the experimental results. According to the ACI Committee
224 Report(3), crack control equation for beams such as Eq. (2) may
underestimate the cracking widths developed in two-way slabs and plates.
Taking this into account, the results calculated by these equations approximate
the experimental results well in all the cases. Such fittings lead to the conclusion
that placing the reinforcement near the surface of the concrete member is an
effective way to control crack width as far as the reinforcement could be placed
with a thin concrete cover.
PREDICTION OF CRACK WIDTH IN CONCRETE
REINFORCED WITH CFN
Cracking in concrete reinforced with two-way reinforcement is controlled
primarily by the stress level in the reinforcement and the spacing in the direction
of applied force (4). As the spacing is already defined, the stress or the strain in
the reinforcement are the most important variables to predict the crack width.
After the development of cracking in a reinforced concrete element, the stress
in the concrete at the cracked section is reduced to zero and is redistributed to
the reinforcement. so that the force in the reinforcement equals the external load
at the cracked section. Between the cracks, the force in a conventional
reinforcing bar decreases and is transferred to the concrete continuously by the
bond action between the concrete and reinforcement. In contrast to such
continuous transfer, the force in an element of CFN changes discontinuously and
at the intersection points of the net, called nodal points, decreases abruptly due
to the resistance of the transverse element against the tensile force.
A simple model for prediction of the stress or strain in the reinforcement is
based on the following assumptions:
(1) most of the stress between adjacent transverse elements, or between a mesh,
is approximated to be constant.
(2) tensile force corresponding to the stress changes in the reinforcement at the
nodal points is transmitted to concrete by the resistance of the transverse
element.
(3) the amount of resisting force of the transverse element is in proportion to the
relative longitudinal displacement, b , between the transverse element and
concrete at the same cross section by the coefficient K, an apparent spring
constant of the concrete.
(4) longitudinal displacement of a nodal point is equal to that of the transverse
element at the same cross section.
(5) elongation of the longitudinal element between a mesh is equal to the
difference between the displacement of the adjacent two nodal points.
The distribution of longitudinal strain in the element and longitudinal strain in
the concrete is schematically represented in Fig. 9. Supposing that the element of
CFN is anchored to the concrete at the nth mesh location from the crack section,
the width of cracking is the summation of all the slip between the element and
the concrete and drying shrinkage on the both sides of the cracked section. The
expression for crack width ( W) can be written as
292
Makizumi, Sakamoto, and Okada
n
w= 2
I(~,+ Sh)L
X
(4)
j ... Q
where Ll£; is the difference between the element tensile strain, £ 1 ,, and the
concrete tensile strain, £,.,;(each strain is the average within the i th mesh
location where i is numbered from the cracked face), Sh is the drying shrinkage
and L is the spacing of the mesh.
The uniform tensile strain in the concrete
is
£r.i
1
A.F .)
,.,, =-(T-E
EA
1 rl.•
£.
I•
(5)
(•
where E is Young's modulus, A is the area of cross section, suffixes c and
represent concrete and fiber, respectively, and T is the applied tensile force.
f
The tensile strain in the element£ 1., is
£ 1,1.
=(b.5.,I 1 )/L
I
(6)
where 5, is the displacement of the ith nodal point. Since the product of the
displacement 5, and the spring constant of concrete K is in proportion to the
stress changes in the element at the ith nodal point
(7)
and the boundary conditions are as follows:
at i=O: £r.o
=
(8)
TfEr AI
T
J
J
c
(9)
,bn+1=0
ati=n:£f.n =£,.·" = E A +EA
c
From these equations the element strain £ r.; can be expressed by the following
relations:
£
£
£
r. 1
p
fA
T-K6
E A
=---1
r r
'
£
=
(l+~)T-(2+~)K6
E A
P
r
'
1
(1+3~+~ 2 )T-(3+4~ +~")KO,
= -'------'---'------'-EA
I
I
( 1+ 6~ + 5~ 2 + ~' )T- (4 + 10~ + 6~ 2 + ~')K6 1
E A
=~-----~-~------~--
r r
(10)
FRP Reinforcement
and b, is varied with the mesh location
at n =2 , b
=
I
atn=3,b=
I
293
,n, as follows:
1
1 +B
--mT
2+B
1+3B+B2
mT
By combining these equations, the width of cracking is obtained as:
W
=
2L
i {E.r.;(1 + E.rA.r)_I_+ Sh}
ECAC
ECA,.
(11)
l'O
Fig. 10 shows the curves calculated by Eq.( 11) with varying the mesh location,
n. The values used in this calculation are as follows:
uP£: effective prestressing stress in PC wire, =95 kgf/mm2
Ac: cross sectional area of concrete symmetric with reinforcement divided by
number of longitudinal strand ofCFN, =(2 x 3 + 3) x 20=180mm2
A.r: area of a longitudinal element ofCFN, =1.414 mm2
E/ Young's modulus ofCFN, =18 tonf/mm2
E,.: Young's modulus of concrete, =4.1 tonf/mm2
L: spacing of a mesh,=20 mm
Sh: drying shrinkage of concrete,= 150 x 10-6
Eq.(ll) fits the data well, especially for the case of n =3 a good agreement
between the calculated results and the experimental results is obtained. Such
agreement shows that the cracking width in concrete reinforced by CFN can be
predicted by Eq.(ll) and the model used in Eq.(ll) is acceptable for estimating
the longitudinal stress in the element of CFN.
CONCLUSIONS
1) Crack width in concrete can be controlled effectively by placing the
reinforcement near the tension surface of concrete. CFN can be available as a
surface reinforcement without providing any corrosion protection.
2) Crack width in CFN reinforced concrete can be predicted by Eq.(ll ). Good
agreement with the calculated results and the experimental data is obtained in
this test.
294
Makizumi, Sakamoto, and Okada
3) The model proposed, considering that the transverse elements play an
important role in resisting the applied tensile force, can estimat the tensile
stress in two-dimensional grids such as CFN and wire mesh.
4) In CFN, due to the resistance of the transverse element against tensile force,
positive anchorage of the reinforcement to the concrete is obtained within a
short length.
ACKNOWLEDGMENTS
The authors wish to thank the FUJI PS Corporation and its Engineering
Division for the PC sheet piles used in this investigation. The financial assistance
which the first author received through the IKET ANI SCIENCE and
TECHNOLOGY FOUNDATION is acknowledged with thanks. Also the
comments and encouragement of Dr. John Bolander Jr., Department of Civil
Engineering in Kyushu University, and Dr. Antonio Nanni, Chairman of this
Symposium, are gratefully acknowledged.
REFERENCES
I. ACI Committee 549, "State-of-the-Art Report on Ferrocement (ACI 549R82), " American Concrete Institute, Detroit, 1982, Chap.4.
2 Japan Society of Civil Engineering, "Model Code for Concrete Structures
(1991 Edit.)," Tokyo, 1991, pp.85-87.
3. ACI Committee 224, "Control of Cracking in Concrete Structures (ACI 224R80), " American Concrete Institute, Detroit, 1980, Chap.4.
4 Nawy E.G., "Crack Width Control in Welded Fabric Reinforced Centrally
Loaded Two-Way Concrete Slabs, " ACI SP-20, "Causes, Mechanism, and
Control of Cracking in Concrete," Detroit, pp.221-235, 1966.
FRP Reinforcement
.
"
•
~
T
295
•,
r'
'
'
c
"
,·
•
JIIII)Uilrlll/ffi'' !! ~~ll)lll~m)n:r;~;ql 111111~nmn 'il,
"
I
0
'
' '
.
5
'
8
oz
T I
I i
:
I
'
1fljl l ln"J'IIP 1!!
:1
10
,,
•
III~IJ!':IIfllllflW!I
., I
"
or
If I
I' ! '
' I
...
"lllllilhl
" '
STAE!m
I I
I 1
I
•
0
Fig. 1-Sample of carbon fiber net (CFN)
PC wire SWPR7A <J>9.3
reinforcing bar 013
S1
0
0
•
•
0
0
reinforcin
S2
0
0
•
•
0
tso S1C
500
0
bar 06
S2C
0
specimens without CFN
specimens using CFN
Fig. 2-Details of the test specimens
I
296
Makizumi, Sakamoto, and Okada
Fig. 3-Placement of the reinforcement and casting of the concrete
Fig. 4-CFN layer on the bottom form at the casting before vibrating
FRP Reinforcement
Fig. 5-Specimen testing set-up
Fig. 6--The measurement of crack width in progress
297
298
Makizumi, Sakamoto, and Okada
0.25r---------------------------------~
0.20
E
E
i
0.15
-"!:
0.10
.IE
-a
u
•
•
•c
11
S1C
52
S2C
53
A 53C
0
•
•
•
•...
0
•A
0.05
0.00
u.4
a
0.1
••
•••
8
11
c
11
c
0.8
A
1.0
1. 2
1.4
Moment, tfm
Fig. 7-The crack width in the concrete with and without CFN reinforcement
S1C
S2C
S3C
E o.t5 --- CEB
-JSCE
E
-·- ACI
s£
o EXP.
:2 0.10
3:
-
.:.:
(,)
I!!
0 0.05
0
.5
1.5 . 5
I
Moment, tfm
Moment, tfm
Moment, tfm
Fig. 8-The experimental results and predictions of the ACI Code,
JSCE Code and CEB Code equations for the crack widths in CFN
reinforced concrete
FRP Reinforcement
CFN element strain Ef
K'=K/EfAr
Fig. 9-Schematic distribution of longitudinal strain in the CFN
element and longitudinal strain in the concrete
S1C
E D.l5
E
£
~
~
.¥
u
o.1
()I! o. 05
o EXP.
S2C
S3C
-
llesh loc.tlon, n
~; ~; /;
O .5
1.5 .5
Moment, tfm
1.5 .5
Moment, tfm
I
1.5
Moment, tfm
Fig. 10-The crack widths in CFN reinforced concrete along with
the predictions of Eq. (11)
299
SP 138-19
Absorbing Capacity of Cushion
System Using Concrete Slab
Reinforced with AFRP Rods
by T. Tamura, H. Mikami,
0. Nakano, and N. Kishi
Synopsis:
This paper presents the results of experimental study on the shock
absorbing performance of proto-type three-layered cushion system. This
system consists of a concrete core slab reinforced with braided Aramid
Fiber Reinforced Plastic (AFRP) rod, and sandwiched between sand
layer (top) and expanded poly-styrol layer (bottom).
In order to study the effect of the rigidity, elongation and bond
strength of reinforcing bars on the shock absorbing performance of a
three-layered cushion system, three types of reinforcing bars were used;
AFRP rod with their surface bonded with silica sand, non-sand surfaced
AFRP rod and deformed steel bar. Furthermore, these results were
compared with the results when single sand layer was used as cushion
material.
The results achieved from these experiments are: ( 1) The transmitted stress of the three-layered cushion system is distributed more effectively than that of single sand layer. (2) The distribution pattern of the
transmitted stress in the three-layered cushion system was affected by
the bond property of the reinforcing bar. (3) Duration time of the
transmitted impact force was affected by the rigidity of the reinforcing
bars.
Keywords: Fiber reinforced plastics; fibers; impact; impact strength; nuclear
reactors; reinforced concrete; shock resistance; slabs; stresses
301
302
Tamura et al
T. Tamura is a assistant manager of Technical Research Institute at
Mitsui Construction Co. Ltd., Chiba, Japan. He is a doctoral student for
degree at Kyusyu Institute of Technology. His research interests include
FRP reinforced concrete and durability of concrete. He is a member of
JSCE and JCI.
H. Mikami is a senior research engineer of Technical Research Institute at Mitsui Construction Co. Ltd., Chiba, Japan. He received his Doctor of Engineering Degree in 1992 from Nippon University. His research
interests include FRP reinforced concrete and steel-concrete composite
structure. He is a member of JSCE and JCI.
0. Nakano is a head of structural engineering of Civil Engineering
Research Institute at Hokkaido Development Bureau, Sapporo, Japan.
He received his Master of Engineering Degree from Hokkaido University
in 1974. His research activities include impact behaviors of rock-shed
structure. He is a member of JSCE and JCI.
N. Kishi is a associate professor of Muroran Institute of Technology,
Muroran, Japan. He received his Doctor of Engineering Degree in 1977
from Hokkaido University. His research activities include impact
behaviors of structural members and stability of flexibly jointed frame.
He is a member of ASCE, SSRC, JSCE and JCI.
INTRODUCTION
Recently, the concern to ensure the greater safety of facilities of
nuclear power plants, rock-sheds or other important structures against
impact loads has been gaining tremendous momentum. There are two
aspects of investigating the impact behavior of a structure: ( 1) investigate the impact behavior of a structure, assuming that the impact loads
directly applied on the structure and (2) investigate a cushion system
that attenuates the impact loads.
The authors have studied both these aspects of the impact behavior
previously (1,2). First of all, the impact behavior of RC slabs, PC slabs
and concrete slabs reinforced with braided AFRP ( Aramid Fiber Reinforced Plastic ) rods subjected to impact loads have studied under elastic
and elasto-plastic regions. Furthermore, the authors have studied the
impact absorbing behavior of sand layer, expanded poly-styrol layer and
three-layered cushion system, where the slab was sandwiched between
sand layer (top) and expanded poly-styrol layer (bottom).
This paper presents the results of experiments on the shock absorbing performance of a three-layered cushion system that consists of a concrete core slab reinforced with braided AFRP rods, and sandwiched
between sand layer and expanded poly-styrol layer.
In order to study the effect of the rigidity, elongation and bond
strength of reinforcing bars in the reinforced concrete core slab on the
shock absorbing performance of a three-layered cushion system, three
types of reinforcing bars were used; AFRP rods with/without bonded
surface with silica sand and deformed steel bars. Furthermore, these
FRP Reinforcement
303
results were compared with the result when single sand layer was used
for the cushion material.
The impact loads were generated by free falling of a 30 kN steel
weight on the center of the cushion system from heights of 10, 20 and 30
meters.
OUTLINE OF EXPERIMENT
Method of Experiment
Figure 1 illustrates the profile and dimensions of the cushion system
used for the experiments. A three-layered cushion system was mounted
on RC foundation with the dimensions of 660 x 660 x 60 em. The cushion
system was subjected to impact loads by free fallin~ of a 30 kN cylindrical steel weight with a flattened spherical bottom ( diameter of cylinder
is 100 em, height of cylinder is 97 em and height of spherical bottom is
17.5 em ) from a predetermined height.
The experimental cases are listed in Table 1. In these experiments,
four numbers of cushion systems ( four types of cushion systems ) were
made and tested, the impact loads were applied in three times ( from
three different heights ) on the center of each cushion system.
Reinforced Concrete Slab
As shown in Table 1, three types of slabs were fabricated and used:
(1) concrete slab reinforced with braided AFRP rods with their surfaces
bonded with No.5 silica sand (AsC), (2) concrete slab reinforced with
braided non-sand bonded AFRP rods (AC) and (3) concrete slab reinforced with deformed steel bars (RC).
All these slabs had identical dimensions of 395x395x20 em and had
the single reinforcement. The reinforcement ratio for all these slabs is 1
%. AFRP rods ( K192S and K192 ) and deformed steel bars (D13) were
arranged at 10 em intervals right angled with each other.
The material properties of these reinforcing bars are listed in Table
2. An AFRP rod has many advantages compared to the conventional
steel reinforcement such as lightweight, non-magnetizable and corrosion
resistant. The elastic elongation of an AFRP rod is about 10 times than
that of a steel bar. The elastic modulus of an AFRP rod is about onethird of a steel bar and its tensile strength is about 2.5 to 3.4 times
greater than that of a steel bar.
The mix proportion of the concrete was designed to produce a
design strength of 20.6 MPa. Concrete was placed and cured on site.
The 28-day compressive strength of the cast-in-place concrete was 23.4
MPa. On commencement of the experiment, the concrete had aged 30
days.
304
Tamura et al
Materials of Top and Bottom Layers
In these experiments, sand layer and expanded poly-styrol layer
with 50cm thickness in each were used as top and bottom layer respectively. The material properties of sand was indicated in Table 3. Sand
was spread and compacted in each 20 em layer with stamping. After
each falling of the steel weight from heights of 10 and 20 m, the portion
into which the weight penetrated was dug up and restored to 50cm by
recompacting the sand in each 20 em layer.
The material properties of expanded poly-styrol used as bottom
layer were indicated in Table 4. In these experiments, blocks of
expanded poly-styrol with dimensions of 200 x 100 x 50 em were arranged
to form the predetermined size of 400 x 400 x 50 em.
Measuring Method
The acceleration of the free-fall steel weight was measured by using
4 accelerometers to calculate the impact force generated when the
weight collided with cushion system.
The impact loads transmitted to the RC foundation were measured
in one direction using 25 load cells located along the centerline of RC
foundation. Diameter and capacity of each load cell were designed as 25
mm and 294 MPa respectively. In these experiments, as it was a main
object to obtain the transmitted stresses, the transmitted impact loads
measured by load cells were converted to the stresses.
Outputs from the individual sensors were recorded by wide-band
data recorders, then A/D conversion was made and the data were finally
processed using workstations.
EXPERIMENTAL RESULTS AND CONSIDERATIONS
Wave Configuration of Transmitted Impact Stress
Figure 2 shows the configurations of the transmitted impact stresses
with respect to time measured by load cells from different experimental
tests. In this figure, horizontal and vertical axis in the plane of paper
show the location of load cells (em) and transmitted impact stresses
(MPa) respectively, and the axis perpendicular to the plane of the paper
shows the time of duration (msec).
The origin of the figure is at the center of loading which is at the
center of RC foundation. The wave patterns on two sides of the center
of loading were almost symmetric to each other. Hence, this figure
shows the wave at load cell located only on one side of the center of
loading between the center and the edge of RC foundation. The total
sampling times in all the cases were 200 msec.
These figures demonstrate the effect of the core slab of the threelayered cushion system on the stress dispersibility. When the weight was
FRP Reinforcement
305
dropped from a 10 m height, the impact stresses were distributed umformly in all of the three cases.
Allowing to fall the weight from a 20 m height, in all the three
cases, stresses at all the points within 70 em from the center of loading
were greater than stresses at points beyond 100 em from the center of
loading. The dispersibility of stresses seems to be somewhat deteriorated
in this case. However, the impact load was still distributed well for this
falling and impact energy generated by weight dropping.
When the weight was dropped from a 30 m height, the stress concentration at the center of loading was increased compared to the case
when the weight was dropped from a 20 m height. The level of stress
concentration was higher in the slab AC than in the slabs AsC and RC.
In the slab AC, the stress at the center of loading is few times higher
than the corresponding value in the slabs AsC and RC. Also the stress
at a point 20 em away from the center of loading was especially big,
because the concrete at the loading point was separated due to the low
bond strength of the rods in the slab AC and the separated concrete
splinters dug themselves down into expanded poly-styrol blocks.
On the other hand, although the dispersibilities of transmitted
impact stresses of the slabs AsC and RC were deteriorated, the stresses
stood around 1MP a and less. These results imply that these slabs still
have the effectiveness as a cushion material. Based on the facts that, 1)
the rigidities of reinforcing rods used in slabs AsC and AC were equal
each other and 2) the bond strength between reinforcing bars and concrete in the slabs RC and AsC was superior to that of in the slab AC, it
seems that the bond strength between the reinforcing bar and concrete is
closely related to the dispersibility of transmitted impact stress.
Comparing the transmitted impact stresses in cases between threelayered cushion systems and single sand layer, the dispersibility of
stresses in the former cases was excellent up to 20 m falling while
stresses in the later were concentrated at the center of loading with an
area of 100 em radius. When the drop height is 10 m, the maximum
transmitted impact stress in case of single sand layer is 3. 7 MP a which is
about 18 to 23 times than those of the three-layered cushion systems.
When the drop height is 30 m, the stress distribution in case of single sand layer was similar to that of the slab AC. Then, the performance of absorbing capacity of three-layered cushion system using the
slab AC with a 30 m fall, is decreased to the level of single sand layer.
In these experiments, the sand layer in three-layered cushion system
was dug up after completion of each experiment and extra sand was
added to make up for the volume reduction of the sand layer caused by
the residual deformation of the slab or the expanded poly-styrol blocks,
so that the top layer could be restored to its initial level. In case of the
slab AC for a 30 m fall, the concrete in the center portion of the slab
was separated and caused residual deformation in slab due to the low
bond strength of the non-sand surfaced AFRP rods. Then, large amount
of sand was filled up. That is the reason why the stress distribution in
this case was similar to that of single sand layer.
306
Tamura et al
Wave Configuration of Impact Force
In this paper, the following two methods were used for the evaluation of impact force generated when the weight dropped on the cushion
system; ( 1) impact force obtained by multiplying the acceleration (G) of
the steel weight by its weight ( hereafter referred to as the weight
impact force ) and (2) impact force obtained by summing up the distributed stresses measured by the load cells ( hereafter referred to as the
transmitted impact force ). Assuming that two transmitted impact
stresses from the origin are the same in both directions, the transmitted
impact force was calculated by integrating the measured stresses on one
side of the origin.
Figure 3 shows the wave configurations of weight and transmitted
impact force with respect to time by solid and broken lines respectively.
From this figure, in cases of three-layered cushion systems, although
both wave configurations are different for each drop height, but the configurations are similar for three cases. Therefore, they may not be
affected by rigidity and bond strength of reinforcing bars.
In the case of the single sand layer, the variations of both waves of
the impact forces with respect to time are independent of the drop
height and are almost the same.
Time Duration of Impact Force
Time durations of both impact forces are listed in Table 5. In case
of three-layered cushion system with a 10 m drop height, time duration
for weight impact force is the same in all cases which is about 80 msec.
However, in the cases of 20 and 30 m drop height, the time durations for
the slabs AsC and AC are about 105 msec which is longer than that of
the slab RC. The time durations of transmitted impact force for threelayered cushion systems with a 10 m drop height are almost similar to
that of weight impact force among them.
For drop heights of 20 and 30 · m, the time durations of the
transmitted impact force for the slabs AsC and AC are longer than that
of the slab RC. These results imply that the effect of rigidity of reinforcing bars on the time duration may be greater than that of bond
strength. That is understood by following reasons: since, in the cases of
the slabs AsC and AC, reinforcing rods have low rigidity and wide elastic region, the membrane action becomes more dominant than the bending action due to increasing cracks. While the displacement and the
range of deformation in slab are increased by the membrane action, time
duration is lengthened. Especially in the case of the slab AC, although
the stress concentration occurred in the center portion of loading point
mentioned in the previous section, the time duration of transmitted
impact force is similar to that of the slab AsC. This behavior implies
that the membrane action by the reinforcing rods in the slab AC is effectively affected to lengthen the time duration.
In the case of single sand layer, the time durations of transmitted
impact force for all drop heights are seemed to be about 50 msec which
is about 5/7 to 3/7th of time durations for three-layered cushion
FRP Reinforcement
307
systems. Based on that 1) the fundamental natural period of impactproof structures like rock-shed etc. is generally between 50 to 60 msec
and 2) time duration of impact force made resonance in the structure is
the half of its fundamental natural period, the lengthened time duration
of transmitted impact force in three-layered cushion system will play significantly important role in reducing the sectional forces of structures.
Maximum Impact Force
Maximum impact forces for each case are listed in Table 6. Since
the stress appeared to be concentrated in the case of the slab AC for a
30 m drop height, the transmitted impact force may be overestimated.
Therefore, this case has been excluded for comparative study.
The weight impact forces for all cases of three-layered cushion systems are observed to have similar tendency among them. The maximum
weight impact forces are about 1200 kN for a 10 m fall and are in the
range of 1900 to 2000 kN for drop heights of 20 and 30 m. On the other
hand, the maximum transmitted impact forces in the slabs AsC and AC
are seemed to be similar. However the impact force in the slab RC is
smaller than those of the slabs AsC and AC. Contrary to the case of
time duration of impact force, these results imply that the maximum
transmitted impact force is affected by the rigidity of reinforcing bars.
Comparing the maximum impact forces between transmitted and weight
ones, the former is smaller than the later. Those proportions are almost
2/3, 1/2 and 1/2 for 10, 20 and 30 m drop respectively.
In the case of single sand layer, the maximum transmitted impact
force is 1.7 to 1.8 times the maximum weight impact force. Except for
the case of the slab AC with a drop height of 30 m, the maximum
transmitted impact force in the three-layered cushion system is smaller (
1/3 to 1/4 ) than that of single sand layer.
CONCLUSIONS
With the aim of studying the shock absorbing performance of a
cushion system using a concrete slab reinforced with braided AFRP rods
sandwiched between sand layer and expanded poly-styrol layer, experiments have conducted on the proto-type model and compared the results
with the cases in which ( 1) deformed steel bars were used as a reinforcing bars of core slab, and (2) a single sand layer was used as a cushion
material. The results achieved through these experiments are summarized in the following paragraphs:
(1) The three-layered cushion system allowed the transmitted stresses
to be distributed more effectively than single sand layer. Then, this
system make the maximum transmitted impact force reduce and
also make the time duration of the impact force lengthen. Threelayered cushion system make the transmitted impact force be
reduced less than 1/3rd of single sand layer. The sectional forces of
main structure can be significantly reduced by using this system.
308
Tamura et al
(2) The distribution of the transmitted stresses in the three-layered
cushion system was affected by the bond property of the reinforcing
bars. Following the occurrence of cracks, the dispersibility of the
transmitted stress in the slab AC became less than that in the slabs
AsC and RC.
(3) The time durations of the transmitted impact force in the slabs AsC
and AC tend to be lengthen more than that of the slab RC. However, the maximum transmitted impact force of the slab RC is
smaller than that of the slabs AsC and AC. Both of these characteristics in these slabs act to be reduced the sectional forces of main
structure.
From these results, it is clear that AFRP rods can be used as reinforcing bar of core RC slab similar to steel bars in three-layered cushion
system.
ACKNOWLEDGEMENT
The authors sincerely thanks to Dr. Ken-ichi G Matsuoka ( Prof.
M uroran Institute of Technology ) and Messrs. Atsushi Matsuoka,
Masahiko Kudo, Yasusi Oyama and Toshiharu Satake ( graduate Students of M uroran Institute of Technology ) who support this investigation.
REFERENCES
(1) H. Mikami, N. Kishi, K. G. Matsuoka and S. G. Nomachi, Impact
Behavior of Concrete Slab Reinforced by Braided AFRP Rods
under Heavy Weight Falling, Journal of Structural Engineering,
JSCE, vol. 37A, pp. 1591-1602, Mar. '91, in Japanese
(2) N. Kishi, 0. Nakano, H. Konno and K. G. Matsuoka, Laboratory
Experiment on Shock-Absorbing Effects of Three-Layered Absorbing System, Journal of Structural Engineering. JSCE, vol. 38A, pp.
1577-1586, Mar. '92, in Japanese
FRP Reinforcement
309
TABLE 1 - EXPERIMENTAL CASES
Type of reinforcing bar
Core slab
Sand-surfaced rod (K192S)
AsC
Non-sand surfaced rod (K192)
AC
Deformed steel bar (D13)
RC
A sand layer with 90cm thickness
Drop height of weight (m)
10
20
30
TABLE 2 - MATERIAL PROPERTIES OF REINFORCING BARS
Reinforcing bar denomination
Material
Nominal diameter (mm)
Nominal sectional area (em 2)
Density (g/cm 3 )
Yield strength (KN)
Tensile strength (KN)
Elastic modulus (GPa)
Elastic elongation (%)
Poisson's ratio
K192S
Aramid
15.5
1.50
K192
Aramid
14.0
1.50
D13
SD30A
12.7
1.27
1.30
1.30
7.85
-
-
188.3
63.7
1.97
0.62
188.3
63.7
1.97
0.62
37.3
55.9 ,...., 75.5
205.9
0.20
0.30
TABLE 3 - MATERIAL PROPERTIES OF SAND
Unit weight
(KN/m 3 )
16.2
Specific
gravity
2.57
Moisture
percentage (%)
1.86
Uniformity
coefficient
4.11
TABLE 4- MATERIAL PROPERTIES OF EXPANDED POLY-STYROL
Density
(N/m 3 )
200
Compressive strength
at 5% strain (MPa)
0.11
Poisson's
ratio
0.05
Max. of elastic
strain(%)
1 or less
Stress at the max.
elastic strain (MPa)
About 0.06
TABLE 5- TI.tvfE DURATIONS OF IMPACT FORCES
Core Slab
AsC
AC
RC
Sand layer
Time duration of weight
impact force (msec)
10m
20m
30m
78.20
106.40
105.90
78.60
111.00
105.60
87.90
79.40
95.00
105.00
113.40
131.60
Time duration of transmitted
impact force (msec)
10m
20m
30m
73.70
128.60
117.20
79.20
133.30
110.10
71.50
103.50
94.60
55.60
55.60
46.40
TABLE 6 - MAXIMUM IMPACT FORCES
Core slab
AsC
AC
RC
Sand layer
Max. weight impact force (KN)
10m
20m
30m
1182
1940
1983
2032
1189
1993
2006
1953
1199
2908
1421
2311
Max. transmitted impact force (KN)
10m
20m
30m
773
945
1167
738
942
2405
682
806
945
2519
4078
5261
FRP Reinforcement
~
~cc~ ~:::·'
--t-~---~l
....
~E
]~
g.
0
!-t. t=~oro~
iRCub
r
S•nd layer I
I
I
$I.,
0
Expuded poly-lstyrol
10
8 :.:
:::::on
Gr•vel (•pprox. IOmm in
T'1
lo.d cell
i
size~
I
7200
(mm)
Fig. !-Profile and dimensions of cushion system
311
312
Tamura et al
(moee)
200
160
120
Slab
40
50
0
(em)
100
50
0
150
100
40
80 [
150
(moee)
H~30m
O 40
50
c..e
100
150
(moee)
200
160
120
80 [
100
150
0
2MPa
(em)
of slab AC
(moee)
200
160
120
H~10m
90cm sand
ao [
0.25MPa
(em)
(moee)
200
160
120
80 [
(em)
Slab RC H·10m
51 b AC H=30m
O 40
Case of sliib AsC
50
(maee)
200
160
120
(em)
k£jL . [,· ·
100
[ 0.25MPa
0
(em)
0.25MPa
"w""
Slab AoC
50
150
(moee)
200
160
120
Slab
50
100
50
100
150
5MPa
(em)
(msee)
200
160
120
(moee)
~
200
~ 160
120
VOem sand Ia er H-20m
O 40
50
100
150
ao [
5MPa
(em)
VOem und
O 40
50
100
150
Case of slab RC
(em)
ao [
2MPa
O 40
50
100
150
80 [
SMPa
(em)
Case of sand layer
Fig. 2-Wave configuration of transmitted impact stress
FRP Reinforcement
- - Weight imp1et force
- - - - Tran1mitted impact force
(KN) Slab AsC H=lOm
~::~1 ~'I
313
- - Weight impact force
- - - - Transmitted impact force
(KN) Slab AC H=lOm
I I ~::~1
f21sl
I I
_::f· r++1 l_::w· rgg; 1 1
2000 Slab AsC
100 ~
H-30m
I P>f}'t I
- 1000 0
(~5~Jslab
_::~I
_:::~i'
_:~jr··
40
80
120
o
1_:::~ I
160
200
0
H=30m
fvf>i. I I
40
80
120
160
time( msec)
time( msec)
Case of alab AsC
Case of alab AC
(~o~J90cm
RC H=lOm
V->i' I I
1_:::~1
200
sand layer H=lOm
P\l I I I
grr-r 1
=
_.L::fm
I_=._g_m..J. . -. L.I_
'--a-b....Jfi-c_>=_H
....
11
2600 Slab AC
ffi 1 1~fmfl'···rr 1
40
ao
120
160
time(msec)
Case of alab RC
200 30
o
ao
160
240
320
400
time(msec)
Case of sand layer
Fig. 3-Wave configuration of weight and transmitted impact force
SP 138-20
Investigation of Bond in
Concrete Member with Fiber
Reinforced Plastic Bars
by E. Makitani, I. Irisawa, and N. Nishiura
The bond tests with beam specimens reported in this
paper were performed in order to research the bond characteristic
of fiber reinforced plastic bars in concrete members which were
made of continuous fibers such as carbon, aramid, glass and vinylon
as well as matrix of epoxy resin. expecially under the condition
when the ends of the bars were not reinforced. It was testified
that pulling-out of the bars whose surfaces were processed into a
spiral shape or covered by cohesive grains of sand did not occur
if the bond length was more than 40d where d was the diameter of
the fiber reinforced plastic bar.
Based on the above test results, using the caron fiber reinforced plastic bars. the tests of lap splice joints were performed by arranging an anchoring position in the center of a
truss type of concrete specimen. It was recognized that the unit
bond strength increased as the length of the lap splice joint increased, but that it decreased for lap splice length between 40d
and 60d.
Synopsis:
Keywords: Bonding; bond strength; carbon; fiber reinforced plastics;
fibers; glass; lap connections; pullout tests; reinforcing steels; slippage;
tensile strength
315
316
Makitani, Irisawa, and Nishiura
ACI and PC! Members E. Maki tani is a professor of Department of
Architecture at Kanto Gakuin University. He gained his B. E.. M.E.
and D. E. degrees at Waseda University, Japan. During 1974 and 1975
years. he was a visiting research associate of civil engineering
at Columbia University to study on the aseismic design of nuclear
power plant under National Science Foundation.
I. Irisawa is a structural research engineer in the Structural
Engineering Section. Institute of Research, Fujita Construction.
Japan. He gained his B. E. and M. E. degrees at Kanto Gakuin University.
N. Nishiura is a student of master course,Depertment of Architecture at Kanto Gakuin University. He is now studying on the mechanical behaviour of RC members with FRP bars.
INTRODUCTION
In recent years.a number of reports on fiber material with regard to increasing the strength of reinforced concrete members in
adaptability have been publised. The reinforced material formed
from inorganic or organic continuous fiber as well as a plastic
matrix in the shape of a rod and mesh or in the shape of threedimensional textile fabrics is developed and used. When the fiber
reinfored plastic bar was used in concrete,
its surface needed
various processing; otherwise. its bond behaviour with adjacent
concrete was deficient. Bond tests on fiber reinforced plastic
bars were performed mainly in the pulling-out test. Contrary to
steel bars,
the bars broke easily at the plastic portion before
pulled out in the vicinity of a grip instrument due to its lower
bearing strength. Hence. reinforcement of the bar in the portion
of the grip instrument and high level test technique were necessitated. But, if the beam type of specimen was applied to the bond
test.
the test results could be derived without reinforcement of
the bar at its end portion.
In this paper, the bond characteristic was investigated by using a beam specimen and inorganic or organic continuous fiber in
the shape of a rod. Furthermore. the truss type of beam specimens
with two anchoring ends and caron fiber reinforced plastic bars
as main reinforcement were tested to research the effect on the
FRP Reinforcement
317
bond characteristic of the bars influenced by the change of lap
splice length. The results were presented as follows.
BOND TEST BY BEAM
The specimen shown in Fig.l is the compound beams made with two
beams in which each had a width of 300mm, a height of 350mm and a
length of 600mm. A steel pinch was used at the top of beams and a
fiber reinforced plastic bar was placed in the centre of the beam
bottom to enable two beams to integrate. The lap splice position
was arranged in the center of the beam with the bond length of
bars at lOd, 20d. and 40d, respectively, where d was the diameter
of the bar. In order to prevent the bars from adhering to the concrete. a plastic vinyl tube was covered around bars and a styrene
board was placed between two beams. The steel bars which were arranged in beams were: main bars of 2-D25 along the lower side of
the beams. 2-D13 bars along the upper side, and stirrups of
[email protected] The fiber reinforced plastic bars made of continuous
fibers such as carbon, glass, aramid, vinylon and plastic resign
matrix were of linear material with a diameter of lOmm.
In order to improve the bond behaviour between concrete and
bars, the surface of fiber reinforced plastic bars must be processed carefully in accordance with their adjacent concrete behaviour. The summary introduction of bars is 1 isted in Table 1. The
tests were based on an evaluation of 29 beams in which bar
surface process. fiber type and bond length were different. The
concrete was mixed with ordinary portland cement. aggregate made
of gravel with a maximum grain size of 25mm and common river sand
by the 60 percent ratio of water to cement. The compressive
strength of concrete for every specimen is shown in Table 2.
The •ethods of loading and •easure•ent
The loading shown in Fig. 1 indicates that two central points
of simple support beams were loaded by two single concentrated
forces produced by the loading test machine with a capacity of
980KN and the load was detected by a load cell of 490KN. In order
to measure the amount of pull-out precisely, the bars were
318
Makitani, Irisawa, and Nishiura
measured with a high sensitive displacement meter which was located at a bar protruding 50mm from the beam end. Also, the relationship between stress and strain of fiber reinforced plastic
bars was obtained from a load and wire strain gauge which
was affixed to the bars.
The tensile force T acting on a fiber reinforced plastic bar
in beam of the specimen was calculated by pure bending moment
M=P·a due to two equal concentrated forces P
T
p . a
(1)
where a is shear span and
denotes the distance between resultant tensile and compressive forces as shown in Fig. 2. This value
of j was calibrated from the beam types of specimens with a steel
bar whose stress-strain relation was already known.
TOE RESULTS AND INVESTIGATION OF TOE BOND TEST
The bond and tension characteristics were obtained from the
tests described in this paper. Table 2 shows the tensile force,
unit bond strength, tensile stress, Young's Modulus and failure
mode at peak loading.
Bond stress-slip curve
Defining the bond stress as an amount in which the tensile
force divided by the perimeter of bars and defining pulling-out
amount as slip,
the relationship between bond stress and slip of
various FRP bars was resulted in Fig. 3 in which the bond length
was lOd and the maximum slip was lOmm. With the above consideration,
the bond behaviour was roughly classified into two groups
as follows; (1). The maximun bond stress occurred at the stage of
very small slip(C2, C4, A2). (2).The maximum bond stress occurred
when the slip was much large(A1. A3, C1. C3, CT1. V1). The difference of the two behaviours was due to the different bar surface
process. The former corresponded to carbon and aramid fiber bars
whose surfaces were covered by grains of sand. The latter corresponded to FRP bars made of carbon, glass and vinylon and aramid
fiber bars in braid.
From the bond stress-slip curves,
it was clear that the bond
FRP Reinforcement
319
stress was divided into the folllowing two kinds; (1). The bond
resistance decreased violently after the maximum bond stress was
reached. This belonged to the carbon fiber bars wound spirally
around it, braid-shape aramid fiber bars and bars covered with
grains of sand. (2). The bond resistance still held after maximum
bond stress was reached. This belonged to the carbon fiber bars
covered with grain sand in surface and glass as well as aramid
fiber bars wound spirally around them. As mentioned above, we observed that the unit bond behaviour could be improved distinctly
if the surface of the fiber reinforced plastic bars was processed.
The relationship between bond length and tensile force at
aaxiaua load
The relationship between bond length and tensile force is shown
in Fig. 4 in which the black marks express the broken bars and the
other marks express the bars which were pulled out from the lap
splice position. The tensile force clearly developed with the increasing of the bond length.For the same type of fiber reinforced
plasic bars, if the tensile force was lower than that of black
marks at maximum loading,
the pulling out of bars occurred.
In
addition, when the slip did not develop in the bond place, the
failure mode was changed from the pulling out type to the rupture
type. For the bond length of 20d, only the bars of glass fiber had
broken. If the bond length was more than 40d, the fiber reinforced
plastic bars of all types had broken. From the relationship between tensile force at maximum load and surface process, we could
discover that the aramid and carbon fiber bars with grains of
sand in the surface in braid shape had a higher tensile force
than the bars without surface process. That was to say, the bond
behaviour could be improved distinctly when the surface of the
bars was covered by grains of sand.
LAP SPLICE JOINT TEST BY BEAM
The specimen shown in Fig. 5 is of the truss type of concrete
member consisting of top and bottom chord members in the beam
320
Makitani, Irisawa, and Nishiura
centre and two anchored ends. The depth of rectangular holes in
the centre was determined by the photo-elasticity test under the
condition that the two chord members were applied by axial force.
The FRP bars were arranged in the central axle of two chord
members reinforced by longitudinal and transversal steel bars. The
lap splice arranged in the center of the bottom chord member was
designed at a length of !Od, 15d, 20d, 25d, 30d, 40d, 50d and 60d,
respectively, where d was the diameter of FRP bar. The slits were
installed in the top and bottom chord members in order that the
tensile forces were not carried by the longitudinal bars.
The plastic vinyl tube was covered on the FRP bar existing in
the place beside the lapped splice in the bottom chord member.
FRP bars made from carbon fiber in pitch and thermosetting
epoxy resin matrix were 8mm in diameter and their surfaces were
processed into the crossed spiral ribs. The tensile strength,
Young's modulus of FRP bars and compressive strength of concrete
were 1860N/mm 2 , 147N/mm 2 and 31N/mm 2 respectively where the material behaviours of the bars were calibrated from the beam type of
specimen mentioned above .
In this paper, eight-specimen tests were performed with various lap splice lengths in the joints.
The aethod of loading and aeasureaent
In terms of the large-scale construction test machine with a
capacity of 2940 KN.
the top and bottom chords arranged in pure
moment position were loaded by 4-point single forces which were
reflected in a wire strain gauge. Considered with this method, the
beam was formed as a truss type and axial force was generated in
top and bottom chords. To measure displacement and crack width in
the slits,displacement meter and rr-type strain gauges were laid.
In addition, in order to measure the FRP bar's strain, strain
gauges were stuck on lap splice joint of bars as well as at the
bottom end of the chord. The tensile force acting on an FRP bar in
the bottom chord in the truss type of specimen was calculated by
the same equation (1) as mentioned in the beam type of specimen.
FRP Reinforcement
321
THE TEST RESULTS AND OBSERVATION
States of cracks and failure
The cracks appeared mainly in slits of the bottom chord ends as
shown in Fig. 6. The reason was the enlarging by pulling-out of FRP
bars in the lap splice joint unti I final fracture. When fracture
occurred, the lap length was changed from lOd to SOd and all fracture modes were pulling-out.
The relationship between tensile stress and displace•ent
The relationship between o and ft is summarized in Fig. 7 in
which o is the crack stretch at the slit of the bottom chord end
and ft is the tensile stress of FRP bars with joint length ranging from lOd to 25d. The amount of o included slip of joint and
elongation of the FRP bar's unbond position. In Fig. 7, it is seen
that there was little difference on ft-o
curves in which lap
length was 20d and 25d, respectively. FRP bar with lap splice
length of 15d was selected as the analysing specimen. Starting
from the free end Fl to the bond end FL of the FRP bars with lap
joint length of 15d, the stress-strain curves are shown in Fig. 8.
Also, in the same figure, the strain at slit FL and tensile stress
curve is included. Then, it was found that the tensile stress of
FRP bars was approximately 500 N/mm 2 and the strain between F3
and F1 increased rapidly which meant the larger slip of bars developed. Second, the F5 in the boundary of the unbond region and
the splice region showed a consistency with F1 at the slit after
the slip occurred. The reason was that concrete resistance in F1
region and bond resistance of FRP bars were lost after the slip
had developed. Third, the FRP bars with splice length of 15d and
tensile stress of 400 N/mm 2 were taken as analysing specimen. The
distribution of strain recovering resistance against slips was
shown in Fig. 9. Hence, the stress level measured at free ends in
lapped splice position was small. It had been proved that the bond
resistance remained relatively large corresponding to its small
strain. The theoretical results obtained from the theory <4 l of
fiber pull out and bond slip were represented by a solid line in
Fig. 9. Hence, it was found from this analysis that the theoretical
values clearly agreed with the experimental ones.
322
Makitani, Irisawa, and Nishiura
The bond
resistance in splice region
The relationship between tensile strength ft and lap splice
length Is is drawn in Fig. 10 when FRP bars were pulled out. It was
found from this figure that the tensile strength increased as the
splice length increased. Then the unit bond strength!' fu in the
splice region was calculated according to the following formula:
!'
f t
fu=
. d
4 I s
( 2)
where d is the diameter of FRP bar.
The relationship between lap splice length and bond strength
calculated from formula (2) was derived as shown in Fig.11. Table
3 summarized the values of unit bond strength for various lap
splice lengths which were obtained by the present test. The conclusion is that bond resistance in the splice region increased as
lap splice length increased. but that it decreased for lap splice
length from 40d to 60d. For these test results, the bond strength
of the FRP bar could be approximately represented as followsct> :
!' fu=O. 77
I s
d
-0.011(
I s
d
( 3)
where Is is the length of lap splice.
The bond strength of FRP bars in the lap splice which was calculated by equation (3) was indicated in Fig.11. It is recognized
from this figure that the theoretical value is approximately
agreed with the test results.
CONCLUSION
With beam and truss type of specimens in this study,
the bond
behaviour of FRP bars was researched and the following conclusion
was obtained.
1. The bond strength of FRP bars was found to increase distinctly
if the surface of bars was processed in spiral type and covered
by sand.
2. With the bond length over 40d, FRP bars were broken even if the
surface was processed.
FRP Reinforcement
323
3. With the length of lap splice more than 40d, the full tensile
force could be transferred.
4. The unit bond strength tended to increase with the increasing
in bond length and lap splice length, but decrease for lap splice
length between 40d and 60d.
5. For the bond tests of beam and truss types,
it was not necessary to reinforce the ends of the FRP bars.
This test method for
researching the bond characteristic was wholly effective.
ACKNOWLEDGEMENT
The authors wish to acknowledge the assistance of Shimizu Construction, Kumagaiguimi Co., Okumuragumi Constrution Co .. and Mitsubishikasei Co .. which provided the FRP bars.
REFERENCES
(1).
Design for Earthquake Resistant Reinforced Concrete Building
Based on Ultimate Strength Concept,
Architectural Institute
of Japan. 1988.
(2). P.M. Ferguson and J. N. Thompson,
Development Length of High
Strength Reinforcing Bars in Bond, ACI Journal. Pro c. Vol. 59,
No.7, July, 1962.
(3). E. Makitani, K. Machida and I. lrisawa, Bond Characteristics
of Fiber Reinforced Composite Material Obtained from Beam
Tes ls,
Proceedings of the Japan Concrete lnst i tute, Vol. 13,
No.2, 1991.
(4). A. E. Naaman. eta!., Fiber Pull out and Bond Slip I : Analytical Study, Journal of Structural Engineering, ASCE, Vol. 17,
No.9, Sep. 1991.
Makitani, Irisawa, and Nishiura
324
TABLE I-CONSTITUTION AND SHAPE OF FRP BAR
Fiber
Denardnat ion
of speci•ens
C-1
C-2
0
Shape of sur face
and cross section
of FRP bars
Fiber
by volu•c
m
C-4
IOd
20d
D.
[lfl')t:;.0'f.!liSl
0
0
~~
@
0 0 0
0
6
~
0 0 0
tB
Canst i tut ion of FRP
bars
30d
0 0 0
0 0 0
Ql]]]QD
Carbon
C-3
Bond length
content
Carbon fiber is •ound
spirally round a bar
Sand is stuch to the.
surface of a bar
A bar is canst i luted
of seven strands
lti sled
A bar is for•ed to
the braid on which
sand is stuch
60
llra•ld
A-1
0
[]]1t[lj']
0
0 0 0
Aruld fiber Is wound
spirally round a bar
A-2
0
~
83
0 0 0
A bar Is Coned to
the braid on which
sand is stuch
A-3
0
0
~·EB
0 0 0
A bar is for•cd to the
onrn:IDD 0
0 0 0
Glass fiber is wound
Glass
G-1
Viny I on
V-1
---Ster!l
UIO
0 moo 0
0
-
0 0 0
-
--I
0 0
-
braid
spiral Jy round a bar
The crossed vi nyron
fibers are wound
round a
b<~r
ocror•cd bar
FRP Reinforcement
TABLE 2-RESULTS OF BOND TEST
Bond
I engt h
Spec i-
Fiber
•ens
(oa)
Tensile
stress
Bond
st rcngth
Fa i I ure
•ode
force
(kN)
C I
c2
c3
C I
26.
33.
29.
31.
0
1
I
1
134
127
67
136
A I
A2
A3
28. 9
30. I
21. 3
112
122
91
GI ass
GI
30. 9
97
881
Vi nylon
vI
21. I
48
505
-
C I
c2
c3
C I
26.
28.
21.
33.
6
9
9
2
17
79
31
81
-
II.
10.
5.
I 0.
A I
A 2
A3
31.3
31. 8
30. I
60
96
60
-
9. I
12. 3
9. 0
GI ass
GI
26. 0
85
770
-
rupture
Vi nylon
vI
28. 3
18
-
1. 0
pull out
Steel
D I0
28. 8
10
562
-
rupture
cI
c2
c3
8
6
9
0
15
51
13
63
-
C I
28.
26.
26.
30.
-
Carbon
13.1
13. 6
I. I
15. 9
A I
A2
A 3
25.1
29. l
29. I
51
12
10
-
16.0
18. 1
11. 9
G1
29. 5
56
-
15. 0
VI nylon
v1
29. 5
31
-
·a. 8
Steel
D 10
27. 1
12
539
4 0
Aramid
Carbon
1 0
ldax iii!UJ
tens i I e
(N/n 2 )
Carbon
2 0
Co•prcss i ve
strength or
concrete
Aram id
Ara.id
Glass
(N/n 2 )
1190
1150
1090
1100
1020
-
-
(N/n 2 )
-
rupture
"
5.1
pu II out
-
rupture
-
"
"
1. 0
pull out
-
rupture
-
"
I
6
6
6
pull oul
"
"
"
"
"
"
pull out
"
"
"
.
"
"
N
N
rupture
325
326
Makitani, Irisawa, and Nishiura
TABLE 3-DATA OBTAINED WITH TRUSS TYPE OF SPECIMENS
Lapped
Spesi•en
Max i111u111
splice
length
load
(co)
(kN)
C-M-10
8(10d)
18.0
C-M-15
12(15d)
C-M-20
16(20d)
C-M-25
Tensile
strength
or
fRP bar
(N/om')
Unit bond
strength
of FRP bar
Fai Jure
- 1ode
(N/n 2 )
346
8.7
30. 9
592
10.2
Pull-out
H. 6
8H
11.0
Pull-out
20(25d)
60.8
1170
11.8
Pull-out
C-M-30
24 (30d)
17. 6
1490
12.5
Pull-out
C-M-40
32(40d)
107.0
2060
13.0
Pull-out
C-M-50
40 (50d)
116.0
2220
11.3
Pull-out
C-M-60
48 (60d)
101.0
1950
8.2
Pull-out
un1l:1111l
'I
~
fRP bar
Pull-out
I
Unbond
/
Unbond
Bond
'
)v.p
lv•P
150 I !50
300
1100
2-013
Metal rotatory device
~
Reinforce1enl
in I oad i tud ina I
direction 2-025
\
Web rei nforce1ent
[email protected] 50
18o
300
I
16o
.
Fig. 1-Equilibrium requirements in the beam-type specimen
FRP Reinforcement
l
1p
p
~
_____.. , c
T
i V=P
.I
a
I•
I
Fig. 2-Details of specimen
20r-----------------------------~
A 2
:,.,,...........
I
....... ,..,.
--------- ....
...... ------~--1
..;.... . . . . .""..
.... ' '
.............. . .
-------- -------.~~.--::-:-::-."1- . .
..a.:.·.:
"'
"'~....
.....
--- .............. __
"'
-::! 10
0
..c
c3
0
2
3
4.
5
6
7
8
9
10
Slip(DUD)
Fig. 3-Unit bond stress and slip curves for bond length of lOd
327
328
Makitani, Irisawa, and Nishiura
150
I
I
I
I
I
I
I
I
I
I
I
I
I •
..-..
:z:
::.:::
Q)
c.J
I-<
100
0
<;-..
•
-~I
/~$
'fj.
.7
'I
I
I
..;_).
'
Q)
I
I
),(r--;,---
•
tl'!"'
'
--1~~~
IZl
c::
Q)
.....,
•
50
E
:::l
I
I
I
'/
•
I
•
I
/
I
,,,/~-··-'
---~~~~-------~---
E
><
:::;;;
"'
I
I
/
!j(
0
10
20
4D
13ond length (em)
Fig. 4-Relationship between maximum tensile force and bond length
Fi:P bar (If)
D'@lOO
[email protected]
= tO
2-D\0
,===~
[email protected]
Z-DIO
r=-=~
1
1
Lu1n. ol lappd
~lboi~U~bDI4
1
4
*?++fF!
I IDD
100
IDD
"'
~"f
IDI
1011
1G 10
?
80
10
1800
tel
to
II
1D
"
"'
IDO
100
101
,,,
IDO
Center
Crou aectlo•
Slit(\Zn 1idth and 50u deptb)
!'
!
[;
tiD
100
1100
t>
"'
~ D~
LJ!LJ
Dlunsinl or IPOCiUl
Fig. 5-Details of specimen for lapped splice test
329
FRP Reinforcement
Fig. 6-State of a crack at the bottom slit
1000.-------~------~----~~----~~------
..
'
''
'
800
600
f t
N/mm
:
'
'
'
2 5d
---------- ,_- .1------------ .J.------------ L..------------ ".. --~--------
ll
l:
i
i
------------1------------;--
::
,
.. ~:
i
7---,/zoct
r------------ r-----------
I
I
:
:,:
,,
I
:
I
, .... -.---~--15 d
~:_ :r-~----------~-----------14 0 0 ----------- -)0~----.--44r~="::
:
:
:
2 )
._..
200
_,-''
I
I
I
. . '..-""
1 0 d:
I
I
:
'
'
'
1.5
2. 0
--- --~-1------------ +----------- +---------- +----------''
'''
0
I
I
:
0. 5
1.0
_______. o ( nm
2. 5
)
Fig. 7-Tensile stress of FRP bar and crack stretch curves at slit
330
Makitani, Irisawa, and Nishiura
700.----.----~-----r----~----.-----.---~
'
-;
600
''
I
~------:
---------:---------:----F-;-I
I
I
I
I
~---------:-------- -~---------
F4 F
- ______ ,.j _________ ,., ________ _;
--
s!
'''
'
'
''
---~---
'
'
---- --~------- -~--------
'
''
'
''
''
'
'
--------,.---------,.-------- _,..---------,.
I
I
I
I
I
I
I
I
'
'
50 0
100 0
FL
I
I
I
'
'
'
I'
I
I
---------
r--------- r --------- r--------- r--------- r-------- -r- --------
0
I
I
I
I
'
1 50 0
2 sao
2000
--+e
(la-
3 aa o
3500
6 )
Fig. 8-Tensile stress and strain curves for lapped splice length of lSd
F5
FL
·~· •
Unhand
50 0
400
ft
-- ''
-~-·--
'
(N/mm 2 )
300
1200
.
------------------- ..' ---'''
F3
F2
F1
•
• •
Length of lapped splices
''
''
'
-"------:-:::::t! ~~E!_e_c_~ ~c:~ 1-, ~-·-~-'!.r -'
'
''
''
'
''
''
'
'
''
'
'
'
'
- - - \ . . - - - - - - - - - - - - - - - - - - - - - - .L - - - - - -L..----- ------ !"--- ---------4------ t-''
''
'
'
'
'''
'
'''
1
!
''
'
'
'
---:----------------------- ~' ------:-------------r----------i------t--
10 0 --
0
F4
I
I
I
'''
'''
''
''
''
-~----
I
I
I
I
I
I
I
FL
I
''
------------------ ..' ----- -t------------ .. --------I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
FS
F4
F3
'''
''
-~------~-'
''
'
''
'
''
•
F2
F1
Posi lion of Measured Strain
Fig. 9-Distribution of strain for tensile force at pulling-out
FRP Reinforcement
3000
2500
ft
( N/mm 2 )
I
331
''
'
''
---- ~---- ~-- -- ~---- ~-- --~---- ---- -~--- ----- -~--- -- ----~' ---: : l : :
:
~
:
1
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
2000 ----r- ---r ---- ~--- -~- --- ~ ------------------r-- ------·---1500
1000
,
I
I
I
:
:
:
:
:
:
:
:
:
:
:
:
I
I
:
:
I
I
:
:
I
:
I
:
-------- ~ --------- ~-- ------ -~- ---
:
I
- - --r- - - - - r--- - -
:
-t-- - - - - - - -
:
l
-r----- -----'"---- -----,.----
l
----r---::!IF----~--- -r ----r---------r---------r---------r ---: l
1
1
I
I
500
--·
I
---- ~---- ~-- -- ~--- -~l : r---- ~
_,_ __ -
I
I
'
:
I
•
I
:
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
:
:
•
'
0
'
'
10d 15d 20d 25d 30d
40d
~Lap
50d
60d
Splice Length
Fig. 10-Relationship between tensile stress at pull-out
and lapped splice length
14
''
''
'
13 -.--- '.. ---- '.. --------- '.... I
I
I
I
I
I
I
I
I
I
I
I
:
l
l
: :
1
11
10
~---
~
I
I
I
I
I
I
I
I
I
1
I
I
-------- '.. --------- ........
-'
I
I
:'
:
--r~----
----r --------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
----: ---ct, --:----:----:---------:---------:- ----- ...
• --r----r--I
''
8 ----t0
''
''
'
:I
'
'
'
Ltu-o • 77 .!...!.._
-o • ou(
d
-~-
~
d
I
) 2 :I
-- -~ -------- -r ---------r------
1
I
I
I
I
:
:
:
:
I
I
I
I
I
lOd 15d 20d 25d 30d
I
:
I
40d
''
''
'''
---------~----
i
:
:
I
:
:
:
. . . -----,.---9 . . ---,... ... --... I --- ... ,..I ---- Ir--- --r---------,.--.
I
I
I
0
'
'
'
-:---------:----- :
•
r----:--:
---- ----r-----r: --
12 ----r ---'ffu
(N/mm 2 )
'''
'''
50d
-~----
'
'
''
-·---'''
''
60d
- - + L a p Splice Lengtb
Fig. 11-Relationship between unit bond stress and lapped splice length
SP 138-21
Bond of GFRP Rebars to
Ordinary-Strength Concrete
by M.R. Ehsani, H. Saadatmanesh,
and S. Tao
Synopsis: An overview of a study on bond of Glass-fiber-reinforced-plastic
(GFRP) rebars to concrete is presented. The 78 specimens to be tested include
several variables, such as the mode of failure (i.e. pullout or splitting), concrete
compressive strength, bar diameter, clear cover distance, and top bar effects.
In addition, the effect of the radius of bend for hooked bars and the extension
on the hooks is being investigated. The study is currently in progress and the
results of the specimens tested to date are presented. Preliminary results
indicate that the top bar effect observed for steel rebars is also present for GFRP
bars. For hooked bars, larger radii of curvature increases the failure load of the
bars significantly.
Keywords: Bending; bond strength; compressive strength; cover; fiber
reinforced plastics; glass fibers; pullout tests; rebars; slippage; splitting
stresses
333
334
Ehsani, Saadatmanesh, and Tao
M. R. Ehsani is Associate Professor of Civil Engineering at the
University of Arizona. He is a Member of ACI Committee 408 where he chairs
a Subcommittee on Bond of FRP Rebars. As a member of Committee 440 he
chairs the Subcommittee on State-of-the-An Repon. Dr. Ehsani is a registered
professional engineer in Arizona and California.
ACI member Hamid Saadatmanesh is Assistant Professor of Civil
Engineering and Engineering Mechanics at the University ofArizona in Tucson.
His research interests include rehabilitation and strengthening of structures and
the application of fiber composites in civil engineering structures. He is the
Secretary of AC/ Committee 440, FRP Tendon and Reinforcements.
S. Tao holds B.S. and M.S. degrees in Civil Engineering from Beijing
Polytechnic University. He is currently enrolled as a Ph.D. student at the
University of Arizona. His research interests include advanced composite
materials.
INTRODUCTION
Corrosion of reinforcement has always been a major problem with
concrete structures. This is particularly serious if the reinforcement is subjected
to high stresses. A combination of high stress and intense corrosion will
produce stress concentrations that may result in rupture of the reinforcement.
Billions of dollars are spent every year to replace or repair concrete structures
that are deteriorated due to the corrosion effects of salt, acids, and other
aggressive elements.
An effective approach to eliminate the corrosion problems in concrete
structures would be to employ corrosion-resistant plastic rebars instead of steel
rebars. Glass-fiber-reinforced-plastic (GFRP) bars offer great potential for use
in reinforced concrete construction under conditions where conventional steelreinforced concrete has resulted in unacceptable serviceability problems.
FIBER REINFORCED PLASTIC REBARS
Fiber-reinforced-plastic (FRP) rebars offer several unique advantages for
solving many engineering problems in areas where conventional materials do not
perform well. Unlike steel rebars, FRP bars are unaffected by electrochemical
deterioration (1). FRP rebars resist the corrosive effects of acids, salts, and
FRP Reinforcement
335
similar aggressive materials under a wide range of temperatures. Other features
of FRP rebars include: High Strength - The ultimate strength of FRP rebars
greatly exceeds that of steel rebars. The ultimate strength of GFRP rebars is
reported at over 100 ksi (690 MPa). The high strength of these bars provides
adequate reinforcing in concrete structures and makes them ideal for use in pretensioned and post-tensioned concrete members. The modulus of elasticity of
GFRP rebars is about one-fourth that of steel. Therefore, attention must be paid
to deflections of structural members reinforced with GFRP rebars.
Fatigue - The fatigue behavior of FRP rebars is very good as compared
with steel rebars. The plastic bars do not fatigue when stressed to no more than
1/2 of their ultimate strength (2).
Low Weight - The specific gravity of FRP rebars is one-fourth that of
steel. The light weight of plastic rebars results in reduced transportation costs
and easier handling on construction sites.
Non-conductivity and Thermal Expansion - Plastic rebars possess
excellent electrical insulating properties. Therefore, they are ideal for use in
airports to help solve radar interference problems, and other applications, such
as hospitals, where presence of steel could cause interference with electrical or
magnetic fields. The coefficient of thermal expansion of GFRP bars is 5.5 x 106 in/in/"F, sufficiently close to that of concrete.
Economy - The cost of GFRP rebars is slightly higher than that of
epoxy-coated steel rebars. However, other features of GFRP rebars, such as
light weight and longer service life, could make them more economical than
epoxy-coated steel rebars.
In spite of the aforementioned advantages, before fiber composite rebars
can be widely accepted for field applications, additional studies must be
undertaken to address several key issues such as creep, relaxation, bond,
ductility, and fatigue behavior of concrete members reinforced with GFRP
rebars.
EXPERIMENTAL STUDY
An extensive investigation of the bond strength of GFRP bars to concrete
is currently underway at the University of Arizona. The pullout and splitting
modes of failure under static loads for some 78 specimens will be studied. In
addition, 24 specimens containing 90° hooks will be tested. All of these
specimens have been constructed, and twenty have been tested to date. The
testing of the remaining specimens is in progress and is scheduled to be
336
Ehsani, Saadatmanesh, and Tao
completed by the end of 1992. A more detailed description of the different
aspects of the experiment is presented in the following sections.
The primary variables for the beam specimens include:
1. Failure Mode - The primary modes of failure for bond tests include
tensile, pullout and tension splitting. Some GFRP bars may fail in tension
outside of the concrete specimen when the ultimate tensile strength of the rebar
is reached. The pullout test is a good measure of the necessary anchorage
length of a bar when it is embedded away from the edges and in the middle of
a mass of concrete. The tension splitting failure mode is more commonly
observed in structures and is predominant for bars with small concrete cover.
Eighteen of the specimens are expected to fail in the pullout mode while thirty
six are expected to fail by tension splitting.
2. Concrete Compressive Strength - The current models for bond
strength of reinforcing steel to concrete are expressed as a function of v'r c.
This indicates that bond failure is initiated by tension failure of concrete. In
general, the deformation patterns on GFRP bars are not as pronounced as those
on steel rebars. Furthermore, the modulus of elasticity for GFRP bars is
typically about 114 that of steel. It is not clear how the difference in mechanical
properties and deformations between GFRP and steel rebars influence the mode
of failure. For these reason, the concrete compressive strength will be varied
at 4,000 and 8,000 psi (28 and 55 MPa). The results presented in this paper are
for those specimens constructed with 4000 psi concrete.
3. Bar Diameter- It is well recognized that the size of reinforcing steel
influences its bond characteristics. The interlocking mechanism between the
aggregate and the ribs or lugs of the GFRP bars may be quite different. Three
common size reinforcement, No. 3, 6, and 9 are used for this investigation.
4. Embedment Length - Straight bars with large embedment lengths fail
by tensile failure of the GFRP bar. For shorter embedment lengths, failure is
usually controlled by excessive slip of the rebar. In order to study the bond
strength over a wide range of service and ultimate loads, the specimens are cast
in groups of three, each having a different embedment length.
5. Clear Cover Distance - The more exact models for calculation of
development length, such as those included in the ACI 318-89 Code (3), do
include the effect of the clear concrete cover, ~. Clearly, larger concrete cover
results in a delay of the splitting of concrete and will enhance the bond
performances. Therefore, for the specimens expected to fail by tension
splitting, the ratios of the clear cover to bar diameter (i.e. ~/db) will be 2, 4,
or 6.
FRP Reinforcement
337
6. Concrete Cast Depth - There is a consensus among researchers that
cement paste and air bubbles which are trapped under a horizontally cast bar
adversely affect bond strength. To investigate this phenomenon, known as the
"top bar" effect, three concrete depths of 8, 24, and 40 inches (200, 610, and
1010 mm) have been selected for the pullout specimens. Some of the beam
specimens also include top bars. The wide range of this variable is expected to
provide sufficient information on the effect of casting position (i.e. top bars) on
bond strength.
For the hook specimens, in addition to concrete compressive strength,
and bar diameter noted above, the effect of the radius of bend and tail length are
being investigated. As shown in Fig. 1, the lead-in portion of the bars were
always protected to prevent them from bonding to concrete. The standard tail
extension for a 90° hook is 12 times the bar diameter. To examine the potential
benefits of longer tails, an extension of 20~ was also used.
For steel rebars, the ACI Code limits the minimum bend radius to
prevent fracture of the bar at the bend. For GFRP bars, however, the bars are
cast in the final position. Consequently, using a smaller radius should not cause
any concern regarding the fracture of the bar. However, due to low shear
strength of the GFRP bars (transverse to the longitudinal axis of the bar), sharp
bends could lead to a reduction in the strength of the hook. In order to examine
this phenomenon, the hooks were cast with an inner radius of bend of either
three times the bar diameter or zero.
Materials
The concrete used for the study is normal weight and was delivered to
the laboratory in a ready-mix truck. The mix proportions for cubic yard of
concrete include: 555 lb (250 kg) Type I cement, 1865 lb (850 kg) aggregate
with a maximum size of 1 inch (25 mm), 1244 lb (565 kg) sand, 308 lb (140
kg) water, and 28 oz of MBL82 Water Reducer produced by Master Builders.
The results of 6 by 12 cylinder tests indicate a compressive strength of 5080 psi
(35 MPa) and 5780 psi (39.8 MPa) for No.3 and No.6 specimens, respectively.
The reinforcing bars are made of E-glass fibers in a polyester matrix and
include a small quantity of fibers which are wrapped around the longitudinal
fibers in a spiral pattern. The function of the wraps is to induce deformations
on the surface of the rebars for enhanced bond behavior.
338
Ehsani, Saadatmanesh, and Tao
Test Setup and Instrumentation
The specimens are being tested in the steel reaction frame shown in Fig.
2. Conduits have been used to prevent the bonding of concrete to the
reinforcement near the ends of the rebar. This test setup results in compressive
forces near the top of the specimen, similar to the case in "real" beams. The
loads are applied using a hydraulic jack.
The specimens are loaded
monotonically and at a consistent rate of loading to failure.
The pullout test specimens will be tested in a simpler frame. In this
case, the jack will rest directly on the specimen, applying a compressive
reaction force to the concrete block. However, because the first few inches of
the bars are isolated by use of conduits, the effect of this compressive force on
the bond stress will be minimal.
The applied loads are transmitted from the jack to the GFRP rebars by
means of a specially constructed set of grips. In earlier studies, the design and
development of suitable grips to pull GFRP rebars embedded in concrete have
presented some difficulties (4). Because glass fiber based composites are very
weak for loads applied transverse to the fiber direction, the portion of the GFRP
rebar in the grip must be protected against crushing. The grips used in this
study are made of sand-coated steel plates and have proven quite effective for
all bar sizes.
The magnitude of the applied load is recorded electronically, using load
cells. The slip of the specimens at the loaded and the free end are measured at
different loads. The slip is measured as the movement of the bar relative to a
fixed reference point, i.e. the concrete specimen, by attaching LVDTs and dial
gages to the loaded and free ends of the reinforcement. The slip measured in
this manner consists of three components, namely: a) elastic deformation of the
bar, b) localized displacements caused by the crushing of the ribs of the rebar,
and c) any rigid body motion of the bar relative to the concrete. To be precise,
the true slip is the third component mentioned above. However, the separation
of these various components during the experimental program is very difficult.
Moreover, because of the low modulus of elasticity and low shear strength of
GFRP bars, the contribution of the first two components to the total measured
slip is fairly large.
Test Results
The testing program is currently in progress. As of this writing, twenty
specimens have been tested. All of the specimens were constructed with 4,000
FRP Reinforcement
339
psi (28 MPa) concrete. Of these, twelve were straight bars, and the remaining
eight were 90° hooks. It is noted that the analysis of the data is being
performed along with the testing. Consequently, only the preliminary results
of these tests are presented below; a more detailed analysis of the data will be
conducted after all specimens have been tested.
a) Straight Bars -- The results of two of the specimens are compared
below. These are No. 3 bars which were expected to fail in splitting mode.
The embedment length for each bar was 4 in. (100 mm), and the ratio of the
clear cover to the bar diameter was 2.
In the first specimen, the rebar was cast with a clear cover of 0. 75 in.
(19 mm) from the bottom of the specimen. Figure 3 shows the load vs. slip for
this specimen. As expected, the loaded end slipped significantly more than the
free end. The solid line represents the net slip, defined as the difference
between the slips at the loaded and free ends. The specimen failed by tension
failure of the rebar at a location 11 in. (280 mm) from the concrete face
(between the grips and the loaded end) at a load of 7,885 lb. (35.05 kN).
The bond stress, u, is defined as:
u =
T
(1)
1f•d•l
where, T = the maximum applied force, d = the diameter of the rebar which
is 0.375 in. (9.5 mm) and 1 = the embedment length which is 4 in. (100 mm).
With these values, the average bond stress for this specimen was 1,673 psi (11.5
MPa). For steel reinforcing bars the allowable bond stress is defined as the
smaller of that associated with a slip of 0.01 in. (0.25 mm) at the loaded end or
0.002 in. (0.05 mm) at the free end (5). Following the same guidelines, the
allowable bond stress controlled by the loaded end was calculated as 287 psi
(1.98 MPa).
In the second specimen, the rebar was cast with approximately 11 in.
(280 mm) of concrete below the bar and a clear cover of 0. 75 in. (19 mm) from
the top of the specimen. Figure 4 shows the load vs. slip relationship for this
specimen. Failure of this specimen also occurred outside of the concrete block
at a distance of 32 in. (810 mm) from the loaded face. The failure load was
8,592lbs. (38.2 kN), corresponding to a bond stress of 1,823 psi (12.57 MPa).
The allowable bond stress was 190 psi (1.31 MPa) and controlled by the free
end of the bar. These two tests indicate that the allowable bond stress in the top
bar is 66% of that for the bottom bar. This is consistent with the current
understanding for the lower bond stresses in steel top bars.
It is interesting to note that none of the specimens which contain No. 3
340
Ehsani, Saadatmanesh, and Tao
or No. 6 bars have failed by splitting. This is most likely due to the smaller
deformations (i.e. ribs) on the GFRP bars compared to steel rebars. For all No.
3 specimens, the GFRP bars broke in tension outside of the concrete. For the
No. 6 specimens, in addition to braking of the bar, two other modes of failure
were observed, namely bar pull-out and slipping. Figure 5 shows the load vs.
slip at the loaded end for two No. 6 bars with a development length of 12 in.
(305 mm). One bar was cast at the bottom of the beam with a clear cover of
2db; the other was cast as a top bar with 22 in. (560 mm) of fresh concrete cast
below it and the same cover of 2db above. The top bar reached a load of 21. 5
kips (95.6 kN) and failed by pull-out and a complete loss of bond. The bottom
bar reached a maximum load of 29.1 kips (129.3 kN). At that point, the
loading was stopped because the resin in the bar was starting to break between
the grips and the specimen. The top bar also had a lower stiffness compared to
the bottom bar. This has been also observed for other top bars.
b) Hooked Bars-- As noted earlier, eight specimens for 90° hooks have
been tested. The schematic of a typical specimen is shown in Fig. 1. These
specimens were constructed with 4000-psi (28 MPa) concrete and include No.
3 and No. 6 bars. For these specimens, only the slip at the loaded end of the
bar was measured. The results for three of the No. 6 hooks are shown in Fig.
6 and a summary of failure loads for all specimens is presented in Fig. 7.
The data indicate that the specimens with larger radii of bend had a
higher stiffness, nearly twice that of specimens with sharp bends. Two distinct
modes of failure have been observed for these specimens. The specimens with
larger radius of bend failed by breaking of the rebar outside of the specimen.
For specimens with sharp bends, the bars broke at a point very close to the bend
inside the specimen. The failure loads for these specimens was always less than
that of their companion specimen with larger bends. Furthermore, as shown in
Fig. 7, the larger tails did not have any significant effect on the failure load of
the hook.
Therefore, these tests indicate that GFRP hooks must be
manufactured with a radius of bend at least equal to the values recommended by
the ACI Code 318-89.
Although the bond strength can be defined for a maximum permissible
slip, this issue does require further examination. GFRP bars in general undergo
larger slips due to the lack of well-defined lugs on their surfaces. It is hoped
that as more data become available from this study, they can be combined with
the results of tests conducted by others to develop new design guidelines for
bond of GFRP rebars to concrete.
FRP Reinforcement
341
SUMMARY AND CONCLUSIONS
Out of a total of 78 specimens, the first twenty have been tested.
Testing of the remaining 58 specimens is currently in progress. Data from these
and tests conducted by others will be used to develop new design
recommendations for bond of GFRP rebars. Preliminary results indicate that
the bond strength of GFRP bars is lower than that of steel. The limited data
indicate that top-bar effect is also present for GFRP rebars.
For hooked bars, the longer the radius of the bend, the higher the
stiffness of the specimen. Furthermore, bars with smaller radii of bend failed
by breaking of the bar at the bend at very small loads. It is recommended that
until additional test results become available, the bond radius of GFRP rebars
be taken as equal to that for steel hooks. The longer extension of hooks,
beyond 12db, did not have any significant effect on the strength of the hooks.
ACKNOWLEDGEMENTS
The authors would like to thank the support of International Grating,
Houston, Texas, for donating the GFRP reinforcing bars used in this study.
REFERENCES
1.
"Kodiak FRP Rebar," (1985).
Texas.
International Grating, Inc., Houston,
·2.
Marshall Vega Corporation, "Fiberglass Reinforced Plastic Rebar," Data
Sheet, Marshall, Arkansas.
3.
ACI Committee 318, "Building Code Requirements for Reinforced
Concrete (ACI 318-89)" American Concrete Institute, Detroit, 1989.
4.
Faza, S.S. and GangaRao H.V.S. (1990). "Bending and Bond Behavior
of Concrete Beams Reinforced with Plastic Rebars", Transportation
Research Record 1290, 185-193.
5.
Park, R., and Paulay, T. (1975). "Reinforced Concrete Structures," John
Wiley & Sons, New York, pp. 407-409.
342
Ehsani, Saadatmanesh, and Tao
R
t
~ I a~C"T ~T
lt
'R
II
l
'
(a) General View of Beem Specimens
(b) General View of Hooked Bar Specimens
Fig. 1-General view of specimens
STEEL FRAME
CONCRETE SPECIMEN
T
BAR BEING
TESTED
FLOOR REACTIONS
Fig. 2-Test frame
343
FRP Reinforcement
LOAD vs. SLIP for 38482
SLLp lmm)
J.O
6.0
9.0
0
,.;
.,a.
o-
•z
Iii.>:
• .>
.,>:a
-.;.
a
.,;
a
a
o+------------r------------r------------r------------+0
o.o
0.1
0.2
o.J
0.1
SLLp lLnl
Fig. 3-Load versus slip for specimen 1
LOAD vs. SLIP for 384T2
SLLp lmm)
J.O
6.0
9.0
q o.o
12.0
~+-----------~-----------L-----------L----------~-.
z
..>:
..,Oa
0
q-.,
!il g
•
.J
.J ..
0
.,;
0
0
"+---------,----------r---------r---------,----------+0
0.0
0.1
0.2
O.J
0.1
o.s
Sli.p li.nl
Fig. 4-Load versus slip for specimen 2
344
Ehsani, Saadatmanesh, and Tao
LOAD
vs.
SLIP
for 4-6bl2Lb2
SLIP lmml
ro._o______~J~.O--------~&._o________9L.0--------~12_.o________l~5.0 0
~
"/"'
"a.
.~
.... //////
.X
-o
o.O
CI
...J
' .··
.•· .
0
a:
oCI
•...J
0
,/
5!
...·
.
..
.;
0
/
.
a:-
..····
SPEC. : 1-66 12TB2
p
p~ =~~JI E-."l.~
~~~ .....
.. ....."lV
..p,...
0~-~~~~···--------~r-----------r-----------~----------+d
I
,•'
0.15
0.00
0.15
O.JO
0.60
SLIP !Lnl
Fig. 5-Load versus slip at the loaded end for top and bottom No. 6 bars
SLIP lmml
0.0
Q
9.0
6.0
J.O
12.0
~i+---------~--------~------~---------L~
0
a:o
o'
..J~
o.o
0.1
O.J
0.2
0.<
0.5
SLIP lcnl
Fig. 6-Load versus slip for three No. 6 hooked bars
FRP Reinforcement
90-DEGREE HOOKS
!7&l
12-3
E2l
12-0
EEa
20-3
O+························································+·V.~···-····~~1··········1
~
Bar Diameter
Fig. ?-Comparison of the date for the hooked bars
345
SP 138-22
FRP Tensile Elements for
Prestressed Concrete
State of the Art,
Potentials and Limits
by F.S. Rostasy
Synopsis: FRP are new materials for structural engineers.
Hence, an overview on ~the important properties of fibers, rnatrix resins and composite elements becomes necessary to show
the assets and draw-backs of FRP and to illustrate their potentials and limits. Besides several other fields, the prestress; ng of concrete seems to become a promising fie 1 d of app l i cation of FRP. In prestressed concrete construction the high
strength and the good corrosion resistance of FRP can be optimally utilized. In this field of application FRP can compete
with prestressing steel especially in such cases in which the
corrosion protection of the prestressing steel becomes expensive or remains tarnished by residual risks. The post-tensioning of concrete requires anchorages with a high mechanical efficiency. The rna in avenues of development are discus sed, the
necessary future research is outlined.
Keywords: Anchorage (structural); fiber reinforced plastics; post tensioning;
prestressed concrete; prestressing steels; pretensioning
347
348
Rostasy
Ferdinand S. Rostasy, born 1932, studies of civil engineering
at the University of Stuttgart; Dr.-Ing. 1958; practical work
and research until 1976, from then on professor of structural
materials at Braunschweig, Germany.
INTRODUCTION
Who and what are they, these new advanced and high-strength fiber reinforced plastics? Looking in awe at todays flood of
techn i ca 1 papers, commit tees and symposia on FRP, they seem to
be the whiz kids of structural materials. Are they really, can
they be used for structural concrete construction, and can they
technically and economically compete with proven materials such
as reinforcing and prestressing steels? To give answers to such
questions is the aim of the lecture.
The FRP dealt with here are the un i direction a 1 fiber composites: endless high-strength fibers embedded in a matrix resin
and shaped to rods, strips or strands, in order to reinforce
and prestress concrete as alternative tensile elements.
FRP are not really new materials, though definitely younger
than prestressing steel. And they were not at all developed for
structural engineering purposes. The building industry is the
lucky beneficiary. FRP - then as multidirectional laminates are widely used today for airplanes, automobiles, chemical apparates and for other purposes. Because of their high static
and dynamic strength as well as their very low density, they
have gradually replaced the metallic components in aircraft
construction. Fig. 1 shows which parts of an airplane component
can be made of FRP.
FRP for sports gear are ubiquitous. There is no tennis racket
without FRP. Certainly the young sportswoman from Fig. 2 cannot
be improved by FRP, but maybe her forehand slam can.
FRP are still rather expensive. What then are their assets if
they are used in aircraft construction, a field of very high
safety standards indeed? Fig. 3 shows some of these assets. No
wonder construction industry became aware of them, as they were
looking keenly for alternative materials of high tensile
strength and with a corrosion resistance superior to prestressing steel. Consequently all over the world, though small in
number, structural concrete companies couregeously ventured
into the deve 1 opment of tendons and into pi 1ot app 1 i cations.
Alliances between constructors, scientists and chemical industries were shaped.
FRP Reinforcement
349
ON FRP AND THEIR COMPONENTS
Fibers and matrices
FRP are strange materials to the structural concrete engineer.
He deserves a compact overview on the properties of FRP, ne ither dwelling on assets nor concealing any draw-backs.
There are many types of fibers though only few are suitable for
high-strength FRP tensile elements. These fibers are: inorganic
glass fibers, organic aramid and carbon fibers.
Fig. 4 shows the rna in mechanical properties of the fibers and
of several commercial FRP elements (G, A, C ... glass, aramid,
carbon fiber; UP ... polyester resin; EP ... epoxy resin). All
fibers are ideally elastic and brittle, all are very strong.
Significant differences in the Young's modulus and tensile failure strain are evident.
The fibers are very thin. Their diameter is in the range of 5
to 25 thousandth of a millimeter, depending on the fiber type.
Besides that they are very sensitive to transverse pressure.
Hence they are parallely embedded in a polymeric matrix resin.
Epoxy, vinyl and polyesther resins are the common resins. By
pultrusion the virtually endless FRP elements can be produced.
Now we have arrived at the engineer's material. More on FRP is
found in (l) .
Shapes of FRP tensile elements
For the prestressing of concrete different types of FRP e l ements with respect to the cross- section and surface texture
were developed. There are circular rods, rectangular strips,
braided elements, multi-wire strands. Fig. 5 shows an example
for Tokyo Rope strand.
With the fiber being extremely thin, several ten thousands of
individual fibers will be densily packed in the FRP element. In
Fig. 5 a cross-section of a bar is inserted. We discover irregularities, the fibers are not ideally straight and parallel.
Moreover, there is a resin skin layer whose adherence to the
core is important as we bond the elements to the concrete or to
the anchorage.
350
Rostasy
Short-term axial tensile strength
Now we have to deal with important mechanical properties such
as the short-term axial tensile strength of FRP elements. FRP
are anisotropic materials. Most important for the practical application is the axial tensile strength of the FRP element,
that is the strength parallel to the fibers.
The matrix with its low strength does not significantly contribute to the strength of the composite. Hence the strength and
the modulus of elasticity of a FRP element mainly depend on the
strength of the fibers and on their volume ratio which is in
the range of 45 to 70 volume percent.
Fig. 6 shows stress-strain lines of several commercial FRP
bars. When comparing these lines with prestressing steel the
ratio of fiber volume must be taken into cons ide ration ( 2 to
4). As we might expect, the FRP are - as the fibers - linear
elastic until failure. With respect to strength they may easily
compete with prestressing steel. Their ultimate strain is always less than the one of high-strength prestressing steel.
Will the ideal elasticity and the lower ultimate strain of FRP
impair the ductile behaviour of a structural concrete member? I
anticipate: not necessarily if suitably designed.
Even if the matrix does not contribute to the tensile strength
of the FRP, is has several indispensable functions. To cite a
few: it protects the sensitive fibers against lateral pressure,
abrasion, chemical attack. It equalizes the variability of fiber strength by bond transfer from broken to unbroken fibers.
It must have an ultimate strain, preferably in excess of FRP.
Multi-axial strength of FRP
If we pre-tension concrete the force of the FRP tensile element
has to be transferred by bond to the concrete. The same mechanism prevails in a post-tensioning anchorage. In either case and especially in the- latter - such transfer entails lateral
pressure and shear on the elements surface in conjunction with
high axial tension. This multi-axial stress condition is not
materia 1- specific, though of greater importance for FRP than
for prestressing steel. Tests show that high lateral pressure
and especially surface notches dramatically reduce the axial
tensile strength of FRP. As will be shown later, this fact will
the governing aspect in anchorage design.
FRP Reinforcement
351
On time, stress and strength
In a prestressed concrete member the FRP elements are essentially subjected to long-term static tensile stresses. Hence we
have to be sure that they can carry these stresses safely in
their prevailing environment.
There exists no material which endures high constant tensile
stresses infinitly long. Each eventually fails by creep rupture. The materials differ with respect to the stresses and
times to failure and other parameters.
Even for the high permissible permanent prestressing steel
stresses of the Eurocode which correspond to 75 % of the characteristic tensile strength prestressing steel does fail by
creep rupture within the service 1ife of the concrete member.
This is not the case for some types of FRP, depending on several parameters.
Fig. 7 shows test results for AFRP ARAPREE flat strips surrounded by an alkaline solution to simulate the pH-value of the
pore water solution of concrete ( 5) . The simultaneous app 1 i cation of high stress and of an aggressive solution reduces the
life expectancy of the fiber. The failure time t
is highly
scattered. We conclude that extensive tests under t~e representative environment of. the prospective application are indispensable. Test results of the failure time must extend into the
range of ten thousand hours and more to reliably forecast the
creep rupture line. By suitable forecast models and by statistical evalution, the characteristic creep rupture line with for
instance a certain probability of survival can be estimated.
The latter- in the figure for a 95% survival probability- is
then the characteristic resistance for the choice of initial
prestress.
From Fig. 7 we deduce a really jolting insight: if we use FRP
for prestressing, we have to assume a reasonably safe upper
boundary value of the service lifetime. In Fig. 8 the consequence for admissible initial prestressing force is shown. At
the end of the service life at the age ts the prestress force
Pm s must be well below the characteristic creep rupture force
Fcctk· The ratio Fcdk/Pm s represents the global safety factor 1
at ts. It will be 1n tn'e range of 1,8 to 2,2, depending on the
type of FRP and on the environment.
Only for a few types of FRP sufficient experimental creep rupture data for alkaline environment are available. These tests
show that glass fiber ~omposites with E- or $-glass fibers are
detrimentally affected by alkaline solutions and sea water.
Less affected are aramid fibers, carbon fiber are not affected
by alkaline solutions.
352
Rostasy
Epoxy matrices are superior to polyesther matrices because of
being less permeable. More regarding durability tests is found
in reference (19).
Dynamic strength
The dynamic tensile strength - expressed as the endured stress
range !J.ac = acu - acl, determined in constant stress amplitude
tests, w1th acu (admissible composite long-term stress; upper
stress) as a tunction of load cycles - varies dependent on the
type of fiber.
Tests showed that the fatigue strength of AFRP and CFRP well
exceeds that of high-strength prestressing wire. In contrast to
this GFRP have a lower fatigue strength (1).
Corrosion resistance
As Fig. 3 had promised, that the corrosion resistance of FRP is
high. Without retracting this statement, some clarification is
necessary. Tests show that CFRP are very resistant to any kind
of environment encountered in practice and are superior to
steel.
The use of unprotected GFRP - made with E- or S-glass fibers which come into frequent or permanent contact with a alkaline
solution is risky. The GFRP may fail by corrosion induced creep
rupture. Alkaline solutions are the pore water of concrete and
sea water. If GFRP are used for bonded tendons, protective
measures become necessary. If GFRP are exposed to normal
weather as external and unbonded tendons, they are not endangered by stress corrosion if ultraviolet radiation is obviated.
AFRP are far less sensitive to alkaline solutions than GFRP.
They are therefore suitable for pre-tensioned concrete.
Creep rupture tests must include the possible adverse effects
of the future environment.
Other properties
There are several other mechanical and non -mechanical properties of FRP which are important and must be made known by experiments prior to application. To quote a few: creep and relaxation, uv-embrittlement, hygral and thermal dilatancy. As is
shown in (6), the significant difference in thermal expansion
between FRP and concrete may cause concern in pre- tensioned
FRP Reinforcement
353
concrete e1ements if a 1ow concrete cover is chosen. As to
other items reference is made to (1).
POST-TENSIONING TENDONS
Requirements of performance and strength
Prestressing steel tendons are susceptible to corrosion if they
are improperly protected and/or subjected to an aggressive environment. Consequently, demands arise to develop and apply
tendons which can be inspected and if warranted be replaced.
Especially for severe environments, FRP tendons are an alternative to prestressing steel tendons due to their inherent corrosion resistance. They will however only then be an alternative
if they can meet certain basic requirements with respect to
their anchorages. Such requirements were stipulated for prestressing steel, but are basically also valid for FRP (7). The
main ones are be summarized as follows:
• Neither the short-term tensile strength nor the dynamic tensile strength of the FRP elements should be significantly affected by the tendon-anchorage assembly, TA.
• The creep rupture strength of the TA should not fall significantly below that of the FRP elements, taking environmental
effects into consideration.
By these requirements several goals are presued. The maximum
utilization of the strength of the FRP material, not only on
the free length of the tendon but also at the TA ensure a high
ultimate strain of the tendon. This is a prerequisite for ductile structural behaviour. FRP are still expensive materials.
Hence they must be utilized at as high a stress as possible, an
economic must.
The fulfillment of the above-mentioned requirements requires
extensive testing as the ultimate proof. Nonetheless, also mechanical models may assist in the design of the TA. The test
procedures for FRP tendons are described in (1), (8), (9).
Force transfer and mechanical efficiency
The prestressing force of the FRP element has to be transfered
to the tendon- anchorage assemb 1y which introduces the force
into the concrete. This transfer inevitably necessitates lateral pressure and shear, acting on the surface of the FRP element.
354
Rostasy
FRP elements very sensitively react to transverse pressure and
espec i a11 y to notches which may sever the outer fibers. Hence
the common methods used for prestressing steel for the force
transfer to the metallic housing of the anchorage such as
teethed steel wedges have to be discarded.
It is obvious the transfer of force from the FRP element to the
anchorage enta i 1s a multi- axi a1 state of stress. There exists
only scarce experimental evidence regarding the triaxial
strength. Fig. 9 schemat i ca 11 y shows a regment of the failure
envelope (10). The maximum tensile strength fct of FRP is attained if neither shear nor lateral normal stresses are acting.
As the latter are needed for the force transfer, a certain reduction of the axi a1 tensile strength of the materia 1 must be
expected and tolerated. This reduction cannot be calculated, it
has to be determined in the tens i 1e rupture test on a tendon
with anchorages on both ends of the tendon. The mechan i ca 1 efficiency factor ryA discloses the reduction:
ryA
meas Flu
:<;
=
cal Fern
with:
meas Flu
cal Fern
measured rupture force of the tendon
with tendon-anchorage assemblies
n Ac fctm ... theoretical mean rupture force of all
FRP elements comprising the tendon,
without TA
As we expect from Fig. 9, the mechanical efficiency factor will
be less than 1. According to (1) it should exceed 0.95, expecia11 y because of economic reasons and in order to en sure adequate ductility of the concrete member. In Fig. 10 the results
of rupture tests on of FRP tendons with TA on both ends are
shown; they are plotted on the force-strain line of the FRP material (black circles). With the mean strength of FRP bars of
the used production 1ot the theoret i ca 1 mean rupture force of
the FRP tendon without TA can be derived (white circles). The
c1oser the density curves are, the higher the mechan i ca 1 efficiency will be.
Fig. 9 also shows that there exists an outer resin skin without
fibers around the inner fiber-reinforced core of the bar.
Hence, the transfer of force requires not only a sufficient intralaminar shear strength within these regions but also a sufficient interlaminar shear strength between them.
FRP Reinforcement
355
Anchorage development
The preceding remarks have shown that FRP call for new ways of
force transfer to the TA. In (11) to (13) the methods known so
far are dealt with. Here only an overview must suffice, with
emphasis rather put on principles than on detail.
The experimental data and the experience on the mechanical behaviour of prestressing steel tendons and anchorages are abundant. When developing FRP tendons this experience cannot be
used. Hence, to minimize costs and time an interactive developmental approach is recommended. It comprises the mechanical modelling of the anchorage and verification testing. In ( 11) to
(12) this approach has be used.
Essentially three types of post-tensioning anchorages have
emerged which differ with respect to the force-transfer mechanism: the clamp anchorage, the wedge- bond anchorage, and the
bond anchorage. Several transfer mechanisms can be combined.
In the clamp anchorage of Fig. ll the FRP bar is clamped between grooved steel plates by prestressed bolts and springs.
The force transfer is of the shear-friction type. Hence intermediary layers become necessary, also for protection. To minimize the differential slip between the FRP bar and the steel
plate the 1 atter can be segmented (12). Such anchorage type
proved to be very efficient for the testing of FRP elements.
In the wedge-bond anchorage of Fig. 12 a certain part of the
force is transfered by resin bond to the steel sleeve. The rest
of force is then anchored by a combination of bond and transverse wedge pressure. This very efficient anchorage has also
been used for multi-bar tendons. Fig. l3 shows as an example
the anchorage for the seven-wire CFRP strands of Tokyo Rope
Mfg. Co. ( 4) .
The most widely used and tested bond anchorage is shown by Fig.
14. The FRP bar is embedded in a steel housing in a polymer resin mortar, which transfers the force by bond to the inner
threaded surface of the housing. Extensive short- term, longterm and dynamic tests have been performed on HLV anchorages (1
to 19 0 7, 5 mm GFRP Polystal bars) of STRABAG. A high efficiency of more than 95% was attained. Fig. 15 shows an HLV anchorage after being tested to failure. This type of anchorage
has been applied for several structures in Europe (14). With a
similar anchorage also rectangular AFRP ARAPREE strips were anchored (15).
A very interesting aspect of HLV tendons is the incoporation of
optical fiber sensors into the Polystal bar for the monitoring
of the prestress (18).
356
Rostasy
PRE-TENSIONED CONCRETE WITH FRP
With CFRP and AFRP being superior to prestressing wire in many
environments which are aggressive to steel, they are suitable
for slender pre-tensioned concrete elements with a low concrete
cover. Pilot applications have proved this statement (e.g. (2),
(16)).
The mechanism of bond to concrete, necessary to introduce the
prestress, is not a composite property of the FRP element. It
depends strongly on the tensile and compressive strength of the
concrete and on the surface texture of the FRP element. Hence,
the bond 1ength can be manipulated vi a the surface texture as
done for prestressing wire with the ribs or indentations.
It was shown in (17) that the transfer bond length is best determined on pretensioned beams. A too short bond length in conjunction with a low cover value may lead to splitting cracks
along the FRP element (6). The differential thermal deformation
between any FRP (the matrix resin expands much more than the
concrete!) and concrete may a1so cause cracking. Further research is needed to clarify the complex interaction.
ON POTENTIALS AND LIMITS
The author refrains from jubilant descriptions and pictures of
proud applications. They will be shown in abundance during this
symposium. More important are the potentials and limits of FRP
for structural concrete.
To begin with the potentials. Are FRP the whiz kids of structural materials? In some ways yes, in others no. They have astonishing assets. The most important one, besides high-strength,
is their high corrosion resistance, though even there, a FRP
specific differentiation is necessary. FRP can be a successful
alternative to prestressing and reinforcing steel in all such
cases where the steel cannot be reliably protected from an aggressive environment or in which the corrosion protection of
steel is very costly. There a many promising fields of application.
This leads straight to the limits. In all such cases where
steel corrosion and/or protection against corrosion cause no
prob 1em or excessive costs, FRP cannot compete with stee 1 . The
initial costs of FRP - depending on the type - are high, though
on a downward trend. Besides that off-set to steel, there are
other drawbacks not to be suppressed. These draw-backs originate from the orthotropy of FRP:
FRP Reinforcement
• high sensitivity against lateral
jury, local bending and abrasion,
pressure,
superficial
• great differences in thermal dilatancy to concrete,
and radially.
357
in-
axially
These draw-backs can be overcome by research and by judiceous
use. With sobriety and tenacity a FRP specific culture of testing, quality control, design, detailing, and execution must mature in the years to come.
liTERATURE
(1)
FIP Commission on Prestressing Materials and Systems:
High-strength Fibers Composite Tensile Elements for
Structural Concrete. State-of-Art Report, July 1992, unpublished.
(2)
Gerritse, A.; Maatjes, E.; Schurhoff, H.J.: Prestressed
Concrete Structures with High-Strength Fibres. IABSE Report Vol. 55, Zurich, 1987, pp. 425/432.
( 3)
Waaser, E. ; Wolff, R. : Ein neuer Werkstoff fur Spann beton. HLV-Hochleistungs-Verbundstab aus Glasfasern. beton
36, 1986, H.7, pp. 245/250.
(4)
Zoch, P.; Kimura, H.; Iwasaki, T.; Heym, M.: Zugelemente
aus
kohlenstofffaserarmiertem
Kunststoff,
eine
neue
Klasse von Vorspannmaterialien. Forschungskolloquium am
I nst i tut fur Werkstoff i m Bauwesen, Un i vers it at Stuttgart, Nov. 1991.
(5)
den Uijl, J.A.: Mechanical Properties of Arapree. Part 4:
Creep and Stress-rupture. Report 25-87-31, TU Delft,
1991.
(6)
de Sitter, W.R.; Vonk, R.A.: Splitting Forces in FRP Pretensioned Concrete. Intl. Symposium on FRP Reinforcement
for Concrete Structures. Vancouver, March 1993.
(7)
FIP Commission on Prestressing Materials and Systems: Recommendations for acceptance of post-tensioning systems.
June 1992, unpublished.
(8)
Hankers, Ch.; Rostasy, F.S.: FRP-tendons for post-tensioned concrete structures. ACMBS-I, October 7-9, 1992,
Sherbrooke, Quebec, Canada.
(9)
Rostasy, F .S.; Budelmann, H.: Principles of Design of FRP
Tendons and Anchorages for Post-tensioned Concrete. Intl.
358
Rostasy
Sympos i urn on FRP Reinforcement for Concrete Structures.
Vancouver, March 1993.
(10)
Hoffman, 0.: The brittle strength of orthotropic materials. J. o. Composite Materials, Vol. 1, 1967, pp.
200/206.
(11)
Kepp, B.: Zum Tragverhalten von Verankerungen fUr hochfeste Stabe aus Gl asfaserverbundwerkstoff al s Bewehrung im
Spannbetonbau. Diss. TU Braunschweig, 1984.
(12)
Faoro, M.: Zum Tragverhalten kunstharzgebundener Glasfaserstabe i m Bereich von Endverankerungen und Ri ssen i m
Beton. Diss. Universitat Stuttgart, 1988.
(13)
Dolan, Ch.W.: Developments in Non-metallic PrestressingTendons. PCI Journal, Sept./Oct. 1990, pp. 80/88.
(14)
Konig, G.; Wolff, R.: Heavy duty composite material for
prestressing of concrete structures. IABSE Symposium, Paris-Versaille, 1987, pp. 419/424.
(15)
Muruyama, Y.; Amana, R.; Okumura, K.: Development of AFRP
Tendon System and its Application to a Stress-Ribbon
Bridge. Ann. Report Vol. 39/1991-10, Kajima Corp., pp.
65/72.
(16)
Yamashita, T.: Construction of Shinmiya Bridge with Carbon Fiber Reinforced Plastic for Prestressed Concrete
Tendon. Journa 1 of Prestressed Concrete (Japan), Vo 1 . 31,
No. 2, March 1989.
(17)
Nanni, A.; Tanigaki, M.: Pretensioned Prestressed Concrete Members with Bonded Fiber Reinforced Plastic Tendons: Development and Flexural Bond Length (Static). ACI
Structural Journal, July-August 1992, pp. 433/441.
(18)
Miesseler, H.-J.; Wolff, R.: Experience with the monitoring of structures using optical fibre sensors. XI. FIPCongress, Hamburg, June 1990, pp. Q 12/17.
(19)
Budelmann, H.; Rostasy, F.S.: Stress Rupture Behaviour of
FRP Elements-Phenomenon, Results and Forecast Models.
Intl. Symposium on FRP Reinforcement for Concrete Structures. Vancouver, March 1993.
FRP Reinforcement
-
CFK
CFK!GFK
GFK
=Metal!
=
Fig. 1-FRP in an airplane component
Fig. 2-Sportswoman and her FRP racket
359
360
Rostasy
high and adjustable tensile strength
high and adjustable Young's modulus
high dynamic strength
good to excellent corrosion resistance
low specific weight
magnetic and electric neutrality
Fig. 3-Main assets of FRP
brand
type
fiber/
matrb
axial I)
tensile
strength
axial I)
Young's
modulus
Vol.-~
GPil
GPa
75
53
3,3
fiber
volume
Vf
axial
ultim.
strain
~
Polystal
G/UP
68
2,65
1,80
Polygon
G/EP
60
2,99
I ,79
93
56
3, I
Arapree
A/EP
45
3,00
I ,35
123
55
2,3
Fibra
A/EP
59
2,35
1,38
110
65
2,0
Leadl ine
C/EP
65
2, 79
1,82
229
149
I ,3
CFCC
(strand)
C/EP
64
3,29
2,12
213
'137
1,6
I) top nulllber related to the fiber cross-section; bottom number to
the composite section
Fig. 4-Main mechanical properties of fibers and FRP elements
FRP Reinforcement
Fig. 5-CFRP bar and strands (Tokio Rope)
2,5 r - - - - - - - - . - - - - . - - - - - - - , - - - - - - - ,
CFRP
1
165mm, Ac= 19,6 mm 2
v1 =65%
4
u:
1,5 1----~"""""'---+------,/r--------i
II
d'
VI
~
AFRP ARAPREE
"til
2,6/20mm,Ac= 52 mm2
I v1 =44"/o
~
"j
_
GFRP Polystal
r,6 7,5mm,
Vt
0
=
r%
2
+-------...,~---1
Ac= 44,2 mm2
I
3
tensile strain in 'ro
Fig. 6-Stress-strain lines of various FRP
4
361
362
Rostasy
1,0r-------------~
ARAPREE f 100K,1.5x20mm
Ac =30 mm 2 , v1=37%
0,8
' ...
'
alkaline solution, 20"C,pH 13
=30 kN , Fern = 36 kN
Fek
'
P,=95%
E
LLu0,7
"'0
~
• tailed
o unfailed
' ' '•
0,6
-
theor. line
1
0,5
0,1
Fed
(
tu )-;;
F= B -t
0
em
10
,t0 =1h
10 2
10 3
10 4
10 5
106
tu; t I In) in [ h]
Fig. 7-Creep rupture of AFRP strips in alkaline
solution-experiments and forecast
short term axial
tensile strength
• creep rupture at tu
ts service life
1=
Fedk
at t.
Pm.s
t hllntul
.
LLU
luk
tu • ts
I In) in [ h]
Fig. 8-Determination of permissible prestressing of FRP
FRP Reinforcement
fiber~
direction~
.-+Ocn
~
fcnt
external
resin skin
inner core
Fig. 9-Failure criterion of ud-FRP under planar stresses and
strength reduction by the anchorage (schematically)
o test result fcu
• test result Ffu
fck and Ffk 5 'fo-fractile
strain on free length of tendon
Ec
Fig. 10-Force-strain line of FRP elements and of
the TA, and test results (schematically)
363
364
Rostasy
clamping forces
clamp
intermediary layer
FRP bar
spring
~~·'
On
?
Fig. 11-Clamp anchorage and stresses (schematically)
Fe~
W
steel cylinder
with inner thread
resin mortar
FRP bar
Fig. 12-Wedge-bond anchorage and stresses (schematically)
FRP Reinforcement
Fig. 13-Wedge-bond anchorage for CFRP strands
Fig. 14-Bond anchorage and stresses (schematically)
365
366
Rostasy
Fig. 15-HLV-anchorage after testing
SP 138-23
Should FRP be Bonded
to Concrete?
by C.J. Burgoyne
Synopsis: The question of whether it is right to bond tendons made of glass,
ararnid or carbon fibres to concrete has not yet been directly addressed. This paper
discusses the various issues involved, and concludes that in many cases, these
tendons should remain unbonded.
All the new materials which have a high enough stiffness and a low enough creep
show a linear elastic response right up to failure, with little or no ductility. This
contrasts with steel, even very high tensile steel, which shows a considerable
reduction in stiffness at high loads. In a bonded beam, when cracks form on the
tension face of the concrete, very high strains are generated across the crack. With
a steel tendon, local yield must occur, with a consequent reduction in cross-section
area, which leads to debonding of the bar on either side of the crack. This allows
the strain at the crack to reduce below its theoretical maximum value. In
calculation, average steel strains are used, which ignore any local increase at the
crack positions, but there are some controversial code rules which limit the
(average) steel strain to less than the material can actually sustain.
When new materials are used, the local yielding mechanism is no longer available,
and the concept of using average strains is no longer justified. In concrete
reinforced with FRP, the whole strain capacity of the fibres is available, and it is
unlikely that fibre failure will occur before the concrete strains become
unacceptable. But in prestressed concrete, much of the fibre strain capacity is
absorbed in the prestress, leaving a tendon very sensitive to high strains in the
vicinity of cracks.
There is a move to increa':e the ductility of beams reinforced or prestressed with
FRP, by the use of (FRP) cages in the compression zone. This will increase the
chances of a bonded tendon snapping before crushing of the concrete occurs.
These mechanisms are not present in unbonded tendons, where high local strains
do not occur, and indeed the change in stress in the tendon is small. It has been
argued, for steel tendons, that this is an economic disadvantage; for FRP tendons,
however, it is shown here to be beneficial.
The intention of raising this matter at the conference is to engender debate on the
topic, before systems start to become widely used.
Keywords: Bonding; cracking (fracturing); ductility; fiber reinforced plastics;
prestressed concrete; prestressing steels; reinforced concrete
367
368
Burgoyne
Dr Chris Burgoyne is a lecturer at Cambridge University, and has been carrying
out research into the properties of aramid ropes for a number of years. His other
research interests include the application of expert systems to the design of
prestressed concrete, and the general philosophy underlying structural design. He
was formerly employed at Imperial College, London.
INTRODUCTION
Prestressing tendons and reinforcing bars, made from new materials such as glass,
aramid or carbon fibres, are now becoming widely available; the properties of the
parallel-lay ropes or FRP pultrusions made from them are now understood. They
are seen as potential replacements for steel in areas where corrosion, weight or the
magnetic properties of steel pose problems.
The temptation is to think of "replacing steel with new materials", without, in many
cases, going back to first principles. We have become so used to the way we make
reinforced and prestressed concrete that we take for granted the reasons why we do
things the way we do. When something changes (in this case the introduction of
new materials), we fail to consider whether our original assumptions remain valid.
This route could lead us to some potentially dangerous or costly mistakes, which
could set back the use of new materials for a very long time. It is in anticipation of
this problem that the paper has been written.
Before we consider what our new structures should look like, it is important to
consider why we build steel reinforced structures the way we do. That requires
some questioning of very basic assumptions.
DESIGN WITH BONDED STEEL
With very few exceptions, reinforced and prestressed concrete structures bond the
steel to the concrete. Why? There are two main reasons; one associated with
corrosion, and the other with the strain capacity of the steel.
Passivation
Steel rusts when exposed to both oxygen and water. This corrosion can be
prevented if the steel is held at a high pH, typically in the range 11-13, when a
passivation reaction takes place at the surface of the steel (Figure 1)(1). Corrosion
is not prevented completely, but a fine surface layer is formed which prevents
further corrosion taking place. Concrete, fortuitously, provides exactly the correct
environment, where the natural alkalinity of the concrete passivates the steel. It is
always assumed that the concrete must be in direct contact with the steel, although
this has not been established with any certainty. A small void next to the steel,
FRP Reinforcement
369
with no access to the outside world, will either be dry or contain a small amount of
water, which will have the same pH as the adjacent concrete. In neither case will
corrosion take place. Grout used in prestressing ducts has far more water in it than
is needed to make the cement hydrate; most is used to make the mix liquid. This
water will eventually evaporate, leaving small voids in the grout. To the best of the
author's knowledge, there is no evidence of corrosion taking place in such small
voids.
Rate of
corrosion
..,._ General wastage
Alkalinity to be
maintained in
:, boiler water
~~--~--~--~--~--~~~-1---L~
2
3
Figure 1.
4
6
7
pH
8
9
10
II
12
Rate of corrosion, as a function of pH.
(taken from reference 1)
If the void is larger, and there is free passage of air and water vapour between the
void and outside, then the passivation is lost and corrosion can occur. The same
depassivation will occur when carbonation, caused by the diffusion of atmospheric
C02, reduces the alkalinity of the concrete. But carbonation, in properly
compacted concrete, will take decades, so this is assumed not to be a problem if
concrete is properly detailed.
The recent decision by the Department of Transport in the UK to ban the use of
grouted post-tensioned concrete bridges is a direct result of concerns about
corrosion taking place in voids within the ducts (2).
Strain Capacity
The other reason why we bond steel to concrete has to do with their strain capacity.
Most concrete, as used in reinforced construction, has an approximately linear
response up to a compressive strain of about 0.0012. This is remarkably similar to
the strain capacity of reinforcing steels. Thus, we can design reinforced concrete
(RC) beams, with the steel taking the tensile strains, and concrete the compressive
strains; the neutral axis of the beam is close to the mid-depth of the beam, making
full use of both materials (Figure 2a). This is a remarkably good arrangement for
two materials which just happened to be both available and relatively cheap at the
same time. Both materials share in carrying the load, and the curvatures that are
induced are minimised by keeping the neutral axis close to the mid-plane.
370
Burgoyne
Prestressed concrete (PSC) works in much the same way. The concrete is usually
of better quality, with a linear strain response up to about 0.0018, and the steel has
a much higher linear strain capacity (about 0.006). If we built reinforced concrete
with prestressing tendons, we would have to under-reinforce the beam to a great
extent to make any significant use of the steel's capacity; very large curvatures
would occur if both materials were to reach their strain limits at the same time
(Figure 2b). We overcome the problem by pre-straining the steel, usually to a
strain of about 0.004. This leads to a capacity to resist additional strain of about
0.002, which is very close to that of the concrete. Once again, we find the
materials in balance, with a neutral axis close to the mid-plane (Figure 2c). There
is the added benefit that the additional strain in the steel increases the stress, which
in tum increases the moment capacity. As we shall see later, this mechanism is not
so effective when the steel is unbonded.
0.0012
0.0012
(b)
(a)
0.0012
0.006
0.0012
(c)
prestrain
0.002
Figure 2.
0.004
Strain distributions in:(a) typical reinforced concrete beam
(b) beam reinforced with untensioned prestressing steel
(c) typical prestressed beam.
FRP Reinforcement
371
So we have two very good reasons for bonding steel to concrete in both reinforced
and prestressed concrete. What happens when the concrete cracks, in normal use
for RC, and when overloaded for PSC?
Cracked Concrete with Bonded Steel
The first thing to note is that the strain in the concrete, at the crack, is technically
unbounded. Consider two points in the concrete, one on each side of the crack,
that were in contact before the crack opened. They are now separated by the width
of the crack, so the strain is infinite. If the steel were perfectly bonded to the
concrete on both sides of the crack, then the steel would also have to have infinite
strain, leading to failure by snapping of even the most ductile steel.
Longitudinal sect1on of axially loaded spec1men
Concrete
l'"'" " "~' '"'
Primary crack
Figure 3.
00
'="'"
Force components on bar
Secondary cracking in vicinity of primary crack.
(taken from reference 3.)
This clearly does not happen, because as soon as the steel starts to yield it gets
thinner since the volume of the steel is sensibly constant. We thus get a small
amount of debonding at the surface of the bar, until the average strain along the
de bonded length of the bar is less than the yield strain. If the bar is deformed, with
raised ribs, there will be secondary cracking in the concrete (3), which also serves
to extend the debonded length and reduce the effective strain (Figure 3).
We take this process for granted. We never try to calculate the actual strain in the
steel, being quite happy to assume a linear variation in strain through the depth of
the beam (as in Figure 4), even though the compression concrete acts
homogeneously, while the concrete in tension is cracked and debonded from the
steel. We assume that the steel strain can be taken as the same as that in the
adjacent concrete, although we know that debonding takes place and that the
concrete strain is wrong. Strain gauges attached to reinforcing bars invariably give
the "wrong" strain. Those placed between cracks give a lower than expected
strain, since the steel and concrete are acting compositely, whereas those that
372
Burgoyne
• •
Figure 4.
Linear strain distribution assumed through beam.
happen to have been placed at crack positions show much higher strains than
expected. The actual strain distribution along the beam in the vicinity of the crack
would be as shown in Figure 5.
The controversial rule in the CEB/FIP model code (4) that limited the additional
(average) strain in the steel to 1%, even though prestressing steels can normally
sustain strains of about 3%, was probably based on the idea that the actual peak
strain would be considerably higher than 1%.
We accept this because it works. There is ample evidence that structures designed
in this way behave as expected. The local yielding and debonding keeps the
situation under control. The strain in both concrete and steel when averaged over
several cracks, does follow the assumptions above, and when the structure is taken
to failure, yielding of the steel becomes general anyway, since we always underreinforce our beams, so our prophecies become self-fulfilling.
Concrete
--------~1~---t_e_n_d_on----~~crack
peak steel
strain
Figure 5.
average
steel strain
.L. __ _
Actual strain distribution along tendon in a cracked beam.
FRP Reinforcement
373
NEW MATERIALS
What will we expect to happen to beams with bonded reinforcement made from
glass, aramid or carbon? The first thing to note is the stress-strain curve. Figure 6
shows typical responses for these materials (5), in the form of fibre; they are all
effectively linear up to a sudden, brittle failure. Bonding the fibres together with
resin will not alter these curves much; the contribution of the resin to the axial
strength is negligible, and when the fibres fail, the bar fails.
So we have lost the essential property of the bar that we rely on to shed load from
our tendon when the concrete cracks. The strain across the crack remains infinite,
at least in theory, and the bar cannot yield to even out the strain in the bar. So what
is the strain in the bar?
Sglass
4000
3000
Prestressing steel
Strain:%
Figure 6. Stress strain curves for glass, aramid and carbon fibres
proposed for use as reinforcement or prestressing.
(taken from reference 5.)
374
Burgoyne
The answer is that it is almost impossible to know. The strain in the concrete on
the face of the crack must be zero, while the strain in the FRP at the crack ought to
be infinite. A short distance away from the crack, the strains in the two materials
must be the same. There is thus a discontinuity (at point A in Figure 7), which can
only be resolved if there is a bond failure crack propagating along the interface.
But how far does it go? In steel, the requirement that the steel stress remains
below yield allows us to have some idea of what is going on, but no such
mechanism exists for FRP.
The controlling factors will thus be surface treatment of the bar, the quality of the
concrete, and the degree of compaction and consolidation in the concrete around
the bar. These are notoriously difficult matters to control, and almost impossible to
test in real structures. Furthermore, the condition of the surface of the bar is
almost entirely a function of the resin, and this will be very sensitive to creep. We
would expect the bar strains to differ between a beam subjected to transient loads,
and one subjected to permanent loads.
)
I
Bond failure
cracks~
l
~-----------------
Figure 7.
VE:f=oo
~fc=O
Strains in the vicinity of a crack.
And yet, we shall have to estimate the strain, and hence the stress, in these bars,
since we shall have to predict the load capacity that the composite (FRP/concrete)
beam will have. Equilibrium considerations will help to a certain extent, but only
in so far as we can estimate the lever arm, and we do that at the moment by
assuming that the strain distribution is linear.
If the strain distribution is non-linear, it is quite possible for the centre of
compression to lie in a different place from that assumed. If it is closer to the
reinforcement than originally assumed, then the lever arm would be lower, and the
reinforcement stress higher.
There is a further problem with variability of the bars, in instances where t?ere are
several reinforcing bars. This is analogous to the bundle theory problems m ropes
(6). The various reinforcing bars will have (slightly) different strengths. ~]early
the weakest will fail first; if it is steel, it can yield - straining at approximately
FRP Reinforcement
375
constant stress- without shedding load. But if it is FRP, it will snap, so that the
remaining bars will have to carry its share of the load. This phenomenon usually
means that the beam will fail when the weakest bar fails.
Under-reinforcement versus Over-reinforcement
We take for granted the fact that we under-reinforce beams with steel
reinforcement, to ensure that the steel yields before the concrete crushes, thus
giving us a more "ductile" response. But snapping of FRP bars will be rather
final; we shall need to ensure, either that the concrete fails first (i.e. overreinforce), or overdesign so that the bars stay well below their strain capacity.
This then leads to consideration of the strain capacity problem. Most new materials
have strain capacities that are up to an order of magnitude higher than those of
reinforcing steels (typically 0.015 or above). They are also expensive, so we want
to make good use of them. This means taking them close to their strain capacities.
But if we do, the strain diagram will be as shown in Figure 8, once more returning
to very high curvatures, with a high neutral axis, and a very under-reinforced
beam.
At this point in the argument, it becomes sensible to think in terms of increasing the
strain capacity of the concrete in compression. It has long been known that
providing a lot of confining reinforcement increases both the strength, and more
particularly the strain capacity, of concrete; up to 0.015 or 0.02 can be achieved
(7). There has never been much economic sense in making use of this with steel
reinforcement, but with high straining FRP, it becomes a more sensible
proposition.
0.0012
0.015
Figure 8.
Strain distribution for FRP reinforced beam.
376
Burgoyne
Research Needed
There is very little data available on the behaviour of beams reinforced or
prestressed with FRP, and what there is concentrates simply on the load capacity.
What is important for future writers of codes of practice, and users of such
materials, is information about the mode of failure, and the behaviour of the tendon
in cracked concrete. This will have to be obtained by careful measurements of
concrete strains, studies of crack patterns, and back-analysis of beams to determine
reinforcement stresses.
BEAMS WITH UNBONDED REINFORCEMENT
Another possibility is to build the beam with unbonded reinforcement. What
happens in this case? For simplicity, it will be assumed that the reinforcement is
attached to the concrete at the ends of the beam, either by prestressing anchorages
or by bond in reinforcement. The strain in the reinforcement is not now affected
by local strains in the concrete, but must respond to changes in length of the whole
beam. The reinforcement strain must be the strain of the concrete adjacent to the
reinforcement, averaged over the whole length of the beam.
If the beam is subjected to a normal working load, there may be cracking in some
regions, but not to any great extent. The strain distribution in the concrete will be
similar to that shown in Figure 9, and the average value (as in the reinforcement),
will be of the order of 50% to 100% of the maximum value, depending on the
shape of the moment diagram.
r==;::
E
Uncracked
;?1
Strain in concrete
adjacent to tendon
,---,---
Strain in
unbonded tendon
Figure 9.
f3.Eu
Strains in an uncracked beam with an unbonded tendon.
FRP Reinforcement
377
When the beam is subjected to its ultimate load, one (or more) hinges will form,
leading to very high concrete strains in that region. But the reinforcement will not
see that peak strain, since it will continue to be subjected to the average value. This
will probably be between 10% and 25% of the maximum value (8), once again
depending on the details of the loading arrangement (Figure 10).
These values will be affected by contact between the reinforcement and the
concrete, such as at the deflectors in prestressed construction, where some sizeable
friction effects may arise. Nevertheless, the reinforcement strain will remain much
lower than the concrete strain.
If the beam is reinforced, the strains mentioned above will be the actual strains in
the reinforcement, but if the beam is prestressed, they will be the additional strain,
over and above that induced by the prestress.
Nothing in the above argument relies on the fact that the reinforcement yields, so
the same logic applies if we are talking about steel or FRP. The argument against
unbonded construction has always been the loss of corrosion protection, and the
loss of the additional stress in the steel. Neither of these will be problems with
FRP.
r==-;;
Cracked
I I ) \
-----r,;Strain in steel
Figure 10.
f3. Eu
Strains in a cracked beam with an unbonded tendon.
PREDICTIONS
The following predictions are made on the basis of the above arguments. It is
assumed that tests and prototype structures will be made in many different
configurations. Most will work, and a few will fail, but eventually those that make
effective use of the materials will become adopted by the industry.
378
Burgoyne
Beams Reinforced with FRP
Beams reinforced with FRP are unlikely to be adopted widely on strength grounds
alone. It will be difficult to get the reinforcement strain sufficiently high without
causing unacceptable curvature in the beam, and very high strains in the concrete.
FRP reinforced beams will find application where their resistance to corrosion is of
prime importance. In that case, it is unlikely that making maximum use of the
material's strength will be important, and stiffness is likely to govern, so the
danger of over-stressing the FRP, and hence snapping the bars, will not arise.
Beams Prestressed with FRP
When we consider prestressed construction, the situation is different. We can take
up as much of the strain capacity in the tendon as we want by pre-straining it.
Economics will dictate that we tension it as highly as possible. The stiffness of the
beam at the working load will not be a problem; that comes almost entirely from the
uncracked concrete. The use of confining links in the compression zone of the
concrete will be economically attractive, so we will be left with the idea that, at
failure, we will want to induce high strains in the concrete next to the tendon.
This leads inevitably to the concept that we will want our tendon to be unbonded.
This will avoid pushing the tendon strain up so high that we snap the tendon. It
will remove problems associated with the difficulty of providing controlled bond
between tendon and concrete, and will give great freedom to the design engineer,
both in form of the concrete cross-section, and in the layout of the prestress.
0.02
FRP strain
I
o.o2
'
25
o/c- 0.005
0
0.008
lprestrain)
Figure 11. Strain distributions in a beam prestressed with an unbonded FRP
tendon, with compression concrete confined to enhance strain capacity.
FRP Reinforcement
379
We can envisage a strain diagram, at failure of the beam, as shown in Figure 11.
The concrete in compression is confined by hoop reinforcement (probably made
from loops of FRP), that can sustain a strain of about 0.02. The tendon will have
been prestrained to about 0.008, and if the neutral axis is at mid-depth, would pick
up an additional strain of about 0.02, which would be very dangerous. But if
unbonded, the extra strain induced in the tendon is likely to be no more than 25%
of this, or about 0.005. This will give a total strain of 0.013, which is well within
the strain capacity of a material such as a high modulus aramid like Kevlar 49.
CONCLUSION
New materials like glass, aramid and carbon fibres offer a great potential to
structural engineers for the construction of non-corroding tendons and
reinforcement. But these materials should not be seen as direct replacements for
steel; they are materials in their own right, and we must consider how they should
best be used.
It has been concluded that beams reinforced with FRP will be unable to use the full
strain capacity of the bars, since the strains induced in the concrete will cause
unacceptable curvatures. It is thus unlikely that there will be problems caused by
bars snapping due to the lack of ductility.
For beams prestressed with FRP, the situation is different. Prestraining the
tendons will make them very sensitive to additional strains induced by beam
curvature. Making the tendon unbonded will offer real assurance that the tendon is
not pushed off the top of its stress-strain curve, causing premature failure.
We need research about the way FRP acts compositely with concrete. We can be
happy about the basic fibre properties themselves, but we must now pass on and
see how these materials can best be used in practice
References
1.
Tretheway, K.R. and Chamberlain, J. Corrosion for students of science
and engineering, Wiley, New York, 1988.
2.
Grouted duct tendon ban poses problems, New Civil Engineer, 8th October
1992.
3.
Beeby, A.W. Cracking and Corrosion, Concrete in the Oceans Technical
Report 1, CIRIA/UEG, 1979.
4.
CEB/FIP. Model code for concrete structures, 1978, Clause 10.4.1.1.
5.
Hollaway, L. and Burgoyne, C.J. Further applications of polymers and
polymer composites, in Polymers & Polymer Composites in Construction,
ed. Hollaway, Thomas Telford, London, 1990.
380
Burgoyne
6.
Burgoyne, C.J. and Flory, J.F. Length effects due to yearn variability in
parallel-lay ropes, Procs MTS 90, Washington DC, 1990.
7.
Harman, T.G. and Slattery, K.T. Advanced composite confinement of
concrete, Procs ACBMS-1, 299-306, Sherbrooke, Quebec, 1992.
8.
Nanni, A.E. and Alkhairi, F.M. Stress at ultimate in unbonded posttensioning tendons:- Part I - evaluation of the state of the art. ACI
Structural Journal, Vol 8, 641-651, 1992.
SP 138-24
Epoxy Socketed Anchors
for Non-Metallic
Prestressing Tendons
by L.E. Holte C.W. Dolan,
and R.J. Schmidt
Synopsis: Finite element analyses and experimental confirmation tests were
conducted to evaluate epoxy socketed anchors for fiber reinforced plastic
prestressing tendons. The studies found that a parabolically varying profile
provides superior performance compared to a conventional linear conic
anchor. It was also found that an anchor with a bond release agent on the
surface between the socket and the resin plug results in a lower peak shear
stress compared to a bonded anchor. The combination of a parabolic anchor
and a bond release agent permits use of a wider range of resins as socketing
agents and is less sensitive to construction tolerances. Additional research
is suggested to optimize material selection, anchor geometry and anchor
construction.
Keywords: Anchors (fasteners); epoxy resins; fiber reinforced plastics; finite
element method; prestressing steels; shear stress
381
382
Holte, Dolan, and Schmidt
Lars E. Holte is a Master of Science candidate in Civil Engineering at the
University of Wyoming. The topic of his master thesis is anchorage of
non-metallic prestressing tendons.
Dr. Charles W. Dolan ,FACI, is an Associate Professor of Civil Engineering
at the University of Wyoming, Laramie, WY. Dr. Dolan chairs joint
ASCE-ACI Committee 423-Prestressed concrete and has over 20 years
prestressed concrete design experience prior to joining the University of
Wyoming faculty.
Dr. Richard J. Schmidt is an Associate Professor of Civil Engineering at the
University of Wyoming, Laramie, WY. Dr. Schmidt specializes in structural
engineering, computational mechanics and nonlinear material modeling.
He is a co-author of the text: Advanced Mechanics ofMaterials, Fifth Edition,
Wiley.
INTRODUCTION
The use of fiber reinforced plastic (FRP) composites for civil
engineering structures is a growing field. The advantages of using synthetic
materials are that synthetic fibers are more resistant to corrosive
environments, they give the designer the ability to tailor the material for a
specific application, and they often have higher strength/weight ratios than
competing materials. One area of growth, in which the resistance to
corrosion is the main advantage, is the use of FRP rods instead of steel in
prestressed concrete structures. There are some unresolved technical
problems with the use ofFRP rods. These include the lack of ductility of the
rods, the relaxation characteristics of the rods, and the thermal sensitivity of
the rods. However, the major technical problem delaying widespread use
of the FRP rods is the lack of a reliable anchoring system for the FRP rods.
This paper presents the results of research to develop a reliable anchoring
system to develop the full tensile capacity of the rods.
To better understand the difficulties with the anchor design, the
possible failure modes should be determined. There are four major failure
modes that have been observed with FRP prestressing tendons. They are:
1. Rupture of the rod within its free length. This is the optimal mode
of failure in that it demonstrates the full tensile capacity of the rod
is developed and the anchor does not induce failure.
2. Bond failure between the epoxy and the rod in the anchoring zone.
This failure mode is a result of ineffective load transfer from the rod
FRP Reinforcement
383
to the epoxy resin plug in the anchor and results from use of an
epoxy with inadequate shear (adhesive) strength or improper
surface preparation of the rod.
3. Shear failure in anchorage zone. With certain anchor geometries,
sharp stress concentrations can be developed that effectively pinch
the rod and lead to shear failure in the fibers of the rod.
4. Long-term creep and excessive deflection. Use of a low-modulus
epoxy, sensitivity to high temperatures, and an anchor with a large
diameter orifice all contribute to creep in the anchorage zone and
the subsequent loss of prestress force.
The objectives of this study were:
•
To study the mechanics of a resin socketed anchor.
•
To evaluate the effects of various geometric and material
parameters on anchor performance.
To develop an improved anchor system by optimal selection of the
various geometric and material parameters.
The research approach involved finite element analysis of the anchor
system to perform the parametric studies and to evaluate design alternatives.
The analyses results were then confirmed through experimental tests. These
tests and analysis examined the short term load characteristics of the anchor
and the rods. Long term behavior and sustained load capacity are not
included.
ANCHOR CONFIGURATIONS
For high strength prestressing steel, a variety of tendon anchorage
systems has been developed. Anchorage of the steel strands is commonly
performed by serrated wedges that grip the tendon or button-heads on the
end of the steel tendon. As a consequence of the ductility of prestressing
steels, the anchorage of prestressing force necessitates only a short anchor
length. In spite of local stress concentrations, almost the full tensile strength
of the prestressing steel can be utilized. However, due to the low shear
strength of synthetic tendons, relative to their tensile strength, traditional
anchoring systems used for high strength prestressing steel are not effective.
Therefore, new ways of anchoring synthetic tendons must be found.
Throughout the world, several researchers are developing glass, aramid, and
graphite fiber systems for pre and post-tensioning of concrete.
384
Holte, Dolan, and Schmidt
Development of a reliable strand chuck is extremely important if
pretensioning and post-tensioning systems are to be developed using bulk
strand tendons, as opposed to custom length. There are five generic
solutions for anchoring synthetic tendons. As illustrated in Figure 1, they
include:
1. Split wedge (Figure 1a). The split wedge system uses metal wedges
in a conic holder to surround the tendon. The wedges compress the
perimeter of the tendon, and the wedges' teeth grip the tendon. A
common failure mode for this system is progressive fracture of the
tendon due to the biting action of the teeth. Japanese (1) and
American (2) researchers claim to have developed split wedge
strand chucks that work on glass and aramid tendons; however, no
details of this anchor are described. Enka (3) uses a plastic wedge
system; however, this anchor is confined to pretensioning
applications.
2. Plug and cone (Figure lb). In the plug and cone system, a bundle
of tendon rods is inserted into a conical anchor housing. Then a
fluted, solid cone is driven into the center of the bundle to splay out
the rods and hold them individually. This system is based upon the
Freyssinet system for steel tendons and has performed well under
both static and fatigue loading of the Parafil aramid tendons (4).
3. Resin sleeve (Figure 1c). Resin sleeve anchors use an epoxy resin
to bond the tendon within a cylindrical, metal sleeve. The inside
and outside surfaces of the sleeve may be threaded to improve bond
with the epoxy resin and to facilitate jacking. The Polystal anchor
(5) uses a threaded cylindrical housing with a sand epoxy matrix.
Tokyo Rope (1) uses a neat resin in a threaded cylindrical housing
to hold their CFCC tendons.
4. Resin potted (Figure ld). The resin potted anchor (2,6,7) is a
variation of both the resin sleeve and split wedge anchor that uses
epoxy resin to bond the tendon inside a conical anchor. The
compressive action of the split wedge anchor can be developed
while the continuous bond of the epoxy resin eliminates the biting
action of the wedge teeth. These anchors have been unreliable in
both long- and short-term testing (6,7). Environmental and
thermal deterioration of these anchors raise serious concerns about
long- term durability.
5. Soft metal overlay (Figure le). Tokyo Rope uses a metal overlay
on carbon cable (1). The metal sleeve is permanently bonded to
the cable and conventional strand chucks are used to grip the
FRP Reinforcement
385
sleeve. The sleeve must be applied to the cable during fabrication
when the cable is prefabricated to length. Substantial flexibility for
field modification is lost with this system.
For this study, only the resin-potted system will be examined further.
The basic load transfer mechanism for this anchor system involves the
transfer of the prestressing force from the rod to the sleeve of the anchor via
normal and shear stresses on the surface of the rod, Figure 2. It is assumed
that no significant load transfer occurs by normal stresses at the end of the
rod. Thus, the entire prestress force must be carried through the tangential
reaction on the lateral surface of the rod over the length of the anchor (8).
The principal objective in anchoring the rod is to obtain uniform distribution
of the shear stress along the rod surface over the anchor length. The ideal
shear stress profile is free of major peaks, and large gradients, and must fall
to zero at the tail end of the anchor.
It has been observed that in reinforced plastic, shear strength increases
with increasing hydrostatic pressure; Mohr-Coulomb behavior (8).
Designing an anchor system that is under internal pressure is a possible way
to improve the load capacity of the anchor. One possible design is a
cone-shaped anchor in which high radial compression is generated as the rod
is loaded. This requires that the resin cone remain unbonded from the
anchor sleeve. Bond release can be achieved during fabrication by
lubricating the surface of the sleeve before the anchor is potted. The radial
stresses on the rod must be high enough to prevent pull-out, but they must
not exceed the level at which compressive damage occurs in either material.
By unbonding the resin cone from the sleeve, the hoop stresses in the resin
cone change from almost zero values to substantial compression, Figure 2.
The relative performance of bonded versus unbonded construction is one of
the parameters that is examined in the finite element and experimental
studies.
FINITE ELEMENT MODEL CHARACTERISTICS
The finite element method was used to predict the mechanical
behavior of the anchor and the different components that make up the
anchor. The finite element models were built using PATRAN (9) and analysis
was performed with FINITE (10). Postprocessing was performed with
PATRAN. All models were axisymmetric and used the standard eight-node,
quadratic isoparametric element, Figure 3.
The finite element model consists of a 0.125" (3.2 mm) diameter
Kevlar reinforced rod, an epoxy resin cone, and an aluminum sleeve. A
length of 4" (100 mm) was used for the anchor based on previous work (11).
The resin and the aluminum are assumed to be isotropic while the Kevlar
rod is assumed to be orthotropic using the properties given in Table 1.
386
Holte, Dolan, and Schmidt
Figure 3a shows a typical anchor model, boundary conditions, and loading.
Figure 3b shows a typical finite element mesh. An aspect ratio of eight or
less was maintained for all elements in the finite element models. The nodes
along the symmetry axis are restricted from moving in the radial direction,
while they can move freely in the x (axial) direction. A nominal load of 10
ksi was uniformly distributed over the end of the rod. The rubber plug at
the front of the anchor orifice used during specimen fabrication was modeled
by assigning the front-most elements of the resin plug a modulus of E = 20
ksi.
For the unbonded model it is assumed that the interface between the
resin cone and the aluminum sleeve is frictionless. This is achieved in the
finite element model by the use of relative constraint equations. By this
approach, nodes along the interface are allowed to displace only along the
tangent to the interface; displacements ofthese nodes normal to the interface
are prevented. While a true frictionless surface might not be obtained, the
effects of friction are expected to be small and any bond that does exist will
be broken under relatively low load levels. Therefore, the results from
analyses of the unbonded and the bonded cones should bound the actual
conditions.
KEY PARAMETERS
The first part of the finite element study focused on identifying and
understanding the critical parameters that influence the anchor behavior.
A series of parametric studies was undertaken. The key parameters studied
were the taper angle 8, the modulus of elasticity of the resin, the radius r; of
the orifice at the front end of the anchor, and bonded verses unbonded
resin-sleeve interface.
A parametric study of the taper angle of the anchor cone was
performed in combination with the study of the influence of the modulus of
elasticity of the resin. Seven different taper angles; 1.5 o, 3 o, 4.5 o, 6 o, 9o, 12 o,
15°, were combined with six different modulii of elasticity; 50, 100, 200, 400,
800, and 1600 ksi (344.7 MPa- 11.0 GPa), for the resin. Both the bonded
and the unbonded interface models were analyzed.
The last part of the parametric study was an examination of the
influence of front radius r;, in a combination with the modulus of elasticity
of the resin on system response. The front radius is the radius of the orifice
in the anchor sleeve. This consists of the radius of the rod plus the gap
between the rod and the anchor sleeve. Four different values of front radius;
3/32, 1/8, 3/16, 5/16 inches (2.38 mm- 7.94 mm), were examined in
FRP Reinforcement
387
combination with six values of the modulus of elasticity (200, 800, 2,000,
5,000, 10,000, 16,000 ksi; 1.38 GPa- 110.3 GPa) for the resin. The orifice
openings represent annular gaps of d/4 to 2d based on the Kevlar rod
diameter d.
INITIAL FINITE ELEMENT RESULTS
The finite element results showed a large reduction in the peak shear
stress in the unhanded interface model, relative to the bonded interface
model. From a deflection plot along the rod (Figure 4) it can be seen that
the bonded model uses only the front quarter of the anchor while the
unhanded model uses most of the anchor length. The figure also shows that
the chosen anchor length of 4" (100 mm) is adequate for all geometric
options. (Figure 4 also includes results of alternative anchor described later
in this paper.)
From Figure 5 it can be seen that for the unhanded model, the
deflection of the loaded end of the rod is dependent on a combination of the
taper angle and the modulus of elasticity of the resin. The figure also shows
that little stiffness is gained by using an angle above 7°. For comparison, a
typical spelter socket for steel cables uses a 15° angle. In the case of the
bonded model, taper angle had little influence on the deflection (these
results not illustrated). In the unhanded model, the peak shear stress was
dependent almost entirely on the modulus of elasticity of the cone, as can
be seen in Figure 6. Results for the bonded model show the same trend: the
peak shear stress increases with increasing modulus of elasticity, but with
higher overall she'ar stress values compared to the unhanded model.
VARIABLE TAPER ALTERNATIVE
The findings from the study of the taper angle and the knowledge
gained by studying the stress contour plots, led to a modified anchor design.
This new anchor has a segmental varying taper. The 4" (100 mm) long
anchor is divided into eight segments with a taper angle starting at 1 o
increasing in even increments of 1 o up to 8 o for the last segment at the tail
end of the anchor. The idea is to reduce the shear stress peak that naturally
exists at the front of the anchor to get a more even distribution of the radial
and shear stresses over the length of the anchor zone. The next refinement
was the use of a parabolic shape on the anchor sleeve, with a oo taper at the
front and an 8 o taper at the tail of the anchor. The results from the finite
element model showed a substantial change in the shape of the shear stress
plot obtained along the surface of the rod (Figure 7). Not only did the peak
shear stress decrease with the variable taper anchors, but the shear stress was
more evenly distributed over the length of the anchor and it decreased to
zero at the tail end of the anchor. However, a segmental or parabolic taper
388
Holte, Dolan, and Schmidt
will give some additional deflection compared to the linear-taper anchor
(Figure 4). Since the deflection is small and may be included in the stressing
operation, this additional deformation is not considered significant.
The parabolic-taper anchor showed results almost identical to the
segmental tapered anchor. The only difference was that the stresses varied
more smoothly through the anchor for the continuous nonlinear-shaped
sleeve. This confirms basic intuition that the segmental varying taper is a
good approximation of the parabolic shape. This fact will be useful in
manufacturing the test anchors. Some models with a cubic variation in taper
were also evaluated. With a cubic taper, the designer can control the taper
angle at a point between the front and the tail of the anchor. The finite
element models of these cubic shaped anchors show marginal improvement
in the distribution of the shear stress. At this time no parametric study has
been performed to optimize the shape of the parabolic- or cubic-tapered
anchors, since the gain seems to be small compared to the gain achieved by
going from a linear taper to a nonlinear taper.
The study of the orifice opening showed that the deflection of the
loaded end of the rod was dependent on both the orifice radius and the
modulus of elasticity. For modulus of elasticity of more than 10,000 ksi (68.9
GPa), the radius had no significant influence. The orifice radius had a strong
influence on the distribution of the shear stress through the anchor. When
the radius increased, the shear stress curve peaked up instead of going to zero
at the front end (Figure 8}. For a 3/32'' (2.38 mm) radius, most of the
deflection takes place in the front part of the anchor. This is because the
front part of the cone is too thin to develop' any shear deformation through
its thickness.
This study and previous studies (6) have shown that by using a softer
modulus of elasticity at the front of the resin cone, the peak shear stress at
the front of the anchor is significantly reduced. By using a soft material for
just the front elements, the peak shear stress was reduced by about 50% for
both bonded and unbonded models (Figure 7).
EXPERIMENTAL
VERIFICATION
The goal of the experimental work was to verify the findings from the
finite element modeling and previous work (11}. Based on the findings from
the finite element study, two anchor designs were selected for testing. As a
baseline, one anchor uses a 7o linear tapered sleeve, and is referred to as the
linear anchor in this report. The second anchor selected for testing was a
segmental varying-taper anchor and is referred to as the segmental anchor.
One set of anchors was made for each design. Each anchor was made from
aluminum alloy, was 4 11 (100 mm) long, and had an outside diameter of 2 11
(50 mm). The orifice at the front of both anchors had a radius of 3/32 11
FRP Reinforcement
389
(2.38 mm). The two sleeves were assembled from 1/2" (12.5 mm) thick rings
(Figure 9). The rings were used for ease of production and to achieve the
ability to readily adjust the internal geometry profile. Figure 10 shows the
resin cores of linear and segmentally tapered anchors. For production, the
sleeve would be produced as one piece. The anchor ends were potted one
at a time. First the inside of the sleeve was wiped, using a swab, with Dow
Corning High Vacuum Grease Lubricant. The tendons used for these test
were Kevlar 49 rods, 0.125" (3.2 mm) in diameter; see Table 1 for more data.
For the tensile test the typical specimen had an overall length of 30" (762
min). The surface of the rod was first roughened using 400 grit sand paper.
Then the surface was degreased with trichloroethane. To protect the rod
from contamination when it was placed into the sleeve, the rod was fitted
with a two-part plastic tube. One part was left in place to seal the gap
between the rod and the sleeve in the front of the anchor, and one part was
removed when the rod was in place. The part left in place also served as the
softening element at the tip of the cone. The anchor and rod assembly were
put in an upright position and the resin was prepared and pored into the
anchor from the tail opening. The resin used for these tests were Sikadur 35
LV, which is a high strength epoxy with low viscosity (Table 2). The epoxy
was cured at room temperature. After the epoxy in the anchor at one end
was cured, the procedure was repeated for the other end.
The anchors were mounted in an Instron 1332 universal testing
machine with a capacity of 55,000 pounds (245 kN). During testing the load
and the stroke data were recorded. In addition, data were taken from an
Instron extensometer with a 3.4" (86.4 mm) gauge length placed on the free
part of the rod to obtain strains in the rod (Figure 11 ). The test results show
that the segmental taper has a 100% success rate, resulting in failure of the
rod within its free length (Table 3). The test results also show that the
segmental taper can develop the full capacity of the rod. The rod samples
tested were over 5 years old and the strength achieved for one of the test
specimens was 119% of the ultimate tensile strength for a Kevlar 49 rod (12).
For each linear taper anchor, failure occurred within the anchor instead of
within the free length. The linear anchor had bond failure and/or shear
failure in the anchoring zone. These are the same problems experienced
in previous work (11). The tests also show that the rod does not have to be
lined up perfectly in the center of the segmental anchor for the anchor to
perform well. This will be advantageous in the case of field ·installation.
The test results correlate with the findings from the finite element
analysis. A segmental anchor gives a much lower value of peak shear stress
and a better distribution of the shear stresses over the length of the anchor.
The bond failure that has been seen in the linear anchors indicates that the
peak shear stress at the front of the linear anchors initiates a progressive
bond failure which propagates towards the tail of the anchor. This failure
is consistent with the capacities of the Sikadur 35 LV resin. The experimental
390
Holte, Dolan, and Schmidt
results for the segmental anchor shows that the deflection in the anchor is
less than predicted by the unbonded finite element model, but larger than
the deflection predicted for the bonded model. The reason for this
difference
could be either attributed to friction between the resin cone and the anchor
sleeve or to an incorrect representation of the resin properties in the finite
element model.
CONCLUSIONS
This study concentrated on the performance of a resin-socketed
anchor for synthetic prestressing tendons. The finite element analysis and
the experimental performance indicate that an anchor with a bond release
between the resin plug and the metal socket produces a substantial reduction
in the interlaminar shear stress at the tendon surface. Furthermore, an
anchor with a segmentally-varying or parabolically-varying taper is superior
to a linear taper. The parabolic taper reduces the peak shear stress at the
front of the anchor, eliminates the possibility of progressive delamination
along the rod interface, and allows the engineer to more closely define the
shear stress distribution within the anchor.
This research concentrated on the behavior of linear and parabolic
anchors using Sikadur 35 high-modulus resin. Analytical studies indicated
that the linear-taper anchor should fail in bond shear at the rod/resin
interface while the variable taper anchor should develop the full tensile
capacity of the rod. The experimental tests confirmed this predicted
behavior. All rods in the variable-taper anchor developed the full tensile
strength of the rod. The stress distribution was sufficient to allow all rods
to break at greater than the rated tensile strength given in the duPont Data
book. For all tests in which the linear-taper anchor was used, bond failure
occurred at the rod/resin interface.
The tests and analyses in this study were conducted on an 118 inch (3
mm) diameter rod. Obtaining a uniform shear stress field implies that the
anchor is suitable for larger diameter rods. The use of larger diameter rods
may be limited by shear transfer within the rod rather than the shear transfer
on the anchor interface.
The segmental-taper or parabolic-taper anchor is far less sensitive to
resin modulus and, theoretically, will work with a wide range of resin
modulii. This allows the designer substantial latitude to select resins suitable
for field performance or specific applications.
FRP Reinforcement
391
CONTINUING RESEARCH
Continuing research includes an examination of the following issues:
•
Evaluation of resin properties for shear strength and modulus for
a linear-taper and parabolic-taper anchors.
Behavior of resin-sleeve anchors and resin selection criteria.
Effect of cyclic loading on the variable- and linear-taper anchors.
•
Effects of elevated temperature on resin-socketed anchors.
•
Performance of anchors with multiple-rod tendons.
Performance of metal wedges in variable-taper sockets.
•
Development of all-composite anchorage systems.
ACKNOWLEDGEMENT
This report documents the findings from a research program is made
possible by a grant from the National Science Foundation, Project no. MSS
9114592, Dr. Ken Chong, Program Director. Sika Corporation provided
epoxy samples and E. I. DuPont de Nemours and Company provided the
Kevlar rods.
REFERENCES
1.
Tokyo Rope Mfg. Co., Ltd. (1990). Carbon Fiber Composite Cable
(Corporate Report).
2.
Iyer, S. L., Khubchandani, A & Feng, J. (1991). "Fiberglass and
Graphite Cables for Bridge Decks," In Srinivasa L. Jyer (Ed.) & Rajan
Sen (Co-ed.), Advanced Composites Materials In Civil Engineering
Structures (pp. 371-382). New York: American Society Civil Engineers.
3.
Enka bv. (1986). Arapree - the Twaron Resin Prestressing Element for
Concrete, Arnem, The Netherlands, (Corporate Report).
4.
Burgoyne, C. J. (Ed.). (1988). "On the Engineering Applications of
Parafil Ropes," Symposium on Engineering Applications of Parafil Ropes.
Department of Civil Engineering, Imperial College of Science and
Technology, London, England.
392
Holte, Dolan, and Schmidt
5.
Wolff, R. & Miesser, H. J. (1989). "New Materials for Prestressing and
Monitoring Heavy Structures." Concrete International. 11(9), 86-89.
6.
National Bureau of Standards. (NBS), (1976). Non-Metallic Antenna
Support Materials Pultruded Rods for Antenna Guys, Catenaries and
Communications Structures. Technical Report AFML-TR-76-42,
Washington, D.C.
7.
Dolan, C. W. (1991). "Kevlar Reinforced Prestressing for Bridge Decks,"
Transportation Research Record 1290.
8.
Patrick, K. & Meier, U. (1991). "CFRP Cables for Large Structures," In
Srinivasa L. Iyer (Ed.) & Rajan Sen (Co-ed.), Advanced Composites
Materials In Civil Engineering Structures (pp. 233-244). New York:
American Society Civil Engineers.
9.
PDA Engineering (1987). PATRAN Plus User Manual, Costa Mesa, CA.
92626.
10. Dodds, R. H. Jr. and Lopez, L. A (1980). '/\ Generalized Software
System for Nonlinear Analysis," International Journal for Advances in
Engineering Software, 2:161-168.
11. Dolan, C. W. (1989). Prestressed Concrete Using Kevlar Reinforced
Tendons. Ph.D. Dissertation, Department of Civil Engineering, Cornell
University, New York.
12. Materials Science Corporation. (1986). DATA MANUAL FOR KEVLAR
49 ARAMID, Spring House, PA.
13. Sika Corporation. (no date). Sikadur 35 HiMod LV, Lindhurst, NJ.
FRP Reinforcement
393
TABLE 1 -TYPICAL DATA FOR UNIDIRECTIONAL
KEVLAR 49 REINFORCED EPOXY
u.s.
Property
S.I.
Customary
11 x 106 psi
78.5 GPa
Transverse Elastic Modulus, E 2 0.8 x 106 psi
5.52 GPa
Shear Modulus, G12
2.07 GPa
Axial Elastic Modulus, E 1
Poisson ratio
0.3
X
106 psi
0.34
Tensile Strength, f1
0.34
200
X
103 psi
1379 MPa
Compressive Strength, f2
4.3
X
103 psi
29.6 MPa
Shear Strength,
6.3
X
103 psi
43.4 MPa
v12
Properties are based on information in the Data Manual (12) for an epoxy
resin rod with a Vf = 60%.
394
Holte, Dolan, and Schmidt
TABLE 2- MECHANICAL PROPERTIES OF SIKADURE 35,
HI-MOD LV RESIN
Property
u.s.
S.I.
Customary
Modulus of Elasticity, 28 day
3.5 x 105 psi
Tensile Properties (ASTM D-638} :
14 day Tensile Strength,
8.4 x 103 psi
4.2%
Elongation at Break
4.1 x 105 psi
Modulus of Elasticity
Flexural Properties (ASTM D-790):
Flexural Strength
14.0 x 103 psi
(Modulus of Rupture)
Tangent Modulus of
Elasticity in Bending
3.7 x 105 psi
Shear Strength (ASTM D-732} :
14 day Shear Strength
5.1 x 103 psi
Deflection Temperature (ASTM D-648):
14 day Deflection Temperature
127 oF
Properties from Sika Data Sheet (13}.
2.41 GPa
57.9 MPa
4.2%
2.83 GPa
96.5 MPa
2.55 GPa
35.2 MPa
52.8 oc
FRP Reinforcement
395
TABLE 3 - TEST RESULTS
Cone
Material
Rod
Material
Failure Failure
Rod
Diameter Load Mode
(lbs)
(in)
Linear
Sikadur 35 LV
Kevlar49
0.125
2782
Bond failure in anchor
226.7
Segmental
Sikadur 35 LV
Kevlar49
0.125
2447
Fracture in rod
199.4.
Segmental
Sikadur 35 LV
Kevlar 49
0.125
2936
Fracture in rod
239.2
Linear
Sikadur 35 LV
Kevlar 49
0.125
1487
Bond failure in anchor
121.2
Segmental
Sikadur 35 LV
Kevlar49
0.125
3355
Fracture in rod
273.4
Linear
Sikadur 35 LV
Kevlar49
0.125
1679
Bond failure in anchor
136.8
Segmental
Sikadur 35 LV
Kevlar 49
0.125
3076
Fracture in rod
250.7
Segmental
Sikadur 35 LV
Kevlar 49
0.125
2874
Fracture in rod
234.2
Test
Anchor
Numb. "JYpe
2
6
22
• Flaw in free length of rod.
Note I in - 25.4 mm, lib
= 4.45 N,
I ksi = 6.89 MPa
Tensile
Stress
(ksi)
396
Holte, Dolan, and Schmidt
a. Split Wedge
Cone
b. Plug and Cone
c. Resin Sleeve
Rod
\
d. Resin Socketed
/Sleeve
Rod
\
e. Soft Metal Overlay
Fig. 1-Anchor types
-Jor- -JDrt
Bonded Model
Unbonded Model
Fig. 2-Anchor configurations
FRP Reinforcement
Kevlar 49 Rod
a. Geometric Model w!Boundary Conditions
-liB
L.
b. Finite Element Mesh
Fig. 3-Geometric model and finite element model
·=
.,g
g
~
0.020 . - - - - - - - - - - - - - - - - - - - - - ,
Parabolic Thper, 8" at Thil-end, Unbonded
O.oi8
Segmental Thper (1"-8"), Unbonded
0.016
unear 7.5" Thper, Unbonded
unear 7.5" Thper. Bonded
0.014
Resin Modulus: 200 ksi
0.012
Applied Load: 10.0 ksi
0.010
0.008
0.006
0.004
0.002
0.000
~'-'-l<CI+J'-+J~'-'+1.+-&.............................................................."-'-........L.J
0.0
0.5
1.0
1.5 2.0
2.5 3.0
3.5 4.0
Distance Along Rod From Thil End, in
4.5
5.0
Fig. 4-Deflection along rod for bonded and unbonded models
397
398
Holte, Dolan, and Schmidt
12.0
•
I
I
I
I
.5 10.0
~
5
.
c
I
I
\
8.0
6.0
1
0
'fi
..,"
Resin Modulus:
•--- 50ksi
• - - 200ksi
•·····
800 ksi
'.
\
\
4.0
~',,,L__
2.0
.........
-..--.,_
2.0
4.0
"
0
0.0
~-----..
0.0
• - 1600ksi
Applied Load: 10.0 ksi
.. ~----------- --=--i-- -- :..:.;:;
:-------- . -
6.0
8.0
10.0
lllper Angle
12.0
14.0
16.0
Fig. 5-Linear taper angle versus deflection at
loaded end of unbonded anchor
4.0
-........ -- ..... ----·---
3.6
3.2
-·- ~~~~n- ~~'!~~~=!~- ~!i. ---
800 ksi
....... 2.8
~-------------------------------------
-
....----------------------------------------!:~:--~~~
::f 2.4
~
en 2.0
t;;
1.6
.c
en 1.2
. -·····--·-·-···---·-·--·=~~:~----~
-1~~-:--=
__ ...... --
"
0.8
0.4
----~~--~--~=--.:..---~-
:..--
0.0
0.0
2.0
4.0
6.0
8.0
10.0
lllper Angle
12.0
14.0
16.0
Fig. 6-Linear taper angle versus peak shear stress for unbonded anchor
I
I
o- - - Bonded 7.5 • Linear lllper
• - Unbonded 7.5" Linear lllper
• · - Unbonded Segmentallllper (1 "-8")
•····· Unbonded Segmentallllper (1"-8")
Resin Modulus: 200 ksi
Applied Load: 10.0 ksi
1.10
]
0.90
~
~
en
...!.
]
0.70
t;;
~ 0.50
0.30
0.10
-0.10
L...L.Ju..J...L...L...L..L.J....L.L.J...W....L...L.Ju..J...L...L...L..L.J....L.L.J...W....L...LJL...LJCL.I...L...L..L.J
0.0
0.5
1.0
1.5
2.0 2.5 3.0 3.5 4.0
Distance Along Rod From lltil End, in
4.5
5.0
Fig. 7-Sbear stress along rod for linear and segmental taper
FRP Reinforcement
1.5
·,;
-"
1.3
a-
1.1
a---·
g
"'
'~
inch radius
1/8 inch radius
·--- 3/32
3/16 inch radius
•----- 5/16 inch radius
Resin Modulus: 200 ksi
Applied Load: 10.0 ksi
::f 0.9
I!
I•
0.7
/!
~
""'
"'
399
.c 0.5
}!
·
;;.:::
0.3
0.1
-
..
-0.1
0.0
0.5
~ .. ~··...
1.0
1.5 2.0
2.5 3.0
3.5 4.0
Distance Along Rod From Thil End, in
4.5
5.0
Fig. 8-Shear stress along rod for different front radius, parabolic anchor
Test Cylinder
1----Rod
Fig. 9-Test anchor with test cylinder
400
Holte, Dolan, and Schmidt
Fig. 10-Linear and segmentally varying resin core
o.5or-------------------,
0.40
Total Deflection
- - - Rod Deflection
·--·· Anchor Deflection
.5
:5 0.30
'ti
~ 0.20
0.10
500
1000
1500
2000
Load,lbs
2500
3000
3500
Fig. 11-Deflection versus load for rod, anchor, and total, segmental anchor
SP 138-25
Behavior of Prestressed
Concrete Beams Using FRP
as External Cable
by H. Mutsuyoshi and A. Machida
Synopsis:
Fiber reinforced plastic (FRP) is a new structural material in the
field of civil engineering. This paper describes the mechanical properties of
prestressed concrete (PC) beams reinforced with FRP as external cables instead
of ordinary steel tendons. PC beams using FRP show almost the same mechanical behavior (such as the load-displacement characteristics and the condition
of the cables) as that of using ordinary steel tendons. An analysis procedure is
shown to provide fairly accurate strength prediction of such PC beams at the
ultimate state. The external cable system using FRP can be applied to actual
PC structures.
Keywords: Beams (supports); cable; fiber reinforced plastics; flexural
strength; prestressed concrete
401
402
Mutsuyoshi and Machida
ACI member Hiroshi Mutsuyoshi is an Associate Professor of Civil Engineering at Saitama University, Urawa, Saitama, Japan. He obtained his doctorate
degree from the University of Tokyo in 1984. His current research interests include application of FRP to concrete structures, external PC, and seismic design
of R/C structures. He is a member of ACI Committee 440.
ACI member Atsuhiko Machida is a Professor of Civil Engineering at Saitama
University, Urawa, Saitama, Japan. He obtained his doctorate degree from
the University of Tokyo in 1976. His current research interests include seismic
design of R/C structures, composite steel-concrete structures, and properties of
fresh concrete.
INTRODUCTION
FRP is made of high strength continuous carbon, aramid or glass fibers
impregnated with resin. FRP has some excellent properties such as 1) high
tensile strength, 2) non-corrosion, and 3) non-magnetization. Recently, the applicability of FRP to concrete structures in place of steel bars or steel tendons
as main reinforcements has been actively studied. Some actual concrete structures reinforced with FRP have been erected in the last few years. In order to
apply FRP effectively to concrete structures, a rational design method has to be
established because some characteristics of FRP are quite different from those
of ordinary steel, and the merits of the mechanical properties of FRP must be
used as effectively as possible for concrete structures.
A PC structure reinforced with FRP as main tendons is clearly one of
the most suitable structures from the points of its high tensile strength and
small relaxation characteristics. Most of the past studies on PC using FRP just
dealt with an inner cable system. On the other hand, PC structures with an
external steel cable have been recently developed and applied actively to actual
structures mainly in Europe. The advantages of PC structures using an external
cable system are as follows (1): 1) dead load of structures can be reduced
because the thickness of the web can be made smaller, 2) reduced complexity
in construction work, 3) old or damaged cables can be easily changed or represtressed, 4) loss of prestress force due to friction of tendons can be reduced.
In order to overcome the mechanical shortcomings of FRP such as lack
of plastic behavior and small elongation at rupture and to make use of FRP
effectively, PC structures reinforced with FRP as external cables can be considered to be a good method. The reason is that breaking of FRP due to stress
concentration can be prevented because the stress acting in FRP is almost uniformly distributed along the cable. Furthermore, since FRP is non-corrosive
material, it can be applied to retrofit of old bridges and water tanks as external
cables. In this study, PC structures reinforced with external FRP cables have
been newly developed, and the fundamental behavior of these PC beams is
investigated.
FRP Reinforcement
403
OUTLINE OF EXPERIMENTS
The test specimens are T-shaped beams as shown in Fig. 1. Deformed
bars (DlO, SD343) 1 were used as longitudinal reinforcement and stirrups. The
ratios of the tensile reinforcement and the shear reinforcement are 0.41% and
1.40%, respectively. Two cables were arranged on the both sides of the beam
web as external cables. An anchorage device of the cable consists of a nut and
a steel tube the outer surface of which is threaded. The loosened strands with
length of about 20cm from each end of the cable, are fixed with resin inside
the steel tube. This anchorage device was used also for uniaxial tensile tests.
To attach the anchorage device on the beam, an assembly made of steel was
installed on the web near the supports. This assembly was fastened to the web
with only prestressed PC bars which passed through the web. The saddles,
whose radius of curvature in the inside surface is 20cm, were installed at the
bottom of the beam to arrange the cables in straight segments between contact
points. In order to investigate the influence of bending angles of the cables on
the behavior of the beam and the cables, apparent bending angles of 7.1 a (Type
1) and 11.3° (Type 2) were investigated. The prestress force was introduced in
the two cables simultaneously with two hydraulic jacks checking the prestress
force by two load cells installed at the end of each cable.
Three kinds of cables were used for the external cables. One is a
prestressing steel strand (SWPR7A; twisted seven wires), another is aramid
fiber (AFRP) and the other is carbon fiber (CFRP). Fig. 2 shows the cables
used in this study. Fig. 3 shows the stress-strain relationships of the cables
obtained from tensile tests. The stress was obtained by dividing the tensile
load by the total sectional area including resin. Table 1 gives the experimental
variables. The introduced prestress forces of a cable are 68.6 kN for Beam No. 1
to 4, and 58.8 kN for Beams No. 5 and 6. These values are less than 50% of
the nominal tensile strength of each cable. The load was applied monotonically
to the beams except for Beam No.6, which was tested under fatigue loading.
Fig. 4 shows the overview of the test.
LOAD-DISPLACEMENT BEHAVIOR
Fig. 5 shows the load-displacement curves of Beam No.1 (Steel), No.3
(CFRP) and No.5 (AFRP). All beams show elastic behavior before the initial
crack occurs, and after the initial cracking, the stiffness of the beams decreased
gradually with increase of the displacement. During the loading, sharp sounds
due to slipping of the cables from the anchorage were often noted. However, it
was confirmed that there was no influence of the slipping on the overall behavior
of the beams and the prestress force of the cables because the amount of the
slipping was very small. The load-displacement curves of the beams reinforced
with FRP (No.3 and No.5) indicate generally almost the same behavior as
that reinforced with ordinary steel before the ultimate states. Concerning the
ultimate failure mode, two types of the failure mode were observed. Beams
No.1 (Steel) and No.2 (AFRP) failed in the compression failure of the concrete
while Beams No.3 (CFRP) and No.4 (CFRP) showed simultaneously both the
compression failure of the concrete and the breaking of the cables. The reason
1 Nominal
diameter is lOmm and nominal yield point is 343 MPa.
404
Mutsuyoshi and Machida
for the difference of the failure mode among the beams (No.1, 2, 3, and 4)
are considered as follows. Though the tensile breaking load of the CFRP is
smaller than those of the other cables, the same amount of the prestress force
was introduced to the cables. Moreover, the bending point of the cable at the
saddle should become a weak point for the tensile force. The breaking loads
of the CFRP cables, as measured by the load cells (109.3 kN for No.3 and
112.6 kN for No.4), were 77% and 80% respectively of the average breaking
load (142.1 kN). To confirm whether this phenomenon is true, the same two
beams (No. 3 and No. 4) were tested. Note that the design strength of FRP
must be carefully determined when FRP is bent at the saddle. It has been found
in this study that the influence of the bending angles (No.2 and No.5) on the
behavior of the beams and the cables is small.
Fig. 6 shows an example of the relationship between the load and the
change of the introduced prestress force in the cables. The introduced prestress
force hardly changed before the initial cracking, and after the initial cracking
it increased markedly with the increase of the applied load. This figure also
indicates that there is no difference in the changes of the prestress force between
both the cables up to the failure.
EVALUATION OF CRACKING LOAD AND FLEXURAL STRENGTH
Bond action between the concrete and the cable is not produced in the
external cables of PC structures. All forces from the concrete to the cable
are transferred only through the anchorage equipment installed on the web.
Therefore, Bernoulli-Euler's assumption, which means that the strain of the
cable is equal to that of the concrete located on the same place as the cable,
is not satisfied. In order to accurately compute the flexural strength of the
external cable PC structures, the idea that the total deformation of the concrete
located just on the cable equals that of the cable is required. However, since
this method is complicated, past equations proposed for unbonded PC members
were used to evaluate the flexural strength of the beams. The equations are
given as follows:
1) Pannel 's equation (2)
Mu = b · d 2 · fc' · qu (1 - 0.9qu)
in which,
qu = (qe
X
10 3
(1)
+.A)/ (1 + 2>.)
qe = PP · IJeJ/J~
>. = 12. PP. Ep. 0.0325/ f~. l
where Mu : ultimate flexural moment (KN-m), b : width of web (m), d :
effective height (m), f~ : compressive strength of concrete (MPa), pp : Apjbd,
AP : area of tendon, IJ ef : effective prestress of tendons (MPa), l : length of
tendon (m), EP : Young's modulus of tendon (GPa).
2) Mattock's equation (3)
IJpu = IJef
+ 1.4 · jjlOO · Pp + 68.6
(2)
FRP Reinforcement
in which,
apu :
405
stress of tendon at ultimate flexural state (MPa).
Table 2 gives the initial cracking load and the flexural strength obtained
from the tests and the calculation based on these equations. The initial cracking
load was calculated from the elastic theory assuming that the prestress force
at cracking is equal to the initial prestress force and the tensile strength of
concrete is 2.744 MPa. The experimental values of the initial cracking loads
in all beams are almost the same and agree well with the analytical ones.
The maximum flexural strengths obtained from Pannel 's equation are different
among the beams and differ from the experimental values. On the other hand,
the calculated strengths by Mattock's equation are generally smaller than the
experimental ones. Comparing the amount of the increase of the applied tensile
force in each cable, there is a good agreement in No.2 (AFRP) and No.5 (AFRP)
between the experimental and calculated values; but the calculated values are
smaller than the experimental ones in the other cables. The difference is due
to the fact that the equations used are derived for unbonded PC members using
ordinary steel tendons. In addition, the reinforcement ratio of the cable, which
significantly influences the flexural strength, depends strongly on how the total
sectional area of the cable is estimated. A definition of the sectional area of
FRP is required if these equations are to be used. In this case, the sectional
area including resin of FRP was used for the calculation. The evaluation of the
flexural strength must be investigated further if FRP is to be used as external
cables for PC structures.
ANALYSIS PROCEDURE FOR ACCURATE FLEXURAL STRENGTH
As mentioned above, Bernoulli-Euler's assumption is not satisfied at any
cross section in external-cable PC beams subjected to flexural moment. In order
to obtain accurate flexural strength of such members, compatibility condition on
deformation, that is the total deformation of the cable equals that of the concrete
located on the cable line, must be satisfied in every stage. A new analytical
method based on the compatibility condition on deformation was developed.
Assmnption for Analysis
The stress-strain curve of the longitudinal steel bars inside the web is
shown in Fig. 7(a). The yield point (294 MPa) and Young's modulus (206 GPa)
were determined from the tensile tests. The stress-strain curve of concrete is
expressed as a parabolic curve and a straight line as shown in Fig. 7(b). The
stress-strain relationships of the cables are shown in Fig. 3. To simplify the
calculation, it was assumed that the external cables are arranged linearly at the
position of Scm from the bottom of the web. Bernoulli-Euler's assumption is
satisfied for the concrete and the longitudinal reinforcements.
Method for Analysis
In unbonded PC and external-cable PC beams subjected to flexural moment, the strain of the cable is not equal to that of the concrete located at the
same position. A coefficient n, representing the ratio of the amount of the
increase in the cable strain from the effective prestressing state to that of the
concrete at an ultimate state, was introduced. The strain of the cable at the
406
Mutsuyoshi and Machida
ultimate state is given by Eq. (3).
Epu
=
Eps
-x)
d
+ Ct ( Ecp + Eu_X_
(3)
in which, Epu : strain of cable at ultimate state, Eps : strain of cable at effective
prestressing state, Ecp : strain of concrete at effective prestressing state, Eu :
strain of concrete at ultimate state (0.0035), a : the ratio of the amount of the
increase in the cable strain to that of the concrete located at the same position
as the cable, x : distance from the extreme compression fiber to the neutral axis
at a critical section.
The ultimate flexural strength can be calculated by the discrete element
method (4). As shown in Fig. 8, the section is divided into n discrete elements
(in this case, n = 160), and the strain distribution of the critical cross section at
the ultimate state is indicated. Assuming the values of a and a curvature when
the compressive extreme fiber reaches the ultimate strain of concrete (0.0035),
the strain at each element can be calculated. Using the calculated strain at
each element and the stress-strain relationships of materials, the stress of each
element and the cable can be obtained. Such procedure is repeated assuming
a new curvature until the equilibrium of the compression force and the tension
force at a critical section is satisfied. When the equilibrium of the internal
forces is satisfied, the ultimate flexural moment is obtained for the assumed a.
Using the calculated ultimate moment, a moment distribution along the beam
can be drawn as shown in Fig. 9. The next step is to calculate the elongation of
the cable between both the anchorages and the total deformation of the concrete
located just on the cable. The shear span of the beam is divided into m = 30
segments. From the moment distribution along the beam, each moment applied
at each cross section is calculated, and also a strain distribution in the cross
section can be obtained by using the discrete element method mentioned above.
Finally, the strain of the concrete located on the same position as the cable in
all the cross sections can be calculated. The following condition of deformation
compatibility is investigated;
(4)
in which, De = I;; (Ecp) i' De : total deformation of the concrete located just on
the cable, DJ = la [ccp + Eu(d- x)/x], D1 : total elongation of cable due to load,
l : cable length, 6 : allowable error (O.Olmm). If De ~ D1, then the assumed
value of a and the calculated moment are correct. If De =/= D1, a new a is
assumed and the above calculation is repeated until the compatibility condition
on deformation is satisfied. Fig. 10 shows the process of the analysis.
Comparison of Calculated and Experimental Results
Table 3 shows the flexural strengths, the values of u and the tension
forces of the cables at the ultimate state obtained from the analysis and the
tests. Comparing the values of a which is 0.35 for No.1 beam (steel cable)
and 0.36 for the other beams (No.2, 3, 4, and 5), one may say that the ratio
of the cable strain to the concrete strain in the critical cross section at the
ultimate state is almost the same independent of the materials used. From the
table, the ratios of the calculated flexural strengths to the experimental ones are
FRP Reinforcement
407
between 0.87-0.99. These values indicate that the calculated values agree well
with the experimental ones. In the tension forces of the cables at the ultimate
state, the calculated values are a little larger than the experimental ones. The
difference between both values may be attributed to the assumptions that the
cables are arranged linearly along the beam. Furthermore, though teflon sheets
were inserted between the saddles and the cables to eliminate friction force,
it is not clear how much friction force was eliminated at the ultimate state.
Although the analytical method is a little complicated, it is possible to evaluate
the behavior of PC beams reinforced with FRP as external cables with sufficient
accuracy.
FATIGUE BEHAVIOR OF PC BEAM WITH FRP AS EXTERNAL CABLES
PC beams using FRP as external cables under static loading showed
almost the same mechanical behavior as those using steel cables. Since FRP
is a non-corrosive material, it can be applied as external cables to retrofit or
strengthen damaged concrete structures. Furthermore, behavior of a PC beam
reinforced with FRP as external cables was investigated under fatigue loading.
CFRP was used for the external cables and the same beam as that used in the
static tests was used. Cracks of width about 0.3mm were introduced in the
web before CFRP cables were installed. The introduced prestress per one cable
was 41% of the nominal breaking load. Fatigue load, whose amplitude was
2.5% (lower load) and 40% (upper load) of the ultimate strength of the beam,
was applied to the beam. Cracks opened when the upper load was applied.
The fatigue loading was stopped at every 5 x 10 5 cycles, and static loads were
applied to the beam to investigate the behavior of the beam under static loading.
Fig. 11 shows both the load-displacement curves before the fatigue load and
after 4x 106 cycles. Fig. 12 shows the relationship between the cable tension and
applied static load on the beam. From the two figures, there was not any change
observed after 4x 106 cycles. Even after lOx JOi cycles, the load-displacement
curve showed almost the same behavior as that before the fatigue loading.
Therefore, within the range of this fatigue test results, PC beams reinforced
with CFRP as external cables have no problems under fatigue loading.
CONCLUSIONS
PC structures using FRP as external cables were newly developed. The
mechanical properties of six PC beams using three kinds of cables (steel, carbon,
aramid) were examined. From the test results, the following conclusions can
be drawn:
1. The load-displacement relationships of the beams reinforced with FRP
give almost the same behavior as that reinforced with ordinary steel.
2. The bending point of the cable at the saddle becomes a weak point for
the tensile force. The breaking load of the CFRP cable, measured by the
load cells, was about 80% of the average breaking load obtained from the
uniaxial tensile tests. Therefore, design strength of FRP must be carefully
determined when FRP is bent at the saddles or the deviators.
408
Mutsuyoshi and Machida
3. Ultimate flexural strength cannot be obtained accurately frorr. the equations proposed for unhanded PC members using steel cables.
4. Using compatibility condition on deformation, that is the elongation of
the cable equals that of the concrete located at the same position at the
ultimate state, the flexural strength and the tension force acting in the cable
can be evaluated accurately. However, this method is too complicated for
design use. A simple method to evaluate the flexural strength must be
established.
5. In order to overcome the mechanical properties of FRP such as lack of
plastic behavior and small elongation at rupture and to make use of FRP
effectively, PC structures reinforced with FRP as external cables can be
considered to be a good method. It was confirmed that FRP can be applied
to actual PC structures as external cables.
ACKNOWLEDGEMENTS
This research work has been supported in part by the Grant-in-Aid for
Scientific Research from the Japanese Ministry of Education. This support is
gratefully acknowledged.
REFERENCES
1. Morimoto, M.: "Evolutions of External Prestressing in Prestressed Concrete Structures," Journal of Prestressed Concrete, Japan, Vol.32, No.5,
1990.
2. Pannel F.: "The ultimate Moment of Resistance of Unhanded Prestressed
Concrete Beams," Magazine of Concrete Research, March, 1969.
3. Mattock, A., Yamazaki, J. and Kattula, B.: "Compressive Study on Prestressed Concrete Beams, With and Without Bond," Journal (~f ACI, Feb.
1971.
4. Park, R. and Paulay, T.: Reinforced Concrete Structures, John Wiley &
Sons, 1975.
5. Muguruma, H., F. Watanabe and Nishiyama, M.: "Ultimate Flexural
Strength of Unhanded PC Members," Journal of Prestressed Concrete,
Japan, Vol.26, No.1, Jan. 1984.
6. Mutsuyoshi, H., Machida, A. and Shiratori, N.: "Application of Carbon Fiber Reinforced Cables to Concrete Structures," IABSE Symposium
Brussels, 1990.
FRP Reinforcement
409
TABLE 1 - EXPERIMENTAL VARIABLES
Beam
No.
Material
Type
Cross Sectional Area (ad)
(Diameter:mm)
1
Steel
I
92. 9(12. 4)
2
AFRP
I
150. 0(14. 0)
3
CFRP
76. 0(12. 5)
4
CFRP
5
AFRP
6
CFRP
n
n
n
n
Introduced
Prestress
(kN)/a cable
Nominal Tensile
Strength of
Cable(kN)
160. 0
188. 2
68.6
142. I
76.0(12.5)
150. 0(14. 0)
76. 0(12. 5)
Concrete
Strength
(MPa)
36.3
188.2
58.8
142. I
Type I and n :Apparent bending angles are 7.1 • and II. 3' respectively.
Cross sectional area includes area of resin.
Diameter means nominal diameter.
TABLE2-COMPARISON OF INITIAL CRACKING LOAD AND ULTIMATE
FLEXURAL STRENGTII BETWEEN EXPERIMENTAL RESULTS AND
VALUES CALCULATED BY EQUATIONS FOR UNBONDED PC MEMBERS
-Beam No.
Cracking load
·
(kN)
Exp.
Flexural Strength
(kN)
Cal.
Exp. ( l1 p)
Pannel
Mattock{ lip)
I
83.3
86.2
217.6 (106.8)
197.0
64.6 (71. 5)
2
88.2
86.2
186.2 ( 68. 6)
287. 1
171. 5 (82. 3)
3
88.2
86.2
196.0 ( 84. 3)
167. 6
162.7 (68. 6)
4
93. 1
86.2
205.8 ( 94. I)
167. 6
162. 7 (68. 6)
5
81. 3
76.4
183.3 ( 81. 3)
253.8
159.7 (82. 3)
l1p:Amount of increase of tensile force in cable from effective prestressing
state
410
Mutsuyoshi and Machida
TABLE 3 - EXPERIMENTAL AND ANALYTICAL ULTIMATE STATE
Total tensile force
of cables (kN)
Flexural Strength
(kN)
Beam No.
a
~ f~or~c)
Exp.
Cal.
I
0.36
10. 3
247. 9
297 9 (I. 20)
217.6
213. 6 (0. 98)
2
0.35
II. 5
218. 5
247 9 (I. 13)
186.2
185. 2 (0. 99)
3
0.35
ll. 8
218. 5
237.2 (I. 09)
196.0
178.4 (0. 91)
4
0.35
ll. 8
235.2
237 2 (I. 00)
205.8
178. 4 (0. 87)
5
0.35
II. 9
196. 0
232. 3 (I. 19)
183. 3
175. 4 (0. 96)
mm)
Exp.
0
0
0
~f(or ~c):Calculated
(
Cal.
total elongation of cable(or concrete) at ultimate state.
):(Calculated value)/(Experimental value)
(mm)
Fig. !-Details of test beam
Steel
Fig. 2-FRP cables used
FRP Reinforcement
2.0
"@'
1.5
a.
Q.
Cl)
Cl)
1.0
LlJ
a:
1Cl)
0.5
0.0
0.0
0.5
2.0
STRAIN (%)
Fig. 3-Stress-strain relations of cables
Fig. 4-0verview of test
411
412
Mutsuyoshi and Machida
240
. l. . , ___ _
200
z
c
0
9
i
160 f--..-+··--.,.~ --+-- ... .....l.-... .. T ........ T ............
'
'
·.:
120
····-----··········l----··········•-·····-······•··············
80
··l·-----------·-r--·-----+--------------.;------------·--1----------·-j-------------··r------------
40
------+--------t-------
0
.
0
5
.
:
St~el (N~.l)
10
15
20
25
30
DISPLACEMENT (mm)
!
35
40
35
40
240
200
z
160
0
120
c
~....
irf
....- ~
"~ v
80
Carbon (No.3)
1!1
40
f.
0
0
5
10
15
20
25
30
DISPLACEMENT (mm)
240
200 f------f---·--+--..
z
c
0
<
9
....-
----+·----+-·---·-"~""""'""~"""'--·1--""
160 f-..........+...........................~""""'T............i., .............t-·············1······-·······
120
80
40
0
0
5
10
15
20
25
30
DISPLACEMENT (mm)
35
40
Fig. 5-Load-displacement curves of PC beams with external cables
FRP Reinforcement
413
2~--~~-,---,---,---,---,---,
· Steel (No. I) ··
200
z
e.
~
~
160
120
~
80
~
~
f- ...........,..............:............ i R· .;.........- ....~ ............ ,_.......... I
00
.............;.................,................ ..
.............,................,..................
20
40
60
80
100
120
1~
CABLE TENSION (kN)
2~----,--.---,---,---.---,---,
._<. . . .,. . . . . . .
200 · Carbon (No.3) .;................; ..............-.
z
160
Cl
120
e.
1
<(
0
-'
80
~
0
0
20
~
60
80
100
120
1~
CABLE TENSION (kN)
2~
Aramid (No.5)
200
z
160
Cl
120
e.
'
···---l---·············
<(
0
-'
80
·······+···--·-r-.. --.....
40
0
0
20
~
60
80
100
t'. . . . . . . .
120
1~
CABLE TENSION (kN)
Fig. 6-Beam load versus cable tension curves of PC beams with external
cables (bold line for right-side cable; dotted line for left-side cable)
414
Mutsuyoshi and Machida
400
f'c
0.85
... 300
~~
a.
~
rJJ
rJJ
rJJ
rJJ
w
a:
w
!ii
a:
frJJ
0.3
.
0.4
STRAIN(%)
STRAIN
(a) steel
(b) concrete
Fig. 7-Stress-strain relations of steel and concrete for analysis
~-+-I\
Strain
at ultimate state
Strain distribution at
\ effective prestressi~g
\state
\
\
\
\
\
\
\
0
\
0
IC:••I
c;
C:ps
:I
c •
Fig. 8-Discrete elements and strain profile at critical cross section
FRP Reinforcement
1111111111!
!1111111111 I
t;; f
Sf
=£f·l
Strain Distribution of Cable
Strain Distribution of Concrete at Cable
Fig. 9-Division of longitudinal element and distribution
of moment, cable strain, and concrete strain
415
416
Mutsuyoshi and Machida
Discrete element in critical cross section
Calculation of ultimate moment Mi
Division of longitudinal element
Calculation of Hi in each cross section
Calculation of strain distribution in each cross section
Total deformation of concrete c5c
No
Fig. tO-Proposed analytical procedure for computing ultimate
flexural strength of unbonded PC and externally-cabled PC members
FRP Reinforcement
417
100
Carbon (No.6) .
z
e
---t_zr~~I=:
60
Cl
c(
9
.
·--··-·-············t·--······· ········-····1··········--···-········!·····7-·--·-r·············-·--··-·
80
40
_
?/
- · - Before Fatigue Loading
'/
i
·~··· After4xUfcycles
·-·-···-~---··4············-··············i········
.
. ......................!
20
/,/
~"/
0
2
0
3
4
5
DISPLACEMENT (mm)
Fig. 11-Load-displacement curves before and after fatigue loading
100
--'- Before Fatigue Loading
• • '• After 4xl(/ cycles
80
z
e
Cl
c(
9
I
60 f-·····-···········-- ····! ·····--········k····-i··-·······--···- ········--!--·····--···-············
I
I
f-............................
!
······--·--····11·-····+··········--··
40
......... ;................
20
0
40
50
60
70
80
CABLE TENSION (kN)
Fig. 12-Applied load-cable tension curves before and after fatigue loading
SP 138-26
Flexural Characteristics of
Prestressed Concrete Beams
with CFRP Tendons
by T. Kato and N. Hayashida
Synopsis: This paper reports on the effects of improvement in
flexural characteristics and deformability(ductility) when using
unhanded CFRP(Carbon Fiber Reinforced Plastic) tendons in
prestressed concrete (PC) beams and bending fatigue characteristics
of bonded type PC beams with CFRP tendons. Based on the results of
flexural loading experiments. with PC beams using unhanded CFRP
tendons. failure modes shifted from CFRP tendons rupture type to
concrete crushing type, while deformation at the ultimate stage was
also changed greatly for the better. It was also succeeded in
ascertaining that effective prestressing force. and tensile
reinforcement quantity and variety are influential as factors
increasing deformability at the ultimate stage. Further. as the
result of bending fatigue tests of bonded type PC beams with CFRP
tendons. it was confirmed that reduction in ultimate flexural loads
of bonded type PC beams due to repetitive loading was not seen and
bending fatigue properties were favorable.
Keywords: Beams (supports); bending; ductility; fatigue (materials); fiber
reinforced plastics; flexural strength; prestressed concrete; prestressing
steels; unbonded prestressing
419
420
Kato and Hayashida
Takehiko Kato is a deputy general manager of Structural Development
Dept., Institute of Construction Technology, Kumagai Gumi Co .. Ltd.
His research interest is behavior of prestressed concrete
structures.
Norimitu Hayashida is a structural engineer of Technical Research &
Development Institute. Kumagai Gumi Co .. Ltd. His research interest
is behavior of prestressed concrete structures.
INTRODUCTION
A comparison of bonded type prestressed concrete (PC) beams using
post-tensioned CFRP (Carbon Fiber Reinforced Plastic) with
conventional PC beams. flexural cracking loads are found to be
roughly the same since they are affected by prestress force. but
yielding loads are slightly larger for the conventional PC beams
due to the influence of Young's modulus of the tendons. However.
when CFRP tendons have been used. flexural strengths increase even
after flexural yielding, and the maximum strengths are determined.
not by crushing of concrete in the compression zone. but in almost
all cases by rupturing of tendons and tensile reinforcement.
Consequently, deformation of beams at the ultimate stage will be
smaller than in case of using conventional PC steel tendons. The
main reasons for this are no yield points for FRP tendons and
elongation at rupture being considerably smaller compared with
conventional PC steel. The CFRP tendon strength increases uniformly
to rupture.
In order to incorporate such a beam in a structure, it is
necessary from the standpoint of safety of the structure for brittle
failure of the beam to be prevented and the beam must be capable
deforming with the frame. To improve deformability (ductility) of a
beam using FRP tendons. it is essential that using unhanded tendons
for no rupture of CFRP tendon at the ultimate load is known
generally as one of the methods for this purpose.
When applying FRP tendons to structures that are subjected to
repetitive loading, it is necessary for the fatigue characteristics
as PC beams to be ascertained.
The results of the flexural loading tests performed to understand
the flexural the characteristics and deformability (ductility) of PC
beams using CFRP tendons and the bending fatigue tests of bonded
type PC beams using CFRP tendons are discussed in this report.
FRP Reinforcement
FLEXURAL
421
LOADING TESTS
Specimens
A list of the specimens is given in Table 1. There was a total of
12 specimens with varieties of tendons and tensile reinforcing bars.
effective prestressing force. bonded or unhanded. and tensile
reinforcement ratio as parameters. The dimensions of the specimens
were cross section of 200 mm x 250 mm. and length of 2500 mm. The
tendons used were stranded type CFRP tendons of 12.5¢.PC steel rods
of 13¢. while tensile reinforcing bars were SD 295 06 steel bars
and CFRP bars of 6 ¢. The material characteristics are given in
table 2 and the specimen configuration in Fig. 1.
~ethod
of Testing
Application of load was by two-point loading of simple-beam type
with bearing span of 2200mm. pure flexure section of 400mm. shear
span ratio of 3. 6. Loading, as indicated in Table 3. was in the
form of load control until occurrence of cracking, after which. it
was displacement control. Measurements were of displacements and
crosssectional curvatures at the middle parts of PC beams. strains
of tendons. tensile reinforcement and concrete. and crack width. The
measuring locations are shown in Fig.2.
Analysis
Flexural cracking moments and flexural failure moments were
calculated using the Architectural Institute of Japan's PC Standards
<JJ approximate calculation equation and e-function method <4 ) and
these were compared with experimental values. Where thee-function
method is an analytical method by which elastoplastic behavior
against bending moment is obtained theoretically assuming the
compressive stress intensity-strain relationship of concrete to be
as shown in Fig 3.
Unbonded PC beams were evaluated using the conformity conditions
for the incremental elongate deformation of tendons accompanying
beam deformations at each time of loading and the total sum of
incremental elongate deformations of concrete at the locations of
tendons over the entire beam lengths.
Test Results
Flexural Cracking Load-- The results of the tests are given in
Table 4. Flexural cracking loads were governed by the initial
422
Kato and Hayashida
prestress force in all specimens. The influence of tendon type.
existence or non-existence of bonding. and quantity and variety of
tensile reinforcement did not affect cracking. Experimental and
analytical values were more or less in agreement.
Cracking Condition-- The final state of cracking is shown in Fig
4. Comparing with bonded and unbonded type in specimens using PC
steel rods and specimens using CFRP tendons. bonded-type specimens
had cracks more dispersed than unbonded-type specimens. Specimens
using CFRP tendons. both bonded and unbonded type, had dispersed
cracks and crack widths were smaller than PC steel rod specimens.
This is thought to have been due to the differences in the
capacities for bonding to concrete and modulus of elasticity. Seen
from the standpoint of the factor of tensile reinforcement.
specimens using CFRP reinforcement had cracks more dispersed than
specimens using ordinary steel reinforcing bars(SD 295). Cracks
were more dispersed as the larger the quantity of tensile
rei nf orcemen t.
'fielding Load and Stiffness after Yielding -- Yielding loads of
beams were obtained from the knees of M ¢ curves of experimental
values and values from strain gauges attached to tensile
rei nf orcemen t. The M- ¢ curves of the various specimens are shown
in Fig. 5 and the envelopes of the various specimen M-¢ curves are
shown in Fig. 6. Comparison of yielding loads by tendon type.
specimens using PC steel rods indicated higher values than
specimens using CFRP tendons. Bonded specimens showed higher
yielding loads than unbonded specimens. As for other factors.
yielding loads were higher for larger quantities of tensile
reinforcement. and the larger prestressing forces.
Next. concerning stiffness after flexural yielding of the beams.
specimens using PC steel rods. both bonded and unbonded. did not
show increases in strength after flexural yielding. With specimens
using CFRP tendons. unbonded specimens did not show increases in
strength after flexural yielding similar to specimens using PC
steel rods. However. with bonded specimens. strengths increased
uniformly after flexural yielding. The reason for this is that CFRP
tendons do not have yield points and exhibit elastic behavior until
rupture. Especially, with specimens using CFRP tendons and CFRP
tensile reinforcement together(GCP~ UCP2. UCP4). yield points do
not exist for both bonded or unbonded specimens. and flexural
strength increases consistently up to ultimate failure.
Ultimate Flexural Load-- At ultimate flexural load. bonded
specimens exhibited higher capacities than unbonded specimens for
both PC steel rods and CFRP tendons irrespective of tendon type.
With specimens using CFRP tendons. unbonded specimens showed higher
values for larger prestress forces. but with bonded specimens.
roughly the same capacities were indicated regardless of prestress
forces. For the influence of tensile reinforcement. specimens using
CFRP reinforcement showed higher capacities than those using steel
bars(SD 295). and capacities increased with larger reinforcement
quantity.
FRP Reinforcement
423
Failure Conditions-- Failures of specimens were all due to
bending. but three types of failure modes. concrete crushing type,
tendon rupture type. and tensile reinforcement rupture type. were
seen depending on the specimen. With specimens using PC steel rods
CPl. P2). both bonded and unbonded specimens failed by concrete
crushing. Examining failure of specimens using CFRP tendons.
bonded specimens CGCPl. GCP2) were tendon rupture type, whereas
with a specimen (GCP3) to which of no prestress in the first place.
crushing of concrete occurred followed immediately by rupturing of
tendons. With specimens using unhanded CFRP tendons. those
(UCP1.3.6. 7) using steel bars (SD 295) for tensile reinforcement
and a specimen CUCP5) with no tensile reinforcement were concrete
crushing type. However. unbonded specimens (UCP2.UCP4) using CFRP 6
¢ rods as tensile reinforcement were CFRP reinforcement rupture
type. That CFRP exhibits elastic behavior until rupture and that
elongation of tendons at rupture is small may be considered as
reasons for rupture of CFRP tendons and tensile reinforcement.
Regarding failure properties. concrete crushing types showed
stable behaviors until deformation limits after yielding. but CFRP
tendon rupture types and tensile reinforcement rupture types were
extremely brittle.
Deformation Properties -- A great improvement in deformation
(ductilitY) at ultimate flexure loads of PC beams was seen when CFRP
tendons were unbonded. Comparing the ultimate deformations of the
various specimens by ratios of curvature at ultimate state to
curvature at yielding (curvature ductility U¢= ¢u/¢y), from the
test results given in Table 4. a specimen (UCP3) using unbonded
CFRP tendons had roughly the same curvature ductility ( U¢=10. 7 to
14. 0) as specimens (Pl. P2) using PC steel rods. However. with
specimens (UCP2. UCP4) using CFRP bars as tensile reinforcement. the
ultimate strengths were decided by the rupturing of CFRP
reinforcement. Strengths were increased as the amounts of CFRP
reinforcement were increased. but curvatures were more or less
constant and improvement effects of ductility were not seen.
The relationship between load increment(~P) in tendon loading and
curvature(¢) of beam during testing of specimens with unbonded
CFRP tendons are shown in Fig 7.
Since CFRP tendons do not have yield points. the loads increment
~p of unbonded CFRP tendons in loading showed a trend of increasing
more or less linearly until failure of the beams. Especially, with
specimens CUCP2. UCP4) using CFRP bars as tensile reinforcement.
unbonded CFRP tendons had larger increment~P of load than specimens
(UCP1.3.5, 6. 7) using steel bars. In specimen (P2) using PC steel
rods. there was increase 1inear 1y in the range where deformation was
small. but when deformation became to larger. there was a trend of
load increment ~p become small as curvature increased.
Analytical Study of Deformability (Ductility) --Analytical
studies were made of factors affecting ultimate deformability of
unhanded PC beams using CFRP tendons. The analyses were performed by
thee-function method using the results of measurements of load
increment ~p of unbonded CFRP tendons. The following may be said
from the results of analyses shown in Fig.8 :
424
Kato and Hayashida
a) When tensile reinforcement ratio is held constant at Pt=0.21%
(steel bars. SD295).and effective prestressing force is varied
between 0. 0 Pu and 0. 6 Pu. cracking load and maximum strength
become greater as the larger effective prestressing force in case
of bonded specimens. but deformation at the ultimate stage is
roughly constant. In case of unhanded specimens. cracking loads and
maximum strengths are same tendency to case of bonded specimens.
but there is a tendency for deformation at the ultimate stage to
become larger as the smaller the prestressing force.
b) With effective prestress constant at 0. 6 Pu. when tensile
reinforcement ratio Pt (steel bars. SD 295) is varied between 0.0%
and 0.49%. yielding load and maximum strength are increased more the
higher the tensile reinforcement ratio for bonded and unhanded
specimens. while ultimate deformation has a tendency to become
larger as the lower the tensile reinforcement ratio.
c) With effective prestressing force constant at 0.6 Pu. when
tensile reinforcement ratio Pt is varied between 0.0% and 0.36%
using CFRP bars and SD 295 steel bars for tensile reinforcement.
maximum strengths are higher when using CFRP bars than when using
SD 295 steel bars for both bonded and unhanded specimens. while
ultimate deformations are greater for specimens using SD 295 steel
bars than those using CFRP bars. Further. there is a tendency for
ultimate deformation to be larger for lower tensile reinforcement
ratio Pt.
BENDING FATIGUE TESTS
Specimens
A list of specimens is given in Table 5. The specimens consisted
of three for static bending and three for bending fatigue. a total
of six specimens with variety of CFRP tendons (strand type, rod
type) and form of tendon arrangement (straight line. curve) as
parameters. The specimen configuration is shown in Fig.9 and
material characteristics are given in Table 6.
Specimen dimensions were cross section of 200mm x 250mm and length
of 2500mm, with two cables of CFRP tendons placed in the cross
section. Prestressing force was 8. 7 tons(85.3 KN) per cable.
~ethod
of Testing
The method of testing was by two-point loading of simple-beam type
with bearing span of 2200 mm and equal bending moment section of
400 mm. The static flexural loading method is shown in Table 7, and
the repetitive loading method of the bending fatigue tests in Table
8.
Repetitive loading was done with loading level divided into two
stages. At the first stage, static loading was applied until
occurrence of cracking, after which repetitive loading of 500,000
FRP Reinforcement
425
cycles was done with upper limit load of the cracking load and lower
limit load as 2 tons (19.612 KN). In the second stage, repetitive
loading of 550,000 cycles was applied with upper limit load 50 to
70% of the failure load obtained from the results of static
flexural loading tests and the lower limit load 2 tons (19.612 KN).
A total of 1.050. 000 cycles were applied. after which one-way
static loading was carried out to failure. The excitation waveform
was a sine waves and the excitation frequency 3Hz. Measurement were
obtained at the designated number of cycles. suspending excitation
at the specified number of cycles on carrying out static loading.
The item of measurement were displacement at beam middle. strains
of tendons and concrete. and condition of cracking.
Test Results
U 1timate Flexural Loads -- The flexural failure loads of
specimensCCFRP N0.4. CFRP N0.5. CFRP N0.6) subjected to repetitive
loading compared with results of static bending tests(CFRP NO.1.
CFRP N0.2. CFRP N0.3) are as shown in Table 9. Where ultimate
flexural loads for each specimens are approximately indicated the
same degree. and reductions in maxi mum 1oad due to repetitive
loading were not seen. Maximum loads were determined for all
specimens by rupturing of CFRP tendons. Experimental values and
calculated values were compared according to the Architectural
Institute of Japan's PC Standards<!> . they were more or less the
same. However. maximum loads of specimens (N0.3. N0.6) in the
experimental value were smaller than the calculated value. this
reason is that on specimens (N0.3. N0.6). in addition to no tensile
reinforcing bars. arrangement of CFRP tendons were curve. these
specimens have larger cracks at turning over the place of CFRP
tendon' s curve.
Condition of Cracking-- The condition of cracking is shown in
Fig. 10. On comparison of the condition of crack occurrence by static
bending and flexural fatigue results. crack widths and numbers of
cracks were roughly the same for each specimens. Cracks that
occurred following the shapes in which tendons were arrayed had been
formed immediately after failure of PC beams. and may be considered
to have occurred due to the effect of prestressing load having been
released as a result of rupturing of CFRP tendons.
Load-Displacement Curve -- The load-displacement curve of the PC
beam subJected to repetitive flexural loading and static bending
test results are shown in Fig. 11.
In repetitive loading at the first and second loading stage, each
specimens (CFRP N0.4. CFRP N0.5. CFRP N0.6) did not show changes in
curve gradients. and hysteresis curves were roughly the same as
static bending test results. However. as the number of cycles of
loading increased. a trend of increase was seen in amount of
displacement at action of upper limit load and in residual
displacement.
426
Kato and Hayashida
Change in Equivalent Rigidity-- Defining the gradient
(load/displacement) of the straight line connecting any point on the
plane at various loads and the origin of the load-displacement
curve as equivalent rigidity. the relationship of equivalent
rigidity and displacement is shown in Fig. 12.
The equivalent rigidities of specimens subjected to repetitive
loading and static bending specimens indicated more or less the
same changes with each specimens. and influences of repetitive
loading were not seen. Further. changes in equivalent rigidities
were constant up to occurrence of cracking. but showed sudden
declines in the range of displacement of 2 to 5 mm. after which a
trend was seen for a gentle decline with increases in displacement
up to ultimate failure.
CONCLUSIONS
The conclusions below were drawn regarding flexural
characteristics and flexural fatigue properties of PC beams using
CFRP tendons.
CD The cracking load of a specimen with unbonded CFRP tendons is
governed by prestressing force similarly to bonded specimens and
~ecimens using conventional PC steel rods.
~The cracking load. yield load. and ultimate flexural loads can be
roughly estimated approximately equation by the conventional the
Architectural Institute of Japan's PC Standards<!> and theefunction method.
aD When CFRP tendons are used in PC beams. bonded specimens will
exhibit brittle failure properties where failure occurs from
rupturing of CFRP tendons. but with unhanded specimens. failure
occurs from crushing of concrete. and the ultimated
eformability(ductility) can be improved to roughly the same degree
as when using conventional PC steel rods.
QD Regarding deformation at the ultimate stage of PC beams using
CFRP tendons. factors such as the size of effective prestressing
force. quantity and variety (CFRP bars. SD 295 steel bars) of
tensile reinforcement have influences.
aD Regarding the bending fatigue characteristics. the maximum
flexural strength. load-displacement curve. and variation in
equivalent rigidity of a PC beam subjected to repetitive flexural
loading in excess of one million cycles are roughly the same as
results of static bending loading. The effect of repetitive loading
not recognizable and it may be ascertained that fatigue properties
are good.
REFERENCES
Architectural lnsti tute of Japan. Standards for Design and
Construction of Prestressed Concrete . with Commentary, 1987.
(1)
FRP Reinforcement
427
(2) Architectural lnsti tute of Japan. Recommended Practice for
Structural Design and Construction of prestressed Reinforced
Concrete(Typeill PC). with Commentary, 1986
(3) Japan Society of Civil Engineers. Application of Continuous Fiber
Reinforcing Material to Concrete Structures .April 1992.
(4) Kat o. Ish i basi. and Kawaguchi : Development of CFRP Tendon
Anchoring device and PC Girder Flexural loading Experiment.
Prestressed Concrete Engineering Association. Sept. 1988
(5) Kato. Hayashida. Noridomi. and Kubo :Study on Method of Improving
Ductility of Beam Member UsingFRP Tendons. Prestressed Concrete
Engineering Association. Vol.3t N~ 1.1992.
(6) Maruyama. Ito. et al. :Flexural Fatigue Characteristics of PC
Beams Using CFRP and AFRP Rods. Concrete Research and Technology,
Vol.12. No.1. 1990.
TABLE 1 - UST OF SPECIMENS OF BENDING TESTS
Effective
Tensile
Prestressing ReinforceForce
ment Ratio
( KN )
(%)
p 1
bonded
0.21 ST
85.31
p 2
unbonded
0.21 ST
GCP1
0.21 ST
85.31
bonded
(0. 6 Pu)
GCP2
0.21 CF
0.0
GCP3
0. 21 ST
UCPl
0.36 ST
UCP2
0.36 CF
85. 31
(0. 6 Pu)
UCP3
0. 21 ST
unbonded
UCP4
0.21 CF
UCP5
0.0
UCP6
56.87
0.21 ST
(0.4 Pu)
UCP7
0. 0
0. 21 ST
ST tensile reinforcing steel bars CSD295 D6)
CF tensile reinforcing CFRP bars (CFRP ¢ 6)
Pu tensile roads of CFRP tendon (12.5 ¢)
Specimen Bonded
or
NO.
Unbonded
T~re
Tendons
PC steel
Rod
(13 ¢)
CFRP
tendon
(12. 5 ¢)
CFRP
tendon
(12. 5 ¢)
428
Kato and Hayashida
TABLE 2- CHARACTERISTIC OF MATERIALS
Materials
Area
(mm 2 )
PC steel bar13 ¢
CFRP tendon 12. 5¢
132. 7
76
ST bar SD295 06
CF bar CFRP ¢6
32. 7
28.3
Materials
Concrete
Compressive
Strength
( N/mm 2 )
40
Yield
Tensile Elastic
Strength Strength Modulus
( N/mm 2 ) ( N/mm 2 ) (KN/mm 2 )
1340
1387
206
-1806
137
206
353
530
-127
1628
Elastic
(KN/mm )
Poisson's
Ratio
25
0. 143
Modulu~
TABLE 3 -LOADING METHOD
Load Stage
NO.1
N0.2
N0.3
NO.4
N0.5
NO.6
step cracking point
step o= 3 mm
step o= 6 mm
step o = 18 mm
step o = 36 mm
step crushing point
Control
Method
load
deformation
deformation
deformation
deformation
deformation
Cycle
2
2
2
2
2
1
Extension
Ratio
( %)
11. 5
1.6
22. 7
1.3
TABLE 4 - RESULT OF FLEXURAL LOADING TESTS
=T==...:.:c:..:..:;r:-:=--t-:-,...:..:..:r=:,...:=~:.:...:c.:..:,:..:..,-...,....--j ~~~nln;
Specimen
NO.
(~.f}¢y)
~····:t·r~···l····it-ii··l····it·ci~····l··it~·:····l··aa:7s~-~:l64.ai·:ss·l····~r:-~·ci···+···it·~~
····i!:~i··l····it~r-+···t~:·······l····:t·:~···l····itii···l····i-i·ri~·-····1···::·!~····1··27:3&~-I:I~\s:ar·l····ii~·::····l····i-t·:·····
····it·:i··l····i·N~··I····t~i······+···~t-~:···l····il:·:I··I····it·~~-·····I··It~-~---·l···so:iis~rs4·:7i·l····i·cit~ci····l····-t·~:-····
····ini··l····it:··l····i:-ii······+·············l········ ····+···············+i·:t~·ci····l···4a:i·s~-I:I6 \a·:94·1······~::-~i····l·················
p 1
P2
GCPl
GCP2
38.83 I
17.46 I
34.32
15.49
GCP3
18 53
· 18.53
· : ··
UCPl
32.75
34.32
4. 62
4. oo
fi: cir ····I·· -:ti~····I·· ·so: 69~-~: f3i:75·I····i~::-~····I· ····:: ·ii
~:·:~·······I· ·--~::-~i·-1 · · i~:-~~-· ·I· ···it ci:· ··· ·l···i~:-~}···1· ··29."i32~J\s: a·s·l···· i i:· ~ci····l···it ~i
.I. .....8.a: ·3434·· .I.· ii:- ~ci······I····~:~~· ·I···· i~:: ··I··
.... 14.71
15.49
9
t~: ·····1··············1··············1·· ······· ·····l·i~:i~····l··s7."7s5·Yss."as·l···-~·it!!····:·····=::·····
UCP3
···iNH····iN:··I···t~:-·····1· -:t-~ 1···~:-iH····ftri~ ·l·ii:·H ·1·25."ii 2~}227:as·l :~:::: ··I· ·iN~
24
1
1................
uc P 4 I·
--=:
:::::~ 4a:os'~P43."s41····it~ci
. o
35.50
34.32
UCP2
15.98
15.49
.....
29.22
3. 73
13.24
24 I
13.14
14 I
6. g,
6. 80
I 37.36
1.......... 1_7:117
12.00
12. oo
I 41.78
I 18.73 '
·I·
17.75
_3_43. 69
235. oo
50.54
19.58
2i.a3··· 13.04 7. 00
25:i·i~f2i.."ia·l····~::-~ci····l····it·:·····
35.60
45.11
15.98
12.00
38.73
···· iHi· ·l·····t ~t-·l····t :6·······1· ···it·ii··l· ····j:·~~·-1· ···i·::-ci:···· · 40.89 ···2s·:i·i~·IPis:i·a·l····m:-~ci····l····iN:·····
uc p s I
UCP7
29.42
· is:·so
=: :
Note
:1)
:2
:3~
39.03
17.55
28.09
50.60
Pailure Modes of
Speciaens u
failure oode 1
a .... =47. 65 ..
failure oode 1
a..r.=42.5~
failure oode 2
a .... =27. 5:.
failure oode 2
a .... =32. 1s.
failure ..de 2
a..r.=51.64•
failure ..de I
a.... =56. 4ha
failure ..de 4
a .... =5o. 91•
failure oode I
a..r.=so.ow
failure ..de 3
a..r.=34.lw
hi lure ..de I
a..r.=47.2w
failure oode I
a.... =52.4&..
failure .. de I
a .... =7o.u..
!t~lY~~e·;~ae0£: s~~g~~~~~ of ~~~~ui:n:Y?: ~~i~f~~~e~!n¥~~cf:Ifu~~ ~3ec~~P~~~~~~= ~?n~~R~~i!~~T.:o~:i~io~~~~t ~~dc~~~s~egfoo
ncrete· 1n !he compre$Sive zone
upp r a ue :Hxpenmental yalue. Th(t low r value :Ana sis value f e-functio ethod.
.
Heexuraf
ult1mate point U01pent s an~tiYsu vafue. the teJI value • :Xrchitecturaf Tostitute of Japans PC Standards
, the right value : e-runchon method.
.
~
:::0
~
:::0
(1)
~
~
3
(1)
::s
.....
cu."'
~
N
\0
430
Kato and Hayashida
TABLE 5 - UST OF SPECIMEN OF BENDING FATIGUE TESTS
SNecimen
0.
CFRP NO. 1
CFRP N0.2
CFRP N0.3
CFRP N0.4
CFRP N0.5
CFRP NO. 6
Variety of
Tendons
bar type
strand type
strand type
bar type
strand type
strand type
Using
Tendons
2-CFRP12¢
2-CFRP 12. 5 ¢
2-CFRP12. 5¢
2-CFRP12 ¢_
2-CFRP12. 5 ¢
2-CFRP12. 5¢
Arramgement
of Tendons
straight 1ine
straight 1ine
curve
straight 1ine
straight line
curve
Variety
of Test
static
test
fatigue
test
TABLE 6- CHARACTERISTIC OF MATERIALS
Diameter Area
T~~e
(mm 2 )
(mm)
Tendon
Strand type
76.0
12.5
tendon
Bar type
113. 1
13.0
tendon
Tensile
Load
( KN )
Tensile Elastic
Strength Modulus
( N/mm 2 ) CKN/mm 2 )
137
1804
137
1.6
172
1520
127
1.3
TABLE 7- LOADING METHOD (STATIC TESTS)
Control
Method
cracking point load
deformation
o= 3 mm
deformation
o= 6 mm
deformation
o = 18 mm
deformation
o = 36 mm
crushing point deformation
Load Stage
NO.1
N0.2
N0.3
N0.4
N0.5
N0.6
step
step
step
step
step
step
Extension
Ratio
(%)
Cycle
2
2
2
2
2
1
FRP Reinforcement
431
TABLE 8 -LOADING METHOD (FATIGUE TESTS)
Specimen
NO.
CFRP
N0.4
CFRP
N0.5
CFRP
N0.6
Loading
Stage
NO.1
N0.2
NO.1
N0.2
NO.1
N0.2
Lower
Load
Upper
Load
19.6
19. 6
19.6
19.6
19.6
19.6
49.0
68. 6
53.9
73.5
44. 1
58.8
( KN)
step
step
step
step
step
step
( KN)
Frequ- Cycle of
Repetitive
ency
(Hz) Loads
500.000
550.000
3
500.000
550.000
3
500.000
550.000
3
TABLE 9- RESULTS OF BENDING STATIC AND FATIGUE TESTS
Specimen
NO.
CFRP NO. 1
CFRP N0.2
CFRP
CFRP
CFRP
CFRP
N0.3
N0.4
N0.5
NO. 6
Variety
of Test
static
test
fatigue
test
Ultimate Flexural Load ( KN)
Exverime~Jjal
al ue I
142.7
120.4
Analltical
Va ue ®
139. 7
123.4
86. 7
137.6
123.4
139.7
121. 5
81. 8
123.4
123.4
CD/®
1. 02
0. 98
0. 71
0.99
0.98
0. 67
Failure Type
of Beam
rupture of
CFRP tendons
rupture of
CFRP tendons
432
Kato and Hayashida
P/2
P/2
~G;Q-,[J'fHU~,~!JtJ tI ktfHtA IJ!H-{ttj.
~
(Unit :nun)
1150!
I
900
2~000
I
900
1150
·
Pt=O. 21X
Pt=O. 36X
Pt=O. OX
Pt=O. 21X
Pt=O. 36X
GJ][GJ]GJ]G]G]
PI P2
WIGm
UCP3 UCP6 UCP7
UCPI
UCPS
GCP2
UCPZ
~N
Fig. !-Dimension of specimens
P/2
~c~p~ 1 "
!2ool
5Q
0
900
P/2
~~~
u,
,. 400
0
900
2S
il5C
2500
(Unit:mm)
Fig. 2-Measuring locations
I
FRP Reinforcement
e-function method
r=
6.75 (
e-0. 812_ e-1. 218)
E
f = E CO
r
1 ------
'
_
a
~ = fC'
fc': maximum compressive stress
of concrete
ceo: strain (a= fc')
0~--~---------------f
Fig. 3-Stress-strain relationship of concrete
433
434
Kato and Hayashida
I
• I
lucPB
Wmax= 7.30 mrn
lucP7
A
Wmax= 5. 79 mrn
A
•
Fig. 4-Final state of cracking
FRP Reinforcement
60 ,-,-,.-..,--,---,--,
60
a 5o
z
···~ ...
lO
- :Experiment
!::. : e-funct ion
aetbod
""
:'50
~
~ (0
~30
"'-
20
c
; 10
"e
0
~
..
0
50 100 150 200 250 300
curvature ¢( x!o-•;. l
p
1
50
z
lO
1- ..:.. J•:·'····'
:....... :
·-i
""
30
20
-c 20
0
B
0 50 100 150 200 250 300
curvature ¢ ( x 10-• /m )
GCP1
50
z
lO
:Experiment
:e-function
oethod
""
30
..... ~- ..
oa-~~~~~_L_J
~
0
50
~
(0
z
lO
~
""
30
~
""
30
"'
20
20
10
"10
e
c
""
0
"
0
60
_;, lO
"'
-
:Experiaent
!::. :e-function
aethod
10
0
e
50 100 150 200 250 300
curvature ¢ ( xto-•;m)
UCP1
60 ,.-,--.,..-...,--,--,--,
:Experiment
!::. :e-function
aethod
~so
~ 10
e
curvature¢( xto-•;m)
UCP5
oL-a--L--~~-L_J
50 100 150 200 250 300
0 50 100 150 200 250 300
curvature ¢ ( xJ0-'1•)
UCP4
curvature ¢( xJo-•/m)
UCP3
60
-;;50
60
!::.
-e
50
z
lO
~
(0
"'
30
~
""
10
""
~
•
•
20
""
20
10
:; 10
0
50 100 150 200 250 300
30
20
~
e
r-.-.---.-,---,--,
-
z lO
""
0
0 50 100 150 200 250 300
' ; !iO
~
c
curvature ¢ ( xlO-'/m)
UCP2
-; 50
50 100 150 200 250 300
"e
60 ,.-.,-.,--,---,--....,.--,
-; 50
·········
30
curvature ¢ ( x W'/m )
GCP3
60
s
50
z lO
""
c
IL.I..L~~'-..L....J....-..J
0
0 50 100 150 200 250 300
curvature ¢( XlO-'/m)
GCP2
60
s
"' 20
"e 10
e
0~~-L--~~_L_J
curvature¢( XlO-'/m)
P2
c
~ 10
..... ~- ...
50 100 150 200 250 300
:20
20
- :Experiment
!::. :e-function
aethod
~ 10
10
s
~
~30
~
30
60 ,-.,...-,.--,---,---,---,
60 ,-.,...-,.--,--,---,-,
s
40
0
0
60 ,-.,...-,.--,--,---,---,
"";!so
0 ........,_.......,:::1--...0:::'-L.-..J
0 50 100 150 200 250 300
curvature ¢ ( xJo-•tm)
UCP5
.
0
B
50 100 150 200 250 300
curvature ¢ ( XIO-'/•)
UCP7
Fig. 5-Moment-curvature (M-¢) curves
435
436
Kato and Hayashida
50
-;;; 40
z
~
30
::e:
-;::; 20
e
0
e 10
Q)
150
200
250
300
curvature ¢ ( x 10- 3 /m)
Fig. 6-Moment-curvature (M-¢1) envelope curves
2 10 0
,..---,---,..---,--""""T--.,.---,
:::.:::
.__.,
0...
<l
'+-<
0
.....,
80
40
c:::
<l.)
s
~
(..)
20
c:::
100
150
200
250
300
curvature¢ ( Xl0- 3 /m)
Fig. ?-Relationship of load increment l!J> of
tendons and curvature ¢1 of beams
FRP Reinforcement
50
r--.---.--,---r--.---,
- bonded type
--- unbonded type
-;;;
40
• _/
,...r (2)
""
~ 30
GJ
""
::E
SD 29~ Pt=O. 2U
~ P-0. 6 Pu
P=O. 4 Pu
P=O. 2 Pu
P=O. 0 Pu
----------®
-----(ID
~ 10
---
E3
~ 10
50
100
150
200
250
OL-~---L--~--~~~~
300
curvature ¢ ( x 10-'/m)
Relationship of effective prestressing
and M- ¢ curve ( Pt= 0. 21", constant)
---G)
- bonded type
--- unbooded type
E3
oL-~---L--~--~~~~
0
........
=
a.>
---~
~---
20
::E
---~·
=
a.>
P=O. 6 Pu
SD 295
Pt=0.49X
Pt=O. 36X
Pt=O. 21X
Pt=O. 0 X
~
•• - --- 00 -
::::::::::.~-- ®
20
~
(1)
437
0
(2)
50
100
150
200
250
300
curvature ¢ ( x 10-'/m)
Relationship of tensile reinforcement
ratios and M- ¢ curve CP=O. 6Pu. constant)
80 r--.---.--,---r--.---,
bonded type
<D
-;;; 60
unbonded type
P=O. 6 Pu
~:CFRP¢6
~
E3
<DPt=O. 36%
®Pt=O. 21"
@
-- :SD295 D6
.<- @ ~Pt=O. 36%
.-:--·
4 Pt=O. 211i
;:-·/ @
5Pt=O.O%
®
~
=
~ 20
60
=
'_::::::::::::::[email protected]
0
----
~
_,:---·
~ 20
0
E3
E3
0'-----'----'----L..--_..___....____,
0
50
100
150
200
250
300
curvature ¢ ( x 1o·• /m )
(3) Relationship of tensile reinforcement
variety, ratios and M- ¢ curve for
bonded specimens (P=O. 6Pu. constant)
P=O. 6 Pu
-:CFRP¢6
<DPt=O. 36%
®Pt=O.Zl%
-- :SD295 06
®Pt=0.36%
@Pt=O.W:
@
@Pt=O.O%
.,._ ...... --
®
0'-----'----'---_.____..___....____,
0
50
100
!50
200
250
300
curvature ¢ ( x 10-'/m )
(4) Relationship of tensile reinforcement
variety, ratios and M- ¢ curve for
unbonded specimens CP=O. 6Pu, constant)
Fig. 8--Moment-curvature (M-¢) curves (analysis)
438
Kato and Hayashida
(Unit :mm)
Fig. 9-Dimension of specimens (bending fatigue tests)
static Bending tests
Fig. 10-Final state of cracking
439
FRP Reinforcement
100
100
80
80
10
60
~ 80
""60
40
~
""
""
40
"" 40
~-.
~
.
c.. 20
20
~
~
100
~
"'
0
"'
1Z
0
16
20
deformationo (mm)
CFRP- NO. I
12
16
20
0
12
deformation o (mm)
CFRP- NO.3
deformation o (mm)
CFRP- N0.2
(I) static Bending tests
100
80
z
""
100
80
80
60
z
60
40
""
"" 60
~
40
~
20
c.. 20
"" 40
c.. 20
.
~
~
"'
12
0
16
20
deformation o (mm)
CFRP- N0.4
~
0
12
16
20
"'
0
12
20
16
deformation o (om)
CFRP- N0.6
deformation 0 ( .. )
CFRP- NO.5
(2) Bending fatigue tests
Fig. 11-Load-deformation cutve
~ 40
"" 30
.:::
:§:
....ll
Ostalic Bending
test (NO. I)
~ 40
;so
t>B~~~~n~Nb~lJ&"e
20
:;, 20
·;: 10
~ 10
~
.
~
=
""
~
~
~
c
>
'j
~50
50
i
10
20
30
deformation
CFRP-NO.
40
50
o(mm)
I, NO. 4
.:="'
=
""
~
0
Ostatic Bending
test (N0.2)
t>Bend ing fa li gue
test (NO.5)
e....
50
40
""!:' 30
Ostatic Bending
lest (N0.3)
t>B~~~~n~Nb~Wue
:;, 20
"-..
~ 10
0 5 10 15 20 25 30 35 40 45 so
deformation (mm)
CFRP-NO. 2, NO. 5
o
.
>
=
""
~
10
20
30
deformation
40
o (mm)
50
CFRP-NO. 3. NO. 6
Fig. 12-Relationship of equivalent rigidity and displacement
SP 138-27
Partially Prestressed Beams
with Carbon Fiber Composite
Strands: Preliminary
Tests Evaluation
by A. E. N aaman, K.H. Tan,
S.M. Jeong, and F.M. Alkhairi
Synopsis. The use of fiber reinforced plastic reinforcement in
reinforced and prestressed concrete structures is gaining increased
attention. This paper describes the results of a preliminary experimental
program in which strands made of carbon fiber composites (trade name
CFCC - Carbon Fiber Composite Cable) were used as pretensioning
reinforcement in two partially prestressed concrete T beams. The
beams were ten foot in length and 12 inches in depth and contained, in
addition to the carbon fiber strands, conventional reinforcing bars.
Experience gained with the stressing, anchoring, and releasing of
CFCC strands is described. Relevant test results regarding loaddeflection response, curvature, stress-increase in the reinforcement with
increased load, cracking and crack widths, and failure modes are
reported, and compared to results obtained from similar tests using
prestressing steel strands. The load deflection response of beams
prestressed with CFCC strands showed generally a trilinear ascending
branch with decreasing slope up to maximum load. Deflections and
crack widths were generally small but increased rapidly upon yielding of
the non-prestressed steel reinforcement. The post-peak response was
characterized by rapid step-wise decrease in load due to successive
failures of the CFCC strands, and stabilization at about the load-carrying
capacity of the remaining steel reinforcing bars. The presence of
reinforcing bars helped the beams sustain large deflections before
crushing of the concrete in the compression zone. Analytical predictions
of the load-deflection response using a nonlinear analysis method were
used and led to reasonable agreement with experimental results.
Keywords: Beams (supports); carbon; composite materials; cracking
(fracturing); deflection; fiber reinforced plastics; fibers; flexure; prestressing:
prestressed concrete; prestressing steels; strains
441
442
Naaman et al
ACI Fellow Antoine E. Naaman, is Professor of Civil Engineering in
the Department of Civil and Environmental Engineering at the University
of Michigan, Ann Arbor.
ACI member Kiang Hwee Tan, is Senior Lecturer in the
Departement of Civil Engineering at the National University of
Singapore, Kent Ridge. He was on sabbatical leave at the University of
Michigan during the Fall of 1991 and the Winter of 1992.
Sang Mo Jeong, is a graduate student and doctoral candidate in the
Department of Civil and Environmental Engineering at the University of
Michigan, Ann Arbor.
ACI member Fadi M. Alkhairi, obtained his Ph.D. degree from the
University of Michigan, Ann Arbor, in 1991, and is currently with T.Y. Lin
International, Alexandria Virginia.
INTRODUCTION
The use of non-metallic reinforcement, particularly fiber
reinforced plastics or polymers, utilizing high performance fibers such as
carbon, glass, aramid (kevlar), and others, is seen primarily as a means
to avoid corrosion problems otherwise encountered in concrete
structures when using conventional steel reinforcing bars or
prestressing tendons. Fiber reinforced plastic (FRP) reinforcements in
the form of bars, tendons, strands, and two or three dimensional
meshes, are being explored in reinforced and prestressed concrete for
two types of applications, namely, new structures and repair or
strengthening of existing structures. For new structures, prestressing,
bonded or unbonded, internal or external, seems a natural field of
application of FRP reinforcement because of their high strength,
corrosion resistance, lower unit weight, easiness in coiling and
handling, good damping and fatigue behavior, and low relaxation
losses [1-3]. Several small bridges have been already built in Germany
and Japan using such reinforcement.
The study presented in this paper deals with the flexural behavior
of partially prestressed concrete beams using carbon fiber composite
strands, trade-named CFCC (Carbon Fiber Composite Cable) as
prestressed reinforcement.
Two T-beams with targeted global
reinforcing index of 0.11 and partial prestressing ratio (PPR) of 0.28,
were prepared and tested under third-point loading. The beam
curvature, load-deflection characteristics, cracking pattern, strain in nonprestressed reinforcement at the critical section, ultimate strength and
slip of the strands were investigated in detail (4). The test results are
FRP Reinforcement
443
compared with analytical predictions using a nonlinear an~lysis
procedure [5] as well as with experimental results from pr~vtously
tested beams [6] with similar reinforcing index and PPR but havmg ste~l
strands as prestressed reinforcement. A summary of the study IS
reported here.
TEST PROGRAM
Test parameters
Two beams, designated as TC6a and TC9, were prepared. The
cross-section dimensions and the loading arrangement for both beams
are shown in Fig. 1. Details of the reinforcement are given in Table 1.
The global reinforcing index, ro, of the beam is defined as [7,8]:
ro = Apsfps/bdefc' + Asfy/bdefc'- As'fy'/bdefc'
(1)
where Aps and As are respectively the areas of prestressed and nonprestressed reinforcement, As' is the area of compression reinforcement,
fps is the stress in prestressed tensile reinforcement at ultimate, fy and fy'
are the yield stress of non-prestressed tensile reinforcement and
compression reinforcement respectively, fc' is the concrete cylinder
compressive stength, b is the width of the flange and de is the effective
depth to the centroid of the force in the tensile reinforcement at nominal
moment resistance. It is given by:
de= (Apsfpsdps + Asfyds)/ (Apsfps + Asfy)
(2)
in which dps and ds are the depths to the prestressed and nonprestressed reinforcement respectively.
The PPR, defined as the ratio of the nominal moment resistance
contributed by the prestressed reinforcement, Mnp. to the nominal
moment resistance contributed by the total flexural reinforcement, Mn, is
given by [6]:
PPR
=
MnpiMn
(3)
For beams with CFCC strands, the concrete strains are within the
linear elastic range when the non-metallic prestressing tendons rupture;
this can be attributed to the low elastic modulus of the tendons
(compared to steel) as well as to their quasi elastic response up to
failure and their relatively low failure strain. Hence, the value of fps
equals fpu and Eqs. (1) to (3) reduce to:
444
N aaman et al
ro = Apsfpu/bdefc' + Asfy/bdefc' - A 5 'fy'/bdefc'
( 4)
de= (Apsfpudps + Asfyds)/ (Apsfpu + A 5 fy)
(5)
PPR = Apsfpu(dp 5 -c/3)/[Apsfpu(dp 5 -c/3)+Asfy(d 5 -c/3)]
(6)
respectively, where c is the neutral axis depth at nominal moment
resistance, that is at the ultimate state in bending.
Note that, since the gloabal reinforcing index and the partial
prestressing ratio are strictly based on force and moment equilibrium of
the section at ultimate, they do not depend on the elastic moduli of the
reinforcing materials used.
The global reinforcing index and the PPR were kept constant at
approximately 0.11 and 0.28 respectively for the two beams prepared
(Table 1). Beam TC6a was originally intended to be prestressed with a
PPR of about 0.57 (that is, Beam TC6 in Table 1) but had a lower PPR
instead due to technical problems in prestressing operation as
described later.
Materials
Ready-mixed concrete with a design strength, fc'. of 6 ksi (42
MPa) and a maximum aggregate size of 3/8 in. (9 mm) was used in the
fabrication of the beams.
The prestressed reinforcement consisted of strands made from
CFCC (Carbon Fiber Composite Cable), each having a diameter of 5/16
in. (8.1 mm), effective section area of 0.047 in.2 (30.4 mm2) and
specified tensile strength, fpu. of 303 ksi (2092 MPa). Information on the
tensile and bond properties of the strands was supplied by the
manufacturer (9}. The stress-strain relation of the strand is nearly a
straight line up to failure with a tensile modulus of 20910 ksi (144000
MPa) and an elongation at break of 1.6%. The average bond strength
measured from pull-out tests using concrete prisms was reported to be
of the order of 1037 psi (7.37 MPa). This translates into a strand
development length to failure of the order of 14 inches (35 em). Thus in
designing the test beams, the application of maximum moment at about
3 ft (90 em) from the ends, guarantees no slippage of the strands up to
their failure. Each strand was provided with diecast sleeves, one at
each of its ends. A third sleeve, which was threaded on the outside, was
also diecast near to one end (Fig. 2). The sleeves so provided enable
the prestressing and anchoring of the strand using standard chucks (jaw
and barrel) and nuts commonly used for prestressing steel strands. Two
such strands were provided in Beam TC9 while four were provided in
Beam TC6a. They were placed in rows of two (Fig. 1).
FRP Reinforcement
445
The reinforcing bars were Grade 60 deformed bars of sizes No. 4
and No. 6. Shear reinforcement consisted of closed stirrups made with
No. 2 plain round bars and spaced according to design requirements in
the shear spans. They were also provided at a spacing of 3 in. (75 mm)
to 4 in. (1 00 mm) within the constant moment region except for the
middle 8 in. (200 mm) length. Marginal compressive reinforcement was
provided in the form of two No. 2 Grade 60 deformed bars to hold the
stirrups in place.
Pretensioning of CFCC Strands
Using a center-hole hydraulic jack, each strand was tensioned to
191 ksi (1317 MPa) or about 63% of its specified tensile strength. The
anchoring of the strand was facilitated by gradually turning a nut on the
threaded sleeve, upon which the jack was removed by either cutting the
strands beyond the jack or, by first removing the released chuck at the
end of the jack then cutting the strand. The latter method was carried
out by gently knocking on the barrel part of the chuck. The procedure
was satisfactory for the two cables in beam TC9.
However, for beam TC6a, knocking the chuck for its removal after
tensioning, resulted in a bond failure between the threaded sleeve and
the tendon at one of the end plates. The knocking had apparently
created a shock wave along the strand which reflected as a tensile
wave, leading to a bond failure between the diecast sleeve and the
tendon. The failure occurred outside the beam length. In view of this
incident, it was decided to leave the lower row of two cables in Beam
TC6a unstressed and pretensioning was thus limited to the upper two
cables only.
Transfer of Prestress
Transfer of prestress was achieved, twenty-one days after the
beams were cast, by means of a large wire cutter. The average
concrete compressive strength as obtained from four 4 x 8 in. (1 00 x 200
mm) cylinders was about 5900 psi (40.8 MPa). No loss in prestressing
force was noted in the strands during this period, as indicated by a load
cell mounted on one of the strands.
However, as soon as the first of the two strands in Beam TC9 was
cut, the resulting shock energy resulted in a combined bond and tensile
failure in the other strand along the threaded sleeve at the end plate
support, outside the concrete section. Thus, the second strand was
essentially released from the prestressing bed. This was confirmed by
the strain readings in the reinforcing bars, which did not change when
the second strand was cut later.
446
Naaman et al
Instrumentation and Test Procedure
Prior to casting of the beams, electrical resistance strain gages
were mounted on the non-prestressed reinforcement and a thin
aluminium strip was placed to act as a crack former at the midspan
section. The fabricated beams were instrumented with linear variable
differential transformers (LVDT) for the measurement of deflections,
curvatures and crack widths at the level of non-prestressed steel at the
midspan, and with dial gages for the monitoring of slip of strands at the
end sections. The test set-up is shown in Fig. 4. Similar instrumentation
and procedure were used and are described in more details in
Reference 6.
Each beam was loaded to about 60% of the theoretical ultimate
load for two cycles before being loaded monotonically to failure in the
third cycle. Beam TC9 was tested 31 days and Beam TC6a 36 days
after the day of casting. The concrete compressive strength (based on
six cylinders) averaged 5950 psi (40.9 MPa) at the time of testing. Test
data were recorded at load intervals of 2 kips (8.9 kN) before yielding of
the non-prestressed steel reinforcement and, thereafter, either at a load
interval of about 1 kip (4.45 kN) or at a deflection increment of 0.05 to
0.1 in. (1.25 to 2.5 mm) as appropriate, using a micro-computer
controlled data acquisition system.
TEST RESULTS AND DISCUSSION
Prestressing Operation
In the preparation of the test beams, it was found that a small
force imparted to the already tensioned CFCC strand could lead to bond
failure at the diecast sleeve, and/or rupture of the strand at the
anchorage. Thus, until further experience is gained, extreme care is
recommended to eliminate any additional force on the strand during the
pretensioning process and the casting of the concrete.
General Behavior of Beams
The test results are presented in graphical form for each beam in
Figs. 5 to 11.
Beam TC9
Beam TC9 was prestressed with two strands at a distance of 2. 75
in. (70 mm) from the bottom fibers. As the load on the beam was
increased, deflection was observed to increase linearly. At about 8 to
10 kips (36 to 44 kN), cracks were observed to form both in the constant
moment region and in the shear spans near the concentrated load
points, with spacings of 3 to 8 in. (75 to 200 mm).
FRP Reinforcement
447
Cracking led to a reduction of the beam stiffness. However, the
beam continued to exhibit linear elastic behavior after cracking. More
cracks developed in the constant moment region and in the shear spans
and propagated towards the compression zone of the beam. The
formation of cracks stabilized at a load of about 16 kips (71 kN). At 30
kips (133kN), that is 60% of the theoretical ultimate load, the crack
spacing averaged 4 ins. (1 00 mm) and crack widths were of magnitude
less than 0.01 in. (0.25 mm). The location of the cracks often coincided
with that of closed stirrups. When unloaded to zero, the residual
deflection of the beam was small and was about 0.05 in. (1.25 mm). The
beam showed similar behavior up to 30 kips (133 kN) for the
subsequent second and third cycles. No slip of the FRP strands was
observed at their ends.
During the third cycle, upon loading the beam beyond 30 kips
(133 kN), more cracks formed, but only in the shear spans. At about 40
kips (180 kN), the non-prestressed reinforcement started to yield, as
indicated by the strain gage readings. The stiffness of the beam
decreased rapidly. There was relatively large displacement with small
increase in the applied load.
A soft splitting sound was heard just before the load peaked at
49.6 kips (221 kN) and then dropped sharply to about 35 kips (156 kN).
This was accompanied by a bang. The dial gage at the end of one of
the strands indicated some very small change in readings, but this was
due to the impact caused by the release in applied load rather than the
slip of the strand. The FRP strands had apparently ruptured within the
span leading to longitudinal cracks in the constant moment region at
their level (Fig. 5).
The load was increased further, leading to large deflection of the
beam; it levelled off once again at about 43 kips (191 kN). Finally, at a
deflection of about 2.0 in. (50 mm), about equal to span over 50, the
concrete crushed explosively in compression at the top of the beam
within the constant moment reg[on near one of the concentrated loads.
The beam had finally failed in flexural compression.
The longitudinal cracks at the level of the strands were found to
extend for a distance of about 20 to 25 in. (500 to 625 mm) on each side
of the midspan section. The beam was later broken up and the rupture
of strands was confirmed at locations near the section of failure.
Beam TC6a
This beam had four CFCC strands placed in two rows; the upper
row at a distance of 5.25 in. (133 mm) and the lower row at a distance of
2.75 in. (70 mm) from the bottom fibers. The lower two strands were not
448
N a am an et al
prestressed. As a result, the global reinforcing index and the PPR of
beam TC6a were 0.1 02 and 0.286 respectively.
Similar to Beam TC9, the deflection of Beam TC6a increased
linearly as load was applied. At about 4 to 6 kips (18 to 27 kN), a crack
formed at the location of the crack former. At loads of 6 to 8 kips (27 to
36 kN), several cracks appeared within the constant moment region and
some within the shear spans near the applied concentrated loads. Their
spacing was about 4 to 8 in. (1 00 to 200 mm). The appearance of
cracks led to a reduction in the beam stiffness; however, the beam
continued to behave in a linear elastic manner. At about 10 kips (44
kN), more cracks appeared in between the primary cracks while the
primary cracks propagated towards the compression zone or towards
the locations of the applied loads.
The formation of cracks within the constant moment region
stabilized at about 10 kips (44 kN). Thereafter, new cracks only
appeared in the shear spans. At about 18 kips (80kN), that is
approximately 60% of the theoretical ultimate load, maximum crack
widths were less than 0.02 in. (0.50 mm) and crack spacing averaged 4
ins (1 00 mm). It was noted that cracks had formed mostly at the location
of closed stirrups. Maximum deflection of the beam was about 0.4 in.
(1 0 mm). When unloaded to zero, the cracks closed up and the residual
deflection was small and about 0.07 in (1.75 mm). Upon reloading, the
cracks opened up again and the load-deflection curve roughly followed
the unloading path of the previous cycle. At 18 kips (80 kN) again, the
deflection and crack width were essentially the same as in the first cycle.
Unloading in the second cycle and reloading for the third cycle did not
result in any deterioration in the beam behavior.
In the third cycle, as the load was increased beyond 18 kips (80
kN), existing cracks continued propagating upwards toward the
compression zone. New cracks were formed, but mostly within the
shear spans. At about 24 kips (1 07 kN), the non-prestressed steel
reinforcement began to yield and deflections and crack widths
increased at a more rapid rate. Fig. 6 shows the cracking pattern in the
constant moment region at about 33 kips (147 kN).
Finally, at about 35 kips (156 kN), two relatively loud bangs were
heard, one following almost immediately the other, and the load
dropped sharply to about 26 kips (116 kN). The upper row of strands
were deemed to have ruptured near the midspan section as confirmed
by the formation of longitudinal cracks at the level of these strands (Fig.
7), within the constant moment region near the location where the
strands had ruptured, in a manner similar to Beam TC9. Also, the
vertical crack at the midspan section was observed to widen excessively
in comparison to other cracks.
FRP Reinforcement
449
Thereafter, the deflection increased rapidly with a small increase
in applied load. At about 32 kips (142 kN), which approximated the
ultimate strength contributed by the two lower strands and the nonprestressed steel reinforcement, another loud bang was heard
(corresponding to the rupture of one CFCC strand) and longitudinal
cracks formed at the level of the lower strands. The load dropped to 25
kips (111 kN). The beam was loaded further and, at about 26 kips (116
kN), another bang indicated the rupture of the last CFCC strand. The
load dropped to 19 kips (85 kN) and then stabilized at about 21.5 kips
(96 kN) up to large deflections. For safety reasons, loading was
terminated when the midspan deflection had reached 2.55 in. (65 mm).
It is however expected that the concrete would have finally crushed at
the top of the beam.
The longitudinal cracks at the level of the strands extended for a
distance of about 22 to 28 in. (550 to 700 mm) on each side of the
midspan section. In a last phase, the region around the midspan
section of the beam was broken with a jack hammer and rupture of all
four strands was confirmed.
Comparison with Analytical Predictions
The flexural response of the beams up to failure can be divided
into two main phases: (a) behavior up to the ultimate load (or moment);
and (b) post-peak behavior. The behavior up to the ultimate load is
characterized by three milestone events, namely: the cracking load (or
moment), the yield load (or moment) at which the non-prestressed steel
reinforcement starts to yield, and the ultimate or maximum load (or
moment). The post-peak behavior is characterized by several levels of
load (or moment) carrying capacity due to remaining tensile
reinforcement.
The method used for the prediction of the beams's behavior up to
the ultimate load, assumes a characteristic yield strength of 65.7 ksi
(453 MPa) for the deformed bars and evaluates the moment of inertia
before cracking using the transformed area of the section. Assuming
strain compatibility, the moment-curvature curves up to the ultimate
moment were derived for the midspan section and plotted in Fig. 8
together with the experimental curves. For comparison purposes, the
reference state was taken as that corresponding to the simply-supported
beam under the effect of final prestress and dead load only. The
predictions are found to compare well with the experimental results
including the yield moment and the ultimate moment.
The load-deflection curves of the beams were derived analytically
by calculating the load for a given moment from statics, and the
deflection from the corresponding curvature. The deflection was
evaluated by doubly integrating the calculated curvatures at cracked
sections over the length of the beam [5]. This method also accounts for
450
N aaman et al
the additional deformation due to shear. The analytical predictions are
compared to the experimental results in Fig. 9.
The prediction method for deflections leads to reasonably good
ageement with the observed deflections, provided the residual
deflection due to the first cycle of loading up to 60% of the ultimate load
is accounted for.
For the post-peak behavior, the residual load-carrying capacities
due to remaining reinforcement were calculated by strain compatibility
[5] and using the ACI stress block (1 0] where approriate.
Tables 2 and 3 compare the calculated characteristic loads, and
corresponding deflections and curvatures, with corresponding
experimental results for Beam TC9 and Beam TC6a respectively. Good
agreement in the load-carrying capacities and the related deflections
and curvatures is generally obtained. Larger discrepancies between
the predicted and experimental values are however observed just
before and after cracking and at yielding of the non-pretressed steel
reinforcement. These are expected as readings could not be taken at
the exact onset of cracking or yielding. The readings were obtained at
the midspan section where a crack former was placed; this may have
led to larger curvatures. The presence of the crack former causes a
stress concentration and may have been responsible for observed loads
at cracking substantially smaller than predicted. Also, the small values
of deflections and curvatures at these loads would mean a larger
percentage error for the same accuracy of measurement. The
deflections obtained experimentally were not adjusted for possible end
support settlement.
The variation of stresses in the prestressed reinforcement with the
applied load are shown in Fig. 10 for the predicted and experimental
values respectively.
The initial stresses in the prestressed
reinforcement were calculated from the effective prestress and verified
from strain readings in the non-prestressed steel. For the purpose of
comparison, both the experimental and the predicted values are plotted
with the same initial stresses. The experimental plot for the stress in the
prestressed reinforcement were obtained indirectly from the measured
curvatures in the concrete and stresses in the non-prestressed
reinforcement. In general, good agreement between the experimental
and predicted results is obtained. Analytical results also indicated that
rupture of the CFCC strands at maximum load occurred before strain
hardening of the reinforcing bars.
Effect of Tendon Type
To evaluate the effects of non-metallic prestressed reinforcement
on the flexural performance of concrete beams, the test results of Beam
TC9 are compared to another previously tested beam [6] using
FRP Reinforcement
451
prestressing steel strands and designated as TS9. Beam TS9 had a
global reinforcing index of 0.095 and a PPR of 0.293.
It is seen from Fig. 11 (a) and (b) that the load-deflection or
moment-curvature relationships of the beams were similar up to the
ultimate load of beam TC9, regardless of the type of prestressed
reinforcement. However, a marked difference in the post-peak behavior
is noted. Beam TS9, being prestressed with steel strands, exhibited
substantial ductility at ultimate whereas Beam TC9, which was
prestressed with CFCC strands, showed a sudden drop in load-carrying
capacity when the ultimate load was reached, due to the rupture of the
strands. The residual load-carrying capacity was slightly larger than that
provided by the remaining reinforcement. Beam TC9, however,
sustained as large deflections as in the case of Beam TS9.
Fig. 11 (c) and (d) respectively show the variations in steel strain
in the non-prestressed reinforcement and the maximum crack width with
load for beams TC9 and TS9. The trends are very similar up to the
ultimate load for both beams. Crack width generally increased rapidly
once the reinforcing bars yielded.
CONCLUSIONS
From the flexural tests conducted on the two T-beams partially
prestressed with CFCC strands, the following conclusions can be
drawn:
1.
Extreme care should be exerted in prestressing fiber reinforced
plastic reinforcements. This is because they exhibit essentially a linear
elastic brittle (non-yielding) response in tension, they have a relatively
low elastic modulus when compared to steel, and thus they release
enormous strain energies at failure. Accidental failure during stressing
and anchoring should be of particular concern. In the preparation of the
test beams, it was found that a small force imparted to the already
tensioned CFCC strand could lead to bond failure at the diecast
sleeves. Thus, particular attention should be paid to eliminate any
additional force on the strand during the pretensioning process and
prior to the transfer of prestress.
2.
The load-deflection response up to the maximum load of beams
prestressed with FRP tendons is essentially trilinear with decreasing
slope; the initial linear portion extends from zero load to first cracking;
the second linear portion extends from first cracking to the yielding of the
reinforcing bars; the third linear portion extends up to the maximum load
at which failure of the FRP tendons occurs. The presence of nonprestressed conventional ductile reinforcing bars dampen the failure of
such beams and provides a minimum residual ductiliy and strength that
may be needed in numerous applications, particularly in seismic zones.
452
N a am an et al
3.
The post-peak load-deflection response of partially prestressed
beams with FRP tendons is characterized by incremental step-like
decreases corresponding each to the failure of one FRP tendon.
Residual strength and ductility is provided by the non-prestressed
reinforcing bars. The failure of each FRP tendons releases sufficient
fracture energy to lead to longitudinal cracks in the concrete, which
extend on each side of the tendon failure section.
4.
Experimental observations and analytical results suggest that
failure of the FRP tendons at maximum load, occurs after yielding of the
reinforcing bars, but before their strain hardening.
5.
The load-deflection response of beams prestressed or partially
prestressed with FRP reinforcement can be predicted with reasonable
accuracy using conventional methods of equilibrium, strain compatibility
and material stress-strain relationships.
Comparison between the beams tested in this investigation, in
which FRP strands were used, and similar beams tested in previous
studies (6) using steel strands, leads to the following conclusions:
6.
For the same global reinforcing index, the cracking load is lower
and the crack width is larger while the beam stiffness after cracking is
smaller for a beam prestressed with CFCC strands than for an ordinary
reinforced concrete beam.
7.
Depending on the partial prestressing ratio, or equivalently, the
amount of reinforcing bars present in the section, the behavior of beams
partially prestressed with CFCC strands can be very similar to that of
beams using steel strands. Cracking and crack widths are also of the
same order; crack widths generally increase very rapidly once the
reinforcing bars yielded.
ACKNOWLEDGEMENT
The CFCC strands used in the investigation were provided by
Tokyo Rope Manufacturing Company, Japan, through C. ITOH and Co.,
LTO, Japan. Their cooperation as well as the particular attention
devoted by Mr. T. lkezawa form C. ITOH and Mr. T. Ohtsuki from Tokyo
Rope are gratefully acknowledged. The assistance of R. Spence and H.
Hammoud in the execution of the test program is also appreciated.
FRP Reinforcement
453
REFERENCES
1.
Dolan, C. W., "Developments in Non-Metallic Prestressing
Tendons", PCI Journal, Vol. 35, No.5, Sep/Oct 1990, pp. 80-88.
2.
Burgoyne, C. J. (Ed.), "Symposium on Engineering Applications
of Parafil Ropes", Department of Civil Engineering, Imperial College of
Science and Technology, London, England, 1988, 91 pp.
3.
Zoch, P., Kimura, H., Iwasaki, T. and Heym, M., "Carbon Fiber
Composite Cables - A New Class of Prestressing Members", preprint,
Transportation Research Board Annual Meeting, Washington D. C., Jan
1991' 20 pp.
4.
Naaman, A.E., Tan, K.H., Jeong, S.M., and Alkhairi, F.M.,
"Partially Prestressed Beams with Carbon Fiber Composite Strands",
Preliminary Progress Report, No. UMCE-92-8, Department of Civil and
Environmental Engineering, University of Michigan, Ann Arbor, February
1992, 51 pages.
5.
Alkhairi, F. M., "Flexural Behavior of Concrete Beams prestressed
with Unbonded Internal and External Tendons", Dissertation submitted
to the University of Michigan in partial fullfilment for the degree of Doctor
of Philosophy, Dec1 991, 415 pp.
6.
Naaman, A. E. and Faunas, M., "Partially Prestressed Beams
under Random-Amplitude Fatigue Loading", Journal of Structural
Engineering, ASCE, Vol. 117, No. 12, Dec 1991, pp. 3742-3761.
7.
Naaman, A. E., "Partially Prestressed Concrete: Review and
Recommendations", PCI Journal, Vol. 30, No. 5, Sep/Oct 1985, pp. 5481.
8.
Naaman, A.E., "Prestressed Concrete Analysis and
Design - Fundamentals", McGraw Hill Book Co., New York, 1982,
670 pages.
9.
Tokyo Rope Manufacturing Co., LTD., "CFCC Technical Data,"
October 1989, Japan, 43 pages.
10.
ACI Committee 318, "Building Code Requirements for
Reinforced Concrete and Commentary (ACI 318-89/ACI 318R89)", American Concrete Institute, Detroit, 1989, 353 pp.
454
N aaman et al
TABLE 1 - DETAILS OF TEST BEAMS
Prestressed
Reinforcement
Non-prestressed
Reinforcement
Aps
(in2)
dps
(in)
As
(in2)
ds
(in)
(1)
(2)
(3)
(4)
(5)
TC9
0.094
(2-CFCC1 x7)
9.25
0.88
(2#6)
TC6'
0.094
(2-CFCC1 x7)
6.75
0.40
(2#4)
0.094
(2-CFCC 1x7)
9.25
0.094
(2-CFCC 1x7)
6.75
Beam
Designation
TC6a
0.094
(2-CFCC1x7)
0.40
( 2 # 4)
Not tested.
de
PPR
(J)
(in)
(6)
(7)
(8)
10.88
10.39
0.266
0.116
11.00
9.04
0.573
0.122
9.25
8.98
0.285
0.102
11.00
Note: See Fig. 1 for notation; 1 in. = 25.4 mm.
455
FRP Reinforcement
TABLE 2 - CHARACIERISTIC LOADS, DEFLECTIONS
AND CURVATURES FOR BEAM TC9
Calc.
(3)/(4)
(3)
(4)
(5)
8.1
10.5
11.6
Beam
Stage
Expt.
(1)
(2)
Crackingb
0.77
0.70
Calc.
(6)/(7)
Expt.
(6)
(7)
(8)
(9)
J4b
o.os•
1.12b
1.87b
1.3oa
1.64 a
0.050b
0.03b
0.063•
0.056b
(x 10·6 in.·1)
Calc.
Expt.
0.082•
Cracking•
TC9
•
8 (in.)
P (kips)
(10)
51'
(9)/(10)
(11)
40b
26 b
51•
45•
0.85b
1.31b
1.00•
1.13•
40.0
46.0
42.2
0.87
0.95
0.565
0.451
0.47
1.25
1.20
343
363
293
0.94
1.17
Ultimate
49.6
50.2
45.9
0.99
1.08
1.096
1.005
0.91
1.09
1.20
821
809
783
1.01
1.05
Residual
Capacity
42.9
37.9
1.13
2.065
2.565
0.81
1786
2101
0.85
Yield
a after cracking:
b before cracking; cat crushing of concrete; • not measured.
Values given by Method 2 are in italics.
Note: 1 in. = 25.4 mm.: 1 kip = 4.448 kN.
TABLE 3 - CHARACIERISTIC LOADS, DEFLECTIONS
AND CURVATURES FOR BEAM TC6a
P (kips)
Beam
(1)
Stage
Expt.
(2)
(3)
(4)
6.0
8.0
8.5
Crackingb
Calc.
(3)/(4)
(5)
0.75
0.71
Cracking•
Yield
TC6a
+ (x
8 (in.)
Expt.
(6)
0.062b
0.110•
Calc.
(7)
10·6 in.·1)
Expt.
(8)
(9)
(10)
(11)
37b
3Qb
26 b
5ga
59 a
1.23b
1.42 b
1.25•
1.25 a
0.85
1.05
0.037b
0.03 b
0.071•
1.6Sb
2.07 b
1.5sa
74•
Calc.
(9)/(10)
(6)/(7)
24.0
26.7
23.9
0.90
1.00
0.606
0.532
0.43
1.14
1.41
357
422
339
Ultimate
35.0
34.5
33.2
1.01
1.05
1.683
1.295
1.35
1.30
1.25
1260
1071
1103
1.18
1.14
At rupture of
top 2 cables
31.5
31.5
1.00
2.125
2.211
0.96
1819
1785
1.02
At rupture of
a lower cable
26.0
24.8
1.05
2.198
2.174
1.01
1940
1756
1.11
Residual
Capacity
21.5
17.6
1.22
5.8Q4C
a after cracking; b before cracking; cat crushing of concrete; • not measured.
Values given by Method 2 are in italics.
Note: 1 in. = 25.4 mm.; 1 kip • 4.448 kN.
4678C
456
N aaman et al
p
3.0 It
3.0 It
3.0 It
12 in
I
4-1/2 in
1
Typical cross section
NON-PREST.
2-CFCC 1 x7 As 2_# 4
As 2-#6
. _ __
_J
~3/4 in
1-1/8 in
Beam TC9
1 in
Beam TC6a
Fig. 1-Cross section dimensions and loading arrangement
FRP Reinforcement
Fig. 2-CFCC with diecast sleeves
Fig. 3-Set-up for pretensioning of tendons
(showing load cell and strain gage reading conditioner)
457
458
N aaman et al
Fig. 4-0verall view of test set-up and instrumentation
Fig. 5-Close-up view of cracking pattern of beam TC9
FRP Reinforcement
1\eam
n
b
Fig. 6-Cracking in constant moment region of
beam TC6a near ultimate load
Fig. 7-Close-up view of cracking pattern of beam TC6a
459
460
N aaman et al
1000
800
c
I
(f)
a.
~
c
Q)
600
400
E
0
::::2:
Beam TC9
W=0.116
PPR=0.266
200
0
0
400
800
1200
1600
2000
Curvature ( x 10- 6 in- 1 )
(a) Beam TC9
c
600
I
(f)
a.
~
......
c
400
Q)
E
0
_______
j
-test
----··----
analysos
I
...
::::2:
Beam TC6a
W=0.102
PPR=0.285
0 U-----------------------~----~--~
0
500
1000
1500
2000
2500
3000
Curvature ( x 10- 6 in- 1 )
(b) Beam TC6a
Fig. 8-Comparison of moment-curvature curves with analytical predictions
FRP Reinforcement
461
60.-----~------~----~------~----·
.................,...
. ....... ; .... .
-a.
( J)
~
Deflection (in)
(a) Beam TC9
40
-
r-----------~----~------------~----,
\
30
( J)
a.
........
. . . _/_,....•.:.:.>···~;,·,..········:·······+ ······""""'l
~
"0
co
20
0
.....J
10
0 ~----~----~----~----~--------~
0
0.5
1. 5
2
2.5
3
Deflection (in)
(b) Beam TC6a
Fig. 9-Comparison of load-deflection curves with analytical predictions
462
Naaman et al
350
Bea;, TC9
300 ·········
(/)
w~0.116
PPR~0.266
250
~
(/)
.2-
200
"0
c
150
~
100
co
50
0
10
0
20
30
40
50
60
Load (kips)
(a) Beam TC9
350
300
(/)
~
(/)
.2"0
c
250
200
150
co
~
100
50
0
0
10
20
30
40
Load (kips)
(b) Beam TC6a
Fig. tO-Comparison of stress in reinforcement versus
applied load curves with analytical predictions
FRP Reinforcement
1000
c
800
'
(f)
CL
~
600
c
Q)
E
400
0
~
200
(w, PPR)
0
0
400
800
1200
1600
2000
Curvature ( x 10- 6 in- 1 )
(a) Moment-curvature curves
60
TS9 ( 0.095, 0.293 )
50
(f)
40
CL
~
"'0
30
co
0
_J
20
..
10
·•·· ····-.
. ........... -~----
(w, PPR)
0
0
0.5
1.5
Deflection
2
(in)
(b) Load-deflection curves
Fig. 11-Effect of tendon type
2.5
463
464
N aaman et al
6000
5000
4000
0
X
3000
(/)
w
2000
:
:
1000
( w, PPR)
0 0
20
10
30
50
40
Load (kips)
(c) Strain in reinforcement versus applied load
0.15 ~----~--~----~----~-----,----,
( w, PPR)
c
0.125
j TS9
0.1
( 0.095, 0.293 )
.kj
H'HH
..c
-D
:········/···
0.075
0.05
...·
:
0.025 I·················
0
10
20
30
40
50
Load (kips)
(d) Crack width versus applied load
Fig. 11 cont.-Effect of tendon type
60
SP 138-28
Properties of Hollow Concrete
Masonry Reinforced with
Fiberglass Composite
by A. Hamid, J. Larralde, and A. Salama
Synopsis:
This paper presents the results of a test program carried out at Drexel University
to study the properties of hollow concrete masonry reinforced with fiber glass
composite as external reinforcement. The research program was planned to
demonstrate the feasibility of using such a technique for hollow masonry walls.
The program included tests of 1/3 scale masonry assemblages to determine the
compressive strength, flexural strength under in- and out-of-plane loads, and inplane splitting tensile strength. Widely available commercial type of fiberglass
mat and matrix resin were used in this investigation. The results indicate that the
proposed retrofitting technique is very effective and has great potential in
improving the strength and deformation properties of hollow concrete masonry.
Keywords: Composite materials; compressive strength; concretes; fiberglass;
flexural strength; masonry: reinforcing materials; tensile strength
465
466
Hamid, Larralde, and Salama
Ahmad A. Hamid is a Professor of Civil Engineering at Drexel University,
Philadelphia. He is a member of ACI 530/ASCES!fMS 402 Masonry Joint
Standards Committee. His research interests include properties of construction
materials, seismic behavior and design of concrete and masonry structures, and
assessment and retrofitting of existing buildings. Jesus Larralde is Assistant
Professor of Civil Engineering at Drexel University. His research interests are in
the area of civil engineering materials and applications of fiber composites in civil
engineering structures. Amr Salama is Professor of Structural Engineering at
Helwan University, Cairo, Egipt. His research interests include mechanics and
properties of materials and design of reinforced concrete structures.
INTRODUCTION
Many existing unreinforced masonry buildings in seismic areas do not meet the
current code requirements for lateral load resistance. Retrofitting of such
structures is necessary to ensure an adequate factor of safety. Masonry is a
composite material made up of units with mortar as a bonding agent. The mortar
bed and head joints represent planes of weakness where failure is initiated by
debonding at the unit-mortar interfaces ( 1-3 ). Thus, hollow masonry is
inherently weak in tension and shear ( 2, 3 ) and has limited resistance to lateral
loads, particularly under low levels of gravity loads which negate the tensile
stresses normal to the bed joints.
The feasibility of a new method for retrofitting concrete block masonry buildings
is discussed in this paper. The method consists of using Fiberglass Composite
Laminae (FCL) adhered to the outside of the masonry surface (Fig. 1). This
study, conducted at Drexel University, was planned to demonstrate the feasibility
of using such a technique for hollow masonry. The program included tests of
1/3-scale masonry assemblages to determine the compressive strength, flexural
strength under in-plane and out-of-plane loads, and splitting tensile strength.
FCL reinforcement fabricated using widely available commercial materials was
used in this investigation.
SPECIMEN FABRICATION AND TESTING PROCEDURES
A total of 27 specimens were fabricated to perform six different tests . The test
specimens are shown in Figure 2. The specimens consisted of 1/3-scale concrete
masonry assemblages fabricated under laboratory controlled conditions. After
fabrication the specimens were allowed to cure for 30 days before testing.
Control specimens without reinforcement were tested as indicated in Table 1.
Other specimens were reinforced with the FCL and were allowed to cure at room
conditions of temperature and humidity for 24 hours prior to testing.
The masonry units used in this program are scale down version of ASTM C140,
Grade N, 6 in. by 8 in. by 16 in. hollow units commonly used in masonry
construction. The unit has a percent solid of 54 % and a compressive strength of
2,800 psi.
FRP Reinforcement
467
These 1/3 scale units, which are manufactured using an in-house block making
machine, have been successfully utilized at Drexel University for the past 15 years
to study the behavior and strength characteristics of concrete masonry elements
(4-6 ).
The FCL reinforcement consisted ofE-glass chopped strand reinforcing mat ( 7)
impregnated with polyester-styrene resin ( 8 ) . The reinforcement was applied in
a hand lay-up process ( see Fig. 1 ). First, a coat of the polyester-styrene resin
was applied on the surface of the assemblages, followed by the application of the
reinforcing mat. A second coat of the resin was then applied on top of the mat
before the first coat started to set. A roller was used to insure complete wetting of
the mat and to intermix the resin from both coats. The resulting thickness of the
FCL was approximately 0.11 in. Thus the percentage of cross sectional area of
FCL with respect to the envelope area of the assemblage was approximately 11
percent. In some of the specimens a layer of commercial, epoxy-based bonding
compound was applied on the surface of the assemblages and was allowed to set
prior to the application of the FCL reinforcement. The purpose of using this
bounding compound is to verify if adequate bond will be developed between the
surface of the masonry assemblage and the FCL and to prevent the potential
uneven surface texture of the assemblages to maintain undisrupted tensile lines in
the FCL. In all cases, the FCL reinforcement was applied on both faces of the
assemblages. The tensile strength and modulus of elasticity of the composite FCL
were determined by direct tensile test of coupon specimens of the reiforcement.
The coupon specimens were fabricated following the same hand lay-up process.
The average tensile strength and modulus of elasticity were 3,276 psi and
196,500 psi, respectively.
TEST RESULTS
A summary of the test results is presented in Table 1. Also, Figure 3 shows the
results graphically. As can be seen, in all cases the FCL reinforcement increased
considerably the load capacity of the concrete masonry assemblages. In all but the
in-plane bending cases, the assemblages failed first before the FCL reinforcement.
It was also observed that bond between the FCL and the assemblages was
maintained without any delamination or debonding in all test cases. It can also be
noticed from the results that there were no appreciable differences between the
FCL reinforced specimens with and without Sikadur. Thus, the FCL alone was
able to develop enough bond strength with the assemblage to prevent bond
failure. The increase in load carrying capacity was more remarkable in the case of
bending, both in-plane and out-of-plane. For the out-of-plane bending, no load
was registered prior to failure in the control specimens in part due to the low
resolution of the testing machine ( 10 lbs), but mostly because of the very low
load capacity. In contrast, with the FCL reinforcement, the load at failure was of
the order of 600 lbs for both cases: specimens with FCL alone and those with
FCL and Sikadur. In the out-of-plane bending tests, while the control specimens
failed as expected by joint debonding, the FCL reinforced specimens failed in
shear away from the joints (Fig. 4) with a shear crack developed parallel to the
length of the specimen and at approximately half the thickness.
468
Hamid, Larralde, and Salama
In the in-plane bending test, the FCL was remarkably effective in preventing
debonding and joint failure along the head joint. As shown in Fig. 5 failure
occurred in the FCL at mid-span away from the head joints.
The specimens tested under axial compression failed by splitting along a plane
parallel to the face of the assemblage and approximately half way the thickness
(see Fig. 6). For the FCL reinforced specimens, the failure load was
approximately three times greater than that in the control specimens without
reinforcement. The increase in the failure load of the specimens with the FCL
reinforcement is probably due to the confining effect of the FCL. The FCL alone
may have contributed to the capacity due to the increase in cross sectional area but
in a small amount, probably not more than 10 percent.
In the in-plane tensile splitting, again, the FCL reinforced assemblages had much
greater load capacity than the control assemblages without reinforcement. In all
cases failure occurred as localized bearing failure at the points of load application.
Therefore, the tensile capacity of the specimens was not achieved. The maximum
recorded loads of the FCL reinforced assemblages were in the order of two to
eight times greater than those of the control assemblages. In none of the cases
debonding was observed but failure occurred by crushing of the portion of the
assemblage at the point of load application.
DISCUSSION
The test results presented herein were obtained using 1/3 scale specimens. The
results clearly demonstrate that using FCL can significantly improve the
compressive, tensile and flexural strengths of hollow masonry. Although only a
few specimens for each case were tested, the load capacity with the reinforcement
is so much greater than the control values that rigurous statistical testing is not
needed to confirm this conclusion. Nonetheless some observations should be
made.
The tested assemblages were reinforced on the entire facial area and on both sides.
The relative amount of reinforcement was therefore high. It is envisioned that in
real applications it would be difficult to reinforce the entire area of the walls
because access may be limited to only portions of the walls. In addition, it could
be prohibitively expensive to reinforce the entire area on both sides. More
realistically, the reinforcement may be applied on portions of the walls only,
following a grid pattern similar to the conventional internal reinforcement with
steel bars. Peculiarly, since the FCL reinforcement is applied externally, it can be
arranged along the diagonals of the walls thus increasing their shear capacity
under horizontal, in-plane loads.
A significant finding from the tests is that the FCL reinforcement was 11-ble to
remain adhered to the assemblages. Neither debonding nor delamination occurred.
It should be noted though that the application of the FCL during the hand lay-up
process was done with the specimens in a horizontal position. This procedure
prevented the runoff of the resin that otherwise would be induced by gravity.
FRP Reinforcement
469
Also, it insured a fairly good impregnation of the assemblages' surface. In real
applications with the masonry walls in a vertical position, the resin may run
down. This should not be of concern nevertheless because practical procedures of
application can be easily devised through a trial-error process. For instance, the
resin may be applied at the same time that the fiberglass mat. That is, the
fiberglass mat can be impregnated with the resin immediately prior to the had layup process.
In a group of specimens, the surface was coated with a commercial, epoxy-based
bonding compound to insure a better bonding between FCL and even specimen
surface. The results showed no difference between these specimens and those
that were not coated prior to the application of the FCL. The ability to maintain
undisturbed tensile lines in the FCL should be of concern. Although the test
results presented herein demonstrate no difference between prepared surface and
unprepared surface, it should be cautioned that the specimens were 1/3 scale thus
reducing the possibility of uneven surface. In actual applications, not only the
area to be covered with the FCL will be larger but also actual field conditions may
not result in a very even, smooth surface on top of which the FCL can be applied
without previous preparation.
Another important aspect that should be addressed as part of the applicability of
FCL reinforcement is the long-term performance. The durability of the FCL
reinforcement exposed to different environments should be investigated to obtain
a successful retrofitting procedure both in the short-term as well as in the longterm. The ability of reinforced composite materials to resist corrosion and
chemical attack has been demonstrated in many applications in the aeronautical
and automotive industries. Thus, it is very likely that adequate long-term
performance and durability can also be obtained in hollow masonry retrofitting.
CONCLUSION
The limited data presented in this paper clearly demonstrates thaL Lhe proposed
strengthening technique using FCL is very effective and has great potential for
retrofitting of concrete masonry structures.
ACKNOWLEDGMENTS
The authors want to express their appreciation to Dr. Frank Ko, Director of the
Fibrous Materials Research Center, Drexel University, and to Mr. Ahmed Morsi,
IMCO Reinforced Plastics, Inc., for the donation of the materials used in this
work.
REFERENCES
I. Hamid, A.A., "Behavior Characteristics of Concrete Masonry," Ph.D. Thesis,
McMaster University, Hamilton, 1978.
2. Hamid, A.A., Drysdale, R.G., and Heidebrecht, A.C., "Shear Strength of
Concrete Masonry Joints," Proceedings, ASCE, V. 105, Sn, July 1979, pp.
1227-1240.
470
Hamid, Larralde, and Salama
3. Drysdale, R.G., Hamid, A.A., and Heidebrecht, A.C., "Tensile Strength of
Concrete Masonry,"Proceedings, ASCE, V. 105, ST7, July 1979, pp. 12611276
4. Becica, I. and Harris, H. G., "Evaluation of Techniques in the Direct Modeling
of Concrete Masonry Structures, " Structural Models Laboratory Report No.
M77-1 Department of Civil Engineering, Drexel University, Philadelphia, June
1976, 96 pp.
5. Harris, H.G. and Becica, I.J., "Direct Small-Scale Modeling of Concrete
Masonry," Advances in Civil Engineering Through Engineering Mechanics,
ASCE, New York, 1977, pp 101-104
6. Abboud, B., Hamid, A.A, and Harris, H.G., "Small-Scale Modeling of
Masonry Structures, "ACI Structural Journal, Proceedings, V. 87, No. 2,
March-April190, pp. 145-155.
7. PPG Industries, Inc., " Fiber Glass Products, " Pittsburgh, PA, 1989.
8. Ashland Chemical, Inc., "Material Data Sheet," Columbus, Oh, 1989.
TABLE 1 - TEST RESULTS
lEST
lYPEOF
SPECIMEN
1. Axial compression
2. Flexure
(ln~plane}
3. Flexure (out-of-plane)
lbs
control
FCL with Sikadur
FCL without S1kadur
3180 4700 5000
9170
11000
9150 13700
control
FCL With S1kadur
F CL Without Sikadur
180 90
1450 1150
1260 1060
control
4.b. Load parallel
to bedjotnls
4.c. Diagonal load
Averaoe
4293
10085
11425
135
1300
1160
shear+ s I1Uina
s littinQ
split1inQ
tens11e debondmg of head !Oint
flexure
flexure
FCL without Sikadur
ne II able
600 525
666 594
control
FCL wtth S1kadur
FCL Without Sikadur
2050
3850
5300
2050
3850
5300
beannQ !allure at supporl
beanng failure at support
440
3600
not tested
440
3600
tensile debonding along bed ·oints
bearing failure at support
FCL without Sikadur
control
FCL wtth S1kadur
F Cl wtlhoul S1kadur
1150, 2050
6950
not tested
1600
6950
sphtllng
bearing fa1lure at support
FCL With Sikadur
4. In-plane Tensile Splitting
4.8. Load perpendtcular
to bed joints
FAILillE M:lDE
ULTIMATE LOAD
IndiVIdual
control
FCL with Sikadur
ne 11 able
563
shear
630
shear
splillino
FRP Reinforcement
Fig. 1-Hand lay-up application of FCL
reinforcement to masonry assemblages
471
472
Hamid, Larralde, and Salama
1. Axial Compression
I.
1~
.
~t.aa•
2. In-plane Flexure
3. Out-of-plane Flexure
4. In-plane Splitting
10.~
10.5"
t
a. Load Perpendicular
to Bed Joints
b. Load Parallel
to Bad Joints
LJ
FCL
Noltoacale
Fig. 2-Test specimens
c. Diagonal Load
FRP Reinforcement
473
[I] Axial Compression
4000
II
In-Plane
0
Out-of-Plane
3000
...
~!:
2000
"'
1000
Control
Without Sibdur
With Sikadur
Type of Specimen
Fig. 3-Graphic comparison of test results
Fig. 4-Photograph of specimen at failure under out-of-plane bending
474
Hamid, Larralde, and Salama
Fig. 5-Photograph of specimen after failure under in-plane bending
..
Fig. 6-Photograph of specimen after failure under axial compression
SP 138-29
Prestressed Concrete Beams
Using Non-Metallic
Tendons and External
Shear Reinforcement
by C.W. Dolan, W. Rider,
M.J. Chajes, and M. DeAscanis
Synopsis:
Four prestressed concrete T-beams were constructed with
aramid-based prestressing tendons made of Fibra rods and parallel fiber Kevlar
rods.
Three beams included an externally applied Kevlar fabric shear
reinforcement. The beams were tested and the performance of beams with
external Kevlar fabric shear reinforcement were compared to a control beam
without external shear reinforcement. The tests indicate that Fibra rods serve
as an effective prestressing tendon with excellent bond characteristics and shear
response predicted by current ACI equations. The tests also indicate that the
externally applied woven fabric can provide the shear resistance. External
shear reinforcement using nonmetallic fabric offers significant opportunities for
strengthening existing structures as well as fabrication of nonmetallic
structures.
Keywords: Beams (supports); bonding; concretes; fabric; fibers; prestressed
concrete; prestressing steels; reinforced concrete; shear properties;
strengthening
475
476
Dolan et al
Dr. Charles W. Dolan, FACI, is an Associate Professor in the Department of
Civil and Architectural Engineering, University of Wyoming, Laramie, WY.
Dr. Dolan has over 20 years of consulting experience and is chairman of
ASCE/ ACI joint committee 423, Prestressed Concrete, and the PCI High
Strength Concrete Committee. He is active in the design and analysis of new
materials for prestressing applications.
Mr. Wade Rider is a MS candidate in the Department of Civil Engineering at
the University of Delaware, Newark, DE and was instrumental in executing
this research.
Michael J. Chajes, MACI is an Assistant Professor in the Department of Civil
Engineering, University of Delaware, Newark, DE. He received his Ph.D.
from the University of California, Davis in 1990. His research interests are in
structural analysis and design including the use of advanced materials in civil
engineering structures. Prof. Chajes is a member of ACI Committee 341,
Earthquake Resistant Concrete Bridges
Mr. Michael DeAscanis is an honors student at the University and participated
in this project as an undergraduate research assistant.
INTRODUCTION
Numerous tests and demonstrations are being conducted to evaluate the
performance of nonmetallic prestressing tendons for prestressed concrete
applications. Test reports on fiberglass tendons (1,2), aramid tendons (3,4)
and carbon tendons (5) have all demonstrated that the nonmetallic materials
will function as short-term prestressing materials. These tests have raised
several issues which require additional study. The three most immediate issues
are: 1) creating sufficient bond between the tendon and the concrete to allow
the full tensile capacity of the tendon to be developed; 2) the use of metallic
materials in the beam; and 3) the long-term durability of the nonmetallic
materials. This research addresses the first two issues.
Nonmetallic prestressing tendons have been suggested as alternatives for
prestressing steel because of their corrosion resistance and their nonmagnetic,
nonelectrical-conducting properties. Nonetheless, many nonmetallic tendon
tests have been conducted using steel shear reinforcement and some
longitudinal nonprestressed steel reinforcement to support the stirrups. While
these steps may be necessary to test the tendon performance, the presence of
the steel obviates some of the advantages of using nonmetallic tendons.
FRP Reinforcement
477
Nonmetallic prestressing strands are not generally commercially
available. Tokyo Rope manufactures a carbon-based tendon called CFCC (5)
and Mitsui manufactures an aramid-based braided reinforcing rod under the
name of Fibra I (6). The Fibra rod is particularly interesting as a prestressing
material because it is made from a resin-impregnated, braided Kevlar2 fiber.
The Kevlar fiber has a high tensile strength and excellent fatigue resistance (7).
The braiding increases the manufacturing complexity, but provides a final
product with potentially excellent bond characteristics.
Strengthening existing structures using FRP plates and external
reinforcement is a subject of worldwide interest (8). When the flexural
capacity is increased, the shear capacity of the structure may also have to be
increased. However, research into improved shear performance is extremely
limited.
This study was undertaken to investigate the following objectives:
(1) assess the ability of Fibra to perform as a prestressing tendon,
(2) examine the bond characteristics of Fibra in a pretensioned beam,
(3) examine the feasibility of using Kevlar fabric as external shear
reinforcement, and
(4) evaluate the shear and flexural response of a beam reinforced with
nonmetallic materials.
NONMETALLIC MATERIAL PROPERTIES
The properties of the nonmetallic materials are examined for the fiber
behavior characteristics and for the mechanical properties of the composite
material. The Kevlar fiber has specific properties which are germane to their
selection as a prestressing material. These properties are summarized below in
Properties of Kevlar. The mechanical properties of the assembled Fibra rod,
Kevlar parallel fiber rod, and the woven fabric are summarized below in
Mechanical Properties of Reinforcement.
1 Fibra is the trade name of an epoxy-impregnated rod woven of Kevlar fibers
and manufactured by the Mitsui Corporation of Japan.
2 Kevlar is the registered tradename of a family of aramid fibers produced by
the E.l. duPont de Nemours and Co. Inc., Wilmington, DE.
478
Dolan et al
Properties of Kevlar
The properties of Kevlar include high strength and stiffness, low creep
and relaxation, moderate thermal stability, low density, nonconductivity, and
high fatigue resistance (9). The morphology of Kevlar fibers is very important
to their mechanical behavior. A layered structure of rod-like crystallites with
their major axis parallel to the fiber axis forms a core of the fiber which is
enclosed by an outer skin. The layers of fibers are also parallel to the fiber
axis and hydrogen bonds are radially oriented. Under stress the individual
crystal alignment is improved and the strength is enhanced by the anisotropic
orientation of the fiber.
The skin-core morphology is particularly beneficial to tensile strength,
however the compression capacity of the fiber is limited by localized buckling
of the core. Thus, aramids are most effective when used in a tensile stress
condition. The failure mode of an aramid/resin composite is characterized by a
high degree of debonding of the fiber from the resin matrix accompanied by
fibrillation. Fibrillation is the formation of many bundles consisting of a small
number of fibers. The fibrillation is a clear indication that the surface shear
capacity of the aramid fibers is low and bond to these fibers is difficult.
Mechanical Properties of Reinforcement
FIBRA Rod - The principle concern of the prestressing tendons and of the
shear fabric is the tensile stress-strain relationship. The Fibra rods supplied by
Mitsui had an effective diameter of 0.31 inches (7.8 mm) and a rated tensile
capacity of 21.1 kips (93.9 kN). A section of the individual rod was epoxy
grouted into a conical anchor and subjected to a static uniaxial load test. This
load test allowed the determination of the stress-strain characteristic for the
composite. Figure 2 gives the stress-strain results of this test through the first
150 ksi (1.05 GPa). Stresses are computed on the gross cross section and will
be lower than the fiber strength because approximately 40 percent of the cross
section is epoxy resin. The ultimate tensile strength of the rod in Figure 2 was
15.8 kips, 209 ksi, (70.3 kN, 1.46 GPa) and is a result of the anchor efficiency
and not the total rod capacity. The effective modulus of elasticity for the Fibra
rod is 6,400 ksi (44.7 GPa).
KEVLAR Rod - Figure 3 provides a stress-strain curve for the parallel fiber
Kevlar reinforced rod. The tensile capacity of these rods is about 220 ksi (1.54
GPa) and is higher than the Fibra for two reasons. The parallel fiber
arrangement is slightly more efficient that the braided Fibra rod and there are
more fibers in a given cross section. The 1/8 inch (3 mm) diameter rod has
FRP Reinforcement
479
five ends of 7100 denier Kevlar fiber cast in a vinylester resin. This efficiency
is reflected in the effective modulus of 12,000 ksi (83.9 GPa). The differences
in strength and effective modulus indicate some of the important issues to be
addressed in specifying nonmetallic reinforcement.
The lower apparent
modulus in the Fibra rod is due to the mechanical arrangement of the fibers
due to braiding. This lower modulus is expected to improve the loss
characteristics of the tendon due to creep and shrinkage of the concrete.
Assemblage of the smooth parallel fiber Kevlar reinforced rods into a
tendon provides a structure similar to a seven-wire strand. Unlike the Fibra
rod, this tendon does not have the woven surface texture generated by the
braiding and has a lower potential bond capacity.
KEVLAR Fabric - A DuPont 328 plain weave fabric weighing 6.8 oz/yd2 (230
gfm2) was used as a wrap. A one and three quarter inch (44 mm) wide strip of
the Kevlar fabric was epoxy bonded to a cardboard tab and tested for its tensile
capacity and its modulus of elasticity. The results of these tests are given in
Figure 4. The tensile strength was 2500 pounds (11 kN) or 1400 pounds per
inch of width (440 N/mm).
TEST PROGRAM
The test program consisted of static testing of four beams in flexure.
Three beams had Fibra tendons, one beam had a tendon manufactured from six
1/8 inch (3 mm) diameter Kevlar reinforced plastic rods. The first beam was
fabricated with no internal or external shear reinforcement and was tested as a
control specimen. The remaining three beams were tested with an externally
applied Kevlar woven fabric for shear reinforcement.
Figure 1 provides the typical cross section and dimensions of the
beams. Table 1 summarizes the beam, strand type, concrete strength, initial
prestress force and strain, and the theoretical flexural and shear capacity. The
flexural capacity was computed using strain compatibility, a Whitney stress
block for the compressive stresses in the concrete, and equilibrium on the cross
section. The tensile capacity of the shear fabric was counted in the flexural
analysis on the soffit of the beam, but the tensile component of the side fabric
was ignored.
The theoretical shear capacities were computed using the equations in
the ACI 318 Building Code (10). The recorded shear capacities were the
minimum values for Vc as computed using the equations given in the ACI
Building Code. Values for Vc• Vci· and Vcw were computed for the beam
without shear reinforcement. These values are summarized in Table 2.
480
Dolan et al
For the remaining beams, the shear capacity was based only on the
external fabric wrap using the relation
Vs = Av*fy*d/s
The concrete contribution was not included and as a result the shear capacity
for externally reinforced beams is conservative but always exceeds the flexural
capacity of the section. The value for Av*fyls in the above equation is the
tensile strength of the Kevlar fabric determined by testing. The fabric shear
capacity was checked against the interlaminar shear capacity of the adhesive
bonding the wrap to the concrete. In all cases, the theoretical bonding capacity
was greater than the fabric strength.
BEAM FABRICATION
The beams were cast in a fixed form placed between two bulkheads on
the laboratory floor at the University of Delaware. A 4-3/8 inch (Ill mm) end
block the full width of the beam was placed at each end of the beam to control
end splitting stresses. The tendons were stressed, and the concrete cast. The
concrete strength was monitored and when the specified strength was attained,
the tendons were detensioned and the beams were removed from the forms.
The tendon load was monitored during the casting operation to assure that there
were no excessive losses in the epoxy socketed anchors. Figure 5 shows a
typical loss diagram for the short-term anchor effects. Retensioning the rod
just prior to casting the concrete helped reduce these anchorage effects.
Beams F2, F3, and Kl each had a Kevlar wrap epoxied to the web of
the beam up to the intersection of the top filet and the flange. The Kevlar
fabric has a texture much like a loosely woven cloth. It is light and easy to
form. Beam F2 had the wrap applied by rolling the epoxy onto the cleaned
concrete surface and rolling the epoxy into the wrap. In all cases Sikadur 35
High Modulus, low viscosity epoxy resin was used. Numerous air pockets
remained under the wrap on beam F2, leading to improvements in the
installation procedure. The wrap was clamped onto the web of Beam Kl while
the epoxy cured. This improved both the surface contact and the bond. The
concrete surface of Beam F3 was wire brushed prior to applying the wrap.
Sandblasting the surface would provide an even better surface preparation,
however, no sandblasting equipment was available for the project. The wrap
was clamped to the web during the epoxy curing time.
FRP Reinforcement
481
TEST SETUP AND INSTRUMENTATION
The beams were tested in a 400 kip (1. 78 MN) capacity Tinius Olsen
universal test machine. The active loading cylinder is at the base of the
machine, necessitating testing the beams in an inverted position. The control
beam was monitored with dial gages, LVDT' s and levels to record the
deflection and rotation performance of the beam. The brittle failure of the
control section, described later, resulted in using dial gages for all the
remaining tests. Dial gages avoided potential damage to the more expensive
sensors. Figure 6 provides a typical beam instrumentation diagram. Figure 7
shows a typical beam test in progress.
TEST PROGRAM AND RESULTS
The beams were tested in the order of F1 (the control beam) F2, K1,
and F3. The results of each test provided some input and modification to the
subsequent tests. Beams F1, F2 and K 1 were loaded monotonically to failure.
For Beam F3, the load was cycled four times to 50% of the predicted failure
load prior to monotonic loading to failure.
Beam Fl
The initial strain (E = .008 in/in - Eu/3) in the Fibra prestressing tendon
was designed to produce a beam where the primary failure mode was either
shear or crushing of the concrete. The tendon was not expected to rupture. A
load transducer was place between the loading head of the Tinius Olsen
machine and the load spreader bar. This load cell assisted in calibration of the
system loads and was removed after the first beam test.
The beam failed in a shear-flexure mode with the shear failure initiating
at the flexural crack, propagating upward to the intersection with the flange,
and finally failing in shear. The failure load was 2349 pounds (10.4 kN),
corresponding to a maximum moment of 40 kip-in (4.5 kN-m) and a shear of
1160 pounds (5.2 kN). The load was approximately 60% of the theoretical
flexural capacity and the shear was slightly greater than the Vci 1 and Vci2
theoretical shear capacity given in Table 2. These results suggest that the Fibra
rod provides nominally the same concrete contribution to shear resistance as
the ACI equation would predict. Since the Fibra would be expected to have a
lower strength for dowel effects than would a steel dowell, the compliance with
482
Dolan et al
the ACI equations is encouraging. Shear strength is highly variable and more
research is needed to fully confirm this behavior.
The beam failed in a sudden brittle mode. Approximately one third of
the beam separated from the main section and broke free. Approximately 8
inches (200 mm) of the Fibra rod was embedded in the broken end. The fact
that the Fibra did not fracture and that the broken segment did not separate
from the main beam indicate the excellent bond characteristics of the Fibra and
the conformance with the prediction that the beam would fail prior to rupture
of the tendon. The load deflection characteristics of the beam are given in
Figure 8. Figure 9 shows the failed beam.
Beam F2
Kevlar fabric was glued to the external surface of Beam F2 between the
endblocks. The tendon was stressed to a strain of 0.012, "'eu/2. Based on the
beam section dimensions and a concrete design strength of 5000 psi (35 MPa),
there was an equal probability that the beam would fail by crushing of the
concrete in the compression zone or by rupture of the tendon. The beam was
loaded monotonically to failure in the same manner as beam Fl.
Failure occurred in flexure at a load of 5470 pounds (24.3 kN),
corresponding to a maximum moment of 91.7 kip-inches (10.1 kN-m) and a
shear of 2735 pounds (12.2 kN). The Kevlar wrap not only provided sufficient
shear capacity to develop the full tensile capacity of the tendon, but also added
some flexural capacity due to the fabric on the beam soffit. The failure load
was 4% above the theoretical flexural capacity of the beam.
The moment-deflection curve for beam F2 is given in Figure 8.
Comparison with the moment-deflection curve for beam F2 indicates the
following behavior. Up to the initial cracking load -- approximately 20 kipinches (2.2 kN-m) for beam Fl and 30 kip-inches (3.3 kN-m) for beam F2 -both beams behave elastically. Beyond the cracking load, beam F2 exhibits
greater stiffness than beam F 1. This is due to the supplemental reinforcement
provided by the Kevlar wrap on the beam soffit. Beam F2 shows considerably
more ductility, i.e., deflection after initial cracking, than beam Fl. The
uniform increase in moment with increased deflection is characteristic of beams
reinforced with materials having no yield point.
FRP Reinforcement
483
Beam Kl
Beam K2 was prestressed with six 118 inch (3 mm) diameter Kevlar
reinforced rods. The rods were stressed to an initial strain of 0.012 in/in
approximately Eu/2, and one rod broke during stressing. Thus, the initial
prestress was less than beams F2 and F3. The rods were coated with an
epoxy, and the wet epoxy was sprinkled with coarse dry sand.
This
significantly improved the bond strength of the tendon. Separate tests on a
single 1/8 inch (3 mm) diameter rod cast in a 1 inch square ( 25.4 mm) beam
section found that sand coating the rod resulted in a bond transfer length of
about 56 rod diameters. This transfer length is comparable to the transfer
length of metallic tendons. The external Kevlar fabric was applied to a
concrete surface which had been cleaned with a wire brush. The fabric was
clamped to the beam while the epoxy cured.
The maximum load on beam K1 was 5454 pounds (24 kN),
corresponding to a maximum moment of 94.1 kip-inches (10.3 kN-m) and a
shear of 2727 pounds (12.1 kN). A loud snap was heard and the load fell to
4363 pounds (8.8 kN). Applying additional displacement to the beam resulted
in a complete flexural failure at 3300 pounds (14. 7 kN). Examination of the
beam after the failure indicated that the initial failure occurred when one of the
six individual rods fractured. Continued application of load and displacement
resulted in the fracture of the remaining rods. The rods broke at different
locations within the beam. The residual strength was due to the bond between
the rod and the concrete as the broken rods were pulled out of the beam. Dial
gages on the ends of the tendons indicated there was no end slippage, further
confirming that the failure was due to tendon fracture.
Figure 10 provides the moment-deflection characteristics of beam Kl.
The behavior is similar to beams Fl and F2. As in beam F2, no indication of
shear failure was present. The actual strength was approximately 20% greater
than the theoretical prediction. The larger variation between predicted and
actual strength may be due to the assumptions of strain conditions in the rod
broken during stressing. If the single rod failure was due to bond failure in the
anchor and some residual strain remained, the capacity of the total tendon
would be higher than predicted and the theoretical results would be closer to
the actual results.
Beam F3
Beam F3 contained one Fibra rod stressed to an initial strain of 0.012
in/in, approximately Eu/2. The Kevlar wrap was applied to a concrete surface
484
Dolan et al
which had been scarified with a wire brush to remove any loose material. The
wrap was clamped to the concrete during the epoxy cure.
Beam F3 was subjected to five cycles of load each reaching 50% of the
predicted ultimate flexural strength of the beam. Figure 11 shows the beam
test in progress. After each cycle, the beam was unloaded and reloaded.
Finally the beam was loaded monotonically to failure. Failure occurred at
5641 pounds (24.3 kN), corresponding to a moment of 95.6 kip-inches (10.5
kN-m) and a shear of 2820 pounds (12.5 kN).
This was 2% above the
The moment-deflection
theoretical flexural capacity of the section.
characteristics of the beam are given in Figure 12. Failure was in flexure with
no indication of shear failure.
OBSERVATIONS AND FINDINGS
While the external fabric successfully provided the shear reinforcement
needed to develop the flexural capacity of the beams, two characteristics of the
fabric performance were noticed. As the beam was loaded beyond the initial
cracking values, numerous cracking sounds were heard. These sounds were
associated with the delamination of the fabric along the side of the beam. The
localized failures led to the fabric wrap peeling off the beam in a single section
with no damage to the fabric. While the shear capacity was checked against
the bond capacity of the adhesive, no allowance was made for microdelamination of the fabric. Superior bond between the fabric and the concrete
may have improved the overall performance of the fabric and postponed the
progressive delamination observed in these tests. Once the fabric delamination
reached a critical point, the tensile contribution of the fabric to the flexural
strength was lost. This led to complete reliance on the tendon and assured that
the tendon ruptured.
The bond between the fabric and the concrete might have been
improved if the concrete surface had been roughened. Sandblasting was
considered, but sandblasting equipment was not available. Consequently, the
tests were conducted with relatively marginal surface preparation.
Additionally, air pockets were observed at the corners as the fabric bent from
the soffit to the web. Chamfers, or application of pressure on the corners,
would improve this condition.
The fabric used to reinforce the concrete had equal fiber content in both
directions. Since the longitudinal flexural strains and cracking were much
greater than the vertical strains, the performance of the fabric might have
improved and the tendency to delaminate been reduced if a non-symmetric
fiber pattern had been used. A fabric could be woven with the desired
FRP Reinforcement
485
longitudinal and vertical distribution of strength. Alternatively, the fabric
could be applied in vertical strips to separate the longitudinal and vertical
motions. The fabric is not run into the compressive zone of the beam because
the flexural compressive strains can buckle the fabric and accelerate
delamination.
The fabric was especially easy to apply. It is light and flexible. Epoxy
can be applied using a brush or a roller. Retrofitting existing structures and
overhead work can be easily accomplished using this fabric. Improved surface
bonding could be accomplished by using a vacuum bonding mechanism during
the resin cure.
The Fibra rods displayed excellent bond and flexural characteristics.
The sudden failure of beam Fl and the ability of the short section of Fibra rod
to retain the beam end indicated that both the dynamic and the static bond is
excellent. Beam F3 showed little deterioration of stiffness and no strength loss
due to the cyclic load. Some stiffness loss may be due to cracking of the
concrete and working of the braided Fibra rod. The final strength of the
section was not diminished, indicating that working of the Fibra rod did not
affect its strength. The softening of the rod due to the resin breakdown leads to
an increased strain with no effective increase in stress. This loosening of the
epoxy bonding around the Fibra fibers in the braided rod may create a
condition similar to yielding in metallic reinforcement.
The issue of ductility with nonmetallic tendons continues to be a topic
of discussion because both the concrete and the tendon are "brittle" materials.
Historically, ductility is defined as the amount of deformation after initial
yielding. Since neither the concrete nor the nonmetallic tendon yield, this
definition is not completely suitable for these structures. If ductility is defined
as the amount of deformation following initial cracking of the beam, then these
beams display considerable ductility. A conventional prestressed concrete
beam will have a moment deflection curve that rapidly changes slope once
cracking in initiated. The slope, representing the stiffness of the section,
approaches zero as the steel tendon reaches its ultimate capacity. Prestressed
beams with nonmetallic tendons typically show a nearly linear increase of
deflection with increased moment until either the tendon ruptures or the
concrete compressive strain is exceeded. Figures 8, 10, and 12 indicate
substantial additional deformation following the initial cracking.
This
deformation is comparable to the deformation in conventionally prestressed
beams and provides ample warning of overload. Thus, the ratio of final
deformation or strain to the deformation or strain at initial cracking is of the
same order of magnitude as conventionally prestressed members. The strain
reserve at failure is a function of the initial design and must be checked to
assure that the concrete compressive capacity of the section may be developed.
486
Dolan et al
The one undesirable characteristic of the Fibra is the need for careful
handling. The Fibra rods are lightweight and easily handled by a single
individual. However, sharp bends -- especially those that crack the epoxy
resin-- produced localized fiber failures. These "notches" occurred as some
durability test samples were being loaded into the test machine. The weight of
the anchor caused a localized bend in the rod. The test specimen failed at the
notch. Failure occurred at less than 50% of the tensile capacity of the rod.
CONCLUSIONS AND RECOMMENDATIONS
External reinforcement successfully allowed the full development of the
flexural capacity of the sections. The strength of the members were accurately
predicted using equations of equilibrium and compatibility. Thus, the Fibra
and Kevlar prestressing rods provided the anticipated strength and
performance. The "ductility" of the member is a function of the strain reserve
available after prestressing. The strain reserve is a function of both the section
dimensions and the nonmetallic tendon. For a fully prestressed member, there
is substantial strength and deflection reserve after initial cracking. Figures
8,10 and 12 indicate that the ductility ratio, as measured by the total tensile
strain at failure to the tensile strain at initial cracking, is substantial.
Nonetheless, a consensus on the definition of ductility ratio needs to be
developed for nonmetallic tendons where there is no clearly defined "yield
zone." The loosening of the resin/fiber bond in the Fibra rod may provide
more ductility than parallel fiber rods. The ability to design ductility into a
nonmetallic prestressing tendons offers interesting research opportunities.
The "concrete contribution" to the shear capacity in beams with
nonmetallic Fibra tendons was accurately predicted by the current ACI
formulas. Since beam Fl failed at slightly more than the values suggested by
ACI, and since shear tests typically have considerable scatter, additional
research is justified to validate the use of the Vci and Vcw equations for Fibra
and other nonmetallic tendons.
The external shear reinforcement performed particularly well in
"retrofitting" the test members. Structural strengthening using externally
applied shear reinforcement may be considered in conjunction with FRP
external reinforcement for increased flexural capacity. Selection of appropriate
epoxies and heat displacement temperatures requires additional research. This
research should be coordinated with an evaluation of the composition of fabric
fiber orientation to assure optimum utilization of the materials.
FRP Reinforcement
487
ACKNOWLEDGMENTS
Support for this research was provided by E.I. duPont and the Delaware
Research Foundation. DuPont also provided the Kevlar fabric for external
shear reinforcement. The Fibra rods were provided by Mitsui Corporation of
Japan. Sika Chemicals donated the epoxy resins used in the prestressing
anchors and for bonding the fabric to the beams.
REFERENCES
1. Iyer, S.L., and Sen, R. eds., Advanced Composite Materials in Civil
Engineering Structures, American Society of Civil Engineers, New York, New
York, 1991.
2. Rostasy, R.S., and Budelmann, "FRP-Tendons for the Post-Tensioning of
Concrete Structures", Iyer, S.L., and Sen, R., eds., Advanced Composite
Materials in Civil Engineering Structures, American Society of Civil
Engineers, New York, New York, 1991, pp. 155.
3. Gerritse, A, and Schurhoff, H.J. Prestressing with Aramid Tendons,
Technical Contribution to FIP lOth Congress, New Delhi, India, 1986.
4. Dolan, C.W., (1991), "Kevlar Reinforced Prestressing For Bridge Decks",
Third Bridge Engineering Conference, Transportation Research Record, No.
1290, Vol. 1, Transportation Research Board Washington, D.C., 1991, p. 68.
5. CFCC TECHNICAL DATA (1989), Tokyo Rope Manufacturing Company,
Ltd., Tokyo, Japan, October, 1989.
6. "Long-Fiber Aramid Reinforcing Bar 'FIBRA-bar"' ,(1986), Technical
Research Institute, Mitsui Construction Co. Ltd., Tokyo, Japan, 1986.
7. Walton, J.M., (1986) and Yeung, Y.T.C., "The Fatigue Performance of
Structural Strands of Pu1truded Composite Rods", Journal of the Institute of
Mechanical Engineers, London, C286/86, 1986, p. 315.
8. Meier, U. (1991), "Strengthening Structures with CFRP Laminates", Iyer,
S.L., and Sen, R., eds., Advanced Composite Materials in Civil Engineering
Structures, American Society of Civil Engineers, New York, New York, 1991,
p. 224.
9. Data Manual for Kevlar 49 Aramid (1987), E.I. duPont de Nemours & Co.
Inc., Wilmington, DE, 1987.
Dolan et al
488
10. Building Code Requirements for Reinforced Concrete (1990), ACI 318-90,
American Concrete Institute, Detroit, MI, 1990.
TABLE 1 -BEAM PROPERTIES AND CHARACTERISTICS
US Conventional units
Test
No.
1
2
3
4
Beam
No.
Tendon
Type
f' c
ksi
Fl
F2
Kl
F3
Fibra
Fibra
Kevlar
Fibra
4790
4990
4390
4790
in/in
P·I
kips
Mn
in-lb
Yn
LB
.008
.012
.012
.012
3.85
5.78
4.95
5.78
81.6
87.7
76.5
94.0
8841
15,7142
15,714
15,714
~::·
P·I
kN
Mn
kN-m
Yn
kN
17.1
25.7
22.0
25.7
8.98
9.65
8.42
10.3
3.9
61.82
61.8
61.8
~::·
I
SI Units
Test
No.
1
2
3
4
Beam
No.
Tendon
Type
f' c
MPa
m/m
F1
F2
K1
F3
Fibra
Fibra
Kevlar
Fibra
33.2
34.9
30.7
33.5
.008
.012
.012
.012
I
Notes:
f' c
= strength of the concrete at the time of the test
~::·
= initial prestressing strain on the tendon
I
P·I
= initial prestressing force
= computed nominal bending capacity of the section
Mn
= computed nominal shear capacity of the section
Yn
1.
Vn is computed as the lesser of Vc• Vci, and Vcw· Beam Fl
had a minimum computed shear value for Vci = 884 pounds
(3. 9 kN) and is used to compute the strength ratio in the text.
2.
Vn is computed on the basis of the external reinforcement
strength and does not include the concrete contribution.
FRP Reinforcement
489
TABLE 2 - SHEAR CAPACITY OF EXTERNAllY REINFORCED BEAMS
US Conventional Units
Beam
Shear
Reinforcement
Yc
Shear
Veil
none
fabric
fabric
fabric
1161
1238
1541
1642
970
990
1859
1941
Fl
F2
Kl
F3
Ca~acit~
(nounds)
Yci2
Yew
Ys
880
880
1240
1240
2510
2850
4544
4770
0
15714
15714
15714
(kNl
Yci2
Yew
Ys
3.92
3.92
5.52
5.52
11.2
12.7
20.2
21.2
0.0
69.9
69.9
69.9
SI Units
Beam
Shear
Reinforcement
Yc
Shear
Veil
none
fabric
fabric
fabric
5.16
6.39
6.86
7.31
4.32
4.41
8.27
8.64
F1
F2
Kl
F3
Ca~acit~
Notes:
The ACI Building Code requires that the shear capacity of the
beam be the lesser of Vci or Vcw· Alternatively, Vc may be
used in lieu of computing Vci or Vcw· Furthermore, the value
of Vci need not be less than 1. 7-.Jf• cbwd. This value is noted as
Vci2 in Table 2.
Yc = (0.6-.Jf'c + 700 YudiMu)bwd
Veil = 0.6-.Jf'cbwd + Yct +ViMcrl Mmax
ACieq. 11-10
AC!eq.ll11
This is the equation for flexural-shear strength of a beam.
Yew = (3.5-.Jf'c + 0.3fpc)bwd + Yp
ACI eq. 11-13
This is the web cracking limitation of shear capacity of the
beam. SI equivalent equations for the above are given in the
appendix to reference (10)
490
Dolan et al
8- 0"
~ 4- 3/8"
Beam Elevation
r- 6" -l
1--
8" ~
H
1-1/2"
Beam Section
Beam Section
Beams K 1 and F3
Beams F 1 and F2
Fig. 1-Beam section and profile
175000
I
150000
125000
fc
I
l
100000
~
!!
,;;
J
75000
50000
25000
0
I
j
1/
0.005
1/
I
0.010
0.015
0.020
0.025
0.030
Strain On/in)
Fig. 2-Fibra rod stress-strain cuiVe
FRP Reinforcement
491
300
......______
250
...m
/
200
..
I
"'"'
s"'
v
150
100
50
/(
L
/
/
0.005
0.01
/
/
/
0.015
0.02
0.025
STRAm
Fig. 3-Kevlar rod stress-strain curve
2500
I
2000
1500
(
j
I
I
c
.....
/
Cl
..:1
1000
500
0
)
v
0.1
0.2
0.3
0.4
0.5
0.6
Elongation (in)
Fig. 4-Fabric load figure
0.7
o. 03
492
Dolan et al
20
18
16
14
..."'m
"'
I
12
10
~
..:1
10
20
30
40
50
60
70
HOURS
Fig. 5-Short term anchor effects
46''
46"
14~ ..
~-LEVEL
~-DIALGAGE
Fig. 6--Beam test instrumentation set-up
80
90
100
FRP Reinforcement
Fig. 7-Beam test in progress
100
_,.
90
..-/
so
/
70
~
60
,.I
...
;g
=
e
Q
::::;
50
40
I
30
/
20
I
----e--
Beam 1
Beam2
.~
/
.~
10
0
0.0
0.5
1.0
1.5
2.0
2.5
Denection (in)
Fig. 8-Moment deflection behavior of beams Fl and F2
493
494
Dolan et al
Fig. 9-Beam Fl after failure - Shear fabric is removed
to expose crack pattern
100
/'
v
L:!_
80
70
:!
...
;g
..
;:
60
50
E
"'
40
,,
~u
90
I
v
/
L
(
30
20
10
0
0.00
0.25
0.50
0.75
1.00
1.25
1.50
DeDection (in)
Fig. 10-Moment-deflection behavior of beam Kl
FRP Reinforcement
Fig. 11-Beam F3 during testing
100
v
90
/
80
/
60
Q.
;g
50
;:
.e
::;
.....
/
70
~
/
(
/
/
40
30
20
10
0.2
0.4
0,6
0.8
1.0
1.2
1.4
1.6
Deflection (in)
Fig. 12-Moment-deflection behavior of beam F3
495
SP 138-30
Flexural Behavior of Masonry
Walls Strengthened with
Composite Fabrics
by M.R. Ehsani, H. Saadatmanesh,
I.H. Abdelghany, and W. Elkafrawy
Synopsis: A new approach for seismic retrofitting of unreinforced masonry
structures is presented. Results from flexural tests of six masonry beams are
reported. The beams were strengthened by epoxying a composite fabric to
their tension face. The variables studied include the type of epoxy, fabric,
mortar, brick, and the finished surface of the wall. It is shown that both the
strength and ductility of the beams were significantly improved. When a
sufficient amount of fabric is used to prevent its premature tension failure, the
integrity of the beam is maintained until the beam fails by reaching the
compressive strength of the masonry.
Keywords: Composite materials; ductility; earthquake resistant structures;
epoxy resins; fabric; flexural strength; masonry; repairs; strength
497
498
Ehsani et al
Mohammad R. Ehsani is Associate Professor of Civil Engineering at the
University ofArizona. He is a Member ofACI Committee 408 where he chairs
a Subcommittee on Bond of FRP Rebars. As a member of Committee 440 he
chairs the Subcommittee on State-of-the-An Repon. Dr. Ehsani is a registered
professional engineer in Arizona and California.
ACI member Hamid Saadatmanesh is Assistant Professor of Civil
Engineering and Engineering Mechanics at the University of Arizona in
Tucson. His research interests include rehabilitation and strengthening of
structures and the application of fiber composites in civil engineering
structures. He is the Secretary of ACI Committee 440, FRP Tendon and
Reinforcements.
I. H. Abdelghany is a graduate student at the University ofArizona. He
received his B.S. in Civil Engineering from Cairo University in 1985. His
interests include seismic behavior of concrete structures.
W. Elkajrawy attended the University of Arizona as a Peace Fellow in
1992. He holds a B.S. degree in Civil Engineering from Cairo University.
INTRODUCTION
One of the most serious problems facing the earthquake engineering
community today is the very large number of older masonry buildings in
seismic regions that were built before any provisions for earthquake loading
were required. These structures are usually constructed from brick or concrete
block and in older cases stone. The units are tied together by a cement mortar
mixture and in some cases, steel or other reinforcement.
While there are several types of masonry structural elements within a
building, the most commonly used and subject to earthquake damages are load
bearing walls. These elements are planes with or without openings which are
designed primarily to carry the vertical loads within the structure. In a seismic
event, however, they must also carry in-plane or out-of-plane horizontal loads
resulting from the earthquake. The damage caused by the lack of or
insufficient reinforcement in these elements has been well documented in
recent earthquakes (1-3).
Unreinforced masonry (URM) buildings are perhaps the most common
type of existing construction worldwide. The high risk of earthquake damage
to older masonry buildings, particularly when unreinforced, and the potential
FRP Reinforcement
499
for a great loss of life and property has led to a wide range of research studies
on masonry structures. One subject that is of great interest to the engineering
community is that of defining methods whereby these structures can be
strengthened or upgraded in lateral load carrying capacity to satisfy modern
seismic design codes.
In the U.S., most of the retrofit techniques developed to date are rather
expensive to implement. Consequently, unless mandated by law, the costs
associated with such strengthening techniques have prevented most building
owners from voluntarily adopting these modifications. The objective of this
research study is to develop a simple and inexpensive technique for seismic
retrofitting of masonry structures.
STRENGTHENING TECHNIQUES
Various methods for strengthening masonry walls have been studied in
recent years. The methods for strengthening of masonry structures can be
divided into two general categories. In one case, significant structural
elements are added to the existing structure, resulting in a substantial change
of dimensions in the original structure. In the other cases, the wall surface(s)
are covered with various coatings to increase its strength and ductility.
The first category of strengthening involves addition of boundary
elements such as reinforced concrete columns (4). The problem with this
technique is the expense associated with the addition of columns and the loss
of valuable floor space. Strengthening may also be achieved by removing one
or more wythes of masonry and replacing the volume with a heavy coat of
pneumatically applied concrete. While this type of upgrading is effective, it
usually requires a great deal of preparation work and often adds considerably
to the weight of the structure which, in turn, may lead to foundation
adjustments.
The second category of strengthening is by surface treatment. A
possible solution is to apply a layer of reinforced shotcrete to the surface of
the unreinforced masonry wall (5). Others have investigated the application
of ferrocement overlays (6). Although effective, these methods are usually
accompanied with a great deal of disturbance to the occupants while the
structure is undergoing rehabilitation.
Recent studies at the University of Arizona have demonstrated that the
strength of concrete beams can be significantly increased by epoxy bonding a
sheet of composite laminate to the tension flange of the beam (7,8). The
method presented here as an extension of the above studies, where for ease of
500
Ehsani et al
application, a thin flexible fabric of glass is epoxied to the masonry wall. The
steps required in strengthening an in-fill frame, for example, include: a)
cleaning the wall surface(s) and if required, filling the mortar joints flush with
the surface of the wall; (b) applying a thin layer of epoxy to the wall surface(s)
and the adjacent frame elements; (c) placing the composite fabric on the
epoxied surfaces and pressing it firmly against the wall; and (d) applying an
additional layer of epoxy to the outer surface of the fabric. If desired, the
edges of the fabric could be bolted to the frame for additional strength. The
surface of the wall could also be covered with plaster. Figure 1 shows a
masonry beam while the fabric is being epoxied to the one of the surfaces.
EXPERIMENTAL STUDY
A study is currently under way to examine the feasibility of this
retrofitting technique. Several small masonry beams have been constructed.
The beams consist of 19 clay bricks, each with a dimension of 2 1/2*4*8 112 in.,
stacked in a single wythe (stack bond). This results in beams which are 8 1/2in. wide, 4-in. high and 57-in. long. The beams are loaded statically to failure
with two concentrated loads over a clear span of 47 in., as shown in Fig. 2.
The designation for the test specimens starts with the letter "B" for Beam,
followed by a combination of four additional letters/numerals, each describing
a variable in the study. These variables include:
1.
Epoxy - The first numeral 1 or 2 in the designation of each
specimen refers to the type of epoxy. Two epoxies are being
investigated. The first one is a two-component epoxy that
performed exceptionally well under previous studies for
strengthening of RIC beams. Among the features of this epoxy
are its high energy absorption, resistance to high humidity, salt
spray, cold and hot environments, and economy. The epoxy
has a consistency similar to cement paste with a pot life of
approximately 1/2 hour. It is fully cured in room temperature
in four hours. A dual-component dispense tool was used to
achieve a uniform mixture of the epoxy as it was being applied
to the wall and fabric. The second adhesive being studied is
also a two-component epoxy which cures at room temperature.
This epoxy has a lower viscosity than the first one and can be
easily spread over the wall surface with a trowel.
2.
Mortar - Because the mortar in older buildings could have
deteriorated and weakened, two different mortar strengths are
being evaluated. The letter "S" refers to a strong mortar with
cement:lime:sand ratios of 1:1,4:3. The letter "W" is used for
FRP Reinforcement
501
the weak mortar having ratios of 1: 1.4:5, respectively.
3.
Fabric - The fourth character (i.e. 1, 2, or 3) refers to the type
of fabric used. Three different fabrics of various strength (i.e.
thickness and weave) have been used to investigate the
possibility of achieving various modes of failure, such as by
tension failure of fabric, by compression failure of brick, etc.
4.
Surface - The last letter refers to the roughness of the face of
the wall where the fabric is attached. This was a function of
the finishing of joints between courses of masonry units before
the mortar had hardened. In both cases, the fabric was epoxied
to the smooth surface of the brick. In one case, the mortar
joint was flush with the outside surface of the wall; these
specimens are designated with the letter "F". In the other case,
a small amount of mortar extruded from the joints; these are
designated with "S". The two surfaces were used to investigate
the effect of the surface finish on bonding of the composite
fabrics.
5.
Brick - All specimens were cast with new clay bricks.
However, because the age of a brick may influence its bonding
characteristics to the epoxy, one specimen was cast with
reclaimed old bricks (Specimen B1S2S-1).
The results for six specimens which have been retrofitted and tested are
presented here.
Materials
As mentioned earlier, two types of mortar were used in this study.
Two- by Four-inch cylinders of the mortar were tested at 28 days and the
compressive strength was calculated as 4650 and 4100 psi for the strong and
weak mortars, respectively. Prisms were also constructed with the new brick
and strong mortar. The 28-day strength of the prisms was calculated as 1870
psi. The prisms failed by compression failure of the bricks; consequently, the
slight change in the mortar strength did not have a significant effect on the
overall strength of the specimens.
Three types of fabrics were used. The first one was a fiberglass fabric
with an acrylic polyvinyl finish which comprises about 6-10% of the product
weight. The fabric weighs 5.6 oz/yd 2 and had a visual 2x4 yarns/in.
construction in the machine (warp) and cross-machine (fill) directions.
502
Ehsani et al
According to the manufacturer, the tensile strength of the fabric as determined
by ASTM-D579 3-inchjaw separation at a crosshead speed of 12 in./min. was
220x270 lbs/in. in the weak and strong directions, respectively. This fabric
was epoxied to the specimens with the strong direction being parallel to the
length of the beam.
The second and third fabrics were unidirectional E-glass. Five samples
of each fabric were tested by the manufacturer in accordance with the out strip
method of ASTM-D1682. The results indicated that the second fabric had
11.3 yams per inch and a tensile strength of 1422 pounds per inch. The
corresponding numbers for the third fabric were 10 and 855.
Test Setup and Instrumentation
The specimens were tested in the frame shown in Fig. 3. Roller
supports were provided at the ends of the specimens and at the points of
application of the load. The midspan deflection of the beams was measured
with two dial gages, placed at both sides of the specimen. Strain gages were
attached to the masonry units on the compressive face of the beam and on top
of the cured fabric on the tension face of the beam.
The specimens were loaded at a constant rate of loading up to failure
of the beam. At 100-pound intervals, loading was temporarily halted while the
strain gages and dial gages were manually read and recorded.
Test Results
Before discussing the results, it is interesting to note that the test
specimens would normally fail under their self weight of approximately 125
pounds. Therefore, prior to strengthening, the specimens were handled very
carefully.
Plots of load vs. midspan deflection for six retrofitted specimens are
presented in Fig. 4. The first fabric, used in Specimen B2S1S, was relatively
weak. Nonetheless, the specimen carried a maximum load of 700 lbs. and a
deflection of 0.27 in. The ultimate load was governed by tension failure of the
fabric. Based on this test it was decided to utilize stronger fabrics in the
remaining tests.
The influence of the strength of the fabric can be readily seen by
comparing Specimens BlS2S and B1S3F, both retrofitted with the same epoxy
FRP Reinforcement
503
(i.e. Type 1). The thicker fabric in B1S2S resulted in a failure load of 2850
lbs. and a deflection of 0.63 in. Failure was initiated by compression crushing
of the bricks near the top of the beam, followed suddenly by diagonal cracking
of the beam in the shear span. Specimen B1S3F had a smaller stiffness due
to the thinner fabric used. This specimen reached a maximum load of 1320
lbs. and a deflection of 0.65 in. At that point, the fabric failed in tension.
Figure 5 shows Specimen B1S2S at the end of the test.
The performance of the second epoxy was superior to that of the first
one. This is evident from comparison of the results for Specimens BlS3F and
B2S3F. Both specimens were retrofitted with the lighter E-glass fabric. The
performance of Specimen B1S3F was discussed above. Specimen B2S3F had
a higher stiffness and reached a load of 1950 lbs at a deflection of 0.98 in.
Both specimens failed by tension failure of the glass fabric. However, the
additional load carried by Specimen B2S3F is attributed to the type of epoxy
used in this specimen.
Comparison of Specimens B1S2S and B1S2S-1 can reveal information
on the performance of the two types of brick used. Specimen B1S2S,
constructed with the new brick, had a larger stiffness and failed at a load of
2850 lbs. Specimen B1S2S-1, which was constructed of old reclaimed brick,
failed at a load of 1400 lbs and at a deflection of 0.48 in. Due to the large
area of the fabric used, both of these specimens failed by compression failure
of the brick. Although no prism tests were performed for the reclaimed brick,
it is believed that the lower strength of this brick resulted in the lower failure
load for the specimen.
The effect of the mortar strength appeared to be negligible in these
specimens. Specimen B1S2S with the strong mortar failed at a load of 2850
lbs. while its companion specimen with weaker mortar, B1W2S, failed at a
load of 3000 lbs. Both of these specimens were retrofitted with the thicker
fabric and failed by compression failure of the masonry. In masonry prism
tests, it was observed that failure was initiated by compression failure of the
clay brick rather than the mortar. Consequently, the slight difference in the
strength of the mortar in these two specimens did not change the mode of
failure and the maximum load carried by both specimens were comparable.
Examination of the specimens during and after the tests indicated that
none of them showed any visible sign of slip or bond failure at the
epoxy/fabric interface.
504
Ehsani et al
FURTHER STUDIES
The proposed method can also be used to increase the shear strength
of unreinforced masonry walls. For both flexural and shear strengthening of
walls, the connection of the fabric to the framing elements can be achieved by
epoxy or a combination of epoxy and mechanical connectors such as steel
angles and bolts. The strength and ductility of these connections has a
significant effect on the overall success of this technique. These topics are
currently under investigation at the University of Arizona.
CONCLUSIONS
The test results indicate that retrofitting of unreinforced masonry
structures with composite fabrics is a very effective technique for increasing
the flexural strength and ductility of these elements. The specimens tested,
carried loads more than twenty times their own weight and exhibited large
deflections, in excess of 1/50 times the span. The strength of the fabric
controlled the mode of failure. When lighter fabrics are used, the maximum
load is that causing tension failure of the fabric. When stronger fabrics are
used, the fabrics maintained the integrity of the specimen until the compressive
strength of the brick was reached. Considering the ease of application of this
method, it appears to have a great potential for seismic retrofitting of masonry
structures.
ACKNOWLEDGEMENTS
Funding for this project has been provided under the National Science
Foundation Grant No. BCS-9201110, Dr. C.S. Liu Program Director. This
support is gratefully acknowledged. However, the views expressed in this
paper are those of the writers and do not necessarily represent the views of the
sponsor.
REFERENCES
1.
Benuska, L., Technical Editor, (1980). "Lorna Prieta Earthquake
Reconnaissance Report," Earthquake Spectra, Supplement to Vol. 6,
pp. 127-149.
FRP Reinforcement
505
2.
Villablanca, P.R., Klinger, R.E., Blondet, M., and Mayes, R.L.
(1990). "Masonry Structures in the Chilean Earthquake of March 3,
1985: Behavior and Correlation with Analysis," The Masonry Society
Journal, 9(1), pp. 20-25.
3.
Rosenblueth, E., and Meli, R. (1986). "The 1985 Earthquake: Causes
and Effects in Mexico City," Concrete International, 8(5), pp. 23-34.
4.
Zezhen, N., Qi, D., Jianyou, C., and Runtao, Y. (1984). "A Study of
Aseismic Strengthening for Multi-Story Brick Buildings by Additional
RIC Columns," Proc., 8WCEE, San Francisco, California, pp.
591-598.
5.
Kahn, L.F. (1984). "Shotcrete Retrofit for Unreinforced Brick
Masonry," Proc., 8WCEE, San Francisco, California, pp. 583-590.
6.
Prawel, S.P., and Reinhorn, A.M. (1985). "Seismic Retrofit of
Structural Masonry using a Ferrocement Overlay," Proc., Third North
American Masonry Conference, Arlington, Texas, 19 pp.
7.
Ehsani, M.R., and Saadatmanesh, H. (1990). "Fiber Composite Plates
for Strengthening Bridge Girders," International Journal of Composite
Structures, 15(4). pp. 343-355.
8.
Saadatmanesh, H., and Ehsani, M.R. (1991). "RIC Beams
Strengthened With GFRP Plates: Experimental Study," Jr. of Str.
Engrg., ASCE, 117(10), pp. 3417-3433.
Fig. !-Composite fabric being epoxied to a masonry beam
506
Ehsani et al
P/2
l
P/2
l
111111111111111111
zs
2S
21 in.
21 in.
530mm
530mm
Fig. 2-Dimensions of the specimens
Fig. 3-Test set-up
II4in.
100mm
FRP Reinforcement
507
Load vs. Deflection
URM Beams
3500.-------.-------r------.-------.-------,
3000 - - - · - - ······---·--·······-·
g.
2000 ·-----··
~
1500
1000
0.2
0.4
0.6
0.8
Midspan Deflection (in)
Fig. 4-Load versus midspan deflection for the specimens
'81 S2S
2'too lbs
!
cJ
./
Fig. 5-Specimen B1S2S at the conclusion of the test
SP 138-31
Flexural Fatigue Behavior of
Prestressed Concrete Beams
Using Aramid-Fiber Tendons
by K. Iwamoto, Y. Uchita
N. Takagi, and T. Kojima
Synopsis ; The fatigue tests of pretensioned concrete
beams with aramid fiber rods as prestressing tendons
were carried out in order to investigate the effects
of upper load ratio and initial tension of tendon on
the fatigue strength and deformation of beams.
The size of the beam was 15xl5x210cm.
Three
levels of initial tension of tendon were chosen in the
lower tendons of beam, approximately 40%, 60% and 70%
of the tensile strength of rod.
The symmetrical two
point loading was applied to the beam.
The upper load
in cyclic loading was varied from 45% to 80% of the
statical ultimate loads, and the lower load was constant at 4.9kN.
The sinusoidal loading of the frequency of 4Hz was applied.
The beams with aramid fiber tendons behaved well
in fatigue performance, comparable to that of the
beams with prestressing wires. The effect of initial
tension of tendon on the fatigue life of the beam was
not observed, and the fatigue life was determined by
the upper load.
The fatigue strength at two million
cycles of loading was more than 65% of the static
ultimate strength of beam.
Keywords: Beams (supports); fatigue (materials); fatigue tests; fibers;
flexural strength; prestressed concrete; prestressing; prestressing steels;
pretensioning; strength; wires
509
510
Iwamoto et al
Kaoru Iwamoto is a chief engineer in Development
Department of Kinki Concrete Industries Co.,Ltd.
He
has been associated with concrete reinforced by
choped-fiber or fiber-tendon.
Yasuo Uchita is a chief research engineer in Technical
Research Center of Kansai Electric Power Co.,Inc. He
has been associated with durability of concrete structures.
Nobuaki Takagi is a research associate in Civil Engineering Department of Ritsumeikan University.
His
main research is the mechanical properties and durability of silica fume concrete.
Takayuki Kojima, ACI member, is a professor in Civil
Engineering Department of Ritsumeikan University.
The
torsion and fatigue of reinforced and prestressed
concrete structures are his present fields of study.
He has published many technical papers and reports.
INTRODUCTION
Fiber reinforced plastic (FRP) rods have developed recently and become one of the new structural
materials.
FRP rods are made of high strength continuous carbon, aramid or glass fibers impregnated with
resin, and have some excellent properties such as high
tensile strength, non-corrosion, non-magnetization and
so on.
It is important to utilize these excellent
properties effectively when the FRP rods are applied
to concrete structures.
The availability of FRP rods,
especially carbon and aramid fiber rod, as the prestressing tendons has been discussed [1].
The flexural behaviour and ultimate strength of prestressed
concrete beams using aramid fiber tendons can be
estimated by the conventional method used in usual
prestressed concrete structures.
On the other hand, the fatigue behaviour of
prestressed concrete beams with FRP rods is not well
known, while some experimental work reported that the
fatigue behaviour of FRP rods alone under the cyclic
tension loading are almost comparable to that of
prestressing steel strand [2].
In this study, the fatigue tests of pretensioned
concrete beams with aramid fiber rods as prestressing
tendons were carried out in order to investigate the
effect of upper load ratio and initial tension of
tendon on the fatigue strength and deformation of
beams.
FRP Reinforcement
511
OUTLINE OF EXPERIMENT
The test program is summarized in Table 1.
Two
kinds of prestressing tendons were used.
One set of
tendons were aramid fiber rods which were braided with
nominal diameters of 6mm and Bmm.
The other was the
prestressing steel wire (SWPD-1) with nominal diameter
of 7mm.
The mechanical properties of prestressing
rods are listed in Table 2.
Three levels of initial
tension of tendon were chosen in the lower tendons of
beam, approximately 40%, 60% and 70% of the tensile
strength of rod.
The initial tension of the upper
tendons of beam was about 60% of that lower tendons.
The square shaped spiral hoop of aramid fiber rod
(diameter; 2mm) was used as the stirrup at an interval
of 35mm.
High early strength portland cement was used.
Mix proportion of concrete is listed in Table 3.
Water-to-cement ratio was 44%, and the maximum size of
coarse aggregate was 13mm.
Compressive strength of
concrete was 59-64MPa and the modulus of elasticity
was 3.3-3.6xl04MPa at the age of 5 days when the
prestress was introduced.
Compressive strength and
modulus of elasticity at the fatigue test (the age of
2 to 4 months) were 64-BOMPa and 3.7-4.0xl04MPa respectively.
In the static and fatigue tests, the beams were
tested under symmetrical two point loading as illustrated in Fig. 1.
The size of beam was 15x15x210cm.
The pre-tensioning system was used as the prestressing
operation. The shear span effective depth ratio (a/d)
was chosen to be 4.0.
Prior to the fatigue test,
three beams (beams name are KB6-7-S, KBB-4-S and PC74-S) were tested under the static loading condition in
order to obtain the cracking load and the ultimate
strength of beam.
In the fatigue test, the sinusoidal loading of
the frequency of 4Hz was applied. The lower load was
4.9kN, and the upper loads were 45-80% of their statical ultimate loads.
In No.1 specimens, the upper
loads were equal to or a little higher than the cracking loads of beams.
In No.2,3,4 specimens, the upper
load ratios were 5-30% higher than that of No.1 specimens. After the given cycles of repetition of loading
; lxl04, 5xlo4, 10xlo4, 20xl04, 40xl04, ---, deflection, strains of concrete and prestressing tendons,
crack widths and curvature were measured at the upper
load.
When the beam did not fail after two million
cycles of loading, the test was followed by the static
loading until the beam failed.
512
Iwamoto et al
RESULTS AND DISCUSSION
Static Test
Table 4 shows the results of the static test.
In
the static test, PC7-4-S beam failed by the crushing
of concrete after yielding of prestressing wire.
Eight flexural cracks were observed in the flexural
span of the beam.
Two types of failure were observed
in the beams with aramid fiber tendons. KB6-7-S beam
failed by the breaking of tendons. Bond cracks along
the tendon in the mid span were observed after the
development of five flexural cracks.
The shear compression failure occurred in KB8-4-S beam after development of diagonal cracks after five flexural cracks.
Fatigue Life and Type of Failure
Typical crack patterns of beams are shown in Fig.
2. Dotted lines in this figure indicate the cracks at
the initial static loading.
The number of cracks in
the fatigue test was less than that in the static
test, with the exception of KB8 series.
It suggests
that the bond performance between tendon and concrete
was deteriorated by the cyclic loading, and the distribution of cracks became worse.
The results of the fatigue test are shown in
Table 5.
In the fatigue test, most of the beams
failed by the fatigue breaking of tendons. The beams
which did not fail at two million cycles of loading,
were tested under the static loading.
Those beams
failed by the breaking of tendons under the static
loading, and at the same time the destruction of cover
concrete in KB8 series and the crushing of concrete in
PC7 series were also observed at the failure of the
beam.
In table 5, the ultimate strength ratio indicates
the ratio of the residual static ultimate strength of
the beam which did not fail after application of two
million loading cycles, to the static ultimate
strength.
The relation between upper load ratio and
number of cycles to failure is shown in Fig. 3.
Specimens in which the upper load was equal to or a
little higher than the cracking load of the beam, i.e.
beam number is l(x-x-1), did not fail at two million
cycles of loading, and showed residual fatigue
strength, exception was beam PC7-6-1.
The residual
ultimate strength ratios of these beams were approximately 1.0.
The upper load ratio is expressed as the percentage to the static ultimate strength of the beam. The
fatigue life of the beam in KB series decreased as
FRP Reinforcement
513
well as that in PC series, when the upper load ratio
became larger.
The fatigue strengths at two million
cycles of loading were about 70% in KB6 series, 65% in
KB8 series and 55% in PC series, when the fatigue
strength was expressed as the upper load ratio.
The
beams in KB series behaved well in fatigue performance, when compared with that in PC series.
In KB6
series, the effect of initial tension of tendon on the
fatigue life was not observed.
In KB8 series in which
the initial tension of tendon was small as 41%, the
fatigue life was a little smaller than that in KB6
series, and was also determined by the upper load
ratio.
Deflection
The relation between load and deflection are
shown in Fig. 4.
The deflection in the fatigue test
was almost similar to that in the static test, although it increased with the number of cycles.
The
difference in load-deflection curve was not remarkable
between the types of prestressing tendons.
The load
at which the inclination in load-deflection curve
changed decreased gradually with the number of cycles,
but its load did not become lower than the calculated
decompression load.
This suggests that the loss of
prestress is small, and the rigidity of the beam does
not decrease largely even under the cyclic loading.
The relation between deflection and number of
cycles expressed in logarithm is shown in Fig. 5.
The
deflection in PC series increased linearly with the
number of cycles until near the failure of beam, but
the large increase in deflection was observed from the
number of cycles of 10,000-500,000 in KB series.
It
suggests that the deterioration of bond between aramid
fiber rod and concrete was more progressive than that
of prestressing wire.
Crack Width
The relation between crack width and number of
cycles of No.1 specimens is shown in Fig. 6, where the
crack width indicates the maximum crack width at the
upper load.
In the reinforced concrete beams in which
aramid fiber rods were used as the longitudinal
bar(not tensioned), the first crack width was very
large (0.3mm-0.5mm), and crack width became remarkably
large value of 2mm at the failure of beam [1].
The
crack width, however, decreased by the introduction of
adequate amount of prestress.
In this experiment, the
first crack width was less than O.lmm in KB series,
and smaller than that in PC series.
Crack width
increased gradually with the number of cycles, but the
width in KB series was smaller than that in PC series,
514
Iwamoto et al
except for KB6-7-l beam.
Compressive Stress of Concrete
Fig. 7 shows
the relation between compressive
strain of concrete and number of cycles.
It is reported that the compression zone of concrete will
become smaller in prestressed concrete beams using
aramid fiber tendons, when compared with prestressed
concrete beams using prestressing wires [3]. But the
clear difference in the compressive strain of concrete
was not observed by the type of prestressing tendons.
The compressive strain of most of beams at the repetition cycles of two million was smaller than lOOOxlQ-6,
when the upper load was almost equal to the cracking
load of beam.
Large increase in compressive strain
was observed as well as the deflection from about the
number of cycles of 10xl04. The maximum strain in the
fatigue test was larger than that in the static test.
For example, the maximum strains in the fatigue test
were about 1.5 times that in the static test in KB6-62 and PC7-6-l beams.
All the beams failed by the breaking of tendons
in the fatigue test, and also in the static test after
two million cycles of loading.
The beam
fatigue
failed only by the crushing of concrete was not observed in this test.
Tensile Stress of Tendon
The relation between the stress range in the rod
measured by the strain gauge and number of cycles to
failure is shown in Fig. 8. The test results of aramid
fiber rods (diameter; 6mm)[2] and prestressing strands
by Toyofuku et al.[4] are also shown in this figure.
In the fatigue test of aramid fiber rods and prestressing strands, the lower loads were 50% and 60% of
tensile strength of rod respectively.
S-N relation of beams in PC series almost coincided with that of prestressing strand alone obtained
by Toyofuku et al., although the stress range calculated from the measured strain varied with repetition
of loading. On the other hand, S-N relation of beams
in KB6 series showed lower values than that of aramid
fiber rods alone.
Factors reducing the strength
include the stress concentration near the crack, and
the wearing between aramid fiber rod and concrete
resulting from the deterioration of bond.
The fatigue
strength of beam with aramid fiber tendons can not be,
therefore, predicted from the fatigue strength of
aramid fiber rod alone.
Further research is necessary
in order to predict the fatigue strength of this kind
of beam.
FRP Reinforcement
515
CONCLUSIONS
The results of this research on the flexural
fatigue behaviour of pretensioned prestressed concrete
beams using aramid fiber tendons can be summarized as
follows :
l) The beams in which the upper load was equal to or
little higher than the cracking load of beam, did not
fail at two million cycles of loading.
The residual
ultimate strength ratio after the fatigue test was
0.9-1.1 of the static strength.
2)
The beams with aramid fiber tendons behaved well
in fatigue performance, comparable to that of the
beams with prestressing wires.
The fatigue strength
at two million cycles of loading was more than 65% of
the static ultimate strength of beam.
3)
The effect of initial tension of tendon on the
fatigue life of the beam was not observed, and the
fatigue life was determined by the upper load.
4)
The remarkable deterioration of bond between
aramid fiber rod and concrete under the cyclic loading
was not observed.
5)
The fatigue strength of the beam with aramid
fiber tendons could not be predicted by the fatigue
strength of aramid fiber rod alone.
REFERENCES
1. K.Iwamoto, K.Sakai, Y.Uchita and T.Kojima, "Flexural Behaviour of Concrete Beams Reinforced by Fiber
Rods", Conference of Kansai Branch JSCE,V-1,1991
2. K.Iwamoto, T.Agawa, Y.Uchita and T.Yoshikawa,
"Various Characteristics of Fiber Rods for Concrete
Reinforcement" ,Conference of Kansai Branch JSCE,V-28,
1990
3.
H.Mikami, M.Kato, T.Tamura and K.Ishibashi,
"Fatigue Behaviour of Concrete Beams Reinforced by
Braid Shaped AFRP Rods",12th Conference of JCI,
pp.1153-1158, 1990
4.
T.Toyofuku and T.Yoneda, "Fatigue Strength of
Steel Prestressing Strand" Concrete Journal, Vol.25,
No.7, pp.37-40, 1987
516
Iwamoto et al
TABLE 1 -TEST PROGRAM
a) Static test
Specimen
KBS-7-S
KB8-4-S
PC7-4-S
Tendon
In it i a I tension
Aramid
Aramid
Steel wire
Pu
*
0. 7 Pu
0. 41 Pu
0. 44 Pu
Tensile strength of rod
b) Fatigue test
Specimen
Tendon
In it i a I
tension
KBS-6-1
-2
-3
Aramid
¢ 6mm
KB6-7-1
-2
-3
*
Prestress
(MPa)
Upper load
ratio (%)
0. 6 Pu
4. 56
60
70
80
Aramid
¢ 6mm
0. 7 Pu
5. 54
70
75
80
KB8-4-1
-2
-3
-4
Aramid
¢ 8mm
0. 41 Pu
5. 54
45
55
65
75
PC7-4-1
-2
Steelwire
0. 44 Pu
5. 54
50
60
PC7-6-1
-2
Steelwire
0.6 Pu
7. 32
60
70
* ; Initial tension of the lower tendon of the beam.
** ; Upper load ratio is expressed as the percentage
**
to
the static ultimate strength of the beam.
TABLE 2 -MECHANICAL PROPERTIES OF RODS
Aramid rod
Diameter (mm)
E. Modulus (MPa)
Tensile strength
Elongation (%)
(kN)
6
6.3X10 4
37.3
2. 4
8
6.5Xl0 4
62. 8
1.9
Steel wire
7
19.6Xl0 4
59. 8
8. 0
Results of tensi Ie-test at authors' laboratory
FRP Reinforcement
517
TABLE 3 -MIX PROPORTION OF CONCRETE
slump
(em)
w/c
(%)
s/a
(%)
water
Unit weight
cement
12± 2
44
49
176
400
(kg/m 3 )
gravel
sand
906
863
Superplasticizer
ex 1. 2%
TABLE 4 - RESULTS OF STATIC TEST
Specimen
Cracking
load
Ultimate
strength
KB6-7-S
KB8-4-S
PC7-4-S
24. 9 kN
26. 7 kN
25. 9 kN
40. 2 kN
62. 8 kN
56. 9 kN
Type of failure
Breaking of tendon
Shear compression failure
Crushing of concrete
TABLE 5 -RESULTS OF FATIGUE TEST
Cracking
load
Type of
failure
Specimen
Upper load
(ratio)
KB6-6-1
KB6-6-2
KB6-6-3
23.9 kN(O. 6)
27. 9 kN (0. 7)
31. 9 kN(O. 8)
23. 7 kN 2, 000, ODD* 36. 3 kN (0. 90)
-24. 1 kN 1. 553, 440
-24. 9 kN
46, 000
T
FT
FT
KB6-7-1
KB6-7-2
KB6-7-3
27. 9 kN (0. 7)
29.9kN(O. 75)
31.9 kN(O. 8)
27. 9 kN 2, 000, 000* 39. 2 kN (0. 98)
-596, 300
27. 7 kN
-27. 1 kN
6. 950
T
FT
FT
KB8-4-1
KB8-4-2
KB8-4-3
KBS-4-4
28. 2kN (0.
34. 5kN (0.
40. 8kN (0.
47.1kN(O.
26.
28.
27.
26.
PC7-4-1
PC7-4-2
PC7-6-1
PC7-6-2
Fatigue
I i fe
Ultimate strength(ratio)
kN 2, 000, 000* 67. 7 kN ( 1. 08)
kN 2, 000, 000* 66. 7 kN (1. 06)
kN 2, 000, 000* 61. 8 kN (0. 98)
-23, 630
kN
T &B
T &B
T &B
FT
29. 4 kN (0. 5)
33. 7 kN(0.6)
29. 2 kN 2, 000, 000* 60. 8 kN (1. 07)
-97,260
27. 5 kN
T &C
FT
34. 1 kN(O. 6)
39.9 kN(O. 7)
31. 6 kN 1, 547, 300
225, 460
31. 8 kN
45)
55)
65)
75)
5
4
0
9
---
FT
FT
* : The beam which did not fail at two mi I I ion cycles.
Type of failure
T : Breaking of tendon
FT
Fatigue breaking of tendon
C Crushing of concrete
B Bonding failure (destruction of cover concrete)
518
Iwamoto et al
500
250
Spiral hoop
( Sgua re shape l
:r-.--+-,~
~:
-
gt
~
L'---T--H~
300
750
(unit: mml
Fig. 1-Test specimen and loading condition
·I
FRP Reinforcement
519
p
Shear compression
failure of concrete
KBB-4-S
(Static test)
p
KB6-7-I
( xl0 4 cycles l
p
KBB-4-2
p
Fatigue breaking
of tendon
PC?-6-1
(N= 1,547,3001
Fig. 2-Typical crack patterns of beam
520
Iwamoto et al
100
90
~
K86-7-3
80
0
-
·-
ol '-
70
ITIIIf
PC7- 6':"2
"0
0
0
~
K86-7-2
KBS-4-4
0
~
K86-6-3
I r
11
60
PC7-4-2
"'
K,Brm-2~ [
Pc~~6~
H-'
n.
n.
11
"KB 6 7
I
l
I l I
I
50
bdeL-
3
K86-6-
)[
KBS-4-2
::J
1
P--I IJ
~-'"Pc7-4-
ll11J4 -I [
40
P-- I I J
10'
Number of
cycles
to
failure
Fig. 3-Relation between upper load ratio and number of cycles to failure
FRP Reinforcement
521
50
KBG-7-S
KBG-6-1
KBG-7-1
40
z
"
0
0
-'
30
20
10
0
10
0
0
15
5
20
10
0
15
Del lection
5
20
10
25
15
lmml
75
60
z
45
"
30
0
0
-'
15
0
0
5
10
0
15
5
0
10
5
15
10
0
Deflection
lmml
Fig. 4-Load-deflection relationship
20
25
522
Iwamoto et al
10
111111_16l~
ifDI
8
'E
f--
.§
6
c:
f--
lr"
0
t
-
~
~~--
--
-- --
0
.
[~r··
~r1-l~J.l
pCi-
II;!J
&-1
--
f""-'
..
i.a
t-""'
fTW
0
10 2
10
.I
~~
--
-·
--
[email protected]
4-1
I"
~
rp
Number of
cycles
Fig. 5-Relation between deflection and number of cycles
I. 0
O•
D•
••••
0.8
'E
.§
;:
'"
KBG-6-1
KBG-7-1
KBB-4-1
PC7-4-I
0.6
~p
"0
~
.X
0
~
0.4
j::•
u
·--
0.2
0
~-
---
I-" 10
1--
kPr
102
---
.,_.,
. !!==:' P'
fl-·1!
I •-
I
....
IIIII
!'"""'
10'
Number of cycles
Fig. 6-Relation between crack width and number of cycles
FRP Reinforcement
523
I 500
:(.
62
I
1200
"'~
KBG-6-Z
u
0
900
0
600
"'
"'>
"'
""'c.
-
-
c
2
J.ll
1
c
u
300
··-It-·
-
lW __
I 1---
I
I
l_-
.
~I£
'!!•
II'.~ 14
I
.. -_..f-+1"7 - I
pc7-G-'
K86-
.. -
[!Y.
j]l=:-
I
I
K88 - 4 - I
.II
'
E
0
u
Ill
I
0
10 2
10
I
104
10 5
Number of cycles
Fig. 7-Relation between compressive strain
of concrete and number of cycles
750
+: KB6
Q:
0
(L
600
..::;;;:-
~
"'
"0
450
- -
2
~
·rnnh ~J ll
~
---- ~ - +
--- ~
PC7tr~~
KBS- 4- 4=tnlfPC7-p
300
~;J r-~mt .:::-r h.-~
K86-6-3
IC7,-6-
2
"'
"'
O:KBB Series_
A.: PC 7 Series
~-
0
"'c
"'
~
Rod olonel2
KB6 Series
KB6 7 3
KB6 7 2
15 0
II
.r- 0-+KBS-4 -2
KB~6-2~
II
in
-KBS-4 -3
-+PC7-4 -I
t1: I I
~-+KB6-6
I I I
... KSS-4 -I
·I
-+K86·7 ·I
0
to•
tO'
Number of cycles lo failure
Fig. 8-Relation between stress range of rods and number of cycles to failure
SP 138-32
Flexural and Shear Behavior
of Prestressed Concrete
Beams Using FRP Rods as
Prestressing Tendons
by A. Yonekura, E. Tazawa,
and H. Nakayama
Synopsis: The flexural and shear behavior of prestressed concrete (PC)
beams using carbon or aramid fiber reinforced plastics (CFRP or AFRP)
rods are experimentally investigated and compared with those using
prestressing steel bars.
Since moduli of elasticity of AFRP and CFRP are about 1/4 and 2/3 of
that of prestressing steel bars, respectively, the deflection of beams
using FRP rods is larger and ultimate flexural and shear strengths of
beams are smaller than those of PC beams using prestressing steel bars
with similar tensile strength of tendons and web reinforcement.
The ultimate flexural and shear strengths and deflection of PC beams
using FRP rods are improved by an increase of prestress in beams.
Keywords: Beams (supports); carbon; deflection; fiber reinforced concretes;
fibers; flexural strength; modulus of elasticity; prestressed concrete;
prestressing steels; shear strength; tensile strength
525
526
Y onekura, Tazawa, and N akayam a
Asuo Yonekura is an associate professor of civil engineering,
Hiroshima University, Japan. He received his M. Sc. from Hiroshima
University in 1967 and his Ph.D. from Tokyo Institute of Technology in
1981. He has published papers on high-strength concrete, silica fume
concrete, flowing concrete, prestressed concrete with FRP rods, etc. He
is a member of ACI.
Ei-ichi Tazawa is a professor of civil engineering, Hiroshima
Universit~ Higashi-hiroshima cit~ Japan.
He received his a Sc. from
the University of Tokyo in 196~ ~Sc. from MIT in 1968 and Ph.~ from
the University of Tokyo in 1978. He had a chief research engineer of
Taisei Corporation in Tokyo. He is a member of ACI.
Hideaki Nakayama is a researcher at Mitsubishi Materials Corporation
in Tokyo. He received his ~Sc. from Hiroshima University in 1992.
INTRODUCTION
New materials such as AFRP, CFRP and GFRP (Glass Fiber Reinforced
Plastics) rods with high tensile strength are being contemplated to be
used as prestressing tendons. Although the flexural properties of PC
beams using FRP rods have been investigated to a some extent (l), the
particular problem is that FRP rods have smaller elongation at rupture
than that of prestressing steel. Therefore, there is possibility of a
brittle failure of PC beams due to rupture of the FRP rods. Although
consideration is given to designing members so that a flexural
compression failure will occur, a distinct design method has not yet
been established. Additionally, the shear behavior of PC beams using FRP
rods as prestressing tendons or stirrups has not been clarified yet.
The objective of this study, therefore, is to examine the flexural
strengl~ flexural failure modes and shear strength of post-tensioned PC
beams using CFRP and AFRP as prestressing tendons. Varying the quantity
of prestressing tendons, prestressing force, type and quantity of axial
reinforcement were used to obtain basic data for the design of PC beams
using FRP rods. In the investigation of shear behavior of PC beams, the
amount of web reinforcement using CFRP or AFRP were varied. The
characteristics of PC beams using CFRP or AFRP rods are compared with
those using prestressing steel.
OUTLINE OF EXPERIMENTS
Materia l_s
The FRP rods used in these experiments are strands made of carbon
fibers (CFRP rods) and rods made of aramid fibers (AFRP rods). The
matrix of FRP are an epoxy resin for CFRP rods and a vinyl ester type
resin for AFRP rods. The mechanical properties of FRP rods are given in
FRP Reinforcement
527
Table 1. Prestressing steel tendons (Type C, No.1 SBPR 110/125) are used
for comparison purposes. Round steel bar of ~9mm (SR-24), deformed bar
D13mm (SD-30), CFRP strands of ~5. ~7. 5 and H2, 5mm and AFRP of H and
~6mm are used for axial
reinforcement, while round steel bar of $6mm
(SR-24) are used for stirrups. In the shear test of PC beams, CFRP
spiral reinforcement of $5mm and AFRP spiral reinforcement of $4mm are
used for stirrups. The mechanical properties of steel bars are given in
Table 2.
The result of bond strength between FRP rods and concrete in the
pull out test by ASTM is shown in Fig. 1. The bond strengths of AFRP
were approximately the same strength as those of deformed steel bar,
because AFRP rods were deformed bar.
High early strength portland cement was used. Mix proportions of
concrete were decided so that the compressive strength of concrete at
transfer prestress (age, 7 days) was about 50 MPa.
Test Variables
The parameters selected for the experimental program were the type
of prestresssing tendons, the type of axial reinforcement, quantities of
prestressing tendons and the amount of initial prestressing force. These
parameters are summerized in Table 3.
Manufacturing and Testing of Specimens
The configuration and dimensions of specimens are shown in Fig. 2.
The specimens consisted of !-shaped cross section beams with spans of
140cm. A prestressing tendon was provided at 6cm from the bottom of the
beam. Axial reinforcements was provided at the four corners as shown in
Fig. 2. Steel stirrups ~6mm were placed for shear reinforcement in case
of flexural tests of beams as shown in Fig. 3. Web reinforcement
consisted of FRP spirals with 8. 5, 11.0 and 13.5cm pitch as shown in
Fig. 3.
Transferring of prestress was done using a 350 KN center hole jack
and grout injected immediately after transfer. The beam loading method
that of two-point loading as shown in Fig. 4. Both flexural and shear
tests were performed. The shear span to effective depth ratio (a/d) was
3. 0 for specimens No. 1-No. 20 and 2. 5 for specimens No. 21-No. 32.
Loading was applied 14 days and prestressing 7 days after construction.
Measurements of strains were made in prestressing tendons, axial and
web reinforcement and concrete, Deflections were measured at the midspan
of beams.
OBSERVED BEHAVIOR AND TEST RESULTS
Calculations of Flexu_ut! Cracking Load and Ultimate Load in Flexure and
shear (2)
528
Y onekura, Tazawa, and Nakayama
The results of flexural and shear tests are provided in Table 4. In
the flexural tests, specimens No. 1-16 and No. 19-20 failed in flexure,
but No. 17 (Cl-CI-So-0) and No.l8(C~~-C-So-O) beams with CFRP rod
without any prestress failed in shear. Flexural failures of beams using
FRP rods were of three types, that is, flexural compression failure,
tension failure) and axial
prestressing tendon rupture(flexural
reinforcement rupture (flexural tension failure). Beams using AFRP rods
and prestressing steel rods had prestressing force monitored by a load
eel I for measurement of effective prestress. llowever, load cells were
not mounted on CFRP rods so that losses of prestress up till loading
were not known. Calculations were made using values immediately after
prestressing. Calculated and measured values were approximately the
same, and it was recognized that even ultimate load can be evaluated by
conventional calculations. Calculations made were as follows.
Flexural cracking load:Pc,
Gce+ab=Mc,/ze -----(1),
Pc,=2Mc,/l
-----(2)
Ultimate strength in flexure:M"(2)
Compressive stress block of concrete at failure is assumed as shown in
Fig. 5, and calculations were made for determining the depth of neutral
axis X by equilibrium Eq. (3). However, in case of specimens using FRP
rods for axial reinforcement, calculations were made without considering
the effect of compressive reinforcement (=Csl. Further, Eq. (4) is for
the case of neutral axis being inside the flange.
C+Cs=Ts+Tr --------------------------------(3)
Mu=Cs(KzX-dsi)+Ts(dsz-KzX)+Tr(dr-KzX) -----(4)
Pu=2Mu/l ---------------------------------(5)
C=K1K3bXa,d -------------------------------(6)
Cs=AsEs(X-dsi) E,u/X -----------------------(7)
Ts=AsEs (dsz-X) E, u/X ----------------------- (8)
Tr=ArEr {E e+ E p ,+ (dr-X) E c u/X} -------------- (9)
where, Mc,:flexural cracking moment, Mu:ultimate moment, !:shear span,
Ze:effective section modulus, ab:modulus of rupture, a,d:compressive
strength of concrete,
Cs:compressive force due to compressive
reinforcement, Ts, T,:tensile forces of tension-side axial reinforcement
and prestressing tendon, respectively, As. Ar:cross sectional areas of
axial reinforcement and prestressing tendon, respectively, Es, Er:modulus
of elasticity of axial reinforcement and prestressing tendon,
respectively, ds1, dsz:distance from compression fiber to centroid of
compression-side and tention-side axial reinforcement, respectively,
dr:distance from compression fiber to centroid of prestressing tendon,
'"":tensile strain of concrete at location of prestressing tendon due to
effective prestressing force.
Ultimate strength in shear:Vu(2)
Vu=2(Vc+Vs+Vr) ----------------------------------(10)
Vc=O. 94{0. 75+1.4(a/d)lPw 1/ 3 G, I/Jd- 1/ 1 bwd --------(11)
Vr=2Mo/a ----------------------------------------(12)
Vs=Aw (Er/Eslf" (sin etcos 8) (z/s) bwd ------------- (13)
Pw=As(Er/Es)/(bwd) ------------------------------(14)
where, Vc:shear strength provided by concrete, V,:shear strength
provided by prestress force, Vs:shear strength provided by shear
0
FRP Reinforcement
529
reinforcement, Mo:decompression moment, Er, Es:moduli of elasticity of
FRP and steel reinforcements, respectively, f":tensile strength of FRP
rod or yield strength of steel bar, Pw:web reinforcement ratio, a:shear
span, bw:web width, d:effective depth, z: lever arm, s:spacing of shear
reinforcement,
Vs is the equation which multiply the equation obtained by truss
analogy by the ratio of modulus of elasticity (Er/Es). Vr is obtained by
the decompression moment.
Cracking and Failure Properties
The cracking properties and strain diagrams of typical beams are
shown in Figs. 6 to 9. Strain distribution diagrams are shown for three
load stages: before occurrence of cracking, after occurrence of
cracking, and just before failure. The values of Cs, Tr and Ts in these
figures indicate the forces carried by compression-side axial steel
reinforcement,
prestressing
tendo~
and
tension-side
axial
reinforcement, respectively. As can be seen in the Figs. 6-8, the
failure patterns differ depending on the variety of the prestressing
tendon. The specimens A,-S-So-2(No. 3) using AFRP rod and c,-S-So-2(No.2)
using CFRP rod showed flexural compression failures,
while
P,-S-S 0 -2(No.l) using steel tendon showed flexural tension failure due
to the yield of steel tendon, with ultimate strengths also differing. In
case of specimen A,-S-So-2(No. 3) with the smallest modulus of
elasticity, tensile strain becomes larger for the same loads as shown in
Fig. 6, and the local ion of the neutral axis determined from the
hypothesis of maintenance of plane sections rises up so that the
compression zone of concrete becomes smaller and compression failure
occurs. This can be observed from the photograph and cracking diagram of
A,-S-So-2(No.3), cracking is shown to have extended inside the flange at
load of 100KN.
The failure pattern of PC beams in flexural strength tests is shown
in Fig. 10. The specimen c,-Co-So-2(No.4) using CFRP strands as axial
reinforcement had its neutral axis rtstng higher than that of
c,-S-So-2(No. 2) using reinforcing steel. However, since the strain at
rupture of CFRP strand was 1.5% and small compared with reinforcing
steel, axial reinforcement was ruptured, with compression fai Jure of
concrete occurring afterward. On the other hand, beam c,-C,-S 0 -2(No. 8)
increased the quantity of axial reinforcement,
since the axial
reinforcement carried tensile force without failure occurring until the
ultimate state, the ultimate strength of Cr-C,-So-2(No. 8) was greatly
increased and flexural compression failure occurred.
Compression failure also occurred in specimen C"-S-So-4(N~ 13) for
which quantities of prestressing tendon and prestressing force were
increased as shown in Fig. 9. The neutral axis obtained from the strain
distribution did not rise as much as that of c,-S-So-2(No. 2). Based on
these facts, the ultimate strength of PC beams using FRP rods showing
flexural compression failure can be increased by increasing the amount
of prestressing force, the quantity of prestressing tendon, and the
quantity of axial reinforcement. This, as shown in Fig. 11, is also
clear from the relationship between ultimate load in frexure and tensile
530
Y onekura, Tazawa, and Nakayama
strength of prestressing tendon. Tensile strength is defined here as the
yield load for prestressing steel bar, and the rupture load for FRP rod.
On the other hand, in case of PC beams using steel prestressing tendons,
the ultimate strength in the flexural compression failure region does
not increase with steel area and the amount of prestressing force. The
influence of the prestressing force level on the ultimate flexural
strength is examined as follows. Figure 12 shows a plot of ultimate
flexural strength of various beams versus prestressing force ratios. The
prestressing force ratio is defined here as the ratios of initial
prestressing force to the rupture load of FRP rod and the yield load of
prestressing steel bar. These data are given in Table 3. The PC beams
using CFRP rod as prestressing tendon show high flexural strength in the
region of flexural compressive failure when prestressing force ratio is
increased. However, when prestressing force ratios are made too high,
the mode of failure changes to flexural tension failure due to rupturing
of prestressing tendons. Therefore, the amount of prestressing force
needs to be carefully selected for high ratio of prestressing force.
The type of shear failure of PC beams with spiral FRP web
reinforcement is caused either by the rupture of spiral reinforcement or
it is a shear compression failure as shown in Figs. 13 and 14. Such
shear failures are very sudden and brittle.
The deflection of PC beams was influenced by several
as described in the following sections.
variables
The load and deflection relationships for PC beams using various
prestressing tendons are plotted in Fig. 15. Deflections after cracking
are larger for beams using FRP rods than those using prestressing steel
bar, especially, for beams with an AFRP rod. In this case, the position
of the neutral axis in the ultimate state becomes higher than those with
a CFRP rod, and the deflection under identical load is slightly larger.
However, as shown in Figs. 16 and 10, when the amount of prestressing
tendon and the prestressing force are increased, the differancc in the
deflections of a beam(No. 15) using a prestressing steel bar and a
beam(No. !6) using an FRP rod is small. This is thought to be because
both show flexural compression failure, and deformation is small in a
beam with large amount of prestressing tendons so that the difference in
deformation according to variety of prestressing tendon becomes smal I.
At !.his time, the local ion of neutral axis for a beam using CFRP rod is
slightly higher than that of a PC beam using prestressing steel bar, but
since the location of the neutral axis is inside lhe web which is of
small width, there is almost no difference in the compressive resultant
at failure, and for this reason also, the difference with the flexural
compression failure strength of PC beam has become small.
FRP Reinforcement
531
The relationships between load and deflection of PC beams using
differrent axial reinforcement are shown in Fig. 17. The tensile
strengths of axial reinforcement of C,-S-So-2(No. 2) using ~9mm steel
round bars and C,-Co-So-2(No. 4) using ~5mm CFRP strand were equal, but
in case of C,-C 0 -So-2(No.4) deflection increased linearly immediately
after initiation of cracking. As for c,-C,-So-2(No. 8) using ~12.5mm CFRP
strand as axial reinforcement, the tensile strength of the reinforcement
was high so that increase in deflection was smal I and ultimate strength
was high as shown in Fig. 10. This indicates that the plastic
deformation properties of beams using CFRP strand as axial reinforcement
become more pronounced than those using reinforcing steel of identical
tensile strength.
Figure 18 shows load and deflection relationship for PC beams with
different prestressing force and identical CFRP rod. The deflection of a
PC beam using CFRP rods for a given load decreases as the prestressing
force is increased. Also, as shown in Table 4, the strength of a PC beam
increases with prestressing force. However, the gradients of the
load-deflection curves after initial cracking were practically unchanged
in all cases. In contrast, the flexural compress ion failure strength was
hardly increased with increased prestressing force for beams using
prestressing steel tendons with large amounts of steel.
As shown in Fig. 19, the shear strengths of PC beams using FRP
tendon and FRP spiral web reinforcement are smaller than those using
steel tendons and steel web reinforcement when the same shear strength
Piovided by web reinforcement. The mode of failure of PC beams using
steel tendons and steel web reinforcements changes from shear failure to
flexural failure by a small increase in web reinforcement resulting a
considerable increase of shear strength as shown in Figs. 13 and 14.
Conversely, the rate of increase of shear strength in PC beams using FRP
rods and FRP spiral web reinforcement is very small as the amount of web
reinforcementis increased. The main reasons to the above are the
deflection properties of beams. When FRP rods are used as prestressing
tendon and web reinforcement, deflect ions become more pronounced than
when using reinforcing steel at the identical load. Thus the shear
failure occurs easily. As shown in Fig. 20, the ultimate shear strengths
of PC beams using FRP rods are increased by increasing prestress,
because the deflection and the principal tensile stress are decreased by
increasing prestress. Therefore, increasing the amount of prestress is a
very effective way of increasing both flexure and shear strength of
prestressed concrete beams.
532
Y onekura, Tazawa, and N akayam a
CONCLUSIONS
The flexural and shear behavior of prestressed concrete beams using
carbon or aramid fiber reinforced plastics (CFRP or AFRP) rods was
experimentally investigated and compared with the behavior of PC beams
using prestressing steel bars.
Based on the data obtained from this research, the following
conclusion may be drawn:
l) In case of PC beams showing flexural compression failure, the
ultimate flexural strength using FRP rods can be increased by increasing
the amount of prestressing force, the quanti t.y of tendon and the
quantity of axial reinforcement. But, the ultimate strength of PC beams
using prestressing steel bars is hardly influenced by these variables.
2) Since moduli of elasticity of AFRP and CFRP are about 1/4 and 2/3 of
that of prestressing steel bars, respectively, the deflection of PC beams
using FRP rods .arc larger than those using prestressing steel bars.
3) In case where flexural compression failure occurs, the amount of
deformation in PC beams using FRP rods at the ultimate stale becomes
greater than when using prestressing steel bars, and flexural cracks in
PC beams rise to the upper flange and prediction of impending failure
becomes possible to an extent.
4) The greater the prestressing force in the same quantity of tendon,
the smaller the deflection of PC beams using FRP rods.
5) When the quantity of tendon is reduced to cause flexural tension
failure of the beam due to rupture of the FRP tendo~ but with
arrangement of ample non prestressed axial reinforcement, it is possible
to prevent sudden rupture of the beam.
6) The ultimate strength in flexure of beams using FRP rods can be
evaluated by conventional calculations.
7) PC beams using FRP rods may be designed for either flexural
compression failure or flexural tension failure.
8) The shear strength of PC beams using FRP rods and FRP web
reinforcement is smaller than those using prestressing steel bars when
the same shear strength is provided by shear reinforcement.
9) The ultimate flexural and shear strength of PC beam using FRP rods
can be improved by increasing prestress force.
REFERENCES
(l) JSCE Subcommittee for Studies on Continuous Fiber, 'Present State of
Technology Concerning Application to the Field of Civil Structures of
Concrete-·Base Composite Material Using Continuous Fibers".
(2) JSCE, 'Revised material of Standard Specification of Concrete',
Library No. 61, 1986. (in Japanese)
Concrete
TABLE 1 -MECHANICAL PROPERTIES OF FRP RODS
:
Diameter Cross sectional
area
(em 2 )
(mm)
*
li.
c
r~r
75
12. 5
17. 8
Rod
25. 0
CF
AF
RP
Rod
~
4.0
0. 0
3xti. 0
I 0. l ( 5. 5)
\lodulus of
elasticity
(1 O'xMPa)
Tensil~
load
(KN)
Tens i Ic Elongation \\'right Cross sectional I
strength
form
1
(MPa)
(%)
(g/m)
I
l. :J2
19600
67000
160000
304500
566000
1940
2200
2110
1970
1950
0. 51
0. 54
0. 54
23200
50900
161100
1900
1900
1900
I. 42
30. 1 (19. 1)
76. 0 (48. 5)
154. 9 (98. 8)
290. 9 {1 ~5. G)
I. 45
I. .'l9
12. 6 ( 8. 2)
28. 3 (18. 4)
84. fi (55. 1)
1.
10
Figures in ( ) indicatr cross-sectional are:; of fiber only
1.5
1.6
1.5
1.5
1.5
7-filament
7-f i lament
153
7-f i lament
294 7-fllament
596 19- f i Iamen t
24
60
3. G
3. 6
11
3. 6
123
18
strand I
strand
strand I
strand
strand
Deformed bar
Deformed bar
Drformed bar
1
1
"rj
::0
"'C
::0
(I)
I
I
I
~
""!
f6
3
(I)
....::s
VI
(.H
(.H
TABLE 2 - MECHANICAL PROPERTIES OF STEEL BARS
~-- --- 11 ~~ame~;r~~~os;_---1 Mod:~~;~fl-;,~-~-d- ;;;eld---~E~o~gati-on- 11
sectional ; elasticity! load
, area
·
I
,
I
r -- ~-1
1
i
Tl __
Prestressi ng s Lee I
bar Type C 1
strength at
Fa i I ure
13
132.7
J
~- 0
1184500
1
17
i
: 220. 2
I
1
•
2. 0
II
270800
1
I
1390
1
10
1
1230
i
I
12
I
1
t
L
,"''"'"'~+- ,: ~. ~;: : ... ::.. i.::~:: I~:~- , ::
,-
--~--
+'
: Round bar i
SR--24
21
I
324. 7
i
2 0
23. 6
I
2. 0
! 12210H- 1300
------1-------r---- --,---6
'
· bar SD-30 !
j
~m~--~~- ~~~~)- ---~~~~x~Pa) -+-~~~-)-f ~~ -t-~~~ ~~
J
No. I SHPR
J 10/125
1
1
l________ ,__j_ ______
1
i
:
-~-----------~---
I
1
I
_______ l __
6700
I
I
I
I
1'
9
---------t-------1
i
183
!
1
i
~--·--~-··----1__·-----~~'
FRP Reinforcement
535
TABLE 3 - DETAILS OF TEST VARIABLES
I
u
Web reinAxial
reinforce- for cement
(PI tch:mm)
ment
Specimen
Tendon
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
P,-s -s.-2
c,-s -s.-2
A,-S -S.-2
c,-c.-s.-2
A,-A.-S.-2
c,-s -s.-3
A,-S -S.-3
c,-c,-s.-2
A,-A -S.-2
Pn-S -s.-2
Cn-S -S.-2
Pn-S -S.-4
Cn-S -S.-4
Cn-S -s.-5
Pw-S,-S.-6
Cm-S,-S.-6
c.-c.-s.-o
Cn-C,-S.-0
Cn-C.-S.-2
Cn-C,-S.-4
m.o
m.5
3d6
~12. 5
3d6
m.s
3d6
m.5
3x•6
H7. o
m.8
H7. o
~17. 8
.17.8
m.o
m.o
m.5
H1. 8
m.8
m.8
21
22
23
24
25
26
27
28
29
30
31
32
c,-c -c.-1
A,-fl -fl.-1
c,-c -c.-2
c,-c -c.-2
c,-c -c.-2
(1,-A -A.-2
(1,-fl-flo. -2
A,-A-A; -2
P,-s,-s.-2
P,-S,-S.-2
c,-c -c.-3
(1,-fl -A.-3
.12.5
3x•6
H2. 5
m.5
m.5
3x•6
3x•6
3x•6
H3
H3
m.s
4-H. 5
4-·6
4-H. 5
4-H-5
4-H. 5
4-·6
4-·6
4-·6
4-013
4-013
4-H. 5
3x~6
4-~6
No.
1: Flexural behavior test,
* Figures in () indicate
load or rupture load.
Specimen names indicate
Variety of prestressing
tendon
A, :AFRP 3x.6
cI : CFRP • I 2. 5
Cu:CFRP .17. 8
Cw:CFRP .25.0
P,:steel tendon ~13
Pn:
·~~ 0
Pru:
•21.0
4-·9
4-·9
4-·9
4-t5
4-~4
4-·9
4-~9
4-m.s
4-~6
4-·9
4-·9
4-·9
4-·9
4-·9
4-·9
4-·9
4-.12.5
4-.12.5
4-H2. 5
4-H2. 5
Prestress- Prestress
lng force
(KN)
(MPa)
*
~6
~6
~6
( 85)
( 85)
( 85)
.6 ( 85)
•6 < 85)
~6 ( 85)
.6 ( 85)
.6 ( 45)
.6 ( 45)
~6 ( 85)
.6 ( 85)
.6 ( 85)
.6 ( 85)
.6 ( 45)
.6 ( 40)
.6 ( 40)
.6 ( 40)
.6 ( 40)
.6 ( 45)
.6 ( 40)
82 (43%)
82 (48%)
82 (51%)
82 (48%)
82 (51%)
100 (59%)
100 (62%)
82 (48%)
82 (51%)
82 (30%)
82 (27%)
164 (61%)
164 (54%)
228 (75%)
246 (54%)
246 (44%)
0 ( 0%)
0 ( 0%)
82 (27%)
164 (54%
7. 2
7. 2
7. 2
7.2
7.2
8. 8
8.8
7. 2
7.2
7.2
7.2
14.4
14.4
20.0
21.6
21. 6
0
0
7.2
14.4
•5 Oto>
56 (33%)
56 (35%)
82 (48%)
82 (48%)
82 (48%)
82 (51%)
82 (51%)
82 (51%)
82 (43%)
82 (43%)
100 (59%)
100 (62%)
4. 9
H(IIO)
•5 < 85)
•5 Oto>
•5 035)
.4 ( 85)
H(llO)
H (135)
•6 OlOl
~6 (135)
•5 Otol
.4 (IJO)
4. 9
7. 2
7.2
7. 2
7. 2
7. 2
7. 2
7.2
7.2
8. 8
8. 8
H; Shear behavior test
P./Pv. where, P, is prestressing force and Pv is yield
Variety of axial
Variety of web
Prestressing
reinforcement
reinforcement
force
Ao :AFRP H
A:AFRP spiral .4
0: 0 KN
A :AFRP .6
C:CFRP spiral ~5
I: 56
S:steel round
2: 82
Co :CFRP
C :CFRP H. 5
s turrup •6
3:100
C, :CFRP U2. 5
4:164
S :steel round bar •9
5:228
S, :steel deformed
6:246
S,:steel deformed bar 013
•s
TABLE 4- DETAILS OF SPECIMENS AND TEST RESULTS
..--Compressive
strength of
concrete
(MPa) #
FI ex. Cracking load (KN)
Exp. /Calc.
Ex perlment
Pc ,-1
Pc,-1/Pc,-2
Calculation
Pc,-2
Ultimate load (KN)
Experiment
Pu-1
Calculation
Pu-2
Exp. /Cal c.
Failure
mode
Pu-1/Pu-2
*
---
I
2
3
4
5
6
7
8
9
10
II
12
13
14
15
16
17
18
19
20
r,-s -s.-2
c,-s -s.-2
A,-S -S.-2
c ,-c.-s.-2
A,-A.-S,-2
c,-s -s.-3
A,-S -S.-3
C,-C,-S,-2
A,-A -S.-2
Pn-S -S,-2
Cn-S -S,-2
Pn-S -S,-4
Cn-S -S.-4
Cn-S -S.-5
r.-s ,-s.-6
c,"-s, -s.-6
c ,-c ,-s.-o
Cn-C,-S,-0
c,-c ,-s,-2
c,-c ,-s,-4
55.9
61. 6
55.4
55. 5
57. 0
62. 3
54. 4
59.6
52. 8
52. 9
61.2
61. 9
61. 8
60. 9
59. 8
64. 8
62. 0
59. 8
55. I
64. 7
(53. 8)
(52. 0)
(56. 0)
(50. 3)
(51. 3)
(54. 5)
(46. I)
(52. 7)
(45. 4)
(48. 2)
(58. 5)
(48. I)
(53. 4)
(53. 0)
(53. 5)
(54. 3)
21
22
23
24
25
26
27
28
29
30
31
32
c,-c -c,-1
A,-A -A,-1
c,-c -c,-2
c ,-c -c,-2
c,-c -c,-2
A,-A -A,-2
A,-A-A," -2
A,-A-A; -2
r ,-s, -s,-2
r ,-s, -s,-2
c,-c -c,-3
A,-A -A,-3
56. 7
62. I
65.5
54.9
54.7
63.5
69. 9
52.0
58. 5
64.3
64.2
61. 8
(49. 5)
(58. 6)
((-
)
)
(55. 2)
(52. 8)
-
( -
)
(51.
(54.
(57.
(65.
(41.
(54.
(56.
(51.
(54.
6)
5)
7)
2)
3)
9)
0)
7)
2)
---
59. 0
55. 0
54. 5
59. 0
58. 0
72. 0
72.0
60. 0
55. 0
55. 0
63. 0
100.0
92. 0
II 0. 0
140. 0
130. 0
58. 7
55.5
50. 9
55. 2
59.3
68.4
63.4
57. 0
56.6
61. 4
61. 8
90. 5
93. 4
112.5
132. 8
121. 2
I. 01
0. 99
I. 07
I. 07
0. 98
I. 05
I. 14
I. 05
0.97
0.90
I. 02
I. I 0
0. 99
0. 98
I. 05
I. 07
20.0
59. 0
78.0
25.9
57. I
84. 6
0. 77
I. 03
0. 92
167.0
130. 0
126. 0
117.5
137. 5
150.0
160. 5
233. 0
144. 0
201. 0
195.5
208. 0
216. 0
200. 0
263. 0
263. 0
196. 0
207. 0
227. 0
251. 0
55. 0
53. 0
65. 0
63. 0
63. 0
68. 0
68. 0
62.0
70. 0
60. 0
70. 0
70. 0
56.2
51. 7
70. 7
72.0
72. 4
64.5
69. 6
65. 6
73. 0
71. 8
78. 3
75.2
0. 98
I. 03
0. 92
0. 88
0. 87
0. 95
0. 97
0. 95
0. 96
0. 84
0. 89
0. 93
178. 6
145. 0
204. 0
183. 6
212. 0
173. 0
161. 0
151. 6
223. 6
215.0
206. 0
163. 0
-
-
-
137.2
130. 2
120. 6
126. 0
133. 3
132. 0
129.4
219. 9
119. 9
184. 9
195. 7
186. 4
218. 4
219. 0
276.4
270. 2
218. 7
206. 5
221. 6
244. 7
1.22
I. 00
I. 04
0. 93
I. 03
I. 14
I. 24
I. 06
I. 20
I. 09
I. 00
I. 12
0. 99
0. 91
0. 95
0. 97
0. 90
I. 00
I. 02
I. 03
FT
FC
FC
RR
RR
TR
TF
FC
FC
FT
FC
FC
FC
TR
FC
FC
sc
sc
FC
FC
148.4
97. 8
197. 8
172.0
162.6
137. 2
127. 8
115. 4
177. 8
160. 6
194.0
132. 4
1.20
I. 48
I. 03
I. 07
I. 30
1.26
I. 26
I. 31
1.25
I. 34
I. 06
1.23
SR
sc
SR
sc
SR
FC
sc
sc
FT
sc
TR
SR
!:Flexural behavior tests, D;Shear behavior tests, #:Compressive strength at loading (compressive strength at prestressing),
FC:Fiexural compression failure, FT:Fiexural tension failure, TR:Tendon failure, RR:Reinforcement rupture, SC:Shear
compression failure, SR:Spiral reinforcement rupture
*
15
ce
0.....
--- 10
~
():CFRP strand ¢5
strand ¢7. 5
4t:CFRP strand ¢12.5
..0..: AFRP rod ~4
'V:AFRP rod ~6
[]:round steel bar ~6
(>:deformed steel bar D!O
....c::
~:CFRP
...........
bD
~
<J..)
~
...........
en
-c:::1
'T1
5
:::0
~
""C
0
:::0
en
o::l
s·
0'
'"'I
8
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
The amount of sliding
Fig. 1-The bond strength of FRP rods and steel bars in the pull-out test by ASTM method
3en
::s
.......
VI
w
-:J
538
Yonekura, Tazawa, and Nakayam a
•
Axial
reioforccaeot
I
70
.l 40
(mm)
150
Fig. 2-Details of a representative test specimen
Flexural test
,
Spr c i mens (No. 1-7,
pitch 8.5ca
~~===:=::;:::::;:::::::;;:::::;::::;::::;:==::;-,
6: pi tcb a. Sc•
ri~I"Tirtt:=:::::;::t: ••6:pitcb
5.5c•
No. 10-13) j
Anchorage
reinforce•ent
I
'
!
.
220
l
I
480
1
pitch 4
Specimens (No. 8 , 9)
I
Specimens (No. 14-20) I
:
I
Shear test
~}t-
...
Specimens with FRP rod (Refer
400
•6:pitcb 4.0c•
Anchorage
reinforccacnt
Table 3)
200
Fig. 3-Arrangement of stirrups
(mm)
FRP Reinforcement
~ Displacc•cnt
•ctcr
I
~~~~~~~~~0
~-.------...------'1]~
'-------.-.48-..-----,.
(em)
Fig. 4-Loading arrangement
'cu
-
./
...:
X£ cu
,
...:
..
II)
c:>.
"C
"C
...
•c
t
~
--
c
I
(ds 2 -X)X
l
I
Tp
Ts
cu
X
Fig. 5-Strain and stress distribution
539
VI
A,-S -So-2(No.3)
~
Cr-S -S 0 -2 (No.2)
~
0
::s
0
3X•6
f(
AFR~
4-•~
Prestress a=7. 2MPa
:.
·f¥"~\
; -------- ~ ~
(Fiexual compresion failure) Pc,= 54. 5KN
Pu =126
100KN
125KN
40KN
~
9l
149•
190•
;::
atl2. 5, 4-•9. Prestress a=7. 2MPa
f? f ,{. ""ft\.}
(Flexural compression failure) Pc,= 55KN
Pu =130
40KN
100KN
128KN
1
192•J1046-J2:19Q.
158•
56•
~749-<>
____
...J
Cs= 3. 66KN Cs=12. 79KN Cs= 14. 20KN
r.=76. 10
r.=93. 40
r.=125. 10
Ts= 0
Ts=34.40
Ts= 34. 40
Fig. 6-Cracking pattern and strain distribution at
beam midspan
59134-
179!1#
2036- ..
:
5451<
:
959-
54JJ,.
192-
Cs= 3. 30KN Cs= 13.30KN Cs=23.40KN
r.=so.so
r.=I02. 90
r.=149. oo
Ts= 0
Ts= 34.40
Ts= 34.40
Fig. ?-Cracking pattern and strain distribution at
beam midspan
'"1
Jl'
~
~
~
Jl'
~
::s
0.
z
~
~
~
3~
l' r-S -So-2 (No. 1)
13.0,
4-~9.
Cu-S -So-4(No.l3)
Prestress a=7.2MPa
17. 8,
#ktt;%0~.}
(Flexural tension failure) Pc,=59KN
Pu =167
lOOKN
40KN
160KN
r
79
13&19JJ.
199-ifgog,.
164J<
95Js.
15781t
~
551Ju
if
4-~9.
Prestress a=l4. 4MPa
~
(Flexural compression failure) Pc,=92KN
Pu =216
80KN
l60KN
200KN
1-fj
17911891•
~
~
~
(b
s·
3074·
Cs= 4. OOKN Cs= l3.40KN Cs= 21. 70KN
Tr=82. 20
Tr=105. 90
Tr=155. 00
Ts= 0
Ts= 34.40
Ts= 34.40
Fig. 8-Cracking pattern and strain distribution
at beam midspan
0'
Cs= 7. 70KN Cs= 21. 20KN
Tr=175. 00
Tr=209. 00
Ts= 0
Ts= 34.40
Cs= 29. 20KN
Tr=264. 00
Ts= 34.40
Fig. 9-Cracking pattern and strain distribution
at beam midspan
§
3(b
::s
......
Vl
~
~
No.4 CT-Co-So-2 $12.5 CFRP, 4-~5.
Prestress ap=7. 2MPa, Pc,=59. OKN,
Pu=117.5KN, Failur~ mode=RR
VI
~
Crack detai I of No. 16
N
~
0
::s
(1)
~
'"i
.P'
~
~
N
~
.P'
C"-C,-So-2 $17.8 CFRP,
a•. =7. 2MPa. Pc ,=59. OKN,
.,....,~
~
FC
::s
0..
z
~
~
~
3
~
Fig. 10-The failure pattern of PC beams in flexure strength tests
FRP Reinforcement
~~--__....D<~C44%)
•<so%>A<54%)
<!.)
(45%)~(30%) .A 05%)
;:;>< 200
<!.)
(75%)Afi-- /O:Flexural compression failure
/.,
<..
(Prestressing steel bar)
L:s_ (51%)
e:Flexural tens ion failure
(Prestressing steel bar)
~:Flexural compression failure
(CFRP rod)
~:Flexural tension failure
(CFRP rod)
;igures in () indicate P,/Pu or P,/Py
c
-o
~
100
200
400
Tensile strength of Tendon (KN)
0
600
Fig. 11-Tensile strength (rupture load or yield load)
of prestressing tendons versus ultimate load
-
0
c
I
-
C1
-
cS-
(FC)
;~
:0 FC:flcxural
co•pressive failure
FT:Flcxural tensile failure
P • - S- :V SC:Sbear co•prcssivc failure
C 1 - S- :Q Tll:Tcodoo rupture
0
0.2
0.4
0.6
8
Prestressing force ratio
Fig. 12-Prestressing force ratio versus ultimate load
543
544
Y onekura, Tazawa, and N akayam a
No.2!
C,-C-Co-1 ~12.5 CFRP, 4-n.5, ~5 Spiral(pitch=l!Omm),
ar=4. 9MPa, Pc,=55. OKN, Pu=l78. 6KN. Failure mode=SR
C,-C-Co-2 ~12. 5 CFRP, ~-F. 5, ~5 Spiral (85),
ar=7. 2MPa, Pc,=65. OK~. Pu=204. OKN, SR
No.24
C,-C-C 0 -2 H2.5CFRP, 4-~7.5, ~5Spiral(IIO),
ar=7. 2MPa, Pc 1 =63. OKN, Pu=183. 6KN, SC
No.25
C,-C-Co-2 H2.5 CFRP, 4-~7.5, ~5 Spiral(J:l:,),
ar=7. 2MPa, Pc,=63. OKN, Pu=212. OKN, SR
No.29
P,-S,-S 0 -2 H3Steel, 4-Dl3, ~6Stirrup(ll0),
ar=7. 2MPa, Pc,=70. OKN, Pu=223. 6KN, FT
Fig. 13-The failure pattern of PC beams
in shear strength tests (CFRP and steel)
FRP Reinforcement
No.22
A,-A-Ao-1 3X~6 AFRP, 4-~6. ~4 Spiral(pitch=llOmm),
Or=4.9MPa, Pc,=53.0KN, Pu=145. OKN, Failure mode=SC
No. 27 A,-A-Ao. -2 3X~6 AFRP, 4-~6, ~4 Spiral (110),
Or=7. 2MPa, Pc,=68.0KN, Pu=161KN, SC
-
No.32
A,-A-Ao-3 3X~6 AFRP, 4-~6, H Spirai(IIO),
or=8. 8Mpa, Pc,=70. OK~ Pu=l63. OKN, SR
No. 30 P ,-s ,-So· -2 H3 Stee I, 4-Dl3, ~6 Stirrup (135),
Or=7. 2MPa, Pc,=60.0KN, Pu=215. OKN, SC
Fig. 14-The failure pattern of PC beams
in shear strength tests (AFRP and steel)
545
546
Yonekura, Tazawa, and Nakayama
~:Ar-S-So-2(No. 3)
CJ:Pr-S-So-2(No. 1)
():Cr-S-So-2(No. 2)
Py =llOKN
I
Pcr=59KN
'"0100
~
0
~
5
0
10
20
15
D e f 1 e c t
1
o n
(mm)
Fig. 15-Load versus deflection for specimens No. 1, 2, 3
'"d
CJ:Pu-S-So-4(No. 12)
():Cu-S-So-4(No. 13)
II:Pm-S,-So-6(No. 15)
tt:Cm-S,-So-6(No. 16)
ro lOO
0
~
0
5
D
e
10
f 1 e c t i o n
5
(mm)
Fig. 16-Load versus deflection for specimens No. 12, 13, 15, 16
FRP Reinforcement
547
~
Cr-Cr-So-2(No. 8)
Cr-S -So-2 (No.2)
D--O C1-C 0 -So-2(No.4)
...-200
~
~
0
......l
5
10
D e f 1 e c t
0
15
1
o n
(mm)
Fig. 17-Load versus deflection for specimens No. 2, 4, 8
200
~100
0---0 Cu-C-So-0 (No. 18)
Dr-----6. Cu-C 1-S 0 -2(No.l9)
[}---{] Cn-Cr-So-4(No. 20)
0
......l
0
5
D e f 1 e c t
10
1
5
o n
(mm)
Fig. 18-Load versus deflection for specimens No. 18, 19, 20
548
Y onekura, Tazawa, and Nakayama
SCF:shear compression failure
SR: spiral rupture
/',;A,-A-Ao( )-2
MCF:flexural compression failure
D:P,-S-S 0 ( )-2
MTF:flexural tension failure
120 s:pitchs of stirrup or spiral
reinforcement
O:C,-C-Co( )-2
>
.....
ro
QsR
Cl)
~
c:
(S=J3. 5 CD)
r...---SR'
__.--0 (S= 8. 5"')
- SCr
(S•!J.O ~ 0(S=JI.Ocro)
1001
SCr
...c::
....,
80
bll
_ _D.wcr
,
/
- - -DSCF
lS. SCr
(S=J I. 0 au)
(S=J3. SOD)
c:
Cl)
.....
....,
cr.>
Cl)
(S= 8. 5 CD)
60~~r--~------~------~--------~-~---
20
40
60
80 (KN)
Shear strength provided by shear reinforcement
Fig. 19-Ultimate shear strength versus shear
strength provided by shear reinforcement
Q TR
0 sc
QS R
80
....,
...c::
S C F
6
SCF:shear compression failure
SR: spiral rupture
TR: tendon rupture
F
6
SR
bll
c:
Cl)
o-o: Cr-C-Co (11. 0) ,_
6 SCF
70
f::s---6 :
A,-A-Ao (11. 0) ·"' T '/
o~r-~----~----~----~--~------60
70
80
90
100
.....
_.....,
Effective prestress force (KN)
Fig. 20-Ultimate shear strength versus effective prestressing force
SP 138-33
Design Concept for Concrete
Members Using Continuous
Fiber Reinforcing Materials
by H. Okamura, Y. Kakuta,
T. Uornoto, and H. Mutsuyoshi
Synopsis: As fibers made of materials such as glass, carbon, aramid and vi nylon have very high resistance to corrosion, more and more attempts are being
made to utilize continuous fiber reinforcing materials (CFRM) in reinforced and
prestressed concrete structures instead of ordinary steel. However, CFRM are
composite materials composed of millions of fibers and binding material, and
have little plastic behavior. The mechanical behavior of reinforced concrete
using CFRM is quite different from conventional reinforced concrete. As of
now, there is no general agreement relating to the methodology to be adopted
in design or testing methods of such fibers. Realizing this problem, the Concrete Committee of the JSCE (Japan Society of Civil Engineers) organized a
subcommittee on CFRM in 1989. The following results have been published
as the committee report in 1992:
1. Design Concept for Concrete Members Using CFRM,
2. Test Methods for Durability of CFRM,
3. Concept for Durability of CFRM,
4. A State-of-the-Art Report on CFRM for Concrete Structures.
This paper describes the design concept for concrete members using CFRM.
Keywords: Composite materials; durability; fiber reinforced plastics; fibers;
flexural strength; limit state design; prestressed concrete; reinforcing
materials; rupture; structural design
549
550
Okamura et al
Fellow ACI member Hajime Okamura is a Professor of Civil Engineering at
University of Tokyo, Tokyo, Japan. His research interests include fatigue and
shear of reinforced concrete members, durability design of reinforced concrete
structures, application of FEM to reinforced concrete, and development of high
performance concrete. He was the chairman of a subcommittee on CFRM.
ACI member Yoshio Kakuta is a Professor of Civil Engineering at Hokkaido
University, Sapporo, Japan. His research interests include cracking, shear and
fatigue of reinforced concrete members, partial prestressed concrete, and application of continuous fiber reinforcing materials to concrete structures.
ACI member Taketo Uomoto is a Professor at the Institute of Industrial Science,
University of Tokyo, Tokyo, Japan. His current research interests include nondestructive test methods, utilization of industrial wastes, corrosion and corrosion
protection of RC structures. He is a member of ACI Committee 440.
ACI member Hiroshi Mutsuyoshi is an Associate Professor of Civil Engineering
at Saitama University, Urawa, Saitama, Japan. His current research interests
include application of FRP to concrete structures, external PC, and seismic
design of R/C structures. He is a member of ACI Committee 440.
INTRODUCTION
In recent years, attention has been focused on durability and other features of CFRM, and technical research and development have been done to
use them as main reinforcements for concrete in place of reinforcing steel bars,
or prestressing steel tendons. This report describes the design concept using
CFRM as reinforcement or tendons in concrete structures. Items not discussed
here are considered to be the same as steel reinforcement in reinforced concrete or prestressed concrete, and JSCE Standard Specification for Design and
Constmction of Concrete Structures (3) shall be applied.
Test methods for determining characteristic values of CFRM have not
yet been established. The mechanical properties of reinforced concrete members
and prestressed concrete members using CFRM as reinforcements are still under
study, and there are many points which are not yet clear. However, from a
practical point of view, recommended ranges of partial safety factors for design
have been provided here tentatively as material factors and member factors for
concrete members using CFRM. These factors will be modified to reliable ones
by further studies and accumulation of test data.
One of the problems in using CFRM is small ductility of members. Not
all stmctures, members, or parts necessarily require large ductiity, but when
using CFRM for structures which had conventionally been designed expecting ductility, it is necessary to thoroughly consider this point. In this report,
designing with high safety factor is recommended.
FRP Reinforcement
551
TYPES OF CONTINUOUS FIBER REINFORCING MATERIALS
The CFRM considered here are reinforcing materials of fiber reinforced
plastic, developed as a substitute for concrete reinforcing steel. Also the basic data concerning characteristic values required for designing are available.
This type of CFRM has different physical properties depending on the material
characteristics of the fiber used, fiber content, and cross-sectional and surface
configurations of the reinforcing material. The fibers generally used are carbon,
aramid, glass, and vinylon. As for binding material, epoxy resin and vinyl
ester resin are mostly used. The configurations of reinforcing material may be
classified to one-dimensional bars, grids, and three-dimensional fabrics. Onedimensional bars include continuous fiber bars in straight form, strand form,
or braid form, and with fiber wound around the surface, or with silica sand
adhering to the surface. The physical characteristics of a CFRM will differ
depending on its configuration, the fiber used, binding material, manufacturing
method, etc., and are thus difficult to determine. Therefore, it is necessary
to obtain characteristic values required for designing by tests. When reliable
characteristic values confirmed by tests are furnished by the manufacturer, it is
permissible to use those values. And in case of a CFRM indicating a specific
physical property, it is necessary to check its influence and confirm its safety.
Fig.l shows the classification of CFRM.
DESIGN CONCEPT
With reinforced concrete or prestressed concrete using steel as main
reinforcement, it is common for the ultimate failure mode to be that of flexural
failure preceding shear failure. Furthermore, it may be said that design for
flexural failure is done so that tensile steel failure due to flexure will occur in
consideration of safety of the structure and of using materials economically and
rationally. However, CFRM are brittle materials with no plastic range, while
concrete is generally a brittle material also.
To briefly comment on shear and flexural failure of reinforced concrete,
it is conceivable for shear failure to be either failure due to crushing of concrete
at the web or compression zone, or failure due to breaking of CFRM used as
shear reinforcement, with both being brittle failures. On the other hand, with
flexural failure, a broad division may be made into failure due to failure of the
CFRM (hereafter referred to as fiber rupture-type flexural failure) and failure
due to failure of the compression zone of concrete (hereafter referred to as
flexural compressive failure), and both failure modes may be said to be brittle
types. Consequently, the problem will be a kind of ultimate failure mode which
will be desirable in case of using CFRM as main reinforcement.
To deal with the problem, the following concepts are adopted to concrete design using CFRM as reinforcements:
1. To increase the safety factor for materials, a sufficient material factor
Tm
shall be used. The design strength fd (or design ultimate load) of CFRM
is to be the characteristic strength fk divided by the material factor Tm as
follows:
552
Okamura et al
2. To obtain sufficient safety of concrete members, an appropriate member
factor Tb shall be used according to the type of failure. The design capacity
of member cross section Rd shall be determined by dividing computed
capacity of member cross section R by a member factor Tb as
DESIGN VALUES OF MATERIALS
Strength
The next following items concerning strength should be considered:
1. CFRM are composite materials consisting of continuous fibers and bind-
ing material. Consequently, when a force acts on CFRM, seen from a
microscopic viewpoint, the local stresses acting on the individual fibers
and the binder such as of resin will be different. However, when the
strength of CFRM as reinforcement of concrete is considered, it is more
rational and convenient to handle the CFRM as a single material. Therefore, the average strength (maximum load/effective cross-sectional area)
of the system as a whole is to be used as the strength of the CFRM.
In case the effective cross-sectional area of the CFRM is unknown, the
maximum carrying load (ultimate load) may be used in lieu of strength.
2. CFRM are generally used as tensile reinforcements. Compressive strength
and shear strength when used as compression or shear reinforcement are
not to be considered.
3. Compared to steel, the variation of tensile strength using identical CFRM
has been recognized to be greater. The variation depends on types, configurations, etc. of the fibers and binders and also the test length and
method of anchoring at the time of testing. Therefore, upon assuming
variation in test values, the characteristic value (J'k) of CFRM is to be a
value guaranteed that the majority of test values will not be lower than
it.
4. When making continuous fibers into a bent shape or when arranging
straight CFRM in curved form, the tensile strength decreases compared to
the uniaxial tensile strength. It has been ascertained through experiments
that the rate of decrease differs according to the ratio of the curvature
radius of the bent shape, or the curved portion, to the diameter of the
CFRM. Therefore, the material strength when continuous fibers are bent
or when CFRM are arranged in curved form, is to be determined based
on the results of appropriate tests.
Design Strength
The material factor T m is to be detennined considering the influences of
the following factors and deviations of experimental data used in tension tests:
damage to CFRM which may occur during construction and transportation,
differences in material characteristics between experimental specimens and the
structure, the effect of material characteristic on the lirnit state, temperature
FRP Reinforcement
553
during use, environmental conditions, etc. Although r m of steel reinforcement
is 1.0, in case of CFRM, the value shall be larger than that of steel and, in
general, Tm may be taken to be between 1.0 to 1.3.
Strain Limit
Unlike steel, CFRM do not have inelastic range, and failure is brittle.
Therefore, in order to consider safety of a concrete member reinforced with
CFRM against failure, it is advisable to newly add a limit strain (elongation)
to design values. The methods for determining characteristic value and design
value of limit strain, and design fatigue value of CFRM are to follow the
methods for strength.
STRUCTURAL ANALYSIS
The following points should be taken into account in conducting analysis
of structures with members using CFRM:
1. Unlike steel, CFRM do not show plastic deformation, and its failure is
brittle. In concrete also, except for when special restraining reinforcement
is provided, it cannot generally be accepted that there will be plastic deformation. Consequently, the mechanical properties of concrete members
reinforced with CFRM are elastic, and it may be considered that there
is hardly any plastic deformation property demonstrated. Therefore, in
calculation of section force for examining the ultimate limit state, it is
advisable not to consider redistribution of bending moments due to plastic
deformation. However, the effect of reduction in rigidity of a member
due to cracks is to be considered.
2. In case when the influence of temperature is conceivable due to difference
in coefficients of thermal expansion of CFRM and concrete, structural
analysis considering thermal stresses is to be performed as necessary.
CONSIDERATION OF ULTIMATE LIMIT STATE
Bending Moment and Axial Force
Fiber Rupture-Type Flexural Failure - In case steel reinforcements are
arranged in multiple layers, it is permissible to evaluate the stress at the centroid
of the reinforcements, but in case of CFRM, fiber rupture-type failure will occur
when the outermost main reinforcement has reached failure strain as shown in
Fig. 2. Also, when different types of CFRM have been used in the same
cross section, or when reinforcing materials with and without bond are used
in combination, it will be necessary to calculate ultimate load taking this into
consideration.
Local Stress on Tensile Reinforcement at Crack Location - Shear force
(dowel action) will occur in tensile reinforcement at a crack located inside a
shear span. When there is a possibility for ultimate load to be lowered by this
dowel action, an appropriate safety factor must be used.
554
Okamura et al
Sustained Load (Static Fatigue) - Regarding static fatigue failure of
CFRM when used as prestressing material in prestressed concrete, it will suffice
to specify the limit value of stress in the prestressing material within a range
which will not be problematic. It will be necessary to examine whether the
tensile strength of the prestressing material after relaxation decreases below the
initial strength.
Compressive Failure in Flexure - It appears that ultimate load can be
estimated with accuracy even when current stress-strain relationship of concrete
is applied. The idea is the same as in the design method for reinforced concrete
or prestressed concrete using steel.
Member Factor - Although the member factor Tb for bending moment
and axial force is 1.15 in case of steel, the value for CFRM may generally be
taken as between 1.15 to 1.3.
Shear Force and Torsion
Failure Mode and Calculation Method for Ultimate Load - Since diagonal tensile failure, compressive failure in shear, failure due to failure of shear
reinforcement, and compressive failure of web concrete are all brittle failures,
design may be done for any one of the failure modes when ultimate limit state
is shear failure mode. However, it will be necessary to use a member factor
considering enough safety according to each failure mode.
Member Factor - Although the member factor Tb for shear force and
torsion for steel reinforcement is taken between 1.15 to 1.3, the value may
generally be taken as between 1.3 to 1.5 in case of CFRM.
LIMIT STATE OF SERVICEABILITY
Consideration of Flexural Cracking
Limit State of Flexural Cracking - One of the limit states of flexural
cracking indicated in (i) to (iii) below is to be selected in accordance with the
function and purpose of the concrete member.
(i) A limit state of stress intensity of concrete due to bending moment and
axial force not reaching tensile stress intensity (tensile stress initiation
limit state).
(ii) A limit state of the stress intensity of concrete due to bending moment
and axial force not exceeding 60% of the design tensile strength of the
concrete (flexural crack initiation limit state).
(iii) A limit state of flexural crack width not exceeding the allowable crack
width (flexural crack width limit state).
Allowable Crack Width - CFRM do not have problems of corrosion.
Accordingly, it is not necessary to set allowable crack width from the point of
view of corrosion. However, excessively large cracks are not desirable since
they will be a problem from the standpoint of aesthetics of a structure, or
FRP Reinforcement
555
will present an appearance causing uneasiness about safety. Consequently, it is
desirable to study conditions with and without cracks, and when crack initiation
is to be allowed, an appropriate allowable crack width shall be decided in
accordance with the type of structure and distance from which it is seen by
the people, so that there will be no problem from the standpoints of aesthetics
and outward appearance. In general, when prestress is not induced in the
main reinforcement, crack width will become fairly large even under low-level
load. Furthermore, in case of combined use with deformed bars, it is necessary
to decide on an allowable crack width considering corrosion of the steel. In
general, allowable crack width may be taken to be between 0.3 to 0.5 mm.
Consideration of Stress Intensity
The tensile stress intensity of prestressing material due to bending moment and axial force are not to exceed the limiting values. The limiting value
of tensile stress intensity of CFRM is to be 60% of the characteristic value of
tensile strength. However, this must be reduced, when necessary, when factors
such as static fatigue are considered. The material coefficient for static fatigue
strength, in general, may be taken as between 1.2 to 1.3, which is different from
that of steel, 1.0.
GENERALSTRUCTURALDETMLS
Bent Shaping of Reinforcing Material
Reinforcing steel can be subjected to bending by plastic fabrication, but
in case of CFRM, bending fabrication cannot be done. Consequently, when
using CFRM as reinforcement having bent portions like stirrups and hoops,
reinforcing material made into the required configurations by bent shaping needs
to be used. The strength of a bent shape of shear reinforcement and design of
ultimate load of member are not independent of each other, and the strength of
the bent shape must be in accordance with the design strength. The strength
of the bent shape portion will differ depending on the kind of the reinforcing
material, method of shaping, bending radius, etc. When the method of bent
shaping has been decided beforehand, the strength of the bent shape portion
may be determined by an appropriate test and this can be used for calculating
ultimate load. When the strength of the bent shape portion is determined by
ultimate load calculations, the bending radius may be selected carrying out
appropriate tests so that the strength of the bent shape portion will be higher
than the design values.
Curved Arrangement of Straight Reinforcing Material
When arranging straight CFRM in curved form. tests must be performed
beforehand regarding the bending radius making possible a curved condition
by elastic deformation without damage. Since secondary stress will occur in
reinforcement at a curved portion, it is necessary to give consideration that a
difference will not occur between the bending radii in design and in construction
along with the effect on strength of the reinforcing material.
556
Okamura et al
REQUIREMENTS FOR EAR1HQUAKE RESISTANCE
According to the current Standard Specification for Design and Construction of Concrete Structures (3), it is necessary that an earthquake-resistant
design of structures should be carried out based on safety during earthquake and
serviceability of the structure after earthquake. Earthquake load is set considering the response characteristics of the ground and structure under the earthquake
design load and plastic deformation capacity, and the degree of damage allowable after the earthquake. However, with a concrete structure reinforced by
CFRM, it is extremely unlikely that plastic deformation occurs. This means
that it is necessary for a structure to be in the elastic range at all times during
an earthquake, while regarding the degree of damage suffered, it is required
for soundness to be maintained. Consequently, when carrying out earthquakeresistant design of a concrete structure reinforced with CFRM, it is necessary
for the structure to have strength greater than the earthquake force acting and
the structure behaving elastically at all times. Furthermore, when safety is to be
examined, appropriate member factors and structural analysis coefficients are
to be used.
·
REFERENCES
(1). "The Design Concept for Concrete Members Using Continuous Fiber
Reinforcing Materials," Concrete Library International, No.20, JSCE,
1992.8.
(2). "State-of-the-Art Report on Continuous Fiber Reinforcing Materials," Concrete Library, No. 72, JSCE, 1992.4. (in Japanese)
(3). Standard Specification for Design and Construction of Concrete Struc-
tures, Part I (Design)," JSCE, 1986. (revised in 1991)
ACKNOWLEDGMENTS
The Design Concept for Concrete Members Using CFRM was completed by the efforts of the committee members. The authors would like to
acknowledge their contributions and also would like to thank the Association of
CCC (Composite Material Using Continuous Fiber for Concrete Reinforcement;
President Minoru Sugita), which entrusted JSCE to organize the subcommittee
on CFRM.
FRP Reinforcement
557
APPENDIX
The content of the committee report titled "The Design Concept for
Concrete Members Using Continuous Fiber Reinforcing Material" (Concrete
Library International No.20, JSCE, 1992.8) is as follows:
Design Concept for Concrete Members
Using Continuous Fiber Reinforcing Materials
(by JSCE Research Subcommittee on CFRM)
CONTENTS
1. Introduction
2. Terminology Concerning Continuous Fibers
3. Types , Methods, and Properties of Continuous Fiber Reinforcing Materials
3.1 Types of Continuous Fiber Reinforcing Materials
3.2 Methods of Use as Concrete Reinforcing Materials
3.3 Mechanical Properties of Continuous Fiber Reinforcing Materials
4. Design Concept
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
Design Values of Materials
Load
Structural Analysis
Consideration of Ultimate Limit State
Limit State of Serviceability
General Structural Details
Design of Members
Requirements for Earthquake Resistance
Consideration of Fatigue Limit State
558
Okamura et al
IClassification of continuous fiber reinforcing material
IClassification by material I
1. Classification by fiber
Inorganic - - [
Fiber
,-PAN-Carbon
Car bon Fiber------<~___
Pitch-Carbon
Alkali
Glass Fiber
B
------<L~~~~~=~t Glass
Aramid Fiber
~i~e~nic ----t-Polyvinylalcohol(PVA)Fiber
thers
2. Classification by binding material
E
Orqanic
Malerial - - - - - - - 1
Epoxy
Vinylester
Unsaturated Polyester
Others
lnorqanic _ _ _ _ _ __
Mate-rial
Special Cement etc.
IClassification by shape I
Continuous
Fiber
Reinforcing
Material
o Surface Treatment
Flat Surface)
One
Straiqh1~ound
Wth S rf
~With Spiral Fiber
Dimensiona1Fornt'
ar
Tieat~en~ce Sand Covered
Bar
Others
No Surface Treatment
~wisted
Po~gonal-[ (Flat surface)
Ba
With Surface Treatment
orm
~~;~ed-rNo Surface Treatment
Lwith Surface Treatment-Sand Covered
Grid
_____,-Plane Grid
Lsolid Glid
Three Dimensional Textile
Shapes
Others ~Spiral Hoop
Lstirrup etc.
Fig. 1-Classification of continuous fiber reinforcing materials
FRP Reinforcement
~
·I
E = breaking strain
(breaking strain : the strain at which the outermost CFRM
reaches fiber breaking load)
Fig. 2-Flexural tension failure at ultimate state in the
case of multi-layered reinforcements
559
SP 138-34
Pultruded FRP Grating
Reinforced Concrete Slabs
by L.C. Bank and Z. Xi
Synopsis: An investigation of the behavior of concrete slabs reinforced
with pultruded FRP (fiber-reinforced plastic) gratings is described. Data
from tests on small-scale and full-scale slab specimens obtained from three
different experimental programs, beginning in September of 1989, are
reviewed. Particular attention is paid to the description of failure modes,
crack patterns, flexural stiffness, and shear response of the slabs.
Analytical methods, based on those developed for steel reinforced
concrete slabs, used to obtain predictions of the ultimate strengths and the
flexural stiffnesses of the slabs, are described. Comparisons between
experimental data and theoretical predictions are presented.
Keywords: Bridge decks; cracking (fracturing); deflection; fiber reinforced
plastics; flexural strength; gratings; reinforced concrete; shear properties;
slabs; stresses
561
562
Bank and Xi
Lawrence (Larry) C. Bank is an Associate Professor in the Department of
Civil Engineering at the Catholic University of America (CUA), located in
Washington, DC. He conducts research on the analysis and design of
composite materials and structures and is a member of ACI Committee
440 - FRP Reinforcement for Concrete.
Zuhan Xi is a Graduate Research Assistant in the Department of Civil
Engineering at CUA. He received his Bachelors degree in Civil
Engineering from Tongji University in Shanghai, China and his Masters
degree from CUA. In 1991/1992 he was a recipient of a Federal Highway
Administration Research Fellowship Grant (GRF).
INTRODUCTION
Of the different composite materials and composite materials
systems being considered for the reinforcement of concrete, those
reinforcement systems that form grids or gratings are of particular interest.
Since composite materials are "manufactured materials" they can be
fabricated in geometric configurations with mechanical and physical
properties that are specifically tailored to their end-uses. This philosophy
has been applied with great success in the aerospace industry where
material and structural "tailoring" has enabled composite material parts to
compete effectively with conventional material parts. Composite materials
in the form of orthogonal gratings and produced by the relatively
inexpensive and automated pultrusion process (1) have the potential to
transfer this material and structural "tailoring" philosophy to the
construction industry in the area of reinforcements for concrete.
In order for composite materials to penetrate the construction
market they must deliver structural performance equivalent or superior to
conventional material systems at lower costs (2). These costs, which
should be assessed over the life-time of the structure, must include
materials costs, construction costs, maintenance costs, repair costs, and
eventually replacement and disposal costs. Currently, composite materials
are being primarily considered as construction materials because of their
potential longevity in adverse environments due to their corrosion
resistance (3),(4). The use of composite materials for this purpose alone
will not be sufficient to offset possible higher materials costs than
conventional materials. Composite material systems that can significantly
reduce construction and maintenance costs due to their material and
geometric "tailoring" must be developed. Such composite material systems
FRP Reinforcement
563
may compete with conventional materials in all markets and not only in
the "corrosion" market. Substitution of composite materials for
conventional materials, such as FRP reinforcing bars, may need to be
avoided since little savings in construction costs will be obtained. In
addition, the different physical and mechanical properties of anisotropic
fiber reinforced polymeric composite materials may make "substitution"
approaches hazardous.
Cognizant of this need to use and develop a composite material
system that by its material, geometric and constructibility properties was
able to deliver to the construction industry the multiple benefits described
above, a research program was initiated to evaluate the suitability of
pultruded FRP gratings for reinforcing concrete. In addition to the
potential corrosion resistance benefits that may be derived from the
non-metallic nature of this reinforcement the pultruded grating provided
numerous other unexpected benefits both in the ease of construction and
in enhanced structural performance. The research to date has used
commercially produced pultruded gratings that were not specifically
produced (or optimized) for reinforcing concretes. Nevertheless, their
suitability as reinforcement substructures for concrete slabs has been
demonstrated and their future optimization seems warranted.
PULTRUDED FRP GRATINGS
Description of Pultruded
Gratin~:s
Pultruded FRP gratings are orthogonal grids consisting of
longitudinal bearing bars (often having distinctive miniature I or T
cross-sections) and transverse bars (cross-rods) having small irregular
shapes consisting of rectangular and circular segments. A photograph of
small pieces of two typical pultruded grating is shown in Fig. 1. The
longitudinal bars are currently produced in heights ranging from 2.54 to
5.08 em (1 to 2 inches) by a number of U.S. manufacturers. The
longitudinal bars are typically spaced at 5.08 to 7.62 em (2 to 3 inches)
on-center and the transverse bars are spaced at 15.24 to 30.48 em (6 to 12
inches) on center. The longitudinal and transverse bars are connected by
proprietary mechanical interlocking systems. The bars are pultruded and
consist of primarily unidirectional E-glass fibers (approximately 70% by
weight) in either Isophthalic Polyester or Vinylester resin matrices.
Grating panels are typically supplied in widths of 1.22 m (4 feet) and
lengths of 3.05 or 6.10 m (10 or 20 feet). By varying the pultrusion die
shapes, the types of fibers used, and the longitudinal and transverse bars
spacings, pultruded grating panels of any bar shape, composite material
564
Bank and Xi
(provided it can be processed), bar spacing, and size (depending on
shipping) can be manufactured.
Mechanical Tests of Pultruded
Gratin~:s
Mechanical tests were performed on samples of two 5.08 em (2 in)
high pultruded FRP gratings (Safe-T-Grate T5020 and Duradek T5000) in
order to obtain the values of the longitudinal modulus (E) of the
pultruded material. (Modulus data is reported by manufactures in a
variety of ways and is intended primarily for grating deflection calculation
purposes. This data could not be used with confidence in the analytical
part of the present study.) Due to the complex geometry of the pultruded
longitudinal bars three different methods (using two different tests) were
used to obtain the "effective" longitudinal modulus of the pultruded
material. Modulus data is the most critical to obtain since, as will be seen
in what follows, pultruded grating reinforced concrete slabs need to be
designed as "over-reinforced" slabs to meet serviceability deflection
requirements. Consequently, they are designed to fail in concrete
compression or concrete shear modes and the strength of the FRP grating
is not that significant. Based on manufacturer's data the results of slab
testing longitudinal strength of the pultruded materials is assumed to be
413.7 MPa (60 ksi) or greater.
Coupon Tensile Tests -- "Dog-bone" coupons with a tapered gage
region measuring 7.62 em (3 in long) by 1.91 em (0.75 in) wide were cut
from the webs of the pultruded longitudinal bars with thickness 0.32 em
(0.125 in). Transverse bars were removed and strain gages placed on both
sides of the coupon in the gage region. Two specimens from each grating
type were tested. Coupons were tested in a universal testing machine.
Failure occurred due to longitudinal splitting (as expected) and the tensile
strength of the material could therefore not be determined from these
tests. All coupons reached stress levels in excess of 206.85 MPa (30 ksi)
prior to failure. The stress-strain response, shown in Fig. 2, was linear to
failure.
Full-size Flexural Tests -- Four point bend tests (interior span of
22.86 em (9 in) and exterior span of 60.96 em (24 in)) were performed on
full-size pieces of the pultruded grating. The full-size specimen width was
25.4 em (10 in) (5 longitudinal bars) and the length was 91.44 em (36 in).
Since the pultruded bars have an irregular geometry and the fiber
distribution within the bar may be different in the flange and web
elements of the bar full-size tests to determine the effective modulus of
the material were deemed appropriate. The effective modulus was
obtained from deflection measurements and from flexural strain
FRP Reinforcement
565
measurements (obtained from the extreme tension and compression
flanges) in the constant moment region. Results of the testing are given in
Table 1.
Differences between values obtained from the tensile and the
flexural tests are attributed to geometry of the pultruded bar, fiber
placement within the bar, and the laminated nature of the material for
which different longitudinal and flexural moduli are predicted by
lamination theory (5). Differences in the deflection and the strain
predictions are most likely due to shear deformation effects resulting from
the short test span (6). It should be noted that when the pultruded grating
is used as a reinforcement in the concrete it may be stressed in tension or
compression through its full depth. This stress distribution is non-uniform
and has a distribution which depends on the distance from the neutral axis
of the slab. (For analytical purposes this distribution is assumed to be
linear). Alternatively, the neutral axis of the slab may fall within the
grating and both tensile and compressive stresses will exist simultaneously
in the grating. In the flexural tests both compressive and tensile stresses
exist. A choice therefore exists as to which E value to use for the
analytical studies. In this work analytical predictions based on the
measured values are presented to show the effect of the modulus of the
pultruded material on the response of the slabs. In preliminary design
studies, however, a value of E = 34.5 GPa (5 x 106 psi) was used to size
the slabs.
EXPERIMENTAL STUDIES OF CONCRETE SLABS
Tests have been conducted on concrete slabs reinforced with
commercially produced pultruded fiber-reinforced plastic gratings. The
details of the tests have been reported in prior publications. In this paper
an attempt is made to assimilate the test data presented previously and to
identify common characteristics of the responses of the various slabs. A
brief review of the test programs is first presented.
Small-scale Tests
To determine the feasibility of reinforcing concrete slabs with FRP
gratings a series of small-scale tests (7) was first performed. Tests were
conducted at The Catholic University of America beginning in September
1989. Both molded and pultruded gratings were used in the study. Of the
nine slabs tested three slabs were reinforced with pultruded gratings
having the distinctive I and T cross-sectional shapes. Slabs measuring
566
Bank and Xi
142.24 em (56 in) long by 30.48 em (12 in) wide by 10.16 em (4 in) deep
and singly reinforced in the tension zone by one grating panel (either
Duradek 1-6000 (Slab # 2), or Safe-T-Grate 1-6015 (Slab #'s 3 and 9))
were tested in four-point bending. The primary purpose of this
investigation was to investigate shear failure modes; the slabs had shear
span to effective depth (a/d) ratios of approximately 3. Another purpose
of the investigation was to determine effective flexural stiffnesses of the
grating reinforced slabs and to determine whether or not grating
reinforced slabs could be designed to achieve flexural stiffness that would
allow the slabs to meet serviceability deflection criteria. Details of the
concrete mixes, grating geometries, fabrication methods, testing procedures
and a description of the failure modes can be found in (7). In what
follows comparisons between experimentally measured and analytical
predictions of ultimate strengths and flexural stiffnesses of these slabs are
shown. The failure modes of the slabs were of particular interest.
Although predicted to fail in shear two slabs (#'s 2 and 3) failed in
concrete compression. The remaining slab ( # 9) which had a higher
concrete strength failed in shear. Other slabs, reinforced with molded
gratings failed primarily in shear modes with distinctive horizontal shear
cracks propagating along the interface between the grating and the
concrete as shown in Fig. 3. The major conclusions of this study were that
the gratings significantly increased the shear resistance of the slabs, that
excellent bond was obtained between the pultruded grating and the
concrete due to mechanical interlocking with the transverse bars (even if
the transverse bars were cut as in slab # 9), and that the flexural
stiffnesses of grating reinforced slabs could achieve acceptable values.
Full-scale Tests
Based on the results of the feasibility study a series of full-scale
tests of simply-supported singly-reinforced pultruded grating reinforced
slabs was conducted at the Turner Fairbank Highway Research Laboratory
of the FHWA (8),(9) beginning in the July of 1991. Six slabs measuring
3.05 m (10ft) long by 1.22 (4ft) wide by 21.59 em (8.5 in) deep were
reinforced with single panels of 5.08 em (2 in) high and 5.08 (2 in)
on-center pultruded gratings (Safe-T-Grate T5020 (Slab #'s Cl. C2, C5),
and Duradek T5000 (Slab #'s C3, C4, C6)). The purpose of this study was
to determine whether grating reinforced concrete slabs could meet
AASHTO (10) serviceability and strength requirements to serve as
highway bridge decks. The investigation focused on flexural stiffness (at
service load), on cyclic versus monotonic loading, on ultimate strength and
on failure modes. Slabs were tested in three-point bending over 2.44 m (8
ft) spans and failed in flexure by compressive crushing of the concrete
(a/d ~ 7). The compressive failure was followed immediately by a
FRP Reinforcement
567
diagonal shear crack which then propagated to the free edge of the slab
(see (8),(9) for details). A photograph of a slab at failure is shown in Fig.
4. Strength and stiffness data for these slabs is presented in what follows.
The performance of the grating reinforced slabs was compared with that
of a steel reinforced slab (Slab # C7 reinforced with #5 bars @ 11.43 em
(4.5 in)). The major conclusions of this study were that full-scale grating
reinforced slabs could be fabricated with no difficulty, that the grating
reinforced slabs gave admissible deflections at service loads, that the
ultimate strengths of the grating reinforced slabs were superior to that of
the steel rebar reinforced slab, that effects of cyclic loads were predictable
and stable, and that once again superior shear strength and bond strength
of pultruded grating reinforced slabs was observed.
The full-scale simply-supported tests were followed in May of 1992
by tests of doubly-reinforced continuous slabs (9),(11 ). The slabs
measured 6.10 m (20ft) long by 1.22 (4ft) wide by 21.59 (8.5 in) deep and
were each reinforced with two 5.08 em (2 in) high pultruded gratings
placed symmetrically about the midplane of the slab with a 2.54 em (1 in)
concrete cover. For the bottom reinforcement the same grating as the
simply-supported slabs was used (Safe-T-Grate T5020 and Ouradek
T5000). For the top reinforcement custom gratings (Safe-T-Grate and
Ouradek) with longitudinal bar spacings of 7.62 em (3 in) on-center were
used to guarantee concrete infiltration between the top reinforcement
grating bars. Three slabs with each grating type (Slab #'s Sl, S2, S3, and
01, 02, 03) were tested in addition to a steel control slab (Slab # ST)
isotropically reinforced with #5 bars @ 11.43 em (4.5 in) both top and
bottom (with a 2.54 em (1 in) concrete cover to provide for comparison
with the grating reinforced slabs). The slabs were placed on three
supports spaced 2.44 m (8 ft) apart and loaded on both spans
simultaneously. The loads were located 1.03 m (3.38 ft) from the middle
support to develop maximum negative moment over the middle support.
The shear span (a/d) to effective depth of the continuous slabs was
approximately 4 for this configuration. Assuming that the slab is a wide
beam, beam theory can be used to calculate the support reactions and the
slab bending moments. The purpose of this study was to investigate
fabrication issues related to doubly-reinforced slabs, to investigate the
behavior of continuous slabs having both negative and positive curvatures,
to observe the response of the pultruded grating in the compression zone
of the slab, to obtain further information on failure modes, and to
investigate post-peak load carrying capacity of pultruded grating reinforced
slabs. In the tests on the simply-supported slabs there were indications
that the slabs could continue to carry load (less than maximum load)
following the concrete failure as the grating itself began to fail. Results of
static tests on all slabs are given in (9). The results for the two cyclic tests
(Slab #'s S2 and 02) were consistent with data obtained from the static
tests and with data obtained from the cyclic tests on the simply-supported
568
Bank and Xi
slabs. Ultimate strength and flexural stiffness data for all slabs are
reported in what follows.
All slabs failed in shear modes in the short shear span region near
the middle support. The nature of the shear crack patterns and the
post-peak response of the grating reinforced slabs was markedly different
from the steel reinforced slab. The steel slab failed catastrophically with
no prior warning at 387 kN (87 kips) (per span). The failure was a typical
flexural-shear failure with a single shear crack running from the loading
point to the middle support (see Fig. 5) . Following failure the load
carrying capacity dropped to zero. The shear failure of the grating
reinforced slabs was precipitated by sounds of cracking at approximately
80% of the ultimate load. This cracking sound, which could not be
detected visually on the concrete surface, is felt to be probably due to
local compressive or shear failure of the pultruded grating material or
concrete microcracking. At failure a series of diagonal shear cracks
developed in the concrete confined in the 6.35 em (2.5 in) deep layer
between the two gratings (see Fig. 6). These cracks then propagated
horizontally along the interfacial boundary between the concrete and the
grating creating large internal delamination planes. Maximum loads
ranging from 378 to 507 kN (85 to 114 kips) (per span) were recorded at
this point. Following the maximum load the slabs continued to carry load
at about 50% of ultimate load while undergoing large deflections (11).
During the post peak regime dramatic progressive global failure of the
entire system could be seen. The pultruded gratings (both top and
bottom) failed due to multiple longitudinal shear failures and bearing
failures at the cross-rod locations brought about by the large rotations due
to the large post-peak deflections. Horizontal cracking of the concrete at
all the interfacial zones continued to develop. A photograph of the shear
failure location well into this post-peak regime is shown in Fig. 7. Major
conclusions of this third series of tests were that doubly-reinforced slabs
could be readily fabricated, that deflections at service loads were
satisfactory (although local curvatures over the middle support were quite
large (see (9),(11) for numerical data)), that ultimate load carrying
capacities were superior to the steel slab and were substantially increased
due to the concrete confinement obtained from the grating, that prior
warning of impending shear failure was given, and that following initial
failure at peak load the slabs all continued to carry significant loads (well
above service loads) while failing in a "ductile-like" manner and absorbing
significant energy.
FRP Reinforcement
569
ANALYTICAL METHODS
Analytical methods were used to predict the behavior of pultruded
grating reinforced concrete slabs. These methods were used to develop
preliminary designs of the full-scale slabs for the experimental studies
described above. Methods used were based on current methods used for
the analysis of steel reinforced concrete slabs ( 12),( 13 ),( 14 ). Analytical
methods to predict the flexural stiffness, El, and the ultimate flexural
strength, M" 11 , of a grating reinforced slab are described in what follows.
ACI formulas to predict nominal shear strengths of the slabs and methods
to predict slab deflection are reviewed. Comparisons between analytical
predictions and experimental data for the slabs tested are presented.
Flexural Stiffness Model
Since serviceability deflection criteria play such a significant role in
the design of FRP reinforced concretes it is important to obtain precise
predictions of the effective slab flexural stiffness (particularly at service
load). The effective slab flexural stiffness is based on ACI practice and is
defined as,
(1)
where Ie is given by ACI formula (9-7) (12) as,
(2)
where Ic, is the second moment (moment of inertia) of the transformed
fully cracked section. Using Eqs. (1) and (2) the effective stiffness of the
grating reinforced slab can be calculated at any load level. Load versus
deflection curves can be generated using these formulae. The applicability
of Eq. (2) which was developed for steel reinforced slabs is still being
evaluated for FRP grating reinforced concrete. Since flexural cracks in
FRP grating reinforced slabs develop at the locations of the transverse
bars (9),(11) they tend to be wider and deeper than cracks in rebar
reinforced slabs that develop bond continuously. Observation of crack
patterns indicates that the fully cracked section develops early in the
loading history. Therefore it is felt that Eq. (2) overestimates the "early"
stiffness of FRP grating reinforced slabs. For this reason Icr is used for
theoretical prediction of the slab flexural stiffness that is presented in what
follows.
570
Bank and Xi
Calculation of In in the case of the pultruded grating reinforced
slab must account for the distinctive I or T shaped geometry of the
longitudinal grating bars. The calculation I,, is facilitated by defining a
"unit cell" having width b, (equal to the on-center spacing of the grating
bars) containing one pultruded longitudinal grating bar in the tension zone
as shown in Fig. 8(a). For this unit cell the height of the neutral axis, Yna•
and the second moment of its transformed fully cracked section r,, are
first calculated. Ic, for the entire slab is then obtained from the
relationship,
(3)
where bw is the width of the entire slab or the unit width (e.g. 1 m). For
given area, AFRP• second moment, IFRP• and height of the neutral axis, YFRP•
of the longitudinal grating bar (from manufacturer's data) the location of
the neutral axis of the fully cracked transformed cross-section (for both the
unit call and the entire slab) is found from the solution to the quadratic
equation,
where the modular ratio, n = EFRr/Eco~cRETE and the remaining variables
are defined in Fig. 8(a). Once the location of the neutral axis of the fully
cracked cross-section has been found r,, of the unit cell is calculated from,
(5)
It is clear from Eqs. (4) and (5) that the shape of the pultruded bar has a
significant effect on Icr· The contribution of the second moment of the
grating bar to the overall second moment of the fully cracked cross-section
is seen explicitly in Eq. (5). As the modular ratio increases this
contribution increases in a direct proportion. In the case of concrete
reinforced with circular cross-section rebars this contribution is neglected
and it is assumed that only the area of the reinforcing material is
significant (i.e. IFRP and YFRP are assumed to be zero, and the depth h or d
is taken to the centroid of the reinforcing bar). Since this paper considers
the use of pultruded gratings for reinforcing slabs in which the height of
the pultruded grating bar relative to the effective depth of the slab is very
significant (about 1/3) Eqs. (4) and (5) are recommended for calculation
of the section properties. In the case of beam reinforcement where this
ratio would be significantly smaller the common assumption of considering
the area of the reinforcement only may be appropriate. For the case of
the doubly reinforced slab a similar equation which accounts for the
FRP Reinforcement
571
presence of the pultruded bar in the compression zone of the slab can be
developed.
Ultimate Strength Model
A model to predict ultimate flexural strength of pultruded grating
reinforced slabs was developed following the models used for steel
reinforced concrete (14). The model is based on an assumed linear strain
distribution through the thickness of the slab and on internal equilibrium
of tensile and compressive forces. The model considers two distinct
failure modes in flexure; compressive failure of the concrete at a
compressive strain of 0.003 and tensile failure of the FRP reinforcing
which depends on the longitudinal strength of the pultruded material.
In the case of compressive failure of the concrete the usual
Whitney rectangular stress distribution (maximum stress of 0.85fc) is
assumed for the concrete and a linear stress distribution based on a linear
stress-strain relation is assumed in the grating bar (see Fig. 8(b) ).
Equilibrium of internal tensile force in the grating bar and the
compressive force in the concrete is obtained by integrating over the
cross-section of the grating bar and the rectangular area of the concrete.
A trial and error method is used: First the location of the neutral axis is
assumed, then the resultant tensile and compressive forces in the section
are obtained by integration. If the forces are unbalanced a new location
of the neutral axis is assumed. The process is repeated until the
difference between the tensile and compressive forces reaches an
acceptably small limit. The ultimate moment of the section, which is
equal to the internal couple due to the resultant tensile and compressive
forces, is then calculated.
In the case of tensile failure of the pultruded grating bar the
non-linear compressive stress distribution in the concrete due to Park and
Paulay (13) is used. This model has been used recently to model to
behavior of reinforced concrete beams strengthened with bonded FRP
plates in (15) and (16). In this work the parameters of the Park and
Paulay model used in (15) are used; the stress distribution in the concrete
is found using the following stress strain relation,
(6)
where ( and E, are the stress and strain in the concrete, respectively. A
trail and error method, similar to the one used for the compressive failure
572
Bank and Xi
fc
=
l
[1 - 300
(7)
(ec - 0.002)]
model is used. First the location of the neutral axis is assumed, then the
strain in the extreme fiber of the pultruded grating bar is set equal to the
. ( eFRP
failure
f m"I ure stram
=
ultimate IF~) an d t he extreme stram
. m
. th e
oFRP
1 """"FRP
concrete is calculated. (If the concrete strain exceeds 0.003 the
compressive failure model is used). The internal forces are obtained by
integration using a linear stress distribution in the grating bar and the
nonlinear stress distribution for the concrete (see Fig. 8(c)). Once the
internal forces have been obtained they are compared and the process
continues as in the case of the compressive failure model until the
ultimate moment is obtained.
Nominal Shear Stren!nh
The nominal shear strength, V", of the grating reinforced slabs is
calculated according to the ACI procedure for beams without shear
reinforcement. ACI formula (11-3) which is recommended for calculation
of the nominal shear strength for continuous beams {14) states,
(8)
The effective depth d for the grating reinforced slab is estimated by
assuming that it is the distance to the centroid of the grating bar, i.e.
d
=
h -
YFRP
(9)
COMPARISONS BE1WEEN THEORY AND EXPERIMENT
Geometric properties obtained from measurement of cross-sectional
dimensions of the pultruded grating bars used in the slabs tested are given
in Table 2. Table 3 shows comparisons between the predicted flexural
stiffness EJrr and experimental values of Elcxp obtained from strain gage
readings on the extreme fibers of the FRP grating bars in tension, E FRP
and on the concrete in compression, E'" The experimental flexural
stiffness value is calculated from the beam moment-curvature relationship
as,
FRP Reinforcement
E
M
Jexp = -
M
573
(10)
1C
The experimental value is reported at a point in the loading history at the
nominal service load (approximately twice the cracking moment of the
slab). For the full-scale slabs the service load was 115.65 kN (26 kips)
(8),( 11 ). Note that for the continuous doubly-reinforced slab data are
given for both the positive moment and the negative moment regions of
the slab. Since the grating bars in the top grating panel are at 7.62 em (3
in) on center and those in the bottom panel are at 5.08 em (2 in) on
center the flexural stiffnesses are different. Note also that in the
theoretical calculation the presence of the grating in the compression zone
is neglected. This would tend to slightly underestimate the theoretical
predictions. It is seen that the theoretical predictions are generally in
reasonable agreement with the experimental data. The dependence of the
slab stiffness of the modulus of the FRP material is quite pronounced (as
is expected) and accurate measurement of the FRP modulus is required.
The experimental data is somewhat erratic; measurement of flexural
stiffness with strain gages placed on the concrete and on the FRP gratings
embedded in the concrete is complicated by numerous factors.
Experimental flexural stiffness obtained from deflection readings in the
pQsitive moment regions were a little more consistent, however such data
cannot be used to obtain the stiffness over the middle support in the
continuous slabs. Nevertheless, all data are of the same order of
magnitude and no gross discrepancies were observed. Experimentally
measured flexural stiffnesses in the negative moment region were generally
less than those in the positive moment region.
Table 4 shows a comparison between the predicted flexural
strengths, nominal shear strengths and the experimental ultimate moments
and shear forces in the slabs. In the case of the continuous slabs data are
given for both the maximum negative and maximum positive moments.
The experimental failure modes are also given. Note that the model only
predicts flexural failure not shear failure, so that slabs failing in shear are
expected to have failed before their theoretical ultimate moments are
reached. According to the theoretical model all predicted flexural failures
are due to concrete compression failure. This is because the slabs are
over-reinforced to obtain desired flexural stiffnesses. For the slabs that
failed in flexure the predicted strengths are in reasonable agreement with
the experimental strength. For the doubly-reinforced slabs that failed in
shear predicted flexural strengths are quite close to the experimental
strengths due to the fact that maximum experimentally observed
compressive strains in the concrete at failure were close to 0.003. As
expected the experimentally measured slab shear strengths at failure are
significantly larger than the nominal shear strength values (for those slabs
574
Bank and Xi
that failed in shear). A comparison is shown between the theoretical
(using Eq. (2) and experimental load deflection curves for slabs C2 and C6
is given in Fig. 9. Different predictions based on the highest measured
values of the FRP stiffnesses shown in Table 1 (Column 3) are given.
Once again reasonable agreement is seen and the dependence of predicted
response on the FRP modulus is demonstrated.
CONCLUSIONS
The feasibility of reinforcing concrete slabs with commercially
produced pultruded FRP gratings has been demonstrated. Experimental
studies have shown that full-scale slabs can be fabricated with
conventional concrete technology. The enhancement of the shear strength
and the post-peak response of the slabs due to the FRP grating has been
seen to be significant. Analytical models to predict flexural stiffnesses and
flexural strengths have been presented. These models have been shown to
give reasonable predictions. The dependence of the predictions obtained
from these models on the modulus of the FRP grating has been
demonstrated. A model to predict the shear strength of FRP grating
reinforced slabs will need to be developed. The use of the nominal shear
strength as a measure of the slab shear strength is overly-conservative.
ACKNOWLEDGEMENTS
This material is based upon work supported by the National
Science Foundation under grants MSM-9003867 and MSS-9114188. Dr. J.
Scalzi serves as Program Director. Support for the second author was
provided by the Federal Highway Administration under the Grants for
Research Fellowships (GRF) program. Donations of FRP gratings from
manufacturers are gratefully acknowledged.
REFERENCES
(1) Strong, A.B., Fundamentals of Composites Manufacturing: Materials
Methods and Applications, Society of Manufacturing Engineers, Dearborn,
Ml, 1989.
(2) Bank, L.C., "Questioning Composites," Civil Engineering, January 1993,
pp. 64-65.
FRP Reinforcement
575
(3) Iyer, S.L., and Sen, R., (eds.), Advanced Composites Materials in Civil
Engineering Structures, ASCE, New York, NY, 1991.
(4) Suprenant, B., (ed.), Serviceability and Durability of Construction
Materials, ASCE, New York, NY, 1990.
(5) Tsai, S.W. and Hahn, H.T., Introduction to Composite Materials,
Technomic, Lancaster, PA, 1985.
(6) Bank, L.C., "Flexural and Shear Moduli of Full-Section Fiber
Reinforced Plastic (FRP) Pultruded Beams," ASTM Journal of Testing and
Evaluation, Vol. 17, 1989, pp. 40-45.
(7) Bank, L.C., Xi, Z., and Mosallam, A.S., "Experimental Study of FRP
Grating Reinforced Concrete Slabs," in Advanced Composites Materials in
Civil Engineering Structures (eds. S. Iyer and R. Sen), ASCE, New York,
NY, 1991, pp. 111-122.
(8) Bank, L.C., Xi, Z., and Munley, E., "Tests of Full-Size Pultruded FRP
Grating Concrete Bridge Decks," in Materials - Performance and
Prevention of Deficiencies and Failure (ed. T.D. White), ASCE, New
York, 1992, pp. 618-631.
(9) Xi, Z., "A Study of Full-Size Pultruded FRP Grating Reinforced
Concrete Bridge Deck Slabs," MS Thesis, The Catholic University of
America, 1992.
(10) AASHTO, Standard Specifications for Highway Bridges, 14th
Edition, Washington, DC, 1989 (with 1990 and 1991 Interims).
(11) Bank, L.C., Xi, Z., and Munley, E., "Performance of
Doubly-Reinforced Pultruded Grating/Concrete Slabs," in Advanced
Composite Materials in Bridges and Structures (eds. K.W. Neale and P.
Labossiere), Canadian Society for Civil Engineering, Montreal, 1992, pp.
351-360.
(12) ACI, Building Code Requirements for Reinforced Concrete, (ACI
318-89), American Concrete Institute, Detroit, MI, 1989.
(13) Park, R., and Paulay, T., Reinforced Concrete Structures, Wiley,
New York, NY, 1975.
(14) Wang, C-K., and Salmon, C.G., Reinforced Concrete Design, 5th
Edition, Harper Collins, New York, NY, 1992.
(15) Triantafillou, T.C., and Plevris, N., "Post-Strengthening of R/C
576
Bank and Xi
Beams with Epoxy Bonded Fiber Composite Materials," in Advanced
Composites Materials in Civil Engineering Structures (eds. S. Iyer and R.
Sen), ASCE, New York, NY, 1991, pp. 245-256.
(16) An, W., Saadatmanesh, H., and Ehsani, M., "RC Beams
Strengthened with FRP Plates. II: Analysis and Parametric Study," ASCE J.
Structural Engineering, Vol. 117, No. 11, 1991, pp. 3434-3455.
TABLE 1 - VALUES OF E FOR PULTRUDED GRATING MATERIAL
web
Grating
~train
deflection
Etemile
Enexu1.tl
Ene:xural
(GPa)
(GPa)
(GPa)
Safe-T-Grate
T5020 2"
38.8
43.7
54.9
Duradek
T5000 2"
31.7
31.1
34.9
Note: 1 GPa = 145.03 ksi
TABLE 2 - GEOMETRIC PROPERTIES OF FRP GRATING BARS
I
Grating
II
YFRP
I
(em)
I
AFRP
(cm 2)
I
IFRP
(cm 4 )
D-1 1"
1.21
1.80
1.46
S-1 1.5"
1.67
1.85
4.37
D-T2"
2.00
3.46
11.07
S-T 2"
1.98
3.50
11.11
Note: 1 em = 0.39 in; 1 cm2 = 0.16 in 2 ; 1 cm 4 = 0.024 in4
D = Duradek; S = Safe-T-Grate
:t from the top flange (see Fig. 8)
I
577
FRP Reinforcement
TABLE 3 - FLEXURAL STIFFNESS COMPARISONS
Slab
Grating
#
f,c
be
Eclcr
Elcxp
(MPa)
(em)
(kN-m 2 /m)*
(kN-m 2 /m)~
2
D-16000
18.29
3.8
621
678
3
S-16015
18.29
3.8
678
687
9
S-16015
24.82
3.8
716
678
C3
D-T5000
34.00
5.08
4642
5206
C4
D-T5000
34.02
5.08
4642
5112
C6
D-T5000
34.02
5.08
4642
3201
C1
S-T5020
34.00
5.08
6713
6873
C2
S-T5020
34.02
5.08
6713
3672
C5
S-T5020
34.00
5.08
6713
4642
S1
S-T**20 1
24.13
5.1
7.6
6421
5178
3983
4237
S2
S-T**20
21.37
5.1
7.6
6317
5093
4849
2947
S3
S-T**20
35.85
5.1
7.6
6751
5395
6195
5348
D1
D-T**OO
24.13
5.1
7.6
4491
3606
4124
4642
D2
D-T**OO
21.37
5.1
7.6
4453
3502
3512
4896
D3
D-T**OO
35.85
5.1
7.6
4604
3625
11015
5395
Note: 1 MPa = 145.03 psi; 1 em = 0.39 in;
1 kN-m2fm = 106.21 kip-in 2/ft
:t: stiffness data are given per unit width of slab (1 m);
t ** = 50 for 5.08 em (2 in) on-center; 66 for 7.62 em (3 in) oncenter (custom grate)
(the numbers indicate %open area of grating on top surface)
578
Bank and Xi
TABLE 4- STRENGTH COMPARISONS
Slab
Grating
Mode
#
vmaxexp
vn
(kN/rn)
(kN/rn)
(kN-rn/rn)
(kN-m/m)
Mmax
exp
Mmax
theory
2
D-16000
Ft
230.6
54.0
53
53
3
S-16015
F
226.2
51.1
52
63
9
S-16015
s
265.6
59.8
61
63
C3
D-T5000
F
179.5
164.9
218
210
C4
D-T5000
F
176.6
164.9
215
210
C6
D-T5000
F
181.0
164.9
221
210
C1
S-T5020
F
202.9
164.9
248
244
C2
S-T5020
F
201.4
164.9
225
244
C5
S-T5020
F
207.2
164.9
252
244
S1
S-T**20
312.3
138.6
131
-190
163
-182
S2
S-T**20
294.8
131.3
124
-180
159
-178
S3
S-T**20
305.0
169.3
129
-186
137
-198
D1
D-T**OO
284.6
138.6
120
-173
137
-152
D2
D-T**OO
239.3
131.3
101
-146
133
-126
D3
D-T**OO
s
s
s
s
s
s
305.0
169.3
128
-186
152
-171
Note: 1 kN/m = 68.53 lb/ft; 1 kN-m/m = 2.70 kip-in/ft
Data are given for unit width (1 m) of slab
t S = shear failure; F = flexural failure (experimental)
FRP Reinforcement
579
Fig. 1-Typical pultruded grating
250
--::s
200
cC
p...
.__.,
150
r.n
r.n
~
~
100
E-<
r.n
·••••• COUPON TEST #4
SAFE-T-GRATE
E=3B543 MPa
50
·· · · · • COUPON TEST #5
DURADEK
E=319234 MPa
2000
4000
6000
MICRO STRAIN
Fig. 2-Coupon tensile test data
8000
580
Bank and Xi
Fig. 3-Typical shear failure in a small-scale specimen
Fig. 4-Flexural shear failure in a full-scale simply-supported specimen
FRP Reinforcement
581
Fig. 5-Sbear failure in a full-scale grating reinforced continuous specimen
Fig. 6-Sbear failure in the steel reinforced full-scale continuous specimen
582
Bank and Xi
..... J
1
Fig. 7-Ultimate grating reinforced slab failure
f•=
f~[l-300(e .~.002))
f•= f![2e .10.002-{E c/0.002) 2 ]
h d
~
(a)
(b)
... z:a:,
(c)
Fig. 8-(a) Unit cell of fully cracked cross-section; (b) stress distributions
for concrete failure model, and; (c) FRP failure model
FRP Reinforcement
583
400
300
200
····· THEORY (SAFE-T-GRATE)
······· THEORY (DURADEK)
o #C2 (SAFE-T-GRATE)
• #C6 (DURADEK)
v
#C7 (STEEL BAR)
100
0~--------~----------~--------~--------~
0
2
3
DEFLECTION AT PT.l (em)
Fig. 9-Load-deflection curves for slabs C2 and C6
4
SP 138-35
Flexural Behavior and Energy
Absorption of Carbon FRP
Reinforced Concrete Beams
by T. Kakizawa, S. Ohno,
and T. Yonezawa
Synopsis,
Research and development ofFRP bars and cables for reinforcements of
concrete structure has recently been carried out. The basic behavior of the concrete
members reinforced with these FRP bars has became well understood. However,
there are still debatable points in terms of the design concept such as the
recommended failure mode or required toughness and ductility. The authors
carried out loading tests of the 16 concrete beams reinforced with carbon FRP bars
and cables in order to discuss the both serviceability and ultimate limit states. The
specimens includes the RC, PC, PPC members. The main factors are bond
properties of the FRP reinforcements and prestress force. The experimental results
show that cracking and deformation behavior vary with the prestress force and
bond property of FRP bars, and that the reasonable serviceability condition will be
achieved by controlling these factors. Also, the failure mode were affected by
these factors and the reinforcing systems, despite these specimens have almost
same reinforcement ratio. In relation to the failure mode, the energy absorption,
which is defined as the area enclosed by load-deflection curve, was measured to
discuss the toughness and ductility for the ultimate limit state. The authors
recommend that the design should take into account the toughness based on the
energy absorbed before the maximum load.
Keywords: Absorption; beams (supports); cable; carbon; cracking
(fracturing); deformation; ductility; failure; fiber reinforced plastics; flexural
strength; prestressed concrete; reinforced concrete
585
586
Kakizawa, Ohno, and Yonezawa
Tadahiro Kakizawa, is a research engineer in Takenaka Research Laboratory. He
received an MS in civil engineering from the University of Tokyo. He is currently
working in the area of development of new structural materials.
Sadatoshi Ohno, is a senior research engineer in the advanced materials group in
Takenaka Research Laboratory. He received his Ph. D. from the University of Surrey, U.K. His research interests include fracture mechanism, alkali aggregate reaction, fiber reinforced composites, and new structural materials.
Toshio Yonezawa, is a chief research engineer in the concrete materials group in
Takenaka Research Laboratory. He received his Ph.D. from the University of
Manchester Institute University, U.K .. He has been extensively involved with research on corrosion problems of steel in concrete, high strength concrete, fiber
reinforced concrete and new materials in construction field.
INTORODUCTION
Substantial effort have recently been made to develop fiber reinforced
plastics (FRP) bars and cables for reinforcement of concrete structures. There is
great interest in the high-strength, rust-free, and non-magnetic properties of such
new materials. With regard to the design of structural members using FRP
reinforcement, it has been reported that flexural behavior can be predicted based
on conventional flexural theory for reinforced concrete. However, the members
reinforced with FRP bars or cables exhibit brittle failure prope11ies since FRP
materials have no plastic region, while conventional reinforced concrete and
prestressed concrete show a ductile failure behavior because of the yield of the
steel reinforcement. From this viewpoint, an appropriate design method for
ultimate limit states of concrete members reinforced with FRP still remains to be
investigated. In related discussion, it has also been reported that the compression
failure mode is preferable for such FRP reinforced concrete members, because the
failure of the member proceeds more gradually at the ultimate state compared with
failures of those governed by brittle FRP breakage( I). In contrast, another opinion
is that alternative design methods which can allow for brittle failure of members
due to FRP breakage should be considered for reasons of economy and rationality
(2).
FRP Reinforcement
587
On the other hand, various studies have worked to improve the ductility
of FRP reinforced concrete by controlling bond properties of the FRP
reinforcement or placing the FRP reinforcement in multiple stages (3). Also,
attempts on improving brittle behavior by constraining the compression zone of
the reinforced members have been reported (4),(5).
However, the important concern is to secure the required ductility both for the
members and the structures being designed, and in order to do this more thorough
discussion of the appropriate design needs to be undertaken.
EXPERIMENTS
In this experiment, small beam specimens of rectangular section were
adopted as shown in Figure I and reinforced concrete (RC), prestressed concrete
(PC), and partially prestressed concrete (PPC) systems using FRP reinforcement
were tested. Cable strand of carbon fiber reinforced plastics (CFRP) were used as
prestressing tendons, and two kind of tensile reinforcement- CFRP cable strands
and CFRP deformed bars- were used. Tables I and 2 give the physical propetties
of the reinforcing materials and the strength of the concrete used, respectively.
Loading tests were carried out on 16 types of specimens with different test
parameters - prestress force, reinforcement type, bonded or unbonded tensile
reinforcement, and prestressing cable -as shown in Table 3. The sectional areas
of FRP reinforcement given in this table are the nominal cross sectional area
including the resin. Specimen No. I is an ordinary reinforced concrete beam
incorporating deformed steel bars, specimen No. 2 is an RC beam using FRP
reinforcement, specimens No. 3 through 6 are PC beams without tensile
reinforcement, and all other specimens are PPC beams. Specimen No. 16,
although made of the same material as specimen No. 13, was made with a 5 mmthick permanent form reinforced with polypropylene fiber net to improve inservice propet1ies. Unidirectional loading was applied to all beams, which have a
span of 170 em and a moment span of 30 em. In the tests, load, deflection, strain in
the concrete and reinforcement, and crack width were measured.
RESULTS OF EXPERIMENT AND DISCUSSION
Cracking and Deformation Properties
Table 4 shows the results of the loading tests. Reasonably close
agreement was obtained between the measured and calculated values of cracking
load. Figure 2 illustrates the cracking patterns of the loaded beams. Provided that
the service load is about one third of the maximum load, no appreciable cracking
was recognized in the PC and PPC specimens and the cracking was very fine even
588
Kakizawa, Ohno, and Yonezawa
when it did occur. The cracks in specimen No.2 are very small, less than 0.1 mm,
at the service load and it is thought that no serious problems would arise in an
actual application as far as cracking is concemed. However, since the specimens
used here were small, it must be taken into consideration that cracks tend to be
smaller than in actual concrete members.
The distribution of the cracks along the beams differed according to the
specimen. When the specimen has no tensile reinforcement in the PC member
(specimens No. 3 through 6) or has a partially unbonded tensile reinforcement in
the PPC members, fewer cracks were observed. In the case of the unbonded PC
specimen (CPC58U), cracks were particularly concentrated in the moment span.
When the working load becomes high (over 13 kN) in this specimen, the
deformation is concentrated only at the center, as demonstrated by the deflection
distributions shown in Figure 3. This may cause local secondary stress at the
deformed area and/or frictional damage contacting with the sheath. These effects
are not desirable because FRP reinforcement may break earlier than expected.
On the other hand, specimens with CFRP deformed bars used as tensile
reinforcement show good crack distributions, and have closer crack spacing than
conventional RC members. However, longitudinal splitting cracks were found
along the reinforcement at the ultimate state. These cracks were related to the
dimensions of the specimen, the concrete cover, and the bond properties of the
reinforcement. The bond properties of FRP reinforcement are greatly affected by
their configuration of deformation, and this remains an area for further study.
In specimen No. 16 (CPRC38UB-NET), where the permanent form
reinforced with polypropylene fiber net was used, the cracks were finely
distributed at a spacing ranging from a few millimeters up to one centimeter,
although these cracks are not illustrated in this paper. This behavior is
advantageous when cracking in the application must be limited, and when the
design calls for a wider range of service conditions.
Ultimate Load and Failure Mode
Figures 4 (a)-(f) show load-deflection curves for the specimens. The
results for specimens reinforced with FRP differed depending on the type of
reinforcement and the differences in bonding. The obtained curves are not
basically different from those of earlier reports. When the results for CPRC24BBYY are compared with those for CPRC24- YR, CPRC38BB- YY with
FRP Reinforcement
589
CPRC38BB- YR, and CPRC38UB- YY with CPRC38UB, the ultimate load is
found to be 15 to 20% higher for specimens using CFPR deformed bars rather than
CFRP cables as the tensile reinforcement, in spite of the almost identical
reinforcement system and same reinforcement ratio. The reason for this difference
is thought to be due to be the bonding characteristics of the reinforcement. When
two kinds of reinforcement with different bond properties are used simultaneously
in the specimen, the actual working stress may be different from the prediction,
which is based on the assumption that plane sections before bending remain plane
after bending.
The calculated and experimental values of ultimate strength were
compared in Table 4. The calculated ultimate strength was derived from the
requirement of strain compatibility and equilibrium of force by repeating
calculation for the divided elements of the beam section. In this calculation, the
relationship between concrete stress and strain was assumed to be expressed by
Umemura's equation(6), as follows.
0"
= O"cu { 6.75 ( e (-0.819
where ~
~)- e ( -1.218 ~))}
= _L_
Ecu
cr: concrete stress,
<Jcu: concrete ultimate stress,
£ : concrete strain, £cu: concrete ultimate strain
Also it was assumed that FRP reinforcement behaves elastically until failure. FRP
reinforced concrete was regarded as having reached its ultimate failure state when
either the prestressing cables, the tensile reinforcement, or the concrete reached
failure strain. For the specimens using unbonded tensile reinforcement or
prestressing cables, the ultimate load was calculated using the average value of
calculated strain over the unbonded region. Calculated values tended to be I 0 to
25% smaller than those measured in experiment, showing an almost similar
tendency. One reason for this may be that the nominal failure strength was adopted
as the strength value for the FRP reinforcement and prestressing cables.
In this study, the amount of reinforcement was planned for all specimens
except Nos. I and 5 such that the value of reinforcing index, q
= PrCcr;cr,k)*, came
to about 0.27-0.3, that is the failure mode became close to the balanced failure
state. As a result, the failure mode varied depending on the presence of bonding
* Pr: reinforcing ratio, crr: 0.87crru , crr": ultimate strength of FRP,
cr,k: concrete strength
590
Kakizawa, Ohno, and Y onezawa
and the arrangement of prestressing cables and reinforcement.
The failure modes predicted by calculation and the results observed from
the experiments are given in Table 4. Although they agree relatively well, there is
a difference between predicted and experimental results for PPC members with
unbonded prestressing cables. Although the reason for this is not clear, some local
stress may have been induced because the measured strain in the prestressing
cables was less than the nominal failure strain at the time of fracture.
According to earlier work, the compression failure mode is slightly
ductile compared with the reinforcement failure mode in FRP reinforced concrete.
However, in this experiment, no great difference could be seen even in the case of
a compression failure. This is because the FRP reinforcement failed during
compression failure when the reinforcement ratio was close to the balanced failure
state (CPC69B and CPC58B) (See Figure 4). Considering these points, additional
reinforcement over the amount required to obtain compression failure needs be
incorporated in order to ensure a compression failure without FRP failure.
However, this still presents problems in terms of economical design. It is also
difficult to specify the failure mode in the design, since this would mean taking
account of changes in the material properties over the whole lifetime and of the
scatter in fracture strength of the FRP reinforcement. In the case of PPC members
with CFRP cables as tensile reinforcement, on the other hand, the beam specimens
deformed by relatively large amounts even after failure of the prestressing cable.
This shows the possibility of designing FRP reinforced concrete members with
such deformation behavior by controlling the amount of reinforcement and the
bonding properties.
Energy Absorption
In terms of design, the ductility required of a structural member should
vary depending on its type, importance, service conditions and so on. In the design
for conventional reinforced concrete, the brittle failure of members has been
avoided considering safety. Therefore, normally designed reinforced concrete for
flexural members are ductile, because their energy absorption is due to the capacity
of the steel to deform; that is, they have adequate ductility and great energy
absorption.
In the case of FRP reinforced concrete members, avoidance of the brittle failure
may not be easy and economical. Also, the ductility factor which is usually
expressed as the ratio of the ultimate deformation to the deformation at first yield is
not considered to be proper to assess the characteristics of FRP reinforced
FRP Reinforcement
591
concrete. The ductility factor is a index to express the deformation capacity and
consequently shows the energy absorption. Therefore, evaluation of energy
absorption would be very important in the design of FRP reinforced concrete,
although a discussion of the proper safety factors has to be continued.
Table 4 and Figure 5 show the absorbed energy, which is defined to be
the area enclosed by the load-deflection curve for each specimen. The energy
absorbed before and after the maximum load is shown separately. The values of
the total energy absorption are unlikely to be affected by the failure mode within
the range of this experiment's conditions. PPC members tended to indicate
greater energy absorption than PC, despite having the same amount of
reinforcement. This may imply an advantage for PPC. When the absorbed energy
after the maximum load is compared, the PC members show no energy absorption
after failure of the FRP cable at the ultimate load because only prestressing cables
were placed in the PC members in this experiment.
On the contrary, the unbonded PC specimen CPC58U, which failed in
compressive mode, showed the same energy absorption before the maximum load
as after the maximum load. In PPC members, some energy is absorbed even after
the prestressing cables have failed, since FRP reinforcement could still sustains a
load; it thus allows for further deformation. Also in the case of PPC members
using FRP reinforcement and prestressing cable of different bond prope11ies, the
amount of absorbed energy after the maximum load tended to be higher when
CFRP cables were used as tensile reinforcement, while the energy absorbed before
the maximum load was higher when CFRP deformed rods were adopted as tensile
reinforcement. Thus it is possible to obtain various energy-absorption prope11ies
by controlling the reinforcing system. However, evaluating the energy absorption
after the maximum load is technically difficult (The theoretical calculation is
thought to be possible, but there are some problems in accuracy. ) If the concrete
members are such that energy absorption is not expected, i.e. only vertical loads
act and the member doesn't need resist the earthquake load, it may be meaningless
to evaluate the energy absorption over such a range in the design.
The energy absorption of FRP reinforced concrete beams and slabs
under the influence of vertical forces should be evaluated by the total energy
absorption up to the maximum load. In the case of the members where an
evaluation of repeated energy loads, such as earthquake loads, is required, fUI1her
detailed studies will be necessary.
592
Kakizawa, Ohno, and Y onezawa
CONCLUSIONS
In this study, the cracking behavior and failure properties and the energy
absorption of FRP reinforced concrete have been investigated experimentally,
with the prestress force and bonding properties of the FRP reinforcement taken as
the experimental factors. The failure mode and deformation behavior are found to
change according to the reinforcing system. The absorbed energy is affected by the
reinforcing system, but little by the failure mode, within the range of these
experiments. PPC members absorb more energy than PC in spite of the same
amount of reinforcement. Based on these results, design criteria were discussed in
connection with energy absorption and failure mode. It is recommended that the
design should take into account the ductility evaluated for the energy absorbed
before maximum load.
REFERENCES
(I) H.Mutsuyoshi, A.Machida, and K.Uehara, "Mechanical properties and
design method of concrete members reinforced by Carbon Fiber Reinforced
Plastics," Proceedings of the Japan Concrete Institute, 12-1, pp. 1117-1122,
1990.
(2) H.Nakai, K.Mukae, H.Asai, and S.Kumagaya, "Analytical study on bending
ultimate state of prestressed concrete beams with FRP rods," Proceedings of the
Japan Concrete Institute, 13-2, pp. 749-754, 1991.
(3) Y.Yamamoto, H.Maruyama, K.Shimizu, and H.Nakamura, "Fractural
properties of concrete members with a multi-stage arrangement of CFRP," Japan
Society of Ci vii Engineering, Proceedings of the 46th Annual Conference, pp.
238-239, 1991.
(4) M.Odera, T.Maruyama, and Y.Ito, "Improvement in compression failure
mode of RC beams using CFRP rods," Japan Society of Civil Engineering,
Proceedings of the 46th Annual Conference, pp. 242-243, 1991.
(5) H.Taniguchi, H.Mutsuyoshi, A.Machida, and T.Kita, "A proposal of
Improvement in failure mode of PC flexural members using FRP," Japan Society
of Civil Engineering, Proceedings of the 46th Annual Conference, pp. 244-245,
1991.
(6) Umemura,"Uitimate strength and plastic behavior of reinforced concrete,"
Transactions of AIJ, 1951.
FRP Reinforcement
593
TABLE 1 - MECHANICAL PROPERTIES OF CFRP REINFORCEMENT
Type
Nominal Ultimate
Strength (kN)
Carbon FRP
96 ( tP 10.5}
Strand cable
57 ( tP 7.5)
Carbon FRP
deformed bar
31 ( <f5mm)
Elastic modulus
(GPa)
140
130
TABLE 2- MECHANICAL PROPERTIES OF CONCRETE
Age
w/c
Compressive Elastic modulus
strength (MPa
(GPa)
28 days 55 °/c
27.3
35.3
TABLE 3 -TEST SPECIMENS
Type of reinforcement
Name of
specimen
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
reinbt:ing
sySieiD
RC·SD
RC
CRC
CPC69B
CPCSBB
PC
CPC38B
CPCSBU
CPRC24BB·YY
CPRC24BIHR
CPRC24UB·YR
CPRC38BB·YY
CPRC38BB-YR PAC
CPRC38UB·YY
CPRC38UB·YR
CPRC38BU·YY
CPRC38BU·YR
CPRC38UB-NET
prestressilg
i:able
type
CFR~
strand
cable
Cr®s
I~
-
reilforoement
type
06(8000)
ratio
Cross
~
26.7
CF deformed bals 78.6
-30.4
55.7
strand 30.4
(%)
1.13
0.69
0.56
55.7
CFRP
cable
Reinfon:ilg
Tensile
-
-
CFRPcable
30.4
CFRP
deformed bals
39.3
0.58
30.4
39.3
-CFRPcable 30.4
CA'iPdeicnned bars 39.3
CFRP cable
30.4
0.54
0.58
0.54
0.58
CFRPdebmed IBs 39.3
0.58
CFRPcable
PS
Tensile
Prestress
foroe
load
cable
bars
(kN)
-
bond
-
bond
'0-:30
0.56
0.54
c:FRP deformed bars
Bond
propel1ies
-
68
0.7
lrilond
57
37
57
o:6
bond
24
0.4
37
0.65
IJ'tlOI1d
bond
ratio
to 0"""
0.4
0.6
lllbOIId
bond
bond
l.rtloR!
0.~- bond
lllbolld
• The tensile reinforcement was unbonded in the region of 110cm at the center
of the beam.
paltially •
lfilond
bond
594
Kakizawa, Ohno, and Yonezawa
TABLE 4 -TEST RESULTS
Name of
No
specinens
Ctacking
load (kN)
~ Calc.
MaxinJm
load (kN)
llliiiiSII8d
Calc.
value
Defteldion
(mm)
load
I !&SO
5.1
4.2
14.6
12.3
49.91
2CA::
3 CPC69II
4 IXS8B
4.1
12.4
12.3
4.2
13.4
11.9
35.3
28.9
29.5
26.0
.25.7
-
25.8
21.99
29.36
s a>c3llll
9.2
11.9
7~YY
9.5
10.4
7.4
20.3
15.3
31.0
15.7
18.0
26.5
23.76
21.36
35.56
s~m
7.5
36.1
27.1
9~m
7.1
7.4
31.6
IO~yy
9.5
9.2
31.4
up.am.m
9.5
9.2
12 Cl'fCB.B.YY
8.4
13 Cl'fCB.B.m
8.0
14 CPfCllllli.YY
8.5
IS CPfCllllll.m
9.1
16,..,...,........., 11.5
9.2
6 CPQi8U
7.4
7.4
9.2
9.2
9.2
9.2
Failure mode
Enelgy absorption
(lctkm)
allhe
before altar
miXilun load
II1IICirun I1111Xi1Tur taal
load
65.7>
Experimental
observation
Calcurated
prediction
Yeild of steel bals
65.7> Yeild ol steel bals
79.3 Failn d IBnSie ban;
57~2 B8iiiiiced iiiiiure
64.4 PS cable failure
Fabe d Tensile IJus
Failure d PS cabl8-
79.3
50.3
64.4
0.0
6.9
0.0
36.4
29.0
0.0
30.2
40.27
71.7
89.9
29.5
3.9
101.2 ~laiure
Failure ollensile bals
93.9 & CQIIllf8SSion lalure
25.5
38.01
74.9
12.1
ol PS cable
87.1 Failure
ancllonsie bals
~laiure
24.5
32.70
70.6
39.8
110.4 Failure ol PS cable
FaiUe of PS cable
36.5
25.5
36.24
87.3
0.0
87.3 Failure ol PS cable
24.3
31.4
25.8
30.8
33.5
25.5
26.29
38.41
25.32
25.33
36.23
44.6 109.5
80.0
10.9
45.0
48.8
84.5
0.0
90.4
4.5
25.5
22.5
24.5
25.5
Balanced laiure
a11ar <Xlrl1>· lalure
36.4 Failure d PS cable
59.2 ~laiure
Failure d PS cable
~laiure
~flib&
CorTpession falJre
and lansile bals
Balanced laiure
ancllansile bals
154.1
90.9
93.8
84.5
94.9
~laiure
Falun! of PS cable
Balanced laiure
Faiure of PS cable
Failure of PS cable
Faiure of PS cable
~laiure
Failure d PS cable
FaiUa of PS cable
~laiure
I I -1 I I
I
t
-
150
+
700
1000
100
.j
PC strand
(CFRP cable)
Fig. 1-Details of test specimens
FRP Reinforcement
No.9 CPRC24UB-YR
No.1 RC-SD
I
zs
595
Yl (d{lliafj [\ \
No.2 CRC
I zs mrCwJh\2Y zs I
zs
No.10 CPRC38BB-YY
I zs )J
arr1 r
l\1-\\ r zs
I
No.11 CPRC38BB-YR
I zs
1r~\ zs I
No.12 CPRC38UB-YY
zs
r!r) t\\
No.6 CPC58U
I zs
zs
1
I zs
zs
zs
No.7 CPRC24BB-YY
1
zs r r 1
ccb rflliYr'\
No.8 CPRC24BB-YR
li
r
crhj 1\ \\1-\J/ zs I
No.13 CPRC38UB-YR
cJdi~\)\ zs
No.14 CPRC38BU-YY
1
Fig. 2-Crack patterns of loaded beams
I
596
Kakizawa, Ohno, and Yonezawa
0
-
E
5
_§_
10
c
15
E
(J)
20
(J)
()
~
Q_
0
-------~-------- -~---'-----..
25
. . . . . . . . . . . . . +. . . . . . . . . . . . . . +.......... =:
-CFRC
30
35
40
...........................
T"""'"""'
~:~~~~
"'('"""' -"*- CPRC38BB-YY
L_~~~J_~~~-L~~~~~~~~
0
42.5
85
127.5
170
Location (em)
Fig. 3-Distribution of displacement at the load of 15 kN
FRP Reinforcement
40r----------------------,
z~30
40r----------------------,
z
CRC
~
CPC69B
20
20
/"/
-o
c
I'
..
... --:~
-----~-<
_;;----
.3 10
RC-50
CPC58B
~/ / _.. ~--• CPC38B
~30
----~--­
597
--~---..--,
CPC58U
OL---~--~--~~~~--~
10
20
30
40
50
Deflection
(mm)
60
0
10
20
30
40
50
Deflection
(mm)
(b)
(a)
z
40r----------------------,
CPRC24BB-YY
CPRC24BB-YR
--.../\
/
~30
60
z~30
//_.
/. .r
20
'
~/~
CPRC24UB-YR
Ok---~~~~--~--~--_J
0
10
20
30
40
50
Deflection
(mm)
OL---~--~~~~--~--~
60
0
10
20
30
40
50
Deflection
(mm)
(c)
60
(d)
40r----------------------,
z~30
CPRC38UB-YY
z
~30
20
20
CPRC38BU-YY~~
OL-~~--~~~~~~~~
0 10 20 30 40 50 60 70 80 90
Deflection
(mm)
CPRC38UB-YR
o~~~~~~--~--~~
0
10
20
30
40
50
Deflection
(mm)
(f)
(e)
Fig. 4--Load-deflection curves
60
598
Kakizawa, Ohno, and Yonezawa
0
20
40
60
80
100
120
Energy Absorption (kN-cm)
Fig. 5-Comparison of energy absorption
140
160
SP 138-36
Theoretical and Experimental
Correlation of Behavior of
Concrete Beams Reinforced
with Fiber Reinforced
Plastic Rebars
by S.S. Faza and H.V.S. GangaRao
Synopsis:
Analysis of the experimental results obtained by testing forty
five concrete specimens reinforced with fiber reinforced plastic (FPP) rebars
is outlined in this paper. Theoretical correlations with experimental results are
conducted in terms of elastic and ultimate bending moment, crack width, and
bond and development length. Emphasis is placed on the beam bending
analysis and design using regular as well as high strength (4 - 10 ksi) concrete
reinforced with FRP rebars by modifying the state-of-the-art design as per the
ACI 318-89 code provisions that are applicable for steel reinforced beams.
However, modifications (from current ACI code) for FRP reinforced beams in
terms of ultimate moment capacity, crack pattern, and development length are
made without deviating significantly from the design philosophy given in the
ACI 318-89 code. Equations for design loads and bending resistance, bond
and development lengths, and crack widths are developed in a simplified form
for practical design applications. Similarities and parallels of these design
equations with current ACI 3 I 8-89 equations are maintained when possible.
Keywords: Beams (supports); bending moments; bonding; composite
materials; cracking (fracturing); fiber reinforced plastics; rebars; reinforcing
materials; strength; structural design
599
600
Faza and GangaRao
Salem S. Faza is a Research Assistant Professor in the Constructed Facilities
Center at West Virginia University. He received his Ph.D. from West
Virginia University in 1991. His research specialities are in advanced
composites application in construction and concrete reinforced with fiber
reinforced plastic rebars. He is a member of ACI Committee 440.
Hota V. S. GangaRao is a Professor and Director of the Constructed
Facilities Center at West Virginia University. He received his Ph.D. from
North Carolina State University. His research interests are in structural
analysis and design of buildings, bridges, transmission towers, railroad ties,
and water tanks.
INTRODUCTION
Use of fiber reinforced plastic (FRP) rebars in concrete members has been
limited due to several factors: poor quality of rebars, surface smoothness
leading to inadequate bond strength, lack of design guidelines, high material
cost, and others. However, the salient characteristics of FRP rebars, mainly
their noncorrosive and nonmagnetic nature, have motivated engineers to
consider the use of FRP rebars in reinforced concrete structures. The
dramatic deterioration of the U.S. constructed facilities due to corrosion of
steel prompted further research in the area of FRP rebars. As a consequence,
substantial improvements in the product have been achieved by rebar
manufacturers. Particularly important are the use of helical wrap and sand
coating on the rebar surface to improve the bond between concrete and FRP
rebar. Yet, current FRP rebar manufacturing and testing techniques are not
standardized, and designs of concrete members reinforced with FRP rebars are
far from attaining the acceptable design levels of concrete members reinforced
with steel rods.
Based on the mechanical properties of FRP rebars obtained by Wu, Faza, and
GangaRao (1), forty-five concrete beams were designed and tested under
bending and bond forces (2). Test variables included concrete strengths (4 10 ksi), type of FRP rebar (smooth, ribbed, sand coated), and rebar size.
The current mathematical models in the American Concrete Institute code (3)
design equations of concrete beams reinforced with mild steel cannot be
applied directly to beams reinforced with FRP rebars for the following
reasons:
1)
2)
3)
4)
Low modulus of elasticity of FRP rebars compared to steel.
Inadequate understanding of bond behavior.
Long term degradation of FRP rebars.
Crack width development and post-cracking behavior of the
concrete beams reinforced with FRP rebars.
FRP Reinforcement
601
OBJECTIVE
Based on the experimental results outlined by Faza and GangaRao (2), the
objective of this paper is to develop theoretical correlations for structural
performance of concrete beams reinforced with FRP rebars. The study was
carried out to investigate the effects of reinforcement configuration in terms of
the type, size, and area of the FRP reinforcement, and concrete compressive
strength on bending and bond resistance, and crack width and propagation.
ULTIMATE MOMENT RESISTANCE
The phenomenon of developing moment resistance in FRP reinforced beams is
identical to that of beams reinforced with steel rods, provided that adequate
bond between FRP and concrete is developed, i.e., the ultimate moment
capacity of FRP reinforced concrete beams can be obtained by satisfying
internal force and moment equilibrium equation (Eq. 1).
{y
Mn = As fy d ( I - 0.59 p
to)
( 1)
where,
Mn
As
J
p
f'
c
b
= nominal moment capacity of a section, in.lb
= area of tension reinforcement, in2
= specific yield strength of reinforcement, psi
= distance from extreme compression fiber to centroid of
tension
= ratio of tension reinforcement =As/bd
= specified compressive strength of concrete, psi
= width of concrete section.
The ultimate resisting moment based on the equiliqrium equation (Eq.
I) is applied for FRP reinforced concrete beams as shown in equation 2, in
which the yield strength of steel is substituted with the ultimate tensile strength
of FRP rebars. The validity of this equation is substantiated through the
experimental results outlined in Faza and GangaRao (2). However, a designer
has to check for bond capacity of the FRP reinforced beams. It should be
noted that to attain full bending resistance modes of failure due to bond, shear
and compression must be avoided
Mn = Af fy d ( 1 - 0.59 p
{yr
V)
(2)
602
Faza and GangaRao
where,
Af = Area of FRP rebars in tension, in2
fyf = Effective yield strength of the FRP rebars, psi.
in which the ultimate tensile strength for FRP rebars is calculated as
fyf = 0.80 ff (ff = experimental rupture failure) in order to account for the
current variation in manufacturing and quality control of the FRP rebars (1).
Previous research on FRP rebars as reinforcement for concrete showed that in
most cases it was not possible for the FRP rebars to develop their full strength
due to high tensile strength associated with these rebars. In order to take
advantage of the high tensile strength of FRP rebars, beams with high strength
concretes (6-10 ksi) were tested to maximize the bending resistance of the
beams (2). The cracking moment of the high strength concrete beams
increased and a substantial decrease in the crack width was noted due to good
bond.
Theoretical ultimate moment capacities of the concrete beams
reinforced with FRP rebars are compared with the observed experimental
values in Table 1.
A brief description of columns 1 to 15 of Table 1 is given below.
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
( 11)
(12)
(13)
(14)
( 15)
Group number
Beam number
Tension reinforcement and type
Area of the FRP reinforcement (in2)
Ultimate tensile stress measured in the FRP rebar (ksi)
Concrete compressive strength of companion cylinders at 28
days, fc', (psi)
Experimental rupture strength of the FRP rebar as obtained by
Wu, Faza, and GangaRao (1) in the tension tests (ksi)
Ultimate tensile strength (fyf = 0.80 x ff) of the FRP rebar, ksi
Observed (experimental) ultimate moment capacity of the tested
beams, Mn = P x L/3, (K-ft)
Calculated moment capacity using measured ultimate stress in
the rebars (column 5), utilizing equation (2), (K-ft)
Calculated moment capacity using experimental rupture strength
values of the rebars (column 7), utilizing equation (2), (K-ft)
Calculated moment capacity using ultimate tensile strength
values of the rebars (column 8), utilizing equation (2), (K-ft)
Ratio of observed ultimate moment capacity to calculated
moment capacity using the actual stress values of the rebars
Ratio of observed ultimate moment capacity to calculated
moment capacity using rxperimental rupture strength of rebars
Ratio of observed ultimate moment capacity to calculated
moment capacity using the ultimate tensile strength of rebars.
FRP Reinforcement
603
In column #13, the ratio of the experimental ultimate moment capacity to the
theoretical ultimate moment capacity in most beams was close to or over 1.00,
with a mean value of 1.095. In cases of beams #7, #4, #9, and #F (column
2), where the primary failure was either in bond between the rebar and
concrete or in shear, their ultimate capacities are much lower than the
theoretical predictions. The use of sand coated rebars with higher strength
concrete as in beams #VHl and #VH2 resulted in a nearly balanced failure of
tension failure of rebars reaching to 102 and 110 ksi (column 5, Table 1),
which is immediately followed by the compression failure in concrete.
FLEXURAL CRACKING
Various limitations on crack widths of steel reinforced beams have been
proposed by different investigators (4). Excessive cracking is undesirable
because it reduces stiffness, enhances the possibility of deterioration, and
causes undesirable appearance. The ACI 318-89 code prescribes rules for the
distribution of flexural reinforcement to control flexural cracking in concrete
beams. Good detailing practice is also required to establish adequate crack
control.
Cracking is expected to occur when the induced tensile stress in the beam
reaches the ultimate concrete tensile stress. The tensile stress in concrete is
transferred to the reinforcing bar through bond forces developed between
concrete and reinforcing bar. The tensile stress in concrete at the cracked
section is relieved, becoming zero at the time the crack occurs; however, the
reinforcing bar must carry the tensile forces at that cracked section. The
neutral axis position must shift upward at the cracked section in order to
maintain equilibrium of forces at that section. Cracking will continue to take
place between old cracks until the concrete stresses do not exceed the concrete
tensile strength. New cracks will cease to occur because of one of the
following reasons:
1)
2)
Excessive slip between the rebar and concrete.
Reduction in the distance between cracks to transfer sufficient
stress to the concrete.
A number of equations have been proposed (4) for the prediction of crack
widths in flexural members reinforced with steel. The ACI 224 committee
report (5) on control of cracking of steel reinforced concrete beams and the
ACI 318-89 code (3) reached the following conclusions:
I)
2)
The reinforcement stress is the most important variable.
The thickness of the concrete cover is an important variable,
604
Faza and GangaRao
but not the only geometric consideration.
The area of concrete surrounding each reinforcing bar is also an
important geometric variable.
The bar diameter is not a major variable.
The size of the bottom crack width is influenced by the amount
of strain gradient from the level of the reinforcement to the
tension face of the beam.
3)
4)
5)
The committee concluded that the most probable maximum crack width can be
predicted by the Gergely-Lutz expression. The maximum flexural crack width
Wmax according to Gergley-Lutz, can be expressed as:
Wmax = 0.076 (3 fs
VdA
X
w-3
(3)
in which
(3 =
Ratio of distances to the neutral axis from the extreme tension
fiber and from the centroid of the main reinforcement. A value
of (3 = 1.20 may be used to estimate the crack widths obtained
in flexure.
de =
Thickness of cover measured to the center of the first layer of
bars (in.).
fs =
Maximum stress (ksi) in the reinforcement at service load level
with 0.6 fy to be used if no computations are available.
A
The effective tension area of concrete surrounding the principal
reinforcement divided by the number of rebars. It is defined as
having the same centroid as the reinforcement (in2) as defined
per the ACI 318-89.
While the above expression can be used to predict the maximum crack width,
the ACI 318-89 code prescribes rules for the distribution of flexural
reinforcement to control flexural cracking in beams. Good detailing practice
is required to lead to adequate crack control. The ACI 318 code specifies:
" When design yield strength fy for tension reinforcement (steel)
exceeds 40,000 psi, cross sections of maximum positive and negative
moment shall be so proportioned that the quality z is given by
z
=
fs
UA
(4)
does not exceed 175 kips per inch for interior exposure and 145 kips
per inch for exterior exposure."
FRP Reinforcement
605
Equation (4) will provide a distribution of the reinforcement bars that will
reasonably control flexural cracking. The equation is written in a form
emphasizing reinforcement details rather than crack width; yet it is based on
the Gergely-Lutz expression (equation 3). The numerical limitations of z are
175 and 145 kips per inch correspond to limiting crack widths of 0.016 and
0.013 in.
Crack Width Analysis of FRP Reinforced Beams
The ACI 224.2R-86 committee report (5) on cracking of steel
reinforced concrete beams recognizes that the expected value of the maximum
crack spacing is about twice that of the average crack spacing. Experimental
results on steel reinforced beams outlined by Bresler (6) have shown that the
average crack spacing value is about twice the cover thickness as measured to
the center of the reinforcing rebar. Therefore, the maximum crack spacing is
equal to about four times the concrete cover thickness. It is recognized by the
committee that crack width may be estimated by multiplying the maximum
crack spacing (four times the concrete cover) with an average strain in the
reinforcement.
The current ACI 224.2R-86 mathematical expressions for predicting crack
widths cannot be directly used to predict the crack widths in high strength
concrete beams reinforced with FRP rebars because of different material
properties associated with higher strength concretes and with FRP rebars. The
effects of the tensile strength of higher strength concrete, the bond strength
between concrete and FRP rebar, and the low modulus of elasticity of FPP
rebars are investigated to establish crack spacing and crack widths.
The low modulus of elasticity of the FRP rebar alone would alter the GergelyLutz expression (equation 3) for predicting crack widths for beams reinforced
with FRP rebars because the current expressions for maximum flexural crack
widths are proportional to the strain in steel reinforcement. The strains in
FRP rebars in the cracking zones are expected to be four times those of steel
because their modulus of elasticity is 7.2 x 106 psi compared to 29 x 106 psi
for steel. by substituting the steel stress with strain, the Gergely-Lutz
expression for beams reinforced with steel is rewritten as:
Wmax = 0.076 {3 Es es
Vd:A
X
10-3
where,
es = strain in the steel rebars.
Es = modulus of elasticity of steel = 29 x 106 psi.
(5)
606
Faza and GangaRao
In order to incorporate FRP rebar properties in the above expression, where
EsiEf ~ 4, substituting FRP strains, the Gergely-Lutz expression can be
modified as follows:
Wmax = 0.076 (3 (4Ef)
if{iA
x I0-3
(6)
w-3
(7)
=
0.30 (3 Ef ef ifd:A x
=
o.3o f3 ff ifd:A x w-3
(8)
where,
Ef = Modulus of elasticity of FRP rebar = 7.2 x 103 ksi
ff = FRP stress (ksi)
ef = Strain in FRP rebar.
Equation (8) would be valid under the assumption that the same crack spacing
is expected to occur while using FRP rebars in lieu of steel rebars. As
mentioned earlier, the crack spacing is governed by the tensile strength of
concrete and bond strength of the reinforcing bar. Therefore, an investigation
of the validity of equation (8), which was developed for regular concrete with
steel reinforcement, is carried out for cases when higher strength concretes are
used with the FRP rebars.
Watstein and Bresler (7) have investigated the relationship between the tensile
strength of concrete, the bond strength and the crack spacing, leading to the
crack width calculation. Their study prescribed the distribution of bond
stresses between adjacent cracks to calculate crack spacing and crack widths.
Assuming that bond stress is constant and when concrete reaches its ultimate
tensile strength, the crack spacing, 1 is defined by Watstein and Bresler as:
1 =(2ft' A) I ( llm
1r
D)
(9)
where,
ft'
= tensile strength of concrete, psi
llm
= maximum bond stress, psi
D
= rebar diameter, in.
The crack width may be approximated by an average strain in a FRP rebar
multiplied by the crack spacing, 1.
Wmax = (ff I Ef) I
(10)
FRP Reinforcement
607
Since crack spacing is governed by the bond stress between the FRP rebar and
concrete (equation 10), it is expected that a higher bond strength will lead to
smaller crack spacing and in turn will result in finer cracks. By examining the
results of the bond experiments (2), bond strength of sand coated rebars is
about two times that of steel rebars. Therefore, it will be expected that the
crack spacing multiplied with strain, would be half of crack width of steel
reinforced beams.
Better bond behavior of sand coated FRP rebars reduces crack spacing to half
the expected crack spacing in a steel reinforced beam. Equation (8) which is
based on multiplying crack spacing with stress overestimates the crack width
when high bond strength is developed with sand coated rebars. The quality of
the surface condition (deformations) and the quality of the sand coating may
vary without a set of standards. In order to accommodate all types of FRP
rebar surface conditions and to achieve a more realistic crack width
calculation, a crack width equation that incorporates the actual bond strength
of FRP rebars is needed, which has to be obtained from the experimental data
until the FRP rebar manufacturing process is standardized.
Based on the assumption that maximum crack width (equation 8) may be
approximated by an average strain in FRP rebar multiplied by expected crack
spacing, and by substituting expected crack spacing in equation (9), the
resulting expression for maximum crack width is shown in equation (11).
W max = (ff I Ef) (2 ft' A) I (Jlm
1r
D)
( 11)
Substituting for Ef = 7.2 x 1Q3 ksi
Wmax = 0.14 ff
Zli'_:!_ ><Io-3
J.lm :rrD
(12)
where,
ft' =
7.s.[i:'
ff = Maximum stress (ksi) in FRP reinforcement at service
load level with 0.5 fyf to be used if no computations are
available.
A = The effective tension area of concrete surrounding the
principal reinforcement divided by the number of rebars. It is
defined as having the same centroid as the reinforcement (in2).
608
Faza and GangaRao
To check the validity of these equations, the experimental crack widths for
beams reinforced with FRP rebars are compared with the theoretical
expressions (8 and 12) in Table 2. The experimental crack widths of beam #
E-D do not correlate well with both equations due to the high bond strength
between the concrete and sand coated rebar. The experimental crack widths
are smaller than the theoretical equations.
Comparisons in Table 2 reveal that equation (12) leads to a better expression
to predict maximum crack widths.
As shown in equation (8) the effects of high strains (that develop as a result of
low modulus of elasticity of FRP rebars together with higher bond strength
when sand coated FRP rebars are used) will produce crack widths that are two
times larger than in the steel reinforced beams. Use of sand coated FRP
rebars should permit higher tolerable crack widths, since the corrosion
problem using the noncorrosive FRP reinforcement is not a major concern. As
a result, the larger crack widths that might develop in the structures should be
considered acceptable and a new tolerable crack width limits may have to
specified.
BOND STRENGm AND DEVEWPMENT LENGTH
Bond strength is a function of:
1.
2.
3.
4.
5.
6.
Adhesion between the concrete and FRP rebars which is
controlled partly by the concrete strength and quality of the
rebars.
Gripping forces resulting from the drying shrinkage of the
surrounding concrete and the rebars.
Frictional resistance to sliding and interlock.
Mechanical anchorage of the rebar through development length,
splicing, hooks, and transverse rebars.
Diameter, shape (with or without wrapping, surface coating)
and spacing of the rebars.
Moment stress gradient in a given zone.
Anchorage and Adhesion Bond
Assume ld to be the length of a rebar embedded in concrete and subjected to
net pulling force dT. If db is the rebar diameter, p. is the average bond stress
and ff is the stress in FRP rebar due to bending in a beam which leads to a
FRP Reinforcement
pull-out force; the anchorage pulling force dT( = J1.
tensile force dT of the rebar cross section,
dT=ndb
4
2
1r
609
db ld) is equal to the
fr
(13)
hence, from equilibrium conditions,
(14)
where,
Jl. = Bond strength (psi)
db = Rebar diameter (in)
ld
Embedment length (in)
ff = Rebar stress (psi)
from which the average bond strength, p., derived as:
(15)
and the development length, ld
(16)
Basic Development Length
It has been verified by earlier tests on steel rebars that bond strength !J., is a
function of the compressive strength of concrete such that
(17)
where K = constant
If the bond strength equals or exceed the yield strength of a rebar with cross
sectional area of Ab = 1r db2t4 , then,
(18)
from Equation 18, the basic development length can be written as:
(19)
610
Faza and GangaRao
where kb is a function of geometric property of a reinforcing bar and the
relationship between bond strength and compressive strength of concrete.
ACI Approach
The development length of rebars in tension is computed as function of rebar
size, yield strength and concrete compressive strength.
The basic
development length is modified according to the requirements of the ACI 31889 code subsections 12.2.3.1, 12.2.3.2, or 12.2.3.3 to reflect the influence of
cover, spacing, transverse reinforcement, casting position, type of aggregate,
and epoxy coating.
The basic development length ldb· as specified by the ACI 318-89 code for
rebars size of #11 and smaller is:
(20)
Where, ldb must not be less than 12 in. and the square root of fc• must be less
than or equal to 100 which corresponds to concrete compressive strength of
10,000 psi.
Basic Development Length for FRP Rebars
For FRP rebars deformed with helical wrap of 45 degrees and sand
coated, the following experimental bond stress (JJ. = PI 1t db ld), values in
Table 3 are based on the average of the experimental results of the cantilever
test setup outlined by Faza and GangaRao (2). In addition a reduction factor
<j> = 0.75, is used:
Using the reduced bond strength values in Table 3, theoretical development
lengths is obtained using the following bond strength and development length
relationship given in equation (16) which results in a minimum development
length of ldb = 6.5 in. for #3 rebar and 10.38 in. for #4 rebar. The
theoretical embedment length calculated above using equation 16, with fyf
taken as the effective yield strength of FRP rebar, fc' taken as 10,000 psi, and
Ab as rebar cross sectional area (in2) are incorporated in the ACI basic
development length equation (20).
K = (6.5 x 100) I (0.11 x 104,00) = 0.057 for #3 rebar
K = (10.38 x 100) I (0.196 x 85,600) = 0.062 for #4 rebar
From the above calculation of the constant K, the ACI basic development
equation is modified, to account for the use of FRP rebars:
FRP Reinforcement
611
(21)
In the above calculations of the basic development length equation, a reduced
bond strength value is assumed. If the maximum experimental bond stress
value is used without a reduction factor ~. the constant K will become 0.0426
for #3 rebar and 0.0465 for #4 rebar.
The ACI 318-89 code provisions for the calculation of development length, ld,
include different safety factors that will increase the development length to a
safe limit. It is therefore recommended that the ACI equation be used for
FRP rebars with no alterations. Until more definitive stress levels are
established by testing many more specimens, the values in Table 3 will be
used for design purposes.
CONCLUSIONS
1. Both the ultimate strength deign method and the working stress (elastic)
design method for flexural design of concrete beams reinforced with FRP
rebars are acceptable. Design equations are developed to establish balanced
reinforcement for both the ultimate design theory and working stress theory,
so that engineers can design concrete beams reinforced with FRP rebars as
required by the ACI 318-89 provisions.
2. In order to estimate the maximum crack width of FRP reinforced concrete
beams, knowledge of the rebar bond strength with concrete is essential before
utilizing the proposed crack width equation (12). Otherwise, as a conservative
design practice, the crack width for beams reinforced with deformed FRP bars
may be estimated to be four times that of steel reinforced concrete beams.
3. The basic development length of FRP rebars should be computed using
equation (21 ), and utilizing the current ACI 318-89 code modifications for
development lengths without any changes.
REFERENCES
1.
Wu, W.P., Faza, S.S., and GangaRao, H.V.S., "Mechanical properties
of fiber reinforced plastic (FRP) bars for concrete reinforcement,"
Constructed Facilities Center Report, 1991.
2. Faza, S.S., and GangaRao, H. V.S., " Bending and Bond Behavior of
Concrete Beams Reinforced with Fiber Reinforced Plastic Rebars"
WVDOH-RP-83 Phase I Report, 1992.
612
Faza and GangaRao
3.
ACI Building Code Requirements for Reinforced Concrete (ACI 318-89),
American Concrete Institute, Detroit, MI, 1989.
4.
Halvorsen, G., Proceedings of the ACI Fall Convention, Lewis H.
Tuthill International Symposium on Concrete and Concrete Construction,
pp. 104, 1987
5.
ACI 224.2R-86, State-of-the-Art report on High Strength Concrete,
American Concrete Institute, Detroit, Ml, 1986.
6.
Bresler, B., "Reinforced Concrete Engineering", Volume 1, Materials,
Structural Elements, Safety, Jjohn Wiley and Sons, New York, 1972.
7.
Watstein, D., and Bresler, B. " Bond and Cracking in Reinforced
Concrete" Reinforced Concrete Engineering, Vol 1, John Wiley and
Sons, New York, 1974.
TABLE 1 - THEORETICAL VERSUS EXPERIMENTAL MOMENT EVALUATIONS
Group
Beam
I
I
Reinf
Area
Reinf
Cone.
Exp.
Ult.
Observed
Calc.
Calc.
Calc.
RATIO
RATIO
RATIO
Sq.
Stress
Strength
Rupture
Tensile
Ult.
Moment
Moment
Moment
Observed
Observed
Observed
in
lksil
!psi)
Tensile
STRENGTH
Moment
USING
USING
USING
vs
vs
vs
Strength
lksi)
K·ft
col 5
col 7
col 8
Calc.
Calc.
Calc.
stress
stress
stress
Moment
Moment
Moment
lk·ftl
lk·ftl
!k·ftl
9 vs 10
9 vs 11
9 vs 12
12
13
14
15
lksil
0.8
Rupture
Strength
IRECT.I
1
2
3
A
2
A
3
A
4
5
313
0.33
313
0.33
VH4
612
A
VH5
B
B
B
5
c
c
c
c
c
6
7
1
4
8
H5
c
6
7
8
102
4200
130.00
76
4200
90.00
0.30
85
10000
313
0.33
72
214
214
214
217
218
217
218
218
0.39
0.39
0.39
9
10
11
104.00
27.75
28.09
35.80
28.64
0.99
0.78
0.97
72.00
24.67
21.26
25.29
20.23
1.16
0.98
1.22
130.00
104.00
31.50
22.92
34.93
27.95
1.37
0.90
1.13
10000
130.00
104.00
21.00
20.79
37.78
30.23
1.01
0.56
0.69
84
4200
80.00
64.00
27.75
27.28
25.96
20.77
1.02
1.07
1.34
69
4200
80.00
64.00
24.67
22.54
26.33
21.06
1.09
0.94
1.17
72
5000
80.00
64.00
16.96
23.80
26.52
21.22
0.71
0.64
0.80
1.20
15
4200
86.00
68.80
16.50
15.58
87.02
69.62
1.06
0.19
0.24
1.57
45
4200
80.00
64.00
40.00
52.32
92.55
74.04
0.76
0.43
0.54
1.20
0
5000
80.00
64.00
41.63
0.00
84.40
67.52
0.49
0.62
1.56
47
6500
80.00
64.00
54.75
57.72
97.63
78.10
0.95
0.56
0.70
1.56
43
6500
80.00
64.00
60.00
52.91
98.71
78.97
1.13
0.61
0.76
614
Faza and GangaRao
TABLE 2 - THEORETICAL VERSUS EXPERIMENTAL CRACK WIDTHS
Beam#
E-VH1
E-VH1
E-VH1
E-VH1
E-F
E-F
E-D
E-D
FRP
Stress
(ksil
8.50
11.30
17.20
37.50
54.00
71.00
32.90
49.50
Exp.
Crack
Width (in)
0.011
0.017
0.02
0.041
0.077
0.096
0.021
0.046
Equation #8
Crack
Width (in)
0.01
0.012
0.018
0.042
0.06
0.08
0.036
0.054
Equation #12
Crack
Width (in)
O.Q11
0.015
0.022
0.049
0.071
0.093
0.04
0.065
TABLE 3 - AVERAGE EXPERIMENTAL BOND STRESS, FORfc' = 10 ksi
Average
Exp. Bond
Strength, psi
Reduction
Factor
Reduced Bond
Strength
,.,.
<P
p.=</Jp.
#3 Sand Coated
2000
0.75
1500
#4 Sand Coated
1375
0.75
1035
FRP Rebar
Size
SP 138-37
Shear Capacity of RC and
PC Beams Using
FRP Reinforcement
by S. Tottori and H. Wakui
Synopsis: Utilizing FRP reinforcement for concrete guideway
structures in the superconductive magnetically levitated train
system(JR MAGLEV) is considered desirable because FRP
reinforcement is diamagnetic. For the design of the guideway
structures using FRP reinforcement, performance of structural RC
and PC members must be clarified. Flexural behavior of these
members can be predicted by conventional design procedures
taking the mechanical properties of FRP reinforcement into
account.
But shear resisting behavior of RC and PC members has not
been clear yet. The reasons are considered as follows:
i )Shear resisting behavior is complicated in itself unlike
flexural behavior.
li )Experimental equation for shear capacity of RC members
using reinforcing steel does not seem to be applicable
because mechanical properties such as Young's modulus,
elongation and the like are different from those of
reinforcing steel.
Under such circumstances, the authors carried out a basic
experiment on shear capacity of rectangular beams using FRP
tendons and FRP shear reinforcement. As the result, the
following are made clear:
)Shear capacity of RC beams without shear reinforcement can
be predicted to some degree by taking account of the tension
stiffness of FRP reinforcement.
li )It seems to be possible to predict contribution of
prestress to shear capacity from decompression moment.
ill )Contribution of FRP shear reinforcement to shear capacity
is smaller than the value calculated by truss analogy. The
reasons seem to be related to the experimental results that
the maximum strain value of FRP shear reinforcement at
shear failure is smaller than the elongation of FRP
reinforcement.
Keywords: Beams (supports); fiber reinforced plastics; flexural strength;
mechanical properties; prestressed concrete; reinforced concrete; shear
properties
615
616
Tottori and Wakui
Seiichi Tottori is a chief researcher of the structural
laboratory in Railway Technical Research Institute(RTRI) Tokyo,
JAPAN, and he received his civil engineering master's degree at
Waseda University, Tokyo, Japan. He is interested in the
development of MAGLEV guideway using advanced materials.
Hajime Wakui is the chief of the structural laboratory in
Railway Technical Research Institute(RTRI) Tokyo, JAPAN, and he
received his civil engineering master's degree at Tokyo
Institute of Technology, Tokyo, Japan. He is interested in the
impact problem on concrete sleepers and limit state design of
railway concrete structures.
INTRODUCTION
Recently the research and development on MAGLEV has been
accelerated and a new test line is currently being constructed
to conduct the final assesment to put it to practical use.
Civil engineers involved in this project are now trying to meet
magnetoelectric requirements in the design of guideway system.
Fig. 1 shows a schematic view of the new guideway system
proposed by the authors. It should be noted that this system
has precast prestressed concrete girders with minimum quantity
of reinforcing steel to satisfy the above-mentioned requirement.
Moreover, in order to make the influence of magnetic field
smaller, the authors are investigating how to replace the
reinforcing steel in the girder with fiber reinforced
plastics(FRP).
FRP tendons with good bond properties to concrete are to be
used in the pretensioning system.
Flexural behavior of PC
girders using FRP tendons can be predicted by conventional
design procedures taking the mechanical properties of FRP
reinforcement into account.
As for shear reinforcement, spiral FRP rods are desirable
in terms of easiness of manufacturing and placing the
reinforcement. But shear capacity of the girders using FRP
reinforcement has not been clear. Then, the evaluation method
of shear capacity is studied using the data of loading tests
carried out by the authors(!) and by the other researchers.
FRP Reinforcement
617
PREMISE FOR EVALUATING SHEAR CAPACITY
As the tension stiffness and bond properties of FRP
reinforcement are different from those of reinforcing steel, it
has been pointed out that shear capacity of RC and PC beams is
smaller than beams using reinforcing steel. This paper stands
on the premise that shear capacity V could be expressed as a
sum of the resistance carried by concrete Vc, the effect of
prestressing Vp and the resistance carried by shear
reinforcement Vs . That is, shear capacity V is calculated by
equation(!).
V=Ve +Vp +Vs
( 1)
SHEAR CAPACITY CARRIED BY CONCRETE
Equations for shear capacity carried by concrete
Equations for shear capacity of reinforced concrete beams
and deep beams without shear reinforcement have been proposed
by Niwa et al. as follows(2).
Shear capacity Ve1
reinforcing steel
Vel=200fe 1/ 3
•
Pw 1/ 3 • d- 1 /
4 •
of reinforced concrete beams using
{0. 75+1.4/(a/d)} • bwd
(kN)
(2)
Shear capacity of Ve2 reinforced concrete deep beams using
reinforcing steel
Vc2=244fc 2/ 3
•
0+/P..) • {1+3.33(r/d)} bwd/ {l+(a/d) 2}
(kN)
(3)
In equations (2) and (3)fe
compressive strength of concrete (Mpa)
Pw
tension reinforcement ratio (=lOOAs /(bw d)J
As
area of tension reinforcement (m 2)
a
shear span (m)
d
ef feet i ve depth (m)
bw
width of web (m)
When Ve2 is larger than Vel, the beam is considered to be
a deep beam and shear capacity Ve is defined as Ve2.
However, in case of using FRP reinforcement, it has been
pointed out, shear capacity is smaller than the value obtained
from the equation (2) or (3). One of the reasons for abovementioned results is that tension stiffness of FRP
618
Tottori and Wakui
reinforcement is smaller than that of reinforcing steel. Then,
it is suggested that shear capacity can be reasonably evaluated
mu 1t i p 1y i n g p w b y (EriEs).
Wh e r e E t a n d E • a r e Yo u n g ' s
modulus of FRP reinforcement and reinforcing steel,
respectively(3).
It could be considered that shear capacity carried by
concrete Vc is equal to a sum of the shear carried by the
compression zone, aggregare interlock and dowel action of
longitudinal reinforcement. Shear carried by the compression
zone and aggregate interlock is considered to be related to the
tension stiffness of the longitudinal reinforcement. Therefore,
it might be rational to a degree to multiply Pw by (EriEs).
However, the shear carried by dowel action of the
longitudinal reinforcement is related to its own flexural
stiffness. As flexural stiffness of FRP reinforcement is much
smaller than that of reinforcing steel as shown in Table 1. it
is not clear whether the shear carried by dowel action can be
evaluated multiplying Pw by (EriE.).
Then, the following tests
concerning dowel action were conducted.
Dowel action of FRP reinforcement
The specimens and the method of loading are shown in Fig.2.
These test conditions are set up with reference to tests
carried out by Krefeld and Thurston(4). Table 1 shows the
mechanical properties of reinforcement used in the loading
tests. Bond properties of FRP reinforcement in Fig.2 are
almost equivalent to those of reinforcing steel. Cracking
patterns and dowel capacity D are shown in Figs. 3 and 4,
respectively. Dowel capacity D is defined as a quarter of the
ultimate load Pt.
From Fig.3, it seems that flexural stiffness of the
longitudinal reinforcement affects the cracking patterns. That
is, cracking does not reach the support of the beam in case of
using FRP reinforcement such as CFRP rods.
Fig.4 shows the relationships between dowel capacity D and
the distance from the surface of the specimen to the center of
the longitudinal reinforcement.
In Fig.4, test results by
Furuichi et al. (5) are shown by dot lines with above-mentioned
results.
In their tests, dimensions of the specimen are the
same as in Fig.2.
But the method of loading is slightly
different from Fig.2. That is, the load is applied not only
from the web of the specimen but also from the upper surface.
By simply comparing these test results without regard to the
difference of the method of loading, the following points are
clarified from Fig.4:
FRP Reinforcement
619
i )In case of using reinforcing steel. dowel capacity 0
increases as the distance C becomes larger.
On the
contrary, in the case of using FRP reinforcement, dowel
capacity 0 seems not to be affected by the distance C.
nevertheless, CFRP rods do not break at all.
ti) As far as C is not larger than 5cm, dowel capacity of the
specimen using FRP reinforcement is about 70% as large as
that of the specimens using reinforcing steel. This ratio
happens to correspond to (EriEs) 1 / 3 •
Therefore, it seems
i n e q u a t i ons ( 2 ) a nd ( 3 )
t h a t mu 1t i p 1y i n g p w b y (E dEs )
approximately evaluates the effect of dowel action on shear
capacity Vc.
Accuracy of equation for shear capacity carried by concrete
Through above-mentioned study, shear capacity of RC beams
using FRP reinforcement without shear reinforcement can be
expressed as follows:
Shear capacity Vet of reinforced concrete beams using FRP
reinforcement
Vct=200fc 1 /
3 •
(Pw • Er/Es)
l/S •
d- 1 /
4 •
{0. 75+1.4/(a/d)} bwd
(kN)
(4)
Shear capacity of V c 2 of reinforced concrete deep beams
using FRP reinforcement
Vc2=244fc2/ 3
•
{l+Jpw(EdEs)} · {1+3. 33(r/d)} bwd/ {l+(a/d) 2} (kN)
(5)
Then, accuracies of the equations(4) and (5) are examined.
Fig. 5 shows the relationships between the ratio of experimental
values of shear capacity to caluculated ones and various
Par am e t e r s such as Pw (E r/E s) , aId and f c . The aver age r at i o
is 1.12 and the coefficient of variation is
of Vc,test/Vc,cal
13.8% for 22 specimens. Through these results, it is clear
that shear capacity V c can be evaluated on the safe side by
equation (4) or (5).
EFFECT OF PRESTRESSING V.
Equation for the effect of prestressing
The effect of prestressing V. could be evaluated by the
following equation.
V•
= 20 (Mo/a)
(kN)
(6)
620
Tottori and Wakui
where Mo is decompression moment(kN · m). In case of the
rectangular cross section, equation(6) can be transformed into
equation(?).
v.
bh
p
3bh
{l+(e/h)}
a/h
(7)
where:
h: height of the specimen
b: width of the specimen
P: prestressing force
e: eccentricity of prestressing force
Therefore, the effect of prestressing could be expressed as
a function of the average compressive stress by prestressing
P/bh, the ratio of the shear span to the height of the specimen
a/h, and the ratio of the eccentricity of prestressing force to
the height of the specimen e/h.
Accuracy of equation for the effect of prestressing
v.
Accuracy of equation(6) or (7) is examined by the test
results of PC rectangular beams using steel or FRP tendons
without shear reinforcement.
Fig. 6 shows the relationships between the ratio of
experimental shear capacity to calculated one and the abovementioned parameters such as P/bh, a/h and e/h. Calculated
shear capacity of PC beams is defined as a sum of Vc by
equation(4) or (5) and V. by the equation(6).
Although the equation(6) tends to slightly overestimate v.
in case of P/bh being bigger, shear capacity seems not to be
affected by the kinds of prestressing tendons. The average
value of Vtast!Vcat
is 0.96 and the coefficient of variation is
12.5% for 43 test specimens.
SHEAR CAPACITY CARRIED BY SHEAR REINFORCEMENT
Strain distribution of shear reinforcement and carrying process
of shear force
Shear tests of RC beams were performed on several specimens
whose parameters were types of spiral FRP rods as shear
reinforcement(see Fig. 7). As for FRP rods, Glass Fiber
Reinforced Plastics(GFRP), Aramid Fiber Reinforced
Plastics(AFRP), Carbon Fiber Reinforced Plastics(CFRP), Vynylon
Fiber Reinforced Plastics(VFRP) were taken up. ·Table 2 shows
FRP Reinforcement
621
the mechanical properties of these FRP rods. From the strain
distribution at the ultimate stage in the loading test and the
carrying process of shear force by shear reinforcement shown in
Figs.8 and 9, respectively, following points are clarified.
Shear force carried by shear reinforcement was measured by
strain gauges attached to the shear reinforcement.
i )Maximum strain value was more than 10- 2 at the ultimate
stage of the loading tests, but it does not reach the value
of elongation at the break of FRP rods.
ti )Increasing tendency of shear force carried by shear
reinforcement after diagonal crack seems to be predictable
by the truss analogy shown in equation(8).
vs =
AsfEsf
s
E
sf(sin a +cos a) z
(8)
where:
Asf
area of shear reinforcement
Esf
Young's modulus of shear reinforcement
e sf
strain of shear reinforcement
a
angle between reinforcement and member axis
s
spacing of shear reinforcement
z
d/1.15
d
effective depth
Evaluation method of shear capacity carried by shear
reinforcement
In RC members using reinforcing steel, it is likely that
shear reinforcement yields at the ultimate state. Therefore,
shear capacity carried by shear reinforcement is calculated by
replacing Esresf with yield strength f Y in equation(8).
Where fy is yield strength of reinforcing steel as shear
reinforcement.
To calculate V. in equation(8), it is necessary to
estimate the Value Of E: sf. From the data on loading tests, the
value of E sf is estimated by equation(9).
bz
where:
b
width of web
= [Esf Asf (sin a+ cos a )/bs].
=EsfPsfCsf
(9)
E
sf
The reI at i onsh ips bet ween V./bz and EsrPsr are shown in
Fi g. 1 0. Fr om t hem, i t f o l l ow s that Vs/bz i s p r o por t i on a I to
EsfPsr
as far as EsrPsr
is limited to a certain value. The
average value and the coefficient of variation of E sr are 0. 01,
28. 8%, respect i v e l y.
Vslbz seems to converge to a c e r t a i n
value as EsfPsr becomes bigger. This phenomenon seems to be due
622
Tottori and Wakui
to that the test specimen tends to crush in a shear-compression
failure as the quantity of shear reinforcement increases.
Fig,11 shows the relationships between the value of e sf
obtained by loading tests and parameters such as a/d, 100Af/bwd
and f c. From these relationships, it seems that there is no
correlation between e sf and above-mentioned parameters, and
that the value of
e sf is about 0.01 as far as the test
specimen breaks into diagonal tension failure.
Accuracy of equation for shear capacity
The calculated value of shear capacity of RC beams is
defined as sum of V c and V s. V c is obtained by equation(4)
or (5). Vs is obtained by substituting 0.01 fore sf in
equation(9).
Fig, 12 shows the relationships between the value of
Vtest!Vcal and Vc/Vcal. The value of Vtest!Vcal approximately
ranges from 0.8 to 1.2. The average value of Vtest!Vcal is 1.01
and the coefficient of variation is 11.5%.
CONCLUSIONS
Conclusions of this investigation are as follows:
i )Dowel capacity of the test specimens using FRP
reinforcement is about 70% of those using reinforcing steel
with almost the same diameter.
This value is about
e qu i v a I en t t o (E r/E.) 1 / 3 •
li )Shear capacity carried by concrete is evaluated by
equation(4) or (5) taking into account the tension stiffness
of longitudinal reinforcement.
ill )It seems to be possible to predict the effect of
prestressing on shear capacity from decompression moment in
case of using FRP tendons as well as in case of using steel
ones.
iv ) I n case o f the d i ago n a I t en s i on f a i I u r e, she a r c a pa c i t y
carried by shear reinforcement is evaluated from
equation(6) by substituting 0.01 for e sf.
REFERENCES
1. Miyata, S., Wakui, H., Tot tori, S. and Terada, T., "Shear
capacity of PC beams with spiral FRP reinforcement''
Proceedings of FIP-XIth International Congress on Prestressed
Concrete, 1990
FRP Reinforcement
2.
Niwa, J., Yamada, K.,
''Revaluation of Equation
concrete Beams without Web
International of JSCE No.9.
623
Yokozawa, K. and Okamura, H.,
for Shear strength of Reinforced
Reinforcement,'' Concrete Library
1987
3. JSCE Research Subcomittee on Continuous Fiber Reinforcing
Materials .• '' Application of Continuous Fiber Reinforcing
Materials to concrete structures'', Concrete Library
International of JSCE No.19, 1992
4. Krefeld, W.J. and Thurston, C.W .• "Contribution of
Longitudinal Steel to Shear Resistance of Reinforced Concrete
Beams'' Journal of the American Concrete Institute, Proc.
V. 63, Marach 1966
5. Furuuchi. H., Kakuta, Y., ''Study on the mechanical model of
dowel action of reinforcing bar" (in Japanese), Proceedings
of the 42th Annual Conference of Japan Society of Civil
Engineers. Part 5, 1987
624
Tottori and Wakui
TABLE 1 - MECHANICAL PROPERTIES OF TENSILE REINFORCEMENT
Reinforcement
Area
(mm 2 )
Young's
modulus
(Gpa)
1
Rebar (D13)
126. 7
206
26. 0
2
Rebar (D16)
198. 6
206
64. 6
3
CFRP rod
(¢
12. 5)
76. 0
137
0. 6
4
CFRP rod
(¢
15. 2)
113. 6
137
1.5
No.
Flexural
stiffness
(Gpa)
TABLE 2 - MECHANICAL PROPERTIES OF SHEAR REINFORCEMENT
Reinforcement
Area
(mm 2 )
Tensile Young's
strength modulus
(Mpa)
(Gpa)
Elongation
at break
(%)
GFRP rod
[GF]
37
696
38
1.8
AFRP rod
[AF]
30
1088
57
2. 2
[CF(l)]
17
1412
110
1.3
[CF(2)]
10
1745
137
1.5
[CF (3)]
18
1245
92
1.4
[VF]
41
598
36
3. 0
CFRP rod
VFRP rod
FRP Reinforcement
625
Precast PC girders (L=l2. 6 m)
Fig. 1-U-shaped guideway system consisting of precast PC girders
100
Fig. 2--Specimen configuration (unit in mm)
626
Tottori and Wakui
Rebar (D13)
CFRP rod(¢12.5)
Fig. 3-Cracking patterns of the specimens
z
12
-=:
10
....,>,
B
0: Rebar (D13)
t::.:CFRP rod (¢12.5)
/
()
""c.
""
6
()
~
4
Q)
3:
0
Cl
_..--"'-Rebar (016).
_....a"
~=-------~4-y\----------~
2
0
3
4
A
CFRP rod ( ¢ 15. 2)
5
6
7
C (em)
Fig. 4-Dowe] capacity
8
FRP Reinforcement
1 .6
-.
§0
0
0
u
1 .0
>
80
Do
0
0
0
0
0
.
..,
""""
+'
>
0
2
3
4 .5
a/d
1 .6
-.
0
0
u
~ 1 .0
..,
OJ0
0
oo
0
0
DO
"""."
+'
>
0 0.2
0.4
0.3
Pw •
-.
(EriEs)
(%)
1 .6
0
8
u
;
0.6
0.5
0
0
1 .0
0
0
~o8
0
..,
"""."
0
0
+'
>
0
20
30
40
fc
50
60
(Mpa)
Fig. 5-VresrfVcal (for RC beams without shear reinforcement)
627
628
Tottori and Wakui
l .6
. l .0
>
u
0
.•
0
--....
0
0
0
0
0
8
10
~
0
0 : Steel tendons
+ : FRP tendons
>
0
0
6
4
2
(Mpa)
P/bh
1 .6
. 1 .0
+
Bo
u
0
+
0
>
--....
.
0
0
~
~
0: Steel tendons
>
0
1.4
+ : FRP
tendons
1.8
2.2
2.6
3.4
3.0
a/h
1 .6
8 B
-• l .0~----------~~h~----~~--~
:j: c 0
:j:
>
--....
0
0
0 : Stee 1 tendons
+ : FRP tendons
0.1
0.2
0.3
0.4
e/h
Fig. 6-J"tes/Vcal (for PC beams without shear reinforcement)
FRP Reinforcement
3100
Tensile reinforcement
Shear reinforcement
0 : Strain gauages
Fig. ?-Specimen configuration (unit in mm)
0.014
0.010
0
40
80
Longitudinal distance from the
center of the specimen (em)
120
Fig. 8-Strain distribution of shear reinforcement
629
630
Tottori and Wakui
z
:::=.
Q)
....tJ
.....0
....
"'
...<::
Q)
50
tr.l
0
100
200
300
400
Applying load (kN)
Fig. 9-Carrying process of shear force
Fig. 10-The relationship between
~/bz
and Esl'sf
FRP Reinforcement
631
0.02
+
0
0
X
!
~
~
0.01
v
v
v
~
~~
+=W,
O"(F,
0
~=cFCil,
ll.=cf(2),
X=Cf(3) I
4
3
2
v
v~l
~
Notes:
J
4.4
a/d
0.02
+
8
X
~
tt
; o. 01 r+:....;sr--;!----------v
~
0
I
v
Notes:
+=w l
o~CT I
~=CF
c1 1I ll.={;f (2) ' X=CF (3) ' V=U'
I
o_~---7----~----~--~----~
0
2
3
4
A r / b d (Er /Es )(%)
5
0.02
+
..
"'
8
X
0.01
!+ v s»
v ~~
Notes:
• D=-<F,
30
v
~
vj
0
+
v
+=r=F< =CHl>,
L>=cF(2),
50
X={F(3),
v~l
70
fc (Mpa)
Fig. 11-The relationship between
Esf
and the value of each variable
632
Tottori and Wakui
1.6
-> 1.0
">.."
+
u
+ "'v
X
X
+
v
"'./>(
v
0
0
4+-81>
v
v
v v
v
~
/ Notes:
[] Ci'
0
0.2
+ r:F
o a=m
0.4
"' Cf"(2)
0.6
X Cf"(3)
v
~l
0.8
V c /V cal
Fig. l2-V,es11Vcal (for RC beams with shear reinforcement)
SP 138-38
Principles of Design of FRP
Tendons and Anchorages for
Post-Tensioned Concrete
by F.S. Rostasy and H. Budelmann
Synopsis: The post-tensioning of concrete members is a promising field of application of FRP tensile elements. Thereby the
high tensile strength and good corrosion resistance of FRP can
be utilized to compete with prestressing steel tendons. In order to be able to compete successfully suitable anchorages for
FRP have to be developed. The report demonstrates the interactive process of experiments and theoretical models for the development of tendon-anchorage assemblies with a high mechanical
efficiency. This process is then illustrated for the example of
a resin-bonded GFRP anchorage.
Keywords: Anchorage (structJral); fiber reinforced plastics; models; post
tensioning; prestressed concrete; prestressing steels; structural design; tensile
strength; tests
633
634
Rostasy and Budelmann
Ferdinand S. Rostasy, born 1932, studies of civil engineering
at the University of Stuttgart; Dr.-Ing. 1958; practical work
and research until 1976, from then on professor of structural
materials at Braunschweig, Germany.
Harald Budelmann, born 1952, studies of civil engineering at
the Technical University of Braunschweig, Ph.D. 1986, senior
research engineer at the TU Braunschweig until 1991, today professor of structural materials at the University of Kassel ,
Germany.
INTRODUCTION
For the post-tensioning of concrete structures the tendons conventionally consist of high-strength prestressing steel elements. For such tendons many types of suitable tendon-anchorages have been developed over the past decades. A great amount
of experience has been gathered with respect to the mechanical
behaviour and practical suitability of the anchorages. Because
of their high strength and excellent corrosion resistance in
many environments, FRP can compete with prestressing steel.
However, in view of the high costs of FRP and because of their
sensitivity against lateral pressure, surface injury etc. new
methods have to be found to transfer the force of the individual FRP element to the anchorage. Concrete industry and research are jointly developing efficient and economic anchorages
for FRP tendons by testing and anal yt i cal modelling. The interactive process of design and experiment will be demonstrated
for the bond anchorage of GFRP rods.
MECHANICAL PROPERTIES OF FRP
Short-term axial tensile strength
FRP tensile elements for post-tensioning tendons are fabricated
as round bars or wires, as braided wires or as strands. The
endless fibers are unidirectionally arranged in the element and
are embedded in a polymeric matrix resin. If FRP are to compete
successfully with high-strength prestressing steel, their axial
tensile strength must be high. This demand can be easily met,
as the glass fibers, aramid fibers and carbon fibers available
today, can be produced to achieve a high tensile strength.
Table 1 presents an overview on the main mechanical properties
FRP Reinforcement
635
of high-strength fibers. Comparison is drawn with prestressing
steel wire.
The fibers are purely elastic materials. Experiments have shown
that the axial tensile strength of an FRP element depends on
the fiber strength and on the fiber volume fraction vf of the
composite section AC' In Fig. 1 the stress-strain lines of the
components and of a specific composite (e.g. GFRP 0 7,5 mm bar,
Polystal) are depicted. The common matrix resins EP and UP do
not significantly cant ri bute to the axial strength f ct and to
the axial Young's modulus Ect of the composite. In spite of
this fact the matrix has several
important functions
(protection of fibers against detrimental mechanical and corrosive effects, etc. (1)). As Fig. 1 shows the behaviour of FRP
is ideal-elastic and brittle. The ultimate tensile strain t:cu
corresponds to that of the fibers, t:fu· The strength of FRP may
easily reach 2 to 3 GPa, hence exceeding prestressing steel.
The ultimate tensile strain is considerably less than that of
prestressing steel.
Table 2 contains an overview on the main mechanical properties
of some commercial FRP elements from producer sources. It is
not known whether the values are mean or characteristic ones.
Other data are given in (1).
Multiaxial short-term strength
The axial force of the individual FRP element has to be transferred to the tendon's anchorage. Irrespective of the design of
the anchorage, the force transfer entails transverse pressure
and axial shear stresses, both acting on the circumference of
the FRP element. Hence, within the anchorage a state of multiaxial stresses exists. Only little is known about the multiaxial strength of FRP, and experimental verification is extremely difficult (1), (2).
Tests show that simultaneous shear and compressive stresses
acting on the surface of the FRP element reduce the axial
strength. It has to be pointed out that such a condition prevails at the entry of and within any type of anchorage for FRP
bars. It is practical to investigate the specific FRP material's susceptibility to such condition in the tendon-anchorage
test.
Static long-term strength
In a pjc-structure the FRP element is subjected to a fairly
constant tensile stress due to prestress with the stress fluctuations caused by live load superimposing. Hence, the static
636
Rostasy and Budelmann
long-term strength must be known. Extensive testing becomes necessary. In (1) an overview is presented.
FRP elements exhibit the phenomenon of creep rupture when subjected to a constant tensile force: The higher the force in relation to the short-term tensile strength, the shorter the
endurance time t until failure will be. The aim of the tests
is the determina~ion of the characteristic endurance line F~lk
(tu) as the relevant resistance of the FRP material which nas
to cover the assumed service life span ts of the structure. The
creep rupture behaviour is greatly influenced by the environment the FRP element is in contact with.
Dynamic strength
The dynamic strength of FRP must be known if the tendon is subjected to frequent stress variations caused by live loads. In
(1) an overview on test results is given. Specific results will
be shown later.
With respect to the use of FRP for tendons several other,
partly non-mechanical properties of the material are important
( 1) .
REQUIREMENTS FOR THE PERFORMANCE AND TESTING
OF POST-TENSIONING TENDONS
Whether a tendon consists of prestressing steel elements or of
FRP elements, both the tendon and the tendon-anchorage assembly
TA have to meet several requirements. These requirements and
test rules for prestressing steel tendons are given in (3). For
FRP tendons i nternat i anal test rules and performance criteria
do not yet exist. FIP recently submitted the following requirements (1):
• The axial tensile strength fct of the FRP elements should not
be significantly reduced by anchorage effects.
• The long-term static stresses (e.g. prestress) and the dynamic stress amplitudes of the service load state should not
reduce the original tensile strength fct of the FRP.
• Adequate creep rupture strength and dynamic strength must be
ensured.
• Environmental effects must not significantly
strength of the FRP during service life.
reduce
the
FRP Reinforcement
637
These requirements are deduced from the demands of structural
safety. However, also economy plays a vital role. Whether a
specific tendon-anchorage assembly meets these requirements can
only be verified by tests although mechanical models may also
assist in the design of theTA. In (1), (4) the test procedures
for the mechanical properties of the material on one hand, and
on the other hand for the short-term rupture force, for the
creep rupture strength and for the dynamic strength of the FRP
tendon are described. In these tests both ends of the test tendon are provided with the TA, with the tendon having a length
of 3 m or more.
Short-term rupture strength of the TA (Tendon-Anchorage Assemb-
ill
The extent to which the axial tensile strength fct of the FRP
element with the cross-section Ac can be utilized in the test
can be expressed by the anchorage efficiency factor:
meas FTu
( 1)
cal Fem
with:
meas FTu- .. measured rupture force of the tendon-anchorage assembly
cal Fem .... calculated mean rupture force of the tendon - consis t i ng of n FRP elements - without anchorages. It
is defined by:
cal Fern
(2)
Prior to the tendon tests the mean axial tensile strength f ctm
of the FRP material must be established. The results of the
tendon tests must satisfy according to (1):
meas nAm
~
0.95
(4)
The scatter of results, expressed by coefficient of variation
Vr, should not significantly exceed that of the material V •
The requirement (4) ensures that the strength FRP is optimally
exploited. This requirement is also essential in view of the
economic competition with prestressing steel, because FRP are
still expensive materials. It also ensures that an adequate
ductile deformation of the post-tensioned bending concrete member can be attained. Results will be presented later on.
638
Rostasy and Budelmann
Static long-term rupture strength of the TA
It was shown in (1) and (4) that the long-term static tensile
strength of both the FRP material and the TA are the relevant
resistances for the choice of the admissible prestressing force
adm Pm 0 at initial stressing. Suitable procedures of testing
and evaluation of data are described in (1). Tests have shown
that a tendon anchorage assembly with a high efficiency TJA in
the short-term rupture test will also exhibit a satisfactory
creep rupture behaviour.
Dynamic strength of the TA
With the permissible prestressing force P~ 0 as the upper force
the &-N curve of the TA has to be establ l'sned over a range of
2·10 load cycles. In sec. 4 some results will be presented.
TEST RESULTS
On Anchorage Development
In view of the fact that all FRP materials react very sensitively to high transverse pressure and to surface notches, the
common ways to anchor prestressing steel (e.g. toothed steel
wedges, threading, etc.) have to be discarded for FRP. An efficient utilization of the FRP strength calls for the controlled,
soft reduction of the force beginning at the entry of the FRP
element into the anchorage without surface damage. In (1), (5)
and (6) the state of anchorage development is outlined.
Such controlled reduction can be achieved by several methods
which will be briefly discussed:
The FRP element can be clamped along their axis between steel
plates and intermediary protective sheets. Transverse pressure is created by prestressed bolts. Anchorage mechanism is
of the shear friction type (6).
• The FRP element is bonded by a resin mortar to a steel sleeve
with inner thread: bond anchorage (7).
• The FRP element can be encased by a tightly fitting, resinfilled steel tube which can then be gripped by a toothed
steel wedge (5).
Also other mechanisms or combinations of several can be applied. In this report only the bond anchorage, specifically the
FRP Reinforcement
639
HLV (Hochleistungsverbund) anchorage of 6 to 19 7,5 mm GFRP
bars (Pol ysta l) of STRABAG will be discussed. This anchorage
has been successfully used for several structures (7).
Anchorage efficiency
Fig. 2 schematically shows the HLV anchorage. Numereous static
tensile rupture tests on tendons have been performed. Optimization of the anchorage design led to a high and consistent anchorage efficiency of 77Am :::: 0.96 with ultimate strains in the
range of 3 %. It is interesting to note that the coefficients
of variation of the axial tensile strength of the FRP and of
the rupture force FTu of the TA were very similar and about
2 %.
Failure commences at the entry of the GFRP bars into the steel
housing with delaminations proceeding into free length of tendon. Fig. 3 shows the TA after failure.
Creep rupture strength of the TA
In Fig. 4 the creep rupture force results vs. failure time of
long-term static tests with HLV tendons (8 or 19 bars) are
plotted. The GFRP bars stem from different production lots. The
tendons were loaded with a constant force. The l eve 1 of force
was chosen as fraction of the tensile rupture force, Eq. ( 2).
The time to failure of the first bar was recorded.
By statistical evaluation of the failure times a lower boundary
line FTlk of the TAwas estimated (characteristic resistance).
For the sake of comparison also the characteristic endurance
line Fc]k(t) of the GFRP bars (anchored in segmented clamping
plates t6-)), thin line, are drawn. There is a rather small difference between the long- term behaviour of the material and
tendon-anchorage assembly.
The creep rupture test must be performed in a test environment
which corresponds to the one of the practical application. The
environment is of great influence (1).
Dynamic strength of the TA
In Fig. 5 the S-N-lines of single bar bond anchorages for a variable upper stress a0 are depicted. Very similar results were
achieved for tendons with 8 to 19 Polystal bars 0 7,5 mm.
640
Rostasy and Budelmann
MECHANICAL MODELS FOR THE DESIGN OF ANCHORAGES
Purpose of models
To base the design of a tendon anchorage exclusively on full
scale prototype testing is a too cumbersome and too costly way.
More efficient is - in the initial phase of development - an
interactive approach by combination of mechanical modelling and
the testing of representative small-size prototypes for model
verification. In the final phase of development the testing of
multi bar anchorages is i nevi table as the final proof of efficiency can only be brought forth by full-scale tests. The sole
reliance on mechanical models and analysis is insufficient.
Mechanical models in conjunction with experiments have already
been applied successfully ((5 to 8)). In the following the
principles of the described approach will be outlined for the
HLV bond anchorage.
Single bar bond anchorage
The analytical investigation of a multibar bond anchor began
with the study of the behaviour of a single GFRP bar (Polystal;
0 7, 5 mm; f ctm "' 1600 MPa) bonded by an unsat. polyesther resin/sand mortar in a cylindrical steel housing with an inner
thread (7). Fig. 7 shows the model anchorage.
The deformation behaviour of the GFRP bar and of the steel cylinder can be assumed elastic. The coupling of the bar to the
mortar can be modelled by a bond stress-bond slip relation. In
Fig. 6 the set-up of the test to establish the r-s relation is
shown. The bond length lb of the bar in the mortar cylinder was
chosen lb"' 2,6 de"' 20 mm. The force F was applied with a constant rate of the slip. The measured r-s line exhibits a maximum rmax· It then drops steeply to a friction bond stress rR
with 1ncreasing slip.
The observed behaviour can be modelled in many ways. Here a
simple trilinear r-s relation was chosen. The suitability of
the chosen relation can be contra ll ed by tests and can be interactively improved. It must be pointed out that this approach
is a crude approximation of the complex triaxial state of
stress on the bar's surface.
The behaviour of the bonded bar can be described by the wellknown differential equation of bond which is derived on basis
of the material laws of the steel , the GFRP bar and the bond,
and by taking into consideration the equilibrium and compatibility conditions. The differential equation is given by Eq.(6):
FRP Reinforcement
641
4(1-nt.L)
r(s(x))
(6)
with:
Ec ... axial tensile Young's modulus of GFRP
Es ... axial Young's modulus of steel housing
n
Ec/Es
1-L
Ac/As
z
0,25
Fig. 7 shows the mechanical model. Eq. (6) can be solved either
analytically or by numerical integration for specific boundary
conditions. In the numerical evaluation the cross-section As
and the Young's modulus Es of the housing, the bond length l b
and the peak value rmax were varied.
In Fig. 7 schematically the distribution of the composite tensile stress ac(x) and of the bond stress r(x) is shown. Two
types of failure may occur: The tensile rupture force of the
GFRP bar Feu = fctAc can be attained, if the bond length is
sufficient. Failure may occur by pull-out; in that case the
failure force Fcur is the maximum force for the specific bond
length which can oe transferred by bond.
Fig. 8 shows the comparison between the numerically determined
values with the model and the experimental results. The failure
force is related to the strength of the material . For short
bond lengths the model underestimates the real behaviour. Both
the test and the model predict that the necessary bond length
to activate the material strength is in the range of 300 mm. In
(8) the single bar anchor was analyzed with the finite element
method using bond link elements with a similar characteristic
as shown in Fig. 6. Very close results were obtained.
Multibar bond anchorage
The distribution of the internal forces in a multibar anchorage
is very complex. It can be approximatively described by a strut
model which is shown in Fig. 9. Sections through the anchorage
after failure exhibit inclined micro-cracks and interlaminar
damage in the bars, both indicated in Fig. 9. It was shown in
(5) that both the inclined strut forces D and their stress components or, r are non-uniformly distributed along the bar's perimeter. Hence, it was necessary to establish a relation between the maximum bond stress rmax and the radial normal pressure in tests. A shear-friction relation of the coulomb type
642
Rostasy and Budelmann
was found. In Fig. 10 the results of calculations and of tests
for bond anchors for with 8 GFRP bars are shown.
CONCLUSION
Anchorages for post-tensioning tendons consisting of FRP elements must be designed in such a way that the relevant mechanical properties of the FRP material can be optimally utilized.
The best approach for the development proved to be an interactive process consisting of mechanical modelling and testing.
The report describes this approach for the bond anchorage of
GFRP bars.
LITERATURE
(1)
FIP Commission on Prestressing Materials and Systems:
High-Strength Fiber Composite Tensile Elements for Structural Concrete. State-of-Art-Report. July 1992, unpubl ished.
(2)
Hoffman, 0.: The brittle strength of orthotropic materials. J. o. Composite Materials, Vol. 1, 1967, pp. 200/206.
(3)
FIP Commission on Prestressing Materials and Systems: Recommendations for acceptance of post- tensioning sys terns.
June 1992, unpublished.
(4)
Hankers, Ch. and Rostasy, F.S.: FRP-tendons for post-tensioned concrete structures. ACMBS-I, October 7-9, 1992;
Sherbrooke, Quebec, Canada.
(5)
Kepp, B.: Zum Tragverhalten von Verankerungen fUr hochfeste Stabe aus Gl asfaserverbundwerkstoff al s Bewehrung im
Spannbetonbau. Diss. TU Braunschweig, 1984.
(6)
Faoro, M.: Zum Tragverhalten kunstharzgebundener Glasfaserstabe in Bereich von Endverankerungen und Rissen im Beton. Diss. Universitat Stuttgart, 1988.
(7)
Konig, G.; Wolff, R.: Heavy duty composite material for
prestressing of concrete structures. IABSE Symposium, Paris-Versaille, pp. 419/424, 1987.
(8)
Keuser, M.; Kepp, B.; Mehlhorn, G.; Rosta.sy, F.S.: Nonlinear Static Analysis of End-Fittings for GFRP-Prestressing Rods. Computers ex: Science, Vol. 17, No. 5-6, pp.
719/730, 1983.
FRP Reinforcement
643
TABLE 1 - OVERVIEW ON MECHANICAL PROPERTIES OF FIBERS
(MEAN VALVES)
fiber
type
for
axial
tensile
axial
Young's
axial
ultim.
FRP
strength modulus
strain
type
density
MPa
GPa
%
t;m3
E-glass
GFRP
2600
75
3,3
2,6
aramid
AFRP
3000
125
2,3
1, 45
carbon
CFRP
3300
215
1, 6
1,70
colddrawn
1700
205
7
7,85
prestr. steel
Rostasy and Budelmann
644
TABLE 2 - OVERVIEW ON THE :MECHANICAL PROPERTIES
OF SO:ME FRP
fiber
volume
brand
fiber/
type
matrix
vf
Vol.-%
Polystal
G/UP
68
G/EP
60
Arapree
A/EP
45
Fibra
A/EP
59
Leadline
C/EP
65
CFCC (strand)
C/EP
64
Polygon
axial 1) axial 1> axi a1
tensile Young's
ultim.
strength modulus
MPa
GPa
2650
75
1800
53
2990
93
1790
56
3000
123
1350
55
2350
110
1380
65
2790
229
1815
149
3290
213
2120
137
strain
%
3,3
3,1
2,3
2,0
1 '3
1,6
1) top number related to the fiber cross-section; bottom number
to the composite section
FRP Reinforcement
645
2500
N
E
..§
2000
z
c
Get =v1 cr 1+(1-vdcrm
• tensile fracture
Ect = VtEt+(1-vtlEm
fct :::vtftt+(1-vtlcr~ - - -
·-----..,---r-
....
......
til
~
c
Q)
I
:
-~-t=t
I :___ I_
glass fiber __ \ /
1000
Q)
"Ui
1
-~----+-1----l
::: Vtftt; if Vt>O,S
I
til
til
I
500
......
---
I
1
\~:
fct
FRP-element 1
Polystal.¢7.5mm
1oi-
--v~ro vot.- 0
matrix resin,vm= 1-vt
cr•.
crrr
0
2
3
5
strain Ec,Em,Et in %
Fig. 1-Stress-strain lines of components and
of the GFRP composite (example)
section,A-A
section, 8-8
AI-Fig. 2-HLV-anchorage for GFRP Polystal bars (schematically)
646
Rostasy and Budelmann
Fig. 3-HLV-anchorage after failure of tendon
z 80
-----
.><
.!;
QJ
u
....
....0 70
:E
c
""
....
c"' 60
estim. Fc 1k(t); material GFRP Polystal ¢7,5 mm
estim. Fr1k(t); TA 8 and 19 bars; tests at 20°C, air
-----
T
I
i
I
I
QJ
~
::I
"C
·;;;
....
QJ
-c
c
50
~ t-eo--
....::I
QJ
0,.0.,0,'17
-
c:!..
u
.. ......
----
I
,_....,.
I
I
I
b
'I
:I
.'
I
I
I
I
I
-
~~~
...
Ll
---- -----
production lots
"E.
40
::I
....
....
,,II
,,'I
I
0
QJ
QJ
;;,,
0T
0
<1,-,[J,"J'
creep rupture
residual strength
unbroken specimen
2
3
4
5
time under loCld in log t [h)
Fig. 4--Creep rupture tests with the HLV-anchorage
6
FRP Reinforcement
120
I
upperstress o 0
...
...
'E
.
.!:
•
eo
~
~
1-
I
.............
Ill
.
884 N/mm2
__
~
. , .. .a.---
•• •• .. . . . . .
<l
~
725 N/mm2
810 N/mm 2
•
~
z
647
M
_....._~_....,
40
....
- - test finished
without rupture
I
~
I
I
105
10 4
16'
load cycles N
Fig. 5-S-N-lines of single Polystal bars, cp 7,5 mm in bond anchorage
(bond length lb = 300 mm)
70
60
50
d
0...
40
f.o-
test result
cal -r- s
100 ----c-j
T
1
T
lb = 20mm
tmax
0
0
..--
1
~
c
......
30
~
Vl
Vl
QJ
t...
~
20
LGFRP-bar
resin/sand mort.
Vl
~
c
0
10
TR
.0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
slip s in mm
Fig. 6-Bond stress-slip relation for a GFRP bar embedded
in a resin mortar; experimental relation and material law
648
Rostasy and Budelmann
mortar
L b ---c=-J
f-=:J
steel housing; A 5 ,E5
~(o)
GFRP.Ac.Ec
tmax
0
Fig. 7-Model of single bar bond anchorage.
Schematic distribution of stresses
T
1max
1max = 36: 40, 44MPa
L--~-----c-
0'40 ' 6
o test results
-calculated
s [mml
tensile failure
1: max [MPa]
44
40
36
100
200
300
bond length Lb in mm
Fig. 8-Measured and calculated failure force of
single bar bond anchors dependent on bond length
FRP Reinforcement
fc -AFc I
H,Hil
I
--+
Fig. 9-Strut model for a multibar bond anchor
8 GFRP bars ¢ 7, 5 mm
l
1: max
variabel
TR
0,8
0,6
bond failure
""'- 1:max [MPa]
0,4
~ ~~
0,2
50
0
100
I range of tests
-
calculated
200
300
bond length Lb in mm
Fig. 10-Measured and calculated failure force of
multibar bond anchors dependent on bond length
649
SP 138-39
Ductile Behavior of Beams
Using FRP as Tendons and
Transverse Reinforcement
by H. Taniguchi, H. Mutsuyoshi,
T. Kita, and A. Machida
Synopsis: It is known that PC members reinforced with FRP as tendons
show brittle failure regardless of the failure mode. The objective of
this paper is to improve the ductility of PC members reinforced with FRP
as tendons. Firstly, the compressive properties of concrete confined
with FRP as transverse reinforcement was investigated. Major
improvement can be made in the stress-strain relationship of concrete
laterally reinforced with FRP, and the concrete members could be given
ductility characteristics similar to those of steel-reinforced members by
confining the concrete with FRP. Secondly, several PC members
reinforced with FRP as tendons and transverse reinforcements were
tested and investigated. It was found that marked improvements could be
made in the ductility of the PC members with FRP tendons by confining
the part of concrete subjected to flexural compression with FRP and
making the members undergo flexural compression failure.
Keywords: Beams (supports); confined concrete; ductility; failure; fiber
reinforced plastics; flexural strength; prestressed concrete; prestressing
steels; stress-strain relationships
651
652
Taniguchi et al
H. Taniguchi is a senior research engineer, Civil Engineering Department,
Technical Research Institute, HAZAMA Corporation, Japan. He obtained his
master of engineering degree from Kanazawa Univ. in 1988. His recent
research has been on new materials and mineral admixtures.
ACI member, H.Mutsuyoshi is an associate professor of Civil Engineering
at Saitama Univ., Urawa, Saitama, Japan. He obtained his doctor degree
from Univ. of Tokyo in 1984. His current research interests include
application of FRP to concrete structures, external PC and seismic
analysis of R/C structures. He is a member of ACI Committee 440
T.Kita is a chief research engineer, Civil Engineering Department,
Technical Research Institute, HAZAMA Corporation, Japan. lie obtained his
master of engineering degree from Ehime Univ. in 1971. Ills recent
research has been on new materials and anti-washout underwater
concrete.
ACI member, A.Machida is a professor of Civil Engineering at Saitama
Unlv., Urawa, Saitama, Japan. He obtains his doctor degree from Univ. of
Tokyo in 1976. His research interests include seismic design for R/C
structures, composite structures consisting of steel and concrete and
properties of fresh concrete.
INTRODUCTION
Fibre reinforced plastics (FRP) are being used in an increasing number
of cases as substitutes for reinforced concrete (RC) steels and
prestressed concrete (PC) steels and a large number of studies have been
published on this matter. When compared with conventional steels, FRP 1)
has greater tensile strengths than PC steel rods, 2) is free from
corrosion which leads to lowering of strengths, 3) is free from
magnetization and 4) is light. While it has these merits, it also has
drawbacks, namely: 5) it has no plasticity regions in its stress-strain
relationships and generally its elongation is small, and 6) its strength
is· reduced in regions subjected to bending during forming. Special
considerations also need to be given to 7) its small Young's modulus and
8) its small linear expansion coefficients, which are neither advantages
nor drawbacks. Several problems in the application of FRP to actual
structures that result from the above dynamic characteristics have been
pointed out, including those to do with the form of failure under
ultimate conditions. Concrete members reinforced with FRP as tendons
show brittle failure regardless of the failure mode. This makes the
determination of the ultimate conditions an important matter. If the
sizes and types of loads are known, plastic properties, or the ductility,
of the structure need not necessarily. When designing structures with
newly developed structural members , one need not be constrained by
conventional design methods for RC; it is desirable rather to develop
new design methods suited to the new materials. However, 1) as
long-standing practice, types of failure for which there are warning
FRP Reinforcement
653
signs are generally used In the design of RC structures, and 2)
redistribution of moments may sometimes be required to prevent the
collapse of the structure, while in seismic structures It Is safe and
rational to ensure a certain amount of ductility. These considerations
make it desirable, if possible, to provide the members with ductility. It
Is with such considerations in the background that studies have been
conducted on methods of Improving the ductility of concrete members
reinforced with FRP. Mutsuyoshl et al.(l) have suggested the possibility
of safe and rational design by using concrete collapse as the ultimate
failure form, while Akiyama et al.(2) have shown that the ductility may be
raised by combining bonded and unbonded FRP or prestressed and
unprestressed FRP. The methods proposed to date, however, have failed
to provide the FRP-reinforced members with the mechanical properties of
the type observed with RC and PC members, which undergo flexural
yielding.
The objective of this study is to Improve ductility of PC members
reinforced with FRP as tendons. First, the compressive properties of
concrete confined with FRP as transverse reinforcements was
investigated. Secondly, several PC members reinforced with CFRP as
tendons and transverse reinforcement were tested and Investigated.
COMPRESSIVE CHARACTERISTICS OF CONCRETE CONFINED WITH FRP
Outline
It is known that the compression-deformation characteristics of
concrete confined by steels tends to give larger values than that of
unconfined concrete(3}. Here, experimental studies were conducted on
the behaviour of concrete under compression when FRP, with its
different dynamic properties from steel, was used to confine the
concrete.
Outline of Test
The specimens were either prisms with sections of 15 em by 15 em or
cylinders with diameters of 15 em, their heights in both cases being 30 em
(Figure 1}. Carbon fibre reinforced plastics {CFRP} and high-strength
steels were used as the transverse reinforcement. The CFRP rods were
either of the 7-piece strand type or the single type {rod-like type}. The
transverse reinforcement was in spirals, forming rectangles 11 em by 11
em or 7.8 em by 7.8 em or circles 11 em in diameter when viewed from
above. Five test parameters of transverse reinforcement pitch, shapes,
types, concrete compressive strengths, and existence of cover were used
as shown in Table 1. The properties of the transverse reinforcement is
given in Table 2.
In the tests, the 1960-kN capacity testing machine shown in Figure 2
was used as the loading device. This was a simple uniaxial compression
tester with pin support at the top. The axial deformation of the
specimens was obtained by measuring the central displacement with a
compressometer with a measurement section of 15 em and the load was
applied until the specimens underwent failure or the axial strain
reached about 3%.
654
Taniguchi et al
Test Results and Discussion
Effects of Transverse Reinforcement Pitch on Stress-Strain
Relationship -- Representative examples of stress-strain curves obtained
In the tests are given In Figure 3. The figure shows the effects of the
reinforcement pitch on specimens with covers confined with circular FRP
reinforcements. It was observed that the yield strength was slightly
larger with plain concrete. As regards the behaviour after yielding,
with specimens confined with FRP, the strength decreases at first due to
the separation of the cover concrete and the rupture of the axial
erection bars, but the confining effect comes into operation once the
axial strain reaches about 1% and the rate of decrease of strength
becomes more gentle despite the complete separation of the cover
concrete. In other words, It was found that marked Improvements could
be achieved In the deformation characteristics of the concrete after
yielding by confining the concrete with FRP. When the pitch of the FRP
transverse reinforcements was altered, it was observed that the rate of
strength reduction decreased with the pitch, and especially at a pitch
of 3 em the strength did not in fact decrease but continued to rise even
at a strain of 2.5%.
Effects of Transverse Reinforcement Shapes -- The effects of the
transverse reinforcement shapes on the stress-strain relationship are
shown in Figure 4. In the behaviour after the strain has reached about
0. 7% when the cover is separated and the confining effect becomes
apparent, It can be seen that the rate of strength decrease is smaller
with specimens confined with circular reinforcements than with specimens
with rectangular reinforcements. This Is thought to be due to the fact
that, whereas uniform action of tensile forces only Is observed during
axial compressive loading with circular reinforcements, with rectangular
reinforcements bending comes Into operation together with the tensile
forces at the corners, and also due to the lowering of the tensile
strength at the bends in the rectangular reinforcements.
Effects of Transverse Reinforcement Types -- The effects of the
differences In the transverse reinforcement types {strand or single
types, same diameters) on the stress-strain relationship are shown In
Figure 5. No clear differences due to the reinforcement types were
observed in the tests despite the differences in the effective
cross-sectional areas and the direction of the fibres in the
reinforcements.
Effects of Concrete Compressive Strength
The effects of the
differences In the compressive strength of the concrete on the
stress-strain relationship are shown In Figure 6. The figure Indicates
that the rate of strength decrease after yielding Is smaller In confined
specimens prepared from concrete with low compressive strengths. This
Is the same tendency as has been observed In concrete confined with
steels.
Effects of Concrete Cover -- Investigations were made on the effects
of the existence of concrete cover on deformation characteristics after
yielding. Shown In Figure 7 are the stress-strain curves for specimens
without cover. When this Is compared with Figure 3, It Is to be noted that
lowering of the strength due to the separation of the cover is not
observed In the plastic region and the strength either shows a linear
FRP Reinforcement
655
increase or is maintained at the same level; with the specimen confined
at 3 em pitches in particular, the strength continued to rise even after
the strain reached around 3%. Specimens confined with high-tension
steels had shown similar behaviour to those confined with FRP at the
same pitches. This fact indicates that steel and FRP provide the same
level of confining effect on concrete despite the differences in their
dynamic characteristics. For an assessment of the effects of concrete
cover, plain concrete was used in the cover, and on the assumption that
the stress-strain relationship will obey the " e" function(4), the
compressive force taken by the core of the confined concrete was
estimated by subtracting the compressive force taken by the cover from
the total compressive force. The stress-strain relationship for the core
as estimated in this way for a specimen with concrete cover and the
stress-strain relationship for a specimen without cover as obtained from
the test are compared in Figure 8. The close agreement between the two
indicates that confined concrete members with cover may be considered
as composite members consisting of the core concrete surrounded by the
transverse reinforcement and the plain concrete in the cover.
Stress-Strain Model of Concrete Confined with FRP
Several proposals have been made to date for model equations for the
stress-strain relationship for concrete confined with steel. These
existing equations, however, are not adequate for estimation of the
stress-strain characteristics of concrete confined with FRP. The main
reasons for this are thought to be the fact that the steel is considered
to have yielded in the plastic region of the confined concrete in the
model equations for specimens with steel and the fact that the elasticity
coefficients of FRP differ from those of steel. An attempt was made,
therefore, to propose an equation for the stress-strain relationship for
concrete confined with FRP reinforcement. The method proposed by
Sakai(5) for concrete confined with steels was used in devising the
model, and corrections were made to parts of the equations proposed by
Sakai using the test results to allow for the differences in the dynamic
characteristics between FRP and steel. The corrected equations are
given below. Figure 9 is a stress-strain model for confined concrete. The
curve equations proposed by Muguruma et al.(6) were adopted in view of
the values for the stress and strain obtained in this way. The
stress-strain model for concrete confined with FRP obtained in this way
is compared with examples of the test values in Figure 10. The proposed
model agrees closely with the test values.
0.015x (ez-ezs)
a r= ------'-----;,_-:-- x Er
0.024X exp (Erx 6x 10- 5 )
a em= (1+ 4
~
am
)0 ' 5 x am
e em= (1 + 65 ~ ) x e m
am
(1)
(2)
(3)
(4)
656
Taniguchi et at
ecu=0.024xexp (Erx6x10-")
(5)
Er= (1-S/D) x (2XAw) /
(6)
(S/D) xEs
where, a r: lateral confining stress, e z: axial strain, a em: maximum
stress of confined concrete, a m:maximum stress of plain concrete, e em:
axial strain at maximum stress of confined concrete, e m: strain at
maximum stress of plain concrete, e cu: ultimate stress of confined
concrete, a u: ultimate stress of plain concrete, Er: lateral confining
rigidity, Es: Young's modulus for transverse reinforcement, Aw: sectional
area of transverse reinforcement, S:transverse reinforcement pitch, D:
diameter of confined core
BENDING CHARACTERISTICS OF PC MEMBERS USING FRP
AS TENDONS AND TRANSVERSE REINFORCEMENT
Outline
It is generally held that when reinforced or prestressed concrete is
subjected to bending, the yielding deformation of steel prevents sudden
collapse of the structure and there will be warning signs before the
structure undergoes failure. An attempt has been made here to provide
members reinforced by FRP with such ductile properties. That is, a way of
thinking opposite to that conventionally used with RC and PC was
adopted and an attempt was made to give the members ductility by
exploiting the deformation capacity of the concrete. As was discussed in
the previous chapter, it was found that major improvements could be
made in the compressive deformation characteristics of concrete by
confining it with FRP. In the experiments described below, the main aim
was to improve the ductile capacity of PC members which use FRP in their
tendons and transverse reinforcement.
Outline of Test
The specimens used in the test were beam members 15 em high, 20 em wide
and 180 em long,as shown in Figure 11. For the tendons, 2 or 3 CFRP rods
were placed at distances of 10 and 11 em from the compressive edge. The
quantities of tendons and the prestress were determined so that the
ultimate failure would take the form of flexural compression failure. The
properties of the CFRP used in the specimens are given in Table 3. The
CFRP rods were made of entwined threads and those with diameters of 12.5
mm and 15.2 mm were used for the tendons, while those used as transverse
reinforcement had a diameter of 5.0 mm. The transverse reinforcement
was in the form of rectangular and circular spirals at 3 em and 5 em
pitches respectively. This transverse reinforcement was positioned in
the 70 em section including the uniform moment segment, while in the
shear section the required amounts of deformed reinforcements (D6) were
placed as shear reinforcements.
Prestress corresponding to 60% of the guaranteed failure load of the
CFRP was introduced 14 days after the placement of the concrete
(compressive strength: 29.4 MPa or more). The prestress was introduced
simultaneously via jigs with a jack by the pre-tension method on the 2 or
3 tendons.
FRP Reinforcement
657
The test parameters are given in Table 4. Unidirectional load was
applied at 2 points at a span of 150 em and with a net bending section of
40 em, and measurements were made on such items as the load, deflection,
variation in the prestress and cracking.
Test Results and Discussion
Crack Patterns and Failure Characteristics -- Examples of crack
patterns are given in Figure 12. Cracks first occur in the tensile side of
the concrete as the load is applied. These cracks develop with the
increase in the load. Cracks then begin to appear also on the
compressive side as one approaches the maximum load, and, with further
loading,
separation of the concrete covering the transverse
reinforcement on the compressive side was observed. This was followed by
the rupture of the compression bars (CFRP rods, diameter: 5.0 mm) placed
to attach the transverse reinforcement and all the cover concrete above
the neutral axis became separated in the final stages. The loading was
continued further in the case of Specimen No. 5 and this resulted in the
separation of all other concrete on the tensile side, with the tendons
only on the tensile side balancing the concrete confined with confining
reinforcements on the compressive side.
It was confirmed from the foregoing that sudden failure of the type
observed in flexural compression failure could be avoided and major
improvements made in the failure characteristics by confining the part
of the concrete subject to flexural compression with FRP.
·Load-Displacement Relationship -- The load-displacement relationships
obtained in the test are shown in Figure 13. When a comparison was made
between specimens with transverse reinforcements (No. 1 to 5) and those
without (No. 6), it was seen that the maximum strength of specimens with
transverse reinforcements was about the same or a little lower than that
of specimens without. As regards the behaviour after yielding, the
strength of unconfined specimens decreased after yielding, showing
typical flexural compression failure characteristics. In the confined
specimens, on the other hand, although the strength shows a sudden drop
at first due to the separation of the cover concrete, it is either
maintained at the same level or shows a very gentle decrease after this
point. In other words, it was found that the plastic deformation
characteristics of the concrete confined with FRP resulted in
deformation characteristics very similar to those observed with failure
of the flexural yielding type. The behaviour after yielding differs
according to the shapes and pitch of the transverse reinforcement. The
effects of these factors on the load-deformation characteristics are
discussed below.
Effects
of
Transverse
Reinforcement-Related
Factors
on
Load-Deformation Characteristics -- The effects of the differences in the
transverse reinforcement pitch are shown in Figure 13(a}. The maximum
strength is about the same in all the specimens regardless of the
reinforcement pitch, suggesting that the differences in the pitch has
little influence on the strength within the range investigated on this
occasion. Similar tendencies are also observed in the lowering of the
strength after yielding. This is thought to be due to the fact that the
similarity of the reinforcement shapes means that the proportion of the
cover concrete in the cross-section remains the same and the adequate
658
Taniguchi et al
confining effect has ensured that separation has not yet occurred in the
concrete within the transverse reinforcement. An examination, however,
of the load-deformation characteristics after yielding reveals that
whereas the strength of Specimen No. 1 with a reinforcement pitch of 5 em
shows a gentle decrease after yielding, that of Specimen No. 2 with a
reinforcement pitch of 3 em shows an increase, indicating that the
confining effect increases as the reinforcement pitch decreases, as was
observed in the compression test results.
The effects of the confinement area on the load-deformation
characteristics are shown in Figure 13(b). Although circular transverse
reinforcement is used in all the specimens here, while the area of the
confined concrete is relatively large and includes the tensile region in
Specimen No. l, the area confined is restricted mainly to the compressive
region in Specimen No. 3. As can be seen in this figure, the rate of
strength decrease is slightly lower with Specimen No. 1, indicating that
the larger the confinement area of compressive concrete, the greater
the confinement effect and the greater the contribution to the
improvement of ductility. Since little lowering of the confinement effect
results from confinement of the concrete in the tensile region including
the tendons, it is thought effective to have the transverse
reinforcement function also as shear reinforcement.
The effects of the transverse reinforcement ratios on the
load-deformation characteristics are shown in Figure 13(c). When the
strengths of the transverse reinforcements used are the same, the
reinforcement effect increases as the area confined by each reinforcing
bar decreases. In the test results here too, the improvement effect is
the greatest in Specimen No. 5 in which 2 rectangular reinforcing bars
are used, showing that the reinforcement effect increases as one
in.creases the transverse reinforcement ratio.
The effects of the transverse reinforcement shapes on the
load-deformation characteristics are shown in Figure 13(d). A comparison
is made in this figure between the results for the specimen with the
smallest confinement effect among those with circular reinforcements and
the results for the specimen with the greatest confinement effect among
those with rectangular reinforcements. The rate of strength decrease
after yielding is more gentle and the confinement effect greater with
circular
reinforcement.
Furthermore,
whereas
rupture
of the
reinforcements was not observed with circular reinforcements,
rectangular reinforcement was found to undergo rupture at the corners.
This was a phenomenon that was also observed in the compression test on
concrete confined with FRP. The lowering of the strengths at the bends
in the rectangular reinforcements are thought to be contributing to the
lowering of the reinforcement effect.
From the foregoing, it was found that major improvements could be
made in the deformation characteristics of members after yielding
through use of concrete confined with FRP and making the failure take
the form of the collapse of the concrete. In other words, it was found to
be possible to provide members reinforced by FRP with properties similar
to those of RC and PC members which undergo flexural yielding using a
concept opposite to that conventionally used with these.
FRP Reinforcement
659
Analysis Method
After it was found through tests on beam members that major
Improvements could be made In the deformation performance and failure
characteristics of the members by confining the compressive concrete
with FRP transverse reinforcement, analytical Investigations were
conducted on this behaviour.
It has already been reported that the bending behaviour of PC members
using FRP in their tendons can be analysed by conventional bending
theory(7). Here, the analysis was carried out as follows using fibre
models.
·Analysis Model -- The following assumptions were made in using fibre
models.
1) The stress-strain curves for the concrete were determined as follows.
The stress-strain curves of the cover concrete will be that of plain
(unconfined) concrete, while those of concrete confined with FRP will
follow the model equation proposed in Section 2.
2) The cover concrete will become separated when Its ultimate strain ( e
u;0.0035) has been reached and will no longer carry load after this. The
separation of the cover concrete results in reduction of the member
cross-section.
The load-deformation relationships were obtained on these assumptions.
Study of Analysis Results -- Examples of test and calculation results
for the load-displacement relationships are given in Figure 14. There Is
generally a close agreement between the test and calculation values
although the rate of the strength decrease is greater in the calculated
values. This Indicates that the load-deformation characteristics of PC
members reinforced with FRP can be obtained with a fair degree of
accuracy by using stress-strain models for confined concrete taking Into
account such factors as the shapes and pitch of the transverse
reinforcements and the concrete strengths.
ESTIMATION OF DEFORMATION ON CAPACITY OF PC MEMBERS USING FRP AS
TENDONS AND TRANSVERSE REINFORCFJIIENT
It was found possible to improve the ductility and failure
characteristics of members by combining FRP transverse reinforcements
with FRP tendons to produce flexural compression failure and to analyse
this behaviour through the method described above. In utilizing
transverse reinforcements on actual structures, there Is a need to
determine the tendon and transverse reinforcement quantities that would
satisfy the ductility requirements of the structures.
Here, equations were proposed for expression of the ductility factor
on the cross-sections of the specimens used in the tests discussed above.
That Is, analytical investigations were conducted .by the above method
with the ductility factor as a function of the mechanical reinforcement
ratio and transverse reinforcement ratio. Ductility factor f.l., mechanical
reinforcement ratio P and transverse reinforcement ratio Pw may be
expressed as follows.
f./. ;
</>
u/¢ Y
(7)
where, ¢ u: curvature at ultimate moment (curvature at ultimate linear
660
Taniguchi et al
strain of confined concrete or ultimate strain of tendon), ¢ y: curvature
when plain concrete reaches ultimate strain (a u=O. 0035)
P = As • fs/b · d · f' c
(8)
where, As: nominal sectional area of FRP tendons, fs: guaranteed rupture
strength of FRP tendons, b: sectional width d: effective height, f' c:
concrete strength
Pw = 2Aw/S · D
(9)
where, Aw: nominal sectional area of transverse reinforcement, S:
transverse reinforcement pitch, D: diameter or length of side of
transverse reinforcement.
The relationships between the ductility factor and the mechanical
reinforcement ratio and transverse reinforcement ratio are shown in
Figure 15. While there are no major differences in the ductility factors
according to the transverse reinforcement ratios at mechanical
reinforcement ratios of around 0.4 or less, there is a clear tendency for
the increase in the ductility factor to become greater in cases with
higher
transverse
reinforcement
ratios
once
the
mechanical
reinforcement ratio exceeds 0.4. At the same time, the ductility factor
increases in all cases with the mechanical reinforcement ratio, and the
tendency is especially marked in the region with low mechanical
reinforcement ratios. This tendency is also more marked in cases with
greater prestress.
Once the required ductility factor of the members is known, therefore,
the combination of mechanical reinforcement ratios and transverse
reinforcement ratios that satisfies this ductility factor can be obtained
from the above relationships. Where a plural number of combinations are
possible, the optimum tendon and transverse reinforcement quantitieE
can be selected taking into account such factors as the shapes of the
members.
CONCLUSION
The study reported above was conducted with the aim of raising the
bending ductility of PC members reinforced with FRP. Experimental and
analytical
investigations
were
conducted
on
the
compressive
characteristics of concrete confined with FRP and the deformation
performance of members which contain concrete confined with FRP. The
following conclusions were drawn from these investigations.
(1) Major improvements can be made in the stress-strain relationships of
the concrete and the concrete members can be given ductility
characteristics similar to those of steel-reinforced members by
confining the concrete with FRP. The improvement effect differs greatly
according to the transverse reinforcement pitch and shape and the
concrete strengths.
(2) In terms of its stress-strain relationships, confined concrete with
cover may be thought of as a composite member consisting of the confined
core concrete and the plain (unconfined)cover concrete. A compressive
stress-strain model for concrete confined with FRP was proposed and
verified.
FRP Reinforcement
661
(3) It was found that marked improvements could be made in the
deformation performance of PC members with FRP tendons by confining the
part of the concrete subjected to flexural compression with FRP and
making the members undergo flexural compression failure. In other
words, it was found possible to provide members reinforced by FRP with
properties similar to those of RC and PC members which undergo flexural
yielding using a concept opposite to that conventionally used with steel
reinforcement.
(4) The behaviour of PC beams using FRP as their tendons and transverse
reinforcement can be analysed using the conventional fibre models by
applying the appropriate stress-strain relationships for the confined
concrete and the cover concrete.
(5) Once the required ductility factors of the members are known, the
optimum tendon and transverse reinforcement quantities can be evaluated
from
the
relationships
between
ductility
factors,
transverse
reinforcement ratios and mechanical reinforcement ratios.
(6) From the above, it is clear that the use of FRP as transverse
reinforcements is a promising method for making improvements in the
deformation performance of PC beam members using FRP in their tendons.
There is a need now to conduct detailed investigations on the design
method, application conditions and structural details with regard to
various confinement conditions, with a view to the application of this
method to actual structures.
REFERENCES
1. Mutsuyoshi,H et al; "Mechanical Properties and Design Method of
Concrete Members Reinforced by Carbon Fiber Reinforced Plastics "
Proc. of JCI, Vol.12, pp.ll17-1122, 1990. (In Japanese)
2. Akiyama. H et al; "Flexural Behavior of Prestressed Concrete Beams
Reinforced with Strip-like Aramid Fiber rods" Proc. of JCI, Vol.12,
pp.1099-1104, 1990. (In Japanese)
3. ex. JCI; "State of the Art Report of The Technical Committee on Design
of Reinforcement for Ductile R/C Members" 1990. (In Japanese)
4. Umemura.H; "Plastic Deformation and Ultimate Strength of Reinforced
Concrete Beam" Transaction of AIJ, No.42, pp.59-70, 1952.
(In Japanese)
5. Sakai, Y ; "Stress-Strain Curves of High-Strength Concrete Confined
by Lateral Reinforcement" Proc. of JCI, Vol.13, pp.43-48, 1991.
(In Japanese)
6. Muguruma, H et al; "Stress-Strain Model of Confined Concrete" Pro c.
of Cement and Concrete, Cement Association of Japan, Vol.34, pp.429
-432, 1980. (In Japanese)
7. Okamoyto, T et al; "Flexural Characteristics of PRC Beam Using Braided
Aramid Fiber Rods " Pro c. of JCI, Vol.10, pp. 671-676, 1988.
(In Japanese)
662
Taniguchi et al
TABLE 1 -
EXPERIMENTAL VARIABLES
Compressive
Specimen Cover strength
Shap~ of Type of
Pitch Shape of
No.
spec1men re1nforcement (em) reinforcement
(MPa)
1
Circle
3
2
Rectangle
Strand
5
3
Circle
Cylinder
4
Circle
7
Rectangle
5
Single
5
6
Circle
Rectangle
7
3
2 9. 4
8
Circle
9
Rectangle
Cover
Strand
5
10
Circle
11
Rectangle
Prism
7
12
Circle
13
Rectangle
Single
5
14
Circle
15
5
Circle
16
Rectangle
Prism
Circle
17
4 9. 0
Strand
5
18
Cylinder
Circle
*
19
20
21
22
23
Non
Cover
Single
Cylinder
*
2 9. 4
Prism
Single
3
5
5
3
1-5
1--
Circle
Rectangle
* High-strength steel
TABLE 2 - MECHANICAL PROPERTIES OF
TRANSVERSE REINFORCEMENTS
Reinforcement Diameter
(mm)
Strand
Single
High-strength
steel
(~')
5.0
5.3
10.1
15.2
5.4
31.0
Tensile
strength
(MPa)
1842
2117
1450
Young's
Modulus
(GPa)
137
139
Elongation
196
11.0
(%)
1.5
1.5
FRP Reinforcement
TABLE 3 - MECHANICAL PROPERTIES OF FRP
Type of C F R P
~
Diameter (nun)
Area (mrn" )
Nominal Breaking Load(kN)
I 2. 5
; I 5. 2
5. 0
I 2. 5
I 5. 2
I 0. I
7 6. 0
I I 3. 6
I 7. 6
I 4 2. I
I 9 8. 9
2 4. 0
I 5 !. 0
2 2 6. 0
Young's Modulus (GPa)
I 37
I37
I37
m
r.:D
EXPERIMENTAL VARIABLES
Concrete
Pitch Strength
ll;ml_
IMPaL
5
29.4
CFRP
Length*
_jcffil_
15.2 x2
10.8
~15.2x2
10.8
3
Specimen Shape of
_&
secti
2
~
Unit mass (g/m)
TABLE 4 -
I
5
~
Introduced Tr~nsverse
Prestress re1nforcement :r~g~~nt
ratio
ratio
lkNl_
240 (60)
0.0031
0.615
29.4
237 (60)
0.0062
0.615
3
~
--.,
~
15.2 X2
9.0
5
29.4
258 (60)
0.0045
0.615
4
~
~
12.5x3
14.0
5
29.4
255 (60)
0. 0029
0.725
5
t=IQ
~12.5x3
7.8
5
29.4
254 (60)
0.0052
0. 725
~15.2x2
-
29.4
238 (60)
-
0.615
6
..
-
( l : Introduced Prestress/Nominal breaking load (\)
* : Diameter or side length of transverse reinforcement
Axis
reinforcement
Transverse
reinforcement
Prism
Cylinder
Fig. 1-Types of compression specimens used
663
664
Taniguchi et al
Plate
Transducer
Strain gage
Plate
Fig. 2-Measurement system used in experiment
36~--~--~--~--~--~
3cm (No.1)
-cu
27
-~-~=--····T············
D.
----Scm
:E
en 18 f-•.............,..
en
.......---
7cm (No.4)
a:
en
(~o. 3)
'f==w~-.--
. .... T... ,..;;;; ~~-~ :.::.-:_ ... -
w
l-
:--~--_.-
9
o~~~~~~~~~~~
0.0
0.5
1.0
1.5
2.0
2.5
STRAIN (%)
Fig. 3-Effect of pitch on concrete stress-strain curve
FRP Reinforcement
36,---,----,----~--~----,
ct~
21
-
18
a.
:lE
(/)
(/)
w
a:
t;
······················
....... .
Circle (No.14)
..
9 _........
Rectangle (N o.13)
0 ~~~~~~~~~~~LL~
0.0
0.5
1.0
1.5
2.0
2.5
STRAIN (%)
Fig. 4-Effect of reinforcement shape on stress-strain curve
-ca 27
a.
:lE
(/) 18
(/)
strand (No.3)
w
a:
t;
9
....
.
oLL~~~~~~~~~~LL~
0.0
0.5
1.0
1.5
2.0
STRAIN (%)
2.5
Fig. 5-Effect of reinforcement type on stress-strain curve
665
666
Taniguchi et al
60.---~--"--~--~----
cu
f =49.0 MPa
a..
-enen
:!:
l-
,.
30
w
a:
...
I
I
I
(No. 18)
:;;:···
~'
:
;
""
en 15
o~~~~~~LLLLLULU~_L
0.0
0.5
1.0
1.5
2.0
STRAIN (%)
2.5
Fig. 6-Effect of concrete strength on stress-strain curve
60,---,--~----~--,---,
3cm (No.19)
cu
a..
:!:
0LWLU~~~~LLLLLUUU~~
0.0
0.5
1.0
1.5
2.0
STRAIN (%)
2.5
Fig. 7-Effect of cover on stress-strain curve
FRP Reinforcement
50~--~----~--~----~--~
Calculation
([No.6)
.....•~~-~-~~-~~-~~-~~-~--~--~·
-C'CI 37
D.
:!:
o~~~~~~~~~~~~~
0.0
0.5
1.0
1.5
2.0
STRAIN (%)
2.5
Fig. 8--Stress-strain cuJVe of core concrete
O'cm ------cn O'm
(fJ
~
O'cu
~
f--t
O'u
(fJ
0
1
I
I
1
I
I
Plain Concrete
e-m cu c-cm
STRAIN
cCU
Fig. 9-Analytical concrete stress-strain
667
668
Taniguchi et al
Exp. (No.19)
-45
ca
a.
::!!:
'~
t /)
t/)
30
tt/)
15 .
w
a:
.
/?.':. ............
Exp.(No.20)
:
!Cal. (~o. 20)
·--·····
·············-····
ouu~~LUUJ~LLUU~~LU
0.0
0.5
1.0
1.5
2.0
STRAIN (%)
2.5
Fig. 10---Analytical and experimental comparison of stress-strain curve
Fig. 11-FRP-reinforced beam specimen
No.1
:zs
No. •I.
:L
+
Fig. 12-Crack pattern
FRP Reinforcement
669
90
0 60
<(
0
....1
30
00
10 20 30 40 50 60
DISPLACEMENT (mm)
00
70
10 20 30 40 50 60
DISPLACEMENT (mm)
70
b)
a)
120,-,--,--,-,--,--,-~
z
90
······[No.5!
m
..>::
. . .~.i.~. ~.~.~-~.i·- . . . .
~ 60
<(
....1
....1
o~~~~~~~~~~
10
20
30 40
50
0 60
:
60
70
!
L.~
0
30
0
90
<(
~·~,
0
z
6
30
o~~~~~~~~~~
0
10 20 30 40 50 60
DISPLACEMENT (mm)
DISPLACEMENT (mm)
c)
d)
Fig. 13-Load-displacement curve of FRP-reinforced beam
670
Taniguchi et al
.-.
z
c.
OL'--L--'-'-'--'--'-L.W--'--'-'--'--'-'-'-'---'-'-'-'-'-'-'-'---'-'-'--'--'-'--'-'-'
0
10
20 30 40 50 60
DISPLACEMENT (mm)
70
Fig. 14-Calculated load-displacement curve
--------·-·r··------------:----------
~8
::!.
0
0.2
0.4
0.6
0.8
1
1.2
Mechanical reinforcement ratio ( p )
Fig. 15-Evaluation of ductility
SP 138-40
Ultimate Strength and
Deformation Characteristics
of Concrete Members Reinforced
with AFRP Rods Under
Combined Axial Tension
or Compression and Bending
by N. Kawaguchi
Synopsis: This paper deals with the theoretical and experimental study on the
ultimate behavior of concrete members reinforced with AFRP rods under combined axial load and bending. This involving the tests on twelve model specimens
subjected to eccentric tension or compression, contributes data on crack formations, ultimate strength, curvature variation and flexural rigidity deterioration. No
marked difference from conventional reinforced concrete members could be
observed. Ductility based on the rotational capacity of cross section and flexural
rigidity after the formation of cracks were both relatively small. In cases where a
certain amount of bond is maintained, even when the members are placed under
axial forces, the estimation at least of the ultimate strength is thought possible by
application of the beam theory with the term for the axial forces taken into
account.
Keywords: Axial loads; bending; compression; cracking (fracturing);
curvature; deformation; deterioration; fiber reinforced plastics; flexural
strength; reinforced concrete; strength; tension
671
672
Kawaguchi
Naotaka Kawaguchi: A professor in the Department of Civil Engineering at
Kokushinkan University, Tokyo, Japan. Research interests are reinforced concrete
members under combined axial load and bending, especially subjected to axial
tension and bending.
INTRODUCTION
The material properties of fibre-reinforced plastics (FRP) used in construction that relate to practical purposes are now fairly well known and investigations
are being conducted today on their application to structural members exploiting
features such as their corrosion resistance, light weight, high strength and nonmagnetic nature. In the study reported below, empirical and theoretical investigations were carried out on the behavior under axial forces and bending, an area
subjected to relatively little research to date, as a part of such investigations for
the application of FRP to concrete members. Model specimens in which aramid
fibre-reinforced plastic (AFRP) bars were used as the main reinforcement were
placed. The observations were made on such items as the conditions of cracking
and failure, ultimate strengths and curvature variation in members subjected to
axial tension or compression and bending.
TEST METHODS
Specimens and Material Oualitv
Specimens with the configuration shown in Figure I were prepared, with
extensions on either side for eccentric loading. The central test section was given
a width of !Scm and height of 20cm. Two braided aramid FRP rods sanded on the
surface, with a nominal diameter of 12mm, were installed on the tensile and compressive sides of the specimens. The geometrical reinforcement ratio of the section was 0.83%. Ordinary steel reinforcements !Omm in diameter were used in
the stirrups and at the corners. The concrete used had the mix shown in Table I
(target strength: 39.2MPa) and a total of 12 specimens in which 6 specimens were
placed at a time were prepared. In each series of 6 specimens, 3 were used for the
eccentric tensile loading test and the remaining 3 for the eccentric compressive
loading test. The direct tension test on the FRP bars gave the following results:
Young's modulus Et= 6.18xl04 MPa, tensile strength Gfi• = 1310MPa, ultimate
tensile strain £Ju = 21210xl0-6. Additionally, the AFRP rod had the value of 0.62
for the Poisson's Ratio, and -6 - -2 x I o- 6;oC for the coefficient of thermal
expansion.
FRP Reinforcement
Loadin~:
673
Method
The eccentricity et was 24cm from the centroid of tensile reinforcement in
eccentric tension, and the eccentricity ec was 20cm from the top of compressive
side of the section in eccentric compression.
In the eccentric tension test, eccentric tension was applied to specimens
placed in horizontal positions via steel rods by centre hole jacks. In the eccentric
compression test, eccentric compression was applied by Amsler testing machines
on specimens placed vertically as in conventional tests on reinforced concrete
columns. All specimens were subjected to eccentric loading in the reverse direction after the occurrence of failure under static conditions; that is, specimens subjected to failure by eccentric tensile loading were then subjected to eccentric
compressive loading, and vice versa. This procedure was adopted by the author
several years ago in tests on reinforced concrete members with the aim of obtaining the basic data for a design method that would not allow the structure as a
whole to collapse even if a part of the structural members undergo rupture.
Measurement Items
With all the specimens. measurements were made on the strain in the FRP
rods and the concrete, and the surface width of the cracks. The strain in the FRP
rods was measured with epoxy foil gauges 2mm in length placed at a total of 3
positions in each member, I at the central cross-section and 2 others at distances
corresponding to the effective depth (16cm) away from the central cross-section.
The concrete strain was measured at the same positions as the FRP rod strain with
contact gauges lOOmm in length, while the vertical strain distribution was measured at 2cm intervals from the top (compression face) of the specimens at each
position for the calculation of the neutral axis and average curvatures.
For the measurement of the surface crack widths, direct-view micrometers
with a sensitivity of 0.01 mm were used at the positions of the axial FRP rods.
TEST RESULTS
Crackin~:
and Failure Conditions
In both the eccentric tension and compression tests, the failure of all the
specimens was due to the crushing of the concrete under compression and no rupture was observed in the FRP rods. Examples of the cracking patterns at the time
of failure are shown in Figures 2 and 3. As with reinforced concrete members( I),
so-called bending failure was seen to have occurred after the development of
bending cracks. Between 5 and 8 cracks occurred in each member at more or less
uniform intervals and the widths of these cracks too were more or less uniform up
to the occurrence of crushing. No intensive widening of cracks was observed
immediately prior to failure in these tests, their widths remaining between 1.5 and
2mm. Axial cracks along the FRP rods were not observed.
67 4
Kawaguchi
Streneth and De{ormation
The measurement results are summarized in Table 2. Because of the difficulty of strain measurement immediately before failure, values corresponding to
95% of the maximum loads are used as the measurement values of the ultimate
strain and curvature. The ultimate strain of the concrete in the compressive zone
were found to be between 4000x1~ and 5000x10--{\ as shown in the table, indicating the validity of using the value of 5000x 10--{\ for the ultimate strain in calculations, for example, of strengths. The relationships between the load and the curvature as obtained from the strain distribution in the compressive zone of the section are shown in Figures 4 and 5. Examples of measurement values for flexural
rigidity as calculated from these are given in Table 3. As was expected from the
material properties of the FRP rods and the reinforcement ratio, development of
cracks resulted in rapid lowering of the flexural rigidity, the rigidity decreasing to
between 5 and 15% of the initial value in these tests, whereas with reinforced
concrete members(2) the decrease was to between 30% and 60% of the initial
values. Formation of plastic hinges was not observed and the increase was more
or less linear after a brief horizontal region after the occurrence of cracks. There
was, however, a large dispersion in the measured values for the curvature, meaning that it is not possible to deduce a general theory from the examples here and
there is a need to obtain additional data, along with the formulation of flexural
rigidity. Examples of the measured strain distribution during the stages immediately prior to failure are given in Figure 6. It can be seen that the average strain
had a linear distribution, or in other words the plane retention rule was in effect at
least up to the ultimate conditions in these tests. In the reverse loading, the