Fracture Mechanics of Concrete and Concrete Structures High Performance, Fiber Reinforced Concrete, Special Loadings and Structural

Applications- B. H. Oh, et al. (eds)

2010 Korea Concrete Institute, ISBN 978-89-5708-182-2

Correlation between tensile and bending behavior of FRC composites

with scale effect
D.J. Kim
Sejong University, Seoul, South Korea

A.E. Naaman & S. El-Tawil

University of Michigan, Ann Arbor, MI, USA

ABSTRACT: This paper describes the results of an experimental test program designed to correlate the tensile
and bending response of fiber reinforced cement composites tested under the same conditions. Several objectives were sought: 1) to correlate tensile and bending behavior of specimens having about the same cross section; 2) to ascertain that a strain hardening composite in tension leads surely to a deflection hardening composite in bending; 3) to observe scale effects on bending behavior; and 4) to verify if some theoretical correlation
between post-cracking tensile strength and bending resistance (modulus of rupture) are validated by experiments. The final objective of this study is to provide hard data needed to determine if the tensile stress-strain
response of fiber reinforced cement composites can be predicted from their load-deflection response, as currently surmised in some test standards and in some finite element studies claiming that tensile response can be
uniquely back-calculated from bending behavior. The test program included several parameters, among
which 2 types of high strength steel fibers (hooked and twisted) with identical volume fractions of fibers (1%),
and three different sizes of cross section for the beams, namely, 5025 mm, 100100 mm, and 150150 mm.
Key observations are described and conclusions drawn.
1 INSTRUCTIONS
Much research has been conducted to increase the
ductility of cement based composites by adding short
fibers, because cement based matrices have innate
weakness in terms of brittle failure under tensile and
flexural loading. To remedy such weakness, Fiber
Reinforced Concrete (FRC) and High Performance
Fiber Reinforced Cementitious Composites (HPFRCC) have been developed. FRC and HPFRCC are
usually first differentiated by their uniaxial tensile response. FRC shows strain softening behavior under
uniaxial tensile load while HPFRCC shows strain
hardening behavior. (Naaman & Reinhardt 1996)
Although all fiber reinforced cement composites can
be simply characterized according to their tensile response, so far there is no standard tensile test method
for fiber reinforced cement composites. Many researchers are still using different tensile test set-ups
e.g., different boundary conditions, sizes and geometries of specimens, gage length, and measurement
techniques. For example, some researchers have
been using bell shaped tensile specimens with hinge
to hinge boundary conditions while others have been
using coupon type specimens with fixed boundary
conditions. Since there is no standard tensile test
method for FRC and HPFRCC, and since such test-

ing is more difficult to carry out, many researchers

have investigated whether the third point flexural test
can be used, as an alternative test method, to obtain
the tensile response of the composite. Note that the
flexural response can be predicted analytically from
the tensile and compressive behavior of the material.
(Naaman (2003), Soranakom (2007) & Mobasher,
2008)) However, it is questioned whether the tensile response of the composite can be uniquely backcalculated from its flexural response. An experimental test program was designed and carried out in order to: 1) correlate the tensile and bending behavior
of FRC specimens with same cross section and to
provide some data to the above question, and 2) to
investigate the scale effect on bending behavior, and
3) to provide data for use by researchers attempting
to predict tensile response from bending response.
2 DEFLECTION HARDENING AND STRAIN
HARDENING
The response of a beam specimen under flexural
loading contains structural effects due to the specimen geometry and loading conditions. For instance,
although the tensile behavior of an FRC specimen
may generate a strain softening response, it may also

J = D ( h,either
T )h a deflection-softening or a deflectiongenerate
(1)
hardening response due to the structural effect under
flexural
Naaman (2003)
suggested
practical
The load.
proportionality
coefficient
D(h,T)a is
called
condition
for
deflection
hardening
response,
when
moisture permeability and it is a nonlinear function
) in tensionTis(Baant
higher
the
post
cracking
strength
( pctemperature
of the
relative
humidity
h and
than
the
first
cracking
strength
(

)
multiplied
with
cc balance requires
& Najjar 1972). The moisture mass

.
athat
factor,
k,
smaller
than
1,
i.e.,
the variation in time of the waterpc massccper The
unit
factor
ranges
between
and 1,wwith
0.4 being
volumek of
concrete
(water1/3content
) be equal
to thea
recommended
firstmoisture
approximation.
Soranakom &
divergence of the
flux J
Mobasher (2007) proposed a closed form solution
forwthe moment-curvature response of FRC, and the
(2)
= Jresults, using their model, indicated that
simulation
t
the direct use of uniaxial tensile stress strain response
under-predicts
response.
The water
content
wthe
canflexural
be the
expressed
as theinThey
sum
explained
this
discrepancy
by
difference
the
of
the
evaporable
water
we (capillary water, water
strain
gradient
profile
and
the
volume
of
the
stressed
vapor, and
adsorbed
water) andand
the non-evaporable
region
between
the tensile
tests.
(Mills
1966,
(chemically
bound)
water
wn flexural
Soranakom
& Mobasher
(2008) also
Pantazopoulo
& size
Millseffect
1995).
It is mentioned
reasonablethat
to
the
brittleness
and
are
more
pronounced
in
assume
that
the
evaporable
water
is
a
function
of
the
flexural
responseh,ofdegree
brittle of
materials,
while
more
relative
humidity,
hydration,
materic, and
accurate
predictions
are
obtained
with
ductile
we=we(h,c,s)
degreeTherefore,
of silica fume
reaction,
s, i.e. FRC
als.
in inverse
estimating
tensile
re=
age-dependent
sorption/desorption
isotherm
sponse
from
their
flexural
response,
the
size
of
flex(Norling
Mjonellshould
1997).beUnder
this assumption
and
ural
specimens
carefully
selected.
It
is
by substituting
Equation
1strong
into size
Equation
2in one
well
known
that
there
is
a
effect
the
obtains
behavior of cementitious composites due to their
brittle behavior (Bazant & Planas (1998), Bazant et
w
w h
al.
(1994)).
Ward &wLie &
(1990)
the
e investigated
&s + w&nbeams(3)
+ ( D h ) =
e
c +
flexural
of
reinforced
mortar
h tbehaviorhof fiber
c
s
different sizes and proposed the ratio between flexural strength and tensile strength as a parameter to
where wthe
e/h is the slope of the sorption/desorption
describe
brittleness of material. The ratio deisotherm
called moisture
capacity).
The
creases as (also
the brittleness
of the material
increases.
governing
equation
(Equation
3)
must
be
completed
Bazant et al. (1994), using extensive laboratory reby appropriate
boundary
sults,
also concluded
thatand
all initial
types conditions.
of brittle failures
relation
betweenexhibit
the amount
of size
evaporable
of The
concrete
structures
a strong
effect.
water
and
relative
humidity
is
called
adsorption
Lepech & Li (2004) investigated size effect in ECC
isotherm (plain
if measured
with increasing
relativity
structural
and reinforced
with steel
bars)
humidity
and
desorption
isotherm
in
the
opposite
beams and reported that there is negligible size
effect
case.
Neglecting
difference
(Xi etHowever,
al. 1994),the
in
in
ECC
comparedtheir
to brittle
concrete.
the
following,
sorption
isotherm
will
be
used
with
results were based only on comparing the equivalent
reference
to bothstrength
sorptionbutand
conditions.
elastic
bending
diddesorption
not consider
the deBy
the
way,
if
the
hysteresis
of
the
moisture
flection capacity.
isotherm
would be
account,of two
different
In evaluating
thetaken
tensileinto
behavior
ductile
fiber
relation,
evaporable
water
vs
relative
humidity,
reinforced cement composites, the strain capacitymust
be used according
to the Therefore,
sign of the variation
of isthea
paramount
parameter.
in this experirelativity
humidity.
Theeffect
shape
of themembers
sorption
mental
program,
the size
in flexural
is
isotherm for not
HPConly
is influenced
by many
parameters,
investigated
with respect
to bending
resisespecially
those
that influence
extentat and
of the
tance
but also
to deflection
capacity
peakrate
stress.

chemical reactions and, in turn, determine pore

structure and pore size distribution (water-to-cement
cement chemical
composition, SF content,
3ratio,
EXPERIMENTAL
PROGRAM
curing time and method, temperature, mix additives,
etc.). types
In theof literature
various
canand
be
Two
high strength
steelformulations
fibers (Hooked
found
to
describe
the
sorption
isotherm
of
normal
Twisted), showing slip hardening response under
concretefiber
(Xipull-out
et al. 1994).
However,
in the
single
testing,
were used
in present
a high
paper
the
semi-empirical
expression
proposed
by
strength cementitious matrix (84MPa) with 1% fiber
Norling
Mjornell
(1997)
is
adopted
because
by volume. Tensile and bending specimens wereit
Proceedings of FraMCoS-7, May 23-28, 2010

explicitly accounts
for the
of tests
hydration
prepared.
Direct tensile
testsevolution
and flexural
were
reactionoutand
SFa content.
This sorption
carried
using
servo-controlled
hydraulicisotherm
testing
reads (MTS810). All the tensile test series led
machine
to a strain-hardening behavior. For the flexural tests,
three different geometries of specimens, S (small),
M

1
(medium), and L (large) were prepared
to investigate

w (h, , s ) =on
G1 (the
c , s )1
+

thee sizeceffect
flexural behavior
of
HPFRCC.

10(g
)h

(4)

3.1 Materials and specimen

10(preparation
g )h

c
1 c

K1 (composition
1
The matrix mix
and proportions
c , s )e

shown in Table 1, and the

properties of fibers

are
are
shown in Table 2. A VMA (Viscosity Modifying
where the
(gelmatrix
isotherm)
represents
the
Agent)
was first
addedterm
to the
to increase
viscosity
physically
water and
the matrix.
second
and
ensure bound
uniform(adsorbed)
fiber distribution
in the
term compressive
(capillary isotherm)
capillary
The
strength ofrepresents
the matrixthewas
measwater.from
This100200
expression
valid only
content
ured
mm is
cylinders
and for
thislow
matrix
was
amount of
SF. The coefficient
G1 represents
aofself-consolidating
mixture
developed the
earlier.

water per unit volume held in the gel pores at 100%

Table
1. Matrix
composition
weight
ratio and compressive
relative
humidity,
and itbycan
be expressed
(Norling
strength.
Mjornell 1997) as
Cement (Type III)

0.80

c
s s
GFly(ash
c s ) = k vg c c + k vg s0.20

(5)

Sand (Flint)

1.00

where
Silica k and k are material
0.07 parameters. From the
maximum amount of water per unit volume that can
Super-Plasticizer
fill
all pores (both capillary0.04pores and gel pores), one
can
calculate K1 as one obtains
VMA
0.012
c
vg
fume

s
vg

Water

0.26

10 g

'
1 c
G 1 e
w0 0.188 c s + 0.2284
s
c (MPa)
s 1

K ( c s ) =
,

Table 2. Properties of
Fiber type

g c c hstudy.

10used

fibers
inthis
1

Hooked

(6)

Twisted

The material
Diameter
(mm) parameters k and k and
0.30* g1 can
be calibrated by fitting experimental data relevant to
Length(evaporable)
(mm)
30
30
free
water content
in concrete
at
various
ages
(Di
Luzio
&
Cusatis
2009b).
Density (g/cc)
7.9
7.9
c
vg
0.38

s
vg

Tensile strength (MPa)

2300
2760**
2.2
Temperature evolution
Elastic modulus (GPa)
Note
that, at early age, since200
the chemical200reactions
*associated
Equivalentwith
diameter
cement hydration and SF reaction
**
Tensile strengththe
of the
fiber after twisting
are exothermic,
temperature
field is not uniform
for non-adiabatic systems even if the environmental
The geometry
of the tensile
specimens and
temperature
is constant.
Heattestconduction
can the
be
test
set-up
are
shown
in
Figure
1.
Two
layers
of
described in concrete, at least for temperaturesteel
not
wire
mesh were
to reinforce
the bell
shaped
exceeding
100Cused
(Baant
& Kaplan
1996),
by
ends
of
the
tensile
specimens
to
minimize
failure
at
Fouriers law, which reads

the grips and out of the gage length. The gage length
was
to be 175mm (=7 inch), between two
q = selected
T
(7)
infrared markers; displacement between the markers
was measured using a non-contacting motion measwhere instrument
q is the (OPTOTRAK
heat flux, TSystem)
is the placed
absolute
uring
at
temperature,
and

is
the
heat
conductivity;
in this
about one meter from the specimen; the measurement
accuracy was 0.001 mm. Beams of three different

geometries were prepared for the flexural tests (Fig.

2): S (50mm25mm300mm); M (100mm100mm300mm);
&, L (150mm150mm450mm). In addition to M and
L type specimens, which are recommended in ASTM
C1609, the S type beam specimens were intentionally
added in this experimental program in order to have
same cross sectional area as the tensile specimens.
3.2 Test setups and procedure

= D ( h , T ) h

The proportionality coefficient D(h,T)

moisture permeability and it is a nonlinea
of the relative humidity h and temperature
& Najjar 1972). The moisture mass balanc
that the variation in time of the water mas
volume of concrete (water content w) be eq
divergence of the moisture flux J
= J

w
t

162.5

OPTOTRAK MARKERS

Apply
Displacement
Support

Apply
Displacement

50.0

Support

GAGE LENGTH

75.0

175.0

200.0

162.5

(a) S type ( 50mm 25mm 300mm)

UNIT : MM
T = 25

125.0

25.0

Figure 1. Tensile test specimen and setup.

Figure 2. Three types of flexural test specimens.

Photos illustrating test set-ups for the flexural

specimens are shown in Figure 3. Detailed information about the flexural test set-up can be found in
Kim et al. (2008). The loading speed for both tensile and flexural tests, i.e. 1.06 mm min , was originally determined from the tensile test by assuming
that the static strain rate is 0.0001 sec which is actually ten times faster than the flexural loading speed

The water content w can be expressed a

of the evaporable water we (capillary wa
vapor, and adsorbed water) and the non-e
(chemically bound) water wn (Mil
Pantazopoulo & Mills 1995). It is reas
assume that the evaporable water is a fu
relative humidity, h, degree of hydration
degree of silica fume reaction, s, i.e. we=w
= age-dependent sorption/desorption
(Norling Mjonell 1997). Under this assum
by substituting Equation 1 into Equati
obtains

w
w h
e mm+100 (mm
(b) M Type
( 100
Dh300
h) =mm) e

h t

w
&c + e &s + w

c
s

where we/h is the slope of the sorption/

isotherm (also called moisture capac
governing equation (Equation 3) must be
by appropriate boundary and initial conditi
The relation between the amount of e
water and relative humidity is called
isotherm if measured with increasing
humidity and desorption isotherm in th
case. Neglecting their difference (Xi et al.
the following, sorption isotherm will be
reference to both sorption and desorption c
By the way, if the hysteresis of the
(c) L type ( 150mm 150 mm 450mm)
isotherm would be taken into account, two
Figure 3. Test set-ups for flexural tests with different specimen
relation, evaporable water vs relative humi
sizes.
be used according to the sign of the varia
relativity
humidity.for The
shape of the
Photos illustrating
test set-ups
the flexural
isotherm
for HPC
is influenced
by many p
specimens are shown
in Figure
3. Detailed
informaespecially
those
that
influence
extent
tion about the flexural test set-up can be found in and
reactions
in turn,
Kim et al. (2008).chemical
The loading
speedand,
for both
ten- determ
structure
and
pore
size
distribution
sile and flexural tests, i.e. 1.06 mm min , was origi- (watercement
chemical
SF
nally determined ratio,
from the
tensile
test by composition,
assuming
curing
time
and
method,
temperature,
mix
that the static strain rate is 0.0001 sec which is acetc.).than
In the
variousspeed
formulatio
tually ten times faster
the literature
flexural loading
found
to
describe
the
sorption
recommended in ASTM C1609. The intention isotherm
of
concretespeed
(Xi etforal.both
1994).
However,
applying same loading
tensile
and in th
the semi-empirical
pro
flexural tests waspaper
to minimize
its effect on expression
the obNorling
Mjornell
(1997)
is
adopted
b
served results.
Proceedings of FraMCoS-7, May 23-28, 2010

(2)

The water content w can be expressed as the sum

Strain upto peak load (%)
of the evaporable
water we (capillary water, water
(a) Twisted fibers 1%
vapor, and adsorbed water) and the non-evaporable
(chemically bound)
water wn (Mills 1966,
Deformation (mm)
0
0.5 Mills 1 1995).1.5 It is reasonable to
Pantazopoulo
&
1
assume that
the evaporable water is a function of
6
relative0.8humidity, h, degree of hydration,
c, and
degree of silica fume reaction, s, i.e.5 we=we(h,c,s)
0.6
= age-dependent
sorption/desorption
isotherm
4
(Norling Mjonell 1997). Under this 3assumption and
0.4
by substituting
Equation 1 into Equation 2 one
2
obtains 0.2
0

0.2

0.4

0.6

0.8

Tensile stress (MPa)

Tensile stress (ksi)

Hooked fibers 1%

w h0
e +0 (0.2D h0.4) = w0.6e 0.8 we1 0
(3)
s
n
Strain
h upto peak
loadc(%)
h t
c
s
(b) Hooked fibers 1%
Figure 4. Tensile behavior of both Twisted and Hooked fibers
reinforced
(Kimslope
et al. of
2008).
the sorption/desorption
where wspecimen
e/h is the

&+

& + w&

isotherm (also called moisture capacity). The

Table
3. Average
experimental
results
the tengoverning
equation
(Equation
3) obtained
must befrom
completed
sile
tests
(Kim
et
al.
2008).
by appropriate boundary and Hooked
initial conditions.
Fiber
type
&
volume
content
1%
The relation between the1%amount ofTwisted
evaporable
water
and relative
First cracking
strengthhumidity is called adsorption
4.264
isotherm
if measured with4.299increasing
relativity
(MPa)
humidity
and
desorption
isotherm
in
the
Post cracking strength (MPa) 5.207
5.499opposite
case.
1994), in
StrainNeglecting
capacity (%)their difference
0.301 (Xi et al.
0.616
theNumber
following,
sorption
isotherm
will
be
used
with
referenceoftocracks
both(EA)
sorption and15desorption23conditions.
11.85 of the
7.74moisture
ByCrack
thespacing
way,(mm)
if the hysteresis
Average crack
49 different
isotherm
wouldwidth
be (m)
taken into37account, two
relation, evaporable water vs relative humidity, must
be Tensile
used according
to thecurves
sign of(strain
the variation
stress-strain
is valid of
up the
to
relativity
humidity.
The
shape
of
the
sorption
peak stress only) of the test series with Twisted and
isothermfiber
for HPC
is influenced
Hooked
are shown
in Figs.by4amany
and parameters,
4b, respecespecially
those
that values
influence
extent parameters
and rate of are
the
tively.
The
average
of tensile
chemical
reactions
and,
in
turn,
determine
pore
estimated from at least three specimens and summastructure
pore3. size
distribution
(water-to-cement
rized
in and
Table
These
parameters
include first
ratio, cement
chemical
composition,
SFstrain
content,
cracking
strength,
post cracking
strength,
cacuring attime
and method,strength,
temperature,
mix of
additives,
pacity
post-cracking
number
cracks
etc.). Inthethegauge
literature
various
formulations
be
within
length,
and average
crack can
width.
found
to
describe
the
sorption
isotherm
of
normal
Crack spacing and average crack width were esticoncretefrom
(Xithe
et al.
1994).
in thethepresent
mated
total
crackHowever,
length within
gage
paper [Kim
the semi-empirical
expression
proposed
by
length
et al. 2008]. Since
the modulus
of rupNorling
Mjornell
(1997)
is
adopted
because
ture (MOR) under bending is highly correlated withit
Proceedings of FraMCoS-7, May 23-28, 2010

10

10

12

1.6

Equiv. bending stress (MPa)

where the first term (gel isotherm) represents the

1.2
physically
bound (adsorbed) water and the8 second
term (capillary
isotherm) represents the capillary
0.8
water. This expression is valid only for low4 content
of SF. 0.4
The coefficient G1 represents the amount of
water per0 unit volumeL/150
held in theL/75gel pores0at 100%
0
0.04
0.08
0.16expressed
0.2
0.24(Norling
relative humidity, and it 0.12
can be
Mjornell 1997) as Deflection (inch)
(a) S type ( 50 mm 25mm 300mm )

c c+ ks s
G ( c s ) = k vg
c Deflection
vg s (mm)

(5)

0
2.8
c
vg
2.4

0.5

1.5

2.5

3.5

parameters. From the

where k and ksvg are material100X100X300
16that can
maximum amount of water per Twisted
unit volume
fibers
Hooked fibers
fill all pores
(both capillary pores and gel pores), one
2
can calculate
K1 as one obtains
12
1.6
1.2
0.8

s + 0.22 s G

0.188

)=
K ( c s0.4

1
1

g c 8c h

g c c h

10

1
L/150

10

L/75
1
0
0.04
0.08
0.12
0.16
c
s
Deflection (inch)
vg
vg

Equiv. bending stress (MPa)

(6)

The material parameters k and k and g1 can

( 100mmexperimental
100mm 300mm
be calibrated(b)byM fitting
data) relevant to
free (evaporable) water content in concrete at
Deflection
(mm) 2009b).
various ages (Di Luzio
& Cusatis
2.8

150X150X450
fibers
2.2 Temperature
evolution Twisted
Hooked fibers
2.4
16
Note that,2 at early age, since the chemical reactions
associated with cement hydration and SF12reaction
1.6
are exothermic,
the temperature field is not uniform
for non-adiabatic
systems even if the environmental
1.2
temperature
is constant. Heat conduction8 can be
described
0.8 in concrete, at least for temperature not
4
exceeding
100C (Baant & Kaplan 1996),
by
0.4
Fouriers law, which reads
L/150
L/75
Equiv. bending stress (MPa)

=0.2 JTwisted fiber 1%

we (h c s ) = G ( c Deflection
s ) (mm)
+

(g
)h
0
1
2
3

e 4 c 5 c 6 (4)

2.8
50X25X300
)h

(g
2.4
c
c 16

( fibers
KTwisted
)e
c
s
Hooked
fibers

2
Equiv. bending stress (ksi)

0.4

Tensile stress (MPa)

Tensile stress (ksi)

The proportionality
coefficient D(h,T) is called
Deformation (mm)
moisture0 permeability
and
it 1.5is a nonlinear function
0.5
1
1
of the relative
humidity h and temperature T (Baant
& Najjar
1972). The moisture mass6 balance requires
0.8
that the variation in time of the water
mass per unit
5
volume0.6 of concrete (water content w4 ) be equal to the
divergence of the moisture flux J

explicitly
accounts
for strength
the evolution
hydration
the
post cracking
tensile
of FRC,ofpost
crackreaction
andvalues
SF content.
This sorption
isotherm
ing
strength
for both Twisted
and Hooked
fireadsshould be noted, i.e., the post cracking strength
bers
with Twisted fibers is about 5.5MPa while that with
Hooked was about 5.2MPa.

Equiv. bending stress (ksi)

(1)

Equiv. bending stress (ksi)

J = Test
D (h,results
T ) h
3.3

= T 0

0.04

0.08

0.12

0.16

0.2

Deflection (inch)

(c) L ( 150mm 150mm 450mm )

0
0.24

(7)

where q is the heat flux, T is the absolute

temperature,
and
is the
heat conductivity;
in reinthis
Figure
5. Flexural
responses
of Twisted
and Hooked fiber
forced specimens.

where, f is equivalent bending stress; P is the applied

load; b is the width of the specimen; and h is the
depth of the specimen.
In referring to Figure 5, it is first observed that all
specimens generated deflection hardening behavior
accompanied by multiple cracks which appeared on
the bottom and side surfaces of the specimens. The
deflection hardening behavior was anticipated since
the tensile behavior of these composites using
Twisted and Hooked fibers produced strain hardening response under uniaxial tensile load (Fig. 4).
Generally, the specimens reinforced with Twisted fibers produced higher equivalent bending strength
than the specimens reinforced with Hooked fibers, as
expected from their tensile response (Fig. 4).
Equivalent bending strengths of specimens reinforced
with Twisted fibers are 15.79MPa for S, 16.63MPa
for M, and 12.70MPa for L series, while those with
of Hooked fibers are 14.02MPa for S, 11.45MPa for
M, and 11.56MPa for L series, respectively.
4 EVALUATION OF THE EXPERIMENTAL
RESULTS
Although the flexural response of all test series
shows deflection hardening behavior regardless of
their different geometry, noticeable difference is
observed in maximum equivalent bending strength,
deflection capacity, and energy absorption capacity according to the different geometry.
Deflection capacity is defined as the deflection
value at maximum equivalent bending stress and
has a strong influence on the energy absorption
capacity in flexure. Since the span length of each
series of specimens is different according to their
geometry (S, M, or L), the deflection values are
normalized by the length of span for comparison as
shown in Figure 6.
In addition, the equivalent bending strength is
normalized by the post cracking tensile strength
since there is difference in tensile strength between
the test series with Twisted or Hooked fibers.
Normalized equivalent bending strength versus
normalized deflection curves are graphically shown in

0.005

0.01

0.015

0.02

The proportionality
150X150X450 coefficient D(h,T)
100X100X300
moisture permeability
50X25X300 and it is a nonlinea
16
of the relative humidity h and
temperature
& Najjar 1972). The moisture mass balanc
that the variation in time of12 the water mas
volume of concrete (water content w) be eq
divergence of the moisture flux J

Twisted fiber
3
2.5
2
1.5
1

Equiv. bending stress (MPa)

(1)

3.5

= J

w
t

0.5

The
water content
w can be expressed a
L/150
L/75
of
the
evaporable
water
w0e (capillary wa
0
0
0.005
0.01
0.015
0.02
vapor, and adsorbed water) and the non-e
Deflection / Span
(chemically
bound) water wn (Mil
(a) Twisted fibers
Pantazopoulo
& Mills 1995). It is reas
assume that the evaporable water is a fu
Deflection
/ Span h, degree of hydration
relative
humidity,
0
0.005
0.01
0.02
degree of silica 0.015
fume reaction,
s, i.e. we=w
3.5
=
age-dependent
sorption/desorption
Hooked fiber
150X150X450
16
(Norling Mjonell
1997). Under
this assum
100X100X300
3
50X25X300
by substituting Equation 1 into Equati
obtains
2.5
12

1.5

w h
e + ( D h) = we
h
8 c
h t

Equiv. bending stress (MPa)

PL
bh 2

Equiv. bending stress / Tensile strength

f =

J Deflection
= D (h, T )/ Span
h

Equiv. bending stress / Tensile strength

The flexural responses of test series with Twisted

and Hooked fibers are shown in Figures 5a, 5b, and
5c according to the size of specimens, respectively.
All of the results shown in Figure 5 are averages
from three to six specimens.
Equivalent elastic bending strength is calculated
from the following equation suggested in ASTM
C1609.

& + e & + w
s

the sorption/
where we/h is the slope of
4
isotherm
(also
called
moisture
capac
0.5
governing equation (Equation 3) must be
L/75
L/150
by appropriate
boundary
and
0
0 initial conditi
0
0.005The 0.01
0.015
0.02
relation between the amount of e
Deflection
Span
water
and /relative
humidity is called
(b) Hooked if
fibers
isotherm
measured with increasing
Figure 6. Equivalenthumidity
bending stress
normalized
deflec- in th
and versus
desorption
isotherm
tion curves.
case. Neglecting their difference (Xi et al.
the
following, sorption
isotherm
will be
Figure 6 and the reference
cracking behavior
of typical
specito
both
sorption
and
desorption
c
mens is illustratedBy
in Figure
7. if the hysteresis of the
the
way,
It can be observed
from would
Figure be
6 that
both
isotherm
taken
intoequivaaccount,
two
lent bending strength
and
deflection
capacity
de- humi
relation,
evaporable
water
vs
relative
crease as the sizebeofused
specimen
increases.
Theofdeaccording
to isthehigher
sign
the varia
flection capacity relativity
of S type specimens
than of the
humidity.
The
shape
L/75, while the deflection
capacity
M and L type
isotherm
for
HPCMoreover,
isof influenced
by
many p
specimens is lower
than
L/150.
theextent
de- and
especially
those
that
influence
flection capacity chemical
of M typereactions
specimens
and,is generally
intheturn,
determ
higher than that ofstructure
L type specimens
although
dif- (waterand
pore
size
distribution
ference is small. ratio,
The maximum
equivalentcomposition,
bending
cement
chemical
SF
strength also shows
strong
size
dependency.
Normalcuring
time
and
method,
temperature,
mix
ized equivalent bending
of Twisted
fibersformulatio
reetc.).
Instrength
theforliterature
various
inforced specimens
are
2.87
S,
3.02
for
M,
and
found
to describe
the sorption
isotherm
2.31 for L sizes,concrete
while that
of
Hooked
fibers
rein(Xi et al. 1994). However, in th
forced specimens are 2.69 for S, 2.20 for M, and
paper
the
semi-empirical expression pro
2.22 for L sizes, respectively. The range of normalNorling
Mjornell
(1997) is adopted b
ized equivalent bending strength is 2.20 to 3.02, and
1

Proceedings of FraMCoS-7, May 23-28, 2010

J =
D (h, Tis)reasonable
h
this
range
according to theoretical pre(1)
dictions (Naaman, 2003)

The proportionality coefficient D(h,T) is called

moisture permeability and it is a nonlinear function
of the relative humidity h and temperature T (Baant
& Najjar 1972). The
mass balance requires
(a) Smoisture
type (Twisted)
that the variation in time of the water mass per unit
volume of concrete (water content w) be equal to the
divergence of the moisture
flux J
(b) S type (Hooked)
= J

(2)

w
t

The water content w can be expressed as the sum

M typew(Twisted)
of the evaporable(c)water
e (capillary water, water
vapor, and adsorbed water) and the non-evaporable
(chemically bound) water wn (Mills 1966,
Pantazopoulo & Mills 1995). It is reasonable to
assume that the evaporable water is a function of
type (Hooked)
relative humidity,(d)hM
, degree
of hydration, c, and
degree of silica fume reaction, s, i.e. we=we(h,c,s)
= age-dependent sorption/desorption isotherm
(Norling Mjonell 1997). Under this assumption and
by substituting Equation 1 into Equation 2 one
obtains
(e) L type (Twisted)

w h
e + ( D h) = we
h

h t

w
&c + e &s + w&n

c
s

(3)

L type of
(Hooked)
the sorption/desorption
where we/h is the(f) slope
Figure
7.
Cracking
behavior
of
flexural
isotherm (also called moisturespecimens.
capacity). The

governing equation (Equation 3) must be completed

the above
results,
is not
clear that the
by Given
appropriate
boundary
and itinitial
conditions.
flexural
responsebetween
can be used
order of
to predict
the
The relation
the in
amount
evaporable
tensile
behavior
of
material,
uniquely.
This
is
bewater and relative humidity is called adsorption
cause
the bending
tests produced
a strong relativity
size deisotherm
if measured
with increasing
pendency
not
only
for
their
bending
strength
but also
humidity and desorption isotherm in the opposite
their
deflection
capacity
and
their
span
length
or span
case. Neglecting their difference (Xi et al. 1994),
in
to
ratio. sorption isotherm will be used with
thedepth
following,
reference to both sorption and desorption conditions.
By the way, if the hysteresis of the moisture
CONCLUSIONS
5isotherm
would be taken into account, two different
relation, evaporable water vs relative humidity, must
This
study
investigated
correlation
betweenoftenbe used
according
to thethesign
of the variation
the
sile
and
bending
behavior
of
FRC
Composites
with
relativity humidity. The shape of the sorption
scale
effect
usingis three
different
isotherm
forbyHPC
influenced
by geometries.
many parameters,
especially those that influence extent and rate of the
ychemical
Althoughreactions
test series and,
with both
Hooked
and Twisted
in turn,
determine
pore
(Torex)
fibers
show
deflection
hardening
behavior
structure and pore size distribution (water-to-cement
undercement
flexural chemical
load, Twisted
(Torex) fiber
genratio,
composition,
SF led
content,
erally time
to higher
equivalent
bending strength
and decuring
and method,
temperature,
mix additives,
flection
capacity.
etc.). In the literature various formulations can be
yfound
Maximum
equivalent
bending isotherm
strength of
type
to describe
the sorption
of Snormal
specimen,
sameHowever,
cross-sectional
as
concrete
(Xiwhich
et al.has
1994).
in the area
present
the
tensile
specimen,
was
almost
three
times
higher
paper the semi-empirical expression proposed by
than theMjornell
post-cracking
strengthbecause
obtainedit
Norling
(1997)tensile
is adopted
Proceedings of FraMCoS-7, May 23-28, 2010

explicitly
accounts
for theThis
evolution
of hydration
from direct
tensile tests.
is very close
to anareaction
SF content.
Thisconditions
sorption (Naaman,
isotherm
lytically and
predicted
best case
reads
2003).
y As the size of specimen decreases, both the equiva the deflection capacity

lent bending strength and

in1

(h, , ) = G ( , ) 1
wcrease.
+
c s 10(g )h
e c s
1
y The ratio of equivalent bending
strength
direct

c to
1 c
e

(4)
tensile strength for all test series ranged from about
range

2.2 to 3, which is wellwithin

)h predicted
10(g the
c
1 c

) e
K1 ( c , s2003
1
analytically (Naaman,
)

suggests that tensile

y This limited investigation

response cannot be uniquely predicted from bending
where
the without
first term
(gel isotherm)
represents
the
response
consideration
to size
effect and
physically
bound
(adsorbed)
water
and
the
second
span length or span to depth ratio.

term (capillary isotherm) represents the capillary

water.
This expression is valid only for low content
It is hoped that additional studies will be carried
the amount
of
of SF.
Thefuture
coefficient
G1 represents
out
in the
to provide
additional data
to further
water
per
unit
volume
held
in
the
gel
pores
at
100%
resolve whether the tensile response can be uniquely
relative humidity,
and it can
be expressed
(Norling
predicted
from the bending
response,
and if so,
which
Mjornell
1997)
as
parameters are needed.
c c+ ks s
G ( c s ) = k vg
c vg s
6 ACKNOWLEDGEMENTS

(5)

k vg are material
parameters.
From
the
where
k vg anddescribed
The
research
herein was
sponsored
by the
maximum
amount
of water
per unitunder
volume
that No.
can
US
National
Science
Foundation
Grant
fill all0754505
pores (both
capillary
pores of
andMichigan.
gel pores), The
one
CMS
to the
University
obtains
can calculate
K1 as one
opinions
expressed
in this
paper are those of the authors and do not necessarily reflect the views of the

sponsor.
g h

c
c

w
s + s G e

c
s

(6)

( ) =
K
REFERENCES
c s

g h
e c c
c

10

0.188

0.22

10

ASTM C 1609/C 1690M-05, 2006, Standard test method for

flexural performance of fiber creinforceds concrete (using
and k vgSociety
and ofg1Testcan
The material
parameters
k vgAmerican
beam
with third-point
loading).
be ing
calibrated
by
fitting
experimental
data
relevant
to
and Materials, pp. 18.
free (evaporable)
content R.,
in 1994,
concrete
at
Bazant,
Z. P., Ozbolt, J.water
& Eligehausen,
Fracture
size effect:
Review
of &
evidence
concrete structures,
various
ages (Di
Luzio
Cusatisfor2009b).
ASCE Journal of Structural Engineering, Vol. 120, No. 8,
pp. 2377-2398.
Bazant,
Z.P. & Planas,
J., 1998, Fracture and Size Effect in
2.2 Temperature
evolution
Concrete and Other quasi-brittle Materials, CRC Press,
Note
that,
at and
early
age, pp.
since
Boca
Raton
London,
616.the chemical reactions
associated
with cement
reaction
Kim,
D. J., Naaman,
A. E. & hydration
S. El-Tawil, and
2008,SF
Comparative
behavior
four fiber reinforced
cementitious
areflexural
exothermic,
theoftemperature
field is not
uniform
composites,
Cement
and
Concrete
Composites,
Vol.30,
forNo.10,
non-adiabatic
systems even if the environmental
pp.917-928.
temperature is constant. Heat conduction can be
Kim, D. J., Naaman, A. E. & S. El-Tawil, 2008, High tensile
described
in concrete,
leastcomposites
for temperature
not
strength strain
hardeningatFRC
with less than
exceeding
100C
(Baant
&
Kaplan
1996),
by
2% fiber content, Proceedings of Second International
Symposium
Ultra reads
High Performance Concrete, GerFouriers
law,onwhich
many, E. Fehling, M. Schmidt and S. Strwald, Co-Editor,
q =Kassel
TUniversity Press GmbH, Heft 10, No. 10, pp. 169(7)
176
Lepech, M. & Li, V. C., 2004, Size Effect in ECC Structural
Members
in Flexure,
Proceedings
Vail,
where
q is
the heat
flux, Tof FRAMCOS-5,
is the absolute
Colorado,
USA,
pp.
1059-1066.
temperature, and is the heat conductivity; in this
Naaman, A. E. & Reinhardt, H. W., 1996, Characterization of
High Performance Fiber Reinforced Cement Composites,

Proceedings of 2nd International Workshop on HPFRCC,

A. E. Naaman and H. W. Reinhardt, eds., RILEM, No. 31,
pp. 1-24.
Naaman, A. E., 2003, Strain Hardening and Deflection Hardening Fiber Reinforced Cement Composites, High Performance Fiber Reinforced Cement Composites
(HPFRCC-4), A.E. Naaman and H.W. Reinhardt, Editors,
RILEM Publications, Pro. 30, pp. 95-113.
Soronakom, C. & Mobasher, B., 2007, Closed-Form MomentCurvature Expressions for Homogenized Fiber-Reinforced
Concrete, ACI Materials Journal, Vol. 104, No. 4, pp.
351-359.
Soronakom, C. & Mobasher, B., 2008, Correlation of tensile
and flexural responses of strain softening and strain hardening cement composites, Cement and Concrete Composites, Vol. 30, pp. 465-477.
Ward, R. & Li, V.C., 1990, Dependence of Flexural Behavior
of Fiber Reinforced Mortar on Material Fracture Resistance and Beam Size, ACI Materials Journal, V. 87, No.
6, pp. 627-637.

= D ( h , T ) h

The proportionality coefficient D(h,T)

moisture permeability and it is a nonlinea
of the relative humidity h and temperature
& Najjar 1972). The moisture mass balanc
that the variation in time of the water mas
volume of concrete (water content w) be eq
divergence of the moisture flux J
= J

w
t

The water content w can be expressed a

of the evaporable water we (capillary wa
vapor, and adsorbed water) and the non-e
(chemically bound) water wn (Mil
Pantazopoulo & Mills 1995). It is reas
assume that the evaporable water is a fu
relative humidity, h, degree of hydration
degree of silica fume reaction, s, i.e. we=w
= age-dependent sorption/desorption
(Norling Mjonell 1997). Under this assum
by substituting Equation 1 into Equati
obtains

w h
e + ( D h) = we
h

h t

w
&c + e &s + w

c
s

where we/h is the slope of the sorption/

isotherm (also called moisture capac
governing equation (Equation 3) must be
by appropriate boundary and initial conditi
The relation between the amount of e
water and relative humidity is called
isotherm if measured with increasing
humidity and desorption isotherm in th
case. Neglecting their difference (Xi et al.
the following, sorption isotherm will be
reference to both sorption and desorption c
By the way, if the hysteresis of the
isotherm would be taken into account, two
relation, evaporable water vs relative humi
be used according to the sign of the varia
relativity humidity. The shape of the
isotherm for HPC is influenced by many p
especially those that influence extent and
chemical reactions and, in turn, determ
structure and pore size distribution (waterratio, cement chemical composition, SF
curing time and method, temperature, mix
etc.). In the literature various formulatio
found to describe the sorption isotherm
concrete (Xi et al. 1994). However, in th
paper the semi-empirical expression pro
Norling Mjornell (1997) is adopted b
Proceedings of FraMCoS-7, May 23-28, 2010