You are on page 1of 29

FIBER REINFORCED CONCRETE

N. Banthia

Synopsis: The usefulness of fiber reinforced concrete (FRC) in


various civil engineering applications is indisputable. Fiber
reinforced concrete has so far been successfully used in slabs on
grade, shotcrete, architectural panels, precast products, offshore
structures, structures in seismic regions, thin and thick repairs,
crash barriers, footings, hydraulic structures and many other
applications. This paper presents a brief state-of-the-art report on
mechanical properties and durability of fiber reinforced concrete.
In particular, issues related to fiber-matrix interaction,
reinforcement mechanisms, standardized testing, resistance to
dynamic loads, and transport properties are discussed.

Keywords: FRC, fiber reinforced concrete, fiber-matrix


interaction, dynamic loads

1
Nemkumar Banthia is a Professor and Distinguished University
Scholar at the University of British Columbia, Vancouver, Canada. A
fellow of the ACI, the Canadian Soc. of Civil Engg., Indian Concrete
Institute and the Canadian Academy of Engineering, his awards include
the Wason Medal of the ACI and the Solutions through Research Award
of the Innovation Council of British Columbia.

INTRODUCTION

Compared to other building materials such as metals and polymers,


concrete is significantly more brittle and exhibits a poor tensile
strength. Based on fracture toughness values, steel is at least 100
times more resistant to crack growth than concrete. Concrete in
service thus cracks easily, and this cracking creates easy access
routes for deleterious agents resulting in early saturation, freeze-
thaw damage, scaling, discoloration and steel corrosion.

The concerns with the inferior fracture toughness of concrete are


alleviated to a large extent by reinforcing it with fibers of various
materials. The resulting material with a random distribution of
short, discontinuous fibers is termed fiber reinforced concrete
(FRC) and is slowly becoming a well accepted mainstream
construction material. Significant progress has been made in the
last thirty years towards understanding the short and long-term
performances of fiber reinforced cementitious materials, and this
has resulted in a number of novel and innovative applications.

There are currently 200,000 metric tons of fibers used for concrete
reinforcement. Table 1 shows the existing commercial fibers and
their properties. Steel fiber remains the most used fiber of all (50% of
total tonnage used) followed by polypropylene (20%), glass (5%)
and other fibers (25%).

2
PERFORMANCE CHARACTERISTICS OF FIBER
REINFORCED CONCRETE

Reinforcement Mechanisms

Concrete carries flaws and micro-cracks both in the material and at


the interfaces even before an external load is applied. These defects
and micro-cracks emanate from excess water, bleeding, plastic
settlement, thermal and shrinkage strains and stress concentrations
imposed by external restraints. Under an applied load, distributed
micro-cracks propagate coalesce and align themselves to produce
macro-cracks. When loads are further increased, conditions of
critical crack growth are attained at the tips of the macro-cracks and
unstable and catastrophic failure is precipitated.

The micro and macro-fracturing processes described above, can be


favorably modified by adding short, randomly distributed fibers of
various suitable materials. Fibers not only suppress the formation of
cracks, but also abate their propagation and growth.

Soon after placement, evaporation of the mix water and the


autogenous process of concrete hydration create shrinkage strains in
concrete. If restrained, this contraction can cause stresses far in
excess of those needed to cause cracking. In spite of every effort,
plastic shrinkage cracking remains a serious concern, particularly in
large surface area placements like slabs on grade, thin surface repairs,
patching and shotcrete linings. With large surface areas, fibers
engage water in the mix and reduce bleeding and segregation. The
result is that there is less water available for evaporation and less
overall free shrinkage [1]. When combined with post-crack bridging
capability of fibers, fibers reduce crack widths and cracks areas when
concrete is retrained [2], (Fig. 1).

In the hardened state, when fibers are properly bonded, they interact
with the matrix at the level of micro-cracks and effectively bridge
these cracks thereby providing stress transfer media that delays their
coalescence and unstable growth (Fig. 2). If the fiber volume fraction
is sufficiently high, this may result in an increase in the tensile

3
strength of the matrix. Indeed, for some high volume fraction fiber
composite [3], a notable increase in the tensile/flexural strength over
and above the plain matrix has been reported (Fig. 3). Once the
tensile capacity of the composite is reached, and coalescence and
conversion of micro-cracks to macro-cracks has occurred, fibers,
depending on their length and bonding characteristics continue to
restrain crack opening and crack growth by effectively bridging
across macro-cracks. This post-peak macro-crack bridging is the
primary reinforcement mechanism in the majority of commercial
fiber reinforced concrete composites.

Based on the discussion above, it emerges that fiber-reinforced


cementitious composites can be classified into two broad categories:
normal performance (or conventional) fiber-reinforced cementitious
composites and high-performance fiber-reinforced cementitious
composites. In FRCs with low to medium volume fraction of fibers,
fibers do not enhance the tensile/flexural strength of the composite
and benefits of fiber reinforcement are limited to energy absorption
or ‗toughness‘ enhancement in the post-cracking regime only. For
high performance fiber reinforced composites, on the other hand,
with a high fiber dosage, benefits of fiber reinforcement are noted in
an increased tensile strength, strain-hardening response before
localization and enhanced ‗toughness‘ beyond crack localization.

Fiber-Matrix Bond

As in any fiber reinforced composite, fiber-matrix bond in FRC is of


critical importance. However, unlike fiber reinforced polymers
(FRPs) used in aerospace and automobile industries where fibers are
employed to enhance strength and elastic modulus, in FRCs,
‗toughness‘ or energy absorption capability is of primary interest.
Therefore, inelastic bond failure mechanisms such as interfacial
crack growth, crack tortuousity and fiber slip are of greater relevance.
Fiber pull-out tests are often performed to assess fiber efficiency in
FRC and in such tests fiber bond and slip are monitored
simultaneously. Fig. 4 shows such a test, and the bond-slip curves
obtained [4].

4
For a fiber embedded in a cementitious matrix and subjected to a
pull-out load (Fig. 5), shear-lag will occur and interfacial debonding
will commence at the point of fiber entry which will slowly
propagate towards the free end of the fiber. Thus, some energy
absorption will occur at the fiber-matrix interface while the bond is
being mobilized and the fiber prepares to slip. Early in the
development of fiber reinforced concrete it became apparent that for
large, macro-fibers with small surface areas, a straight fiber will pull-
out at low values of interfacial stress and will generate stress in fiber
far below its tensile strength. Most commercial macro-fibers of steel
and other materials (polypropylene, for example) are now deformed
to enhance their bond with the surrounding matrix. However, even
here there is a limit. If deformed excessively, fibers may develop
stresses that exceed their strength and fracture in the process (Fig. 3).
The energy absorption in such cases is limited, and although some
fiber slippage may precede fracture, poor toughening ensues. For
maximized fiber efficiency, a pull-out mode of fiber failure where
pull-out occurs at a fiber stress close to its tensile strength is
preferred. It is important to mention that fiber failure mode is highly
dependent on the angle at which fiber is inclined with respect to the
direction of the pull-out force.

Fundamental Fracture Studies and Modeling

In the case of classically brittle materials like glass, Linear Elastic


Fracture Mechanics applies and fracture can be completely defined
by a single parameter called the ‗critical stress intensity factor‘, KC.
In micro-fracturing, strain-softening material like concrete, one
parameter description of fracture is not possible and multi-
parameter fracture criterion have been proposed [5,6]. In the case
of fiber reinforced concrete, in addition to crack closing pressure
due to aggregate interlocking, fiber bridging occurs behind the tip
of a propagating crack where fibers undergo bond-slip processes
and provide additional closing pressures. The fracture processes in
fiber reinforced cement composites are therefore even more
complex and advanced models are needed to simulate these
processes. Attempts have been made to model fracture in FRC
using the cohesive crack model [7] as well as the J-integral [8].
However, strictly speaking, these are only crack initiation criteria

5
and fail to define conditions for continued crack growth. To define
both crack initiation and growth, there is now general agreement
that a continuous curve of fracture conditions at the crack tip is
needed as done in an R-curve [9]. An R-curve (Figure 6) is a
significantly more suitable representation of fracture in FRCs, as
one can monitor variations in the stress intensity as the crack
grows and derive a multi-parameter fracture criterion.

A contoured double cantilever specimen is often used to obtain R-


Curves for FRC. A typical test is shown in Fig. 7, and the resulting
R-Curves are shown in Fig. 8 [10].

Standardized Tests for Toughness Measurements

Characterization of toughness (or energy absorption capability) of


FRC through standardized testing remains a hotly debated topic.
There is still no general agreement on how the toughness of fiber-
reinforced concrete should be measured [11-13]. ASTM has two
standards [14,15] and Japan Society for Civil Engineering (JSCE)
Standard SF-4 [16] is also often used. These three techniques and
their analysis schemes are compared in Table 2. Unfortunately, they
all treat toughness differently and there is little cross relationship
between the toughness parameters they produce [17].

Blast and Impact Resistance

Since 9/11, there has been an increased interest in developing


materials with enhanced resistance to explosive and impact loads.
Testing has clearly demonstrated that the ideal way to enhance the
impact resistance of concrete is by fiber reinforcement. Fibers
enhance the post-fracture stress transfer capability in concrete,
enhance dynamic fracture toughness, decrease dynamic crack
velocities and increase the absorption of energy under impact loads.

Drop Weight Impact Tests [18] are generally performed to measure


the resistance of fiber reinforced concrete to impact loads. In most
modern impact test systems, sufficient instrumentation is provided
such that loads, deformations, and velocities are simultaneously
measured; these are needed for a later analysis of the data. One

6
major issue that needs to be dealt with is that of inertial loading.
Specimen accelerates during a test and all measurements are made
while the specimen is still under acceleration. One commonly
adopted technique is to carry out direct measurements of
accelerations, and then use the principle of virtual work to derive
expressions for generalized inertial loads. For a simply supported
plate specimen impacted in the center, the generalized inertial load
(Pi(t)) is given by [19]:

hl 2  .x
Pi (t)  u(x, l/2, t).cosec (1)
4 l
where, l is width (also length) of the plate,  is the mass density, h
is the thickness of the plate and ü(x,y,t) = acceleration at any point
(x,y) on the plate at time t. Once the generalized inertial load is
obtained, the plate can be modelled as a Single Degree of Freedom
(SDOF) system and the generalized bending load can be obtained
from the Equation of dynamic equilibrium,
Pb (t)  Pt (t)  Pi (t) (2)

Similar expressions have been developed for beam specimens [18]


and other geometries [20].

For fiber reinforced concrete, while an improvement in impact


properties is widely reported, on a worrisome note, steel fibers are
reported [21] to fracture across cracks at high rates of loading and
thus produce a brittle response at very high strain-rates. As seen in
Fig. 9 and 10, SFRC may show increased brittleness under very
high strain rates.

The exact reasons of the observed brittleness of some FRC


materials under impact can be understood only via fundamental
testing of bond-slip mechanisms, fracture studies and modeling
[22]. In a recent study [23], a model proposed by Armelin and
Banthia [24] was adopted to predict the load-displacement
response of beam under impact (Fig. 11). The compressive strain,
o, at the top-most fiber of the specimen leads to an axial
shortening, o, as shown. This in turn leads to stress, c, in the

7
uncracked concrete. On the other hand, it results in fiber slippage,
wi, below the neutral axis and corresponding forces, fi, as the fibers
pull-out. Thus, the flexural load carried during the post-crack
phase is obtained by satisfying the equilibrium of moments:
2M e
P (3)
l

The equilibrating moment, Me, may be calculated by summing the


moments generated by concrete stresses and the individual
moments generated by the N individual fibers bridging the crack
below the neutral axis. It follows from Fig. 11, that
c' N

  c b.dy   f i  0 (equilibrating forces)


0 1
(4)

c' N
M e    c b.dy  y    f i y i  (equilibrating moments) (5)
0 1

where b is the width of the beam, c’ is the depth of the uncracked


section and y is the distance from the neutral axis.

In the model, the pull-out force in each fiber (fi) is expressed as a


function of the crack width, wi, according to the average pull-out
force versus slip (or crack width) relationship obtained
experimentally at the full embedment length, le=l/2. To enable this,
single fibers must be pulled out from a concrete matrix at various
inclinations with respect to the pull-out load. The bond-slip
response is then represented using the Ramberg-Osgood
formulation so that the force carried by each fiber may be
expressed in terms of its orientation, i and the slip, wi, as follows:
 
 1 A 
f i  i , wi   E p wi  A   (6)
 
1
 1  Bwi  
C
 
C

where the constants A, B, C and Ep, are obtained for each


orientation through the Ramberg-Osgood formulation.
Recognizing that the average force in the fibers at a layer which is

8
at a distance ‗y‘ from the neutral axis is averaged over the entire
range of embedment and inclination that is possible, the value of
‗fi‘ in Equations 4 and 5 may be computed as follows (24):

1  f 0 w f w 1 
fi    f 22.5 w  f 45 w  f 67.5 w  90   f geometry w (7)
2  2 2 4 
or
1  f w f w 1 
f i   0  f15 w  f 30 w  f 45 w  f 60 w  f 75 w  90   f geometry w (8)
2  2 2 6 

The resulting model prediction is plotted together with the


experimental flexural response under impact in Fig. 12. Not
surprisingly, the analytical response is monotonic and predictions
are not excellent. Nevertheless the model can approximate the
experimental response, and predict both peak load and toughness.
For proper predictions, one needs to somehow engage information
on crack velocities and changes in the stress-intensity factors at the
tip of a fast moving crack. Such attempts are currently underway.

Bio-Inspired FRCs for Longer Service Life

The decreasing life span of concrete structures is becoming an


issue of greater and greater importance for societies. The primary
problem is the corrosion of the steel in reinforced concrete
structures. Chloride penetration and carbonation are the primary
reasons for such corrosion and any measures aimed at mitigating
the ingress of chlorides or CO2 into the body of concrete are
expected to significantly enhance the durability of concrete
structures. These deleterious agents enter the body of concrete
through one of the three transport mechanisms: diffusion, capillary
sorption and permeability—of these, the permeability is considered
as the dominant mechanism. Any measures adopted to reduce
permeability of concrete will therefore help in preserving
durability. Results have indicated that permeability, in turn, is
highly dependent upon cracking in concrete and an increase in the
crack width will not only produce a highly permeable concrete (Fig.
13) but also enhance the possibility of rebar corrosion, Figure 14,
Bentur [25].

9
There is an increased interest these days in bio-inspired fiber
reinforced cementitious composites. These are based on bio-
degradable and recyclable natural fibers (such as cellulose) and are
both environmentally friendly and sustainable. Current research at
The University of British Columbia is focussing on such
composites, and the preliminary indications are that such bio-
inspired materials are highly promising in building and
regenerating a sustainable infrastructure.

The influence of an externally applied stress on the permeability of


concrete remains poorly understood. Banthia and co-workers [26]
& [27], developed a novel technique of measuring the permeability
of concrete under an applied stress and investigated the benefits of
fiber reinforcement. The permeability cell was mounted directly in
a 200 kN hydraulic Universal Testing Machine (UTM) such that a
uniform compressive stress could be applied directly on the
concrete specimen housed in the cell. The water collected was
related to the coefficient of water permeability (Kw) by applying
Darcy‘s law:
QL
Kw  (9)
Ah
Kw = Coefficient of water permeability (m/s), Q = Rate of Water
Flow (m3/s), L = Thickness of specimen wall (m), A = Permeation
area (m2) and h = Pressure head (m).

Their data are plotted in Fig. 15. Notice that under conditions of
no-stress, fibers reduce the permeability of concrete, and the
reduction appears to be proportional to the fiber volume fraction.
Data further indicates that stress has a significant influence on the
permeability of concrete. When stress was first increased to 0.3fu,
both plain and FRC showed a decrease in the permeability.
However, when the stress was increased to 0.5fu, plain and FRC
showed very different trends. At 0.5fu, the permeability of plain
concrete increased substantially over that of the unstressed
specimen, but for FRC, while there was an increase in the
permeability over 0.3fu, the permeability still stayed below that of
the unstressed specimen.

10
The above observations can be related to cracking. At 0.3fu, it is
conceivable that in both plain and FRC, there is no discernible
cracking that can affect the flow of water. However, at 0.3fu, the
stress-strain response for both plain and FRC would become non-
linear indicating the presence of cracking. As given by the
Poiseuille Law, Edvardsen [28], the flow of water through cracks
is proportional to the cube of the crack width. In the case of FRC,
one can expect the fibers to suppress cracking and hence maintain
the rate of flow similar to an unstressed specimen. When combined
with the phenomenon of ‗pore compression‘, this implies that the
permeability of FRC under stress can in fact be lower than that of
an unstressed specimen.

Bhargava and Banthia [27] extended the permeability data


described above towards service life prediction. Most service life
prediction models for concrete involve the use of diffusion
coefficients Tutti [29]. Unfortunately, studies relating different
transport coefficients are rare. In particular, experimental data
relating permeability and diffusion coefficient is lacking, and only
a theoretical correlation can be established between these two
coefficients via a correlation constant, as follows:

Empirical equations for the permeability coefficient were proposed


by Hedegaard et al. [30] and for diffusion coefficient were
proposed by Hansen et al. [31] as follows:

  c  0.31 f  
K w  exp 4.3   4.0  (10)
  w  

 c  0.3 f 
  7.0 
D  1.7 x10  w  (11)

where,

Kw= water permeability coefficient (m/s)


D= Chloride ion diffusion coefficient, in cm2/s
c= cement content of concrete, in kg/m3

11
w= water content of concrete, in kg/m3
f= fly-ash content of concrete, in kg/m3

By substituting the values of c, w and f for the concrete mixture


used in the permeability tests in Eqs. (10) and (11), one obtains
Kw=1.07x10-10 (m/s) and D = 7.89x10-13 (m2/s).

Further, the permeability K (m2) of a single straight pore with


effective pore radius reff embedded in a medium of cross-sectional
area A can be related to effective pore radius by assuming Hagen-
Poiseuille’s law to be valid for small pores.
reff4
K (12)
8A
where reff is the effective pore radius defined as the radius of the
effective pores which take part in the transport. Also, the diffusion
coefficient can be related to the area fraction of effective pores as,
reff2
D  Do a eff  Do (13)
A

where a eff = is the area fraction of effective pores


Do = is the diffusion coefficient in a bulk fluid

Assuming that the effective pore radius in Eqs. (12) and (13) is the
same, a general relationship between permeability K (m2) and
diffusion coefficients D (m2/s) emerges,
2
reff
K D (14)
8 Do
Further, it is to be noted that an interconnected pore system is
necessary for a continuous network of flow paths to be available
for various transporting media. In saturated conditions, the steady
state flow coefficient can be related to the water permeability
coefficient as the two processes occur simultaneously,

12
K (15)
Kw 

g

Using Eqs. (14) and (15), the water permeability coefficient K w


(m/s) and the diffusion coefficient D (m2/s) can be related as,
2
reff g
Kw  D (16)
8 Do
Where Kw as before is the water permeability coefficient (m/s),
D is the diffusion coefficient (m2/s),
reff is the effective pore radius,
 is the viscosity of water (Ns/m2),
 is the density of water (kg/m3) and,
g is the gravity (m/s2)

This equation corresponds to Katz-Thompson Equation, Garboczi


[32], and is based on the assumption that the effective radius
affecting the permeability and the diffusion coefficient is the same.
Equation (16) can be further modified to consider the effect of
stress and the fibers on concrete. Since the permeability coefficient
is proportional to the fourth power of effective pore radius Eq. (12)
and since the normalized permeability coefficient is related to the
water permeability coefficient of unstressed plain concrete through
the previously defined factors F and S, describing, respectively, the
influence of fiber reinforcement and stress Bhargava & Banthia
[27], the effective pore radius can be modified to:

*
rnormalized  F 0.25 S 0.25 reff (17)

where, r* normalized is the effective pore radius corresponding to


normalized permeability values and reff in this case is the effective
pore radius of plain concrete under zero stress condition.

Substituting Eq. (17) into Eq. (16), we get a modified equation


which relates normalized water permeability to diffusion
coefficient as,

13
K normalized  CF 0.5 S 0.5 D (18)

2
reff g
where C = is a constant proportional to second power of the
8 Do
effective pore radius of plain concrete under zero stress condition.

For plain concrete and zero stress condition F=S=1 and for this
case:
K normalized  K w plain  unstressed  CxD (19)

Substituting the empirical values of the water permeability


coefficient Kw=1.07x10-10 m/s and the chloride ion diffusion
coefficient D = 7.89x10-13 m2/s, as obtained previously, the value
of constant C for the concrete in question can be calculated:
C = 135.62 m-1 (20)
The constant C computed above takes into consideration the
effective pore radius of plain concrete under zero stress condition
and properties of the chloride ion diffusion coefficient. The
calculated chloride ion diffusion coefficients are given in Table 3.

The Durability Factor, D, for a given concrete under a given stress


level can be defined as the ratio of its expected service life to that
of companion plain concrete under zero stress. Using Tuutti‘s
model [29], ingress of chlorides is estimated by a one-dimensional
diffusion process using the Fick‘s Second Law of diffusion. For
non-steady state condition, the chloride concentration C at a
location x and at a time t is given by Crank [33].

C   C 
 D  (21)
t x  x 
Here, the diffusion coefficient D may be a constant or a function of
other variables such as chloride concentration, location, time,
temperature, etc.

14
For a simple case with known geometry and boundary conditions
where the diffusion coefficient D can be assumed to be a constant,
solution to Eq. (21) is given by Newman [34]:
  x 
C ( x, t )  C s 1  erf   (22)
  2 Dt 
2 z t 2
erf ( z )  e dt (23)
 0
where,
erf is a standard error function,
x is effective concrete cover depth,
C s is the concentration of the chloride ions at the outside surface
of the concrete and is assumed to be constant with time. That is,
C  C s for x = 0 and for any t
C i is the concentration at the depth of the reinforcement; assumed
to be zero at t =0.
C t is the threshold concentration required to initiate steel
reinforcement corrosion. The initiation period is accomplished
when C i  C t and,
t = time
Eq. (22) can be solved by using a normal standard distribution
Bertolini et al. [35]:
erf ( z)  2 N ( z 2 )  1 (24)

t2
1 z 2 2
N (z 2)   e dt (25)
2 

The initiation time can thus be calculated by assuming a constant


diffusion coefficient for concrete, a known surface chloride content
(dictated by the environment), the thickness of the concrete cover
and critical chloride ion content at which onset of corrosion is
expected.

15
Solving the above equation for Ct = threshold concentration of
chloride ions = 0.50 % (based on the mass of cement), C s =chloride
ions concentration at the surface of concrete = 0.70 % (based on
the mass of cement), x = 25 mm, and diffusion coefficients, D,
from Table 1:
x2
t  ti  (26)
0.2678D
Notice that a lower value of 0.50% threshold concentration of
chloride ions was chosen due to the presence of fly-ash in concrete
which is known to increase the rate of corrosion. The above
equation predicts that service life of any concrete is proportional to
x2, and holds an inverse relationship with the chloride ion diffusion
coefficient. Therefore doubling the concrete cover increases
service life of concrete by a factor of 4, whereas a 10-fold
reduction in diffusion coefficient will result in a 10-fold increase in
the predicted service life. Substituting the values of diffusion
coefficient from Table 3 into Eq. 26 for different concrete types
and stress conditions, the Durability Factors were computed and
are plotted in Fig. 16. Notice in Figure 16 that as per the model,
fiber reinforcement can be effective in enhancing the durability of
concrete under both stressed and unstressed conditions.

CONCLUDING REMARKS

A brief state-of-the-art report on fiber reinforced concrete is


presented. Our understanding of fiber-matrix interaction,
reinforcement mechanisms and performance characteristics is fairly
advanced. Fiber reinforced concrete is a promising material to be
used in the Middle-East for sustainable and long-lasting concrete
structures. Its performance has already been proven in other hot and
arid climates and in other chemically deleterious environments.

16
REFERENCES

[1]. Weiss, W.J., Shah, S.P., ―Recent trends to reduce shrinkage cracking in
concrete pavements,‖ Proceedings of the Airfield Pavement Conference,
Aircraft/Pavement Technology: In the Midst of Change, pp. 217-228; 1997.
[2]. Banthia, N., Gupta R., and Mindess, S. Developing crack resistant FRC
overlay materials for repair applications, NSF Conference, Bergamo, Italy;
2004.
[3]. Banthia, N. and Sheng, J., Fracture Toughness of Micro-Fiber Reinforced
Cement Composites, Cement and Concr. Composites., 18: pp. 251-269;
1996.
[4]. Bindiganavile, V and Banthia, N., Polymer and Steel Fiber Reinforced
Cementitious Composites under Impact Loading, Part 1: Bond-Slip
Response, American Concrete Institute, Materials Journal, Vol. 98, No.1,
pp. 10-16: 2001.
[5]. Hillerborg, A. et al, Cement and Concrete Research, 6, pp. 773-82; 1976.
[6]. Jeng, Y.S. and Shah, S.P., ASCE J of Engineering Mechanics, 111(10): pp.
1227-41;1985.
[7]. Hillerborg, A., Cement and Concrete Composites, 2, pp. 177-84; 1980.
[8]. Mindess, S. et al, Cement and Concrete Research, 7, 731-742; 1977.
[9]. Mobasher, B., Ouyang, C. and Shah, S.P., Int. J. of Fracture, 50, pp. 199-
219; 1991.
[10]. Banthia, N., and Genois, I., Controlled Crack Growth Tests for Optimization
of Micro-Fiber Reinforced Cement Composites, ACI, Special Publication:
Application and Testing of Fracture Mechanics Concepts (Ed. C.
Vipulanandan), SP-201, pp. 55-74; 2000.
[11]. Banthia, N. and Trottier, J.-F., ―Test Methods of Flexural Toughness
Characterization: Some Concerns and a Proposition,‖ Concrete Int.: Design
& Construction, American Concrete Institute, Materials Journal, 92(1), pp.
48-57; 1995.
[12]. Barr, B., Gettu, R., Al-Oraimi, S.K.A., and Bryars, L.S., ―Toughness
Measurements-the Need to Think Again,‖ Cement and Concrete Composites,
18, pp. 281-297; 1996.
[13]. Gopalaratman, V.S., et al, ―Fracture Toughness of Fiber Reinforced Concrete‖,
ACI Materials Journal, July-August 1991, pp. 339-353, and Johnston, C.D.,
Discussion of above paper, ACI Materials Journal, pp. 304-309; May-June
1992.

17
[14]. ASTM C 1609/C 1609M – 05, Standard Test Method for Flexural
Performance of Fiber-Reinforced Concrete (Using Beam With Third-Point
Loading), ASTM International, PA, United States.
[15]. ASTM C 1399-98, ―Test Method for Obtaining Average Residual-Strength of
Fiber-Reinforced Concrete,‖ Annual Book of ASTM Standards, Vol. 04.02.
Japan Society of Civil Engineers; 2004.
[16]. Standard SF-4 Method of Test for Flexural Strength and Flexural Toughness
of Fiber Reinforced Concrete, Japan Society of Civil Engineers; pp. 58-66;
1984.
[17]. Banthia, N. and Mindess, S., Toughness Characterization of Fiber Reinforced
Concrete: Which Standard to Use? ASTM, J. of Testing and Evaluation, 32(2),
pp. 138-139; 2004.
[18]. Banthia, N., Mindess, S., Bentur, A. and Pigeon, M., Expt. Mech. 29 (2): pp.
63-69; 1989.
[19]. Gupta, P. et al., Journal of Materials in Civil Engrg., ASCE, 12 (1) 81-90;
2000.
[20]. Bindiganavile, V. and Banthia, N., Generating Dynamic Crack Growth
Resistance Curves for Fiber Reinforced Concrete, Experimental Mechanics
in press. 2004.
[21]. Bindiganavile, V and Banthia, N., Polymer and Steel Fiber Reinforced
Cementitious Composites under Impact Loading, Part 2: Flexural
Toughness, American Concrete Institute, Materials Journal, Vol. 98(1): pp.
17-24; 2001.
[22]. Kaadi, G.W., MS Thesis, The University of Illinois, Chicago, (1983).
[23]. Banthia, N., Impact and Blast Protection with Fiber Reinforced Concrete,
Conference Proceedings - BEFIB, Veronna, Italy, RILEM, 39, pp. 31-44;
2004.
[24]. Armelin, H. and Banthia, N., ACI Mat. J., 94(1): pp. 18-31; 1997.
[25]. Bentur, A., et al. 2005. Comprehensive Approach for the Design of
Concrete for Durability and Long Term Performance of Structures.
ConMat05 Mindess Symposium Proc., University of British Columbia (Ed.
Banthia, Bentur and Shah), 10 pp.
[26]. Banthia, N. and Bhargava, A., 2007. American Concrete Institute, Materials
Journal, 104(1), pp. 303-309.
[27]. Bhargava, A. and Banthia, N. 2008. RILEM, Materials and Structures, 41:
363-372.
[28]. Edvardsen, C., 1999. ACI Materials Journal, 96(4): 448-454.

18
[29]. Tuutti, K., 1982. Corrosion of steel in concrete. Swedish Cement and
Concrete Research Institute, Stockholm, Sweden. CBI Research Report no.
4.82.
[30]. Hedegaard, S.E. and Hansen, T.C., 1992. Materials and Structures, 25: 381-
387.
[31]. Hansen, T.C., Jensen, J., Johannesson, T. 1986. Cement and Concrete
Research, 16(5): 782-784.
[32]. Garboczi, J., Cement and Concrete Research, 20 (4) (1990) 590-601.TITLE
[33]. Crank J., ―Mathematics of diffusion‖, Oxford: Clarendon Press, 1956.
[34]. Newman, A.B. 1970. American Institute of Chemical Engineers, Vol. 27
pages.
[35]. Bertolini, L., Elsener, B., Pedeferri, P., and Polder, R., WILEY-VCH Verlag
GmbH and Co. kGaA, Weinheim (2004).

19
TABLES

Table 1. Properties of Fibers used as Reinforcement in Concrete

Fiber type Tensile Tensile Tensile Fiber Alkali


strength modulus strain (%) diameter stability,
(MPa) (GPa) (max-min) (m) (relative)
Asbestos 600-3600 69-150 0.3-0.1 0.02-30 excellent
Carbon 590-4800 28-520 2-1 7-18 excellent
Aramid 2700 62-130 4-3 11-12 good
Polypropylene 200-700 0.5-9.8 15-10 10-150 excellent
Polyamide 700-1000 3.9-6.0 15-10 10-50 -
Polyester 800-1300 up to 15 20-8 10-50 -
Rayon 450-1100 up to 11 15-7 10-50 fair
Polyvinyl
800-1500 29-40 10-6 14-600 good
Alcohol
Polyacrylonitrile 850-1000 17-18 9 19 good
Polyethylene 400 2-4 400-100 40 excellent
Polyethylene
---------- ---------- ---------- 1-20 excellent
pulp (oriented)
High Density
2585 117 2.2 38 excellent
Polyethylene
Carbon steel 3000 200 2-1 50-85 excellent
Stainless steel 3000 200 2-1 50-85 excellent
AR- Glass 1700 72 2 12-20 good

20
Table 2. Description of Test Methods
Standards ASTM C 1609-08 JSCE SF-4 ASTM C 1399-98

Test
Specimen

Initial Loading Reloading


Test A beam specimen is quasi-statically loaded at its third-points to A stable narrow crack is first created in the specimen by applying a
Description failure and the resulting load vs. net center point deflection is flexural load in series with a steel plate under controlled conditions.
plotted for further analysis. The plate is then removed, and the specimen is reloaded in flexure to
obtain the post-crack load vs. net displacement curve.

First Crack Initial Loading Curve


Load, N

Load, N
A
B C D
E Reloading Curve (Pre-cracked Beam)
Typical
Curve F G H P0.5 P0.75 P1.0 P1.25
I J
oδ 3δ 5.5δ 10.5δ δtb = L/150
Net Deflection, mm
Net Deflection, mm

Px,y = Load at displacement Flexural Toughness (Tb) = Area OAEJ


y for a x mm section
Flexural Toughness Factor (FT)
f150,0.75 (MPa): Residual Tb x L
FT =
strength at P150,0.75  tb x bd 2 Average Residual Strength
Analysis
f150,3.0 (MPa): Residual FT
strength at P150,3.0 Re 2(%)  x100 ARS = ((P0.5+P0.75+P1.0+P1.25)/4) x L/bd2
MOR
Toughness 150,3.0 (J): The MOR = Modulus of Rupture
energy to a net deflection of b = Breadth of the Beam
1⁄150 of the span (3.0 mm d = Depth of the Beam
for a 150 mm specimen)

21
Table 3. Computed Values of Chloride Ion Diffusion Coefficient
Fiber Applied Normalized Chloride
Volume Stress water ion
Fraction Level permeability F S diffusion
Vf coefficient coefficient
Knormalizedx10- Dx10-13
10
(m/s) (m2/s)
0.0fu 1.66 1 1 12.24
0.0% 0.3fu 103 1 0.62 9.64
0.5fu 2.30 1 1.38 14.43
0.0fu 0.95 0.57 1 9.27
0.1% 0.3fu 0.53 0.57 0.57 6.85
0.5fu 0.71 0.57 0.76 7.95
0.0fu 0.60 0.36 1 7.37
0.3% 0.3fu 0.32 0.36 0.53 5.40
0.5fu 0.45 0.36 0.75 6.38
0.0fu 0.30 0.18 1 5.21
0.5% 0.3fu 0.10 0.18 0.33 3.02
0.5fu 0.18 0.18 0.62 3.97

22
FIGURES

2.5 F1 F2
F3 F4

Average Crack Width (mm)


2.0 F5 F6
F7

1.5

1.0

0.5

0.0
0.0% 0.1% 0.2% 0.3% 0.4%
Volume Fraction (%)

Fig. 1. (Left) Plastic Shrinkage Crack Control Efficiency with


Increasing Fiber Volume Fraction from Top to Bottom and (Right)
Maximum Crack Width for Various Fibers (F1-F7).


peak


Micro crack Macro crack
Formation Formation

Fig. 2. Fiber Reinforcement Before and After the Creation of


a Macro-Crack (Left) and Crack Bridging by Fibers (Right).

23
Fig. 3. (Left) A CFRC Composite in Tension and (Right) Stress-Strain
Curves Showing Strain-Hardening at High Fiber Volume Fractions.

0.35

0.3

0.25 Excessively Deformed Fiber (Fracture)


Load (kN)

0.2

0.15 Deformed Fiber (Pull-Out)

0.1
Undeformed
0.05

0
0 5 10 15
Slip (mm)

Fig. 4. (Left) A Fiber Pull-Out Test and (Right) Bond-Slip Pull-Out


Curves for Various Deformed Fibers. (Notice fiber fracture in an
excessively deformed fiber.)

24
Fig. 5. Shear-Lag in a Bonded
Fiber with Inelastic Mechanisms

Fig. 6. R-Curve Representation of


Fracture in FRC.

Fig. 7. A CDCB Before and


After Fracture.

Fig. 8. R-Curves Generated from CDCB Tests


Shown in Figure 7

25
90
80 Steel Fiber
Toughness (Nm) Polypropylene
70
Fiber
60
50
40
30
20
10
0
200 500 750 1000
Drop Height (mm)

Fig. 9. Impact Resistance of Steel FRC and Polypropylene FRC. (Note


the increase in brittleness in SFRC at high rates of loading.)

350
300
250 Low Strain-Rate Impact
Load (kN)

200
150 High Strain-Rate Impact

100
Quasi-Static
50
0
0 0.2 0.4 0.6 0.8 1
Deflection (mm)

Fig. 10. Impact Response of SFRC Beams.


(Notice brittleness at high strain-rates.)

26
0
0

uncracked
strains

c

section
displacements

f2 f1
2 ƒi = f(w i , i , li )

fi.. f3
wi 3
i

Fig. 11. Schematic View of Forces and Stresses Acting on the


Cracked Section of an SFRC Beam.

45
Experimental Response Analytical Response
40 Peak Load
38 34
(kN)
35 JSCE
Toughness 6.30 7.05
30 Factor (MPa)
load (kN)

25

20

15

10

5 Experimental
Analytical
0
0 0.5 1 1.5 2 2.5 3
deflection (mm)

Fig. 12. Model Predictions under Impact Loading. (Note that


predicted response is monotonic, but predicts both peak load and
toughness.)

27
Fig. 13. Effect of Crack Width on Permeability [1].

Fig. 14. The Effect of Crack Width on Corrosion Potential;


a Potential below –280mV Indicates Corrosion Initiation
and below –400mV Indicates Active Corrosion.

28
-1 0
. 2.5

Normalized Permeability Coefficient x 10


0.0% Fiber
2 0.1% Fiber
0.3% Fiber
1.5 0.5% Fiber
(m/s)
1

0.5

0
0 0.1 0.2 0.3 0.4 0.5
Stress Level ( f u )

Fig. 15. Normalized Permeability Coefficients.

4.5
4
Durability Factor, D

3.5 0.0% Fiber


3 0.1% Fiber
2.5 0.3% Fiber
2 0.5% Fiber

1.5
1
0.5
0
0 0.2 0.4 0.6
Stress Level (f u )

Fig. 16. Durability Factors. (Notice durability enhancements with


fibre reinforcement.)

29