You are on page 1of 21

Materials and Structures (2009) 42:12611281

DOI 10.1617/s11527-009-9529-4

ORIGINAL ARTICLE

Fibre reinforced concrete: new design perspectives


Marco di Prisco Giovanni Plizzari Lucie Vandewalle

Received: 11 July 2007 / Accepted: 11 June 2009 / Published online: 10 September 2009
 RILEM 2009

Abstract Although the use of Fibre Reinforced the main concepts behind the structural rules for FRC
Concrete (FRC) for structural applications is contin- structural design.
uously increasing, it is still limited with respect to its
potentials, mainly due to the lack of International Keywords Fibre Reinforced Concrete 
Building Codes for FRC structural elements. Within Constitutive laws  Safety factors 
fib (Federation Internationale du Beton), the Special Characteristic length  Structural behavior 
Activity Group 5 is preparing a New fib Model Code Redundancy  Structural design
that aims to update the previous CEB-FIP Model
Code 90, published in 1993, that can be considered as
the reference document for Eurocode 2. The New
Model Code includes several innovations and 1 Introduction
addresses among other topics, new materials for
structural design. In this respect, FRC will be Fibre Reinforced Concrete (FRC) is a composite
introduced. The Technical Groups fib TG 8.3 Fibre material that is characterized by an enhanced post-
reinforced concrete and fib TG 8.6 Ultra high cracking tensile residual strength, also defined as
performance FRC are preparing some sections of toughness in the following, due to the fibre rein-
the New Model Code, including regular and high forcement mechanisms provided by fibres bridging
performance FRC. This paper aims to briefly explain the crack surfaces.
To enhance concrete toughness for structural
applications, high-modulus fibres can be used to
M. di Prisco (&) substitute, partially or totally, conventional reinforce-
Department of Structural Engineering, Politecnico di
Milano, Piazza Leonardo da Vinci 32, 20133 Milan, Italy ment. Other types of fibres, having usually a low
e-mail: marco.diprisco@polimi.it modulus and a small size (length of few millimetres
and diameter of few microns) can be used to reduce
G. Plizzari shrinkage cracking and to enhance fire resistance.
Department of Civil Engineering, Architecture, Land
and Environment, University of Brescia, Brescia, Italy Hybrid system of fibres (HyFRC) is a more recent
e-mail: plizzari@ing.unibs.it development in fibre reinforced cement-based tech-
nology, based on the purpose of optimizing several
L. Vandewalle material performances [14]. In a HyFRC, two or more
Department of Civil Engineering, K. U. Leuven,
Heverlee, Belgium different types of fibres can be properly combined to
e-mail: lucie.vandewalle@bwk.kuleuven.be produce a composite, whose mechanical and physical
1262 Materials and Structures (2009) 42:12611281

performances take benefits from each type of fibres and reinforcement, while RILEM TC162-TDF produced
from a possible synergistic response. design guidelines for typical structural elements [7,
Since fibres can be considered as a smeared 8]. Afterwards, recommendations were produced by
reinforcement, FRC can be properly used to assure other Countries, as France [9], Sweden [10], Ger-
a more favourable crack distribution and to limit many [11], Austria [12] and Italy [13].
crack width at Serviceability Limit States (SLS). Due to a better knowledge of FRC and the recent
Fibres can also be used as partial substitution of developments worldwide of guidelines for structural
conventional reinforcement (rebars or welded mesh) design, the fib Special Activity Group 5 (SAG 5),
and, in some structures, they can totally substitute who is preparing the new fib Model Code, decided to
rebars. These structures are, in general, characterized introduce some sections on FRC. Working Groups
by a high degree of structural redundancy and TG 8.3 (Fibre Reinforced Concrete) and TG 8.6
represent a type of applications where fibres can be (Ultra High Performance Fibre Reinforced Con-
a more effective reinforcement system, since a cost crete) of fib are preparing these sections of the new
saving can be achieved. fib Model Code concerning FRC design rules for
During the last three decades, a wide research has providing a guidance to engineers to properly (and
been performed on material properties of FRC, both safely) design FRC structural elements both at
at fresh and hardened state [14]. Research on serviceability (SLSs) and ultimate limit states
structural response of FRC elements was mainly (ULSs), based on the state-of-the-art knowledge.
developed during the last fifteen years. As a conse- This paper aims to present some principles gov-
quence, there is still a lack of international Building erning structural design of FRC elements made of
Codes for structural design of FRC elements and this regular concrete that were mainly introduced by fib
may explain the limited utilisation of FRC among TG 8.3: the main concepts were derived from some
practitioners, even though a number of design national guidelines for FRC structural design [11, 13]
guidelines were recently developed. In fact, clear and from the guidelines proposed by RILEM TC162-
and simple design rules from Building Codes are TDF [7, 8]. The principles discussed herein are
strongly required by designers, who hardly accept to mainly related to Steel Fibre Reinforced Concrete
voluntary assume responsibilities by adopting guide- (SFRC) having a softening post-cracking behaviour
lines or, even worse, research results available in in uniaxial tension (Fig. 1a), even though they can be
scientific papers. extended to hardening materials (Fig. 1b). Since
Early design considerations were produced by ACI hardening behaviour is mainly obtained with a very
544 [5] and even in ACI 318 [6] some new rules were high performance concrete matrix and high fibre
just introduced with reference to minimum shear contents, fib TG 8.3 is cooperating with TG 8.6 for

Fig. 1 Typical Load (P) (a) (b)


deformation (d) curve for P P P P
FRC: post-cracking
softening (a) and hardening
(b) behaviour P P


Materials and Structures (2009) 42:12611281 1263

the design rules concerning Ultra High Performance different rules may be adopted in FRC structures
Fibre Reinforced Concrete. (smaller concrete covers for instance).
Since fibre reinforcement mechanisms are mainly
activated after cracking of the concrete matrix, fibres
2 FRC classification have marginal influence in the behaviour of
uncracked elements. Therefore, concrete tensile
Classification is an important requirement for struc- strength is related to the matrix strength and is not
tural materials. When referring to regular concrete, influenced by fibres.
designers choose its strength, workability or exposi- The residual (post-cracking) tensile strength,
tion classes that have to be provided by concrete which represents an important design parameter for
producers. When using FRC, compressive strength is FRC structures, is the mechanical property more
not particularly influenced by the presence of fibres influenced by fibre reinforcement.
(up to a volume fraction of 1%); therefore, the Due to the well known difficulties in performing
classification for plain concrete can be also adopted uniaxial tensile tests [16], bending tests with small
for FRC. It is well known that fibres reduce notched beams are best candidates to be a standard
workability of fresh concrete, but workability classes test method for the FRC classification. Since bending
for plain concrete can also be adopted for FRC. Some behaviour is markedly different from uniaxial-tension
studies are still needed for exposition classes since behaviour, it may happen that softening materials in
fibres may reduce the crack width [14]; therefore, for tension present a hardening behaviour in bending [17]
the exposition classes described in the EN 206 [15], (Fig. 2). In fact, in bending tests, the linear variability

pc

cc
STRESS

HPFRCC typical
(Strain-hardening and
multiple cracking)

FRC typical;
Matrix strain-softening

HPFRCC cc STRAIN pc
(or elongation)
Tensile Strain Hardening:
V f (V fcri ) tension

DFRCC

FRC Deflection Hardening:


COMPOSITES (V fcri ) bending V f (V fcri ) tension

MOR
fr
Tensile Strain Softening:
V f < (V fcri ) tension
f cc
(Deflection-hardening
LOAD

and multiple cracking)

Deflection Softening:
V f < (V fcri ) bending
Matrix Deflection-softening

c u
DEFLECTION

Fig. 2 Softening and hardening concepts for the classification of FRC [17]
1264 Materials and Structures (2009) 42:12611281

of the strain profile along the uncracked cross section Actually, European standard EN 14651 (2005)
favours a more stable propagation of the cracks, thus requires four different values of the residual strength
often inducing a significant hardening in bending even (fR1, fR2, fR3, fR4; Fig. 3), corresponding to different
if, in uniaxial tension, the material shows a softening values of the Crack Mouth Opening Displacement
behaviour after cracking of the concrete matrix. (CMOD = 0.5, 1.5, 2.5 and 3.5 mm, respectively) of
Material classification for FRC is based on nom- the notched specimen. For structural design the use of
inal properties that characterize its post-cracking four different values may be an obstacle for the
tensile strength. These properties are determined acceptance of the FRC design formulation. Therefore,
from bending tests according to EN 14651 [18] but it was assumed that fR1 and fR3 can characterize the
other type of tests (with beam or plate specimens) can FRC residual strength for SLS and ULS analysis,
be accepted, if correlation factors with the parameters respectively.
of EN 14651 can be demonstrated. To further simplify the classification, material
Post-cracking strength in hardening or softening behaviour at ULS can be related to the behaviour at
materials varies with the increasing of deformation or SLS using the fR3/fR1 ratio. Therefore, FRC toughness
crack opening. Since the residual strength varies with can be classified by using a couple of parameters: the
the imposed deflection (or crack mouth opening first one is a number representing the fR1 class, while
displacement) in the specimen, at least two deforma- the second one is a letter representing the fR3/fR1 ratio
tion values should be considered: the first one should (Fig. 4).
be significant for SLS verifications, while the second The strength classes for fR1k (characteristic value
one should be significant for ULS verifications. of fR1) are defined by the following values:

Fig. 3 Typical curve of the N


nominal stress versus fL
[MPa]
CMOD for FRC

f R1
f R2

f R3
f R4

CMOD [mm]

CMOD1 = 0.5 CMOD2 = 1.5 CMOD3 = 2.5 CMOD4 = 3.5

Fig. 4 Typical example of 5 N


FRC classification fR1k =2.2 MPa
[MPa] fR3k /fR1k = 0.82 2b
fR3k =1.8 MPa
4

3 d
c
fR1k
2 b
fR3k
a
CMOD 1

CMOD 2

CMOD 3

CMOD 4

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
CMOD [mm]
Materials and Structures (2009) 42:12611281 1265

1:0; 1:5; 2:0; 2:5; 3:0; 4:0; 5:0; 6:0; 7:0; 8:0 MPa) and isotropic reinforcement contribution, which is an
useful assumption for design purposes, safety factors
The fR3k/fR1k ratio can be represented with letters need to be used that take into account relevant aspects
a, b, c, d, corresponding to the values: like cast conditions, structure size and geometry. To
show how large the scattering can be in the real
a if 0:5  fR3k =fR1k \0:8
applications, five case-studies of FRC are analyzed in
b if 0:8  fR3k =fR1k \1:1 the following.
c if 1:1  fR3k =fR1k \1:4 The first case-study, represented in Fig. 5, is a top
d if 1:4  fR3k =fR1k slab made of Self Compacting SFRC (Vf = 0.34%;
hooked-ends, low-carbon, fibre length lf = 60 mm,
This classification properly represents the most fibre diameter, df = 0.9 mm). This slab is cast over
common cases of softening FRCs, but it can also be hollow core elements [19], simply supported on
adopted for hardening FRCs. foundation beams resting on piles (Fig. 5a, b, i) to
By using the proposed classification, a material build an industrial pavement designed to support
having, for example, fR1k = 2.2 MPa and significant dynamic loads (distributed loads of
fR3k = 1.8 MPa is classified as 2b (Fig. 4). 40 kN/m2, Fig. 5a). The SFRC was pumped and the
Since brittleness must be avoided in structural tests adopted to control the whole production process
behaviour, fibre reinforcement can be used as a showed an acceptable scattering on the cubic com-
substitution (even partially) of conventional rein- pressive strength, Rc, (variation coefficient &8%,
forcement (at ULS), if the following relationships are Fig. 5c), a reasonable dispersion on the peak flexural-
fulfilled: tensile strength, flf, (var. coeff. &14%, Fig. 5e), but a
fR1k =fLk [ 0:4 1 very high dispersion for both residual parameters
(26.8% for feq(00.6), Fig. 5f, and 34.6% for feq(0.63),
fR3k =fR1k [ 0:5 2
Fig. 5g), being feq(00.6) (similar to fR1) and feq(0.63)
where fLk is the characteristic value of the nominal (similar to fR3) determined according to UNI 11039 [20].
strength, corresponding to the peak load (or the A second case study [21] refers to a simply
highest load value in the interval 00.05 mm), supported SFRC beam characterized by a cross
determined from the beam test required by EN section of 300 9 300 mm (Fig. 6a). The beam is
14651 (Fig. 3). prefabricated and the fibre content is 50 kg/m3
(Vf = 0.64%). No longitudinal reinforcement is
introduced in this beam, since it serves as a reference
3 Partial safety factors beam for other post-tensioned beams. Low-carbon
hooked-ends steel fibres were used (lf = 60 mm;
As already mentioned, FRC often presents a post- df = 0.8 mm). The more controlled casting proce-
cracking softening behaviour. In design context, the dure allowed the producer to guarantee a smaller
main difference between FRC and traditional RC variation in the material properties characterized
elements is due to the presence of a distributed according to UNI 11039 [20]. The variation was 11%
reinforcement (fibres) that provides a residual post and 22% for serviceability (feq(00.6)) and ultimate
cracking strength to FRC (Fig. 2). (feq(0.63)) residual strength parameters, respectively
In the uncracked phase, FRC can be assumed as a (Table 1). These variations are similar to other
homogenous and isotropic material. However, after investigations related to prefabricated productions.
cracking, available research shows that the FRC It was verified that the increase of fibre content
residual strength depends significantly on the number contributes for a more homogeneous dispersion.
of fibres crossing the active cracks and on their FRC is a cementitious composite to be considered
orientation [21, 44]. Since fibre distribution and as a homogeneous material; therefore, the specimens
orientation are affected by several factors of difficult used to characterize its behaviour should have a
control (flowability, viscosity, filling ability), the volume that can be representative of the FRC real
scatter of the residual strength parameters can be application, since volume dimension affects the
significant. To assume that fibres have a homogeneous heterogeneity grade, which is a measure of the
1266 Materials and Structures (2009) 42:12611281

Fig. 5 Industrial pavement


[19]: a industrial pavement
before FRC topping cast;
b design sketch;
c dispersion and average
curve for nominal bending
strength versus CTOD;
d test set-up for notched
specimens according to
UNI 11039 [20];
e dispersion of compressive
strength (Rc); f dispersion
for first cracking nominal
strength; g residual strength
at SLSfeq(00.6); h residual
strength at ULSfeq(0.63);
i schematic representation
of the structure that forces
FRC post-cracking residual
behaviour towards average
instead of characteristic
behaviour

topological variability. This concept is easily under- A change of casting and handling procedure as
standable by evaluating the number of fibres in well as a change of the mixer imposed by the larger
certain regions, as shown in Fig. 6d, by adopting a cast volume can drastically improve mechanical
representative volume equal to 100 mm side cube, characteristics in bending, as it is highlighted in the
and considering a single cross section at right angle third example. The thin curve in Fig. 6b represents
with the longitudinal direction. Although the number the average response computed on the basis of
of fibres can significantly change in the different seventeen tests carried out during the production of
regions, by enlarging the size of the representative the retaining structures shown in Figs. 7 and 8 [22,
volume the variation coefficient related to the number 23] (Table 2), consisting of post-tensioned SFRC
of fibres in the fractured area can be reduced. framed panels used to stabilize a critical slope. The
Materials and Structures (2009) 42:12611281 1267

Fig. 5 continued

mix design to produce the beams of the previous advantages offered by this material, that it does not
case-study was also used to build the framed panels. require a minimum cover to protect fibres which can
However, when compared to the results obtained in be assimilated to smeared reinforcement.
previous case-study, the tests carried out for quality The fibres used were straight steel fibres with a
control purposes in this third case-study have pro- diameter of 0.16 mm and a length of 13 mm: the fibre
vided higher peak strength values and smaller content was 100 kg/m3. The cementitious composite
variation coefficient, as Fig. 6b (thin line) and 7 was characterized by an average cubic compressive
evidence. To improve fibre reinforcement effective- strength of 116 MPa. In these thin elements conven-
ness, it is also possible to use thin slabs with the aim tional reinforcement is not an effective reinforcement
of increasing the favourable fibre orientation effect, (apart the use of pre-stress wires) and is too
passing from a 3D to a 2D random distribution. The susceptible to corrosion phenomena. Therefore, the
thin slab used in precast roofing systems, shown in use of fibre reinforcement is quite appropriate but, in
Fig. 9, aims to benefit from this favourable effect general, FRC should present a hardening response in
[24]. In this fourth case-study, a nominal thickness of bending. The representativeness of standard beam
only 25 mm highlights one of the most important specimens, with a cross section of 150 9 150 mm,
1268 Materials and Structures (2009) 42:12611281

Fig. 6 SFRC beams [21]: (a) 300 1250 1000 1250


a geometry and set-up
adopted for beam tests;
b average curve of the
nominal stress versus Crack

300

300
Tip Opening Displacement
(CTOD) and dispersion of
the prismatic notched beam
250 3000 250
tests made of the same mix
and tested according to UNI 3500
11039 [20] (Fig. 5d); c load
versus displacement of the (b) (c)
full size beams; d fibre
number computed in nine
regions 100 9 100 mm of
the cracked cross section for
the three tested beams

(d)

Table 1 Bending strengths


Test fIf (MPa) fpeak (MPa) feq(00.6) (MPa) feq(0.63) (MPa)
of the specimens tested
according to UNI 11039 S0A 3.899 6.18 5.559 2.879
[20] and cast during beam
production (Fig. 6; [21]): S0B 4.350 6.80 6.063 3.890
first cracking strength (fIf), S0C 4.736 6.64 5.600 4.014
peak strength (fpeak), S1A 4.334 7.02 6.216 4.545
average strengths computed
S1B 4.231 4.98 4.623 2.496
in 00.6 mm (feq(0-0.6)) and
0.63 (feq(0.63)) crack S1C 4.195 6.28 5.602 3.661
opening ranges S2A 4.575 6.53 5.825 3.672
S2B 5.003 5.57 4.920 2.409
S2C 4.131 5.27 4.650 2.809
Average 4.384 6.142 5.451 3.375
Variation coeff. (%) 7.69 11.61 10.79 22.19
Materials and Structures (2009) 42:12611281 1269

response. In fact, the notch favours a stable crack


propagation and significantly modifies the first-
cracking process, especially if the fibre content is
not high enough to change the mechanical behaviour
at the peak; moreover, often fibres cumulate in the
bottom of the specimen and a notch could subtract a
significant bending resource to the specimen. For all
these reasons, an unnotched prismatic specimen, cast
with the same procedure used for the structure and of
the same thickness, is preferred. The specimen is
tested by means of a four point bending set-up that
favours a crack propagation starting from the weakest
cross section between the load points. The transducer
applied between two points of the bottom surface of
Fig. 7 Retaining structures tests: average curve of nominal
the specimen measures the relative displacement
stress rN versus CTOD and dispersion of the characterization
tests on the notched beam [22, 23] between these points (COD); the real concrete
deformability is obtained by taking into account the
distance between the axis of the transducer and the
bottom surface of the specimen, according to a plane
section kinematic approach. This type of thin spec-
imen can be better representative of the behaviour of
thin FRC structures; therefore it can be named
structural specimen as first suggested in the Italian
guidelines on Steel Fibre Reinforced Concrete [25].
The example proposed in Fig. 9 can be used to
better understand the behaviour of thin elements. This
example is composed of three thin plates 25 mm thick,
1.2 m wide and 2.5 m long that were tested in bending
according to a four point bending set-up (Fig. 9d, e).
The material was first characterized by means of
standard specimens (Fig. 9a). Afterwards, 14 pris-
matic specimens were extracted from a fourth plate,
sawn at right angle with respect to the longitudinal axis
of the plate. The specimens were 600 mm long and
150 mm wide. Therefore, the longitudinal stress
induced by bending were also at right angles with
respect to the longitudinal stresses induced by bending
Fig. 8 Retaining structure: a general view of the slope; b, c
in the large plates. Anyway, it should be noticed that
the single prefabricated panel respectively in the final location
and during handling [23] casting procedures were not repeatable and, therefore,
the formworks were randomly filled (usually starting
from the middle region to the borders). Other speci-
for the characterization of the behaviour of this type mens sawn and tested parallel to the plate longitudinal
of elements is, however, too questionable [18]. At the axis showed similar results. The comparison between
same time, fibre dispersion and orientation are the average curves rNCTOD obtained from the
strongly affected by casting procedure that, in thin standard notched specimen and curves rNCOD
elements, is different from that used in standard obtained from unnotched structural specimen high-
specimens. Furthermore, standard specimens are lights a significant difference in terms of peak strength
notched and it is difficult to guarantee the hardening and toughness; the large variation of the residual
behaviour of the material (in statistical terms) by their strength at ULS, feq2, (41%; Table 3) in the range
1270 Materials and Structures (2009) 42:12611281

Table 2 Bending post-


Test j (gg) fIf (MPa) feq(00.6) (MPa) feq(0.63) (MPa)
cracking strengths obtained
according to UNI 11039 NB1t2 38 5.37 4.55 2.64
[20] and cast during
retaining structure NB2t2 38 5.94 7.67 4.53
production (Fig. 8ac): age NB2t1 35 5.16 6.23 3.03
first cracking strength NB2t2 35 5.13 8.39 4.99
(fIf), peak strength (fpeak),
NB3t1 49 5.03 9.30 6.00
average strengths computed
in 00.6 mm (feq(00.6)) and NB3t2 32 5.16 8.50 5.69
0.63 (feq(0.63)) crack NB3t3 32 5.87 5.91 3.74
opening ranges NB4t1 46 6.51 7.93 5.56
NB4t2 34 5.99 8.88 5.07
NB4t3 34 6.38 8.70 4.97
NB5t1 58 5.81 8.93 4.96
NB5t2 44 6.06 7.82 4.37
NB6t1 69 7.22 8.83 3.80
NB6t2 42 6.00 7.86 5.01
NB7t1 40 6.06 8.45 3.31
NB7t2 83 5.96 7.81 4.98
Average 5.94 8.39 4.87
Variation coeff. (%) 10 6.3 15.1

2.43.6 mm, mainly reflects the random fibre orienta- in the bottom of the bucket could cause a reduction of
tion [26] and a rough casting procedure: it was the first fibre content in the last plate: in fact, the bucket was
attempt of slab casting and no specialized features progressively emptied by its bottom.
were considered to reduce the possible scattering. The last observation concerns the standard devi-
Moreover, while the CTOD in a notched specimen ation related to the global response of the structure. In
only represents the width of a single crack induced in a the first case-study, the hollow core structure forces
notched cross section, the displacement recorded by the top slab to have a similar bottom tensile strain and
the transducer applied to the bottom surface of the this implies that the different regions characterized by
structural specimen (COD) is the result of the crack a size comparable to that of the representative volume
propagation process of several cracks formed in the are forced to work as parallel springs (Fig. 5i; [22,
gage length of the transducer (200 mm; Fig. 9b). 27]). This implies that the global response was close
Furthermore, due to the diffuse crack pattern that to the average mechanical behaviour of the standard
generally occurs in this type of FRC structural specimens rather than to the characteristic one.
specimens, beside the relative displacement of the Furthermore, the standard deviation of the global
crack edges, the COD includes also the strain of the response is smaller than that exhibited by standard
uncracked concrete between cracks. specimens due to stress redistribution. Both the small
Finally, the bending behaviour of the thin plates, beam (Fig. 6b, c) and the full-scale plate (Fig. 9c, f)
tested according to a four point bending set-up responses confirm this evidence. The third full size
(Fig. 9d, e), is shown in Fig. 9f. The global response thin plate (Fig. 9f) is the only exception, even if this
of two plates exhibits an average peak strength of case was probably affected by a reduced fibre content
12 MPa and a corresponding strain of 0.005, close to in the plate. A similar trend was reported by Cominoli
the average response shown by the unnotched [28] in the tests he carried out on ground slabs
specimens. 3 9 3 m and 150 mm thick, made of concrete C30/
Only the third plate exhibited a reduced ductility, 37 and reinforced with a volume fraction of 0.38% of
probably caused by the casting procedure adopted. In hooked-end steel fibres having a length of 50 mm and
fact, this plate was the last to be cast by means of the a diameter of 0.75 mm. The slabs were supported on
same bucket. A possible accumulation of steel fibres steel springs with a spacing of about 400 mm to
Materials and Structures (2009) 42:12611281 1271

(a) 16 (d)
N UNI test average
[MPa]

12

0
0 1 2 3 4 5
CTODm [mm]

long top 1
(b) (long top 2) (e)

long bottom 1
vert B1 (long bottom 2) vert C1
(vert B2) (vert C2)

*measures in mm
vert B (1)
(c) 25 (f) 8
N vert D (1)
[MPa] vert B (2)
20 vert D (2)
6 vert B (3)
vert D (3)
Load [kN]

15

4
10

5 2
20C
20C average
0
0 1 2 3 0
COD [mm] 0 40 80 120
Average Deflection[mm]

Fig. 9 High Performance FRC thin plates [24]: a notched bottom transducers 1, 2 in Fig. 9b); d, e geometry and set-up of
beam tests according to UNI 11039 [20]; b geometry of the the tests on thin full scale plates; f load versus average
unnotched structural specimens; c nominal stress rN versus deflection measured in the middle section and close to the
average relative displacement at the bottom fibre (longitudinal loading points; g, h, i photographs during and after testing

reproduce a Winkler soil. The results shown in mechanism, and the ability of the structure to provide
Fig. 10a and c confirm that the dispersion of the a stress redistribution, demonstrated by the significant
mechanical response of six notched beam tests difference between maximum load and cracking load
(carried out according to UNI 11039 [20]), is much (Pcr, Fig. 10c).
larger than the scatter obtained in the three full size On the basis of the previous considerations, the choice
SFRC slabs on elastic supports. of the safety coefficients should take into account:
Table 4 includes the values of the scatter of the
the representativeness of the specimens used to
residual strength parameters (feq(00.6) and feq(0.63))
characterize the mechanical response of the
obtained in the notched beam tests and the scatter of
material, in relation to the structure considered;
the maximum load of the slabs. The smaller scatter of
the number of specimens used for mechanical
the maximum load is mainly related to two main
characterization;
variables: the fractured volume involved in the failure
1272 Materials and Structures (2009) 42:12611281

Fig. 9 continued

Table 3 Panel thickness and main experimental results from four point bending tests on structural specimen [24]
No. Test Thickness Max. load fIf fIf.av feq1 feq1.av (MPa) feq2 feq2.av(MPa)
(h) (mm) (kN) (MPa) (MPa) (std) (MPa) (std) (MPa) (std)

1 B1 27 5.90 12.94 11.05 (1.51) 18.71 14.09 (2.41) 23.78 12.14 (4.98)
2 B2 26 4.56 12.42 16.67
3 B3 25 4.34 11.85 16.67 19.31
4 B4 24 2.39 9.70 11.52 8.47
5 B5 25 3.10 9.80 13.52 12.73
6 B7 25 3.24 10.93 13.84 12.79
7 B8 25 3.03 12.40 13.83 9.48
8 B9 25 3.70 10.80 15.59 14.03
9 B10 26 3.17 11.38 12.30 10.56
10 B11 23 3.11 12.43 16.14
11 B12 25 2.67 11.10 11.59 8.82
12 B13 26 2.54 7.40 10.51 6.97
13 B14 25 3.10 11.94 14.51 11.02
14 B15 26 3.04 9.66 11.88 7.66
feq1 is the average nominal strength in COD range 0.30.5 mm; feq2 is the average nominal strength in COD range 2.43.6 mm

the stress redistribution capacity of the structure required by EN14651 [18] can be regarded as a reliable
under consideration; size. For all the other cases, like thin walled structures or
the fractured volume involved in the failure ultra high performance materials, where the casting flow
mechanism. can significantly affect the mechanical response of the
material and make the standard specimen unreliable for
In the new Model Code, these aspects are simplified
the mechanical characterization (due to fibre orientation
by considering a unified partial safety factor for cracked
and dispersion), different specimens are required. In
FRC in tension (cF) equal to 1.5 for massive structures,
order to obtain a more reliable material characterization,
where the representative volume of the standard beam
Materials and Structures (2009) 42:12611281 1273

(a) (b)
5
N
[MPa] 4

average
1

0
0 1 2 3
CTODm [mm]
(c) 350
(d)
P
[kN] 300
250

200

150

100 Slab P20 50/0,75 Vf=0,38 %


Pcr
Slab P21 50/0,75 Vf=0,38 %
50
Slab P22 50/0,75 Vf=0,38 %
0
0 1 2 3 4 5 6
[mm]

Fig. 10 Slabs supported by distributed steel springs: a, b results from bending tests on standard specimens according to UNI 11039
[20]; c load versus deflection curve of the three slabs; d general view of the slab during handling procedure [28]

Table 4 Experimental results from bending tests on standard fFtsd:red fFtsd =K fFtud:red fFtud =K 3a; b
specimens and from tests on full-scale slabs on grade [28]
where fFtsd and fFtud are the design values of the
Specimen Thickness Max. load Scatter (%)
(mm) (kN)
residual post cracking strength at service and at
ultimate conditions, respectively (see Sect. 4).
Slabs 150 301.4 2.30 Coefficient K plays the same role of g factor in the
291.4 equation proposed by ISO 2394/98 and first proposed
288.2 for SFRC structures by UNI 11188 [25]. In these
Beams 150 feq(00.6) feq(0.63) Standards, by considering a log-normal model of
18.0 23.1 random distribution of strength, and adding the
UNI 11039 parameters: feq(00.6),av = 2.83 MPa, feq(0.63),av = uncertainties referring to the geometry and to the
3.00 MPa, feq(00.6),k = 1.95 MPa, feq(0.63),k = 1.81 MPa design model of the cross section, respectively, the
partial safety factor is evaluated as:

these should be thinner specimens, notched or unnot- eabdr kdF


cF 4
ched, able to better represent the effective structural g
behaviour by simulating the real casting conditions [29, with
30]. In these special cases when standard specimens are q
used a coefficient K is introduced, affecting both the dr d2F d2g d2m 5
local and the global mechanical response. A value of 1.1
for global behaviour and 1.5 for local behaviour are where dF s=fm is the variation coefficient of SFRC
proposed at the serviceability and ultimate limit states, tensile strength (ratio between standard deviation and
according to the equations: mean value); dg ; dm are the variation coefficients of
1274 Materials and Structures (2009) 42:12611281

geometry and model adopted (both assumed for local check and, as a consequence, g is reduced to
dg = dm = 0.05 [25]); b fm  fmin =s is the reli- 0.8. b-coefficient can be set equal to 3 because its
ability index (assumed b = 3.8 [25]); a is the value, coupled to the high variation coefficient
sensitivity value of the material on the section selected (equal to 0.25), involves a minimum value
resistance (assumed a = 0.8 [25]); k = (fm - fk)/s always larger than 0.25 fm. In the case studies
is the coefficient representative of the characteristic investigated, no minimum strength fmin lower than
value at 5% fractile (assumed 1.645 [25]); g is the 0.25 fm was measured.
conversion factor assumed g = 0.9 ([25]) that takes When a local check has to be carried out, the main
into account the representativeness of the mechanical difference can be related to the a coefficient, because
parameter used to identify the mechanical character- no other mechanism can be considered and, therefore,
istics of the composite in the casting. the traditional value 0.8, conventionally assumed for
For structural purpose, in UNI 11188 the limitation RC structures, is recommended, while the coefficient
dF B 0.30 is assumed and the actual value of dF is b can be increased from 3 to 4 to consider even the
suggested to be deduced from qualification tests. In lack of fibres!
any case, dF should not be lower than 0.10. For thin walled structures or ultra high perfor-
Following the same procedure, a different choice mance materials, a coefficient g = 0.8 for global
for the values of the coefficients can be proposed. In check can be introduced, and 0.75 for local check,
fact, by adopting a log-normal model of random because the representativeness of the standard spec-
distribution, it is possible to assume a smaller value for imen has to be regarded as significantly less reliable.
k coefficient depending on the variation coefficient. Moreover, to consider that, in some cases, no
The variation coefficient has to be assumed equal for conventional reinforcement is used, a coefficient can
both SLS and ULS; it generally ranges between 0.10 be assumed equal to the traditional value 0.8, while
(SLS) and 0.35 (ULS). A value equal to 0.25 can be the same values chosen for massive structures can be
assumed as a reasonable value and, in a log normal introduced with reference to b. If a local check has to
distribution, is associated to a k value close to 1.4. be performed, a coefficient b = 4 (that means plain
In massive FRC structures, coefficient a can be set concrete, because dF is assumed 0.25!) is suggested
equal to 0.7 instead of 0.8, because FRC is generally and, therefore, the associated safety coefficient cF**
coupled with conventional reinforcement or it is used becomes equal to 1.5 cF. If a structural specimen
in structures where large volumes are involved in the able to reproduce the casting procedure as in the
failure mechanism and the structures are character- structure is used, the coefficient g can be set equal to 1
ized by a significant redundancy. The conversion (or prudentially 0.9), thus justifying reduced safety
factor g, that takes into account the representativeness coefficients. The safety coefficients for the various
of the nominal specimen according to EN14651 [18] cases discussed are specified in Table 5.
can be reduced to 0.85 in order to take into account In automatic casting procedure and for self com-
that, often, the reference volume can be relatively pacting FRC, the orientation factor could be much
small for large macro-fibres or the casting procedure better controlled [26, 30] and the material could be
could be not realistic with reference to fibre disper- treated as an orthotropic material with a better fibre
sion and orientation. This effect is more pronounced efficiency and a further reduction of the safety

Table 5 Parameters suggested and related safety coefficients for massive and thin-walled structures
Structure Check Test dF a b g k cF

Massive Global EN14651 0.25 0.7 3 0.85 1.4 1.43


Local EN14651 0.25 0.8 4 0.80 1.4 2.02
Thin-walled Global EN14651 0.25 0.8 3 0.80 1.4 1.64
Local EN14651 0.25 0.8 4 0.75 1.4 2.16
Global Structural 0.25 0.8 3 1 (0.9) 1.4 1.31 (1.46)
Local Structural 0.25 0.8 4 1 (0.9) 1.4 1.61 (1.79)
Materials and Structures (2009) 42:12611281 1275

coefficients since smaller variation coefficients dF can simpler assumptions suggested in relation to specific
be taken into account. cases [13].
An alternative characterization test refers to the
use of ASTM round plate resting on three supports as
reference nominal test [31, 32]. This test reduces the 4 Constitutive laws in uniaxial tension
scattering because the volume involved in the failure
mechanism is higher and the three radial cracks The FRC post-cracking tensile behaviour is modelled
propagate along three different directions, which in mode I crack propagation by a stress-crack opening
generally form at internal angles of about 120. In law. Two simplified stress-crack opening constitutive
this way, any fibre orientation is generally averaged laws may be deduced from the bending test results in
and the material can be better treated as a homoge- order to simulate a rigid-plastic behaviour or a linear
neous isotropic material. If a slab has to be designed, post-cracking behaviour (hardening or softening), as
this nominal test could be more representative and, schematically shown in Fig. 11. In this figure fFts
therefore, a value g = 0.9 could be assumed since the represents the serviceability residual strength, defined
failure mechanism involves the formation of yield as the post-cracking strength for a crack opening
lines oriented in several directions. value representative of the SLS analysis, and fFtu is
The last important question concerns with the lack the residual strength representative of the ULS
of knowledge of the heterogeneity grade in redundant analysis. In Fig. 11 wu is the crack opening corre-
structures able to redistribute stresses. For these sponding to the ULS considered, while the crack
structures, the stable propagation is controlled by opening associated to fFts is set equal to zero.
structural redundancy and the dispersion in the When considering softening materials, the definition
structure response is relatively small, if compared of a stressstrain law is based on the identification of a
with that of the material strength (see Figs. 6b, c and crack width and on the corresponding structural char-
10a, c). The homogeneity assumption for the material acteristic length, lcs, of the structural element. This basic
and the consequent characteristic post-cracking concept was first introduced by Bazant, with the
strength value, as determined from beam specimens, designation of crack band width [35, 36], then extended
often becomes too conservative because the stable and refined in the context of its use in kinematic models
crack propagation allows a multiple crack propagation [3739] and finally its genesis was related to the intrinsic
and the large fractured area implies a structure characteristics of the material [40, 41].
response mainly governed by the average values of Under the framework of the simulation of concrete
the material properties rather than by the character- fracture initiation and propagation by smeared crack
istic values [33, 34]. To take into account this material models, the lcs is used to assure results that are not
resource, a model coefficient cRd can be adopted: dependent on the finite element mesh refinement
( ) [4245]. In this context, lcs represents a bridge to
1   1 Xk;i
Rd  R Xd;i ; ad;i R ; ad;i 6 connect continuous mechanics governed by stress
cR d cRd cm;i strain (re) constitutive relationships, as it is common
in smeared crack models, and fracture mechanics
where R{} is a specific function related to the
considered mechanical model (i.e. bending, shear,
etc.); cRd is a partial factor covering uncertainties of
the assumed model; Xd,i is the design value of the
material property; ad,i is the nominal value of the
geometrical parameters involved in the model by
considering the tolerances; cm is the partial factor for
the material property.
The coefficient cRd can be smaller than unity, when
considering the favourable effect of material hetero- Fig. 11 Simplified constitutive laws: stresscrack opening
geneity, and can be computed on the basis of suitable (continuous and dashed lines refer to softening and hardening
finite element computations [34] or according to materials, respectively)
1276 Materials and Structures (2009) 42:12611281

governed by stresscrack opening (rw), as it is yh 10


common in discrete crack models, firstly proposed by
Hillerborg et al. [46] in concrete mechanics can be assumed due to the very reduced extension of
framework. the compressed region. The same assumption can be
The introduction of the lcs allows the designer to taken for slabs.
define the strain as: Tensile strain hardening materials are characterized
by the occurrence of diffuse crack patterns [47].
e w=cs 7 Therefore, the identification of crack openings is not
In elements with conventional reinforcement (re- necessary since a conventional stressstrain law may
bars), the characteristic length, lcs, may be evaluated be directly adopted, by dividing the relative displace-
as: ment by the gauge length of the displacement trans-
ducer, if a bending test is carried out. Any situation
cs minfsrm ; yg 8 where the bending response is partially hardening and
being: srm the mean value of the crack spacing; y the partially softening has to be split in a pre-peak
distance between the neutral axis and the tensile the continuous response and in a post-peak softening
most stretched fibres of the tensile zone of the cross response. A softening response after the peak does not
section, evaluated in the cracked phase by assuming imply any multi-localization and, therefore, it has to be
neglecting the peak-tensile strength of uncracked considered as a localized behaviour, thus requiring the
FRC, and for a load configuration corresponding to characteristic length approach.
SLS.
The ultimate tensile strength fFtu in the linear 4.1 Rigid-plastic model
model depends on the required ductility that is related
to the allowed crack width. The ultimate crack width The rigid-plastic model identifies a unique reference
can be calculated as: value, fFtu, based on the ultimate behaviour. The
rigid-plastic model takes the static equivalence into
wu lcs eFu 9 account (Fig. 12a), and fFtu results from the assump-
by assuming eFu equal to 2% for a neutral axis tion that the whole compressive force is concentrated
crossing the cross section and 1% for a neutral axis in the top fibre of the section. The value of fFtu is
external to the cross section. computed as:
In sections without traditional reinforcement under fR3
bending, under combined tensileflexural or under fFTu 11
3
combined compressiveflexural forces, with result-
Equation 11 is obtained by rotational equilibrium at
ing force external to the section, the simplified
ULS, by assuming wu = CMOD3 (Fig. 13a), and a
expression

(a) (b) = E x (c) f Fts


C
x


y M M

= w i1 /l cs f Fts 0.5 f R3 - 0.2 f R1

Fig. 12 Simplified models adopted to compute: a the ultimate b the tensile strength (fFts) by means of a elasto-plastic model;
tensile strength in uniaxial tension (fFtu) by means of a rigid- c fFtu for the linear model by means of a rigid-linear model
plastic model and the residual nominal bending strength (fR3);
Materials and Structures (2009) 42:12611281 1277

Fig. 13 a Typical results


from a bending test on a
softening material and
b linear post-cracking
constitutive law

constant stress block along the section, as shown in for a crack opening displacement corresponding to the
Fig. 12a. The residual strength parameter fR3 of Eq. 11 SLS:
is obtained from rotational equilibrium, by assuming a
linear elastic stress distribution along the notched cross 3FCMOD1  L
fR1 15
section, as in the pre-peak condition, for a crack 2b h2sp
opening displacement corresponding to the ULS:
being FCMOD1 the force at CMOD1 (Fig. 13a).
3FCMOD3  L The strength value fFts corresponding to the crack
fR3 12
2b h2sp opening CMOD1 is determined from equilibrium,
with the assumption that the compressive stress
being F(CMOD3) the force at CMOD3 (Fig. 13a), L, distribution is linear (Fig. 12b) and that the tensile
b and hsp the span length, the width and the depth of behaviour is elasto-plastic until a crack opening
the net notched cross section of the beam specimen. displacement corresponding to the SLS.
The variability introduced in the results by elastic
4.2 Linear post-cracking model modulus is regarded as negligible and a common
value is assumed.
The linear post-cracking model is characterized by The stress value corresponding to crack opening
the fFts and fFtu parameters. Both parameters have to CMOD3 is determined from equilibrium, with the
be defined through residual values of flexural strength assumption that the compressive stress resultant is
by using the following equations [4850]: applied on the extrados chord (Fig. 12c) and that the
fFTs 0:45fR1 13 tensile behaviour is rigid-linear.
wu Once the linear stress-crack opening relationship
fFTu fFTs  fFTs  0:5fR3 0:2fR1  0 is identified, by introducing the characteristic length
CMOD3
it is possible to deduce the design relationship
14
stressstrain, as shown in Fig. 14 with reference to
Equation 14 defines fFtu as the residual strength softening [48] and hardening materials. It should be
associated to an ultimate crack opening wu that does underlined that, in order to be consistent with
not necessarily coincide with CMOD3, and it is Eurocodes assumptions, the tensile strength of plain
obtained by considering a linear constitutive law concrete should be neglected for softening materi-
between points with abscissa CMOD1 and CMOD3, als. On the contrary, when a F.E. approach is
up to the point with abscissa wu (Fig. 13). adopted, the first abrupt post-peak branch, corre-
The residual strength parameter fR1 of Eqs. 13 and sponding to unstable fracture propagation and
14 is obtained from rotational equilibrium, by assum- connected to localization of a single crack, could
ing a linear elastic stress distribution along the be useful for reproducing strain localization in the
notched cross section, as in the pre-peak condition, structure [34].
1278 Materials and Structures (2009) 42:12611281

Fig. 14 Stressstrain
relations for softening (a)
and hardening (b) FRC

5 Basic aspects for design fibres provide an additional reinforcement, ductility is


generally provided by conventional reinforcement
Fibre reinforcement is especially suitable for stati- that plays a major contribution to the tensile strength.
cally indeterminate structures of high stress distribu- For hardening FRCs (in uniaxial tension), fibres can
tion capacity. In structures where both localized and be used as the only reinforcement (without rebars)
diffuse stresses are present, which is the usual case, a also in statically determined structural elements. In
reinforcement system composed of rebars and fibres structures made of linear elements without traditional
may represent an optimized reinforcement solution. reinforcement, ductility requires that FRC has a
In structural elements, where fibres aim to hardening behaviour in tension.
substitute (even partially) conventional reinforce- In any case, the maximum load of the FRC
ment, some restrictions on the minimum residual structure (Pu) should be bP1, being P1 the load at
strength are applied (Eqs. 1, 2). The residual strength cracking initiation and b a coefficient higher than 1.
becomes significant in structures characterized by a In addition, a minimum structural ductility (ratio
high degree of redundancy, where a remarkable stress between the ultimate displacement, du, and the
redistribution occurs. For this reason, in structures displacement at cracking initiation, dc), should be
without rebars, where fibres completely substitute guaranteed.
conventional reinforcement, a minimum level of As an example, when referring to bending, ULS is
redundancy of the structural member is required. reached when one of the following conditions is
On the contrary, in structures with rebars, where obtained (Fig. 15):

cu f cd f cd
x x

f Fts / F NSd
y
Asl M Rd
su

Fu hardening softening f Ftu / F


Fig. 15 ULS for bending moment and axial force: use of the simplified stressstrain relationship (k e g coefficient in accordance
with current codes)
Materials and Structures (2009) 42:12611281 1279

attainment of the maximum compressive strain in the crack width and increase the tension stiffening of
the FRC, ecu; concrete between cracks. Crack control is particularly
attainment of the maximum tensile strain in the important for durability issues. In structures with
steel (if present), esu; rebars, a reduced minimum concrete cover is
attainment of the maximum tensile strain in the expected when fibre reinforcement is present.
FRC, eFu.
For a design axial force (NSd ), the ultimate bending
6 Concluding remarks
moment (MRd ) can be evaluated by means of the
translation and rotation equilibrium equations under
A very important milestone within the research of
the assumption of a plane section.
FRC is the implementation of this material in the
The fundamental hypotheses for the ULS analysis
coming fib Model Code that, in the next future, will
of FRC sections are:
lead to a probable development of structural rules in
sections remain plane; FRC elements in Eurocodes or in national codes. In
perfect bond conditions are assumed between fact, new structural materials are easily used by
steel bars and surrounding concrete; designers only when design rules are present in
the tensile stresses in FRC are derived from the building codes.
design stress/strain relationship (Fig. 14); The main advantages from FRC come from the
for the compressive stresses, the equations used presence of a reinforcement (fibres) distributed in the
for plain concrete can be also applied for FRC. structural element that enhances the resistance to
diffused tensile stresses and guarantees a reinforce-
The evaluation of the ultimate bending moment
ment even in the large concrete covers, often
can be done with reference to the strain and stress
necessary for fire resistance or durability issues, or
distributions shown in Fig. 15 that correspond to the
in the corners of the structural elements.
FRC stress/strain relationships and the constitutive
Fibre reinforcement goes into the structural ele-
laws for rebars (if present) in accordance with the
ment together with the concrete, thus avoiding costs
current codes.
for handling and placing the reinforcement. Further-
Fibres in a beam represent an additional distrib-
more, fibres require a smaller storage area (normally
uted reinforcement that enhances shear resisting
a tank) with respect to conventional reinforcement
mechanisms both in elements with and without
(rebars or welded mesh).
transverse reinforcement.
Besides the resistance to ULSs, fibre reinforce-
The use of the minimum amount of conventional
ment can significantly enhance structural behaviour
shear reinforcement (stirrups) can be prevented if the
at service conditions, by limiting cracking pheno-
following condition is fulfilled [51]:
mena or structural deformation.
p In any case, although the level of knowledge on
fck
fFtuk  16 FRC tremendously increased during the last 10 years,
20
further research is needed to verify and optimize the
being fck the characteristic value of the concrete proposed design rules, to investigate the long term
compressive strength and fFtuk the characteristic value behaviour of different FRCs and other open issues.
of fFtu. This condition allows to reduce the crack width Furthermore, a new generation of FRCs will soon
growth and favours the diffusion of inclined crack enter into the market. They are based on a cocktail
pattern and, as a consequence, can ensure a sufficient of different fibre types (for material and/or geometry)
member ductility. When a great amount of longitudi- to contemporary enhance different structural perfor-
nal reinforcement in the compressive zone is present, mances (Hybrid systems of fibres).
adequate stirrups reinforcement shall be applied in
order to avoid buckling of the compressed rebars. Acknowledgements A special acknowledgement goes to
Besides the structural resistance at ULS, a signif- Prof. Joost Walraven for the fruitful discussions and to Prof.
Falkner who shared its large design experience. The Authors
icant contribution can be provided by FRC at SLS are also grateful to all the members of fib Task Groups TG 8.3
conditions. In fact, fibre reinforcement may reduce and TG 8.6 for the constructive discussions during the several
1280 Materials and Structures (2009) 42:12611281

meetings where many ideas presented in this paper took a 18. EN 14651 (2004) Test method for metallic fibre concrete
definitive shape. measuring the flexural tensile strength (limit of propor-
tionality, residual). Varenna, Italy
19. di Prisco M, Colombo M, Dozio D, Mauri M (2006) SFRC
References ground slab: an experience on hollow core slabs. In: Pro-
ceedings of the 16th national conference (CTE 06),
1. Rossi P, Chanvillard G (eds) (2000) Fibre Reinforced Parma, 2006, pp. 965978 (in Italian)
Concretes. In: RILEM Proceedings of the 5th RILEM 20. UNI 11039 (2003) Concrete reinforced with steel fibres.
Symposium (BEFIB 2000), PRO15, BEFIB 2000. RILEM Part II: test method for the determination of first cracking
Publications S.A.R.L., Bagneux, France strength and ductility indexes. Italian Board for Standard-
2. di Prisco M, Felicetti R, Plizzari G (eds) (2004) Fibre- ization, Italy
Reinforced Concrete. In: RILEM Proceedings of the 6th 21. di Prisco M, Dozio D (2008) Post-tensioned SFRC beams
RILEM Symposium (BEFIB 2004), PRO39, BEFIB 2004. in Fibre Reinforced Concrete: design and applications. In:
RILEM Publications S.A.R.L., Bagneux, France di Prisco et al (eds) Proceedings of the 8th RILEM Sym-
3. Reinhardt HW, Naaman AE (eds) (2007) High Perfor- posium (BEFIB 08). RILEM Publications S.A.R.L., Bag-
mance Fibre Reinforced Cement Composites (HPFRCC5). neux, France, pp 899910
Rilem Publication S.A.R.L., PRO53 22. di Prisco C, di Prisco M, Mauri M, Scola M (2006) A new
4. Gettu R (ed) (2008) Fibre Reinforced Concrete: design and design for stabilizing ground slopes. In: Proceedings of the
applications, BEFIB 2008, Bagneux, France, RILEM 2nd fib congress, Napoli (Italy), June 58, 2006, ID 4-1 on
Publications S.A.R.L., PRO60 CD-ROM
5. ACI Committee 544 (1996) Design considerations for steel 23. di Prisco M, di Prisco C, Dozio D, Galli A, Lapolla S
Fiber Reinforced Concrete, ACI 544.4R-88. American (2008) Assessment and control of a SFRC retaining
Concrete Institute, ACI Farmington Hills structure: mechanical issues. In: Binda et al (eds) On site
6. ACI Committee 318 (2008) Building code and commen- assessment of concrete, masonry and timber structures.
tary, Report ACI 318-08/318R-08. American Concrete Proceedings of the international conference (SACOMA-
Institute, Farmington Hills TIS). RILEM Publications S.A.R.L., Bagneux, France,
7. Vandewalle L et al (2002) Recommendation of Rilem PRO 59, Varenna, vol 1, pp 539604
TC162-TDF: test and design methods for steel fibre rein- 24. di Prisco M, Lamperti M, Lapolla S, Khurana RS (2008)
forced concrete: design of steel fibre reinforced concrete HPFRCC thin plates for precast roofing. In: Proceedings of
using the rw method: principles and applications. Mater second international symposium on ultra high performance
Struct 35:262278 concrete, Kassel, Germany, 2008, pp 675682
8. Vandewalle L et al (2003) Recommendation of Rilem 25. UNI 11188 (2004) Design, production and control of Steel
TC162-TDF: test and design methods for steel fibre rein- Fibre Reinforced structural elements. Italian Board of
forced concrete: re-design method (final recommenda- Standardization, Italy
tion). Mater Struct 36:560567 26. Ferrara L, Dozio D, di Prisco M (2007) On the connections
9. AFGC-SETRA (2002) Ultra High Performance Fibre- between fresh state behavior, fiber dispersion and tough-
Reinforced Concretes, Interim recommendations. AFGC ness properties of steel fiber reinforced concrete. In:
Publication, France Reinhardt HW, Naaman AE (eds) High Performance Fiber
10. Stalfiberbetong (1995) rekommendationer for konstruction, Reinforced Cement Composites (HPFRCC5). Proceedings
utforande och provning Betongrapport n.4. Svenska Bet- of the 5th international RILEM Workshop, Mainz, Ger-
ongforeningen, Betongrapport many, PRO 53. RILEM Publication SARL, pp 249258
11. Deutscher Ausschuss fur Stahlbeton (DAfStb) (2007) 27. di Prisco M, Colombo M (2006) FRC and HPFRC com-
Guidelines for steel fiber reinforced concrete23th posites: from constitutive behaviour to structural applica-
Draftrichtlinie StahlfaserbetonDIN 1045 Annex parts tions, in Measuring, monitoring, and modelling concrete
14, August 2007 properties. In: Konsta-Gdoutos MS (ed) Proceedings of the
12. Faserbeton R (2002) Osterreichische Vereinigung fur international symposium dedicated to Prof. S. P. Shah,
Beton-und Bautechnik. OBBV, Wien Alexandroupolis (Greece). Springer, Dordrecht, pp 5968
13. CNR-DT 204 (2006) Guidelines for design, construction 28. Cominoli L (2007) Studio sul calcestruzzo fibrorinforzato
and production control of fiber reinforced concrete struc- per applicazioni industriali: dalle proprieta del materiale al
tures. National Research Council of Italy, Italy comportamento strutturale, Ph.D. Thesis, University of
14. Vandewalle L (2000) Cracking behaviour of concrete Brescia (ISBN 978-88-96225-10-3), 431 pp
beams reinforced with a combination of ordinary rein- 29. Stahli P, van Mier JGM (2007) Manufacturing, fibre
forcement and steel fibers. Mater Struct 33(227):164170 anisotropy and fracture of hybrid fibre concrete. Eng Fract
15. UNI EN 206-1 (2006) Concretepart 1: specification, Mech 74:223242
performance, production and conformity. CEN, Brussels 30. Stahli P, Custer R, van Mier JGM (2008) On flow prop-
16. Hordijk DA (1991) Local approach to fatigue of concrete. erties, fibre distribution, fibre orientation and flexural
Dissertation, Delft University of Technology behaviour of FRC. Mater Struct 41(1):189196
17. Naaman AE, Reinhardt HW (2003) High performance fiber 31. Lambrechts AN (2004) The variation of steel fibre rein-
reinforced cement compositesHPFRCC4. In: RILEM forced concrete characteristics. Study on toughness results
Proceedings, PRO 30, RILEM Publications S.A.R.L., 20022003. In: di Prisco et al (eds) Fiber Reinforced
Bagneaux, France Concrete: from theory to practice. International workshop
Materials and Structures (2009) 42:12611281 1281

on advances in Fiber Reinforced Concrete. Starrylink Ed- 43. Kooiman AG (2000) Modelling steel fibre reinforced
itrice, Brescia, Italy, pp 135148 concrete for structural design. PhD thesis, Technical Uni-
32. Minelli F, Plizzari G (2007) Fiber reinforced concrete char- versity Delft, Optima Grafische Communicatie, Rotterdam
acterization: round panel vs. beam test toward a harmoniza- 44. Barros JAO, Cunha VMCF, Ribeiro AF, Antunes JAB
tion. In: Proceedings of 3rd central European congress on (2005) Post-cracking behaviour of steel fibre reinforced
Concrete Engineering, Visegrad, Hungary. September 1718. concrete. Mater Struct 38(275):4756
Publishing Company of Budapest University of Technology 45. di Prisco M, Iorio F, Plizzari G (2003) HPSFRC pre-
Hungary, Budapest, 2007, pp 213220 stressed roof elements. In: Schnutgen B, Vandewalle L
33. di Prisco M, Failla C, Plizzari GA, Toniolo G (2004) (eds) Test ad design methods for steel fibre reinforced
Italian guidelines on SFRC. In: di Prisco M et al (eds) concretebackground ad experiences. PRO 31, RILEM,
Fiber-Reinforced Concrete: from theory to practice. Star- Bochum, Germany, pp 161188
rylink, Bergamo, pp 3972 46. Hillerborg A, Modeer M, Peterson PE (1976) Analysis of
34. Dozio D (2008) SFRC structures: identification of the crack formation and crack growth by means of fracture
uniaxial tension characteristic constitutive law. Department mechanics and finite elements. Cement Concr Res 6:
of Structural Engineering, Politecnico di Milano, Milano 773782
35. Bazant ZP, Oh BH (1983) Crack band theory for fracture 47. JSCE (2008) Recommendations for design and construc-
of concrete. Mater Struct 16(93):155177 tion of high performance fiber reinforced cement com-
36. Bazant ZP, Cedolin L (1983) Finite element modeling of posites with multiple fine cracks (HPFRCC). In: Rokugo K
crack band propagation. ASCE J Struct Eng 109(1):6992 (ed) Concrete engineering series, p 82
37. Bazant ZP, Pijaudier-Cabot G (1988) Non-local continuum 48. di Prisco M, Felicetti R, Lamperti M, Menotti G (2004) On
damage, localization instability and convergence. ASME J size effect in tension of SFRC thin plates. In: Li VC et al
Appl Mech 55:287293 (eds) Fracture mechanics of concrete structures. BL Sch-
38. Bazant ZP, Novak D (2003) Stochastic models for defor- mick and AD Pollington, USA, pp 10751082
mation and failure of quasibrittle structures: recent 49. di Prisco M, Ferrara L, Colombo M, Mauri M (2004) On
advances and new directions. In Bicanic N et al (eds) the identification of SFRC costitutive law in uniaxial ten-
Computational modelling of concrete structures. Swets & sion. In: di Prisco et al (eds) Fiber Reinforced Concrete.
Zeitlinger, Lisse, pp 583598 Proceedings of 6th RILEM symposium (BEFIB 04), PRO
39. de Borst R, Mulhaus HB, Pamin J, Sluys LJ (1992) 39. RILEM Publications S.A.R.L., Bagneaux, France,
Computational modeling of localization of deformation. In: pp 827836
Owen DRJ et al (eds) Proceedings on computational 50. Colombo M (2006) FRC bending behaviour: a damage
plasticity fundamentals and applications, part II. Pineridge model for high temperatures. PhD thesis, Department of
Press, Swansea, pp 483508 Structural Engineering, Politecnico di Milano, StarryLink,
40. Fokwa D, Berthaud Y (1993) Heterogeneous materials: Brescia, Italy
experimental analysis of localization and the influence of 51. Minelli F, Plizzari G, Vecchio FJ (2007) Influence of steel
size of the heterogeneities on the behaviour in tension. fibers on full-scale RC beams under shear loading. In: High
Mater Struct 26:136143 performance concrete, brick-masonry and environmental
41. di Prisco M, Felicetti R, Gambarova PG (1999) On the aspects. Sixth international conference of fracture
evaluation of the characteristic length in high strength mechanics of concrete and concrete structures FRAM-
concrete. In: Azizinamini A, Darwin D, French C (eds) COS6. Taylor & Francis Group (UK), London/Catania, 3,
High strength concrete. ASCE, Kona, Hawaii, pp 377390 pp 15231531
42. Ferrara L, di Prisco M (2001) Mode I fracture behavior in
concrete: non-local damage modeling. ASCE J Eng Mech
127(7):678692

You might also like