Virtual and Physical Prototyping

ISSN: 1745-2759 (Print) 1745-2767 (Online) Journal homepage: https://www.tandfonline.com/loi/nvpp20

Ductility of 3D printed concrete reinforced with

short straight steel fibers

F. P. Bos, E. Bosco & T. A. M. Salet

To cite this article: F. P. Bos, E. Bosco & T. A. M. Salet (2019) Ductility of 3D printed concrete
reinforced with short straight steel fibers, Virtual and Physical Prototyping, 14:2, 160-174, DOI:
10.1080/17452759.2018.1548069

To link to this article: https://doi.org/10.1080/17452759.2018.1548069

© 2018 The Author(s). Published by Informa

UK Limited, trading as Taylor & Francis
Group

Published online: 19 Nov 2018.

Submit your article to this journal

Article views: 4036

View related articles

View Crossmark data

Citing articles: 34 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=nvpp20
VIRTUAL AND PHYSICAL PROTOTYPING
2019, VOL. 14, NO. 2, 160–174
https://doi.org/10.1080/17452759.2018.1548069

Ductility of 3D printed concrete reinforced with short straight steel fibers

a a
F. P. Bos , E. Bosco and T. A. M. Saleta,b
a
Department of the Built Environment, Eindhoven University of Technology, Eindhoven, the Netherlands; bWitteveen+Bos consulting
engineers, Deventer, the Netherlands

ABSTRACT ARTICLE HISTORY

With the number of 3D printed concrete structures rapidly increasing, the demand for concepts that Received 16 October 2018
allow for robust and ductile printed objects becomes increasingly pressing. An obvious solution Accepted 11 November 2018
strategy is the inclusion of fibers in the printed material. In this study, the effect of adding short
KEYWORDS
straight steel fibers on the failure behaviour of Weber 3D 115-1 print mortar has been studied 3D printing; concrete; CMOD;
through several CMOD tests on cast and printed concrete, on different scales. The experiments steel fiber; numerical
have also been simulated numerically. The research has shown that the fibers cause an modelling
important increase in flexural strength, and eliminate the strength difference between cast and
printed concrete that exists without fibers. The post-peak behaviour, nevertheless, has to be
characterised as strongly strain-softening. In the printed specimens, a strong fiber orientation in
the direction of the filament occurs. However, this has no notable effect on the performance in
the tested direction: cast and printed concrete with fibers behave similarly in the CMOD test. For
the key parameters, no scale effect was found for the specimens with fibers, contrary to the ones
without. Numerical modelling of the test by using the Concrete Damage Plasticity material model of
Abaqus, with a Thorenfeldt-based constitutive law in compression and a customised constitutive
law in tension, results in a reasonable fit with the experimental results.

1. Introduction
With digitally fabricated concrete (DFC) quickly outgrow-
ing the laboratory phase, as showcased by recent
examples such as a pedestrian bridge in Spain (3ders
2017a), a bicycle bridge in the Netherlands (Salet et al.
2018), an office hotel in Denmark (3dprinthuset 2017),
and a workshop in Dubai (3ders 2017b), the quest for
tensile capacity and ductility has become a major issue
(Salet et al. 2017; Asprone et al. 2018) because conven-
tional reinforcement techniques are either incompatible
with DFC, or largely cancel out its potential benefits.
Several innovative concepts have been presented that
provide ductility and tensile capacity to printed concrete
in an automated fashion, including robotised placement
of mesh-reinforcement and automatically entrained
cable reinforcement (Hack and Lauer 2014; Hack et al.
2015; Bos et al. 2017). External passive reinforcement
and 3D printed steel reinforcement have also been
suggested 2020 (Asprone et al. 2018; Mechtcherine
et al. 2018). These concepts require further development
to be able to fully appreciate their possibilities.
An obvious strategy to address this issue is to include
fibers in the print material. In conventional concrete
construction glass, steel, polymer, and other fibers have
been used for decades for several purposes, for instance
to reduce shrinkage cracking, but also to reduce or
obviate the need for conventional reinforcement bars.
In such a structural capacity, steel fibers are commonly
used. Steel fiber reinforced concrete (SFRC) is particularly
beneficial in applications in which conventional
reinforcement is difficult to apply, for instance when a
high reinforcement ratio is required, in underwater con-
crete, or in slender and geometrically complex elements
or in areas of high load concentrations (e.g. due to loca-
lised connections). These latter two examples may typi-
cally occur in DFC.
Given these considerations, the current study explores
the structural performance of SFRC applied in a particular
DFC process, namely the ‘3D Concrete Printing (3DCP)’
Fused Deposition Modelling (FDM) process that is
being developed at the Eindhoven University of Technol-
ogy (TU/e) and is similar to facilities at several other
research institutes and commercial enterprises (Bos
et al. 2016; Khoshnevis and Dutton 1998; Khoshnevis
2004; Lim et al. 2012; Wu, Wang, and Wang 2016). First,
the features of the printing with steel fibers in the

CONTACT F.P. Bos f.p.bos@tue.nl Department of the Built Environment, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, the
Netherlands
© 2018 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/
4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in
any way.

particular facility are discussed. Then, the theoretical fra-
mework is introduced. The experimental work, consisting
of Crack Mouth Opening Displacement (CMOD) tests on
differently sized cast and printed specimens, is sub-
sequently presented. As part of this study, a continuous
2D numerical model was developed that has been
used to recreate the experiments. The modelling issues
are discussed, particularly with regard to the material
model, as well as the correspondence between numeri-
cal and experimental results.
2. Fibers in 3D printed concrete
2.1. The use of structural fibers in concrete
printing
The 3DCP facility is schematically shown in Figure 1. Steel
fibers could not be applied in the dry mix, as the set-up
makes use of a rotor-stator pump with rubber lining and
narrow cavities that is unable to pass such fibers without
blocking and damaging the pump. Therefore, an
additional device was developed, capable of applying
fibers near to the mixed material near the print nozzle
(number 6 in Figure 1). In the future, it should be possible
to adjust the fiber content by location in the print object
to obtain optimised material use, as suggested (for non-
printed applications) by Plückelmann, Song, and Breiten-
bücher (2017).
In the current study, short 6 mm straight fibers
(Bekaert Dramix OL 6/.16) have been used as they were
expected not to significantly impact the workability (Mar-
kovic 2006), and to be able to pass the nozzle opening
without clogging as might occur with larger types of
fibers (various nozzles are used, typically the opening is
10 × 40–60 mm2).
The addition of fibers to achieve ductility in 3D con-
crete printing is an obvious solution strategy that has
been explored on a very small scale by Hambach and
Volkmer (2017), who added 3–6 mm basalt, glass and
carbon fibres to a printable mixture, and Panda,
Chandra Paul, and Jen Tan (2017), who compared glass
fibers of different lengths (3, 6 and 8 mm) and varying
VOL% of fibres. Both studies reported a significant
Figure 1. Schematic representation of TU/e 3DCP set-up. Steel fibers can not be pumped, and should therefore be added to the
filament near the print head (6).
VIRTUAL AND PHYSICAL PROTOTYPING 161
increase in flexural tensile strength, as well as an orien-
tation effect of the fibers in the direction of the
filament flow, but neither discussed the effects on duct-
ility. The influence of several fibers on the flexural per-
formance and interface strength has been studied by
Nematollahi et al. (2018), while Weng et al. (2018)
tested the flexural and compressive strength of a single
mixture with poly-vinyl alcohol (PVA) fibers, and Zareiyan
and Khoshnevis (2018), among other things, studied the
influence of polypropylene (PP) fibers on the interface
adhesion. Furthermore, printable engineered cementi-
tious composites (ECCs) and strain-hardening cementi-
tious composites (SHCCs) and are under development,
as shown by Soltan and Li (2018) and Ogura, Nerella,
and Mechtcherine (2018). These groups used PVA and
high-density polyethylene (HDPE) fibers, respectively,
and obtained strain-hardening performance.
2.2. Structural performance of concrete reinforced
with straight steel fibers
As shown e.g. by Markovic (2006), it should be expected
that normal strength concrete with short straight steel
fibers will show increased tensile strength due to
delayed macro-crack development. Furthermore, a
strong strain-softening post-crack behaviour should be
anticipated, as the fiber-matrix adhesion is insufficient
to generate a pull-out resistance in the order of magni-
tude of the fiber strength, i.e. the fiber efficiency is
«100%. The increased tensile strength, fracture tough-
ness, and failure strain may nevertheless be beneficial
for many practical applications, and improve structural
performance in statically indeterminate structures.
The performance of fiber reinforced concrete (FRC)
has been reported to depend on fiber properties (geo-
metry, material, aspect ratio), matrix (cementitious
material, filler and aggregates, water/cement ratio), as
well as the matrix-fiber combination (fiber distribution,
fiber orientation) (Markovic 2006; Naaman 2003). With
regard to fiber orientation, it should be noted that,
since the fiber resistance is governed by pull-out rather
than yielding as for conventional reinforcement bars, a
162 F. P. BOS ET AL.
full parallel alignment of the fiber in the direction of the
principle tensile stress is not necessarily beneficial.
Results found in literature are not entirely consistent
on this issue. Several studies found an increase of resist-
ance up to 20% for inclination angles of up to 60° (Van
Gysel 2000; Robins, Austin, and Jones 2002), whereas
another work found an important decrease of resistance
leading up to 50% at 60° (Bartos and Duris 1994). Yet,
others noted that the post-fracture resistance and
energy absorption significantly reduces for inclination
angles over 30° (Banthia and Trottier 1994; Armelin and
Banthia 1997; Svec et al. 2012), supported by a study on
the effect of the magnetic orientation of fibers (Wijffels
et al. 2017). When corrected for a fixed limited value of
fiber slip, several studies indicate an optimal pull-out
resistance is reached at inclination angles of 10–15° (Mar-
kovic 2006). It is conceivable the effect on inclination
angle is not independent of fiber geometry, as the resist-
ance of deformed fibers is governed by the plastic defor-
mation required for pull-out. Judging from the variety of
results found in literature, it is hard to predict whether
the notable orientation of fibers found by both
Hambach and Volkmer (2017) and Panda, Chandra Paul,
and Jen Tan (2017) in 3DCP due to the concrete flow,
will be beneficial to the post-fracture performance or not.
To assess the structural performance of FRC, RILEM
(Rilem Technical Committees 2002) has proposed the
use of the Crack Mouth Opening Displacement (CMOD)
test. This test has been codified in the EN 14651 (Neder-
lands Normalisatie Instituut 2007) and was adopted by
the fib Model Code 2010 (International Federation for
Structural Concrete 2013) as the standard method to
characterise FRC for structural calculations. In this stan-
dardised 3-point bending test, the increase of the
notch width in a beam, known as the crack mouth
opening, is measured during load application and used
to obtain an overall tensile strength value as well as an
indication of residual resistance after the Limit of Propor-
tionality (LOP) has been reached. The residual resistance
fR,j is determined at prescribed levels of CMODj. Here,
j = 1, 2, 3, 4 which corresponds to 0.5, 1.5, 2.5, and
3.5 mm of crack mouth opening respectively (e.g.
CMOD3 = 2.5 mm). The residual resistances fR,1 and fR,3
are particularly important, as the fib Model Code 2010
uses these values to classify the performance of a particu-
lar FRC composition. Both the LOP and residual strengths
are calculated through linear elastic theory, as in:
3FL l 3Fj l
f
fct,L = and fR,j = (1)
2bh2sp 2bh2sp
with FL force at the limit of proportionality, and Fj the force
at subsequent predescribed CMOD levels 0.5, 1.5, 2.5 and
3.5 mm. The geometrical features of the specimen (the
height hsp , the width and the length l) are defined in
Figure 3.
As the behaviour after fracture is generally highly non-
linear, and often strain-softening, particularly the calcu-
lation of residual strength should be considered as an
apparent value.
The presence of the notch eliminates edge effects and
ensures the development of a single localised crack that
can be measured unambiguously. However, it influences
the flexural resistance in comparison to unnotched speci-
mens, which obviously, are more similar to actual appli-
cations. Contrary to concrete without fibers where the
notch induces a peak stress that leads to earlier failure,
the notch has actually been reported to have a beneficial
effect in FRC, particularly with non-straight fibers (Stähli
and Van Mier 2004). Other studies have shown less pro-
nounced results for FRC with straight fibers (Chanvillard
and Rigaud 2003). Even though these issues should be
heeded when translating experimental results to an
application environment, the notched test has been
used in this study for its practical merit of a predictable
crack location and broad acceptance.
2.3. Scale effect
In this study, experimental tests have been performed on
cast FRC specimens of the size specified by the EN 14651,
as well as on cast and printed FRC specimens scaled
down with a factor of 0.27, to match the scale of
beams specified in the EN 196-1 (Nederlands Normalisa-
tie Instituut 2016) for flexural tests on cements and
mortars.
Plain concrete exhibits a scale effect for tensile and
flexural strength. Smaller specimens tend to show a
higher resistance as the strain energy release rate,
which governs crack growth in brittle materials, is less
in such specimens (Hillerborg, Modéer, and Petersson
1976; Bažant 1984; Bažant, Ožbolt, and Eligehausen
1994). Yoo et al. (2016) have noted the scale effect is
more pronounced in high strength concrete (HSC) than
in normal strength concrete (NSC). This should be attrib-
uted to the increased brittleness. They furthermore con-
cluded that an increase in size led to a decrease in
flexural performance, but an increase in fiber volume
content reduced the sensitivity to this effect in SFRC
and amorphous metallic FRC. In concretes of higher duct-
ility, the size effect seems to disappear completely.
Lepech and Li (2003) found no size effect in a range of
tested Engineered Cementitious Composite beams
without and with conventional reinforcement (ECC and
R/ECC). Mahmuda, Yang, and Hassan (2013) concluded
the size effect in ultra-high performance fiber reinforced

concrete (UHPFRC) beams is small. Considering that the
scale effect is related to the strain energy release on
the one hand, and the fracture surface energy on the
other, this is in line with expectations.
A study by Giaccio, Tobes, and Zerbino (2008) was par-
ticularly aimed at the scale effect in the CMOD test, and
intended to determine whether smaller specimens could
be used to characterise FRC behaviour, arguing that this
would simplify experimental testing and moreover corre-
sponds better to some practical uses. They applied a scale
factor of 0.7, not only to the EN 14651 specimen size, but
also to the CMOD values at which relevant stress values
are recorded (e.g. the strength fR1 was recorded at a
CMOD1 of 0.5 mm × 0.7 = 0.35 mm). When this correction
was applied, it was concluded that, as long as the fiber
length and aggregate size are compatible with the
mold, the proposed downscaling was allowable.
The EN 14651 allows testing of samples with aggre-
gates of up to 32 mm in size, and fiber lengths of up to
60 mm. The applied material Weber 3D 115-1 has a
maximum grain size of 1 mm, and the length of the
applied fibers is 6 mm. Therefore, the ratios of the
maximum to applied aggregate size and fiber length
are significantly smaller (0.03 and 0.1, respectively) than
the scale factor (0.27).
For several reasons, scaling down the specimen size in
this study, was considered desirable as well as justifiable.
Most importantly, the reduced specimen size results in
section dimensions that bear resemblance to the sec-
tions of elements printed in the TU/e 3DCP facility.
Testing reduced size specimens therefore corresponds
best with the actual intended use. Furthermore, it
results in a significantly improved research and material
efficiency. Regardless of these considerations, the scale
effect should be small anyway, as the concrete quality
is low and a significant VOL% of fibers is applied.
3. Experimental research
3.1. Experimental set-up and test series
Three CMOD tests have been performed: CMOD-1 on
cast, full scale specimens, CMOD-2 on cast, downsized
specimens, and CMOD-3 on printed, downsized speci-
mens. For each CMOD test, one set of specimens
without and one set with fibers has been tested. With
the exception of one, each set consisted of 3 specimens.
Figure 2 schematically shows the specimen types, and
Figure 3(a) details the experimental set-up. In Figure 3
(b), a printed specimen is shown during testing, in the
scaled down version of the test set-up.
CMOD-1 was performed in accordance with the EN
14651. However, given the typical size of a filament
VIRTUAL AND PHYSICAL PROTOTYPING 163
section (which roughly corresponds to the size of the
nozzle opening), the EN 14651 specimen dimensions
are unsuitable to evaluate the performance of printed
concrete from the 3DCP process. The process is
intended for more filigree structures, which should be
reflected in the dimensions of experimental specimens.
Therefore, the set-up and specimens of CMOD-2 and -3
were scaled down by a factor of 40/150 (≈ 0.27), so
that the specimen section size w × h measures 40 ×
40 mm2. The nominal specimen dimensions are given
in Table 1.
In all tests, the CMOD was measured with a clip gage
and the vertical displacement with (a) linear variable
differential transformer(s) (LVDT(s)). In CMOD-1, one
LVDT was mounted on an aluminium rod connected to
the specimens at mid-height above the supports, and
measured the vertical displacement of a strip glued
onto the specimen directly under the load point. In
CMOD-2 and -3, this set-up was impossible due to the
dimensions of the specimens. Alternatively, two LVDTs
were used to measure the vertical displacement of
glued on strips at each side of the crack mouth
opening, and averaged.
The tests were performed on an Instron universal test
rig, controlled by the crack mouth opening displacement,
which was set at 0.05 mm/min until CMOD = 0.1 mm was
reached, and then raised to 0.2 mm/min. The tests were
stopped at specimen failure (without fiber) or when a
CMOD ≥ 4 mm had been reached (with fiber).
3.2. Material
All specimens were prepared with Weber 3D 115-1 print
mortar as the matrix material. This mixture was custom
designed for the 3DCP process by Saint Gobain Weber
Beamix and is globally described by Bos et al. (2016).
Some material properties derived from uni-axial
tension tests (Slager 2017) are given in Table 2.
Bekaert DRAMIX ® OL 6/.16 high strength straight steel
fibers were added to the B-sets of specimens in a quan-
tity of 150 kg/m3 (corresponding to 2.1 VOL%). The basic
fiber properties are provided by Bekaert (2017), and
listed in Table 3. Table 4 provides an overview of speci-
men sets.
3.3. Specimen preparation
The material for CMOD-1 was mixed in a standard gravity
mixer, with a water/mixture ratio of 0.2. The form work
was initially filled for 90% (in the casting order
specified by the EN 14651) and compacted for 10 s on
a vibrating table, after which the remaining 10% was
added and the specimen was compacted again for
164 F. P. BOS ET AL.

Figure 2. Schematic comparison of specimen types: 1A – full scale, cast, no fibers; 1B – full scale, cast, fibers; 2A – reduced scale, cast, no
fibers; 2B – reduced scale, cast, fibers; 3A – reduced scale, printed, no fibers; 3B – reduced scale, printed, fibers.

10 s. The top was scraped flat. The specimens were
demoulded after 24 h and stored under water for the
rest of the curing period. The notch was applied with a
saw one day before testing. The specimens were tested
at an age of 28 days.
The specimen preparation of the CMOD-2 specimens
was similar. However, instead of a gravity mixer, a hand
mixer in a bucket was used due to small amount of
material required.
The printed specimens of CMOD-3 were prepared
with a beta version of the fiber addition device. Practi-
cally, this means the fibers were added to the mix that
was prepared as for CMOD-2. Subsequently, the wet
mix with the fibers was introduced in a screw-pump

Figure 3. CMOD test set-up: (a) general scheme based on the EN 14651, and (b) printed specimen being tested in scaled down version
of CMOD test.
 VIRTUAL AND PHYSICAL PROTOTYPING 165

Table 1. Nominal specimen dimensions for each CMOD test. to the main bending stresses in the CMOD test. This is
CMOD b = h [mm] hsp [mm] l sp [mm] hsp /l sp [-] x [mm]* shown by Figure 4(a–c), showing a spare piece of fiber
1 150 125 500 4 2 reinforced printed material, sawed perpendicular
2 40 33 130 4 2
3 40 33 130 4 2
(Figure 4(a)) and parallel (Figure 4(b,c)) to the print
* the notch width could not be scaled. direction.
The load-CMOD curves that directly result from the
tests have been recalculated into stress-CMOD curves
Table 2. Weber 3D 115–1 experimentally obtained material through Equation (1). Figures 5–7 show the stress-
properties (without fibers, uni-axial tension test). CMOD curves for CMOD-1, -2, and -3, respectively, with
Property Symbol Value Unit and without fibers.
Compressive strength, mean value fcm 23.2 N/mm2 The flexural tensile strength ft , LOP fL and residual
Compressive strength, 0.95 fck 20.7 N/mm2 strengths fR,1 and fR,3 according to the EN 14651, are
characteristic value
Ultimate strain 1cu3 0.0035 – listed in Table 5. For the scaled down CMOD-2 and
Tensile strength, mean value fctm 1.49 N/mm2 CMOD-3 tests, also the residual strengths at scaled
Specific Fracture energy GF 0.031 N/mm
E-modulus, tangent E0 17,336 N/mm2 CMOD levels are given (fR,1∗0.27 and fR,3∗0.27 ). These have
E-modulus, secant Ecm 25,717 N/mm2 also been indicated in the graphs. The size dependent
specific fracture energy Gf , calculated over the displace-
ment from the peak strength (fracture onset) onwards, is
Table 3. Bekaert Dramix OL 6/.16 steel fiber properties, also included in Table 5. Figure 8(a,b) summarize the key
according to Declaration of performance. properties in a comparative column chart.
Property Symbol Value Unit Finally, Table 6 lists the average post-cracking classifi-
Length l 6 mm cation of the fiber reinforced series according to the fib
Diameter d 0.15 mm
Aspect ratio l/d 40 – Model Code 2010. For the scaled down specimens, a
Tensile strength ft 3000 N/mm2 characterisation both based on fR,3 /fR,1 as well as on
E-modulus E 200,000 N/mm2
fR,3∗0.27 /fR,1∗0.27 is provided.

device that was connected to a hose and print nozzle

3.5. Discussion
that was manually moved over a deposition plate to
create layered specimens in a fashion similar to the auto- 3.5.1. Effect of fibers and printing
mated 3DCP process. The specimens were stored under The experimental results show that the addition of fibers
water after initial hardening and sawed to size before significantly increases the flexural strength (Figure 8(a)).
testing, as their dimensions exceeded the 40 × 40 mm2 The increase is most striking for the printed specimens,
dimensions. in which it is twice as high as in the cast, full-scale speci-
mens, reaching over 500%. This is mainly due to the low
flexural bending strength of the printed specimens
3.4. Experimental results without fibers, which is only about half the strength of
As reported by other researchers, a strong fiber orien- the cast specimens (see further discussion below). In all
tation in the direction of the print filament was visually cases, the fibers introduce a residual strength in the
observed in the printed specimens, compared to the specimens that is modest in comparison to other fiber
cast ones. The fibers are thus generally oriented parallel reinforced concrete compositions, but nevertheless a sig-
nificant improvement over concrete without fibers. The
obtained post-fracture behaviour is strongly strain-soft-
Table 4. Overview of specimen series.
ening, to an extent that a classification according to
Matrix
CMOD Set Qty. material Fibers Manufacturing the fib Model Code 2010 is impossible (fR,3 /fR,1 , 0, 5),
1 1A 3 Weber 3D – Cast if the CMOD for j = 1 and j = 3 is not adjusted for scale.
115-1 The increase in flexural and residual strength,
1B 3 Weber 3D 150 kg/m3 Dramix Cast
115-1 OL 6/.16 obviously, also translates in an importantly enlarged
2 2A 2 Weber 3D – Cast specific fracture energy (Figure 8(b)), by a factor of >30
115-1
2B 3 Weber 3D 150 kg/m3 Dramix Cast in the cast full-scale elements, up to a factor of >100
115-1 OL 6/.16 for the scaled-down specimens (printed and cast).
3 3A 3 Weber 3D – Printed
115-1
For specimens without fibers, a significant decrease in
3B 3 Weber 3D 150 kg/m3 Dramix Printed flexural strength was found for printed specimens in com-
115-1 OL 6/.16 parison to cast ones (36%, compare series 2A and 3A,
166 F. P. BOS ET AL.

Figure 4. Sawn open spare piece of fiber reinforced, printed concrete. Figure (a) was sawn open perpendicular to the print direction.
The fibers are slightly bended due to the sawing. In Figures (b and c), showing the parallel sections, the fibers appear as long lines,
indicating orientation of the fibers in the print direction.

Figure 5. Stress-CMOD curves for CMOD-1A and -1B (blue tones and red tones, respectively).

Figure 6. Stress-CMOD curves for CMOD-2A and -1B (blue tones and red tones, respectively).
 VIRTUAL AND PHYSICAL PROTOTYPING 167

Figure 7. Stress-CMOD curves for CMOD-3A and -3B (blue tones and red tones, respectively).

Figure 8(a) and Table 5). As the specimen orientation in
the tests was such that the results were not influenced
by the layered structure of the printed specimens, this
difference is expected to be caused by the lack of compac-
tion in the printing process. Remarkably though, the
flexural strength reduction vanishes for cast and printed
concrete with fibers (+6%, compare 2B and 3B). Likewise,
the specific fracture energy and residual strengths fR,1 and
fR,3 are within small margins. The fiber orientation caused
by the printing process does not seem to have a positive
or negative effect compared to randomly oriented fibers,
at least for the tested direction.
Thus, the addition of fibers to this printed concrete
shows significant benefits. It eliminates the detrimental
effect of a lack of compaction. Even though the material
behaviour is still strongly strain-softening, the increase in
flexural strength and fracture energy has advantages,
particularly for secondary structural applications (e.g.
façade elements) and connections (redistribution of
peak stresses).
3.5.2. Scale effect
Contrary to expectations, this study found the flexural
bending strength of the cast specimens without fibers
was reduced for the scaled-down series (26% reduction,
compare series 1A and 2A), rather than enlarged. This is
possibly caused by the geometry and accuracy of the
notch which had a larger influence on the smaller speci-
mens. For the cast specimens with fibers, a reduction was
also found, but it was much smaller (11%, compare 1B to
2B) and solely caused by the underperformance of a
single specimen – see Figure 6. The small number of
specimens in each series does not allow for any final con-
clusions on this issue.

Table 5. Experimental results.

CMOD test properties
CMOD Spec. f t [N/mm2] Gf [N/mm] f LOP f R,1 f R,1∗0.27 f R,3 f R,3∗0.27
1A (cast, full size, no fibers) 1A.1 2.42 0.025 2.42 0.06 * – *
1A.2 2.51 0.034 2.51 – * – *
1A.3 2.29 0.025 2.29 0.06 * – *
Average 2.41 0.028 2.41 n/a * n/a *
1B (cast, full size, fibers) 1B.1 6.65 1.16 4.84 4.84 * 1.40 *
1B.2 6.25 1.07 4.87 4.25 * 1.21 *
1B.3 5.95 1.03 4.59 4.40 * 1.27 *
Average 6.28 1.09 4.77 4.48 * 1.29 *
2A (cast, scaled, no fibers) 2A.1 1.67 0.015 1.67 – 0.28 – –
2A.2 1.91 0.013 1.91 – 0.21 – –
Average 1.79 0.014 1.79 n/a 0.25 n/a n/a
2B (cast, scaled, fibers) 2B.1 4.02 0.81 4.00 2.66 3.85 1.10 2.39
2B.2 6.32 1.31 5.92 4.61 6.22 1.95 4.15
2B.3 6.45 1.30 6.15 4.56 6.29 1.93 4.09
Average 5.60 1.14 5.36 3.94 5.45 1.66 3.54
3A (printed, scaled, no fibers) 3A.1 0.93 0.0049 0.93 – 0.05 – –
3A.2 1.12 0.0083 1.12 – 0.20 – –
3A.3 1.37 0.0100 1.37 – 0.20 – –-
Average 1.14 0.0077 1.14 n/a 0.15 n/a n/a
3B (printed, scaled, fibers) 3B.1 6.55 1.09 6.28 4.26 6.10 1.58 3.89
3B.2 6.46 0.86 5.50 3.94 6.32 1.68 3.41
3B.3 4.86 0.97 4.39 3.62 4.71 1.54 3.49
Average 5.96 0.97 5.39 3.94 5.71 1.60 3.60
*The scaled residual strength values do not apply to the full scale specimens.
168 F. P. BOS ET AL.

Figure 8. a Average results: flexural strength, residual strengths R,1 and R,3. b. Average specific fracture energy Gf .

The residual strength at j = 1 is quite similar for both rather than the peak strength. Note finally that edge
the full and reduced scale. At j = 3, the residual strength effects in the experimental tests influence the strain
in the reduced scale specimens is slightly larger. energy release that drives crack growth over a distance
The specific fracture energy Gf of the specimens known as the ligament length al , creating an additional
without fiber is highly scale dependent (compare 1A source of size-dependency (Abdalla and Karihaloo 2003).
and 2A). The obtained value from the full scale test is Although the residual strength ratio fR,3 /fR,1 is
comparable to the value obtained earlier from uni-axial higher for the smaller fiber reinforced specimens due
tensile testing (Table 2), but for the scaled down speci- to a higher value of fR,3 , this does not result in a
mens, it is only half. This can be explained from the different classification of post-fracture behaviour.
definition of the size dependent specific fracture However, this changes when scaled down values for
energy Gf in a CMOD test. This can be calculated as j = 3 (CMOD = 2.5 mm · 0.27 = 0.675 mm) and j = 1
the energy exerted by the external load divided by the (CMOD = 0.5 mm · 0.27 = 0.135 mm) are applied, as
fracture surface area perpendicular to the load direction suggested by Giaccio, Tobes, and Zerbino (2008). In
(Abdalla and Karihaloo 2003), i.e. that case, an ‘a’ classification is found for the scaled

down specimens.
1
Gf = Fd d (2)
hsp b
4. Numerical modelling
From Eq. (2) it follows that Gf is largely determined by the
Even though the printed SFRC obtained in this study
product of the flexural strength and the corresponding
shows strongly strain softening behaviour, the increased
displacement, as the post-fracture branch is only very
tensile strength and ultimate strain can have merit in
short. Based on standard mechanics equations
structural applications, for instance in the design of struc-
3Fl Fl
(s = ; d= ), it can be determined that tural joints. In order to assess the structural performance
2bh2 48Ebh3
of this material in arbitrary geometries, a suitable calcu-
an increase in scale with a factor 2, leads to a section
lation method has to be determined. Therefore, the
A = bh that is 4 times as high, but also results in an
experiments were simulated through Finite Element
increase of Fd with a factor of 8, i.e. a net increase with
Modelling (FEM) and evaluated on their agreement
a factor 8/4 = 2 – if only the peak would determine Gf .
with the experimental results.
The fiber reinforced specimens, on the other hand,
show hardly any scale effect with regard to specific frac-
ture energy (compare 1B and 2B), because its value is 4.1. Material model
mainly determined by the post-fracture behaviour,
The combination of isotropic damage and plasticity is a
frequently used modelling framework to describe the
Table 6. Post-cracking classification according to the fib Model
complex non-linear response of concrete. Several
Code 2010.
models in the literature have been developed along
Classification full scale Classification reduced scale
CMOD f R,3 /f R,1 letter f R,3∗0.27 /f R,1∗0.27 letter
this strategy, allowing to capture the failure response
1B 0.29 – * –
of the material both in tension and in compression –
2B 0.42 – 0.65 a see for instance (Lubliner et al. 1989; Peerlings et al.
3B 0.41 – 0.63 a 1998; Voyiadjis, Taqieddin, and Kattan 2008; Nguyen

and Korsunsky 2008; Grassl and Jirasek 2006; Poh and
Swaddiwudhipong 2009). In this paper, the Concrete
Damaged Plasticity (CDP) model proposed by Lubliner
et al. (1989) has been adopted. This is used to describe
both concrete with and without fibres, in the assumption
that, when fibres are present, they are uniformly distrib-
uted and the material can be thus considered to be hom-
ogenous. The CDP model is available as a material model
in the commercial software Abaqus, with the option of
implementing separate user-defined hardening or soft-
ening post-cracking curves for compression and
tension that come into effect after a linear elastic branch.
4.2. Constitutive relation in compression
A commonly used constitutive law for concrete in
compression is the relation proposed by Thorenfeldt,
Tomaszewicz, and Jensen (1987) given as
 
1c n
sc (1c ) = fcm (3)
1c1 n − (1 − (1c /1c1 )nk )
with: n = 0, 80 + fck /17 and

1, if 0 , 1c , 1c1
k= , where 1c1 is the
0, 67 + fck /62, if 1c ≥ 1c1
compressive strain at the peak stress.
It consists of two quadratic branches, before and after
cracking. As the CDP model includes a linear elastic pre-
cracking branch, rather than a quadratic one, the Thoren-
feldt law cannot be incorporated completely. Thus, in the
CDP model three branches were defined in compression,
as illustrated by Figure 9: (i) a linear elastic first branch
with a limit at the characteristic compressive strength,
(ii) a quadratic hardening branch up to the mean com-
pressive strength and corresponding to the first part of
the Thorenfeldt law, and (iii) a quadratic softening
branch after the mean compressive strength that
Figure 9. Thorenfeldt material model compared to the CDP
model in Abaqus, in compression.
VIRTUAL AND PHYSICAL PROTOTYPING 169
corresponds with the post-cracking part of the Thoren-
feldt law. The input values were based on the values
listed in Table 2, and have been used as the constitutive
relation in compression for all the analyses (with and
without fibers). Although the compressive properties
would be influenced by the presence of fibers, this is
not expected to significantly influence the response of
the FE-model as it is governed by tensile failure.
4.3. Constitutive relation in tension
Several constitutive laws have been proposed to model
the behaviour of concrete in tension. These essentially
require two input data: the (size-independent) specific
fracture energy GF and the shape of the stress–strain
softening branch (Rots et al. 1985). The specific fracture
energy GF can be defined as the integral of the stress-
separation relation


GF = s dw (4)
Here, s and w represent the traction and the crack
opening in the normal direction, respectively. The size-
independent specific fracture energy (4) can be obtained
from the size-dependent one (2), by performing a set of
experiments on test specimens of different scales or with
different notch sizes (Abdalla and Karihaloo 2003).
In continuum damage models, as the CDP used in this
analysis, the crack is smeared over a band of a certain
width wc . Within a finite element setting, this is therefore
related to the element size. In order to avoid dependency
of the results on the selected discretization, the fracture
energy must be released over this width, i.e. GF is con-
stant and does not depend on the finite element mesh
(Rots et al. 1985; Bažant and Cedolin 1979; Hordijk
1991). Under this premise, the assumption is made
here that the strain is constant along the crack band,
yielding
w = wc 1f (5)
This assumption has been proven to hold for lower order
finite elements and for particular cases of symmetry (Rots
et al. 1985), consistently with the numerical test pre-
sented in this paper. For more general loading conditions
or higher order finite element types, assumption (5) does
not generally hold. Non-local/gradient enhanced soften-
ing models (Peerlings et al. 1998; Poh and Swaddiwudhi-
pong 2009) are more appropriate in the latter situation.
As for the strain-softening curve, for plain concrete,
exponential functions are frequently used (Hordijk
1991; Cornelissen et al. 1986; Krätzig and Pölling 2004),
as they have showed good agreement with experimental
results. For two reasons, however, these cannot be used
170 F. P. BOS ET AL.
for fiber reinforced concrete. First, the specific fracture
energy is much higher than for plain concrete and there-
fore it would require to be established separately.
Second, contrary to plain concrete, there is no generally
accepted continuous post-fracture function for fiber
reinforced concrete (Awinda, Chen, and Barnett 2016),
because this would highly depend on the specifics
(fiber type, volume, etc.) of any FRC. Therefore, the
tensile post-fracture behaviour was modelled by insert-
ing it in the CDP model in a tabular format, with values
for stress and strain.
In this regard, the fictitious strain 1f in the crack was
determined according to assumption (5), taking the
crack opening w to be equal to the experimentally
acquired CMOD values, and the crack band width wc to
√
be equal to the characteristic mesh size, lc = Ae with
Ae = Atot /ne the average element area (total mesh area
divided by the number of elements).
The stress values for the tabular input have been
determined with linear elastic mechanics through
relations Eq. (1), as used by the fib Model Code 2010 to
obtain the (residual) strength levels for values of
CMOD. Although the stress at the specimen edge can
be calculated from linear elastic mechanics up to fracture
onset, the post-cracking stress will be distributed over
the height in a non-linear and a principally undetermin-
able fashion. Also, it should be noted that this approach
is the opposite of the spectrum compared to the deter-
mination of the specific fracture energy Gf through
Eq. (2), which assumes a constant distribution of fracture
energy over the section height. It will depend on the
actual crack mouth opening which approach resembles
the actual distribution state more closely. Depending
on the separation and the material behaviour, the
actual stress at the outer edge can easily be 50% or
less of the linear elastic calculated value. Nevertheless,
the fib Model Code 2010 provides guidelines to recalcu-
late these values into general constitutive laws that can
be used in structural modelling. These approaches are
elaborated by Di Prisco, Colombo, and Dozio (2013)
and depend on stress distributions, structural character-
istic length (lcs , essentially identical to the crack band
width wc ), and Young’s modulus – all of which have to
be assumed and are functions of the strain. Similar reser-
vations have been noted by Singh (2017).
Considering the above remarks, the numerical model-
ling of general structural geometries should ideally be
based on a constitutive law derived from direct uni-
axial tensile tests (as e.g. performed by Mahmuda,
Yang, and Hassan 2013), because this results in a less
ambiguous apparent stress-separation relation. Awinda,
Chen, and Barnett (2016) also came to this conclusion,
but in reference to Petersson (1981) note that tensile
tests are in general difficult to perform and go on to
argue that modelling based on material properties
obtained from flexural tests is allowable to simulate
those tests (in that case in a discrete cracking model).
This approach was adopted in this study. Note that
even uni-axial tensile tests do not result in pure material
properties: Rots and de Borst (1989) showed that asym-
metric crack propagation under uni-axial tension has
consequences for the derivation of stress–strain
relations.
4.4. Geometry and boundary conditions
The experiment was numerically modelled in 2D, using
the commercially available FE-code Abaqus. As exper-
imental research Giaccio, Tobes, and Zerbino (2008)
has shown the specimen thickness has no notable
effect on the flexural performance, the application of a
2D model was considered to be appropriate.
The CMOD-1, CMOD-2 and CMOD-3 experiments
were modelled, with their respective 2D geometries
(which are identical for CMOD-2 and -3). The following
boundary conditions apply. The displacement of the
left bottom corner is fixed in the vertical and horizontal
directions; the right bottom corner is vertically con-
strained. The loading was applied as a prescribed displa-
cement at the top edge above the notch, using the
relation between CMOD and vertical displacement
given by the EN 14651.
To mesh the geometries, bilinear plane stress
elements with reduced integration (CPS4R) were used.
The mesh properties are given in Table 7. A bias-factor
of 1 was applied to minimise the influence of non-con-
stant mesh element sizes.
4.5. Results and discussion
Figure 10(a–d) show the deformed configuration of the
CMOD-1B (full scale, cast, with fibers) test. The contour
plot refer to the sxx field. The crack localisation and
development is clearly visible, and follows a trend signifi-
cantly close to those of the experimental results.
Figures 11 and 12 illustrate the stress–strain curves
(continuous line) obtained from the numerical simu-
lations, for both the specimens without and with fibers.
The average experimental results are shown for
Table 7. Applied meshes and properties.
Geometry Bias factor l c [mm] ne [-] Ae,ave [mm2]
CMOD-1 1 1.78 24,528 3.056
CMOD-2 1 0.91 6,457 0.989
CMOD-3 1 0.91 6,457 0.989

Figure 10a–f. Deformed configurations predicted from the FEM model (scale factor 10). The contour variable shows the stress sxx
distribution in CMOD-1B.
comparison with dashed lines. In general, the numerical
results are fairly capturing those of the experimental tests.
In the numerical model, the peak stress is reached at a
slightly lower value of CMOD than in the experimental
test. This can be explained by the fact that the CDP
model assumes a linear-elastic trajectory up to fracture,
whereas in reality micro-cracks reduce the stiffness
before fracture onset. The flexural strength peak is
found accurately, which is to be expected as a linear
elastic stress distribution over the specimen height is
still valid at this point.
Figure 11. Stress-CMOD curves for CMOD-1A, -2A, and -3A (full
size, scaled down cast and scaled down printed, respectively,
without fibers), average experimental (dashed lines) and numeri-
cal (continuous lines).
VIRTUAL AND PHYSICAL PROTOTYPING 171
Beyond the peak, the numerical curves generally over-
estimate the trend of the experimental ones, especially in
the case of fibre-reinforced specimens – Figure 12. This
discrepancy grows with increasing values of CMOD. The
overestimation is to be attributed to the modelling
assumptions discussed earlier. For the fibre-reinforced
material, translating the experimental results in terms of
load-CMOD curves into stress–strain relations, by assum-
ing relation (5) to hold, may introduce a first source of mis-
match. Additionally, taking a linear elastic stress
distribution over the beam height overestimates the
Figure 12. Stress-CMOD curves for CMOD-1B, -2B, and -3B (full
size, scaled down cast and scaled down printed, respectively,
with fibers), average experimental (dashed lines) and numerical
(continuous lines).
172 F. P. BOS ET AL.
stress level in the ultimate fiber once the actual stress dis-
tribution moves towards a more plastic like behaviour.
Thus, this effect grows stronger for higher values of CMOD.
The presented numerical modelling method, thus,
results in a reasonably accurate simulation of the
CMOD test up to a moderate extent after fracture. To
improve the description in the post-peak response, it
would be necessary to base the constitutive model on
data obtained from uni-axial tensile tests and consist-
ently to use more tailored damage models. This will be
considered in future work.
5. Conclusions
The effect of adding short straight steel fibers on the
failure behaviour of Weber 3D 115-1 print mortar has
been studied through several CMOD tests on cast and
printed concrete, on different scales. The experiments
have also been simulated numerically. The research has
shown the fibers cause an important increase in flexural
strength, and eliminate the strength difference between
cast and printed concrete without fibers. The post-peak
behaviour, nevertheless, has to be characterised as
strongly strain-softening. In the printed specimens, a
strong fiber orientation in the direction of the filament
occurs. However, this has no notable effect on the per-
formance in the tested direction: cast and printed con-
crete with fibers behave similarly in the CMOD test. For
the key parameters, no scale effect was found for the
specimens with fibers, contrary to the ones without
fibers. Thus, it appears a characterisation of material per-
formance based on a scaled down version of the standar-
dised CMOD test can be justified. Numerical modelling of
the test by using the Concrete Damage Plasticity material
model of Abaqus, with a Thorenfeldt-based constitutive
law in compression and a customized constitutive law in
tension, based on tabulated stress and strain values deter-
mined from the CMOD test and mesh geometry, result in a
reasonable fit with the experimental results up to a
modest extent over the onset of fracture. For larger separ-
ations, however, an accurate model needs to be based on
constitutive behaviour obtained from uni-axial tensile
testing, as the stress values cannot be determined unam-
biguously from the CMOD test.
Acknowledgements
The assistance in the 3DCP research of Master track students
Structural Design at the TU/e Department of the Built Environ-
ment is highly valued. In particular, we appreciate the efforts of
W. Raedts and C. van der Krift as this study draws on their MSc
graduation theses (digitally available from the TU/e library
repository). The TU/e research programme on 3DCP is co-
funded by a partner group of enterprises and associations,
that on the date of writing consisted of (alphabetical order)
Ballast Nedam, BAM Infraconsult bv, Bekaert, Concrete Valley,
CRH, Cybe, Saint-Gobain Weber Beamix, SGS Intron, SKKB,
Van Wijnen, Verhoeven Timmerfabriek, and Witteveen+Bos.
Their support is gratefully acknowledged. In particular, the
authors would like thank SG Weber Beamix and Bekaert NV
for supplying the printing concrete and reinforcement fibers,
respectively.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
F. P. Bos http://orcid.org/0000-0002-6666-2395
E. Bosco http://orcid.org/0000-0003-1985-2150
References
3ders. 2017a. Accessed September 26, 2017. http://www.3ders.
org/articles/20161214-spain-unveils-worlds-first-3d-printed-
pedestrian-bridge-made-of-concrete.html.
3ders. 2017b. Accessed September 26, 2017. http://www.3ders.
org/articles/20170602-cybe-construction-completes-
3dprinting-of-168-sq-m-rdrone-laboratory-in-dubai.htm.
3dprinthuset. 2017. Accessed December 30, 2017. https://
3dprinthuset.dk/europes-first-3d-printed-building/.
Abdalla, H. M., and B. L. Karihaloo. 2003. “Determination of Size-
Independent Specific Fracture Energy of Concrete from
Three-Point Bend and Wedge Splitting Tests.” Magazine of
Concrete Research. 55 (2): 133–141.
Armelin, H. S., and N. Banthia. 1997. “Predicting the Flexural
Postcracking Performance of Steel Fiber Reinforced
Concrete from the Pullout of Single Fibers.” ACI Materials
Journal 94 (1): 18–31.
Asprone, D., F. Auricchio, C. Menna, and V. Mercuri. 2018. “3D
Printing of Reinforced Concrete Elements: Technology and
Design Approach.” Construction and Building Materials. 165:
218–231. doi:10.1016/j.conbuildmat.2018.01.018.
Asprone, D., C. Menna, F. P. Bos, T. A. M. Salet, J. Mata-Falcón,
and W. Kaufmann. 2018. “Rethinking Reinforcement for
Digital Fabrication with Concrete (DFC).” Cement and
Concrete Research 112: 111–121. doi:10.1016/j.cemconres.
2018.05.020.
Awinda, K., J. Chen, and S. J. Barnett. 2016. “Investigating
Geometrical Size Effect on the Flexural Strength of the
Ultra High Performance Fibre Reinforced Concrete Using
the Cohesive Crack Model.” Construction and Building
Materials 105: 123–131.
Banthia, N., and J. Trottier. 1994. “Concrete Reinforced with
Deformed Steel Fibers, Part I: Bond-Slip Mechanisms.” ACI
Materials Journal 91 (5): 435–446.
Bartos, P. J. M., and M. Duris. 1994. “Inclined Tensile Strength of
Steel Fibres in a Cement-Based Composite.” Composites 25
(10): 945–952.
Bažant, Z. P. 1984. “Size Effect in Blunt Fracture: Concrete,
Rock, Metal.” Journal of Engineering Mechanics 110 (4):
518–535.

Bažant, Z. P., and L. Cedolin. 1979. “Blunt Crack Band
Propagation in Finite Element Analysis.” Journal of the
Engineering Mechanics Division 105 (2): 297–315.
Bažant, Z. P., Joško Ožbolt, and Rolf Eligehausen. 1994. “Fracture
Size Effect: Review of Evidence for Concrete Structures.”
Journal of Structural Engineering. 120 (8): 2377–2398.
Bekaert, K. M. 2017. “EC Declaration of Performance DRAMIX®
OL 6/.16, 2017.” Accessed December 30, 2017. https://
www.bekaert.com/en/product-catalog/content/dop/dramix-
ol-ce-documents?filterLanguage=en&filterText=6.
Bos, F., Z. Ahmed, E. Jutinov, and T. Salet. 2017. “Experimental
Exploration of Metal Cable as Reinforcement in 3D Printed
Concrete.” Materials 10: 1314. doi:10.3390/ma10111314.
Bos, F., R. Wolfs, Z. Ahmed, and T. Salet. 2016. “Additive
Manufacturing of Concrete in Construction: Potentials
and Challenges of 3D Concrete Printing.” Virtual and Physical
Prototyping. 11: 209–225. doi:10.1080/17452759.2016.1209867.
Chanvillard, G., and S. Rigaud. 2003. “Complete Characterisation
of Tensile Properties of Ductal® UHPFRC According to the
French Recommendations.” In International Workshop High
Performance Fiber Reinforced Cement Composites, edited by
A. E. Naaman, and H. W. Reinhardt, 21–34. France: RILEM
Publications.
Cornelissen, H. A. W., D. A. Hordijk, and H. W. Reinhardt. 1986.
“Experimental Determination of Crack Softening Characteristics
of Normalweight and Lightweight Concrete.” Heron 31 (2):
45–56.
Di Prisco, M., M. Colombo, and D. Dozio. 2013. “Fibre-Reinforced
Concrete in the fib Model Code 2010: Principles, Models and
Test Validation.” Structural Concrete 14 (4), doi:10.1002/suco.
201300021.
Giaccio, G., J. M. Tobes, and R. Zerbino. 2008. “Use of Small
Beams to Obtain Design Parameters of Fibre Reinforced
Concrete.” Cement and Concrete Composites 30: 297–306.
Grassl, P., and M. Jirasek. 2006. “Plastic Model with Non-Local
Damage Applied to Concrete.” International Journal for
Numerical and Analytical Methods in Geomechanics 30: 71–90.
Hack, N., and W. V. Lauer. 2014. “Mesh-Mould: Robotically
Fabricated Spatial Meshes as Reinforced Concrete
Formwork.” Architectural Design 84 (3): 44–53.
Hack, N., W. V. Lauer, F. Gramazio, and M. Kohler. 2015. “Mesh
Mould: Robotically Fabricated Metal Meshes as Concrete
Formwork and Reinforcement.” In Proceedings of the 11th
International Symposium on Ferrocement and 3rd ICTRC
International Conference on Textile Reinforced Concrete,
edited by W. Brameshuber, 347–359. doi:10.1007/978-3-
319-26378-6_15.
Hambach, M., and D. Volkmer. 2017. “Properties of 3D-Printed
Fiber-Reinforced Portland Cement Paste.” Cement and
Concrete Composites 79: 62–70.
Hillerborg, A., M. Modéer, and P.-E. Petersson. 1976. “Analysis of
Crack Formation and Crack Growth in Concrete by Means of
Fracture Mechanics and Finite Elements.” Cement and
Concrete Research 6 (6): 773–781.
Hordijk, D. A. 1991. “Local Approach to Fatigue of Concrete.”
(PhD Thesis). Delft University of Technology.
International Federation for Structural Concrete. 2013. Fib
Model Code for Concrete Structures 2010. Berlin: Ernst &
Sohn. ISBN 978-3-433-03061-5.
Khoshnevis, B. 2004. “Automated Construction by Contour
Crafting-Related Robotics and Information Technologies.”
Automation in Construction 13: 5–19.
VIRTUAL AND PHYSICAL PROTOTYPING 173
Khoshnevis, B., and R. Dutton. 1998. “Innovative Rapid
Prototyping Process Makes Large Sized, Smooth Surfaced
Complex Shapes in a Wide Variety of Materials.” Materials
Technology 13: 53–56.
Krätzig, W., and R. Pölling. 2004. “An Elasto-Plastic Damage Model
for Reinforced Concrete with Minimum Number of Material
Parameters.” Computers and Structures 82: 1201–1215.
Lepech, M., and V. C. Li. 2003. “Preliminary Findings on Size
Effect in ECC Structural Members in Flexure.” Brittle Matrix
Composites 7 (1001): 48109–42125.
Lim, S., R. A. Buswell, T. T. Le, S. A. Austin, A. G. F. Gibb, and
T. Thorpe. 2012. “Developments in Construction-Scale
Additive Manufacturing Processes.” Automation in
Construction 21: 262–268.
Lubliner, J., J. Oliver, S. Oller, and E. Oñate. 1989. “A Plastic-
Damage Model for Concrete.” International Journal of Solids
and Structures 25 (3): 299–326.
Mahmuda, G. H., Z. Yang, and A. M. T. Hassan. 2013.
“Experimental and Numerical Studies of Size Effects of Ultra
High Performance Steel Fibre Reinforced Concrete (UHPFRC)
Beams.” Construction and Building Materials 48: 1027–1034.
Markovic, I. 2006. “High-Performance Hybrid-Fibre Concrete -
Development and Utilisation.” PhD thesis., Delft: Delft
University of Technology.
Mechtcherine, V., J. Grafe, V. N. Nerella, E. Spaniol, M. Hertel,
and U. Füssel. 2018. “3D-printed Steel Reinforcement for
Digital Concrete Construction – Manufacture, Mechanical
Properties and Bond Behaviour.” Construction and Building
Materials 179: 125–137. doi:10.1016/j.conbuildmat.2018.05.
202.
Naaman, A. E. 2003. “Engineered Steel Fibers with Optimal
Properties for Reinforcement of Cement Composites.”
Journal of Advanced Concrete Technology 1 (3): 241–252.
Nederlands Normalisatie Instituut. 2007. NEN-EN 14651+A1 (en)
Test Method for Metallic Fibered Concrete – Measuring the
Flexural Tensile Strength (Limit of Proportionality (LOP),
Residual). Delft: NEN.
Nederlands Normalisatie Instituut. 2016. NEN-EN 196-1:2016 en,
Methods of Testing Cement - Part 1: Determination of Strength.
Delft: NEN.
Nematollahi, B., M. Xia, J. Sanjayan, and P. Vijay. 2018.
“Effect of Type of Fiber on Inter-Layer Bond and Flexural
Strengths of Extrusion-Based 3D Printed Geopolymer.”
In Proceedings of the 2nd International Conference on
Advanced Manufacturing and Materials, 119–126. ICAMM,
Tokyo, Japan.
Nguyen, G. D., and A. M. Korsunsky. 2008. “Development of an
Approach to Constitutive Modelling of Concrete: Isotropic
Damage Coupled with Plasticity.” International Journal of
Solids and Structures 45: 5483–5501.
Ogura, H., V. Nerella, and V. Mechtcherine. 2018. “Developing
and Testing of Strain-Hardening Cement-Based Composites
(SHCC) in the Context of 3D-Printing.” Materials 11: 1375.
doi:10.3390/ma11081375.
Panda, B., S. Chandra Paul, and M. Jen Tan. 2017. “Anisotropic
Mechanical Performance of 3D Printed Fiber Reinforced
Sustainable Construction Material.” Materials Letters 209:
146–149.
Peerlings, R. H. J., R. de Borst, W. A. M. Brekelmans, and M. G. D.
Geers. 1998. “Gradient-Enhanced Damage Modelling of
Concrete Fracture.” Mechanics of Cohesive-Frictional Materials.
3: 323–342.
174 F. P. BOS ET AL.
Petersson, P. E. 1981. Crack Growth and Development of Fracture
Zones in Plain Concrete and Similar Materials. Lund: Division
of Building Materials, Lund Institute of Technology Lund.
Plückelmann, S., F. Song, and R. Breitenbücher. 2017. “Hybrid
Concrete Elements with Splitting Fiber Reinforcement
under Two-Dimensional Partial-Area Loading.” In High Tech
Concrete: Where Technology and Engineering Meet, edited
by M. Luković, and D. A. Hordijk, 347–355. Cham: Springer
International Publishing.
Poh, L. H., and S. Swaddiwudhipong. 2009. “Over-nonlocal
Gradient Enhanced Plastic-Damage Model for Concrete.”
International Journal of Solids and Structures 46 (25–26):
4369–4378.
Rilem Technical Committees. 2002. “Test and Design Methods
for Steel Fibre Reinforced Concret.” Materials and Structures
35: 579–582.
Robins, P., S. Austin, and P. Jones. 2002. “Pull-out Behaviour
of Hooked Steel Fibres.” Materials and Structures 35 (8):
434–442.
Rots, J. G., P. Nauta, G. M. A. Kusters, and J. Blaauwendraad.
1985. “Smeared Crack Approach and Fracture Localization
in Concrete.” Delft: Heron 30 (1): 1–48.
Rots, J., and R. de Borst. 1989. “Analysis of Concrete Fracture in
‘Direct’ Tension.” International Journal of Solids and
Structures 25 (12): 1381–1394.
Salet, T. A. M., F. P. Bos, R. J. M. Wolfs, and Z. Y. Ahmed. 2017. “3D
Concrete Printing – A Structural Engineering Perspective.” In
High Tech Concrete: Where Technology and Engineering Meet,
edited by M. Luković, and D. A. Hordijk, xliii–xlvii. Cham:
Springer International Publishing.
Salet, T. A. M., Z. Y. Ahmed, F. P. Bos, and H. L. M. Laagland. 2018.
Design of a 3D Printed Concrete Bridge by Testing.” Virtual
and Physical Prototyping 13 (3): 222–236.
Singh, H. 2017. “Steel Fiber Reinforced Concrete – Behavior,
Modelling and Design.” Springer Transactions in Civil and
Environmental Engineering. doi:10.1007/978-981-10-2507-5_2.
Slager, G. J. 2017. “Influence of the Interface between Layers on
the Tensile Properties of 3D-Printed Concrete.” MSc
Graduation Thesis, (A-2017.194), Eindhoven: TU/e.
Soltan, D. G., and V. C. Li. 2018. “A Self-Reinforced Cementitious
Composite for Building-Scale 3D Printing.” Cement and
Concrete Composites 90: 1–13. doi:10.1016/j.cemconcomp.
2018.03.017.
Stähli, P., and J. G. M. Van Mier. 2004. “Three-Fibre Type Hybrid-
Fibre Concrete.” In RILEM. 5th International Symposium on
Fracture Mechanics of Concrete and Concrete Structures
(FraMCoS), 1105–1112. ISBN: 0-87031-135-2.
Svec, O., J. Skoecek, J. Stang, J. F. Olesen, and L. N. Thrane. 2012.
“Application of the Fluid Dynamics Model to the Field of
Fibre Reinforced Self-Compacting Concrete.” In Proceedings
of the Numerical Modeling Strategies for Sustainable
Concrete Structures (NMSSC), edited by P. Rossi, and J.-L.
Tailhan, 1–9. Aix en Provence, France.
Thorenfeldt, E., A. Tomaszewicz, and J. J. Jensen. 1987.
“Mechanical Properties of High-Strength Concrete and
Applications in Design.” In Proc. Symp. Utilization of High-
Strength Concrete. Tapir, Trondheim, Norway.
Van Gysel, A. 2000. “Studie van het uittrekgedrag van staalve-
zels ingebed in een cementmatrix met toepassing op staal-
vezelbeton onderworpen aan buiging.” PhD Thesis,
Laboratory Magnel for Concrete Research, University of
Ghent.
Voyiadjis, G. Z., Z. N. Taqieddin, and P. I. Kattan. 2008.
“Anisotropic Damage–Plasticity Model for Concrete.”
International Journal of Plasticity 24: 1946–1965.
Weng, Y., S. Qian, M. Li, and M. J. Tan. 2018. “3D Printable High
Performance Fiber Reinforced Cementitious Composites for
Large-Scale Printing.” In 3rd International Conference on
Progress in Additive Manufacturing, edited by C. K. Chua,
W. Y. Yeong, M. J. Tan, E. J. Liu, and S. B. Tor. Singapore:
Nanyang Technological University.
Wijffels, M. J. H., R. J. M. Wolfs, A. S. J. Suiker, and T. A. M. Salet.
2017. “Magnetic Orientation of Steel Fibres in Self-
Compacting Concrete Beams: Effect on Failure Behaviour.”
Cement and Concrete Composites 80: 342–355.
Wu, P., J. Wang, and X. Wang. 2016. “A Critical Review of the Use
of 3-D Printing in the Construction Industry.” Automation in
Construction 68: 21–31.
Yoo, D. Y., N. Banthia, J.-M. Yang, and Y.-S. Yoon. 2016. “Size
Effect in Normal- and High-Strength Amorphous Metallic
and Steel Fiber Reinforced Concrete Beams.” Construction
and Building Materials 121: 676–685.
Zareiyan, B., and B. Khoshnevis. 2018. “Effects of Mixture
Ingredients on Interlayer Adhesion of Concrete in Contour
Crafting.” Rapid Prototyping Journal 24 (3): 584–592. doi:10.
1108/RPJ-02-2017-0029.