Construction and Building Materials 65 (2014) 620–629

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Influences of geometric patterns of 3D spacer fabric on tensile behavior

of concrete canvas
Fangyu Han a, Huisu Chen a,⇑, Kefeng Jiang a, Wulong Zhang b, Tao Lv b, Yujie Yang b
a
Jiangsu Key Laboratory of Construction Materials, School of Materials Science & Engineering, Southeast University, Nanjing 211189, China
b
Architectural Engineering Institute of the General Logistic Department, P.L.A., Xi’an, China

h i g h l i g h t s

 A promising concrete reinforcement material was introduced for new concrete applications.
 Influences of geometric patterns of 3D spacer fabric on the tensile behavior of concrete canvas were investigated.
 Concrete canvas reinforced by 3D spacer fabric with one solid outer textile substrate exhibited improved tensile behaviors.
 3D spacer fabric reinforcement was a better option than spacer yarns alone.

a r t i c l e i n f o a b s t r a c t

Article history: In this study, influences of geometric patterns of 3D spacer fabric on the tensile behavior of concrete
Received 20 February 2014 canvas (CC) were investigated. Five 3D spacer fabrics with different geometric patterns were investigated.
Received in revised form 14 May 2014 Tensile stress–strain curves and crack propagating patterns of CCs were obtained in warp and weft
Accepted 15 May 2014
directions through experiments. Experimental results revealed that the CC samples reinforced by the
Available online 9 June 2014
3D spacer fabrics with one solid outer textile substrate exhibited improved tensile behaviors in terms
of tensile strength, reinforcing efficiency factor and crack pattern. Moreover, for CCs, 3D spacer fabric
Keywords:
reinforcement was a better option than spacer yarns alone.
3D spacer fabric
Geometric pattern
Ó 2014 Elsevier Ltd. All rights reserved.
Concrete canvas
Tensile behavior
Crack pattern

1. Introduction
As a new composite material, textile reinforced concrete (TRC)
has various outstanding properties such as larger load-bearing
capacity, excellent ductility, thinner thickness, light-weight of
components, resistance to corrosion and no magnetic disturbances
[1–5]. Textile used as reinforcement can significantly improve ten-
sile strength of concrete. However, the tensile strength, ductility
and cracking pattern of TRC depends not only on its components,
but also the bonding between reinforcement and the matrix which
is influenced by the geometries of textile including weft yarns
spacing [5], stitches size [6,7], and bundle size of yarns [6,8].
The 3D spacer fabric is a more attractive product when it is used
for concrete applications [9,10]. As a special textile, 3D spacer
fabric presents various advantages and shows a superior behavior
as it could reinforce the cementitious composites in three
⇑ Corresponding author. Tel.: +86 25 52090645.
E-mail addresses: chenhs@seu.edu.cn (H. Chen), musult@126.com (T. Lv).
directions and provide defined positioning of the two outer sub-
strates [10]. 3D spacer fabric has been successfully applied in civil
engineering [11,12] such as sandwich panels, façade elements and
claddings etc. The superiority of applying 3D spacer fabric in con-
crete is the possibility to produce concrete elements with desired
cross-section and reinforcing yarns along the thickness of the ele-
ments. Because of good stability of 3D spacer fabric, it also allows
the design of extremely-thin-structural concrete elements with
outstanding mechanical performance [13–15].
Among various engineering applications of 3D spacer fabric
reinforced cementitious composites, concrete canvas (CC, the con-
cept was first proposed by Brewin and Crawford in 2005) is one of
the most promising products [16]. For conventional 3D spacer fab-
ric reinforced cementitious composites, one needs to mix water
with dry powder firstly, then cast the fresh mixture into the mold
with 3D spacer fabric, and demold till it is harden. However, CC has
different preparation process. In initial stage, CC is a flexible 3D
spacer fabric impregnated with cement powder. Like soft cloth,
CC can closely cover the surface of arbitrary structure or element

http://dx.doi.org/10.1016/j.conbuildmat.2014.05.041
0950-0618/Ó 2014 Elsevier Ltd. All rights reserved.
 F. Han et al. / Construction and Building Materials 65 (2014) 620–629
before using. Then, one just needs to spray or add water to the top
surface of CC. After CC hardens, a thin, durable water-proof and
fire-proof composite layer forms. Its shape is completely the same
as the outer profile of the structure or element where CC is cov-
ered. Therefore, CC can be quickly, efficiently and widely used in
civil engineering such as a cover of prefabricated shelter, a track-
way for vehicles, pedestrians or protection layer for pipe and
lining.
In order to apply CC in practice, its mechanical properties
should be investigated thoroughly. As a load-bearing element,
the geometric patterns of 3D spacer fabric would undoubtedly
affect the mechanical properties of CC. However, to the authors’
knowledge, the influences of geometric patterns of 3D spacer fabric
on the tensile behavior of CC have barely investigated in the
literatures.
Therefore, in this study, we studied the influences of geometric
patterns of 3D spacer fabric on tensile behavior of CC. Five
PET-based 3D spacer fabrics with different geometric patterns
were investigated. Tests have been done to obtain warp and weft
tensile stress–strain curves and the crack propagating patterns of
CCs.
2. Experimental program
2.1. 3D spacer fabrics
A typical 3D spacer fabric and the inside components are presented in Fig. 1. The
warp yarns are inserted into the stitches and assembled together with the weft
yarns, by this way a grid net can be produced and the meshes in the net can be knit-
ted in various shapes. In addition, two different kinds of spacer yarns can be
inserted into the structure. One of them, spacer yarn I, is vertical to the outer textile
substrates, the other one, spacer yarn II, is inclined to the outer textile substrates.
The warp direction is along machine direction while the weft direction is along
the orthogonal direction of warp direction [11].
In this study, five 3D spacer fabrics with different geometric patterns of outer
textile substrates and different amounts of spacer yarns were investigated to study
the tensile behavior of CCs. As described in Table 1, T20 is a 3D spacer fabric with
20 mm thickness whose two outer textile substrates are identical mesh fabric. It has
both spacer yarns I and II, and the pore shape of both substrates is regular triangle.
The warp/weft yarn and spacer yarn are made by 342 dtex PET multifilament and
379 dtex PET monofilament, respectively. The other four are 3D spacer fabrics with
15 mm thickness (denoted as N15) in which only one of their outer textile sub-
strates is mesh fabric and the other is solid fabric. The architecture of solid fabric
of N15 is tricot, where the architecture of solid fabric of N15-IV is much looser than
that of the others. Only spacer yarn II is inserted into their structure. The pore shape
of mesh fabric is regular rhombus for N15-I and N15-II, and regular square for N15-
III and N15-IV. The warp/weft yarn of solid fabric and mesh fabric are made by 396
dtex and 339 dtex PET multifilament, respectively. The spacer yarn is made by 495
dtex PET monofilament. In addition, all warp/weft yarns of mesh fabrics of 3D
spacer fabrics are in twisted form. The other structural parameters of the fabrics
are given in Table 2.
2.2. Specimens preparation and curing
Raw materials for the matrix used in this study were anhydrite and calcium
sulphoaluminate cement (CSA). Their compositions are given in Table 3. The CSA
cement used contains 65.5% ye’elimite and its Blaine specific surface is 442 m2/
kg. The anhydrite has a Blaine specific surface of 387 m2/kg. Their particle size dis-
tributions determined by laser diffraction (Microtrac S3500) are given in Fig. 2.
Fig. 1. A typical 3D spacer fabric: (a) global view and (b) side view.
621
Optimized mix proportion of matrix is shown in Table 4, in which anhydrite is
20% of CSA by weight and water/binder ratio was fixed at 0.45. The selected
water/binder ratio was calculated based on the mass of sample before and after
spraying in preliminary exploration work, where water-spray stopped immediately
if the water just penetrates through the bottom layer of 3D spacer fabric.
The preparation procedure was conducted at temperature of 25 °C and relative
humidity of 75%. Firstly, CSA and anhydrite were poured into a Thunderbird
ARM-02 mixer and stirred for 10 min at 94 r/min. Then, the powder mixture
was gradually placed and vibrated into the mold with 3D spacer fabric of
400 mm  100 mm  20 mm (i.e. T20 type fabric) or 400 mm  100 mm  15 mm
(i.e. N15 type fabrics) until the 3D spacer fabric was fully impregnated with powder.
In the same way, the unreinforced control samples with the sizes of
400 mm  100 mm  20 mm were also prepared. To assess the bonding strength
between yarns and the matrix, a type of dog-bone mold with dimensions of
235 mm  25 mm  25 mm was adopted (as shown in Fig. 3). A slot with 2 mm
width was located in the middle of mold. A PVC slice with a single yarn passing
through the center is insert into the slot. The embedded lengths of the single yarn
on both sides of slice are 15 mm and 80 mm, respectively. Finally, tap water with
temperature of 24.3 °C was sprayed into the mold until the water/cement ratio
reached 0.45. Specimens were demolded after final setting, and moved into the
standard curing room (T = 20 ± 2 °C and RH P 95%). To check whether water pene-
trates through this whole thickness of the sample, we cut along the cross section of
the sample after hardened, and the result is given in Fig. 4. It can be seen from Fig. 4
that the density of the paste in bottom layer is similar to the top layer. Therefore, it
demonstrates that the current water-spaying method may guarantee water pene-
trate through the whole thickness of sample. To obtain the suitable curing age for
tensile test, the time-dependent compressive strength of cubic CSA-based CC sam-
ple with dimensions of 15 mm  15 mm  15 mm was preliminarily investigated
and the result is given in Fig. 5. It revealed that 10-day compressive strength of
the sample almost reaches the maximum mechanical strengths value. In addition,
the mechanical characteristics of commercial CC made by Concrete Canvas Ltd
was evaluated after 10-day curing. For the sake of comparison, the tensile behavior
of CCs would be conducted after 10 days curing. According to the difference in
the type of 3D spacer fabrics as shown in Table 1, the CC samples were labeled as
T20-CC, N15-I-CC, N15-II-CC, N15-III-CC and N15-IV-CC, respectively.
2.3. Test methods
2.3.1. Tensile tests of yarns and 3D spacer fabrics
The tensile tests of yarns were conducted on XL-2 yarn tensile tester with 30 N
capacity. The stroke-control rate was set at 50 mm/min and the initial distance
between the two clamps was set at 250 mm. Each group has 10 samples. The tensile
tests of 3D spacer fabrics were conducted on CMT4104 tensile machine with 10 kN
capacity. The stroke-control rate was set at 10 mm/min. The sizes of specimens are
350 mm  70 mm  20 mm for T20 and 350 mm  70 mm  15 mm for N15. The
tensile tests of 3D spacer fabrics were separately conducted in warp and weft direc-
tions considering the anisotropy. Each group has three samples. Before tests, all
specimens were cured under temperature of 20 ± 2 °C and relative humidity of
65.0 ± 4.0% for 12 h. Tensile stress–strain curves were obtained for all yarns and
3D spacer fabrics, but only the average values and standard deviation were reported
in this study.
2.3.2. Yarns pull-out tests
The bonding strength between yarns and the matrix were tested by using a
CMT4103 electromechanical universal tensile machine with 1 kN load cell. The
stroke-control rate was set at 0.5 mm/min. Five parallel samples of each group were
tested and the average values and standard deviation were presented. Almost all
yarns were pull out from the side with shorter embedded length of yarn. The curves
of pull-out load P per unit remaining embedded length versus slip displacement Ds
were obtained. The remaining embedded length is the difference between the ini-
tial embedded length L and the slip displacement Ds. Considering the very short
region of embedded yarns that adhesive bonding strength exerts on, it is assumed
that the yarns are held in matrix only by frictional bonding with no adhesive
bonding.
622 F. Han et al. / Construction and Building Materials 65 (2014) 620–629

Table 1
Structure and view of 3D spacer fabrics.

3D Face view of outer Side view of spacer yarns Back view of outer
spacer textile substrates Warp direction Weft direction textile substrates
fabrics

T20

N15-I

N15-II

N15-III

N15-IV

Table 2
Structural parameters of 3D spacer fabrics.

3D spacer fabric T20 N15-I N15-II N15-III N15-IV

Density (kg/m3) 57.5 94.6 48.6 52.3 90
Amount of spacer yarn (spacer yarns/cm2) 140 70 35 35 70

Table 3
Mineralogical and chemical compositions of CSA and anhydrite (wt.%).

Mineralogical compositions CSA Anhydrite Chemical compositions CSA Anhydrite

C4A3 S 65.51 – SiO2 8.5 0.68

CT 6.48 – Al2O3 32.6 0.40
C2S-b 16.52 – Fe2O3 2.7 –
C4AF 2.12 – CaO 41.7 39.0
C12A7 4.27 – MgO 3.5 1.95
CS 0.45 92.64 SO3 9 49.5
C S H2 – 6.80 TiO2 1.5 –
C S H0.5 – 0.56 L.O.I 0.5 5.46
CaO 1.52 –
CaMg(CO)2 3.14 –

2.3.3. Tensile tests of CCs
Due to the anisotropy of 3D spacer fabrics, the tensile behavior of CCs in
warp and weft direction should be different and both were investigated in this
study.
The tensile tests of CCs were performed on a CMT4104 electromechanical
universal testing machine with 10 kN load cell. The stroke-control rate was set at
1 mm/min. The sample sizes of CCs used for uniaxial tensile tests are shown in
Fig. 6, the thickness of specimens is identical with that of the corresponding 3D
spacer fabric. To avoid crack growth in the vicinity of clamps, a double-edge notch
with dimensions of 3 mm  15 mm was designed for the unreinforced control sam-
ples and T20-CCs. Metal plates were glued onto both ends of specimens by using
epoxy resin to alleviate localized damage and minimize deformation of matrix at
clamps. A hole with 8 mm diameter was drilled at the center of the metal plates.
Load was transferred through the dowels at the center of metal plates to prevent
torsional and bending moments of specimens owing to the possible misalignments
of the constraints. The average values and standard deviation of at least three par-
allel samples were acquired for different CCs in warp and weft direction, respec-
tively. Typical stress–strain curves of different CCs were compared. Testing was
stopped after samples failed or when the maximum strain reached 10% in warp
direction and 25% in weft direction. First crack strength and maximum bridging
stress were acquired for all CCs. The first crack strength rfc is defined as the applied
tensile stress at which a matrix crack spreads throughout the cross section of the
sample under tension. The maximum bridging stress rB is defined as the maximum
stress that bridging fibers can transfer across the crack of specimen.
 F. Han et al. / Construction and Building Materials 65 (2014) 620–629
Fig. 2. Particle size distribution of CSA cement and anhydrite.
Table 4
Mixture proportion of matrix (kg/m3).
Cement Anhydrite
750 150
Fig. 3. Schematic of dog-bone specimen for yarns pull-out tests.
3. Results and discussion
3.1. Tensile properties of yarns
The tensile stress–strain curves of all yarns are shown in Fig. 7.
The corresponding results are summarized in Table 5. The ultimate
elongation of yarns of N15 is obviously lower than that of T20,
while the strength of yarns of N15 is relatively higher than that
of T20 except that of mesh fabric (MF). The warp/weft yarns of
solid fabric (SF) for N15 have the highest strength followed by
the spacer yarns, and then by the warp/weft yarns of mesh fabric
(MF). The strength of spacer yarns for T20 is higher than that of
its warp/weft yarns.
3.2. Tensile properties of 3D spacer fabrics
Fig. 8 presents tensile behaviors of all the 3D spacer fabrics. The
behaviors can be divided into two main stages by a jamming tran-
sition phase [17]. The first stage is slipping of the warp/weft yarns
at the interlock point between adjacent loops. A change of the loop
shape occurs by yarn bending and results in a linear behavior of the
3D spacer fabric under low loads as demonstrated in stretched
stage of Fig. 8, where the N15-IV is chosen as a representative
example for N15. This phenomenon maintains until the jamming
transition phase happens, in which a transverse compression of
the warp/weft yarns appears due to the warp/weft yarns rear-
rangement at the cross-over. The warp/weft yarns bending tends
toward a limit while the tensile load increases, during this stage,
623
Fig. 4. Cross section of the hardened samples.
Water
405
Fig. 5. Time-dependent compressive strength of CC sample.
rupture of mesh fabric of N15 was observed due to its relative
low strength as shown in Fig. 8. At the end of this stage, the stitch
was fully jammed. A fast increase of the load was thus observed
which brings about a second almost linear domain of the stress–
strain curve. In the end, only the solid fabric carries the load for
N15 due to the rupture of mesh fabric during tensile test as illus-
trated in failure state of Fig. 8.
As shown in Fig. 8, the tensile strength and ultimate elongation
of T20 is the lowest in warp and weft direction except the ultimate
elongation in weft direction. The much larger elongation of T20 in
weft direction is attributed to the detachment between weft yarns
and warp yarns of both outer textile substrates during tensile pro-
cess as illustrated in stretched state of Fig. 8(b). The higher tensile
strength of N15 is resulted from the particular geometric pattern of
outer textile substrates where solid fabric can bear much higher
load than mesh fabric. The tensile properties of N15 are almost
analogous except N15-IV. The lower strength of N15-IV is attrib-
uted to the looser architecture of solid fabric. These results indicate
that the geometric patterns of outer textile substrates of 3D spacer
fabric play a vital role in loading capacity and the solid fabric dom-
inate the tensile strength and ultimate elongation of N15.
3.3. Bond behavior of yarns
The shear force versus slip curves in Fig. 9 presents the debond-
ing characteristics of various yarns with the matrix. Instead of pull-
ing out from the matrix, the warp/weft yarns ruptured at small slip
displacement which indicates that the critical pull-out length for
all the warp/weft yarns are very short. The critical pull-out length
of yarn is defined as the maximum embedded length of a yarn to be
624 F. Han et al. / Construction and Building Materials 65 (2014) 620–629

3.4. Tensile behavior of CCs

3.4.1. Influence of structural configurations of 3D spacer fabrics

Fig. 6. Schematic of tensile specimen.
Fig. 7. Tensile stress–strain curves of various yarns.
pulled out from a matrix without rupture which is related to the
tensile strength of the yarn as well as the yarn/matrix bonding
strength. It should be noted that the outer textile substrates are
normally exposed on the surface of hardened cement matrix rather
than anchored into matrix, and the warp/weft yarns of outer textile
substrates are continuous and assembled together. Therefore, the
contribution of warp/weft yarns of outer textile substrates on the
bridging stress mainly comes from the tensile strength of these
yarns themselves.
As shown in Fig. 9, the spacer yarns are pulled out from matrix
and two different stages in the descending part of curves can be
observed: quickly descending stage and gently descending one.
The latter stage tends to be a horizontal line. The frictional bonding
strength depends on the average value of the height in the gently
descending stage of the curves. The results are presented in Table 6.
And the frictional bonding strength of spacer yarn of T20 is slightly
higher than that of N15.
To investigating the influence of structural configurations of 3D
spacer fabrics on the tensile behavior of CCs, two kind 3D spacer
fabric (T20 and N15-I) with different structural configurations
were considered. In addition, the tensile behavior of the unrein-
forced control samples was also investigated to compare with
the uniaxial tensile behavior of CCs. The experimental results are
given in Figs. 10–12 and the tensile properties of CCs are summa-
rized in Table 6.
As shown in Fig. 10, as expected, only one single crack occurred
in the unreinforced control samples and their ultimate tensile
strength was apparently lower than that of the specimens cast pro-
duced by traditional vibration-compacted process. The average
ultimate tensile strength of the unreinforced control samples is
0.63 MPa.
Due to the structural configurations of 3D spacer fabrics and
loading directions, the tensile behavior of CCs exhibits a high diver-
sity. Irrespective of loading directions, T20-CCs behave similarly
and two different stages are clearly identified in the tensile
stress–strain curves as shown in Fig. 11: linear elastic stage and
yarns pull-out stage. However, one more stage of solid fabric fail-
ure is observed in the tensile stress–strain curves of N15-I-CCs in
Fig. 12 where the same behavior is evident for N15-I-CCs in each
direction. Difference between stages is attributed to the variations
in the structural configurations of 3D spacer fabrics. Due to one of
outer textile substrates of N15-I is solid fabric, It can bear much
higher load than T20 as demonstrated in Fig. 8, and even higher
than the maximum bridging stress induced by 3D spacer fabric
during yarns pull-out stage. This led to the appearance of solid fab-
ric failure stage. It is particularly noted that the matrix was com-
pletely out of function and only the solid fabric carried the load
in this stage. At this moment, the tensile behavior of N15-I-CCs
was totally determined by the tensile performance of solid fabric,
and the specimen failed when the ultimate tensile strength of
the reinforcement reached.
In the linear elastic stage, the load is primarily carried by the
matrix until the matrix cracks and the slope reflects the elastic
modulus of CCs. Once the first crack strength of CC is reached,
the whole force is transferred through the crack to the 3D spacer
fabric. The first crack strength of CC is attributed to two aspects:
matrix strength and fiber bridging effect [18]. As demonstrated
in Table 6, compared to the unreinforced control samples, a slight
increase of first crack strength of the both CCs is normally observed
due to the weak bonding of yarns with matrix as well as the low
Young’s modulus of PET. Due to the lower bridging stress induced
by spacer yarns in warp direction, the first crack strength of N15-I-
CCs in weft direction is higher than that in warp direction. It is
attributed to the layout pattern of spacer yarns as shown in Table 1
where the spacer yarns are almost vertical to load direction in
warp direction. However, due to the larger yarns volume fraction
in warp direction, the situation is reversed when it comes to the
first crack of T20-CCs.

Table 5
Properties and geometries of various yarns of different 3D spacer fabrics.

3D spacer fabrics Component (–) Diameter (mm) Fineness (dtex) Ultimate elongation (%) Tensile strength (MPa) Young’s modulus (GPa)
T20 Warp/weft yarn 0.18b 342 19.3 (0.6)a 445 (6)a 2.36 (0.05)a
Spacer yarn 0.20 379 18.1 (1.8)a 567 (9)a 2.99 (0.07)a
N15 Warp/weft yarn_MF 0.14b 339 9.2 (0.6)a 443 (8)a 4.85 (0.19)a
Warp/weft yarn_SF 0.12b 396 6.1 (0.5)a 731 (24)a 11.39 (0.46)a
Spacer yarn 0.18 495 9.8 (0.8)a 705 (26)a 7.23 (0.65)a
a
Numbers in parentheses are the standard deviations.
b
Determined by their fineness and density.
 F. Han et al. / Construction and Building Materials 65 (2014) 620–629 625

Fig. 8. Tensile stress–strain curves of 3D spacer fabrics: (a) warp direction and (b) weft direction.

by both components. In this stage, irrespective of the structural

Fig. 9. Shear force per unit remaining embedded length versus slip displacement of
pull-out tests for various yarns in matrix.
Because of the bonding between 3D spacer fabric and matrix,
the reinforcement redistributes the load after the first crack, thus
brings about the yarns pull-out stage where the load is carried
configurations of 3D spacer fabrics and loading directions, the
maximum bridging stress of the both CCs is similar as demon-
strated in Table 6. Apparently, the maximum bridging stress of
CCs is closely related to the bonding behavior between 3D spacer
fabric and matrix. However, since the spacer yarns of 3D spacer
fabric are interlocked together with two outer textile substrates,
the bonding mechanism of 3D spacer fabric with matrix is quite
complex. It is mainly attributed to two aspects: one is inter-
anchorage effect of spacer yarns with warp/weft yarns during
pull-out test, the other is the contribution of outer textile sub-
strates of 3D spacer fabric, in which the outer textile substrates
are normally exposed on the surface of hardened cement matrix
rather than anchored into matrix.
In the yarns pull-out stage, T20-CCs produce only one single
crack which propagated from the first crack as shown in Fig. 11
while N15-I-CCs produce multiple cracking in weft direction as
shown in Fig. 12(b). The difference observed demonstrates that
the solid fabric of N15-I plays a crucial role on multiple cracking.
However, only one single crack occurred in warp direction for
N15-I-CCs as shown in Fig. 12(a), it indicates that the occurrence
of multiple cracking for N15-I-CCs is not merely caused by the solid
fabric. As discussed above, a higher bridging stress induced by
N15-I is obtained in weft direction due to the layout pattern of
spacer yarns. Probably, the interaction of spacer yarns and solid

Table 6
Tensile strength and reinforcing efficiency factor of various CCs.

CCs Yarn volume fraction Orientation angle of yarn Bond strength At first crack At maximum bridging stress
Sb (%) Wc (%) h (–) u (–) ss (MPa) rfc (MPa) rB (MPa) g (–)
T20-CCs Warp 3.98 0.28/0.28d 0.31p 0/0.33pd 0.44 (0.03)a 0.78 (0.09)a 1.16 (0.06)a 0.235
Weft 1.99 0.28/0.28d 0.33p 0.5p/0.17pd 0.74 (0.07)a 1.08 (0.08)a 0.378
N15-I-CCs Warp 2.91 0.35/0.35d 0.37p 0.17p/0.17pd 0.61 (0.13)a 1.14 (0.06)a 0.197
Weft 2.91 0.35/0.35d 0.39p 0.33p/0.33pd 0.67 (0.09)a 1.04 (0.02)a 0.294
N15-II-CCs Warp 1.46 0.35/0.35d 0.36p 0.17p/0.17pd 0.55 (0.07)a 1.07 (0.11)a 0.192
Weft 1.46 0.35/0.35d 0.38p 0.33p/0.33pd 0.39 (0.02)a 0.64 (0.07)a 0.96 (0.04)a 0.285
N15-III-CCs Warp 1.46 0.35/0.35d 0.27p 0/0.5pd 0.78 (0.05)a 0.97 (0.05)a 0.279
Weft 1.46 0.35/0.35d 0.29p 0.5p/0d 0.95 (0.12)a 1.10 (0.07)a 0.319
N15-IV-CCs Warp 2.91 0.35/0.35d 0.28p 0/0.5pd 0.68 (0.11)a 1.12 (0.07)a 0.293
Weft 2.91 0.35/0.35d 0.30p 0.5p/0d 0.49 (0.03)a 0.57 (0.06)a 0.151
a
Numbers in parentheses are the standard deviations.
b
Spacer yarns.
c
Warp/weft yarns of single mesh fabric.
d
Left and right side of the semicolon are volume or orientation angle of warp yarns and weft yarns, respectively.
626 F. Han et al. / Construction and Building Materials 65 (2014) 620–629
Fig. 10. Tensile stress–strain curves of the unreinforced control samples.
fabric leads to multiple cracking of N15-I-CCs in weft direction
during tensile test. Compared Figs. 11 and 12, it is noted that the
multiple cracking phenomenon in weft direction significantly
improves the ductility of CCs. This phenomenon indicates that
N15 is a better reinforcement for CCs in terms of crack pattern.
3.4.2. Influence of the existence of outer textile substrates
Moreover, in order to analyze the contribution of outer textile
substrates of 3D spacer fabric to the tensile behavior of CCs, the
tensile behavior of CCs without outer textile substrates were inves-
tigated and the results are also given in Figs. 11 and 12. The
method of removing outer textile substrates of 3D spacer fabric
is to polish it out of the CC sample by abrasion paper before tests.
As can be seen in Fig. 11, the tensile behavior is similar for the
specimens without outer textile substrates, irrespective of loading
directions. Compared with the specimens with outer textile
substrates, two different stages are also clearly identified in the
tensile stress–strain curves of the specimens without outer
textile substrates: linear elastic stage and yarns pull-out stage.
The similar situation goes to the crack pattern, only one single
crack occurs in the specimens without outer textile substrates in
both directions. These results indicate that the tensile behavior of
T20-CCs is not controlled by the outer textile substrates of T20.
Fig. 11. Tensile stress–strain curves of T20-CCs with or without outer textile substrates: (a) warp direction and (b) weft direction.
As shown in Fig. 12, the specimens without outer textile sub-
strates behave similarly, irrespective of loading directions. How-
ever, the tensile behavior of the specimens without outer textile
substrates is significantly different from that of the specimens with
outer textile substrates. Due to the lack of solid fabric, the tensile
stress–strain curves of the specimens without outer textile sub-
strates do not have the stage of solid fabric failure any more in both
directions. Another distinct difference is that they do not exhibit
multiple cracking in weft direction because of the lack of solid fab-
ric as well. These differences observed indicate that the tensile
behavior of N15-I-CCs is remarkably controlled by the outer textile
substrates of N15-I.
In Figs. 11 and 12, it can be seen that similar first crack strength
of the specimens with or without outer textile substrates is
obtained in both directions. This may be due to the weak bonding
between outer textile substrates with hardened cement matrix in
which the outer textile substrates are normally exposed on the sur-
face of hardened cement matrix rather than anchored into matrix.
If the outer textile substrates of CCs are removed, a great decrease
of maximum bridging stress and ductility would be clearly
observed in both directions, as expected. These results suggest that
3D spacer fabric is a more advisable reinforcement than spacer
yarns alone, especially for N15-I-CCs.
3.4.3. Influence of geometric configurations of 3D spacer fabrics
As discussed above, the N15 is a better reinforcement for CCs in
terms of crack pattern. To further investigate the influence of geo-
metric configurations on the tensile behavior of CCs, four different
types of N15 were considered. The differences in the geometric
configurations of selected 3D spacer fabrics are the amount of
spacer yarns, orientation angle of warp/weft yarns and the tight-
ness of solid fabric architecture as shown in Tables 1 and 2. The
experimental results are shown in Fig. 13 and tensile properties
are summarized in Table 6 as well.
Irrespective of the geometric configurations of 3D spacer fabric,
the tendency of tensile stress–strain curves of N15-CCs are similar.
By comparing the tensile strength of N15-I-CCs, N15-II-CCs and
N15-III-CCs in Table 6, a slight improvement of both first crack
strength and maximum bridging stress are observed with the
increase of the amount of spacer yarns, which is attributed to
the higher bridging stress induced by spacer yarns. Moreover, as
the orientation angle of warp/weft yarns decreases, both first crack
strength and maximum bridging stress increase. This is due to the
 F. Han et al. / Construction and Building Materials 65 (2014) 620–629 627

Fig. 12. Tensile stress–strain curves of N15-I-CCs with or without outer textile substrates: (a) warp direction and (b) weft direction.

Fig. 13. Tensile stress–strain curves of N15-CCs: (a) warp direction and (b) weft direction.

larger contribution of warp/weft yarns strength with the decreased
orientation angle.
Table 7 shows the average crack spacing xd and the ultimate
strain eu of N15-I-CCs, N15-II-CCs and N15-III-CCs when multiple
cracking terminates. It can be seen from Table 7 that increasing
the amount of spacer yarns leads to a decrease of about 23% of
the xd and an increase of about 8% of the eu. Furthermore, with
the orientation angle of warp/weft yarns decreases, the xd
decreases about 12% and the eu increases 5%. These are attributed
to that more flaws of matrix would be activated during multiple
cracking when the bridging stress induced by spacer yarns
Table 7
Ultimate crack spacing and ultimate strain of CCs reinforced by N15-I, N15-II and
N15-III when multiple cracking terminates.
CCs xd (mm)
N15-I-CCs 6.767 (0.673)a
N15-II-CCs 8.738 (0.367)a
N15-III-CCs 7.710 (0.382)a
a
Numbers in parentheses are the standard deviations.
and/or the contribution of warp/weft yarns strength increase.
Therefore, it could be concluded that the ductility of CCs can be
significantly improved by increasing the amount of spacer yarns
and/or decreasing the orientation angle of warp/weft yarns of 3D
spacer fabric.
Furthermore, it is worth noting that the degree of multiple
cracking is also dependent on the tensile performance of solid fab-
ric of N15. As illustrated in Fig. 13(b), the failure of N15-IV-CCs at
lower tensile stress in weft direction, compared with other N15-
CCs, reveals that the degree of multiple cracking does not reach
saturation (according to Kabele [19], the saturation of multiple
cracking is defined as a state when all multiple cracks continue
to open but no new cracks are generated) due to the lower tensile
strength of N15-IV. In addition, as shown in Fig. 14, the average
ultimate crack width when multiple cracking terminates was
much larger than ECC materials (less than 60 lm) [20] and it is
about 0.56 mm (measured by digital calipers in the middle of crack
eu (–)
along the side section of samples as shown in Fig. 14). Due to the
0.090 (0.004)a large crack width, the ultimate strains of N15-CCs were very high.
0.083 (0.006)a
However, it also significantly increased water permeability
0.087 (0.006)a
(squared increased with crack width) [21], and thereby reducing
the mechanical and durability performances of CCs.
628 F. Han et al. / Construction and Building Materials 65 (2014) 620–629
Fig. 14. Multiple cracking pattern of N15-CCs in weft direction.
3.5. Reinforcing efficiency factors
As demonstrated in Table 6, the difference in the maximum
bridging stress of various CCs is little except N15-IV-CCs in weft
direction. To compare the reinforcing efficiency of different 3D
spacer fabrics on the similar post-cracking tensile strength of
CCs, an reinforcing efficiency factor g is proposed which is defined
as the ratio of the maximum bridging stress of CCs to the total
maximum stress contribution from each components of 3D spacer
fabric along loading direction. As discussed above, the contribution
of warp/weft yarns and spacer yarns to the maximum bridging
stress mainly comes from their tensile strength and bonding
strength. Moreover, considering the fact that solid fabric is not bro-
ken at maximum bridging stress of N15-CCs due to its higher ulti-
mate tensile strength, it is assumed that the contribution of solid
fabric to the maximum bridging stress of N15-CCs is identical with
that of the mesh fabric. Therefore, the reinforcing efficiency factor
g of 3D spacer fabric can be calculated by Eq. (1)
rB
g¼
2V w rw cos u þ rsm
In Eq. (1), rB is the maximum bridging stress of CCs (MPa) obtained
from uniaxial tensile tests, rsm is the maximum bridging stress
induced by spacer yarns (MPa), rw is the tensile strength of warp/
weft yarns (MPa), Vw is the volume fraction of warp/weft yarns of
single mesh fabric (–), u is the orientation angle of warp/weft yarns
along the loading direction (–). According to Li [22], rsm can be cal-
culated by Eq. (2) for oriented fiber system
"    2 #
4ss V s cos h Ls 2do Ls 2do
rsm ¼ 1  1
ds 2 Ls 4 Ls
where ts is frictional bonding strength of spacer yarn with matrix
(MPa), Ls is spacer yarn length (mm), Vs is spacer yarn volume frac-
tion (–), ds is spacer yarn diameter (m), do = (Ls2 ts)/[(1 + g) Esds], and
it corresponds to the slip displacement at which debonding is com-
pleted along the full length of the embedded spacer yarn segment
(m). In addition g = (VsEs)/(VmEm), where Es and Em are Young’s mod-
ulus of spacer yarn and matrix (MPa), respectively, h is the orienta-
tion angle of spacer yarn along the loading direction as shown in
Fig. 15 (–), it can be expressed as Eq. (3).
Fig. 15. Simplified modeling of spacer yarn II of 3D spacer fabric.
In fact, a certain degree of symmetrical deflection exists as
shown in Fig. 15 because all spacer yarns bend slightly due to
the gravity of spacer yarns and outer textile substrates. Generally,
the orientation distribution of spacer yarn II as shown in Fig. 1
requires a three-dimensional description, and the arrangement
of spacer yarn II can be considered for simplicity of calculation
as shown in Fig. 15(a). In Fig. 15(a), the X-, Y- and Z-axes indicate
the direction of warp, weft and through-the-thickness of 3D
spacer fabric, respectively. Here, the projected orientation angle
of spacer yarn II in the X–Z plane along warp direction and Y–Z
plane along weft directions is directly considered, respectively,
as shown in Fig. 15(b and c). And it could be calculated as follow-
ing [23]:
h h
hwarp ¼ tan1 ; hweft ¼ tan1 ð3Þ
m w
where h is thickness of 3D spacer fabric (m), m and w are segment
width (m). In addition, the orientation angle of spacer yarn I of T20
along warp direction is identical with hwarp.
The reinforcing efficiency factors of various CCs in Table 6
reveals that T20-CCs presented better efficiency than N15-CCs in
each direction except N15-III-CCs and N15-IV-CCs in warp direc-
tion. And the efficiency of CCs in weft direction is generally better
than that in warp direction which suggests that the layout pattern
of spacer yarns is closely related to the reinforcing efficiency factor.
By comparing the reinforcing efficiency factors of N15-I-CCs,
N15-II-CCs and N15-III-CCs, a small decrease of reinforcing
efficiency factor is obtained with increasing the amount of spacer
yarn. Moreover, as the orientation angle of warp/weft yarns
decreases, the reinforcing efficiency factor increases very much.
ð1Þ This is attributed to the larger contribution of warp/weft yarns
strength with decreasing orientation angle along loading direction.
These results indicate that N15 with less amount of spacer yarn
and warp/weft yarns alignment along loading direction is a more
efficient reinforcement for CCs.
4. Conclusions
The tensile behavior of CCs and influences of the geometric pat-
terns of 3D spacer fabric on tensile behavior were investigated.
ð2Þ Five PET-based 3D spacer fabrics with different geometric patterns
were investigated. The tensile stress–strain curves and the crack
propagating patterns of CCs along the warp and weft directions
were obtained and the experimental results were analyzed. The
following conclusions could be drawn:
(I) Due to one outer textile substrate of N15 is solid fabric, the
tensile stress–strain curves of N15-CCs have an obvious
stage of solid fabric failure compared with T20-CCs. The ten-
sile behavior of N15-CCs is significantly determined by the
tensile performance of the solid fabric.
 F. Han et al. / Construction and Building Materials 65 (2014) 620–629
(II) The first crack strength of N15-CCs in weft direction is
higher than that in warp direction whereas the situation is
reversed when it comes to the first crack of T20-CCs. N15-
CCs produce multiple cracking in weft direction which is
probably attributed to the interaction between solid fabric
and spacer yarns. The degree of multiple cracking is closely
related to the tensile properties of solid fabric.
(III) The tensile behavior of N15-CCs is greatly influenced by
outer textile substrates, but it is not the case for T20-CCs.
3D spacer fabric is a more advisable reinforcement than
spacer yarns alone, especially for N15-CCs.
(IV) N15 is a more efficient reinforcement for CCs. And the effi-
ciency of CCs in weft direction is better than that in warp
direction.
(V) N15-CCs exhibit the improved ductility and strengthened
tensile strength when increasing the amount of spacer yarns,
however, a small decrease of efficiency factor is observed.
Furthermore, the N15 with warp/weft yarns alignment along
loading direction is a better choice in the light of tensile
strength, ductility and reinforcing efficiency factor.
N15-CCs exhibit better tensile behaviors in terms of tensile
strength, reinforcing efficiency factor, and crack pattern. It is also
easy to be produced (mixture cannot leak out of solid fabric) and
packed (only mesh fabric need to be packed). Therefore, 3D spacer
fabric with one solid outer textile substrate is a more advisable
embodiment and efficient reinforcement for CCs.
Acknowledgements
The authors greatly appreciate reviewer’s helpful comments in
improving the quality of our paper. In addition, the financial
support from Architectural Engineering Institute of the General
Logistic Department of P.L.A. is gratefully acknowledged.
References
[1] Isley F. The use of high performance textiles in construction projects. J Ind Text
2002;31(3):205–21.
[2] Tsesarsky M, Peled A, Katz A, Anteby I. Strengthening concrete elements by
confinement within textile reinforced concrete (TRC) shells – static and impact
properties. Constr Build Mater 2013;44:514–23.
629
[3] Hegger J, Voss S. Investigations on the bearing behaviour and application
potential of textile reinforced concrete. Eng Struct 2008;30(7):2050–6.
[4] Koeckritz U, Cherif C, Weiland S, Curbach M. In situ polymer coating of open
grid warp knitted fabrics for textile reinforced concrete application. J Ind Text
2010;40(2):157–69.
[5] Colombo IG, Magri A, Zani G, Colombo M, Prisco M. Textile reinforced concrete:
experimental investigation on design parameters. Mater Struct 2013;46(11):
1933–51.
[6] Peled A, Cohen Z, Pasder Y, Roye A, Gries T. Influences of textile characteristics
on the tensile properties of warp knitted cement based composites. Cem Concr
Compos 2008;30(3):174–83.
[7] Peled A. Strain hardening behavior of textile reinforced concrete (TRC). In:
Filho RDT, Silva FA, Koenders EAB, Fairbairn EMR, editors. Proceedings of
SHCC2-Rio conference. Bagneux: RILEM Publications; 2011. p. 45–52.
[8] Hartig J, Häußler-Combe U, Schicktanz K. Influence of bond properties on the
tensile behaviour of textile reinforced concrete. Cem Concr Compos
2008;30(10):898–906.
[9] Vassiliadis S, Kallivretaki A, Psilla N, Provatidis CH, Mecit D, Roye A. Numerical
modelling of the compressional behaviour of warp-knitted spacer fabric. Fibres
Text East Eur 2009;17(5):56–61.
[10] Armakan DM, Roye A. A study on the compression behavior of spacer fabrics
designed for concrete applications. Fiber Polym 2009;10(1):116–23.
[11] Gries T, Roye A, Offermann P, Peled A. Textile reinforced concrete-state-of-the-
art report of RILEM TC 201-TRC. Bagneux: RILEM Publications; 2006. p. 11–27.
[12] Roye A, Gries T. 3-D textiles for advanced cement based matrix reinforcement.
J Ind Text 2007;2(3):163–73.
[13] Mecit D, Roye A. Investigation of a testing method for compression behavior of
spacer fabrics designed for concrete applications. Text Res J 2009;79(10):
867–75.
[14] Naaman AE. Evolution in ferrocement and thin reinforced cementitious
composites. Arab J Sci Eng 2012;37(2):421–41.
[15] Zhu D, Mobasher B, Peled A. Experimental study of dynamic behavior of
cement-based composites. J Sustain Cem – Based Mater 2013;2(1):1–12.
[16] Concrete Canvas Ltd. Concrete canvas. <http://concretecanvas.de/index.html>.
22.7.12.
[17] Balea L, Dusserre G, Bernhart G. Mechanical behaviour of plain-knit reinforced
injected composites effect of inlay yarns and fibre type. Composites Part B
2014;56:20–9.
[18] Li VC, Leung KY. Steady-state and multiple cracking of short random fiber
composites. ASCE J Eng Mech 1992;11(118):2246–64.
[19] Kebele P. Analytical model of multiple cracking in fiber reinforced
cementitious composites under uniaxial tension. In: Wu ZS, Abe M, editors.
Proceedings of the first international conference on structural health
monitoring and intelligent infrastructure. Tokyo: Tylor & Fracis; 2003. p.
791–8.
[20] Li VC, Wang S, Wu C. Tensile strain-hardening behavior of polyvinyl alcohol
engineered cementitious composite (PVC-ECC). ACI Mater J 2001;98(6):492–3.
[21] Akhavan A, Shafaatian SMH, Rajabipour F. Quantifying the effects of crack
width, tortuosity, and roughness on water permeability of cracked mortars.
Cem. Concr. Res 2012;42(2):313–20.
[22] Li VC. Postcrack scaling relations for fiber reinforced cementitious composites.
J Mater Civil Eng 1992;4(1):41–57.
[23] Yip J, Ng SP. Study of three-dimensional spacer fabrics: physical and
mechanical properties. J Mater Process Tech 2008;206(1/3):359–64.