Professional Documents
Culture Documents
5 dry shrinkage ضغط شدnn PDF
5 dry shrinkage ضغط شدnn PDF
2014–10:57am] [1–20]
//blrnas3.glyph.com/cenpro/ApplicationFiles/Journals/SAGE/3B2/JITJ/Vol00000/140081/APPFile/SG-
JITJ140081.3d (JIT) [PREPRINTER stage]
Article
0(00) 1–20
! The Author(s) 2014
of concrete canvas
Abstract
Concrete canvas has taken attention for their rapidly deployable hardened character-
istic property in civil engineering. However, the drying shrinkage of concrete canvas has
not been addressed yet in the literatures. In this study, a theoretical model was pre-
sented for studying influences of 3D spacer fabric on drying shrinkage of concrete
canvas. The model was based on assumption that drying shrinkage restraint provided
by 3D spacer fabric is joint action of each component of 3D spacer fabric separately.
To calibrate this model, the drying shrinkage of two concrete canvases reinforced by
PET-based 3D spacer fabric with one solid outer textile substrate was experimented.
Moreover, a simplified expression of maximum tensile stress generated in the matrix of
both concrete canvases was obtained for evaluating their risk of drying shrinkage-
induced cracking. The results showed that drying shrinkage strain of concrete canvas
samples became lower due to the restraint provided by 3D spacer fabric and a satis-
factory correlation between model predictions and experimental results was found at
later age. For both concrete canvases, a greater restraint was found in warp direction,
thereby resulting in a larger tensile stress generated in the matrix. Furthermore, the
restraint on the drying shrinkage of concrete canvas was provided mostly by spacer
yarns and thereby it contributed to the most of maximum tensile stress generated in
the matrix of concrete canvas.
Keywords
3D spacer fabric, concrete canvas, drying shrinkage model, experimental verification,
risk evaluation
1
Jiangsu Key Laboratory of Construction Materials, School of Materials Science & Engineering, Southeast
University, Nanjing, China
2
Architectural Engineering Institute of the General Logistic Department, P.L.A., Xi‘an, China
Corresponding author:
Huisu Chen, Jiangsu Key Laboratory of Construction Materials, School of Materials Science & Engineering,
Southeast University, Nanjing 211189, China.
Email: chenhs@seu.edu.cn
Introduction
Recently, an increasing interest occurred in the use of 3D spacer fabric as novel
component reinforcement for cement-based composite materials [1–3]. 3D spacer
fabric, composed by two components of outer textile substrates and spacer yarns, is
a special type of textile with 3D yarn architecture and 3D textile architecture [4].
Due to the beneficial structure of an armature system of highly oriented yarns, 3D
spacer fabric has been shown to exhibit a superior behavior as it can be easily
placed in a mold and impregnated to its whole depth by cement-based matrix in
one step [5]. Meanwhile, 3D spacer fabric can be used for producing concrete
elements with reinforcing yarns in three directions and defined positioning of the
two outer substrates [4]. 3D spacer fabric has been applied successfully in civil
engineering, such as prefabricated wall panels and shell concrete members [1,2].
In addition, a more attractive type of cementitious composite, i.e. concrete
canvas (CC), is manufactured through 3D spacer fabric [6–8]. The significant dif-
ference between CC and conventional 3D spacer fabric reinforced cementitious
composites is the preparation process. Before set of the cement powder, CC is a
flexible 3D spacer fabric impregnated with cement powder that can cover the sur-
face of arbitrary structure or element like a soft cloth, whatever its shape. Then,
one just needs to spray or add water from the top surface of the CC for construc-
tion. Once set, a thin composite layer, that is durable, water-resistant and fire
resistant, forms. Han’s work [7,8] demonstrated that compared with the matrix
without fabric, 3D spacer fabric provided an efficient reinforcement on both com-
pressive and tensile behavior of the CC. Thus, CC can be quickly, efficiently and
widely used as a structural element to form a cover for prefabricated shelters, paths
for vehicles or for pedestrians, or a protective layer for pipes, linings, etc. [6].
It is well known that cracking in cementitious composites significantly influences
their service life. In addition to the action of external load, the main reason for
initial crack formation in the cementitious composites structures is partially attrib-
uted to its drying shrinkage [9,10], which is particularly significant in the case of
thin-wall slab element. It is also recognized that fibers in cementitious composites
provide significant restraint on their drying shrinkage, which has been comprehen-
sively studied in the literatures [11–13]. In fact, the restraint provided by fibers is
mainly realized by bonding strength through the contact interface between fibers
and matrix [14,15]. Once cement matrix undergoes shrinkage, fibers are subjected
to compressive stress whereas cement matrix is subjected to tensile stress due to the
shear stress produced on their contact interface. In the case of CC, both spacer
yarns and outer textile substrates of the 3D spacer fabric can provide restraint on
its drying shrinkage, and the influence of 3D spacer fabric should be investigated
thoroughly. However, to the author’s current knowledge, the drying shrinkage of
CC, one of most concentrated factors, has not been addressed yet.
This study aimed to investigate the influence of 3D spacer fabric on the drying
shrinkage of CC. A theoretical model of the drying shrinkage of CC was presented.
The drying shrinkage of two CCs reinforced by PET-based 3D spacer fabric with
one solid outer textile substrate was tested to verify the rationality of this model.
Han et al. 3
The initial cracking point strain of CC subjected to uniaxial static strain was
around 1.2*102–3.9*102. The previous works showed that the magnitude of
drying shrinkage strain of CC was lower than 1*103 [16]. It is lower than the
initial cracking strain. In other words, the inter-anchorage effect of spacer yarns
with outer textile substrates was not reflected yet during the drying shrinkage
period [7]. Therefore, it is assumed that the restraint provided by 3D spacer
fabric on the drying shrinkage of CC came from its two components separately,
i.e. outer textiles substrates and spacer yarns, as shown in Figure 1. As it was a
thin-wall and anisotropic element, this study mainly focused on the drying shrink-
age of CC in the warp and weft directions.
If only the spacer yarns was focused, a certain degree of symmetrical deflection
of spacer yarns existed in the warp direction, as depicted in Figure 1(a), due to the
gravity effect of spacer yarns and outer textile substrates. Generally, the orientation
distribution of spacer yarns required a three-dimensional description as shown in
Figure 2(a). Due to the symmetrical structure, the ‘‘V-shape’’ spacer yarns can be
simplicity as a straight fiber by two steps as shown in Figure 2(a) for the sake of
simplicity of calculation [17]. In Figure 2(a), the X-, Y- and Z-axes indicated the
direction of warp, weft and through-the-thickness of 3D spacer fabric, respectively.
Here, as shown in Figure 2(b) and (c), the projected orientation angle and projected
length of spacer yarns in the X-Z plane along warp direction and Y-Z plane along
weft direction were considered directly. The projected orientation can be calculated
as following
h
¼ sin1 ð1Þ
Lfs
Han et al. 5
Figure 3. Schematic diagram of restraint provided by spacer yarns. (a) Equivalent element
resulted from transformation of spacer yarns and (b) enlarged representative cylinder.
interfacial bond strength [14]. Therefore, it was assumed that the restraint provided
by spacer yarns with projected length Lfs and volume content Vfs in CC was
equivalent to the restraint in an idealized composite with aligned spacer yarn
distribution, effective spacer yarn length Lfse and spacer yarn spacing 2Rs, as
given in Figure 3(a) [18]. The effective spacer yarn length Lfse of spacer yarns
along warp or weft directions was given by
where is orientation angle of spacer yarns along warp or weft directions [–].
The spacer yarn spacing 2Rs could be solved through the representative cylinder
with radius Rs and length (Lfse þ 2Rs) isolated from CC, as shown in Figure 3(b).
Based on the assumption that the total volume of CC was equal to the sum of all
individual cylinder volumes, the spacer yarn spacing 2Rs along shrinkage direction
could be calculated as following
2Rs r2fs
R2s 1 þ ¼ ð3Þ
Lfs cos Vfs cos
where rfs is spacer yarn radius (mm). Equation (3) provided the general relation
between spacer yarn spacing, spacer yarn volume content, length, radius and orien-
tation angle.
Thus, the restraint provided by spacer yarns on the drying shrinkage of CC
could be calculated through the representative volume element with a single
spacer yarn in center, as shown in Figure 3(b) [18]. In Figure 3(b), the shrinkage
of cylinder cement matrix adjacent to the spacer yarn was considered to be influ-
enced by the spacer yarn, whereas the left and the right parts of length Rs each were
assumed to be free of its influence because the contact area at the tip of fiber is too
small for the matrix to transfer any significant force in the fiber. Moreover, the
tensile stress exerted on cement matrix was a function of location within restrain
zone. Accordingly, the expression of free drying shrinkage of CC under the work of
spacer yarns along shrinkage direction was given by [18,19]
Z Lfse
1 mas
"ts ¼ "m dx ð4Þ
Lfse þ 2Rs 0 Em
where "ts and "m are drying shrinkage strain of CC under the work of spacer yarns
and pure matrix [–], Em is tensile elastic modulus of matrix (MPa), mas is tensile
stress generated in cement matrix due to the work of spacer yarns (MPa).
According to Zhang et al. [18,19], the tensile stress mas could be solved from
shear-lag theory based on several simplifying assumptions: (1) the matrix and
reinforcing spacer yarns are both elastic materials; (2) the interface between the
matrix and spacer yarns is infinitesimally thin; (3) there is no slip between the
spacer yarns and the matrix at the interface; (4) the shrinkage strain in matrix
"m, at a distance Rs from the reinforcement, is equal to the free shrinkage of the
matrix. Therefore, the tensile stress mas was given by
2 3
L 2x
Afs cosh s fse2rfs
mas ¼ Efs "m 41 5 ð5Þ
Am L
cosh s fse 2rfs
where Afs and Am are cross-sectional area of reinforcing spacer yarn and matrix
(mm2), respectively, Efs is elastic modulus of spacer yarn (MPa). The first negative
sign in equation (5) indicated that the matrix stress is opposite to that exerted on
the spacer yarns. The parameter s was given as
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Em
s ¼ ð6Þ
ð1 þ vm ÞEfs log Rs =rfs
Vfs cos 1
"ts ¼ "m 1 s 1 tanhðs s Þ ð7Þ
1 Vfs cos ð1 þ 1=s1 Þ s s
where
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Efs 1 Lfs cos
s ¼ ; s ¼ ; s ¼ ;
Em ð1 þ vm Þs log s2 2rfs
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð8Þ
Lfs cos 1
s1 ¼ ; s2 ¼
2Rs Vfs cos ð1 þ 1=s1 Þ
Han et al. 7
where subscript w stand for the warp/weft yarns of MF. ’ is orientation angle of
warp/weft yarns of MF along warp or weft directions [–].
Equation (9) provided a general relationship between the drying shrinkage
strains of CC along warp or weft directions under the interaction of 3D spacer
fabric with pure matrix. Equation (9) revealed that the drying shrinkage strain of
CC is less than that of pure matrix and it is influenced by the properties of spacer
yarn (including spacer yarn length, radius, volume content, orientation angle and
elastic modulus), the warp/weft yarn (including warp/weft yarn length, radius,
volume content, orientation angle and elastic modulus) and matrix (including the
drying shrinkage behavior and the corresponding elastic modulus).
As the hydration goes on, the matrix of CC was subjected to tensile stress due to
the restraint provided by 3D spacer fabric. As discussed above, the total tensile
stress generated in the matrix of CC can be calculated by equation (10)
8 2 3 2 39
Lfse 2xs Lfwe 2xw
"m < cosh s 2rfs cosh w 2rfw =
ma ¼ Afs Efs 41 5 þ Afw Efw 41 5
Am : cosh fs
L
cosh fw
L ;
s 2rfs w 2rfw
ð10Þ
It was found from equation (10) that time-dependent tensile stress generated in
matrix is closely related to the drying shrinkage of pure matrix, as well as to the
restraint provided by 3D spacer fabric. In addition, it also should be noted that the
drying shrinkage-induced maximum tensile stress generated in matrix is critical for
the crack ignition and propagation in the CC slab when comparing to the ultimate
tensile strength of matrix.
Experimental verification
To verify the rationality of the model given in equation (9), the drying shrinkage
experiments of two CC reinforced by PET-based 3D spacer fabric with one solid
outer textile substrate and pure matrix were conducted. Also, the tensile elastic
modulus of pure matrix was tested as one input parameter.
Raw materials
As shown in Table 1, two types of 3D spacer fabrics with different geometric
patterns, labeled as N1 and N2 were investigated. For both 3D spacer fabrics,
the thickness was 15 mm and the architecture of SF was tricot. The warp/weft
yarn of SF and MF were made by 396 dtex and 339 dtex PET multifilaments,
respectively. The spacer yarn was made by 495 dtex PET monofilament.
However, the pore shape of MF was rectangle for N1 and rhombus for N2. The
properties of the components of both 3D spacer fabrics, as determined by a yarn
tensile tester with 30 N capacity (XL-2), are identical, and summarized in Table 2.
As shown in Table 3, the main differences of both 3D spacer fabrics were orienta-
tion angles as well as projected lengths of spacer yarns and warp/weft yarns.
Raw materials for the matrix used in this study were anhydrite and calcium
sulphoaluminate cement (CSA). The mineralogical and chemical compositions of
the CSA cement and the anhydrite were determined by X-ray fluorescence
(ThermoFisher Scientific ARL QUANTX) and X-ray diffraction (Bruker-AXS
Discover 8), respectively. The results are given in Table 4. The detail on the opti-
mized process of matrix formula can be found in Ref. [8], the final optimized mix
proportion of matrix used in this paper is shown in Table 5, in which the anhydrite
was 20% of the CSA by weight and water-to-binder ratio was fixed at 0.45. The
selected water to binder ratio was calculated based on the mass of the sample
before and after water-spraying in preliminary exploration work, where water-
spraying stopped immediately if the water slightly penetrated through the
bottom of 3D spacer fabric.
Han et al. 9
N1
N2
3D
3D
spacer Efs Efw Vfs Vfw Lfs Lfw rfs rfw vm
fabrics (GPa) (GPa) (%) (%) (–) ’ (–) (mm) (mm) (mm) (mm) (–)
N1
Warp 15.91 10.54 2.91 0.35/0.35a 0.28p 0/0.5pb 19.47 160.00/ 0.09 1.08 0.23
40.00c
a b
Weft 15.91 10.54 2.91 0.35/0.35 0.30p 0.5p/0 18.54 40.00/ 0.09 1.08 0.23
160.00c
N2
Warp 15.91 10.54 2.91 0.35/0.35a 0.37p 0.17p/ 16.34 78.58/ 0.09 1.08 0.23
0.17pb 78.58c
a
Weft 15.91 10.54 2.91 0.35/0.35 0.39p 0.33p/ 15.94 46.47/ 0.09 1.08 0.23
0.33pb 46.47c
a
Left and right side of the slash are volume fraction of warp yarns and weft yarns of MF, respectively.
b
Left and right side of the slash are orientation angle of warp yarns and weft yarns of MF, respectively.
c
Left and right side of the slash are length of warp yarns and weft yarns of MF, respectively.
The drying shrinkage of control samples and CC samples in warp and weft
directions were measured using contraction rack and dial gauge as shown in
Figure 4, according to Chinese specification JC/T 603-2004 [20]. The initial
length of the samples was measured at 3 hours after water-spraying and regular
length measurements were made thereafter up to 90 days age. After different curing
ages (3 d, 7 d, 14 d, 28 d, 90 d), the tensile elastic modulus of pure matrix were tested
by an uniaxial tensile test machine (CMT 5105) equipped with a 100 kN load cell
according to Chinese national standard GB/T 50081-2002 [21], as shown in
Figure 5. The strain was acquired by a dynamic data acquisition instrument
(TMR-7200). Three parallel samples were tested for each point. The average
values and the corresponding standard deviation were presented. The detail on
the influence of geometric pattern of 3D spacer fabric on tensile behavior of CC
can be found in Ref. [7].
Han et al. 11
Experimental results
The drying shrinkage strain versus curing age curves of CC and control samples are
shown in Figure 6. In all cases, the drying shrinkage strain increases with curing age
and develops fast within 10 curing days. After 30 curing days, the drying shrinkage
curves are prone to flatten. This is consistent with the trend of compressive
strengths development, as shown in Figure 7 [7,8]. Compared to the control sam-
ples, the drying shrinkage strain of CC samples becomes lower due to the restraint
provided by 3D spacer fabric. In addition, the drying shrinkage of CC samples in
warp direction is less than that in weft direction. It revealed that the restraint
provided by 3D spacer fabric in warp direction is stronger than that in weft direc-
tion, which is consistent with the result of mechanical anisotropy of CC samples [8].
This is attributed to the layout pattern of spacer yarn as shown in Table 1; the
orientation angle of spacer yarn in the warp direction is lower than that in the weft
direction after transforming as shown in Table 3. Consequently, a greater restraint
was obtained with a lower orientation angle of spacer yarn as described in equation
(9). In each direction, the drying shrinkage strain of N1-CCs is less than that of
N2-CCs, where the mostly impact factors shown in equation (9) are also the pro-
jected orientation angles and projected length of spacer yarn and warp/weft yarn.
Due to the lower projected orientation angle and larger projected length of spacer
yarn and warp/weft yarn as presented in Table 3, a greater restraint was obtained
for N1-CCs and then resulting in a less drying shrinkage strain. Figure 8 shows the
tensile elastic modulus of pure matrix at different curing ages. Similar to drying
shrinkage of CCs, it also increases with curing age and develops from rapidly to
gently after 10 days.
As the drying shrinkage strain "m and tensile elastic modulus Em of pure matrix
were two key input parameters for the presented drying shrinkage model, it should
obtain the expressions of "m and Em, and they could be acquired by fitting the
experimental data. The famous models of drying shrinkage of cement matrix in the
literatures included CEB-FIP model, ACI model and BP model [22]. The CEB-FIP
model, as shown in equation (11), is adopted in numerous concrete codes across the
world and is valid for concrete with a compressive strength between 12 and
80 MPa. The BP model, as shown in equation (12), is semi-empirical but the
most powerful one which considers almost every parameter affecting the shrinkage
behavior. The ACI model, as shown in equation (13), was one of the most simple
and the most accurate in predicting time-dependent strains [23]
where "s is the time-dependent drying shrinkage [–], t is the duration of drying
shrinkage (days), "so is a parameter depending on the type of cement, the mean
Han et al. 13
compressive strength of concrete at the age of initial shrinkage and relative humid-
ity [–], s is function of time [–]
where "su is the ultimate shrinkage [–], Kh is a parameter concerning the humidity of
the environment [–], Q is function of time [–]
ta
"s ðtÞ ¼ "su ð13Þ
f þ ta
where f and a are considered constants for a given member shape and size which
define the time-ratio part.
Han et al. 15
Therefore, the ACI model was chosen to fit the experimental data. Referring to
the ACI 209R-92 shrinkage model [24], the fitting result of "m was expressed as
t
"m ¼ 841:365 ð14Þ
t þ 1:894
t
Em ¼ 19:739 ð15Þ
t þ 1:828
Model verification
Substituting the relevant parameters in Table 3, the expressions of free drying
shrinkage strain of control samples, and tensile elastic modulus of pure matrix
into the equation (9), would theoretically obtain the drying shrinkage strain of
CCs in the warp and weft directions. The comparisons between model predictions
and experimental results are shown in Figure 9. It can be seen that theoretical
predictions are in satisfactory agreement with experimental results at later age,
except the weft direction of N1-CCs. However, in the weft direction of N1-CCs,
a good agreement between theoretical prediction and experimental result is found
when ignoring the restraint contributed from warp/weft yarns, as shown in
Figure 9(a). It revealed that the warp/weft yarns of MF do not provide restraint
on the drying shrinkage of N1-CCs in the weft direction. The reason may be
Figure 9. Comparisons of model predictions with experimental results of CCs. (a) N1-CC and
(b) N2-CC. The double arrow lines and corresponding values represent the fraction of restraint
provided by spacer yarns.
attributed to the form of weft yarns. As shown in Table 1, the weft yarns of MF of
N1 are in straight form rather than twisted form. Generally, the warp/weft yarns in
straight form mainly improved the bonding effect between outer textile substrates
and cement matrix. This made their rigidity rather low when they are subjected to
compressive stress [25]. By comparing the contribution of the restraint provided by
each component of 3D spacer fabric to the drying shrinkage of both CC samples,
as shown in Figure 9, it can be seen that the restraint on their drying shrinkage is
provided mostly by spacer yarns. This is attributed to its larger volume faction and
higher Young’s modulus.
Han et al. 17
deriving the tensile elastic modulus of matrix, as shown in equation (15), the time-
dependent ultimate tensile strength of matrix can be expressed as
t
mu ¼ 1:132 ð17Þ
t þ 2:119
Afs Afw
ma, max ¼ "m Efs ð1 KVs Þ þ Efw ð1 KVw Þ ð18Þ
Am Am
where
1 1
KVs ¼ ; KVw ¼ ð19Þ
L L
cosh s 2rfesfs cosh w 2rfew
fw
Here, KVs and KVt are two influence factors depending on yarn properties, which
are closely related to the volume fraction, length, radius and orientation angle of
yarns. For the components of 3D spacer fabric in this study, the contribution
of these two influence factors to the ma,max can be negligible, as demonstrated
in Figure 10, which are applied with components parameters of N2.
where
Figure 11. Analysis on risk of drying shrinkage-induced cracking of CCs. The double arrow
lines and corresponding values represent the fraction of maximum tensile stress contributed
by spacer yarns.
Han et al. 19
Conclusions
An analytical model is presented for studying the influence of 3D spacer fabric on
drying shrinkage of CC samples. A satisfactory correlation between the model
predictions and experimental results is found at later age. Due to the introduction
of 3D spacer fabric, the drying shrinkage strain of CC samples becomes lower than
the control samples. A greater restraint is found in warp direction due to the
smaller orientation angle. The restraint on the drying shrinkage of both CC sam-
ples is provided mostly by spacer yarns, which is attributed to its larger volume
faction and higher Young’s modulus. Furthermore, a simplified expression of
drying shrinkage-induced maximum tensile stress generated in the matrix is derived
for evaluating the risk of cracking for both CC samples. A larger drying shrinkage-
induced maximum tensile stress generated in matrix is found in warp direction due
to the greater restraint. Moreover, spacer yarns contribute to the most of maximum
tensile stress generated in the matrix of CCs.
Acknowledgement
The authors gratefully acknowledge financial support from the Architectural Engineering
Institute of the General Logistic Department of P.L.A.
Funding
This research received no specific grant from any funding agency in the public,
commercial, or not-for-profit sectors.
References
[1] Gries T, Roye A, Offermann P, et al. Textile reinforced concrete-state-of-the-art report
of RILEM TC 201-TRC. Report, Aachen, France, Autumn, 2006.
[2] Roye A and Gries T. 3-D textiles for advanced cement based matrix reinforcement.
J Ind Text 2007; 2: 163–173.
[3] Mecit D and Roye A. Investigation of a testing method for compression bahavior of
spacer fabrics designed for concrete applications. Text Res J 2009; 79: 867–875.
[4] Armakan DM and Roye A. A study on the compression behavior of spacer fabrics
designed for concrete applications. Fiber Polym 2009; 10: 116–123.
[5] Vassiliadis S, Kallivretaki A, Psilla N, et al. Numerical modelling of the compressional
behaviour of warp-knitted spacer fabric. Fibres Text East Eur 2009; 17: 56–61.