Journal of The Textile Institute
Publicat ion det ails, including inst ruct ions for aut hors and subscript ion informat ion:
ht t p:/ / www.t andfonline.com/ loi/ t j t i20
Study of changes in 3D-woven multilayer interlock
fabric preforms while forming
a
b
Y. Nawab , X. Legrand & V. Koncar
b
a
Inst it ut de Recherche en Génie Civil et Mécanique, Universit é de Nant es, Saint -Nazaire,
France
b
ENSAIT, GEMTEX, Roubaix, France
Available online: 10 Apr 2012
To cite this article: Y. Nawab, X. Legrand & V. Koncar (2012): St udy of changes in 3D-woven mult ilayer int erlock fabric
preforms while forming, Journal of The Text ile Inst it ut e, DOI:10.1080/ 00405000.2012.676267
To link to this article: ht t p:/ / dx.doi.org/ 10.1080/ 00405000.2012.676267
PLEASE SCROLL DOWN FOR ARTI CLE
Full t erm s and condit ions of use: ht t p: / / www.t andfonline.com / page/ t erm s- and- condit ions
This art icle m ay be used for research, t eaching, and privat e st udy purposes. Any subst ant ial or syst em at ic
reproduct ion, redist ribut ion, reselling, loan, sub- licensing, syst em at ic supply, or dist ribut ion in any form t o
anyone is expressly forbidden.
The publisher does not give any warrant y express or im plied or m ake any represent at ion t hat t he cont ent s
will be com plet e or accurat e or up t o dat e. The accuracy of any inst ruct ions, form ulae, and drug doses should
be independent ly verified wit h prim ary sources. The publisher shall not be liable for any loss, act ions, claim s,
proceedings, dem and, or cost s or dam ages what soever or howsoever caused arising direct ly or indirect ly in
connect ion wit h or arising out of t he use of t his m at erial.
The Journal of The Textile Institute
iFirst, 2012, 1–7
Study of changes in 3D-woven multilayer interlock fabric preforms while forming
Y. Nawaba*, X. Legrandb and V. Koncarb
a
Institut de Recherche en Génie Civil et Mécanique, Université de Nantes, Saint-Nazaire, France; bENSAIT, GEMTEX,
Roubaix, France
(Received 4 October 2011; final version received 12 March 2012)
Multilayer woven reinforcements are increasingly employed in the domain of composite materials. Delamination
occurrence and resultant failure of a laminated composite piece subjected to high vibrations, is an issue of much
concern in aeronautics. The situation becomes more complex in case of bended/curved pieces. In order to improve
through the thickness mechanical properties, 3D-woven multilayer interlock fabric is used as composite reinforcement. Structural changes, i.e. thickness change, relative slippage of layers, change in tow aspect ratio and change
of orientation of the tows columns, etc. occur in such fabrics during the forming process. These changes may lead
to the gradient of the resin amount in composite, internal stresses and variations of mechanical properties in the
piece. No significant research has been conducted on this aspect. Lack of knowledge or neglecting these changes
may lead to prejudicial estimations of ultimate mechanical properties and fracture analysis. In the present article,
the changes that occurred in 5-layer and 13-layer 3D-woven multilayer interlock fabrics have been studied, when
moulded at five different angles and two different bending radii. A significant change in thickness, tow aspect
ratio, tow orientation and relative layer slippage is observed.
Keywords: textile composite; multilayer interlock fabric; mouldability; tow orientation
Introduction
Laminated woven composites are widely used in the
aeronautical, naval and automotive applications due to
their uniform in-plane mechanical and thermophysical
properties but their through the thickness properties
are poor (Gowayed & Fan, 2001; Miravete, 1999).
Delamination of plies occurs under high vibrations,
especially, in the aeronautics, resulting into the piece
failure. The situation becomes more complex, if such
pieces are curved/angled. In order to improve the
interlayer fracture resistance, and through the thickness properties of composites, 2D+ or multilayer interlock woven preform are used. Multilayer interlock
preforms can be fabricated on modified conventional
looms, but for certain 3D-woven fabrics, special
machines are required. In order to weave multilayer
interlock preforms, weaving is done such that the
numbers of threads from different fabric layers are
used to bind the layers with each other.
Multilayer interlock woven fabrics are categorized
into four types (Gu & Zhili, 2002; Sheng & van Hoa,
2003) depending on the orientation of binding threads
among the layers, and their penetration depth into the
layers. Layer-to-layer binding may have more fibre
*Corresponding author. Email: yasir.nawab@univ-nantes.fr
ISSN 0040-5000 print/ISSN 1754-2340 online
Copyright Ó 2012 The Textile Institute
http://dx.doi.org/10.1080/00405000.2012.676267
http://www.tandfonline.com
volume fraction than through the thickness binding.
Moreover, angle interlock bound fabrics are more
resistant to delamination than orthogonal interlocked
fabrics (Potluri & Sagar, 2008; Yi & Ding, 2004).
These preforms are impregnated with resin inside
a mould to convert them into a composite. The shape
and complexity of the composite may vary as there
are various numbers of ways, in which a preform
may be moulded. There is a real need to understand
the moulding process and the ability of a particular
woven preform to adopt to the shape of the mould,
especially considering the variety of ways in which
preforms can be formed. Since there is no standard
method to measure mouldability, it has no standard
definition. However, some authors described it in the
terms of shear rigidity (Chen, Lo, Tayyar, & Day,
2002). It can also be defined as “the ability of a preform to adopt to the shape of a mould during the
composite formation”. When composites are formed
from complex multilayer interlock woven reinforcements, certain changes in preform geometry are
observed. These changes are directly related to
“mouldability” of the preform. Since a textile material
is essentially hierarchical in nature (Lomov, Gusakov,
Huysmans, Prodromou, & Verpoest, 2000; Lomov
2
Y. Nawab et al.
et al., 2001; Long, 2005; Verpoest & Lomov, 2005),
the changes that are visible at macro scale are always
accompanied by the changes that occur at the lower
scales. Thus, these hierarchical changes can be
observed at the micro (change in fibre/filament distribution), meso (Boisse, Zouari, & Gasser, 2005)
(change in tow-aspect ratio) and the macro scales
(change in thickness of the preform). These changes
depend on thickness of perform, i.e. number of layers, moulding angle, moulding geometry (moulding
radius) and tool-part interaction (which sides of the
preform are in contact with the mould). These
changes certainly affect the mechanical properties of
the final composite piece and should be taken into
account while modelling; otherwise they may lead to
the prejudicial estimations. Thus, mouldability can be
characterized in terms of changes that occur in the
preform at micro, meso and macro scale, when it is
impregnated to form a composite.
In the present study, multilayer woven preforms
were moulded at different angles and radiuses and
then impregnated in resin in order to fix the preform
and the tows in the moulded shape. Then they were
cut into fine slices, polished and analysed with a
microscope to observe the changes in the preform at
microscopic, and mesoscopic scale. It was observed
that changes that occur at micro and meso scales –
changes related to filament distribution and tow geometry – have direct correlation with the macro scale
changes, i.e. angle of bending, and thickness of the
preform. Thus identification of the mouldability can
be carried out by employing a statistical analysis at
the different hierarchical scales of the fabric. It is
found that a thick and dense multilayer fabric has a
uniform surface as compared to low density and thin
fabric. Tow-aspect ratio is an indication of compression on the tow layers. For an angled moulded fabric,
tow-aspect ratio of external layers was found greater
than the tow-aspect ratio at internal layers. Thickness
of the fabric also changes on moulding and its value
depends on the moulding angle, moulding radius and
the distance from bending point. There is also an
interlayer tow slippage, which depends on the
coupling of compressive and axial stresses.
Materials and methods
Multilayer orthogonal layer-to-layer fabrics with 5- and
13-layers were woven on a purpose-built shuttle weaving machine. Weave design of 5-layer fabric is shown
in Figure 1. Weave design of 13-layer fabric can be
obtained by extending 5-layer design on the same
pattern. Dobby shedding mechanism was installed on
the machine for separation of warp threads during
weaving.
Warp tows were coming directly to the machine
from a V-type creel, on which carbon warp bobbins
were mounted. Carbon tows of 200 Tex, 6 K in the
warp and the glass tows “RC14 320 P109” in the weft
were used. Reed filling was done so that the warp columns were perpendicular to the weft columns, i.e. for
5-layer fabric, number of tows per dent were kept
equal to 4, and for 13-layer fabric, equal to 12. Warp
and weft-tow densities of 5-layer fabric were found
equal to 16 tows/cm and 14 tows/cm, respectively. In
case of the 13-layer fabric, warp and weft densities
were found equal to 24 tows/cm, and 130 tows/cm,
respectively. Width of the woven fabric in both cases
was kept equal to 50 cm.
Fabric was cut into samples of 7 1 cm, both in
the direction of warp and weft, from different places
of the fabric. For the fabric moulding, aluminium
moulds bended at angles 0°, 45°, 90°, 135° and 180°
were used. The samples were moulded manually at
these angles, and at two radii of curvatures: R1 (8 mm)
and R2 (12.5 mm). These fabric samples were impregnated with transparent low viscosity epoxy resin in a
silicon batch. Objective was to block the fabric in
moulded position to observe the changes due to the
forming process.
Figure 2 shows a moulded (angle 90° and radius
12.5 mm) fabric sample after the curing. The resin
used in this study, polymerize at room temperature,
and have negligible chemical shrinkage. Therefore,
Figure 1. Weave design of 5-layer multilayer interlock
fabric with orthogonal layer–layer binding (Texgen).
Figure 2. A fabric sample (90°) moulded at radius
12.5 mm after curing.
The Journal of The Textile Institute
cure-induced changes in the preform, i.e. cure
shrinkage and residual deformation, etc. are negligible.
In order to avoid any changes due to pressure, resin
infusion was done at the atmospheric pressure. So the
complete fabric impregnation was only made possible
by the low viscosity of resin.
The composite pieces were cut in thin slices and
then polished by using the polish papers on polishing
machine. Final polishing was done on a gold plate
“Cameo® Disk Gold” by using diamond liquid. These
thin slices were then viewed under Bel® photonic
microscope for the measurements. For maximum visibility, the images are taken in patches at the same
zoom level without displacing the sample, and then
combined to get the full image of piece. In order to
make the study more precise, regions vertex (C) ± 1
cm, and vertex (C) ± 2 cm (Figure 3) were marked on
the moulded samples. Straight fabric sample herein
designated as 0° was taken as the reference for the
changes.
Results and discussion
In the following sections, results on the measurement
of thickness of the fabric, tow-aspect ratios of the layers, interlayer slippage and orientation of the tow columns are presented. Measurements were taken on five
different samples for each thickness, angle and radius
to verify the repeatability of the results. Standard deviations of these measurements were also calculated and
are shown on their respective graphs.
Thickness of the fabric
To achieve uniform mechanical characteristics,
thickness of the preform should be uniform. Since the
fabric is the combination of two groups of wavy structures (warp and weft yarns), resultant fabric structure
has troughs and crests (Figure 4). This non-
Figure 3. Sample of the polished fabric piece with marks
C ± 1 and C ± 2 cm.
3
Figure 4. Weft-wise cross-sectional view of the 5-layer
fabric.
uniformity may result in resin-rich areas in the region
of troughs at the surface causing poor surface mechanical characteristics compared to laminated composites
having relatively uniform tow arrangements.
In this study, it was found that the thickness of the
multilayer fabrics is not homogeneous. It is maximum
at the points of intersection of the warp and weft tows
(crest) and minimum at the centre in between them
(trough). The difference between maximum and minimum thickness of fabric depends on tow properties,
tow density, number of layers in fabric and weave
design. For our weave design, a difference of 43% (of
minimum thickness) was observed between maximum
and minimum thickness of 5-layer fabric, whereas in
case of 13-layer fabric it was just 5%. Thus, it can be
concluded that high density and thicker fabrics have
more uniform preform structure as compared to low
density and thinner fabrics. So an extreme care should
be taken while selecting multilayer fabric as composite
reinforcement.
Tow-aspect ratio
Aspect ratio of the tows is “the ratio of distance
between boundaries of tow along major axis to the
distance along minor axis”. If this ratio is higher, then
the tows will be more lenticular. This ratio can also be
used as an index of compression on the tows. If a
layer is subjected to compression, the aspect ratio of
the tows in that layer will be increased, and vice
versa.
In this study, it is noted that on moulding, the
aspect ratio of tows of external layers of fabric is
more than the ones of internal layers. It may be concluded that compression of tows at external fabric layers is more than compression of tows at internal
layers. This difference may result into gradients of
resin penetration through the thickness, internal stresses in piece and unstable final shape.
Figure 5 shows variation of the aspect ratio of layers with moulding angle. Tow-aspect ratio is uniform
at all the layers of 0° fabric. However, for angled
geometries, it increases with the increasing angle.
Therefore, it can be said that compression increases
with increasing moulding angle, and reaches the
maximum value at 135°. It is also found that the
4
Y. Nawab et al.
8
0°
18
7
135° R1
6
45° R1
5
180°
14
12
Slippage (%)
Aspect ratio
16
90°R1
4
3
2
1
0
10
-1
-2
8
0
1st outside
2nd
3rd
4th
30
5th inside
Fabric layers
Figure 5. Graph showing aspect ratios of layers at
different angles.
compression on the tows is lesser in case of moulding
at greater radius of curvature, and vice versa.
Interlayer slippage
On moulding, the fabric is submitted to the compressive
and axial stresses. The layers of fabric slip with respect
to each other as fabric structure is not rigid. These stresses are coupled so if compressive stress is greater, there
will be no or lesser slippage whereas a greater slippage
is observed if axial stresses are greater.
On moulding, the distance among the tows (X1) at
external layers increases and that at inside (X2)
decreases or tends to decrease, due to the axial
constraints.
The structure of a 5-layer fabric is not very compact. The slippage of tows in this case, can be better
understood by dividing the bending/moulding in two
phases: 0–90° and 90–180°. In first phase, when the
bending angle is increases from 0 to 90°, the tows
move easily in the beginning thanks to the looseness
of the structure. On approaching the 90°, tightness of
structure increases, and slippage becomes lesser. From
90 to 180°, the slippage of the tows depends on the
competition between the in-plane (tensile force along
fabric surface) and out of plane (compression force
perpendicular to fabric surface) components of the
bending force (Figure 3 and 9). In the second phase,
effective force causes the tows to move in that tight
structure until 135°, where the compression force
becomes maximum (see the aspect ratio at 135° in
Figure 5). Further movement of tows (135–180°)
becomes very difficult due to presence of this maximum compressive force.
Figure 6 shows the changes in the distance between
tows on moulding at radius R1 and in the region centre
± 1 cm. In case of a 13-layer fabric, there is no
significant movement in phase 1 due to compactness of
60
90
120
150
180
Moulding angle (degree)
Figure 6. Distance (centre ± 1 cm) between the columns of
weft on moulding at radius R1 and R2.
structure, but from 90 to 180°, the distance at outside
increased in the same way as in case of the
5-layer fabric. No significant change at the inside layers
is observed. This is due to fact the tows are very close
and did not get space to become closer.
For a given thickness, the slippage “G” is a function of moulding angle “h” and bending radius “R”
and its value can be found by using second degree
regression equation:
G ¼ 0:281 þ 2:7 10 2 h þ 7:2 10 5 (h
þ 2:10 10 2 R
(R
9:26 10
90)2
3
9:6875)2 :
In (2) “G” is the slippage in “mm”, h is the moulding
angle in degrees and R is the radius of moulding in
mm.
Change in thickness on moulding
Thickness of a composite is supposed to change during fabrication due to chemical shrinkage of resin (Li,
Potter, Wisnom, & Stringer, 2004) and compactness of
fibres due to pressure applied. Here, we noted that
moulding of multilayer interlock fabric is itself a
source of thickness change.
On moulding, the thickness of the fabric changes
mainly due to the re-arrangement of the tows, interlayer slippage, slippage of the tows within a layer and
compression on the tows on moulding. In this study,
average of the minimum and maximum thickness of
the fabric was taken as reference to calculate changes
in thickness on moulding.
It is noted that thickness change on moulding is a
function of initial thickness “T” of the piece, bending/
moulding angle “h”, bending/moulding radius and the
distance from the bending point (vertex). When moulded at radius R1 and R2, slight change in thickness
from 0 to 45° was observed in case of both 5- and
The Journal of The Textile Institute
5
Tensile component of bending force
Compressive component
of bending force
Vertex/centre column
Figure 7. Variation of thickness (13 layers) with radius
and moulding angle.
13-layer fabric but afterwards it increased considerably
and reached maximum value at 135° as shown in
Figure 7.
It was also observed that decrease in the aspect
ratio of tows, and their redistribution is the main cause
of the increase of thickness of 13-layer fabrics. Interlayer slippage was not significant due to higher density of this fabric. As shown is Figure 7, the thickness
of fabric moulded at radius R2 (12.5 mm) is greater
than thickness at radius R1 (8 mm). It is due to compression on the tows, which will be lesser in case of
radius R2. So it can be concluded that keeping all
other factors same, the final thickness of the moulded
fabric is directly proportional to moulding radius. It is
also observed that thickness is lesser in the region vertex ± 1 cm due to more compression in this region as
compared to region centre ± 1–2 cm.
Orientation of the tows
Position and orientation of the tows columns in a
woven composite piece play a very important role to
define its mechanical characteristics and thickness. In
multilayer orthogonal layer-to-layer woven fabric, the
yarn columns are perpendicular to each other
(Figure 8) but on moulding their position and
orientation change with respect to the piece surface
(Figure 9). This change in orientation/position affects
not only the thickness and aspect ratio but also the
mechanical properties.
Figure 9. Cross-sectional view of 13-layer fabric moulded
at 180°.
The new orientation is in more order in case of thick
and high-density fabric as compared to thin and lowdensity fabric. Prediction of final tow position is essential to model the mechanical properties of the composite. Figure 9 shows the 13-layer fabric moulded at 180°.
It can be noted that the tows column at the vertex/centre
of fabric did not change its orientation (a = 90°) on
moulding but as we go away from the centre, the orientation angle of columns starts decreasing. At a certain
distance from the centre, this orientation angle gained
minimum value and from that point the columns
remained at that orientation for the rest of the length.
Consider that a fabric of thickness T is moulded
along (outside) a mould of radius R and moulding
angle h. Due to greater outer length, the outer layer
will displace with respect to inner layer by distance
“x” as shown in Figure 10, and at that point the tows
will bend at an angle a with the fabric surface.
The tow column at OO’ being the central column,
will not change its position/orientation but the tow
column at B at distance L from vertex will change its
orientation and will bend at an angle a Figure 10.
Displacement “x” of the tow column depends on
bending radius R, fabric thickness “T” and distances
“L” from centre tow column and can be written as:
x¼L
RL=(R þ T ):
(2)
The maximum value of x is denoted by X and found
equal to “Th/2” where h is the moulding angle in radians. The value of angle of orientation a can be found
by using values of thickness, tow displacement “x”
and slippage G by using Equation (3)
a ¼ tan 1 ½T=(X þ G):
Figure 8. Cross-sectional view of 13-layer fabric with
perpendicular warp and weft columns.
Orientation angle
(3)
Figure 11 shows the predicted value of angle of
orientation of tow columns of a 13-layer fabric
6
Y. Nawab et al.
studied. It is noted that mouldability-based changes
are so significant that they may have prominent effect
on ultimate mechanical properties of the composite
and neglecting them while modelling these properties
may result into prejudicial calculations. Basic rules
were defined for mouldability-based changes, keeping
in view the finding of this study. An empirical model
was proposed to predict the tow-column orientation in
thick and high-density multilayer interlock fabric. The
results of this model are found coherent with the
experimental data. In this article, only orthogonal
layer-to-layer interlock fabrics are studied. Different
weave designs, i.e. orthogonal through the thickness,
layer-to-layer angle interlock, etc. will certainly affect
the compactness/tightness of fabric structure, and the
mouldability-based changes in it. Further research is
required to fully understand the forming behaviour. A
work is already in progress.
a
O’
θ
α
Figure 10. Schematic view of fabric moulded at angle h.
90
Angle of Orientation
10
Angle of orientation
Error %
80
5
70
60
0
50
40
-5
Distance 'L' from vertex
Figure 11. Graph of calculated values of angle of
orientation a and % error.
(R = 12.5 mm and moulding angle 180°) at different
distances from the centre/vertex point, calculated by
using Equation (3).
The percentage of error is the difference of calculated value of the orientation angle from the experimental value and is presented on the same graph.
Its value is less than ± 5%, which is within the
acceptable limits.
Conclusion
In this study, 5- and 13-layers multilayer orthogonal
interlock fabrics were studied during moulding at two
radiuses, and at five moulding angles. Changes in the
fabric, i.e. change in thickness, interlayer slippage,
changes in tow-aspect ratio and in orientation of the
tows as a function of moulding angle and radius were
References
Boisse, P., Zouari, B., & Gasser, A. (2005). A mesoscopic
approach for the simulation of woven fibre composite
forming. Composites Science and Technology, 65(3–4),
429–436.
Chen, X., Lo, W.Y., Tayyar, A.E., & Day, R.J. (2002).
Mouldability of angle-interlock woven fabrics for technical applications. Textile Research Journal, 72(3),
195–200. doi: 10.1177/004051750207200302.
Gowayed, Y., & Fan, H. (2001). Fatigue behavior of textile
composite materials subjected to tension–tension loads.
Polymer Composites, 22(6), 762–769. doi: 10.1002/
pc.10578.
Gu, H., & Zhili, Z. (2002). Tensile behavior of 3D woven
composites by using different fabric structures. Materials & Design, 23(7), 671–674. doi: 10.1016/s0261-3069
(02)00053-5.
Li, C., Potter, K., Wisnom, M.R., & Stringer, G. (2004).
In-situ measurement of chemical shrinkage of MY750
epoxy resin by a novel gravimetric method. Composites
Science and Technology, 64(1), 55–64.
Lomov, S.V., Gusakov, A.V., Huysmans, G., Prodromou,
A., & Verpoest, I. (2000). Textile geometry preprocessor
for meso-mechanical models of woven composites.
Composites Science and Technology, 60(11), 2083–
2095. doi: 10.1016/s0266-3538(00)00121-4.
Lomov, S.V., Huysmans, G., Luo, Y., Parnas, R.S., Prodromou, A., Verpoest, I., et al. (2001). Textile composites:
Modelling strategies. Composites Part A: Applied
Science and Manufacturing, 32(10), 1379–1394.
Long, A. (2005). Design and manufacture of textile composites. Cambridge: Woodhead.
Miravete, A. (1999). 3-D textile reinforcements in composite
materials. Cambridge: Woodhead.
Potluri, P., & Sagar, T.V. (2008). Compaction modelling of
textile preforms for composite structures. Composite
Structures, 86(1–3), 177–185.
Sheng, S.Z., & van Hoa, S. (2003). Modeling of 3D angle
interlock woven fabric composites. Journal of Thermoplastic Composite Materials, 16(1), 45–58. doi:
10.1177/0892705703016001206.
The Journal of The Textile Institute
Verpoest, I., & Lomov, S.V. (2005). Virtual textile composites software wisetex: Integration with micro-mechanical, permeability and structural analysis. Composites
Science and Technology, 65(15–16), 2563–2574. doi:
10.1016/j.compscitech.2005.05.031.
7
Yi, H.L., & Ding, X. (2004). Conventional approach on
manufacturing 3D woven preforms used for composites.
Journal of Industrial Textiles, 34(1), 39–50. doi:
10.1177/1528083704045847.