Original article

Textile Research Journal

2014, Vol. 84(6) 638–654

Three-dimensional circular various weave ! The Author(s) 2014

Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
patterns in woven preform structures DOI: 10.1177/0040517513499437
trj.sagepub.com

Kadir Bilisik1, Nesrin Sahbaz Karaduman1,

Nedim Erman Bilisik2 and Havva Esra Bilisik2

Abstract
The aim of this study was to develop three-dimensional (3D) fully interlaced representative circular woven preform
structures and to understand the effects of weave pattern and number of layers on 3D circular woven structures. Various
3D circular woven preforms were developed. Data generated from these structures included yarn-to-yarn space, density,
yarn angle, yarn length and crimp.
It was shown that the weave patterns affected the 3D circular woven preform structures. The yarn-to-yarn spaces in the
3D fully interlaced circular structures were high compared to the traditional 3D orthogonal circular woven structures in
fabric circumference (fabric outside surface) due to the interlacement of the yarn sets. The 3D plain, twill and satin
structures resulted in axial angle (a) in fabric length; circumferential angle (c), and interlaced radial angle (ri) in fabric
circumference and fabric diameter due to the axial-circumferential and axial-radial interlacements. The weave patterns
slightly affected the yarn angles. On the other hand, it was observed that the number of layers considerably affected the
radial arc length and the radial length in wall thickness in the 3D circular woven structure.
The interlacement on 3D plain, twill and satin circular woven structures resulted in axial crimp, circumferential crimp and
radial crimp. The crimps in the 3D fully interlaced circular woven structures slightly depended on the types of weave
pattern and the number of layers.

Keywords
Three-dimensional fully interlaced preform, three-dimensional plain circular preform, three-dimensional twill circular
preform, three-dimensional satin circular preform, radial crimp, circumferential and axial crimps

to make multi-directional preforms, but there is no cir-

Introduction
cumferential and true radial reinforcement.18 A few
Three-dimensional (3D) textile structural composites special processes have also been developed to make
are used in various civil and defense applications due 3D preforms: these are sleeve wrapping and fiber place-
to their better specific properties,1–6 and 3D reinforce- ment. However, there is no radial reinforcement in the
ment can be considered as alternative materials since preform.17
they are delamination-free and damage tolerant.3,7–9 A 3D non-interlaced circular woven fabric was
Textile preforms are made by weaving, braiding, knit- developed.19 The fabric included axial, circumferential
ting, stitching and nonwoven techniques, and can be and only +bias yarns. These are woven on the surface
chosen, generally, based on the end-use require-
ments.2,10–12 Generally, 3D circular preforms are fabri-
cated by a modified flat weaving loom or in some cases 1
Department of Textile Engineering, Faculty of Engineering, Erciyes
by specially designed automated looms and manufac- University, Turkey
tured to a near-net shape to reduce scrap.11–16 2
Department of Electronic Engineering, Faculty of Engineering, Istanbul
Traditional filament winding and pultrusion techniques Kultur University, Turkey
were used to make one or multi-directional 3D circular
Corresponding author:
performs. However, there is no radial reinforcement in Kadir Bilisik, Department of Textile Engineering, Faculty of Engineering,
the thickness direction.17 On the other hand, two- Erciyes University, 38039 Talas-Kayseri, Turkey.
dimensional (2D) triaxial and 3D braiding were utilized Email: kadirbilisik@gmail.com

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

Bilisik et al. 639

of a mandrel with the resulting shape being determined It was also claimed that the density and the placement
by the action of the jacquard.19 A 3D circular woven- of z-yarn in the 3D woven preform influenced the in-
knitted preform, which has axial, circumferential and plane properties of the 3D woven structure.38 Bilisik9
radial yarns, was developed.20 All yarn sets are cylin- showed that the increasing of the fiber density in the
drically orthogonal inside the structure except for the warp or bias and filling fiber sets resulted in a high
outer and inner surfaces in which the yarn sets are total preform volume fraction and that porosity in the
interlaced to provide structural integrity.20 A 3D circu- crossing points of fiber sets in the preform was reduced.
lar preform, in which axial and circumferential yarns Bilisik17 also found that the parameters related with the
were wound on the mold and at a later stage the pre- multiaxis 3D circular woven fabric-process were bias
form was stitched to put some fiber in the radial direc- orientation, radial and circumferential yarn insertion,
tion, was fabricated for space applications.21 However, beat-up and take-up. The bias yarns on the outer and
stitching can damage the yarn sets. The 3D preform inner surfaces of the structure formed helical paths and
technology called AutoweaveTM was developed to be there was a slight angle difference between them, espe-
used in the exit cone for aerospace applications.22 The cially in the producing of thick wall preforms. There was
preform has axial, circumferential yarns and a pul- a certain relation between preform density (fiber volume
truded radial rod and is suitable for only part manu- fraction), bias yarn orientation and take-up rate. Cox
facturing. The winding principle was employed to form et al.7 stated that a low volume fraction 3D woven pre-
the structure.22 However, preparing the radial rod skel- form may perform well under impact load compared to a
eton in the mandrel is labor intensive and not suitable tight volume fraction 3D woven perform. On the other
to make a continuous circular preform. Another type of hand, it was explained that under the impact load, 3D
3D circular preform structure was developed with a woven composites confined the impact energy around
different method for the automotive industry where the local region due to the z-yarn.38
axial, circumferential and radial yarns were used.23 As seen in the literature, only a few studies have been
The axial yarns were stationary, but the winding prin- carried out on 3D circular non-interlaced woven pre-
ciple was employed to insert the circumferential yarns form structures. Therefore, the aim of this study was to
and the radial yarn bobbins were exchanged between develop 3D fully interlaced circular woven fabrics, and
each adjacent radial row in order to insert the radial to examine some of their geometrical base structural
yarn.23 This is also suitable for part manufacturing and features for possible soft and rigid material
cannot be used in continuous circular preform fabrica- applications.
tion. Recently, a multiaxis 3D circular loom was devel-
oped to make 3D circular non-interlaced woven
preform under the various unit cells.24–26 Materials and methods
Three-dimensional flat woven lattice structures were Three-dimensional fully interlaced circular woven
developed by Fukuta et al.27,28 and Khokar.29 They
have three sets of yarns. These are the warp, weft and
preform and method
z-yarn (binder). These are interlaced with each other An experimental rig was constructed to make a repre-
based on a plain weave pattern. The weft yarns in the sentative 3D fully interlaced circular woven preform
lateral direction interlace around each warp yarn, (plain, twill and satin) based on the 3D weaving
which occupies the longitudinal direction, and the method. This method involves five distinct steps to
binder yarns in the vertical direction also interlace weave two to five layer preforms. These are axial yarn
around each warp yarn.27–29 However, twill or satin let-off, 2D circular shedding, multiple circumferential
or any derived interlace pattern other than plain yarn insertions and beat-up, which was carried out by
cannot be produced. open reed. Figure 1 shows the 3D circular plain weave
Three-dimensional flat and circular woven preforms (a1–a5). The 3D plain, twill and satin circular woven
are classified based on various parameters. These par- preforms, which can be defined as axial-circumferential
ameters depend on fiber type and formation, fiber and axial-radial yarns, were interlaced based on plain,
orientation and interlacements and micro and macro twill and satin weave patterns.
unit cells.11,12,15 Various studies have been published In order to make the representative 3D circular plain
on the geometrical and mechanical properties of 3D woven preform, the axial yarns must be arranged in a
circular and flat orthogonal30–36 and a few interlaced matrix of circular rows and radial columns within the
flat woven fabric preforms.10,11,37,38 It was reported required circular cross-section. It looks like a circular
that the take-up rate of the 3D weaving affected the network in which the yarns are laid in between the
yarn-to-yarn distance in the dry preform, which even- adjacent axial yarns as shown in Figure 1 (a1). The
tually influenced the directional and total volume frac- first step is the one-step sequential movement of an
tion as well as z-yarn angle of 3D woven fabrics.38 even and odd number of axial layers in the radial

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

640 Textile Research Journal 84(6)

Figure 1. Three-dimensional (3D) weaving method to make representative fully interlaced circular woven preform; 3D circular plain
woven preform (a1–a5).

column direction (a2). This was carried out by using the circumferential yarn, which was mounted on each indi-
2D circular shedding unit (not shown). The second step vidually rotated ring, was used. In addition, the radial
is circumferential yarn insertion between each axial carriers reciprocated linearly to the radial corridor of
layer in the circular row direction (a3). The third step the 2D shedding plane on the rig, thus crossing the
is the one-step sequential movement of an even and odd radial yarns in the preform structure (crossing
number of axial layers in the circular row direction (a4). shedding).
This was again fulfilled by the 2D circular shedding. Three-dimensional representative circular woven
The fourth step is radial yarn insertion between each preform structures made by using polyethylene tubes
axial layer in the radial column direction (a4). After (diameter, 5 mm and wall thickness, 1 mm) using the
these steps (a2–a4) were repeated, the 3D circular 3D weaving method are shown in Figure 2. Individual
plain woven preform structure was achieved (a5). The shuttles for circumferential insertion and crossing shed-
steps were repeated depending on preform length ding were utilized to fabricate the 3D representative
requirements. The representative traditional 3D circu- woven preforms. The 3D representative woven pre-
lar woven orthogonal structure was also made in order forms were formed by using light beat-up (open reed)
to compare the 3D fully interlaced circular woven pre- in order to obtain perfect alignment in, generally, three
form structures. The 3D orthogonal circular woven cylindrical directions.
perform unit cell has basically three yarn sets, namely The 3D circular woven structures were designed
axial, circumferential and radial yarns. Axial yarns as plain (3DCWP) (a1 and a2) in which axial-
were arranged in a matrix of circular rows and radial circumferential and axial-radial yarns were interlaced
columns. Circumferential yarns were laid down in a 1/1 plain pattern; twill (3DCWT) (b1 and b2) in
between each adjacent axial yarn row. They were which the axial-circumferential yarn was interlaced in a
single end and deposited through the preform length. 2/2 twill weave pattern, whereas the axial-radial yarn
Radial ends were positioned between each axial row was interlaced in a 1/1 plain pattern at two layers and 3/
through the preform thickness and they locked all 2 ribs at five layers due to the limitation of the number
other yarn sets to ensure the structural integrity of of layers in preform wall thickness; satin (3DCWS) (c1
the preform. During the formation of the developed and c2) in which the axial-circumferential yarn was
3D fully interlaced and orthogonal circular woven interlaced in a 1/4 satin weave pattern, whereas the
preform structures, the individual shuttle for axial-radial yarn was interlaced in a 1/1 plain pattern

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

Bilisik et al. 641

Figure 2. Three-dimensional representative fully interlaced and traditional orthogonal circular woven preform structures where a
polyethylene tube was used instead of high-modulus yarns. General outside and inside view of five-layer circular plain woven preform,
respectively (a1 and a2); general outside and inside view of twill woven preform, respectively (b1 and b2); general outside and inside
view of satin woven preform, respectively (c1 and c2); general outside and inside view of orthogonal woven preform, respectively
(d1 and d2).

at two layers and 3/2 ribs at five layers due to the limi- length: 0–300 mm and precision: 0.2 mm). The angle
tation of the number of layers in the preform wall thick- measurements of the 3D representative woven preform
ness and traditional orthogonal (3DCWO) (d1 and d2) structures were performed using a manual angle instru-
woven patterns. The structures were produced in vari- ment (measurement angle: 0–180 and precision: 0.2 ).
ous numbers of layers at the same inside diameter. The
measurements of the 3D representative circular woven
preform structures were not repeated due to the limited
Dimensions
number of samples. On the other hand, the measure- The outside diameter (Sd), length (Sl) and wall thick-
ments of the 3D representative woven preform struc- ness (St) measurements of the representative 3D circu-
tures were repeated five times on the outside surface of lar woven preform structures were carried out. The
each structure. The average values of the various meas- precision of dimensional measurements was 0.01 mm.
urements were presented. On the other hand, the outside diameter, wall thickness
and length measurements were performed on the best
Parameters measured in 3D representative circular representative part in the 3D circular woven preform
structures. Yarn-to-yarn distance on axial-axial (a-a),
woven preform structures circumferential-circumferential (c-c) and radial-radial
Measurements of the 3D representative fully interlaced (r-r) in the outside surface of circumference, length
circular woven preforms were performed in normal and cross-sections (diameter) of the 3D representative
conditions (out-of-loom conditions). The measure- woven structures were measured as shown in Figure 3.
ments of the representative circular woven preforms
in normal conditions were carried out in force-free
Angles
environments. The measurements on traditional 3D
representative orthogonal circular woven preforms The measured angles on the various 3D representative
were also fulfilled in order to compare them with 3D circular woven preform structures were defined as fol-
representative fully interlaced woven preforms. The lows: a was the axial angle, which was between the
dimensional measurements of the 3D representative axial and circumferential yarns in the z-direction (fabric
woven preform structures were performed using length); ar was the trajectory axial angle, which was
the Newman digital caliper (Measurement length: between the axial and the radial yarns in the radial (r)
0–150 mm and precision: 0.01 mm, China). The direction (fabric diameter); c was the circumferential
uncrimped yarn length and arc-length measurements angle, which was between the circumferential and the
were performed using a flexible ruler (measurement axial yarns in the -direction (fabric circumference);

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

642 Textile Research Journal 84(6)

Figure 3. Measured yarn-to-yarn space on three-dimensional representative fully interlaced circular woven preform structure.

cr was the trajectory circumferential angle, which

Density
was between the circumferential and the radial yarns
in the r-direction (fabric diameter); r was the radial The densities of the axial, circumferential and radial
angle, which was between the radial and the axial yarns yarns were measured on the outside surface and
in the r-direction (fabric diameter); ra was the tra- cross-section of the 3D representative circular woven
jectory radial angle, which was between the radial and preform structure. The precision of density measure-
the axial yarns in the z-direction (fabric length); ri ments was 0.2 ends/5 cm. Also, the total yarn ends for
was the interlaced radial angle, which was between the each woven structure were found in the following rela-
radial and the axial yarns in the r-direction (fabric tions as
diameter); ric was the trajectory interlaced radial
angle, which was between the radial and the circumfer- Total axial ends, Ae ¼ N  M ð2Þ
ential yarns in the -direction (fabric circumference);
h was the helix angle, which was between the circum- Total circumferential ends, Ce ¼ N þ 1
ferential and the axial yarns in the z-direction (fabric ðfor a traditional 3D orthogonal circular woven structureÞ
circumference) for all the 3D circular woven structures. ð3Þ
The helix angle is expressed with regards to actual helix
angle (ha) in Equation (1). All the angles on 3D rep- Ce ¼ Nðfor a 3D fully interlaced circular woven structureÞ
resentative circular woven preform structures are ð4Þ
shown in Figure 4. It was found that a, c and
ri were due to the fully interlaced structures. Total radial yarn ends, Re ¼ M þ 1 ð5Þ
However, r was partly due to light beat-up and h
was partly due to the nature of the 3D circular structure where N is the number of yarns in the circular row
(that is, continuous circumferential yarn insertion) direction and M is the number of yarns in the radial
and partly due to light beat-up, as shown in Figures column direction.
4(d)–(f), respectively. The angles between the axial
and circumferential yarns, and axial and radial yarns
were measured on the outside surface of the 3D repre-
Yarn lengths
sentative circular woven preform structures by means The measured yarn lengths in the 3D representative
of a 4 mm rod. The precision of angle measurements circular woven preform structures are shown in
was 0.2 degrees: Figure 5. These include uncrimped axial length (la),
uncrimped circumferential length (lc), radial arc
h ¼ 90 þ ha ðdegreeÞ ð1Þ length on the outside surface of the structure (lra),
radial length in structure wall thickness (lr) and
where h is the helix angle and ha is the actual helix uncrimped total radial length (lrt). The precision of
angle. yarn length measurements was 0.2 mm. Uncrimped

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

Bilisik et al. 643

Figure 4. Measured angles on various three-dimensional representative circular preform structures; axial angle (a) (a), trajectory
axial angle (ar) (a); circumferential angle (c) (c), trajectory circumferential angle (cr) (c); radial angle (r) (d), trajectory radial
angle (ra) (d); interlaced radial angle (ri) (b), trajectory interlaced radial angle (ric) (b); helix angle (h) (e) and (f).

Figure 5. Measured yarn length on the three-dimensional representative fully interlaced circular woven preform structure:
uncrimped axial length (la) (a); uncrimped circumferential length (lc) (c); radial arc length in structures surface (lra) (b) and (c); radial
length in structures wall thickness (lr) (b) and (c); structures length (Sl) (a); structure’s outside diameter and outside surface length,
respectively (Sd, Ssl) (a); structure’s wall thickness (St) (a).

total radial length was calculated in the following rela- Therefore, the required radial yarn in the weaving can
tions as be easily determined.

lrt ¼ lra þ lr ð6Þ

Crimp
Equation (5) is simple and it provided the relations Crimps in the 3D fully interlaced representative circular
between each segment length and the total length of woven preform structure were calculated based on the
the radial in the 3D circular woven structures. structure dimensions and the uncrimped representative

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

644 Textile Research Journal 84(6)

yarn lengths. The following relations could be in Figure 6 (b1, b2 and b3). In the 3DCWS structure,
expressed as the axial and circumferential yarns were interlaced as a
1/4 (skipping step 2) satin pattern in the in-plane direc-
Cað%Þ ¼ ðla  SlÞ  100=Sl ð7Þ tion (Z), whereas the axial and radial yarns were inter-
laced as a 1/1 plain pattern in the out-of-plane direction
Ccð%Þ ¼ ðlc  SslÞ  100=Ssl ð8Þ (R) for two layers and 3/2 ribs (skipping step 1, plain
derivatives) for five layers due to the limitation of the
Crð%Þ ¼ ðlrt  St  100Þ=St ð9Þ number of axial layers of the rig. The perspective and
cross-sectional view of the 3DCWS (1/4 satin in axial-
where Ca is the axial crimp(%), la is the uncrimped circumferential and 3/2 ribs in axial-radial) representa-
axial length (cm), Sl is the structures length (cm), Cc tive structure and computer-aided drawings are shown
is the circumferential crimp (%), lc is the uncrimped in Figure 6 (c1, c2 and c3).
circumferential length (cm), Ssl represents the struc- On the other hand, a traditional 3D circular woven
tures outside surface length (cm), Cr is the radial orthogonal (3DCWO) representative structure was
crimp (%), lrt is the uncrimped total radial length made for comparison purposes with the 3D fully inter-
(cm) and St is the structure’s wall thickness. The preci- laced circular woven structures. The perspective and
sions of dimensional and yarn length measurements cross-sectional view of the 3DCWO representative
were 0.01 and 0.2 mm, respectively. woven structure and computer-aided drawings are
shown in Figure 6 (d1, d2 and d3).

Results and discussion Preform yarn-to-yarn space results

Preform structure results The yarn-to-yarn space and dimensional specifications
Three-dimensional fully interlaced circular woven of various 3D representative circular woven preform
structures were developed for two and five layers. structures are presented in Table 1. Figure 7(a) shows
Three-dimensional fully interlaced woven structures the relationship between yarn-to-yarn distance and the
can be defined if all three yarn sets (axial, circumferen- 3D circular woven preform structures in the -z plane
tial and radial) are interlaced with each other based on (fabric length) for two and five layers. Figure 7(b)
any specific weave pattern, for instance, plain or twill or shows the relationship between the yarn-to-yarn dis-
satin. The resulting structure was considered as a 3D tance and the 3D circular woven preform structures
fully interlaced circular woven structure. The 3D fully in the -r plane (fabric cross-section or fabric diameter)
interlaced circular woven structures were divided into for two and five layers. Figure 7(c) shows the relation-
three basic patterns by analogy with traditional 2D ship between the yarn-to-yarn distance and the 3D cir-
woven fabrics as 3D circular woven plain (3DCWP), cular woven preform structures in the -r plane (fabric
3D circular woven twill (3DCWT) and 3D circular circumference) for two and five layers.
woven satin (3DCWS), as shown in Figure 6. In the yarn-to-yarn space of two- and five-layer
In the 3DCWP structure, the axial and circumferen- structures to fabric length as shown in Figure 7(a),
tial yarns were interlaced as a 1/1 plain pattern in the and Table 1, the circumferential-to-circumferential
in-plane direction (Z), whereas the axial and radial and the radial-to-radial spaces in all 3D woven preform
yarns were interlaced as a 1/1 plain pattern in the structures were almost equal but they were higher than
out-of-plane direction (R) for two and five layers. the yarn-to-yarn spaces in fabric cross-section and
The perspective and cross-sectional view of the fabric circumference. This was because of the process-
3DCWP representative structure and computer-aided ing parameters. For instance, light beat-up was used
drawings using Unigraphics NX6 are shown in during weaving under constant take-up rate. Three-
Figure 6 (a1, a2 and a3). In the 3DCWT structure, dimensional fully interlaced structures have slightly
the axial and circumferential yarns were interlaced as high yarn-to-yarn spaces compared to the traditional
a 2/2 (skipping step 1) twill pattern in the in-plane dir- 3D representative orthogonal woven structures due to
ection (Z), whereas the axial and radial were interlaced interlacement. It seemed that the yarn-to-yarn spaces
as a 1/1 plain pattern in the out-of-plane direction (R) were proportional to the interlacement occurring
for two layers and 3/2 ribs (skipping step 1, plain between the axial-circumferential and the axial-radial
derivatives) for five layers due to the limitation of the yarns. On the other hand, increasing the number of
number of axial layers of the rig. The perspective and layers from two to five did not significantly affect the
cross-sectional view of the 3DCWT (2/2 twill in axial- yarn-to-yarn spaces in all 3D preform structures.
circumferential and 3/2 ribs in axial-radial) representa- In the yarn-to-yarn spaces of all 3D woven pre-
tive structure and computer-aided drawings are shown form structure to the fabric cross-section, as shown

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

Bilisik et al. 645

Figure 6. Developed representative three-dimensional fully interlaced and orthogonal circular woven preform structures. General
view of five-layer actual and computer-aided drawing of 3DCWP plain (a1–a3), 3DCWT twill (b1–b3), 3DCWS satin (c1–c3) and
traditional 3DCWO orthogonal (d1–d3) circular woven preform structures.

in Figure 7(b) and Table 1, the axial-to-axial space 1.0 cm in two-layer structures and from 0.8 to 0.8 cm
varied from 0.8 to 1.2 cm in two-layer structures and in five-layer structures. The radial-to-radial space
from 0.8 to 1.0 cm in five-layer structures. The circum- varied from 1.2 to 1.8 cm in two-layer structures and
ferential-to-circumferential space varied from 0.8 to from 1.5 to 2.5 cm in five-layer structures. It was

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

646 Textile Research Journal 84(6)

Table 1. Yarn-to-yarn space and the dimensional specifications of various three-dimensional representative circular woven preform
structures

Yarn-to-yarn distance in outside surface

Total yarns Structure (Circumference  length  cross-section, cm)

Outside
diameter  wall Axial Circumferential Radial
Fabric Number of thickness Axial – axial circumferential Radial – radial
structure layers Axial Circumferential Radial (Sd  St, cm) (a-a) (c-c) (r-r)

3DCWP 2 2  24 2  10 1  24 12.2  3.0 1.8  –  1.0 –  4.5  0.8 1.8  4.5  1.8
5 5  24 5  10 1  24 14.5  6.0 2.5  –  1.0 –  4.0  0.8 2.5  4.0  2.5
3DCWT 2 2  24 2  10 1  24 9.1  3.0 1.2  –  1.2 –  4.7  1.0 1.2  4.7  1.2
5 5  24 5  10 1  24 14.2  5.8 1.6  –  1.0 –  4.3  0.8 1.6  4.3  1.6
3DCWS 2 2  24 2  10 1  24 9.0  3.1 1.2  –  1.2 –  4.7  1.0 1.2  4.7  1.2
5 5  24 5  10 1  24 12.5  6.6 1.8  –  1.0 –  4.3  0.8 1.8  4.3  1.8
3DCWO 2 2  24 3  10 1  24 13.5  3.2 1.5  –  0.8 –  4.0  0.8 1.5  4.0  1.5
5 5  24 6  10 1  24 14.0  5.3 1.5  –  0.8 –  3.2  0.8 1.5  3.2  1.5

considered that the values of the axial and circumfer- 3D orthogonal woven preform structures due to inter-
ential yarn-to-yarn spaces in all 3D woven preform lacement in the radial direction of the preform struc-
structures were close to each other except for radial ture. On the other hand, the a angle in the 3D woven
space. structures was lower than the r angle due to interlace-
In the yarn-to-yarn spaces of all 3D woven preform ment. In addition, the r angle of all 3D woven preform
structures to fabric circumference, as shown in structures in five layers was lower than those of two
Figure 7(c) and Table 1, both the axial-to-axial and layers. This is because the radial yarn became steep
radial-to-radial spaces varied from 1.2 to 1.8 cm in when the number of layers increased.
two-layer structures and from 1.5 to 2.5 cm in five- In the ri and c angles of all 3D woven preform
layer structures. It was considered that the values of structures to fabric circumference, as shown in Figure
axial and circumferential yarn-to-yarn spaces in each 8(b) and Table 2, the ri angle varied from 6 to 7
3D woven preform structure were almost equal. degrees in two-layer structures and from 8 to 9 degrees
However, the axial and radial yarn-to-yarn spaces of in five-layer structures. The c angle varied from 4 to 7
all 3D circular woven structures in five layers were degrees in two-layer structures and from 7 to 8 degrees
slightly higher than those in two layers due to the in five-layer structures. It was also found that the ri
increasing number of layers. and c angles in all the 3D woven structures in five-
layer structures were slightly higher than those in
two-layer structures. Also, the a, c and ri angles
Preform structure angle results mainly depended on weave pattern, whereas the r
The measured angles of the various 3D representative angle depended on the weaving processing parameters,
circular woven preform structures are presented in such as beat-up and take-up rate. Also, increasing the
Table 2. Figure 8(a) shows the relationship between number of layers tended to decrease the r angle. In the
the yarn angle and the 3D circular woven preform h angle of two- and five-layer structures to fabric cir-
structures in the r-z plane (fabric length) for two and cumference, as seen in Figure 8(c) and Table 2, the h
five layers. Figure 8(b) shows the relationship between angle in all the 3D circular woven structures in five-
the yarn angle and the 3D circular woven preform layer structures was slightly higher than those in two-
structures in the -r plane (fabric circumference) for layer structures. It appears that weave patterns did not
two and five layers. Figure 8(c) shows the relationship affect the h angle, but the number of layers in 3D cir-
between the yarn angle and the 3D circular woven pre- cular woven structures did so slightly.
form structures in the -z plane (fabric circumference)
for two and five layers.
Preform structure density results
In the r and a angles of the two- and five-layer
structures to fabric length, as seen in Figure 8(a) and The density results of the various 3D representative
Table 2, the r angle in 3D fully interlaced circular circular woven preform structures are presented in
woven structures was slightly higher than in traditional Table 3. Figure 9(a) shows the relationship between

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

Bilisik et al. 647

Figure 7. (a) Relationship between the yarn-to-yarn distance and the three-dimensional (3D) circular woven preform structures in
the -z plane (fabric length). (b) Relationship between the yarn-to-yarn distance and the 3D circular woven preform structures in the
-r plane (fabric cross-section). and (c) Relationship between the yarn-to-yarn distance and the 3D circular woven preform structures
in the -r plane (fabric circumference).

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

648 Textile Research Journal 84(6)

Table 2. Measured angles of the various three-dimensional representative circular woven preform structures
Total yarns Axial angle Circumferential angle Radial angle Helix angle
Fabric Number
structure of layers Axial Circumferential Radial (a ) (ar ) (c ) (cr ) (r ) (ra ) (ri ) (ric ) (h ) (ha )

3DCWP 2 2  24 2  10 1  24 13 77 4 86 55 35 6 84 97 7
5 5  24 5  10 1  24 8 82 7 83 40 50 8 82 95 5
3DCWT 2 2  24 2  10 1  24 10 80 7 83 68 22 7 83 105 15
5 5  24 5  10 1  24 8 82 7 83 27 63 9 81 102 12
3DCWS 2 2  24 2  10 1  24 7 83 5 85 55 35 6 84 97 7
5 5  24 5  10 1  24 5 85 8 82 35 55 8 82 95 5
3DCWO 2 2  24 3  10 1  24 0 90 0 90 50 40 0 90 100 10
5 5  24 6  10 1  24 0 90 0 90 25 65 0 90 95 5

the density and the 3D circular woven preform struc- yarn length and the 3D circular woven preform struc-
tures in the r-z plane (fabric length) for two and five tures for two and five layers.
layers. Figure 9(b) shows the relationship between the As seen in Figure 10 and Table 4, the radial arc
density and the 3D circular woven preform structures lengths of the 3D fully interlaced woven structures in
in the -r plane (fabric circumference) for two and two and five layers were slightly higher than those of
five layers. traditional 3D orthogonal woven structures due to
As seen in Figure 9(a) and Table 3, the circumferen- interlacement. In addition, it was found that the
tial and radial densities (ends per 10 cm) of all 3D pre- radial lengths in wall thickness of the 3D circular
form woven structures in two and five layers to fabric satin woven structures in two and five layers were
length were almost equal to each other. This indicated higher than those of all the other 3D woven preform
that all the 3D structures were uniformly and consist- structures. Also, the total radial lengths of all 3D cir-
ently fabricated. This is of importance to achieve regu- cular woven structures in five layers were higher than
lar directional properties in the 3D circular woven those in two-layer structures. This indicated that when
structure for either soft or rigid applications. the number of layers increased, the total radial lengths
As seen in Figure 9(b) and Table 3, the axial and of the 3D circular woven structure increased.
radial densities (ends per 10 cm) of all the 3D fully
interlaced preform woven structures in two and five
layers to fabric circumference were slightly irregular
Preform structure crimp results
due to the interlacement between the axial-circumfer- The axial crimp, circumferential crimp and radial crimp
ence and the axial-radial yarns. On the other hand, the values of the various 3D representative circular woven
circumference and the radial densities of all 3D circular preform structures, especially for the fully interlaced
preform woven structures in two and five layers to circular woven structures, are presented in Table 5.
fabric length were over half compared to the axial Figure 11 shows the relationship between the yarn
and radial densities of all the 3D circular preform crimp and the 3D circular woven preform structures
woven structures in two and five layers to fabric for two and five layers.
circumference. As seen in Figure 11 and Table 5, the radial crimps
(Cr) of the 3D fully interlaced structures in two layers
were higher than those in five layers due to the small
Preform structure yarn length results
structure wall thickness/total radial yarn length ratio.
The radial yarn length and the uncrimped axial and On the other hand, the radial crimps of the 3D fully
circumferential yarn length values of the various 3D interlaced structures in two and five layers varied from
representative circular woven preform structures are 53.33 to 58.06 (%) and from 8.33 to 10.60(%), respect-
presented in Table 4. The results of the radial yarn ively. Also, it was noted that the Cr of the 3D satin
length included radial arc length, which appeared on woven structure in two and five layers was slightly
both the outside and inside surface of the 3D circular higher than those of the 3D plain and twill structures.
woven preform structures, and radial length in wall The axial (Ca) and the circumferential crimp (Cc) of the
thickness, which was placed in the out-of-plane direc- 3D fully interlaced circular woven structures varied
tion of the 3D structure, and total radial yarn length. from 6.5 to 8.0 (%) and from 2.80 to 18.54 (%) in
Figure 10 shows the relationship between the radial two layers, respectively. However, the Ca and Cc of

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

Bilisik et al. 649

Figure 8. (a) Relationship between the yarn angle and the three-dimensional (3D) circular woven preform structures in the r-z plane
(fabric length). (b) Relationship between the yarn angle and the 3D circular woven preform structures in the -r plane (fabric
circumference). and (c) Relationship between the yarn helix angle and the 3D circular woven preform structures in the -z plane
(fabric circumference).

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

650 Textile Research Journal 84(6)

Table 3. Measured density of the various three-dimensional representative circular woven preform structures

Structure Fabric density on outside surface

Total yarns Length  outside (circumference  length, ends/10 cm)
diameter  wall
thickness
Fabric Number (Sl  Sd Circumferential Radial
structure of layers Axial Circumferential Radial  St, cm) Axial yarn yarn yarn

3DCWP 2 2  24 2  10 1  24 20  12.7  3.0 7.0  – –  3.0 7.0  3.0

5 5  24 5  10 1  24 20  14.5  6.0 6.0  – –  3.0 6.0  3.0
3DCWT 2 2  24 2  10 1  24 20  9.1  3.0 10.0  – –  3.0 10.0  3.0
5 5  24 5  10 1  24 20  14.2  5.8 7.0  – –  3.0 7.0  3.0
3DCWS 2 2  24 2  10 1  24 20  9.0  3.1 8.0  – –  3.0 8.0  3.0
5 5  24 5  10 1  24 20  12.5  6.6 6.0  – –  3.0 6.0  3.0
3DCWO 2 2  24 3  10 1  24 20  13.5  3.2 7.0  – –  3.0 7.0  3.0
5 5  24 6  10 1  24 20  14.0  5.3 7.0  – –  4.0 7.0  3.0

Figure 9. (a) Relationship between the density and the three-dimensional (3D) circular woven preform structures in the r-z plane
(fabric length). and (b) Relationship between the density and the 3D circular woven preform structures in the -r plane (fabric
circumference).

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

Bilisik et al. 651

Table 4. Measured yarn lengths of the various three-dimensional representative circular woven preform structures
Total yarns Uncrimped  radial yarn length

Uncrimped Uncrimped Radial arc Radial length

axial yarn circumferential length on on wall
Fabric Number length yarn length outside surface thickness Total radial
structure of layers Axial Circumferential Radial (la/Sl, cm) (lc/Ssl, cm) (lra, cm) (lr, cm) length (lrt, cm)

3DCWP 2 2  24 2  10 1  24 21.6 41.0 1.8 2.8 4.6

5 5  24 5  10 1  24 21.5 47.0 2.3 4.2 6.5
3DCWT 2 2  24 2  10 1  24 21.3 30.7 1.6 3.0 4.6
5 5  24 5  10 1  24 21.2 46.0 2.0 4.4 6.4
3DCWS 2 2  24 2  10 1  24 21.3 33.5 1.4 3.5 4.9
5 5  24 5  10 1  24 21.2 41.0 2.3 5.0 7.3
3DCWO 2 2  24 3  10 1  24 20.0 42.4 1.6 3.0 4.6
5 5  24 6  10 1  24 20.0 44.0 1.2 4.5 5.7

Figure 10. Relationship between the radial yarn length and the three-dimensional (3D) circular woven preform structures.

Table 5. Measured crimp of the various three-dimensional representative circular woven preform structures
Uncrimped yarn length Structure Crimp

Outside
Outside Wall surface
Fabric Number Axial Circumferential Radial Length diameter Thickness length Axial Circumferential Radial
structure of layers (la, cm) (lc, cm) (lrt, cm) (Sl, cm) (Sd, cm) (St, cm) (Ssl, cm) (Ca, %) (Cc, %) (Cr, %)

3DCWP 2 21.6 41.0 4.6 20 12.7 3.0 39.88 8.00 2.80 53.33
5 21.5 47.0 6.5 20 14.5 6.0 45.53 7.50 3.22 8.33
3DCWT 2 21.3 30.7 4.6 20 9.1 3.0 28.57 6.50 7.45 53.33
5 21.2 46.0 6.4 20 14.2 5.8 44.59 6.00 3.16 10.34
3DCWS 2 21.3 33.5 4.9 20 9.0 3.1 28.26 6.50 18.54 58.06
5 21.2 41.0 7.3 20 12.5 6.6 39.25 6.00 4.45 10.60
3DCWO 2 20.0 42.4 4.6 20 13.5 3.2 42.39 0.00 0.00 0.00
5 20.0 44.0 5.7 20 14.0 5.3 43.96 0.00 0.00 0.00

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

652 Textile Research Journal 84(6)

Figure 11. Relationship between the yarn crimps and the three-dimensional (3D) circular woven preform structures.

3D fully interlaced circular woven structures varied depended on the take-up rate, which influenced the dir-
from 6.0 to 7.5 (%) and from 3.16 to 4.45 (%) in ectional volume fraction of the 3D circular woven pre-
five-layer structures, respectively. It was found form. When the yarn sets in the 3D structure
that the radial crimp of the 3D fully interlaced pre- were interlaced, the axial and radial densities in the -
form structures was higher than those of the axial r plane were slightly irregular, which eventually
crimp and the circumferential crimp in two and affected the local properties of the 3D circular pre-
five layers. form structures. However, depending on soft or
rigid requirements, the axial and the circumferential,
and the radial densities in both the r-z plane and -r
General results plane can be tailored based on the processing
parameters.
Preform space
Yarn-to-yarn spaces in all the developed 3D woven pre-
form structures will affect the directional and total fiber
Preform yarn length
volume fractions. The basic processing parameters The radial length in all the 3D circular woven preform
for the yarn-to-yarn spaces were the beat-up and the structures was identified as the most critical yarn length
take-up rate, which mainly affect the circumferential- due to the placement of the 3D circular woven preform
to-circumferential and the radial-to-radial spaces. wall thickness. When all the yarn sets in the 3D woven
These spaces in the preform can also affect matrix infil- preform were interlaced fully, the radial arc length
tration during composite fabrication, for instance, slightly increased. On the other hand, weave pattern
directional porosity distribution in the preform, types slightly affected the radial arc length. The radial
infiltration time and fiber alignment during length in wall thickness depended on the number of
consolidation.12,35,38 layers and the yarn interlacements. In general, yarn
interlacement and the number of layers influence all
the yarn lengths in 3D woven circular preform
Preform angle
structures.
Generally, the a, c and ri angles can be affected by
weave pattern, whereas the r angle can be affected by
the weaving processing parameters, such as beat-up and
Preform crimp
take-up rate and the number of layers. The 3D fully interlaced circular woven structures had
axial crimp, circumferential crimp and radial crimp. On
the other hand, the traditional 3D circular orthogonal
Preform density
woven structures had no directional crimps. It was
The circumferential and radial densities of all the 3D found that the directional crimps of the 3D circular
circular woven preforms in the r-z plane mainly woven structures slightly depended on the types of

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

Bilisik et al. 653

weave pattern. The radial crimp was also reversely depended on the types of weave pattern and the
affected by the number of layers. number of layers.
In light of these findings, we plan to conduct future
Relevance of possible results of research on 3D fully interlaced circular woven preform
three-dimensional fully interlaced structures using high-modulus fibers for use in possible
soft ballistic applications.
circular woven preform
Yarn interlacements between the axial-circumferential Funding
and axial-radial in 3D fully interlaced circular woven
This research received no specific grant from any funding
structures can improve the energy absorption capabil- agency in the public, commercial or not-for-profit sectors.
ities of 3D soft circular woven preform structures for
one-piece soft ballistic applications. Also, the axial-cir-
cumferential and axial-radial interlacements in 3D fully Acknowledgments
interlaced circular woven preform structures can also This work is the product of the principle author’s continued
result in better damping properties for semi-rigid27,28 research efforts at the various textile material research centers
(flexible) and rigid composite22,33,38 structural and labs at the University of Leeds, North Carolina State
University at Raleigh and Erciyes Universities since 1988.
applications.

References
Conclusions
1. Dow MB and Dexter HB. Development of stitched,
Three-dimensional fully interlaced circular woven pre- braided and woven composite structures. The ACT
form structures were developed. Basically, weave pat- Program and at Langley Research Center (1985 to
tern and number of layers were considered as 1997), NASA/TP-97-206234, 1997.
processing parameters. Three-dimensional plain, twill 2. Kamiya R, Cheeseman BA, Popper P, et al. Some recent
advances in the fabrication and design of three dimen-
and satin representative circular woven preform struc-
sional textile preforms: a review. Compos Sci Technol
tures were made under these parameters. The data
2000; 60: 33–47.
included the yarn-to-yarn spaces, yarn angles, yarn 3. Ko FK and Chou TW. Textile structural composites.
length and crimps in the 3D circular woven representa- New York: Elsevier, 1989.
tive preform structures. The data obtained from the 4. Chou TW. Microstructural design of fiber composites.
developed 3D circular representative preform struc- Cambridge: Cambridge University Press, 1992.
tures were measured under out-of-loom conditions in 5. Hearle JWS. Textiles for composites. Textil Horiz 1994;
the force-free environment. 11: 11–15.
We found that the yarn-to-yarn spaces in the 3D 6. Mouritz AP, Bannister MK, Falzon PJ, et al. Review of
fully interlaced structures were high compared to applications for advanced three dimensional fiber textile
those of the traditional 3D orthogonal woven struc- composites. Composites Part A 1999; 30: 1445–1461.
tures due to the directional interlacement where the dir- 7. Cox BN, Dadkhah MS, Morris WL, et al. Failure mech-
anisms of 3D woven composites in tension, compression
ectional fiber volume fraction can be affected. It was
and bending. ACTA Metall Mater 1993; 42: 3967–3984.
found that the axial angle (a) in fabric length, the cir-
8. Bilisik K. Multiaxis 3D woven preform and properties of
cumferential angle (c) and the interlaced radial angle multiaxis 3D woven and 3D orthogonal woven carbon/
(ri) in fabric circumference were due to the axial- epoxy composites. J Reinf Plast Compos 2010; 29:
circumferential and axial-radial interlacements. These 1173–1186.
angles depended on the type of weave pattern, whereas 9. Bilisik K. Multiaxis 3D weaving: comparison of devel-
the radial angle (r) in fabric length mainly depended oped two weaving methods-tube–rapier weaving versus
on take-up rate. It was also observed that the interlace- tube-carrier weaving and effects of bias yarn path to the
ment in the 3D circular woven structure caused slight preform properties. Fibers Polym 2010; 11: 104–114.
irregular axial and radial densities in fabric circumfer- 10. Khokar N. Noobing: a nonwoven 3D fabric-forming pro-
ence. All yarn lengths in the 3D circular woven preform cess explained. J Textil Inst 2002; 93: 52–74.
structures depended on the weave pattern and the 11. Bilisik K. Three dimensional (3D) weaving and braiding.
PhD Thesis, University of Leeds, Leeds, 1991.
number of layers.
12. Bilisik K. Multiaxis three dimensional (3D) weaving for
The interlacement on three yarn sets results in axial composites: a review. Textil Res J 2012; 82: 725–743.
crimp, circumferential crimp and radial crimp. 13. Brandt J, Drechsler K and Filsinger J. Advanced textile
Probably the unique feature of this research was the technologies for the cost effective manufacturing of high
radial crimp in the 3D circular woven preform. In add- performance composites. In: RTO AVT specialist meeting
ition, it was found that the directional crimps of the 3D on low cost composite structures, Norway, RTO-MP-
fully interlaced circular woven structures slightly 069(II), 2001.

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015

654 Textile Research Journal 84(6)

14. Mohamed MH. Three dimensional textiles. Am Sci 1990; and circular woven preform (radial crossing weaving).
78: 530–541. Textil Res J 2009; 79: 1067–1084.
15. Chen X, Taylor LW and Tsai LJ. An overview on fabri- 27. Fukuta K, Onooka R, Aoiki E, et al. Three dimensionally
cation of three-dimensional woven textile preforms for latticed flexible structure composites. Patent No 4336296,
composites. Textil Res J 2011; 81: 932–944. USA, 22 June 1982.
16. Crawford JA. Recent developments in multidirectional 28. Fukuta K, Aoiki E, Nagatsuka Y and Kitano T. Method
weaving. NASA Conference Publication No. 2420, 1985. for formation of three dimensional woven fabric and appar-
17. Bilisik K. Multiaxis three dimensional (3D) circular atus thereof. Patent 4615256, USA, 7 October 1986.
woven preforms – ‘‘radial crossing weaving’’ and 29. Khokar N. 3D-Weaving: theory and practice. J Textil
‘‘radial in-out weaving’’: preliminary investigation of Inst 2001; 92: 193–207.
feasibility of weaving and methods. J Textil Inst 2010; 30. Tan P, Tong L and Steven GP. Modelling for predicting
101: 967–987. the mechanical properties of textile composites-A review.
18. Bilisik K. Three dimensional braiding for composites: a Composites Part A 1997; 28A: 903–922.
review. Textil Res J 2013; 83: 1414–1436. 31. Naik RA. TEXCAD - Textile composite analysis for
19. Emerson PD, Davis J, Oakfield JC, et al. Circular weav- design, NASA Contractor Report 4639, 1994.
ing apparatus, products and process. Patent 3719210, 32. Cox BN and Dadkhah MS. The macroscopic elasticity of
USA, 1973. 3D woven composites. J Compos Mater 1995; 29:
20. Banos J, Cantagrel JC, Cahuzac G, et al. Method and 785–819.
machine for 3D weaving for obtaining woven hollow 33. Cox BN and Flanagan G. Handbook of analytical meth-
reinforcements of revolution. Patent 4183232, USA, 1980. ods for textile composites, NASA Contractor Report
21. Bompard B, Brochier J and Bruyere A. Method of and 4750, 1997.
apparatus for the production of bodies of parts of 3D 34. Pochiraju K and Chou TW. Three-dimensional woven
fabric. Patent 4080915, USA, 1978. and braided composites - I. A model for anisotropic stiff-
22. Rolincik PG. AutoWeave, a unique automated 3D weav- ness prediction. Polym Compos 1999; 20: 565–580.
ing technology. SAMPE Q 1987; 32: 40–47. 35. Verpoest I and Lomov SV. Virtual textile composites
23. Yasui Y, Anahara M and Omori H. 3D fabric and method software WiseTex: integration with micro-mechanical,
for making the same. Patent 5091246, USA, 1992. permeability and structural analysis. Compos Sci
24. Bilisik AK. Multiaxial three dimensional (3D) circular Technol 2005; 65: 2563–2574.
woven fabric. Patent 6129122, USA, 2000. 36. Buchanan S, Grigorash A, Quinn JP, et al. Modelling the
25. Bilisik AK. Multiaxial three dimensional (3D) circular geometry of the repeat unit cell of three-dimensional
weaving and multiaxial 3D circular woven preforms for weave architectures. J Textil Inst 2010; 101: 679–685.
composites. In: Haddad YM and Firstov M (eds) pro- 37. Khokar N. Woven 3D fabric material. Patent 6338367,
ceedings of NATO advanced research workshop series, USA, 15 January 2002.
Advanced multilayered and fiber reinforced composites, 38. Bilisik K, Karaduman NS, Bilisik NE, et al. Three dimen-
Kluwer academic publisher, June 1998. sional (3D) fully interlaced woven preforms for compos-
26. Bilisik K and Mohamed MH. Multiaxis three dimen- ites. Textil Res J, Epub ahead of print 21 May 2013. DOI:
sional (3D) flat woven preform (tube rapier weaving) 10.1177/0040517513487791.

Downloaded from trj.sagepub.com at FRESNO PACIFIC UNIV on January 4, 2015