Composites: Part A 43 (2012) 1717–1729

Contents lists available at SciVerse ScienceDirect

Composites: Part A
journal homepage: www.elsevier.com/locate/compositesa

Design and manufacturing of an L-shaped thermoplastic composite beam

by braid-trusion
Louis Laberge Lebel ⇑, Asami Nakai
Department of Advanced Fibro Science, Kyoto Institute of Technology, Matsugasaki, Sakyo-Ku, Kyoto 606-8585, Japan

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

Article history: Braid-trusion is a manufacturing process for composite materials in which a braiding machine is coupled
Received 6 April 2011 with a pultrusion die to continuously produce beams with constant cross-section and off-axis fiber ori-
Received in revised form 25 January 2012 entation. This study presents a geometrical model of the tri-axial braid which allows the design of
Accepted 1 March 2012
braided preforms that achieve correct filling of the pultrusion die at the same time as it limits fiber fric-
Available online 10 March 2012
tion on die walls. The typic design parameters are listed and used for the manufacturing of a braid-truded
thermoplastic composite beam where fibers are aligned at ±69° as well as 0° with respect to the beam
Keywords:
axis. Tensile mechanical characterization and cross-section observations are also presented.
Braiding
Pultrusion
Ó 2012 Elsevier Ltd. All rights reserved.
Thermoplastic resin
Consolidation

1. Introduction adjusted to a given die geometry so that the composite is properly

Braiding is a traditional textile technique that has found several
new industrial applications. High modulus fibers, such as carbon fi-
bers, can be braided and consolidated with a polymer matrix to
form structural members that possess outstanding mechanical
properties. Braided composites have received great attention from
the automotive industry for its enormous potential as energy
absorption members in crushing situations [1].
However, to achieve successful implementation of braided
structures in the automotive industry, several challenges must be
overcome without sacrificing the product performance. The most
important one is achieving high production rate. Another challenge
with increasing importance is the production of environmentally
friendly primary structures. Thermoplastic composite pultrusion
is a continuous manufacturing process that produces constant
cross-section beams of structural quality using recyclable matrices
[2]. Inorganic fiber reinforcement (e.g., glass fiber, carbon fibers) is
pulled through a heated die jointly with a thermoplastic polymer
matrix. The heated die has a taper region where the molten poly-
mer impregnates the fiber bed. There is also a constant cross-sec-
tion region that gradually cools the matrix and forms the beam
at its constant cross-section dimensions. The operating parame-
ters, i.e. the pull speed and the die temperature, have to be
⇑ Corresponding author. Address: 23-102 St-Joseph East Blvd. Montreal, QC,
Canada H2T 1H3. Tel.: +1 514 849 3842.
E-mail addresses: louislabergelebel@gmail.com (L.L. Lebel), nakai@kit.ac.jp
(A. Nakai).
consolidated at die exit [3].
The fibers are usually oriented parallel to the direction of the
pultrusion. Few processes exist to control the fiber orientation of
pultruded products. Some early works reported the pultrusion of
braided fabrics, hence, giving a possibility of introducing oriented
fibers with respect to the beam axis in this continuous process
[4,5]. The ‘‘braid-trusion’’ combines a braiding machine as the
material feeding system with the pultrusion die system. However,
recent investigations have demonstrated that the braided preform
must be carefully designed to be able to control the fiber orienta-
tion in the braid-truded beam [6].
Several models have been developed to describe the braid
geometry. Brunnschweiler adapted Pierce’s theory for woven fab-
rics to braided fabrics in the 1950s with the objective of estimating
the crimp path of braided yarns and the locked configurations [7–
9]. The ‘‘Handbook of Industrial Braiding’’, published in 1989, also
provides good indication of the important parameters of braided
fabrics using simple equations [10]. Pastor et al. developed a mod-
eling procedure of textile reinforcements using the Bezier curve
and the Lagrange’s principle of minimum work [11]. Robitaille
et al. developed a formal procedure to represent the interlacing
path of yarns in braided fabrics based on a series of vectors in
the fabric unit cell [12]. Other studies presented a computer pro-
gram and an experimental verification for modeling the braided
fabric meso-structure [13,14]. These models provide good approx-
imation of the braided fabric structure but are often mathemati-
cally complex or hidden behind proprietary code. Moreover, the
braid structure is always represented by a flat unit cell, which does

1359-835X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.compositesa.2012.03.001
1718 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729

not represent the circular nature of tubular braids. Also, these BY fiber angle (hin), the average braid angle (hBraid), and the outside
models do not evaluate the deformation of the braided fabric when BY fiber angle (hout), which are all calculated with the same braid
it is pulled through a pultrusion die. Hence, there is a need to de- pitch value.
velop a simple geometrical braid model that can be readily adapt-    
pDby;in pDBraid
able to braid-trusion. hin ¼ arctan < hBraid ¼ arctan < hout
p p
This study presents a geometrical model for tri-axial tubular  
braids. This model is then used to describe the phenomena occur- pDBy;out
¼ arctan ð1Þ
ring in the braid-trusion manufacturing process. Using this theo- p
retical description, a design procedure is proposed for the
The braiding machine controls the braid pitch. The pitch dis-
successful design of braided fabric for the braid-trusion manufac-
tance is determined by the ratio of the pulling speed of the braid
turing process. Finally, an experimental braid-trusion is performed
structure over the rotation speed of the BY carriers around the
using the suggested design procedure.
braider head [15]. The braid diameter can be controlled using a
forming mandrel. If no mandrel is used, the braid diameter will re-
1.1. The braid model duce until the braid reaches a ‘‘locked-state’’, often referred to as
the tensile jammed position [8]. At this locked-state, the contact
The three-axial tubular braided fabric possesses yarns continu- between compressed BYs and MEYs prevent the further reduction
ously spiraling along the braid axis, called the braiding yarns (BYs). of the braid diameter when the preform is subjected to a pulling
The BYs are divided into two sets intertwined in opposite direc- force on the axial direction.
tions along the braid axis. Another set of yarns, called the mid- Fig. 1b shows the BY path in one braid repeat length. The BY
dle-end-yarns (MEYs), is laid inside the braid along the axis path is assumed here to be constituted of a cosine that can be ex-
direction and is locked-in by the two sets of BYs. The MEYs are pressed using Eq. (2), where the value NBY is the total amount of
introduced for dimensional stability, and longitudinal mechanical BYs.
properties improvement [10]. Finally, a third set of yarns, called  
ðbBy þ bMEY Þ pNBY cos h 2p
the core yarns (CYs) can be placed inside the braid along the braid y¼ cos ; with 0 6 x 6 ð2Þ
2 2 NBY cos h
axis. All the yarns are assumed to have an elliptic cross-section
characterized by a given width and thickness (aY, bY). In the regular The maximum crimp angle a is calculated at half of the plait
braided fabric, each BY continuously passes under two other yarns length from the derivative of the BY cosine path, as shown by Eq.
and then over two yarns of the opposing set. Fig. 1 shows the (3).
braided architecture and a representation of the BY path.  
NBY
The pitch length (p) is the distance for one BY to execute a com- a ¼ arctan ðbBY þ bMEY Þp cos h ð3Þ
4p
plete rotation around the braid axis (see Fig. 1a). Fig. 2 introduces
the several types of braid diameters that can be defined. The inter- The length of the cosine braiding yarn centerline of Eq. (2) can
nal diameter (Din) is also equal to the mandrel diameter when the be calculated using numerical integration (see Appendix A). The
braid is formed onto a cylindrical support. The diameter of internal crimp ratio of the BY can be calculated for one repeating unit of
BYs (DBY,in) is the diameter of the circle passing at the middle of the the cosine curve according to Eq. (4). The length of the cosine
innermost BYs. The braid diameter (DBraid) is the diameter of the braiding yarn path (LBY) is divided by the spiral length of the braid-
circle passing through all the MEY centers. The diameter of outside ing yarn projected onto the average braid diameter cylinder.
BYs (DBY,out) is the diameter of the circle passing through the out-
LBY cos h
ermost BYs’ centers. Finally, the outside diameter (Dout) is the Rc ¼ ð4Þ
p
diameter of the virtual cylinder enclosing the braid on its external
surface. Defining different braid diameters leads to the definition of The unconsolidated yarn cross-section area (AY) refers to the
different fiber angles depending on which diameter the fiber is yarn material area (SY) by a yarn fiber packing ratio (RP) according
positioned. Eq. (1) shows three different fiber angles: the internal to Eq. (5). The yarn fiber packing is assumed here to be between 0.5

Fig. 1. Braid architecture: (a) localization of the braiding yarn, middle-end yarn and core yarn for one pitch length and (b) the fiber path is represented by a cosine. (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729 1719

Fig. 2. Definition of the braid diameters. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

and 0.7. The unconsolidated yarn cross-section area can also be cal-
culated using the area of its elliptical shape and using the yarn
dimensions.
SY paY bY
AY ¼ ¼ ð5Þ
RP 4 that cuts the BYs between the MEYs. This gap depends on the
Note that in Eq. (5), both letters ‘‘A’’ and ‘‘S’’ represent surface quan-
tities, but ‘‘A’’ represents the unconsolidated area (including the
spacing between fibers in the yarn) and ‘‘S’’ represents the material
area. The same logic applies for other surface area quantities in this
text.
It is generally accepted that the specific volume of the yarns in-
side the fabric remains constant. However, the cross-section
dimensions of the yarn are assumed to vary depending on the
yarn’s position inside the fabric [4]. That is because the constraints
applied on one yarn by the surrounding yarns also vary inside the
braid structure depending on the yarn’s position. The constraints
applied on one yarn come from several sources such as tension,
bending, and friction. However, the approach chosen here will be
purely geometric and rely on the critical points inside the braid
where the yarns are subject to contact with one another. At these
critical points, the yarn cross-section dimensions are modified in
the event of a contact. If there is no contact between the yarns at
a particular critical point, a gap between yarns is then created.
Fig. 3 describes the selected critical points and the possible gaps in-
side the braided fabric.
The first critical gap is situated along the braid axis on a plane
width of the BY situated at mid-diameter (aBYP), the pitch, the num-
ber of BY (NBY) and the braiding angle (h). This gap is calculated by
following equation.
p aBYP
GapBYP ¼  ð6Þ
NBY sin h
The second critical gap separates the BYs at the surface of the braid
and depends on the BY width at surface (aBYS), the braid angle, and
the pitch divided by half of the amount of BYs. This gap is calculated
by following equation.
2p aBYP
GapBYS ¼  ð7Þ
N BY sin h
The third gap describes the separation between the BYs and MEYs at
the braid diameter. This gap depends on the BYs crimped section
projected on the cylindrical surface having a diameter equal to

Fig. 3. Description of the gaps between yarns in the braid. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this
article.)
1720
the braid average diameter. This BY section has a width equal to the
BY width at pitch, but the thickness is the BY thickness at pitch
(bBYP) projected on the crimp angle. This gap, calculated by
Eq. (8), is the difference between the circumference of the braid
divided by half the number of BYs, the width of the MEY and the
distance between two parallel lines enclosing the crimped section
of the BYs.
2pDBraid
GapBYMEY ¼  aMEY
NBY
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
!2
u
u b
2
t
 a2BYP cos2 hBraid þ BYP
sin hBraid
sin aMax
These three gaps can be used to calculate yarn cross-section dimen-
sions using a calculation algorithm (see Appendix B). The necessary
initial data are BY and MEY unconsolidated area as well as their
maximum and minimum aspect ratios. The braid diameter, the
pitch and the number of BY are also needed. The first step is to
set all the yarn dimensions to the maximum aspect ratio dimen-
sions. Then, in a calculation loop, the crimp angle and new yarn
dimensions are calculated using Eqs. (6)–(8) and setting the gaps
to zero. The dimensions are compared to maximum and minimum
aspect ratio dimensions. If the dimensions are outside the minimum
to maximum aspect ratio range, the new dimensions are then set to
the closest aspect ratio extreme. The values for the next iteration
are the average between the values from the previous and the
current calculation step. This loop is repeated until the yarn
dimensions do not significantly change from a calculation step to
another.
1.2. Braid-truded beam material content
The material cross-sectional area within the preform is the vol-
ume of material in one pitch length normalized by the pitch length,
as calculated by following equation.
NBY SBY RC
Spreform ¼ þ NMEY SMEY þ NCY SCY
cos hBraid
Fig. 4. Expansion of the circumference of the consolidated braid over the braid perimeter value upon consolidation in the pultrusion die. (For interpretation of the references
to colour in this figure legend, the reader is referred to the web version of this article.)
L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729
The factors influencing the braided preform cross-section area
are the total yarn filaments cross-section area (SBY, SMEY, SCY), the
number of yarns in each layer (NBY, NMEY, NCY), the braid angle,
and the crimp ratio. The total yarn filament cross-section area is
calculated using the fineness (F, in tex or g/km) divided by the den-
sity (q) of each of the reinforcement and matrix constituents of the
yarn.
Ff Fm
SY ¼ Sf þ S m ¼ þ ð10Þ
qf qm
ffi
The preform material area calculated using Eq. (9) can be nor-
2
ð8Þ
malized by the die cross-section area to determine the die filling
ratio (RF). The fiber volume content calculation of a braid-truded
beam is obtained by a development similar to Eq. (9), dividing
the reinforcement area of the preform by the die cross-section
area, as shown in following equation.
 
NBY Sf ;BY RC 1
V f ;beam ¼ þ NMEY Sf ;MEY þ N CY Sf ;CY ð11Þ
cos hBraid ADie
1.3. Deformation of the braid during pultrusion
During pultrusion, the external fibers suffer extreme friction
onto the die walls. This friction can force the BYs to slide onto
the MEY on the axial direction of the braid, hence increasing the
pitch length of the braid. This phenomenon has been observed in
previous studies [6] and can lead to a dramatic stop of the braid-
trusion process. By changing the pitch length, the material exiting
the die will be less than the material entering the die, causing an
accumulation of BY at the die entrance. This material accumulation
at the die entrance will sooner or later stop the braid-trusion pro-
cess. To fix the pitch length of the braid when passing through a
pultrusion die, the preform can be braided at the locked-state. At
this state, the tension and bending forces applied on the yarns,
the compression of the yarn cross-section, and the friction forces
generated at fiber–fiber contact all contribute to an interlocking
ð9Þ action of the BY onto the MEY.
 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729
When compressed inside the pultrusion die, the thickness of the
braid, and so the amplitude of the BY cosine path, will reduce. As
the pitch length does not change, the braid structure spreads along
its circumferential direction to accommodate the constant yarn
length. The expansion of the braid circumference from the com-
pression action is shown in Fig. 4. This phenomenon implies that
the yarns used in the braided pultruded preform have to be able
to spread during consolidation. It has been observed that jute tows,
spun and twisted from stable fibers, break during pultrusion be-
cause the tows could not deform from a circular cross-section
shape to a flattened elliptical shape [16].
Another implication of the flattening braid structure is that the
fiber angle in the pultruded beam will be higher than the fiber an-
gle in the braid. The spreading along the circumferential dimension
and the braiding angle increase are dependent on the compression
ratio of the braid, which is the consolidated braid thickness divided
by the braid thickness. Calculation software is again needed to find
this consolidated braid circumference and consolidated fiber angle,
following a calculation algorithm provided in Appendix C.
1.4. Modulus calculation
The modulus is calculated using the classical lamination theory
[17]. The fiber and matrix properties, the beam fiber volume frac-
tion, the beam fiber angle and the area of every component of the
beam are needed. Eleven laminae are needed in the following ori-
entation sequence: [h°, h°, 0°, h°, h°, 0°, h°, h°, 0°, h°, h°].
2. Design of a braid-truded beam
The obvious criteria for the design of braid-truded beams are
the mechanical requirements, as well as the shape and outside
dimensions of the beam cross-section. Depending on the case,
the importance of each of these criteria can vary. For example, if
a die of fixed dimensions is already available or the beam cross-
section shape and dimensions are already fixed, all other parame-
ters must be adjusted to reach the desired mechanical properties.
However, several constraints must be considered. First, the yarn
precursors, such as commingled yarns, powder impregnated yarns
or parallel hybrid yarns, usually have a fixed fineness and rein-
forcement/matrix ratio when they are acquired from the manufac-
turer. Second, the die filling ratio must be over 100% to achieve
resin flow that will impregnate the fibers. If the filling ratio is too
small, only the reinforcing fiber bed will get compressed during
its passage in the consolidation die. Third, the number of braided
layers should be kept to a minimum for easy manufacturing. Each
braided layer needs an independent circular braider head or a ma-
chine having a complicated braiding tracks arrangement. Forth, as
stated earlier, the braid must be formed at the tensile locked-state.
Finally, the braid must be designed with a circumference
ðCBraid;Preform Þ that spreads during consolidation and reaches the cir-
cumference of the braid layer in the final beam ðCBraid;Beam Þ. This can
be represented by following equation.
CBraid;Preform < CBraid;Consolidated ¼ CBraid;Beam
This will ensure that the BY fibers have an adequate length for
the die dimensions. If the BY fibers are too long, friction on the die
walls increases dramatically impairing the pultrusion process. This
also destroys the braid architecture by changing the local fiber an-
gle or even the pitch length. Conversely, if the consolidated braid
perimeter is shorter than the layer perimeter, the reinforcing fibers
do not occupy the whole beam cross-section.
1721
2.1. Design example
To demonstrate how the geometrical model presented can
guide the design of pull-truded beams, an L-shape cross-section
of 30 mm side length and 3 mm thickness was used. The BYs avail-
able for braid-trusion were composed of a 400 tex carbon tow and
a 560 tex nylon tow. The MEYs were composed of a 1200 tex car-
bon tow and a 700 tex nylon tow. Only one braided layer was used
and the interior of the braid was filled with CYs composed of a
400 tex carbon tow and a 560 tex nylon tow. The shape of the L-
section and material arrangement are schematized in Fig. 5.
The thickness of the braided layer can be calculated from the
braided layer material area (two first terms of Eq. (9)) and the L-
Shaped side dimension, as shown in the following equation:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
C Die  C 2Die  SBraid;Beam
tBraid;Beam ¼ ð13Þ
2
The braided layer circumference in the L-shape (CBraid,Beam) is calcu-
lated using following equation.
CBraid;Beam ¼ 4cDie  4t Braid;Beam ð14Þ
The model was used to calculate the tensile modulus, the die fill
ratio and the beam fiber angle and the braid diameter for 25 differ-
ent pitch conditions (from 40 mm to 160 mm) and 21 different CY
amounts (from 70 to 90), giving a total of 525 braiding conditions.
In every condition the braid diameter was adjusted to obtain a
match between the consolidation circumference and the braided
layer circumference in the beam (CBraid,consolidated = CBraid,Beam). This
calculation was performed using the algorithms provided in the
appendices. The results are shown in Fig. 6 with four 3D graphs
of 15 continuous lines on the pitch axis for several amounts of CYs.
The Beam Fiber Angle varies almost linearly with the pitch dis-
tance from 35° to 70° for all the different amounts of core yarns
(see Fig. 6a). This is well explained by the fact that the calculated
outside braid circumference for all conditions does not vary much
in the 111 mm to 118 mm range (see Fig. 6b). The fill ratio var-
ies from 70% to 120% (see Fig. 6c). The highest fill ratio is
achieved using 90 CY and the lowest pitch of 40 mm. The predicted
beam tensile modulus varies from 37.5 GPa to 52.5 GPa (see
Fig. 6d). The highest predicted tensile modulus is obtained with
the highest amount of CYs (90) and the highest pitch distance
(160 mm). This result shows that the fiber angle has a stronger ef-
fect on the tensile modulus than the fill ratio over the range of con-
ditions evaluated.
ð12Þ
Fig. 5. Geometry and layout of the braid-truded beam.
1722 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729
Fig. 6. 3D plots of calculated braid-trusion conditions according to the braid model for different pitches and different CY amounts: (a) fiber angle in the beam, (b) necessary
braid outside circumference to achieve correct filling of the pultrusion die, (c) fill ratio, (d) predicted tensile modulus. In every graph, the mark (O) represents the selected
condition in the experimental section.
Fig. 7. Pultrusion die geometry and pultrusion settings.
Even though several braid conditions are evaluated and pre-
sented in Fig. 6, only one condition can be successfully braid-
truded due to the necessity of braiding at the low-lock condition
to prevent fiber angle change in the pultrusion die. This low-lock
condition can be found in an experimental procedure where braids
are done at different pitches and the external diameter of the braid
is measured to build a calibration curve.
3. Experimental
The reinforcing fibers were 400 tex carbon tows (T700SC-6000-
50C, Toray Industries, qf = 1.8 g/cc) and the polyamide 6,6 (PA66)
matrix was also available in the form of fibers (Leona, Asahi Kasei
Fiber, qf = 1.14 g/cc), bought in tows of 140 tex. Prior to the braid-
ing, yarns were obtained by mixing few tows of reinforcement and
fibers in a parallel hybrid arrangement. The BYs were composed of
one carbon tow and four PA66 tows for a total BY fineness of
960 tex. The MEYs were obtained mixing three carbon tows and
five PA66 tows for a total fineness of 1900 tex. The CYs were the
same as the BYs.
Five single layer braids without mandrel were produced using
48 BYs and 24 MEYs in a regular braid arrangement. The braiding
machine (Muratec, Murata Machinery) had a constant yarn carrier
rotation speed of 1.75 RPM. The braid take-up set speed varied
from 70 mm/min to 105 mm/min. The braiding outside circumfer-
ence was measured wrapping a tape over the braided preform
right after the braiding operation, without releasing the tension
in the braid. A mark was done on the tape prior to wrapping; the
tape was then enrolled without tension over the braid until the
tape was superposing itself. Another mark was done over the first
one and the circumference was taken as the distance between the
two marks on the unrolled tape. The braiding pitch was measured
as the distance along the mandrel length for one fiber to make a
complete revolution around the mandrel. From each of the braids,
a pitch length was cut and weighed to evaluate the braid fineness
(g/km).
A 2-m preform composed of 48 BY, 24 BY, and 75 CY was then
braided for pultrusion. The braider rotation speed was 1.75 RPM
and the braid take-up speed was set to 79 mm/min. The pitch
and the outside circumference of the preform were measured.
Fig. 7 shows the schematic of the L-Shape pultrusion die geom-
etry, consisting of the preheater, the pultrusion die and an air
blower placed at the die exit. The braided preform was pulled
through this forming assembly by a pulling machine. The ‘‘L’’ thick-
ness dimensions were large in the preheater, reduced at the con-
solidation die entrance and gradually further reduced in the
 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729
taper region until a final thickness of 3 mm was reached. The
whole system had a total of eight separate heating zones. The pre-
heating temperature is set to be slightly under the melting point of
the PA66 (260 °C). The entrance region is set to the processing tem-
perature of 300 °C. Towards the end of the pultrusion die, the tem-
perature is decreased to 270 °C. An air-cooling system was used to
quench the pultruded material immediately after exiting the die. A
video of the braid-trusion system is available and accompanies the
electronic version of this manuscript.
The braided preform was pultruded at a speed of 21.4 mm/
min. The temperature both at the surface and inside the beam
was recorded using thermocouples (type K) and a data acquisition
system (DCS-100A, Kyowa). The fiber pitch was measured on the
beam after pultrusion at every 20 mm by measuring the axial dis-
tance of 24 individual carbon tows. The cross-section images of all
the pultruded beams were acquired with a microscope (PME3,
Olympus) equipped with a digital camera (Powershot A540, Ca-
non). Rectangular coupons of 20 mm  200 mm were cut from
each pultruded beam for tensile testing (55R4206, Instron). The
coupons were fixed in the hydraulic gripping heads using tabs of
layered stainless steel mesh (100 count) and an aluminum plate
of 1 mm thickness. The constant speed of testing was set to
1 mm/min. Deformation of samples was recorded using strain
gages (KFG-10-120-C1-11, Kyowa). All reported experimental val-
ues were averaged over three measurements.
4. Results and discussion
4.1. Braid model verification
Pitch, outside circumference and fineness measurements made
on the five braids during the calibration experiments are presented
in Table 1. Using the pitches and outside circumferences, braid
finenesses have been calculated using the model for a compaction
ratio of 0.7. The percentage error between the model and the
experimental value is also shown in the table.
The measured braid finenesses correspond well to the values gi-
ven by the model, with the mean percentage error around 2.91%.
This good comparison between the measured and calculated val-
ues indicates that the evaluation of the crimp of the model is close
to reality. It is possible to adjust the yarn compaction ratio param-
eter of the model to reach calculated fineness values closer to real-
ity. If the value of 0.58 is used instead of the 0.7 compaction ratio
that was used for the calculation of Table 1, the mean percentage
error decreases to a near zero value.
The outside circumference values can be fitted with a quadratic
equation to obtain an equation of the outside diameter function of
the braid pitch. The calibration curve drawn from this equation,
shown in Fig. 8, was then used to find the braid pitch settings that
would produce a braided preform that fits the pultrusion die. Here,
‘‘fit’’ relates to the concept of matching the consolidation circum-
ference of the braid to the circumference of the braid in the pultru-
sion die (CBraid,Consolidated = CBraid,Beam). The selected condition for
the braid-trusion experiment was found using the braid model
and is presented in Table 2. This braiding condition is also shown
Table 1
Measurements on braiding calibrations experiments and comparison with the tri-axial braid model.
Measured pitch (mm) Measured outside circumference (mm)
38.5 ± 0.1 128.3 ± 0.8
44.2 ± 0.3 112.2 ± 0.3
49.8 ± 0.3 107.0 ± 0.5
53.8 ± 0.3 102 ± 1
59.7 ± 0.3 96.8 ± 0.3
1723
Fig. 8. Quadratic fit of the braid outside circumference as a function of the braid
pitch. The square mark on the Quadratic fit curve is the selected experimental
condition for braid-trusion.
by a circle in Fig. 6. The braid selected had a pitch of 44.8 mm
and an outside circumference of 113.5 mm according to the cali-
bration curve (see the square mark in Fig. 8). The braid circumfer-
ence at the average braid diameter is 104.7 mm giving an average
fiber angle in the braid of 66.8°. At such a high angle, the crimp ra-
tio had a value of 1.104, indicating a highly compact braid struc-
ture with great fiber waviness. The preform fineness, including
the 48 BYs, the 24 MEYs and the 75 CYs was 246,920 tex. The pre-
form area was 180.4 mm2, higher than the 171 mm2 area of the die
exit. Finally, the preform fiber volume content is 35%. The model
also calculates the properties of the braid-truded beam. The pitch
of the braided fiber was also 44.8 as the BYs should not move along
the beam axis direction during pultrusion. The braid consolidation
circumference was 115.6 mm. This value was higher than the braid
average circumference, successfully modeling the spreading of the
braid around the circumferencial direction during the braid com-
pression in the pultrusion die. The braid circumference in die
was equal to the braid consolidation circumference. This matching
is the result of the convergence algorithm (see Appendix C). The
braided fiber angle increased to 68.8° due to the spreading of the
braid along its circumferential direction without changing the
pitch length. The braid layer, including the BYs and the MEYs,
had a thickness of 1.08 mm in the beam. The CY layer, at the center
of the beam, had a thickness of 0.84 mm. Consequently, twice the
thickness of the braid layer, plus the thickness of the CY layer was
equal to the L-shape thickness of 3 mm. The die fill ratio was 105%,
as the calculated preform material area was higher than the die
exit area. This fill ratio increased the fiber volume content to
37%, because only matrix is flushed out of the die entrance. Finally,
according to calculation, the pultruded beam had an axial tensile
modulus of 39.3 GPa and perpendicular (hoop) tensile modulus
of 42.2 GPa. The hoop modulus was higher than the axial modulus
due to the high braid angle.
Measured fineness (tex) Calculated fineness (tex) Percentage error on fineness ()
214,793 209,171 2.7%
179,670 175,777 2.2%
163,180 158,528 2.9%
153,180 148,013 3.5%
141,704 137,249 3.2%
1724 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729

Table 2
Yarn types, preform parameters, and braid-truded beam characteristics calculated with the braid geometrical model using the outside circumference of 113.5 mm.

Type (–) Reinforcement Matrix Fiber volume content (–) Amount (–)
2 2
Material (–) Fineness (tex) Area (mm ) Material (–) Fineness (tex) Area (mm )
Yarns
BY Carbon 400 0.222 Nylon 620 0.491 31% 48
MEY Carbon 1200 0.667 Nylon 700 0.614 52% 24
CY Carbon 400 0.222 Nylon 620 0.491 31% 75

Pitch Preform outside circ. Braid average circ. Fiber angle By crimp ratio Preform fineness Preform area Fiber volume content
(mm) (mm) (mm) (°) (–) (tex) (mm2) (–)
Preform
44.8 113.5 104.7 66.8 1.104 246,920 180.4 35%
Braid consol. Braid circ. in Fiber Braid lay. Core lay. Fill Fiber volume Beam axial tensile Beam perp. tensile
circ. (mm) die (mm) angle (°) thick. (mm) thick. (mm) ratio (–) content (–) modulus (GPa) modulus (GPa)
Beam
44.8 115.6 115.6 68.8 1.08 0.84 105% 37% 39.3 42.2

4.2. Braid-trusion
The measurement of temperature inside the braided preform
during pultrusion is shown at Fig. 9. The solid line represents the
temperature settings, also indicated in Fig. 7.
The surface and inside temperature reached a plateau at
200 mm from the entrance of the preheater, which corresponds
to 40% of the length of the preheater. The value measured was
250 °C at the surface of the preform and 230 °C inside the pre-
form. These temperature values were slightly under the set pre-
heater value of 260 °C, probably due to heat losses from the
preheater openings. After entrance in the die, the braid reached
the set process temperature of 300 °C at 135 mm of the die
entrance. It can then be estimated that the process time occurred
from this point over to the end of the taper region because pressure
is applied on the preform mostly when the die cavity changes
dimensions. A 5 min and 22 s flow time was calculated. After the
250 mm taper region, the temperature of the beam gradually
dropped to 270 °C. The exit temperature slightly over the melting
temperature of the nylon ensured that no matrix would stick on
the die walls, which would have caused sloughing imperfections
[18]. After the die exit, the temperature of the beam quickly
dropped to room temperature due to contact with cool airflow.
Table 3 shows the measurements of the preform before pultru-
sion and from the beam after pultrusion. The braid pitch and out-
side circumference were slightly higher than the values
recommended by the calibration procedure. This almost insignifi-
cant difference is attributed to the tension variation in the yarns
caused by the braiding machine. The pitch was also measured after
pultrusion, directly on the beam surface. The pitch dimension did
not change when considering the standard deviation on the mea-
surements. The tensile modulus was 36 GPa and the tensile
strength was 469 MPa. The tensile modulus achievement ratio
can be calculated as a ratio of the measured value over the pre-

Fig. 9. Temperature tracking of the material at the surface and inside the preform
Fig. 10. Cross-section image of the braid-truded beam.
during pultrusion.

Table 3
Braid-trusion preform and pultrusion characterization.

Property Braid pitch (mm) Braid outside circ. (mm) Beam pitch (mm) Tensile strength (MPa) Tensile modulus (GPa)
Calculated theoretical values (see Table 2) 44.8 113.5 44.8 N/A 39.3
Experimental values 43.8 ± 0.3 110.5 ± 2.1 46.4 ± 4.7 469 ± 40 36 ± 2
 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729 1725

dicted value (see tensile modulus column in Table 3). In this case, p Braid pitch
the achievement ratio was 92%, indicating a good approximation of NBY Amount of BY
the fiber orientation, and the fiber layout in the beam by the model n Number of subdivision for BY length calculation
presented in this study. Output variables
Fig. 10 shows the entire cross-section of the beam. The fibers L Yarn length in one rotation around the mandrel
are well spread throughout the cross-section of the beam. Little re- RC Crimp ratio
sin-rich regions are present at the L-shaped beam tips. A die design
with rounded corners would be preferable to accommodate the cir-
cular nature of the braid. This would allow the braid to deform bet-
ter and completely fill the die cross-section. Voids are visible at Algorithm
several locations. This suggests that the resin flow time was insuf- Initialize the first yarn horizontal coordinate to 0
ficient to fill the fiber bundles. A better mixing of the matrix and xi=1 = 0
reinforcement filaments would be desirable to reduce the flow dis- Calculate the first yarn vertical coordinate
tance to achieve full impregnation [19]. A void free beam might bBYS þ bMEY
yi¼1 ¼
then be obtained with the same manufacturing conditions. The 2
For the rest of the n subdivisions i for the calculation of the
higher concentration of voids at the tips of the L-shape indicates
length (1 < i 6 n)
that some macro-flow occurred around the fiber bundles. How-
Calculate the next horizontal coordinate
ever, this marco-flow was not sufficient to flush excess air bubbles
2p i
from the beam before the end of the taper area of the pultrusion xi ¼ 
N BY cos hBraid n
die. Again, a better mixing of the resin and the reinforcement fibers
Calculate the next vertical coordinate
prior to the braiding operation would reduce air entrapment in the  
ðbBYS þ bMEY Þ i
beam due to a faster impregnation. Finally, the surface finish ap- yi ¼ cos p
2 n
pears to be rough on the cross-section. This is explained both by
Calculate the incremental length, and add the result to the
the presence of air bubbles escaping through the surface when
previously calculated fiber length
the beam exits the pultrusion die and matrix shrinkage. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
L ¼ L þ ðyi  yi1 Þ2 þ ðxi  xi1 Þ2
5. Conclusion Calculate the total BY length in one rotation around the
mandrel
This study introduced a geometrical model to represent the tri- L ¼ L N2BY
axial braid architecture and to calculate its deformation inside an Calculate the crimp ratio
L-Shaped die during the braid-trusion manufacturing process. L
The design criteria for successful braid-trusion manufacturing RC ¼
p  cos hBraid
were listed and used to create a preform. A thermoplastic compos-
ite beam reinforced with carbon filaments, some oriented along
the axis of the beam, and also some oriented with a 68.6°, was then
Appendix B. Calculation algorithm for the BY and MEY
manufactured as an example. This beam could be produced in a
dimensions
continuous manner using the braid-trusion manufacturing process.
A better mixing of the reinforcement and matrix filaments inside
This algorithm calculates the yarn dimensions in the braid for a
the precursor yarns is desirable to achieve better consolidation.
given set of yarn and braid properties. The braid structure is regu-
Moreover, the beam surface finish could be improved with a prop-
lar (2  2).
er cooling process involving contact with a forming/cooling die.
The model and the design procedure developed in this study could
Input variables
help manufacturers, such as those from the automotive industry, to
ABY Unconsolidated area of the BY
produce braid-truded thermoplastic composite products.
RaBY,Max Maximum BY aspect ratio
RaBY,min Minimum BY aspect ratio
Acknowledgments AMEY Unconsolidated area of the MEY
RaMEY,Max Maximum MEY aspect ratio
The authors wish to acknowledge the financial support from the RaMEY,min Minimum MEY aspect ratio
Japanese Society for the Promotion of Science, the Toyota Motors DBraid Braid average diameter
Corporation and from the New Energy and Industrial Technology p Braid pitch
Development Organization of Japan (Project No. 06A43505c). NBY Number of braiding yarns

Output variables
Appendix A. Calculation algorithm for the length of the braiding
aBYS BY width at interior surface
yarn aBYP BY width at average braid diameter
aMEY MEY width
This algorithm calculates the real braiding yarn length in one aMax Maximum BY crimp angle
rotation around the braid fabric tube and the crimp ratio of the
braiding yarn.

Input variables Algorithm

bBYS BY thickness at surface Calculate the maximum and minimum width of the BY
bMEY MEY thickness
hBraid Average braid angle
(continued on next page)
1726 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi p
4ABY RaBY;Max 4ABY RaBY;min aBYP;New ¼ sin hBraid
aBY;Max ¼ ; aBY;min ¼ NBY
p p
Calculate the maximum and minimum width of the MEY Compare with the minimum and maximum aspect ratio BY
widths
If aBYP;New < aBY;min
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4AMEY RaMEY;Max 4AMEY RaMEY;min Then aBYP;New ¼ aBY;min
aMEY;Max ¼ ; aMEY;min ¼ If aBYP;New > aBY;Max
p p
Then aBYP;New ¼ aBY;Max
Calculate the braid average angle
Calculate a new BY width at surface value
 
pDBraid sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 2
hBraid ¼ arctan
p 2pDBraid bBYP 2
aMEY;New ¼  a2BYP cos2 hBraid þ sin hBraid
NBY sin aMax
While the errors on yarn widths are still over 1010
If it is the first iteration
Compare with the minimum and maximum aspect ratio
Set the yarn width at maximum aspect ratio value
MEY widths
If aMEY;New < aMEY;min
aBYS ¼ aBY;Max ; aBYP ¼ aBY;Max ; aMEY ¼ aMEY;Max Then aMEY;New ¼ aMEY;min
If aMEY;New > aMEY;Max
For all the other iterations
Then aMEY;New ¼ aMEY;Max
Set the yarn width to the convergence values
Calculate the errors on the yarn widths

aBYS ¼ aBYS;Next ; aBYP ¼ aBYP;Next ; aMEY ¼ aMEY;Next

eBYS ¼ aBYS  aBYS;New ; eBYP ¼ aBYP  aBYP;New ;
Calculate the yarn thicknesses
eMEY ¼ aMEY  aMEY;New
Calculate the yarn widths for the next iteration
4ABY 4ABY 4AMEY
bBYS ¼ ; bBYP ¼ ; bMEY ¼
paBYS paBYP paMEY
eBYS eBYP
Calculate the max crimp angle aBYS;Next ¼ aBYS  ; aBYP;Next ¼ aBYP  ;
2 2
eMEY
  aMEY;Next ¼ aMEY 
NBY 2
aMax ¼ arctan ðbBY þ bMEY Þp cos hBraid
4p
Calculate the braid diameter passing through inside BY
centers
Appendix C. Calculation algorithm for the successful braid-
trusion manufacturing
DBY;in ¼ DBraid  bBYS  bMEY
This algorithm calculates the braid parameters to achieve suc-
Calculate of the braid angle at this diameter
cessful braid-trusion for one set of material and a L-Shaped die.
  Input variables
pDBY;in
hBY;in ¼ arctan qf Reinforcement density
p
qm Matrix density
Calculate a new BY width near the internal surface qP Yarn packing ratio
Ramax Maximum yarn aspect ratio (can be yarn specific)
Ramin Minimum yarn aspect ratio (can be yarn specific)
2p
aBYS;New ¼ sin hBY;in Ff Yarn reinforcement fineness (Yarn type specific)
NBY
Fm Yarn matrix fineness (Yarn type specific)
Compare with the minimum and maximum aspect ratio BY p Braid pitch
widths NBY Number of braiding yarns
If NMEY Number of braiding yarns
NCY Number of braiding yarns
aBYS;New < aBY;min
CDie L-Shaped die side length
Then TDie L-Shaped die ‘‘L’’ thickness
aBYS;New ¼ aBY;min Output variables
If Various output variables such as the different diameters and
circumferences of the braid both inside the preform, the BY
aBYS;New > aBY;Max yarn angle, the fiber volume contents
Then
aBYS;New ¼ aBY;Max Algorithm
Calculate a new BY width at pitch Calculate the yarn reinforcement areas
 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729 1727

F f ;BY F f ;MEY F f ;CY Calculate the error between the braid diameter target and the
Sf ;BY ¼ ; Sf ;MEY ¼ ; Sf ;CY ¼
qf qf qf braid outside diameter:
eD=DTarget  Dout
Calculate the yarn matrix areas
Calculate the diameter value for the next iteration:
F m;BY F m;MEY F m;CY
Sm;BY ¼ ; Sm;MEY ¼ ; Sm;CY ¼
qm qm qm eD
DNext ¼ DBraid 
Calculate the yarn total material areas 2

SBY ¼ Sf ;BY þ Sm;BY ; SMEY ¼ Sf ;MEY þ Sm;MEY ; SCY ¼ Sf ;CY þ Sm;CY Calculate the braid thickness:
tBraid,perform = 2bBYS + 2bMEY
Calculate the yarn unconsolidated cross-section areas Calculate the fiber length (LBY,perform) and the crimp ratio
(RC,perform) in the braided layer of the preform with
SBY SMEY SCY
ABY ¼ ; AMEY ¼ ; ACY ¼ algorithm provided at Appendix A and using the following
RP RP RP variables:
Calculate the die area  BY thickness at surface (bBYS)
 MEY thickness (bMEY)
 Fiber angle (hBraid)
ADie ¼ ð2C Die  T Die ÞT Die
 Pitch (p)
Attribute the die outside circumference to the braid target  Number of braiding yarns (NBY)
outside diameter Calculate the total consolidated fiber area of only the BYs,
only the MEYs, and only the CYs:
DTarget ¼ 4C Die
While the error on the consolidation circumference is not NBY Sf ;BY RC
under 1010 Sf ;BY;Preform ¼ ;
cos hBraid
While the error on the target diameter value is not under
1010
If it is the first iteration Sf ;MEY;Preform ¼ NMEY Sf ;MEY
Initialize the braid diameter to be equal to the target
diameter:
Sf ;CY;Preform ¼ NCY Sf ;CY
DBraid ¼ DTarget Sum the fiber consolidated area of the preform’s BYs, MEYs,
Or Else, at all the other iterations and CYs:
Attribute the next diameter value for the braid diameter:
Sf ;Preform ¼ Sf ;BY;Preform þ Sf ;MEY;Preform þ Sf ;CY;Preform
DBraid ¼ DNext Calculate the total consolidated material (fiber and matrix)
Calculate the BY width at surface, the BY width at pitch, the area of only the BYs, only the MEYs, and only the CYs:
MEY width, and also the maximum crimp angle with the
algorithm presented at Appendix B and using the following NBY SBY RC
variables: SBY;Preform ¼ ;
cos hBraid

 Unconsolidated BY area (ABY)

 Minimum BY aspect ratio (RaBY,min) SMEY;Preform ¼ NMEY SMEY
 Maximum BY aspect ratio (RaBY,Min)
 Unconsolidated MEY area (AMEY)
 Minimum BY aspect ratio (RaMEY,min) SCY;Preform ¼ NCY SCY
 Maximum BY aspect ratio (RaMEY,Max)
 Braid pitch (p) Sum the material area of the preform’s BYs, MEYs, and CYs:
 Braid diameter (DBraid)
 Amount of BYs (NBY) SPreform ¼ SBY;Preform þ SMEY;Preform þ SCY;Preform

Calculate the BY thickness at surface, the BY thickness at pitch Calculate the preform fineness:
and the MEY thickness:
NBY F BY RC
F Preform ¼ þ N MEY F MEY þ NCY F CY
4ABY 4ABY 4AMEY cos hBraid
bBYS ¼ ; b ¼ ; b ¼
paBYS BYP paBYP MEY paMEY Calculate the preform fiber volume fraction:
Calculate the braid outside diameter:
Dout = DBraid + bMEY + 2bBYS Sf ;Preform
Calculate the average braid angle: V f ;Preform ¼
SPreform
  Calculate the beam fiber volume fraction:
pDBraid
hBraid ¼ arctan
p (continued on next page)
1728 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729

Sf ;Preform  Number of braiding yarns (NBY)

V f ;Beam ¼ Calculate the error on the BY fiber length:
ADie
Calculate the fill ratio:
eL ¼ LBY;preform  LBY;Consolidated
SPreform Adjust the braided layer consolidated circumference using
RFill ¼ the error calculated:
ADie
Calculate the material area of only the braided layer in the eL
LBraid;Consolidated ¼ LBraid;Consolidated 
beam: 2 sin hBraid;Consolidated
NBY  Sf ;BY  RC NMEY  Sf ;MEY Re-calculate the fiber angle in the consolidated state:
SBraid;Beam ¼ þ
cos hBraid  V f ;Beam V f ;Beam
 
Calculate the thickness dimension of the braided layer in CBraid;Consolidated
hBraid;Beam ¼ arctan
the beam if the braided layer occupies all the die p
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Calculate the error on the consolidation circumference:
C Die  C 2Die  SBraid;Beam
t Braid;Beam ¼
2
eC ¼ CBraid;Beam  CBraid;Consolidated
Calculate the L-shape side dimension of the beam core layer
Adjust the outside diameter of the braid to achieve a fit of
the preform in the consolidation die:
C CY;Beam ¼ C BL;Beam  2t BL;Beam
Initialize the L-shape thickness dimension of the beam core eC
layer DTarget ¼ Dout 
2p
Calculate the beam modulus using classical laminate theory
T CY;Beam ¼ T Die  2tBraid;Beam
Calculate a beam braided layer circumference if the braided
layer fits the L-shape: Appendix D. Supplementary material

CBraid;Beam ¼ 4C Die  4t Braid;Beam Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.compositesa.
Calculate the thickness of the braiding yarn at surface in the 2012.03.001.
beam (consolidated state):
References
t Braid;Beam
bBYS;Beam ¼ bBYS [1] Okano M, Nakai A, Hamada H. Axial crushing performance of braided
t Braid;Preform
composite tubes. Int J Crashworthiness 2005;10(3):287–94.
[2] Joshi SC, Lam YC. Integrated approach for modelling cure and crystallization
Calculate the thickness of the middle-end-yarn in the beam kinetics of different polymers in 3D pultrusion simulation. J Mater Process
(consolidated state): Technol 2006;174(1–3):178–82.
[3] Hussain F, Roy S, Narasimhan K, Vengadassalam K, Lu H. E-glass-polypropylene
pultruded nanocomposite: manufacture, characterization, thermal and
tBraid;Beam mechanical properties. J Thermoplast Compos Mater 2007;20(4):411–34.
bMEY;Beam ¼ bMEY [4] Bechtold G, Kameo K, Friedrich K, Hamada H. Pullbraiding of commingled GF/
t Braid;Preform
PP yarn: influence of processing parameters. Proceedings 4th international
Initialize the consolidated circumference equal to the beam symposium for textile composites (TexComp). Kyoto, Japan; 1998. p. O/28/1-5.
[5] Michaeli W, Juerss D. Thermoplastic pull-braiding: pultrusion of profiles with
braided layer circumference: braided fibre lay-up and thermoplastic matrix system (PP). Compos A Appl Sci
Manuf 1996;27(1):3–7.
[6] Lebel LL, Tanaka Y, Torun AR, Ohtani A, Nakai A. Consolidation quality and
CBraid;Consolidated ¼ CBraid;Beam mechanical properties of thermoplastic composite manufactured by pull-
braiding 8th joint Canada–Japan workshop on composites. Montreal,
Initialize the consolidated fiber angle using the beam Canada: DEStech Publications Inc; 2010.
braided layer circumference and the pitch: [7] Brunnschweiler D. Braids and braiding. J Text Inst 1953;44(9):666–86.
[8] Brunnschweiler D. Structure and tensile properties of braids. J Text Inst Trans
1954;45(1):55–77.
  [9] Peirce FT. The geometry of cloth structure. J Text Inst Trans 1937;28(3):45–96.
CBL;Beam
hBraid;Consolidated ¼ arctan [10] Ko FK, Pastore CM, Head AA. Handbook of industrial braiding. Convington,
p KY: Atkins and Pearce; 1989.
[11] Pastore CM, Birger AB, Clyburn E. Geometrical modelling of textile
While the error on the BY fiber length in the beam is still reinforcements. In: Poe CC, Harris CE, editors. Mechanics of textile
over 1010 composites conference. Hampton, Virginia; 1995. p. 597–623.
[12] Robitaille F, Clayton BR, Long AC, Souter BJ, Rudd CD. Geometric modelling of
Calculate the fiber length and the crimp ratio using the industrial preforms: woven and braided textiles. Proceedings of the
algorithm provided at Appendix A and using the following Institution of Mechanical Engineers, Part L, vol, 213 [J Mater: Des Appl];
variables: 2009. p. 69–82.
[13] Lomov SV, Huysmans G, Verpoest I. Hierarchy of textile structures and
 Consolidated BY thickness at surface (bBYS,Beam)
architecture of fabric geometric models. Text Res J 2001;71(6):534–43.
 Consolidated MEY thickness (bMEY,Beam) [14] Lomov SV, Parnas RS, Ghosh SB, Verpoest I, Nakai A. Experimental and
 Consolidated fiber angle (hBraid,Consolidated) theoretical characterization of the geometry of two-dimensional braided
 Pitch (p) fabrics. Text Res J 2002;72(8):706–12.
 L.L. Lebel, A. Nakai / Composites: Part A 43 (2012) 1717–1729 1729

[15] Laberge-Lebel L, Hoa SV. Manufacturing of braided thermoplastic composites
with carbon/nylon commingled fibers. J Compos Mater 2007;41(9):1101–21.
[16] Lebel LL, Daisuke H, Tanaka Y, Nakai A. Braid-trusion of jute/polylactic acid
composites. 25th American society for composite technical conference and
14th US–Japan conference on composite materials. Dayton, Ohio, USA; 2010.
[17] Hyer MW. Stress analysis of fiber-reinforced composite materials. Boston:
WCB McGraw-Hill; 1998.
[18] Carlsson A, Astrom BT. Experimental investigation of pultrusion of glass fibre
reinforced polypropylene composites. Compos A Appl Sci Manuf 1998;29(5–
6):585–93.
[19] Bernet N, Michaud V, Bourban PE, Manson JAE. An impregnation model for the
consolidation of thermoplastic composites made from commingled yarns. J
Compos Mater 1999;33(8):751–72.