Professional Documents
Culture Documents
Di
Milano
Facoltà
di
Ingegneria
dei
Processi
Industriali
Degree
in
Materials
Engineering
3D
Reinforcement
of
Composite
Materials
Supervisors:
Valter
Carvelli
(Politecnico
di
Milano)
Giulio
Ventura
(Politecnico
di
Torino)
Carlo
Poggi
(Politecnico
di
Milano)
Corinna
A
Conway
750155
Academic
Year
2010
-‐
2011
2
Dedicated
to
my
little
brother,
Michael
Conway
Who
taught
me
how
precious
life
truly
is
and
what
it
means
to
follow
your
dreams.
And
to
My
Parents,
Robert
and
Theresa
Conway
Who
made
this
all
possible
through
their
love
and
3
4
Summary 14
1.1
Introduction
16
1.2
Fiber
Reinforced
Composites
(FRC)
16
1.2.1
The
Matrix
17
1.2.2
The
Fibers
17
1.3
Reinforcement
Architectures
19
1.3.1
2D
Composites
20
1.3.2
3D
Composites
21
References
22
Chapter
2
Reinforcement
Fabric
Manufacturing
23
2.1
Introduction
23
2.2
Weaving
23
2.2.1
2D
Weaving
23
2.2.2
3D
Weaving
25
2.2.3
3D
Orthogonal
Non-‐Woven,
Multiaxial
Weaving
and
Distance
Fabrics
26
2.3
Braiding
27
2.3.1
2D
Braiding
27
2.3.2
Two
and
Four
Step
3D
Braiding
27
2.3.3
Multilayer
Interlock
Braiding
28
2.4
Knitting
28
2.5
Stitching
29
2.6
Z-‐Pinning
30
5
References 31
Chapter
3
Composite
Manufacturing
32
3.1
Introduction
32
3.2
Composite
Consolidation
techniques
32
3.2.1
Resin
Transfer
Molding
32
3.2.2
Resin
Film
Infusion
32
3.2.3
SCRIMP
34
3.3
Consolidation
Equipment
34
3.3.1
Tooling
(mold)
35
3.3.2
Heating
and
Cooling
35
3.3.3
Injection
Equipment
36
3.4
Optimization
36
References
39
Chapter
4
Textile
Fiber
Reinforcement
Properties
40
4.1
Introduction
40
4.2
In-‐Plane
Shear
40
4.3
In-‐Plane
Biaxial
Tension
42
References
44
Chapter
5
Composite
Modeling
45
5.1
Introduction
45
5.2
Fundamentals
46
5.3
Representative
Volume
48
5.4
Rule
of
Mixtures
50
5.5
Basic
Models
for
2D
Woven
Composites
51
5.5.1
Mosaic
Model
51
5.5.2
Undulation
Model
53
6
Chapter
6
3D
Woven
Composites
63
6.1
Introduction
63
6.2
3D
Woven
Composites
63
6.2.1
Microstructure
Features
and
Crimp
63
6.2.2
Tensile
Properties
66
6.2.3
Compressive
Properties
67
6.2.4
Flexural
and
Interlaminar
Shear
Properties
68
6.2.5
Interlaminar
Fracture
68
6.2.6
Impact
Damage
Tolerance
69
References
71
Chapter
7
3D
Braided,
Knitted,
Stitched
and
Z-‐Pinned
Composites
72
7.1
Introduction
72
7.2
3D
Braided
Composites
72
7.2.1
In-‐Plane
Properties
72
7.2.2
3D
vs.
2D
Braided
Composites
73
7.3
3D
Knit
Composites
73
7.3.1
In-‐Plane
Properties
73
7.3.2
Interlaminar
Fracture
and
Impact
Toughness
75
7.4
Stitched
Composites
75
7.4.1
In-‐Plane
Mechanical
Properties
76
7.4.2
Fracture
Toughness
and
Impact
Damage
Tolerance
76
7.5
Z-‐Pinned
Composites
77
7
8
Figures
Figure
1.1:
Depiction
of
fiber
type,
and
non-‐woven
composite
architectures
18
Figure
1.2:
Layup
sequence
18
Figure
1.3:
Stress
vs.
Strain
graph
comparing
Carbon
(green),
Glass
(purple)
and
Aramid
(red)
fiber
properties
19
Figure
1.4:
Braided,
woven
and
knit
fabric
structures
20
Figure
1.5:
Comparison
of
in-‐plane
and
through-‐thickness
Properties
21
Figure
2.1:
Traditional
weaving
loom
24
Figure
2.2:
Illustration
of
yarn
structure,
and
common
weave
Patterns
25
Figure
2.3:
3D
weave
geometries
26
Figure
2.4:
Illustrating
the
ability
to
weave
slits
into
the
fabric
capable
of
creating
three-‐dimensional
structures
26
Figure
2.5:
Multilayer
interlock
braided
fabric
28
Figure
2.6:
3D
knit
fabric
29
Figure
2.7:
Stitched
fabric
30
Figure
2.8:
Z-‐pinning
process
30
Figure
3.1:
RTM
33
Figure
3.2:
RFI
34
Figure
3.3:
Autoclave
for
composite
consolidation.
Image
provided
by
AAC
research
35
Figure
3.4:
Viscosity
vs.
Time
–
temperature
dependence
of
thermoset
TGDDM
resin.
Image
taken
from
Understanding
of
Rheology
of
Thermosets
37
Figure
4.1:
Biaxial
tension
test
(left)
and
Picture
frame
test
(right)
41
Figure
4.3:
Illustrating
non-‐linear
behavior
of
woven
fabric
42
Figure
4.4:
Biaxial
testing
Machine
42
Figure
4.5:
Biaxial
testing
sample
43
Figure
4.6:
Clamps
43
Figure
5.1:
Illustrating
the
microscopic
heterogeneity
of
a
composite
structure.
Fibers
shown
in
grey
and
matrix
in
blue.
45
Figure
5.2:
Illustrating
the
three
planes
of
symmetry
that
make
composites
orthotropic
materials.
Planes
are
shown
in
yellow.
46
9
10
11
12
Abstract
Composite
materials
present
a
unique
opportunity
to
engineer
a
material
in
order
to optimize
its
physical,
thermal
and
mechanical
properties
for
specific
applications.
Offering
many
advantages
such
a
relatively
high
specific
strength,
stiffness,
fatigue
resistance
and
corrosion
resistance
with
respect
to
weight.
Due
to
their
exceptional
qualities,
composites
can
be
found
in
many
applications,
from
aircrafts,
helicopters
and
spacecrafts
to
submarines,
automobiles
and
sporting
goods.
However
their
wide
spread
use
has
been
inhibited
by
their
high
cost,
poor
delamination
toughness,
and
poor
impact
damage
resistance.
Many
prospects
have
been
investigated
as
methods
for
improving
these
characteristics,
however
composites
reinforced
with
3D
fabric
architectures
appear
to
be
the
most
promising
solution.
Here
an
investigation
of
3D
fabric
architectures
(3D
woven,
braided,
knit,
stitched
and
z-‐pinned),
manufacturing
methods,
and
composite
properties
are
reviewed
in
order
to
have
a
better
understanding
of
the
pros
and
cons
of
such
a
material
as
well
as
potential
improvements
and
opportunities.
As
expected
3D
composites
solve
many
of
the
problems
faced
by
2D
composites,
however
these
improvements
are
accompanied
by
the
deterioration
of
in-‐plane
properties.
Many
3D
composites
show
potential
for
applications
unsuited
for
2D
composites,
however
optimization
of
3D
fabric
manufacturing,
composite
production,
and
in-‐
and
out-‐
of-‐plane
properties
needs
further
investigation.
13
Summary
I
materiali
compositi
presentano
una
opportunità
unica
per
progettare
un
materiale
in
modo
da
ottimizzare
le
sue
proprietà
fisiche,
termiche
e
meccaniche
per
applicazioni
specifiche.
Questi
materiali
offrono
molti
vantaggi
come:
una
relativamente
alta
resistenza
specifica,
rigidezza,
resistenza
a
fatica
e
resistenza
a
corrosione
se
confrontati
col
peso.
A
causa
delle
loro
qualità
eccezionali,
i
compositi
possono
essere
trovati
in
molte
applicazioni:
dagli
aeroplani,
elicotteri
e
veicoli
spaziali
ai
sottomarini,
automobili
e
merci
sportive.
Purtroppo
la
loro
larga
diffusione
è
limitata
dal
loro
alto
costo,
piccola
durezza
alla
laminazione
e
piccola
resistenza
all'impatto.
Molti
aspetti
sono
stati
investigati
come
metodi
per
migliorare
queste
caratteristiche,
comunque
i
compositi
rinforzati
con
architetture
di
tessuti
3D
appaiono
essere
la
soluzione
più
promettente.
Un
approfondimento
sui
metodi
di
produzione,
modellazione,
e
le
proprietà
dei
compositi
di
tessuti
3d,
3d
intreccaiti,
cuciti
e
z-‐appuntati
architetture
di
rinforzo
è
stata
eseguita
in
modo
da
capire
meglio
i
pro
e
contro
di
questi
materiali
così
come
possibili
miglioramenti
e
opportunità.
Molti
miglioramenti
sono
ancora
necessari
per
la
produzione
di
tessuti
di
rinforzo
3D.
I
tessuti
e
i
tessuti
intrecciati
possono
essere
prodotti
usando
speciali
macchine
o
modificando
i
tradizionali
macchinari
2D.
Comunque
le
architetture
e
i
costi
e
i
volumi
di
produzione
sono
correntemente
limitati
dalle
tecnologie
disponibili.
RTM
(Resin
transfer
molding)
and
RFI
(Resin
film
infusion)
or
SCRIMP
sono
i
metodi
più
efficienti
per
la
corrente
produzione
di
compositi
3D.
Ogni
metodo
ha
diversi
benefici
e
limitazioni.
Una
revisione
di
base
degli
attuali
metodi
di
prova
e
modellazione
per
i
compositi
3D
è
presentata
nei
capitoli
4
e
5.
Come
ci
si
aspetta
i
compositi
3D
risolvono
molti
dei
problemi
che
hanno
i
compositi
2D,
comunque
questi
miglioramenti
sono
accompagnati
dalla
deteriorazione
delle
proprietà
nel
piano.
L'ottimizzazione
della
manifattura
dei
tessuti
3D,
la
produzione
di
compositi,
e
le
proprietà
nel
piano
e
fuori
dal
piano
hanno
bisogno
di
ulteriori
investigazioni,
comunque
molti
compositi
3D
mostrano
potenzialità
per
diverse
applicazioni
per
cui
non
possono
essere
usati
compositi
2D.
Per
esempio,
cuciture
e
z-‐pinning
mostrano
eccezionali
potenzialità
per
il
rinforzo
dei
giunti,
mentre
I
tessuti
a
maglia
3D
mostrano
un
eccellemte
resistenza
all'impatto
e
sono
di
particolare
interesse
per
l'uso
nelle
protesi
ma
non
sono
utilizzabili
per
applicazioni
strutturali.
14
15
1.2.1
The
Matrix
Selection
of
the
constituent
materials
depends
on
the
desired
properties
of
the
final
product.
The
two
most
common
matrix
materials
for
FR
composites
are
thermoplastics
and
thermosets.
Thermoplastics
are
characterized
by:
high
application
temperatures,
high
toughness
and
ease
of
repair.
On
the
down
side
thermoplastics
require
high
processing
temperatures,
and
can
be
difficult
to
handle
due
to
their
high
viscosity.
Thermoset
matrix
materials
are
characterized
by
their
low
viscosity
and
low
processing
temperature
with
drawbacks
in
application
temperature,
and
toughness
(see
Table
1.1).
Matrix
Material
Cost
Application
Temperature
Toughness
PAI
(TP)
>25
€/kg
>
300°C
Medium-‐High
PEEK
(TP)
>25
€/kg
>
300°C
High
Polyimide
(TS)
10-‐25
€/kg
200-‐300°C
High
PES
(TP)
10-‐25
€/kg
200-‐300°C
Low-‐Medium
Epoxy
(TS)
2.5
–
10
€/kg
120-‐200°C
Low
Phenolic
(TS)
2.5
–
10
€/kg
120-‐200°C
Low
PBT
(TP)
2.5
–
10
€/kg
120-‐200°C
Low-‐Medium
PA
(TP)
2.5
–
10
€/kg
120-‐200°C
Medium-‐High
Polyester
(TS)
<2.5
€/kg
<120°C
Low
PP
(TP)
<2.5
€/kg
<120°C
Table
1.1:
Matrix
materials
costs,
application
temperature
and
toughness.
(TP
–
thermoplastic
and
TS
–
thermoset)
3
1.2.2
The
Fibers
The
fibers
of
the
FR
composite
can
be
varied
in
size,
shape,
length,
direction,
architecture,
and
material
in
order
to
engineer
a
composite
to
the
have
specific
properties.
The
length
of
the
reinforcing
fibers
can
be
whiskers
(short/staple)
or
continuous
(filament)
(Figure
1.1),
and
usually
have
an
ovular
or
circular
cross-‐sectional
shape,
although
almost
any
shape
is
possible.
Whisker
reinforcement
fibers
are
used
to
create
non-‐woven,
non-‐structural
composites.
When
randomly
oriented
in
the
matrix
material
they
create
an
isotropic
composite,
while
orienting
the
fibers
can
give
more
strength
in
the
orientation
direction,
generating
an
anisotropic
composite.
16
Figure
1.1:
Depiction
of
fiber
type,
and
non-‐woven
composite
architectures
On
the
other
hand,
using
filament
fibers
makes
it
possible
to
engineer
the
reinforcement
architecture.
This
can
be
achieved
through
the
prepreg
lay-‐up,
or
by
using
woven,
braided,
stitched,
or
z-‐pinned
fabrics.
A
prepreg
is
a
unidirectional
fiber
sheet
impregnated
with
uncured
matrix
resin.
The
layup
of
the
prepregs
determines
the
fiber
orientations
within
the
composite
(see
figure
1.2).
Fibers
may
be
oriented
in
one
direction
(e.g.
0°/0°/0°/0°)
giving
unidirectional
characteristics,
or
in
multiple
directions
(e.g.
0°/+45°/-‐45°/90)
creating
quasi-‐isotropic
properties.1
Figure
1.2:
Layup
sequence3
17
The
fiber
material
is
also
very
important.
Glass
fiber
reinforcements
are
the
most
common
due
to
their
low
cost
and
high
strength,
however
limitations
are
found
in
the
low
modulus,
high
density/weight,
low
fatigue
and
wear
resistance,
and
sensitivity
to
humid
environments.
Within
the
glass
fibers
there
are
different
types
that
can
be
selected
based
on
the
desired
properties
and
environmental
conditions
(see
Table
1.2)
Type
General
Characteristics
E
Low
cost,
General
purpose
S/R
High
stiffness
and
strength
D
Good
dielectric
properties
A/AR
Alkali
resistance
E-‐CR
Acid
Resistance
C
Good
chemical
resistance
Table
1.2:
Glass
type
and
defining
characteristics
Carbon
fibers
are
becoming
more
popular
and
are
of
high
interest
due
to
their
high
modulus,
high
strength,
and
low
density/weight,
however
they
are
still
extremely
expensive.
Figure
1.3:
Stress
vs.
Strain
graph
comparing
Carbon
(green),
Glass
(purple)
and
Aramid
(red)
fiber
properties.
3
Aramid
fibers
have
advantages
in
its
high
toughness,
high
strength
and
low
cost,
but
suffer
from
low
UV
resistance,
and
low
compressive
strength
(although
the
low
compressive
strength
can
be
used
to
an
advantage
in
certain
applications).
These
are
the
three
most
common
fiber
reinforcement
materials,
whose
tensile
behavior
are
compared
in
Figure
1.3.
However
it
is
important
to
note,
that
boron,
basalt
and
ceramic
fibers
have
also
been
used
to
a
much
lesser
extent.1,2
1.3
Reinforcement
Architectures
Using
more
complex
reinforcement
architectures
gives
another
engineering
possibility.
Woven,
knit,
braided,
stitched,
and
z-‐pinned
architectures
provide
18
the
most
interesting
opportunities
at
the
moment
(figure
1.4).
Within
each
fabric
production
process
there
are
many
different
architectures
that
can
be
achieved.
For
example
with
3D
weaving
we
can
produce
angle
interlock
weaves,
orthogonal
weaves
or
through-‐thickness
interlock
weaves.
The
properties
of
each
of
these
fabrics
differ
greatly,
therefore
the
fabric
itself
can
be
engineered
for
the
desired
properties.1,2
Figure
1.4:
Braided,
woven
and
knit
fabric
structures
1.3.1
2D
Composites
2D laminated composites are among the most common composites used in the
market today. In applications requiring high performance properties filament fibers
are selected over whiskers and are oriented in the x-, y-directions of the composite.
Some of the major disadvantages of 2D composites lie in their high cost, and low
through-thickness mechanical properties due to the lack of z-directional fibers.
Therefore the mechanical properties in the through-thickness direction are determined
by the mechanical properties of the resin and the fiber-resin interface. A comparison
of the in-plane and through-thickness strengths of 2D laminates, seen in figure 1.5
below, reveals that the through-thickness properties are usually less than 10% of the
in-plane properties and therefore cannot be used in applications supporting high
through-thickness or inter-laminar shear loads.2
19
20
References
1. M. Sc. Badawi, Said Sobhey A. M. Development
of
the
Weaving
Machine
and
3D
Woven
Spacer
Fabric
Structures
for
Light
Weight
Composites
Materials.
Dresden Technical University. 2007
2. Tong,
L.
Mouritz,
A.P.
and
Bannister,
M.K.
3D
Fibre
Reinforced
Polymer
Composites.
Elsevier
Science
Ltd.
Oxford,
UK.
2002.
3. Poggi,
Carlo.
Composites
For
Structural
Application.
Politecnico
di
Milano.
Course,
2nd
Semester
2011
21
22
Figure
2.1:
Traditional
weaving
loom3
The
weaving
process
starts
by
threading
or
feeding
the
warp
yarns,
those
that
run
in
the
machine
direction
–
0
degrees,
into
the
loom
from
the
source
yarns.
The
source
yarns
are
run
through
a
series
of
rollers
in
order
to
maintain
and
control
the
tension.
These
yarns
are
then
fed
through
the
lifting
mechanism.
The
lifting
mechanism
lifts
the
warp
yarns
in
order
to
create
a
space,
or
shed,
for
the
weft
yarns
to
be
inserted.
The
weft
yarns
are
those
running
horizontal
to
the
machine,
or
in
the
90
degree
direction.
The
sequence
in
which
the
warp
yarns
are
lifted
and
the
weft
yarns
inserted
creates
the
pattern
of
the
fabric
(see
Figure
2.2).
It
is
important
to
note
that
the
fabric
architecture
greatly
influences
the
mechanical
properties
and
drapability
of
the
fabric
and
is
highly
dependent
on
the
weave
pattern,
fiber
or
tow
size,
weft
and
warp
yarn
concentration,
yarn
tension,
and
tightness
of
the
tows.1,2,3
Plain
weave
being
the
stiffest
(least
drapable)
and
weakest,
while
satin
is
the
strongest
and
most
drapable.3
23
Figure
2.2:
Illustration
of
yarn
structure,
and
common
weave
patterns
Picking
is
the
process
of
inserting
the
weft
yarns
in
the
shed
created
by
the
lifting
mechanism.2
This
can
take
place
in
a
number
of
different
ways.
The
oldest
technique
for
insertion
of
the
weft
yarns
is
through
the
use
of
a
shuttle
to
transport
the
yarns
through
the
shed.
This
technique
is
slow,
but
creates
a
closed
edge
fabric.
Open
edged
fabrics
can
be
produced
at
much
quicker
rates,
using
a
mechanical
arm,
rapier,
or
high-‐pressure
air
or
water
to
transport
the
weft
yarns
through
the
shed.3
The
next
step
is
the
beating
up
process,
in
which
the
inserted
weft
yarns
are
compacted
using
a
comb-‐like
devise,
the
reed.
Finally,
in
order
to
have
a
continuous
process
the
fabric
is
advanced
forward
by
a
series
of
positively
driven
rollers,
this
is
called
take-‐up.
This
process
is
continued
until
the
desired
length
of
fabric
is
created.
The
fabric
can
be
produced
continuously
and
cut
into
the
lengths
needed.
Also
different
types
of
yarns
can
be
used
for
warp
and
weft
to
help
created
a
fabric
better
suited
for
the
intended
use.
1,2,3
2.2.2
3D
Weaving
The
major
difference
between
2D
and
3D
woven
fabrics
is
the
need
of
multiple
layers
of
warp
yarns
in
the
3D
fabrics.
This
tends
to
be
a
major
disadvantage,
as
the
need
for
a
large
number
of
warp-‐ends
and
the
time
required
to
prepare
the
loom
can
be
very
costly.
Therefore,
at
the
moment,
most
3D
woven
fabrics
are
used
in
the
production
of
narrow
products
reducing
the
number
of
warp
yarns
required.
As
stated
above,
the
traditional
weaving
equipment
can
be
easily
altered
to
create
a
3D
woven
fabric.
The
first
modification
is
to
use
a
lifting
mechanism
with
multiple
eyes,
allowing
for
layered
warp
yarns.
Jacquard
looms
are
normally
selected
for
the
production
of
3D
woven
fabrics,
given
the
distinct
advantage
of
improved
control
of
the
lifting
mechanism.
With
the
multiple
layers
of
warp
yarns,
comes
the
creation
of
multiple
sheds.
This
allows
for
multiple
insertions
of
the
weft
yarns
at
the
same
time,
and
is
the
second
modification
needed
in
order
to
have
3D
weaving.
1,2,3
In
the
formation
of
a
3D
woven
fabric,
pockets
are
formed
between
any
four
adjacent
warp
yarns.
These
pockets
can
be
filled
with
stuffer
yarns
that
do
not
interlace
with
the
weft
yarns.
The
pockets
can
be
filled
according
to
mechanical
needs
and
in
this
way
the
fabric
can
be
further
engineered
to
specific
applications.
In
order
to
maximize
performance,
majority
of
the
yarns
are
designed
to
lay
flat,
and
only
select
warp
yarns
are
used
to
bind
the
layers
together.
Examples
of
the
weaving
architectures
capable
of
being
produced
using
the
3D
weaving
procedure
are
given
in
the
figure
2.3.
Please
note
that
24
these
fabric
architectures
are
idealized
and
not
possible
in
reality.
Woven
materials
can
be
produced
in
types
of
solid,
shell,
tubular
and/or
combinations
of
these.
1,2,3
Figure
2.3:
3D
weave
geometries2
As
with
2D
weaving,
3D
weaving
is
limited
to
yarn
placement
in
the
0
and
90
degree
directions.
Therefore
its
use
is
limited
to
components
that
are
not
subjected
to
extensive
shear
and
torsion
stresses.
An
advantage
of
the
weaving
loom
is
its
capability
of
producing
fabrics
with
slits
that
can
then
be
opened
into
three-‐dimensional
structures
(see
figure
2.4).
This
can
be
used
to
produce
I
beams
and
boxes
using
flat
fabric
and
have
already
been
used
in
civil
engineering
components.
3
Figure
2.4:
Illustrating
the
ability
to
weave
slits
into
the
fabric
capable
of
creating
three-‐
dimensional
structures.3
Examples
of
3D
weaving
equipment
include,
3WEAVE
created
by
3tex.
This
machine
allows
for
the
use
of
multiple
filling
layers
at
a
time,
use
of
carbon,
aramid,
glass,
polyethylene,
steel
fibers,
etc.,
produce
a
fabric
thickness
up
to
25.4
mm,
and
a
fabric
width
of
1830mm.
2
2.2.3
3D
orthogonal
Non-Woven,
Multiaxial
Weaving,
and
Distance
Fabrics
3D
orthogonal
non-‐wovens
are
those
fabrics
produced
from
the
same
equipment
used
to
produce
3D
woven
fabrics,
but
that
do
not
contain
interlacing
25
26
Figure
2.5:
Multilayer
interlock
braided
fabric3
2.4
Knitting
At
the
moment
knitting
is
the
least
known
and
studied
of
the
fabric
production
techniques
for
use
as
composite
reinforcements.
However,
current
conventional
knitting
machines
are
already
capable
of
producing
3D
27
Figure
2.6:
3D
knit
fabric3
The
traditional
forms
of
knitting
are
either
warp
of
weft
knitting.
In
weft
knitting
there
is
only
a
single
yarn
fed
into
the
machine
at
a
90-‐degree
direction
with
respect
to
the
fabric
production.
The
yarn
forms
a
line
of
interlocking
loops
to
form
the
knit
fabric.
While
with
warp
knitting,
there
are
a
number
of
yarns
feed
into
the
machine
at
the
0
degree
direction
with
respect
to
the
fabric
production.
With
warp
knitting,
multiple
types
of
yarns
can
easily
be
knit
together,
however
more
yarn
bundles
will
be
needed
and
therefore
can
be
more
costly.
The
interlocking
of
the
loops
is
achieved
through
a
needle
bed,
a
row
of
closely
spaced
needles
that
pull
the
yarns
through
the
previously
knit
loops.
Machines
with
two
or
more
needle
beds
are
capable
of
creating
3D
knit
fabrics.
3
2.5
Stitching
Stitching
is
the
simplest
and
cheapest
of
the
methods
for
producing
a
3D
fabric
architecture.
The
process
involves
the
insertion
of
a
needle
carrying
a
z-‐
directional
yarn
through
layers
of
2D
fabric,
in
effect
stitching
the
layers
together
and
creating
a
3D
architecture
(see
figure
2.7).
The
z-‐binding
yarns
are
most
commonly
aramid.
This
is
due
to
their
high
toughness
as
well
as
equipment
requirements.
Current
stitching
machinery
may
be
used
with
aramid
yarns
28
without
further
alterations.
However
attention
must
be
given
to
their
tendency
to
absorb
moisture
and
insufficient
binding
to
many
common
polymer
resins.3
Figure
2.7:
Stitched
fabric3
Creating
3D
architectures
through
stitching
provides
many
benefits.
Among
those
is
the
possibility
to
use
the
process
with
traditional
2D
woven,
braided,
knit,
etc.
prepregs.
This
allows
for
a
great
degree
of
flexibility
in
the
fabric
lay-‐up;
using
different
material
layers,
as
well
as
different
yarn
directions.
Also,
stitching
can
be
placed
only
in
the
areas
that
require
reinforcement
in
the
z-‐
direction,
as
well
as
complex
stitching
patterns
by
using
current
embroidery
machinery
and
software.
Another
great
advantage
is
the
ability
to
create
complex
3D
shapes
by
stitching
different
component
parts
together.3
The
main
disadvantages
with
this
process
lay
in
the
reduction
of
the
in-‐
plane
properties.
This
is
due
to
local
fiber
damaged
caused
by
the
needle
insertion,
increased
crimp
induced
by
the
z-‐directional
yarns,
and
resin-‐rich
pockets
formed
by
the
bunching
of
fibers
contained
by
the
stitching
yarns.3
2.6
Z-pinning
Z-‐pinning
is
used
as
an
alternative
method
to
stitching.
The
process
uses
pre-‐cured
reinforcement
fibers,
which
are
embedded
in
a
thermoplastic
foam
and
placed
on
top
of
the
prepreg
or
dry
fabric.
The
prepreg
and
foam
are
then
prepared
for
curing.
During
the
curing
process,
the
thermoplastic
foam
collapses
and
the
pressure
slowly
drives
the
reinforcing
fibers
into
the
component
(see
figure
2.8).
With
z-‐pinning,
there
is
less
crimping
induced
by
the
z-‐directional
reinforcing
fibers
as
well
as
less
damage
to
the
yarns
in
the
prepreg,
while
still
maintaining
the
high
level
of
control
over
reinforcement
placement.3
Figure
2.8:
Z-‐pinning
process3
29
References
1. Long,
A.C.
Design
and
Manufacture
of
Textile
Composites.
Woodhead
Publishing
Limited,
Cambridge
England.
2005.
2. M. Sc. Badawi, Said Sobhey A. M. Development
of
the
Weaving
Machine
and
3D
Woven
Spacer
Fabric
Structures
for
Light
Weight
Composites
Materials.
Dresden Technical University. 2007
3. Tong,
L.
Mouritz,
A.P.
and
Bannister,
M.K.
3D
Fibre
Reinforced
Polymer
Composites.
Elsevier
Science
Ltd.
Oxford,
UK.
2002.
30
Composite
Manufacturing
3
3.1
Introduction
There
are
many
different
ways
in
which
to
consolidate
the
preform
to
create
the
final
composite
component,
however
not
all
of
these
processes
are
suited
for
3D
preform
consolidation.
Methods
such
as
hand
impregnation,
pultrusion,
and
commingled
yarns
greatly
distort
the
fabric
architecture
during
composite
consolidation,
significantly
diminishing
the
final
mechanical
properties
of
the
component.
Therefore,
in
order
to
reap
the
benefits
of
the
3D
preform
production
technologies,
the
correct
consolidation
technology
must
be
chosen.
At
this
moment
the
only
manufacturing
process
that
is
successful
in
consolidating
3D
fiber
performs
is
Liquid
Molding
(LCM).
This
is
due
to
its
high
flexibility
regarding
component
shape.
For
preforms
of
complex
geometries
LCM
offers
opportunity
to
produce
a
high
quality
component
for
a
relatively
low
cost.
1,2
31
and
resin
polymerization
rate.
Using
these
variants,
injection
time
and
length
can
be
determined
and
maximized
with
economical
considerations.
However,
the
tooling
used
in
RTM
is
often
expensive
due
to
the
high-‐pressure
requirements,
and
component
size
is
limited
due
to
maximum
injection
length
(two
meters
is
generally
the
limit)
and
financial
considerations.
1,2
Figure
3.1:
RTM
Variations
to
RTM
include
Vacuum
assisted
RTM
(VARTM)
where
a
vacuum
is
applied
to
aid
in
consolidation,
air
removal
and
increase
the
velocity
of
resin
infiltration,
and
Structural
Reaction
Injection
Molding
(SRIM)
where
higher
injection
pressures
are
used
to
decrease
production
time.
1,2
3.2.2
Resin
Film
Infusion
Resin
Film
Infusion
(RFI)
is
an
alternative
to
the
RTM
method.
In
RFI
the
resin
is
present
in
the
form
of
a
film
instead
of
a
liquid.
The
resin
film
is
placed
on
the
tool
surface,
over
which
the
preform
is
placed.
On
top
of
the
preform
a
release
film
(to
allow
for
easy
component
removal)
and
breather
material
(in
order
to
form
a
vacuum)
are
added.
Everything
is
bagged,
vacuumed
and
placed
in
an
autoclave
to
be
heated
under
pressure
(see
figure
3.2).
The
resin
film
melts
and
is
sucked
up
into
the
preform
through
capillary
action,
thus
being
absorbed
in
the
thickness
direction.
The
pressure
can
be
varied
in
order
to
compact
the
component
to
the
desired
fiber
volume
fraction.
1,2
32
Figure
3.2:
RFI2
RFI
has
many
advantages
and
disadvantages
of
RTM.
The
advantages
of
RFI
consist
of
the
relatively
low
tooling
costs
and
the
loss
of
the
maximum
injection
length
limitations.
However,
RFI
has
limitations
in
the
thickness
of
the
component.
Therefore
RFI
is
usually
used
with
thinner
larger
components
while
RTM
is
suited
for
smaller
thicker
components.
Another
disadvantage
of
RFI
is
the
relatively
high
costs
of
the
resin
film,
which
can
cost
up
to
two
times
the
price
of
the
pure
resin,
as
well
as
their
difficulty
to
handle.
1,2
3.2.3
SCRIMP
Seemann
Composite
Resin
Infusion
Process
(SCRIMP)
is
a
mixture
of
both
the
RTM
and
the
RFI
consolidation
processes.
SCRIMP
uses
a
liquid
resin
from
an
external
source,
like
with
RTM,
and
impregnates
the
preform
in
the
thickness
direction,
like
with
RFI.
To
achieve
resin
absorption
in
the
thickness
direction,
a
resin
distribution
medium
is
used.
This
medium
allows
the
resin
to
flow
quickly
over
the
preform
surface,
spreading
over
the
entire
surface
and
then
being
absorbed
in
a
similar
fashion
to
RFI
through
the
thickness
of
the
component.
The
preparation
is
similar
to
RFI,
with
the
layering
of
the
components
and
sealing
in
a
vacuum
bag.
The
prepared
setup
is
then
placed
under
vacuum
and
the
resin
is
sucked
into
the
freeform
through
a
resin
inlet
port.
The
vacuum
created
pressure
gradient
provides
the
driving
force
for
resin
infusion
and
no
other
injection
equipment
is
needed.
This
process
has
an
advantage
in
that
tooling
costs
are
cut
similar
to
RFI,
as
well
as
cost
reductions
in
the
raw
materials
as
in
RTM.
The
limitations
of
thickness
and
maximum
length
are
also
overcome
in
this
process.
1,2
3.3
Consolidation
Equipment
Just
as
for
the
selection
of
consolidation
techniques,
the
selection
of
equipment
to
optimize
composite
consolidation
is
based
off
of
many
variants
such
as
production
quantities
and
qualities,
material
selection,
process,
etc.
In
33
this
section
we
will
briefly
discuss
the
equipment
used
for
the
three
main
processes
described
above.
3.3.1
Tooling
(mold)
For
the
RTM
process
a
closed
mold
is
used,
meaning
that
the
preform
is
completely
enclosed
by
a
mold,
while
for
the
RFI
and
SCRIMP
processes
a
one
sided
mold
or
open
mold
is
used.
The
most
important
consideration
is
the
material
used
to
produce
the
mold.
This
depends
on
cost
and
production
volume.
For
low
production
volume,
wood
and
plaster
are
generally
used
to
make
the
mold
due
to
the
ease
of
mold
production
and
low
costs.
However
for
large
production
volumes
(10,000s)
metals
such
as
steel
and
aluminum
are
chosen.
For
high
production
volumes,
it
is
more
cost
efficient
to
use
metals
due
to
their
high
durability
the
need
for
repair
or
replacement
is
greatly
reduced.
Also
metals
tend
to
produce
higher
quality
surface
finishes
and
allow
for
a
wider
range
of
processing
temperatures.
3.3.2
Heating
and
Cooling
As
with
all
the
other
equipment,
the
heating
and
cooling
systems
are
dependent
on
the
consolidation
process.
For
RFI
and
SCRIMP,
using
an
open
mold,
it
is
more
cost
effective
to
use
an
external
heating
sources,
such
as
a
convection
oven,
autoclave
(figure
3.3)
or
other
similar
heating
devices.
The
heating
system
selected
will
depend
on
component
size,
shape,
required
heating
rate
and
curing
temperature.
Cooling
is
generally
achieved
through
air-‐cooling.
2,4
Figure
3.3:
Autoclave
for
composite
consolidation.
Image
provided
by
AAC
research
For
RTM
heating
using
external
sources
becomes
too
expensive.
Here
it
becomes
more
cost
effective
to
use
an
integrated
heating
system.
This
system
consists
of
a
series
of
internal
channels
that
allow
for
temperature
controlled
34
water
or
oil
to
flow
through
the
mold.
The
heat
is
transferred
between
the
water
and
the
mold
to
control
the
temperature
of
the
mold.
2
3.3.3
Injection
Equipment
Injection
equipment
is
specific
to
the
RTM
process
as
it
is
the
only
process
requiring
the
pressurized
injection
of
resin
into
the
mold.
This
equipment
generally
consists
of
a
resin
storage
area,
resin
feed
apparatus
and
delivery
nose
and
is
highly
dependent
on
resin
choice
and
resin
handling
requirements.
The
first
option
to
consider
is
whether
to
have
the
resin
injection
controlled
by
constant
pressure
or
constant
flow
rate.
Constant
pressure
injection
is
beneficial
in
the
sense
that
pressures
can
be
controlled
and
therefore
will
not
exceed
equipment
capabilities,
however
with
constant
pressure
the
flow
rate
will
decrease
as
the
preform
becomes
impregnated
with
resin
and
if
resin
cure
rates
are
not
controlled
defects
may
form
in
the
final
component.
Constant
flow
rate
ensures
that
the
preform
is
impregnated
at
a
constant
rate
and
therefore
pre-‐
curing
is
no
longer
a
problem,
however
pressures
required
to
maintain
flow
rate
may
be
extremely
high,
requiring
expensive
equipment.
2
The
second
option
to
consider
is
whether
to
have
resin
and
hardener
injected
together
or
separately.
Premixed
and
simultaneously
injected
resin
and
hardener
have
better
mixing
and
curing
control.
The
equipment
is
simpler
and
cheaper
with
a
higher
flexibility
to
change
between
resins
and
lower
maintenance
costs.
However,
having
the
resin
and
hardener
premixed
runs
the
risk
of
curing
occurring
in
the
reservoir,
therefore
only
limited
amounts
can
be
stored
in
the
reservoir
at
a
time
and
often
leads
to
excess
waste.
Therefore
it
is
generally
better
suited
for
components
produced
in
low
volume
or
using
different
resin
systems.
On
the
other
hand,
separately
injected
resin/hardener
systems
reduce
waste
as
this
system
mixes
only
the
required
amounts
at
any
given
time.
However
it
is
difficult
to
switch
between
different
resins
and
due
to
the
increased
complexity
of
the
equipment,
maintenance
costs
are
increased.
Therefore
this
system
is
generally
used
in
production
lines
where
large
numbers
of
components
are
produced
and
flexibility
is
not
as
important.2
3.4
Optimization
Optimization
of
the
consolidation
process
is
very
important
to
maintain
product
quality,
reduce
waste,
and
reduce
costs.
Optimization
involves
the
correct
selection
of
materials,
equipment,
and
processing
requirements.
As
stated
above,
resin
selection
is
the
most
important
determinant
of
injection
equipment.
The
selection
of
a
resin
is
based
off
of
both
component
requirements
(mechanical,
environmental,
health,
and
costs)
and
the
manufacturing
process.
Here
we
will
focus
on
the
consideration
given
to
the
manufacturing
process
when
selecting
a
resin.
The
fist
most
important
consideration
is
the
viscosity
of
the
resin.
The
viscosity
must
allow
for
complete
infusion
of
the
preform
without
the
need
of
excessive
pressure.
Generally
the
pressure
range
must
fall
between
100kPa
–
700kPa,
and
obviously
pressure
is
determined
by
the
viscosity
of
the
resin
combined
with
the
permeability
of
the
preform
which
is
a
factor
of
fiber
35
volume
and
injection
distance,
however
it
is
a
general
rule
that
resin
viscosity
should
not
exceed
500cps
during
molding.2
Using
Darcy’s
law
we
can
relate
pressure,
flow
rate
and
resin
viscosity,
preform
permeability,
pressure
gradient,
and
injection
distance.1,2
Flow
Rate
=
[(Permeability
x
Cross
Section)/Viscosity]
x
(Pressure
drop/Length)
Re-‐written
in
variables:
〈u〉
=
(-‐K/η)⋅∇〈P〉f
(eq.
3.1)
〈u〉
=
resin
velocity
vector
averaged
over
fluid
volume
K
=
permeability
tensor
of
the
textile
preform
η
=
resin
viscosity
∇〈P〉f
=
pressure
gradient
averaged
over
the
fluid
volume
As
thermosets
are
the
most
commonly
used
resins
for
LCM
another
important
factor
to
consider
is
the
relationship
between
viscosity,
temperature
and
setting
time.
This
plays
an
important
role
in
the
selection
of
pressure
vs.
flow
rate
injection
equipment.1,2
Figure
3.4:
Viscosity
vs.
Time
–
temperature
dependence
of
thermoset
TGDDM
resin
Image
taken
from
Understanding
of
Rheology
of
Thermosets3
As
seen
in
the
illustration
(Figure
3.4),
initial
viscosity
of
thermoset
resins
decrease
with
increasing
temperatures,
however
curing
rate
increases
with
increasing
temperature,
causing
the
viscosity
to
increase
over
time.
Therefore
it
is
important
to
find
the
correct
balance
between
the
viscosity,
temperature
and
curing
time.2,3
The
architecture
of
the
textile
preform
forms
a
complex
network
of
channels
through
which
the
resin
flows.
Certain
architecture
types
can
create
preferential
flow
directions,
which
can
lead
to
entrapped
air.
Using
equation
3.1
above
it
is
possible
to
derive
an
equation
describing
the
mold
filling
process
by
taking
the
partial
derivative
of
flow
rate
with
respect
to
time:1,2
36
The
resin
is
assumed
incompressible,
therefore:
δ〈u〉/δt
=
0
Viscosity
is
assumed
constant,
therefore:
δ(K⋅∇〈P〉f)/δt
=
(δK/δt)⋅∇〈P〉f+
K⋅(δ∇〈P〉f/δt)
=
0
Boundary
Conditions:
Mold
walls:
n
(K∇)
=
0
Flow
front:
P
=
0
Injection
gates:
P=Pi
In
order
to
minimize
defects
in
the
resin
injection
process
it
is
common
to
us
Liquid
Molding
Simulation
(LIM).
LIMs
use
the
above
equations
with
boundary
conditions
to
preform
a
finite
element
analysis
simulating
resin
flow
in
the
mold
cavity.
The
variables
are
most
often:
Mold
geometry,
resin
and
preform
properties,
gate
location,
and
injection
conditions.
Using
LIMs
it
is
possible
to
optimize
resin,
gate
location,
and
determine
minimum
fill
time,
possible
disturbances
and
problems
in
the
filling
process
and
injection
conditions
to
cut
costs
and
reduce
defects.1
37
References
1. Long,
A.C.
Design
and
Manufacture
of
Textile
Composites.
Woodhead
Publishing
Limited,
Cambridge
England.
2005.
2. Tong,
L.
Mouritz,
A.P.
and
Bannister,
M.K.
3D
Fibre
Reinforced
Polymer
Composites.
Elsevier
Science
Ltd.
Oxford,
UK.
2002.
3. Franck,
A.J.
Understanding
the
Rehology
of
Thermosets.
TA
Instruments
2004.
4. http://www.aac-‐
research.at/products/products_AAC_CompositePolymer_de.html
38
39
Figure
4.1:
Biaxial
tension
test
(left)
and
Picture
frame
test
(right).8
The
picture
frame
test
is
usually
employed
to
test
high
shear
angles
and
produces
uniform
shear
throughout
the
sample.
The
fabric
sample
is
mounted
into
a
hinged
frame
paying
close
attention
to
ensure
the
fibers
are
parallel
to
the
frame,
as
a
slight
variation
in
fiber
direction
will
cause
drastically
different
results
(see
figure
4.1).
The
two
opposite
corners
are
pulled
in
opposite
directions.
In
order
to
achieve
pure
shear
the
yarns
in
the
frame
should
be
free
to
rotate,
however
this
is
extremely
hard
to
achieve
and
therefore
in
most
cases
the
fabric
edges
are
held
with
a
firm
grip.
Having
a
firm
grip
causes
the
fibers
near
the
frame
to
bend
leading
to
a
difference
between
the
shear
angle
in
the
fabric
and
in
the
frame.
In
order
to
reduce
the
effects
of
the
firm
grips,
it
is
common
to
use
a
cross-‐shaped
sample
and
remove
the
parallel
yarns
near
the
frame.
Direct
measurement
of
axial
load
and
shear
angle
is
possible
through
the
following
relationships:7,
8,
10
(4.1)
(4.2)
Shear
force
(Fs)
is
determined
by
the
axial
force
(Fx),
the
slide
length
of
the
shear
frame
(l)
and
the
frame
angle
(Φ).
While
frame
angle
can
be
determined
directly
from
crosshead
displacement
(Dx).
Shear
angle
can
then
be
determined
from
the
frame
angle.
(4.3)
Photo
camera
and
video
cameras
are
employed
to
measure
yarn
width,
pitch
and
estimate
lock
angle.
As
mentioned
above,
lock
angle
characterizes
the
deformation
limit,
or
the
maximum
deformation
before
wrinkling
occurs.
In
40
order
to
determine
the
wrinkling
point,
samples
are
marked
with
horizontal
lines,
wrinkling
occurs
when
the
lines
buckle.7,
8,
10
4.3
In-Plane
Biaxial
Tension
Woven
fabrics
are
composed
of
perpendicular
interlacing
yarns,
which
are
in
turn
composed
of
fibers.
The
cross
sectional
area
of
the
fibers
is
so
small
that
it
is
assumed
they
are
only
subject
to
tensile
stress
in
the
fiber
direction.
The
interlacing
of
the
yarns
causes
them
to
become
wavy,
or
crimped.
When
placed
under
tension
the
yarns
begin
to
straighten.
If
tension
is
applied
in
only
one
direction
the
yarns
in
that
direction
will
straighten
completely
while
the
crimp
of
the
interlacing
yarns
will
increase
to
accommodate
the
straightening
of
the
other
yarns.
On
the
other
hand,
if
both
sets
of
yarns
are
placed
under
tension
a
crimp
equilibrium
will
be
reached.
This
is
a
biaxial
phenomenon.
Due
to
the
biaxial
behavior
and
yarn
undulations
the
tensile
behavior
of
a
woven
fabric
is
non-‐linear
at
low
tensions
(see
figure
4.3).1,
3,
5,
7,
8
Figure
4.3:
Illustrating
non-‐linear
behavior
of
woven
fabric8
Characterization
of
the
biaxial
tension
behavior
of
woven
fabrics
is
achieved
using
a
biaxial
testing
machine.
The
machine
is
equipped
with
four
independently
controlled
axes.
These
axes
are
computer
directed
allowing
for
independent
control
of
direction,
distance
and
speed
of
axes
displacement.
Figure
4.4:
Biaxial
testing
Machine10
41
Figure
4.5:
Biaxial
testing
sample8
The
fabric
sample,
in
the
shape
of
a
cross,
is
attached
using
free
moving
clamping
rigs
(figure
4.4
and
4.5).
The
clamps
are
equipped
with
load
cells
to
record
forces
(see
figure
4.6).
Deformations
are
recorded
using
optical
techniques.
A
common
technique
is
to
use
a
monochromatic
CCD
video
camera
to
record
the
changes
in
fabric
geometry
and
use
digital
image
correlation
(DIC)
to
quantify
the
data.
DIC
requires
a
camera
to
be
positioned
perpendicular
to
the
fabric
surface
and
the
assumption
that
there
are
no
out
of
plan
deformations.
DIC
records
the
change
in
the
surface
of
the
sample
by
a
series
of
images
taken
of
the
sample
at
different
deformation
stages.
This
requires
points
of
reference,
with
textiles
this
can
be
naturally
created
from
weave
and
surface
textures,
or
a
speckled
paint
pattern
can
be
applied.
6
Figure
4.6:
Clamps10
In
the
end,
the
DIC
and
the
forces
recorded
on
by
the
clamps
gives
a
mapping
of
the
deformation
concentrations
for
the
induced
stresses.
This
data
is
important
in
knowing
how
the
fabric
reacts
under
biaxial
tension,
and
can
lead
to
a
better
understanding
of
how
the
fabric
will
behave
both
in
the
composite
as
well
as
during
forming.
42
References
1. Buet-‐Gautier,
K.
and
Boisse,
P.
Experimental
Analysis
and
Modeling
of
Biaxial
Mechanical
Behavior
of
Woven
Composite
Reinforcements.
Experimental
Mechanics.
Vol.
41,
No.
3,
September
2001.
2. Bogdanovich,
A.E.,
and
Pastore,
C.M.
Mechanics
of
Textile
and
Laminated
Composites.
Chapman
&
Hall,
1996.
3. Gasser,
A.,
Boisse,
P.,
Hanklar,
S.
Mechanical
Behavior
of
dry
Fabric
Reinforcements.
3D
Simulations
Versus
Biaxial
Tests.
Elsevier.
Computational
Material
Science
17
(2000)
7-‐20.
4. Ko,
Frank
K.,
and
Chou,
Tsu-‐Wei.
Textile
Structural
Composites
North
Holland,
1989.
5. Launay,
Jean,
Lahmar,
Fathia,
Boisse,
Philippe,
Vacher,
Pierre.
Strain
Measurement
in
tests
on
Fibre
Fabric
by
Image
Correlation
Method.
Advanced
Composites
Letters,
Vol.
11,
No.
1,
2001.
6. Lomov,
S.V.,
Boisse,
Ph.,
Deluycker,
E.,
Morestin,
F.,
Vancloster,
K.,
Vandepitte,
D.,
Verpoest,
I.,
Willems,
A.
Full-field
strain
measurements
in
textile
deformability
studies.
Composites:
Part
A
39
(2008)
1232-‐1244.
7. Lomov,
S.V.,
Willems,
A.,
Barburski,
M.,
Stoilova,
Tz.,
Verpoest,
I.
Experimental
Textile
Mechanics:
Characterization
of
Deformability
of
Reinforcements
for
Textile
Composites.
http://www.mtm.kuleuven.ac.be/research/c2/poly/index.htm
8. Long,
A.C.
Design
and
Manufacture
of
Textile
Composites.
Woodhead
Publishing
Limited,
Cambridge
England.
2005.
9. Luo,
Y.
and
Verpoest,
I.
Biaxial
tension
and
ultimate
deformation
of
knitted
fabric
reinforcements.
Department
of
Metallurgy
and
Materials
Engineering,
Katholieke
Universiteit
Leuven,
Belgium.
Composites:
Part
A
33
(2002)
197-‐203.
10. Quaglini,
Virginio,
Corazza,
Carola,
Poggi,
Carlo.
Experimental
Characterization
of
Orthotropic
Technical
Textiles
Under
Uniaxial
and
Biaxial
loading.
Composites:
Part
A
39
(2008)
1331-‐1342.
43
Composite
Modeling
5
5.1
Introduction
Composites
are
composed
of
a
reinforcing
textile
or
fiber
and
matrix
materials.
Their
properties
depend
on
the
mechanical
properties
of
each
individual
component
as
well
as
the
interaction
between
them.
For
simplicity
sake
let
us
consider
a
composite
embedded
with
unidirectional
parallel
fibers.
On
a
microscopic
scale
the
composite
would
be
considered
heterogeneous,
due
to
the
alteration
between
sections
of
fiber
and
matrix
(see
figure
5.1).
However,
due
to
the
scale
in
which
composites
are
used
and
the
impracticality
of
modeling
a
heterogeneous
microstructure,
composites
are
usually
assumed
to
have
a
homogeneous
macrostructure.
Therefore,
when
modeling
composite
behavior,
it
is
extremely
important
to
select
a
representative
volume
that
accommodates
all
demonstrated
properties
of
the
composite
on
the
microscopic
level.12,
18,
23
Figure
5.1:
Illustrating
the
microscopic
heterogeneity
of
a
composite
structure.
Fibers
shown
in
grey
and
matrix
in
blue.
It
is
also
important
to
distinguish
between
isotropic,
anisotropic
and
orthotropic
behaviors.
Composites
are
known
for
their
anisotropy
or
differing
mechanical
properties
depending
on
orientation,
and
fall
under
orthotropic
materials
as
they
generally
have
three
axes
of
symmetry
around
which
the
properties
remain
the
same
(see
figure
5.2).
12,
23
44
Figure
5.2:
Illustrating
the
three
planes
of
symmetry
that
make
composites
orthotropic
materials.
Planes
are
shown
in
yellow.
Modeling
of
composite
materials
is
a
very
useful
tool
in
determining
the
mechanical
properties,
as
experimental
methods
tend
to
be
extremely
expensive
and
impractical.
In
this
chapter
we
will
review
the
fundamentals
on
which
composite
modeling
is
based
as
well
as
some
of
the
more
widely
used
methods
for
modeling
3D
woven
composites.
12,23
5.2
Fundamentals
To
begin
let
us
review
the
fundamentals
that
are
required
for
composite
modeling.
The
first
being
the
generalized
Hooke’s
law
-‐
the
linear
constitutive
relationship
between
stress
and
strain.
It
is
represented
by
the
following:
[σ]
=
[C][ε]
(5.1)
(5.2)
45
[C]
is
the
elastic
stiffness
constant
and
[S]
is
the
inverse
matrix
of
[C].
Leading
to
the
following
relationships:
(5.3)
(5.4)
(5.5)
Where
ui
(i=1,2,3)
are
the
displacements
in
Cartesian
coordinates
and
xj
(j=1,2,3)
are
the
coordinates.
The
above
relationships
can
be
further
simplified
for
orthotropic
materials
as
they
have
the
three
planes
of
symmetry.
Becoming:
(5.6)
(5.7)
46
Where
E1,
E2,
E3
are
the
three
elastic
moduli,
G12,
G23,
G31
are
the
three
shear
moduli,
and
ν12,
ν23,
ν31
are
the
three
independent
Poisson’s
ratios.
Remember
that:
(5.8)
Therefore
the
constitutive
relationship
for
elastic
behavior
of
orthotropic
materials
becomes:
ε1
=
1/E1
(σ1
-‐
ν12σ2
-‐
ν13σ3)
ε2
=
1/E2
(σ2
-‐
ν21σ1
-‐
ν23σ3)
(5.9)
ε3
=
1/E3
(σ3
-‐
ν31σ1
-‐
ν32σ2)
γ12
=
τ12/G12
γ23
=
τ23/G23
γ13
=
τ13/G13
5.3
Representative
Volume
As
discussed
previously,
it
is
very
important
to
select
a
good
representative
volume.
The
representative
volume
must
be
large
enough
to
include
all
microstructural
features.
Think
of
it
as
a
type
of
monomer,
or
building
block
that
when
replicated
is
capable
of
reconstructing
the
composite.
By
determining
the
mechanical
properties
of
the
representative
volume,
we
can
use
continuum
mechanics
to
reproduce
the
properties
of
the
material
as
a
whole.12,
23
For
a
representative
volume
subject
to
a
homogeneous
macroscopic
stress
or
strain
and
no
body
forces,
the
average
stress
and
strains
are
defined
as
the
sum
of
the
micro-‐stresses
and
micro-‐strains
in
the
representative
volume,
divided
by
the
volume
and
can
be
represented
by
the
following
equations:
(5.10)
σij
and
εij
are
the
true
stresses
and
strains
(micro
stresses
and
strains)
in
the
representative
volume
V.
12,
23
The
boundary
conditions
for
iso-‐strain
and
iso-‐stress
on
the
representative
volume
are
expressed
by
the
following:
Iso-‐strain:
(5.11)
47
Iso-‐stress: (5.12)
When
there
is
perfect
interfacial
bonding
between
the
representative
volumes,
then
=
and
intrinsically
=
.
nj
is
the
unit
normal
vector
pointing
away
from
the
surface
of
the
representative
volume.
23
By
applying
the
homogeneous
boundary
conditions
to
the
stress-‐strain
relationship
we
can
define
the
effective
properties
of
the
representative
volume
as
follows:
Iso-‐strain:
(5.13)
(5.14)
Iso-‐stress:
(5.15)
(5.16)
Solutions
for
true
stress
and
strain
can
be
obtained
with
either
analytical
or
numerical
approaches.
Analytical
approaches
are
cheaper
and
faster,
but
require
a
large
number
of
assumptions
and
therefore
may
exclude
certain
characteristics
and
their
input.
Finite
element
methods
(numerical)
allow
for
more
accurate
modeling,
however
can
be
quite
expensive.12,
23
48
5.4
Rule
of
Mixtures
The
rule
of
mixtures
is
very
useful
in
many
analytical
approaches
for
determining
the
longitudinal
modulus
of
unidirectional
composites
and
the
major
Poisson’s
ratio
from
the
properties
of
the
individual
components.
However
for
determining
the
transverse
and
shear
moduli
modifications
must
be
made.
The
assumptions
made
with
the
rule
of
mixtures
are
that
the
composite
is
subject
to
uniform
or
iso-‐stress
and
iso-‐strain
conditions,
no
transverse
stresses,
and
that
the
load
carried
by
the
fiber
and
matrix
is
proportional
to
both
their
moduli
and
cross-‐sectional
area.
For
determining
longitudinal
modulus
it
is
also
assumed
that
the
fiber
and
matrix
are
elastic
bodies
acting
in
parallel
resulting
in
the
following
relationships:
(5.17)
(5.18)
Vf
and
Vm
representing
the
volume
fractions
of
the
fiber
and
matrix
respectively,
and
Ef,
Em,
νf,
νm
represent
the
elastic
moduli
and
Poisson’s
ratios
of
the
fiber
and
matrix
components.13,
23
To
approximate
the
transverse
elastic
properties
of
a
composite,
the
fiber
and
matrix
components
are
assumed
to
be
elastic
bodies
in
series,
however
this
is
an
inaccurate
method
for
approximation
because
in
reality
the
transverse
elastic
properties
lay
between
the
series
and
parallel
models.12,
23
(5.19)
In
order
to
better
approximate
the
transverse
properties
of
a
composite
from
the
individual
component
properties,
Halpin-‐Tsai
equations
offer
a
good
approximation
by
taking
into
account
the
fact
that
the
properties
lay
somewhere
between
a
parallel
and
series
model.12,
23
(5.20)
(5.21)
49
Where:
M
represents
E2,
G12,
G23
of
the
composite
Mf/m
represents
Ef/m
or
Gf/m
of
the
fiber
or
matrix
Vf
is
the
fiber
volume
fraction
ξ
is
a
constant
representing
the
way
the
load
is
shared
between
the
fiber
and
the
matrix.
(ξ=0
series,
ξ=∞
parallel)
5.5
Basic
Models
for
2D
Woven
Composites
The
first
step
to
applying
a
model
is
to
choose
the
representative
volume
of
a
woven
composite.
Fortunately
woven
textiles
are
formed
by
a
repetitive
geometric
pattern
represented
by
the
weave
pattern.
This
makes
selection
of
the
representative
volume,
or
unit
cell,
relatively
easy
as
it
is
the
smallest
volume
that
represents
the
weave
pattern.
This
pattern
is
represented
by
the
number
of
interlacing
weft
( )
and
warp
( )
yarns.
23
For
plain
weave
= =2
For
twill
weave
= =4
For
Satin
weave
(5-‐harness)
= =5
This
pattern
or
weave
geometry
can
be
seen
in
figure
5.3
below,
demonstrating
the
most
common
weave
geometries.
Figure
5.3:
Examples
of
common
weave
geometries
Let
us
begin
by
discussing
several
basic
1D
methods
for
modeling
2D
woven
composites:
Mosaic
model
and
Fiber
Undulation
model.
These
methods
are
considered
1D
as
they
either
do
not
consider
fiber
undulation
or
consider
fiber
undulation
in
only
one
direction.
12,
23
5.5.1
Mosaic
Model
The
mosaic
model
idealizes
the
composite
as
a
grouping
of
asymmetrical
cross-‐ply
laminates
that
can
then
be
modeled
using
the
classical
laminate
theory
(CLT)
neglecting
shear
deformation
in
the
thickness
direction.
50
(5.22)
(5.23)
Figure
5.4:
CLT
modeling
of
a
layered
composite
Using
the
above
assumptions
the
constitutive
equations
are
given
by
the
following:
(5.24)
51
(5.25)
N
–
Stress
resultant
M
–
Bending
moment
ε
-‐
In-‐plane
strains
(mid-‐plane)
k
–
Curvature
(mid-‐plane)
Aij
–
In-‐plan
stretching
stiffness
Bij
–
Bending/stretching
coupling
Dij
–
Bending
stiffness
Qij
–
Elastic
constants
of
a
lamina
Subscript
k
referring
to
the
kth
layer
and
z
referring
to
the
distance
between
the
mid-‐plane
and
the
layers
boundary
as
seen
in
figure
5.4.
If
we
assume
an
iso-‐strain
field
in
the
middle
plane,
and
that
our
composite
can
be
modeled
as
an
idealized
asymmetrical
cross-‐ply
laminate,
then
we
can
express
the
stiffness
constants
by
the
following
equations:
(the
simplification
is
only
applicable
for
uniform
weave
geometries
and
not
hybrid
weaves)
2,
3,
11,
16,
18,
23
(5.26)
5.5.2
Undulation
Model
While
the
mosaic
model
is
good
for
predicting
the
upper
and
lower
bounds
for
effective
stiffness
and
compliance
constants
for
a
unit
cell
of
a
woven
composite,
it
does
not
take
into
consideration
non-‐uniform
stresses
and
strains
that
tend
to
concentrate
in
the
fiber
interlacing
regions
or
fiber
undulation.
The
fiber
undulation
model
was
created
as
a
more
accurate
prediction
model.
2,
3,
23
52
Figure
5.5:
Fiber
undulation
model
It
is
assumed
that
the
fiber
undulation
in
weft
and
warp
yarns
can
be
expressed
using
the
following
sinusoidal
equations:
Weft:
(5.27)
Warp:
when
a0
<
x
<
a/2
(5.28)
au
representing
the
length
of
the
undulation
and
a0=(a-‐
au)/2
and
a2=(a+
au)/2.
Undulation
in
the
warp
direction
is
neglected.
Refer
to
figure
5.5
for
definitions
of
other
variables.
As
can
be
seen
in
figure
5.5,
the
unit
cell
in
the
fiber
undulation
model
consists
of
two
straight
cross-‐ply
regions
and
one
undulated
region.
The
two
straight
cross-‐ply
regions
can
be
model
as
before,
while
the
elastic
stiffness
constants
of
the
undulated
weft
yarns
can
be
expressed
in
terms
of
weft
elastic
constants
(Qij)
and
the
undulation
angle
θ.
The
undulation
angle
is
represented
by
the
following
equation:
53
(5.29)
And
the
elastic
stiffness
constants
of
the
undulated
weft
yarn
are
expressed
by:
(5.30)
(5.31)
Ishikawa
and
Chou6
obtained
the
following
formulas:
(5.32)
where
superscripts
F,
W
and
M
represent
weft,
warp
and
matrix
respectively.
The
average
effective
properties
of
the
unit
cell
can
be
determined
by
the
sum
of
the
average
straight
cross-‐ply
regions
and
the
average
undulated
region.
Here
we
have
the
average
compliance
equations,
inverted
matrices:
54
(5.33)
It
is
important
to
note
that
this
model
only
takes
into
consideration
undulation
in
the
weft
yarn,
not
in
the
warp
direction.
To
take
into
consideration
yarn
undulation
in
both
directions
a
2D
model
is
needed.
Further,
to
take
into
consideration
yarn
undulation
in
both
directions
as
well
as
out-‐of-‐plane
elastic
constants
a
3D
model
is
needed.
Reviews
of
2D
models
can
be
found
in
work
completed
by
Naik
and
colleagues14,
15
and
for
3D
models
in
work
completed
by
Hahn
and
Pandey16
and
Vandeurzen
et
al.24,
25,
26
for
3D
models.
5.6
Models
for
3D
Woven
Composites
There
are
many
different
ways
to
model
3D
woven
composites,
many
depend
on
the
geometry
of
the
woven
fabric,
be
it
3D
orthogonal
interlock,
3D
through-‐thickness
angle
interlock
or
layer-‐to
layer
interlock
(see
figure
5.6).
The
three
main
models
we
will
be
discussing
here
will
be
the
orientation
averaging
model,
mixed
iso-‐stress
and
iso-‐strain,
and
finite
element
applications.2,
3,
11,
12,
16,
23
Figure
5.6:
Types
of
3D
woven
fabrics
55
(5.38)
Where
σ
is
the
width
of
the
distributions.
In
order
to
reduce
the
Young’s
ξ
modulus
and
Poisson
ratio
caused
by
waviness,
a
waviness
knockdown
factor
is
introduced:
This
method
provides
respectable
predictions
of
the
in-‐plane
macroscopic
elastic
constants
and
a
reasonable
estimation
of
the
through-‐thickness
elastic
constants.
2,
3,
22
5.6.2
Iso-strain
and
Iso-stress
model
For
prediction
of
mechanical
and
thermo-‐elastic
properties
of
3D
orthogonal
and
angle-‐interlock
composite
materials,
Tan
et
al.20,
21,
22
proposed
a
mixed
iso-‐stress
and
iso-‐strain
based
unit
cell
modeling
method.
For
a
3D
orthogonal
fabric
we
consider
the
structure
to
be
simplified
into
rectangular
cross-‐sectional
shapes
in
three
mutually
orthogonal
directions.
An
example
of
a
unit
cell
is
shown
in
figure
5.7
(a).
56
Figure
5.7:
a)
Unit
cell
for
the
mixed
iso-‐strain
and
iso-‐stress
model.
b)
Division
of
the
unit
cell
into
4
blocks
The
Unit
cell
is
further
divided
into
smaller
blocks,
which
can
then
be
divided
again
to
create
individual
unidirectional
composite
blocks
(see
figure
5.7
(b)).
The
properties
of
each
unidirectional
block,
A,
B,
C,
D,
E,
F,
G
can
be
determined
using
methods
described
previously.
To
determine
the
overall
properties
of
the
unit
cell
it
is
possible
to
assemble
the
blocks
individually
in
the
x-‐,
y-‐
,
and
z-‐directions
(see
figure
5.8)
Figure
5.8:
Possible
assembly
directions
of
block
A
and
B
57
For
an
x-‐assemblage
of
NA
number
of
A
blocks
and
NB
number
of
B
blocks
in
the
strip,
the
overall
material
properties
are
represented
by
the
following
equations:
(5.40)
Where
VA
and
VB
are
the
volume
fractions
of
block
A
and
block
B
in
the
strip,
and
subscripts
S,
A,
B
refer
to
strip,
block
A
and
block
B
respectively.
For
a
y-‐assemblage
of
A
and
B
blocks:
(5.41)
For
a
z-‐assemblage
of
A
and
B
blocks:
58
(5.42)
In
figure
5.7
b)
it
is
seen
that
the
blocks
A
and
B,
D
and
E,
and
F
and
G
can
be
evaluated
using
the
x-‐assembly
equations.
Strips
1
and
2,
and
3
and
4
can
then
be
modeled
using
the
y-‐assembly
equations.
Finally
the
top
and
bottom
planes
can
be
evaluated
using
the
z-‐assembly
equations,
estimating
the
overall
properties
of
the
complete
unit
cell.
2,
3,
12,
23
5.6.3
Finite
Element
Model
Finite
Element
Modeling
(FEM)
is
probably
the
most
accurate
method
for
predicting
elastic
and
failure
behavior
of
textile
composites.
It
allows
for
detailed
modeling
of
complex
geometries
and
varying
material
properties.
FEM
works
by
breaking
down
the
composite
structure
into
small
regions
whose
constitutive
properties
can
be
easily
evaluated.
Using
brick,
wedge
and
tetrahedral
elements
it
is
possible
to
generate
a
mesh
depicting
in
detail
the
true
geometry
of
the
unit
cell
as
seen
in
figure
5.9.
Determination
of
true
fabric
geometry
in
the
composite
is
of
course
is
limited
by
current
measurement
technologies
and
may
not
be
cost
effective.
Figure
5.9:
Example
of
a
3D
FE
model
of
a
unit
cell
of
a
3D
orthogonal
Woven
composite
material.22
Yarns
are
modeled
as
orthotropic,
with
respect
to
the
principle
axes,
with
cross
sectional
shapes
of
rectangular,
circular,
elliptical
or
lenticular,
while
the
matrix
is
assumed
homogeneous
and
isotropic.
It
is
important
to
note
that
accuracy
of
FEM
depends
on
how
accurately
the
fabric
geometry
is
modeled.
2,
3,
11,
33
59
References
1.
Aboudi
J.,
1991,
Mechanics
of
Composite
Materials
-‐
A
Unified
Micromechanical
Approach,
Elsevier,
Amsterdam,
The
Netherlands.
2. Ansar,
mahmood,
Wang
Xinwei,
Zhou
Chouwei,
2011,
Modeling
strategies
of
3D
woven
composites:
A
review.
Cmoposite
structures
93,
Elsevier.
3. Bogdanovich,
A.E.,
and
Pastore,
C.M.
Mechanics
of
Textile
and
Laminated
Composites.
With
applications
to
structural
analysis.
Chapman
&
Hall,
London,
UK
1996.
4. Chou
T.W.
and
F.K.
KO,1989,
Textile
Structural
Composites,
Volume
3
Composite
MaterialsSeries,ElsevierSciencePublishers,Amsterdam,NewYork,U.S.A.,
1989.
5.
Hahn
H.T.
and
R.
Pandey,
1994,
A
micromechanics
model
for
thermoplastic
properties
of
plain
weave
fabric
composites,
J.
Eng.
Mat.
&
Tech.,
116:
517-‐423
6. Ishikawa
T.
and
T.W.
Chou,
1982a,
Elastic
behavior
of
woven
hybrid
composites,
J.
Comp.
Mat.
162-‐19.
7. Ishikawa
T.
and
T.W.
Chou,
1982b,
Stiffness
and
strength
behavior
of
woven
fabric
composites,
J.
Mat.
Sci.,
17:3211-‐3220.
8. Ishikawa
T.
and
T.W.
Chou,
1983a,
One-‐dimensional
Micromechanical
analysis
of
Woven
Fabric
Composites,
AIAA
J.,
21:1714-‐1721.
9. Ishikawa
T.
and
T.W.
Chou,
1983b,
In-‐plane
thermal
expansion
and
thermal
bending
coefficients
of
fabric
composites,
J.
Comp.
Mat.
17:92-‐
104.
10. Ishikawa
T.
and
T.W.
Chou,
1983c,Nonlinear
behavior
of
woven
fabric
composites,
J.
Comp.
Mat.
17:399-‐413.
11. Ko,
Frank
K.,
and
Chou,
Tsu-‐Wei.
Textile
Structural
Composites
12. Long,
A.C.
Design
and
Manufacture
of
Textile
Composites.
Woodhead
Publishing
Limited,
Cambridge
England.
2005.
13. Mori
T.
and
Tanaka
K.,
1973,Average
stresses
in
matrix
and
average
elastic
energy
of
materials
with
misfitting
inclusions,
Acta
Metalurgica,
21571-‐574.
14. Naik
N.K.
and
V.K.
Ganesh,
1992,
Prediction
of
on-‐axes
elastic
properties
of
plain-‐weave
fabric
composites,
Comp.
Sci.
&
Tech.,
45:135-‐152.
15. Naik
N.K.
and
P.S.
Shembekar,
1992a,Elastic
behavior
of
woven
fabric
composites:
I-‐
lamina
analysis,
J.
Comp.
Mat.
26:2197-‐2225.
16. Onate,
Eugenio
and
Kroplin,
Bern.
Textile
Composites
and
Inflatable
Structures.
Springer,
The
Netherlands.
2005.
17. Shembekar
P.S.
and
N.K.
Naik,
1992,Elastic
behaviourof
woven
fabric
composites:
II-‐
Laminate
analysis,
J.
Comp.
Mat.,
26:2226-‐2246.
18. Stig,
Fredrik.
An
Introduction
to
the
Mechanics
of
3D-Woven
Fiber
Reinforced
Composites.
KTH
School
of
Engineering
Sciences,
Stockholm,
Sweden.
April
2009.
19. Tan
P.,
L.
Tong
and
G.P.
Steven,
1997a,
Modeling
for
predicting
the
mechanical
properties
of
textile
composites
-‐
A
review,
Composites,
28A:903-‐922.
20. Tan
P.,
L.
Tong
and
G.P.
Steven,
1998,Modelling
approaches
for
3D
orthogonal
woven
composites,
J.
Rein.
Plastics
&
Comp.,
17545-‐577.
60
21. Tan
P.,
L.
Tong
and
G.P.
Steven,
1999a,
Models
for
predicting
thermo
mechanical
properties
of
3D
orthogonal
woven
composites,
J.
Rein.
Plastics
&
Comp.,
18:151.
22. Tan
P.,
L.
Tong
and
G.P.
Steven,
1999b,Micro-‐mechanics
models
for
mechanical
and
thermo-‐mechanical
properties
of
3D
angle
interlock
woven
composites,
Comp.,
30A:637-‐648.
23. Tong,
L.
Mouritz,
A.P.
and
Bannister,
M.K.
3D
Fiber
Reinforced
Polymer
Composites.
Elsevier
Science
Ltd.
Oxford,
UK.
2002.
24. Vandeurzen
P.,
J.
Ivens
and
I.
Verpoest,
1996a,
A
three-‐dimensional
micromechanical
analysis
of
woven
fabric
composites:
I.
Geometric
analysis,
Comp.
Sci.
&
Tech.,
56:
1303-‐1315.
25. Vandeurzen
P.,
J.
Ivens
and
I.
Verpoest,
1996b,
A
three-‐dimensional
micromechanical
analysis
of
woven
fabric
composites:
11.
Elastic
analysis,
Comp.
Sci.
&
Tech.,
56:
1317-‐1327.
26. Vandeurzen
P.,
J.
Ivens
and
I.
Verpoest,
1998,
Micro-‐stress
analysis
of
woven
fabric
composites
by
multilevel
decomposition, J. Comp. Mat.,
32:623-651.
61
3D
Woven
Composites
6
6.1
Introduction
3D
composites
have
been
developed
in
order
to
combat
the
problems
faced
by
2D
composites:
reduce
fabrication
costs,
increase
through-‐thickness
mechanical
properties
and
improve
impact
damage
tolerance.
In
3D
composites,
fibers
are
aligned
not
only
in
the
x-‐,
y-‐direction,
but
also
in
the
through-‐thickness
(z-‐)
direction.
By
placing
fibers
in
the
z-‐direction
mechanical
performance,
impact
damage
and
ease
of
processing
can
all
be
improved.14
3D
composites
consist
of
several
types,
including
woven,
knit,
braided,
stitched,
and
z-‐pinned.
The
first
3D
composite
to
be
manufactured
was
a
braided
carbon-‐carbon
composite
in
the
1960s
to
be
used
in
rocket
motor
components.
Soon
after
3D
woven
carbon-‐carbon
composites
were
used
for
the
brake
components
of
jet
aircrafts.
3D
composites
were
meant
to
replace
high-‐
temperature
metal
alloy
parts
to
improve
durability,
heat
distortion
and
weight.
These
early
3D
composites
may
not
have
been
FR
composites,
however
the
idea
still
remains
valid.14
6.2
3D
Woven
Composites
At
the
moment
3D
woven
composites
are
used
in
only
a
few
niche
markets.
They
show
a
great
potential
for
applications
in
the
aerospace,
marine,
infrastructure,
military
and
medical
fields,
due
to
their
benefits:
potential
reduced
fabrication
costs,
increased
design
flexibility,
improved
impact
resistance
and
through-‐thickness
mechanical
properties.
However
there
are
still
many
challenges
impeding
their
application,
such
as
current
production
cost
due
to
small
scale
production,
cost
of
certification
of
new
materials
for
load-‐bearing
structures,
and
uncertainty
of
benefits.
Only
once
these
barriers
are
overcome
will
3D
woven
composites
make
a
name
for
itself
in
the
world
of
composites.
Here
in
this
section
we
will
discuss
microstructural
features
and
their
effect
on
mechanical
properties,
delamination
resistance,
impact
damage,
as
well
as
the
different
failure
mechanisms.14
6.2.1
Microstructure
Features
and
Crimp
The
microstructure
is
determined
mainly
by
the
fiber
architecture
of
the
62
Figure
6.1:
Tensile
strength
at
different
stages
of
the
weaving
process9
This
degradation
of
tensile
strength
is
seen
in
figure
6.1
above.
A
reduction
in
tensile
strength
of
about
30%
is
seen
at
the
later
stages
of
the
weaving
process.
The
extent
of
damage
and
reduction
of
properties
depends
on
the
weave,
loom,
as
well
as
yarn
material,
type
(twisted
or
untwisted),
and
diameter.
Glass
fibers
display
the
highest
loss
of
strength
in
comparison
with
carbon
or
Kevlar.14
Distortion
of
fibers
from
their
idealized
architecture
is
due
to
the
interlacing
of
the
yarns
during
weaving.
Misalignment
and
waviness
in
3D
performs
is
much
higher
than
that
of
the
2D
performs,
as
much
as
25%
higher,
and
can
cause
a
large
degradation
in
composite
properties.5
Crimping,
or
yarn
waviness,
is
possibly
the
most
important
factor
in
determining
the
mechanical
properties
of
the
final
composite.
The
more
crimped
the
yarns,
the
lower
the
strength
of
the
composite.
When
a
crimped
structure
is
subject
to
a
tensile
force
the
yarns
begin
to
straighten,
however
the
yarns
do
not
straighten
at
the
same
rate,
as
the
crimping
is
not
uniform
throughout
the
fabric.
Therefore
stress
concentrate
in
specific
regions
causing
those
regions
to
fail
earlier
than
others.
Crimping
also
introduces
shear
stresses
into
the
matrix
(see
figure
6.2).11,
14
Crimp
has
been
defined
in
several
ways
from
crimp
%,
crimp
angle
as
well
as
defining
the
geometry
of
crimp.
Describing
the
geometry
of
crimp
is
important
when
modeling
the
composite
properties
as
discussed
previously
in
chapter
5.
The
simplest
definition
of
crimp
is
the
crimp
%
or
the
ratio
of
the
difference
between
the
yarn
length
and
fabric
length
over
fabric
length:
63
An
additional
concern
for
3D
woven
composites
is
the
crimp
induced
by
the
z-‐
binder
yarns.
They
cause
excess
bunching
of
the
surface
fibers
where
they
cross
over
the
in-‐plane
yarns.
This
reads
to
regions
rich
in
fiber
content
and
creates
problems
for
the
resin
to
enter.
These
regions
lack
resin
and
have
increased
porosity.
At
the
same
time,
the
fiber
bunching
creates
spaces
fiber
poor
that
allow
for
excess
resin
to
enter.
The
result
is
fiber
rich
and
resin
rich
areas
on
the
composite
surface.14
Figure
6.2:
Illustration
of
the
crimping
in
2D
woven
fabrics5
Figure
6.3:
difference
between
Idealized
z-‐binder
geometry
(a)
and
actual
(b)14
Misalignment
of
the
z-‐binder
can
also
occur
due
to
high
tensile
stresses
during
weaving
as
well
as
excessive
pressure
during
consolidation.
As
can
be
seen
in
figures
6.3
and
6.4
the
actual
z-‐binder
geometry
is
much
different
than
the
idealized.
This
is
important
to
note,
as
tensile
behavior
is
greatest
when
the
fiber
direction,
so
if
the
fiber
is
not
aligned
properly,
the
mechanical
behavior
will
be
greatly
reduced.14
64
Figure
6.4:
Top
and
cross
sectional
view
12
6.2.2
Tensile
Properties
The
tensile
properties
of
3D
woven
composites
are
only
recently
beginning
to
become
understood,
and
are
often
compared
with
those
of
2D
woven
composites
with
similar
fiber
content.
The
Young’s
modulus
of
3D
woven
composites
is
commonly
lower
than
the
modulus
of
the
equivalent
2D
composite,
however
other
studies
show
the
modulus
being
higher.1,9
It
is
important
to
note
that
the
Young’s
modulus
is
not
significantly
influenced
by
the
z-‐binder
content
or
fiber
structure,
and
is
most
probably
influenced
by
the
fiber
content
(increasing
with
increasing
fiber
content)
and
degree
of
fiber
waviness
(decreasing
with
increasing
waviness),
and
can
be
accurately
predicted
using
the
block
laminate
and
unit
cell
models.14
A
feature
displayed
by
3D
woven
composites,
but
not
by
2D,
is
an
onset
of
plastic
deformation
at
relatively
low
tensile
stresses,
displaying
a
reduction
of
stiffness
from
20-‐50%.
This
softening
is
due
to
the
onset
of
plastic
deformation
in
the
most
heavily
distorted
load-‐bearing
tows,
distortion
of
yarns
being
caused
by
the
z-‐binders.
The
critical
tensile
stress
at
which
plastic
deformation
of
these
yarns
occurs
can
be
estimated
by
the
following
equation:14
f s [τ13 ]
σa =
[ξ]
€ 65
Where
fs
is
the
volume
fraction
of
the
load-‐bearing
tows,
τ13
is
the
axial
shear
strength
of
the
tow,
and
ξ
is
the
fiber
waviness
parameter,
defined
as
the
average
misalignment
angle
for
90%
of
all
load-‐bearing
tows.
In
order
to
minimize
softening
it
is
necessary
to
minimize
the
in-‐plane
fiber
waviness
or
use
a
high
yield
shear
strength
resin.14
As
tensile
stresses
increase
above
the
critical
value,
the
matrix
of
3D
woven
composites
begin
to
crack,
z-‐binders
begin
to
de-‐bond.
Failure
occurs
though
rupture
of
load-‐bearing
tows.
3D
woven
composites
generally
have
a
lower
tensile
strength
compared
with
2D.
The
lower
tensile
strength
is
most
likely
due
to
fiber
damage
from
weaving,
fiber
waviness
and
pinching
of
the
surface
tows.
Prediction
of
tensile
failure
strength
of
3D
woven
composites
is
difficult
as
it
relies
heavily
on
fiber
damage
and
crimping
which
are
hard
to
accurately
measure.14
6.2.3
Compressive
properties
In
most
cases
it
has
been
found
that
the
compression
modulus
of
3D
is
lower
than
that
of
2D
woven
laminate
with
similar
fiber
content,
due
to
higher
crimping
and
increased
waviness
of
the
3D
fibers.3
Studies
for
compressive
strength
determination
are
inconclusive
as
they
show
both
an
increase
and
a
decrease
in
strength.
The
cause
of
the
increased
compressive
strength
is
unclear.
However
the
decreased
compressive
strength
is
due
to
kinking
of
the
load-‐
bearing
tows
(see
figure
6.5).
Kinking
initiates
in
regions
with
low
resistance
to
permanent
shear
deformation
such
as
defects
or
misalignments,
which
are
much
more
prevalent
in
3D
woven
composites.
Kinking
arises
when
plastic
shear
flow
of
the
matrix
surrounds
the
axial
tow,
causing
a
rotation
until
break
of
the
individual
tows.14
Figure
6.5:
Kinking
failure
in
compression
Kinking
in
3D
woven
composites
is
usually
concentrated
in
the
surface
regions,
where
the
most
severely
distorted
tows
are
located.
Kink
bands
develop
as
discrete
geometric
flaws,
which
inhibit
catastrophic
failure
and
instead
cause
the
material
to
fail
gradually
under
increasing
strain,
leading
to
significantly
higher
66
Figure
6.6:
Mode
I
delamination
cracking14
This
dramatic
increase
in
delamination
toughness
is
caused
by
the
necessity
for
the
crack
tip
to
pass
through
the
z-‐binders.
The
de-‐bonding
of
the
z-‐binders
as
the
crack
propagates
absorbs
a
small
amount
of
energy.
The
majority
of
the
toughening
is
caused
by
the
bridging
zone
(figure
6.6),
where
the
z-‐binders
support
a
large
portion
of
the
applied
force
in
turn
reducing
the
stresses
acting
on
the
crack
tip
and
increasing
de-‐bonding
toughness.
Further
toughening
occurs
from
crack
branching,
due
to
the
induced
toughness
of
the
z-‐
binders,
and
the
energy
absorbed
due
to
z-‐binder
fracture
and
de-‐bonding.
67
Figure
6.7:
Effect
of
impact
velocity
on
delamination
damage
of
2D
and
3D
woven
composites2
68
Figure
6.8:
Effect
of
impact
energy
on
flexural
strength15
Figure
6.9:
Effect
of
impact
energy
on
the
compressive
strength1
69
References
1. Arendts
F.J.,
K.
Drechler
and
J.
Brandt,
1993,
Advanced
textile
structural
composites
–
status
and
outlook,
Proc.
Of
the
Int.
Conf.
on
Advanced
Composite
Materials,
ed.T.
Chandra
and
A.K.
Dhingra,
409-‐416.
2. Billaut
F.
and
0
.
Roussel,
1995,
Impact
resistance
of
3-‐D
graphite/epoxy
composites,
Proc.
of
the
11”
Int.
Conf.
on
Comp.
Mat.,
Vol.
5,
Ed.
A.
Pourartip
and
K.
Street,
Whistler,
BC,
Canada,
14-‐18
August
1995,
Woodhead
Publishing
Ltd.,
551-‐558.
3. Brandt
J.,
K.
Drechsler
and
Fa-‐J.Arendts,
1996,
Mechanical
performance
of
composites
based
on
various
three-‐dimensional
woven-‐fibre
preforms,
Comp.
Sci.
&
Tech.,
56:
381-‐386.
4. Chou
S.,
H.-‐C.
Chen
and
H.-‐E.
Chen,
1992,
Effect
of
weave
structure
on
mechanical
fracture
behavior
of
three-‐dimensional
carbon
fiber
fabric
reinforced
epoxy
resin
composites,
Comp.
Sci.
&
Tech.,
45:
23-‐35.
5. Cox
B.N.,
M.S.
Dadhkak,
W.L.
Morris
and
J.G.
Flintoff,
1994,
Failure
mechanisms
of
3D
woven
composites
in
tension,
compression
and
bending,
Acta
Metal.
et
Mat.,
42:3967-‐3984.
6. Guenon
V.A.,
T.-‐W.
Chou
and
J.W.
Gillespie,
1989,
Toughness
properties
of
a
three-‐dimensional
carbon-‐epoxy
composite,
J.
Mat
Sci.,
24:4168-‐4175
7. James
B.
and
Howlett
S.,
1997,
Enhancement
of
post
impact
structural
integrity
of
GFRP
composite
by
through-‐thickness
reinforcement,
Proc.
of
the
2ndEuropean
Fighting
Vehicle
Symposium,
27-‐29
May,
Shrivenham,
UK.
8. KO
F.K.
and
D.
Hartman,
1986,
Impact
behavior
of
2-‐D
and
3-‐D
glass/epoxy
composites,
SAMPE
J.,
July/August,
26-‐30.
9. Lee
B.,
K.H.
Leong
and
I.
Herszberg,
2001,
The
effect
of
weaving
on
the
tensileproperties
of
carbon
fibre
tows
and
woven
composites,
J.
Rein.
Plastics
&
Comp.,20:
652-‐670.
10. Mouritz
A.
P.,
C.
Baini
and
I.
Herszberg,
1999,
Mode
I
interlaminar
fracture
toughness
properties
of
advanced
textile
fibre
glass
composites,
Composites,30:859-‐870.
11. Stig,
Fredrik.
An
Introduction
to
the
Mechanics
of
3D-Woven
Fiber
Reinforced
Composites.
KTH
School
of
Engineering
Sciences,
Stockholm,
Sweden.
April
2009.
12. Tan
P.,
L.
Tong,
G.P.
Steven
and
T.
Ishikawa.
2000a.
Behavior
of
3D
orthogonal
woven
CFRP
composites.
I:
Experimental
investigation,
Composites,
31A:259-‐271.
13. Tanzawa
Y.,
N.
Watanabe
and
T.
Ishikawa,
1997,
Interlaminar
delamination
toughness
and
strength
of
3-‐D
orthogonal
interlocked
fabric
composite,
Roc.
of
the
Eleventh
Int.
Conf.
on
Composite
Materials,
Ed.
M.L.
Scott,
14-‐18
July,
Gold
Coast,
Australia,
V-‐47
to
V-‐57.
14. Tong,
L.
Mouritz,
A.P.
and
Bannister,
M.K.
3D
Fibre
Reinforced
Polymer
Composites.
Elsevier
Science
Ltd.
Oxford,
UK.
2002.
15. Voss
S.,
A.
Fahmy
and
H.
West,
1993,
Impact
tolerance
of
laminated
and
3-‐
dimensionally
reinforced
graphite-‐epoxy
panels,
Proc.
of
the
Int.
Conf.
on
Advanced
Composite
Materials,
Ed
T.
Chandra
and
A.K.
Dhingra,
591-‐596.
70
71
Table
7.1:
Reported
results
from
Macander
et
al.
for
effects
of
braid
pattern
and
edge
conditions11
7.2.2
3D
vs.
2D
Braided
Composites
It
is
often
beneficial
to
compare
the
properties
of
new
materials
to
older
materials
in
order
to
have
a
better
idea
of
the
benefits
and
challenges
presented.
However,
this
often
difficult
as
it
is
can
be
unclear
as
to
what
variables
should
be
considered.
Gause
tested
a
1x1
and
1x1x1/2F
3D
braided
composite
against
a
24-‐ply
laminate
of
AS1/3501
prepreg
designed
to
mimic
the
3D
architecture.
The
results
for
the
3D
composite
were
as
follows:
decreased
tensile
strength
in
all
directions,
increase
in
longitudinal
compressive
properties
and
tensile
modulus,
and
increased
ability
to
retain
tensile
properties
in
fabrics
containing
open
holes.
Further
studies
are
needed
to
further
characterize
3D
braided
composites.15
7.3
3D
Knit
Composites
Of
all
the
fabric
reinforcements
used
for
composite
construction,
knit
fabric
is
the
least
understood.
Knitting
creates
a
fabric
by
looping
yarns
together,
creating
a
highly
curved
yarn
structure,
a
structure
with
relatively
low
structural
strength.
The
advantages
of
3D
knit
performs
lie
in
their
high
drapability
and
impact
damage
resistance.
They
are
ideally
suited
to
produce
non-‐structural
parts
with
complex
geometries.
Due
to
its
exceptional
impact
damage
resistance,
it
is
being
considered
for
the
potential
use
in
medical
prosthesis,
bicycle
helmets,
and
automobile
doors.15
7.3.1
In-Plane
Properties
The
in-‐plane
tensile
properties
of
3D
knit
composites
have
been
studied
the
most.
However,
little
information
was
given
on
knit
architecture
regarding
72
loop
length,
shapes
and
densities,
which
could
give
valuable
insight
into
the
mechanisms
behind
the
outcomes.
Tests
have
shown
that
the
tensile
strength
of
knitted
composites
is
similar
to
those
of
composites
constructed
with
non-‐woven
randomly
oriented
fibers,
and
that
increases
in
fiber
volume
increased
tensile
strength.8,9,13
Wu
and
Leong
studied
the
effect
of
knit
architecture
on
the
mechanical
properties.10,16
They
found
that
it
was
possible,
to
change
the
tensile
properties
from
quasi-‐isotropic
to
strongly
anisotropic
by
changing
the
knit
geometry.
The
tensile
properties
increased
with
increasing
fiber
orientation,
as
expected.
The
results
are
listed
in
tables
7.2
and
7.3
below,
and
figure
7.1
depicts
the
fiber
architectures
used
in
the
tests.
15
Table
7.2:
Tensile
properties
of
warp
knit
with
varying
knit
architectures16
Table
7.3:
Tensile
properties
of
weft
knit
with
varying
knit
architectures
10
Figure
7.1:
Warp
knit
(a)
Denbigh,
(b)
1x3
single
cord,
and
(c)
1x4
single
cord
architectures15
The
tensile
properties
can
also
be
controlled
by
changing
the
loop
length
or
stitch
density
(as
loop
length
increases
stitch
density
decreases).
Leong10
found
that
the
modulus
is
dependent
on
fiber
volume
fraction,
while
tensile
strength
and
failure
stain,
in
course
and
wale
directions,
decreases
with
decreasing
loop
length
for
both
weft
and
warp
knit
(see
figure
7.2).
On
the
other
hand
Wu16
found
the
exact
opposite;
in
the
course
direction
he
found
that
tensile
strength
increases
with
decreasing
loop
length
and
that
loop
length
had
no
effect
73
in
the
wale
direction.
This
shows
that
the
properties
of
the
knit
fabric
are
dependent
on
many
parameters
and
further
testing
is
needed.10,15,16
Figure
7.2:
Wale
and
course
directions
as
well
as
warp
and
weft
fabric
structure.
There
has
not
been
much
investigation
on
the
compressive
properties
of
knit
composites.
However,
with
what
little
has
been
done,
it
has
been
found
that
knit
composites
have
a
compressive
strength
better
than
the
tensile
strength,
while
compressive
modulus
is
about
the
same.
The
compressive
properties
do
not
depend
on
loop
length
and
are
much
more
isotropic
compared
to
tensile
properties.15
7.3.2
Interlaminar
Fracture
and
Impact
Toughness
Here
is
where
the
true
benefits
of
3D
knit
fabrics
lie.
Mouritz12
found
the
fracture
toughness
of
the
knit
fabrics
to
be
significantly
higher
than
that
for
2D
and
3D
woven,
2D
braided
and
stitched
composites.
This
is
due
to
the
looped
yarn
structure,
which
causes
crack
deflection
and
excessive
crack
branching,
consequently
increasing
the
toughness.
This
can
be
controlled
through
stitch
density,
as
density
increases
toughness
decreases,
due
to
the
tighter
structure
of
densely
stitched
knit
fabric.
Tight
structures
do
not
allow
for
intermingling,
which
is
responsible
for
crack
deflection.11,12,15
The
impact
performance
of
knit
composites
under
low
to
medium
energy
is
also
of
great
interest.
Chou3
found
that
knit
composites
are
capable
of
absorbing
up
to
2.4
times
more
impact
energy
than
woven
composites.
This
makes
knit
composites
attractive
candidates
for
components
requiring
high
impact
absorption
or
non-‐structural
damage
prone
parts.3,15
7.4
Stitched
Composites
Stitching
involves
sewing
the
laminates
in
the
z-‐direction
with
a
high
strength
thread.
The
high
strength
thread
is
inserted
through
a
stack
of
fabric
using
an
industrial
grade
sewing
machine.
The
through-‐thickness
reinforcement
of
stitched
composites
is
between
1-‐5%
and
is
comparable
to
woven,
braided
and
knitted
3D
composites.
The
act
of
stitching
the
fabric
prepregs
together
improves
the
delamination
resistance,
and
impact
damage
tolerance.
Stitching
can
also
be
used
to
connect
separate
composite
components
together,
74
eliminating
the
need
for
mechanical
fasteners,
reducing
costs
and
improving
the
joints
mechanical
properties.
Due
to
the
simple
nature
of
stitching,
it
is
possible
to
only
stitch
the
areas
that
would
benefit
from
the
z-‐direction
reinforcement,
further
reducing
costs.
Up
to
this
point
stitching
has
proven
to
be
a
simple,
flexible
and
low-‐cost
method
for
producing
3D
composites,
however
there
are
limitations
to
the
complexity
of
the
components.15
7.4.1
In-plane
Mechanical
Properties
During
the
stitching
process
damage
can
occur.
This
damage
is
the
most
important
factor
to
consider
when
considering
the
mechanical
properties.
The
most
common
forms
of
damage
are
fiber
breakage,
fiber
misalignment,
and
crimping,
all
causing
sever
decreases
in
mechanical
properties.
Other
damage
can
occur
from
micro
cracking
at
stitch
insertion
sites
and
fiber
compaction.
Due
to
the
increase
in
fiber
damage
caused
by
stitching,
stitched
composites
tend
to
have
slightly
lower
tension,
compaction
and
flexural
properties
than
unstitched
composites,
although
there
are
many
contradictions
between
data.
As
stitched
composites
have
been
extensively
studied,
there
are
databases
containing
detailed
information
on
tension,
compression,
and
bending
modulus
and
strength
of
various
materials.15
The
interlaminar
shear
properties
of
stitched
composites
have
not
been
extensively
studied.
The
few
studies
that
have
been
carried
out
show
contradicting
results
and
are
therefore
inconclusive.15
7.4.2
Fracture
Toughness
and
Impact
Damage
Tolerance
Stitched
composites
show
a
large
improvement
in
mode
I
delamination
resistance.
The
mechanism
behind
the
fracture
toughness
is
similar
to
that
for
3D
woven
composites
(see
figure
7.3).
When
a
crack
forms
and
propagates
through
the
stitches
a
bridging
zone
is
formed,
where
the
stitches
exert
a
closing
force,
lowering
strain
in
the
crack
tip.
Stitched
composites
also
show
a
large
increase
in
mode
II
delamination
toughness,
however
not
as
great
as
in
mode
I.7,15
Figure
7.3:
Illustrating
mode
I
interlaminar
toughening
mechanism
of
stitched
composites7
Due
to
the
high
fracture
toughness
of
stitched
composites
it
is
not
surprising
that
they
have
good
impact
damage
tolerance.
Most
studies
for
75
impact
damage
tolerance
have
been
for
ballistic
projectiles
and
explosion
shock
wave
as
stitched
composites
have
been
of
particular
interest
to
the
military.
In
both
cases
an
increase
in
post-‐impact
mechanical
properties
was
observed.15
7.5
Z-Pinned
Composites
Z-‐pinning
involves
the
insertion
of
short
metal
wire
or
pultruded
composite
pins
in
the
z-‐direction.
The
pins
can
be
inserted
in
uncured
prepreg
taps
or
dry
fabrics
(see
figure
7.4).
For
further
details
on
the
z-‐pinning
process
please
refer
to
Chapter
2.
As
this
is
a
relatively
new
technology
the
full
potential
and
application
possibilities
are
currently
being
investigated.
At
this
point
z-‐
pinning
has
demonstrated
the
ability
to
increase
joint
strength,
remove
the
need
for
fasteners
and
create
a
more
evenly
distributed
the
load
over
a
join
region.
However,
there
is
not
much
data
on
mechanical
properties
such
as
flexural
strength,
shear
strength,
fatigue,
etc.
The
data
that
will
be
reported
here
will
consist
of
tensile
and
compressive
strength,
and
delamination
properties.15
Figure
7.4:
Depiction
of
z-‐pinned
architecture
at
insertion
site14
7.5.1
Tensile
and
Compressive
Strength
Steeves
and
Fleck
investigated
the
effect
of
z-‐pinning
on
tensile
and
compressive
strength.14
Their
results
showed
an
average
decrease
in
tensile
strength
of
about
27%,
while
the
modulus
remained
same.
This
decrease
in
tensile
strength
is
believed
to
be
caused
by
the
stress
concentration
and
fiber
damage
at
the
pin
insertion
sites.
Freitas
showed
the
same
results
for
z-‐pinned
materials
containing
more
than
1.5%
pinning.
Below
1.5%
the
tensile
properties
remained
almost
unchanged.4,5,6,14,15
The
effect
of
z-‐pinning
on
compressive
properties
has
proven
to
be
very
similar
that
of
tensile,
with
a
30%
decrease
in
strength.14
However,
it
was
also
found
that
there
is
a
strong
correlation
between
fiber
misalignment
and
decrease
in
composite
strength.
This
correlation
is
caused
by
weaving
of
in-‐plane
yarns
around
the
z-‐pin
(see
figure
7.5),
as
weaving
increases,
compressive
strength
decreases.
Weaving
increases
as
the
insertion
angle
of
the
z-‐pin
increases.
Insertion
angles
of
0°
showed
the
least
weaving,
while
weaving
increased
for
23°
and
45°
specimens.4,5,6,14,15
76
Figure
7.5:
Depiction
of
weaving
and
deflection
caused
by
z-‐pins14
7.5.2
Delamination
resistance
Z-‐pinned
composites
show
a
high
resistance
to
interlaminar
cracking
and
through-‐thickness
failure.
Cartei
and
Partridge
investigated
the
properties
under
mode
I
and
II
failure.
For
mode
I
failure
the
composite
showed
a
rapid
increase
in
delamination
resistance
with
increasing
z-‐fiber
content
and
decreasing
pin
diameter.1,2,15
The
mechanisms
behind
the
increased
resistance
are
similar
to
that
of
3D
woven
and
stitched
composites.
The
presence
of
the
z-‐directional
fibers
creates
the
bridging
zone,
which
relieves
strain
from
the
crack
tip.
For
more
details
on
the
mechanism
see
chapter
6.15
77
References
1. Cartei
D.D.R.
and
I.K.
Partridge,
1999a,
Delamination
behavior
of
Z-‐pinned
laminates,
Proc.
12*
Int.
Conf.
Comp.
Mat.,
5-‐9
July,
Paris.
2. Cartei
D.D.R.
and
I.K.
Partridge,
1999b,
Delamination
behavior
of
Z-‐pinned
laminates,
Proc.
2"'
ESIS
TC4
Conf.,
Ed.
J.
G.
Williams,
13-‐15
Sept.,
CH-‐
Les
Diablerets,
Elsevier.
3. Chou
S.
and
C-‐J.
Wu,
1992,
A
study
of
the
physical-‐properties
of
epoxy-‐resin
composites
reinforced
with
knitted
glass-‐fiber
fabrics,
J.
Reinforced
Plastics
&
Comp.,
11:1239-‐1250.
4. Freitas
G.,
T.
Frusco,
T.
Campbell,
J.
Harris
and
S.
Rosenberg,
1996,
Z-‐Fiber
technology
and
products
for
enhancing
composite
design,
Proc.
of
the
83'
Meeting
of
the
AGARD
SMP
on
"BoltedKionded
Joints
in
Polymeric
Composites",
Sep.
2-‐3,
Florence,
Italy,
pp.17-‐1
-‐
17-‐8.
5. Freitas
G.,
C.
Magee,
P.
Dardzinski
and
T.
Fusco,
1994,
Fiber
insertion
process
for
improved
damage
tolerance
in
aircraft
laminates,
J.
Advanced
Mat.,
25:36-‐43.
6. Freitas
G.,
C.
Magee,
J.
Boyce
and
R.
Bott,
1991,
Service
tough
composite
structures
usingz-‐fiberprocess,Proc.9"
DoD/NASA/FAAConf.FibrousComp.,LakeTahoe,
Nevada,
Nov.
7. He
M.
and
B.N.
Cox,
1998,
Crack
bridging
by
through-‐thickness
reinforcement
in
delaminating
curved
structures,
Comp.
29:377.
8. Hohfeld
J.,
M.
Drews
and
R.
Kaldenhoff,
1994,
Roc.
of
the
31d
Int.
Conf.
of
Flow
Processes
in
Composite
Materials,
July
7-‐9,
120.
9. Huang
Z.M.,
Y.
Zhang
and
S
.
Ramakrishna,
2001,
Modeling
of
the
progressive
failure
behavior
of
multilayer
knitted
fabric-‐reinforced
composite
laminates,
Comp.
Sci.
&
Tech.,
61:2033-‐2046.
10. Leong
K.
H.,
P.
J.
Falzon,
M.
K.
Bannister
and
I.
Herszberg.
1998,
An
investigation
of
the
mechanical
performance
of
weft-‐knit
Milano-‐rib
glass/epoxy
composites,
Comp.
Sci.
&Tech.,
58:239-‐251.
11. Macander
A.B.,
R.M.
Crane
and
E.T
Camponeschi,
1986,
Fabrication
and
mechanical
properties
of
multidimensionally
(X-‐D)
braided
composite
materials,
Composite
Materials:
Testing
and
Design
(7
Conf.),
ASTM
STP
893,
American
Society
for
Testing
and
Materials,
Philadelphia,
422-‐443.
12. Mouritz
A.
P.,
C.
Baini
and
I.
Herszberg,
1999,
Mode
I
interlaminar
fracture
toughness
properties
of
advanced
textile
fibre-‐glass
composites,
Composites,
30:859-‐870.
13. Ramakrishna
S.,
N.
K.
Cuong
and
H.
Hamada,
1997,
Tensile
properties
of
plain
weft
knitted
glass
fiber
fabric
reinforced
epoxy
composites,
J.
Rein.
Plastics
&
Comp.,
16:946-‐966.
14. Steeves
A.,
Craig,
Norman
A.
Fleck,
2006,
In-‐plane
properties
of
composite
laminates
with
through-‐thickness
pin
reinforcement,
International
journal
of
solids
and
structures
43:
3197-‐3212.
15. Tong,
L.
Mouritz,
A.P.
and
Bannister,
M.K.
3D
Fibre
Reinforced
Polymer
Composites.
Elsevier
Science
Ltd.
Oxford,
UK.
2002.
16. Wu
W-‐L.,
M.
Kotaki,
A.
Fujita,
H.
Hamada,
M.
Inoda
and
Z-‐I.
Maekawa,
1993,
Mechanical-‐properties
of
warp-‐knitted,
fabric-‐reinforced
composites,
J.
Reinforced
Plastics
&
Comp.,
12:1096-‐1110.
78
Concluding
Remarks
8
In
this
work
the
different
3D
reinforcement
architectures
have
been
discussed
in
some
detail.
As
a
concluding
remark,
it
is
helpful
to
summarize
the
pros
and
cons
associated
with
them.
In
General,
3D
reinforcement
architectures
are
characterized
by
reduced
in-‐plane
mechanical
properties,
increased
delamination
toughness
and
increased
impact
damage
resistance.
The
reduction
of
the
in-‐plane
mechanical
properties
is
due
to
increased
fiber
damage
during
processing
(especially
for
stitching)
and
increased
crimp.
The
improved
out-‐of-‐
plane
mechanical
properties
are
attributable
to
the
z-‐directional
fibers.
At
this
time,
the
manufacturing
of
3D
composites
is
still
being
explored,
however
there
are
high
hopes
for
the
future
that
this
type
of
reinforcement
architecture
will
be
able
to
improve
manufacturing
efficiency
and
ease,
eventually
reducing
costs.
Some
of
the
defining
characteristics
of
specific
3D
architectures
should
also
be
acknowledged.
Knit
fabrics
are
of
particular
interest
due
to
their
exceptional
impact
resistance,
however
are
unsuitable
for
structural
application.
Stitched
and
z-‐pinned
are
among
the
simplest
and
most
inexpensive
methods
for
producing
3D
architectures
greatly
improving
mechanical
properties
in
joints
connections.
While
3D
braided
and
woven
allow
for
3D
shapes
to
be
produced
inherently
providing
a
means
for
enhanced
production
techniques.
3D
composites
and
their
reinforcement
architectures
need
to
be
further
studied
and
improvements
to
both
the
reinforcement
as
well
as
production
techniques
need
to
be
investigated
before
wide
spread
adoption
is
possible.
However,
3D
composites
are
a
very
promising
solution
to
many
of
the
problems
faced
by
2D
composites
and
can
be
expected
to
become
more
prevalent
in
the
future.
79