Modeling the Dimensional Variations

of Composites Using Effective

Coefficients of Thermal Expansion

CHENSONG DONG*
Curtin University of Technology, GPO Box 1987, Perth, WA 6845, Australia

ABSTRACT: With the increasing demands for energy efficiency and environmental
protection, composite materials have become an important alternative to traditional
materials. Composite materials offer many advantages over traditional materials
including low density, high strength, high stiffness to weight ratio, excellent durability,
and design flexibility. Despite all these advantages, composite materials have not
been as widely used as expected because of the complexity and cost of the manufactur-
ing process. One of the main causes is associated with poor dimensional control.
‘Spring-in’ is a common dimensional variation of composite angled parts. Due to
the layerwise structure of composites, traditionally the spring-in was calculated
numerically by defining each lamina as an element along the through-thickness
direction and assigning its material properties and orientation. This method is usually
too time-consuming for practical applications. In this article, the effective coefficients
of thermal expansion of symmetric laminates were derived. Using the effective CTE,
spring-in can be calculated by a simple mathematical formula. This approach was
validated by finite element analysis and experiments. The effective CTE are very useful
in 3-D finite element analysis. The effective CTE provide a feasible approach to the
3-D finite element analysis. Compared with the layerwise approach, the number of
elements is significantly reduced. This was illustrated by an example.

KEY WORDS: composites, dimensional variation, spring-in, effective coefficient of

thermal expansion.

INTRODUCTION

HE LAST FOUR decades have seen tremendous advancement in the science and tech-
T nology of fiber-reinforced composites. The low density, high strength, high stiffness to
weight ratio, excellent durability, and design flexibility of composites are the primary
reasons for their use in many structural components in aircraft, automotive, marine,
and other industries [1]. Despite all these advantages, composite materials have not

*E-mail: [email protected]
Figures 4, 5, 7 and 8 appear in color online: http://jcm.sagepub.com

Journal of COMPOSITE MATERIALS, Vol. 43, No. 22/2009 2639

0021-9983/09/22 2639—14 $10.00/0 DOI: 10.1177/0021998309345296
ß The Author(s), 2009. Reprints and permissions:
http://www.sagepub.co.uk/journalsPermissions.nav
Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015
2640 C. DONG

been as widely used as expected because of the complexity and cost of the manu-
facturing process. One of the main causes is associated with poor dimensional control.
For example, NASA is investigating the feasibility of SCRIMP to produce aircraft quality
heavily loaded primary structures. Additional technology development is required to
achieve dimensional control and acceptable fiber volume fraction for thick structural
elements [2].
Typical manufacturing processes are injecting resin into the fiber preform in the mold
and curing at some temperature (often higher than room temperature), such as RTM and
VARTM processes. During the curing process, the resin shrinks due to the chemical
reaction (crosslinking). After curing, the part is cooled down to the room temperature.
Due to the anisotropic material nature, dimensional variations are induced by the fiber-
resin CTE mismatch and the curing shrinkage of resin.
For asymmetric laminates, warpage such as saddle shape is induced. The residual
stresses and warpage can be predicted by the classical lamination theory [3,4]. For
symmetric laminates, no warpage is induced for flat parts. However, spring-in is induced
in angled parts, as shown in Figure 1. Hahn and Pagano [5] performed an elastic
analysis of the residual stresses in a thermoset matrix composite. Radford and
Diefendorf [6] developed a simple mathematical formula to predict the spring-in of
curved shaped parts, which was used by Huang and Yang [7] in their experimental
studies. Kollar [8] presented an approximate analysis of spring-in. Jain and Mai [9,10]
developed a mechanics-based model using modified shell theory. Yoon and Kim [11]
developed a computational method by applying the characterized properties to the
classical lamination theory. All of these studies considered simple structures such as
L-shaped. In order to predict the spring-in of more complex structures, numerical sim-
ulation tools of finite element method or finite difference method are often employed.
Wang et al. [12] conducted a finite element analysis of spring-in using ABAQUS. Ding et
al. [13] developed a 3-D finite element analysis procedure to predict spring-in resulting
from anisotropy for both thin and thick angled composite shells.
The mechanical behavior of composite materials is better represented by a viscoelas-
tic model. Clifford et al. [14] developed a 3-D thermo-viscoelastic model based on FEA
to predict the residual stress and dimensional stability of large complex composite
parts. Weitsman [15] developed a method for evaluating the residual stresses in
cross-ply graphite-epoxy laminates which incorporated linear viscoelasticity throughout
the cooldown stage. Wang et al. [16] studied residual stresses and warpage in multi-
directional woven glass/epoxy laminates. The in-plane stresses and curvatures were

Figure 1. Spring-in.

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

Modeling the Dimensional Variations of Composites 2641

calculated using a convolution integration form in the lamination theory expression. All
the aforementioned studies neglected the stresses developed before cooldown. A few
studies also addressed the residual stresses developed before cooldown, i.e. during the
curing process. White and Hahn [17,18] studied the residual stress development during
the curing of thin laminates numerically and experimentally. A 2-D finite difference
thermo-chemical model was developed. The cure-dependent material properties were
incorporated into a generalized plane-strain finite element model to predict the residual
stress development in a graphite-epoxy composite. Li et al. [19] used a plane-strain,
linearly elastic finite element model with temperature-dependent matrix properties to
analyze the evolution of residual stresses in graphite-PEEK composites during curing.
Wiersma et al. [20] developed a thermo-elastic model and extended it into a thermo-
viscoelastic model. A plane-strain finite element process model COMPRO was devel-
oped to simulate the spring-in and warpage in the autoclave process [21,22]. Zhu et al.
[23] developed a fully 3-D coupled thermo-chemo-viscoelastic finite element model to
simulate the heat transfer, curing, and residual stress development during the manu-
facturing cycle of thermoset composite parts. Wisnom et al. [24] presented a shear-lag
analysis for the spring-in of curved thermoset composites.
From the literature, it can be seen that the process-induced dimensional variations for
composites such as spring-in were studied either analytically or numerically. The use of
analytical models is limited to simple geometries such as L-shaped. When numerical
techniques such as finite element analysis are used, 3-D models are often necessary because
of the three dimensional anisotropy. Due to the layerwise structure, each lamina is repre-
sented with an element in the through-thickness direction. Thus, a large number of ele-
ments are required. Depending on the stacking sequence, the material properties and/or
orientation for each layer need to be defined. This is time-consuming and inconvenient
for practical applications. In this study, the effective coefficients of thermal expansion
were derived by combining all the laminae into a lumped layer. Using the effective
CTE, spring-in can be calculated by a simple mathematical formula. This approach was
validated by finite element analysis and experiments. The effective CTE provide a feasible
approach to the 3-D finite element analysis. Compared with the layerwise approach, the
number of elements is significantly reduced. The effectiveness of this approach was illu-
strated by an example.

DERIVATION OF EFFECTIVE COEFFICIENTS OF THERMAL EXPANSION

As shown in Figure 2, in order to describe the elastic properties of a thin lamina, two
right-handed coordinate systems, namely the 1-2-z system and the x-y-z system, are
defined. Both are in the plane of the lamina, and the z axis is normal to this plane. In
the 1-2-z system, axis 1 is along the fiber length and represents the longitudinal direction of
the lamina, and axis 2 is normal to the fiber length and represents the transverse direction
for the lamina. Together, they constitute the principal material directions in the plane of
the lamina. In the x-y-z system, x and y axes represent the loading directions. Fibers are
aligned at an angle  with the positive x direction [1].

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

2642 C. DONG

2
y

1
q

Figure 2. Definition of principal material axes and loading axes for a lamina.

For a thin lamina in the 1-2-z coordinate system, its stiffness matrix can be expressed by
C. The elements of C are given in [25]. The stiffness matrix in the x-y-z system is
C ¼ TCTT , where T is the transformation matrix and is given by
0 1
cos2  sin2  0 0 0 2 cos  sin 
B sin2  cos2  0 0 0 2 cos  sin  C
B C
B C
B 0 0 1 0 0 0 C
T¼B B C ð1Þ
0 0 0 cos   sin  0 C
B C
B C
@ 0 0 0 sin  cos  0 A
 cos  sin  cos  sin  0 0 0 cos2   sin2 
The CTE in the x-y-z system are related to the longitudinal CTE a11 and transverse CTE
a22 as

xx ¼ 11 cos2  þ 22 sin2 

yy ¼ 11 sin2  þ 22 cos2 
ð2Þ
xy ¼ ð11  22 Þ cos  sin 
zz ¼ 22
For a  lamina, the constitutive relationship when only thermal stresses are considered is
2 32 3 2 3
C11 C12 C13 0 0 C16 "xx  xx
T 0
6C 76 " 
T 7 6 7
6 12 C22 C23 0 0 C26 76 yy  yy 7 607
6 76 7 6 7
6 C13 C23 C33 0 0 C36 7 6 "  zz
T 7 6 0 7
6 76 zz 7¼6 7 ð3Þ
6 0 0 0 C44 C45 7
0 76 2"yz 7 607
6 6 7 6 7
6 76 7 6 7
4 0 0 0 C45 C55 0 54 2"xz 5 405
C16 C26 C36 0 0 C66 2"xy  2xy
T 0

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

Modeling the Dimensional Variations of Composites 2643

Equation (3) can be rewritten as

2 32 3 2 3
C11 C12 C13 0 0 C16 "xx C11 xx þ C12 yy þ C13 zz þ 2C16 xy
6 76 7 6 7
6 76 7 6 7
6 C12 C22 C23 0 0 C26 7 6 "yy 7 6 C12 xx þ C22 yy þ C23 zz þ 2C26 xy 7
6 76 7 6 7
6 76 7 6 7
6 76 7 6 7
6 C13 C23 C33 0 0 C36 76 "zz 7 6 C13 xx þ C23 yy þ C33 zz þ 2C36 xy 7
6 76 7 6 7
6 76 7¼6 7
T
6 76 2" 7 6 7
6 0 0 0 C44 C45 0 76 yz 7 6 0 7
6 76 7 6 7
6 76 7 6 7
6 0 0 0 C45 C55 7 6 7
0 76 2"xz 7 6 6 0 7
6 7
4 54 5 4 5
C16 C26 C36 0 0 C66 2"xy C16 xx þ C26 yy þ C36 zz þ 2C66 xy
ð4Þ
For a general laminate consisting of a number of laminae, all the laminae are bonded
together and have the same strains. Thus, the constitutive relationship when only thermal
stresses are considered is
2 e e e e 32 3 2 3
C11 C12 C13 0 0 C16 "xx Txx
6 e 76 7 6 7
6C e e
0 C26 7
e
6 12 C22 C23 0 76 "yy 7 6
7 6 Tyy 7
7
6 766 7 6 7
6 e e e e 7
6 C13 C23 C33 0 0 C36 76 6 "zz 77
6 Tzz 7
6 7
6 76 7¼6 7
T ð5Þ
6 76
0 76 2"yz 7 6 0 7
7 6
e e
6 0 0 0 C44 C45 7
6 76 7 6 7
6 e e 76 7 6 7
6 0 0 0 C C 0 7 4 2" xz 5 4 0 5
4 45 55 5
e e e e 2"xy Txy
C C
16 C 026 0 C 36 66
e
where Cij is the equivalent stiffness of the laminate [26]. Txx, Tyy, Tzz, and Txy are given as
n 
P         
Txx ¼ C11 i ðxx Þi þ C12 i yy i þ C13 i ðzz Þi þ2 C16 i xy i
i¼1
n 
P         
Tyy ¼ C12 i ðxx Þi þ C22 i yy i þ C23 i ðzz Þi þ2 C26 i xy i
i¼1
ð6Þ
n 
P         
Tzz ¼ C13 ð Þ þ C23 i yy i þ C33 i ðzz Þi þ2 C36 i xy i
i xx i
i¼1
n 
P         
Txy ¼ C16 i ðxx Þi þ C26 i yy i þ C36 i ðzz Þi þ2 C66 i xy i
i¼1
e e
For a symmetric laminate consisting of cross-plies and/or angle-plies, C16 ¼ C26 ¼
e e
C36 ¼ C45 ¼ 0, and Txy ¼ 0. The derivation is given in Appendix. Thus, Equation (5) is
reduced to
2 e e e 32 3 2 3
C11 C12 C13 "xx Txx
6 e e e 7 6 7 6 7
6C 76 7 6 7
4 12 C22 C23 54 "yy 5 ¼ 4 Tyy 5
T ð7Þ
e e e
C13 C23 C33 "zz Tzz

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

2644 C. DONG

The effective coefficients of thermal expansion are given as

 
 Txx Tyy Tzz 
 
 e e e 
C 
 12 C22 C23 
 e 
"xx  C13 C23 C33 
e e
e
xx ¼ ¼ 

T  Ce11 Ce12 Ce13 
 
 e e e 
C
 12 C22 C23 
 e 
C C
e
C 
e
13 23 33
 e e e 
 C11 C12 C13 
 
 
 Txx Tyy Tzz 

 e 
"yy C C23
e
C33 
e
eyy ¼ ¼  13 e  ð8Þ

T  Ce11 e
C12 C13 

 e e e 
C C22 C23 
 12
 e 
C C23
e
C 
e
13 33
 e e e 
 C11 C12 C13 
 
 e 
C e
C22
e
C23 
 12 
 
" T Tyy Tzz 
zz xx
ezz ¼ ¼ e e e 

T  C11 C12 C13 

 e e e 
C C22 C23 
 12
 e 
C C23
e
C33 
e
13

SPRING-IN MODEL BASED ON EFFECTIVE COEFFICIENTS

OF THERMAL EXPANSION

The starting materials for liquid composite processing are fiber preforms and liquid
resin. The processing can be divided into three stages [27]. In Stage I, the resin is fully
uncured and behaves as a viscous fluid. Stage II is the curing of the resin stage, the resin is
heated up to a temperature usually above the glass transition temperature, where a sig-
nificant increase in modulus and a reduction in specific volume begin to occur. Stage III
marks the end of the curing process, where the resin exhibits traditional viscoelastic behav-
ior at elevated temperatures and approaches elastic behavior at lower temperatures. After
demolding, the part is usually post-cured at an elevated temperature above Tg. During this
process, the stresses induced in the curing process can be significantly relaxed [28—30]. For
some resin systems, such as epoxy resins, the curing strain during conversion can be
neglected in the analysis of residual stress [4]. Thus, the residual stresses and dimensional
variations are assumed to be induced during the cooling down period. Since the mechan-
ical properties of resin are highly dependent on temperature, the mechanical properties
and CTE of the laminates are also dependent on temperature.

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

Modeling the Dimensional Variations of Composites 2645

αT > αI

αT

180- φ r ′ αI
r
180- φ ′

φ′
φ

Figure 3. Mechanism of spring-in.

As shown in Figure 3, the mechanism of spring-in is illustrated by using an L-shaped

part. The included angle is denoted as  and the radius of the curved section is denoted
as r. The original arc length is given by s ¼ rð180  Þ.
When the part is cooled down from the curing temperature Tc to the room temperature
T0, the radius and arc length are
 Z T0 
r0 ¼ r 1 þ T dT
Tc
 ZT  ð9Þ
0
s0 ¼ r 1 þ I dT ð180  Þ
Tc

where aI and aT are the in-plane and through-thickness CTE, respectively. The included
angle is given by
RT
0 s0 1 þ Tc0 I dT
 ¼ 180  0 ¼ 180  RT ð180  Þ ð10Þ
r 1 þ 0 T dT Tc

The spring-in is given by

R T0
TcðT  I ÞdT

 ¼ RT ð180  Þ ð11Þ
1 þ Tc0 T dT

Equation (11) shows that because of the mismatch of the in-plane CTE aI and the through-
thickness CTE aT, the included angle decreases after the composite part is cured, which is
commonly called ‘spring-in’.
It is well known that resin properties are highly temperature-dependent. Since temper-
ature mainly affects the modulus of matrix, and fibers are often much stiffer than matrix,
the effective CTE are not significantly affected by temperature. Thus, in this article, for
simplicity, the CTE were assumed to be independent of temperature, and Equation (11) is
simplified as
ðT  I ÞðT0  Tc Þ

 ¼ ð180  Þ ð12Þ
1 þ T ðT0  Tc Þ
For the calculation of spring-in of a cylindrical shell, aI ¼ axx and aT ¼ azz.
For the purpose of validation, L-shaped carbon/epoxy laminates were made by the resin
transfer molding process. The material is T300 carbon fibre/CYCOM934 epoxy prepreg.

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

2646 C. DONG

Various stacking sequences including [0]4, [30/30]s, [45/45]s, [60/60]s, [90]4, [0/90]2s,
and [90/0]2s were considered. The nominal thickness of each lamina was approximately
0.2 mm and the fiber volume fraction was determined to be 65%. The material properties
are as shown in Table 1. In the actual processing of composites, chemical shrinkage is also
induced. Thus, the modified coefficient of thermal expansion was introduced to include the
effect of chemical shrinkage, i.e.

0 ¼   ð1=
TÞ"c ð13Þ
First, the effective coefficients of thermal expansion calculated using Equation (8) were
validated by finite element analysis using a representative volume element. A commercial
FEA package MSC.Marc Mentat was employed in this study. Figure 4 shows the mesh for
[90/0]2s laminates. By applying symmetric boundary conditions only one quarter of the
unit cell was modeled in FEA. The boundary conditions used in FEA are as follows: along
the planes x, y, and z ¼ 0, the model was restricted to move in the x, y, and z directions,
respectively. Along the opposite planes, the periodic boundary conditions were applied,
which were achieved by defining links in MSC.Marc Mentat. The model underwent a unit
temperature drop. The displacements in the x, y, and z directions were obtained accord-
ingly. Figure 4 shows the displacement in the x direction, the units being mm. The CTE of
the composite in the direction i is given by

li
ii ¼ ð14Þ
li
T

Table 1. Material properties of T300/CYCOM 934 carbon epoxy unidirectional prepreg.

E11 (GPa) E22 (GPa) E33 (GPa) m12 m13 m23 G12 (GPa)

150.76 7.93 7.93 0.2525 0.2525 0.30 3.7

6  6  6 
a11 (10 / C) a22 (10 / C) a33 (10 / C) ec11 ec22 ec33
0.8 27.62 27.62 1.25  104 7.056  103 7.056  103

–2.262e–019
90° lamina
–4.447e–007
0° lamina
–8.895e–007

–1.334e–006

–1.779e–006

–2.224e–006

–2.668e–006

–3.113e–006

–3.558e–006

–4.003e–006

Z –4.447e–006 Z

X Y X Y

Figure 4. Prediction of effective CTE by FEA. Left: mesh; right: displacement. The contour shows the
displacement in the x direction in mm.

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

Modeling the Dimensional Variations of Composites 2647

The effective coefficients of thermal expansion of cross-ply laminates calculated from FEA
and Equation (8) are as shown in Table 2. They are in excellent agreement.
The spring-in of cross-ply laminates was calculated by the current model and FEA,
respectively. 3-D FEA was conducted, as shown in Figure 5. Each lamina was represented
by an element in the through-thickness direction. Experimentally, the spring-in was also
measured from the sample parts fabricated. For each stacking sequence, two parts were
made and the spring-in was measured at various locations. The average values are pre-
sented. The spring-in values calculated by FEA and Equation (12), and from experiments
are as shown in Table 3. Again, excellent agreement is found.

Table 2. Effective coefficients of thermal expansion of cross-ply

laminates calculated by current model and FEA.

axx (aI) (106/8C) azz (aT) (106/8C)

Stacking sequence Current model FEA Current model FEA

[0/90]2s 4.45 4.45 92.21 92.20

[90/0]2s 4.45 4.45 92.21 92.20

Z X

Figure 5. 3-D mesh for L-shaped carbon/epoxy laminates.

Table 3. Spring-in of cross-ply laminates calculated by current

model and FEA, and from experiments.

Spring-in (8)

Stacking sequence Current model FEA Experiment

[0/90]2s 1.250 1.220 1.245

[90/0]2s 1.250 1.238 1.240

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

2648 C. DONG

For other stacking sequences, the spring-in was calculated by the current model and
FEA, respectively, and measured from the sample parts fabricated. The spring-in values
calculated by the current model, FEA, and from experiments are as shown in Figure 6.
Overall, the spring-in from the current model is in excellent agreement with that from
FEA. They are both in good agreement with experimental data.

3-DIMENSIONAL APPLICATIONS

An attractive advantage of using effective coefficients of thermal expansion is: in 3-D

finite element analysis, the number of elements is significantly reduced and the material
properties definition process is significantly simplified. Thus, it provides a practical
approach for the design of composite structures. An example is shown in Figure 7, con-
structed by introducing curvature to the third direction of an L-shaped part. The materials
were assumed to be those as shown in Table 1. The stacking sequence was [0/90]3s. One
quarter of the part was modeled in FEA. 27,864 elements and 32,110 nodes were used in
the layerwise FEA and 1044 and 1710 nodes were used in the FEA using the equivalent

2
Current model
FEA
1.5 Experiment
Δ φ (°)

0.5

0
0 15 30 45 60 75 90
θ (°)

Figure 6. Spring-in calculated by the current model, FEA, and from experiments.

Z X

Figure 7. A 3-D surface.

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

Modeling the Dimensional Variations of Composites 2649

4.626e–001 4.310e–001

4.164e–001 3.879e–001

3.701e–001 3.448e–001

3.239e–001 3.017e–001

2.776e–001 2.586e–001

2.313e–001 2.155e–001

1.851e–001 1.724e–001

1.388e–001 1.293e–001

9.253e–002 8.619e–002

4.626e–002 4.310e–002

4.919e–014 Y 1.812e–013 Y

z X z X

Figure 8. Dimensional variations calculated from FEA. The contour shows the total displacement in mm.
Left: by layerwise approach; right: using equivalent material properties.

material properties, respectively. The dimensional variations calculated from FEA are as
shown in Figure 8. The total displacements from the layerwise approach and using the
equivalent material properties are 0.46 mm and 0.43 mm, respectively. The relative error is
6.5% and thus the accuracy is satisfactory.

CONCLUSIONS

In this article, the effective coefficients of thermal expansion of symmetric laminates

were derived. Using the effective CTE, spring-in can be calculated by a simple mathemat-
ical formula. This approach was validated by finite element analysis and experiments.
Good agreement is found. The effective CTE are very useful in 3-D finite element analysis.
The effective CTE provide a feasible approach to the 3-D finite element analysis.
Compared with the layerwise approach, the number of elements is significantly reduced.
The effectiveness was illustrated by an example.

ACKNOWLEDGMENT

The author is supported by the Curtin Research Fellowship.

APPENDIX

The elements of the stiffness matrix have the following relationship.

C11 ðÞ ¼ C11 ðÞ
C12 ðÞ ¼ C12 ðÞ
C13 ðÞ ¼ C13 ðÞ
C16 ðÞ ¼ C16 ðÞ
C22 ðÞ ¼ C22 ðÞ

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

2650 C. DONG

C23 ðÞ ¼ C23 ðÞ

C26 ðÞ ¼ C26 ðÞ

C33 ðÞ ¼ C33 ðÞ

C36 ðÞ ¼ C36 ðÞ

C44 ðÞ ¼ C44 ðÞ

C45 ðÞ ¼ C45 ðÞ

C55 ðÞ ¼ C55 ðÞ

C66 ðÞ ¼ C66 ðÞ

The CTE have the following relationship.

xx ðÞ ¼ xx ðÞ

yy ðÞ ¼ yy ðÞ

zz ðÞ ¼ zz ðÞ

xy ðÞ ¼ xy ðÞ
Thus, for a symmetric laminate consisting of cross-plies and/or angle-plies,
e e e e
C16 ¼ C26 ¼ C36 ¼ C45 ¼ 0
Txy ¼ 0

REFERENCES

1. Mallick, P.K. (1993). Fiber-Reinforced Composites: Materials, Manufacturing, and Design, Marcel
Dekker, New York.
2. Shuart, M.J., Johnston, N.J., Dexter, H.B., Marchello, J.M. and Grenoble, R.W. (1998).
Automated Fabrication Technologies for High Performance Polymer Composites, NASA
Langley Technical Report Server.
3. Cowley, K.D. and Beaumont, P.W.R. (1997). The Measurement and Prediction of Residual
Stresses in Carbon-fibre/Polymer Composites, Composites Science and Technology, 57(11):
1445—1455.
4. Theriault, R.P. and Osswald, T.A. (1999). Processing Induced Residual Stress in Asymmetric
Laminate Panels, Polymer Composites, 20(3): 493—509.
5. Hahn, H.T. and Pagano, N.J. (1975). Curing Stress in Composite Laminates, Journal of
Composite Materials, 9: 91—105.
6. Radford, D.W. and Diefendorf, R.J. (1993). Shape Instabilities in Composite Resulting from
Laminate Anistropy, Journal of Reinforced Plastics and Composites, 12(1): 58—75.
7. Huang, C.K. and Yang, S.Y. (1997). Warping in Advanced Composite Tools with Varying
Angles and Radii, Composites Part A, 28(9—10): 891—893.
8. Kollar, L.P. (1994). Approximate Analysis of the Temperature Induced Stresses and
Deformations of Composite Shells, Journal of Composite Materials, 28(5): 392—414.

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

Modeling the Dimensional Variations of Composites 2651

9. Jain, L.K., Lutton, B.G., Mai, Y.W. and Paton, R. (1997). Stresses and Deformations Induced
During Manufacturing. Part II: A Study of the Spring-in Phenomenon, Journal of Composite
Materials, 31(7): 696—719.
10. Jain, L.K. and Mai, Y.W. (1997). Stresses and Deformations Induced During Manufacturing.
Part I: Theoretical Analysis of Composite Cylinders and Shells, Journal of Composite Materials,
31(7): 673—695.
11. Yoon, K.J. and Kim, J.-S. (2001). Effect of Thermal Deformation and Chemical Shrinkage on
the Process Induced Distortion of Carbon/Epoxy Curved Laminates, Journal of Composite
Materials, 35(3): 253—263.
12. Wang, J., Kelly, D. and Hillier, W. (2000). Finite Element Analysis of Temperature Induced
Stresses and Deformations of Polymer Composite Components, Journal of Composite Materials,
34(17): 1456—1471.
13. Ding, Y., Chiu, W.K. and Liu, X.L. (2001). Anisotropy Related ‘Spring-in’ of Angled
Composite Shells, Polymers & Polymer Composites, 9(6): 393—401.
14. Clifford, S., Jansson, N., Yu, W., Michaud, V. and Manson, J.-A. (2006). Thermoviscoelastic
Anisotropic Analysis of Process Induced Residual Stresses and Dimensional Stability in Real
Polymer Matrix Composite Components, Composites Part A, 37(4): 538—545.
15. Weitsman, Y. (1979). Residual Thermal Stresses due to Cool-down of Epoxy-Resin Composites,
Journal of Applied Mechanics, 46(3): 563—567.
16. Wang, T.M., Daniel, I.M. and Gotro, J.T. (1992). Thermoviscoelastic Analysis of
Residual Stresses and Warpage in Composite Laminates, Journal of Composite Materials,
26(6): 883—899.
17. White, S.R. and Hahn, H.T. (1992). Process Modeling of Composite Materials: Residual Stress
Development During Cure. Part I: Model Formulation, Journal of Composite Materials, 26(16):
2402—2422.
18. White, S.R. and Hahn, H.T. (1992). Process Modeling of Composite Materials: Residual Stress
Development During Cure. Part II: Experimental Validation, Journal of Composite Materials,
26(16): 2423—2454.
19. Li, M.C., Wu, J.J., Loos, A.C. and Morton, J. (1997). A Plane-strain Finite Element Model for
Process-induced Residual Stresses in a Graphite/PEEK Composite, Journal of Composite
Materials, 31(3): 212—243.
20. Wiersma, H.W., Peeters, L.J.B. and Akkerman, R. (1998). Prediction of Springforward in
Continuous-fiber/Polymer L-shaped Parts, Composites Part A, 29(11): 1333—1342.
21. Fernlund, G., Osooly, A., Poursartip, A., Vaziri, R., Courdji, R., Nelson, K., George, P.,
Hendrickson, L. and Griffith, J. (2003). Finite Element Based Prediction of Process-induced
Deformation of Autoclaved Composite Structures Using 2D Process Analysis and 3D Structural
Analysis, Composite Structures, 62(2): 223—234.
22. Johnston, A., Vaziri, R. and Poursartip, A. (2001). A Plane Strain Model for Process-induced
Deformation of Laminated Composite Structures, Journal of Composite Materials, 35(16):
1435—1469.
23. Zhu, Q., Geubelle, P.H., Li, M. and Tucker III, C.L. (2001). Dimensional Accuracy of
Thermoset Composites: Simulation of Process-induced Residual Stresses, Journal of
Composite Materials, 35(2): 2171—2205.
24. Wisnom, M.R., Potter, K.D. and Ersoy, N. (2007). Shear-lag Analysis of the Effect of Thickness
on Spring-in of Curved Composites, Journal of Composite Materials, 41(11): 1311—1324.
25. Chou, T.-W. (1992). Microstructural Design of Fiber Composites, Cambridge University Press,
Cambridge, UK.
26. Chen, H.J. and Tsai, S.W. (1996). Three-dimensional Effective Moduli of Symmetric Laminates,
Journal of Composite Materials, 30(8): 906—917.
27. Bogetti, T.A. and Gillespie Jr, J.W. (1992). Process-induced Stress and Deformation in Thick-
section Thermoset Composite Laminates, Journal of Composite Materials, 26(5): 626—660.

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015

2652 C. DONG

28. Genidy, M.S., Madhukar, M.S. and Russell, J.D. (2000). A New Method to Reduce
Cure-induced Stresses in Thermoset Polymer Composites Part II: Closed Loop Feedback
Control System, Journal of Composite Materials, 34(22): 1905—1925.
29. Madhukar, M.S., Genidy, M.S. and Russell, J.D. (2000). A New Method to Reduce
Cure-induced Stresses in Thermoset Polymer Composites Part I: Test Method, Journal of
Composite Materials, 34(22): 1882—1904.
30. Russell, J.D., Madhukar, M.S., Genidy, M.S. and Lee, A.Y. (2000). A New Method to Reduce
Cure-induced Stresses in Thermoset Polymer Composites Part III: Correlating Stress History to
Viscosity Degree of Cure and Cure Shrinkage, Journal of Composite Materials, 34(22):
1926—1947.

Downloaded from jcm.sagepub.com at R.M.I.T. Libraries Bundoora on March 14, 2015