J Thermal Stresses 2007 Berger Etal

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/237613825
Numerical and Analytical Approaches for Calculating the Effective

ThermoMechanical Properties of Three-Phase Composites
Article in Journal of Thermal Stresses · July 2007

DOI: 10.1080/01495730701415665
CITATIONS READS
21 356
7 authors, including:
Srihari Kurukuri Ulrich Gabbert

National Research Council Canada, London, Ontario, Canada Otto-von-Guericke-Universität Magdeburg
35 PUBLICATIONS 389 CITATIONS 344 PUBLICATIONS 4,061 CITATIONS
SEE PROFILE SEE PROFILE
Julián Bravo-Castillero
Universidad Nacional Autónoma de México
206 PUBLICATIONS 2,807 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Biomechanical mdel for FESS View project
Modeling of Damage Detection on Fiber Reinforced Composite Polymer plates View project
All content following this page was uploaded by Ulrich Gabbert on 28 October 2020.
The user has requested enhancement of the downloaded file.

This article was downloaded by:[Otto-von-guericke-universitaet]
On: 10 September 2007
Access Details: [subscription number 770633723]
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Thermal Stresses

Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713723680
Numerical and Analytical Approaches for Calculating the
Effective Thermo-Mechanical Properties of Three-Phase
Composites
Online Publication Date: 01 August 2007
To cite this Article: Berger, H., Kurukuri, S., Kari, S., Gabbert, U.,
Rodriguez-Ramos, R., Bravo-Castillero, J. and Guinovart-Diaz, R. (2007) 'Numerical
and Analytical Approaches for Calculating the Effective Thermo-Mechanical
Properties of Three-Phase Composites', Journal of Thermal Stresses, 30:8, 801 -
817
To link to this article: DOI: 10.1080/01495730701415665
URL: http://dx.doi.org/10.1080/01495730701415665
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction,
re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly
forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents will be
complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be
independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or
arising out of the use of this material.
© Taylor and Francis 2007
Journal of Thermal Stresses, 30: 801–817, 2007
Downloaded By: [Otto-von-guericke-universitaet] At: 15:51 10 September 2007
Copyright © Taylor & Francis Group, LLC

ISSN: 0149-5739 print/1521-074X online
DOI: 10.1080/01495730701415665
NUMERICAL AND ANALYTICAL APPROACHES FOR

CALCULATING THE EFFECTIVE THERMO-MECHANICAL
PROPERTIES OF THREE-PHASE COMPOSITES
H. Berger, S. Kurukuri, S. Kari, and U. Gabbert

Otto-von-Guericke-University of Magdeburg, Faculty of Mechanical
Engineering, Institute of Mechanics, Magdeburg, Germany
R. Rodriguez-Ramos and J. Bravo-Castillero

Facultad de Matemática y Computación, Universidad de La Habana,
Vedado, Habana 4, Cuba
R. Guinovart-Diaz
Instituto Tecnológico y de Estudios Superiores de Monterrey, Campus Estado
de México, Carretera Lago de Guadalupe, Estado de México, México
The aim of this article is to envisage the effective thermo-mechanical properties

of three phase composites made up of coated unidirectional cylindrical fibers using
homogenization techniques. The main focus is on square arrangements of cylindrical
fibers in composite. The numerical approach is based on the micro-mechanical unit
cell modeling technique using finite element method (FEM) with appropriate boundary
conditions and it allows the extension to composites with arbitrary geometrical
inclusion configurations, providing a powerful tool for fast calculation of their effective
properties. The results obtained from the numerical technique are compared with those
obtained by means of the analytical asymptotic homogenization method (AHM) for
different fiber volume fractions. In order to analyze the interphase effect, the effective
properties are compared with the results obtained from some theoretical approaches
reported in the literature.
Keywords: Effective thermo-mechanical properties; FEM; Homogenization; Interphase; Three-phase

composite.
INTRODUCTION
Thermo-elastic composites constitute an important class of materials with
a wide variety of applications ranging from aerospace structures and electronic
printed circuit boards to recreational and commercial equipment. Some of the
most important and useful properties of these composites are lightweight, high
Received 2 November 2006; accepted 1 March 2007.

This work has been supported by DFG Graduiertenkolleg 828 “Micro-Macro Interactions in
Structured Media and Particle Systems” at the University of Magdeburg. Thanks also to the project
PNCB IBMFQ 09-2004 and FENOMEC, PAPIIT, DGAPA, UNAM-Mexico, under grant IN101705.
These supports are greatly acknowledged.
Address correspondence to H. Berger, University of Magdeburg, Institute of Mechanics,
Universitätsplatz 2, Magdeburg, D-39106, Germany. E-mail: berger@mb.uni-magdeburg.de
801
802 H. BERGER ET AL.
strength and stiffness, excellent frictional properties, good resistance to fatigue

and retention of these properties at high temperatures. The combination of these
properties has placed thermo-elastic composites at first rank among materials used
for heat shields, leading edges, re-entry tips, rocket nozzles and brakes for military
and advanced civilian aircrafts. The effective thermo-mechanical properties of the
composite depend upon properties of the constituents and the fiber volume fraction.
Many authors have developed techniques to study the elastic behavior of fibrous
composites. They take into account the existence of an intermediate layer between
the matrix and the fiber. These thin layers are called interphases or interfacial zones
between fiber and matrix. The effective utilization of the fiber reinforced composites
depends on efficient load transfer from the matrix to fibers through these interphases.
These interphases are formed due to, for example, chemical reaction between the
matrix and fiber materials or the use of protective coatings on the fiber during
manufacturing. Although small in thickness, interphases can significantly affect the
overall mechanical properties of the fiber-reinforced composites. It is the weakest link
in the load path, and consequently most failures in fiber reinforced composites, such
as debonding, fiber pullout, and matrix cracking, occur in or near this region. Thus,
it is crucial to fully understand the mechanism and effects of the interphases on the
overall material properties of fiber reinforced composites.
Without seeking to make a deep revision of the results reached in the past two
decades, we will show a summary of some of the many reviewed articles, some of
which possess results that can be used in comparisons with the predictions obtained
in our model.
Several authors have developed techniques to study the elastic behavior of
multiphase fibrous composites. Hill [1] and Hashin’s bounds [2, 3] provide widely
established benchmarks for validating the predicted effective properties of multiphase
fiber reinforced composites with arbitrary phase geometry. It was shown by Walpole
[4] that the thin coating on an inclusion has a pronounced effect on the fields just
outside the inclusion. The stress field in a coated continuous fiber composite subjected
to thermo-mechanical loading has been considered by Mikata et al. [5]. Theocaris et al.
[6] described a model to predict the influence of the interphase in fibrous composites.
Their model was based on a correct version of Kerner’s model, which is conveniently
modified by introducing an interphase layer between the fiber and the matrix.
Two models to approximate the thermo-elastic response of a composite
body reinforced by coated fibers oriented in various directions were developed in
Pagano et al. [7, 8]. The fundamental representative volume element (RVE) is a
three-phase concentric circular cylinder under prescribed displacement components
and surface tractions. The analysis leads to estimating how a coating applied
to the fiber influences the effective thermo-elastic properties of fiber reinforced
composites. Hashin’s [9] imperfect interface conditions are defined in terms of
linear relations between interface tractions and displacement jumps. All the thermo-
elastic properties of unidirectional fiber composites with such interface conditions
are evaluated on the basis of a generalized self-consistent scheme model. The effect
of interphase on the transverse properties of elasto-plastic composites is studied
employing the finite element method in Yeh [10].
In Sutcu [11] a simple recursive algorithm is presented that considers only two
concentric cylinders at a time in order to calculate five effective elastic constants and
two linear thermal expansion coefficients for a uni-axially aligned composite that
contains an arbitrary number of coatings on its fibers.
THERMO-MECHANICAL PROPERTIES OF THREE-PHASE COMPOSITES 803
In Dasgupta and Bhandarkar [12] a method to obtain the transversely

isotropic effective thermo-mechanical properties of unidirectional composites
reinforced with coated cylindrical fibers is discussed. In this work, the method
proposed in Benveniste et al. [13] is extended to composites with multiply-coated
cylindrical reinforcements, and a generalized self-consistent scheme is proposed for
obtaining the transverse shear properties. Chu and Rokhlin [14] reported a method
for the inverse determination of effective elastic moduli of mesophases using a
multiphase generalized self-consistent model.
Lagache et al. [15] determined numerically the effect of a mesophase using
a finite element formulation in order to solve the local problems derived from
the homogenization method. The effect of an interphase on the behavior of a
glass/epoxy composite based on the Theocaris interphase model and taking a
hexagonal cell for the three-phase composite was discussed in Chouchaoui and
Benzeggagh [16]. Agbossou and Pastor [17] developed a thermal self-consistent
model for n-layered fiber composites subjected to a uniform temperature. They
propose analytical and semi-analytical models for determining the thermal behavior
of composites. Liu et al. [18] have modeled interphases in fiber reinforced composites
under transverse loading using the boundary element method. The interphases are
regarded as elastic layers between the fibers and the matrix.
Hashin [19] reported that the imperfect interphase conditions are equivalent
to the effect of a thin elastic interphase, and high accuracy of the method is proved
by comparison of solutions of several problems in terms of the explicit presence
of the interphase as a third phase. For two-phase composites, Rodriguez-Ramos
et al. [20] and Guinovart-Diaz et al. [21], using the asymptotic homogenization
method (AHM), obtained analytical formulae for the global elastic constants
of a binary fiber composite with perfect interfaces (which means that tractions
and displacements are continuous across the interface) in a periodic structure,
particularly with square and hexagonal distributions of fibers in the matrix.
There are some studies using the finite element approach for the micro-
mechanical unit cell modeling to investigate more complex structures, for instance,
Golanski et al. [22], Terada and Kikuchi [23], Berger et al. [24], Kari et al. [25] and Xia
et al. [26]. However, in many cases of interest the perfect interphase is not an adequate
model, and it is necessary to include in unit cell modeling one or more interphases
separating the reinforcement inclusion phase from the host matrix phase.
In the present study, the previously stated experiences of the authors for two-
phase composites are extended, considering now a third phase between the fiber
and matrix. In this work, effective thermo-mechanical properties of three-phase
composites for different volume fractions of reinforced fiber material are predicted
using micro-mechanical unit cell modeling. The unit cell is analyzed using the finite
element method and appropriate periodic boundary conditions for different loading
conditions. In this model the perfect adhesion between the phases and the matrix is
considered.
BASIC EQUATIONS: HOMOGENIZATION METHOD IN THERMOELASTICITY

Many industrial and engineering materials as well as the majority of natural
materials are inhomogeneous, i.e., they consist of dissimilar constituents (or
“phases”) that are distinguishable at some (small) length scale. Each constituent
shows different material properties and/or material orientations and may itself be
inhomogeneous at some smaller length scale(s).
To determine the macroscopic overall characteristics of heterogeneous media
is an essential problem in many engineering applications. From the time and cost
view points, performing straightforward experimental measurements on a number
of material samples, for various phase properties, volume fractions and loading
histories is a hardly feasible task. On the other hand, due to usually enormous
difference in length scales, it is impossible, for instance, to generate a finite element
mesh that accurately represents the microstructure and also allows the numerical
solution of the macroscopic structural component within a reasonable amount
of time on today’s computational systems. To overcome this problem several
homogenization techniques have been created to obtain a suitable constitutive
model to be inserted at the macroscopic level, i.e., homogenization is a mechanics
based modeling scheme that transforms a body of a heterogeneous material into a
constitutively equivalent body of a homogeneous continuum.
In principle, the transformation model should be built on the basis of
the composite microstructure, along with the relevant physical laws. A set of
effective properties is obtained for the equivalent homogeneous continuum. When
a composite specimen is under external load, micro stresses and strains are induced
throughout the specimen. Ideally, the micro fields should be computed exactly,
given the specimen and its fiber/matrix microstructure. Through the concept of
homogenization, the composite specimen is regarded as a body of an effective
homogeneous material, whose mechanical behavior is described by a definitive
constitutive law.
Let us consider a stationary thermo-mechanical problem in a heterogeneous
medium. Let the position of a typical point of the body be denoted by three
coordinates x1 x2 x3 of a Cartesian system of axes and let the periodic cell of
the structure, Y , be defined by the inequalities, −hi /2 < yi < hi /2, where yi = xi /
denotes the local fields, and 0 < 1, hi ∼ 1. The linear equation of equilibrium
in quasi-static and stationary problem is given by
ijj + Xi = 0 (1)
where the subscripts assume the values 1, 2, and 3, the comma denotes partial
differentiation, and the summation convention is applied. Xi is the body force. The
stress and strain tensors are related to the temperature change T by the
constitutive relations
ij x = Cijkl x/kl x − x/Tx

ij = Cijkl x/ kl x/ (2)
Eq. (2) is known as the Duhamel–Neumann law. The constitutive laws contain the
components of the elasticity four-order tensor Cijkl x/ and the components of the
thermal-stress second-order tensor ij x/, and ij x/ are the thermal expansion
coefficients. These coefficients are functions of the fast variable yi = xi /, because
they express the rapid change of properties of the composite or the high level of the
heterogeneity.
The strain tensor is defined by the Cauchy relations
1
kl x = ukl + ulk (3)
2
where uk are the components of displacement.
Eqs. (1) and (2) perform a system for solving the thermo-mechanical problem
in which the influence of the temperature is induced by external body forces
arising from temperature difference T . To this system we must adjoin appropriate
boundary and initial conditions, which will be omitted here for simplicity. All
coefficients Cijkl ij and kl are considered to be piecewise-smooth Y -periodic
functions of the fast coordinates yi . The objective of the homogenization approach
is to obtain a closed system of equations with constant coefficients, equivalent to
the given system (1). These new coefficients are considered to represent the global
properties of the composite.
To solve this classical problem a two-scale homogenization technique has
been applied. Rodriguez-Ramos et al. [20] and Guinovart-Diaz et al. [21], derived
analytical expressions for the macroscopic effective properties of two-phase fiber
reinforced composite with perfect interphases using the asymptotic homogenization
method (AHM), which means that tractions and displacements are continuous
across the interface in a periodic structure. For the prediction of effective elastic
coefficients of three-phase composites in Afonso et al. [27] by means of closed form
of analytical solutions based on a combination of the modified shear-lag model
and the method of cells are considered. It does not require detailed knowledge
of the microstructure. Guinovart-Diaz et al. [28] derived a recursive asymptotic
homogenization scheme to predict the effective elastic properties of multi-phase
composite materials.
Recently, analytical expressions for effective material coefficients have been
derived in Guinovart-Diaz et al. [29] for three-phase, thermo-elastic composites.
A simple closed form of effective properties for a three-phase unidirectional
transversely isotropic composite is presented. By using homogenization schemes for
periodic media, the local problems are solved and effective thermo-elastic properties
moduli are determined. These expressions for the effective thermo-elastic constants
can be found in Guinovart-Diaz et al. [29], and are not given here for sake of
brevity. The present work is closely connected with these formulations. But here the
focus is set on the development of an equivalent numerical homogenization tool that
also can be used for more complex composites to exceed the limits of the analytical
methods.
NUMERICAL HOMOGENIZATION OF RVE USING

FINITE ELEMENT ANALYSIS
Representative Volume Element
For many composites, the macrostructure can be seen as a periodic array
of repeated unit cells. Also for particulate reinforced composites, a repeated unit
cell can still be constructed after assuming a uniform distribution and the same
geometry for the reinforcing phase. Therefore, in most micro-mechanical analyses
Figure 1 (a) Schematic diagram of a periodic unit cell; (b) Notation of surfaces.
the repeated unit cell is chosen as the RVE for the composites as shown in
Figure 1. In the present work, micro-mechanical analysis method is applied to
periodic RVE.
The micro-mechanical method provides the effective thermo-mechanical
properties of three phase composites from the known properties of their constituents
(fiber, matrix and interphase) for different volume fractions using periodic
representative volume element (RVE) or a unit cell model. The purpose of present
article is to establish unit cells for three-phase composites and to demonstrate the
usage of proper boundary conditions to account for the periodic nature of the stress
and strain fields in the composite. The finite element method has been extensively
used in the literature to analyze a periodic unit cell, to determine the effective
properties of fiber reinforced composites. Here the unit cells are discretized and
analyzed using the finite element method to predict the effective thermo-mechanical
properties of unidirectional periodic coated (interphase) cylindrical fiber composites
for different volume fractions.
This constitutive law can be determined based on the detailed fields in the
selected unit cell through an “averaging” procedure. Specifically, if the exact micro
fields ij and ij in the unit cell are known under the applied load, the averaged
stresses and strains over the unit cell are given by
1
¯ ij = dV
V V ij
(4)
1
¯ ij = dV
V V ij
where V is the volume of the RVE. The averages are then treated as the effective
stress and strain fields in the homogenized RVE. The relations between ij and ij
determine the “effective” constitutive law.
The constitutive relation presented in Eq. (2) can also be expressed in matrix
form as shown in Eq. (5).
    
11 C11 C12 C13 C14 C15 C16 11 − 11 T
22  C21 C22 C23 C24 C25 C26  22 − 22 T 
    
33  C31 C32 C33 C34 C35 C36  33 − 33 T 
 =   (5)
23  C41 C42 C43 C44 C45 C46  23 − 23 T 
    
13  C51 C52 C53 C54 C55 C56  31 − 31 T 
12 C61 C62 C63 C64 C65 C66 12 − 12 T
In the preceding, Cijkl is the effective stiffness of the homogenized composite. The
number of independent constants in Cijkl is determined by the assumed symmetry.
For these particular composites, transversely isotropic symmetry is assumed. For
a transversely isotropic material, thermal expansion cannot induce shear, but the
expansion in the three directions need not be equal. Consequently the equation has
the form
    
11 C11 C12 C13 0 0 0 11 − 11 T
22   C11 C13 0 0 0   
    22 − 11 T 
33   C33 0 0 0  33 − 33 T 
 
 =  (6)
23   C44 0 0   23 
    
13   symm
C44 0   31 
12 C66 12
For instance, the first column of stiffness matrix is obtained when we impose
the boundary conditions in such a way that the macroscopic strain 11 is not
equal to zero and all other strains and T are zero. Once all the independent
constants of stiffness matrix are obtained, coefficient of thermal expansion (CTE)
kl can be obtained by solving the constitutive equation with all kl equal to zero
and nonzero T (by providing temperature difference to the unit cell as a body
force and fix all the faces in all directions). Due to temperature difference there
may be stresses developed inside the RVE. From the above averaging relations
stresses ¯ ij are calculated. Once we know the ¯ ij and stiffness matrix C and
temperature difference T , we can solve for the effective coefficient of thermal
expansion kl .
For transversely isotropic symmetry, the engineering constants are related to

components of the stiffness tensor by
2
Ea = C11 C33 + 2C13
2
C12 − 2C11 C13
2
− C33 C12
2
/ C11 C33 − C13
2
2 2
Et = C11 C33 + 2C13
2
C12 − 2C11 C13
2
− C33 C12
2
/ C11 − C12
2

a = C12 C33 − C13
2
/ C11 C33 − C13
2
(7)

t = C11 C13 − C12 C13 / C11 C33 − C132
G23 = G13 = C44 G12 = C66
Boundary Conditions
One of the most important issue in the finite element analysis of periodic
a RVE is an appropriate application of the periodic boundary conditions. The
displacement field for the periodic structure can be expressed as
ui x1 x2 x3 = 0ij xj + u∗i x1 x2 x3 (8)
In the above, 0ij is the global (average) strain tensor of the periodic structure and
the first term on the right side represents a linear distributed displacement field. The
second term on the right side u∗i x1 x2 x3 is a periodic function from one unit cell
to another. It represents a modification to the linear displacement field due to the
heterogeneous structure of the composites.
Since the periodic array of the repeated unit cells represents a continuous
physical body, two continuities must be satisfied at the boundaries of the
neighboring unit cells. One is that the displacements must be continuous, i.e.,
the adjacent unit cells cannot be separated or penetrated into each other at the
boundaries after the deformation. The second condition implies that the traction
distributions at the opposite parallel boundaries of a unit cell must be the same.
In this manner, the individual unit cell can thus be assembled as a physically
continuous body followed by Xia et al. [30].
Obviously, the assumption of displacement field in the form of Eq. (8) meets
the first of the above requirements. Unfortunately, it cannot be directly applied
to the boundaries since the periodic part, u∗i x1 x2 x3 is generally unknown. For
any unit cell, its boundary surfaces must always appear in parallel pairs, the
displacements on a pair of parallel opposite boundary surfaces can be written as
∗
i = ij xj + ui
0 k+
uk+ (9)
∗
i = ij xj + ui
0 k−
uk− (10)
where indices k+ and k− identify the kth pair of two opposite parallel boundary
surfaces of a repeated unit cell. Note that u∗i x1 x2 x3 is the same at the two
parallel boundaries (periodicity), therefore, the difference between the above two
equations is
k+
i − ui = ik xj − xj = 0ij xjk
0
uk+ k− k−
(11)
Since xjk are constants for each pair of the parallel boundary surfaces, with
specified 0ij , the right side becomes constant and such equations can be easily applied
in the finite element analysis as nodal displacement constraint equations. Eq. (11) is
a special type of displacement boundary conditions. Instead of giving known values
of boundary displacements, it specifies the displacement differences between two
opposite boundaries. Obviously, it becomes easier to adopt this form in a finite element
procedure, instead of applying Eq. (8) directly as the boundary conditions.
To apply the constraint equations (11) for instance in FEM, it is required
to produce the same meshing at each two paired boundary surfaces. Then each
constraint equation in (11) contains only two displacement components of the
paired nodes. The number of the constraint equations is usually quite large, certain
preprocessing program can be used to produce the data depending on the individual
FEM code used.
FINITE ELEMENT SIMULATIONS

All the finite element calculations were done with commercial finite element
program ANSYS within the framework of the small displacements theory, and the
materials are assumed to behave as linear elastic and isotropic solids. The finite
element mesh is created using three dimensional multi-field 20-node tetrahedral
elements. To ensure equal mesh configurations on opposite surfaces for applying
periodic boundary conditions three surfaces are first meshed with plane elements.
Then the plane mesh configurations are copied to the opposite surfaces and the
three dimensional mesh is generated based on the pre-meshed surfaces. With these
identical nodal configurations on opposite surface the periodic boundary conditions
can be applied as constraint equations between the appropriate nodal pairs.
For the evaluation of effective coefficients, the boundary conditions have to be
applied to the unit cell in such a way that, except one component of macroscopic strain
field vector in Eq. (6), all other strain components and T are equal to zero. Then each
effective coefficient can be easily determined by multiplying the corresponding row of
material matrix with the strain field vector. In the next subsections, implementations
of different boundary conditions to estimate all the effective coefficients of composite
are explained in detail with the help of Figure 1(b).
Evaluation of C11 and C12

For the calculation of C11 and C12 , we impose the boundary conditions in such
away that the macroscopic strain ¯ 11 is not equal to zero and all other strains and
T are zero in Eq. (6). This can be achieved by applying the appropriate constraint
equations to the different surfaces of the unit cell. For instance, consider a RVE
with unit size xjK+ − xjK− = 1 and ¯ 11 = 0
05, then Eq. (11) reduces as follows

1 − u1 =
uA+ ¯ 11 x1A+ − x1A− = 0
05
A−
(12)

2 − u2 =
uB+ ¯ 22 x2B+ − x2B− = 0 since ¯ 22 = 0
B−
(13)

3 − u3
uC+ = ¯ 33 x3C+ − x3C− = 0 since ¯ 33 = 0
C−
(14)
Here u1 u2 and u3 are the displacements in x1 x2 and x3 directions, respectively.

For the calculation of average stresses and strains ¯ 11 ¯ 22 and ¯ 11 according

to Eq. (6), the integral is replaced by a sum over averaged element values multiplied
by the respective element volume. Using these averaged values the coefficients
C11 and C12 can be calculated from the matrix Eq. (6). Due to zero strains and
temperature fields except ¯ 11 , the first row becomes ¯ 11 = C11 ¯ 11 . Then C11 can be
calculated as the ratio of ¯ 11 /¯11 . Similarly, C12 can be evaluated as the ratio of
¯ 22 /¯11 from the second row of matrix Eq. (6).

For the calculation of the effective coefficients C13 and C33 we have similar
conditions like for the calculation of the effective coefficients C11 and C12 . But now
prescribed displacements in the form of constraint equations in x3 direction must
be applied on parallel surfaces along x3 direction and prescribe zero displacements
in the form of constraint equations on all other parallel surfaces, i.e., x1 and x2
directions. For instance, if we consider the applied far filed strain ¯ 33 = 0
05 on the
unit size RVE, then applied constraint equations are as follows

1 − u1 =
uA+ ¯ 11 x1A+ − x1A− = 0
A−
(15)

2 − u2 =
uB+ ¯ 22 x2B+ − x2B− = 0
B−
(16)

3 − u3
uC+ = ¯ 33 x3C+ − x3C− = 0
05
C−
(17)
Using the average strains and stresses calculated from Eq. (4) and from the third
row of stiffness matrix Eq. (6), we get C33 as the ratio of ¯ 33 /¯33 and similarly, C13
can be evaluated as the ratio of ¯ 11 /¯33 from first row of matrix Eq. (6).

To evaluate the effective coefficient C66 the in-plane shear strain ¯ 12 may have
a non-zero value in strain-temperature field vector of Eq. (6) only. This can be
achieved by applying the appropriate constraint equations to the different surfaces
of the unit cell. For illustration, consider the unit size of RVE and ¯ 12 = 0
05, then
Eq. (11) reduced as follows
1 − u1 = 0
uA+ 2 − u2 = 0
05
uA+ 3 − u3 = 0
uA+
A− A− A−
(18)
uB+
1 − uB−
1 = 0
05 uB+
2 − uB−
2 = 0 uB+
3 − uB−
3 =0 (19)
Using the calculated non-zero average strain and stress values ¯ 12 and ¯ 12 then from
the sixth row in the matrix Eq. (6) we get ¯ 12 = C66 ¯ 12 and consequently, C66 can be
computed as the ratio of ¯ 12 /¯12 .
For the evaluation of C44 the out-of-plane shear strain ¯ 23 or ¯ 31 may have
a non-zero value only. Because of transverse isotropic symmetry, we can consider
either x1 − x3 or x2 − x3 planes as out-of-shear plane. For example, if we consider
shear strain ¯ 23 = 0
05 with unit size RVE, then the applied displacement constraint
equations are as follows
1 − u1 = 0
uB+ 2 − u2 = 0
uB+ 3 − u3 = 0
05
uB+
B− B− B−
(20)
1 − u1
uC+ = 0 uC+
2 − u2 = 0
05 uC+
3 − u3 =0
C− C− C−
(21)
From the non-zero average strain and stress, C44 can be computed as the ratio of
¯ 23 /¯23 .
Evaluation of 11 and 33

As we already know all the independent stiffness constants in matrix Eq. (6),
now in order to evaluate the effective coefficients of thermal expansion, we impose
the boundary conditions in such a way that, all strain fields of strain vector in Eq.
(6) are set to be zero and non-zero temperature field T is applied by providing
temperature difference to the unit cell as a body force and fix all the faces in all
directions as follows here.
1 = u1 = u2 = u2 = u3 = u3 = 0
uA+ A− A+ A− A+ A−
(22)
uB+
1 = uB−
1 = uB+
2 = uB−
2 = uB+
3 = uB−
3 =0 (23)
uC+
1 = uC−
1 = uC+
2 = uC−
2 = uC+
3 = uC−
3 = 0 and (24)
T = 1 (25)
Due to this temperature loading, stresses are induced inside the RVE. From Eq. (6)
we can evaluate the average stresses developed in all directions. Once we know the
¯ 11 , ¯ 22 and ¯ 33 with independent stiffness constants and temperature difference T
in Eq. (6), we obtain a system of three linear equations from which we can calculate
all effective thermal coefficients.
RESULTS AND DISCUSSION

The effective thermo-mechanical properties of transversely distributed
unidirectional cylindrical fiber composites are evaluated for different fiber volume
fractions ranging from 10% up to 70% with 10% intervals using FEM and AHM
approaches. All AHM results are calculated with formulae reported in Guinovart-
Diaz [29]. The material properties presented in Benveniste et al. [31] are used in the
micro-mechanical unit cell modeling given in the Table 1. In all these calculations
constant interphase volume fraction equal to 1.07% is considered, i.e., with increase
Table 1 Material properties of composite constituents
Thermal expansion
Constituent Young’s modulus (GPa) Poisson’s ratio coefficient (10−6 )/K
Nickel matrix 214 0.3113 13.3

Tungsten fiber 345 0.2778 5.0
Carbon coating 34.48 0.2022 3.3
of fiber volume fraction, the thickness of the interphase is reduced. From these
discrete fractions, graphs are interpolated and shown as comparison between
AHM and FEM in Figures 2 and 3. The results show a good agreement between
calculations by AHM and FEM.
Figure 2 Effective elastic stiffness constants: comparison FEM and AHM.

Figure 3 Effective coefficients of thermal expansion (CTE): comparison between FEM and AHM.
Now for further validation of proposed unit cell modeling using finite element
method, we compare the obtained results with Guinovart-Diaz et al. [29], Pagano-
Tandon [7] and Periodic Medium Homogenization (PMH) data reported in Table 5
of Lagache et al. [15]. In the PMH method, the asymptotic homogenization
method is applied to determine the effective properties of composite with hexagonal
distribution of the fibers and the local problems are solved by the finite element
method. In these comparisons we use a composite material made of Nicalon fiber
(volume fraction = 0.6) and barium magnesium aluminosilicate (BMAS) matrix with
an isotropic interphase material (volume fraction = 0.08), which is the result of
intermixing phenomena and migration of coupling agents inside the matrix as
reported in Lagache et al. [15]. The material properties of the constituents are listed
in Table 2.
Here we investigate the effect of the thickness of coating material on the
behavior of the Nicalon /BMAS composite system by considering the two different
volume fractions (0 and 0.08) of interphase coating material. Table 3 shows that,
with less than 8% of interphase volume fraction, the effective moduli Et , Ga and
Gt of the composite decrease abruptly lower than 50% of the composite properties
without interphase. Also, we can observe a good agreement between the different
approaches.
Table 2 Material properties of BMAS/Nicalon composite with interphase
Constituent Young’s modulus (GPa) Shear modulus (GPa)
BMAS matrix 106.0 43.0

Nicalon fiber 200.0 77.0
Interphase material 3.45 1.33
Table 3 Comparison between FEM, AHM and Pagano and Tandon [7]
and PMH of Lagache et al. [15]
Model Ea Et Ga Gt
FEM (V2 = 0) 162.571 153

506 61
369 60
428
AHM (V2 = 0) 162.578 153
508 61
363 60
431
Pagano & Tandon [7] (V2 = 0) 162.530 152
840 60
615 60
357
FEM (V2 = 0
08) 154.310 58
112 22
818 22
506
AHM (V2 = 0
08) 154.307 58
105 22
823 22
512
Pagano & Tandon [7] (V2 = 0
08) 154.296 58
332 23
057 22
850
PMH, Lagache et al. [15] (V2 = 0
08) 154.131 58
438 22
909 22
668
Table 4 illustrates the comparison between our approach (FEM), AHM,

numerical PMH and Self-Consistent Scheme (SCS) double approximations reported
in Table 2 of Lagache et al. [15]. The properties of the different media used in
the computation are E1 = 4 GPa, 1 = 2 = 0
34, E3 = 84 GPa and 3 = 0
22 and the
interphase Young’s modulus E2 is constant and taken in a range between 4 and
12 GPa. The fiber volume fraction is equal to V3 = 0
5 and the interphase volume
content is V2 = 0
1246. The Young’s and shear moduli of the composite increases
with increase of coating Young’s modulus for the different models. In this case,
some matrix material is replaced by a harder material. Also it can be observed that
the values for the elastic moduli derived from FEM model are close to those derived
from AHM, SCS and PMH models.
Table 4 The sensitivity of effective moduli to the elastic modulus of the coating E2 ,
which is revealed by different models
E2 Model Ea Et Ga Gt
4 FEM 44
0182 10
3241 4
0310 3
5917
AHM 44
0184 10
3247 4
0241 3
5990
SCS 44
02 10
31 4
020 3
590
PMH 44
0190 10
7550 4
0487 3
8052
6 FEM 44
2548 11
2647 4
4218 3
9210
AHM 44
2714 11
2615 4
4206 3
9360
SCS 44
27 11
37 4
412 3
931
PMH 44
2294 11
6448 4
3832 4
1420
8 FEM 44
5214 11
8014 4
6427 4
1342
AHM 44
5243 11
8095 4
6541 4
1356
SCS 44
52 11
79 4
642 4
130
PMH 44
4397 12
1702 4
5753 4
3444
10 FEM 44
7641 12
1786 4
8145 4
2689
AHM 44
7771 12
1761 4
8083 4
2705
SCS 44
78 12
12 4
793 4
261
12 FEM 45
0312 12
4394 4
9153 4
3625
AHM 45
0298 12
4395 4
9180 4
3681
SCS 45
03 12
40 4
900 4
353
PMH 44
8603 12
7665 4
7885 4
5771
CONCLUSIONS
A unit cell model is employed to predict the effective thermo-mechanical
properties of three-phase coated unidirectional cylindrical fibers using
homogenization techniques for different fiber volume fractions. The numerical
approach is based on the finite element method. Longitudinal and transversal
effective thermo-mechanical coefficients have been calculated with the finite
element model and compared with analytical solutions based on the asymptotic
homogenization method. The numerical results demonstrate that the developed
FEM approach is very accurate and efficient for the analysis of unit cell models of
fiber reinforced composites, with the presence of the interphases. The present work
has laid down a foundation for further applications of micro-mechanical finite
element analysis for problems, such as an investigation of stress field around the
fiber in order to understand the onset and the development of inelastic behavior
such as plastic deformation and possible damage. Furthermore the proved reliability
of the introduced FEM approach opens new possibilities to investigate composites
with arbitrary geometrical types of inclusions which cannot be covered by most
other homogenization methods.
REFERENCES
1. R. Hill, Theory of Mechanical Properties of Fiber-Strengthened Materials, I Elastic
Behavior. J. Mech. Phys. Solids, vol. 12, pp. 199–212, 1964.
2. Z. Hashin, On Elastic Behavior of Fiber-Reinforced Materials of Arbitrary Transverse
Phase Geometry. J. Mech. Phys. Solids, vol. 13, pp. 119–134, 1965.
3. Z. Hashin, Analysis of Properties of Fiber Composites with Anisotropic Constituents.
J. Appl. Mech., vol. 46, pp. 543–550, 1979.
4. L. J. Walpole, A Coated Inclusion in an Elastic Medium. Math. Proc. Comb. Phil. Soc.,
vol. 83, pp. 495–506, 1978.
5. Y. Mikata and M. Taya, Stress Field in and Around a Coated Short Fiber in an Infinite
Matrix Subjected to Uniaxial and Biaxial Loadings. ASME J. Appl. Mech., vol. 52,
pp. 19–24, 1985.
6. P. S. Theocaris and A. G. Varias, The Influence of the Mesophase on the Transverse
and Longitudinal Moduli and the Major Poisson’s Ratio in Fibrous Composites.
Colloid Polym. Sci., vol. 264, pp. 561–569, 1986.
7. N. J. Pagano and G. P. Tandon, Elastic Response of Multidirectional Coated-Fiber
Composites, Compos. Sci. Technol., vol. 31, pp. 273–293, 1988.
8. N. J. Pagano and G. P. Tandon, Thermo-Elastic Model for Multi-Directional Coated-
Fiber Composites: Traction Formulation. Compos. Sci. Technol., vol. 38, pp. 247–269,
1990
9. Z. Hashin, Thermo-Elastic Properties of Fiber Composites with Imperfect Interface.
Mech. Mater., vol. 8, pp. 333–348, 1990.
10. J. R. Yeh, The Effect of Interphase on the Transverse Properties of Composites. Int. J.
Solid Stress., vol. 29, no. 20, pp. 2493–2502, 1992.
11. M. Sutcu, A Recursive Concentric Cylinder Model for Composites Containing Coated
Fibers. Int. J. Solid Struct., vol. 29, pp. 197–213, 1992.
12. A. Dasgupta and S. M. Bhandarkar, A Generalized Self-Consistent Mori-Tanaka
Scheme for Fiber-Composites with Multiple Interphases. Mech. Mater., vol. 14,
pp. 67–82, 1992.
13. Y. Benveniste, G. J. Dvorak, and T. Chen, Stress Fields in Composites with Coated
Inclusions. Mech. Mater., vol. 7, pp. 305–317, 1989.
14. Y. C. Chu and S. I. Rokhlin, Determination of Fiber–Matrix Interphase Moduli from
Experimental Moduli of Composite with Multi-Layered Fibers. Mech. Mater., vol. 21,
pp. 191–215, 1995.
15. M. Lagache, A. Agbossou, and J. Pastor, Role of Interphase on Elastic Behavior of
Composite Materials: Theoretical and Experimental Analysis. J. Comp. Mater., Vol. 28
no. 12, pp. 1141–1157, 1994.
16. C. S. Chouchaoui and M. L. Benzeggagh, The Effect of Interphase on the Elastic
Behavior of a Glass/Epoxy Bundle. Compos. Sci. Technol., vol. 57, pp. 617–622, 1997.
17. A. Agbossou and J. Pastor, Thermal Stresses and Thermal Expansion Coefficients of n-
Layered Fiber-Reinforced Composites, Compos. Sci. Technol., vol. 57, pp. 249–260, 1997.
18. Y. J. Liu, N. Xu, and J. F. Luo, Modeling of Interphases in Fiber-Reinforced
Composites under Transverse Loading Using the Boundary Element Method. J. Appl.
Mech., vol. 67, pp. 41–49, 2000.
19. Z. Hashin, Thin Interphase/Imperfect Interface in Elasticity with Application to
Coated Fiber Composites. J. Mech. Phys. Solids, vol. 50, pp. 2509–2537, 2002.
20. R. Rodriguez Ramos, F. J. Sabina, R. Guinovart Diaz, and J. Bravo Castillero, Closed-
Form Expressions for the Effective Coefficients of Fiber-Reinforced Composite with
Transversely Isotropic Constituents. I: Elastic and Square Symmetry. Mech. Mater.,
vol. 33, pp. 223–235, 2001.
21. R. Guinovart-Diaz, J. Bravo-Castillero, R. Rodriguez-Ramos, and F. J. Sabina, Closed-
Form Expressions for the Effective Coefficients of a Fiber-Reinforced Composite with
Transversely Isotropic Constituents-I: Elastic and Hexagonal Symmetry. J. Mech. Phys.
Solids, vol. 49, pp. 1445–1462, 2001.
22. D. Golanski, K. Terada, and N. Kikuchi, Macro and Micro Scale Modeling of Thermal
Residual Stresses in Metal Matrix Composite Surface Layers by the Homogenization
Method. Comput. Mech., vol. 19, no. 3, pp. 188–202, 1997.
23. K. Terada and N. Kikuchi, Global-Local Constitutive Modeling of Composite Materials
by the Homogenization Method, Mater. Sci. Res. Int., vol. 2, no. 2, pp. 73–80, 1996.
24. H. Berger, S. Kari, U. Gabbert, R. Rodriguez-Ramos, R. Guinovart, J. A. Otero, and
J. Bravo-Castillero, An Analytical and Numerical Approach for Calculating Effective
Material Coefficients of Piezoelectric Fiber Composites, Int. J. Solid Struct., vol. 42,
pp. 5692–5714, 2005.
25. S. Kari, H. Berger, R. Rodriguez-Ramos, and U. Gabbert, Computational Evaluation
of Effective Material Properties of Composites Reinforced by Randomly Distributed
Spherical Particles. Composite Struct., vol. 77, pp. 223–231, 2007.
26. Z. Xia, C. Zhou, Q. Yong, and X. Wang, On selection of repeated unit cell model
and application of unified periodic boundary conditions in micro-mechanical analysis
of composites. Int. J. Solid. Struct., vol. 43, pp. 266–278, 2006.
27. J. C. Afonso and G. Ranalli, Elastic Properties of Three-Phase Composites: Analytical
Model Based on the Modified Shear-Lag Model and the Method of Cells. Compo. Sci. &
Tech., vol. 65, pp. 1264–1275, 2005.
28. R. Guinovart-Diaz, R. Rodriguez-Ramos, J. Bravo-Castillero, F. J. Sabina, J. A. Otero
and G. A. Maugin, A Recursive Asymptotic Homogenization Scheme for Multi-Phase
Fibrous Elastic Composites. Mech. Mater., vol. 37, pp. 1119–1131, 2005.
29. R. Guinovart-Diaz, R. Rodriguez-Ramos, J. Bravo-Castillero, F. J. Sabina and

G. A. Maugin, Closed-Form Thermo-Elastic Moduli of a Periodic Three-Phase Fiber-
Reinforced Composite. J. Therm. Stresses, vol. 28, pp. 1067–1093, 2005.
30. Z. Xia, Y. Zhang, and F. Ellyin, A Unified Periodical Boundary Conditions for
Representative Volume Elements of Composites and Applications. Int. J. Solid. Struct.,
vol. 40, pp. 1907–1921, 2003.
31. Y. Benveniste, G. J. Dvorak, and T. Chen. Stress Fields in Composites with Coated
Inclusions. Mech. Mater., vol. 7, pp. 305–317, 1989.
View publication stats

J Thermal Stresses 2007 Berger Etal

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

J Thermal Stresses 2007 Berger Etal

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Numerical and Analytical Approaches for Calculating the Effective

Article in Journal of Thermal Stresses · July 2007

Srihari Kurukuri Ulrich Gabbert

SEE PROFILE SEE PROFILE

Biomechanical mdel for FESS View project

The user has requested enhancement of the downloaded file.

Journal of Thermal Stresses

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

Copyright © Taylor & Francis Group, LLC

NUMERICAL AND ANALYTICAL APPROACHES FOR

H. Berger, S. Kurukuri, S. Kari, and U. Gabbert

R. Rodriguez-Ramos and J. Bravo-Castillero

The aim of this article is to envisage the effective thermo-mechanical properties

Keywords: Effective thermo-mechanical properties; FEM; Homogenization; Interphase; Three-phase

Received 2 November 2006; accepted 1 March 2007.

strength and stiffness, excellent frictional properties, good resistance to fatigue

In Dasgupta and Bhandarkar [12] a method to obtain the transversely

BASIC EQUATIONS: HOMOGENIZATION METHOD IN THERMOELASTICITY

ij x = Cijkl x/kl x − x/Tx

The strain tensor is deﬁned by the Cauchy relations

NUMERICAL HOMOGENIZATION OF RVE USING

For transversely isotropic symmetry, the engineering constants are related to

G23 = G13 = C44 G12 = C66

ui x1 x2 x3 = 0ij xj + u∗i x1 x2 x3 (8)

FINITE ELEMENT SIMULATIONS

Evaluation of C11 and C12

Here u1 u2 and u3 are the displacements in x1 x2 and x3 directions, respectively.

For the calculation of average stresses and strains ¯ 11 ¯ 22 and ¯ 11 according

Evaluation of C13 and C33

Evaluation of C44 and C66

equations are as follows

Evaluation of 11 and 33

RESULTS AND DISCUSSION

Table 1 Material properties of composite constituents

Nickel matrix 214 0.3113 13.3

Figure 2 Effective elastic stiffness constants: comparison FEM and AHM.

Table 2 Material properties of BMAS/Nicalon composite with interphase

Constituent Young’s modulus (GPa) Shear modulus (GPa)

BMAS matrix 106.0 43.0

FEM (V2 = 0) 162.571 153

Table 4 illustrates the comparison between our approach (FEM), AHM,

29. R. Guinovart-Diaz, R. Rodriguez-Ramos, J. Bravo-Castillero, F. J. Sabina and

View publication stats

You might also like

ij x = Cijkl x/kl x − x/Tx

ui x1 x2 x3 = 0ij xj + u∗i x1 x2 x3 (8)

For the calculation of average stresses and strains ¯ 11 ¯ 22 and ¯ 11 according