hjb/101130

COMPOSITES ENGINEERING, Part II

317.003

CONTINUUM MICROMECHANICS OF MATERIALS

Helmut J. B ohm Institut f ur Leichtbau und Struktur-Biomechanik TU Wien

c H.J. B ohm, ILSB/TUW, 1993, 2010

Contents
Notes Remarks on Composites 1 Introduction 1.1 Length Scales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Homogenization and Localization: Basic Notions . . . . . . . . . . . . 1.3 Material Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Basic Modeling Strategies in Continuum Micromechanics of Materials 2 Some Basic Analytical Approaches for Thermoelastic 2.1 Rules of Mixture . . . . . . . . . . . . . . . . . . . . . 2.2 The VFD Model of Dvorak and Bahei-el-Din . . . . . . 2.3 The CSA and CCA Models . . . . . . . . . . . . . . . . . . . . . . . . . . . iv v 1 1 2 3 5 10 10 11 12 14 14 17 19 22 25 28 28 31 32 33 33 35 36 37

Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 Mean Field Methods 3.1 General Relations between Mean Fields in Thermoelastic Two-Phase Materials 3.2 Mist Strains: Eshelbys Solution . . . . . . . . . . . . . . . . . . . . . . . 3.3 Dilute Inhomogeneous Inclusions . . . . . . . . . . . . . . . . . . . . . . . 3.4 MoriTanaka Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 Self-Consistent Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Other Mean Field and Related Estimates . . . . . . . . . . . . . . . . . . . 3.7 Mean Field Methods for Elastoplastic Composites . . . . . . . . . . . . . . 3.8 Mean Field Methods for Nonaligned Composites . . . . . . . . . . . . . . . 3.9 Mean Field Methods for Diusion-Type Problems . . . . . . . . . . . . . . 4 Bounding Methods 4.1 Hill and HashinShtrikman-Type Bounds . . . . . . . . . . . . . . . 4.2 Improved Bounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Bounds for the Nonlinear Behavior . . . . . . . . . . . . . . . . . . 4.4 Comparisons Between Some Mean Field and Bounding Predictions . . . . . . . . . . . . . . . . .

5 General Remarks on Modeling Approaches Based on Discrete Microgeometries 43 6 Periodic Microeld Models 6.1 Basic Concepts of Unit Cell Models . . . . . . . . . . . . . . . . . . . . . . 6.2 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii 47 47 49

6.3 6.4 6.5 6.6 6.7 6.8

Application of Loads and Evaluation of Microelds . . . . . . Unit Cell Models for Continuous Fiber Reinforced Composites Unit Cell Models for Short Fiber Reinforced Composites . . . Unit Cell Models for Particle Reinforced Composites . . . . . Unit Cell Models for Woven and Laminated Composites . . . Unit Cell Models for Porous and Cellular Materials . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

53 57 61 64 68 70 73 76 80 83

7 Embedded Cell Models 8 Windowing Approaches 9 Multi-Scale Models 10 Closing Remarks

iii

Notes
In the following, Nye notation is used for mechanical variables, i.e., tensors of order 4 are written as 6 6-quasi-matrices, and stress-like as well as strain-like tensors of order 2 as 6-quasi-vectors. These 6-vectors are connected to index notation by the relations 11 (1) 11 (1) (2) 22 (2) 22 (3) 33 (3) 33 = , = = 12 = 12 (4) (4) (5) 13 (5) 13 23 (6) 23 (6)

where ij = 2ij are the shear angles. Tensors of order 4 are denoted by bold upper case letters, stress- and strain-like tensors of order 2 by bold lower case Greek letters, and 3vectors by bold lower case letters. All other variables are taken to be scalars. Note that the use of the present notation requires that the 4th order tensors show orthotropic or higher symmetry and that the coecients of Eshelby and concentration tensors may dier compared to index notation. The contraction between a tensor of order 2 and a 3-vector is denoted by the symbol , where [ n]i = ij nj , and the tensorial product between two tensors of order 2 as well as the dyadic product between two vectors are denoted by the symbol , where [ ]ijkl = ij kl , and [a n]ij = ai nj , respectively. A superscript T denotes the transpose of a tensor or vector. Constituents (phases) are denoted by superscripts, with (p) standing for a general phase, for a matrix, (i) for inhomogeneities, and (f) for bers.

(m)

References to equations and section numbers beginning with a roman I are understood to pertain to the rst part of the lecture Composites Engineering, which is given by Prof. F.G.Rammerstorfer. For information on the research work in the eld of continuum micromechanics of materials carried out at the Institut f ur Leichtbau und Struktur-Biomechanik see the web pages http://ilsb.tuwien.ac.at/ilfb/ilfb ra13.html.

iv

Remarks on Composites
Inhomogeneous materials consist of dissimilar constituents (or phases) that are distinguishable at some length scale(s), called the characteristic length(s) of the inhomogeneities (e.g., the diameters of or distances between the reinforcements of composites). Each constituent shows dierent material properties (as, e.g., in composite and porous materials) and/or material orientations (as, e.g., in polycrystals) and may again be inhomogeneous at some smaller length scale. There are numerous schemes for the classication and nomenclature of inhomogeneous and multiphase materials. The following list, which is neither comprehensive nor complete, gives a number of classication schemes which are relevant to the present course. Production route: natural composites (e.g. wood, bone, alloys containing precipitates or inclusions, in-situ composites, polycrystalline materials, many porous and cellular materials) articial composites (composite materials in the strict sense1 ), foams, syntactic foams, functionally graded materials (FGMs) Matrix material (for articial composites): polymer matrix composites (PMCs) thermoset or thermoplastic matrix; often referred to in terms of the reinforcement: GFRP, CFRP, hybrid composites . . . metal matrix composites (MMCs) often taken to include intermetallic matrix composites (IMCs), which are brittle at room temperature brittle matrix composites ceramic matrix composites (CMCs), glass matrix composites, carboncarbon composites (CCCs)

Such materials are designed to take advantage of the dierent properties of the constituents to achieve some overall behavior(s). Production-related issues of many composite materials and their usage in lightweight structures are covered in the ILSB course Leichtbau mit faserverst arkten Kunststoffen given by I.C. SkrnaJakl.

 Mesoscale topology (for articial composites): laminates composites containing woven, knitted, braided reinforcements Microscale topology: matrixinclusion topologies interwoven topologies (e.g. duplex steels; open cell foams) layered topologies Phase arrangement statistics statistically homogeneous phase arrangements statistically inhomogeneous (e.g. graded) phase arrangements Shape of the reinforcements (for matrixinclusion topologies): continuously reinforced; typically unidirectional (UD) bers long ber reinforced; may signify either continuous bers or bers markedly exceeding the critical length (compare eqn. (I1.41)) short ber reinforced: aspect ratios are clearly larger than 1; length typically comparable to or exceeding the critical length particle reinforced: aspect ratios close to 1 (equiaxed) platelet reinforced: aspect ratios clearly smaller than 1 hybrid reinforcements Orientation of the reinforcements (for matrixinclusion topologies): aligned, unidirectional nonaligned (e.g. due to production processes); orientations may be described by orientation distribution functions (ODFs) random, planar random Relative size of the reinforcements (matrixinclusion topologies): monodisperse sizes (e.g., equal ber diameters) bidisperse sizes (e.g., two clusters of particle radii) polydisperse sizes

vi

Chapter 1 Introduction
1.1 Length Scales

Descriptions of the properties of composite materials have to account for at least two (but often three or more) length scales2 : Macroscale: length scale of the structure, component or sample Mesoscale: intermediate length scale (e.g., lamina level in layered composites) Microscale: length scale of the inhomogeneities, e.g., reinforcement diameters or distances In many cases, the constituents themselves are inhomogeneous at some smaller length scale, being, e.g., polycrystalline or porous. Figure 1.1 shows a schematic depiction of the various length scales in a hypothetical structure consisting of a stiened shell made of a metal matrix composite. The subject of the present lectures is continuum micromechanics, i.e., the study of mechanical properties of inhomogeneous materials within the framework of continuum mechanics while directly accounting for the phase arrangement at the microscale. Even though the emphasis is put on the thermomechanical behavior of two-phase composites, there is a large body of literature applying analogous or related methods both to other physical properties, such as thermal conduction, and to a wide range of other inhomogeneous materials. In micromechanical approaches, the stress and strain elds in an inhomogeneous material are split into contributions corresponding to the dierent length scales, which may be termed fast and slow variables. It is assumed that these length scales are suciently dierent so that the stress and strain elds at the smaller length scales (fast variables, microelds), (x) and (x), inuence the macroscopic behavior at the larger length scales only via their volume averages, but not via their uctuations, and
The nomenclature used here is far from universal, the naming of the length scales of inhomogeneous materials being notoriously inconsistent in the literature.
2

MACRO (structural)

MICRO 1 (composite) SUBMICRO 1 (nano)

MESO (laminate) MICRO 2 (polycrystal) SUBMICRO 2 (atomistic)

Figure 1.1: Schematic depiction of length scales in a hypothetical laminated MMC. gradients of the stress and strain elds at larger length scales (slow variables, macroelds), and , as well as compositional gradients are not signicant at the smaller length scales, where these elds appear constant and can be cast into the form of uniform far eld or applied stresses and strains. Viewed from the macroscale the behavior of a material that is inhomogeneous at some microscale can be described via an energetically equivalent homogenized continuum, provided the separation between the two length scales is suciently large and the above conditions are met. If these relations between fast and slow variables are not fullled to a sucient degree (e.g., near free surfaces of anisotropic materials or at macroscopic interfaces adjoined by at least one inhomogeneous material), continuum micromechanics must be used with extreme care and appropriate procedures must be employed.

1.2

Homogenization and Localization: Basic Notions

For regions of components or samples that do not exhibit signicant macroscopic stress or strain gradients, the microscopic strain and stress elds, (x) and (x), and the corresponding macroscopic responses, and , can be formally linked by localization relations (x) = A(x) (x) = B(x) and by homogenization relations of the type 1 s 1 = s = 1 2s s 1 (x) d = s s (x) d = (u(x) n (x) + n (x) u(x)) d
s

(1.1)

t(x) x d
s

(1.2)

Here s and s stand for the volume and the surface of a volume element, u(x) are the deformation vectors, t(x) = (x) n (x) the surface traction vectors, n (x) the unit surface normal vectors, and is the dyadic product of vectors. The tensors A(x) and B(x) in eqn. (1.1) are called mechanical strain and stress concentration tensors (Hill, 1963), respectively. The surface integral formulation for given above holds only for perfect interfaces between the constituents, in which case the mean strains and stresses in a control volume, and , are fully determined by the surface displacements and tractions. Otherwise correction terms involving the displacement jumps across gaps or cracks must be accounted for, see, e.g., Nemat-Nasser and Hori (1993). In the absence of body forces the microstresses (x) are self-equilibrated (but not necessarily zero). Equation (1.1) applies only to elastic behavior, but can be easily modied to cover thermoelastic behavior, compare eqn. (3.5), and the extension to the nonlinear range (e.g., to elastoplastic materials described by secant or incremental plasticity) is formally rather straightforward, compare section 3.7. The microgeometries of real inhomogeneous materials are at least to a certain extent random and, in nearly all practically relevant cases, are highly complex. Accordingly, exact expressions for A(x), B(x), (x) and (x) cannot realistically be provided for nontrivial composites and approximations have to be introduced. Typically, these approximations are based on the ergodic hypothesis, the heterogeneous material being assumed to be statistically homogeneous. Accordingly, suciently large subvolumes selected randomly within a sample are taken to give rise to the same eective material properties which, in turn, correspond to the overall material properties3 . Ideally, the homogenization volume should be chosen to be an appropriate representative volume element (RVE), which is a volume element that is of sucient size to contain all information necessary for describing the behavior of the composite. For discussions of RVEs and the boundary conditions applied to them see, e.g., (Hashin, 1983; Markov, 2000; Bornert et al., 2001; Kanit et al., 2003). In practice, however, it may be very dicult to identify RVEs and, accordingly, approximations to them are used in discrete microeld analysis. Such volume elements used for homogenization, denoted here as s , should be suciently large to allow a meaningful sampling of the microelds and suciently small for the inuence of macroscopic gradients to be negligible and for an analysis of the microelds to be possible.

1.3

Material Symmetries

The homogenized behavior of many multiphase materials can be idealized as being either statistically isotropic (e.g., composites reinforced by spherical particles, randomly oriented particles of general shape or randomly oriented bers, many polycrystals, many porous and
This requirement was symbolically denoted as MICROMESOMACRO by Hashin (1983), where MICRO and MACRO have their usual meanings and MESO stands for the size of the homogenization volume. Note that some inhomogeneous materials, e.g., graded materials, are not statistically homogeneous (and consequently non-ergodic) so that special procedures may be required to model them.
3

cellular materials, random mixtures of two phases) or statistically transversely isotropic (composites reinforced with aligned bers or platelets, composites reinforced with nonaligned inhomogeneities showing a planar random or some other axisymmetric orientation distribution function, etc.), compare Hashin (1983). Statistically isotropic multiphase materials show the same macroscopic behavior in all directions, and their eective elasticity and thermal expansion tensors take the forms (in Nye notation) E11 E12 E12 0 0 0 E12 E11 E12 0 0 0 E12 E12 E11 0 0 0 (1.3) = E= 0 . 0 0 0 E44 0 0 0 0 0 0 0 E44 0 1 0 0 0 0 0 0 E44 = 2 (E11 E12 ) Two independent parameters (chosen among moduli such as the eective Youngs modulus E , the eective Poisson number , the eective shear modulus G, the eective bulk modulus K =E/3(1 2 ), or the eective Lam e constants) are sucient for describing their overall linear elastic behavior and one is required for the eective thermal expansion behavior in the linear range, the coecient of thermal expansion = 11 .

The eective elasticity and thermal expansion tensors for statistically transversely isotropic materials have the structure E11 E12 E12 0 0 0 l E12 E22 E23 0 0 0 q E12 E23 E22 0 q 0 0 , (1.4) E= = 0 0 0 0 E44 0 0 0 0 0 0 0 E44 0 1 0 0 0 0 0 E66 = 2 (E22 E23 ) 0

where 1 is the axial direction and 23 is the transverse plane of isotropy. Generally, the thermoelastic behavior of transversely isotropic materials is described by ve independent elastic constants and two independent coecients of thermal expansion. Appropriate elastic parameters for this purpose are, e.g., the axial and transverse eective Youngs moduli, El and Eq , the axial and transverse eective shear moduli, Glq and Gqt , the axial and transverse eective Poisson numbers, lq and qt , as well as the eective transverse bulk 2 modulus KT =El /2[(1 qt )(El /Eq ) 2lq ]. The transverse (in-plane) properties are related via Gqt =Eq /2(1 + qt ), but for general transversely isotropic materials there is no linkage between the axial properties El , Glq and lq . Both an axial and a transverse eective coecient of thermal expansion, l = 11 and q = 22 , are required. For the special case of transversely isotropic aligned brous two-phase materials the relations El =
(f) El

+ (1 )E

(m)

4(lq (m) )2

(f)

lq = lq + (1 ) (m)

(f)

(f) (f) (m) (m) (1/KT 1/KT )2 KT KT (f) lq (m) 1 1 + (m) (f) (f) (m) KT 1/KT 1/KT KT KT

1 KT (1.5)

(Hill, 1964) allow the eective moduli El and lq to be expressed by KT , some constituent properties, and the ber volume fraction . Equations (1.5) can be used to reduce the number of independent eective elastic parameters required for describing unidirectional continuously reinforced composites to three. It is worth noting that the overall material symmetries of inhomogeneous materials and their eect on various physical properties can be treated in full analogy to the symmetries of crystals as discussed, e.g., by Nye (1957). The eects of the overall symmetry of the phase arrangement on the overall mechanical behavior of inhomogeneous materials can be marked4 , especially on the post-yield and other nonlinear responses to mechanical loads, compare Nakamura and Suresh (1993) or Weissenbek (1994). Accordingly, even though the overall behavior of actual multiphase materials may deviate to some minor degree from the ideal material symmetries discussed above, it is important to stay as close as possible to the appropriate overall symmetry in any modeling eort. Layer-type (generalized plane stress) material tensors for use with lamination and layered shell theories (which are important for mesoscopic and macroscopic descriptions of components made of composites materials, compare part I of this lecture) can be obtained from the fully three-dimensional material tensors given in eqns. (1.3) and (1.4) as l 0 0 0 E11 ,L E12 ,L q E12 ,L E22 ,L 0 0 0 (1.6) L = EL = 0 , 0 0 0 0 E44 ,L 0 0 0 0 0 E55 ,L 0 0 0 0 0 E66 ,L with Eij,L = Eij
Ej 3 E , E33 i3

compare Dorninger (1989).

1.4

Basic Modeling Strategies in Continuum Micromechanics of Materials

Homogenization methods aim at nding a volume elements responses to prescribed mechanical loads (typically far eld stresses or far eld strains) or temperature excursions and to deduce from them the overall thermomechanical tensors or moduli. The most straightforward application of such studies is materials characterization, i.e., simulating the material response under simple loading conditions such as uniaxial tensile tests. Analysis of this kind can be performed by all approaches described here (with the possible exception of bounding methods). Many homogenization procedures can also be employed directly as micromechanically based constitutive material models at higher length scales, i.e., they can link general macroscopic stress tensors to the corresponding strain tensors. This, of course, requires the capability of describing the overall response of an inhomogeneous material under any loading
Overall properties described by tensors of order 2, e.g., the thermal expansion response, are much less sensitive to symmetry eects than the elastic behavior, compare Nye (1957).
4

condition and for any loading history. Compared to semi-empirical constitutive laws, as proposed, e.g., by Davis (1996), micromechanically based constitutive models have both a clear physical basis and an inherent capability for zooming in on the local phase stresses by using localization procedures. The latter allow the local response of the phases to be found when the macroscopic response or state of a sample or structure is known, e.g., for predicting plastic yielding of the constituents or for assessing the initiation and progress of microscopic damage under given macroscopic loads. In addition to materials characterization and constitutive modeling, there are a number of other important applications of continuum micromechanics, among them studying local phenomena in inhomogeneous materials, e.g., the nucleation and growth of cracks, the stresses at intersections between macroscopic interfaces and free surfaces, the interactions between phase transformations and microstresses, and the elds in the vicinity of crack tips. For the latter behaviors details of the microstructure tend to be of major importance and can even determine the macroscopic response, an extreme case being the mechanical strength of brittle inhomogeneous materials. Because for realistic phase distributions the analysis of the spatial variations of the microelds in suciently large volume elements is typically beyond present capabilities, approximations have to be used. The majority of the resulting modeling approaches may be classied as falling into two groups. The rst of them comprises methods that describe the microgeometries of inhomogeneous materials on the basis of (limited) statistical information: Mean Field Approaches (MFAs) and related methods (see chapter 3): The microelds within each constituent of an inhomogeneous material are approximated by their phase averages (p) and (p) , i.e., piecewise (phase-wise) uniform stress and strain elds are employed. Such models typically use information on the microscopic topology, the reinforcement shape and orientation, and, in some cases, on statistical descriptors of details of the phase geometry. The localization relations then take the form
(p) (p)

(p) = A (p) = B

(1.7)

and the homogenization relations can be written as

(p)

= =

1 (p) 1 (p)

(x) d
(p)

with with

= =

V (p) V (p)

(p)

(p)

(x) d
(p)

(p)

(1.8)

where (p) stands for a given phase of the composite, (p) is the corresponding phase volume, and V (p) = (p) / k (k) is the volume fraction of the phase. Note that, in and B are not contrast to eqn. (1.1), for MFAs the phase concentration tensors A functions of the spatial coordinates within the volume element. Mean Field Approaches tend to be formulated in terms of the phase concentration tensors, eqn. (1.7), and they pose relatively low computational requirements. They 6

have been highly successful in describing the thermoelastic response of composites and other inhomogeneous materials; their use in modeling elastoplastic inhomogeneous materials is the subject of ongoing research. Bounding Methods (see chapter 4): Variational principles are used to obtain upper and (in many cases) lower bounds on the overall elastic tensors, elastic moduli, secant moduli, and other physical properties of inhomogeneous materials. Bounds aside from their intrinsic interest are important tools for assessing other models of inhomogeneous materials. In addition, in most cases one of the bounds provides good estimates for the physical property under consideration, even if the bounds are rather slack (Torquato, 1991). No variational bounds are available for the microelds. Many bounding methods are closely related to mean eld estimates. The second group of approximations are based on studying discrete microstructures and includes the following groups of models, compare the sketches in g. 1.2. Periodic Microeld Approaches (PMAs) or Periodic Homogenization (see chapter 6): The inhomogeneous material is approximated by an innitely extended periodic model material. The corresponding periodic microelds are typically obtained by analyzing unit cells (which may describe microgeometries of a wide range of complexity) via analytical or numerical methods. Periodic homogenization can handle constitutive modelling for both linear and nonlinear behaviors and, at present, is the most exible method of continuum micromechanics. The high resolution of the microelds provided by PMAs can be highly useful for studying the initiation of damage at the microscale. However, because they inherently give rise to periodic congurations of all relevant elds, PMAs are not suited to investigating phenomena involving damage that is localized rather than diuse (e.g., the interaction of the microstructure with a macroscopic crack). Periodic microeld approaches can give detailed information on the local stress and strain elds within a given unit cell. Embedded Cell or Embedding Approaches (ECAs; see chapter 7): The behavior of the inhomogeneous material is approximated by a model consisting of a core, i.e., a resolved discrete phase arrangement, that is embedded within some outer region to which far eld loads are applied. The material properties of this outer region may be described by some macroscopic constitutive law, they can be determined (quasi) self-consistently from the behavior of the core, or the embedding region may take the form of a coarse description and/or discretization of the phase arrangement. ECAs are highly useful for materials characterization, and they are usually the best choice for modeling regions of special interest, such as crack tips and their surroundings, in inhomogeneous materials. Embedding approaches can be used when the length scales are not well resolved, e.g., when then far eld loads or the composition of the material are not uniform. They can resolve local stress and strain elds in the core region at high detail but tend to be computationally expensive. Windowing Approaches (see chapter 8): Subregions of simple shape (windows) are chosen at random from a given phase arrangement and are subjected to macrohomogeneous and/or mixed uniform boundary conditions. The former type of boundary conditions give rise to lower and upper estimates and bounds for the moduli or tensors describing the overall behavior of a given inhomogeneous material, whereas the 7

PHASE ARRANGEMENT

PERIODIC APPROXIMATION, UNIT CELL

EMBEDDED CONFIGURATION

WINDOW

Figure 1.2: Schematic sketch of a random matrixinclusion microstructure and of the volume elements used by a periodic microeld method (which employs a slightly dierent periodic model microstructure), an embedding scheme and a windowing model for studying this inhomogeneous material. latter lead to estimates. At present the main elds of applications of windowing approaches involve linear behaviors of inhomogeneous materials. For small inhomogeneous samples (i.e., samples the size of which exceeds the microscale by not more than, say, an order of magnitude) the full microstructure can be modeled and studied, see, e.g., Papka and Kyriakides (1994). In such samples boundary condition eects can play a prominent role. Since no transitions between length scales are involved in problems of this type, they are not micromechanical methods in the strict sense. Further descriptions of inhomogeneous materials, such as rules of mixtures (isostrain and isostress models) and semi-empirical formulae like the HalpinTsai equations (Halpin and Kardos, 1976), which in most cases have a rather weak physical background and limited predictive capabilities, are given a short discussion in chapter 2, where also a number of physically based models, such as expressions for the self-similar composite sphere assemblages (CSA) of Hashin (1962) and composite cylinder assemblages (CCA) of Hashin and Rosen (1964), are introduced. For studying materials that are inhomogeneous at a number of (suciently dierent) length scales (e.g., materials in which well dened clusters of inhomogeneities are present) hierarchical procedures that use homogenization at more than one level are a natural exten8

sion of the above concepts. Such multi-scale models are the subjects of a short discussion in chapter 9. All of the continuum micromechanical methods discussed in the present course notes are suitable for homogenization problems. Mean eld and periodic microeld procedures can be used for localization tasks without major restrictions, whereas bounding methods are not targeted at localization. Embedding and windowing methods show boundary layers in the vicinity of the surface of the volume elements, where local elds are perturbed by the embedding material or by the prescribed uniform boundary conditions. The present course notes concentrate on classical composites, i.e., inhomogeneous materials showing matrixinclusion microtopology, and they are limited to the special case of two-phase composites. Most of the expressions and methods can, however, be extended to multi-phase materials.

Chapter 2 Some Basic Analytical Approaches for Thermoelastic Composites

The analytical approaches described in the present section are not mean eld methods, but some of them can be used within a mean eld framework. They are, in general, considerably simpler than mean eld methods proper, but tend to be less accurate and/or less exible.

2.1

Rules of Mixture

In the most general case, rule of mixture expressions for some scalar eective physical property of a two-phase composite take the form = ((i) ) + (1 )((m) )
1

(2.1)

where (i) and (m) denote inhomogeneities (bers or particles) and matrix, respectively, and the exponent must be chosen to obtain a good t to experimental data. The rule of mixture expressions given in sections I1.1.1 and I1.3.2a are special cases of eqn.(2.1) with = 1 (Voigt models) and = 1 (Reuss models), respectively, which, in contrast to most other choices of , have clear physical interpretations. Voigt-type expressions correspond to full strain coupling of the phases (springs in parallel, isostrain models, arithmetic averages) and Reuss-type expressions to full stress coupling (springs in series, isostress models, harmonic averages), i.e., they describe the in-plane and outof-plane behavior, respectively, of a layered material made up of two materials having the same Poisson number. The usefulness of Voigt- and Reuss-type expressions for actual composites depends strongly on the given microtopology and material properties. They are closely related to the Hill bounds, eqn. (4.1), and, with few exceptions, their predictions lie outside the HashinShtrikman bounds discussed in section 4.1. Even though it neglects Poisson eects, the Voigt-type expression El = El + (1 )E (m)
(f)

(2.2)

is usually a very good approximation to the axial Youngs modulus of continuously reinforced unidirectional composites. It can also be extended to nonlinear material behavior of 10

the phases, giving reasonable results, e.g., for the axial response of continuously reinforced unidirectional MMCs. Furthermore, the assumption of strain coupling gives rise to the following simple, but useful expression for the axial thermal expansion behavior of long ber reinforced UD composites l = El l + (1 )E (m) l El + (1 )E (m)
(f) (f) (f) (m)

(2.3)

Reuss-type models for the overall behavior of particle reinforced materials or for the transverse and shear properties of continuously reinforced composites typically generate excessively soft macroscopic responses. Rules of mixture can in principle be used to generate eective elastic tensors (and consequently, using eqn.(3.13), concentration tensors) that may be employed in a mean eld framework. Because they do not intrinsically account for the relationships between the engineering moduli, however, such procedures are either inconsistent or non-unique. Another semi-empirical description of the overall properties of composites is given by the HalpinTsai equations (Halpin and Kardos, 1976), = (m) 1 + 1 with = (i) (m) (i) + (m) , (2.4)

(compare section I1.1.1), which can be viewed as a t giving the correct behavior for the limiting cases = 0 and = 1. The tting parameter depends on the microgeometry of the composite and on the physical property to be described. The HalpinTsai equations have been used mostly in assessing experimental results and are rather limited in their predictive capabilities. For them, too, predictions for dierent engineering moduli are not necessarily consistent in terms of the materials overall symmetry.

2.2

The VFD Model of Dvorak and Bahei-el-Din

An approach that is closely related to the rules of mixture, but gives consistent and unique overall material tensors for unidirectional continuously reinforced composites is known as the Vanishing Fiber Diameter (VFD) model (Dvorak and Bahei-el Din, 1982). The physical interpretation of the VFD model is a composite containing aligned and continuous, but innitely thin bers (which strongly inuence the axial eective behavior, but aect the transverse behavior of the composite only via the macroscopic Poisson eect) in a matrix. VFD expressions are obtained by employing Voigt-type formulae for the eective axial Youngs modulus and the eective axial Poisson number, whereas Reuss-type expressions are used for the axial and transverse shear moduli and for the transverse bulk modulus El = E (f) + (1 )E (m) lq = (f) + (1 ) (m) Glq = Gqt = [/G(f) + (1 )/G(m) ]1 KT = [/KT + (1 )/KT ]1 11
(f) (m)

(2.5)

These equations do not necessarily fulll Hills relations, eqn (1.5). The VFD model tends to underestimate the eective transverse stiness of composites (typically some predictions lie outside the HashinShtrikman bounds, see section 4.1) and the mean hydrostatic stresses in the matrix. Due to its simplicity, however, it has been a popular description for continuously reinforced elastoplastic and viscoelastoplastic UD composites, giving good results for ber dominated properties and reasonable predictions for the strain hardening behavior in matrix dominated deformation modes.

2.3

The CSA and CCA Models

The composite sphere assemblage (CSA) model of Hashin (1962) and the composite cylinder assemblage (CCA) model of Hashin and Rosen (1964) are of considerable interest because they give exact expressions for some engineering moduli of special, but fairly realistic, phase arrangements that represent particle reinforced and aligned continuously reinforced composites, respectively. It may be noted that even though the derivations of the CCA and CSA methods do not involve phase averaged microstresses and strains, the results can be directly interpreted in terms of mean elds. Both CSA and CCA are based on analyzing control volumes tightly packed with either composite spheres or composite cylinders of dierent diameters. The cores of these composite spheres/cylinders consist of the reinforcement, and the matrix is placed in a concentric shell of the appropriate thickness to give the desired volume fractions, compare g. 2.1. These representative volume elements are subjected to suitable homogeneous boundary conditions, from which the boundary conditions for a single cylinder or sphere are deduced. The appropriate dierential equations governing elastic deformation of the composite spheres or cylinders are then solved and the overall moduli are obtained.

Figure 2.1: Microgeometry corresponding to the CCA and CSA methods (Hashin, 1983) 12

If this process is carried out for the radial deformation of composite spheres, the CSA expression for the eective bulk modulus K of a particle reinforced composite results as K = K (m) + 1/(K (i) K (m) ) + 3(1 )/(3K (m) + 4G(m) ) . (2.6)

For continuously reinforced unidirectional composites, the load cases of uniform axial extension, radial deformation in the transverse plane, as well as uniform axial shearing displacements and tractions can be handled, giving rise to the CCA expressions El = lq =
(f) El

+ (1 )E + (1 )

(m)

+ +

4 (1 )(lq (m) )2 (1 )/KT + /KT + 1/G(m) (1 )(lq (m) )(1/KT 1/KT ) (1 )/KT + /KT + 1/G(m)
(f) (m) (f) (m) (f) (f) (m)

(f)

(f) lq (m)

(m)

KT = K T + Glq = G(m)

1/(KT KT ) + (1 )/(KT + G(m) ) . + (f) 1/(Glq G(m) ) + (1 )/2G(m)

(f)

(m)

(m)

(2.7)

Shear loading of composite spheres and transverse shear loading as well as transverse extension of composite cylinders cannot be handled within this framework, so that the moduli G (in the case of particle reinforced composites) and Eq , Gqt as well as qt (in the case of continuously reinforced UD composites) must be evaluated by other means, the usual choice being a three-phase self-consistent scheme, see section 3.5.

13

Chapter 3 Mean Field Methods

Mean eld methods in continuum micromechanics aim at obtaining the overall properties of inhomogeneous materials, such as their overall elasticity and compliance tensors, E and C, respectively, and their overall tensor of coecients of thermal expansion (CTEs), , in terms of the appropriate phase properties and of information on the phase topology and geometry. The descriptions are based on phase averaged stress and strain elds in the constituents, in terms of which localization is also carried out.

3.1

General Relations between Mean Fields in Thermoelastic Two-Phase Materials

In the present chapter, a number of general relations between the phase averaged elds, the overall material tensors and the phase concentration tensors are given and mean eld models for dilute and non-dilute composites are introduced. For thermoelastic materials, the overall stressstrain relations take the form = E + T = C + T

(3.1)

where = E is the overall specic thermal stress tensor (i.e., the overall stress response of the fully constrained material to a purely thermal unit load) and T is the (spatially homogeneous) temperature dierence with respect to some stress-free reference temperature. The phase averaged stresses and phase averaged strains in the materials constituents can be dened as
(m)

(m)

1 (m) 1 = (m)

(x)d
(m)

(i)

(x)d
(m)

(i)

1 (i) 1 = (i)

(x)d
(i)

(x)d
(i)

(3.2)

where (m) and (i) are the volumes taken up by matrix and reinforcement, respectively, with s = (m) + (i) . The phases themselves are taken to behave thermoelastically, i.e., 14

(m) (m)

= E(m) (m) + (m) T = C(m) (m) + (m) T

(i) (i)

= E(i) (i) + (i) T = C(i) (i) + (i) T

(3.3)

where (m) = E(m) (m) and (i) = E(i) (i) . From the denition of volume averaging, eqn. (3.2), the relations between the phase averaged elds, = =
(i)

+ (1 ) (m) = a (i) + (1 ) (m) = a

(3.4)

are immediately obtained, where = V (i) = (i) /s stands for the volume fraction of the reinforcements and 1 = V (m) = (m) /s for the volume fraction of the matrix. a and a are the far eld (applied) homogeneous stress and strain elds, respectively, with a = C a . Perfect interfaces between the phases are assumed in expressing the macroscopic strain of the composite as the weighted sum of the phase averaged strains. The phase averaged strains and stresses can be related to the overall strains and stresses (Hill, 1963), respec, , and , B by the phase strain and stress concentration tensors A tively, which are dened for thermoelastic inhomogeneous materials by the expressions
(m) (m)

(m) + (m) T =A (m) T (m) + =B

(i) (i)

(i) + (i) T =A (i) T (i) + =B

(3.5)

compare eqn. (1.7) for the purely elastic case. By using eqns. (3.4) and (3.5), the phase averaged phase strain and stress concentration tensors (which will be called strain and stress concentration tensors, respectively, from now on) can be shown to fulll the relations (i) + (1 )A (m) = I A (i) + (1 )B (m) = I B (i) + (1 ) (m) = o (i) + (1 ) (m) = o

(3.6)

where I stands for the rank 4 (symmetric) unit tensor and o for the rank 2 null tensor. The eective elasticity and compliance tensors, respectively, of the composite can be evaluated from the properties of the phases and from the concentration tensors using the relationships (i) + (1 )E(m) A (m) E = E(i) A (i) = E(i) + (1 )[E(m) E(i) ]A (m) = E(m) + [E(i) E(m) ]A (i) + (1 )C(m) B (m) C = C(i) B (i) = C(i) + (1 )[C(m) C(i) ]B (m) = C(m) + [C(i) C(m) ]B 15

(3.7)

(3.8)

and the eective thermal expansion coecient tensor, , and the specic thermal stress tensor, , can be obtained as + (i) ] + (1 )[C(m) = [C(i) + (m) ] (m) = (i) + (1 )(m) + (1 )[C(m) C(i) ] (i) = (i) + (1 )(m) + [C(i) C(m) ] .
(i) (m)

(3.9)

(i) + (i) ] + (1 )[E(m) (m) + (m) ] = [E(i) = (i) + (1 )(m) + (1 )[E(m) E(i) ] (m) = (i) + (1 )(m) + [E(i) E(m) ] (i) .

(3.10)

The above expressions can be derived by inserting eqns. (3.3) and (3.5) into eqns. (3.4) and comparing with eqns. (3.1). In addition, can be obtained as (i))T (i) + (1 )(B (m) )T (m) = (B an expression known as the Levin (1967) formula. By invoking the principle of virtual work relations were developed (Benveniste and Dvo (p) , rak, 1990; Benveniste et al., 1991) which link the thermal strain concentration tensors, (p) , and the thermal stress concentration to the mechanical strain concentration tensors, A (p) , to the mechanical stress concentration tensors, B (p) , respectively, tensors, (m) ][E(i) E(m) ]1 [(m) (i) ] (m) = [I A (i) ][E(m) E(i) ]1 [(i) (m) ] (i) = [I A (m) = [I B (m) ][C(i) C(m) ]1 [(m) (i) ] (i) = [I B (i) ][C(m) C(i) ]1 [(i) (m) ] , (3.11)

(3.12)

Furthermore, by using eqns. (3.7) and (3.8), the elastic strain and stress concentration tensors can be obtained from the eective plus the phase elasticity and compliance tensors, respectively, as (m) = A (m) B 1 [E(m) E(i) ]1 [E E(i) ] 1 1 [C(m) C(i) ]1 [C C(i) ] = 1 (i) = 1 [E(i) E(m) ]1 [E E(m) ] A 1 (i) = [C(i) C(m) ]1 [C C(m) ] . (3.13) B

Analogous expressions for the thermal strain and stress concentration tensors result as (m) = (i) (m) (i) 1 [E(m) E(i) ]1 [ (i) (1 )(m) ] 1 1 (i) = [E E(m) ]1 [ (i) (1 )(m) ] 1 [C(m) C(i) ]1 [ (i) (1 )(m) ] = 1 1 (i) [C C(m) ]1 [ (i) (1 )(m) ] . = 16

(3.14)

In addition, the following set of equations can be derived, which provide relations between the stress and strain concentration tensors of each phase (m) = C(m) B (m) E(i) [I + (1 )(C(m) C(i) )B (m) E(i) ]1 = C(m) B (m) E A (i) = C(i) B (i) E(m) [I + (C(i) C(m) )B (i) E(m) ]1 = C(i) B (i) E A (m) = E(m) A (m) C(i) [I + (1 )(E(m) E(i) )A (m) C(i) ]1 = E(m) A (m) C B (i) = E(i) A (i) C(m) [I + (E(i) E(m) )A (i) C(m) ]1 = E(i) A (i) C . B (3.15)

(3.16)

From eqns. (3.12), (3.15) and (3.16) it is evident that the knowledge of one elastic phase concentration tensor is sucient for describing the full thermoelastic behavior of a twophase inhomogeneous material within the mean eld framework5 . Finally, the energetic equivalence in elasticity between the behaviors on the micro- and macroscales can be expressed by the relation 1 1 T = 2 2 T (x) (x) d =

1 2

(3.17)

where the (x) are general statically admissible stress elds and the (x) are general kinematically admissible strain elds. Equation (3.17) is known as Hills macrohomogeneity condition or the MandelHill condition. It states that the volume average of the elastic energy density evaluated on the basis of the microscopic elds must be equal to the elastic energy density of the homogenized material.

3.2

Mist Strains: Eshelbys Solution

A large proportion of the mean eld descriptions used in continuum micromechanics of materials are based on the work of Eshelby (1957), who studied the stress and strain distributions in homogeneous media that contain a subregion that spontaneously changes its shape and/or size (undergoes a transformation) so that it no longer ts into its previous space in the parent medium. Eshelbys results show that if an elastic homogeneous ellipsoidal inclusion (i.e., an inclusion consisting of the same material as the matrix) in an innite matrix is subjected to a homogeneous strain t (called the stress-free strain, unconstrained strain, eigenstrain, or transformation strain), the stress and strain states in the constrained inclusion are uniform, i.e., (i) = (i) and (i) = (i) (Eshelby, 1957). The uniform strain in the constrained inclusion (the constrained strain), c , is related to the stress-free strain t by the expression c = St (3.18)

where S is referred to as the Eshelby tensor. For eqn. (3.18) to hold, t may be any kind of eigenstrain which is uniform over the inclusion (e.g., a thermal strain or a strain due to some phase transformation which involves no changes in the elastic constants of the inclusion).
Similarly, n1 elastic phase concentration tensors must be known for describing the overall thermoelastic behavior of an n-phase material.
5

17

For spheroidal inclusions (i.e., ellipsoids of rotation) in an isotropic matrix, S can be evaluated analytically and depends only on the Poissons ratio of the homogeneous material (or, in the case of inhomogeneous inclusions, on the Poissons ratio of the matrix) and on the aspect ratio a of the inclusion. Under these conditions the nonzero components of the Eshelby tensor can be expressed as S (1, 1) = S (2, 2) = S (3, 3) = S (1, 2) = S (1, 3) = S (2, 1) = S (3, 1) = S (2, 3) = S (3, 2) = S (4, 4) = S (5, 5) = S (6, 6) = 1 3a2 1 3a2 (m) (m) 1 2 + g (a) 1 2 + 2(1 (m) ) a2 1 a2 1 3 9 1 a2 1 2 (m) g (a) + (m) 2 (m) 2 8(1 ) a 1 4(1 ) 4(a 1) 1 3 1 1 2 (m) + 2 g (a) + 1 2 (m) + (m) 2 2(1 ) a 1 2(a 1) 3a2 1 a2 1 (1 2 (m) ) g (a) 2(1 (m) ) a2 1 4(1 (m) ) a2 1 1 a2 3 g (a) 1 2 (m) + (m) 2 2 4(1 ) 2(a 1) 4(a 1) 1 a2 + 1 1 3(a2 + 1) (m) (m) 1 2 1 2 g (a) 2(1 (m) ) a2 1 2 a2 1 3 1 a2 2(a2 1) + 1 2 (m) g (a) (3.19) (m) 2 2(1 ) 4(a 1)

(Tandon and Weng, 1984). Here the 1-direction is the axis of rotation of the spheroid and a stands for the aspect ratio of the inclusions (i.e., a is given by the length of the axis of rotation of the spheroids divided by their diameter, so that for continuous cylindrical bers a , for spherical inclusions a = 1, and for innitely thin circular disks or platelets a 0). The function g (a) is given by the expressions a g= 2 [a(a2 1)1/2 arcosh a] (a 1)3/2 for berlike (prolate) inclusions (a > 1) and a [arccos a a(1 a2 )1/2 ] g= (1 a2 )3/2 for disklike (oblate) inclusions (a < 1). For the special case of spherical inclusions (a = 1) in an isotropic matrix, the only nonzero components of the Eshelby tensor are 7 5 (m) 15(1 ) 5 (m) 1 S (1, 2) = S (2, 1) = S (1, 3) = S (3, 1) = S (2, 3) = S (3, 2) = 15(1 (m) ) 2(4 5 (m) ) S (4, 4) = S (5, 5) = S (6, 6) = 15(1 (m) ) S (1, 1) = S (2, 2) = S (3, 3) =

(3.20)

and for inclusions in the form of continuous bers (a ) of circular cross section they take the form 18

S (2, 2) = S (3, 3) = S (2, 3) = S (3, 2) = S (2, 1) = S (3, 1) = S (4, 4) = S (5, 5) = S (6, 6) =

5 4 (m) 8(1 (m) ) 4 (m) 1 8(1 (m) ) (m) 2(1 (m) ) 1 2 3 4 (m) 4(1 (m) )

(3.21)

Equations (3.19) to (3.21) follow the conventions of Nye notation as discussed in the introductory notes, with the shear terms in the strain vector being the shear angles ij = 2ij . A similar notation is employed by Pedersen (1983), but a number of standard works dealing with mistting inclusions, such as Mura (1987), use dierent conventions and consequently give dierent expressions for S (4, 4), S (5, 5) and S (6, 6).

3.3

Dilute Inhomogeneous Inclusions

Mean eld methods for dilute inhomogeneous matrixinclusion composites typically aim at making use of Eshelbys expressions for the elds in a homogeneous inclusion subjected to an eigenstrain by using the concept of an equivalent homogeneous inclusion. This strategy involves replacing the actual inhomogeneous inclusion (often referred to as an inhomogeneity), which has dierent material properties than the matrix and which is subjected to a given unconstrained eigenstrain t , with a (ctitious) equivalent homogeneous inclusion on which a (ctitious) equivalent eigenstrain is made to act. This equivalent eigenstrain must be chosen in such a way that the same stress and strain elds are obtained in the actual inhomogeneous inclusion and in the ctitious homogeneous inclusion. Following the strategy outlined above, the case of inhomogeneous inclusions is handled by investigating an equivalent homogeneous inclusion subjected to an appropriate equivalent eigenstrain . The latter is chosen in such a way that the inhomogeneous inclusion and the equivalent homogeneous inclusion attain the same stress state (i) and the same constrained strain c (Eshelby, 1957; Withers et al., 1989). When (i) is expressed in terms of the elastic strain in the inhomogeneity or inclusion, this condition translates into the equality (i) = E(i) (c t ) = E(m) (c ) . (3.22) Here c t and c are the elastic strains in the inhomogeneity and the equivalent homogeneous inclusion, respectively. Obviously, in the general case the stress-free strains will be dierent for the equivalent inclusion and the real inhomogeneity, t = . A typical route (but not the only one) for obtaining mean eld descriptions of inhomogeneous materials subjected to an eigenstrain consists of rst expressing the appropriate 19

equivalent eigenstrain of an equivalent homogeneous inclusion in terms of the constituents material properties, the shape of the inclusion, and the stress-free strain of the inhomogeneous inclusion. This equivalent eigenstrain is then used to generate the expressions for the average phase stresses and strains. Because eqn. (3.18) was derived for a general homogeneous inclusion problem, it will also hold for the equivalent homogeneous inclusion, for which it takes the form c = S . (3.23) By inserting this expression into eqn. (3.22) the relation (i) = E(i) (S t ) = E(m) (S I) (3.24)

is obtained, which can be rearranged to express the equivalent eigenstrain acting on the homogeneous inclusion in terms of the known stress-free eigenstrain t of the real inclusion as = [(E(i) E(m) )S + E(m) ]1 E(i) t . (3.25) By substituting eqn. (3.25) into the right hand side of eqn. (3.24), the stress in the inhomogeneity, (i) , is nally obtained as (i) = E(m) (S I)[(E(i) E(m) )S + E(m) ]1 E(i) t = E(m) Rdil E(i) t , where the tensor Rdil is dened as Rdil = (S I)[(E(i) E(m) )S + E(m) ]1 . (3.27)

(3.26)

As intended, eqn. (3.26) expresses the average stress in an inhomogeneous ellipsoidal inclusion in terms of the Eshelby tensor S (and thus of the inhomogeneity aspect ratio a), of the elastic tensors of matrix and inhomogeneity, and of the stress-free strain t , all of which are known. If a uniform external stress a is applied to an inhomogeneous elastic matrixinclusion system, the stress in the inhomogeneity, (i) , will be a superposition of this applied stress and of some additional stress caused by the constraining eect of the surrounding matrix on the inhomogeneity. Such problems can be treated by extending of the strategy followed in eqns. (3.23) to (3.27) and introducing an equivalent homogeneous inclusion subjected to both the applied stress a and to a suitable equivalent eigenstrain . Again, this equivalent eigenstrain is chosen in such a way that the stress, (i) , and the constrained strain, c , in the inhomogeneity are the same in the inhomogeneous and the equivalent homogeneous cases. Alternatively, a uniform far eld strain a may be used as far eld load. The equivalent eigenstrain depends on both the applied stress a (or applied strain a ) and, if present, on the eigenstrain of the inclusion, t , i.e., = ( a , t ) or = (a , t ). By writing the stress in the inhomogeneity as E(i) (a + c t ) and that in the equivalent homogeneous inclusion as E(m) (a + c ), i.e., in terms of the elastic tensors and the elastic strains, the relation (i) = E(i) (a + c t ) = E(m) (a + c ) 20 (3.28)

is obtained. Inserting Eshelbys result in the form of eqns. (3.23) to (3.27) and solving for the equivalent eigenstrain yields = [(E(i) E(m) )S + E(m) ]1 [(E(m) E(i) )a + E(i) t ] , (3.29)

which, as required, contains a contribution due to the applied strain a as well as a term due to the unconstrained eigenstrain t . Substituting eqn. (3.29) into the r.h.s. of eqn. (3.28) allows to express the stress in the inhomogeneity as (i) = E(m) {I + (S I)[(E(i) E(m) )S + E(m) ]1 (E(m) E(i) )}a + E(m) (S I)[(E(i) E(m) )S + E(m) ]1 E(i) t .

(3.30)

Because for a single inhomogeneity in an innite matrix the relationship a = C a = C(m) a holds, eqn. (3.30) can be rewritten as (i) = E(m) {I + (S I)[(E(i) E(m) )S + E(m) ]1 (E(m) E(i) )}C(m) a + E(m) (S I)[(E(i) E(m) )S + E(m) ]1 E(i) t = E(m) [I + Rdil (E(m) E(i) )]C(m) a + E(m) Rdil E(i) t where Rdil is dened by eqn. (3.27). (i) , By comparing the denition of the inhomogeneity stress concentration tensor B eqns. (1.7) and (3.5), with eqn. (3.31), by equating the eective stress in the composite, , with the applied stress, a , and by setting the transformation strain to zero, t = 0, an expression for the stress concentration tensor of a dilute composite can be directly obtained as (i) = E(m) [I + Rdil (E(m) E(i) )]C(m) B . (3.32) dil (i) , can be found by The thermal stress concentration tensor for the inhomogeneities, setting a = = 0 and inserting the thermal mismatch strain for the unconstrained eigenstrain in eqn. (3.31), i.e., t = ((i) (m) )T , and then comparing with eqn. (3.2), which gives (i) = E(m) Rdil E(i) ((i) (m) ) . (3.33) dil The matrix stress concentration tensors corresponding to eqns. (3.32) and (3.33) can be obtained easily by using eqn. (3.6). Alternative expressions for the stress and strain concentration tensors of dilute composites can be obtained by studying an equivalent inclusion in the absence of transformation strains, i.e., t = 0 (Hill, 1965; Benveniste, 1987). Under this condition, the strain in the inhomogeneity becomes (i) = (i) = a + c = a + S (3.34) and the equivalent to eqn. (3.28) takes the form (i) = E(i) (a + c ) = E(m) (a + c ) which, together with eqn. (3.34), provides the expression E(i)
(i)

(3.31)

(3.35)

= E(m) ( 21

(i)

(3.36)

Solving eqn. (3.36) for and substituting into eqn. (3.34) leads to the expression
(i)

= [I + SC(m) (E(i) E(m) )]1 a

(3.37)

from which the dilute inhomogeneity strain concentration tensor follows directly as (i) = [I + SC(m) (E(i) E(m) )]1 A dil . (3.38)

By setting (i) = C(i) (i) and using a = C(m) a , the dilute stress concentration tensor for the inhomogeneities is found from eqn. (3.37) as (i) = E(i) [I + SC(m) (E(i) E(m) )]1 C(m) B dil . (3.39)

The expressions for stress and strain concentration tensors given in this section were derived for a single inclusion in an innite matrix and hold for inhomogeneities that are dilutely dispersed in the matrix and thus do not feel any eects of their neighbors (i.e., they are loaded by the unperturbed applied stress a ). Accordingly, the dilute concentration (i) , are independent of the inhomogeneity (i) and (i) , B tensors for the inhomogeneities, A dil dil dil volume fraction . The inhomogeneity volume fraction does, however, enter the corresponding expressions for the matrix concentration tensors obtained via eqn. (3.6). The overall elastic and thermal expansion tensors of dilute matrixinclusion systems can be easily obtained by plugging the concentration tensors derived above, eqns. (3.32), (3.33), (3.38) and/or (3.39), into the relations (3.7) to (3.9). It must be kept in mind, however, that the resulting expressions only hold for dilute composites with inhomogeneity volume fractions in the range 0.1. It is worth noting that, within the framework of the equivalent inclusion approach, eqns. (3.19) to (3.21) can be used for any inhomogeneous inclusion in an isotropic matrix regardless of the inhomogeneitys material symmetry. Analytical expressions for the Eshelby tensor have also been given for spheroidal inclusions when the matrix shows transversely isotropic (Withers, 1989) or cubic material symmetry (Mura, 1987), provided the material axes of both constituents are aligned with the orientations of the inhomogeneities. In cases where no analytical solutions are available, the Eshelby tensor can evaluated numerically, compare Gavazzi and Lagoudas (1990). In general, the stress and strain elds outside an inclusion in an innite matrix are not uniform on the microscale (Eshelby, 1959) and can be described via the exterior point Eshelby tensor, see, e.g., Ju and Sun (1999). In mean eld theories, which aim to link the average elds in matrix and inclusions with the overall response of inhomogeneous materials, however, it is only the average matrix stresses and strains that are of interest.

3.4

MoriTanaka Estimates

Theoretical descriptions of the overall thermoelastic behavior of composites with reinforcement volume fractions of more than a few percent must account for the interaction between inhomogeneities, i.e., for the eects of the surrounding inhomogeneities on the stress and 22

strain elds experienced by a given ber or particle. One way for achieving this consists of approximating the stresses acting on an inhomogeneity, which may be viewed as perturbation stresses caused by the presence of other inhomogeneities superimposed on the applied far eld stress, by an appropriate average matrix stress. The idea of combining such a concept of an average matrix stress with Eshelby-type equivalent inclusion approaches goes back to Brown and Stobbs (1971) as well as Mori and Tanaka (1973). Eective eld theories of this type are generically referred to as MoriTanaka methods. By construction they do not invoke explicit (e.g., pair-wise) interactions between individual inhomogeneities, but rather operate on a level of collective interactions. It was pointed out by Benveniste (1987) that in the isothermal case the central assumption involved in MoriTanaka approaches can be denoted as
(i) (i)

(i) (m) = A dil (i) (m) = B dil

(3.40)

Thus, the methodology developed for dilute inhomogeneities is retained and the interactions with the surrounding inhomogeneities are accounted for by suitably modifying the stresses or strains acting on each inhomogeneity. Equation (3.40) may thus be viewed as a modication of eqn. (1.7) in which the macroscopic strain or stress, or , is replaced by the phase averaged matrix strain or stress, (m) and (m) , respectively. In a further step suitable expressions for (m) and/or (m) must be introduced into the scheme. Following Benveniste (1987), inserting eqns. (3.40) into eqns. (3.4) leads directly to relations between the macroscopic strains and stresses on the one hand and the inhomogeneity strains and stresses on the other hand = =
(i)

+ (1 ) + (1 )

(m)

+ (1 )I] = [ A dil
(i)

(i)

(m)

(i) + (1 )I][A (i) ]1 = [ A dil dil

(i) (m)

(i) + (1 )I] = [ B dil

(i)

(m)

(i) + (1 )I][B (i) ]1 = [ B dil dil Equations (3.41) can be rearranged into the form
(i) (i)

(3.41)

(i) [(1 )I + A (i) ]1 = A dil dil (i) (i) 1 = Bdil [(1 )I + Bdil]

(3.42)

from which MoriTanaka-expressions for the inhomogeneity strain and stress concentration tensors for non-dilute composites follow directly as (i) = A (i) [(1 )I + A (i) ]1 A M dil dil (i) (i) 1 (i) BM = Bdil [(1 )I + Bdil ]

(3.43)

Finally, the corresponding expressions for the matrix strain and stress concentration tensors can be found by inserting eqns. (3.43) into eqns. (3.41) to give (m) = [(1 )I + A (i) ]1 A M dil (i) 1 (m) BM = [(1 )I + Bdil] 23

(3.44)

Equations (3.43) and (3.44) automatically fulll conditions (3.6) and may be evaluated with any strain and stress concentration tensors pertaining to dilute inhomogeneities in a matrix. If the equivalent inclusion expressions, eqns. (3.38) and (3.39), are employed, the strain and stress concentration tensors for the non-dilute composite take the form (m) = {(1 )I + [I + SC(m) (E(i) E(m) )]1 }1 A M (m) BM = {(1 )I + E(i) [I + SC(m) (E(i) E(m) )]1 C(m)

(3.45)

The expressions for the eective elasticity and compliance tensors of the composite quoted in (Benveniste, 1987; Benveniste and Dvorak, 1990) can be recovered as [(1 )I + A ]1 EM = E(m) + [E(i) E(m) ]A dil dil (m) (i) (m) (i) (i) ]1 CM = C + [C C ]B [(1 )I + B
dil dil (i) (i)

(3.46)

by substituting eqns. (3.45) into eqns. (3.7) and (3.8). Finally, the specic thermal stress and thermal expansion coecient tensors follow from eqns. (3.10) and (3.9) as (i) ]1 [E(i) E(m) ]1 [(i) (m) ] (i) [(1 )I + A M = (m) + [E(i) E(m) ]A dil dil (m) (i) (m) (i) (i) ]1 [C(i) C(m) ]1 [(i) (m) ] M = + [C C ]B [(1 )I + B
dil dil

. (3.47)

Benvenistes interpretation of the MoriTanaka approach, eqns. (3.43) to (3.47), is equivalent to a method developed by Tandon and Weng (1984), which directly gives expressions for the overall elasticity tensor of two-phase inhomogeneous materials as ET = E(m) I [(E(i) E(m) ) S (S I) + E(m) ]1 [E(i) E(m) ]
1

(3.48)

Because eqn. (3.48) does not explicitly use the compliance tensor of the inhomogeneities, C(i) , it can be modied in a straightforward way to describe the macroscopic stiness of porous materials by setting E(i) 0, giving rise to the relationship Epor = E(m) I + (I S)1 1
1

(3.49)

which, however, should not be used for void volume fractions that are in excess of, say, = 0.256. For realistic stiness ratios between inhomogeneities and matrix (elastic contrasts), a number of further MoriTanaka-type theories (which use dierent derivations and obtain dierent formulas for the concentration tensors), evaluate to the same numbers for the concentration and overall elastic tensors as Benvenistes approach, among them Pedersens Mean Field Theory (Pedersen, 1983) and Wakashimas method (Wakashima et al., 1988).
MoriTanaka theories are based on the assumption that the shape of the inhomogeneities can be described by ellipsoids of a given aspect ratio throughout the deformation history. In porous materials with high void volume fractions deformation at the microscale takes place mainly by bending and buckling of cell walls or struts (Gibson and Ashby, 1988). Such eects cannot be described by MoriTanaka models, which, consequently, tend to overestimate the eective stiness of cellular materials by far.
6

24

As is evident from their derivation, MoriTanaka-type theories at all volume fractions describe composites consisting of aligned ellipsoidal inhomogeneities embedded in a matrix, i.e., inhomogeneous materials with a distinct matrixinclusion microtopology. More precisely, it was shown by Ponte Casta neda and Willis (1995) that MoriTanaka methods are a special case of so-called HashinShtrikman variational estimates in which the spatial arrangement of the inhomogeneities follows an ellipsoidal distribution with the same aspect ratio and orientation as the inhomogeneities themselves, compare g. 3.1.

Figure 3.1: Sketch of ellipsoidal inhomogeneities in an aligned ellipsoidally distributed spatial arrangement as used implicitly in MoriTanaka-type approaches (a=2.0). MoriTanaka predictions for the overall Youngs and shear moduli of composites reinforced by aligned or spherical reinforcements tend to be on the low side (see the comparisons in section 4.1), but generally are highly useful estimates. For discussions of the range of validity of MoriTanaka theories for elastic inhomogeneous two-phase materials see, e.g., Christensen et al. (1992). In addition to standard MoriTanaka methods describing microgeometries with aligned inhomogeneities having the same material properties, procedures have been developed for extending the MoriTanaka method to nonaligned or hybrid composites (i.e., materials with more than one inclusion phase), compare section 3.8. MoriTanaka-type theories can be implemented into computer programs in a straightforward way. Because they are explicit algorithms, all that is required are matrix additions, multiplications, and inversions plus expressions for the Eshelby tensor. This makes them important tools for evaluating the stiness and thermal expansion properties of inhomogeneous materials that show a matrixinclusion topology with aligned inhomogeneities or voids. MoriTanaka-type approaches for thermoelastoplastic materials are discussed in section 3.7.

3.5

Self-Consistent Estimates

An alternative way of extending the expressions for the elastic properties of dilute two-phase materials derived in section 3.3 to non-dilute volume fractions starts by rewriting eqn. (3.38) for an inhomogeneity that is surrounded by the eective medium (i.e., the composite 25

 (i) A (i) (E, C) and itself) instead of the matrix (i.e., E(m) E and C(m) C, so that A dil dil (i) B (i) (E, C)), compare g. 3.2. The results of the above formal procedure may be B dil dil inserted into eqn. (3.7), giving rise to the relationship (i) ESC = E(m) + (E(i) E(m) )A dil = E(m) + (E(i) E(m) )[I + SC(E(i) E)]1

(3.50)

Equation (3.50) can be interpreted as an implicit nonlinear system of equations for the unknown elastic tensors E=ESC and C=CSC , which describe the eective medium. This system can be solved by self-consistent iterative schemes of the type En+1 = E(m) + [E(i) E(m) ][I + Sn Cn (E(i) En )]1 Cn+1 = En+1
1

(3.51)

The Eshelby tensor Sn in eqn. (3.51) describes the response of an inhomogeneity embedded in the n-th iteration of the eective medium; it must be recomputed for each iteration7 .

i m m

(m) tot

11111111 00000000 00000000 11111111 00000000 11111111 00000000 11111111 00000000 11111111 i 00000000 11111111 00000000 11111111 00000000 11111111 eff 00000000 11111111
a

11111111 00000000 00000000 11111111 00000000 11111111 00000000 11111111 00000000 11111111 i 00000000 11111111 00000000 11111111 m 00000000 11111111 eff 00000000 11111111
a

dilute

MTM

CSCS

GSCS

Figure 3.2: Schematic comparison of the Eshelby method for dilute composites, Mori Tanaka approaches and classical as well as generalized self-consistent schemes. This approach is known as the two-phase or classical self-consistent scheme (CSCS), see, e.g., Hill (1965), and its predictions dier noticeably from those of MoriTanaka methods in being close to one of the HashinShtrikman bounds (compare section 4.1) for low reinforcement volume fractions, but close to the other bound for reinforcement volume fractions approaching unity (compare gs. 4.1 to 4.7). Generally, two-phase self-consistent schemes are best suited to describing the overall properties of two-phase materials that do not show a matrixinclusion microtopology at some or all of the volume fractions of interest. Essentially, the microstructures described by the CSCS are characterized by interpenetrating or granular phases around = 0.5, with one of the materials acting as the matrix for 0 and the other for 1. Consequently, the CSCS is typically not the best choice for describing composites showing a matrixinclusion topology, but is well suited to studying Functionally Graded Materials (FGMs) in which the volume fraction of a constituent can vary from 0 to 1 through the thickness of a sample. Multi-phase versions of the CSCS are important methods for describing polycrystals. Being implicit and requiring the iterative solution of nonlinear equations, the CSCS is computationally more demanding than MoriTanaka approaches.
For aligned non-spherical inhomogeneities the eective medium shows transversely isotropic behavior, so that an appropriate procedure must be used for evaluating the Eshelby tensor.
7

26

A more elaborate self-consistent approach, the three-phase or generalized self-consistent scheme (GSCS) of Christensen and Lo (1979, 1986), describes inhomogeneities surrounded by a matrix layer (of a thickness appropriate for obtaining the required reinforcement volume fraction) embedded in an eective medium, compare g. 3.2. By considering the dierential equations describing the elastic response of such three-phase regions under appropriate boundary conditions, equations for the overall elastic moduli can be obtained. For composites with spherical reinforcements the predictions for the eective bulk modulus K obtained with such a procedure correspond to the CSA results, and for composites reinforced by unidirectional continuous bers the CCA results for the eective transverse bulk modulus KT , the eective axial shear modulus Glq and the eective axial Youngs modulus El are recovered. Beyond these results, in the case of a matrix reinforced with spherical inhomogeneities (which shows statistically isotropic overall behavior and is a very useful model for composites reinforced by approximately equiaxed particles) the following quadratic equation is obtained for the eective shear modulus GGSC GGSC B+D =0 A+ (m) G G(m) with 7 10 A = 8[G(i) /G(m) 1](4 5 (m) )1 3 2[63(G(i) /G(m) 1)2 + 21 3 ] 3 +252[G(i) /G(m) 1]2 3 25[G(i) /G(m) 1](7 12 (m) + 8 (m) )2 +4(7 10 (m) )2 3 B = 4[G(i) /G(m) 1](1 5 (m) )1
5 10 3 5

(3.52)

+ 4[63(G(i) /G(m) 1)2 + 21 3 ] 3

504[G(i) /G(m) 1]2 3 + 150[G(i) /G(m) 1](3 (m) ) (m) 2 +3(15 (m) 7)2 3 D = 4[G(i) /G(m) 1](5 (m) 7)1
5 10 3

2[63(G(i) /G(m) 1)2 + 21 3 ] 3

2

1 2 3

+252[G(i) /G(m) 1]2 3 + 25[G(i) /G(m) 1]( (m) 7)2 (7 + 5 (m) )2 3 = [G(i) /G(m) 1](7 10 (m) )(7 + 5 (i) ) + 105( (i) (m) ) = [G(i) /G(m) 1](7 + 5 (i) ) + 35(1 (i) ) = [G(i) /G(m) 1](8 10 (m) ) + 15(1 (m) ) .

The analogous expression for the overall transverse shear modulus of composites reinforced by aligned continuous bers (which show statistically transversely isotropic overall behavior) takes the form GGSC,qt GGSC,qt B+D =0 A+ (m) G G(m) with A = 3 (1 )2 [8[G(f) /G(m) 1][G(f) /G(m) + f ] +[(G(f) /G(m) )m + f m ((G(f) /G(m) )m f ) 3 ] [m (G(f) /G(m) 1) ((G(f) /G(m) )m + 1)] B = 6 (1 )2[G(f) /G(m) 1][G(f) /G(m) + f ] +[(G(f) /G(m) )m + (G(f) /G(m) 1) + 1] [(m 1)(G(f) /G(m) + f ) 2 3 ((G(f) /G(m) )m f )] + (m + 1)[G(f) /G(m) 1][G(f) /G(m) + f + ((G(f) /G(m) )m f ) 3 ] 27
2

(3.53)

D = 3 (1 )2 [G(f) /G(m) 1][G(f) /G(m) + f ] +[(G(f) /G(m) )m + (G(f) /G(m) 1) + 1] [G(f) /G(m) + f + ((G(f) /G(m) )m f ) 3 ] m = 3 4 (m) f = 3 4 (f) . Generalized self-consistent schemes for general aligned spheroidal inhomogeneities can be derived from energy considerations (Huang and Hu, 1995), the resulting expressions being again implicit and somewhat unhandy. For standard composites, i.e., materials consisting of sti inhomogeneities embedded in a soft matrix, some of the eective elastic moduli (K , KT ) obtained from the above scheme coincide with the corresponding MoriTanaka results, whereas others (e.g., E and G) are higher, compare gs. 4.1 to 4.7.

3.6

Other Mean Field and Related Estimates

The most important mean eld approaches besides eective eld and self-consistent methods are dierential schemes (McLaughlin, 1977; Norris, 1985). These eective medium theories may be envisaged as involving repeated cycles of adding small concentrations of inhomogeneities to a material and then homogenizing. Following Hashin (1988) the overall elastic tensors can then be described by the dierential equations 1 dED dil = (E(i) ED )A d 1 dCD 1 dil = (C(i) CD )B d 1

(3.54)

with the initial conditions ED =E(m) and CD =C(m) , respectively, at =0. In analogy to dil and B dil depend on ED . In general, eqns. (3.54) must be integrated nueqn. (3.50) A merically, e.g., by RungeKutta schemes. Dierential schemes describe matrixinclusion microgeometries with polydisperse distributions of the sizes of the inhomogeneities (corresponding to the repeated homogenization steps. Torquato (1991) developed second order estimates for the thermoelastic properties of two-phase materials, which correspond to the same phase arrangements and use the same three-point microstructural parameters and as discussed for improved bounds in section 4.2. These estimates lie between the corresponding bounds and give the best analytical predictions currently available for the overall thermoelastic response of inhomogeneous materials that show the appropriate microstructures and are free of aws.

3.7

Mean Field Methods for Elastoplastic Composites

The MoriTanaka theories discussed in section 3.4 can be extended in a relatively straightforward way to describing inhomogeneous materials that contain at least one elastoplastic or thermoelastoplastic constituent.

28

In the literature two main lines of development of mean eld approaches to modeling elastoplastic inhomogeneous materials can be found. On the one hand, secant plasticity (deformation theory) concepts have been used, see, e.g., Tandon and Weng (1988) or Dunn and Ledbetter (1997), and, on the other hand, incremental plasticity models have been developed, compare, e.g., Lagoudas et al. (1991), Karayaka and Sehitoglu (1993) or Pettermann et al. (1999). Whereas methods based on secant plasticity are limited to radial (or approximately radial) loading paths of the constituents in stress space8 and, accordingly, are not suitable for use as micromechanically based constitutive models, incremental plasticity approaches are not subject to limitations in this respect. However, especially for matrix dominated deformation modes, the secant formulations are typically superior to mean eld algorithms based on incremental plasticity in the post-yield regime, where the latter algorithms show a tendency towards overestimating overall strain hardening. This problem has been an active eld of research, compare Suquet (1997), and, for statistically isotropic composites, recent algorithmic modications have led to marked improvements in this respect (Doghri and Ouaar, 2003). Incremental mean eld methods can be formulated on the basis of strain and stress rate tensors for elastoplastic phases (p) , d (p) and d (p) , which can be expressed in analogy to eqn. (3.5) as d d
(p) (p) (p) (p) t dT = A t d + (p) (p) = B t d + t dT

(3.55)

where d stands for the macroscopic strain rate tensor, d for the macroscopic stress (p) (p) (p) (p) t , B rate tensor, and dT for a homogeneous temperature rate. A t , t , and t are instantaneous strain and stress concentration tensors, respectively. Assuming the inhomogeneities to show elastic and the matrix to show elastoplastic material behavior9 , the overall instantaneous (tangent) stiness tensor can be written in terms of the phase properties and the instantaneous concentration tensors as
(m) (m) Et = E(i) + (1 )[Et E(i) ]A t (m) (i) (i) (m) 1 = [C + (1 )[Ct C ]Bt ]

(3.56)

compare eqns. (3.7) and (3.8). Using the MoriTanaka formalism of Benveniste (1987), the instantaneous matrix concentration tensors take the form
(m) (m) (m) A = {(1 )I + [I + St Ct (E(i) Et )]1 }1 t (m) (m) (m) (m) B = {(1 )I + E(i) [I + St Ct (E(i) Et )]1 Ct }1 t

(3.57)

compare eqn. (3.45). Expressions for the instantaneous thermal concentration tensors and instantaneous coecients of thermal expansion can be derived in analogy to the correIn deformation plasticity elastoplastic behavior is approximated by nonlinear elastic behavior. As a consequence such descriptions are restricted to monotonous loading and radial loading paths in stress space at the constituent level, a condition that is typically violated to some extent in the phases of elastoplastic inhomogeneous materials, even if the overall loading paths are perfectly radial, see, e.g., Pettermann (1997). This eect is due to changes in the accommodation of the phase stresses and strains in inhomogeneous materials upon yielding of a constituent. 9 Analogous expressions can be derived for elastoplastic inhomogeneities in an elastic matrix or, in general, for composites containing any required number of elastoplastic phases.
8

29

sponding thermoelastic relations. It should be noted that eqns. (3.57) use the instantaneous Eshelby tensor St , which depends on the current state of the matrix material and in general has to be evaluated numerically. Formulations of eqns. (3.55) to (3.57) that are directly suitable for implementation as micromechanically based constitutive models at the integration point level within Finite Element codes can be obtained by replacing rates such as d (p) with the corresponding nite increments, e.g., (p) . It is worth noting that in the resulting incremental Mori Tanaka methods no assumptions on the overall yield loci and the overall ow potential are made, the eective material behavior being entirely determined by the incremental mean eld equations and the constitutive behavior of the phases. As a consequence, mapping of the stresses onto the yield surface cannot be handled at the level of the homogenized material and radial return mapping must be applied to the matrix at the microscale instead. Accordingly, the constitutive equations describing the overall behavior cannot be integrated directly (as is the case for homogeneous elastoplastic materials), and iterative algorithms are required. Extended versions of such algorithms can also handle thermal expansion eects10 and temperature dependent material parameters. Mean eld based descriptions of elastoplastic inhomogeneous materials tend to overestimate the macroscopic ow stress of composites because, in general, the phase average of the mean equivalent stress in an elastoplastic phase is greater than the equivalent stress generated from the corresponding phase averages of the stress components, i.e., the relation 2 e
(p)

3 T s s 2

(p)

3 T s 2

(p)

(p)

(3.58)

holds, where s stands for the deviatoric part of the stress tensor11 . For isotropic overall behavior and isotropic constituent properties this diculty can be mitigated by evaluating the mean von Mises equivalent stresses from energy considerations (Qiu and Weng, 1992) or by using statistically based theories (Buryachenko,1996). Furthermore, because they assume elastoplastic phases to yield as a whole once the corresponding equivalent stress (p) e exceeds the yield stress, mean eld approaches predict sharp transitions from elastic to plastic states instead of the actual, gradual progress of yielded regions at the microscale. Incremental MoriTanaka methods, especially when they incorporate the improvements introduced, e.g., by Doghri and Ouaar (2003), oer a combination of useful accuracy, exibility in terms of inhomogeneity geometries, and relatively low computational requirements. Self-consistent schemes can also be combined with secant and incremental approaches to describing elastoplastic inhomogeneous materials, see, e.g., Hill (1965), Hutchinson (1970) or Berveiller and Zaoui (1981). These procedures (as well as their weaknesses and strengths) are closely related to those discussed above for the MoriTanaka method.
Elastoplastic inhomogeneous materials such as metal matrix composites show a hysteretic thermal expansion response, i.e., there are no thermal expansion coecients in the strict sense. 11 Evaluating the von Mises equivalent stresses from the phase averaged stress components also leads to predictions by standard mean eld methods that materials with spherical reinforcements will not yield under overall hydrostatic loads and homogeneous temperature loads. This is in contradiction to experimental and other theoretical results.
10

30

In addition, mean eld approaches have been employed to obtain estimates for the nonlinear response of inhomogeneous materials due to microscopic damage or to combinations of damage and plasticity, see, e.g., Tohgo and Chou (1996) or Guo et al. (1997). As an alternative to directly extending mean eld theories into secant or incremental plasticity, they can also be combined with Dvoraks transformation eld analysis (Dvorak, 1992) in order to obtain descriptions of the overall behavior of inhomogeneous materials in the plastic range, see, e.g., Plankensteiner (2000). Such approaches are well suited for being used as material models in FE codes, are very attractive in terms of computational requirements, but tend to strongly overestimate the overall strain hardening of elastoplastic composites because they use elastic accommodation of microstresses and strains throughout the loading history.

3.8

Mean Field Methods for Nonaligned Composites

The statistics of the microgeometries of nonaligned and hybrid matrixinclusion composites can be typically described on the basis of orientation distribution functions (ODFs) and aspect ratio or length distribution functions (LDFs) of the reinforcements. MoriTanaka methods have been extended to describing the elastic behavior of such (i)(, , ) for materials, the typical starting point being dilute concentration tensors B dil reinforcements having orientations with respect to the global coordinate system that are described by the Euler angles , and . A phase averaged dilute ber stress concentration tensor can be obtained by orientational averaging as Bdil =
(i) 2 0 0 2 0

(i)(, , ) (, , ) d d d B dil

(3.59)

where the orientation distribution function (, , ) is normalized to give = 1. The core statement of the MoriTanaka approach, eqn. (3.40), can then be written in the form
(i)

= Bdil

(i)

(m)

= Bdil BM = BM
(m) (i)

(m)

(i)

(3.60)

which allows the phase concentration tensors BM and BM to be evaluated in analogy to eqns. (3.43) to (3.45). The stress (i) obtained from eqn. (3.60) is an average over all inhomogeneities, irrespective of their orientation, and, accordingly, contains only limited information. The average stresses in inhomogeneities of a given orientation (, , ), which (i) B(m) . is of considerably more interest, can be evaluated as (i) = B M dil Extended MoriTanaka methods for modeling the elastic behavior of microstructures that contain nonaligned inhomogeneities were developed by a number of authors, see, e.g., Benveniste (1990), Dunn and Ledbetter (1997), Pettermann et al. (1997) or Mlekusch (1999). These models dier mainly in the algorithms employed for orientational or congurational averaging. Methods of this type tend to be very useful for many practical applications. They are, however, of a somewhat ad-hoc nature and can lead to non-symmetric 31

eective stiness tensors12 in certain situations (Benveniste et al., 1991; Ferrari, 1991). The latter unphysical behavior also appears in multi-phase MoriTanaka models in which aligned inhomogeneities show both dierent material behaviors and dierent aspect ratios. For the special case of randomly oriented bers or platelets of a given aspect ratio, symmetrized dilute strain concentration tensors may be constructed, giving rise to the so-called Wu tensors (Wu, 1966). They can be combined with MoriTanaka methods (Benveniste, 1987) or classical self-consistent schemes (Berryman, 1980) to describe composites with randomly oriented phases of the matrixinclusion and interwoven or granular types, respectively. Due to the overall isotropic behavior of composites reinforced by randomly oriented bers or platelets, their elastic behavior must comply with the standard Hashin Shtrikman bounds (Hashin and Shtrikman, 1963), compare section 4.1. A simpler approach to describing the macroscopic behavior of composites with nonaligned reinforcements is the use of laminate analogies. In such models the composite is approximated as a laminate consisting of unidirectionally reinforced layers that are appropriately oriented for handling one ber orientation each (Fu and Lauke, 1998). The ber orientation distributions are, accordingly, approximated by orientational averaging at the level of the elastic tensors. Within such a laminate analogy approach mean eld expressions can be used for the individual layers to handle ber aspect ratios. The resulting schemes can be useful especially for describing composites with planar random ber orientations (Huang, 2001).

3.9

Mean Field Methods for Diusion-Type Problems

There are strong conceptual and algorithmic analogies between inhomogeneity problems describing elastostatic behavior and diusive transport processes, see, e.g., Hashin (1983). Specically, the eects of dilute inhomogeneous inclusions in diusion-type problems (e.g., heat conduction, electrical conduction, diusion of moisture) as well as in some coupled problems (e.g., inhomogeneous materials with at least one piezoelectric constituent) can be described in analogy to eqns. (3.18) to (3.39) via a depolarization or diusion Eshelby tensor13 . For non-dilute cases, analoga of the MoriTanaka, self-consistent and dierential schemes can be derived, see, e.g., Hatta and Taya (1986), Miloh and Benveniste (1988), Dunn and Taya (1993), Chen (1997) as well as Torquato (2002).

The reason for these diculties lies in the aligned ellipsoidal symmetry of the phase arrangement that is inherent to MoriTanaka methods. 13 In contrast to the mechanical Eshelby tensor, which is of order 4, the depolarization tensor is of order 2, and for the case of spheroidal inhomogeneities embedded in an isotropic matrix depends only on the formers aspect ratio. Explicit expressions for the elements of depolarization tensors are given, e.g., by Hatta and Taya (1986).

12

32

Chapter 4 Bounding Methods

Whereas mean eld methods, unit cell approaches, embedding strategies and windowing models can typically be used for both homogenization and localization tasks, bounding methods are restricted to homogenization. Discussions in this section again are restricted to materials consisting of two perfectly bonded constituents. Accordingly, they do not account for any aws (e.g., porosity) that may degrade the macroscopic behavior of actual composites Rigorous bounds for the overall elastic properties of inhomogeneous materials can be obtained from appropriate variational (minimum energy) principles. In the following, only outlines of bounding methods are given; formal treatments were given, e.g., by NematNasser and Hori (1993), Ponte Casta neda and Suquet (1998), Markov (2000), Bornert et al. (2001), Milton (2002) as well as Torquato (2002).

4.1

Hill and HashinShtrikman-Type Bounds

Classical expressions for the minimum potential energy and the minimum complementary energy in combination with uniform stress and strain trial functions lead to the simplest variational bounding expressions, the upper bounds of Voigt (1889) and the lower bounds of Reuss (1929). In their tensorial form they are known as the Hill bounds and can be written as 1 V (p) C(p) E V (p) E(p) . (4.1)
(p) (p)

These bounds, while universal and very simple, do not contain any information on the microgeometry of an inhomogeneous material beyond the phase volume fractions, and are typically too slack to be of much practical use14 , but in contrast to the HashinShtrikman and higher order bounds they also hold for (non-pathological) control volumes that are too small to be proper RVEs.
The bounds on the Youngs moduli obtained from eqn. (4.1) are equivalent to Voigt and Reuss expressions in terms of the corresponding phase moduli (compare section 2.1) only if the constituents have equal Poissons ratios. Due to the homogeneous stress and strain assumptions used for obtaining the Hill bounds, (p) =I and B (p) =I, respectively. the corresponding phase strain and stress concentration tensors are A V R
14

33

Much tighter bounds can be obtained from variational expressions due to Hashin and Shtrikman (1962), which are based on the principle of minimum potential energy. They are formulated in terms of a homogeneous reference material and of polarization elds (x) that describe the dierence between the microscopic stress elds in the inhomogeneous material (elasticity tensor E(x)) and the reference material (elasticity tensor E0 ), = E = E0 + (x) with (x) = (E(x) E0 ) (x) . (4.2)

Exact polarization elds are highly heterogeneous, but simplied approximations may be employed for constructing trial elds for use in variational theorems. By choosing phasewise constant polarizations and optimizing them, HashinShtrikman-type bounds are obtained. The HashinShtrikman bounds proper (Hashin and Shtrikman, 1963) apply to inhomogeneous materials with statistically isotropic overall symmetry and are given in terms of bounds on the eective bulk and shear moduli. For two-phase materials that fulll the condition (K (m) K (i) )(G(m) G(i) ) > 0 the bounds on the eective moduli take the form
L.B. KHS = K (m) + U . B. KHS . B. GL HS . B. GU HS

+ 3(1 )/(3K (m) + 4G(m) ) 1 = K (i) + (m) (i) 1/(K K ) + 3/(3K (i) + 4G(i) ) = G(m) + (i) (m) (m) 1/(G G ) + 6(1 )(K + 2G(m) )/[5G(m) (3K (m) + 4G(m) )] 1 . (4.3) = G(i) + (m) (i) (i) 1/(G G ) + 6 (K + 2G(i) )/[5G(i) (3K (i) + 4G(i) )] 1/(K (i) K (m) )

For obtaining bounds on the eective Youngs moduli from these data see Hashin (1983), and for bounding expressions on the Poisson numbers see Zimmerman (1992). Another set of HashinShtrikman bounds were developed for statistically transversely isotropic composites reinforced by aligned continuous bers15 . For the overall moduli KT , El , lq , and Glq they are equivalent to the corresponding CCA expressions16 , see eqn. (2.7). For the transverse shear modulus Gqt the expressions
. B. (m) GL + qt HS = G

(f) 1/(Gqt

G(m) ) + (1

. B. (m) 1+ GU qt HS = G (f)

(1 + )
(m) (f)

(m) )(KT (m) 2

+ 2G(m) )/2G(m) (KT + G(m) )

(m)

[1 + 3(1 )2 (m) /( 3 + 1)] Gqt > G(m)

for KT > KT ,
15

These HashinShtrikman bounds apply to transversely isotropic two-phase materials in which the phases are continuous in the axial direction. 16 For KT , El and Glq the CCA expressions are lower bounds and for lq they may be either bound. In order to obtain the corresponding other bounds the roles of inhomogeneities and matrix have to be exchanged in eqn. (2.7).

34

. B. GL qt HS

= G

(m)

1+

(1 + (m) ) [1 + 3(1 )2 (m) /( 3 (m) )]

(f) (m) (m) (m) 2

. B. (m) + GU qt HS = G

1/(Gqt G(m) ) + (1 )(KT + 2G(m) )/2G(m) (KT + G(m) )

(f)

for KT < KT , hold, where

Gqt < G(m)

(f)

(4.4)

= ( (m) (f) )/(1 + (f) ) = ( + (m) )/( 1) (f) = KT /(KT + 2Gqt ) (m) = 1/(3 4 (m) ) = Gqt /G(m) Further developments led to the Willis (1977) bounds, which pertain to aligned ellipsoidal microgeometries, compare g. 3.1 (and thus to transversely isotropic overall behavior). They include the HashinShtrikman bounds for isotropic and continuously reinforced composites as special cases. From the practical point of view it is of interest that for two-phase materials MoriTanaka estimates correspond to one of the Willis bounds, compare Weng (1990), and the other bound can be obtained after a color inversion (i.e., after exchanging the roles of inhomogeneities and matrix)17 . Accordingly, the bounds can be evaluated for fairly general phase geometries by matrix algebra such as eqns. (3.46) and (3.48). Hashin Shtrikman-type bounds can also be derived for simple periodic phase arrangements, see, e.g., Nemat-Nasser and Hori (1993).
(f) (f) (f) (f)

4.2

Improved Bounds

When more complex trial functions are used in variational bounding, their optimization requires statistical information on the phase arrangement in the form of n-point correlation functions18 . This way improved bounds can be generated that are signicantly tighter than HashinShtrikman-type expressions (which in this context can be classied as two-point bounds). Three-point bounds for statistically isotropic two-phase materials can be formulated in such a way that the information on the phase arrangement statistics is contained in two three-point microstructural parameters, ( ) and ( ). In this framework, the Beran and Molyneux (1966) bounds on the eective bulk modulus of macroscopically isotropic
Put more precisely, the lower Willis bounds are obtained from MoriTanaka methods when the more compliant material is used as matrix phase, and the upper bounds when it is used as inclusion phase. 18 For discussions on statistical descriptions of phase arrangements see, e.g., Torquato (1991), Pyrz and Bochenek (1998) as well as Torquato (2002).
17

35

materials can be written as

L.B. K3P

4 (1 )(1/K (i) 1/K (m) )2 1/K + 3 1/G 4 1/K K 3 (1 )(K (i) K (m) )2 +4 G 3K

U . B. K3P =

(4.5)

and the Phan-Thien and Milton (1983) bounds for the overall shear modulus take the form
. B. GL = 3P

G G

6 (1 )(1/G(i) 1/G(m) )2 + 1 6G 6 (1 )(1/G(i) 1/G(m) )2 + 6G (4.6)

. B. GU = 3P

with 6 2 1 1 1 21 + + G K G G K G 2 3 G 6 K + 7G 5 G = 2 K G +5 G = 5 f f f f

128 99 + K G

+45

1 G

= (1 )f (m) + f (i) = f (m) + (1 )f (m) = (1 )f (m) + f (i) = (1 )f (m) + f (i)

. (4.7)

( ) and ( ) can in principle be evaluated for any given microstructure. Analytical expressions or tabulated data in terms of the reinforcement volume fraction are available for a number of generic microgeometries of practical importance, among them statistically homogeneous isotropic materials containing identical, bidisperse and polydisperse impenetrable (hard) spheres (describing matrixinclusion composites) as well as monodisperse interpenetrating spheres (Boolean models that can describe many interpenetrating phase arrangements), and statistically homogeneous transversely isotropic materials reinforced by impenetrable or interpenetrating aligned cylinders. References to some three-point bounds and to a number of expressions for ( ) and ( ) applicable to matrixinclusion composites can be found in section 4.4, where results from mean eld and bounding approaches are compared. For a review of higher order bounds for elastic (as well as other) properties of inhomogeneous materials see Torquato (1991).

4.3

Bounds for the Nonlinear Behavior

Analoga to the Hill bounds for nonlinear inhomogeneous materials were introduced by Bishop and Hill (1951). For polycrystals the nonlinear equivalents to Voigt and Reuss 36

expressions are usually referred to as Taylor and Sachs bounds, respectively. In analogy to mean eld estimates for elastoplastic material behavior, nonlinear bounds are typically obtained by evaluating a sequence of linear bounds. Talbot and Willis (1985) extended the HashinShtrikman variational principles to obtain one-sided bounds (i.e., upper or lower bounds, depending on the material combination) on the nonlinear mechanical behavior of inhomogeneous materials. An important development was the derivation by Ponte Casta neda (1992) of a variational principle that allows upper bounds on specic stress-strain responses of elastoplastic inhomogeneous materials to be generated on the basis of upper bounds on the elastic tensors19 by using a series of inhomogeneous reference materials, the properties of which have to be obtained by optimization procedures for each strain level. Essentially, the variational principle guarantees the best choice for the comparison material at a given load. The Ponte Casta neda bounds are closely related to mean eld approaches using improved secant plasticity methods (compare section 3.7). The study of bounds like the development of improved estimates for the overall nonlinear mechanical behavior of inhomogeneous materials has been an active eld of research during the past decades, see the recent reviews by Suquet (1997), Ponte Casta neda and Suquet (1998) as well as Willis (2000).

4.4

Comparisons Between Some Mean Field and Bounding Predictions

In order to give some idea of the type of predictions that can be obtained by dierent mean eld (and related) approaches and by bounding methods for the thermomechanical responses of inhomogeneous thermoelastic materials, results for the overall elastic moduli and coecients of thermal expansion are presented as functions of the reinforcement volume fraction in this section. All comparisons are based on E-glass particles or bers embedded in an epoxy matrix. The elastic contrast of these constituents is c 21 and the thermal expansion contrast takes a value of approximately 0.14, see table 4.1.

Table 4.1: Constituent material parameters of epoxy matrix and E-glass reinforcements used in generating gs. 4.1 to 4.7. matrix reinforcements E [GPa] 3.5 74.0 [ ] [1/K] 0.35 36.0 0.2 4.9

Figures 4.1 and 4.2 show predictions for the overall Youngs and shear moduli of a particle reinforced composite using the above material parameters. The MoriTanaka results (MTM) as mentioned before coincide with the lower HashinShtrikman (H/S)
The Ponte Casta neda bounds are rigorous for nonlinear elastic inhomogeneous materials and, on the basis of deformation theory, are very good approximations for materials with at least one elastoplastic constituent. Applying the Ponte Casta neda variational procedure to elastic lower bounds does not necessarily lead to a lower bound for the inelastic behavior.
19

37

EFFECTIVE YOUNGs MODULUS [GPa]

80.0

DS CSCS 3PE (mono/h) GSCS 3PUB (mono/h) 3PLB (mono/h) H/S UB H/S LB; MTM Hill bounds

0.0
0.0

40.0

0.2

0.4

0.6

0.8

1.0

PARTICLE VOLUME FRACTION [ ]

Figure 4.1: Bounds and estimates for the eective Youngs moduli of glass/epoxy particle reinforced composites as functions of the particle volume fraction.

EFFECTIVE SHEAR MODULUS [GPa]

DS CSCS 3PE (mono/h) GSCS 3PUB (mono/h) 3PLB (mono/h) H/S UB H/S LB; MTM Hill bounds

0.0
0.0

10.0

20.0

30.0

0.2

0.4

0.6

0.8

1.0

PARTICLE VOLUME FRACTION [ ]

Figure 4.2: Bounds and estimates for the eective shear moduli of glass/epoxy particle reinforced composites as functions of the particle volume fraction. bounds, whereas the classical self-consistent scheme (CSCS) shows a typical behavior in that it is close to one HashinShtrikman bound at low volume fractions, approaches the other at high volume fractions, and displays a transition behavior in the form of a sigmoid curve in-between. The three-point bounds (3PLB and 3PUB) shown in gs. 4.1 and 4.2 are

38

based on formalisms developed by Beran and Molyneux (1966), Milton (1981) as well as Phan-Thien and Milton (1983). Like the three-point estimates (3PE) they were evaluated for impenetrable spherical particles of equal size and use expressions for the statistical parameters ( ) and ( ) that were reported by Miller and Torquato (1990) and Torquato et al. (1987), respectively. These expressions are available for reinforcement volume fractions up to =0.6. The results from the generalized self-consistent scheme (GSCS) lie between the MoriTanaka predictions and the three-point estimates, whereas the dierential scheme (DS) provides the stiest predictions except the upper bounds. Predictions for the coecients of thermal expansion of statistically isotropic inhomogeneous materials are presented in g. 4.3, the bounds being obtained by evaluating a version of Levins formula, eqn.(3.11), with bounds on the eective bulk modulus. Here the HashinShtrikman bounds and the three-point bounds are used as the basis of this procedure. The CTEs corresponding to the self-consistent schemes were obtained with the mean eld formalism discussed in section 3.1. Because they are based on the same estimates for the eective bulk modulus, the results given for the GSCS and the Mori Tanaka-scheme agree with the upper Levin/HashinShtrikman bounds.

40.0

EFFECTIVE CTE [1/K x 10 -6 ]

0.0

20.0

DS CSCS 3PE (mono/h) GSCS 3PUB (mono/h) 3PLB (mono/h) H/S UB; MTM H/S LB

0.0

0.2

0.4

0.6

0.8

1.0

PARTICLE VOLUME FRACTION [ ]

Figure 4.3: Bounds and estimates for the eective CTEs of glass/epoxy particle reinforced composites as functions of the particle volume fraction. In gs. 4.3 to 4.7 results are presented for the overall transverse Youngs moduli, overall axial and transverse shear moduli, as well as overall axial and transverse coecients of thermal expansion of composites reinforced by aligned continuous bers20 . The results for the three-point bounds shown follow the formalism of Silnutzer (1972) and Milton (1981), correspond to the case of aligned impenetrable circular cylinders of equal diameter and
In the case of the axial Youngs moduli the estimates and bounds are nearly indistinguishable from each other and from the rule of mixture result, eqn.(2.2), for the material parameters and scaling used in g. 4.4 and are, accordingly, not given here.
20

39

EFFECTIVE YOUNGs MODULUS [GPa]

80.0

DS CSCS 3PE (mono/h) GSCS 3PB LB (mono/h) 3PB LB (mono/h) H/S UB H/S LB; MTM Hill bounds

0.0
0.0

40.0

0.2

0.4

0.6

0.8

1.0

FIBER VOLUME FRACTION [ ]

Figure 4.4: Bounds and estimates for the eective transverse Youngs moduli of glass/epoxy ber reinforced composites as functions of the ber volume fraction.

EFFECTIVE SHEAR MODULUS [GPa]

DS CSCS 3PE (mono/h) GSCS 3PUB (mono/h) 3PLB (mono/h) H/S UB H/S LB; MTM Hill bounds

0.0
0.0

10.0

20.0

30.0

0.2

0.4

0.6

0.8

1.0

FIBER VOLUME FRACTION [ ]

Figure 4.5: Bounds and estimates for the eective axial shear moduli of glass/epoxy ber reinforced composites as functions of the ber volume fraction. use statistical parameters evaluated by Torquato and Lado (1992) for ber volume fractions 0.7. Generally, a qualitatively similar behavior to the particle reinforced case is evident. It is noteworthy, however, that the overall transverse CTEs in g. 4.7 at low ber volume fractions exceeds the CTEs of both constituents. Such behavior is typical for

40

EFFECTIVE SHEAR MODULUS [GPa]

DS CSCS 3PE (mono/h) GSCS 3PUB (mono/h) 3PLB (mono/h) H/S UB H/S LB; MTM Hill bounds

0.0
0.0

10.0

20.0

30.0

0.2

0.4

0.6

0.8

1.0

FIBER VOLUME FRACTION [ ]

Figure 4.6: Bounds and estimates for the eective transverse shear moduli of glass/epoxy ber reinforced composites as functions of the ber volume fraction.

40.0

EFFECTIVE CTE [1/K x 10 -6 ]

DS tr. DS ax. CSCS tr. CSCS ax. 3PE tr. (mono/h) MTM, GSCS tr. MTM, GSCS ax.

0.0
0.0

20.0

0.2

0.4

0.6

0.8

1.0

FIBER VOLUME FRACTION [ ]

Figure 4.7: Estimates for the eective axial and transverse CTEs of glass/epoxy ber reinforced composites as functions of the ber volume fraction. continuously reinforced composites and is caused by the marked axial constraint enforced by the bers. In gs. 4.1 to 4.7 the classical self-consistent scheme as expected is not in good agreement with the three-point bounds shown, which explicitly correspond to matrix 41

inclusion topologies. Considerably better agreement with the CSCS is found by using three-point parameters of the interpenetrating sphere or cylinder type (which can also describe cases where both phases percolate, but are not as symmetrical with respect to the constituents as the CSCS). From a practical point of view it is worth noting that despite their sophistication improved bounds (and higher order estimates) may give overly optimistic predictions for the overall moduli because they describe ideal composites, whereas in actual materials it is practically impossible to avoid aws such as porosity.

42

Chapter 5 General Remarks on Modeling Approaches Based on Discrete Microgeometries

In the present context, micromechanical approaches based on discrete microgeometries encompass unit cell, embedding and windowing methods. Broadly speaking, these approaches trade o restrictions to the generality of the microstructures that can be studied for the capabilities of using ne grained geometrical models and of resolving details of the stress and strain elds at the length scale of the inhomogeneities. The main elds of application of these methods are studying the nonlinear behavior of inhomogeneous materials and evaluating the microelds at high resolution. Even though in many cases the latter type of information is not necessary for describing the macroscopic behavior of inhomogeneous materials, their damage and failure behavior can depend on details of the microgeometry. There are two, often complementary, philosophies for modeling inhomogeneous materials via discrete microgeometries. One of them is based on studying generic phase arrangements, which may range from simple periodic arrays of bers or particles to highly complex microgeometries involving a considerable number of reinforcements, the positions of which are chosen to approximate some statistical phase distribution. Complex phase arrangements of the latter type can be generated by appropriate statistically based algorithms such as random insertion methods or numerical annealing procedures, which give rise to quasi-random and statistically reconstructed microstructures21 , respectively. Such microgeometries tend to employ idealized inclusion shapes, equiaxed particles embedded in a matrix, for example, being represented by spheres22 . Alternatively, microgeometries may be chosen to follow as closely as possible the phase arrangement of a given sample of the material to be modeled, obtained, e.g., from metAs understood here, statistically reconstructed microstructures are equivalent to real ones in a statistical sense, i.e., they show equal or very similar phase distribution statistics. For algorithms for reconstructing matrixinclusion and more general microgeometries see, e.g., Rintoul and Torquato (1997) and Torquato (1998). 22 For uniform boundary conditions it can be shown that the overall elastic behavior of matrixinclusiontype composites can be bounded by approximating the actual shape of particles by inner and outer envelopes of smooth (e.g., ellipsoidal) shape. This is known as the Hill modication theorem (Hill, 1963; Huet et al., 1990).
21

43

allographic sections, serial sections or tomographic data (Terada and Kikuchi, 1996; Li et al., 1999; Babout et al., 2004; Chawla and Chawla, 2006). The resulting descriptions are termed real structure models. Unless simple periodic phase arrangements are considered, for both of the above approaches to obtaining microgeometries the question immediately arises as to how complex (and thus large) the model geometry must be in order to adequately capture the physical behavior of the material to be studied. For the case of elastic statistically isotropic composites with matrixinclusion topology and sphere-like particles, Drugan and Willis (1996) estimated that for approximating the overall moduli with errors of less than 5% or less than 1%, respectively, volume elements with sizes of roughly two or ve inclusion diameters are sucient for any volume fraction23 . Alternatively, adequate sizes of model geometries for elastic studies may be estimated on the basis of experimentally obtained correlation lengths (Bulsara et al., 1999; Jeulin, 2001), by comparing statistical distribution functions of actual and model microgeometries (Zeman and Sejnoha, 2001), or by using windowing approaches (compare section 8) to bounding the overall response from above and below. For nonlinear matrix behavior a number of numerical studies (Zohdi, 1999; Jiang et al., 2001; B ohm and Han, 2001) have indicated that substantially larger volume elements may be necessary for satisfactorily approximating the required overall symmetry and for obtaining good agreement between the responses of (nominally) statistically equivalent phase arrangements, especially at elevated overall inelastic strains. This indicates that the size of satisfactory multi-inclusion unit cells depends markedly on the phase material behavior24 . The majority of published micromechanical studies of discrete microstructures have employed standard numerical engineering methods for resolving the microelds, studies using Finite Dierence (FD) algorithms, compare Adams and Doner (1967), spring lattice models, compare Ostoja-Starzewski (1996), the Boundary Element Method (BEM), compare Achenbach and Zhu (1989), the Finite Element Method (FEM), techniques using Fast Fourier Transforms (FFT), compare Moulinec and Suquet (1994), and Discrete Fourier Transforms (DFT), compare M uller (1996), as well as FE-based discrete dislocation models (Cleveringa et al., 1997) having been reported. A number of more specialized approaches are discussed in connection with periodic microeld analysis, see section 6. Generally speaking, spring lattice models tend to have advantages in handling pure traction boundary conditions and in modeling the progress of microcracks due to local (brittle) failure. Boundary elements tend to be at their best in studying geometrically complex linear elastic problems. For all the above methods the characteristic length of the discretization (mesh size) must be considerably smaller than the microscale of a given problem in order to obtain spatially well resolved results. At present, the FEM is the most popular numerical scheme for evaluating discrete microgeometries, especially in the nonlinear range, where its exibility and capability of
23 Note that a volume element fullling a condition of this type is not necessarily an RVE as discussed in section 1.2. 24 For elastoplastic matrices, weaker strain hardening translates into unit cells requiring a higher number of inhomogeneities, and very large volume elements may be necessary when the matrix shows strain softening and/or localization.

44

supporting a wide range of constitutive models for the constituents and for the interfaces between them are especially appreciated. An additional asset of the FEM in the context of continuum micromechanics is its ability to handle discontinuities in the stress and strain components (which typically occur at interfaces between dierent constituents) in a natural way via appropriately placed element boundaries. Applications of the FEM to micromechanical studies tend to fall into four main groups, compare g. 5.1. In most published works the phase arrangements are discretized by an often high number of standard continuum elements, the mesh being designed in such a way that element boundaries (and, where appropriate, special interface elements) are positioned at all interfaces between constituents. Such an approach has the advantage that in principle any microgeometry can be handled and that readily available commercial FE packages may be used. However, the actual modeling of complex phase congurations in many cases requires sophisticated and/or specialized preprocessors for generating the mesh, a task that has been dicult to automate, and the resulting stiness matrices may show unfavorable conditioning due to suboptimal element shapes. Alternatively, a smaller number of special hybrid elements may be used, which are specically formulated to model the deformation, stress, and strain elds in an inhomogeneous region consisting of a single inhomogeneity or void together with the surrounding matrix on the basis of some appropriate analytical theory. The most highly developed approach of this type at present is the Voronoi Finite Element Method (Ghosh et al., 1996), in which the mesh for the hybrid elements is obtained by Voronoi tessellation based on the positions of the reinforcements. Large planar multi-inclusion arrangements can be analyzed this way using a limited number of (albeit rather complex) elements, and good accuracy as well as signicant gains in eciency have been claimed. Computational strategies of this type are specically tailored for inhomogeneous materials with matrixinclusion topologies.
a) b) c) d)

Figure 5.1: Sketch of FEM approaches used in micromechanics: a) discretization by standard elements, b) special hybrid elements, c) pixel/voxel discretization, d) multiphase elements 45

Especially when the phase arrangements to be studied are based on digital images of actual microgeometries, a third approach to discretizing microgeometries of interest. It consists of using a regular square or hexahedral mesh that has the same resolution as the digital image, each element being assigned to one of the constituents by operations such as thresholding of the grey values of the corresponding pixel or voxel, respectively. Such meshes have the advantage of allowing a straightforward automatic model generation from appropriate experimental data (metallographic sections, tomographic scans) and of avoiding ambiguities in smoothing the digital data (which are generally present if a standard FE mesh is employed to discretize experimental data of this type). Obviously voxel element strategies lead to ragged phase boundaries, which may give rise to some oscillatory behavior of the solutions (Niebur et al., 1999) and can lead to very high local stress maxima (Terada et al., 1997). Such digital image based (DIB) models are, however, claimed not to cause unacceptably large errors in the predicted eective behavior, even for relatively coarse discretizations, at least in the linear elastic range (Guldberg et al., 1998). Good spatial resolution, i.e., a high number of pixels or voxels, of course, is very benecial to such models. A fourth approach also uses regular FE meshes, but assigns phase properties at the integration point level of standard elements (multiphase elements), see, e.g., Schmauder et al. (1996). Essentially, this amounts to trading o ragged boundaries at element edges for smeared-out (and typically degraded) microelds within those elements that contain a phase boundary, because stress or strain discontinuities within elements cannot be adequately handled by standard FE shape functions. With respect to the element stinesses the latter concern can be much reduced by overintegrating elements containing phase boundaries, which leads to good approximations of integrals involving non-smooth displacements by numerical quadrature, see Zohdi and Wriggers (2001). The resulting stress and strain distributions, however, remain smeared-out approximations in elements that contain phase boundaries. A fairly recent development for studying microgeometries involving a large number of inhomogeneities (tens to thousands) by Finite Element methods involves programs specially geared towards solving micromechanical problems. Such codes may be based on matrix-free iterative solvers such as Conjugate Gradient (CG) methods, analytical solutions for the microelds (e.g. constant strain approximations corresponding to the upper Hill bounds, eqn. (4.1)), being used as starting solutions to speed convergence25 . For studies involving such programs see, e.g., Gusev (1997) or Zohdi and Wriggers (2001).

Essentially, in such a scheme the initial guess gives a good estimate of long wavelength contributions to the solution, and the CG iterations take care of short wavelength variations.

25

46

Chapter 6 Periodic Microeld Models

Periodic Microeld Approaches (PMAs) aim to describe the macroscopic and microscopic behavior of inhomogeneous materials by studying model materials that have periodic microstructures. The rst two parts of this chapter cover some basic concepts of unit cell based PMAs. Following this, applications to some types of composites and other inhomogeneous materials are discussed.

6.1

Basic Concepts of Unit Cell Models

Periodic microeld approaches analyze the behavior of innite (one, two- or three-dimensional) periodic arrangements of constituents making up a given inhomogeneous material under the action of far eld mechanical loads or uniform temperature elds26 . The most common approach to studying the stress and strain elds in such periodic congurations is based on partitioning the microgeometry into periodically repeating unit cells to which the investigations may be limited without loss of information or generality, at least for static analysis27 . A wide variety of unit cells have been employed in published PMA studies, ranging from simple periodic arrays of inhomogeneities to highly complex phase arrangements, such as multi-inclusion cells (supercells). For some simple periodic phase arrangements it is also possible to nd analytical solutions via series expansions that make explicit use of the periodicity (Sangani and Lu, 1987) or via appropriate potential methods (Wang et al., 2000). Even though most PMA studies in the literature have used standard numerical engineering methods as described in chapter 5, some more specialized approaches to evaluating
Standard PMAs cannot handle free boundaries or macroscopic gradients in mechanical loads, temperatures or composition in any direction in which the elds are periodic. If such gradients or free boundaries are to be studied, however, in many cases layer models can be generated that are non-periodic in one direction and periodic in the other(s), see, e.g., Reiter and Dvorak (1998) and Weissenbek et al. (1997). 27 Periodic phase arrangements, especially simply periodic microstructures, typically have to be used with extreme care in dynamic analysis, because they act as lters that exclude all waves with frequencies that do not fall within certain bands, see, e.g., Suzuki and Yu (1998). In addition, due to the boundary conditions required for obtaining periodicity, unit cell analysis can only handle wavelengths that are smaller than or equal to the appropriate cell dimension, and mechanical waves are locked in within the cell instead of being allowed to pass through and continue into the far eld. For the same reason, unit cell based methods typically can resolve only subsets of the buckling modes and buckling loads in stability analysis (depending on the geometry of the cell), so that considerable care is required in using them for such purposes.
26

47

the mechanical behavior of periodic geometries are worth mentioning. One of them, known as the Method of Cells (Aboudi, 1989), discretizes unit cells that correspond to square arrangements of square bers into four subcells, within each of which displacements are approximated by low-order polynomials. Traction and displacement continuity conditions at the faces of the subcells are imposed in an average sense and analytical and/or semianalytical approximations to the deformation elds are obtained in the elastic and inelastic ranges. Even though they use highly idealized microstructures, are restricted in handling axial shear and provide only limited information on the microscopic stress and strain elds, the resulting models pose relatively low computational requirements. The Method of Cells has been used as a constitutive model for analyzing structures made of continuously reinforced composites, see, e.g., Arenburg and Reddy (1991). Developments of the algorithm led to the Generalized Method of Cells (Aboudi, 2004), which allows ner discretizations of unit cells for ber and particle reinforced composites, reinforcement and matrix being essentially split into a number of subregions of rectangular or hexahedral shape. For some comparisons with microelds obtained by Finite Element based unit cells see Pahr and Arnold (2002). Another micromechanical approach, the Transformation Field Analysis of Dvorak (1992), allows the prediction of the nonlinear response of inhomogeneous materials based on either mean eld descriptions (compare the remarks in section 3.7) or on periodic microeld models for the elastic behavior. Provided suciently ne discretizations of the phases are employed the use of essentially elastic accommodation has been found to be acceptable, and high computational eciency is claimed for the method (Dvorak et al., 1994). A further solution strategy for PMAs uses numerically evaluated equivalent inclusion approaches that explicitly account for interacting inhomogeneities, see, e.g., Fond et al. (2001). In typical periodic microeld approaches strains and stresses are approximated as the sum of uniform macroscopic contributions (slow variables), and , and periodically varying microscopic uctuations (fast variables), and , respectively, i.e., (z) = (z ) = + (z ) + (z )

(6.1)

Here z is a microscopic coordinate that has sucient resolution for describing the variations on the microscale. For suitably dened unit cells, UC , eqn. (1.2) implies that the microscopic uctuations must fulll the relations 1 UC (z) d = 0
UC

and

1 UC

(z) d = 0
UC

(6.2)

As a consequence of eqn. (6.1), in periodic microeld approaches each unit of periodicity (unit cell) contributes the same increment of the displacement vector u and the macroscopic displacements vary (multi)linearly. An idealized depiction of such a situation is presented in g. 6.1, which shows the periodic variations of the strains s (s) = s + s (s) and of the corresponding displacements us (s) = s s + us (s) along the length s of a generic one-dimensional periodic two-phase material consisting of constituents A and B. The periodicity of the strains and of the 48

displacements is immediately apparent, the unit of periodicity and the corresponding displacement increment being marked as cU and us , respectively.

, u

<s> s us <s> s

us

cu A B A B A B A

s
Figure 6.1: Schematic depiction of the variation of the strains s (s) and the displacements us (s) along a generic one-dimensional composite (coordinate s) consisting of constituents A and B. Symmetry points of s (s) and us (s) are indicated by small circles.

6.2

Boundary Conditions

The proper use of unit cell based methods requires that the cells together with the boundary conditions (BC) prescribed on them generate valid tilings both for the undeformed geometry and for all deformed states pertinent to the problem (i.e., gaps and overlaps between neighboring unit cells as well as unphysical constraints on the deformations must not be possible). In order to achieve this, the boundary conditions for the unit cells must be specied in such a way that all deformation modes appropriate for the load cases to be studied can be attained. The three major types of boundary conditions used in periodic microeld analysis are periodicity, symmetry, and antisymmetry (or point symmetry) BC28 . Generally, for a given periodic phase arrangement unit cells are non-unique, the range of possibilities being especially wide when point or mirror symmetries are present in the microgeometry. As an example, g. 6.2 depicts a hexagonal array of planar circular inhomogeneities and some of the unit cells that can be used to study the mechanical behavior of this arrangement. In g. 6.3 the three basic types of boundary conditions are sketched for two-dimensional unit cells. The sides and corner points of the cells are annotated and an extension of this nomenclature to three-dimensional cases is shown in g. 6.5.
In addition, free surface boundary conditions may be used for layer-type models. For a more formal treatment of boundary conditions for unit cells see, e.g., Michel et al. (1999).
28

49

B A

D E C

G F H I J

periodic boundary symmetry boundary point symmetry boundary

symmetry center (pivot point)

Figure 6.2: Periodic hexagonal array of circular inhomogeneities in a matrix and 10 unit cells that can be used to describe the mechanical responses of this arrangement under loads acting parallel to the coordinate axes. The most general boundary conditions for unit cells are periodicity BC, which can handle any possible deformed state of the cell and, consequently, of the inhomogeneous material to be modeled. In g. 6.2 cells A to E belong to this group. Because such unit cells tile the computational space by translation, neighboring cells (and, consequently, opposite faces of a given cell) must t into each other like the pieces of a jigsaw puzzle in both undeformed and deformed states. Periodic phase arrangements can be described by sets of periodicity vectors pn , where the number of periodicity vectors, N , equals the spatial dimension of the problem. The minimum volume of unit cells pertaining to such a set of periodicity vectors is well dened. There is, however, a wide variety shapes of minimum volume unit cells for a given periodic phase arrangement, as is indicated by g. 6.4. The surface of any unit cell for use with periodicity boundary conditions must consist of at least N pairs of faces (or parts of faces) k . The surfaces making up such a pair must be identical but are shifted relative to each other by shift vectors ck which are linear combinations of the periodicity vectors. In g. 6.4 pairs of faces are marked by being drawn with the same line style.
NW NE N 111111 u 000000 111111111111111 000000000000000 000000 111111 u 000000000000000 111111111111111 000000000000000 111111111111111 000000000000000 111111111111111 W 000000000000000 111111111111111 E 000000000000000 111111111111111 000000000000000 111111111111111 u 000000000000000 SW 111111111111111 SE 000000000000000 111111111111111 000000000000000 111111111111111 S z
NW NE SE

z2

111 000 u 000 111 000 111 NW 000 NE 111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 W 1111111111111 E 0000000000000 0000000000000 1111111111111 0000000000000 1111111111111 u SW 1111111111111 SE 0000000000000
uNW

NE

SE

u 0000 N 1111 1111111111111 0000000000000 NE 0000 1111 0000000000000 1111111111111 0000 1111 U 0000000000000 1111111111111 E 0000000000000 1111111111111 W 1111111111111 0000000000000 P u 0000000000000 1111111111111 0000000000000 1111111111111 u 00000 11111 L 0000000000000 00000 11111 SW 1111111111111 0000000000000 1111111111111 S uNW
U

NW

Figure 6.3: Sketch of periodicity, symmetry, and antisymmetry boundary conditions as used with two-dimensional unit cells. 50

p2 p1+ p2 p1

Figure 6.4: Six dierent minimum unit cells for a two-dimensional periodic matrix inclusion medium with two (slightly) non-orthogonal translation vectors p1 and p2 . Paired faces (or parts of faces) are marked by identical line styles and inhomogeneities inside the unit cells are set o by darker shading. In the small strain regime the most general formulation of periodicity boundary conditions for the pair of surfaces k can be given as u(sk + ck ) u(sk ) = ck , (6.3)

where sk are position vectors on the faces making up the pair. For the special case of quadrilateral two-dimensional unit cells, such as the one shown in g. 6.3 (left) this leads to the expressions uN ( s1 ) = uS ( s1 ) + uNW uE ( s2 ) = uW ( s2 ) + uSE , (6.4)

where s k denotes the position in a local coordinate system on the pair of faces making up k and vertex SW is assumed to be xed. Equations (6.4) directly imply that uNE = uNW + uSE .

For numerical analysis the two faces making up a pair k must be discretized in a compatible way, i.e., the nodal points on them must be positioned at equal values of the face coordinates s k . Equations (6.3) then become sets of linear constraints each of which links three nodal displacement DOFs29 . Comparing eqns.(6.3) and (6.4) shows that the displacements of the master nodes, SE and NW, contain the information on the macroscopic strain tensor . In addition, the displacements of the master nodes and of faces S and W fully control the displacements of the slave faces N and E.
In principle, all variables (i.e., for mechanical analysis the displacements, strains and stresses) must be linked by appropriate periodicity conditions. When a displacement based FE code is used, however, such conditions can be specied explicitly only for the displacement components (including, where appropriate, rotational DOFs). The periodicity of the stresses and strains is fullled approximately, mainly because in typical implementations nodal stresses and strains are not averaged across cell boundaries, even though they ought to be. The periodicity of stresses leads to antiperiodic tractions at unit cell boundaries.
29

51

Conditions analogous to eqn. (6.4) can be specied for any appropriately chosen regular space-lling two-dimensional cell having an even number of sides (squares, rectangles, and hexagons) and for three-dimensional cells with an even number of faces (cubes, hexahedra, rhombic dodekahedra, and regular tetrakaidekahedra), compare g. 6.5. The scheme can also be extended to unit cells of less regular shape, compare Cruz and Patera (1995). Periodicity BC generally are the least restrictive option for unit cell models using phase arrangements obtained by statistically based algorithms (supercells). In practice FEbased unit cell analysis with periodicity boundary conditions can be rather expensive in terms of computing time and memory requirements, because the linear equations between three boundary DOFs corresponding tend to markedly degrade the band structure of the system matrix, especially in three-dimensional problems. Compared to other discrete microstructure approaches, periodic homogenization typically shows the fastest convergence in terms of cell sizes, see, e.g., El Houdaigui et al. (2007).

Figure 6.5: Sketch of three-dimensional unit cell with nomenclature of vertices, edges and faces. For two-dimensional rectangular and hexahedral unit cells in which all faces of the cell coincide with symmetry planes of the phase arrangement and for load cases that do not interfere with these mirror symmetries, periodicity BC simplify to symmetry boundary conditions. For the cell shown in g. 6.3 (center) these take the form uE ( s2 ) = uSE vN ( s1 ) = uNW uW ( s2 ) = 0 vS ( s1 ) = 0 , (6.5)

where u and v stand for the displacement components in 1- and 2-directions, respectively. Again the vertices SE and NW play the role of master nodes, but the constraint equations are much simpler than eqn.(6.4), involving only 2 DOFs. Volume elements using symmetry boundary conditions must be rectangles or regular hexahedra, compare cell G in g. 6.2; the extension of eqn.(6.5) to three-dimensional problems is straightforward. The load cases that can be handled with symmetry BC are restricted to uniform thermal loads, normal mechanical loads acting on one or more pairs of faces, and combinations of the above30 . A modied version of symmetry boundary conditions can be used to model the bending of layer models, see, e.g., Weissenbek et al. (1997). Symmetry boundary conditions
Because the load cases that can be handled include uniaxial loading normal to the faces, loading by extensional shear (obtained, e.g., in the two-dimensional case by applying normal stresses a to the vertical and a to the horizontal faces), biaxial loading, hydrostatic loading and thermal loading, symmetry BC in many cases are sucient for materials characterization.
30

52

are very useful for describing relatively simple microgeometries, but tend to be restrictive in modeling phase arrangements generated by statistically based algorithms because inhomogeneities crossed by the cell boundaries must be symmetric with respect to the cell faces. Antisymmetry boundary conditions are even more limited in terms of the microgeometries that they can handle, because they require the presence of centers of point symmetry (pivot points). In contrast to symmetry boundary conditions, however, unit cells employing them on all faces are subject to few restrictions to the load cases that can be handled. Among the unit cells shown in g. 6.2, cells G and H use point symmetry BC on all faces and can handle any in-plane deformation31 . Alternatively, antisymmetry BC can be combined with symmetry BC to obtain very small unit cells that are restricted to loads acting normal to the symmetry faces, compare unit cells I and J in g. 6.2. The sketch in g. 6.3 (right) also shows such a small unit cell, the antisymmetry boundary conditions being applied to the E-side only, where a pivot point P is present. For this conguration, the boundary conditions can be denoted as uU ( sP ) + uL (s P ) = 2uP vN ( s1 ) = vNW = 2vP

vS ( s1 ) = 0

uW ( s2 ) = 0

(6.6)

where uU ( sP ) and uL (s P ) are the deformation vectors of pairs of points U and L that are positioned symmetrically on face E (where a local coordinate s P centered on the pivot P is dened in analogy to g. 6.11) and the displacements of which are point symmetric with respect to P. The phase arrangement, the undeformed geometry and the discretization of such a face must also be antisymmetric with respect to the pivot point P. Three-dimensional unit cells employing combinations of symmetry and point symmetry BC can be used to advantage for studying cubic arrays of particles, see, e.g., Weissenbek et al. (1994).

6.3

Application of Loads and Evaluation of Microelds

The primary practical challenge in using periodic microeld approaches for modeling inhomogeneous materials lies in choosing and generating suitable unit cells that in combination with appropriate boundary conditions allow an accurate representation of the actual microgeometries within available computational resources. These unit cells must then be subjected to appropriate macroscopic stresses, strains and temperature excursions. Whereas loads of the latter type generally do not pose a problem, applying far eld stresses or strains is not necessary trivial (note that the variations of the stresses along the faces of a unit cell in general are not known a priori, so that it is not possible to prescribe appropriate boundary tractions via distributed loads).
Unit cells used with antisymmetry boundary conditions may have odd numbers of faces. Triangular unit cells similar to cell H in g. 6.2 were, e.g., used by Teply and Dvorak (1988) to study the transverse mechanical behavior of hexagonal arrays of bers. Rectangular cells with point symmetries on each boundary were introduced by Marketz and Fischer (1994) for perturbed square arrangements of inhomogeneities.
31

53

Asymptotic Homogenization
The most versatile and elegant strategy for linking the macroscale and microscale in unit cell analysis is based on a mathematical framework known as homogenization theory or asymptotic homogenization, see, e.g., Suquet (1987). Macroscopic and microscopic coordinates, Z and z, respectively, are explicitly introduced. The macroscopic coordinates, Z, correspond to the standard coordinates x, whereas the microscopic coordinates z are scaled up as zi = xi / (6.7) in order to zoom in on the local behavior. Here the ratio between the characteristic lengths of the macroscale, L, and the microscale, , is used as scaling parameter = /L 1. The displacement eld in the unit cell can then be represented by an asymptotic expansion of the type ui (Z, z, ) = ui (Z) + ui (Z, z) + 2 ui (Z, z) + H.O.T.
(0) (1) (0) (1) (2)

(6.8)

where the ui are the eective or macroscopic displacements and ui stands for the periodically varying displacement perturbations due to the microstructure32 . Using the chain rule, i.e., 1 f (Z(x), z(x), ) f+ f , (6.9) x Z z the strains can be related to the displacements (in the small strain regime) as ij (Z, z, ) = 1 2 + 2
(1)

(0) u + Zj i (1) u + Zj i

(0) u + Zi j (1) u + Zi j

(1) u + zj i (2) u + zj i ,

(1) u zi j (2) u zi j

+ H.O.T. (6.10)

= ij (Z, z) + ij (Z, z) + H.O.T.

(0) (0)

(2)

where terms of the type ij = 1 u are deleted due to the underlying assumption that zj i the variations of slow variables are negligible at the microscale. The strains of order O (0 ) can be split into slow and fast contributions,

ij =

(1)

(0) 1 (0) ui + u 2 Zj Zi j
(1)

and

ij =

(1)

(1) 1 (1) ui + u 2 zj zi j

(6.11)

respectively. The stresses can be expanded into the expression ij (Z, z, ) = ij (Z, z) + ij (Z, z) + H.O.T.
(2)

(6.12)

Using the two-scale assumption and, as a consequence, eqn. (6.9), the equilibrium equations take the form 1 + ij (Z, z, ) + fi (Z) = 0 Zj zj
32

(6.13)

The nomenclature used in eqns. (6.7) to (6.17) follows typical usage in asymptotic homogenization. It is more general than but can be directly compared to the one used in eqns. (6.1) to (6.6), where no macroscopic coordinate Z is employed. Zero order terms can generally be identied with macroscopic quantities and rst order terms generally correspond to periodically varying microscopic quantities.

54

the fi being macroscopic body forces. By inserting eqn. (6.12) into this expression and sorting the resulting terms by order of a hierarchical system of partial dierential equations is obtained (1) =0 zj ij (1) (2) ij + + fi = 0 Zj zj ij 1 ) 0 ) , (6.14)

(order (order

the rst of which gives rise to a boundary value problem at the unit cell level that is referred to as the micro equation. Using eqn. (6.11) and elastic stressstrain relations, (1) an ansatz for the strains ij can be made in the form ij = Iijmn + ij + ij =
(1) (1) (1)

imn (1) mn zj

(6.15)

so that the microstresses appearing in the micro equation can be expressed as ij = Eijkl (z) Iklmn +
(1)

kmn (1) mn zl

(6.16)

where Eijkl (z) is the phase-level elasticity tensor, Iijkl is the 4th-order unit tensor, and the characteristic function ijk (z) describes the deformation modes of the unit cell. The second equation of the system (6.14), the macro equation, after volume averaging over the unit cell, allows to link the macroscopic and microscopic elds. The homogenized elasticity tensor is then obtained as Eijkl = 1 UC Eijkl (z) Iklmn +
UC

kmn (z) d zl

(6.17)

Analogous expressions can be derived for the tangent modulus tensors used in nonlinear analysis, compare Ghosh et al. (1996). These relations can be used as the basis of Finite Element algorithms that solve for the characteristic function ijk , a task that typically has required special analysis codes. For detailed discussions of asymptotic homogenization methods within the framework of FEM-based micromechanics see, e.g., Hollister et al. (1991), Ghosh et al. (1996) or Hassani and Hinton (1999). An asymptotic homogenization procedure for elastic composites that uses standard elements within a commercial FE package is described by Banks-Sills and Leiderman (1997). On the basis of homogenization theory macroscopic problems can be solved without explicitly specifying homogenized constitutive laws, so that simultaneous two-scale analysis can be carried out, see, e.g., Ghosh et al. (1996), Feyel (2003) or Terada et al. (2003).

Method of Macroscopic Degrees of Freedom

When asymptotic homogenization is not used, it is good practice to apply far eld stresses and strains to a given unit cell via concentrated nodal forces or prescribed displacements, 55

respectively, at the master nodes and/or pivots, an approach termed the method of macroscopic degrees of freedom by Michel et al. (1999). Employing the divergence theorem and using the nomenclature and conguration of g. 6.3 (left), for load controlled analysis the forces to be applied to the master nodes SE and NW, PSE and PNW , of a two-dimensional unit cell with periodicity boundary conditions can be shown to be given by the surface integrals PSE =
E

ta (z) d

PNW =
N

ta (z) d

(6.18)

Here ta (s) = a n (s) stands for the homogeneous surface traction vector corresponding to the applied (far eld) stress eld33 at some given point s on the cells surface UC , and n (s) is the local unit normal vector to the appropriate face. Equation (6.18) can be generalized to require that each master node is loaded by a force corresponding to the surface integral of the surface traction vectors over the face slaved to it via an equivalent of eqns. (6.18), compare Smit et al. (1998). An analogous procedure holds for three-dimensional cases. For geometrically nonlinear analysis eqns. (6.18) must be applied to the current conguration. For applying far eld strains to periodic volume elements, the displacements to be prescribed to the master nodes must be obtained from the macroscopic strains via appropriate displacementstrain relations. For example, using the notation of eqns. (6.4), the displacements to be prescribed to the master nodes NE and SW of the unit cell shown in g. 6.3 (left) can be evaluated from eqn. (6.3) as
a uSE = a vSE = 12 c1 11 c1 a a uNW = 21 c2 vNW = 22 c2

(6.19)

for an applied strain a and linear displacementstrain relations34 . Typically strain controlled unit cell models are easier to handle than stress controlled ones. In general, the overall stress and strain tensors within a unit cell can be evaluated by volume averaging (e.g., using some numerical integration scheme) or by using the equivalent surface integrals given in eqn. (1.2), i.e., = 1 UC 1 = UC 1 UC UC 1 (z) d = 2UC UC (z) d = t z d
UC

(u n + n u) d
UC

(6.20)

In the case of rectangular or hexahedral unit cells that are aligned with the coordinate axes, averaged engineering stress and strain components can, of course, be evaluated by dividing the applied or reaction forces at the master nodes by the appropriate surface areas and by dividing the displacements at the master nodes by the appropriate cell lengths, respectively.
Note that the ta (s) are not identical with the actual local values of the tractions, t(s), at the cell boundaries, but are equal to them in an integral sense over the cell face. 34 The displacements at node NE are fully determined by those of the master nodes, compare eqns. (6.4), so that one of the variables vSE and uNW is not independent. Analogous relations hold for three-dimensional cases.
33

56

In order to obtain three-dimensional homogenized elastic tensors with the method of macroscopic degrees of freedom six suitable, linearly independent load cases must be solved for. For evaluating phase averaged quantities from unit cell analysis, it is good practice to use direct volume integration according to eqn. (1.8) in analogy to the left hand terms in eqn. (6.20)35 . In many FE codes this can be done by approximate numerical quadrature according to N 1 1 f ( z ) d fl l (6.21) f = l=1 Here fl and l are the function value and the integration weight (in terms of the volume of the integration point), respectively, associated with the l-th integration point within a given integration volume that contains N integration points. When macroscopic values or phase averages are to be generated of variables that are nonlinear functions of the stress or strain components (e.g., equivalent stresses, equivalent strains or stress triaxialities), only direct volume averaging should be used, because evaluation of nonlinear variables on the basis of averaged components may lead to unacceptable inaccuracies, compare also eqn. (3.58).

6.4

Unit Cell Models for Continuous Fiber Reinforced Composites

Composites reinforced by continuous aligned bers typically show a statistically transversely isotropic overall behavior and can be studied well with periodic homogenization. Materials characterization with the exception of the overall axial shear behavior can be carried out with two-dimensional unit cell models employing generalized plane strain elements that use a global degree of freedom to describe the axial deformation of the whole model36 . Such elements are implemented in a number of commercial FE codes. For handling the overall axial shear response, however, special elements (Adams and Crane, 1984) or threedimensional analysis with appropriate periodicity boundary conditions (Pettermann and Suresh, 2000) are required. It is worth noting that in all the above cases generalized plane strain states of some type are described, i.e., the axial strains are constant over the unit cell although out-of-plane deformations are present. Three-dimensional analysis is necessary for studying composites reinforced by continuous aligned bers when the eects of ber misalignment or of broken bers are to be studied, see, e.g., Mahishi (1986).
It is of some practical interest that volume averaged and phase averaged microelds obtained from unit cells must fulll all relations given in section 3.1 and can thus be conveniently used to check the consistency of a given model by inserting them into eqns. (3.4). Note, however, that for nite deformations appropriate stress and strain measures must be used for this purpose (Nemat-Nasser, 1999). 36 Because the axial stiness of composites reinforced by continuous aligned bers can usually be satisfactorily described by Voigt-type models, compare eqn. (2.2), unit cell models of such materials have tended to concentrate on the transverse behavior. Plain strain models, however, can describe neither the axial constraints nor the axial components of microstresses and microstrains.
35

57

PH0

PS0

CH1

MS5

RH2

CS7

CH3

CS8

Figure 6.6: Eight periodic ber arrangements of ber volume fraction =0.475 for modeling continuously ber reinforced composites (B ohm and Rammerstorfer, 1995). Basic generalized plane strain models of continuously reinforced composites make use of simple periodic ber arrangements as shown in g. 6.6, all of which can be described by
EPS.EFF.PLASTIC 2.25E-02 1.75E-02 1.25E-02 7.50E-03 2.50E-03 INC.800

Figure 6.7: Microscopic distributions of the accumulated equivalent plastic strain in the matrix of a transversely loaded unidirectional continuously reinforced ALTEX/Al MMC ( =0.453) as predicted by a multi-ber unit cell (arrangement DN). 58

rather small unit cells using symmetry and/or point symmetry BC. The simplest among these arrangements are the periodic hexagonal (PH0) and periodic square (PS0) arrays. Models with hexagonal symmetry (PH0,CH1,RH2,CH3) give rise to transversely isotropic thermoelastic overall behavior, whereas the other ber arrangements shown in g. 6.6 have tetragonal (PS0,CS7,CS8) or monoclinic (MS5) overall symmetry. Elastoplastic behavior of the matrix causes the macroscopic symmetries of the ber arrangements to degrade under most loads due to the low symmetry of the local material properties, the material state depending on the load history a given point has undergone, compare also g. 6.8. In many cases simple periodic microgeometries do not provide fully satisfactory descriptions of ber reinforced materials, most of which show at least some randomness in the ber positions. Much improved models can be obtained by periodic multi-ber unit cells that employ quasi-random ber positions. Such models can either use symmetry BC, compare Nakamura and Suresh (1993) and g. 6.7, or periodicity BC (Zeman and Sejnoha, 2007). In table 6.1 thermoelastic moduli of an aligned continuously reinforced ALTEX/Al MMC as predicted by bounding methods, MFAs and unit cells methods using arrangements PH0, CH1 and PS0 (compare g. 6.6) as well as DN (see g. 6.7) are listed. For this material combination all unit cell results (even PS0, which is not transversally isotropic) Table 6.1: Overall thermoelastic moduli of a unidirectional continuously reinforced ALTEX/Al MMC ( =0.453) as predicted by the HashinShtrikman (HS) and three-point (3PB) bounds, by the MoriTanaka method (MTM) and the generalized self-consistent scheme (GSCS), by the dierential scheme (DS), by Torquatos three-point estimates (3PE), as well as by unit cell analysis using periodic arrangements shown in g. 6.6 (PH0, CH1, PS0) and the multi-ber cell displayed in g. 6.7 (DN). For arrangement PS0 responses in the 0 and 45 , and for arrangement DN responses in the 0 and 90 directions are listed. El [GPa] bers 180.0 matrix 67.2 HS/lo 118.8 HS/hi 119.3 3PB/lo 118.8 3PB/hi 118.9 MTM 118.8 GSCS 118.8 DS 118.8 3PE 118.8 PH0 118.8 CH1 118.7 PS0/00 118.8 PS0/45 118.8 DN/00 118.8 DN/90 118.8
Eq [GPa] 180.0 67.2 103.1 107.1 103.8 104.5 103.1 103.9 103.9 103.9 103.7 103.9 107.6 99.9 104.8 104.6 lq [] 0.20 0.35 0.276 0.279 0.278 0.279 0.279 0.279 0.278 0.279 0.279 0.279 0.279 0.279 0.278 0.278 qt l q [] [K 1 106 ] [K 1 106 ] 0.20 6.0 6.0 0.35 23.0 23.0 0.277 0.394 0.326 0.347 0.342 11.84 16.46 0.337 11.84 16.46 0.339 11.94 16.35 0.338 11.89 16.40 0.340 11.84 16.46 0.338 11.90 16.42 0.314 11.85 16.45 0.363 11.85 16.45 0.334 11.90 16.31 0.333 11.90 16.46

59

fall within the HashinShtrikman bounds, but the predictions for the square arrangement show clear in-plane anisotropy and do not follow the three-point bounds. The results for the multi-ber arrangement indicate some minor deviation from transversely isotropic macroscopic behavior. The ber arrangements shown in gs. 6.6 and 6.7 give nearly identical results for the overall thermoelastoplastic behavior of continuously reinforced composites under axial mechanical loading, and the predicted overall axial and transverse responses under thermal loading are also very similar. The overall behavior under transverse mechanical loading, however, depends markedly on the phase arrangement, see g. 6.8. For ber arrangements of tetragonal or lower symmetry (e.g., PS0, MS5, CS7 and CS8) the predicted transverse stinesses depend strongly on the loading direction, and the behavior of the hexagonal arrangements is sandwiched between the sti (0 ) and the compliant (45 ) responses of periodic square arrangements in both the elastic and elastoplastic ranges. Multi-ber unit cells such as g. 6.7 that approach statistical transverse isotropy tend to show noticeably stronger strain hardening compared to periodic hexagonal arrangements.
6.00E+01

5.00E+01

Fiber

PS/00 DN PH PS/45 Matrix

APPLIED STRESS [MPa]

1.00E+01

2.00E+01

3.00E+01

4.00E+01

0.00E+00

DN/00 PS0/45 PS0/00 PH0/90 PH0/00 Al99.9 MATRIX ALTEX FIBER

0.00E+00 5.00E-04 1.00E-03 1.50E-03 2.00E-03 2.50E-03 3.00E-03 3.50E-03 4.00E-03

STRAIN []

Figure 6.8: Transverse elastoplastic response of a unidirectional continuously reinforced ALTEX/Al MMC ( =0.453, elastoplastic matrix with linear hardening) to transverse uniaxial loading as predicted by unit cell models PH0, PS0 and DN. The distributions of microstresses and microstrains in bers and matrix typically depend markedly on the ber arrangement, especially under thermal and transverse mechanical loading. In the plastic regime, the microscopic distributions of equivalent stresses and equivalent plastic strains, of hydrostatic stresses, and of stress triaxialities tend to be strongly inhomogeneous, compare g. 6.7. As a consequence, the onset of ductile damage in the matrix (or of brittle damage in brittle matrix composites), of brittle failure of the bers and of interfacial decohesion at the bermatrix interfaces show a strong dependence on the ber arrangement.

60

6.5

Unit Cell Models for Short Fiber Reinforced Composites

The overall symmetry of short ber reinforced composites in many cases is isotropic (for random ber orientations) or transversely isotropic (for aligned bers, planar random bers and other ber arrangements with axisymmetric orientation distributions). However, processing conditions can give rise to a wide range of ber orientation distributions and, consequently, lower overall symmetries (Allen and Lee, 1990). The thermoelastic and thermoelastoplastic behavior of aligned short ber reinforced composites has been successfully estimated by MoriTanaka methods, which can also be extended to nonaligned bers and reinforcements showing an aspect ratio distribution, compare section 3.8. Such mean-eld approaches are, however, limited in resolving details of ber arrangements, especially for inelastic material behavior. At present the most powerful tools for studying the inuence of ber shapes and orientations, of clustering eects, of the interaction of bers of dierent sizes, and of local stress and strain elds between neighboring bers are periodic microeld methods. In contrast to continuously reinforced composites, the phase arrangements of discontinuously reinforced materials are inherently three-dimensional. The simplest threedimensional unit cell models of aligned short ber reinforced composites have used periodic square arrangements of non-staggered or staggered aligned bers37 , see, e.g., Levy and Papazian (1991) and compare g. 6.9. Somewhat more complex three-dimensional models can be based on periodic arrays of alternatingly tilted misaligned bers (Srensen et al., 1995). Such approaches are relatively simple but rather restrictive in terms of ber

Figure 6.9: Three-dimensional unit cells for modeling non-staggered (left) and staggered (right) square arrangements of aligned short bers. Shaded cells require symmetry BC, cells outlined in bold can use periodicity BC.
Such square arrangements give rise to tetragonal overall symmetry, and, consequently, the transverse overall properties are direction dependent.
37

61

arrangements that can be handled. When larger unit cells supporting periodicity boundary conditions are used the full thermomechanical behavior of the composites can be studied. For many materials characterization studies, a more economical alternative to the above three-dimensional unit cells are axisymmetric models describing the axial behavior of nonstaggered or staggered arrays of aligned cylindrical short bers in an approximate way. The basic idea behind these models is to replace unit cells for square or hexagonal arrangements by circular composite cylinders of equivalent cross sectional area (and volume fraction) as sketched in 6.10. The resulting axisymmetric cells are not unit cells in the strict sense, because they overlap and are not space lling. In addition, they do not have the same transverse ber spacing as the corresponding three-dimensional arrangements and cannot be used to study the response to most transverse mechanical loading conditions. However, they have the advantage of signicantly reduced computational requirements. Symmetry boundary conditions are used for the top and bottom faces of these cells, and the BCs at the outer (circumferential) surface are chosen to maintain the same cross sectional area along the axial direction for an aggregate of cells.

Figure 6.10: Periodic arrays of aligned non-staggered (top) and staggered (bottom) short bers and corresponding axisymmetric cells (left: cross sections in transverse plane; right: sections parallel to bers). In the case of non-staggered bers this can be easily done by specifying symmetry-type boundary conditions for the outer surfaces, whereas for staggered arrangements a pair of cells with dierent ber positions is considered, for which the total cross sectional area is required to be independent of the axial coordinate. Using the nomenclature of g. 6.11 and the notation of eqns. (6.4) to (6.6), this leads to nonlinear relations for the radial displacements u and linear constraints for the axial displacements v at the outer surface, (rU + uU )2 + (rL + uL )2 = 2(rP + uP )2 and vU + vL = 2vP , (6.22)

respectively, where r is the radius of the undeformed cell. By choosing the two cells making up the pair to be antisymmetric with respect to a pivot point P the arrangement can be 62

N
NW

NE

~ s

N
NW

NE

U E P

E W
z
SW

W
z

L s ~
r SE UNDEFORMED

SW

L
r
SE

DEFORMED

Figure 6.11: Axisymmetric cell for staggered arrangement of short bers: undeformed and deformed shapes. described by a single cell with an antisymmetric outer (E-) face, U and L being nodes on this face that are positioned symmetrically with respect to the pivot, P. Nonlinear constraints such as eqn. (6.22), however, are not widely available and may be cumbersome to use in a given FE code. For most applications the BC for the radial displacements in eqn. (6.22) can be linearized without major loss in accuracy, so that antisymmetry boundary conditions analogous to eqn. (6.6) are obtained for the outer surface, u U + u L = 2u P and vU + vL = 2vP , (6.23)

For many years axisymmetric unit cell models of the types shown in g. 6.10 have been the workhorses of PMA studies of short ber reinforced composites, see, e.g., (Povirk et al., 1992) or Tvergaard (1994). Typically, descriptions using staggered arrangements allow a wider range of microgeometries to be covered, compare B ohm et al. (1993) or Tvergaard (2003), and give more realistic descriptions of actual composites. Both staggered and non-staggered axisymmetric models can be extended to study a considerable range of arrangements incorporating bers of dierent size and/or aspect ratio by foregoing the use of point symmetric geometries and coupling two or more dierent cells via the condition of keeping the cross sectional area of the aggregate independent of the axial coordinate (B ohm et al., 1993). Three-dimensional unit cells and axisymmetric cells have been used successfully for studying the nonlinear thermomechanical behavior of aligned short ber reinforced MMCs, e.g., with respect to the their stressstrain responses, to the pseudo-Bauschinger eect, and to thermal residual stresses. They have provided valuable insight into causes and eects of matrix, interface and ber damage. Over the past 15 years unit cells containing bers that follow some prescribed orientation distribution function (and where appropriate, ber size and/or aspect ratio distribution functions), have become feasible for non-dilute ber volume fractions. Such approaches, however, tend to pose considerable challenges in generating appropriate ber arrangements at non-dilute volume fractions due to geometrical frustration. The rst studies of this type were restricted to the linear elastic range, where the BEM has been found to answer well, see, e.g., Banerjee and Henry (1992). Meshing such congurations for use with the FEM can be dicult, compare Shephard et al. (1995), and analyzing the me63

Figure 6.12: Unit cell for a composite reinforced by randomly oriented short bers (B ohm et al., 2002). The nominal ber volume fraction is = 0.15 and the 15 cylindrical bers in the cell have a nominal aspect ratio of a = 5. chanical response of the resulting cells requires considerable computing power, especially for nonlinear material behavior. As an example, g. 6.12 shows a unit cell that contains 15 randomly oriented cylindrical bers of aspect ratio 5, supports periodicity boundary conditions, and is discretized by tetrahedral elements.

6.6

Unit Cell Models for Particle Reinforced Composites

Particle reinforced composites are typically assumed to show a statistically isotropic overall response, and mean eld models are available for describing the overall thermoelastic behavior for the case of spherical inhomogeneities. Extensions of the mean eld solutions into the nonlinear range are available, compare section 3.7, but they tend to be subject to limitations in predicting the overall thermomechanical response in the post-yield regime. In addition, mean eld methods for particle reinforced composites cannot account for many particle shape, clustering, and size distribution eects and cannot resolve local uctuations of the stress and strain elds. A diculty in using unit cell approaches to modeling particle reinforced materials with overall isotropic behavior is due to the fact that even for spherical inhomogeneities no simple periodic phase arrangement exists that is inherently elastically isotropic. Together with the wide variation in microgeometries and particle shapes in actual materials, this causes studies using generic phase arrangements to be subject to potentially tricky tradeos between keeping computational requirements at manageable levels (favoring simple particle shapes combined with two-dimensional models or simple three-dimensional microgeometries) and obtaining suciently realistic models for a given purpose (often leading to requirements for unit cells containing a high number of particles of complex shape at 64

randomly selected positions). In many respects, however, periodic microeld models of particle reinforced composites are subject to similar constraints and use analogous approaches as work on short ber reinforced composites. Most three-dimensional unit cell studies of generic microgeometries for particle reinforced composites have been based on simple cubic (sc), face centered cubic (fcc) and body centered cubic (bcc) arrangements of spherical, cylindrical and cube-shaped particles, compare Hom and McMeeking (1991). By invoking the symmetries of these arrangements and using symmetry as well as antisymmetry boundary conditions, relatively simple unit cells for materials characterization can be obtained38 , compare Weissenbek et al. (1994) and see gs. 6.13 and 6.14. In addition, work employing hexagonal or tetrakaidekahedral particle arrangements has been reported, see, e.g., Rodin (1993). Of the above arrangements, simple cubic models are probably the easiest to handle, but they show a marked anisotropy, much more so than fcc and bcc congurations.

1-part.cell 1-particle cell 2-particle cell

2-particle cell

2-particle cell

Reference Volume

1-particle cell

Figure 6.13: Simple cubic, face centered cubic, and body centered cubic arrangements Weissenbek et al. (1994).

Figure 6.14: Some unit cells for particle reinforced composites using cubic arrangements of inhomogeneities: s.c. arrangement of cubes, b.c.c. arrangement of cylinders and f.c.c. arrangement of spheres Weissenbek et al. (1994). During the past 15 years three-dimensional studies based on more complex phase arrangements have begun to appear in the literature. Gusev (1997) used Finite Element
Again, considerably larger unit cells with periodicity rather than symmetry boundary conditions are required for unrestricted modeling of the full thermomechanical response of cubic phase arrangements.
38

65

methods in combination with unit cells containing up to 64 statistically positioned particles to describe the overall behavior of elastic particle reinforced composites. Hexahedral unit cells containing up to 10 particles in a perturbed cubic conguration (Watt et al., 1996) as well as cube shaped cells incorporating at least 15 spherical particles in quasirandom arrangements (B ohm et al., 1999; B ohm and Han, 2001), compare g. 6.15, or clusters of particles (Segurado et al., 2003) were proposed for studying elastoplastic particle reinforced MMCs and related materials. Three-dimensional simulations involving high numbers of particles have been reported for investigating elastic composites (Michel et al., 1999), for studying brittle matrix composites that develop damage (Zohdi and Wriggers, 2001), and for rubber reinforced polymers (Fond et al., 2001).

Figure 6.15: Unit cell for a particle reinforced MMC ( =0.2) containing 20 spherical particles in a quasi-random arrangement suitable for using periodicity BC (B ohm and Han, 2001). In analogy to short ber reinforced materials, compare g. 6.10, axisymmetric cell models with staggered or non-staggered particles can be used for materials characterization of particle reinforced composites. By appropriate choice of the dimensions of the axisymmetric cells, sc, fcc and bcc arrangements of particles can be approximated. Axisymmetric cell models have been a mainstay of PMA modeling of materials containing particulate inhomogeneities, see, e.g., Bao et al. (1991). Due to their relatively low computational requirements, planar unit cell models of particle reinforced materials have also been used to a considerable extent. Typically, plane stress models (which actually describe reinforced sheets or the stress states at the surface of inhomogeneous bodies) show a more compliant and plane strain models (which correspond to reinforcement by aligned continuous bers rather than particles) show a stier overall response than three-dimensional descriptions, compare table 6.2. With respect to the overall behavior, plane stress analysis may be preferable to plane strain analysis, compare Weissenbek (1994), but no two-dimensional model gives satisfactory results in terms of the predicted microstress and microstrain distributions (B ohm and Han, 2001). Axisymmetric cell models typically provide considerably better results than planar ones.

66

Table 6.2: Overall thermoelastic properties of a particle reinforced SiC/Al MMC (spherical particles, =0.2) as predicted by the HashinShtrikman (HS) and third order (3PB) bounds, by the MoriTanaka method (MTM), by the generalized self-consistent scheme (GSCS), by the dierential scheme (DS), by Torquatos three-point estimates (3PE), as well as by unit cell analysis using three-dimensional cubic arrangements, axisymmetric cells approximating sc, fcc and bcc arrangements, two-dimensional multi-particle models based on plane stress (2/D PST) and plane strain (2/D PSE) kinematics, as well as the three-dimensional multi-particle model shown in g. 6.15. E [GPa] 429.0 67.2 90.8 114.6 91.4 93.9 90.8 91.5 92.7 91.8 95.4 88.1 87.9 85.5 98.7 92.4 E [100] E [110] [GPa] [GPa] 96.4 90.5 89.0 91.7 90.0 90.6 l [] [K 1 106 ] 0.17 4.3 0.35 23.0 0.286 16.8 0.340 18.6 0.323 18.5 0.328 18.6 0.329 18.6 0.327 18.6 0.326 18.4 0.327 18.6 18.7 18.6 18.6 18.6/18.6 18.1/19.0 18.4/18.8 0.334 0.500 0.326

particles matrix HS/lo HS/hi 3PB/lo 3PB/hi MTM GSCS DS 3PE sc fcc bcc axi/sc axi/fcc axi/bcc 2/D PST 2/D PSE 3/D

In table 6.2 bounds, MFA results, and a number of PMA predictions for the overall thermoelastic moduli of a particle reinforced SiC/Al MMC are compared, the loading directions for the cubic arrangements being identied by Miller indices. It is noteworthy that for the loading directions considered here, which do not span the full range of possible responses of the cells none of the cubic arrangements gives results that fall within the third order bounds for identical spherical particles. The overall anisotropy of the simple cubic arrangement is marked, whereas the body centered and face centered arrays deviate much less from overall isotropy. The predictions of the axisymmetric analysis can be seen to be of comparable quality in terms of the elastic moduli to those of the corresponding cubic arrays, but typically give an anisotropic thermal expansion behavior. The results listed for the three-dimensional multi-particle models are ensemble averages over a number of unit cells and loading directions, and they show very good agreement with the threepoint bounds and estimates, whereas there are marked dierences to the plane stress and plane strain models. 67

Similarly to ber reinforced MMCs, particle reinforced composites typically display highly inhomogeneous microscopic stress and strain distributions, especially for the matrix in the nonlinear range39 . For example, g. 6.16, shows the predicted equivalent plastic strains of the elastoplastic matrix inside a multi-particle unit cell model of an MMC. The concentrations of plastic strains can lead to shear bands when the matrix material displays softening behavior.
SECTION FRINGE PLOT EPS.EFF.PLASTIC INC.12 5.0000E-02 4.0000E-02 3.0000E-02 2.0000E-02 1.0000E-02

SCALAR MIN: 6.6202E-04 SCALAR MAX: 8.1576E-02

Figure 6.16: Predictions for the equivalent plastic strains in a particle reinforced MMC ( =0.2) subjected to uniaxial tensile loading obtained by a unit cell with 20 spherical particles in a quasi-random arrangement (B ohm and Han, 2001). If the particles are irregular in shape or if their volume fraction markedly exceeds a value of 0.5 (as is the case, e.g., for cermets such as WC/Co), generic unit cell models employing relatively regular particle shapes may not result in very satisfactory microgeometries. For such materials a typical approach consists of basing PMA models on a real structure obtained from a metallographic section, compare Fischmeister and Karlsson (1977). Due to the nature of the underlying experimental data, models of this type often take the form of planar analysis, the limitations of which have been discussed above.

6.7

Unit Cell Models for Woven and Laminated Composites

Periodic microeld methods play an important role in studying composites with woven, braided or knitted reinforcements, where the reinforcing phase takes the form of textile-like structures consisting of bundles of bers (tows). Unit cell models of woven composites are
Plane stress models tend to predict much higher levels of equivalent plastic strains in the matrix (and thus weaker hardening) than do plane strain and generalized plane strain models of the same reinforcement volume fraction, and the results from three-dimensional arrangements typically lying between those of the above groups of planar models.
39

68

Tows

Matrix

Composite

Figure 6.17: A unit cell for modeling a plain weave lamina via symmetry boundary conditions. The tow region (left), the matrix region (center) and the assembled unit cell are shown. typically based on modeling ber bundles as a mesophase with smeared out material properties, which, in turn, are obtained from analyzing unidirectionally continuously reinforced composites40 . Such matrixtow level unit cells in many cases describe one woven ply, so that free surfaces are used at the bottom and top faces. Symmetry boundary conditions can be used for materials characterization, but for full homogenization modied periodicity boundary conditions are required, which introduce macroscopic rotational degrees of freedom that allow to describe warping and twisting of the woven lamina or laminae. Unit cells for woven composites tend to be fairly complex geometrically, even when only symmetry BC are used, compare g. 6.17. They tend to be computationally expensive, especially for nonlinear constituent behavior. Another group of composite materials that can be studied to advantage by unit cell methods are laminates consisting of plies the thickness of which is not much greater than the ber diameter and which, accordingly, contain only few bers in thickness direction. Figure 6.18 shows unit cells for modeling the mechanical behavior of monolament rein

90 o plies

0o ply

Figure 6.18: Two unit cells for modeling cross ply laminates and one for modeling angle ply laminates.

40

This modeling strategy obviously is a type of multiscale modeling.

69

forced cross ply composites with two (left) and one (center) ber layers per ply, respectively, and for modeling angle ply laminates (right).

6.8

Unit Cell Models for Porous and Cellular Materials

Elastoplastic porous materials have been the subject of a considerable number of PMA studies due to their relevance to the ductile damage and failure of metallic materials41 . Generally, modeling concepts for porous materials are closely related to those employed for particle reinforced composites, the main dierence being that the shapes of the voids may evolve signicantly through the loading history42 . Three-dimensional unit cells based on cubic arrangements of voids and axisymmetric cells as discussed in section 6.6 have been used in the majority of the pertinent PMA studies. In cellular materials, such as foams and cancellous bone, the volume fraction of the solid phase is low (often amounting to no more than a few percent) and the void phase may be topologically connected (open cell foams), unconnected (closed cell foams) or both of the above (e.g., hollow sphere foams). Such materials often display a small linear range, and at higher strains gross shape changes of the cells typically take place. This behavior is especially pronounced for compressive loading, where elastic buckling (e.g., in polymer foams), plastic buckling (e.g., in metal foams) and brittle failure (e.g., in cancellous bone) of struts or cell walls can play major roles. Accordingly, unit cells for such materials often require special provision for handling large deformations of and contact between cell walls43 . In addition, care must be taken that boundaries on which symmetry BCs are specied do not coincide with cell walls that can be expected to show instabilities. Finally, models must be suciently large so that nontrivial deformation patterns can develop, compare Daxner (2003). The geometrically most simple cellular materials are regular honeycombs, which can be modeled by planar hexagonal cell models. Somewhat less ordered two-dimensional arrangements have been used for studying the crushing behavior of soft woods (Holmberg et al., 1999), and highly irregular planar arrangements, compare g. 6.19, can be used to study many aspects of the geometry dependence of the mechanical response of cellular materials, such as metallic foams.
Most constitutive models describing ductile damage and failure of metals are based on micromechanical considerations, among them Rice and Tracey (1969), Gurson (1977), Tvergaard and Needleman (1984) as well as Gologanu et al. (1997). 42 The evolution of the shapes of initially spheroidal voids under non-hydrostatic loads has been the subject of intensive studies by mean eld type methods, compare Kailasam et al. (2000). Such models are based on the assumption that initially spherical pores will stay ellipsoids throughout the deformation history, which axisymmetric cell analysis (G ar ajeu et al., 2000) has shown to be an excellent approximation for axisymmetric tensile load cases. For compressive loading, however, initially spherical pores may evolve into markedly dierent shapes (Segurado et al., 2002). 43 The standard analytical models for the thermomechanical behavior of cellular materials (Gibson and Ashby, 1988) use analytical unit cells.
41

70

Figure 6.19: Planar periodic unit cell for studying irregular cellular materials (Daxner, 2003). Three-dimensional studies of closed cell foams have used generic microgeometries based on cubic arrangements of spherical voids (Hollister et al., 1991), of truncated cubes plus small cubes (Santosa and Wierzbicki, 1998), of rhombic dodekahedra, and of regular tetrakaidekahedra (Simone and Gibson, 1998), compare g. 6.20, as well as on random arrangements of multiple spherical pores (Smit et al., 1999). Analogous model geometries have been reported for modeling open cell foams. The eects of details of the microgeometries of cellular materials (e.g., thickness distributions and geometrical imperfections or aws of cell walls or struts), which can considerably inuence the overall behavior (a typical example being the very small elastic ranges of metallic foams), have been an active eld of research over the past decade, see, e.g., Grenestedt (1998) and Daxner (2003). Unit cell models of cellular materials with realistic microgeometries often are rather complex and numerically demanding due to these materials tendency to deform by local mechanisms and instabilities. However, analytical

Figure 6.20: Foam microstructure modeled by regular tetrakaidekahedra (truncated octahedra). 71

solutions have been reported for some simple periodic phase arrangements, compare, e.g., Warren and Kraynik (1991). The modeling of a special cellular material, cancellous (or spongy) bone, has attracted considerable research interest for the past 25 years. Cancellous bone shows a wide range of microstructures, which can be idealized as beam or beamplate congurations (Gibson, 1985). The solid phase of cancellous bone is an inhomogeneous material at a lower length scale. In studying the mechanical behavior of cancellous bone, large three-dimensional unit cell models based on tomographic scans of actual samples and using voxel-based discretization schemes have become fairly widely used, compare Hollister et al. (1991).

72

Chapter 7 Embedded Cell Models

Like periodic microeld methods, Embedded Cell Approaches (ECAs) aim at predicting the microelds in inhomogeneous materials at high spatial resolution. For this purpose they use models consisting of a core (or local heterogeneous region), the conguration of which can range from rather simple to highly detailed phase arrangements, that is embedded in an outer region serving mainly for transmitting the applied loads, compare g. 7.1. This modeling strategy avoids some of the drawbacks of PMAs, especially the requirement that the geometry and all microelds must be strictly periodic44 . In modeling the core, similar considerations hold as for PMAs, especially with respect to the overall symmetry and to two-dimensional vs. three-dimensional analysis. Of course, a material description must be chosen for the outer region that is compatible to the smeared-out behavior of the core, so that errors in the accommodation of stresses and strains are avoided. Some care is also required with respect to spurious boundary layers which may occur at the interfaces between the core and the surrounding material45 . Although some analytical methods such as classical and generalized self-consistent schemes, see section 3.5, may be viewed as embedding schemes, most ECAs with more complex cores have been based on numerical engineering methods. Three basic types of embedding approaches can be found in the literature. One of them uses discrete phase arrangements in both the core region and in the surrounding material, the latter, however, being discretized by a much coarser FE mesh, see, e.g., Sautter et al. (1993). Such models, which in some ways approach descriptions of a full sample with a rened mesh in an interior region, to a large extent avoid boundary layers between core and outer regions, but tend to be relatively expensive computationally. In the second group of embedding methods the behavior of the outer regions is described via appropriate smeared-out constitutive models. In the simplest case these take the form of semi-empirical or micromechanically based constitutive laws that are prescribed a priori for the embedding zone and which must be chosen to correspond closely to the overall
Embedding analysis can be used without intrinsic restrictions at or in the vicinity of free surfaces and interfaces, it can handle gradients in composition and loads, and it can be employed for studying the interaction of macrocracks with microstructures. Also, the requirement of suciently separated length scales, compare section 1.1, does not necessarily apply to embedded models. 45 These interfaces are a consequence of the modeling approach only and do not have any physical background. Such boundary layers typically have a thickness of, say, an inhomogeneity diameter for elastic materials, but they may be longer ranged for nonlinear material behavior.
44

73

Figure 7.1: Schematic depiction of the arrangement of core and embedding region in an embedded cell model of a tensile test specimen. behavior of the core. This way, conceptually simple models are obtained that are very well suited to studying local phenomena such as the stress and strain distributions in the vicinity of crack tips or at macroscopic interfaces in composites (Chimani et al., 1997) or the growth of cracks in inhomogeneous materials (van der Giessen and Tvergaard, 1994; Wulf et al., 1996; Motz et al., 2001; Gonz alez and LLorca, 2007). A closely related approach uses appropriate coupling conditions (which may be implemented via the boundary conditions of the micromodel) to link a macromodel employing smeared-out material data with a micromodel describing discrete constituents (V aradi et al., 1999). The third type of embedding schemes employ the homogenized thermomechanical response of the core to determine the eective behavior of the surrounding medium, giving rise to models of the self-consistent type, which are mainly employed for materials characterization. The use of such approaches is, of course, predicated on the availability of suitable parameterizable constitutive laws for the embedding material that can follow the cores instantaneous homogenized behavior with high accuracy for all load cases and for any loading history. This requirement can typically be fullled easily in the linear range, see e.g., Chen (1997), but leads to considerable diculties when at least one of the constituents shows elastoplastic or viscoplastic material behavior46 . Accordingly, approximations have to be used (the consequences of which may be dicult to assess in view of the nonlinearity and path dependence of elastoplastic material behavior) and models of this type are best termed quasi-self-consistent schemes. Such approaches were discussed, e.g., by Bornert et al. (1994) and by Dong and Schmauder (1996). They are the only embedding methods that do micromechanics in the strict sense in that they handle scale transitions.
Typically, the eective yielding behavior of elastoplastic composites shows some dependence on the rst stress invariant, and, for low plastic strains, the homogenized response of the core tends to be strongly inuenced by the fractions of the elastoplastic constituent(s) that have actually yielded. In addition, in many cases anisotropies of the yielding and hardening behavior are introduced by the phase geometry (e.g., aligned bers) and by the phase arrangement of the core. Finding constitutive laws that, on the one hand, can satisfactorily account for such phenomena and, on the other hand, have the capability of being easily adapted to the instantaneous responses of the core by adjusting free parameters, poses a major obstacle for quasi-self-consistent schemes in nonlinear regimes.
46

74

For materials characterization embedded cells can be loaded by homogeneous stresses or strains applied to the outer boundaries. If cracks or similar problems are to be studied, however, it is preferable to impose displacement boundary conditions corresponding to the far eld behavior of suitable analytical solutions (e.g., to the displacement eld of a crack tip). Alternatively, complete samples can be considered in simulated experiments, as sketched in g. 7.1. Eective and phase averaged stresses and strains from embedded cell analysis are best evaluated with eqn. (6.20) or its equivalents, and in the case of geometrically complex phase arrangements it is good practice to use only the central regions of the core for this purpose in order to avoid possible boundary layers.

75

Chapter 8 Windowing Approaches

The aim of windowing methods is to obtain estimates or bounds for the macroscopic properties of inhomogeneous materials on the basis of non-periodic volume elements that are referred to as mesoscopic test windows or, shorter, as windows. These volume elements have simple shapes, are extracted at random positions and with random orientations from an inhomogeneous medium, and are smaller than RVEs, compare g. 8.1. Because the results of windowing pertain to individual samples rather than to a material, they are referred to as apparent (rather than eective) properties.

Figure 8.1: Schematic depiction of a composite and four rectangular windows of equal size. Windowing methods are based on a surface integral version of the Hill condition, eqn. (3.17), that takes the form t(x) n (x)
T

u(x) x d = 0 ,

(8.1)

see, e.g., Hazanov (1998). In general there are four ways of fullling this equation, three of them being based on uniform boundary conditions (Hazanov and Amieur, 1995; OstojaStarzewski, 2006). First, the traction term in eqn. (8.1) can be set to zero by specifying appropriate Neumann boundary conditions for the tractions t(x). This can be achieved by prescribing 76

a given macroscopically homogeneous stress tensor a to all faces of the volume element, t(x) = a n (x) x , (8.2)

leading to statically uniform boundary conditions (SUBC). Second, the right hand term in eqn. (8.1) can be set to zero by imposing a given macroscopically homogeneous strain tensor a on all boundaries, u(x) = a x x , (8.3)

resulting in kinematically uniform boundary conditions (KUBC). Third, mixed uniform boundary conditions (MUBC) may be specied, which enforce the scalar product under the integral to vanish separately for each face k making up the surface of the volume element, [t(x) n (x)
T

u(x) x d = 0

x k

(8.4)

This involves appropriate combinations of traction and strain components that are uniform over a given face of the volume element, but not macroscopically homogeneous. Mixed uniform boundary conditions that fulll eqns. (8.1) and (8.4) must be orthogonal in the uctuating contributions (Hazanov and Amieur, 1995). Finally, the uctuations of non-uniform boundary elds can be made to cancel out by pairing parallel faces of the volume element such that they show identical uctuations but surface normals of opposite orientations. This strategy, which does not involve homogeneous elds, is used by the periodicity boundary conditions discussed in chapter 6. Macrohomogeneous boundary conditions following eqns. (8.2) and (8.3) can be shown to give rise to lower and upper estimates, respectively, for the overall elastic stiness of a given mesoscopic volume element (Nemat-Nasser and Hori, 1993). Ensemble averages of such estimates obtained from windows of comparable size provide lower and upper bounds on the overall eective tensors of these volume elements. These bounds are sometimes referred to as mesoscale bounds. By denition, for RVEs the lower and upper estimates and bounds on the overall elastic properties must coincide (Hill, 1963). Accordingly, hierarchies of bounds that are generated from sets of windows of dierent sizes (Huet, 1999) can bring out the eects of the size of the volume elements and indicate the size of proper representative volume elements. Equations (8.4) can be fullled by a range of dierent MUBC, resulting in dierent estimates for the apparent macroscopic tensors. A specic set of mixed uniform boundary conditions that avoids prescribing nonzero boundary tractions was proposed by Pahr and Zysset (2008) for obtaining the apparent elastic tensors of cellular materials. Table 8.1 lists these six load cases for volume elements that have the shape of right hexahedra aligned with the coordinate system. Their edge lengths in the 1-, 2- and 3-directions are c1 , c2 and c3 , respectively. The components of the prescribed strain tensor are denoted as a ij and those a of the prescribed traction vector as i . When applied to periodic volume elements with orthotropic phase arrangements, these MUBC were found to give the same predictions for the macroscopic elasticity tensor as periodic homogenization. Accordingly, they are called 77

Table 8.1: The six linearly independent uniform strain load cases making up the periodicity compatible mixed boundary conditions (PMUBC) proposed by Pahr and Zysset (2008) and the pertinent boundary conditions for thermal loading; East, West, North, South, Top, and Bottom denote the faces of the hexahedral volume element, compare g. 6.5, and the ci are the lengths of its edges.
East u 1 = a 11 c1 /2 a a 2 = 3 =0 u1 = 0 a a 2 = 3 =0 u1 = 0 a a 2 = 3 =0 a u2 = 21 c1 /2 a u3 = 0, 1 =0 a u3 = 31 c1 /2 a u2 = 0, 1 =0 u1 = 0 a a 2 = 3 =0 u1 = 0 a a 2 = 3 =0 West u 1 = a 11 c1 /2 a a 2 = 3 =0 u1 = 0 a a 2 = 3 =0 u1 = 0 a a 2 = 3 =0 u 2 = a 21 c1 /2 a u3 = 0, 1 =0 a u3 = 31 c1 /2 a u2 = 0, 1 =0 u1 = 0 a a 2 = 3 =0 u1 = 0 a a 2 = 3 =0 North u2 = 0 a a 1 = 3 =0 a u2 = 22 c2 /2 a a 1 = 3 =0 u2 = 0 a a 1 = 3 =0 a u1 = 12 c2 /2 a u3 = 0, 2 =0 u2 = 0 a a 1 = 3 =0 a u3 = 32 c2 /2 a u1 = 0, 2 =0 u2 = 0 a a 1 = 3 =0 South u2 = 0 a a 1 = 3 =0 u 2 = a 22 c2 /2 a a 1 = 3 =0 u2 = 0 a a 1 = 3 =0 u 1 = a 12 c2 /2 a u3 = 0, 2 =0 u2 = 0 a a 1 = 3 =0 u 3 = a 32 c2 /2 a u1 = 0, 2 =0 u2 = 0 a a 1 = 3 =0 Top u3 = 0 a a 1 = 2 =0 u3 = 0 a a 1 = 2 =0 a u3 = 33 c3 /2 a a 1 = 2 =0 u3 = 0 a a 1 = 2 =0 a u1 = 13 c3 /2 a u2 = 0, 3 =0 a u2 = 23 c3 /2 a u1 = 0, 3 =0 u3 = 0 a a 1 = 2 =0 Bottom u3 = 0 a a 1 = 2 =0 u3 = 0 a a 1 = 2 =0 u 3 = a 33 c3 /2 a a 1 = 2 =0 u3 = 0 a a 1 = 2 =0 u 1 = a 13 c3 /2 a u2 = 0, 3 =0 a u2 = 23 c3 /2 a u1 = 0, 3 =0 u3 = 0 a a 1 = 2 =0

Tensile 1 Tensile 2 Tensile 3 Shear 12 Shear 13 Shear 23 Thermal Loading

periodicity compatible mixed uniform boundary conditions (PMUBC). The concept of periodicity compatible mixed uniform boundary conditions can be extended to thermoelasticity by adding a load case that constrains all displacements normal to the faces of the volume element, sets all in-plane tractions to zero, and applies a uniform temperature increment T , see table 8.1. This allows to evaluate the volume averaged specic thermal stress tensor , from which the apparent thermal expansion tensor can be obtained as = C . It has been shown that the PMUBC listed in table 8.1 give valid results for suborthotropic periodic unit cells (Pahr and B ohm, 2008), which indicates that they are also applicable to non-periodic volume elements provided the sub-orthotropic contributions to their overall symmetry is relatively small. The PMUBC accordingly oer an attractive option for evaluating estimates of the macroscopic elasticity and thermal expansion tensors of periodic and non-periodic volume elements. The generation of lower and upper estimates by windowing using SUBC and KUBC can be shown to be valid in the context of nonlinear elasticity and deformation plasticity (Jiang et al., 2001). PMUBC, however, rely on the superposition principle for obtaining eective tensors and, accordingly, cannot be used for generating tangent tensors. They can, however, be employed for some materials characterization tasks for elastoplastic inhomogeneous materials (Pahr and B ohm, 2008). Like embedding methods windowing gives rise to perturbed boundary layers near the surfaces of the volume element, which may inuence phase averages of microelds. The principal strength of windowing methods lies in providing an approach to studying the linear behavior of non-periodic volume elements.

78

79

Chapter 9 Multi-Scale Models

The micromechanical methods discussed in chapters 3, 6, and 8 are designed to handle a single scale transition between a lower and a higher length scale (microscale and macroscale), overall responses being obtained by homogenization and/or local elds by localization. Many inhomogeneous materials, however, show more than two clearly distinct characteristic length scales, typical examples being laminated and woven composites, materials in which there are well dened clusters of particles, as well as most biomaterials. In such cases an obvious modeling strategy is a hierarchical (or multi-scale) approach that uses a sequence of scale transitions, i.e., the material response at any given length scale is described on the basis of the homogenized behavior of the next lower one47 as depicted schematically in g. 9.1.

scale transition #1

scale transition #2

MACROSCALE (sample)

MESOSCALE (particle clusters)

MICROSCALE (particles in matrix)

Figure 9.1: Schematic representation of a multi-scale approach to studying a material consisting of clustered inhomogeneities in a matrix. Two scale transitions, macromeso and mesomicro, are used. A multi-scale model can be viewed as a sequence of scale transitions and suitable micromechanical models, i.e., mean eld, unit cell, windowing and, to some extent, embedding approaches, may be used as building blocks at any level within hierarchical schemes. Such
Describing the material behavior at lower length scales by a homogenized model implies that characteristic lengths dier by, say, two orders of magnitude or more, so that valid volumes can be dened for homogenization. It is not technically correct to employ hierarchical approaches within bands of more or less less continuous distributions of length scales, as can be found, e.g., in some metallic foams with highly disperse cell sizes.
47

80

multi-scale modeling strategies have the additional advantage of allowing the behavior of the constituents at all lower length scales to be assessed via the corresponding localization relations. Among the continuum mechanical multi-scale descriptions of the thermomechanical behavior of inhomogeneous materials reported in the literature, some combine mean eld methods at the higher length scale with mean eld (Hu et al., 1998; Tszeng, 1998) or periodic microeld approaches (Gonz alez and LLorca, 2000) at the lower length scale. The most common strategy for multi-scale modeling, however, uses Finite Element based unit cell or embedding methods at the topmost length scale, which implies that the homogenized material models describing the lower level(s) of the hierarchy must take the form of micromechanically based constitutive laws that can be evaluated at each integration point. This requirement typically does not give rise to extreme computational workloads in the elastic range, where the superposition principle can be used to obtain the full homogenized elasticity and thermal expansion tensors with a limited number of modeling runs. For simulating the thermomechanical response of nonlinear inhomogeneous materials, however, essentially a full micromechanical submodel has to be maintained and solved at each integration point in order to account for the history dependence of the local responses48 . Although within such a framework the use of sophisticated unit cell based models at the lower length scales tends to be an expensive proposition in terms of computational requirements, considerable work of this type has been reported (Belsky et al., 1995; Ghosh et al., 1996; Smit et al., 1998; Feyel, 2003). Lower (but by no means negligible) computational costs can be achieved by using constitutive models based on mean eld approaches, e.g., incremental MoriTanaka methods. In recent advanced models the type of micromechanical model used for the lowest scale transition is automatically adapted to the local geometry and/or gradients of the elds, embedding methods being used at the most critical locations (Raghavan and Ghosh, 2004). Figure 9.2 shows a result obtained by applying a multiscale approach that uses an incremental MoriTanaka method, compare section 3.7, at the integration point level of a meso-scale unit cell model for describing the elastoplastic behavior of a cluster-structured high speed steel. The particle-rich and particle poor regions used for the description at the mesoscale are treated as particle reinforced MMCs with appropriate reinforcement volume fractions. Such an approach not only predicts macroscopic stressstrain curves, but also allows the mesoscopic distributions of phase averaged microscopic variables to be evaluated, compare Plankensteiner (2000). Finally, it is worth noting that multi-scale approaches are not limited to using the standard methods of continuum micromechanics as discussed above. Especially the capability of the Finite Element method of handling highly complex constitutive descriptions has been used to build multi-scale descriptions that employ, among others, material models based on crystal plasticity (McHugh et al., 1993), discrete dislocation plasticity (Cleveringa et al., 1997) and atomistics (Tadmor et al., 2000).

Some approaches, however, have aimed a relaxing this requirement by using suitable precomputed results for the lower homogenization step, compare G anser (1998).

48

81

Eps_eff,p_(m) 3.5000E-03 3.0000E-03 2.5000E-03 2.0000E-03 1.5000E-03

SCALAR MIN: -6.8514E-04 SCALAR MAX: 1.3667E-02

Figure 9.2: Phase averaged microscopic equivalent plastic strains in the matrix within the inhomogeneity-poor regions of a cluster-structured high speed steel under mechanical loading as predicted by a mesoscopic unit cell model combined with an incremental Mori Tanaka model at the microscale (Plankensteiner, 2000).

82

Chapter 10 Closing Remarks

Methods of continuum micromechanics of materials have enjoyed considerable success in the past ve decades in furthering the understanding of the thermomechanical behavior of inhomogeneous materials and in providing predictive tools for engineers and materials scientists. However, they are subject to some practical limitations that should be kept in mind when employing them. All the methods discussed in chapters 3 to 8 implicitly use the assumption that the constituents of the inhomogeneous material to be studied can be treated as homogeneous, which, of course, is not necessarily true. When the length scale of the inhomogeneities in a constituent is much smaller than the length scale of the phase arrangement to be studied, multi-scale models as discussed in section 9 can be used. No rigorous analytical theory appears to be available at present for handling scale transitions in materials that do not fulll this requirement (e.g., particle reinforced composites in which the grain size of the matrix is comparable to or even larger than the size of the particles). Typically the best that can be done in such cases is either to use suciently large models with resolved microgeometries (a computationally expensive proposition) or to employ homogenized phase properties and be aware of the approximation that is introduced. A major practical diculty in the use of continuum micromechanics approaches (which in many cases is closely related to the questions mentioned above) has been obtaining appropriate constitutive models and material parameters for the constituents. Typically, available data pertain to the behavior of the bulk materials (as measured from homogeneous macroscopic samples), whereas what is actually required are parameters (and, in some cases, constitutive theories) describing the in-situ response of the phases at the microscale. In fact, the dearth of dependable material parameters is one of the reasons why predictions of the strength of inhomogeneous materials by micromechanical methods tend to be a considerable challenge. Another point that should be kept in mind is that continuum micromechanical descriptions in most cases do not have absolute length scales unless these are introduced via the constitutive models of the constituents. Absolute length scales can be provided explicitly via discrete dislocation models, via gradient or nonlocal constitutive laws and damage models, or implicitly, e.g., by adjusting the phase material parameters to account for grain sizes via the HallPetch eect. Analysts should also be aware that absolute length 83

scales may be introduced inadvertently into a model by mesh dependence eects of discretizing numerical methods, an important issue being strain localization due to softening material behavior of a constituent. When multi-scale models are used special care may be necessary to avoid introducing inconsistent length scales at dierent modeling levels. In addition it is worth noting that usually the macroscopic response of inhomogeneous materials is much less sensitive to the phase arrangement (and to modeling approximations) than are the distributions of the microelds. Consequently, whereas good agreement in the overall behavior of a given model with benchmark theoretical results or experimental data typically indicates that the phase averages of stresses and strains are described satisfactorily, this does not necessarily imply that the stress and strain distributions have been captured correctly. It is important to be aware that work in the eld of micromechanics of materials invariably involves nding viable compromises in terms of the complexity of the models, which, on the one hand, have to be able to account (at least approximatively) for the physical phenomena relevant to the given problem, and, on the other hand, must be suciently simple to allow solutions to be obtained within the relevant constraints of time, cost, and computational resources. Obviously, actual problems cannot be solved without recourse to various approximations and tradeos the important point is to be aware of them. It is worth keeping in mind that there is no such thing as a best micromechanical approach to all applications and that models, while indispensable in understanding the behavior of inhomogeneous materials, are just models and do not reect the full complexities of real materials.

84

Bibliography
J. Aboudi. The generalized method of cells and high-delity generalized method of cells micromechanical models A review. Mech.Adv.Mater.Struct., 11:329366, 2004. J. Aboudi. Micromechanical analysis of composites by the method of cells. Appl.Mech.Rev., 42:193221, 1989. J.D. Achenbach and H. Zhu. Eect of interfacial zone on mechanical behavior and failure of ber-reinforced composites. J.Mech.Phys.Sol., 37:381393, 1989. D.F. Adams and D.A. Crane. Finite element micromechanical analysis of a unidirectional composite including longitudinal shear loading. Comput.Struct., 18:11531165, 1984. D.F. Adams and D.R. Doner. Transverse normal loading of a uni-directional composite. J.Compos.Mater., 1:152164, 1967. D.H. Allen and J.W. Lee. The eective thermoelastic properties of whisker-reinforced composites as functions of material forming parameters. In G.J. Weng, M. Taya, and H. Ab e, editors, Micromechanics and Inhomogeneity, pages 1740, New York, NY, 1990. SpringerVerlag. R.T. Arenburg and J.N. Reddy. Analysis of metal-matrix composite structures I. Micromechanics constitutive theory. Comput.Struct., 40:13571368, 1991. L. Babout, Y. Br echet, E. Maire, and R. Foug` eres. On the competition between particle fracture and particle decohesion in metal matrix composites. Acta mater., 52:45174525, 2004. P.K. Banerjee and D.P. Henry. Elastic analysis of three-dimensional solids with ber inclusions by BEM. Int.J.Sol.Struct., 29:24232440, 1992. L. Banks-Sills and V. Leiderman. Macro-mechanical material model for ber-reinforced metal matrix composites. In A.S. Khan, editor, Physics and Mechanics of Finite Plastic and Viscoplastic Deformation, pages 183184, Fulton, MD, 1997. NEAT Press. G. Bao, J.W. Hutchinson, and R.M. McMeeking. Particle reinforcement of ductile matrices against plastic ow and creep. Acta metall.mater., 39:18711882, 1991. V. Belsky, M.W. Beall, J. Fish, M.S. Shephard, and S. Gomaa. Computer-aided multiscale modeling tools for composite materials and structures. Int.J.Comput.Syst.Engng., 6: 213223, 1995.

85

Y. Benveniste. A new approach to the application of MoriTanakas theory in composite materials. Mech.Mater., 6:147157, 1987. Y. Benveniste. Some remarks on three micromechanical models in composite media. J.Appl.Mech., 57:474476, 1990. Y. Benveniste and G.J. Dvorak. On a correspondence between mechanical and thermal eects in two-phase composites. In G.J. Weng, M. Taya, and H. Ab e, editors, Micromechanics and Inhomogeneity, pages 6582, New York, NY, 1990. SpringerVerlag. Y. Benveniste, G.J. Dvorak, and T. Chen. On diagonal and elastic symmetry of the approximate eective stiness tensor of heterogeneous media. J.Mech.Phys.Sol., 39: 927946, 1991. M.J. Beran and J. Molyneux. Use of classical variational principles to determine bounds for the eective bulk modulus in heterogeneous media. Quart.Appl.Math., 24:107118, 1966. J.G. Berryman. Long-wavelength propagation in composite elastic media, II. Ellipsoidal inclusions. J.Acoust.Soc.Amer., 68:18201831, 1980. M. Berveiller and A. Zaoui. A simplied self-consistent scheme for the plasticity of twophase metals. Res.Mech.Lett., 1:119124, 1981. J.F.W. Bishop and R. Hill. A theory of the plastic distortion of a polycrystalline aggregate under combined stress. Phil.Mag., 42:414427, 1951. H.J. B ohm and W. Han. Comparisons between three-dimensional and two-dimensional multi-particle unit cell models for particle reinforced metal matrix composites. Modell.Simul.Mater.Sci.Engng., 9:4765, 2001. H.J. B ohm and F.G. Rammerstorfer. Fiber arrangement eects on the microscale stresses of continuously reinforced MMCs. In R. Pyrz, editor, MicrostructureProperty Interactions in Composite Materials, pages 5162, Dordrecht, The Netherlands, 1995. Kluwer Academic Publishers. H.J. B ohm, F.G. Rammerstorfer, and E. Weissenbek. Some simple models for micromechanical investigations of ber arrangement eects in MMCs. Comput.Mater.Sci., 1: 177194, 1993. H.J. B ohm, A. Eckschlager, and W. Han. Modeling of phase arrangement eects in high speed tool steels. In F. Jeglitsch, R. Ebner, and H. Leitner, editors, Tool Steels in the Next Century, pages 147156, Leoben, 1999. Montanuniversit at Leoben. H.J. B ohm, A. Eckschlager, and W. Han. Multi-inclusion unit cell models for metal matrix composites with randomly oriented discontinuous reinforcements. Comput.Mater.Sci., 25:4253, 2002. M. Bornert, E. Herv e, C. Stolz, and A. Zaoui. Self consistent approaches and strain heterogeneities in two-phase elastoplastic materials. Appl.Mech.Rev., 47:66S76, 1994.

86

M. Bornert, T. Bretheau, and P. Gilormini, editors. Homog en eisation en m ecanique des mat eriaux. Editions Herm` es, Paris, 2001. L.M. Brown and W.M. Stobbs. The work-hardening of coppersilica. I. A model based on internal stresses, with no plastic relaxation. Phil.Mag., 23:11851199, 1971. V.N. Bulsara, R. Talreja, and J. Qu. Damage initiation under transverse loading of unidirectional composites with arbitrarily distributed bers. Compos.Sci.Technol., 59:673 682, 1999. N. Chawla and K.K. Chawla. Microstructure based modeling of the deformation behavior of particle reinforced metal matrix composites. J.Mater.Sci., 41:913925, 2006. T. Chen. Exact moduli and bounds of two-phase composites with coupled multield linear responses. J.Mech.Phys.Sol., 45:385398, 1997. C.M. Chimani, H.J. B ohm, and F.G. Rammerstorfer. On stress singularities at free edges of bimaterial junctions A micromechanical study. Scr.mater., 36:943947, 1997. R.M. Christensen and K.H. Lo. Solutions for eective shear properties in three phase sphere and cylinder models. J.Mech.Phys.Sol., 27:315330, 1979. R.M. Christensen and K.H. Lo. Erratum to Christensen and Lo, 1979. J.Mech.Phys.Sol., 34:639, 1986. R.M. Christensen, H. Schantz, and J. Shapiro. On the range of validity of the moritanaka method. J.Mech.Phys.Sol., 40:6973, 1992. H.H.M. Cleveringa, E. van der Giessen, and A. Needleman. Comparison of discrete dislocation and continuum plasticity predictions for a composite material. Acta mater., 45: 31633179, 1997. M.E. Cruz and A.T. Patera. A parallel Monte-Carlo nite element procedure for the analysis of multicomponent random media. Int.J.Num.Meth.Engng., 38:10871121, 1995. L.C. Davis. Flow rule for the plastic deformation of particulate metal matrix composites. Comput.Mater.Sci., 6:310318, 1996. T. Daxner. Computational Simulation of the Thermal Conductivity of MMCs under Consideration of the InclusionMatrix Interface. Reihe 18, Nr.285. VDIVerlag, D usseldorf, FRG, 2003. I. Doghri and A. Ouaar. Homogenization of two-phase elasto-plastic composite materials and structures. Int.J.Sol.Struct., 40:16811712, 2003. M. Dong and S. Schmauder. Modeling of metal matrix composites by a self-consistent embedded cell model. Acta mater., 44:24652478, 1996. K. Dorninger. Entwicklung von nichtlinearen FE-Algorithmen zur Berechnung von Schalenkonstruktionen aus Faserverbundstoen. Reihe 18, Nr.65. VDIVerlag, D usseldorf, FRG, 1989.

87

W.J. Drugan and J.R. Willis. A micromechanics-based nonlocal constitutive equation and estimates of representative volume element size for elastic composites. J.Mech.Phys.Sol., 44:497524, 1996. M.L. Dunn and H. Ledbetter. Elasticplastic behavior of textured short-ber composites. Acta mater., 45:33273340, 1997. M.L. Dunn and M. Taya. Micromechanics predictions of the eective electroelastic moduli of piezoelectric composites. Int.J.Sol.Struct., 30:161175, 1993. G.J. Dvorak. Transformation eld analysis of inelastic composite materials. Proc.Roy.Soc.London, A437:311327, 1992. G.J. Dvorak and Y.A. Bahei-el Din. Plasticity analysis of brous composites. J.Appl.Mech., 49:327335, 1982. G.J. Dvorak, Y.A. Bahei-el Din, and A.M. Wafa. The modeling of inelastic composite materials with the transformation eld analysis. Modell.Simul.Mater.Sci.Engng., 2:571 586, 1994. F. El Houdaigui, S. Forest, A.F. Gourgues, and D. Jeulin. On the size of the representative volume element for isotropic elastic copper. In Y.L. Bai and Q.S. Zheng, editors, IUTAM Symposium on Mechanical Behavior and Micro-Mechanics of Nanostructures Materials, pages 171180, Dordrecht, 2007. SpringerVerlag. J.D. Eshelby. The determination of the elastic eld of an ellipsoidal inclusion and related problems. Proc.Roy.Soc.London, A241:376396, 1957. J.D. Eshelby. The elastic eld outside an ellipsoidal inclusion. Proc.Roy.Soc.London, A252: 561569, 1959. M. Ferrari. Asymmetry and the high concentration limit of the MoriTanaka eective medium theory. Mech.Mater., 11:251256, 1991. F. Feyel. A multilevel nite element method (FE2 ) to describe the response of highly non-linear structures using generalized continua. Comput.Meth.Appl.Mech.Engng., 192: 32333244, 2003. H.F. Fischmeister and B. Karlsson. Plastizit atseigenschaften grob-zweiphasiger Werkstoe. Z.Metallkd., 68:311327, 1977. C. Fond, A. Riccardi, R. Schirrer, and F. Montheillet. Mechanical interaction between spherical inhomogeneities: An assessment of a method based on the equivalent inclusion. Eur.J.Mech. A/Solids, 20:5975, 2001. S.Y. Fu and B. Lauke. The elastic modulus of misaligned short-ber-reinforced polymers. Compos.Sci.Technol., 58:389400, 1998. H.P. G anser. Large Strain Behavior of Two-Phase Materials. Reihe 5, Nr.528. VDIVerlag, D usseldorf, FRG, 1998.

88

M. G ar ajeu, J.C. Michel, and P. Suquet. A micromechanical approach of damage in viscoplastic materials by evolution in size, shape and distribution of voids. Comput.Meth.Appl.Mech.Engng., 183:223246, 2000. A.C. Gavazzi and D.C. Lagoudas. On the numerical evaluation of Eshelbys tensor and its application to elastoplastic brous composites. Comput.Mech., 7:1219, 1990. S. Ghosh, K.H. Lee, and S. Moorthy. Two scale analysis of heterogeneous elastic-plastic materials with asymptotic homogenization and Voronoi cell nite element model. Comput.Meth.Appl.Mech.Engng., 132:63116, 1996. L.J. Gibson. The mechanical behavior of cancellous bone. J.Biomech., 18:317328, 1985. L.J. Gibson and M.F. Ashby. Cellular Solids: Structure and Properties. Pergamon Press, Oxford, UK, 1988. M. Gologanu, J.B. Leblond, G. Perrin, and J. Devaux. Recent extensions of Gursons model for porous ductile materials. In P. Suquet, editor, Continuum Micromechanics, pages 61130, Vienna, 1997. SpringerVerlag, CISM Courses and Lectures Vol. 377. C. Gonz alez and J. LLorca. A self-consistent approach to the elasto-plastic behavior of two-phase materials including damage. J.Mech.Phys.Sol., 48:675692, 2000. C. Gonz alez and J. LLorca. Virtual fracture testing of composites: A computational micromechanics approach. Engng.Fract.Mech., 74:11261138, 2007. J.L. Grenestedt. Inuence of wavy imperfections in cell walls on elastic stiness of cellular solids. J. Mech. Phys. Sol., 46:2950, 1998. R.E. Guldberg, S.J. Hollister, and G.T. Charras. The accuracy of digital image-based nite element models. J.Biomech.Engng., 120:289295, 1998. G. Guo, J. Fitoussi, D. Baptiste, N. Sicot, and C. Wol. Modelling of damage behavior of a short-ber reinforced composite structure by the nite element analysis using a micromacro law. Int.J.Dam.Mech., 6:278299, 1997. A.L. Gurson. Continuum theory of ductile rupture by void nucleation and growth: Part I Yield criteria and ow rules for porous ductile media. J. Engng. Mater. Technol., 99: 215, 1977. A.A. Gusev. Representative volume element size for elastic composites: A numerical study. J.Mech.Phys.Sol., 45:14491459, 1997. J.C. Halpin and J.L. Kardos. The HalpinTsai equations: A review. Polym.Engng.Sci., 16:344351, 1976. Z. Hashin. The elastic moduli of heterogeneous materials. J.Appl.Mech., 29:143150, 1962. Z. Hashin. Analysis of composite materials A survey. J.Appl.Mech., 50:481505, 1983. Z. Hashin. The dierential scheme and its application to cracked materials. J.Mech.Phys.Sol., 36:719733, 1988. 89

Z. Hashin and B.W. Rosen. The elastic moduli of ber-reinforced materials. J.Appl.Mech., 31:223232, 1964. Z. Hashin and S. Shtrikman. A variational approach to the theory of the elastic behavior of multiphase materials. J.Mech.Phys.Sol., 11:127140, 1963. Z. Hashin and S. Shtrikman. On some variational principles in anisotropic and nonhomogeneous elasticity. J.Mech.Phys.Sol., 10:335342, 1962. B. Hassani and E. Hinton. Homogenization and Structural Topology Optimization. SpringerVerlag. London, 1999. H. Hatta and M. Taya. Equivalent inclusion method for steady state heat conduction in composites. Int.J.Engng.Sci., 24:11591172, 1986. S. Hazanov. Hill condition and overall properties of composites. Arch. Appl. Mech., 68: 385394, 1998. S. Hazanov and M. Amieur. On overall properties of elastic bodies smaller than the representative volume. Int. J. Engng. Sci., 33:12891301, 1995. R. Hill. Elastic properties of reinforced solids: J.Mech.Phys.Sol., 11:357372, 1963. Some theoretical principles.

R. Hill. Theory of mechanical properties of bre-strengthened materials: I. Elastic behaviour. J.Mech.Phys.Sol., 12:199212, 1964. R. Hill. A self-consistent mechanics of composite materials. J.Mech.Phys.Sol., 13:213222, 1965. S.J. Hollister, D.P. Fyhrie, K.J. Jepsen, and S.A. Goldstein. Application of homogenization theory to the study of trabecular bone mechanics. J.Biomech., 24:825839, 1991. S. Holmberg, K. Persson, and H. Petersson. Nonlinear mechanical behaviour and analysis of wood and bre materials. Comput.Struct., 72:459480, 1999. C.L. Hom and R.M. McMeeking. Plastic ow in ductile materials containing a cubic array of rigid spheres. Int.J.Plast., 7:255274, 1991. G.K. Hu, G. Guo, and D. Baptiste. A micromechanical model of inuence of particle fracture and particle cluster on mechanical properties of metal matrix composites. Comput.Mater.Sci., 9:420430, 1998. J.H. Huang. Some closed-form solutions for eective moduli of composites containing randomly oriented short bers. Mater.Sci.Engng.A, 315:1120, 2001. Y. Huang and K.X. Hu. A generalized self-consistent mechanics method for solids containing elliptical inclusions. J.Appl.Mech., 62:566572, 1995. C. Huet. Coupled size and boundary-condition eects in viscoelastic heterogeneous and composite bodies. Mech. Mater., 31:787829, 1999.

90

C. Huet, P. Navi, and P.E. Roelfstra. A homogenization technique based on Hills modication theorem. In G.A. Maugin, editor, Continuum Models and Discrete Systems, pages 135143, Harlow, UK, 1990. Longman. J.W. Hutchinson. Elastic-plastic behavior of polycrystalline metals and composites. Proc.Roy.Soc.London, A319:247272, 1970. D. Jeulin. Random structure models for homogenization and fracture statistics. In D. Jeulin and M. Ostoja-Starzewski, editors, Mechanics of Random and Multiscale Microstructures, pages 3391, Vienna, 2001. SpringerVerlag, CISM Courses and Lectures Vol. 430. M. Jiang, M. Ostoja-Starzewski, and I. Jasiuk. Scale-dependent bounds on eective elastoplastic response of random composites. J.Mech.Phys.Sol., 49:655673, 2001. J.W. Ju and L.Z. Sun. A novel formulation for the exterior point Eshelbys tensor of an ellipsoidal inclusion. J.Appl.Mech., 66:570574, 1999. M. Kailasam, N. Aravas, and P. Ponte Casta neda. Porous metals with developing anisotropy: Constitutive models, computational issues and applications to deformation processing. Comput.Model.Engng.Sci., 1:105118, 2000. T. Kanit, S. Forest, I. Gallier, V. Mounoury, and D. Jeulin. Determination of the size of the representative volume element for random composites: Statistical and numerical approach. Int.J.Sol.Struct., 40:36473679, 2003. M. Karayaka and H. Sehitoglu. Thermomechanical deformation modeling of Al2xxxT4/SiCp composites. Acta metall.mater., 41:175189, 1993. D.C. Lagoudas, A.C. Gavazzi, and H. Nigam. Elastoplastic behavior of metal matrix composites based on incremental plasticity and the MoriTanaka averaging scheme. Comput.Mech., 8:193203, 1991. V.M. Levin. On the coecients of thermal expansion of heterogeneous materials. Mech.Sol., 2:5861, 1967. A. Levy and J.M. Papazian. Elastoplastic nite element analysis of short-ber-reinforced SiC/Al composites: Eects of thermal treatment. Acta metall.mater., 39:22552266, 1991. M. Li, S. Ghosh, O. Richmond, H. Weiland, and T.N. Rouns. Three dimensional characterization and modeling of particle reinforced metal matrix composites, Part I: Quantitative description of microstructural morphology. Mater.Sci.Engng.A, 265:153173, 1999. J.M. Mahishi. An integrated micromechanical and macromechanical approach to fracture behavior of ber-reinforced composites. Engng.Fract.Mech., 25:197228, 1986. F. Marketz and F.D. Fischer. Micromechanical modelling of stress-assisted martensitic transformation. Modell.Simul.Mater.Sci.Engng., 2:10171046, 1994. K. Markov. Elementary micromechanics of heterogeneous media. In K. Markov and L. Preziosi, editors, Heterogeneous Media: Micromechanics Modeling Methods and Simulations, pages 1162, Boston, MA, 2000. Birkh auser. 91

P.E. McHugh, R.J. Asaro, and C.F. Shih. Computational modeling of metal-matrix composite materials I. Isothermal deformation patterns in ideal microstructures. Acta metall.mater., 41:14611476, 1993. R. McLaughlin. A study of the dierential scheme for composite materials. Int.J.Engng.Sci., 15:237244, 1977. J.C. Michel, H. Moulinec, and P. Suquet. Eective properties of composite materials with periodic microstructure: A computational approach. Comput.Meth.Appl.Mech.Engng., 172:109143, 1999. C.A. Miller and S. Torquato. Eective conductivity of hard sphere suspensions. J.Appl.Phys., 68:54865493, 1990. T. Miloh and Y. Benveniste. A generalized self-consistent method for the eective conductivity of composites with ellipsoidal inclusions and cracked bodies. J.Appl.Phys., 63: 789796, 1988. G.W. Milton. The Theory of Composites. Cambridge University Press, Cambridge, 2002. G.W. Milton. Bounds on the electromagnetic, elastic, and other properties of twocomponent composites. Phys.Rev.Lett., 46:542545, 1981. B. Mlekusch. Thermoelastic properties of short-bre-reinforced thermoplastics. Compos.Sci.Technol., 59:911923, 1999. T. Mori and K. Tanaka. Average stress in the matrix and average elastic energy of materials with mistting inclusions. Acta metall., 21:571574, 1973. C. Motz, R. Pippan, A. Ableidinger, H.J. B ohm, and F.G. Rammerstorfer. Deformation and fracture behavior of ductile aluminium foams in the presence of notches under tensile loading. In J. Banhart, M.F. Ashby, and N.A. Fleck, editors, Cellular Metals and Metal Foaming Technology, pages 299304, Bremen, FRG, 2001. Verlag MIT Publishing. H. Moulinec and P. Suquet. A fast numerical method for computing the linear and nonlinear mechanical properties of composites. C.R.Acad.Sci.Paris, s erie II, 318:14171423, 1994. W.H. M uller. Mathematical versus experimental stress analysis of inhomogeneities in solids. J.Phys.IV, 6:1139C1148, 1996. T. Mura. Micromechanics of Defects in Solids. Martinus Nijho, Dordrecht, 1987. T. Nakamura and S. Suresh. Eects of thermal residual stresses and ber packing on deformation of metal-matrix composites. Acta metall.mater., 41:16651681, 1993. S. Nemat-Nasser. Averaging theorems in nite deformation plasticity. Mech.Mater., 31: 493523, 1999. S. Nemat-Nasser and M. Hori. Micromechanics: Overall Properties of Heterogeneous Solids. NorthHolland, Amsterdam, 1993.

92

G.L. Niebur, J.C. Yuen, A.C. Hsia, and T.M. Keaveny. Convergence behavior of highresolution nite element models of trabecular bone. J.Biomech.Engng., 121:629635, 1999. A.N. Norris. A dierential scheme for the eective moduli of composites. Mech.Mater., 4: 116, 1985. J.F. Nye. Physical Properties of Crystals, Their Representation by Tensors and Matrices. Clarendon, Oxford, UK, 1957. M. Ostoja-Starzewski. Material spatial randomness: From statistical to representative volume element. Probab. Engng. Mech., 21:112131, 2006. M. Ostoja-Starzewski. Towards scale-dependent constitutive laws for plasticity and fracture of random heterogeneous materials. In A. Pineau and A. Zaoui, editors, Micromechanics of Plasticity and Damage of Multiphase Materials, pages 379386, Dordrecht, The Netherlands, 1996. Kluwer. D.H. Pahr and S.M. Arnold. The applicability of the generalized method of cells for analyzing discontinuously reinforced composites. Composites B, 33:153170, 2002. D.H. Pahr and H.J. B ohm. Assessment of mixed uniform boundary conditions for predicting the mechanical behavior of elastic and inelastic discontinuously reinforced composites. Comput. Model. Engng. Sci., 34:117136, 2008. D.H. Pahr and P.K. Zysset. Inuence of boundary conditions on computed apparent elastic properties of cancellous bone. Biomech. Model. Mechanobiol., 7:463476, 2008. S.D. Papka and S. Kyriakides. In-plane compressive response and crushing of honeycomb. J.Mech.Phys.Sol., 42:14991532, 1994. O.B. Pedersen. Thermoelasticity and plasticity of composites I. Mean eld theory. Acta metall., 31:17951808, 1983. H.E. Pettermann. Derivation and Finite Element Implementation of Constitutive Material Laws for Multiphase Composites Based on MoriTanaka Approaches. Reihe 18, Nr.217. VDIVerlag, D usseldorf, 1997. H.E. Pettermann and S. Suresh. A comprehensive unit cell model: A study of coupled eects in piezoelectric 13 composites. Int.J.Sol.Struct., 37:54475464, 2000. H.E. Pettermann, H.J. B ohm, and F.G. Rammerstorfer. Some direction dependent properties of matrixinclusion type composites with given reinforcement orientation distributions. Composites B, 28:253265, 1997. H.E. Pettermann, A.F. Plankensteiner, H.J. B ohm, and F.G. Rammerstorfer. A thermoelasto-plastic constitutive law for inhomogeneous materials based on an incremental MoriTanaka approach. Comput.Struct., 71:197214, 1999. N. Phan-Thien and G.W. Milton. New third-order bounds on the eective moduli of n-phase composites. Quart.Appl.Math., 41:5974, 1983. 93

A.F. Plankensteiner. Multiscale Treatment of Heterogeneous Nonlinear Solids and Structures. Reihe 18, Nr.248. VDIVerlag, D usseldorf, FRG, 2000. P. Ponte Casta neda. Bounds and estimates for the properties on nonlinear inhomogeneous systems. Phil.Trans.Roy.Soc., A340:531567, 1992. P. Ponte Casta neda and P. Suquet. Nonlinear composites. In E. van der Giessen and T.Y. Wu, editors, Advances in Applied Mechanics 34, pages 171302, New York, NY, 1998. Academic Press. P. Ponte Casta neda and J.R. Willis. The eect of spatial distribution on the eective behavior of composite materials and cracked media. J.Mech.Phys.Sol., 43:19191951, 1995. G.L. Povirk, S.R. Nutt, and A. Needleman. Analysis of creep in thermally cycled Al/SiC composites. Scr.metall.mater., 26:46166, 1992. R. Pyrz and B. Bochenek. Topological disorder of microstructure and its relation to the stress eld. Int.J.Sol.Struct., 35:24132427, 1998. Y.P. Qiu and G.J. Weng. A theory of plasticity for porous materials and particle-reinforced composites. J.Appl.Mech., 59:261268, 1992. P. Raghavan and S. Ghosh. Adaptive multi-scale modeling of composite materials. Comput.Model.Engng.Sci., 5:151170, 2004. T.J. Reiter and G.J. Dvorak. Micromechanical models for graded composite materials: II. Thermomechanical loading. J.Mech.Phys.Sol., 46:16551673, 1998. A. Reuss. Berechnung der Fliegrenze von Mischkristallen auf Grund der Plastizit atsbedingung f ur Einkristalle. ZAMM, 9:4958, 1929. J.R. Rice and D.M. Tracey. On the ductile enlargement of voids in triaxial stress elds. J.Mech.Phys.Sol., 17:201217, 1969. M. Rintoul and S. Torquato. Reconstruction of the structure of dispersions. J. Colloid Interf. Sci., 186:467476, 1997. G.J. Rodin. The overall elastic response of materials containing spherical inhomogeneities. Int.J.Sol.Struct., 30:18491863, 1993. A. Sangani and W. Lu. Elastic coecients of composites containing spherical inclusions in a periodic array. J.Mech.Phys.Sol., 35:121, 1987. S.P. Santosa and T. Wierzbicki. On the modeling of crush behavior of a closed-cell aluminum foam structure. J.Mech.Phys.Sol., 46:645669, 1998. M. Sautter, C. Dietrich, M.H. Poech, S. Schmauder, and H.F. Fischmeister. Finite element modelling of a transverse-loaded bre composite: Eects of section size and net density. Comput.Mater.Sci., 1:225233, 1993.

94

S. Schmauder, J. Wulf, T. Steinkop, and H. Fischmeister. Micromechanics of plasticity and damage in an Al/SiC metal matrix composite. In A. Pineau and A. Zaoui, editors, Micromechanics of Plasticity and Damage of Multiphase Materials, pages 255262, Dordrecht, The Netherlands, 1996. Kluwer. J. Segurado, E. Parteder, A. Plankensteiner, and H.J. B ohm. Micromechanical studies of the densication of porous molybdenum. Mater. Sci. Engng., A333:270278, 2002. J. Segurado, C. Gonz alez, and J. LLorca. A numerical investigation of the eect of particle clustering on the mechanical properties of composites. Acta mater., 51:23552369, 2003. M.S. Shephard, M.W. Beall, R. Garimella, and R. Wentorf. Automatic construction of 3/D models in multiple scale analysis. Comput.Mech., 17:196207, 1995. N. Silnutzer. Eective Constants of Statistically Homogeneous Materials. PhD thesis, University of Pennsylvania, Philadelphia, PA, 1972. A.E. Simone and L.J. Gibson. Deformation characteristics of metal foams. Acta mater., 46:21392150, 1998. R.J.M. Smit, W.A.M. Brekelmans, and H.E.H. Meijer. Prediction of the mechanical behavior of non-linear heterogeneous systems by multi-level nite element modeling. Comput.Meth.Appl.Mech.Engng., 155:181192, 1998. R.J.M. Smit, W.A.M. Brekelmans, and H.E.H. Meijer. Prediction of the large-strain mechanical response of heterogeneous polymer systems: Local and global deformation behaviour of a representative volume of voided polycarbonate. J.Mech.Phys.Sol., 47:201 221, 1999. N.J. Srensen, S. Suresh, V. Tvergaard, and A. Needleman. Eects of reinforcement orientation on the tensile response of metal matrix composites. Mater.Sci.Engng.A, 197: 110, 1995. P. Suquet. Eective properties of nonlinear composites. In P. Suquet, editor, Continuum Micromechanics, pages 197264, Vienna, 1997. SpringerVerlag. P.M. Suquet. Elements of homogenization for inelastic solid mechanics. In E. SanchezPalencia and A. Zaoui, editors, Homogenization Techniques in Composite Media, pages 194278, Berlin, 1987. SpringerVerlag. T. Suzuki and P.K.L. Yu. Complex elastic wave band structures in three-dimensional periodic elastic media. J.Mech.Phys.Sol., 46:115138, 1998. E.B. Tadmor, R. Phillips, and M. Ortiz. Hierarchical modeling in the mechanics of materials. Int.J.Sol.Struct., 37:379390, 2000. D.R.S. Talbot and J.R. Willis. Variational principles for inhomogeneous non-linear media. J.Appl.Math., 35:3954, 1985. G.P. Tandon and G.J. Weng. The eect of aspect ratio of inclusions on the elastic properties of unidirectionally aligned composites. Polym.Compos., 5:327333, 1984. 95

G.P. Tandon and G.J. Weng. A theory of particle-reinforced plasticity. J.Appl.Mech., 55: 126135, 1988. J.L. Teply and G.J. Dvorak. Bounds on overall instantaneous properties of elastic-plastic composites. J.Mech.Phys.Sol., 36:2958, 1988. K. Terada and N. Kikuchi. Microstructural design of composites using the homogenization method and digital images. Mater.Sci.Res.Int., 2:6572, 1996. K. Terada, T. Miura, and N. Kikuchi. Digital image-based modeling applied to the homogenization analysis of composite materials. Comput.Mech., 20:331346, 1997. K. Terada, I. Saiki, K. Matsui, and Y. Yamakawa. Two-scale kinematics and linearization for simultaneous two-scale analysis of periodic heterogeneous solids at nite strain. Comput.Meth.Appl.Mech.Engng., 192:35313563, 2003. K. Tohgo and T.W. Chou. Incremental theory of particulate-reinforced composites including debonding damage. JSME Int.J.Srs.A, 39:389397, 1996. S. Torquato. Random Heterogeneous Media. SpringerVerlag. New York, NY, 2002. S. Torquato. Random heterogeneous media: Microstructure and improved bounds on eective properties. Appl.Mech.Rev., 44:3775, 1991. S. Torquato. Morphology and eective properties of disordered heterogeneous media. Int.J.Sol.Struct., 35:23852406, 1998. S. Torquato and F. Lado. Improved bounds on the eective moduli of random arrays of cylinders. J.Appl.Mech., 59:16, 1992. S. Torquato, F. Lado, and P.A. Smith. Bulk properties of two-phase disordered media. IV. Mechanical properties of suspensions of penetrable spheres at nondilute concentrations. J.Chem.Phys., 86:63886392, 1987. T.C. Tszeng. The eects of particle clustering on the mechanical behavior of particle reinforced composites. Composites B, 29:299308, 1998. V. Tvergaard. Debonding of short bres among particulates in a metal matrix composite. Int.J.Sol.Struct., 40:69576967, 2003. V. Tvergaard. Fibre debonding and breakage in a whisker-reinforced metal. Mater.Sci.Engng.A, 190:215222, 1994. V. Tvergaard and A. Needleman. Analysis of the cup-cone fracture in a round tensile bar. Acta metall., 32:157169, 1984. E. van der Giessen and V. Tvergaard. Development of nal creep failure in polycrystalline aggregates. Acta metall.mater., 42:952973, 1994. K. V aradi, Z. N eder, K. Friedrich, and J. Fl ock. Finite-element analysis of a polymer composite subjected to ball indentation. Compos.Sci.Technol., 59:271281, 1999.

96

 W. Voigt. Uber die Beziehung zwischen den beiden Elasticit ats-Constanten isotroper K orper. Ann. Phys., 38:573587, 1889. K. Wakashima, H. Tsukamoto, and B.H. Choi. Elastic and thermoelastic properties of metal matrix composites with discontinuous bers or particles: Theoretical guidelines toward materials tailoring. In The KoreaJapan Metals Symposium on Composite Materials, pages 102115, Seoul, Korea, 1988. The Korean Institute of Metals. J. Wang, J.H. Andreasen, and B.L. Karihaloo. The solution of an inhomogeneity in a nite plane region and its application to composite materials. Compos.Sci.Technol., 60:7582, 2000. W.E. Warren and A.M. Kraynik. The nonlinear elastic properties of open-cell foams. J.Appl.Mech., 58:376381, 1991. D.F. Watt, X.Q. Xu, and D.J. Lloyd. Eects of particle morphology and spacing on the strain elds in a plastically deforming matrix. Acta mater., 44:789799, 1996. E. Weissenbek. Finite Element Modelling of Discontinuously Reinforced Metal Matrix Composites. Reihe 18, Nr.164. VDIVerlag, D usseldorf, Germany, 1994. E. Weissenbek, H.J. B ohm, and F.G. Rammerstorfer. Micromechanical investigations of arrangement eects in particle reinforced metal matrix composites. Comput.Mater.Sci., 3:263278, 1994. E. Weissenbek, H.E. Pettermann, and S. Suresh. Numerical simulation of plastic deformation in compositionally graded metalceramic structures. Acta mater., 45:34013417, 1997. G.J. Weng. The theoretical connection between MoriTanaka theory and the Hashin ShtrikmanWalpole bounds. Int.J.Engng.Sci., 28:11111120, 1990. J.R. Willis. The overall response of nonlinear composite media. Eur.J.Mech. A/Solids, 19: 165184, 2000. J.R. Willis. Bounds and self-consistent estimates for the overall moduli of anisotropic composites. J.Mech.Phys.Sol., 25:185202, 1977. P.J. Withers. The determination of the elastic eld of an ellipsoidal inclusion in a transversely isotropic medium, and its relevance to composite materials. Phil.Mag., A59: 759781, 1989. P.J. Withers, W.M. Stobbs, and O.B. Pedersen. The application of the Eshelby method of internal stress determination to short bre metal matrix composites. Acta metall., 37: 30613084, 1989. T.T Wu. The eect of inclusion shape on the elastic moduli of a two-phase material. Int.J.Sol.Struct., 2:18, 1966. J. Wulf, T. Steinkop, and H. Fischmeister. FE-simulation of crack paths in the real microstructure of an Al(6061)/SiC composite. Acta mater., 44:17651779, 1996. 97

 J. Zeman and M. Sejnoha. Numerical evaluation of eective elastic properties of graphite ber tow impregnated by polymer matrix. J.Mech.Phys.Sol., 49:6990, 2001. J. Zeman and M. Sejnoha. From random microstructures to representative volume elements. Modell.Simul.Mater.Sci.Engng., 15:325S335, 2007. R.W. Zimmerman. HashinShtrikman bounds on the Poisson ratio of a composite material. Mech.Res.Comm., 19:563569, 1992. T.I. Zohdi. A model for simulating the deterioration of structural-scale material responses of microheterogeneous solids. In D. Gross, F.D. Fischer, and E. van der Giessen, editors, Euromech Colloqium 402 Micromechanics of Fracture Processes, pages 9192, Darmstadt, FRG, 1999. TU Darmstadt. T.I. Zohdi and P. Wriggers. A model for simulating the deterioration of structural-scale material responses of microheterogeneous solids. Comput.Meth.Appl.Mech.Engng., 190: 28032823, 2001.

98