Professional Documents
Culture Documents
MICROMECHANICS
Helmut J. BOHM
hjb@ilsb.tuwien.ac.at
Definitions
Material Symmetries
Models Overview
Part I
Introduction
Definitions
Material Symmetries
Models Overview
Definitions
Material Symmetries
Models Overview
Introductory Remarks
Present lecture deals with numerical approaches to modeling
the material behavior of inhomogeneous materials using
continuum methods
referred to as Continuum Micromechanics
Main emphasis is put on
composite materials
thermoelastic and thermoelastoplastic material behavior
evaluation by Finite Element method
(quasi)static behavior
Subject is treated from an engineering point of view
mainly aimed at modeling aspects
Definitions
Material Symmetries
Models Overview
Basics
material is inhomogeneous:
at least two distinct constituents (phases)
material shows at least two distinct length scales
behavior at all relevant length scales can be described by
continuum mechanics
Definitions
Material Symmetries
Models Overview
Length Scales
Many inhomogeneous materials can be studied at a number of
different length scales
macroscale: sample or structure
microscale: distinguishable constituents
mesoscale(s): intermediate scale(s), e.g., laminae
MACRO (structural)
MICRO 1 (composite)
SUBMICRO 1 (nano)
MESO (laminate)
MICRO 2 (polycrystal)
SUBMICRO 2 (atomistic)
Definitions
Material Symmetries
Models Overview
Outline
Definitions
Material Symmetries
Models Overview
Definitions
Material Symmetries
Models Overview
Composites
Polycrystals
Graded materials
Definitions
Material Symmetries
Models Overview
Definitions
Material Symmetries
Models Overview
production related
(functionally) graded
monodisperse
bidisperse, polydisperse (described by size distribution)
10
Definitions
Material Symmetries
Models Overview
by inhomogeneity shape
(typically described via aspect ratio a)
Definitions
Material Symmetries
Models Overview
random
planar random
general (described via orientation distribution function,
ODF; texture)
by inhomogeneity material
12
Definitions
Material Symmetries
Models Overview
Classification of Composites
Classification (composites)
by matrix material
by matrix behavior
Definitions
Material Symmetries
Models Overview
Inhomogeneous Materials
14
Definitions
Material Symmetries
Models Overview
Inhomogeneous Materials
15
Definitions
Material Symmetries
Models Overview
Inhomogeneous Materials
16
Definitions
Material Symmetries
Models Overview
Inhomogeneous Materials
Polycrystals
17
Definitions
Material Symmetries
Models Overview
Inhomogeneous Materials
Metallic foams
18
Definitions
Material Symmetries
Models Overview
Inhomogeneous Materials
Trabecular bone
19
Definitions
Material Symmetries
Models Overview
Inhomogeneous Materials
20
Definitions
Material Symmetries
Models Overview
Nomenclature
(Elasticity-like) tensors of order 4:
E Eijkl
written as 66-quasi-matrices (Voigt/Nye notation)
symmetric (elasticities, . . .) or not (Eshelby tensors, . . .)
example: orthotropic elasticity tensor in principal
orientation
E=
0
0
0 E44 0
0
0
0
0
0 E55 0
0
0
0
0
0 E66
21
Definitions
Material Symmetries
Models Overview
Nomenclature
Strain- and stress-like tensors of order 2:
ij
written as 6-quasi-vectors (Voigt/Nye notation)
strains use shear angles ij = 2ij
(engineering notation)
11
11
11
22 22
22
33 33
33
=
=
212 = 12
12
213 13
13
23
223
23
Note: different authors may use different ordering of
components
22
Definitions
Material Symmetries
Models Overview
Nomenclature
Conductivity-like tensors of order 2:
K kij
example: orthotropic conductivity tensor
k11 0
0
K = 0 k22 0
0
0 k33
Physical vectors:
n ni
23
Definitions
Material Symmetries
Models Overview
Nomenclature
Most tensor operations are treated as matrix/vector operations
Additional operations
[a n]ij = ai nj (dyadic product)
[ ]ijkl =ij kl
[ n]i = ij nj (contraction)
24
Definitions
Material Symmetries
Models Overview
Nomenclature
Macroscopic and microscopic quantities
Meso/microlevel: phases denoted by superscripts
(m) (i)
, , ...
(p) (f)
25
Definitions
Material Symmetries
Models Overview
Scale Transitions
Homogenization (coarse graining):
small scale large scale
reduce number of DOFs of system
find energetically equivalent homogeneous material
Localization (fine graining):
large scale small scale
find local response to macroscopic (load) state
26
Definitions
Material Symmetries
Models Overview
Scale Transitions
Homogenization relations: typically involve volume averaging
continuous displacements required for surface integral
formulation for hi
Z
Z
1
1
(x) d =
u(x) n + n u(x) d
hi =
s s
2s s
Z
Z
1
1
hi =
(x) d =
t(x) x d
s s
s s
Formal localization relations
(position dependent) concentration tensors A(x), B(x)
(x) = A(x) hi
(x) = B(x) hi
27
Definitions
Material Symmetries
Models Overview
Volume Elements
Homogenization volume s must be of suitable size
micro s macro (Hashin,1983)
in ideal case s is representative volume element
(RVE)
Representative volume elements
geometrical RVE:
depends solely on phase geometry
physical RVE:
depends on phase geometry and and physical parameter
considered
RVEs allow evaluating effective properties in strict sense
smaller volume elements: SVEs
28
Definitions
Material Symmetries
Models Overview
Outline
Definitions
Material Symmetries
Models Overview
29
Definitions
Material Symmetries
Models Overview
Material Symmetries
Material symmetries play important role in description of
material behavior (constitutive behavior)
discussed in following for linear elastic behavior
elasticity and compliance tensors (order 4) link stress and
strain tensors (order 2)
= E
= C
C = E1
30
Definitions
Material Symmetries
Models Overview
Material Symmetries
Isotropic materials
2 independent elastic moduli required for elasticity tensor
1 CTE required for thermal expansion tensor
=
E=
0
0
0
0 E44 0
0
0
0
0
0
0 E44 0
0
0
0
0
0
0 E44
1
E44 = (E11 E12 )
2
31
Definitions
Material Symmetries
Models Overview
Material Symmetries
Isotropic materials: elastic moduli
2
E11 (E11 + E12 ) 2E12
E =
E11 + E12
E12
=
E11 + E12
G = E44
1
K =
(E11 + 2E12 )
3
Isotropic materials: relations between moduli
E
G=
2(1 + )
32
Definitions
Material Symmetries
Models Overview
Material Symmetries
Isotropic materials: conduction, diffusion
conduction behavior described by 1 conduction coefficient
conductivity tensor:
k 0 0
K = 0 k 0
0 0 k
33
Definitions
Material Symmetries
Models Overview
Material Symmetries
Transversally isotropic materials
5 independent elastic moduli required for elasticity tensor
2 CTEs required for thermal expansion tensor
elasticity tensor in principal orientation:
A
E11 E12 E12 0
0
0
T
T
E12 E23 E22 0
0
0
=
E=
0
0
0
0
E
0
0
44
0
0
0
0
0 E44 0
0
0
0
0
0
0 E66
1
E66 = (E22 E23 )
2
34
Definitions
Material Symmetries
Models Overview
Material Symmetries
Transversally isotropic materials: moduli
2
2E12
EA = E11
E22 + E23
2
2
2
E11 E23
+ E22 E12
2E12 E23
ET = E22
2
E11 E22 E12
E12
AT =
E22 + E23
2
E11 E23 E12
TT =
2
E11 E22 + E12
GAT = E44
GTT = E66
35
Definitions
Material Symmetries
Models Overview
Material Symmetries
Transversally isotropic materials: moduli
1
KT =
(E22 + E23 )
2
1
(E11 + 2E22 + 4E12 + 2E23 )
K =
9
Transversally isotropic materials: relations between moduli
ET
GTT =
2(1 + TT )
EA
KT =
2[(1 TT )(EA /ET ) 2AT 2 ]
36
Definitions
Material Symmetries
Models Overview
Material Symmetries
Hills (1964) connections for special case of (macroscopically
transversally isotropic) UD LFRC
3 instead of 5 independent parameters for transversally
isotropic elasticity
(f)
4(AT (m) )2
(f)
EA = EA + (1 )E (m) +
(f)
(m)
(1/KT 1/KT )2
(f)
AT =
A =
(f)
AT
with
+ (1 )
(f)
KT
1
(m)
KT
(m)
KT
AT (m)
(f)
(m)
1/KT 1/KT
37
Definitions
Material Symmetries
Models Overview
Material Symmetries
Transversally isotropic materials: conduction
conduction behavior described by 2 conduction
coefficients, kA and kT
conductivity tensor in principal orientation:
kA 0 0
K = 0 kT 0
0 0 kT
38
Definitions
Material Symmetries
Models Overview
Material Symmetries
Materials with cubic symmetry
isotropy condition not fulfilled in general:
E44 6= 21 (E11 E12 )
Definitions
Material Symmetries
Models Overview
Material Symmetries
Tetragonal materials
transverse isotropy condition not fulfilled in general:
E66 6= 21 (E22 E23 )
40
Definitions
Material Symmetries
Models Overview
Material Symmetries
Orthotropic materials
9 independent elastic moduli required for elasticity tensor
3 CTEs required for thermal expansion tensor
3 conductivities required for conductivity tensor
elasticity tensor in principal orientation:
1
E11 E12 E13 0
0
0
3
E13 E23 E33 0
0
0
=
E=
0
0
0
0 E44 0
0
0
0
0
0 E55 0
0
0
0
0
0
0 E66
41
Definitions
Material Symmetries
Models Overview
Material Symmetries
Orthotropic materials (contd.)
42
Definitions
Material Symmetries
Models Overview
Material Symmetries
macroscopic statistically isotropic elastic symmetry typically
found in
composites reinforced by particles, randomly oriented
fibers, randomly oriented platelets
polycrystals with equiaxed, randomly oriented crystallites
materials containing equiaxed pores
Note: there are no isotropic monocrystals
43
Definitions
Material Symmetries
Models Overview
Material Symmetries
macroscopic statistically transversally isotropic elastic
symmetry typically found in
composites reinforced by aligned or planar random fibers,
platelets; fibers with axisymmetric ODF
polycrystals with appropriate texture, aligned columnar
crystallites
materials containing aligned, columnar pores
some monocrystals
44
Definitions
Material Symmetries
Models Overview
Material Symmetries
macroscopic statistically orthotropic elastic symmetry typically
found in
polycrystals with more general texture
biocomposites, e.g., bone, wood
some monocrystals
cubic and tetragonal elastic symmetry typically found in
some monocrystals
results from certain simply periodic unit cell models
(square arrays, cubic arrays . . .)
45
Definitions
Material Symmetries
Models Overview
Material Symmetries
Lower elastic symmetries typically found in
some monocrystals
numerical results from volume elements that are too small
to be proper RVEs
46
Definitions
Material Symmetries
Models Overview
Material Symmetries
Elastic symmetries of inhomogeneous materials tend to be
broken when inelastic regimes are entered
regions of different material behavior within materially
nonlinear phases, e.g., elastoplasticity with hardening
47
Definitions
Material Symmetries
Models Overview
Material Symmetries
Distribution of equivalent stress predicted for hexagonal arrangement of parallel cylindrical voids
48
Definitions
Material Symmetries
Models Overview
Outline
Definitions
Material Symmetries
Models Overview
49
Definitions
Material Symmetries
Models Overview
50
Definitions
Material Symmetries
Models Overview
51
Definitions
Material Symmetries
Models Overview
Scale Transitions
In two-scale problems, fields can be split into
slow contributions (hi, hi) and fast contributions
(fluctuations, (x) and (x))
(x) = hi + (x)
(x) = hi + (x)
52
Definitions
Material Symmetries
Models Overview
Scale Transitions
Standard continuum micromechanical models have no absolute
length scale
absolute length scale may be introduced via behavior of
constituents, interfaces
in numerical models discretization may implicitly
introduce length scale(s)
Relative length scales, e.g., different particle sizes, are fully
handled
53
Definitions
Material Symmetries
Models Overview
54
Definitions
Material Symmetries
Models Overview
55
Definitions
Material Symmetries
Models Overview
56
Definitions
Material Symmetries
Models Overview
57
Definitions
Material Symmetries
Models Overview
58
Definitions
Material Symmetries
Models Overview
PHASE ARRANGEMENT
PERIODIC APPROXIMATION
59
Definitions
Material Symmetries
Models Overview
60
Definitions
Material Symmetries
Models Overview
PHASE ARRANGEMENT
WINDOW
61
Definitions
Material Symmetries
Models Overview
62
Definitions
Material Symmetries
Models Overview
PHASE ARRANGEMENT
EMBEDDED CONFIGURATION
63
Definitions
Material Symmetries
Models Overview
Modeling Strategies
PHASE ARRANGEMENT
EMBEDDED CONFIGURATION
PERIODIC APPROXIMATION,
UNIT CELL
WINDOW
64
Definitions
Material Symmetries
Models Overview
no scale transition
65
Definitions
Material Symmetries
Models Overview
66
Part II
Numerical Models
67
68
General Remarks
Note: Full field methods can give individual high-quality
solutions, but they do not provide information on
microstructural parameters
69
General Remarks
Numerical micromechanics concentrates on studying
specific, discrete phase arrangements
by numerical continuum methods.
This requires
governing differential equations
appropriate volume elements
Outline
Governing Equations
Boundary Conditions
Volume Elements
Num.Engng.Methods
FEM
FEM in Micromechanics
Damage Modeling
71
Governing Equations
Modeling of static mechanical behavior is based on
conservation of momentum, leading to stress equilibrium
in Cauchy stresses for static case, small strain regime, no
body forces
ij = 0 div (x) = o
xi
constitutive relations
ij = Eijkl kl
t
dij = Eijkl
dkl
t
ij = Eijkl
kl
kl =
1
uk +
ul
2 xl
xk
72
Governing Equations
Simplest case: mechanical behavior for small strain elasticity
(micromechanics: position dependent material behavior)
Eijkl
uk = 0
xi
xl
Well-known special case for homogeneous isotropic elasticity
with body forces:
NavierCauchy equations
G 2 uk
2 ui
(K + )
+G
+ bi = 0
3 xk xi
xk xk
More complex material behaviors (e.g., elastoplasticity) cannot
be denoted in closed form.
73
Governing Equations
Modeling of static thermal conduction without heat sources:
divergence-free heat flux q
qi = 0 div q(x) = 0
xi
Constitutive behavior links heat flux and negative temperature
gradient via the conductivity tensor K (Fouriers law)
T
qi = kij
xj
For isotropic conductivity the resulting differential equation in
terms of the temperature T is
k
T =0
xi xi
74
Governing Equations
Mathematical descriptions relevant to many problems of
continuum micromechanics are partial differential equations
(boundary value problems): mechanostatics, time-independent
conduction (and analogous diffusion problems)
Requirements on numerical methods
must be able to handle underlying PDEs
must be able to handle required boundary conditions
inhomogeneous material behavior must be accounted for
interfacial effects may have to be accounted for
75
Governing Equations
Variables in elastic and conduction problems
physical problem
direct variable
generalized intensity
generalized flux
generalized property
inverse property
elasticity
displacement field
u [m]
strain field
[]
stress field
[Pa]
elasticity
E [Pa]
compliance
C
thermal conduction
temperature field
T [K]
thermal gradient field
d [Km1 ]
heat flux field
q [Wm2 ]
thermal conductivity
K [Wm1 K1 ]
thermal resistivity
R
76
Outline
Governing Equations
Boundary Conditions
Volume Elements
Num.Engng.Methods
FEM
FEM in Micromechanics
Damage Modeling
77
Boundary Conditions
Micromechanics of elastic materials is based on MandelHill
condition (Hill,1963)
hiT hi = h T i
Z
h1 Z
ih 1 Z
i
1
ij (x) d
ij (x) d =
[ij (x) ij (x)] d
s s
s s
s s
78
Boundary Conditions
By using integral formulations for and , an integral form of
MandelHill condition can be obtained as
(Hazanov/Amieur,1995)
Z
T
t(x) hi n (x) u(x) hi x d = 0
79
Boundary Conditions
Integral form of MandelHill condition
Z
T
t(x) hi n (x) u(x) hi x d = 0
80
Boundary Conditions
For use in micromechanics numerical methods must be
capable of handling the following boundary conditions
kinematically uniform BC are Dirichlet BC
81
Boundary Conditions
Boundary conditions (contd.)
for models involving structural elements (beams, plates,
shells) boundary conditions for rotational degrees of
freedom must be considered
requirements for handling BC in thermal conduction (and
other diffusion-type problems) are analogous to those
pertinent to mechanical problems
82
Outline
Governing Equations
Boundary Conditions
Volume Elements
Num.Engng.Methods
FEM
FEM in Micromechanics
Damage Modeling
83
Volume Elements
structureproperty relationships play central role in continuum
micromechanics choosing suitable volume elements is of
major importance in numerical models
volume elements must reliably reproduce the
characteristics of microgeometry important for a given
problem
84
Volume Elements
Volume elements used in numerical micromechanics take the
form of
synthetic phase geometries
Volume Elements
Highly idealized VEs
describe artificial, usually periodic model material
useful for studying principles underlying behavior,
developing algorithms
no pretense of being representative of phase arrangements
of actual materials
86
Volume Elements
Volume elements approaching realistic microstructures
aim at quantitative results and direct comparisons with
experiments
degree of representativeness and, consequently, size of the
volume element is an issue
87
88
89
90
1.96 S(Y )
err(Y ) =
nSVE
91
93
94
tin
ra
pa
Se
Se
pa
ra
tin
ax
is
ne
Li
Separating plane
95
96
97
98
Volume Elements
Volume elements for other inhomogeneous materials
polycrystals
cellular materials
cancellous bone
99
Outline
Governing Equations
Boundary Conditions
Volume Elements
Num.Engng.Methods
FEM
FEM in Micromechanics
Damage Modeling
100
101
102
103
104
div E0 (x)+ (x) = o
0 (x, y) (y)dy
(x) =
s
Z
0 (x, y) (y)dy
(x) = hi
s
105
Outline
Governing Equations
Boundary Conditions
Volume Elements
Num.Engng.Methods
FEM
FEM in Micromechanics
Damage Modeling
107
u
(x) trial function containing free parameters
R(x) residual (error)
w(x) weighting function, of same type as trial function
108
hei
u(x) =
Xn
heii
N(x) i U
i=1
heii N (x):
iU :
shape functions
(unknown) values of displacements at nodal points
109
K: stiffness matrix
U: vector of unknown nodal displacements
F: vector of nodal forces
110
112
113
Outline
Governing Equations
Boundary Conditions
Volume Elements
Num.Engng.Methods
FEM
FEM in Micromechanics
Damage Modeling
116
117
Standard FE meshes
118
119
Inhomogeneous element
120
121
Pixel mesh
122
123
124
Multiphase elements
125
no well-defined interfaces
126
XFEM mesh
127
128
130
131
132
133
1
hf i =
1X
fl l
f (x)d
l=1
134
l=1
135
136
PROBABILITY DENSITY []
3
2.5
2
1.5
1
0.5
0
0
10
12
14
137
PDD MATRIX
vol.avg. MATRIX
PDD SPH
vol.avg. SPH
PROBABILITY DENSITY []
1.5
0.5
0
0
10
12
14
Predicted phase-level distributions of von Mises equivalent stress in both phases of PRC ( = 0.2) secondary
peak and corresponding matrix regions in model
138
139
(i) hi
hi(i) = A
(i) hi
hi(i) = B
140
141
142
2.2124
0.9226
0.9255
E =
0.0015
0.0163
0.0015
0.9226
2.2182
0.9257
0.0052
0.0000
0.0000
2.2302 0.0023
0.0028
0.0054
0.0094 0.0038
0.0023 0.6596
0.6623
0.0029
0.0028 0.0094
0.6550
0.0054 0.0038 0.0029
2.028 0.009
0.854 0.002
0.857 0.004
E =
0.000 0.002
0.003 0.004
Ensemble average (5 SVEs, each with 20 tetrahedra, = 0.2), normalized w.r.t. E (m)
E,iso
2.030
0.853
0.853
=
0.000
0.000
0.000
0.853
2.030
0.853
0.000
0.000
0.000
0.853
0.853
2.030
0.000
0.000
0.000
0.000
0.000
0.000
0.588
0.000
0.000
0.000
0.000
0.000
0.000
0.588
0.000
0.000
0.000
0.000
0.000
0.000
0.588
144
awkward to mesh
additional information required for actual geometries at
(nearly) touching inhomogeneities
147
Outline
Governing Equations
Boundary Conditions
Volume Elements
Num.Engng.Methods
FEM
FEM in Micromechanics
Damage Modeling
150
on microscale
151
153
very flexible
parameter hungry (e.g., using tables of parameters in
terms of local stress triaxiality = vol /dev )
bit of a mixed blessing: which parameters are material
parameters and which are modeling parameters?
154
155
157
Accumulated plastic strains ahead of notch in LFRC predicted by embedding model (Gonzalez/Llorca,2006)
159
160
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Part III
Periodic Homogenization
161
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
162
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
General Remarks
Periodic homogenization (Periodic Microfield Approach,
PMA)
uses unit cells (repeating unit cells, RUC): volume
elements describing periodic phase arrangements
163
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
164
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
(z) = hi + (z)
(z) = hi + (z)
165
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
166
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
<s >
s
us
< s> s
us
cu
A
s
Periodic microfields in one-dimensional case
167
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
168
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
p1
p2
p1
Different minimum unit cells for a 2-dimensional periodic matrixinclusion medium with two (slightly)
non-orthogonal translation vectors p1 and p2
170
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
p1
p1
p2
3 minimum unit cells for a 2-dimensional periodic matrixinclusion medium detail
171
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
172
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
173
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
B
A
E
C
G
F
periodic boundary
symmetry boundary
point symmetry boundary
symmetry center
(pivot point)
174
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
NE
111111
000000
000000
111111
u
NE
z2
uSE
SW
z1
SE
1
0
0
1
u
0
1
0
1
NW
NW
1111
0000
u
0000
1111
0000
1111
0000 NE
1111
NE
E
uSE
SW
1
0
0
1
0
NW 1
uNW
111
000
000 NE
111
000
111
0000
1111
u
U
0000
1111
U
uP
u
11111
00000
00000L
11111
L
SE
SW
SE
175
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
NE
~
s1
NW
N
c^1
c1
W
s~
E
s~2
~
s1
SW
W
S
c2
SE
c^2
SE
SW
Periodicity BC in 2D
179
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
ck = ck + uk = ck + hi ck
c1 = xE (
s2 ) xW (
s2 ) = xSE xSW
c1 = xE (
s2 ) + uE (
s2 ) xW (
s2 ) + uW (
s2 )
= c1 + uE (
s2 ) uW (
s2 )
= c1 + uSE uSW
so that
hi ck = uSE uSW
macroscopic degrees of freedom of unit cell can be carried
by specific nodes (master nodes)
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
uE (
s2 ) = uW (
s2 ) + uSE
181
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
in 2D: rectangles
in 3D: right hexahedra
Extensional shear
182
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
uW (
s2 ) = 0
vS (
s1 ) = 0
NE
N
W
s~2
c1
E
~
s1
SW
NE
NW
c2
E
S
SE
SW
SE
Symmetry BC in 2D
183
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
uU (
sP ) + uL (
sP ) = 2uP
vN (
s1 ) = vNW = 2vP
vS (
s1 ) = 0
uW (
s2 ) = 0
master nodes carry macroscopic DOFs (P,NW)
NE
N
W
~
s
c1
NW
E
~
s
~
sP
~
s1 S
SW
NE
N
E
c2
SE
SW
Antisymmetry BC on E-face in 2D
S
SE
186
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
187
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
188
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
189
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
190
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
f (Z(x), z(x), )
x
1
f+
f
Z
z
(1)
(2)
(0)
z ui
+H.O.T.
=0
192
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
1
ui +
uj
ij =
2 xj
xi
expand into
ij (Z, z, ) =
+
+
=
1 n (0)
(0)
(1)
ui +
uj
u +
+
2
Zj
Zi
zj i
(1)
(2)
n (1)
ui +
uj
u +
+
2
Zj
Zi
zj i
H.O.T.
(1)
(2)
ij (Z, z) + ij (Z, z) + H.O.T.
(1) o
u
zi j
(2) o
u
zi j
193
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
(1)
(1)
ij (Z, z) = ij + ij
194
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
expansion of stresses
(1)
(2)
+ H.O.T.
(1)
=0
zj ij
(order
1 )
(2)
(1)
ij +
+ bi = 0
Zj
zj ij
(order
0 )
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
(1)
ij
(1)
ij
imn (1)
= Iijmn +
mn
zj
196
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
(1)
kmn (1)
ij = Eijkl (z) Iklmn +
mn
zl
Eijkl =
Eijkl (z) Iklmn +
kmn (z) d
UC UC
zl
197
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
ta (s) d
PSE =
ta (s) d
PNW =
199
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
PNW = N taN
200
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
a
0
12
a
nN (s) =
tNW =
0
1
Z a
a
N 12
12
PNW =
d =
0
0
N
analogous in 3D case
201
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
202
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
analogous in 3D case
203
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
205
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
results that look O.K. for one unit cell do not necessarily
make physical sense for a periodic assembly
206
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
207
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
208
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
209
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
W
E
Homogenization of porous layer-like structure
210
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
211
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
212
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
214
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
leads to 2D analysis
axial shear loading: inhomogeneous out-of-plane
deformations must be allowed
special GPE or 3D models required
modeling of fiber damage requires 3D volume elements
periodic hexagonal (PHA): elastically transversally
isotropic
periodic square (PSA): not transversally isotropic
multi-fiber arrangements
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
CH3
CS8
RH2
CH1
PH0
CS7
MS5
PS0
216
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
217
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
EA
[GPa]
180.0
67.2
118.8/119.3
118.8/118.9
118.8
118.8
118.8
118.8
118.8
118.7
118.8
118.8
118.8
118.8
ET
[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
AT
[]
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
TT
[]
0.20
0.35
0.277/0.394
0.326/0.347
0.342
0.337
0.339
0.338
0.340
0.338
0.314
0.363
0.334
0.333
A
[K 1 106 ]
6.0
23.0
11.84/12.47
11.85/11.98
11.84
11.84
11.94
11.89
11.84
11.90
11.85
11.85
11.90
11.90
T
[K 1 106 ]
6.0
23.0
15.77/16.46
16.31/16.45
16.46
16.46
16.35
16.40
16.46
16.42
16.45
16.45
16.31
16.46
219
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
PS/00
Fiber
DN
4.00E+01
PH
3.00E+01
PS/45
2.00E+01
Matrix
1.00E+01
5.00E+01
6.00E+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 []
220
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
INC.22
4.0000E-02
3.0000E-02
2.0000E-02
1.0000E-02
1.0000E-08
0.0E+00
Distribution of equivalent plastic strain in the matrix of an MMC under transverse uniaxial loading
(CH1, = 0.475)
221
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
INC.800
1.5000E-02
1.2000E-02
9.0000E-03
6.0000E-03
3.0000E-03
0.0E+00
Distribution of equivalent plastic strain in the matrix of an MMC under transverse uniaxial loading
(DS, = 0.453)
222
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
INC.800
2.5000E+02
2.0000E+02
1.5000E+02
1.0000E+02
5.0000E+01
0.0E+00
Distribution of maximum principal stress in the fibers of an MMC under transverse uniaxial loading
(DS, = 0.453)
223
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
x3
x2
Distribution of temperature gradients in a Gr/Al MMC under transverse thermal loading (DS, = 0.453)
224
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
90 o plies
0o ply
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
226
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
227
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
228
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Tows
Matrix
Composite
229
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
230
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
3
2
Maximum principal stress in tows (twill weave) under thermal loading (Marte,2004)
231
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
232
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
233
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
3D unit cell models for aligned SFRC with non-staggered and staggered fibers
234
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Axisymmetric cell models for aligned SFRC with non-staggered and staggered fibers
235
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
P
A= const
236
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
~s
NW
NE
N
NW
E
P
W
z
z
SW
L s
~
SE
r
undeformed
SW
SE
deformed
237
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
238
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
240
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
3D unit cells for SFRC reinforced by randomly oriented fibers ( = 0.15, a = 5) color coded for Weibull fracture
probabilities under uniaxial tension
241
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
MAX.PRINC.STRESS / APPL.STRESS []
MAX.PRINC.STRESS / APPL.STRESS []
0.5
0
0
10
20
30
40
50
60
ANGLE [deg]
70
80
90
4
3.5
3
2.5
2
1.5
1
UC mean
MTM
UC
0.5
0
0
10
20
30
40
50
60
ANGLE [deg]
70
80
90
Comparison of maximum principal stress in fibers predicted by extended MTM and 3D unit cell models
( = 0.15, a = 5, spheroids (left) and cylinders (right), uniaxial tensile loading)
242
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
x2
x1
x3
3D unit cell for SFRC reinforced by (nearly) planar randomly oriented short fibers ( = 0.207, a = 10)
243
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
x3
x2
x1
Distribution of temperature gradients in an MMC reinforced by (nearly) transversely oriented short fibers under
in-plane thermal loading (hot spots are marked)
244
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
245
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
axisymmetric cells
planar cells
3/D cells
246
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
2-particle cell
2-particle
cell
2-particle cell
Reference
Volume
1-particle cell
247
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
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]
[GPa]
96.4
89.0
90.0
E [110]
[GPa]
90.5
91.7
90.6
[]
0.17
0.35
0.286/0.340
0.323/0.328
0.329
0.327
0.326
0.327
A
[K 1 106 ]
4.3
23.0
16.8/18.6
18.5/18.6
18.6
18.6
18.4
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
249
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
1.0
0.8
a)
[110]
0.6
0.4
0.2
b)
[100]
Distribution of equivalent plastic strain predicted with unit cell model based on sc arrangement of spheres
250
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
0.0
Distribution of equivalent plastic strain predicted with multi-particle unit cell model
251
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
252
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
500
400
MTM, voids
MTM, perfect
MTM/RTA, hom
MTM/RTA, inh
3PB, hom
PMA
300
200
100
0
1
10
100
Particle diameter [um]
1000
10000
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
1.45
1.4
1.35
1.3
1.25
1.2
1.4
1.5
1.6
1.7
1.8
1.9
2
2.1
NORMALIZED EFFECTIVE SHEAR MODULUS []
HSB
3PB
SPH
TKD
OCT
CUB
2.2
2.3
TET
Berryman plot of elastic moduli predicted for different particle shapes at volume fraction = 0.2 and elastic
contrast 10
254
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
SPH
OCT
CUB
TET
/E (m)
ENI
GNI /G(m)
ERMT
/E (m)
GRMT /G(m)
EMPC
/E (m)
GMPC /G(m)
1.347
1.382
1.393
1.429
1.367
1.407
1.418
1.458
1.415
1.454
1.467
1.508
1.442
1.484
1.500
1.545
1.428
1.468
1.478
1.525
1.455
1.514
1.500
1.564
normalized moduli
255
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
-0.5
-1
-0.01
-0.005
0.005
0.01
0.015
OCT
0.5
-0.5
-1
0.02
-0.01
-0.005
CUB
0.5
-0.5
-1
-0.01
-0.005
0.005
0.01
MACROSCOPIC STRAIN []
0.005
0.01
0.015
0.02
0.015
0.02
MACROSCOPIC STRAIN []
NORMALIZED MACROSCOPIC STRESS []
MACROSCOPIC STRAIN []
0.015
0.02
TET
0.5
-0.5
-1
-0.01
-0.005
0.005
0.01
MACROSCOPIC STRAIN []
Predicted macroscopic responses of SiC/Al composites ( = 0.2) subjected to pure shear loading cycle
257
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
3.5
3
2.5
2
1.5
1
0.5
0
-0.005
-0.004
-0.003
-0.002
-0.001
3.5
MATRIX OCT
INHOMOG. OCT
3
2.5
2
1.5
1
0.5
0
-0.005
3.5
MATRIX CUB
INHOMOG. CUB
3
2.5
2
1.5
1
0.5
0
-0.005
-0.004
-0.003
-0.002
-0.001
MACROSCOPIC STRAIN []
-0.004
-0.003
-0.002
-0.001
MACROSCOPIC STRAIN []
NORMALIZED EQUIVALENT STRESS []
MACROSCOPIC STRAIN []
3.5
MATRIX TET
INHOMOG. TET
3
2.5
2
1.5
1
0.5
0
-0.005
-0.004
-0.003
-0.002
-0.001
MACROSCOPIC STRAIN []
Predicted evolution of von Mises equivalent stress in both phases of SiC/Al composites ( = 0.2) during
volumetric loading cycle
258
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
259
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
PMA: Polycrystals
Unit cell models of polycrystals
Voronoi microstructures may have to be modified
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
PMA: Polycrystals
Polycrystal example (continued)
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
PMA: Polycrystals
Polycrystal example (continued)
Critical resolved shear stresses in (partially remeshed) model of polycrystals under plane strain compression (Quey
et al.,2011)
262
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Outline
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
263
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
264
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
265
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
266
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
267
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
269
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
270
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Open cell lattice structures: structural and unit cell models (Luxner et al.,2005)
272
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
273
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
Unit cell for hollow strut foam: undeformed, compressed, sheared (Daxner et al.,2007)
274
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
275
Unit Cells
Loading
Issues
LFRC
Weaves
SFRC
PRC
Polycrystals
Cellular
276
Windowing Models
Multi-Scale Models
Checking Models
Part IV
Other Models
277
Windowing Models
Multi-Scale Models
Checking Models
278
Windowing Models
Multi-Scale Models
Checking Models
Outline
Windowing Models
Embedded Cell Models
Multi-Scale Models
Checking Models
279
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Basic idea of windowing models: estimate or bound
macroscopic properties on the basis of randomly selected
samples of microstructure that are too small to be RVEs:
windows
280
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
281
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
integral MandelHill condition can be fulfilled by
macrohomogeneous stresses (SUBC) and strains (KUBC)
t(x) = a n (x) or u(x) = a x
282
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
mesoscale bounds obtained from volume elements of
different sizes give hierarchies of bounds
283
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
alternative way of fulfilling integral MandelHill condition:
uniform stress or strain BC at level of individual faces
(MUBC)
T
[t(x) hi n (x) u(x) hi x d = 0
x k
284
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
a specific set of MUBC gives close agreement with
periodic homogenization for orthotropic or higher elastic
symmetry:
periodicity compatible mixed uniform boundary conditions
(PMUBC; Pahr/Zysset,2008)
285
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
Tensile 1
Tensile 2
Tensile 3
Shear 12
Shear 13
Shear 23
Thermal
Loading
East
u1 = a11 c1 /2
a
2 = 3a = 0
u1 = 0
2a = 3a = 0
u1 = 0
2a = 3a = 0
u2 = a21 c1 /2
u3 = 0, 1a = 0
u3 = a31 c1 /2
u2 = 0, 1a = 0
u1 = 0
2a = 3a = 0
u1 = 0
2a = 3a = 0
West
u1 = a11 c1 /2
a
2 = 3a = 0
u1 = 0
2a = 3a = 0
u1 = 0
2a = 3a = 0
u2 = a21 c1 /2
u3 = 0, 1a = 0
u3 = a31 c1 /2
u2 = 0, 1a = 0
u1 = 0
2a = 3a = 0
u1 = 0
2a = 3a = 0
North
u2 = 0
= 3a = 0
u2 = a22 c2 /2
1a = 3a = 0
u2 = 0
1a = 3a = 0
u1 = a12 c2 /2
u3 = 0, 2a = 0
u2 = 0
1a = 3a = 0
u3 = a32 c2 /2
u1 = 0, 2a = 0
u2 = 0
1a = 3a = 0
1a
South
u2 = 0
= 3a = 0
u2 = a22 c2 /2
1a = 3a = 0
u2 = 0
1a = 3a = 0
u1 = a12 c2 /2
u3 = 0, 2a = 0
u2 = 0
1a = 3a = 0
u3 = a32 c2 /2
u1 = 0, 2a = 0
u2 = 0
1a = 3a = 0
1a
Top
u3 = 0
= 2a = 0
u3 = 0
1a = 2a = 0
u3 = a33 c3 /2
1a = 2a = 0
u3 = 0
1a = 2a = 0
u1 = a13 c3 /2
u2 = 0, 3a = 0
u2 = a23 c3 /2
u1 = 0, 3a = 0
u3 = 0
1a = 2a = 0
1a
Bottom
u3 = 0
= 2a = 0
u3 = 0
1a = 2a = 0
u3 = a33 c3 /2
1a = 2a = 0
u3 = 0
1a = 2a = 0
u1 = a13 c3 /2
u2 = 0, 3a = 0
u2 = a23 c3 /2
u1 = 0, 3a = 0
u3 = 0
1a = 2a = 0
1a
The 6 uniform strain load cases and the thermal load case defining PUBC in thermoelasticity
286
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing: analogous concepts hold for thermal conduction
SUBC, KUBC, PMUBC
Thermal 1
Thermal 2
Thermal 3
East
T = da1 l1 /2
q1a = 0
q1a = 0
West
T = da1 l1 /2
q1a = 0
q1a = 0
North
q2a = 0
T = da2 l2 /2
q2a = 0
South
q2a = 0
T = da2 l2 /2
q2a = 0
Top
q3a = 0
q3a = 0
T = da3 l3 /2
Bottom
q3a = 0
q3a = 0
T = da3 l3 /2
287
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
Comparison of estimates for macroscopic elastic properties obtained by windowing and periodic homogenization for
a PRC
288
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
SUBC, KUBC, PMUBC give rise to boundary layers near
surface of volume elements
289
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
Unit cell with PBC and PMUBC under tensile transverse loading (G
andara,2008)
290
Windowing Models
Multi-Scale Models
Checking Models
Windowing Models
Windowing (contd.)
CT-based voxel model of window of cancellous bone two load cases using windowing with PMUBC (Pahr,2013)
291
Windowing Models
Multi-Scale Models
Checking Models
Outline
Windowing Models
Embedded Cell Models
Multi-Scale Models
Checking Models
292
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
Basic idea: highly resolved core (kernel, motif) in
smeared-out embedding region
core: any spatially resolved, inhomogeneous phase
arrangement
RVE or SVE
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
Embedded cells (contd.)
loads are introduced via embedding region
294
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
Embedded cells (contd.)
295
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
2
3
1
ECA model of open cell foam (beam model for core, continuum model for embedding
296
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
Embedded cells (contd.)
embedding cells always give rise to boundary layers close
to interface in both core and embedding regions
297
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
S22
VALUE
-3.24E+02
-2.01E+02
-7.90E+01
+4.36E+01
+1.66E+02
+2.89E+02
+4.11E+02
+5.34E+02
+6.56E+02
+7.79E+02
+9.01E+02
+1.02E+03
2
3
ECA model of progressing crack tip in open cell foam, stresses in struts (Ableidinger,2000)
298
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
Embedded cells (contd.)
self-consistent evaluation of properties of embedding
region
in elastoplasticity:
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
Embedded cells (contd.)
embedded cells using self-consistent embedding are
micromechanical method in the strict sense because they
involve a scale transition
300
Windowing Models
Multi-Scale Models
Checking Models
Embedded Cells
Embedded cells (contd.)
can provide highly resolved local models without
periodicity constraints
301
Windowing Models
Multi-Scale Models
Checking Models
Outline
Windowing Models
Embedded Cell Models
Multi-Scale Models
Checking Models
302
Windowing Models
Multi-Scale Models
Checking Models
Multi-Scale Models
Designator multi-scale modeling is used in different ways:
hierarchical models: sequential or concurrent
homogenization (and localization) for handling problems
involving multiple scale transitions
303
Windowing Models
Multi-Scale Models
Checking Models
Multi-Scale Models
Hierarchical models for microgeometries with suitable
intermediate length scales
building block approach
Windowing Models
Multi-Scale Models
Checking Models
Multi-Scale Models
Hierarchical models using concurrent homogenization
usually FE-based at higher length scale
at each integration point a micromechanical model is run
providing a stress answer for a given strain input
305
Windowing Models
Multi-Scale Models
Checking Models
Multi-Scale Models
Hierarchical multi-Scaling
scale
transition
#2
scale
transition
#1
MACROSCALE
(sample)
MESOSCALE
(particle clusters)
MICROSCALE
(particles in matrix)
306
Windowing Models
Multi-Scale Models
Checking Models
Multi-Scale Models
Hierarchical multi-Scaling
EPS.EFF.PLAST.MTX.
3.5000E-03
3.0000E-03
2.5000E-03
2.0000E-03
1.5000E-03
Phase averaged microscopic equivalent plastic strains in matrix within inhomogeneity poor regions
307
Windowing Models
Multi-Scale Models
Checking Models
Multi-Scale Models
Two-sale models using nested FE computations at two length
scales
often called FE2 models (Feyel,1999)
concurrent evaluation at macroscale (structure) and
microscale (periodic homogenization with prescribed
strains from integration points of macromodel)
308
Windowing Models
Multi-Scale Models
Checking Models
Multi-Scale Models
Two-sale models using nested FE computations at two length
scales (contd).
depending on position within structure conditions may be
very different
try to use micromechanical method that is most efficient
for given conditions
an implementation might use
Windowing Models
Multi-Scale Models
Checking Models
Outline
Windowing Models
Embedded Cell Models
Multi-Scale Models
Checking Models
310
Windowing Models
Multi-Scale Models
Checking Models
Checking of Models
Checking and debugging of models: Important issues
discretization: meshing and mesh resolution
Windowing Models
Multi-Scale Models
Checking Models
Checking of Models
Two important aspects in verifying simulations:
simulations aim at closely approximating and helping in
explaining experimental data
Windowing Models
Multi-Scale Models
Checking Models
Checking of Models
Checking and debugging of models (contd.)
comparisons with reference solutions
313
Windowing Models
Multi-Scale Models
Checking Models
Checking of Models
Checking and debugging of models (contd.)
comparisons with reference solutions
PonteCasta
neda (1992) bounds and second order
secant schemes (PonteCasta
neda/Suquet,1998) can
give excellent reference values for monotonic
elastoplastic behavior
all bounds become slack at high phase contrasts
314
Windowing Models
Multi-Scale Models
Checking Models
Checking of Models
Checking and debugging of models (contd.)
General rule: given errors tend to have more marked
consequences in nonlinear than in linear regimes
315
Windowing Models
Multi-Scale Models
Checking Models
General Issues
Models should be as complex as necessary and as simple as
possible
Sources of difficulties: free parameters
material parameters are required in general
316