CHAP 3

FEA for Nonlinear Elastic Problems

Nam-Ho Kim

1
 Introduction
• Linear systems
– Infinitesimal deformation: no significant difference between the
deformed and undeformed shapes
– Stress and strain are defined in the undeformed shape
– The weak form is integrated over the undeformed shape
• Large deformation problem
– The difference between the deformed and undeformed shapes is
large enough that they cannot be treated the same
– The definitions of stress and strain should be modified from the
assumption of small deformation
– The relation between stress and strain becomes nonlinear as
deformation increases
• This chapter will focus on how to calculate the residual
and tangent stiffness for a nonlinear elasticity model

2
 Introduction
• Frame of Reference
– The weak form must be expressed based on a frame of reference
– Often initial (undeformed) geometry or current (deformed)
geometry are used for the frame of reference
– proper definitions of stress and strain must be used according to
the frame of reference
• Total Lagrangian Formulation: initial (undeformed)
geometry as a reference
• Updated Lagrangian Formulation: current (deformed)
geometry
• Two formulations are theoretically identical to express
the structural equilibrium, but numerically different
because different stress and strain definitions are used

3
 Table of Contents
• 3.2. Stress and Strain Measures in Large Deformation
• 3.3. Nonlinear Elastic Analysis
• 3.4. Critical Load Analysis
• 3.5. Hyperelastic Materials
• 3.6. Finite Element Formulation for Nonlinear Elasticity
• 3.7. MATLAB Code for Hyperelastic Material Model
• 3.8. Nonlinear Elastic Analysis Using Commercial Finite
Element Programs
• 3.9. Fitting Hyperelastic Material Parameters from Test
Data
• 3.9. Summary
• 3.10.Exercises
4
3.2
Stress and Strain Measures

5
 Goals – Stress & Strain Measures

• Definition of a nonlinear elastic problem

• Understand the deformation gradient?

• What are Lagrangian and Eulerian strains?

• What is polar decomposition and how to do it?

• How to express the deformation of an area and volume
• What are Piola-Kirchhoff and Cauchy stresses?

6
 Mild vs. Rough Nonlinearity

• Mild Nonlinear Problems (Chap 3)

– Continuous, history-independent nonlinear relations between
stress and strain
– Nonlinear elasticity, Geometric nonlinearity, and deformation-
dependent loads

• Rough Nonlinear Problems (Chap 4 & 5)

– Equality and/or inequality constraints in constitutive relations
– History-dependent nonlinear relations between stress and strain
– Elastoplasticity and contact problems

7
 What Is a Nonlinear Elastic Problem?
• Elastic (same for linear and nonlinear problems)
– Stress-strain relation is elastic
– Deformation disappears when the applied load is removed
– Deformation is history-independent
– Potential energy exists (function of deformation)
• Nonlinear
– Stress-strain relation is nonlinear
(D is not constant or do not exist)
– Deformation is large
• Examples
– Rubber material
– Bending of a long slender member
(small strain, large displacement)
8
 Reference Frame of Stress and Strain
• Force and displacement (vector) are independent of the
configuration frame in which they are defined (Reference
Frame Indifference)
• Stress and strain (tensor) depend on the configuration
• Total Lagrangian or Material Stress/Strain: when the
reference frame is undeformed configuration
• Updated Lagrangian or Spatial Stress/Strain: when the
reference frame is deformed configuration
• Question: What is the reference frame in linear
problems?

9
 Deformation and Mapping
• Initial domain W0 is deformed to Wx
– We can think of this as a mapping from W0 to Wx
• X: material point in W0 x: material point in Wx
• Material point P in W0 is deformed to Q in Wx
x  Xu x  ( X, t)  X  u( X, t)

displacement F
Wx
W0 u
Q
P

X x

,  1 : One-to-one mapping
Continuously differentiable
10
 Deformation Gradient
• Infinitesimal length dX in W0 deforms to dx in Wx
• Remember that the mapping is continuously differentiable

x Wx Q'
dx  dX  dx  FdX P'
X W0 u dx
dX Q
P

• Deformation gradient:
xi u 1  [ij ],
Fij  F  1  1  0u
Xj X  
0  , x 
– gradient of mapping F X x

– Second-order tensor, Depend on both W0 and Wx

– Due to one-to-one mapping: det F  J  0. dX  F 1 d x

– F includes both deformation and rigid-body rotation

11
 Example – Uniform Extension
• Uniform extension of a cube in all three directions
x1  1X1 , x2  2X2 , x3  3X3

• Continuity requirement: i  0Why?

• Deformation gradient: 0 0
 1
F   0 2 0 
 0 0 3 

• 1  2 : uniform
3 expansion (dilatation) or contraction
• Volume change
– Initial volume: dV0  dX1dX2dX3
– Deformed volume:
dVx  dx1dx2dx3  123dX1dX2dX3  123dV0

12
 Green-Lagrange Strain
• Why different strains?
2 2
• Length change: dx  dX  d x T d x  d XT d X
 dXT FT FdX  dXT dX
 dXT (FT F  1)dX

Ratio of length change

• Right Cauchy-Green Deformation Tensor

C  FT F
• Green-Lagrange Strain Tensor dx
dX
1
E (C  1) The effect of rotation is eliminated
2

To match with infinitesimal strain

13
 Green-Lagrange Strain cont.
• Properties:
– E is symmetric: ET = E
1  ui uj 
– No deformation: F = 1, E = 0 ij    

2  Xj Xi 

1  u uT uT u 
E    
2  X X X X 
Displacement gradient

 21 0 u   0 uT   0 uT  0 u 
Higher-order term

1
– When
 0 u  1
,
E
2
 
 0 u   0 uT  ε

– E = 0 for a rigid-body motion, but

ε0

14
 Example – Rigid-Body Rotation
• Rigid-body rotation
x1  X1 cos   X2 sin 
x2  X1 sin   X2 cos  a
x3  X3

• Approach 1: using deformation gradient

 cos   sin  0  1 0 0
F   sin  cos  0  FT F   0 1 0 
 0 0 1   0 0 1 

E  21 (FT F  1)  0

Green-Lagrange strain removes rigid-body rotation from deformation

15
 Example – Rigid-Body Rotation cont.
• Approach 2: using displacement gradient
u1  x1  X1  X1 (cos   1)  X2 sin 
u2  x2  X2  X1 sin   X2 (cos   1)
u3  x3  X3  0
 cos   1  sin  0 
 0 u   sin  cos   1 0 
 0 0 0 

 2(1  cos ) 0 0
 0 uT  0 u   0 2(1  cos ) 0 
 0 0 0 

E  21 ( 0 u   0 uT   0 uT  0 u)  0
16
 Example – Rigid-Body Rotation cont.
• What happens to engineering strain?
u1  x1  X1  X1 (cos   1)  X2 sin 
u2  x2  X2  X1 sin   X2 (cos   1)
u3  x3  X3  0

 cos   1 0 0
   0 cos   1 0 
 0 0 0 

Engineering strain is unable to take care of rigid-body rotation

17
 Eulerian (Almansi) Strain Tensor
2 2
• Length change: dx  dX  dxT dx  dXT dX
 dxT dx  dxT F T F 1dx
 dxT (1  F T F 1 )dx
 dxT (1  b1 )dx

• Left Cauchy-Green Deformation Tensor

b  F FT b–1: Finger tensor

• Eulerian (Almansi) Strain Tensor

1
e  (1  b1 )
2
Reference is deformed (current) configuration

18
 Eulerian Strain Tensor cont.

• Properties
– Symmetric
u
– Approach engineering strain when  1
x
– In terms of displacement gradient

1  u uT uT u  
e     x 
2  x x x x  x
1

2

 x u   x uT   x uT  x u  Spatial gradient

• Relation between E and e

E  FT eF

19
 Example – Lagrangian Strain
• Calculate F and E for deformation in the figure
• Mapping relation in W0
 4
3
 X   NI (s, t)XI  (s  1) Y
 I 1 4 Deformed element
 4
2.0
 1
 Y   NI (s, t)YI  2 (t  1)
 I 1 1.0
Undeformed
element
• Mapping relation in Wx X
0.7 1.5
 4
 x(s, t)   NI (s, t)xI  0.35(1  t)
 I 1
 4

 y(s, t)   NI (s, t)yI  s  1
 I 1
20
 Example – Lagrangian Strain cont.
• Deformation gradient
Wx
x x s Q'
F  W0
P'
u dx
X s X dX Q
 0 .35   4 / 3 0  P
  0 
 1 0  2  x(s ,t )
 0 0.7  X(s ,t )
 
 4 / 3 0  Reference
domain (s, t)
• Green-Lagrange Strain

1 T  0.389 0  Tension in X1 dir.

E  (F F  1)   
2  0 0.255  Compression in X2 dir.

21
 Example – Lagrangian Strain cont.
• Almansi Strain
 0.49
T 0 
b  FF   
 0 1.78 
 0.52 0 
e 1
2  1 b 
1

0 0.22 
Compression in x1 dir.
Tension in x2 dir.
 
• Engineering Strain
 1 0.7 
0 u  F  1   
 1.33  1 
 1 0.32 
 1
2  0 u  0 u 
T

0.32 1 
Artificial shear deform.
Inconsistent normal deform.
 
Which strain is consistent with actual deformation?

22
 Example – Uniaxial Tension
• Uniaxial tension of incompressible material (l1 = l > 1)
• From incompressibility x1  1X1
123  1  2  3   1/2 x2  2X2
x3  3X3
• Deformation gradient and deformation tensor
 2
  0 0   0 0 
    1

F   0  1/2 0  C0  0 
0 1/2  0 0  1 
 0    
• G-L Strain
 2  1 0 0 
1 1

E  0  1 0 
2 1 
 0 0   1 
23
 Example – Uniaxial Tension
• Almansi Strain (b = C)

  2 0 0  1   2 0 0 
  1 
b1  0  0 e  0 1 0 
2
 0 0    0 0 1   
 

• Engineering Strain

  1 0 0 
 1/2 
 0  1 0 
10%
strain
 0 0  1/2
 1 

• Difference
1 2 1
E11  (  1) e11  (1   2 ) 11    1
2 2 24
 Polar Decomposition

• Want to separate deformation from rigid-body rotation

• Similar to principal directions of strain

• Unique decomposition of deformation gradient
F  QU  VQ

– Q: orthogonal tensor (rigid-body rotation)

– U, V: right- and left-stretch tensor (symmetric)

• U and V have the same eigenvalues (principal stretches),

but different eigenvectors

25
 Polar Decomposition cont.

d x  Q  U  dX F  VQ
 V  Q  dX
e1 e3

Q e2 V
λ3e3
E3 λ1e1

E1 E2 λ3E3 λ2e2
U
λ2E2 Q
λ1E1

F  QU
• Eigenvectors of U: E1, E2, E3
• Eigenvectors of V: e1, e2, e3
• Eigenvalues of U and V:l1, l2, l3 26
 Polar Decomposition cont.
• Relation between U and C
U2  C U C
– U and C have the same eigenvectors.
– Eigenvalue of U is the square root of that of C
• How to calculate U from C?
• Let eigenvectors of C be   [E1 E2 E3 ]
T C
• Then,   where
 12 0 0
  Deformation tensor in
0 22 0
principal directions
0 0 23 


27
 Polar Decomposition cont.
Useful formulas
• And U    T 3
C  i Ei  Ei
 2

i1
 1 0 0 3
   0 2 0  U  iEi  Ei
i1
 0 0 3  3
Q  ei  Ei
i1
3
• General Deformation b  i ei  ei
 2

dx  FdX  b  QUdX  b i1

3
1. Stretch in principal directions V  iei  ei
2. Rigid-body rotation i1
3
3. Rigid-body translation F  iei  Ei
i1 28
 Generalized Lagrangian Strain
• G-L strain is a special case of general form of Lagrangian
strain tensors (Hill, 1968)
1
Em 
2m

U2m  1 

29
 Example – Polar Decomposition
• Simple shear problem X 2, x 2
2
 x1  X1  kX2 k
 3
 x2  X2
x  X
 3 3 X 1, x 1
1 k
• Deformation gradient F   
 0 1 
 1 2 
 1 k  3
• Deformation tensor C  FT F    2 
2 7
 k k  1   3 3


• Find eigenvalues and eigenvectors of C X2 E1

E2
1  3, 2  1 3 60o
X1
E1   1
2 2
3
 
, E2   2
3 1
2 
30
 Example – Polar Decomposition cont.
3 0 
• In E1 – E2 coordinates C      
 0 1 3 
 1 2  3 2
• Principal Direction Matrix   [E1 E2 ]   
 3 2 1 2 

• Deformation tensor in principal directions

  T  C  
• Stretch tensor
 3 0 
  
 0 1 3 
 32
T 12 
U       
 1 2 5 2 3 
31
 Example – Polar Decomposition cont.
• How U deforms a square?
X2, x2
 1   3 2 0   1 2 
U    , U      
0   1 2   1  5 2 3 
• Rotational Tensor
30o X 1, x 1
1
 32 12 
Q  FU   
  1 2 3 2 
X2, x2
 3 2  1   1 2   1.15 
Q    , Q    
 1 2  0  5 2 3   1 
– 30o clockwise rotation
30o X 1, x 1
5 3 6 12 
V  F  QT  
 1 2 3 2 
32
 Example – Polar Decomposition cont.
• A straight line X2  X1 will
tan deform to
X1  x1  kx2, X2  x2
X2, x2
 x2  (x1  kx2 ) tan 
 x1   tan1   k  x2
• Consider a diagonal line: q = 45o
25o
x2 1 X1, x1
tan      24.9
x1 1  k
X2, x2
• Consider a circle

X12  X22  r2 X1, x1

(x1  kx2 )2  x22  r2
x12  2kx1x2  (1  k2 )x22  r2 Equation of ellipse
33
 Deformation of a Volume
• Infinitesimal volume by three vectors
– Undeformed: dV0  dX1  (dX2  dX3 )  erstdXr1dXs2dXt3
– Deformed: dVx  dx1  (dx2  dx3 )  eijk dxi1dxj2dxk3

dVx  eijk dxi1dxj2dxk3

 xj  dX2
 xi 2  xk
1   dx2
 eijk  dXr   dXs   dXt3  dX1
 Xr   Xs   Xt
  dX 3
dx3
dx1
xi xj xk
 eijk dXr1dXs2dXt3
Xr Xs Xt
 erstJ dXr1dXs2dXt3
 JdV0
From Continuum Mechanics

J  det F  123 eijk air ajs akt  erst det a

34
 Deformation of a Volume cont.
• Volume change
dVx  J dV0

• Volumetric Strain
dVx  dV0
 J 1
dV0

• Incompressible condition: J = 1
• Transformation of integral domain

 x
f d   0
fJ d

35
 Example - Uniaxial Deformation of a Beam
• Initial dimension of L0×h0×h0 deforms to L×h×h

x1  1X1 1  L / L0 L0
x2  2X2 2  h / h0
x3  3X3 3  h / h0
h0
• Deformation gradient h0

 1 0 0 J  det F  123
F   0 2 0  2 L
L  h  LA
 0 0   h
3  L0  h0  L0A0
h

• Constant volume
L0 L0
J 1  h  h0 A  A0
L L
36
 Deformation of an Area
• Relationship between dS0 and dSx
NdS0  dX1  dX2 NdS
i
1 2
0  eijk dXj dXk

ndSx  dx1  dx2 nrdSx  erstdxs1dxt2

Xj Xk 1 2
NdS
i 0  eijk dxsdxt
xs xt

Xi Xi Xi Xj Xk 1 2

 NdS
i 0  eijk dxsdxt
xr xr xr xs xt

F(X)

N n
dX 2 dx2
X dS0 x dS
x
S0 dx1 Sx
dX1

Undeformed Deformed
37
 Deformation of an Area cont..
• Results from Continuum Mechanics
xr xs xt
eijk F  erst
Xi Xj Xk

1 Xi Xj Xk

erst F  eijk .
xr xs xt

nrdSx
• Use the second relation:

Xi Xi Xj Xk 1 2 1

NdS
i 0  eijk dxs dxt  erst F dxs1dxt2
xr xr xs xt

ndSx  JF T
 NdS0 F T  N
n  F T  N  n
F T  N

dSx  J F( x) T N( X) dS0
38
 Stress Measures
• Stress and strain (tensor) depend on the configuration
• Cauchy (True) Stress: Force acts on the deformed config.
f
– Stress vector at Wx: t  lim  n
Sx  0 Sx

Cauchy Stress, sym

– Cauchy stress refers to the current deformed configuration as a

reference for both area and force (true stress)
Undeformed configuration Deformed configuration

∆f
N
∆S0
P ∆Sx
P n

39
 Stress Measures cont.
• The same force, but different area (undeformed area)

f
T  lim  PT N
S0  0 S0 First Piola-Kirchhoff Stress
Not symmetric

– P refers to the force in the deformed configuration and the area

in the undeformed configuration
• Make both force and area to refer to undeformed config.
df  ndSx  PT NdS0 ndSx  JF T  NdS0

df  (JF T NdS0 )  PT NdS0

P  J F 1  : Relation between s and P

40
 Stress Measures cont.
• Unsymmetric property of P makes it difficult to use
– Remember we used the symmetric property of stress & strain
several times in linear problems
• Make P symmetric by multiplying with F-T
1
S  P  F T  JF 1    F T  F  S  FT
J
Second Piola-Kirchhoff Stress, symmetric
– Just convenient mathematical quantities
• Further simplification is possible by handling J differently

  J  F  S  F T

Kirchhoff Stress, symmetric

41
 Stress Measures cont.
• Example

 x
 :  d x   0
 : J d 0   0
 :  d 0

Integration can be done in W0

• Observation
– For linear problems (small deformation):   E  e
– For linear problems (small deformation):     P  S
– S and E are conjugate in energy
– S and E are invariant in rigid-body motion

42
 Example – Uniaxial Tension
F
• Cauchy (true) stress: 11 , s22 = s33 = s12 = s23 = s13 = 0
A
• Deformation gradient:
L0
 11 0 0 
 
F 1  0 21 0 , J 1
 0 0 31  h0

h0
• First P-K stress
1 F 1 F A F L
P11  (JF 11   
A 1 A A0 A0 h
h
• Second P-K stress F

1 T F 1 F A2 FA F
S11  (JF F )11     No clear physical
A 12 A A02 A02 A0 1 meaning

43
 Summary
• Nonlinear elastic problems use different measures of
stress and strain due to changes in the reference frame
• Lagrangian strain is independent of rigid-body rotation,
but engineering strain is not
• Any deformation can be uniquely decomposed into rigid-
body rotation and stretch
• The determinant of deformation gradient is related to the
volume change, while the deformation gradient and
surface normal are related to the area change
• Four different stress measures are defined based on the
reference frame.
• All stress and strain measures are identical when the
deformation is infinitesimal
44
3.3
Nonlinear Elastic Analysis

45
 Goals
• Understanding the principle of minimum potential energy
– Understand the concept of variation
• Understanding St. Venant-Kirchhoff material
• How to obtain the governing equation for nonlinear elastic
problem
• What is the total Lagrangian formulation?
• What is the updated Lagrangian formulation?
• Understanding the linearization process

46
Numerical Methods for Nonlinear Elastic Problem
• We will obtain the variational equation using the principle
of minimum potential energy
– Only possible for elastic materials (potential exists)
• The N-R method will be used (need Jacobian matrix)
• Total Lagrangian (material) formulation uses the
undeformed configuration as a reference, while the
updated Lagrangian (spatial) uses the current
configuration as a reference
• The total and updated Lagrangian formulations are
mathematically equivalent but have different aspects in
computation

47
 Total Lagrangian Formulation
• Using incremental force method and N-R method
– Total No. of load steps (N), current load step (n)
n 1
f  nf  fn
• Assume that the solution has converged up to tn
• Want to find the equilibrium state at tn+1
Last converged configuration
Undeformed configuration (known)
Current configuration
(known) n
W (unknown)
0
W n
u n
P W
n+1
0
P ∆u
n+1
P

X x
Iteration

48
 Total Lagrangian Formulation cont.
• In TL, the undeformed configuration is the reference
• 2nd P-K stress (S) and G-L strain (E) are the natural choice
• In elastic material, strain energy density W exists, such
that
W
stress 
strain

• We need to express W in terms of E

49
 Strain Energy Density and Stress Measures
• By differentiating strain energy density with respect to
proper strains, we can obtain stresses
• When W(E) is given
W(E)
S Second P-K stress
E
• When W(F) is given
W W E W
 :  F  F  S  PT First P-K stress
F E F E
• It is difficult to have W(e) because e depends on rigid-
body rotation. Instead, we will use invariants in Section
3.5

50
 St. Venant-Kirchhoff Material
• Strain energy density for St. Venant-Kirchhoff material
W(E)  21 E : D : E Contraction operator: a : b  aijbij
• Fourth-order constitutive tensor (isotropic material)
D  1  1  2I
E E
– Lame’s constants:  
(1  )(1  2) 2(1  )

– Identity tensor (2nd order): 1  [ij ]

– Identity tensor (4th order): Iijkl  21 (ik  jl  il jk )

I : a  a, 2nd-order sym. a
1 : a  tr(a)  aii  a11  a22  a33

– Tensor product: a  a  aijakl (4th-order)

51
 St. Venant-Kirchhoff Material cont.
• Stress calculation
– differentiate strain energy density
W(E)
S  D : E  tr(E)1  2E
E
– Limited to small strain but large rotation

E  21 (FT F  1)  21 (UT Q T Qu  1)  21 (U2  1)

– Rigid-body rotation is removed and only the stretch tensor

contributes to the strain
W W
– Can show S  2
E C
Deformation tensor

52
 Example Y

• E = 30,000 and n = 0.3 Deformed element

2.0

• G-L strain:  0.389 0 

E
 0 0.255  1.0
Undeformed
element
• Lame’s constants: 0.7 1.5
X

E E
  17,308   11, 538
(1  )(1  2) 2(1  )

• 2nd P-K Stress:

1 0  .389 0 
S  tr(E)1  2E  (.389  .255)    2  0
0 1   .255 
 11,296 0 

 0 3, 565 

1 T  1, 872 0 
  FSF  
J  0 21,516 
53
 Example – Simple Shear Problem
• Deformation map X 2, x 2

x1  X1  kX2 , x2  X2 , x3  X3

1 k 1 T 1 0 k 
F  E  (F F  1)   
 0 1  2 2  k k2 
X 1, x 1

• MaterialEproperties E
  40MPa   40MPa
(1  )(1  2) 2(1  )
20
Cauchy
• 2nd P-K stress 10
k 2
2k 

Shear stress
S  tr(E)1  2E  20   MPa 2nd P-K
2
 2k 3k  0

1 T
 5k2  3k 4 2k  3k3  -10
  FSF  20   MPa
J  2k  3k
3
3k 2

-20
-0.4 -0.2 0.0 0.2 0.4
Shear parameter k 54
 Boundary Conditions
• Boundary Conditions
u  g, on h Essential (displacement) boundary

t  PT N, on  s Natural (traction) boundary

You can’t use S

• Solution space (set)


V  u u  [H1 ( )]3, u h  g 

• Kinematically admissible space


Z  u u  [H1 ()]3, u h
0 

55
 Variational Formulation
• We want to minimize the potential energy (equilibrium)
Pint: stored internal energy
Pext: potential energy of applied loads

( u)  int ( u)   ext ( u)
  W(E) d   uT fb d   s uT t d
0 0 o

• Want to find u   that minimizes the potential energy

– Perturb u in the direction of ū   proportional to t
u  u  u
– If u minimizes the potential, P(u) must be smaller than P(ut) for all
possible ū

56
 Variational Formulation cont.
• Variation of Potential Energy (Directional Derivative)
d
( u, u )  ( u  u ) We will use “over-bar” for variation
d  0
– P depends on u only, but P depends on both u and ū
– Minimum potential energy happens when its variation becomes
zero for every possible ū
– One-dimensional example

P(u)

At minimum, all directional

ū u ū derivatives are zero

57
 Example – Linear Spring

k u
f

• Potential energy: (u)  1 k  u2  f  u

2

• Perturbation: (u  u)  1 k  (u  u)2  f  (u  u)

2
d
• Differentiation:  (u  u)   k  (u  u)  u  f  u
d
• Evaluate at original state:
d
 (u  u)   k uu  fu  0
d  0

Variation is similar to differentiation !!!

58
 Variational Formulation cont.
• Variational Equation
W(E) T b T
( u, u )  0 E : E d   0 u f d   so t d  0
u

for all ū  
– From the definition of stress

 S : E d   u T fb d    s u T t d 
0 0 o

Variational equation in TL formulation

– Note: load term is similar to linear problems

– Nonlinearity in the strain energy term
• Need to write LHS in terms of u and ū

59
 Variational Formulation cont.
• How to express strain variation
E( u)  21 (C  1)  1
2  0 u  0 uT  0uT 0u 
d
E ( u, u )  E( u  u )
d  0
  0u  0u T  0u T 0 u  0 uT 0u 
1
2

 21  (1   0 uT ) 0 u   0 u T (1   0 u) 
 21  FT  0 u   0 u T F 

E ( u, u )  sym(0 u T F)

E ( u, uis
Note: E(u) is nonlinear, but ) linear

60
 Variational Formulation cont.
• Variational Equation

 S : E d   u T fb d    s u T t d  for all ū  

0 0 o

a( u, u ) ( u )

Energy form Load form

a( u, u )  ( u ), u  Z
• Linear in terms of strain if St. Venant-Kirchhoff material
is used
• Also linear in terms of ū
• Nonlinear in terms of u because displacement-strain
relation is nonlinear
61
 Linearization (Increment)
• Linearization process is similar to variation and/or
differentiation
– First-order Taylor series expansion
– Essential part of Newton-Raphson method
• Let f(xk+1) = f(xk + Duk), where we know xk and want to
calculate Duk
k 1 k df( x)
f( x )  f( x )   uk  H.O.T.
dx
• The first-order derivative is indeed linearization of f(x)

d f
L[f]  f( x  u)   u Linearization
d  0 x

d f
f  f  f( x  u )  u Variation
d  0 x
62
 Linearization of Residual
• We are still in continuum domain (not discretized yet)
• Residual R( u)  a( u, u )  ( u )
• We want to linearize R(u) in the direction of Du
– First, assume that u is perturbed in the direction of Du using a
variable t. Then linearization becomes
T
R( u  u)  R 
L[R( u)]    u
  0  u 

– R(u) is nonlinear w.r.t. u, but L[R(u)] is linear w.r.t. Du

– Iteration k did not converged, and we want to make the residual at
iteration k+1 zero
T
k
 R( u ) 
k 1 k k
R( u )   u  R( u )0
 u  63
 Newton-Raphson Iteration by Linearization
• This is N-R method (see Chapter 2)
k T
 R( u )  k k
   u  R( u ) f
 u  f(xk)
x

f(xk+1)
• Update state uk 1  uk  uk Duk
xk+1 xk x
xk 1  X  uk 1
 R( uk ) 
• We know how to calculate R(uk), but how about  u ? 
 
 
[R( u)]  [a( u, u )  ( u )]
u u
– Only linearization of energy form will be required
– We will address displacement-dependent load later

64
 Linearization cont.
• Linearization of energy form

L[a( u, u )]  L  0 S : E d     [S : E  S : E ] d

   0

– Note that the domain is fixed (undeformed reference)

– Need to express in terms of displacement increment Du
• Stress increment (St. Venant-Kirchhoff material)
S
S  : E  D : E
E

• Strain increment (Green-Lagrange strain)

E  21 ( FT F  FT F)

 x   ( X  u)  u
F            0 u
 X   X  X
65
 Linearization cont.
• Strain increment E  21 ( FT F  FT F)
 21 ( 0 uT F  FT  0 u)
 sym( 0 uT F)
!!! Linear w.r.t. Du
• Inc. strain variation E  [sym( 0 u T F)]
 sym( 0 u T F)
 sym( 0 u T  0 u) !!! Linear w.r.t. Du

• Linearized energy form

L[a( u, u )]   
0
[E : D : E  S : E ] d  a * ( u; u, u )

– Implicitly depends on u, but bilinear w.r.t. Du and ū

– First term: tangent stiffness
– Second term: initial stiffness
66
 Linearization cont.
• N-R Iteration with Incremental Force
– Let tn be the current load step and (k+1) be the current iteration
– Then, the N-R iteration can be done by

a * ( n u k ; u k , u )  ( u )  a( n u k , u ), u  Z

– Update the total displacement

n k 1
u  n uk  uk

• In discrete form
{ d }T [ nKTk ]{ dk }  { d }T { nRk }

• What are [ nKTk ] { nRk }

and ?

67
 Example – Uniaxial Bar
• Kinematics du  u2, du
 u2
dX dX F = 100N
2
du 1  du  1
1 2
2
E11     u  (u )
dX 2  dX  2
2 2
x
L0=1m

• Strainduvariation
du du
E11    u2 (1  u2 )
dX dX dX

W  1 2 
• Strain
W(E11 ) energy )2
 21 E  (E11density and
S11  stress 11
 E  E  E u  (u
 2 2 2  )
E 11  

L0

a(u, u) 
S11E11AdX  S11AL0 (1  u2 )u2
• Energy and load forms
0
(u)  u2F

R  u2  S11AL0 (1  u2 )  F   0, u2
68
 Example – Uniaxial Bar
• Linearization
S11  EE11  E(1  u2 )u2 E11  u2 u2
L0
0  E11  E  E11  S11  E11  AdX
*
a (u; u,u) 
 EAL0 (1  u2 )2 u2 u2  S11AL0u2 u2
• N-R iteration

[E(1  u2k )2  S11

k
]AL0 u2k  F  S11
k
(1  u2k )AL0

u2k 1  usk  u2k

69
 Example – Uniaxial Bar

(a) with initial stiffness

Iteration u Strain Stress conv
0 0.0000 0.0000 0.0000 9.999E01
1 0.5000 0.6250 125.00 7.655E01
2 0.3478 0.4083 81.664 1.014E02
3 0.3252 0.3781 75.616 4.236E06

(b) without initial stiffness

Iteration u Strain Stress conv
0 0.0000 0.0000 0.0000 9.999E01
1 0.5000 0.6250 125.00 7.655E01
2 0.3056 0.3252 70.448 6.442E03
3 0.3291 0.3833 76.651 3.524E04
4 0.3238 0.3762 75.242 1.568E05
5 0.3250 0.3770 75.541 7.314E07

70
 Updated Lagrangian Formulation
• The current configuration is the reference frame
– Remember it is unknown until we solve the problem
– How are we going to integrate if we don’t know integral domain?
• What stress and strain should be used?
– For stress, we can use Cauchy stress (s)
– For strain, engineering strain is a pair of Cauchy stress
– But, it must be defined in the current configuration
T
1  u u 
     sym( x u)
2  x x 
 

71
 Variational Equation in UL
• Instead of deriving a new variational equation, we will
convert from TL equation
 T 
1 1 u T u
  F  S  FT E   FF 
J 2  X X 
 
 S  J F 1    F  T T
1 T  T u u 1 
 F F  F F
2  X X 
 
 T T
1 T X u u X 
 F   F
2  x X X x 
Similarly  
 T
 E  F T    F 1 T u u 
 F   F
T 2  x x 
1  u u   
    
2  x x   FT    F
 
72
 Variational Equation in UL cont.
• Energy Form
a( u, u )   S : E d   (JF 1F T ) : (FT F) d
0 0

Fik1klFjl1Fmi mnFnj  mk nlkl mn  mn mn

 0
S : E d   0
 :  Jd   x
 :  d

– We just showed that material and spatial forms are

mathematically equivalent
• Although they are equivalent, we use different notation:

a( u, u )   x
 :  d Is this linear or nonlinear?

• Variational Equation

a( u, u )  ( u ), u  Z What happens to load form?

73
 Linearization of UL
• Linearization of ax ( u, u ) will be challenging because we don’t
know the current configuration (it is function of u)
• Similar to the energy form, we can convert the linearized
energy form of TL
• Remember a * ( u; u, u )   
0
[E : D : E  S : E ] d 0 
• Initial stiffness term
1  u T u u T u 
S : E  J(F 1F T ) :   
2  X X X X 
 
1

1 1 um um um um 
 JFik klFjl   
2  Xi Xj Xi Xj 
1  um um um um 
 Jkl  
2  xk xl xk xl  kl ( u, u )

74
 Linearization of UL cont.
• Initial stiffness term
S : E  J : ( u, u ) ( u, u )  sym( x u T x u)

• Tangent stiffness term

(E : D : E)  (FT    F) : D : (FT    F)
 Fki klFljDijmnFpm pqFqn
1 
 J kl  FkiFljDijmnFpmFqn  pq
J  4th-order spatial
constitutive tensor
E : D : E  J  : c : 
1
where cijkl  FirFjsFkmFlnDrsmn
J

75
 Spatial Constitutive Tensor
• For St. Venant-Kirchhoff material
D  (1  1)  2I Drsmn  rs mn  (rmsn  rn sm )
• It is possible to show
1
cijkl   bijbkl  (bikbjl  bilbjk )  .
J
• Observation
– D (material) is constant, but c (spatial) is not
– S  D : E,   c:

76
 Linearization of UL cont.
• From equivalence, the energy form is linearized in TL and
converted to UL
L[a( u, u )]   0
[  : c :    : ]J d

a * ( u; u, u )   [  : c :    : ] d
x

• N-R Iteration
a * ( n uk ; uk , u )  ( u )  a( n uk , u ), u  Z
• Observations
– Two formulations are theoretically identical with different
expression
– Numerical implementation will be different
– Different constitutive relation
77
 Example – Uniaxial Bar
• Kinematics F = 100N
du u2 du u2 1 2
 , 
dx 1  u2 dx 1  u2
x
L0=1m

• Deformation gradient: F11  dx  1  u2 , J  1  u2

dX

• Cauchy stress:   1 F S F  E(u  1 u2 )(1  u )

11 11 11 11 2 2 2
J 2

• Strain variation:  (u)  F T E F 1  u2

11 11 11 11
1  u2

• Energy & load forms: a(u, u)  L

0 1111 (u)Adx  11Au2 (u)  u2F

• Residual:
R  u2  11A  F   0, u2

78
 Example – Uniaxial Bar
• Spatial constitutive relation: c1111  1 F11F11F11F11E  (1  u2 )3 E
J
L
• Linearization: 0 11 (u)c111111 ( u)Adx  EA(1  u2 )
2
u2 u2
L 11A
0 1111 ( u, u)Adx 
1  u2
u2u2

L
0  11 (u)c111111 (u)  11(u,u)  Adx
*
a (u; u, u) 
11
 EA(1  u2 )2 u2 u2  Au2 u2
1  u2

Iteration u Strain Stress conv

0 0.0000 0.0000 0.000 9.999E01
1 0.5000 0.3333 187.500 7.655E01
2 0.3478 0.2581 110.068 1.014E02
3 0.3252 0.2454 100.206 4.236E06
79
Section 3.5
Hyperelastic Material Model

80
 Goals
• Understand the definition of hyperelastic material
• Understand strain energy density function and how to use
it to obtain stress
• Understand the role of invariants in hyperelasticity
• Understand how to impose incompressibility
• Understand mixed formulation and perturbed Lagrangian
formulation
• Understand linearization process when strain energy
density is written in terms of invariants

81
 What Is Hyperelasticity?
• Hyperelastic material - stress-strain relationship derives
from a strain energy density function
– Stress is a function of total strain (independent of history)
– Depending on strain energy density, different names are used,
such as Mooney-Rivlin, Ogden, Yeoh, or polynomial model
• Generally comes with incompressibility (J = 1)
– The volume preserves during large deformation
– Mixed formulation – completely incompressible hyperelasticity
– Penalty formulation - nearly incompressible hyperelasticity
• Example: rubber, biological tissues
– nonlinear elastic, isotropic, incompressible and generally
independent of strain rate
• Hypoelastic material: relation is given in terms of stress
and strain rates 82
 Strain Energy Density
• We are interested in isotropic materials
– Material frame indifference: no matter what coordinate system is
chosen, the response of the material is identical
– The components of a deformation tensor depends on coord. system
– Three invariants of C are independent of coord. system
• Invariants of C
I1  tr(C )  C11  C22  C33  12  22  32 No deformation
I1 = 3
I2  21  (tr C )2  tr(C 2 )   1222  2223  2312 I2 = 3
I3 = 1
I3  det C  122223

– In order to be material frame indifferent, material properties

must be expressed using invariants
– For incompressibility, I3 = 1
83
 Strain Energy Density cont.
• Strain Energy Density Function
– Must be zero when C = 1, i.e., l1 = l2 = l3 = 1

W(I1 , I2 , I3 )   Amnk (I1  3)m (I2  3)n (I3  1)k
m  n  k 1

– For incompressible material


W(I1 , I2 )   Amn (I1  3)m (I2  3)n
m n 1

– Ex: Neo-Hookean model


W(I1 )  A10 (I1  3) A10 
– Mooney-Rivlin model 2

W(I1 , I2 )  A10 (I1  3)  A01 (I2  3)

84
 Strain Energy Density cont.
• Strain Energy Density Function
– Yeoh model

W1 (I1 )  A10 (I1  3)  A20 (I1  3)2  A30 (I1  3)3

– Ogden model Initial shear modulus
N N
1

W1 ( 1, 2, 3 )   i 1 i  2 i  3 i  3
i 1 i
  
    ii
2 i 1

– When N = 1 and a1 = 1, Neo-Hookean material

– When N = 2, a1 = 2, and a2 = 2, Mooney-Rivlin material

85
 Example – Neo-Hookean Model
• Uniaxial tension with incompressibility
1   2  3  1 / 
• Energy density
2
W  A10 (I1  3)  A10 (12  22  32  3)  A10 (2   3)

• Nominal stress
W  1   1 
P  2A10    2     1    2 
     (1  ) 
50
Linear elastic
0

-50
Nominal stress

-100

Neo-Hookean
-150

-200

-250
-0.8 -0.4 0 0.4 0.8
Nominal strain
86
 Example – St. Venant Kirchhoff Material
• Show that St. Venant-Kirchhoff material has the following
strain energy density
 2
W(E)   tr(E)   tr(E2 )
2
W(E) tr(E) tr(E2 )
S  tr(E) 
E E E
• First term
tr(E)
tr(E)  1 : E 1
E
tr(E)
tr(E)  1(1 : E)  (1  1) : E
E
• Second term
EijEji
 ik  jlEji  Eij jk il  Elk  Elk  2Elk
Ekl
87
 Example – St. Venant Kirchhoff Material cont.
• Therefore
tr(E) tr(E2 )
S  tr(E) 
E E
 (1  1) : E  2E
  (1  1)  2I  : E

88
 Nearly Incompressible Hyperelasticity
• Incompressible material
– Cannot calculate stress from strain. Why?
• Nearly incompressible material
– Many material show nearly incompressible behavior
– We can use the bulk modulus to model it

• Using I1 and I2 enough for incompressibility?

– No, I1 and I2 actually vary under hydrostatic deformation
– We will use reduced invariants: J1, J2, and J3

J1  I1I31/3 J2  I2I32/3 J3  J  I31/2

• Will J1 and J2 be constant under dilatation?

89
 Locking
• What is locking
– Elements do not want to deform even if forces are applied
– Locking is one of the most common modes of failure in NL analysis
– It is very difficult to find and solutions show strange behaviors
• Types of locking
– Shear locking: shell or beam elements under transverse loading
– Volumetric locking: large elastic and plastic deformation
• Why does locking occur?
– Incompressible sphere under hydrostatic pressure

Pressure No unique pressure

for given displ.
p sphere

Volumetric strain
90
 How to solve locking problems?
• Mixed formulation (incompressibility)
– Can’t interpolate pressure from displacements
– Pressure should be considered as an independent variable
– Becomes the Lagrange multiplier method
– The stiffness matrix becomes positive semi-definite

Displacement

Pressure

4x1 formulation

91
 Penalty Method
• Instead of incompressibility, the material is assumed to be nearly
incompressible
• This is closer to actual observation
• Use a large bulk modulus (penalty parameter) so that a small volume
change causes a large pressure change
• Large penalty term makes the stiffness matrix ill-conditioned
• Ill-conditioned matrix often yields excessive deformation
• Temporarily reduce the penalty term in the stiffness calculation
• Stress calculation use the penalty term as it is

Unique pressure
Pressure

for given displ. 1 

 7 
10
[K]   
 1 
 
 1
Volumetric strain
92
 Example – Hydrostatic Tension (Dilatation)

 x1  X1  0 0   2 0 0 
  
 x2  X2 F   0  0  C0 2 0 
 x  X 0
 3 3  0 0  
 0 2 

• Invariants
I1  32 I2  3 4 I3  6 I1 and I2 are not constant
• Reduced invariants

J1  I1I31/3  3 J1 and J2 are constant

J2  I2I32/3  3
J3  I31/2  3

93
 Strain Energy Density
• Using reduced invariants

W(J1 , J2 , J3 )  WD (J1, J2 )  WH (J3 )

– WD(J1, J2): Distortional strain energy density

– WH(J3): Dilatational strain energy density

• The second terms is related to nearly incompressible

behavior
K
WH (J3 )  (J3  1)2
2
   23 
– K: bulk modulus for linear elastic material
Abaqus: W
1 2
(J
H 3 )  (J  1)
2D 3
94
 Mooney-Rivlin Material
• Most popular model
– (not because accuracy, but because convenience)

W(J1, J2, J3 )  WD (J1, J2 )  WH (J3 )

K
 A10 (J1  3)  A01 (J2  3)  (J3  1)2
2
– Initial shear modulus ~ 2(A10 + A01)
– Initial Young’s modulus ~ 6(A10 + A01) (3D) or 8(A10 + A01) (2D)
– Bulk modulus = K
• Hydrostatic pressure
W WH
p   K(J3  1)
J3 J3

– Numerical instability for large K (volumetric locking)

95
 Mooney-Rivlin Material cont.
• Second P-K stress

W W J1 W J2 W J3

S   
E J1 E J2 E J3 E a
a,E 
S  A10J1,E  A01J2,E  K(J3  1)J3,E E

– Use chain rule of differentiation

J1  I1I31/3
J1,E  (I31/3 )I1,E  31 I1 (I34/3 )I3,E J2  I2I32/3

J2,E  (I32/3 )I2,E  2 I (I 5/3 )I J3  I31/2

3 2 3 3,E

J3,E  21 (I31/2 )I3,E

I1,E  21 I1,E  21
I2,E  4(1  trE)1  4E I2,E  2(I1 1  C )
I3,E  (2  4trE)1  4E  [ 94 eimn ejrsEmrEns ] I3,E  2I3C 1
96
 Example
• Show I1,E  21, I2,E  2(I1 1  C ), I3,E  2I3C 1
• Let I1  tr(C ), I2  21 tr(CC ), I3  31 tr(CCC )
• Then I1  I1 , I2  21 I12  I2 , I3  I3  61 I13  I1I2
• Derivatives
I1 I2 I3
 ij,  Cji,  Cjk Cki
Cij Cij Cij

I1 I2 I3

 ij,  I1ij  Cji,  I3Cji1
Cij Cij Cij

and
 
2
C E

97
 Mixed Formulation
• Using bulk modulus often causes instability
– Selectively reduced integration (Full integration for deviatoric
part, reduced integration for dilatation part)
• Mixed formulation: Independent treatment of pressure
WH (J3, p)  p(J3  1)
– Pressure p is additional unknown (pure incompressible material)
– Advantage: No numerical instability
– Disadvantage: system matrix is not positive definite
• Perturbed Lagrangian formulation
1 2
WH (J3, p)  p(J3  1)  p
2K

– Second term make the material nearly incompressible and the

system matrix positive definite
98
 Variational Equation (Perturbed Lagrangian)
• Stress calculation
1 2
W(J1 , J2, J3 )  A10 (J1  3)  A01 (J2  3)  p(J3  1)  p
2K
S  A10J1,E  A01J2,E  pJ3,E

• Variation of strain energy density

W  W,E E  W,pp
p
 S : E  (J3  1  )p
K
r  ( u, p)
• Introduce a vector of unknowns:
a(r, r )   0
 S : E  pH  d

p
H  J3  1  Volumetric strain
K

99
 Example – Simple Shear
• Calculate 2nd P-K stress for the simple shear deformation
– material properties (A10, A01, K) X2, x2
1 1 0 1 1 0
F   0 1 0  C  FT F   1 2 0 
 0 0 1   0 0 1  45o

I1  4, I2  4, I3  1 X1, x1

I1,E  21
 6 2 0 
I2,E  2(I1 1  C )   2 4 0 
 0 0 6 
 4 2 0 
I3,E  2I3C 1   2 2 0 
 0 0 2 
100
 Example – Simple Shear cont.

 5 4 0 
J1  I1I31/3 4 4 2
J1,E  I1,E  I3,E   4 1 0 
3 3
 0 0 1 
J2  I2I32/3  4
 7 5 0 
2
J2,E  I2,E  83 I3,E   5 2 0 
J3  I31/2  1 3
 0 0 1 

S  A10J1,E  A01J2,E  K(J3  1)J3,E

 5A10  7A01 4A10  5A01 0 

2
  4A10  5A01 A10  2A01 0 

3
 0 0 A10  A01 
Note: S11, S22 and S33 are not zero

101
 Stress Calculation Algorithm
• Given: {E} = {E11, E22, E33, E12, E23, E13}T, {p}, (A10, A01)
{1}  {1 1 1 0 0 0}T { C }  2{E }  {1}

I1  C1  C2  C3
I2  C1C2  C1C3  C2C3  C4 C4  C5 C5  C6C6
I3  (C1C2  C4C4 )C3  (C4 C6  C1C5 )C5  (C4 C5  C2C6 )C6

{I1,E }  2{1 1 1 0}
{I2,E }  2{C2  C3 C3  C1 C1  C2 C4 C5 C6 }
{I3,E }  2{C2C3  C52 C3C1  C62 C1C2  C42
C5C6  C3C4 C6C4  C1C5 C4C5  C2C6 }

{J1,E }  I31/3 {I1,E }  31 I1I34/3 {I3,E }

{J2,E }  I32/3 {I2,E }  23 I2I35/3 {I3,E }
{J3,E }  21 I31/2 {I3,E },
For penalty method, use
{ S }  A10 {J1,E }  A01 {J2,E }  p{J3,E } K(J3 – 1) instead of p
102
 Linearization (Penalty Method)
• Stress increment
S  W,E,E : E  D : E
• Material stiffness

S
D  A10J1,EE  A01J2,EE  K(J3  1)J3,EE  KJ3,E  J3,E
E

• Linearized energy form

a * ( u; u, u )    E : D : E  S : E  d
0

103
 Linearization cont.
• Second-order derivatives of reduced invariants
 31 1  43 4  73 1  34
J1,EE  I1,EE I3  I3 (I1,E  I3,E  I3,E  I1,E )  I1I3 I3,E  I3,E  I1I3 I3,EE
3 9 3
 23 2  53 10  83 2  53
J2,EE  I2,EE I3  I3 (I2,E  I3,E  I3,E  I2,E )  I2I3 I3,E  I3,E  I2I3 I3,EE
3 9 3
1  32 1  21
J3,EE   I3 I3,E  I3,E  I3 I3,EE
4 2

I1,EE  0
I2,EE  41  1  I
I3,EE  4I3C 1  C 1  I3C 1 IC 1

104
 MATLAB Function Mooney
• Calculates S and D for a given deformation gradient

%
% 2nd PK stress and material stiffness for Mooney-Rivlin material
%
function [Stress D] = Mooney(F, A10, A01, K, ltan)
% Inputs:
% F = Deformation gradient [3x3]
% A10, A01, K = Material constants
% ltan = 0 Calculate stress alone;
% 1 Calculate stress and material stiffness
% Outputs:
% Stress = 2nd PK stress [S11, S22, S33, S12, S23, S13];
% D = Material stiffness [6x6]
%

105
 Summary
• Hyperelastic material: strain energy density exists with
incompressible constraint
• In order to be material frame indifferent, material
properties must be expressed using invariants
• Numerical instability (volumetric locking) can occur when
large bulk modulus is used for incompressibility
• Mixed formulation is used for purely incompressibility
(additional pressure variable, non-PD tangent stiffness)
• Perturbed Lagrangian formulation for nearly
incompressibility (reduced integration for pressure term)

106
Section 3.6
Finite Element Formulation for
Nonlinear Elasticity

107
 Voigt Notation
• We will use the Voigt notation because the tensor
notation is not convenient for implementation
– 2nd-order tensor  vector
– 4th-order tensor  matrix
• Stress and strain vectors (Voigt notation)

{ S }  {S11 S22 S12 }T

{E }  {E11 E22 2E12 }T

– Since stress and strain are symmetric, we don’t need 21 component

108
 4-Node Quadrilateral Element in TL
• We will use plane-strain, 4-node quadrilateral element to
discuss implementation of nonlinear elastic FEA
• We will use TL formulation
• UL formulation will be discussed in Chapter 4

X2 4 t
3 (–1,1) (1,1)

1 2
X1 (–1,–1) (1,–1)

Finite Element at Reference Element

undeformed domain

109
 Interpolation and Isoparametric Mapping
• Displacement interpolation
Ne
Nodal displacement vector (uI, vI)
u  NI (s)uI
I 1
Interpolation function
• Isoparametric mapping
– The same interpolation function is used for geometry mapping
Nodal coordinate (XI, YI)
Ne
X  NI (s)XI
I 1

N1  41 (1  s)(1  t)
N2  41 (1  s)(1  t) Interpolation (shape) function

N3  41 (1  s)(1  t) • Same for all elements

• Mapping depends of geometry
N4  41 (1  s)(1  t)
110
 Displacement and Deformation Gradients
• Displacement gradient
Ne Ne
u NI (s)
X
  X uI ui,j   NI,j (s)uIi
I 1 I 1

 0 u  {u1,1 u1,2 u2,1 u2,2 }T

NI ( s)
– How to calculate ?
X

• Deformation gradient
{F }  {F11 F12 F21 F22 }T  {1  u1,1 u1,2 u2,1 1  u2,2 } T

– Both displacement and deformation gradients are not symmetric

111
 Green-Lagrange Strain
• Green-Lagrange strain

1
 E11   u1,1  2 (u1,1u1,1  u2,1u2,1 ) 
   
{E }   E22    u2,2  21 (u1,2u2,1  u2,2u2,2 ) 
 2E   
 12 u
  1,2  u2,1  u u
1,2 1,1  u u
2,1 2,2 

– Due to nonlinearity, {E }  [B]{ d}

{ S }  [D]{E }
– For St. Venant-Kirchhoff material,

   2  0
[D]      2 0 
 0 0  

112
 Variation of G-R Strain
• Although E(u) is nonlinear, E ( u, u ) is linear

E ( u, u )  sym( 0 u T F) {E }  [BN ]{ d }

 
 F11N1,1 F21N1,1 F11N2,1 F21N2,1  F11N4,1 F21N4,1 
 
 
[BN ]   F12N1,2 F22N1,2 F12N2,2 F22N2,2  F12N4,2 F22N4,2 
 
 
F N
 11 1,2 F N
21 1,2 F N
11 2,2 F N
21 2,2 F N
11 4,2 F N
21 4,2 

 F N F12N4,1 F22N4,1 
 12 1,1 F22N1,1 F12N2,1 F22N2,1
Function of u
Different from linear strain-displacement matrix

113
 Variational Equation
• Energy form
a( u, u )   0
S : E d

 { d }T  [BN ]T { S } d
0

 { d }T {Fint }
• Load form

( u )   u T fb d   S u T t d
0 0

 uIT   
Ne
 NI ( s)fb d   S NI (s)t d
0 0
I 1
 { d }T { Fext }

• Residual
{ d }T { Fint ( d)}  { d }T {F ext }, { d }  Zh
114
 Linearization – Tangent Stiffness
• Incremental strain { E }  [BN ]{ d}

• Linearization
T  T  { d}
0 E : D : E d   { d }
  0 N
[B ] [D][BN ] d 


T  T  { d}
0 S : E d   { d }
  0 G
[B ] [ ][BG ] d 


 S11 S12 0 0 
S S22 0 0 
[]   12
 0 0 S11 S12 
 
 0 0 S12 S22 

 N1,1 0 N2,1 0 N3,1 0 N4,1 0 

 
 N1,2 0 N2,2 0 N3,2 0 N4,2 0 
[BG ] 
 0 N 0 N2,1 0 N3,1 0 N4,1 
1,1
 
 0 N2,1 0 N2,2 0 N3,2 0 N4,2  115
 Linearization – Tangent Stiffness
• Tangent stiffness
T T
[KT ]  0  N
 [B ] [D][BN ]  [B 
G [][BG ]  d0
]

• Discrete incremental equation (N-R iteration)

{ d }T [KT ]{ d}  { d }T {F ext  Fint }, { d }  Zh

– [KT] changes according to stress and strain

– Solved iteratively until the residual term vanishes

116
 Summary
• For elastic material, the variational equation can be
obtained from the principle of minimum potential energy
• St. Venant-Kirchhoff material has linear relationship
between 2nd P-K stress and G-L strain
• In TL, nonlinearity comes from nonlinear strain-
displacement relation
• In UL, nonlinearity comes from constitutive relation and
unknown current domain (Jacobian of deformation
gradient)
• TL and UL are mathematically equivalent, but have
different reference frames
• TL and UL have different interpretation of constitutive
relation.
117
Section 3.7
MATLAB Code for
Hyperelastic Material Model

118
 HYPER3D.m
• Building the tangent stiffness matrix, [K], and the residual
force vector, {R}, for hyperelastic material
• Input variables for HYPER3D.m
Variable Array size Meaning
MID Integer Material Identification No. (3) (Not used)
PROP (3,1) Material properties (A10, A01, K)
UPDATE Logical variable If true, save stress values
LTAN Logical variable If true, calculate the global stiffness matrix
NE Integer Total number of elements
NDOF Integer Dimension of problem (3)
XYZ (3,NNODE) Coordinates of all nodes
LE (8,NE) Element connectivity

119
function HYPER3D(MID, PROP, UPDATE, LTAN, NE, NDOF, XYZ, LE)
%***********************************************************************
% MAIN PROGRAM COMPUTING GLOBAL STIFFNESS MATRIX AND RESIDUAL FORCE FOR
% HYPERELASTIC MATERIAL MODELS
%***********************************************************************
%%
global DISPTD FORCE GKF SIGMA
%
% Integration points and weights
XG=[-0.57735026918963D0, 0.57735026918963D0];
WGT=[1.00000000000000D0, 1.00000000000000D0];
%
% Index for history variables (each integration pt)
INTN=0;
%
%LOOP OVER ELEMENTS, THIS IS MAIN LOOP TO COMPUTE K AND F
for IE=1:NE
% Nodal coordinates and incremental displacements
ELXY=XYZ(LE(IE,:),:);
% Local to global mapping
IDOF=zeros(1,24);
for I=1:8
II=(I-1)*NDOF+1;
IDOF(II:II+2)=(LE(IE,I)-1)*NDOF+1:(LE(IE,I)-1)*NDOF+3;
end
DSP=DISPTD(IDOF);
DSP=reshape(DSP,NDOF,8);
%
%LOOP OVER INTEGRATION POINTS
for LX=1:2, for LY=1:2, for LZ=1:2
E1=XG(LX); E2=XG(LY); E3=XG(LZ);
INTN = INTN + 1;
%
% Determinant and shape function derivatives
[~, SHPD, DET] = SHAPEL([E1 E2 E3], ELXY);
FAC=WGT(LX)*WGT(LY)*WGT(LZ)*DET; 120
 % Deformation gradient
F=DSP*SHPD' + eye(3);
%
% Computer stress and tangent stiffness
[STRESS DTAN] = Mooney(F, PROP(1), PROP(2), PROP(3), LTAN);
%
% Store stress into the global array
if UPDATE
SIGMA(:,INTN)=STRESS;
continue;
end
%
% Add residual force and tangent stiffness matrix
BM=zeros(6,24); BG=zeros(9,24);
for I=1:8
COL=(I-1)*3+1:(I-1)*3+3;
BM(:,COL)=[SHPD(1,I)*F(1,1) SHPD(1,I)*F(2,1) SHPD(1,I)*F(3,1);
SHPD(2,I)*F(1,2) SHPD(2,I)*F(2,2) SHPD(2,I)*F(3,2);
SHPD(3,I)*F(1,3) SHPD(3,I)*F(2,3) SHPD(3,I)*F(3,3);
SHPD(1,I)*F(1,2)+SHPD(2,I)*F(1,1)
SHPD(1,I)*F(2,2)+SHPD(2,I)*F(2,1) SHPD(1,I)*F(3,2)+SHPD(2,I)*F(3,1);
SHPD(2,I)*F(1,3)+SHPD(3,I)*F(1,2)
SHPD(2,I)*F(2,3)+SHPD(3,I)*F(2,2) SHPD(2,I)*F(3,3)+SHPD(3,I)*F(3,2);
SHPD(1,I)*F(1,3)+SHPD(3,I)*F(1,1)
SHPD(1,I)*F(2,3)+SHPD(3,I)*F(2,1) SHPD(1,I)*F(3,3)+SHPD(3,I)*F(3,1)];
%
BG(:,COL)=[SHPD(1,I) 0 0;
SHPD(2,I) 0 0;
SHPD(3,I) 0 0;
0 SHPD(1,I) 0;
0 SHPD(2,I) 0;
0 SHPD(3,I) 0;
0 0 SHPD(1,I);
0 0 SHPD(2,I);
0 0 SHPD(3,I)];
end 121
 %
% Residual forces
FORCE(IDOF) = FORCE(IDOF) - FAC*BM'*STRESS;
%
% Tangent stiffness
if LTAN
SIG=[STRESS(1) STRESS(4) STRESS(6);
STRESS(4) STRESS(2) STRESS(5);
STRESS(6) STRESS(5) STRESS(3)];
SHEAD=zeros(9);
SHEAD(1:3,1:3)=SIG;
SHEAD(4:6,4:6)=SIG;
SHEAD(7:9,7:9)=SIG;
%
EKF = BM'*DTAN*BM + BG'*SHEAD*BG;
GKF(IDOF,IDOF)=GKF(IDOF,IDOF)+FAC*EKF;
end
end; end; end;
end
end

122
 Example Extension of a Unit Cube
• Face 4 is extended with a stretch ratio l = 6.0
• BC: u1 = 0 at Face 6, u2 = 0 at Face 3, and u3 = 0 at Face 1
• Mooney-Rivlin: A10 = 80MPa, A01 = 20MPa, and K = 107
X2
% Nodal coordinates
XYZ=[0 0 0;1 0 0;1 1 0;0 1 0;0 0 1;1 0 1;1 1 1;0 1 1]; Face 1
% 4 3
% Element connectivity
LE=[1 2 3 4 5 6 7 8];
% 8 7 Face 4
% No external force
EXTFORCE=[]; Face 6 2
% X1
% Prescribed displacements [Node, DOF, Value] 1
SDISPT=[1 1 0;4 1 0;5 1 0;8 1 0; % u1=0 for Face 6
1 2 0;2 2 0;5 2 0;6 2 0; % u2=0 for Face 3 5 6
1 3 0;2 3 0;3 3 0;4 3 0; % u3=0 for Face 1 X3 Face 3
2 1 5;3 1 5;6 1 5;7 1 5]; % u1=5 for Face 4
%
% Load increments [Start End Increment InitialFactor FinalFactor]
TIMS=[0.0 1.0 0.05 0.0 1.0]';
%
% Material properties
MID=-1;
PROP=[80 20 1E7];
123
 Example Extension of a Unit Cube
Time Time step Iter Residual
0.05000 5.000e-02 2 1.17493e+05
Not converged. Bisecting load increment 2
 
Time Time step Iter Residual
0.02500 2.500e-02 2 2.96114e+04
3 2.55611e+02 6000
4 1.84747e-02
5 1.51867e-10 5000
 
Time Time step Iter Residual 4000

Stress
0.05000 2.500e-02 2 2.48106e+04
3 1.69171e+02 3000
4 7.67766e-03
5 2.39898e-10 2000
 
Time Time step Iter Residual 1000
0.10000 5.000e-02 2 8.45251e+04
3 1.88898e+03 0
4 8.72537e-01 1 2 3 4 5 6
Stretch ratio
5 1.86783e-07
...
Time Time step Iter Residual
1.00000 5.000e-02 2 8.55549e+03
3 8.98726e+00
4 9.88176e-06
5 1.66042e-09 124
 Hyperelastic Material Analysis Using ABAQUS
• *ELEMENT,TYPE=C3D8RH,ELSET=ONE
– 8-node linear brick, reduced integration with hourglass control,
hybrid with constant pressure
• *MATERIAL,NAME=MOONEY
*HYPERELASTIC, MOONEY-RIVLIN
80., 20.,
– Mooney-Rivlin material with A10 = 80 and A01 = 20
• *STATIC,DIRECT
– Fixed time step (no automatic time step control)
y

z 125
 Hyperelastic Material Analysis Using ABAQUS
*HEADING *HYPERELASTIC, MOONEY-RIVLIN
- Incompressible hyperelasticity (Mooney- 80., 20.,
Rivlin) Uniaxial tension *STEP,NLGEOM,INC=20
*NODE,NSET=ALL UNIAXIAL TENSION
1, *STATIC,DIRECT
2,1. 1.,20.
3,1.,1., *BOUNDARY,OP=NEW
4,0.,1., FACE1,3
5,0.,0.,1. FACE3,2
6,1.,0.,1. FACE6,1
7,1.,1.,1. FACE4,1,1,5.
8,0.,1.,1. *EL PRINT,F=1
*NSET,NSET=FACE1 S,
1,2,3,4 E,
*NSET,NSET=FACE3 *NODE PRINT,F=1
1,2,5,6 U,RF
*NSET,NSET=FACE4 *OUTPUT,FIELD,FREQ=1
2,3,6,7 *ELEMENT OUTPUT
*NSET,NSET=FACE6 S,E
4,1,8,5 *OUTPUT,FIELD,FREQ=1
*ELEMENT,TYPE=C3D8RH,ELSET=ONE *NODE OUTPUT 126
 Hyperelastic Material Analysis Using ABAQUS
• Analytical solution procedure
– Gradually increase the principal stretch l from 1 to 6
– Deformation gradient
 0 0 
 
F  0 1 /  0 
0 0 1 /  


– Calculate J1,E and J2,E

– Calculate 2nd P-K stress
S  A10J1,E  A01J2,E

1
F  S  FT
  stress
– Calculate Cauchy
J

11  11  22

– Remove the hydrostatic component of stress 127
 Hyperelastic Material Analysis Using ABAQUS
• Comparison with analytical stress vs. numerical stress

128
Section 3.9
Fitting Hyperelastic Material
Parameters from Test Data

129
 Elastomer Test Procedures
• Elastomer tests
– simple tension, simple compression, equi-biaxial tension, simple
shear, pure shear, and volumetric compression

70
uni-axial
bi-
60
pure
axialshear

50
Nominal stress

40

30

20

10

0
0 1 2 3 4 5 6 7
Nominal strain
130
 Elastomer Tests
• Data type: Nominal stress vs. principal stretch

F L F
L

Simple tension test

Pure shear test F
F

F
L

Volumetric compression test

Equal biaxial test
131
 Data Preparation
• Need enough number of independent experimental data
– No rank deficiency for curve fitting algorithm
• All tests measure principal stress and principle stretch

Experiment Type Stretch Stress

Uniaxial tension Stretch ratio l = L/L0 Nominal stress TE = F/A0

Equi-biaxial Stretch ratio l = L/L0 in y- Nominal stress TE = F/A0

tension direction in y-direction

Pure shear test Stretch ratio l = L/L0 Nominal stress TE = F/A0

Volumetric test Compression ratio l = L/L0 Pressure TE = F/A0

132
 Data Preparation cont.
• Uni-axial test 1  , 2   3  1 / 
U
T   2(1   3 )(A10   A01 )

T  2 3   A10 
T(A10 ,A01 , )  { x } { b}   2(    ) 2(1   )   
 A01 

• Equi-biaxial test     , 3  1 /  2
1 2
1 U
T   2(   5 )(A10  2A01 )
2 
• Pure shear test   ,   1,   1 / 
1 2 3

U
T   2(   3 )(A10  A01 )


133
 Data Preparation cont.
• Data Preparation

Type 1 1 1  4 4  4
 1 2 3  i i1  NDT
T E T1E T2E T3E  Ti E TiE1 E
 TNDT

• For Mooney-Rivlin material model, nominal stress is a

linear function of material parameters (A10, A01)

134
 Curve Fitting for Mooney-Rivlin Material
• Need to determine A10 and A01 by minimizing error
between test data and model
NDT
 
2
minimize TkE  T(A10 , A01, k )
A10 ,A01 k 1

• For Mooney-Rivlin, T(A10, A01, lk) is linear function

– Least-squares can be used  A10 
{ b}   
 A01 
 T1E   T1   x(1 )T 
 E   T   T

E  T2   2   x(1 ) 
{T }    {T}    {b}  [ X]{b}

       

 TE   TNDT   x( T
 NDT   NDT ) 
135
 Curve Fitting cont.
• Minimize error(square)

{ e }T { e }  { T E  T }T { T E  T }
 { T E  Xb}T { T E  Xb}
 { T E }T { T E }  2{b}T [ X]T { T E }  {b}T [ X]T [ X]{b}

• Minimization  Linear regression equation

[ X]T [ X]{b}  [ X]T { T E }

136
 Stability of Constitutive Model
• Stable material: the slope in the stress-strain curve is
always positive (Drucker stability)
• Stability requirement (Mooney-Rivlin material)

d : D : d   0
• Stability check is normally performed at several specified
deformations (principal directions)
d1d1  d2d2  0

 D11 D12   d1 

 d1 d2  D   0
 21 D22   d2 
• In order to be P.D.
D11  D22  0
D11D22  D12D21  0
137