Lecture3 PartB

Total lagrangian formulation of a bar element and
path following methods
Prof. Dr. Eleni Chatzi

Dr. Giuseppe Abbiati, Dr. Konstantinos Agathos
Lecture 5 - 26 October, 2017
Institute of Structural Engineering, ETH Zürich
October 26, 2017
Institute of Structural Engineering Method of Finite Elements II 1

Outline
1 Introduction
2 Weak form
3 Finite element formulation
4 Total Lagrangian formulation of a bar element
5 Path-following Methods

Learnig goals
Understanding how geometrically nonlinear finite elements are

formulated
Gaining a basic understanding of path following solution

methods
Gaining a basic understanding of how the above are

implemented
Gaining a basic understanding of how the above can be used in

commercial software (ABAQUS)

Significance of the lecture
Applications:
Structures undergoing large displacements and/or

rotations such as cables, arches and shells
Materials such as elastomers and biological/soft tissue
Modeling of plastically deforming materials

Weak form
We recall:
R R R
Weak form (TL): S : δEdV = F · δudV + T · δudA
V V A
Linearized weak form:
Z Z Z Z Z
i
e·D·δedV + S : δηdV = F·δudV + T·δudA− Si : δedV
V V V A V
where:
h T i
1
e = 2 ∇∆u + ∇∆uT + ∇ ui ∇∆u + ∇∆uT ∇ui
h i
1
η = 2 ∇∆uT ∇∆u

Finite element formulation
Coordinates and displacements are interpolated from the
corresponding nodal quantities using FE interpolation:
N
X N
X
Xi = NJ XiJ , xi = NJ xiJ
J=1 J=1
N
X N
X
ui = NJ uiJ , ∆ui = NJ ∆uiJ
J=1 J=1
or in matrix form:
X = NXn , x = Nxn
ui = Nuni , ∆u = N∆un
Index J and superscript n refer to nodal quantities, NJ are the FE

shape functions.
Typically isoparametric elements are used:
NJ = NJ (ξ1 , ξ2 , ξ3 )
where ξ1 , ξ2 , ξ3 are the isoparametric coordinates. Then:
        

∂·   ∂X1 ∂X2 ∂X3   ∂·   ∂·  
∂· 
 ∂ξ   ∂ξ ∂ξ1 ∂ξ1   ∂X   ∂X   ∂ξ 
 1  1  1  1  1
        
 ∂·   ∂X1 ∂X2 ∂X3   ∂·   ∂·  −1  ∂· 
 =  ⇒ =J  
 ∂ξ   ∂ξ ∂ξ2 ∂ξ2   ∂ξ 
  ∂X2 
   ∂X 
 2  2  2  2
        
 ∂·   ∂X1 ∂X2 ∂X3   ∂·   ∂·   ∂· 
∂ξ3 ∂ξ3 ∂ξ3 ∂ξ3 ∂X3 ∂X3 ∂ξ3
| {z }
J
The same transformation can be used for the current configuration.

Utilizing the above, the quantities ∇u, ∇∆u can be computed.
For example:
N
∂u1 X ∂NJ J
= u
∂X1 J=1 ∂X1 i
∂NJ ∂NJ ∂NJ ∂NJ

= J−1
11 + J−1
12 + J−1
13
∂X1 ∂ξ1 ∂ξ2 ∂ξ3
where J−1 −1
ij is the element ij of matrix Jij .

Utilizing the above, the linear and non-linear part of the strain
increment can be written in terms of the nodal displacement
increments:
e = BL ∆un ⇒ δe = BL δ∆un
1
η = ∆unT BT n
NL BNL ∆u ⇒ δη = δ∆u
nT T
BNL BNL ∆un
2

By taking into account that:
u = ui + ∆u ⇒ δu = δ∆u
Substituting the above in the weak form we obtain:
Z Z
nT
δ∆u BT
L DBL dV ∆u
n
+ δ∆u nT
BT i n
NL S BNL dV ∆u =
V V
Z Z Z
nT T nT
δ∆u N FdV + δ∆u N TdA − δ∆unT
T
BT i
L Ŝ dV
V A V

u = ui + ∆u ⇒ δu = δ∆u
Z Z
nT
BT n nT
BT i n

δ∆u

L DBL dV ∆u +
δ∆u

NL S BNL dV ∆u =
V V
Z Z Z
nT T nT T nT BT i

δ∆u
N FdV +
δ∆u
N TdA −
δ∆u

L Ŝ dV
V A V

u = ui + ∆u ⇒ δu = δ∆u
Z Z
BT
L DBL dV ∆u
n
+ BT i n
NL S BNL dV ∆u =
V V
Z Z Z
T
N FdV + NT TdA − BT i
L Ŝ dV
V A V

u = ui + ∆u ⇒ δu = δ∆u
KT
z
 }| {
 
Z Z 
 BT DBL dV + BT Si BNL dV  ∆un =
 
 L NL 
 
V V 
| {z } | {z }
KL KNL
−Ri
zZ Z }| Z {
T T
N FdV + N TdA − BT i
L Ŝ dV
V A V
| {z } | {z }
fext fint
The tangent stiffness matrix, residual and force vectors are then
written as:
KT = KL + KNL
Z Z
KL = BT
L DBL dV , KNL = BT i
NL S BNL dV
V V
−Ri = fext − fint i

Z Z Z
T T i
fext = N FdV + N TdA, fint = BT i
L Ŝ dV
V A V

Substituting the above, the linearized equilibrium equation can be

obtained in terms of the nodal unknowns:
KT ∆un = −Ri
KT ∆un = fext − fint i

Bar element
Next the tangent stiffness matrix and force vector are developed for
the two noded bar element.
It is convenient to write:
u12 = u11 + l cos θ − L
u22 = u21 + l sin θ

Bar element
FE shape function for a two noded bar:

1−ξ 1+ξ
N=
2 2
L
Jacobian: J =
2
Shape function derivative:
N ∂N 1h i
N,1 = = J −1 = −1 1
∂X1 ∂ξ L

Bar element
By considering displacements along both directions:
 
u11
1−ξ 1+ξ
     
 1
 u1   2 0 0    u2 

 = 2 ·
1−ξ 1 + ξ  2
 
u2 0 0  u1 

2 2  
u22
| {z }
N
and
 
  u11
∂u1    
 1
 
 ∂X1  1  −1 0 1  u2 
0   
 =  · 
 ∂u2  L 0 −1 0 1  2
 u1 
∂X1 | {z }  
N,1 u22

Bar element
Displacement derivative:
 
1
∂ui 1h i
 ui  ui2 − ui1
= N,1 uni = −1 1 · =
∂X1 L ui2 L
For i = 1:
∂u1 u 2 − u11 u 1 + l cos θ − L − u11 l cos θ − L l cos θ

= 1 = 1 = = −1
∂X1 L L L L
For i = 2:
∂u2 u 2 − u21 u 1 + l sin θ − u21 l sin θ

= 2 = 2 =
∂X1 L L L

Bar element
∂ (·)
By taking into account that = 0 the Green-Lagrange strain at
∂X2
the previous step (only component 11 is of interest) assumes the
form:
∂u1 1 ∂u1 2 1 ∂u2 2

E11 = + +
∂X1 2 ∂X1 2 ∂X1
2
l cos θ 1 l cos θ 1 l sin θ 2

= −1 + −1 +
L 2 L 2 L
2
l −L 2
=
2L2

Bar element
∂ (·)
By taking into account that = 0 the linear part of the strain
∂X2
increment (only component 11 is of interest) assumes the form:
∂∆u1 ∂u1 ∂∆u1 ∂u2 ∂∆u2

e11 = + +
∂X1 ∂X1 ∂X1 ∂X1 ∂X1

Bar element
1

e11 = −1 0 1 0 +
L
| {z }
∂∆u1
∂X1
l cos θ 1

+ −1 −1 0 1 0 +
L L
| {z }| {z }
∂u1 ∂∆u1
∂X1 ∂X1
 
∆u11
 
 ∆u 1 
 
l sin θ 1

2
+ 0 −1 0 1

 
L L 2
}  ∆u1 

| {z }| {z
∂u2 ∂∆u2
 
∂X1 ∂X1 ∆u22

Bar element
 
∆u11
 
i  ∆u 1 
 
l h 2
e11 = − cos θ − sin θ cos θ sin θ 

2

|L }  ∆u12 
 
{z
BL  
∆u22
| {z }
∆un

Bar element
The nonlinear part of the strain increment is:
∂∆u1
 
" 2 2 #
1 ∂∆u1 ∂∆u2 ∂∆u1 ∂∆u2

 ∂X 
η= + = · 1 
2 ∂X1 ∂X1 ∂X1 ∂X1  ∂∆u2 
∂X1
 
∆u11
∂∆u1
     
1
 
 ∂X 
1 
1  −1 0 1 0   ∆u2 

 ∂∆u2  = L  ·
 

0 −1 0 1 2
 ∆u1 
 
∂X1 | {z }  
BNL =N,1 ∆u22
| {z }
∆un

Bar element
If a linear stress strain relation is employed then:
l 2 − L2
S11 = E E11 = E
2L2
where E is Young’s modulus
Then:
∂S11
=E
∂E11

Bar element
Linear part of the stiffness matrix:

Z
KL = BT
L E BL dV
V
Employing that V = AL, where A the cross section of the bar in the
reference configuration, we obtain:
cos 2 θ −cos 2 θ
 
cosθsinθ −cosθsinθ
l2  2
sin θ −sinθcosθ −sin2 θ 
KL = EA
 
2
L3 cos θ sinθcosθ 
 

Symm sin2 θ

Bar element
Similarly the nonlinear part of the tangent stiffness matrix is:

Z
KNL = BT
NL S11 BNL dV
V
By carrying out the integration we obtain:
 
1 0 −1 0
l 2 − L2  0 1 0 −1 
KNL = EA
 
2L3 −1 0 1 0 
 

0 −1 0 1

Bar element
Finally the force vector can be obtained as:

Z
f= BT
L S11 dV
V
Performing the integration:
−cosθ
 
 
l 2 − L2  −sinθ 
fint = EA
 
2L2
 
 cosθ 
 
sinθ

The Newton-Raphson method
We recall the linearized equilibrium equation:
KT ∆un = −Ri = fext − fint i
Typically external loads are scaled by a factor λ:
KT ∆un = −Ri = λfext − fint i

For a given value of λ, the above corresponds to a Newton-Raphson

iteration where:
∂Ri
R = Ri + ∆un = fint i − λfext + KT ∆un = 0
∂un
The iteration is repeated until
|R| <= tol

where tol is the tolerance.

Typically:
Different (increasing) values are given to the load factor
Displacements are obtained using the Newton-Raphson method
The response of the structure is obtained for progressively

increasing loading
The load-displacement curve of the structure can be obtained

The above procedure is called load control and can be illustrated as

follows:
Increment load
Apply load increment
Solve with
Newton-Raphson
Repeat for next

increment


follows:
Increment load
Solve with
Newton-Raphson
Repeat for next

increment


follows:
Increment load
Solve with
Newton-Raphson
Repeat for next

increment


follows:
Increment load
Solve with
Newton-Raphson
Repeat for next

increment


follows:
Increment load
Solve with
Newton-Raphson
Repeat for next

increment


follows:
Increment load
Solve with
Newton-Raphson
Repeat for next

increment

In many cases the solution path is not monotonic
Phenomena such as snap-through and snap-back can be present
The Newton-Raphson method will either fail or not provide the

full path in those cases
Such phenomena are often of interest in practice e.g.when the

overcritical behavior of a structure needs to be known

Example
Solution path involving snap-through:

Example

Example

Example

Path-following methods
In this category of methods:
It is attempted to follow the full solution path including

snap-back and snap-through phenomena
The load factor is included as an additional unknown in the

nonlinear system of equations
An additional equation (constraint) is added in the system
The choice of this constraint leads to different methods

After adding ∆λ as an unknown and the constraint as an equation

the new system of equations is obtained:
" # " #
R (u, λ) 0
=
g (u, λ) 0
where:
u is the vector of nodal unknowns
λ is the load factor
R is the residual of the Newton-Raphson method
g is the additional constraint

The above system can be solved by employing the Newton-Raphson

method. The linearization of the system yields:
 
∂R ui , λi ∂R ui , λi

" #
 R ui , λi + ∆u + ∆λ  0
∂u ∂λ =
 
i i i i 0


i i
∂g u , λ ∂g u , λ 
g u ,λ + ∆u + ∆λ
∂u ∂λ
where:
∆u is the nodal displacement increment
∆λ is the load factor increment

The above system can be solved by employing the Newton-Raphson

method. The linearization of the system yields:
" #" # " #
KT −fext ∆u Ri
=
hT s ∆λ −g i
where:
∆u is the nodal displacement increment
∆λ is the load factor increment
∂g
h Is the gradient of g with respect to u: h =
∂u
∂g
s is the derivative of g with respect to λ: s =
∂λ

The coefficient matrix in this system is not symmetric.
To exploit the symmetry and sparsity of KT the following

partitioning procedure is employed:
∆uI = KT −1 fext , ∆uII = KT −1 Ri
g i + hT ∆uII
∆λ = − , ∆u = ∆λ∆uI + ∆uII
s + hT ∆uI

The choice of the additional constraint g is crucial
Different alternatives exist
In some cases problem specific constraints are needed
Some widely used alternatives are demonstrated in the following

Path-following methods: Load control
Setting g = λ − λ̄ results in load control:
∂g ∂g
g = λ − λ̄ ⇒ h = = 0, s = =1
∂u ∂λ
g i = λi − λ̄
" #" # " #

KT −fext ∆u Ri
=
0 1 ∆λ λ̄ − λi

∆λ = λ̄ − λi , ∆u = KT −1 λ̄fext − fint i

Path-following methods: Displacement control
Setting g = T · u − ū results in displacement control:
∂g ∂g
g = T · u − ū ⇒ h = = TT , s = =0
∂u ∂λ
g i = T · ui − ū
" #" # " #

KT −fext ∆u Ri
=
T 0 ∆λ ū − T · ui
h i
In the above T = 0 0 . . . 1 . . . 0 0 where the entry 1
corresponds to a selected dof

Displacement control in a path Displacement control in a path

involving snap-through involving snap-back





Path-following methods: Arc-length
q
Setting g = (u − u0 )T · (u − u0 ) + (λ − λ0 )2 − ∆s results in arc
length control (Crisfield 1981):
q
gi = (ui − u0 )T · (ui − u0 ) + (λi − λ0 )2 − ∆s
ui − u0 λi − λ0

∂g ∂g
h= = , s= =
∂u g ∂λ g
 
KT −fext " # " #
T  ∆u Ri
 u − u0
i i 0
λ − λ  ∆λ = −g i

g g
In the above, superscript 0 refers to the beginning of the step

We observe that for i = 0 the system becomes:
 
KT −fext " # " #
T  ∆u Ri
 u − u0
0 0 0
λ − λ  ∆λ = −g i

g g

" #" # " #

KT −fext ∆u Ri
=
0 0 ∆λ −g i

" #" # " #

KT −fext ∆u Ri
=
0 0 ∆λ −g i
The coefficient matrix is singular!
To obtain the load factor at the first iteration a predictor step is

introduced.

For the predictor step the system is solved for the external loads:
∆up = KT −1 fext
Then the increment of the load factor is computed as:
∆s
∆λp = ±
k∆up k

The sign of the increment is determined by the current stiffness

parameter:
fext T ∆u
κ=
∆uT ∆u

Setting g = (∆up )T u − u1 + ∆λp λ − λ1 results in an

alternative method for arc length control (Riks 1972):

g i = (∆up )T ui − u1 + ∆λp λi − λ1
∂g ∂g
h= = ∆up , s = = ∆λp
∂u ∂λ
" #" # " #

KT −fext ∆u Ri
=
(∆up )T ∆λp ∆λ −g i
In the above, superscript 1 refers to the predictor step

Illustration of arc-length methods:
Crisfield: Riks:

Lecture3 PartB

Uploaded by

Copyright:

Available Formats

Lecture3 PartB

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture3 PartB

Uploaded by

Copyright:

Available Formats

Total lagrangian formulation of a bar element and

path following methods

Prof. Dr. Eleni Chatzi

Institute of Structural Engineering, ETH Zürich

October 26, 2017

Institute of Structural Engineering Method of Finite Elements II 1

3 Finite element formulation

4 Total Lagrangian formulation of a bar element

Institute of Structural Engineering Method of Finite Elements II 2

Understanding how geometrically nonlinear finite elements are

Gaining a basic understanding of path following solution

Gaining a basic understanding of how the above are

Gaining a basic understanding of how the above can be used in

Institute of Structural Engineering Method of Finite Elements II 3

Structures undergoing large displacements and/or

Materials such as elastomers and biological/soft tissue

Modeling of plastically deforming materials

Institute of Structural Engineering Method of Finite Elements II 4

Linearized weak form:

Institute of Structural Engineering Method of Finite Elements II 5

Index J and superscript n refer to nodal quantities, NJ are the FE

Typically isoparametric elements are used:

The same transformation can be used for the current configuration.

Utilizing the above, the quantities ∇u, ∇∆u can be computed.

∂NJ ∂NJ ∂NJ ∂NJ

Institute of Structural Engineering Method of Finite Elements II 8

Institute of Structural Engineering Method of Finite Elements II 9

By taking into account that:

Institute of Structural Engineering Method of Finite Elements II 10

By taking into account that:

Institute of Structural Engineering Method of Finite Elements II 10

By taking into account that:

Institute of Structural Engineering Method of Finite Elements II 10

−Ri = fext − fint i

Institute of Structural Engineering Method of Finite Elements II 11

Substituting the above, the linearized equilibrium equation can be

KT ∆un = fext − fint i

Institute of Structural Engineering Method of Finite Elements II 12

u12 = u11 + l cos θ − L

u22 = u21 + l sin θ

Institute of Structural Engineering Method of Finite Elements II 13

FE shape function for a two noded bar:

Shape function derivative:

Institute of Structural Engineering Method of Finite Elements II 14

Institute of Structural Engineering Method of Finite Elements II 15

∂u1 u 2 − u11 u 1 + l cos θ − L − u11 l cos θ − L l cos θ

∂u2 u 2 − u21 u 1 + l sin θ − u21 l sin θ

Institute of Structural Engineering Method of Finite Elements II 16

∂u1 1 ∂u1 2 1 ∂u2 2

Institute of Structural Engineering Method of Finite Elements II 17

∂∆u1 ∂u1 ∂∆u1 ∂u2 ∂∆u2

Institute of Structural Engineering Method of Finite Elements II 18

Institute of Structural Engineering Method of Finite Elements II 19

Institute of Structural Engineering Method of Finite Elements II 20

The nonlinear part of the strain increment is:

Institute of Structural Engineering Method of Finite Elements II 21

If a linear stress strain relation is employed then:

Institute of Structural Engineering Method of Finite Elements II 22

Linear part of the stiffness matrix: