FINITE ELEMENT : MATRIX FORMULATION

Georges Cailletaud
Ecole des Mines de Paris, Centre des Matériaux
UMR CNRS 7633

Contents 1/67
Contents

Discrete versus continuous

Element
Interpolation
Element list

Global problem
Formulation
Matrix formulation
Algorithm
Continuous→ Discrete→Continuous

How much rain ?

Discrete versus continuous 3/67

Continuous→ Discrete→Continuous

Geometry discretization

Discrete versus continuous 4/67

Continuous→ Discrete→Continuous

Unknown field discretization

Discrete versus continuous 5/67

Continuous→ Discrete→Continuous

Use elements

Discrete versus continuous 6/67

Finite Element Discretization

Replace continuum formulation by a discrete representation for unknowns

and geometry
Unknown field: X
ue (M) = Nie (M)qei
i

Geometry: X
x(M) = Ni∗e (M)x(Pi )
i

Interpolation functions and shape functions Ni∗e such as:

Nie
X
∀M, Nie (M) = 1 and Nie (Pj ) = δij
i

Isoparametric elements iff Nie ≡ Ni∗e

Discrete versus continuous 7/67

Contents

Discrete versus continuous

Element
Interpolation
Element list

Global problem
Formulation
Matrix formulation
Algorithm
2D-mapping

Subparametric Superparametric Isoparametric

element element element

Geometry Geometry Geometry

Unknown field Unknown field Unknown field

more field nodes more geometrical nodes same number of

than geometrical nodes than field nodes geom and field nodes

Rigid body displacement not represented for superparametric element that has nonlinear edges !
The location of the node at the middle of the edge is critical for quadratic edges

Element 9/67
Shape function matrix, [N] – Deformation matrix, [B]

Field u, T , C
Gradient ∼ε, grad(T ),. . .
Constitutive equations σ
∼
=Λ
∼
:∼
ε, q = −kgrad(T )
∼

Conservation div (σ
∼
) + f = 0, . . .
First step: express the continuous field and its gradient wrt the
discretized vector

Element 10/67
Deformation matrix [B] (1)

Knowing: X
ue (M) = Nie (M)qei
i

Deformation can be obtained from the nodal displacements, for

instance in 2D, small strain:
∂ux ∂N1 (M) e ∂N2 (M) e
εxx = = q1x + q2x + . . .
∂x ∂x ∂x
∂uy ∂N1 (M) e ∂N2 (M) e
εyy = = q1y + q2y + . . .
∂y ∂y ∂y
∂ux ∂uy ∂N1 (M) e ∂N2 (M) e
2εxy = + = q1x + q1y + . . .
∂y ∂x ∂y ∂x

Element 11/67
Deformation matrix [B] (2)

Matrix form, 4-node quadrilateral

e
 
q1x
e 
 
e e N1 0 N2 0 N3 0 N4 0  q1y
{u} = [N]T {q} =  
0 N1 0 N2 0 N3 0 N4  ... 
e
q4y

e e
{ε} = [B]T {q}
e
 
  q1x
N1,x 0 N2,x 0 N3,x 0 N4,x 0  e 
= 0 N1,y 0 N2,y 0 N3,y 0 N4,y   q1y 
 ... 
N1,y N1,x N2,y N2,x N3,y N3,x N4,y N4,x e
q4y
Shear term taken as γ = 2ε12

Element 12/67
Reference element

η
η
1

ξ
ξ
−1 1
y

x −1

Actual geometry Reference element

Physical space (x, y ) Parent space (ξ, η)

Z Z +1 Z +1
f (x, y )dxdy = f∗ (ξ, η)Jdξdη
Ω −1 −1

J is the determinant of the partial derivatives ∂x/∂ξ. . . matrix

Element 13/67
Remarks on geometrical mapping

The values on an edge depends only on the nodal values on the same
edge (linear interpolation equal to zero on each side for 2-node lines,
parabolic interpolation equal to zero for 3 points for 3-node lines)
Continuity...
The mid node is used to allow non linear geometries
Limits in the admissible mapping for avoiding singularities

Element 14/67
Mapping of a 3-node line

1.5
x2 =0
1

0.5 x2 =1/2
x
0 x2 =1

-0.5

-1
-1 -0.5 0 0.5 1
ξ

Physical segment: x1 =-1 x3 =1 −1 6 x2 6 1

Parent segment: ξ1 =-1 ξ3 =1 ξ2 =0
“ ”
x = ξ + 1 − ξ 2 x2

Element 15/67
Jacobian and inverse jacobian matrix

∂x ∂x  
 
   
dx  ∂ξ ∂η  dξ = [J] dξ
=  ∂y ∂y  dη
dy dη
∂ξ ∂η

∂ξ ∂ξ  
 
   
dξ  ∂x ∂y  dx = [J]−1 dx
=  ∂η ∂η  dy
dη dy
∂x ∂y

Since (x, y ) are known from Ni (ξ, η) and xi ,

−1
[J] is computed from the known quantities in [J], using also:

∂x ∂y ∂y ∂x
J = Det ([J]) = −
∂ξ ∂η ∂ξ ∂η

Element 16/67
Expression of the inverse jacobian matrix

∂y ∂y
 
1  ∂η −
−1
[J] =  ∂x ∂ξ 
J − ∂x 
∂η ∂ξ
For a rectangle [±a, ±b] in the ”real world”, the mapping function is
the same for any point inside the rectangle. The jacobian is a
diagonal matrix, with ∂x/∂ξ = a, ∂y /∂η = b, and the determinant
value is ab
For any other shape, the ”mapping” changes according to the
location in the element
For computing [B], one has to consider ∂Ni /∂x and ∂Ni /∂y :

∂Ni ∂Ni ∂ξ ∂Ni ∂η

= +    
∂x ∂ξ ∂x ∂η ∂x then ∂Ni /∂x −1 ∂Ni /∂ξ
= [J]
∂Ni ∂Ni ∂ξ ∂Ni ∂η ∂Ni /∂y ∂Ni /∂η
= +
∂y ∂ξ ∂y ∂η ∂y
Element 17/67
Contents

Discrete versus continuous

Element
Interpolation
Element list

Global problem
Formulation
Matrix formulation
Algorithm
2D solid elements

Type shape interpol # of polynom

of disp nodes terms
C2D3 tri lin 3 1, ξ, η
C2D4 quad lin 4 1, ξ, η, ξη
C2D6 tri quad 6 1, ξ, η, ξ 2 , ξη, η 2
C2D8 quad quad 8 1, ξ, η, ξ 2 , ξη, η 2 , ξ 2 η, ξη 2
C2D9 quad quad 9 1, ξ, η, ξ 2 , ξη, η 2 , ξ 2 η, ξη 2 , ξ 2 η 2

Element 19/67
3D solid elements

Type shape interpol # of polynom

of disp nodes terms
C3D4 tetra lin 4 1, ξ, η, ζ
C3D6 tri prism lin 6 1, ξ, η, ζ, ξη, ηζ
C3D8 hexa lin 8 1, ξ, η, ζ, ξη, ηζ, ζξ, ξηζ
C3D10 tetra quad 10 1, ξ, η, ζ, ξ 2 , ξη, η 2 , ηζ, ζ 2 , ζξ
C3D15 tri prism quad 15 1, ξ, η, ζ, ξη, ηζ, ξ 2 ζ, ξηζ, η 2 ζ, ζ 2 ,
ξζ 2 , ηζ 2 , ξ 2 ζ 2 , ξηζ 2 , η 2 ζ 2
C3D20 hexa quad 20 1, ξ, η, ζ, ξ 2 , ξη, η 2 , ηζ, ζ 2 , ζξ,
ξ 2 η, ξη 2 , η 2 ζ, ηζ 2 , ξζ 2 , ξ 2 ζ, ξηζ,
ξ 2 ηζ, ξη 2 ζ, ξηζ 2
C3D27 hexa quad 27 ξ i η j ζ k , (i, j, k) ∈ 0, 1, 2

Element 20/67
Isoparametric representation
Example: 2D plane stress elements with n nodes

Element geometry
Xn n
X n
X
1= Ni x= N i xi y= N i yi
i=1 i=1 i=1
Displacement interpolation
X n n
X
ux = Ni uxi uy = Ni u y i
i=1 i=1

Matrix form 
    N1
1 1 1 1 ... 1  
 x   x1 x2 x3 ... xn   N2 
     N3 
 y  =  y1 y2 y3 ... yn   
    . 
ux  ux1 ux2 ux3 ... uxn 
 
 . 
uy uy 1 uy 2 uy 3 ... uyn
Nn

Element 21/67
The linear triangle

IFEM–Felippa
Terms in 1, ξ, η
Element 22/67
The bilinear quad

IFEM–Felippa
Terms in 1, ξ, η, ξη

Element 23/67
The quadratic triangle

IFEM–Felippa
Terms in 1, ξ, η, ξ 2 , ξη, η 2
Element 24/67
The biquadratic quad

IFEM–Felippa
Terms in 1, ξ, η, ξ 2 , ξη, η 2 , ξ 2 η, ξη 2 , ξ 2 η 2

Element 25/67
The 8-node quad

IFEM–Felippa
1
Corner nodes: Ni = (1 + ξξi )(1 + ηηi )(ξξi + ηηi − 1)
4
1
Mid nodes, ξi = 0: Ni = (1 − ξ 2 )(1 + ηηi )
2
1
Mid nodes, ηi = 0: nI = (1 − η 2 )(1 + ξξi )
2
Terms in 1, ξ, η, ξ 2 , ξη, η 2 , ξ 2 η, ξη 2

Element 26/67
Approximated field

1
ξ η
Polynomial basis
ξ2 ξη η2
3 2 2
ξ ξ η ξη η3

Examples :

C2D4 (1 + ξi ξ)(1 + ηi η)
C2D8, corner 0.25(−1 + ξi ξ + ηi η)(1 + ξi ξ)(1 + ηi η)
C2D8 middle 0.25(1. − ξ 2 )(1. + ηi η)

Element 27/67
The 2-node infinite element

Displacement is assumed to be q1 at node 1 and q2 = 0 at node 2

x ξ

x1 x2 → ∞ 1 2
Interpolation
1−ξ 1+ξ
N1 = N2 =
2 2
Geometry
1+ξ
N1∗ such as x = x1 + α N2∗ = 0
1−ξ
ξ =?
Resulting displacement interpolation
u(x) =??

Element 28/67
The 2-node infinite element

Displacement is assumed to be q1 at node 1 and q2 = 0 at node 2

x ξ

x1 x2 → ∞ 1 2
Interpolation
1−ξ 1+ξ
N1 = N2 =
2 2
Geometry
1+ξ
N1∗ such as x = x1 + α N2∗ = 0
1−ξ
x − x1 − α
ξ=
x − x1 + α
Resulting displacement interpolation
u(x) =?

Element 29/67
The 2-node infinite element

Displacement is assumed to be q1 at node 1 and q2 = 0 at node 2

x ξ

x1 x2 → ∞ 1 2
Interpolation
1−ξ 1+ξ
N1 = N2 =
2 2
Geometry
1+ξ
N1∗ such as x = x1 + α N2∗ = 0
1−ξ
x − x1 − α
ξ=
x − x1 + α
Resulting displacement interpolation
αq1
u(x) = N1 (x) q1 = N1 (ξ(x)) q1 =
x − x1 + α
Element 30/67
Connecting element

7 6 5
Connection between
a linear and a
4
quadratic quad
lin. quad.

1 2 3
Quadratic interpolation with node number 8 in the middle of 1–7:
u(M) = N1 q1 + N8 q8 + N7 q7
On edge 1–7, in the linear element, the displacement should verify:
q8 =?
Overloaded shape function in nodes 1 and 7 after suppressing node
8:
u(M) =??

Element 31/67
Connecting element

7 6 5
Connection between
a linear and a
4
quadratic quad
lin. quad.

1 2 3
Quadratic interpolation with node number 8 in the middle of 1–7:
u(M) = N1 q1 + N8 q8 + N7 q7
On edge 1–7, in the linear element, the displacement should verify:
q8 = (q1 + q7 )/2
Overloaded shape function in nodes 1 and 7 after suppressing node
8:
u(M) =??

Element 32/67
Connecting element

7 6 5
Connection between
a linear and a
4
quadratic quad
lin. quad.

1 2 3
Quadratic interpolation with node number 8 in the middle of 1–7:
u(M) = N1 q1 + N8 q8 + N7 q7
On edge 1–7, in the linear element, the displacement should verify:
q8 = (q1 + q7 )/2
Overloaded shape function in nodes 1 and 7 after suppressing node
8:    
N8 N8
u(M) = N1 + q1 + N7 + q7
2 2
Element 33/67
Contents

Discrete versus continuous

Element
Interpolation
Element list

Global problem
Formulation
Matrix formulation
Algorithm
Thermal conduction

Strong form:
”GIVEN r : Ω → R, a volumetric flux,
φd : Γf → R, a surface flux,
Td : Γu → R, a prescribed temperature,
FIND T : Ω → R, the temperature, such as:”
in Ω φi,i =r
on Γu T = Td
on ΓF −φi ni = Φd

Constitutive equation (Fourier, flux (W /m2 ) proportional to the

temperature gradient)

φi = −κij T ,j conductivity matrix: [κ] (W /m.K )

Global problem 35/67

Thermal conduction (2)

Weak form:
S, trial solution space, such as T = T d on Γu
V, variation space, such as δT = 0 on Γu
”GIVEN r : Ω → R, a volumetric flux,
Φd : Γf → R, a surface flux,
T d : Γu → R, a prescribed temperature,
FIND T ∈ S such as ∀δT ∈ V
Z Z Z
− φi δT ,i dΩ = δTrdΩ + δT Φd dΓ
Ω Ω ΓF

”For any temperature variation compatible with prescribed temperature

field around a state which respects equilibrium, the internal power
variation is equal to the external power variation: δT ,i φi is in W /m3 ”
T is present in φi = −κij T ,j

Global problem 36/67

Elastostatic

Strong form:

volume Ω with prescribed volume forces f d : σij,j + fi = 0

surface ΓF with prescribed forces Fd : Fid = σij nj
surface Γu with prescribed displacements ud : ui = uid
Constitutive equation: σij = Λijkl εkl = Λijkl uk,l
So that: Λijkl uk,lj + fi = 0

Global problem 37/67

Principle of virtual power

Weak form:

volume V with prescribed volume forces : f d

surface ΓF with prescribed forces : Fd
surface Γu with prescribed displacements : ud
Virtual displacement rate u̇ kinematically admissible (u̇ = u̇d on Γu )
The variation u̇ is such as: u̇ = 0 on Γu . Galerkin form writes, ∀u̇:
Z Z Z
d
σ
∼
:∼ε̇ dΩ = f u̇ dΩ + Fd u̇ dS
Ω Ω ΓF

Global problem 38/67

Discrete form of virtual power

Application of Galerkin approach for continuum mechanics:

virtual displacement rate u̇ ≡ w h ; σ

∼
≡ u h ,x

{u̇ e }, nodal displacements allow us to compute u̇ and ∼

ε̇:

u̇ = [N]{u̇ e } ; ε̇ = [B]{u̇ e }
∼

Galerkin form writes, ∀{u̇ e }:

X Z  X Z
e e
{σ({u }).[B].{u̇ } dΩ = (
f d .[N].{u̇ e } dΩ
elt Ω elt Ω
Z
+ Fd .[N].{u̇ e } dS )
ΓF

Global problem 39/67

Internal and external forces
In each element e:
Internal forces:
Z Z
e
{Fint }= {σ({u e }).[B] dΩ = [B]T {σ({u e }) dΩ
Ω Ω
External forces:
Z Z
e
{Fext }= f d .[N]dΩ + Fd .[N]dS
Ω ΓF

e
The solution of the problem: {Fint ({u e })} = {Fext
e
} with Newton
iterative algorithm will use the jacobian matrix :
e
∂{Fint }
[K e ] =
∂{u e }
Z
∂{σ} ∂{ε}
= [B]T . . dΩ
Ω ∂{ε} ∂{u e }
Z
∂{σ}
= [B]T . .[B] dΩ
Ω ∂{ε}

Global problem 40/67

Linear and non linear behavior

Applying the principle of virtual power ≡ Stationnary point of

Potential Energy
For elastic behavior
Z
[K e ] = [B]T .[Λ∼
] .[B] dΩ
∼
Ω

is symmetric, positive definite (true since σ

∼
and ∼
ε are conjugated)
 
∂{σ}
For non linear behavior, one has to examine [Lc ] = . Note
∂{ε}
that [Lc ] can be approached (quasi-Newton).
e
{Fext } may depend on {u e } (large displacements).

Global problem 41/67

Elastostatic, strong and weak form, a summary

BCu Displacements Body forces

u f

Kinematics Equilibrium

Strains Constitutive Stresses BCs

ε equations σ

STRONG WEAK
d
BCu: u = u on Γu BCu: u h = u d on Γu
Kinematics: ε = [B] u in Ω Kinematics: ε = [B] u h in Ω
Constitutive equation: Constitutive equation:
σ = Λε σ = Λε
Equilibrium: [B] σ + f = 0 Equilibrium: δΠ = 0
BCs: σn = F on ΓF BCs: δΠ = 0

Global problem 42/67

Contents

Discrete versus continuous

Element
Interpolation
Element list

Global problem
Formulation
Matrix formulation
Algorithm
Matrix–vectors formulation of the weak form of the
problem

[K ] {q} = {F }
Thermal conduction:
Z Z Z
T
[K ] = [B] [κ] [B] dΩ {F } = [N] rdΩ + [N] Φd dΓ
Ω Ω ∂Ω
Elasticity:
Z Z Z
T
[K ] = [B] [Λ] [B] dΩ {F } = [N] f d dΩ + [N] Fd dΓ
Ω Ω ∂Ω
In each element e:
Internal forces:
Z Z
e
{Fint }= {σ({u e }).[B] dΩ = [B]T {σ({u e }) dΩ
Ω Ω
External forces:
Z Z
d
e
{Fext } = f .[N]dΩ + Fd .[N]dS
Global problem Ω ΓF 44/67
The stiffness matrix

Example of a 4-node quad and of a 20-node hexahedron ()

T
[B] [D] [B] [K ]
3 (6) 3 (6) 8 (60) 8 (60)

8 3 3 = 8
(60) (6) (6) (60)

The element stiffness matrix is a square matrix, symmetric, with no zero

inside.
Its size is equal to the number of dof of the element.

Global problem 45/67

Nodal forces (1)

Z
T
e
{Fext }= [N] F d dS
ΓF

η
η
Ft Fn
7 1

6 5

8
ξ
4 ξ
−1 1

1 3
2
−1

Fx ds = Ft dx − Fn dy
   
dx dξ
with = [J]
Fy ds = Fn dx + Ft dy dy dη

Global problem 46/67

Nodal forces (2)
∂x ∂y
Integration on edge 5–7: dx =
dξ dy = dξ
∂ξ ∂ξ
Components 9, 10, for the nodes 5; 11, 12 for nodes 6; 13, 14 for nodes 7

Z 1  
e ∂x ∂y
Fext (2i − 1) = e Ni − FnFt dξ
−1 ∂ξ ∂ξ
Z 1  
e ∂x ∂y
Fext (2i) = e Ni F n + Ft dξ
−1 ∂ξ ∂ξ
Example, for a pressure Fn = p, and no shear (Ft = 0) on the 5–7 edge
of a 8-node rectangle

−a 6 x 6 a y =b represented by − 1 6 ξ 6 1 η = 1
∂x ∂y
=a =0
∂ξ ∂ξ
N5 = ξ(1 + ξ)/2 N6 = 1 − ξ 2 N7 = −ξ(1 − ξ)/2

Global problem 47/67

Nodal forces (3)

(4)

(1) (1)

7 6 5 7 6 5

Z 1
1 ap
F10 = F5y = e ξ(1 + ξ)padξ =
−1 2 3
Z1
4ap
F12 = F6y = e (1 − ξ 2 )padξ =
−1 3

The nodal forces at the middle node are 4 times the nodal forces at
corner nodes for an uniform pressure (distribution 1–2–1–2–1... after
adding the contribution of each element)

Global problem 48/67

Nodal forces (4)

Axisymmetric 8-node quad

z
(2)

(1)

7 6 5
7 6 5
Face of a 20-node hexahedron
(4) (−1) (4)

(−1) (−1)
(4)
(4)

(−1)

Global problem 49/67

Nodal forces (5)

Face of a 27-node hexahedron

who knows ?

Face of a 15-node hexahedron

(3)
(3)

(3)

Global problem 50/67

Assembling the global matrix

6 7
4 3

C
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
Local versus global numbering

Global problem 51/67

Assembling the global matrix

F1A q1A
0 1 0 1
F1 = q1 =
BF2
B = F2A +F1B C Bq2 = q2A = q1B C
+F2B
C B
= q2B
C
BF3 = C Bq3 = C
B C B C
BF4
B = F3A +F4B C
+F1 CC
Bq4
B = q3A = q4B C
= q1 CC
BF5
B = +F3B +F2C C
C
Bq5
B = = q3B = q2C C
C
@F6 = C
+F A @q6 = =q AC
4 4
F7 = +F3C q7 = = q3C
6 7
4 3

C
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3

Global problem 52/67

Assembling the global matrix

11

6 7
4 3

C
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 53/67

Assembling the global matrix

11 12

22
21

6 7
4 3

C
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 54/67

Assembling the global matrix

11 12 13

22 23
21

6 7
32 4 3
31
33
C
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 55/67

Assembling the global matrix

11 12 13

22 23
21
11

6 7
32 4 3
31
33
C
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 56/67

Assembling the global matrix

11 12 13

22 23
21
11

6 7
32 4 3
31
33
C
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 57/67

Assembling the global matrix

11 12 13

22 23
21 12
11

21 22

6 7
32 4 3
31
33
C
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 58/67

Assembling the global matrix

11 12 13

22 23
21 12 13
11

21 22 23

6 7
32 4 3
31
33
C
33
31 32
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 59/67

Assembling the global matrix

11 12 13

22 23
21 12 13
11 14

21 22 24 23

6 7
32 44 43 3
31 42
11 4
41 33
C
34 33
31 32
1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 60/67

Assembling the global matrix

11 12 13

22 23
21 12 13
11 14

21 22 24 23

6 7
32 44 43 3
31 42
11 4
41 33 12
C
34 33
31 32
21 22 1 2
4 5
4 3
3

B
A
1 2 1 2
1 2 3
1

Global problem 61/67

Assembling the global matrix

11 12 13

22 23
21 12 13
11 14

21 22 24 23

6 7
32 44 43 3
31 42
11
13 4
41 33 12
C
34 33
31 32 23
21 22 1 2
4 5
4 3
3

B
A
31 32 33
1 2 1 2
1 2 3
1

Global problem 62/67

Assembling the global matrix

11 12 13

22 23
21 12 13
11 14

21 22 24 23

6 7
32 44 43 3
31 42
11
14 13 4
41 33 12
C
34 33
31 32 24 23
21 22 1 2
4 5
4 3
42 44 43 3
41
B
A
31 32 34 33
1 2 1 2
1 2 3
1

Global problem 63/67

Contents

Discrete versus continuous

Element
Interpolation
Element list

Global problem
Formulation
Matrix formulation
Algorithm
Global algorithm

For each loading increment, do while k{R}iter k > EPSI :

iter = 0; iter < ITERMAX ; iter + +
1 Update displacements: ∆{u}iter +1 = ∆{u}iter + δ{u}iter
2 Compute ∆{ε} = [B].∆{u}iter +1 then ∆ε ∼
for each Gauss point
∆σ
3 Integrate the constitutive equation: ∆ε∼
→ ∆σ∼
, ∆αI , ∼
∆ε∼
4 Compute int and ext forces: {Fint ({u}t + ∆{u}iter +1 )} , {Fe }
5 Compute the residual force: {R}iter +1 = {Fint } − {Fe }
6 New displacement increment: δ{u}iter +1 = −[K ]−1 .{R}iter +1

Global problem 65/67

Convergence

Value of the residual forces < R , e.g.

!1/n
X
||{R}||n = Rin ; ||{R}||∞ = max |Ri |
i
i

Relative values:

||{R}i − {R}e ||
<
||{R}e ||
Displacements
 
{U} − {U}k n < U
k+1

Energy
 T
{U}k+1 − {U}k . {R}k < W

Global problem 66/67