Finite Element Method
Formulations
Computational Methods
Idealization
Mathematical Models
A model is a symbolic device built to simulate
and predict aspects of behavior of a system.
Abstraction of physical reality
Dicretization
Finite Element Discretization
The problem is discretized
Physical meaning is conserved on elements
Interpretation is easier
Solution
1. Linear System Solution Algorithms
Gauss-Jordan Elimination
LU Factorization
Jacobi method
Gauss-Seidel method
2. Error Estimation and Convergence
Analysis
Interpretation
1. Physical Interpretation:
The continous physical model is divided into
finite pieces called elements and laws of
nature are applied on the generic element.
The results are then recombined to
represent the continuum.
2. Mathematical Interpretation:
The differetional equation representing the
system is converted into a variational form,
which is approximated by the linear
combination of a finite set of trial functions.
FEM Notation
Elements are defined by the following
properties:
1. Dimensionality
2. Nodal Points
3. Geometry
4. Degrees of Freedom
5. Nodal Forces
(Non homogeneous RHS of the DE)
Element Types
Variational Formulation
of Bar Element
(Energy Approach)
Axially Loaded Member
Axially Loaded Member
Axially Loaded Member
Strain Energy
Hookes Law:
( x) E ( x)
where
du
( x)
dx
Strain Energy Density:
1
( x) ( x)
2
Strain Energy (continued)
Integrating over the Volume of the Bar:
1
1 L
1 L
U dV p dx ( EAu ' )u ' dx
2 V
2 0
2 0
1 L
u ' EAu ' dx
2 0
All quantities may depend on x.
External Energy
Due to applied external loads
1. The distributed load q(x)
2. The point end load P. This can be
included in q.
External Energy:
L
W qu dx
0
VARIATIONAL FORMULATION
The differential equations for the one
dimensional bar can also be described
as follows :
d
du
L[u ] ( p ( x) ) q ( x),
dx
dx
0 x 1
u (0) u (1) 0.
Recall in Lecture Notes No.1:
-1/dx (a du/dx) = f
for 0 < x < L
(1)
Weighted Residual Methods
It follows that:
L[u] q 0
Multiplying this by a weight function v and
integrating over the whole domain we
obtain:
1
( L[u ] q)v dx (v, L[u ] q) 0
0
Equation (2) is called variational form
( 2)
Approximating Function
We can replace u and v in the formula with
their approximation function i.e.
n
u ( x) U ( x) u j N j ( x)
j 1
v( x) V ( x) d jy j ( x)
j 1
The functions Nj and yj are of our choice and
are meant to be suitable to the particular
problem.
U is called trial function and V is called test function
The Residual
Replacing v and u with respectively V and U
1
( L[u] q)v dx (v, L[u] q) 0
0
(2)
becomes
(V , r ) 0,
V Sn
r ( x) L[U ] q
r(x) is called the residual (as the name of the method
suggests)
The vanishing inner product shows that the residual is
orthogonal to all functions V in the test space.
Applying Discretization:
N
Substituting
V ( x) d jy j ( x)
j 1
(V , L[U ] q) 0
into
and exchanging summations and integrals we
obtain
N
d
j 1
(y j , L[U ] q ) 0
j 1, 2 , ... , n
As the inner product equation is satisfied for all
choices of V in Sn, the above equation has to be
valid for all choices of dj which implies that
(y j , L[U ] q ) 0
j 1, 2 , ... , n
Galerkins Method
One obvious choice of y j would be taking it equal
to N j
This Choice leads to the Galerkins Method
( N j , L[U ] q ) 0
j 1, 2 , ... , n
This form of the problem is called the strong form of
the problem. Because the so chosen test space has
more continuity than necessary.
Therefore it is worthwile for this and other reasons to
convert the problem into a more symmetrical form
This can be acheived by integrating by parts the
initial strong form of the problem.
The Strong Form
Let us remember the initial form of the problem
L[u ]
d
du
( p ( x) ) q ( x),
dx
dx
0 x 1
u (0) u (1) 0.
The Weak Form
1
(v, L[u ] q) v [( pu ' )'q] dx
0
The Weak Form
1
(v, L[u ] q ) v( pu ' )' dx vq dx
Integrating by parts
Recall:
d (ab) ab ad (b)
av
b pu '
v [( pu ' )'q] dx vpu' (v' pu ') dx
1
v [( pu ' )'q] dx (v' pu 'vq) dx vpu' 0 0
1
The Weak Form
The problem can be rewritten as
A(v , u) (v , q) 0
1
where
A(v , u ) (v' pu ') dx
0
(v , q ) (vq) dx
0
The integration by parts eliminated the second derivatives
from the problem making it possible less continouity than the
previous form. This is why this form is called weak form of the
problem.
A(v,u) is called Strain Energy.
Solution Space
Now that derivative of v comes into the picture v needs
to have more continoutiy. As we want to keep symmetry
its appropriate to choose functions that produce
bounded values of
1
A(u , u ) p (u ' ) 2 dx
0
As p is necessarily smooth functions the following
restriction is sufficient
1
2
(
u
'
)
dx
0
We require v and u to satisfy boundary conditions.
Linear System of Equations
The solution now takes the form
A(v , u ) (v , q )
More explicitly substituting U and V (remember we
chose them to have the same base) and swapping
summations and integrals we obtain
N
u
k 1
A( N j , N k ) ( N j , q ) ,
j 1, 2 , ... , n
Connection to the Physical System
Mechanical
Formulation
1 L
U u ' EAu ' dx
2 0
L
Mathematical
Formulation
A(v , u ) v' pu ' dx
1
W qu dx
(v , q )
U W 0
A(v , u) (v , q) 0
