More Details on
Variational Formulation
Axially Loaded Member
Axially Loaded Member
The Strong Form
Let us remember the initial form of the problem
L[u ]
d
du
( p ( x) ) z ( x)u q ( x),
dx
dx
0 x 1
u (0) u (1) 0.
1
(v, L[u ] q) v [( pu ' )' zu q] dx
0
Integrating by parts
1
v [( pu' )' zu q] dx (v' pu'vzu vq) dx vpu'
0
1
0
The Weak Form
The problem can be rewritten as
A(v , u) (v , q) 0
where
A(v , u ) (v' pu 'vzu) 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.
Finite Element Discretization
Let us take the initial value problem with constant
coefficients
pu ' ' zu q( x),
0 x 1
p, z 0
u (0) u (1) 0.
As a first step let us divide the domain in n
subintervals with the following mesh
0 x0 x1 ... xn 1
Each subinterval
element.
( x j 1 , x j ), j 1 : n
is called finite
The Trial Basis
The problem at this point can be easily solved
using the previously derived Galerkins Method
N
u
k 1
A( N j , N k ) ( N j , q ) ,
j 1, 2 , ... , n
A little more work is needed to convert this problem
into matrix notation
Matrix Form of the Problem
Restricting U over the typical finite element we can write
U ( x) u j 1 N j 1 ( x) u j N j ( x)
x [ x j 1 , x j ]
Which in turn can be written as
U ( x) [u j 1
u j 1
N j 1 ( x)
u j ]
[ N j 1 ( x) N j ( x)]
(
)
u
N
x
j
j
x [ x j 1 , x j ]
in the same way
V ( x) [d j 1
N j 1 ( x)
d j 1
d j ]
[ N j 1 ( x) N j ( x)] d
N
(
x
)
j
x [ x j 1 , x j ]
Matrix Form of the Problem
Taking the derivative
U ' ( x) [u j 1
1 / h j
u j 1
u j ]
[1 / h j 1 / h j ] u
h
1
/
j
j
h j x j x j 1
x [ x j 1 , x j ]
Derivative of V is analogus
V ' ( x) [d j 1
1 / h j
d j 1
[1 / h j 1 / h j ]
d j ]
1
/
h
d
j
x [ x j 1 , x j ]
Matrix Form of the Problem
The variational formula can be elementwise defined as
follows:
N
[ A (V ,U ) (V , q)
j 1
]0
A j (V , U ) A (V , U ) A (V , U )
S
j
M
j
A (V , U )
S
j
xj
x j 1
A (V , U )
M
j
xj
x j 1
xj
pV 'U ' dx
zVU dx
(V , q ) Vq dx
x j 1
The Element Stiffnes Matrix
Substituting U,V,U and V into these formulae we obtain
A (V , U ) [d j 1
S
j
A (V , U ) [d j 1
S
j
xj
d j ]
x j1
xj p
d j ]
x j1 h 2
j
u j 1
1 / h j
[1 / h j 1 / h j ] dx
p
1
/
h
u
j
u j 1
1 1 u j 1
1 1 dx u [d j 1 d j ]K j u
j
p
Kj
hj
1 1
1 1
The Element Mass Matrix
The same way
x j N j 1
A (V , U ) [d j 1 d j ] z
[ N j 1
x j1 N j
u j 1
S
A j (V , U ) [d j 1 d j ]M j
u
j
M
j
zh j 2 1
Mj
6 1 2
u j 1
N j ] dx
uj
External Work Integral
The external work integral cannot be evaluated for every
xj
function q(x)
(V , q) Vq dx
x j 1
We can consider a linear interpolant of q(x) for simplicity.
q ( x) q j 1 N j 1 ( x) q j N j ( x),
x [ x j 1 , x j ]
Substituting and evaluating the integral
N j 1
q j 1
dx [d j 1 , d j ] l j
(V , q) j [d j 1 , d j ]
[ N j 1 , N j ]
x j 1
Nj
qj
h j 2q j 1 q j
Element load vector
lj
6 q j 1 2q j
xj
Assembling
Now the task is to assemble the elements into
the whole system in fact we have to sum each
integral over all the elements
For doing so we can extend the dimension of
each element matrix to n and then put the 2x2
matrix at the appropriate position inside it
With all matrices and vectors having the same
dimension the summation looks like
Assembling
N
S
T
A
d
Ku
j
j 1
u1
u
u 2
...
un 1
d1
d
d 2
...
d n 1
1 1
1 2 1
1 2 1
p
K
... ... ...
h
1 2 1
1 1
Assembling
Doing the same for the Mass Matrix
N
M
T
A
d
Mu
j
j 1
2 1
1 4 1
zh 1 4 1
M
... ... ...
6
1 4 1
1 2
Assembling
Doing the same for the Load Vector
N
(V , q)
j 1
dT f
2q0 q1
F0
q 4q q F
1
2
0
1
h q1 4q2 q3 F2
f
...
2
...
qn 2 4qn 1 qn Fn 1
qn 1 2qn Fn applied nodal forces