CH09 Variational Formulation
CH09 Variational Formulation
and
Variational Formulations in FEM
Jayadeep U. B.
M.E.D., NIT Calicut
Ref.:
1. Forray, Marvin J., Variational Calculus in Science and
Engineering, McGraw Hill International Edn.
2. Reddy, J. N., An Introduction to the Finite Element Method,
McGraw Hill International Edition.
3. Zienkiewicz, O. C., and Morgan, K., Finite Elements and
Approximation, John Wiley & Sons.
Introduction
The Fundamental Idea:
Consider the problem of the bending of a cantilever under
loads as in figure below:
x
w(x)
The displacement function w(x) at equilibrium will be the one,
corresponding to which, the internal forces, generated due to
resistance of beam to deform, and the external forces are
equal.
From an energy perspective, the equilibrium of the system
corresponds to minimum potential energy.
Lecture - 01
Introduction contd.
Assuming that the potential energy (P.E.) of the initial system
to be zero, P.E. of the final system has two parts:
Strain energy stored in the cantilever beam
Reduction in P.E. of the force
If the stiffness of the beam and the force are constant and
known, the P.E. becomes a function of only the displacement,
which by itself is a function.
For example, in this problem of the cantilever, the admissible
displacements, satisfying the specified B.C. are the functions
dw
with:
w(0) 0 &
dx
x 0
Introduction contd.
Hence, the analysis problem is solved, if we can find the
displacement function, which minimizes the P.E. of the
system, from a space of admissible functions.
To be more precise, P.E. is a scalar valued function, in the
form of an integral over the complete domain, of the
displacement function.
Such a quantity is called as a Functional.
Variational Calculus is the calculus of Functionals (mainly
concerned with extremization problems).
General form of a functional, with only one independent
variable , one dependent variable and its first order derivative
w.r.t. the independent variable:
x1
I ( y ) F ( x, y, y ')dx,
x0
Lecture - 01
dy
where, y y ( x), y '
dx
du
Necessary condition:
dx
d 2u
Sufficient condition:
dx 2
0
x x0
0
x x0
If the second order derivative also is zero, higher order derivatives need to be considered.
Lecture - 01
u u ( x0 x ) u ( x0 ) 0 for x 0
(Provided u is differentiable at x0 )
du
dx
0
x x0
Sufficiency condition:
i.e.:
Lecture - 01
du
dx
du
dx
x x0 x
x
x x0
d 2u
2 0
dx
u u ( x, y ) in a specified domain.
Necessary conditions for a minimum at x = c, y = d:
u
x
0 &
x c , y d
u
y
0
x c , y d
u
x 2
2u
xy
u
xy
should be Positive Definite.
2
u
y 2
u K u 0
T
u
K u 0
T
if and only if u 0 .
Variational Calculus
The Beginnings Brachistochrone Problem:
The word Brachistochrone means shortest-time. The problem
was formulated by John Bernoulli in 1696.
The problem is to find the path corresponding to shortest-time
for a particle sliding from point (x0, y0) to (x1, y1) in the
vertical plane, under the action of gravity.
x
(x0, y0)
(x1, y1)
g, y
Ans.: Cycloidal Curve
Lecture - 02
x x
1
1
ds
ds
dt , Total time, t dt
v
v
x x0
x x0
ds 1 y ' dx & v 2 g y y0
2
t
Lecture - 02
1
2g
x1
x0
1 y '
dx
y y0
2
Lecture - 02
I ( y ) F ( x, y, y ')dx,
x0
B.C.: y ( x0 ) y0
&
dy
dx
y ( x1 ) y1
y ( x) u ( x) ( x), where,
( x0 ) ( x1 ) 0 & scalar quantity 0.
Variation in y, y
Lecture - 02
I ( ) F ( x, u , u ' ')dx,
x0
dI
d
0
d
d
F ( x, u , u ' ')dx 0
x0
F
F
' dx 0
u
u ' '
x0
x1
Lecture - 02
x1
F
F
x u u ' ' dx 0
0
The second term inside the integral can be evaluated using
integration by parts: x
x
x
F
F
x u ' ' dx u '
0
x0
d
dx
x0
1
F
dx
u '
We have: ( x0 ) ( x1 ) 0
x1
x1
F
d F
' dx
dx
u '
dx u '
x0
x0
Substituting:
Lecture - 02
x1
F d F
x u dx u ' dx 0
0
( x) ( x)dx 0,
x0
( x) 0 in x0 x x1.
Lecture - 02
F d F
F d F
x u dx u ' dx 0 u dx u' 0
0
This equation is called the Euler Equation or Euler-Lagrange
Equation in Variational Calculus.
For the stationarity of a functional, Euler equation must be
satisfied, along with the B.C.:
u ( x0 ) y0
& u ( x1 ) y1
I ( y ) 1 ( y ') 2 dx F 1 ( y ') 2
x0
, we have:
2
u
u '
1 u '
d
dx
u'
0
2
1 u '
Therefore, we get:
u'
1 u '
u ( x) is a straight line.
Lecture - 02
F d F
0, where, i 1, 2,..., n
ui dx ui '
Integrals with Higher Derivatives:
x1
I ( y ) F x, y , y ',..., y ( n ) dx
x0
Lecture - 03
w
w
I F x, y, w, wx , wy d , where, wx
& wy
x
y
F F
F
0
u x u x
y u y
Lecture - 03
I ( y ) F ( x, y, y ')dx,
where, y y ( x ), y '
x0
B.C.: y ( x0 ) y0
&
dy
dx
y ( x1 ) is not specified.
I ( ) F ( x, u , u ' ')dx
x0
Lecture - 03
dI
F
F
0
' dx 0
d
u
u '
x0
The second term inside the integral can be evaluated using
integration by parts:
x1
F
F
x u ' ' dx u '
0
x1
x0
x1
d F
dx
dx u '
x0
We have: ( x0 ) 0
x1
x1
F
F
d F
'
dx
dx
x u '
u ' x1 x0 dx u '
0
x
1
F d F
Substituting: F
dx 0
u ' x1 x0 u dx u '
Lecture - 03
F d F
0 &
u dx u '
F
0
u ' x1
The second requirement above is called the natural boundary
condition (This is same as the natural B.C., we have seen in
the W.R. formulation).
Lecture - 03
Area of C.S. = A
Youngs Modulus = E
x=L
1
2
2
EA
bAy
dx
20
2
0
0
Lecture - 03
F
bA &
u
d F
d
F d
u dx
d 2u
E 2 b 0
dx
F
EAu ' x L 0
u ' x L
This is same as the natural B.C. that the force acting at the
lower end of the rod is zero.
Lecture - 03
I ( y ) F x, y, y ' dx
a
d
d
d
dy
y
'
dx
dx
dx
dx
In other words, the derivative of variation with respect to an
independent variable is same as variation of the derivative.
Lecture - 04
F
F
F
y
y'
y
y '
I ( y ) F x, y , y ' dx F x, y, y ' dx
F
F
I
y
y' dx
y
y '
a
b
Lecture - 04
F d F
I
y dx Boundary terms
y dx y '
a
b
(since y '
d
y)
dx
I 0
Rayleigh-Ritz Method
If there is a functional corresponding to a problem, the
solution is the function, which makes the functional
stationary. If we have the functional:
I F ( x, y, , ', ",...)d
; $ am N m , with & N m 0; 1 m M on
m 1
am
0 for m 1, 2,..., M
Necessary condition: 0
Strain Energy (3D):
1
U x x y y z z xy xy yz yz zx zx d
2
W Xu Yv Zw d
Lecture - 04
t u t v t w d
x
Ne
e 1
e 1
e U e W e
The elemental contribution to the strain energy:
1
1
T
T
U d D d
2
2
e
e
e
1 e T
T
u B D B u e d
2
e
Elemental contribution to potential of loads:
W
e
u N X d u N t d
e T
Lecture - 04
e T
B D B d
T
&
f N X d N t d
T
K U f
Ne
K
e 1
Lecture - 04
K e
&
Ne
f f e
e 1
2
2
1
I k d Q d q d
2 x
y
Necessary condition: I 0
Using the finite element method:
Ne
I Ie
e 1
I
0
Lecture - 04
; j N j
j 1
Substitution gives:
n
N i
e
i
x
i 1 x
1
I
2 x
Lecture - 04
x
q
Q
d
d d
k
k
y
q
y
N j
n
N i x
1 N i
i
j d
n
j 1 2 e
y N j
x
i 1
q
Q
i d i d
e k
k
eq
N i
K
x
e
e
ij
&
Lecture - 04
N j
j
Ni N
y
y
x
Q
fi d
k
e
eq
e
q
d
k
Assignment No. 3
Exercise Problem No: 3.2, 3.5, 3.14, 6.2 & 6.3 in the book:
Finite Elements and Approximation by Zienkiewicz, O. C.,
and Morgan, K.
Due Date: As announced in the class.
Any suitable assumptions can be made, but clearly state the
assumptions and their justifications along with the answers.
Lecture - 04
Extremize: I F x, y, y ' dx
with:
x0
x1
K G x, y, y ' dx constant
x0
For 1 ( , ) to be stationary,
1 M r d M d 0
Since this should be true for any arbitrary variation, all the
terms in R.H.S. above must be zero. The first term gives:
0 in
Using the Fundamental Lemma of variational calculus, the
second term gives:
M r 0 on
And since the Lagrange Multiplier is an arbitrary parameter,
third term gives:
M 0 on
Lecture - 05
on .
1
T
(u ) U W d
2
Using FE formulation, the functional (P.E.) becomes a
function of nodal displacements:
1
T
T
(u ) u B D B u d
2
Re-writing:
1
T
(u ) u K u ,
2
where K B
D B d
u1 u1
Lecture - 05
1
(u ) u1 u2 L
2
K11
K
12
K 22
K1N
K2N
uN
K12
K1N
K 2 N
M
K NN
u1
u
2
u N
1 (u, )
Lecture - 05
1
u1 u2 L
2
K12
K 22
M
L
L
O
K1N
K2N
uN
K11
K
12
M
K1N
K 2 N
M
K NN
u1
u
2
u1 u1
M
u N
1 1
u1 u2 L
u1 2
K11
K
u N 12 K11
M
K1N
K12 L
K1N
Simplifying:
K11
Lecture - 05
K12 L
u1
u
2
K1N 0
M
u N
u1
u
2
0
M
u N
K11
K12 L
K1N
u1
u
2
1 M
0
u
N
1
u1 u1 0
1
Lecture - 05
0 L
u1
u
2
0 0 M
u1
u
N
K11
K
12
M
K1N
1
K12
K 22
M
K2 N
0
L
L
O
L
0
K1N
K2N
M
K NN
0
1
0
0
0
0
u1
u
2
M
u N
0
0
0
0
u 1
Lecture - 05
T
1 (u ) u K u u1 u1
2
2
where is the penalty parameter.
1
K11
u1
Lecture - 05
K12 L
u1
u
2
K1N u1 u1 0
M
u N
K11
K12 L
K11
K
12
K1N
K12
K 22
M
K2N
L
L
O
L
u1
u
2
K1N u1
M
u N
K1N
K 2 N
M
K NN
u1
u
2
u N
u 1
Lecture - 05
Concluding Remarks
The variational calculus gives an alternate method to arrive at
the finite element stiffness matrices, if a functional
corresponding to the problem could be found.
The scope of Galerkin WR method is clearly wider, but there
are some specific merits for variational formulation:
Variational formulation is physically more meaningful as
compared to WR statements (e.g.: minimization of P.E.)
Many methods used in FEA, which were known to work,
were given proper mathematical reasoning by the use of
variational calculus.
Other methods, like Least-square method, can also be used for
the finite element formulation. However, they are not widely
used like the Galerkin and Variational methods.
Lecture - 05