2.

Seminar Energy Methods, FEM Class of 2013

i
... internal potential energy stored in the body/system

e
... potential energy due to external loads, e.g., body forces, traction forces, etc.
problem denition: nd
i
dependent on the deformations and rotations in the elastic body
(simplication: static problem, elastic body)
dw
i
dg
g
F

i
=
_
V
w
i
dV (2)
w
i
... volume specic strain energy density
w
i
=
x
_
0

x
d
x
+
y
_
0

y
d
y
+
z
_
0

z
d
z
+
xy
_
0

xy
d
xy
+
yz
_
0

yz
d
yz
+
xz
_
0

xz
d
xz
(3)
dw
i
= : d
with =
_

x

xy

xz

yx

y

yz

zx

zy

z
_

_
and =
_

x
1
2

xy
1
2

xz
1
2

yx

y
1
2

yz
1
2

zx
1
2

zy

z
_

_
Solution steps
1. chose specic stress condition (e.g. plane stress, plane strain)
2. set stress-strain relationship (e.g. Hookes law)
3. set strain-displacement relationship (e.g. truss, beam, linear Bernoulli)
4. nd displacement eld over the body which leads to minimum potential of the total system
under consideration of boundary conditions
1
Minimisation (Extremum) Principal
= (
i
+
e
) = 0 (4)
variation: (.) =
(.)
u
u with displacement eld u
exact fullment of extremum conditions leads to Eulerian dierential equations strong
local form of equilibrium conditions
approximate fullment of extremum conditions leads to weak form (Ritz-method)
y
N
(x) =
0
(x) +
n

i=1
a
i

i
(x) (5)
y
N
(x) ... approximation/ansatz-function for displacement

0
(x) ... function for particular solution u = 0 at B (boundary)

i
(x) ... homogeneous solution function for u = 0 at B
a
i
... unknowns
requirements:
y
N
needs to full kinematic boundary conditions (displacement boundary conditions)
y
N
doesnt necessarily satisfy natural boundary conditions (traction boundary condi-
tions)
2 Example
Formulate the potential of internal energy depending on the nodal displacement components of
node 1 (linear-elastic material).
x,u
y,v
1
2 3
" "
I II
~
x,u
y,v x,u y,v
n
n
n
~ ~
~
E, A, I = const.
L = L = L
I II
Wanted:
i
= f
_
u(1) , v (1) , (1)
_
simplication: GA =
2
3 Solution
Wanted: internal potential energy for a planar Bernoulli-Beam in form of
i
= f
_
u(1) , v (1) , (1)
_
.
specic stress state (solution step 1)

x
= 0
y
=
z
= 0

xz
=
xy
= 0 (no torsion)

yz
= 0 (no shearing)
stress-strain relationship (solution step 2)

x
= E
x
(6)
internal potential for each beam (in local coordinates):

i
=
_
V
_
x
0

x
d
x
d V = E
_
V
_
x
0

x
d
x
d V =
1
2
E
_
V

2
x
d V (7)
strain-displacement relationship (solution step 3)
strains in linear regime (small strains):

x
=
d u
d x
(8)
kinematic hypothesis due to Bernoulli
y,v
n
u =u(y=0)
0
u(y)=u - y
0
nq
from u(y) = u
0
y and =
d v
d x
follows

x
=
d
d x
_
u(y = 0) y
_
=
d u
d x

d
2
v
d x
2
y = u

y (9)
inserting (9) in (7)

i
=
1
2
E
_
V
(u

y)
2
d V
=
1
2
E
_
V
_
u

2
2 u

y + (v

y)
2
_
d V (10)
with dV = dxdA and
_
A
dA = A and
_
A
ydA = 0 and
_
A
y
2
dA = I
z

i
=
1
2
EA
_
x
u

2
d x +
1
2
EI
z
_
x
v

2
d x (11)
3
ansatz-functions for displacement eld (solution step 4)
Hermite-polynoms
x
1
L
2
1
H H
2
2
x
1
H
4
3
4
1
H
L
x
H
4
4
4
2
H
L
H
2
1
= 1
x
L
H
4
1
= 1 3
_
x
L
_
2
+ 2
_
x
L
_
3
H
4
2
= L
_
x
L
2
_
x
L
_
2
+
_
x
L
_
3
_
H
2
2
=
x
L
H
4
3
= 3
_
x
L
_
2
2
_
x
L
_
3
H
4
4
= L
_

_
x
L
_
2
+
_
x
L
_
3
_
boundary conditions
node 2: u(2) = v (2) = (2) = 0
node 3: u(3) = v (3) = (3) = 0
ansatz functions for beams I and II (local coordinates)
u
I
=

H
2
1
u
I
(2) + H
2
2
u
I
(1)
u
II
= H
2
1
u
II
(1) +

H
2
2
u
II
(3)
v
I
=

H
4
1
v
I
(2) +

H
4
2
(2) + H
4
3
v
I
(1) + H
4
4
(1)
v
II
= H
4
1
v
II
(1) + H
4
2
(1) +

H
4
3
v
II
(3) +

H
4
4
(3) (12)
transformation local-global coordinate system
"
"
u
I
u
II
~
u
~
v
v
I
~
u
~
v
" "
v
II
u
I
(1) = cos u(1) sin v (1)
u
II
(1) = cos u(1) + sin v (1)
v
I
(1) = cos v (1) + sin u(1)
v
II
(1) = cos v (1) sin u(1)
ansatz functions for beams I and II (global coordinates)
u
I
= H
2
2

_
cos u(1) sin v (1)
_
u
II
= H
2
1

_
cos u(1) + sin v (1)
_
v
I
= H
4
3

_
cos v (1) + sin u(1)
_
+ H
4
4
(1)
v
II
= H
4
1

_
cos v (1) sin u(1)
_
+ H
4
2
(1) (13)
3 global unknowns u(1), v (1), (1) to be determined by MPE

i
= f ( u(1) , v (1) , (1))
total internal potential by sum of both beams and insertion of ansatz-functions

i
=
I
i
+
II
i
=
1
2
EA
_
x
u

I
2
d x +
1
2
EI
_
x
v

I
2
d x +
1
2
EA
_
x
u

II
2
d x +
1
2
EI
_
x
v

II
2
d x (14)
4
derivations of ansatz-functions
u

I
= H
2
2

_
cos u(1) sin v (1)
_
u

II
= H
2
1

_
cos u(1) + sin v (1)
_
v

I
= H
4
3

_
cos v (1) + sin u(1)
_
+ H
4
4

3
(1)
v

II
= H
4
1

_
cos v (1) sin u(1)
_
+ H
4
2

3
(1) (15)
with
H
2
1

=
1
L
H
2
2

=
1
L
H
4
1

=
6
L
2
+ 12
x
L
3
H
4
2

=
4
L
+ 6
x
L
2
H
4
3

=
6
L
2
12
x
L
3
H
4
4

=
2
L
+ 6
x
L
2
(16)
insertion into
i
and integration
5
Appendix: Derivation of an ansatz-function
i
k
u(i)
u(k)
v(k)
v(i)
x
n(i)
n(k)
Idea: Build up the ansatz-function by power series
with the nodal displacements as free parameters, i.e.
unknowns.
axial displacement u(x)
two unknowns u(i) and u(k) linear approach with two parameters a and b
u(x) = a x + b
boundary conditions:
u(x = 0) = u(i) b = u(i)
u(x = L) = u(k) a =
u(k) u(i)
L
lead to
u(x) =
u(k) u(i)
L
x + u(i)
=
_
1
x
L
_
. .
H
2
1
u(i) +
x
L
..
H
2
2
u(k)
vertical displacement v (x)
4 unknowns v (i), v (k), (i) and (k) cubic approach with 4 parameters
v (x) = a x
3
+ b x
2
+ c x + d
(x) = v

(x)
= 3a x
2
+ 2b x + c
boundary conditions
v (x = 0) = v (i) d = v (i)
(x = 0) = (i) c = (i)
v (x = L) = v (k) a L
3
+ b L
2
+ (i) L + v (i) = v (k)
(x = L) = (k) 3a L
2
+ 2b L + (i) = (k)
6
lead to
a =
(k) + (i)
L
2
2
v (k) v (i)
L
3
b =
(k) + 2(i)
L
+ 3
v (k) v (i)
L
2
nally leads to
v (x) =
_
(k)
L
2
+
(i)
L
2
2
v (k)
L
3
+ 2
v (i)
L
3
_
x
3
+
_

(k)
L
2
(i)
L
+ 3
v (k)
L
2
3
v (i)
L
2
_
x
2
+(i) x + v (i)
=
_
2
x
3
L
3
3
x
2
L
2
+ 1
_
v (i) +
_
2
x
3
L
3
+ 3
x
2
L
2
_
v (k)
+
_
x
3
L
2
2
x
2
L
+ x
_
(i) +
_
x
3
L
2
+
x
2
L
_
(k)
= H
4
1
v (i) + H
4
3
v (k) + H
4
2
(i) + H
4
4
(k)
7