Lecture 2

Variational Principles
in
Computational Solid
Mechanics
Somenath Mukherjee

Gangan Prathap

Scientist,
Structural Technologies Division,
National Aerospace Laboratories (NAL),
Bangalore,
Karnataka, India

Director,
National Institute of Science and
Information Resources (NISCAIR),
New Delhi,
India
1

Lecture 2
Variational Principles in Computational
Solid Mechanics
Chapters
1.

Principle of virtual work. Reciprocal Theorem.

2.

The Least Action Principle in dynamics.

(Hamiltons principle and the Euler-Lagrange equation)

3.

Weighted Residual (Galerkins) Method and the principle

of minimum potential energy.

4.

Element formulations using variational principles.

2

Lecture 2
Chapter 1
Principle of virtual work.
Reciprocal Theorem.

Archimedes was the first person to use the

concept of virtual work

1.1 Virtual work (over virtual displacements)

What is virtual displacement?
Virtual displacement is an imaginary displacement of any system
in frozen (fixed) time and space, over and above the actual
displacement.
In particular, virtual displacement can be imposed upon an
actual displacement at equilibrium.

What is virtual work?

Virtual work is the work done by the forces (external and internal)
traveling through virtual displacements.

1.2 Principle of Virtual work

For any arbitrary yet admissible set of virtual
displacements, superposed over the equilibrium
configuration, the net virtual work done by the
forces (external and internal) is zero.
Equilibrium

Zero virtual work

Wext Wint = 0

(1.1)

Archimedes was the first person to use the concept

of Virtual Work in his calculations for the Lever.
2

L1
1

L2

F1

F2
Equilibrium: Virtual Work Principle

F F =0
11 2 2
"Give me a place to
stand on, and I will
move the Earth."

Geometric Compatibility
Together

1= 2
L L
1 2

F1 L1= F2 L2
7

Example 1
Derive the equations of equilibrium for the bar system.
Suppose that actual orthogonal displacement components
of load point at equilibrium are u and v.

L2

L4

L3

L1

P
v

Actual strain (at equilibrium)

in the i-th bar is

u. cos i v. sin i
i =
+
Li
Li
Actual stress (at equilibrium) in the
i-th bar is

u. cos i v. sin i
i = Ei i = E i
+
Li
Li
8

Apply a virtual displacement

u along direction u.
L2

L4

L3

L1

By Principle of virtual work,

Wext = Wint

P
u.P cos =

i i .( Ai Li )

i =1
4

u. cos i v. sin i
= Ei
+
Li
Li
i =1

u. cos i
Li

( Ai Li )

Ei Ai
P cos =
u cos 2 i + v sin i cos i
i =1 Li

)
9

Apply a virtual displacement

v along direction v.
L2

L4

L3

L1

By Principle of virtual work,

Wext = Wint

v.P sin =

4
i =1
4

i i .( Ai Li )

u. cos i v. sin i
= Ei
+
Li
Li
i =1

v. sin i
Li

Ei Ai
P sin =
u sin i cos i + v sin 2 i
i =1 Li

( Ai Li )

)
10

L2

L4

L3

L1

u
v

The equation of equilibrium for the bar system is

4

Ei Ai
cos 2 i
i =1 Li
4
Ei Ai
(sin i . cos i )
i =1 Li

u
P cos
Ei Ai
(sin i . cos i )
i =1 Li
=
4
Ei Ai
sin 2 i
v
P sin
i =1 Li
4

Solving this equation gives displacements u and v, from which

11
the stresses in the bars can be determined.

1.3 Conservative Systems

In a conservative system, the net work done by
the forces in any closed path is zero.
In a conservative system, the net work from one
equilibrium state to another is path independent.
2

dW = 0

dW = dW
1

(1.2)

2
1
1

dW = 0

dW is path independent
1
12

1.4 Symmetry in Linear Elastic Conservative Systems

In a linear elastic conservative system, the stiffness
and flexibility matrices are symmetric.

kij = k ji

fij = f ji

(1.3)

Example 2
Demonstrate the invariance of work done in a bar with
changing order of loading.

1
l

2
P

Q
L
13

1
l

2
P

Case 1. First load point 1 by P and then load 2 by Q.

Work done by loads:

WPQ =

1
1
1
1 QL
Pl
Ql
P.u1,1 + P.u1, 2 + Q.u2,2 = P
+P
+ Q
2
2
EA
2 EA
2
EA

Case 2. First load point 2 by Q and then load 1 by P.

Work done by loads:

1
1
1 QL
1
Pl
Pl
+ P
WQP = Q.u2,2 + Q.u2,1 + P.u1,1 = Q
+Q
2
2
2
2 EA
EA
EA
WPQ = WQP
This demonstrates the invariance of work done with changing order of loading.
f12 =

Symmetry in flexibility

f 21 =

u1, 2
Q
u2,1
P

l
EA

l
EA

f12 = f 21
14

1.5 Maxwell Bettis Reciprocal Theorem

On a given structure (linear and conservative
system), the virtual work done by load system A
over virtual displacement for load system B is
equal to the virtual work done by load system B
over virtual displacement for load system A.

{ B } {FA } = { A } {FB }
T

Proof

(1.4)

{ B }T {FA } = { B }T [ K ]{ A } = { A }T [ K ]{ B }
= { A }T {FB }
because stiffness [K] is symmetric
(linear conservative system)

15

Example 3
Determine the support reactions of the portal
frame under the loading shown.
P
L/2

L/2

C
VA
MA

VB
A
HA

B
MB

HB
16

L/2

L/2

{2} {F1} = {1} {F2}

T

C
VA
HA

System 1
A

VB
B

MB

MA

yC,2

yA,2

y A, 2 .V A yC , 2 .P = 0
HB

VA =

P. yC , 2
y A, 2

System 2
17

L/2

L/2

{3} {F1} = {1} {F3}

T

C
VA
HA

System 1
A

VB

x A,3 .H A yC ,3 .P = 0

B
MB

MA
yC,3

HB

HA =

P. yC ,3
x A, 3

System 3
xA,3
18

L/2

L/2

{4} {F1} = {1} {F4}

T

C
VA
HA

System 1
A

VB

A, 4 .M A yC , 4 .P = 0

B
MB

MA
yC,4

HB

MA =

P. yC , 4

A, 4

System 4
A,4
19

L/2

L/2

yC,2

C
VA
HA

System 1
A

V A = VB =

B
MB

MA

xA,3

VB yA,2

System 2
P. yC , 2
y A, 2

HB

yC,3

yC,4

System 3

System 4

P. yC ,3
H A = H B =
x A, 3

M A = M B =

P. yC , 4

A, 4

A,4
20

1.6 Consequence of the Reciprocal Theorem

- The unit load method
1

BM2

BM1
M=Px

m=x

System 1

System 2

By Reciprocal Theorem:
But .P=(slope2).(BM1)

=
0

=
0

.1= .P
m( x )
dx M ( x)
EI
L

M ( x).m( x)
Px.x
PL3
dx =
dx =
EI
EI
3EI
0

(1.5)
21

Lecture 2
Variational Principles
in
Computational Solid
Mechanics
Somenath Mukherjee

Gangan Prathap

Scientist,
Structural Technologies Division,
National Aerospace Laboratories (NAL),
Bangalore,
Karnataka, India

Director,
National Institute of Science and
Information Resources (NISCAIR),
New Delhi,
India
1

Lecture 2
Chapter 2
The Least Action Principle in dynamics
(Hamiltons principle
and the Euler-Lagrange equation)

Nature always acts by the shortest paths.

. - Pierre de Fermat

The pioneers of Analytical Mechanics

&
Variational Calculus

Maupertuis
Fermat
Galileo

Lagrange

Newton

Hamilton
Euler

2.1 The Least Action Principle

Early History of Variational Calculus
Archimedes (287-212 BC)
For the first time in history the principle of virtual work (from virtual displacements)
was discovered and employed by Archimedes to derive the Lever Rule.
F1L1 = F2 L2
[F11 + F2 ( 2 ) = 0
1 / L1 = 2 / L2 ]
Pierre de Fermat (1601-1665 AD)
Discovered the Principle of Least time for optical path. Derived Snells Law of
Refraction (and reflection) of light rays using this principle.
B
dl
Nature always acts by the shortest paths.

=0
1 sin 1 = 2 sin 2
v
A
Pierre Louis Moreau de Maupertuis (1698-1759AD)
Defined a parameter called the Action: (Action=Path Integral of the momentum)
He postulated that Nature always behaves in a way so as to optimize (minimize) this
B
B
action.
(Euler presented the mathematical proof of this postulate.) I = mv.dl ; mv.dl = 0
A

Johann Bernoulli (1667-1748)

Presented the Brachistochrone problem (finding the path of fastest descent between
two fixed points) in Acta Eruditorum and its solution is 1682. Solutions were provided
independently by
Isaac Newton, Jacob Bernoulli (elder brother of Johann) and L Hospital.
5

Archimedes Lever
Equilibrium of any system under given forces
conforms to the vanishing of the Net Virtual work
done by these forces traveling over the
geometrically compatible virtual displacements:
F1L1 = F2 L2

[F11 + F2 ( 2 ) = 0

L1

1 / L1 = 2 / L2 ]

L2
F2

F1
Fermats Optical Path
From one fixed point to another, Light travels
along the path of least time joining these points.
T AB

Length
=
=
Speed

a 2 + (d x) 2
b2 + x2
+
v1
v2

dT AB
(x d )
x
+
=
dx
v1 a 2 + ( d x) 2 v 2 b 2 + x 2

Medium 1

For the optimized path of least time

dT AB
=0
dx

(d x)
v1 a + (d x)
2

sin 1 sin 2
=
v1
v2

c =speed of light in vacuum

v =speed of light in medium

x
v2 b + x
2

Medium 2

sin 1
sin 2
=c
v1
v2

1 sin 1 = 2 sin 2
sin = Consant

d
=

c
v

b
2

B
x
6

The Brachistochrone
The Path of Fastest Descent under gravity from one point to another.
Given two fixed points A and B.
What is the path for shortest time of fall from A to B?

2 R

2R
v= (2gy)

y
For the path for shortest time of fall from A to B
B

dl
=0
v
A

sin
= k (Consant )
v
Speed : v = 2 gy

x=R(
-sin )
y=R(1-cos
)
The Brachistochrone is a Cycloid
between points A and B.
Radius of generating wheel R=0.25/(k*k*g)
7

Hamiltons Principle and the Euler-Lagrange Equations

Hamiltons Principle of Least Action:
Nature determines the path of any particle from point 1 to point 2 in a way
so as to minimize the Action I
L=Lagrangian, T=Kinetic Energy, V=Potential Energy
2

I = L(q,q, t )dt

L = T V

q = q (t )

( for

virtual changes

about any

path) = I + 2 I + ..

For the Action to be a minimum, any

virtual changes about the minimum point
should lead to a net increase in the Action,
i.e.
I

( for

virtual changes

I > 0

I = 0

about

min) 0 .

2I > 0

i.e. The first variation of the Action ( I)

should vanish for small perturbations
(virtual displacements) about the minimum point.

First variation for the Lagrangian L (for virtual displacement in frozen time is)

L =

n
i =1

L
L
qi + qi
q
qi

subject to

qi (t = t1 ) = qi (t = t 2 ) = 0

I = L(qi ,qi , t ) dt

Action

L = T V

qi = qi (t )

: generalised

co ordinates

n 2

L
L
First variation for the Action for any I = L(q ,q , t )dt =
+

q
qi dt
i i
i
q
path 1-2 is
i =1 1 qi
1
For the critical path, I=0
0 =

i =1 1

0 =

i =1 1
2

qi

L
qi

n
L
qi dt +
qi
i =1

L
qi
q
n

2
1

d L
qi dt

dt
q

L
L d L
qi dt +
qi

qi dt qi
qi
i =1
=0

Path A

Critical Path

1
Path B

qi (t = t1 ) = qi (t = t2 ) = 0

t1

t2

Euler-Lagrange Equation (Conservative Systems)

d L
L

=0
dt q i qi
9

Euler-Lagrange Equation (Non-Conservative Systems)

d L
L

= Qi
dt q i qi
Qi = Non conservative

generalised

force

What is a generalized force?

Virtual work=(Virtual physical displacement)

=(Virtual generalized displacement)

Physical force

Generalized force

= (qi ).Qi

10

2.2 Applications of the Least Action Principle

Free Vibration of Single Degree of freedom systems
The Pendulum
Kinetic Energy: T = mx 2 = m(l )
1
2

1
2

Potential Energy:

Euler-Lagrange Equation (Conservative Systems)

Equation of Motion

ml 2 + mgl. sin = 0

V = mgl (1 cos )

d L L

== 0
dt

L = T V

g
l

+ . sin = 0

Linear Analysis (small oscillations)

g
+ =0
l

= A. sin(nt + )

Natural frequency and Time Period

n =

g
(rad / sec)
l

fn =

1
2

g
( Hz )
l

1
l
= 2
(sec)
fn
g
11

2.3 Some simple applications of the Least Action Principle

Free Vibration of Single Degree of freedom systems
Example 1 The Spring-Mass model
Kinetic Energy:

1
T = mx 2
2

Euler-Lagrange Equation (Conservative Systems)

L d L

=0
x dt x

mx + kx = 0

Equation of Motion (free vibration)

Solution (free vibration)

V=

Potential Energy:

1 2
kx
2

L = T V

k
m

x = A. sin(nt + )

x(t)

Natural frequency and Time Period

n =

k
(rad / sec)
m

fn =

1
2

k
( Hz )
m

1
m
= 2
(sec)
fn
k
12

Free Vibration of Multi-degree of freedom systems

Example 2 The Spring-Mass model for a 2 DOF system
1
1
2
m
x
+
m2 x2 2
Kinetic Energy:
1 1
2
2
1
1
Potential Energy: V = k1 x12 + k 2( x2 x1 ) 2
2
2
T=

d L L

=0
dt x1 x1

L = T V

Euler-Lagrange Equation (Conservative Systems) d L L = 0

dt x2 x2
Equation of Motion (free vibration)

Solution

m1 0
0 m2

m1

x1

m2

x2

k1 + k 2

k2

x1

k2

k2

x2

k1

0
=
0

m1

x1
A1
=
. sin(nt + )
x2
A2
A1
k1 + k 2
+
k2
A2

Eigenvalues and Eigen-modes

k2
k2
12 :

m2

x2(t)

x1(t)

k1 + k 2
det
k2

A1
0
=
A2
0

{1} =

k2

A1,1
A2,1

2 2 :

0
k2
2 m1
=0
n
0 m2
k2

{2 } =

A1, 2
A2, 2

13

The Spring-Mass model for a 2 DOF system (Continued)

Orthogonality of the Eigen-modes (Normal modes)
T
{1} [ K ]{2 } = {2 }T [ K ]{1} = 0
{1}T [ K ]{1} = k11 {2 }T [ K ]{2 } = k22

{1}T [ M ]{2 } = {2 }T [ M ]{1} = 0

{1}T [ M ]{1} = m11

{2 }T [ M ]{2 } = m22

Natural Frequencies (rad/sec)

k
1 = 11
m11

kii : generalized modal stiffness for mode i

mii:: generalized modal mass for mode i

k 22
2 =
m22

Estimation of dynamic response for

Modal superposition

x1

[M ]

x1
x2

+ [K ]

x1
x2

F1 (t )

F2 (t )

= q1 (t ){1}+ q2 (t ){2 } = [{1}, {2 }]

x2

[]

q1
q2

= [ ]{q}

Pre-multiply equation of motion by modal matrix

x1
x1
q1
q1
Q1 (t )
T
T
T F1 (t )
T
T
[ ] [M ]
+ [ ] [K ]
= [ ]
[ ] [M ][ ]
+ [ ] [K ][ ]
=
x2
x2
F2 (t )
q2
q2
Q2 (t )

m11
0
0 m22

q1
k11 0
+
q2
0 k 22
kii
/\/\/\/\/\

mii

q1
Q1 (t )
=
q2
Q2 (t )
qi(t)
Qi(t)

mii qi + kii qi = Qi (t )
Qi (t ) = {i }T

F1 (t )
F2 (t )
14

Modal Decoupling:

The Spring-Mass model: equal stiffnesses and masses

Equation of Motion (free vibration):
Solution:

m 0
0 m

x1
A1
=
. sin(nt + )
x2
A2
Characteristic Equation:

det

x1
x2

2k

2k
k

k
k

x1
x2

0
0

=0

Natural frequencies (rad/sec) and Eigen-modes

k

1 = (0.618)

k
:
m

{1 } =

1
1.618

2 = (1.618)

k
:
m

{ 2 } =

1
0.618

m
x1(t)

x2(t)

Observation: The natural frequencies and mode shapes for this particular
system are related to the Golden Numbers :
(1+ 5)/2=1.618 and (1- 5)/2=-0.618

(!!!)

15

Stationary Property of the Rayleigh Quotient at the Normal Modes of Vibration

The Spring-Mass model: equal stiffnesses and masses
Solution:

x1
x2

X1
X2

. sin(t )

x1
X
= 1 . cos(t )
x2
X2

Velocity Vector :

Conservation of Energy: KE max = PE max

1
1
1
1
m 2 X 12 + m 2 X 2 2 = kX 1 2 + k ( X 2 X 1 ) 2
2
2
2
2

k
2 =
m
Rayleigh Quotient
k

X2
X1

+1

+1

k
m

m
x1(t)
d 2
=0
X2
d
X1

X2
1
X1

X2
X1

x2(t)
2

X2
1 = 0
X1

X2
= 1.618, 0.618
X1

Observation: The Rayleigh Quotient is extremum (stationary) at the normal modes.

16

Example 3 A rigid car chassis as a 2 DOF system

a0

b0

CM
1

k1

hCM

k2

Case 1: Point of reference for displacements : Center of mass, CM

Expressions for kinetic energy T, potential energy V and the Lagrangian L are

1
1
mhCM 2 + I CM 2
2
2
1
1
1
1
V = k112 + k 2 2 2 = k1{a0 hCM }2 + k 2 {b0 + hCM }2
2
2
2
2
1
1
1
1
L = T - V = mhCM 2 + I CM 2 k1{a0 hCM }2 k 2 {b0 + hCM }2
2
2
2
2
T=

17

Example 3 A rigid car chassis as a 2 DOF system (Continued)

a0

b0

CM
1

hCM

k1

k2

Case 1: Point of reference for displacements : Center of mass, CM

Lagranges Equations

d L L

=0
dt h h

d L L

=0
dt

Equations of free vibration in matrix form

hCM

I CM

(k1a0 k 2b0 ) hCM

k1 + k 2
0
=
(k1a0 k 2b0 ) k1a0 2 + k 2b0 2
0

Note:
For displacements at the center of mass, the system equations are inertially decoupled.
But the system equations are in general stiffness coupled.

b
k
If however, the spring stiffness values be in the ratio 1 = 0
k 2 a0
then the system is also stiffness decoupled.

k1a0 k 2 b0 = 0
18

Example 3 A rigid car chassis as a 2 DOF system (Continued)

as

S CM

k1

bs

hs

k2

Case 2: A special point of reference S for displacements so that

bs k1
=
as k 2

k1as k 2bs = 0

(k1 )as (k 2 )bs = 0

This means that under an external force F acting the point S, one can have
pure translation (no rotation) of the rigid chassis without any rotation
(equal and opposite moments are induced by the springs about point S).
as

k2

k1

k1

bs

S CM

k2

19

Example 3 A rigid car chassis as a 2 DOF system (Continued)

as

bs

S CM

k1

hs

k2

Case 2: A special point of moment balance S for displacements

T=

1
1
1
mhCM 2 + I CM 2 = m hS + e
2
2
2

)2 + 12 I

1
2

1
2

2
mhs 2 + mehs + I S 2
CM =

I s = I CM + me 2
1
1
1
1
k112 + k 2 2 2 = k1{as hs }2 + k2 {bs + hs }2
2
2
2
2
1
1
2
2 1
2 1
L = T - V = mhs + mehs + I s k1{as hs } k 2 {bs + hs }2
2
2
2
2

V=

20

Example 3 A rigid car chassis as a 2 DOF system (Continued)

as

bs

S CM

k1

hs

k2

Case 2: A special point of moment balance S for displacements

Lagranges Equations

d L L

=0
dt h h

Apply the condition

&

bs k1
=
as k 2

d L L

=0
dt

k1as k 2bs = 0

Equations of free vibration in matrix form

m me
me I CM

k1 + k2
+
0

hs

0
k1as 2 + k 2bs 2

0
=
0

hs

Note:
For displacements at the special point S for balance of moments from spring forces
the system equations are stiffness decoupled.
But the system equations are in general inertially coupled.
If CM and S coincide then system is both stiffness wise and inertially decoupled. 21

2.4 Constrained Systems; Lagranges Multipliers

A system of n degrees of freedom (DOFs) is holonomic if all its DOFs are mutually independent.
A system is non-holonomic when some of its n degrees of freedom (DOFs) are dependent on the
others. This means that it has some r constraints so that the actual number of its independent
degrees of freedom is (n-r).
We need to find the stationary points of the functional

t2

I = L(qi , qi , t )dt

Subject to r constraints

i = 1,2,...n

t1
t2

f j (qi ) = constant

f j (qi , qi , t )dt = constant

or

j = 1,2...r

t1

Procedure:
Modify the Lagrangian

L* = L +

j f j

j =1

Equation of motion can be had from

t2

I = 0 where

I = L * ( q i , q i , t ) dt
t1

Modified Lagranges Equations:

d L
L

+
dt qi
qi

r
j =1

f j
qi

=0

or

= Qi : Non conservative

force
22

Example 4 The Atwood Machine

Find the expressions for accelerations of the two unequal masses suspended
from the machine by an in-extensible string.
T

Method 1: Using Newtons Law

m1a = T m1g .............(i )

Light pulley
m1

m2 a = m2 g T ...........(ii )
(i ) + (ii )

m2

y
m1
m2

m1g

(m1 + m2 )a = (m2 m1 ) g
a=

m2g

( m2 m1 )
g
( m2 + m1 )

x = a

m2 m1

y=a

Method 2: Using Lagranges Equations

1
1
1
1
1
m1x 2 + m2 y 2 = m1x 2 + m2 ( x 2 ) = (m1 + m2 ) x 2
2
2
2
2
2
V = m1gx m2 gy = m1 gx m2 g { x (R + )} = (m1 m2 ).gx + (constant)1
KE =

x + y + R = = fixed
length of
y = x (R + ) = x constant
and

y = x and

y = x

string

L = KE - V
1
(m1 + m2 ) x 2 [ (m1 m2 ).gx + (constant)1 ]
2
d L L

= 0 (m1 + m2 ) x (m1 m2 ) g = 0
dt x x
(m m2 )
(m m1 )
x= 1
g
y = x = 2
g
(m1 + m2 )
(m1 + m2 )

L=

23

Example 4 The Atwood Machine (continued)

Light pulley

Method 3: Using Lagranges Equations with Lagranges

Multipliers
Constraint Equation

x + y + R = = fixed

length of

f ( x, y ) = x + y = constant

string

y = x

Lagrangian L
1
1
m2
KE = m1x 2 + m2 y 2
V = m1gx m2 gy
2
2
One equation from
constraint
1
2 1
2
L = KE - V = m1x + m2 y + (m1gx + m2 gy )
2
2
Lagranges Equations
m1

f
d L L
=0

+
dt x x
dx

m1 x m1 g + .1 = 0

f
d L L

+
=0
y
dt y y

m2 y m2 g + .1 = 0

(m m )

2
Solving
x= 1
g
(m1 + m2 )
all three
equations: = m1 ( g x) = T

y = x =

(m2 m1 )
g
(m1 + m2 )

Two more equations

from Lagrange
Equations

The Lagrange Multiplier is the

constraint force,
24
viz. tension T is string

Example 5 A simple constrained minimal problem

y
y=2x

Minimize the function

z (x, y) = x2 + y2 - 2xy - 4x
Subject to the constraint

-4

y 2x = 0
Using Direct
Substitution

Using Lagranges
Multiplier

z1 = x 2 + y 2 2 xy 4 x + ( y 2 x)

y = 2x

z = x 2 + y 2 2 xy 4 x = x 2 4 x At stationary
z
= 2x 4
x
At stationary
z
=0
x

point

( xm , ym ) = (2,4)

zmin = xm 2 4 xm = 4

Using Penalty Method

(This type is used to
describe many FE
formulations)
z 2 = x 2 + y 2 2 xy 4 x + ( y 2 x) 2

point (minima ) At stationary

z1
= 2 x 2 y 4 2 = 0
x
z1
= 2 x + 2 y + = 0
y
z1
= 2 x + y = 0

Solving ;
xm = 2 , ym = 4 ( z1 ) m = 4

point

(minima )

z 2
= 2 x 2 y 4 4 ( y 2 x) = 0
x
z 2
= 2 x + 2 y + 2 ( y 2 x) = 0
y
Solving ;
xm = 2 1 +

ym = 2 2 +

As ,
x m 2,
y m 4 ( z 2 ) m 4

25

Lecture 2
Variational Principles
in
Computational Solid
Mechanics
Somenath Mukherjee

Gangan Prathap

Scientist,
Structural Technologies Division,
National Aerospace Laboratories (NAL),
Bangalore,
Karnataka, India

Director,
National Institute of Science and
Information Resources (NISCAIR),
New Delhi,
India
1

Lecture 2
Chapter 3

Weighted Residual (Galerkins) method and

the principle of minimum potential energy.

Lord Rayleigh

James Clerk Maxwell

3.1 The self adjoint operator

Consider a differential equation:

(u) = f
Force (kinetic)
f Specified
Boundary

Displacement (kinematic)
Specified Boundary

Use a trial function v as virtual displacement

v. (u)..d = u .

v. (u)..d = u .

If

(v).d +

Domain

{d (v) f (u ) d (u ) f * (v)}d

=d + f

(v).d

is self adjoint operator

(3.1)

This is the consequence of the Reciprocal Theorem (valid for conservative

systems only).
A conservative system is characterized by a differential equation of self- adjoint
operator. A conservative system has a symmetric stiffness matrix
4 of
(or compliance matrix). The energy of a conservative system is independent
changing order of loads.

Example 1. Show that the linear differential operator for the axially
loaded bar is self-adjoint.
x,u(x)

d
du

EA
=q
dx
dx
L

d
du
v. (u)..d = v
EA
dx
dx
0

du
= v EA
dx
L

x=L
x =0

d
dv
= u
EA
dx
dx
0

du
dx = v EA
dx

du
dv
du
EA
dx = v EA
+
dx
dx
dx
0
du
dx + v EA
dx

v. (u)..d = u .

d
d
=
EA
dx
dx

d
du
v
EA
dx
dx
0

x=L
x =0

q
x=L

+
x =0
x= L

dv
du
EA
dx
dx
dx
0

+
x =0

dv
u EA
dx

dv
EA
.u
dx

x=L

+ u
x =0

d
dv
EA
dx
dx

dx

x=L

x =0

(v).d + {vf (u ) uf * (v)}=d + f

L

d
dv
dx = u
EA
dx
dx
0

dx

Reciprocal Theorem
5

3.2 Weighted Residual Method

for a system of differential equation of equilibrium with
self-adjoint operator
u h .f .d
(u) = f
Consider a differential equation:

Use an admissible trial function uh as virtual displacement, and

u is the analytical solution

u .{ (u) - f }.d = a(u , u ) (u , f ) u .R B

h

B is the domain boundary. Boundary terms vanish when either

uh =0 (specified kinematic boundary condition) or reaction R=0
at boundary B.

0 = a(u h , u ) (u h , f )
a(u , u ) = (u , f )
h

(3.2)

For a self-adjoint operator, a(u,v) is symmetric bilinear functional

u h . (u)..d = u .

( u h ).d

a(u h , u ) = a(u, u h )

(3.3)
6

Case 1
If the trial function is a variation of the exact displacement

uh = u
a (u , u ) = (u , f )

Here

1
a (u , u ) = W
2

1
a (u, u ) is the analytical strain energy U of the system
2

and W is the virtual work done by the external forces.

If we define a system potential energy (for a conservative system)
1
= a(u, u ) (u, f ) = U W
2
(3.4)
Then equilibrium condition :
=0

Principle of stationary potential energy

Case 2
If the exact function (of differential equation) is replaced by
N
N
the trial function
h
h
u ( x ) = aii ( x )
u ( x) = (ai ).i ( x)
i =1

Thus
Weighted residual statement is
h

u . (u ) - f .d = 0

ai j .

i =1

(u h ) - f = e
N

aii ( x) - f .d = 0

j = 1,2...N

i =1

(3.5)
Galerkins Method
Rayleigh-Ritz Method

a(u h , u h ) = (u h , f )

e=residual

1
a(u h , u h ) = W
2

If we define a system potential energy (for a conservative system)

1
= a(u h , u h ) (u h , f ) = U W
2

Then equilibrium condition :

=0

(3.6)
8

Rayleigh Ritz Method (RRM) or Galerkins Method (GM);

Which should we use?
For conservative systems, the equations of equilibrium (with any
given approximate function for displacement) determined by both
Rayleigh- Ritz and Galerkins Methods are identical.
For non-conservative systems, we cannot adopt the Rayleigh-Ritz Method
(RRM), since the reciprocal theorem (symmetry; equation (3.1)) does not
hold, and consequently we cannot define the potential energy functional
for the system.
Such non-conservative systems are described by differential equations with
Non-self-adjoint operators.
However, the virtual work principle remains valid even for non-conservative
systems. Thus Galerkins method (GM) of weighted residuals can be applied
even to non-conservative systems to establish the finite element equations
(force-displacement relation) of equilibrium from the corresponding
differential equations of non-self-adjoint differential operators .

Message: Use RRM or GM for conservative systems

Use only GM for non-conservative systems

Example 2. Weighted residual statements for the axially loaded bar.

x,u
L
d
du
d
d

EA
=q
=
EA
q
dx
dx
dx
dx
L

u .{ (u) - q}.d = u
h

0 = u

du
EA
dx

0
x=L
x =0

0 = a (u h , u ) (u h , q )

du h
du
a (u , u ) =
EA
dx
dx
dx
0

du
d
EA
q dx

dx
dx

L
du h
du
EA
dx u h q.dx 1 a (u , u ) = 1 du EA du dx
+
dx
dx
0
0
2
2 0 dx
dx
L

1
1 du h
du h
h h
a (u , u ) =
EA
dx
2
2 0 dx
dx

Weighted residual statement is

ai j .

aii ( x) - q .d = 0

j = 1,2...N

Galerkins Method

i =1

Rayleigh-Ritz Method

=0
L

1
1 du h
du h
h h
h
EA
dx u h q.dx
= a (u , u ) (u , q) =
2
2 0 dx
dx
0

= U W
10

3.3 The principle of minimum potential energy and

equations of equilibrium
For a conservative (self-adjoint) system, we can identify
a potential energy functional , so that at equilibrium,
any admissible virtual displacement leads to

=0
=0

Virtual work

Equations of Equilibrium

11

Incompatible elements
in harmony
(equilibrium)

Compatible elements
not in harmony
12

The principle of minimum potential energy

(for searching equilibrium)
Of all the possible admissible displacements (satisfying
the kinematic boundary conditions) the one that
corresponds to equilibrium (of forces) is the one that makes
the potential energy stationary.
(i.e. any virtual displacement about equilibrium (of forces)
brings about a vanishing first variation of potential
Energy)

=0 for u

(3.7)

A stable equilibrium makes the potential energy a minimum

13

The principle of stationary complementary energy

(for searching compatibility)
Of all the possible admissible forces (in equilibrium) the
one that corresponds to displacement compatibility is the
one that makes the complementary energy stationary.
(i.e. any virtual change of forces in equilibrium brings about
a vanishing first variation of complementary energy).

*=0 for F

(3.8)

Complementary energy and its variation is

*=U*-W

*= U*- W(for F)
14

3.4 Castiglianos Theorems

The principle of minimum potential energy
(for searching equilibrium)
= U

Piui = 0

i=

U
Pi =
ui

(3.9)

The principle of stationary complementary energy

(for searching compatibility)
Stress
* = U *
ui =

uiPi = 0

U *
Pi

For a linear structure,

U*
(3.10)

U* = U

U
Strain

Complementary energy/volume equals strain energy/volume

15

Example 3. Derive the equations of equilibrium and boundary

conditions for the axially loaded bar with spring supported ends.
x,u

/\/\/\
k0

L
q

x=0
Potential energy functional
=U-W
L
L
1 du
1
du
2 1
2
=
EA
dx + k0u0 + k Lu L uq.dx
2 0 dx
2
2
dx
0
L

/\/\/\
kL
x=L

du
du
.EA
dx + + k0u0 .u0 + k Lu L .u L u.q.dx
=
dx
dx
0
0
du
.u
= EA
dx

x= L
x =0

du
EA + k0u0
dx

d
du
+ k0u0 .u0 + k Lu L .u L u
EA
dx
dx
0
du
u0 + EA + k Lu L
dx
x =0

u.q.dx

L
x= L

u L + u
0

d
du
q dx
EA
dx
dx
16

=0

Equilibrium
0=

du
EA + k0u0
dx

du
u0 + EA + k Lu L
dx
x =0

L
x= L

u L + u
0

d
du
q dx
EA
dx
dx

For non-trivial arbitrary, yet admissible virtual displacements

d
du
q =0

EA
dx
dx
Either u 0 = 0
Either u L = 0

or
or

For specified kinematic

boundary conditions

EA
EA

Differential equation for axially

loaded bar

(3.11)

du
+ k 0u0
dx

du
+ k LuL
dx

=0
x =0

=0

Boundary conditions
at the ends x=0 & x=L

(3.12)

x=L

Force boundary conditions with spring ends

Force boundary conditions with free ends

du
EA
dx

= 0,
x =0

EA

du
dx

=0
x17
=L

Example 4. Derive the equations of equilibrium and boundary

conditions for the Euler beam with spring supported ends.

r0

rL

x=0

k0

kL

Potential energy functional

=U-W

d 2w
dw
1 d 2w
1
2 1
2 1
=
+
+
+
EI
dx
k
w
k
w
r
0 0
0
L L
dx
2 0 dx 2
2
2
2
dx 2
L

=
0

d 2w
dx

.EI

d 2w
dx

x=L

dw
1
+ rL
dx
x =0 2

wq.dx
x= L

dx wq.dx
0

+ k0 w0 .w0 + k L wL .wL + r0

dw
dx

x =0

dw
dx

x =0

+ rL

dw
dx

x=L

dw
dx

x= L
18

= w.
0

d2
dx

EI

d 2w
dx

q dx

d 2w

dw
EI 2 + r0
dx
dx

x =0

d
d 2w
EI 2 + k0 w
dx
dx

dw

dx

+ EI
x =0

d 2w
dx 2

+ rL

dw
dx

x= L

d
d 2w
wx =0 + EI 2 + k L w
dx
dx
x =0

dw
dx

x =0

wx = L
x= L

For non-trivial, arbitrary, yet admissible virtual displacements,

Equilibrium

dx 2

=0.

dw
Either
dx

=0

or

EI

dx

x =0

EI

d w
dx 2

q =0

Differential equation
for the Euler beam

(3.13)

dw
Either
dx

=0

Either w L = 0

or

EI

d 2w
dx

x=L

Either w0 = 0

d 2w

or

or

dw
dx

+ r0

+ rL

d
d 2w
EI
+ k0 w
2
dx
dx

dw
dx

=0
x =0

=0
x=L

=0
x =0

d
d 2w

+ kLw
EI
2
dx
dx

=0
x=L

Boundary conditions at the ends x=0 & x=L

(3.14)

19

3.5 Constrained media problems

(use of penalty methods)

Elementary beam theory as constrained media problem

Principal assumption in Eulers simple beam theory
(valid for thin beams):
Plane sections normal to the neutral axis before bending
remain normal to it even after bending.
This means that
(a) bending rotation is equal to the slope of the neutral axis.
(b) shear deformation zx is (assumed) zero.

Euler beam model

Shear flexible beam model

x
dw/dx
dw/dx

z,w
=dw/dx

NA

zx=

- dw/dx

20

Elementary beam theory as constrained media problem (continued)

Euler beam model

1
d
=
EI
2
dx
0

z,w
=dw/dx

NA

The Euler beam has

infinite shear rigidity
But the practice of using a
large shear rigidity for
thin beams creates a
problem called Shear
Locking in shear-flexible
beam elements.

1
dw
dx qwdx +

2
dx
0
0

=0

Equilibrium Equations
EI

dw/dx

d 2
dx

dw
=0
dx

d d 2 w
=q

dx dx 2
Combining

i.e.

....(i )
d
dw
=q

dx
dx

....(ii )

(i ) & (ii )

d 2
d
...(iii )
EI 2 q = 0
dx
dx
Boundary conditions at x = 0 &

x=L

d
= 0 or = 0
dx
dw
Either
= 0 or w = 0

dx
dw
As
dx
Either

EI

Equation (iii ) reduces to

EI

d 4w
dx

q 0 21

Lecture 2
Variational Principles
in
Computational Solid
Mechanics
Somenath Mukherjee

Gangan Prathap

Scientist,
Structural Technologies Division,
National Aerospace Laboratories (NAL),
Bangalore,
Karnataka, India

Director,
National Institute of Science and
Information Resources (NISCAIR),
New Delhi,
India
1

Lecture 2
Chapter 4
Element formulations
using variational principles

4.1 Applications of the Rayleigh-Ritz method

x

q
L

EI

Example 1. Determine the deflection curves and bending curvature

distributions of a simply supported uniform beam under u.d.l. q using
the Rayleigh-Ritz method.
Use the following admissible displacement functions for approximations,
satisfying the essential (kinematic) boundary conditions:

w(x=0)=0 & w(x=L)=0

(i ) w( x) = a.x( L x)
x
(ii ) w( x) = a. sin
L

q
Le=L/2

Le=L/2

qL2
8 EI

qL
12 EI

(i ) w( x) = a.x( L x)
L

d 2w
dx

= 2a

1
d w
qL3
2
EI
dx q.w.dx = 2a ( EIL) a
=
2
6
20
dx
0
2

qL3
=0
Equilibrium : = 4a (a).( EIL) (a).
6
2

a=

qL
24 EI

w=

qL
x( L x)
24 EI

Exact Curvature
Approximate
Curvature=2a
Areas under the
curvature curves are
Equal. WHY?
Aexact

2 L qL2
qL3
=

=
3 8EI 12 EI

Aapprox

qL2
qL3
= L
=
12 EI 12 EI

Aexact = Aapprox4

q
Le=L/2

Le=L/2

qL2
8 EI

Exact Curvature

(ii ) w( x) = a. sin
L

L
2

d 2w
dx

=a

sin

Areas under the

curvature curves

x
L

d w
qL
1
2

EI
dx
q
w
dx
a
EIL
a
(
)
2
.
.
4

20
dx 2
L
4
0

Equilibrium : = 0
a=

4qL4
5

EI

2a(a).
w=

4qL4
5

EI

sin

4
4

4L

( EIL) ( 2a)

Aexact

qL

=0

qL3
2 L qL2
=

=
3 8 EI 12 EI

Aapprox =
Aapprox

4qL4

5 EI

2L

sin
0

qL3
=
12 EI

Aexact = Aapprox

x
L

dx

= EI
0

EI

h 2

(superscript

approximated function)
wh ( x) = a.x( L x)

h = 2a

wh ( x) = a. sin(x / L)

h = a.( / L) 2 sin(x / L)

dx

dx

2
2

= EI

d w
dx

dx

Deflection
Error of the Energy of
at beam
the Error
Energy
center x=L/2
2
h 2
h 2

signifies

Exact

d w

L
0

Displacement function
and curvature (strain)

q
w( x) =
x( L3 2 x 2 L + x 3 )
24 EI
q
=
x( L x)
2 EI

qL4
(0.01041)
EI

q 2 L5
(1440) EI

qL4
(0.01307)
EI

1
4 q 2 L5

240 6 EI

qL4
(0.01302)
EI

q 2 L5
(1440) EI
1
4 q 2 L5

240 6 EI

0
6

4.2 Element formulations using the principle of

minimum potential energy
4.2.1 The bar element (2 nodes, 2 DOF)
e
L
R1
R2
1
x
u1,F1
u2,F2

=-1

=1

EAe

Displacement function
(C0 continuity)
h

u =

N i ( ).ui =[N1

2x
Le

i =1

N 2 ( ) =

1
N1 ( ) =
2

[N]: Shape function matrix

Strain

=
h

u 2 u1
e

Le

Le

1
2

1 1

u1
= [ N ]{ e }
u2

N2 ]

N1 ( ) =

u1
= [ B ]{ e }
u2

[B]: Strain-displacement matrix

1+
2

2
1+
N 2 ( ) =
2

(4.1)
(4.2)
7

R1

The bar element

Le

R2

u1,F1

=-1

Stress

h = EAe h = [ D][ B]{ e }

u2,F2

EAe

=1

[ D] = EAe

Strain Energy of element

x= L

{ } { }dx ={

1
h
Ue =
2 x =0

=1

e T

1
1
L
[ B]T [ D][ B] d { e } = { e }T [ K e ]{ e }
2
2 =1
2

=1

e 1
1
L
EA
e
T
[K ] =
[ B] [ D][ B] d =
2
L 1 1
=1

where

(4.3)

Work done by external loading q(x)

L

{u } q ( x).dx ={ }

W =

e T

x =0
1

where

L
[ N ] q( ). d ={ e }T {F e }
2
= 1
T

L
F1e
{F } =
[ N ] q ( ). d =
,
e
2
F2
= 1
e

Fi =

L
N i ( ).q( ). d
2
= 1

8
(4.4)

Le

R1

The bar element

R2

u1,F1

u2,F2

=-1

Potential Energy of element is

e = U e W e
1
= { e }T [ K e ]{ e } { e }T {F e } { e }T {R h,e }
2
e
Equilibrium of element demands that

=1

EAe

{ R h ,e } =

Reaction vector

= 0

R1
R2

0 = { e }T [ K e ]{ e } { e }T {F e } { e }T {R h,e }

[ K e ]{ e } = {F e } + {R h,e }

(4.5)

=1

L
EAe 1 1
[K ] =
[ B ] [ D ][ B ] d =
L 1 1
2
= 1
e

Element Stiffness

Generalized force vector

L
Fi =
N i ( ).q ( ). d
2
= 1
e

4.2.2 The Euler beam element (2 nodes, 4 DOF)

w2,F2

w1,F1
Le

EI

1,M1

2,M2

=-1

2x
Le

=1

Displacement function (C1 continuity)

h

w =

N i ( ). i =[N1

N2

N3

w1
N4 ]

i =1

1
w2

= [ N ]{ e }

(4.6)

1
N1 = 2 + ( 2 3)
4

Le
N 2 = ( + 1)( 1)2
8

1
N 3 = 2 ( 2 3)
4

[N]: Shape function matrix

Le
N 2 = ( + 1)2 ( 1)
8

10

The Euler beam element

Strain and stress resultant vectors

=
h

d 2 wh
dx

(L )

d 2 wh

e 2

(L )

e 2

h = M h = EI e h = [ D][ B]{ e }

[6

[B]: Strain-displacement matrix

Le (3 1) 6

Le (3 + 1) { e } = [ B]{ e }

(4.7)

[ D] = EI e

Strain Energy of element

x= L

{ } { }dx ={

1
Ue =
h
2 x =0

=1

e T

1
1
L
[ B]T [ D][ B] d { e } = { e }T [ K e ]{ e }
2 =1
2
2

=1

where

12 ( EI e / Le )

[K ] =
e

Stiffness
matrix

L
[K ] =
[ B] [ D][ B] d
2
= 1
e

6 ( EI e / Le )
3

12 ( EI e / Le )
2

6 ( EI e / Le )

6 ( EI e / Le )

12 ( EI e / Le )

4 ( EI e / Le )

6 ( EI e / Le )

6 ( EI e / Le )

12 ( EI e / Le )

2 ( EI e / Le )

6 ( EI e / Le )

6 ( EI e / Le )

2 ( EI e / Le )

6 ( EI e / Le )

4 ( EI e / Le )

(4.8)
11

The Euler beam element

Work done by external loading q(x)
L

L
[ N ] q( ). d ={ e }T {F e }
W = {w } q( x).dx ={ }
2
x =0
=1
e

e T

F1
1

where

M1
L
{F } =
[ N ] q( ). d =
F2
2
=1
M2
e

(4.9)

Potential Energy of element is

e = U e W e
1 e T e
= { } [ K ]{ e } { e }T {F e } { e }T {R h,e }
2
e

Equilibrium of element demands that

= 0

[ K e ]{ e } = {F e } + {R h,e }

R1
{ R h ,e } =

R2
R3
R4

Reaction vector

(4.10)
12

4.3 The general methods of element formulation

The Rayleigh Ritz Method
Steps

The Galerkin Method

1. Use a displacement interpolation {u h } = [ N ]{ e }

2. Develop strain and stress resultants.

Differential eqn.

= [ B]{ }
h

(u) = q

h = [ D] h = [ D][ B]{ e }

3. Generate potential function

Virtual work

e = U e W e
1
= { e }T [ K e ]{ e } { e }T {F e } { e }T {R h,e }
2

= 0
e

4. Use PMPE

[ K e ] = [ B]T [ D][ B]dx = [ B]T [ D][ B] det[ J ]d

e

{F } = [ N ] q ( x).dx = [ N ] q( ). det[ J ]d
e

(4.12)
[ K e ]{ e } = {F e } + {R h,e }

e
T

{ ({u })- q}.d = 0

[ N ]T .

(4.11)

13

4.4 Some specimens from the Lagrangian family of C0 elements

Displacement approximation in an element through Lagrangian interpolation
h

u = N1 ( x)u1 + N 2 ( x)u2 + .....N n ( x)un =

N i ( x).ui =[ N ]{ e }

i =1
n

N1 ( x) =

( x2 x)( x3 x)...( xn x)
=
( x2 x1 )( x3 x1 )...( xn x1 )

( x j x)
j =2
n

( x j x1 )
j =2

n
i =1

N i ( ) = 1

( x j x)
N i ( x) =

j i
n

( x j xi )
j i

(4.13)

Such elements are of C0 continuity because across the element

boundary, only the function interpolated is continuous, but its
derivatives are not continuous.
n
i =1

N i ( ) = 1

Ensures rigid body displacement of element

14

(a) For the two noded bar element, a linear Lagrangian is used for
displacement, hence a constant strain over the element results:

R1

Le

R2

u1,F1

=-1

2x
e

=1

EAe

u =

N i ( ).ui =[N1

N i ( ) = 1

2
N 2 ( ) =

1+
2

1
u1

N2 ]

u2

i =1
2

1
2

u2,F2

Displacement function
h

N1 ( ) =

= [ N ]{ e }

N1 ( ) =

1
2

[N]: Shape function matrix

N 2 ( ) =

1+
2

(4.14)

i =1

Strain

=
h

u 2 u1
e

u1
= [ B ]{ e }
u2

[B]: Strain-displacement matrix

(4.15)
15

(b) For the three noded bar element, a quadratic Lagrangian is used for
displacement, hence a linear strain over the element results:
x

=-1

Le

=1
3 =0

Element Stiffness Matrix

7

1
EAe
dx
e
T
e
[ B] EA [ B ]
d = e
K ]=
d
2 = 1
3L

1
7 8
8 8 16

Element force vector

=

uh =

2x
Le

N i ( ).ui = [N1

i =1
3

1/ 6
dx
{F e Applied } = [ N ]T q.
d = qLe 1 / 6
d
1
2/3
q=force p.u.length

u1

N2

N 3 ] u2 = [ N ]{ e }
u3

(4.17)

1
1
N1 = (1 ), N 2 = (1 + ), N 3 = (1 2 )
2
2

(4.16)

N i ( ) = 1

i =1

2 du h
2 dN1
du h du h d
=
= e
= e
=
dx
d dx L d
L d
h

h = [ B]{ e }

dN 2
d

dN 3
{ e } =
d

2
e

1 2
,
2

2
e

1 + 2
,
2

2
e

( 2 ) { e }
16

Example 2. A robotic arm is rotating

at a constant angular velocity
=30 rad/sec. Determine the axial stress
distribution from centrifugal forces (using
two quadratic elements).

Assembly of stiffness matrices

[ KG ] =

7 8 1
8 16 8

[K e ] =

e =1

107 0.6
1
3 21
0

7
1 8 (1)
7

10
0
.
6
1
7 8 (3)
[K 1] =
3 21
8 8 16 (2)
7
1 8 (3)
7
10
0
.
6

1
7 8 (5)
[K 2 ] =
3 21
8 8 16 (4)

0 (1)
0 ( 2)

8 14 8 1 (3)
0 8 16 8 (4)

Element stiffness matrices

0
0

7 (5)

Element force vectors

1

dx
d
{F applied } = [ N ] q.
d
1
e

dx Le
=
d
2

2x
e

1
17

Computation of element force vectors

with averaged body force
f

Average

( x e mean ) 2
g

lb/in3

0.2836 (10.5) 30 2
f Average =
= 6.94
32.2 12
1 / 6 (1)
{F 1applied } = f (1) Average Ae Le 1 / 6 (2)
2 / 3 (3)
(1)

0.2836 (31.5) 30 2
f Average =
= 20.81
32.2 12
1 / 6 (3)
{F 2 applied } = f ( 2) Average Ae Le 1 / 6 (5)
2 / 3 ( 4)
( 2)

The global (applied) force vector after force assembly

{F G Applied } = [14.57 58.26 58.26 174.79 43.70]T

18

Solving for displacements

(u1=0). Eliminating u1
8

16

u2

58.26

58.26
107 0.6 8 14 8 1 u3
=
0 8 16 8 u4
174.79
63
0
1 8 7 u5
43.7

{ G } = 10 3 0 0.5735 1.0706 1.4147 1.5294]T

Element strains and stresses

h = [ B]{ e } =

h = E h = 107

2
e

Exact

1 2
,
2

1 2
,
2

1 + 2
,
2

2
2
e

2 L2
2g

2
e

1 + 2
,
2

x
1
L

(2 ) { e }
2
e

(2 ) { e

Question
Why does the stress obtained from
the present FE solution does not
19
coincide with the best-fit ?

Question
Why does the stress obtained from the present
solution does not coincide with the best-fit ?
From the differential equation

du
d
EA
= q (x )
dx
dx

Exact stress (quadratic) is

=E

du
1
1
= e (q)dx = e ( fAe )dx
dx
A e
A e

2 L2

x
1
L

Exact
=
With the consistent force vector, we
2g
can get the FE stress as the linear best-fit
to the analytical stress in each element. This implies that the true body force f should
be linearly varying (and not constant average)
within the element
This is the consistent force vector
2 x
2 x e
e
A
f =
q = fA =
for each element.
g
g
Le

Le
e
T
T
d
{F applied } = [ N ] q( x).dx = [ N ] q( ).
2
xe L
0
1
q( ) =

x( )
g

A =

[ x

+ L ( + 1) / 2] e
A
g
e

Le

1
=-1

3
=0

2
20
=1

Question
Centrifugal body force, f

Why does the stress obtained from the present

solution does not coincide with the best-fit ?

x
0

From the differential equation

du
Exact
d

EA
= q (x )
FE (averaged body force)
dx
dx
FE (consistent force
with actual linearly
Exact stress (quadratic) is
varying body force)

Axial
Stress

=E

du
1
1
= e (q)dx = e ( fAe )dx
dx
A e
A e

2 L2

x
1
L

Exact
=
With the consistent force vector, we
2g
can get the FE stress as the linear best-fit
to the analytical stress in each element. This implies that the true body force f should
be linearly varying (and not constant average)
within the element
This is the consistent force vector
2 x
2 x e
e
A
f =
q = fA =
for each element.
g
g
Le

Le
e
T
T
d
{F applied } = [ N ] q( x).dx = [ N ] q( ).
2
xe L
0
1
q( ) =

x( )
g

A =

[ x

+ L ( + 1) / 2] e
A
g
e

Le

1
=-1

3
=0

2
21
=1

4.5 Some specimens from the Isoparametric family of elements

Displacement approximation and Geometry of element described by the
same set of Lagrangian interpolation functions

{u h } = [ N ]{ e }

(4.18)

{x} = [ N ]{x e }

Advantages:
Curved/inclined outlines/surfaces of elements can be suitably modeled.
(a) The aforementioned two-noded bar element is a linear
isoparametric element.
Le

=-1
h

u =
x=

2
i =1
2

N i ( ).ui =[N1

N i ( ).xi =[N1

i =1

N2 ]
N2 ]

u1
= [ N ]{ e }
u2

x1
= [ N ]{x e }
x2

EAe

=1

1
N1 ( ) =
2

1+
N 2 ( ) =
2
22

(b) The three-noded (quadratic) bar element with quadratic geometry

description

uh =

N i ( ).ui = [N1

N2

N3 ]

i =1

u1
u2 = [ N ]{ e }
u3

1
1
N1 = (1 ), N 2 = (1 + ), N 3 = (1 2 )
2
2
x1
3
x = N i ( ).xi = [N1 N 2 N 3 ] x2 = [ N ]{x e }
i =1
x3
du h d du h (du h / d )
=
=
=
= [ B]{ e }
dx
dx d
dx / d
h

(4.19)

Le
x

=-1

=1

3 =0

23

(c) The four-noded (linear) quadrilataral element

for plane stress problem

y,v

Node,i

4
2
x,u

1
(1 )(1 ),
4
1
N3 = (1 + )(1 + ),
4
N1 =

Shape functions

Displacement
N1
uh
=
0
vh

0
N1

N2
0

0
N2

N3
0

0
N3

N4
0

0
N4

u1
v1
u2
v2
= [ N ]{ e }
u3
v3
u4
v4

-1

-1

+1

-1

+1

+1

-1

+1

1
(1 + )(1 )
4
1
N 2 = (1 )(1 + )
4

N2 =

Geometry
4

x
y

N i xi
=

i =1
4

= [ N ]{ X e }
N i yi

i =1

(4.20)

(4.21)
24

(c) The four-noded (linear) quadrilataral element

for plane stress problem
Element stress

xh
{ h } = y h =
xy h

E
(1 2 )

1
0

xh
y h = [ D][ B ]{ e }

(1 ) / 2 xy h

1
u1

Element strain
h
xh
{ h } = y h
xy h

v1
u
u2
x

N
x
N
x
N
x
N
x
/
0
/
0
/
0
/
0
1
2
3
4
v2
v h
=
=
N1 / y
N 2 / y
N 3 / y
N 4 / y
0
0
0
0
= [ B]{ e }
u
y
N1 / y N1 / x N 2 / y N 2 / x N 3 / y N 3 / x N 4 / y N 4 / x 3
h
h
u
v
v3
+
y
x
u4
v4

Jacobian
x /
( x, y )
[J ] =
=
x /
( , )

y /
y /

i =1
4
i =1

(Ni / )xi
(Ni / )xi

4
i =1
4
i =1

(Ni / ) yi
(Ni / ) yi

(4.22)
(4.23)
25

(c) The four-noded (linear) quadrilataral element

for plane stress problem

y,v

3
Ni / x
1 N i /
= [J ]
Ni /
N i / y

4
2
x,u

Element stiffness matrix:

1 1

[K e ] =

[ B]T [ D][ B].t.dxdy =

e

[ B ]T [ D][ B].t det[ J ].dd

(4.24)

11

t=element thickness

26

Example 3. Plane stress analysis of a plate with a hole using quad4 elements

y
x

xy
27

4.6 Numerical Integration by Gauss Quadrature

If f( ) is a polynomial of degree (2n-1) or less, then it is exactly integrated
within the domain 1 1 by the n-point Gaussian quadrature rule.
f ( ) = p2 n ( ) = a1 + a2 + a3 3 ..... + a2n 2 n1
1

f ( )d =
1

n
i =1

wi f ( i )

1 1

(4.25)

28