Principles of Minimum Potential

Energy and Rayleigh-Ritz Method

 Axially loaded elastic bar
y A(x) = cross section at x
b(x) = body force distribution
(force per unit length)
x E(x) = Youngs modulus
x u(x) = displacement of the bar at
x=0 x=L x

Differential equation governing the response of the bar

d du
AE b 0; 0 x L
dx dx
 y

F
x
x
x=0 x=L
An example of Boundary conditions
u 0 at x 0 u 0 at x 0
du
u 1 at x L EA F at x L
dx
d 2u d 2u b
AE 2 b; 2
;
dx dx AE

Differential equation + Boundary conditions = Strong form

of the boundary value problem
 Real Life

o Real Physical problems are Differential equations (strong)

governed by Ordinary or formulation
Partial differential
equations which satisfy
Variational (weak) formulation
Boundary conditions
Initial conditions

Approximate the weak form using

finite elements
 Potential energy

The potential energy of an elastic body is defined as

Strain energy (U) potential energy of loading W

Virtual work results from fictitious arbitrary

displacements that are assumed to occur instantaneously;
dui are known as virtual displacements

Virtual work done by external forces as a result of a

virtual displacement field is given by:

W F0 dV u
V
 Principle of Stationary Potential Energy

o The virtual change in the potential energy of the structure as a

result of a virtual displacement field is given by:

W U
o A structure is in stable static equilibrium its potential energy is
minimum;
o Static equilibrium means that the displacement field is such
that
W U 0
 Principle of Minimum Potential Energy

W F0 dV u
V
Among all admissible displacements that a body can have, the
one that minimizes the total potential energy of the body
satisfies the strong formulation

Admissible displacements: these are any displacement that

satisfies boundary conditions of the original problem and
continuity requirements
 The Principle of Minimum Potential Energy - Why

o The Principle of Minimum Potential Energy and the strong

formulation (PDE) are exactly equivalent statements of the
same problem (find displacements/deformation/stresses).

o The exact solution (uexact) that satisfies the strong form, renders
the potential energy of the system a minimum.

o So, why use the Principle of Minimum Potential Energy?

it is much less demanding than the strong formulation.
Requires only the first derivative to be finite
The admissible displacement needs to satisfy only the displacement
boundary condition. Force boundary condition is satisfied automatically.
 Assumed Displacement Fields

o Finite element formulation, takes as its starting point,

not the strong formulation, but the Principle of
Minimum Potential Energy.
o In general it is very hard if not impossible to solve for
displacement fields using the PDEs that relates strains
and stresses in a structure
o The minimization of the potential energy is also
difficult to perform exactly.
o An approximate solution can be found by assuming a
displacement field with a number of unknown
parameters and solving for the parameters using the
principle of stationary potential energy
 Equation of Motion ( deformable body)
W U
0
u u u
1
U dV W uT b dV+ uT t d
V 2
V

Boundary of the
t b structure

Traction on Body
external surface forces
1
dV+ uT b dV+
uT t d
V 2 V
u
1D x ; E x
x
Equation of Motion ( deformable body)
1
( dV) (
uT b dV) ( uT t d)
V 2 V
+ + 0
u u u u

u (displacement) is the only unknown

F

u(x,y,z) is a continuous function

(infinite values)
 Rayleigh-Ritz Method
o The assumed approximate displacement fields can be
written as: l
ai f i x , y , z
u x , y , z i 1
m
u x v x , y , z b j g j x , y , z
j 1
w x , y , z n
c k hk x , y , z
k 1
where fi, gj, hk can be any family of functions (e.g.
polynomials or sinuisoids)
o Classical form Approximate field is defined over the
entire region of interest
o Finite Element Form- Approximate field is defined in
piecewise fashion
 Rayleigh-Ritz Method (contd)

o The functions fi, gj, hk are known as basis

functions
o The undetermined coefficients ai, bj, ck are
known as generalized coordinates (unknown)
o Basis functions must satisfy the constraints or
boundary conditions in the given problem;
o Such functions are known as kinematically
admissible functions
 Rayleigh-Ritz Method (contd)

o Once an approximate displacement field has

been assumed the potential energy of the
structure becomes a function of the
undetermined coefficients

ai , b j , c k i 1 l , j 1 m , k 1 n

o Applying the principle of stationary potential

energy to this function results in:
l n m

ai b j ck 0
i 1 ai j 1 b j k 1 c k
 Rayleigh-Ritz Method (contd)

o If the ai, bj, ck are independent then the dai, dbj,

dck are linearly independent; this means

0 i 1 l , j 1 m , k 1 n
ai b j c k

o which results in l+m+n linear equations for

l+m+n unknowns ai, bj, ck
o The solution for ai, bj, ck yields the approximate
displacement field u(x)
 Strain energy of a linear spring

F x
k k
1 F
u k
u
F = Force in the spring
u = deflection of the spring
k = stiffness of the spring

Hookes Law
F = ku
 Strain energy of a linear spring

dU Differential strain energy of the

F spring for a small change in
displacement (du) of the spring
dU Fdu
u u+du For a linear spring
dU kudu
The total strain energy of the spring
u 1 2
U k u du k u
0 2
 Strain energy of a nonlinear spring

dU dU Fdu
F The total strain energy of the spring

u u+du

u
U F du Area under the force displacement curve
0
 Potential energy of a linear spring

x
k Potential energy of the loading
F
W Fu
k
u

Strain energy (U) potential energy of loading W

1 2
ku Fu
2
 Principle of minimum potential energy
k1 k2 F
x
d1x d 2x d 3x

Total potential energy of the system

1 1
system k 1 (d2 x ) k 2 (d3 x d2 x )2 Fd3x
2

2 2
Find the equilibrium > minimise the potential energy
system
k 1d2 x k 2 (d3 x d2 x ) 0 Equation (1)
d2 x
system
k 2 (d3 x d2 x ) F 0 Equation (2)
d3 x
 Principle of minimum potential energy

In matrix form,

k 1 k 2 k 2 d2 x 0
k
2 k 2 d3 x F

Already derived.
 Axially loaded elastic bar
A(x) = cross section at x
b(x) = body force distribution
x (force per unit length)
x E(x) = Youngs modulus
x=0 x=L u(x) = displacement of the bar
at x

du
Axial strain
dx
du
Axial stress E E
dx
2
1 1 du
Strain energy per unit volume of the bar dU E
2 2 dx
Strain energy of the bar
1 L1
U dU 2
dV
x 0 2
Adx since dV=Adx
 Axially loaded elastic bar
Strain energy of the bar
2
L1 1 L du
U A dx EA dx
0 2 2 0 dx

Potential energy of the loading

L
W bu dx Fu(x L)
0

Potential energy of the axially loaded bar

2
1 L du L
EA dx bu dx Fu(x L)
2 0 dx 0
 Example
E=1
F=2 x
A=1
x=1
x=0 x=2
 Example
E=A=1
F=2 x

x=1
x=0 x=2
2
1 du
2
The potential energy (u) dx Fu(x
1)
2 0 dx
Potential Energy
Strain Energy of load F applied
at x 1

Let us assume a polynomial admissible displacement field

2
u a0 a 1 x a 2 x
 Example
admissible displacement must satisfy the displacement
boundary conditions

u(x 0) a0 0
u(x 2) a0 2a1 4a2 0
Hence, we obtain
a0 0
a1 2a2

the admissible displacement is

u a0 a1 x a 2 x 2
a2 2 x x 2
 Example

o Applying the Rayleigh Ritz principle to obtain

the unknown (a2)
2
1 2 du
(u) dx Fu(x 1)
2 0 dx
2
1 d
2

2 0 dx


a2 2 x x 2 dx F a2 2 x x 2 evaluated
at x 1

4 2
a2 2a2
3
8 3
0 a 2 2 0 a2
a2 3 4
 Example

o The approximate solution to this problem, using

the Rayleigh-Ritz principle is
u a0 a 1 x a2 x 2
a2 2x x 2
3
2x x 2
4

o The exact answer to this problem is

x for 0 x 1
uexact
2 x for 1 x 2
The displacement solution :

Exact solution
1

0.8 Approximate
solution

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
x
The stress within the bar:

1.5
Exact Stress

0.5
Approximate
stress
Stress

-0.5

-1

-1.5

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

x
 How can you improve the approximation?

Pascals Triangle:
Degree of polynomial

0
1
1
x y
2
x2 xy y2
3
x3 x2 y x y2 y3

x4 x3 y x2 y2 x y3 y4 4

x5 x4 y x3 y2 x2 y3 x y4 y5 5
 Example

o An aluminum (E = 7.17 1010 Pa) rod of variable cross section is subjected

to a point load of 453.6 N at its narrower end. Determine the displacement
field in the rod and the deflection at its loaded end using the Rayleigh-Ritz
method. Compare your results with the exact solution.

453.6 N
Al = 0. 00016 m2

Ar = 8. 10-5 m2
0.305 m

PL Al
ux ln
E Al Ar A Al Ar x
l
L
 Solution

o Assume a quadratic displacement function that is consistent

with the constraints (i.e. is kinematically admissible):
2
u x a1 x a2 x
o Note that this function is admissible because it satisfies the
only boundary condition of u(0) = 0.
o The constants a1 and a2 are undetermined parameters
o A higher order polynomial or a function of different type could
have been assumed.
 Solution (contd)

o Using the definition of linear strain we can write for the strain
at a point along the rod:
du
x a1 2a2 x
dx
o The unit strain energy can now be written:

1 T 1 2
U 0 E E x
2 2
1 2
E a1 2a2 x
2
 Solution (contd)

o The strain energy stored in the rod can be computed from:

U U 0dV
o In this case V

Ar Al
dV A x dx Ar x dx
L
 Solution (contd)

o Thus the expression for the total strain energy becomes:

L
1 2 Ar Al
U E a1 2a2 x Ar x dx
2 0 L
o Substituting the given values for the various parameters we
obtain

2 2 6
U 1.32a 28.1a1a2 211a
1 2 10 J
 Solution (contd)

o In this case there is only one external applied force which

causes a displacement at the free end of the rod. The work
done by this force is:
2
W Pu L P a1L a2L
o Substituting the given values we obtain:

3
W 1.35a1 16.2a2 10 J
 Solution (contd)

o Thus the potential energy of the system is:

U W
1.32a12 28.1a1a2 211a22 106 1.35a1 16.2a2 10 3 J

o From the principle of stationary potential energy we obtain:

0
a1

0
a2
 Solution (contd)

o Solving these equations we obtain the approximate

displacement function

2 6
u x 9x 0.37 x 10 m
 Comparison with Exact Results

o The plot below compares the approximate solution obtained

by the Rayleigh-Ritz method to the exact solution
-3
x 10
7
Exact
Approximate
6

4
u (in)

0
0 2 4 6 8 10 12
x (in)