UNIT- I

INTRODUCTION
1. EXPLAIN THE GENERAL STEPS OF THE FINITE ELEMENT ANALYSIS
This section presents the general procedure of finite element analysis. For simplicity's sake, we will
consider only the structural problems.
The following two general methods are associated with the finite element analysis.
They are:
(i) Force method.
(ii) Displacement or stiffness method.
In force method, internal forces are considered as the unknowns of the problem. In displacement or
stiffness method, displacements of the nodes are considered as the Unknowns of the problem.
Among these two approaches, displacement method is more desirable because its formulation is simpler
for most structural analysis problems. So, a vast majority of general purpose finite element programs have
used the displacement formulation for solving structural problems.
(i) One dimensional elements: A bar and beam elements are considered as one dimensional element.
The simplest line element also known as linear element has two nodes, one at each end as shown in
Fig.l.2.

(ii) Two dimensional elements:Triangular and rectangular elements are considered as two dimensional
elements. These elements are loaded by forces in their own plane. The simplest two dimensional
elements have comer nodes as shown in Fig.l .3.

(iii) Three dimensional elements:The most common three dimensional elements are tetrahedral and
hexahedral (Brick) elements. These elements are used for three dimensional stress analysis problems. The
simplest three dimensional elements have
corner nodes as shown in Fig.l.4.

(iv) Axisymmetric elements: The axissymmetric element is developed by rotating a triangle or

quadrilateral about a fixed axis located in the plane of the element through 360 o. it is shown in fig. 1.5.
When the geometry and loading of the problems are axisymmetric, these elements are used.

Step 2: Numbering of Nodes and Elements: The nodes and elements should be numbered after
discretization process. The numbering process is most important since it decide the size of the stiffness
matrix and it leads the reduction of memory requirement. While numbering the nodes, the following
condition should be satisfied.
Maximum node number Minimum node number = Minimum.
Longer side Numbering process:

Shorter side Numbering process:

From the above equation, we came to know, shorter side numbering process is followed in the
finite element analysis and it reduces the memory requirements.
Step 3 : Selection of Displacement Function or Interpolation Function
It involves choosing a displacement function within each element. Polynomials of linear,
quadratic and cubic form are frequently used as displacement functions because they are simple to work
within finite element formulation.

The polynomial type of interpolation functions are mostly used due to the following reasons.

1. It is easy to formulate and computerize the finite element equations.

2. It is easy to perform differentiation or integration.
3. The accuracy of the results can be improved by increasing the order of the polynomial.
Fig 1.7 shows the polynomial approximation in one dimension.
Let us consider (x) is a field variable.
Case (i): Linear Polynomial
One dimension problem (x) = a0 + a1x
Two dimension problem (x,y) = a0 + a1x + a2y
Three dimension problem (x,y,z) = a0 + a1x + a2y + a3z
Case (ii): Quadratic Polynomial
One dimension problem (x) = a0 + a1x + a2x2
Two dimension problem (x,y) = a0 + a1x + a2y + a3x2 + a4y2+ a5xy
Three dimension problem (x,y,z) = a0 + a1x + a2y + a3z + a4x2 + a5y2+ a6z2 + a7xy + a8yz + a9xz
Step 4: Define the material behavior by using Strain Displacement and Stress Strain
Relationships
Strain Displacement and Stress Strain relationships are necessary for deriving the equations
for each finite element.
In case of one dimensional deformation, the strain-displacement relationship is given by,

du
dx

Where, u = Displacement field variable along x direction.

e = Strain.
The stress-strain relationship is given by

Ee
Where, = Stress in x direction
E = Modulus of elasticity or Youngs modulus
Step 5: Derivation of element stiffness matrix and equations:
The finite element equation is in matrix form as,

In compact matrix form as,

F k u
e

Where, e is a element, {F} is the vector of element nodal forces, [k] is the element stiffness
matrix and [u} is the element displacement vector.
This equation can be derived by any one of the following methods.
(i) Direct Equilibrium Method: This method is much easier to apply for line or onedimensional
elements.
(ii) Variational Method: This method is most easily adaptable to the determination ofelement
equations for complicated elements (i.e., element having large number of degrees offreedom)
likeaxisymmetric stress element, plate bending element and two or threedimensional solid stress element.
(iii) Weighted Residual Method: This method is (Galerkin's method) useful fordeveloping the
element equations in thermal analysis problems. They are especially usefulwhen a functional such as
potential energy is not readily available.
Step 6: Assemble the element equations to obtain the global or total equations:
The individual element equations obtained in step 5 are added together by using a methodof
superposition i.e., direction stiffness method. The final assembled or global equation whichis in the form
of

F K u
Where {F} = Global force vector,
[K] = Global stiffness matrix
{u} = Global displacement vector
Step 7: Applying boundary conditions:
From equation (1.5), we know that, global stiffness matrix [K] is a singular matrixbecause its
determinant is equal to zero. In order to remove this singularity problem, certainboundary conditions are

applied so that the structure remains in place instead of moving as arigid body. The global equation (1.5)
to be modified to account for the boundary conditionsof the problem.
Step 8: Solution for the unknown displacements:
A set of simultaneous algebraic equations formed in step 6 can be written in expandedmatrix form as
follows:

These equation can be solved and unknown displacements { u } are calculated by usingGaussian
elimination method or Gauss-Seidel method.
Step 9: Computation of the element strains and stresses from the nodal displacements, {u}:
In structural stress analysis problem, stress and strain are important factors. From thesolution of
displacement vector { u }, stress and strain value can be calculated.
In case of one dimensional deformation, the strain-displacement relationship is given by,

Strain, e

du
dx

u 2 u1
x 2 x1

Where, u1 and u2 are displacement at node 1 and 2

x1 x2 = Actual length of the element.
From that, we can find the strain value.
By knowing the strain, stress value can be calculated by using the relation,
Stress, = E e
where, E = Young's Modulus.
e =Strain.
Step 10: Interpret the results (Post Processing):

Analysis and evaluation of the solution results is referred to as post-processing. Postprocessor computer
programs help the user to interpret the results by displaying them ingraphical form.
Steps 1 to 10 are summarized as follows:

1. The differential equation of a physical phenomenon is given by,

d2 y
500x 2 0,
2
0 x 1
dx
y a1 x x 4
Trail function,

y 0 0

Boundary conditions are,

y 1 0

Calculate the value of the parameter a 1 by the following methods:(i) Point collocation; (ii)
Subdomain collocation; (iii) Least squares; (iv) Galerkin.

Given:

Differential equation,

d2y
500x 2 0,
0 x 1
dx 2

y a1 x x

Trail function,

.(1)

y 0 0

Boundary conditions are,

y 1 0

To find: The value of parameter a1 by,

(i) Point collocation method,
(ii) Subdomain collocation method,
(iii) Least squares method,
(iv) Galerkins method
Solution: First we have to verify, whether the trial function satisfies the boundary conditions or not.
Trial function is, y = a1(x-x4)
When x = 0, y = 0
x = 1, y = 0
Hence it satisfies the boundary conditions.
(i) Point collocation method: y = a1(x-x4)

dy
a1 1 4x 3
dx
d2 y
a1 12x 2
2
dx
d2 y
a112x 2
2
dx

d2 y
dx 2
Substituting

value in given differential equation (1),

Residual, R = -12a1x2+500x2

.(2)

In point collocation method, residuals are set to zero,

R = -12a1x2+500x2 = 0

.(3)

In this problem, we have to find only one parameter, a 1. So, only one collocation point is needed.
The point may be chosen between 0 and 1. Let us take 1/2.

x
Substituting

1
2
in equation (3),

1
1
R 12a1 500
2
2
1
1
12a1 500 0
4
2
3a1 125 0

a1 41.66
.(4)

Hence the trail function is, y = 41.66(x-x4)

(ii) Subdomain collocation method:
1

Rdx 0
0

This method requires,

Substitute R value,

12a x

x 3
12a1
3

500x 2 dx 0
1

x3
500
3 0

0
0
12a1
500
1 0 1 0 0
3
3
12a1 500

0
3
3
12a1 500 0
12a1 500
a1

500
12

a1 41.66
.(5)
Trail function is, y = 41.66 (x-x4)

(iii) Least squares method:

1

R dx
2

This method requires, I =

1

I
R
R
dx
a1 0 a1
It can also be written as,
We know that, R = -12a1x2+500x2

R
12x 2
a1
R
a1
Substitute R and

values in equation (6),

.(6)

I
12a1x 2 500x 2 12x 2 dx
a1 0
I
0
a1
The requirement is,
1

12a x

500x 2 12x 2 dx 0

6000x 4 dx 0

144a x
1

x 5
144a1
5

x5
6000
5 0

0
0
144a1
6000
1 0
1 0 0
5
5
28.8 a1 = 1200
a1 = 41.66

The trail function is, y = 41.66(x-x4)

(iv) Galerkins method: In tis method, the trail function itself is considered as the weighing function, w i
1

w Rdx 0
i

.(8)
Here, the trail function is y = wi = a1(x-x4)
Substitute wi and R values in equation (8),

a x x 12a x
4

500x 2 dx 0

a1 12a1 x 3 500x 3 12a 1x 6 500x 6 dx 0

0

1
1
1
1
x 4
x 4
x7
x7
a1 12a1 500 12a1 500
0
7 0

4 0
4 0 7 0

12a1
500
12a
500
1 0 1 0 1 (1 0) (1 0) 0
4
4
7
7

-3a1 + 125 + 1.714a1 - 71.428 = 0

- 1.286a1 = -53.572
a1 = 41.66

(9)

The trail function is, y = 41.66(x-x4)

From equations (4), (5), (7) and (9), we know that the value of parameter a 1 is same for all the four
methods.

2. Determine the displacements of nodes 1 and 2 in the spring system shown in Fig. 1. Use minimum
potential energy principle to assemble equations of equilibrium.

Given:
k1 = 60 N/m,

F1 = 100 N

k2 = 75 N/m,

F2 = 80 N

k3 = 100 N/m
To find: Displacements of nodes 1 and 2.
Solution: Let u1 and u2 be the displacements of nodes 1 and 2. Then the extensions of springs are,
1 = u1, 2 = u1,3 = u2 u1
We know that,
Minimum of potential energy principle,
= Strain energy work done

1
1
1
k112 k 2 22 k 332 100u1 80u 2
2
2
2

1
1
1
2
k1u12 k 2 u12 k 3 u 2 u1 100u1 80u 2
2
2
2
(1)

0
u1
Now,

k1u1 k 2 u1 k 3 u 2 u1 1 100 0

k1 k 2 k 3 u1 k 3u 2 100
(2)

0
u 2
Similarly,

k 3 u 2 u1 80 0

k 3 u1 k 3u 2 80
(3)

Arranging equation (2) and (3) in matrix form,

k1 k 2 k 3

k 3

k 3 u 1
100

k 3 u 2
80
60 75 100 100 u 1 100

100
100 u 2 80

235 100 u 1 100

100 100 u 80

235u1 100u 2 100

.(4)

100u1 100u 2 80
.(5)

135u1 180
180
135
u1 1.333
u1

Substitute the u1 value in equation (5)

100(1.333) 100u 2 80
u 2 2.133
Result: Displacements of nodes,
u1 = 1.333 m
u2 = 2.133 m

3. A simply supported beam subjected to uniformly distributed load entire span. Determine the
bending moment and deflection at mid span by using Rayleigh Ritz method and compare with exact
solutions.

To find: 1. Deflection and Bending moment at mid span.

2. Compare with exact solutions.
Solution: We know that, for simply supported beam, the Fourier series,

a _ sin

n 1,3

nx
l
is the approximating function.

To make this series more simple let us consider only two terms.

y a1 sin

x
3x
a 3 sin
l
l

Deflection,

(1)

where a1, a2 are Ritz parameter.

We know that,
Total potential energy of the beam, = U-H
Where, U = Strain energy,
H = work done by external force.
The strain energy, U of the beam due to bending is given by,
(2)

dy
x
3x 3
a1 cos a 2 cos

dx
l l
l
l
dy a1
x a 2 3
3x

cos

cos
dx
l
l
l
l
2
a
d y
x
3
3x 3
1 sin
a 2 sin

2
dx
l
l l
l
l
l

(2)

d2 y
a 1 2
x
9 2
3x

sin

a
sin
2
2
2
2
dx
l
l
l
l
(3)

d2 y
dx 2
Substituting

value in equation (3),

2

EI a 2
x a 2 92
3x
U 1 2 sin
2 sin
dx
2 0
l
l
l
l
l

EI a 2
x a 2 92
3x
U 1 2 sin
2 sin
dx
2 0 l
l
l
l
l

EI 4
x
3x
U 4 a1 sin
9a 2 sin
dx
2 l 0
l
l
l

EI 4
x
3x
x
3x
U 4 a12 sin 2
81a 2 2 sin 2
2a1 sin 9a 2 sin
dx
2 l 0
l
l
l
l
[(a+b)2 = a2 + b2 +2ab]
l

EI 4 2 2 x
3x
x
3x
U 4 a1 sin
81a 2 2 sin 2
18a 1a 2 sin sin dx
2 l 0
l
l
l
l
(5)
l

2
2
a1 sin
0

x
1
2x
dx a12 1 cos
dx
l
2
l
0

a12
2x
1 cos
dx

2 0
l

a12

2x
dx
l

2x
sin

l
x 0 2xl

l 0

a12

dx cos

a12
l 0 0 0
2

2
Q sin x

1 cos 2x
2

a12l
2
(6)
l

3x
a 22l
0 81a 2 sin l 81 2
2

Similarly,
l

(7)

18a1a 2 sin
0

x
3x
x
3x
sin
dx 18a1a 2 sin sin
dx
l
l
l
l
0
l

18a1a 2 sin
0

3x
x
sin dx
l
l

1
2x
4x
18a1a 2 cos
cos
dx
2
l
l
0

Q sin A sin B

cos A B cos(A B)

l
l
18a1a 2
2x
4x
cos
dx

cos
dx

2 0
l
l
0

18a1a 2
2

2x
4 x
sin
sin

l
l
2
4

l 0
l
0
l

9a1a 2 0 0 0
l

18a a

1 2

sin

Q sin 2 0;sin 4 0;sin 0 0

x
3x
sin
dx
l
l
0

Substitute (6), (7) and (8) in equation (5),

EI 4 a12l 81a 22l

0
2 l 4 2
2

(8)

EI 4 a12 l 81a 22l

2 l4 2
2

EI 4 2
a 81a 22
3 1
4 l
U=

(9)

We know that,
l

x
3x

0 wydx 0 w a1 sin l a 2 sin l dx

Work done by external force, H =
l

x
3x

0 wydx 0 w a1 sin l a 2 sin l dx

l

x
3x

w a1 sin
a 2 sin
dx
l
l

l
x
3x
w a1 sin
dx a 2 sin
dx
l
l
0
0

x
3 x

cos
cos

l
l
a 2

3

l 0
l

w a1

0
l
l
a1l
a 2 l
x
3 x
w
cos
cos
l 0 3
l 0

a l
a1l

w
1 1 2 1 1
3

2a1l 2a1l

2wl
a 2
a1

3

Substitute (9) and (10) values in equation (2)

=U-H

(10)

EI 4 2
a1 81a 22
4 l3
=

2wl
a 2
a1

3

(11)

For stationary value of , the following conditions must be satisfied.

0
a1

0
a 2
and

EI4
2wl
3 (2a1 )
0
a1
4l

EI4
2wl
(2a1 )
3
4l

3
4wl
a1
EI5
EI4
2wl 1
3 (162a1 )
0
a 2
4l
3
EI4
2wl 1
(162a1 )

3
4l
3
Similarly,

a2

2wl
4l3

3 162EI4

a2

4wl4
243EI5

y a1 sin

x
3x
a 3 sin
l
l

We know that,
Substituting a1 and a2 values,

4wl3
x
4wl4
3x
sin

sin
5
5
EI
l 243EI
l

(12)

l
2

We know that, maximum deflection occurs at

x
Substitute

l
2
in equation (12),

l
l
3
4
4wl
2 4wl sin
2

sin
5
5
EI
l
243EI
l
3
4
4wl

4wl
3

sin
sin
5
5
EI
2 243EI
2
3
4
4wl
4wl

5
EI 243EI5
4wl3
1

1
5
EI
243
4wl3
3.98wl3

0.9958

EI5
EI5
wl3
0.0130
EI5
3

y max
y max
y max
y max
y max
y max

4. A bar of uniform cross section is clamped at one end and left free at the other end and it is
subjected to a uniform axial load P as shown in Fig. Calculate the displacement and stress in a
barby using two terms polynomial and three terms polynomial. Compare with exact solutions.

To find: 1. Displacement of the bar, u.

2. Stress in the bar,
By using two terms and three terms polynomial.
Solution: We know that, Polynomial function for displacement is,
u = a0+a1x+a2x2+a3x3+a4x4+ +anxn
Case (i): Considering two terms of polynomial,
i.e., u = a0 + a1x

(1)

Apply boundary condition,

at x = 0, u = 0,
0 = a0 + 0
a0 = 0
substituting a0 value in equation (1),
u = a 1x

(2)

du
a1
dx
We know that,
Total potential energy of the bar, = U H
Where, U = strain energy of the bar,
H = work done by external force of the bar

(3)

EA du
U
dx
2 0 dx
Strain energy,
l

EA 2

a1 dx
2 0
EAa12
l
x 0
2
EAa12 l
U
2

H Pdx uAdx
Work done by external force,
l

A udx A a1xdx
0

x 2

Aa1
H

Aa1l
2

Aa1 2
l
2

Substitute U and H values in equation (3),

EAa12l Aa1l2

2
2

For stationary value of , the following condition must be satisfied.

0
a1
i.e.

EA 2a1 l Al2

0
2
2
Al2
EAa1l
0
2
Al2
EAa1l
2
l
a1
2E
Substitute a1 value in equation (2),

u a1 x

l
x
2E

l
x
2E
u u1 u 0

l
l0
2E

We know that, extension of the bar,

[at x = l, u = u1, at x = 0, u = u0 = 0]

l2
2E

l2
2E

Extension or displacement of the bar,

du
l
E
dx
2E

Stress in the bar,

Stress in the bar,

Case (ii) : Considering three terms of polynomial,

l
2

u a 0 a1 x a 2 x 2
i.e.,
Apply boundary condition,

at x = 0, u = 0,
0 = a0 + 0 + 0
a0 = 0

Substitute a0 value in equation (8)

u = a1x + a2x2

.(9)

du
a1 2a 2 x
dx
We know that, Total potential energy of the bar, = U H
2

U=

(10)

EA du
EA
2
a1 2a 2 x dx
dx

2 0 dx
2 0

Strain energy,
l

EA
2

a12 2a 2 x 2a1 2a1a 2x dx

2 0
l

EA
a12 dx 4a 22 x 2dx 4a1a 2 xdx

2 0

x 3
x 2
EA 2
l

a1 x 0 4a 22 4a1a 2
2
3 0
3 0

4a 22 3
4a a
EA 2
a
l

l 0 1 2 l 2 0

2
3
2

EA 2 4a 22 3
U
a1 l
l 2a1a 2 l 2

2
3

(11)
Work done by external force,

H Pdx uAdx
0

A udx A a1x a 2 x 2 dx
l

A a1xdx a 2 x 2dx
0

x 2
x3
A a1 a 2
2 0 2 0
a
a

A 1 l 2 0 2 l3 0
3
2

a1 2 a 2 3
l l
3
2

H A

(12)

Substitute (11) and (12) values in (10),

=UH

EA 2 4a 22 3
a
a
a1 l
l 2a1a 2 l 2 A 1 l 2 2 l3

2
3
3
2

(13)

For stationary value of , the following conditions must be satisfied.

0
a1
i.e.,

0
a 2
and

 l2
EA
2a1l 0 2a 2l2 A 0 0

a1
2
2

EA
2a1l 2a 2l2 A
2

l2
0
2

Al2
0
2
Al2

EA a1l a 2l 2
EA a1l a 2l 2
a1l a 2 l 2

a1 a 2 l

l 2
2E

l
2E
(14)

0
a 2
Similaraly,

EA
8a 2 3
l3
2

2a
l

A
0

0
1

a 2
2
3
3

EA 8a 2 3
Al3
2
l 2a1l
2 3
3

8 3
2Al3
2
a 2 l 2a1l
3
3EA
8 3
2l3
a 2 l 2a1l2
3
3E
4 3
l3
2
a 2l a1l
3
3E
4
l
a 2 l a1
3
3E
4
l
a1 a 2 l
3
3E

a1 1.333a 2 l

l
3E
(15)

a1 2a 2 l

l
2E

Solving (14) and (15),

a1 1.333a 2 l

l
3E

a 2l 1.333a 2l

l l

2E 3E

l 1 1
l 3 2

E 2 3
E
6
l
0.333a 2 l
6E

0.333a 2
6E

a2
2E
0.333a 2 l

Substituting a2 value in equation (14),

a1
l

2E
2E
l l
a1

2E 2E
l l
a1

2E 2E
2l
a1
2E
l
a1
E

u a 1x a 2 x 2
We know that,

u
Substitute a1 and a2 values,

l
l 2
x
x
E
2E

x 2
lx

E
2
(16)

At x = l, u = u1 substitute in equation (16),

u1

2 l2
l
E
2

u1

l2

E 2

l 2
u u1 u 0
0
2E
We know that, extension of the bar,
[at x = l, u = u1, at x = 0, u = u0 = 0]

l2
2E
l2
u
2E

Extension or displacement of the bar,

(17)

From equation (16), we know that,

x 2
lx

E
2

du
2x

l l x
dx E
2
E

du

E l x l x
dx
E

Stress in the bar,

l x
Stress in the bar,
Exact solution:We know that, actual extension of the bar,

(18)

Pdx
Adx
l

AE 0 AE
0

x 2
xdx
E0
E 2
l

l2

E 2

l2
2E
(19)

From equation (6), (17) and (19), we know that total extension of the bar obtained is exact in both
the cases.