FINITE ELEMENT ANALYSIS

In this method, prototypes can be used. If we want to change the dimensions of the
prototype, we have to disassemble the entire prototype and reassemble it and then
testing should be carried out. It needs man power and materials. So, it is time
consuming and costly process.
2. Analytical Methods or Theoretical Analysis
In these methods, problems are expressed by mathematical differential equations. It
gives quick and closed form solutions. It is used only for simple geometries and
idealized support and loading conditions.
3. Numerical Methods
Analytical solutions can be obtained only for certain simplified situations. For
problems involving complex material properties and boundary conditions, the
engineer prefers numerical methods that gives approximate but acceptable solutions.
The following three methods are coming under numerical solutions.
(i) Functional Approximation.
(ii) Finite Difference Method (FDM).
(iii) Finite Element Method (FEM).
(i) Functional Approximation:
• The classical methods such as Rayleigh-Ritz methods (variational approach) and
Galerkin methods (weighted residual methods) are based on functional
approximation but vary in their procedure for evaluating the unknown parameters.
• Rayleigh-Ritz method is useful for solving complex structural problems,
encountered in finite element analysis.
• Weighted residual method is useful for solving non-structural problems.
(ii) Finite Difference Method (FDM):
• Finite difference method is useful for solving heat transfer fluid mechanics and
structural mechanics problems. It is a general method. It is applicable to any
phenomenon for which differential equation along with the boundary conditions
are available. It works well for two dimensional regions with boundaries parallel
to the coordinate axes.
 • The starting point in the finite difference method is that the differential equation
must be known before solving. After that, the region is subdivided into a
convenient number of divisions. The differential equation is applied successively
at the various points of the subdivided region, a set of simultaneous equations are
generated which upon solving lead to approximate solution to the problem. This
is the essence of finite difference method.
• This method is difficult to use when regions have curved or irregular boundaries
and it is difficult to write general computer programs.
(iii) Finite Element Method (FEM) or Finite Element Analysis (FEA):
• Finite element method is a numerical method for solving problems of Engineering
and mathematical physics.
• In this method, a body or a structure in which the analysis to be carried out is
subdivided into smaller elements of finite dimensions called finite elements. Then
the body is considered as an assemblage of these elements connected at a finite
number of joints called ‘Nodes’ or Nodal points. The properties of each type of
finite element is obtained, assembled together and solved as whole to get solution.
• In other words, in the finite element method, instead of solving the problem for
the entire body in one operation, we formulate the equations for each finite
element and combine them to obtain the solution of the whole body.
• Finite element method is used to solve physical problems involving complicated
geometrics, loading and material properties which cannot be solved by analytical
methods. This method is extensively used in the field of structural mechanics,
fluid mechanics, heat transfer, mass transfer, electric and magnetic fields
problems.

Figure 1.1 Shows the Finite Element Discretization

 • Based on applications, the finite element problems are classified as follows:
(i) Structural problems
(ii) Non-structural problems.
(i) Structural problems:
In structural problems, displacement at each nodal point is obtained. By using
these displacement solutions, stress and strain in each element can be calculated.
(ii) Non-Structural problems
In non-structural problems, temperature or fluid pressure at each nodal point is
obtained. By using these values, properties such as heat flow, fluid flow etc.., for each
element can be calculated.

1.2 HISTORICAL BACKGROUND OF FEM

• Basis ideas of the finite element analysis were developed by aircraft engineers in
the early 1940s. These were primarily the matrix methods of analysis.
• The modern development of the finite element method began in the year of 1945
in the field of structural engineering with the work by Hrennikoff.
• IN 1947 Levy introduced the flexibility of force method and in 1953 he suggested
stiffness method which could be a promising alternative for use in analyzing
statically redundant aircraft structures.
• By using energy principles, Argyris and Kelsey developed matrix structural
analysis methods in 1954. This development illustrated the important role that
energy principles would play in the finite element method.
• The term finite element was first introduced by Clough in 1960 in the plane stress
analysis and he used both triangular and rectangular elements in that analysis.
• Most of the finite element work upto early 1960s dealt with small strains and small
displacements, elastic material behavior and static loadings. In 1961, Turner
considered large deflection and thermal analysis problems. In 1962, Gallagher
introduced material no—linearities problems, whereas buckling problems were
initially treated by Gallagher and padlog in 1963. In 1968, Zinkiewicz extended
the method to visco elasticity problems.
 • Weighted residual methods was first introduced by Szabo and Lee in 1969 for
structural analysis and then by Zinkiewicz and parekh in 1970 for transient field
problems.
• During the decades of the 1960s, and 1970s, the finite element method was
extended to applications in shell bending, plate bending, heat transfer analysis,
fluid flow analysis and general three dimensional problems in structural analysis.
• From the early 1950s to present, enormous advances have been made in the
application of finite element method to solve complicated engineering problems.
It is curious to mote that the present day finite element method does not have its
root in one discipline. The mathematicians continue to put the finite element
method on sound theoretical ground whereas the engineers continue to find
interesting extensions in various branches of engineering. These concurrent
developments have made the finite element method as one of the most powerful
approximate methods.

1.3 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.
(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, displacement 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.
• We now present the steps along with explanations used in the finite element
method formulation.
1.6 Basic of Finite Element Method

Step 1: Discretization of structure

The art of subdividing a structure into a convenient number of smaller elements
is known as discretization.
Smaller elements are classified as follows:
(i) One dimensional elements.
(ii) Two dimensional elements
(iii) Three dimensional elements
(iv) Axisymmetric elements.
(i) One dimensional elements:
A bar and beam elements are considered as one dimensional elements. The simplest
line element also known as linear element has two nodes, one at each end as shown
in Figure 1.2.

Figure 1.2 Bar Element.

(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 corner nodes as shown in Figure 1.3.

Figure 1.3 Simple Two Dimensional Elements

(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 Figure 1.4.

Figure 1.4 Simple Three Dimensional Elements

(iv) Axisymmetric elements:
The axisymmetric element is developed by rotating a triangle or quadrilateral about
a fixed axis located in the plane of the element through 3600. It is shown in Figure
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 satisfies.

Figure 1.5
 Figure 1.6 (a)
{Maximum node number} – {Minimum node number} = Minimum
It is explained in the Figure 1.6(a) and (b).
Longer Side Numbering Process:

[Note: Number with circle denotes element. Number without circle denotes node]
Considering element (3),
Maximum node number = 10
Minimum node number = 3
Difference = 7 …(1.1)
Shorter Side Numbering Process:

Figure 1.6 (b)

 Considering the same element (3).

Maximum node number = 14

Minimum node number = 9
Difference = 5 …(1.2)
From equation (1.1) and (1.2), 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 a Displacement Function or Interpolation Function
• It involves choosing a displacement function within each element. Polynomial of
linear, quadratic and cubic form are frequently used as displacement functions
because they are simple to work within finite element formulation.

Figure 1.7 Polynomial Approximation in One Dimension

• 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.
Figure 1.7 shows the polynomial approximation in one dimension.
 Let us consider  (x) is a fixed variable.
Case (i): Linear Polynomial:
One dimensional problem  (x) = 𝑎 + 𝑎 𝑥
Two dimensional problem  (x, y) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑦
Three dimensional problem  (x, y, z) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑧
Case (ii): Quadratic Polynomial:
One dimensional problem  (x) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥
Two dimensional problem  (x, y) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑥𝑦
Three dimensional problem  (x, y,z) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑧 + 𝑎 𝑥 +
𝑎 𝑦 + 𝑎 𝑧 + 𝑎 𝑥𝑦 + 𝑎 𝑦𝑧 + 𝑎 𝑥𝑧
Step 4: Define the Material Behavior by using Strain- Displacement and Stress –
Strain Relationship
• 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
𝑒= … (1.3)
𝑑𝑥
Where, u → Displacement field variable along x direction.
e → Strain.
The stress-strain relationship is given by,
=Ee …(1.4)
Where,  → Stress in x direction.
E → Modulus of elasticity or Young’s modulus.
Step 5 : Derivation of element stiffness matrix and equations:
The finite element equation is in matrix form as,
 𝐹 𝑘 𝑘 𝑘 ⋯ 𝑘 𝑢
⎧𝐹 ⎫ ⎡ ⎤
𝑘 ⋯ 𝑘 ⎧ 𝑢 ⎫
⎢𝑘
⎪ ⎪ 𝑘 ⎥ ⎪𝑢 ⎪
𝐹 = ⎢𝑘 𝑘 𝑘 ⋯ 𝑘 ⎥
⎨⋮ ⎬ ⎢ ⋮ ⋮ ⎥⎨ ⋮⎬
⎪ ⎪ ⎪ ⎪
⎩𝐹 ⎭ ⎣𝑘 ⋯ ⋯ ⋯ 𝑘 ⎦ ⎩𝑢 ⎭
In compact matrix from as,
{ 𝐹 } = [𝑘 ]{ 𝑢 }

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 one
dimensional elements.
(ii) Variational Method : This method is most easily adaptable to the determination of
element equations for complicated elements (i.e., element having large number of
degrees of freedom) like axisymmetric stress element, plate bending element and
two or three dimensional solid stress element.
(iii) Weighted Residual Method: This method is (Galerkin’s method) useful for
developing the element equations in thermal analysis problems. They are
especially useful when 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 method of superposition i.e., direction stiffness method. The final assembled or global
equation which is in the form of
{ 𝐹} = [𝐾]{𝑢} … (1.5)

Where, { F } → Global force vector.

[ F ] → Global stiffness matrix.

{ F } → Global displacement vector.

Step 7: Applying boundary conditions:
From equation (1.5), we know that, global stiffness matrix [ K ] is a singular
matrix because its determination is equal to zero. In order to remove this singularity
problem, certain boundary conditions are applied so that the structure remains in place
instead of moving as a rigid body. The global equation (1.5) to be modified to account
for the boundary conditions of the problem.
Step 8: Solution for the unknown displacements:
A set of simultaneous algebraic equations formed in step 6 can be written in
expanded matrix form as follows:
𝑘 𝑘 𝑘 ⋯ 𝑘 𝑢
𝐹 ⎡ ⎤⎧ ⎫
𝑘 ⋯ 𝑘 𝑢
⎢𝑘
⎧𝐹 ⎫ 𝑘
⎪ ⎪ ⎥ ⎪𝑢 ⎪
𝐹 ⎢𝑘 𝑘 𝑘 ⋯ 𝑘 ⎥
= ⎢𝑘 𝑘 𝑘 ⋯ 𝑘 ⎥ 𝑢
⎨⋮ ⎬ ⎢⋯ ⋯ ⋯ ⋯ ⋯ ⎥⎨ ⋮ ⎬
⎪⋮ ⎪
⎩𝐹 ⎭ ⎢⋯ ⋯ ⋯ ⋯ ⋯ ⎥⎪ ⋮ ⎪
⎣𝑘 𝑘 𝑘 ⋯ 𝑘 ⎦ ⎩𝑢 ⎭

These equation can be solved and unknown displacements { u } are calculated by

using Gaussian 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
the solution of displacement vector {u}, stress and strain value can be calculated.
In case of one dimensional deformation, the strain-displacement relationship is
given by,
𝑑𝑢
𝑆𝑡𝑟𝑎𝑖𝑛, 𝑒 = [From equation (1.3)]
𝑑𝑥
𝑢 −𝑢
=
𝑥 −𝑥
Where, 𝑢 and 𝑢 are displacement at node 1 and 2.
𝑥 − 𝑥 = 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.
Post processor computer programs help the user to interpret the results by displaying them
in graphical form.
Steps 1 to 10 are summarized as follows:
1.4 DISCRETIZATION
1.4.1 Introduction
In this chapter, we are going to learn about discretization, node, assembly, system
etc. To make this easier to understand, let us compare these words with the parts over
human body. Apart from flesh, our body consists of bones. They are hands, legs gingers,
thigh bones, etc. these parts are connected together at different places, so that when
movement takes place, we do not feel any pain. Nature has assembled in such a way that
every human being is able to sustain certain amount of load without experiencing stain.
Similarly any structure like an automobile, ship, Aeroplane, etc., consists of
several components assembled together.
Now let us study about ‘Element’. The characteristics of an element are as follows:
(i) It is a small portion of a system
(ii) It has definite shape.
(iii) It should have minimum two nodes.
(iv) Nodes are placed where connection is made to another element.
(v) Loads act only at the nodes.
Examples:

Figure 1.8 Truss element

 Figure 1.9

1.4.2 Discretization
The art of subdividing a structure into a convenient number of smaller components
is known as Discretization. These smaller components are then put together. The process
of uniting the various elements together is called Assemblage. The assemblage of such
elements then represents the original body.
Discretization can be classified as follows:
(i) Natural.
(ii) Artificial (continuum).

1.4.3 Natural Discretization

In structural analysis, a truss is considered as a natural system. The various
members of the truss constitute the elements. These elements are connected at various
joints known as nodes.
Nodal Points: Each kind of finite element has a specific structural shape and is
interconnected with the adjacent elements by nodal points or nodes.

Figure 1.10 Natural discretization of truss

Nodal forces: The forces that act at each nodal point are called nodal forces.
Degrees of freedom: When the force or reaction act at nodal point, node is subjected to
deformation. This deformation includes displacements, rotations, and/or strains. These
are collectively known as degrees of freedom or simply we can say nodal displacement
is called degrees of freedom.
In Fig.1.10, the truss consists of 9 elements and 6 nodes. There are four freely
moving and two extreme constrained nodes. The truss is a natural system as there is no
possibility either to increase or decrease the number of elements and the nodes.

1.4.4 Artificial Discretization (Continuum)

Continuum is generally considered to be a single mass of material as found in a
forging, concrete dam, deep beam, plate and so on.
Unlike the truss element which is physically present in the truss, in a continuum, the
following three elements exist only in our imagination.
1. Triangular element.
2. Rectangular element.
3. Quadrilateral element.
They are shown in Figure 1.11.

Figure 1.11.
Discretization using triangular element is shown in Figure 1.12. & Figure 1.13.
 Finite Element Analysis 1.17

Figure 1.12. Discretization using triangular elements

Figure 1.13.
Fig 1.14 shows a deep beam. In Fig.1.15, it is shown how it is discretized using
simple rectangular elements.
1.18 Basic of Finite Element Method

Figure 1.14. Deep beam Figure 1.15. Deep beam discretization using
Rectangular elements
Fig 1.16 (a) shows a planar continuum subjected to uniformly distributed load on
the top.
Fig.1.16 (b), the continuum id discretized into eight triangular elements. The
discretization shown is only one way. We can subdivide the continuum into triangular
elements in a number of ways. Alternative way is shown in fig .1.17.

Figure 1.16. Figure 1.17. Alternative way of discretization

1.4.5 Discretization Process

The following points to be considered while analyzing the discretization process.
(i) Type of elements:
• The type of elements to be used will be evident from the physical problem.
 Finite Element Analysis 1.19

• The structure, shown in Fig 1.18 is discretized by using line of bar elements.

Figure 1.18.
• The finite element idealization can be done by using three dimensional rectangular
element in stress analysis of short beam problem which is shown in Fig.1.19.

Figure 1.19. (a) Short beam (b) Discretization using three-dimensional elements
• A thin wall sheet shown in Fig.1.20 (a), which can be discretized by several types
f elements as shown in Fig. 1.20(b).
1.20 Basic of Finite Element Method

Figure 1.20. (a) Original shell (b) Discretization using different types of elements

• The choice of the element to be used for discretization depends upon the following
factors.
(i) Number of degrees of freedom needed.
(ii) Expected accuracy.
(iii) Necessary equations required.

• However, in certain problems, the given structure cannot be discretized by using

only one type of elements. In such cases, we can use two or more types of elements
for discretization.
Example: Air craft wing.
(ii) Size of elements:

• The size of elements influences the convergence of the solution of the problem
directly. So, it should be chosen with more care.
 Finite Element Analysis 1.21

• If the size of the element is small, the final solution is more accurate. But the
computational time for the smaller size element is more when compared to larger
size element.
• Another characteristic related to the size of elements that affects the finite element
problem solution is the “ Aspect ratio” of the elements.
• Aspect ratio is defined as the ratio of the largest dimension of the element to the
smallest dimension. The conclusion of many researchers is that the aspect ratio
should be close to unity as possible. For a two dimensional rectangular element,
the aspect ratio is conveniently defined as length to breadth ratio. Aspect ratio
closer to unity yields better results.
(iii) Location of nodes:
• If the structure has no abrupt changes in geometric, load, boundary conditions and
material properties, the structure can be divided into equal subdivisions. So, the
spacing of the nodes are uniform.
• If there are any discontinuities in geometric, load, boundary conditions and
material properties of the structure, nodes should be introduced at these
discontinuities as shown in the following figures.

Figure 1.21. Geometric discontinuities

Figure 1.22. Discontinuity in loading

 1.22 Basic of Finite Element Method

Figure 1.23. Discontinuity of boundary conditions

Figure 1.24. Material discontinuity

(iv) Number of elements:
The number of elements to be selected for discretization depends upon the
following factors:
1. Accuracy desired.
2. Size of the elements.
3. Number of degrees of freedom involved.
• If the number of elements in the structure is increased, the final solution of the
problem is expected to be more accurate. But the use of large number of elements
involves a large number of degrees of freedom, it leads the storage problem in the
available computer memory.

1.5 INITIAL VALUE AND BOUNDARY VALUE PROBLEMS

A differential equation along with subsidiary conditions on the unknown function
and its c=derivatives, all given at the same value of the independent variable, constitutes an
initial value problem. The subsidiary conditions are initial conditions. If the subsidiary
 Finite Element Analysis 1.23

conditions are given at more than one value of the independent variable, the problem is a
boundary value problem and the conditions are boundary conditions.
For Example:
• The problem 𝑦 + 2 𝑦 = 𝑒 ; 𝑦(𝜋) = 1, 𝑦 (𝜋) = 2 is an initial-value problem,
because both the subsidiary conditions are given at x = .
• The problem 𝑦 + 2 𝑦 = 𝑒 ; 𝑦(0) = 1, 𝑦(1) = 1 is an boundary-value
problem, because the two subsidiary conditions are given at the different values x
= 0 and x = 1.

1.6 PROBLEM BASED ON INITIAL VALUE PROBLEM

Example 1.1 Find the solution of the initial value problem.
y’ + y =0; y(3) = 2
Given : Differential equation, y’ + y = 0
Boundary condition at y(3) = 2
Solution: Differential equation, y’ + y = 0 …(1)
Boundary condition at y(3) = 2
…(2)
 𝑥 = 3, 𝑦 = 2
Using Auxiliary equation,

+1=0 ∵𝑦 = =

 = -1 …(3)
We know that, complementary function or characteristic function,
𝑦(𝑥) = 𝑐 𝑒 … (4)
Applying the boundary condition (2) in equation (4),
𝑦(3) = 𝑐 𝑒
2=𝑐 𝑒 [∵ 𝑦(3) = 2]
𝑐 = 2𝑒
By substituting equation (5) in equation (4)
𝑦(𝑥) = 2𝑒 𝑒
1.24 Basic of Finite Element Method

As the solution of the initial-value problem.

Result: y(x) = 2e3 e –x, as the solution of the initial-value problem.

Example 1.2

Find a solution of the initial-value problem + − 2 𝑦 = 0, Boundary

conditions y(0) =2, y’(0) =5.
Given:
Differential equation,
𝑑 𝑦 𝑑𝑦
+ − 2𝑦 = 0
𝑑𝑥 𝑑𝑥
Boundary conditions are 𝑦(0) = 2 , 𝑦 (0) = 5
Solution: differential equation,
𝑑 𝑦 𝑑𝑦
+ − 2𝑦 = 0 . . . (1)
𝑑𝑥 𝑑𝑥
Boundary conditions are 𝑦(0) = 2 , 𝑦 (0) = 5 ...(2)
Using auxiliary equation, 2 +  -2 = 0
( - 1)( + 2) = 0
 1 = 1, 2 = -2 …(3)
We know that, complementary function, 𝑦(𝑥) = 𝐴 𝑒  1𝑥 + 𝐵𝑒 
Put  = 1,  = −2 in the above equation,
𝑦(𝑥) = 𝐴 𝑒 + 𝐵𝑒 . . . (4)
( )
𝑦 = 𝐴 𝑒 − 2𝐵𝑒 … (5)
Applying the boundary conditions (2) in equation (4) and (5),we get
y(0) = 2  x = 0, y = 2
𝑦(0) = 𝐴𝑒 + 𝐵𝑒
2 = A+ B
A+B=2 …(6)
 Finite Element Analysis 1.25

Similarly, y'(0) = x = 0, y = 5
( )
𝑦 = 𝐴𝑒 − 2 𝐵𝑒
5=A-2B
A–2B=5 …(7)
Solving the equation (6) and (7),
A+B =2
A -2 B = 5
(-) (+) (-)
3 B = -3
B = -1
Substitute the B = -1 value in equation (6),
A -1 = 2
A=2+1
A=3
By substituting A and B values in equation (4),
𝑦(𝑥) = 3 𝑒 − 𝑒
Result: General solution 𝑦(𝑥) = 3 𝑒 − 𝑒

1.7 PROBLMES SOLVED ON BOUNDARY-VALUE PROBLEM

EXAMPLE 1.3
Find a solution of a boundary-value problem 𝑦" + 𝑦 = 0 𝑤𝑖𝑡ℎ 𝑦 (0) =
0 𝑎𝑛𝑑 𝑦(𝜋/6) = 4.
Given: Differential equation, y” + y = 0
Boundary conditions are y(0) = 0, y(/6) = 4
Solution: Differential equation, y” + y = 0 …(1)
Boundary conditions are y(0) = 0, y (/6) = 4
Using auxiliary equation, 2 + 1 = 0
1.26 Basic of Finite Element Method

𝑑 𝑑
∵𝑦 = = ; 𝑦" = =
𝑑𝑥 𝑑𝑥

 = −1

 = √−1

 = ±𝑖

 = ∝ ±𝑖𝛽 [𝐻𝑒𝑟𝑒, ∝ = 0, 𝛽 = 1]
We know that, complementary functions are,
𝑦(𝑥 ) = 𝑒 ∝ [𝑐 cos 𝛽𝑥 + 𝑐 sin 𝛽𝑥]

𝑦(𝑥 ) = 𝑒 [𝑐 cos 𝑥 + 𝑐 sin 𝑥]

𝑦(𝑥 ) = 𝑐 cos 𝑥 + 𝑐 sin 𝑥 … (2)

Applying boundary conditions in equation (2),
𝑦(0) = 0  𝑦(0) = 𝑐 cos(0) + 𝑐 sin(0)
0=𝑐 +0 [∵ cos 0 = 1, sin 0 = 0]
c1 = 0
similarly,
𝑦(𝜋⁄6) = 4  𝑦 (𝜋/6) = 𝑐 cos(𝜋/6) + 𝑐 sin(𝜋/6)

√3 1
4 =𝑐 +𝑐
2 2
1
𝑝𝑢𝑡 𝑐 = 0, 𝑐 =4
2
𝑐 =8
Substitute the 𝑐 𝑎𝑛𝑑 𝑐 values in equation (2)
y(x) = 0 + 8 sin x
y(x) = 8 sin x
Result : General equation, y(x) = 8 sin x
 Finite Element Analysis 1.27

Example 1.4
Find a solution of a boundary-value problem 𝑦" + 4𝑦 = 0 𝑤𝑖𝑡ℎ 𝑦(𝜋/8) =
0, 𝑦(𝜋/6) = 1.
Given : Differential equation, y” + 4y = 0 …(1)
Boundary conditions are y(/8) = 0 and y(/6) = 1 …(2)
Solution: Differential equation, y” + 4y = 0 …(1)
Boundary conditions are y(/8) = 0 and y(/6) = 1
Using auxiliary equation, 2 + 4 = 0
 = −4
 = √−4
 = ±2𝑖
 = ∝ ±𝑖𝛽 [∵ 𝐻𝑒𝑟𝑒, ∝ = 0, 𝛽 = 2]
We know that, complementary functions,
𝑦(𝑥 ) = 𝑒 ∝ [𝑐 cos 𝛽𝑥 + 𝑐 sin 𝛽𝑥]
𝑦(𝑥 ) = 𝑒 [𝑐 cos2 𝑥 + 𝑐 sin 2𝑥]
𝑦(𝑥 ) = 𝑐 cos 2𝑥 + 𝑐 sin2 𝑥 … (3)
Applying first boundary conditions in equation (3),
y(/8) = 0

 𝑦(/8) = 𝑐 cos 2(/8) + 𝑐 cos 2(/8)

𝑦(/8) = 𝑐 cos (/4) + 𝑐 𝑠𝑖𝑛(/4)

1 1
0=𝑐 × √2 + 𝑐 × √2
2 2

√2 √2
𝑐 +𝑐 =0
2 2
1 1
(𝑜𝑟) 𝑐 +𝑐 =0
√2 √2
1.28 Basic of Finite Element Method

Applying second boundary conditions in equation (3),

y(/6) = 1
 𝑦(/6) = 𝑐 cos 2(/6) + 𝑐 sin 2(/6)
 
1 = 𝑐 cos + 𝑐 sin
3 3
√3 1
𝑐 +𝑐 =1
2 2
Solving the equation (4) and (5),
Equation (4) 
1 1
𝑐 +𝑐 =0
√2 √2
Equation (5) 

√3 1
𝑐 +𝑐 =1
2 2
Get, 𝑐 = 2.732
Substitute the 𝑐 value in equation (5)

√3 1
2.732 +𝑐 =1
2 2
𝑐 = −2.732
From these c and c value, substitute the equation (3)
𝑦(𝑥) = 2.732 cos 2 𝑥 + (−2.732) sin 2 𝑥
𝑦(𝑥) = 2.732 [cos 2 𝑥 − sin 2 𝑥]
Result: General equation, 𝑦(𝑥) = 2.732 [cos 2 𝑥 − sin 2 𝑥]

1.8 EIGEN VALUE PROBLEM [BOUNDARY VALUE PROBLEM]

When applied to the boundary- value problem, has the form
𝑦" + 𝑃(𝑥 + )𝑦′ + 𝑄(𝑥 + )𝑦 = 0
 Finite Element Analysis 1.29

Non-trivial solutions may exit for certain values of . Those values of  for which
no-trivial solutions do exists are called eigen values, the corresponding non-trivial
solutions are called eigen functions.
For example: for the axial vibration of bar, to find u (x) and  that satisfies the
partial differentiation equation and boundary conditions are,
𝐴 𝑦 " +  𝑦 ′ = 0, 𝑓𝑜𝑟 0 < 𝑥 < 𝐿
𝑦(0) = 0, 𝑦 (𝐿) = 0

1.9 PROBLMES SOLVED ON EIGEN VALUE PROBLEM

[BOUNDARY VALUE PROBLEM]
Example 1.5

Find a eigen values and eigen function y” + 4 y +  y = 0; with boundary

conditions are y(0) = 0, y(1) + y (1) = 0

Given: Differential equation,

𝑦" − 4 𝑦′ + 4 𝑦 = 0
Boundary conditions are, y(0) = 0
y(1) + y’(1) = 0
Solution: Given differential equation,

𝑦" − 4 𝑦′ + 4 𝑦 = 0 … (1)
Boundary conditions are, y(0) = 0
y(1) + y’(1) = 0 …(2)
The auxiliary equation is,

𝑚 −4𝑚+4 = 0 … (3)

 = (𝑚 − 2  )(𝑚 − 2 ) = 0

 = (𝑚 − 2  ) = 0 (𝑚 − 2 ) = 0

 = 𝑚 = 2 ,𝑚 = 2 
1.30 Basic of Finite Element Method

We know that, complementary function is,

𝑦(𝑥) = 𝑐 𝑒 +𝑐 𝑥𝑒 …(4)

𝑦(𝑥) = 𝑐 𝑒  +𝑐 𝑥𝑒 

Differentiate with respect to “x” in equation (4),

𝑦 (𝑥) = 2  𝑐 𝑒  +𝑐 [𝑥2𝑒  +𝑒  ]
𝑦 (𝑥) = 2  𝑐 𝑒  +𝑐 𝑥2𝑒  +𝑒  … (5)
Applying first boundary condition in equation (5),
y(0) = 0, x = 0, y = 0
 𝑦(0) = 𝑐 𝑒 + 𝑐 (0)𝑒
𝑐 =0
Applying second boundary conditions in equation (5),
𝑦(1) + 𝑦 (1) = 0
We get, 𝑦(1) = 𝑐 𝑒  +𝑐 𝑒 

𝑦 ( )
=2𝑐 𝑒  + 𝑐 (2  𝑒  + 𝑒 )

We know that,
 𝑐 𝑒  +𝑐 𝑒  +2𝑐 𝑒  + 𝑐 (2  𝑒  + 𝑒 ) =0
𝑐 (1 + 2 ) + 𝑐 (2 + 2 ) = 0
If now follows that C1 = 0 and either C2 = 0 (or)  = -1
The choice of C2 = 0.
The result in the trivial solution y = 0.
The choice of  = -1.
The result in the non-trivial solution, i.e., 𝑦 = 𝑐 𝑒 , where 𝑐  arbitrary
Thus the boundary value problem has eigen value  = -1 and the eigen function
𝑦=𝑐 𝑒
Result: Eigen value and eigen functions, 𝑦 = 𝑐 𝑒
 Finite Element Analysis 1.31

Example 1.6
Find a eigen values and eigen function 𝑦” + 4  𝑦 + 4  𝑦 = 0; with boundary
conditions are 𝑦 ( ) = 0, 𝑦(2) + 2𝑦 (2) = 0.
Given : Differential equation,
𝑦" − 4 𝑦′ + 4 𝑦 = 0
Boundary conditions are, 𝑦 (1) = 0, 𝑦(2) + 2𝑦 (2) = 0.
Solution: Differential equation,
𝑦” + 4  𝑦 + 4  𝑦
=0 … (1)
Boundary conditions are, y′(1) = 0
y (2) +2 y′ (2) = 0 …(2)
The auxiliary equation is, 𝑚 − 4  𝑚 + 4  = 0 … (3)
(𝑚 − 2  )(𝑚 − 2 ) = 0
𝑚 = 2  ,𝑚 = 2 
We know that, complementary functions are,
𝑦(𝑥) = 𝑐 𝑒  +𝑐 𝑥𝑒  … (4)
Differentiate with respect to “x” in equation (4),
𝑦 (𝑥) = 2  𝑐 𝑒  +𝑐 𝑥2𝑒  +𝑒  … (5)
Applying first boundary condition in equation (5), we get,
y'(1) = 0
 𝑦 (1) = 2  𝑐 𝑒  +𝑐 2𝑒  +𝑒  =0
 2𝑐 𝑒  + 𝑐 [2  𝑒  +𝑒 ] =0
 2  𝑐 + 𝑐 [2  + 1] = 0 … (6)
Applying second boundary conditions in equation (5) and (4),
𝑦(2) + 2 𝑦 (2) = 0
 𝑦(2) = 𝑐 𝑒  + 2𝑐 𝑒 
1.32 Basic of Finite Element Method

 𝑦′(2) = 2  𝑐 𝑒  + 𝑐 (4  𝑒  + 𝑒 )

Adding both equations,

 𝑐 𝑒  + 2𝑐 𝑒  +2 𝑐 𝑒  +𝑐 4𝑒  +𝑒  =0

 𝑐 𝑒  + 2𝑐 𝑒  + 4 𝑐 𝑒  +2𝑐 4𝑒  +𝑒  =0

 𝑐 𝑒  + 2𝑐 𝑒  +4𝑐 𝑒  +8𝑐 𝑒  +2𝑐 𝑒  =0

 𝑐 + 2𝑐 + 4  𝑐 + 8  𝑐 + 2 𝑐 = 0

 (1 + 4 )𝑐 + (8  + 4)𝑐 = 0 … (7)

Solving equations (6) and (7),
 2𝑐 +𝑐 (2 +1) =0
 (1 + 4 ) 𝑐 + (8  + 4) 𝑐 = 0
Set determination is equal to zero,
2 1+2
=0
1+4 4+8
 2  (4 + 8 ) − (1 + 2 )(1 + 4 ) = 0
 8  (1 + 2 ) − (1 + 2 )(1 + 4 ) = 0
 (1 + 2 ) − [8  − (1 + 4 )] = 0
 (1 + 2 ) − (1 + 4 ) = 0
 (1 + 2 ) = 0 𝑎𝑛𝑑 (4  − 1) = 0
1 1
 = − 𝑎𝑛𝑑  =
2 4
When,  = − 𝑎𝑛𝑑  = the result has non-trivial solution.

It follows that eigen values are  = − 𝑎𝑛𝑑  = and the corresponding eigen
functions are,
𝑦 =𝑐 𝑥𝑒 and
/
𝑦 = 𝑐 (−3 + 𝑥) 𝑒
 Finite Element Analysis 1.33

Result: Eigen value,  = − 𝑎𝑛𝑑  =

Eigen functions, 𝑦 =𝑐 𝑥𝑒 𝑎𝑛𝑑

/
𝑦 = 𝑐 (−3 + 𝑥) 𝑒

1.10 SOLUTION OF EIGEN VALUE PROBLEMS [MATRIX

APPROACH]
There are three methods to solve eigen value problems, They are,
1. Determinant methods
2. Transformation methods
3. Vector iteration methods

1.10.1 Determinant Methods

These methods are primarily based on the equations,
{[𝐾] −  [𝑚]} { 𝑢} = 0
If the eigen vector is non-trivial, the required condition is,
det|[𝐾] −  [𝑚]| = 0
 |[𝐾] −  [𝑚]| = 0 … (1.6)
Trial value of  is taken and the determinant | [K] -  [m] | = 0 is computer. The
curve is drawn by taking several trial values.

Figure 1.25. Determinant-based Method

1.34 Basic of Finite Element Method

Due to heavy computational cost and several iterations are required to determine
all the eigen values, the determinant based methods are not implemented in practice.

1.10.2 Transformation Methods

This method is used to transform the eigen value problems,
Let, [𝐾 ]{𝑢} =  { 𝑢} … (1.7)
Transform [ K ] into a diagonal matrix by using a series of matrix transformations,
[ 𝐾 ] = [ 𝑇] [ 𝐾 ] [ 𝑇 ] … (1.8)
Where, [ T ] is the transformation matrix, which is usually an orthogonal matrix.
𝑖. 𝑒., [ 𝑇] = [ 𝑇 ]
When we transform [ K ] completely into diagonal matrix, then the elements on the
diagonals are considered as eigen values,
 0
[ 𝑇] [ 𝐾 ][ 𝑇 ] =  … (1.9)
0 
Where,  ,  and  are eigen values.

1.10.3 Vector iteration Methods

• Vector iteration methods are normally available in many commercial finite
element software packages.
• In this method, trial eigen vector is assumed and repeated matrix manipulation is
performed to compute the desired eigen vector.

1.11 SOLVED PROBLEMS ON EIGEN VALUES [MATRIX

APPROACH]
4 −20 −10
Example 1.7 Find the eigen values of −2 10 4
6 −30 −13
Solution:
Step 1: To find characteristic equation,
 Finite Element Analysis 1.35
4 −20 −10
Let the given matrix be 𝐴 = −2 10 4
6 −30 −13
The characteristic equation is
 −𝑎  +𝑎  −𝑎 =0 … (1)
Where, 𝑎 = Sum of leading diagonal elements
= 4 + 10 = 13
𝑎 =1 … (2)
𝑎 = Sum of minors of the leading diagonal elements
10 4 4 −10 4 −20
= + +
−30 −13 6 −13 −2 10
= -130 + 120 – 52 + 60 + 40 – 40
𝑎 = −2 … (3)
4 −20 −10
𝑎 = |𝐴| = −2 10 4
6 −30 −13
= 4[-130+ 120] +20 [26-24] – 10[ 60 -60]
= -40 + 40 + 0
𝑎 =0 … (4)
Substitute the 𝑎 , 𝑎 𝑎𝑛𝑑 𝑎 values in equation (1),
 − −2=0
Step 2: To find eigen values:
 − −2=0
( −  − 2) = 0
𝑊ℎ𝑒𝑛  = 0,  − −2=0
1 ± √1 + 8 1± 3
 = =
2 2
= 2 or -1
1.36 Basic of Finite Element Method

 Eigen values are  = 0, -1, 2

Result: Eigen values,  = 0,  = −1,  = 2.

1.12 WEIGHTED RESUDUAL METHODS

1.12.1 Introduction
The method of weighted residuals is a powerful approximate procedure applicable
to several problems. For structural problems, potential energy functional can be easily
formed, so, Rayleigh-Ritz method is used. On the other hand, for non-structural
problems, the differential equation of the phenomenon can be easily formulated. For such
type of problems, the method of weighted residuals becomes very useful. There are many
types of weighted residuals, of them four are very popular. They are.
(i) Point collocation method.
(ii) Subdomain collocation method.
(iii) Least squares method.
(iv) Galerkin’s method.
Among these four methods, the Galerkin approach has the widest choice and is
used in finite element analysis.

1.12.2. General Procedure

Our interest is to find y, which is the solution for the differential equation. If it is
not possible to find a solution, we assume an approximate function for y. When we
substitute the approximate solution in the differential equation, we can get residual and
that residual can be expresses as,
𝑅(𝑥 ; 𝑎 , 𝑎 , 𝑎 ) = 0
Where 𝑎 , 𝑎 are unknown parameters present in assumed trial function.
The assumed trial function can be expressed as follows:
𝑦 = 𝑓(𝑥; 𝑎 , 𝑎 , 𝑎 , … … , 𝑎 )
Trial function y must exactly satisfy the boundary conditions.
The method of weighted residuals needs the parameters 𝑎 , 𝑎 , 𝑎 , … … , 𝑎 to be
determined by satisfying the following equation.
 Finite Element Analysis 1.37

𝑤 𝑅(𝑥; 𝑎 , 𝑎 , 𝑎 , … . , 𝑎 )𝑑𝑥 = 0 … (1.10)

Where, 𝑤 is a function of x and known as weighting function.

Dis a domain, R is a residual.

1.12.3. Point Collocation Method

In the collocation method, also called point collocation, residuals are set to zero
at n different locations Xi, and the weighting function wi is denoted as (x- xi)

 𝑤 =  (𝑥 − 𝑥 )

Substituting 𝑤 value in equation (1.10),

 (𝑥 − 𝑥 ) 𝑅(𝑥; 𝑎 , 𝑎 , 𝑎 , … . , 𝑎 )𝑑𝑥 = 0 … (1.11)

The 𝑥 ′𝑠 are referred to as collocation points and are selected by the discretion of
the analyst.

In equation (1.11), term ∫  (𝑥 − 𝑥 ) = 1

𝑆𝑜, 𝑅(𝑥; 𝑎 , 𝑎 , 𝑎 , … . , 𝑎 ) = 0

1.12.4 Subdomain Collocation Method

In this method, the weighting functions (𝑤 ) are chosen to be unity over a portion
of the domain and zero elsewhere. It is given as follows:
1 for 𝑥 in 𝐷
𝑤 =
0 for 𝑥 not in 𝐷
1 for 𝑥 in 𝐷
𝑤 =
0 for 𝑥 not in 𝐷

⋮ ⋮ ⋮
1 for 𝑥 in 𝐷
𝑤 =
0 for 𝑥 not in 𝐷
Where D is a domain.
1.38 Basic of Finite Element Method

1.12.5 Least Squares Method

In this method, the integral of the weighted square of the residual over the domain is
required to be minimum.

𝑖. 𝑒. , 𝐼= [𝑅(𝑥; 𝑎 , 𝑎 , 𝑎 , … … . 𝑎 )] 𝑑𝑥 = minimum

𝑤ℎ𝑒𝑟𝑒, 𝐼 = 𝑓(𝑎 , 𝑎 , … … 𝑎 )
The requirement is
𝜕𝐼
= 0, 𝑖 = 1, 2, 3, … . . 𝑛
𝜕𝑎

1.12.6 Galerkin’s Method

In this method, the trial function, 𝑁 (𝑥), itself is considered as the weighting
functions; that is,
𝑤 = 𝑁 (𝑥)
Substitute 𝑤 value in equation (1.10),

 𝑁 (𝑥)𝑅(𝑥; 𝑎 , 𝑎 , … . . 𝑎 )𝑑𝑥 = 0 … (1.12)

i = 1, 2, 3, ….. n

1.12.7 Solved Problems – Weighted Residual Method

Example 1.8
The following differential equation is available for a physical phenomenon
AE + q = 0 with the boundary conditions

y(0) = 0
𝑑𝑦
=0
𝑑𝑥
Find the value of f(x) using the weighted residual method.

Given: Differential equation, 𝐴 𝐸 +𝑞 = 0

 Finite Element Analysis 1.39

Boundary conditions are y(0) = 0

𝑑𝑦
=0
𝑑𝑥
To find: f(x)
Solution: Assume a trial solution.
𝐿𝑒𝑡 𝑦(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 … . (1)
Apply first boundary condition, i.e., substitute x =0 and y =0 in equation (1).
(1) 0 = 𝑎 + 0 + 0
𝑎 =0
Apply second boundary condition,
𝑦(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥
𝑑𝑦
= 𝑎 +2𝑎 𝑥
𝑑𝑥
𝑑𝑦
𝐴𝑡 𝑥 = 𝐿, =0
𝑑𝑥
 0=𝑎 +2𝑎 𝐿
 𝑎 = −2 𝑎 𝐿
Solving 𝑎 𝑎𝑛𝑑 𝑎 value in equation (1),
(1)  𝑦(𝑥) = −2 𝑎 𝑥 𝐿 + 𝑎 𝑥
𝑦(𝑥) = 𝑎 [𝑥 − 2 𝑥𝐿] … . (2)
𝑑𝑦
 = 𝑎 (2𝑥 − 2 𝐿)
𝑑𝑥
𝑑 𝑦
= 2𝑎
𝑑𝑥
We know that, residual,
𝑑 𝑦
𝑅 =𝐴𝐸 +𝑞 = 0
𝑑𝑥
 𝐴 𝐸 (2 𝑎 ) + 𝑞 = 0
1.40 Basic of Finite Element Method

 𝐴 𝐸 2𝑎 = −𝑞
−𝑞
 𝑎 =
2𝐴𝐸
Substitute 𝑎 value in equation (2),
−𝑞
 𝑦(𝑥) = [𝑥 − 2 𝑥 𝐿]
2𝐴𝐸
𝑞
𝑦(𝑥) = [2 𝑥 𝐿 − 𝑥 ]
2𝐴𝐸
Result: Final solution,
𝑞
𝑦(𝑥) = [2 𝑥 𝐿 − 𝑥 ]
2𝐴𝐸

Example 1.9
The governing differential equation for the fully developed lamina flow is given
by  + 𝜌 𝑔 𝑐𝑜𝑠 𝜃 = 0.

𝑑𝑢
if boundary conditions are = 0, 𝑢(𝐿) = 0,
𝑑𝑥
Find the velocity distribution, u(x).
Solution: Differential equation,
𝑑 𝑦
 + 𝜌 𝑔 𝑐𝑜𝑠 𝜃 = 0
𝑑𝑥
𝑑𝑢
Boundary conditions are =0
𝑑𝑥
u(L) = 0
To find: Velocity distribution, u(x).
Solution: Assume a trial function.
𝐿𝑒𝑡 𝑢(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 … … (1)
Apply first boundary condition,
𝑑𝑢
𝑖. 𝑒., = 0 𝑎𝑡 𝑥 = 0.
𝑑𝑥
 Finite Element Analysis 1.41
𝑑𝑢
 =𝑎 +2𝑎 𝑥
𝑑𝑥
𝑑𝑢
𝐴𝑇 𝑥 = 0 , =0
𝑑𝑥
 𝑎 =0
Apply second boundary condition,
i.e., at x = L, u(x) = 0

 𝑢(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥

 0= 𝑎 +𝑎 𝐿+𝑎 𝐿

Substitute 𝑎 = 0

 0= 𝑎 +𝑎 𝐿

 𝑎 = −𝑎 𝐿

Solving 𝑎 𝑎𝑛𝑑 𝑎 value in equation (1),

𝑢(𝑥) = −𝑎 𝐿 + 0 + 𝑎 𝑥

= 𝑎 [𝑥 − 𝐿 ]

𝑢(𝑥) = 𝑎 [𝑥 − 𝐿 ] … (2)
𝑑𝑢
 = 𝑎 [2 𝑥]
𝑑𝑥
𝑑 𝑢
= 2𝑎
𝑑𝑥
We know that, residual,

𝑑 𝑦
𝑅= + 𝜌 𝑔 𝑐𝑜𝑠 𝜃 = 0
𝑑𝑥
  (2 𝑎 ) + 𝜌 𝑔 𝑐𝑜𝑠 𝜃 = 0
− 𝜌 𝑔 𝑐𝑜𝑠𝜃
𝑎 =
2
1.42 Basic of Finite Element Method

Substitute 𝑎 value in equation (2),

− 𝜌 𝑔 𝑐𝑜𝑠𝜃
 𝑢(𝑥) = [𝑥 − 𝐿 ]
2
𝜌 𝑔 𝑐𝑜𝑠𝜃
𝑢(𝑥) = [𝐿 −𝑥 ]
2
Result: Velocity distribution
𝜌 𝑔 𝑐𝑜𝑠𝜃
𝑢(𝑥) = [𝐿 −𝑥 ]
2
Example 1.10
Find the solution for the following differential equation.
𝑑 𝑢
𝐸𝐼 −𝑞 = 0
𝑑𝑥

The boundary conditions are u(0) = 0, (0) = 0

𝑑 𝑢 𝑑 𝑢
(𝐿) = 0, (𝐿) = 0
𝑑𝑥 𝑑𝑥
Given: The governing differential equation
𝑑 𝑢
𝐸𝐼 −𝑞 = 0
𝑑𝑥

The boundary conditions are u(0) = 0, (0) = 0

𝑑 𝑢 𝑑 𝑢
(𝐿) = 0, (𝐿) = 0
𝑑𝑥 𝑑𝑥
Solution: Assume a trial function.
𝐿𝑒𝑡 𝑢(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 … … (1)
Apply first boundary condition,
i.e., at x = 0, u(x) = 0
 0 =𝑎 +0+0+0+0
 𝑎 =0
 Finite Element Analysis 1.43

Apply second boundary condition,

𝑑𝑢
i. e. , at 𝑥 = 0, = 0
𝑑𝑥
𝑑𝑢
 = 0+a +2𝑎 +3𝑎 𝑥 +4𝑎 𝑥
𝑑𝑥

0= 𝑎 +0+0+0

 𝑎 =0
Apply third boundary condition,
𝑑 𝑢
i. e. , at x = L, = 0
𝑑𝑥

𝑑 𝑢
 = 2 a + 6 𝑎 + 12 𝑎 𝑥
𝑑𝑥

 0 = 2 𝑎 + 6 𝑎 𝐿 + 12 𝑎 𝐿

 2 𝑎 = −6 𝑎 𝐿 − 12 𝑎 𝐿

 𝑎 = −[3 𝑎 𝐿 + 6 𝑎 𝐿 ]
Apply fourth boundary condition,
𝑑 𝑢
i. e. , at 𝑥 = L, = 0
𝑑𝑥

𝑑 𝑢
 = 0 + 6 𝑎 + 24 𝑎 𝑥
𝑑𝑥

 0 = 6 𝑎 + 24 𝑎 𝐿

 6 𝑎 = −24 𝑎 𝐿

 𝑎 = −4 𝑎 𝐿
Substitute 𝑎 , 𝑎 , 𝑎 𝑎𝑛𝑑 𝑎 values in equation (1),
𝑢(𝑥 ) = 0 + 0 − [3 𝑎 𝐿 + 6 𝑎 𝐿 ]𝑥 − 4 𝑎 𝐿 𝑥 + 𝑎 𝑥
1.44 Basic of Finite Element Method

= −[3 𝑎 𝐿 + 6 𝑎 𝐿 ]𝑥 − 4 𝑎 𝐿 𝑥 + 𝑎 𝑥

= −[3 (−4 𝑎 𝐿) × 𝐿 + 6 𝑎 𝐿 ]𝑥 − 4 𝑎 𝐿 𝑥 + 𝑎 𝑥

[∵ 𝑎 = −4 𝑎 𝐿]
= 12 𝑎 𝐿 𝑥 − 6 𝑎 𝐿 𝑥 − 4 𝑎 𝐿 𝑥 + 𝑎 𝑥

= 𝑎 [12 𝐿 𝑥 − 6 𝐿 𝑥 − 4 𝐿 𝑥 + 𝑥 ]

𝑢(𝑥) = 𝑎 [6 𝐿 𝑥 − 4 𝐿 𝑥 + 𝑥 ] … (2)

𝑑𝑢
 = 𝑎 [6 𝐿 (2𝑥) − 12 𝐿 𝑥 + 4 𝑥 ]
𝑑𝑥

𝑑 𝑢
 = 𝑎 [6 𝐿 (2) − 24 𝐿 𝑥 + 12 𝑥 ]
𝑑𝑥

𝑑 𝑢
 = 𝑎 [0 − 24𝐿 + 24𝑥]
𝑑𝑥

𝑑 𝑢
 = 𝑎 [0 − 0 + 24]
𝑑𝑥

𝑑 𝑢
= 24 𝑎
𝑑𝑥
We know that, Residual,
𝑑 𝑢
𝑅=𝐸𝐼 −𝑞 =0
𝑑𝑥

 𝐸 𝐼 (24 𝑎 ) − 𝑞 = 0

 𝐸 𝐼 24 𝑎 = 𝑞
𝑞
 𝑎 =
24 𝐸 𝐼
Substitute 𝑎 value in equation (2),
𝑞
 𝑢(𝑥) = [6𝐿 𝑥 −4𝐿𝑥 +𝑥 ]
24 𝐸 𝐼
 Finite Element Analysis 1.45
𝑞
 𝑢(𝑥) = [𝑥 −4𝐿𝑥 +6𝐿 𝑥 ]
24 𝐸 𝐼
Result: Final solution
𝑞
𝑢(𝑥) = [𝑥 −4𝐿𝑥 +6𝐿 𝑥 ]
24 𝐸 𝐼

Example 1.11
The following differential equation is available for a physical phenomenon.
𝑑 𝑢
𝐴𝐸 +𝑎𝑥 =0
𝑑𝑥
The boundary conditions are 𝑢(0) = 0, 𝐴 𝐸 = 0. By using Galerkin’s
technique, find the solution of the above differential equation.
Given: Differential equation,
𝑑 𝑢
𝐴𝐸 +𝑎𝑥 =0
𝑑𝑥
Boundary condition 𝑢(0) = 0, 𝐴 𝐸 =0

To find: u(x) by using Galerkin’s technique.

Solution: Assume a trial function.
𝐿𝑒𝑡 𝑢(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 … … (1)
Apply first boundary condition, i.e., at x = 0, u(x) = 0
(1) 0=𝑎 +0+0+0
 𝑎 =0
Apply second boundary condition,
𝑑𝑢
i. e. , at 𝑥 = L, 𝐴 𝐸 =0
𝑑𝑥
𝑑𝑢
 = 0+a +2𝑎 𝑥+3𝑎 𝑥
𝑑𝑥
 0= a +2𝑎 𝐿+3𝑎 𝐿
1.46 Basic of Finite Element Method

 𝑎 = −(2 𝑎 𝐿 + 3 𝑎 𝐿 )
Substitute 𝑎 𝑎𝑛𝑑 𝑎 value in equation (1),
𝑢(𝑥 ) = 0 + −(2 𝑎 𝐿 + 3 𝑎 𝐿 )𝑥 − 𝑎 𝑥 + 𝑎 𝑥

= −2 𝑎 𝐿𝑥 + 3 𝑎 𝐿 𝑥 + 𝑎 𝑥 + 𝑎 𝑥

= 𝑎 [𝑥 − 2 𝐿𝑥] + 𝑎 [𝑥 − 3 𝐿 𝑥]

𝑢(𝑥) = 𝑎 [𝑥 − 2 𝐿𝑥] + 𝑎 [𝑥 − 3 𝐿 𝑥] … (2)

We know that, Residual,
𝑑 𝑢
𝑅 = 𝐴𝐸 +𝑎𝑥 . . (3)
𝑑𝑥
𝑑𝑢
(2)  = 𝑎 [2 𝑥 − 2𝐿] + 𝑎 [3𝑥 − 3 𝐿 ]
𝑑𝑥

𝑑 𝑢
= 𝑎 [2] + 𝑎 [6 𝑥]
𝑑𝑥

𝑑 𝑢
=2𝑎 +6𝑎 𝑥
𝑑𝑥

Substitute value in equation (3),

(3) 𝑅 = 𝐴 𝐸 (2 𝑎 + 6 𝑎 𝑥) + 𝑎 𝑥
Residual, 𝑅 = 𝐴 𝐸 (2 𝑎 + 6 𝑎 𝑥) + 𝑎 𝑥 … (4)
From Galerkin’s technique,

𝑤 𝑅 𝑑𝑥 = 0 … (5)

where, w = weighting function

From equation (2), we know that,
𝑤 = (𝑥 − 2 𝐿 𝑥)
𝑤 = (𝑥 − 3 𝐿 𝑥)
 Finite Element Analysis 1.47

Substitute 𝑤 , 𝑤 and R values in equation (5),

(5) (𝑥 − 2 𝐿 𝑥)[ 𝐴 𝐸 (2 𝑎 + 6 𝑎 𝑥) + 𝑎 𝑥]𝑑𝑥 = 0 … (6)

(𝑥 − 3 𝐿 𝑥)[ 𝐴 𝐸 (2 𝑎 + 6 𝑎 𝑥) + 𝑎 𝑥]𝑑𝑥 = 0 … (7)

(6) (𝑥 − 2 𝐿 𝑥)[ 𝐴 𝐸 (2 𝑎 + 6 𝑎 𝑥) + 𝑎 𝑥]𝑑𝑥 = 0

(𝑥 − 2 𝐿 𝑥)[ 2 𝑎 𝐴 𝐸 + 6 𝑎 + 𝐴 𝐸 𝑥 + 𝑎 𝑥 ]𝑑𝑥 = 0

[2 𝑎 𝐴 𝐸 𝑥 + 6 𝑎 𝐴 𝐸 𝑥 + 𝑎 𝑥 − 4 𝑎 𝐴 𝐸 𝐿 𝑥 − 12 𝑎 𝐴 𝐸 𝐿 𝑥

− 2 𝑎 𝐿 𝑥 ] 𝑑𝑥 = 0
𝑥 𝑥 𝑥 𝑥 𝑥
 2𝑎 𝐴𝐸 +6𝑎 𝐴𝐸 +𝑎 − 4 𝑎 𝐴 𝐸 𝐿 − 12 𝑎 𝐴 𝐸 𝐿
3 4 4 𝑥 3
𝑥
−2𝑎𝐿 =0
3

𝐿 𝐿 𝐿 𝐿 𝐿 𝐿
2𝑎 𝐴𝐸 +6𝑎 𝐴𝐸 +𝑎 −4𝑎 𝐴𝐸 − 12 𝑎 𝐴 𝐸 −2𝑎 =0
3 4 4 𝑥 3 3
2 3 𝑎𝐿 2
 𝑎 𝐴 𝐸 𝐿 + 𝑎 𝐴 𝐸𝐿 + −2𝑎 𝐴𝐸𝐿 −4𝑎 𝐴𝐸 𝐿 − 𝑎𝐿 = 0
3 2 4 3
2 3 𝑎𝐿 2
 𝐴 𝐸𝑎 𝐿 − 2 + 𝑎 𝐴 𝐸𝐿 −4 + − 𝑎𝐿 = 0
3 2 4 3
−4 5 2 1
 𝐴𝐸𝐿 𝑎 − 𝐴𝐸𝐿 𝑎 = − 𝑎𝐿
3 2 3 4
−4 5 5
 𝐴𝐸𝐿 𝑎 − 𝐴𝐸𝐿 𝑎 = 𝑎𝐿
3 2 12
1.48 Basic of Finite Element Method
4 5 −5
 𝐴𝐸𝑎 𝐿 + 𝐴𝐸 𝑎 𝐿 = 𝑎𝐿 … (8)
3 2 12
Equation (7),

 (𝑥 − 3 𝐿 𝑥)[ 𝐴 𝐸 (2 𝑎 + 6 𝑎 𝑥) + 𝑎 𝑥]𝑑𝑥 = 0

 [𝑥 − 3 𝐿 𝑥][ 2 𝑎 𝐴 𝐸 + 6 𝑎 𝐴 𝐸 𝑥 + 𝑎 𝑥 ]𝑑𝑥 = 0

 [ 2 𝐴 𝐸 𝑎 𝑥 + 6 𝐴 𝐸 𝑎 𝑥 + 𝑎 𝑥 − 6 𝐴 𝐸 𝑎 𝐿 𝑥 − 18 𝐴 𝐸 𝑎 𝐿 𝑥

− 3 𝑎 𝐿 𝑥 ]𝑑𝑥 = 0

𝑥 𝑥 𝑎𝑥 𝑥 𝑥
 2𝐴𝐸𝑎 + 6𝐴 𝐸 𝑎 + −6𝐴𝐸𝑎 𝐿 − 18𝐴𝐸 𝑎 𝐿
4 5 5 2 3

𝑥
−3𝑎𝐿 =0
𝑥

1 6 1
 𝐴 𝐸 𝑎 𝑥 + 𝐴𝐸 𝑎 𝑥 𝑎 𝑥 − 3𝐴𝐸 𝑎 𝐿 𝑥 − 6 𝐴𝐸 𝑎 𝐿 𝑥 − 𝑎 𝐿 𝑥
2 5 5

1 6 1
 𝐴𝐸 𝑎 𝐿 + 𝐴𝐸 𝑎 𝐿 + 𝑎 𝐿 − 3 𝐴𝐸 𝑎 𝐿 (𝐿 ) − 6 𝐴𝐸 𝑎 𝐿 (𝐿 )
2 5 5
− 𝑎 𝐿 (𝐿 ) = 0

1 6 1
 𝐴𝐸 𝑎 𝐿 + 𝐴𝐸 𝑎 𝐿 + 𝑎 𝐿 − 3 𝐴𝐸 𝑎 𝐿 − 6 𝐴𝐸 𝑎 𝐿 − 𝑎 𝐿 = 0
2 5 5
1 6 1
 𝐴 𝐸𝑎 𝐿 − 3 + 𝐴𝐸 𝑎 𝐿 −6 +𝑎𝐿 −1 =0
2 5 5
−5 24 4
𝐴𝐸𝐿 − 𝐴𝐸 𝑎 𝐿 = 𝑎 𝐿
2 5 5
 Finite Element Analysis 1.49
5 24 −4
 𝐴𝐸 𝑎 𝐿 + 𝐴𝐸 𝑎 𝐿 = 𝑎𝐿 … (9)
2 5 5

Solving equations (8) and (9),

4 5 −5
Equation (8) 𝐴 𝐸 𝑎 𝐿 + 𝐴𝐸 𝑎 𝐿 = 𝑎𝐿
3 2 12
5 24 −4
Equation (9) 𝐴𝐸 𝑎 𝐿 + 𝐴𝐸 𝑎 𝐿 = 𝑎𝐿
2 5 5

Multiplying equation (8) by L and equation (9) by ,

20 25 −25
𝐴𝐸𝑎 𝐿 + 𝐴𝐸 𝑎 𝐿 = 𝑎𝐿
6 4 24
20 96 −16
𝐴𝐸 𝑎 𝐿 + 𝐴𝐸 𝑎 𝐿 = 𝑎𝐿
6 15 15

25 96 16 25
Subtracting, − 𝐴𝐸 𝑎 𝐿 = − 𝑎𝐿
4 15 15 24
375 − 384 384 − 375
𝐴𝐸 𝑎 𝐿 = 𝑎𝐿
60 360
9 9
 − 𝐴𝐸 𝑎 𝐿 = 𝑎𝐿
60 360

 − 0.15 𝐴𝐸 𝑎 = 0.025 𝑎
𝑎
 𝑎 = −0.1666
𝐴𝐸
−𝑎
 𝑎 = … (10)
6𝐴𝐸
Substituting 𝑎 value in equation (8),
4 5 −𝑎 −5
𝐴𝐸 𝑎 𝐿 + 𝐴𝐸 𝐿 = 𝑎𝐿
3 2 6 𝐴𝐸 12
1.50 Basic of Finite Element Method
4 −5 5 −𝑎
𝐴𝐸 𝑎 𝐿 = 𝑎 𝐿 − 𝐴𝐸𝐿
3 12 2 6 𝐴𝐸
4 −5 5
𝐴𝐸 𝑎 𝐿 = 𝑎𝐿 + 𝑎𝐿
3 12 2
4
𝐴𝐸 𝑎 𝐿 = 0
3

 𝑎 =0

Substitute 𝑎 𝑎𝑛𝑑 𝑎 values in equation (2),

−𝑎
 𝑢(𝑥) = 0 × [𝑥 − 2 𝐿 𝑥] + [𝑥 − 3 𝐿 𝑥] = 0
6 𝐴𝐸
𝑎
 𝑢(𝑥) = [3𝐿 𝑥−𝑥 ]
6 𝐴𝐸
𝑎
𝐑𝐞𝐬𝐮𝐥𝐭: 𝑢(𝑥) = [3𝐿 𝑥−𝑥 ]
6 𝐴𝐸

Example 1.12
The following differential equation for the long cylinder of radius R with heat
generation q0 is given by
𝑑 𝑇 𝐼 𝑑𝑇 𝑞
+ + =0
𝑑𝑟 𝑟 𝑑𝑟 𝑘
The boundary conditions are T(R) = Tw
𝑑𝑇
𝑞 𝜋 𝑅 𝐿 = (−𝑘)(2𝜋 𝑅 𝐿)
𝑑𝑟
Find the temperature distribution T as a function of radial location r.
Given: Differential equation

𝑑 𝑇 𝐼 𝑑𝑇 𝑞
+ + =0
𝑑𝑟 𝑟 𝑑𝑟 𝑘
Boundary conditions T(R) = Tw
 Finite Element Analysis 1.51
𝑑𝑇
𝑞 𝜋 𝑅 𝐿 = (−𝑘)(2𝜋 𝑅 𝐿)
𝑑𝑟
To find: Temperature distribution, T(r)
Solution: Assume a trial solution.
𝐿𝑒𝑡 𝑇 = 𝑎 + 𝑎 (𝑟 − 𝑅) + 𝑎 (𝑟 − 𝑅 ) … . (1)

Apply first boundary condition, i.e., at r = R, T = Tw

(1) 𝑇 = 𝑎 + 𝑎 (𝑅 − 𝑅) + 𝑎 (𝑅 − 𝑅)

𝑇 =𝑎

𝑎 =𝑇
Apply second boundary condition, i.e.,
𝑑𝑇
𝑞 𝜋 𝑅 𝐿 = (−𝑘)(2𝜋 𝑅 𝐿)
𝑑𝑟

𝑑𝑇
(1)  = 0 + 𝑎 (1 − 0) + 𝑎 2 (𝑟 − 𝑅)(1 − 0)
𝑑𝑟

= 𝑎 + 2 𝑎 (𝑟 − 𝑅)

𝑑𝑦
= 𝑎 + 2 𝑎 (𝑅 − 𝑅)
𝑑𝑥

𝑑𝑦
 =𝑎
𝑑𝑥
We know that,

𝑑𝑇
−𝑘 2𝜋 𝑅 𝐿 =𝑞 𝜋𝑅 𝐿
𝑑𝑟

 − 𝑘 2𝜋 𝑅 𝐿(𝑎 ) = 𝑞 𝜋 𝑅 𝐿

𝑞 𝜋𝑅 𝐿
 𝑎 =
−𝑘 2𝜋 𝑅 𝐿
1.52 Basic of Finite Element Method
−𝑞 𝑅
 𝑎 =
2𝑘
Solving 𝑎 𝑎𝑛𝑑 𝑎 value in equation (1),
−𝑞 𝑅
 𝑇=𝑇 + (𝑟 − 𝑅) + 𝑎 (𝑟 − 𝑅)
2𝑘
𝑞 𝑅
𝑇=𝑇 − (𝑟 − 𝑅) + 𝑎 (𝑟 − 𝑅) … (2)
2𝑘
We know that, Residual,
𝑑 𝑇 1 𝑑𝑇 𝑞
R + + … (3)
𝑑𝑟 𝑟 𝑑𝑟 𝑘
𝑑𝑇 𝑞 𝑅
(2)  =0− + 𝑎 2(𝑟 − 𝑅)
𝑑𝑟 2𝑘

𝑑 𝑇
= 0 + 2𝑎 (1 − 0)
𝑑𝑟

= 2𝑎
1 −𝑞 𝑅
(3)  𝑅 = 2𝑎 + + 2 𝑎 (𝑟 − 𝑅) … (4)
𝑟 2𝑘
From Galerkin’s technique

𝑤 𝑅 𝑑𝑟 = 0

𝑞 1 −𝑞 𝑅
 2 𝐿 (𝑟 − 𝑅) 2 𝑎 + + + 2 𝑎 (𝑟 − 𝑅) 𝑟 𝑑𝑟 = 0
𝑘 𝑟 2𝑘

𝑞 𝑞 𝑅 2𝑎
 2 𝐿 (𝑟 − 𝑅) 2 𝑎 + − + (𝑟 − 𝑅) 𝑟 𝑑𝑟 = 0
𝑘 2𝑘𝑟 𝑟

𝑞 𝑟 𝑞 𝑅
 2 𝐿 (𝑟 − 𝑅) 2 𝑎 𝑟 + − + 2 𝑎 (𝑟 − 𝑅) 𝑑𝑟 = 0
𝑘 2𝑘
 Finite Element Analysis 1.53

𝑞 𝑟 𝑞 𝑅
 2 𝐿 (𝑟 − 𝑅) 2𝑎 𝑟+ − + 2𝑎 𝑟−2𝑎 𝑅 𝑑𝑟 = 0
𝑘 2𝑘

… (5)
Now apply Bernoull’s formula,

𝑢 𝑣 𝑑𝑥 = 𝑢 𝑣 − 𝑢 𝑣 + 𝑢"𝑣 + 𝑢′′′ 𝑣 + ⋯

u = Differentiate

v = integrate
Differentiating u with respect to r,
𝑢 = (𝑟 − 𝑅)

𝑢′ = 2 (𝑟 − 𝑅)

𝑢′′ = 2

𝑢′′′ = 0
Integrating v with respect to r,
𝑞 𝑞
𝑣 = 2𝑎 𝑟+ 𝑟− 𝑅 + 2𝑎 𝑟 − 2 𝑎 𝑅
𝑘 2𝑘

𝑟 𝑞 𝑟 𝑞 𝑟
𝑣 = 2𝑎 + − 𝑅 𝑟 + 2𝑎 −2𝑎 𝑅𝑟
2 𝑘 2 2𝑘 2

𝑟 𝑞 𝑟 𝑞 𝑟 𝑟 𝑟
 𝑣 = 𝑎 + − 𝑅 +𝑎 −2𝑎 𝑅
3 2𝑘 3 2𝑘 2 3 2

𝑎 𝑟 𝑞 𝑟 𝑞 𝑟 𝑎 𝑟 𝑟
𝑣 = + − 𝑅 + − 𝑎 𝑅
3 3 6𝑘 4 4𝑘 3 3 4 3

𝑎 𝑟 𝑞 𝑞 𝑅𝑟 𝑎 𝑟 𝑎 𝑅𝑟
= + − + −
12 24 𝑘 12 𝑘 12 3
Now substitute the above values in the Bernoulli’s formula
1.54 Basic of Finite Element Method

𝑢 𝑣 𝑑𝑥 = 𝑢 𝑣 − 𝑢 𝑣 + 𝑢"𝑣 + 𝑢′′′ 𝑣 + ⋯

𝑞 𝑟 𝑞 𝑅𝑟
= (𝑟 − 𝑅) 𝑎 𝑟 + − +𝑎 𝑟 −2𝑎 𝑟𝑅
2𝑘 2𝑘
𝑎 𝑟 𝑞 𝑟 𝑞 𝑅𝑟 𝑎 𝑟
= 2(𝑟 − 𝑅) + − +
3 6𝑘 4𝑘 3
𝑎 𝑟 𝑞 𝑟 𝑞 𝑅𝑟 𝑎 𝑟
− 𝑎 𝑅 𝑟 +2 + − +
12 24 𝑘 12 𝑘 12
𝑎 𝑅𝑟
−
3
Substitute in equation (5),
𝑞 𝑟 𝑞 𝑅𝑟
2𝜋 𝐿 (𝑟 − 𝑅) 𝑎 𝑟 + − +𝑎 𝑟 −2𝑎 𝑟𝑅
2𝑘 2𝑘
𝑎 𝑟 𝑞 𝑟 𝑞 𝑅𝑟 𝑎 𝑟
− 2(𝑟 − 𝑅) + − +
3 6𝑘 4𝑘 3
𝑎 𝑟 𝑞 𝑟 𝑞 𝑅𝑟 𝑎 𝑟
− 𝑎 𝑅 𝑟 +2 + − +
12 24 𝑘 12 𝑘 12
𝑅𝑟
− 𝑎 =0
3

𝑎 𝑅 𝑞 𝑅 𝑞 𝑅 𝑎 𝑅 𝑎 𝑅
 2𝜋𝐿 0−0+2 + − + − −0 =0
12 24 𝑘 12 𝑘 12 3

𝑎 𝑅 𝑞 𝑅 𝑞 𝑅 𝑎 𝑅 2𝑎 𝑅
 2𝜋𝐿 + − + − =0
6 12 𝑘 6𝑘 6 3

2𝑎 𝑅 𝑞 𝑅 𝑞 𝑅 2𝑎 𝑅
 2𝜋𝐿 + − − =0
6 12 𝑘 6𝑘 3

𝑎 𝑅 𝑞 𝑅 2𝑎 𝑅
 − − =0
3 12 𝑘 3

−𝑎 𝑅 𝑞 𝑅
=
3 12 𝑘
 Finite Element Analysis 1.55

−3 𝑞 𝑅
 𝑎 =
𝑅 × 12 𝑘
−𝑞
𝑎 =
4𝑘
Substitute 𝑎 value in equation (2),
−𝑞 𝑅 −𝑞
(2) 𝑇=𝑇 + (𝑟 − 𝑅) + (𝑟 − 𝑅)
2𝑘 4𝑘

𝑞 𝑅𝑟 𝑞 𝑅 𝑞
=𝑇 − + − [ 𝑟 + 𝑅 − 2 𝑟 𝑅]
2𝑘 2𝑘 2𝑘

𝑞 𝑅𝑟 𝑞 𝑅 𝑞 𝑟 𝑞 𝑅 𝑞 2𝑟𝑅
=𝑇 − + − − +
2𝑘 2𝑘 4𝑘 4𝑘 4𝑘

𝑞 𝑅𝑟 𝑞 𝑅 𝑞 𝑟 𝑞 𝑅 𝑞 𝑟𝑅
=𝑇 − + − − +
2𝑘 2𝑘 4𝑘 4𝑘 4𝑘

𝑞 𝑅 1 1 𝑞 𝑟
=𝑇 + − −
𝑘 2 4 4𝑘

𝑞 𝑅 2 𝑞 𝑟
=𝑇 + −
𝑘 8 4𝑘

𝑞 𝑅 𝑞 𝑟
=𝑇 + −
4𝑘 4𝑘
𝑞
𝑇 =𝑇 + [𝑅 − 𝑟 ]
4𝑘
𝑞
 𝑇−𝑇 = [𝑅 − 𝑟 ]
4𝑘

Example 1.13
The following differential equation is available for a physical phenomenon.
𝑑 𝑦
+ 50 = 0, 0 ≤ 𝑥 ≤ 10
𝑑𝑥
Trial function is, 𝑦 = 𝑎 (10 − 𝑥)
Boundary conditions are, y(0) = 0
1.56 Basic of Finite Element Method

y(10) = 0
Find the value of the parameter 𝑎 by the following methods:
(i) Point collocation; (ii) Subdomain collocation; (iii) Least squares; (iv) Galerkin.
Given:
Differential equation,
𝑑 𝑦
+ 50 = 0, 0 ≤ 𝑥 ≤ 10 … (1)
𝑑𝑥
Trial function is, 𝑦 = 𝑎 𝑥(10 − 𝑥)
Boundary conditions are, y(0) = 0
y(10) = 0
To find: The value of the parameter 𝑎 by,
(i) Point collocation method; (ii) Subdomain collocation method;
(iii) Least squares method; (iv) Galerkin’s method.
Solution:
First we have to verify, whether the trial function satisfies the boundary conditions
or not.
Trial function is, 𝑦 = 𝑎 𝑥 (10 − 𝑥)
When x = 0, y = 0
x = 10, y = 0
Hence it satisfies the boundary conditions.
(i) Point Collocation method: 𝑦 = 𝑎 𝑥 (10 − 𝑥)

 𝑦 = 𝑎 (10𝑥 − 𝑥 )

𝑑𝑦
 = 𝑎 (10 − 2𝑥)
𝑑𝑥

𝑑 𝑦
 =−2𝑎
𝑑𝑥
 Finite Element Analysis 1.57

Substitute value in given differential equation (1),

 Residual, R = −2 a + 50 … (2)
In point collocation method, residuals are set to zero.
 R = −2 a + 50 = 0
−2 a = −50
a = 25 … (3)
Hence the trial function is, y = 25 x (10- x)
(ii) Subdomain collocation method:

This method requires, 𝑅 𝑑𝑥 = 0

Substitue R value,  [−2 𝑎 + 50] 𝑑𝑥 = 0

 [−2 𝑎 𝑑𝑥 + 50 𝑑𝑥] = 0

 [−2 𝑎 𝑥 + 50 𝑥] =0
 − 20 𝑎 = −500
𝑎 = 25 … (4)
Hence the trial function is, y = 25 x (10- x)
(iii) Least squares method:

This method requires I= 𝑅 𝑑𝑥

𝜕𝐼 𝜕𝐼
It can also be written as, = 𝑅 𝑑𝑥 . . (5)
𝜕𝑎 𝜕𝑎
1.58 Basic of Finite Element Method

We know that, 𝑅 = −2 𝑎 + 50
𝜕𝐼
= −2
𝜕𝑎
𝜕𝐼
Substitue R and value in equation (5),
𝜕𝑎

𝜕𝐼
 (−2 𝑎 + 50)(−2) 𝑑𝑥
𝜕𝑎

The requirement is, =0

 (−2 𝑎 + 50)(−2) 𝑑𝑥 = 0

 (−2 𝑎 + 50)𝑑𝑥 = 0

 [−2 𝑎 𝑑𝑥 + 50 𝑑𝑥 ] = 0

 [−2 𝑎 + 50 𝑥] =0

 − 2 a (10) + 50(10) − [0] = 0

 − 2 a + 500 = 0

 − 2 a = −500

 a = 25 . . . (6)

Therefore, the trial; function becomes, y = 25 x (10 –x).

(iv) Galerkin’s method:
In this method, the trial function itself is considered as the weighting function , wi
 Finite Element Analysis 1.59

 𝑤 𝑅 𝑑𝑥 = 0 … (7)

Here, the trial function is, 𝑦 = 𝑤 = 𝑎 𝑥(10 − 𝑥)

Substitute 𝑤 𝑎𝑛𝑑 𝑅 Values in equation (7),

 𝑎 𝑥(10 − 𝑥) × (−2 𝑎 + 50) 𝑑𝑥 = 0

 𝑎 𝑥(10 − 𝑥) × (−2 𝑎 + 50) 𝑑𝑥 = 0

 𝑎 (10𝑥 − 𝑥 )(−2 𝑎 + 50) 𝑑𝑥 = 0

 𝑎 [−20 𝑎 𝑥 + 500 𝑥 + 2 𝑎 𝑥 − 50 𝑥 ]𝑑𝑥 = 0

𝑥 𝑥 𝑥 𝑥
 𝑎 −20 𝑎 + 500 + 2 𝑎 − 50 =0
2 2 3 3

−20 𝑎 500 2𝑎 50
 [10 − 0] + [10 − 0] + [10 − 0] − [10 − 0] = 0
2 2 3 3
2𝑎 50
 − 10a [100] + 250 [100] + [ 1000] − [ 1000] = 0
3 3

 − 1000a + 25000 + 666.66 𝑎 − 16666.66 = 0

 − 333.33 a = −8333.33

 a = 25 … (8)
The trial function is, y = 25 x (10 – x)
1.60 Basic of Finite Element Method

From equation (3), (4), (6) and (8) we know that the value of parameter 𝑎 is
same for all the four methods.
Result: Parameter, 𝑎 (for all the four methods) = 25

Example 1.14
The differential equation of a physical phenomenon is given by,
𝑑 𝑦
+ 500 𝑥 = 0, 0 ≤ 𝑥 ≤ 1
𝑑𝑥
Trial function is, 𝑦 = 𝑎 (𝑥 − 𝑥 )
Boundary conditions are, y(0) = 0
y(1) = 0
calculate the value of the parameter 𝑎 by the following methods:
(i) Point collocation; (ii) Subdomain collocation; (iii) Least squares; (iv) Galerkin.
Given:
𝑑 𝑦
Differential equation + 500 𝑥 = 0, 0 ≤ 𝑥 ≤ 1 … (1)
𝑑𝑥
Trial function is, 𝑦 = 𝑎 (𝑥 − 𝑥 )
Boundary conditions are, y(0) = 0
y(1) = 0
To find: The value of the parameter 𝑎 by,
(i) Point collocation method;
(ii) Subdomain collocation method;
(iii) Least squares method;
(iv) Galerkin’s method.
Solution:
First, we have to verify whether the trial function satisfies the boundary conditions
or not.
Trial function is, 𝑦 = 𝑎 (𝑥 − 𝑥 )
 Finite Element Analysis 1.61

When x = 0, y = 0
x = 1, y = 0
Hence it satisfies the boundary conditions.
(i) Point Collocation method: 𝑦 = 𝑎 (𝑥 − 𝑥 )
𝑑𝑦
 = 𝑎 (1 − 4𝑥 )
𝑑𝑥

𝑑 𝑦
 = 𝑎 (0 − 12𝑥 )
𝑑𝑥

𝑑 𝑦
 = − 12 𝑎 𝑥
𝑑𝑥

Substitute value in given differential equation (1),

 Residual, R = −12 a 𝑥 + 500 𝑥 … (2)

In point collocation method, residuals are set to zero.
 R = −12 a 𝑥 + 500 𝑥 = 0 … (3)
In this problem, we have to find only one parameter, 𝑎 . So, only one collocation
point is needed. The point may be chosen between 0 and 1. Let us take .

Substituting 𝑥 = in equation (3)

1 1
 R = −12 a + 500 =0
2 2
1 1
 − 12 a + 500 =0
4 4

 − 3 a + 125 = 0

 a = 41.66 … (4)

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

1.62 Basic of Finite Element Method

(ii) Subdomain collocation method:

This method requires, 𝑅 𝑑𝑥 = 0

Substitue R value,

 (−12 𝑎 𝑥 + 500 𝑥 ) 𝑑𝑥 = 0

𝑥 𝑥
 − 12 𝑎 + 500 =0
3 3
−12 𝑎 500
 [1 − 0] + [1 − 0] = 0
3 3
−12 𝑎 500
 + =0
3 3
 − 12 𝑎 + 500 = 0
 − 12 𝑎 = −500
500
 𝑎 =
12
 𝑎 = 41.66 … (5)
Trial function is, y = 41.66 (x - x4)
(iii) Least squares method:

This method requires I= 𝑅 𝑑𝑥

𝜕𝐼 𝜕𝑅
It can also be written as, = 𝑅 𝑑𝑥 . . (6)
𝜕𝑎 𝜕𝑎

We know that, 𝑅 = −12 𝑎 𝑥 + 500 𝑥

𝜕𝑅
= −12𝑥
𝜕𝑎
 Finite Element Analysis 1.63
𝜕𝑅
Substitue R and value in equation (6),
𝜕𝑎

𝜕𝐼
 = (−12 𝑎 𝑥 + 500 𝑥 )(−12 𝑥 ) 𝑑𝑥
𝜕𝑎

The requirement is, =0

 (−12 𝑎 𝑥 + 500 𝑥 )(−12 𝑥 ) 𝑑𝑥 = 0

 (144 𝑎 𝑥 − 6000 𝑥 )𝑑𝑥 = 0

 (144 𝑎 𝑥 − 6000 𝑥 )𝑑𝑥 = 0

𝑥 𝑥
 144 𝑎 + 6000 =0
5 5

144 𝑎 6000
 [1 − 0] + [1 − 0] = 0
5 5

 28.8 a = 1200

 a = 41.66 . . . (7)
The trial function is, y = 41.66 (x –x4).
(iv) Galerkin’s method: In this method, the trial function itself is considered as the
weighting function , wi

 𝑤 𝑅 𝑑𝑥 = 0 … (8)

Here, the trial function is, 𝑦 = 𝑤 = 𝑎 (𝑥 − 𝑥 )

1.64 Basic of Finite Element Method

Substitute 𝑤 𝑎𝑛𝑑 𝑅 Values in equation (8),

 𝑎 (𝑥 − 𝑥 )(−12 𝑎 𝑥 + 500 𝑥 ) 𝑑𝑥 = 0

 𝑎 (𝑥 − 𝑥 )(−12 𝑎 𝑥 + 500 𝑥 ) 𝑑𝑥 = 0

 𝑎 (−12 𝑎 𝑥 + 500 𝑥 + 12 𝑎 𝑥 + 500 𝑥 ) 𝑑𝑥 = 0

𝑥 𝑥 𝑥 𝑥
 𝑎 −12 𝑎 + 500 + 12 𝑎 − 500 =0
4 4 7 7

−12 𝑎 500 12 𝑎 500

 [1 − 0] + [1 − 0] + [1 − 0] − [1 − 0] = 0
4 4 7 7

 − 3a + 125 + 1.714 𝑎 − 71.428 = 0

 − 1.286 a = −53.572

 a = 41.66 … (9)

Trial function is, y = 41.66 (x - x4)

From equation (4), (5), (7) and (9) we know that the value of parameter 𝑎 is
same for all the four methods.
Result: Parameter, 𝑎 (For all the four methods) = 41.66.

Example 1.15
The differential equation of a physical phenomenon is given by,
𝑑 𝑦
+ 500 𝑥 = 0; 0 ≤ 𝑥 ≤ 1
𝑑𝑥
By using the trial function, 𝑦 = 𝑎 (𝑥 − 𝑥 ) + 𝑎 (𝑥 − 𝑥 ), calculate the value of
the parameter 𝑎 𝑎𝑛𝑑 𝑎 by the following methods:
 Finite Element Analysis 1.65

(i) Point collocation; (ii) Subdomain collocation; (iii) Least squares; (iv) Galerkin.
The boundary conditions are: y(0) = 0
y(1) = 0
Given:
𝑑 𝑦
Differential equation + 500 𝑥 = 0, 0 ≤ 𝑥 ≤ 1 … (1)
𝑑𝑥
Trial function is, 𝑦 = 𝑎 (𝑥 − 𝑥 ) + 𝑎 (𝑥 − 𝑥 )
Boundary conditions are: y(0) = 0
y(1) = 0
To find: The value of the parameter 𝑎 𝑎𝑛𝑑 𝑎 by,
(i) Point collocation method;
(ii) Subdomain collocation method;
(iii) Least squares method;
(iv) Galerkin’s method.
Solution:
First we have to verify, whether the trial function satisfies the boundary conditions
or not.
Trial function is 𝑦 = 𝑎 (𝑥 − 𝑥 ) + 𝑎 (𝑥 − 𝑥 )

When x = 0, y = 0

x = 1, y = 0
Hence it satisfies the boundary conditions.
Residual, R: 𝑦 = 𝑎 (𝑥 − 𝑥 ) + 𝑎 (𝑥 − 𝑥 )

𝑑𝑦
= 𝑎 (1 − 3 𝑥 ) + 𝑎 (1 − 5 𝑥 )
𝑑𝑥

𝑑 𝑦
= 𝑎 (−6𝑥) + 𝑎 (−20 𝑥 )
𝑑𝑥
1.66 Basic of Finite Element Method

𝑑 𝑦
= −6 𝑎 𝑥 − 20 𝑎 𝑥
𝑑𝑥

Substitute value in given differential equation (1),

Residual, R = −6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥 … (2)

The interval 0 to 1 is divided into two domains 0 𝑡𝑜 𝑎𝑛𝑑 𝑡𝑜 1.

(i) Point Collocation method: In point collocation method, residuals are set to zero.
 R = −6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥 = 0 … (3)
𝟏
Domain (1): Limits is 𝟎 𝐭𝐨 𝟐: In domain (1), we can choose an arbitrary point. Let it be

So, put 𝑥 = in equation (3).

1 1 1
 R = −6 𝑎 − 20 𝑎 + 500 =0
3 3 3
20 500
 −2 𝑎 − 𝑎 + =0
27 9

 −2 𝑎 − 0.741 𝑎 = −55.55

 2 𝑎 − 0.741 𝑎 = 55.55

 𝑎 + 0.3705 𝑎 = 27.775 … (4)

𝟏
Domain (2) : Limits is 𝟐 𝒕𝒐 𝟏: In domain (2), we can choose 𝑥 = and substituting the
same in equation (3).
2 2 2
 R = −6 𝑎 − 20 𝑎 + 500 =0
3 3 3
8 4
 −4 𝑎 − 20 𝑎 × + 500 × = 0
27 9

 −4 𝑎 − 5.925 𝑎 + 222.22 = 0
 Finite Element Analysis 1.67

 4 𝑎 + 5.925 𝑎 = 222.22

 𝑎 + 1.481 𝑎 = 55.555 … (5)

Solving equation (4) and (5)
− 𝑎 − 0.3705 𝑎 = −27.775
𝑎 + 1.481 𝑎 = 55.55
1.111 𝑎 = 55.55
𝑎 = 25
Substituting 𝑎 value in equation (4) or (5).
(5)  a + 1.481(25) = 55.555
a + 37.025 = 55.555
 a = 18.53
Hence the trial function is, y = 18.53 (x- x3) + 25(x- x5)
(ii) Subdomain collocation method:

This method requires, 𝑅 𝑑𝑥 = 0

The interval 0 to 1 is divided into two domains 0 𝑡𝑜 𝑎𝑛𝑑 𝑡𝑜 1.

/

𝐅𝐨𝐫 𝐝𝐨𝐦𝐚𝐢𝐧 (𝟏): 𝑅 𝑑𝑥 = 0

Substitue R value,
/

 (−6 𝑎 𝑥 + 20𝑎 𝑥 + 500 𝑥 ) 𝑑𝑥 = 0

𝑥 𝑥 𝑥
 −6𝑎 − 20 𝑎 + 500 =0
2 4 3
1.68 Basic of Finite Element Method

−6 𝑎 1 20 𝑎 1 500 1
 −0 − −0 + −0 =0
2 2 4 2 3 2
−6 𝑎 20 𝑎 500
 − + =0
8 64 24
 − 0.75 𝑎 − 0.3125 𝑎 + 20.83 = 0
 − 0.75 𝑎 − 0.3125 𝑎 = 20.83
 𝑎 + 0.4166 𝑎 = 27.773 … (6)

𝐅𝐨𝐫 𝐝𝐨𝐦𝐚𝐢𝐧 (𝟐): 𝑅 𝑑𝑥 = 0

/

Substitue R value,

𝑥 𝑥 𝑥
 −6𝑎 − 20 𝑎 + 500 =0
2 /
4 /
3 /

−6 𝑎 1 20 𝑎 1 500 1
 1− − 1− + 1− =0
2 2 4 2 3 2

−6 𝑎 20 𝑎 500
 [0.75] − [0.9375] + [0.875] = 0
2 4 3

 − 2.25 𝑎 − 4.6875 𝑎 + 145.83 = 0

 2.25 𝑎 + 4.6875 𝑎 = 145.83

 𝑎 + 2.083 𝑎 = 64.813 … (7)

Solving equation (6) and (7),
−𝑎 − 0.4166 𝑎 = −27.773

𝑎 + 2.083 𝑎 = 64.813

1.6664 𝑎 = 37.04

𝑎 = 22.23
 Finite Element Analysis 1.69

Substituting 𝑎 value in equation (6) or (7).

(7)  a + 2.083(22.23) = 64.813

a + 46.305 = 64.813

 a = 18.50
Hence the trial function is, y = 18.50 (x- x3)+ 22.23(x- x5)
(iii) Least squares method:

This method requires I= 𝑅 𝑑𝑥

𝜕𝐼 𝜕𝑅
It can also be written as, = 𝑅 𝑑𝑥 . . (8)
𝜕𝑎 𝜕𝑎

For domain (1):

/
𝜕𝐼 𝜕𝑅
= 𝑅 𝑑𝑥 . . (9)
𝜕𝑎 𝜕𝑎

We know that, 𝑅 = −6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥

𝜕𝑅
= −6 𝑥
𝜕𝑎
𝜕𝑅
Substitue R and value in equation (9),
𝜕𝑎
/
𝜕𝐼
 = (−6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥 ) × (−6 𝑥) 𝑑𝑥
𝜕𝑎

The requirement is, =0

/

 (−6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥 ) × (−6 𝑥) 𝑑𝑥 = 0

1.70 Basic of Finite Element Method
/

 (36 𝑎 𝑥 + 120 𝑎 𝑥 + 3000 𝑥 ) 𝑑𝑥 = 0

/ / /
𝑥 𝑥 𝑥
 36 𝑎 + 120 𝑎 − 3000 =0
3 5 4

36 𝑎 1 120 𝑎 1 3000 1
 −0 + −0 − −0 =0
3 2 5 2 4 2

1 1 1
 12a × + 24 𝑎 × − 750 × =0
8 32 16

 1.5 a + 0.75 𝑎 = 46.875

 a + 0.5 𝑎 = 31.25 . . . (10)

For domain (2):
/
𝜕𝐼 𝜕𝑅
= 𝑅 𝑑𝑥 . . (11)
𝜕𝑎 𝜕𝑎

We know that, 𝑅 = −6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥

𝜕𝑅
= −20 𝑥
𝜕𝑎
𝜕𝑅
Substitue R and value in equation (11),
𝜕𝑎

𝜕𝐼
 (−6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥 ) × (−20 𝑥 )
𝜕𝑎
/

The requirement is, =0

 (−6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 ) × (−20 𝑥 ) = 0

/
 Finite Element Analysis 1.71

 120 𝑎 𝑥 + 400 𝑎 𝑥 − 10000 𝑥 = 0

/

𝑥 𝑥 𝑥
 120 𝑎 + 400 𝑎 − 10000 =0
5 /
7 /
6 /

120 𝑎 1 400 𝑎 1
 (1) − + (1) −
5 2 7 2

10000 1
− (1) − −0 =0
6 2

 24a [1 − 0.03125] + 57.142 a [1 − 0.00781] − 1666.66[1 − 0.01562] = 0

 23.25 a + 56.695 𝑎 − 1640.626 = 0

 a + 2.438 𝑎 = 70.564 . . . (12)

Solving equation (10) and (12),
−𝑎 − 0.5 𝑎 = −31.25

𝑎 + 2.438 𝑎 = 70.564

1.938 𝑎 = 39.314

𝑎 = 20.28

Substituting 𝑎 value in equation (10) or (12).

(10)  a + 0.5(20.28) = 31.25

a + 10.14 = 31.25

 a = 21.11

Hence the trial function is, y = 21.11 (x- x3)+ 20.28(x- x5)
(iv) Galerkin’s method: In this method, the trial function itself is considered as the
weighting function , wi
1.72 Basic of Finite Element Method

 𝑤 𝑅 𝑑𝑥 = 0

𝐅𝐨𝐫 𝐝𝐨𝐦𝐚𝐢𝐧 (𝟏): 𝑤 𝑅 𝑑𝑥 = 0 … (13)

Here, the trial function is, 𝑦 = 𝑤 = 𝑥 − 𝑥

Residual, R value is, R = −6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥
Substitute 𝑤 𝑎𝑛𝑑 𝑅 Values in equation (13),
/

 (𝑥 − 𝑥 )(−6 𝑎 𝑥 + 20 𝑎 𝑥 + 500 𝑥 ) 𝑑𝑥 = 0

 [−6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥 + 6 𝑎 𝑥 + 20 𝑎 𝑥 − 500 𝑥 ] 𝑑𝑥 = 0

/
𝑥 𝑥 𝑥 𝑥 𝑥
 −6𝑎 − 20 𝑎 + 500 +6𝑎 + 20 𝑎
3 5 4 5 7
/
𝑥
− 500 =0
6

1 1 1 1
 − 2𝑎 − 0 − 4𝑎 − 0 + 125 − 0 + 1.2 𝑎 −0
2 2 2 2
1 1
+ 2.857 a − 0 − 83.33 −0 =0
2 2

 − 0.25 a − 0.125 𝑎 + 7.8125 + 0.0375 𝑎 + 0.0223 𝑎 − 1.299 = 0

 − 0.2125a − 0.1027 𝑎 + 6.5135 = 0

 0.2125 a + 0.1027 𝑎 = 6.5135

 Finite Element Analysis 1.73

 a + 0.4832 𝑎 = 30.651 … (14)

𝐅𝐨𝐫 𝐝𝐨𝐦𝐚𝐢𝐧 (𝟐): 𝑤 𝑅 𝑑𝑥 = 0 … (15)

/

Here, the trial function is, 𝑦 = 𝑤 = (𝑥 − 𝑥 )

R = −6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥
Substitute 𝑤 𝑎𝑛𝑑 𝑅 Values in equation (15),

 (𝑥 − 𝑥 ) × (−6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥 ) 𝑑𝑥 = 0
/

 −6 𝑎 𝑥 − 20 𝑎 𝑥 + 500 𝑥 + 6 𝑎 𝑥 + 20 𝑎 𝑥 − 500 𝑥 = 0
/

𝑥 𝑥 𝑥 𝑥
 −6𝑎 − 20 𝑎 + 500 +6𝑎
3 /
5 /
4 /
5 /

𝑥 𝑥
+ 20 𝑎 − 500 =0
9 /
8 /

1 1 1 1
 − 2𝑎 1 − −4𝑎 1− + 125 1 − + 0.857 𝑎 1 −
2 2 2 2
1 1
+ 2.22 a 1− − 62.5 1 − =0
2 2

 − 1.75 a − 3.875 𝑎 + 117.187 + 0.850 𝑎 + 2.215 𝑎 − 62.225 = 0

 − 0.9 a − 1.659 𝑎 + 54.932 = 0

 0.9 a + 1.659 𝑎 = 54.932

 a + 1.843 𝑎 = 61.035 … (16)

Solving equation (14) and (16)

1.74 Basic of Finite Element Method

−𝑎 − 0.4832 𝑎 = −30.651
𝑎 + 1.843 𝑎 = 61.035
1.3598 𝑎 = 30.384
𝑎 = 22.34
Substituting 𝑎 value in equation (16),
 a + 1.843(22.34) = 61.035

a + 41.173 = 61.035

 a = 19.862

Hence the trial function is, y = 19.862 (x- x3)+ 22.34(x- x5)
Result: 𝑎 𝑎
(i) Point collocation: 18.53 25
(ii) Subdomain collocation: 18.50 22.23
(iii) Least squares method: 21.11 20.28
(iv) Galerkin’s method: 19.862 22.34

Example 1.16
The following differential equation is available for a physical phenomenon:
𝑑 𝑦
− 10 𝑥 = 5; 0 ≤ 𝑥 ≤ 1
𝑑𝑥
The boundary conditions are: y(0) = 0
y(1) = 0
By using Galerkin’s method of weighted residuals to find an approximate solution of the
above differential equation and also compare with exact solution.
Given: Differential equation,
𝑑 𝑦
− 10 𝑥 = 5 . . (1)
𝑑𝑥
boundary conditions are: y(0) = 0
 Finite Element Analysis 1.75

y(1) = 0
To find : Approximate solution by using Galerkin’s method.
Solution: We know that, a single trial function, which satisfies the stated boundary
condition is,
𝑦 = 𝑎 𝑥(𝑥 − 1) = 𝑎 (𝑥 − 𝑥)
𝑑𝑦
 = 𝑎 (2𝑥 − 1)
𝑑𝑥
𝑑 𝑦
 = 𝑎 (2 − 0)
𝑑𝑥
𝑑 𝑦
= 2𝑎
𝑑𝑥

Substitute value in given differential equation (1),

 Residual, 𝑅 = 2𝑎 − 10 𝑥 − 5
In Galerkin’s method, the trial function itself is considered as the weighting function, wi.

 𝑤 𝑅 𝑑𝑥 = 0 … (2)

Here, the trial function is, 𝑦 = 𝑤 = 𝑎 𝑥(𝑥 − 1)

Substitute 𝑤 𝑎𝑛𝑑 𝑅 Values in equation (2),

 𝑎 𝑥(𝑥 − 1) × (2 𝑎 − 10 𝑥 − 5) 𝑑𝑥 = 0
/

 a [2 𝑎 𝑥 − 10 𝑥 − 5 𝑥 − 2 𝑎 𝑥 + 10 𝑥 + 5 𝑥]𝑑𝑥 = 0
/

𝑥 𝑥 𝑥 𝑥 𝑥 𝑥
 2𝑎 − 10 −5 −2𝑎 + 10 +5 =0
3 5 3 2 4 2
1.76 Basic of Finite Element Method
10 5 2𝑎 10 5
 2 𝑎 [1 − 0] − [1 − 0] − [1 − 0] − [1 − 0] + [1 − 0] + [1 − 0]
5 3 2 4 2
=0

2a
 − 2 − 1.666 − 𝑎 + 2.5 + 2.5 = 0
3

 0.666 a − 𝑎 + 1.3342 = 0

 − 0.334 a = −1.334

 a =4

So, the approximate solution is obtained as,

y = 4 x(x - 1) …(3)
we can compare this result with exact solution.
𝑑 𝑦
= 10 𝑥 + 5 (𝑔𝑖𝑣𝑒𝑛)
𝑑𝑥

𝑑𝑦 𝑑 𝑦 10 𝑥
 = = +5𝑥+𝑐
𝑑𝑥 𝑑𝑥 3

𝑑𝑦 10 𝑥 5𝑥
 𝑦= = + +𝑐 𝑥+𝑐
𝑑𝑥 3 ×4 2

 𝑦 = 0.833 𝑥 + 2.5 𝑥 + 𝑐 𝑥 + 𝑐 … (4)

Apply boundary conditions, when x = 0, y = 0

(4)  0 = c2
 c2 = 0
When x =1, y = 0
(4)  0 = 0.833 + 2.5 + c1 + c2
0 = 3.333 + c1 + 0 [∵ 𝑐 = 0]
 c1 = -3.333
 Finite Element Analysis 1.77

Substitute c1 and c2 values in equation (4),

 𝑦 = 0.833 𝑥 + 2.5 𝑥 − 3.333 𝑥 … (5)
Result: Approximate solution, y = 4 x (x -1)
Exact solution, 𝑦 = 0.833 𝑥 + 2.5 𝑥 − 3.333 𝑥

Example 1.17
The following differential equation of a physical phenomenon is given by
𝑑 𝑦
− 10 𝑥 = 5
𝑑𝑥
Obtain two term Galerkin solution by using the trial functions:
𝑁 (𝑥) = 𝑥(𝑥 − 1); 𝑁 (𝑥) = 𝑥 (𝑥 − 1); 0 ≤ 𝑥 ≤ 1
The boundary conditions are: y(0) = 0
y(1) = 0
Given: Differential equation,
𝑑 𝑦
− 10 𝑥 = 5 . . . (1)
𝑑𝑥
Trial functions, 𝑁 (𝑥) = 𝑥(𝑥 − 1); 𝑁 (𝑥) = 𝑥 (𝑥 − 1)
 𝑦 = 𝑎 𝑥 (𝑥 − 1) + 𝑎 𝑥 (𝑥 − 1) . . . (1)
boundary conditions are: y(0) = 0
y(1) = 0
To find: Approximate solution by using Galerkin’s method.
Solution: First we have to verify, whether the trial function satisfies the boundary
conditions or not.
Trial function 𝑦 = 𝑎 𝑥 (𝑥 − 1) + 𝑎 𝑥 (𝑥 − 1)
When x = 0, y = 0
x =1, y = 0
Hence its satisfies the boundary conditions.
Residual, R: 𝑦 = 𝑎 𝑥(𝑥 − 1) + 𝑎 𝑥 (𝑥 − 1)
1.78 Basic of Finite Element Method

𝑦 = 𝑎 (𝑥 − 𝑥) + 𝑎 (𝑥 − 𝑥 )

𝑑𝑦
 = 𝑎 (2𝑥 − 1) + 𝑎 (3 𝑥 − 2 𝑥)
𝑑𝑥

𝑑 𝑦
 = 𝑎 (2 − 0) + 𝑎 (6𝑥 − 2)
𝑑𝑥

𝑑 𝑦
= 2𝑎 + 6 𝑎 𝑥 − 2 𝑎
𝑑𝑥

Substitute value in given differential equation (1),

 Residual, 𝑅 = 2𝑎 + 6 𝑎 𝑥 − 2 𝑎 − 10 𝑥 − 5 . . . (3)
In Galerkin’s method, the trial function itself is considered as the weighting
function, wi.

 𝑤 𝑅 𝑑𝑥 = 0

Or we can write,

𝑥(𝑥 − 1)(2 𝑎 + 6 𝑎 𝑥 − 2 𝑎 − 10 𝑥 − 5) 𝑑𝑥 = 0 … (4)

and

𝑥(𝑥 − 1)(2 𝑎 + 6 𝑎 𝑥 − 2 𝑎 − 10 𝑥 − 5) 𝑑𝑥 = 0 … (5)

Solving equation (4),

(𝑥 − 𝑥)(2 𝑎 + 6 𝑎 𝑥 − 2 𝑎 − 10 𝑥 − 5) 𝑑𝑥 = 0
 Finite Element Analysis 1.79

 [2 𝑎 𝑥 + 6𝑎 𝑥 − 2𝑎 𝑥 − 10𝑥 − 5𝑥 − 2 𝑎 𝑥 − 6𝑎 𝑥 + 2 𝑎 𝑥 + 10𝑥

+ 5𝑥]𝑑𝑥 = 0

𝑥 𝑥 𝑥 𝑥 𝑥 𝑥
 2𝑎 +6𝑎 −2𝑎 − 10 −5 − 2𝑎
3 4 3 5 3 2

𝑥 𝑥 𝑥 𝑥
−6𝑎 +2𝑎 + 10 +5 =0
3 2 4 2

2𝑎 6𝑎 2𝑎 10 5 2 𝑎 6𝑎 2𝑎 10 5
 + − − − − − + + + =0
3 4 3 5 3 2 3 2 4 2

 0.666 a + 1.5𝑎 − 0.666𝑎 − 2 − 1.666 − 𝑎 − 2 𝑎 + 𝑎 + 2.5 + 2.5 = 0

 0.333 a − 0.1666 𝑎 + 1.334 = 0

 − 0.333 a + 0.1666 𝑎 = 1.334

 a + 0.50 𝑎 = 4 . . (6)
Solving equation (5),

(𝑥 − 𝑥)(2 𝑎 + 6 𝑎 𝑥 − 2 𝑎 − 10 𝑥 − 5) 𝑑𝑥 = 0

 (𝑥 − 𝑥 )(2 𝑎 + 6 𝑎 𝑥 − 2 𝑎 − 10 𝑥 − 5) 𝑑𝑥 = 0

 2 𝑎 𝑥 + 6𝑎 𝑥 − 2𝑎 𝑥 − 10𝑥 − 5𝑥 − 2 𝑎 𝑥 − 6𝑎 𝑥 + 2 𝑎 𝑥 + 10𝑥

+ 5𝑥 = 0
1.80 Basic of Finite Element Method

𝑥 𝑥 𝑥 𝑥 𝑥 𝑥
 2𝑎 +6𝑎 −2𝑎 − 10 −5 − 2𝑎
4 5 4 6 4 3

𝑥 𝑥 𝑥 𝑥
−6𝑎 +2𝑎 + 10 +5 =0
4 3 5 3

2𝑎 6𝑎 2𝑎 10 5 2 𝑎 6𝑎 2𝑎 10 5
 + − − − − − + + + =0
4 5 4 6 4 3 4 3 5 3

 0.5 a + 1.2𝑎 − 0.5𝑎 − 1.666 − 1.25 − 0.666𝑎 − 1.5𝑎 + 0.666𝑎 + 2

+ 1.666 = 0

 − 0.166 a − 0.133 𝑎 + 0.75 = 0

 − 0.166 a + 0.133 𝑎 = 0.75

 a + 0.8012 𝑎 = 4.518 … (7)

Solving equation (6) and (7)
−𝑎 − 0.50 𝑎 = −4
𝑎 + 0.8012 𝑎 = 4.518
0.3012 𝑎 = 0.518
 𝑎 = 1.719
Substituting 𝑎 value in equation (7),
 a + 0.8012(1.719) = 4.518
a = 3.140
So, the two term approximate solution is,
𝑦 = 3.140 𝑥(𝑥 − 1) + 1.719 𝑥 (𝑥 − 1)
= 3.140 𝑥 − 3.140 𝑥 + 1.719 𝑥 − 1.719 𝑥
𝑦 = 1.719 𝑥 + 1.421 𝑥 − 3.140 𝑥
Result: The two term Galerkin’s approximate solution is,
𝑦 = 1.719 𝑥 + 1.421 𝑥 − 3.140 𝑥
 Finite Element Analysis 1.81

Example 1.18
The differential equation of a physical phenomenon is given by
𝑑 𝑦
+ 𝑦 = 4𝑥, 0 ≤ 𝑥 ≤ 1
𝑑𝑥
The boundary conditions are: y(0) = 0
y(1) = 1
Obtain two term approximate solution by using Galerkin method of weighted residuals.
Given: Differential equation,
𝒅𝟐 𝒚
+ 𝒚 = 𝟒𝒙, 𝟎 ≤ 𝒙 ≤ 𝟏 . . . (1)
𝒅𝒙𝟐
Boundary conditions are: y(0) = 0
y(1) = 1
To find: One term approximate solution by using Galerkin’s method.
Solution: Here, the boundary conditions are not homogeneous. So, we assume a trial
function as,
𝑦 = 𝑎 𝑥 (𝑥 − 1) + 𝑥
First we have to verify, whether the trial function satisfies the boundary conditions or not.
𝑦 = 𝑎 𝑥 (𝑥 − 1) + 𝑥 … (2)

When x = 0, y = 0

x =1, y = 1
Hence its satisfies the boundary conditions.
Residual, R: 𝑦 = 𝑎 𝑥(𝑥 − 1) + 𝑥 = 𝑎 (𝑥 − 𝑥) + 𝑥

𝑑𝑦
 = 𝑎 (2𝑥 − 1) + 1
𝑑𝑥

𝑑 𝑦
 = 𝑎 (2)
𝑑𝑥
1.82 Basic of Finite Element Method

𝑑 𝑦
= 2𝑎
𝑑𝑥

Substitute value in given differential equation.

(1) 2𝑎 +𝑦 = 4𝑥
Substitute y value,
 2 𝑎 + 𝑎 𝑥(𝑥 − 1) + 𝑥 = 4 𝑥
 𝑅𝑒𝑠𝑖𝑑𝑢𝑎𝑙, 𝑅 = 2 𝑎 + 𝑎 𝑥(𝑥 − 1) + 𝑥 − 4 𝑥
In Galerkin’s method, the trial function itself is considered as the weighting function, wi.

 𝑤 𝑅 𝑑𝑥 = 0 … (3)

Here, the trial function is, 𝑦 = 𝑤 = 𝑥(𝑥 − 1)

Substitute 𝑤 and R values in equation (3),

 𝑥(𝑥 − 1)[2 𝑎 + 𝑎 𝑥(𝑥 − 1) + 𝑥 − 4𝑥] 𝑑𝑥 = 0

 𝑥(𝑥 − 1)[2 𝑎 + 𝑎 𝑥 − 𝑎 𝑥 + 𝑥 − 4𝑥] 𝑑𝑥 = 0

 (𝑥 − 𝑥)[2 𝑎 + 𝑎 𝑥 − 𝑎 𝑥 − 3𝑥] 𝑑𝑥 = 0

 [2 𝑎 𝑥 + 𝑎 𝑥 − 𝑎 𝑥 − 3𝑥 − 2 𝑎 𝑥 − 𝑎 𝑥 + 𝑎 𝑥 + 3 𝑥 ]𝑑𝑥 = 0

𝑥 𝑥 𝑥 𝑥 𝑥 𝑥
 2𝑎 +𝑎 −𝑎 −3 − 2𝑎 −𝑎
3 5 4 4 2 4
𝑥 𝑥 𝑥
− 𝑎 + 𝑎 +3 =0
4 3 3
 Finite Element Analysis 1.83
2𝑎 𝑎 𝑎 3 2𝑎 𝑎 𝑎 3
 + − − − − + + =0
3 5 4 4 2 4 3 3
 0.666 a + 0.2𝑎 − 0.25𝑎 − 0.75 − 𝑎 − 0.25 𝑎 − 0.333 𝑎 + 1 = 0
−0.301 a = −0.25
a = 0.830
So, the one term approximate solution is,
𝑦 = 0.830 𝑥(𝑥 − 1) + 𝑥
= 0.830 𝑥 − 0.830 𝑥 + 𝑥
𝑦 = 0.830 𝑥 + 0.17𝑥
Result: The two term Galerkin’s approximate solution is,
𝑦 = 0.830 𝑥 + 0.17 𝑥

Example 1.19
Find the deflection at the centre of a simply supported beam of span length ‘l’
subjected to uniformly distributed load throughout its length as shown in Fig.(i0, using
(a) point collocation method, (b) Sub-domain method, (c) Least squares method, and (d)
Galerkin’s method.

Fig. (i)
Given:
1.84 Basic of Finite Element Method

To find: Deflection at the centre “C” by using

(i) Point collocation method
(ii) Subdomain collocation method
(iii) Least squares method
(iv) Galerkin’s method
Solution: The differential equation governing the deflection of beam subjected to
uniformly distributed load is given by,
𝑑 𝑦
𝐸𝐼 − 𝜔 = 0, 0 ≤ 𝑥 ≤ 1 … (1)
𝑑𝑥
The boundary conditions are y = 0 at x = 0 and x = 1, where y is the deflection,
𝑑 𝑦
𝐸𝐼 = 0 𝑎𝑡 𝑥 = 0 𝑎𝑛𝑑 𝑥 = 1
𝑑𝑥
𝑑 𝑦
Where, 𝐸𝐼 = 𝑀(𝐵𝑒𝑛𝑑𝑖𝑛𝑔 𝑚𝑜𝑚𝑒𝑛𝑡)
𝑑𝑥
E → Young’s modulus.
I → Moment of inertia of the beam.
Let us select the trial function for deflection as,
𝜋𝑥
𝑦 = 𝑎 𝑠𝑖𝑛 … (2)
𝑙
Hence it satisfies the boundary conditions,
𝑑𝑦 𝜋 𝜋𝑥
 = 𝑎 . cos
𝑑𝑥 𝑙 𝑙

𝑑 𝑦 𝜋 𝜋𝑥
 = −𝑎 . sin
𝑑𝑥 𝑙 𝑙

𝑑 𝑦 𝜋 𝜋𝑥
 = −𝑎 . cos
𝑑𝑥 𝑙 𝑙

𝑑 𝑦 𝜋 𝜋𝑥
 =𝑎 . sin … (3)
𝑑𝑥 𝑙 𝑙
 Finite Element Analysis 1.85

Substituting the equation (3) in the governing equation (1),

𝜋 𝜋𝑥
𝐸𝐼 𝑎 . sin −𝜔 =0
𝑙 𝑙
𝜋 𝜋𝑥
Take, Residual, 𝑅=𝐸𝐼𝑎 . sin − 𝜔
𝑙 𝑙
(a) Point collocation method:
In this method, the residuals are set to zero.
𝜋 𝜋𝑥
 𝑅=𝐸𝐼𝑎 . sin − 𝜔=0
𝑙 𝑙

𝜋 𝜋𝑥
𝐸𝐼𝑎 . sin = 𝜔
𝑙 𝑙
To get maximum deflection, take 𝑥 = (i.e., at the center of beam)

𝜋 𝜋𝑥 𝑙
Then, 𝐸𝐼𝑎 . sin = 𝜔
𝑙 𝑙 2

𝜋 𝜋
𝐸𝐼𝑎 = 𝜔 ∵. sin =1
𝑙 2

𝜔 𝑙
𝑎=
𝜋 𝐸𝐼

Substitute “a” value in the trial function equation (2),

𝜔 𝑙 𝜋𝑥
𝑦= sin
𝜋 𝐸𝐼 𝑙

1 𝜔 𝑙 𝜋 𝑙
𝐴𝑡 𝑥 =  𝑦 = sin
2 𝜋 𝐸𝐼 𝑙 2

𝜔 𝑙 𝜋
𝑦 = ∵. sin =1
𝜋 𝐸𝐼 2

𝜔 𝑙
𝑦 =
97.4 𝐸 𝐼
1.86 Basic of Finite Element Method

(b) Sub-domain collocation method:

In this method, the integral of the residual over the sub-domain is set to zero.

𝑅 𝑑𝑥 = 0

Substitute R value,

𝜋 𝜋𝑥
 𝑎𝐸𝐼 sin − 𝜔 𝑑𝑥 = 0
𝑙 𝑙

𝜋𝑥
𝜋 − cos
 𝑎𝐸𝐼 𝑙 −𝜔 𝑥 =0
𝑙 𝜋
𝑙

𝜋 𝜋𝑥 𝜋
 𝑎𝐸𝐼 − cos −𝜔 𝑥 =0
𝑙 𝑙 𝑙

𝜋
 − 𝑎𝐸𝐼 ( cos 𝜋 − cos 0) −𝜔 𝑙 = 0
𝑙
𝜋
 − 𝑎𝐸𝐼 (−1 − 1) = 𝜔 𝑙
𝑙
[ ∵ 𝑐𝑜𝑠𝜋 = −1 ; 𝑐𝑜𝑠 0 = 1]
𝜔𝑙 𝜔𝑙
 𝑎= =
2𝜋 𝐸𝐼 62 𝐸 𝐼
Substitute “a” value in the trial function equation (2),
𝜔𝑙 𝜋𝑥
𝑦= 𝑠𝑖𝑛
62𝐸 𝐼 𝑙

1 𝜔 𝑙 𝜋 𝑙
𝐴𝑡 𝑥 = , 𝑦 = sin
2 62 𝐸 𝐼 𝑙 2

𝜔 𝑙
𝑦 =
62 𝐸 𝐼
 Finite Element Analysis 1.87

(c) Least squares method:

In this method, the functional

𝐼= 𝑅 𝑑𝑥 𝑖𝑠 𝑚𝑖𝑛𝑖𝑚𝑢𝑚.

𝜋 𝜋𝑥
𝐼= 𝑎𝐸𝐼 sin −𝜔 𝑑𝑥
𝑙 𝑙

𝜋 𝜋𝑥 𝜋 𝜋𝑥
= 𝑎 𝐸 𝐼 𝑠𝑖𝑛 + 𝜔 − 2 𝑎 𝐸 𝐼𝜔 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙 𝑙 𝑙

2𝜋𝑥
𝐼 𝜋 1 − 𝑐𝑜𝑠
= 𝑎 𝐸 𝑙 +𝜔
𝑙 2

𝜋 𝜋𝑥
− 2 𝑎 𝐸 𝐼𝜔 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙

2𝜋𝑥
1 − cos 2 𝑥 𝜋𝑥 1 − 𝑐𝑜𝑠
∵ sin 𝑥 =  sin = 𝑙
2 𝑙 2

𝜋 1 2𝜋𝑥 𝑙
= 𝑎 𝐸 𝐼 𝑥 − sin +𝜔 𝑥
𝑙 2 𝑙 2𝜋
𝜋 𝜋𝑥 𝑙
− 2 𝑎 𝐸 𝐼𝜔 − cos
𝑙 𝑙 𝜋

𝜋 1 1
=𝑎 𝐸 𝐼 𝑙− (sin 2𝜋 − sin 0) + 𝜔 𝑙
𝑙 2 2𝜋
𝜋 𝑙
− 2 𝑎 𝐸 𝐼𝜔 . (− cos 𝜋 − 𝑐𝑜𝑠0)
𝑙 𝜋
[∵ 𝑠𝑖𝑛 2𝜋 = 0; sin 0 = 0; cos 𝜋 = 0 ; cos 0 = 1 ]
𝜋 𝑙 𝜋
=𝑎 𝐸 𝐼 . + 𝜔 𝑙 − 2 𝑎 𝐸 𝐼𝜔 (−1 − 1)
𝑙 2 𝑙
1.88 Basic of Finite Element Method

𝜋
𝑎 𝐸 𝐼
= 𝑙 + 𝜔 𝑙 − 4 𝑎 𝐸 𝐼𝜔 𝜋
2𝑙 𝑙
𝜕𝐼
𝑁𝑜𝑤, =0
𝜕𝑎
2𝑎𝐸 𝐼 𝜋 𝜋
= = 4 𝐸 𝐼𝜔
2𝑙 𝑙
2𝑎𝐸 𝐼 𝜋 𝜋
= 4 𝐸 𝐼𝜔
𝑙 𝑙
4𝜔𝑙
∴ 𝑎=
𝜋 𝐸𝐼
Hence the trial function,
4𝜔𝑙 𝜋𝑥
𝑦= sin
𝜋 𝐸𝐼 𝑙
1
At x = , maximum defelction,
2
4𝜔𝑙 𝜋𝑥 𝑙 𝜋
𝑦 = sin ∵ sin =1
𝜋 𝐸𝐼 𝑙 2 2
𝜔𝑙
𝑦 =
76.5 𝐸 𝐼
(d) Galerkin’s method:
In this method,

𝑦 𝑅 𝑑𝑥 = 0.

𝑥  𝑥
 𝑎 𝑠𝑖𝑛 𝑎𝐸𝐼 𝑠𝑖𝑛 − 𝜔 𝑑𝑥 = 0
𝑙 𝑙 𝑙

 𝑥 𝑥
 𝑎 𝐸𝐼 sin − 𝑎 𝜔 sin 𝑑𝑥 = 0
𝑙 𝑙 𝑙
 Finite Element Analysis 1.89

𝜋 1 2𝜋𝑥 𝜋𝑥
= 𝑎 𝐸𝐼 1 − cos − 𝑎 𝜔 𝑠𝑖𝑛 𝑑𝑥 = 0
𝑙 2 𝑙 𝑙

𝜋 1 1 2𝜋𝑥 𝑙 𝜋𝑥
 𝑎 𝐸𝐼 𝑥− sin +𝑎𝜔 cos =0
𝑙 2 2𝜋 𝑙 𝜋 𝑙

𝜋 1 𝑙
𝑎 𝐸𝐼 −2𝑎𝜔 =0
𝑙 2 𝜋
2𝜔𝑙 2 𝑙
∴ 𝑎= .
𝜋 𝐸𝐼𝜋
4𝜔𝑙
𝑎=
𝜋 𝐸𝐼
Hence, the trial function,
4𝜔𝑙 𝜋𝑥
𝑦= . 𝑠𝑖𝑛
𝜋 𝐸𝐼 𝑙
1
At x = , maximum deflection,
2
4𝜔𝑙 𝜋𝑥 𝑙 4𝜔𝑙
𝑦 = 𝑠𝑖𝑛 =
𝜋 𝐸𝐼 𝑙 2 𝜋 𝐸𝐼
𝜔𝑙
𝑦 =
76.5 𝐸 𝐼
Verification: we know that, simply supported beam is subjected to uniformly distributed
load, maximum deflection is,
5 𝜔𝑙 𝜔𝑙
𝑦 = = 0.01
384 𝐸 𝐼 𝐸𝐼

Result: Maximum deflection at 𝑥 =

(a) Point collocation method:

𝜔 𝑙 𝜔 𝑙
𝑦 = = 0.01
97.4 𝐸 𝐼 𝐸𝐼
1.90 Basic of Finite Element Method

(b) Sub-domain collocation method:

𝜔 𝑙 𝜔 𝑙
𝑦 = = 0.01
62 𝐸 𝐼 𝐸𝐼
(c) Least squares method:
𝜔 𝑙 𝜔 𝑙
𝑦 = = 0.01
62 𝐸 𝐼 𝐸𝐼
(d) Galerkin’s method:
𝜔𝑙 𝜔 𝑙
𝑦 = = 0.01
76.5 𝐸 𝐼 𝐸𝐼

Example 1.20
Find the deflection at the centre of a clamped beam subject to uniformly
distributed load throughout its length as shown in Fig. (i). use point collection method.
Take trial function as 𝑦 = 𝑎(𝑥 − 2𝑙𝑥 + 𝑙 𝑥 ).

Fig, (i)
Given:

Fig, (ii)
Solution: The differential equation governing the deflection of the beam subjected to udl
is given by
𝑑 𝑦
𝐸𝐼 − 𝜔 = 0, 0 ≤ 𝑥 ≤ 𝑙 … . (1)
𝑑𝑥
 Finite Element Analysis 1.91

Since this is the clamped beam (i.e., fixed end beam), the boundary conditions are,
Deflection, y = 0 at x = 0 and x =1
𝑑𝑦
Slope, θ = 𝐸𝐼 = 0 𝑎𝑡 𝑥 = 0 𝑎𝑛𝑑 𝑥 = 1
𝑑𝑥
The trial function should satisfy the above boundary conditions
We know that,
Trial function, 𝑦 = 𝑎(𝑥 − 2𝑙𝑥 + 𝑙 𝑥 ) …(2)

𝑑𝑦
⟹ = 𝑎(5𝑥 − 8𝑙𝑥 + 3𝑙 𝑥 )
𝑑𝑥

𝑑 𝑦
⟹ = 𝑎(20𝑥 − 24𝑙𝑥 + 6𝑙 𝑥)
𝑑𝑥

𝑑 𝑦
⟹ = 𝑎(60𝑥 − 48𝑙 + 6𝑙 )
𝑑𝑥

𝑑 𝑦
⟹ = 𝑎(120𝑥 − 48𝑙)
𝑑𝑥

Substituting value in equation (1), we get

𝑎 𝐸𝐼 (120𝑥 − 48𝑙) − 𝜔 = 0
Take Residual, 𝑅 = 𝑎 𝐸𝐼 (120𝑥 − 48𝑙) − 𝜔
In point collocation method, the residual is set to zero
i.e., 𝑎 𝐸𝐼 (120𝑥 − 48𝑙) − 𝜔 = 0

To get maximum deflection, take x =

𝑙
⟹ 𝑎 𝐸𝐼 120 × − 48𝑙 = 𝜔
2

120𝑎 𝐸𝐼𝑙 = 𝜔
𝜔
𝑎 =
120 𝐸𝐼𝑙
1.92 Basic of Finite Element Method
𝜔
Hence, the trial function 𝑦 = (𝑥 − 2𝑙𝑥 + 𝑙 𝑥 )
120 𝐸𝐼𝑙

At x = , Maximum deflection,
𝜔 𝑙 𝑙 𝑙
𝑦 = − 2𝑙 +𝑙
120 𝐸𝐼𝑙 2 2 2

𝜔𝑙
𝑦 =
384𝐸𝐼

Result: Maximum deflection at x =

𝜔𝑙
𝑦 =
384𝐸𝐼

1.13. THE GENERAL WEIGHTED RESIDUAL STATEMENT

After understanding the basic technique and successfully solved a few problems,
the general weighted residual statement can be written as

𝑤 𝑅 𝑑𝑥 = 0 𝑓𝑜𝑟 𝑖 = 1, 2, … . . , 𝑛 … (1.13)

Where 𝑤 = 𝑁 … (1.14)
The better result will be obtained by considering more terms in polynomial and
trigonometric series.
SOLVED PROBLEMS- GENERAL WEIGHTED RESIDUAL METHOD
Example 1.21
Consider a 1 mm diameter, 50 mm long aluminium pin-fin as shown in Fig. (i)
used to enhance the heat transfer from a surface wall maintained at 300oC. the governing
differential equation and the boundary conditions are given by.

Fig, (i)
 Finite Element Analysis 1.93

𝑑 𝑇 𝑃ℎ
𝐾 = (𝑇 − 𝑇 )
𝑑𝑥 𝐴
𝑇(0) = 𝑇 = 300 𝐶
dT
(L) = 0(insulated tip)
dx
Where, K = Thermal conductivity
P = Perimeter
A = Cross – sectional area
h = Convective heat transfer coefficient
𝑇 = Wall temperature
𝑇 = Ambient temperature
Let, K = 200 W/moC for aluminium, h = 20 W/m2 oC, 𝑇 = 30 𝐶. Estimate the
temperature distribution in the fin using the Galerkin weighted residual method.
Given: Diameter, d = 1 mm = 1  10-3 m
Length, L = 50 = 50  10-3 m
Wall temperature, Tw = 300oC
Governing differential equation,
𝑑 𝑇 𝑃ℎ
𝐾 = (𝑇 − 𝑇 )
𝑑𝑥 𝐴
𝑇(0) = 𝑇 = 300 𝐶
dT
(L) = 0(insulated tip)
dx
Thermal conductivity, K = 200 W/m oC
Heat transfer coefficient, h = 20 W/m2 oC
Ambient temperature, 𝑇 = 30oC
To find: Temperature distribution using Galerkin method.
Solution: Assume a trial solution. Let,
𝑇(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 …(1)
1.94 Basic of Finite Element Method

The boundary conditions are

𝑇(0) = 𝑇 = 300 𝐶 …(a)
dT
(L) = 0 … (b)
dx
From equation (a)  x = 0, T = 300oC
Applying these values in equation (1),
 𝑎 = 300
From equation (b) 
dT
𝑥 = 𝐿, =0
dx
Differentiate equation (1),
dT
= 𝑎 + 𝑎 2𝑥 … (2)
dx

0 = 𝑎 + 𝑎 (2𝐿)

𝑎 = −2𝐿𝑎
Substitute 𝑎 and 𝑎 value in equation (1),
𝑇(𝑥) = 300 + (−2𝐿𝑎 )𝑥 + 𝑎 𝑥

𝑇(𝑥) = 300 + 𝑎 (𝑥 − 2𝐿𝑥) …(3)

We know that
𝑑 𝑇 𝑃ℎ
𝐾 = (𝑇 − 𝑇 )
𝑑𝑥 𝐴

𝑑 𝑇 𝜋(1 × 10 ) × 20
200 = 𝜋 (𝑇 − 30) [∵ 𝑃 = 𝜋𝐷]
𝑑𝑥 (1 × 10 )
4

𝑑 𝑇
= 400(𝑇 − 30) … (4)
𝑑𝑥
Substitute ‘T’ value from equation (3)
 Finite Element Analysis 1.95

𝑑 𝑇
= 400[(300 + 𝑎 (𝑥 − 2𝐿𝑥) − 30)]
𝑑𝑥
= 400[270 + 𝑎 (𝑥 − 2𝐿𝑥)] … (5)
From equation (2),
dT
= 𝑎 + 𝑎 2𝑥
dx
dT
 = 2𝑎 … (6)
dx
Substitute the equation (6) in equation (5),

2𝑎 = 400[270 + 𝑎 (𝑥 − 2𝐿𝑥)]

2𝑎 − 400[270 + 𝑎 (𝑥 − 2𝐿𝑥)] = 0

Take, Residual, 𝑅 = 2𝑎 − 400[270 + 𝑎 (𝑥 − 2𝐿𝑥)] … (8)

To minimize the residual, take weight function as

W(x) = x2 – 2 Lx

 (𝑥 − 2𝐿𝑥)𝑅 𝑑𝑥 = 0 … (9)

Substitute the equation (8) in equation (9),

(𝑥 − 2𝐿𝑥)[2 𝑎 − 400(270 + 𝑎 (𝑥 − 2 𝐿 𝑥)]𝑑𝑥 = 0

(𝑥 − 2𝐿𝑥)[(2 𝑎 − 400𝑎 𝑥 + 800𝑎 𝐿𝑥]𝑑𝑥 = 0

 [(2 𝑎 𝑥 − 108000 − 400𝑎 𝑥 + 800𝑎 𝐿𝑥 − 4𝑎 𝑙𝑥 + 21600 𝑙𝑥

+ 800𝑎 𝑙𝑥 − 1600𝑎 𝑙 𝑥 )]𝑑𝑥 = 0

1.96 Basic of Finite Element Method

𝑥 𝑥 𝑥 𝑥 𝑙𝑥 𝑙𝑥
 2𝑎 − 108000 − 400𝑎 + 800𝑎 𝑙 − 4𝑎 + 21600
3 3 5 4 2 2
𝑙𝑥 𝑙 𝑥
+ 800𝑎 − 1600𝑎 =0
4 3

𝑙 𝑙 1 𝑙 𝑙 𝑥 𝑙
 2𝑎 − 108000 − 400𝑎 + 800𝑎 − 4𝑎 + 21600
3 3 5 4 3 2
𝑙 𝑙
+ 800𝑎 − 1600𝑎 =0
4 3
2𝑎 10800 𝑎 𝑙 𝑎 𝑙 4𝑎 21600 𝑙
 − − 400 + 800 − + + 800𝑎
3 3 5 4 2 2 4
𝑙
− 1600𝑎 =0
3
1600𝑙
 𝑎 [0.6667 − 80 𝑙 + 200𝑙 − 2 + 200 𝑙 − = −72000
3
Substitute L = 50  10-3 m
 𝑎 [0.6667 − 0.2 + 0.5 − 2 + 0.5 − 1.3333 = −72000
 𝑎 [−1.8666] = −72000
𝑎 = 38572.80
Galerkin solution, T(x) = 300 + 38572.80 (x2 – 2 L x )
Result: Galerkin solution, T(x) = 300 + 38572.80 (x2 – 2 L x )

1.14 VARIATIONAL (WEAK) FORM OF THE WEIGHTED

RESUDUAL STATEMENT
We know that the general weighted residual statement is,

𝑤 𝑅 𝑑𝑥 = 0 … (1.15)

In this variational method, integration is carried out by parts. It reduces the

continuity requirement on the trial function assumed in the solution. So, it is referred to
as the week form. In this method, it is possible to have a wider choice of trial functions.
 Finite Element Analysis 1.97

1.15 COMPARISON OF DIFFERENTIAL EQUATION, WEIGHTED

RESIDUAL STATEMENT AND WEAK FORMULATION OF
WEIGHTED RESUDUAL STATEMENT
1.15.1 Differential Equation
Consider a uniform rod subjected to uniform axial load 𝑞 as shown in Fig. 1.26

qo

Fig, 1.26 Uniform rod

The deformation of the bar is governed by the differential equation,
𝑑 𝑢
𝐴𝐸 +𝑞 = 0 … (1.16)
𝑑𝑥

With the boundary conditions,

u (0) = 0

𝑑𝑢
𝐴𝐸 =𝑃 … (1.17)
𝑑𝑥

1.15.2 Weighted Residual Statement

In order to find the solution for the aforementioned problem, the weighted
residual statement can be developed as follows:

𝑑 𝑢
𝑤(𝑥) 𝐴 𝐸 + 𝑞 𝑑𝑥 = 0 … (1.18)
𝑑𝑥

With the boundary conditions,

u (0) = 0

𝑑𝑢
𝐴𝐸 =𝑃 … (1.19)
𝑑𝑥
1.98 Basic of Finite Element Method

1.15.3 Observations on the Weighted Residual Statement

 Weighted residual statement can be developed for any form of differential
equations like linear, non-linear, ordinary, partial, etc.
 The weighted residual statement is developed only for differential equation and it
is not suitable for boundary conditions.
 The trail solution should satisfy all the boundary conditions and it should be
differentiable as many times as needed in the original differential equation.

1.15.4 Weak form of Weighted Residual Statement

By performing integration by parts, the weak form of weighted residual statement
of the aforementioned problem is obtained as follows:

𝑑𝑢 𝑑𝑢 𝑑𝑤
𝑤(𝑥)𝐴 𝐸 − 𝐴𝐸 . . 𝑑𝑥 + 𝑤(𝑥) 𝑞𝑑𝑥 = 0 … (1.20)
𝑑𝑥 𝑑𝑥 𝑑𝑥

With the boundary conditions,

u (0) = 0

𝑑𝑢
𝐴𝐸 =𝑃
𝑑𝑥

When we apply the boundary conditions, the equation (1.20) reduces to

𝑑𝑢 𝑑𝑤
𝐴𝐸 . . 𝑑𝑥 = 𝑤(𝑥) 𝑞𝑑𝑥 + 𝑤(𝑙)𝑃 … (1.21)
𝑑𝑥 𝑑𝑥

𝟏.15.5 Observations on the Weak Form

 A much wider choice of trial functions can be used.
 The weak form can be developed for any higher order differential equation.
 Natural boundary conditions are directly applied in the differential equation.
 The trial solution should satisfy the essential boundary conditions.
 Finite Element Analysis 1.99

1.16 PRINCIPLE OF STATIONARY TOTAL POTENTIAL (PSTP)

1.16.1. Potential Energy in Elastic Bodies
Potential energy is the capacity to do the work by the forces acting on deformable
bodies. The forces acting on a body may be classified as external forces and internal
forces. External forces are the applied loads while internal forces are the stresses
developed in the body. Hence the total potential energy is the sum of internal and external
potential energies.
Consider a spring-mass system as shown in Fig. 1.27. Let its stiffness (load per
unit deflection) be k and length L. Due to a force P let it extend by u.

Fig. 1.27
The load P moves down by distance u. Hence it loses its capacity to do work by
P u. the external potential energy in this case is given by,
H=-Pu …(1.22)
When the load has reached equilibrium position after extension of spring by u, the
force in spring is K u. But when extension was zero the resisting force was also zero.

Hence, the average force =

The energy stored in the spring due to strain

= Average force  Deflection
𝐾𝑢
= 𝑢
2
1
= K𝑢
2
1.100 Basic of Finite Element Method

 Total potential energy in the spring,

1
𝜋= K𝑢 −𝑃𝑢 … (1.23)
2

1.16.2 Principle of Minimum Potential Energy

From the expression for total potential energy,
𝜋 =𝑈+𝐻 … (1.24)
We know that, Potential energy of external force H is equal but opposite to total
virtual work done by external force.
Thus, 𝐻 = −𝐻

∴ 𝛿𝐻 = −𝛿𝐻

⇒ 𝛿𝜋 = 𝛿𝑈 − 𝛿𝐻 … . (1.25)

In, principle of virtual work, 𝛿𝑈 = 𝛿𝐻

∴ 𝛿𝜋 = 0
Hence, we can conclude that a deformable body is in equilibrium when the
potential energy is having stationary value.
Hence, the principal of minimum potential energy states “Among all the
displacement equations that internal compatibility and the boundary condition those that
also satisfy the equations of equilibrium make the potential energy a minimum is a stable
system.”

1.17 SOLVED PROBLESM – POTENTIAL ENERGY APPROCH

Example 1.22
The spring assembly is shown in Fig (i). assemble the finite element equation by
using direct approach and potential energy approach.

Fig. (i)
 Finite Element Analysis 1.101

Given:

Fig. (ii)
To find: Global stiffness matrix for the spring system.
Solution: Consider the free body diagram of nodes 1, 2 and 3 as shown in Fig. (iii). Let
the displacement of nodes by 𝑢 , 𝑢 and 𝑢 . The extension of spring 1, 2 and 3 are,
We know that, Displacement,
𝛿 = 𝑢 , 𝛿 = 𝑢 − 𝑢 , 𝛿 = 𝑢 − −𝑢 … (1)
The equilibrium equations are (from Fig. (iii))
−𝑘 𝛿 + 𝑘 𝛿 + 𝐹 = 0 … (2)

−𝑘 𝛿 + 𝑘 𝛿 + 𝐹 = 0 … (3)

−𝑘 𝛿 + 𝐹 = 0 … (4)

Substitute 𝛿 , 𝛿 and 𝛿 values in equations (2), (3) and (4).

Equation (2)  −𝑘 𝑢 + 𝑘 (𝑢 − 𝑢 ) = −𝐹

𝑘 𝑢 − (𝑢 − 𝑢 ) = −𝐹

(𝑘 + 𝑘 )𝑢 − 𝑘 𝑢 = −𝐹 …(5)

Equation (3)  −𝑘 (𝑢 − 𝑢 ) + 𝑘 (𝑢 − 𝑢 ) = −𝐹

𝑘 (𝑢 − 𝑢 ) − 𝑘 (𝑢 − 𝑢 ) = 𝐹

−𝑘 𝑢 + (𝑘 + 𝑘 )𝑢 − 𝑘 𝑢 = −𝐹 … (6)

Equation (4)  −𝑘 (𝑢 − 𝑢 ) = −𝐹

𝑘 (𝑢 − 𝑢 ) = 𝐹
1.102 Basic of Finite Element Method

−𝑘 𝑢 + 𝑘 𝑢 = 𝐹 … . (7)
Arranging equations (5), (6) and (7) in matrix form
(𝑘 + 𝑘 ) −𝑘 0 𝑢 𝐹
−𝑘 (𝑘 + 𝑘 ) −𝑘 𝑢 = 𝐹 … (8)
0 −𝑘 𝑘 𝑢 𝐹

Now, let us see the potential energy approach. Total potential energy in the system is,
1 1 1
𝜋 = 𝑘 𝛿 + 𝑘 𝛿 + 𝑘 𝛿 −𝐹 𝑢 −𝐹 𝑢 −𝐹 𝑢
2 2 2
1 1 1
= 𝑘 𝑢 + 𝑘 (𝑢 − 𝑢 ) + 𝑘 (𝑢 − 𝑢 ) − 𝐹 𝑢 − 𝐹 𝑢
2 2 2
−𝐹 𝑢
∂π
Apply, =0
∂u

 −𝑘 𝑢 + 𝑘 (𝑢 − 𝑢 )(−1) − 𝐹 = 0

−𝑘 𝑢 − 𝑘 (𝑢 − 𝑢 ) − 𝐹 = 0

(𝑘 + 𝑘 )𝑢 − 𝑘 𝑢 = 𝐹 …(9)
∂π
Apply, =0
∂u

 𝑘 (𝑢 − 𝑢 ) + 𝑘 (𝑢 − 𝑢 )(−1) = 𝐹

−𝑘 𝑢 − (𝑘 + 𝑘 )𝑢 − 𝑘 𝑢 = 𝐹 … (10)

∂π
Apply, =0
∂u

 𝑘 (𝑢 − 𝑢 ) − 𝐹 = 0

−𝑘 𝑢 + 𝑘 𝑢 = 𝐹 … (11)
Equation (9), (10) and (11) in matrix form,
 Finite Element Analysis 1.103

(𝑘 + 𝑘 ) −𝑘 0 𝑢 𝐹
−𝑘 (𝑘 + 𝑘 ) −𝑘 𝑢 = 𝐹 … (12)
0 −𝑘 𝑘 𝑢 𝐹

Result: Finite element equation

(𝑘 + 𝑘 ) −𝑘 0 𝑢 𝐹
−𝑘 (𝑘 + 𝑘 ) −𝑘 𝑢 = 𝐹
0 −𝑘 𝑘 𝑢 𝐹

Example 1.23
Determine the displacements of nodes 1 and 2 in the spring system shown in Fig.
(i). use minimum of potential energy principle to assemble equations of equilibrium.

Fig. (i)
Given:

Fig. (ii)
𝑘 = 60 N/m, 𝐹 = 100 N
𝑘 = 75 N/m, 𝐹 = 80 N
𝑘 = 100 N/m
1.104 Basic of Finite Element Method

To find: Displacements of nodes 1 and 2.

Solution: Let 𝑢 and 𝑢 be the displacements of nodes 1 and 2. Then the extensions of
springs are,
𝛿 = 𝑢 ,𝛿 = 𝑢 − 𝑢 ,𝛿 = 𝑢 − 𝑢
We know that,
Minimum of potential energy principle,
π = Strain energy − Work done

1 1 1
𝜋 = 𝑘 𝛿 + 𝑘 𝛿 + 𝑘 𝛿 − 100𝑢 − 80𝑢
2 2 2
1 1 1
= 𝑘 𝑢 + 𝑘 𝑢 + 𝑘 (𝑢 − 𝑢 ) − 100𝑢 − 80 … (1)
2 2 2
∂π
∴ Now, =0
∂u

 𝑘 𝑢 + 𝑘 𝑢 + 𝑘 (𝑢 − 𝑢 )(−1) − 100 = 0

 (𝑘 + 𝑘 + 𝑘 )𝑢 − 𝑘 𝑢 = 0 …(2)

∂π
Similarly, = 0 𝑘 (𝑢 − 𝑢 ) − 80 = 0
∂u

−𝑘 𝑢 + 𝑘 𝑢 = 80 … (3)

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

𝑘 +𝑘 +𝑘 −𝑘 𝑢 100
𝑢 =
−𝑘 𝑘 80

Substituting the values of 𝑘 , 𝑘 and 𝑘 , we get,

60 + 75 + 100 −100 𝑢 100

=
−100 100 𝑢 80
235 −100 𝑢 100
=
−100 100 𝑢 80
 Finite Element Analysis 1.105

235𝑢 − 100𝑢 = 100

−100𝑢 + 100𝑢 = 80

135𝑢 = 180

180
𝑢 =
135

𝑢 = 1.333

Substitute the 𝑢 – value in equation (5)

−100(1.333) + 100𝑢 = 80

𝑢 = 2.133

Result: Displacements of nodes,

𝑢 = 1.333𝑚

𝑢 = 2.133𝑚

1.18 RAYLEIGH – RITZ METHOD (VARIATIONAL APROACH)

1.18.1. Introduction
 Rayleigh – Ritz method is an integral approach method which is useful for solving
complex structural problems, encountered in finite element analysis. This method
is possible only if a suitable functional is available, otherwise Galerkin’s method
of weighted residual is used. By using this method stiffness matrices and
consistent load vector can be assembled easily. This method is mostly used for
solving solid mechanics problems.
 The phrase “Variational methods” refers to methods that make use of variational
principles, such as the principles of virtual work and the principle of minimum
potential energy in solid and structural mechanics, to determine the approximate
solutions of the problems.
 In Rayleigh – Ritz method for continuous system we deal with the following
functional.
1.106 Basic of Finite Element Method

Potential energy, 𝜋 = 𝑓(𝑦, 𝑦 , 𝑦 )𝑑𝑥 … (1.26)

 In out terminology, a functional is an integral expression that implicitly contains

the governing differential equations for a particular problem/
 Total potential energy of the structure is given by,
Internal External
π = Potential − Potential
energy energy

= Strain Energy – Work done by External Forces

π=U−H
 In this method, the approximating functions must satisfy the boundary conditions
and should be easy to use. Polynomials are generally used and sometimes sine and
cosine terms are also used as approximating function.
 In general, any exact function can be represented as a polynomial or trigonometric
series with undetermined constants as shown below.
y=a +𝑎 𝑥+𝑎 𝑥 +𝑎 𝑥 +⋯
or
𝜋𝑥 3𝜋𝑥
y = 𝑎 sin + 𝑎 sin +⋯
𝑙 𝑙
The constants a , 𝑎 , 𝑎 are unknowns known as Ritz parameters of the curve.
When the parameters are infinite, the particular polynomial tends to match the exact
value. So, the accuracy depends upon the number of parameters chosen.
 The following two conditions must be fulfilled by the approximating function.
1. It should satisfy the geometric boundary conditions.
2. The function must have atleast one Ritz parameter.
 In general, a Rayleigh-Ritz solution is rarely exact except in some special simple
cases, but it becomes more accurate with the use of more parameters.
 This method can be understood clearly by solving the following examples.
 Finite Element Analysis 1.107

1.18.2 Solved Problems (on Rayleigh – Ritz Method)

Example 1.25
A simply supported beam subjected to uniformly distributed load over entire span.
Determine the bending moment and deflection at midspan by using Rayleigh-Ritz method
and compare with exact solutions.
Given:

To find: 1. Deflection and Bending moment at midspan.

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

𝑛𝜋𝑥
𝑦= 𝑎 sin is the approximating function.
𝑙
,

To make this series more simple let us consider only two terms.
𝜋𝑥 3𝜋𝑥
Deflection, 𝑦 = 𝑎 sin + 𝑎 sin … (1)
𝑙 𝑙
Where, 𝑎 , 𝑎 are Ritz Parameters.
We know that,
Total potential energy of the beam,  = U – H …(2)
Where, U → Strain energy.
H → Work done by external force.
The strain energy, U, of the beam due to bending is given by,

𝐸𝐼 𝑑 𝑦
𝑈= 𝑑𝑥 … (3)
2 𝑑𝑥
1.108 Basic of Finite Element Method
𝑑𝑦 𝜋𝑥 𝜋 3𝜋𝑥 3𝜋
= 𝑎 cos × + 𝑎 cos
𝑑𝑥 𝑙 𝑙 𝑙 𝑙
𝑑𝑦 𝑎 𝜋 𝜋𝑥 𝑎 3𝜋 3𝜋𝑥
= 𝑐𝑜𝑠 + cos
𝑑𝑥 𝑙 𝑙 𝑙 𝑙

𝑑 𝑦 −𝑎 𝜋 𝜋𝑥 𝜋 3𝜋 3𝜋𝑥 3𝜋
⇒ = 𝑠𝑖𝑛 × −𝑎 sin ×
𝑑𝑥 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙

−𝜋 𝑎 𝜋𝑥 9𝜋 3𝜋𝑥
= 𝑠𝑖𝑛 −𝑎 sin
𝑙 𝑙 𝑙 𝑙

𝑑 𝑦 −𝑎 𝜋 𝜋𝑥 𝑎 9𝜋 3𝜋𝑥
= 𝑠𝑖𝑛 − sin … (4)
𝑑𝑥 𝑙 𝑙 𝑙 𝑙

Substituting value in equation (3),

𝐸𝐼 −𝑎 𝜋 𝜋𝑥 𝑎 9𝜋 3𝜋𝑥
⇒𝑈= 𝑠𝑖𝑛 − sin 𝑑𝑥
2 𝑙 𝑙 𝑙 𝑙

𝐸𝐼 𝑎 𝜋 𝜋𝑥 𝑎 9𝜋 3𝜋𝑥
= 𝑠𝑖𝑛 + sin 𝑑𝑥
2 𝑙 𝑙 𝑙 𝑙

𝐸𝐼 𝜋 𝜋𝑥 3𝜋𝑥
= × 𝑎 𝑠𝑖𝑛 + 9𝑎 sin 𝑑𝑥
2 𝑙 𝑙 𝑙

𝐸𝐼 𝜋 𝜋𝑥 3𝜋𝑥
𝑈= × 𝑎 𝑠𝑖𝑛 + 81𝑎 sin
2 𝑙 𝑙 𝑙

𝜋𝑥 3𝜋𝑥
+ 2𝑎 𝑠𝑖𝑛 9𝑎 sin 𝑑𝑥
𝑙 𝑙

[∵ (𝑎 + 𝑏) = 𝑎 + 𝑏 + 2𝑎𝑏]
 Finite Element Analysis 1.109

𝐸𝐼 𝜋 𝜋𝑥 3𝜋𝑥
𝑈= × 𝑎 𝑠𝑖𝑛 + 81𝑎 sin
2 𝑙 𝑙 𝑙

𝜋𝑥 3𝜋𝑥
+ 18𝑎 𝑎 𝑠𝑖𝑛 sin 𝑑𝑥 … (5)
𝑙 𝑙

Since, ∫ 𝑎 𝑠𝑖𝑛 𝑑𝑥 = 𝑎 ∫ 1− 𝑑𝑥 ∵ sin 𝑥 =

𝑎 cos 2𝜋𝑥
= 1− 𝑑𝑥
2 𝑙

𝑎 2𝜋𝑥
= 𝑑𝑥 − 𝑐𝑜𝑠 𝑑𝑥
2 𝑙

2𝜋𝑥
𝑎 𝑠𝑖𝑛
= (𝑥) − 𝑙
2 2𝜋
𝑙

𝑎 𝑙 2𝜋𝑥
= 1−0− 𝑠𝑖𝑛 − sin 0
2 2𝜋 𝑙

𝑎 𝑙 𝑎 𝑙
= 1− (0 − 0) = [∵ sin 2𝜋 = 0; sin 0 = 0]
2 2𝜋 2

𝜋𝑥 𝑎 𝑙
𝑎 𝑠𝑖𝑛 𝑑𝑥 = … (6)
𝑙 2

Similarly,

3𝜋𝑥 1 6𝜋𝑥 1 − cos 2𝑥

81 𝑎 𝑠𝑖𝑛 = 81𝑎 1 − cos 𝑑𝑥 ∵ sin 𝑥 =
𝑙 2 𝑙 2

81𝑎 6𝜋𝑥
= 𝑑𝑥 − 𝑐𝑜𝑠 𝑑𝑥
2 𝑙
1.110 Basic of Finite Element Method

6𝜋𝑥
81𝑎 𝑠𝑖𝑛
= (𝑥) − 𝑙
2 6𝜋
𝑙

81𝑎 𝑙 6𝜋𝑥
= 1−0− 𝑠𝑖𝑛 − sin 0
2 6𝜋 𝑙

𝑎 𝑙
= 1− (sin 6𝜋 − sin 0) [∵ sin 6𝜋 = 0; sin 0 = 0]
2 6𝜋

3𝜋𝑥 81𝑎 𝑙
⇒ 81 𝑎 𝑠𝑖𝑛 = … (7)
𝑙 2

𝜋𝑥 3𝜋𝑥 𝜋𝑥 3𝜋𝑥
18 𝑎 𝑎 𝑠𝑖𝑛 𝑠𝑖𝑛 = 18 𝑎 𝑎 𝑠𝑖𝑛 𝑠𝑖𝑛
𝑙 𝑙 𝑙 𝑙

3𝜋𝑥 𝜋𝑥
= 18 𝑎 𝑎 𝑠𝑖𝑛 𝑠𝑖𝑛
𝑙 𝑙

1 2𝜋𝑥 4𝜋𝑥
= 18 𝑎 𝑎 𝑐𝑜𝑠 − 𝑐𝑜𝑠 𝑑𝑥
2 𝑙 𝑙

cos(𝐴 − 𝐵) − cos(𝐴 + 𝐵)
∵ sin 𝐴 sin 𝐵 =
2

18 𝑎 𝑎 2𝜋𝑥 4𝜋𝑥
= 𝑐𝑜𝑠 𝑑𝑥 − 𝑐𝑜𝑠 𝑑𝑥
2 𝑙 𝑙

2𝜋𝑥 4𝜋𝑥
18 𝑎 𝑎 sin sin
= 𝑙 − 𝑙
2 2𝜋 4𝜋
𝑙 𝑙

= 9𝑎 𝑎 [0 − 0] = 0
 Finite Element Analysis 1.111
[∵ sin 2𝜋 = 0; sin 4𝜋 = 0; sin 0 = 0]

𝜋𝑥 3𝜋𝑥
⇒ 18 𝑎 𝑎 𝑠𝑖𝑛 𝑠𝑖𝑛 =0 … (8)
𝑙 𝑙

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

𝐸𝐼 𝜋 𝑎 𝑙 81𝑎 𝑙
(5) ⇒ 𝑈= + +0
2 𝑙 2 2

𝐸𝐼 𝜋 𝑙
𝑈= [𝑎 + 81𝑎 ]
2 𝑙

EI π
Strain energy, U = [𝑎 + 81𝑎 ] … (9)
2 𝑙
We know that,
Work done by external force,

𝜋𝑥 3𝜋𝑥
𝐻= 𝜔 𝑦 𝑑𝑥 = 𝜔 𝑎 𝑠𝑖𝑛 + 𝑎 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙

𝜋𝑥 3𝜋𝑥
=𝜔 𝑎 𝑠𝑖𝑛 + 𝑎 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙

𝜋𝑥 3𝜋𝑥
=𝜔 𝑎 𝑠𝑖𝑛 𝑑𝑥 + 𝑎 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙

𝜋𝑥 3𝜋𝑥
−cos −cos
=𝜔 𝑎 𝑙 −𝑎 𝑙
𝜋 3𝜋
𝑙 𝑙
𝑎 𝑙 𝜋𝑥 𝑎 𝑙 3𝜋𝑥
=𝜔 cos − cos
𝜋 𝑙 3𝜋 𝑙

𝑎 𝑙 𝑎 𝑙
=𝜔 [(−1) − 1] − [(−1) − 1]
𝜋 3𝜋
1.112 Basic of Finite Element Method

2𝑎 𝑙 2𝑎 𝑙 [∵ cos 0 = 1;
=𝜔 + cos 𝜋 = −1
𝜋 3𝜋 cos 3𝜋 = −1]
2𝜔𝑙 𝑎
= 𝑎 +
𝜋 3
2𝜔𝑙 𝑎
𝐻= 𝑎 + … (10)
𝜋 3
Substitute (9) and (10) values in equation (2).
(2)  =U–H
𝐸𝐼 𝜋 2𝜔𝑙 𝑎
⇒ = (𝑎 + 81𝑎 ) − 𝑎 + … (11)
4 𝑙 𝜋 3
For stationary value of , the following conditions must be satisfied.
𝜕𝜋 𝜕𝜋
= 0 𝑎𝑛𝑑 =0
𝜕𝑎 𝜕𝑎

𝜕𝜋 𝐸𝐼 𝜋 2𝜔𝑙
⇒ = (2𝑎 ) − =0
𝜕𝑎 4 𝑙 𝜋

𝐸𝐼 𝜋 2𝜔𝑙
⇒ (2𝑎 ) =
4 𝑙 𝜋

4𝜔𝑙
⇒ 𝑎 =
𝐸𝐼𝜋

𝜕𝜋 𝐸𝐼 𝜋 2𝜔𝑙 1
Similarly, = (162𝑎 ) − =0
𝜕𝑎 4 𝑙 𝜋 3

𝐸𝐼 𝜋 2𝜔𝑙 4𝑙 4𝜔𝑙
⇒ (162𝑎 ) = × =
4 𝑙 𝜋 162 𝐸𝐼𝜋 243 𝐸𝐼𝜋

2𝜔𝑙 4𝑙 4𝜔𝑙
⇒ 𝑎 = × =
3𝜋 162 𝐸𝐼𝜋 243 𝐸𝐼𝜋

4𝜔𝑙
𝑎 =
243 𝐸𝐼𝜋
 Finite Element Analysis 1.113

We know that,
𝜋𝑥 3𝜋𝑥
𝑦 = 𝑎 𝑠𝑖𝑛 + 𝑎 𝑠𝑖𝑛
𝑙 𝑙
Substituting 𝑎 , and 𝑎 values,
4𝜔𝑙 𝜋𝑥 4𝜔𝑙 3𝜋𝑥
⇒ 𝑦= sin + 𝑠𝑖𝑛 … (12)
𝐸𝐼𝜋 𝑙 243 𝐸𝐼𝜋 𝑙

We know that, maximum deflection occurs at x = .

Substitute in x = equation (12),

𝑙 𝑙
4𝜔𝑙 𝜋×2 4𝜔𝑙 3𝜋 × 2
⇒ 𝑦 = sin + 𝑠𝑖𝑛
𝐸𝐼𝜋 𝑙 243 𝐸𝐼𝜋 𝑙

4𝜔𝑙 𝜋 4𝜔𝑙 3𝜋
⇒ 𝑦 = sin + 𝑠𝑖𝑛
𝐸𝐼𝜋 2 243 𝐸𝐼𝜋 2

4𝜔𝑙 4𝜔𝑙
𝑦 = −
𝐸𝐼𝜋 243 𝐸𝐼𝜋

4𝜔𝑙 1 𝜋 3𝜋
= 1− ∵ sin = 1; sin = −1
𝐸𝐼𝜋 243 2 2

4𝜔𝑙 3.98𝜔𝑙
= (0.9958) =
𝐸𝐼𝜋 𝐸𝐼 𝜋

𝜔𝑙
⇒ 𝑦 = 0.0130 … (13)
𝐸𝐼
We know that, simply supported beam subjected to uniformly distributed load,
maximum deflection is,
5 𝜔𝑙
𝑦 =
384 𝐸𝐼

𝜔𝑙
𝑦 = 0.0130 … (14)
𝐸𝐼
1.114 Basic of Finite Element Method

From equations (13) and (14), we know that, exact solution and solution obtained
by using Rayleigh-Ritz method are same.
Bending moment at Mid span
We know that,
d y
Bending moment, M = EI … (15)
dx
From equation (4), we know
d y 𝑎 𝜋 𝜋𝑥 𝑎 9𝜋 3𝜋𝑥
=− 𝑠𝑖𝑛 + sin
dx 𝑙 𝑙 𝑙 𝑙
Substituting 𝑎 , and 𝑎 values,
d y 4𝜔𝑙 𝜋 𝜋𝑥 4𝜔𝑙 9𝜋 3𝜋𝑥
⇒ =− × 𝑠𝑖𝑛 + × sin
dx 𝐸𝐼 𝜋 𝑙 𝑙 243 𝐸𝐼 𝜋 𝑙 𝑙

Maximum bending occurs at x = .

𝑙 𝑙
d y 4𝜔𝑙 𝜋 𝜋2 4𝜔𝑙 9𝜋 3𝜋 2
⇒ =− × 𝑠𝑖𝑛 + × sin
dx 𝐸𝐼 𝜋 𝑙 𝑙 243 𝐸𝐼 𝜋 𝑙 𝑙

4𝜔𝑙 𝜋 𝜋 4𝜔𝑙 9𝜋 3𝜋
=− × 𝑠𝑖𝑛 + × sin
𝐸𝐼 𝜋 𝑙 2 243 𝐸𝐼 𝜋 𝑙 2

4𝜔𝑙 𝜋 4𝜔𝑙 9𝜋
=− (1) + × (−1)
𝐸𝐼 𝜋 𝑙 243 𝐸𝐼 𝜋 𝑙

𝜋 3𝜋
∵ sin = 1; sin = −1
2 2

4𝜔𝑙 36𝜔𝑙 𝜋
=− −
𝐸𝐼 𝜋 243 𝐸𝐼 𝜋

4𝜔𝑙 36𝜔𝑙
=− −
𝐸𝐼 𝜋 243 𝐸𝐼 𝜋
 Finite Element Analysis 1.115

4𝜔𝑙 0.148𝜔𝑙 𝜔𝑙
=− + = −3.852
𝐸𝐼 𝜋 𝐸𝐼 𝜋 𝐸𝐼 𝜋

d y 𝜔𝑙
= −0.124
dx 𝐸𝐼

Substituting value in bending moment equation,

𝜔𝑙
(15) ⇒ 𝑀 = 𝐸𝐼 × (−0.124)
𝐸𝐼

⇒ 𝑀 = −0.124𝜔𝑙 … (16)
[Negative sign indicates downward load]
We know that, for simply supported beam subjected to uniformly distributed load,
maximum bending moment is,
𝜔𝑙
𝑀 =
8

𝑀 = 0.125 𝜔𝑙 … (17)
From equation (16) and (17), we know that, exact solution and solution obtained
by using Rayleigh-Ritz method are almost same. In order to get accurate result, more
terms in Fourier series should be taken.

Example: 1.26
A beam AB of span ‘l’ simply supported at ends and carrying a concentrated load
W at the centre ‘C’ as shown in Fig. Determine the deflection at midspan by using
Rayleigh-Ritz method and compare with exact solutions.

Fig. (i)
1.116 Basic of Finite Element Method

Given:

To find: Deflection at midspan, 𝑦 .

Solution: From example (1.24). we know that,
𝜋𝑥 3𝜋𝑥
Deflection, 𝑦 = 𝑎 𝑠𝑖𝑛 + 𝑎 𝑠𝑖𝑛 … (1)
𝑙 𝑙
Total potential energy of the beam,
=U–H …(2)
Where, U → Strain energy.
H → Work done by external force.
The strain energy, U, of the beam due to bending is given by,

𝐸𝐼 𝑑 𝑦
𝑈= 𝑑𝑥 … (3)
2 𝑑𝑥

From equation (9) in previous example problem (1.18), we know that,

𝐸𝐼 𝜋
𝑈= [𝑎 + 81𝑎 ] … (4)
4 𝑙
Work done by external force, H = 𝑊 𝑦 …(5)
We know,
𝜋𝑥 3𝜋𝑥
Deflection, 𝑦 = 𝑎 sin + 𝑎 sin
𝑙 𝑙

In the span, deflection is maximum at x =

𝑙 𝑙
𝜋×2 3𝜋 × 2
⇒ 𝑦 = 𝑎 sin + 𝑎 𝑠𝑖𝑛
𝑙 𝑙
 Finite Element Analysis 1.117
𝜋 3𝜋
= 𝑎 sin + 𝑎 𝑠𝑖𝑛
2 2
𝑦 =𝑎 −𝑎 …(6)
𝜋 3𝜋
∵ sin = 1; sin = −1
2 2
Substitute 𝑦 value in equation (5),
⇒ 𝐻 = 𝑊(𝑎 − 𝑎 ) … (7)
Substitute U and H values in equation (2),
𝐸𝐼 𝜋
⇒ = (𝑎 + 81𝑎 ) − 𝑊(𝑎 − 𝑎 ) … (8)
4 𝑙
For stationary value of , the following conditions must be satisfied.
𝜕𝜋 𝜕𝜋
= 0 𝑎𝑛𝑑 =0
𝜕𝑎 𝜕𝑎

𝜕𝜋 𝐸𝐼 𝜋
⇒ = (2𝑎 ) − 𝑊 = 0
𝜕𝑎 4 𝑙

𝐸𝐼 𝜋
⇒ (2𝑎 ) − 𝑊 = 0
4 𝑙

𝐸𝐼 𝜋
⇒ (𝑎 ) = 𝑊
2 𝑙

2𝑙 𝑊
⇒ 𝑎 =
𝐸𝐼𝜋

𝜕𝜋 𝐸𝐼 𝜋
Similarly, = (162𝑎 ) + 𝑊 = 0
𝜕𝑎 4 𝑙

𝐸𝐼 𝜋
⇒ (162𝑎 ) + 𝑊 = 0
4 𝑙

81𝐸𝐼 𝜋
⇒ 𝑎 = −𝑊
2 𝑙
1.118 Basic of Finite Element Method

2𝑙 𝑊
⇒ 𝑎 = … (10)
81𝐸𝐼𝜋
We know that,
Maximum deflection, 𝑦 =𝑎 −𝑎
2𝑙 𝑊 2𝑙 𝑊 2𝑙 𝑊 1
⇒ 𝑦 = + = 1+
𝐸𝐼𝜋 81𝐸𝐼𝜋 𝐸𝐼𝜋 81

2𝑙 𝑊
= (1.0123)
𝐸𝐼𝜋

2.0246 𝑙 𝑊 𝑊𝑙
= = 0.0207
𝐸𝐼𝜋 𝐸𝐼

𝑊𝑙
⇒ 𝑦 = … (11)
48.1𝐸𝐼
We know that, simply supported beam subjected to point load at centre, maximum
deflection is,
𝑊𝑙
𝑦 = … (12)
48𝐸𝐼
From equations (11) and (12), we know that, exact solution and solution obtained
by using Rayleigh-Ritz method are same. In order to get accurate result, more terms in
Fourier series should be taken.
Example 1.27
A simply supported beam subjected to uniformly distributed load over entire span
and it is subjected to a point load at the centre of the span. Calculate the bending moment
and deflection at midspan by using Rayleigh-Ritz method and compare with exact
solution.
Given:

Fig. (i)
 Finite Element Analysis 1.119

To find: 1. Deflection and Bending moment at midspan.

2. Compare with exact solutions.
Solution: From Example 1.24, we know that,
𝜋𝑥 3𝜋𝑥
Deflection, 𝑦 = 𝑎 sin + 𝑎 sin … (1)
𝑙 𝑙
Total potential energy of the beam is given by,
=U–H …(2)
Total strain energy U of the beam due to bending is given by,

𝐸𝐼 𝑑 𝑦
𝑈= 𝑑𝑥 … (3)
2 𝑑𝑥

From equation (9) in Example 1.18, we know that,

𝐸𝐼 𝜋
𝑈= [𝑎 + 81𝑎 ] … (4)
2 𝑙
Work done by external force,

𝐻= 𝜔 𝑦 𝑑𝑥 + 𝑊 𝑦 … (5)

From equation (10) in Example 1.24, we know that,

2𝜔𝑙 𝑎
𝜔 𝑦 𝑑𝑥 = 𝑎 + … (6)
𝜋 3

We know that,
𝜋𝑥 3𝜋𝑥
𝑦 = 𝑎 sin + 𝑎 sin
𝑙 𝑙

In the span, deflection is maximum at x =

𝑙 𝑙
𝜋×2 3𝜋 × 2
⇒ 𝑦 = 𝑎 sin + 𝑎 𝑠𝑖𝑛
𝑙 𝑙
1.120 Basic of Finite Element Method
𝜋 3𝜋
= 𝑎 sin + 𝑎 𝑠𝑖𝑛
2 2
𝑦 =𝑎 −𝑎 …(7)
𝜋 3𝜋
∵ sin = 1; sin = −1
2 2
2𝜔𝑙 𝑎
(5) ⇒ 𝐻= 𝑎 + + 𝑊(𝑎 − 𝑎 ) … (8)
𝜋 3
Substituting U and H values in equation (2) get
𝐸𝐼 𝜋 2𝜔𝑙 𝑎
⇒ = (𝑎 + 81𝑎 ) − 𝑎 + + 𝑊(𝑎 − 𝑎 )
4 𝑙 𝜋 3

𝐸𝐼 𝜋 2𝜔𝑙 𝑎
⇒ = (𝑎 + 81𝑎 ) − 𝑎 + − 𝑊(𝑎 − 𝑎 ) … (9)
4 𝑙 𝜋 3
For stationary value of , the following conditions must be satisfied.
𝜕𝜋 𝜕𝜋
= 0 and =0
𝜕𝑎 𝜕𝑎

𝜕𝜋 𝐸𝐼 𝜋 2𝜔𝑙
⇒ = (2𝑎 ) − −𝑊 =0
𝜕𝑎 4 𝑙 𝜋

𝐸𝐼 𝜋 2𝜔𝑙
⇒ 𝑎 − −𝑊 =0
2 𝑙 𝜋

𝐸𝐼 𝜋 2𝜔𝑙
⇒ (𝑎 ) = +𝑊
2 𝑙 𝜋

2𝑙 2𝜔𝑙
⇒ 𝑎 = +𝑊 … (10)
𝐸𝐼𝜋 𝜋

𝜕𝜋 𝐸𝐼 𝜋 2𝜔𝑙 1
Similarly, = (162𝑎 ) − +𝑊 =0
𝜕𝑎 4 𝑙 𝜋 3

𝐸𝐼 𝜋 2𝜔𝑙
⇒ (162𝑎 ) − +𝑊 =0
4 𝑙 3𝜋
 Finite Element Analysis 1.121

𝐸𝐼 𝜋 2𝜔𝑙
⇒ (162𝑎 ) = −𝑊
4 𝑙 3𝜋

4𝑙 2𝜔𝑙
⇒ 𝑎 = −𝑊
162𝐸𝐼𝜋 3𝜋

2𝑙 2𝜔𝑙
⇒ 𝑎 = −𝑊 … (11)
81𝐸𝐼𝜋 3𝜋
From equation (7),
We know that,
Maximum deflection, 𝑦 =𝑎 −𝑎

2𝑙 2𝜔𝑙 2𝑙 𝑊 2𝜔𝑙
⇒ 𝑦 = −𝑊 − −𝑊
𝐸𝐼𝜋 3𝜋 81𝐸𝐼𝜋 3𝜋

4𝜔𝑙 1 2𝑊𝑙 1
⇒ 𝑦 = 1− − 1+
𝐸𝐼𝜋 243 81𝐸𝐼𝜋 81

3.98𝜔𝑙 202𝑊𝑙
= −
𝐸𝐼𝜋 81𝐸𝐼𝜋

𝑙 𝜔 𝑊𝑙
= 0.0130 + 0.0207
𝐸𝐼𝜋 𝐸𝐼

𝑙 𝜔 𝑊𝑙
⇒ 𝑦 = 0.0130 + 0.0207 … (12)
𝐸𝐼𝜋 𝐸𝐼
We know that, simply supported beam subjected to uniformly distributed load,
maximum deflection is,
5 𝜔𝑙
𝑦 =
384 𝐸𝐼
Simply support beam subjected to point load at centre, maximum deflection is,
𝑊𝑙
𝑦 =
48𝐸𝐼
So, Total deflection
1.122 Basic of Finite Element Method

5 𝜔𝑙 𝑊𝑙
𝑦 = +
384 𝐸𝐼 48𝐸𝐼

𝜔𝑙 𝑊𝑙
𝑦 = 0.0130 + 0.0208 … (13)
𝐸𝐼 𝐸𝐼
From equations (12) and (13), we know that, exact solution and solution obtained
by using Rayleigh-Ritz method are same.
Bending moment at Mid span
We know that,
d y
Bending moment, M = EI … (14)
dx
From equation (4), we know
d y 𝑎 𝜋 𝜋𝑥 𝑎 9𝜋 3𝜋𝑥
=− 𝑠𝑖𝑛 + sin
dx 𝑙 𝑙 𝑙 𝑙
Substituting 𝑎 , and 𝑎 values from equation (10) and (11)
d y 2𝑙 2𝜔𝑙 𝜋 𝜋𝑥
⇒ =− + 𝑊 × sin
dx 𝐸𝐼 𝜋 3𝜋 𝑙 𝑙
2𝑙 2𝜔𝑙 9𝜋 3𝜋𝑥
+ −𝑊 × × sin
81𝐸𝐼 𝜋 3𝜋 𝑙 𝑙

Maximum bending occurs at x = .

𝑙
2𝑙 2𝜔𝑙 𝜋 𝜋2
=− + 𝑊 × sin
𝐸𝐼 𝜋 3𝜋 𝑙 𝑙

𝑙
2𝑙 2𝜔𝑙 9𝜋 3𝜋 2
+ −𝑊 × × sin
81𝐸𝐼 𝜋 3𝜋 𝑙 𝑙

2𝑙 2𝜔𝑙 𝜋 2𝑙 2𝜔𝑙 9𝜋
=− + 𝑊 × (1) + −𝑊 × (−1)
𝐸𝐼 𝜋 3𝜋 𝑙 81𝐸𝐼 𝜋 3𝜋 𝑙
 Finite Element Analysis 1.123
𝜋 3𝜋
∵ sin = 1; sin = −1
2 2
2𝑙 2𝜔𝑙 2𝑙 2𝜔𝑙
=− +𝑊 − −𝑊
𝐸𝐼 𝜋 3𝜋 9𝜋 𝐸𝐼 3𝜋

4𝜔𝑙 2𝜔𝑙 4𝜔𝑙 2𝜔𝑙

=− + − +
𝐸𝐼 𝜋 𝐸𝐼 𝜋 27 𝜋 𝐸𝐼 9𝜋 𝐸𝐼

4𝜔𝑙 1 2𝜔𝑙 1
=− 1− + 1+
𝐸𝐼 𝜋 27 𝐸𝐼 𝜋 9

3.851𝜔𝑙 𝑊𝑙
=− + 2.222
𝐸𝐼 𝜋 𝐸𝐼 𝜋

d y 𝜔𝑙 𝑊𝑙
= − 0.124 + 0.225
dx 𝐸𝐼 𝐸𝐼

Substituting value in bending moment equation,

𝑑 𝑦 𝜔𝑙 𝑊𝑙
(14) ⇒ 𝑀 = 𝐸𝐼 = −𝐸𝐼 0.124 + 0.225
𝑑𝑥 𝐸𝐼 𝐸𝐼

⇒ 𝑀 = −(0.124𝜔𝑙 + 0.225𝑊𝑙) … (15)

[Note: Negative sign indicates downward load]
We know that, for simply supported beam subjected to uniformly distributed load,
maximum bending moment is,
𝜔𝑙
𝑀 =
8
Simply supported beam subjected to point load at centre, maximum bending
moment is,
𝑊𝑙
𝑀 =
4
Total bending moment,
1.124 Basic of Finite Element Method

𝜔𝑙 𝑊𝑙
𝑀 = +
8 4
𝑀 = 0.125𝜔𝑙 + 0.25 𝑊𝑙 … (16)
From equation (15) and (16), we know that, exact solution and solution obtained
by using Rayleigh-Ritz method are almost same. In order to get accurate result, more
terms in Fourier series should be taken.

Example 1.28
A bar 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 bar by using two terms polynomial and three terms polynomial. Compare
with exact solutions.

Given:

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 = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 + ⋯+ 𝑎 𝑥
Case (i): Considering two terms of polynomial,
𝑖. 𝑒. , 𝑢 = 𝑎 + 𝑎 𝑥 …(1)
Apply boundary condition,
at x = 0, u = 0
 Finite Element Analysis 1.125

 0=𝑎 +0
 𝑎 =0
Substituting 𝑎 value in equation (1),
 𝑢=𝑎 𝑥 …(2)
𝑑𝑢
=𝑎
𝑑𝑥
We know that,
Total potential energy of the beam,  = U – H …(3)
Where, U → Strain energy of the bar.
H → Work done by external force of the bar.

𝐸𝐴 𝑑𝑢
Strain energy, 𝑈 = 𝑑𝑥
2 𝑑𝑥

𝐸𝐴 𝑑𝑢
= 𝑑𝑥
2 𝑑𝑥

𝐸𝐴𝑎
= [𝑥]
2
𝐸𝐴𝑎 𝑙
𝑈= … (4)
2
Work done by external force,

𝐻= 𝑃 𝑑𝑥 = 𝜌 𝑢 𝐴 𝑑𝑥 [∵ 𝐿𝑜𝑎𝑑, 𝑃 = 𝜌 𝑢 𝐴]

=𝜌𝐴 𝑢 𝑑𝑥 = 𝜌 𝐴 𝑎 𝑥 𝑑𝑥 [∵ 𝑢 = 𝑎 𝑥]

𝑥 𝜌𝐴𝑎
= 𝜌 𝐴𝑎 = [𝑙 ]
2 2
1.126 Basic of Finite Element Method

𝜌𝐴𝑎 𝑙
𝐻= … (5)
2
Substitute U and H values in equation (3)
𝐸𝐴 𝐴 𝑎 𝑙 𝜌 𝐴 𝑎 𝑙
(3) ⇒ 𝜋= −
2 2
For stationary value of , the following condition must be satisfied.
𝜕𝜋
𝑖. 𝑒., =0
𝜕𝑎
𝐸𝐴 (2𝑎 )𝑙 𝜌 𝐴 𝑙
⇒ − =0
2 2
𝜌𝐴𝑙
⇒ 𝐸𝐴 𝑎 𝑙 − =0
2
𝜌𝐴𝑙
⇒ 𝐸𝐴 𝑎 𝑙 =
2
𝜌𝑙
⇒ 𝑎 =
2𝐸
Substitute 𝑎 value in equation (2),
𝜌𝑙
⇒ 𝑢=𝑎 𝑥= ×𝑥
2𝐸
𝜌𝑙
𝑢= ×𝑥
2𝐸
We know that, Extension of the bar,
𝜌𝑙
𝛿𝑢 = 𝑢 − 𝑢 = ×𝑙−0
2𝐸
[∵ 𝐴𝑡 𝑥 = 𝑙, 𝑢 = 𝑢 , 𝐴𝑡 𝑥 = 0, 𝑢 = 𝑢 = 0]
𝜌𝑙
=
2𝐸
Extension or displacement of the bar,
𝜌𝑙
𝛿𝑢 = … (6)
2𝐸
 Finite Element Analysis 1.127

Stress in the bar,

𝑑𝑢 𝜌𝑙 𝜌𝑙
𝜎=𝐸 =𝐸× ∵𝑢= ×𝑥
𝑑𝑥 2𝐸 2𝐸
𝜌𝑙
𝜎= … (7)
2
Case (ii): Considering three terms of polynomial,
𝑖. 𝑒. , 𝑢 = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 … (8)
Apply boundary condition, at x = 0, u = 0
 0=𝑎 +0+0
 𝑎 =0
Substituting 𝑎 value in equation (1),
(8)  𝑢 =𝑎 𝑥+𝑎 𝑥 …(9)
𝑑𝑢
= 𝑎 + 2𝑎 𝑥
𝑑𝑥
We know that,
Total potential energy of the beam,  = U – H …(10)

𝐸𝐴 𝑑𝑢
Strain energy, 𝑈 = 𝑑𝑥
2 𝑑𝑥

𝐸𝐴
= (𝑎 + 2𝑎 𝑥) 𝑑𝑥
2

𝐸𝐴
= [𝑎 + (2𝑎 𝑥) + 2𝑎 2𝑎 𝑥] 𝑑𝑥
2

[∵ (𝑎 + 𝑏) = 𝑎 + 𝑏 + 2𝑎𝑏]

𝐸𝐴
= [𝑎 𝑑𝑥 + 4𝑎 𝑥 𝑑𝑥 + 4𝑎 2𝑎 𝑥 𝑑𝑥]
2
1.128 Basic of Finite Element Method

𝐸𝐴 𝑥 𝑥
= 𝑎 (𝑥) + 4𝑎 + 4𝑎 𝑎
2 3 3

𝐸𝐴 4𝑎 4𝑎 𝑎
= [𝑎 (𝑙 − 0) + (𝑙 − 0) + (𝑙 − 0)]
2 3 2
𝐸𝐴 4𝑎
𝑈= 𝑎 𝑙+ (𝑙 ) + 2𝑎 𝑎 (𝑙 ) … (4)
2 3
Work done by external force,

𝐻= 𝑃 𝑑𝑥 = 𝜌 𝑢 𝐴 𝑑𝑥 [∵ 𝐿𝑜𝑎𝑑, 𝑃 = 𝜌 𝑢 𝐴]

=𝜌𝐴 𝑢 𝑑𝑥 = 𝜌 𝐴 𝑎 𝑥 + 𝑎 𝑥 𝑑𝑥

= 𝜌 𝐴 [𝑎 𝑥 𝑑𝑥 + 𝑎 𝑥 𝑑𝑥]

𝑥 𝑥
=𝜌𝐴 𝑎 +𝑎
2 3

𝑎 𝑎
=𝜌𝐴 [𝑙 − 0] + [𝑙 − 0]
2 3

𝑎 𝑎
𝐻 =𝜌𝐴 𝑙 + 𝑙 … (12)
2 3
Substitute (11) and (12) values in (10),
(10) ⇒ 𝜋 =𝑈−𝐻

𝐸𝐴 4𝑎
𝜋= 𝑎 𝑙+ (𝑙 ) + 2𝑎 𝑎 (𝑙 )
2 3
𝑎 𝑎
−𝜌𝐴 𝑙 + 𝑙 … (13)
2 3
 Finite Element Analysis 1.129

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

𝜕𝜋 𝜕𝜋
𝑖. 𝑒., = 0 𝑎𝑛𝑑 =0
𝜕𝑎 𝜕𝑎

𝜕𝜋 𝐸𝐴 4𝑎 𝑎 𝑎
⇒ = 𝑎 𝑙+ (𝑙 ) + 2𝑎 𝑎 (𝑙 ) − 𝜌 𝐴 𝑙 + 𝑙 =0
𝜕𝑎 2 3 2 3

𝐸𝐴 𝑙
⇒ [2𝑎 𝑙 + 2𝑎 (𝑙 )] − 𝜌 𝐴 =0
2 2

𝜌𝐴𝑎 𝑙
⇒ 𝐸𝐴 [𝑎 𝑙 + 𝑎 𝑙 ] − =0
2
𝜌𝑙
⇒ 𝑎 +𝑎 𝑙 = … (14)
2𝐸
𝜕𝜋
Similarly, =0
𝜕𝑎

𝐸𝐴 8𝑎 𝑙
⇒ 0+ (𝑙 ) + 2𝑎 (𝑙 ) − 𝜌 𝐴 0 + =0
2 3 3

𝐸𝐴 8 𝜌 𝐴𝑙
⇒ 𝑎 (𝑙 ) + 2𝑎 (𝑙 ) =
2 3 3

8 2𝜌 𝐴𝑙
⇒ 𝑎 (𝑙 ) + 2𝑎 (𝑙 ) =
3 3𝐸𝐴

8 2𝜌 𝑙
⇒ 𝑎 (𝑙 ) + 2𝑎 (𝑙 ) =
3 3𝐸

4 𝜌𝑙
⇒ 𝑎 (𝑙 ) + 𝑎 (𝑙 ) =
3 3𝐸
4 𝜌𝑙
⇒ 𝑎 + 𝑎 𝑙=
3 3𝐸
𝜌𝑙
⇒ 𝑎 + 1.333𝑎 𝑙 = … (15)
3𝐸
1.130 Basic of Finite Element Method

Solving (14) and (15),

𝜌𝑙
𝑎 +𝑎 𝑙 =
2𝐸
𝜌𝑙
𝑎 + 1.333 𝑎 𝑙 =
2𝐸
𝜌𝑙 𝜌𝑙
𝑎 𝑙 − 1.333 𝑎 𝑙 = −
2𝐸 3𝐸
𝜌𝑙 1 1 𝜌𝑙 3−2
−0.333 𝑎 𝑙 = − =
𝐸 2 3 𝐸 6
𝜌𝑙
−0.333 𝑎 𝑙 =
6𝐸
𝜌
⇒ −0.333 𝑎 =
6𝐸
−𝜌
⇒ 𝑎 =
2𝐸

Substituting 𝑎 value in equation (14)

−𝜌 𝜌𝑙
(14) ⇒ 𝑎 + 𝑙 =
2𝐸 2𝐸
𝜌𝑙 𝜌𝑙
𝑎 − =
2𝐸 2𝐸
𝜌𝑙 𝜌 𝑙
𝑎 = +
2𝐸 2𝐸
𝜌𝑙
𝑎 =
𝐸

We know that, 𝑢 =𝑎 𝑥+𝑎 𝑥

Substitute 𝑎 and 𝑎 values,

𝜌𝑙 𝜌
𝑢= 𝑥− 𝑥
𝐸 𝐸
 Finite Element Analysis 1.131

𝜌 𝑥
𝑢= 𝑙𝑥 − … (16)
𝐸 2
At x = l, u = 𝑢 substitute in equation (16),
𝜌 𝑙
⇒ 𝑢 = 𝑙 −
𝐸 2
𝜌 𝑙
𝑢 = ×
𝐸 2
We know that, Extension of the bar,
𝜌𝑙
𝛿𝑢 = 𝑢 − 𝑢 = −0
2𝐸
[∵ 𝐴𝑡 𝑥 = 0, 𝑢 = 𝑢 = 0]
𝜌𝑙
=
2𝐸

Extension or displacement of the bar, 𝛿𝑢 = … (17)

From equation (16), we know that,

𝜌 𝑥
𝑢= 𝑙𝑥 −
𝐸 2
𝑑𝑢 𝜌 2𝑥 𝜌
= 𝑙𝑥 − = (𝑙 − 𝑥)
𝑑𝑥 𝐸 2 𝐸
Stress in the bar,
𝑑𝑢 𝜌
𝜎=𝐸 = 𝐸 × (𝑙 − 𝑥)
𝑑𝑥 𝐸
𝜎 = 𝜌(𝑙 − 𝑥) … (18)
Exact Solution: We know that, actual extension of the bar,

𝑃𝑑𝑥 𝑃𝐴𝑥
𝛿𝑙 = = 𝑑𝑥 [∵ 𝑃 = 𝜌 𝐴𝑥]
𝐴𝐸 𝐴𝐸
1.132 Basic of Finite Element Method

𝜌 𝜌 𝑥 𝜌 𝑙
= 𝑥 𝑑𝑥 = =
𝐸 𝐸 2 𝐸 2

𝜌𝑙
𝛿𝑙 =
2𝐸
From equation (6), (17) and (19), we know that total extension of the bar obtained
is exact in both the cases.
Result:
1. Displacement of the bar,
𝜌𝑙
𝛿𝑢 = [Two terms polynomial]
2𝐸
𝜌𝑙
𝛿𝑢 = [Three terms polynomial]
2𝐸
2. Stress in the bar,
𝜌𝑙
𝜎= [Two terms polynomial]
2
𝜎 = 𝜌(𝑙 − 𝑥) [Three terms polynomial]

Example 1.29
A linear elastic spring is subjected to a force of 1000 N as shown in Fig. Calculate
the displacement and the potential energy of the spring.

Given: Force, F = 1000 N

Stiffness, k = 500 N/mm
To find: 1. Displacement, x.
2. Potential energy, .
 Finite Element Analysis 1.133

Solution: we know that,

Total potential energy of the beam,  = U – H …(1)

Where, U = Strain energy = (kx)  x

H = Work done by external force = Fx

Substitute U and H values in equation (1)
1
⇒ 𝜋= (𝑘, 𝑥) × 𝑥 − 𝐹𝑥
2
1
𝜋= 𝑘𝑥 − 𝐹𝑥 … (2)
2
∂𝜋
For stationary value of π, =0
∂x
1
⇒ × 2𝑘𝑥 − 𝐹 = 0
2

⇒ 𝑘𝑥 − 𝐹 = 0

⇒ 500(𝑥) − 1000 = 0

⇒ 500(𝑥) = 1000

⇒ 𝑥 = 2𝑚𝑚

Substitute x values in equation (2),

1 1
(2) ⇒ 𝜋 = 𝑘𝑥 − 𝐹𝑥 = (500)(2) − 1000(2)
2 2

𝜋 = −1000 𝑁 − 𝑚𝑚
Result:
1. Displacement, x = 2mm
2. Potential energy,  = -1000 N – mm
1.134 Basic of Finite Element Method

Example 1.30
Consider a 1 mm diameter, 50 mm long aluminium pin-fin as shown in Fig.(i)
used to enhance the heat transfer from a surface wall maintained at 300 oC. Calculate the
temperature distribution in a pin fin by using Rayleigh-Ritz method, Take, k = 200 w/moC
for aluminium h = 20 w/moC, 𝑇 = 30 𝐶.

Fig. (i)
𝑑 𝑇 𝑃ℎ
𝑘 = (𝑇 − 𝑇 )
𝑑𝑥 𝐴

𝑇(0) = 𝑇 = 300 𝐶

𝑑𝑇
𝑞 = 𝑘𝐴 (𝐿) = 0 (Insulated tip)
𝑑𝑥

Given: The governing differential equation,

𝑑 𝑇 𝑃ℎ
𝑘 = (𝑇 − 𝑇 )
𝑑𝑥 𝐴

Diameter, d = l mm = 1  10-3 m
Length, L = 50 mm = 50  10-3 m
Thermal conductivity, k = 200 w/moC
Heat transfer coefficient, h = 20 w/moC
Fluid temperature, 𝑇 = 30oC
Boundary conditions, T (0) = TW = 300oC
𝑑𝑇
𝑞 = 𝑘𝐴 (𝐿) = 0
𝑑𝑥
To find: Ritz parameters.
 Finite Element Analysis 1.135

Solution: The equivalent functional representation is given by

 = Strain Energy – Work done
=U–W

1 𝑑𝑇 1 𝑃ℎ
= 𝑘 𝑑𝑥 + (𝑇 − 𝑇 ) 𝑑𝑥 − 𝑞 𝑇 … (1)
2 𝑑𝑥 2 𝐴

1 𝑑𝑇 1 𝑃ℎ
= 𝑘 𝑑𝑥 + (𝑇 − 𝑇 ) 𝑑𝑥 … (2)
2 𝑑𝑥 2 𝐴

[∵ 𝑞 = 0]
Assume a trial function, let T(x) =𝑎 + 𝑎 𝑥 + 𝑎 𝑥 … (3)

Apply boundary condition, at x = 0, T(x) = 300

300 = 𝑎 + 𝑎 (0) + 𝑎 (0)

𝑎 = 300
Substituting 𝑎 value in equation (3),
T(x) = 300 + 𝑎 𝑥 + 𝑎 𝑥 …(4)

𝑑𝑇
⇒ = 𝑎 + 2𝑎 𝑥 … (5)
𝑑𝑥
Substitute the equation (4), (5) in (2)

1 1 𝑃ℎ
= 𝑘(𝑎 + 2𝑎 𝑥) 𝑑𝑥 + (300 + 𝑎 𝑥 + 𝑎 𝑥 − 30) 𝑑𝑥
2 2 𝐴

1 1 𝑃ℎ
= 𝑘(𝑎 + 2𝑎 𝑥) 𝑑𝑥 + (270 + 𝑎 𝑥 + 𝑎 𝑥 ) 𝑑𝑥
2 2 𝐴

∵ (𝑎 + 𝑏) = 𝑎 + 𝑏 + 2𝑎𝑏;
(𝑎 + 𝑏 + 𝑐) = 𝑎 + 𝑏 + 𝑐 + 2𝑎𝑏 + 2𝑏𝑐 + 2𝑐𝑎
1.136 Basic of Finite Element Method

𝑘
= [𝑎 + 4𝑎 𝑥 + 4𝑎 𝑎 𝑥]
2

𝑃ℎ
+ [(270) + 𝑎 𝑥 + 𝑎 𝑥 + 540𝑎 𝑥 + 2𝑎 𝑎 𝑥
2𝐴

+ 540𝑎 𝑥 )]𝑑𝑥

×
𝑘 4𝑎 𝑥 4𝑎 𝑎 𝑥
= 𝑎 + +
2 3 2

𝑃ℎ 𝑎 𝑥 𝑎 𝑥 540𝑎 𝑥
+ 72900𝑥 + + +
2𝐴 3 5 2

×
2𝑎 𝑎 𝑥 540𝑎 𝑥
+ +
4 3

𝑘 4𝑎 (50 × 10 ) 4𝑎 𝑎 (50 × 10 )
= (50 × 10 )𝑎 + +
2 3 2
𝑃ℎ 𝑎 (50 × 10 )
+ 72900(50 × 10 ) +
2𝐴 3
𝑎 (50 × 10 ) 540𝑎 (50 × 10 )
+ +
5 2
2𝑎 𝑎 (50 × 10 ) 540𝑎 (50 × 10 )
+ +
4 3

200
= [50 × 10 𝑎 + 1.666 × 10 𝑎 + 50 × 10 𝑎 𝑎 ]
2
 × 10 × 20
+  + [3645 + 4.166 × 10 𝑎
2 × 4 × (10 )

+ 6.25 × 10 𝑎 + 0.675𝑎 + 3.125 × 10 𝑎 𝑎

+ 0.0225𝑎 ]
 Finite Element Analysis 1.137

 = [5𝑎 + 0.0166𝑎 + 0.5𝑎 𝑎 ] + [14.58 × 10 + 1.66𝑎

+ 2.5 × 10 𝑎 + 2700𝑎 + 0.125𝑎 𝑎 + 900𝑎 ]

 = [6.66𝑎 + 0.0191𝑎 + 0.625𝑎 𝑎 + 27000𝑎 + 900𝑎

+ 14.58 × 10 ]

𝜕𝜋
Apply, = 10
𝜕𝑎

⇒ 13.32𝑎 + 0.625𝑎 + 27000 = 0

⇒ 13.32𝑎 + 0.625𝑎 = −27000 … (6)

𝜕𝜋
Apply, =0
𝜕𝑎

⇒ 0.625𝑎 + 0.382𝑎 + 900 = 0

⇒ 0.625𝑎 + 0.382𝑎 = −900 … (7)

Solve the equations (6) and (7),
13.32𝑎 + 0.625𝑎 = −27000

0.625𝑎 + 0.382𝑎 = −900

(6)  0.625
⇒ 8.325𝑎 + 0.3906𝑎 = −16875
(7)  -13.32
⇒ − 8.325𝑎 − 0.50886𝑎 = +11988
−0.1182𝑎 = −4887
𝑎 = 41345
Substituting 𝑎 value in equation (6)
13.32𝑎 + 0.625(41345) = −27000
⇒ 𝑎 = 3967.01
1.138 Basic of Finite Element Method

Substitute 𝑎 and 𝑎 values in equation (3)

𝑇 = 300 − 3967.01𝑥 + 41345𝑥
Result: Temperature distribution in pin-fin
𝑇 = 300 − 3967.01𝑥 + 41345𝑥
Example 1.31
Using Rayleigh-Ritz method determine the expressions for displacement and
stress in fixed bar subjected to axial force P as shown in Fig. (i). Draw the displacement
and stress variation diagram. Take three terms in displacement function.

Fig. (i)
Given:

Fig. (ii)
To draw: Displacement and stress variation diagram
Solution: we know that, polynomial function for three terms,
𝑖. 𝑒. , 𝑢 = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 …(1)
This function has to satisfy the boundary condition,
(i) at x = 0, u = 0
(ii) at x = l, u = 0
 Finite Element Analysis 1.139

Apply boundary condition (i), we get

𝑎 =0 …(2)
Apply boundary condition (ii), we get
0=𝑎 +𝑎 𝑙+𝑎 𝑙 …(3)
From equations (2) and (3) we get,
0=𝑎 +𝑎 𝑙+𝑎 𝑙
𝑎 = −𝑎 𝑙 …(4)
Substituting 𝑎 and 𝑎 values in equation (1),
𝑢 =0−𝑎 𝑙𝑥+𝑎 𝑥

𝑢 = 𝑎 [−𝑙 𝑥 + 𝑥 ] …(5)

𝑙 𝑙 𝑙
At x = , 𝑢 =𝑎 −𝑙 +
2 2 2

𝑎 𝑙
𝑢 =− … (6)
4
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.
 Potential energy

𝐸𝐴 𝑑𝑢
π= 𝑑𝑥 − 𝑝 𝑢 … (7)
2 𝑑𝑥

We know that, 𝑢 = 𝑎 [−𝑙 𝑥 + 𝑥 ]

𝑑𝑢
⇒ = 𝑎 (−𝑙 + 2𝑥)
𝑑𝑥
1.140 Basic of Finite Element Method

𝐸𝐴 𝑎 𝑙
Now, π= (𝑎 (−𝑙 + 2𝑥)) 𝑑𝑥 − 𝑝 −
2 4

𝐸𝐴 𝑙
= 𝑎 (𝑙 + 4𝑥 − 4 𝑙 𝑥)𝑑𝑥 − 𝑝 𝑎
2 4

𝐸𝐴 4𝑥 𝑙
= 𝑎 𝑙 𝑥+ −2𝑙𝑥 +𝑝𝑎
2 3 4

𝐸𝐴 4𝑙 𝑙
= 𝑎 𝑙 + − 2 𝑙 (𝑙) +𝑝𝑎
2 3 4

𝐸𝐴 𝑙 𝑙
π= 𝑎 +𝑝𝑎
2 3 4

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

𝜕𝜋
𝑖. 𝑒., =0
𝜕𝑎

𝐸𝐴 𝑎 𝑙 𝑙
2 +𝑝 =0
2 3 2

𝑙 𝑙
𝐸𝐴𝑎 =𝑝
3 4
3𝑝
𝑎 =− … (8)
4𝐸𝐴𝑙
Substitute 𝑎 value in equations (5) and (6)
−3𝑝
𝑢= [−𝑙 𝑥 + 𝑥 ]
4𝐸𝐴𝑙
3𝑝𝑙
𝑢 =
16
 Finite Element Analysis 1.141

Stress in the bar,

𝑑𝑢
𝜎=𝐸
𝑑𝑥

= 𝐸𝑎 (−𝑙 + 2𝑥)

3𝐸𝑝
= (𝑙 − 2𝑥)
4𝐴𝑙
3𝑝
𝜎= (𝑙 − 2𝑥)
4𝐴𝑙
We know that,
3𝑝
At 𝑥 = 0 ⇒ 𝜎 =𝜎 =
4𝐴
1
At 𝑥 = ⇒ 𝜎 =𝜎 =0
2

3𝑝
At 𝑥 = 1 ⇒ 𝜎 =𝜎 =−
4𝐴
The variation of displacement and stress diagram are shown in figure.

Fig. Variation of displacement

Fig. Variation of stress

1.142 Basic of Finite Element Method

Example 1.32

Consider the differential equation for a problem such as + 300𝑥 = 0; 0 ≤

𝑥 ≤ 1 with the boundary conditions, y(0) = y(1) = 0, the functional corresponding to this
problem to be extremized is given by

1 𝑑𝑦
𝐼= − + 300𝑥 𝑦 𝑑𝑥
2 𝑑𝑥

Find the solution of the problem using Rayleigh Ritz method using a one term
solution is y = a x (1 – x3).
Given: Differential equation,
𝑑 𝑦
+ 300𝑥 = 0; 0≤𝑥≤1
𝑑𝑥

Boundary conditions: y(0) = y(1) = 0

(i) x = 0, y = 0

(ii) x = 1, y = 0

1 𝑑𝑦
𝐼= − + 300𝑥 𝑦 𝑑𝑥
2 𝑑𝑥

Trial function, y = a x (1 – x3)

i.e., y = a x – a x4 … (1)
It satisfies the boundary conditions,
x = 0, y = 0
x = 1, y = 0
Differential equation,
𝑑 𝑦
+ 300𝑥 = 0 … (2)
𝑑𝑥
 Finite Element Analysis 1.143
𝑑𝑦
⇒ = 𝑎 − 4𝑎𝑥
𝑑𝑥
𝑑𝑦
= (𝑎 − 4𝑎𝑥 ) … (3)
𝑑𝑥

1 𝑑𝑦
We know that, 𝐼= − + 300𝑥 𝑦 𝑑𝑥 … (4)
2 𝑑𝑥

Substitute the equation (1), (3) in (4)

1
⇒ 𝐼= − (𝑎 − 4𝑎𝑥 ) + 300𝑥 (𝑎 𝑥 – 𝑎 𝑥 ) 𝑑𝑥
2

1
= − (𝑎 + 16𝑎 𝑥 − 8𝑎 𝑥 )
2

+ (300𝑎 𝑥 – 300𝑎 𝑥 ) 𝑑𝑥

1 𝑥 𝑥
=− 𝑎 𝑥 + 16𝑎 − 8𝑎
2 7 4

𝑎𝑥 𝑎𝑥
+ 300 – 300
4 7

1 16 300 300
=− 𝑎 + 𝑎 − 2𝑎 + 𝑎– 𝑎
2 7 4 7

𝑎 8 300 300
𝐼=− − 𝑎 +𝑎 + 𝑎– 𝑎
2 7 4 7
𝜕𝐼
Apply, =0
𝜕𝑎
2𝑎 8 300 300
⇒ − − (2𝑎) + 2𝑎 + − =0
2 7 4 7
1.144 Basic of Finite Element Method
16 300 300
−𝑎 − 𝑎 + 2𝑎 + − =0
7 4 7
−16𝑎 + 7𝑎 2100 + 1200
𝑖. 𝑒., =−
7 28
7
𝑖. 𝑒., −9𝑎 = −900 ×
28

a = 25

Hence the solution is, y = 25 x (1 – x3)

Result: Solution, y = 25 x (1 – x3)

1.19. APPLICATION TO BAR ELEMENT

1.19.1 Bar Element Formulated from the stationarity of a Functional

Fig. 1.31 Typical bar Element

Consider a bar element with nodes 1 and 2 as shown in Fig. 1.31. 𝑢 and 𝑢 are
the displacements at the respective nodes. So, 𝑢 and 𝑢 are considered as degrees of
freedom of this bar element. [Refer chapter 2, Section 2.6.3, equation no. (2.21)].
Node: Degrees of freedom is nothing but noded displacements.]
𝑢 =𝑁 𝑢 +𝑁 𝑢 …(1.27)
𝑥
Where, 𝑁 = 1 −
𝑙
𝑥
𝑁 =
𝑙
Substitute the 𝑁 , 𝑁 values in equation (1.27)
 Finite Element Analysis 1.145
𝑥 𝑥
𝑢 = 1− 𝑢 + 𝑢 … (1.28)
𝑙 𝑙
The strain energy stored within the element is given by,

𝐴𝐸 𝑑𝑢
𝑈= 𝑑𝑥 … (1.29)
2 𝑑𝑥

𝐴𝐸 𝑢 − 𝑢
𝑈= 𝑑𝑥
2 𝑙

𝐴𝐸 𝑢 − 𝑢
𝑈= (𝑥)
2 𝑙
𝐴𝐸 𝑢 − 𝑢
𝑈= (𝑙) … (1.30)
2 𝑙
When there is a distributed force 𝑞 acting at each point on the element and
concentrated forces F at the nodes, the potential of the external forces is given by,

𝐻= 𝑞 𝑢 𝑑𝑥 + 𝐹 𝑢 + 𝐹 𝑢

=𝑞 𝑙 𝑑𝑥 + 𝐹 𝑢 + 𝐹 𝑢 ∵𝑢=

𝑙
𝐻=𝑞 𝑑𝑥 + 𝐹 𝑢 + 𝐹 𝑢 … (1.31)
2
Thus the total potential energy,
π=𝑈−𝐻

𝐴𝐸 𝑢 − 𝑢 𝑙
π= −𝑞 (𝑢 + 𝑢 ) 𝑑𝑥 − 𝐹 𝑢
2 𝑙 2
−𝐹 𝑢 … (1.32)

𝜕π
Apply, =0
𝜕𝑢
1.146 Basic of Finite Element Method
𝐴𝐸 𝑞 𝑙
⇒− × 2(𝑢 − 𝑢 ) − −𝐹 =0
2𝑙 2
𝐴𝐸 𝑞 𝑙
(𝑢 − 𝑢 ) − −𝐹 =0
𝑙 2
𝐴𝐸 𝑞 𝑙
⇒ (𝑢 − 𝑢 ) = +𝐹 … (1.33)
𝑙 2
𝜕π
Similarly, =0
𝜕𝑢

𝐴𝐸 𝑞 𝑙
⇒ × 2(𝑢 − 𝑢 ) − −𝐹 =0
2𝑙 2
𝐴𝐸 𝑞 𝑙
(𝑢 − 𝑢 ) = +𝐹 … (1.34)
𝑙 2

Arrange the equation (1.33) and 1.34) in matrix form, we get,

𝑞 𝑙
𝐴𝐸 1 −1 𝑢 𝐹
= 2 + … (1.35)
𝑙 −1 1 𝑢 𝑞 𝑙 𝐹
2
[k] {u} = {F}

1.19.2 One – dimensional Heat Transfer Elements Based on the stationary of

a Functional
Consider a bar element with nodes 1 and 2 as shown in Fig. 1.32. T1 and T2 are
the temperatures at the respective nodes. So, T1 and T2 considered as degrees if freedom
of this bar element.

Fig. 1.32. Heat transfer element

 Finite Element Analysis 1.147

We know that, 𝑇(𝑥) = 𝑁 𝑇 + 𝑁 𝑇 …(1.36)

𝑥 𝑥
𝑇(𝑥) = 1 − 𝑇 + 𝑇 … (1.37)
𝑙 𝑙
𝑥 𝑥
∵ 𝑁 = 1 − ;𝑁 =
𝑙 𝑙
The strain energy stored within the element is given by,

1 𝑑𝑇
𝑈= 𝑘 𝑑𝑥 … (1.38)
2 𝑑𝑥

Potential energy of external force is given by

𝐻= 𝑞 𝑇 𝑑𝑥 + 𝑄 𝑇 + 𝑄 𝑇 … (1.39)

The total potential energy,

π =𝑈−𝐻

1 𝑑𝑇
= 𝑘 𝑑𝑥 − 𝑞 𝑇 𝑑𝑥 − 𝑄 𝑇 − 𝑄 𝑇 … (1.40)
2 𝑑𝑥

From equation (1.37),

𝑑𝑇 1 1
=− 𝑇 + 𝑇
𝑑𝑥 𝑙 𝑙
𝑑𝑇 1
= (𝑇 − 𝑇 ) … (1.41)
𝑑𝑥 𝑙
Substitute the equation (1.41) in equation (1.40),

1 1
π= 𝑘 (𝑇 − 𝑇 ) 𝑑𝑥 − 𝑞 𝑇 𝑑𝑥 − 𝑄 𝑇 − 𝑄 𝑇
2 𝑙

𝑘 𝑞 𝑙
π= ((𝑇 − 𝑇 ) − (𝑇 + 𝑇 ) − 𝑄 𝑇 − 𝑄 𝑇
2𝑙 2
1.148 Basic of Finite Element Method

We know that,
𝜕π
Apply, =0⇒
𝜕𝑇
𝑘 𝑞 𝑙
× 2(𝑇 − 𝑇 )(−1) − −𝑄𝑇 =0
2𝑙 2
𝑘 𝑞 𝑙
(𝑇 − 𝑇 ) = +𝑄𝑇 … (1.42)
𝑙 2
𝜕π
Similarly, =0⇒
𝜕𝑇
𝑘 𝑞 𝑙
× 2(𝑇 − 𝑇 ) − −𝑄 𝑇 =0
2𝑙 2
𝑘 𝑞 𝑙
(𝑇 − 𝑇 ) = +𝑄 𝑇 … (1.43)
𝑙 2
Arrange the equation (1.42) and 1.43) in matrix form, we get,
𝑞 𝑙
𝑘 1 −1 𝑇 𝑄
= 2 + … (1.44)
𝑙 −1 1 𝑇 𝑞 𝑙 𝑄
2
[k] {u} = {F}

1.20 ADVANTAGES OF FINITE ELEMENT METHOD

1. One of the major advantages of FEM over approximate methods is the fact that
FEM can handle irregular geometry in a convenient manner.
2. It handles general load conditions without difficulty.
3. Non-homogeneous materials can be handled easily.
4. All the various types of boundary conditions are handled.
5. Dynamic effects are included.
6. Very the size of the elements to make it possible for using small elements where
necessary.
7. Higher order elements may be implemented.
 8. Altering the element model with different loads, boundary conditions and other
changes in the model can be done easily and cheaply.

1.21 DISADVANTAGES OF FEM

1. It requires a digital computer and fairly extensive software.
2. It required longer execution time compared with finite difference method.
3. Output result will vary considerably, when the body is modeled with fine mesh
when compared to body modeled with course mesh.
4. In finite difference method, the governing differential equation of the
phenomenon must be known whereas finite element method does not require to
express fully.

1.22 APPLICATIONS OF FINITE ELEMENT ANALYSIS

The finite element can be used to analysis both structural and non-structural
problems.
In structural problems, displacement at each nodal point is obtained. Bu using
these displacement solutions, stress and strain in each element can be calculated.
Typical structural problems include:
1. Stress analysis including truss and frame analysis.
2. Stress concentration problems typically associated with holes, fillets or other
changes in geometry in a body.
3. Buckling analysis: Example: Connecting rod subjected to axial compression.
4. Vibration analysis: Example: A beam subjected to different types of loading.
In non-structural problems, temperature or fluid pressure at each nodal point is
obtained. By using these values, properties such as heat flow, fluid flow, etc., for each
element can be calculated.
Non – structural problems include:
1. Heat transfer analysis:
Example: Steady state thermal analysis on composite cylinder.
2. Fluid flow analysis.
 Example: Fluid flow through pipes.
3. Distribution of electric or magnetic potential
Example: Modeling of electromagnetic field of motor.
Recently finite element analysis is used in some biomechanical engineering
problems (which may include stress analysis) typically include analysis of human spine,
skull, hip joints, heart, eye, etc.
 UNIT 2
ONE DIMENSIONAL PROBLEMS

2.1. INTRODUCTION
Bar and beam elements are considered as one dimensional elements. These
elements are often used to model trusses and frame structures.
A bar is a member which resist only axial loads, whereas a beam can resist
axial, lateral, and twisting loads. A truss is an assemblage of bars with pin joints and
a frame is an assemblage of beam elements.
In this unit, one dimensional elements and step-by-step procedure for the
analysis of bars, trusses and beams are discussed. The total potential energy, stress-
strain and strain- displacement relationships are used in developing the finite
element method for a one dimensional problem. The basic one dimensional
procedure is same for two and three dimensional problems.

2.2. STRESS, STRAIN, DISPLACEMENT AND LOADING

In one dimensional problems, stress (σ), strain (e), displacement (u) and loading
depends only on the variable x. So, the vectors u, σ and e can be written as,
u = u(x)
σ= σ(x)
e = e(x)
The stress-strain relationship is given by.
σ = Ee
where,
σ → Stress, N/mm² Fig. 2.1 A bar is subjected to loading
2.2 One Dimensional Problems

e → Strain.
E→ Young's modulus, N/mm²
The strain-displacement relationship is given by,
𝑑𝑢
e=
𝑑𝑥
The differential volume can be written as,
dV = A dx
There are three types of loading acts on the body. They are:
(i) Body force (f).
(ii) Traction force (T).
(iii) Point load (P).
Body Force (f)
A body force is a distributed force acting on every elemental volume of the
body. Unit: Force per unit volume.
Example: Self weight due to gravity.
Traction Force (T)
A traction force is a distributed force acting on the surface of the body. Unit:
Force per unit area but for one dimensional problems, unit is force per unit length.
Examples: Frictional resistance, viscous drag, surface shear, etc.
Point Load (P)
Point load is a force acting at a particular point which causes displacement.

2.3. FINITE ELEMENT MODELLING

Finite element modelling consists of the following:
(i) Discretization of structure.
(ii) Numbering of nodes.
(i) Discretization
The art of subdividing a structure into a convenient number of smaller
components known as discretization.
 Finite Element Analysis 2.3

Consider a bar as shown in Fig.2.2. The first step is to model the bar as a
stepped at Let us model the bar using 5 finite elements, each having a uniform cross
section as shown Fig.2.3. Every element connects two nodes. Five element, six node
model element is shown in Fig.2.4.

Fig. 2.2 Fig. 2.3 Fig. 2.4

The element numbers are circled to distinguish them from node numbers. The
cross-sectional area, traction forces and body forces are constant within each element.
But, these are differing in magnitude from element to element. Better results are
obtained by increase number of finite elements.
(ii) Numbering of nodes
In one dimensional problem, each node is allowed to move only in ±x direction.
So, each node has one degrees of freedom. (Degrees of freedom is nothing but a nodal
displacement). A six node finite element model is shown in Fig.2.5. It has six degrees of
freedom. Load is considered as positive if it is acting along the +x direction.

Fig. 2.5
In the element connectivity table, the heading 1 and 2 refer to local node
numbers of an element and the corresponding node numbers on the structure are called
global numbers. Connectivity thus establishes the local-global correspondence.

Fig. 2.6 (a)

2.4 One Dimensional Problems

Fig. 2.6 (b) Connectivity bar

2.4. CO-ORDINATES
The co-ordinates are generally classified as follows:
(i) Global co-ordinates,
(ii) Local co-ordinates.
(iii) Natural co-ordinates.

2.4.1. Global Co-ordinates

The points in the entire structure are defined using co-ordinate system is known
as co-ordinate system.
Example:

Fig. 2.7 One dimensional bar

Fig. 2.8 Two dimensional triangular element

 Finite Element Analysis 2.5

2.4.2. Local Co-ordinates

In finite element method, separate co-ordinate is used for each element. It is very
used for deriving element properties. But the final equations are to be formed only by
global ordinate systems.

Fig. 2.9 Local co-ordinates system

2.6 One Dimensional Problems

2.4.3. Natural Co-ordinates

A natural co-ordinate system is used to define any point inside the element by a set
of dimensionless numbers whose magnitude never exceeds unity. This system is very
useful in assembling of stiffness matrices.
(1) Natural Co-ordinates in One Dimension

Fig. 2.10. Natural co-ordinates for a line element

Consider a two noded line element as shown in Fig.2.10. Any point p inside the
line element is identified by two natural co-ordinates L1 and L2 and the Cartesian co-
ordinate x. Node 1 and node 2 have the Cartesian co-ordinates x1 and x2 respectively.
We know that,
Total weightage of natural co-ordinates at any point is unity.
i.e.. L1 + L2 = = 1
Any point x within the element can be expressed as a linear combination of the
nodal co-ordinates of nodes 1 and 2 as,
L1x1+L2x2 =x
Arrange equation (2.1) and (2.2) in matrix form,
𝐿 1
1 1
=
𝑥 𝑥
𝐿 𝑥
𝐿 1
1 1 1 𝑥 1 1
 = =
𝑥 𝑥 (𝑥 − 𝑥 ) −𝑥 1
𝐿 𝑥 𝑥
𝑎 𝑎 1 𝑎 −𝑎
𝑁𝑜𝑡𝑒: 𝑎 𝑎 =
(𝑎 . 𝑎 ) − (𝑎 . 𝑎 ) −𝑎 𝑎
 Finite Element Analysis 2.7
1 𝑥 − 𝑥
= −𝑥 + 𝑥
𝑥 −𝑥

1 𝑥 − 𝑥
= 𝑥 − 𝑥
𝑥 −𝑥

1 𝑥 − 𝑥
=
𝑙 𝑥 − 𝑥

{∵ 𝑥 − 𝑥 𝑖𝑠 𝑡ℎ𝑒 𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑡ℎ𝑒 𝑒𝑙𝑒𝑚𝑒𝑛𝑡, 𝑙}

𝑥 −𝑥
𝐿 𝑙
=
𝐿 𝑥−𝑥
𝑙
The variation of L1 and L2 is shown in Fig.2.12 and Fig.2.13. L1 is one at node 1
and it is zero at node 2 whereas L2 is one at node 2 and it is zero at node 1.

Fig. 2.11 Fig. 2.12. Variation of L1 Fig. 2.13 Variation of L2

Integration of polynomial terms in natural co-ordinates can be performed by
using the simple formula,

𝛼! 𝛽!
(𝐿 ) (𝐿 )𝛼 𝑑𝑥 = ×𝑙
(𝛼 + 𝛽 + 1)!

Where,  is the factorial of .

Natural Co-ordinate, 𝜀
In one dimensional problem, the following type of natural co-ordinate is also
used. Consider a one dimensional element as shown in Fig.2.14.
In the local number scheme, the first node will be numbered 1 and the second
node 2. c is the Centre of nodes and 2 and p is the point referred.
The natural coordinator 𝜀 for any point in the element is defined as,
2.8 One Dimensional Problems
𝑝𝑐
𝜀 = 𝑥 −𝑥
2

Fig. 2.14.
𝑝𝑐
𝜀= [∵ 𝑥 − 𝑥 = 𝑙]
𝑙
2
2 2
= 𝑝 𝑐 = (𝑥 − 𝑥 ) [∵ 𝑝𝑐 = 𝑥 − 𝑥 ]
𝑙 𝑙
2 𝑥 +𝑥 𝑥 +𝑥
= × 𝑥− [∵ 𝑥 = ]
𝑙 2 2
2 𝑥 +𝑥
= × 𝑥−
𝑙 2
2 𝑥 −𝑥 +2𝑥
= × 𝑥−
𝑙 2
2 𝑙+2𝑥
= × 𝑥−
𝑙 2
2 𝑙
𝜀 = 𝑥− +𝑥
𝑙 2
𝜀𝑙 𝑙
 =𝑥− −𝑥
2 2
𝜀𝑙 𝑙
 + = 𝑥−𝑥
2 2
𝑙
 ( + 1) = 𝑥 − 𝑥 … (2.4)
2
 Finite Element Analysis 2.9

Applying boundary conditions,

At node 1, x=x1
1
(2.4)  (1 + 𝜀) = 0
2
 1+ 𝜀 =0
 𝜀 = −1
At node 2, x=x2
𝑙
(2.4)  (1 + 𝜀) = 𝑥 − 𝑥
2
𝑙
 (1 + 𝜀) = 𝑙
2
 1+𝜀 = 2
 𝜀=1

Fig. 2.15 Variation of natural co-ordinate, 

Natural Co-ordinates in Two Dimensions
Consider a triangular element having 3 nodes as shown in Fig.2.l6.
Let p is the point inside the element and it has 3 co-ordinates L1 , L2 and L3
From the definition of natural co-ordinates, we know that,
𝐿 +𝐿 +𝐿 =1 … (2.5)
2.10 One Dimensional Problems

𝐿 𝑥 +𝐿 𝑥 +𝐿 𝑥 =𝑥 … (2.6)
𝐿 𝑦 +𝐿 𝑦 +𝐿 𝑦 =𝑦 … (2.7)
Assemble the above equations in matrix form,
1 1 1 𝐿 1
𝑥 𝑥 𝑥 𝐿 = 𝑥
𝑦 𝑦 𝑦 𝐿 𝑦

𝐿 1 1 1 1
𝐿 = 𝑥 𝑥 𝑥 𝑥 . . . (2.8)
𝐿 𝑦 𝑦 𝑦 𝑦

+ − +
⎡1 1 1⎤
⎢− + 1⎥
𝐿𝑒𝑡 𝐷 = ⎢𝑥 ⎥
𝑥 ⎥
⎢ 𝑥
⎢+ − +⎥
⎣𝑦 𝑦 𝑦 ⎦

𝐶
𝐷 = … (2.9)
|𝐷|

Fig. 2.16.
Coefficients of matrix D:
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 )
𝐶= 𝑦 −𝑦 𝑦 −𝑦 𝑦 −𝑦
𝑥 −𝑥 𝑥 −𝑥 𝑥 −𝑥
 Finite Element Analysis 2.11
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
 𝐶 = 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 … . (2.10)
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
1 1 1
𝐷= 𝑥 𝑥 𝑥
𝑦 𝑦 𝑦
|𝐷| = 1 (𝑥 𝑦 − 𝑥 𝑦 ) − 1 (𝑥 𝑦 − 𝑥 𝑦 )
+ 1(𝑥 𝑦 − 𝑥 𝑦 ) … (2.11)
Substitute CT and |D| values in equation (2.9)
1
(2.9)  𝐷 =
(𝑥 𝑦 − 𝑥 𝑦 ) − (𝑥 𝑦 − 𝑥 𝑦 ) + (𝑥 𝑦 − 𝑥 𝑦 )
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
× 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
Substitute D-1 values in equation (2.8),
𝐿 1 1 1 1
 𝐿 = 𝑥 𝑥 𝑥 𝑥
𝐿 𝑦 𝑦 𝑦 𝑦
𝐿 1
 𝐿 =
(𝑥 𝑦 − 𝑥 𝑦 ) − (𝑥 𝑦 − 𝑥 𝑦 ) + (𝑥 𝑦 − 𝑥 𝑦 )
𝐿
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 1
× 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 𝑥 … (2.12)
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 𝑦
The area of the triangle ABC can be expressed as a function of the x , y co-
ordinates of the nodes 1 , 2 and 3.

1 1 𝑥 𝑦
𝐴= 1 𝑥 𝑦
2 1 𝑥 𝑦
1
= [1(𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )]
2
1
= [𝑥 𝑦 − 𝑥 𝑦 − 𝑥 𝑦 − 𝑥 𝑦 + 𝑥 𝑦 − 𝑥 𝑦 ]
2
2.12 One Dimensional Problems
1
𝐴 = [𝑥 𝑦 − 𝑥 𝑦 − (𝑥 𝑦 − 𝑥 𝑦 ) + (𝑥 𝑦 − 𝑥 𝑦 )]
2
 (𝑥 𝑦 − 𝑥 𝑦 ) − (𝑥 𝑦 − 𝑥 𝑦 ) + (𝑥 𝑦 − 𝑥 𝑦 ) = 2 𝐴 … (2.13)
Substitute (2.13) value in equation (2.12),
𝐿 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 1
1
 𝐿 = 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 𝑥 … (2.12)
𝐿 2𝐴 𝑥 𝑦 − 𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 𝑦
Integration of polynomial terms in natural co-ordinates for two dimensional
elements can be performed by using the formula,
𝛼! 𝛽! 𝛾!
(𝐿 )∝ (𝐿 ) (𝐿 ) 𝑑𝐴 = × 2𝐴 … (2.14)
(𝛼 + 𝛽 + 𝛾 + 2)!

2.4.4. Solved problems on natural co-ordinates

Example 2.1

Calculate the value of 𝐿 𝐿 𝐿 𝑑𝐴.

Solution:

𝐿 𝐿 𝐿 𝑑𝐴 … (1)

We know that
𝛼! 𝛽! 𝛾!
(𝐿 )∝ (𝐿 ) (𝐿 ) 𝑑𝐴 = × 2𝐴 … (2)
(𝛼 + 𝛽 + 𝛾 + 2)!
Compare equation (1) and (2),
 = 1,  = 1,  = 1.
1! 1! 1!
𝐿 𝐿 𝐿 𝑑𝐴 = × 2𝐴
(1 + 1 + 1 + 2)!
1×1 ×1
= × 2𝐴
5!
1
= × 2𝐴
5×4 ×3 ×2×1
 Finite Element Analysis 2.13
𝐴
𝐿 𝐿 𝐿 𝑑𝐴 =
60

Example 2.2
Determine the value of ∮ 𝐿 ( 𝐿 ) (𝐿 ) 𝑑𝐴.

𝐒𝐨𝐥𝐮𝐭𝐢𝐨𝐧: we know that, 𝐿 ( 𝐿 ) (𝐿 ) 𝑑𝐴.

𝛼! 𝛽! 𝛾!
(𝐿 )∝ (𝐿 ) (𝐿 ) 𝑑𝐴 = × 2𝐴
(𝛼 + 𝛽 + 𝛾 + 2)!

Here, ∝ = 1, 𝛽 = 2 𝑎𝑛𝑑 𝛾 = 3 .

1! × 1! × 1! 1! × 2! × 3!
𝐿 𝐿 𝐿 𝑑𝐴 = × 2𝐴 = × 2𝐴
(1 + 2 + 3 + 2)! 8!

1×2 × 1×3 ×2×1

= × 2𝐴
8×7×6×5×4 ×3 ×2×1

𝐴
𝐿 𝐿 𝐿 𝑑𝐴 =
1680

Example 2.3

Calculate the value of ∮ 𝐿 𝐿 𝑑𝑥.

Solution: we know that,

𝛼! 𝛽!
𝐿∝ 𝐿 𝑑𝑥 = × 𝑙 [From equaiton no. 2.3]
(𝛼 + 𝛽 + 1)!

Here, ∝ = 1, 𝛽 = 1

1! × 1! 1×1 1
𝐿 𝐿 𝑑𝑥 = ×𝑙 = ×𝑙 = ×𝑙
(1 + 1 + 1)! 3! 3×2×1

𝑙
𝐿 𝐿 𝑑𝑥 =
6
2.14 One Dimensional Problems

Example 2.4

Determine the value of ∫ 𝐿 𝑑𝑥.

Solution: we know that,

𝛼! 𝛽!
𝐿∝ 𝐿 𝑑𝑥 = ×𝑙
(𝛼 + 𝛽 + 1)!

Here, ∝ = 3, 𝛽 = 0
3! 0! (3 × 2 × 1)
𝐿 𝐿 𝑑𝑥 = ×𝑙 = ×𝑙
(3 + 0 + 1)! 4!
3×2 ×1
= ×𝑙
4× 3×2×1
𝑙
𝐿 𝐿 𝑑𝑥 =
4

2.5 SHAPE FUNCTIONS

2.5.1. Introduction
If the values of the field variable are computed only at nodes, how are values
obtained at other nodal points within a finite element? This is a most important point
of finite element analysis.
The values of the field variable computed at the nodes are used to approximate
the values at non-nodal points by interpolation of the nodal values.

Fig.2.17
Consider the three noded triangular element as shown in Fig.2.17.
The nodes are exterior and at any point within the element the field variable is
described by the following approximate relation.
 Finite Element Analysis 2.15

∅(𝑥, 𝑦) = 𝑁 (𝑥, 𝑦)∅ + 𝑁 (𝑥, 𝑦)∅ + 𝑁 (𝑥, 𝑦)∅

where ∅ , ∅ , ∅ are the values of the field variable at the nodes, and N1, N2 and
N3 are the interpolation functions. N1, N2 and N3 are also called as shape functions
because they are used to express the geometry or shape of the element. Shape function
has unit value at one nodal point and zero value at other nodal points.
In one dimensional problem, the basic field variable is displacement.

𝑢= 𝑁𝑢

For two noded bar element, the displacement at any point within the element is given by,

𝑢= 𝑁𝑢 = 𝑁 𝑢 +𝑁 𝑢

where u1 and u2 are nodel displacements.

Fig. 2.18
In two dimensional stress analysis problem, the basic field variable is displacement.

So, 𝑢= 𝑁𝑢

𝑣= 𝑁𝑣

For three noded triangular element, the displacement at any point within the
element is given by,

𝑢= 𝑁𝑢 =𝑁 𝑢 +𝑁 𝑢 +𝑁 𝑢

𝑢= 𝑁 𝑢𝑣 = 𝑁 𝑣 + 𝑁 𝑣 + 𝑁 𝑣

Where, u1, u2, u3, v1, v2, v3 are nodal displacements.

In general, shape functions need to satisfy the following:
1. First derivatives should be finite within an element.
2. Displacement should he continuous across the element boundary.
2.16 One Dimensional Problems

The characteristics of shape function are:

1. The shape function has unit value at its own nodal point and zero value at other
node points.
2. The sum of shape function is equal to one.
3. The shape functions for two dimensional elements are zero along' each side that
the node does not touch.
4. The shape functions are always polynomials of the same type as the original
interpolation equations.

2.5.2. Polynomial Shape Functions

Polynomials are generally used as shape function due to the following reasons.
1. Differentiation and integration of polynomials are quite easy.
2. It is easy to formulate and computerize t h e finite element equations.
3. The accuracy of the results can be improved by increasing the order of the
polynomial.
The approximation of a non-linear one dimensional function by using
polynomials of different order is shown in Fig.2.19.

Fig. 2.19 Approximation of a function by polynomials of different order

Let us consider displacement u is a field variable.
Case (i): linear polynomial
For one dimensional problem, 𝑢 = 𝑎 + 𝑎 𝑥
For two dimensional problem, 𝑢(𝑥, 𝑦) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑦
For three dimensional problem, 𝑢(𝑥, 𝑦, 𝑧) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑧
 Finite Element Analysis 2.17

Case(ii): Quadratic polynomial

For one dimensional problem,
𝑢 =𝑎 +𝑎 𝑥+𝑎 𝑥
For two dimensional problem,
𝑢(𝑥, 𝑦) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑥𝑦
For three dimensional problem,
𝑢(𝑥, 𝑦, 𝑧) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑧 + 𝑎 𝑥 + 𝑎 𝑦 + 𝑎 𝑧 + 𝑎 𝑥𝑦
+ 𝑎 𝑦𝑧 + 𝑎 𝑥𝑧

2.5.3. Derivation of the displacement function u and shape function N for

one dimensional Linear bar element based on global co-ordinate approach
Consider a bar element with nodes 1 and 2 as shown in Fig.2.20. u1and u2 are
the displacements at the respective nodes. So, u1 and u2 are considered as degrees of
freedom of this bar element.
[Note: Degrees of freedom is nothing but nodal displacements.]

Fig 2.20 Two noded bar element

Since the element has got two degrees of freedom, it will have two generalized
co- ordinates.
𝑢 =𝑎 +𝑎 𝑥 … (2.15)
a0 and a1 are global or generalized co-ordinates.
Writing the equation (2.15) in matrix form,
𝑎
𝑢 = [1 𝑥] . . (2.16)
𝑎
𝐴𝑡 𝑛𝑜𝑑𝑒 1, 𝑢 = 𝑢 , 𝑥 = 0
2.18 One Dimensional Problems

𝐴𝑡 𝑛𝑜𝑑𝑒 2, 𝑢 = 𝑢 , 𝑥 = 𝑙
Substitute the above value in equation (2.15),
𝑢 =𝑎 … (2.17)
𝑢 =𝑎 +𝑎 𝑙 … (2.18)
Arranging the equation (2.17),(2.18) in matrix form,
𝑢 1 0 𝑎
𝑢 𝑎
= 1 𝑙 … (2.19)
↓ ↓ ↓
𝑢∗ 𝐶 𝐴
Where u*  Degrees of freedom.
C  Connectivity matrix.
A  Generalized or global co-ordinates matrix.
𝑎 1 0 𝑢
 𝑎 = 𝑢
1 𝑙

1 1 −0 𝑢
= 𝑢
𝑙−0 −1 𝑙

𝑎 𝑎 1 𝑎 −𝑎
𝑁𝑜𝑡𝑒 𝑎 𝑎 = × −𝑎 𝑎
(𝑎 𝑎 −𝑎 𝑎 )

𝑎 1 1 0 𝑢
 𝑎 = 𝑢
𝑙 −1 𝑙
𝑎
Substitute 𝑎 values in equation (2.16),

1 𝑙 0 𝑢
 𝑢 = [ 1 𝑥]
𝑙 −1 1 𝑢
1 0 𝑢
= [ 1 𝑥] 𝑙
𝑙 −1 1 𝑢
1 𝑢
= [1 𝑥 0 + 𝑥] 𝑢
𝑙
 Finite Element Analysis 2.19
[∵ 𝑀𝑎𝑡𝑟𝑖𝑥 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑐𝑎𝑡𝑖𝑜𝑜𝑛 (1 × 2) × (2 × 2) = (1 × 2)]
𝑙−𝑥 𝑥 𝑢
𝑢= 𝑢 … (2.20)
𝑙 𝑙
𝑢
𝑢 = [𝑁 𝑁] 𝑢

Displacement function

[𝑢 = 𝑁 𝑢 + 𝑁 𝑢 ] … . (2.21)

𝑙−𝑥 𝑥
Where, shape function, 𝑁 = ; Shape function, N =
𝑙 𝑙

We may note that N1 and N2 obey the definition of shape function, i.e., the shape
function will have a value equal to unity at node to which it belongs and zero value at
other nodes
Checking: At node 1, x=0.

𝑙−𝑥 𝑙−0
 𝑁 = =
𝑙 𝑙

𝑁 =1

𝑥 0
 𝑁 = =
𝑙 𝑙

𝑁 =0

At node 2, x =1

𝑙−𝑥 𝑙−0
 𝑁 = =
𝑙 𝑙

𝑁 =0

𝑥 0
 𝑁 = =
𝑙 𝑙

𝑁 =1
2.20 One Dimensional Problems

2.6. THE POTENTIAL-ENERGY APPROACH

We know that
The general expression for the potential energy is given by,

1
𝜋= 𝜎 𝑒 𝐴 𝑑𝑥 − 𝑢 𝑓𝐴 𝑑𝑥 − 𝑢 𝑇𝑑𝑥 − 𝑢𝑃
2

Where,   Stress, N/mm2

e  Strain
A  Area, mm2
u  Displacement, mm
f  Body force, N
T  Traciton force, N
P  Point load, N
When the continuum has been discretized into finite elements, the expression for
 becomes as follows:

1
𝜋= 𝜎 𝑒 𝐴 𝑑𝑥 − 𝑢 𝑓𝐴 𝑑𝑥 − 𝑢 𝑇 𝑑𝑥 − 𝑄𝑃
2

The above equation can be written as,

𝜋= 𝑈 − 𝑢 𝑓𝐴 𝑑𝑥 − 𝑢 𝑇 𝑑𝑥 − 𝑄𝑃 … . (1)

1
where, strain energy, 𝑈 = 𝜎 𝑒 𝐴 𝑑𝑥
2
Stiffness matrix for a bar element:
We know that,
1
Strain energy, 𝑈 = 𝜎 𝑒 𝐴 𝑑𝑥 … (2)
2
 Finite Element Analysis 2.21

we know that,
Strain, 𝑒 = 𝐵 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎
Strain, 𝜎 = 𝐸 𝑒 𝑌𝑜𝑢𝑛𝑔 𝑠 𝑚𝑜𝑑𝑢𝑙𝑢𝑠, 𝐸 =
𝑆𝑡𝑟𝑎𝑖𝑛, 𝑒
 𝑆𝑡𝑟𝑎𝑖𝑛, 𝜎 = 𝐸 × 𝐵𝑢
Subsitute  and e values in equation (2),
1 1
𝑈 = (𝐸 𝐵 𝑢) (𝐵 𝑢) 𝐴 𝑑𝑥 = 𝐸 𝐵 𝑢 𝐵 𝑢 𝐴 𝑑𝑥
2 2
1
= 𝑢 𝐸𝐵 𝐵 𝑢 𝐴 𝑑𝑥
2

1
= 𝑢 𝐴𝐸𝐵 𝐵𝑢 𝑑𝑥
2

1
= 𝑢 [𝐴 𝐸 𝐵 𝐵 𝑢 [𝑥] ]
2
1
𝑈 = 𝑢 [𝐴 𝐸 𝐵 𝐵 𝑢 (𝑙)] … (3)
2
From equation (2.31), we know that,
Strain displacement matrix,
−1 1
[ 𝐵] =
𝑙 𝑙
−1
 [ 𝐵] = 𝑙
1
𝑙
Substitute B, BT values in equation (2.94)
−1
1 −1 1
 𝑈 = 𝑢 . 𝐴𝐸 𝑙 𝑢𝑙
2 1 𝑙 𝑙
𝑙
2.22 One Dimensional Problems
1 1 −1 1
= 𝑢 . 𝐴𝐸 × × [−1 1]𝑢 𝑙
2 𝑙 1 𝑙
1 𝐴𝐸 −1 [−1
= 𝑢 1]𝑢
2 𝑙 1
1 𝐴𝐸 1 −1
𝑈 = 𝑢 𝑢
2 𝑙 −1 1
[∵ (2 × 1) × (1 × 2) = 2 × 2]
The above equation is in the form of
1
𝑈 = 𝑢 [𝐾 ]𝑢
2
Where, Stiffness matrix,
𝐴𝐸 1 −1
[𝐾 ] = … (4)
𝑙 −1 1

2.7. STIFFNESSS MATRIX [K]

In order to get an expression for the stiffness matrix in finite element method, let
us review the strain energy expression in structural mechanics.
Consider ω1, ω2…. ωn are nodal displacement parameters or otherwise known as
degrees of freedom, W1,W2,…Wn are the corresponding nodal loads acting at degrees
of freedom. {ω} and {W} are column matrix.
𝑊
⎧ ⎫
⎪𝑊 ⎪
{𝑊} = 𝑊
⎨ : ⎬
⎪ ⎪
⎩𝑊 ⎭
𝜔
⎧𝜔 ⎫
⎪ ⎪
{𝜔} = 𝜔
⎨ ⋮ ⎬
⎪ ⎪
⎩𝜔 ⎭
We know that, {W} = [K] {𝜔 ∗} …… (2.22)
Where, W = Nodal loads.
K = stiffness matrix.
 Finite Element Analysis 2.23

𝜔 = degrees of freedom
From equation (2.22), we know that, nodal loads and the corresponding degrees
of freedom of freedom are linked through stiffness matrix.
We know that.
Work done, P= strain energy
1 1 1 1
 𝑃= 𝑊 𝜔 + 𝑊 𝜔 + 𝑊 𝜔 + ⋯……+ 𝑊 𝜔
2 2 2 2
We can write this equation in matrix form,
𝜔
⎧𝜔 ⎫
1 ⎪ ⎪
i. e. , 𝑃 = [𝑊 𝑊 𝑊 … … . . 𝑊 ] 𝜔
2 ⎨ ⋮ ⎬
⎪ ⎪
⎩𝜔 ⎭
1
𝑃= {𝑊} {𝜔∗ } … (2.23)
2
[Note: [ ]  Row matrix; { }  Column matrix.
1
𝑃= {[ K]{𝜔}} {𝜔∗ }
2
1
𝑃= [𝐾] {𝜔} {𝜔∗ }
2
1
𝑃 = {𝜔∗ } [𝐾] {𝜔∗ } … (2.24)
2
[ K is a symmetric matrix. So, [K] T= [ K] ]
Equation (2.24) is a strain energy equation for a structure.
Our aim is to find the expression for stiffness matrix [K]. Let us consider one
dimensional element. u1, u2, u3,…..,un are the degrees of freedom of that element.
We know that,
Strain, { 𝑒} = [ 𝐵] {𝑢∗ } … (2.25)

 {𝑒} = [𝐵] { 𝑢∗ } … (2.26)

Where, { e} is a strain matrix [ Column matrix].
2.24 One Dimensional Problems

[ B ] is a strain-displacement matrix [Row matrix].

{ u*} is a degree of freedom [ column matrix].
We know that,
𝑆𝑡𝑟𝑒𝑠𝑠 {𝜎} = [𝐸]{ 𝑒}
{𝜎} = [𝐷]{ 𝑒} … (2.27)
𝑊ℎ𝑒𝑟𝑒, [ 𝐸] = [𝐷] = 𝑌𝑜𝑢𝑛𝑔 𝑠 𝑚𝑜𝑑𝑢𝑙𝑢𝑠.
Strain energy expression is given by,

1
𝑈= { 𝑒} {𝜎} 𝑑𝑣 … (2.28)
2

Substitute {e}T and {} values,

1
 𝑈= [𝐵] {𝑢∗ } [𝐷]{𝑒} 𝑑𝑣
2

1 ∗ 1
{𝑢 } [𝐵] [𝐷]{𝑒} 𝑑𝑣
2 2

Substitute {e } value,
1
 𝑈= { 𝑢∗ } [𝐵] [ 𝐷][𝐵] {𝑢∗ } 𝑑𝑣
2

1
𝑈= { 𝑢∗ } [𝐵] [ 𝐷][𝐵] {𝑢∗ } 𝑑𝑣 {𝑢∗ } … (2.29}
2
From equation (2.24), we know that,
1 ∗
𝑃= {𝜔 } [ 𝐾]{ 𝜔∗ } … (2.24)
2
Comparing equation (2.29) and (2.24),
 {𝜔∗ } = {𝑢∗ }
{𝜔∗ } = {𝑢∗ }
 Finite Element Analysis 2.25

[𝐾 ] = [ 𝐵] [𝐷][𝐵]𝑑𝑣

So, stiffness matrix, [𝐾 ] = [ 𝐵] [𝐷][𝐵]𝑑𝑣 … (2.30)

Where, [B]  strain displacement relationship matrix.

[D]  Elasticity matrix or Stress-strain displacement relationship matrix.

In one dimensional problem,
𝑑𝑢
Strian, 𝑒=
𝑑𝑥

Where, u  Displacement function.

[D] = [E] = E= Young’s modulus.

In Beam problem, Strain,

𝑑 𝑢
𝑒 = 𝐶𝑢𝑟𝑣𝑎𝑡𝑢𝑟𝑒 =
𝑑𝑥

[D] = [ E I ] = Flexural rigidity.

2.7.1. Properties of Stiffness Matrix

1. It is a symmetric matrix.
2. The sum of elements in any column must be equal to zero.
3. It is an unstable element. So, the determinant is equal to zero.
4. The dimension of the global stiffness matrix [ K] is N x N, where N is the
number of nodes. This follows from the fact that each node has only one degree
of freedom.
5. The diagonal coefficients are always positive and relatively large when
compared to the off-diagonal values in the same row.
2.26 One Dimensional Problems

2.7.2. Derivation of Stiffness Matrix for One Dimensional Linear Bar

Element
Consider a one-dimensional bar element with nodes 1 and 2 as shown in
Fig.2.21. Let u1 and u2 be the nodal displacement parameters or otherwise known as
degrees of freedom.

Fig. 2.21. A bar element with two nodes

We know that,

Stiffness matrix [𝐾] = [ B] [D][B]dv [From equation no. (2.30)]

In one dimensinal bar element,

Displacement function, 𝑢 = 𝑁 𝑢 + 𝑁 𝑢 [From equation no. (2.21)]
𝑙−𝑥
𝑤ℎ𝑒𝑟𝑒, 𝑁 =
𝑙
𝑥
𝑁 =
𝑙
We know that,
𝑑𝑁 𝑑𝑁
Strain − Displacement matrix, [𝐵] =
𝑑𝑥 𝑑𝑥
−1 1
[𝐵] = … (2.31)
𝑙 𝑙
−1
 [𝐵] = 𝑙 … (2.32)
1
𝑙
 Finite Element Analysis 2.27

In one dimensional problems, [D] = [E] = E =Young’s modulus …(2.33)

Substitute [B], [B]T and [D] values in stiffness matrix equation. [Limit is 0 to l].
−1 1 −1
 [𝐾] = 𝑙 × 𝐸 × −1 1 𝑑𝑣 = 𝑙 𝑙 𝐸 𝑑𝑣
1 𝑙 𝑙 −1 1
𝑙 𝑙 𝑙
[ ∵ Matrix multiplication (2 × 1) × (1 × 2) = (2 × 2)]
1 −1
= 𝑙 𝑙 𝐸 𝐴 𝑑𝑥 [𝑑𝑣 = 𝐴 𝑑𝑥]
−1 1
𝑙 𝑙
1 −1 1 −1
=𝐴𝐸 𝑙 𝑙 𝑑𝑥 = 𝐴 𝐸 𝑙 𝑙 [𝑥]
−1 1 −1 1
𝑙 𝑙 𝑙 𝑙
1 −1 1 −1
=𝐴𝐸 𝑙 𝑙 (𝑙 − 0) = 𝐴 𝐸 𝑙 𝑙 𝑙
−1 1 −1 1
𝑙 𝑙 𝑙 𝑙
𝐴 𝐸 𝑙 1 −1
=
𝑙 −1 1
𝐴 𝐸 1 −1
[𝐾]= … (2.34)
𝑙 −1 1
The properties of a stiffness matrix are satisfied.
1. It is symmetric.
2. The sum of elements in any column is equal to zero.

2.7.3. Derivation of Finite Element Equation for one Dimensional Linear

Bar Element
We know that, General force equation is,
{F} = [K] {u} ... (2.35)
where, {F} is a element force vector [Column matrix].
2.28 One Dimensional Problems

[K] is a stiffness matrix [Row matrix].

{u} is a nodal displacement [Column matrix].
For one dimensional bar element, stiffness matrix [K ] is given by,
𝐴𝐸 1 −1
[𝐾]= [From equation no. (2.34)]
𝑙 −1 1
For one noded bar element,
𝐹
{𝐹} =
𝐹
𝑢
{𝑢} = 𝑢

Substitute [K ] {F} and {u} values in equation (2.35),

𝐹 𝐴𝐸 1 −1 𝑢
 = … (2.36)
𝐹 𝑙 −1 1 𝑢
This is a Finite Element Equation for One Dimensional two noded elements.

2.7.4. Assembling The Stiffness Equations or Global Equations

Consider a bar as shown in Fig.2.22(a). This bar can be equally divided into 4
elements as shown in Fig.2.22(b).

Fig. 2.22 (a) Fig. 2.22 (b)

Now the bar has 4 elements with 5 nodes
[Note: A number with circle denotes element and without circle denotes nodes].
We know that,
Finite element equation for two noded bar element is,
 Finite Element Analysis 2.29

𝐹 𝐴𝐸 1 −1 𝑢
=
𝐹 𝑙 −1 1 𝑢
For element (1) (Nodes 1, 2):

Finite element equation is,

𝑎 𝑎 𝑢
𝐹 𝐴𝐸 1
= −1 … (2.37)
𝐹 𝑙 𝑎 𝑎
𝑢
−1 1
For element (2) (Nodes 2, 3):

Finite element equation is,

𝑎 𝑎 𝑢
𝐹 𝐴𝐸 1
= −1 … (2.38)
𝐹 𝑙 𝑎 𝑎
𝑢
−1 1
For element (3) (Nodes 3, 4):

Finite element equation is,

𝑎 𝑎 𝑢
𝐹 𝐴𝐸 1
= −1 … (2.39)
𝐹 𝑙 𝑎 𝑎
𝑢
−1 1
2.30 One Dimensional Problems

For element (4) (Nodes 4, 5):

Finite element equation is,

𝑎 𝑎 𝑢
𝐹 𝐴𝐸 1
= −1 … (2.40)
𝐹 𝑙 𝑎 𝑎
𝑢
−1 1
Assembling the equations (2.37), (2.38), (2.39) and (2.40),
𝐹 𝑎 𝑎 𝑎 𝑎 𝑎 𝑢
⎧ ⎫ ⎡ 1 ⎤ ⎧ ⎫
−1 0 0 0
⎪𝐹 ⎪ ⎢𝑎 𝑎 𝑎 𝑎 𝑞 ⎥ ⎪𝑢 ⎪
⎪ ⎪ ⎢ −1
⎪ ⎪ 1 + 1 −1 0 0 ⎥⎪ ⎪
𝐴𝐸 ⎢𝑎 𝑎 𝑎 𝑎 𝑎 ⎥ 𝑢
𝐹 =
⎨ ⎬ 𝑙 ⎢ 0 −1 1 + 1 −1 0 ⎥⎨ ⎬
⎢𝑎 𝑎 𝑎 𝑎 𝑎 ⎥ 𝑢
⎪𝐹 ⎪ ⎢ 0 ⎪ ⎪
⎪ ⎪ 0 −1 1 + 1 −1 ⎥ ⎪ ⎪
⎪ ⎪ ⎢𝑎 𝑎 𝑎 𝑎 𝑎 ⎥
⎩𝐹 ⎭ ⎣ 0 0 0 −1 1 ⎦ ⎩𝑢 ⎭

𝐹 1 −1 0 0 0 𝑢
⎧𝐹 ⎫ ⎡−1 2 −1 0 ⎤ ⎧𝑢 ⎫
⎪ ⎪ 𝐴𝐸 ⎢ 0 ⎪ ⎪
𝐹 = ⎢ 0 −1 2 −1 0 ⎥⎥ 𝑢
⎨𝐹 ⎬ 𝑙 0 −1 2 −1⎥ ⎨ 𝑢 ⎬
⎪ ⎪ ⎢0 ⎪ ⎪
⎩𝐹 ⎭ ⎣0 0 0 −1 1 ⎦ ⎩𝑢 ⎭

[Note : The bar has 5 nodes and each node has one degree of freedom. So, the global
stiffness matrix size is 5 × 5].
1 −1 0 0 0
⎡−1 2 −1 0 0⎤
[𝐾] ⎢ 0 −1 2 −1 0 ⎥
= ⎢ ⎥
⎢0 0 −1 2 −1⎥
⎣0 0 0 −1 1 ⎦
 Finite Element Analysis 2.31

2.8 THE LOAD OR FORCE VECTOR {F}

Consider a vertically hanging bar of length l, uniform cross-section A, density ρ
and young’s modulus E. This bar is subjected to self-weight Xb.

Fig. 2.23. Vertically hanging bar will self-weight

The element nodal force vector is given by,

{𝐹} = [ 𝑁] 𝑋 … (2.41)

We know that, self weight due to loading force,

𝑋 = ρ 𝐴 𝑑𝑥 … (2.42)
For one dimensional bar element, the displacement function is given by,
𝑢 =𝑁 𝑢 +𝑁 𝑢 [From equaiton no. (2.21)]
𝑙−𝑥
𝑤ℎ𝑒𝑟𝑒, 𝑁 =
𝑙
𝑥
𝑁 =
𝑙
𝑙−𝑥 𝑥
 [𝑁] =
𝑙 𝑙
𝑙−𝑥
 [𝑁] = 𝑙 … (2.43)
𝑥
𝑙
Substitute Xb and [N]T values in equation (2.41),
𝑙−𝑥 𝑙−𝑥
 {𝐹} = 𝑙 𝜌 𝐴 𝑑𝑥 = 𝜌 𝐴 𝑙 𝑑𝑥
𝑥 𝑥
𝑙 𝑙
2.32 One Dimensional Problems
𝑥 𝑥 𝑑𝑥
1− 𝑑𝑥 −
=𝜌𝐴 𝑙 𝑑𝑥 = 𝜌 𝐴 𝑙
𝑥 𝑥 𝑑𝑥
𝑙 𝑙
𝑥 𝑙 𝑙
⎧𝑥 −⎫ ⎧𝑙 − ⎫ 𝑙−
=𝜌𝐴 2 𝑙 = 𝜌𝐴 2𝑙 = 𝜌𝐴 2
⎨ 𝑥 ⎬ ⎨ 𝑙 ⎬ 𝑙
⎩ 2𝑙 ⎭ ⎩ 2𝑙 ⎭ 2
𝑙
= 𝜌𝐴 2
𝑙
2
𝜌𝐴𝑙 1
Force vector, [𝐹] = . . (2.44)
2 1

2.9. SOLVED PROBLEMS ON LINEAR BAR ELEMENTS

Example 2.5
A two noded truss element is shown in fig(i). The nodal displacement are
u1=5mm and u2=8mm. calculate the displacement at x=l/4, l/3 and l/2.

Fig. (i)
Given : Displacement , u1 = 5 mm
u2 = 8 mm

Fig. (ii)
 Finite Element Analysis 2.33

To find: Displacement u at
𝑙 𝑙 𝑙
𝑥= , 𝑎𝑛𝑑 .
4 3 2
Solution: Displacement function for two noded truss element is given by,
𝑢 =𝑁 𝑢 +𝑁 𝑢 [From equation (2.21)]
𝑙−𝑥
where, 𝑁 =
𝑙
𝑥
𝑁 =
𝑙
𝑙−𝑥 𝑥
 𝑢= 𝑢 + 𝑢 … (1)
𝑙 𝑙
1
Substitute 𝑥 = , 𝑢 = 5 𝑎𝑛𝑑 𝑢 = 8 𝑖𝑛 equaiton (1),
4
1 𝑙
𝑙−4
 𝑢 = ×5+ 4 ×8
𝑙 𝑙

1 1
= 1− ×5+ ×8
4 4
𝑙
𝑢 = 5.75 𝑚𝑚 𝑎𝑡 𝑥 =
4
1
Substitute 𝑥 = , 𝑢 = 5 𝑎𝑛𝑑 𝑢 = 8 𝑚𝑚 𝑖𝑛 𝑒𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (1),
3
𝑙 𝑙
𝑙−3
(1) 𝑢 = ×5+ 3 ×8
𝑙 𝑙

1 1
= 1− ×5+ ×8
3 3
1 1
= 1− ×5+ ×8
4 4
𝑙
𝑢 = 6 𝑚𝑚 𝑎𝑡 𝑥 =
3
2.34 One Dimensional Problems
1
Substitute 𝑥 = , 𝑢 = 5 𝑎𝑛𝑑 𝑢 = 8 𝑚𝑚 𝑖𝑛 𝑒𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (1),
2
𝑙 𝑙
𝑙−2
(1) 𝑢 = ×5+ 2 ×8
𝑙 𝑙

1 1
= 1− ×5+ ×8
2 2
𝑙
𝑢 = 6.5 𝑚𝑚 𝑎𝑡 𝑥 =
2
Result:
𝑙
𝑢 = 5.75 𝑚𝑚 𝑎𝑡 𝑥 =
4
𝑙
𝑢 = 6 𝑚𝑚 𝑎𝑡 𝑥 =
3
𝑙
𝑢 = 6.5 𝑚𝑚 𝑎𝑡 𝑥 =
2

Example 2.6
A one dimensional bar is shown in fig.(i). calculate the following:
(i) Shape function N1 and N2 at point P
(ii) If U1=3mm and U2=-5mm, calculate the displacement u at point P.

Fig. (i)
Given: 𝑥 = 20 𝑚𝑚, 𝑥 = 36 𝑚𝑚, 𝑢 = 3 𝑚𝑚, 𝑢 = −5 𝑚𝑚, 𝑥 = 24 𝑚𝑚

To find: 1. Shape function N1 and N2 at point P.

2. Displacement u at point P.
 Finite Element Analysis 2.35

Solution:

Fig. (ii)
We know that,
Actual length of the bar, 𝑙 = 𝑥 − 𝑥 = 36 − 20
𝑙 = 16 𝑚𝑚
The distance between point 1 and point P is
x = 24 – 20
 x = 4 mm
Displacement function for two noded bar element is given by
𝑢 =𝑁 𝑢 +𝑁 𝑢 … (1)
[From equation no.(2.21)]
𝑙−𝑥
where, 𝑁 =
𝑙
𝑥
𝑁 =
𝑙
16 − 4
 𝑁 =
16
𝑁 = 0.75 𝑚𝑚
4
 𝑁 =
16
𝑁 = 0.25 𝑚𝑚
Substitute 𝑁 , 𝑁 , 𝑢 𝑎𝑛𝑑 𝑢 𝑣𝑎𝑙𝑢𝑒𝑠 𝑖𝑛 𝑒𝑞𝑢𝑎𝑖𝑡𝑜𝑛 𝑛𝑜. (1),
(1) 𝑢 = 𝑁 𝑢 + 𝑁 𝑢 = (0.75)(3) + 0.25(−5)
𝑢 = 1 𝑚𝑚
2.36 One Dimensional Problems

Result: 1. Shape function, N1 = 0.75 mm and N2 = 0.25 mm

2. Displacement u at point P is 1mm

Example 2.7
Consider a bar as shown in fig. (i). cross-sectional area of the bar is 750mm2 and
young’s modulus is 2*105 N/mm2. If u1=0.5mm and u2=0.625mm, calculate the
following: (i) Displacement at point, P (ii) Strain, e (iii) Stress, (iv) Elements stiffness
matrix,[K] (v)strain energy, U

Fig. (i)
Given: Area A = 750 mm2
Young’s Modulus, E = 2 × 105 N/mm2
Displacement, 𝑢 = 0.5 mm
𝑢 = 0.625 𝑚𝑚
Distance, 𝑥 = 375 𝑚𝑚
𝑥 = 575 𝑚𝑚
𝑥 = 500 𝑚𝑚
Solution: We know that, strain,
𝑒 = [𝐵] {𝑢∗ } [From equaiton no. (2.25)]
Where, [B] is a strain-displacement matrix.
{𝑢∗ } is a degree of freedom
−1 1
 [𝐵] = [From equaiton no. (2.31)]
𝑙 𝑙
−1 1
=
200 200
−1 1 𝑢
Strain, 𝑒 = [𝐵]{𝑢∗ } =
200 200 𝑢
 Finite Element Analysis 2.37
−1 1 0.5
=
200 200 0.625
−1 1
= × 0.5 + × 0.625
200 200
Strain, 𝑒 = 6.25 × 10
We know that,
𝑆𝑡𝑟𝑎𝑖𝑛,  = 𝐸 𝑒 = 2 × 10 × 6.25 × 10
 = 125 𝑁/𝑚𝑚
For one dimensional bar element, stiffness matrix is given by,
𝐴𝐸 1 −1 750 × 2 × 10 1 −1
[𝐾 ] = =
𝑙 −1 1 200 −1 1

[𝐾 ] = 7.5 × 10 1 −1
−1 1
1 ∗
We know that, Strain, 𝑈 = {𝑢 } [ 𝐾]{ 𝑢∗ }
2
…….. [From equation no. (2.24)]
1 1 −1 𝑢
= [ 𝑢 𝑢 ] × 7.5 × 10 𝑢
2 −1 1
1 1 −1 0.5
= [ 0.5 0.625] × 7.5 × 10
2 −1 1 0.625
1 0.5 −0.625
= [ 0.5 0.625] × 7.5 × 10
2 −0.5 +0.625

[∵ (2 × 2 ) × (2 × 1) = 2 × 1]

1 −0.125
= × 7.5 × 10 [0.5 0.625]
2 0.125
1
= × 7.5 × 10 [0.5 × (−0.125) + 0.625 × 0.125 ]
2

Strain energy, U =5859.37 N-mm

2.38 One Dimensional Problems

Result:
(i) u = 0.5781 mm
(ii) 𝑒 = 6.25 × 10
(iii)  = 125 𝑁/𝑚𝑚2
(iv) [𝐾 ] = 7.5 × 10
(v) U = 5859.37 N-mm

Example 2.8
A steel bar of length 800 mm is subjected to all axial load of 3 kN as shown
in Fig. (i). Find the elongation of the bar, neglecting self weight.

Fig. (i)
Take E = 2 × 105 N/mm2, A = 300 mm2
Given: Length, l = 800 mm
Load, F = 3 kN = 3 × 103 N
Young’s modulus, E = 2 × 105 N/mm2
Area, A = 300 mm2
To find: Elongation, u
Solution: we can divide the bar into two elements as shown in Fig.(ii).

Fig. (ii)
 Finite Element Analysis 2.39

Now the bar has 2 elements with 3 nodes.

[Note: Number with circle denotes Element & Number without circle denotes Node]
We can find the displacement at node 1, node 2 and node 3 .

Displacement at node 1 is u1, node 2 is u2 and node 3 is u3.

For one dimensional two noded bar element, the finite element equation is,
𝐹 𝐴𝐸 1 −1 𝑢
= [From equaiton no. (2.36)]
𝐹 𝑙 −1 1 𝑢
For element 1: (Nodes 1, 2):
Finite element equation is,
𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

300 × 2 × 10 1 −1 𝑢 𝐹
=
400 −1 1 𝑢 𝐹
𝑎 𝑎 𝑢 𝐹
150 × 10 1 −1 = … . . (1)
𝑎 𝑎
−1 1 𝑢 𝐹
For element 2: (Nodes 2, 3):
Finite element equation is,
2.40 One Dimensional Problems
𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

300 × 2 × 10 1 −1 𝑢 𝐹
 =
400 −1 1 𝑢 𝐹
𝑎 𝑎 𝑢 𝐹
 150 × 10 1 −1 = … . . (2)
𝑎 𝑎
−1 1 𝑢 𝐹
Assemble the finite elements, i.e., assemble the finite element equation (1) and (2).
𝑎 𝑎 𝑎
⎡ 1 −1 0 ⎤ 𝑢 𝐹
⎢𝑎 𝑎 𝑎 ⎥ 𝑢
 150 × 10 ⎢ −1 = 𝐹
1 + 1 −1 ⎥ 𝑢
⎢𝑎 𝑎 𝑎 ⎥ 𝐹
⎣ 0 −1 1 ⎦
1 −1 0 𝑢 𝐹
 150 × 10 −1 2 −1 𝑢 = 𝐹
0 −1 1 𝑢 𝐹 … (3)
↓
[𝐾] 𝑔𝑙𝑜𝑏𝑎𝑙
[Note : The rod has 3 nodes. Each node has single degree of freedom. So, the global
stiffness matrix [K] size is 3 × 3.
It may be noted the stiffness matrix properties are satisfied.
1 −1 0
[𝐾 ] = −1 2 −1
0 −1 1
1. It is symmetric.
2. The sum of elements in any column is equal to zero.
 Finite Element Analysis 2.41

Applying Boundary Conditions

(i) Displacement at node 1 is zero. i.e., u1 = 0.
(ii) 3 x 103 N load is acting at node 3, i.e., F3=3 x 103 N. Self-weight is neglected,
so, F1= F2 = 0.
Substitute u 1 , F 1 , F2 and F3 values in equation (3),
1 −1 0 0 0
 150 × 10 −1 2 −1 𝑢 = 0
0 −1 1 𝑢 3 × 10
Here, u1 = 0. So, neglect first row and first column of [K] matrix, Hence, the final
reduced equation is,
2 −1 𝑢 0
 150 × 10 𝑢 =
−1 1 3 × 10

150 × 10 (2𝑢 − 𝑢 ) = 0 … (4)

150 × 10 (−𝑢 + 𝑢 ) = 3 × 10 … (5)

𝑆𝑜𝑙𝑣𝑖𝑛𝑔, 150 × 10 (𝑢 ) = 3 × 10
𝑢 = 0.02 𝑚𝑚
Substitute u2 value in equation (4),
 150 × 10 (2 × 0.02 − 𝑢 ) = 0

 2 × 0.02 − 𝑢 = 0

 2 × 0.02 = 𝑢

 𝑢 = 0.04 𝑚𝑚
Verification: We know that, Total elongation,
𝑝𝐿
𝐿 =
𝐴𝐸
3 × 10 × 800
=
300 × 2 × 10
 𝐿 = 0.04 𝑚𝑚
2.42 One Dimensional Problems

Result:
1. Elongation or displacement at node 1, u1 =0
2. At node 2, u2 = 0.02 mm
3. At node 3, u3 = 0.04 mm

Example 2.9
Solve the problem explained in Example 2.8, by dividing the bar into 4
equal parts.
Solution:

Fig. (i)
The bar has 4 elements with 5 nodes as shown in Fig.(i).
We can find the displacement at node 1, node 2, node 3, node 4 and node 5. i.e.,
u1, u2, u3, u4 and u5.
Finite element equation for one dimensional two noded bar element is given by,
𝐹 𝐴𝐸 1 −1 𝑢
=
𝐹 𝑙 −1 1 𝑢
For element 1 (Nodes 1, 2): Finite element equation is,
𝐴𝐸 1 −1 𝑢 𝐹
= [∵ 𝐴 = 𝐴 = 𝐴 = 𝐴 = 300 𝑚𝑚 ]
𝑙 −1 1 𝑢 𝐹
300 × 2 × 10 1 −1 𝑢 𝐹
=
200 −1 1 𝑢 𝐹
 Finite Element Analysis 2.43
𝑎 𝑎
1 −1 𝑢 𝐹
300 × 2 × 10 𝑎 𝑎 𝑢 = … (1)
𝐹
−1 1

For element 2 (Nodes 2, 3): Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

300 × 2 × 10 1 −1 𝑢 𝐹
𝑢 =
200 −1 1 𝐹
𝑎 𝑎
1 −1 𝑢 𝐹
300 × 10 𝑎 𝑎 𝑢 = … (2)
𝐹
−1 1
For element 3 (Nodes 3, 4):
𝐴 𝐸 1 −1 𝑢 𝐹
 =
𝑙 −1 1 𝑢 𝐹

300 × 2 × 10 1 −1 𝑢 𝐹
 =
200 −1 1 𝑢 𝐹
2.44 One Dimensional Problems
𝑎 𝑎
1 −1 𝑢 𝐹
300 × 10 𝑎 𝑎 𝑢 = … (3)
𝐹
−1 1
For element 4: (Nodes 4, 5):
𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹
300 × 2 × 10 1 −1 𝑢 𝐹
=
200 −1 1 𝑢 𝐹

𝑎 𝑎
1 −1 𝑢 𝐹
300 × 10 𝑎 𝑎 𝑢 = … (4)
𝐹
−1 1
Assemble the finite elements, i.e., assemble the finite element equation (1), (2),
(3) and (4)
𝑢 𝐹
𝑎 𝑎 𝑎 𝑎 𝑎 ⎧ ⎫
⎡ 1 −1 0 0 0 ⎤⎧ ⎫ ⎪
⎢𝑎 𝑎 𝑎 𝑎 𝑎 ⎥ ⎪𝑢 ⎪ 𝐹⎪
⎪ ⎪ ⎪ ⎪
⎢ −1 1 + 1 −1 0 0 ⎥⎪ ⎪ ⎪ ⎪
⎢𝑎 𝑎 𝑎 𝑎 𝑎 ⎥ 𝑢 𝐹
300 × 10 ⎢ ⎥ = … (5)
0 −1 1 + 1 −1 0 ⎨ ⎬ ⎨ ⎬
⎢𝑎 𝑎 𝑎 𝑎 𝑎 ⎥ 𝑢
⎢ 0 ⎪ ⎪ ⎪𝐹 ⎪
0 −1 1 + 1 −1 ⎥ ⎪ ⎪ ⎪ ⎪
⎢𝑎 𝑎 𝑎 𝑎 𝑎 ⎥ ⎪𝑢 ⎪
⎣ 0 ⎪𝐹 ⎪
0 0 −1 1 ⎦⎩ ⎭
⎩ ⎭
[Note: the rod has 5 nodes. Each node has single degree of freedom. So, the global
stiffness matrix [K] size is 5 × 5.
It may be noted that the stiffness matrix properties are satisfied.
1. [ K] matrix is symmetric.
2. The sum of the elements in any column is equal to zero.
 Finite Element Analysis 2.45

Applying Boundary Conditions:

Displacement at node 1 is zero. i.e., u1 = 0 self –weight is neglected.
𝑆𝑜, 𝐹 =𝐹 =𝐹 =𝐹 =0

3 × 10 𝑁 𝑙oad is acting at node 5 .

𝑆𝑜, 𝐹 = 3 × 10 𝑁

𝑆𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑒 𝑢 , 𝐹 , 𝐹 , 𝐹 , 𝐹 𝑎𝑛𝑑 𝐹 𝑣𝑎𝑙𝑢𝑒𝑠 𝑖𝑛 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (5).

1 −1 0 0 0 0 0
⎡−1 2 −1 0 ⎧ ⎫ ⎧ ⎫
0⎤ ⎪𝑢 ⎪ ⎪ 0 ⎪
⎢ 0 −1 2 −1 0 ⎥
 150 × 10 ⎢ ⎥ 𝑢 = 0
⎢0 0 −1 2 −1⎥ ⎨𝑢 ⎬ ⎨ 0 ⎬ … (6)
⎪ ⎪ ⎪ ⎪
⎣0 0 0 −1 1 ⎦ ⎩𝑢 ⎭ ⎩3 × 10 ⎭
↓
[𝐾]
In equation (6), u1 = 0. So, neglect first row and first column of [ K ] matrix.
2 −1 0 0 𝑢 0
−1 2 −1 0 𝑢 0
 300 × 10 𝑢 = … (7)
0 −1 2 −1 0
0 0 −1 1 𝑢 3000
By using Gaussian Elimination method, we can find the solution of equation (7).
2 −1 0 0 0
Let , −1 2 −1 0 0
0 −1 2 −1 0
0 0 −1 0 3000
1 −1/2 0 0 0
−1 2 −1 0 0
0 −1 2 −1 0
0 0 −1 1 3000 

1 −1/2 0 0 0
0 3/2 −1 0 0
0 −1 2 −1 0
0 0 −1 1 3000 
2.46 One Dimensional Problems
1 −1/2 0 0 0
0 1 −2/3 0 0
0 −1 2 −1 0
0 0 −1 1 3000  ×

1 −1/2 0 0 0
0 1 −2/3 0 0
0 0 4/3 −1 0
0 0 −1 1 3000 

1 −1/2 0 0 0
0 1 −2/3 0 0
0 0 1 −3/4 0
0 0 −1 1 3000  ×

1 −1/2 0 0 0
0 1 −2/3 0 0
0 0 1 −3/4 0
0 0 −1 1/4 3000 

Equation (7) becomes,

1 −1/2 0 0 𝑢 0
0 1 −2/3 0 𝑢 0
300 × 10 −3/4 𝑢 =
0 0 1 0
0 0 0 1/4 𝑢 3000
1
 300 × 10 0𝑢 +0𝑢 +0𝑢 + 𝑢 = 3000
4
1
 300 × 10 𝑢 = 3000
4
 𝑢 = 0.04 𝑚𝑚
3
 300 × 10 0𝑢 +0𝑢 +𝑢 − 𝑢 =0
4
3
 𝑢 − 𝑢 = 00
4
3
 𝑢 = (0.04)
4
 𝑢 = 0.03 𝑚𝑚
 Finite Element Analysis 2.47
2
 300 × 10 0𝑢 +0𝑢 − 𝑢 +0𝑢 =0
3
2
 𝑢 − 𝑢 = 00
3
2
 𝑢 = (0.03)
3
 𝑢 = 0.02 𝑚𝑚
1
 300 × 10 𝑢 − 𝑢 +0 𝑢 +0𝑢 =0
2
1
 𝑢 − 𝑢 =0
2
1 1
 𝑢 = 𝑢 = × 0.02
2 2
 𝑢 = 0.01 𝑚𝑚
Result: Displacement
𝑢 =0
𝑢 = 0.01 𝑚𝑚
𝑢 = 0.02 𝑚𝑚
𝑢 = 0.03 𝑚𝑚
𝑢 = 0.04 𝑚𝑚

Example 2.10
Consider a bar as shown in Fig. (i). An axial load of 200 kN is applied at
point p. Take Al = 2400 mm2, E1 = 70 × 109 N/m2, A2 = 600 mm2, E2 = 200 x 10
N/m2 .
Calculate the following:
(i) The nodal displacement at point P,
(ii) Stress in each material.
(iii) Reaction force
2.48 One Dimensional Problems

Fig. (i)
Given:

Fig. (ii)
Area of element (1), A1 = 2400 mm2
Area of element (2), A2 = 600 mm2
Length of element (1), l1 = 300 mm
Length of element (2), l2 = 400 mm
Young’s modulus of element (1), E1 = 70 × 109 N/m2 = 70 × 103 N/mm2
Young’s modulus of element (1), E2 = 200 × 109 N/m2 = 200 × 103 N/mm2
Point load, p = 200 Kn = 200 × 103 N
To find:
(i) Nodal displacement at point p, i.e., u2
(ii) Stress in each material, 1 and 2.
(iii)Reaction force, R1, R2.
Soltuion: Finite element equaiton for one dimensional two noded bar element is given
by,
𝐹 𝐴𝐸 1 −1 𝑢
= 𝑢 [From equaiton no. 2.36]
𝐹 𝑙 −1 1
 Finite Element Analysis 2.49

For element 1: (Nodes 1, 2):

Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
𝑢 =
𝑙 −1 1 𝐹

2000 × 70 × 10 1 −1 𝑢 𝐹
 𝑢 =
300 −1 1 𝐹
𝑎 𝑎
5.6 −5.6 𝑢 𝐹
 1 × 10 𝑎 𝑎 𝑢 = … (1)
𝐹
−5.6 5.6

For element 2: (Nodes 2, 3): Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
𝑢 =
𝑙 −1 1 𝐹

600 × 200 × 10 1 −1 𝑢 𝐹
 𝑢 =
400 −1 1 𝐹
𝑎 𝑎
3 −3 𝑢 𝐹
 1 × 10 𝑎 𝑎 𝑢 = … (2)
𝐹
−3 3

Assemble the finite elements. i.e., assemble the finite element equations (1) and (2).
2.50 One Dimensional Problems
𝑎 𝑎 𝑎
⎡ 5.6 −5.6 0 ⎤ 𝑢 𝐹
⎢𝑎 𝑎 𝑎 ⎥ 𝑢
 1 × 10 ⎢ = 𝐹
−5.6 5.6 + 3 −3 ⎥ 𝑢
⎢𝑎 𝑎 𝑎 ⎥ 𝐹
⎣ 0 −3 3 ⎦

5.6 −5.6 0 𝑢 𝐹
 1 × 10 −5.6 8.6 −3 𝑢 = 𝐹
0 −3 3 𝑢 𝐹 … (3)
↓
[𝐾]

[Note: The bar has 3 nodes. Each node has single degree of freedom. So, the
global stiffness matrix [ K ] size is 3× 3. The properties of the stiffness matrix are also
satisfied.
(i) [ K] matrix is symmetric.
(ii) The sum of elements in any column is equal to zero.
Applying boundary conditions:
Displacements at node 1 and node 3 are zero. So, 𝑢 − 1 = 𝑢 =
0. 𝐴 𝑙𝑜𝑎𝑑 𝑜𝑓 200 × 10 𝑁 is acting at node 2. So, 𝐹 = 200 × 10 𝑁. Self-weight is
neglected. i.e., 𝐹 = 𝐹 = 0. Substitute 𝑢 , 𝑢 𝑎𝑛𝑑 𝐹 , 𝐹 𝑎𝑛𝑑 𝐹 values in equaiton (3).
5.6 −5.6 0 0 0
(3)  1 × 10 −5.8 8.6 −3 𝑢 = 2 × 10
0 −3 3 0 0

In the above equation, u1 = 0. So, neglect first row and first column of [ K]
matrix. u3 = 0, so, neglect third row and third column of [ K] matrix. The final reduced
equation is,
1 × 10 [8.6]{ 𝑢 } = { 2 × 10 }

 8.6 × 10 𝑢 = 2 × 10

8.6 𝑢 = 2

𝑢 = 0.2325 𝑚𝑚
 Finite Element Analysis 2.51

Stress in each element:

𝑑𝑢
We know that, 𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 =𝐸
𝑑𝑥
𝑢 −𝑢 (0.2325 − 0)
For element (1), 𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = 𝐸 = 70 × 10 ×
𝑙 300
𝜎 = 54.25 𝑁/𝑚𝑚
𝑢 −𝑢
For element (2), 𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = 𝐸
𝑙
(0 − 0.2325)
= 200 × 10 ×
400
𝑁
 𝜎 = −116.25 (Compressive stress is acting)
𝑚𝑚
Reaction force: we know that,
Reaction force, {𝑅} = [𝐾]{ 𝑢∗ } − {𝐹}
𝑅 5.6 −5.6 0 𝑢 𝐹
 𝑅 = 1 × 10 −5.6 8.6 −3 𝑢 − 𝐹
𝑅 0 −3 3 𝑢 𝐹
𝑅 5.6 −5.6 0 0 0
 𝑅 = 1 × 10 −5.6 8.6 −3 0.2325 − 2 × 10
𝑅 0 −3 3 0 0
0 − 5.6 (0.2325) + 0 0
= 1 × 10 0 + 8.6 (0.2325) + 0 2 × 10
0 − 3 (0.2325) + 0 0
−1.302 0
= 1 × 10 2 − 2 × 10
−0.6975 0
−1.302 × 10 0
= 2 × 10 − 2 × 10
−0.6975 × 10 0
𝑅 −1.302 × 10
𝑅 = 0
𝑅 −0.6975 × 10
2.52 One Dimensional Problems

 𝑅 = −1.302 × 10
𝑅 = 0𝑁
𝑅 = −0.6975 × 10 𝑁
We know that, Reaction force is equivalent and opposite to the applied force.
𝐕𝐞𝐫𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧: 𝑅 + 𝑅 + 𝑅 = −1.302 × 10 − 0 − 0.6975 × 10
= −200 × 10 𝑁(𝐴𝑝𝑝𝑙𝑖𝑒𝑑 𝑓𝑜𝑟𝑐𝑒)
Result:
(i) Nodal displacement at point p, i.e., u2 = 0.2325 mm
(ii) Stress in each material,
1 = 54.25 N/mm2 (tensile)
2 = -116.25 N/mm2 (Compressive)
(iii) Reaction force, 𝑅 = −1.302 × 10
𝑅 = 0
𝑅 = −0.6975 × 10 𝑁

Example 2.11
A thin steel plate of uniform thickness 25 mm is subjected to a point load of
420 N at mid depth as shown in Fig.(i). The plate is also subjected to self-weight. If
Young's modulus, E = 2 x 105 N/mm2 and unit weight density, ρ = 0.8 x 10- 4
N/mm3, calculate the following: (i) Displacement at each nodal point. (ii) Stresses in
each element.

Fig. (i)
 Finite Element Analysis 2.53

Given:

Fig. (ii)
Thickness, t = 25 mm
For element (1): Area, A1 = 100 × 25 = 2500 mm2
For element (2): Area, A2 = 80 × 25 = 2000 mm2
Point load, p = 420 N
Young’s modulus, E = 2 × 105 N/mm2
Unit weight density, ρ = 0.8 × 10-4 N/mm2
To find: (i) Displacement at each nodal points, 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢 .
(ii) Stress in each element, 1 and 2.
Solution: The Steel plat is subjected to self-weight. So, we have to find the body force
acting at nodal point 1, 2 and 3.
𝜌𝐴𝑙
we know thta, body force vector, {F} = [ From equation no. (2.44)]
2
𝐹 𝜌 𝐴 𝑙 1
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟏): force vector, =
𝐹 2 1
0.8 × 10 × 2500 × 200 1 1
= = 20
2 1 1
𝐹 20
= … (1)
𝐹 20
2.54 One Dimensional Problems

𝐹 𝜌 𝐴 𝑙 1
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐): force vector, =
𝐹 2 1
0.8 × 10 × 2000 × 200 1 1
= = 16
2 1 1
𝐹 16
= … (2)
𝐹 16
Assembling the force vector, i.e., assemble the equation (1) and (2)
𝐹 20 20
 𝐹 = 20 + 16 = 36
𝐹 16 16
A Point load of 420 N is acting at mid depth i.e., at nodal point 2, as shown in
Fig(ii) So, add 420 N in F2 vector.
𝐹 20
 𝐹 = 36 + 420
𝐹 16
𝐹 20
Global force vector 𝐹 = 456 … (3)
𝐹 16
Finite element equation for one dimensional plate element is given by,
𝐹 𝐴𝐸 1 −1 𝑢
= [From equaiton no. (2.36)]
𝐹 𝑙 −1 1 𝑢
For element 1: (Nodes 1, 2):

Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹
2500 × 2 × 10 1 −1 𝑢 𝐹
 =
200 −1 1 𝑢 𝐹
 Finite Element Analysis 2.55

1 2
12.5 −12.5 1 𝑢 𝐹
 2 × 10 = … (4)
−12.5 12.5 2 𝑢 𝐹
For element 2: (Nodes 2, 3): Finite element equation is,
𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹
2000 × 2 × 10 1 −1 𝑢 𝐹
 =
200 −1 1 𝑢 𝐹

2 3
12.5 −12.5 2 𝑢 𝐹
 2 × 10 = … (5)
−12.5 12.5 3 𝑢 𝐹
Assemble the finite elements i.e., assemble the finite element equations (4) and (5).
1 2 3
12.5 −12.5 0 1 𝑢 𝐹
 2 × 10 −12.5 12.5 + 10 −10 2 𝑢 − 𝐹
0 −10 10 3 𝑢 𝐹
12.5 −12.5 0 𝑢 𝐹
 2 × 10 −12.5 22.5 −10 𝑢 − 𝐹 . . (6)
0 −10 10 𝑢 𝐹
Apply the boundary conditions i.e., at node 1, displacement u1 = 0. Substitute u1,
F1, F2 and F3 values in equation (6).
12.5 −12.5 0 0 20
 2 × 10 −12.5 22.5 −10 𝑢 − 456
0 −10 10 𝑢 16
In the above equation, u1 = 0. So, neglect first now and first column of
[K]matrix. The reduced equation is,
22.5 −10 𝑢 456
 2 × 10 =
−10 10 𝑢 16
2.56 One Dimensional Problems

 2 × 10 (22.5 𝑢 − 10 𝑢 ) = 456 … . (7)

2 × 10 (−10 𝑢 + 10 𝑢 ) = 16 … . (8)
Solving 2 × 10 (12.5 𝑢 ) = 472
 𝑢 = 1.888 × 10 𝑚𝑚
Substitute 𝑢 value in equation (7),
 2 × 10 [−10(1.888 × 10 ) + 10 𝑢 ] = 16

 − 10(1.888 × 10 ) + 10 𝑢 = 8 × 10

 − 10 𝑢 = 1.968 × 10

𝑢 = 1.968 × 10 𝑚𝑚

𝑑𝑢
we know that, stress, =𝐸
𝑑𝑥
𝑢 −𝑢 1.888 × 10 −0
For element (1): 𝜎 = 𝐸 × = 2 × 10 ×
𝑙 200
𝜎 = 0.188 𝑁/𝑚𝑚
𝑢 −𝑢
For element (2): 𝜎 = 𝐸 ×
𝑙
1.968 × 10 − 1.888 × 10
= 2 × 10 ×
200
𝜎 = 0.008𝑁/𝑚𝑚
Result: (i) Displacement at each nodal points:
𝑢 =0
𝑢 = 1.888 × 10 𝑚𝑚
𝑢 = 1.968 × 10 𝑚𝑚
(ii) Stresses in each element:
1 = 0.188 N/mm2
2 = 0.008 N/mm2
 Finite Element Analysis 2.57

Example 2.12
The three bar assemblage is shown in fig (i). a force of 2500 N is appied in the x
direction at node 2. The length of each element is 750 mm. Take E = 4 ×105 N/mm2 and
A=1200mm2 for elments 1 and 2.
Take E=2 × 105 N/mm2 and A=1200mm2 for element 3. Nodes 1 and 4 are fixed.

Fig. (i)
Calculate the following: (i) Global stiffness matrix.
(ii) Displacemnets of nodes 2 and 3.
(iii) Reactions at nodes 1 and 4.
Given:

[Note: Number with circle denoted elements & number without circle denotes nodes]
𝑙 = 750 𝑚𝑚; 𝑙 = 750 𝑚𝑚; 𝑙 = 750 𝑚𝑚;
𝐹 = 2500 𝑁; 𝐸 = 4 × 10 𝑁/𝑚𝑚 = 𝐸 ;
𝐴 = 600 𝑚𝑚 = 𝐴 ; 𝐸 = 2 × 10 𝑁/𝑚𝑚 ; 𝐴 = 1200 𝑚𝑚
To find: (i) Global stiffness matrix [ K].
(ii) Displacemnets of nodes 2 and 3 (u2, u3).
(iii) Reactions at nodes 1 and 4 (R1, R4).
Solution: Finite element equation for one dimensional two noded bar element is given
by,.
2.58 One Dimensional Problems

𝐹 𝐴𝐸 1 −1 𝑢
= 𝑢
𝐹 𝑙 −1 1
[ From equation no (2.36)]
For element 1: (Nodes 1, 2): Finite element equation is,
𝐴 𝐸 1 −1 𝑢 𝐹
𝑢 =
𝑙 −1 1 𝐹

600 × 4 × 10 1 −1 𝑢 𝐹
 𝑢 =
750 −1 1 𝐹

1 −1 𝑢 𝐹
 3.2 × 10 𝑢 =
−1 1 𝐹

1 2
3.2 −3.2 1 𝑢 𝐹
 1 × 10 = … (1)
−3.2 3.2 2 𝑢 𝐹
For element 2: (Nodes 2, 3): Finite element equation is,
𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

600 × 4 × 10 1 −1 𝑢 𝐹
 =
7500 −1 1 𝑢 𝐹

1 −1 𝑢 𝐹
 3.2 × 10 =
−1 1 𝑢 𝐹
 Finite Element Analysis 2.59

2 3
3.2 −3.2 2 𝑢 𝐹
 1 × 10 = … (2)
−3.2 3.2 3 𝑢 𝐹
For element 3: (Nodes 3, 4): Finite element equation is,
𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

1200 × 2 × 10 1 −1 𝑢 𝐹
 =
750 −1 1 𝑢 𝐹

1 −1 𝑢 𝐹
 3.2 × 10 =
−1 1 𝑢 𝐹

3 4
3.2 −3.2 3 𝑢 𝐹
 1 × 10 = … (3)
−3.2 3.2 4 𝑢 𝐹
Assemble the finite elements i.e., assemble the finite element equations (1), (2) and (3).
1 2 3 4
3.2 −3.2 0 0 1 𝑢 𝐹
−3.2 3.2 + 3.2 −3.2 0 2 𝑢 𝐹
 1 × 10 𝑢 =
0 −3.2 3.2 + 3.2 −3.2 3 𝐹
0 0 −3.2 3.2 4 𝑢 𝐹

3.2 −3.2 0 0 𝑢 𝐹
−3.2 6.4 −3.2 0 𝑢 𝐹
 1 × 10 𝑢 =
0 −3.2 6.4 −3.2 𝐹
𝑢 … (4)
0 0 −3.2 3.2 𝐹
↓
[𝐾]
[Note: The bar has 4 nodes. Each node has single degree of freedom. So, the global
stiffness matrix [K] size is 4×4. The properties of the stiffness matrix are also satisfied.]
2.60 One Dimensional Problems

Applying boundary conditions:

(i) Displacements at node 1 and node 4 are zero  𝑢 = 𝑢 = 0.
(ii) 2500 N load is acting at node 2  𝐹 = 2500 𝑁.
(iii)Self-weight is neglected  𝐹 = 𝐹 = 𝐹 = 0
Substitute 𝑢 , 𝑢 , 𝐹 , 𝐹 , 𝐹 , 𝐹 𝑣𝑎𝑙𝑢𝑒𝑠 𝑖𝑛 𝑒𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (4).
3.2 −3.2 0 0 0 0
 1 × 10 −3.2 6.4 −3.2 0 𝑢 = 2500
0 −3.2 6.4 −3.2 𝑢 0
0 0 −3.2 3.2 0 0
In the above equation, u1 = 0. So, neglect first now and first column of [K]matrix. u4 =
0, so, neglected fourth row and fourth column of [K] matrix. The final reduced equation
is,
6.4 −3.2 𝑢 2500
1 × 10 =
−3.2 6.4 𝑢 0

1 × 10 (6.4 𝑢 − 3.2 𝑢 ) = 2500 … . (5)

1 × 10 (−3.2 𝑢 + 6.4 𝑢 ) = 0 … . (6)

Multiple the equation (6) by 2,
1 × 10 (6.4 𝑢 − 3.2 𝑢 ) = 2500

1 × 10 (−6.4 𝑢 + 12.8 𝑢 ) = 0 … . (7)

Solving 1 × 10 (9.6 𝑢 ) = 2500

 𝑢 = 2.604 × 10 𝑚𝑚
Substitute 𝑢 value in equation (5),
 1 × 10 [6.5 𝑢 3.2(2.604 × 10 )] = 2500

 6.5 𝑢 − 8.33 × 10 = 0.025

 𝑢 = 5.127 × 10 𝑚𝑚
we know that, sReaction force, {𝑅} = [𝐾 ]{𝑢∗ } − {𝐹}
 Finite Element Analysis 2.61
𝑅 3.2 −3.2 0 0 𝑢 𝐹
𝑅 3.2 6.4 −3.2 0 𝑢 𝐹
 = 1 × 10 𝑢 −
𝑅 0 −3.2 6.4 −3.2 𝐹
𝑅 0 0 0 3.2 𝑢 𝐹
𝑅 3.2 −3.2 0 0 0 0
𝑅 3.2 6.4 −3.2 0 5.127 × 10 2500
 = 1 × 10 −
𝑅 0 −3.2 6.4 −3.2 2.604 × 10 0
𝑅 0 0 −3.2 3.2 0 0
0 − 3.2 × 5.127 × 10 + 0 + 0
= 1 × 10 0 + 6.4 × 5.127 × 10 − 3.2 × 2.604 × 10 +0
0 + 3.2 × 5.127 × 10 − 6.4 × 2.604 × 10 +0
0 + 0 × 3.2 × 2.604 × 10 + 0
0
− 2500
0
0
−0.0164 0
0.02448 2500
= 1 × 10 −
2.50 × 10 0
−8.33 × 10 0
−0.0164 × 10 0 −1640 0
= 0.02448 × 10 − 2500 = 2448 − 2500
2.50 × 10 × 10 0 25 0
−8.33 × 10 × 10 0 −833 0
𝑅 −1640
𝑅
= −52
𝑅 25
𝑅 −833
 𝑅 = −1640 𝑁; 𝑅 = −52𝑁; 𝑅 = 25 𝑁; 𝑅 = −833 𝑁
We know that, reaction force is equivalent and opposite to the applied force.
Verification: 𝑅 + 𝑅 + 𝑅 + 𝑅 = −1640 − 52 + 25 − 833
= -2500 N (Applies force)
3.2 −3.2 0 0
Result: (i) [𝐾] = 1 × 10 3.2 6.4 −3.2 0
0 −3.2 6.4 −3.2
0 0 −3.2 3.2
2.62 One Dimensional Problems

(ii) 𝑢 = 5.127 × 10 𝑚𝑚
𝑢 = 2.604 × 10 𝑚𝑚
(iii) R1 = - 1640 N
R4 = -833 N
[Note: By strictly solving, we will get R2 =0 and R3 = 0. But there is some error in the
solution due to numeriacal approximation. If we take more number of elements, this
error approaches to zero. So, we will get R1 + R4 = -2500 N (Applied force).

Example 2.13
Consider the bar as shown in fig (i).
Calculate the following: (i) nodal displacement (ii) element stresses, (iii) support
reactions

Fig. (i)
Take E = 2 × 105 N/mm2; p = 400 kN.
Given:

Fig. (ii)
[Note: number with circles denotes elements & Number without circle denotes nodes]
Area of element (1), A1 = 300 mm2
Area of element (2), A2 = 300 mm2
Area of element (3), A3 = 600 mm
 Finite Element Analysis 2.63

Length of element (1), l1 = 200 mm

Length of element (2), l2 = 200 mm
Length of element (3), l3 = 400 mm
Young’s modulus E = 2 × 105 N/mm2
Point load at node 2, p = 400 kN = 400 × 103 N
To find:
(i) Nodal displacement 𝑢 , 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢 .
(ii) Stress in each material, 1, 2 and 3.
(iii) Reaction at the support, R1 and R4.
Soltuion: Finite element equaiton for one dimensional two noded bar element is given
by,
𝐹 𝐴𝐸 1 −1 𝑢
= 𝑢 [From equaiton no. 2.36]
𝐹 𝑙 −1 1

For element 1: (Nodes 1, 2): Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
𝑢 =
𝑙 −1 1 𝐹

300 × 2 × 10 1 −1 𝑢 𝐹
𝑢 =
300 −1 1 𝐹

1 −1 𝑢 𝐹
3 × 10 𝑢 =
−1 1 𝐹

1 2
3 −3 1 𝑢 𝐹
 1 × 10 =
−3 3 2 𝑢 𝐹
2.64 One Dimensional Problems

For element 2: (Nodes 2, 3): Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
𝑢 =
𝑙 −1 1 𝐹
300 × 2 × 10 1 −1 𝑢 𝐹
𝑢 =
400 −1 1 𝐹

1 −1 𝑢 𝐹
3 × 10 𝑢 =
−1 1 𝐹

2 3
3 −3 2 𝑢 𝐹
 1 × 10 = … (2)
−3 3 3 𝑢 𝐹

For element 3: (Nodes 3, 4): Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
𝑢 =
𝑙 −1 1 𝐹
1 −1 𝑢 𝐹
 3 × 10 𝑢 =
−1 1 𝐹

3 4
3 −3 3 𝑢 𝐹
 1 × 10 = … (3)
−3 3 4 𝑢 𝐹
Assemble the finite elements. i.e., assemble the finite element equations (1) and (2).
1 2 3 4
3 −3 0 0 1 𝑢 𝐹
−3 3 + 3 −3 0 2 𝑢 𝐹
 1 × 10 𝑢 =
0 −3 3 + 3 −3 3 𝐹
0 0 −3 3 4 𝑢 𝐹
3 −3 0 0 𝑢 𝐹
−3 6 −3 0 𝑢 𝐹
 1 × 10 𝑢 = … (4)
0 −3 6 −3 𝐹
0 0 −3 3 𝑢 𝐹
Applying boundary conditions:
(𝑖) Node 1 and node 4 are fixed. So, 𝑢 = 𝑢 = 0.
(𝑖𝑖) 400 × 10 𝑁 𝑖𝑠 𝑎𝑐𝑡𝑖𝑛𝑔 𝑎𝑡 𝑛𝑜𝑑𝑒 2. 𝑠𝑜, 𝐹 = 400 × 10 𝑁.
 Finite Element Analysis 2.65
(𝑖𝑖) 𝑆𝑒𝑙𝑓 − 𝑤𝑒𝑖𝑔ℎ𝑡 𝑖𝑠 𝑛𝑒𝑔𝑙𝑒𝑐𝑡𝑒𝑑. 𝑆𝑜, 𝐹 = 𝐹 = 𝐹 = 0.
Substitute 𝑢 , 𝑢 𝑎𝑛𝑑 𝐹 , 𝐹 , 𝐹 𝑎𝑛𝑑 𝐹 𝑣𝑎𝑙𝑢𝑒𝑠 𝑖𝑛 𝑒𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (4).
3 −3 0 0 0 0
(4)  1 × 10 −3 6 −3 0 𝑢 = 400 × 10
0 −3 6 −3 𝑢 0
0 0 −3 3 0 0
In the above equation, u1 = 0. So, neglect first row and first column of [K] matrix. u4 =
0, so, delete fourth row and fourth column of [K] matrix. Hence the equation reduced to,
6 −3 𝑢
1 × 10 𝑢 = 400 × 10
−3 6 0
 1 × 10 (6 𝑢 − 3 𝑢 ) = 400 × 10 … (5)
1 × 10 (−3 𝑢 + 6 𝑢 ) = 0 … (6)
Multiply the equation (6) by 2.
 1 × 10 (6 𝑢 − 3 𝑢 ) = 400 × 10 … (5)
1 × 10 (−6 𝑢 + 12 𝑢 ) = 0 … (7)
Solving, 1 × 10 (6 𝑢 ) = 400 × 10
 𝑢 = 0.4444 𝑚𝑚
Substitute u3 value in equation (5),
 1 × 10 [6 𝑢 − 3 (0.4444) = 400 × 10 ]
 𝑢 = 0.8888 𝑚𝑚
𝑑𝑢
We know that, Stress, 𝜎=𝐸
𝑑𝑥
𝑢 −𝑢
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟏): 𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 =𝐸
𝑙
(0.8888 − 0)
= 2 × 10 ×
200
𝜎 = 888.88 𝑁/𝑚𝑚 [Tensile stress]
(𝑢 − 𝑢 )
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐), : Stress, 𝜎 =𝐸
𝑙
2.66 One Dimensional Problems
(0.4444 − 0.8888)
= 2 × 10 ×
200
𝜎 = −444.44 𝑁/𝑚𝑚 [Compressive stress ]
Reaction force: we know that,
(𝑢 − 𝑢 ) (0 − 0.4444)
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟑), : 𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = 𝐸 = 2 × 10 ×
𝑙 400
𝜎 = −222.222 𝑁/𝑚𝑚 [Compressive stress ]
We know that,
Reaction force, {𝑅} = [𝐾]{ 𝑢∗ } − {𝐹}
𝑅 3 −3 0 0 𝑢 𝐹
𝑅 −3 6 −3 0 𝑢 𝐹
 = 1 × 10 𝑢 −
𝑅 0 −3 6 −3 𝐹
𝑅 0 0 −3 3 𝑢 𝐹
3 −3 0 0 0 0
= 1 × 10 −3 6 −3 0 0.8888 − 400 × 10
0 −3 6 −3 0.4444 0
0 0 −3 3 0 0
0 − 3 × 0.8888 + 0 + 0 0
= 1 × 10 0 + 6 × 0.8888 − 3 × 0.4444 + 0 − 400 × 10
0 − 3 × 0.8888 + 6 × 0.4444 + 0 0
0 + 0 − 3 × 0.4444 + 0 0
−2.6667 × 10 0
= 4 × 10 − 400 × 10
0 0
−1.3333 × 10 0
𝑅 −2.6667 × 10
𝑅 0
=
𝑅 0
𝑅 −1.3333 × 10
 𝑅 = −2.6667 × 10 𝑁
𝑅 = 0
𝑅 = 0
𝑅 = −1.3333 × 10 𝑁
 Finite Element Analysis 2.67

We know that, Reaction force is equivalent and opposite to the applied force.
𝑽𝒆𝒓𝒊𝒇𝒊𝒄𝒂𝒕𝒊𝒐𝒏: 𝑅 + 𝑅 + 𝑅 + 𝑅 = −2.6667 × 10 + 0 + 0 − 1.3333 × 10
= −4 × 10 𝑁(𝐴𝑝𝑝𝑙𝑖𝑒𝑑 𝑓𝑜𝑟𝑐𝑒)
Result:
(i) Nodal displacement:
𝑢 =0
𝑢 = 0.8888 𝑚𝑚
𝑢 = 0.4444 𝑚𝑚
𝑢 =0
(ii) Element Stresses:
1 = 888.88 N/mm2 (Tensile)
2 = -444.44 N/mm2 (Compressive)
3 = -222.22 N/mm2 (Compressive)
(iii) Reaction force, 𝑅 = −2.6667 × 10 𝑁
𝑅 = −1.3333 × 10 𝑁

Example 2.14
Consider a taper steel plate of uniform thickness, t = 25 mm as shown in Fig.
The Young’s modulus of the plate, E = 2 × 105 N/mm2 and weight density, ρ = 0.82 ×
10-4 N/mm3. In addition to its self-weight, the plate is subjected to a point load p =
100 N at its midpoint. Calculate the following by modeling the plate with two finite
elements:
(i) Global force vector {F}
(ii) Global stiffness matrix [K]
(iii) Displacements in each element
(iv) Stresses in each element
(v) Reaction force at the support
2.68 One Dimensional Problems

Fig. (i)
Given: In this problem, the area of the element is varying at each cross- section. If we
consider this area variation, the problem will be tedious. So, the given taper bar is
considered as stepped bas as shown in Fig.(ii).

Fig. (ii)
Solution: Area at node 1, A = Width × Thickness = W × t = 150 × 25
𝐴 = 3750 𝑚𝑚
Area at node 2, 𝐴 = 𝑊 × 𝑡
𝑊 +𝑊 150 + 75
×𝑡 = × 25
2 2
[∵ 𝑡 = 𝑡 = 𝑡 = 25 𝑚𝑚]
 Finite Element Analysis 2.69

𝐴 = 2812.5 𝑚𝑚
Area at node 3, 𝐴 = 𝑊 × 𝑡 = 75 × 25
𝐴 = 1875 𝑚𝑚
Average area of element (1):
𝐴𝑟𝑒𝑎 𝑎𝑡 𝑛𝑜𝑑𝑒, 1 + 𝐴𝑟𝑒𝑎 𝑎𝑡 𝑛𝑜𝑑𝑒, 2
𝐴 =
2
3750 + 2812.5
=
2
𝐴 = 3281.25 𝑚𝑚
Average area of element (2):
𝐴𝑟𝑒𝑎 𝑎𝑡 𝑛𝑜𝑑𝑒, 2 + 𝐴𝑟𝑒𝑎 𝑎𝑡 𝑛𝑜𝑑𝑒, 3
𝐴 =
2
2812.5 + 1875
=
2
𝐴 = 2343.75 𝑚𝑚
Young’s modulus, E = 2 × 105 N/mm2
Weight density,  = 0.82 × 10-4 N/mm3; Length, l = 300 mm
Point load, p = 100 N
To find: (i) Global force vector {F}.
(ii) Global stiffness matrix, [K].
(iii) Displacement in each element.
(iv) Stresses in each element.
(v) Reaction force at the support.
Solution: the steel plate is subjected to self-weight. So, we have to find the body force
acting at nodal points 1, 2 and 3.
We know that,
𝜌𝐴𝑙 1
Body force vector, {𝐹} = [From equation no. (2.44)]
2 1
2.70 One Dimensional Problems

𝐹 𝜌 𝐴̅ 𝑙 1
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟏), Force vector, =
𝐹 2 1

0.82 × 10 × 3281.25 × 300 1

= ×
2 1
1
= 40.359 ×
1
𝐹 40.359
= … (1)
𝐹 40.359

𝐹 𝜌 𝐴̅ 𝑙 1
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐), Force vector, =
𝐹 2 1

0.82 × 10 × 2343.75 × 300 1

= ×
2 1
1
= 28.828 ×
1
𝐹 28.828
= … (2)
𝐹 28.8289
Assembling the force vector, i.e., assemble the equation (1) and (2),
𝐹 40.359 40.359
𝐹 = 40.359 + 28.828 − 69.187
𝐹 28.828 28.828
A point load of 100 N is acting at node 2 as shown in Fig. So, add 100 N in F2 vector.
𝐹 40.359
𝐹 = 69.187 + 100
𝐹 28.828
𝐹 40.359
Golbal force vector, 𝐹 = 169.187 … (3)
𝐹 28.828
Finite element equation for one dimensional plate element is given by,
𝐹 𝐴𝐸 1 −1 𝑢
= 𝑢
𝐹 𝑙 −1 1
 Finite Element Analysis 2.71

For element 1: (Nodes 1, 2): Finite element equation is,

𝐴̅ 𝐸 1 −1 𝑢 𝐹
𝑢 =
𝑙 −1 1 𝐹
3281.25 × 2 × 10 1 −1 𝑢 𝐹
𝑢 =
300 −1 1 𝐹
1 −1 𝑢 𝐹
10.937 × 2 × 10 𝑢 =
−1 1 𝐹
1 2
10.937 −10.937 1 𝑢 𝐹
2 × 10 = … (4)
−10.937 10.937 2 𝑢 𝐹
For element 2: (Nodes 2, 3): Finite element equation is,
𝐴̅ 𝐸 1 −1 𝑢 𝐹
𝑢 =
𝑙 −1 1 𝐹
2343.75 × 2 × 10 1 −1 𝑢 𝐹
𝑢 =
300 −1 1 𝐹
1 −1 𝑢 𝐹
7.8125 × 2 × 10 𝑢 =
−1 1 𝐹
1 2
7.8125 −7.8125 2 𝑢 𝐹
 2 × 10 = … (5)
−7.8125 7.8125 3 𝑢 𝐹
Assemble the finite element equations (4) and (5).
1 2 3
10.937 −10.937 0 1 𝑢 𝐹
 2 × 10 −10.937 10.937 + 7.8125 −7.8125 2 𝑢 = 𝐹
0 −7.8125 7.8125 3 𝑢 𝐹
10.937 10.937 0 𝑢 𝐹
 2 × 10 −10.937 18.749 −7.8125 𝑢 = 𝐹
0 −7.8125 7.8125 𝑢 𝐹 … (6)
↓
[𝐾]
Applying boundary conditions, i.e., at node 1, displacement, 𝑢 = 0. Substitute
𝐹 , 𝐹 𝑎𝑛𝑑 𝐹 in equation (6).
2.72 One Dimensional Problems

10.937 10.937 0 0 40.359

 2 × 10 −10.937 18.749 −7.8125 𝑢 = 169.187
0 −7.8125 7.8125 𝑢 28.828
In the above equation, u1 = 0. So, neglect first row and first column of [K] matrix. The
reduced equation is,
18.749 −7.8125 𝑢 169.187
2 × 10 𝑢 =
−7.815 7.8125 28.828
 2 × 10 (18.749 𝑢 − 7.8125 𝑢 ) = 169.187 … (7)
2 × 10 (−7.8125 𝑢 + 7.8125 𝑢 ) = 28.828 … (8)
Solving, 2 × 10 (10.936 )𝑢 = 198.015
 𝑢 = 9.053 × 10 𝑚𝑚
Substitute u2 value in equation (7),
 2 × 10 [18.749 (9.053 × 10 ) − 7.8125 𝑢 ] = 169.187
 18.749 × 9.053 × 10 − 7.8125 𝑢 = 8.459 × 10
−7.8125 𝑢 = −8.514 × 10
 𝑢 = 10.898 × 10 𝑚𝑚
𝑑𝑢
We know that, Stress, 𝜎=𝐸
𝑑𝑥
𝑢 −𝑢
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟏): 𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = 𝐸
𝑙
(9.053 × 10 − 0)
= 2 × 10 ×
300
𝜎 = 0.060 𝑁/𝑚𝑚
(𝑢 − 𝑢 )
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐), : Stress, 𝜎 = 𝐸 ×
𝑙
2 × 10 × (10.8982 × 10 − 9.053 × 10 )
=
300
𝜎 = 0.0123 𝑁/𝑚𝑚
 Finite Element Analysis 2.73

Reaction force: we know that,

Reaction force, {𝑅} = [𝐾]{ 𝑢∗ } − {𝐹}
𝑅 10.937 −10.937 0 𝑢 𝐹
 𝑅 = 2 × 10 −10.937 18.749 −7.8125 𝑢 − 𝐹
𝑅 0 −7.8125 7.8125 𝑢 𝐹
10.937 −10.937 0 0
= 2 × 10 −10.937 18.749 −7.8125 9.053 × 10
0 −7.8125 7.8125 10.898 × 10
40.359
− 169.187
28.828
0 − 10.937 × 9.053 × 10 + 0
= 2 × 10 0 + 18.749 × 9.053 × 10 − 7.8125 × 10.898 × 10
0 − 7.8125 × 9.053 × 10 + 7.8125 × 10.898 × 10
40.359
− 169.187
28.828
−9.901 × 10 40.359
= 2 × 10 8.459 × 10 − 169.187
1.441 × 10 28.828
−198.02 40.359
= 169.187 − 169.187
28.828 28.828
𝑅 −238.379
𝑅 = 0
𝑅 0
 𝑅 = −2.38.379 𝑁
𝑅 = 0𝑁
𝑅 = 0𝑁
𝑅 = −1.3333 × 10 𝑁
We know that, Reaction force is equivalent and opposite to the applied force.
𝐕𝐞𝐫𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧: R + R + R = −238.379 + 0 + 0 = −238.379 N
𝐀𝐩𝐩𝐥𝐢𝐞𝐝 𝐟𝐨𝐫𝐜𝐞, F + F + F = 40.359 + 169.187 + 28.828 = 238.37 N
2.74 One Dimensional Problems

Result:
40.359
(𝑖){𝐹} = 169.187
28.828
10.397 −10.397 0
(𝑖𝑖) [𝐾] = 2 × 10 −10.397 18.749 −7.8125
0 −7.8125 7.8125
(i) 𝑢 = 0
𝑢 = 9.053 × 10 𝑚𝑚
𝑢 = 10.898 × 10 𝑚𝑚
(ii) 1 = 0.060 N/mm2
2 = 0.0123 N/mm2
(iii) Reaction force, 𝑅 = −238.379𝑁
𝑅 = 0𝑁
𝑅 = 0𝑁
Example 2.15:
For a tapered plate of uniform thickness t = 10 mm as shown in Fig.(i),
find the displacements at the nodes by forming into two element model. The
bar has mass density, ρ = 7800 kg/m3, Young's modulus, E = 2 105 MN/m2.
In addition to self- weight, the plate is subjected to a point load p = 10 kN at its
Centre. Also determine the reaction force at the support.
[Anna University, ME-Aeronautical Engg-Dec.2006}

Fig. (i)
 Finite Element Analysis 2.75

Given: In this problem, the area of the element is varying at each cross-section. If we
consider this area variation, the problem, will be tedious. So, the given taper bar is
considered as stepped bar as shown in Fig. (ii).

Fig. (ii)
Area at node 1, A = Width × Thickness = W × t
= 80 × 10
𝐴 = 800 𝑚𝑚
W +W
Area at node 1, A = W × t ×t
2
80 + 40
= × 10
2
𝐴 = 600 𝑚𝑚 [∵ 𝑡 = 𝑡 = 𝑡 = 10 𝑚𝑚]
Area at node 3, A = W × t = 40 × 10
𝐴 = 400 𝑚𝑚
Area at nose, 1 + Area at node, 2
Average area of element (1), A =
2
800 + 600
𝐴̅ =
2
𝐴̅ = 700 𝑚𝑚
2.76 One Dimensional Problems
Area at nose, 2 + Area at node, 3
Average area of element (2), A =
2
600 + 400
𝐴̅ =
2
𝐴̅ = 500 𝑚𝑚
Mass density, ρ = 7800 kg⁄m
Weight density, ρ = 7800 × 9.81 N⁄m
= 76518 𝑁⁄𝑚
= 76518 × 10 𝑁⁄𝑚𝑚
= 7.6518 × 10 𝑁⁄𝑚𝑚
Young, s modulus, E = 2 × 10 MN⁄m
= 2 × 10 × 10 𝑁⁄𝑚
= 2 × 10 × 10 × 10 𝑁⁄𝑚𝑚
𝐸 = 2 × 10 𝑁⁄𝑚𝑚
Point load, p = 10 kN = 10 × 10 N
To find: (i) Displacement at each nodes.
(ii) Reaction force at the support.
Solution: The plate is subjected to self-weight. So, we have to find the body force
acting at nodal points 1, 2, 3.
We know that,
𝜌𝐴𝑙 1
Body force vector, {F} = [From equation no. (2.44)]
2 1
𝐹 𝜌 𝐴̅ 𝑙 1
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟏): Force vector =
𝐹 2 1
7.6518 × 10 × 700 × 150 1
=
2 1
1
= 4.017
1
 Finite Element Analysis 2.77
𝐹 4.017
= … (1)
𝐹 4.017
𝐹 𝜌 𝐴̅ 𝑙 1
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐): Force vector =
𝐹 2 1
7.6518 × 10 × 500 × 150 1
=
2 1
1
= 2.869
1
𝐹 2.869
= … (2)
𝐹 2.869
Assembling the force vector, i.e., assemble the equation (1) and (2),
𝐹 4.017 4.017
 𝐹 = 4.017 + 2.869 = 6.886
𝐹 2.869 2.869
A point load of 10 × 103 N is acting at node 2 as shown in Fig. So, add 10,000 N in F2
vector.
𝐹 4.017 4.017
 𝐹 = 6.886 + 10,000 = 10006.886
𝐹 2.869 2.869
𝐹 4.017
Global force vector, 𝐹 = 10006.886 . . (3)
𝐹 2.869
Finite element equation for one dimensional plate element is given by,
𝐹 𝐴𝐸 1 −1 𝑢
=
𝐹 𝑙 −1 1 𝑢
For element 1: (Nodes 1,2 ): Finite element equation is,
𝐴̅ 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹
700 × 2 × 10 1 −1 𝑢 𝐹
=
150 −1 1 𝑢 𝐹
1 −1 𝑢 𝐹
4.666 × 2 × 10 =
−1 1 𝑢 𝐹
2.78 One Dimensional Problems

1 2
1 −1 1 𝑢 𝐹
2 × 10 = … (4)
−1 1 2 𝑢 𝐹
For element 2: (Nodes 2, 3 ): Finite element equation is,
𝐴̅ 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹
500 × 2 × 10 1 −1 𝑢 𝐹
 =
150 −1 1 𝑢 𝐹
1 −1 𝑢 𝐹
 3.333 × 2 × 10 =
−1 1 𝑢 3
1 2
1 −1 1 𝑢 𝐹
 2 × 10 = … (5)
−1 1 2 𝑢 𝐹
Assemble the finite element equations (4) and (5).
1 2 3
4.666 −4.666 0 1 𝑢 𝐹
 2 × 10 −4.666 4.666 + 3.333 −3.333 2 𝑢 = 𝐹
0 −3.333 3.333 3 𝑢 𝐹
4.666 −4.666 0 𝑢 𝐹
 2 × 10 −4.666 7.999 −3.333 𝑢 = 𝐹
0 −3.333 3.333 𝑢 𝐹 … (6)
↓
[𝐾]
Apply the boundary conditions, i.e., at nose 1, displacement u1= 0. Substituting
u1, F1, F2 and F3 values in equation (6),
4.666 −4.666 0 0 4.017
 2 × 10 −4.666 7.999 −3.333 𝑢 = 10006.886
0 −3.333 3.333 𝑢 2.869
In the above equation u1= 0. So, neglect first row and first column of [K]
matrix. The reduced equation is,
7.999 −3.333 𝑢 10,006.886
2 × 10 =
−3.333 3.333 𝑢 2.869
 2 × 10 (7.999 𝑢 − 3.333 𝑢 ) = 10,006.886 … (7)
 Finite Element Analysis 2.79

2 × 10 (−3.333 𝑢 + 3.333 𝑢 ) = 2.869 … (8)

Solving, 2 × 10 (4.666 𝑢 ) = 10009.755
 𝑢 = 0.01073 𝑚𝑚
Substitute u2 value in equation (8),
 2 × 10 [−3.333(0.01073) + 3.333 𝑢 ] = 2.869
 − 3.333(0.01073) + 3.333 𝑢 = 1.4345 × 10
 𝑢 = 0.01073 𝑚𝑚
We know that, Reaction force, {R} = [K]{ u∗ } − {F}
𝑅 4.666 −4.666 0 𝑢 𝐹
 𝑅 = 2 × 10 −4.666 7.999 −3.333 𝑢 − 𝐹
𝑅 0 −3.333 3.333 𝑢 𝐹
4.666 −4.666 0 0 4.017
= 2 × 10 −4.666 7.999 −3.333 0.01073 − 10,006.886
0 −3.333 3.333 0.01073 2.869
0 + 4.666 × 0.01073 + 0 4.017
= 2 × 10 0 + 7.999 × 0.01073 − 3.333 × 0.01073 − 10,006.886
0 − 3.333 × 0.01073 + 3.333 × 0.01073 2.869
−0.050 4.017
= 2 × 10 0.050 − 10,006.886
0 2.869
−10,000 4.017
= 10,000 − 10,006.886
0 2.869
𝑅 −10004.017
𝑅 = −6.886
𝑅 −2.869
 𝑅 = −10004.07 𝑁
𝑅 = −6.886 𝑁
𝑅 = −2.869 𝑁
We know that, Reaction force is equivalent and opposite to the applied force.
𝑽𝒆𝒓𝒊𝒇𝒊𝒄𝒂𝒕𝒊𝒐𝒏: 𝑅 + 𝑅 + 𝑅 = −10004.017 + 6.886 − 2.869
2.80 One Dimensional Problems

= −10013.772 𝑁
𝑨𝒑𝒑𝒍𝒊𝒆𝒅 𝒇𝒐𝒓𝒄𝒆, 𝐹 + 𝐹 + 𝐹 = 4.017 + 10,006.886 + 2.869
= 10013.772 𝑁
Result:
(i) Displacement at each node:
𝑢 =0
𝑢 = 0.01073 𝑚𝑚
𝑢 = 0.01073 𝑚𝑚
(ii) Reaction force at the support:
𝑅 = −10004.017 𝑁

Example 2.16
Using two finite elements, find the stress distribution in a uniformly tapering bar
of circular cross-sectional area 3 cm2 and 2cm2 at their ends, length 100mm, subjected
to an axial tensile load of 50 N at smaller end and fixed at larger end Take the value of
young’s modulus 2*105 N/mm2. [Anna university ME.- jan,2006]
Given:
In this problem, the area of the element is varying at each cross-section. If we
consider this area variation, the problem will be tedious. So, the given taper bar is
considered as stepped bar as shown in Fig.(i).

Fig. (i)
 Finite Element Analysis 2.81

Area at node 1, A = 3 cm
Area at node 3, A = 2 cm
A +A 3+2
Area at node 2, A = = = 2.5 cm
2 2
Area at nose, 1 + Area at node, 2
Average area of element (1), A =
2
3 + 2.5
𝐴̅ =
2
𝐴̅ = 2.75 𝑐𝑚 = 2.75 × 10 𝑚𝑚
𝐴̅ = 275 𝑚𝑚
Average area of element (2),
Area at nose, 2 + Area at node, 3
A =
2
2.5 + 2
= = 2.25 cm
2
A = 225 mm
Tensile load at node, 3 = 50 N
Young’s modulus, E = 5 × 105 N/mm2
To find: Stress distribution: (i) Stress in element (1), 1.
(ii) Stress in element (2), 2.
Solution: Finite element equation for one dimensional two noded bar element is given
by,
𝐹 𝐴𝐸 1 −1 𝑢
=
𝐹 𝑙 −1 1 𝑢
For element 1: (Nodes 1, 2 ): Finite element equation is,
𝐴̅ 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹
275 × 2 × 10 1 −1 𝑢 𝐹
=
50 −1 1 𝑢 𝐹
2.82 One Dimensional Problems

1 2
11 −11 1 𝑢 𝐹
 2 × 10 = … (1)
−11 11 2 𝑢 𝐹
For element 2: (Nodes 2, 3 ): Finite element equation is,
𝐴̅ 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹
225 × 2 × 10 1 −1 𝑢 𝐹
=
50 −1 1 𝑢 𝐹
2 3
9 −9 2 𝑢 𝐹
1 × 10 = … (2)
−9 9 3 𝑢 𝐹
Assemble the finite element equations (1) and (2).
1 2 3
11 −11 0 1 𝑢 𝐹
 1 × 10 −11 11 + 9 −9 2 𝑢 = 𝐹
0 −9 9 3 𝑢 𝐹
11 −11 0 𝑢 𝐹
 1 × 10 −11 11 + 9 −9 𝑢 = 𝐹
0 −9 9 𝑢 𝐹 … (3)
↓
[𝐾]
Apply boundary conditions:
(i) At nose 1, displacement u1= 0.
(ii) Self-weight is neglected and 50 N is acting at node 3. So, 𝐹 = 0 , 𝐹 = 0, 𝐹 =
50 𝑁. Substituting u1, F1, F2 and F3 values in equation (3),
11 −11 0 0 0
 1 × 10 −11 20 −9 𝑢 = 0
0 −9 9 𝑢 50
In the above equation u1= 0. So, neglect first row and first column of [K] matrix.
Hence, the equation reduces to,
20 −9 𝑢 0
1 × 10 =
−9 9 𝑢 50
 Finite Element Analysis 2.83

 1 × 10 (20 𝑢 − 9 𝑢 ) = 0 … (4)
1 × 10 (−9 𝑢 + 9 𝑢 ) = 50 … (5)
Solving, 1 × 10 (11 𝑢 ) = 50
 𝑢 = 4.545 × 10 𝑚𝑚
Substitute u2 value in equation (4),
 1 × 10 [20 (4.545 × 10 ) − 9 𝑢 ] = 0
 20(4.545 × 10 ) = 9 𝑢
 𝑢 = 1.01 × 10 𝑚𝑚
𝑑𝑢
We know that, 𝑆𝑡𝑟𝑒𝑠𝑠 𝜎 =𝐸
𝑑𝑥
𝑢 −𝑢
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟏): 𝑆𝑡𝑟𝑒𝑠𝑠 𝜎 = 𝐸
𝑙
𝑢 −𝑢
=𝐸
𝑙
2 × 10 (4.545 × 10 − 0)
=
50
𝜎 = 0.1818 𝑁/𝑚𝑚
𝑢 −𝑢
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐): Stress 𝜎 = 𝐸 ×
𝑙
2 × 10 (1.01 × 10 − 4.545 × 10 )
=
50
𝜎 = 0.222 𝑁/𝑚𝑚
𝑽𝒆𝒓𝒊𝒇𝒊𝒄𝒂𝒕𝒊𝒐𝒏: 𝑤𝑒 𝑘𝑛𝑜𝑤 𝑡ℎ𝑎𝑡,
Load
Stress, σ =
Area
P 50
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟏): σ = =
A 275
𝜎 = 0.1818 𝑁/𝑚𝑚
2.84 One Dimensional Problems
P 50
𝐅𝐨𝐫 𝐞𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐): σ = =
A 225
𝜎 = 0.222 𝑁/𝑚𝑚

Example 2.17
A rod subjected to an axial load P = 600 KN is applied as shown in Fig.
Divide the domain into two elements. Determine the following:
(a) Displacement at each node.
(b) Stresses in each element.
(c) Reactions at each node point.
Take A = 250 mm2, E = 2 x l0^5 N/mm^2 [AU. Dec. 2005]

Fig. (i)

Fig. (ii)
In this problem, we should first determine whether contact occurs between the
bar and the wall. To do this, assume that the wall does not exist. The deformation at
node is given by,
𝑃𝐿 600 × 10 × 150
𝛿𝐿 = =
𝐴𝐸 250 × 2 × 10
𝛿𝐿 = 1.8 𝑚𝑚
The gap between the wall and node 3 is 1.2 mm. So, the contact occurs between
the bar and the wall.
 Displacement, u = 1.2 𝑚𝑚
 Finite Element Analysis 2.85

Axial load, P = 600 kN = 600 × 10 N

Area, A = 250 mm
Young s modulus , E = 2 × 10 mm
To find: (a) Displacement at each node, u1, u2 and u3.
(b) Stresses in each element, 1 and 2.
(c) reactions at each node point, R1, R2 and R3.
Solution:

Finite element equation for one dimensional two noded bar element is given by,
𝐹 𝐴𝐸 1 −1 𝑢
=
𝐹 𝑙 −1 1 𝑢
For element 1: (Nodes 1, 2):

Finite element equation is,

𝐴𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

250 × 2 × 10 1 −1 𝑢 𝐹
=
150 −1 1 𝑢 𝐹

1 −1 𝑢 𝐹
3.333 × 10 =
−1 1 𝑢 𝐹
2.86 One Dimensional Problems

1 2
3.333 −3.333 1 𝑢 𝐹
1 × 10 = … (1)
−3.333 3.333 2 𝑢 𝐹
For element 2: (Nodes 2, 3 ): Finite element equation is,
𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

250 × 2 × 10 1 −1 𝑢 𝐹
 =
150 −1 1 𝑢 𝐹
1 2
3.333 −3.333 2 𝑢 𝐹
1 × 10 = … (2)
−3.333 3.333 3 𝑢 𝐹
Assemble the finite element equations (1) and (2).
1 2 3
3.333 −3.333 0 1 𝑢 𝐹
 1 × 10 −3.333 3.333 + 3.333 −3.333 2 𝑢 = 𝐹
0 −3.333 3.333 3 𝑢 𝐹
3.333 −3.333 0 𝑢 𝐹
 1 × 10 −3.333 6.666 −3.333 𝑢 = 𝐹
0 −3.333 3.333 𝑢 𝐹 … (3)
↓
[𝐾]

Apply boundary conditions:

(i) At nose 1, displacement u1= 0.
(ii) At node 3, u3 = 1.2 mm (given).
(iii)Point load, = 600 × 103 N is acting at node 2 and Self-weight is neglected.
So, 𝐹 = 0 , 𝐹 = 600 × 10 𝑁 , 𝐹 = 0.
 Finite Element Analysis 2.87

Apply u1, u3, F1, F2 and F3 values in equation (3),

3.333 −3.333 0 0 0
 1 × 10 −3.333 6.666 −3.333 𝑢 = 6 × 10
0 −3.333 3.333 1.2 0
In the above equation u1= 0. So, neglect first row and first column of [K] matrix.
Hence, the equation reduces to,
6.666 −3.333 𝑢
1 × 10 = 6 × 10
−3.333 3.333 1.2 0
 1 × 10 [6.666 𝑢 − 3.333(1.2)] = 6 × 10
 6.666 𝑢 − 3.9996 = 6
 𝑢 = 1.50 𝑚𝑚
We know that,
𝑑𝑢
𝑆𝑡𝑟𝑒𝑠𝑠 𝜎 = 𝐸.
𝑑𝑥
𝐸(𝑢 − 𝑢 )
𝑭𝒐𝒓 𝒆𝒍𝒆𝒎𝒆𝒏𝒕 (𝟏): 𝑆𝑡𝑟𝑒𝑠𝑠 𝜎 =
𝑙
2 × 10 (1.5 − 0)
=
150
𝜎 = 2000 𝑁/𝑚𝑚 [𝑇𝑒𝑛𝑠𝑖𝑙𝑒 𝑠𝑡𝑟𝑒𝑠𝑠]
𝐸(𝑢 − 𝑢 )
𝑭𝒐𝒓 𝒆𝒍𝒆𝒎𝒆𝒏𝒕 (𝟐): Stress 𝜎 =
𝑙
2 × 10 (1.2 − 1.5)
=
150
= −400 𝑁/𝑚𝑚 [Compressive stress]
We know that,
Reaction force, {R} = [K]{ u∗ } − {F}
𝑅 3.333 −3.333 0 𝑢 𝐹
 𝑅 = 1 × 10 −3.333 6.666 −3.333 𝑢 − 𝐹
𝑅 0 −3.333 3.333 𝑢 𝐹
2.88 One Dimensional Problems
3.333 −3.333 0 0 0
= 1 × 10 −3.333 6.666 −3.333 1.50 − 6 × 10
0 −3.333 3.333 1.2 0
0 + 3.333(1.50) + 0 0
= 1 × 10 0 + 6.6666 (1.50) − 3.3333 (1.2) − 6 × 10
0 − 3.333(1.50) − 3.3333 (1.2) 0
−5 0
= 1 × 10 6 − 6 × 10
1 0
−5 × 10 0
= 6 × 10 − 6 × 10
−1 × 10 0
𝑅 −5 × 10
𝑅 = 0
𝑅 −1 × 10
 𝑅 = −5 × 10 𝑁 = −500 𝑘𝑁
𝑅 =0
𝑅 = −1 × 10 𝑁 = −100 𝑘𝑁
We know that, reaction force is equivalent and opposite to the applied force.
𝐕𝐞𝐫𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧: R + R + R = −500 kN + 0 − 100 kN,
= −600 𝑘𝑁 [𝐴𝑝𝑝𝑙𝑖𝑒𝑑 𝑓𝑜𝑟𝑐𝑒]
Result: (𝑎). 𝑢 = 0
𝑢 = 1.5 𝑚𝑚
𝑢 = 1.2 𝑚𝑚
(𝑏). 𝜎 = 2000 𝑁/𝑚𝑚 [𝑇𝑒𝑛𝑠𝑖𝑙𝑒]
𝜎 = −400 𝑁/𝑚𝑚 [𝑐𝑜𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑣𝑒]
(𝑐). 𝑅 = −500 𝑘𝑁
𝑅 = 0 𝑘𝑁
𝑅 = −100 𝑘𝑁
 Finite Element Analysis 2.89

2.10. SPRINGS
Consider a spring element with nodes 1, 2 as shown in Fig.2.2S. Let k be the
spring constant and I be the length of the spring. F1 and F2 are the nodal forces acting at
node1 and 2 respectively. u1 and u2 are the displacements at the respective nodes.

Fig. 2.25.
By the sign convention for nodal forces and equilibrium, we have
𝐹 = −𝑇 𝑎𝑛𝑑 𝐹 = 𝑇
Where T  Tensile force.
We know that,
Tensile force, T = k u
Where, k  spring constant, N/m
u  Change in deformation, m.
 F1 = - k u

= 𝑘(𝑢 − 𝑢 )

𝐹 = 𝑘(𝑢 − 𝑢 ) … (2.76)

 F2 = - k u

𝐹 = 𝑘(𝑢 − 𝑢 ) … (2.77)

Arranging equation (2.76) and (2.77) in matrix form,

𝐹 𝑘−𝑘 𝑢
 =
𝐹 −𝑘 𝑘 𝑢
𝐹 1 −1 𝑢
=𝑘 … (2.78)
𝐹 −1 1 𝑢
This is a finite element equation for spring element.
2.90 One Dimensional Problems

2.10.1 Solved Problems

Example 2.18
A spring assemblage with arbitrarily numbered nodes are shown in Fig. The
nodes 1 and 2 are fixed and a force of 500 kN is applied at node 4 ill the x
direction. Calculate the following:
(i) Global stiffness matrix.
(ii) Nodal displacements.
(iii) Reactions at each nodal point.

Fig. (i)
𝑇𝑎𝑘𝑒 𝑆𝑝𝑟𝑖𝑛𝑔 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡, 𝑘 = 100 𝑘𝑁⁄𝑚 ; 𝑘 = 200 𝑘𝑁 ⁄𝑚 ; 𝑘 = 300 𝑘𝑁⁄𝑚 ;
Given:

Fig. (ii)
Nodal force, F4 = 500 kN
Spring constant, k1 = 100 kN/m
k2 = 200 kN/m
k3 = 300 kN/m
To find: (i) Global stiffness matrix [K].
(ii) Nodal displacements, 𝑢 , 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢 .
(iii) Reactions at each nodal point, 𝑅 , 𝑅 , 𝑅 𝑎𝑛𝑑 𝑅 .
 Finite Element Analysis 2.91

Solution: we know that,

Finite element equation for spring element is
𝐹 1 −1 𝑢
=𝑘 … [From equation no. (2.78)]
𝐹 −1 1 𝑢
For element 1: (Nodes 1, 3): Finite element equation is,
1 −1 𝑢 𝐹
𝑘 𝑢 =
−1 1 𝐹

1 −1 𝑢 𝐹
 100 𝑢 =
−1 1 𝐹

1 3
100 −100 1 𝑢 𝐹
 = … (1)
−100 100 3 𝑢 𝐹
For element 2: (Nodes 3, 4): Finite element equation is,
1 −1 𝑢 𝐹
𝑘 𝑢 =
−1 1 𝐹

1 −1 𝑢 𝐹
200 𝑢 =
−1 1 𝐹

3 4
200 −200 3 𝑢 𝐹
 = … (2)
−200 200 4 𝑢 𝐹
For element 3: (Nodes 4, 2): Finite element equation is,
1 −1 𝑢 𝐹
𝑘 𝑢 =
−1 1 𝐹

1 −1 𝑢 𝐹
300 𝑢 =
−1 1 𝐹

4 2
300 −300 4 𝑢 𝐹
= … (3)
−300 300 2 𝑢 𝐹
Assemble equations (1), (2) and (3)
2.92 One Dimensional Problems

1 2 3 4
100 0 −100 0 1 𝑢 𝐹
0 300 0 −300 2 𝑢 𝐹
𝑢 =
−100 0 200 + 100 −200 3 𝐹
0 −300 −200 200 + 300 4 𝑢 𝐹
100 0 −100 0 𝑢 𝐹
0 300 0 −300 𝑢 𝐹
 𝑢 =
−100 0 300 −200 𝐹
𝑢 … (4)
0 −300 −200 500 𝐹
↓
[𝐾]
Applying boundary conditions:
At node 1, 𝑢 = 0
At node 2, 𝑢 = 0
Nodal forces, F = F = F = 0
𝐹 = 500 𝑘𝑁 [𝐺𝑖𝑣𝑒𝑛]
Substitute 𝐹 , 𝐹 , 𝐹 𝑎𝑛𝑑 𝐹 . 𝑢 𝑎𝑛𝑑 𝑢 values in equation (4).
100 0 −100 0 0 0
0 300 0 −300 0 0
 2 × 10 𝑢 = 0
−100 0 300 −200
0 −300 −200 500 𝑢 𝑢
In the above equation, u1 = 0. So, neglect first row and first column of [ K]
matrix. 𝑢 = 0, so, neglect second row and second column of [K] matrix. Hence, the
equation reduces to
300 −200 𝑢 0
𝑢 =
−200 500 500
300 𝑢 − 200 𝑢 = 0 … (5)
−200 𝑢 + 500 𝑢 = 500 … (6)
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (5) × 2  600 𝑢 − 400 𝑢 = 0
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (6) × 3  − 600 𝑢 + 1500 𝑢 = 1500
Solving, 1100 𝑢 = 1500
 Finite Element Analysis 2.93
1500
𝑢 =
1100
𝑢 = 1.364 𝑚
Substitute u4 value in equation (5),
 300 𝑢 − 200(1.364) = 0
 𝑢 = 0.9091 𝑚
We know that,
Reaction force, {R} = [K]{ u∗ } − {F}
𝑅 100 0 −100 0 𝑢 𝐹
𝑅 0 300 0 −300 𝑢 𝐹
 = 𝑢 −
𝑅 −100 0 300 −200 𝐹
𝑅 0 −300 −200 500 𝑢 𝐹
100 0 −100 0 0 0
= 0 300 0 −300 0 − 0
−100 0 300 −200 0.9091 0
0 −300 −200 500 1.364 500
0 + 0 − 100(0.9091) + 0 0
⎡ ⎤
0 + 0 + 0 − 300(1.364)
=⎢ ⎥− 0
⎢ 0 + 0 + 300(0.9091) − 200(1.364) ⎥ 0
⎣0 + 0 + 200(0.9091) + 500(1.364)⎦ 500
−90.91 0
= −409.20 − 0
0 0
500 500
𝑅 −90.91
𝑅
= −409.20
𝑅 0
𝑅 0
 𝑅 = −90.91 𝑘𝑁
𝑅 = −409.20 𝑘𝑁
𝑅 = 0
𝑅 = 0
2.94 One Dimensional Problems

Result:
100 0 −100 0
(i) [K] = 0 300 0 −300
−100 0 300 −200
0 −300 −200 500
(ii) 𝑢 =𝑢 =0
𝑢 = 0.9091 𝑚,
𝑢 = 1.364 𝑚
(iii) 𝑅 = −90.91 𝑘𝑁
𝑅 = −409.20 𝑘𝑁
𝑅 = 0
𝑅 = 0

Example 2.19

For the bar assemblages shown in fig .(i), determine the nodal displacements, the
forces in each element and the reactions

Fig. (i)

E = 70 GPa, A = 2  10-4 m2, k = 2000 kN/m

Given:

Fig. (ii)
 Finite Element Analysis 2.95

Point load, P = 8kN

Young’s modules, E = 70 GPa = 70  109 Pa

= 70  109 N/m2

= 70  106 kN/m2

Area, A = 2  10-4 m2

Spring constant, k = 2000 kN/m

Length of the bar, l1 = 2 m

l2 = 2 m

To find:

(i) Nodal displacements, 𝑢 , 𝑢 , 𝑢 and 𝑢

(ii) Nodal forces in each element.

(iii) Nodal Reactions in each element 𝑅 , 𝑅 , 𝑅 and 𝑅

Solution: Finite element equation for one dimensional two noded bar element is given
by
𝐹 𝐴𝐸 1 −1 𝑢
=
𝐹 𝑙 −1 1 𝑢

For element 1: (Nodes 1, 2):

Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

2 × 10 × 70 × 10 1 −1 𝑢 𝐹
⇒ =
2 −1 1 𝑢 𝐹

1 −1 𝑢 𝐹
⇒ 7 × 10 =
−1 1 𝑢 𝐹
2.96 One Dimensional Problems
𝟏 𝟐 𝑢 𝐹
⇒ 10 7 −7 𝟏 𝑢 = … (1)
𝐹
−7 7 𝟐

For element 2: (Nodes 2, 3): Finite element equation is,

𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

2 × 10 × 70 × 10 1 −1 𝑢 𝐹
⇒ =
2 −1 1 𝑢 𝐹

1 −1 𝑢 𝐹
⇒ 7 × 10 =
−1 1 𝑢 𝐹

𝟐 𝟑 𝑢 𝐹
⇒ 10 7 −7 𝟐 𝑢 = … (2)
𝐹
−7 7 𝟑

For element 3: (Nodes 3, 4): Finite element equation for spring element is given by,

F 1 −1 u
=k [From equation no. 2.78)]
F −1 1 u

1 −1 u F
⇒ 𝑘 =
−1 1 u F

1 −1 u F
⇒ 2000 =
−1 1 u F

𝟑 𝟒 u F
⇒ 10 2 −2 𝟑 u = … (3)
F
−2 2 𝟒

Assemble the equation (1), (2) and (3),

1 2 3 4
7 −7 0 0 𝟏 𝑢 𝐹
−7 7 + 7 −7 0 𝟐 𝑢 𝐹
⇒ 10 𝑢 =
0 −7 7+2 −2 𝟑 𝐹
0 0 −2 2 𝟒 𝑢 𝐹
 Finite Element Analysis 2.97

7 −7 0 0 𝑢 𝐹
−7 14 −7 0 𝑢 𝐹
⇒ 10 𝑢 = … (4)
0 −7 9 −2 𝐹
0 0 −2 2 𝑢 𝐹

Applying boundary conditions:

Ad node 1, 𝑢 = 0

Ad node 4, 𝑢 = 0

Nodal force 𝐹 = 𝐹 = 𝐹 = 0

F = 8 kN [Given]

Substitute 𝑢 , 𝑢 , 𝐹 , 𝐹 and 𝐹 values in equation (4)

7 −7 0 0 0 0
⇒ 10 −7 14 −7 0 𝑢 = 8
0 −7 9 −2 𝑢 0
0 0 −2 2 0 0
↓
[K]

In the above equation, 𝑢 = 0. So, delete first row and first column of [K] matrix.
𝑢 = 0, so, delete fourth row and fourth column of [ K ] matrix. Hence the equation
reduces to,
14 −7 𝑢 8
10 𝑢 =
−7 9 0
⇒ 10 (14𝑢 − 7𝑢 ) = 8 … (5)

⇒ 10 (−7𝑢 + 9𝑢 ) = 0 … (6)

10 (14𝑢 − 7𝑢 ) = 8

𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (6) × 2 ⇒ 10 (−14𝑢 + 18𝑢 ) = 0

Solving, 10 (11u ) = 8

8
⇒ 𝑢 =
11 × 10
2.98 One Dimensional Problems

⇒ 𝑢 = 0.727 × 10 𝑚

Substitute 𝑢 value in equation (6),

⇒ 10 (−7𝑢 + 9 × 0.727 × 10 ) = 0

⇒ 𝑢 = 0.935 × 10 𝑚

We know that, Reaction force, {R} = [ K ] { u*} – {F}

𝑅 7 −7 0 0 𝑢 𝐹
𝑅 −7 14 −7 0 𝑢 𝐹
⇒ = 10 𝑢 −
𝑅 0 −7 9 −2 𝐹
𝑅 0 0 −2 2 𝑢 𝐹

7 −7 0 0 0 0
−7 14 −7 0 0.935 × 10
= 10 − 8
0 −7 9 −2 0.727 × 10 0
0 0 −2 2 0 0

0 − 7 × 0.935 × 10 + 0 + 0 0
= 10 0 + 14 × 0.935 × 10 − 7 × 0.727 × 10 + 0 − 8
0 − 7 × 0.935 × 10 + 9 × 0.727 × 10 + 0 0
0 + 0 − 2 × 0.727 × 10 + 0 0

−6.546 × 10 0 −6.546 0
= 10 8 × 10 − 8 = 8 − 8
0 0 0 0
−1.455 × 10 0 −1.455 0

𝑅 −6.546
𝑅 8
⇒ =
𝑅 0
𝑅 −1.455

⇒ 𝑅 = −6546 𝑘𝑁

𝑅 =0

𝑅 =0

𝑅 = −1.455 𝑘𝑁
 Finite Element Analysis 2.99

We know that, Reaction force is equivalent and opposite to the applied force.

Verification: 𝑅 + 𝑅 + 𝑅 + 𝑅 = −6.546 + 0 + 0 − 1.455

= −8 𝑘𝑁 [Applied force]

Forces in each element:

For element 1: (Nodes 1, 2):

Finite element equation for element (1) is

𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹

2 × 10 × 70 × 10 1 −1 0 𝐹
⇒ =
2 −1 1 0.935 × 10 𝐹

7 −7 0 𝐹
⇒ 10 =
−7 7 0.935 × 10 𝐹

⇒ 10 (0 − 7 × 0.935 × 10 ) = 𝐹

⇒ 10 (0 + 7 × 0.935 × 10 ) = 𝐹

⇒ 𝐹 = −6.546 𝑘𝑁

𝐹 = 6.546 𝑘𝑁

For element (1): Force at node 1, 𝐹 = −6.546 𝑘𝑁

Force at node 2, 𝐹 = 6.546 𝑘𝑁

For element 2: (Nodes 2,3):

Finite element equation for element (2) is

𝐴 𝐸 1 −1 𝑢 𝐹
=
𝑙 −1 1 𝑢 𝐹
2.100 One Dimensional Problems

2 × 10 × 70 × 10 1 −1 0.935 × 10 𝐹
⇒ =
2 −1 1 0.727 × 10 𝐹

7 −7 0.935 × 10 𝐹
⇒ 10 =
−7 7 0.727 × 10 𝐹

⇒ 10 (0 − 7 × 0.935 × 10 − 7 × 0.727 × 10 ) = 𝐹

⇒ 10 (0 − 7 × 0.935 × 10 + 7 × 0.727 × 10 ) = 𝐹

⇒ 𝐹 = 1.455 𝑘𝑁

𝐹 = −1.455 𝑘𝑁

For element (2): Force at node 2, 𝐹 = 1.455 𝑘𝑁

Force at node 3, 𝐹 = −1.455 𝑘𝑁

For element 3: (Nodes 3, 4):

Finite element equation for spring element is given by,

1 −1 u F
𝑘 =
−1 1 u F

1 −1 0.727 × 10 F
2000 =
−1 1 0 F

2 −2 0.727 × 10 F
10 =
−2 2 0 F

⇒ 10 (2 × 0.727 × 10 − 0) = 𝐹

⇒ 10 (−2 × 0.727 × 10 + 0) = 𝐹

⇒ 𝐹 = 1.455 𝑘𝑁

𝐹 = −1.455 𝑘𝑁

For element (3): Force at node 3, 𝐹 = 1.455 𝑘𝑁

Force at node 4, 𝐹 = −1.455 𝑘𝑁

 Finite Element Analysis 2.101

Result: (i) Nodal displacements

𝑢 =0

𝑢 = 0.935 × 10 m

𝑢 = 0.727 × 10 m

𝑢 =0

(ii) Nodal forces in each element:

For element (1): F1 = - 6.546 kN; F2 = 6.548 kN

For element (2): F2 = 1.455 kN

F3 = - 1.455 kN

For element (3): F3 = 1.455 kN

F4 = - 1.455 kN

(iii) Nodal reactions in each element:

𝑅 = −6546 𝑘𝑁

𝑅 =0

𝑅 =0

𝑅 = −1.455 𝑘𝑁

2.11. TRUSSES
2.11.1. Introduction
A truss is defined as a structure, made up of several bars, riveted or welded
together. The following assumptions are made while finding the forces in a truss.
(i) All the members are pin jointed.
(ii) The truss is loaded only at the joints.
2.102 One Dimensional Problems

(iii) The self-weight of the members are neglected unless stated.

A two dimensional truss element is shown in Fig.2.26

Fig. 2.26. A two dimensional truss element

2.11.2. Stiffness Matrix [ K ] for a Truss Element

Consider a two noded bar element as shown in Fig.2.27, for the analysis of
trusses. This element is subjected to only axial forces. So, the displacements are only
in the axial directions.
The nodal displacement for this bar element is given by,
𝑢
{𝑢 } =
𝑢

Fig. 2.27 Two noded bar element

 Finite Element Analysis 2.103

From the Fig. 2.27, we know that,

𝑢 = 𝑢 cos 𝜃 + 𝑢 sin 𝜃
𝑢 = 𝑢 cos 𝜃 + 𝑢 sin 𝜃
Consider l and m are the direction consines, So, we take l = cos  and m = sin .
⇒ 𝑢 =𝑢 𝑙+𝑢 𝑚 …(2.79)
𝑢 =𝑢 𝑙+𝑢 𝑚 …(2.80)
The above equations can now be writer in matrix form as,
𝑢
𝑢 𝑙 𝑚 0 0 𝑢
= 𝑢
𝑢 0 0 𝑙 𝑚
𝑢
⇒ {𝑢 } = [ 𝐿]{𝑢} …(2.81)
𝑙 𝑚 0 0
Where, [L]= and it is called transformation matrix.
0 0 𝑙 𝑚
Referring to Fig. 2.28, let (𝑥 , 𝑦 ) and (𝑥 , 𝑦 ) be the co-ordinates of nodes 1 and
2 respectively. We can find l, m and 𝑙 values by using the following formulae.

Fig. 2.28 Direction cosines

𝑥 −𝑥
𝑙 = cos 𝜃 = … (2.82)
𝑙
𝑦 −𝑦
𝑚 = sin 𝜃 = … (2.83)
𝑙

𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 ) … (2.84)
2.104 One Dimensional Problems

From Fig. 2.28., we know that the truss is also a one dimensional two noded bar
element. The stiffness matrix for two noded bar element is given by,
𝐴 𝐸 1 −1
[𝐾 ] = … (2.85)
𝑙 −1 1
We know that,
1
Strain energy, 𝑈 = {𝑢 } [𝐾 ]{𝑢 } … (2.86)
2
From equation (2.81), we know that,
{𝑢 } = [ 𝐿] {𝑢}
Substitute {𝑢 } value in equation (2.86),

1
⇒ 𝑈 = ([𝐿]{𝑢 }) [𝐾 ][𝐿]{𝑢}
2
1
= [𝐿] {𝑢} [𝐾 ] [𝐿] {𝑢}
2
1
𝑈= {𝑢} {𝑢}[𝐾] … (2.87)
2
1
𝑊ℎ𝑒𝑟𝑒, [𝐾] = [𝐿] [𝐾 ] [𝐿]
2

Element stiffness matrix in global co-ordinates.

1
[𝐾] = [𝐿] [𝐾 ] [𝐿] … (2.88)
2

Substitute [L] value from equation (2.81) and [𝐾 ] value from equation (2.85).
𝑙 0
⇒ [𝐾] = 𝑚 0 𝐴 𝐸 1 −1 𝑙 𝑚 0 0
0 𝑙 𝑙 −1 1 0 0 𝑙 𝑚
0 𝑚
𝑙 0
𝑙 𝑚 0 0 [𝐿]
∵ [𝐿] = ; = 𝑚 0
0 0 𝑙 𝑚 0 𝑙
0 𝑚
 Finite Element Analysis 2.105
𝑙 0
𝐴 𝐸 𝑚 0 1 −1 𝑙 𝑚 0 0
⇒ [𝐾] =
𝑙 0 𝑙 −1 1 0 0 𝑙 𝑚
0 𝑚
𝑙 0
𝐴 𝐸 𝑚 0 𝑙−0 𝑚−0 0−𝑙 0−𝑚
=
𝑙 0 𝑙 −𝑙 + 0 −𝑚 + 0 0 + 𝑙 0+𝑚
0 𝑚
[∵ (2 × 2) × (2 × 4) = 2 × 4]

𝑙 0
𝐴 𝐸 𝑚 0 𝑙 𝑚 −𝑙 −𝑚
=
𝑙 0 𝑙 −𝑙 −𝑚 𝑙 𝑚
0 𝑚
𝑙 𝑙𝑚 − 0 −𝑙 + 0 −𝑙𝑚 + 0
𝐴 𝐸 𝑙𝑚 − 0 𝑚 − 0 −𝑚𝑙 + 0 −𝑚 + 0
=
𝑙 0−𝑙 0 − 𝑙𝑚 0+𝑙 0 + 𝑙𝑚
0 − 𝑙𝑚 0 − 𝑚 0 + 𝑙𝑚 0+𝑚
[∵ (4 × 2) × (2 × 4) = 4 × 4]

𝑙 𝑙𝑚 −𝑙 −𝑙𝑚
𝐴 𝐸 𝑙𝑚 𝑚 −𝑚𝑙 −𝑚
[𝐾] = … (2.89)
𝑙 −𝑙 −𝑙𝑚 𝑙 𝑙𝑚
−𝑙𝑚 −𝑚 𝑙𝑚 𝑚
It may be noted that the stiffness matrix properties are satisfied.
1. [K] matrix is symmetric.
2. The sum of elements in any column is equal to zero.

2.11.3 Finite Element Equation for a Two Noded Truss element

We know that, finite element general equation is,
{F}=[K]{u} …(2.90)
Where, { F } is a element force vector [ Column Matrix ].
{ K } is a stiffness matrix [ Row Matrix ].
2.106 One Dimensional Problems

{ u } is a nodal displacements [ Column Matrix ].

We know that, For truss element, stiffness matrix,
𝑙 𝑙𝑚 −𝑙 −𝑙𝑚
𝐴 𝐸 𝑙𝑚 𝑚 −𝑚𝑙 −𝑚
[𝐾] =
𝑙 −𝑙 −𝑙𝑚 𝑙 𝑙𝑚
−𝑙𝑚 −𝑚 𝑙𝑚 𝑚
Substitute [ K ] value in equation (2.90),
𝐹 𝑙 𝑙𝑚 −𝑙 −𝑙𝑚 𝑢
𝐹 𝐴 𝐸 𝑙𝑚 𝑚 −𝑚𝑙 −𝑚 𝑢
⇒ = 𝑢 … (2.91)
𝐹 𝑙 −𝑙 −𝑙𝑚 𝑙 𝑙𝑚
𝐹 −𝑙𝑚 −𝑚 𝑙𝑚 𝑚 𝑢

This is a finite element equation for a truss element.

2.11.4 Solved Problems

EXAMPLE 2.20
For the two bar truss shown in fig.(i), determine the displacements of node 1 and the
stress in element 1-3.

Fig. (i)
Given: Yound’s modulus, E = 70 GPa = 70  109 Pa
= 70  109 N/m2
= 70  109 N/mm2
Area, A = 200 mm2
Point load at node 1 = 12 kN = 12  103 N
 Finite Element Analysis 2.107

Fig. (ii)
To find: 1. Displacements of node, 1
2. stress in element, (2)
Solution: Consider node 1 as the origin.
The co-ordinates of various nodes are given below.
𝑥 𝑦
Node 1 = (0, 0)
𝑥 𝑦
Node 2 = (-500, 0)
𝑥 𝑦
Node 3 = (400, -300)

For element (1): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (−500 − 0) + (0 − 0)
𝑙 = 500 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 =
𝑙
−500 − 0
=
500
𝑙 = −1
𝑦 −𝑦
𝑚 =
𝑙
0−0
=
500
2.108 One Dimensional Problems

𝑚 =0

For element (2): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (400 − 0) + (−300 − 0)
𝑙 = 500 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 =
𝑙
400
=
500
𝑙 = 0.8
𝑦 −𝑦
𝑚 =
𝑙
−300 − 0
=
500
𝑚 = −0.6

Fig. (iii)
For element (1): Displacements 𝑢 , 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢
Stiffness matrix [ K ] for a truss element is given by,
1 2 3 4
𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚 1
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥ 2
[𝐾] = ⎢ −𝑙
𝑙 −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥ 3
⎢
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦4

[From equation no. (2.89)]

 Finite Element Analysis 2.109
1 0 −1 0
200 × 70 × 10 0 0 0 0
=
500 −1 0 1 0
0 0 0 0
1 2 3 4 Global numbers
1 0 −1 0 𝟏
0 0 0 0 𝟐
[𝐾] = 28 × 10 …(1)
−1 0 1 0 𝟑
0 0 0 0 𝟒

For element (2): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔

1 2 5 6
𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚 1
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥ 2
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥ 5
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦6

0.64 −0.48 −0.64 0.48

200 × 70 × 10 −0.48 0.36 0.48 −0.36
=
500 −0.64 −0.48 0.64 −0.48
0.48 −0.36 −0.48 0.36

1 2 5 6
0.64 −0.48 −0.64 0.48 1
−0.48 0.36 0.48 −0.36 2
[𝐾] = 28 × 10 … (2)
−0.64 −0.48 0.64 −0.48 5
0.48 −0.36 −0.48 0.36 6

Assemble the stiffness matrix [ K ], i.e., assemble the equation (1) and (2)
1 2 3 4 5 6
1 1 1
+ + 0 0 0.64 0.48
0.64 -0.48
[ K ] = 28  103
1 1 2
+ + 0 0 0.48 -0.36
-0.48 0.36
2.110 One Dimensional Problems

-1 0 1 0 0 0 3
0 0 0 0 0 0 4
-0.64 0.48 0 0 0.64 -0.48 5
0.48 -0.36 0 0 -0.48 0.36 6

1.64 −0.48 −1 0 −0.64 0.48

⎡−0.48 0.36 0 0 0.48 −0.36 ⎤
⎢ −1 0 1 0 0 0 ⎥⎥
⇒ [𝐾] = 28 × 10 ⎢ … (3)
⎢ 0 0 0 0 0 0 ⎥
⎢−0.64 0.48 0 0 0.64 −0.48 ⎥
⎣−0.48 −0.36 0 0 −0.48 0.36 ⎦

[Note: The given truss element has 3 nodes and each node has 2 degree of
freedom. So, total degrees of freedom are 6 (𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢 ). Hence the
stiffness matrix size is 6  6.]
We know that, General finite element equation is
{F}=[K]{u}

[K]{u}={F}

1.64 −0.48 −1 0 −0.64 0.48 𝑢 𝐹

⎡−0.48 0.36 ⎤ ⎧ ⎫ ⎧𝐹 ⎫
0 0 0.48 −0.36 𝑢
⎢ −1 0 1 0 0 0 ⎥⎥ ⎪𝑢 ⎪ ⎪𝐹 ⎪
28 × 10 ⎢ = … (4)
⎢ 0 0 0 0 0 0 ⎥ ⎨𝑢 ⎬ ⎨𝐹 ⎬
⎢−0.64 0.48 0 0 0.64 −0.48 ⎥ ⎪𝑢 ⎪ ⎪𝐹 ⎪
⎣−0.48 −0.36 0 0 −0.48 0.36 ⎦ ⎩𝑢 ⎭ ⎩𝐹 ⎭

Applying boundary conditions [Refer Fig. (iii)]:

(i) Node 2 is fixed. So, 𝑢 = 𝑢 = 0
(ii) Node 3 is fixed. So, 𝑢 = 𝑢 = 0
(iii) A point load of 12  103 N is acting at node 1 in downward direction.
So, F2 = - 12  103 N
(iv) Self-weight is neglected. So, F1 = F3 = F4 = F5 = F6 = 0
 Finite Element Analysis 2.111

Substitute the above values in equation (4)

1.64 −0.48 −1 0 −0.64 0.48 𝑢 0
⎡−0.48 0.36 0 0 0.48 −0.36 ⎤ ⎧ 𝑢 ⎫ ⎧−12 × 10 ⎫
⎪ ⎪
⎢ −1 0 1 0 0 0 ⎥⎥ ⎪ 0 ⎪ 0
⇒ 28 × 10 ⎢ = 0
⎢ 0 0 0 0 0 0 ⎥ ⎨0⎬ ⎨ ⎬
⎢−0.64 0.48 0 0 0.64 −0.48 ⎥ ⎪ 0 ⎪ ⎪ 0 ⎪
⎣−0.48 −0.36 0 0 −0.48 0.36 ⎦ ⎩ 0 ⎭ ⎩ 0 ⎭
In the above equation 𝑢 = 𝑢 = 𝑢 = 𝑢 = 0. So, delete third row third
column, fourth row, fourth column, fifth row fifth column and sixth row sixth column of
[ K ] matrix. Hence the equation reduces to
1.64 −0.48 𝑢 0
⇒ 28 × 10 𝑢 =
−0.48 0.36 −12 × 10

⇒ 28 × 10 (1.64𝑢 − 0.48 𝑢 ) = 0 … (5)

⇒ 28 × 10 (−0.48 𝑢 − 0.36 𝑢 ) = −12 × 10 … (6)

Equation (6)  3.4166 

28 × 10 (−1.64 𝑢 − 1.23 𝑢 ) = −41 × 10 … (7)

28 × 10 (1.64 𝑢 − 0.48 𝑢 ) = 0 … (5)

Solving, 28 × 10 (0.75 𝑢 ) = −41 × 10

⇒ 𝑢 = −1.952 𝑚𝑚
Substitute 𝑢 value in equation (5),

28 × 10 (1.64𝑢 − 0.48 (−1.952)) = 0

⇒ 1.64 𝑢 − 0.48 (−1.952) = 0

⇒ 𝑢 = −0.571 𝑚𝑚
For element (1):
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚] 𝑢
𝑙
𝑢
2.112 One Dimensional Problems

For element (1): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔

𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚 ] 𝑢
𝑙
𝑢
−0.571
70 × 10
= [−0.8 −0.6 0.8 −0.6 ] −1.952
500 0
0
= 140 [(−0.8) × (−0.571) + 0.6 × (−1.952) + 0
+ 0]
𝜎 = −100 𝑁/𝑚𝑚 [compressive stress]
Result: 1. Displacement of node 1
𝑢 = −0.571 𝑚𝑚
𝑢 = −1.952 𝑚𝑚
2. Stress in element (2)
𝜎 = −100 𝑁/𝑚𝑚 [compressive stress]
Example 2.21
Consider a three bar truss as shown in Fig.(i). It i s given that E = 2  105
N/mm2. Calculate the following:
(i) Nodal displacements.
(ii) Stress in each member.
(iii) Reactions at the support.
Take Area of element (1) = 2000mm2
Area of element (2) = 2500 mm2
Area of element (3) = 2500 mm2
Given: Young's modulus, E = 2 X105 N/mm2
Area of element (1) = 2000 mm2
Area of element (2) = 2500 mm2
Area of element (3) = 2500 mm2
 Finite Element Analysis 2.113

Point load, P = - 250 kN = - 250  103 N

= - 2.5  105 N [∵ Load is downward direction]

Fig. (ii)
[Node: Number with circle = Element; Number without circle = Node]
To find:
(i) Nodal displacements, 𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢 .
(ii) Stress in each member, 𝜎 , 𝜎 𝑎𝑛𝑑 𝜎
(iii) Reactions at the support 𝑅 , 𝑅 , 𝑅 , 𝑅 , 𝑅 𝑎𝑛𝑑 𝑅 .
Solution: Consider node 1 as the origin.
The co-ordinates of various nodes are given below:
𝑥 𝑦
Node 1 = (0, 0)
𝑥 𝑦
Node 2 = (1000, 0)
𝑥 𝑦
Node 3 = (500, 500)

For element (1): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (1000 − 0) + (0 − 0)
𝑙 = 1000 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 =
𝑙
 1000 − 0
=
1000
𝑙 =1
𝑦 −𝑦
𝑚 =
𝑙
0−0
=
1000
𝑚 =0

For element (2): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (500 − 1000) + (500 − 0)

𝑙 = 707.107 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 =
𝑙
500 − 1000
=
707.107
𝑙 = −0.707
𝑦 −𝑦
𝑚 =
𝑙
500 − 0
=
707.107
𝑚 = 0.707

For element (3): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (500 − 0) + (500 − 0)
𝑙 = 707.107 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 =
𝑙
500 − 0
=
707.107
𝑙 = 0.707
 𝑦 −𝑦
𝑚 =
𝑙
500 − 0
=
707.107
𝑚 = 0.707

Fig. (iii)
For element (1): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟑 𝒂𝒏𝒅 𝒖𝟒 [Refer Fig. (iii)]
Stiffness matrix [ K ] for a truss element is given by,
𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
[𝐾] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

(1) 1×0 −(1) −(1 × 0)

⎡ ⎤
2000 × 2 × 10 1 × 0 (0) −(1 × 0) −(0) ⎥
= ⎢
1000 ⎢−(1) −(1 × 0) (1) 1 × 0⎥
⎣ 0 −(0) 1×0 (0) ⎦

1 0 −1 0
0 0 0 0
= 4 × 10
−1 0 1 0
0 0 0 0

1 2 3 4
1 0 −1 0 𝟏
0 0 0 0 𝟐
[𝐾] = 1 × 10 …(1)
−1 0 1 0 𝟑
0 0 0 0 𝟒
For element (2): Displacements 𝟏 , 𝒖𝟐 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔 :
𝒖 𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

(−0.707) −0.707 × 0.707 (−0.707) 0.707 × 0.707

2500 × 2 × 10 ⎡−0.707 × 0.707 (0.707) 0.707 × 0.707 (−0.707)
⎤
= ⎢ ⎥
707.107 ⎢ −(−0.707) −0.707 × 0.707 (−0.707) −0.707 × 0.707⎥
⎣ 0.707 × 0.707 −(0.707) −0.707 × 0.707 (0.707) ⎦

0.4998 −0.4998 −0.4998 0.4998

−0.4998 0.4998 0.4998 −0.4998
= 28 × 10
−0.4998 0.4998 0.4998 −0.4998
0.4998 −0.4998 −0.4998 0.4998

3 4 5 6
3.534 −3.534 −3.534 3.534 𝟑
−3.534 3.534 3.534 −3.534 𝟒 …(2)
[𝐾] = 1 × 10
−3.534 3.534 3.534 −3.534 𝟓
3.534 −3.534 −3.534 3.534 𝟔

For element (3): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔 :

𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

(−0.707) 0.707 × 0.707 −(0.707) −0.707 × 0.707

2500 × 2 × 10 ⎡−0.707 × 0.707 (0.707) 0.707 × 0.707 −(0.707) ⎥
⎤
= ⎢
707.107 ⎢ −(0.707) −0.707 × 0.707 (0.707) 0.707 × 0.707⎥
⎣−0.707 × 0.707 −(0.707) 0.707 × 0.707 (0.707) ⎦

0.4998 0.4998 −0.4998 −0.4998

0.4998 0.4998 −0.4998 −0.4998
= 28 × 10
−0.4998 −0.4998 0.4998 0.4998
−0.4998 −0.4998 0.4998 0.4998
 Finite Element Analysis 2.117

1 2 5 6
3.534 −3.534 −3.534 3.534 1
−3.534 3.534 3.534 −3.534 2 …(3)
[𝐾] = 1 × 10
−3.534 3.534 3.534 −3.534 5
3.534 −3.534 −3.534 3.534 6

Assemble the stiffness matrix [ K ], i.e., assemble the equation (2) and (3)
1 2 3 4 5 6
4 0
+ + -4 0 -3.534 -3.534 1
3.534 3.534
0 0
+ + 0 0 -3.534 -3.534 2
3.534 3.534
4 0
-4 0 + - -3.534 3.534 3
3.534 3.534
[ K ] = 1  105
0 0
0 0 - + 3.534 -3.534 4
3.534 3.534
-3.534 -3.534
-3.534 -3.534 -3.534 -3.534 + + 5
3.534 3.534
-3.534 -3.534
-3.534 -3.534 -3.534 -3.534 + + 6
3.534 3.534

7.534 3.534 −4 0 −3.534 −3.534

⎡ −0.48 0.36 0 0 −3.534 −3.534 ⎤
⎢ −4 0 7.534 −3.534 −3.534 ⎥
3.534 ⎥
⇒ [𝐾] = 1x10 ⎢ … (4)
⎢ 0 0 −3.534 3.534 3.534 −3.534 ⎥
⎢−3.534 −3.534 −3.534 3.534 7.068 0 ⎥
⎣−3.534 −3.534 3.534 −3.534 0 7.068 ⎦
 [Note: The given truss element has 3 nodes and each node has 2 degree of
freedom. So, total degrees of freedom are 6 (𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢 ). Hence the
stiffness matrix size is 6  6.]
We know that, General finite element equation is
{F}=[K]{u}
[K]{u}={F}

7.534 3.534 −4 0 −3.534 −3.534 𝑢 𝐹

⎡ −0.48 ⎧𝐹 ⎫
0.36 0 0 −3.534 −3.534 ⎤ ⎧𝑢 ⎫
⎪ ⎪
⎢ −4 0 ⎥ ⎪𝑢 ⎪ 𝐹
7.534 −3.534 −3.534 3.534 ⎥
1 × 10 ⎢ 𝑢 =
⎢ 0 0 −3.534 3.534 3.534 −3.534 ⎥ ⎨𝑢 ⎬ ⎨𝐹 ⎬
⎢−3.534 −3.534 −3.534 3.534 7.068 0 ⎥ ⎪ ⎪ ⎪𝐹 ⎪
⎣−3.534 −3.534 3.534 −3.534 0 7.068 ⎦ ⎩𝑢 ⎭ ⎩𝐹 ⎭
… (5)
Applying boundary conditions [Refer Fig. (iii)]:
i) Node 1 is fixed. So, 𝑢 = 𝑢 = 0
ii) Node 2 is moving in x direction. So, 𝑢 ≠ 0 𝑎𝑛𝑑 𝑢 = 0
iii) At node 3, Point load of 250  103 N is acting in downward direction.
So, F6 = - 250  103 N = -2.5  105 N
iv) Self-weight is neglected. So, F1 = F2 = F3 = F4 = F5 = 0.
Substitute 𝑢 , 𝑢 , 𝑢 , F1, F2, F3, F4, F5 and F6 values in equation (5).
7.534 3.534 −4 0 −3.534 −3.534 0
⎡ −0.48 0.36 0 0 −3.534 −3.534 ⎤ ⎧ 0 ⎫
⎢ −4 ⎥⎪ ⎪
0 7.534 −3.534 −3.534 3.534 ⎥ 𝑢
1 × 10 ⎢
⎢ 0 0 −3.534 3.534 3.534 −3.534 ⎥ ⎨ 0 ⎬
⎢−3.534 −3.534 −3.534 3.534 7.068 0 ⎥ ⎪𝑢 ⎪
⎣−3.534 −3.534 3.534 −3.534 0 7.068 ⎦ ⎩𝑢 ⎭
0
⎧ 0 ⎫
⎪ ⎪
0
= 0
⎨ ⎬
⎪ 0 ⎪
⎩−2.5 × 10 ⎭
In the above equation 𝑢 = 0. So, delete first row first column of [ K ] matrix.
𝑢 = 0 𝑎𝑛𝑑 𝑢 = 0. Hence delete second row, second column, fourth row and fourth
 Finite Element Analysis 2.119

column of [K] matrix.

The final reduced equation is,
7.534 −3.534 3.534 𝑢 0
1 × 10 −3.534 7.068 0 𝑢 = 0
3.534 0 7.068 𝑢 −2.5 × 10
7.534 −3.534 3.534 𝑢 0
−3.534 7.068 0 𝑢 = 0 … (6)
3.534 0 7.068 𝑢 −2.5
By using Gaussian elimination method, we can find 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢 values.
7.534 −3.534 3.534 0
𝐿𝑒𝑡, −3.534 7.068 0 0
3.534 0 7.068 −2.5
7.534 −3.534 3.534 0 𝑅
⇒ −3.534 7.068 0 0 𝑅 →
7.534
3.534 0 7.068 −2.5
1 −0.469 0.469 0
⇒ 0 5.410 1.657 0 𝑅 → 𝑅 + 3.534𝑅
3.534 0 7.068 −2.5
1 −0.469 0.469 0 𝑅
⇒ 0 5.410 1.657 0 𝑅 →𝑅 −
3.2649
0 0 4.902 −2.5

Now, we obtained the required triangular matrix

1 −0.469 0.469 𝑢 0
⇒ 0 5.410 1.657 𝑢 = 0
0 0 4.902 𝑢 −2.5

⇒ 4.902 𝑢 = −2.5

𝑢 = −0.5099 𝑚𝑚

⇒ 5.410 𝑢 + 1.657 𝑢 = 0

5.410 𝑢 + 1.657 (−0.5099) = 0

𝑢 = 0.1562 𝑚𝑚
2.120 One Dimensional Problems

⇒ 𝑢 − 0.469 𝑢 + 0.469 𝑢 = 0

𝑢 − 0.469 (0.1562) + 0.469(−0.5099) = 0

𝑢 − 0.3124 = 0

𝑢 = 0.3124 𝑚𝑚

We know that,
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚] 𝑢
𝑙
𝑢
For element (1):
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚 ] 𝑢
𝑙
𝑢
0
2 × 10 0
= [−1 −0 1 0 ]
1000 0.3124
0
2 × 10
= [0 − 0 + 0.3124 + 0]
1000
2 × 10
= × 0.3124
1000
𝜎 = 62.48 𝑁/𝑚𝑚 [Tensile stress]
For element (2):
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚 ] 𝑢
𝑙
𝑢
𝑢
2 × 10 𝑢
= [0.707 −0.707 −0.707 0.707 ] 𝑢
707.107
𝑢
 Finite Element Analysis 2.121
0.3124
0
= 282.842[0.707 −0.707 −0.707 0.707 ] 0.1562
−0.5099
= 282.842[0.707 × 0.3124 − 0.707 × 0 − 0.707 × 0.1562
+ 0.707 ](−0.5099)
= 282.842(−0.250062)
𝜎 = −70.729 𝑁/𝑚𝑚 [Compressive stress]
For element (3):
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚 ] 𝑢
𝑙
𝑢
0
2 × 10 0
= [−0.707 −0.707 0.707 0.707 ]
707.107 0.1562
−0.5099
= 282.842[0 − 0 + 0.707(0.1562) + 0.707 ](−0.5099)
= 282.842(−0.250065)
𝜎 = −70.729 𝑁/𝑚𝑚 [Compressive stress]
Reaction forces: we know that
Reaction force, { R } = [ K ] {u*} – { F }
𝑅 7.534 3.534 −4 0 −3.534 −3.534 𝑢
⎧𝑅 ⎫ ⎡ −0.48 ⎤ ⎧ ⎫
⎪ ⎪ 0.36 0 0 −3.534 −3.534 𝑢
⎢ −4 ⎪ ⎪
𝑅 0 7.534 −3.534 −3.534 3.534 ⎥⎥ 𝑢
= 1 × 10 ⎢
⎨𝑅 ⎬ ⎢ 0 0 −3.534 3.534 3.534 −3.534 ⎥ ⎨𝑢 ⎬
⎪𝑅 ⎪ ⎢−3.534 −3.534 −3.534 3.534 7.068 0 ⎥ ⎪𝑢 ⎪
⎩𝑅 ⎭ ⎣ −3.534 −3.534 3.534 −3.534 0 7.068 ⎦ ⎩𝑢 ⎭
𝐹
⎧𝐹 ⎫
⎪ ⎪
𝐹
−
⎨𝐹 ⎬
⎪𝐹 ⎪
⎩𝐹 ⎭
2.122 One Dimensional Problems
𝑢
⎧𝑢 ⎫
⎪𝑢 ⎪
⇒ {𝑅 } = 1 × 10 [7.534 3.534 − 4 0 − 3.534 − 3.534 ] 𝑢 − {𝐹 }
⎨𝑢 ⎬
⎪ ⎪
⎩𝑢 ⎭
0
0 ⎧⎫
⎪⎪
= 1 × 10 [7.534 3.534 − 4 0 − 3.534 − 3.534 ] 0.3124 − {0}
⎨ 0 ⎬
⎪ 0.1562 ⎪
⎩−0.5099⎭
= 1 × 10 [0 + 0 − 4 × 0.3124 + 0 − 3.534 × 0.1562
+ (−3.534) × (−0.5099) ]
{𝑅 } = 0
𝑢
⎧𝑢 ⎫
⎪𝑢 ⎪
⇒ {𝑅 } = 1 × 10 [3.534 3.534 0 0 − 3.534 − 3.534 ] 𝑢 − {𝐹 }
⎨𝑢 ⎬
⎪ ⎪
⎩𝑢 ⎭
0
0⎧ ⎫
⎪ ⎪
= 1 × 10 [3.534 3.534 0 0 − 3.534 − 3.534 ] 0.3124 − {0}
⎨ 0 ⎬
⎪ 0.1562 ⎪
⎩−0.5099⎭
= 1 × 10 [0 + 0 + 0 − 3.534 × 0.1562 + (−3.534) × (−0.5099) ]
{𝑅 } = 1.249 × 10 𝑁
𝑢
⎧𝑢 ⎫
⎪𝑢 ⎪
⇒ {𝑅 } = 1 × 10 [−4 0 7.534 − 3.534 − 3.534 3.534] 𝑢 − {𝐹 }
⎨𝑢 ⎬
⎪ ⎪
⎩𝑢 ⎭
0
⎧0 ⎫
⎪ ⎪
= 1 × 10 [−4 0 7.534 − 3.534 − 3.534 3.534 ] 0.3124 − {0}
⎨ 0 ⎬
⎪ 0.1562 ⎪
⎩−0.5099⎭
 Finite Element Analysis 2.123
= 1 × 10 [0 + 0 + 7.534 × 3.534 + 0 + (−3.534) × 0.1562
+ 3.534 × (−0.5099) ]
{𝑅 } = 0
𝑢
⎧𝑢 ⎫
⎪𝑢 ⎪
⇒ {𝑅 } = 1 × 10 [0 0 − 3.534 3.534 − 3.534 − 3.534] 𝑢 − {𝐹 }
⎨𝑢 ⎬
⎪ ⎪
⎩𝑢 ⎭
0
⎧ 0 ⎫
⎪ ⎪
= 1 × 10 [0 0 − 3.534 3.534 3.534 − 3.534 ] 0.3124 − {0}
⎨ 0 ⎬
⎪ 0.1562 ⎪
⎩−0.5099⎭
= 1 × 10 [0 + 0 − 3.534 × 0.3124 + 0 + 3.534 × 0.1562
+ (3.534) × (−0.5099) ]
{𝑅 } = 1.249 × 10 𝑁
𝑢
⎧𝑢 ⎫
⎪𝑢 ⎪
⇒ {𝑅 } = 1 × 10 [−3.534 − 3.534 − 3.534 3.534 7.068 0] 𝑢 − {𝐹 }
⎨𝑢 ⎬
⎪ ⎪
⎩𝑢 ⎭
0
0 ⎫⎧
⎪⎪
= 1 × 10 [−3.534 − 3.534 − 3.534 3.534 7.068 0 ] 0.3124 − {0}
⎨ 0 ⎬
⎪ 0.1562 ⎪
⎩−0.5099⎭
= 1 × 10 [0 + 0 + (−3.534) × 0.3124 + 0 + 7.068 × 0.1562 + 0 ]
{𝑅 } = 0
𝑢
⎧𝑢 ⎫
⎪𝑢 ⎪
⇒ {𝑅 } = 1 × 10 [−3.534 − 3.534 3.534 − 3.534 0 7.068] 𝑢 − {𝐹 }
⎨𝑢 ⎬
⎪ ⎪
⎩𝑢 ⎭
2.124 One Dimensional Problems
0
0 ⎧
⎫
⎪
⎪
= 1 × 10 [−3.534 − 3.534 3.534 − 3.534 0 7.068] 0.3124
⎨ 0 ⎬
⎪ 0.1562 ⎪
⎩−0.5099⎭
− 1 × 10 {−2.5}
= 1 × 10 [0 + 0 + 3.534 × 0.3124 + 0 + 0 + 7.068 × (−0.5099)]
− 1 × 10 {−2.5}
= −2.5 × 10 + 2.5 × 10
{𝑅 } = 0
We know that, Reaction force is equivalent and opposite to the applied force.
Verification: R1+R2+R3+R4+R5+R6 = 0 + 1.249  105 + 0 + 1.249 105+0+0
= 2.5 105N [Applied force]
Result: 1. Nodal displacement
𝑢 =0
𝑢 =0
𝑢 = 0.3124
𝑢 =0
𝑢 = 0.1562
𝑢 = −0.5099
2. Stress in each member:
𝜎 = 62.48 𝑁/𝑚𝑚 [Tensile]
𝜎 = −70.729 𝑁/𝑚𝑚 [compressive]
𝜎 = −70.729 𝑁/𝑚𝑚 [compressive]
3. Reaction forces:
𝑅 =0
𝑅 = 1.249 × 10 𝑁
𝑅 =0
 Finite Element Analysis 2.125

𝑅 = 1.249 × 10 𝑁
𝑅 =0
𝑅 =0

Example 2.22
Consider a four bar truss as shown i n Fi g . (i). It is given that E = 2 x10 5
N/mm2 and Ae = 625 mm2 for all elements.
(i) Determine the element stiffness matrix for each element.
(ii) Assemble the structural stiffness matrix [K] for the entire truss.
(iii) Solve for the nodal displacement.

Fig. (i)
Given: Young's Modulus, E = 2 X105 N/mm2
Area of each element (1), Ae = 625mm2
Load acting at node 3 = - 12000 N [∵ Load is downward direction]
Load acting at node 2 = 10000 N

Fig.(ii)
2.126 One Dimensional Problems

[Node: Number with circle denotes element & Number without circle denotes Node]
To find:
1. Element stiffness matrix for each element
2. Global stiffness matrix [ K ]
3. Nodal displacements 𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢
Solution: Consider node 1 as the origin.
The co-ordinated of various nodes are given below:
𝑥 𝑦
Node 1 = (0, 0)
𝑥 𝑦
Node 2 = (1000, 0)
𝑥 𝑦
Node 3 = (1000, 750)
𝑥 𝑦
Node 4 = (0, 750)

For element (1): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 ) [From equation No. (2.84)]

= (1000 − 0) + (0 − 0)
𝑙 = 1000 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 = [From equation No. (2.82)]
𝑙
1000 − 0
=
1000
𝑙 =1
𝑦 −𝑦
𝑚 = [From equation No. (2.83)]
𝑙
0−0
=
1000
𝑚 =0
 Finite Element Analysis 2.127

For element (2): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (1000 − 1000) + (750 − 0)

𝑙 = 750 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 =
𝑙
1000 − 1000
=
750
𝑙 =0
𝑦 −𝑦
𝑚 =
𝑙
750 − 0
=
750
𝑚 =1

For element (3): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (1000 − 0) + (750 − 0)
𝑙 = 1250 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 =
𝑙
1000 − 0
=
1250
𝑙 = 0.8
𝑦 −𝑦
𝑚 =
𝑙
750 − 0
=
1250
𝑚 = 0.6

For element (4): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (1000 − 0) + (750 − 750)

2.128 One Dimensional Problems

𝑙 = 1000 𝑚𝑚
𝑥 −𝑥
Direction consines, 𝑙 =
𝑙
1000 − 0
=
1000
𝑙 =1
𝑦 −𝑦
𝑚 =
𝑙
750 − 750
=
1000
𝑚 =0

Fig. (iii)
For element (1): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟑 𝒂𝒏𝒅 𝒖𝟒
Stiffness matrix [ K ] for a truss element is given by,
𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
[𝐾] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

(1) 1 × 0 −(1) 0
625 × 2 × 10 0 0 0 0
=
1000 −(1) 0 (1) 0
0 0 0 0
 Finite Element Analysis 2.129
1 0 −1 0
0 0 0 0
= 1.25 × 10
−1 0 1 0
0 0 0 0

1 2 3 4
1.25 0 −1.25 0 𝟏
0 0 0 0 𝟐
[𝐾] = 1 × 10 …(1)
−1.25 0 1.25 0 𝟑
0 0 0 0 𝟒

For element (2): Displacements 𝒖𝟑 , 𝒖𝟒 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔 :

𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

0 0 0 0
625 × 2 × 10 0 1 0 −1
=
750 0 0 0 0
0 −1 0 1
0 0 0 0
0 1 0 −1
= 1.666 × 10
0 0 0 0
0 −1 0 1

3 4 5 6
0 0 0 0 𝟑
0 1.666 0 −1.666 𝟒
[𝐾] = 1 × 10 …(2)
0 0 0 0 𝟓
0 −1.666 0 1.666 𝟔

For element (3): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔 :

𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦
2.130 One Dimensional Problems
(0.8) 0.8 × 0.6 −(0.8) −0.8 × 0.6
625 × 2 × 10 ⎡ 0.8 × 0.6 (0.6) −0.8 × 0.6 −(0.6) ⎥
⎤
= ⎢
750 ⎢ −(0.8) −0.8 × 0.6 (0.8) 0.8 × 0.6⎥
⎣−0.8 × 0.6 −(0.6) 0.8 × 0.6 (0.6) ⎦

1 2 5 6
0.64 0.48 −0.64 0.48 𝟏
0.48 0.36 −0.48 −0.36 𝟐 …(3)
[𝐾] = 1 × 10
−0.64 −0.48 0.64 0.48 𝟓
−0.48 −0.36 0.48 0.36 𝟔

For element (3): Displacements 𝒖𝟕 , 𝒖𝟖 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔 :

𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

(1) 0 −(1) 0
625 × 2 × 10 0 0 0 0
=
750 −(1) 0 1 0
0 0 0 0
1 0 −1 0
0 0 0 0
= 1.25 × 10
−1 0 1 0
0 0 0 0

7 8 5 6
1.25 0 −1.25 0 𝟕
0 0 0 0 𝟖 …(4)
[𝐾] = 1 × 10
−1.25 0 1.25 0 𝟓
0 0 0 0 𝟔
 Finite Element Analysis 2.131

Assemble the stiffness matrix [ K ], i.e., assemble the equation (1), (2), (3) and (4)
1 2 3 4 5 6 7 8
1.25 0 0 0
+ + -1.25 0 -0.64 -0.48 1
0.64 0.48
0 0 0 0
+ + 0 0 -0.48 -0.36 2
0.48 0.36
1.25 0 0 0
-1.25 0 + + 0 0 3
0 0
0 0 0 0
-
0 0 + + 0 4
[ K ] = 1  105 1.666
0 1.666
0 0 -1.25 0
+ +
-0.64 -0.48 0 0 0.64 0.48 5
+ +
1.25 0
0 1.666 0 0
+ +
-
-0.48 -0.36 0 0.48 0.36 6
1.666
+ +
0 0
0 0 0 0 -1.25 0 1.25 0 7
0 0 0 0 0 0 0 0 8
1.89 0.48 −1.25 0 −0.64 −0.48 0 0
⎡ 048 0.36 0 0 −0.48 −0.36 0 0⎤
⎢−1.25 0 1.25 0 0 0 0 0⎥
⎢ ⎥
⇒ [𝐾] = 1 × 10 ⎢ 0 0 0 1.666 0 −1.666 0 0⎥
⎢−0.64 −0.48 0 0 1.89 0.48 −1.25 0 ⎥
⎢−0.48 −0.36 0 −1.666 0.48 2.026 0 0⎥
⎢ 0 0 0 0 −1.25 0 1.25 0 ⎥
⎣ 0 0 0 0 0 0 0 0⎦
… (5)
2.132 One Dimensional Problems

We know that, General finite element equation is

{F}=[K]{u}

[K]{u}={F}

1.89 0.48 −1.25 0 −0.64 −0.48 0 0 𝑢

⎡ 048 0.36 0 0 −0.48 −0.36 0 0 ⎤ ⎧𝑢 ⎫
⎢−1.25 0 1.25 0 0 0 0 0⎥ ⎪ 𝑢 ⎪
⎢ ⎥ ⎪𝑢 ⎪
⇒ 1 × 10 ⎢ 0 0 0 1.666 0 −1.666 0 0⎥
⎢−0.64 −0.48 0 0 1.89 0.48 −1.25 0 ⎥ ⎨𝑢 ⎬
⎢−0.48 −0.36 0 −1.666 0.48 2.026 0 0 ⎥ ⎪𝑢 ⎪
⎢ 0 0 0 0 −1.25 0 1.25 0 ⎥ ⎪𝑢 ⎪
⎣ 0 0 0 0 0 0 0 0 ⎦ ⎩𝑢 ⎭
𝐹
⎧𝐹 ⎫
⎪𝐹 ⎪
⎪ ⎪
𝐹
= … (6)
⎨𝐹 ⎬
⎪𝐹 ⎪
⎪𝐹 ⎪
⎩𝐹 ⎭

Applying boundary conditions [Refer Fig. (iii)]:

1. Node 1 is fixed. So, u1 = u2 = 0.
2. Node 4 is fixed. So, u7 = u8 = 0.

3. Node 2 is moving in x direction. So, 𝑢 ≠ 0 𝑎𝑛𝑑 𝑢4 = 0.

4. At node 3, point load of 12000 N is acting in downward direction. So, F6 =

12000 N.
5. At node 2, point load of 10000 N is acting in x direction. So, F3 = 10000N.
6. Self-weight is neglected. So, F1 = F2 = F4 = F5= F7 = F8 = 0.
Substitute boundary conditions values in equation (6). u1 = u2 = u4 = u7 = u8 = 0.
So, delete first row first column, second row second column, fourth row fourth column,
seventh row seventh column and eighth row eighth column of [K] matrix.
 Finite Element Analysis 2.133

The final reduced equation is,

1.25 0 0 𝑢 10,000
1 × 10 0 1.89 0.48 𝑢 = 0
0 0.48 2.026 𝑢 −12,000

The above equation can be solved by using Gaussian elimination method.

1.25 0 0 10,000
𝐿𝑒𝑡, 0 1.89 0.48 0
0 0.48 2.026 −12,000

1 0 0 8,000 𝑅
⇒ 0 1.89 0.48 0 𝑅 →
−12,000 1.25
0 0.48 2.026

1 0 0 8,000
⇒ 0 1.89 0.48 0 𝑅 → 𝑅 − 0.25396𝑅
0 0.48 2.026 −12,000

Now, we obtained the required triangular matrix

1 0 0 𝑢 8000
⇒ 1 × 10 0 1.89 0.48 𝑢 = 0
0 0 1.9041 𝑢 −12,000

⇒ 1 × 10 (1.9041) 𝑢 = −2.5

𝑢 = −0.063 𝑚𝑚

⇒ 1 × 10 (1.89 𝑢 + 0.48 𝑢 ) = 0

1.89 𝑢 = −0.48 (−0.063)

𝑢 = 0.016 𝑚𝑚

⇒ 1 × 10 𝑢 = 8000

𝑢 = 0.08 𝑚𝑚
2.134 One Dimensional Problems

Result: (1) Element stiffness matrix for each element:

1.25 0 −1.25 0
[𝐾] = 1 × 10 0 0 0 0
−1.25 0 1.25 0
0 0 0 0
0 0 0 0
[𝐾] = 1 × 10 0 1.666 0 −1.666
0 0 0 0
0 −1.666 0 1.666
0.64 0.48 −0.64 −0.48
[𝐾] = 1 × 10 0.48 0.36 −0.48 −0.36
−0.64 −0.48 0.64 0.48
−0.48 −0.36 0.48 0.36
1.25 0 −1.25 0
[𝐾] = 1 × 10 0 0 0 0
−1.25 0 1.25 0
0 0 0 0

(2) Global stiffness matrix:

1.89 0.48 −1.25 0 −0.64 −0.48 0 0

⎡ 048 0.36 0 0 −0.48 −0.36 0 0⎤
⎢−1.25 0 1.25 0 0 0 0 0⎥
⎢ ⎥
⇒ [𝐾] = 1 × 10 ⎢ 0 0 0 1.666 0 −1.666 0 0⎥
⎢−0.64 −0.48 0 0 1.89 0.48 −1.25 0 ⎥
⎢−0.48 −0.36 0 −1.666 0.48 2.026 0 0⎥
⎢ 0 0 0 0 −1.25 0 1.25 0 ⎥
⎣ 0 0 0 0 0 0 0 0⎦

(3) Displacements:

𝑢 =0
𝑢 =0
𝑢 = 0.08 𝑚𝑚
𝑢 =0
𝑢 = 0.016 𝑚𝑚
𝑢 = −0.063 𝑚𝑚
 Finite Element Analysis 2.135

𝑢 =0
𝑢 =0

Example 2.23
For the plane truss shown in fig., determine the horizontal and vertical
displacements of nodal and the stresses in each element. all elements have E=201 Gpa
and A=4  10-4 m2

Fig. (i)
Given:
Youn’s modulus, E = 201 GPa = 201  109 Pa = 201  109 N/m2
Area of each element, A = 4  10-4 m2
Load acting at node, 1 = - 20 kN = - 20  103 N

[∵ Load is downward direction]

Load acting at node 1 = - 10 kN = - 10  103 N
[∵ Load is acting opposite to x direction]

Fig. (ii)
[Node: Number with circle denotes element & Number without circle denotes Node]
2.136 One Dimensional Problems

To find:
(i) Nodal displacements 𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢
(ii) Stress in each element, 𝜎 , 𝜎 𝑎𝑛𝑑 𝜎 .
Solution: Consider node 1 as the origin.
The co-ordinated of various nodes are given below:
𝑥 𝑦
Node 1 = (0, 0)
𝑥 𝑦
Node 2 = (0, 0)
𝑥 𝑦
Node 4 = (3, 0)
At node 3: 𝑥 = 3 × cos 45 = 2.121 𝑚
𝑦 = 3 × sin 45 = 2.121 𝑚
𝑥 𝑦
Node 4 = (2.121, 2.121)

For element (1): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (0 − 0) + (3 − 0)
𝑙 = 3 𝑚𝑚
𝑥 −𝑥 0−0
Direction consines, 𝑙 = =
𝑙 3
𝑙 =0
𝑦 −𝑦 3−0
𝑚 = =
𝑙 3
𝑚 =1

For element (2): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (2.121 − 0) + (2.121 − 0)
𝑙 = 2.9995 = 3𝑚
 Finite Element Analysis 2.137
𝑥 −𝑥 0−0
Direction consines, 𝑙 = =
𝑙 3

𝑙 =0

𝑦 −𝑦 3−0
𝑚 = =
𝑙 3

𝑚 =1

For element (2): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (2.121 − 0) + (2.121 − 0)

𝑙 = 2.9995 = 3 𝑚

𝑥 −𝑥 2.121 − 0
Direction consines, 𝑙 = =
𝑙 3

𝑙 = 0.707

𝑦 −𝑦 2.121 − 0
𝑚 = =
𝑙 3

𝑚 = 0.707

For element (3): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (3 − 0) + (0 − 0)

𝑙 = 3 𝑚𝑚

𝑥 −𝑥 3−0
Direction consines, 𝑙 = =
𝑙 3

𝑙 = 1 𝑚𝑚
𝑦 −𝑦
𝑚 =
𝑙
2.138 One Dimensional Problems
0−0
=
3

𝑚 =0

Fig. (iii)
For element (1): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟑 𝒂𝒏𝒅 𝒖𝟒
Stiffness matrix [ K ] for a truss element is given by,
𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
[𝐾] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

[From equation no. (2.89)]

0 0 0 0
4 × 10 × 201 × 10 0 1 0 −(1)
=
1000 0 0 0 0
0 0 0 1

1 2 3 4
0 0 0 0 𝟏
0 1 0 −1 𝟐
[𝐾] = 268 × 10 …(1)
0 0 0 0 𝟑
0 −1 0 1 𝟒
 Finite Element Analysis 2.139

For element (2): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔 :

𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

1 2 5 6
0.499 0.499 −0.499 0.499 1
0.499 0.499 −0.499 −0.499 2 …(2)
[𝐾] = 268 × 10
−0.499 −0.499 0.499 0.499 5
−0.499 −0.499 0.499 0.499 6

For element (3): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟕 𝒂𝒏𝒅 𝒖𝟖 :

𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦

(1) 0 −(1) 0
4 × 10−4 × 201 × 109 0 0 0 0
=
3 −(1) 0 1 0
0 0 0 0

1 2 7 8
0.64 0.48 −0.64 0.48 1
0.48 0.36 −0.48 −0.36 2 …(3)
[𝐾] = 268 × 10
−0.64 −0.48 0.64 0.48 7
−0.48 −0.36 0.48 0.36 8
2.140 One Dimensional Problems

Assemble the stiffness matrix [ K ], i.e., assemble the equation (1), (2) and (3).

1 2 3 4 5 6 7 8
0 0 -1 0
+ +
- -
0.499 0.499 0 0 1
0.499 0.499
+ +
1 0
0 1 0 0
+ +
- -
0.499 0.499 0 -1 2
[ K ] = 1  105
0.499 0.499
+ +
0 0
0 0 0 0 0 0 0 0 3
0 -1 0 1 0 0 0 0 4
- - 0 0
0 0 0.499 0.499 5
0.499 0.499
- - 0 0
0 0 0.499 0.499 6
0.499 0.499
-1 0 0 0 0 0 1 0 7
0 0 0 0 0 0 0 0 8

⇒ [𝐾]
1.499 0.499 0 0 −0.64 −0.499 −1 0
⎡ 0.499 1.499 0 −1 −0.48 −0.499 0 0 ⎤
⎢ 0 0 0 0 0 0 0 0 ⎥⎥
⎢
= 268 × 10 ⎢ 0 −1 0 1 0 0 0 0 ⎥ … (4)
⎢−0.499 −0.499 0 0 0.499 0.499 0 0⎥
⎢−0.499 −0.499 0 0 0.499 0.499 0 0⎥
⎢ −1 0 0 0 0 0 1 0⎥
⎣ 0 0 0 0 0 0 0 0⎦
 Finite Element Analysis 2.141

We know that, General finite element equation is

[K]{u}={F}
1.499 0.499 0 0 −0.64 −0.499 −1 0 𝑢
⎡ 0.499 1.499 0 −1 −0.48 −0.499 0 0⎤ ⎧𝑢 ⎫
⎢ 0 0⎥ ⎪ ⎪
⎢
0 0 0 0 0 0 ⎥ ⎪𝑢 ⎪
0 −1 0 1 0 0 0 0⎥ 𝑢
⇒ 268 × 10 ⎢ 𝑢
⎢−0.499 −0.499 0 0 0.499 0.499 0 0⎥ ⎨𝑢 ⎬
⎢−0.499 −0.499 0 0 0.499 0.499 0 0⎥ ⎪ ⎪
⎢ −1 0 0 0 0 0 1 0⎥ ⎪𝑢 ⎪
⎣ 0 0 0 0 0 0 0 0⎦ ⎩𝑢 ⎭
𝐹
⎧𝐹 ⎫
⎪𝐹 ⎪
⎪ ⎪
𝐹
= … (5)
⎨𝐹 ⎬
⎪𝐹 ⎪
⎪𝐹 ⎪
⎩𝐹 ⎭
Applying boundary conditions [Refer Fig. (iii)]:
(i) Node 2 is fixed. So, u3 = u4 = 0
(ii) Node 3 is fixed. So, u5 = u6 = 0.
(iii) Node 4 fixed. So, u7 = u8 = 0
(iv) At node 1, point load of 20 103N is acting in downward direction. So, F=-20
103 N.
(v) At node1, point load of 20 103N is acting opposite to x direction. So, F=-10 
103 N.
(vi) Self-weight is neglected. So, F3=F4=F5=F6=F7=F8=0.
Substitute the above boundary condition values in equation no.(5), u3 = u4 = u5 =
u6 = u7 = u8 = 0. So, delete third row third column, fourth row fourth column, fifth row
fifth column, sixth row sixth column, seventh row seventh column, and eighth row
eighth column of [K] matrix.
The final reduced equation is,
1.499 0.499 𝑢 𝐹
268 × 10 𝑢 =
0.499 1.499 𝐹
2.142 One Dimensional Problems
1.499 0.499 𝑢
⇒ 268 × 10 𝑢 = −10 × 10
0.499 1.499 −20 × 10

⇒ 268 × 10 (1.499 𝑢 + 0.499 𝑢 ) = −10 × 10 … (6)

268 × 10 (0.499 𝑢 + 1.499 𝑢 ) = −20 × 10 … (7)

Equation (7)  3.005 

268 × 10 (1.499 𝑢 + 4.504 𝑢 ) = −60.1 × 10 … (8)

Solving equation (6) and (8),

⇒ 268 × 10 (1.499 𝑢 + 0.499 𝑢 ) = −10 × 10

268 × 10 (1.499 𝑢 + 4.504 𝑢 ) = −60.1 × 10

268 × 10 (−4.005 𝑢 ) = −50.1 × 10

⇒ 𝑢 = −4.66 × 10 𝑚

Substitute 𝑢 value in equation (6).

268 × 10 (1.499 𝑢 + 0.499(−4.66 × 10 ) = −10 × 10

⇒ 𝑢 = −9.379 × 10 𝑚
We know that,
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚] 𝑢
𝑙
𝑢
For element (1):
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚 ] 𝑢
𝑙
𝑢
−9.379 × 10
201 × 10
= [0 −1 0 1 ] −4.66 × 10
1000 0
0
 Finite Element Analysis 2.143

−9.379 × 10
= 67 × 10 [0 −1 0 1 ] −4.66 × 10
0
0
= 67 × 10 [0 + 4.66 × 10 +0+0
𝜎 = 31.22 × 10 𝑁/𝑚
For element (2):
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚 ] 𝑢
𝑙
𝑢

−9.379 × 10
201 × 10
= [−7.07 −0.707 0.707 0.707 ] −4.66 × 10
1000 0
0

= 67 × 10 [−0.707 × (9.379 × 10 )
+ (−0.707) × (−4.66 × 10 ) + 0 + 0]

𝜎 = 26.52 × 10 𝑁/𝑚
For element (3):
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚 ] 𝑢
𝑙
𝑢
−9.379 × 10
201 × 10
= [−1 0 1 0 ] −4.66 × 10
1000 0
0
= 67 × 10 [−9.379 × 10 + (−1) + 0 + 0 + 0]
𝜎 = 6.28 × 10 𝑁/𝑚
Result: (i) Nodal displacements:

𝑢 = −9.379 × 10 𝑚
2.144 One Dimensional Problems

𝑢 = −4.66 × 10 𝑚
𝑢 =0
𝑢 =0
𝑢 =0
𝑢 =0
𝑢 =0
𝑢 =0
(ii) Stresses in each element:
𝜎 = 31.22 × 10 𝑁/𝑚
𝜎 = 26.52 × 10 𝑁/𝑚
𝜎 = 6.28 × 10 𝑁/𝑚

Example 2.24
For the three-bar truss shown in fig (i), determine the displacements of node 1
and the stress in element 3.

Fig (i)
Given:
Area of cross section, A of each member} = 250 mm2
Youn’s modulus, E = 200 GPa = 200  109 Pa = 200  109 N/m2
 Finite Element Analysis 2.145

= 200  103 N/m2

Area of each element, A = 4  10-4 m2

Fig. (ii)
To find:
1. Displacements of node 1.
2. Stress in each element, 3.
Solution: Consider node 1 as the origin.
The co-ordinated of various nodes are given below:
𝑥 𝑦
Node 1 = (0, 0)
𝑥 𝑦
Node 2 = (-450, 600)
𝑥 𝑦
Node 3 = (800, 600)
𝑥 𝑦
Node 4 = (450, 600)

For element (1): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (−450 − 0) + (600 − 0)
𝑙 = 750 𝑚𝑚
2.146 One Dimensional Problems
𝑥 −𝑥 −450 − 0
Direction consines, 𝑙 = =
𝑙 750

𝑙 = −0.6

𝑦 −𝑦 600 − 0
𝑚 = =
𝑙 750

𝑚 = 0.8

For element (2): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (800 − 0) + (600 − 0)

𝑙 = 1000 𝑚

𝑥 −𝑥 450 − 0
Direction consines, 𝑙 = =
𝑙 750

𝑙 = −0.6

𝑦 −𝑦 600 − 0
𝑚 = =
𝑙 3

𝑚 = 0.8

For element (2): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (800 − 0) + (600 − 0)

𝑙 = 1000 𝑚

𝑥 −𝑥 800 − 0
Direction consines, 𝑙 = =
𝑙 1000

𝑙 = 0.8

𝑦 −𝑦 600 − 0
𝑚 = =
𝑙 1000
 Finite Element Analysis 2.147

𝑚 = 0.6

For element (3): Length, 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (450 − 0) + (600 − 0)

𝑙 = 750 𝑚𝑚

𝑥 −𝑥 450 − 0
Direction consines, 𝑙 = =
𝑙 750

𝑙 = 0.6 𝑚𝑚

𝑦 −𝑦 600 − 0
𝑚 = =
𝑙 750

𝑚 = 0.8

Fig. (iii)
For element (1): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟑 𝒂𝒏𝒅 𝒖𝟒
Stiffness matrix [ K ] for a truss element is given by,
1 2 3 4
𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚 1
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥ 2
[𝐾] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥ 3
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦4
2.148 One Dimensional Problems
(−0.6) −0.6 × 0.8 −(−0.6) −0.6 × 0.8
250 × 200 × 10 0.6 × 0.8 (0.8) 0.6 × 0.8 −(0.8)
=
750 (−0.6) 0.6 × 0.8 (−0.6) −0.6 × 0.8
0.6 × 0.8 −(0.8) −0.6 × 0.8 (0.8)
1 2 3 4
24 −32 −24 32 1
−32 42.666 32 −42.666 2
[𝐾] = 1 × 10 …(1)
−24 32 24 −32 3
32 −42.666 −32 42.666 4

For element (2): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟓 𝒂𝒏𝒅 𝒖𝟔 :

1 2 5 6
𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚 1
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥ 2
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥ 5
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦6

(0.8)2 −0.6 × 0.8 −(0.8)2 −0.8 × 0.6 ⎤

3 ⎡
250 × 200 × 10 (0.6)2 −0.8 × 0.6 −(0.6)2 ⎥
= ⎢ 0.8 × 0.6
1000 ⎢ −(0.8)2 −0.8 × 0.6 (0.8)2 0.8 × 0.6⎥
⎣−0.8 × 0.6 −(0.6)2 −0.6 × 0.8 (0.6)2 ⎦

1 2 5 6
32 24 −32 −24 1
24 18 −24 −18 2
[𝐾] = 1 × 10 …(2)
−32 −24 32 24 5
−24 −18 24 18 6

For element (3): Displacements 𝒖𝟏 , 𝒖𝟐 , 𝒖𝟕 𝒂𝒏𝒅 𝒖𝟖 :

1 2 7 8
𝑙 𝑙 𝑚 −𝑙 −𝑙 𝑚 1
⎡ ⎤
𝐴 𝐸 ⎢𝑙 𝑚 𝑚 −𝑙 𝑚 −𝑚 ⎥ 2
Stiffness matrix [ K ] = ⎢ −𝑙
𝑙 ⎢ −𝑙 𝑚 𝑙 𝑙 𝑚 ⎥⎥ 7
⎣−𝑙 𝑚 −𝑚 𝑙 𝑚 𝑚 ⎦8
 Finite Element Analysis 2.149
(0.6)2 0.6 × 0.8 −(0.6)2 −0.6 × 0.8 ⎤
3 ⎡
250 × 200 × 10 (0.8)2 −0.6 × 0.8 −(0.8)2 ⎥
= ⎢ 0.6 × 0.8
750 2 (0.6)2
⎢ −(0.6) −0.6 × 0.8 0.6 × 0.8⎥
⎣−0.6 × 0.8 −(0.8)2 0.6 × 0.8 (0.8)2 ⎦

1 2 7 8
24 32 −24 −32 1
32 42.666 −32 −42.666 2
[𝐾] = 268 × 10 …(3)
−24 −32 24 32 7
−32 −42.666 32 42.666 8

Assemble the stiffness matrix [ K ], i.e., assemble the equation (1), (2) and (3).

1 2 3 4 5 6 7 8
24 -32
+ +
32 24 -24 32 -32 -24 -24 -32 1
+ +
24 32
-32 42.666
+ +
-
[ K ] = 1  105 24 18 32 -24 -18 -32 -42.66 2
42.666
+ +
32 42.666
-24 32 24 -32 0 0 0 0 3
-
32 -32 42.666 0 0 0 0 4
42.666
-32 -24 0 0 32 24 0 0 5
-24 -18 0 0 24 18 0 0 6
-24 -32 0 0 0 0 24 32 7
-
-32 0 0 0 0 32 42.666 8
42.666
2.150 One Dimensional Problems

80 24 −24 32 −32 −24 −24 −32

⎡ 24 103.332 32 −42.666 −24 −18 −32 −42.666 ⎤
⎢−24 32 24 −32 0 0 0 0 ⎥
⎢ ⎥
⇒ [𝐾] = 1 × 10 ⎢ 32 −42.666 −32 42.666 0 0 0 0 ⎥
⎢−32 −24 0 0 32 24 0 0 ⎥
⎢−24 −18 0 0 24 18 0 0 ⎥
⎢−24 −32 0 0 0 0 24 32 ⎥
⎣−32 −42.666 0 0 0 0 32 42.666 ⎦

We know that, General finite element equation is

[K]{u}={F}
80 24 −24 32 −32 −24 −24 −32 𝑢
⎡ 24 103.332 32 −42.666 −24 −18 −32 −42.666 ⎤ ⎧𝑢 ⎫
⎢−24 32 24 −32 0 0 0 0 ⎥ ⎪
⎪𝑢 ⎪
⎪
⎢ 32 −42.666 −32 42.666 0 0 0 0 ⎥ 𝑢
⇒ 1 × 10 ⎢−32 0 0 32 0 ⎥ 𝑢
−24 24 0 ⎨𝑢 ⎬
⎢−24 −18 0 0 24 18 0 0 ⎥
⎢−24 ⎥ ⎪
⎪𝑢 ⎪
⎪
−32 0 0 0 0 24 32
⎣−32 −42.666 0 0 0 0 32 42.666 ⎦ ⎩𝑢 ⎭
𝐹
⎧𝐹 ⎫
⎪𝐹 ⎪
⎪ ⎪
𝐹
= … (4)
𝐹
⎨ ⎬
⎪𝐹 ⎪
⎪𝐹 ⎪
⎩𝐹 ⎭
Applying boundary conditions [Refer Fig. (iii)]:
1. Node 2 is fixed. So, u3 = u4 = 0
2. Node 4 is fixed. So, u7 = u8 = 0.
3. Node 3 fixed. So, u5 = u6 = 0
4. At node 1, point load of 18 103N is acting in downward direction.
So, F=-18103 N.
5. Self-weight is neglected. So, F1=F3=F4=F5=F6=F7=F8=0.
Substitute the above values in equation (4), 𝑢 = 𝑢 = 𝑢 = 𝑢 = 𝑢 = 𝑢 =
0 delete third row third column, fourth row fourth column, fifth row fifth column, sixth
 Finite Element Analysis 2.151

row sixth column, seventh row seventh column and eighth row eighth column of [K]
matrix. Hence the final reduced equation is,
80 24 𝑢 𝐹
1 × 10 =
24 103.332 𝑢 𝐹
80 24 𝑢 0
⇒ 1 × 10 𝑢 =
24 103.332 −18 × 10
⇒ 80 𝑢 + 24 𝑢 = 0 … (5)
24 𝑢 + 103.332 𝑢 = 0 … (6)
Equation (6)  3.3334 
⇒ 80 𝑢 + 344.44 𝑢 = −60 … (7)
80 𝑢 + 24 𝑢 = 0 … (5)
Solving, 320.44 𝑢 = 0
⇒ 𝑢 = −0.187 𝑚𝑚
Substitute 𝑢 value in equation (5).
⇒ 80 𝑢 + 24 (−0.187 = −0
⇒ 𝑢 = 0.0561 𝑚𝑚
We know that,
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚] 𝑢
𝑙
𝑢
For element (3):
𝑢
𝐸 𝑢
𝑆𝑡𝑟𝑒𝑠𝑠, 𝜎 = [−𝑙 −𝑚 𝑙 𝑚 ] 𝑢
𝑙
𝑢
0.561
200 × 10
= [−0.6 −0.8 0.6 0.8 ] −0.187
750 0
0
= 266.6667[−0.6 × 0.561 + (−0.8) × (−0.187) + 0 + 0]
2.152 One Dimensional Problems

𝜎 = −49.86 𝑁/𝑚
Result: (i) Displacements of node 1:

𝑢 = 0.561 𝑚𝑚
𝑢 = −0.187 𝑚𝑚
(ii) Stresses in element(3):
𝜎 = −49.86𝑁/𝑚

2.12. BENDING Of BEAMS

2.12.1. Introduction
Beam is a structural member which is supported along the
length and subjected to external forces or loads acting
transversely i.e., perpendicular to the centre line. Beam is
sufficiently long when compared to the lateral dimensions.

2.12.2. Types of Beams

The following are the important types of beams.
(a) Cantilever beam
(b) Simply supported beam
(c) Overhanging beam
(d) Fixed beam and
(e) Continuous beam
(a) Cantilever Beam
A beam with one end free and the other end fixed is
called cantilever beam. Refer Fig.2.29(a).
(b) Simply Supported Beam (SSB)
A beam supported or resting freely on the supports at its both ends is called SSB.
It is shown in Fig.2.29(b).
 Finite Element Analysis 2.153

(c) Overhanging Beam

If the one or both the end portions are extended beyond the support, then it is
called overhanging beam. Refer Fig.2.29(c).
(d) Fixed Beam
A beam whose both ends are fixed or built into the walls, is called a fixed
beam. Refer Fig.2.29(d).
(e) Continues beam
A beam which has more than two supports is called continuous beam. Refer
Fig.2.29( e)

2.12.3. Transverse Loading on Beams

A load which is acting vertically downward on the
horizontal beam is called transverse load.

2.13.4. Types of Transverse Load

A beam may be subjected to the following types of loads.
(a) Point or concentrated load.
(b) Uniformly Distributed Load (UDL)
(c) uniformly Varying Load (UVL)
(a) Point or Concentrated Load
A load (W) which is acting at a particular point is
called point load. Refer Fig.2.30(a).
(b) Uniformly Distributed Load (UDL)
A load which is spread over a beam in such a manner
that the rate of loading 'w' is uniform throughout the length.
Refer Fig.2.30(b).
(c) Uniformly Varying Load (UVL)
A load which is spread over a beam in such a manner that the rate of loading
uniformly varies from point to point along the beam. The load is zero at one end and
increases uniformly to the other end. It is also called as triangular load. Refer
Fig.2.30(c).
2.154 One Dimensional Problems

2.12.5. Derivation of shape Function for Beam Element [Fourth Order

Beam Equation]
Consider the beam element as shown in Fig.2.31. The beam is of length L with
axial local co-ordinate x and transverse local co-ordinate y. The local transverse nodal
displacements are given by dly and d2y' The rotations are given by $1 and $2' The local
nodal forces are given by Fly and F 2y- The bending moments are given by ml and m.

Fig. 2.31. Beam element with positive nodal displacements, rotations, forces, and
moments
At all nodes, the following sign conversions are used.
(i) Moments are positive in the counterclockwise direction.
(ii) Rotations are positive in the counterclockwise direction.
(iii) Forces are positive in the positive y direction
(iv) Displacements are positive in the positive y direction.
Fig.2.32 indicates the sign conventions used in simple beam theory for
positive shear forces F and bending moments m

Fig. 2.32 Beam theory sign conventions for shear forces and bending moments
Assume the transverse displacement variation through the element length to be
𝑣(𝑥) = 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 … (2.100)
 Finite Element Analysis 2.155

We express v in terms of the nodal degrees of freedom 𝑑 , 𝑑 ϕ 𝑎𝑛𝑑 ϕ as

follows:
At x = 0,
𝑣(0) = 𝑎 = 𝑑 … (2.101)
𝑑𝑣( )
= 3𝑎 𝑥 + 2𝑎 𝑥 + 𝑎
𝑑𝑥
𝑑𝑣( )
=𝑎 =ϕ … (2.102)
𝑑𝑥
When x = L,
𝑣(𝐿) = 𝑎 𝐿 + 𝑎 𝐿 + 𝑎 𝐿 + 𝑎 = 𝑑 … (2.103)
𝑑𝑣( )
= 3𝑎 𝐿 + 2𝑎 𝐿 + 𝑎 = ϕ … (2.104)
𝑑𝑥
Finding 𝑎 and 𝑎 in terms of 𝑑 , 𝑑 ϕ 𝑎𝑛𝑑 ϕ by using the above equations
(2.101), (2.102), (2.103) and (2.104).
(2.103) ⇒ 𝑑 =𝑎 𝐿 +𝑎 𝐿 +𝑎 𝐿+𝑎
=𝑎 𝐿 +𝑎 𝐿 +𝑎 𝐿+𝑑 [∵ 𝑎 = 𝑑 ]
⇒ (𝑑 −𝑑 )=𝑎 𝐿 +𝑎 𝐿 +𝑎 𝐿+ϕ 𝐿
⇒ (𝑑 −𝑑 −ϕ 𝐿 )=𝑎 𝐿 +𝑎 𝐿
1
⇒ 𝑑 −𝑑 −ϕ 𝐿 =𝑎 𝐿 +𝑎 𝐿 … (2.105)
𝐿
(2.104) ⇒ ϕ = 3𝑎 𝐿 + 2𝑎 𝐿 + 𝑎
= 3𝑎 𝐿 + 2𝑎 𝐿 + ϕ [∵ 𝑎 = ϕ ]
⇒ ϕ − ϕ = 3𝑎 𝐿 + 2𝑎 𝐿 … (2.106)
Equation (2.105)  3
3
⇒ 𝑑 −𝑑 − ϕ 𝐿 = 3𝑎 𝐿 + 3𝑎 𝐿 … (2.107)
𝐿
Solving equation (2.106) and (2.107)
ϕ − ϕ = 3𝑎 𝐿 + 2𝑎 𝐿
2.156 One Dimensional Problems
3
𝑑 −𝑑 − ∅ 𝐿 = 3𝑎 𝐿 + 3𝑎 𝐿
𝐿
3
Subtracting, ϕ −ϕ − 𝑑 − 𝑑 − ∅ 𝐿 = −𝑎 𝐿
𝐿
3 3
ϕ −ϕ − 𝑑 −𝑑 + ∅ 𝐿 = −𝑎 𝐿
𝐿 𝐿
3
ϕ −ϕ − 𝑑 −𝑑 + 3∅ = −𝑎 𝐿
𝐿
3
ϕ − 2ϕ − 𝑑 −𝑑 = −𝑎 𝐿
𝐿
1 3
(ϕ + 2ϕ ) − 𝑑 −𝑑 = −𝑎
𝐿 𝐿
1 3
⇒ (ϕ + 2ϕ ) + 𝑑 −𝑑 = −𝑎
𝐿 𝐿
−1 3
⇒ (ϕ + 2ϕ ) − 𝑑 −𝑑 =𝑎
𝐿 𝐿
−3 1
⇒ 𝑎 = 𝑑 −𝑑 − (2ϕ + ϕ )
𝐿 𝐿
…(2.108)
Substitute 𝑎 value in equation (2.104)
−3 1
⇒ ϕ = 3𝑎 𝐿 + 2𝐿 𝑑 −𝑑 − (2ϕ + ϕ ) + 𝑎
𝐿 𝐿
6
= 3𝑎 𝐿 − 𝑑 −𝑑 − 2(2ϕ + ϕ ) + ϕ
𝐿
[∵ 𝑎 = ϕ ]
6
⇒ ϕ − ϕ = 3𝑎 𝐿 − 𝑑 −𝑑 − 4ϕ − 2ϕ
𝐿
6
⇒ 3ϕ + 3ϕ = 3𝑎 𝐿 − 𝑑 −𝑑
𝐿
6
⇒ 3𝑎 𝐿 = 3ϕ + 3ϕ + 𝑑 −𝑑
𝐿
 Finite Element Analysis 2.157
2
𝑎 𝐿 = ϕ + 3ϕ + 𝑑 −𝑑
𝐿
1 2
⇒ 𝑎 = (ϕ + ϕ ) + 𝑑 −𝑑
𝐿 𝐿
2 1
⇒ 𝑎 = 𝑑 −𝑑 + (ϕ + ϕ ) … (2.109)
𝐿 𝐿
Substitute 𝑎 , 𝑎 , 𝑎 𝑎𝑛𝑑 𝑎 value in equation (2.100)
2 1
𝑣(𝑥) = 𝑑 −𝑑 + (ϕ + ϕ ) 𝑥
𝐿 𝐿
−3 1
+ 𝑑 −𝑑 − (2ϕ + ϕ ) 𝑥 + ϕ 𝑥
𝐿 𝐿
+𝑑 … (2.110)

∵ 𝑎 = ϕ ;𝑎 = 𝑑
In matrix form, v (x) = [ N ] { d }
𝑑
⎧ ⎫
ϕ
⇒ 𝑣(𝑥) = [𝑁 𝑁 𝑁 𝑁]
⎨𝑑 ⎬
⎩ϕ ⎭

⇒ 𝑣(𝑥) = 𝑁 𝑑 +𝑁 ϕ +𝑁 𝑑 +𝑁 ϕ … (2.111)
Where N1, N2, N3 and N4 are shape functions for beam element.
1
𝑁 = (2𝑥 − 3𝑥 𝐿 + 𝐿 ) ⎫
𝐿 ⎪
1
𝑁 = (𝑥 𝐿 − 2𝑥 𝐿 + 𝑥𝐿 )⎪
𝐿 … (2.112)
1 ⎬
𝑁 = (−2𝑥 + 3𝑥 𝐿)
𝐿 ⎪
1 ⎪
𝑁 = (𝑥 𝐿 − 𝑥 𝐿 ) ⎭
𝐿
Verification: we know
𝑣(𝑥) = 𝑁 𝑑 +𝑁 ϕ +𝑁 𝑑 +𝑁 ϕ … (2.111)
2.158 One Dimensional Problems

Substituting N1, N2, N3 and N4 values,

1 1
⇒ 𝑣(𝑥) = (2𝑥 − 3𝑥 𝐿 + 𝐿 )𝑑 + (𝑥 𝐿 − 2𝑥 𝐿 + 𝑥𝐿 )ϕ
𝐿 𝐿
1 1
+ (−2𝑥 + 3𝑥 𝐿)𝑑 + (𝑥 𝐿 − 𝑥 𝐿 )ϕ
𝐿 𝐿
2𝑥 3𝑥 𝐿 𝑥 𝐿 2𝑥 𝐿
⇒ 𝑣(𝑥) = 𝑑 − 𝑑 +𝑑 + ϕ − ϕ + 𝑥ϕ
𝐿 𝐿 𝐿 𝐿
2𝑥 3𝑥 𝐿 𝑥 𝐿ϕ 𝑥 𝐿
− 𝑑 + 𝑑 + − ϕ
𝐿 𝐿 𝐿 𝐿
2𝑥 3𝑥 𝑥 2𝑥
𝑣(𝑥) = 𝑑 − 𝑑 +𝑑 + ϕ − ϕ + 𝑥ϕ
𝐿 𝐿 𝐿 𝐿
2𝑥 3𝑥 𝑥 ϕ 𝑥 ϕ
− 𝑑 + 𝑑 + −
𝐿 𝐿 𝐿 𝐿
2 1
⇒ 𝑣(𝑥) = 𝑑 −𝑑 + (ϕ + ϕ ) 𝑥
𝐿 𝐿
−3 1
+ 𝑑 −𝑑 − (2ϕ + ϕ ) 𝑥 + ϕ 𝑥 + 𝑑
𝐿 𝐿
[Same as equation no. (2.110)]

2.12.6. Stiffness Matrix [K] for beam element

The stiffness matrix [k] for beam element is derived by using a direct
equilibrium approach and beam theory sign conversions.
We know that,
Transverse displacement
2 1
𝑣(𝑥) = 𝑑 −𝑑 + (ϕ + ϕ ) 𝑥
𝐿 𝐿
−3 1
+ 𝑑 −𝑑 − (2ϕ + ϕ ) 𝑥 + ϕ 𝑥
𝐿 𝐿
+𝑑
𝑑𝑣( ) 2 1
⇒ = 3𝑥 𝑑 −𝑑 + (ϕ + ϕ )
𝑑𝑥 𝐿 𝐿
−3 1
+ 2𝑥 𝑑 −𝑑 − (2ϕ + ϕ ) + ϕ
𝐿 𝐿
 Finite Element Analysis 2.159

𝑑 𝑣( ) 2 1
⇒ = 6𝑥 𝑑 −𝑑 + (ϕ + ϕ )
𝑑𝑥 𝐿 𝐿
−3 1
+2 𝑑 −𝑑 − (2ϕ + ϕ ) … (2.13)
𝐿 𝐿
𝑑 𝑣( ) 2 1
⇒ =6 𝑑 −𝑑 + (ϕ + ϕ ) … (2.14)
𝑑𝑥 𝐿 𝐿
Put x = 0 in equation (2.113)
𝑑 𝑣( ) −3 1
⇒ = 0+2 𝑑 −𝑑 − (2ϕ + ϕ )
𝑑𝑥 𝐿 𝐿
−6 2
= 𝑑 −𝑑 − (2ϕ + ϕ )
𝐿 𝐿
1
= −6𝐿𝑑 + 6𝐿𝑑 − 4𝐿 ϕ − 2𝐿 ϕ … (2.115)
𝐿
Put x = L in equation (2.113)
𝑑 𝑣(𝐿) 2 1
⇒ = 6𝐿 𝑑 −𝑑 + (ϕ + ϕ )
𝑑𝑥 𝐿 𝐿
−3 1
+2 𝑑 −𝑑 − (2ϕ + ϕ )
𝐿 𝐿
12𝐿 6𝐿 6
= 𝑑 −𝑑 + (ϕ + ϕ ) − 𝑑 −𝑑
𝐿 𝐿 𝐿
2
− (2ϕ + ϕ )
𝐿
1
= 12𝐿𝑑 − 12𝐿𝑑 + 6𝐿 ϕ + 6𝐿 ϕ − 6𝐿𝑑 + 6𝐿𝑑
𝐿
− 4𝐿 ϕ − 2𝐿 ϕ
𝑑 𝑣(𝐿) 1
= 6𝐿𝑑 + 2𝐿𝑑 − 6𝐿𝑑 + 4𝐿 ϕ … (2.116)
𝑑𝑥 𝐿
Put x = 0 in equation (2.114)
𝑑 𝑣(0) 2 1
⇒ =6 𝑑 −𝑑 + (ϕ + ϕ )
𝑑𝑥 𝐿 𝐿
1
= 12𝑑 − 12𝑑 − 6𝐿ϕ + 4𝐿 ϕ
𝐿
2.160 One Dimensional Problems

𝑑 𝑣(0) 1
= 12𝑑 + 6𝐿ϕ − 12𝑑 + 6𝐿ϕ … (2.117)
𝑑𝑥 𝐿
Put x = L in equation (2.113)
𝑑 𝑣(𝐿) 2 1
⇒ =6 𝑑 −𝑑 + (ϕ + ϕ )
𝑑𝑥 𝐿 𝐿
1
= 12𝑑 − 12𝑑 + 6𝐿ϕ − 6𝐿 ϕ
𝐿
𝑑 𝑣(𝐿) 1
= 12𝑑 + 6𝐿ϕ − 12𝑑 + 6𝐿ϕ … (2.118)
𝑑𝑥 𝐿
We know that,
𝑑 𝑣(0)
Nodal force, 𝐹 = 𝐸𝐼
𝑑𝑥
𝐸𝐼
⇒ 𝐹 = 12𝑑 + 6𝐿ϕ − 12𝑑 + 6𝐿ϕ
𝐿
[From equation no. (2.117)]
𝑑 𝑣(0)
Bending moment, 𝑚 = −𝐸𝐼
𝑑𝑥
𝐸𝐼
=− −6𝐿𝑑 − 4𝐿 ϕ + 6𝐿𝑑 − 2𝐿 ϕ
𝐿
[From equation no. (2.115)]
𝐸𝐼
𝑚 = 6𝐿𝑑 + 4𝐿 ϕ − 6𝐿𝑑 + 2𝐿 ϕ
𝐿
𝑑 𝑣(0)
Nodal force, 𝐹 = −𝐸𝐼
𝑑𝑥
−𝐸𝐼
⇒ 𝐹 = 12𝑑 + 6𝐿ϕ − 12𝑑 + 6𝐿ϕ
𝐿
[From equation no. (2.118)]
𝐸𝐼
⇒ 𝐹 = −12𝑑 − 6𝐿ϕ + 12𝑑 − 6𝐿ϕ
𝐿
𝑑 𝑣(𝐿)
Bending moment, 𝑚 = −𝐸𝐼
𝑑𝑥
 Finite Element Analysis 2.161
𝐸𝐼
⇒ 𝑚 = 6𝐿𝑑 + 2𝐿 ϕ − 6𝐿𝑑 + 4𝐿 ϕ
𝐿
[From equation no. (2.116)]
Arranging to the above equation 𝐹 , 𝑚 , 𝐹 , 𝑚 in matrix form,

𝐹 12 6𝐿 −12 6𝐿 ⎧𝑑 ⎫
𝑚 𝐸𝐼 6𝐿 4𝐿 −6𝐿 2𝐿 ϕ
⇒ = … (2.120)
𝐹 𝐿 −12 −6𝐿 12 −6𝐿 ⎨𝑑 ⎬
𝑚 6𝐿 2𝐿 −6𝐿 4𝐿 ⎩ ϕ ⎭
12 6𝐿 −12 6𝐿
[K] = 6𝐿 4𝐿 −6𝐿 2𝐿
−12 −6𝐿 12 −6𝐿
6𝐿 2𝐿 −6𝐿 4𝐿
Where E = Young’s modulus
I = Moment of inertia
L = Length of the beam

2.12.7. Nodal Forces and Bending Moments

The nodal forces (F1y and F2y) and bending moments (m1 and m2) for different types of
loads on fixed beams are given in Table 2.1

Loading case 𝑭𝟏𝒚 𝒎𝟏 𝑭𝟐𝒚 𝒎𝟐

1.
𝑊 𝑊𝐿 𝑊 𝑊𝐿
− − −
2 8 2 8

2.
−𝑊𝑏 (𝐿 + 2𝑎) −𝑊𝑎𝑏 −𝑊𝑏 (𝐿 + 2𝑎) 𝑊𝑎 𝑏
𝐿 𝐿 𝐿 𝐿

3.
−𝛼(1 𝛼(1
-W -W
− 𝛼)𝑊𝐿 − 𝛼)𝑊𝐿
2.162 One Dimensional Problems

4.

−𝑤𝐿 −𝑤𝐿 𝑤𝐿 𝑤𝐿
2 12 2 12

−7𝑤𝐿 −𝑤𝐿 −3𝑤𝐿 𝑤𝐿

5.
20 20 20 30

6. −𝑤𝐿 −5𝑤𝐿 −𝑤𝐿 5𝑤𝐿

4 96 4 96
7. −13𝑤𝐿 −11𝑤𝐿 −3𝑤𝐿 5𝑤𝐿
32 192 32 192

2.13.8. Solved Problems

EXAMPLE 2.25
A fixed beam of length 2 L m carries a uniformly distributed load of w (N/m) which
run over a length of L m from the fixed end, as shown in Fig.(i). Calculate the rotation at point B.

Fig.(i)
Given:

Fig.(ii)
To find: Rotation at point B, 2.
Solution: we can divide the beam into two elements as shown in Fig. (iii)
 Finite Element Analysis 2.163

Fig.(iii)
For element (1) (Nodes 1, 2, 3, 4 i.e., 𝑑 , ϕ , 𝑑 , ϕ

Fig.(iv)
Finite element equation is

12 6𝐿 −12 6𝐿 ⎧𝑑 ⎫ 𝐹
𝐸𝐼 6𝐿 4𝐿 −6𝐿 2𝐿 ϕ 𝑚
= … (1)
𝐿 −12 −6𝐿 12 −6𝐿 ⎨𝐹 ⎬ 𝐹
6𝐿 2𝐿 −6𝐿 4𝐿 ⎩ ϕ ⎭ 𝑚

For uniformly distributed load,

−𝑤𝐿 −𝑤𝐿 −𝑤𝐿 𝑤𝐿
𝐹 = ; 𝑚 = ;𝐹 = ; 𝑚 =
2 12 2 12
[Refer table 2.1]
1 2 3 4
−𝑤𝐿
⎧ ⎫
⎪ 2 ⎪
12 6𝐿 −12 6𝐿 1 ⎧𝑑 ⎫ ⎪
⎪−𝑤𝐿 ⎪
⎪
𝐸𝐼 6𝐿 4𝐿 −6𝐿 2𝐿 2 ϕ 12
(1) ⇒ = … (2)
𝐿 −12 −6𝐿 12 −6𝐿 3 ⎨𝐹 ⎬ ⎨ − 𝑤𝐿 ⎬
6𝐿 2𝐿 −6𝐿 4𝐿 4⎩ϕ ⎭ ⎪ 2 ⎪
⎪ ⎪
⎪ 𝑤𝐿 ⎪
⎩ 12 ⎭
2.164 One Dimensional Problems

For element (2) (Nodes 3, 4, 5, 6 i.e., 𝑑 , ϕ , 𝑑 , ϕ

Fig( iv)
Finite element equation is

12 6𝐿 −12 6𝐿 ⎧𝑑 ⎫ 𝐹
𝐸𝐼 6𝐿 4𝐿 −6𝐿 2𝐿 ϕ 𝑚
= … (3)
𝐿 −12 −6𝐿 12 −6𝐿 ⎨ 𝐹 ⎬ 𝐹
6𝐿 2𝐿 −6𝐿 4𝐿 ⎩ ϕ ⎭ 𝑚

There is no load and moment on element (2). So, 𝐹 =𝐹 = 0 and 𝑚 = 𝑚 = 0.

12 6𝐿 −12 6𝐿 ⎧𝑑 ⎫ 0
6𝐿 4𝐿 −6𝐿 2𝐿 ϕ
= 0 … (4)
−12 −6𝐿 12 −6𝐿 ⎨ 𝐹 ⎬ 0
6𝐿 2𝐿 −6𝐿 4𝐿 ⎩ ϕ ⎭ 0
Assemble the finite elements, i.e., assemble the equations (2) and (4),
1 2 3 4 5 6
6𝐿 𝑑
12 6𝐿 −12 0 0 1⎧ ⎫
⎡ 6𝐿 4𝐿 −6𝐿 2𝐿 ⎤ ϕ ⎪
0 0 2⎪ ⎪
𝐸𝐼 ⎢−12 −6𝐿 12 + 12 −6𝐿 + 6𝐿 −12 6𝐿 ⎥ 3 𝑑 ⎪
⎢ ⎥
𝐿 ⎢ 6𝐿 2𝐿 −6𝐿 + 6𝐿 4𝐿 + 4𝐿 −6𝐿 2𝐿 ⎥ 4 ⎨ ϕ ⎬
⎢ 0 0 −12 −6𝐿 12 −6𝐿⎥ 5 ⎪ ⎪𝑑 𝑦⎪ ⎪
⎣ 0 0 6𝐿 2𝐿 −6𝐿 4𝐿 ⎦ 6 ⎩ ϕ ⎭
−𝑤𝐿
⎧ ⎫
⎪ 2 ⎪
⎪ −𝑤𝐿 ⎪ 1
⎪ 12 ⎪ 2
3
= −𝑤𝐿 + 0 4
⎨ 2 ⎬
⎪ 𝑤𝐿 ⎪5
⎪ 12 + 0 ⎪ 6
⎪ 0 ⎪
⎩ 0 ⎭
 Finite Element Analysis 2.165
−𝑤𝐿
⎧ ⎫
⎪ 2 ⎪
𝑑
12
⎡ 6𝐿
6𝐿 −12 6𝐿 0 0 ⎧
⎤ ⎫ ⎪−𝑤𝐿 ⎪
4𝐿 −6𝐿 2𝐿 0 0 ⎪ ϕ ⎪
𝐸𝐼 ⎢−12 −6𝐿 ⎪ ⎪ ⎪ 12 ⎪
24 0 −12 6𝐿 ⎥⎥ 𝑑
⇒ ⎢ 6𝐿 0 = −𝑤𝐿 … (5)
𝐿 ⎢ 2𝐿 8𝐿 −6𝐿 2𝐿 ⎥ ⎨ ϕ ⎬ ⎨ 2 ⎬
⎢ 0 0 −12 −6𝐿 12 −6𝐿⎥ ⎪⎪𝑑 𝑦⎪
⎪ ⎪ 𝑤𝐿 ⎪
⎣ 0 0 6𝐿 2𝐿 −6𝐿 4𝐿 ⎦ ⎩ ϕ ⎭ ⎪
12 ⎪
⎪ 0 ⎪
⎩ 0 ⎭
Applying boundary conditions (Refer Fig. (ii)
(i) A is fixed. So, displacement 𝑑 and rotation ϕ are zero.
(ii) At B, displacement 𝑑 =0
(iii) C is fixed. So, displacement 𝑑 and rotation ϕ = 0
Substitute the above values in equation (5)
−𝑤𝐿
⎧ ⎫
⎪ 2 ⎪
12
⎡ 6𝐿
6𝐿 −12 6𝐿 0 0
⎤ ⎧
0
⎫ ⎪−𝑤𝐿 ⎪
4𝐿 −6𝐿 2𝐿 0 0 ⎪0⎪ ⎪ 12 ⎪
𝐸𝐼 ⎢−12 −6𝐿 24 0 −12 6𝐿 ⎥⎥ 0
⇒ ⎢ 6𝐿 0 = −𝑤𝐿
𝐿 ⎢ 2𝐿 8𝐿 −6𝐿 2𝐿 ⎥ ⎨0⎬ ⎨ 2 ⎬
⎢ 0 0 −12 −6𝐿 12 −6𝐿⎥ ⎪0⎪ ⎪ 𝑤𝐿 ⎪
⎣ 0 0 6𝐿 2𝐿 −6𝐿 4𝐿 ⎦ ⎩0⎭ ⎪
12 ⎪
⎪ 0 ⎪
⎩ 0 ⎭
In the above equation 𝑑 =ϕ =𝑑 = 𝑑 𝑦 = ϕ = 0.
So, delete first row first column, second row second column, third row third
column, fifth row fifth column and sixth row sixth column of [K] matrix. Hence the
equation reduces to
𝐸𝐼 𝑤𝐿
[8𝐿 ]ϕ =
𝐿 12
𝑤𝐿
⇒ ϕ =
96 𝐸𝐼
2.166 One Dimensional Problems

𝑤𝐿
Slope or rotation, ϕ =
96 𝐸𝐼
𝑤𝐿
Result: Slope or rotation at B, ϕ =
96 𝐸𝐼

Example 2.26
For the beam and loading shown in fig (i), calculate the rotations at B and C.

Fig. (i)
E = 210 GPa; I = 6  106 mm4
Given:

Fig. (ii)
Young’s modulus, E = 210 GPa
= 210  109 Pa
E = 210  109 N/m2
Moment of inertia, I = 6  106 mm4
I = 6  10-6 m4
For element 1, Length, L = 1 m
For element 2, Length, L = 1 m
 Finite Element Analysis 2.167

Uniformly distributed load, w = 12 kN/m

w = 12  103 N/m
To find:
1. Slope or rotation at B, ϕ
2. Slope or rotation at C, ϕ
Solution: we can divide the beam into two elements as shown in Fig. (iii).

Fig. (iii)
For element (1) (Nodes 1, 2, 3, 4 i.e., 𝑑 , ϕ , 𝑑 , ϕ )

Fig.(iv)
Finite element equation is

12 6𝐿 −12 6𝐿 ⎧𝑑 ⎫ 𝐹
𝐸𝐼 6𝐿 4𝐿 −6𝐿 2𝐿 ϕ 𝑚
= … (1)
𝐿 −12 −6𝐿 12 −6𝐿 ⎨𝑑 ⎬ 𝐹
6𝐿 2𝐿 −6𝐿 4𝐿 ⎩ ϕ ⎭ 𝑚

There is no load and moment on element (2). So, 𝐹 =𝐹 = 0 and 𝑚 = 𝑚 = 0.

2.168 One Dimensional Problems

1 2 3 4
12 6 −12 6 1 ⎧𝑑 ⎫
210 × 10 × 6 × 10 6 4 −6 2 2 ϕ
(1) ⇒
(1) −12 −6 12 −6 3 ⎨𝑑 ⎬
6𝐿 2 −6 4 4⎩ϕ ⎭
0 1
= 0 2 … (2) [∵ 𝐿 = 1𝑚]
0 3
0 4

For uniformly distributed load,

−𝑤𝐿 −𝑤𝐿 −𝑤𝐿 𝑤𝐿
𝐹 = ; 𝑚 = ;𝐹 = ; 𝑚 = [Refer Table 2.1]
2 12 2 12

−12 × 10 × 1
⇒𝐹 = = −6000𝑁 [∵ 𝑤 = 12 × 10 𝑁/𝑚; 𝐿 = 1𝑚]
2

−12 × 10^3 × (1)

⇒𝑚 = = −1000𝑁 − 𝑚]
12

−12 × 10 × 1
⇒𝐹 = = −6000𝑁 − 𝑚]
12

12 × 10 × (1)
⇒𝑚 = = 1000𝑁 − 𝑚]
12

3 4 5 6
12 6 −12 6 3 ⎧𝑑 ⎫
210 × 10 × 6 × 10 6 4 −6 2 4 ϕ
(3) ⇒
(1) −12 −6 12 −6 5 ⎨𝑑 ⎬
6𝐿 2 −6 4 6⎩ϕ ⎭
−6000 3
= −1000 4 … (4)
−6000 5
1000 6
 Finite Element Analysis 2.169

Assemble the finite elements i.e., assemble the finite element equations (2) and (4)
1 2 3 4 5 6
𝑑
12 6 −12 6 0 0 1⎧ ⎫
⎡ 6 4 −6 2 0 ⎤
0 2⎪ ϕ ⎪
⎢−12 ⎪ ⎪
210 × 10 × 6 × 10
⎢ −6 12 + 12 −6 + 6 −12 6 ⎥⎥ 3 𝑑
(1) ⎢ 6 2 −6 + 6 4+4 −6 2 ⎥4⎨ ϕ ⎬
⎢ 0 0 −12 −6 12 −6⎥ 5 ⎪
⎪𝑑 𝑦⎪
⎪
⎣ 0 0 6 2 −6 4 ⎦6⎩ ϕ ⎭
0 1
⎧ ⎫
⎪ 0 ⎪23
= 0 − 6000
⎨0 − 1000 ⎬4
⎪ −6000 ⎪ 5
⎩ 1000 ⎭ 6
𝑑
12 6 −12 6 0 0 ⎧ ⎫ 0
⎡ 6 4 −6 2 0 ⎤
0 ⎪ ϕ ⎪ ⎧ ⎫
⎢−12 ⎥ ⎪ ⎪ ⎪ 0 ⎪
−6 24 0 −12 6 ⎥ 𝑑
1.26 × 10 ⎢ 6 0 8 = −6000 … (5)
⎢ 2 −6 2 ⎥ ⎨ ϕ ⎬ ⎨−1000⎬
⎢ 0 0 −12 −6 12 −6⎥ ⎪ ⎪ ⎪
⎪𝑑 𝑦⎪ ⎩
−6000⎪
⎣ 0 0 6 2 −6 ⎦
4 ⎩ϕ ⎭ 1000 ⎭

Applying boundary conditions (Refer Fig. (ii)

(i) A is fixed. So, displacement 𝑑 and rotation ϕ are zero.
(ii) At B, displacement 𝑑 = 0
(iii) At C, vertical displacement 𝑑 =0
Substitute the above values in equation (5)
12 6 −12 6 0 0 0 0
⎡ 6 ⎤ ⎧0 ⎫ ⎧
4 −6 2 0 0 ⎪ ⎪ ⎪ 0 ⎫⎪
𝐸𝐼 ⎢−12 −6 24 0 −12 6 ⎥⎥ 0
⇒ ⎢ 6 = −6000
𝐿 ⎢ 2 0 8 −6 2 ⎥ ⎨ϕ ⎬ ⎨−1000⎬
⎢ 0 0 −12 −6 12 −6⎥ ⎪ 0 ⎪ ⎪−6000⎪
⎣ 0 0 6 2 −6 4 ⎦ ⎩ϕ ⎭ ⎩ 1000 ⎭
In the above equation 𝑑 =ϕ =𝑑 = 𝑑 𝑦 = 0.
So, delete first row first column, second row second column, third row third
column, fifth row fifth column of [K] matrix. Hence the equation reduces to
2.170 One Dimensional Problems

8 2 ϕ −1000
1.26 × 10 =
2 4 ϕ 1000
Matrix multiplication,
⇒ 1.26 × 10 [8ϕ + 2ϕ ] = −1000 … (6)
⇒ 1.26 × 10 [2ϕ + 4ϕ ] = 1000 … (7)
Equation (7)  (-4) 
⇒ 1.26 × 10 [−8ϕ − 16ϕ ] = −4000 … (8)
Equation (6))  1.26 × 10 [14ϕ ] = −1000
Solving, 1.26 × 10 [−14ϕ ] = −1000
⇒ ϕ = 2.834 × 10 rad
Substitute ϕ value in equation (6).
⇒ 1.26 × 10 [−8ϕ + 2(2.834 × 10 )] = −1000
⇒ ϕ = −1.70 × 10 rad
Result: (i) Slope or rotation at B, ϕ = −1.70 × 10 rad
(ii) Slope or rotation at C, ϕ = 2.834 × 10 rad

Example 2.27
Find the deflection at the point load and the slopes at the ends for the steel shaft which
is simply supported at the bearing A and B as shown in Fig. (i).

Fig (i)
Take E = 200 GPa.
Given:
Young’s modulus, E = 200 GPa
= 200  109 Pa (or) 200  109 N/m2
 Finite Element Analysis 2.171

E = 200  109 N/m2

= 2  105 N/mm2
Moment of inertia,
Element 1 and 2, I1 = 1.25  105 mm4
Element 3, I2 = 1.25  105 mm4

Fig (ii)
For element 1, Length, L1 = 150 mm
For element 2, Length, L2 = 75 m
For element 3, Length, L3 = 125 m
Point load at node 2, F2 = 3000 N
To find:
i) Deflection at the point load
ii) Slope at the ends
Solution: we can divide the beam into three elements as shown in Fig. (iii).

Fig. (iii)
2.172 One Dimensional Problems

For element (1) (Nodes 1, 2, 3, 4 i.e., 𝑑 , ϕ , 𝑑 , ϕ )

Fig.(iv)
Finite element equation is

12 6𝐿 −12 6𝐿 ⎧𝑑 ⎫ 𝐹
𝐸𝐼 6𝐿 4𝐿 −6𝐿 2𝐿 ϕ 𝑚
= … (1)
𝐿 −12 −6𝐿 12 −6𝐿 ⎨𝑑 ⎬ 𝐹
6𝐿 2𝐿 −6𝐿 4𝐿 ⎩ ϕ ⎭ 𝑚

There is no load and moment on element (1).

So, 𝐹 =𝐹 = 0, 𝑚 = 0, 𝐹 = −3000 𝑁, 𝑚 = 0.

1 2 3 4
12 900 −12 900 𝟏 ⎧𝑑 ⎫
2 × 10 × 1.25 × 10 900 9 × 10 900 4.5 × 10 𝟐 ϕ
⇒
150 −12 −900 12 −900 𝟑 ⎨𝑑 ⎬
900 4.5 × 10 −900 9 × 10 𝟒⎩ϕ ⎭
𝐹
= 0 … (2)
−3000
0

1 2 3 4
8.88 666 −8.88 666 1 ⎧𝑑 ⎫ 0
666 66600 −666 33300 2 ϕ 0
10 = … (2)
−8.88 −666 8.88 −666 3 ⎨𝑑 ⎬ −3000
666 33300 −666 66600 4⎩ϕ ⎭ 0
 Similarly,
For element (2) (Nodes 3, 4, 5, 6 i.e., 𝑑 , ϕ , 𝑑 , ϕ )

12 450 −12 450 ⎧𝑑 ⎫ 𝐹

2 × 10 × 1.25 × 10 450 22500 −450 11250 ϕ 𝑚
⇒ =
150 −12 −450 12 −450 ⎨𝑑 ⎬ 𝐹
450 11250 −4500 22500 ⎩ ϕ ⎭ 𝑚

Fig. (v)
There is no load and moment on element (1).
So, 𝐹 = 0, 𝑚 = 0, 𝐹 = 0, 𝑚 = 0.

12 450 −12 450 ⎧𝑑 ⎫ 𝐹

10 450 22500 −450 11250 ϕ =
𝑚
… (4)
−12 −450 12 −450 ⎨𝑑 ⎬ 𝐹
450 11250 −4500 22500 ⎩ ϕ ⎭ 𝑚

3 4 5 6
71.16 2668.5 −71.16 2668.5 𝟑 ⎧𝑑 ⎫ 0
2668.5 133425 −2668.5 66712.5 𝟒 ϕ
⇒ 10 = 0 … (4)
−71.16 −2668.5 71.16 −2668.5 𝟓 ⎨𝑑 ⎬ 0
2668.5 66712.5 −2668.5 133425 𝟔⎩ϕ ⎭ 0

Similarly,
Element (3) (Nodes 5, 6, 8, i.e., 𝑑 , ϕ , 𝑑 , ϕ )

Fig. (vi)
 12 750 −12 750 ⎧𝑑 ⎫ 𝐹
2 × 10 × 4 × 10 750 62500 −750 31250 ϕ 𝑚
⇒ =
125 −12 −750 12 −750 ⎨𝑑 ⎬ 𝐹
750 31250 −750 62500 ⎩ ϕ ⎭ 𝑚

There is no load and moment on element (3).

So, 𝐹 = 0, 𝐹 = 𝐹 , 𝑚 = 𝑚 0.
5 6 7 8
4.92 307.5 −4.92 307.5 5 ⎧𝑑 ⎫ 0
⇒ 10 307.5 25625 −307.5 12812.5 6 ϕ = 𝐹
0
… (5)
−4.92 −307.5 4.92 −307.5 7 ⎨𝑑 ⎬
307.5 12812.5 −307.5 25625 8⎩ϕ ⎭ 0
Assemble the finite elements, i.e., assemble the finite element equations (3), (4) and (5).
1 2 3 4 5 6 7 8
8.88 666 −8.88 666 0 0 0 0 1
⎡ 666 66600 −666 33300 0 0 0 0 ⎤2
⎢−8.88 −666 80.04 2002.5 −71.61 2668.5 0 0 ⎥3
⎢ ⎥
 10 ⎢ 666 33300 2002.5 200025 −2668.5 66712.5 0 0 ⎥4
⎢ 0 0 −71.16 −2668.5 76.08 −2361 −4.92 307.5 ⎥ 5
⎢ 0 0 2668.5 66712.5 −2361 159050 −307.5 12812.5⎥ 6
⎢ 0 0 0 0 −4.92 −307.5 4.92 −307.5 ⎥ 7
⎣ 0 0 0 0 3075 12812.5 −307.5 25625 ⎦ 8
𝑑
⎧ ⎫ 𝐹
⎪∅ ⎪ ⎧
0
⎫
⎪𝑑 ⎪ ⎪−3000⎪
⎪ ⎪ ⎪ ⎪
∅ 0
× = … (6)
⎨ 𝑑 ⎬ ⎨ 0 ⎬
⎪∅ ⎪ ⎪ 0 ⎪
⎪𝑑 ⎪ ⎪ 𝐹 ⎪
⎪ ⎪ ⎩
⎩∅ ⎭ 0 ⎭

Applying boundary conditions,

𝑑 = 0, ∅ = ∅ , 𝑑 = 𝑑 , ∅ =∅
𝑑 = 𝑑 , ∅ =∅ , 𝑑 = 0, ∅ = ∅
 Since, 𝑑 = 0, 𝑑 = 0,neglect first row column, and seventh row seventh
column in global stiffness matrix(6). Hence the equation reduces to
∅
66600 −666 33300 0 0 0 ⎧𝑑 ⎫
⎡ −666 80.04 2002.5 −71.16 2668.5 0 ⎤⎪
⎢33300 ⎪ ⎪
⎪
2002.5 200025 −2668.5 66712.5 0 ⎥⎥ ∅
 10 ⎢
⎢ 0 −71.16 −2668.5 76.08 −2361 307.5 ⎥ ⎨𝑑 ⎬
⎢ 0 2668.5 66712.5 −2361 159050 12812.5⎥ ⎪ ⎪∅ ⎪
⎪
⎣ 0 0 0 307.5 12812.5 25625 ⎦ ⎩ ∅ ⎭
0
⎧−3000⎫
⎪ 0 ⎪
= 0 … (7)
⎨ 0 ⎬
⎪ 0 ⎪
⎩ 0 ⎭
From equation (7)
10 66600 ∅ − 666 𝑑 + 33300 ∅ =0 … . (8)

10 −666 ∅ + 80.04 𝑑 + 2002.5 ∅ − 71.16 𝑑 + 2668.5 ∅

= −3000 … (9)
10 33300 ∅ + 2002.5 𝑑 + 20025 ∅ − 2668.5 𝑑 + 66712.5 ∅
= −3000 … (10)
10 −71.16 𝑑 − 2668.5 ∅ + 76.08 𝑑 − 2361 ∅ + 307.5 ∅
=0 … (11)
10 2668.5 𝑑 + 66712.5 ∅ − 2361 𝑑 + 159050 ∅ + 12812.5 ∅
=0 … (12)
10 307.5 𝑑 + 12812.5 ∅ + 25625 ∅ =0 … (13)
−66600 ∅ + 666 𝑑 𝑦
𝐄𝐪𝐮𝐚𝐢𝐭𝐨𝐧 (𝟖)  ∅ =
33300
∅ = −2 ∅ + 0.02 𝑑
−307.5 𝑑 − 25625 ∅
𝐸𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (13)  ∅ =
12812.5
∅ = −0.024 𝑑 −2 ∅
Substituting the values of ∅ , ∅ in equation (9), (10), (11), (12), we get
From equation (9) 

10 −666 ∅ + 80.04 𝑑 + 2002.5 (−2 ∅ + 0.02 𝑑 − 71.16 𝑑

+ 2668.5 −0.024 𝑑 −2∅ = −3000

 − 666 ∅ + 80.04 𝑑 − 4005 ∅ + 40.05 𝑑 − 71.16 𝑑 − 64.044 𝑑

− 5337 ∅ = −0.3

 − 4671 ∅ + 120.1 𝑑 − 135.2 𝑑 − 5337 ∅ = −0.3 … (14)

We know that,
From equation (10) 
10 33300 ∅ + 2002.5 𝑑 + 200025(−2 ∅ + 0.02 𝑑 − 2668.5 𝑑
+ 6671.5 −0.024 𝑑 −2∅ = 0

 − 366750 ∅ + 6003 𝑑 − 4269.6 𝑑 − 133425 ∅ = 0 … (15)

From equation (11) 

10 −71.16 𝑑 − 2668.5 (−2 ∅ + 0.02 𝑑 + 76.08 𝑑
− 2361 −0.024 𝑑 − 2 ∅ ) + 307.5 𝑑 = 0

 5337 ∅ − 124.5 𝑑 + 132.7 𝑑 + 5029.5 𝑑 =0 … (16)

From equation (12) 

10 2668.5 𝑑 + 66712.5(−2 ∅ + 0.02 𝑑 − 2361 𝑑
− 159050 −0.024 𝑑 − 2 ∅ ) + 12812.5 ∅ = 0

 − 133425 ∅ + 4002.8 𝑑 − 6178.2 𝑑 − 305288 ∅ = 0 … (17)

From equation (17) 

133425 ∅ + 4002.8 𝑑 + 305288 ∅
𝑑 =
−6178.2
𝑑 = −21.6 ∅ + 0.65 𝑑 − 49.4 ∅ … (18)
Substitute the equation (18) in equations (14), (15) and (16),
From equation (14) 
−4671 ∅ + 120.1 𝑑 − 135.2 −21.6 ∅ + 0.65 𝑑 − 49.4 ∅
− 5337 ∅ = −0.3

 − 1750.7 ∅ + 32.28 𝑑 + 1341.9 ∅ = −0.3 … (19)

From equation (15) 

−366750 ∅ + 6003 𝑑 − 4269.6 −21.6 ∅ + 0.65 𝑑 − 49.4 ∅
− 133425 ∅ = 0

 − 274527 ∅ + 3355.8 𝑑 + 77493 ∅ = 0 … (20)

From equation (16) 

−5337 ∅ − 124.5 𝑑 − 132.7 −21.6 ∅ + 0.65 𝑑 − 49.4 ∅
+ 5029.5 ∅ = 0

 2470.7 ∅ − 38.2 𝑑 − 1525.9 ∅ = 0 … (21)

2470.7 ∅ − 38.2 𝑑
∅ = … (22)
1525.9

 ∅ = 1.62 ∅ − 0.025 𝑑 … (23)

Substitute the equation (23) in equations (20), we get
Form equation (20) 
− 274527 ∅ + 3355.8 𝑑 + 77493 1.62 ∅ − 0.025 𝑑 =0

 − 148988 ∅ + 1418.5 𝑑 =0

∴ − 148988 ∅ = −1418.5 𝑑

𝑑 = 105 ∅ … (24)
Substitute equation (240 and (25) in equation (19),
2.178 One Dimensional Problems

Equation (19) 
−1750.7 ∅ + 32.2 (105 ∅ ) + 1341.6 [1.62 ∅ − 0.025 (105 ∅ )]
= −0.3
281.7 ∅ = −0.3
∅ = −0.0011 𝑟𝑎𝑑
∴ Subsittue the value (∅ ) in equaiton (24),
𝑑 = 105(−0.0011)
𝑑 = −0.112 𝑚𝑚
∴ 𝑆𝑢𝑏𝑠𝑖𝑡𝑡𝑢𝑒 𝑡ℎ𝑒 𝑣𝑎𝑙𝑢𝑒 (∅ ) 𝑎𝑛𝑑 (𝑑 )𝑖𝑛 𝑒𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (23),
∅ = 1.62(−0.0011) − 0.025(−0.112)
∅ = 0.0010 𝑟𝑎𝑑
Result:
1. Deflection at the loaded point, 𝑑 = −0.112 𝑚𝑚
2. Slope at the left bearing, ∅ = −0.0011 𝑟𝑎𝑑
3. Slope at the right bearing ∅ = 0.0010 𝑟𝑎𝑑

2.13. QUADRATIC BAR ELEMENT

2.13.1. Derivation of Shape Functions for One-Dimensional Quadratic
Bar Element
Consider a quadratic bar element with nodes 1,2 and 3 as shown in Fig.(i). u1,
u2 and u3 are the displacements at the respective nodes. So, u1, u2 and u3 are considered
as degrees of freedom of this quadratic bar element.
Since the element has got three nodal displacements, it will have three
generalized coordinates.
𝑢 =𝑎 +𝑎 𝑥+𝑎 𝑥 … (1)
Where, 𝑎 − 0 , 𝑎 , 𝑎 are global or generalized coordinates. Writing the equation
(2.47) in matrix form,
 Finite Element Analysis 2.179
𝑎
𝑢 = 1𝑥𝑥 𝑎
[ ] … (2)
𝑎
At node 1, u = u , x=0
At node 2, u = u , x=l
l
At node 3, u = u , x=
2

Fig. (i). Quadratic bar element

Substitute the above values in equation (2.47),
𝑢 =𝑎 … (3)
𝑢 =𝑎 +𝑎 𝑙+𝑎 𝑙 … (4)
𝑙 𝑙
𝑢 =𝑎 +𝑎 +𝑎 … (5)
2 2
Substitute the equation (3) in equation (4) and (5),
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (2.50)  𝑢 = 𝑢 + 𝑎 𝑙 + 𝑎 𝑙 … (6)
𝑎 𝑙 𝑎 𝑙
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (2.51)  𝑢 = 𝑢 + + … (7)
2 4
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (2.52)  𝑢 − 𝑢 = 𝑎 𝑙 + 𝑎 𝑙 … (8)
𝑎 𝑙 𝑎 𝑙
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (2.53)  𝑢 − 𝑢 = + … (9)
2 4
Arranging the equation (2.54) and (2.55) in matrix form,

𝑢 −𝑢 𝑙 𝑙 𝑎
𝑢 −𝑢 = 𝑙 𝑙 𝑎
2 4
2.180 One Dimensional Problems

𝑎 𝑙 𝑙 𝑢 −𝑢
 𝑎 = 𝑙 𝑙 𝑢 −𝑢
2 4

𝑙
1 𝑙 𝑢 −𝑢
= 4
𝑙 𝑙 −𝑙 𝑢 −𝑢
(4 − 2) 𝑙
2

𝑎 𝑎 1 𝑎 −𝑎
𝑁𝑜𝑡𝑒: ∵ 𝑎 𝑎 = × −𝑎 𝑎
(𝑎 𝑎 −𝑎 𝑎 )

𝑙
𝑎 1 −𝑙 𝑢 −𝑢
 = 4 … (10)
𝑎 −𝑙 −𝑙 𝑢 −𝑢
4 1
2
−4 𝑙
 𝑎 = (𝑢 − 𝑢 ) − 𝑙 (𝑢 − 𝑢 ) … (11)
𝑙 4

−4 −𝑙
 𝑎 = (𝑢 − 𝑢 ) − 𝑙(𝑢 − 𝑢 ) … (12)
𝑙 4
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (2.57)
−4 𝑙 𝑢 𝑙 𝑢
𝑎 = − −𝑙 𝑢 +𝑙 𝑢
𝑙 4 4

−4 𝑙 𝑢 4𝑙 𝑢 4𝑙 𝑢 4𝑙 𝑢
= + + −
4𝑙 4𝑙 𝑙 𝑙
−𝑢 𝑢 4𝑢 4𝑢
= + + −
𝑙 𝑙 𝑙 𝑙
−3 𝑢 𝑢 4𝑢
𝑎 = − + … (13)
𝑙 𝑙 𝑙
Equation (2.58) 
−4 −𝑙 𝑢 𝑙
𝑎 = + 𝑢 +𝑙𝑢 −𝑙𝑢
𝑙 4 2
 Finite Element Analysis 2.181
4𝑙𝑢 4𝑙 4𝑙 4𝑙
= − 𝑢 − 𝑢 − 𝑢
2𝑙 2𝑙 𝑙 𝑙
2𝑢 2 4 4
= − 𝑢 𝑢 − 𝑢
𝑙 𝑙 𝑙 𝑙
2 2𝑢 4
𝑎 = 𝑢 − + 𝑢 … (14)
𝑙 𝑙 𝑙
Arranging the equation (2.49), (2.59) and (2.60) in matrix form,
1 0 0
𝑎 ⎡−3 −1 4 ⎤ 𝑢
𝑎 ⎢ ⎥
= ⎢ 𝑙 𝑙 𝑙 ⎥ 𝑢 … (15)
𝑎 ⎢2 2 −4⎥ 𝑢
⎣𝑙 𝑙 𝑙 ⎦
Substitute the equation (2.61) in equation (2.48),
1 0 0
⎡−3 −1 4 ⎤ 𝑢
⎢ ⎥
{𝑢} = [ 1 𝑥 𝑥 ]⎢ 𝑙 𝑙 𝑙 ⎥ 𝑢 … (16)
⎢2 2 −4⎥ 𝑢
⎣𝑙 𝑙 𝑙 ⎦
𝑢
3 2𝑥 −𝑥 2 𝑥 4𝑥 4𝑥
{𝑢} = 1− 𝑥+ + + 𝑢
𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑢
𝑢
{𝑢} = [𝑁 𝑁 𝑁 ] 𝑢
𝑢

{𝑢} = 𝑁 𝑢 + 𝑁 𝑢 𝑁 𝑢 … (17)

Where, shape functions,

3𝑥 2𝑥
𝑁 = 1− +
𝑙 𝑙
−𝑥 2 𝑥
𝑁 = +
𝑙 𝑙
4𝑥 4𝑥
𝑁 = −
𝑙 𝑙
2.182 One Dimensional Problems

2.13.2. Solved Problems on Quadratic Bar Element

Example 2.28
A Steel bar of length 800mm is subjected to an axial load of 3kN as shown in fig
(i). Find the nodal displacement of the bar and load vectors,
Take E=2*105N/mm2, A=300mm2, and mass density  =7800kg/m3

Fig. (i).
Given: Length, l = 800 mm
Load, F = 3 kN = 3 × 103 N
Young’s modulus, E = 2 × 105 N/mm2
Area, A = 300 mm2
Mass desnity,  = 7800 kg/m2 – 76518 N/m3 = 7.6518 × 10-5 N/mm3
To find: (i) Nodal displacements
(ii) Load vectors
Solution: we can divide the bar into two elements as shown in Fig. (ii).

Fig. (ii)
 Finite Element Analysis 2.183

The bar is subjected to self-weight. So, we have to find the body force acting at
nodal points 1, 2 and 3.
We know that,
Force vector for quadratic bar element,
1
⎧ ⎫
𝐹 ⎪6⎪
1
𝐹 = 𝜌𝐴𝑙
𝐹 ⎨6⎬
⎪2⎪
⎩3⎭
1
⎧ ⎫
⎪6⎪
1
= 7.6518 × 10 × 300 × 800
⎨6⎬
⎪2⎪
⎩3⎭
1
⎧ ⎫
⎪6⎪
1
= 18.364
⎨6⎬
⎪2⎪
⎩3⎭

𝐹 3.06
𝐹 = 3.06 . . (1)
𝐹 12.24
At point load of 3 ×103 N is acting at node 3 as shown in Fig. (ii). So, add 3 ×103
N in F3 vector,
𝐹 3.06
 𝐹 = 3.06
𝐹 12.24 + 3 × 10

𝐹 3.06
 𝐹 = 3.06 … (2)
𝐹 3012.24
∴ Finite element equation for one dimensional quadratic bar element is given by,
2.184 One Dimensional Problems
𝐹 𝐸𝐴 7 1 −8 𝑢
 𝐹 = 1 7 −8 𝑢 … (3)
3𝑙
𝐹 −8 −8 16 𝑢

Substitute the equation (2) in equation (3),

3.06 2 × 10 × 300 7 1 −8 𝑢
3.06 = 1 7 −8 𝑢
3 × 800
3012.24 −8 −8 16 𝑢
3.06 7 1 −8 𝑢
3.06 = 25000 1 7 −8 𝑢 … (4)
3012.24 −8 −8 16 𝑢

Applying the boundary conditions, i.e.,

At node 1, displacement, u1 = 0 (fixed)

At node 2, displacement, u2 = u2

At node 3, displacement, u3 = u3

Substituting 𝑢 , 𝑢 𝑎𝑛𝑑 𝑢 values in equation (4),

3.06 7 1 −8 𝑢
 3.06 = 25000 1 7 −8 𝑢
3012.24 −8 −8 16 𝑢

In the above equation 𝑢 = 0, so, neglect first row and first column of [ k]
matrix. The reduced equation is,

7 −8 𝑢 3.06
= 25000 =
−8 16 𝑢 3012.24

 25000 [7𝑢 − 8 𝑢 ] = 3.06 . . . (5)

 25000 [−8 𝑢 + 16 𝑢 ] = 3012.24 . . . (6)

∴ 𝐸𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (6) × 7  25000[ −56 𝑢 + 112 𝑢 ] = 2.108 × 10

𝐸𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (5) × 8  25000[ 56 𝑢 − 56 𝑢 ] = 24.48

 Finite Element Analysis 2.185

𝑠𝑜𝑙𝑣𝑖𝑛𝑔 25000 (56 𝑢 ) = 2.1104 × 10

 𝑢 = 0.0150 𝑚𝑚

Substitute 𝑢 value in equation (5),

 25000[ 7 𝑢 − 8(0.015)] = 3.06

 𝑢 = 0.0171 𝑚𝑚

Result:

𝐹 3.06
Load vector, 𝐹 = 3.06
𝐹 3012.24

Nodal displacement, 𝑢 = 0

𝑢 = 0.0171 𝑚𝑚

𝑢 = 0.0150 𝑚𝑚
 UNIT 3
TWO DIMENSIONAL PROBLEMS

3.1. INTRODUCTION
This chapter considers the two dimensional finite element. Two dimensional
elements are defined by three or more nodes in a two dimensional plane (i.e., x, y plane).
The basic element useful for two dimensional analysis is the triangular element. The
simplest two dimensional elements have corner nodes as shown in Fig.3.1. A quadrilateral
(special forms of rectangle and parallelogram) element can be obtained by assembling
two or four triangular elements, as shown in Fig.3.2. They are often used to model a wide
range of engineering problems.

Fig. 3.1. Two dimensional elements

Fig. 3.2. A quadrilateral element as an assemblage of two or four triangular

elements
3.2 Two Dimensional Problems

The two dimensional analysis of hydraulic cylinder rod end with plane strain
triangular elements is shown in Fig.3.3.

Fig. 3.2. Two dimensional analysis of hydraulic cylinder rod end

The two dimensional finite element formulation follows the same steps which is
used in the one dimensional problems. The displacements and distributed body force
values are functions of the position indicated by (x, y).
𝑢
The displacement vector u is given by, u =
𝑣
Where u and v are the x and y components of u respectively.
The stresses and strains are given as,
𝜎
𝜎= 𝜎
𝜏

𝑒
𝑒= 𝑒
𝛾

Where,   Normal stress

  Shear stress

e  Normal strain

  Shear strain.
𝐹𝑥
Body force is given by, F=
𝐹𝑦
 Finite Element Analysis 3.3

3.2. PLANE STRESS AND PLANE STRAIN

The two dimensional element is extremely important for the following two analysis.
(i) Plane stress analysis.
(ii) Plane strain analysis.
(i) Plane Stress Analysis
Plane stress is defined to be a state of stress in which the normal stress (σ) and
shear stress (τ) directed perpendicular to the plane are assumed to be zero.
Generally, members that are thin (those with a small z dimension compared to the
in-plane x and y dimensions) and whose loads act only in the x-y plane can be considered
to be under plane stress.
Plates with holes and plates with fillets are coming under plane stress analysis
problems.

Fig. 3.4 Plane stress problems: (a) plate with hole; (b) plate with fillet
where, T→ Surface tractions (i.e., pressure acting on the surface edge or face of a
member, unit→ Force/Area → N/m²)
Normal stress, σz = 0
Shear stress τxz and τyz = 0
(ii) Plane strain analysis
Plane strain is defined to be a state of strain in which the strain normal to the xy
plane and the shear strains are assumed to be zero.
Dams and pipes subjected to loads that remain constant over their lengths are
coming under plane strain analysis problems.
3.4 Two Dimensional Problems

Fig. 3.5 Plane strain problems; (a) dam subjected to horizontal loading; (b) pipe
subjected to a vertical load
Here, Normal stress, ez = 0
Shear stresses, γxz and γyz = 0

3.3. FINITE ELEMENT MODELLING

Finite element modelling consists of the following:
(i) Discretization of structure.
(ii) Numbering of nodes.
(i) Discretization
The art of subdividing a structure into a convenient number of smaller components
is known as discretization. In two dimensional problems, three kinds of finite elements
are used.
They are:
(i) Triangular element.
(ii) Rectangular element.
(iii) Quadrilateral element.
In truss, the above three elements are physically present. But in a continuum, the
above three elements exist only in our imagination. The continuum shown in Fig.3.6 is
 Finite Element Analysis 3.5

discretized into eight triangular element as shown in Fig.3.7. The points where the corners
of the triangles meet are called nodes. Each triangle formed by three nodes and three sides
is called an element.

Fig. 3.6. Continuum Fig. 3.7. Discretized into eight triangular elements
The element numbers are circled to distinguish from node numbers. The cross-
section area, traction force and body force are constant within each element. But these are
differ in magnitude from element to element. Better results are obtained by increasing the
number of elements.
In Fig.3.7, the triangular elements fill the entire region except a small region at
the boundary. This unfilled region can be eliminated by choosing smaller elements or
elements with curved boundaries.
(ii) Numbering of Nodes

Fig. 3.8. Triangular element

3.6 Two Dimensional Problems

In one dimensional problem, each node is allowed to move only in ±x direction.

But in two dimensional problem, each node is permitted to move in the two directions
i.e., x and y. Hence each node has two degrees of freedom (Nodal displacements). A three
node finite element model is shown in Fig.3.8 has six degrees of freedom.
The element connectivity table is given for Fig.3.7. The heading 1 and 2 refer to
the local node numbers of an element and the corresponding node numbers on the body
are called global numbers. Connectivity thus establishes the local-global correspondence.

3.4. CONSTANT STRAIN TRIANGULAR (CST) ELEMENT

A three noded triangular element is known as constant strain triangular (CST)
element which is shown in Fig.3.9. It has six unknown displacement degrees of freedom
(u1 v1, u2 v2, u3 v3). The element is called CST because it has a constant strain throughout
it.

Fig. 3.9. Constant strain triangular element

3.5. SHAPE FUNCTION DERIVATION FOR THE CONSTANT

STRAIN TRIANGULAR ELEMENT (CST)
We begin this section with the development of the shape function for a basic two
dimensional finite element, called constant strain triangular element (CST).
We consider this CST element because its derivation is the simplest among the
available two dimensional elements.
 Finite Element Analysis 3.7

Fig. 3.10. Three noded CST element

Consider a typical CST element with nodes 1, 2 and 3 as shown in Fig.3.10 Let
the nodal displacements be u1 , v1, u2 , v2, u3 and v3
𝑢
⎧𝑢 ⎫
⎪𝑢 ⎪
Displacement {𝑢}= 𝑣
⎨ ⎬
⎪𝑣 ⎪
⎩𝑣 ⎭
since the CST element has got two degrees of freedom at each node (u, v) the total
degrees of freedom is 6. Hence it has 6 generalized coordinates,
Let, 𝑢 =𝑎 +𝑎 𝑥+𝑎 𝑦 …(3.1)
𝑣 =𝑎 +𝑎 𝑥+𝑎 𝑦 …(3.2)
Where, 𝑎 , 𝑎 , 𝑎 , 𝑎 , 𝑎 and 𝑎 are global or generalized co-ordinates.
 𝑢 =𝑎 +𝑎 𝑥 +𝑎 𝑦
𝑢 =𝑎 +𝑎 𝑥 +𝑎 𝑦
𝑢 =𝑎 +𝑎 𝑥 +𝑎 𝑦
Write the above equations in matrix form,
𝑢 1 𝑥 𝑦 𝑎
𝑢 = 1 𝑥 𝑦 𝑎
𝑢 1 𝑥 𝑦 𝑎
3.8 Two Dimensional Problems

𝑎 1 𝑥 𝑦 𝑢
⇒ 𝑎 = 1 𝑥 𝑦 𝑢 … (3.3)
𝑎 1 𝑥 𝑦 𝑢

+ − +
⎡1 𝑥 𝑦 ⎤
⎢− + −⎥
𝐿𝑒𝑡 𝐷 = ⎢1 𝑥 ⎥
𝑦 ⎥
⎢
⎢+ − +⎥
⎣1 𝑥 𝑦 ⎦
𝐶
We know, 𝐷 = … (3.4)
|𝐷|
Find the co-factors of matrix D.
𝑥 𝑦
𝐶 =+ 𝑥 𝑦 = (𝑥 𝑦 − 𝑥 𝑦 )

1 𝑦
𝐶 =− = (𝑦 − 𝑦 ) = 𝑦 − 𝑦
1 𝑦
1 𝑥
𝐶 =+ =𝑥 −𝑦
1 𝑥
𝑥 𝑦
𝐶 =− 𝑥 𝑦 = (𝑥 𝑦 − 𝑥 𝑦 )

1 𝑦
𝐶 =+ =𝑦 −𝑦
1 𝑦
1 𝑥
𝐶 =− = −(𝑥 − 𝑥 ) = 𝑥 − 𝑥
1 𝑥
𝑥 𝑦
𝐶 =+ 𝑥 𝑦 =𝑥 𝑦 −𝑥 𝑦

1 𝑦
𝐶 =− = (𝑦 − 𝑦 ) = 𝑦 − 𝑦
1 𝑦
1 𝑥
𝐶 =+ =𝑥 −𝑥
1 𝑥
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑦 − 𝑦 ) (𝑥 − 𝑦 )
⇒ 𝐶 = (𝑥 𝑦 − 𝑥 𝑦 ) (𝑦 − 𝑦 ) (𝑥 − 𝑥 )
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑦 − 𝑦 ) (𝑥 − 𝑥 )
 Finite Element Analysis 3.9

(𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 )
⇒ 𝐶 = (𝑦 − 𝑦 ) (𝑦 − 𝑦 ) (𝑦 − 𝑦 ) … (3.5)
(𝑥 − 𝑦 ) (𝑥 − 𝑥 ) (𝑥 − 𝑥 )
1 𝑥 𝑦
We know, 𝐷= 1 𝑥 𝑦
1 𝑥 𝑦
|𝐷| = 1(𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 ) … (3.6)
Substitute CT and D values in equation (3.4),
1
⇒ 𝐷 =
(𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ))
× (𝑦 −𝑦 ) (𝑦 − 𝑦 ) (𝑦 − 𝑦 )
(𝑥 −𝑦 ) (𝑥 − 𝑥 ) (𝑥 − 𝑥 )
Substitute D-1 values in equation (3.3),
𝑎 1 𝑥 𝑦 𝑢
⇒ 𝑎 = 1 𝑥 𝑦 𝑢
𝑎 1 𝑥 𝑦 𝑢
𝑎
⇒ 𝑎
𝑎
1
=
(𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 )) 𝑢
× (𝑦 − 𝑦 ) (𝑦 − 𝑦 ) (𝑦 − 𝑦 ) 𝑢 … (3.7)
(𝑥 − 𝑦 ) (𝑥 − 𝑥 ) (𝑥 − 𝑥 ) 𝑢

The area of the triangle can be expressed as a function of the x, y co-ordinates of

the nodes 1, 2 and 3.

1 1 𝑥 𝑦
⇒ 𝐴= 1 𝑥 𝑦
2 1 𝑥 𝑦
1
|𝐴| = [1(𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )]
2
⇒ 2 𝐴 = (𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 ) … (3.8)
3.10 Two Dimensional Problems

Substitute 2A values in equation (3.7),

𝑎
⇒ 𝑎
𝑎
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 )) 𝑢
1
= × (𝑦 − 𝑦 ) (𝑦 − 𝑦 ) (𝑦 − 𝑦 ) 𝑢 … (3.9)
2𝐴 𝑢
(𝑥 − 𝑦 ) (𝑥 − 𝑥 ) (𝑥 − 𝑥 )
𝑎 𝑝 𝑝 𝑝 𝑢
1
⇒ 𝑎 = × 𝑞 𝑞 𝑞 × 𝑢 … (3.10)
𝑎 2𝐴 𝑟 𝑟 𝑟 𝑢
Where, 𝑝 =𝑥 𝑦 −𝑥 𝑦 ; 𝑝 =𝑥 𝑦 −𝑥 𝑦 ; 𝑝 =𝑥 𝑦 −𝑥 𝑦
𝑞 =𝑦 −𝑦 ; 𝑞 =𝑦 −𝑦 ; 𝑞 =𝑦 −𝑦
𝑟 =𝑥 −𝑥 ; 𝑟 =𝑥 −𝑥 ; 𝑟 =𝑥 −𝑥
From equation (3.1), we know that,
𝑢 =𝑎 +𝑎 𝑥+𝑎 𝑦
We can write this equation in matrix form,
𝑎
𝑢 = [1 𝑥 𝑦] 𝑎
𝑎
𝑎
Substitute 𝑎 value, from equation no. (3.10)
𝑎
𝑝 𝑝 𝑝 𝑢
1
⇒ 𝑢 = [1 𝑥 𝑦] × 𝑞 𝑞 𝑞 × 𝑢
2𝐴 𝑟 𝑟 𝑟 𝑢
𝑝 𝑝 𝑝 𝑢
1
= [1 𝑥 𝑦] 𝑞 𝑞 𝑞 𝑢
2𝐴 𝑟 𝑟 𝑟 𝑢
1
= [𝑝 + 𝑞 𝑥 + 𝑟 𝑦 𝑝 +𝑞 𝑥+𝑟 𝑦 𝑝 + 𝑞 𝑥 + 𝑟 𝑦]
2𝐴
𝑢
× 𝑢
𝑢
 Finite Element Analysis 3.11

[∵ (1 × 3) × (3 × 3) = 1 × 3
𝑢
𝑝 +𝑞 𝑥+𝑟 𝑦 𝑝 +𝑞 𝑥+𝑟 𝑦 𝑝 +𝑞 𝑥+𝑟 𝑦
𝑢= [ ]× 𝑢
2𝐴 2𝐴 2𝐴 𝑢
The above equation is in the form of
𝑢
𝑢 = [𝑁 𝑁 𝑁] 𝑢 … (3.11)
𝑢
𝑣
Similarly, 𝑣 = [𝑁 𝑁 𝑁] 𝑣 … (3.12)
𝑣
𝑝 +𝑞 𝑥+𝑟 𝑦
Where, Shape Function, 𝑁 =
2𝐴
𝑝 +𝑞 𝑥+𝑟 𝑦
𝑁 =
2𝐴
𝑝 +𝑞 𝑥+𝑟 𝑦
𝑁 =
2𝐴
Assembling the equations (3.11) and (3.12) in matrix form,

𝑢
⎧𝑣 ⎫
𝑢(𝑥, 𝑦) 𝑁 0 𝑁 0 𝑁 0 ⎪𝑢 ⎪
Displacement function, u = = × 𝑣 … (3.13)
𝑣(𝑥, 𝑦) 0 𝑁 0 𝑁 0 𝑁 ⎨ ⎬
⎪𝑢 ⎪
⎩𝑣 ⎭
3.6 STRAIN – DISPLACEMENT MATRIX [ B ] FOR CST ELEMENT
Displacement function for CST element is given by,
𝑢
⎧𝑣 ⎫
𝑢(𝑥, 𝑦) 𝑁 0 𝑁 0 𝑁 0 ⎪𝑢 ⎪
u= = × 𝑣
𝑣(𝑥, 𝑦) 0 𝑁 0 𝑁 0 𝑁 ⎨ ⎬
⎪𝑢 ⎪
⎩𝑣 ⎭
or we can write
u=N 𝑢 +𝑁 𝑢 +𝑁 𝑢
3.12 Two Dimensional Problems

u =N 𝑣 +𝑁 𝑣 +𝑁 𝑣
The strain components for CST element are,
𝜕𝑢
Normal strain, e =
𝜕𝑥
𝜕𝑁 𝜕𝑁 𝜕𝑁
⇒e = 𝑢 + 𝑢 + 𝑢
𝜕𝑥 𝜕𝑥 𝜕𝑥
𝜕𝑣
Normal strain, e =
𝜕𝑦
𝜕𝑁 𝜕𝑁 𝜕𝑁
⇒e = 𝑣 + 𝑣 + 𝑣
𝜕𝑦 𝜕𝑦 𝜕𝑦
𝜕𝑢 𝜕𝑣
Shear strain, 𝛾 = +
𝜕𝑦 𝜕𝑥
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
⇒𝛾 = 𝑢 + 𝑢 + 𝑢 + 𝑣 + 𝑣 + 𝑣
𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑦
Arranging the strains e , 𝑒 and 𝛾 in matrix form,
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝑢
⎡ 0 0 0 ⎤
⎢ 𝜕𝑥 𝜕𝑥 𝜕𝑥 ⎥ ⎧ 𝑣 ⎫
𝑒
𝑒 ⎢ 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎪𝑢 ⎪
⇒ =⎢ 0 0 0 × … (3.14)
𝛾 𝜕𝑦 𝜕𝑦 𝜕𝑦 ⎥ ⎨𝑣 ⎬
⎢𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ 𝑢
⎢ 𝜕𝑦 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎪ ⎪
𝜕𝑦 𝜕𝑦 ⎩𝑣 ⎭
⎣ 𝜕𝑥 𝜕𝑥 𝜕𝑥 ⎦
From equation (3.11) or (3.12), we know that,
𝑝 +𝑞 𝑥+𝑟 𝑦
Shape Function, 𝑁 =
2𝐴
𝑝 +𝑞 𝑥+𝑟 𝑦
𝑁 =
2𝐴
𝑝 +𝑞 𝑥+𝑟 𝑦
𝑁 =
2𝐴
Partial differentiation,
𝜕𝑁 𝑞 𝜕𝑁 𝑞 𝜕𝑁 𝑞
= ; = ; =
𝜕𝑥 2𝐴 𝜕𝑥 2𝐴 𝜕𝑥 2𝐴
 Finite Element Analysis 3.13
𝜕𝑁 𝑟 𝜕𝑁 𝑟 𝜕𝑁 𝑟
= ; = ; =
𝜕𝑦 2𝐴 𝜕𝑦 2𝐴 𝜕𝑦 2𝐴
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
Substitute , , , , and values in equation (3.14),
𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑦
𝑢
⎧𝑣 ⎫
𝑒 𝑞 0 𝑞 0 ⎪
𝑒 1 𝑞 0 𝑢 ⎪
(3.14) ⇒ = 0 𝑟 0 𝑟 0 𝑟
2𝐴 𝑣
𝛾 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞 ⎨𝑢 ⎬
⎪ ⎪
⎩𝑣 ⎭
The above equation is in the form of { e } = [ B ] { u }
𝑞 0 𝑞 0 𝑞 0
Where, [ B ] = Strain – Displacement matrix = 0 𝑟 0 𝑟 0 𝑟 … (3.15)
𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
𝑤ℎ𝑒𝑟𝑒, 𝑞 =𝑦 −𝑦
𝑞 =𝑦 −𝑦
𝑞 =𝑦 −𝑦
[ From equation no. (3.10)]
𝑟 =𝑥 −𝑥
𝑟 =𝑥 −𝑥
𝑟 =𝑥 −𝑥

3.7. STRESS-STRAIN RELATIONSHIP MATRIX OR

CONSTITUTTIVE MATRIX [D] FOR TWO DIMENSIONAL
ELEMENT
Consider a three dimensional body which is subjected to the stresses 𝜎 , 𝜎 and
𝜎 independently as shown in Fig. 311.
3.14 Two Dimensional Problems

Hooke’s law states that when a material is loaded within its elastic limit, the stress
is directly proportional to the strain.
i.e., stress  strain
e
 = Ee
𝜎
𝑒=
𝐸
Where, e = Strain
 = Stress, N/mm2
E = Yound’s modulus or Modulus of elasticity, N/mm2
The stress in the x direction produces a positive strain in x direction as shown in
Fig. 3.12.

Fig. 3.12
𝜎
Strain, e = … (3.16)
𝐸

Fig. 3.13
Fig. 3.13 shows the positive stress in the y direction produces a negative strain in
the x direction as a result of Poisson’s effect which is given by,
 Finite Element Analysis 3.15
𝑣𝜎
−e =
𝐸
−𝑣𝜎
⇒ e = … (3.17)
𝐸
Where, v  Poisson’s ratio.
Similarly, the stress in the z direction produces a negative strain in the x direction
as shown in Fig. 3.14.

Fig. 3.14
𝑣𝜎
−e =
𝐸
−𝑣𝜎
⇒ e = … (3.18)
𝐸
By applying superposition principle to the equations (3.16), (3.17) and (3.18),
we get
𝜎 𝜎 𝜎
e = −𝑣 −𝑣 … (3.19)
𝐸 𝐸 𝐸
This is a strain equation in x direction.
Similarly, the strains in y and z directions can be calculated as follows:
𝜎 𝜎 𝜎
Strain in y direction, e = −𝑣 + 𝑣 − 𝑣 … (3.20)
𝐸 𝐸 𝐸
𝜎 𝜎 𝜎
Strain in z direction, e = −𝑣 − 𝑣 + 𝑣 … (3.21)
𝐸 𝐸 𝐸
Solving equations (3.19), (3.20) and (3.21) for the normal stresses ( 𝜎 , 𝜎 and 𝜎 ),
we get.
𝐸
𝜎 = 𝑒 (1 − 𝑣) + 𝑣𝑒 + 𝑣𝑒 … (3.22)
(1 + 𝑣)(1 − 2𝑣)
3.16 Two Dimensional Problems
𝐸
𝜎 = 𝑣 𝑒 (1 − 𝑣)𝑒 + 𝑣𝑒 … (3.23)
(1 + 𝑣)(1 − 2𝑣)
𝐸
𝜎 = 𝑣 𝑒 + 𝑣𝑒 + (1 − 𝑣)𝑣𝑒 … (3.24)
(1 + 𝑣)(1 − 2𝑣)
The shear stress and shear strain relationship is given by,
 = G
Where,   Shear stress
  Shear strain
G  Modulus of rigidity or Shear modulus
The expressions for the three different sets of shear stresses are,
 = 𝐺

 = 𝐺

 = 𝐺
E
Where, G → Modulus of rigidity =
2(1 + v)
E
⇒ = 
2(1 + v)

E 1 − 2𝑣
 = × × … (3.25)
(1 + v)(1 − 2v) 2

E
⇒ = ×
2(1 + v)

E 1 − 2𝑣
 = × × … (3.26)
(1 + v)(1 − 2v) 2

E
⇒ = ×
2(1 + v)

E 1 − 2𝑣
 = × × … (3.27)
(1 + v)(1 − 2v) 2
 Finite Element Analysis 3.17

Assembling the equations (3.22). (3.23), (3.24), (3.25), (3.26) and (3.27) in matrix form,
𝜎 0 0 0 𝑒
⎡(1 − 𝑣) 𝑣 𝑣
0 0 0 ⎤⎧𝑒 ⎫
⎧𝜎 ⎫ ⎢ 𝑣 (1 − 𝑣) 𝑣
⎪𝜎 ⎪ E 0 0 0 ⎥⎪ 𝑒 ⎪
⎢ 𝑣 𝑣 (1 − 𝑣) 1 − 2𝑣 ⎥
⇒ 𝜏 = ⎢ 0 0 ⎥ 
⎨ ⎬ (1 + v)(1 − 2v) ⎢ 0 0 0 2 1 − 2𝑣 0 ⎥ ⎨ ⎬
𝜏
⎪ ⎪ 0 0 0 1 − 2𝑣 ⎥ ⎪ ⎪
⎢ 0 2
⎩𝜏 ⎭ ⎣ 0 0 
0 0 0 2 ⎦⎩ ⎭

…(3.28)
The above equation is in the form of
{}=[D]{e}
The above equation (3.28) gives a three dimensional stress-strain relationship for
an isotropic body.
Where, [D] is stress-strain relationship matrix or constitutive matrix.
0 0 0
⎡(1 − 𝑣) 𝑣 𝑣
0 0 0 ⎤
⎢ 𝑣 (1 − 𝑣) 𝑣 0 0 0 ⎥
E ⎢ 𝑣 𝑣 (1 − 𝑣) 1 − 2𝑣 ⎥
[𝐷] = 0 0 ⎥ … (3.29)
(1 + v)(1 − 2v) ⎢ 0 0 0 1 − 2𝑣 0 ⎥
⎢ 0 0 2
⎢ 0 0 1 − 2𝑣 ⎥
2
⎣ 0 0 0 0 0 2 ⎦
Where, E = Modulus of Elasticity or Young’s modulus
V = Poisson’s ratio

3.7.1 Plane Stress

For two dimensional plane stress problems, the normal stress, 𝜎 and shear
stresses 𝜏 , 𝜏 are zero.
i.e., 𝜎 =𝜏 = 𝜏 =0
The shear strains  ,  are zero but ez are not equal to zero

i.e.,  = =0

Substitute 𝜎 = 0 in equation (3.19),

𝜎 𝜎
⇒e = −𝑣 … (3.30)
𝐸 𝐸
3.18 Two Dimensional Problems

Substitute 𝜎 = 0 in equation (3.20),

𝜎 𝜎
⇒ e = −𝑣 + … (3.31)
𝐸 𝐸
Substitute in equation (3.30) and (3.31),
𝜎 𝑣𝜎
e = −
𝐸 𝐸
𝜎 𝜎
𝑣e = −𝑣 +𝑣 [Equation (3.31) × 𝑣]
𝐸 𝐸

𝜎 𝑣 𝜎
e + 𝑣e = −
𝐸 𝐸
𝜎
⇒ e + 𝑣e = (1 − 𝑣 )
𝐸
𝐸
⇒ 𝜎 = e + 𝑣e … (3.32)
(1 − 𝑣 )
Substitute in equation (3.30) and (3.31),
𝜎 𝜎
𝑣e =𝑣 −𝑣 [Equation (3.30) × 𝑣]
𝐸 𝐸
𝜎 𝜎
e = −𝑣 +
𝐸 𝐸
𝜎 𝜎
𝑣e +𝑒 =𝑣 +𝑣
𝐸 𝐸
𝜎
𝑣 e + 𝑒 = (1 − 𝑣 )
𝐸
𝐸
𝜎 = 𝑣e +𝑣e … (3.33)
(1 − 𝑣 )
We know that, Shear stress, 𝜏 =𝐺

Where, G  Modulus of rigidity = ( )

  Shear strain

𝑣  Poisson’s ratio
 Finite Element Analysis 3.19
𝐸
⇒𝜏 = 
2(1 + 𝑣)
𝐸 1 − 2𝑣
𝜏 = × × … (3.34)
(1 − 𝑣) 2
Assembling the equations (3.32), (3.33) and (3.34) in matrix form,

𝜎 1 𝑣 0 𝑒
E 𝑣 1 0 𝑒
⇒ 𝜎 = 1−𝑣 … (3.35)
𝜏 (1 − 𝑣) 
0 0
2
The above equation is in the form of
{}=[D]{e}
The above equation (3.35) gives a two dimensional stress-strain relationship for
plane stress problems.
Where, [D] is stress-strain relationship matrix or constitutive matrix.
1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣 … (3.36)
1−𝑣
0 0
2
Where, E = Modulus of Elasticity or Young’s modulus
V = Poisson’s ratio

3.7.2 Plane strain

For plane strain, we assume the following strains to be zero.
𝑒 = = =0

The shear stresses 𝜏 =𝜏 = 0, but 𝜎 ≠ 0.

From equation (3.28), we know that,
𝜎 0 0 0 𝑒
⎡(1 − 𝑣) 𝑣 𝑣
0 0 0 ⎤⎧𝑒 ⎫
⎧𝜎 ⎫ ⎢ 𝑣 (1 − 𝑣) 𝑣
⎪𝜎 ⎪ E 0 0 0 ⎥⎪ 𝑒 ⎪
⎢ 𝑣 𝑣 (1 − 𝑣) 1 − 2𝑣 ⎥
⇒ 𝜏 = ⎢ 0 0 ⎥ 
⎨ ⎬ (1 + v)(1 − 2v) ⎢ 0 0 0 2 1 − 2𝑣 0 ⎥⎨ ⎬
𝜏
⎪ ⎪ ⎢
0 0 0 0 2 1 − 2𝑣 ⎥ ⎪ ⎪
⎩𝜏 ⎭ ⎣ 0 0 
0 0 0 2 ⎦⎩ ⎭
3.20 Two Dimensional Problems

In the above equation, 𝑒 = 0, so, delete third row and third column of [D] matrix.
 =0, so, delete fifth row and fifth column of [D] matrix.  =0, hence, delete sixth row
and sixth column of [D] matrix. The final reduced equation is,

𝜎 (1 − 𝑣) 𝑣 0 𝑒
E 𝑣 (1 − 𝑣) 0 𝑒
𝜎 = … (3.37)
𝜏 (1 + v)(1 − 2v) 1 − 2𝑣 
0 0
2
The above equation is in the form of
{}=[D]{e}
The above equation (3.37) gives a two dimensional stress-strain relationship for
plane strain problems.
Where, [D] = Stress-strain relationship or constitutive matrix.
(1 − 𝑣) 𝑣 0
[𝐷] =( 𝑣 (1 − 𝑣) 0 … (3.36)
)( )
0 0

Where, E  Young’s modulus

v  Poisson’s ratio

3.8 STIFNESS MATRIX EQUATION FOR TWO DIMENSIONAL

ELEMENT (CST ELEMENT)
We know that,

Stiffness matrix, [K] = [𝐵] [𝐷][𝐵]𝑑𝑉 [From chapter 2]

[K] = [𝐵] [𝐷][𝐵]𝑉

⇒ [K] = [𝐵] [𝐷][𝐵]𝐴 𝑡 [∵ 𝑉 = 𝐴 × 𝑡]
Stiffness matrix, [K] = [𝐵] [𝐷][𝐵]𝐴 𝑡 … (3.39)

1 1 𝑥 𝑦
Where, A ⟶ Area of the traingular element = 1 𝑥 𝑦
2 0 𝑥 𝑦
 Finite Element Analysis 3.21

t  Thickness of element
[B] = Strain – Displacement matrix

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 [From equation no. (3.15)]
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
Where 𝑞 = 𝑦 − 𝑦 ; 𝑞 =𝑦 −𝑦 ; 𝑞 =𝑦 −𝑦
𝑟 =𝑥 −𝑥 ; 𝑟 =𝑥 −𝑥 ; 𝑟 = 𝑥 −𝑥
[ D ] = Stress-Strain relationship matrix
For plane stress problems,
1 𝑣 0
[𝐷] = 𝑣 1 0 [From equation no. (3.36)]
0 0

For plane strain problems,

(1 − 𝑣) 𝑣 0
E 𝑣 (1 − 𝑣) 0
[𝐷] =
(1 + v)(1 − 2v) 1 − 2𝑣
0 0
2
[From equation no. (3.38)]
Where, E = Young’s modulus or Modulus of Elasticity
V = Poisson’s ratio

3.9. LINEAR STRAIN TRIANGULAR (LST) ELEMENT

A six noded triangular element is known as Linear Strain Triangular (LST)
element which is shown in Fig.3.15. It has twelve unknown displacement degrees of
freedom. The displacement functions of the element are quadratic instead of linear as in
the CST.
The procedures for development of the stiffness matrix equations for the LST
element follow the same steps as those used for the CST element. But the number of
equations used for developing shift matrix equation is 12 instead of 6. It is a tedious
process to solve those equations. Hence, we will use a computer, to solve many of the
mathematical equations.
3.22 Two Dimensional Problems

Fig. 3.15. Linear strain triangular element

LST element is preferred than the CST element for plane stress applications when
relatively small numbers of nodes are used. LST element is not preferred when large
numbers of nodes are used since the cost of formation of the element stiffnesses, equation
bandwidth are high compared to CST element. Computer modelling for large number of
nodes are also difficult for LST element.

3.10. FORMULAE USED

1. For constant strain triangle (CST) element
Shape function, N1 + N2 + N3 = 1
Co-ordinate, x = 𝑁 𝑥 + 𝑁 𝑥 + 𝑁 𝑥
Co-ordinate, y = 𝑁 𝑦 + 𝑁 𝑦 + 𝑁 𝑦
Or
Co-ordinate, x = (𝑥 − 𝑥 )𝑁 + (𝑥 − 𝑥 )𝑁 + 𝑥
Co-ordinate, y = (𝑦 − 𝑦 )𝑁 + (𝑦 − 𝑦 )𝑁 + 𝑦

1 1 𝑥 𝑦
2. Area of the traingular element, A= 1 𝑥 𝑦
2 0 𝑥 𝑦
3. Strain – Displacement matrix for CST element is,
 Finite Element Analysis 3.23

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞

Where 𝑞 = 𝑦 − 𝑦 ; 𝑞 =𝑦 −𝑦 ; 𝑞 =𝑦 −𝑦

𝑟 =𝑥 −𝑥 ; 𝑟 =𝑥 −𝑥 ; 𝑟 = 𝑥 −𝑥
4. Stress-Strain relationship matrix for plane stress problem,
1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣
1−𝑣
0 0
2
Where, v  Poisson’s ratio
E  Young’s modulus
5. Stress-Strain relationship matrix for plane stress problem,
(1 − 𝑣) 𝑣 0
E 𝑣 (1 − 𝑣) 0
[𝐷] =
(1 + v)(1 − 2v) 1 − 2𝑣
0 0
2
6. Element Stiffness matrix for CST element,
[K] = [𝐵] [𝐷][𝐵]𝐴 𝑡

7. Element stress, {  } = [ D ] { e }
𝑢
⎧𝑣 ⎫
𝜎 ⎪𝑢 ⎪
𝜎 = [ 𝐷 ][ 𝐵 ] 𝑣
𝜏 ⎨ ⎬
⎪𝑣 ⎪
⎩𝑣 ⎭

Where, 𝜎 , 𝜎  Normal stresses

𝜏  Shear stress

𝑢, 𝑣  Nodal displacements
3.24 Two Dimensional Problems

8. Maximum normal stress,

𝜎 +𝜎 𝜎 −𝜎
𝜎 =𝜎 = + +𝜏
2 2

Minimum normal stress,

𝜎 +𝜎 𝜎 −𝜎
𝜎 =𝜎 = + +𝜏
2 2

9. Principal angle,
2𝜏
tan 2𝜃 =
𝜎 −𝜎
10. Element strain,
𝑢
⎧𝑣 ⎫
⎪𝑢 ⎪
{𝑒} = [ 𝐵 ]{ 𝑢 } = [ 𝐵 ] 𝑣
⎨ ⎬
⎪𝑣 ⎪
⎩𝑣 ⎭

3.11. SOLVED PROBLEMS-CST ELEMENTS

Example 3.1
Determine the shape functions N1, N2 and N3 at the interior point P for the
triangular element shown in Fig.(i).

Fig. (i)
 Finite Element Analysis 3.25

Given
𝑥 = 2, 𝑦 = 3
𝑥 = 7, 𝑦 = 4
𝑥 = 4, 𝑦 = 7
𝑥 = 3.5, 𝑦 = 5

To find: Shape functions N1, N2 and N3 at the interior point, P.

Solution: We know that,
x = (𝑥 − 𝑥 )𝑁 + (𝑥 − 𝑥 )𝑁 + 𝑥 …(1)
y = (𝑦 − 𝑦 )𝑁 + (𝑦 − 𝑦 )𝑁 + 𝑦 …(2)
Substitute the co-ordinates values,
(1)  3.5 = (2 − 4)N + (7 − 4)N + 4 …(3)
(2)  5 = (3 − 7)N + (4 − 7)N + 7 …(4)
Equation (3) becomes,
3.5 = −2𝑁 + 3𝑁 + 4
 -0.5 = −2𝑁 + 3𝑁
 2𝑁 − 3𝑁 = 0.5 …(5)
Equation (3) becomes, 5 = −4𝑁 − 3𝑁 + 7
 -2 = −4𝑁 − 3𝑁
 4𝑁 + 3𝑁 = 2 …(6)
3.26 Two Dimensional Problems

Solving equation (5) and (6),

2𝑁 − 3𝑁 = 0.5
4𝑁 + 3𝑁 = 2
Solving 6N1=2.5
N1=0.4166
Substituting N1 value in equation (5) or equation (6),
2𝑁 − 3𝑁 = 0.5
2× 0.4166 − 3𝑁 = 2
 N2 = 0.11111
We know that, N1 + N2 N3 = 1
 0.4166 +0.1111+N3=1
 N3=0.4723
Result: Shape functions at the interior point, P.
N1= 0.4166
N2 = 0.11111
N3 = 0.4723

Example 3.2
The nodal co-ordinates of the triangular element are shown in Fig.(i). At the
interior point P, the x co-ordinate is 3.5 and N₁ = 0.4, calculate N2, N3 and the y co-
ordinate at point P.
 Finite Element Analysis 3.27

Given
𝑥 = 1, 𝑦 = 2
𝑥 = 4, 𝑦 = 3
𝑥 = 6, 𝑦 = 4
𝑥 = 3.5, 𝑦 = 0.4

To find: 1. Shape functions N2 and N3

2. Co-ordinate y.
Solution: We know that, Sum of shape function is equal to 1.
⇒𝑁 +𝑁 +𝑁 =1

⇒ 0.4 + 𝑁 + 𝑁 = 1

⇒ 𝑁 + 𝑁 = 0.6

We know that, Co-ordinate, x = 𝑁 𝑥 + 𝑁 𝑥 + 𝑁 𝑥

3.5 = 0.4(1) + N × (4) + 𝑁 × (6)

3.1 = 4 N + 6𝑁

⇒ 4𝑁 + 6𝑁 = 3.1 …(2)

Solving equation (1) and (2,

-4𝑁 − 4𝑁 = -2.4
3.28 Two Dimensional Problems

4𝑁 + 6𝑁 = 3.1

2N3 = 0.7
N3= 0.35
Substituting N3 value in equation (1) or (2),
𝑁 + 𝑁 = 0.6
𝑁 + 0.35 = 0.6
N2 = 0.25
We know that, Co-ordinate, y = 𝑁 𝑦 + 𝑁 𝑦 + 𝑁 𝑦
= 0.4  2 + 0.25  3 + 0.35  4
y = 2.95
Result: 1. Shape functions
N2 = 0.11111
N3 = 0.4723
2. Co-ordinate, y = 2.95

Example 3.3
Determine the x and y co-ordinates of point P for the triangular element shown in
Fig.(i). The shape functions N1 and N₂ are 0.2 and 0.3 respectively.

Fig. (i)
Given
𝑥 = 1, 𝑦 =1 𝑥 = 4, 𝑦 =2
 Finite Element Analysis 3.29

𝑥 = 3, 𝑦 =6 𝑁 = 0.2, 𝑁 = 0.3

To find: x and y Co-ordinate of point P.

Solution: We know that,
Sum of shape function is equal to 1.
⇒𝑁 +𝑁 +𝑁 =1

⇒ 0.2 + 0.3 + 𝑁 = 1

⇒ 𝑁 = 0.5

x Co-ordinate at point P is,

x=𝑁 𝑥 + 𝑁 𝑥 +𝑁 𝑥

= 0.2(1) + 0.3 × (4) + 0.5 × (3)

x = 2.9

y Co-ordinate at point P is,

y=𝑁 𝑦 + 𝑁 𝑦 +𝑁 𝑦

= 0.2(1) + 0.3 × (2) + 0.5 × (6)

y = 3.8

Result: Co-ordinate of point P

x = 2.9

y = 3.8
3.30 Two Dimensional Problems

Example 3.4
For the constant strain triangular element shown in Fig.(i), assemble strain-
displacement matrix. Take t=20 mm and E = 2 x 105 N/mm².

Fig. (i)
Given:

Fig. (ii)
𝑥 = 100, 𝑦 = 100
𝑥 = 400, 𝑦 = 100
𝑥 = 200, 𝑦 = 400
Young’s modulus, E = 2  105 N/mm2
Thickness, t = 20 mm
To find: Strain – Displacement matrix [ B ]
Solution: We know that,
Strain – Displacement matrix
 Finite Element Analysis 3.31

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 … (1)
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
[From equation no. (3.15)]
Where 𝑞 = 𝑦 − 𝑦 = 100 − 400 = −300
𝑞 = 𝑦 − 𝑦 = 400 − 100 = 300
𝑞 = 𝑦 − 𝑦 = 100 − 100 = 0
𝑟 = 𝑥 − 𝑥 = 200 − 400 = −200
𝑟 = 𝑥 − 𝑥 = 100 − 400 = −200
𝑟 = 𝑥 − 𝑥 = 400 − 100 = 300
Substitute the above values in equation (1)
⇒ [𝐵] =
−300 0 300 0 0 0
0 −200 0 −100 0 300 … (2)
−200 −300 −100 300 300 0
Where, A = Area of the element

1 1 𝑥 𝑦
= 1 𝑥 𝑦
2 0 𝑥 𝑦

1 1 100 100
= 1 400 100
2
0 200 400
1
= × [1(400 × 400 − 200 × 100) − 100(400 × 1 − 100 × 1)
2
+ 100(200 × 1 − 400 × 1)
A = 45, 000 mm2
Substitute a value in equation (2),

1 −300 0 300 0 0 0
⇒ [𝐵] = 0 −200 0 −100 0 300
2A
−200 −300 −100 300 300 0
3.32 Two Dimensional Problems

1 −3 0 3 0 0 0
[𝐵] = 0 −2 0 −1 0 3
900
−2 −3 −1 3 3 0
Result: Strain – Displacement matrix

1 −3 0 3 0 0 0
[𝐵] = 0 −2 0 −1 0 3
900
−2 −3 −1 3 3 0

Example 3.5
Determine the stiffness matrix for the constant strain triangular (CST) element
shown in Fig.(i). The co-ordinates are given in units of millimeters. Assume plane stress
conditions. Take E = 210 GPa, v=0.25 and t = 10 mm

Fig. (i)
Given:
𝑥 = 20, 𝑦 = 30
𝑥 = 80, 𝑦 = 30
𝑥 = 50, 𝑦 = 120

Fig. (ii)
 Finite Element Analysis 3.33

Young’s modulus, E = 210 Gpa = 2  109 Pa

= 210  109 N/m2 = 210  103 N/mm2
E = 2.1  105 N/mm2
Poisson’s ratio v = 0.25
Thickness, t = 10 mm
Assume plane stress condition.
To find: Stiffness matrix [K]
Solution: We know that,
Stiffness matrix, [ K ] = [ B ]T [ D ] [ B ] A t …(1)
[From equation no. (3.39)]
Where, A = Area of the element

1 1 𝑥 𝑦
= 1 𝑥 𝑦
2 0 𝑥 𝑦

1 1 20 30
= 1 80 30
2
0 50 120
1
= × [1(80 × 120 − 50 × 30) − 20(120 − 30) + 30(50 − 80)]
2
1
= × [8100 − 1800 − 900]
2
A = 2700 mm2 …(2)
We know that,
Strain – Displacement matrix

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 … (3)
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
[From equation no. (3.15)]
Where 𝑞 = 𝑦 − 𝑦 = 30 − 120 = −90
3.34 Two Dimensional Problems

𝑞 = 𝑦 − 𝑦 = 120 − 30 = 90
𝑞 = 𝑦 − 𝑦 = 30 − 30 = 0
𝑟 = 𝑥 − 𝑥 = 50 − 80 = −30
𝑟 = 𝑥 − 𝑥 = 20 − 50 = −30
𝑟 = 𝑥 − 𝑥 = 80 − 20 = 60
Substitute the above values in equation (1)

1 −90 0 90 0 0 0
⇒ [𝐵] = 0 −30 0 −30 0 60
2A
−30 −90 −30 90 60 0
Substitute area, A value,

1 −90 0 90 0 0 0
= 0 −30 0 −30 0 60
2 × 2700
−30 −90 −30 90 60 0
30 −3 0 3 0 0 0
= 0 −1 0 −1 0 2
2 × 2700
−1 −3 −1 3 2 0
−3 0 3 0 0 0
[𝐵] = 5.555 × 10 0 −1 0 −1 0 2 … (4)
−1 −3 −1 3 2 0
We know that,
Stress-Strain relationship matrix [ D ] for plane stress problem is
1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣 [From equation no. (3.36)]
1−𝑣
0 0
2
1 0.25 0
2.1 × 10 0.25 1 0
= 1 − 0.25
1 − (0.25)
0 0
2
2.1 × 10 1 0.25 0
= 0.25 1 0
0.9375
0 0 0.375
 Finite Element Analysis 3.35

2.1 × 10 × 0.25 4 1 0
= 1 4 0
0.9375
0 0 1.5
4 1 0
[𝐷] = 56 × 10 1 4 0
0 0 1.5
4 1 0
⇒ [𝐷][𝐵] = 56 × 10 1 4 0 × 5.555
0 0 1.5
−3 0 3 0 0 0
× 10 0 −1 0 −1 0 2
−1 −3 −1 3 2 0
4 1 0 −3 0 3 0 0 0
= 311.08 1 4 0 0 −1 0 −1 0 2
0 0 1.5 −1 −3 −1 3 2 0
−12 + 0 + 0 0 − 1 + 0 12 + 0 + 0 0−1+0 0+0+0 0+2+0
= 311.08 −3 + 0 + 0 0−4+0 3+0+0 0−4+0 0+0+0 0+8+0
0 + 0 − 1.5 0 + 0 − 4.5 0 + 0 − 1.5 0 + 0 + 4.5 0 + 0 + 3 0+0+0

𝑁𝑜𝑡𝑒: [(3 × 3) × (3 × 6) = 3 × 6]

−12 −1 12 −1 0 2
= 311.08 −3 −4 3 −4 0 8 … (6)
−1.5 −4.5 −1.5 4.5 3 0

We know that,

−3 0 3 0 0 0
[𝐵] = 5.555 × 10 0 −1 0 −1 0 2 [From equation no. 4]
−1 −3 −1 3 2 0
−3 0 −1
⎡0 −1 −3⎤
⎢3 0 −1⎥⎥
[𝐵] = 5.555 × 10 ⎢
⎢0 −1 3 ⎥
⎢0 0 2⎥
⎣0 2 0⎦
3.36 Two Dimensional Problems

⇒ [𝐵] [𝐷][𝐵]
−3 0 −1
⎡
0 −1 −3⎤
3 ⎢
0 −1⎥⎥
= 5.555 × 10 ⎢
0 −1 3 ⎥
⎢
0 0
⎢ 2⎥
0 ⎣
2 0⎦
−12 −1 12 −1 0 2
× 311.08 −3 −4 3 −4 0 8
−1.5 −4.5 −1.5 4.5 3 0
= 5.555 × 10
36 + 0 + 1.5 3 + 1 + 4.5 −36 + 0 + 1.5 3 + 0 − 4.5 0+0−3 −6 + 0 + 0
⎡ 0 + 3 + 4.5 0 + 4 + 13.5 0 − 3 + 4.5 0−4+0 0+0−9 0−8+0 ⎤
⎢ 0+0−3 6 + 0 + 0 ⎥⎥
−36 + 0 − 1.5 −3 + 0 + 4.5 36 + 0 + 1.5 −3 + 0 − 4.5
× 311.08 ⎢
⎢ 0 + 3 − 4.5 0 + 4 − 13.5 0 − 3 − 4.5 0 + 4 + 13.5 0+0+9 0−8+0 ⎥
⎢ 0+0−3 0+0−9 0+0−3 0+0+9 0+0+6 0+0+0 ⎥
⎣ 0−6+0 0−8+0 0+6+0 0−8+0 0+0+0 0 + 16 + 0 ⎦

37.54 7.5 −34.5 −1.5 −3 −6

⎡ 7.5 17.5 1.5 −9.5 −9 −8⎤
⎢
[𝐵] [𝐷][𝐵] = 1.78 ⎢ 34.5 1.5 37.5 −7.5 −3 6 ⎥⎥
⎢ −1.5 −9.5 −7.5 17.5 9 −8⎥
⎢ −3 −9 −3 9 6 0⎥
⎣ −6 −8 6 −8 0 16 ⎦
Substitute [𝐵] [𝐷][𝐵] and A, t values in equation (1),
Stiffness matrix [K]
37.54 7.5 −34.5 −1.5 −3 −6
⎡ 7.5 17.5 1.5 −9.5 −9 −8⎤
⎢ 34.5 1.5 37.5 −7.5 −3 6 ⎥⎥
= 1.78 ⎢ × 2700 × 10𝑁/𝑚𝑚
⎢ −1.5 −9.5 −7.5 17.5 9 −8⎥
⎢ −3 −9 −3 9 6 0⎥
⎣ −6 −8 6 −8 0 16 ⎦
37.54 7.5 −34.5 −1.5 −3 −6
⎡ 7.5 17.5 1.5 −9.5 −9 −8⎤
⎢ 34.5
[K] = 46.656 × 10 ⎢ 1.5 37.5 −7.5 −3 6 ⎥⎥
𝑁/𝑚𝑚
⎢ −1.5 −9.5 −7.5 17.5 9 −8⎥
⎢ −3 −9 −3 9 6 0⎥
⎣ −6 −8 6 −8 0 16 ⎦
 Finite Element Analysis 3.37

Result: Stiffness matrix [ K ]

37.54 7.5 −34.5 −1.5 −3 −6
⎡ 7.5 17.5 1.5 −9.5 −9 −8⎤
⎢ 34.5
[K] = 46.656 × 10 ⎢ 1.5 37.5 −7.5 −3 6 ⎥⎥
𝑁/𝑚𝑚
⎢ −1.5 −9.5 −7.5 17.5 9 −8⎥
⎢ −3 −9 −3 9 6 0⎥
⎣ −6 −8 6 −8 0 16 ⎦

Example 3.6
For the plane stress element shown in Fig.(i), the nodal displacements are:
u1 = 2.0 mm ; v1 = 1.0 mm

u2 = 0.5 mm ; v2 = 0.0 mm

u3 = 3.0 mm ; v3 = 1.0 mm

Fig. (i)
Determine the element stresses σx, σy, τxy, σ1, and σ2, and the principal angle θp
Let E=210 GPa, v=0.25 and t = 10 mm. All co-ordinates are in millimeters.
Given: Nodal displacements:
u1 = 2.0 mm ; v1 = 1.0 mm

u2 = 0.5 mm ; v2 = 0.0 mm

u3 = 3.0 mm ; v3 = 1.0 mm
3.38 Two Dimensional Problems

Fig. (ii)
𝑥 = 20 𝑚𝑚; 𝑦 = 30 mm
𝑥 = 80 𝑚𝑚 𝑦 = 30 mm
𝑥 = 50𝑚𝑚 𝑦 = 120 mm
Young’s modulus, E = 210 GPa = 2  109 Pa
= 210  109 N/m2 = 210  103 N/mm2
E = 2.1  105 N/mm2
Poisson’s ratio v = 0.25
Thickness, t = 10 mm
To find: 1. Element stresses
(a) Normal stress, 𝜎
(b) Normal stress, 𝜎
(c) Shear stress, 𝜏
(d) Maximum normal stress, 𝜎
(e) Minimum normal stress, 𝜎
2. Principal angle, 𝜃
Solution: We know that,
A = Area of the element

1 1 𝑥 𝑦 1 1 20 30
= 1 𝑥 𝑦 = 1 80 30
2 0 𝑥 𝑦 2
0 50 120
 Finite Element Analysis 3.39
1
= × [1(80 × 120 − 50 × 30) − 20(120 − 30) + 30(50 − 80)]
2
1
= × [8100 − 1800 − 900]
2
A = 2700 mm2 …(1)
We know that,
Strain – Displacement matrix

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 … (2)
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
[From equation no. (3.15)]
Where 𝑞 = 𝑦 − 𝑦 = 30 − 120 = −90
𝑞 = 𝑦 − 𝑦 = 120 − 30 = 90
𝑞 = 𝑦 − 𝑦 = 30 − 30 = 0
𝑟 = 𝑥 − 𝑥 = 50 − 80 = −30
𝑟 = 𝑥 − 𝑥 = 20 − 50 = −30
𝑟 = 𝑥 − 𝑥 = 80 − 20 = 60
Substitute the above values in equation (2)

1 −90 0 90 0 0 0
⇒ [𝐵] = 0 −30 0 −30 0 60
2A
−30 −90 −30 90 60 0
Substitute area, A value,

1 −90 0 90 0 0 0
⇒ [𝐵] = 0 −30 0 −30 0 60
2 × 2700
−30 −90 −30 90 60 0
30 −3 0 3 0 0 0
= 0 −1 0 −1 0 2
2 × 2700
−1 −3 −1 3 2 0
−3 0 3 0 0 0
[𝐵] = 5.555 × 10 0 −1 0 −1 0 2 … (3)
−1 −3 −1 3 2 0
3.40 Two Dimensional Problems

We know that,
Stress-Strain relationship matrix [ D ] for plane stress problem is
1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣 [From equation no. (3.36)]
1−𝑣
0 0
2
1 0.25 0
2.1 × 10 0.25 1 0
= 1 − 0.25
1 − (0.25)
0 0
2
2.1 × 10 1 0.25 0
= 0.25 1 0
0.9375
0 0 0.375
2.1 × 10 × 0.25 4 1 0
= 1 4 0
0.9375
0 0 1.5
4 1 0
[𝐷] = 56 × 10 1 4 0 …(4)
0 0 1.5
4 1 0
⇒ [𝐷][𝐵] = 56 × 10 1 4 0 × 5.555
0 0 1.5
−3 0 3 0 0 0
× 10 0 −1 0 −1 0 2
−1 −3 −1 3 2 0
4 1 0 −3 0 3 0 0 0
= 311.08 1 4 0 0 −1 0 −1 0 2
0 0 1.5 −1 −3 −1 3 2 0
−12 + 0 + 0 0 − 1 + 0 12 + 0 + 0 0−1+0 0+0+0 0+2+0
= 311.08 −3 + 0 + 0 0−4+0 3+0+0 0−4+0 0+0+0 0+8+0
0 + 0 − 1.5 0 + 0 − 4.5 0 + 0 − 1.5 0 + 0 + 4.5 0 + 0 + 3 0 + 0 + 0
−12 −1 12 −1 0 2
[𝐷][𝐵] = 311.08 −3 −4 3 −4 0 8 … (5)
−1.5 −4.5 −1.5 4.5 3 0
We know that,
stress, {  } = [ D ] { e }
 Finite Element Analysis 3.41
𝑢
⎧𝑣 ⎫
⎪𝑢 ⎪
= [ 𝐷 ][ 𝐵 ] 𝑣
⎨ ⎬
⎪𝑣 ⎪
⎩𝑣 ⎭
2
⎧1⎫
−12 −1 12 −1 0 2 ⎪ ⎪
0.5
= 311.08 −3 −4 3 −4 0 8 ×
−1.5 −4.5 −1.5 4.5 3 0 ⎨0.0⎬
⎪3⎪
⎩1.0⎭
(−12 × 2) + (−1 × 1) + (12 × 0.5) − (1 × 0) + (0 × 3) + (2 × 1)
= 311.08 (−3 × 2) − (4 × 1) + (3 × 0.5) − (4 × 0) + (0 × 8) + (8 × 1)
−(1.5 × 2) − (4.5 × 1) − (1.5 × 0.5) + (4.5 × 0) + (3 × 3) + (0 × 1)
𝑁𝑜𝑡𝑒: [(3 × 6) × (6 × 1) = 3 × 1]

−17
{  } = 311.08 −0.5
0.75
𝜎 −5288.36
𝜎 = −155.54 …(6)
𝜏 233.31

 Normal stress, 𝜎 = -5288.36 N/mm2

Normal stress, 𝜎 = -155.54 N/mm2

Shear stress, 𝜏 = 233.21 N/mm2
We know that,
Maximum normal stress,

𝜎 +𝜎 𝜎 −𝜎
𝜎 =𝜎 = + +𝜏 … (7)
2 2

−5288.36 − 155.54 −5288.36 − 155.54

= + + (233.31)
2 2
3.42 Two Dimensional Problems

𝜎 = −144.956 𝑁/𝑚𝑚
Minimum normal stress,

𝜎 +𝜎 𝜎 −𝜎
𝜎 =𝜎 = + +𝜏 … (7)
2 2

−5288.36 − 155.54 −5288.36 − 155.54

= + + (233.31)
2 2

𝜎 = −5298.9 𝑁/𝑚𝑚
We know that, Principal angle,
2𝜏
tan 2𝜃 = … (9)
𝜎 −𝜎
2𝜏
 2𝜃 = tan
𝜎 −𝜎
2 × 233.31
= tan
−5288.36 − 155.54
2𝜃 = −5.194
 𝜃 = −2.59
Result: 1. Element stresses
(a) Normal stress, 𝜎 = - 5288.36 N/mm2
(b) Normal stress, 𝜎 = - 155.54 N/mm2
(c) Shear stress, 𝜏 = 233.31 N/mm2
(d) Maximum normal stress, 𝜎 = - 144.956 N/mm2
(e) Minimum normal stress, 𝜎 = - 5298.9 N/mm2
2. Principal angle, 𝜃 = -2.59o

Example 3.7
For the triangular element shown in Fig.(i). Obtain the strain- displacement
relation matrix [B] and determine the strains ex, ey and 
 Finite Element Analysis 3.43

Fig.(i)
Nodal Displacements are
u1 = 0.001 ; v1 = -0.004
u2 = 0.003 ; v2 = 0.002
u3 = -0.002 ; v3 = 0.005 All coordinates are in millimetres
Given: Nodal Displacements:
u1 = 0.001 ; v1 = -0.004
u2 = 0.003 ; v2 = 0.002
u3 = -0.002 ; v3 = 0.005
𝑥 = 1 𝑚𝑚; 𝑦 = 1 𝑚𝑚
𝑥 = 8 𝑚𝑚 𝑦 = 4 𝑚𝑚
𝑥 = 2 𝑚𝑚 𝑦 = 7 𝑚𝑚
To find: Element stresses
(a) Normal stress, 𝑒
(b) Normal stress, 𝑒
(c) Shear stress, 𝛾
Solution: We know that,
A = Area of the element

1 1 𝑥 𝑦 1 1 1 1
= 1 𝑥 𝑦 = 1 8 4
2 0 𝑥 𝑦 2
1 2 7
3.44 Two Dimensional Problems
1
= × [1(56 − 8) − 1(7 − 4) + 1(2 − 8)]
2

A = 19.5 mm …(1)
We know that,
Strain – Displacement matrix

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 … (2)
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
Where 𝑞 = 𝑦 − 𝑦 = 4 − 7 = −3

𝑞 =𝑦 −𝑦 = 7−1= 6

𝑞 = 𝑦 − 𝑦 = 1 − 4 = −3

𝑟 = 𝑥 − 𝑥 = 2 − 8 = −6

𝑟 = 𝑥 − 𝑥 = 1 − 2 = −1

𝑟 = 𝑥 −𝑥 = 8−1 =7
Substitute the above values in equation (2)

1 −3 0 6 0 −3 0
⇒ [𝐵] = 0 −6 0 −1 0 7
2A
−6 −3 −1 6 7 −3
Substitute area, A value,

1 −3 0 6 0 −3 0
⇒ [𝐵] = 0 −6 0 −1 0 7
2 × 19.5
−6 −3 −1 6 7 −3
We know that,
𝑢
⎧𝑣 ⎫
⎪𝑢 ⎪
Element strain, { 𝑒 } = [𝐵 ][ 𝑢 ] 𝑣
⎨ ⎬
⎪𝑣 ⎪
⎩𝑣 ⎭
 Finite Element Analysis 3.45
0.001
⎧−0.004⎫
1 −3 0 6 0 −3 0 ⎪
0.003
⎪
= 0 −6 0 −1 0 7 ×
39
−6 −3 −1 6 7 −3 ⎨ 0.002 ⎬
⎪−0.002⎪
⎩ 0.005 ⎭
(−3 × 0.001) + 0 + (6 × 0.003) + 0 + (3 × 0.002) + 0
1
= 0 + (6 × 0.004) + 0 − 1(1 × 0.002) + 0 + (7 × 0.005)
39
(−6 × 0.001) + (3 + 0.004) − (1 × 0.003) + (6 × 0.002) + 7(−0.002) + (−3 × 0.005)

1 0.0021
{𝑒} = 0.057
39
−0.014
𝑒 5.38 × 10
 𝑒 = 1.4615 × 10
𝛾 3.589 × 10

 𝑒 = 5.38  10-4
𝑒 = 1.4615  10-3

𝛾 = 3.589  10-4
Result: Element stresses
(a) Normal stress, 𝑒 = 5.38  10-4
(b) Normal stress, 𝑒 = 1.4615  10-3

(c) Shear stress, 𝛾 = 3.589  10-4

Example 3.8
The two dimensional propped beam shown in Fig.(i) is divided into two CST
elements. Determine the nodal displacements and element stresses using plane stress
conditions. Body force is neglected in comparison with the external forces.
Take, Thickness, t = 10 mm
Young's modulus, E = 2 × 105 N/mm²
Poisson's ratio, v = 0.25
3.46 Two Dimensional Problems

Fig. (i)
Given:

Fig. (ii)
Thickness, t = 10 mm
Young's modulus, E = 2 × 105 N/mm²
Poisson's ratio, v = 0.25
To find: (i) Nodal displacements 𝑢 𝑣 , 𝑢 𝑣 , 𝑢 , 𝑣 and 𝑢 , 𝑣
(ii) Element stress, 𝜎 and 𝜎
Solution:
Consider element (1): (Nodal displacements 𝑢 𝑣 , 𝑢 , 𝑣 , 𝑢 , 𝑣 )
 Finite Element Analysis 3.47

Take node 1 as origin

𝑥 𝑦
For node 1: (0, 0)
𝑥 𝑦
For node 3: (1000, 750)
𝑥 𝑦
For node 4: (0, 750)
Solution: We know that,
Stiffness matrix [ K ] = [B]T [D] [B] A t …(1)
[From equation no. (3.39)]
Where, A = Area of the element

1 1 𝑥 𝑦 1 1 1 1
= 1 𝑥 𝑦 = 1 1000 750
2 0 𝑥 𝑦 2
1 0 750

1 1000 × 750
= × 1 × (1000 × 750 − 0) =
2 2

A = 375 103 mm2 …(2)

Strain – Displacement matrix

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 … (3)
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
[From equation no. (3.15)]
Where 𝑞 = 𝑦 − 𝑦 = 750 − 750 = 0
𝑞 = 𝑦 − 𝑦 = 750 − 0 = 750
𝑞 = 𝑦 − 𝑦 = 0 − 750 = −750
𝑟 = 𝑥 − 𝑥 = 0 − 1000 = −1000
𝑟 = 𝑥 −𝑥 = 0−0 =0
𝑟 = 𝑥 − 𝑥 = 1000 − 0 = 1000
3.48 Two Dimensional Problems

Substitute the above values in equation (3)

1 0 0 750 0 −750 0
⇒ [𝐵] = 0 −1000 0 0 0 1000
2A
−1000 0 0 750 1000 −750
Substitute area, A value,

1 0 0 750 0 −750 0
⇒ [𝐵] = 0 −1000 0 0 0 1000
2 × 37510
−1000 0 0 750 1000 −750
250 0 0 3 0 −3 0
[𝐵] = 0 −4 0 0 0 4
2 × 37510
−4 0 0 3 4 −3
Stress-Strain relationship matrix [ D ] for plane stress problem is
1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣 [From equation no. (3.36)]
1−𝑣
0 0
2
1 0.25 0
2.1 × 10 0.25 1 0
= 1 − 0.25
1 − (0.25)
0 0
2
2.1 × 10 × 1 0.25 0
= 0.25 1 0
0.9375
0 0 0.375

2.1 × 10 × 0.25 4 1 0
= 1 4 0
0.9375
0 0 1.5
4 1 0
[𝐷] = 2 × 10 × 0.2667 1 4 0 …(5)
0 0 1.5
4 1 0
⇒ [𝐷][𝐵] = 2 × 10 × 0.2667 1 4 0
0 0 1.5
250 0 0 3 0 −3 0
× 0 −4 0 0 0 4
2 × 375 × 10
−4 0 0 3 4 −3
 Finite Element Analysis 3.49

2 × 10 × 0.2667 × 250 4 1 0 −3 0 3 0 0 0
= 1 4 0 0 −1 0 −1 0 2
2 × 375 × 10
0 0 1.5 −1 −3 −1 3 2 0
0+0+0 0 − 4 + 0 12 + 0 + 0 0 + 0 + 0 −12 + 0 + 0 0 + 4 + 0
= 17.78 0 + 0 + 0 0 − 16 + 0 3 + 0 + 0 0+0+0 −3 + 0 + 0 0 + 16 + 0
0+0∓ 0+0+0 0+0+0 0 + 0 + 4.5 0+0+6 0 + 0 − 4.5

0 −4 12 0 −12 4
[𝐷][𝐵] = 17.78 0 −16 3 0 −3 16 … (6)
−6 0 0 4.5 6 −4.5

We know that,

250 0 0 3 0 −3 0
[𝐵] = 0 −4 0 0 0 4 [From equation no. (4)]
2 × 37510
−4 0 0 3 4 −3
0 0 −4
⎡ 0 −4 0 ⎤
250 ⎢3 0 0 ⎥⎥
[𝐵] = ⎢
2 × 37510 ⎢ 0 0 3⎥
⎢−3 0 4⎥
⎣0 4 −3⎦
0 0 −4
⎡ 0 −4 0 ⎤
⎢3 0 0 3 0 −3 0
250 0 0 ⎥⎥
⇒ [𝐵] [𝐷][𝐵] = × 17.78 ⎢ 0 −4 0 0 0 4
2 × 375 × 10 ⎢0 0 3⎥
⎢−3 0 4 ⎥ −4 0 0 3 4 −3
⎣0 4 −3⎦

24 0 0 −18 −24 18
⎡ 0 64 −12 0 12 −64 ⎤
⎢ 0 −12 36 0 −36 12 ⎥
= 5.927 × 10 ⎢ ⎥
⎢−18 0 0 13.5 18 −13.5 ⎥
⎢−24 12 −36 18 36 + 24 −12 − 18 ⎥
⎣ 18 −64 12 −13.5 −12 − 18 +64 + 13.5⎦
24 0 0 −18 −24 18
⎡ 0 64 −12 0 12 −64 ⎤⎥
⎢
3⎢ 0 −12 36 0 −36 12 ⎥
[𝐵] [𝐷][𝐵] = 5.927 × 10
⎢−18 0 0 13.5 18 −13.5⎥
⎢−24 12 −36 18 60 −30 ⎥
⎣ 18 −64 12 −13.5 −30 77.5 ⎦
3.50 Two Dimensional Problems

Substitute [𝐵] [𝐷][𝐵] and A, t values in equation (1),

Stiffness matrix, [𝐾]
24 0 0 −18 −24 18
⎡
⎢
0 64 −12 0 12 −64 ⎤⎥
3⎢ 0 −12 36 0 −36 12 ⎥ × 375 × 103 × 10
= 5.927 × 10
⎢−18 0 0 13.5 18 −13.5⎥
⎢−24 12 −36 18 60 −30 ⎥
⎣ 18 −64 12 −13.5 −30 77.5 ⎦
24 0 0 −18 −24 18
⎡ 0 64 −12 0 12 −64 ⎤⎥
⎢
4⎢ 0 −12 36 0 −36 12 ⎥
= 2.22 × 10
⎢−18 0 0 13.5 18 −13.5⎥
⎢−24 12 −36 18 60 −30 ⎥
⎣ 18 −64 12 −13.5 −30 77.5 ⎦
53.28 0 0 −39.96 −53.28 39.96
⎡
⎢
0 142.08 −26.64 0 26.64 −142.08⎤⎥
[𝐾] = 1 × 104⎢ 0 −26.64 79.92 0 −79.92 26.64 ⎥
⎢−39.96 0 0 29.97 39.96 −29.97 ⎥
⎢−53.28 26.64 −79.92 39.96 133.2 −66.6 ⎥
⎣ 39.96 −142.08 26.64 −29.97 −66.6 172.05 ⎦

For element (1), nodal displacements are 𝑢 𝑣 , 𝑢 , 𝑣 , 𝑢 , 𝑣 . [Refer Fig. (iii)]

𝑢 𝑣 𝑢 𝑣 𝑢 𝑣
24 0 0 −18 −24 18 𝑢
⎡ 0 64 −12 0 12 −64 𝑣⎤
⎢ 0 −12 36 0 −36 12 ⎥⎥ 𝑢
Stiffness matrix, [𝐾] = 1 × 10 ⎢
⎢−18 0 0 13.5 18 −13.5⎥ 𝑣
⎢−24 12 −36 18 60 −30 ⎥ 𝑣
⎣ 18 −64 12 −13.5 −30 77.5 ⎦ 𝑣
…(7)
Consider element (2): (Nodal displacements 𝑢 𝑣 , 𝑢 , 𝑣 , and 𝑢 , 𝑣 )

Fig. (iv)
 Finite Element Analysis 3.51

Take node 1 as origin

𝑥 𝑦
For node 1: (0, 0)
𝑥 𝑦
For node 2: (1000, 0)
𝑥 𝑦
For node 3: (1000, 750)
We know that,
Stiffness matrix for element (2) [ K ]2 = [B]T [D] [B] A t …(8)
Where, A = Area of the triangular element

1 1 𝑥 𝑦 1 1 0 0
= 1 𝑥 𝑦 = 1 1000 0
2 0 𝑥 𝑦 2
1 0 750

1 1000 × 750
= × 1 × (1000 × 750 − 0) =
2 2

A = 375 103 mm2

Strain – Displacement matrix

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 … (9)
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
Where 𝑞 = 𝑦 − 𝑦 = 0 − 750 = −750
𝑞 = 𝑦 − 𝑦 = 750 − 0 = 750
𝑞 =𝑦 −𝑦 = 0−0= 0
𝑟 = 𝑥 − 𝑥 = 1000 − 1000 = 0
𝑟 = 𝑥 − 𝑥 = 0 − 1000 = −1000
𝑟 = 𝑥 − 𝑥 = 1000 − 0 = 1000

1 −750 0 750 0 0 0
⇒ [𝐵] = 0 0 0 −1000 0 1000
2A
0 −750 −1000 750 1000 0
3.52 Two Dimensional Problems

Substitute area, A value,

1 −750 0 750 0 0 0
⇒ [𝐵] = 0 0 0 −1000 0 1000
2 × 375 × 10
0 −750 −1000 750 1000 0
250 −3 0 3 0 0 0
[𝐵] = 0 0 0 −4 0 4
2 × 375 × 10
0 −3 −4 3 4 0
Stress-Strain relationship matrix [ D ] for plane stress problem is
1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣
1−𝑣
0 0
2
1 0.25 0
2.1 × 10 0.25 1 0
= 1 − 0.25
1 − (0.25)
0 0
2

2.1 × 10 × 1 0.25 0
= 0.25 1 0
0.9375
0 0 0.375

2.1 × 10 × 0.25 4 1 0
= 1 4 0 … (11)
0.9375
0 0 1.5

250 × 2 × 10 × 0.25 4 1 0 −3 0 3 0 0 0
[𝐷][𝐵] = 1 4 0 0 0 0 −4 0 4
2 × 375 × 10 × 0.9375
0 0 1.5 0 −3 −4 3 4 0
−12 0 12 −4 0 4
= 17.78 −3 0 3 −16 0 16 … (12)
0 −4.5 −6 4.5 6 0

We know that,

250 −3 0 3 0 0 0
[𝐵] = 0 0 0 −4 0 4 [From equation no. (10)]
2 × 375 × 10
0 −3 −4 3 4 0
 Finite Element Analysis 3.53
−3 0 0
⎡ 0 0 −3⎤
250 ⎢ 3 0 −4⎥⎥
[𝐵] = ⎢
2 × 375 × 10 ⎢ 0 −4 3 ⎥
⎢ 0 0 4⎥
⎣ 0 4 0⎦
−3 0 0
⎡ 0 0 −3⎤ −12
⎢ 0 12 −4 0 4
[𝐵] [𝐷][𝐵] =
250
⎢ 3 0 −4⎥⎥
−3 0 3 −16 0 16
2 × 37510 ⎢ 0 −4 3 ⎥
⎢ 0 0 4⎥ 0 −4.5 −6 4.5 6 0
⎣ 0 4 0⎦
36 0 −36 12 0 −12
⎡ 0 13.5 18 −13.5 −18 0 ⎤
⎢−36 18 60 −30 −24 12 ⎥⎥
= 5.927 × 10 ⎢
⎢ 12 −13.5 −30 77.5 18 −64⎥
⎢ 0 −18 −24 18 24 0 ⎥
⎣ 12 0 12 −64 0 64 ⎦

Substitute [𝐵] [𝐷][𝐵] and A, t values in equation (8),

36 0 −36 12 0 −12
⎡
⎢ 0 13.5 18 −13.5 −18 0 ⎤⎥
3 ⎢−36 18 60 −30 −24 12 ⎥ × 375 ×
Stiffness matrix, [𝐾]2 = 5.927 × 10
⎢ 12 −13.5 −30 77.5 18 −64⎥
⎢ 0 −18 −24 18 24 0 ⎥
⎣ 12 0 12 −64 0 64 ⎦
103 × 10

36 0 −36 12 0 −12
⎡ 0 13.5 18 −13.5 −18 0 ⎤⎥
⎢
= 2.22 × 104 ⎢−36 18 60 −30 −24 12 ⎥
⎢ 12 −13.5 −30 77.5 18 −64⎥
⎢ 0 −18 −24 18 24 0 ⎥
⎣ 12 0 12 −64 0 64 ⎦
79.92 0 −79.92 26.64 0 −26.64
⎡ 0 29.97 39.96 −29.97 −39.96 0 ⎤
⎢
[𝐾] = 1 × 104 ⎢−79.92 39.96 133.2 −66.6 −53.28 26.64 ⎥⎥
⎢ 26.64 −29.97 −66.6 172.05 39.96 −142.08⎥
⎢ 0 −39.96 −53.28 39.96 53.28 0 ⎥
⎣−26.64 0 26.64 −142.08 0 142.08 ⎦
3.54 Two Dimensional Problems

For element (2), nodal displacements are 𝑢 𝑣 , 𝑢 , 𝑣 and 𝑢 , 𝑣 [Refer Fig. (iv)]
Stiffness matrix, [𝐾] 𝑢 𝑣 𝑢 𝑣 𝑢 𝑣
79.92 0 −79.92 26.64 0 −26.64 𝑢
⎡ 0 29.97 39.96 −29.97 −39.96 0 ⎤ 𝑣
⎢−79.92 39.96 133.2 −66.6 −53.28 26.64 ⎥⎥ 𝑢
= 1 × 10 ⎢ 𝑣
⎢ 26.64 −29.97 −66.6 172.05 39.96 −142.08⎥
⎢ 0 −39.96 −53.28 39.96 53.28 0 ⎥ 𝑢
⎣−26.64 0 26.64 −142.08 0 142.08 ⎦ 𝑣
…(13)
Global stiffness Matrix [ K ]
Assemble the stiffness matrix equations (7) and (13),
Global stiffness Matrix [ K ] =
𝑢 𝑣 𝑢 𝑣 𝑢 𝑣 𝑢 𝑣
0 0
53.28 -39.96
+ -79.92 26.64 + -53.28 39.96 𝑢
+79.92 -26.64
0 0
+ 142.08 -26.64 0
0+0 + 39.96 -29.97 + + 26.64 -142.08 𝑣
29.97 -39.96 0
-79.92 39.96 133.2 -66.6 -53.28 26.64 0 0 𝑢
1 104 26.64 -29.97 -66.6 172.05 39.96 -142.08 0 0 𝑣
0 -26.64 79.92 0
+ + -53.28 39.96 + + -79.92 26.64 𝑢
0 -39.96 53.28 0
-39.96 0 0 29.97
+ + 26.64 -142.08 + + 39.96 -29.97 𝑣
-26.64 0 0 142.08
-53.28 26.64 0 0 -79.92 39.96 133.2 -66.6 𝑢
39.96 -142.08 0 0 26.64 -29.97 -66.6 172.05 𝑣
 Finite Element Analysis 3.55

Global Stiffness matrix, [𝐾] =

𝑢 𝑣 𝑢 𝑣 𝑢 𝑣 𝑢 𝑣
133.2 0 −79.92 26.64 0 −66.6 −53.28 39.96 𝑢
⎡ 0 172.05 39.96 −29.97 −66.6 0 ⎤ 𝑣
⎢−79.92 26.64 −142.08⎥
39.96 133.2 −66.6 −53.28 26.64 0 0 𝑢
⎢ ⎥ 𝑣
1 × 10 ⎢ 26.64 −29.97 −66.6 −53.28 26.64 −142.08 0 0 ⎥
⎢ 0 −66.6 −53.28 39.96 133.2 0 −79.92 26.64 ⎥ 𝑣
⎢ −66.6 0 26.64 −142.08 0 172.08 39.96 −29.97 ⎥ 𝑣
⎢−53.28 26.64 0 0 −79.92 39.96 133.2 −66.6 ⎥ 𝑢
⎣ 39.96 −142.08 0 0 26.64 −29.97 −66.6 172.05 ⎦ 𝑣

We Know that, general force equation is

{ F } = [ K ] {u }
𝐹
⎧𝐹 ⎫ 133.2 0 −79.92 26.64 0 −66.6 −53.28 39.96 𝑢
⎡ 0 ⎤
⎪ ⎪ 172.05 39.96 −29.97 −66.6 0 26.64 −142.08 ⎧ 𝑣 ⎫
⎪𝐹 ⎪ ⎢−79.92 39.96 133.2 −66.6 −53.28 26.64 0 0 ⎥ ⎪𝑢 ⎪
⎪ ⎪ ⎢ ⎥ ⎪𝑣 ⎪
𝐹 26.64 −29.97 −66.6 −53.28 26.64 −142.08 0 0
= 1 × 10 ⎢ ⎥
⎨𝐹 ⎬ ⎢ 0 −66.6 −53.28 39.96 133.2 0 −79.92 26.64 ⎥ ⎨𝑢 ⎬
⎪𝐹 ⎪ ⎢ −66.6 0 26.64 −142.08 0 172.08 39.96 −29.97 ⎥ ⎪𝑣 ⎪
⎪𝐹 ⎪ ⎢−53.28 26.64 0 0 −79.92 39.96 133.2 −66.6 ⎥ ⎪𝑢 ⎪
⎪ ⎪ ⎣ 39.96 −142.08 0 0 26.64 −29.97 −66.6 172.05 ⎦ ⎩ 𝑣 ⎭
⎩𝐹 ⎭

…(14)
Applying boundary conditions [Refer Fig. (ii)]
1. Node, 1 and Node 4 are fixed. So, 𝑢 , 𝑣 and 𝑢 , 𝑣 are zero. i.e., 𝑢 = 𝑣 =
𝑢 = 𝑣 = 0.
2. Node 2, is moving in x direction. So 𝑢 ≠ 0 but, 𝑣 = 0.
3. At node 3, a point load of 75000N is acting in x direction. So, F3-x = 75,000N.
4. Body force is neglected. So, the remaining forces are zero.
i.e., 𝐹 =𝐹 =𝐹 =𝐹 =𝐹 =𝐹 =𝐹 =𝐹 = 0.

Substitute the above values in equation (14),

0 133.2 0 −79.92 26.64 0 −66.6 −53.28 39.96 0
⎧ ⎫ ⎡ 0 ⎤
0
⎢−79.92
172.05 39.96 −29.97 −66.6 0 26.64 −142.08⎥ ⎧ 0 ⎫
⎪ 0 ⎪ 39.96 133.2 −66.6 −53.28 26.64 0 0 ⎪𝑢 ⎪
⎪ ⎪ ⎢ ⎥⎪ ⎪
0 −66.6
= 1 × 10 ⎢ 26.64 −29.97 −53.28 26.64 −142.08 0 0 ⎥ 𝑢0
⎨75 × 10 ⎬ ⎢ 0 −66.6 −53.28 39.96 133.2 0 −79.92 26.64 ⎥ ⎨ ⎬
⎪ 0 ⎪ ⎢ −66.6 0 26.64 −142.08 0 172.08 39.96 −29.97 ⎥ ⎪𝑣 ⎪
⎪ 0 ⎪ ⎢−53.28 26.64 0 0 −79.92 39.96 133.2 −66.6 ⎥ ⎪ 0 ⎪
⎩ 0 ⎭ ⎣ 39.96 −142.08 0 0 26.64 −29.97 −66.6 172.05 ⎦ ⎩ 0 ⎭
3.56 Two Dimensional Problems

In the above equation 𝑢 , 𝑣 , 𝑣 , 𝑢 , 𝑣 are zero. So delete the corresponding row

and column of [K] matrix. Hence the equation reduces to
0 133.2 −53.28 26.64 𝑢
75 × 10 = 1 × 10 −53.28 133.2 0 𝑢
0 26.64 0 172.05 𝑣
0 133.2 −53.28 26.64 𝑢
⇒ 7.5 = −53.28 133.2 0 𝑢 …(15)
0 26.64 0 172.05 𝑣
0 133.2 −53.28 26.64 𝑢
⇒ 18.75 = 0 279.72 0 𝑢
18.75 0 0 172.05 𝑣

⇒ −4349.81 𝑣 = 18.75

⇒ 𝑣 = −0.00431 𝑚𝑚
⇒ 279.72𝑢 + 26.64𝑣 = 18.75

⇒ 279.72𝑢 + 26.64 × (−0.00431) = 18.75

⇒ 𝑢 = 0.067 𝑚𝑚

⇒ 133.2𝑢 − 53.28𝑢 + 26.64𝑣 = 0

⇒ 133.2𝑢 − 53.28(0.067) + 26.64(−0.00431) = 0

⇒ 𝑢 = 0.02766 𝑚𝑚

Nodal displacements:
𝑢 = 0𝑚𝑚 𝑣 = 0 𝑚𝑚
𝑢 = 0.02766𝑚𝑚 𝑣 = 0 𝑚𝑚
𝑢 = 0.067 𝑚𝑚 𝑣 = −0.00431 𝑚𝑚
𝑢 = 0 𝑚𝑚 𝑣 = 0 𝑚𝑚
Stress in each element:
We know that, stress, {  } = [D] [B] { u }
 Finite Element Analysis 3.57

For element (1): [Refer Fig. (iii)].: (Nodal displacements equation 𝒖𝟏 , 𝒗𝟏 , 𝒖𝟑 , 𝒗𝟑

and 𝒖𝟒 , 𝒗𝟒 )
𝑢
⎧𝑣 ⎫
0 −4 12 0 −12 4 ⎪𝑢 ⎪
⇒ stress, {  } = 17.78 0 −16 3 0 −3 16 𝑣
−6 0 0 4.5 6 −4.5 ⎨𝑢
⎬
⎪ ⎪
⎩𝑣 ⎭
[From equation no. (6)]
0
⎧ 0 ⎫
0 −4 12 0 −12 4 ⎪ ⎪
0.067
= 17.78 0 −16 3 0 −3 16 ×
−6 0 0 4.5 6 −4.5 ⎨−0.00431⎬
⎪ 0 ⎪
⎩ 0 ⎭
12 × 0.067
{  }1 = 17.78 3 × 0.067
4.5 × −0.00431
𝜎 14.295
𝜎 = 3.574 𝑁/𝑚𝑚
𝜏 −0.345

 Where, 𝜎 , 𝜎  Normal stress,

𝜏  Shear stress
For element (2): [Refer Fig. (iv)].: (Nodal displacements equation 𝒖𝟏 , 𝒗𝟏 , 𝒖𝟐 , 𝒗𝟐 and
𝒖 𝟑 , 𝒗𝟑 )
Stress, { 2 } = [D]2 [B]2 { u }
𝑢
⎧𝑣 ⎫
−12 0 12 −4 0 4 ⎪𝑢 ⎪
= 17.78 −3 0 3 −16 0 16 𝑣
0 −4.5 −6 4.5 6 0 ⎨ ⎬
⎪𝑢 ⎪
⎩𝑣 ⎭
[From equation no. (12)]
3.58 Two Dimensional Problems
0
⎧ 0 ⎫
−12 0 12 −4 0 4 ⎪ ⎪
0.02766
= 17.78 −3 0 3 −16 0 16 ×
⎨ 0 ⎬
0 −4.5 −6 4.5 6 0 0.067
⎪ ⎪
⎩−0.00431⎭
(12 × 0.02766) + (4 × −0.00431)
{  }2 = 17.78 (3 × 0.02766) + (16 × −0.00431)
(−6 × −0.00431) + (6 × 0.067)
𝜎 5.595
𝜎 = 0.2492 𝑁/𝑚𝑚
𝜏 4.196
Result:
1. Nodal displacements: 𝑢 = 0, 𝑣 = 0, 𝑢 = 0.2766𝑚𝑚; 𝑣 = 0, 𝑢 = 0.067𝑚𝑚;
𝑣 = −0.00431 𝑚𝑚; 𝑢 = 0 𝑎𝑛𝑑 𝑣 = 0.
𝜎 14.295
2. For element (1) 𝜎 = 3.574 𝑁/𝑚𝑚
𝜏 −0.345
𝜎 5.595
3. For element (2) 𝜎 = 0.2492 𝑁/𝑚𝑚
𝜏 4.196

Example 3.9
Determine the nodal displacements of nodes 1 and 2 and the element stresses for
the two dimensional loaded plate as shown in Fig.(i). Assume plane stress condition.
Take, v= 0.25, E = 2 x 105 N/mm², Thickness = 15 mm.

Fig. (i)
 Finite Element Analysis 3.59

Body force is neglected in comparison with external forces.

Given:

Fig. (ii)
Thickness, t = 10 mm
Young's modulus, E = 2 × 105 N/mm²
Poisson's ratio, v = 0.25
Point load acting at node 2 is 100 kN.
To find: (i) Nodal displacements at node 1 and 2. i.e., 𝑢 𝑣 , and 𝑢 𝑣
(ii) Element stress, 𝜎 and 𝜎
Solution:
Consider element (1): (Nodal displacements 𝑢 𝑣 , 𝑢 , 𝑣 , 𝑢 , 𝑣 )

Fig. (iii)
3.60 Two Dimensional Problems

Take node 1 as origin

𝑥 𝑦
For node 1: (0, 0)
𝑥 𝑦
For node 3: (0, 50)
𝑥 𝑦
For node 3: (-75, 0)
Solution: We know that,
Stiffness matrix [ K ] = [B]T [D] [B] A t …(1)
[From equation no. (3.39)]
Where, A = Area of the element

1 1 𝑥 𝑦 1 1 0 0
= 1 𝑥 𝑦 = 1 0 50
2 0 𝑥 𝑦 2
1 −75 0
1 3750
= × 1 × (0 × 75 × 50) =
2 2
A = 1875 mm2 …(2)
Strain – Displacement matrix

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 … (3)
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
[From equation no. (3.15)]
Where 𝑞 = 𝑦 − 𝑦 = 50 − 0 = 50
𝑞 =𝑦 −𝑦 = 0−0= 0
𝑞 = 𝑦 − 𝑦 = 0 − 50 = −50
𝑟 = 𝑥 − 𝑥 = −75 − 0 = −75
𝑟 = 𝑥 − 𝑥 = 0 + 75 = 75
𝑟 = 𝑥 −𝑥 = 0−0 =0
 Finite Element Analysis 3.61

Substitute the above values in equation (3)

1 50 0 0 0 −50 0
⇒ [𝐵] = 0 −75 0 75 0 0
2A
−75 50 75 0 0 −50
Substitute area, A value,

1 50 0 0 0 −50 0
⇒ [𝐵] = 0 −75 0 75 0 0
2 × 1875
−75 50 75 0 0 −50
50 1 0 0 0 −1 0
[𝐵] = 0 −1.5 0 1.5 0 0 … (4)
3750
−1.5 1 1.5 0 0 −1
Stress-Strain relationship matrix [ D ] for plane stress problem is
1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣 [From equation no. (3.36)]
1−𝑣
0 0
2
1 0.25 0
2 × 10 0.25 1 0
= 1 − 0.25
1 − (0.25)
0 0
2
2 × 10 1 0.25 0
= 0.25 1 0
0.9375
0 0 0.375
2 × 10 × 0.25 4 1 0
= 1 4 0
0.9375
0 0 1.5
4 1 0
[𝐷] = 2 × 10 × 0.2667 1 4 0 …(5)
0 0 1.5
4 1 0
⇒ [𝐷][𝐵] = 2 × 10 × 0.2667 1 4 0
0 0 1.5
50 1 0 0 0 −1 0
× 0 −1.5 0 1.5 0 0
3750
−1.5 1 1.5 0 0 −1
3.62 Two Dimensional Problems

2 × 10 × 0.2667 × 50 4 1 0 1 0 0 0 −1 0
= 1 4 0 0 −1.5 0 1.5 0 0
3750
0 0 1.5 −1.5 1 1.5 0 0 −1
4 −1.5 0 1.5 −4 0
= 711.2 1 −4 × 1.5 0 4 × 1.5 −1 0
1.5 × −1.5 1.5 1.5 × −1.5 0 0 −1.5
4 −1.5 0 1.5 −4 0
[𝐷][𝐵] = 711.2 1 −6 0 6 −1 0 … (6)
−2.25 1.5 2.25 0 0 −1.5
We know that,

50 1 0 3 0 −1 0
[𝐵] = 0 −1.5 0 1.5 0 0 [From equation no. (4)]
3750
−1.5 1 1.5 0 0 −1
1 0 −1.5
⎡ 0 −1.5 1 ⎤
50 0⎢ 0 1.5 ⎥⎥
[𝐵] = ⎢
3750 ⎢ 0 1.5 0 ⎥
⎢−1 0 0 ⎥
⎣0 0 −1 ⎦
1 0 −1.5
⎡
⎢ 0 −1.5 1 ⎤⎥ 4 −1.5 0 1.5 −4 0
50
⇒ [𝐵] [𝐷][𝐵] = × 711.2 ⎢ 0 0 1.5 ⎥ 1 −6 0 6 −1 0
3750 ⎢ 0 1.5 0 ⎥ −2.25 1.5 2.25
⎢−1 0 0 −1.5
0 0 ⎥
⎣ 0 0 −1 ⎦
7.375 −3.75 −3.375 1.5 −4 2.25
⎡ −3.75 10.5 2.25 −9 1.5 −1.5 ⎤
⎢ 2.25 0 ⎥
[𝐵] [𝐷][𝐵] = 9.483 ⎢−3.375 3.375 0 −2.25⎥
⎢ 1.5 −9 0 9 −1.5 0 ⎥
⎢ −4 1.5 0 −1.5 4 0 ⎥
⎣ 2.25 −1.5 −2.25 0 0 1.5 ⎦
Substitute [𝐵] [𝐷][𝐵] and A, t values in equation (1),
Stiffness matrix, [𝐾]
7.375 −3.75 −3.375 1.5 −4 2.25
⎡ −3.75 10.5 2.25 −9 1.5 −1.5 ⎤
⎢−3.375 2.25 3.375 0 0 −2.25⎥⎥
= 9.483 ⎢ × 1875 × 15
⎢ 1.5 −9 0 9 −1.5 0 ⎥
⎢ −4 1.5 0 −1.5 4 0 ⎥
⎣ 2.25 −1.5 −2.25 0 0 1.5 ⎦
 Finite Element Analysis 3.63
7.375 −3.75 −3.375 1.5 −4 2.25
⎡ −3.75 10.5 2.25 −9 1.5 −1.5 ⎤
⎢ 2.25 0 ⎥
−3.375 3.375 0 −2.25⎥
= 26.67 × ⎢
⎢ 1.5 −9 0 9 −1.5 0 ⎥
⎢ −4 1.5 0 −1.5 4 0 ⎥
⎣ 2.25 −1.5 −2.25 0 0 1.5 ⎦
196.69 −100.01 −90.01 40.01 −106.68 60
⎡−100.01 280.04 60 −240.03 40.01 −40.01⎤⎥
⎢
4 ⎢ −90.01 60 90.01 0 0 −60 ⎥
[𝐾] = 1 × 10
⎢ 40.01 −240.03 0 240.03 −40.01 0 ⎥
⎢−106.68 40.01 0 −40.01 106.68 0 ⎥
⎣ 60 −40.01 −60 0 0 40.01 ⎦
For element (1), nodal displacements are 𝑢 𝑣 , 𝑢 , 𝑣 , 𝑢 , 𝑣 . [Refer Fig. (iii)]
Stiffness matrix, [𝐾] =
𝑢 𝑣 𝑢 𝑣 𝑢 𝑣
196.69 −100.01 −90.01 40.01 −106.68 60 𝑢
⎡−100.01 280.04 60 −240.03 40.01 ⎤
−40.01 𝑣
⎢ −90.01 60 90.01 0 0 −60 ⎥⎥ 𝑢
= 1 × 10 ⎢
⎢ 40.01 −240.03 0 240.03 −40.01 0 ⎥ 𝑣
⎢−106.68 40.01 0 −40.01 106.68 0 ⎥𝑢
⎣ 60 −40.01 −60 0 0 40.01 ⎦ 𝑣
…(7)
Consider element (2): (Nodal displacements 𝑢 𝑣 , 𝑢 , 𝑣 , and 𝑢 , 𝑣 )

Fig. (iv)
Take node 3 as origin
3.64 Two Dimensional Problems

𝑥 𝑦
For node 1: (0, 0)
𝑥 𝑦
For node 3: (0, -50)
𝑥 𝑦
For node 3: (75, 0)
We know that,
Stiffness matrix [ K ] = [B]T [D] [B] A t …(8)
Where, A = Area of the triangular element

1 1 𝑥 𝑦 1 1 0 0
= 1 𝑥 𝑦 = 1 0 −50
2 0 𝑥 𝑦 2
1 75 0
1 3750
= × 1 × (0 + (50 × 75) =
2 2
A = 1875 mm2
Strain – Displacement matrix

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟 … (9)
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
Where 𝑞 = 𝑦 − 𝑦 = −50 − 0 = −50

𝑞 =𝑦 −𝑦 = 0−0= 0

𝑞 = 𝑦 − 𝑦 = 0 + 50 = 50

𝑟 = 𝑥 − 𝑥 = 75 − 0 = 75

𝑟 = 𝑥 − 𝑥 = 0 − 75 = −75

𝑟 = 𝑥 −𝑥 = 0−0 =0

1 −50 0 0 0 50 0
⇒ [𝐵] = 0 75 0 −75 0 0
2A
75 −50 −75 0 0 50
 Finite Element Analysis 3.65

Substitute Area, value,

1 −50 0 0 0 50 0
⇒ [𝐵] = 0 75 0 −75 0 0
2 × 1875
75 −50 −75 0 0 50

50 −1 0 0 0 1 0
[𝐵] = 0 1.5 0 −1.5 0 0 … (10)
3750
1.5 −1 −1.5 0 0 1

Stress-Strain relationship matrix [ D ] for plane stress problem is

1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣 [From equation no. (3.36)]
1−𝑣
0 0
2
1 0.25 0
2 × 10 0.25 1 0
= 1 − 0.25
1 − (0.25)
0 0
2

2 × 10 × 1 0.25 0
= 0.25 1 0
0.9375
0 0 0.375

2 × 10 × 0.25 4 1 0
= 1 4 0 … (11)
0.9375
0 0 1.5

250 × 2 × 10 × 0.25 4 1 0 −1 0 0 0 1 0
[𝐷][𝐵] = 1 4 0 0 1.5 0 −1.5 0 0
0.9375 × 3750
0 0 1.5 1.5 −1 −1.5 0 0 1
−4 1.5 0 −1.5 4 0
= 711.2 −1 6 0 −6 1 0 … (12)
2.25 −1.5 −2.25 0 0 1.5

We know that,

50 −1 0 0 0 1 0
[𝐵] = 0 1.5 0 −1.5 0 0
3750
1.5 −1 −1.5 0 0 1
3.66 Two Dimensional Problems
−1 0 1.5
⎡ 0 1.5 −1 ⎤
50 ⎢ 0 0 −1.5⎥⎥
[𝐵] = ⎢
2 × 375 × 10 ⎢ 0 −1.5 0 ⎥
⎢ 1 0 0 ⎥
⎣ 0 0 1 ⎦
−1 0 1.5
⎡0 1.5 −1 ⎤ −4
⎢0 1.5 0 −1.5 4 0
50 0 −1.5⎥⎥
[𝐵] [𝐷][𝐵] = × 711.2 ⎢ −1 6 0 −6 1 0
3750 ⎢ 0 −1.5 0 ⎥
2.25 −1.5 −2.25 0 0 1.5
⎢1 0 0 ⎥
⎣0 0 1 ⎦

7.375 −3.75 −3.375 1.5 −4 2.25

⎡ −3.75 10.5 2.25 −9 −1.5 −1.5 ⎤
⎢ 2.25 0 ⎥
−3.375 3.375 0 −2.25⎥
= 9.483 ⎢
⎢ 1.5 −9 0 9 −1.54 0 ⎥
⎢ −4 1.5 0 −1.5 4 0 ⎥
⎣ 2.25 −1.5 −2.25 0 0 0 ⎦

Substitute [𝐵] [𝐷][𝐵] and A, t values in equation (1),

Stiffness matrix, [𝐾]

⎡
7.375 −3.75 −3.375 1.5 −4 2.25 ⎤
⎢ −3.75 10.5 2.25 −9 −1.5 −1.5 ⎥
= 9.483 ⎢−3.375 2.25 3.375 0 0 −2.25⎥ × 1875
⎢ 1.5 −9 0 9 −1.54 0 ⎥
⎢ −4 1.5 0 −1.5 4 0 ⎥
⎣ 2.25 −1.5 −2.25 0 0 1.5 ⎦
× 15

7.375 −3.75 −3.375 1.5 −4 −2.5

⎡
⎢
−3.75 10.5 2.25 −9 −1.5 −1.5 ⎤⎥
= 26.67 × 104 ⎢−3.375 2.25 3.375 0 0 −2.25⎥
⎢ 1.5 −9 0 9 −1.54 0 ⎥
⎢ −4 1.5 0 −1.5 4 0 ⎥
⎣ 2.25 −1.5 −2.25 0 0 1.5 ⎦
196.69 −100.01 −90.01 40.01 −106.68 60
⎡−100.01 280.04 60 −240.03 40.01 −40.01⎤
⎢ −60 ⎥⎥
[𝐾] = 1 × 10 ⎢ −90.01 60 90.01 0 0
⎢ 40.01 −240.03 0 240.03 −40.01 0 ⎥
⎢−106.68 40.01 0 −40.01 106.68 0 ⎥
⎣ 60 −40.01 −60 0 0 40.01 ⎦
 Finite Element Analysis 3.67

For element (2), nodal displacements are 𝑢 𝑣 , 𝑢 , 𝑣 and 𝑢 , 𝑣 [Refer Fig. (iv)]
𝑢 𝑣 𝑢 𝑣 𝑢 𝑣
196.69 −100.01 −90.01 40.01 −106.68 60 𝑢
⎡−100.01 280.04 60 −240.03 40.01 −40.01⎤ 𝑣
⎢ −90.01 60 90.01 0 0 −60 ⎥⎥ 𝑢
Stiffness matrix, [𝐾] = 1 × 10 ⎢ 𝑣
⎢ 40.01 −240.03 0 240.03 −40.01 0 ⎥
⎢−106.68 40.01 0 −40.01 106.68 0 ⎥ 𝑢
⎣ 60 −40.01 −60 0 0 40.01 ⎦ 𝑣

…(13)
[Node: Take, Nodal displacements 𝑢 𝑣 , 𝑢 , 𝑣 and 𝑢 , 𝑣 respective to (𝑥 𝑦 )(𝑥 𝑦 )
and (𝑥 𝑦 )
Global stiffness Matrix [ K ]
Assemble the stiffness matrix equations (7) and (13),
Global stiffness Matrix [ K ] =
𝑢 𝑣 𝑢 𝑣 𝑢 𝑣 𝑢 𝑣
-
196.69 -100.01 -90.01 40.01 0 0 60 𝑢
106.68
-100.01 280.04 60 -240.03 0 0 40.01 -40.01 𝑣
106.68 0 0 -40.01
-
-90.01 60 + + 40.01 + + 𝑢
106.68
90.01 0 0 60
0 40.01 -60 0
40.01 -240.03 + + 60 -40.01 + + 𝑣
0 240.03 -40.01 0
1 104
-
0 0 60 196.69 -100.01 -90.01 40.01 𝑢
106.68
-
0 0 40.01 -40.01 280.04 60 -240.03 𝑣
100.01
0 -60 90.01 0
-106.68 40.01 + + -90.01 60 + + 𝑢
0 -40.01 106.68 0
0 0 240.03
-40.01
60 -40.01 + 40.01 -240.03 + + 𝑣
-60
0 0 40.01
3.68 Two Dimensional Problems

Global Stiffness matrix, [𝐾] =

196.69 −100.01 −90.01 40.01 0 0 −106.68 60
⎡−100.01 280.04 60 −240.03 0 0 ⎤
40.01 −40.01
⎢ −90.01 60 196.69 0 −106.68 40.01 0 −100.01⎥⎥
⎢
1 × 10 ⎢ 40.01 −240.03 0 240.04 60 −40.01 −100.01 0 ⎥
⎢ 0 0 −106.68 60 196.69 −100.01 −90.01 40.01 ⎥
⎢ 0 0 40.01 −40.01 −100.01 280.04 60 −240.03⎥
⎢−106.68 40.01 0 −100.01 −90.01 60 196.69 0 ⎥
⎣ 60 −40.01 −100.01 0 40.01 −240.03 0 280.04 ⎦

We Know that, general force equation is

{ F } = [ K ] {u }
𝐹
⎧𝐹 ⎫ 196.69 −100.01 −90.01 40.01 0 0 −106.68 60 𝑢
⎡−100.01 280.04 ⎤
⎪ ⎪ 60 −240.03 0 0 40.01 −40.01 ⎧ 𝑣 ⎫
⎪𝐹 ⎪ ⎢ −90.01 −100.01⎥⎥ ⎪ ⎪
⎪ ⎪ ⎢
60 196.69 0 −106.68 40.01 0 ⎪𝑢 ⎪
𝐹 𝑣
= 1 × 10 ⎢ 40.01 −240.03 0 240.04 60 −40.01 −100.01 0 ⎥
⎨𝐹 ⎬ ⎢ 0 0 −106.68 60 196.69 −100.01 −90.01 40.01 ⎥ ⎨𝑢 ⎬
⎪𝐹 ⎪ ⎢ 0 0 40.01 −40.01 −100.01 280.04 60 −240.03⎥ ⎪𝑣 ⎪
⎪𝐹 ⎪ ⎢−106.68 40.01 0 −100.01 −90.01 60 196.69 0 ⎥ ⎪𝑢 ⎪
⎪ ⎪ ⎣ 60 −40.01 −100.01 0 40.01 −240.03 0 280.04 ⎦ ⎩ 𝑣 ⎭
⎩𝐹 ⎭

…(14)
Applying boundary conditions [Refer Fig. (ii)]
1. Node 3 and Node 4 are fixed. So, 𝑢 , 𝑣 and 𝑢 , 𝑣 are zero.
i.e., 𝑢 = 𝑣 = 𝑢 = 𝑣 = 0.
2. Node 1, is moving in x direction. So 𝑢 ≠ 0 but, 𝑣 = 0.
3. At node 2, a point load of 100 103 N is acting in (-y) direction.
So, F2-y = -100 103 N.
4. Body force is neglected. So, the remaining forces are zero.
i.e., 𝐹 =𝐹 =𝐹 =𝐹 =𝐹 =𝐹 =𝐹 =𝐹 = 0.
Substitute the above values in equation (14),
0 196.69 −100.01 −90.01 40.01 0 0 −106.68 60 𝑢
⎧ ⎫ ⎡−100.01 280.04 ⎤
0 60 −240.03 0 0 40.01 −40.01 ⎧ 0 ⎫
⎪ ⎪ ⎢ −90.01 −100.01⎥⎥ ⎪ 𝑢
⎪ 0 ⎪ ⎢
60 196.69 0 −106.68 40.01 0 ⎪ ⎪ ⎪
0 𝑣
= 1 × 10 ⎢ 40.01 −240.03 0 240.04 60 −40.01 −100.01 0 ⎥
⎨−100 × 10 ⎬ ⎢ 0 0 −106.68 60 196.69 −100.01 −90.01 40.01 ⎥ ⎨ 0 ⎬
⎪ 0 ⎪ ⎢ 0 0 40.01 −40.01 −100.01 280.04 60 −240.03⎥ ⎪ 0 ⎪
⎪ 0 ⎪ ⎢−106.68 40.01 0 −100.01 −90.01 60 196.69 0 ⎥ ⎪0⎪
⎩ 0 ⎭ ⎣ 60 −40.01 −100.01 0 40.01 −240.03 0 280.04 ⎦ ⎩ 0 ⎭
 Finite Element Analysis 3.69

In the above equation 𝑣 =0, are zero. So delete second row and second column of
[K] matrix. Similarly, 𝑢 , 𝑣 , 𝑢 𝑎𝑛𝑑 𝑣 are zero. So, delete fifth row fifth column, sixth
row sixth column, seventh row seventh column and eight row eight column of [K] matrix.
Hence the equation reduces to:
0 196.69 −90.01 40.01 𝑢
0 = 1 × 10 −90.01 196.69 0 𝑢
−100 × 10 40.01 0 280.04 𝑣
0 196.69 −90.01 40.01 𝑢
⇒ 0 = −90.01 196.69 0 𝑢 …(15)
−10 40.01 0 280.04 𝑣
0 196.69 −90.01 40.01 𝑢
𝑢 𝑅 ⟶ 𝑅 + 2.1852𝑅
⇒ 0 = 0 339.797 40.01
𝑅 ⟶ 𝑅 − 4.916𝑅
49.16 0 −90.01 −1336.67 𝑣
0 196.69 −90.01 40.01 𝑢
⇒ 0 = 0 339.797 40.01 𝑢 𝑅 ⟶ 𝑅 − 3.775𝑅
185.584 0 −90.01 −5006.056 𝑣

⇒ −5006.056 𝑣 = 185.584

⇒ 𝑣 = −0.0371 𝑚𝑚

⇒ 339.797𝑢 + 40.01𝑣 = 0

⇒ 339.797𝑢 + 40.01 × (−0.0371) = 0

⇒ 𝑢 = 0.00436 𝑚𝑚

⇒ 196.69𝑢 − 90.01𝑢 + 40.01𝑣 = 0

⇒ 196.69𝑢 − 90.01(0.00436) + 40.01(−0.0371) = 0

⇒ 𝑢 = 0.00954 𝑚𝑚

Nodal displacements:
𝑢 = 0.00954𝑚𝑚 𝑣 = 0 𝑚𝑚
𝑢 = 0.00436𝑚𝑚 𝑣 = −0.0371 𝑚𝑚
𝑢 =0 𝑣 =0
3.70 Two Dimensional Problems

𝑢 =0 𝑣 =0
Stress in each element:
We know that, stress, {  } = [D] [B] { u }
For element (1): [Refer Fig. (iii)].: (Nodal displacements equation 𝒖𝟏 , 𝒗𝟏 , 𝒖𝟐 , 𝒗𝟐 and
𝒖 𝟒 , 𝒗𝟒 )
𝑢
⎧𝑣 ⎫
−4 1.5 0 −1.5 4 0 ⎪𝑢 ⎪
⇒ stress, {  } = 711.2 −1 6 0 −6 1 0 𝑣 ⎬
2.25 −1.5 −2.25 0 0 1.5 ⎨⎪𝑢 ⎪
⎩𝑣 ⎭
[From equation no. (6)]
0.00954
0 ⎧ ⎫
−4 1.5 0 −1.5 4 0 ⎪ ⎪
0.00436
= 711.2 −1 6 0 −6 1 0 ×
2.25 −1.5 −2.25 0 0 1.5 ⎨−0.0371⎬
⎪ 0 ⎪
⎩ 0 ⎭
(4 × 0.067) + (1.5 × −0.0371)
{  }1 = 711.2 0.00954 + (6 × −0.0371
(−2.25 × 0.00954) + (2.25 × 0.00436
𝜎 −12.44
𝜎 = −151.53 𝑁/𝑚𝑚
𝜏 −8.289
 Where, 𝜎 , 𝜎  Normal stress,

𝜏  Shear stress
For element (2): [Refer Fig. (iv)].: (Nodal displacements equation 𝒖𝟑 , 𝒗𝟑 , 𝒖𝟒 , 𝒗𝟒 and
𝒖 𝟐 , 𝒗𝟐
Stress, { 2 } = [D]2 [B]2 { u }
𝑢
⎧𝑣 ⎫
−4 1.5 0 −1.5 4 0 ⎪𝑢 ⎪
= 711.2 −1 6 0 −6 1 0 𝑣
2.25 −1.5 −2.25 0 0 1.5 ⎨
⎪𝑢
⎬
⎪
⎩𝑣 ⎭
 Finite Element Analysis 3.71

[From equation no. (12)]

0
0 ⎧ ⎫
−4 1.5 0 −1.5 4 0 ⎪ ⎪
0
= 711.2 −1 6 0 −6 1 0 ×
⎨ 0 ⎬
2.25 −1.5 −2.25 0 0 1.5 ⎪ 0.00436 ⎪
⎩−0.0371⎭
4 × 0.00436
{  }2 = 711.2 1 × 0.00436
1.5 × −0.0371
𝜎 12.403
𝜎 = 3.100 𝑁/𝑚𝑚
𝜏 −39.578
Result:
1. Nodal displacements:
𝑢 = 0.00954, 𝑣 = 0,
𝑢 = 0.00436𝑚𝑚; 𝑣 = −0.0371,
𝑢 =0 𝑣 = 0 𝑚𝑚;
𝑢 =0 𝑣 = 0.
2. Element stresses:
𝜎 −12.44
For element (1) 𝜎 = −151.53 𝑁/𝑚𝑚
𝜏 −8.289
𝜎 12.403
For element (2) 𝜎 = 3.10 𝑁/𝑚𝑚
𝜏 −39.578
Tutorial: A thin plate is subjected to surface traction as shown in Fig. (i). Calculate the
global stiffness matrix.
3.72 Two Dimensional Problems

Take, t = 25 mm, E = 2  105 N/mm2 and v = 0.30

Assume plane stress condition.
Given: Thickness, t = 25 mm
Young’s modulus, E = 2  105 N/mm2
Poisson’s ratio, v = 0.30
Breadth, b = 250 mm
Length, l = 500 mm
Tensile surface traction, T = 0.4 N/mm2
The tensile surface traction is converted into nodal force.
1 1
⇒𝐹= 𝑇𝐴 = × 𝑇 × (𝑏 × 𝑡)
2 2
1
= × 0.4 × 250 × 25
2
Nodal force, F= 1250 N

Fig.(ii) Discretized plate

3.12 ELASTICITY EQUATIONS

Elasticity equations are used for solving structural mechanics problems. These
equations must be satisfied if an exact solution to a structural mechanics problem is to be
obtained.
There are four basic sets of elasticity equations. They are:
(i) Strain - Displacement relationship equations
 Finite Element Analysis 3.73

(ii) Stress - Strain relationship equations

(iii) Equilibrium equations
(iv) Compatibility equations

3.12.1. Strain-Displacement Relationship Equations

Consider a two dimensional element PQRS as shown in Fig.3.16. When an
external force acts on the element, it undergoes deformation and it becomes P'Q'R'S'.
Displacement in the x direction is u and y direction is v.

Fig. 3.16. Two dimensional element before and after deformation

We know that, the strain is equal to the ratio of change in length to the original
length of the body. Considering the element PQ in x direction,
𝑃 𝑄 − 𝑃𝑄
⇒ Stain, 𝑒 = … (3.40)
𝑃𝑄
We know that, 𝑃𝑄 = 𝑑𝑥 … (3.41)
𝜕𝑢 𝜕𝑣
and (𝑃 𝑄 ) = 𝑑𝑥 + + 𝑑𝑥 + 𝑑𝑥
𝜕𝑥 𝜕𝑥
calculating 𝑃 𝑄 using the binomial theorem and neglecting the higher order
elements i.e., and .
3.74 Two Dimensional Problems
𝜕𝑢
⇒ 𝑃 𝑄 = 𝑑𝑥 + + 𝑑𝑥 … (3.42)
𝜕𝑥
Substituting equations (3.42) and (3.41) in equation (3.40)
𝜕𝑢
𝑑𝑥 + 𝑑𝑥 − 𝑑𝑥
(3.40) ⇒ 𝑒 = 𝜕𝑥 … (3.43)
𝑑𝑥
𝜕𝑢
𝑑𝑥 1 + −1
𝜕𝑥
= … (3.44)
𝑑𝑥
𝜕𝑢
𝑒 = … (3.45)
𝜕𝑥
Similarly, considering the element PS in y direction,
𝜕𝑣
⇒ 𝑒 = … (3.46)
𝜕𝑦
The shear strain 𝛾 is obtained by using by using the following relation,
𝜕𝑢 𝜕𝑣
𝛾 = + … (3.47)
𝜕𝑦 𝜕𝑥
Equations (3.45), (3.46) and 3.47) are strain displacement relationships for two
dimensional element.
For three dimensional element, the displacement in z-direction is w. strain –
displacement equations are obtained by extending the two dimensional derivations.
𝜕𝑤
Stain in z direction, 𝑒 = … (3.48)
𝜕𝑧
𝜕𝑢 𝜕𝑤
Shear Stains are, 𝛾 = + … (3.49)
𝜕𝑧 𝜕𝑥
𝜕𝑣 𝜕𝑤
𝛾 = + … (3.50)
𝜕𝑧 𝜕𝑦
Equilibrium equations:
Consider a three dimensional element as shown in Fig. 3.17
It is subjected to normal stresses 𝜎 , 𝜎 and 𝜎 , shear stresses 𝜏 , 𝜏 and 𝜏 , and
body forces 𝐵 , 𝐵 and 𝐵 as shown in Fig. 3.17 (b). the stresses acting on the element are
 Finite Element Analysis 3.75

assumed to be constant as they act on the width of each face. But they are varying from
one face to the opposite. For example, 𝜎 is acting on the left vertical face, whereas 𝜎 +
dx is acting on the right vertical face.

Adding all the forces acting on the element in x – direction,

⇒ Σ𝐹 = 0

𝜕𝜎 𝜕𝜏
⇒ 𝜎 + 𝑑𝑥 𝑑𝑦𝑑𝑧 − 𝜎 𝑑𝑦𝑑𝑧 + 𝜏 + 𝑑𝑦 𝑑𝑥 𝑑𝑧 − 𝜏 𝑑𝑥 𝑑𝑧
𝜕𝑥 𝜕𝑦
𝜕𝜏
+ 𝜏 + 𝑑𝑧 𝑑𝑥 𝑑𝑦 − 𝜏 𝑑𝑥 𝑑𝑦 + 𝐵 𝑑𝑥 𝑑𝑦 𝑑𝑧 = 0
𝜕𝑧

𝜕𝜎 𝜕𝜏
⇒ 𝜎 𝑑𝑦𝑑𝑧 + 𝑑𝑥𝑑𝑦𝑑𝑧 − 𝜎 𝑑𝑦𝑑𝑧 + 𝜏 𝑑𝑥 𝑑𝑧 + 𝑑𝑦𝑑𝑥 𝑑𝑧 − 𝜏 𝑑𝑥 𝑑𝑧
𝜕𝑥 𝜕𝑦
𝜕𝜏
+ 𝜏 𝑑𝑥 𝑑𝑦 + 𝑑𝑧 𝑑𝑥 𝑑𝑦 − 𝜏 𝑑𝑥 𝑑𝑦 + 𝐵 𝑑𝑥 𝑑𝑦 𝑑𝑧 = 0
𝜕𝑧
𝜕𝜎 𝜕𝜏 𝜕𝜏
⇒ 𝑑𝑥𝑑𝑦𝑑𝑧 + 𝑑𝑥 𝑑𝑦 𝑑𝑧 + 𝑑𝑥 𝑑𝑦𝑑𝑧 + 𝐵 𝑑𝑥 𝑑𝑦 𝑑𝑧 = 0 … (3.51)
𝜕𝑥 𝜕𝑦 𝜕𝑧
3.76 Two Dimensional Problems

Divided by 𝑑𝑥𝑑𝑦𝑑𝑧
𝜕𝜎 𝜕𝜏 𝜕𝜏
⇒ + + +𝐵 =0 … (3.52)
𝜕𝑥 𝜕𝑦 𝜕𝑧
Similarly, adding all the forces acting on the element in the y and z – directions,
𝜕𝜎 𝜕𝜏 𝜕𝜏
⇒ + + +𝐵 =0 … (3.53)
𝜕𝑦 𝜕𝑧 𝜕𝑥
𝜕𝜎 𝜕𝜏 𝜕𝜏
⇒ + + +𝐵 =0 … (3.54)
𝜕𝑧 𝜕𝑥 𝜕𝑦
Equations (3.52), (3.53) and (3.54) are equilibrium equations for three
dimensional element.

3.13. AXISYMMETRIC ELEMENTS

In previous chapters, we have been concerned with one dimensional elements and
two dimensional elements. In this chapter, we consider a special two dimensional element
called the axisymmetric element.
Many three dimensional problems in engineering exhibit symmetry about an axis
of rotation. Such types of problems are known as axisymmetric problems. These problems
can be solved by using two dimensional finite elements. These elements are most
conveniently described in cylindrical (r, θ, z) co-ordinates. The required conditions for a
problem to be axisymmetric are as follows:
1. The problem domain must be symmetric about the axis of revolution, which is
conventionally taken as the z-axis.
2. All boundary conditions must be symmetric about the axis of revolution.
3. All loading conditions must be symmetric about the axis of revolution.
An axisymmetric solid is generated by revolving a plane figure about an axis in
the plane.
Finite elements for axisymmetric solids are pictured as triangular element or
quadrilateral element as shown in Fig.3.18 and 3.19. But these shapes are actually cross-
sections of ring elements.
We begin with the development of the stiffness matrix for the simplest
axisymmetric element, the triangular torus, whose vertical cross-section is a plane
triangle.
 Finite Element Analysis 3.77

Fig. 3.18. Three-node axisymmetric triangular element

Fig. 3.19 . Four-node axisymmetric quadrilateral element

3.13.1. Axisymmetric Formulation

Consider a typical axisymmetric triangular element with nodes 1, 2 and 3 as
shown in Fig.3.20.
In two dimensional problems, the displacements and distributed body force values
are indicated by x-y plane. But in case of axisymmetric problems, these values are
indicated by r-z plane as shown in Fig.3.20.
3.78 Two Dimensional Problems

Fig. 3.20. Typical axisymmetric element

For two dimensional problem, the displacement vector u is given by,
𝑢
𝑢(𝑥, 𝑦) =
𝑣
Where, u and v are the x and y components of u respectively.
In case of axisymmetric problems, the displacement vector u is given by,
𝑢
𝑢(𝑟, 𝑧) =
𝑤
Where, u and w are r and z components of u respectively.
The stresses and strains for two dimensional element are given by,
𝜎
stress, {𝜎} = 𝜎
𝜏
𝑒
strain, {𝑒} = 𝑒
𝛾
in case of axisymmetric element, stresses and strains are given by,
𝜎
𝜎
stress, {𝜎} = 𝜎
𝜏
where, 𝜎 → Radial stress
𝜎 → Longitudinal stress
 Finite Element Analysis 3.79

𝜎 → Circumferential stress
𝜏 → Shear stress
𝑒
𝑒
strain, {𝑒} = 𝑒
𝛾
where, 𝑒 → Radial stress
𝑒 → Longitudinal stress
𝑒 → Circumferential stress
𝛾 → Shear stress
For two dimensional problem, body force is given by,
𝐹
F= 𝐹

In case of axisymmetric problem,

𝐹
F=
𝐹

3.13.2 Derivation of shape function for Axisymmetric element (Triangular

Element)
Consider an axisymmetric triangular element with nodes 1, 2 and 3 as shown in
Fig. 3.18 Let the nodal displacements be 𝑢 , 𝑤 , 𝑢 , 𝑤 and 𝑢 , 𝑤 .
𝑢
⎧𝑤 ⎫
⎪𝑢 ⎪
Displacement, {𝑢} = 𝑤
⎨ ⎬
⎪𝑢 ⎪
⎩𝑤 ⎭
Since the triangular element has two degrees of freedom at each node, it has 6
generalized co-ordinates.
Displacement functions, 𝑢 =𝑎 +𝑎 𝑟+𝑎 𝑧 …(3.55)
𝑤 =𝑎 +𝑎 𝑟+𝑎 𝑧 …(3.56)
Where, 𝑎 , 𝑎 , 𝑎 , 𝑎 , 𝑎 and 𝑎 are global or generalized co-ordinates.
3.80 Two Dimensional Problems

Let 𝑢 =𝑎 +𝑎 𝑟+𝑎 𝑧
𝑢 =𝑎 +𝑎 𝑟 +𝑎 𝑧
𝑢 =𝑎 +𝑎 𝑟 +𝑎 𝑧
Write the above equations in matrix form,
𝑢 𝑎
1 1 𝑟 𝑧
𝑢 = 1 𝑟 𝑧 𝑎
𝑢 2 0 𝑟 𝑧 𝑎
𝑎 1 𝑟 𝑧 𝑢
𝑎 = 1 𝑟 𝑧 𝑢 …(3.57)
𝑎 0 𝑟 𝑧 𝑢
+ − +
⎡1 𝑟 𝑧 ⎤
⎢− + +⎥
𝐿𝑒𝑡, 𝐷 = ⎢ 1 ⎥
⎢ 𝑟 𝑧 ⎥
⎢+ − +⎥
⎣1 𝑟 𝑧 ⎦
𝐶
𝐷= … (3.58)
|𝐷|
Find the co-factors of matrix D.
𝑟 𝑧
𝐶 =+ 𝑟 𝑧 = (𝑟 𝑧 − 𝑟 𝑧 )

1 𝑧
𝐶 =− = −(𝑧 − 𝑧 ) = (𝑧 − 𝑧 )
1 𝑧
1 𝑟
𝐶 =+ = +(𝑟 − 𝑧 )
1 𝑟
𝑟 𝑧
𝐶 = − 𝑟 𝑧 = (𝑟 𝑧 − 𝑟 𝑧 ) = 𝑟 𝑧 − 𝑟 𝑧

1 𝑧
𝐶 =+ = (𝑧 − 𝑧 )
1 𝑧
1 𝑟
𝐶 =− = −(𝑟 − 𝑟 ) = (𝑟 − 𝑟 )
1 𝑟
𝑟 𝑟
𝐶 =+ 𝑟 𝑧 =𝑟 𝑧 −𝑟 𝑧
 Finite Element Analysis 3.81
1 𝑧
𝐶 =− = −(𝑧 − 𝑧 ) = (𝑧 − 𝑧 )
1 𝑧
1 𝑟
𝐶 =+ = (𝑟 − 𝑧 )
1 𝑟
(𝑟 𝑧 − 𝑟 𝑧 ) (𝑧 − 𝑧 ) 𝑟 − 𝑧
⇒ 𝐶 = (𝑟 𝑧 − 𝑟 𝑧 ) 𝑧 − 𝑧 𝑟 −𝑟
𝑟 𝑧 −𝑟 𝑧 𝑧 −𝑧 𝑟 −𝑧
𝑟 𝑧 −𝑟 𝑧 𝑟 𝑧 −𝑟 𝑧 𝑟 𝑧 −𝑟 𝑧
C = 𝑧 −𝑧 𝑧 −𝑧 𝑧 −𝑧 … (3.59)
𝑟 −𝑧 𝑟 −𝑟 𝑟 −𝑧
We know that,
1 𝑟 𝑧
𝐷= 1 𝑟 𝑧
0 𝑟 𝑧
1 𝑟 𝑧
|𝐷| = 1 𝑟 𝑧
0 𝑟 𝑧
|𝐷| = 1(𝑟 𝑧 − 𝑟 𝑧 ) − 𝑟 (𝑧 − 𝑧 ) + 𝑧 (𝑟 − 𝑟 ) …(3.60)
Substitute CT and D values in equation (3.58),
𝑟 𝑧 −𝑟 𝑧 𝑟 𝑧 −𝑟 𝑧 𝑟 𝑧 −𝑟 𝑧
1
(4.20) ⇒ 𝐷 = × 𝑧 −𝑧 𝑧 −𝑧 𝑧 −𝑧
(𝑟 𝑧 − 𝑟 𝑧 ) − 𝑟 (𝑧 − 𝑧 ) + 𝑧 (𝑟 − 𝑟 ) 𝑟 −𝑧 𝑟 −𝑟 𝑟 −𝑧

Substitute D-1 value in equation (3.57),

𝑎 1 𝑟 𝑧 𝑢
𝑎 = 1 𝑟 𝑧 𝑢
𝑎 0 𝑟 𝑧 𝑢
1
=
(𝑟 𝑧 − 𝑟 𝑧 ) − 𝑟 (𝑧 − 𝑧 ) + 𝑧 (𝑟 − 𝑟 )
𝑟 𝑧 −𝑟 𝑧 𝑟 𝑧 −𝑟 𝑧 𝑟 𝑧 −𝑟 𝑧 𝑢
× 𝑧 −𝑧 𝑧 −𝑧 𝑧 −𝑧 𝑢 … (3.61)
𝑟 −𝑧 𝑟 −𝑟 𝑟 −𝑧 𝑢
The area of the triangle can be expressed as a function of the r, z co-ordinates of
the nodes 1, 2 and 3.
3.82 Two Dimensional Problems

1 1 𝑟 𝑧
𝐴= 1 𝑟 𝑧
2 0 𝑟 𝑧

1
𝐴= 1(𝑟 𝑧 − 𝑟 𝑧 ) − 𝑟 (𝑧 − 𝑧 ) + 𝑧 (𝑟 − 𝑟 )
2

⇒ 2𝐴 = (𝑟 𝑧 − 𝑟 𝑧 ) − 𝑟 (𝑧 − 𝑧 ) + 𝑧 (𝑟 − 𝑟 ) ...(3.62)

Substitute equation (3.62) in equation (3.61),

𝑎 𝑟 𝑧 −𝑟 𝑧 𝑟 𝑧 −𝑟 𝑧 𝑟 𝑧 −𝑟 𝑧 𝑢
1
𝑎 = 𝑧 −𝑧 𝑧 −𝑧 𝑧 −𝑧 𝑢 … (3.63)
𝑎 2A 𝑟 −𝑧 𝑟 −𝑟 𝑟 −𝑧 𝑢
𝑎 𝛼 𝛼 𝛼 𝑢
1
⇒ 𝑎 = 𝛽 𝛽 𝛽 𝑢 … (3.64)
𝑎 2A 𝛾 𝛾 𝛾 𝑢

Where, 𝛼 = 𝑟 𝑧 − 𝑟 𝑧 ; 𝛼 =𝑟 𝑧 −𝑟 𝑧 ; 𝛼 =𝑟 𝑧 −𝑟 𝑧

𝛽 =𝑧 −𝑧 𝛽 =𝑧 −𝑧 𝛽 =𝑧 −𝑧

𝛾 =𝑟 −𝑧 𝛾 =𝑟 −𝑟 𝛾 = 𝑟 −𝑧
From equation (3.55), we know that,
𝑢 =𝑎 +𝑎 𝑟+𝑎 𝑧
We can write this equation in matrix form,
𝑎
𝑢 = [1 𝑟 𝑧] 𝑎
𝑎
𝛼 𝛼 𝛼 𝑢
1
= [1 𝑟 𝑧] × 𝛽 𝛽 𝛽 × 𝑢 [From equation(3.64)]
2A 𝛾 𝛾 𝛾 𝑢
𝛼 𝛼 𝛼 𝑢
1
= [1 𝑟 𝑧] 𝛽 𝛽 𝛽 × 𝑢
2A 𝛾 𝛾 𝛾 𝑢
 Finite Element Analysis 3.83
1
= [𝛼 + 𝛽 𝑟 + 𝛾 𝑧 𝛼 +𝛽 𝑟+𝛾 𝑧 𝛼 + 𝛽 𝑟 + 𝛾 𝑧]
2A
𝑢
× 𝑢
𝑢
[ Note: (1  3)  (3  3) = (1  3)]
𝑢
𝛼 +𝛽 𝑟+𝛾 𝑧 𝛼 +𝛽 𝑟+𝛾 𝑧 𝛼 +𝛽 𝑟+𝛾 𝑧
⇒𝑢= × 𝑢
2𝐴 2𝐴 2𝐴 𝑢
The above equation is in the form of
𝑢
𝑢 = [𝑁 𝑁 𝑁 ] 𝑢 ..(3.65)
𝑢
𝑤
Similarly, 𝑤 = [𝑁 𝑁 𝑁 ] 𝑤 ..(3.66)
𝑤
𝛼 +𝛽 𝑟+𝛾 𝑧
Where, Shape function, 𝑁 =
2𝐴
𝛼 +𝛽 𝑟+𝛾 𝑧
𝑁 =
2𝐴
𝛼 +𝛽 𝑟+𝛾 𝑧
𝑁 =
2𝐴
We can write equations (3.65) and (3.66) as follows:
𝑢 =𝑁 𝑢 +𝑁 𝑢 +𝑁 𝑢 …(3.67)

𝑤 =𝑁 𝑤 +𝑁 𝑤 +𝑁 𝑤 …(3.68)
Assembling the equations (3.67) and (3.68) in matrix form,
Displacement function,
𝑢
⎧𝑤 ⎫
𝑢(𝑟, 𝑧) 𝑁 0 𝑁 0 𝑁 0 ⎪𝑢 ⎪
𝑢(𝑟, 𝑧) = = …(3.69)
𝑤(𝑟, 𝑧) 0 𝑁 0 𝑁 0 𝑁 ⎨𝑤 ⎬
⎪𝑢 ⎪
⎩𝑤 ⎭
3.84 Two Dimensional Problems

3.13.3 Strain – Displacement Matrix [B] for Axisymmetric Triangular

Element
Displacement function for axisymmetric triangular element is given by,
Displacement function,
𝑢
⎧𝑤 ⎫
𝑢(𝑟, 𝑧) 𝑁 0 𝑁 0 𝑁 0 ⎪𝑢 ⎪
𝑢(𝑟, 𝑧) = =
𝑤(𝑟, 𝑧) 0 𝑁 0 𝑁 0 𝑁 ⎨𝑤 ⎬
⎪𝑢 ⎪
⎩𝑤 ⎭
or
we can write, 𝑢 =𝑁 𝑢 +𝑁 𝑢 +𝑁 𝑢 …(3.70)
𝑤 =𝑁 𝑤 +𝑁 𝑤 +𝑁 𝑤 …(3.71)
The strain components are,
𝜕𝑢 𝜕𝑁 𝜕𝑁 𝜕𝑁
Radial strain, 𝑒 = = 𝑢 + 𝑢 + 𝑢
𝜕𝑟 𝜕𝑟 𝜕𝑟 𝜕𝑟
𝜕𝑁 𝜕𝑁 𝜕𝑁
⇒𝑒 = 𝑢 + 𝑢 + 𝑢 … (3.72)
𝜕𝑟 𝜕𝑟 𝜕𝑟
𝑢
Circumferential strain, 𝑒 =
𝑟
𝑁 𝑁 𝑁
⇒𝑒 = 𝑢 + 𝑢 + 𝑢 … (3.73)
𝑟 𝑟 𝑟
𝜕𝑤
Longitudinal strain, 𝑒 =
𝜕𝑧
𝜕𝑁 𝜕𝑁 𝜕𝑁
⇒𝑒 = 𝑤 + 𝑤 + 𝑤 … (3.74)
𝜕𝑧 𝜕𝑧 𝜕𝑧
𝜕𝑢 𝜕𝑤
Shear strain, 𝛾 = +
𝜕𝑧 𝜕𝑟
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
𝛾 = 𝑢 + 𝑢 + 𝑢 + 𝑤 + 𝑤
𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑟 𝜕𝑟
𝜕𝑁
+ 𝑤 … (3.75)
𝜕𝑟
 Finite Element Analysis 3.85

Arranging equations ( 3.72), (3.73), (3.74) and (3.75) in matrix form,

𝜕𝑁 𝜕𝑁 𝜕𝑁
⎡ 0 0 0 ⎤
⎢ 𝜕𝑟 𝜕𝑟 𝜕𝑟 ⎥ 𝑢
𝑒 𝑁 𝑁 𝑁 ⎧𝑤 ⎫
⎢ 0 0 0 ⎥ ⎪𝑢 ⎪
𝑒
⇒ 𝑒 =⎢ 𝑟 𝑟 𝑟 ⎥ 𝑤 … (3.76)
⎢ 0 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎨ ⎬
𝛾 0 0 ⎪𝑢 ⎪
⎢𝜕𝑁 𝜕𝑧 𝜕𝑁 𝜕𝑧 𝜕𝑁 𝜕𝑧 ⎥
𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎩𝑤 ⎭
⎢
𝜕𝑧 𝜕𝑧 𝜕𝑧
⎣ 𝜕𝑟 𝜕𝑟 𝜕𝑟 ⎦
From equation (3.65) or (3.66), we know that,
𝛼 +𝛽 𝑟+𝛾 𝑧
Shape function, 𝑁 =
2𝐴
𝛼 +𝛽 𝑟+𝛾 𝑧
𝑁 =
2𝐴
𝛼 +𝛽 𝑟+𝛾 𝑧
𝑁 =
2𝐴
𝜕𝑁 𝛽
Partial differentiation ⇒ =
𝜕𝑟 2𝐴
𝜕𝑁 𝛽
=
𝜕𝑟 2𝐴
𝜕𝑁 𝛽
=
𝜕𝑟 2𝐴
𝑁 1 𝛼 𝛾 𝑧
= +𝛽 +
𝑟 2𝐴 𝑟 𝑟
𝑁 1 𝛼 𝛾 𝑧
= +𝛽 +
𝑟 2𝐴 𝑟 𝑟
𝑁 1 𝛼 𝛾 𝑧
= +𝛽 +
𝑟 2𝐴 𝑟 𝑟
𝜕𝑁 𝛾
=
𝜕𝑧 2𝐴
3.86 Two Dimensional Problems
𝜕𝑁 𝛾
=
𝜕𝑧 2𝐴
𝜕𝑁 𝛾
=
𝜕𝑧 2𝐴
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝑁 𝑁 𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
Substitute , , , , , , , , and values in Equation (3.76).
𝜕𝑟 𝜕𝑟 𝜕𝑟 𝑟 𝑟 𝑟 𝜕𝑧 𝜕𝑧 𝜕𝑧
𝑢
𝛽 0 𝛽 0 𝛽 0 ⎧ ⎫
𝑒 ⎡𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤ 𝑤
𝑒 1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0 ⎥ ⎪𝑢 ⎪
= 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 ⎥ 𝑤
𝑒 2𝐴 ⎢ 0 𝛾 0 𝛾 0 𝛾 ⎥ ⎨𝑢 ⎬
𝛾 ⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦⎪ ⎩𝑤 ⎭
⎪

The above equation is in the form of,

{e}=[B]{u}
[ B] = Strain – Displacement matrix
𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
= 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 ⎥ . . (3.77)
2𝐴 ⎢ 0 𝛾 0 𝛾 0 𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
Where, 𝛼 = 𝑟 𝑧 − 𝑟 𝑧 ; 𝛼 =𝑟 𝑧 −𝑟 𝑧 ; 𝛼 =𝑟 𝑧 −𝑟 𝑧
𝛽 =𝑧 −𝑧 𝛽 =𝑧 −𝑧 𝛽 =𝑧 −𝑧
𝛾 =𝑟 −𝑧 𝛾 =𝑟 −𝑟 𝛾 = 𝑟 −𝑧

3.13.4 Stress – Strain Relationship Matrix [D] for Axisymmetric Triangular

Element
By using Hooke’s law, we derived the following normal stresses equations.
𝐸
𝜎 = 𝑒 (1 − 𝑣) + 𝑣𝑒 + 𝑣𝑒
(1 + 𝑣)(1 − 2𝑣)
𝐸
𝜎 = 𝑣𝑒 (1 − 𝑣) + 𝑒 + 𝑣𝑒
(1 + 𝑣)(1 − 2𝑣)
𝐸
𝜎 = 𝑣𝑒 + 𝑣𝑒 + (1 − 𝑣)𝑒
(1 + 𝑣)(1 − 2𝑣)
 Finite Element Analysis 3.87
𝐸 1 − 2𝑣
𝜏 = ×𝛾
(1 + 𝑣)(1 − 2𝑣) 2
Substitute x = r and y =  in the above equations,
𝐸
 Radial stress, 𝜎 = [𝑒 (1 − 𝑣) + 𝑣𝑒 + 𝑣𝑒 ] … (3.78)
(1 + 𝑣)(1 − 2𝑣)
𝐸
Circumferential stress, 𝜎 = [𝑣𝑒 (1 − 𝑣) + 𝑒 + 𝑣𝑒 ] … (3.79)
(1 + 𝑣)(1 − 2𝑣)
𝐸
Longitudinal strain, 𝜎 = 𝑣𝑒 + 𝑣𝑒 + (1 − 𝑣)𝑒 … (3.80)
(1 + 𝑣)(1 − 2𝑣)
𝐸 1 − 2𝑣
Shear strain, 𝜏 = ×𝛾 … (3.81)
(1 + 𝑣)(1 − 2𝑣) 2
Arranging the above equations, (3.78), (3.79), (3.80) and (3.81) in matrix form,

𝜎 0 𝑒
⎡1 − 𝑣 𝑣 𝑣
0 ⎤ 𝑒
𝜎 𝐸 ⎢ 𝑣
= 1−𝑣 𝑣 0 ⎥ … (3.82)
𝜎 (1 + 𝑣)(1 − 2𝑣) ⎢ 𝑣 𝑣 1 − 𝑣 1 − 2𝑣 ⎥ 𝑒
𝜏 ⎢ 0 0 ⎥ 𝛾
⎣ 0
2 ⎦
The above equation is in the form of,
{}=[D]{E}
Where [ D] = Stress - Strain relationship matrix
0
⎡1 − 𝑣 𝑣 𝑣
0 ⎤
𝐸 ⎢ 𝑣
= 1−𝑣 𝑣 0 ⎥ … (3.83)
(1 + 𝑣)(1 − 2𝑣) ⎢ 𝑣 𝑣 1 − 𝑣 1 − 2𝑣 ⎥
⎢ 0 0 ⎥
⎣ 0
2 ⎦

3.13.5 Assemblage of the Element stiffness Matrix [K]

We know that,

Stiffness matrix, [𝐾] = [𝐵] [𝐷][𝐵]𝑑𝑉 = [𝐵] [𝐷][𝐵] 𝑑𝑉

= [𝐵] [𝐷][𝐵]𝑉
3.88 Two Dimensional Problems

Stiffness matrix, [𝐾] = 2𝜋𝑟 𝐴 [𝐵] [𝐷][𝐵] …(3.84)

[∵ 𝑉 = 2𝜋𝑟𝐴]
Where,
𝑟 +𝑟 +𝑟
Co − ordinate, 𝑟 =
3
A = Area of the triangular element = ½ (b  h)
[B] = Strain – Displacement matrix
𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
= 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟
2𝐴 ⎢ 0 𝛾 0 𝛾 0
⎥
𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
Where, 𝛼 = 𝑟 𝑧 − 𝑟 𝑧 ; 𝛼 =𝑟 𝑧 −𝑟 𝑧 ; 𝛼 =𝑟 𝑧 −𝑟 𝑧
𝛽 =𝑧 −𝑧 𝛽 =𝑧 −𝑧 𝛽 =𝑧 −𝑧
𝛾 =𝑟 −𝑧 𝛾 =𝑟 −𝑟 𝛾 = 𝑟 −𝑧
𝑟 +𝑟 +𝑟
and Co − ordinate, 𝑟 =
3
[D] = Stress- Strain relationship matrix
0
⎡1 − 𝑣 𝑣 𝑣
0 ⎤
𝐸 ⎢ 𝑣
= 1−𝑣 𝑣 0 ⎥
(1 + 𝑣)(1 − 2𝑣) ⎢ 𝑣 𝑣 1 − 𝑣 1 − 2𝑣 ⎥
⎢ 0 0 ⎥
⎣ 0
2 ⎦
Where, E  Young’s modulus
v  Poisson’s ratio

3.14. SOLVED PROBLEMS -AXISYMMETRIC ELEMENT

Example 3.10
The nodal co-ordinates for an axisymmetric triangular element are given below
r1 = 10 mm ; z1 = 10 mm
r2 = 30 mm ; z2 = 10 mm
 Finite Element Analysis 3.89

r3 = 30 mm ; z3 = 40 mm
Evaluate [B] Matrix for that element
Given:
r1 = 10 mm ; z1 = 10 mm
r2 = 30 mm ; z2 = 10 mm
r3 = 30 mm ; z3 = 40 mm
To find:
Strain displacement matrix [B]
Solution: We know that,
Strain - Displacement matrix
𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
[𝐵] = 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 ⎥ … (1)
2𝐴 ⎢ 0 𝛾 0 𝛾 0 𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
Where, A = Area of the triangular element

1 1 𝑟 𝑧
= 1 𝑟 𝑧
2 1 𝑟 𝑧
1
= (𝑟 𝑧 − 𝑟 𝑧 ) − 𝑟 (𝑧 − 𝑧 ) + 𝑧 (𝑟 − 𝑟 )
2
1
= [(30 × 40) − (30 × 10) − 10(40 − 10) + 10(30 − 30)]
2
1
= × 600
2
A = 300 mm2
𝑟 +𝑟 +𝑟 10 + 30 + 30
Co − ordinate, 𝑟= =
3 3
r = 23.334 mm
3.90 Two Dimensional Problems
𝑧 +𝑧 +𝑧 10 + 10 + 40
𝑧= =
3 3
z = 20 mm
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (30 × 40) − (30 × 10)
𝜶𝟏 = 𝟗𝟎𝟎 𝒎𝒎𝟐
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (30 × 10) − (10 × 40)
𝜶𝟐 = −𝟏𝟎𝟎 𝒎𝒎𝟐
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (10 × 10) − (30 × 10)
𝜶𝟑 = −𝟐𝟎𝟎 𝒎𝒎𝟐
𝛽 = 𝑧 − 𝑧 = 10 − 40
𝜷𝟏 = −𝟑𝟎 𝒎𝒎
𝛽 = 𝑧 − 𝑧 = 40 − 10
𝜷𝟐 = 𝟑𝟎 𝒎𝒎
𝛽 = 𝑧 − 𝑧 = 10 − 10
𝜷𝟑 = 𝟎
𝛾 = 𝑟 − 𝑧 = 30 − 30
𝜸𝟏 = 𝟎
𝛾 = 𝑟 − 𝑟 = 10 − 30
𝜸𝟐 = −𝟐𝟎 𝒎𝒎
𝛾 = 𝑟 − 𝑧 = 30 − 10
𝛾 = 20 𝑚𝑚
𝛼 𝛾 𝑧 900 0 × 20
⇒ +𝛽 + = + (−30) + = 8.571 𝑚𝑚
𝑟 𝑟 23.334 23.334
𝛼 𝛾 𝑧 −100 (−20 × 20)
⇒ +𝛽 + = + (30) + = 8.571 𝑚𝑚
𝑟 𝑟 23.334 23.334
𝛼 𝛾 𝑧 −200 20 × 20
⇒ +𝛽 + = +0+ = 8.571 𝑚𝑚
𝑟 𝑟 23.334 23.334
 Finite Element Analysis 3.91

Substitute, A, 𝛽 , 𝛽 , 𝛽 , +𝛽 + , +𝛽 + , +𝛽 + , 𝛾 , 𝛾 and 𝛾
values in equation (1),
−30 0 30 0 0 0
1 8.571 0 8.571 0 8.571 0
(1) ⇒ [𝐵] =
2 × 300 0 0 0 −20 0 20
0 −30 −20 30 20 0
−0.05 0 0.05 0 0 0
[𝐵] = 0.0142 0 0.0142 0 0.0142 0
0 0 0 −0.0333 0 0.0333
0 −0.05 −0.0333 0.05 0.0333 0

Result: Strain-Displacement matrix

−0.05 0 0.05 0 0 0
[𝐵] = 0.0142 0 0.0142 0 0.0142 0
0 0 0 −0.0333 0 0.0333
0 −0.05 −0.0333 0.05 0.0333 0

Example 3.11
The nodal co-ordinates for an axisymmetric triangular element are given below

r1 = 20 mm ; z1 = 40 mm

r2 = 40 mm ; z2 = 40 mm

r3 = 30 mm ; z3 = 60 mm

Evaluate [B] Matrix for that element

Given: Coordinates

r1 = 20 mm ; z1 = 40 mm

r2 = 40 mm ; z2 = 40 mm

r3 = 30 mm ; z3 = 60 mm

To find: Strain displacement matrix [B]

3.92 Two Dimensional Problems

Solution:

𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
[𝐵] = 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 ⎥ … (1)
2𝐴 ⎢ 0 𝛾 0 𝛾 0 𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
[From equation no. (4.39)]
Where, A = Area of the triangular element

1 1 𝑟 𝑧
= 1 𝑟 𝑧 [From equation no. (4.24)]
2 1 𝑟 𝑧
1
= (𝑟 𝑧 − 𝑟 𝑧 ) − 𝑟 (𝑧 − 𝑧 ) + 𝑧 (𝑟 − 𝑟 )
2
1
= [(40 × 60) − (30 × 40) − 20(60 − 40) + 40(30 − 40)]
2
1
= [1200 − 400 − 400]
2
A = 200 mm2
𝑟 +𝑟 +𝑟 20 + 40 + 30
Co − ordinate, 𝑟= =
3 3
r = 30 mm
𝑧 +𝑧 +𝑧 40 + 40 + 60
𝑧= =
3 3
z = 46.667 mm
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (40 × 60) − (30 × 40)
𝜶𝟏 = 𝟏𝟐𝟎𝟎 𝒎𝒎𝟐
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (30 × 40) − (20 × 60)
𝜶𝟐 = 𝟎
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (20 × 40) − (40 × 40)
𝜶𝟑 = −𝟖𝟎𝟎 𝒎𝒎𝟐
 Finite Element Analysis 3.93

𝛽 = 𝑧 − 𝑧 = 40 − 60
𝜷𝟏 = −𝟐𝟎 𝒎𝒎
𝛽 = 𝑧 − 𝑧 = 60 − 40
𝜷𝟐 = 𝟐𝟎 𝒎𝒎
𝛽 = 𝑧 − 𝑧 = 40 − 40
𝜷𝟑 = 𝟎
𝛾 = 𝑟 − 𝑧 = 30 − 40
𝜸𝟏 = −𝟏𝟎 𝒎𝒎
𝛾 = 𝑟 − 𝑟 = 20 − 30
𝜸𝟐 = −𝟏𝟎 𝒎𝒎
𝛾 = 𝑟 − 𝑧 = 40 − 20
𝛾 = 20 𝑚𝑚

𝛼 𝛾 𝑧 1200 −10 × 46.667

⇒ +𝛽 + = + (−20) + = 40 − 20 − 15.556
𝑟 𝑟 30 30

= 4.444 𝑚𝑚

𝛼 𝛾 𝑧 0 (−10 × 46.667)
⇒ +𝛽 + = + (20) + = 4.444 𝑚𝑚
𝑟 𝑟 30 30
𝛼 𝛾 𝑧 −800 20 × 46.667
⇒ +𝛽 + = +0+ = 4.444 𝑚𝑚
𝑟 𝑟 30 30

Substitute, A, 𝛽 , 𝛽 , 𝛽 , +𝛽 + , +𝛽 + , +𝛽 + , 𝛾 , 𝛾 and 𝛾
values in equation (1),
−20 0 20 0 0 0
1 4.444 0 4.444 0 4.444 0
(1) ⇒ [𝐵] =
2 × 200 0 −10 0 −10 0 20
−10 −20 −10 20 20 0
−0.05 0 0.05 0 0 0
[𝐵] = 0.0111 0 0.0111 0 0.0111 0
0 −0.025 0 −0.025 0 0.05
−0.025 −0.05 −0.025 0.05 0.05 0
3.94 Two Dimensional Problems

Result: Strain-Displacement matrix

−0.05 0 0.05 0 0 0
[𝐵] = 0.0111 0 0.0111 0 0.0111 0
0 −0.025 0 −0.025 0 0.05
−0.025 −0.05 −0.025 0.05 0.05 0

Example 3.12
For the element shown in fig (i), determine the stiffness matrix. Take E= 200 Gpa
and v = 0.25. The coordinated shown in fig (i) are in millimetres.

Fig. (i)
Given:
r1 = 0 mm ; z1 = 0 mm
r2 = 50 mm ; z2 = 0 mm
r3 = 50 mm ; z3 = 50 mm
Young’s Modulus, E = 200 GPa
= 2 x 105 N/mm2
Poisson’s ratio, v = 0.25
To find:
Element stiffness matrix [K].
Solution: For axisymmetric triangular element, stiffness matrix [K] is given by,
[𝐾] = 2𝜋𝑟 𝐴 [𝐵] [𝐷][𝐵] …(1)
 Finite Element Analysis 3.95

Where, A = Area of the triangular element

1 1 𝑟 𝑧
= 1 𝑟 𝑧
2 1 𝑟 𝑧
We can calculate as,
Area of the triangular element,
1 1
A = × Breadth × Height = × 50 × 50
2 2
A = 1250 𝑚𝑚 …(2)
𝑟 +𝑟 +𝑟 0 + 50 + 50
Co − ordinate, 𝑟= =
3 3
r = 33.333 mm …(3)
𝑧 +𝑧 +𝑧 0 + 0 + 50
𝑧= =
3 3
z = 16.666 mm …(4)
we know that,
Stress - Strain relationship matrix
0
⎡1 − 𝑣 𝑣 𝑣
0 ⎤
𝐸 ⎢ 𝑣
= 1−𝑣 𝑣 0 ⎥
(1 + 𝑣)(1 − 2𝑣) ⎢ 𝑣 𝑣 1 − 𝑣 1 − 2𝑣 ⎥
⎢ 0 0 ⎥
⎣ 0
2 ⎦
0
⎡1 − 0.25 0.25 0.25
0 ⎤
2 × 10 ⎢ 0.25 1 − 0.25 0.25 ⎥
⇒ [𝐷] = ⎢ 0
(1 + 0.25)(1 − (2 × 0.25)) 0.25 0.25 1 − 0.25 1 − 2(0.25)⎥
⎢ ⎥
⎣ 0 0 0 ⎦
2
0.75 0.25 0.25 0 3 1 1 0
= 320 × 10 0.25 0.75 0.25 0 = 320 × 10 × 0.25 1 3 1 0
0.25 0.25 0.75 0 1 1 3 0
0 0 0 0.25 0 0 0 1
3.96 Two Dimensional Problems
3 1 1 0
[𝐷] = 80 × 10 1 3 1 0
1 1 3 0
0 0 0 1
We know that, Strain - Displacement matrix
𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
[𝐵] = 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 ⎥ … (6)
2𝐴 ⎢ 0 𝛾 0 𝛾 0 𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
Where,
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (50 × 50) − (50 × 0)
𝜶𝟏 = 𝟐𝟓𝟎𝟎 𝒎𝒎𝟐
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (50 × 0) − (0 × 50)
𝜶𝟐 = 𝟎
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (0 × 0) − (50 × 0)
𝜶𝟑 = 𝟎
𝛽 = 𝑧 − 𝑧 = 0 − 50
𝜷𝟏 = −𝟓𝟎 𝒎𝒎
𝛽 = 𝑧 − 𝑧 = 50 − 0
𝜷𝟐 = 𝟓𝟎 𝒎𝒎
𝛽 =𝑧 −𝑧 = 0−0
𝜷𝟑 = 𝟎
𝛾 = 𝑟 − 𝑧 = 50 − 50
𝜸𝟏 = 𝟎
𝛾 = 𝑟 − 𝑟 = 0 − 50
𝜸𝟐 = − 𝟓𝟎 𝒎𝒎
𝛾 = 𝑟 − 𝑧 = 50 − 0
𝛾 = 50 𝑚𝑚
 Finite Element Analysis 3.97
𝛼 𝛾 𝑧 2500
⇒ +𝛽 + = + (−50) + 0 = 25 𝑚𝑚
𝑟 𝑟 23.334
𝛼 𝛾 𝑧 (−20 × 20)
⇒ +𝛽 + = 0 + (30) + = 25 𝑚𝑚
𝑟 𝑟 23.334
𝛼 𝛾 𝑧 50 × 16.666
⇒ +𝛽 + = 0+0+ = 25 𝑚𝑚
𝑟 𝑟 33.333

Substitute, A, 𝛽 , 𝛽 , 𝛽 , +𝛽 + , +𝛽 + , +𝛽 + , 𝛾 , 𝛾 and 𝛾
values in equation (6),
−50 0 50 0 0 0
1 25 0 25 0 25 0
[𝐵] =
2𝐴 0 0 0 −50 0 50
0 −50 −50 50 50 0
Substitute Area A, value
−50 0 50 0 0 0
1 25 0 25 0 25 0
[𝐵] = … (7)
2 × 1250 0 0 0 −50 0 50
0 −50 −50 50 50 0
1 1 3 0
⇒ [𝐷][𝐵] = 80 × 10 3 1 1 0
1 3 1 0
0 0 0 1
−50 0 50 0 0 0
1 25 0 25 0 25 0
×
2 × 1250 0 0 0 −50 0 50
0 −50 −50 50 50 0
3 1 1 0 −2 0 2 0 0 0
1 3 1 0 × 1 1 0 1 0 1 0
= 80 × 10
1 1 3 0 2 × 1250 0 0 0 −2 0 2
0 0 0 1 0 −2 −2 2 2 0
3 1 1 0 −2 0 2 0 0 0
= 800 1 3 1 0 1 0 1 0 1 0
1 1 3 0 0 0 0 −2 0 2
0 0 0 1 0 −2 −2 2 2 0
3.98 Two Dimensional Problems
−6 + 1 + 0 + 0 0+0+0+0 6+1+0+0 0+0−2+0 0+1+0+0 0+1+2+0
= 800 −2 + 3 + 0 + 0 0+0+0+0 2+3+0+0 0+0−2+0 0+3+0+0 0+0+2+0
−2 + 1 + 0 + 0 0+0+0+0 2+1+0+0 0+0−6+0 0+1+0+0 0+0+6+0
0+0+0+0 0+0+0−2 0+0+0−2 0+0+0+2 0+0+0+2 0+0+0+0
−5 0 7 −2 1 2
[𝐷][𝐵] = 800 1 0 5 −2 3 2 …(8)
−1 0 3 −6 1 6
0 −2 −2 2 2 0
We know that,
−50 0 50 0 0 0
1 25 0 25 0 25 0
[𝐵] =
2 × 1250 0 0 0 −50 0 50
0 −50 −50 50 50 0
−2 0 2 0 0 0
25 1 0 1 0 1 0
[𝐵] =
2 × 1250 0 0 0 −2 0 2
0 −2 −2 2 2 0
−2 1 0 0
⎡ 0 0 0 −2⎤
⎢ 2 1 0 −2⎥⎥
[𝐵] = 0.01 ⎢ …(9)
⎢ 0 0 −2 2 ⎥
⎢ 0 1 0 2⎥
⎣ 0 0 2 0⎦
−2 1 0 0
⎡0 0 0 −2⎤⎥ −5 0 7 −2 1 2
⎢
[𝐵] [𝐷][𝐵] = 0.01 ⎢ 2 1 0 −2⎥ × 800 1 0 5 −2 3 2
⎢0 0 −2 2 ⎥ −1 0 3 −6 1 6
⎢0 1 0 2⎥ 0 −2 −2 2 2 0
⎣0 0 2 0⎦
−2 1 0 0
⎡ 0 0 0 −2⎤ −5 0 7 −2 1 2
⎢
= 0.01 × 800 ⎢
2 1 0 −2⎥⎥ 1 0 5 −2 3 2
⎢ 0 0 −2 2 ⎥ −1 0 3 −6 1 6
⎢0 1 0 2⎥ 0 −2 −2 2 2 0
⎣0 0 2 0⎦
10 + 1 + 0 + 0 0 + 0 + 0 + 0 −14 + 5 + 0 + 0 4 − 2 + 0 + 0 −2 + 3 + 0 + 0 −4 + 2 + 0 + 0
⎡ 0+0+0+0 0+0+0+4 0+0+0+4 0+0+0−4 0+0+0−4 0+0+0+0 ⎤
⎢−10 + 1 + 0 + 0 0 + 0 + 0 + 4 14 + 5 + 0 + 0 −4 − 2 + 0 − 0 2 + 3 + 0 − 4 4 + 2 + 0 + 0 ⎥⎥
= 800 ⎢ 0 + 0 + 2 + 0 0+0+0−4 0+0−6−4 0 + 0 + 12 + 4 0 + 0 + 0 + 2 0 + 0 − 12 + 0 ⎥
⎢
⎢ 0+1+0+0 0+0+0−4 0+5+0−4 0−2+0+4 0+3+0+4 0+2+0+0 ⎥
⎣ 0+0−2+0 0+0+0+0 0+0+6+0 0 + 0 − 12 + 0 0 + 0 + 2 + 0 0 + 0 + 12 + 0 ⎦
 Finite Element Analysis 3.99
11 0 −9 2 1 −2
⎡0 4 4 −4 −4 0 ⎤
⎢−9 4 23 −10 1 6 ⎥⎥
[𝐵] [𝐷][𝐵] = 8 ⎢
⎢ 2 −4 −10 16 2 −12⎥
⎢ 1 −4 1 2 7 2 ⎥
⎣−2 0 6 −12 2 12 ⎦
Substitute [𝐵] [𝐷][𝐵] value in equation (1),
11 0 −9 2 1 −2
⎡0 4 4 −4 −4 0 ⎤
⎢ 23 −10 6 ⎥⎥
⇒ [𝐾] = 2𝜋𝑟𝐴 × 8 ⎢−9 −4 4 1
⎢ 2 −10 16 2 −12⎥
⎢ 1 −4 1 2 7 2 ⎥
⎣−2 0 6 −12 2 12 ⎦
11 0 −9 2 1 −2
⎡0 4 4 −4 −4 0 ⎤
⎢−9 4 23 −10 1 6 ⎥⎥
= 2 × 𝜋 × 33.333 × 1250 × 8 ⎢ 2 −4 −10 16 2 −12⎥
⎢
⎢1 −4 1 2 7 2 ⎥
⎣−2 0 6 −12 2 12 ⎦
11 0 −9 2 1 −2
⎡0 4 4 −4 −4 0 ⎤
⎢−9 4 23 −10 1 6 ⎥⎥
Stiffness matrix [𝐾] = 2.094 × 10 ⎢ 2 −4 −10 16 2 −12⎥
⎢
⎢ 1 −4 1 2 7 2 ⎥
⎣−2 0 6 −12 2 12 ⎦
It may be noted that stiffness matrix [K] is symmetric.
Result:
11 0 −9 2 1 −2
⎡0 4 4 −4 −4 0 ⎤
⎢−9 4 23 −10 1 6 ⎥⎥
Stiffness matrix [𝐾] = 2.094 × 10 ⎢ 2 −4 −10 16 2 −12⎥ 𝑁/𝑚𝑚
⎢
⎢ 1 −4 1 2 7 2 ⎥
⎣−2 0 6 −12 2 12 ⎦

Example 3.13
For the axisymmetric element shown in Fig. (i), determine the element stresses.
Take E = 2.1 × 105 N/mm² and v=0.25.
The co-ordinates shown in Fig. (i) are in millimeters. The nodal displacements are:
3.100 Two Dimensional Problems

u1 = 0.05 mm ; w1 = 0.03 mm
u2 = 0.02 mm ; w2 = 0.02 mm
u3 = 0 mm ; w3 = 0 mm

Given:
r1 = 0 mm ; z1 = 0 mm
r2 = 60 mm ; z2 = 0 mm
r3 = 30 mm ; z3 = 50 mm
Nodal displacements:
u1 = 0.05 mm ; w1 = 0.03 mm
u2 = 0.02 mm ; w2 = 0.02 mm
u3 = 0 mm ; w3 = 0 mm
Young’s modulus, E = 2.1  105 N/mm2
Poisson’s ratio, v = 0.25
To find: Element stresses:
(i) Radial stress, 𝜎
(ii) Circumferential stress, 𝜎
(iii) Longitudinal stress, 𝜎
(iv) Shear stress, 𝜏
 Finite Element Analysis 3.101

Solution: we know that,

Stress {  } = [ D ] [ B ] { u }
𝑢
𝑒 ⎧𝑤 ⎫
𝑒 ⎪𝑢 ⎪
⇒ 𝑒 = [𝐷][𝐵] 𝑤 …(1)
⎨ ⎬
𝛾 ⎪𝑢 ⎪
⎩𝑤 ⎭
Area of the triangular element,
1 1
A = × Breadth × Height = × 60 × 50
2 2
A = 1500 𝑚𝑚 …(2)
𝑟 +𝑟 +𝑟 0 + 60 + 30
Co − ordinate, 𝑟= =
3 3
r = 30 mm …(3)
𝑧 +𝑧 +𝑧 0 + 0 + 50
𝑧= =
3 3
z = 16.667 mm …(4)
we know that,
Stress - Strain relationship matrix
0
⎡1 − 𝑣 𝑣 𝑣
0 ⎤
𝐸 ⎢ 𝑣
= 1−𝑣 𝑣 0 ⎥
(1 + 𝑣)(1 − 2𝑣) ⎢ 𝑣 𝑣 1 − 𝑣 1 − 2𝑣 ⎥
⎢ 0 0 ⎥
⎣ 0
2 ⎦
0
⎡1 − 0.25 0.25 0.25
0 ⎤
2.1 × 10 ⎢ 0.25 1 − 0.25 0.25 ⎥
⇒ [𝐷] = ⎢ 0
(1 + 0.25)(1 − (2 × 0.25)) 0.25 0.25 1 − 0.25 1 − 2(0.25)⎥
⎢ ⎥
⎣ 0 0 0 ⎦
2
0.75 0.25 0.25 0 3 1 1 0
= 336 × 10 0.25 0.75 0.25 0 = 336 × 10 × 0.25 1 3 1 0
0.25 0.25 0.75 0 1 1 3 0
0 0 0 0.25 0 0 0 1
3.102 Two Dimensional Problems
3 1 1 0
[𝐷] = 80 × 10 1 3 1 0 …(5)
1 1 3 0
0 0 0 1
We know that, Strain - Displacement matrix
𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
[𝐵] = 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 ⎥ … (6)
2𝐴 ⎢ 0 𝛾 0 𝛾 0 𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
Where,
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (60 × 50) − (30 × 0)
𝜶𝟏 = 𝟑𝟎𝟎𝟎 𝒎𝒎𝟐
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (30 × 0) − (0 × 50)
𝜶𝟐 = 𝟎
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (0 × 0) − (60 × 0)
𝜶𝟑 = 𝟎
𝛽 = 𝑧 − 𝑧 = 0 − 50
𝜷𝟏 = −𝟓𝟎 𝒎𝒎
𝛽 = 𝑧 − 𝑧 = 50 − 0
𝜷𝟐 = 𝟓𝟎 𝒎𝒎
𝛽 =𝑧 −𝑧 = 0−0
𝜷𝟑 = 𝟎
𝛾 = 𝑟 − 𝑧 = 30 − 60
𝜸𝟏 = −𝟑𝟎 𝒎𝒎
𝛾 = 𝑟 − 𝑟 = 0 − 30
𝜸𝟐 = −𝟑𝟎 𝒎𝒎
𝛾 = 𝑟 − 𝑧 = 60 − 0
𝛾 = 60 𝑚𝑚
 Finite Element Analysis 3.103
𝛼 𝛾 𝑧 3000 (−30 × 16.667)
⇒ +𝛽 + = + (−50) + = 33.33 𝑚𝑚
𝑟 𝑟 23.334 30
𝛼 𝛾 𝑧 (−30 × 16.667)
⇒ +𝛽 + = 0 + 50 + = 33.333 𝑚𝑚
𝑟 𝑟 30
𝛼 𝛾 𝑧 60 × 16.667
⇒ +𝛽 + = 0+0+ = 33.33 𝑚𝑚
𝑟 𝑟 30

Substitute, A, 𝛽 , 𝛽 , 𝛽 , +𝛽 + , +𝛽 + , +𝛽 + , 𝛾 , 𝛾 and 𝛾
values in equation (6),
−50 0 50 0 0 0
1 33.33 0 33.33 0 33.33 0
[𝐵] =
2 × 1500 0 −30 0 −30 0 60
−30 −50 −30 50 60 0
−50 0 50 0 0 0
[𝐵] = 3.3333 × 10 33.33 0 33.33 0 33.33 0 … (7)
0 −30 0 −30 0 60
−30 −50 −30 50 60 0
3 1 1 0
⇒ [𝐷][𝐵] = 80 × 10 1 3 1 0 × 3.3333
1 1 3 0
0 0 0 1
−50 0 50 0 0 0
× 10 33.33 0 33.33 0 33.33 0
0 −30 0 −30 0 60
−30 −50 −30 50 60 0
3 1 1 0 −50 0 50 0 0 0
= 28 1 3 1 0 × 33.33 0 33.33 0 33.33 0
1 1 3 0 0 −30 0 −30 0 60
0 0 0 1 −30 −50 −30 50 60 0
−116.67 −30 183.33 −30 33.33 60
[𝐷][𝐵] = 28 49.99 −30 149.99 −30 99.99 60 …(8)
−16.67 −90 83.33 −90 33.33 180
−30 −50 −30 50 60 0
Substitute [𝐵] [𝐷][𝐵] value 𝑢 , 𝑤 , 𝑢 , 𝑤 and 𝑢 , 𝑤 values in equation (1),
equation (1) becomes,
3.104 Two Dimensional Problems
0.05
𝜎 −116.67 −30 183.33 −30 33.33 60 ⎧ ⎫
𝜎 ⎪0.03⎪
0.02
𝜎 = 28 49.99 −30 149.99 −30 99.99 60
−16.67 −90 83.33 −90 33.33 180 ⎨0.02⎬
𝜏 −30 −50 −30 50 60 0 ⎪ 0 ⎪
⎩ 0 ⎭
𝜎 −3.666
𝜎 4
⇒ 𝜎 = 28
−3.666
𝜏 −2.6
𝜎 −102.65
𝜎 112
𝜎 = 28
−102.65
𝜏 −72.8

⇒ Radial stress, 𝜎 = -102.65 N/mm2

Circumferential stress, 𝜎 = 112 N/mm2

Longitudinal stress, 𝜎 = -102.65 N/mm2

Shear stress, 𝜏 = -72.8 N/mm2

Result: Element stresses:
Radial stress, 𝜎 = -102.65 N/mm2

Circumferential stress, 𝜎 = 112 N/mm2

Longitudinal stress, 𝜎 = -102.65 N/mm2

Shear stress, 𝜏 = -72.8 N/mm2

Example 3.14:
A long hollow cylinder of inside diameter 100 mm and outside diameter 140 mm
is subjected to an internal pressure of 4 N/mm2 as shown in Fig. (i). By using two elements
on the 15 mm length shown in Fig. (i), calculate the displacements at the inner radius.
 Finite Element Analysis 3.105

Take E = 2  105 N/mm2 and v = 0.3.

Given:

Fig(ii)
Inner diameter, de = 100 mm
Inner radius, re = 50 mm
Outer diameter, De = 140 mm
3.106 Two Dimensional Problems

Outer radious, Re = 70 mm
Internal Pressure, P = 4 N/mm2
Length , le = 15 mm
Young’s modulus, E = 2.1  105 N/mm2
Poisson’s ratio, v = 0.3
To find: Nodal displacements: 𝑢 , 𝑤 , 𝑢 , 𝑤 , 𝑢 , 𝑤 , 𝑢 , 𝑤
Solution: For element (1) (Nodal displacements: 𝑢 , 𝑤 , 𝑢 , 𝑤 , 𝑢 , 𝑤 , 𝑢 , 𝑤 )

Fig. (iii)
Co-ordinates:
At node 1: r1 = 50 mm ;
z1 = 15 mm
At node 2: r2 = 50 mm ;
z2 = 0 mm
At node 4: r3 = 70 mm ;
z3 = 15 mm
𝑟 +𝑟 +𝑟 50 + 50 + 70
We know that, 𝑟= =
3 3
 Finite Element Analysis 3.107

r = 56.6667 mm …(1)
𝑧 +𝑧 +𝑧 15 + 0 + 15
𝑧= =
3 3
z = 10 mm …(2)
Area of the triangular element,
1 1
A = × Breadth × Height = × (70 − 50) × 15
2 2
1
A = × 20 × 15
2
A = 150 𝑚𝑚 …(3)
we know that,
stiffness for axisymmetric triangular element,
[𝐾] = 2𝜋𝑟 𝐴 [𝐵] [𝐷][𝐵] …(4)

Stress - Strain relationship matrix

0
⎡1 − 𝑣 𝑣 𝑣
0 ⎤
𝐸 ⎢ 𝑣
[𝐷] = 1−𝑣 𝑣 0 ⎥
(1 + 𝑣)(1 − 2𝑣) ⎢ 𝑣 𝑣 1 − 𝑣 1 − 2𝑣 ⎥
⎢ 0 0 ⎥
⎣ 0
2 ⎦
0
⎡1 − 0.3 0.3 0.3
0 ⎤
2 × 10 ⎢ 0.3 1 − 0.3 0.3 ⎥
⇒ [𝐷] = ⎢ 0
(1 + 0.3)(1 − (2 × 0.3)) 0.3 0.25 1 − 0.3 1 − 2(0.3)⎥
⎢ ⎥
⎣ 0 0 0
2 ⎦
0.7 0.3 0.3 0
2 × 10 0.3 0.7 0.3 0
=
0.5 0.3 0.3 0.7 0
0 0 0 0.2
0.7 0.3 0.3 0
[𝐷] = 3846.6153 × 10 0.3 0.7 0.3 0
0.3 0.3 0.7 0
0 0 0 0.2
3.108 Two Dimensional Problems

We know that, Strain - Displacement matrix

𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
[𝐵] = 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 ⎥ … (5)
2𝐴 ⎢ 0 𝛾 0 𝛾 0 𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
Where,
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (50 × 15) − (70 × 0)
𝜶𝟏 = 𝟕𝟓𝟎 𝒎𝒎𝟐
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (70 × 15) − (50 × 15)
𝜶𝟐 = 𝟑𝟎𝟎 𝒎𝒎𝟐
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (50 × 0) − (50 × 15)
𝜶𝟑 = −𝟕𝟓𝟎 𝒎𝒎𝟐
𝛽 = 𝑧 − 𝑧 = 0 − 15
𝜷𝟏 = −𝟏𝟓 𝒎𝒎
𝛽 = 𝑧 − 𝑧 = 15 − 15
𝜷𝟐 = 𝟎
𝛽 = 𝑧 − 𝑧 = 15 − 0
𝜷𝟑 = 𝟏𝟓 𝒎𝒎
𝛾 = 𝑟 − 𝑧 = 70 − 50
𝜸𝟏 = 𝟐𝟎 𝒎𝒎
𝛾 = 𝑟 − 𝑟 = 50 − 70
𝜸𝟐 = −𝟐𝟎 𝒎𝒎
𝛾 = 𝑟 − 𝑧 = 50 − 50
𝛾 = 0 𝑚𝑚
𝛼 𝛾 𝑧 750 20 × 10
⇒ +𝛽 + = + (−15) + = 1.7647 𝑚𝑚
𝑟 𝑟 56.6667 56.6667
 Finite Element Analysis 3.109
𝛼 𝛾 𝑧 300 −20 × 10
⇒ +𝛽 + = +0+ = 1.7647 𝑚𝑚
𝑟 𝑟 56.6667 56.6667
𝛼 𝛾 𝑧 −750
⇒ +𝛽 + = + 15 + 0 = 1.7647 𝑚𝑚
𝑟 𝑟 56.6667

Substitute, A, 𝛽 , 𝛽 , 𝛽 , +𝛽 + , +𝛽 + , +𝛽 + , 𝛾 , 𝛾 and 𝛾
values in equation (6),
−15 0 0 0 15 0
1 1.7647 0 1.7647 0 1.7647 0
[𝐵] =
2 × 150 0 0 0 −20 0 0
20 −15 −20 0 0 15
−15 0 0 0 15 0
[𝐵] = 3.3333 × 10 1.7647 0 1.7647 0 1.7647 0 … (6)
0 0 0 −20 0 0
20 −15 −20 0 0 15
0.7 0.3 0.3 0
⇒ [𝐷][𝐵] = 384.6153 × 10 0.3 0.7 0.3 0 × 3.3333
0.3 0.3 0.7 0
0 0 0 0.2
−15 0 0 0 15 0
× 10 1.7647 0 1.7647 0 1.7647 0
0 0 0 −20 0 0
20 −15 −20 0 0 15
−9.9706 6 0.5294 −6 11.0294 0
[𝐷][𝐵] = 1.282 × 10 −3.2647 6 1.2353 −6 5.7353 0
−3.9706 14 0.5294 −14 5.0294 0
4 −3 −4 0 0 3

We know that,
−15 0 0 0 15 0
[𝐵] = 3.3333 × 10 1.7647 0 1.7647 0 1.7647 0 … (6)
0 0 0 −20 0 0
20 −15 −20 0 0 15
3.110 Two Dimensional Problems
−15 1.7647 0 20
⎡ 0 0 20 −15⎤
⎢ 0 1.7647 0 −20⎥⎥
[𝐵] = 3.3333 × 10 ⎢
⎢ 0 0 −20 0 ⎥
⎢ 15 1.7647 0 0 ⎥
⎣ 0 0 0 15 ⎦
−15 1.7647 0 20
⎡
⎢
0 0 20 −15⎤⎥
[𝐵] [𝐷][𝐵] = 3.333 × 10−3 ⎢ 0 1.7647 0 −20⎥ × 1.28
⎢ 0 0 −20 0 ⎥
⎢ 15 1.7647 0 0 ⎥
⎣ 0 0 0 15 ⎦
−9.9706 6 0.5294 −6 11.0294 0
× 10 −3 −3.2647 6 1.2353 −6 5.7353 0
−3.9706 14 0.5294 −14 5.0294 0
4 −3 −4 0 0 3
223.7978 −139.4118 −85.7611 79.412 −155.32 60
⎡ −139.412 325 70.588 −280 −100.588 −45⎤
⎢ 82.18 10.1211 −60⎥⎥
[𝐵] [𝐷][𝐵] = 4.2733 ⎢ −85.7612 70.588 −10.588
⎢ 79.412 −280 −10.588 280 −100.588 0 ⎥
⎢−155.3202 −100.5882 10.1210 −100.588 175.5621 0 ⎥
⎣ 60 −45 −60 0 0 45 ⎦

Substitute [𝐵] [𝐷][𝐵] value in equation (1),

[ K ]1 = 2    56.6667  150  4.2733 
223.7978 −139.4118 −85.7611 79.412 −155.32 60
⎡ −139.412 325 70.588 −280 −100.588 −45⎤
⎢ −85.7612 82.18 10.1211 −60⎥⎥
⎢ 70.588 −10.588
⎢ 79.412 −280 −10.588 280 −100.588 0 ⎥
⎢−155.3202 −100.5882 10.1210 −100.588 175.5621 0 ⎥
⎣ 60 −45 −60 0 0 45 ⎦
[𝐾]
223.7978 −139.4118 −85.7611 79.412 −155.32 60
⎡ −139.412 325 70.588 −280 −100.588 −45⎤
⎢ −85.7612 70.588 82.18 −10.588 10.1211 −60⎥⎥
= 228.2246 × 10 ⎢ 79.412 −280 −10.588 280 −100.588 0 ⎥
⎢
⎢−155.3202 −100.5882 10.1210 −100.588 175.5621 0 ⎥
⎣ 60 −45 −60 0 0 45 ⎦
 Finite Element Analysis 3.111
𝑢 𝑤 𝑢 𝑤 𝑢 𝑤
51.076 −31.817 −19.573 18.124 −35.448 13.693 𝑢
⎡−31.817 74.173 16.110 −63.903 22.957 −10.270⎤ ⎧𝑤 ⎫
⎢−19.573 ⎪𝑢 ⎪
16.110 18.755 −2.416 2.310 −13.693⎥
[𝐾] = 10 ⎢ 18.124 −63.903 −2.416 63.903 −22.957 0 ⎥ 𝑤
⎨ ⎬
⎢−35.448 22.957 2.310 −22.957 40.068 0 ⎥ ⎪𝑢 ⎪
⎣ 13.693 −10.270 −13.693 0 0 10.270 ⎦ ⎩𝑤 ⎭
… (7)
For element (1) (Nodal displacements: 𝑢 , 𝑤 , 𝑢 , 𝑤 , 𝑢 , 𝑤 )

Fig. (iv)
Co-ordinates:
At node 2: r1 = 50 mm ;
z1 = 0 mm
At node 3: r2 = 70 mm ;
z2 = 0 mm
At node 4: r3 = 70 mm ;
z3 = 15 mm
𝑟 +𝑟 +𝑟 50 + 70 + 70
We know that, 𝑟= =
3 3
r = 63.333 mm
3.112 Two Dimensional Problems
𝑧 +𝑧 +𝑧 0 + 0 + 15
𝑧= =
3 3

z = 5 mm

Area of the triangle,

1 1
A = × Breadth × Height = × (70 − 50) × 15
2 2

A = 150 𝑚𝑚

we know that,

stiffness for axisymmetric triangular element,

[𝐾] = 2𝜋𝑟 𝐴 [𝐵] [𝐷][𝐵] …(8)

Stress - Strain relationship matrix

0
⎡1 − 𝑣 𝑣 𝑣
0 ⎤
𝐸 ⎢ 𝑣
[𝐷] = 1−𝑣 𝑣 0 ⎥⎥
(1 + 𝑣)(1 − 2𝑣) ⎢ 𝑣 𝑣 1 − 𝑣 1 − 2𝑣
⎢ 0 0 ⎥
⎣ 0
2 ⎦
0
⎡1 − 0.3 0.3 0.3
0 ⎤
2 × 10 ⎢ 0.3 1 − 0.3 0.3 ⎥
⇒ [𝐷] = ⎢ 0 ⎥
(1 + 0.3)(1 − (2 × 0.3)) 0.3 0.25 1 − 0.3 1 − 2(0.3)⎥
⎢
⎣ 0 0 0
2 ⎦

0.7 0.3 0.3 0

2 × 10 0.3 0.7 0.3 0
=
0.5 0.3 0.3 0.7 0
0 0 0 0.2
0.7 0.3 0.3 0
[𝐷] = 384. .6153 × 10 0.3 0.7 0.3 0 …(9)
0.3 0.3 0.7 0
0 0 0 0.2
 Finite Element Analysis 3.113

We know that, Strain - Displacement matrix

𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
[𝐵] = 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟 ⎥ … (10)
2𝐴 ⎢ 0 𝛾 0 𝛾 0 𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
Where,
𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (70 × 15) − (70 × 0)

𝜶𝟏 = 𝟏𝟎𝟓𝟎 𝒎𝒎𝟐

𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (70 × 0) − (50 × 15)

𝜶𝟐 = −𝟕𝟓𝟎 𝒎𝒎𝟐

𝛼 = 𝑟 𝑧 − 𝑟 𝑧 = (50 × 0) − (70 × 15)

𝜶𝟑 = 𝟎

𝛽 = 𝑧 − 𝑧 = 0 − 15

𝜷𝟏 = −𝟏𝟓 𝒎𝒎

𝛽 = 𝑧 − 𝑧 = 15 − 0

𝜷𝟐 = 𝟏𝟓 𝒎𝒎

𝛽 =𝑧 −𝑧 = 0−0

𝜷𝟑 = 𝟎

𝛾 = 𝑟 − 𝑧 = 70 − 70

𝜸𝟏 = 𝟎

𝛾 = 𝑟 − 𝑟 = 50 − 70

𝜸𝟐 = −𝟐𝟎 𝒎𝒎

𝛾 = 𝑟 − 𝑧 = 70 − 50
3.114 Two Dimensional Problems

𝛾 = 20 𝑚𝑚
𝛼 𝛾 𝑧 1050
⇒ +𝛽 + = + (−15) + 0 = 1.579 𝑚𝑚
𝑟 𝑟 63.3333
𝛼 𝛾 𝑧 −750 −20 × 5
⇒ +𝛽 + = + 15 + = 1.579 𝑚𝑚
𝑟 𝑟 63.3333 63.3333
𝛼 𝛾 𝑧 20 × 5
⇒ +𝛽 + = 0+0+ = 1.579 𝑚𝑚
𝑟 𝑟 63.3333

Substitute, A, 𝛽 , 𝛽 , 𝛽 , +𝛽 + , +𝛽 + , +𝛽 + , 𝛾 , 𝛾 and 𝛾
values in equation (10), we get
−15 0 15 0 0 0
1 1.579 0 1.579 0 1.579 0
[𝐵] =
2 × 150 0 0 0 −20 0 20
0 −15 −20 15 20 0
−15 0 15 0 0 0
[𝐵] = 3.3333 × 10 1.579 0 1.579 0 1.579 0
0 0 0 −20 0 20
0 −15 −20 15 20 0
0.7 0.3 0.3 0
⇒ [𝐷][𝐵] = 384.6153 × 10 0.3 0.7 0.3 0 × 3.3333
0.3 0.3 0.7 0
0 0 0 0.2
−15 0 15 0 0 0
× 10 1.579 0 1.579 0 1.579 0
0 0 0 −20 0 20
0 −15 −20 15 20 0
−10.0263 0 10.9737 −6 0.4737 6
[𝐷][𝐵] = 1.282 × 10 −3.3947 0 5.6053 −6 1.1053 6
−4.0263 0 4.9737 −14 0.4737 14
0 −3 −4 3 4 0
We know that,
−15 0 15 0 0 0
[𝐵] = 3.3333 × 10 1.579 0 1.579 0 1.579 0
0 0 0 −20 0 20
20 −15 −20 15 20 15
 Finite Element Analysis 3.115
−15 1.579 0 0
⎡ 0 0 0 −15⎤
⎢ 0 −20⎥⎥
[𝐵] = 3.3333 × 10 ⎢ 15 1.579
⎢ 0 0 −20 15 ⎥
⎢ 0 1.579 0 20 ⎥
⎣ 0 0 20 0 ⎦
−15 1.579 0 0
⎡
⎢
0 0 0 −15⎤⎥
[𝐵] [𝐷][𝐵] = 3.333 × 10−3 ⎢ 15 1.579 0 −20⎥ × 1.28
⎢ 0 0 −20 15 ⎥
⎢ 0 1.579 0 20 ⎥
⎣ 0 0 20 0 ⎦
−10.0263 0 10.9737 −6 0.4737 6
× 10−3 −3.3947 0 5.6053 −6 1.1053 6
−4.0263 0 4.9737 −14 0.4737 14
0 −3 −4 3 4 0
145.034 0 −155.755 80.526 −5.360 −80.526
⎡ 0 45 60 −45 −160 0 ⎤
⎢−155.755 60 253.456 −159.474 −71.149 99.474 ⎥
[𝐵] [𝐷][𝐵] = 4.2733 ⎢ ⎥
⎢ 80.526 −45 −159.474 325 50.526 −280 ⎥
⎢ −5.360 −60 −71.149 50.526 81.745 9.474 ⎥
⎣ −80.526 0 99.474 −280 9.474 280 ⎦

Substitute [𝐵] [𝐷][𝐵] value in equation (8), we get,

[ K ]2 = 2    63.3333  150  4.2733 
145.034 0 −155.755 80.526 −5.360 −80.526
⎡ 0 45 60 −45 −160 0 ⎤
⎢−155.755 60 253.456 −159.474 −71.149 99.474 ⎥⎥
⎢
⎢ 80.526 −45 −159.474 325 50.526 −280 ⎥
⎢ −5.360 −60 −71.149 50.526 81.745 9.474 ⎥
⎣ −80.526 0 99.474 −280 9.474 280 ⎦

[ 𝐾 ] = 255.074
145.034 0 −155.755 80.526 −5.360 −80.526
⎡ 0 45 60 −45 −160 0 ⎤
⎢−155.755 60 253.456 −159.474 −71.149 99.474 ⎥⎥
× 10 ⎢ 80.526 −45 −159.474 325 50.526 −280 ⎥
⎢
⎢ −5.360 −60 −71.149 50.526 81.745 9.474 ⎥
⎣ −80.526 0 99.474 −280 9.474 280 ⎦
3.116 Two Dimensional Problems
𝑢 𝑤 𝑢 𝑤 𝑢 𝑤
36.9994 0 −39.729 20.540 −1.367 −20.540 𝑢
⎡ 0 11.478 15.304 −11.478 −15.304 0 ⎤ ⎧ 𝑤 ⎫
⎢−39.729 15.304 ⎥ ⎪ 𝑢 ⎪
64.650 −40.678 −18.148 25.373
[ 𝐾 ] = 10 ⎢ 20.540 −11.478 −40.678 82.899 12.887 −71.421⎥ 𝑤
⎨ ⎬
⎢ −1.367 −15.304 −18.148 12.887 20.851 2.417 ⎥ ⎪ 𝑢 ⎪
⎣ 20.540 0 25.373 −71.421 2.417 71.421 ⎦ ⎩𝑤 ⎭
… (11)
Assemble the equation (7) and (11),
Global stiffness Matrix [ K ] =
𝑢 𝑤 𝑢 𝑤 𝑢 𝑤 𝑢 𝑤
- -
51.076 -31.817 18.124 13.639
19.573 35.448
+ + + 0 0 + 𝑢
+ +
0 0 0 0
0 0
-31.817 74.173 16.110 -63.903 22.957 -10.270
+ + + + 0 0 + + 𝑤
0 0 0 0 0 0
-
-19.573 16.110 18.755 -2.416 20.540 2.310 -13.693
39.729
+ + + + + - - 𝑢
+
0 0 36.994 0 0 1.367 20.540
0
106 
-
18.124 -63.903 -2.416 63.903 0 0 0
22.957
+ + + + + - + 𝑤
-
0 0 0 11.478 15.304 11.478 0
15.304
0 0
0 0 0 0
+ +
0 0 + + + + 𝑢
- -
15.304 64.650 -40.678 25.373
39.729 18.148
-
20.540 -11.478 82.899 12.887 -71.421
40.678
0 0 + + + + + 𝑤
+
0 0 0 0 0
0
 Finite Element Analysis 3.117
0
-35.448 22.957 2.310 -22.957 0 40.068 0
+
+ + - - + + + 𝑢
-
0 0 1.367 15.304 12.887 20.851 2.417
18.148
-
13.693 -10.270 0 0 0 0 10.270
13.693
+ + + + + + + 𝑤
-
0 0 0 25.373 -71.421 2.417 71.421
20.540
[Combining equation (7)+(11)]
Global stiffness Matrix [ K ] =
𝑢 𝑤 𝑢 𝑤 𝑢 𝑤 𝑢 𝑤
51.076 -31.817 -19.573 18.124 0 0 -35.448 13.639 𝑢
-31.817 74.173 16.110 -63.903 0 0 22.957 -10.270 𝑤

-19.573 16.110 55.749 -2.416 -39.729 20.540 0.943 -34.233 𝑢

106  18.124 -63.903 -2.416 75.381 15.304 -11.478 -38.261 0 𝑤

0 0 -39.729 15.304 64.650 -40.678 -18.148 25.373 𝑢
0 0 20.540 -11.478 -40.678 82.899 12.887 -71.421 𝑤

-35.448 22.957 0.943 -38.261 -18.148 12.887 60.919 2.417 𝑢

13.693 -10.270 -34.233 0 25.373 -71.421 2.417 81.691 𝑤
We know that, { F } = [ K ] { u }
𝐹 51.076 −31.817 −19.573 18.124 0 0 𝑢
⎧𝐹 ⎫ −35.448 13.6390
⎡−31.817 74.173 ⎤
⎪ ⎪ ⎢−19.573 16.110
16.110 −63.903 0 0 22.957 −10.270⎥ ⎧𝑤 ⎫
⎪𝐹 ⎪ ⎢
55.749 −2.416 −39.729 20.540 0.943 −34.233⎥ ⎪ ⎪
⎪𝑢 ⎪
𝐹 18.124 −63.903 −2.416 75.381 15.304 −11.478 −38.261 0 𝑤
= 10 ⎢ ⎥
⎨𝐹 ⎬ ⎢ 0 0 −39.729 15.304 64.650 −40.678 −18.148 25.373 ⎥ ⎨ 𝑢 ⎬
⎪𝐹 ⎪ ⎢ 0 0 20.540 −11.478 −40.678 82.899 12.887 −71.421⎥ ⎪𝑤 ⎪
⎪𝐹 ⎪ ⎢−35.448 22.957 0.943 −38.261 −18.148 12.887 60.919 2.417 ⎥ ⎪ 𝑢 ⎪
⎩𝐹 ⎭ ⎣ 13.693 −10.270 −34.233 0 25.373 −71.421 2.417 81.691 ⎦ ⎩𝑤 ⎭

Forces: we know that,

2𝜋 𝑟 × 𝑙 × 𝑃 2 × 𝜋 × 50 × 15 × 4
Forces, 𝐹 =𝐹 = =
2 2
𝐹 =𝐹 = 9424.77 𝑁
3.118 Two Dimensional Problems

The remaining forces are zero. i.e., 𝐹 , 𝐹 , 𝐹 , 𝐹 , 𝐹 and 𝐹 are zero.

Displacements [Refer Fig. (ii)]
1. Node 1 is moving in r direction. So. 𝑢 ≠ 0, but 𝑤 = 0.
2. Node 2 is moving in r direction. So. 𝑢 ≠ 0, but 𝑤 = 0.
3. Node 3 and 4 are fized. So. 𝑢 , 𝑤 , 𝑢 and 𝑤 are zero.
Substitute nodal forces and nodal displacements values in equation (12),
9424.77
⎧ 0 ⎫
⎪9424.77⎪
⎪ ⎪
(12) ⇒ 0
⎨ 0 ⎬
⎪ 0 ⎪
⎪ 0 ⎪
⎩ 0 ⎭
51.076 −31.817 −19.573 18.124 0 0 −35.448 13.6390 𝑢
⎡−31.817 74.173 16.110 −63.903 0 0 ⎤ ⎧0⎫
⎢−19.573 16.110 22.957 −10.270⎥
55.749 −2.416 −39.729 20.540 0.943 −34.233⎥ ⎪𝑢 ⎪
⎪ ⎪
⎢ 0
= 10 ⎢ 18.124 −63.903 −2.416 75.381 15.304 −11.478 −38.261 0 ⎥
⎢ 0 0 −39.729 15.304 64.650 −40.678 −18.148 25.373 ⎥ ⎨0⎬
⎢ 0 0 20.540 −11.478 −40.678 82.899 12.887 −71.421⎥ 0
⎪
⎪0⎪ ⎪
⎢−35.448 22.957 0.943 −38.261 −18.148 12.887 60.919 2.417 ⎥
⎣ 13.693 −10.270 −34.233 0 25.373 −71.421 2.417 81.691 ⎦ ⎩0⎭

Delete second row, second column, fourth row, fourth column, fifth row, fifth
column, sixth row, sixth column, seventh row, seventh column and eight row and eight
column of the above matrix. hence the equation reduces to.
9424.77 51.076 −19.573 𝑢
= 10
9424.77 −19.573 55.749 𝑢

9424.77 = 10 (51.076 𝑢 − 19.573𝑢 ) …(13)

9424.77 = 10 (−19.573 𝑢 + 55.749𝑢 ) …(14)

𝐸𝑞𝑛. (13) × 19.573 ⇒ 184.47 × 10 = 10 (999.710 𝑢 − 383.102𝑢 )

𝐸𝑞𝑛. (14) × 51.076 ⇒ 481.379 × 10 = 10 (−999.710 𝑢 + 2847.4352𝑢 )

Solving, 665.849 × 10 = 10 (2464.333)𝑢

⇒ 𝑢 = 2.70 × 10 𝑚𝑚
 Finite Element Analysis 3.119

Substitute 𝑢 value in equation (13) or (14)

⇒ 9424.77 = 10 (51.076 𝑢
− 19.573 × 2.70 × 10 𝑚𝑚)

⇒ 𝑢 = 2.88 × 10 𝑚𝑚

Result: Displacements:
𝑢 = 2.88 × 10 𝑚𝑚 𝑤 =0

𝑢 = 2.70 × 10 𝑚𝑚 𝑤 =0

𝑢 =0 𝑤 =0

𝑢 =0 𝑤 =0
3.120 Two Dimensional Problems

TWO MARKS QUESTIONS & ANSWERS

1. How do you define two dimensional elements?

Two dimensional elements are defined by three or more nodes in a two dimensional
plane (i.e., x, y plane). The basic element useful for two dimensional analysis is the
triangular element.
2. What is CST element?
Three noded triangular element is known Constant Strain Triangle (CST), which is
shown in Fig.(i). It has six unknown displacement degrees of freedom (u1 v1 , u2 v2 ,
u3 v3). The element is called CST because it has a constant strain throughout it.

3. What is LST element?

Six noded triangular element is known as Linear Strain Triangle (LST), which is
shown in Fig.(ii). It has twelve unknown displacement degrees of freedom. The
displacement functions for the element are quadratic instead of linear as in the CST.
 Finite Element Analysis 3.121

4. What is QST element?

Ten noded triangular element is known as Quadratic Strain Triangle (QST) which is
shown in Fig.(iii). It is also called cubic displacement triangle.

5. What is meant by plane stress analysis?

Plane stress is defined to be a state of stress in which the normal stress (σ) and shear
stress (τ) directed perpendicular to the plane are assumed to be zero.
6. Define plane strain analysis.
Plane strain is defined to be a state of strain in which the strain normal to the xy plane
and the shear strains are assumed to be zero.
7. Write a displacement function equation for CST element.
𝑢
⎧𝑣 ⎫
𝑢(𝑥, 𝑦) 𝑁 0 𝑁 0 𝑁 0 ⎪𝑢 ⎪
Displacement function, u = = × 𝑣
𝑣(𝑥, 𝑦) 0 𝑁 0 𝑁 0 𝑁 ⎨ ⎬
⎪𝑢 ⎪
⎩𝑣 ⎭
Where, N1, N2, N3 are shape functions.
8. Write a strain-displacement matrix for CST element.
Strain-Displacement matrix for CST element is,

1 𝑞 0 𝑞 0 𝑞 0
[𝐵] = 0 𝑟 0 𝑟 0 𝑟
2A 𝑟 𝑞 𝑟 𝑞 𝑟 𝑞
Where, A = Area of the element
3.122 Two Dimensional Problems

𝑞 =𝑦 −𝑦 ; 𝑞 =𝑦 −𝑦 ; 𝑞 =𝑦 −𝑦
𝑟 =𝑥 −𝑥 ; 𝑟 =𝑥 −𝑥 ; 𝑟 = 𝑥 −𝑥
9. Write down the stress-strain relationship matrix for plane stress condition.
For plane stress problems, stress-strain relationship matrix is,
1 𝑣 0
E 𝑣 1 0
[𝐷] = 1−𝑣
1−𝑣
0 0
2
Where, E = Young’s modulus
v = Poisson’s ratio
10. Write down the stress-strain relationship matrix for plane strain condition.
For plane strain problems, Stress-strain relationship matrix is,
1−𝑣 𝑣 0
𝐸 𝑣 1−𝑣 0
[𝐷] = 1 − 2𝑣
(1 + 𝑣)(1 − 2𝑣)
0 0
2
11. Write down the stiffness matrix equation for two dimensional CST element.
Stiffness matrix [ K ] = [B]T [D] [B] A t …(1)
Where, [B]  Strain – Displacement matrix
[D]  Stress-strain matrix
A  Area of the element
t  Thickness of the element
12. Write down the expression for the shape functions for a constant strain
triangular element.
For CST element,
𝑝 +𝑞 𝑥+𝑟 𝑦
Shape Function, 𝑁 =
2𝐴
𝑝 +𝑞 𝑥+𝑟 𝑦
𝑁 =
2𝐴
 Finite Element Analysis 3.123
𝑝 +𝑞 𝑥+𝑟 𝑦
𝑁 =
2𝐴
where, p = 𝑥 𝑦 − 𝑥 𝑦 ; 𝑝 = 𝑥 𝑦 − 𝑥 𝑦 ; 𝑝 =𝑥 𝑦 −𝑥 𝑦
𝑞 =𝑦 −𝑦 ; 𝑞 =𝑦 −𝑦 ; 𝑞 =𝑦 −𝑦
𝑟 =𝑥 −𝑥 ; 𝑟 =𝑥 −𝑥 ; 𝑟 =𝑥 −𝑥
13. What is axisymmetric element?
Many three dimensional problems in engineering exhibit symmetry about an axis of
rotation. Such types of problems are solved by a special two dimensional element
called as axisymmetric element.
14. What are the conditions for a problem to be axisymmetric?
1. The problem domain must be symmetric about the axis of revolution.
2. All boundary conditions must be symmetric about the axis of revolution.
3. All loading conditions must be symmetric about the axis of revolution.
15. Write down the displacement equation for an axisymmetric triangular element.
Displacement function,
𝑢
⎧𝑢 ⎫
𝑢(𝑟, 𝑧) 𝑁 0 𝑁 0 𝑁 0 ⎪𝑢 ⎪
𝑢(𝑟, 𝑧) = =
𝑤(𝑟, 𝑧) 0 𝑁 0 𝑁 0 𝑁 ⎨𝑤 ⎬
⎪𝑤 ⎪
⎩𝑤 ⎭
16. Write down the shape functions for an axisymmetric triangular element.
𝛼 +𝛽 𝑟+𝛾 𝑧
Where, Shape function, 𝑁 =
2𝐴
𝛼 +𝛽 𝑟+𝛾 𝑧
𝑁 =
2𝐴
𝛼 +𝛽 𝑟+𝛾 𝑧
𝑁 =
2𝐴
Where, 𝛼 = 𝑟 𝑧 −𝑟 𝑧 ; 𝛼 =𝑟 𝑧 −𝑟 𝑧 ; 𝛼 =𝑟 𝑧 −𝑟 𝑧
𝛽 =𝑧 −𝑧 𝛽 =𝑧 −𝑧 𝛽 =𝑧 −𝑧
3.124 Two Dimensional Problems

𝛾 =𝑟 −𝑧 𝛾 =𝑟 −𝑟 𝛾 = 𝑟 −𝑧
17. Give the Strain-Displacement matrix equation for an axisymmetric triangular
element.
Strain -Displacement matrix,
𝛽 0 𝛽 0 𝛽 0
⎡𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 𝛼 𝛾 𝑧 ⎤
1 ⎢ +𝛽 + 0 +𝛽 + 0 +𝛽 + 0⎥
[𝐵] = 𝑟 𝑟 𝑟 𝑟 𝑟 𝑟
2𝐴 ⎢ 0 𝛾 0 𝛾 0
⎥
𝛾 ⎥
⎢
⎣ 𝛾 𝛽 𝛾 𝛽 𝛾 𝛽⎦
𝑟 +𝑟 +𝑟
Where, Co − ordinate, 𝑟 =
3
𝑧 +𝑧 +𝑧
𝑧=
3
18. Write down the Stress-Strain relationship matrix for an axisymmetric
triangular element.
Stress − Strain relationship matrix, [D]
0
⎡1 − 𝑣 𝑣 𝑣
0 ⎤
𝐸 ⎢ 𝑣
= 1−𝑣 𝑣 0 ⎥⎥
(1 + 𝑣)(1 − 2𝑣) ⎢ 𝑣 𝑣 1 − 𝑣 1 − 2𝑣
⎢ 0 0 ⎥
⎣ 0
2 ⎦
Where, E  Young’s modulus
v  Poisson’s ratio
19. Give the stiffness matrix equation for an axisymmetric triangular element.
Stiffness matrix [ K ] = 2 r A [B]T [D] [B]
𝑟 +𝑟 +𝑟
Where, Co − ordinate, 𝑟
3
A Area of the triangular element matrix
20. What are the ways in which a three dimensional problem can be reduced to a
two dimensional approach?
1. Plane stress: One dimension is too small when compared to other two
dimensions.
 Finite Element Analysis 3.125

Example: Gear - Thickness is small.

2. Plane strain: One dimension is too large when compared to other two
dimensions.
Example: Long pipe [Length is long compared to diameter]
3. Axisymmetric: Geometry is symmetric about the axis.
Example: Cooling tower
 UNIT 4
ISOPARAMETRIC ELEMENTS

4.1. ISOPARAMETRIC ELEMENTS

4.1.1. Introduction
The finite element method is a powerful technique for analysing engineering
problems involving complex and irregular geometrics. However, the two and three
dimensional elements (triangle, rectangle, brick) discussed in previous chapters cannot
be used efficiently for irregular geometrics
Consider a continuum shown in Fig.4.1(a) and it is discretized by using triangular
elements which is shown in Fig. 4.1(b)

Fig.4.1(a) Continum Fig.4.1(b) Continum is discretized by

triangular elements
It is difficult to represent the curved boundaries by straight edges elements. A
large number of elements may be used to obtain reasonable resemblance between original
body and the assemblage. In order to overcome this drawback, isoparametric elements
are used. I,e. for problems involving curved boundaries, a family of elements known as
"isoparametric elements" can be used.
The isoparametric concept was first brought out by Taig and latter on generalized
by B.M. Irons for mapping the curved boundaries. They brought out the concept of
4.2 Isoparametric Elements

mapping for regular triangular, rectangular elements and brick elements from natural co-
ordinate system to global cartesian system as shown in Fig.4.2, 4.3 & 4.4.

Fig.4.2. Concept of mapping in isoparametric elements (Triangular element)

Fig.4.3

Fig.4.4
In this chapter, method of co-ordinate transformation from natural co-ordinate
system to global co-ordinate system, shape function and stiffness matrix for four noded
quadrilateral element, numerical integration are presented.

4.2. NATURAL CO-ORDINATES

A natural co-ordinate system is used to define any point inside the element by a
set of dimensionless numbers whose magnitude never exceeds unity. This system is very
useful in assembling of stiffness matrices.
 Finite Element Analysis 4.3

(1) Natural Co-ordinates in One Dimension

Fig.4.5 Natural Co-ordinates for a line element

Consider a two noded line element as shown in Fig.4.5. Any point p inside the
line element is identified by two natural co-ordinates L, and L, and the cartesian co-
ordinate x. Node 1 and node 2 have the cartesian co-ordinates x, and x, respectively.
We know that,
Total weightage of natural co-ordinates at any point is unity.
i.e, L1 + L2 = 1 ...(4.1)
Any point x within the element can be expressed as a linear combination of the
nodal co-ordinates of nodes 1 and 2 as,
L1𝑥 + L2𝑥 = x
Arrange equation (5.1) and (5.2) in matrix form,
1 1 𝐿 1
=
𝑥 𝑥 𝐿 𝑥
𝐿 1 1 1
=
𝐿 𝑥 𝑥 𝑥
1 𝑥 −1 1
=
(𝑥 − 𝑥 ) −𝑥 1 𝑥
𝑎 𝑎 1 𝑎 −𝑎
𝑁𝑜𝑡𝑒: 𝑎 𝑎 =
( 𝑎 . 𝑎 ) − (𝑎 . 𝑎 ) −𝑎 𝑎
1 𝑥 −𝑥
= −𝑥 + 𝑥
𝑥 −𝑥
1 𝑥 −𝑥
= 𝑥−𝑥
𝑥 −𝑥
4.4 Isoparametric Elements
1 𝑥 −𝑥
= ∴ x − x, is the length of the element, 1]
𝑙 𝑥−𝑥
𝑥 −𝑥
𝐿
= 𝑥 −𝑙 𝑥
𝐿
𝑙
The variation of L, and L is shown in Fig.4.7 and Fig.4.8. L, is one at node 1 and
it is zero at node 2 whereas L is one at node 2 and it is zero at node 1.

Fig.4.6 Fig.4.7 Variation of L1 Fig.4.8 Variation of L2

Integration of polynomial terms in natural co-ordinates can be performed by using
the simple formula,

𝛼! 𝛽!
(𝐿 ) (𝐿 ) 𝑑𝑥 = ×𝑙
(𝛼 + 𝛽 + 1)!

Where, α! is the factorial of α

Natural Co-ordinate, ε
In one dimensional problem, the following type of natural co-ordinate is also used.
Consider a one dimensional element as shown in Fig.4.9.
In the local number scheme, the first node will be numbered 1 and the second
node 2.e the centre of nodes 1 and 2 and p is the point referred.
The natural co-ordinator & for any point in the element is defined as,
𝑝𝑐
ε= 𝑥 −𝑥
2

Fig.4.9.
 Finite Element Analysis 4.5
𝑝𝑐
ε = [∵ 𝑥 − 𝑥 = 𝑙]
𝑙
2
2 2
= 𝑝 𝑐 = (x − x ) [⸪pc = x − xc]
l l
2 𝑥 +𝑥 𝑥 +𝑥
= × 𝑥− ∵𝑥 =
l 2 2
2 𝑥 +𝑥
= × 𝑥−
l 2
2 𝑥 − 𝑥 + 2𝑥
= × 𝑥−
l 2
2 𝑙 + 2𝑥
= × 𝑥−
l 2
2 𝑙
𝜀− × 𝑥− +𝑥
1 2
𝜀𝑙 𝑙
= 𝑥− −𝑥
2 2
𝜀𝑙 𝑙
+ =𝑥−𝑥
2 2
𝑙
(𝜀 + 1) = 𝑥 − 𝑥
2

Applying boundary conditions,

At node 1, 𝑥=𝑥
𝑙
⇒ (1 + 𝜀) = 0
2
1+ ε = 0
ε = -1
At node 2, 𝑥 = 𝑥
𝑙
⇒ (1 + 𝜀) = 𝑥 − 𝑥
2
4.6 Isoparametric Elements
𝑙
(1 + 𝜀) = 1
2
1+ ε = 2
ε=1

Fig.4.10. Variation of natural co-ordinates, ε

Natural Co-ordinates in Two Dimensions
Consider a triangular element having 3 nodes as shown in Fig.4.11.
Let p is the point inside the element and it has 3 co-ordinates L₁, L2 and L3.
From the definition of natural co-ordinates, we know that,
𝐿 + 𝐿 + 𝐿 = 1
𝐿 𝑥 + 𝐿 𝑥 + 𝐿 𝑥 = 𝑥
𝐿 𝑦 + 𝐿 𝑦 + 𝐿 𝑦 = 𝑦
Assemble the above equations in matrix form,
1 1 1 𝐿 1
𝑥 𝑥 𝑥 𝐿 = 𝑥
𝑦 𝑦 𝑦 𝐿 𝑦
𝐿 1 1 1 1
𝐿 = 𝑥 𝑥 𝑥 𝑥
𝐿 𝑦 𝑦 𝑦 𝑦
1 1 1
Let D = 𝑥 𝑥 𝑥
𝑦 𝑦 𝑦
 Finite Element Analysis 4.7

𝐶
D =
|𝐷|

Fig.4.11
Coefficients of matrix D:
𝑥 𝑥
C = + 𝑦 𝑦 = 𝑥 𝑦 −𝑥 𝑦
𝑥 𝑥
C = − 𝑦 𝑦 = − (𝑥 𝑦 − 𝑥 𝑦 ) = 𝑥 𝑦 − 𝑥 𝑦
𝑥 𝑥
C = + 𝑦 𝑦 =𝑥 𝑦 −𝑥 𝑦

1 1
C = − = − (𝑦 − 𝑦 ) = 𝑦 − 𝑦
𝑦 𝑦
1 1
C = + = 𝑦 −𝑦
𝑦 𝑦
1 1
C = − = − (𝑦 − 𝑦 ) = 𝑦 − 𝑦
𝑦 𝑦
1 1
C = + =𝑥 −𝑥
𝑥 𝑥
1 1
C = − = − (𝑥 − 𝑥 ) = 𝑥 − 𝑥
𝑥 𝑥
4.8 Isoparametric Elements
1 1
C = + = 𝑥 −𝑥
𝑥 𝑥
𝑥 𝑦 −𝑥 𝑦 𝑥 𝑦 −𝑥 𝑦 𝑥 𝑦 −𝑥 𝑦
𝐶= 𝑦 −𝑦 𝑦 −𝑦 𝑦 −𝑦
𝑥 −𝑥 𝑥 −𝑥 𝑥 −𝑥
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
𝐶 = 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
1 1 1
D = 𝑥 𝑥 𝑥
𝑦 𝑦 𝑦
|D| = 1(𝑥 𝑦 − 𝑥 𝑦 ) − 1(𝑥 𝑦 − 𝑥 𝑦 ) + 1(𝑥 𝑦 − 𝑥 𝑦 )
Substitute CT and |D| values in equation
1
⇒ 𝐷 =
(𝑥 𝑦 − 𝑥 𝑦 ) − (𝑥 𝑦 −𝑥 𝑦 ) + (𝑥 𝑦 − 𝑥 𝑦 )
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
× 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥
Substitute |D| values in equation,
𝐿 1 1 1 1
𝐿 = 𝑥 𝑥 𝑥 𝑥
𝐿 𝑦 𝑦 𝑦 𝑦
𝐿 1
𝐿 =
𝐿 (𝑥 𝑦 − 𝑥 𝑦 ) − (𝑥 𝑦 − 𝑥 𝑦 ) + (𝑥 𝑦 − 𝑥 𝑦 )
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 1
× 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 𝑥
𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 𝑦
The area of the triangle ABC can be expressed as a function of the x, y co-
ordinates of the nodes 1, 2 and 3.

1 1 𝑥 𝑦
𝐴= 1 𝑥 𝑦
2 1 𝑥 𝑦
 Finite Element Analysis 4.9
1
= [1(𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )]
2
1
= [𝑥 𝑦 − 𝑥 𝑦 − 𝑥 𝑦 + 𝑥 𝑦 + 𝑥 𝑦 − 𝑥 𝑦 ]
2
1
= [𝑥 𝑦 − 𝑥 𝑦 − (𝑥 𝑦 − 𝑥 𝑦 ) + (𝑥 𝑦 − 𝑥 𝑦 )]
2
(𝑥 𝑦 − 𝑥 𝑦 ) − (𝑥 𝑦 − 𝑥 𝑦 ) + (𝑥 𝑦 − 𝑥 𝑦 ) = 2A

𝐿 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 1
1
𝐿 = 𝑥 𝑦 −𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 𝑥
𝐿 2A 𝑥 𝑦 − 𝑥 𝑦 𝑦 −𝑦 𝑥 −𝑥 𝑦
Integration of polynomial terms in natural co-ordinates for two dimensional
elements can performed by using the formula,
𝛼! 𝛽! 𝛾!
(𝐿 ) (𝐿 ) (𝐿 ) 𝑑𝐴 = × 2A
(𝛼 + 𝛽 + 𝛾 + 2)!

4.3. SOLVED PROBLEMS ON NATURAL CO-ORDINATES

Example 4.1

Calculate the value of 𝐿 𝐿 𝐿 𝑑𝐴.

Solution:

L L L 𝑑𝐴 … (1)

We know that,
𝛼! 𝛽! 𝛾!
(𝐿 ) (𝐿 ) (𝐿 ) 𝑑𝐴 = × 2A … (2)
(𝛼 + 𝛽 + 𝛾 + 2)!
Compare equation (1) and (2),
𝛼 = 1, 𝛽 = 1 𝑎𝑛𝑑 𝛾 = 1.
1! 1! 1!
L L L 𝑑𝐴 = × 2A
(1 + 1 + 1 + 2)!
4.10 Isoparametric Elements
1×1×1
= × 2A
5!
1
= × 2A
5×4×3×2×1
A
L L L 𝑑𝐴 =
60

Example 4.2

𝐃𝐞𝐭𝐞𝐫𝐦𝐢𝐧𝐞 𝐭𝐡𝐞 𝐯𝐚𝐥𝐮𝐞 𝐨𝐟 𝑳𝟏 (𝑳𝟐 )𝟐 (𝑳𝟑 ) 𝟑 𝒅𝑨.

𝑨

Solution: We know that,

𝛼! 𝛽! 𝛾!
(𝐿 ) (𝐿 ) (𝐿 ) 𝑑𝐴 = × 2A
(𝛼 + 𝛽 + 𝛾 + 2)!
Here, 𝛼 = 1, 𝛽= 2 and 𝛾= 3.
1! 2! 3! 1×2×3
L L L 𝑑𝐴 = × 2A = × 2A
(1 + 2 + 3 + 2)! 8!
1×2×1×3×2×1
= × 2A
8×7×6×5×4×3×2×1

A
L L L 𝑑𝐴 =
1680

Example 4.3
𝟏

𝐂𝐚𝐥𝐜𝐮𝐥𝐚𝐭𝐞 𝐭𝐡𝐞 𝐯𝐚𝐥𝐮𝐞 𝐨𝐟 𝑳𝟏 𝑳𝟐 𝒅𝒙.

𝟎

Solution: We know that,

𝛼! 𝛽!
𝐿 𝐿 𝑑𝑥 = × 𝑙
(𝛼 + 𝛽 + 1)!

Here 𝛼 = 1, 𝛽 = 1.
 Finite Element Analysis 4.11

1! 1! 1
𝐿 𝐿 𝑑𝑥 = × 𝑙= × 𝑙
(1 + 1 + 1)! 3!

1
= × 𝑙
3×2×1

1
𝐿 𝐿 𝑑𝑥 =
16

Example 4.4
𝟏
𝐃𝐞𝐭𝐞𝐫𝐦𝐢𝐧𝐞 𝐭𝐡𝐞 𝐯𝐚𝐥𝐮𝐞 𝐨𝐟 𝑳𝟏 𝟑 𝒅𝒙
𝟎

Solution: We know that,

𝛼! 𝛽!
𝐿 𝐿 𝑑𝑥 = × 𝑙
(𝛼 + 𝛽 + 1)!

Here, 𝛼 = 3, 𝛽 = 0
3! 0! (3 × 2 × 1)
𝐿 𝐿 𝑑𝑥 = × 𝑙 = × 𝑙
(3 + 0 + 1)! 4!
3×2×1
= × 𝑙
4×3×2×1
l
𝐿 𝐿 𝑑𝑥 =
4

4.4. ISOPARAMETRIC, SUPERPARAMETRIC AND SUBPARAMETRIC

ELEMENTS
Isoparametric Element
We know that, shape functions are used for defining the geometry and
displacements of the element. Consider a element shown in Fig.4.12.
4.12 Isoparametric Elements

Fig.4.12.
• Nodes used for defining geometry.
A Nodes used for defining displacements.
In this element, all the eight nodes are used in defining geometry as well as
displacements.
If the number of nodes used for defining the geometry is same as number of for
defining the displacements, then, it is known as isoparametric element.
Superparametric Element
Consider a element shown in Fig. 4.13.

Fig.4.13.
 Nodes used for defining geometry.
 A Nodes used for defining displacements.
In this element, eight nodes are used to define the geometry and four nodes are
used to define the displacements. If the number of nodes used for defining the geometry
 Finite Element Analysis 4.13

is more than number of nodes used for defining the displacements, then, it is known as
superparametric element.
Subparametric Element
Consider a element shown in Fig.4.14.

Fig.4.14
 Nodes used for defining geometry.
 A Nodes used for defining displacements.
In this element, four nodes are used to define the geometry and eight nodes are
used to define the displacements. If the number of nodes used for defining the geometry
is less than number of nodes used for defining the displacements, then it is known as
subparametric element.

4.5. ONE DIMENSIONAL SHAPE FUNCTIONS FOR

ISOPARAMETRIC FORMULATION OF THE BAR ELEMENT
Consider a bar element with nodes 1 and 2 as shown in Fig.4.15. 𝑢1 and u2 are the
displacements at the respective nodes. So, u1 and u2 are considered as degree of freedom
of this bar element.

Fig. 4.15 Linear bar element

4.14 Isoparametric Elements

Fig.4.16 Natural co-ordinate system

First, the natural coordinate "ε" is attached to the element, with the origin located
at the centre of the element, as shown in Fig.4.16.
The "ε" axis need not be parallel to the x-axis - this is only for convenience.
We consider the bar element to have two degrees of freedom-axial displacements
u1 and u2 at each node associated with the global axis.
When the "ε" and "x" axes are parallel to each other, the "ε" and "x" coordinates
can be related by,
𝐿
𝑥=𝑥 + 𝜀 … (4.4)
2
Where 𝑥 → is the global coordinate of the element centroid using the global
coordinates 𝑥 𝑎𝑛𝑑 𝑥 .
(𝑥 + 𝑥 )
𝑥 = … (4.5)
2
Substitute the equation (4.5) in equation (4.4),
𝑥 +𝑥 𝐿
𝑥= + 𝜀
2 2
𝑥 +𝑥 𝐿
𝑤𝑒 𝑘𝑛𝑜𝑤 𝑡ℎ𝑎𝑡 𝑥= + 𝜀 [∵ 𝐿 = (𝑥 − 𝑥 )]
2 2
𝑥 +𝑥 𝑥 +𝑥
𝑥= + 𝜀
2 2
𝑥 +𝑥 𝑥 +𝑥
𝜀=𝑥−
2 2
 Finite Element Analysis 4.15

We can express the natural coordinate ‘’ in terms of the global coordinates as,
(𝑥 + 𝑥 ) 2
𝜀= 𝑥− × … (4.6)
2 (𝑥 − 𝑥 )
Since, the element has got two degrees of freedom, it will have two generalized
coordinates,
𝑥 =𝑎 +𝑎 𝜀 … . (4.7)
Where,  is such that -1    1. Solving for the 𝑎 ′𝑠 in terms of 𝑥 𝑎𝑛𝑑 𝑥 , we obtain
1
𝑥 = [(1 − 𝜀)𝑥 + (1 + 𝜀)𝑥 ] … . (4.8)
2
Writing the equation (5.19) in matrix form,
1 𝑥
{𝑥} = [(1 − 𝜀)(1 + 𝜀)] 𝑥
2
𝑥
{𝑥} = [𝑁 𝑁 ] 𝑥

Displacement function, 𝑥 = 𝑁 𝑢 + 𝑁 𝑢

1−𝜀
where, shape function, N =
2
1+𝜀
N₂ =
2
We may note, that N, and N₂ obey the definition of shape function i.e., the shape
function will have a value equal to unity at the node to which it belongs and zero value at
other
4.16 Isoparametric Elements

Fig. 4.17. Shape function variations with natural coordinates

(a) shape function N1, (b) shape function N₂, and (c) Linear displacement field

4.6. SHAPE FUNCTIONS FOR 4 NODED RECTANGULAR

PARENT ELEMENT BY USING NATURAL CO-ORDINATE
SYSTEM AND CO-ORDINATE TRANSFORMATION [TWO
DIMENSIONAL]

Fig. 4.18. Four noded rectangular parent element

Consider a four noded rectangular element as shown in Fig 4.18. The parent
element defined in e and co-ordinates Le, natural co-ordinates e is varying from -1 to 1
and is also varying-1 to 1.
We know that,
Shape function value is unity at its node and its value is zero at other nodes.
At node 1: (Co-ordinates ε = -1, 𝜂 = -1)
Shape function, N1 = 1 at node 1.
 Finite Element Analysis 4.17

N1 = 0 at nodes 2, 3, and 4.
N1 has to be in the form of N1 = C (1 - ε) (1 – η) …(4.20)
where, C is constant.
Substitute ε = -1 and η = -1 in equation (4.20).
⇒ N1 = C (1+1)(1+1)
=> N1 = 4C
=> 1 = 4C
1
=> C =
4
Substitute C value in equation (4.20),
1
𝑁 = (1 − ε) (1 − η) … (4.21)
4
At node 2: (Co-ordinates ε = 1, η = -1)
Shape function N₂ =1 at node 2.
N1= 0 at node 1, 3 and 4.
N2 has to be in form of, N2 = C (1+ε) (1-η) …(4.22)
Substitute ε = 1 and η = -1 in equation (4.22),
N2 = C(1+1) (1+1)
N2 = 4C
1= 4C [⸪ N = 1]
1
C =
4
Substitute C value in equation (4.22)
1
 𝑁 = (1 + )( 1 − ) … (4.23)
4
At node 3: (Co – ordinates  = 1 ,  = 1 )
Shape functions N3 = 1 at node 3.
N3 = 0 at node 1, 2 and 4.
4.18 Isoparametric Elements

N3 has to be in the from of N3 = C ( 1 +  ) ( 1 + ) …(4.24)

Substitute  = 1 and  =1 in equation (4.24)
 𝑁 = 𝐶(1 + 1)(1 + 1)
𝑁 =4𝐶
1= 4𝐶 [∵ 𝑁 = 1]
1
 𝐶=
4
Substitute C value in equation (4.24)
1
 𝑁 = (1 + )( 1 − ) … (4.25)
4
At node 4: (Co – ordinates  = -1 ,  = 1 )
Shape functions N4 = 1 at node 4.
N4 = 0 at node 1, 2 and 3
N4 has to be in the from of N4 = C ( 1 -  ) ( 1 + ) …(4.26)
Substitute  = -1 and  =1 in equation (4.26)
 𝑁 = 𝐶(1 + 1)(1 + 1)
𝑁 =4𝐶
1=4𝐶 [∵ 𝑁 = 1]
1
𝐶 =
4
Substitute C value in equation (4.26)
1
 𝑁 = (1 − )( 1 + ) … (4.27)
4
𝑢
Consider a point p with co-ordinate (, ). If displacement function 𝑢 =
𝑣
represents the displacement components of a point located at (, ) then,
𝑢 =𝑁 𝑢 +𝑁 𝑢 +𝑁 𝑢 +𝑁 𝑢
and
 Finite Element Analysis 4.19

𝑣 =𝑁 𝑣 +𝑁 𝑣 +𝑁 𝑣 +𝑁 𝑣
It can be written in matrix form as,
𝑢
⎧𝑣 ⎫
⎪𝑢 ⎪
⎪ ⎪
𝑢 𝑁 0 𝑁 0 𝑁 0 𝑁 0 𝑣
𝑢= = 𝑢 … (4.28)
𝑣 0 𝑁 0 𝑁 0 𝑁 0 𝑁 ⎨ ⎬
⎪𝑣 ⎪
⎪𝑢 ⎪
⎩𝑣 ⎭
In the isoparametric formulation i.e., for global system, the co-ordinates of the
nodal points are (𝑥 , 𝑦 ), (𝑥 , 𝑦 ), (𝑥 , 𝑦 )𝑎𝑛𝑑 (𝑥 , 𝑦 ). In order to get mapping, the co-
ordinate of point p is defined as
𝑥 =𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥
𝑦 =𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦

Fig 4.19. Four noded quadrilateral element or mapped element

The above equations can be written in matrix form as,
𝑥
⎧𝑦⎫
⎪𝑥 ⎪
⎪ ⎪
𝑥 𝑁 0 𝑁 0 𝑁 0 𝑁 0 𝑦
𝑢= 𝑦 = … (4.29)
0 𝑁 0 𝑁 0 𝑁 0 𝑁 ⎨𝑥 ⎬
⎪𝑦 ⎪
⎪𝑥 ⎪
⎩𝑦 ⎭
4.20 Isoparametric Elements

4.7 ELEMENT STIFFNESS MATRIX EQUATION FOR 4 NODED

ISOPARAMETRIC QUADRILATERAL ELEMENT
Assembling element stiffness matrix for isoparametric element is a tedious
process since it involves co-ordinate transformation from natural co-ordinate system to
global co-ordinate system.

Fig. 4.20
The displacement function u for parent rectangular element is given by,
𝑢
⎧𝑣 ⎫
⎪𝑢 ⎪
𝑢 ⎪ ⎪
𝑁 0 𝑁 0 𝑁 0 𝑁 0 𝑣
𝑢= =
0 𝑁 0 𝑁 0 𝑁 0 𝑁 ⎨𝑢 ⎬
𝑣
⎪𝑣 ⎪
⎪𝑢 ⎪
⎩𝑣 ⎭
The displacement function u for parent rectangular element is given by,
𝑥
⎧𝑦⎫
⎪𝑥 ⎪
𝑥 ⎪ ⎪
𝑁 0 𝑁 0 𝑁 0 𝑁 0 𝑦
𝑢= = 𝑥
𝑦 0 𝑁 0 𝑁 0 𝑁 0 𝑁 ⎨ ⎬
𝑦
⎪ ⎪
⎪𝑥 ⎪
⎩𝑦 ⎭
We have to express the derivatives of a function in x, y co-ordinates in terms of
its derivatives in ,  co-ordinates. This can be done as follows:
 Finite Element Analysis 4.21

𝐿𝑒𝑡 𝑓 = 𝑓(𝑥, 𝑦)
𝑓 = 𝑓[𝑥(, ), 𝑦(, )]
The relationship between natural co-ordinates and global co-ordinates can be
calculated by using chain rule of partial differentiation.
𝜕𝑓 𝜕𝑓 𝜕𝑥 𝜕𝑓 𝜕𝑦
 = × + ×
𝜕𝜀 𝜕𝑥 𝜕𝜀 𝜕𝑦 𝜕𝜀

𝜕𝑓 𝜕𝑓 𝜕𝑥 𝜕𝑓 𝜕𝑦
= × + ×
𝜕 𝜕𝑥 𝜕 𝜕𝑦 𝜕
Arranging the above equations in matrix from,
𝜕𝑓 𝜕𝑥 𝜕𝑦 𝜕𝑓
⎧ ⎫ ⎡ ⎤⎧ ⎫
𝜕𝜀 𝜕𝜀 𝜕𝜀 ⎥ 𝜕𝑥
 = ⎢
⎨𝜕𝑓 ⎬ ⎢𝜕𝑥 𝜕𝑦⎥ ⎨𝜕𝑓 ⎬
⎩𝜕⎭ ⎣𝜕 𝜕⎦ ⎩𝜕𝑦⎭
𝜕𝑓 𝜕𝑓
⎧ ⎫ ⎧ ⎫
𝜕𝜀 𝜕𝑥
 𝜕𝑓 = [ 𝐽 ] 𝜕𝑓 … (4.30)
⎨ ⎬ ⎨ ⎬
⎩𝜕⎭ ⎩𝜕𝑦⎭
Where J is the Jacobain matrix
𝜕𝑥 𝜕𝑦
⎡ ⎤
𝜕𝜀 𝜕𝜀 ⎥ 𝐽 𝐽
[𝐽]= ⎢
𝜕𝑥 𝜕𝑦  [𝐽]= … (4.31)
⎢ ⎥ 𝐽 𝐽
⎣𝜕 𝜕⎦
𝜕𝑥 𝜕𝑦
Where, 𝐽 ;𝐽 =
𝜕𝜀 𝜕𝜀
𝜕𝑥 𝜕𝑦
𝐽 ;𝐽 =
𝜕 𝜕
We know that,
𝑥 =𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥
… (4.32)
𝑦 = 𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦
𝜕𝑥 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
𝐽 = = 𝑥 + 𝑥 + 𝑥 + 𝑥 … (4.33)
𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀
4.22 Isoparametric Elements
𝜕𝑦 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
𝐽 = = 𝑦 + 𝑦 + 𝑦 + 𝑦 … (4.34)
𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀
𝜕𝑥 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
𝐽 = = 𝑥 + 𝑥 + 𝑥 + 𝑥 … (4.35)
𝜕 𝜕 𝜕 𝜕 𝜕

𝜕𝑦 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
𝐽 = = 𝑦 + 𝑦 + 𝑦 + 𝑦 … (4.36)
𝜕 𝜕 𝜕 𝜕 𝜕
We know that,
Shape functions for quadrilateral element are:
1
𝑁 = (1 − 𝜀)(1 − 𝜂)
4
1
𝑁 = (1 + 𝜀)(1 − 𝜂)
4
1
𝑁 = (1 + 𝜀)(1 + 𝜂)
4
1
𝑁 = (1 − 𝜀)(1 + 𝜂)
4

[From equations (4.21), (4.23), (4.25) and (4.27)]

𝜕𝑁 1 1
 = (−1) × (1 − 𝜂) = × −(1 − 𝜂) … (4.37)
𝜕𝜀 4 4
𝜕𝑁 1 1
 = (1) × (1 − 𝜂) = × (1 − 𝜂) … (4.38)
𝜕𝜀 4 4
𝜕𝑁 1 1
 = (1) × (1 + 𝜂) = (1 + 𝜂) … (4.39)
𝜕𝜀 4 4
𝜕𝑁 1 1
 = (−1) × (1 + 𝜂) = × −(1 + 𝜂) … (4.40)
𝜕𝜀 4 4
𝜕𝑁 1 1
 = (1 − 𝜀)(−1) = × −(1 − 𝜀) … (4.41)
𝜕 4 4
 Finite Element Analysis 4.23
𝜕𝑁 1 1
 = (1 + 𝜀)(−1) = × −(1 + 𝜀) … (4.42)
𝜕 4 4

𝜕𝑁 1 1
 = (1 + 𝜀)(1) = × (1 + 𝜀) … (4.43)
𝜕 4 4

𝜕𝑁 1 1
 = (1 − 𝜀)(1) = × (1 + 𝜀) … (4.44)
𝜕 4 4
∂N ∂N ∂N ∂N ∂N ∂N ∂N ∂N
Substitute , , , , , , , and values in equaitons
∂ε ∂ε ∂ε ∂ε ∂ ∂ ∂ ∂
Equation (4.33) becomes,
1
𝐽 = [−(1 − 𝜂)𝑥 + (1 − 𝜂)𝑥 + (1 + 𝜂)𝑥 − (1 + 𝜂)𝑥 ] … (4.45)
4
Equation (4.34) becomes,
1
𝐽 = [−(1 − 𝜂)𝑦 + (1 − 𝜂)𝑦 + (1 + 𝜂)𝑦 − (1 + 𝜂)𝑦 ] … (4.46)
4
Equation (4.35) becomes,
1
𝐽 = [−(1 − )𝑥 − (1 + )𝑥 + (1 + )𝑥 + (1 − )𝑥 ] … (4.47)
4
Equation (4.36) becomes,
1
𝐽 = [−(1 − )𝑦 − (1 + )𝑦 + (1 + )𝑦 + (1 − )𝑦 ] … (4.48)
4
From equation (4.31), we know that,
𝐽 𝐽
Jacobain matrix, [ 𝐽 ] = … (4.49)
𝐽 𝐽
Where,
1
𝐽 = [−(1 − 𝜂)𝑥 + (1 − 𝜂)𝑥 + (1 + 𝜂)𝑥 − (1 + 𝜂)𝑥 ]
4
1
𝐽 = [−(1 − 𝜂)𝑦 + (1 − 𝜂)𝑦 + (1 + 𝜂)𝑦 − (1 + 𝜂)𝑦 ]
4
1
𝐽 = [−(1 − )𝑥 − (1 + )𝑥 + (1 + )𝑥 + (1 − )𝑥 ]
4
4.24 Isoparametric Elements
1
𝐽 = [−(1 − )𝑦 − (1 + )𝑦 + (1 + )𝑦 + (1 − )𝑦 ]
4
From equation (4.30), we know that,
𝜕𝑓 𝜕𝑓
⎧ ⎫ ⎧ ⎫
𝜕𝜀 𝜕𝑥
 𝜕𝑓 = [ 𝐽 ] 𝜕𝑓
⎨ ⎬ ⎨ ⎬
⎩𝜕⎭ ⎩𝜕𝑦⎭
𝜕𝑓 𝜕𝑓
⎧ ⎫ ⎧ ⎫
𝜕𝜀 𝐽 𝐽 𝜕𝑥
 =
⎨𝜕𝑓 ⎬ 𝐽 𝐽 ⎨𝜕𝑓 ⎬
⎩𝜕⎭ ⎩𝜕𝑦⎭
𝜕𝑓 𝜕𝑓
⎧ ⎫ ⎧ ⎫
𝜕𝑥 𝐽 𝐽 𝜕
 =
⎨𝜕𝑓⎬ 𝐽 𝐽 ⎨𝜕𝑓 ⎬
⎩𝜕𝑦⎭ ⎩𝜕⎭
𝜕𝑓 𝜕𝑓
⎧ ⎫ ⎧ ⎫
𝜕𝑥 1 𝐽 −𝐽 𝜕
 = … (4.50)
⎨𝜕𝑓⎬ | 𝐽| −𝐽 𝐽 ⎨𝜕𝑓 ⎬
⎩𝜕𝑦⎭ ⎩𝜕⎭
The strain-displacement relations are,
𝜕𝑢
⎧ ⎫
𝑒 ⎪ 𝜕𝑥 ⎪
⎪ 𝜕𝑣 ⎪
𝑒= 𝑒 = … (4.51)
𝛾 ⎨ 𝜕𝑦 ⎬
⎪𝜕𝑢 𝜕𝑣
⎪ + ⎪ ⎪
⎩𝜕𝑦 𝜕𝑥⎭
Substituting f = u in equation (5.50).
𝜕𝑢 𝜕𝑢
⎧ ⎫ ⎧ ⎫
𝜕𝑥 1 𝐽 −𝐽 𝜕
 𝜕𝑢 = … (4.52)
⎨ ⎬ | 𝐽| −𝐽 𝐽 ⎨𝜕𝑢⎬
⎩𝜕𝑦⎭ ⎩𝜕⎭
 Finite Element Analysis 4.25
𝜕𝑣 𝜕𝑣
⎧ ⎫ ⎧ ⎫
𝜕𝑥 1 𝐽 −𝐽 𝜕
𝑆𝑖𝑚𝑖𝑙𝑎𝑟𝑙𝑦  𝜕𝑣 = … (4.53)
⎨ ⎬ | 𝐽| −𝐽 𝐽 ⎨𝜕𝑣 ⎬
⎩𝜕𝑦⎭ ⎩𝜕⎭
Equations (4.51), (4.52) and (4.53) yield,
𝜕𝑢
⎧ ⎫
⎪ 𝜕 ⎪
𝑒 𝐽 −𝐽 0 0 ⎪𝜕𝑢 ⎪
1 𝜕
Strain, {𝑒} = 𝑒 = 0 0 −𝐽 𝐽 … (4.54)
𝛾 | 𝐽| −𝐽 𝐽 −𝐽 ⎨𝜕𝑣 ⎬
𝐽
⎪ 𝜕 ⎪
⎪𝜕𝑣 ⎪
⎩𝜕⎭
We know that, 𝑢 =𝑁 𝑢 +𝑁 𝑢 +𝑁 𝑢 +𝑁 𝑢
𝑦 =𝑁 𝑣 +𝑁 𝑣 +𝑁 𝑣 +𝑁 𝑣
𝜕𝑢 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
 = 𝑢 + 𝑢 + 𝑢 + 𝑢 … (4.55)
𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀
𝜕𝑢 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
 = 𝑢 + 𝑢 + 𝑢 + 𝑢 … (4.56)
𝜕 𝜕 𝜕 𝜕 𝜕

𝜕𝑣 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
 = 𝑣 + 𝑣 + 𝑣 + 𝑣 … (4.57)
𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀
𝜕𝑣 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
 = 𝑣 + 𝑣 + 𝑣 + 𝑣 … (4.58)
𝜕 𝜕 𝜕 𝜕 𝜕
Assembling the equations (4.55), (4.56), (4.57) & (4.58) in matrix form,

𝜕𝑢 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
⎧ ⎫ ⎡ 0 0 0 0⎤ 𝑢
⎪ 𝜕 ⎪ ⎢ 𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀 ⎥ ⎧𝑣 ⎫
⎪𝜕𝑢⎪ ⎢𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎪𝑢 ⎪
0 0 0 0 ⎪ ⎪
𝜕 ⎢ 𝜕 𝜕 𝜕 𝜕 ⎥ 𝑣
 ⎢ … (4.59)
⎨𝜕𝑣 ⎬ 𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎨𝑢 ⎬
⎢ ⎥ 𝑣
⎪ 𝜕 ⎪ ⎢ 0 𝜕𝜀 0 𝜕𝜀 0 𝜕𝜀 0 𝜕𝜀 ⎥ ⎪ ⎪
⎪𝜕𝑣 ⎪ ⎪𝑢 ⎪
⎢ 0 𝜕𝑁 0 𝜕𝑁 0 𝜕𝑁 0 𝜕𝑁 ⎥ ⎩ 𝑣
⎩𝜕⎭ ⎭
⎣ 𝜕 𝜕 𝜕 𝜕 ⎦
4.26 Isoparametric Elements

1 𝐽 −𝐽 0 0
 𝑆𝑡𝑟𝑎𝑖𝑛, {𝑒} = 0 0 −𝐽 𝐽
| 𝐽| −𝐽 𝐽 −𝐽
𝐽
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
⎡ 0 0 0 0⎤ 𝑢
⎢ 𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀 ⎥ ⎧𝑣 ⎫
⎢𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎪𝑢 ⎪
0 0 0 0 ⎪ ⎪
⎢ 𝜕 𝜕 𝜕 𝜕 ⎥ 𝑣
× ⎢ … (4.60)
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎨𝑢 ⎬
⎢ ⎥ 𝑣
⎢ 0 𝜕𝜀 0 𝜕𝜀 0 𝜕𝜀 0 𝜕𝜀 ⎥ ⎪
⎪𝑢
⎪
⎪
⎢ 0 𝜕𝑁 0 𝜕𝑁 0 𝜕𝑁 0 𝜕𝑁 ⎥ ⎩ 𝑣 ⎭
⎣ 𝜕 𝜕 𝜕 𝜕 ⎦

We know that, 𝑆𝑡𝑟𝑎𝑖𝑛, {𝑒} = [ 𝐵] { 𝑢 }

𝑢
⎧𝑣 ⎫
⎪𝑢 ⎪
⎪ ⎪
𝑣
 {𝑒} = [ 𝐵] 𝑢 … (4.61)
⎨ ⎬
⎪𝑣 ⎪
⎪𝑢 ⎪
⎩𝑣 ⎭

Comparing equations (4.60) and (4.61),

Strain − Displacement matrix[ 𝐵]

1 𝐽 −𝐽 0 0
= 0 0 −𝐽 𝐽
| 𝐽| −𝐽 𝐽 −𝐽
𝐽
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁
⎡ 0 0 0 0⎤ 𝑢
⎢ 𝜕𝜀 𝜕𝜀 𝜕𝜀 𝜕𝜀 ⎥ ⎧𝑣 ⎫
⎢𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎪𝑢 ⎪
0 0 0 0 ⎪ ⎪
⎢ 𝜕 𝜕 𝜕 𝜕 ⎥ 𝑣
× ⎢ … (4.62)
𝜕𝑁 𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥ ⎨𝑢 ⎬
⎢ ⎥ 𝑣
⎢ 0 𝜕𝜀 0 𝜕𝜀 0 𝜕𝜀 0 𝜕𝜀 ⎥ ⎪
⎪𝑢
⎪
⎪
⎢ 0 𝜕𝑁 0 𝜕𝑁 0 𝜕𝑁 0 𝜕𝑁 ⎥ ⎩ 𝑣 ⎭
⎣ 𝜕 𝜕 𝜕 𝜕 ⎦
 Finite Element Analysis 4.27
∂N ∂N ∂N ∂N ∂N ∂N ∂N ∂N
Substitute , , , , , , , and values in equation (4.62),
∂ε ∂ε ∂ε ∂ε ∂ ∂ ∂ ∂
[Refer equations (4.37) to (4.44)]

1 𝐽 −𝐽 0 0 1
[𝐵]= 0 0 −𝐽 𝐽 ×
| 𝐽| −𝐽 𝐽 −𝐽 4
𝐽
−(1 − 𝜂) 0 (1 − 𝜂) 0 (1 + 𝜂) 0 −(1 + 𝜂) 0
⎡ ⎤
−(1 − 𝜀) 0 −(1 + 𝜀) 0 (1 + 𝜀) 0 (1 − 𝜀) 0
×⎢ ⎥ … (4.63)
⎢0 −(1 − 𝜂) 0 (1 − 𝜂) 0 (1 + 𝜂) 0 −(1 + 𝜂)⎥
⎣ 0 −(1 − 𝜀) 0 −(1 + 𝜀) 0 (1 + 𝜀) 0 (1 − 𝜀) ⎦
This is a strain-displacement relationship matrix [B] equation for isoparametric
quadrilateral element.
We know that, General element stiffness matrix equation,

[𝐾]= [ 𝐵] [𝐷][𝐵]

For Isoparametric quadrilateral element,

 [𝐾]=𝑡 [ 𝐵] [𝐷] [𝐵] 𝜕𝑥 𝜕𝑦 . . . (4.64)

[ ∵ 𝜕𝑥 𝜕𝑦 =| 𝐽| 𝜕𝜀 𝜕𝜂]
Where, t  Thickness of the element
| J |  Determinant of the J
,   Natural co-ordinates
[B]  Strain- Displacement relationship matrix
[ D]  Stress- Strain relationship matrix
For two dimensional problems,
Stress-Strain relationship matrix,
0
𝐸 1 𝑣
0
[𝐷] = 𝑣 1 1 − 𝑣 [For plane stress conditions]
1−𝑣 0
0
2
4.28 Isoparametric Elements
0
𝐸 1−𝑣 𝑣
0
[𝐷] = 𝑣 1−𝑣 1−2𝑣
(1 + 𝑣)(1 − 2𝑣) 0
0
2
[For plane stress conditions]
Where, E  Young’s modulus
v  Poisson’s ratio

4.8 ELEMENT FORCE VECTOR

The element force vector is given by,
𝐹
{𝐹} = [𝑁] … (4.65)
𝐹
Where, N is the shape function.
Fx is a load or force on x direction.
Fy is a force on y direction.

4.9 SOLVED PROBLEMS – ISOPARAMETRIC ELEMENTS

Example.4.5
Evaluate the Cartesian co-ordinate of the point P which has local co-ordinates
 = 0.6 and  = 0.8 as shown in Fig. (i).

Fig. (i).
Given: Natural co-ordinates of point P
 = 0.6
 Finite Element Analysis 4.29

 = 0.8
Cartesian co-ordinates of point 1, 2, 3 and 4.
𝑥 = 3; 𝑦 =2
𝑥 = 9; 𝑦 =4
𝑥 = 6; 𝑦 =8
𝑥 = 4; 𝑦 =5

Fig. (ii).
To find: Cartesian co-ordinates of the point P(x, y).
Solution: We know that,
Shape functions for quadrilateral element are:
1
𝑁 = (1 − 𝜀)(1 − 𝜂)
4
1
𝑁 = (1 + 𝜀)(1 − 𝜂)
4
1
𝑁 = (1 + 𝜀)(1 + 𝜂)
4
1
𝑁 = (1 − 𝜀)(1 + 𝜂)
4
Substitute  and  values in the above equations,
1
 𝑁 = (1 − 0.6)(1 − 0.8) = 0.02
4
4.30 Isoparametric Elements
1
 𝑁 = (1 + 0.6)(1 − 0.8) = 0.08
4
1
 𝑁 = (1 + 0.6)(1 + 0.8) = 0.72
4
1
 𝑁 = (1 − 0.6)(1 + 0.8) = 0.18
4
We know that,
co − ordinate, 𝑥 =𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥
= 0.02 (3) + 0.08 (9) + 0.72 (6) + 0.18(4)
x = 5.82
Similarly, Co-ordinate, 𝑦 = 𝑁 𝑦 + 𝑁 𝑦 + 𝑁 𝑦 + 𝑁 𝑦
= 0.02 × (2) + 0.08 (4) + 0.72 (8) + 0.18(5)
y = 7.02
x y
Result: The Cartesian co-ordinates of point P are (5.82, 7.02).

Example 4.6
For the isoparametric four noded quadrilateral element shown in fig.(i), determine
the Cartesian co-ordinates of point P which has local co-ordinates ε = 0.5 and  = 0.5.

Fig. (i)
Given : Natural co-ordinates of point P
 = 0.5
 = 0.5
 Finite Element Analysis 4.31

Fig (ii)
Cartesian co-ordinates of point 1, 2, 3 and 4,
𝑥 = 1; 𝑦 =1

𝑥 = 5; 𝑦 =1

𝑥 = 6; 𝑦 =6

𝑥 = 1; 𝑦 =4
To find: Cartesian co-ordinates of point P (x, y).
Solution: we know that,
Shape functions for quadrilateral element are,
1
𝑁 = (1 − 𝜀)(1 − 𝜂)
4
1
𝑁 = (1 + 𝜀)(1 − 𝜂)
4
1
𝑁 = (1 + 𝜀)(1 + 𝜂)
4
1
𝑁 = (1 − 𝜀)(1 + 𝜂)
4
Substitute  and  values in the above equations,
1
 𝑁 = (1 − 0.5)(1 − 0.5) = 0.0625
4
4.32 Isoparametric Elements
1
 𝑁 = (1 + 0.5)(1 − 0.5) = 0.1875
4
1
 𝑁 = (1 + 0.5)(1 + 0.5) = 0.5625
4
1
 𝑁 = (1 − 0.5)(1 + 0.5) = 0.1875
4
We know that,
co − ordinate, 𝑥 = 𝑁 𝑥 + 𝑁 𝑥 + 𝑁 𝑥 + 𝑁 𝑥

= 0.0625 × 1 + 0.1875 × 5 + 0.5625 × 6 + 0.1875 × 1

x = 4.5625
Similarly,
Co-ordinate, 𝑦 =𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦

= 0.0625 × 1 + 0.1875 × 1 + 0.5625 × 6 + 0.1875 × 4

y = 4.375
x y
Result: The Cartesian co-ordinates of point P are (4.5625, 4.375).

Example 4.7
For the isoparametric quadrilateral element shown in Fig.(i), determine the local
co-ordinates of the point P which has cartestian co-ordinates (7,4).

Fig. (i)
 Finite Element Analysis 4.33

Given:

Fig. (ii)
Cartesian co-ordinates of point P,
x = 7; y=4
Cartesian co-ordinates of point 1, 2, 3 and 4,
𝑥 = 1; 𝑦 =1
𝑥 = 5; 𝑦 =1
𝑥 = 6; 𝑦 =6
𝑥 = 1; 𝑦 =4
To find: Local co-ordinates of the point P, i.e.,  and .
Solution: we know that,
Shape functions for quadrilateral element are,
1
𝑁 = (1 − 𝜀)(1 − 𝜂)
4
1
𝑁 = (1 + 𝜀)(1 − 𝜂)
4
1
𝑁 = (1 + 𝜀)(1 + 𝜂)
4
1
𝑁 = (1 − 𝜀)(1 + 𝜂)
4
4.34 Isoparametric Elements

Cartesian co − ordinate of point 𝑃(𝑥, 𝑦),

𝑥 =𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 … (1)

𝑦 = 𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 … (2)

Substitute 𝑁 , 𝑁 , 𝑁 , 𝑁 , 𝑥 , 𝑥 , 𝑥 𝑎𝑛𝑑 𝑥 values in equations (1),

1
 7= [(1 − 𝜀)(1 − 𝜂) × 3 + (1 + 𝜀)(1 − 𝜂) × 6(1 + 𝜀)(1 + 𝜂) × 8 + (1
4
− 𝜀)(1 + 𝜂) × 2]

 28 = [(1 − 𝜂 − 𝜀 + 𝜀𝜂 )3 + (1 − 𝜂 + 𝜀 − 𝜀𝜂)6
+ (1 + 𝜂 + 𝜀 + 𝜀𝜂)8 (1 + 𝜂 − 𝜀 − 𝜀𝜂)2]

 28 = 3 − 3 𝜂 − 3𝜀 + 3𝜀𝜂 + 6 − 6𝜂 + 6𝜀 − 6𝜀𝜂 + 8 + 8𝜂 + 8𝜀 + 8𝜀𝜂 + 2

+ 2𝜂 − 2𝜀 − 2𝜀𝜂

28 = 19 + 𝜂 + 9𝜀 + 3𝜀𝜂

 𝜂 + 9𝜀 + 3𝜀𝜂 = 9 … . (3)

subsitute N , N , N , N , y, y , y , y and y vlaues in equaiton (2),

1
4= [(1 − 𝜀)(1 − 𝜂) × 1 + (1 + 𝜀)(1 − 𝜂) × 1 + (1 + 𝜀)(1 + 𝜂) × 6
4
+ (1 − 𝜀)(1 + 𝜂) × 5]

1
= [1 − 𝜂 − 𝜀 + 𝜀𝜂 + 1 − 𝜂 + 𝜀 − 𝜀𝜂 + 6 + 6𝜂 + 6𝜀 + 6𝜀𝜂 + 5 + 5𝜂
4
− 5𝜀 − 5𝜀𝜂]

1
4 = [ 13 + 9𝜂 + 𝜀 + 𝜀𝜂 ]
4

 16 = 13 + 9𝜂 + 𝜀 + 𝜀𝜂

 9𝜂 + 𝜀 + 𝜀𝜂 = 3 … (4)
 Finite Element Analysis 4.35

Equation (4) multiplied by (-3),

− 27 𝜂 − 3 𝜀 − 3𝜀𝜂 = −9 … (5)
Solving equation (3) and (5)
𝜂 + 9𝜀 + 3𝜀𝜂 = 9
− 27 𝜂 − 3 𝜀 − 3𝜀𝜂 = −9
− 26 𝜂 + 6 𝜀 = 0
− 26 𝜂 = − 6 𝜀
 𝜀 = 4.3333  … (6)
Substitute  value in equation (3),

(3) 𝜂 + 9 (4.3333 𝜂) + 3 (4.3333 𝜂) × 𝜂 = 9

𝜂 + 39𝜂 + 13 𝜂 = 9

 13 𝜂 + 40𝜂 = 9

 13 𝜂 + 40𝜂 − 9 = 0

−40 ± (40) − 4(13)(−9)

𝜂= 
2(13)

−𝑏 ± √𝑏 − 4 𝑎𝑐
𝑎 𝑥 + 𝑏 𝑥 + 𝑐 = 0; 𝑅𝑜𝑜𝑡𝑠 ∶
2𝑎
−40 + 45.475
=
26
𝜂 = 0.210587
Substitute  value in equation (6),
  = 4.33333 × 0.210587
 = 0.912545
Result : Local co-ordinates of the point P,
 = 0.210587
 = 0.912545
4.36 Isoparametric Elements

Example 4.8
Evaluate [J] at  =  = ½ for the linear quadrilateral element shown in fig.(i).

Fig. (i)
Given: Natural co-ordinates at point, P.
1 1
𝜀= = 0.5 ; 𝜂= = 0.5
2 2

Fig. (ii)
Cartesian co-ordinates of point 1, 2, 3 and 4,
𝑥 = 4; 𝑦 =4
𝑥 = 7; 𝑦 =5
𝑥 = 8; 𝑦 = 10
𝑥 = 3; 𝑦 =8
 Finite Element Analysis 4.37

To find: 1 Jacobian matrix [J]

Solution: Jacobian matrix for quadrillaternal element is given by,
𝐽 𝐽
[𝐽]=
𝐽 𝐽

1
𝐽 = [−(1 − 𝜂)𝑥 + (1 − 𝜂)𝑥 + (1 + 𝜂)𝑥 − (1 + 𝜂)𝑥 ] … (1)
4
1
𝐽 = [−(1 − 𝜂)𝑦 + (1 − 𝜂)𝑦 + (1 + 𝜂)𝑦 − (1 + 𝜂)𝑦 ] … (2)
4
1
𝐽 = [−(1 − 𝜀)𝑥 − (1 + 𝜀)𝑥 + (1 + 𝜀)𝑥 + (1 − 𝜀)𝑥 ] … (3)
4
1
𝐽 = [−(1 − 𝜀)𝑦 − (1 + 𝜀)𝑦 + (1 + 𝜀)𝑦 + (1 − 𝜀)𝑦 ] … (4)
4

Subsitutute η, ε, x , x , x , x , y , y , y and y values in equaiton (1), (2), (3)and (4).

1
(1) 𝐽 = [−(1 − 0.5)4 + (1 − 0.5)7 + (1 + 0.5)8 − (1 + 0.5)3]
4
𝐽 = 2.25
1
(2) 𝐽 = [−(1 − 0.5) × 4 + (1 − 0.5) × 5 + (1 + 0.5)10 − (1 + 0.5)8]
4
𝐽 = 0.875
1
(3) 𝐽 = [−(1 − 0.5)4 − (1 + 0.5)7 + (1 + 0.5)8 + (1 − 0.5)3]
4
𝐽 = 0.25
1
(4) 𝐽 = [−(1 − 0.5) × 4 − (1 + 0.5) × 5 + (1 + 0.5)10 + (1 − 0.5)8]
4
𝐽 = 2.375
𝐽 𝐽 2.25 0.875
 [ 𝐽] = =
𝐽 𝐽 0.25 2.375
2.25 0.875
Result: Jacobian matrix, [ 𝐽] =
0.25 2.375
4.38 Isoparametric Elements

Example 4.9
A four noded rectangular element is shown in Fig.(i). Determine the following
1. Jacobian Matrix 2. Strain- Dsiplacement matrix 3. Element Stresses.

Fig. (i)
𝑇𝑎𝑘𝑒 𝐸 = 2 × 10 𝑁/𝑚𝑚 𝑣 = 0.25; 𝑢
= [0, 0, 0.003, 0.004, 0.006, 0.004, 0. 0 ] 𝜀 = 0; 𝜂 = 0
Assume plane stress condition.
Given:

Fig. (ii)
Cartesian co-ordinates of point 1, 2, 3 and 4,
𝑥 = 0; 𝑦 =0
𝑥 = 2; 𝑦 =0
𝑥 = 2; 𝑦 =1
𝑥 = 0; 𝑦 =1
 Finite Element Analysis 4.39

Young s modulus, 𝐸 = 2 × 10 𝑁/𝑚

Poisson’s ratio, v = 0.25
0
⎧ 0 ⎫
⎪0.003⎪
⎪ ⎪
0.004
Displacements, 𝑢=
⎨0.006⎬
⎪0.004⎪
⎪ 0 ⎪
⎩ 0 ⎭
Natural co-ordinates,  = 0 ,  = 0
To find:
1. Jacobian matrix, J.
2. Strain –Displacement matrix, [B].
3. Element stress, .
Solution: Jacodian matrix for quadrilateral element is given by,
𝐽 𝐽
[𝐽]=
𝐽 𝐽

1
𝐽 = [−(1 − 𝜂)𝑥 + (1 − 𝜂)𝑥 + (1 + 𝜂)𝑥 − (1 + 𝜂)𝑥 ] … (1)
4
1
𝐽 = [−(1 − 𝜂)𝑦 + (1 − 𝜂)𝑦 + (1 + 𝜂)𝑦 − (1 + 𝜂)𝑦 ] … (2)
4
1
𝐽 = [−(1 − 𝜀)𝑥 − (1 + 𝜀)𝑥 + (1 + 𝜀)𝑥 + (1 − 𝜀)𝑥 ] … (3)
4
1
𝐽 = [−(1 − 𝜀)𝑦 − (1 + 𝜀)𝑦 + (1 + 𝜀)𝑦 + (1 − 𝜀)𝑦 ] … (4)
4
Subsitutute 𝑥 , 𝑥 , 𝑥 , 𝑥 , 𝑦 , 𝑦 , 𝑦 , 𝑦 , ε and η values in equaiton (1), (2), (3)and (4).
1
(1) 𝐽 = [0 + 2 + 2 − 0]
4
𝐽 =1
4.40 Isoparametric Elements
1
(2) 𝐽 = [0 + 0 + 1 − 1]
4
𝐽 =0
1
(3) 𝐽 = [0 − 2 + 2 + 0]
4
𝐽 =0
1
(4) 𝐽 = [− 0 − 0 + 1 + 1]
4
𝐽 = 0.5
𝐽 𝐽
 [𝐽]=
𝐽 𝐽

Jacobian matrix, [𝐽]= 1 0 … (5)

0 0.5
 |𝐽| = 1 × 0.5 − 0
|𝐽| = 0.5 … (6)

1 𝐽 −𝐽 0 0 1
 [𝐵]= 0 0 −𝐽 𝐽 ×
| 𝐽| −𝐽 𝐽 −𝐽 4
𝐽
−(1 − 𝜂) 0 (1 − 𝜂) 0 (1 + 𝜂) 0 −(1 + 𝜂) 0
⎡ ⎤
−(1 − 𝜀) 0 −(1 + 𝜀) 0 (1 + 𝜀) 0 (1 − 𝜀) 0
×⎢ ⎥
⎢0 −(1 − 𝜂) 0 (1 − 𝜂) 0 (1 + 𝜂) 0 −(1 + 𝜂)⎥
⎣ 0 −(1 − 𝜀) 0 −(1 + 𝜀) 0 (1 + 𝜀) 0 (1 − 𝜀) ⎦
Substitute 𝐽 , 𝐽 , 𝐽 , 𝐽 , |𝐽|, 𝜂 𝑎𝑛𝑑 𝜀 𝑣𝑎𝑙𝑢𝑒𝑠,
−1 0 1 0 1 0 −1 0
1 0.55 0 0 0 1 −1 0 −1 0 1 0 1 0
 [𝐵]= 0 0 0 1 ×
0.5 4 0 −1 0 1 0 1 0 −1
0 1 0.5 0
0 −1 0 −1 0 1 0 1

1 −0.5 0 0.5 0 0.5 0 −0.5 0

= 0 −1 0 −1 0 1 0 1
0.5 × 4
−1 −0.5 −1 0.5 1 0.5 1 −0.5

1 −1 0 1 0 1 0 −1 0
= 0 −2 0 −2 0 2 0 2
0.5 × 4
−2 −1 −2 1 2 1 2 −1
 Finite Element Analysis 4.41
−1 0 1 0 1 0 −1 0
[𝐵] = 0.25 0 −2 0 −2 0 2 0 2 … (7)
−2 −1 −2 1 2 1 2 −1
We know that,
Element stress, 𝜎 = [𝐷][𝐵]{ 𝑢} … (8)
For plane stress condition,
Stress – strain relationship matrix,
1 𝑣 0
𝐸 𝑣 1 0
[𝐷] = 1−𝑣
1−𝑣
0 0
2
1 0.25 0
2 × 10 0.25 1 0
= 1 − 0.25
1 − (0.25)
0 0
2
1 0.25 0
= 213.33 × 10 0.25 1 0
0 0 0.375
4 1 0
= 213.33 × 10 × 0.25 1 4 0
0 0 1.5
4 1 0
[ 𝐷] = 53.333 × 10 1 4 0 … (9)
0 0 1.5
Substitute [D], [B] and { u } values in equation (8),
4 1 0
 { } = 53.333 × 10 1 4 0
0 0 1.5
0
⎧ 0 ⎫
⎪0.003⎪
−1 0 1 0 1 0 −1 0 ⎪ ⎪
0.004
× 0.25 0 −2 0 −2 0 2 0 2
0.006⎬
−2 −1 −2 1 2 1 2 −1 ⎨
⎪0.004⎪
⎪ 0 ⎪
⎩ 0 ⎭
4.42 Isoparametric Elements

= 53.333 × 10
0
⎧ 0 ⎫
⎪0.003⎪
−4 −2 4 −2 4 2 −4 2 ⎪ ⎪
0.004
× 0.25 −1 −8 1 −8 1 8 −1 8
0.006⎬
−3 −1.5 −3 1.5 3 1.5 3 −1.5 ⎨
⎪0.004⎪
⎪ 0 ⎪
⎩ 0 ⎭
= 13.333
0 + 0 + (4 × 0.003) + (−2 × 0.004) + (4 × 0.006) + (2 × 0.004) + 0 + 0
× 10 0 + 0 + (1 × 0.003) + (−8 × 0.004) + (1 × 0.006) + (8 × 0.004) + 0 + 0
0 + 0 + (−3 × 0.003) + (1.5 × 0.004) + (3 × 0.006) + (1.5 × 0.004) + 0 + 0
0.036
{𝜎 } = 13.333 × 10 0.009
0.021
480
{𝜎 } = 120 𝑁/𝑚
280
Result:

1 0
1. Jacobian matrix, [ 𝐽 ] =
0 0.5
−1 0 1 0 1 0 −1 0
2. 𝑆train − Displacement matrix, [𝐵] = 0.25 0 −2 0 −2 0 2 0 2
−2 −1 −2 1 2 1 2 −1

480
3. 𝑆𝑡𝑟𝑒𝑠𝑠, {𝜎 } = 120 𝑁/𝑚
280

Example:4.10
Consider a quadrilateral element as shown in fig.(i). The local co-ordinates are
ϵ=0.5 and ή=1/2. Evaluate Jacobian Matrix and Strain – Displacement Matrix.
 Finite Element Analysis 4.43

Fig. (i)
Given: Local co-ordinates
 = 0.5
 = 0.5

Fig. (ii)
𝑎
𝑡𝑎𝑛 40 =
15
 a = 12.586 mm

Fig. (iii)
4.44 Isoparametric Elements

The cartesian co-ordinates are:

𝑥 = 0; 𝑦 =0
𝑥 = 50 𝑚𝑚; 𝑦 =0
𝑥 = 50 + 12.536 = 62.586 𝑚𝑚; 𝑦 = 15 𝑚𝑚
𝑥 = 𝑎 = 12.586 𝑚𝑚; 𝑦 = 15 𝑚𝑚
To find:
1. Jacobian matrix, [ J ].
2. Strain –Displacement matrix, [B].
Solution: Jacodian matrix for quadrilateral element is given by,
𝐽 𝐽
[𝐽]=
𝐽 𝐽
1
𝐽 = [−(1 − 𝜂)𝑥 + (1 − 𝜂)𝑥 + (1 + 𝜂)𝑥 − (1 + 𝜂)𝑥 ] … (1)
4
1
𝐽 = [−(1 − 𝜂)𝑦 + (1 − 𝜂)𝑦 + (1 + 𝜂)𝑦 − (1 + 𝜂)𝑦 ] … (2)
4
1
𝐽 = [−(1 − 𝜀)𝑥 − (1 + 𝜀)𝑥 + (1 + 𝜀)𝑥 + (1 − 𝜀)𝑥 ] … (3)
4
1
𝐽 = [−(1 − 𝜀)𝑦 − (1 + 𝜀)𝑦 + (1 + 𝜀)𝑦 + (1 − 𝜀)𝑦 ] … (4)
4
Subsitutute 𝑥 , 𝑥 , 𝑥 , 𝑥 , 𝑦 , 𝑦 , 𝑦 , 𝑦 , ε and η values in equaiton (1), (2), (3)and (4).
1
(1) 𝐽 = [−(1 − 0.5)(0) + (1 − 0.5) × 50 + (1 + 0.5) × 62.586
4
− (1 + 0.5)12.586]
𝐽 = 25
1
(2) 𝐽 = [0 + 0 + (1 + 0.5)15 − (1 + 0.5)15]
4
𝐽 =0
1
(3) 𝐽 = [−(1 − 0.5) × 0 − (1 + 0.5) × 50 + (1 + 0.5) × 62.586
4
+ (1 − 0.5) × 12.586 ]
 Finite Element Analysis 4.45

𝐽 = 6.293
1
(4) 𝐽 = [− 0 − 0 + (1 + 0.5)15 + (1 − 0.5)15]
4
𝐽 = 7.5
𝐽 𝐽
 [𝐽]=
𝐽 𝐽
25 0
Jacobian matrix, [ 𝐽 ] = … (5)
6.293 75
 |𝐽| = (25 × 7.5) − (0 × 6.293)
|𝐽| = 187.5
We know that,
Strain –Displacement matrix for quadrilateral element is,

1 𝐽 −𝐽 0 0 1
 [𝐵]= 0 0 −𝐽 𝐽 ×
| 𝐽| −𝐽 𝐽 −𝐽 4
𝐽
−(1 − 𝜂) 0 (1 − 𝜂) 0 (1 + 𝜂) 0 −(1 + 𝜂) 0
−(1 − 𝜀) 0 −(1 + 𝜀) 0 (1 + 𝜀) 0 (1 − 𝜀) 0
×
0 −(1 − 𝜂) 0 (1 − 𝜂) 0 (1 + 𝜂) 0 −(1 + 𝜂)
0 −(1 − 𝜀) 0 −(1 + 𝜀) 0 (1 + 𝜀) 0 (1 − 𝜀)
Substitute 𝐽 , 𝐽 , 𝐽 , 𝐽 , |𝐽|, 𝜂 𝑎𝑛𝑑 𝜀 𝑣𝑎𝑙𝑢𝑒𝑠,
1 7.5 0 0 0
 [𝐵]= 0 0 −6.293 25
187.5
−6.293 25 7.5 0
−(1 − 0.5) 0 (1 − 0.5) 0 (1 + 0.5) 0 −(1 + 0.5) 0
1 −(1 − 0.5) 0 −(1 + 0.5) 0 (1 + 0.5) 0 (1 − 0.5) 0
×
4 0 −(1 − 0.5) 0 (1 − 0.5) 0 (1 + 0.5) 0 −(1 + 0.5)
0 −(1 − 0.5) 0 −(1 + 0.5) 0 (1 + 0.5) 0 (1 − 0.5)

1 7.5 0 0 0
= 0 0 −6.293 25
187.5 × 4
−6.293 25 7.5 0
−0.5 0 0.5 0 1.5 0 −1.5 0
× −0.5 0 −1.5 0 1.5 0 0.5 0
0 −0.5 0 0.5 0 1.5 0 −1.5
0 −0.5 0 −1.5 0 1.5 0 0.5
4.46 Isoparametric Elements
1
=
750
𝟕. 𝟓 × −(𝟎. 𝟓) 𝟎 𝟕. 𝟓 × 𝟎. 𝟓 𝟎 𝟕. 𝟓 × 𝟏. 𝟓 𝟎 𝟕. 𝟓 × −(𝟏. 𝟓) 𝟎
⎡ ⎤
−𝟔. 𝟐𝟗𝟑 × (−𝟎. 𝟓) −𝟔. 𝟐𝟗𝟑 × 𝟎. 𝟓 −𝟔. 𝟐𝟗𝟑 × 𝟏. 𝟓 −𝟔. 𝟐𝟗𝟑 × −𝟏. 𝟓
⎢ ⎥
⎢ 𝟎 + 𝟎 + 𝟎 + 𝟎 + ⎥
× ⎢ 𝟐𝟓 × (−𝟎. 𝟓) 𝟐𝟓 × −𝟏. 𝟓 𝟐𝟓 × 𝟏. 𝟓 𝟐𝟓 × 𝟏. 𝟓 ⎥
⎢ −𝟔. 𝟐𝟗𝟑 × (−𝟎. 𝟓) −𝟔. 𝟐𝟗𝟑 × 𝟎. 𝟓 −𝟔. 𝟐𝟗𝟑 × 𝟏. 𝟓 −𝟔. 𝟐𝟗𝟑 × −𝟏. 𝟓 ⎥
⎢ + 𝟕. 𝟓 × −𝟎. 𝟓 + 𝟕. 𝟓 × 𝟎. 𝟓 + 𝟕. 𝟓 × 𝟏. 𝟓 + 𝟕. 𝟓 × −𝟏. 𝟓 ⎥
⎣ 𝟐𝟓 × (−𝟎. 𝟓) 𝟐𝟓 × −𝟏. 𝟓 𝟐𝟓 × 𝟏. 𝟓 𝟐𝟓 × 𝟏. 𝟓 ⎦

1
=
750
−3.75 0 3.75 0 11.25 0 −11.25 0
× 0 −9.3535 0 −40.6465 0 28.0605 0 21.939
−9.3535 −3.75 −40.6465 3.75 28.0605 7.875 21.939 −7.875
1
[𝐵] =
750
−3.75 0 3.75 0 11.25 0 −11.25 0
× 0 −9.3535 0 −40.6465 0 28.0605 0 21.939
−9.3535 −3.75 −40.6465 3.75 19.6423 7.875 21.939 −7.875
Result:
25 0
1. Jacobian matrix, [ J ] =
6.293 7.5
2. Strain − Displacement matrix,
1
[B] =
750
−3.75 0 3.75 0 11.25 0 −11.25 0
× 0 −9.3535 0 −40.6465 0 28.0605 0 21.939
−9.3535 −3.75 −40.6465 3.75 19.6423 7.875 21.939 −7.875

Example 4.11
Establish the strain-displacement matrix for the linear quadrilateral element
as shown in Fig.(i) at Gauss point r = 0.57735 and s = 0.57735.

Fig. (i)
 Finite Element Analysis 4.47

Given:

Fig. (ii)
Cartesian co-ordinates of point 1, 2, 3 and 4 are:
𝑥 = 1; 𝑦 =1

𝑥 = 5; 𝑦 =2

𝑥 = 4; 𝑦 =5

𝑥 = 2; 𝑦 =4

Natural co-ordinates,  = 0.57735

 = -0.57735
To find: Strain –Displacement matrix, [B]
Solution: Jacodian matrix for quadrilateral element is given by,
𝐽 𝐽
[𝐽]=
𝐽 𝐽

1
𝐽 = [−(1 − 𝜂)𝑥 + (1 − 𝜂)𝑥 + (1 + 𝜂)𝑥 − (1 + 𝜂)𝑥 ] … (1)
4
1
𝐽 = [−(1 − 𝜂)𝑦 + (1 − 𝜂)𝑦 + (1 + 𝜂)𝑦 − (1 + 𝜂)𝑦 ] … (2)
4
4.48 Isoparametric Elements
1
𝐽 = [−(1 − 𝜀)𝑥 − (1 + 𝜀)𝑥 + (1 + 𝜀)𝑥 + (1 − 𝜀)𝑥 ] … (3)
4
1
𝐽 = [−(1 − 𝜀)𝑦 − (1 + 𝜀)𝑦 + (1 + 𝜀)𝑦 + (1 − 𝜀)𝑦 ] … (4)
4

Subsitutute 𝑥 , 𝑥 , 𝑥 , 𝑥 , 𝑦 , 𝑦 , 𝑦 , 𝑦 , ε and η values in equaiton (1), (2), (3)and (4).

1
(1) 𝐽 = [−(1 + 0.57735)(1) + (1 + 0.57735)(5) + (1 − 0.57735)(4)
4
− (1 − 0.57735)(2)]

𝐽 = 1.78867

1
(2) 𝐽 = [−(1 + 0.57735)(1) + (1 + 0.57735)(2) + (1 − 0.57735)(5)
4
− (1 − 0.57735)(4)]

𝐽 = 0.5

1
(3) 𝐽 = [−(1 − 0.57735)(1) − (1 + 0.57735)(5) + (1 + 0.57735)(4)
4
+ (1 − 0.57735)(2) ]

𝐽 = −0.2886

1
(4) 𝐽 = [−(1 − 0.57735)(1) − (1 + 0.57735)(2) + (1 + 0.57735)(5)
4
+ (1 − 0.57735)(4)]

𝐽 = 1.5

𝐽 𝐽
 [𝐽]=
𝐽 𝐽

|𝐽| = 1.7886 0.5

−0.2886 1.5

|𝐽| = 2.827
 Finite Element Analysis 4.49

We know that,
Strain –Displacement matrix for quadrilateral element is,

1 𝐽 −𝐽 0 0 1
 [𝐵]= 0 0 −𝐽 𝐽 ×
| 𝐽| −𝐽 𝐽 −𝐽 4
𝐽

−(1 − 𝜂) 0 (1 − 𝜂) 0 (1 + 𝜂) 0 −(1 + 𝜂) 0
−(1 − 𝜀) 0 −(1 + 𝜀) 0 (1 + 𝜀) 0 (1 − 𝜀) 0
×
0 −(1 − 𝜂) 0 (1 − 𝜂) 0 (1 + 𝜂) 0 −(1 + 𝜂)
0 −(1 − 𝜀) 0 −(1 + 𝜀) 0 (1 + 𝜀) 0 (1 − 𝜀)

Substitute 𝐽 , 𝐽 , 𝐽 , 𝐽 , |𝐽|, 𝜂 𝑎𝑛𝑑 𝜀 𝑣𝑎𝑙𝑢𝑒𝑠,

1 1.5 −0.5 0 0
 [𝐵]= 0 0 0.2886 1.7886
2.827
0.2886 1.7886 1.5 −0.5
−1.57735 0 1.57735 0 0.42265 0 −0.42265 0
1 −0.42265 0 −1.57735 0 1.57735 0 0.42265 0
×
4 0 −1.57735 0 1.57735 0 0 −0.42265
1.57735
0 −0.42265 0 −1.57735 0 0 0.42265

1 −2.1547 0 3.1547 0 −0.1547 0 −0.8453 0

= 0 −1.2111 0 −2.3660 0 2.9432 0 0.63397
2.827 × 4
−1.2111 −2.1547 −2.3660 3.1457 2.9432 −0.1547 0.63397 −0.8453

−2.1547 0 3.1547 0 −0.1547 0 −0.8453 0

[𝐵] = 0.0884 0 −1.2111 0 −2.3660 0 2.9432 0 0.63397
−1.2111 −2.1547 −2.3660 3.1457 2.9432 −0.1547 0.63397 −0.8453

Result:
1. 𝑆𝑡𝑟𝑎𝑖𝑛 − 𝐷𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑚𝑎𝑡𝑟𝑖𝑥, [𝐵]
−2.1547 0 3.1547 0 −0.1547 0 −0.8453 0
= 0.0884 0 −1.2111 0 −2.3660 0 2.9432 0 0.63397
−1.2111 −2.1547 −2.3660 3.1457 2.9432 −0.1547 0.63397 −0.8453

Example 4.12
For the isoparametric quadrilateral element shown in Fig. (i), the Cartesian co-
ordinates of point P are (6,4). The loads 10 kN and 12 kN are acting in x and y directions
on that point P. Evaluate the nodal equivalent forces.
4.50 Isoparametric Elements

Fig. (i)
Given: Cartesian co-ordinates of point P,
𝑥 = 6; 𝑦=4

Fig. (ii)
The Cartesian co-ordinates of point 1, 2, 3 and 4 are:
𝑥 = 2; 𝑦 =1
𝑥 = 8; 𝑦 =4
𝑥 = 6; 𝑦 =6
𝑥 = 3; 𝑦 =5
Loads, 𝐹 = 10 𝑘𝑁, 𝐹 = 12 𝑘𝑁
To find: Nodal equivalent forces for x and y directions,
𝑖. 𝑒. , 𝐹 , 𝐹 , 𝐹 , 𝐹 , 𝐹 , 𝐹 , 𝐹 , 𝐹 ,
 Finite Element Analysis 4.51

Solution: we know that,

Shape functions for quadrilateral element are,
1
𝑁 = (1 − 𝜀)(1 − 𝜂) … (1)
4
1
𝑁 = (1 + 𝜀)(1 − 𝜂) … (2)
4
1
𝑁 = (1 + 𝜀)(1 + 𝜂) … (3)
4
1
𝑁 = (1 − 𝜀)(1 + 𝜂) … (4)
4
Cartesian co-ordinates of point, P(x, y).
𝑥 =𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 … (5)

𝑦 =𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 … (6)

Substitute 𝑥, 𝑥 , 𝑥 , 𝑥 , 𝑥 , 𝑁 , 𝑁 , 𝑁 𝑎𝑛𝑑 𝑁 values in equation (5),

1
6 = [(1 − 𝜀)(1 − 𝜂)2 + (1 + 𝜀)(1 − 𝜂)8 + (1 + 𝜀)(1 + 𝜂)6
4
+ (1 − 𝜀)(1 + 𝜂)3]

24 = [(1 − 𝜂 − 𝜀 + 𝜀𝜂)2 + (1 − 𝜂 + 𝜀 − 𝜀𝜂)8 + (1 + 𝜂 + 𝜀 + 𝜀𝜂)6

+ (1 + 𝜂 − 𝜀 − 𝜀𝜂)3]

24 = [2 − 2𝜂 − 2𝜀 + 2𝜀𝜂 + 8 − 8𝜂 + 8𝜀 − 8𝜀𝜂 + 6 + 6𝜂 + 6𝜀 + 6𝜀𝜂 + 3

+ 3𝜂 + 3𝜀 − 3𝜀𝜂]

24 = 19 − 𝜂 + 9𝜀 − 3𝜀𝜂

5 = −𝜂 + 9𝜀 − 3𝜀𝜂

 9𝜀 − 𝜂 − 3𝜀𝜂 = 5 … (7)
Substitute 𝑦, 𝑦 , 𝑦 , 𝑦 , 𝑦 , 𝑁 , 𝑁 , 𝑁 𝑎𝑛𝑑 𝑁 values in equation (6),
4.52 Isoparametric Elements
1
4 = [(1 − 𝜀)(1 − 𝜂)1 + (1 + 𝜀)(1 − 𝜂)4 + (1 + 𝜀)(1 + 𝜂)6
4
+ (1 − 𝜀)(1 + 𝜂)5]

16 = [1 − 𝜂 − 𝜀 + 𝜀𝜂 + (1 − 𝜂 + 𝜀 − 𝜀𝜂) × 4 + (1 + 𝜂 + 𝜀 + 𝜀𝜂) × 6
+ (1 + 𝜂 − 𝜀 − 𝜀𝜂) × 5]

= [1 − 𝜂 − 𝜀 + 𝜀𝜂 + 4 − 4𝜂 + 4𝜀 − 4𝜀𝜂 + 6 + 6𝜂 + 6𝜀 + 6𝜀𝜂 + 5
+ 5𝜂 − 5𝜀 − 5𝜀𝜂]

16 = [16 + 6 𝜂 + 4𝜀 − 2𝜀𝜂]

 4𝜀 + 6𝜂 − 2𝜀𝜂 = 0 … (8)

Equation (7) multiplied by 2 and equation (8) multiplied by (-3)

18𝜀 − 2 𝜂 − 6𝜀𝜂 = 10 … (9)

−12𝜀 − 18𝜂 + 6𝜀𝜂 = 0 … (10)

Solving, 6𝜀 − 20𝜂 = 0

 − 20𝜂 = 10 − 6𝜀

 20𝜂 = 6𝜀 − 10

6𝜀 − 10
 𝜂=
20

 𝜂 = 0.3 − 0.5 … (11)

Substituting  values in equation (7),

 9𝜀 − (0.3𝜀 − 0.5) − 3𝜀(0.3𝜀 − 0.5) = 5

 9𝜀 − 0.3𝜀 − 0.5 − 0.9 + 1.5 = 5

 10.2𝜀 − 0.9 − 4.5 = 0

 0.9 − 10.2𝜀 + 4.5 = 0

 Finite Element Analysis 4.53

10.2 ± (−10.2) − 4(0.9)(4.5)

 𝜀=
2(0.9)

10.2 − 9.372
=
1.8

𝜀 = 0.46
Substitute  value in equation (11),
  = 0.3(0.46)-0.5

 = -0.362

Substitute  and  values in equation (1), (2), (3) and (4).

1
(1)  𝑁 = (1 − 0.46)(1 + 0.362)
4

𝑁 = 0.18387

1
(2)  𝑁 = (1 + 0.46)(1 + 0.362)
4

𝑁 = 0.49713

1
(3)  𝑁 = (1 + 0.46)(1 − 0.362)
4

𝑁 = 0.23287

1
(4)  𝑁 = (1 − 0.46)(1 − 0.362)
4

𝑁 = 0.08613

We know that,

𝐹
Element force vector, {𝐹} = [𝑁] 𝐹 … (12)
4.54 Isoparametric Elements
𝐹 𝑁
𝐹 𝑁
 = {𝐹 }
𝐹 𝑁
𝐹 𝑁
𝐹 0.18384
𝐹 0.49713
 = {10}
𝐹 0.23287
𝐹 0.08613
𝐹 0.18384
𝐹 0.49713
 = 𝑘𝑁
𝐹 0.23287
𝐹 0.08613
𝐹 𝑁
⎧ ⎫
𝐹 𝑁
Similarly, = {𝐹 }
⎨𝐹 ⎬ 𝑁
⎩𝐹 ⎭ 𝑁

0.18387
0.49713
= {12}
0.23287
0.08613
𝐹 2.20644
⎧ ⎫
𝐹 5.96556
 = 𝑘𝑁
⎨𝐹 ⎬ 2.79444
⎩𝐹 ⎭ 1.03356

Result: Nodal forces for x directions,

𝐹 1.8387
𝐹 4.9713
 = 𝑘𝑁
𝐹 2.3287
𝐹 0.8613
Nodal forces for y directions,
𝐹 2.20644
⎧ ⎫
𝐹 5.96556
 = 𝑘𝑁
𝐹
⎨ ⎬ 2.79444
⎩𝐹 ⎭ 1.03356
 Finite Element Analysis 4.55

Example 4.13
Evaluate the Jacobian matrix for the isoparametric quadrilateral element shown in
Fig. (i).

Fig. (i).
Given: Cartesian co-ordinates:
𝑥 = 1; 𝑦 =0
𝑥 = 2; 𝑦 =0
𝑥 = 2.5; 𝑦 = 1.5
𝑥 = 1.5; 𝑦 =1

Fig. (ii).
To find: Jacobian matrix, [ J ]
Solution: Shape functions for isoparametric quadrilateral element are given by,
1
𝑁 = (1 − 𝜀)(1 − 𝜂) … (1)
4
1
𝑁 = (1 + 𝜀)(1 − 𝜂) … (2)
4
4.56 Isoparametric Elements
1
𝑁 = (1 + 𝜀)(1 + 𝜂) … (3)
4
1
𝑁 = (1 − 𝜀)(1 + 𝜂) … (4)
4
We know that, for quadrilateral element, Cartesian co-ordinates are:

𝑥 =𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 … (5)

𝑦 =𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 … (6)

Substitute 𝑁 , 𝑁 , 𝑁 , 𝑁 , 𝑥 , 𝑥 , 𝑥 , 𝑥 , 𝑦 , 𝑦 , 𝑦 𝑎𝑛𝑑 𝑦 values in equation (5) and (6).

1
𝑥 = [(1 − 𝜀)(1 − 𝜂) × 1 + (1 + 𝜀)(1 − 𝜂) × 2 + (1 + 𝜀)(1 + 𝜂) × 2.5
4
+ (1 − 𝜀)(1 + 𝜂) × 1.5] . . . (7)
1
𝑦= [(1 − 𝜀)(1 − 𝜂) × 0 + (1 + 𝜀)(1 − 𝜂) × 0 + (1 + 𝜀)(1 + 𝜂) × 1.5
4
+ (1 − 𝜀)(1 + 𝜂) × 1] . . . (8)
Simplifying equation (7),
1
𝑥= [1 + 𝜂 − 𝜀 + 𝜀𝜂 + 2 − 2𝜂 + 2𝜀 − 2𝜀𝜂 + 2.5 + 2.5 𝜂 + 2.5𝜀
4
+ 2 .5 𝜀𝜂 + 1.5 + 1.5𝜂 − 1.5 𝜀 − 1.5 𝜀𝜂]

1
= [7 + 𝜂 + 2𝜀 + 0 𝜀𝜂]
4
1
 𝑥= [7 + 𝜂 + 2𝜀] … (9)
4
𝜕𝑥 1
 = [0 + 0 + 2]
𝜕𝜀 4
𝜕𝑥 1
= … (10)
𝜕𝜀 2
𝜕𝑥 1
 = [0 + 1 + 0]
𝜕𝜂 4
 Finite Element Analysis 4.57
𝜕𝑥 1
= … (11)
𝜕𝜂 4

Substitute equation (8),

1
𝑦 = [0 + 0 + 1.5 + 1.5 𝜂 + 1.5 𝜀 + 1.5 𝜀𝜂 + 1 + 𝜂 − 𝜀 − 𝜀𝜂]
4
1
𝑦 = [2.5 + 2.5 𝜂 + 0.5 𝜀 + 0.5 𝜀𝜂] … (12)
4
𝜕𝑦 1 0.5
 = [0 + 0 + 0.5 + 0.5 𝜂] = [ 1 + 𝜂]
𝜕𝜀 4 4
𝜕𝑦
= 0.125[1 + 𝜂] … (13)
𝜕𝜀
𝜕𝑦 1 0.5
 = [0 + 2.5 + 0.5 𝜀] = [ 5 + 𝜀]
𝜕𝜂 4 4

𝜕𝑦
= 0.125[5 + 𝜀] … (14)
𝜕𝜂

𝐽 𝐽
We know that, Jacobian matrix, [ 𝐽] =
𝐽 𝐽

𝜕𝑥 𝜕𝑦
Where, 𝐽 = ; 𝐽 =
𝜕𝜀 𝜕𝜀
𝜕𝑥 𝜕𝑦
𝐽 = ; 𝐽 =
𝜕𝜂 𝜕𝜂

1
0.125(1 + 𝜂)
 [ 𝐽] = 2
1
0.125(1 + 𝜀)
4
1
0.125(1 + 𝜂)
𝐑𝐞𝐬𝐮𝐥𝐭: Jacobian matrix, [ 𝐽] = 2
1
0.125(1 + 𝜀)
4
4.58 Isoparametric Elements

Example 4.14
For the element shown in Fig. (i), determine the Jacobian matrix

Fig. (i)
Given: Cartesian co-ordinates:
𝑥 = 0; 𝑦 =0
𝑥 = 5; 𝑦 = −2
𝑥 = 6; 𝑦 =5
𝑥 = 3; 𝑦 =4

Fig. (ii).
To find: Jacobian matrix, [ J ]
Solution: Shape functions for isoparametric quadrilateral element are given by,
 Finite Element Analysis 4.59
1
𝑁 = (1 − 𝜀)(1 − 𝜂) … (1)
4
1
𝑁 = (1 + 𝜀)(1 − 𝜂) … (2)
4
1
𝑁 = (1 + 𝜀)(1 + 𝜂) … (3)
4
1
𝑁 = (1 − 𝜀)(1 + 𝜂) … (4)
4

We know that, for quadrilateral element, Cartesian co-ordinates are:

𝑥 =𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 +𝑁 𝑥 … (5)

𝑦 =𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 +𝑁 𝑦 … (6)

Substitute 𝑁 , 𝑁 , 𝑁 , 𝑁 , 𝑥 , 𝑥 , 𝑥 , 𝑥 , 𝑦 , 𝑦 , 𝑦 𝑎𝑛𝑑 𝑦 values in equation (5) and (6).

1
𝑥 = [(1 − 𝜀)(1 − 𝜂) × 0 + (1 + 𝜀)(1 − 𝜂) × 5 + (1 + 𝜀)(1 + 𝜂) × 6
4
+ (1 − 𝜀)(1 + 𝜂) × 3] . . . (7)

1
𝑦= [(1 − 𝜀)(1 − 𝜂) × 0 + (1 + 𝜀)(1 − 𝜂) × (−2)
4
+ (1 + 𝜀)(1 + 𝜂) × 5 + (1 − 𝜀)(1 + 𝜂) × 4] . . . (8)

Simplifying equation (7),

1
𝑥= [5 − 5𝜂 + 5𝜀 − 5𝜀𝜂 + 6 + 6𝜂 + 6𝜀 + 6𝜀𝜂 + 3 + 3𝜂 − 3𝜀 − 3𝜀𝜂]
4
1
𝑥= [14 + 4𝜂 + 8𝜀 − 2 𝜀𝜂] … (9)
4
𝜕𝑥 1 1 4
 = [0 + 0 + 8 − 2𝜂] = [ 8 − 2𝜂] = [2 − 0.5𝜂]
𝜕𝜀 4 4 4
𝜕𝑥
= 2 − 0.5𝜂 … (10)
𝜕𝜀
4.60 Isoparametric Elements
𝜕𝑥 1 4
 = [0 + 4 + 0 − 2𝜀] = [1 − 0.5 𝜀]
𝜕𝜂 4 4
𝜕𝑥
= 1 − 0.5 𝜀 … (11)
𝜕𝜂

Substitute equation (8),

1
𝑦 = [+0 − 2 + 2𝜂 − 2𝜀 + 2𝜀𝜂 + 5 + 5𝜂 + 5𝜀 + 5𝜀𝜂 + 4 + 4𝜂 − 4𝜀
4
− 4𝜀𝜂]

1
𝑦 = [7 + 11 𝜂 − 𝜀 + 3 𝜀𝜂] … (12)
4
𝜕𝑦 1
 = [0 + 0 − 1 + 3 𝜂]
𝜕𝜀 4
𝜕𝑦
= 0.75𝜂 − 0.25 … (13)
𝜕𝜀
𝜕𝑦 1
 = [0 + 11 − 0 + 3 𝜀]
𝜕𝜂 4

𝜕𝑦
= 2.75 + 0.75 𝜀 … (14)
𝜕𝜂
We know that,
𝐽 𝐽
Jacobian matrix, [ 𝐽] =
𝐽 𝐽
𝜕𝑥 𝜕𝑦
Where, 𝐽 = ; 𝐽 =
𝜕𝜀 𝜕𝜀
𝜕𝑥 𝜕𝑦
𝐽 = ; 𝐽 =
𝜕𝜂 𝜕𝜂

2 − 0.5𝜂 0.75𝜂 − 0.25

 [ 𝐽] =
1 − 0.5𝜀 2.75 + 0.75𝜀
2 − 0.5𝜂 0.75𝜂 − 0.25
Result: Jacobian matrix, [ 𝐽] =
1 − 0.5𝜀 2.75 + 0.75𝜀
 Finite Element Analysis 4.61

4.11. NUMERICAL INTEGRATION [GAUSSIAN QUADRATURE]

AND APPLICATION TO PLANE STRESS PROBLEMS
The Gauss quadrature is one of the numerical integration methods to calculate the
definite integrals. In finite element analysis, Gauss quadrature method is mostly preferred.
In this method, the numerical integration is achieved by the following expression,

𝑓(𝑥) 𝑑𝑥 = 𝑤 𝑓(𝑥)

where, 𝑤 is weight function.

𝑓(𝑥 ) is values of the function at pre-determined sampling points.
Function 𝑓(𝑥 ) is calculated at several sampling points i.e. n = 1, 2, 3,… and each
value of 𝑓(𝑥 ) is multiplied by weight function 𝑤 Finally, all the terms are added, it gives
the value of integration.
Table 4.1 shows the location of Gauss sampling points 𝑓(𝑥 ) and corresponding
weight function w for different number of points (n).
Table 4.1 gives Gauss points for integration from -1 to +1,

𝑓(𝑥) 𝑑𝑥 = 𝑤 𝑓(𝑥 )

Table 4.1.
Number of Corresponding weights,
Location, . 𝒙𝒊
points, n 𝒘𝒊

1. 𝑥 = 0.000 …. 2.000

1
2. 𝑥 ,𝑥 = ± = ±0.577350269189 1.000
3

5
3 = 0.555555 …
3. 𝑥 ,𝑥 = ± = ±0.774596669241 9
5
4.62 Isoparametric Elements
8
= 0.888888 …
9
𝑥 = 0.000
𝑥 , 𝑥 = ±0.8611363116 0.3478548451
4.
𝑥 , 𝑥 = ±0.3399810436 0.6521451549

4.12. SOLVED PROBLEMS-GAUSSIAN QUADRATURE

Example 4.16

Evaluate ∫ (x + x ) dx by applying 3 point gaussian quadrature

Given

Integral I= (𝑥 + 𝑥 ) 𝑑𝑥

𝑓(𝑥) = 𝑥 + 𝑥
To find: Evaluate the integral by using Gaussian quadrature with three Gauss points.
Solution: We know that, for three point Gaussian quadrature

3
𝑥 = = 0.774596669
5

𝑥 =0

3
𝑥 =− = −0.774596669
5

5
𝑤 = = 0.555555
9
8
𝑤 = = 0.888888
9
5
𝑤 = = 0.555555 [Refer Table 4.1]
9
 Finite Element Analysis 4.63

We know that, 𝑓(𝑥) = 𝑥 + 𝑥

𝑓(𝑥 ) = (𝑥 ) + (𝑥 )

= (0.774596669) + (0.774596669)

𝑓(𝑥 ) = 0.96

𝑤 𝑓(𝑥 ) = 0.555555 × 0.96

𝑤 𝑓(𝑥 ) = 0.5333 … (1)

𝑓(𝑥 ) = (𝑥 ) + (𝑥 ) = (0) + (0)2

𝑓(𝑥 ) = 0

𝑤 𝑓(𝑥 ) = 0.888888 × 0

𝑤 𝑓(𝑥 ) = 0 … (2)

𝑓(𝑥 ) = (𝑥 ) + (𝑥 )

= (−0.774596669) + (−0.774596669)

𝑓(𝑥 ) = 0.96

𝑤 𝑓(𝑥 ) = 0.555555 × 0.96

𝑤 𝑓(𝑥 ) = 0.5333 … (3)

Adding equation (1),(2) and (3),

𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 0.5333 + 0 + 0.5333 = 1.0666
Result:

(𝑥 + 𝑥 ) 𝑑𝑥 = 1.0666

Verification:

𝑥 𝑥
(𝑥 + 𝑥 ) 𝑑𝑥 = +
5 3
4.64 Isoparametric Elements
1 1
= [(1 ) − (−1 )] + [(1 ) − (−1 )]
5 3
1 1
= [1 + 1] + [1 + 1] = 1.0666
5 3

Example 4.17
𝟏

𝐄𝐯𝐚𝐥𝐮𝐚𝐭𝐞 𝐭𝐡𝐞 𝐢𝐧𝐭𝐞𝐠𝐫𝐚𝐥 (𝐱 𝟒 − 𝟑𝐱 + 𝟕) 𝐝𝐱

𝟏

Given:

Integral I = (𝑥 − 3𝑥 + 7) 𝑑𝑥

𝑓(𝑥) = (𝑥 − 3𝑥 + 7)
To find: Evaluate the integral by using Gaussian quadrature.
Solution: We know that, the given integrand is a polynomial of order 2. So for exact
integration,
2n -1 = 4
2n = 5
5
n = = 2.5
2
The calculated number of sampling points should be rounded upto the nearest
integer value. So , n =2.5≈3, i.e., in this problem, we should use three sampling points.
For three point Gaussian quadrature,

3
𝑥 = = 0.774596669
5

𝑥 =0

3
𝑥 =− = −0.774596669
5
 Finite Element Analysis 4.65
5
𝑤 = = 0.555555
9
8
𝑤 = = 0.888888
9
5
𝑤 = = 0.555555 [Refer Table 4.1]
9

We know that,
𝑓(𝑥) = (𝑥 − 3𝑥 + 7)

𝑓(𝑥 ) = (𝑥 − 3𝑥 + 7)

= (0.774596669) − 3(0.774596669) + 74

𝑓(𝑥 ) = 5.036209992

𝑤 𝑓(𝑥 ) = 0.555555 × 5.036209992

𝑤 × 𝑓(𝑥 ) = 2.797891

𝑓(𝑥 ) = (𝑥 − 3𝑥 + 7) = (0)4 − 3(0) + 7

𝑓(𝑥 ) = 7

𝑤 𝑓(𝑥 ) = 0.888888 × 7

𝑤 𝑓(𝑥 ) = 6.222216

𝑓(𝑥 ) = (𝑥 − 3𝑥 + 7)

= (−0.774596669) ˗ 3(−0.774596669) + 7

𝑓(𝑥 ) = 9.683790008

𝑤 𝑓(𝑥 ) = 0.555555 × 9.683790008

𝑤 𝑓(𝑥 ) = 5.379877
4.66 Isoparametric Elements

Adding equation (1),(2) and (3),

𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 )

= 2.797891 + 6.222216 + 5.379877

= 14.399984

Result:

(𝑥 − 3𝑥 + 7) 𝑑𝑥 = 14.399984

Verification:

𝑥 𝑥
(𝑥 − 3𝑥 + 7) 𝑑𝑥 = −3 + 7[𝑥]
5 2

1 3
= [(1 ) − (−1 )] − [(1 ) − (−1 )] + 7 [ 1 − (1)]
5 2
1
= [2] – 0 + 7(2) = 14.4
5

Example 4.18
𝟏
Evaluate the integral by gaussian quadrature ∫ 𝟏 𝐱 𝟐 𝐝𝐱

Given:
𝟏
Integral, I = 𝒙𝟐 𝒅𝒙
𝟏

𝑓(𝑥) = 𝑥
To find: Evaluate the integral by using Gaussian quadrature.
Solution: We know that, the given integrand is a polynomial of order 2. So for exact
integration,
2n -1 = 2
2n = 3
 Finite Element Analysis 4.67
3
n = = 1.5
2
The calculated number of sampling points should be rounded upto the nearest
integer value. So , n =1.5≈2, i.e., in this problem, we should use two sampling points.
For three point Gaussian quadrature,

1
𝑥 =+ = 0.577350269
3

1
𝑥 =− = −0.577350269
3

𝑤 =1
𝑤 =1 [Refer Table 4.1]
We know that,
𝑓(𝑥) = 𝑥

𝑓(𝑥 ) = 𝑥 = (0.577350269)

𝑓(𝑥 ) = 0.333333333

𝑤 𝑓(𝑥 = 1 × 0.333333333)

𝑤 𝑓(𝑥 ) = 0.333333333 … . (1)

𝑓(𝑥 ) = 𝑥 = (−0.577350269)

𝑓(𝑥 ) = 0.333333333

𝑤 𝑓(𝑥 ) = 1 × 0.333333333

𝑤 𝑓(𝑥 ) = 0.333333333 … . (2)

Adding (1),(2)
𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 0.333333333 + 0.333333333 = 0.666666666
4.68 Isoparametric Elements

Result:

𝑥 𝑑𝑥 = 0.666666666

Example 4.19

Evaluate the integral I = ∫ ( 2 + x + x ) dx and compare with exact solution.

Given:
𝟏

Integral, I = ( 𝟐 + 𝒙 + 𝒙𝟐 ) 𝒅𝒙
𝟏

To find: Evaluate the integral by using Gaussian quadrature.

Solution: We know that, the given integrand is a polynomial of order 2.
So, 2n – 1 = 2
2n = 3
n = 1.5 ≈ 2
We should use two sampling points.
For two point Gaussian quadrature,

1
𝑥 =+ = 0.577350269
3

1
𝑥 =− = −0.577350269
3

𝑤 =1
𝑤 =1 [Refer Table 4.1]
We know that,
𝑓(𝑥) = 2 + 𝑥 + 𝑥
𝑓(𝑥 ) = 2 + 𝑥 + 𝑥
= 2 + (0.577350269) + (0.577350269)
 Finite Element Analysis 4.69

𝑓(𝑥 ) = 2.9106836

𝑤 𝑓(𝑥 ) = 1 × 2.9106836

𝑤 𝑓(𝑥 ) = 2.9106836 … . (1)

𝑓(𝑥 ) = 2 + 𝑥 + 𝑥

= 2 − (0.577350269) + (−0.577350269)2

𝑓(𝑥 ) = 1.755983

𝑤 𝑓(𝑥 ) = 1 × 1.755983

𝑤 𝑓(𝑥 ) = 1.755983 … . (2)

Adding (1) & (2),

𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 2.9106836 + 1.755983 = 4.666666

( 2 + 𝑥 + 𝑥 ) 𝑑𝑥 = 4.666666

Exact solution:

1 1
( 2 + 𝑥 + 𝑥 ) 𝑑𝑥 = 2[𝑥] + [𝑥 ] + [𝑥 ]
2 3

1 1
= 2[1 − (1)] + [1 – (1)] + [1 – (1)]
2 3
= 4.666666
Result:

( 2 + 𝑥 + 𝑥 ) 𝑑𝑥 = 4.666666 [By two point Gaussian quadrature]

( 2 + 𝑥 + 𝑥 ) 𝑑𝑥 = 4.666666 [By exact method]

4.70 Isoparametric Elements

Example 4.20
𝟏 𝐱
Evaluate ∫ 𝟏
𝐞 𝐝𝐱 by applying 3 point gaussian quadrature.

Given:

Integral, I = 𝑒 𝑑𝑥

𝑓(𝑥) = 𝑒 𝑑𝑥
To find: Evaluate the integral by using Gaussian quadrature.
Solution: We know that, for three point Gaussian quadrature

3
𝑥 = = 0.774596669
5

𝑥 =0

3
𝑥 =− = −0.774596669
5

5
𝑤 = = 0.555555
9
8
𝑤 = = 0.888888
9
5
𝑤 = = 0.555555 [Refer Table 4.1]
9
We know that,
𝑓(𝑥) = 𝑒
.
𝑓(𝑥 ) = 𝑒 = 𝑒

𝑓(𝑥 ) = 0.460889634

𝑤 𝑓(𝑥 ) = 0.555555 × 0.460889634

𝑤 𝑓(𝑥 ) = 0.25604954 … . (1)

 Finite Element Analysis 4.71

𝑓(𝑥 ) = 𝑒 = 𝑒

𝑓(𝑥 ) = 1

𝑤 𝑓(𝑥 ) = 0.888888 × 1

𝑤 𝑓(𝑥 ) = 0.888888 … . (2)

( . )
𝑓(𝑥 ) = 𝑒 = 𝑒
.
𝑓(𝑥 ) = 𝑒

𝑓(𝑥 ) = 2.169716837

𝑤 𝑓(𝑥 ) = 0.555555 × 2.169716837

𝑤 𝑓(𝑥 ) = 1.205397037 … . (3)

Adding equations (1),(2) and (3),

𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 )
= 0.25604954 + 0.888888 + 1.205397037
= 2.350334
Result:

𝑒 𝑑𝑥 = 2.350334

Example 4.21
𝟏
𝐱
Evaluate 𝐜𝐨𝐬 𝟐 𝐝𝐱 by applying 3 point gaussian quadrature.
𝟏

Given:

𝑥
Integral, I = cos 𝑑𝑥
2

𝑥
𝑓(𝑥) = cos
2
4.72 Isoparametric Elements

To find: Evaluate the integral by using Gaussian quadrature.

Solution: We know that, for three point Gaussian quadrature

3
𝑥 = = 0.774596669
5

𝑥 =0

3
𝑥 =− = −0.774596669
5

5
𝑤 = = 0.555555
9
8
𝑤 = = 0.888888
9
5
𝑤 = = 0.555555 [Refer Table 4.1]
9

We know that,
𝑥
𝑓(𝑥) = cos
2
𝑥 0.774596669
𝑓(𝑥 ) = cos = cos rad
2 2

𝑓(𝑥 ) = cos = 0.9259328256

𝑤 𝑓(𝑥 ) = 0.555555 × 0.925932825

𝑤 𝑓(𝑥 ) = 0.51440661 … . (1)

𝑥 0
𝑓(𝑥 ) = cos = cos rad
2 2

𝑓(𝑥 ) = 1

𝑤 𝑓(𝑥 ) = 0.888888 × 1
 Finite Element Analysis 4.73

𝑤 𝑓(𝑥 ) = 0.888888 … . (2)

𝑥 −0.774596669
𝑓(𝑥 ) = cos = cos rad
2 2

𝑓(𝑥 ) = 0.925932825

𝑤 𝑓(𝑥 ) = 0.555555 × 0.925932825

𝑤 𝑓(𝑥 ) = 0.51440661 … . (3)

Adding equations (1),(2) and (3),

𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 0.51440661 + 0.888888 + 0.51440661
= 1.91770
Result:

𝑥
cos 𝑑𝑥 = 1.91770
2

Example 4.22

Evaluate cos dx by applying 3 point gaussian quadrature and compare with

exact solution.
Given:

𝜋𝑥
Integral, I = cos 𝑑𝑥
2

𝜋𝑥
𝑓(𝑥) = cos
2
To find: Evaluate the integral by using 3 Gaussian quadrature and compare with exact
solution
Solution: We know that, for three point Gaussian quadrature
3
𝑥 = = 0.774596669
5
4.74 Isoparametric Elements

𝑥 =0

3
𝑥 =− = −0.774596669
5

5
𝑤 = = 0.555555
9
8
𝑤 = = 0.888888
9

𝑤 = = 0.555555 [Refer Table 4.1]

We know that,
𝜋𝑥
𝑓(𝑥) = cos
2
𝜋𝑥 𝜋 × (0.774596669)
𝑓(𝑥 ) = cos = cos
2 2

𝑓(𝑥 ) = 0.346711

𝑤 𝑓(𝑥 ) = 0.555555 × 0.346711

𝑤 𝑓(𝑥 ) = 0.1926174 … . (1)

𝜋𝑥 𝜋×0
𝑓(𝑥 ) = cos = cos rad
2 2

𝑓(𝑥 ) = 1

𝑤 𝑓(𝑥 ) = 0.888888 × 1 … . (2)

𝜋𝑥 𝜋 × (−0.774596669)
𝑓(𝑥 ) = cos = cos
2 2

𝑓(𝑥 ) = 0.346711

𝑤 𝑓(𝑥 ) = 0.555555 × 0.346711

 Finite Element Analysis 4.75

𝑤 𝑓(𝑥 ) = 0.1926174 … . (3)

Adding equations (1),(2) and (3),

𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 0.1926174 + 0.888888 + 0.1926174
= 1.2741228

𝜋𝑥
cos 𝑑𝑥 = 1.2741228
2

Exact solution:
𝜋𝑥
𝜋𝑥 sin 2
cos = 𝜋
2
2
2 𝜋 −𝜋
= × sin − sin = 1.2732395
𝜋 2 2
Result:

1. cos = 1.2741228 [By three points gauss quadrature]

2. cos = 1.2732395 [By exact method]

Example 4.23

Evaluate x + cos dx by applying 3 point gaussian quadrature and compare with

exact solution.
Given:

x
Integral, I = x + cos dx
2

x
f(x) = x + cos dx
2
4.76 Isoparametric Elements

To find: Evaluate the integral by using 3 Gaussian quadrature and compare with exact
solution
Solution: We know that, for three point Gaussian quadrature

3
𝑥 = = 0.774596669
5

𝑥 =0

3
𝑥 =− = −0.774596669
5

5
𝑤 = = 0.555555
9
8
𝑤 = = 0.888888
9
5
𝑤 = = 0.555555 [Refer Table 4.1]
9

We know that,
𝑓(𝑥) = x
𝑥
𝑓(𝑥) = 𝑥 + cos
2
𝑥
𝑓(𝑥 ) = 𝑥 + cos
2
0.774596669
= (0.774596669) + cos rad
2

𝑓(𝑥 ) = 1.5259328

𝑤 𝑓(𝑥 ) = 0.555555 × 1.5259328

𝑤 𝑓(𝑥 ) = 0.8477396 … . (1)

 Finite Element Analysis 4.77
𝑥 0
𝑓(𝑥 ) = 𝑥 + cos = (0)2 + cos rad
2 2
𝑓(𝑥 ) = 1

𝑤 𝑓(𝑥 ) = 0.888888 × 1

𝑤 𝑓(𝑥 ) = 0.888888 … . (2)

𝑥
𝑓(𝑥 ) = 𝑥 + cos
2
−0.774596669
= (−0.774596669) + cos rad
2
𝑓(𝑥 ) = 1.5259328
𝑤 𝑓(𝑥 ) = 0.555555 × 1.5259328
𝑤 𝑓(𝑥 ) = 0.8477396 … . (3)
Adding equations (1),(2) and (3),
𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 0.8477396 + 0.888888 + 0.8477396

𝑥
𝑥 + cos 𝑑𝑥 = 2.58436
2

Exact solution:
𝑥
𝑥 𝑥 sin 2
𝑥 + cos 𝑑𝑥 = +
2 3 1
2
1 1 −1
= [1 − (−1) ] + 2 sin − sin rad
3 2 2
= 2.58436
Result:

1. 𝑥 + cos 𝑑𝑥 = 2.58436 [By three points gauss quadrature]

2. 𝑥 + cos 𝑑𝑥 = 2.58436 [By exact method]

4.78 Isoparametric Elements

Example 4.24

Evaluate dx by applying 3 point gaussian quadrature.

Given:

cos 𝑥
Integral, I = 𝑑𝑥
1− 𝑥

cos 𝑥
𝑓(𝑥) = 𝑑𝑥
1− 𝑥
To find: Evaluate the integral by using 3 Gaussian quadrature.
Solution: We know that, for three point Gaussian quadrature.

3
𝑥 = = 0.774596669
5

𝑥 =0

3
𝑥 =− = −0.774596669
5

5
𝑤 = = 0.555555
9
8
𝑤 = = 0.888888
9
5
𝑤 = = 0.555555 [Refer Table 4.1]
9
We know that,
cos 𝑥
𝑓(𝑥) =
1− 𝑥
cos 𝑥 cos(0.774596669)
𝑓(𝑥 ) = = rad
1− 𝑥 1 − (0.774596669)
𝑓(𝑥 ) = 1.78675798
𝑤 𝑓(𝑥 ) = 0.555555 × 1.78675798
 Finite Element Analysis 4.79

𝑤 𝑓(𝑥 ) = 0.99264233 … . (1)

cos 𝑥 cos(0)
𝑓(𝑥 ) = =
1− 𝑥 1 − (0)
𝑓(𝑥 ) = 1
𝑤 𝑓(𝑥 ) = 0.888888 × 1
𝑤 𝑓(𝑥 ) = 0.888888 … . (2)
cos 𝑥 cos(−0.774596669)
𝑓(𝑥 ) = = rad
1− 𝑥 1 − (−0.774596669)
𝑓(𝑥 ) = 1.78675798
𝑤 𝑓(𝑥 ) = 0.555555 × 1.78675798
𝑤 𝑓(𝑥 ) = 0.99264233 … . (3)
Adding equations (1),(2) and (3),
𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 0.99264233 + 0.888888 + 0.99264233
= 2.87417
Result:

cos 𝑥
𝑑𝑥 = 2.87417
1− 𝑥

Example 4.25

Evaluate I = 3e + x + dx using one point and two point gaussian

quadrature and compare with exact solution.

Given:

1
Integral, I = 3𝑒 + 𝑥 + 𝑑𝑥
𝑥+2

1
𝑓(𝑥) = 3𝑒 + 𝑥 +
𝑥+2
4.80 Isoparametric Elements

To find: 1. Evaluate the integral by using one point and two point gaussian quadrature.
2. Compare with exact solution.
Solution:
One point gaussian quadrature: We know that, for one point Gaussian quadrature.
𝑥 =0; 𝑤 =2
1
𝑓(𝑥) = 3𝑒 + 𝑥 +
𝑥+2
1
𝑓(𝑥 ) = 3𝑒 + 𝑥 +
𝑥 +2
1
= 3𝑒 + 0 +
0+2
𝑓(𝑥 ) = 3.5
𝑤 𝑓(𝑥 ) = 2 × 3.5
𝑤 𝑓(𝑥 ) = 7

1
3𝑒 + 𝑥 + 𝑑𝑥 = 𝟕 for one point Gauss quadrature
𝑥+2

Two point gaussian quadrature: We know that, for two point Gaussian quadrature.

1
𝑥 =+ = 0.577350269
3

1
𝑥 =− = −0.577350269
3

𝑤 =1
𝑤 =1 [Refer Table 4.1]
1
𝑓(𝑥) = 3𝑒 + 𝑥 +
𝑥+2
1
𝑓(𝑥 ) = 3𝑒 + 𝑥 +
𝑥 +2
 Finite Element Analysis 4.81

.
1
= 3𝑒 + (0.577350269) +
0.577350269 + 2

𝑓(𝑥 ) = 6.065265

𝑤 𝑓(𝑥 ) = 1 × 6.065265

𝑤 𝑓(𝑥 ) = 6.065265 … . (1)

1
𝑓(𝑥 ) = 3𝑒 + 𝑥 +
𝑥 +2

.
1
= 3𝑒 + (−0.577350269) +
−0.577350269 + 2

𝑓(𝑥 ) = 2.7203987

𝑤 𝑓(𝑥 ) = 1 × 2.7203987

𝑤 𝑓(𝑥 ) = 2.7203987 … . (2)

Adding (1) and (2),
𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 6.065265 + 2.7203987 = 8.7859

1
3𝑒 + 𝑥 + 𝑑𝑥 = 8.7859 for two point Gauss quadrature
𝑥+2

Exact solution:

1
I = 3𝑒 + 𝑥 + 𝑑𝑥
𝑥+2

𝑥
= 3[𝑒 ] + + [ln(𝑥 + 2)]
3
1
= 3 [𝑒 − (𝑒 )] + [1 − (−1) ][ln(1 + 2 ) − ln(−1 + 2)]
3
1
= 3 [2.718 – 0.3678] + [1 + 1] + ln(3) − ln(1)
3
4.82 Isoparametric Elements

3𝑒 + 𝑥 + 𝑑𝑥 = 8.8158

Result:
1. One point gaussian quadrature

1
3𝑒 + 𝑥 + 𝑑𝑥 = 𝟕
𝑥+2

2. Two point gaussian quadrature

3𝑒 + 𝑥 + 𝑑𝑥 = 8.7859

3. Exact solution:

3𝑒 + 𝑥 + 𝑑𝑥 = 8.8158

Example 4.25
𝟏 𝟏
Evaluate 𝐈 = ∫ 𝟏 ∫ 𝟏(𝟐𝐱 𝟐 + 𝟑 𝐱𝐲 + 𝟒𝐲 𝟐 ) 𝐝𝐱 𝐝𝐲 using Gauss integration.

Given:
𝟏
𝟏
Integral, I = (𝟐𝒙𝟐 + 𝟑 𝒙𝒚 + 𝟒𝒚𝟐 ) 𝒅𝒙 𝒅𝒚
𝟏
𝟏

𝑓(𝑥, 𝑦) = [2𝑥 + 3 𝑥𝑦 + 4𝑦 ]
To find: Evaluate the integral by using Gaussian quadrature.
Solution: We know that, the given integral is a polynomial of order 2. So, for exact
integration,
So, 2n – 1 = 2
2n = 3
n = 1.5 ≈ 2
We should use two sampling points.
 Finite Element Analysis 4.83

For two point Gaussian quadrature,

𝑥 = 0.57735, 𝑦 = 0.57735

𝑥 = −0.57735, 𝑦 = −0.57735

𝑤 =1

𝑤 =1 [Refer Table 4.1]

For two points scheme, the above equation can be written as,

𝑓(𝑥, 𝑦)𝑑𝑥 𝑑𝑦

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

+ 𝑤 𝑓(𝑥 , 𝑦 ) … (1)
We know that,
𝑓(𝑥, 𝑦) = [2𝑥 + 3 𝑥𝑦 + 4𝑦 ]

𝑤 𝑓(𝑥 , 𝑦 ) = 𝑤 (2𝑥 + 3 𝑥 𝑦 + 4𝑦 )

= 12 [2 (0.57735)2 + 3 (0.57735)(0.57735) + 4 (0.57735)2]

𝑤 𝑓(𝑥 , 𝑦 ) = 3 … (2)

𝑤 𝑤 𝑓(𝑥 , 𝑦 ) = 𝑤 𝑤 (2𝑥 + 3 𝑥 𝑦 + 4𝑦 )

= 1×1 [2 (0.57735)2 + 3 (0.57735)(-0.57735) + 4 (-0.57735)2]

𝑤 𝑤 𝑓(𝑥 , 𝑦 ) = 1 …(3)

𝑤 𝑤 𝑓(𝑥 , 𝑦 ) = 𝑤 𝑤 (2𝑥 + 3 𝑥 𝑦 + 4𝑦 )

= 1×1 [2 (-0.57735)2 + 3 (-0.57735)(0.57735) + 4 (0.57735)2]

𝑤 𝑤 𝑓(𝑥 , 𝑦 ) = 1 … (4)

𝑤 𝑓(𝑥 , 𝑦 ) = 𝑤 (2𝑥 + 3 𝑥 𝑦 + 4𝑦 )

= 12 [2 (-0.57735)2 + 3 (-0.57735)(-0.57735) + 4 (-0.57735)2]

4.84 Isoparametric Elements

𝑤 𝑓(𝑥 , 𝑦 ) = 3 … (5)

Substitute the equation (2),(3),(4) and (5) in equation (1),

(2𝑥 + 3 𝑥𝑦 + 4𝑦 )𝑑𝑥 𝑑𝑦 = 3 + 1 + 1 + 3 = 8

(2𝑥 + 3 𝑥𝑦 + 4𝑦 )𝑑𝑥 𝑑𝑦 = 8

Verification: The exact solution of integral is,

(2𝑥 + 3 𝑥𝑦 + 4𝑦 ) 𝑑𝑥 𝑑𝑦

2 3
= 𝑥 + 𝑦𝑥 + 4𝑦 𝑥 𝑑𝑦
3 2

2 3
= (1 + 1) + 𝑦(1 − 1) + 4𝑦 (1 + 1) 𝑑𝑦
3 2

4
= + 8𝑦 𝑑𝑦
3

4 8 4 8
= 𝑦+ 𝑦 = (1 + 1) + (1 + 1)
3 3 3 3
8 16 24
= + = =8
3 3 3

(2𝑥 + 3 𝑥𝑦 + 4𝑦 ) 𝑑𝑥 𝑑𝑦 = 8

Result:

The Integral (2𝑥 + 3 𝑥𝑦 + 4𝑦 ) 𝑑𝑥 𝑑𝑦 = 8

 Finite Element Analysis 4.85

Example 4.27
Integrate the function f(r)= 1+r+r2+r3 between the limits -1 and +1 using
i) Exact Method
ii) Gauss integration method and compare the two results
Given function , f(r)= 1+r+r2+r3
To find: Evalute the integral by using gauss integration method and compare with exact
method.
Solution:
We know that, the given integrated is a polynomial of order 3 .
So, 2 n -1 = 3
 2n =4
 n =2
We should use two sampling points.
For two point Gaussian quadratic,

1
𝑟 = + = 0.577350269
3

1
𝑟 = − = −0.577350269
3

𝑤 =1

𝑤 =1

We know that, f(r)= 1+r+r2+r3

𝑓(𝑟 ) = 1 + 𝑟 + 𝑟 + 𝑟

= 1 + 0.577350269 + (0.577350269) + (0.577350269)

𝑓(𝑟 ) = 2.1031336
4.86 Isoparametric Elements

𝑤 𝑓(𝑟 ) = 1 × 2.1031336

𝑤 𝑓(𝑟 ) = 2.1031336 … (1)

𝑓(𝑟 ) = 1 + 𝑟 + 𝑟 + 𝑟

= 1 + (0.577350269) + (−0.577350269)
+ (0.577350269)

𝑓(𝑟 ) = 0.5635329

𝑤 𝑓(𝑟 ) = 1 × 0.5635329

𝑤 𝑓(𝑟 ) = 0.5635329 … (2)

Adding (1) and (2) ,

𝑤 𝑓(𝑟 ) + 𝑤 𝑓(𝑟 ) = 2.1031336 + 0.5635326 = 2.666666

 (1 + 𝑟 + 𝑟 + 𝑟 ) = 2.666666

Exact method:

𝑟 𝑟 𝑟
(1 + 𝑟 + 𝑟 + 𝑟 ) = 𝑟 + + +
2 3 4

1 1 1
= [𝑟] − [𝑟 ] + [𝑟 ] + [𝑟 ]
2 3 4
1 1 1
= [1 − (−1)] + [(1) − (−1) ] + [1 − (−1) ] + [1
2 3 4
− (−1) ]

1 1 1
= 2 + (0) + (1 + 1) + (0)
2 3 4

(1 + 𝑟 + 𝑟 + 𝑟 ) = 2.666666
 Finite Element Analysis 4.87

Result:

1. (1 + 𝑟 + 𝑟 + 𝑟 )𝑑𝑟 = 2.666666 [By ginuss integration]

2. (1 + 𝑟 + 𝑟 + 𝑟 )𝑑𝑟 = 2.666666 [By exact integration]

Example 4.28

Evaluate the integral I = ∫ (a + a x + a x + a x )dx using Gauss integration

𝐆𝐢𝐯𝐞𝐧: Integral, I = (a + a x + a x + a x )dx

𝑓(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥62 + 𝑎 𝑥63

To find: evaluate the integral by using Gauss integration.
Solution: we know that, the given integrand is a polynomial of order 3. So, for exact
integration,
2n–1=3
2n=4
n=2
We should use two sampling points,
For two point Gaussian quadrature,

1
𝑥 = + = 0.577350269
3

1
𝑥 = − = −0.577350269
3

𝑤 =1
𝑤 =1
4.88 Isoparametric Elements

We know that 𝑓(𝑥) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥

 𝑓(𝑥 ) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥

= 𝑎 + 𝑎 (0.577350269) + (0.33333333)𝑎
+ (0.192450089)𝑎

 𝑤 𝑓(𝑥 ) = 1 (𝑎 + 0.577350269 𝑎 + 0.33333333 𝑎

+ 0.192450089 𝑎 ) … (1)

𝑓(𝑥 ) = 𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥

= 𝑎 + 𝑎 (−0.577350269) + 𝑎 (−0.577350269)
+ 𝑎 (−0.577350269)

𝑓(𝑥 ) = 𝑎 − (0.577350269 ) 𝑎 + 0.33333333 𝑎 − 0.192450089 𝑎

 𝑤 𝑓(𝑥 ) = 1 [ 𝑎 − 0.577350269 𝑎 + 0.33333333 𝑎

+ 0.192450089 𝑎 ] … (2)

Adding equation (1) and (2),

𝑤 𝑓(𝑥 ) + 𝑤 𝑓(𝑥 ) = 2𝑎 + 0.666666 𝑎

 (𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 )𝑑𝑥 = 2𝑎 + 0.666666 𝑎

2
 (𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 )𝑑𝑥 = 2 𝑎 + 𝑎
3

Verification:

𝑥 𝑥 𝑥
(𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 )𝑑𝑥 = 𝑎 𝑥 + 𝑎 +𝑎 +𝑎
2 3 4

𝑎 𝑎 𝑎
= 𝑎 [𝑥] + [𝑥62 ] + [𝑥 ] + [𝑥 ]
2 3 4
 Finite Element Analysis 4.89
𝑎 𝑎 𝑎
= 𝑎 [1 − (−1)] + [1 − (−1) ] + [1 − (−1) ] + [1 − (−1) ]
2 3 4
2
= 2𝑎 + 𝑎
3

2𝑎
𝑹𝒆𝒔𝒖𝒍𝒕: (𝑎 + 𝑎 𝑥 + 𝑎 𝑥 + 𝑎 𝑥 )𝑑𝑥 = 2 𝑎 +
3
 UNIT 5
APPLICATION TO HEAT TRANSFER
AND DYNAMIC ANALYSIS

5.1 HEAT TRANSFER

Heat transfer can be defined as the transmission of energy from one region to
another region due to temperature difference. A knowledge of the temperature
distribution within a body is important in many engineering problems. There are three
modes of heat transfer.
They are:
(i) Conduction
(ii) Convection
(iii) Radiation
(i) Conduction
Heat conduction is a mechanism of heat transfer from a region of high
temperature to a region of low temperature within a medium (solid, liquid or gases) or
between different medium in direct physical contact.
In conduction, energy exchange takes place by the kinematic motion or direct
impact of molecules. Pure conduction is found only in solids.
(ii) Convection
Convection is a process of heat transfer that will occur between a solid surface
and a fluid medium when they are at different temperatures.
Convection is possible only in the presence of fluid medium.
5.2 Application to Heat Transfer and Dynamic Analysis

(iii) Radiation
The heat transfer from one body to another without any transmitting medium is
known as radiation. It is an electromagnetic wave phenomenon.

5.2 DERIVATION OF TEMPERATURE FUNCTION (T) AND

SHAPE FUNCTION (N) FOR ONE DIMENSIONAL HEAT
CONDUCTION ELEMENT
Consider a bar element with nodes 1 and 2 as shown in Fig.5.1. T 1, and T₂ are
the temperatures at the respective nodes. So, T 1, and T₂ are considered as degrees of
freedom of this bar element.
Since the element has got two degrees of freedom, it will have two generalized
co-ordinates.

Fig.5.1
Writing the equation in matrix form,
𝑎
𝑇 = [1𝑥] 𝑎

At node 1, T = T1, x = 0
At node 1, T = T2, x = 1
Substitute the above values in equation
 𝑇 =𝑎
 𝑇 =𝑎 +𝑎 𝑙
Assembling the equations in matrix form
𝑇 1 0 𝑎
=
𝑇 1 𝑙 𝑎
 Finite Element Analysis 5.3
𝑎 1 0 𝑇
 𝑎 =
1 𝑙 𝑇

1 1 0 𝑇
=
𝑙−0 1 𝑙 𝑇

𝑎 𝑎 1 𝑎 𝑎−
𝑁𝑜𝑡𝑒: 𝑎 𝑎 = × −𝑎 𝑎
(𝑎 𝑎 −𝑎 𝑎 )

𝑎 1 𝑙 0 𝑇
 𝑎 =
𝑙 −1 𝑙 𝑇
a
Substitute a values in equation

1 𝑙 0 𝑇
 𝑇 = [ 1 𝑥]
𝑙 −1 1 𝑇

1 𝑙 0 𝑇
= [ 1 𝑥]
𝑙 −1 1 𝑇

1 𝑇
= [ 𝑙 − 𝑥 0 + 𝑥]
𝑙 𝑇

[ ∵ 𝑚𝑎𝑡𝑟𝑖𝑥 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑐𝑎𝑡𝑖𝑜𝑛 (1 × 2)(2 × 2) = (1 × 2)]

𝑙−𝑥 𝑥 𝑇
𝑇=
𝑙 𝑙 𝑇

𝑇
𝑇 = [𝑁 𝑁]
𝑇

Temperature function, 𝑇 =𝑁 +𝑇 +𝑁 𝑇

𝑙−𝑥
where, shape functions, 𝑁 =
𝑙
𝑥
𝑁 =
𝑙
5.4 Application to Heat Transfer and Dynamic Analysis

5.3 DERIVATION OF STIFFNESS MATRIX FOR ONE

DIMENSIONAL HEAT CONDUCTION ELEMENT
We know that,
Consider a one dimensional bar element with nodes 1 and 2 as shown in Fig.5.2.
Let T and T₂ be the temperatures at the respective nodes and k be the thermal conductivity
of the material.

Fig. 5.2
We know that

Stiffness matrix [ 𝐾] = [𝐵] [𝐷][𝐵] 𝑑𝑣

In one dimensional element,

Temperature function, 𝑇 =𝑁 𝑇 +𝑁 𝑇
𝑙−𝑥
where, shape functions, 𝑁 =
𝑙
𝑥
𝑁 =
𝑙
We know that,
𝑑𝑁 𝑑𝑁
strain − Displacement matrix, [𝐵] =
𝑑𝑥 𝑑𝑥
−1 1
 [𝐵] =
𝑙 𝑙
−1
 [𝐵] = 𝑙
1
𝑙
 Finite Element Analysis 5.5

In one dimensional heat conduction problems,

[ D] = [ K ] =k = Thermal conductivity of the material
Substitute [ B ], [𝐵] and [ D] values in stiffness matrix equation
−1

Stiffness matrix [𝐾 ]= 𝑙 × 𝑘 × −1 1 𝑑𝑣
for heat conduction 1 𝑙 𝑙
𝑙
1 −1
= 𝑙 𝑙 𝑘 𝑑𝑣
−1 1
𝑙 𝑙

[ ∵ 𝑚𝑎𝑡𝑟𝑖𝑥 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑐𝑎𝑡𝑖𝑜𝑛 (2 × 1)(1 × 2) = (2 × 2)]

1 −1
= 𝑙 𝑙 𝑘 𝐴 𝑑𝑥 [∵ 𝑑𝑣 = 𝐴 × 𝑑𝑥]
−1 1
𝑙 𝑙

=𝐴𝑘 ∫ 𝑑x

1 −1
=𝐴𝑘 𝑙 𝑙 [ 𝑥]
−1 1
𝑙 𝑙
1 −1
=𝐴𝑘 𝑙 𝑙 (𝑙 − 0)
−1 1
𝑙 𝑙
1 −1
=𝐴𝑘𝑙 𝑙 𝑙
−1 1
𝑙 𝑙
5.6 Application to Heat Transfer and Dynamic Analysis
𝐴 𝑘 𝑙 1 −1
=
𝑙 −1 1
𝐴 𝑘 𝑙 1 −1
[K ]=
𝑙 −1 1
Where, A = Area of the element, m2
K = Thermal conductivity of the element, W/mK
l = Length of the element, m

5.4 FINITE ELEMENT EQUATION OF ONE DIMENSIONAL HEAT

CONDUCTION PROBLEMS
We know that,
General force equation is {𝐹} = [ 𝐾 ]{ 𝑇}
Where, {F} is a element force vector [Column matrix]
[KC] is a stiffness matrix [Row matrix]
{T} is a model temperature [Column matrix]
For one dimensional heat conduction problems, stiffness matrix, [ K ] is given by
𝐴 𝑘 1 −1
[𝐾 ] =
𝑙 −1 1
Consider a two noded element as shown in Fig. 5.4
𝐹
Force vector { 𝐹} =
𝐹
𝑇
Nodal temperature { 𝑇} =
𝑇

Fig .5.3
 Finite Element Analysis 5.7

Subsitute [K ] { F} and{ T} values in equaiton

𝐹 𝐴 𝑘 1 −1 𝑇
 =
𝐹 𝑙 −1 1 𝑇
Case (i) One dimensional heat conduction with free end convection
Consider a one dimensional element with nodes 1 and 2 as shown in T1 and T2 are
the temperatures at the respective nodes. Assume
convection occurs only from the right end of the element as
shown in Fig.5.4
Stiffness matrix [𝐾 ] for one dimensional heat
conduction element is given by
𝐴 𝑘 1 −1
[𝐾 ] =
𝑙 −1 1
The convection term contribution to the stiffness matrix is given by Fig.5.4

[𝐾 ] = ℎ [𝑁] [𝑁]𝑑𝐴

Where h = Heat transfer coefficient, W/m2k

N = Shape factor
We know that,
𝑙−𝑥 𝑥
Shape factor, [ 𝑁] = [ 𝑁 𝑁 ] =
𝑙 𝑙
At node 2, x=l
 [𝑁] = [𝑁 𝑁 ] = [ 0 1]

 [𝑁] = 0
1
Substitute [N] and [ N ]T values in equation,
0 [
 [𝐾 ] = ℎ 0 1 ] 𝑑𝐴
1
0 0
[𝐾 ] =ℎ 𝑑𝐴
0 1
5.8 Application to Heat Transfer and Dynamic Analysis
(2 × 1) × (1 × 2) = (2 × 2)
0 0
[𝐾 ] =ℎ𝐴
0 1

Stiffness matrix [ 𝐾] = [ 𝐾 ] + [𝐾 ]

𝐴 𝑘 1 −1 0 0
[𝐾] = +ℎ𝐴
𝑙 −1 1 0 1

The convection force form the free end of the element is obtained from the
following relation,
𝑁 (𝑥 = 𝑙)
{𝐹 } =ℎ𝑇 𝐴
𝑁 (𝑥 = 𝑙)

{𝐹 } 0
=ℎ𝑇 𝐴
1
We know that, general force equation is
{𝐹} = [𝐾] {𝑇}
Substitute {𝐹} 𝑎𝑛𝑑 [𝐾] values,
0 𝐴 𝑘 1 −1 0 0 𝑇
 ℎ𝑇 𝐴 = +ℎ𝐴
1 𝑙 −1 1 0 1 𝑇

𝐴 𝑘 1 −1 0 0 𝑇 0
 +ℎ𝐴 = ℎ𝑇 𝐴
𝑙 −1 1 0 1 𝑇 1

Where, A = Area of the element, m2

k = Thermal conductivity of the element, W/mK
l = Length of the element
h = Heat transfer coefficient, W/m2K
T = Fluid temperature, K
T = Temperature, K
This is a finite element equation for one dimensional heat conduction element
with free end convection.
 Finite Element Analysis 5.9

Case (ii) One dimensional element with conduction, convection and internal heat
conduction, convection and internal heat generation.
Consider a rod with nodes 1 and 2 as shown in Fig.5.5. This rod is subjected to
conduction, convection and internal heat generation.

Fig.5.5.
We know that, heat conduction part of the stiffness matrix [ K ] for the one
dimensional element is
𝐴 𝑘 1 −1
[𝐾 ] =
𝑙 −1 1
Heat convection part of the stiffness matrix [ K ] for the one dimensional element
is given by

[𝐾 ]= ℎ [𝑁] [𝑁] 𝑑𝑆

=ℎ𝑃 [ 𝑁] [ 𝑁] 𝑑𝑥

[∵ 𝑑𝑆 = 𝑃 × 𝑑𝑥 𝑤ℎ𝑒𝑟𝑒 𝑃 = 𝑃𝑒𝑟𝑖𝑚𝑒𝑡𝑒𝑟 𝑜𝑓 𝑡ℎ𝑒 𝑒𝑙𝑒𝑚𝑒𝑛𝑡]

𝑙−𝑥
𝑙 𝑙−𝑥 𝑥
=ℎ𝑃 𝑥 𝑑𝑥
𝑙 𝑙
𝑙
𝑥 𝑥 𝑥
⎡ 1− − ⎤
=ℎ𝑃 ⎢ 𝑙 𝑙 𝑙 ⎥ 𝑑𝑥
⎢ 𝑥 𝑥 𝑥 ⎥
⎣ 𝑙−𝑙 𝑙 ⎦
5.10 Application to Heat Transfer and Dynamic Analysis

𝑥
⎡ 1− 𝑙 𝑥 𝑥 ⎤
⎢ − ⎥
1 2𝑙 3𝑙 ⎥
= ℎ 𝑃 ⎢3 × − 𝑙
⎢ ⎥
⎢𝑥 𝑥 𝑥 ⎥
−
⎣2 𝑙 3 𝑙 ⎦
3𝑙

𝑙
⎡ 1− ⎤
⎢ 𝑙 − 𝑙 𝑙
−
𝑙
− 0⎥
−3 −3 2𝑙 3𝑙
= ℎ𝑃⎢ ⎥
⎢ 𝑙 𝑙 ⎥
⎢ 𝑙 𝑙 𝑙 ⎥
⎣ 2𝑙−3𝑙 −0 3𝑙
−0 ⎦

𝑙 𝑙 𝑙
−
=ℎ𝑃 3 2 3
𝑙 𝑙 𝑙
−
2 3 3
𝑙 𝑙
=ℎ𝑃 3 6
𝑙 𝑙
6 3
ℎ𝑃𝑙 2 1
[𝐾 ]=
6 1 2

Stiffness matrix [ 𝐾 ] = [ 𝐾 ] + [𝐾 ]

𝐴 𝑘 1 −1 ℎ𝑃𝑙 2 1
[𝐾] = +
𝑙 −1 1 6 1 2
Force matrix due to heat generation is given by,

𝐹 = [ 𝑁] 𝑄 𝑑𝑉

= [𝑁] × 𝑄 × 𝐴 × 𝑑𝑥 [∵ 𝑑𝑉 = 𝐴 × 𝑑𝑥]
 Finite Element Analysis 5.11

= 𝑄×𝐴 [𝑁] 𝑑𝑥

𝑙−𝑥
= 𝑄×𝐴 𝑙 𝑑𝑥
𝑥
𝑙
𝑥
𝑙−
= 𝑄×𝐴 𝑙 𝑑𝑥
𝑥
𝑙
𝑥
⎧𝑥− ⎫
= 𝑄×𝐴 2𝑙
⎨ 𝑥 ⎬
⎩ 2𝑙 ⎭

𝑙
⎧𝑙− − 0⎫
= 𝑄×𝐴 2𝑙
⎨ 𝑙 ⎬
⎩ 2𝑙−0 ⎭
𝑙 𝑙
⎧ ⎫
= 𝑄×𝐴 2𝑙 = 𝑄×𝐴 2
⎨𝑙 ⎬ 𝑙
⎩2 𝑙 ⎭ 2
𝑙 𝑙
𝐹 = 𝑄×𝐴×
2 𝑙
Force matrix due to convection is given by

{𝐹 } = ℎ 𝑇 [ 𝑁] 𝑑𝑆

= ℎ 𝑇 [ 𝑁] 𝑃 × 𝑑𝑥 [∵ 𝑑𝑆 = 𝑃 × 𝑑𝑥]

=𝑃ℎ𝑇 [𝑁] 𝑑𝑥
5.12 Application to Heat Transfer and Dynamic Analysis
𝑙−𝑥
=𝑃ℎ𝑇 𝑙 𝑑𝑥
𝑥
𝑙
𝑥
⎧𝑥 −
⎫
=𝑃ℎ𝑇 2 𝑙
⎨ 𝑥 ⎬
⎩ 2𝑙 ⎭

𝑙
⎧𝑙 −
− 0⎫
=𝑃ℎ𝑇 2 𝑙
⎨ 𝑙 ⎬
⎩ 2𝑙−0 ⎭
𝑙
=𝑃ℎ𝑇 2
𝑙
2
𝑃ℎ𝑇 1 1
{𝐹 } =
2 1
Adding equations
Force matrix, {𝐹} = 𝐹 + {𝐹 }

𝑄𝐴𝑙 1 𝑃ℎ𝑇 1 1
= +
2 1 2 1

𝑄𝐴𝑙+𝑃ℎ𝑇 𝑙 1
{𝐹} =
2 1

We know that, General force equaiton is

{𝐹} = [ 𝐾]{ 𝑇}
Substitute { F } and [ K ] values,

𝑄𝐴𝑙+𝑃ℎ𝑇 𝑙 1 𝐴 𝑘 1 −1 ℎ𝑃𝑙 2 1 𝑇
 +
2 1 𝑙 −1 1 6 1 2 𝑇
 Finite Element Analysis 5.13

𝐴 𝑘 1 −1 ℎ𝑃𝑙 2 1 𝑇 𝑄𝐴𝑙+𝑃ℎ𝑇 𝑙 1
 + =
𝑙 −1 1 6 1 2 𝑇 2 1

Where, A = Area of the element, m2

k = Thermal conductivity of the element, W/mk
l = Length of the element, m
h = Heat transfer coefficient, W/m2k
P = Perimeter, m
T = Temperature, K
Q = Heat generation, W
T = Fluid temperature, K
This is a finite element equation for one dimensional element which is subjected
to conduction, convection and internal heat generation.

5.5 SOLVED PROBLEMS - HEAT TRANSFER [ONE

DIMENSIONAL]
Example 5.1
A wall of 0.6 m thickness having thermal conductivity of 1.2 W/mK. The wall
is to be insulated with a material of thickness 0.06 m having an average thermal
conductivity of 0.3 W/mK. The inner surface temperature is 1000 °C and outside of the
insulation is exposed to atmospheric air at 30°C with heat transfer coefficient of
35W/m²K. Calculate the nodal temperatures.
Given:
Thickness of the wall, l1 = 0.6 m
Thermal conductivity of the wall, k1 = 1.2 W/Mk
Thickness of the insulation, l2 = 0.06 m
Thermal conductivity of the insulation, k2 = 03 W/mK
Inner surface temperature, T1 = 10000C+ 273 = 1273 K.
Atmospheric air temperature, 𝑇 = 30 𝐶 + 273 = 303 𝐾
5.14 Application to Heat Transfer and Dynamic Analysis

Heat transfer coefficient at outer side, h = 35 W/m2k

To find: Nodal temperatures, (T2 and T3)

Solution:
For element 1: (Nodes 1, 2)
Finite element equation is
𝐴 𝑘 1 −1 𝑇 𝐹
=
𝑙 −1 1 𝑇 𝐹
For unit area,
𝐴 =1𝑚
1.2 1 −1 𝑇 𝐹
=
0.6 −1 1 𝑇 𝐹
1 2
1 −1 1 𝑇 𝐹
 = … (1)
−1 1 2 𝑇 𝐹
For element 2: (Nodes 2, 3)

This element is subjected to both conduction and convection. So, finite element
equation is
 Finite Element Analysis 5.15
𝐴 𝑘 1 −1 0 0 𝑇 0
+ℎ𝐴 =ℎ𝑇 𝐴
𝑙 −1 1 0 1 𝑇 1

1 × 0.3 1 −1 0 0 𝑇 0
+ 35 × 1 = 35 × 303 × 1
0.06 −1 1 0 1 𝑇 1

5 −5 0 0 𝑇 0
+ =
−5 5 0 35 𝑇 10.605 × 10

2 3

5 −5 𝟐 𝑇 0
= … . (2)
−5 5 𝟑 𝑇 10.605 × 10

Assemble the finite elements, i.e., assemble the finite element equations (1) and (2).

2 −2 0 𝑇 0
−2 7 −5 𝑇 0
0 −5 40 𝑇 = 10.605 × 10 … (3)
↓ ↓ ↓
[𝐾] [𝑇] [ 𝐹]

To solve the above equation, the following steps to be followed.

Step 1: The first row and first column of the stiffness matrix [ K] have been set equal to
0 except for the main diagonal, which has been set equal to 1.

1 0 0 𝑇 0
 0 7 −5 𝑇 = 0
0 −5 40 𝑇 10.605 × 10
Step 2: the first row of the force matrix is replaced by the known temperature at node 1,
i.e., T1.

1 0 0 𝑇 1273
 0 7 −5 𝑇 = 0
0 −5 40 𝑇 10.605 × 10
Step 3: The second row, first column of stiffness matrix [ K ] value (From equation no.3)
is multiplied by known temperature at node 1, i.e., -2 × 1273 = - 2546. This value (as
positive digit, i.e., 2546) has been added to the second row of the force matrix.
5.16 Application to Heat Transfer and Dynamic Analysis

1 0 0 𝑇 1273
0 7 −5 𝑇 = 2546 … (4)
0 −5 40 𝑇 10.605 × 10
Solving equation (4),

 7 𝑇 − 5 𝑇 = 2546 . . (5)

−5 𝑇 + 40 𝑇 = 10.605 × 10 . . (6)

Equation (5) × 8,

 56 𝑇 − 40 𝑇 = 20.368 × 10 . . (7)

Equation (6)  − 5 𝑇 + 40 𝑇 = 10.605 × 10

51 𝑇 = 30.973 × 10

𝑇 = 607.313 𝐾

Substitute T2 value in equation (5),

 7 × 607.313 − 5 𝑇 = 2546

𝑇 = 341.03 𝐾

Result: Nodal temperatures: T1 = 1273 K

T2 = 607.313 K

T3 = 341.03 K

Example 5.2
A furnace wall is made up of three layers, inside layer with thermal conductivity
8.5 W/mK, the middle layer with conductivity 0.25 W/mK, the outer layer with
conductivity 0.08 W/mK. The respective thickness of the inner, middle and outer layer
are 25 cm, 5 cm and 3 cm respectively. The inside temperature of the wall is 600 °C
and outside of the wall is exposed to atmospheric air at 30°C with heat transfer
coefficient of 45 W/m²K. Determine the nodal temperatures.
 Finite Element Analysis 5.17

Given:

Thermal conductivity of the inner layer, k1 = 8.5 W/mk

Thermal conductivity of the middle layer, k2 = 0.25 W/mK
Thermal conductivity of the outer layer, k3 = 0.08 W/mK
Inner thickness, l1 = 25 cm = 0.25 m
Middle layer thickness, l2 = 5 cm = 0.05 m
Outer layer thickness, l3 = 3 cm = 0.03 m
Inner temperature of the wall, T1 = 6000C+ 273 = 873 K.
Atmospheric air temperature, 𝑇 = 30 𝐶 + 273
= 303 𝐾
Heat transfer coefficient at outer side, h = 45 W/m2k
To find: Nodal temperatures, (T2 ,T3 and T4)
Solution:
For element 1: (Nodes 1, 2)
Finite element equation is

𝐴 𝑘 1 −1 𝑇 𝐹
=
𝑙 −1 1 𝑇 𝐹

For unit area, A1 = 1 m2

8.5 1 −1 𝑇 𝐹
 =
0.25 −1 1 𝑇 𝐹
5.18 Application to Heat Transfer and Dynamic Analysis

1 2
34 −34 𝟏 𝑇 𝐹
 = … (1)
−34 34 𝟐 𝑇 𝐹
For element 2: (Nodes 2, 3)
Finite element equation is
𝐴 𝑘 1 −1 𝑇 𝐹
=
𝑙 −1 1 𝑇 𝐹
1 × 0.25 1 −1 𝑇 𝐹
 =
0.05 −1 1 𝑇 𝐹
2 3
5 −5 2 𝑇 𝐹
 = … (2)
−5 5 3 𝑇 𝐹
For element 2: (Nodes 2, 3)
This element is subjected a both conduction and convection. So, finite element
equation is

𝐴 𝑘 1 −1 0 0 𝑇 0
+ℎ𝐴 =ℎ𝑇 𝐴
𝑙 −1 1 0 1 𝑇 1
1 × 0.08 1 −1 0 0 𝑇 0
+ 45 × 1 = 45 × 303 × 1
0.03 −1 1 0 1 𝑇 1
2.666 −2.666 0 0 𝑇 0
+ =
−2.666 2.666 0 45 𝑇 13.635 × 10
3 4
2.666 −2.666 3 𝑇 0
= … . (3)
−2.666 2.666 4 𝑇 13.635 × 10
 Finite Element Analysis 5.19

Assemble the finite elements, i.e., assemble the finite element equations (1), (2) and (3).
1 2 3 4
34 −34 0 0 𝟏 𝑇 𝐹
−34 34 + 5 −5 0 𝟐 𝑇 𝐹
=
0 −5 5 + 2.666 −2.666 𝟑 𝑇 𝐹
0 0 −2.666 47.666 𝟒 𝑇 𝐹
In this problem, there is no heat generation and there is no convection except from
the right end.

𝑆𝑜, {𝐹 } = {𝐹 } = {𝐹 } = 0

𝑎𝑛𝑑 {𝐹 } = 13.635 × 10

34 −34 0 0 𝑇 0
−34 39 −5 0 𝑇 0
0 −5 7.666 −2.666 𝑇 0
= 13.635 × 10 … (4)
0 0 −2.666 47.666 𝑇
↓ ↓ ↓
[𝐾] [𝑇] [ 𝐹]

To solve the above equation, the following steps to be followed.

Step 1: The first row and first column of the stiffness matrix [ K] have been set equal to
0 except for the main diagonal, which has been set equal to 1.

1 0 0 0 𝑇 0
0 39 −5 0 𝑇 0
 =
0 −5 7.666 −2.666 𝑇 0
0 0 −2.666 47.666 𝑇 13.635 × 10
Step 2: the first row of the force matrix is replaced by the known temperature at node 1,
i.e., T1.

1 0 0 0 𝑇 873
0 39 −5 0 𝑇 0
 =
0 −5 7.666 −2.666 𝑇 0
0 0 −2.666 47.666 𝑇 13.635 × 10
Equation (5) becomes,
5.20 Application to Heat Transfer and Dynamic Analysis

1 0 0 0 𝑇 873
0 1 −0.128 0 𝑇 761.076
=
0 0 1 −0.376 𝑇 541.614
0 0 0 46.655 𝑇 13.635 × 10

 46.655 𝑇 = 15.076 × 10

 𝑇 = 323.21 𝐾

 𝑇 − 0.379 𝑇 = 541.614

𝑇 − 0.379(323.21) = 541.614

𝑇 = 664.11 𝐾

 𝑇 − 0.218 (𝑇 ) = 761.076

𝑇 − 0.128(664.11) = 761.076

𝑇 = 846.08 𝐾
Result: Nodal temperatures:
𝑇 = 873 𝐾

𝑇 = 846.08 𝐾

𝑇 = 664.11 𝐾

𝑇 = 323.21 𝐾

Verification:
Heat flow through composite wall is given by
∆𝑇
𝑄=
𝑅
[From HMT data book, C. P. Kothandaraman, Page No. 43 & 44]
Wher𝑒, ∆𝑇 =𝑇 −𝑇
1
𝑅 =
1 𝑙 𝑙 𝑙 𝐼
+ + + +
ℎ 𝐴 𝑘 𝐴 𝑘 𝐴 𝑘 𝐴 ℎ 𝐴
 Finite Element Analysis 5.21
𝑇 −𝑇
 𝑄 =
1 𝑙 𝑙 𝑙 𝐼
+ + + +
ℎ 𝐴 𝑘 𝐴 𝑘 𝐴 𝑘 𝐴 ℎ 𝐴
Heat transfer coefficient at inner side is not given. So, neglect that term.
873 − 303
 𝑄 = [∵ 𝐴 = 1𝑚 ]
0.25 0.05 0.03 1
+ + +
8.5 0.25 0.08 45

 𝑄 = 909.62 𝑊/𝑚

We know that, interface temperature

𝑇 −𝑇 𝑇 −𝑇 𝑇 −𝑇 𝑇 −𝑇 𝑇 −𝑇
𝑄= = = = = … (6)
𝑅 𝑅 𝑅 𝑅 𝑅

𝑇 −𝑇
 𝑄=
𝑅

𝑇 −𝑇 𝑙
 𝑄= [∵ 𝑅 = ]
𝑙 𝑘 𝐴
𝑘 𝐴

873 − 𝑇
 909.62 = [∵ 𝐴 = 1 𝑚 ]
0.25
8.5

 𝑇 = 846.24 𝐾

𝑇 −𝑇
(5)  𝑄=
𝑅

𝑇 −𝑇 𝑙
= [∵ 𝑅 = ]
𝑙 𝑘 𝐴
𝑘 𝐴

846.24 − 𝑇
909.62 = [∵ 𝐴 = 1 𝑚 ]
0.05
0.25

 𝑇 = 664.31 𝐾
5.22 Application to Heat Transfer and Dynamic Analysis
𝑇 −𝑇
(6) 𝑄=
𝑅

𝑇 −𝑇 𝑙
= [∵ 𝑅 = ]
𝑙 𝑘 𝐴
𝑘 𝐴

664.31 − 𝑇
909.62 = [∵ 𝐴 = 1 𝑚 ]
0.03
0.08

 𝑇 = 323.20 𝐾

Example 5.3
An aluminium alloy fin of 7 mm thick and 50 mm long protrudes from a wall,
which is maintained at 120°C. The ambient air temperature is 22 °C. The heat transfer
coefficient and thermal conductivity of the fin material are 140 W/m²K and 55 W/mK
respectively. Determine the temperature distribution of fin.
Given:

Thickness, t = 7 mm = 0.007 m
Length, l = 50 mm = 0.050 m
Base temperature, T1 = 1200C + 273 = 393 K
Ambient temperature, T = 220C+ 273= 295 K
Heat transfer coefficient, h = 140 W/m2 K
Thermal conductivity, K = 55 W/mK
To find: Temperature distribution, (𝑇 , 𝑇 , 𝑇 , 𝑇 , … . . )𝑖𝑛 𝑡ℎ𝑒 𝑓𝑖𝑛.
 Finite Element Analysis 5.23

Solution: For simplicity’s sake, discretize the fin into two equal size element.

For element 1: (Nodes 1, 2)

We know that,
Finite element equation is
𝐴 1 −1 ℎ𝑃𝑙 2 1 𝑇 𝑄𝐴𝑙 +𝑃ℎ𝑇 𝑙 1
+ = … (1)
𝑙 −1 1 6 1 2 𝑇 2 1

where P = Perimeter = 2 × l (Approximately)

= 2 × 0.050

A = Area = Length × Thickness

= 0.050 × 0.007

𝐴 = 3.5 × 10 𝑚

Substitute A, P, k, h, l1 and T values in equation (1).

3.5 × 10 × 55 1 −1 140 × 0.1 × 0.025 2 1 𝑇

(1) +
0.025 −1 1 6 1 2 𝑇

𝑄 𝐴 𝑙 + 140 × 0.1 × 295 × 0.025𝑃 ℎ 𝑇 𝑙 1

=
2 1
𝑄𝐴𝑙
Heat generation Q is not given. So, neglect that term,
2
5.24 Application to Heat Transfer and Dynamic Analysis

1 −1 2 1 𝑇 1
 0.77 + 0.0583 = 51.625
−1 1 1 2 𝑇 1

0.77 −0.77 0.116 0.0583 𝑇 51.625

 + =
−0.77 0.77 0.0583 0.116 𝑇 51.625
1 2
0.886 −0.7117 1 𝑇 51.625
 = … (2)
−0.7117 0.886 2 𝑇 51.625
For element 2: (Nodes 2, 3)

Since all the parameters in elements (1) and element (2) are same, the finite
element equation becomes
2 3
0.886 −0.7117 𝟐 𝑇 51.625 𝟐
= … (3)
−0.7117 0.886 𝟑 𝑇 51.625 𝟑
Assemble the finite elements (2) and (3).
1 2 3
0.886 −0.7117 0 𝟏 𝑇 51.625
−0.7117 0.886 + 0.886 −0.7117 𝟐 𝑇 = 51.625 + 51.625
0 −0.7117 0.886 𝟑 𝑇 51.625

0.886 −0.7117 0 𝑇 51.625

−0.7117 0.886 + 0.886 −0.7117 𝑇 51.625 + 51.625
0 −0.7117 0.886 𝑇 = 51.625 … (4)
↓ ↓ ↓
[𝐾] [𝑇] [ 𝐹]
To solve the above equation, the following steps to be followed.
Step 1: The first row and first column of the stiffness matrix [K] have been set equal to
0 except for the main diagonal, which has been set equal to 1.
 Finite Element Analysis 5.25

1 0 0 𝑇 51.625
 0 1.772 −0.7117 𝑇 = 103.25
0 −0.7117 0.886 𝑇 51.625
Step 2: The first row of the force matrix is replaced by the known temperature at node 1,
i.e., T1.
1 0 0 𝑇 393
 0 1.772 −0.7117 𝑇 = 103.25
0 −0.7117 0.886 𝑇 51.625
Step 3: The second row, first column of stiffness matrix [ K ] value (From equation no.4)
is multiplied by known temperature at node 1, i.e., -0.7117 × 393 = - 279.69. This value
(as positive digit, i.e., 279.69) has been added to the second row of the force matrix.
1 0 0 𝑇 393
 0 1.772 −0.7117 𝑇 = 103.25 + 279.69
0 −0.7117 0.886 𝑇 51.625

1 0 0 𝑇 393
 0 1.772 −0.7117 𝑇 = 382.94 … (5)
0 −0.7117 0.886 𝑇 51.625

Solving equation (5),

 1.772 𝑇 − 0.7117 𝑇 = 382.94 . . . (6)

−0.7117 𝑇 + 0.886 𝑇 = 51.625 . . . (7)

Equation (7) × 2.4898,
−1.772 𝑇 + 2.2059 𝑇 = 128.536 . . . (8)

𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (6)  1.772 𝑇 − 0.7117 𝑇 = 382.94

1.4942 𝑇 = 511.476

 𝑇 = 342.31 𝐾

Substitute T3 value in equation (6),

 1.772(𝑇 ) − 0.7117 × 342.31 = 382.94

 𝑇 = 353.59 𝐾
5.26 Application to Heat Transfer and Dynamic Analysis

Result: Temperature distribution:

T1 = 393 K
T2 = 353.59 K
T3 = 342.31 K
Verification: Since the length of the fin is 50 mm, it is treated as short fin. Assume end
is insulated.
Temperature distribution for short fin, end insulated is given by
𝑇−𝑇 cos ℎ 𝑚 [𝑙 − 𝑙]
(9) =
𝑇 −𝑇 cos ℎ ( 𝑚 𝑙)

𝑇−𝑇 1
= … (10)
𝑇 −𝑇 cos ℎ ( 𝑚 𝑙)

ℎ 𝑃
where 𝑚=
𝑘 𝐴

P = Perimeter = 2 × l (Approximately)

= 2 × 0.050 = 0.1 𝑚

A = Area = Length × Thickness

= 0.050 × 0.007

A = 35 × 10-4 m2

ℎ 𝑃
𝑚=
𝑘 𝐴

140 × 0.1
=
55 × 3.5 × 10

𝑚 = 26.96 𝑚
 Finite Element Analysis 5.27
𝑇−𝑇 1
(10) =
𝑇 −𝑇 cos ℎ ( 26.96 × 0.050)

𝑇 − 295 1
 = [∵ Base Temperature 𝑇 = 𝑇 = 393 𝐾]
393 − 295 2.05

𝑇 − 295 = 47.8

𝑇 = 342.8 𝐾
Temperature at the end of the fin.
𝑇 = 𝑇 = 342.8 𝐾
(ii) Temperature at the middle of the fin:
1
Put 𝑥 = in equation (9),
2
𝑙
𝑇−𝑇 cos ℎ 𝑚 [𝑙 − 2]
(9) =
𝑇 −𝑇 cos ℎ (𝑚 𝑙)
0.050
𝑇−𝑇 cos ℎ 26.96[0.050 − 2 ]
= [∵ 𝑇 = 𝑇 ]
𝑇 −𝑇 cos ℎ (26.96 × 0.050)
𝑇 − 295 1.234
 =
393 − 295 2.049
𝑇 − 295
= 0.6025
393 − 295
𝑇 = 354.04 𝐾
Temperature at the middle of the fin,
𝑇 / 𝑇 = 354.04 𝐾

Example 5.4
A steel rod of diameter d = 2 cm, length L = 5 cm and thermal conductivity k=
50 W/m °C is exposed at one end to a constant temperature of 320°C. The other end is
in ambient air of temperature 20°C with a convection coefficient of h=100 W/m 2 0C.
Determine the temperature at the midpoint of the rod.
5.28 Application to Heat Transfer and Dynamic Analysis

Given:
Diameter, d = 2 cm = 0.02 m
Length, l = 5 mm = 0.50 m
Thermal conductivity, k = 50 W/m0C
One end temperature, 𝑇 = 320 𝐶 + 273 = 593 𝐾
Thermal air conductivity, 𝑇 = 20 𝐶 + 273 = 293 𝐾
Convection coefficient, h = 100 W/m2 oC

To find: Temperature at the mid point of the rod (𝑇 ).

Solution: Discretize the rod into two equal size element.

For element 1: (Nodes 1, 2)

We know that, Finite element equation is

𝐴 𝑘 1 −1 ℎ𝑃𝑙 2 1 𝑇 𝑄𝐴𝑙 +𝑃ℎ𝑇 𝑙 1
+ = … (1)
𝑙 −1 1 6 1 2 𝑇 2 1
where P = Perimeter = π d = π × 0.02
𝑃 = 0.0628 𝑚
 Finite Element Analysis 5.29
𝜋
𝐴 = 𝐴𝑟𝑒𝑎 = 𝑑
4
𝜋
= (0.02)
4
𝐴 = 3.14 × 10 𝑚
𝑄𝐴𝑙
Heat generation Q is not given. So, neglect that term,
2
𝐴 𝑘 1 −1 ℎ𝑃𝑙 2 1 𝑇 𝑃ℎ𝑇 𝑙 1
(1) + =
𝑙 −1 1 6 1 2 𝑇 2 1
Substitute A, P, k, h, l1 and T values,
3.14 × 10 × 50 1 −1 100 × 0.0628 × 0.025 2 1 𝑇
(1) + ×
0.025 −1 1 6 1 2 𝑇
0.0628 × 100 × 293 × 0.025 1
=
2 1
1 −1 2 1 𝑇 1
 0.628 + 0.0262 = 23
−1 1 1 2 𝑇 1
0.628 −0.628 0.0524 0.0262 𝑇 23
 + =
−0.628 0.628 0.0262 0.0524 𝑇 23
1 2
0.6804 −0.6018 𝟏 𝑇 23 𝟏
 = … (2)
−0.6018 0.6804 𝟐 𝑇 235 𝟐
For element 2: (Nodes 2, 3)

Since all the parameters in elements (1) and element (2) are same, the finite
element equation becomes
2 3
0.6804 −0.6018 𝟐 𝑇 23 𝟐
 = … (3)
−0.6018 0.6804 𝟑 𝑇 23 𝟑
5.30 Application to Heat Transfer and Dynamic Analysis

Assemble the finite elements (2) and (3).

1 2 3
0.6804 −0.6018 0 𝟏 𝑇 23 𝟏
−0.6018 0.6804 + 0.6804 −0.6018 𝟐 𝑇 = 23 + 23 𝟐
0 −0.6018 0.6804 𝟑 𝑇 23 𝟑

0.6804 −0.6018 0 𝑇 23
−0.6018 1.3608 −0.6018 𝑇 46
0 −0.6018 0.6804 𝑇 = 23 … (4)
↓ ↓ ↓
[𝐾] [𝑇] [ 𝐹]
To solve the above equation, the following steps to be followed.
Step 1: The first row and first column of the stiffness matrix [ K] have been set equal to
0 except for the main diagonal, which has been set equal to 1.
1 0 0 𝑇 23
 0 1.3608 −0.6018 𝑇 = 46
0 −0.6018 0.6804 𝑇 23
Step 2: The first row of the force matrix is replaced by the known temperature at node 1,
i.e., T1.
1 0 0 𝑇 593
 0 1.3608 −0.6018 𝑇 = 46
0 −0.6018 0.6804 𝑇 23
Step 3: The second row, first column of stiffness matrix [ K ] value (From equation no.4)
is multiplied by known temperature at node 1, i.e., -0.7117 × 393 = - 279.69. This value
(as positive digit, i.e., 356.867) has been added to the second row of the force matrix.
1 0 0 𝑇 593
 0 1.3608 −0.6018 𝑇 = 146 + 356.867
0 −0.6018 0.6804 𝑇 23
1 0 0 𝑇 593
 0 1.3608 −0.6018 𝑇 = 402.867 … (5)
0 −0.6018 0.6804 𝑇 23
Solving equation (5),
 1.3608 𝑇 − 0.6018 𝑇 = 402.867 … (6)
 Finite Element Analysis 5.31

−0.6018 𝑇 + 0.6804 𝑇 = 23 . . . (7)

Equation (6) × 1.1306,

1.5385 𝑇 − 0.6804 𝑇 = 455.485 . . (8)

𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (7)  − 0.6018 𝑇 + 0.6804 𝑇 = 23

0.9367 𝑇 = 478.485

 𝑇 = 510.819 𝐾

Result: Temperature at the midpoint of the rod, 𝑇 = 510.819 𝐾

Example 5.5
Calculate the temperature distribution in a one dimension fin with physical
properties given in Fig.(i). The fin is rectangular in shape and is 120 mm long. 40 mm
wide and 10 mm thick. Assume that convection heat loss occurs from the end of the fin.
Use two elements. Take k = 0.3 W/mm°C; h=1×10 -3 W/mm²°C, T∞=20°C

Fig. (i)
Given data:
Length, l = 120 mm = 0.050 m
Wide, W = 40 mm = 0.040 m
Thickness, t = 10 mm = 0.010 m
Thermal conductivity, K = 0.3 W/mmoC
= 0.3 × 10 𝑊 ⁄𝑚 𝐶 = 300 𝑊 ⁄𝑚 𝐶
Heat transfer coefficient, ℎ = 1 × 10 𝑊 ⁄𝑚𝑚 𝐶
= 1 × 10 × 10 𝑊 ⁄𝑚 𝐶
ℎ = 1000 𝑊 ⁄𝑚 𝐶
5.32 Application to Heat Transfer and Dynamic Analysis

Ambient temperature, T = 200C + 273 = 293 K

One end temperature, T1 = 1200C + 273 = 393 K
To find: Temperature distribution, (𝑇 , 𝑇 𝑎𝑛𝑑 𝑇 )
Solution: Discretize the fin into two equal size element.

For element 1: (Nodes 1, 2):

We know that, Finite element equation is

𝐴 𝑘 1 −1 ℎ𝑃𝑙 2 1 𝑇 𝑄𝐴𝑙 +𝑃ℎ𝑇 𝑙 1
+ = … (1)
𝑙 −1 1 6 1 2 𝑇 2 1

where P = Perimeter = 2 (W + t) (for rectangular)

= 2 [ 0.040 + 0.010]

𝑃 = 0.1 𝑚

𝐴 = 𝐴𝑟𝑒𝑎 = 𝑊𝑖𝑑𝑡ℎ × 𝑇ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠

= 0.040 × 0.010

𝐴 = 4 × 10 𝑚
𝑄𝐴𝑙
Heat generation Q is not given. So, neglect that term,
2
𝐴 𝑘 1 −1 ℎ𝑃𝑙 2 1 𝑇 𝑃ℎ𝑇 𝑙 1
(1) + =
𝑙 −1 1 6 1 2 𝑇 2 1

Substitute A, P, k, h, l1 and T values in equation (1).

 Finite Element Analysis 5.33

4 × 10 × 300 1 −1 1000 × 0.1 × 0.060 2 1 𝑇

 +
0.060 −1 1 6 1 2 𝑇
0.1 × 1000 × 293 × 0.060 1
=
2 1
2 −2 2 1 𝑇 879
 + =
−2 2 1 2 𝑇 879
1 2
4 −1 𝟏 𝑇 879 𝟏
 = … (2)
−1 4 𝟐 𝑇 879 𝟐
For element 2: (Nodes 2, 3)

Since all the parameters in elements (1) and element (2) are same, the finite
element equation becomes
2 3
4 −1 𝟐 𝑇 879 𝟐
 = … (3)
−1 4 𝟑 𝑇 879 𝟑
Assemble the finite elements (2) and (3).
1 2 3
4 −1 0 𝟏 𝑇 879 𝟏
−1 4 + 4 −1 𝟐 𝑇 = 879 + 879 𝟐
0 −1 4 𝟑 𝑇 879 𝟑
4 −1 0 𝑇 879
−1 4 + 4 −1 𝑇 1758
0 −1 4 𝑇 = 879 … (4)
↓ ↓ ↓
[𝐾] [𝑇] [ 𝐹]
To solve the above equation, the following steps to be followed.
Step 1: The first row and first column of the stiffness matrix [K] have been set equal to
0 except for the main diagonal, which has been set equal to 1.
5.34 Application to Heat Transfer and Dynamic Analysis

1 0 0 𝑇 879
 0 8 −1 𝑇 = 1758
0 −1 4 𝑇 879
Step 2: The first row of the force matrix is replaced by the known temperature at node 1,
i.e., T1.
1 0 0 𝑇 393
 0 8 −1 𝑇 = 1758
0 −1 4 𝑇 879
Step 3: The second row, first column of stiffness matrix [ K ] value (From equation no.4)
is multiplied by known temperature at node 1, i.e., -1 × 393 = - 393. This value (as positive
digit, i.e., 393) has been added to the second row of the force matrix.
1 0 0 𝑇 393
 0 8 −1 𝑇 = 393 + 1758
0 −1 4 𝑇 879

1 0 0 𝑇 393
 0 8 −1 𝑇 = 2151 … (5)
0 −1 4 𝑇 879
Solving equation (5),
 8 𝑇 − 𝑇 = 2151 . . . (6)

− 𝑇 + 4 𝑇 = 879 . . (7)
Equation (7) × 8,
 − 8 𝑇 + 32 𝑇 = 7032 . . (8)

𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (6)  8 𝑇 − 𝑇 = 2151

31 𝑇 = 9183

 𝑇 = 296.226 𝐾
Substitute T3 value in equation (6),
 8 (𝑇 ) − 296.226 = 2151
 𝑇 = 305.90 𝐾
 Finite Element Analysis 5.35

Result: Temperature distribution:

T1 = 393 K
T2 = 305.90 K
T3 = 296.226 K

5.6 HEAT TRANSFER IN 2-DIMENSION (THERMAL PROBLEMS)

5.6.1 Shape Function Derivation for Heat Transfer in 2D Element
We begin this section with the development of the shape function for a basic two
dimensional triangular element. We consider this triangular element because its
derivation is the simplest among the available two dimensional elements.

Fig 5.6 Three noded triangular element

Consider a typical triangular element with nodes 1, 2 and 3 as shown in Fig. 5.16.
Let the nodal displacements be 𝑢 , 𝑢 , 𝑢 , 𝑣 , 𝑣 𝑎𝑛𝑑 𝑣 .
𝑢
⎧𝑢 ⎫
⎪𝑢 ⎪
Displacement {𝑢} = 𝑣
⎨ ⎬
⎪𝑣 ⎪
⎩𝑣 ⎭
Since the triangular element has got two degrees of freedom at each node ( u, v),
the total degrees of freedom is 6. Hence it has 6 generalized coordinates.
𝐿𝑒𝑡 , 𝑢 =∝ +∝ 𝑥 +∝ 𝑦
𝑣 =∝ +∝ 𝑥 +∝ 𝑦
5.36 Application to Heat Transfer and Dynamic Analysis

Where, ∝ , ∝ , ∝ , ∝ , ∝ 𝑎𝑛𝑑 ∝ are global or generalized co-ordinates.

 𝑢 =∝ +∝ 𝑥 +∝ 𝑦

𝑢 =∝ +∝ 𝑥 +∝ 𝑦

𝑢 =∝ +∝ 𝑥 +∝ 𝑦
Write the above equations in matrix form,
𝑢 1 𝑥 𝑦 ∝
𝑢 = 1 𝑥 𝑦 ∝
𝑢 1 𝑥 𝑦 ∝

∝ 1 𝑥 𝑦 𝑢
 ∝ = 1 𝑥 𝑦 𝑢
∝ 1 𝑥 𝑦 𝑢

+ − +
⎡1 𝑥 𝑦 ⎤
⎢− + −⎥
𝐿𝑒𝑡 𝐷 =⎢ 𝑦 ⎥⎥
⎢1 𝑥
⎢+ − +⎥
⎣1 𝑥 𝑦 ⎦

𝐶
We know, 𝐷 =
|𝐷|
Find the co-factors of matrix D.
𝑥 𝑦
𝑐 = + 𝑥 𝑦 = (𝑥 𝑦 − 𝑥 𝑦 )

1 𝑦
𝑐 = − = −(𝑦 − 𝑦 ) = 𝑦 − 𝑦
1 𝑦

1 𝑥
𝑐 = + = (𝑥 − 𝑥 )
1 𝑥
𝑥 𝑦
𝑐 = − 𝑥 𝑦 = −(𝑥 𝑦 − 𝑥 𝑦 ) = 𝑥 𝑦 − 𝑥 𝑦

1 𝑦
𝑐 = + =𝑦 −𝑦
1 𝑦
 Finite Element Analysis 5.37
1 𝑥
𝑐 = − = −(𝑥 − 𝑥 ) = 𝑥 − 𝑥
1 𝑥
𝑥 𝑦
𝑐 = + 𝑥 𝑦 =𝑥 𝑦 −𝑥 𝑦

1 𝑦
𝑐 = − = −(𝑦 − 𝑦 ) = 𝑦 − 𝑦
1 𝑦

1 𝑥
𝑐 = + = 𝑥 − 𝑥
1 𝑥

(𝑥 𝑦 − 𝑥 𝑦 ) (𝑦 − 𝑦 ) (𝑥 − 𝑥 )
 𝐶 = (𝑥 𝑦 − 𝑥 𝑦 ) (𝑦 − 𝑦 ) (𝑥 − 𝑥 )
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑦 − 𝑦 ) (𝑥 − 𝑥 )

(𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 )
 𝐶 = (𝑦 − 𝑦 ) (𝑦 − 𝑦 ) (𝑦 − 𝑦 )
(𝑥 − 𝑥 ) (𝑥 − 𝑥 ) (𝑥 − 𝑥 )

1 𝑥 𝑦
We know that, 𝐷 = 1 𝑥 𝑦
1 𝑥 𝑦

|𝐷| = 1 (𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )

Subsitute C and D values in equaiton

1
 𝐷 =
(𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 )
× 𝑦 −𝑦 𝑦 −𝑦 𝑦 −𝑦
𝑥 −𝑥 𝑥 −𝑥 𝑥 −𝑥

Subsitute D value in equaiton

∝ 1 𝑥 𝑦 𝑢
 ∝ = 1 𝑥 𝑦 𝑢
∝ 1 𝑥 𝑦 𝑢
5.38 Application to Heat Transfer and Dynamic Analysis
∝
1
 ∝ =
∝ (𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )
(𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) 𝑢
× 𝑦 −𝑦 𝑦 −𝑦 𝑦 −𝑦 𝑢
𝑥 −𝑥 𝑥 −𝑥 𝑥 −𝑥 𝑢
The area of the triangle can be expressed as a function of the x, y co-ordinates of
the nodes 1, 2 and 3.

1 1 𝑥 𝑦
 𝐴= 1 𝑥 𝑦
2 1 𝑥 𝑦
1
|𝐴| = [1(𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )]
2
 2 𝐴 = (𝑥 𝑦 − 𝑥 𝑦 ) − 𝑥 (𝑦 − 𝑦 ) + 𝑦 (𝑥 − 𝑥 )
Substitute 2A values in equation
∝
1 (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) (𝑥 𝑦 − 𝑥 𝑦 ) 𝑢
 ∝ = 𝑦 −𝑦 𝑦 −𝑦 𝑦 −𝑦 𝑢
∝ 2𝐴 𝑥 −𝑥 𝑥 −𝑥 𝑥 −𝑥 𝑢
∝ 𝑎 𝑎 𝑎 𝑢
1
 ∝ = 𝑏 𝑏 𝑏 × 𝑢
∝ 2𝐴 𝑐 𝑐 𝑐 𝑢
𝑤ℎ𝑒𝑟𝑒, 𝑎 = 𝑥 𝑦 − 𝑥 𝑦 ; 𝑎 = 𝑥 𝑦 − 𝑥 𝑦 ; 𝑎 = 𝑥 𝑦 − 𝑥 𝑦
𝑏 = 𝑦 − 𝑦 ;𝑏 = 𝑦 − 𝑦 ;𝑏 = 𝑦 − 𝑦
𝑐 = 𝑥 − 𝑥 ;𝑐 = 𝑥 − 𝑥 ;𝑐 = 𝑥 − 𝑥
𝑢 =∝ +∝ +∝ 𝑦
We can write this equation in matrix form.
∝
𝑢 = [ 𝑙 𝑥 𝑦] ∝
∝
∝
Substitute ∝ value,
∝
 Finite Element Analysis 5.39
𝑎 𝑎 𝑎 𝑢
1
 𝑢 =[𝑙𝑥𝑦] × 𝑏 𝑏 𝑏 × 𝑢
2𝐴 𝑐 𝑐 𝑐 𝑢
𝑎 𝑎 𝑎 𝑢
1
= [𝑙𝑥𝑦] 𝑏 𝑏 𝑏 × 𝑢
2𝐴 𝑐 𝑐 𝑐 𝑢
𝑢
1
= [𝑎 + 𝑏 𝑥 + 𝑐 𝑦 𝑎 + 𝑏 𝑥 + 𝑐 𝑦 𝑎 + 𝑏 𝑥 + 𝑐 𝑦] × 𝑢
2𝐴 𝑢
[ ∵ (1 × 3) × (3 × 3) = 1 × 3 ]
𝑢
𝑎 +𝑏 𝑥+ 𝑐 𝑦 𝑎 +𝑏 𝑥+𝑐 𝑦 𝑎 +𝑏 𝑥+𝑐 𝑦
𝑢= 𝑢
2𝐴 2𝐴 2𝐴 𝑢
The above equation is in the form of
𝑢
𝑢 = [𝑁 𝑁 𝑁 ] 𝑢
𝑢
𝑣
Similarly, 𝑣 = [𝑁 𝑁 𝑁 ] 𝑣
𝑣
𝑎 +𝑏 𝑥+ 𝑐 𝑦
Where, shape function, 𝑁 =
2𝐴
𝑎 +𝑏 𝑥+ 𝑐 𝑦
𝑁 =
2𝐴
𝑎 + 𝑏 𝑥 + 3𝑦
𝑁 =
2𝐴
Assemble the equation in matrix form,
𝑢
⎧𝑣 ⎫
𝑢(𝑥, 𝑦) 𝑁 0 𝑁 0 𝑁 0 ⎪𝑢 ⎪
Displacement funciton, 𝑢 = = × 𝑣
𝑣(𝑥, 𝑦) 0 𝑁 0 𝑁 0 𝑁 ⎨ ⎬
⎪𝑢 ⎪
⎩𝑣 ⎭
5.40 Application to Heat Transfer and Dynamic Analysis

5.7 STIFFNESS MATRIX AND LOAD VECTOR FOR HEAT

TRANSFER IN TWO-DIMENSIONAL ELEMENT

Fig. 5.7 Triangular element with nodal temperature

Triangular element is the basic element for solution of two-dimensional heat
transfer problems. Consider the three-noded triangular element with nodal temperatures
T1, T2 and T3 as shown in Fig 5.7.
The temperature function is given by,
𝑇(𝑥, 𝑦) = 𝑁 𝑇 + 𝑁 𝑇 + 𝑁 𝑇
We know that,
𝑎 +𝑏 𝑥+ 𝑐 𝑦
shape functions, 𝑁 = ⎫
2𝐴 ⎪
𝑎 +𝑏 𝑥+ 𝑐 𝑦
𝑁 =
2𝐴 ⎬
𝑎 + 𝑏 𝑥 + 𝑐 𝑦⎪
𝑁 = ⎭
2𝐴
We know that,

Stiffness matrix [K ] = [𝐵] [𝐷] [𝐵] 𝑑𝑣

We know that
𝜕𝑁 𝜕𝑁 𝜕𝑁
⎡ ⎤
𝜕𝑥 𝜕𝑥 𝜕𝑥 ⎥
Strain − Displacement matrix, [𝐵] = ⎢𝜕𝑁 𝜕𝑁 𝜕𝑁 ⎥
⎢
⎣ 𝜕𝑦 𝜕𝑦 𝜕𝑦 ⎦
 Finite Element Analysis 5.41

By partial differentiation,
𝜕𝑁 𝑏 𝜕𝑁 𝑐
= = ⎫
𝜕𝑥 2𝐴 𝜕𝑥 2𝐴⎪
𝜕𝑁 𝑏 𝜕𝑁 𝑐
= =
𝜕𝑥 2𝐴 𝜕𝑥 2𝐴⎬
𝜕𝑁 𝑏 𝜕𝑁 𝑐 ⎪
= = ⎭
𝜕𝑥 2𝐴 𝜕𝑥 2𝐴
Substitute the equation
𝑏 𝑏 𝑏
[𝐵] = 2𝐴 2𝐴 2𝐴
𝑐 𝑐 𝑐
2𝐴 2𝐴 2𝐴
1 𝑏 𝑏 𝑏
[𝐵] =
2𝐴 𝑐 𝑐 𝑐

1 𝑏 𝑐
[𝐵] = 𝑏 𝑐
2𝐴
𝑏 𝑐
We know that,
𝑘 0
Stress − strain matrix, [𝐷] =
0 𝑘
Assuming a unit thickness, the elements volume can be expressed as dv = Da

1 𝑏 𝑐
𝑘 0 1 𝑏 𝑏 𝑏
 [𝐾 ] = 𝑏 𝑐
0 𝑘
× 𝑑𝐴
2𝐴 2𝐴 𝑐 𝑐 𝑐
𝑏 𝑐
𝑏 𝑐
1 𝑘 0 𝑏 𝑏 𝑏
= 𝑏 𝑐 𝑑𝐴
4𝐴 0 𝑘 𝑐 𝑐 𝑐
𝑏 𝑐
𝑏 𝑐
1 𝑘 0 𝑏 𝑏 𝑏
= 𝑏 𝑐 𝑑𝐴
4𝐴 0 𝑘 𝑐 𝑐 𝑐
𝑏 𝑐

1 𝑏 𝑐
𝑘 0 𝑏 𝑏 𝑏
= 𝑏 𝑐
4𝐴 0 𝑘 𝑐 𝑐 𝑐
𝑏 𝑐
5.42 Application to Heat Transfer and Dynamic Analysis
𝑏 𝑘 +0 0+𝑐 𝑘
1 𝑏 𝑏 𝑏
= 𝑏 𝑘 +0 0+𝑐 𝑘
4𝐴 𝑐 𝑐 𝑐
𝑏 𝑘 +0 0+𝑐 𝑘

(𝑏 𝑘 + 𝑐 𝑘 ) (𝑏 𝑏 𝑘 + 𝑐 𝑐 𝑘 ) (𝑏 𝑏 𝑘 + 𝑐 𝑐 𝑘 )
1
[𝐾 ] = (𝑏 𝑏 𝑘 + 𝑐 𝑐 𝑘 ) (𝑏 𝑘 + 𝑐 𝑘 ) (𝑏 𝑏 𝑘 + 𝑐 𝑐 𝑘 )
4𝐴
(𝑏 𝑏 𝑘 + 𝑐 𝑐 𝑘 ) (𝑏 𝑏 𝑘 + 𝑐 𝑐 𝑘 ) (𝑏 𝑘 + 𝑐 𝑘 )

For an isotropic material with 𝑘 = 𝑘 = 𝑘 ,

Stiffness matrix for conduction,
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
1
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
To determine the stiffness matrix for convection,

[𝐾 ] = ℎ [ 𝑁 ] [ 𝑁 ] 𝑑𝑠

𝑁
=ℎ 𝑁 [ 𝑁 𝑁 𝑁 ] 𝑑𝑠
𝑁
𝑁 𝑁𝑁 𝑁𝑁
=ℎ 𝑁𝑁 𝑁 𝑁𝑁 𝑑𝑠
𝑁𝑁 𝑁𝑁 𝑁
Let the edge 1-2 of element lies on the boundary as shown in Fig. 5.8. So that N3
= 0 along this edge.

Fig. 5.8 Heat loss by convection from sides 1-2

 Finite Element Analysis 5.43

Substitute N3 = 0 in equation

𝑁 𝑁𝑁 0
[𝐾 ]ℎ = 𝑁𝑁 𝑁 0 𝑑𝑠
0 0 0

subsitute N = L , N = L and N = l , along the edge 1 − 2, N = l = 3.

𝐿 𝐿 𝐿 0
Hence,  [𝐾 ]= ℎ 𝐿 𝐿 𝐿 0 𝑑𝑠
0 0 0

Where, s- denotes the direction along the edge 1-2.

𝛼! 𝛽!
We know that 𝐿∝ 𝐿 𝑑𝑠 = 𝑠
(𝛼 + 𝛽 + 1)!

2!
Therefore, 𝐿 𝑑𝑠 = 𝑠
(2 + 1)!

1×2
= 𝑠
1×2 ×3
𝑠
𝐿 𝑑𝑠 =
3

1! 1!
Similarly, 𝐿 𝐿 𝑑𝑠 = 𝑠
(1 + 1 + 1)!

1
= 𝑠
1×2 ×3
𝑠
𝐿 𝐿 𝑑𝑠 =
6

2!
Similarly, 𝐿 𝑑𝑠 = 𝑠
(2 + 1)!

1 ×2
= 𝑠
1×2 ×3
5.44 Application to Heat Transfer and Dynamic Analysis
𝑠
𝐿 𝑑𝑠 =
3
𝑠 𝑠
⎡ 3 0⎤
⎢𝑠 6 ⎥
 [𝐾 ] =ℎ ⎢ 𝑠
0⎥
⎢ 6 3 ⎥
⎣ 0 0 0⎦
[Direction along the edge (1-2)]
𝑠 𝑠
⎡ 3 0⎤
ℎ 𝑠 ⎢𝑠 6 ⎥
[𝐾 ] = ⎢ 𝑠
6 0⎥
⎢ 6 3 ⎥
⎣ 0 0 0⎦
Now, consider the edge 2-3 of element lies on the boundary
Hence, N = L = 0, N = L , N = L .
Substitute the 𝑁 , 𝑁 𝑎𝑛𝑑 𝑁 values in equation we get
0 0 0
[𝐾 ] =ℎ 0 𝐿 𝐿 𝐿 𝑑𝑠
0 𝐿 𝐿 𝐿
0 0 0
⎡ 𝑠 𝑠 ⎤
[𝐾 ] ⎢0 ⎥
=ℎ ⎢ 3 6 ⎥
⎢0 𝑠 𝑠 ⎥
⎣ 6 3 ⎦
[Direction along the edge (2-3)]

ℎ 𝑠 0 0 0
= 0 2 1
6
0 1 2
Similarly, let the edge 3 -1 of elements lies on the boundary.
Hence, N = L , N = L , N = L .
Substitute the 𝑁 , 𝑁 𝑎𝑛𝑑 𝑁 values in equation
𝐿 0 𝐿 𝐿
[𝐾 ] =ℎ 0 0 0 𝑑𝑠
𝐿 𝐿 0 𝐿
 Finite Element Analysis 5.45
𝑠 𝑠
⎡ 3 0
⎢ 6 ⎤⎥
[𝐾 ] =ℎ ⎢ 0 0 0 ⎥
⎢𝑠 0
𝑠 ⎥
⎣ 6 3 ⎦
ℎ 𝑠 2 0 1
[𝐾 ] = 0 0 0
6
1 0 2
Stiffness matrix for convection,
[ 𝐾 ] = [𝐾 ] + [𝐾 ] + [𝐾 ]

ℎ 𝑠 2 1 0 ℎ 𝑠 0 0 0
[𝐾 ]= 1 2 0 + 0 2 1
6 6
0 0 0 0 1 2
ℎ 𝑠 2 0 1
+ 0 0 0
6
1 0 2
Stiffness matrix for 2 –dimensional heat transfer element is given by,
[𝐾]= [𝐾 ]+[𝐾 ]

(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾]= (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 +𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 +𝑐 𝑐 ) (𝑏 + 𝑐 )
ℎ 𝑠 2 1 0 ℎ 𝑠 0 0 0
+ 1 2 0 + 0 2 1
6 6
0 0 0 0 1 2
ℎ 𝑠 2 0 1
+ 0 0 0
6
1 0 2

5.8 FORCE VECTOR OR LOAD VECTOR, (F)

The force vector for 2 – dimensional heat transfer element is given by
𝑁
[𝐹 ] = 𝑞 [ 𝑁] 𝑑𝑣 = 𝑞 𝑁 𝑑𝐴
𝑁
5.46 Application to Heat Transfer and Dynamic Analysis
𝐿
[𝐹 ] = 𝑞 𝐿 𝑑𝐴
𝐿
𝐿
= 𝑞 𝐿 𝑑𝐴 [∵ 𝑑𝑣 = 𝑑𝐴 (𝑢𝑛𝑖𝑡 𝑡ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠)]
𝐿
By using area co- ordinates system,
𝛼! 𝛽! 𝛾!
𝐿∝ 𝐿 𝐿 𝑑𝐴 = × 2𝐴
(𝛼 + 𝛽 + 𝛾 + 2)!
1! 1
We know that, 𝐿 𝑑𝐴 = × 2𝐴 = × 2𝐴
(1 + 2)! 1×2 ×2
𝐴
𝐿 𝑑𝐴 =
3
1! 1
Similarly, 𝐿 𝑑𝐴 = × 2𝐴 = × 2𝐴
(1 + 2)! 1×2 ×3
𝐴
𝐿 𝑑𝐴 =
3
1! 1
Similarly, 𝐿 𝑑𝐴 = × 2𝐴 = × 2𝐴
(1 + 2)! 1×2 ×3
𝐴
𝐿 𝑑𝐴 =
3
Substitute the equations and values in equation

𝑞 𝐴 1
{𝐹 } = 1
3
1

𝑆𝑖𝑚𝑖𝑙𝑎𝑟𝑙𝑦, {𝐹 } = 𝑞 [𝑁] 𝑑𝑠

𝑁
= 𝑞 𝑁 𝑑𝑠
𝑁
If the edge 1-2 lies on s2,
 Finite Element Analysis 5.47

subsitute N = L , N = L and N = l , along the edge 1 − 2, N = L = 0.

𝐿
{𝐹 } = 𝑞 𝐿 𝑑𝑠
𝐿

By using surface edges,

𝛼! 𝛽!
𝐿∝ 𝐿 𝑑𝑠 = 𝑠
(𝛼 + 𝛽 + 1)!
We know that,
1! 1 𝑠
𝐿 𝑑𝑠 = 𝑠= 𝑠=
(1 + 1)! 2! 2

1! 𝑠
𝑠𝑖𝑚𝑖𝑙𝑎𝑟𝑙𝑦, 𝐿 𝑑𝑠 = 𝑠=
(1 + 1)! 2

𝑞 𝑠 1
{𝐹 } = 1
2
0

Similarly, the vector {F3} can be obtained as,

{𝐹 } = ℎ 𝑡 [𝑁] 𝑑𝑠

If the edge 1-2 lies on s3,

subsitute N = L , N = L and N = L , along the edge 1 − 2 𝑁 = 𝐿 = 0.
𝑁
{𝐹 } = ℎ 𝑇 𝑁 𝑑𝑠
𝑁

𝐿
=ℎ𝑇 𝐿 𝑑𝑠
𝐿

𝐿
=ℎ𝑇 𝐿 𝑑𝑠
0
5.48 Application to Heat Transfer and Dynamic Analysis

ℎ 𝑇 𝑠 1
{𝐹 } = 1
2
0

5.9. SOLVED PROBLEMS [TWO DIMENSIONAL]

Example 5.6
Find the temperature distribution in a square region with uniform
energy generation as shown in Fig.(i). Assume that there is no temperature variation in
z-direction. Take k=30 W/cm °C, length = 10 cm, T=50°C, q= 100 W/cm³.
Given:

Fig. (i)
k = 30 W /CmoC; l = 10 cm; T = 50Oc; q = 100W/cm3
To find: Temperature distribution
Solution:

Fig. (ii)
 Finite Element Analysis 5.49

Element (1):
1 𝑥 𝑦
𝐴𝑟𝑒𝑎 𝑜𝑓 𝑡𝑟𝑖𝑎𝑛𝑔𝑙𝑒 𝐴, = 1 𝑥 𝑦
1 𝑥 𝑦

1 1 0 0
= 1 5 0
2
1 0 5
= 12.5 cm2
Length of edge 2 -3,

𝐿 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (5 − 0) + (0 − 5)
= 7.07 cm
Stiffness matrix of element (1),
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
Where,
𝑏 = (𝑦 − 𝑦 ) = 0 − 5 = −5
𝑏 = (𝑦 − 𝑦 ) = 5 − 0 = 5
𝑏 = (𝑦 − 𝑦 ) = 0 − 0 = 0
𝑐 = (𝑥 − 𝑥 ) = 0 − 5 = −5
𝑐 = (𝑥 − 𝑥 ) = 0 − 0 = 0
𝑐 = (𝑥 − 𝑥 ) = 5 − 0 = 5
5.50 Application to Heat Transfer and Dynamic Analysis

Substituting the above values in [K1],

30 50 −25 −25
[𝐾 ] = −25 25 0
4 × 12.5
−25 0 25
1 2 4
30 −15 −15 𝟏
[𝐾 ] = −15 15 0 𝟐
−15 0 15 𝟒
Element (2):

1 1 𝑥 𝑦
Area of triangle 𝐴, = 1 𝑥 𝑦
2 1 𝑥 𝑦

1 1 0 5
= 1 5 0
2
1 5 5
A = 12.5 cm2
Length of edge 1 -2,

𝐿 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (5 − 0) + (0 − 5)
= 7.07 cm
Stiffness matrix of element (2),
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
Where,
𝑏 = (𝑦 − 𝑦 ) = 0 − 5 = −5
𝑏 = (𝑦 − 𝑦 ) = 5 − 5 = 0
𝑏 = (𝑦 − 𝑦 ) = 5 − 0 = 5
𝑐 = (𝑥 − 𝑥 ) = 5 − 5 = 0
 Finite Element Analysis 5.51

𝑐 = (𝑥 − 𝑥 ) = 0 − 5 = −5
𝑐 = (𝑥 − 𝑥 ) = 5 − 0 = 5
Substituting the above values in [K2],
4 2 5
15 0 −15 𝟒
[𝐾 ] = 0 15 −15 𝟐
−15 −15 30 𝟓
Element (3):

1 1 x y
Area of triangle A, = 1 x y
2 1 x y

1 1 5 0
= 1 10 0
2
1 5 5
= 12.5 cm2
Length of edge (2 -3),

L = (x − x ) + (y − y )

= (5 − 10) + (5 − 0)
= 7.07 cm
Stiffness matrix of element (3),
(b + c ) (b b + c c ) (b b + c c )
k
[K ] = (b b + c c ) (b + c ) (b b + c c )
4A
(b b + c c ) (b b + c c ) (b + c )
Where,
b = (y − y ) = 0 − 5 = −5
b = (y − y ) = 5 − 0 = 5
b = (y − y ) = 0 − 0 = 0
c = (x − x ) = 5 − 10 = −5
c = (x − x ) = 5 − 5 = 0
5.52 Application to Heat Transfer and Dynamic Analysis

c = (x − x ) = 10 − 5 = 5
Substituting the above values in [K3],

30 50 −25 −25
[K ] = −25 25 0
4 × 12.5
−25 0 25
2 3 5
30 −15 −15 𝟐
[K ] = −15 15 0 𝟑
−15 0 15 𝟓
Element (4):

1 1 𝑥 𝑦
Area of triangle 𝐴, = 1 𝑥 𝑦
2 1 𝑥 𝑦

1 1 5 5
= 1 10 0
2
1 10 5
= 12.5 cm2
Length of edge (1 -2),

𝐿 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (10 − 5) + (0 − 5)
= 7.07 cm
Stiffness matrix of element (4),
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
Where,
𝑏 = (𝑦 − 𝑦 ) = 0 − 5 = −5
𝑏 = (𝑦 − 𝑦 ) = 5 − 5 = 0
𝑏 = (𝑦 − 𝑦 ) = 5 − 0 = 5
𝑐 = (𝑥 − 𝑥 ) = 10 − 10 = 0
 Finite Element Analysis 5.53

𝑐 = (𝑥 − 𝑥 ) = 5 − 10 = −5
𝑐 = (𝑥 − 𝑥 ) = 10 − 5 = 5
Substituting the above values in [K4],
5 3 6
15 0 −15 𝟓
[𝐾 ] = 0 15 −15 𝟑
−15 −15 30 𝟔
Element (5):

1 1 𝑥 𝑦
Area of triangle 𝐴, = 1 𝑥 𝑦
2 1 𝑥 𝑦

1 1 0 5
= 1 5 5
2
1 0 10

= 12.5 cm2
Length of edge (2 -3),

𝐿 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (0 − 5) + (10 − 5)

= 7.07 cm
Stiffness matrix of element (5),
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
Where,
𝑏 = (𝑦 − 𝑦 ) = 5 − 10 = −5
𝑏 = (𝑦 − 𝑦 ) = 10 − 5 = 5
𝑏 = (𝑦 − 𝑦 ) = 5 − 5 = 0
𝑐 = (𝑥 − 𝑥 ) = 0 − 5 = −5
5.54 Application to Heat Transfer and Dynamic Analysis

𝑐 = (𝑥 − 𝑥 ) = 0 − 0 = 0
𝑐 = (𝑥 − 𝑥 ) = 5 − 0 = 5
Substituting the above values in [K5],
4 5 7
30 −15 −15 𝟒
[𝐾 ] = −15 15 0 𝟓
−15 0 15 𝟕
Element (6):

1 1 𝑥 𝑦
Area of triangle 𝐴, = 1 𝑥 𝑦
2 1 𝑥 𝑦

1 1 0 10
= 1 5 5
2
1 5 10
= 12.5 cm2
Length of edge (1 -2),

𝐿 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (5 − 0) + (5 − 10)
= 7.07 cm
Stiffness matrix of element (6),
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
Where,
𝑏 = (𝑦 − 𝑦 ) = 5 − 10 = −5
𝑏 = (𝑦 − 𝑦 ) = 10 − 10 = 0
𝑏 = (𝑦 − 𝑦 ) = 10 − 5 = 5
𝑐 = (𝑥 − 𝑥 ) = 5 − 5 = 0
𝑐 = (𝑥 − 𝑥 ) = 0 − 5 = −5
 Finite Element Analysis 5.55

𝑐 = (𝑥 − 𝑥 ) = 5 − 0 = 5
Substituting the above values in [K6],
7 5 8
15 0 −15 𝟕
[𝐾 ] = 0 15 −15 𝟓
−15 −15 30 𝟖
Element (7):

1 1 𝑥 𝑦
Area of triangle 𝐴, = 1 𝑥 𝑦
2 1 𝑥 𝑦

1 1 5 5
= 1 10 5
2
1 5 10
= 12.5 cm2
Length of edge (2 -3),

𝐿 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (5 − 10) + (10 − 5)
= 7.07 cm
Stiffness matrix of element (7),
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
Where,
𝑏 = (𝑦 − 𝑦 ) = 5 − 10 = −5
𝑏 = (𝑦 − 𝑦 ) = 10 − 5 = 5
𝑏 = (𝑦 − 𝑦 ) = 5 − 5 = 0
𝑐 = (𝑥 − 𝑥 ) = 5 − 10 = −5
𝑐 = (𝑥 − 𝑥 ) = 5 − 5 = 0
𝑐 = (𝑥 − 𝑥 ) = 10 − 5 = 5
5.56 Application to Heat Transfer and Dynamic Analysis

Substituting the above values in [K7],

5 6 8
30 −15 −15 𝟓
[𝐾 ] = −15 15 0 𝟔
−15 0 15 𝟖
Element (8):
1 𝑥 𝑦
1
Area of triangle 𝐴, = 1 𝑥 𝑦
2 1 𝑥 𝑦

1 1 5 10
= 1 10 5
2
1 10 10
= 12.5 cm2
Length of edge (1 -2),

𝐿 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (10 − 5) + (5 − 10)
= 7.07 cm
Stiffness matrix of element (8),
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
Where,
𝑏 = (𝑦 − 𝑦 ) = 5 − 10 = −5
𝑏 = (𝑦 − 𝑦 ) = 10 − 10 = 0
𝑏 = (𝑦 − 𝑦 ) = 10 − 5 = 05
𝑐 = (𝑥 − 𝑥 ) = 10 − 10 = 0
𝑐 = (𝑥 − 𝑥 ) = 5 − 10 = −5
𝑐 = (𝑥 − 𝑥 ) = 10 − 5 = 5
 Finite Element Analysis 5.57

Substituting the above values in [K8],

8 6 9
15 0 −15 𝟖
[𝐾 ] = 0 15 −15 𝟔
−15 −15 30 𝟗
Assemble the stiffness matrix,
[K]=
1 2 3 4 5 6 7 8 9
30 -15 0 -15 0 0 0 0 0 1
15 +
-15 15 + -15 0 -15 -15 0 0 0 0 2
30
15 +
0 -15 0 0+0 -15 0 0 0 3
15
15 +
-15 0 0 15 + -15- 15 0 -15 0 0 4
30
15+15+
-15 - -15 - -15-
0 0+0 15+30+ 0+0 -15 - 15 0 5
15 15 15
30+15
30 +
-15 -
0 0 -15 0 15 0 0+0 -15 6
15
+15
15
0 0 0 -15 0+0 0 -15 0 7
+15
-15 - 30 + 15 +
0 0 0 0 0+0 -15 -15 8
15 15
0 0 0 0 0 -15 0 - 15 30 9
5.58 Application to Heat Transfer and Dynamic Analysis

1 2 3 4 5 6 7 8 9
30 −15 0 −15 0 0 0 0 0 𝟏
⎡−15 60 −15 0 −30 0 0 0 0 ⎤𝟐
⎢ 0 −15 30 0 0 −15 0 0 0 ⎥𝟑
⎢ ⎥
⎢ −15 0 0 60 −30 0 −15 0 0 ⎥𝟒
[ 𝐾] = ⎢ 0 −30 0 −30 120 −30 0 −30 0 ⎥𝟓
⎢ 0 0 −15 0 −30 60 0 0 −15⎥ 𝟔
⎢ 0 0 0 −15 0 0 30 −15 0 ⎥𝟕
⎢ 0 0 0 0 −30 0 −15 60 −15⎥ 𝟖
⎣ 0 0 0 0 0 −15 0 −15 30 ⎦ 𝟗
Load vector for element (1),

𝑞𝐴 1
{ 𝐹} = 1
2
1
𝐹 𝑞𝐴 1
{ 𝐹} = 𝐹 = 1
𝐹 3
1
100 × 12.5 1
= 1
3
1
1
𝐹 416.6 𝟏
 { 𝐹} = 𝐹 = 416.6 𝟐
𝐹 416.6 𝟒
Similarly, Load vector for element (2),
𝐹 𝑞𝐴 1
{ 𝐹} = 𝐹 = 1
𝐹 3
1
100 × 12.5 1
= 1
3
1
1
𝐹 416.6 4
𝐹 = 416.6 2
𝐹 416.6 5
 Finite Element Analysis 5.59

Similarly, Load vector for element (3),

𝐹 𝑞𝐴 1
{ 𝐹} = 𝐹 = 1
𝐹 3
1
100 × 12.5 1
= 1
3
1
1
𝐹 416.6 𝟐
𝐹 = 416.6 𝟑
𝐹 416.6 𝟓
Similarly, Load vector for element (4),
𝐹 𝑞𝐴 1
𝐹 = 1
𝐹 3
1
100 × 12.5 1
= 1
3
1
1
𝐹 416.6 𝟓
𝐹 = 416.6 𝟑
𝐹 416.6 𝟔
Similarly, Load vector for element (5),
𝐹 𝑞𝐴 1
𝐹 = 1
𝐹 3
1
100 × 12.5 1
= 1
3
1
1
𝐹 416.6 𝟒
𝐹 = 416.6 𝟓
𝐹 416.6 𝟕
5.60 Application to Heat Transfer and Dynamic Analysis

Similarly, Load vector for element (6),

𝐹 𝑞𝐴 1
𝐹 = 1
𝐹 3
1
100 × 12.5 1
= 1
3
1
1
𝐹 416.6 𝟕
𝐹 = 416.6 𝟓
𝐹 416.6 𝟖
Similarly, Load vector for element (7),
𝐹 𝑞𝐴 1
𝐹 = 1
𝐹 3
1
100 × 12.5 1
= 1
3
1
1
𝐹 416.6 𝟓
𝐹 = 416.6 𝟔
𝐹 416.6 𝟖
Similarly, Load vector for element (8),
𝐹 𝑞𝐴 1
𝐹 = 1
𝐹 3
1
100 × 12.5 1
= 1
3
1
1
𝐹 416.6 𝟖
𝐹 = 416.6 𝟔
𝐹 416.6 𝟗
 Finite Element Analysis 5.61

Assemble the load vector or force matrix,

1
𝐹 416.6 𝟏
⎧𝐹 ⎫ ⎧1249.8⎫
⎪𝐹 ⎪ 𝟐
⎪ ⎪ ⎪ 833.2 ⎪ 𝟑
⎪ ⎪
⎪𝐹 ⎪ ⎪1249.8⎪ 𝟒
𝐹 = 2499.6 𝟓
⎨𝐹 ⎬ ⎨1249.8⎬ 𝟔
⎪𝐹 ⎪ ⎪ 833.2 ⎪ 𝟕
⎪ ⎪ ⎪ ⎪
⎪𝐹 ⎪ ⎪1249.8⎪ 𝟖
⎩𝐹 ⎭ ⎩ 416.8 ⎭ 𝟗

Globl matrix, {F} = [ K ]{ T}

1 2 3 4 5 6 7 8 9
416.6 30 −15 0 −15 0 0 0 0 0 𝟏 ⎧𝑇 ⎫
⎧1249.8⎫ ⎡−15 60 −15 0 −30 0 0 0 0 ⎤ 𝟐 ⎪𝑇
⎢ 0 ⎪
⎪ 833.2 ⎪
⎪ ⎪ −15 30 0 0 −15 0 0 0 ⎥ 𝟑 ⎪𝑇 ⎪
⎢ ⎥
⎪1249.8⎪
⎢ −15 0 0 60 −30 0 −15 0 0 ⎥ 𝟒 ⎪𝑇 ⎪
2499.6 = ⎢ 0 −30 0 −30 120 −30 0 −30 0 ⎥𝟓 𝑇
⎨1249.8⎬ ⎢ 0 0 −15 0 −30 60 0 0 −15⎥ 𝟔 ⎨𝑇 ⎬
⎪ 833.2 ⎪ ⎢ 0 0 0 −15 0 0 30 −15 0 ⎥ 𝟕 ⎪𝑇 ⎪
⎪ ⎪ ⎪ ⎪
⎪1249.8⎪ ⎢ 0 0 0 0 −30 0 −15 60 −15⎥ 𝟖 ⎪𝑇 ⎪
⎩ 416.8 ⎭ ⎣ 0 0 0 0 0 −15 0 −15 30 ⎦ 𝟗 ⎩𝑇 ⎭
Applying boundary conditions,
𝑇 = 𝑇 , 𝑇 = 𝑇 , 𝑇 = 50 𝐶, 𝑇 = 𝑇 , 𝑇 = 𝑇 ,
𝑇 = 50 𝐶, 𝑇 = 50 𝐶, 𝑇 = 50 𝐶, 𝑇 = 50 𝐶.
 𝑇 = 𝑇 = 𝑇 = 𝑇 = 𝑇 = 50 𝐶
1 2 3 4 5 6 7 8 9
416.6 30 −15 0 −15 0 0 0 0 0 𝟏 𝑇
⎡−15 60 ⎧ ⎫
⎧1249.8⎫ −15 0 −30 0 0 0 0 ⎤ 𝟐 ⎪𝑇 ⎪
⎪ 833.2 ⎪ ⎢ 0 −15 30 0 0 −15 0 0 0 ⎥ 𝟑 ⎪50⎪
⎪ ⎪ ⎢ ⎥
⎪1249.8⎪ 60 −15 0 0 ⎥ 𝟒 ⎪𝑇 ⎪
⎢−15 0 0 −30 0
2499.6 = ⎢ 0 −30 0 −30 120 −30 0 −30 0 ⎥𝟓 𝑇
⎨1249.8⎬ ⎢ 0 0 −15 0 −30 60 0 0 −15⎥ 𝟔 ⎨50⎬
⎪ 833.2 ⎪ ⎢ 0 0 0 −15 0 0 30 −15 0 ⎥ 𝟕 ⎪50⎪
⎪ ⎪ ⎪ ⎪
⎪1249.8⎪ ⎢ 0 0 0 0 −30 0 −15 60 −15⎥ 𝟖 ⎪50⎪
⎩ 416.8 ⎭ ⎣ 0 0 0 0 0 −15 0 −15 30 ⎦ 𝟗 ⎩50⎭
5.62 Application to Heat Transfer and Dynamic Analysis

We know that,
 30𝑇 − 15 𝑇 − 15 𝑇 = 416.4 … (1)

 − 15𝑇 + 60 𝑇 − 15 (50) − 30 𝑇 = 1249.8

−15𝑇 + 60 𝑇 − 30𝑇 = 1999.8 … (2)

 − 15𝑇 + 30 (50) − 15(50) = 833.2 … (3)

 − 15𝑇 + 60 𝑇 − 30𝑇 − 15(50) = 1249.8 … (4)

 − 30𝑇 − 30𝑇 + 120 𝑇 − 30(50) − 30(50) = 2499.6 … (5)

 − 15(50) − 30𝑇 + 60(50) − 15(50) = 1249.8 … (6)

 − 15𝑇 + 30(50) − 15(50) = 833.2 … (7)

 − 30𝑇 − 15(50) + 60(50) − 15(50) = 1249.8 … (8)

Equaiton (7) − 15 T = 833.2 + 15 (50) − 30(50)

𝑇 = 119.4 𝐶

Substitute the values T4 and T5 in equation (4),

 − 15𝑇 − 30(119.4) + 120(105.6) − 30(50) − 15(50) = 1249.8

−15𝑇 = 1249.8 + 30(119.4) − 120(105.6) + 30(50) − 15(50)

𝑇 = 133.31 𝐶
Substitute the values T1 and T4 in equation (1),
 30(133.3) − 15 𝑇 − 15 (119.4) = 416.6
𝑇 = 119.42 𝐶
Result: Temperature distribution,
𝑇 = 133.31 𝐶
𝑇 = 119.42 𝐶
𝑇 = 50 𝐶
 Finite Element Analysis 5.63

𝑇 = 119.42 𝐶
𝑇 = 105.6 𝐶
𝑇 = 50 𝐶
𝑇 = 50 𝐶
𝑇 = 50 𝐶
𝑇 = 50 𝐶

Example 5.7
Compute the element matrix and vectors for the element shown in Fig.(i) when
the edges 2-3 and 3-1 experience convection heat loss.

Fig. (i).
Given: k = 60 W/cm K
q0 = 50 W/cm3
(𝑥 , 𝑦 ) = 4, 6
(𝑥 , 𝑦 ) = 8, 10
(𝑥 , 𝑦 ) = 12, 8
To find: (1) Element matrix [ KC ] and [ Kh] (2) force vector
Solution: We know that, Stiffness matrix for conduction,
(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
5.64 Application to Heat Transfer and Dynamic Analysis

Where,

1 1 𝑥 𝑦
Area of triangle 𝐴, = 1 𝑥 𝑦
2 1 𝑥 𝑦

1 1 4 6
= 1 8 10
2
1 12 8
A = 12.5 cm2
Where,
𝑏 = (𝑦 − 𝑦 ) = (10 − 8) = 2
𝑏 = (𝑦 − 𝑦 ) = (8 − 6) = 2
𝑏 = (𝑦 − 𝑦 ) = (6 − 10) = −4
𝑐 = (𝑥 − 𝑥 ) = (12 − 8) = 4
𝑐 = (𝑥 − 𝑥 ) = (4 − 12) = −8
𝑐 = (𝑥 − 𝑥 ) = (8 − 4) = 4
Substituting the above values in [KC],
(4 + 16) (4 − 32) (−8 + 16)
60
[𝐾 ] = (4 − 32) (4 + 64) (−8 − 32)
4 × 12
(−8 + 16) (−8 − 32) (16 + 16)
25 −35 10
[𝐾 ] = −35 85 −50 . . (1)
−10 −50 40
We know that, stiffness matrix for convection,

ℎ 𝑠 2 0 1 ℎ 𝑠 0 0 0
[𝐾 ] = 0 0 0 + 0 2 1
6 6
1 0 2 0 1 2
𝑤ℎ𝑒𝑟𝑒, 𝑠 = 𝐿𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑒𝑑𝑔𝑒 1 − 3

𝑠 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

𝑠 = (4 − 12) + (6 − 8)
𝑠 = 8.25 𝑐𝑚
 Finite Element Analysis 5.65

Similarly, s = Length of edge 3 − 2

𝑠 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

𝑠 = (12 − 8) + (8 − 10)

𝑠 = 4.47 𝑐𝑚
Substitution of these values in equation [ Kh]

15 × 8.25 2 0 1 10 × 4.47 0 0 0
[𝐾 ] = 0 0 0 + 0 2 1
6 6
1 0 2 0 1 2
41.25 0 20.62
[𝐾 ] = 0 14.90 7.45 … (2)
20.62 7.45 56.15

q A 1
Force vector, = 1 at side s
3
1

50 × 12 1
= 1
3
1
200
{𝐹 } = 200 … (3)
200

𝑞𝑠 1
{𝐹 } = 1 𝑎𝑡 𝑠𝑖𝑑𝑒 𝑠 .
2
0

{𝐹 } = 0 … (4)

[ ∵ No boundary heat flux is specified]

ℎ 𝑠 0 ℎ 𝑠 1
{𝐹 } = 1 + 0
2 2
1 1
5.66 Application to Heat Transfer and Dynamic Analysis

10 × 40 × 4.47 0 15 × 40 × 8.25 1
{𝐹 } = 1 + 0
2 2
1 1
2475
{𝐹 } = 894
3369
Result:
(1) Element matrix:
25 −35 10
Conduciton: [K ] = −35 85 −50
10 −50 40
41.25 0 20.62
Conveciton: [K ] = 0 14.90 7.45
20.62 7.45 56.15
200
(2)Force vector: {F } = 200
200
{𝐹 } = 0
2475
{𝐹 } = 894
3369

Example 5.8
Find the temperature distribution in a square region with uniform energy
generation as shown in Fig.(i). Assume that there is no temperature variation in the z-
direction. Take k=30 W/cm °C, 1= 10 cm, T=50°C, q = 100 W/cm³.

Fig. (i).
 Finite Element Analysis 5.67

Given: k = 30 W/cm OC
l = 10 cm
T = 500C
Q = 100 W/cm3
To find: Temperature distribution.
Solution: Converting into 2 elements
Element 1:

1 1 𝑥 𝑦 1 1 0 0
Area of triangle, 𝐴 = 1 𝑥 𝑦 = 1 10 0
2 1 𝑥 𝑦 2
1 10 10
𝐴 = 50 𝑐𝑚
Length at side, 1-3

 𝑙 = (𝑥 − 𝑥 ) + (𝑦 − 𝑦 )

= (0 − 10) + (10 − 0)
𝑙 = 14.14 𝑐𝑚
We know that,
𝑏 = (𝑦 − 𝑦 ) = (0 − 10) = −10

𝑏 = (𝑦 − 𝑦 ) = (10 − 0) = 10

𝑏 = (𝑦 − 𝑦 ) = (0 − 0) = 0

𝑐 = (𝑥 − 𝑥 ) = (10 − 10) = 0

𝑐 = (𝑥 − 𝑥 ) = (0 − 10) = −10

𝑐 = (𝑥 − 𝑥 ) = (10 − 0) = 10

(𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
𝑘
[𝐾 ] = (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 )
4𝐴
(𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 𝑏 + 𝑐 𝑐 ) (𝑏 + 𝑐 )
5.68 Application to Heat Transfer and Dynamic Analysis

30 100 −100 0
[𝐾 ] = −100 200 −100
4 × 50
0 −100 100

1 2 3
15 −15 0 𝟏
[𝐾 ] = −15 30 −15 𝟐
0 −15 15 𝟑
Similarly,
Element 2:
1
Area of triangle 𝐴 ,= 𝑏ℎ
2
1
= × 10 × 10
2
1
𝐴 = (100) = 50 𝑐𝑚
2
We know that,
𝑏 = (𝑦 − 𝑦 ) = (10 − 10) = 0

𝑏 = (𝑦 − 𝑦 ) = (10 − 0) = 10

𝑏 = (𝑦 − 𝑦 ) = (0 − 10) = −10

𝑐 = (𝑥 − 𝑥 ) = (10 − 0) = 10

𝑐 = (𝑥 − 𝑥 ) = (0 − 10) = −10

𝑐 = (𝑥 − 𝑥 ) = (0 − 0) = 0

Stiffness matrix for element (2),

30 100 −100 0
[𝐾 ] = −100 200 −100
4 × 50
0 −100 100
1 4 3
15 −15 0 𝟏
[𝐾 ] = −15 30 −15 𝟒
0 −15 15 𝟑
 Finite Element Analysis 5.69

Assembling [ K ] matrix,
1 2 3 4
30 −15 0 −15 𝟏
[ 𝐾 ] = −15 30 −15 0 𝟐
0 −15 30 −15 𝟑
−15 0 −15 30 𝟒
𝑞𝐴 1
Load vector, {𝐹} = 1
3
1
Element (1),
𝐹 100 × 50 1 1666.67 1
𝐹 = 1 = 1666.67 2
𝐹 3
1 1666.67 3
Similarly,
Element (2),
𝐹 1666.67 1
𝐹 = 1666.67 4
𝐹 1666.67 3
Assembling load vector,
𝐹 3.3 × 10
𝐹
= 1.6 × 10
𝐹 3.3 × 10
𝐹 1.6 × 10
Global matrix, {F} = [K][T]
Temperature distribution at node 2, 3 and 4 are
𝑇 = 𝑇 = 𝑇 = 𝑇 = 50 𝐶
3.3 × 10 30 −15 0 −15 𝑇
1.6 × 10 = −15 30 −15 0 = 50
3.3 × 10 0 −15 30 −15 50
1.6 × 10 −15 0 −15 30 50
We know that,
3.3 × 10 = (30 𝑇 ) − (15 × 50) + (0 × 50) − (15 × 50)
𝑇 = 160 𝐶
5.70 Application to Heat Transfer and Dynamic Analysis

Result: Temperature distribution,

𝑇 160
𝑇
= 50 𝐶
𝑇 50
𝑇 50

5.10 DYNAMIC ANALYSIS

5.10.1 Introduction
It provides an elementary introduction to time-dependent problems. It provides
the basic equations necessary for structural dynamics analysis and develop both the
lumped and the consistent mass matrix involved in the analysis of bar, beam and spring
elements.
We discuss method the formulation and solution of vibration problems using finite
element

5.10.2 Fundamentals of Vibration

Any motion which repeats itself after an interval of time is called vibration or
oscillation or periodic motion.
All bodies possessing mass and elasticity are capable of producing vibrations.
Vibration problems, in practice, occur wherever there are rotating or moving parts in a
machinery. The study of vibration is concerned with oscillatory motions of the bodies and
the forces associated with them.
Illustration: Consider a spring-mass system constrained to move in a rectilinear manner
along the axis of the spring, as shown in Fig.5.9.
When the mass is displaced from its equilibrium position A, the internal forces in
the form of elastic or strain energy are present in the body; and hence the mass reaches
position B.
At release, these forces bring the mass to its original position. At the equilibrium
position A, the whole of the elastic or strain energy is converted into kinetic energy due
to which the mass continues to move in the opposite direction to position C.
At C, the whole of the kinetic energy is again converted into elastic or strain
energy due to which the body again returns to the equilibrium position A.
 Finite Element Analysis 5.71

Fig. 5.9. Vibration of a spring-mass system

In this way, vibratory motion is repeated indefinitely and exchange of energy
takes place.
Similarly, the swinging of simple pendulum is another example of vibration as the
motion of ball is to and fro from its mean position repeatedly.

5.10.3 Causes of Vibrations:

The main causes of vibration are as follows:
1. Unbalanced forces in the machine. These forces are produced from within
machine itself.
2. Elastic nature of the system.
3. Self-excitations produced by the dry friction between the two mating surfaces.
4. External excitations applied on the system.
5. Winds may cause the vibrations in certain systems such as transmission and
telephone lines under certain conditions.
6. Earthquakes also cause vibrations and are greatly responsible for the failure of and
dams, many buildings, etc.

5.10.4 Effects of Vibrations

(i) Negative effects: The existence of vibrating elements in any mechanical produces
unwanted noise, high stresses, wear, poor reliability and premature failure of one
more of the parts. In addition to this, vibrations are a great source of human
discomfort in the form of physical and mental strains.
5.72 Application to Heat Transfer and Dynamic Analysis

(ii)Positive effects: Inspite of the harmful effects, the vibratory systems are built into
the system machines. Examples are almost all musical instruments, vibrating
conveyors, vibrating screens, shakers, stress relievers, etc.

5.10.5 Methods of Elimination/Reduction of the Undesirable Vibrations

The undesirable vibrations can be eliminated or reduced by one or more of the
following methods.
 By removing the causes of vibration.
 By resting the machinery on proper type of isolators.
 By using shock absorbers.
 By using dynamic vibration absorbers.
 By using the screens (if noise is the objection).

5.10.6. Terminology Used in Vibratory Motion

The terms commonly used in the study of vibrations are presented in
1. Periodic motion: A motion which repeats itself after equal interval of time.
2. Time period (tp): It is the time taken by a motion to repeat itself. It is also called
as period of vibration, and is measured in seconds.
3. Cycle: It is the motion completed during one time period.
4. Frequency (f): It is the number of cycles completed in one second. It is expressed
in hertz (Hz). It is a reciprocal of time period. Mathematically, 𝑓 = Hz.

5. Natural frequency: Frequency of free vibration of the system.

6. Amplitude (X): The maximum displacement of a vibrating body from the mean
position.
7. Resonance: When the frequency of the external force is equal to the natural
frequency of a vibrating body, the amplitude of vibration becomes excessively
large. This phenomenon is known as resonance.
8. Damping: It is the resistance to the motion of a vibrating body.
 Finite Element Analysis 5.73

5.10.7 Simple Harmonic Motion

 Since most of the vibrating systems follow simple harmonic motion (SHM),
therefore is essential to have proper understanding of SHM related basic concepts.
 A body is said to have simple harmonic motion (SHM), if it moves or vibrates
about a position such that its acceleration is always proportional to its distance
from the position and is directed towards the mean position or equilibrium
position.
Consider a particle 'P' moving around a circle with a uniform angular velocity o
rad/s as shown in fig.5.10

.
Fig 5.10 Simple harmonic motion of a particle moving around a circle
Displacement of particle 'P' from mean position after time 't', as shown in Fig.5.10,
is given by
x = X sin of
X = Maximum displacement (or amplitude) of particle from mean position.
Velocity of particle after time ‘t’ is given by
𝑑𝑥
𝑣= − 𝜔 𝑋 cos 𝜔𝑡
𝑑𝑡
Acceleration of particle after time ‘t’ is
𝑑 𝑥
𝑎= = −𝜔 𝑋 sin 𝜔𝑡 = − 𝜔 𝑥 [since 𝑥 = X sin 𝜔𝑡]
𝑑𝑡
𝑑 𝑥
or ±𝜔 𝑥 =0
𝑑𝑡
The above equation is known as differential equation or fundamental equation of
S.H.M
5.74 Application to Heat Transfer and Dynamic Analysis

Time period and frequency:

2𝜋
Time period, 𝑡 =
𝜔
1 𝜔
Frequency, 𝑓 = =
𝑡 2𝜋

5.10.8 Types of Vibrations

Vibrations may be classified according to:
(a) the actuating force on the body, and
(b) the stresses in the supporting medium, as shown in Fig.5.11.

Fig. 5.11. Types of vibrations

I. According to the Actuating Force
1. Free or Natural Vibrations
 If the periodic motion continues after the cause of original disturbance (i.e., initial
displacement) is removed, then the body is said to be under free or natural
vibrations.
 The frequency of the free vibrations is called free or natural frequency.
 Example: Oscillation of a simple pendulum
2. Forced Vibrations
 When the body vibrates under the influence of external force, then the body is said
to be under forced vibrations.
 Finite Element Analysis 5.75

 The vibrations have the same frequency as the applied force.

 Examples: Vibrations in machine tools, electric bells, vibratory conveyors, etc.
3. Damped Vibrations
When there is a reduction in amplitude over every cycle of vibration, the motion
is said to be damped vibration. That is, if the vibratory system has a damper, the motion
of the system will be opposed by it and the energy of the system will be dissipated in
friction.
On the contrary, the system having no damper is known as undamped vibration.
If the damper is connected with free vibrating body to control vibrations, then it
is called free damped vibrations. If the damper is connected with forced vibrating body
to control vibrations, then it is called forced damped vibrations.
Examples: Vibrations in all machinery in actual use are damped in nature.
4. Undamped vibrations
If no energy is lost or dissipated in friction or other resisting force during
vibration, then such vibration is known as undamped vibration.
In other words, the system having no damper produces undamped vibrations.
In the vibratory system, if the amount of external excitation is known in
magnitude, it causes deterministic vibration.
II. According to Motion of System with Respect to Axis
Consider a vibrating body, e.g., a rod, shaft or spring. Fig.5.12 shows a heavy disc
carried on one end of a weightless shaft, the other end being fixed. This system can
execute any one of the following types of vibrations.
1. Longitudinal Vibrations
When the particles of the shaft or disc moves parallel to the axis of the shaft, then
the vibrations are known as longitudinal vibrations, as shown in Fig. 5.12 (a).
2. Transverse Vibrations
When the particles of the shaft or disc move approximately perpendicular to the
axis of the shaft, then the vibrations are known as transverse vibrations, as shown in Fig.
5.12 (b).
5.76 Application to Heat Transfer and Dynamic Analysis

3. Torsional Vibrations
When the particles of the shaft or disc move in a circle about the axis of the shaft,
then the vibrations are known as torsional vibrations, as shown in Fig. 5.12 (c).

5.11 EQUATION OF MOTION USING LAGRANGE’S APPROACH

𝜕 𝑑𝑇 𝜕𝑇 𝜕𝜋 𝐹
− + =
𝜕𝑡 𝑑𝑢̇ 𝜕𝑢 𝜕𝑢 𝑢
𝐹
Where = Generalised force in the coordinates 𝑢
𝑢
𝜕𝑢
𝑢̇ =
𝜕𝑡

5.12 Formulation of Finite Element Equations

We know that, shape functions for a typical finite element, the displacement of
an interior point can be written in terms of the nodal degree of freedom as,
{ 𝑢} = 𝑁 𝑢 + 𝑁 𝑢 … (1)

Differentiating with time the velocity at the point is given by,

{𝑢̇ } = 𝑁 𝑢 + 𝑁 𝑢 … (2)

For the bar or truss element,

𝑥 𝑥 𝑢
𝑢= 1− 𝑢 … (3)
𝑙 𝑙

𝑥 𝑥 𝑢̇
𝑢̇ = 1− … (4)
𝑙 𝑙 𝑢̇

For the beam element,

𝑣
𝜃
𝑣 = [𝑁 𝑁 𝑁 𝑁 ] 𝑣 … (5)
𝜃
 Finite Element Analysis 5.77

⎧𝑣̇ ⎫
⎪𝜃̇ ⎪
𝑣̇ = [𝑁 𝑁 𝑁 𝑁 ] … (6)
⎨𝑣̇ ⎬
⎪ ⎪
⎩𝜃̇ ⎭
For a two-dimensional element, each point can have u and v displacements,
𝑢
⎧𝑣 ⎫
𝑢 𝑁 0 𝑁 0 𝑁 0 ⎪𝑢 ⎪
{ 𝑢} = =
𝑣 0 𝑁 0 𝑁 0 𝑁 ⎨𝑣 ⎬
⎪𝑢 ⎪
⎩𝑣 ⎭
𝑢̇
⎧𝑣 ̇ ⎫
⎪ ⎪
𝑁 0 𝑁 0 𝑁 0 𝑢̇
{ 𝑢̇ } = 𝑢̇ =
0 𝑁 0 𝑁 0 𝑁 ⎨𝑣 ̇ ⎬
𝑣̇
⎪𝑢 ̇ ⎪
⎩𝑣 ̇ ⎭
The kinetic of an element mass, m = p v within the element is given by,
1
𝑇= 𝑚𝑣
2
1
𝑑𝑇 = (𝑑𝑚) (𝑣)
2
1
𝑑𝑇 = 𝜌 𝑑𝑣 (𝑣 ) [∵ 𝑚 = 𝑝 𝑣; 𝑑𝑚 = 𝑝 𝑑𝑣]
2
1
𝑑𝑇 = {𝑢̇ } {𝑢̇ }𝜌 𝑑𝑣 … (7)
2

where, { 𝑢} = [ 𝑁 ]{𝑢}

{ 𝑢̇ } = [ 𝑁 ]{𝑢̇ } … (8)
Substitute the equation (7) in equation (8),
1
𝑑𝑇 = {𝑢̇ } [𝑁] [𝑁] {𝑢̇ } 𝜌 𝑑𝑣 … (9)
2
5.78 Application to Heat Transfer and Dynamic Analysis

Integrating the above equation,

1
𝑇= {𝑢̇ } 𝜌[𝑁] [𝑁] 𝑑𝑣 {𝑢̇ }
2

1
𝑇= {𝑢̇ } [𝑚] {𝑢̇ } … (10)
2

where, [m] = ρ[N] [N] is the consistent mass matrix for the element,

The total potential energy of the system,

1
 = {𝑢} [𝐾]{𝑢} − {𝑢} {𝐹} … (11)
2
1
𝐸𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (10) 𝑇 = {𝑢̇ } [𝑚]{𝑢̇ } … (12)
2
𝜕𝑇
 = 0,
𝜕𝑢
𝜕𝑇
= [𝑚]{𝑢̇ },
𝜕{𝑢̇ }

𝑑 𝜕𝑇
= [𝑚]{𝑢 ̈ },
𝑑𝑡 𝜕{𝑢̇ }

We know that,

𝜕𝜋
= [ 𝐾]{𝑢} − [𝐹]
𝜕{𝑢}

Substituting the above values in Lagrange’s equation of motion,

[𝑚]{𝑢̈ } + [𝐾]{𝑢} − [𝐹] = 0

𝑖. 𝑒., [𝑚]{𝑢̈ } + [𝐾]{𝑢} = {𝐹} … (13)

 Finite Element Analysis 5.79

5.13. CONSISTENT MASS MATRIX FOR VARIOUS ELEMENT

For the bar element: shape Functions are,
𝑥 𝑥
𝑁 = 1 − ,𝑁 =
𝑙 𝑙

Mass matrix, [𝑚] = 𝜌[𝑁] [𝑁] 𝑑𝑣 … (14)

𝑁
=𝜌𝐴 [𝑁 𝑁 ] 𝑑𝑥 [∵ 𝑑𝑣 = 𝐴 𝑑𝑥]
𝑁

𝑥
1− 𝑥 𝑥
=𝜌𝐴 𝑙 1− 𝑑𝑥
𝑥 𝑙 𝑙
𝑙
𝑥 𝑥 𝑥
⎡ 1− − ⎤
=𝜌𝐴 ⎢ 𝑙 𝑙 𝑙 ⎥ 𝑑𝑥
⎢𝑥 𝑥 𝑥 ⎥
⎣𝑙−𝑙 𝑙 ⎦

𝑥
⎡ 1− 𝑙 𝑥 𝑥 ⎤
⎢ − ⎥
−1 2𝑙 3𝑙 ⎥
= 𝜌 𝐴⎢ 3 𝑙
⎢ ⎥
⎢𝑥 𝑥 𝑥 ⎥
−
⎣2 𝑙 3 𝑙 ⎦
3𝑙
𝑙 𝑙 𝑙
−
=𝜌𝐴 3 2 3
𝑙 𝑙 𝑙
−
2 3 3
𝑙 𝑙
=𝜌𝐴 3 6
𝑙 𝑙
6 3
𝜌𝐴𝑙 2 1
[𝑚] = . . (15)
6 1 2
5.80 Application to Heat Transfer and Dynamic Analysis

For the Beam Element: Shape functions are,

3𝑥 2𝑥
𝑁 = 1− +
𝑙 𝑙
2𝑥 𝑥
𝑁 =𝑥− +
𝑙 𝑙
3𝑥 2𝑥
𝑁 = −
𝑙 𝑙
𝑥 𝑥
𝑁 = − +
𝑙 𝑙
We know that

Mass matrix, [𝑚] = 𝜌[𝑁] [𝑁] 𝑑𝑣 … (16)

𝑁
𝑁
=𝜌𝐴 = [𝑁 𝑁 𝑁 𝑁 ]𝑑𝑥 … (17)
𝑁
𝑁
Substituting 𝑁 , 𝑁 , 𝑁 𝑎𝑛𝑑 𝑁 values and after performing all the integrations,
the beam element mass matrix is obtained as
156 22 𝑙 54 −13 𝑙
𝜌 𝐴 𝑙 22 𝑙 4𝑙 13 𝑙 −3/𝑙
[𝑚] = . . (18)
420 54 13 𝑙 156 −22 𝑙
−13 𝑙 −3 𝑙 −22 𝑙 4/𝑙

5.14 LUMPED MASS MATRIX

Fig. 5.13 Mass lumping for bar and beam element

 Finite Element Analysis 5.81

The lumped mass matrix for a bar element is given by

𝜌𝐴𝑙 1 0
[𝑚] = … (19)
2 0 1
The lumped mass matrix for a beam element is given by
1 0 0 0
𝜌𝐴𝑙 0 0 0 0
[𝑚] = … (20)
2 0 0 1 0
0 0 0 0

5.15 FORMULAE USED

1. Longitudinal Vibration of Bar
Finite element equation,
{[𝐾] − [𝑚 ]𝜔 {𝑢} = {𝐹}

EA 1 −1
Where, Stiffness matrix, {[K] =
L 1 1
𝜌𝐴𝐿 2 1
[𝑚] = for consistent mass matrix
6 1 2
𝜌𝐴𝐿 1 0
[𝑚] = for lumped mass matrix
2 0 1
2. Transverse Vibration of Beam
Finite element equation,
{[𝐾] − [𝑚 ]𝜔 {𝑢} = {𝐹}

12 6l −12 6l
EA 6l 4 l −6l 2l
Where, Stiffness matrix, {[K] =
l −12 −6 l 12 −6 l
6l 2l −6 l 4l
156 22 l 54 −13 l
ρ A L 22 l 4l 13 l −3 l
Mass matrix, [m] =
420 54 13 l 156 −22 l
−13 l −3 l −22 l 4 l
For consistent mass matrix
5.82 Application to Heat Transfer and Dynamic Analysis
1 0 0 0
𝜌𝐴𝐿 0 0 0 0 𝑓𝑜𝑟 𝑙𝑢𝑚𝑝𝑒𝑑 𝑚𝑎𝑠𝑠 𝑚𝑎𝑡𝑟𝑖𝑥
[𝑚] =
2 0 0 1 0
0 0 0 0

5.16 SOLVED PROBLMES

Example 5.9
Find the natural frequency of longitudinal vibration of the unconstrained stepped
bar as shown in Fig.(i).

Fig. (i).
Given:

Fig. (ii).
Element (1) Element (2)
Area, A = 2 A Area, A = 2 A

Length, L = Length, L =

Young , s modulus, E = E Young , s modulus, E = E

Density, ρ = ρ Density, ρ = ρ
To find: Natural frequencies of the rod.
Solution: The bar with two element and 3 nodes area as shown in Fig.(ii). The stiffness
matrix of the two elements are,
 𝐴 𝐸 1 −1
[𝐾 ] =
𝐿 −1 1

2 𝐴 𝐸 1 −1
=
𝐿 −1 1
2
4 𝐴 𝐸 1 −1
=
𝐿 −1 1

1 2
2 𝐴 𝐸 1 −1 𝟏
[𝐾 ] =
𝐿 −1 1 𝟐

Similarly, Stiffness matrix for element (2),

𝐴 𝐸 1 −1
[𝐾 ] =
𝐿 −1 1

𝐴 𝐸 1 −1
=
𝐿 −1 1
2

2 3
2 𝐴 𝐸 1 −1 𝟐
[𝐾 ] = … (2)
𝐿 −1 1 𝟑

Assemble the stiffness matrix,

1 2 3
2𝐴𝐸 2 −2 0 𝟏
[𝐾 ] = −2 3 −1 𝟐 … (3)
𝐿
0 −1 1 𝟑

Mass matrix for Element (1),

𝜌 𝐴 𝐿 2 1
[𝑚 ] =
6 1 2

[From equation (15)]

 𝐿
𝜌×2𝐴×2 2 1
[𝑚 ] =
6 1 2

1 2
𝜌𝐴𝐿 2 1 𝟏
=
6 1 2 𝟐
1 2
𝜌𝐴𝐿 4 2 𝟏
[𝑚 ] = … (4)
12 2 4 𝟐
Similarly, mass matrix for Element (2),
𝐿
𝜌 𝐴 𝐿 2 1 𝜌 𝐴 × 2 1
[𝑚 ] = = 2
6 1 2 6 1 2

2 3
𝜌𝐴𝐿 2 1 𝟐
[𝑚 ] = … (5)
12 1 2 𝟑

1 2 3
𝜌𝐴𝐿 4 2 0 𝟏
Assemble the mass matrix, [𝑚] = 2 6 1 𝟐 … (6)
12
0 1 2 𝟑
Since, the bar is unconstrained (no degrees of freedom is fixed), the finite element
equation is
{[𝐾] − [𝑚]𝜔 }{𝑢} = {𝑃}
Substitute [K] and [m] values
𝑢 𝑃
2𝐴 𝐸 2 −2 0 𝜌𝐴𝐿 4 2 0
−2 3 −1 − 𝜔 2 6 1 𝑢 − 𝑃
𝐿 12 𝑢 𝑃
0 −1 1 0 1 2
Applying boundary conditions,
𝑃 =𝑃 =𝑃 =0
[No degrees of freedom is fixed]
We set the determinant of the coefficient matrix equal to zero, we have
 2𝐴 𝐸 2 −2 0 𝜌𝐴𝐿 4 2 0
−2 3 −1 − 𝜔 2 6 1 =0 … (7)
𝐿 12
0 −1 1 0 1 2
2𝐴 𝐸
Divide both sides by
𝐿
2 −2 0 𝜌𝐴𝐿 4 2 0
𝜔
12
−2 3 −1 − 2 6 1 =0
2𝐴 𝐸
0 −1 1 𝐿 0 1 2
2 −2 0 𝜌𝐿 𝜔 4 2 0
−2 3 −1 − 2 6 1 =0 … (8)
24 𝐴
0 −1 1 0 1 2
𝜌𝐿 𝜔
Take, 𝛽 =
24 𝐸
Equation (8) can be rewritten as,
2 −2 0 4 2 0
−2 3 −1 − 𝛽 2 6 1 =0
0 −1 1 0 1 2

2(1 − 2𝛽 ) −2(1 − 𝛽 ) 0
 −2(1 + 𝛽 ) 3(1 − 2 𝛽 ) −(1 + 𝛽 ) = 0
0 −1(1 + 𝛽 ) (1 − 2 𝛽 )

 2(1 − 2𝛽 )[3 (1 − 2 𝛽 ) − (1 + 2 𝛽 ) ]
+ 2(1 + 𝛽 )[ −2(1 + 𝛽 )(1 − 2 𝛽 )] = 0

Bu simplifying the above equation, we get

 18[𝛽 (1 − 2𝛽 )(𝛽 − 2) = 0 … (9)
The roots of equation (9) give the natural frequencies of the bar.
𝜌𝐿 𝜔
We know that,  =
24 𝐸
when,  =0𝜔 = 0𝜔 =0

1 12 𝐸 𝐸
when,  = 𝜔 =  𝜔 = 3.46 𝑟𝑎𝑑/𝑠
2 𝜌𝐿 (𝜌 𝐿 )
 48 𝐸 𝐸
when,  =2𝜔 =  𝜔 = 6.92 𝑟𝑎𝑑/𝑠
𝜌𝐿 (𝜌 𝐿 )
𝑁𝑎𝑡𝑢𝑟𝑎𝑙 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑖𝑒𝑠 𝑎𝑟𝑒, 𝜔 = 0

𝐸
𝜔 = 3.46
(𝜌 𝐿 )

𝐸
𝜔 = 6.92
(𝜌 𝐿 )
Result: Natural frequencies of longitudinal vibration,
𝜔 =0

𝐸
𝜔 = 3.46 𝑟𝑎𝑑/𝑠
(𝜌 𝐿 )

𝐸
𝜔 = 6.92 𝑟𝑎𝑑/𝑠
(𝜌 𝐿 )

Example 5.11
For the bar as shown in Fig.(i) with length 2 L, modulus of elasticity E, mass
density p, and cross-sectional area A, determine the first two natural frequencies.

Fig. (i).
Given:

Fig. (ii).
Length, L = 2 L
Young s modulus, E = E
 Finite Element Analysis 5.87

Mass Density, ρ = ρ
Cross-sectional area, A = A
To find: Natural frequencies.
Solution: We can divide the bar with two as shown in Fig.(iii).

Fig. (iii).
Stiffness matrix for element (1):
1 2
𝐴 𝐸 1 −1 1
[𝐾 ] =
𝐿 −1 1 2
Similarly,
2 3
𝐴 𝐸 1 −1 𝟐
𝑬𝒍𝒆𝒎𝒆𝒏𝒕 (𝟐): [𝐾 ] =
𝐿 −1 1 𝟑
Assembling the element matrix,
1 2 3
𝐴 𝐸 1 −1 0 𝟏
[𝐾] = −1 2 −1 𝟐 … (1)
𝐿
0 −1 1 𝟑
Lumped mass matrix or consistent mass matrix can be used for solving the problem.
Lumped mass matrix for Element (1):
1 2
𝜌𝐴𝐿 1 0 𝟏
[𝑚 ] = [𝐹𝑟𝑜𝑚 𝑒𝑞𝑢𝑎𝑖𝑡𝑜𝑛 (15)]
2 0 1 𝟐
Similarly
Element (2),
1 2
𝜌𝐴𝐿 1 0 𝟏
[𝑚 ] =
2 0 1 𝟐
 𝜌𝐴𝐿 1 0 0 𝟏
Assemble the mass matrix, [𝑚] = 0 2 0 𝟐 … (2)
2
0 0 1 𝟑
Global matrix, for bar element,
{[𝐾] − 𝜔 [𝑚]}{𝑢} = {𝑃}

𝑢 𝑃
𝐴 𝐸 1 −1 0 𝜌𝐴𝐿 1 0 0
−1 2 −1 − 𝜔 0 2 0 𝑢 − 𝑃 … (3)
𝐿 12 𝑢 𝑃
0 −1 1 0 0 1
Applying boundary conditions,
𝑢 = 0 (𝑓𝑖𝑥𝑒𝑑), 𝑃 =0
𝑢 =𝑢 𝑃 =0
𝑢 =𝑢 𝑃 =0

𝐴 𝐸 1 −1 0 𝜌𝐴𝐿 1 0 0 0 0
 −1 2 −1 − 𝜔 0 2 0 𝑢 − 0
𝐿 12 𝑢
0 −1 1 0 0 1 0

In the above equation, 𝑢 = 0, so, neglect first row and first column of [K] and
[m]matrix. The final reduced equation is,
𝐴 𝐸 2 −1 𝜌𝐴𝐿 2 0 𝑢
− 𝜔 𝑢 =0 … (4)
𝐿 −1 1 12 0 1

To obtain a solution to the set of homogeneous equation in equation (4), we set

the determinant of the coefficient matrix equal to zero.
𝐴 𝐸 2 −1 𝜌𝐴𝐿 2 0
− 𝜔 =0 … (5)
𝐿 −1 1 12 0 1

Divide both sides by AL,

𝐴 𝐸 2 −1 𝜌𝐴𝐿 2 0
−  =0 [∵  = 𝜔 ]
𝜌 𝐴 𝐿 −1 1 2 ×𝜌𝐴𝐿 0 1

𝐸 2 −1  2 0
− =0
𝜌 𝐿 −1 1 2 0 1
 𝐸
Take, 𝜇=
𝜌 𝐿
2 −1  2 0
𝜇 − =0
−1 1 2 0 1
(2 𝜇 − ) −𝜇
 =0 … (6)
−𝜇 𝜇−
2

 (2𝜇 − )( 𝜇 − ) [−𝜇 ] = 0
2

 2𝜇 −𝜇−𝜇 + −𝜇 =0
2


 𝜇 − 2𝜇  + =0
2

 − 2𝜇  + 𝜇 = 0 … (7)
2
By solving the quadratic equation (7),

4
4 − 2  −𝑏 ± √𝑏 − 4 𝑎 𝑐
 = −(−2) ± [∵  = ]
2 2𝑎
2

 = 2 ± √2

 =  [2 ± √2 ]

  = 3.41 ,  = 0.585 ,

𝑁𝑎𝑡𝑢𝑟𝑎𝑙 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑖𝑒𝑠 𝑎𝑟𝑒,

we know that  = ω = 0

 𝜔 = √
 𝜔 = 3.41 
𝜔 = 1.85  𝑟𝑎𝑑/𝑠
 𝐸 𝐸
∴ 𝜔 = 1.85 𝑟𝑎𝑑/𝑠 [∵ 𝜇 = ]
𝜌 𝐿 𝜌𝐿

Similarly, 𝜔 = 0.585 

𝜔 = 0.76  𝑟𝑎𝑑/𝑠𝑒𝑐

𝐸
𝜔 = 0.76 𝑟𝑎𝑑/𝑠
𝜌 𝐿

Result: Natural frequencies are,

𝐸
𝜔 = 1.85 𝑟𝑎𝑑/𝑠
𝜌 𝐿

𝐸
𝜔 = 0.76 𝑟𝑎𝑑/𝑠
𝜌 𝐿

Example 5.12
Consider a uniform cross-section bar as shown in Fig.(i) of length “L” made up
of a material whose Young’s modulus and density are given by E and . Estimate the
natural frequencies of axial vibration of the bar using both lumped and consistent mass
matrix.

Fig. (i). A uniform bar

Given: 𝐿𝑒𝑛𝑔𝑡ℎ, 𝐿 = 𝐿
𝑌𝑜𝑢𝑛𝑔 𝑠 𝑚𝑜𝑑𝑢𝑙𝑢𝑠, 𝐸 = 𝐸
𝐷𝑒𝑛𝑠𝑖𝑡𝑦, 𝜌 = 𝜌
To find: Natural frequencies of axial vibration of the bar by using both lumped and
consistent mass matrix.
 Finite Element Analysis 5.91

Solution:
Lumped mass matrix [One element]

Fig. (ii).
Using just one element for the entire rod i.e., L =L and using lumped mass
matrix, we have
[ 𝐾 ] − 𝜔 [𝑚] {𝑢} = 0

[ 𝐾 ]{𝑢} − 𝜔 [𝑚] {𝑢}

𝜌𝐴𝐿
𝐴 𝐸 1 −1 𝑢 0 𝑢
− 𝜔 2
𝐿 −1 1 𝑢 𝜌𝐴𝐿 𝑢
0
2
𝜌𝐴𝐿 1 0
[∵ [𝑚] = ]
2 0 1

Applying boundary conditions,

𝑢 = 0 (𝑓𝑖𝑥𝑒𝑑), 𝑢 =𝑢
𝜌𝐴𝐿
𝐴 𝐸 1 −1 0 0 0
− 𝜔 2
𝐿 −1 1 𝑢 𝜌𝐴𝐿 𝑢
0
2
In the above equation, 𝑢 = 0, so, neglect first row and first column of [K] and
[m] matrix. The final reduced equation is,
𝐴𝐸 𝜌𝐴𝐿
(𝑢 ) = 𝜔 (𝑢 )
𝐿 2
𝐸 𝜌 𝐿
=𝜔
𝐿 2
 2𝐸
𝐻𝑒𝑛𝑐𝑒, 𝜔 =
𝜌𝐿

1.414 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠 … (1)
𝐿 𝜌

Consistent mass matrix (one element):

We know that, with one element and consistent mass matrix,
{[ 𝐾 ] − 𝜔 [ 𝑚]}{𝑢} = 0

[ 𝐾 ]{𝑢} = 𝜔 [ 𝑚]

𝐴 𝐸 1 −1 𝑢 𝜌𝐴𝐿 2 1 𝑢
− 𝜔
𝐿 −1 1 𝑢 6 1 2 𝑢

𝜌𝐴𝐿 2 1
∵ [𝑚] =
6 1 2
Applying boundary conditions, 𝑢 = 0(𝑓𝑖𝑥𝑒𝑑)
𝑢 =𝑢
𝐴 𝐸 1 −1 0 𝜌𝐴𝐿 2 1 0
− 𝜔
𝐿 −1 1 𝑢 6 1 2 𝑢
In the above equation, 𝑢 = 0, so, neglect first row and first column of [K] and
[m ] matrix. The final reduced equation is,
𝐴𝐸 𝜌𝐴𝐿
(𝑢 ) = 𝜔 . (2 𝑢 )
𝐿 6
𝐸 𝜌 𝐿
=𝜔 .
𝐿 3

3𝐸
Therefore, 𝜔 =
𝜌𝐿

1.732 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠 … (2)
𝐿 𝜌
 Finite Element Analysis 5.93

Lumped mass matrix (Two element):

Divide the rod into 2 elements as shown in Fig. (iii).

Fig. (iii)
Stiffness matrix for bar element,
Element (1):
1 2
𝐴 𝐸 1 −1 1 𝐿
[𝐾 ] = [∵ 𝐿 = ]
𝐿 −1 1 2 2

1 2
𝐴 𝐸 1 −1 1
=
𝐿
(2) −1 1 2

Similarly,
1 2
2 𝐴 𝐸 1 −1 𝟏
[𝐾 ] =
𝐿 −1 1 𝟐
𝐴 𝐸 1 −1
𝐄𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐): [𝐾 ] =
𝐿 −1 1

𝐴 𝐸 1 −1 𝐿
= [∵ 𝐿 = ]
𝐿
(2) −1 1 2

2 3
2 𝐴 𝐸 1 −1 𝟐
[𝐾 ] =
𝐿 −1 1 𝟑
5.94 Application to Heat Transfer and Dynamic Analysis

Assemble the stiffness matrix,

1 2 3
2 𝐴 𝐸 1 −1 0 𝟏
[𝐾] = −1 2 −1 𝟐 … (3)
𝐿
0 −1 1 𝟑
Lumped mass matrix for bar element
Element (1):
𝜌𝐴𝐿 1 0
[𝑚 ] =
2 0 1
𝜌𝐴𝐿 1 0 𝟐 𝐿
[𝑚 ] = ∵𝐿 =
2 0 1 𝟑 2
1 2
𝜌𝐴𝐿 1 0 𝟏
Similarly, [𝑚 ] =
4 0 1 𝟐
Element (2),
𝜌𝐴𝐿 1 0
[𝑚 ] =
2 0 1
2 3
𝜌𝐴𝐿 1 0 𝟐 𝐿
[𝑚 ] = ∵𝐿 =
2 0 1 𝟑 2
Assemble the lumped mass matrix,
1 2 3
𝜌𝐴𝐿 1 0 0 𝟏
[𝑚] = 0 2 0 𝟐 … (4)
4
0 0 1 𝟑
We know that, Global matrix, for axial vibration of the bar,
[𝐾] − 𝜔 [𝑚] {𝑢} = 0 … (5)
[𝐾]{𝑢} = 𝜔 [𝑚]{𝑢}
𝑢 0 𝑢
2 𝐴 𝐸 1 −1 0 𝜌𝐴𝐿 1 0
−1 2 −1 𝑢 = 𝜔 0 2 0 𝑢 … (6)
𝐿 4
0 −1 1 𝑢 0 0 1 𝑢
 𝑢 𝑃
2 𝐴 𝐸 1 −1 0 𝜌𝐴𝐿 1 0 0
−1 2 −1 − 𝜔 0 2 0 𝑢 − 𝑃 … (3
𝐿 12 𝑢 𝑃
0 −1 1 0 0 1
Applying boundary conditions,
𝑢 = 0 (𝑓𝑖𝑥𝑒𝑑),
𝑢 =𝑢
𝑢 =𝑢

2 𝐴 𝐸 1 −1 0 0 𝜌𝐴𝐿 1 0 0 𝑢
 −1 2 −1 𝑢 =𝜔 0 2 0 𝑢
𝐿 4
0 −1 1 𝑢 0 0 1 𝑢
In the above equation, 𝑢 = 0, so, neglect first row and first column of [K] and
[m] matrix. The final reduced equation is,
2 𝐴 𝐸 2 −1 𝑢 𝜌𝐴𝐿 2 0 𝑢
= 𝜔 … (7)
𝐿 −1 1 𝑢 12 0 1 𝑢

Divide both sides by  A L.

2 𝐴 𝐸 2 −1 𝑢 𝜌𝐴𝐿 2 0 𝑢
= 𝜔
𝜌 𝐴 𝐿 −1 1 𝑢 4𝜌𝐴𝐿 0 1 𝑢

𝐸 4 −2 1 2 0 𝑢
−𝜔 × =0 … (8)
𝜌 𝐿 −2 2 4 0 1 𝑢

To obtain a solution to the set of homogeneous equation in equation (8), we set

the determinant of the coefficient matrix equal to zero.
𝐸 4 −2 1 2 0
−𝜔 × =0
𝜌 𝐿 −2 2 4 0 1

𝑇𝑎𝑘𝑒, 𝜔 = 

𝐸
=𝜇
𝜌 𝐿

4 −2 0.5 0
𝜇 −  =0
−2 2 0 0.25
 (4 𝜇 − 5 ) (−2 𝜇)
 =0
(−2 𝜇) (2 𝜇 − 0.25 )

 [(4 𝜇 − 0.5 )(2 𝜇 − 0.25 )] − (4 𝜇 ) = 0

8 𝜇 − 𝜇  − 𝜇 + 0.125  − 4 𝜇 =0

0.125  − 2𝜇  + 4 𝜇 = 0

By solving the quadratic equation (7),

2 ± (−2) − 4(0.125)(4  ) −𝑏 ± √𝑏 − 4 𝑎 𝑐
= ∵ =
(2𝑋0.125) 2𝑎

2 ± 4 − 2 
 =
0.25

2 ± √2
 =
0.25
2 ± 1.42 
=
0.25

  = 8  ± 5.68 ,

 = 13.653 ,  = 2.343

13.65 𝐸 2.343
∴ =  =
𝜌𝐿 𝜌𝐿

We know that, natural frequencies are,

 =𝜔

 𝜔 = √

13.65 𝐸 3.695 𝐸
∴ 𝜔 = = 𝑟𝑎𝑑/𝑠 … (10)
𝜌𝐿 𝐿 𝜌
 2.343 𝐸 1.531 𝐸
Similarly, 𝜔 = = 𝑟𝑎𝑑/𝑠 … (11)
𝜌𝐿 𝐿 𝜌

Consistent mass Matrix (Two Elements)

We know that,

2𝐴𝐸 1 −1 0
Global stiffness matrix, [ K] = −1 2 −1 [From equation no. (3)]
𝐿
0 −1 1
We know that,
Consistent mass matrix for element (1):
𝜌𝐴𝐿 2 1
[𝑚 ]=
6 1 2
1 2
𝜌𝐴𝐿 2 1 1 𝐿
= ∵𝐿 =
12 1 2 2 2
Similarly,
Element (2):
𝜌𝐴𝐿 2 1
[𝑚 ]=
6 1 2
2 3
𝜌𝐴𝐿 2 1 2 𝐿
= ∵𝐿 =
12 1 2 3 2
Assemble the mass matrix, [m]
1 2 3

𝜌𝐴𝐿 2 1 0 𝟏
[ 𝑚] = 1 4 1 𝟐
12
0 1 2 𝟑
Global matrix, for axial vibration of bar,
{[𝐾] − 𝜔 [𝑚]}{𝑢} = 0
 𝑢
2 𝐴 𝐸 1 −1 0 𝜌𝐴𝐿 2 1 0
−1 2 −1 − 𝜔 1 4 0 𝑢 =0
𝐿 4 𝑢
0 −1 1 0 1 2
Applying boundary conditions,
𝑢 = 0 (𝑓𝑖𝑥𝑒𝑑),
𝑢 =𝑢
𝑢 =𝑢

2𝐴𝐸 1 −1 0 0 𝜌𝐴𝐿 2 1 0 0
 −1 2 −1 𝑢 =𝜔 1 4 0 𝑢 =0
𝐿 4
0 −1 1 𝑢 0 1 2 𝑢
In the above equation, 𝑢 = 0, so, neglect first row and first column of [K] and
m] matrix. The final reduced equation is,
2 𝐴 𝐸 2 −1 𝜌𝐴𝐿 4 1 𝑢
−𝜔 𝑢 =0
𝐿 −1 1 12 1 2

Divide both sides by  A L.

2 𝐴 𝐸 2 −1 𝜌𝐴𝐿 4 1 𝑢
−𝜔 𝑢 =0
𝜌 𝐴 𝐿 −1 1 12𝜌 𝐴 𝐿 1 2

𝐸 4 −2 1 4 1 𝑢
−𝜔 𝑢 =0 … (12)
𝜌 𝐿 −2 2 12 1 2

To obtain a solution to the set of homogeneous equation in equation (12), we set the
determinant of the coefficient matrix equal to zero.
𝐸 4 −2 0.333 0.0833
−𝜔 =0
𝜌 𝐿 −2 2 0.0833 0.1666
𝐸
𝑇𝑎𝑘𝑒,  = 𝜔 𝑎𝑛𝑑 = 𝜇
𝜌𝐿
4 −2 0.333 0.0833
𝜇 −  =0
−2 2 0.0833 0.1666
(4 𝜇 − 0.333 ) (−2 𝜇 − 0.0833 )
 =0
(−2 𝜇 = 0.0833 ) (2 𝜇 − 0.166 )
 [(4 𝜇 − 0.333 )(2 𝜇 − 0.166 )] − [(−2 𝜇 − 0.0833 )(−2 𝜇
− 0.833 ) = 0

 [ 8 𝜇 − 0.664 𝜇  − 0.666 𝜇 + 0.0552  − (4𝜇 + 0.1666𝜇 

+ 0.1666 𝜇  + 0.00693  )] = 0

 4 𝜇 − 2.656 𝜇  − 0.04827  =0

 − 0.04827 − 2.656 𝜇  + 4 𝜇 = 0

 0.04827 + 2.656 𝜇  − 4 𝜇 = 0 … (13)

By solving the quadratic equation (13),

−2.656 ± (2.656 ) − 4(0.04827)(−4  )

=
2(0.04827)

−𝑏 ± √𝑏 − 4 𝑎 𝑐
[∵  = ]
2𝑎

−2.656  ± 5.05  − 0.772 

=
00965
−2.656  ± 0.4119 
= = 27.52 ± 4.268 
0.0965

 =  [−27.52 ± 4.268]

 = 32.69  = 2.59 

Thus the natural frequencies are,

 =𝜔

𝜔 = √

∴𝜔 = 2.59 𝜇
5.100 Application to Heat Transfer and Dynamic Analysis

𝜔 = 1.604 𝜇

𝐸
𝜔 = 1.60
𝜌𝐿

1.60 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠
𝐿 𝜌

𝐸
Similarly, 𝜔 = 32.69 𝜇 = 5.71 𝜇 = 5.71
𝜌𝐿

5.71 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠
𝐿 𝜌

Result: Natural frequencies are,

(1) Using one element,
(a) Lumped Mass matrix

1.414 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠
𝐿 𝜌

(b) Consistent mass Matrix

1.732 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠
𝐿 𝜌

(2) Using two elements

(a) Lumped Mass matrix

3.69 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠
𝐿 𝜌

1.53 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠
𝐿 𝜌
 (b) Consistent mass Matrix

1.60 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠
𝐿 𝜌

5.71 𝐸
𝜔 = 𝑟𝑎𝑑/𝑠
𝐿 𝜌

Example 5.13
For the one-dimensional bar shown in Fig.(i), determine the natural frequencies
of longitudinal vibration using two elements of equal length. Take E = 2 × 10 N/
mm , ρ = 0.8 × 10 N/mm , and L = 400 mm.

Fig. (i)
Given:

Fig. (ii)
For element (1): 𝐿 = 200 𝑚𝑚
𝐴 = 600 𝑚𝑚
For element (1): 𝐿 = 200 𝑚𝑚
𝐴 = 600 𝑚𝑚
𝐸 = 2 × 10 𝑁/𝑚𝑚
𝜌 = 2 × 10 𝑁/𝑚𝑚
To find: Natural frequencies of longitudinal bar.
5.102 Application to Heat Transfer and Dynamic Analysis

Solution: the bar has 2 elements with 3 nodes s shown I Fig.(ii).

We know that,
Stiffness matrix for,
Element (1):
𝐴 𝐸 1 −1
[𝐾 ] =
𝐿 −1 1
600 × 2 × 10 1 −1
=
200 −1 1
1 2
[𝐾 ] = 10 6 −6 1
−6 6 2
Similarly,
𝐴 𝐸 1 −1
𝐄𝐥𝐞𝐦𝐞𝐧𝐭 (𝟐): [𝐾 ] =
𝐿 −1 1

600 × 2 × 10 1 −1
=
2 −1 1

2 3
6 −6 𝟐
[𝐾 ] = 10
−6 6 𝟑
Assemble the stiffness matrix,
1 2 3
6 −6 0 𝟏
[𝐾] = 10 −6 12 −6 𝟐 … (1)
0 −6 6 𝟑
We know that, mass matrix for
Element (1):
𝜌𝐴 𝐿 2 1
[𝑚 ] =
6 1 2

0.8 × 10 × 600 × 200 2 1

=
6 1 2
 1 2
3.2 1.6 𝟏
Similarly, [𝑚 ] = 1.6 3.2
𝟐
Element (2),
𝜌𝐴 𝐿 2 1
[𝑚 ] =
2 1 2
0.8 × 10 × 600 × 200 2 1
=
6 1 2
2 3
[𝑚 ] = 3.2 1.6 𝟐
1.6 3.2 𝟑
Assemble the mass matrix,
1 2 3
3.2 1.6 0 𝟏
[𝑚] = 1.6 6.4 1.6 𝟐 … (3)
0 1.6 3.2 𝟑
Global matrix, for longitudinal bar element,
{[𝐾] − 𝜔 [𝑚]}{𝑢} = 0

6 −6 0 3.2 1.6 0 𝑢
10 −6 12 −6 − 𝜔 1.6 6.4 1.6 𝑢 =0
0 −6 6 0 1.6 3.2 𝑢

Applying boundary conditions,

𝑢 = 0 (𝑓𝑖𝑥𝑒𝑑),
𝑢 =𝑢
𝑢 =𝑢
6 −6 0 3.2 1.6 0 0
10 −6 12 −6 − 𝜔 1.6 6.4 1.6 𝑢 =0
0 −6 6 0 1.6 3.2 𝑢
In the above equation, 𝑢 = 0, so, neglect first row and first column of [K] and
[m] matrix. The final reduced equation is,
 12 −6 6.4 1.6 𝑢
10 − 𝜔 𝑢 =0 … (4)
−6 6 1.6 3.2
To obtain a solution to the set of homogeneous equation in equation (4), we set
the determinant of the coefficient matrix equal to zero.
12 −6 6.4 1.6
10 − 𝜔 =0 … (5)
−6 6 1.6 3.2

(12 × 10 − 6.4𝜔 ) (−6 × 10 − 1.6𝜔 )

 =0
(−6 × 10 − 1.6 𝜔 ) (6 × 10 − 3.2 𝜔 )

 [(12 × 10 − 6.4𝜔 )(6 × 10 − 3.2 𝜔 )] − [(−6 × 10

− 1.6 𝜔 )(−6 × 10 − 1.6 𝜔 )] = 0

 [(7.2 × 10 − 3.84 × 10 𝜔 − 3.84 × 10 𝜔 + 20.48 𝜔 )]

− [(3.6 × 10 + 9.6 × 10 𝜔 + 9.6 × 10 𝜔 + 2.56 𝜔 )] = 0

 3.6 × 10 − 9.6 × 10 𝜔 + 17.92 𝜔 = 0

𝐿𝑒𝑡,  = 𝜔

17.92  − 9.60 × 10 + 3.6 × 10 =0

 − 53.57 × 10  + 2 × 10 =0 … (6)

By solving the quadratic equation (6),

53.57 × 10 ± (53.57 × 10 ) − 4(2 × 10 )

 =
2(1)

−𝑏 ± √𝑏 − 4 𝑎 𝑐
[∵  = ]
2𝑎
  = −40218.683

  = −497281.316
We know that, natural frequencies,
𝜔 =

 𝜔 = √
∴ 𝜔 =  = √40218.683

𝜔 = 200.54 𝑟𝑎𝑑/𝑠

Similarly, 𝜔 =  = √497281.316

𝜔 = 705.181 𝑟𝑎𝑑/𝑠

Result: Natural frequencies of longitudinal are,

𝜔 = 200.54 𝑟𝑎𝑑/𝑠

𝜔 = 705.181 𝑟𝑎𝑑/𝑠