Welcome to

Synchronous classes of
Finite Element Method
19EMEC301

Prof Gururaj Fattepur

 Content Hrs
Unit - 1
1. Introduction to FEM:
FEM paradigm : History, present/future, Research, Application, stress at a point, stress components on
arbitrary plane, Equilibrium equations, compatibility equations, Generalized Hook's law, Plane stress 07
and plain strain, principle of minimum potential energy and virtual work, RR method, FEM steps,
Advantages , disadvantages and limitations.
2 Interpolation Functions For General Element Formulation :
Discreatisation process, types of elements, size of elements, location of node, node numbering scheme and
08
mesh requirements in finite element method, Galerkin's methods with numericals, polynomial form of
interpolation functions, convergence requirements.
Unit - 2

3. FEA analysis:
Pascal triangle, shape functions (1D, 2D, LST, CST, Quad, Higher order elements), Stiffness matrix and
08
its properties. Elimination approach, Penalty approach and Thermal effect based practical engineering
problems.
4. Advanced FEA analysis:
Multi-point constraint, Iso-parametric and Axi-symmetric elements. Practical aspects of industrial 07
machine components, Field issues related to structural applications using higher order polynomials.
Unit - 3

5. Post processing techniques:

Validate and interpret the results, Average and Un-average stresses, Special tricks for post processing, 04
Design modification, CAE Reports.
6. Experimental Validation and Data Acquisition:
Strain gauge, Photo elasticity, Load cells, Torque Sensors/Transducers, Dynamic tests, Acceleration test, 04
Fatigue life measurement, Natural Frequency measurements.
 References
Text Books
• K. H. Huebner, D. L. Dewhirst, D. E. Smith and T. G. Byrom, The Finite Element
Method for Engineers, 4th edition, Wiley, New York, 2001.
• T. R. Chandraputala and A. D. Belegundu, Introduction to Finite Elements in
Engineering, Third Edition, Prentice Hall of India, 2004.
• Nitin Ghokale, Practical finite element analysis, Finite to infinite, 2008.
Reference Books:
• Introduction to the Finite Element Method, by N. S. Ottosen and H. Petersson.
Prentice-Hall, Englewood Cliffs, 1992.
• S. S. Rao, Finite Element Method in Engineering , Fourth Edition, Elsevier Publishing,
2007
 Pre-requisite knowledge
• What is stress? Internal resistance to external force
ζ = Force/Area
• Units for stress 1 Pa = 106 N / m2 or = 106 N / mm2
1 Mpa = 1 N / mm2
1 bar = 0.1 N / mm2
• What is Moment of Inertia (I) and polar moment of inertia (J)?
I second moment of area about a plane axis x or y (mm4)
J is second moment of area about perpendicular axis z. (mm4)
• Type of force and deflection formulae.
• Uni-axial stress : Tensile / compression (ζ = F/A)
Bending (ζ = M y / l) , shear stress ηxy = F/A
Torsion stress η = T r / J
• Bi-axial stress: VonMises and Tresca criteria
• Tri-axial stress: 3 principal stresses ηmax = (ζ1 – ζ2) / 2
• Matrices
 This 1st day Presentation Covers
3 videos content from Week 1

Week 1-chunk 1- video Week 1-chunk 2- video Week 1-chunk 3 -video

• FEM Paradigm : History • FEM Advantages • Types of forces
• FEM Present /Future • Tools /Software • Work potential
• FEM Research • Steps involved in FEM • Stress tensor
• FEM applications
• FEM / FEA difference
• FEM/BEM/FVM/FDM
 Introduction to FEA
Methods to Solve Any Engineering Problem

Analytical Method Numerical Method Experimental Method

Classical Approach Mathematical Approach Actual Measurement

100% Accurate Results Approximate, Assumptions Made Time Consuming , Needs expensive setup

Applicable only for Simple problems like Applicable to real life complicated problems Applicable only if physical prototype is
Cantilever , simply supported beams and available
Cylinders etc..

Complete in itself Results can not be believed blindly and Results can not be believed blindly and
must be verified by experimental methods Minimum 2 or more prototypes must be
and Hand Calculations. tested.
Although applicable to simple shaped Finite Element Method: Linear , Nonlinear , -Strain Gauge
geometries only , Analytical methods are Buckling , Thermal, Dynamics & Fatigue - Photo elasticity
considered as Closed form solutions i.e. analysis - Vibration measurement (accelerometers)
100% Accurate Boundary Element Method: Acoustics / - Sensors for Temp & pressure etc…
NVH analysis - Fatigue test
Finite Volume Method: CFD
(Computational Fluid Dynamics) &
Computational Electromagnetic
Finite Difference Method: Thermal & Fluid
Flow analysis (in combination with FVM)
Different Numerical Methods

Finite Element Method (FEM) :

Very Popular Method based upon discretization of component into Finite number of blocks (elements)
Applications : Linear, Nonlinear, Thermal, Dynamics , Buckling and Fatigue Analysis

Boundary Element Method (BEM) :

It‟s a very powerful and efficient technique to solve acoustics and NVH problems
Just like Finite Element Method, it also requires Nodes and Elements but as the name suggest, it considers only
the outer boundary of the domain

Finite Volume Method (FVM) :

All Computational Fluid Dynamics (CFD) soft wares are based upon FVM.
Unit Volume is considered in Finite Volume Method (similar to Elements in Finite Element Method)
Variable properties at nodes are Pressure , Velocity , Area , Mass etc.
It is based on Navier – Stoke equations ( Mass ,Momentum and Energy Conservation equations)

Finite Difference Method (FDM) :

Finite Element and Finite Difference share many common things.
In general, Finite difference Method is described as a way to solve difference equation.
It uses Taylor‟s series to convert differential equation into algebraic equation. Higher order terms neglected.

Is it possible to use all the above listed methods (FEA ,BEM , FVM, FDM)
to solve same problem (say Cantilever problem)?
Answer : YES ! But the difference is in Accuracy achieved , programming ease and time
required to obtain the solution
 History of FEM
• 1940 Basic idea of FEA developed by Aerospace engineers using matrix method
• 1943 variational method by Courant
• 1945 Hrennikoff (field of structural engineering)
• 1947 Levy introduced force method
• 1953 Levy (stiffness method)
• 1954 Argyris and Kesley (Metrix structural analysis)
• 1956 Turner,Clough, Martin,Topper (Stiffness method)
• 1960 Clough ( Finite element, plane problems)
• 1961 Turner( large deflection thermal problems)
• 1965 NASTRAN developed
• 1968 Zinkiewicz ( Visco elasticity problems)
• 1977 Alberto peano developed fiesta a P-code software
• 1982 Probe, a First industrial software implemented by Barna Szabo and Kent myers
• 1987 Mechanica developed by Rasna corp.
• 2001 A. Duster proved P version code is most efficient in plasticity
• 2006 ASME guide for V&V for computational solid mechanics released
• 2008 NASA released standard for models and simulations
• 2012 Barna Szabo and Ricordo Actis introduced simulation goverance
• 2013 Simscale officially released cloud based 3D simulation
FEM Philosophy …in layman terms

The Philosophy of FEA can be explained with a small example such as “

Measuring the Perimeter of a Circle”

If one need to evaluate the perimeter of a circle without using the conventional
formula (2*pi*r), FEA approach is analogous to Dividing the circle into a number of
segments and joining the points using Straight lines

Since it is very easy to measure the length of straight line. Measure the length of
one line and multiply it by No. of lines to get the perimeter.

Approximate results….isn’t it ?

What if we want to achieve more

accurate result?
 Introduction to “Finite Element Analysis”

Concept of Discretization (Meshing)

Concept of FEM is all about Discretization (Meshing) i.e. Dividing a

big structure/component into small discrete Blocks (Nodes and
Element concept)
But why do we do this Meshing ???

No. of Points = ∞ No. of Points = 8

DoF per point = 6 DoF per point = 6
Total No of Equations to be solved Total No of Equations to be solved
=∞*6=∞ = 8 * 6 = 48

From Infinite to Finite…Hence the

Term “Finite Element Method”
Why Finite Element Method ?

FEM is the most widely applied computer simulation method in Engineering.

It is very closely integrated with CAD/CAM applications.

It is very well proven , tested and validated method for simulating any complex
practical scenario in the area of Structural ,Thermal ,Vibration etc..
Are FEA and FEM different ?

Finite Element Analysis (FEA) and Finite Element Method (FEM) both are
one & the same.

FEA is a method/process based upon FEM

Term “FEA” is more popular in industries while “FEM” at Education

centers
 Steps involved in FEM software
 Understanding the Problem (Thermal , Structural, Dynamic etc)
 Element Selection (Solid , Shell , Beams etc)
 Meshing ( Discretization) Pre
 Deciding the Boundary Conditions (Constraints , Connections
Processing
etc..)
 Load Application (Point , Surface , Body loads etc..)

 Solution (Solver , Sub step / Time step , Nonlinearity etc) Solution

 In-Depth study & interpretation of Analysis Results

 Post processing of Results (Deflection , Stress , Strain etc..)
 Report Preparation
Post
 Observation and Conclusion from the Analysis Processing
 Suggestion and Recommendation for Design Changes, if
required.
 Basic steps involved in FEM
1. Discretisation of the continuum
2. Selection of approximating function
3. Formation of elemental matrices. ( ke )
4. Formation of global matrices ( K )
5. Incorporation of boundary conditions
6. Calculation of unknown nodal
displacements
7. Calculation of strain and stress
Available Commercial FEA Tools/Software Packages
 Applications of FEM
Applications can be classified in to three category

1. Equilibrium or steady state or time-independent problems

• Stress distribution in solid mechanics
• Heat flux distribution in Thermal problems
• Pressure distribution in Fluid mechanics problems.
2. Eigen value problems ( Extension of equillibrium problems)
• Natural frequency/buckling load in solid mechanics
• Stability of laminar flow in fluid mechanics
• Resonance characteristics in electrical
3. Transient or time dependent problems
• Time Dependent problems
Application of FEM in Practical Engineering

• Mechanical / Aerospace / Civil Engineering / Automobile Engineering

• Structural Analysis ( Static / Dynamic , Linear / Non-Linear )
• Thermal Analysis ( Steady State / Transient )
• Electromagnetic Analysis
• Geomechanics
• Biomechanics
• etc….
Practical Applications of FEA

• Aerospace Domain

• Automotive Domain
Practical Applications of FEA

• Hi-Tech /Electronics

• Medical Devices
Practical Applications of FEA

and many more ….

Advantages of FEA

• Cost
• Design Cycle time
• No. of Prototypes
• Testing
• Design Optimization
 Total work potential of a body

• Sum of work potential on a body due it

forces exerted on it.
 Types of Forces
1. Internal forces: Acting inside the body Example: Internal stresses
2. External forces: Acted by external agency on the body.
There are three types of external forces.

A) Body forces : These forces are acting on every element of the body .
Unit is N / mm3 .
Example: weight , magnetic force, centrifugal force, inertia force.

Work Potential (WP) is

work done by body force.
B) Surface forces/ Traction : Force acting on the surface of the body.
Unit is N / mm2 Example: Push / Pull , hammering etc.
But for one dimension, it is acting on the length of the body.
Unit is N/mm Example: surface tension, friction, viscous drag

Hence, Force vector is given by

C) Point Load: It is the force acting on a point on the body.

Total work potential of a body is given by

Stress at point
Point is a body in which all dimensions dx, dy, dz
are tending to zero
Stress acting on each positive plane can be divided in to 3
components σxx , σyy, σzz are called Normal
X plane : σxx , τxy , τxz components
Y plane: σyy , τyx , τyz τxy , τxz , τyx , τyz ,τzx , τzy are
called Shear components
Z plane: σzz , τzx , τzy
All these components arranged in matrix form is called as
Stress tensor at pont ( i,j)
 Welcome to
Synchronous classes of
Finite Element Method
19EMEC301

Prof Gururaj Fattepur

 This Presentation Covers
3 videos content from Week 1

Week 1-chunk 4- video Week 1-chunk 5- video Week 1-chunk 6 -video

• Equations of • Steps of FEM • Total Potential Energy
Equilibrium •Problems on spring
•Linear Constitutive
equations
•Plane stress and plain
strain condition
• Problems
 Re-look : Stress tensor

In this stress tensor ηxy = ηyx , ηxz = ηzx , ηyz = ηzy

Hence it is symmetrical about diagonal axis.
If the element in equilibrium ------then there must be
opposite stresses.
if the element is under deformation ---- the stresses on +ve
faces must be differential with stresses on –ve faces.
 Stresses on a typical element

Let the intensity of body forces in x, y, z directions will be X,Y,Z . Hence ,

Total body forces on the element will be force X volume
i) X . dx . dy . dz in x direction
ii) Y . dx . dy . dz in y direction
iii) Z . dx . dy . dz in z direction
 Equations of Equilibrium

In Mechanics, conditions for equilibrium

∑F =0 in x,y,z directions ∑M=0 about x,y,z axes
For ∑Fx = 0 , there are 7 force components on the element
Similarly we have , ∑Fy = 0 , ∑Fz = 0
Equations of equilibrium are,

Stress vector is ,

Corresponding strain vector is,

 Strain displacement equations

Taking displacement components as u,v,w in x,y,z directions respectively

The normal strain in x direction in terms of displacement component is given by

The shear strain in xy plane is given by,

Simplifying to first order accuracy by dropping second order terms, we get

normal strain and shear strain equations as follows
 Hookes law
1. For uni-axial force : Force in one axis Hence y and z
terms in above equations become zero.
Hence, Stress strain
ζxx = Dxx εxx where,
Dxx is proportionality constant known as stiffness constant
for force in x plane in x direction.
In uni-axial case there will be 1 stress and 1 strain
x,y,z axis can also be denoted as 1,2,3 and stiffness coefficient as k
if we denoted the K11 in matrix form it will 1 X 1 dimension.

examples : Tensile test, compression test , torsion test etc.

2. For Bi-axial stress. : Forces applied in x,y directions
Hence there will be 4 stress and strain components
The stress tensor will be ζxx ηxy 0
ηyx ζyy 0
0 0 0

ζI , ζ2 are principle stresses. ζI > ζ2

( principle stresses are stresses on those
Planes on which shear stress will be zero)

Max. principle stress theory: ( for brittle materials)

When ζI = ζyield ( in uni-axial loading) then yield occurs
VonMises theory : (For ductile materials)
When shear strain energy = shear strain energy at yield point, then
yield occurs
3. For tri-axial stress: Forces applied in x,y,z directions
There will be 6 stress components and 6 strain components
Hence the constant of elasticity will be in 6X 6 matrix

This is called as “Generalized Hookes law”

it is symmetrical about diagonal ,hence only 21 material prosperities is
considered for linear elastic Anisotropic materials
For orthotropic materials 4,5,6 planes are symmetrical about 1,2,3
hence only 9 material constants are considered.
Hence for isotropic, homogeneous, linear elastic materials
Normal strains are given by

Where ν = poison's ratio

E= young‟modulus

Shear strains are given by

Similarly we have, Normal stresses

Shear stresses are given by

The material property matrix D is given by
 Plain stress condition
It is defined to be a state of stress in which the normal stress (ζ) and
shear stress (η) directed to perpendicular to the plane are assumed
to be zero.
In this, the third dimension of the body compared to other two is
small.
 Plain strain condition
It is a state of strain in which normal strain (ε) and shear strain (γ)
normal to xy plane are assumed to be zero.
In this, the third dimension of the body compared to other two is very
large.
Differences between plain stress and plain strain
3. Axi-symmetric condition
These are the bodies generated by rotating a line or curve about an axis

µ = ν ( in some books used )

 Problems on plane stress and plain strain

1. Plain stress problem

 Problems on plane stress and plain strain

1. Plain strain problem

Basic Steps in Finite element Method;
1. Discretisation of the continuum
2. Selection of approximating functions
3. Formation of the element stiffness matrix
4. Formation of overall stiffness matrix
5. Formation of the element loading matrix
6. Formation of the overall element loading matrix
7. Formation of the overall equilibrium equation
8. Incorporation of boundary conditions
9. Calculations of unknown nodal displacements
10. Calculation of strain and stresses

SCHOOL OF MECHANICAL ENGINEERING

1. Discretisation of the continuum:
It is to devide the continuum in to smaller regions of finite
dimensions called as “Finite elements”

Node is a point on the element where input or output of

information can be done.
Continuum means a continuous body
There are different shaped elements
2. Selection of approximating functions
It is also known as interpolation model or displacement
function. It is a mathematical function which represent the
variation of the displacement within the element.
There are two types of functions in use
A) Polynomial function:

B) Trigonometric function:
3. Formation of the element stiffness matrix:
This is the process of finding elemental stiffness matrix for
each element . (ke )

4. Formation of overall stiffness matrix.

The overall or global stiffness matrix can be
found out by adding all elemental matrices
k = k1 + k2 + -----+ kn
5 Formation of the element loading matrix (fe )
6 Formation of the overall element loading matrix
This matrix F can be find out by adding all elemental
loading matrices.
7. Formation of the overall equilibrium equation
Over all equilibrium equation can be formulated as

[ k ] [Q] = [F]
dimensions [mXm] [mX1]=[mX1]
8. Incorporation of boundary conditions:
All known values in the problem are filled in these matrices
to reduce the number of unknown. The known values will be
in the form of inputs. Finally
The number of simultaneous equations = no. of unknowns
9. Calculations of unknown nodal displacements.
Solving the simultaneous equations gives unknown values
in the form of nodal displacements.
10 Calculation of strain and stresses: Nodal stress / strains and
global stress / strain can be find out by using these
displacements
 Total Potential Energy in the body

The Function of function is called Functional

Strain energy is given by
For entire body
 Work Potential (WP)
A) Work Potential due to body force

B) Work Potential due to Traction

C) Work potential due to point load

Total work potential is

 Total Potential Energy in the body
For entire body
 Principle of Minimum Potential Energy (MPE)
“ Of all the displacement configuration, a body can
assume which satisfy the compatibility conditions
and boundary conditions , the configuration which
satisfying equilibrium condition is one which will
have the minimum potential energy”

in this body Y attains equillibrium

early as it is having lesser PE
 Problems on springs
Example 1: For the given spring using MPE, determine the nodal
displacements taking F1 = 75 N F2 = 100N
Solution
 Problems on springs
Example 2: For the given spring find the nodal displacements.
 Welcome to
Synchronous classes of
Finite Element Method
19EMEC301

Monday, Tuesday, Wednesday

10.00 A.M to 11.00 AM
 This Presentation Covers
4 videos content from Week 1

Week 1-chunk 7- video Week 1-chunk 8- video Week 1-chunk 9&10 -video
• Limitations of FEM • Applications of MPE • Methods to use MPE
•Steps in Rayleigh-Ritz
• Problems
 Limitations of FEM software
1. Maximum Data storage : The calculations in FEM are carried out in
matrices which include a large data set. And in each iterations the
variables are to be stored in the memory. This requires a large
memory in RAM as well as Hard disk. A good configuration computer
is necessary which cost more.
2. Error of approximation: All numerical methods are based on
approximation. This approximation introduced because of assuming
continuum object in to finite elements, assuming a interpolation
function, Selecting a suitable element for the given shape of the object
etc. There is limit kept for this error.
For pristine materials error < 5%
For Composite materials error <20%
3. Time consuming:
The meshing involve discretization of the continuum. The quality of the
mesh decides the error in the solution. If fine meshing opted for less
error, it needs lot of time. If it contains many iterations , the design
process becomes tedious.

4. Computer configuration: In micro / nano , Bio-medical , Bio-material

applications special capability in configuration required as they involve
high amount of computation and complexity. For this purpose special
FEM software's developed which need higher configurations for the
computing machines
 Applications of Minimum Potential Energy

MPE can be applied in linear displacement problems:

Let a Force (F ) applied on a bar as shown
L
Because of force there is
F
a change in the length in
one direction. This change
L+ δ l
is called displacement
F

In this case, change in

F
dimension in more than
one direction. This change
is called Deformation.

F
Consider a Bar with axial load F
In classic mechanics
A,E, L
δ = PL/AE = FL / AE
F
δ = F / (AE/L) = F/K
L+δl

F
In MPE method
Total PE = Strain energy + Work Potential
= ½ (K ) (x2) - F . X
(- ve sign because work done on the material)
Under minima condition First derivative will be equated to zero
MPE = d (½ (K ) (x2) - F . X ) = 0
dx
K x –F = 0 Hence x= F/K = δ
Hence MPE method can be applied to linear displacement problems
 Using MPE , Two methods are formulated
1. An variational method known as Rayleigh-Ritz method
2. A weighted residual method known as Galerkin‟s method
Rayleigh-Ritz method
In this method , an approximate solution of the following type is
assumed for the field variable.

where, fi (x) are known independent functions called trial functions

defined for the entire domain, ai are unknown parameters to find.
This solution is to be substituted in to the functional, then
functional becomes function of unknown parameter ai . Then potential
energy minimized by using necessary conditions.
Let Potential energy functional is П , then

This yield „n‟ equations with „n‟ number of inknown „ai and fi „ soling
those equations we get unknown values
 General steps for Rayleigh-Ritz method

1. Formulate the PE functional for the given problem

2. Assume a trial displacement function for the given problem, which
should satisfy the boundary conditions.
3. Substitute the displacement function in to PE functional.
4. Minimize the PE functional so as to obtain the equilibrium
condition.
5. Determine the unknown displacements hence the strains and
stresses.
 Problems on Rayleigh-Ritz method

By R-R method, for a bar of C/S are A , young's modulus E, subjected

to uniaxial loading P show that, at a distance of x from fixed end is
U = ( P/AE) x
hence determine the end deflection and stress to which the bar is
subjected to .

Let follow the steps of Rayleigh-Ritz method

1. Formulate the PE functional :

2. Assume a polynomial displacement function

3. Substitute the displacement function in to PE functional

4. Minimize PE functional
5. Determination of displacement ,strain and stress
substituting the value of a1 in previous equation
hence prooved

6. Find the end deflection and stress

Problem no. 2: Using the R-R method find the stress and displacement
at the mid point of the given bar whose A =100mm2 and E = 70Gpa
Assume displacement model to be 2nd order polynomial.

1. Formulate the PE functional

2. Assume a polynomial 2nd order function

3. Substitute the displacement function in to PE functional

4. Minimize PE functional

5. Determination of displacement and stress

Problem no. 3: Using R-R method, determine the maximum deflection
of the beam fixed at both the ends subjected to a uniformly distributed
load P0 over entire length as shown

1. Formulate the PE functional

2. Assume a polynomial function

3. Substitute the displacement function in to PE functional

4. Minimizing the PE functional