Introduction to

Finite Element Method

Dr Sana Zulfiqar
 Overview
• Governing equations for various physical phenomena are
described by partial differential equations (PDEs)
 Overview

• For simple problems, analytical solutions can be found

• But where geometric and material non-linearities, e.g.
complex components or boundary conditions, then have
to revert to numerical approximation e.g. by Finite
Element Method
• The FE method is a technique to find approximate
solutions to PDEs
• Can be used to solve problems in solid mechanics,
structural mechanics, fluid mechanics, electro-magnetics,
thermo-mechanics, acoustics, even finance problems.
• Will focus on solid mechanics problems in this course
 Overview

• Based on discretization, assembly and

then solution of system equation
• While imposing kinematic constraint
(compatibility), material constitutive
behavior and equilibrium
• Complicated domains can be handled
with ease
• Virtual investigations of design
• or to observe phenomena that cannot
be observed experimentally, e.g. stress
field inside a solid component
 Why Learn FEM?

• Knowledge of how the method works will enhance your

analysis skills and provide better understanding of
numerical analysis in general
• The intelligent use of commercial FE software and
correct interpretation of the output requires basic
knowledge of the method
• Possible to extend commercial software packages by
implementing user defined sub-routines.
• Widespread use in industry for virtual experiments
during design stage.
• Extremely cost effective and fast compared to physical
testing.
 Mechanics

• The field of mechanics can be subdivided in four

main areas

Theoretical
Applied
Mechanics
Computational
Experimental
 Computational Mechanics

• Computational mechanics can be classified based

on the physical domain

Nano and Micromechanics

Continuum Mechanics:
Computational Solids and Structures
Mechanics Fluids
Multiphysics

Systems
Computational Mechanics
 Computational Solid and
Structural Mechanics
• Computational solid and structural mechanics can
be subdivided in two major areas

Computational Statics
Solid and Structural
Mechanics (CSM) Dynamics
 CSM Statics

• Statics may be classified in following manner

Linear
CSM Statics
Nonlinear
 CSM Linear Statics
• A numerical method is required to solve the CSM
linear statics problems using computers
• Several spatial discretization methods are available

Finite Element Method

Finite Difference Method

CSM Linear Statics Boundary Element Method

Finite Volume Method

Mesh Free Method

 CSM Linear Statics by FEM
• For finite element method discretization, choose a
formulation based on primary unknown variable and a
solution method
Displacement

Formulation Equilibrium (Force)

for FEM Mixed
Model
Hybrid

Stiffness
Solution for
Flexibility
FEM Model
Mixed
Summary of What’s Covered in
this Course

• Computational Structural Mechanics

• Linear Statics Problems

• Spatially discretized

• Solved by the stiffness

Sinking of the Sleipner A offshore
platform
• Platform produces oil and gas in the North Sea
• Supported on the seabed at a water depth of 82 m
• Concrete gravity base structure consisting of 24 cells
• Four cells are elongated to shafts supporting the platform
deck
• First concrete base structure for Sleipner A sprang a leak and
sank under a controlled ballasting operation (August 1991)
• The failure involved a total economic loss of about $700
million
• Investigation traced the error to inaccurate finite element
approximation of the linear elastic model
• Shear stresses were underestimated by 47%, leading to
insufficient design
• In particular, certain concrete walls were not thick enough
 Titan Incident 2023

• Went under catastrophic implosion. An implosion is the opposite

of an explosion. In an explosion, the force acts outwards, but in
an implosion the force acts inwards. When a submersible is deep
in the ocean it experiences the force on its surface due to water
pressure. When this force becomes larger than the force hull can
withstand, the vessel implodes violently.
• The key is the design of the hull that protects the vessel against
the large external water pressure that is trying to crush the hull.
Much of the existing technology is based on steel, titanium and
aluminum. The performance of these materials under extreme
stress is well understood.
• However, the Titan’s hull had an experimental design. It used
mostly carbon fibers, which have the advantage of being lighter
than titanium or steel, so Titan could have more space for
passengers. Properties of carbon fibers for deep sea
applications are, however, not that well understood. It can
crack and break suddenly.
Titan 2023
 Finite Element Method
• FEM is a numerical method for solving problems for
engineering and mathematical physics

• Typical classes of problems

• Structural analysis
• Heat transfer
• Fluid flow
• Mass transport
• Electromagnetic potential
• etc.
• Can solve problems involving complicated geometries,
loadings and material properties for which analytical
solutions are not possible
 Finite Element Method

• Finite element formulation results in a system of

simultaneous algebraic equations
• Numerical methods yield approximate values of the
unknowns at discrete number of points in continuum
• Continuum or body is divided into an equivalent system of
smaller bodies or units known as finite elements
• Finite elements are interconnected at points common to two
or more elements known as nodal points or nodes
• This process of representing a body with finite elements is
known as discretisation
• Equations are formulated for each element and then
combined to obtain solution for whole body
Idealization For A Simple
Structure
Physical FEM
 History
• In 1943, Richard Courant proposed breaking a continuous
system into triangular elements
• In 1950s, a team from Boeing demonstrated that complex
surfaces could be analysed with a matrix of triangular
shapes
• The term “finite element” was coined by Dr Ray Clough in
1960s
• In early 1960s, the MacNeal-Schwendle Corporation (MSC)
developed a general purpose FEA code. It had limit of
68000 degrees of freedom
• After completion of NASA contract, MSC continued
development of its own code MSC/NASTRAN. The original
NASTRAN became available to public and formed basis for
dozens of FE packages
• Around the same time, ANSYS, MARC and SAP were
introduced
 History
• In later decade of 1970s, computer aided design (CAD) was introduced

• In 1980s, CAD progressed from 2D drafting tool to a 3D surfacing tool.

The use of FEM and CAD on the same workstation with
interchangeable geometry using IGES or DXF was possible

• By the end of 1990s, PCs became a major force in the analysis. The
technology became accessible to wide majority of engineers

• In 2000s, the commercial FEA codes rapidly developed. User interfaces

became more closer to CAD software. Pre and post analysis became
much more easier. Computing power in form of compute clusters
became available

• The latest trends include multiphysics capabilities, damage and failure

modelling, complex nonlinear material models, XFEM etc. Use of
multicore high end compute clusters has become a norm
Matrix Notation
 Matrix Notation

• Matrices used in FEM to

• Simplify formulation of element stiffness equations
• For purposes of long-hand solutions of various problems
• For programming the method for computers

• Notation used
• Force acting at various points or nodes
Direction of
force
Node
number

• Nodal displacements are given by

• and are nodal displacements in and directions

 Matrix Notation
• Thus the nodal forces and nodal displacements can be
represented in form of column matrices of size n x 1
• The element and global stiffness matrices are represented
by square matrices

[ ] [ ]
𝑘11 𝑘11 ⋯ 𝑘1𝑛 𝐾 11 𝐾 11 ⋯ 𝐾 1𝑛
[ 𝑘 ]= 𝑘21 𝑘22 ⋯ 𝑘2𝑛 𝑘𝑛2 ¿ ⋯ ¿ 𝑘𝑛𝑛¿ [ 𝐾 ]= 𝐾 21 𝐾 22 ⋯ 𝐾 2𝑛 𝐾 𝑛 2¿ ⋯ ¿ 𝐾 𝑛𝑛¿
⋮ ⋮ ¿ ¿ ⋮ ⋮ ¿ ¿
• In structural theory, elements and are often referred to as
stiffness influence coefficients
• The global nodal forces and global nodal displacements are
related through global stiffness matrix
{ 𝐹 }= [ 𝐾 ] { 𝑑 }
 Matrix Notation
{ 𝐹 }= [ 𝐾 ] { 𝑑 }
• The equation is called global stiffness equation and
represents a set of simultaneous equations
• In expanded form

{ }[ ]{ }
𝐹 1𝑥 𝑢1
𝐾 11 𝐾 11 ⋯ 𝐾 1 𝑛
𝐹 1𝑦 𝑣1
= 𝐾 21 𝐾 22 ⋯ 𝐾 2𝑛 𝐾 𝑛2 ¿ ⋯ ¿ 𝐾 𝑛𝑛 ¿
⋮ ⋮
⋮ ⋮ ¿ ¿
𝐹 𝑛𝑧 𝑤𝑛

• Assume a structure to be forced in a displaced configuration

then

𝐹 1 𝑥 =𝐾 11 , 𝐹 1 𝑦 =𝐾 21 , …, 𝐹 𝑛𝑧 =𝐾 𝑛 1
General Steps of
FEM
 General Steps of FEM
• For simplicity, we will consider the structural problem for now
• Engineers seeks to determine displacements and stresses
throughout a structure, which is in equilibrium and subjected
to applied loads
• Two general approaches traditionally associated with FEM
• Force or flexibility method: uses internal forces as unknowns,
result is a set of algebraic equations for determining unknown
forces
• Displacement or stiffness method: uses displacements of nodes
as unknowns
• For computational purposes, displacement or stiffness method
is simpler to formulate for most of the problems and will be
discussed further
 General Steps of FEM

• Another general method used to develop governing

equations for both structural and non-structural
problems is variational method

• Variational method can be based on a number of

principles
• Theorem of minimum potential energy: applies to
materials behaving in a linear-elastic manner
• Principle of virtual work: applies to materials behaving in
linear-elastic as well as in nonlinear manner
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4 Step 1 & 8 are generally

decided by Analyst
• Define the strain / • Derive the element
Steps 2-7 are carried out
displacement and stiffness matrix and
stress / strain equations automatically by computer
relationships program
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4
• Define the strain / • Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
Step 1: Discretise and Select the Element
Types

• Dividing the body into an equivalent system of finite elements with

associated nodes
• Choose most appropriate element type to model most closely the
actual physical behaviour
• Primary Engineering Judgements
• Total number of elements used
• Variation in size and type of elements within a given body
• Elements
• Small enough to give useable results
• Large enough to reduce computational effort
• Small element (and possibly higher order elements) are generally
desirable where the results are changing rapidly such as where
changes in geometry occurs
• Large elements can be used where results are relatively constant
Step 1: Discretise and Select the Element
Types

• The discretised body or mesh is often created with

mesh-generation programs or preprocessors

• Choice of Elements by Analyst

• Physical makeup of the body under actual loading
conditions
• How close to actual behaviour the analyst wants the
results to be?
• Appropriateness of one, two or three dimensional
idealisations is necessary
• Choice of most appropriate element for a particular
problem
 Commonly Used Elements

Two node line element Higher order line element

• Line Elements
• Consists of bar and beam elements
• Have a cross-sectional area but represented by a line
segments
• Cross-sectional area within an element may vary in
some element types
• Often used to model trusses and frame structures
• Simplest line element has two nodes while higher order
elements have three or more nodes
 Commonly Used Elements

• Simple two dimensional elements with corner nodes

(typically used to represent plane stress / strain) and
higher order two dimensional elements with
intermediate nodes along the side
Two Dimensional Elements

• Basic two dimensional elements are loaded by forces in

their own plane (plane stress / plane strain conditions)

• Triangular or quadrilateral

• Simple two dimensional elements have corner nodes

only (linear elements) with straight sides

• There are higher order elements with midside nodes

(called quadratic elements) and curved sides

• Elements can have variable thickness or be constant

• Often used in a wide range of engineering problems

 Commonly Used Elements

• Simple three dimensional elements (typically used to

represent three dimensional stress state) and higher
order three dimensional elements with intermediate
nodes along edges
 Commonly Used Elements

• Simple axisymmetric triangular and quadrilateral

elements used for axisymmetric problems
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4
• Define the strain / • Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
Step 2: Select a Displacement
Function
• Involves choosing a displacement function within each
element

• Defined using nodal values of the element

• Linear, quadratic and cubic polynomials are frequently used

functions as they are simple to use in FEM formulation

• Trigonometric series can also be used

• For a two dimensional element, displacement function is a

function of the coordinates in its plane (e.g. x-y plane)
Step 2: Select a Displacement
Function
• The functions are expressed in terms of nodal
unknowns (in two dimensional problem, in terms of x-y
components)

• Same general displacement function can be used

repeatedly for each element

• Hence, in FEM a continuous quantity such as

displacement throughout the body is approximated by
a discrete model composed of a set of piece-wise
continuous functions defined within each finite domain
or finite element
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4
• Define the strain / • Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
Step 3: Define the Strain / Displacement
and Stress / Strain Relationships

• Necessary for deriving the equations for each finite element

• In case of one dimensional deformation, say x-direction

• Stresses must be related to strains by a constitutive law.

Simplest relationship is given by Hooke’s law
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4
• Define the strain / • Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
 Step 4: Derive the Element
Stiffness Matrix and Equations
Several methods are used for determining the
element stiffness matrix
Direct Equilibrium or Stiffness Method

• Stiffness matrix and element equations relating

nodal forces to nodal displacements are obtained
using force equilibrium conditions
• Most easily adaptable to line or one dimensional
elements
• Can be used to illustrate development for spring,
bar, and beam elements
 Step 4: Derive the Element
Stiffness Matrix and Equations
Work or Energy Methods
• For two and three dimensional problems, it is much easier to
use energy methods to develop stiffness matrix

• Methods used
• Principle of virtual work (using virtual displacements)
• Applicable to any material behaviour
• Can be used when a potential function does not exist
• Principle of minimum potential energy (will be used later)
• Applicable only to elastic materials
• Castigliano’s Theorem
• Applicable only to elastic materials
• All three methods yield identical element equations for
linear-elastic materials
 Step 4: Derive the Element
Stiffness Matrix and Equations
• For extending the FEM outside the stress analysis field, a functional
analogous to the one used with principle of minimum potential
energy is helpful in determining stiffness matrix
• Functional
• A function of another function
• A function that takes functions as arguments
• An integral expression that implicitly contains differential equations that
describe the problem

• Let denote the functional and denote a function then

• A more general form of a functional depending on two

independent variables and

𝜋=∬ 𝐹 ( 𝑥,𝑦,𝑢,𝑣,𝑢,𝑥 ,𝑢,𝑣 ,𝑣,𝑥 ,𝑣,𝑦 ,𝑢,𝑥𝑥 ,…𝑣, 𝑦𝑦 ) 𝑑𝑥𝑑𝑦

 Step 4: Derive the Element
Stiffness Matrix and Equations

Method of Weighted Residuals

• Particularly popular method is Galerkin’s Method

• Yields the same results as the energy methods wherever

applicable

• Especially useful when a functional such as potential energy is

not readily available

• Allows FEM to be applied directly to any differential equation

• Other methods include least squares, collocation, subdomain

weighted residual methods
 Step 4: Derive the Element
Stiffness Matrix and Equations

• Using any of the above methods, we get the equations to

describe the behaviour of an element

{ }[ ]{ }
𝑓1 𝑑1
Vector of 𝑘11 𝑘12 𝑘13 … 𝑘1 𝑛 Vector of
element
𝑓2 𝑑2
𝑘21 𝑘22 𝑘23 … 𝑘2 𝑛 unknown
nodal forces 𝑓3 = … ¿ 𝑘𝑛𝑛 ¿ 𝑑3
𝑘31 𝑘32 𝑘33 … 𝑘3 𝑛 element
⋮ ⋮ nodal dofs
⋮ ¿ ¿ 𝑘𝑛1 ¿
𝑓𝑛 𝑑𝑛
Element Stiffness Matrix
• In compact form
{ 𝑓 }=[ 𝑘 ] { 𝑑 }
• Generalised displacements may include such
quantities as actual displacements, slopes or even
curvatures
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4
• Define the strain / • Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
Step 5: Assemble to Obtain the Global
Equations and Introduce Boundary
Conditions

• The individual element nodal equilibrium equations

generated in Step 4 are assembled into the global nodal
equilibrium equations

• The direct method of superposition (called the direct stiffness

method), based on nodal force equilibrium, can be used

• Implicit in the direct stiffness method is the concept of

continuity or compatibility

• Compatibility means that the structure remains together and

no tears occur anywhere inside the structure

• The final assembled equation can be written as

{ 𝐹 }= [ 𝐾 ] { 𝑑 }
Step 5: Assemble to Obtain the Global
Equations and Introduce Boundary
Conditions

• For most problems, global stiffness matrix is square and

symmetric

• It can be shown that global stiffness matrix is singular i.e. its

determinant is equal to zero

• To remove singularity, boundary conditions (constraints or

supports) are applied so structure remains in place and no
rigid body motion occurs

• Applied known loads are accounted for in the global force

matrix
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4
• Define the strain / • Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
 Step 6: Solve for Unknown Degrees of
Freedom (or Generalised Displacements)

• After modification to account for boundary conditions, a set

of simultaneous algebraic equations is obtained

{ }[ ]{ }
𝐹1 𝑑1
𝐾 11 𝐾 12 𝐾 13 … 𝐾 1𝑛
𝐹2 𝑑2
𝐾 21 𝐾 22 𝐾 23 … 𝐾 2𝑛
𝐹3 = … ¿ 𝐾 𝑛𝑛 ¿ 𝑑 3
𝐾 31 𝐾 32 𝐾 33 … 𝐾 3𝑛
⋮ ⋮
⋮ ¿ ¿ 𝐾𝑛 1 ¿
𝐹𝑛 𝑑𝑛

• Where is the structure total number of unknown nodal

degrees of freedom
• Can be solved for s using elimination method (such as
Gauss’s method) or an iterative method (such as Gauss-
Seidel method)
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4
• Define the strain / • Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
Step 7: Solve for the Element
Strains and Stresses

• For structural stress analysis, important secondary

quantities of strain and stress can be obtained from
displacements

• Typical relationships between strain and

displacements and between stress and strain can
be used
 General Steps of FEM

Step 1 Step 6 Step 7

• Discretise and select • Solve for unknown • Solve for element
the element types degree of freedom strains and stresses
(displacements)

Step 2 Step 5 Step 8

• Select a displacement • Assemble to obtain global • Interpret the results
equations and apply
function boundary conditions

Step 3 Step 4
• Define the strain / • Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
Step 8: Interpret the Results

• Final goal is to interpret and analyse the results for

use in the design / analysis process

• Determination of locations in structures where large

deformations and large stresses occur is generally
important

• Post processor computer programs help the user to

interpret the results by displaying them in graphical
form
 Summary

• Finite element method is a numerical method to

solve complex engineering / physics problems

• The method can be used to solve problems in various

fields

• Versatility of FEM lies in its ability to solve complex

geometries, materials and boundary conditions