Chapter 1

Using ANSYS for
Finite Element
Analysis
Using ANSYS for
Finite Element
Analysis
A Tutorial for Engineers
Volume I
Wael A. Altabey, Mohammad Noori,

and Libin Wang
MOMENTUM PRESS, LLC, NEW YORK

Using ANSYS for Finite Element Analysis: A Tutorial for Engineers,
Volume I
Copyright © Momentum Press®, LLC, 2018.
All rights reserved. No part of this publication may be reproduced, stored

in a retrieval system, or transmitted in any form or by any means—
electronic, mechanical, photocopy, recording, or any other—except for
brief quotations, not to exceed 400 words, without the prior permission
of the publisher.
First published by Momentum Press®, LLC

222 East 46th Street, New York, NY 10017
www.momentumpress.net
ISBN-13: 978-1-94708-320-2 (print)

ISBN-13: 978-1-94708-321-9 (e-book)
Momentum Press Sustainable Structural Systems Collection
Collection ISSN: 2376-5119 (print)

Collection ISSN: 2376-5127 (electronic)
Cover and interior design by Exeter Premedia Services Private Ltd.,

Chennai, India
10 9 8 7 6 5 4 3 2 1
Printed in the United States of America

Abstract
Finite Element Method (FEM) is a well-established technique for analyz-

ing the behavior and the response of structures or mechanical components
under static, dynamic, or thermal loads. Over the past two decades the use
of finite element analysis as a design tool has grown rapidly. Easy to use
commercial software have become common tools in the hands of students
as well as practicing engineers. The objective of this two volume book is
to demonstrate the use of one of the most commonly used Finite Element
Analysis software, ANSYS, for linear static, dynamic, and thermal anal-
ysis through a series of tutorials and examples. Some of the topics and
concepts covered in these tutorials include development of beam, frames,
and grid equations; 2-D elasticity problems; dynamic analysis; and heat
transfer problems. We are hoping these simple, yet, fundamental tutorials
will assist the users with the better understanding of finite element
modeling, how to control modeling errors, the safe use of the FEM in sup-
port of designing complex load bearing components and structures. There
are many good textbooks currently used for teaching the fundamentals of
finite element methods. There are also detailed users manuals available
for commercial software (ANSYS). However, those sources are useful for
advanced students and users. Therefore, there was a need to develop a
tutorial that would supplement a course in basic finite element or can be
used by practicing engineers who may not have the advanced training in
finite element analysis. That is the gap addressed by this book.
Keywords
ANSYS, composite materials, Dynamics, Failure analysis, Fatigue loads,

FEM, optimization, statistics
Contents
List of Figures ix
Preface xi
1 Introduction to Finite Element Analysis 1
1.1 Finite Element Method 1
1.2 Review Topics 15
1.3 General Steps of FEM 26
1.4 The Most Common Finite Element Types 27
1.5 Types of Element Formulation Methods 27
1.6 Derivation of Spring Element Equations Using
Direct Method 28
1.7 Bar Element Formulation Using Direct Method 31
1.8 Examples of Linear FEM 34
1.9 FEA: Modeling, Errors, and Accuracy 48
2 Static Analysis Using ANSYS 53
2.1 Overview of Structural Analysis 53
2.2 Static Analysis Procedure 54
3 Geometric Modeling 65
3.1 Typical Steps Involved in Model Generation
Within ANSYS 65
3.2 Importing Solid Models Created in CAD Systems 66
3.3 Solid Modeling 68
3.4 Tutorial 1: Solid Modeling Using 2D Primitives 69
3.5 Tutorial 2: Solid Modeling Using 3D Primitives 79
viii • Contents
4 Static Analysis Using Line Elements 91

4.1 Tutorial 3: Static Analysis Using Truss Elements 91
4.2 Tutorial 4 (a): Static Analysis Using BEAM Elements 112
4.3 Tutorial 4 (b): Static Analysis Using Beam Elements
with Distributed Load 115
5 Static Analysis Using Area Elements 121
5.1 Tutorial 5: Static Analysis Using Area Elements:
Plane Problem (Bracket) 121
5.2 Tutorial 6: Static Analysis Using Area Elements:
Plane Problem (Wrench) 141
6 Static Analysis Using Volume Elements 147
6.1 Tutorial 7: Static Analysis Using Volume Elements:
Component Design 147
6.2 Tutorial 8: Static Analysis Using Volume Elements:
Assembly Design 153
7 Thermal Stress Analysis 161
7.1 Tutorial 9: Thermal Analysis of Mechanical Structure 161
7.2 Tutorial 10 (a): Thermal-Stress Analysis-Sequential
Coupled Field 168
7.3 Tutorial 10 (b): Thermal-Stress Analysis: Direct-Coupled
Field177
Summary 185
Bibliography 187
Index 189
List of Figures
Figure 1.1. A diagram of the two common branches of the

general modeling solution. 2
Figure 1.2. The thermomechanical stresses in an air-cooled
turbine blade depicted. 3
Figure 1.3. The complex three-dimensional geometry of
the blade along with the combined thermal and
mechanical loadings. 3
Figure 1.4. The mathematical FEM. 5
Figure 1.5. The physical FEM. 6
Figure 1.6. A two dimensional FE model for a gear tooth. 7
Figure 1.7. Linear and nonlinear behavior of the structure. 9
Figure 1.8. Three-dimensional body. 17
Figure 1.9. Equilibrium of elemental volume. 18
Figure 1.10. (a) Plane stress (b) Plane strain. 21
Figure 1.11. Outline of an FE analysis project. 49
Figure 1.12. (a) A tapered bar loaded by axial force P,
(b) Discretization of the bar into four uniform
two-node elements of equal length. 50
Figure 1.13. Lateral midpoint displacement versus time for a
beam loaded by a pressure pulse. The material is
elastic-perfectly plastic. Plots were generated by
various users and various codes. 51
Preface
The finite element method (FEM) is a well-established technique for

analyzing the structural behavior of mechanical components and systems.
In recent years, the use of finite element analysis as a design tool has
grown rapidly. Easy-to-use commercial software have become common
tools in the hands of students, as well as practicing engineers.
The objectives of this work include:
• To teach students the basic concepts in the linear FEM as related to

solving engineering problems in solids and heat transfer.
• To provide students with a working knowledge of finite element
analysis tools and their use in mechanical design.
• The topics covered in this course include: introduction to finite
element; finite element formulation; introduction to a general
FE software (ANSYS); development of beam, frames, and grid
equations; 2D elasticity problems; dynamic analysis; and heat

transfer problems.
Gain insight into appropriate use of finite element modeling, under-

stand how to control modeling errors, benefit from hands-on exercise at
the computer workstation, and understand the safe use of FEM in support
of designing complex load-bearing components and structures.
There are many good textbooks already in existence that cover the
theory of FEMs. Similarly, there are detailed user manuals available
for commercial software (ANSYS). But, these are useful for advanced
students and users. Therefore, there was a need to develop a computer
session manual in line with the flow of the course and utilizing the software
platform available in the department. Students will be able to acquire the
required level of understanding and skill in modeling, analysis, validation,
and report generation for various design problems.
xii • Preface
This work could also be very helpful for the students of senior design
(mechanical system design) and (FEA for large deformation problems). In
addition, it could be used for computer sessions of short courses on stress
analysis techniques and finite element analysis offered by the Mechanical
Engineering department.
After giving a brief introduction to the finite element analysis and
modeling, various guided tutorials have been included in this manual.
Several new tutorials have been developed and others adapted from
different sources including ANSYS manuals in two volumes, ANSYS
workshops and Internet resources. Tutorials have been arranged in each
volume according to the flow of the course and cover topics such as
solid modeling using 2D and 3D primitives available in ANSYS, static
structural analysis (truss, beam, 2D and 3D structures), dynamic analysis
(harmonic and modal analysis), and thermal analysis.
Wael A. Altabey, Assistant Professor, Department of Mechanical
Engineering, Faculty of Engineering, Alexandria University, Alexandria
(21544), Egypt. Postdoctoral follow, International Institute for Urban
Systems Engineering, Southeast University, Nanjing, Jiangsu (210096),
China.
Mohammad Noori, Professor, Department of Mechanical Engineering,
California Polytechnic State University, San Luis Obispo, CA 93405,
USA.
E-mail: wael.altabey@gmail.com
mnoori@outlook.com
Chapter 1
Introduction to Finite
Element Analysis
1.1 Finite Element Method
The field of mechanics can be subdivided into three major areas: theo-
retical, applied, and computational. Theoretical mechanics deals with
fundamental laws and principles of mechanics studied for their intrinsic
scientific value. Applied mechanics transfers this theoretical knowledge
to scientific and engineering applications, especially through the con-
struction of mathematical models of physical phenomena. Computational
mechanics solves specific problems by simulation through numerical
methods implemented on digital computers.
One of the most important advances in applied mathematics in the
20th century has been the development of the finite element method as
a general mathematical tool for obtaining approximate solutions to
boundary-value problems. The theory of finite elements draws on almost
every branch of mathematics and can be considered as one of the richest
and most diverse bodies of the current mathematical knowledge.
1.1.1 Mathematical Modeling of Physical

Systems
In general, engineering problems are mathematical models of physical

situations. Two main goals of engineering analysis are to be able to iden-
tify the basic physical principle(s) and fundamental laws that govern the
behavior of a system or a control volume and to translate those princi-
ples into a mathematical model involving an equation or equations that
can be solved accurately to predict qualitative and quantitative behavior
of the system. The resulting mathematical model is frequently a single
2 • Using ANSYS for Finite Element Analysis
d ifferential equation or a set of differential equations with a set of cor-

responding boundary and initial conditions whose solution should be
consistent with and accurately represent the physics of the system. These
governing equations represent balance of mass, force, or energy. When
possible, the exact solution of these equations renders detailed behavior of
a system under a given set of conditions.
In situations where the system is relatively simple, it may be possible
to analyze the problem by using some of the classical methods learned
in elementary courses in ordinary and partial differential equations. Far
more frequently, however, there are many practical engineering problems
for which we cannot obtain exact solutions. This inability to obtain an
exact solution may be attributed to either the complex nature of governing
differential equations or the difficulties that arise from dealing with the
boundary and initial conditions. To deal with such problems, we resort to
numerical approximations. In contrast to analytical solutions, which show
the exact behavior of a system at any point within the system, numerical
solutions approximate exact solutions only at discrete points, called nodes.
Due to the complexity of physical systems, some approximation must
be made in the process of turning physical reality into a mathematical
model. It is important to decide at what points in the modeling process
these approximations are made. This, in turn, determines what type of ana-
lytical or computational scheme is required in the solution process. Let us
Physical
problem
Simplified model Complicated model
Exact solution
Approximate
for approximate
solution for
model
exact model
FEM approach
Figure 1.1. A diagram of the two common branches of the general modeling
solution.
Introduction to Finite Element Analysis • 3
consider a diagram of the two common branches of the general modeling

solution process as shown in Figure 1.1.
For many real-world problems, the second approach is in fact the
only possibility. For instance, suppose that the aim is to find the thermo-
mechanical stresses in an air-cooled turbine blade depicted in Figure 1.2.
Figure 1.3 shows the complex three-dimensional geometry of the
blade along with the combined thermal and mechanical loadings, which
makes the analysis of the blade a formidable task. Nevertheless, many
powerful commercial finite element packages are available that can be
implemented to perform this task with relative ease.
Figure 1.2. The thermomechanical stresses

in an air-cooled turbine blade depicted.
Temperature
max Equivalent
stress
min
max
min
Figure 1.3. The complex three-dimensional geometry of the blade along

with the combined thermal and mechanical loadings.
1.1.2 Basic Concept of Numerical Methods
The basic concept of these methods is based on the idea of building a

complicated object with simple blocks, or, dividing a complicated object
into small and manageable pieces. Application of this simple idea can be
found everywhere in everyday life, as well as in engineering. Examples
include Lego (kids’ play), buildings, and approximation of the area
of a circle:
Element Si
1 2
Area of one triangle = Si = R sin qi
2
N
1 2  2p 
Area of one circle = ∑Si = R N sin  
i =1 2 N
Where N = total number of triangles (elements). The first step of any

numerical procedure is discretization. This process divides the medium of
interest into a number of small subregions and nodes.
There are two common classes of numerical methods: finite differ-
ence methods and finite element methods (FEMs). With finite difference
methods, the differential equation is written for each node, and the deriv-
atives are replaced by difference equations. This approach results in a set
of simultaneous linear equations. Although finite difference methods are
easy to understand and employ in simple problems, they become difficult
to apply to problems with complex geometries or complex boundary
conditions. This situation is also true for problems with nonisotropic
properties. By contrast, FEM uses integral formulations, rather than dif-
ference equations to create a system of algebraic equations. Moreover,
an approximate continuous function is assumed to represent the solution
for each element. The complete solution is then generated by connect-
ing or assembling the individual solutions, allowing for continuity at the
interelemental boundaries.
Thus, FEM is a numerical analysis technique for obtaining approxi-
mate solutions to a wide variety of engineering problems.
1.1.3 A Brief History of FEM
Finite element analysis (FEA) was first developed in 1943 by R. Courant,

who utilized the Ritz method of numerical analysis and minimization
of variational calculus to obtain approximate solutions to vibration
systems. Shortly thereafter, a paper published in 1956 by M. J. Turner,
R. W. Clough, H. C. Martin, and L. J. Topp established a broader definition
of numerical analysis. The paper centered on the “stiffness and deflection
of complex structures.”
By the early 1970s, FEA was limited to expensive mainframe com-
puters generally owned by the aeronautics, automotive, defense, and
nuclear industries. Since the rapid decline in the cost of computers and the
phenomenal increase in computing power, FEA has been developed to an
incredible precision. Present day super computers are now able to produce
accurate results for all kinds of parameters.
1.1.4 The FEM Analysis Process
A model-based simulation process using FEM consists of a sequence of

steps. This sequence takes two basic configurations depending on the
environment in which FEM is used. These are referred to as the mathe-
matical FEM and the physical FEM.
The mathematical FEM as depicted in Figure 1.4, the centerpiece in
the process steps of the mathematical FEM is the mathematical mode,
which is often an ordinary or partial differential equation in space and
time. Using the methods provided by the variational calculus, a discrete
finite element model is generated from the mathematical model. The
resulting FEM equations are processed by an equation solver, which
provides a discrete solution. In this process, we may also think of an
ideal physical system, which may be regarded as a realization of the
mathematical model. For example, if the mathematical model is the

Verification
Mathematical discretization + solution error
model
FEM
Solution
Physical Discrete
Complicated model
problem solution
Verification
solution error
Figure 1.4. The mathematical FEM.

Mathematical
model
Solution
Physical Discrete
Complicated model
problem solution
Verification
solution error
Simulation error = Modeling + Solution error
Figure 1.5. The physical FEM.
Poisson’s equation, realizations may be a heat conduction problem. In

mathematical FEM, this step is unnecessary and indeed FEM discretiza-
tions may be constructed without any reference to physics.
The concept of error arises when the discrete solution is substituted
in the mathematical and discrete models. This replacement is generically
called verification. The solution error is the amount by which the discrete
solution fails to satisfy the discrete equations. This error is relatively
unimportant when using computers. More relevant is the discretization
error, which is the amount by which the discrete solution fails to satisfy
the mathematical model.
The physical FEM is depicted in Figure 1.5; in the physical FEM pro-
cess, the centerpiece is the physical system to be modeled. The processes of
idealization and discretization are carried out concurrently to produce the dis-
crete model. Indeed, FEM discretizations may be constructed and adjusted
without reference to mathematical models, simply from experimental mea-
surements. The concept of error arises in the physical FEM in two ways,
known as verification and validation. The verification is the same as in the
mathematical FEM: the discrete solution is replaced into the discrete model
to get the solution error. As noted earlier, this error is not generally important.
Validation tries to compare the discrete solution against observation
by computing the simulation error, which combines modeling and solu-
tion errors. As the latter is typically insignificant, the simulation error in
practice can be identified with the modeling error. Comparing the discrete
solution with the ideal physical system would in principle quantify the
modeling errors.
1.1.5 How Does FEA Work?
FEA uses a complex system of points called nodes, which make a grid
called a mesh. Figure 1.6 shows a two-dimensional FE model for a gear
Typical
element
Typical
node
Figure 1.6. A two dimensional FE model for a gear tooth.
tooth. All nodes and elements lie in the plane of the paper. This mesh
is programmed to contain the material and structural properties, which
define how the structure will react to certain loading conditions. Nodes
are assigned at a certain density throughout the material depending on
the anticipated stress levels of a particular area. Regions that will receive
large amounts of stress usually have a higher node density than those that
experience little or no stress. Points of interest may consist of: fracture
point of previously tested material, fillets, corners, complex detail, and
high-stress areas. The mesh acts like a spider web, in that from each node,
there extends a mesh element to each of the adjacent nodes. This web of
vectors is what carries the material properties to the object creating many
elements (theory).
A wide range of objective functions (variables within the system)
is available for minimization or maximization: mass, volume, tempera-
ture, strain energy, stress strain, force, displacement, velocity, accelera-
tion, synthetic (user-defined). There are multiple loading conditions that
may be applied to a system. Point, pressure, thermal, gravity, and centrif-
ugal static loads, thermal loads from solution of heat transfer analysis,
enforced displacements, heat flux and convection, point, pressure, and
gravity dynamic loads. Each FEA program may come with an element
library or one is constructed over time. Some sample elements are: rod
elements, beam elements, plate or shell or composite elements, shear
panel, solid elements, spring elements, mass elements, rigid elements,
and viscous damping elements. Many FEA programs also are equipped
with the capability to use multiple materials within the structure such as:
i sotropic: identical throughout, orthotropic: identical at 90 degrees, general

anisotropic: different throughout.
The basic ingredients of any FEA of structural problems are:
• Geometry
• Material model (constitutive model)
• Loading
• Support or boundary conditions
1.1.6 Types of Structural Analysis
Structural analysis is probably the most common application of the FEM.

Structure is defined as something that is designed to bear load. There-
fore, the term structural (or structure) implies not only civil engineer-
ing structures such as bridges and buildings, but also naval, aeronautical,
and mechanical structures such as ship hulls, aircraft bodies, and machine
housings, as well as mechanical components such as pistons, machine
parts, and tools.
FEA has become a solution to the task of predicting failure due to
unknown stresses by showing problem areas in a material and allowing
designers to see all of the theoretical stresses within. This method of prod-
uct design and testing is far superior to the manufacturing costs, which
would accrue if each sample was actually built and tested. FEA consists
of a computer model of a material or design that is stressed and analyzed
for specific results. It is used in new product design and existing product
refinement. A company is able to verify whether a proposed design will be
able to perform to the client’s specifications prior to manufacturing or con-
struction. Modifying an existing product or structure is utilized to qualify
the product or structure for a new service condition. In case of structural
failure, FEA may be used to help determine the design modifications to
meet the new condition. There are generally two types of analysis that
are used in industry: 2D modeling and 3D modeling. While 2D modeling
conserves simplicity and allows the analysis to be run on a relatively nor-
mal computer, it tends to yield less accurate results. 3D modeling, how-
ever, produces more accurate results while sacrificing the ability to run on
all, but the fastest computers effectively. Within each of these modeling
schemes, the programmer can insert numerous algorithms (functions),
which may make the system behave linearly or nonlinearly. Linear sys-
tems are far less complex and generally do not take into account plastic
deformation. Nonlinear systems do account for plastic deformation, and
many also are capable of testing a material all the way to fracture.
F
Linear
Nonlinear
Figure 1.7. Linear and nonlinear behavior of the structure.
Structural analysis consists of linear and nonlinear behavior of the

structural material, as shown in Figure 1.7. Linear behavior assumes that
the material is not plastically deformed. Nonlinear behavior consists of
stressing the material past its elastic capabilities (into the plastic range).
The stresses in the material then vary with the amount of deformation.
Vibrational analysis is used to test a material against random vibrations,
shock, and impact. Each of these incidences may act on the natural vibra-
tional frequency of the material, which, in turn, may cause resonance and
subsequent failure. Fatigue analysis helps designers to predict the life of a
material or structure by showing the effects of cyclic loading on the spec-
imen. Such analysis can show the areas where crack propagation is most
likely to occur. Failure due to fatigue may also show the damage tolerance
of the material. Heat transfer analysis models the conductivity or thermal
fluid dynamics of the material or structure. This may consist of a steady-
state or transient transfer. Steady-state transfer refers to constant thermal
properties in the material that yield linear heat diffusion.
Thus, some of the common types of structural analysis performed
using FEA are as follows.
1.1.6.1 Static Analysis (Linear and Nonlinear)
A static analysis calculates the effects of steady loading conditions on a

structure, while ignoring inertia and damping effects, such as those caused
by time-varying loads. A static analysis can, however, include steady
inertia loads (such as gravity and rotational velocity) and time-varying
loads that can be approximated as static equivalent loads (such as the
static equivalent wind and seismic loads commonly defined in many
building codes).
Static analysis is used to determine the displacements, stresses,
strains, and forces in structures or components caused by loads that do
not induce significant inertia and damping effects. Steady loading and
response conditions are assumed; that is, the loads and the structure’s
response are assumed to vary slowly with respect to time. The kinds of
loading that can be applied in a static analysis include:
• Externally applied forces and pressures

• Steady-state inertial forces (such as gravity or rotational velocity)
• Imposed (nonzero) displacements
• Temperatures (for thermal strain)
• Fluences (for nuclear swelling)
1.1.6.2 Transient Dynamic Analysis (Linear and Nonlinear)
Transient dynamic analysis (sometimes called time history analysis) is

a technique used to determine the dynamic response of a structure under
the action of any general time-dependent loads. You can use this type of
analysis to determine the time-varying displacements, strains, stresses,
and forces in a structure, as it responds to any combination of static, tran-
sient, and harmonic loads. The time scale of the loading is such that the
inertia or damping effects are considered to be important. If the inertia
and damping effects are not important, you might be able to use a static
analysis instead.
1.1.6.3 Modal Analysis
You use modal analysis to determine the vibration characteristics (natural

frequencies and mode shapes) of a structure or a machine component
while it is being designed. It also can be a starting point for another,
more detailed, dynamic analysis, such as a transient dynamic analysis,
a harmonic response analysis, or a spectrum analysis.
1.1.6.4 Harmonic Analysis
Any sustained cyclic load will produce a sustained cyclic response (a

harmonic response) in a structural system. Harmonic response analysis
gives you the ability to predict the sustained dynamic behavior of your
structures, thus enabling you to verify whether or not your designs will
successfully overcome resonance, fatigue, and other harmful effects of
forced vibrations.
1.1.6.5 Spectrum Analysis
Spectrum analysis is one in which the results of a modal analysis are used
with a known spectrum to calculate displacements and stresses in the
model. It is mainly used in place of a time history analysis to determine
the response of structures to random or time-dependent loading condi-
tions such as earthquakes, wind loads, ocean wave loads, jet engine thrust,
rocket motor vibrations, and so on.
1.1.6.6 Buckling Analysis
Buckling analysis is a technique used to determine buckling loads,

critical loads at which a structure becomes unstable, and buckled

mode shapes—the characteristic shape associated with a structure’s
buckled response.
1.1.6.7 Nonlinear Analysis
If a structure experiences large deformations, its changing geometric

configuration can cause the structure to respond nonlinearly.
Nonlinear stress–strain relationships are a common cause of non
linear structural behavior.
Many factors can influence a material’s stress–strain properties,
including load history (as in elastoplastic response), environmental condi-
tions (such as temperature), and the amount of time that a load is applied
(as in creep response).
1.1.6.8 Contact Problems
Contact problems are highly nonlinear and require significant computer

resources to solve. It is important that you understand the physics of
the problem and take the time to set up your model to run as efficiently
as possible.
Contact problems present two significant difficulties. First, you gen-
erally do not know the regions of contact until you have run the problem.
Depending on the loads, material, boundary conditions, and other factors,
surfaces can come into and go out of contact with each other in a largely
unpredictable and abrupt manner. Second, most contact problems need to
account for friction. There are several friction laws and models to choose
from, and all are nonlinear. Frictional response can be chaotic, making
solution convergence difficult.
In addition to these two difficulties, many contact problems must also
address multi-field effects, such as the conductance of heat and electrical
currents in the areas of contact.
1.1.6.9 Fracture Mechanics
Cracks and flaws occur in many structures and components, sometimes

leading to disastrous results. The engineering field of fracture mechanics
was established to develop a basic understanding of such crack propaga-
tion problems.
Fracture mechanics deals with the study of how a crack or flaw in a
structure propagates under applied loads. It involves correlating analytical
predictions of crack propagation and failure with experimental results.
The analytical predictions are made by calculating fracture parame-
ters such as stress intensity factors in the crack region, which you can
use to estimate crack growth rate. Typically, the crack length increases
with each application of some cyclic load, such as cabin pressurization-
depressurization in an airplane. Further, environmental conditions such
as temperature or extensive exposure to irradiation can affect the fracture
propensity of a given material.
1.1.6.10 Composites
Composite materials have been used in structures for a long time. In recent
times, composite parts have been used extensively in aircraft structures,
automobiles, sporting goods, and many consumer products.
Composite materials are those containing more than one bonded
material, each with different structural properties. The main advantage of
composite materials is the potential for a high ratio of stiffness to weight.
Composites used for typical engineering applications are advanced fiber
or laminated composites, such as fiberglass, glass epoxy, graphite epoxy,
and boron epoxy.
FEA allows you to model composite materials with specialized ele-
ments called layered elements. Once you build your model using these
elements, you can do any structural analysis (including nonlinearities such
as large deflection and stress stiffening).
1.1.6.11 Fatigue
Fatigue is the phenomenon in which a repetitively loaded structure frac-

tures at a load level less than its ultimate static strength. For instance, a
steel bar might successfully resist a single static application of a 300 kN
tensile load, but might fail after 1,000,000 repetitions of a 200 kN load.
The main factors that contribute to fatigue failures include:
• Number of load cycles experienced

• Range of stress experienced in each load cycle
• Mean stress experienced in each load cycle
• Presence of local stress concentrations
A formal fatigue evaluation accounts for each of these factors as

it calculates how used up a certain component will become during its
anticipated lifecycle.
1.1.6.12 Other Types of Engineering Analysis
Other types of engineering that can be solved using FEA are:
• Heat transfer (linear and nonlinear)

• Fluid flow (laminar and turbulent)
• Crash dynamics
• Electromagnetics
• Optimization
• Kinematics
1.1.7 Advantages and Disadvantages
Advantages:
• Model irregularly shaped bodies quite easily

• Handle general load conditions without difficulty
• Model bodies composed of several different materials because the
element equations are evaluated individually
• Handle unlimited number and kinds of boundary conditions
• Vary the size of the elements to make it possible to use small
elements where necessary
• Alter the FE model relatively easily and cheaply

• Corrective measures can be taken before any hardware is built
(also, the number of prototypes can be reduced)
• Physically meaningful to engineers and resemble actual structure
Disadvantages
• Experience and judgment needed

• Requires good computing facility
1.1.8 Areas of Application
1.1.8.1 Structural
• A1 Stress analysis
• A2 Buckling
• A3 Vibration analysis
• A4 …..
1.1.8.2 Nonstructural
• B1 Heat transfer
• B2 Fluid flow
• B3 Distribution of electric or magnetic
• B4 Lubrication
• B5 …..
1.1.8.3 Examples of Multi-physics Applications
• A + B1 Thermal analysis
• B1 + B2 Convective flow
• A + B1 + B2 Manufacturing processes
1.1.8.4 Common Fields of Application
• Mechanical engineering
• Aerospace engineering
• Civil engineering
• Automobile engineering
• Electrical engineering
• Chemical engineering
• Geomechanics
• Biomechanics
1.2 Review Topics
1.2.1 Matrix Operations
Matrix Algebra
• A matrix is an m × n array of numbers arranged in m rows and n
columns.
• m = n A square matrix.
• m = 1 A row matrix.
• n = 1 A column matrix.
• aij Element of matrix a row i, column j
Multiplication of a matrix by a scalar:
[a] = k [c] aij = k cij
Addition of matrices: Matrices must be of the same order (m × n). Add

them term by term:
[c] = [a] + [b] cij = aij + bij
Multiplication of two matrices: If [a] is m × n, then [b] must have n rows:
[c] = [a] [b]
n
cij = ∑aie � bie
e =1
Transpose of a matrix: Interchange of rows and columns:
[aij] = [aji]T
• If [a] is m × n, then [a]T is n × m

• If [a] = [a]T, then [a] is symmetric. [a] must be a square matrix
The identity matrix (or unit matrix) is denoted by the symbol [I]:
[a ][ I ] = [ I ][a ] = [a]
1 0 0 
[ I ] = 0 1 0
0 0 1 
Differentiating a matrix:
d  daij 
dx
[ a] =  
 dx 
Example of differentiating a matrix:
1  a11 a12   x 
U= [x y]   
a22   y 
2  a21
 ∂U 
 ∂x   a a12   x 
 ∂U  = 
11
 
   a21 a22   y 
 ∂y 
Integrating a matrix:
[a ]� dx = ∫a� ij dx
The inverse of a matrix is such that:
[a ][a ]−1 = [ I ]
To find the inverse of a matrix, first find the determinant of the matrix:
|a| = determinant of the matrix [a]
Next, calculate the cofactors of [a]. Cofactors of [aij] are given by:
cij = (−1)i + j d
Where matrix d is the first minor of [aij] and is matrix [a] with row i and
column j deleted.
Finally, determine the inverse
 aij 
−1
=
[c ]T
  a
1.2.2 Elasticity Equations
1.2.2.1 Stress Equilibrium Equations
A three-dimensional body occupying a volume V and having a surface S

is shown in Figure 1.8. Points in the body are located by x, y, and z coor-
dinates. The boundary is constrained on some region, where displacement
is specified. On part of the boundary, distributed force per unit area T,
also called traction, is applied. Under the force, the body deforms. The
deformation of a point (x = [x y z]T) is given by the three components of
its displacement: u = [u v w]T
The distributed force per unit volume, for example, the weight per
unit volume, is the body force vector f given by:
T
f =  f x fy f z 
The body force acting on the elemental volume dV is shown in

igure 1.8. The surface traction T may be given by its component values
F
at points on the surface:
y
σy
x
z
τ yx
τ yz xy
τx
τ zy
τzx τxz
Loaded material body
σz
Figure 1.8. Three-dimensional body.

T
T = Tx Ty Tz 
Examples of traction are distributed contact force and action of

p ressure. A concentrated load P acting at a point i is represented by its
three components:
T
Pi =  Px Py Pz 
i
The stresses acting on the elemental volume dV are shown in

igure 1.9. When the volume dV shrinks to a point, the stress tensor is rep-
F
resented by placing its components in a (3 × 3) symmetric matrix. How-
ever, we represent stress by the six independent components as follows:
s = s x sy sz txy t yz txz 
Where s x , s y , and s z are normal stresses and txy , t yz , and tzx are shear
stresses. Let us consider equilibrium of the elemental volume shown in
Figure 1.9. First, we get forces on faces by multiplying the stresses by
the corresponding areas. Writing ∑ Fzx = 0, ∑ Fy = 0, and ∑ Fz = 0 and
recognizing dV = dx dy dz, we get the equilibrium equations:
∂s x ∂txy ∂txz
+ + + Fx = 0
∂x ∂y ∂z
∂txy ∂s y ∂t yz
+ + + Fy = 0
∂x ∂y ∂z
∂txz ∂t yz s z
+ + + Fz = 0
∂x ∂y ∂z
y σyy
σyx
σyz σxy
σzy σxx
σzx σxz
x
σzz
Figure 1.9. Equilibrium of elemental volume.

1.2.2.2 Strain–Displacement Relationship
Strains can be calculated by differentiating displacement functions.

Differentiation of a function is possible only if it is continuous. There-
fore, the strain–displacement relations are also known as compatibility
equations and are given as follows:
∂u ∂u ∂v
ex = gzy = +
∂x ∂y ∂x
∂v ∂u ∂w
ey = gxz = +
∂y ∂z ∂x
∂w ∂w ∂v
ez = g yz = +
∂z ∂y ∂z
In matrix form:
∂ 
 ∂x 0 0
 
 ∂
 ex   0 0
∂y 
e   
 y   ∂
e   0 0 u 
 z =  ∂z   
 v 
gxy   ∂ ∂
   ∂y 0   w
∂x 
gxz   
g   ∂ ∂
 yz   0
∂z ∂x 
 
0 ∂ ∂
 ∂z ∂y 
1.2.2.3 Stress-Strain Relationships
For linear elastic materials, the stress–strain relations come from the gen-
eralized Hooke’s law. For isotropic materials, the two material properties
are Young’s modulus (or modulus of elasticity) E and Poisson’s ratio ν.
For a three-dimensional case, the state of stress at any point in relation to
the state of strain as follows:
1 − v v v 0 0 0 
  e
s x   v 1 − v v 0 0 0  x 
s   v v 1 − v 0 0 0  e y 
 y   
s   1 − 2v  
 z =  0 0 0 0 0   ez 
 g
txy   2
  xy 
   1 − 2v  
t yz   0 0 0 0 0  g yz 
2 
t    
 zx  1 − 2v  gzx 
 0 0 0 0 0 
 2 
In matrix notation:
{s} = [ D] {e}
Where [D] is known as a stress–strain matrix or material properties matrix
and is given by:
1 − v v v 0 0 0 
 v 1− v v 0 0 0 
 
 v v 1− v 0 0 0 
 
 1 − 2v
E
[ D ] = (1 + v)(1 − 2v)  0 0 0 0 0 
2 
 1 − 2v 
 0 0 0 0 0 
 2 
 1 − 2v 
 0 0 0 0 0 
 2 
1.2.2.4 Special Cases
One dimension: In one dimension, we have normal stress along the x-axis
and the corresponding normal strain. Stress–strain relations are simply to:
{σx}=[E]{εx}
Where [D] = [E]

Plane stress: A thin planar body subjected to in-plane loading on its edge
surface is said to be in plane stress. A ring press-fitted on a shaft as shown
in Figure 1.10(a) is an example. Here, stresses s x , txz, and tzy are set as
zero. The Hooke’s law relations then give us:
P P
z
σx = 0
τxz = 0
τzy = 0
(a)
P P
z
εz = 0
γzx = 0
γyz = 0
(b)
Figure 1.10. (a) Plane stress (b) Plane strain.
 
 s  1 u 0   ex 
x
    
s y  = u 1 0  e y 
    
txy  0 0 1 − u  gxy 
 2 
Plane strain: If a long body of uniform cross-section is subjected to trans-

verse loading along its length, a small thickness in the loaded area, as
shown in Figure 1.10(b), can be treated as subjected to plane strain. Here
e z, gzx, and g yz are taken as zero. Stress s z may not be zero in this case. The
stress–strain relations can be obtained directly as:
 
s  1 u 0   ex 
x
  E   
s y  = u 1 0  e y 
  (1 + v ) (1 − 2v )  1 − u  gxy 
txy  0 0  
 2 
1.2.3 Solution of Set of Linear Algebraic

Equations
1.2.3.1 Cramer’s Rule
Consider a set of linear algebraic equations written as follows:

In matrix notation: [a]{x} = {c}
n
Or, in index notation: ∑ j =1 aij x j = ci
Let matrix  d (i)  be matrix [a] with column i replaced by [c]. Then:
 
d (i )
xi =
a
Example:
Consider a set of three linear algebraic equations given as follows:
− x1 + 3 x2 − 2 x3
2 x1 − 4 x2 + 2 x3
4 x2 + x3
In matrix form:
 −1 3 −2   x1  2 
 2 −4 2   x  = 1 
  2  
 0 4 1   x3  3 
Solution using Cramer’s rule:
−1 3 −2
a = 2 −4 2 = ( −1) ( −4 − 8) − (3) ( 2 − 0) + ( −2) (8 − 0) = 12 − 6 − 16 = 10
0 4 1
d (1) 2 3 −2
1 −41
x1 = = 1 −4 2 = = 4.1
a (−10) −10
3 4 1
d ( 2) −1 2 −2
1 −11
x2 = = 2 1 2 = = 1.1
a (−10) −10
0 3 1
d (3) −1 3 2
1 −14
x3 = = 2 −4 1 = = 1.4
a (−10) −10
0 4 3
1.2.3.2 Inversion
[a ]{ x} = {c}
[a ]−1 [a ]{ x} = [a ]−1 {c}
[ I ]{ x} = [a ]−1 {c}
∴ { x} = [ a ] {c}
−1
Example:
 −1 3 −2   x1  2 
 2 −4 2   x  = 1 
   2    using inversion,
 0 4 1   x3  3 
 x1   1.2 1.1 0.2  2   4.1 

      
∴  x2 x  =  0.2 0.1 0.2  1  =  1.1 
 x   −0.8 −0.4 0.2  3  −1.4 
 3      
1.2.3.3 Gaussian Elimination
General system of n equations with n unknowns:
 a11 a1n   x1   c1 
   =  
    
 an1 ann   xn  cn 
Steps in Gaussian elimination:

1. Eliminate the coefficient of x1 in every equation except the first one.
Select a11 as the pivot element.
• Add the multiple −a21 / a11 of the first row to the second row.
• Add the multiple −a31 / a11 of the first row to the third row.
• Continue this procedure through the nth row
After this step:
 a11 a12 a1n 

 ´ 
 x1   c1 
´
a2 n   x2  c 
´
0 a22
   =  2
    
 `  x   
ann   n  cn 
´
 0 an 2
2. Eliminate the coefficient of x2 in every equation below the second

′
one. Select a22 as the pivot element.
′ ′
• Add the multiple a32 / a22 of the second row to the third row.
′ ′
• Add the multiple a42 / a22 of the second row to the fourth row.
• Continue this procedure through the nth row.
After this step:
 a11 a12 a13 a1n  c

 ´ 
 x1   1 
´ ´  ´ 
0 a22 a23 a2 n   x2  c 
     2 
0 0 a33
''
a3'' n   x3  =  c '' 
 
    
3

 ´   xn   
 0 an 2 an'' 3 ann
''
 cn 
3. Repeat the process for the remaining rows until we have a triangu-
larized system of equations:
 a11 a12 a13 a14 a1n   c1 

 ´ ´ ´ ´   x1   ´ 
0 a22 a23 a24 a2 n   x2   c 
    2 
0 0 a33
''
a34
''
a3'' n   x3   c '' 
   =  3 
0 0 0 a44
'''
a4'''n   x4   c ''' 
4
    
    
 ´ x 
n −1   n  c n −1 
 0 an 2 an'' 3 an''' 4 ann   n 
Solve using back substitution:
cnn −1
xn = n −1
ann
 n 
1
xi =
aii 

a1, n +1 − ∑ air xr 

r = i +1
Example:
 2 2 1   x1  9 
 2 1 0   x  = 4 
  2  
1 1 1   x3  6 
1. Eliminate the coefficient of x1 in every equation except the first one.

Select a11 as the pivot element.
• Select a11 = 2 as the pivot element.
• Add the multiple − a21 / a11 = − 2 / 2 = − 1 of the first row to the
second row.
1
• Add the multiple − a31 / a11 = − = − 0.5 of the first row to the
2
third row.
2 2 1   x1   9 
 0 −1 −1   x  =  −5 
  2  
 0 0 0.5  x3  1.5
2. Eliminate the coefficient of x2 in every equation below the second

′
one. Select a22 as the pivot element (already done in this example).
2 2 1   x1   9 
 0 −1 −1   x  =  −5 
  2  
 0 0 0.5  x3  1.5
3. Solve using back substitution:
x3 =
c3''
=
( 3 2) = 3
a33
''
( 12)
x2 =
( −5 + 3) = 2
−1
x1 =
(9 − 2 ( 2) − 3) = 1
2
1.3 General Steps of FEM
The objective of structural analysis is usually to determine the displace-

ments and stresses throughout the structure, which is in equilibrium and
is subjected to applied loads. At any point in the continuum body, there
are 15 unknowns (three displacements, six stresses, and six strains). To
determine these 15 unknowns, we have 15 equations in three-dimensional
case shown as follows:
Unknowns Unknowns governing equations

Displacements (u,v,w) 3 Stress equilibrium equations 3
Stresses 6 Compatibility equations 6
 ex e y ez gxy g yz gxz 

Strains 6 Stress–strain equations 6
 ex e y ez gxy g yz gxz 

15 15
There are two general approaches associated with the finite element
method to solve the governing equations: force (or flexibility) method and
displacement (or stiffness) method. The force method uses internal forces
as the unknown of the problem, whereas the displacements are the sys-
tem variable in displacement method. The displacement method is more
desirable because its formulation is simpler for most structural analysis
problems. Furthermore, a vast majority of general-purpose finite element
programs have incorporated the displacement formulation for solving
structural problems. Consequently, only the displacement method will be
used throughout this course.
The basic steps involved in any FEA consist of the following:
Preprocessing phase (build the FE model, loads, and constraints)
1. Discretize and select element type.

2. Select a displacement function.
3. Define strain/displacement and stress/strain relationships.
4. Derive element stiffness matrix and equations.
5. Assemble equations and introduce Boundary Conditions (BCs).
Solution phase: (assemble and solve the system of equations)

6. Solve for the unknown degrees of freedom.
Post-processing Phase: (sort and display the results)
7. Solve for element stresses and strains.

8. Interpret the results.
1.4 The Most Common Finite Element Types
The basic idea of FEA is to make calculations at only limited (finite)

number of points and then interpolate the results for the entire domain
(line, surface, or volume). Any continuous object has infinite degrees of
freedom, and it is just not possible to solve the problem in this format.
FEM reduces the degrees of freedom from infinite to finite with the help
of discretization or meshing (nodes and elements).
1D (line) element 2D (plane) element 3D (solid) element
(Spring, truss, beam,

pipe, etc.)
(Membrane, plate,
shell, etc.) (3-D fields, temperature,
displacement, stress,
flow velocity)
1.5 Types of Element Formulation

Methods
The element characteristic matrix has different names in different problem

areas. In structural mechanics, it is called a stiffness matrix; it relates nodal
displacements to nodal forces. There are three important ways to derive an
element characteristic matrix.
1.5.1 Direct (Equilibrium) Method
The direct method is based on physical reasoning. It is limited to very

simple elements (spring, bar, and beam), but is worth studying because it
enhances our physical understanding of FEM. According to this method,
the stiffness matrix and element equations relating nodal forces to nodal
displacements are obtained using force equilibrium conditions for a basic
element, along with force/deformation relationships.
1.5.2 Work or Energy Methods
To develop the stiffness matrix and equations for two- and three-
dimensional elements, it is much easier to apply a work or energy method.
These are based on variational calculus. The variational method is appli-
cable to problems that can be stated by certain integral expressions such
as the expression for potential energy. The principle of virtual work (using
virtual displacement), the principle of minimum potential energy, and
Castigliano’s theorem are methods frequently used for the purpose of der-
ivation of element equations. The principle of virtual work is applicable
for any material behavior, whereas the principle of minimum potential
energy and Castigliano’s theorem are applicable only to elastic materials.
For the purpose of extending, FEM outside the structural stress anal-
ysis field, a functional (a scalar function of other functions) analogous to
the one to be used with the principle of minimum potential energy is quite
useful in deriving the element stiffness matrix and equations.
1.5.3 Methods of Weighted Residuals
Weighted residual methods are particularly suited to problems for which

differential equations are known, but no variational statement or func-
tional is available. For stress analysis and some other problem areas, the
variational method and the most popular weighted residual method (the
Galerkin method) yield identical finite element formulations.
1.6 Derivation of Spring Element

Equations using Direct Method
To understand the FE formulation, we start with the concept “Everything

important is simple.” Figure shows a spring element
1 k 2
xˆ
ƒˆ1x ,dˆ1x ƒˆ2x ,dˆ2x
L
Two nodes: Node 1, node 2

Local nodal displacements: dˆ1x ,dˆ2x (inch, m, mm)
Local nodal forces: ƒˆ ,fˆ (lb, newton)
1x 2x
Spring constant (stiffness) K (lb/in, N/m, N/mm)
The stiffness equation for a single spring element in a local coordinate

system can be written as:
∧  ∧ 
 f 1x   k11 k12   d 1x   ∧   ∧   ∧
 ∧ =   ⇒  f  = k d 
 f   k21 k22   ∧       
 2x  d
 2x 
Step 1: Select the element type

Consider linear spring element subjected to tensile forces
k
2
1 ˆx
L
1 k 2
T
T x̂
dˆ1x dˆ2x
Step 2: Select a displacement function
• Degree of freedom (DOF) per node = 1

• Number of nodes per element = 2
• Total (DOF) per element = 2 × 1 = 2
• Number of coefficients = 2
1
x y
x2 xy y2
x3 x2 y x y2 y3
x4 x3 y x2 y2 x y3 y4
Assume a linear displacement function:
∧ ∧
u = a1 + a2 x
Write in matrix form:
∧  ∧   a1 
u = 1 x   
   a2 
 
Express u^ as a function of nodal displacements (  d 1x , d 2 x  ).
∧ ∧
Apply boundary condition:

∧ ∧ ∧
At x∧ = 0 u∧ = d 1x ∴ u∧ (0) = a1 + a2 ( 0) = d 1x ∴ a1 = d 1x
∧ ∧
∧ ∧ d 2 x − d 1x
At x = L u = d 2 x ∴ u ( L ) = a1 + a2 ( L ) = d 2 x ∴ a1 =
∧ ∧ ∧
L
Substituting values of coefficients:

∧
 ∧ ∧   ∧ 
∧  ∧ ∧  ∧
∧ d 2 x − d 1x  ∧  x ∧ x x x   d 1x 
x = 1 −  d 1x +   d 2 x = 1 −
∧
∴ u = d 1x +   
 L   L  L  L L ∧ 
        d 2 x 
 ∧
∴ u = [ N ] d 
∧
 
∧ ∧
x x
Where N1 = 1 − and N 2 =
L L
N1 and N2 are called shape functions or interpolation functions. They
express the shape of the assumed displacements. The sum of all shape
functions at any point within an element should be equal to 1.
N1 = 1 N2 = 0 at node 1
N1 = 0 N2 = 1 at node 2
N1 + N2 = 1
N1 N2 N1 N2
1 2 1 2
1 2
L
L L
Step 3: Define the strain/displacement and stress/strain relationships

∧ ∧
Deformation, d = u∧ ( L) − u∧ ( 0) = d 2 x − d 1x
 ∧ ∧ 
From the force/deformation relationship: T = k d = k  d 2 x − d 1x 
 
Where T is the tensile force and δ is the total elongation.
Step 4: Derive the element stiffness matrix and equations

Consider the equilibrium of forces for the spring.
∧  ∧ ∧ 
At node 1, f 1x = −T = k  d 1x − d 2 x 
 
At node 2, f
∧
2x
 ∧ ∧ 
= T = k  d 2 x − d 1x 
 
In matrix form,
 ∧  ∧ 
− k   d 1x    ∧ ( e)   ∧( ) 
  ∧ ( e)  
e
 f 1x   k 
 ∧ =   ⇒  f  =  k  d 
 f  −k k   ∧       
 2x  d 2 x       
Note k is symmetric. Is k singular or nonsingular? That is, can we solve

the equation? If not, why?
Step 5: Assemble the element equations to obtain the global equations
and introduce the boundary conditions
(e) 
N ∧
Global stiffness matrix: [ K ] = ∑ e =1  k 
 
e
N  ∧ 
( )
Global load vector: { F } = ∑ e =1  f 
 
∴ { F } = [ K ] {d }
This vector does not imply a simple summation of the element matrices,
but rather denotes that these element matrices must be assembled properly
satisfying compatibility conditions.
Step 6: Solve for nodal displacements
Displacements are then determined by imposing boundary conditions, such
as support conditions, and solving a system of equations, {F} = [K]{d},
simultaneously.
Step 7: Solve for element forces
Once displacements at each node are known, then substitute back into
element stiffness equations to obtain element nodal forces.
1.7 Bar Element Formulation using

Direct Method
Step 1: Select the element type
y
Tˆ x ˆ (force/length)
Cx
ˆy T
x̂, û
2
L dˆ2x , f̂2x
1
θ
T
x
dˆ1x , fˆ1x
Assumptions:
∧ ∧
• The bar cannot sustain shear force, that is, f 1 y = 0, f 2 y = 0.
• Any effect of transverse displacement is ignored.
• Hooke’s law applies, that is, s x = E e x.
• No intermediate applied loads.
Two nodes: Node 1, Node 2

Local nodal displacements: ∧ ∧
d 1x , d 2 x (inch, m, mm)
Local nodal forces: ∧ ∧
f 1x , f 2 x (lbs., Newton)
Length L (inch, m, mm)
Cross-sectional area A (Sq. inch, Sq. m, Sq. mm)
Modulus of elasticity E (psi, Pa, MPa)
The stiffness equation for a single spring element in a local coordinate

system can be written as:
∧  ∧ 
 f 1x   k11 k12   d 1x   ∧   ∧   ∧
 ∧ =   ⇒  f  = k d 
 f   k21 k22   ∧       
 2x  d 2 x 
Step 2: Select a displacement function
• Degree of freedom (DOF) per node = 1

• Number of nodes per element = 2
• Total (DOF) per element = 2 × 1 = 2
• Number of coefficients = 2
Assume a linear displacement function:
∧ ∧
u = a1 + a2 x
Write in matrix form.

∧  ∧   a1 
u = 1 x   
  a2 
∧ ∧
Express u^ as a function of nodal displacements ( d 1x , d 2 x )
Apply boundary condition:
∧ ∧ ∧
At x∧ = 0 u∧ = d 1x ∴ u∧ ( 0) = a1 + a2 ( 0) = d 1x ∴ a1 = d 1x
∧ ∧
∧ ∧ d 2 x − d 1x
At x = L u = d 2 x ∴ u ( L) = a1 + a2 ( L) = d 2 x ∴ a1 =
∧ ∧ ∧
L
Substituting values of coefficients:
∧
 ∧   ∧ 
∧ ∧  ∧  ∧
∧ d 2 x − d 1x  ∧  x ∧  x ∧ x x   d 1x 
d 2 x = 1 −
∧
∴u = d 1x +  x = 1− d 1x +  
 L   L  L  L L ∧ 
        d 2 x 
 ∧
∴ u = [ N ] d 
∧
 
∧ ∧
x x
Where N1 = 1 − and N 2 =
L L
Step 3: Define the strain/displacement and stress/strain relationships
From the definition of strain, the strain/displacement relationship can be
derived as:
 ∧    ∧   ∧  ∧ 
d u  d  d 1x    dN1 dN 2   d 1x   1 1   d 1x 
∴ {e} =  =   [ N1 N2 ]    =   =  − L  
 dx   dx   d∧2 x    dx dx   ∧  L   ∧ 
      d 2 x  d 2 x 
 ∧
∴ {e} = [ B ]  d 
 
Where e is known as strain–displacement matrix.

From Hooke’s law, the stress/strain relationship is: s x = E e x.
Step 4: Derive the element stiffness matrix and equations
To derive the stiffness equation using direct method, we employ equilib-
rium condition at each node.
 ∧ ∧ 
∧ d 2 x − d 1x  AE  ∧ ∧ 
At node 1, f 1x = − T = − As x = − A( E e x ) = − AE  =  d 1x − d 2 x 
 L  L  
 
∧ AE  ∧ ∧ 
At node 2, f 2 x = T =  d 2 x − d 1x 
L  
In matrix form,
∧  ∧ 
 f 1x  AE  1 −1 d 1x   ∧ (e )   ∧ (e )   ∧ (e ) 
 =    ⇒  f  =  k  d 
 f∧  L  −1 1  d∧2 x       
 2x   
Steps 5 to 7 are same as before.
1.8 Examples of Linear FEM
Example 1
For the structure shown in the figure, determine the nodal displacements,
the forces in each element, and the reactions.
2 3
2
1 1 2 3 15 kN 4 E = 210 GPa
3m A = 3 × 10–4 m2
4 5
2
3m
EA
k= the stiffness of element
L
All elements of the previous figure have the same material and
dimensions.
EA  1 −1
k (1) = k ( 2) = k (3) = k ( 4) = k =
L  −1 1 
210 *103 * 300  1 −1

k=  −1 1 
3000  
 21 −21
k=  *10
3
 −21 21 
The global stiffness matrix (K)
K = k1 + k2 + k3 + k4
  21 −21 0 0 0  0 0 0 0 0  0 0 0 0 0  0 0 0 0 0 
  −21 21 0 0 0  0 21 −21 0 0  0 21 0 −21 0  0 21 0 0 −21
 
K = 10 *   0
3
0 0 0 0  + 0 −21 21 0 0  + 0 0 0 0 0  + 0 0 0 0 0 
       
 0 0 0 0 0  0 0 0 0 0  0 −21 0 0 0  0 0 0 0 0 
  0 0 0 0  0 0 0  0 0 0  0 −21 0 21 
 0 0 0 0 0 0
 21 −21 0 0 0 
 −21 84 −21 −21 −21

K = 10 *  0
3
−21 21 0 0 
 
 0 −21 0 21 0 
 0 −21 0 0 21 
The nodal displacement equations:
{F}=[K ]{d}
 F1x   d1x 
F  d 
 2 x   2 x 
 F3 x  = [ K ]  d3 x 
F  d 
 4x   4x 
 F5 x   d5 x 
 F1x   21 −21 0 0 0  0 
15000   −21 84 −21 −21 −21 d 2 x 
    
 F3 x  = 10 *  0 −21 21 0  0 
3
0
 F   
 0 −21 0 21 0  0 
 4x   
 F5 x   0 −21 0 0 21   0 
F1x = −21*103 * d 2 x (1)
15000 = 84 *103 * d 2 x (2)

F3 x = −21*103 * d 2 x (3)
F4 x = −21*103 * d 2 x (4)
F5 x = −21*103 * d 2 x (5)
From equation (2), d2x = 0.17857 mm

From equations (1–3–4–5), the reactions, F1x = F3x = F4x = F5x = –3.75kN
The force in each element, {f} = [k]{d}
The first element between nodal 1, 2:
 f1(x1)  3  21 −21 d1x = 0 

 (1)  = 10 *   
 f 2 x   −21 21   d 2 x 
f1(x1) = −21*103 * d 2 x
f1(x1) = −3.75 kN
f 2(1x) = − f1(x1) = 3.75 kN
The second element between nodal 2, 3:
 f 2(x2)  3  21 −21  d 2 x 
 ( 2)  = 10 *   
 f3 x   −21 21  d3 x = 0 
f 2(x2) = 21*103 * d 2 x
f 2(x2) = 3.75 kN
f3(x2) = − f 2(x2) = −3.75 kN
The third element between nodal 2, 4:

 f 2(x2)  3  21 −21  d 2 x 
 ( 2)  = 10 *   
 f 4 x   −21 21  d 4 x = 0 
f 2(x2) = 21*103 * d 2 x
f 2(x2) = 3.75 kN
f 4(x2) = − f 2(x2) = −3.75 kN
The fourth element between nodal 2, 5:
 f 2(x2)  3  21 −21  d 2 x 
 ( 2)  = 10 *   
 f5 x   −21 21  d5 x = 0 
f 2(x2) = 21*103 * d 2 x
f 2(x2) = 3.75 kN
f5(x2) = − f 2(x2) = −3.75 kN

Example 2
Derive a finite element to solve this problem.
1000 lb
60°
3 2
1000 lb
30° 1
3
4
To derive this problem must be tack the rotational effect by multiply-

ing the local stiffness matrix in rotational matrix (R).
The rotational matrix in this problem can be derived as:
 c2 cs −c 2 −cs 
 
 cs s2 −cs −s2 
= 
2
 −c −cs c2 cs 
 
 − cs −s2 cs s2 
Where: c2 = cos2 (β), s2 = sin2 (β), cs = cos (β)*sin (β).

β = The angle between x-axis and the element C.C.W.
c2 cs s2
β1 = 120° 0.25 −0.433 0.75
β2 = 180° 1 0 0
β3 = 210° 0.75 0.433 0.25
The property of this matrix is a symmetric matrix. This matrix is

multiplied with the local stiffness matrix of each element.
Ei Ai
k (i ) =
Li
[ Ri ] i = 1, 2, 3
Where: [Rj]= [R] when derived with βi .
i =3
[ K ] = ∑k (i ) i = 1, 2, 3
i =1
And, the next steps as the same in example 1.
Example 3
Analyze a thin plate of uniform thickness t = 5 mm using an appropriate
finite element formulated in part (a) and (b). The recommended mesh
for the plate is shown in the figure. Determine the displacement at point
(250, 125) and stresses in the element A only. Use E = 210 GPa and
ʋ = 0.30.
4 3
(4)
(1), A 5
(3) 250 mm
(2)
1 2
500 mm 30°
(a) 40 kN
The governing global matrix equation is:

{F} = [K]{d}
Expanding matrices in global matrix equation, we obtain
 F1x   R1x   d1x   0 

F   d 
 1 y   R1 y 
  0 
 1y   
 F2 x  20000  d2 x  d2 x 
     
 F2 y   34641 d 2 y 
 
d 2 y 
F   R  d   0 
 3x   3x   3x 
 = = [ K ]  = [K ]
 F3 y   R3 y   d3 y 
   
 0 
F   R  d   0 
 4x   4x   4x   
 F4 y   R4 y  d 4 y   0 
d 
 F5 x   0 
 
 d5 x 
   
 5x 
 F5 y   0   d5 y   d5 y 
   
Where [K] is an 10 × 10 before deleting rows and column to account for

the fixed boundary support at nodes 1, 3, and 4.
Assemblage of stiffness matrix
The stiffness matrix for element is:
[k] = tA[B]T [D] [B]
Element (1)
m=4
(1) j=5
i=1
For element (1), we have coordinates and xi = 0, yi = 0, xm = 0, ym = 250,

xj = 250 and yj = 125 because the global axes are set up at node 1, and
1
A= bh
2
1
A = ( 250)( 250) = 31250 mm 2
2
t = 5 mm
The matrix [B] is given by:
 bi 0 bj 0 bm 0
1  
[B] = 2 A  0 gi 0 gj 0 gm 
 
 gi bi gj bj gm bm 
Where: bi = y j − ym = 125 − 250 = − 125

b j = ym − yi = 250 − 0 = 250
bm = yi − y j = 0 − 125 = −125
gi = xm − x j = 0 − 250 = −250
g j = xi − xm = 0 − 0 = 0
gm = x j − xi = 250 − 0 = 250
Therefore, substituting in [B] matrix,
 −125 0 250 0 −125 0 

1 
[ B ] = 62500  0 −250 0 0 0 250 
 −250 −125 0 250 250 −125
For plane stress, the [D] matrix is conveniently expressed here as:
 
1 u 0 
E  
[D] = 2 
u 1 0 
(1 − u ) 
1− u 
0 0 
 2 
With ʋ = 0.3 and E = 210 GPa, we obtain
 1 0.3 0 
210000 
[D] = 0 .3 1 0 
0.91 
 0 0 0.35
 −125 0 −250 
 0
 −250 −125 
 1 0.3 0 
T 210000  250 0 0 
Then, [ B ] [ D ] =   0.3 1 0 
62500(0.91)  0 0 250  
 0 0 0.35
 −125 0 250 
 
 0 250 −125 
 −125 −37.5 −87.5 

 −75 −250 −43.75
 
 250 75 0 
∴ [ B ] [ D ] = 3.6923 
T

 0 0 87.5 
 −125 −37.5 87.5 
 
 75 250 −43.75
The stiffness matrix for element (1) is:
 −125 −37.5 −87.5 

 −75 −250 −43.75
   −125 0 250 0 −125 0 
 250 0  1 
250 
75
[ ]
k = ( 5)( 31250 )( 3 .6923)  ×  0 −250 0 0 0
 0 0 87.5  62500
 −250 −125 0 250 250 −125
 −125 −37.5 87.5 
 
 75 250 −43.75
 37500 20313 −31250 −21875 −6250 1563 

 20313 67969 −18750 −10938 −1563 −57031
 
 −31250 −18750 6250 0 −31250 18750 
[k ] = 9.2308  −21875 −10938 0 21875 21875 −10938


 −6250 −1563 −31250 21875 37500 −20313
 
 1563 −57031 18750 −10938 −20313 67969 
Element (2)
m=5
(2)
i=1 j=2
1
A= bh
2
1
A= (500)(125) = 31250 mm2
2
For element (2), we have coordinates xi = 0, yi = 0, xm = 250, ym = 125,

bi = y j − ym = −125
b j = ym − yi = 125
bm = yi − y j = 0
gi = xm − x j = −250
g j = xi − xm = −250
gm = x j − xi = 500
 −125 0 125 0 0 0 
[ B ] = 62500  0 −250 0 −250 0 500
1 
 −250 −125 −250 125 500 0 
 1 0.3 0 
[ D ] = 0.91 0.3 1 0 
210000 
 0 0 0.35
 −125 0 −250 
 0 −250 −125
  1 0.3 0 
210000  125 −250  
0 
0
Then, [ B ] [ D ] =
T
  0.3 1
62500(0.91)  0 −250 125  
 0 0 0.35
 0 0 500 
 
 0 500 0 
 −125 −37.5 −87.5 

 −75 −250 −43.75
 
 125 37.5 −87.5 
∴ [ B ] [ D ] = 3.6923 
T

 −75 −250 43.75 
 0 0 175 
 
 150 500 0 
 −125 −37.5 −87.5 

 −75 −250 −43.75
   −125 0 125 0 0 0 
 125 37.5 −87.5 
[k ] = (5)(31250)(3.6923)  −75 −250 43.75  × 62500  0 −250 0 −250 0 500
1
   −250 −125 −250 125 500 0 
 0 0 175 
 
 150 500 0 
 37500 20313 6250 −1560 −43750 −18750 

 20313 67970 1560 57030 −21880 −125000 

 6250 1560 37500 −20310 −43750 18750 
[k ] = 9.2308  −1560 57030 −20310 67970 21880 −125000
 
 −43750 −21880 −43750 21880 87500 0 
 
 −18750 −1125000 18750 −125000 0 250000 
Element (3)
j=3
m=5 (3)
i=2
1
A = bh
2
1
A = ( 250)( 250) = 31250 mm 2
2

bi = y j − ym =125
b j = ym − yi =125
bm = yi − y j = − 250
gi = xm − x j = − 250
g j = xi − xm = 250
gm = x j − xi = 0
 125 0 125 0 −250 0 

1 
[ B ] = 62500  0 −250 0 250 0 0 
 −250 125 250 125 0 −250 
 1 0.3 0 
210000 
[D] = 0 .3 1 0 
0.91 
 0 0 0.35
 125 0 −250 
 0 −250 125 
  1 0.3 0 
210000  125 250  
[ B ] [ D ] = 62500(0.91)  0 250 125  0.3 1 0 
T 0
  0 0 0.35
 −250 
0 0 
 
 0 0 −250 
 125 37.5 −87.5

 −75 −250 43.75 

 125 37.5 87.5 
∴ [ B ] [ D ] = 3.6923 
T

 75 250 43.775 
 −250 −75 0 
 
 0 0 −87.5
Then, The stiffness matrix for element (3) is:

 125 37.5 −87.5
 −75 −250 43.75 
  125 0 125 0 −250 0 
 125 37.5 87.5  1  
[k ] = (5)(31250)(3.6923)  75 250 ×
43.75  62500 
0 −250 0 250 0 0 
  −250 125 250 125 0 −250 
 −250 −75 0 
 
 0 0 −87.5
 37500 −20313 −6250 −1563 −31250 21875 

 −20313 67979 1563 −57031 18750 −10938

 −6250 1563 37500 −20313 −31250 −21875
[k ] = 9.2308  −1563 −57031 −20313 67979 −18750 −10938


 −31250 18750 −31250 −18750 62500 0 
 
 21875 −10938 −21875 −10938 0 21875 
Element (4)
m=4 j=3
(4)
i=5
1
A = bh
2
1
A = (500)(125) = 31250 mm 2
2

xj = 500 and yj = 250, because the global axes are set up at node 1, and
bi = y j − ym = 0
b j = ym − yi =125
bm = yi − y j = −125
gi = xm − x j = − 500
g j = xi − xm = 250
gm = x j − xi = 250
 0 0 125 0 −125 0 
1 
[B] = 0 −500 0 250 0 250 
62500 
 −500 0 250 125 250 125 
 1 0.3 0 
[ D ] = 0.91 0.3 1 0 
210000 
 0 0 0.35
 0 0 −500 
 0 −500 0 
  1 0.3 0 
210000  125 250  
0 
0
Then, [ B ] [ D ] =
T
  0.3 1
62500(0.91)  0 250 125  
 0 0 0.35
 −125 0 250 
 
 0 250 −125
 0 0 −175 
 −150 −500 0 

 125 37.5 87.5 
∴ [ B ] [ D ] = 3.6923 
T

 75 250 43.75 
 −125 −37.5 87.5 
 
 75 250 −43.75
 0 0 −175 
 −150 −500 0 
  0 0 125 0 −125 0 
 125 87.5 
[k ] = (5)(31250)(3.6923)  75 250 43.75  × 62500  0 −500 0 250 0 250
37.5 1
   −500 0 250 125 250 125 
 −125 −37.5 87.5 
 
 75 250 −43.75
 87500 0 −43750 −21880 −43750 21880 

 0 250000 −18750 −125000 18750 −125000 

 −43750 −18750 37500 20310 6250 −1560 
[k ] = 9.2308  −21880 −125000 20310 67979 1560 57030 
 
 −43750 18750 6250 1560 37500 −20310 
 
 21880 −125000 −1560 57030 −20310 67970 
The stiffness matrix
 37500 20313 0 0 0 0 −6250 1563 −31250 −21875

 20313 67969 0 0 0 0 −1563−57031 −18750 −10938

 0 0 0 0 0 0 0 0 0 0 
 
 0 0 0 0 0 0 0 0 0 0 
 0 0 0 0 0 0 0 0 0 0 
 k (1)  = 9.2308 ∗  
   0 0 0 0 0 0 0 0 0 0 
 −6250 −1563 0 0 0 0 37500 −20313 −31250 21875 
 
 1563 −57031 0 0 0 0 −20313 67969 18750 −10938
 
 −31250 − 18750 0 0 0 0 −31250 18750 6250 0 
 −21875 −10938 0 0 0 0 21875 −10938 0 21875 
 37500 20313 6250 −1560 0 0 0 0 −43750 −18750 

 20313 67 9 70 1560 57030 0 0 0 0 −21880 −125000 

 6250 1560 37500 −20310 0 0 0 0 −43750 18750 
 
 −1560 57030 −20310 67970 0 0 0 0 21880 −125000 
 0 0 0 0 0 0 0 0 0 0 
( )
 k  = 9.2308 ∗ 
2

   0 0 0 0 0 0 0 0 0 0 
 0 0 0 0 0 0 0 0 0 0 
 
 0 0 0 0 0 0 0 0 0 0 
 
 −43750 − 21880 −43750 21880 0 0 0 0 87500 0 
 −18750 −125000 18750 −125000 0 0 0 0 0 250000 
0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 
 
0 0 37500 −20313 −6250 −1563 0 0 −31250 21875 
 
0 0 −20313 67979 1563 −57031 0 0 18750 −10938
0 0 −6250 1563 37500 −20313 0 0 −31250 −21875
( )
 k  = 9.2308 ∗ 
3

  0 0 −1563 −57031 −20313 67979 0 0 −18750 −10938
0 0 0 0 0 0 0 0 0 0 
 
0 0 0 0 0 0 0 0 0 0 
 
0 0 −31250 18750 −31250 −18750 0 0 62500 0 
0 0 21875 −10938 −21875 −10938 0 0 0 21875 
0 0 0 0 0 0 0 0 0  0
0 0 0 0 0 0 0 0 0  0
 
0 0 0 0 0 0 0 0 0 0 
 
0 0 0 0 0 0 0 0 0 0 
0 0 0 0 37500 20310 6250 −1560 −43750 −18750 
( )
 k  = 9.2308 ∗ 
4

  0 0 0 0 20310 67979 1560 57030 −221880 −125000 
0 0 0 0 6250 1560 37500 −20310 −43750 18750 
 
0 0 0 0 −1560 57030 −20310 67970 21880 −125000 
 
0 0 0 0 −43750 −21880 −43750 21880 87500 0 
0 0 0 0 −18750 −125000 18750 −125000 0 250000 
The total stiffness matrix can be obtained:
[ K ] =  K (1)  +  K (2)  +  K (3)  +  K (4) 
[ K ] = 9.2308 ∗
 75000 40626 6250 −1560 0 0 −6250 1563 −75000 −40625 
 40626 135939 1560 57030 0 0 −1563 −57031 −40630 −135938

 6250 1560 75000 −40623 −6250 −1563 0 0 −75000 40625 
 
 −1560 57030 −40623 135949 1563 −57031 0 0 40630 −135938
 0 0 −6250 1563 75000 0 6250 −1560 −75000 −40625 
 
 0 0 −1563 −57031 0 135958 1560 57030 −40630 −135938
 −6250 −1563 0 0 6250 1560 75000 −40623 −75000 40625 
 
 1563 −57031 0 0 −1560 57030 −40623 135939 40630 −135938
 
 −75000 −40630 −75000 40630 −75000 −40630 −75000 40630 243750 0 
 −40625 −135938 40625 −135938 −40625 −135938 40625 −135938 0 543750 
By applying the governing global matrix equation, we get:
{F} = [K]{d}
 R1x   75000 40626 6250 −1560 0 0 −6250 1563 −75000 −40625   0 
 R   
 1 y   40626 1359 −1563 −57031 −40630 −135938  0 
939 1560 57030 0 0
 
20000   6250 1560 75000 −40623 −6250 −1563 0 0 −75000 40625   2x 
    
 34641  −1560 57030 −40623 135949 1563 −57031 0 0 40630 −135938  2y 
 R   0 −6250 −1560 −75000 −40625   0 
 3x   0 1563 75000 0 6250
 =   
 R3 y   0 0 −1563 −57031 0 135958 1560 57030 −40630 −135938  0 
 R   −6250 −1563 0 0 6250 1560 750000 −40623 −75000 40625   0 
 4x    
 R4 y   1563 −57031 0 0 −1560 57030 −40623 135939 40630 −135938   0 
  −75000 −40630 −75000 40630 −75000 −40630 −75000 40630 243750 0  
0 5x
    
−40625 −135938 40625 −135938 −40625 −135938 40625 −135938
 0  
0 543750   5y 

The previous governing global matrix equation has 10 equations with 10

unknowns, which are (R1x, R1y, R3x, R3y, R4x, R4y, d2x, d2y, d5x, d5y); we can get
them by solving the following equations.
By using MATLAB, we can solve them and the results are as follows:
R1x = − 7.6722 kN , R1 y = 13.675 kN , R1 = 15.680 kN
R3 x = − 13.483 kN , R3 y = − 42.210 kN , R3 = 44.311 kN
R4 x = − 4.4911 kN , R4 y = − 6.1025 kN , R4 = 7.5770 kN
d 2 x = 0.0632 mm, d 2 y = 0.0513 mm
d5 x = 0.0109 mm, d5 y = 0.0081 mm
The displacement at point (250, 125) = (0.0109, 0.0081) mm

We can determine the stresses in each element by using equation:
{s} = [ D ][ B ]{d }
The stress for element A, we then have
 d1x 
  d 
1 u  1y 
0   b1 0 b5 0 b4 0 
E   1   d5 x 
{s} = u 1 0 × g1 g5 g4   
(1 − u )
0 0 0
2 2A  d5 y
 1− u   g1 b1 g5 b5 g4 b4   
0 0  d 
 2   4x 
d 4 y 
Substituting numerical values for matrix [D], [B] given by the analysis of
element A and the appropriate part of {d}, we can obtain:
 0 
 0 
 1 0.3 0   −125 0 250 2 −125 0  
210000  × 0  0.0109 
{s} = 0 .3 1 0 −250 0 0 0 250 
 0.0081
0.91( 62500)   
 0 0 0.35  −250 −125 0 250 250 −125  
 0 
 
 0 
 s  10.0615
 x   
 s y  =  3.0185  Mpa
   2.6169 
txy   
1.9 FEA: Modeling, Errors, and Accuracy
Modeling is the simulation of a physical structure or physical process by

means of a substitute analytical or numerical construct. It is not simply
preparing a mesh of nodes and elements. Modeling requires that the phys-
ical action of the problem be understood well enough to choose suitable
A problem Anticipate A problem Preprocess

must be physical must be prepare the
solved behavior solved FE model
plan how
FE results
will be Plan revised
Is FE Yes FE model
checked to
analysis using insight
see if they Solve
required? provided by
are equations of
reasonable the current the FE model
FE model
Analytical or
experimental No
solution
Postprocess
Are results reasonable?
display FE
Yes Are error estimates small?
results
Stop Does model revision do little
to alter computed results? Computer
software
Figure 1.11. Outline of an FE analysis project.
kinds of elements, and enough of them, to represent the physical action

adequately. Figure 1.11 gives an outline of an FEA project.
1.9.1 Modeling Error
Whatever the analysis method, we do not analyze the actual physical

problem; rather, we analyze a mathematical model of it. Thus, we intro-
duce modeling error. For example, in the elementary beam theory, we
represent a beam by a line (its axis) and typically ignore deformations
associated with transverse shear. This is an excellent approximation for
slender beams, but not for very short beams. Or, for axial load problem
of Figure 1.12, we would probably assume that a state of uniaxial stress
prevails throughout the bar, which is proper if taper is slight, but improper
if taper is pronounced. Real structures are not so easily classified, as they
are often built of parts that would be idealized mathematically in differ-
ent ways and have cutouts, stiffeners, and connectors whose behavior
is uncertain.
The foregoing considerations must be addressed in order to decide what
types of elements to use and how many of them. If a beam is deep, trans-
verse shear deformation may become important and should be included
in beam elements. If a beam is very deep, two- or three-dimensional ele-
ments are more appropriate than beam elements. If a beam has a wide
cross-section, the plate theory may be more appropriate than beam theory
(then, of course, choose plate elements rather than beam elements). If an
hO
hO + ∆h P
x x, u
LT 4@L = LT
(a) (b)
Figure 1.12. (a) A tapered bar loaded by axial force P, (b) Discretization of the
bar into four uniform two-node elements of equal length.
axisymmetric pressure vessel has a thick wall, one should regard it as a

solid of revolution, rather than a shell of revolution, and choose axisym-
metric solid elements, rather than axisymmetric shell elements.
1.9.2 Discretization Error
Let us now consider the axially tapered bar of Figure 1.4 in more detail
and describe how the FEM implements the mathematical model. We will
assume that a satisfactory mathematical model is based on a state of uniax-
ial stress. An analytical solution is then rather easy, but we pretend not to
know it and ask for an FE solution instead. We discretize the mathematical
model by dividing it into two node elements of constant cross-section, as
shown in Figure 1.4b. Each element has length L, accounts only for a con-
stant uniaxial stress along its length, and has an axial deformation given
by the elementary formula PL/AE. For each element, A may be taken as
constant and equal to the cross-sectional area of the tapered bar at an x
coordinate corresponding to the element center. The displacement of load
P is equal to the sum of the element deformations. Intuitively, we expect
that the exact displacement be approached as more and more elements
are used to span the total length LT. However, even if many elements are
used there is an error, known as discretization error, which exists because
the physical structure and the mathematical model each has infinitely
many degrees of freedom (DOF) (namely, the displacements of infinitely
many points), while the FE model has a finite number of DOF (the axial
displacements of its nodes).
How Many Elements Are Enough?
Imagine that we carry out two FEAs, the second time using a more refined
mesh. The second FE model will have lesser discretization error than the
first, and will also represent the geometry better if the physical object has
curved surfaces. If the two analyses yield similar solutions, we suspect

that results are not much in error. Or, we might establish a sequence of
solutions by solving the problem more than twice, using a finer mesh each
time. By studying how the sequence converges, we may be able to state
with some confidence that the results from the finest mesh are in error by
less than, say, 5 percent.
1.9.3 Numerical Error
After the analyst has introduced modeling error and discretization error, the
computer introduces numerical error by rounding or truncating numbers
as it builds matrices and solves equations. Usually, the numerical error is
small, but some modeling practices can greatly increase it.
FE computer programs have become widely available, easier to use,
and can display results with attractive graphics. Even an inept user can
produce some kind of answer. It is hard to disbelieve FE results because of
the effort needed to get them and the polish of their presentation. But, any
model, good or bad, can produce smooth and colorful stress contours. It
is possible that most FEAs are so flawed that they cannot be trusted. Even
a poor mesh, inappropriate element types, incorrect loads, or improper
supports may produce results that appear reasonable on casual inspection.
200 mm 20 mm
1-ms presuure pulse 4 mm
10
3c
2
3c
5 2 9
5
Displacement, mm
7
10
3b 9
10
0 2
7
8 5
1
1 9
7 3b
–5 8
5 8
1
10
–10
0 1 2 3 4 5 6 7
Figure 1.13. Lateral midpoint displacement versus time for a beam loaded by a
pressure pulse. The material is elastic-perfectly plastic. Plots were generated by
various users and various codes [2].
A poor model may have defects that are not removed by refinement of
the mesh.
A responsible user must understand the physical nature of the
problem and the behavior of finite elements well enough to prepare a
suitable model and evaluate the quality of the results. Competence in
using FE for stress analysis does not imply competence in using FE for
(say) magnetic field problems. The engineer who uses the software, not
the software vendor, even if the results are affected by the errors in the
software, takes responsibility for the results produced.
Figure 1.13 is an example of discrepancies that may appear. A pressure
pulse is applied to a straight beam with hinge supports. The loading causes
the material to yield and the beam to vibrate. Analysis seeks to track the lat-
eral displacement of the midpoint as a function of time. The results plotted
come from 10 reputable analysis codes and were obtained by users regarded
as expert. Yet, if any of the curves is correct, we c annot tell which one it is.
Admittedly, the problem is difficult. The results indicate strong sensitivities
of both physical and computational nature. This example reminds us that
any analysis program is based on theory and approximation, and that a user
may push the program beyond its range of validity.
Index
A deformation, 132
addition of matrices, 15 element material properties,
algebraic equation. See linear 126
algebraic equations geometric properties,
ANSYS. See also geometric 125–126
modeling; static analysis listing the stresses, 130–132
model geometry loads, applying, 128–129
analysis type and analysis meshing, 126–127
options, 61–62 saving, 127
applying loads and obtaining solving, 129–130
solution, 61, 62–63 stresses, 133–134
initiate solution, 63 title, giving, 122
load step options, 63 type of element, 125
preprocessing plane problem (wrench)
analysis title, 55 ANSYS solution, 142–146
element real constants, 57–59 geometry, 142
element types, 56–57 loading, 142
jobname, 54–55 material properties, 142
linear material properties, methodology, 142
59–60 assembly design (static analysis)
material properties, 59 ANSYS solution, 154–160
nonlinear material properties, boundary condition, 154
60–61 geometry, 154
units, 55 loading, 154
results, 63–64 material properties, 154
applied mechanics, 1 methodology, 154
area elements, static analysis
plane problem (bracket), B
121–122, 134–141 bar element formulation, 31–33
analysis type, 127 beam elements, static analysis.
Boolean operations, 122–124 See also distributed loads on
constraints, 127–128 beam elements
deflection, 133 analysis type, 113
190 • Index
constraints, 113 distributed loads on beam

deformation, 115 elements. See also beam
element density, 113 elements, static analysis
element material properties, 113 analysis type, 118
gravity, 113–114 constraints, 118
keypoints, 112 deformation, 119–120
lines, creating, 113 element material properties, 117
meshing, 113 element types, 116
preprocessor menu, 112 keypoints, 116
real constants, 113 lines, 116
solving, 115 loads, applying, 118–119
title, giving, 112 meshing, 117
type of element, 113 plot elements, 117–118
bracket (plane problem). See area preprocessor menu, opening, 116
elements, static analysis real constants, 117
buckling analysis, 11 solving, 119
title, giving, 116
C dynamic analysis. See transient
CAD model, 66–68 dynamic analysis
compatibility equation, 19
component design (static analysis) E
ANSYS solution, 148–153 elasticity equations
constraints, 147 one dimension, 20
geometry, 147 plane strain, 21
loading, 147 plane stress, 20–21
material properties, 147 strain-displacement relationship,
methodology, 148 19
composites, 12 stress equilibrium equations,
computational mechanics, 1 17–18
contact problems, 11–12 stress-strain relationships, 19–20
Cramer’s rule, 22–23 elementary beam theory, 49
element characteristic matrix
D direct method, 27
differentiating a matrix, 16 variational method, 28
direct-coupled field (thermal weighted residual methods, 28
analysis) element real constants, 57–59
ANSYS solution, 177–184 engineering analysis, 13
loads and material properties,
177 F
direct (equilibrium) method, 27 fatigue, 13
bar element formulation, 31–33 fatigue analysis, 9
spring element equations, 28–31 FEA program, 7–8
discretization error, 50–51 finite difference method, 4
displacement/stiffness method, 26 finite element analysis (FEA)
Index • 191
discretization error, 50–51 L

history of, 5 linear algebraic equations
modeling error, 49–50 Cramer’s rule, 22–23
numerical error, 51–52 Gaussian elimination, 23–26
finite element method (FEM) inversion, 23
bar element formulation, linear behavior, 9–10
31–33 linear material properties, 59–60
description, 1 linear vs. nonlinear static analysis,
element characteristic matrix, 54
27–28 load step options, 63
examples, 34–48
FEA, 6–8 (see also finite element M
analysis (FEA)) material properties, 59
mathematical model, 1–3, 5–6 mathematical model, 1–3, 5–6
numerical methods, 4 matrix algebra, 15–17
physical, 6 mechanical structure (thermal
spring element equations, analysis)
28–31 ANSYS solution, 162–168
steps of, 26–27 loads and material properties,
structural analysis (see structural 161
analysis of FEM) modal analysis, 10
types, 27 modeling, 48–49
force / flexibility method, 26 modeling error, 49–50
fracture mechanics, 12 multi-physics application, 14
multiplication of a matrix by a
G scalar, 15
Galerkin method, 28 multiplication of two matrices, 15
Gaussian elimination, 23–26
geometric modeling N
solid model, 68–69 nonlinear analysis, 11
steps in, 65–66 nonlinear behavior, 9–10
3D, 79–89 nonlinear material properties,
2D, 69–78 60–61
using CAD system, 66–68 nonstructural application, 14
numerical error, 51–52
H numerical methods, 4
harmonic analysis, 10
heat transfer analysis, 9 O
one dimension elasticity equation,
I 20
identity matrix, 16
integrating a matrix, 16 P
inverse of a matrix, 16 physical FEM, 6
inversion, 23 plane strain elasticity equation, 21
192 • Index
plane stress elasticity equation, linear material properties,

20–21 59–60
post-processing phase, 27 material properties, 59
preprocessing phase, 26 nonlinear material properties,
60–61
S units, 55
sequential-coupled field (thermal results, 63–64
analysis) steady-state transfer, 9
ANSYS solution, 169–177 stiffness matrix, 27
loads and material properties, strain-displacement relationship,
169 19
solid modeling, 68–69 stress equilibrium equations,
3D, 79–89 17–18
2D, 69–78 stress-strain relationships, 19–20
solution phase, 26 structural analysis, 53
spectrum analysis, 10 structural analysis of FEM, 8–9
spring element equations, 28–31 advantages, 13–14
static analysis, 9–10 application, 14–15
definition of, 53–54 buckling analysis, 11
linear vs. nonlinear, 54 composites, 12
loadings in, 54 contact problems, 11–12
using area elements (see area disadvantages, 14
elements, static analysis) engineering types, 13
using beam elements (see beam fatigue, 13
elements, static analysis) fracture mechanics, 12
using truss elements (see truss harmonic analysis, 10
elements, static analysis) modal analysis, 10
using volume elements (see nonlinear analysis, 11
volume elements, static spectrum analysis, 10
analysis) static analysis, 9–10
static analysis procedure transient dynamic analysis, 10
model geometry structural application, 14
analysis type and analysis
options, 61–62 T
applying loads and obtaining theoretical mechanics, 1
solution, 61, 62–63 thermal analysis
initiate solution, 63 direct-coupled field
load step options, 63 ANSYS solution, 177–184
preprocessing loads and material properties,
analysis title, 55 177
element real constants, mechanical structure
57–59 ANSYS solution, 162–168
element types, 56–57 loads and material properties,
jobname, 54–55 161
Index • 193
sequential-coupled field solving, 104–105

ANSYS solution, 169–177 title, giving, 91–92
loads and material properties, 2D model, 8, 69–78
169
3D model, 8, 79–89 V
time history analysis. see transient variational method, 28
dynamic analysis vibrational analysis, 9
transient dynamic analysis, 10 volume elements, static analysis
transpose of a matrix, 15 assembly design
truss elements, static analysis ANSYS solution, 154–160
analysis type, 101 boundary condition, 154
axial stress, 110–111 geometry, 154
constraints, applying, 101–103 loading, 154
deflection, 107–109 material properties, 154
deformation, 106–107 methodology, 154
element material properties, 98 component design
forming lines, 94–95 ANSYS solution, 148–153
geometric properties, 96–97 constraints, 147
keypoints, 92–94 geometry, 147
listing, 109–110 loading, 147
listing the stresses, 111–112 material properties, 147
loads, applying, 103–104 methodology, 148
meshing, 95–96, 99, 100
plot numbering, 100 W
quitting ANSYS, 112 weighted residual methods, 28
reaction forces, 106 wrench (plane problem). See area
saving, 100–101 elements, static analysis

Chapter 1

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 1

Uploaded by

Copyright:

Available Formats

Using ANSYS for

Wael A. Altabey, Mohammad Noori,

MOMENTUM PRESS, LLC, NEW YORK

Copyright © Momentum Press®, LLC, 2018.

All rights reserved. No part of this publication may be reproduced, stored

First published by Momentum Press®, LLC

ISBN-13: 978-1-94708-320-2 (print)

Momentum Press Sustainable Structural Systems Collection

Collection ISSN: 2376-5119 (print)

Cover and interior design by Exeter Premedia Services Private Ltd.,

Printed in the United States of America

Finite Element Method (FEM) is a well-established technique for analyz-

ANSYS, composite materials, Dynamics, Failure analysis, Fatigue loads,

4 Static Analysis Using Line Elements 91

Figure 1.1. A diagram of the two common branches of the

The finite element method (FEM) is a well-established technique for

• To teach students the basic concepts in the linear FEM as related to

Gain insight into appropriate use of finite element modeling, under-

1.1 Finite Element Method

1.1.1 Mathematical Modeling of Physical

In general, engineering problems are mathematical models of physical

d­ ifferential equation or a set of differential equations with a set of cor-

Simplified model Complicated model

consider a diagram of the two common branches of the general modeling

Figure 1.2. The thermomechanical stresses

Figure 1.3. The complex three-dimensional geometry of the blade along

1.1.2 Basic Concept of Numerical Methods

The basic concept of these methods is based on the idea of building a

Where N = total number of triangles (elements). The first step of any

1.1.3 A Brief History of FEM

Finite element analysis (FEA) was first developed in 1943 by R. ­Courant,

1.1.4 The FEM Analysis Process

A model-based simulation process using FEM consists of a sequence of

Figure 1.4. The mathematical FEM.

Figure 1.5. The physical FEM.

Poisson’s equation, realizations may be a heat conduction problem. In

1.1.5 How Does FEA Work?

Figure 1.6. A two dimensional FE model for a gear tooth.

i­ sotropic: identical throughout, orthotropic: identical at 90 degrees, general

1.1.6 Types of Structural Analysis

Structural analysis is probably the most common application of the FEM.

Figure 1.7. Linear and nonlinear behavior of the structure.

Structural analysis consists of linear and nonlinear behavior of the

1.1.6.1 Static Analysis (Linear and Nonlinear)

A static analysis calculates the effects of steady loading conditions on a

• Externally applied forces and pressures

1.1.6.2 Transient Dynamic Analysis (Linear and Nonlinear)

Transient dynamic analysis (sometimes called time history analysis) is

1.1.6.3 Modal Analysis

You use modal analysis to determine the vibration characteristics ­(natural

Any sustained cyclic load will produce a sustained cyclic response (a

1.1.6.6 Buckling Analysis

Buckling analysis is a technique used to determine buckling loads,

If a structure experiences large deformations, its changing geometric

Contact problems are highly nonlinear and require significant computer

Cracks and flaws occur in many structures and components, sometimes

Fatigue is the phenomenon in which a repetitively loaded structure frac-

• Number of load cycles experienced

A formal fatigue evaluation accounts for each of these factors as

1.1.6.12 Other Types of Engineering Analysis

Other types of engineering that can be solved using FEA are:

• Heat transfer (linear and nonlinear)

1.1.7 Advantages and Disadvantages

• Model irregularly shaped bodies quite easily

4 Static Analysis Using Line Elements 91

Figure 1.1. A diagram of the two common branches of the

1.1.1 Mathematical Modeling of Physical

d ifferential equation or a set of differential equations with a set of cor-

Finite element analysis (FEA) was first developed in 1943 by R. Courant,

i sotropic: identical throughout, orthotropic: identical at 90 degrees, general

You use modal analysis to determine the vibration characteristics (natural

1.2.3 Solution of Set of Linear Algebraic

 a11 a12 a1n 

 a11 a12 a13 a1n  c

 a11 a12 a13 a14 a1n   c1 

1.5 Types of Element Formulation

1.6 Derivation of Spring Element