Prem K.

Kythe
Dongming Wei

An Introduction
to Linear and Nonlinear
Finite Element Analysis
A Computational Approach

With 152 illustrations

Springer Science+Business Media, LLC

Prem K. Kythe Dongming Wei
Department of Mathematics Department of Mathematics
University of New Orleans University of New Orleans
New Orleans, LA 70148-0001 New Orleans, LA 70148-0001
U.S.A. U.S.A.

Library of Congress Cataloging-in-Publication Data

Kythe, Prem K.
An introduction to linear and nonlinear finite element analysis : a computational approach
Iby Prem K. Kythe, Dongming Wei.
p. cm.
Inc\udes bibliographical references and index.
ISBN 978-1-4612-6466-8 ISBN 978-0-8176-8160-9 (eBook)
DOI 10.1007/978-0-8176-8160-9
1. Structural analysis (Engineering) 2. Finite element method. 1. Wei, Dongming, 1958-
II. Title

TA646.K98 2003
620'.0042-<ic22 2003057713
CIP

AMS Subject Classifications: 65N30,65-01

Printed on acid-free paper.

©2004 Springer Science+Business Media New York
$®
Originally published by Birkhauser Boston in 2004
Softcover reprint of the hardcover 1st edition 2004
AII rights reserved. This work may not be translated or copied in whole or in part without the written
permission of the publisher (Springer Science+Business Media, LLC),
except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with
any form of information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed is forbidden.

The use of general descriptive names, trade names, trademarks, etc., in this publication, even ifthe
former are not especially identified, is not to be taken as a sign that such names, as understood by the
Trade Marks and Merchandise Marks Act, may accordingly be used freely by anyone.

ISBN 978-1-4612-6466-8 SPIN 10888337

Typeset by TechType Works, Inc., Gretna, LA.

Mathematica® is a registered Trademark ofWolfram Research, Inc.
MA TLA~ is a registered trademark of MathWorks, Inc.

9 8 7 6 543 2 1
Contents

Preface ix
Notation xvii

1 Introduction 1
1.1 Historical Sketch 1
1.2 Euler-Lagrange Equations 2
1.2.1 Functionals 2
1.2.2 Boundary Conditions 3
1.3 Weak Variational Form 3
1.4 Galerkin Method 9
1.5 Rayleigh-Ritz Method 11
1.6 Exercises 18

2 One-Dimensional Shape Functions 25

2.1 Local and Global Linear Shape Functions 25
2.2 Local and Global Quadratic Shape Functions 28
2.3 Parametric Coordinates 31
2.4 Hermite Shape Functions 31
2.5 Exercises 34

3 One-Dimensional Second-Order Equation 37

3.1 Galerkin Finite Element Method 37
3.1.1 Linear Elements for Second-Order Equation " 38
3.1.2 Local and Global Matrices 42
vi CONTENTS

3.1.3 Quadratic Elements for Second-Order Linear Problems 45

3.1.4 Mixing Two Different Elements 47
3.2 Two Dependent Variables 63
3.5 Exercises 66

4 One-Dimensional Fourth-Order Equation 75

4.1 Euler-Bernoulli Beam Equation 75
4.1.1 Finite Element Equation 75
4.1.2 Boundary Conditions 77
4.2 Exercises 86

5 Two-Dimensional Elements 89
5.1 Linear Three-Node Triangular Elements 89
5.2 Bilinear Four-Node Rectangular Elements 91
5.3 Global Shape Functions 94
5.4 Triangular Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.5 Shape Functions on the Sides of a Triangle 99
5.5.1 Linear Shape Functions on the Sides of a Triangle 99
5.5.2 Quadratic Shape Functions on the Sides of a Triangle 101
5.6 Exercises 105

6 Two-Dimensional Problems 107

6.1 Single Dependent Variable Problems 107
6.1.1 Local Weak Formulation 108
6.1.2 Finite Element Equation 109
6.1.3 Evaluation of Stiffness Matrix and Load Vector . . . . . . . . . . . . . 110
6.1.4 Evaluation of Boundary Integrals. . . . . . . . . . . . . . . . . . . . . . . . .. III
6.1.5 Assembly of Element Matrices 112
6.2 Exercises . . . . . . . . . . . . . . . . . . . . . 129

7 More Two-Dimensional Problems 137

7.1 Heat Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
7.2 Torsion 144
7.3 Seepage 146
7.4 Fluid Flows 152
7.5 Exercises . .... . . . . . . . . . . . . ..... .. .... . . . . . . .. . ... . . ... . .. .. .. 164

8 Axisymmetric Heat Transfer 175

8.1 Radial Symmetry 175
8.2 Linear Elements 182
8.3 Linear Elements for Heat Transfer in Fluids 194
8.4 Nonlinear Heat Transfer 196
8.4.1. Gauss-Legendre Integration Method 198
8.5 Exercises 202
 CONTENTS vii

9 Transient Problems 207

9.1 Classical Methods 207
9.2 One-Dimensional Transient Problems 212
9.3 Time-Dependent Heat Conduction 217
9.3.1 Derivation of Finite Element Equations 218
9.3.2 Numerical Time Integration 221
9.4 Two-Dimensional Transient Problems 228
9.5 Exercises 233

10 Single Nonlinear One-Dimensional Problems 241

10.1 Newton 's method 241
10.2 Radiation Heat Transfer 243
10.3 Stress Analysis of Plastic Rods 252
10.4 Power-Law Pressure Driven Flow between Two Plates 256
10.5 Mixing-Length Equation for Turbulent Flow in Pipes 259
10.6 Rayleigh-Ritz and Nonlinear Gradient Methods 261
10.7 Exercises 266

11 Plane Elasticity 269

11.1 Stress-Strain Relations 269
11.2 Constant-Strain Triangular Element 271
11.3 Virtual Displacement Finite Element Model . . . . . . . . . 272
11.4 Weak Form Finite Element Model 273
11.5 Stiffness Matrix and Load Vector 275
11.5.1 Linear CST Element 275
11.5.2 Assembly of CST Elements 276
11.5.3 Bilinear Rectangular Element 284
11.5.4 Two or More Bilinear Rectangular Elements 288
11.6 Exercises 291

12 Stokes Equations and Penalty Method 297

12.1 Equality-Constrained Programs and Lagrange Multipliers 297
12.2 Penalty Formulation for Linear Stokes Equation 299
12.3 Penalty Linear Triangular Stokes Element 301
12.4 Penalty Bilinear Rectangular Stokes Element 309
12.5 Penalty Linear Triangular Power-law Stokes Element 311
12.6 Solutions by Conjugate Gradient Methods 313
12.7 Exercises 321

13 Vibration Analysis 323

13.1 Hamiltonian Principle 323
13.2 Free Axial Vibrations of an Elastic Rod " 327
13.3 Free Vibrations of a Euler Elastic Beam 330
13.4 Free In-Plane Vibrations of an Elastic Plate 334
viii CONTENTS

13.5 Axial Vibrations of a Plastic Rod 342

13.6 Eigenvalue Problems 344
13.6.1 Hyperbolic Equations 344
13.6.2 Parabolic Equations 346
13.6.3 Helmholtz Equation 347
13.7 Exercises 347

14 Computer Codes 351

14.1 Mathematica Codes 351
14.2 Ansys Codes 363
14.3 Matlab Codes 368
14.4 Fortran Codes 374

A Integration Formulas 393

B Special Cases 397

C Temporal Approximations 405

C.l First-Order Derivative . . . . . . . 405
C.2 Second-Order Derivative 406

D Isoparametric Elements 409

D.l Introduction 409
D.Ll Local ~ -Syatem 409
D.l .2 Triangular Elements 411
D.1.3 Rectangular Elements 412
D.2 Curvilinear Coordinates 414
D.2.1 One-Dimensional Integrals 416
D.2.2 Two-Dimensional Integrals 417
D.3 Pascal Triangle 418

E Green's Identities 421

F Gaussian Quadrature 423

G Gradient-Based Methods 429

G.l Method of Steepest Descent 429
G.2 Conjugate Gradient Method 431

Bibliography 433

Index 439
Preface

This textbook on linear and nonlinear finite element methods has evolved out of
teaching of a sequence of two courses at the University of New Orleans for the past
12 years to a variety of students from the engineering, physics, geophysics, and
mathematics departments. The first of these courses is an introduction designed for
undergraduate and beginning graduate students, whereas the second is a follow-
up mainly for graduate students. In the past we used a number of textbooks and
selected material at different times on a trial basis, and we developed our own
notes and a repertoire of representative engineering and technological problems
that cater to the disciplines of our students.
As everyone knows the market is replete with textbooks on finite element
analysis. Then why do we need another textbook on this subject? To answer this
question, we next describe how we have used existing books and why we felt a
need to write our own textbook .
(i) Although there are many excellent books on the market, very few are de-
signed for students from different disciplines and use widely available modern
software such as Mathematica, Matlab, Fortran , and commercial finite element
codes such as Ansys.
(ii) In our view, the finite element technology oftoday has grown into a basic tool
for almost every field in engineering and applied sciences. Finite element courses
have become more and more popular in both undergraduate and graduate curricula
in colleges of engineering and sciences throughout the world. The new generation
of undergraduate and graduate students is more accustomed to using computers
for everyday homework assignments and research than ever before. There are
a few introductory finite element textbooks that provide a good combination of
x PREFACE

methodology, applications, computer implementation, and hands-on projects.

(iii) Of the books we have used, some incorporate commercial codes into
the text. A few books introduce applications of the finite element method in
electromagnetics together with other traditional applications in solid and fluid
mechanics. There are a few other books that deal exclusively with the finite
element analysis of electromegnetic problems. Such books are aimed at special
readers in specialized fields of interest and are not suitable for a general curriculum.
(iv) We have found only one specialized book on finite element analysis dealing
with non-Newtonian fluid flows, but it is designed for chemical engineers. Other
books do not consider this topic at all. Even in books with traditional applications,
only a few give details about vibrational finite element analysis, although there are
other books that deal exclusively with this topic.
(v) We believe that the basic theory of finite element analysis should be made
accessible to all undergraduate students in engineering and sciences with a min-
imum background in mathematics, computer science, physics, and any specific
engineering field. Our experience with the books we have used is that only a few
are designed to serve a broad range of engineering and science majors and to ex-
pose students to hands-on application projects using the computational software
tools that are now widely available on campuses.
(vi) We have found that in our introductory course it is very helpful to devote
sufficient time to cover equations in one space dimension and include problems
with applications in the following areas: structural, thermal , Newtonian and non-
Newtonian fluids, mechan ical vibration , electromagnetics, variational principles,
and thermal elastic coupled systems . The underlying mathematics for the first
course are kept at a minimum by using only basic techniques, such as the Gauss -
Jordan elimination method and the backward substitution method for solving linear
systems of algebraic equation, integration by parts formula s in one or two dimen-
sions , chain rule for differentiation, and change of variable formula for integration
and Newton 's method for solving nonlinear systems of algebraic equations.
(vii) In the second course we introduce advanced techniques for solving discrete
algebraic systems, viz., line search , method of steepest descent, linear and nonlinear
conjugate gradient methods, Newton's methods and penalty method.
During the past decade we have accumulated valuable experience and absorbed
many excellent ideas from other authors in teaching these two service courses. The
result is a general service applied finite element textbook that not only combines
all the above ideas and experience but is diverse enough in applications and simple
enough in theory for students in all engineering and science fields. We have kept the
mathematical theory to a minimum, just sufficient for understanding the beginner's
whys and hows of finite element methods.
 PREFACE xi

Salient Features

This textbook presents basic theoretical material and practical applications for
solving problems of interest in engineering and technology. The book covers a
two-semester course and is designed for junior/senior undergraduate and first-year
graduate students from engineering, geophysics, physics, and applied mathematics
who have had a course in calculus of several variables, differential equations as
well as linear algebra and have some knowledge of computers.
This book
• Fulfills the need for a service course taught to students with diverse background.
• Provides thorough instructions and training to these diverse students, unlike
most textbooks that are designed with specific applications, e.g., for structural
engineers , mechanical engineers, electrical engineers, chemical engineers, geo-
physicists, or applied mathematicians.
• Involves students in the use of widely available modern software such as Mathe-
matica, Matlab, and the use of commercial finite element codes such as Ansys .
• Deals with problems of varied types, including linear and nonlinear problems
in radiation heat transfer, mechanics of elastic and plastic media, continuum
mechanics , non-Newtonian fluid flows, and electromagnetics.
• Provides accessibility to all students with diverse backgrounds with a minimum
of mathematical analysis .
• Serves a broad range of engineering and science students and exposes them to
applications from different disciplines , unlike books that deal with only one
specific fields.
• In the first-semester course, presents one-dimensional equations and their appli-
cations in the following areas: structural analysis, heat conduction, fluid flows,
mechanical vibration , electromagnetics, variational principles, and thermoe-
lastic coupled systems .
• In the second-semester course , presents advanced techniques for solving dis-
crete algebraic systems , viz., line search, method of steepest descent, conjugate
gradient methods, Newton's methods , and penalty method.
• Offers supporting and other interesting material via the Internet at the authors'
web sites .
• Provides service, accessibility, and applications toward the development of the
subject.
xii PREFACE

Overview

A general description of the topics covered in the book is as follows. After a brief
historical sketch of the development of the finite element method and its subsequent
development, Chapter 1 introduces the weak variational formulation of a boundary
value problem. Galerkin and Rayleigh-Ritz weighted residual methods. This in-
troduction lays the foundation of the Galerkin finite element method developed in
subsequent chapters. Chapter 2 is devoted to the development of one-dimensional
global and local linear and quadratic Legendre interpolation shape functions and
the Hermite interpolation shape functions that are used with fourth-order differen-
tial equations in the beam problems. The Galerkin method for a one-dimensional
second-order equation, also known as the potential or the bar equation, is first
developed in Chapter 3, using linear and quadratic elements. Problems are solved
by mixing these two types of elements, and a one-dimensional problem with two
dependent variables is solved in this chapter. Problems with a one-dimensional
fourth-order equation, also known as the beam equation, are covered in Chap-
ter 4. Two-dimensional problems start from Chapter 5, which is devoted to the
development of linear three-node triangular and bilinear four-node rectangular cle-
ments. Triangular coordinate system is introduced, and linear and quadratic shape
functions on the sides of a triangle are developed in this chapter.
Chapter 6 deals with two-dimensional problems involving a single space vari-
able; the finite element equation is developed, stiffness matr ix and the load vector
are evaluated, element matrices are assembled, and boundary integrals are eval-
uated for different types of boundary value problems taken from industrial and
technological areas . More two-dimensional boundary value problems, mainly
from heat transfer, torsion, seepage, and fluid flows, are presented in Chapter 7.
Axisymmetric linear and nonlinear heat transfer problems in solids and fluids
are presented in Chapter 8. Transient problem and numerical time integration
methods are presented in Chapter 9. Single nonlinear one-dimensional equations
and the related finite element method are discussed in Chapter 10. Plane elasticity
problems are introduced in Chapter 11; principle of virtual work, linear triangu-
lar elements and bilinear rectangular elements are introduced, the concept of two
degrees of freedom at each node is explained, and simple problems mostly involv-
ing two elements are presented. Chapter 12 deals with the Stokes equation and
the penalty method; linear triangular and bilinear rectangular Stokes elements are
derived, and solution by the conjugate gradient method is explained in detail. The
corresponding nonlinear Stokes elements for power-law non-Newtonian flows are
also introdu ced together with the penalty method and some nonlinear conjugate
gradient methods. Chapter 13 deals with the vibration analysis ; the Hamiltonian
principle is introduced, and the free axial vibrations of an elastic rod , free vibra-
 PREFACE xiii

tions of an Euler elastic rod, free in-plane vibrations of an elastic plate, and axial
vibrations of a plastic rod are discussed. The chapter ends with solutions of eigen-
value problems that arise from hyperbolic and parabolic equations and that lead to
the Helmholtz equation. New codes in Mathematica, Ansys, Matlab, and Fortran
are given in Chapter 14.
There are six appendices, which contain useful supplemental information. Ap-
pendix A contains integration formulas that are used in the development of the
subject. Appendix B contains special cases of linear triangular and bilinear rectan-
gular stiffness matrix and load vector that arise in certa in special situations. For-
mulas for numerical solution of time-dependent problems are given in Appendix
C. Information about some basic isoparametric elements, together with curvilinear
coordinates and the Pascal triangle, is collected in Appendix D. Green 's identities
are available in Appendix E, Gaussian quadrature is presented in Appendix F, and
gradient-based methods, viz., the method of steepest descent and the linear and
nonlinear conjugate gradient methods, are discussed in Appendix G.
As mentioned earlier, this book contains both linear and nonlinear finite element
methods . The nonlinear development of two-dimensional heat transfer problems
and the related Gauss-Legendre integration method are presented in Chapter 8.
The finite element method for solving single nonlinear one-dimensional equations
is developed in Chapter 10. It introduces the Newton's method and discusses
problems involving radiational heat transfer, stress analysis of plastic rods, power-
law pressure driven flows, mixing-length equation for turbulent flows in pipes, and
the Rayleigh-Ritz and nonlinear gradient methods . Since the subject of nonlinear
finite element methods is advanced, we have taken care to include topics that can
be easily understood, especially by the students for whom this book is primarily
written.
The book is a bit heavy on heat transfer problems. It has been our experi-
ence that such problems are easy for instructors to explain and the students to
comprehend their elegance and simplicity.
There are 87 examples and 148 exercises in the book ; hints , answers, and
solutions for most of the exercises are available in the book. A liberal use of
illustrations is made to clarify the nature of the problems; as a result there are 152
figures and illustrations in the book.
The bibliography at the end of the book contains literature that has been con-
sulted as well as references that are cited . The book ends with a detailed subject
index.
This is an introductory textbook which provides a good combination of method-
ology, applications, computer implementation, and hands-on projects for students
with diverse background from engineering, technology, physics, geophysics, and
applied mathematics.
xiv PREFACE

To Instructors

As mentioned above, the chapters present balanced two-semester course material,

which can be tailored to the needs of different levels of instruction. Although the
following table outlines some suggested curricula at the two levels, the instructor is
free to adapt whatever chapters or sections in any order depending on the limitations
of the course.

First Semester Second Semester

Chapter 1 Chapter 8
Chapter 2 Chapter 9
Chapter 3 Chapter 10
Chapter 4 Chapter 11
Chapter 5 Chapter 12
Chapter 6 Chapter 13
Chapter?

Chapter 14 is to be used in both courses, as it pertains to the topics and sections

needed in each course.

Computational Aspects

Computer codes using Mathematica, Ansys, Matlab, and Fortran are presented in
Chapter 14. They cover linear as well as nonlinear problems, which are discussed
in detail. Our students have freely used and modified some well-known Fortran
programs, for example, by Taylor and Hughes (1981), Reddy (1984, 1993) and
Segerlind (1984), and certain other codes available in private domain. They have
generated these modifications for certain specific problems and used them to obtain
computational solutions. We have encouraged our students to study these and other
authors in order to use and modify their codes. It has helped them understand the
subject much better than using a single textbook.
We have provided some new codes in Mathematica, Ansys, Matlab, and For-
tran, as they arise during the course of the presentation of the textual material.
Mathematica 3.0 or higher is used to verify most of the solutions. For various
 PREFACE xv

other useful items and suggestions, check either author's website :

www.math.uno.edu/fac/pkythe.html www.math .uno.edu/fac/dwei.html

Acknowledgments

The authors thank all the students who have participated in our finite element
courses since 1990. Although the list is very large, we will take liberty of mention-
ing names of certain students who, by their hard work and perseverance, excelled
in the study of this subject, benefited not only themselves but other students,
and kept the course running with full steam throughout these years . These dedi-
cated students are Antonis Antoniou, Pravan Kumar Bachu, Xiang Bian, Rambabu
Chennupati, Gordon C. Cuculla III, Shawn Davis, Andrea M. Dunham, Lang
Yan Fook , Karunakar R. Gaddam, Patrick Louis Garrity, Praveen Ghantasala,
Monikonda Shekar Goud, Wael M. Khairy, Gloria Ladnier, Haibiao Luo , Larry
Manalo, William A. Myers, Yashawant Mummaneni, Chakvartula 1. S. V. Nallani,
Vuong Hun Nguyen, Tien Szu Pan, Madahv K. Parimi, Srinivasa Rao Pasala, Srid-
har Rao Pasupuleti, Yezdi Parvez Pavri, Khoa Pham, Suresh K. Sivapuram, Atluri
Srinivas, Edupuganti Suneel, Jifeng Ru , Annette R. Schulte, Reddy P. Shashikanth,
Michael Sheaf, Ved P. Tandle, Nuthan K. Vajinapalli, Chandra Rao Valipe, Satya-
narayana Rao Valipe, Prakash Vyakaran, Bonnie Williams, Qin Xu, Zhidong Xu,
Xuan Yu, Keshu Zhang, and Peng Zang.
The authors thank Dr. Vivek Sahai of Cheng Fluid Systems, Mountain View,
CA , for some valuable advice. They thank Dr. Carroll F. Blakemore for help and
support, and some other colleagues for useful suggestions. The contribution of
the Senior Production Editor and the copy editor at Birkhauser Boston for the fine
work of editing the manus cript, and of the people at TechType Works , Inc. , Gretna,
Louisiana, for typesetting the manuscript is sincerely acknowledged.

New Orleans, Louisiana Prem K. Kythe

July 2003 Dongming Wei
Notation

Standard notation is used throughout the book. Since matrices and vectors occupy
most of the text, the usual convention is adopted. All matrices are printed in
uppercase boldface and all vectors in lowercase boldface letters , except U , F , Q,
and T, which represent the displacement vector, load (force) vector, flux vector,
temperature vector, and the error vector at the global nodal points, respectively,
and F (e) and R (e), which denote the load vector and the error vector for an element
n (e) , respectively. This practice does not lend to any confusion in the conte xt of
their use. Also, sometimes to save space on a page, a column vector is printed as
the transpose of its corresponding row vector.
A list of the notation , abbreviations and acronyms used in this book is given
below.

a elastic modulus in R"

ai , bi, c, coefficients in a linear triangular shape functions
a row vector for ai
A area; surface area of a fin; cross-sectional area
A (e) area of a two-dimensional clement
As total lateral surface area (of a fin)
AT, (Af transpo se of a matrix A
b(w, u) bilinear form
b, coefficients in a linear triangular shape functions
b row vector for bi
B matrix
B(e) matrix of the values of bi , c,
const constant
coefficients in a linear triangular element
xviii NOTATION

specific heat at constant pressure

Cv specific heat at constant volume
row vector for Ci
c.,
C
material (elastic) constants
C connectivity matrix ; matrix of the elastic constants Ci j
dof degree(s) of freedom
D dimension of the geometry in fluid flow
D matrix of first partial derivative of shape functions
e.g. for example
e finite element number e = 1, .. . ,NE
erfc complementary error function
E modulus of elasticity, Young's modulus
Ek elastic kinetic energy
Ep elastic potential energy
El flexural rigidity
Eq(s) equation(s) (when followed by an equation number)
f prescribed function; heat source; recharge (aquifer)
fee) value of f on an element nee)
fo uniform load
j axial body force
fx,fy body forces
Fo ,FL shear forces at x = 0 and x = L, repectively
P Froude number
FE finite element
f vector
F force or load vector (F = f + Q)
F (e) force or load vector of an element nee)
g prescribed function; shear modulus (torsion); gravity
g vector; body forces
h thickness of a fin or a plate
h vector
H(x,xo) Heaviside unit function (= 0 if x < xo, and = 1 if x > xo)
H square matrix
unit vector in the x direction
i.e. that is
iff if and only if
I moment of inert ia
l(u) functional ; total energy of an elastic mechanical system
1m modified Bessel function of the first kind and order m
Imn integral defined in Example 5.2
j unit vector in the y direction
J current
J Jacobian matrix
k thermal conductivity; permeability coefficient (aquifer)
 NOTATION xix

kx ,ky thermal conductivity in the x and y direction

k unit vector in the z direction
k(e) value of k on an element e
K constant value of a metal property; consistency coefficient
K(e) stiffness matrix of an element n(e)
K global stiffness matrix
Kb matrix
l(w) linear functional
l( e) length of the interval [xie) , x~e)J
i.i« : length of nonuniformone-dimensional consecutive elements
L length of an interval; length unit; linear operator
i: Lagrange function; Laplace transform
c:' inverseof Laplace transform
L matrix
M bending moment; polar moment of a cross-sectional area
Mo, ML bending moment at x = a and x = L, respectively
M matrix
M(e) matrix (radially symmetric element)
n power-law index
n outward normal vector
direction cosines of n
a a a
n·\? = nx ax + n y oy + n z oz
NE number of elements
NL number of local nodes on an element
P
. I au
pressure; penmeter; a so = U x = ax
psi lbs/irr'
P vertical point load; vector
Pn(X) Legendre polynomials of degree n
Pi search direction
q heat source; also = uy =
au
oy
qn heat flux
q rate of heat generation
q temperature gradient
Ql shear force
Q2 bending moment
Q~ e) shear force at node i of an element n( e)
Q(e) vector of secondary degrees of freedom (boundary terms)
Q part of F corresponding to the natural boundary conditions
Qb vector
r radial distance
r(u) scalar residual (error) in the Galerkin method
(r,e ,z) cylindrical polar coordinates
xx NOTAT I ON

R radius
Rn Euclidean n-space
R+ set of positive real numbers
Rj errors (j = 1, . . . , n + 1)
Re Reynolds number
~ real part
R(e) residual or error vector
R global error vector
8 variable of the Laplace transform; arc length ; second(s) (time)
(8, t) nondimensional coordinates on the unit square
sym symmetric (matrix)
s(e) square matrix
t time
ix ,iy prescribed secondary variables
T temperature, temperature distr ibution
t; base temperature of a fin
Too amb ient temperature
T temperature vector
T( e) temperature vector for an element n( e)
U dependent variable; stress function; displacement; mean velocity
Uo prescribed velocit y
Uoo free stream velocity
(e)
Ui value of U at node i
(e )
Ua linear interpolation function for the interval [x ~e) , x~e) ]
U velocity vector, displacement vector
U vector of the first time derivatives of u
U vector of the second time derivatives of u
u( e) approximation of u on an element n (e)
u ,v velocity components of u or v , in x and y direction
ii , V approximate solution for u or v
Uo inlet velocity
Ue nodal value of U at a global node e, e = 1, . . . , NE
u (e) (e) (e) (e) (e)]
coIumn vector [u 1 ... uN vI .. . V N
U vector of global values of displacement U
Vr radial component of velocity v
V three-dimensional solid or volume
av boundary surface of a solid V
v velocity vector
W test function
Wi weights in Gaussian quadrature
Wl ,W2 ,W3 three different test functions
W work
global coordinate
end points of a line element
 NOTATION xxi

local coordinate ( = x - x~e» )

initial guess (Newton 's method)
set of nonnegative integers
a thermal capacitance
fJ film coefficient, convective heat transfer coefficient
fJPY Dai-Yuan parameter

,fJr
s Hestenes-Stiefel parameter
specific heat ration (= cp / c,J
r boundary of a domain (= a n)
f 1,f z disjoint portions of boundary I' (f 1 U f z = an)
b tolerance
bi j Kronecker delta (= 1 if i = j ; -1 if i =1= j)
D.t time step
emissivity
strain ; penalty parameter
strain vector
shear strain
e strain ( = - ~~); angle of twist (torsion)
{) nondimensional temperature
= lu~e) -u~~ll , i = 1, 2, .. . , N
von Karman constant
eigenvalue
dynamic viscosity of a fluid
kinematic viscosity of a fluid (= J.L / p); Poisson's ratio (elasti city)
~ isoparametric one-dimensional coordinate, -1 < :::
~ 1
~, 'TJ isoparametric two-dimensional coordinates, -1 ::::: ~ , 'TJ ::::: 1
~i Gaussian points
6 ,6,6 trilinear coordinates
p density
a stress; Stefan-Boltzmann constant
aD uniform load
a stress vector
T shear stress
Txy , Tyz average shear stresses
¢ velocity potential of a flow
¢~ e) basis function ; interpolation shape function
<I>e global shape functions for e = 1, . . . , N + 1
lj>(e) vector of the shape functions for an element n (e)
X (e) characteristic function for the interval [x~e), x~e) ] or n (e)
'IjJ stream function
w radian frequency
n domain
n(e) general finite element
xxii NOTATION

In(e)1 area or volume of an element n(e)

an boundary of the domain n
an(e) boundary of an element n(e)
o null vector
o matrix of differential operators
I-D one-dimensional
2-D two-dimensional
3-D three-dimensional
.8 .8 k 8
\l gra d =1-+J-+ -
8x 8y 8z
I x] norm of a vector x
II\lull norm of gradient vector (= Jui + u~ + u~ )
\If gradient matrix of f
. (8P. Bp, P
\lp pressure gra dlent 8x 1 + 8yJ + 88z k)
\l·F divergence of a vector F (div F)
2 2
\l2 a2 a a )
Laplacian ( = 8x 2 + 8 y 2 + 8z 2
AlB set A minus set B
apprimately equal to
empty set
 An Introduction
to Linear and Nonlinear
Finite Element Analysis