Solid Mechanics

Extract from on-line books of lecture notes on solid mechanics of prof. P. Kelly – Department of
Engineering Science, University of Auckland
The complete text is available at
http://homepages.engineering.auckland.ac.nz/~pkel015/SolidMechanicsBooks/index.html
Table of Contents
Cap 1
1. Introduction 1
1.1 What is Solid Mechanics? 3
1.2 What is in this Book? 7
2. Statics of Rigid Bodies 9
2.1 The Fundamental Concepts and Principles of Mechanics 11
2.2 The Statics of Particles 14
2.3 The Statics of Rigid Bodies 19
3. Stress 29
3.1 Surface and Contact Stress 31
3.2 Body Forces 39
3.3 Internal Stress 42
3.4 Equilibrium of Stress 52
3.5 Plane Stress 62
4. Strain 85
4.1 Strain 87
4.2 Plane Strain 101
4.3 Volumetric Strain 109
5. Material Behaviour and Mathematical Modelling 113
5.1 Mechanics Modelling 115
5.2 The Response of Real Materials 119
5.3 Material Models 134
5.4 The Continuum 137
6. Linear Elasticity 141
6.1 The Linear Elastic Model 143
6.2 Homogeneous Problems in Linear Elasticity 152
7. Applications of Elasticity 165
7.1 One Dimensional Axial Deformations 167
7.2 Torsion 177
7.4 Beams 192
Answers to Selected Problems:
Chapter 2
Chapter 3
Chapter 4
Chapter 5
Chapter 6
Chapter 7

Cap 2
1. Differential Equations for Solid Mechanics 1
1.1 The Equations of Motion 3
1.2 The Strain-Displacement Relations 9
1.3 Compatibility of Strain 17
7. Three Dimensional Elasticity 187
7.1 Vectors, Tensors and the Index Notation 189
7.2 Analysis of Three Dimensional Stress and Strain 201
7.3 Governing Equations of Three Dimensional Elasticity 220
Answers to Selected Problems:
Chapter 1
Chapter 7
Cap 3
Appendix insights
1. Vectors and Tensors 1
1.1 Vector Algebra 3
1.2 Vector Spaces 10
1.3 Cartesian Vectors 15
1.4 Matrices and Element Form 22
1.5 Coordinate Transformation of Vector Components 24
1.6 Vector Calculus 1 - Differentiation 30
1.7 Vector Calculus 2 – Integration 50
1.8 Tensors 67
1.9 Cartesian Tensors 75
1.10 Special Second Order Tensors & Properties of Tensors 84
2. Kinematics 199
2.1 Motion 201
2.2 Deformation and Strain 207
2.3 Deformation and Strain: Further Topics 233
2.4 Material Time Derivatives 239
2.5 Deformation Rates 243
2.6 Deformation Rates: Further Topics 253
2.7 Small Strain Theory 258
2.9 Rigid Body Rotations 276
3. Stress and the Balance Principles 315
3.1 Conservation of Mass 317
3.2 The Momentum Principles 327
3.3 The Cauchy Stress Tensor 330
3.4 Properties of the Stress Tensor 335
3.6 The Equations of Motion and Symmetry of Stress 351
3.7 Boundary Conditions and the Boundary Value Problem 356
1 Introduction
This brief chapter introduces the subject of Solid Mechanics and the contents of this book
(Book I) and the books which follow (Books II-V).
1
2
Section 1.1
1.1 What is Solid Mechanics?

Solid mechanics is the study of the deformation and motion of solid materials under the
action of forces. It is one of the fundamental applied engineering sciences, in the sense
that it is used to describe, explain and predict many of the physical phenomena around us.
Here are some of the wide-ranging questions which solid mechanics tries to answer:
When will this cliff How does the

collapse? heart contract
and expand as it
is pumped?
2
1
How long will a tuning
When will these fork vibrate for?
gears wear out?
3
How will the San Andreas fault

in California progress? How Pacific
will the ground move during an Plate
earthquake?
American
Plate
why does nature

use the materials it
how do you build a bridge 5
does? Knee
which will not collapse?
Solid mechanics is a vast subject. One reason for this is the wide range of materials
which falls under its ambit: steel, wood, foam, plastic, foodstuffs, textiles, concrete,
biological materials, and so on. Another reason is the wide range of applications in which
these materials occur. For example, the hot metal being slowly forged during the
manufacture of an aircraft component will behave very differently to the metal of an
automobile which crashes into a wall at high speed on a cold day.
Solid Mechanics Part I 3 Kelly

Section 1.1
Here are some examples of Solid Mechanics of the cold, hot, slow and fast …
7 8
how did this Antarctic ice how much will this glacier move in
fracture? one year?
what materials can withstand what damage will occur during a
extreme heat? car crash?
Here are some examples of Solid Mechanics of the small, large, fragile and strong …
what affects the quality of paper? how strong is an eggshell and what
(shown are fibers 0.02mm thick) prevents it from cracking?
how will a ship withstand wave how thick should a dam be to
slamming? withstand the water pressure?
1.1.1 Aspects of Solid Mechanics

The theory of Solid Mechanics starts with the rigid body, an ideal material in which the
distance between any two particles remains fixed, a good approximation in some
applications. Rigid body mechanics is usually subdivided into
 statics, the mechanics of materials at rest, for example of a bridge taking the weight
a car
 dynamics, the study of bodies which are changing speed, for example of an
accelerating and decelerating elevator
Following on from statics and dynamics usually comes the topic of Mechanics of
Materials (or Strength of Materials). This is the study of some elementary but very
relevant deformable materials and structures, for example beams and pressure vessels.
Elasticity theory is used, in which a material is assumed to undergo small deformations
when loaded and, when unloaded, returns to its original shape. The theory well
approximates the behaviour of most real solid materials at low loads, and the behaviour of
the “engineering materials”, for example steel and concrete, right up to fairly high loads.

Section 1.1
More advanced theories of deformable solid materials include

 plasticity theory, which is used to model the behaviour of materials which undergo
permanent deformations, which means pretty much anything loaded high enough
 viscoelasticity theory, which models well materials which exhibit many “fluid-
like” properties, for example plastics, skin, wood and foam
 viscoplasticity theory, which is a combination of viscoelasticity and plasticity
Some other topics embraced by Solid Mechanics, are

 rods, beams, shells and membranes, the study of material components which can
be approximated by various model geometries, such as “very thin”
 vibrations of solids and structures
 composite materials, the study of components made up of more than one material,
for example fibre-glass reinforced plastics
 geomechanics, the study of materials such as rock, soil and ice
 contact mechanics, the study of materials in contact, for example a set of gears
 fracture and damage mechanics, the mechanics of crack-growth and damage in
materials
 stability of structures
 large deformation mechanics, the study of materials such as rubber and muscle
tissue, which stretch fairly easily
 biomechanics, the study of biological materials, such as bone and heart tissue
 variational formulations and computational mechanics, the study of the
numerical (approximate) solution of the mathematical equations which arise in the
various branches of solid mechanics, including the Finite Element Method
 dynamical systems and chaos, the study of mechanical systems which are highly
sensitive to their initial position
 experimental mechanics
 thermomechanics, the analysis of materials using a formulation based on the
principles of thermodynamics
10 11 12
13 14

Section 1.1
Images used:
1. http://www.allposters.com/-sp/Haute-Ville-on-Cliff-Edge-Bonifacio-South-Corsica-Corsica-France-
Mediterranean-Europe-Posters_i8943529_.htm
2. http://www.natureworldnews.com/articles/1662/20130430/uk-begin-clinical-trials-gene-therapy-treat-heart-
failure.htm
3. Microsoft Clip Art
4. http://geology.com/articles/san-andreas-fault.shtml
5. http://www.pbs.org/wgbh/buildingbig/wonder/structure/sunshineskyway1_bridge.html
6. http://www.healio.com/orthopedics/sports-medicine/news/online/%7B3ea83913-dac4-4161-bc4c-
8efdad467a74%7D/microfracture-returned-44-of-high-impact-athletes-to-sport
7. http://allabout.co.jp/gm/gl/16461/
8. http://www.autoblog.com/2007/06/22/brilliance-bs6s-adac-crash-test-is-anything-but/
9. http://www.worldwideflood.com/ark/anti_broaching/anti-broaching.htm
12. http://www.editinternational.com/photos.php?id=47a887cc8daaa
13. http://edition.cnn.com/2008/US/07/06/nose.cone/index.html?iref=mpstoryview
14. http://rsna.kneadle.com/secure/Radiology/Modalities/CT/Clinical/MDCT-Peds-Skull-Frac.html

Section 1.2
1.2 What is in this Book?

This book is divided into five “sub-books”:
I. An Introduction to Solid Mechanics

II. Engineering Sold Mechanics
III. The Finite Element Method
IV. Foundations of Continuum Solid Mechanics
V. Material Models in Continuum Solid Mechanics
One can take a “bottom up” approach or a “top down” approach to the subject. In the
former, one looks at the particular – a restricted set of ideal geometries and materials, and
a restricted set of models and equations. One then builds upon this knowledge
incrementally, upwards and outwards. This would be the approach taken if one began at
the start of Book I and worked through all the books more or less sequentially. This is the
course taken by most engineering students, who would work through (a subset of the)
material over three to four years. Alternatively, one can begin very generally – consider
all the relevant equations and only later simplify these to the particular application under
study. This approach would involve starting at the beginning of Book IV and then
working through Book V.
The principal advantage of the bottom-up approach is that one can begin the journey with
only a limited mathematical knowledge. One can also develop a very physical feel for the
subject, and over 95% of applications in the real world can probably be attacked using
material from the first three books. The advantage of the top-down approach is that it
gives a lofty perspective of the subject at the outset, although the mathematics required is
not easy.
The aim of Book I is to cover the essential concepts involved in solid mechanics, and the
basic material models. It is primarily aimed at the Engineering or Science undergraduate
student who has, perhaps, though not necessarily, completed some introductory courses
on mechanics and strength of materials. Apart from giving a student a good grounding in
the fundamentals, it should act as a stepping stone for further study into Books II to V and
into some of the more specialised topics mentioned in §1.1. The philosophy adopted in
Book I is as follows:
 The mathematics is kept at a fairly low level; in particular, there are few
differential equations, very little partial differentiation and there is no tensor
mathematics
 The critical concepts – the ones which make what follows intelligible, and which
students often “miss” – are highlighted
 The physics involved, and not just the theory, is given attention
 A wide range of material models are considered, not just the standard Linear
Elasticity
The outline of Book I is as follows: Chapter 2 covers the essential material from a typical
introductory course on mechanics; it serves as a brief review for those who have seen the
material before, and serves as an introduction for those who are new to the subject.
Chapters 3-8 cover much of the material typical of that included in a Strength of
Materials or Mechanics of Materials course, and includes the elementary beam theory and

Section 1.2
energy methods. The latter part of the book, Chapters 10-12 cover the more advanced
material models, namely viscoelasticity, plasticity and viscoplasticity.
In Book II, differential equilibrium and strain is introduced, allowing for more complex
problems to be tackled, including problems of contact mechanics, fracture mechanics and
elastodynamics, the study of wave propagation and vibrations, and more complex
problems of plasticity theory and viscoplasticity.
In Book III, the Finite Element Method, the standard method of obtaining
approximate/numerical solutions to the equations of Solid Mechanics, is examined.
In Book IV, tensor mathematics is introduced, allowing one to analyse the mechanics of
solid materials without making any approximations, for example regarding the strain in
materials.
Finally, in Book V, material models are described.

2 Statics of Rigid Bodies
Statics is the study of materials at rest. The actions of all external forces acting on such
materials are exactly counterbalanced and there is a zero net force effect on the material:
such materials are said to be in a state of static equilibrium.
In much of this book (Chapters 6-8), static elasticity will be examined. This is the study
of materials which, when loaded by external forces, deform by a small amount from some
initial configuration, and which then take up the state of static equilibrium. An example
might be that of floor boards deforming to take the weight of furniture. In this chapter, as
an introduction to this subject, rigid bodies are considered. These are ideal materials
which do not deform at all.
The chapter begins with the fundamental concepts and principles of mechanics –
Newton’s laws of motion. Then the mechanics of the particle, that is, of a very small
amount of matter which is assumed to occupy a single point in space, is examined.
Finally, an analysis is made of the mechanics of the rigid body.
The material in this chapter covers the essential material from a typical introductory
course on statics. Although the concepts presented in this chapter serve mainly as an
introduction for the later chapters, the ideas are very useful and important in themselves,
for example in the design of machinery and in structural engineering.
9
10
Section 2.1
2.1 The Fundamental Concepts and Principles of

Mechanics
2.1.1 The Fundamental Concepts

The four fundamental concepts used in mechanics are space, time, mass and force1. It is
not easy to define what these concepts are. Rather, one “knows” what they are, and they
take on precise meaning when they appear in the principles and equations of mechanics
discussed further below.
The concept of space is associated with the idea of the position of a point, which is
described using coordinates ( x, y, z ) relative to an origin o as illustrated in Fig. 2.1.1.
y
particle p

o x
Figure 2.1.1: a particle in space
The time at which events occur must be recorded if a material is in motion. The concept
of mass enters Newton’s laws (see below) and in that way is used to characterize the
relationship between the acceleration of a body and the forces acting on that body.
Finally, a force is something that causes matter to accelerate; it represents the action of
one body on another.
2.1.2 The Fundamental Principles

The fundamental laws of mechanics are Newton’s three laws of motion. These are:
Newton’s First Law:

if the resultant force acting on a particle is zero, the particle remains at rest (if originally
at rest) or will move with constant speed in a straight line (if originally in motion)
By resultant force, one means the sum of the individual forces which act; the resultant is
obtained by drawing the individual forces end-to-end, in what is known as the vector
1
or at least the only ones needed outside more “advanced topics”

Section 2.1
polygon law; this is illustrated in Fig. 2.1.2, in which three forces F1 , F2 , F3 act on a
single particle, leading to a non-zero resultant force2 F.
F1
F3 F1 F
F2 F
 
F1
F
F2
F2
F3 F3
(a) (b) (c) (d)
Figure 2.1.2: the resultant of a system of forces acting on a particle; (a)

three forces acting on a particle, (b) construction of the resultant F, (c)
an alternative construction, showing that the order in which the forces
are drawn is immaterial, (d) the resultant force acting on the particle
Example (illustrating Newton’s First Law)
In Fig. 2.1.3 is shown a floating boat. It can be assumed that there are two forces acting
on the boat. The first is the boat’s weight Fg , that is its mass times the acceleration due
to gravity g. There is also an upward buoyancy force Fb exerted by the water on the boat.
Assuming the boat is not moving up or down, these two forces must be equal and
opposite, so that their resultant is zero.
Fg
Fb
Figure 2.1.3: a zero resultant force acting on a boat
The resultant force acting on the particle of Fig. 2.1.2 is non-zero, and in that case one
applies Newton’s second law:
2
the construction of the resultant force can be regarded also as a principle of mechanics, in that it is not
proved or derived, but is taken as “given” and is borne out by experiment

Section 2.1
Newton’s Second Law:

if the resultant force acting on a particle is not zero, the particle will have an acceleration
proportional to the magnitude of the resultant force and in the direction of this resultant
force:
F  ma (2.1.1)
where3 F is the resultant force, a is the acceleration and m is the mass of the particle. The
units of the force are the Newton (N), the units of acceleration are metres per second
squared (m/s2), and those of mass are the kilogram (kg); a force of 1 N gives a mass of 1
kg an acceleration of 1 m/s2.
If the water were removed from beneath the boat of Fig. 2.1.3, a non-zero resultant force
would act, and the boat would accelerate at g m/s2 in the direction of Fg .
Newton’s Third Law:

each force (of “action”) has an equal and opposite force (of “reaction”)
Again, considering the boat of Fig. 2.1.3, the water exerts an upward buoyancy force on
the boat, and the boat exerts an equal and opposite force on the water. This is illustrated
in Fig. 2.1.4.
 Fb
Fb
surface of water
(a) (b)
Figure 2.1.4: Newton’s third law; (a) the water exerts a force on the
boat, (b) the boat exerts an equal and opposite force on the water
Newton’s laws are used in the analysis of the most basic problems and in the analysis of
the most advanced, complex, problems. They appear in many guises and sometimes they
appear hidden, but they are always there in a Solid Mechanics problem.
3
vector quantities, that is, quantities which have both a magnitude and a direction associated with them, are
represented by bold letters, like F here; scalars are represented by italics, like m here. The magnitude and
direction of vectors are illustrated using arrows as in Fig. 2.1.2

Section 2.2
2.2 The Statics of Particles
2.2.1 Equilibrium of a Particle

The statics of particles is the study of particles at rest under the action of forces. Such
particles can be analysed using Newton’s first law only. This situation is referred to as
equilibrium, which is defined as follows:
Equilibrium of a Particle
A particle is in equilibrium when the resultant of all the forces acting on that particle is
zero
In practical problems, one will want to introduce a coordinate system to describe the
action of forces on a particle. It is important to note that a force exists independently of
any coordinate system one might use to describe it. For example, consider the force F in
Fig. 2.2.1. Using the vector polygon law, this force can be decomposed into
combinations of any number of different individual forces; these individual forces are
referred to as components of F. In particular, shown in Fig 2.2.1 are three cases in which
F is decomposed into two rectangular (perpendicular) components, the components of F
in “direction x” and in “direction y”, Fx and Fy .
y y y
F F Fy
x
Fy F
Fy Fx
 Fx
x  
Fx
x
(a) (b) (c)
Figure 2.2.1: A force F decomposed into components Fx and Fy

using three different coordinate systems
By resolving forces into rectangular components, one can obtain analytic solutions to
problems, rather than relying on graphical solutions to problems, for example as done in
Fig. 2.1.2. In order that the resultant force F on a body be zero, one must have that the
resultant force in the x and y directions are zero individually1, as illustrated in the
following example.
1
and in the z direction if one is considering a three dimensional problem

Section 2.2
Example
Consider the particle in Fig. 2.2.2, subjected to forces F1 , F2 , F3 . The particle is in

equilibrium and so by definition the resultant force is zero, F  0 . The forces are
decomposed into horizontal and vertical components F1x , F2 x , F3 x and F1 y , F2 y , F3 y . The
horizontal forces may be added together to get a single horizontal force Fx , which must
equal zero. This force Fx should be evaluated using the vector polygon law but, since the
individual forces F1x , F2 x , F3 x all lie along the same line, one need only add together the
magnitudes of these vectors, which involves simply an addition of scalars:
F1x  F2 x  F3 x  0 . Similarly, one has F1 y  F2 y  F3 y  0 . These equations could be
used to evaluate, for example, the force F1 , if only F2 and F3 were known.
F1
F1 y

F3 x F1x
x
F2 x
F3 F3 y
F2 y
F2
`
Figure 2.2.2: Calculating the resultant of three forces by

decomposing them into horizontal and vertical components
In general then, if a set of forces F1 , F2 , , Fn act on a particle, the particle is in

equilibrium if and only if
F x  0, F y  0, F z 0 Equations of Equilibrium (particle) (2.2.1)
These are known as the equations of equilibrium for a particle. They are three
equations and so can be used to solve problems involving three “unknowns”, for example
the three components of one of the forces. In two-dimensional problems (as in the next
example), they are a set of two equations.
Example
Consider the system of two cables attached to a wall shown in Fig. 2.2.3. The cables
meet at C, and this point is subjected to the two forces shown. To evaluate the forces of
tension arising in the cables AC and BC, one can draw a free body diagram of the particle

Section 2.2
C, i.e. the particle is isolated and all the forces acting on that particle are considered, Fig
2.2.3b.
y
A
FAC
100 N 100 N
C
 4
60 o
3 FBC
C x
C
B  FAC
120 N 120 N
(a) (b) (c)
Figure 2.2.3: Calculating the tension in cables; (a) the cable system,
(b) a free-body diagram of particle C, (c) cable AC in equilibrium
The equations of equilibrium for particle C are
F x   FBC  FAC cos 60  100 cos   0,

F y  FAC cos 30  100 sin   120  0
leading to FAC  46.2 N, FBC  36.9 N .
The cable exerts a tension/pulling force on particle C and so, from Newton’s third law, C
must exert an equal and opposite force on the cable, as illustrated in Fig. 2.2.3c.
■
The concept of the free body is essential to Solid Mechanics, for the most simple and
most complex of problems. Again and again, problems will be solved by considering
only a portion of the complete system, and analysing the forces acting on that portion
only.
2.2.2 Rough and Smooth Surfaces

Fig 2.2.4a shows a particle in equilibrium, sitting on a rough surface and subjected to a
force F. Such a surface is one where frictional forces are large enough to prevent
tangential motion. The free body diagram of the particle is shown in Fig. 2.2.4b. The
friction reaction force is R f and the normal reaction force is N and these lead to the
resultant reaction force R which, by Newton’s first law, must balance F.
When a particle meets a smooth surface, there is no resistance to tangential movement.

The particle is subjected to only a normal reaction force, and thus a particle in equilibrium
can only sustain a purely normal force. This is illustrated in Fig. 2.2.4c.

Section 2.2
F F F
Rf
N R
(a) (b) (c)
Figure 2.2.4: a particle sitting on a surface; (a) a rough surface, (b)

a free-body diagram of the particle in (a), (c) a smooth surface
2.2.3 Problems
1. A 3000kg crate is being unloaded from a ship. A rope BC is pulled to position the
crate correctly on the wharf. Use the Equations of Equilibrium to evaluate the
tensions in the crane-cable AB and rope. [Hint: create a free body for particle B.]
10o
A
cable
B
15o
rope
C
2. A metal ring sits over a stationary post, as shown in the plan view below. Two forces
act on the ring, in opposite directions. Draw a free body diagram of the ring including
the reaction force of the post on the ring. Evaluate this reaction force. Draw a free
body diagram of the post and show also the forces acting on it.
200 N 50 N
3. Two cylindrical barrels of radius 500mm are placed inside a container, a cross
section of which is shown below. The mass of each barrel is 10kg. All surfaces are
smooth. Draw free body diagrams of each barrel, including the reaction forces
exerted by the container walls on the barrels, the weight of each barrel, which acts
through the barrel centres, and the reaction forces of barrel on barrel. Apply the
Equations of Equilibrium to each barrel. Evaluate all forces. What forces act on the
container walls?

Section 2.2
1.5m

Section 2.3
2.3 The Statics of Rigid Bodies

A material body can be considered to consist of a very large number of particles. A rigid
body is one which does not deform, in other words the distance between the individual
particles making up the rigid body remains unchanged under the action of external forces.
A new aspect of mechanics to be considered here is that a rigid body under the action of a
force has a tendency to rotate about some axis. Thus, in order that a body be at rest, one
not only needs to ensure that the resultant force is zero, but one must now also ensure that
the forces acting on a body do not tend to make it rotate. This issue is addressed in what
follows.
2.3.1 Moments, Couples and Equivalent Forces

When one swings a door on its hinges, it will move more easily if (i) one pushes hard, i.e.
if the force is large, and (ii) if one pushes furthest from the hinges, near the edge of the
door. It makes sense therefore to measure the rotational effect of a force on an object as
follows:
The tendency of a force to make a rigid body rotate is measured by the moment of that
force about an axis. The moment of a force F about an axis through a point o is defined
as the product of the magnitude of F times the perpendicular distance d from the line of
action of F and the axis o. This is illustrated in Fig. 2.3.1.
Rigid body
axis F
o line of action
of force
d 
point of application
of force
Figure 2.3.1: The moment of a force F about an axis o (the axis goes
“into” the page)
The moment M o of a force F can be written as
M 0  Fd (2.3.1)
Not only must the axis be specified (by the subscript o) when evaluating a moment, but
the sense of that moment must be given; by convention, a tendency to rotate
counterclockwise is taken to be a positive moment. Thus the moment in Fig. 2.3.1 is
positive. The units of moment are the Newton metre (Nm)
Note that when the line of action of a force goes through the axis, the moment is zero.

Section 2.3
It should be emphasized that there is not actually a physical axis, such as a rod, at the
point o of Fig. 2.3.1; in this discussion, it is imagined that an axis is there.
Two forces of equal magnitude and acting along the same line of action have not only the
same components Fx , Fy , but have equal moments about any axis. They are called
equivalent forces since they have the same effect on a rigid body. This is illustrated in
Fig. 2.3.2.
Rigid body
F1 line of action
o
of force
d 
 F2
Figure 2.3.2: Two equivalent forces
Consider next the case of two forces of equal magnitude, parallel lines of action separated
by distance d, and opposite sense. Any two such forces are said to form a couple. The
only motion that a couple can impart is a rotation; unlike the forces of Fig. 2.3.2, the
couple has no tendency to translate a rigid body. The moment of the couple of Fig. 2.3.3
about o is
M o  Fd 2  Fd1  Fd (2.3.2)
o
d1
F  d
d2  F
Figure 2.3.3: A couple
The sign convention which will be followed in most of what follows is that a couple is
positive when it acts in a counterclockwise sense, as in Fig. 2.3.3.
It is straight forward to show the following three important properties of couples:

(a) the moment of Fig. 2.3.3 is also Fd about any axis in the rigid body, and so can be
represented by M, without the subscript. In other words, this moment of the couple is
independent of the choice of axis. {see ▲Problem 1}
(b) any two different couples having the same moment M are equivalent, in the sense that
they tend to rotate the body in precisely the same way; it does not matter that the

Section 2.3
forces forming these couples might have different magnitudes, act in different
directions and have different distances between them.
(c) any two couples may be replaced by a single couple of moment equal to the algebraic
sum of the moments of the individual couples.
Example
Consider the two couples shown in Fig. 2.3.4a. These couples can conveniently be
represented schematically by semi-circular arrows, as shown in Fig. 2.3.4b. They can
also be denoted by the letter M, the magnitude of their moment, since the magnitude of
the forces and their separation is unimportant, only their product. In this example, if the
body is in static equilibrium, the couples must be equal and opposite, M 2   M 1 , i.e. the
sum of the moments is zero and the net effect is to impart zero rotation on the body.
Note that the curved arrow for M 2 has been drawn counterclockwise, even though it is
negative. It could have been illustrated as in Fig. 2.3.4c, but the version of 2.3.4b is
preferable as it is more consistent and reduces the likelihood of making errors when
solving problems (see later).
d2  F2
M 2   F2 d 2
F2

F1 
d1
 F1 M 1  F1d1 M 2   F2 d 2
(a) (b) (c)
Figure 2.3.4: Two couples acting on a rigid body
A final point to be made regarding couples is the following: any force is equivalent to (i)
a force acting at any (other) point and (ii) a couple. This is illustrated in Fig. 2.3.5.
Referring to Fig. 2.3.5, a force F acts at position A. This force tends to translate the rigid
body along its line of action and also to rotate it about any chosen axis. The system of
forces in Fig. 2.3.5b are equivalent to those in Fig. 2.3.5a: a set of equal and opposite
forces have simply been added at position B. Now the force at A and one of the forces at
B form a couple, of moment M say. As in the previous example, the couple can
conveniently be represented by a curved arrow, and the letter M. For illustrative
purposes, the curved arrow is usually grouped with the force F at B, as shown in Fig.
2.3.5c. However, note that the curved arrow representing the moment of a couple, which
can be placed anywhere and have the same effect, is not associated with any particular
point in the rigid body.

Section 2.3
B B
F F  F
F  F
 d  M  Fd
A A
(a) (b) (c)
Figure 2.3.5: Equivalents force/moment systems; (a) a force F, (b)

an equivalent system to (a), (c) an equivalent system involving a
force and a couple M
Note that if the force at A was moved to a position other than B, the moment M of Fig.
2.3.5c would be different.
Example
Consider the spanner and bolt system shown in Fig. 2.3.6. A downward force of 200N is
applied at the point shown. This force can be replaced by a force acting somewhere else,
together with a moment. For the case of the force moved to the bolt-centre, the moment
has the magnitude shown in Fig. 2.3.6b.
20 cm
200 N
M  40 N m 40 N m
200 N
(a) (b) (c)
Figure 2.3.6: Equivalent force and force/moment acting on a

spanner and bolt system
As mentioned, it is best to maintain consistency and draw the semi-circle representing the
moment counterclockwise (positive) and given a value of  40 as in Fig. 2.3.6b; rather
than as in Fig. 2.3.6c.
■
Example
Consider the plate subjected to the four external loads shown in Fig. 2.3.7a. An
equivalent force-couple system F-M, with the force acting at the centre of the plate, can
be calculated through

Section 2.3
F x  200 N, F y  100 N
M o  (100)(100)  (50 / 2)(100)  (50 / 2)(100)  (200)(50)   7071.07 Nmm
and is shown in Fig. 2.3.7b. A resultant force R can also be derived, that is, an
equivalent force positioned so that a couple is not necessary, as shown in Fig. 2.3.7.c.
100 N
R  223.61 N
50 N o
F  223.61 N
45
o o
50 mm
100 mm  d  31.62 mm 
50 N
45o
200 N M  7071.07 Nmm
200 mm
(a) (b) (c)
Figure 2.3.7: Forces acting on a plate; (a) individual forces, (b) an

equivalent force-couple system at the plate-centre, (c) the resultant
force
The force systems in the three figures are equivalent in the sense that they tend to impart
(a) the same translation in the x direction, (b) the same translation in the y direction, and
(c) the same rotation about any given point in the plate. For example, the moment about
the upper left corner is
Fig 2.3.7a: (100)(0)  (50 / 2)(50)  (50 / 2)(150)  (200)(100)

Fig 2.3.7b:  (223.61)(89.44)  7071
Fig 2.3.7c:  (223.61)(57.82)
all leading to M  12928.93 Nmm about that point.

■
2.3.2 Equilibrium of Rigid Bodies

The concept of equilibrium encountered earlier in the context of particles can now be
generalized to the case of the rigid body:
Equilibrium of a Rigid Body

A rigid body is in equilibrium when the external forces acting on it form a system of
forces equivalent to zero

Section 2.3
The necessary and sufficient conditions that a (two dimensional) rigid body is in
equilibrium are then
F x  0, F y  0, M o  0 Equilibrium Equations (2D Rigid Body) (2.3.3)
that is, there is no resultant force and no resultant moment. Note that the x  y axes and
the axis of rotation o can be chosen completely arbitrarily: if the resultant force is zero,
and the resultant moment about one axis is zero, then the resultant moment about any
other axis in the body will be zero also.
2.3.3 Joints and Connections

Components in machinery, buildings etc., connect with each other and are supported in a
number of different ways. In order to solve for the forces acting in such assemblies, one
must be able to analyse the forces acting at such connections/supports.
One of the most commonly occurring supports can be idealised as a roller support, Fig.
2.3.8a. Here, the contacting surfaces are smooth and the roller offers only a normal
reaction force (see §2.2.2). This reaction force is labelled R y , according to the
conventional x  y coordinate system shown. This is shown in the free-body diagram of
the component.
hinge
roller pin
Rx
y Rx
Ry Ry
Ry
M
x
(a) (b) (c)
Figure 2.3.8: Supports and connections; (a) roller support, (b) pin
joint, (c) clamped
Another commonly occurring connection is the pin joint, Fig. 2.3.8b. Here, the
component is connected to a fixed hinge by a pin (going “into the page”). The
component is thus constrained to move in one plane, and the joint does not provide
resistance to this turning movement. The underlying support transmits a reaction force

Section 2.3
through the hinge pin to the component, which can have both normal ( R y ) and tangential
( R x ) components.
Finally, in Fig. 2.3.8c is shown a fixed (clamped) joint. Here the component is welded
or glued and cannot move at the base. It is said to be cantilevered. The support in this
case reacts with normal and tangential forces, but also with a couple of moment M, which
resists any bending/turning at the base.
Example
For example, consider such a component loaded with a force F a distance L from the
base, as shown in Fig. 2.3.9a. A free-body diagram of the component is shown in Fig.
2.3.9b. The known force F acts on the body and so do two unknown forces R x , R y , and
a couple of moment M. The unknown forces and moment will be called reactions
henceforth. If the component is static, the equilibrium equations 2.3.3 apply; one has,
taking moments about the base of the component,
F x  F  R x  0, F y  R y  0, M o   FL  M  0
and so
R x   F , R y  0, M  FL
The moment is positive and so acts in the direction shown in the Figure.
F F
L L
Rx o
Ry
M
(a) (b)
Figure 2.3.9: A loaded cantilevered component; (a) loaded

component, (b) free body diagram of the component
The reaction moment of Fig. 2.3.9(b) can be experienced as follows: take a ruler and hold
it firmly at one end, upright in your right hand. Simulate the applied force now by
pushing against the ruler with a finger of your left hand. You will feel that, to maintain
the ruler “vertical” at the base, you need to apply a twist with your right hand, in the
direction of the moment shown in Fig. 2.3.9(b).
Note that, when solving this problem, moments were taken about the base. As mentioned
already, one can take the moment about any point in the column. For example, taking the
moment about the point where the force F is applied, one has

Section 2.3
M F  Rx L  M  0
This of course leads to the same result as before, but the final calculation of the forces is
now slightly more complicated; in general, it is easier if the axis is chosen to coincide
with the point where the reaction forces act – this is because the reaction forces do not
then appear in the moment equation:  M o   FL  M  0 .
■
For ease of discussion, from now on, “couples” such as that encountered in Fig. 2.3.9 will
simply be called “moments”.
All the elements are now in place to tackle fairly complex static rigid body problems.
Example
Consider the plate subjected to the three external loads shown in Fig. 2.3.10a. The plate
is supported by a roller at A and a pin-joint at B. The weight of the plate is assumed to be
small relative to the applied loads and is neglected. A free body diagram of the plate is
shown in Fig 2.3.10b. This shows all the forces acting on the plate. Reactions act at A
and B: these forces represent the action of the base on the plate, preventing it from
moving downward and horizontally. The equilibrium equations can be used to find the
reactions:
F  F  0  F  0
x xB xB
 F   F  150  100  50  F  0  F  F  200 N

y yA yB yA yB
 M  (150)(50)  (100)(120)  (50)(200)  F (200)  0 

A yB FyB  47.5 N ,
 FyA  152.5 N
150 N 100 N 150 N 100 N

50 mm 70 mm
80 mm
50 N 50 N
150 mm
100 mm
 FxB
A B
FyA FyB
200 mm
(a) (b)
Figure 2.3.10: Equilibrium of a plate; (a) forces acting on the plate,

(b) free-body diagram of the plate
The resultant moment was calculated by taking the moment about point A. As mentioned
in relation to the previous example, one could have taken the moment about any other

Section 2.3
point in the plate. The “most convenient” point about which to take moments in this
example would be point A or B, since in that case only one of the reaction forces will
appear in the moment equilibrium equation.
■
In the above example there were three unknown reactions and three equilibrium equations
with which to find them. If the roller was replaced with a pin, there would be four
unknown reactions, and now there would not be enough equations with which to find the
reactions. When this situation arises, the system is called statically indeterminate. To
find the unknown reactions, one must relax the assumption of rigidity, and take into
account the fact that all materials deform. By calculating deformations within the plate,
the reactions can be evaluated. The deformation of materials is studied in the following
chapters.
To end this Chapter, note the following:

(i) the equilibrium equations 2.3.3 result from Newton’s laws, and are thus as valid for
a body of water as they are for a body of hard steel; the external forces acting on a
body of still water form a system of forces equivalent to zero.
(ii) as mentioned already, Newton’s laws apply not only to a complete body or
structure, but to any portion of a body. The external forces acting on any free-body
portion of static material form a system of forces equivalent to zero.
(iii) there is no such thing as a rigid body. Metals and other engineering materials can
be considered to be “nearly rigid” as they do not deform by much under even fairly
large loads. The analysis carried out in this Chapter is particularly relevant to these
materials and in answering questions like: what forces act in the steel members of a
suspension bridge under the load of self-weight and traffic? (which is just a more
complicated version of the problem of Fig. 2.2.3 or Problem 3 below).
(iv) if the loads on the plate of Fig. 2.3.10a are too large, the plate will “break”. The
analysis carried out in this Chapter cannot answer where it will break or when it will
break. The more sophisticated analysis carried out in the following Chapters is
necessary to deal with this and many other questions of material response.
2.3.4 Problems
1. A plate is subjected to a couple Fd , with d  20cm , as shown below left. Verify

that the couple can be moved to the position shown below right, and the effect on the
plate is the same, by showing that the moment about point o in both cases is
M  20 F .
30 cm 20 cm
100 cm 20 cm 100 cm
F F
F
F 30 cm
o o
100 cm 100 cm

Section 2.3
2. What force F must be applied to the following static component such that the tension
in the cable, T, is 1kN? What are the reactions at the pin support C?
150 mm
T 250 mm
150 mm F
C
3. A machine part is hinged at A and subjected to two forces through cables as shown.
What couple M needs to be applied to the machine part for equilibrium to be
maintained? Where can this couple be applied?
100 N
100 mm
A
 
75 mm M

50 N

3 Stress
Forces acting at the surfaces of components were considered in the previous chapter. The
task now is to examine forces arising inside materials, internal forces. Internal forces are
described using the idea of stress. There is a lot more to stress than the notion of “force
over area”, as will become clear in this chapter. First, the idea of surface (contact) stress
distributions will be examined, together with their relationship to resultant forces and
moments. Then internal stress and traction will be discussed. The means by which
internal forces are described is through the stress components, for example  zx ,  yy , and
this “language” of sigmas and subscripts needs to be mastered in order to model sensibly
the internal forces in real materials. Stress analysis involves representing the actual
internal forces in a real physical component mathematically. Some of the limitations of
this are discussed in §3.3.2.
Newton’s laws are used to derive the stress transformation equations, and these are
then used to derive expressions for the principal stresses, stress invariants, principal
directions and maximum shear stresses acting at a material particle. The practical case
of two dimensional plane stress is discussed.
29
30
Section 3.1
3.1 Surface and Contact Stress

The concept of the force is fundamental to mechanics and many important problems
can be cast in terms of forces only, for example the problems considered in Chapter 2.
However, more sophisticated problems require that the action of forces be described
in terms of stress, that is, force divided by area. For example, if one hangs an object
from a rope, it is not the weight of the object which determines whether the rope will
break, but the weight divided by the cross-sectional area of the rope, a fact noted by
Galileo in 1638.
3.1.1 Stress Distributions

As an introduction to the idea of stress, consider the situation shown in Fig. 3.1.1a: a
block of mass m and cross sectional area A sits on a bench. Following the
methodology of Chapter 2, an analysis of a free-body of the block shows that a force
equal to the weight mg acts upward on the block, Fig. 3.1.1b. Allowing for more
detail now, this force will actually be distributed over the surface of the block, as
indicated in Fig. 3.1.1c. Defining the stress to be force divided by area, the stress
acting on the block is
mg
 (3.1.1)
A
The unit of stress is the Pascal (Pa): 1Pa is equivalent to a force of 1 Newton acting
over an area of 1 metre squared. Typical units used in engineering applications are
the kilopascal, kPa (10 3 Pa ), the megapascal, MPa ( 10 6 Pa ) and the gigapascal, GPa
( 10 9 Pa ).

mg mg
(a) (b) (c)
Figure 3.1.1: a block resting on a bench; (a) weight of the block, (b) reaction of
the bench on the block, (c) stress distribution acting on the block
The stress distribution of Fig. 3.1.1c acts on the block. By Newton’s third law, an
equal and opposite stress distribution is exerted by the block on the bench; one says
that the weight force of the block is transmitted to the underlying bench.
The stress distribution of Fig. 3.1.1 is uniform, i.e. constant everywhere over the
surface. In more complex and interesting situations in which materials contact, one is
more likely to obtain a non-uniform distribution of stress. For example, consider the
case of a metal ball being pushed into a similarly stiff object by a force F, as

Section 3.1
illustrated in Fig. 3.1.2.1 Again, an equal force F acts on the underside of the ball,
Fig. 3.1.2b. As with the block, the force will actually be distributed over a contact
region. It will be shown in Part II that the ball (and the large object) will deform and
a circular contact region will arise where the ball and object meet2, and that the stress
is largest at the centre of the contact surface, dying away to zero at the edges of
contact, Fig. 3.1.2c (  1   2 in Fig. 3.1.2c). In this case, with stress  not constant,
one can only write, Fig. 3.1.2d,
F   dF    dA (3.1.2)
A A
The stress varies from point to point over the surface but the sum (or integral) of the
stresses (times areas) equals the total force applied to the ball.
F F
contact Small
region region dA
F 2 
1
(a) (b) (c) (d)
Figure 3.1.2: a ball being forced into a large object, (a) force applied to ball, (b)
reaction of object on ball, (c) a non-uniform stress distribution over the
contacting surface, (d) the stress acting on a small (infinitesimal) area
A given stress distribution gives rise to a resultant force, which is obtained by

integration, Eqn. 3.1.2. It will also give rise to a resultant moment. This is examined
in the following example.
Example
Consider the surface shown in Fig. 3.1.3, of length 2m and depth 2m (into the page).
The stress over the surface is given by   x kPa, with x measured in m from the left-
hand side of the surface.
The force acting on an element of length dx at position x is (see Fig. 3.1.3b)
dF   dA   x kPa    dx m  2 m 
1
the weight of the ball is neglected here
2
the radius of which depends on the force applied and the materials in contact

Section 3.1
The resultant force is then, from Eqn. 3.1.2
2
F   dF  2  xdx kPa m 2  4 kN  
A 0
The moment of the stress distribution is given by
M 0   dM     l dA (3.1.3)
A A
where l is the length of the moment-arm from the chosen axis.
Taking the axis to be at x  0 , the moment-arm is l  x , Fig. 3.1.3b, and
2
M x 0   dM  2  x  x dx kPa m 3    16
3
kN m
A 0
Taking moments about the right-hand end, x  2 , one has
2
8
M x  2   dM  2  x  2  x  dx kPa m 3   kN m
3
 
A 0
 (x)
x  (x)
x
2m dx
(a) (b)
Figure 3.1.3: a non-uniform stress acting over a surface; (a) the stress
distribution, (b) stress acting on an element of size dx
3.1.2 Equivalent Forces and Moments
Sometimes it is useful to replace a stress distribution  with an equivalent force F,

i.e. a force equal to the resultant force of the distribution and one which also give the
same moment about any axis as the distribution. Formulae for equivalent forces are
derived in what follows for triangular and arbitrary linear stress distributions.

Section 3.1
Triangular Stress Distribution
Consider the triangular stress distribution shown in Fig. 3.1.4. The stress at the end is
 0 , the length of the distribution is L and the thickness “into the page” is t. The
equivalent force is, from Eqn. 3.1.2,
L
x 1
F  t 0  dx   0 Lt (3.1.4)
0
L 2
which is just the average stress times area. The point of action of this force should be
such that the moment of the force is equivalent to the moment of the stress
distribution. Taking moments about the left hand end, for the distribution one has,
from 3.1.3,
L
1
M o  t  x ( x)dx   0 L2 t
0
3
Placing the force at position x  xc , Fig. 3.1.4, the moment of the force is
M o   0 Lt / 2xc . Equating these expressions leads to the position at which the
equivalent force acts:
2
xc  L. (3.1.5)
3
equivalent force
 (x)
0
x
o xc
L
Figure 3.1.4: triangular stress distribution and equivalent force
Note that the moment about any axis is now the same for both the stress distribution
and the equivalent force.
■
Arbitrary Linear Stress Distribution
Consider the linear stress distribution shown in Fig. 3.1.5. The stress at the ends are
 1 and  2 and this time the equivalent force is
L
F  t   1  ( 2   1 )( x / L)dx  Lt  1   2  / 2 (3.1.6)
0

Section 3.1
Taking moments about the left hand end, for the distribution one has
L
M o  t  x ( x)dx  L2 t  1  2 2  / 6
0
The moment of the force is M o  Lt  1   2 xc / 2 . Equating these expressions leads

to
L 1  2 2 
xc  (3.1.7)
3 1   2 
Eqn. 3.1.5 follows from 3.1.7 by setting  1  0 .
equivalent force
 (x)
2
1
o xc
L
Figure 3.1.5: a non-uniform stress distribution and equivalent force
The Centroid
Generalising the above cases, the line of action of the equivalent force for any
arbitrary stress distribution  (x) is
t  x ( x)dx  xdF
xc   Centroid (3.1.8)
t   ( x)dx F
This location is known as the centroid of the distribution.
Note that most of the discussion above is for two-dimensional cases, i.e. the stress is
assumed constant “into the page”. Three dimensional problems can be tackled in the
same way, only now one must integrate two-dimensionally over a surface rather than
one-dimensionally over a line.
Also, the forces considered thus far are normal forces, where the force acts
perpendicular to a surface, and they give rise to normal stresses. Normal stresses are
also called pressures when they are compressive as in Figs. 3.1.1-2.

Section 3.1
3.1.3 Shear Stress

Consider now the case of shear forces, that is, forces which act tangentially to
surfaces.
A normal force F acts on the block of Fig. 3.1.6a. The block does not move and, to
maintain equilibrium, the force is resisted by a friction force F  mg , where  is
the coefficient of friction. A free body diagram of the block is shown in Fig. 3.1.6b.
Assuming a uniform distribution of stress, the stress and resultant force arising on the
surfaces of the block and underlying object are as shown. The stresses are in this case
called shear stresses.
F F
F
(a) (b)
Figure 3.1.6: shear stress; (a) a force acting on a block, (b) shear stresses arising
on the contacting surfaces
3.1.4 Combined Normal and Shear Stress

Forces acting inclined to a surface are most conveniently described by decomposing
the force into components normal and tangential to the surface. Then one has both
normal stress  N and shear stress  S , as in Fig. 3.1.7.

N
S
Figure 3.1.7: a force F giving rise to normal and shear stress over the contacting
surfaces
The stresses considered in this section are examples of surface stresses or contact
stresses. They arise when materials meet at a common surface. Other examples
would be sea-water pressurising a material in deep water and the stress exerted by a
train wheel on a train track.

Section 3.1
3.1.5 Problems
1. Consider the surface shown below, of length 4cm and unit depth (1cm into the
page). The stress over the surface is given by   2  x kPa, with x measured in
cm from the surface centre.
(a) Evaluate the resultant force acting on the surface (in Newtons).
(b) What is the moment about an axis (into the page) through the left-hand end of
the surface?
(c) What is the moment about an axis (into the page) through the centre of the
surface?
y x
 ( x)
2. Consider the surface shown below, of length 4mm and unit depth (1mm into the
page). The stress over the surface is given by   x MPa, with x measured from
the surface centre. What is the total force acting on the surface, and the moment
acting about the centre of the surface?
y
x
3. Find the reaction forces (per unit length) at the pin and roller for the following
beam, which is subjected to a varying pressure distribution, the maximum
pressure being  ( x)  20 kPa (all lengths are in cm – give answer in N/m)
[Hint: first replace the stress distribution with three equivalent forces]
4 12 8
4 16 4

Section 3.1
4. A block of material of width 10cm and length 1m is pushed into an underlying

substrate by a normal force of 100 N. It is found that a uniform triangular normal
stress distribution arises at the contacting surfaces, that is, the stress is maximum
at the centre and dies off linearly to zero at the block edges, as sketched below
right. What is the maximum pressure acting on the surface?
100 N typical cross-

section

1m
stress
10 cm distribution

Section 3.2
3.2 Body Forces

Surface forces act on surfaces. As discussed in the previous section, these are the forces
which arise when bodies are in contact and which give rise to stress distributions. Surface
forces also arise inside materials, acting on internal surfaces, Fig. 3.2.1a, as will be
discussed in the following section.
To complete the description of forces acting on real materials, one needs to deal with
forces which arise even when bodies are not in contact; one can think of these forces as
acting at a distance, for example the force of gravity. To describe these forces, one can
define the body force, which acts on volume elements of material. Fig. 3.2.1b shows a
sketch of a volume element subjected to a magnetic body force and a gravitational body
force Fg .
contact
F force
F internal
surface
volume
element
F e.g. air
pressure Fg
(a) (b)
Figure 3.2.1: forces acting on a body; (a) surface forces acting on surfaces, (b) body
forces acting on a material volume element
3.2.1 Weight
The most important body force is the force due to gravity, i.e. the weight force. In
Chapter 2 there were examples involving the weight of components. In those cases it was
simply stated that the weight could be taken to be a single force acting at the component
centre (for example, Problem 3 in §2.2.3). This is true when the component is
symmetrical, for example, in the shape of a circle or a square. However, it is not true in
general for a component of arbitrary shape.
In what follows, the important case of a flat object of arbitrary shape will be examined.
The weight of a small volume element V of material of density  is dF  gV and

the total weight is

Section 3.2
F   gdV (3.2.1)
V
Consider the general two-dimensional case, Fig. 3.2.2, where material elements of area
Ai (and constant thickness t) are subjected to forces Fi  tgAi .
y z
x
 xc , yc 
Ai
Fi  tgAi
t
Fg
Figure 3.2.2: Resultant Weight on a body
The resultant, i.e. equivalent, weight force due to all elements, for a component with
uniform density, is
F   dF  tg  dA  gtA ,
where A is the cross-sectional area.
The resultant moments about the x and y axes, which can be positioned anywhere in the
body, are M x  tg  ydA and M y  tg  xdA respectively; the moment M x is shown
in Fig. 3.2.3. The equivalent weight force is thus positioned at ( x c , y c ) , Fig. 3.2.2, where
xc 
 xdA , yc 
 ydA Centroid of Area (3.2.2)
A A
The position ( x c , y c ) is called the centroid of the area. The quantities  xdA ,  ydA , are
called the first moments of area about, respectively, the y and x axes.

Section 3.2
z
y Ai
x
y M x
Ai
y o
tgAi y
tgAi
(a) (b)
Figure 3.2.3: The moment Mx; (a) full view, (b) plane view
3.2.2 Problems
1. Where does the resultant force due to gravity act in the triangular component shown
below? (Gravity acts downward in the direction of the arrow shown, perpendicular to
the component’s surface)
90 o
1m
1m

Section 3.3
3.3 Internal Stress

The idea of stress considered in §3.1 is not difficult to
conceptualise since objects interacting with other objects are
encountered all around us. A more difficult concept is the
idea of forces and stresses acting inside a material, “within
the interior where neither eye nor experiment can reach” as
Euler put it. It took many great minds working for centuries
on this question to arrive at the concept of stress we use
today, an idea finally brought to us by Augustin Cauchy,
who presented a paper on the subject to the Academy of
Sciences in Paris, in 1822.
Augustin Cauchy
3.3.1 Cauchy’s Concept of Stress
Uniform Internal Stress
Consider first a long slender block of material subject to equilibrating forces F at its ends,
Fig. 3.3.1a. If the complete block is in equilibrium, then any sub-division of the block
must be in equilibrium also. By imagining the block to be cut in two, and considering
free-body diagrams of each half, as in Fig. 3.3.1b, one can see that forces F must be
acting within the block so that each half is in equilibrium. Thus external loads create
internal forces; internal forces represent the action of one part of a material on another
part of the same material across an internal surface. If the material out of which the block
is made is uniform over this cut, one can take it that a uniform stress   F / A acts over
this interior surface, Fig. 3.3.1b.
F F F F
F
imaginary F F N
F
cut  F
F A S
F
F F F F
(a ) ( b) (c)
Figure 3.3.1: a slender block of material; (a) under the action of external forces F,
(b) internal normal stress σ, (c) internal normal and shear stress

Section 3.3
Note that, if the internal forces were not acting over the internal surfaces, the two half-
blocks of Fig. 3.3.1b would fly apart; one can thus regard the internal forces as those
required to maintain material in an un-cut state.
If the internal surface is at an incline, as in Fig. 3.3.1c, then the internal force required for
equilibrium will not act normal to the surface. There will be components of the force
normal and tangential to the surface, and thus both normal (  N ) and shear (  S ) stresses
must arise. Thus, even though the material is subjected to a purely normal load, internal
shear stresses develop.
From Fig. 3.3.2a, the normal and shear stresses arising on an interior surface inclined at
angle  to the horizontal are {▲Problem 1}
F F
N  cos 2  , S  sin  cos  (3.3.1)
A A
F
internal
point
F
FN
 FS  internal
surface
 
A
F
(a ) ( b)
Figure 3.3.2: stress on inclined surface; (a) decomposing the force into normal and
shear forces, (b) stress at an internal point
Although stress is associated with surfaces, one can speak of the stress “at a point”. For
example, consider some point interior to the block, Fig 3.3.2b. The stress there evidently
depends on which surface through that point is under consideration. From Eqn. 3.3.1a,
the normal stress at the point is a maximum F / A when   0 and a minimum of zero
when   90 o . The maximum normal stress arising at a point within a material is of
special significance, for example it is this stress value which often determines whether a
material will fail (“break”) there. It has a special name: the maximum principal stress.
From Eqn. 3.3.1b, the maximum shear stress at the point is  F / 2 A and arises on
surfaces inclined at  45 o .
Non-Uniform Internal Stress
Consider a more complex geometry under a more complex loading, as in Fig. 3.3.3.
Again, using equilibrium arguments, there will be some stress distribution acting over any
given internal surface. To evaluate these stresses is not an easy matter, and much of Part

Section 3.3
II is devoted to doing just that. Suffice to say here that they will invariably be non-
uniform over a surface, that is, the stress at some particle will differ from the stress at a
neighbouring particle.
F2 F3
F4
N
F1
S
Figure 3.3.3: a component subjected to a complex loading, giving rise to a non-

uniform stress distribution over an internal surface
Traction and the Physical Meaning of Internal Stress
All materials have a complex molecular microstructure and each molecule exerts a force
on each of its neighbours. The complex interaction of countless molecular forces
maintains a body in equilibrium in its unstressed state. When the body is disturbed and
deformed into a new equilibrium position, net forces act, Fig. 3.3.4a. An imaginary plane
can be drawn through the material, Fig. 3.3.4b. Unlike some of his predecessors, who
attempted the extremely difficult task of accounting for all the molecular forces, Cauchy
discounted the molecular structure of matter and simply replaced the molecular forces
acting on the plane by a single force F, Fig 3.3.4c. This is the force exerted by the
molecules above the plane on the material below the plane and can be attractive or
repulsive. Different planes can be taken through the same portion of material and, in
general, a different force will act on the plane, Fig 3.3.4d.
F F
(a ) ( b) (c ) (d )
Figure 3.3.4: a multitude of molecular forces represented by a single force; (a)

molecular forces, a plane drawn through the material, replacing the molecular
forces with an equivalent force F, a different equivalent force F acts on a different
plane through the same material
The definition of stress will now be made more precise. First, define the traction at some
particular point in a material as follows: take a plane of surface area S through the point,
on which acts a force F. Next shrink the plane – as it shrinks in size both S and F get
smaller, and the direction in which the force acts may change, but eventually the ratio
F / S will remain constant and the force will act in a particular direction, Fig. 3.3.5. The

Section 3.3
limiting value of this ratio of force over surface area is defined as the traction vector (or
stress vector) t:
F
t  lim (3.3.2)
S  0 S
F
S
F
a plane passing through some S
point in the material
F
S
Figure 3.3.5: the traction vector - the limiting value of force over area, as the surface
area of the element on which the force acts is shrunk
An infinite number of traction vectors act at any single point, since an infinite number of
different planes pass through a point. Thus the notation lim S 0 F / S is ambiguous.
For this reason the plane on which the traction vector acts must be specified; this can be
done by specifying the normal n to the surface on which the traction acts, Fig 3.3.6. The
traction is thus a special vector – associated with it is not only the direction in which it
acts but also a second direction, the normal to the plane upon which it acts.
n1 same point with different planes n2

passing through it
(defined by different normals)
F F
n1 different forces act on n2
different planes through
S the same point
S
F F
t (n1 )  lim t (n 2 )  lim
S 0 S S 0 S
Figure 3.3.6: two different traction vectors acting at the same point

Section 3.3
Stress Components
The traction vector can be decomposed into components which act normal and parallel to
the surface upon which it acts. These components are called the stress components, or
simply stresses, and are denoted by the symbol  ; subscripts are added to signify the
surface on which the stresses act and the directions in which the stresses act.
Consider a particular traction vector acting on a surface element. Introduce a Cartesian

coordinate system with base vectors i, j, k so that one of the base vectors is a normal to
the surface, and the origin of the coordinate system is positioned at the point at which the
traction acts. For example, in Fig. 3.3.7, the k direction is taken to be normal to the
plane, and t ( k )  t x i  t y j  t z k .
t (nˆ ) z t (k )
tz
k
j
i
y
ty tx
x
Figure 3.3.7: the components of the traction vector
Each of these components t i is represented by  ij where the first subscript denotes the
direction of the normal to the plane and the second denotes the direction of the
component. Thus, re-drawing Fig. 3.3.7 as Fig. 3.3.8: t ( k )   zx i   zy j   zz k . The first
two stresses, the components acting tangential to the surface, are shear stresses, whereas
 zz , acting normal to the plane, is a normal stress1.
z t (k )
 zz
k
j
i
y
 zy  zx
x
Figure 3.3.8: stress components – the components of the traction vector
The traction vector shown in Figs. 3.3.7, 3.3.8, represents the force (per unit area) exerted
by the material above the surface on the material below the surface. By Newton’s third
1
this convention for the subscripts is not universally followed. Many authors, particularly in the
mathematical community, use the exact opposite convention, the first subscript to denote the direction and
the second to denote the normal. It turns out that both conventions are equivalent, since, as will be shown
later, the stress is symmetric, i.e.  ij   ji

Section 3.3
law, an equal and opposite traction must be exerted by the material below the surface on
the material above the surface, as shown in Fig. 3.3.9 (thick dotted line). If t ( k ) has stress
components  zx ,  zy ,  zz , then so should t ( k ) : t (  k )   zx (  i )   zy (  j)   zz (  k )   t ( k ) .
z
t (k )
 zx
 zy k  zz
j
 zz i
y
 zy  zx
( k )
t
x
Figure 3.3.9: equal and opposite traction vectors – each with the same stress
components
Sign Convention for Stress Components
The following convention is used:
The stress is positive when the direction of the normal and the direction of
the stress component are both positive or both negative
The stress is negative when one of the directions is positive and the other is
negative
According to this convention, the three stresses in Figs. 3.3.7-9 are all positive.
Looking at the two-dimensional case for ease of visualisation, the (positive and negative)
normal stresses and shear stresses on either side of a surface are as shown in Fig. 3.3.10.
Normal stresses which “pull” (tension) are positive; normal stresses which “push
(compression) are negative. Note that the shear stresses always go in opposite directions.
 yy  yy
 yx  yx
y
 yx  yx
 yy x  yy
(a ) (b)
Figure 3.3.10: stresses acting on either side of a material surface: (a) positive
stresses, (b) negative stresses
Examples of negative stresses are shown in Fig. 3.3.11 {▲Problem 4}.

Section 3.3
x
y
i t ( j)
y e2j
 yy
k j  yx
k  yz
 yz i
x
z  yx z
 yy t (  j)
t (  j)   yx i   yy j   yz k t ( j)   yx i   yy j   yz k
(a ) ( b)
Figure 3.3.11: examples of negative stress components
3.3.2 Real Problems and Saint-Venant’s Principle

Some examples have been given earlier of external forces acting on materials. In reality,
an external force will be applied to a real material component in a complex way. For
example, suppose that a block of material, welded to a large object at one end, is pulled at
its other end by a rope attached to a metal hoop, which is itself attached to the block by a
number of bolts, Fig. 3.3.12a. The block can be idealised as in Fig 3.3.12b; here, the
precise details of the region in which the external force is applied are neglected.
(a ) F
(e) F
( b) F
(f ) F
stress the stress
(c) F
same differs here
 F/A F/2
(g )
F/2
(d ) F
Figure 3.3.12: a block subjected to an external force: (a) real case, (b) ideal model,
(c) stress in ideal model, (d) stress in actual material, (e) the stress in the real
material, away from the right hand end, is modelled well by either (f) or (g)
According to the earlier discussion, the stress in the ideal model is as in Fig. 3.3.12c. One
will find that, in the real material, the stress is indeed (approximately) as predicted, but

Section 3.3
only at an appreciable distance from the right hand end. Near where the rope is attached,
the force will differ considerably, as sketched in Fig.3.3.12d.
Thus the ideal models of the type discussed in this section, and in much of this book, are
useful only in predicting the stress field in real components in regions away from points
of application of loads. This does not present too much of a problem, since the stresses
internal to a structure in such regions are often of most interest. If one wants to know
what happens near the bolted connection, then one will have to create a complex model
incorporating all the details and the problem will be more difficult to solve.
It is an experimental fact that if two different force systems are applied to a material, but
they are equivalent force systems, as in Fig. 3.3.12(f,g), then the stress fields in regions
away from where the loads are applied will be the same. This is known as Saint-
Venant’s Principle. Typically, one needs to move a distance away from where the loads
are applied roughly equal to the distance over which the loads are applied.
3.3.3 Problems
1. Derive Eqns. 3.3.1.
2. The four sides of a square block are subjected to equal forces S, as illustrated. The
length of each side is l and the block has unit depth (into the page). What normal and
shear stresses act along the (dotted) diagonal? [Hint: draw a free body diagram of the
upper left hand triangle.]
S
S S
S
3. A shaft is concreted firmly into the ground. A thick steel rope is looped around the
shaft and a force is applied normal to the shaft, as shown. The shaft is in static
equilibrium. Draw a free body diagram of the shaft (from the top down to ground
level) showing the forces/moments acting on the shaft (including the reaction forces at
the ground-level; ignore the weight of the shaft). Draw a free body diagram of the
section of shaft from the top down to the cross section at A. Draw a free body
diagram of the section of shaft from the top to the cross section at B. Roughly sketch
the stresses acting over the (horizontal) internal surfaces of the shaft at A and B.
B
F
A
ground

Section 3.3
4. In Fig. 3.3.11, which of the stress components is/are negative?

5. Label the following stress component acting on an internal material surface. Is it a
positive or negative stress?
z
acting parallel
to surface
6. Label the following shear stresses. Are they positive or negative?
7. Label the following normal stresses. Are they positive or negative?
8. By the definition of the traction vector t which acts on the x  z plane,

t ( j)   yx i   yy j   yz k . Sketch these three stress components on the figure below.

Section 3.3

Section 3.4
3.4 Equilibrium of Stress

Consider two perpendicular planes passing through a point p. The stress components
acting on these planes are as shown in Fig. 3.4.1a. These stresses are usually shown
together acting on a small material element of finite size, Fig. 3.4.1b. It has been seen
that the stress may vary from point to point in a material but, if the element is very small,
the stresses on one side can be taken to be (more or less) equal to the stresses acting on
the other side. By convention, in analyses of the type which will follow, all stress
components shown are positive.
p
x
 yy
 yy
 yx
 yx  xy  xy
 xx  xx
 yx  xx  xy
 xx
 xy
 yx
 yy
 yy
(a ) (b)
Figure 3.4.1: stress components acting on two perpendicular planes through a point;
(a) two perpendicular surfaces at a point, (b) small material element at the point
The four stresses can conveniently be written in the matrix form:
 xx  xy 
 ij     (3.4.1)
 yx  yy 
It will be shown below that the stress components acting on any other plane through p can
be evaluated from a knowledge of only these stress components.
3.4.1 Symmetry of the Shear Stress

Consider the material element shown in Fig. 3.4.1b, reproduced in Fig. 3.4.2a below. The
element has dimensions is x  y and is subjected to uniform stresses over its sides.
The resultant forces of the stresses acting on each side of the element act through the side-
centres, and are shown in Fig. 3.4.2b. The stresses shown are positive, but note how

Section 3.4
positive stresses can lead to negative forces, depending on the definition of the x  y
axes used. The resultant force on the complete element is seen to be zero.
 yy Fy   yy x
 yx Fx   yx x
 xy Fy   xy y
 xx y Fx   xx y
x
 xy
 xx Fx   xx y
Fy   xy y
 yx Fx   yx x
 yy Fy   yy x
(a) (b)
Figure 3.4.2: stress components acting on a material element; (a) stresses, (b)
resultant forces on each side
By taking moments about any point in the block, one finds that {▲Problem 1}
 xy   yx (3.4.2)
Thus the shear stresses acting on the element are all equal, and for this reason the  yx
stresses are usually labelled  xy , Fig. 3.4.3a, or simply labelled  , Fig. 3.4.3b.
 xy 
 xy  xy  
 xy 
(a ) (b)
Figure 3.4.3: shear stress acting on a material element
3.4.2 Three Dimensional Stress

The three-dimensional counterpart to the two-dimensional element of Fig. 3.4.2 is shown
in Fig. 3.4.4. Again, all stresses shown are positive.

Section 3.4
 yy
 yx
y  yz  xy
 zy
 xz  xx
x
 zx
z
 zz
Figure 3.4.4: a three dimensional material element
Moment equilibrium in this case requires that
 xy   yx ,  xz   zx ,  yz   zy (3.4.3)
The nine stress components, six of which are independent, can now be written in the
matrix form
 xx  xy  xz 
 
ij

  yx  yy  yz 

(3.4.4)
 zx  zy  zz 

A vector F has one direction associated with it and is characterised by three components
( Fx , Fy , Fz ) . The stress is a quantity which has two directions associated with it (the
direction of a force and the normal to the plane on which the force acts) and is
characterised by the nine components of Eqn. 3.4.4. Such a mathematical object is called
a tensor. Just as the three components of a vector change with a change of coordinate
axes (for example, as in Fig. 2.2.1), so the nine components of the stress tensor change
with a change of axes. This is discussed in the next section for the two-dimensional case.
(The concept of a tensor will be examined more closely in Books II and especially IV.)
3.4.3 Stress Transformation Equations

Consider the case where the nine stress components acting on three perpendicular planes
through a material particle are known. These components are  xx ,  xy , etc. when using
x, y , z axes, and can be represented by the cube shown in Fig. 3.4.5a. Rotate now the
planes about the three axes – these new planes can be represented by the rotated cube
shown in Fig. 3.4.5b; the axes normal to the planes are now labelled x, y , z and the
corresponding stress components with respect to these new axes are  xx  ,  xy
 , etc.

Section 3.4
y
y
x
 yy x  yy
 yx
z
z  yx
 xy  xx
 yz  xy
 yz
 xz
 zy  zy
 xx  xz
 zx
 zx
 zz  zz
(a) (b)
Figure 3.4.5: a three dimensional material element; (a) original element, (b) rotated
element
There is a relationship between the stress components  xx ,  xy , etc. and the stress
components  xx ,  xy
 , etc. The relationship can be derived using Newton’s Laws. The
equations describing the relationship in the fully three-dimensional case are very lengthy
– they will be discussed in Books II and IV. Here, the relationship for the two-
dimensional case will be derived – this 2D relationship will prove very useful in
analysing many practical situations.
Two-dimensional Stress Transformation Equations
Assume that the stress components of Fig. 3.4.6a are known. It is required to find the
stresses arising on other planes through p. Consider the perpendicular planes shown in
Fig. 3.4.65b, obtained by rotating the original element through a positive
(counterclockwise) angle  . The new surfaces are defined by the axes x   y  .

Section 3.4
 yy  yy
  xx

 xx
 xx  xx
y  y

x
x  yy   yy
(a ) (b)
Figure 3.4.6: stress components acting on two different sets of perpendicular

surfaces, i.e. in two different coordinate systems; (a) original system, (b) rotated
system
To evaluate these new stress components, consider a triangular element of material at the
point, Fig. 3.4.7. Carrying out force equilibrium in the direction x  , one has (with unit
depth into the page)
F x :  xx
 AB   xx OB cos    yy OA sin    OB sin    OA cos   0 (3.4.5)
Since OB  AB cos , OA  AB sin  , and dividing through by AB ,
 xx   xx cos 2    yy sin 2    sin 2 (3.4.6)
y
y B

  xx
 xx
x

 x
O A

 yy
Figure 3.4.7: a free body diagram of a triangular element of material

Section 3.4
The forces can also be resolved in the y  direction and one obtains the relation
   ( yy   xx ) sin  cos    cos 2 (3.4.7)
Finally, consideration of the element in Fig. 3.4.8 yields two further relations, one of
which is the same as Eqn. 3.4.6.
y
y
 yy
  x
 xx
 
 x
o
 yy
Figure 3.4.8: a free body diagram of a triangular element of material
In summary, one obtains the stress transformation equations:
 xx  cos 2  xx  sin 2  yy  sin 2 xy

 yy  sin 2  xx  cos 2  yy  sin 2 xy 2D Stress Transformation Equations (3.4.8)
 xy  sin  cos  ( yy   xx )  cos 2 xy
These equations have many uses, as will be seen in the next section.
In matrix form,
 xx  xy   cos  sin    xx  xy  cos   sin  

   (3.4.9)
 yx  yy   sin  cos    yx  yy   sin  cos  
Body Force, Acceleration and Non-Uniform Stress
Here, it will be shown that the Stress Transformation Equations are valid also when (i)
there are body forces, (ii) the body is accelerating and (iii) the stress and other quantities
are not uniform.

Section 3.4
Suppose that a body force Fb   Fb  x i   Fb  y j acts on the material and that the material
is accelerating with an acceleration a  a x i  a y j . The components of body force and
acceleration are shown in Fig. 3.4.9 (a reproduction of Fig. 3.4.7). The body force will
vary depending on the size of the material under consideration, e.g. the force of
gravity Fb  mg will be larger for larger materials; therefore consider a quantity which is
independent of the amount of material: the body force per unit mass, Fb / m . Then, Eqn
3.4.5 now reads
F x :  xx
 AB   xx OB cos    yy OA sin    OB sin    OA cos 
(3.4.10)
  Fb / m  x m cos    Fb / m  m sin   ma x cos   ma y sin   0
where m is the mass of the triangular portion of material. The volume of the triangle is
2
AB / sin 2 so that, this time, when 3.4.10 is divided through by AB , one has
 xx   xx cos2    yy sin 2    sin 2

(3.4.11)
 AB   Fb / m  x / 2sin    Fb / m  / 2 cos   a x / 2sin   a y / 2cos  
where  is the density. Now, as the element is shrunk in size down to the vertex O,
AB  0 , and Eqn. 3.4.6 is recovered. Thus the Stress Transformation Equations are
valid provided the material under consideration is very small; in the limit, they are valid
“at the point” O.
y
y B
ay   xx
 xx ax
 Fb  y x

 Fb  x
O x
 A
 yy
Figure 3.4.9: a free body diagram of a triangular element of material, including a

body force and acceleration
Finally, consider the case where the stress is not uniform over the faces of the triangular
portion of material. Intuitively, it can be seen that, if one again shrinks the portion of
material down in size to the vertex O, the Stress Transformation Equations will again be

Section 3.4
valid, with the quantities  xx

 ,  xx ,  yy etc. being the values “at” the vertex. To be more
precise, consider the  xx stress acting over the face OB in Fig. 3.4.10. No matter how
the stress varies in the material, if the distance OB is small, the stress can be
approximated by a linear stress distribution, Fig. 3.4.10b. This linear distribution can
itself be decomposed into two components, a uniform stress of magnitude  xx o
(the value
of  xx at the vertex) and a triangular distribution with maximum value  xx . The
resultant force on the face is then OB  xxo   xx / 2  . This time, as the element is
shrunk in size,  xx  0 and Eqn. 3.4.6 is again recovered. The same argument can be
used to show that the Stress Transformation Equations are valid for any varying stress,
body force or acceleration.
 xx
B B
 xx
O  xxo O
Figure 3.4.10: stress varying over a face; (a) stress is linear over OB if OB is small,
(b) linear distribution of stress as a uniform stress and a triangular stress
Three Dimensions Re-visited
As the planes were rotated in the two-dimensional analysis, no consideration was given to
the stresses acting in the “third dimension”. Considering again a three dimensional block,
Fig. 3.4.11, there is only one traction vector acting on the x  y plane at the material
particle, t. This traction vector can be described in terms of the x, y, z axes as
t   zx i   zy j   zz k , Fig 3.4.11a. Alternatively, it can be described in terms of the
x, y, z axes as t   zx i   zy j   zz k , Fig 3.4.11b.

Section 3.4
y
y
x

x
z  z
 zy t  zy t
 zx
 zx
 zz
 zz
(a) (b)
Figure 3.4.11: a three dimensional material element; (a) original element, (b) rotated
element (rotation about the z axis)
With the rotation only happening in the x  y plane, about the z axis, one has
 zz   zz , k  k . One can thus examine the two dimensional x  y plane shown in Fig.
3.4.12, with
 zx i   zy j   zx i   zy j . (3.4.12)
Using some trigonometry, one can see that
 zx   zx cos    zy sin 

. (3.4.13)
 zy   zx sin    zy cos 
y
y
 zx
t  zy
 x  zy
j i

o x
 zx
Figure 3.4.12: the traction vector represented using two different coordinate systems

Section 3.4
3.4.4 Problems
1. Derive Eqns. 3.4.2 by taking moments about the lower left corner of the block in Fig.
3.4.2.
2. Suppose that the stresses acting on two perpendicular planes through a point are
 xx  xy   2  1
 
 ij    
 yx  yy   1 1 
Use the stress transformation formulae to evaluate the stresses acting on two new
perpendicular planes through the point, obtained from the first set by a positive
rotation of 30o . Use the conventional notation x   y  to represent the coordinate
axes parallel to these new planes.

Section 3.5
3.5 Plane Stress

This section is concerned with a special two-dimensional state of stress called plane stress.
It is important for two reasons: (1) it arises in real components (particularly in thin
components loaded in certain ways), and (2) it is a two dimensional state of stress, and thus
serves as an excellent introduction to more complicated three dimensional stress states.
3.5.1 Plane Stress

The state of plane stress is defined as follows:
Plane Stress:
If the stress state at a material particle is such that the only non-zero stress components act in
one plane only, the particle is said to be in plane stress.
The axes are usually chosen such that the x  y plane is the plane in which the stresses act,
Fig. 3.5.1.
 yy y
 yy
 xy  xy
 xx
Figure 3.5.1: non-zero stress components acting in the x – y plane
The stress can be expressed in the matrix form 3.4.1.
Example
The thick block of uniform material shown in Fig. 3.5.2, loaded by a constant stress  o in
the x direction, will have  xx   0 and all other components zero everywhere. It is
therefore in a state of plane stress.

Section 3.5
o
Figure 3.5.2: a thick block of material in plane stress
3.5.2 Analysis of Plane Stress

Next are discussed the stress invariants, principal stresses and maximum shear stresses
for the two-dimensional plane state of stress, and tools for evaluating them. These quantities
are useful because they tell us the complete state of stress at a point in simple terms. Further,
these quantities are directly related to the strength and response of materials. For example,
the way in which a material plastically (permanently) deforms is often related to the
maximum shear stress, the directions in which flaws/cracks grow in materials is often related
to the principal stresses, and the energy stored in materials is often a function of the stress
invariants.
Stress Invariants
A stress invariant is some function of the stress components which is independent of the
coordinate system being used; in other words, they have the same value no matter where the
x  y axes are drawn through a point. In a two dimensional space there are two stress
invariants, labelled I 1 and I 2 . These are
I 1   xx   yy
Stress Invariants (3.5.1)
I 2   xx yy   xy2
These quantities can be proved to be invariant directly from the stress transformation
equations, Eqns. 3.4.8 {▲Problem 1}. Physically, invariance of I 1 and I 2 means that they
are the same for any chosen perpendicular planes through a material particle.
Combinations of the stress invariants are also invariant, for example the important quantity
 xx   yy
1
2
I1  1
4
I 12  I 2   1
4
 xx   yy    xy2
2
(3.5.2)
2

Section 3.5
Principal Stresses
Consider a material particle for which the stress, with respect to some x  y coordinate
system, is
 xx  xy   2  1
  (3.5.3)
 yx  yy   1 1 
The stress acting on different planes through the point can be evaluated using the Stress
Transformation Equations, Eqns. 3.4.8, and the results are plotted in Fig. 3.5.3. The original
planes are re-visited after rotating 180o .
Stress at Point
 yy  xx
3
1
Stress
0
0 30 60 90 120 150 180
-1  xy
-2
Degrees
Figure 3.5.3: stresses on different planes through a point
It can be seen that there are two perpendicular planes for which the shear stress is zero, for 
 58o and    58  90  . In fact it can be proved that for every point in a material there are
o
two (and only two) perpendicular planes on which the shear stress is zero (see below). These
planes are called the principal planes. It will also be noted from the figure that the normal
stresses acting on the planes of zero shear stress are either a maximum or minimum. Again,
this can be proved (see below). These normal stresses are called principal stresses. The
principal stresses are labelled  1 and  2 , Fig. 3.5.4.
2
1
zero shear stress
on these planes
Principal (principal planes)
stresses 1
2
Figure 3.5.4: principal stresses

Section 3.5
The principal stresses can be obtained by setting  xy  0 in the Stress Transformation
Equations, Eqns. 3.4.8, which leads to the value of  for which the planes have zero shear
stress:
2 xy
tan 2  Location of Principal Planes (3.5.4)
 xx   yy
For the example stress state, Eqn. 3.5.3, this leads to
1
  arctan 2 
2
and so the perpendicular planes are at   31.72 148.28 and   58.3 .
Explicit expressions for the principal stresses can be obtained by substituting the value of 
from Eqn. 3.5.4 into the Stress Transformation Equations, leading to (see the Appendix to
this section, §3.5.8)
 1  12 ( xx   yy )  1
4 ( xx   yy ) 2   xy2
Principal Stresses (3.5.5)
 2  12 ( xx   yy )  1
4 ( xx   yy ) 2   xy2
For the example stress state Eqn.3.5.3, one has
3 5 3 5
1   2.62, 2   0.38
2 2
Note here that one uses the symbol  1 to represent the maximum principal stress and  2 to
represent the minimum principal stress. By maximum, it is meant the algebraically largest
stress so that, for example,  1  3 .
From Eqns. 3.5.2, 3.5.5, the principal stresses are invariant; they are intrinsic features of the
stress state at a point and do not depend on the coordinate system used to describe the stress
state.
The question now arises: why are the principal stresses so important? One part of the answer
is that the maximum principal stress is the largest normal stress acting on any plane through a
material particle. This can be proved by differentiating the stress transformation formulae
with respect to ,

Section 3.5
d xx
  sin 2 ( xx   yy )  2 cos 2 xy
d
d yy
  sin 2 ( xx   yy )  2 cos 2 xy (3.5.6)
d
d xy
  cos 2 ( xx   yy )  2 sin 2 xy
d
The maximum/minimum values can now be obtained by setting these expressions to zero.
One finds that the normal stresses are a maximum/minimum at the very value of  in Eqn.
3.5.4 – the value of  for which the shear stresses are zero – the principal planes.
Very often the only thing one knows about the stress state at a point are the principal stresses.
In that case one can derive a very useful formula as follows: align the coordinate axes in the
principal directions, so
 xx   1 ,  yy   2 ,  xy  0 (3.5.7)
Using the transformation formulae with the relations sin 2   12 (1  cos 2 ) and
cos 2   12 (1  cos 2 ) then leads to
1 1
 xx  ( 1   2 )  ( 1   2 ) cos 2
2 2
1 1
 yy  ( 1   2 )  ( 1   2 ) cos 2 (3.5.8)
2 2
1
 xy   ( 1   2 ) sin 2

2
Here,  is measured from the principal directions, as illustrated in Fig. 3.5.5.
principal directions

2
1
1 2
Figure 3.5.5: principal stresses and principal directions

Section 3.5
The Third Principal Stress
Although plane stress is essentially a two-dimensional stress-state, it is important to keep in

mind that any real material is three-dimensional. The stresses acting on the x  y plane are
the normal stress  zz and the shear stresses  zx and  zy , Fig. 3.5.6. These are all zero (in
plane stress). It was discussed above how the principal stresses occur on planes of zero shear
stress. Thus the  zz stress is also a principal stress. Technically speaking, there are always
three principal stresses in three dimensions, and (at least) one of these will be zero in plane
stress. This fact will be used below in the context of maximum shear stress.
 zz  0
 zy  0
 zx  0
 yy y
 yy
 xy  xy
 xx
Figure 3.5.6: stresses acting on the x – y plane
Maximum Shear Stress
Eqns. 3.5.8 can be used to derive an expression for the maximum shear stress.
Differentiating the expression for shear stress with respect to  , setting to zero and solving,
shows that the maximum/minimum occurs at   45 , in which case
1 1
 xy   45
  1   2 ,  xy   45
  1   2 
2 2
or
max xy  
1
1   2 Maximum Shear Stress (3.5.9)
2
Thus the shear stress reaches a maximum on planes which are oriented at  45 to the
principal planes, and the value of the shear stress acting on these planes is as given above.
Note that the formula Eqn. 3.5.9 does not let one know in which direction the shear stresses
are acting but this is not usually an important issue. Many materials respond in certain ways
when the maximum shear stress reaches a critical value, and the actual direction of shear

Section 3.5
stress is unimportant. The direction of the maximum principal stress is, on the other hand,
important – a material will in general respond differently according to whether the normal
stress is compressive or tensile.
The normal stress acting on the planes of maximum shear stress can be obtained by
substituting   45 back into the formulae for normal stress in Eqn. 3.5.8, and one sees that
 xx   yy  ( 1   2 ) / 2 (3.5.10)
The results of this section are summarised in Fig. 3.5.7.
2  1   2  / 2
max( xy )
 1  2 / 2
 1   2  / 2
1
  45o
1
 1   2  / 2
 1   2  / 2 2
Figure 3.5.7: principal stresses and maximum shear stresses acting on the x – y plane
The maximum shear stress in the x  y plane was calculated above, Eqn. 3.5.9. This is not
necessarily the maximum shear stress acting at the material particle. In general, it can be
shown that the maximum shear stress is the maximum of the following three terms (see the
Appendix to this section, §3.5.8):
1 1 1
1   2 , 1   3 , 2  3
2 2 2
The first term is the maximum shear stress in the 1  2 plane, i.e. the plane containing the  1
and  2 stresses (and given by Eqn. 3.5.9). The second term is the maximum shear stress in
the 1  3 plane and the third term is the maximum shear stress in the 2  3 plane. These are
sketched in Fig. 3.5.8 below.

Section 3.5
max   12  1   2 max   12  1   3 max   12  2   3

2 3 3
1 1 2
(a) 1-2 plane (a) 1-3 plane (a) 2-3 plane
Figure 3.5.8: principal stresses and maximum shear stresses
In the case of plane stress,  3   zz  0 , and the maximum shear stress will be
1 1 1 
max   1   2 , 1 , 2  (3.5.11)
2 2 2 
3.5.3 Stress Boundary Conditions

When solving problems, information is usually available on what is happening at the
boundaries of materials. This information is called the boundary conditions. Information is
usually not available on what is happening in the interior of the material – information there
is obtained by solving the equations of mechanics.
A number of different conditions can be known at a boundary, for example it might be

known that a certain part of the boundary is fixed so that the displacements there are zero.
This is known as a displacement boundary condition. On the other hand the stresses over a
certain part of the material boundary might be known. These are known as stress boundary
conditions – this case will be examined here.
General Stress Boundary Conditions
It has been seen already that, when one material contacts a second material, a force, or
distribution of stress arises. This force F will have arbitrary direction, Fig. 3.5.9a, and can be
decomposed into the sum of a normal stress distribution  N and a shear distribution  S , Fig.
3.5.9b. One can introduce a coordinate system to describe the applied stresses, for example
the x  y axes shown in Fig. 3.5.9c (the axes are most conveniently defined to be normal and
tangential to the boundary).

Section 3.5
S  xy
F
N
y
 yy
contact
x
region
(a) (b) (c)
Figure 3.5.9: Stress boundary conditions; (a) force acting on material due to contact
with a second material, (b) the resulting normal and shear stress distributions, (c)
applied stresses as stress components in a given coordinate system
Figure 3.5.10 shows the same component as Fig. 3.5.9. Shown in detail is a small material
element at the boundary. From equilibrium of the element, stresses  xy ,  yy , equal to the
applied stresses, must be acting inside the material, Fig. 3.5.10a. Note that the tangential
stresses, which are the  xx stresses in this example, can take on any value and the element
will still be in equilibrium with the applied stresses, Fig. 3.5.10b.
 xx
 xy
 yy
 yy
 xy
 xx
(a) (b)
Figure 3.5.10: Stresses acting on a material element at the boundary, (a) normal and
shear stresses, (b) tangential stresses
Thus, if the applied stresses are known, then so also are the normal and shear stresses acting
at the boundary of the material.
Stress Boundary Conditions at a Free Surface
A free surface is a surface that has “nothing” on one side and so there is nothing to provide
reaction forces. Thus there must also be no normal or shear stress on the other side (the
inside).
This leads to the following, Fig. 3.5.11:

Section 3.5
Stress boundary conditions at a free surface:

the normal and shear stress at a free surface are zero
This simple fact is used again and again to solve practical problems.
Again, the stresses acting normal to any other plane at the surface do not have to be zero –
they can be balanced as, for example, the tangential stresses  T and the stress  in Fig.
3.5.11.
T
T
T
S
 N
 N S  0
Figure 3.5.11: A free surface - the normal and shear stresses there are zero
Atmospheric Pressure
There is something acting on the outside “free” surfaces of materials – the atmospheric
pressure. This is a type of stress which is hydrostatic, that is, it acts normal at all points, as
shown in Fig. 3.5.12. Also, it does not vary much. This pressure is present when one
characterises a material, that is, when its material properties are determined from tests and so
on, for example, its Young’s Modulus (see Chapter 5). The atmospheric pressure is therefore
a datum – stresses are really measured relative to this value, and so the atmospheric pressure
is ignored.
Figure 3.5.12: a material subjected to atmospheric pressure

Section 3.5
3.5.4 Thin Components

Consider a thin component as shown in Fig. 3.5.13. With the coordinate axes aligned as
shown, and with the large face free of loading, one has  zx   zy   zz  0 . Strictly
speaking, these stresses are zero only at the free surfaces of the material but, because it is
thin, these stresses should not vary much from zero within. Taking the “z” stresses to be
identically zero throughout the material, the component is in a state of plane stress1. On the
other hand, were the sheet not so thin, the stress components that were zero at the free-
surfaces might well deviate significantly from zero deep within the material.
x
 yy
 xy

 xx
Figure 3.5.13: a thin material loaded in-plane, leading to a state of plane stress
When analysing plane stress states, only one cross section of the material need be considered.
This is illustrated in Fig. 3.5.14.
1
it will be shown in Book II that, when the applied stresses  xx ,  yy ,  xy vary only linearly over the thickness
of the component, the stresses  zz ,  zx ,  zy are exactly zero throughout the component, otherwise they are only
approximately zero

Section 3.5
y y
x x
z
3 – dimensional 2 – dimensional cross-section of

material material
Figure 3.5.14: one two-dimensional cross-section of material
Note that, although the stress normal to the plane,  zz , is zero, the three dimensional sheet of
material is deforming in this direction – it will obviously be getting thinner under the tensile
loading shown in Fig. 3.5.14.
Note that plane stress arises in all thin materials (loaded in –plane), no matter what they are
made of.
3.5.5 Mohr’s Circle

Otto Mohr devised a way of describing the state of stress at a point using a single diagram,
called the Mohr's circle.
To construct the Mohr circle, first introduce the stress coordinates  ,  , Fig. 3.5.15; the
abscissae (horizontal) are the normal stresses  and the ordinates (vertical) are the shear
stresses  . On the horizontal axis, locate the principal stresses  1 ,  2 , with  1   2 . Next,
draw a circle, centred at the average principal stress  ,    1   2  / 2,0  , having radius
 1   2  / 2 .
The normal and shear stresses acting on a single plane are represented by a single point on
the Mohr circle. The normal and shear stresses acting on two perpendicular planes are
represented by two points, one at each end of a diameter on the Mohr circle. Two such
diameters are shown in the figure. The first is horizontal. Here, the stresses acting on two
perpendicular planes are  ,    1 ,0 and  ,    2 ,0 and so this diameter represents
the principal planes/stresses.

Section 3.5
 yy
 xy
 B  xx
1 1 
 xx  ( 1   2 )  ( 1   2 ) cos 2 A
2 2
A
1 1
( xx   yy ) 2  ( xy ) 2  ( 1   2 )
4 2
2
 1

2
1
 xy   ( 1   2 ) sin 2
2  2
B
1 1
 yy  ( 1   2 )  ( 1   2 ) cos 2 1
2 2
1 1
( xx   yy )  ( 1   2 )
2 2
Figure 3.5.15: Mohr’s Circle
The stresses on planes rotated by an amount  from the principal planes are given by Eqn.
3.5.8. Using elementary trigonometry, these stresses are represented by the points A and B
in Fig. 3.5.15. Note that a rotation of  in the physical plane corresponds to a rotation of 2
in the Mohr diagram.
Note also that the conventional labeling of shear stress has to be altered when using the Mohr
diagram. On the Mohr circle, a shear stress is positive if it yields a clockwise moment about
the centre of the element, and is "negative" when it yields a negative moment. For example,
at point A the shear stress is "positive" (   0 ), which means the direction of shear on face A
of the element is actually opposite to that shown. This agrees with the formula
1
 xy   ( 1   2 ) sin 2 , which is less than zero for  1   2 and   90 o . At point B the
2
shear stress is "negative" (   0 ), which again agrees with formula.

Section 3.5
3.5.6 Stress Boundary Conditions (continued)

Consider now in more detail a surface between two different materials, Fig. 3.5.16. One says
that the normal and shear stresses are continuous across the surface, as illustrated.
y
 yy( 2 )
2 x
 yy(1)   yy( 2)
 xy( 2)  xy(1)   xy( 2)
1
 xy(1)
 yy(1)
Figure 3.5.16: normal and shear stress continuous across an interface between two
different materials, material ‘1’ and material ‘2’
Note also that, since the shear stress  xy is the same on both sides of the surface, the shear
stresses acting on both sides of a perpendicular plane passing through the interface between
the materials, by the symmetry of stress, must also be the same, Fig. 3.5.17a.
 xy
 xx( 2)
 xx(1)
 xy
(a ) ( b)
Figure 3.5.17: stresses at an interface; (a) shear stresses continuous across the
interface, (b) tangential stresses not necessarily continuous
However, again, the tangential stresses, those acting parallel to the interface, do not have to
be equal. For example, shown in Fig. 3.5.17b are the tangential stresses acting in the upper
material,  xx( 2 ) - they balance no matter what the magnitude of the stresses  xx(1) .
Description of Boundary Conditions
The following example brings together the notions of stress boundary conditions, stress
components, equilibrium and equivalent forces.

Section 3.5
Example
Consider the plate shown in Fig. 3.5.18. It is of width 2a , height b and depth t. It is
subjected to a tensile stress r, pressure p and shear stresses s. The applied stresses are
uniform through the thickness of the plate. It is welded to a rigid base.
y
r
s
s
p
b
2a
Figure 3.5.18: a plate subjected to stress distributions
Using the x  y axes shown, the stress boundary conditions can be expressed as:
 xx ( a, y )   p
Left-hand surface:  , 0 yb
  xy (a, y )   s
 yy ( x, b)   r
Top surface:  ,  a  x  a
 xy ( x, b)   s
  xx ( a, y )  0
Right-hand surface:  , 0 yb
 xy ( a, y )   s
Note carefully the description of the normal and shear stresses over each side and the signs of
the stress components.
The stresses at the lower edge are unknown (there is a displacement boundary condition
there: zero displacement). They will in general not be uniform. Using the given x  y axes,
these unknown reaction stresses, exerted by the base on the plate, are (see Fig 3.5.19)

Section 3.5
 yy ( x,0)
Lower surface:  ,  a  x  a
 xy ( x,0)
Note the directions of the arrows in Fig. 3.5.19, they have been drawn in the direction of
positive  yy ( x,0),  xy ( x,0) .
 xy (x,0)
 yy (x,0)
Figure 3.5.19: unknown reaction stresses acting on the lower edge
For force equilibrium of the complete plate, consider the free-body diagram 3.5.20; shown
are the resultant forces of the stress distributions. Force equilibrium requires that
a
 Fx  bpt  2ast  t   xy ( x,0)dx  0

a
a
F y  2art  t   yy ( x,0)dx  0
a
2art
2ast
bpt
bst
bst
Figure 3.5.20: a free-body diagram of the plate in Fig. 3.5.18 showing the known
resultant forces (forces on the lower boundary are not shown)
For moment equilibrium, consider the moments about, for example, the lower left-hand
corner. One has

Section 3.5
a
 M 0  bpt (b / 2)  2ast (b)  2art (a)  bst (2a)  t   yy ( x,0)  a  x dx  0

a
If one had taken moments about the top-left corner, the equation would read
M 0  bpt (b / 2)  2art (a)  bst (2a )

a a
 t   xy ( x,0)  b dx  t   yy ( x,0)  a  x dx  0
a a
■
3.5.7 Problems
1. Prove that the function  x   y , i.e. the sum of the normal stresses acting at a point, is a
stress invariant. [Hint: add together the first two of Eqns. 3.4.8.]
2. Consider a material in plane stress conditions. An element at a free surface of this

material is shown below left. Taking the coordinate axes to be orthogonal to the surface
as shown (so that the tangential stress is  xx ), one has
 xx  xy   xx 0
  
 yx  yy   0 0
(a) what are the two in-plane principal stresses at the point? Which is the maximum and
which is the minimum?
(b) examine planes inclined at 45o to the free surface, as shown below right. What are the
stresses acting on these planes and what have they got to do with maximum shear
stress?
y y
y x
45
o
 yy
 xy
 xx  xx  xy  xx
3. The stresses at a point in a state of plane stress are given by

 xx  xy   1 3
 
 yx  yy   3 2
(a) Draw a little box to represent the point and draw some arrows to indicate the
magnitude and direction of the stresses acting at the point.
(b) What relationship exists between Oxy and a second coordinate set Ox y  , such that
the shear stresses are zero in Ox y  ?
(c) Find the two in-plane principal stresses.

Section 3.5
(d) Draw another box whose sides are aligned to the principal directions and draw some
arrows to indicate the magnitude and direction of the principal stresses acting at the
point.
(e) Check that the sum of the normal stresses at the point is an invariant.
4. A material particle is subjected to a state of stress given by

  0
 
 ij    0
 0 0 0
Find the principal stresses (all three), maximum shear stresses (see Eqn. 3.5.11), and the
direction of the planes on which these stresses act.
5. Consider the following state of stress (with respect to an x, y , z coordinate system):
0  0
 0 0
 
 0 0 0
(a) Use the stress transformation equations to derive the stresses acting on planes
obtained from the original planes by a counterclockwise rotation of 45o about z axis.
(b) What is the maximum normal stress acting at the point?
(c) What is the maximum shear stress? On what plane(s) does it act? (See Eqn. 3.5.11.)
6. Consider the two dimensional stress state

   0
ij
0
 
Show that this is an isotropic state of stress, that is, the stress components are the same
on all planes through a material particle.
7. (a) Is a trampoline (the material you jump on) in a state of plane stress? When someone
is actually jumping on it?
(b) Is a picture hanging on a wall in a state of plane stress?
(c) Is a glass window in a state of plane stress? On a very windy day?
(d) A piece of rabbit skin is stretched in a testing machine – is it in a state of plane stress?
8. Consider the point shown below, at the boundary between a wall and a dissimilar
material. Label the stress components displayed using the coordinate system shown.
Which stress components are continuous across the wall/material boundary? (Add a
superscript ‘w’ for the stresses in the wall.)

Section 3.5
9. A thin metal plate of width 2b , height h and depth t is loaded by a pressure distribution
p( x) along  a  x   a and welded at its base to the ground, as shown in the figure
below. Write down expressions for the stress boundary conditions (two on each of the
three edges). Write down expressions for the force equilibrium of the plate and moment
equilibrium of the plate about the corner A.
y
p(x)
2a
A x
2b
3.5.8 Appendix to §3.5

A Note on the Formulae for Principal Stresses
To derive Eqns. 3.5.5, first rewrite the transformation equations in terms of 2 using
sin 2   12 (1  cos 2 ) and cos 2   12 (1  cos 2 ) to get
 xx  12 (1  cos 2 ) xx  12 (1  cos 2 ) yy  sin 2 xy

 yy  12 (1  cos 2 ) xx  12 (1  cos 2 ) yy  sin 2 xy
 xy  12 sin 2 ( yy   xx )  cos 2 xy
Next, from Eqn. 3.5.4,

Section 3.5
2 xy  xx   yy
sin 2  , cos 2 
( xx   yy ) 2  4 xy2 ( xx   yy ) 2  4 xy2
Substituting into the rewritten transformation formulae then leads to
 xx  12 ( xx   yy )  1
4 ( xx   yy ) 2   xy2
 yy  12 ( xx   yy )  1
4 ( xx   yy ) 2   xy2
 xy  0
Here  xx   yy so that the maximum principal stress is  1   xx and the minimum principal
stress is  2   yy . Here it is implicitly assumed that tan 2  0 , i.e. that 0  2  90 or
180  2  270 . On the other hand one could assume that tan 2  0 , i.e. that
90  2  180 or 270  2  360 , in which case one arrives at the formulae
 xx  12 ( xx   yy )  1
4 ( xx   yy ) 2   xy2
 yy  12 ( xx   yy )  1
4 ( xx   yy ) 2   xy2
The results can be summarised as Eqn. 3.5.5,
 1  12 ( xx   yy )  1
4 ( xx   yy ) 2   xy2
 2  12 ( xx   yy )  1
4 ( xx   yy ) 2   xy2
These formulae do not tell one on which of the two principal planes the maximum principal
stress acts. This might not be an important issue, but if this information is required one needs
to go directly to the stress transformation equations. In the example stress state, Eqn. 3.5.3,
one has
 xx  cos 2  (2)  sin 2  (1)  sin 2 (1)

 yy  sin 2  (2)  cos 2  (1)  sin 2 (1)
For   31.72 148.28 ,  xx  2.62 and  yy  0.38 . So one has the situation shown
below.

Section 3.5
y
 yy
 xx
  31.72 o x
If one takes the other angle,   58.3 , one has  xx  0.38 and  yy  2.62 , and the situation
below
x
 xx
y
  58.3o
 yy
A Note on the Maximum Shear Stress
Shown below left is a box element with sides perpendicular to the 1, 2, z axes, i.e. aligned
with the principal directions. The stresses in the new x, y  axis system shown are given by
Eqns. 3.5.8, with  measured from the principal directions:
1 1
 xx  ( 1   2 )  ( 1   2 ) cos 2
2 2
1 1
 yy  ( 1   2 )  ( 1   2 ) cos 2

2 2
1
 xy   ( 1   2 ) sin 2
2
Now as well as rotating around in the 1  2 plane through an angle  , rotate also in the x, z
plane through an angle  (see below right). This rotation leads to the new stresses

Section 3.5
ˆ xx  cos2  xx  sin 2  zz  sin 2 xz

ˆ zz  sin 2  xx  cos2  zz  sin 2 xz
ˆ xz  sin  cos  ( zz   xx )  cos 2 xz
In plane stress,  zz   xz  0 , so one has the stresses
ˆ xx  cos2  xx , ˆ yy  sin 2  xx , ˆ xy   12 sin 2 xx
z z
y
2 ẑ
x̂
x 
 x
1
The shear stress can be written out in full:
ˆ xy  ,     12 sin 2  12 ( 1   2 )  12 ( 1   2 ) cos 2  .
This is a function of two variables; its minimum value can be found by setting the partial
derivatives with respect to these variables to zero. Differentiating,
ˆ xy /    cos 2  12 ( 1   2 )  12 ( 1   2 ) cos 2 

ˆ xy /    12 sin 2  ( 1   2 ) sin 2 
Setting to zero gives the solutions sin 2  0 , cos 2  0 , i.e.   0 ,   45o . Thus the
maximum shear stress occurs at 45o to the 1  2 plane, and in the 1  z , i.e. 1  3 plane (as in
Fig. 3.5.8b). The value of the maximum shear stress here is then ˆ xy  12  1 , which is the
expression in Eqn. 3.5.11.

4 Strain
The concept of strain is introduced in this Chapter. The approximation to the true strain
of the engineering strain is discussed. The practical case of two dimensional plane
strain is discussed, along with the strain transformation formulae, principal strains,
principal strain directions and the maximum shear strain.
85
86
Section 4.1
4.1 Strain
If an object is placed on a table and then the table is moved, each material particle moves
in space. The particles undergo a displacement. The particles have moved in space as a
rigid body. The material remains unstressed. On the other hand, when a material is acted
upon by a set of forces, it changes size and/or shape, it deforms. This deformation is
described using the concept of strain. The study of motion, without reference to the
forces which cause such motion, is called kinematics.
4.1.1 One Dimensional Strain
The Engineering Strain
Consider a slender rod, fixed at one end and stretched, as illustrated in Fig. 4.1.1; the
original position of the rod is shown dotted.
A B A B
fixed   
Figure 4.1.1: the strain at a point A in a stretched slender rod; AB is a line element in
the unstretched rod, AB is the same line element in the stretched rod
There are a number of different ways in which this stretching/deformation can be

described. Here, what is perhaps the simplest measure, the engineering strain, will be
used. To determine the strain at point A, Fig. 4.1.1, consider a small line element AB
emanating from A in the unstretched rod. The points A and B move to A and B when
the rod has been stretched. The (engineering) strain  at A is then1
AB  AB
 ( A)  (4.1.1)
AB
The strain at other points in the rod can be evaluated in the same way.
If a line element is stretched to twice its original length, the strain is 1. If it is unstretched,
the strain is 0. If it is shortened to half its original length, the strain is 0.5 . The strain is
often expressed as a percentage; a 100% strain is a strain of 1, a 200% strain is a strain of
2, etc. Most engineering materials, such as metals and concrete, undergo very small
strains in practical applications, in the range 106 to 102 ; rubbery materials can easily
undergo large strains of 100%.
Consider now two adjacent line elements AE and EB (not necessarily of equal length),
which move to AE  and E B , Fig. 4.1.2. If the rod is stretching uniformly, that is, if all
1
this is the strain at point A. The strain at B is evidently the same – one can consider the line element AB to
emanate fom point B (it does not matter whether the line element emanates out from the point to the “left”
or to the “right”)

Section 4.1
line elements are stretching in the same proportion along the length of the rod, then
AE  / AE  E B / EB , and  ( A)   ( E ) ; the strain is the same at all points along the rod.
A B A B
fixed     
E E
Figure 4.1.2: the strain at a point A and the strain at point E in a stretched rod
In this case, one could equally choose the line element AB or the element AE in the
calculation of the strain at A, since
AB  AB AE   AE
 ( A)  
AB AE
In other words it does not matter what the length of the line element chosen for the
calculation of the strain at A is. In fact, if the length of the rod before stretching is L0 and
after stretching it is L , Fig. 4.1.3, the strain everywhere is (this is equivalent to choosing
a “line element” extending the full length of the rod)
L  L0
 (4.1.2)
L0
fixed
L0
L
Figure 4.1.3: a stretched slender rod
On the other hand, when the strain is not uniform, for example AE  / AE  E B / EB ,
then the length of the line element does matter. In this case, to be precise, the line
element AB in the definition of strain in Eqn. 4.1.1 should be “infinitely small”; the
smaller the line element, the more accurate will be the evaluation of the strain. The
strains considered in this book will be mainly uniform; non-uniform strain will be dealt
with in detail in Book II.
Displacement, Strain and Rigid Body Motions
To highlight the difference between displacement and strain, and their relationship,
consider again the stretched rod of Fig 4.1.1. Fig 4.1.4 shows the same rod: the two
points A and B undergo displacements u ( A)  AA , u ( B )  BB . The strain at A, Eqn
4.1.1, can be re-expressed in terms of these displacements:

Section 4.1
u ( B )  u ( A)
 ( A)  (4.1.3)
AB
In words, the strain is a measure of the change in displacement as one moves along the
rod.
A B A B
fixed   
u ( A)
u( B )
Figure 4.1.4: displacements in a stretched rod
Consider a line element emanating from the left-hand fixed end of the rod. The
displacement at the fixed end is zero. However, the strain at the fixed end is not zero,
since the line element there will change in length. This is a case where the displacement
is zero but the strain is not zero.
Consider next the case where the rod is not fixed and simply moves/translates in space,
without any stretching, Fig. 4.1.5. This is a case where the displacements are all non-zero
(and in this case everywhere the same) but the strain is everywhere zero. This is in fact a
feature of a good measure of strain: it should be zero for any rigid body motion; the strain
should only measure the deformation.
original position new position
Figure 4.1.5: a rigid body translation of a rod
Note that if one knows the strain at all points in the rod, one cannot be sure of the rod’s
exact position in space – again, this is because strain does not include information about
possible rigid body motion. To know the precise position of the rod, one must also have
some information about the displacements.
The True Strain
As mentioned, there are many ways in which deformation can be measured. Many
different strains measures are in use apart from the engineering strain, for example the
Green-Lagrange strain and the Euler-Almansi strain: referring again to Fig. 4.1.1, these
are
2 2 2 2
AB  AB AB  AB
Green-Lagrange  ( A)
 2
, Euler-Alamnsi  ( A)
 2
(4.1.4)
2 AB 2 AB

Section 4.1
Many of these strain measures are used in more advanced theories of material behaviour,
particularly when the deformations are very large. Apart from the engineering strain, just
one other measure will be discussed in any detail here: the true strain (or logarithmic
strain), since it is often used in describing material testing (see Chapter 5).
The true strain may be defined as follows: define a small increment in strain to be the
change in length divided by the current length: d  t  dL / L . As the rod of Fig. 4.11
stretches (uniformly), this current length continually changes, and the total strain thus
defined is the accumulation of these increments:
L
dL  L
t  
L0
L
 ln   .
 L0 
(4.1.5)
If a line element is stretched to twice its original length, the (true) strain is 0.69. If it is
unstretched, the strain is 0. If it is shortened to half its original length, the strain is 0.69 .
The fact that a stretching and a contraction of the material by the same factor results in
strains which differ only in sign is one of the reasons for the usefulness of the true strain
measure.
Another reason for its usefulness is the fact that the true strain is additive. For example, if
a line element stretches in two steps from lengths L1 to L2 to L3 , the total true strain is
 L3   L2   L3 
 t  ln    ln    ln   ,
 L2   L1   L1 
which is the same as if the stretching had occurred in one step. This is not true of the
engineering strain.
The true strain and engineering strain are related through (see Eqn. 4.1.2, 4.1.5)
 t  ln1    (4.1.6)
One important consequence of this relationship is that the smaller the deformation, the
less the difference between the two strains. This can be seen in Table 4.1 below, which
shows the values of the engineering and true strains for a line element of initial length
1mm, at different stretched lengths. (In fact, using a Taylor series expansion,
 t  ln 1       12  2  13  3   , for small  .) Almost all strain measures in use are
similar in this way: they are defined such that they are more or less equal when the
deformation is small. Put another way, when the deformations are small, it does not
really matter which strain measure is used, since they are all essentially the same – in that
case it is sensible to use the simplest measure.

Section 4.1
L0 (mm) L (mm)  t
1 2 1 0.693
1 1.5 0.5 0.405
1 1.4 0.4 0.336
1 1.3 0.3 0.262
1 1.2 0.2 0.182
1 1.1 0.1 0.095
1 1.01 0.01 0.00995
1 1.001 0.001 0.000995
Table 4.1: true strain and engineering strain at different stretches
It should be emphasised that one strain measure, e.g. engineering or true, is not more
“correct” or better than the other; the usefulness of a strain measure will depend on the
application.
4.1.2 Two Dimensional Strain

The two dimensional case is similar to the one dimensional case, in that material
deformation can be described by imagining the material to be a collection of small line
elements. As the material is deformed, the line elements stretch, or get shorter, only now
they can also rotate in space relative to each other. This movement of line elements is
encompassed in the idea of strain: the “strain at a point” is all the stretching, contracting
and rotating of all line elements emanating from that point, with all the line elements
together making up the continuous material, as illustrated in Fig. 4.1.6.
before deformation after deformation
Figure 4.1.6: a deforming material element; original state of line elements and their
final position after straining
It turns out that the strain at a point is completely characterised by the movement of any
two mutually perpendicular line-segments. If it is known how these perpendicular line-
segments are stretching, contracting and rotating, it will be possible to determine how any
other line element at the point is behaving, by using a strain transformation rule (see
later). This is analogous to the way the stress at a point is characterised by the stress
acting on perpendicular planes through a point, and the stress components on other planes
can be obtained using the stress transformation formulae.

Section 4.1
So, for the two-dimensional case, consider two perpendicular line-elements emanating
from a point. When the material that contains the point is deformed, two things (can)
happen:
(1) the line segments will change length and
(2) the angle between the line-segments changes.
The change in length of line-elements is called normal strain and the change in angle
between initially perpendicular line-segments is called shear strain.
As mentioned earlier, a number of different definitions of strain are in use; here, the
following, most commonly used, definition will be employed, which will be called the
exact strain:
Normal strain in direction x : (denoted by  xx )

change in length (per unit length) of a line element originally lying in the x direction
Normal strain in direction y : (denoted by  yy )

change in length (per unit length) of a line element originally lying in the y direction
Shear strain: (denoted by  xy )

(half) the change in the original right angle between the two perpendicular line
elements
Referring to Fig. 4.1.7, the (exact) strains are
AB  AB AC   AC 1
 xx  ,  yy  ,  xy      . (4.1.7)
AB AC 2
C
y
deforms

C B

A
A B
x
Figure 4.1.7: strain at a point A
These 2D strains can be represented in the matrix form
  xy 
    xx  yy 
(4.1.8)
 yx
As with the stress, the strain matrix is symmetric, with, by definition,  xy   yz .

Section 4.1
Note that the point A in Fig. 4.1.7 has also undergone a displacement u(A). This
displacement has two components, u x and u y , as shown in Fig. 4.1.8 (and similarly for
the points B and C).
A
u( A)
uy
A
ux
Figure 4.1.8: displacement of a point A
The line elements not only change length and the angle between them changes – they can
also move in space as rigid-bodies. Thus, for example, the normal and shear strain in the
three examples shown in Fig. 4.1.9 are the same, even though the displacements occurring
in each case are different – strain is independent of rigid body motions.
Figure 4.1.9: rigid body motions
The Engineering Strain
Suppose now that the deformation is very small, so that, in Fig. 4.1.10, AB   AB * –
here AB * is the projection of AB  in the x  direction. In that case,
AB *  AB
 xx  .
B
(4.1.9)
AB
Similarly, one can make the approximations
AC *  AC 1  B* B C *C  
 yy  ,  xy     , (4.1.10)
AC 2  AB AC 
the expression for shear strain following from the fact that, for a small angle, the angle
(measured in radians) is approximately equal to the tan of the angle.

Section 4.1
y C* C

deforms
B
C

A B*
A B
x
Figure 4.1.10: small deformation
This approximation for the normal strains is called the engineering strain or small strain
or infinitesimal strain and is valid when the deformations are small. The advantage of
the small strain approximation is that the mathematics is simplified greatly.
Example
Two perpendicular lines are etched onto the fuselage of an aircraft. During testing in a
wind tunnel, the perpendicular lines deform as in Fig. 4.1.10. The coordinates of the line
end-points (referring to Fig. 4.1.10) are:
C : (0.0000,1.0000) C  : (0.0025,1.0030)
A : (0.0000,0.0000) A : (0.0000,0.0000)
B : (1.0000,0.0000) B : (1.0045,0.0020)
The exact strains are, from Eqn. 4.1.9, (to 8 decimal places)
2 2
AB*  B* B
 xx   1  0.00450199
AB
2 2
AC *  C *C 
 yy   1  0.00300312
AC
1  B * B   C *C   
 xy   arctan    arctan     0.00224178
2  AB*   AC *  
    
The engineering strains are, from Eqns. 4.1.10-11,
AB* AC * 1  B B C C  
* *
 xx   1  0.0045,  yy   1  0.003,  xy      0.00225

AB AC 2  AB AC 
 
As can be seen, for the small deformations which occurred, the errors in making the
small-strain approximation are extremely small, less than 0.11% for all three strains.
■

Section 4.1
Small strain is useful in characterising the small deformations that take place in, for
example, (1) engineering materials such as concrete, metals, stiff plastics and so on, (2)
linear viscoelastic materials such as many polymeric materials (see Chapter 9), (3) some
porous media such as soils and clays at moderate loads, (4) almost any material if the
loading is not too high.
Small strain is inadequate for describing large deformations that occur, for example, in
many rubbery materials, soft tissues, engineering materials at large loads, etc. In these
cases the more precise definition 4.1.7 (or a variant of it), as developed and used in Book
III, is required. That said, the engineering strain and the concepts associated with it are an
excellent introduction to the more involved large deformation strain measures.
In one dimension, there is no distinction between the exact strain and the engineering
strain – they are the same. Differences arise between the two in the two-dimensional case
when the material shears (as in the example above), or rotates as a rigid body (as will be
discussed further below).
Engineering Shear Strain and Tensorial Shear Strain
The definition of shear strain introduced above is the tensorial shear strain  xy . The
engineering shear strain2  xy is defined as twice this angle, i.e. as    , and is often
used in Strength of Materials and elementary Solid Mechanics analyses.
4.1.3 Sign Convention for Strain

A positive normal strain means that the line element is lengthening. A negative normal
strain means the line element is shortening.
For shear strain, one has the following convention: when the two perpendicular line
elements are both directed in the positive directions (say x and y ), or both directed in the
negative directions, then a positive shear strain corresponds to a decrease in right angle.
Conversely, if one line segment is directed in a positive direction whilst the other is
directed in a negative direction, then a positive shear strain corresponds to an increase in
angle. The four possible cases of shear strain are shown in Fig. 4.1.11a (all four shear
strains are positive). A box undergoing a positive shear and a negative shear are also
shown, in Figs. 4.1.11b,c.
2
not to be confused with the term engineering strain, i.e. small strain, used throughout this Chapter

Section 4.1
(a) (b) (c)
Figure 4.1.11: sign convention for shear strain; (a) line elements undergoing positive
shear, (b) a box undergoing positive shear, (c) a box undergoing negative shear
4.1.4 Geometrical Interpretation of the Engineering Strain

Consider a small “box” element and suppose it to be so small that the strain is
constant/uniform throughout - one says that the strain is homogeneous. This implies that
straight lines remain straight after straining and parallel lines remain parallel. A few
simple deformations are examined below and these are related to the strains.
A positive normal strain  xx  0 is shown in Fig. 4.1.12a. Here the undeformed box
element (dashed) has elongated. Knowledge of the strain alone is not enough to
determine the position of the strained element, since it is free to move in space as a rigid
body. The displacement over some part of the box is usually specified, for example the
left hand end has been fixed in Fig. 4.1.12b. A negative normal strain acts in Fig. 4.1.12c
and the element has contracted.
y y y
fixed
x x x
(a ) ( b) ( c)
Figure 4.1.12: normal strain; (a) positive normal strain, (b) positive normal strain
with the left-hand end fixed in space, (c) negative normal strain
A case known as simple shear is shown in Fig. 4.1.13a, and that of pure shear is shown
in Fig. 4.1.13b. In both illustrations,  xy  0 . A pure (rigid body) rotation is shown in
Fig. 4.1.13c (zero strain).

Section 4.1
y y y
x x x
(a ) ( b) ( c)
Figure 4.1.13: (a) simple shear, (b) pure shear, (c) pure rotation
Any shear strain can be decomposed into a pure shear and a pure rotation, as illustrated in
Fig. 4.1.14.
 
arbitrary shear strain pure shear pure rotation
Figure 4.1.14: shear strain decomposed into a pure shear and a pure rotation
4.1.5 Large Rotations and the Small Strain

The example in section 4.2 above illustrated that the small strain approximation is good,
provided the deformations are small. However, this is provided also that any rigid body
rotations are small. To illustrate this, consider a square material element which
undergoes a pure rigid body rotation of  , Fig. 4.1.15. The exact strains remain zero.
The small shear strain remains zero also. However, the small normal strains are seen to
be  xx   yy  cos   1 . Using a Taylor series expansion, this is equal to
 xx   yy   2 / 2   4 / 24   . Thus, when  is small, the rotation-induced strains are
of the magnitude/order  2 . If  is of the same order as the strains themselves, i.e. in the
range 106  102 , then  2 will be very much smaller than  and the rotation-induced
strains will not introduce any inaccuracy; the small strains will be a good approximation
to the actual strains. If, however, the rotation is large, then the engineering normal strains
will be wildly inaccurate. For example, when   45o , the rotation-induced normal
strains are  0.3 , and will likely be larger than the actual strains occurring in the
material.

Section 4.1


Figure 4.1.15: an element undergoing a rigid body rotation
As an example, consider a cantilevered beam which undergoes large bending, Fig. 4.1.16.
The shaded element shown might well undergo small normal and shear strains. However,
because of the large rotation of the element, additional spurious engineering normal
strains are induced. Use of the precise definition, Eqn. 4.1.7, is required in cases such as
this.
Figure 4.1.16: Large rotations of an element in a bent beam
4.1.6 Three Dimensional Strain

The above can be generalized to three dimensions. In the general case, there are three
normal strains,  xx ,  yy ,  zz , and three shear strains,  xy ,  yz ,  zx . The  zz strain
corresponds to a change in length of a line element initially lying along the z axis. The
 yz strain corresponds to half the change in the originally right angle of two perpendicular
line elements aligned with the y and z axes, and similarly for the  zx strain. Straining in
the y  z plane (  yy ,  zz ,  yz ) is illustrated in Fig. 4.1.17 below.
Figure 4.1.17: strains occurring in the y – z plane
The 3D strains can be represented in the (symmetric) matrix form

Section 4.1
 xx  xy  xz 
    yx  yy  yz 

(4.1.11)
  zx  zy  zz 
As with the stress (see Eqn. 3.4.4), there are nine components in 3D, with 6 of them being
independent.
4.1.7 Problems
1. Consider a rod which moves and deforms (uniformly) as shown below.
(a) What is the displacement of the left-hand end of the rod?
(b) What is the engineering strain at the left-hand end of the rod
5cm 7cm 8cm

original position new position
2. A slender rod of initial length 2cm is extended (uniformly) to a length 4cm. It is then
compressed to a length of 3cm.
(a) Calculate the engineering strain and the true strain for the extension
(b) Calculate the engineering strain and the true strain for the compression
(c) Calculate the engineering strain and the true strain for one step, i.e. an extension
from 2cm to 3cm.
(d) From your calculations in (a,b,c), which of the strain measures is additive?
3. An element undergoes a homogeneous strain, as shown. There is no normal strain in

the element. The angles are given by   0.001 and   0.002 radians. What is the
(tensorial) shear strain in the element?

x
4. In a fixed x  y reference system established for the test of a large component, three
points A, B and C on the component have the following coordinates before and after
loading (see figure):

Section 4.1
C : (0.0000,1.5000) C  : (0.0025,1.5030)
A : (0.0000, 0.0000) A : (0.0000, 0.0000)
B : (2.0000, 0.0000) B  : (2.0045, 0.0000)
Determine the actual strains and the small strains (at/near point A). What is the error
in the small strain compared to the actual strains?
5. Sketch the deformed shape for the material shown below under the following strains
(A, B constant):
(i)  xx  A  0 (taking  yy   xy  0 ) – assume that the right-hand edge is fixed
(ii)  yy  B  0 (with  xx   xy  0 ) – assume that the lower edge is fixed
(iii)  xy  B  0 (with  xx   yy  0 ) – assume that the left-hand edge is fixed
6. The element shown below undergoes the change in position and dimensions shown
(dashed square = undeformed). What are the three engineering strains  xx ,  xy ,  yy ?
1 50
1 2 1

Section 4.2
4.2 Plane Strain

A state of plane strain is defined as follows:
Plane Strain:
If the strain state at a material particle is such that the only non-zero strain components act
in one plane only, the particle is said to be in plane strain.
The axes are usually chosen such that the x  y plane is the plane in which the strains are
non-zero, Fig. 4.2.1.
y
 xx ,  yy ,  xy
Figure 4.2.1: non-zero strain components acting in the x – y plane
Then  xz   yz   zz  0 . The fully three dimensional strain matrix reduces to a two

dimensional one:
 xx  xy  xz 
   xx  xy 
 yx  yy  yz     yy 
(4.2.1)
 zx  zy  zz   yx

4.2.1 Analysis of Plane Strain

Stress transformation formulae, principal stresses, stress invariants and formulae for
maximum shear stress were presented in §4.4-§4.5. The strain is very similar to the
stress. They are both mathematical objects called tensors, having nine components, and
all the formulae for stress hold also for the strain. All the equations in section 3.5.2 are
valid again in the case of plane strain, with  replaced with  . This will be seen in what
follows.
Strain Transformation Formula
Consider two perpendicular line-elements lying in the coordinate directions x and y , and
suppose that it is known that the strains are  xx ,  yy ,  xy , Fig. 4.2.2. Consider now a
second coordinate system, with axes x , y  , oriented at angle  to the first system, and
consider line-elements lying along these axes. Using some trigonometry, it can be shown
that the line-elements in the second system undergo strains according to the following

Section 4.2
(two dimensional) strain transformation equations (see the Appendix to this section,
§4.2.5, for their derivation):
 xx  cos 2  xx  sin 2  yy  sin 2 xy

 yy  sin 2  xx  cos 2  yy  sin 2 xy Strain Transformation Formulae (4.2.2)
 xy  sin  cos  ( yy   xx )  cos 2 xy
y
y
x

x
Figure 4.2.2: A rotated coordinate system
Note the similarity between these equations and the stress transformation formulae, Eqns.
3.4.8. Although they have the same structure, the stress transformation equations were
derived using Newton’s laws, whereas no physical law is used to derive the strain
transformation equations 4.2.2, just trigonometry.
Eqns. 4.2.2 are valid only when the strains are small (as can be seen from their derivation
in the Appendix to this section), and the engineering/small strains are assumed in all
which follows. The exact strains, Eqns. 4.1.7, do not satisfy Eqn. 4.2.2 and for this reason
they are rarely used in 2D analyses – when the strains are large, other strain measures,
such as those in Eqns. 4.1.4, are used.
Principal Strains
Using exactly the same arguments as used to derive the expressions for principal stress,
there is always at least one set of perpendicular line elements which stretch and/or
contract, but which do not undergo angle changes. The strains in this special coordinate
system are called principal strains, and are given by (compare with Eqns. 3.5.5)
1  1
2  xx   yy   1
4 ( xx   yy ) 2   xy2
Principal Strains (4.2.3)
2  1
2  xx   yy  
1
4 ( xx   yy ) 2   xy2
Further, it can be shown that 1 is the maximum normal strain occurring at the point, and
that  2 is the minimum normal strain occurring at the point.
The principal directions, that is, the directions of the line elements which undergo the
principal strains, can be obtained from (compare with Eqns. 3.5.4)
2 xy
tan 2  (4.2.4)
 xx   yy

Section 4.2
Here,  is the angle at which the principal directions are oriented with respect to the x
axis, Fig. 4.2.2.
Maximum Shear Strain
Analogous to Eqn. 3.5.9, the maximum shear strain occurring at a point is
1
 xy max   1   2  (4.2.5)
2
and the perpendicular line elements undergoing this maximum angle change are oriented
at 45 o to the principal directions.
Example (of Strain Transformation)
Consider the block of material in Fig. 4.2.3a. Two sets of perpendicular lines are etched
on its surface. The block is then stretched, Fig. 4.2.3b.
y y
fixed
(a ) x

   x
fixed
( b)
 
Figure 4.2.3: A block with strain measured in two different coordinate systems
This is a homogeneous deformation, that is, the strain is the same at all points. However,
in the x  y description,  xx  0 and  yy   xy  0 , but in the x   y  description, none of
the strains is zero. The two sets of strains are related through the strain transformation
equations.
■
Example (of Strain Transformation)
As another example, consider a square material element which undergoes a pure shear, as
illustrated in Fig. 4.2.4, with
 xx   yy  0,  xy  0.01

Section 4.2
o x
Figure 4.2.4: A block under pure shear
From Eqn. 4.2.3, the principal strains are 1  0.01,  2  0.01 and the principal
directions are obtained from Eqn. 4.2.4 as   45 . To find the direction in which the
maximum normal strain occurs, put   45 in the strain transformation formulae to find
that 1   xx
  0.01 , so the deformation occurring in a piece of material whose sides are
aligned in these principal directions is as shown in Fig. 4.2.5.
0.01a
a
y a x
45o
o
Figure 4.2.5: Principal strains for the block in pure shear
The strain as viewed along the principal directions, and using the x  y system, are as
shown in Fig. 4.2.6.
agrees
after
with
before
Figure 4.2.6: Strain viewed from two different coordinate systems

Section 4.2
Note that, since the original x  y axes were oriented at  45 to the principal directions,
these axes are those of maximum shear strain – the original  xy  0.01 is the maximum
shear strain occurring at the material particle.
■
4.2.2 Thick Components

It turns out that, just as the state of plane stress often arises in thin components, a state of
plane strain often arises in very thick components.
Consider the three dimensional block of material in Fig. 4.2.7. The material is
constrained from undergoing normal strain in the z direction, for example by preventing
movement with rigid immovable walls – and so  zz  0 .
Rigid Walls
Figure 4.2.7: A block of material constrained by rigid walls
If, in addition, the loading is as shown in Fig. 4.2.7, i.e. it is the same on all cross sections
parallel to the y  z plane (or x  z plane) – then the line elements shown in Fig. 4.2.8
will remain perpendicular (although they might move out of plane).
 xz  0
e zx  0
 yz  0
Figure 4.2.8: Line elements etched in a block of material – they remain

perpendicular in a state of plane strain
Then  xz   yz  0 . Thus a state of plane strain will arise.

Section 4.2
The problem can now be analysed using the three independent strains, which simplifies
matters considerable. Once a solution is found for the deformation of one plane, the
solution has been found for the deformation of the whole body, Fig. 4.2.9.
Figure 4.2.9: three dimensional problem reduces to a two dimensional one for the
case of plane strain
Note that reaction stresses  zz act over the ends of the large mass of material, to prevent
any movement in the z direction, i.e.  zz strains, Fig. 4.2.10.
 zz
z
Figure 4.2.10: end-stresses required to prevent material moving in the z direction
A state of plane strain will also exist in thick structures without end walls. Material
towards the centre is constrained by the mass of material on either side and will be
(approximately) in a state of plane strain, Fig. 4.2.10.
material in a state of
plane strain
Figure 4.2.10: material in an approximate state of plane strain
The concept of Plane Strain is useful when solving many types of problem involving thick
components, even when the ends of the mass of material are allowed to move (as in Fig.
4.2.10); this idea will be explored in the context of generalised plane strain and
associated topics in Book II.

Section 4.2
4.2.3 Mohr’s Circle for Strain

Because of the similarity between the stress transformation equations 3.4.8 and the strain
transformation equations 4.2.2, Mohr’s Circle for strain is identical to Mohr’s Cirlce for
stress, section 3.5.5, with  replaced by  (and  replaced by  xy ).
4.2.4 Problems
1. In Fig. 4.2.3, take   30o and  xx  0.02 .

(a) Calculate the strains  xx ,  yy ,  xy .
(b) What are the principal strains?
(c) What is the maximum shear strain?
(d) Of all the line elements which could be etched in the block, at what angle  to
the x axis are the perpendicular line elements which undergo the largest angle
change from the initial right angle?
2. Consider the undeformed rectangular element below left which undergoes a uniform
strain as shown centre.
(a) Calculate the engineering strains  xx ,  yy ,  xy
(b) Calculate the engineering strains  xx ,  yy ,  xy . Hint: use the two half-diagonals
EC and ED sketched; by superimposing points E , E  (to remove the rigid body
motion of E), it will be seen that point D moves straight down and C moves left,
when viewed along the x, y axes, as shown below right.
(c) Use the strain transformation formulae 4.2.2 and your results from (a) to check
your results from (b). Are they the same?
(d) What is the actual unit change in length of the half-diagonals? Does this agree
with your result from (b)?
1 1.01 y
C D C C
C D
y E x x
1 0.99
y E
D
x
A B A B E, E
D
3. Repeat problem 2 only now consider the larger deformation shown below:
(a) Calculate the engineering strains  xx ,  yy ,  xy
(b) Calculate the engineering strains  xx ,  yy ,  xy
(c) Use the strain transformation formulae 2.4.2 and your results from (a) to check
your results from (b). Are they accurate?
(d) What is the actual unit change in length of the half-diagonals? Does this agree
with your result from (b)?

Section 4.2
1
C D
1.5
y E x C D
1
y 0.5
x E
A B A B
Derivation of the Strain Transformation Formulae
Consider an element ABCD undergoing a strain  xx with  yy   xy  0 to ABC D as

shown in the figure below.
D C C
x
E*
E  E

x
A B B
In the x  y coordinate system, by definition,  xx  BB / AB . In the x  y system, AE

moves to AE  , and one has
EE * cos  EE  BB
 xx    cos 2 
AE AB / cos  AB
which is the first term of the first of Eqn. 4.2.2. The remainder of the transformation
formulae can be derived in a similar manner.

Section 4.3
4.3 Volumetric Strain

The volumetric strain is the unit change in volume due to a deformation. It is an
important measure of deformation and is discussed in what follows.
4.3.1 Two-Dimensional Volumetric Strain

Analogous to Eqn 3.5.1, the strain invariants are
I1   xx   yy
Strain Invariants (4.3.1)
I 2   xx yy   xy2
Using the strain transformation formulae, Eqns. 4.2.2, it will be verified that these
quantities remain unchanged under any rotation of axes.
The first of these has a very significant physical interpretation. Consider the deformation
of the material element shown in Fig. 4.3.1a. The unit change in volume, called the
volumetric strain, is
V ( a  a )(b  b)  ab

V ab
 (1   xx )(1   yy )  1 (4.3.2)
  xx   yy   xx yy
If the strains are small, the term  xx yy will be very much smaller than the other two
terms, and the volumetric strain in that case is given by
V
  xx   yy Volumetric Strain (4.3.3)
V
y y d
b b
b b
c
x x
a a a a
(a) (b)
Figure 4.3.1: deformation of a material element; (a) normal deformation, (b) with
shearing
Since by Eqn. 4.3.1 the volume change is an invariant, the normal strains in any
coordinate system may be used in its evaluation. This makes sense: the volume change

Section 4.3
cannot depend on the particular axes we choose to measure it. In particular, the principal
strains may be used:
V
 1   2 (4.3.4)
V
The above calculation was carried out for stretching in the x and y directions, but the
result is valid for any arbitrary deformation. For example, for the general deformation
shown in Fig. 4.3.1b, the volumetric strain is V / V   xx   yy   xx yy   c / a  d / b  ,
which again reduced to Eqns 4.3.3, 4.3.4, for small strains.
An important consequence of Eqn. 4.3.3 is that normal strains induce volume changes,
whereas shear strains induce a change of shape but no volume change.
4.3.2 Three Dimensional Volumetric Strain

A slightly different approach will be taken here in the three dimensional case, so as not to
simply repeat what was said above.
Consider the element undergoing strains  xx ,  xy , etc., Fig. 4.3.2a. The same deformation
is viewed along the principal directions in Fig. 4.3.2b, for which only normal strains arise.
The volumetric strain is:
V ( a  a )(b  b)( c  c )  abc


V abc
 (1  1 )(1   2 )(1   3 )  1 (4.3.5)
 1   2   3
and the squared and cubed terms can be neglected because of the small-strain assumption.
Since any elemental volume such as that in Fig. 4.3.2a can be constructed out of an
infinite number of the elemental cubes shown in Fig. 4.3.3b, this result holds for any
elemental volume irrespective of shape.

Section 4.3
y
x( 2 )
x (1 )
c
c
x
b a principal
b a directions
z
x(3)
(a ) ( b)
Figure 4.3.2: A block of deforming material; (a) subjected to an arbitrary strain; (a)
principal strains

5 Material Behaviour and
Mechanics Modelling
In this Chapter, the real physical response of various types of material to different types
of loading conditions is examined. The means by which a mathematical model can be
developed which can predict such real responses is considered.
113
114
Section 5.1
5.1 Mechanics Modelling
5.1.1 The Mechanics Problem

Typical questions which mechanics attempts to answer were given in Section 1.1. In the
examples given, one invariably knows (some of) the forces (or stresses) acting on the
material under study, be it due to the wind, water pressure, the weight of the human body,
a moving train, and so on. One also often knows something about the displacements
along some portion of the material, for example it might be fixed to the ground and so the
displacements there are zero. A schematic of such a generic material is shown in Fig.
5.1.1 below.
Known forces
acting on boundary
?
Known
displacements at
boundary
Figure 5.1.1: a material component; force and displacement are known along some
portion of the boundary
The basic problem of mechanics is to determine what is happening inside the material.
This means: what are the stresses and strains inside the material? With this information,
one can answer further questions: Where are the stresses high? Where will the material
first fail? What can we change to make the material function better? Where will the
component move to? What is going on inside the material, at the microscopic level?
Generally speaking, what is happening and what will happen?
One can relate the loads on the component to the stresses inside the body using
equilibrium equations and one can relate the displacement to internal strains using
kinematics relations. For example, consider again the simple rod subjected to tension
forces examined in Section 3.3.1, shown again in Fig. 5.1.2. The internal normal stress
 N on any plane oriented at an angle  to the rod cross-section is related to the external
force F through the equilibrium equation 3.3.1:  N  F cos2  / A , where A is the cross-
sectional area. Similarly, if the ends undergo a separation/displacement of   l  l0 , Fig.
5.1.2b, the strain of any internal line element, at orientation  , is  N   cos2  / l0 .
However, there is no relationship between this internal stress and internal strain: for any
given force, there is no way to determine the internal strain (and hence displacement of
the rod); for any given displacement of the rod, there is no way to determine the internal
stress (and hence force applied to the rod). The required relationship between stress and
strain is discussed next.

Section 5.1
N 
l0 l
internal
surface

F
(a ) ( b)
Figure 5.1.2: a slender rod; (a) internal stress due to external force, (b) internal
strain due to gross displacement of rod (dotted = before straining)
5.1.2 Constitutive Equation

Stress was discussed in Chapter 3 and strain in Chapter 4. In all that discussion, no
mention was made of the particular material under study, be it metallic, polymeric,
biological or foodstuff (apart from the necessity that the strain be small when using the
engineering strain). The concept of stress and the resulting theory of stress
transformation, principal stresses and so on, are based on physical principles (Newton’s
Laws), which apply to all materials. The concept of strain is based, essentially, on
geometry and trigonometry; again, it applies to all materials. However, it is the
relationship between stress and strain which differs from material to material.
The relationship between the stress and strain for any particular material will depend on
the microstructure of that material – what constitutes that material. For this reason, the
stress-strain relationship is called the constitutive relation, or constitutive law. For
example, metals consist of a closely packed lattice of atoms, whereas a rubber consists of
a tangled mass of long-chain polymer molecules; for this reason, the strain in a metal will
be different to that in rubber, for any given stress.
The constitutive equation allows the mechanics problem to be solved – this is shown
schematically in Fig. 5.1.3.

Section 5.1
External External
Loads Displacements
Equilibrium Kinematics
Internal Internal
Stresses Constitutive Strains
Equation
Figure 5.1.3: the role of the constitutive equation in the equations of mechanics
Example Constitutive Equations
A constitutive equation will be of the general form
  f   . (5.1.1)
The simplest constitutive equation is a linear elastic relation, in which the stress is
proportional to the strain:
  . (5.1.2)
Although no real material satisfies precisely Eqn. 5.1.2, many do so approximately – this
type of relation will be discussed in Chapters 6-8. More complex relations can involve
the rate at which a material is strained or stressed; these types of relation will be
discussed in Chapter 10.
More on constitutive equations will follow in Section 5.3.
5.1.3 Mathematical Model

Some of the questions asked earlier can be answered using experimentation. For
example, one could use a car-crash test to determine the weakest points in a car.
However, one cannot carry our multiple tests for each and every possible scenario –
different car speeds, different obstacles into which it crashes, and so on; it would be too
time-consuming and too expensive. The only practical way in which these questions can
be answered is to develop a mathematical model. This model consists of the various
equations of equilibrium and the kinematics, the constitutive relation, equations
describing the shape of the material, etc. (see Fig. 5.1.3). The mathematical model will
have many approximations to reality associated with it. For example, it might be
assumed that the material is in the shape of a perfect sphere, when in fact it only
resembles a sphere. It may be assumed that a load is applied at a “point” when in fact it is
applied over a region of the material’s surface. Another approximation in the
mathematical model is the constitutive equation itself; the relation between stress and

Section 5.1
strain in any material can be extremely complex, and the constitutive equation can only
be an approximation of the reality.
Once the mathematical model has been developed, the various equations can be solved
and the model can then be used to make a prediction. The prediction of the model can
now be tested against reality: a set of well-defined experiments can be carried out – does
the material really move to where the model says it will move?
Simple models (simple constitutive relations) should be used as a first step. If the
predictions of the model are wildly incorrect, the model can be adjusted (made more
complicated), and the output tested again.
The equations associated with simple models can often be solved analytically, i.e. using a
pen and paper. More complex models result in complex sets of equations which can only
be solved approximately (though, hopefully, accurately) using a computer.

Section 5.2
5.2 The Response of Real Materials

The constitutive equation was introduced in the previous section. The means by which
the constitutive equation is determined is by carrying out experimental tests on the
material in question. This topic is discussed in what follows.
5.2.1 The Tension Test

Consider the following key experiment, the tensile test, in which a small, usually
cylindrical, specimen is gripped and stretched, usually at some given rate of stretching. A
typical specimen would have diameter about 1cm and length 5cm, and larger ends so that
it can be easily gripped, Fig. 5.2.1a. Specialised machines are used, for example the
Instron testing machine shown in Fig. 5.2.1b.
Test specimen
1
2
Figure 5.2.1: the tension test; (a) test specimen, (b) testing machine
As the specimen is stretched, the force required to hold the specimen at a given
displacement/stretch is recorded1.
The Engineering Materials
For many of the (hard) engineering materials, the force/displacement curve will look
something like that shown in Fig. 5.2.2. It will be found that the force is initially
proportional to displacement as with the linear portion OA in Fig. 5.2.2. The following
observations will also be made:
(1) if the load has not reached point A, and the material is then unloaded, the
force/displacement curve will trace back along the line OA down to zero force and
zero displacement; further loading and unloading will again be up and down OA .
(2) the loading curve remains linear up to a certain force level, the elastic limit of the
material (point A). Beyond this point, permanent deformations are induced2; on
1
the very precise details of how the test should be carried out are contained in the special standards for
materials testing developed by the American Society for Testing and Materials (ASTM)

Section 5.2
unloading to zero force (from point B to C), the specimen will have a permanent
elongation. An example of this response (although not a tension test) can be seen
with a paper clip – gently bend the clip and it will “spring back” (this is the OA
behaviour); bend the clip too much (AB) and it will stay bent after you let go (C).
(3) above the elastic limit (from A to B), the material hardens, that is, the force
required to maintain further stretching, unsurprisingly, keeps increasing. (However,
some materials can soften, for example granular materials such as soils).
(4) the rate (speed) at which the specimen is stretched makes little difference to the
results observed (at least if the speed and/or temperature is not too high).
(5) the strains up to the elastic limit are small, less than 1% (see below for more on
strains).
Force B
A Parallel
elastic 
limit
unload
0 C Displacement
Permanent
displacement after
removal of load
Figure 5.2.2: force/displacement curve for the tension test; typical response for
engineering materials
Stress-Strain Curve
There are two definitions of stress used to describe the tension test. First, there is the
force divided by the original cross sectional area of the specimen A0 ; this is the nominal
stress or engineering stress,
F
n  (5.2.1)
A0
Alternatively, one can evaluate the force divided by the (smaller) current cross-sectional
area A, leading to the true stress
2
if the tension tests are carried out extremely carefully, one might be able to distinguish between a point
where the stress-strain curve ceases to be linear (the proportional limit) and the elastic limit (which will
occur at a slightly higher stress)

Section 5.2
F
 (5.2.2)
A
in which F and A are both changing with time. For small elongations, within the linear
range OA , the cross-sectional area of the material undergoes negligible change and both
definitions of stress are more or less equivalent.
Similarly, one can describe the deformation in two alternative ways. As discussed in
Section 4.1.1, one can use the engineering strain
l  l0
 (5.2.3)
l0
or the true strain
l
 t  ln   (5.2.4)
 l0 
Here, l 0 is the original specimen length and l is the current length. Again, at small
deformations, the difference between these two strain measures is negligible.
The stress-strain diagram for a tension test can now be described using the true
stress/strain or nominal stress/strain definitions, as in Fig. 5.2.3. The shape of the
nominal stress/strain diagram, Fig. 5.2.3a, is of course the same as the graph of force
versus displacement. C here denotes the point at which the maximum force the specimen
can withstand has been reached. The nominal stress at C is called the Ultimate Tensile
Strength (UTS) of the material.
After the UTS is reached, the specimen “necks”, that is, the specimen begins to deform
locally – with a very rapid reduction in cross-sectional area somewhere about the centre
of the specimen until the specimen breaks, as indicated by the asterisk in Fig. 5.2.3. The
appearance of a test specimen at each of these stages of the stress-strain curve is shown
top of Fig. 5.2.3a.
For many materials, it will be observed that there is little or no volume change during the
permanent deformation phase, so A0l0  Al and    N 1    . This nominal stress to
true stress conversion formula will only be valid up to the point of necking.

Section 5.2
n  C
C 
 t
(a ) (b)
Figure 5.2.3: typical stress-strain curve for an engineering material; (a) engineering
stress and strain, (b) true stress and strain
The stress-strain curves for mild steel and aluminium are shown in Fig. 5.2.4. For mild
steel, the stress at first increases after reaching the elastic limit, but then decreases. The
curve contains a distinct yield point; this is where a large increase in strain begins to
occur with little increase in required stress3, i.e. little hardening. There is no distinct yield
point for aluminium (or, in fact, for most materials), Fig. 5.2.4b. In this case, it is useful
to define a yield strength (or offset yield point). This is the maximum stress that can be
applied without exceeding a specified value of permanent strain. This offset strain is
usually taken to be 0.1 or 0.2% and the yield strength is found by following a line parallel
to the linear portion until it intersects the stress-strain curve.
elastic yield yield

n limit point strength
  

offset

strain
(a ) (b)
Figure 5.2.4: typical stress-strain curves for (a) mild steel, (b) aluminium
3
this is also called the lower yield point; the upper yield point is then the higher stress value just above
the elastic limit

Section 5.2
The Young’s Modulus
The slope of the stress-strain curve over the linear region, before the elastic limit is
reached, is the Young’s Modulus E:

E (5.2.5)

The Young’s Modulus has units of stress and is a measure of how “stiff” a material is.
Eqn. 5.2.5 is a constitutive relation (see Eqn. 5.1.2); it is the one-dimensional linear
elastic constitutive relation.
Use of the Tension Test Data
What is the data from the tension test used for? First of all, it is of direct use in many
structural applications. Many structures, such as bridges, buildings and the human
skeleton, are composed in part of relatively long and slender components. In service,
these components undergo tension and/or compression, very much like the test specimen
in the tension test. The tension test data (the Young’s Modulus, the Yield Strength and
the UTS) then gives direct information on the amount of stress that these components can
safely handle, before undergoing dangerous straining or all-out failure.
More importantly, the tension test data (and similar test data – see below) can be used to
predict what will happen when a component of complex three-dimensional shape is
loaded in a complex way, nothing like as in the simple tension test. This can be put
another way: one must be able to predict the world around us without having to resort to
complex, expensive, time-consuming materials testing – one should be able to use the test
data from the tension test (and similar simple tests) to achieve this. How this is actually
done is a major theme of mechanics modelling and these Books.
Test data for a number of metals are listed in Table 5.2.1 below. Note that although some
materials can have similar stiffnesses, for example Nickel and Steel, their relative
strengths can be very different.
Young’s 0.2% Yield Ultimate Tensile

Modulus Strength Strength
E (GPa) (MPa) (MPa)
Ni 200 70 400
Mild steel 203 220 430
Steel (AISI 1144) 210 540 840
Cu 120 60 400
Al 70 40 200
Al Alloy (2014-T651) 73 415 485
Table 5.2.1: Tensile test data for some metals (at room temperature)
Data as listed above should be treated with caution – it should be used only as a rough
guide to the actual material under study; the data can vary wildly depending on the purity
and precise nature of the material. For example, the tensile strength of glass as found in a

Section 5.2
typical glass window is about 50MPa. For fine glass fibres as used in fibre-reinforced
plastics and composite materials, the tensile strength can be 4000MPa. In fact, glass is a
good reminder as to why the tensile values differ from material to material – it is due to
the difference in microstructure. The glass window has many very fine flaws and cracks
in it, invisible to the naked eye, and so this glass is not very strong; very fine slivers of
glass have no such flaws and are extremely strong – hence their use in engineering
applications.
The Poisson’s Ratio
Another useful material parameter is the Poisson’s ratio  .4 As the material stretches in
the tension test, it gets thinner; the Poisson’s ratio is a measure of the ease with which it
thins:
w / w0 
   w (5.2.6)
l / l0 
Here, w  w  w0 , w0 are the change in thickness and original thickness of the specimen,
Fig. 5.2.5; l  l  l0 , l0 are the change in length and original length of the specimen;
 w  ( w  w0 ) / w0 is the strain in the thickness direction. A negative sign is included
because w is negative, making the Poisson’s ratio a positive number. (It is implicitly
assumed here that the material is getting thinner by the same amount in all directions; see
below in the context of anisotropy for when this is not the case.)
Most engineering materials have a Poisson’s Ratio of about 0.3. Values for a range of
materials are listed in Table 5.2.2 further below.
l l0 l l0 y
w0
x
Figure 5.2.5: Change in dimensions of a test specimen
Recall from Section 4.3 that the volumetric strain is given by the sum of the normal
strains. There is no harm in re-calculating this for the tensile test specimen of Fig. 5.2.5.
One has V / V  w2l / w02l0  1 , so that, assuming the strains are small so that the terms
 w ,  w2 and  w2 can be neglected, V / V    2 w (this is the sum of the normal
4
this is the Greek letter nu, not the letter “v”

Section 5.2
strains,  xx   yy   zz , Fig. 5.2.4). Using the definition of the Poisson’s ratio, Eqn. 5.2.6,
one has
V
  1  2  (5.2.7)
V
A material which undergoes little volume change thus has a Poisson’s Ratio close to 0.5;
rubber and other soft tissues, for example biological materials, have Poisson’s Ratios very
close to 0.5. A material which undergoes zero volume change (  0.5 ) is called
incompressible (see more on incompressibility in Section 5.2.4 below). At the other
extreme, materials such as cork can have Poisson’s Ratios close to zero. The reason for
this can be seen from the microstructure of cork shown in Fig. 5.2.6; when tested in
compression, the hexagonal honeycomb structure simply folds down, with no necessary
lateral expansion.
Figure 5.2.6: Microstructure of Cork
Auxetic materials are materials which have a negative Poisson’s Ratio; when they are
stretched, they get thicker. Examples can be found amongst polymers, foams, rocks and
biological materials. These materials obviously have a very particular microstructure. A
typical example is the network microstructure shown in Fig. 5.2.7.
(a ) (b)
Figure 5.2.7: Auxetic material (a) before loading, (b) after loading

Section 5.2
Ductile and Brittle Materials
The engineering materials can be grouped into two broad classes: the ductile materials
and the brittle materials. The ductile materials undergo large permanent deformations,
stretching and necking before failing5. The term ductile rupture is usually reserved for
materials which fail in this way. The separate pieces of the specimen pull away from
each other gradually, leaving rough surfaces. A simple measure of ductility is the
engineering strain at failure. The brittle materials are generally more stiff and strong, but
fail without undergoing much permanent deformation – the tension specimen undergoes a
sudden clean break – a fracture. The UTS in the case of a brittle material is the same as
the failure/fracture stress. Ceramics and glasses are extremely brittle – they fractures
suddenly without undergoing any permanent deformation. The difference is illustrated
schematically in Fig. 5.2.8 below.
clean
n fracture
brittle 
Moderate
ductile
rupture
ductile 
Severe
ductile
rupture

Figure 5.2.8: the difference between ductile and brittle materials
Ductility will depend on temperature – a very cold metal will tend to shatter suddenly,
whereas it will stretch more easily when hot.
Soft Materials
Tension test data for non-engineering materials can be very different to that given above.
For example, the typical response of a “soft” material, such as rubber, is shown in Fig.
5.2.9. For many soft materials, the elastic limit (or yield strength) can be very high on the
stress-strain curve, close to failure. Most of the curve is elastic, meaning that when one
unloads the material, the unloading curve traces over the loading curve back down to zero
stress and zero strain: the material does not undergo any permanent deformation6. Note
that the stress-strain curve is non-linear (curved), unlike the straight line elastic portion
for a typical metal, Fig. 5.2.2-4, so these materials do not have a single Young’s Modulus
through which their response can be described.
5
the term ductile is used for a specimen in tension; the analogous term for compression is malleability – a
malleable material is easily “squashed”
6
here, as elsewhere, these statements should not be taken literally; a real rubber will undergo some
permanent deformation, only it will often be so small that it can be discounted, and an unload curve will
never “exactly” trace over a loading curve

Section 5.2
n 
load
unload
Figure 5.2.9: typical load/unload curves for rubber
5.2.2 Compression Tests

Many materials are used, or designed for use, in compression only, for example soils and
concrete. These materials are tested in compression. A common testing method for
concrete is to place a cylindrical specimen between two parallel plates and bring the
plates together. The typical response of concrete is shown in Fig. 5.2.10a; at failure, the
concrete crushes catastrophically, as in the specimen shown in Fig. 5.2.10b. Nominal
stresses in the region 20-70MPa are typical and a good concrete would strain to much less
than 1% at failure.
n 
 5
(a) (b)
Figure 5.2.10: typical compressive response of concrete; (a) stress-strain curve, (b)
specimen at failure
For many materials, e.g. metals, a compression test will lead to similar results as the
tensile stress. The yield strength in compression will be approximately the same as (the
negative of) the yield strength in tension. If one plots the true stress versus true strain
curve for both tension and compression (absolute values for the compression), and the
two curves more or less coincide, this would indicate that the behaviour of the material
under compression is broadly similar to that under tension. However, if one were to use
the nominal stress and strain, then the two curves would not coincide even if the real
tensile/compressive behaviour was similar (although they would of course in the small-
strain linear region); this is due to the definition of the engineering strain/stress.

Section 5.2
5.2.3 Shear Tests
In the shear test, the material is subjected to a shear strain   2 xy by applying a shear
stress7    xy , Fig. 5.2.11a. The resulting shear stress-strain curve will be similar to the
tensile stress-strain curve, Fig. 5.2.11b. The shear stress at failure, the shear strength,
can be greater or smaller than the UTS. The shear yield strength, on the other hand, is
usually in the region of 0.5-0.75 times the tensile yield strength. In the linear small-strain
region, the shear stress will be proportional to the shear strain; the constant of
proportionality is the shear modulus G:

G (5.2.8)

For many of the engineering materials, G  0.4 E .
  


(a) (b)
Figure 5.2.11: the shear test; (a) specimen subjected to shear stress and shear strain
(dotted = undeformed), (b) shear stress-strain curve
5.2.4 Compressibility
In the confined compression test, a sample is placed in a container and a piston is used
to compress it at some pressure p, Fig. 5.2.12a. This test can be used to determine how
compressible a material is. When a material is compressed by equal pressures on all
sides, the ratio of applied pressure p to (unit) volume change, i.e. volumetric strain
V / V , is called the Bulk Modulus K, Fig. 5.2.12b (this is not quite the situation in Fig.
5.2.12a – the reaction pressures on the side walls will only be about half the applied
surface pressure p; see Section 6.2):
p
K  (5.2.9)
V / V
The negative sign is included since a positive pressure implies a negative volumetric
strain, so that the Bulk Modulus is a positive value.
7
there are many ways that this can be done, for example by pushing blocks of the material over each other,
or using more sophisticated methods such as twisting thin tubes of the material (see Section 7.2)

Section 5.2
p p
 V / V
(a) (b)
Figure 5.2.12: the confined compression test; (a) specimen subjected to confined
compression, (b) pressure plotted against volume change
A material which can be easily compressed has a low Bulk Modulus. As mentioned
earlier, a material which cannot be compressed at all is called incompressible ( K   ).
No real material is incompressible, but some can be regarded as incompressible so as to

make the mechanics modelling easier. For example, the Shear Modulus of rubber is very
much smaller than its Bulk Modulus, Table 5.2.2. Essentially, this means that the shape
of rubber can be easily changed as compared to its volume. Thus, in applications where a
rubber component is being deformed or subjected to arbitrary stressing, it is perfectly
reasonable to simply assume that rubber is incompressible. The same applies, only more
so, to water; the Shear Modulus is effectively zero and there is no resistance to change in
shape (which will be observed on pouring a glass of water on to the ground); it is thus
regarded almost always as completely incompressible. On the other hand, even though
the Bulk Modulus of the metals and other engineering materials is very much larger than
that of water or rubber, they are still regarded as compressible in applications – the
extremely small changes in volume are significant.
Young’s Modulus E Shear Modulus G Bulk Modulus K Poissons

(GPa) (GPa) (GPa) Ratio
Ni 200 76 180 0.31
Mild steel 203 78 138 0.30
Steel (AISI 1144) 210 80 140 0.31
Cu 120 46 142 0.34
Al 70 26 76 0.35
Rubber 14.9  104 5  104 1 0.49
14 14
Water  10  10 2.2
Table 5.2.2: Moduli and Poisson’s Ratios for a number of materials
5.2.5 Cyclic Tests

Many materials are subjected to complex loading regimes when in service, not simply a
one-off stretching, shearing or compression. A classic example are the wings of an
aircraft which are continually loaded in tension, then compression, then tension and so on,
as in Fig. 5.2.13. Anything moving back and forward is likely to be subjected to this
tension/compression-type cyclic loading. Another example would be the stresses
experienced by cardiac tissue in a pumping heart.

Section 5.2
Figure 5.2.13: cyclic loading; alternating between tension (positive stress) and
compression (negative stress) over time t
Cyclic tests can be carried out to determine the response of materials to such loading
cycles. An example is shown in Fig. 5.2.14a, the stress-strain response of a Stainless
Steel. The Steel is first cycled between two strain values (one positive, one negative,
differing only in sign) a number of times. The stress is seen to increase on each
successive cycle. The strain is then increased for a number of further cycles, and so on.
One does not have to move from tension to compression; many materials cycle in only
tension or compression. For example, the response to cyclic (compressive) loading of
polyurethane foam is shown in Fig. 5.2.14b (note how the loading curve is similar to that
in 5.2.9).
6 7
(a) (b)
Figure 5.2.14: cyclic loading; (a) cyclic straining of a Stainless Steel, (b) cyclic
loading (in compression) of a polyurethane foam
5.2.6 Other Tests

There are other important tests, for example the Vickers and Brinell hardness tests, and
the three-point bending test. The hardness tests will be discussed in Book II. The
bending test is discussed in section 7.4.9, in the context of beam theory. Another two
very important tests, the creep test and the stress relaxation test, will be discussed in
Chapter 10.

Section 5.2
5.2.7 Isotropy and Anisotropy

Many materials have a strong direction-dependence. The classic example is wood, which
has a clear structure –along the grain, along which fine lines can be seen, and against the
grain, Fig. 5.2.15. The wood is stiffer and stronger along the grain than against the grain.
A material which has this direction-dependence of mechanical (and physical) properties is
called anisotropic.
Figure 5.2.15: Wood
Fig. 5.2.16 shows stress-strain curves for human ligament tissue; in one test, the ligament
is stretched along its length (the longitudinal direction), in the second, across the width
of the ligament (the transverse direction). It can be seen that the stiffness is much higher
in the longitudinal direction. Another example is bone – it is much stiffer along the
length of the bone than across the width of the bone. In fact, many biological materials
are strongly anisotropic.
Figure 5.2.16: Anisotropic response of human ligament
A material whose properties are the same in all directions is called isotropic. In
particular, the relationship between stress and strain at any single location in a material is
the same in all directions. This implies that if a specimen is cut from an isotropic material
and subjected to a load, it would not matter in which orientation the specimen is cut, the

Section 5.2
resulting deformation would be the same – as illustrated in Fig. 5.2.17. Most metals and
ceramics can be considered to be isotropic (see Section 5.4).
o
o
o
o
Figure 5.2.17: Illustration of Isotropy; the stress-strain response is the same no

matter in what “direction” the test specimen is cut from the material
Anisotropy will be examined in more detail in §6.3. It will be shown there, for example,
that an anisotropic material can have a Poisson’s ratio greater than 0.5.8
5.2.8 Homogeneous Materials

The term homogeneous means that the mechanical properties are the same at each point
throughout the material. In other words, the relationship between stress and strain is the
same for all material particles. Most materials can be assumed to be homogeneous.
In engineering applications, it is sometimes beneficial to design materials/components

which are specifically not homogeneous, i.e. inhomogeneous. Such materials whose
properties vary gradually throughout are called Functionally Graded Materials, and
have been gaining popularity since the 1980s-90s in advanced technologies.
Note that a material can be homogeneous and not isotropic, and vice versa –
homogeneous refers to different locations whereas isotropy refers to the same location.
5.2.9 Problems
1. Steel and aluminium can be considered to be isotropic and homogeneous materials. Is
the composite sandwich-structure shown here isotropic and/or homogeneous?
Everywhere in the sandwich?
steel
aluminiu
steel
8
cork was mentioned earlier and it was pointed out that it has a near-zero Poisson’s Ratio; actually, cork is
quite anisotropic and the Poisson’s Ratio in other “directions” will be different (close to 1.0)

Section 5.2
Images used:
1. http://site.metacos.com/main/3108/index.asp?pageid=84386&t=&AlbumID=0&page=2
2. http://travisarp.wordpress.com/2012/01/24/how-do-they-do-that-tenderness/
3. http://www.ara.com/Projects/SVO/popups/weld_geometry.html
4. http://10minus9.wordpress.com/2010/03/23/10minus9-interview-philip-moriarty-part-2/
5. http://info.admet.com/blog/?Tag=Compression%20Test
6. Chaboche JL, On some modifications of kinematic hardening to improve the description of ratcheting effects, Int.
J. Plasticity 7(7), 661-678, 1991.
7. Shen Y, Golnaraghi F, Plumtree A, Modelling compressive cyclic stress-strain behaviour of structural foam, Int J
Fatigue, 23(6), 491-497, 2001.
8. http://www.sufaparket.com/solid-parquet/product-profile-solid/sungkai
9. Quapp KM, Weiss JA, Material characterization of human medial collateral ligament, J Biomech Engng, 120(6),
757-763, 1998.

Section 5.3
5.3 Material Models

The response of real materials to various loading conditions was discussed in the previous
section. Now comes the task of creating mathematical models which can predict this
response. To this end, it is helpful to categorise the material responses into ideal models.
There are four broad material models which are used for this purpose: (1) the elastic
model, (2) the viscoelastic model, (3) the plastic model, and (4) the viscoplastic model.
These models will be discussed briefly in what follows, and in more depth throughout the
rest of this book.
5.3.1 The Elastic Model

An ideal elastic material has the following characteristics:
(i) the unloading stress-strain path is the same as the loading path
(ii) there is no dependence on the rate of loading or straining
(iii) it does not undergo permanent deformation; it returns to its precise original shape
when the loads are removed
Typical stress-strain curves for an ideal elastic model subjected to a tension (or
(compression) test are shown in Fig. 5.3.1. The response of a linear elastic material,
where the stress is proportional to the strain, is shown in Fig. 5.3.1a and that for a non-
linear elastic material is shown in Fig. 5.3.1b.
From the discussion in the previous section, the linear elastic model will well represent
the engineering materials up to their elastic limit (see, for example, Figs. 5.2.2-4). It will
also represent the complete stress-strain response up to the point of fracture of many very
brittle materials. The model can also be used to represent the response of almost any
material, provided the stresses are sufficiently small.
The non-linear elastic model is useful for predicting the response of soft materials like
rubber and biological soft tissue (see, for example Fig. 5.2.9).
 
load
unload
 
(a ) ( b)
Figure 5.3.1: The Elastic Model; (a) linear elastic, (b) non-linear elastic
It goes without saying that there is no such thing as a purely elastic material. All
materials will undergo at least some permanent deformations, even at low loads; no
material’s response will be exactly the same when stretched at different speeds, and so on.

Section 5.3
However, if these occurrences and differences are small enough to be neglected, the ideal
elastic model will be useful.
Note also that a prediction of a material’s response may be made with accuracy using the
elastic model in some circumstances, but not in others. An example would be metal; the
elastic model might well be able to predict the response right up to high stress levels
when the metal is cold, but not so well when the temperature is high, when inelastic
effects may not be so easily disregarded (see below).
5.3.2 Viscoelasticity
When solid materials have some “fluid-like” characteristics, they are said to be
viscoelastic. A fluid is something which flows easily when subjected to loading – it
cannot keep to any particular shape. If a fluid is one (the “viscous”) extreme and the
elastic solid is at the other extreme, then the viscoelastic material is somewhere in
between.
The typical response of a viscoelastic material is sketched in Fig. 5.3.2. The following
will be noted:
(i) the loading and unloading curves do not coincide, Fig. 5.3.2a, but form a
hysteresis loop
(ii) there is a dependence on the rate of straining d / dt , Fig. 5.3.2b; the faster the
stretching, the larger the stress required
(iii) there may or may not be some permanent deformation upon complete unloading,
Fig. 5.3.2a
  1 2
load 1  2
unload
 
possible permanent
deformation
(a ) ( b)
Figure 5.3.2: Response of a Viscoelastic material in the Tension test; (a) loading and
unloading with possible permanent deformation (non-zero strain at zero stress), (b)
different rates of stretching
The effect of rate of stretching shows that the viscoelastic material depends on time. This
contrasts with the elastic material; it makes no difference whether an elastic material is
loaded to some given stress level for one second or one day, or quickly or slowly, the
resulting strain will be the same. This rate effect can be seen when you push your hand
through water – it is easier to do so when you push slowly than when you push fast.

Section 5.3
Depending on how “fluid-like” or “solid-like” a material is, it can be considered to be a

viscoelastic fluid, for example blood or toothpaste, or a viscoelastic solid, for example
Silly Putty™ or foam. That said, the model for both and the theory behind each will be
similar.
Viscoelastic materials will be discussed in detail in Chapter 10.
5.3.3 Plasticity
Plasticity has the following characteristics:
(i) The loading is elastic up to some threshold limit, beyond which permanent
deformations occur
(ii) The permanent deformation, i.e. the plasticity, is time independent
This plasticity can be seen in Figs. 5.2.2-4. The threshold limit – the elastic limit – can be
quite high but it can also be extremely small, so small that significant permanent
deformations occur at almost any level of loading. The plasticity model is particularly
useful in describing the permanent deformations which occur in metals, soils and other
engineering materials. It will be discussed in further detail in Chapter 11.
5.3.4 Viscoplasticity
Finally, the viscoplastic model is a combination of the viscoelastic and plastic models. In
this model, the plasticity is rate-dependent. One of the main applications of the model is
in the study of metals at high temperatures, but it is used also in the modeling of a huge
range of materials and other applications, for example asphalt, concrete, clay, paper pulp,
biological cells growth, etc. This model will be discussed in Chapter 12.

Section 5.4
5.4 Continuum Models and Micromechanics

The models mentioned in the previous section are continuum models. What this means
is explained in what follows.
5.4.1 Stress and Scale

In the definition of the traction vector, §3.3.1, it was assumed that the ratio of force over
area would reach some definite limit as the area S of the surface upon which the force
F acts was shrunk to zero. This issue can be explored further by considering Fig. 5.4.1.
Assume first that the plane upon which the force acts is fairly large; it is then shrunk and
the ratio F / S tracked. A schematic of this ratio is shown in Fig. 5.4.2. At first (to the
right of Fig. 5.4.2) the ratio F / S undergoes change, assuming the stress to vary within
the material, as it invariably will if the material is loaded in some complex way.
Eventually the plane will be so small that the ratio changes very little, perhaps with some
small variability  . If the plane is allowed to get too small, however, down below some
distance h * say and down towards the atomic level, where one might encounter
“intermolecular space”, there will be large changes in the ratio and the whole concept of a
force acing on a single surface breaks down.
force
force
plane
point
plane
point
Figure 5.4.1: A force acting on an internal surface; allowing the plane on which the
force acts to get progressively smaller
In a continuum model, it is assumed that the ratio F / S follows the dotted path shown in
Fig. 5.4.2; a definite limit is reached as the plane shrinks to zero size. It should be kept in
mind that the traction in a real material should be evaluated through
F
t  lim (5.4.1)
S  ( h *) 2 S

Section 5.4
where h * is some minimum dimension below which there is no acceptable limit. On the
other hand, it is necessary to take the limit to zero in the mathematical modelling of
materials since that is the basis of calculus (this will become more necessary in Book II,
where calculus is used more extensively).
F
S
getting closer to the
“point”
continuum h*
approximation
more or less constant F/S

Molecular level – with some variability  F/S changing as one moves
away from “point”
Figure 5.4.2: the change in traction as the plane upon which a force acts is reduced
in size
In a continuum model, then, there is a minimum sized element one can consider, say of
size V  (h*) 3 . When one talks about the stress on this element, the mass of this
element, the density, velocity and acceleration of this element, one means the average of
these quantities throughout or over the surface of the element – the discrete atomic
structure within the element is ignored and is “smeared” out into a continuum element.
The continuum element is also called a representative volume element (RVE), an

element of material large enough for the heterogeneities to be replaced by homogenised
mean values of their properties. The order of the dimensions of RVE’s for some common
engineering materials would be approximately (see the metal example which follows)
Metal: 0.1mm
Polymers/composites: 1mm
Wood: 10mm
Concrete: 100mm
One does not have any information about what is happening inside the continuum
element – it is like a “black box”. The scale of the element (and higher) is called the
macroscale – continuum mechanics is mechanics on the macroscale. The scale of
entities within the element is termed the microscale – continuum models cannot give any
information about what happens on the microscale.

Section 5.4
5.4.2 Example: Metal

Metal, from a distance, appears fairly uniform. With the help of a microscope, however,
it will be seen to consist of many individual grains of metal. For example, the metal
shown in Fig. 5.4.3 has grains roughly 0.05mm across, and each one has very individual
properties (the crystals in each grain are aligned in different directions).
Figure 5.4.3: metal grains
If one is interested in the gross deformation of a moderately sized component of this

metal, it would be sufficient to consider deformations that are averaged over volumes
which are large compared to individual grains, but small compared to the whole
component. A minimum dimension of, say, h *  0.5mm for the metal of Fig. 5.4.3 would
suffice, and this would be the macro/micro-scale boundary, with a minimum surface area
of dimension (h * ) 2 for the definition of stress.
When one measures physical properties of the metal “at a point”, for example the density,
one need only measure an average quantity over an element of the order, say, (0.5mm) 3
or higher. It is not necessary to consider the individual grains of metal – these are inside
the “black box”. The model will return valuable information about the deformation of the
gross material, but it will not be able to furnish any information about movement of
individual grains.
It was shown how to evaluate the Young’s Modulus and other properties of a metal in
Section 5.2.1. The test specimens used for such tests are vastly larger than the continuum
elements discussed above. Thus the test data is perfectly adequate to describe the
response of the metal, on the macroscale.
What if the response of individual grains to applied loads is required? In that case a
model would have to be constructed which accounted for the different mechanical
properties of each grain. The metal could no longer be considered to be a uniform
material, but a complex one with many individual grains, each with different properties
and orientation. The macro/micro boundary could be set at about h *  0.1m . There are

Section 5.4
now two problems which need to be dealt with: (1) experiments such as the tensile test
would have to be conducted on specimens much smaller than the grain size in order to
provide data for any mathematical model, and (2) the mathematical model will be more
complex and difficult to solve.
5.4.3 Micromechanical Models

Consider the schematic of a continuum model shown in Fig. 5.4.4 below. One can
determine the material’s properties, such as the Young’s modulus E, through
experimentation, and the resulting mathematical continuum model can be used to make
predictions about the material’s response. With the improved power of computers,
especially since the 1990s, it has now become possible to complement continuum models
with micromechanical models. These models take into account more fine detail of the
material’s structure (for example of the individual grains of the metal discussed earlier).
Usually, one will have a micromechanical model of a small (typical) RVE of material.
This then provides information regarding the properties of the RVE to be included in a
continuum model (rather than having a micromechanical model of the complete material,
which is in most cases still not practical). The means by which the properties at the micro
scale are averaged (for example into a “smeared out” single E value) and passed “up” to
the continuum model is through homogenisation theory. Such micromechanical models
can provide further insight into material behaviour than the simpler continuum model.
continuum
element
E macro-to-
micro scale
Figure 5.4.4: continuum model and micromechanical model
5.4.4 Problems
1. You want to evaluate the stiffness E of a metal for inclusion in a mechanics model.
What minimum size specimen would you use in your test - 10 m , 0.1mm, 5mm or
5cm ?
2. Individual rice grains are separate solid particles. However, rice flowing down a
chute at a food processing plant can be considered to be a fluid, and the flow of rice
can be solved using the equations of mechanics. What minimum dimension h*
should be employed for measurements in this case to ensure the validity of a
continuum model of flowing rice?

6 Linear Elasticity
The simplest constitutive law for solid materials is the linear elastic law, which assumes a
linear relationship between stress and engineering strain. This assumption turns out to be
an excellent predictor of the response of components which undergo small deformations,
for example steel and concrete structures under large loads, and also works well for
practically any material at a sufficiently small load.
The linear elastic model is discussed in this chapter and some elementary problems
involving elastic materials are solved. Anisotropic elasticity is discussed in Section 6.3.
141
142
Section 6.1
6.1 The Linear Elastic Model
6.1.1 The Linear Elastic Model

Repeating some of what was said in Section 5.3: the Linear Elastic model is used to
describe materials which respond as follows:
(i) the strains in the material are small1
(ii) the stress is proportional to the strain,    (linear)
(iii) the material returns to its original shape when the loads are removed, and the
unloading path is the same as the loading path (elastic)
(iv) there is no dependence on the rate of loading or straining
From the discussion in the previous chapter, this model well represents the engineering
materials up to their elastic limit. It also models well almost any material provided the
stresses are sufficiently small.
The stress-strain (loading and unloading) curve for the Linear Elastic solid is shown in
Fig. 6.1.1a. Other possible responses are shown in Figs. 6.1.1b,c. Fig. 6.1.1b shows the
typical response of a rubbery-type material and many biological tissues; these are non-
linear elastic materials (see Book IV). Fig. 6.1.1c shows the typical response of
viscoelastic materials (see Chapter 10) and that of many plastically and viscoplastically
deforming materials (see Chapters 11 and 12).
hysteresis
   loop
load
unload
  
(a ) ( b) (c )
Figure 6.1.1: Different stress-strain relationships; (a) linear elastic, (b) non-linear
elastic, (c) viscoelastic/plastic/viscoplastic
It will be assumed at first that the material is isotropic and homogeneous. The case of an
anisotropic elastic material is discussed in Section 6.3.
1
if the small-strain approximation is not made, the stress-strain relationship will be inherently non-linear;
the actual strain, Eqn. 4.1.7, involves (non-linear) squares and square-roots of lengths

Section 6.1
6.1.2 Stress-Strain Law
Consider a cube of material subjected to a uniaxial tensile stress  xx , Fig. 6.1.2a. One
would expect it respond by extending in the x direction,  xx  0 , and to contract laterally,
so  yy   zz  0 , these last two being equal because of the isotropy of the material. With
stress proportional to strain, one can write
1 
 xx   xx ,  yy   zz    xx (6.1.1)
E E
 xx  xx
 xx  xx
(a ) ( b)
Figure 6.1.2: an element of material subjected to a uniaxial stress; (a) normal strain,
(b) shear strain
The constant of proportionality between the normal stress and strain is the Young’s
Modulus, Eqn. 5.2.5, the measure of the stiffness of the material. The material parameter
 is the Poisson’s ratio, Eqn. 5.2.6. Since  yy   zz   xx , it is a measure of the
contraction relative to the normal extension.
Because of the isotropy/symmetry of the material, the shear strains are zero, and so the
deformation of Fig. 6.1.2b, which shows a non-zero  xy , is not possible – shear strain can
arise if the material is not isotropic.
One can write down similar expressions for the strains which result from a uniaxial
tensile  yy stress and a uniaxial  zz stress:
1 
 yy   yy ,  xx   zz    yy
E E (6.1.2)
1 
 zz   zz ,  xx   yy    zz
E E
Similar arguments can be used to write down the shear strains which result from the
application of a shear stress:
1 1 1
 xy   xy ,  yz   yz ,  xz   xz (6.1.3)
2 2 2

Section 6.1
The constant of proportionality here is the Shear Modulus  , Eqn. 5.2.8, the measure of
the resistance to shear deformation (the letter G was used in Eqn. 5.2.8 – both G and 
are used to denote the Shear Modulus, the latter in more “mathematical” and “advanced”
discussions) .
The strain which results from a combination of all six stresses is simply the sum of the
strains which result from each2:
 xx 
1
E
 
 xx    yy   zz  ,  xy 
1
2
 xy ,  xz 
1
2
 xz ,
 yy
1
E
 
  yy    xx   zz  ,  yz 
1
2
 yz (6.1.4)
 zz
1
E

  zz    xx   yy 
These equations involve three material parameters. It will be proved in §6.3 that an
isotropic linear elastic material can have only two independent material parameters and
that, in fact,
E
 (6.1.5)
21   
This relation will be verified in the following example.
Example: Verification of Eqn. 6.1.5
Consider the simple shear deformation shown in Fig. 6.1.3, with  xy  0 and all other
strains zero. With the material linear elastic, the only non-zero stress is  xy  2 xy .
y x
y
Figure 6.1.3: a simple shear deformation
2
this is called the principle of linear superposition: the "effect" of a sum of "causes" is equal to the sum
of the individual "effects" of each "cause". For a linear relation, e.g.   E , the effects of two causes
 1 ,  2 are E1 and E 2 , and the effect of the sum of the causes  1   2 is indeed equal to the sum of the
individual effects: E (1   2 )  E1  E 2 . This is not true of a non-linear relation, e.g.   E , since
2
E ( 1   2 )
2
 E12  E 22

Section 6.1
Using the strain transformation equations, Eqns. 4.2.2, the only non-zero strains in a
second coordinate system x   y  , with x  at   45 o from the x axis (see Fig. 6.1.3), are
 xx   xy and  yy   xy . Because the material is isotropic, Eqns 6.1.4 hold also in this
second coordinate system and so the stresses in the new coordinate system can be
determined by solving the equations
 xx   xy 
1
E
 
 xx    yy   zz  ,  xy  0 
1
2
 xy ,  xz  0 
1
2
 xz ,
1
 
 yy   xy   yy    xx   zz  ,  yz  0 
E
1
2
 yz (6.1.6)
 zz  0 
1
E

 zz    xx   yy  
which results in
E E
 xx    xy ,  yy    xy (6.1.7)
1 v 1 v
But the stress transformation equations, Eqns. 3.4.8, with  xy  2 xy , give
 xx  2  xy and  yy  2 xy and so Eqn. 6.1.5 is verified.
■
Relation 6.1.5 allows the Linear Elastic Solid stress-strain law, Eqn. 6.1.4, to be written
as
 xx 
1
E

 xx    yy   zz  
 yy
1

  yy    xx   zz 
E

 zz
1
E

  zz    xx   yy  
Stress-Strain Relations (6.1.8)
1 
 xy   xy
E
1 
 xz   xz
E
1 
 yz   yz
E
This is known as Hooke’s Law. These equations can be solved for the stresses to get

Section 6.1
 xx 
E
(1   )(1  2 )

(1   ) xx    yy   zz  
 yy 
E
(1   )(1  2 )

(1   ) yy    xx   zz  
 zz 
E
(1   )(1  2 )

(1   ) zz    xx   yy  
Stress-Strain Relations (6.1.9)
E
 xy   xy
1 
E
 xz   xz
1 
E
 yz   yz
1 
Values of E and  for a number of materials are given in Table 6.1.1 below (see also
Table 5.2.2).
Material E (GPa) 
Grey Cast Iron 100 0.29
A316 Stainless Steel 196 0.3
A5 Aluminium 68 0.33
Bronze 130 0.34
Plexiglass 2.9 0.4
Rubber 0.001-2 0.4-0.49
Concrete 23-30 0.2
Granite 53-60 0.27
Wood (pinewood)
fibre direction 17 0.45
transverse direction 1 0.79
Table 6.1.1: Young’s Modulus E and Poisson’s Ratio ν for a selection of materials at
20oC
Volume Change
Recall that the volume change in a material undergoing small strains is given by the sum
of the normal strains (see Section 4.3). From Hooke’s law, normal stresses cause normal
strain and shear stresses cause shear strain. It follows that normal stresses produce
volume changes and shear stresses produce distortion (change in shape), but no volume
change.
6.1.3 Two Dimensional Elasticity

The above three-dimensional stress-strain relations reduce in the case of a two-
dimensional stress state or a two-dimensional strain state.

Section 6.1
Plane Stress
In plane stress (see Section 3.5),  xz   yz   zz  0 , Fig. 6.1.5, so the stress-strain

relations reduce to
 xx 
1
E

 xx   yy 
 yy
1

  yy   xx
E

1 
 xy   xy
E
 xx 
E
1  2

 xx   yy 
 yy 
E
1  2

 xx   yy 
E
 xy   xy
1 
Stress-Strain Relations (Plane Stress) (6.1.10)
with

 zz  
E
 xx 
  yy ,  xz   yz  0
(6.1.11)
 zz   xz   yz  0
x
z
Figure 6.1.5: Plane stress
Note that the  zz strain is not zero. Physically,  zz corresponds to a change in thickness
of the material perpendicular to the direction of loading.
Plane Strain
In plane strain (see Section 4.2),  xz   yz   zz  0 , Fig. 6.1.6, and the stress-strain
relations reduce to

Section 6.1
1 
 xx 
E

(1   ) xx   yy 
1 
 yy 
E

  xx  (1   ) yy 
1 
 xy   xy
E
 xx 
E
(1   )(1  2 )
 
(1   ) xx   yy
 yy 
E
(1   )(1  2 )
 
(1   ) yy   xx
E
 xy   xy
1 
Stress-Strain Relations (Plane Strain) (6.1.12)
with
 zz   xz   yz  0
(6.1.13)
 zz    xx   yy ,  xz   yz  0
Again, note here that the stress component  zz is not zero. Physically, this stress
corresponds to the forces preventing movement in the z direction.
 zz
Figure 6.1.6 Plane strain - a thick component constrained in one direction
Similar Solutions
The expressions for plane stress and plane strain are very similar. For example, the plane
strain constitutive law 6.1.12 can be derived from the corresponding plane stress
expressions 6.1.10 by making the substitutions
E 
E ,   (6.1.14)
1   2
1  
in 6.1.10 and then dropping the primes. The plane stress expressions can be derived from
the plane strain expressions by making the substitutions
1  2  
E  E ,   (6.1.15)
1    2
1  

Section 6.1
in 6.1.12 and then dropping the primes. Thus, if one solves a plane stress problem, one
has automatically solved the corresponding plane strain problem, and vice versa.
6.1.4 Problems
1. Consider a very thin sheet of material subjected to a normal pressure p on one of its
large surfaces. It is fixed along its edges. This is an example of a plate problem, an
important branch of elasticity with applications to boat hulls, aircraft fuselage, etc.
(a) write out the complete three dimensional stress-strain relations for this case (both
cases, stress in terms of strain, strain in terms of stress) - simplify the relations
using the fact that the sheet is thin, the stress boundary condition on the large face
and the coordinate system shown (just substitute in appropriate values for  xz ,  yz
and  zz )
(b) assuming that the through thickness change in the sheet can be neglected, show
that
p    xx   yy 
y
p
x
z
2. A strain gauge at a certain point on the surface of a thin A5 Aluminium component

(loaded in-plane) records strains of  xx  60 μm,  yy  30 μm,  xy  15 μm .
Determine the principal stresses. (See Table 6.1.1 for the material properties.)
3. Use the stress-strain relations to prove that, for a linear elastic solid,
2 xy 2 xy

 xx   yy  xx   yy
and, indeed,
2 xz 2 xz 2 yz 2 yz
 , 
 xx   zz  xx   zz  yy   zz  yy   zz
Note: from Eqns. 3.5.4 and 4.2.4, these show that the principal axes of stress and
strain coincide for an isotropic elastic material
4. Consider the case of hydrostatic pressure in a linearly elastic solid:

 p 0 0 
  
 ij   0  p 0 
 0 0  p 
as might occur, for example, when a spherical component is surrounded by a fluid
under high pressure, as illustrated in the figure below. Show that the volumetric strain
is equal to

Section 6.1
31  2 
p ,
E
so that the Bulk Modulus, Eqn. 5.2.9, is
E
K
3 1  2 
5. Consider again Problem 2 from §3.5.7.

(a) Assuming the material to be linearly elastic, what are the strains? Draw a second
material element (superimposed on the one shown below) to show the deformed
shape of the square element – assume the displacement of the box-centre to be
zero and that there is no rotation. Note how the free surface moves, even though
there is no stress acting on it.
(b) What are the principal strains  1 and  2 ? You will see that the principal
directions of stress and strain coincide (see Problem 3) – the largest normal stress
and strain occur in the same direction.
y
 xx  xx
6. A thin linear elastic plate is subjected to a uniform compressive stress  0 as shown

below. Show that the slope of the plate diagonal shown after deformation is given by
b  1   0 / E 
tan        
a  1 0 / E 
What is the magnitude of  for a steel plate ( E  210GPa ,   0.3 ) of dimensions
20  20 cm 2 with  0  1 MPa ?
0 0


Section 6.2
6.2 Homogeneous Problems in Linear Elasticity

A homogeneous stress (strain) field is one where the stress (strain) is the same at all
points in the material. Homogeneous conditions will arise when the geometry is simple
and the loading is simple.
6.2.1 Elastic Rectangular Cuboids

Hooke’s Law, Eqns. 6.1.8 or 6.1.9, can be used to solve problems involving
homogeneous stress and deformation. Hoooke’s law is 6 equations in 12 unknowns (6
stresses and 6 strains). If some of these unknowns are given, the rest can be found from
the relations.
Example
Consider the block of linear elastic material shown in Fig. 6.2.1. It is subjected to an
equi-biaxial stress of  xx   yy    0 .
Since this is an isotropic elastic material, the shears stresses and strains will be all zero for
such a loading. One thus need only consider the three normal stresses and strains.
There are now 3 equations (the first 3 of Eqns. 6.1.8 or 6.1.9) in 6 unknowns. One thus
needs to know three of the normal stresses and/or strains to find a solution. From the
loading, one knows that  xx   and  yy   . The third piece of information comes
from noting that the surfaces parallel to the x  y plane are free surfaces and so  zz  0 .
From Eqn. 6.1.8 then, the strains are
 
 xx   yy  (1   ) ,  zz  2 ,  xy   xz   yz  0
E E
As expected,  xx   yy and  zz  0 .
y 
 
x
z 
Figure 6.2.1: A block of linear elastic material subjected to an equi-biaxial stress

Section 6.2
6.2.2 Problems
1. A block of isotropic linear elastic material is subjected to a compressive normal stress
 o over two opposing faces. The material is constrained (prevented from moving) in
one of the direction normal to these faces. The other faces are free.
(a) What are the stresses and strains in the block, in terms of  o ,  , E ?
(b) Calculate three maximum shear stresses, one for each plane (parallel to the faces
of the block). Which of these is the overall maximum shear stress acting in the
block?
2. Repeat problem 1a, only with the free faces now fixed also.

7 Applications of
Elasticity
The linear elastic model was introduced in the previous chapter and some elementary
problems involving elastic materials were solved there (in particular in section 6.2). In
this Chapter, five important, practical, theories are presented concerning elastic materials;
they all have specific geometries and are subjected to particular types of load. In §7.1,
the geometry is that of a long slender bar and the load is one which acts along the length
of the bar; in §7.2, the geometry is that of a long slender circular bar and the load is one
which twists the bar; in §7.3 the geometry is that of a thin-walled cylindrical or spherical
component, and the load is normal to these walls; in §7.4 the geometry is that of a long
and slender beam, and the load is transverse to the beam length. Finally, in §7.5, the
geometry is a column, fixed at one end and loaded at the other so that it deflects. These
five particular situations allow for simplifications (or approximations) to be made to the
full three-dimensional linear elastic stress-strain relations; this allows one to write down
simple expressions for the stress and strain and so solve some important practical
problems analytically.
165
166
Section 7.1
7.1 One Dimensional Axial Deformations

In this section, a specific simple geometry is considered, that of a long and thin straight
component loaded in such a way that it deforms in the axial direction only. The x-axis is
taken as the longitudinal axis, with the cross-section lying in the x  y plane, Fig. 7.1.1.
x
z
Figure 7.1.1: A slender straight component; (a) longitudinal axis, (b) cross-section
7.1.1 Basic relations for Axial Deformations

Any static analysis of a structural component involves the following three considerations:
(1) constitutive response
(2) kinematics
(3) equilibrium
In this Chapter, it is taken for (1) that the material responds as an isotropic linear elastic
solid. It is assumed that the only significant stresses and strains occur in the axial
direction, and so the stress-strain relations 6.1.8-9 reduce to the one-dimensional equation
 xx  E xx or, dropping the subscripts,
  E (7.1.1)
Kinematics (2), the study of deformation, was the subject of Chapter 4. In the theory
developed here, known as axial deformation, it is assumed that the axis of the
component remains straight and that cross-sections that are initially perpendicular to the
axis remain perpendicular after deformation. This implies that, although the strain might
vary along the axis, it remains constant over any cross section. The axial strain occurring
over any section is defined by Eqn. 4.1.2,
L  L0
 (7.1.2)
L0
This is illustrated in Fig. 7.1.2, which shows a (shaded) region undergoing a compressive
(negative) strain.
Recall that individual particles/points undergo displacements whereas regions/line-

elements undergo strain. In Fig. 7.1.2, the particle originally at A has undergone a
displacement u ( A) whereas the particle originally at B has undergone a displacement
u (B) . From Fig. 7.1.2, another way of expressing the strain in the shaded region is (see
Eqn. 4.1.3)

Section 7.1
u ( B)  u ( A)
 (7.1.3)
L0
L0
(a) A B x
u(A)
u(B)
(b)   x
L
Figure 7.1.2: axial strain; (a) before deformation, (b) after deformation
Both displacements u ( A) and u (B) of Fig. 7.1.2 are positive, since the particles displace
in the positive x direction – if they moved to the left, for consistency, one would say they
underwent negative displacements. Further, positive stresses are as shown in Fig. 7.1.3a
and negative stresses are as shown in Fig. 7.1.3b. From Eqn. 7.1.1, a positive stress
implies a positive strain (lengthening) and a compressive stress implies a negative strain
(contracting)
 0  0  0  0
x x
(a) (b)
Figure 7.1.3: Stresses arising in the slender component; (a) positive (tensile) stress,
(b) negative (compressive) stress
Equilibrium, (3), will be considered in the individual examples below.
Note that, in the previous Chapter, problems were solved using only the stress-strain law
(1). Kinematics (2) and equilibrium (3) were not considered, the reason being the
problems were so simple, with uniform (homogeneous) stress and strain (as indeed also in
the first example which follows). Whenever more complex problems are encountered,
with non-uniform stress and strains, (3) and perhaps (2) need to be considered to solve for
the stress and strain.
7.1.2 Structures with Uniform Members

A uniform axial member is one with cross-section A and modulus E constant along its
length, and loaded with axial forces at its ends only.

Section 7.1
Example
Consider the bar of initial length L shown in Fig. 7.1.4, subjected to equal and opposite
end-forces F. The free-body (equilibrium) diagram of a section of the bar shown in Fig.
7.1.4b shows that the internal force is also F everywhere along the bar. The stress is thus
everywhere   F / A and the strain is everywhere
F
 (7.1.4)
EA
and, from Eq. 7.1.2, the bar extends in length by an amount
FL
 (7.1.5)
EA
Note that although the force acting on the left-hand end is negative (acting in the  x
direction), the stress there is positive (see Fig. 7.1.3).
F F
(a)    
A B C D
F F
(b)
Figure 7.1.4: A uniform axial member; (a) subjected to axial forces F, (b) free-body
diagram
Displacements need to be calculated relative to some datum displacement1. For example,

suppose that the displacement at the centre of the bar is zero, u ( B )  0 , Fig. 7.1.4. Then,
from Eqn. 7.1.3,
F L
u (C )  u ( B)   (C  B) 
EA 4
F L
u ( D)  u ( B)   ( D  B)  (7.1.6)
EA 2
F L
u ( A)  u ( B)   ( A  B)  
EA 2
■
1
which is another way of saying that one can translate the bar left or right as a rigid body without affecting
the stress or strain – but it does affect the displacements

Section 7.1
Example
Consider the two-element structure shown in Fig. 7.1.5. The first element is built-in at
end A, is of length L1 , cross-sectional area A1 and Young’s modulus E1 . The second
element is attached at B and has properties L2 , A2 , E 2 . External loads F and P are
applied at B and C as shown. An unknown reaction force R acts at A. This can be
determined from the force equilibrium equation for the structure:
RF P 0 (7.1.7)
As usual, the reaction is first assumed to act in the positive (x) direction. With R known,
the stress  (1) in the first element can be evaluated using the free-body diagram 7.1.5b,
and  (1) using Fig. 7.1.5c:
PF P
 (1)  ,  ( 2)  (7.1.8)
A1 A2
and so the strain is
PF P
 (1)  ,  ( 2)  (7.1.9)
E1 A1 E 2 A2
Note that the stress and strain are discontinuous at B 2.
A L1 L2
R F P
(a)   
B C
R  PF PF
(b)
P P
(c) 
Figure 7.1.5: A two-element structure (a) subjected to axial forces F and P, (b,c)
free-body diagrams
For each element, the total elongations  i are
2
this result, which can be viewed as a violation of equilibrium at B, is a result of the one-dimensional
approximation of what is really a two-dimensional problem

Section 7.1
 1  u ( B)  u ( A) 
P  F L1
E1 A1
(7.1.10)
PL2
 2  u (C )  u ( B) 
E 2 A2
If P  F , then  1  0 as expected, with R  0 and   0 .
Thus far, the stress and strain (and elongations) have been obtained. If one wants to
evaluate the displacements, then one needs to ensure that the strains in each of the two
elements are compatible, that is, that the elements fit together after deformation just like
they did before deformation. In this example, the displacements at B and C are
u ( B )  u ( A)   1 , u (C )  u ( B)   2 (7.1.11)
A compatibility condition, bringing together the separate relations in 7.1.11, is then
P  F L1 PL2
u (C )  u ( A)   (7.1.12)
E1 A1 E 2 A2
ensuring that u (B) is unique. As in the previous example, the displacements can now be
calculated if the displacement at any one (datum) point is known. Indeed, it is known that
u ( A)  0 .
■
Example
Consider next the similar situation shown in Fig. 7.1.6. Here, both ends of the two-
element structure are built-in and there is only one applied force, F, at B. There are now
two reaction forces, at ends A and C, but there is only one equilibrium equation to
determine them:
R A  F  RC  0 (7.1.13)
Any structure for which there are more unknowns than equations of equilibrium, so that
the stresses cannot be determined without considering the deformation of the structure, is
called a statically indeterminate structure3.
3
See the end of §2.3.3

Section 7.1
A L1 L2 C
RA F RC
(a)   
B
 RA  F  RC F  RC
(b) 
RC RC
(c) 
Figure 7.1.6: A two-element structure built-in and both ends; (a) subjected to an
axial force F, (b,c) free-body diagrams
In terms of the unknown reactions, the strains are
 (1) RA F  RC  ( 2) RC
 (1)    ,  ( 2)   (7.1.14)
E1 E1 A1 E1 A1 E2 E 2 A2
and, for each element, the total elongations are
R A L1 RC L2
1  , 2  (7.1.15)
E1 A1 E 2 A2
Finally, compatibility of both elements implies that the total elongation  1   2  0 .

Using this relation with Eqn. 7.1.13-14 then gives
L2 E1 A1 L1 E 2 A2
RA  F , RC   F (7.1.16)
L1 E 2 A2  L2 E1 A1 L1 E 2 A2  L2 E1 A1
The displacements can now be evaluated, for example,
1
u ( B)   F (7.1.17)
E1 A1 / L1  E 2 A2 / L2
so that a positive F displaces B to the right and a negative F displaces B to the left.
■
Note the general solution procedure in this last example, known as the basic force
method:
Equilibrium + Compatibility of Strain in terms of unknown Forces

 Solve equations for unknown Forces

Section 7.1
The Stiffness Method
The stiffness method (also known as the displacement method) is a slight modification
of the above solution procedure, where the final equations to be solved involve known
forces and unknown displacements only:
Equilibrium in terms of Displacement

 Solve equations for unknown Displacements
If one deals in displacements, one does not need to ensure compatibility (it will
automatically be satisfied); compatibility only needs to be considered when dealing in
strains (as in the previous example).
Example (The Stiffness Method)
Consider a series of three bars of cross-sectional areas A1 , A2 , A3 , Young’s moduli

E1 , E 2 , E3 and lengths L1 , L2 , L3 , Fig. 7.1.7. The first and third bars are built-in at points
A and D, bars one and two meet at B and bars two and three meet at C. Forces PB and PC
act at B and C respectively.
The force is constant in each bar, and for each bar there is a relation between the force Fi ,
and elongation,  i , Eqn. 7.1.5:
Ai E i
Fi  k i  i where ki  (7.1.18)
Li
Here, k i is the effective stiffness of each bar. The elongations are related to the
displacements,  1  u B  u A etc., so that, with u A  u D  0 ,
F1  k1u B , F2  k 2 u C  u B , F3   k 3u C (7.1.19)
There are two degrees of freedom in this problem, that is, two nodes are free to move.
One therefore needs two equilibrium equations. One could use any two of
 F1  PB  PC  F3  0,  F1  PB  F2  0,  F2  PC  F3  0 (7.1.20)
In the stiffness method, one uses the second and third of these; the second is the “node B”
equation and the third is the “node C” equation. Substituting Eqns. 7.1.19 into 7.1.20
leads to the system of two equations
 k1  k 2 u B  k 2 u C   PB
(7.1.21)
 k 2 u B  k 2  k 3 u C   PC
which can be solved for the two unknown nodal displacements

Section 7.1
A B C D
PB PC
(a)    
F1 PB F2
(b)  
F2 PC F3
(c)  
F1 F1 F3 F3
(d)  
Figure 7.1.7: three bars in series; (a) subjected to external loads, (b,c,d) free-body
diagrams
Note that it was not necessary to evaluate the reactions to obtain a solution. Once the
forces have been found, the reactions can be found using the free-body diagram of Fig.
7.1.7d.
The stiffness method is a very systematic procedure. It can be used to solve for structures
with many elements, with the two equations 7.1.21 replaced by a large system of
equations which can be solved numerically using a computer.
7.1.3 Structures with Non-uniform Members

Consider the structure shown in Fig. 7.1.8, an axial bar consisting of two separate
components bonded together. The components have Young’s moduli E1 , E 2 and cross-
sectional areas A1 , A2 . The bar is subjected to equal and opposite forces F as shown, in
such a way that axial deformations occur, that is, the cross-sections remain perpendicular
to the x axis throughout the deformation.
Since there are only axial deformations, the strain is constant over a cross-section.
However, the stress is not uniform, with  1  E1 and  2  E 2  ; on any cross-section,
the stress is higher in the stiffer component. The resultant force acting on each
component is F1  E1 A1 and F2  E 2 A2  . Since F1  F2  F , the total elongation is
FL
 (7.1.22)
E1 A1  E 2 A2

Section 7.1
L F1 F1
F2 F2
F E1 F
 E2 
A B
Figure 7.1.8: A bar consisting of two separate materials bonded together
7.1.4 Resultant Force and Moment

Consider the force and moments acting over any cross-section, Fig. 7.1.9. The resultant
force is the integral of the stress times elemental area over the cross section, Eqn. 3.1.2,
F    dA (7.1.23)
A
There are two moments; the moment M y about the y axis and M z about the z axis,
M y   z dA, M z    y dA (7.1.24)
A A
Positive moments are defined through the right hand rule, i.e. with the thumb of the
right hand pointing in the positive y direction, the closing of the fingers indicates the
positive M y ; the negative sign in Eqn. 7.1.24b is due to the fact that a positive stress with
y  0 would lead to a negative moment M z .
y y
dA My
z
 
y
z F z
Mz
(a) (b)
Figure 7.1.9: Resultants on a cross-section; (a) resultant force, (b) resultant

moments
Consider now the case where the stress is constant over a cross-section. Since it is
assumed that the strain is constant over the cross-section, from Eqn. 7.1.1 this will occur
when the Young’s modulus is constant. In that case, Eqns. 7.1.23-24 can be re-written as
F  A, M y    zdA, M z    ydA (7.1.25)

A A

Section 7.1
The quantities  zdA and 

A A
ydA are the first moments of area about, respectively,
the y and z axes. These are equal to z A and yA , where ( y , z ) are the coordinates of
the centroid of the section (see Eqn. 3.2.2). Taking the x axis to run through the
centroid, y  z  0 results in M y  M z  0 . Thus, a resultant axial force which
acts through the centroid of the cross-section ensures that there is no
moment/rotation of that cross-section, the main assumption of this section.
For the non-uniform member of Fig. 7.1.8, since the resultant of a constant stress over an
area is a force acting through the centroid of that area, the forces F1 , F2 act through the
centroids of the respective areas A1 , A2 . The precise location of the total resultant force F
can be determined by taking the moments of the forces F1 , F2 about the y and z axes, and
equating this to the moment of the force F about these axes.
7.1.5 Problems
1. Consider the rigid beam supported by two deformable bars shown below. The bars
have properties L1 , A1 and L2 , A1 and have the same Young’s modulus E. They are
separated by a distance L. The beam supports an arbitrary load at position x, as
shown. What is x if the beam is to remain horizontal after deformation.
L1 L2

Section 7.2
7.2 Torsion
In this section, the geometry to be considered is that of a long slender circular bar and the
load is one which twists the bar. Such problems are important in the analysis of twisting
components, for example lug wrenches and transmission shafts.
7.2.1 Basic relations for Torsion of Circular Members

The theory of torsion presented here concerns torques1 which twist the members but
which do not induce any warping, that is, cross sections which are perpendicular to the
axis of the member remain so after twisting. Further, radial lines remain straight and
radial as the cross-section rotates – they merely rotate with the section.
For example, consider the member shown in Fig. 7.2.1, built-in at one end and subject to
a torque T at the other. The x axis is drawn along its axis. The torque shown is positive,
following the right-hand rule (see §7.1.4). The member twists under the action of the
torque and the radial plane ABCD moves to ABC D .
B C

 x
A C D
T
Figure 7.2.1: A cylindrical member under the action of a torque
Whereas in the last section the measure of deformation was elongation of the axial
members, here an appropriate measure is the amount by which the member twists, the
rotation angle  . The rotation angle will vary along the member – the sign convention is
that  is positive in the same direction as positive T as indicated by the arrow in Fig.
7.2.1. Further, whereas the measure of strain used in the previous section was the normal
strain  xx , here it will be the engineering shear strain  xy (twice the tensorial shear strain
 xy ). A relationship between  (dropping the subscripts) and  will next be established.
As the line BC deforms into BC  , Fig. 7.2.1, it undergoes an angle change  . As

defined in §4.1.2, the shear strain  is the change in the original right angle formed by
BC and a tangent at B (indicated by the dotted line – this is the y axis to be used in  xy ).
If  is small, then
CC  R ( L)
    tan    (7.2.1)
BC L
1
the term torque is usually used instead of moment in the context of twisting shafts such as those
considered in this section

Section 7.2
where L is the length, R the radius of the member and  (L) means the magnitude of
 at L. Note that the strain is constant along the length of the member although  is
not. Considering a general cross-section within the member, as in Fig. 7.2.2, one has
R ( x)
   (7.2.2)
x
B E
 
E
A F
T
Figure 7.2.2: A section of a twisting cylindrical member
The shear strain at an arbitrary radial location r, 0  r  R , is
r ( x )
 (r )  (7.2.3)
x
showing that the shear strain varies from zero at the centre of the shaft to a
maximum R ( L) / L  R ( x) / x  on the outer surface of the shaft.
Considering a free-body diagram of any portion of the shaft of Fig. 7.2.1, a torque T
acts on all cross-sections. This torque must equal the resultant of the shear stresses
acting over the section, as schematically illustrated in Fig. 7.2.3a. The elemental
force acting over an element with sides dr and rd is dA  rdrd , Fig. 7.2.3b,
and so the resultant moment about r  0 is
2 R R
T   r  (r )drd  2  r  (r )dr
2 2
(7.2.4)
0 0 0
Hooke’s law is
  G (7.2.5)
where G is the shear modulus (the  of Eqn. 6.1.5). But  / r is a constant and so
therefore also is  / r (provided G is) and Eqn. 7.2.4 can be re-written as
 (r )  R
  (r ) J
T 2  r dr  
3
(7.2.6)
r  0  r
The quantity in square brackets is called the polar moment of inertia of the cross-
section (also called the polar second moment of area) and is denoted by J. For this
circular cross-section it is given by

Section 7.2
R
R 4 D 4
J  2  r 3 dr   (7.2.7)
0
2 32
where D is the diameter. In general, for a cross-section of arbitrary shape,
J   r 2 dA Polar Moment of Area (7.2.8)

A
where dA is an element of area and the integration is over the complete cross-section.
  (r )
dA  rdrd
r
(a) (b)
Figure 7.2.3: Shear stresses acting over a cross-section; (a) shear stress, (b)
moment for an elemental area
From Eqn. 7.2.6, the shear stress at any radial location is given by
rT
 (r )  (7.2.9)
J
From Eqn. 7.2.1, 7.2.5 and 7.2.9, the angle of twist at the end of the member – or the
twist at one end relative to that at the other end – is
TL
 (7.2.10)
GJ
Example
Consider the problem shown in Fig.7.2.4, two torsion members of lengths L1 , L2 ,

diameters d1 , d 2 and shear moduli G1 ,G2 , built-in at A and subjected to torques TB and
TC . Equilibrium of moments can be used to determine the unknown torques acting in
each member:
 T1  TB  TC  0,  T2  TC  0 (7.2.11)
so that T1  TB  TC and T2  TC .

Section 7.2
TB TC

(a) x
D
A L1 B L2 C
(b) T1
(c) T2
Figure 7.2.4: A structure consisting of two torsion members; (a) subjected to torques
TB and TC , (b,c) free-body diagrams
The shear stresses in each member are therefore
r TB  TC  rTC
1  , 2  (7.2.12)
J1 J2
where J 1  d14 / 32 and J 2  d 24 / 32 .
From Eqn. 7.2.10, the angle of twist at B is given by  B  T1 L1 / G1 J 1 . The angle of twist
at C is then
T2 L2
C   B (7.2.13)
G2 J 2
■
Statically indeterminate problems can be solved using methods analogous to those used in
the section 7.1 for uniaxial members.
Example
Consider the structure in Fig. 7.2.5, similar to that in Fig. 7.2.4 only now both ends are
built-in and there is only a single applied torque, TB .

Section 7.2
TB

(a) x
D
A L1 B L2 C
(b) T1
TB T2
(c) T1 T1 T2 T2
Figure 7.2.5: A structure consisting of two torsion members; (a) subjected to a

Torque TB , (b) free-body diagram, (c) separate elements
Referring to the free-body diagram of Fig. 7.2.5b, there is only one equation of
equilibrium with which to determine the two unknown member torques:
 T1  TB  T2  0 (7.2.14)
and so the deformation of the structure needs to be considered. A systematic way of

dealing with this situation is to consider each element separately, as in Fig. 7.2.5c. The
twist in each element is
T1 L1 TL
1  , 2  2 2 (7.2.15)
G1 J 1 G2 J 2
The total twist is zero and so 1   2  0 which, with Eqn. 7.2.14, can be solved to obtain
L2 G1 J 1 L1G2 J 2
T1   TB , T2   TB (7.2.16)
L1G2 J 2  L2 G1 J 1 L1G2 J 2  L2 G1 J 1
The rotation at B can now be determined,  B  1   2 .

■
7.2.2 Stress Distribution in Torsion Members

The shear stress in Eqn. 7.2.9 is acting over a cross-section of a torsion member.
From the symmetry of the stress, it follows that shear stresses act also along the
length of the member, as illustrated to the left of Fig. 7.2.6. Shear stresses do not act
on the surface of the element shown, as it is a free surface.

Section 7.2
Any element of material not aligned with the axis of the cylinder will undergo a
complex stress state, as shown to the right of Fig. 7.2.6. The stresses acting on an
element are given by the stress transformation equations, Eqns. 3.4.8:
 xx   sin 2 ,  yy   sin 2 ,  xy   cos 2 (7.2.17)
 yy
 xx

x

  xy x
 0
y y
T
Figure 7.2.6: Stress distribution in a torsion member
From Eqns. 3.5.4-5, the maximum normal (principal) stresses arise on planes at   45 o
and are  1   and  2   . Thus the maximum tensile stress in the member occurs at
45 o to the axis and arises at the surface. The maximum shear stress is simply  , with
  0.
7.2.3 Problems
1. A shaft of length L and built-in at both ends is subjected to two external torques, T at
A and 2T at B, as shown below. The shaft is of diameter d and shear modulus G.
Determine the maximum (absolute value of) shear stress in the shaft and determine
the angle of twist at B.
A B
T B 2T
L/4 L/4 L/2

Section 7.4
7.4 The Elementary Beam Theory

In this section, problems involving long and slender beams are addressed. As with
pressure vessels, the geometry of the beam, and the specific type of loading which will be
considered, allows for approximations to be made to the full three-dimensional linear
elastic stress-strain relations.
The beam theory is used in the design and analysis of a wide range of structures, from
buildings to bridges to the load bearing bones of the human body.
7.4.1 The Beam

The term beam has a very specific meaning in engineering mechanics: it is a component
that is designed to support transverse loads, that is, loads that act perpendicular to the
longitudinal axis of the beam, Fig. 7.4.1. The beam supports the load by bending only.
Other mechanisms, for example twisting of the beam, are not allowed for in this theory.
applied force
applied pressure
cross section
roller support pin support
Figure 7.4.1: A supported beam loaded by a force and a distribution of pressure
It is convenient to show a two-dimensional cross-section of the three-dimensional beam

together with the beam cross section, as in Fig. 7.4.1. The beam can be supported in
various ways, for example by roller supports or pin supports (see section 2.3.3). The
cross section of this beam happens to be rectangular but it can be any of many possible
shapes.
It will assumed that the beam has a longitudinal plane of symmetry, with the cross
section symmetric about this plane, as shown in Fig. 7.4.2. Further, it will be assumed
that the loading and supports are also symmetric about this plane. With these conditions,
the beam has no tendency to twist and will undergo bending only1.
longitudinal plane
of symmetry
Figure 7.4.2: The longitudinal plane of symmetry of a beam
1
certain very special cases, where there is not a plane of symmetry for geometry and/or loading, can lead
also to bending with no twist, but these are not considered here

Section 7.4
Imagine now that the beam consists of many fibres aligned longitudinally, as in Fig.
7.4.3. When the beam is bent by the action of downward transverse loads, the fibres near
the top of the beam contract in length whereas the fibres near the bottom of the beam
extend. Somewhere in between, there will be a plane where the fibres do not change
length. This is called the neutral surface. The intersection of the longitudinal plane of
symmetry and the neutral surface is called the axis of the beam, and the deformed axis is
called the deflection curve.
y
fibres contracting
z neutral surface
fibres extending
Figure 7.4.3: the neutral surface of a beam
A conventional coordinate system is attached to the beam in Fig. 7.4.3. The x axis
coincides with the (longitudinal) axis of the beam, the y axis is in the transverse direction
and the longitudinal plane of symmetry is in the x  y plane, also called the plane of
bending.
7.4.2 Moments and Forces in a Beam

Normal and shear stresses act over any cross section of a beam, as shown in Fig. 7.4.4.
The normal and shear stresses acting on each side of the cross section are equal and
opposite for equilibrium, Fig. 7.4.4b. The normal stresses  will vary over a section
during bending. Referring again to Fig. 7.4.3, over one part of the section the stress will
be tensile, leading to extension of material fibres, whereas over the other part the stresses
will be compressive, leading to contraction of material fibres. This distribution of normal
stress results in a moment M acting on the section, as illustrated in Fig. 7.4.4c. Similarly,
shear stresses  act over a section and these result in a shear force V.
The beams of Fig. 7.4.3 and Fig. 7.4.4 show the normal stress and deflection one would
expect when a beam bends downward. There are situations when parts of a beam bend
upwards, and in these cases the signs of the normal stresses will be opposite to those
shown in Fig. 7.4.4. However, the moments (and shear forces) shown in Fig. 7.4.4 will
be regarded as positive. This sign convention to be used is shown in Fig. 7.4.5.

Section 7.4
cross-section in
(a ) beam
( b)  

V
(c) M M
Figure 7.4.4: stresses and moments acting over a cross-section of a beam
V M
positive positive
bending shearing
M
V
(a ) ( b) (c)
Figure 7.4.5: sign convention for moments and shear forces
Note that the sign convention for the shear stress in the beam theory conflicts with the
sign convention for shear stress used in the rest of mechanics, introduced in Chapter 3.
This is shown in Fig. 7.4.6.
y 
x 
Mechanics Beam Theory
(in general)
Figure 7.4.6: sign convention for shear stress in beam theory
The moments and forces acting within a beam can in many simple problems be evaluated
from equilibrium considerations alone. Some examples are given next.

Section 7.4
Example 1
Consider the simply supported beam in Fig. 7.4.7. From the loading, one would expect
the beam to deflect something like as indicated by the deflection curve drawn. The
reaction at the roller support, end A, and the vertical reaction at the pin support2, end B,
can be evaluated from the equations of equilibrium, Eqns. 2.3.3:
R Ay  P / 3, RBy  2 P / 3 (7.4.1)
2l / 3 deflection
P curve
A B
l
RAy RBy
Figure 7.4.7: a simply supported beam
The moments and forces acting within the beam can be evaluated by taking free-body
diagrams of sections of the beam. There are clearly two distinct regions in this beam, to
the left and right of the load. Fig. 7.4.8a shows an arbitrary portion of beam representing
the left-hand side. A coordinate system has been introduced, with x measured from A.3
An unknown moment M and shear force V act at the end. A positive moment and force
have been drawn in Fig. 7.4.8a. From the equilibrium equations, one finds that the shear
force is constant but that the moment varies linearly along the beam:
P P 2l
V  , M  x (0 x ) (7.4.2)
3 3 3
2l / 3 P
A M A M
P/3 V P/3 V
x x
(a ) ( b)
Figure 7.4.8: free body diagrams of sections of a beam
2
the horizontal reaction at the pin is zero since there are no applied forces in this direction; the beam theory
does not consider such types of load
3
the coordinate x can be measured from any point in the beam; in this example it is convenient to measure
it from point A

Section 7.4
Cutting the beam to the right of the load, Fig. 7.4.8b, leads to
2P 2P 2l
V  , M  l  x  (  xl ) (7.4.3)
3 3 3
The shear force is negative, so acts in the direction opposite to that initially assumed in
Fig. 7.4.8b.
The results of the analysis can be displayed in what are known as a shear force diagram
and a bending moment diagram, Fig. 7.4.9. Note that there is a “jump” in the shear
force at x  2l / 3 equal to the applied force, and in this example the bending moment is
everywhere positive.
V M
2Pl
P
9
3
2P
 2l / 3
3
l l
(a ) ( b)
Figure 7.4.9: results of analysis; (a) shear force diagram, (b) bending moment
diagram
■
Example 2
Fig. 7.4.10 shows a cantilever, that is, a beam supported by clamping one end (refer to
Fig. 2.3.8), and loaded by a force at its mid-point and a (negative) moment at its end.
5 kN
3m 3m
VA 4 kNm
MA
Figure 7.4.10: a cantilevered beam loaded by a force and moment
Again, positive unknown reactions M A and V A are considered at the support A. From
the equilibrium equations, one finds that
M A  11 kNm, V A  5 kN (7.4.4)

Section 7.4
As in the previous example, there are two distinct regions along the beam, to the left and
to the right of the applied concentrated force. Again, a coordinate x is introduced and the
beam is sectioned as in Fig. 7.4.11. The unknown moment M and shear force V can then
be evaluated from the equilibrium equations:
V  5 kN, M  11  5 x kNm 0  x  3 (7.4.5)

V  0, M  4 kNm 3  x  6
3m
5 kN
11  5 M 11  5 M
A V A V
x x
(a ) ( b)
The results are summarized in the shear force and bending moment diagrams of Fig.
7.4.12.
V M
11
4
5
6m 3m
(a ) ( b)
diagram
In this example the beam experiences negative bending moment over most of its length.
■
Example 3
Fig. 7.4.13 shows a simply supported beam subjected to a distributed load (force per unit
length). The load is uniformly distributed over half the length of the beam, with a
triangular distribution over the remainder.

Section 7.4
40 N/m
A B C
6m 6m
Figure 7.4.13: a beam subjected to a distributed load
The unknown reactions can be determined by replacing the distributed load with statically
equivalent forces as in Fig. 7.4.14 (refer to §3.1.2). The equilibrium equations then give
R A  220 N, RC  140 N (7.4.6)
240 N 120 N
3m 3m 2m 4m
RA RC
Figure 7.4.14: equivalent forces acting on the beam of Fig. 7.4.13
Referring again to Fig. 7.4.13, there are two distinct regions in the beam, that under the
uniform load and that under the triangular distribution of load. The first case is
considered in Fig. 7.4.15.
40 N/m
x V
220
Figure 7.4.15: free body diagram of a section of a beam
The equilibrium equations give
V  220  40 x , M  220 x  20 x 2 0  x  6 (7.4.7)
The region beneath the triangular distribution is shown in Fig. 7.4.16. Two possible
approaches are illustrated: in Fig. 7.4.16a, the free body diagram consists of the complete

Section 7.4
length of beam to the left of the cross-section under consideration; in Fig. 7.4.16b, only
the portion to the right is considered, with distance measured from the right hand end, as
12  x . The problem is easier to solve using the second option. From Fig. 7.4.16b then,
with the equilibrium equations, one finds that
V  140  10(12  x) 2 / 3 , M  140(12  x)  10(12  x) 3 / 9 6  x  12 (7.4.8)
M M V
x
V 12  x
220 140
(a ) ( b)
The results are summarized in the shear force and bending moment diagrams of Fig.
7.4.17.
V M
220 600
 140
6m 6m 6m 6m
(a ) ( b)
diagram
7.4.3 The Relationship between Loads, Shear Forces and

Bending Moments
Relationships between the applied loads and the internal shear force and bending moment
in a beam can be established by considering a small beam element, of width x , and
subjected to a distributed load p(x) which varies along the section of beam, and which is
positive upward, Fig. 7.4.18.

Section 7.4
p( x) p
V (x)
M (x) M ( x  x)
A
V ( x  x)
x x  x
x
Figure 7.4.18: forces and moments acting on a small element of beam
At the left-hand end of the free body, at position x, the shear force, moment and
distributed load have values F (x) , M (x) and p(x) respectively. On the right-hand end,
at position x  x , their values are slightly different: F ( x  x) , M ( x  x) and
p( x  x) . Since the element is very small, the distributed load, even if it is varying, can
be approximated by a linear variation over the element. The distributed load can
therefore be considered to be a uniform distribution of intensity p(x) over the length x
together with a triangular distribution, 0 at x and p say, a small value, at x  x .
Equilibrium of vertical forces then gives
1
V ( x)  p( x)x  px  V ( x  x)  0
2
(7.4.9)
V ( x  x)  V ( x) 1
  p( x)  p
x 2
Now let the size of the element decrease towards zero. The left-hand side of Eqn. 7.4.9 is
then the definition of the derivative, and the second term on the right-hand side tends to
zero, so
dV
 p(x) (7.4.10)
dx
This relation can be seen to hold in Eqn. 7.4.7 and Fig. 7.4.17a, where the shear force
over 0  x  6 has a slope of  40 and the pressure distribution is uniform, of intensity
 40 N/m . Similarly, over 6  x  12 , the pressure decreases linearly and so does the
slope in the shear force diagram, reaching zero slope at the end of the beam.
It also follows from 7.4.10 that the change in shear along a beam is equal to the area
under the distributed load curve:
x2
V ( x 2 )  V  x1    p( x)dx (7.4.11)
x1
Consider now moment equilibrium, by taking moments about the point A in Fig. 7.4.18:

Section 7.4
x 1 x
 M ( x)  V ( x)x  M ( x  x)  p( x)x  px 0
2 2 3
(7.4.12)
M ( x  x)  M ( x) x x
  V ( x)  p ( x)  p
x 2 6
Again, as the size of the element decreases towards zero, the left-hand side becomes a
derivative and the second and third terms on the right-hand side tend to zero, so that
dM
 V (x) (7.4.13)
dx
This relation can be seen to hold in Eqns. 7.4.2-3, 7.4.5 and 7.4.7-8. It also follows from
Eqn. 7.4.13 that the change in moment along a beam is equal to the area under the shear
force curve:
x2
M ( x 2 )  M  x1    V ( x)dx (7.4.14)
x1
7.4.4 Deformation and Flexural Stresses in Beams

The moment at any given cross-section of a beam is due to a distribution of normal stress,
or flexural stress (or bending stress) across the section (see Fig. 7.4.4). As mentioned,
the stresses to one side of the neutral axis are tensile whereas on the other side of the
neutral axis they are compressive. To determine the distribution of normal stress over the
section, one must determine the precise location of the neutral axis, and to do this one
must consider the deformation of the beam.
Apart from the assumption of there being a longitudinal plane of symmetry and a neutral
axis along which material fibres do not extend, the following two assumptions will be
made concerning the deformation of a beam:
1. cross sections which are plane and are perpendicular to the axis of the undeformed
beam remain plane and remain perpendicular to the deflection curve of the deformed
beam. In short: “plane sections remain plane”. This is illustrated in Fig. 7.4.19. It
will be seen later that this assumption is a valid one provided the beam is sufficiently
long and slender.
2. deformation in the vertical direction, i.e. the transverse strain  yy , may be neglected in
deriving an expression for the longitudinal strain  xx . This assumption is summarised
in the deformation shown in Fig. 7.4.20, which shows an element of length l and
height h undergoing transverse and longitudinal strain.

Section 7.4
y plane in deformed beam remains

perpendicular to the deflection curve
x
deflection curve
Figure 7.4.19: plane sections remain plane in the elementary beam theory
dh
h dl dh
 xx  ,  yy   0
l h
l
dl
Figure 7.4.20: transverse strain is neglected in the elementary beam theory
With these assumptions, consider now the element of beam shown in Fig. 7.4.21. Here,
two material fibres ab and pq , of length x in the undeformed beam, deform to a b 
and p q  . The deflection curve has a radius of curvature R. The above two assumptions
imply that, referring to the figure:
p a b  a bq    / 2 (assumption 1)

ap  a p  , bq  b q  (assumption 2) (7.4.15)
Since the fibre ab is on the neutral axis, by definition a b   ab . However the fibre
pq , a distance y from the neutral axis, extends in length from x to length x  . The
longitudinal strain for this fibre is
x   x ( R  y )  R y
 xx    (7.4.16)
x R R
As one would expect, this relation implies that a small R (large curvature) is related to a
large strain and a large R (small curvature) is related to a small strain. Further, for y  0
(above the neutral axis), the strain is negative, whereas if y  0 (below the neutral axis),
the strain is positive4, and the variation across the cross-section is linear.
4
this is under the assumption that R is positive, which means that the beam is concave up; a negative R
implies that the centre of curvature is below the beam

Section 7.4

p
 q y
x q y p x b
 b x
a

a
before after
Figure 7.4.21: deformation of material fibres in an element of beam
To relate this deformation to the stresses arising in the beam, it is necessary to postulate
the stress-strain law for the material out of which the beam is made. Here, it is assumed
that the beam is isotropic linear elastic5. Since there are no forces acting in the z
direction, the beam is in a state of plane stress, and the stress-strain equations are (see
Eqns. 6.1.10)
 xx 
1
E

 xx   yy 
 yy
1

  yy   xx
E

(7.4.17)

 zz  
E

 xx   yy 
1 
 xy   xy ,  xz   yz  0
E
Yet another assumption is now made, that the transverse normal stresses,  yy , may be
neglected in comparison with the flexural stresses  xx . This is similar to the above
assumption #2 concerning the deformation, where the transverse normal strain was
neglected in comparison with the longitudinal strain. It might seem strange at first that
the transverse stress is neglected, since all loads are in the transverse direction. However,
just as the tangential stresses are much larger than the radial stresses in the pressure
vessel, it is found that the longitudinal stresses in a beam are very much greater than the
transverse stresses. With this assumption, the first of Eqn. 7.4.17 reduces to a one-
dimensional equation:
 xx   xx / E (7.4.18)
5
the beam theory can be extended to incorporate more complex material models (constitutive equations)

Section 7.4
and, from Eqn. 7.4.16, dropping the subscripts on  ,
E
  y (7.4.19)
R
Finally, the resultant force of the normal stress distribution over the cross-section must be
zero, and the resultant moment of the distribution is M, leading to the conditions
E
0    dA  
R A
ydA
A
(7.4.20)
E 
M    ydA   y 2 dA    y 2 dA
A
RA yA
and the integration is over the complete cross-sectional area A. The minus sign in the
second of these equations arises because a positive moment and a positive y imply a
compressive (negative) stress (see Fig. 7.4.4).
The quantity A
ydA is the first moment of area about the neutral axis, and is equal to yA ,
where y is the centroid of the section (see, for example, §3.2.1). Note that the horizontal
component of the centroid will always be at the centre of the beam due to the symmetry
of the beam about the plane of bending. Since the first moment of area is zero, it follows
that y  0 : the neutral axis passes through the centroid of the cross-section.
The quantity A
y 2 dA is called the second moment of area or the moment of inertia
about the neutral axis, and is denoted by the symbol I. It follows that the flexural stress is
related to the moment through
My
  Flexural stress in a beam (7.4.21)
I
This is one of the most famous and useful formulas in mechanics.
The Moment of Inertia
The moment of inertia depends on the shape of a beam’s cross-section. Consider the
important case of a rectangular cross section. Before determining the moment of inertia
one must locate the centroid (neutral axis). Due to symmetry, the neutral axis runs
through the centre of the cross-section. To evaluate I for a rectangle of height h and
width b, consider a small strip of height dy at location y, Fig. 7.4.22. Then
h / 2
bh 3
I   y dA  b  y dy 
2 2
(7.4.22)
A h / 2
12

Section 7.4
This relation shows that the “taller” the cross-section, the larger the moment of inertia,
something which holds generally for I. Further, the larger is I, the smaller is the flexural
stress, which is always desirable.
dy
y
h  y0
centroid
Figure 7.4.22: Evaluation of the moment of inertia for a rectangular cross-section
For a circular cross-section with radius R, consider Fig. 7.4.23. The moment of inertia is
then
2 R
 R4
I   y dA 
2
 r
3
sin  drd 
2
(7.4.23)
A 0 0
4
r dA  rdrd
y
Figure 7.2.23: Moment of inertia for a circular cross-section
Example
Consider the beam shown in Fig. 7.4.24. It is loaded symmetrically by two concentrated
forces and has a circular cross-section of radius 100mm. The reactions at the two
supports are found to be 100N. Sectioning the beam to the left of the forces, and then to
the right of the first force, one finds that
V  100, M  100 x 0  x  250

(7.4.24)
V  0, M  25000 250  x  l/ 2
where l is the length of the beam.

Section 7.4
250 mm 250 mm
r  100 mm
Figure 7.4.24: a loaded beam with circular cross-section
The maximum tensile stress is then
M max ( y max ) 25000r

 max     31.8 MPa (7.4.25)
I r 4 / 4
and occurs at all sections between the two loads.

■
7.4.5 Shear Stresses in Beams

In the derivation of the flexural stress formula, Eqn. 7.4.21, it was assumed that plane
sections remain plane. This implies that there is no shear strain and, for an isotropic
elastic material, no shear stress, as indicated in Fig. 7.4.25.
shear stresses would

produce an angle change
before deformation
Figure 7.4.25: a section of beam before and after deformation
This fact will now be ignored, and an expression for the shear stress  within a beam will
be developed. It is implicitly assumed that this shear stress has little effect on the
calculation of the flexural stress.
As in Fig. 7.4.18, consider the equilibrium of a thin section of beam, as shown in Fig.
7.4.26. The beam has rectangular cross-section (although the theory developed here is
strictly for rectangular cross sections only, it can be used to give approximate shear stress
values in any beam with a plane of symmetry). Consider the equilibrium of a section of
this section, at the upper surface of the beam, shown hatched in Fig. 7.4.26. The stresses
acting on this section are as shown. Again, the normal stress is compressive at the
surface, consistent with the sign convention for a positive moment. Note that there are no
shear stresses acting at the surface – there may be distributed normal loads or forces
acting at the surface but, for clarity, these are not shown, and they are not necessary for
the following calculation.
From equilibrium of forces in the horizontal direction of the surface section:

Section 7.4
   
   dA     dA  bx  0 (7.4.26)
 A  x A  x  x
The third term on the left here assumes that the shear stress is uniform over the section –
this is similar to the calculations of §7.4.3 – for a very small section, the variation in
stress is a small term and may be neglected. Using the bending stress formula, Eqn.
7.4.21,
M ( x  x)  M ( x) y
 dA  b  0 (7.4.27)
A
x I
and, with Eqn. 7.4.13, as x  0 ,
VQ
 Shear stress in a beam (7.4.28)
Ib
where Q is the first moment of area A

ydA of the surface section of the cross-section.
b
  
V ( x)
M ( x) M ( x  x) 

V ( x  x)
x x  x
x
Figure 7.4.26: stresses and forces acting on a small section of material at the surface
of a beam
As mentioned, this formula 7.4.28 can be used as an approximation of the shear stress in a
beam of arbitrary cross-section, in which case b can be regarded as the depth of the beam
at that section. For the rectangular beam, one has
b  h2 
h/2
Q  b  ydy    y 2  (7.4.29)
y
2 4 
so that
6V  h 2 
   y 2 
3 
(7.4.30)
bh  4 

Section 7.4
The maximum shear stress in the cross-section arises at the neutral surface:
3V 3V
 max   (7.4.31)
2bh 2 A
and the shear stress dies away towards the upper and lower surfaces. Note that the
average shear stress over the cross-section is V / A and the maximum shear stress is
150% of this value.
Finally, since the shear stress on a vertical cross-section has been evaluated, the shear
stress on a longitudinal section has been evaluated, since the shear stresses on all four
sides of an element are the same, as in Fig.7.4.6.
Example
Consider the simply supported beam loaded by a concentrated force shown in Fig. 7.4.27.
The cross-section is rectangular with height 100 mm and width 50 mm . The reactions at
the supports are 5 kN and 15 kN . To the left of the load, one has V  5 kN and
M  5x kNm . To the right of the load, one has V  15 kN and M  30  15 x kNm .
The maximum shear stress will occur along the neutral axis and will clearly occur where
V is largest, so anywhere to the right of the load:
3Vmax
 max   4.5 MPa (7.4.32)
2A
1.5 m 20 kN 0.5 m
Figure 7.4.27: a simply supported beam
As an example of general shear stress evaluation, the shear stress at a point 25 mm below
the top surface and 1 m in from the left-hand end is, from Eqn 7.4.30,   1.125 MPa .
The shear stresses acting on an element at this location are shown in Fig. 7.4.28.
1m
  1.125 MPa
Figure 7.4.28: shear stresses acting at a point in the beam

■

Section 7.4
7.4.6 Approximate nature of the beam theory

The beam theory is only an approximate theory, with a number of simplifications made to
the full equations of elasticity. The accuracy of the theory is briefly explored in this
section.
When a beam is in pure bending, that is when the shear force is everywhere zero, the full
elasticity solution shows that plane sections do actually remain plane and the beam theory
is exact. For more complex loadings, plane sections do actually deform. For example, it
will be shown in Book II that the initially plane sections of a cantilever subjected to an
end force, Fig. 7.4.29, do not remain plane. Nevertheless, the beam theory prediction for
normal and shear stress is exact in this simple case.
deformed section not plane
Figure 7.4.29: a cantilevered beam loaded by a force and moment
Consider next a cantilevered beam of length l and rectangular cross section, height h and
width b, subjected to a uniformly distributed load p. With x measured from the
cantilevered end, the shear force and moment are given by V  p(l  x) and
 
M  ( pl 2 / 2) 1  2 x / l  2( x / l ) 2 . The shear stress is
6 p  h2 
  3   y 2 l  x  (7.4.33)
bh  4 
which turns out to be exact. The flexural stresses at the cantilevered end, at the upper
surface, are
2
 3 l 
   (7.4.34)
p 4h
The exact solution is, however (see Book II),
2
 3 l  1
    (7.4.35)
p 4h 5
It can be seen that the beam theory is a good approximation for the case when l / h is
large, in which case the term 1/5 is negligible.

Section 7.4
In summary, for most configurations, the elementary beam theory formulae for flexural
stress and transverse shear stress are accurate to within about 3% for beams whose length-
to-height ratio is greater than about 4.
7.4.7 Beam Deflection

Consider the deflection curve of a beam. The displacement of the neutral axis is denoted
by v, positive upwards, as in Fig. 7.4.30. The slope at any point is then given by the first
derivative, dv / dx .
For any type of material, provided the displacement is small, it can be shown that the
radius of curvature R is related to the second derivative d 2 v / dx 2 through (see the
Appendix to this section, §7.4.10)
1 d 2v
 (7.4.36)
R dx 2
and for this reason d 2 v / dx 2 is called the curvature of the beam. Using Eqn. 7.4.19,
   Ey / R , and the flexural stress expression, Eqn. 7.4.21,    My / I , one has the
moment-curvature equation
d 2v
M ( x)  EI moment-curvature equation (7.4.37)
dx 2
Figure 7.4.30: the deflection of a beam
With the moment known, this differential equation can be integrated twice to obtain the
deflection. Boundary conditions must be supplied to obtain constants of integration.
Example
Consider the cantilevered beam of length L shown in Fig. 7.4.31, subjected to an end-
force F and end-moment M 0 . The moment is found to be M ( x)  F ( L  x)  M 0 , with x
measured from the clamped end. The moment-curvature equation is then
d 2v
EI  ( FL  M 0 )  Fx
dx 2
dv 1
 EI  ( FL  M 0 ) x  Fx 2  C1 (7.4.38)
dx 2
1 1
 EIv  ( FL  M 0 ) x 2  Fx 3  C1 x  C 2
2 6

Section 7.4
The boundary conditions are that the displacement and slope are both zero at the clamped
end, from which the two constant of integration can be obtained:
v(0)  0  C2  0
(7.4.39)
v(0)  0  C1  0
L F
M0
Figure 7.4.31: a cantilevered beam loaded by an end-force and moment
The slope and deflection are therefore
1 1 1  dv 1  1 2
v  ( FL  M 0 ) x 2  Fx 3 ,  ( FL  M 0 ) x  2 Fx  (7.4.40)
EI  2 6  dx EI
The maximum deflection occurs at the end, where
1 1 1 3
v ( L)   2 M 0 L2
 FL  (7.4.41)
EI 3 
The term EI in Eqns. 7.4.40-41 is called the flexural rigidity, since it is a measure of the
resistance of the beam to deflection.
Example
Consider the simply supported beam of length L shown in Fig. 7.4.32, subjected to a
uniformly distributed load p over half its length. In this case, the moment is given by
3 1 2 L
 8 pLx  2 px 0 x
2

M ( x)   (7.4.42)
 1 L
 8 pL( L  x) xL
 2

Section 7.4
L/2 L/2
Figure 7.4.32: a simply supported beam subjected to a uniformly distributed load

over half its length
It is necessary to apply the moment-curvature equation to each of the two regions

0  x  L / 2 and L / 2  x  L separately, since the expressions for the moment in these
regions differ. Thus there will be four constants of integration:
d 2v 3 1 d 2v 1 2 1
EI 2
 pLx  px 2 EI  pL  pLx
dx 8 2 dx 2 8 8
dv 3 1 dv 1 2 1
 EI  pLx 2  px 3  C1  EI  pL x  pLx 2  D1
dx 16 6 dx 8 16
3 1 1 1
 EIv  pLx 3  px 4  C1 x  C 2  EIv  pL2 x 2  pLx 3  D1 x  D2
48 24 16 48
(7.4.43)
The boundary conditions are: (i) no deflection at pin support, v(0)  0 and (ii) no
deflection at roller support, v( L)  0 , from which one finds that C 2  0 and
D2   pL4 / 24  D1 L . The other two necessary conditions are the continuity conditions
where the two solutions meet. These are that (i) the deflection of both solutions agree at
x  L / 2 and (ii) the slope of both solutions agree at x  L / 2 . Using these conditions,
one finds that
9 pL3 17 pL3
C1   , C2   (7.4.44)
384 384
so that
 wL4   x  x
3
x 
4
L
  9   24   16   0 x
 384 EI   L  L  L   2
v (7.4.45)
 wL 1  17 x   24 x   8 x  
4 2 3
L
 384 EI         xL
 
  L   L   L   2
The deflection is shown in Fig. 7.4.33. Note that the maximum deflection occurs in
0  x  L / 2 ; it can be located by setting dv / dx  0 there and solving.

Section 7.4
x/L
3 4
x x x
 9   24   16 
L L L
1
384 EI
v
pL4
2
2 3
x x x
1  17   24   8 
L L L
Figure 7.4.33: deflection of a beam

■
7.4.8 Statically Indeterminate Beams

Consider the beam shown in Fig. 7.4.34. It is cantilevered at one end and supported by a
roller at its other end. A moment is applied at its centre. There are three unknown
reactions in this problem, the reaction force at the roller and the reaction force and
moment at the built-in end. There are only two equilibrium equations with which to
determine these three unknowns and so it is not possible to solve the problem from
equilibrium considerations alone. The beam is therefore statically indeterminate (see the
end of section 2.3.3).
M0
Figure 7.4.34: a cantilevered beam supported also by a roller
More examples of statically indeterminate beam problems are shown in Fig. 7.4.35. To
solve such problems, one must consider the deformation of the beam. The following
example illustrates how this can be achieved.

Section 7.4
Figure 7.4.35: examples of statically indeterminate beams
Example
Consider the beam of length L shown in Fig. 7.4.36, cantilevered at end A and supported
by a roller at end B. A moment M 0 is applied at B.
A M0 M0
B
Figure 7.4.36: a statically indeterminate beam
The moment along the beam can be expressed in terms of the unknown reaction force at
end B: M ( x)  RB ( L  x)  M 0 . As before, one can integrate the moment-curvature
equation:
d 2v
EI  R B ( L  x)  M 0
dx 2
1
 R B L  M 0 x  RB x 2  C1
dv
 EI (7.4.46)
dx 2
1 1
 EIv  R B L  M 0 x 2  RB x 3  C1 x  C 2
2 6
There are three boundary conditions, two to determine the constants of integration and
one can be used to determine the unknown reaction RB . The boundary conditions are (i)
v(0)  0  C 2  0 , (ii) dv / dx(0)  0  C1  0 and (iii) v( L)  0 from which one finds
that R B  3M 0 / 2 L . The slope and deflection are therefore

Section 7.4
 x  3  x  2 
M B L2
v      
4 EI L   L  
(7.4.47)
dv M B L   x   x 
2
 3   2 
dx 4 EI   L   L 
One can now return to the equilibrium equations to find the remaining reactions acting on
the beam, which are R A   RB and M A  M 0  LR B
■
7.4.9 The Three-point Bending Test

The 3-point bending test is a very useful experimental procedure. It is used to gather data
on materials which are subjected to bending in service. It can also be used to get the
Young’s Modulus of a material for which it might be more difficult to get via a tension or
other test.
A mouse bone is shown in the standard 3-point bend test apparatus in Fig. 7.4.37a. The
idealised beam theory model of this test is shown in Fig. 7.4.37b. The central load is P,
so the reactions at the supports are P / 2 . The moment is zero at the supports, varying
linearly to a maximum PL / 4 at the centre.
F
L/2 L/2
vmax
1
(a) (b)
Figure 7.4.37: the three-point bend test; (a) a mouse bone specimen, (b) idealised
model
The maximum flexural stress then occurs at the outer fibres at the centre of the beam: for
a circular cross-section,  max  FL /  R 3 . Integrating the moment-curvature equation,
and using the fact that the deflection is zero at the supports and, from symmetry, the slope
is zero at the centre, the maximum deflection is seen to be vmax  FL3 / 12 R 4 E . If one
plots the load F against the deflection vmax , one will see a straight line (initially, before
the elastic limit is reached); let the slope of this line be Ê . The Young’s modulus can
then be evaluated through
L3 ˆ
E E (7.4.48)
12 R 4

Section 7.4
With   E , the maximum strain is  max  FL /  ER 3  12 Rvmax / L2 . By carrying the

test on beyond the elastic limit, the strength of the material at failure can be determined.
7.4.10 Problems
1. The simply supported beam shown below carries a vertical load that increases
uniformly from zero at the left end to a maximum value of 9 kN/m at the right end.
Draw the shearing force and bending moment diagrams
9 kN/m
6m
2. The beam shown below is imply supported at two points and overhangs the supports
at each end. It is subjected to a uniformly distributed load of 4 kN/m as well as a
couple of magnitude 8 kN m applied to the centre. Draw the shearing force and
bending moment diagrams
8 kN m
4 kN/m
1m 2m 2m 1m
3. Evaluate the centroid of the beam cross-section shown below (all measurements in
mm)
75 75
50
225
50
50 100
4. Determine the maximum tensile and compressive stresses in the following beam (it
has a rectangular cross-section with height 75 mm and depth 50 mm)

Section 7.4
1 kN
5m
5. Consider the cantilever beam shown below. Determine the maximum shearing stress
in the beam and determine the shearing stress 25 mm from the top surface of the beam
at a section adjacent to the supporting wall. The cross-section is the “T” shape shown,
for which I  40  10 6 mm 4 .
[note: use the shear stress formula derived for rectangular cross-sections – as
mentioned above, in this formula, b is the thickness of the beam at the point where the
shear stress is being evaluated]
50
50 kN
125
2m
50
200
6. Obtain an expression for the maximum deflection of the simply supported beam
shown here, subject to a uniformly distributed load of w N/m .
w N/m
7. Determine the equation of the deflection curve for the cantilever beam loaded by a
concentrated force P as shown below.
a P
8. Determine the reactions for the following uniformly loaded beam clamped at both
ends.

Section 7.4
w N/m
Curvature of the deflection curve
Consider a deflection curve with deflection v(x) and radius of curvature R(x) , as shown
in the figure below. Here, deflection is the transverse displacement (in the y direction) of
the points that lie along the axis of the beam. A relationship between v(x) and R(x) is
derived in what follows.
y, v( x)
R (x )
v(x)
First, consider a curve (arc) s. The tangent to some point p makes an angle  with the x
– axis, as shown below. As one move along the arc,  changes.
y
p ds
dy
 dx
Define the curvature  of the curve to be the rate at which  increases relative to s,
d

ds
Thus if the curve is very “curved”,  is changing rapidly as one moves along the curve
(as one increase s) and the curvature will be large.

Section 7.4
From the above figure,
dy ds (dx) 2  (dy ) 2
tan  ,   1  (dy / dx) 2 ,
dx dx dx
so that
d d dx d arctandy / dx  dx 1 d 2 y dx
   
ds dx ds dx ds 1  dy / dx 2 dx 2 ds
d 2 y / dx 2

1  dy / dx 2 3/ 2
Finally, it will be shown that the curvature is simply the reciprocal of the radius of
curvature. Draw a circle to the point p with radius R . Arbitrarily measure the arc length
s from the point c, which is a point on the circle such that cop   . Then arc length
s  R , so that
d 1
 
ds R
o o d
R
 R
c p d
 ds
x x
Thus
d 2v
1 dx 2
 3
R
  dv  2  2
1    
  dx  
If one assumes now that the slopes of the deflection curve are small, then dv / dx  1 and
1 d 2v

R dx 2
Images used:
1. http://www.mc.vanderbilt.edu/root/vumc.php?site=CenterForBoneBiology&doc=20412

Answers to Selected Problems: Chapter 2
2.2
1. Fcable  31.4 kN, Frope  56.4 kN

2. 150 N
3 10 g / 3  56.6 N
2.3
3. F  600 N , R xC  1000 N, R yC  F
4. M  6.25 Nm , i.e. clockwise. Anywhere.
A1
3.1
1. (a) F  0.8 N , (b) M  0.0213 Nm , (c) M  0.0053 Nm

2. F  0, M  0.0053 Nm
3. RA  2000N/m, RB  1600N/m
4. 2kPa
3.2
1 1
1. xc  , yc 
3 3
3.3
2.  N  S / l,  s  0
3. at A: No stress at B.
4.  yz is negative
5.  yz (positive)
6. bottom left:  xz (negative), top:  zx (negative), bottom right:  yz (positive)
7. bottom left:  xx (positive), top:  zz (positive), bottom right:  yy (positive)
3.4
2.  xx  0.884,  22
  2.116,  12  0.933
3.5
2. (a)  1   xx ,  2  0 , (b)  xx   yy   xx / 2,  xy   xx / 2

3. (b)   32 o , (c)  1  3.85,  2  2.85
4.  1  2 ,  2  0,  3  0 , Max shear is  , the original planes are planes of maximum
shearing stress.
5. (b)  1   , (c) Max shear is 12  , acts on planes oriented at 45o to the principal planes
acting in the 2-3 plane and the 1-3 plane
8.  xx(2)   xx( w ) ,  xy(2)   xy( w )
b b a
 t   xy ( x,0)dx  0,  t   yy ( x,0)dx  t  p( x)dx  0
b b a
9. b a
 t  ( x  b) yy ( x,0)dx  t  ( x  b) p( x)dx  0
b a
A1
4.1
1.
(a) 12cm
(b) 0.6
2.
(a)   1.0 ,  t  0.693
(b)   0.25 ,  t  0.288
(c)   0.5 ,  t  0.405
3.
 xy  0.0015
4.
Small strains:  xx  0.00225,  yy  0.002,  xy  0.00083
Actual strains:  xx  0.00225,  yy  2.001386  103 ,  xy  8.3166823  104
Errors: 0%, 0.069%, 0.200%
5.
(i)
y
x
(ii)
y
x
(iii)
y
6.
 xx  0,  yy  3.8053  10 3 ,  xy  1.6646  104
A1
4.2
1.
(a)
 xx  0.02,  yy   xy  0
 xx  0.015,  yy  0.005,  xy  8.66  103
(b) 1  0.02,  2  0
(c) max  xy   0.01
(d) 45 degrees.
2.
(a)  xx  0.01,  yy  0.01,  xy  0
(b)  xx
  0,  yy  0,  xy  0.01
(c) the same as (b) using   45
(d) 1.0001  1 . Close to (b).
3.
(a)  xx  0.5,  yy  0.5,  xy  0
(b)
 xx  0,  yy  0,  xy  0.5
(c) the same as (b) using   45
(d) 2  1  0.414 . Not the same as (b).
A2
5.2
1. Each component treated separately would be homogeneous and isotropic; the complete
structure is not homogeneous; it is not isotropic along the interfaces between the separate
components.
5.4
1. 5mm
A1
6.1
Q.2
 xx  15.92308 MPa,  yy  11.07692 MPa,  xy  2.42308 MPa
 1  16.927,  2  10.073,  3  0
5.
(a)  xx   0 / E ,  yy   zz   0 / E ,  xy   xz   yz  0
(b) 1   0 / E ,  2   3   0 / E
6. 3.1  10 6 rads
6.2
1.
normal strains: 0,   o (1   2 ) / E ,  o (1   ) / E
normal stresses:   0 ,   0 , 0 , (b)  o / 2
2
normal strains: 0,   o (1  2 )(1   ) /(1   ) E , 0
normal stresses:   0 /(1   ),   0 ,   0 /(1   ) ]
6.3
2.  2  0.6969 MPa,  3  0.5360 MPa,  1  8.463  10 5
 t 1   f   f 1   t 
3. 2  p , 3  p
1   t f 1   t f
4.
(a)  1  48.25,  2  36.75,  6  4.82
(b)  1  0.0317,  2  0.1623,  6  3.4486  10 5
(c) G  69.9kPa , E 2  200Pa,  21  0.0904
(d) principal stresses (   0 ): 50, 35, 0
principal strains   20.02 ):  p1  0.1623,  p 2  0.0317
(e) No.
5.   48.2 o .
A1
7.1
1. L / 1  A1 L2 / A2 L1 
7.2
1.  max  20T / d 3 ,  B  12 LT / Gd 4
7.3
1. 0.25 MPa
2. t / 2
3. (a) 15MPa, 30MPa, (b) 15MPa, (c) 750N
4. 0.13, 0.03, -0.07 (  10 3 ), 0.015mm
7.4
1. V  9  0.75 x 2 kN , M  9 x  0.25 x 3 kN m
2. M  2 x 2 ;  2 x 2  10 x  10;  2 x 2  10 x  2;  2 x 2  24 x  72 kN m
3. y  152.3 mm
4. 107 MPa
5. 4.78 MPa;  0.43 MPa
6.  max  5wL4  / 384 EI 
7. EIv  ( Pa 3 / 6)( x / a ) 2  3  x / a  ; EIv  ( Pa 3 / 6)  3x / a  1
8. R  wL / 2; M   wL2 / 12
A1
1 Differential Equations
for Solid Mechanics
Simple problems involving homogeneous stress states have been considered so far,
wherein the stress is the same throughout the component under study. An exception to
this was the varying stress field in the loaded beam, but there a simplified set of elasticity
equations was used. Here the question of varying stress and strain fields in materials is
considered. In order to solve such problems, a differential formulation is required. In this
Chapter, a number of differential equations will be derived, relating the stresses and body
forces (equations of motion), the strains and displacements (strain-displacement
relations) and the strains with each other (compatibility relations). These equations are
derived from physical principles and so apply to any type of material, although the latter
two are derived under the assumption of small strain.
1
2
Section 1.1
1.1 The Equations of Motion

In Part I, balance of forces and moments acting on any component was enforced in order
to ensure that the component was in equilibrium. Here, allowance is made for stresses
which vary continuously throughout a material, and force equilibrium of any portion of
material is enforced.
One-Dimensional Equation
Consider a one-dimensional differential element of length Δx and cross sectional area A,

Fig. 1.1.1. Let the average body force per unit volume acting on the element be b and the
average acceleration and density of the element be a and ρ . Stresses σ act on the
element.
b, a
σ ( x + Δx)
A
σ (x)
x x + Δx
Δx
Figure 1.1.1: a differential element under the action of surface and body forces
The net surface force acting is σ ( x + Δx) A − σ ( x) A . If the element is small, then the
body force and velocity can be assumed to vary linearly over the element and the average
will act at the centre of the element. Then the body force acting on the element is AbΔx
and the inertial force is ρAΔxa . Applying Newton’s second law leads to
σ ( x + Δx) A − σ ( x) A + bΔxA = ρaΔxA

σ ( x + Δx) − σ ( x) (1.1.1)
→ + b = ρa
Δx
so that, by the definition of the derivative, in the limit as Δx → 0 ,
dσ
+ b = ρa 1-d Equation of Motion (1.1.2)
dx
which is the one-dimensional equation of motion. Note that this equation was derived
on the basis of a physical law and must therefore be satisfied for all materials, whatever
they be composed of.
The derivative dσ / dx is the stress gradient – physically, it is a measure of how rapidly

the stresses are changing.
Example
Consider a bar of length l which hangs from a ceiling, as shown in Fig. 1.1.2.
Solid Mechanics Part II 3 Kelly

Section 1.1
z
l
Figure 1.1.2: a hanging bar
The gravitational force is F = mg downward and the body force per unit volume is thus
b = ρg . There are no accelerating material particles. Taking the z axis positive down, an
integration of the equation of motion gives
dσ
+ ρg = 0 → σ = − ρgz + c (1.1.3)
dz
where c is an arbitrary constant. The lower end of the bar is free and so the stress there is
zero, and so
σ = ρg (l − z ) (1.1.4)
Two-Dimensional Equations
Consider now a two dimensional infinitesimal element of width and height Δx and Δy
and unit depth (into the page).
Looking at the normal stress components acting in the x −direction, and allowing for
variations in stress over the element surfaces, the stresses are as shown in Fig. 1.1.3.
σ xx ( x, y + Δy ) σ xx ( x + Δx, y + Δy )
Δy
Δx
σ xx ( x, y ) σ xx ( x + Δx, y )
Figure 1.1.3: varying stresses acting on a differential element
Using a (two dimensional) Taylor series and dropping higher order terms then leads to the
linearly varying stresses illustrated in Fig. 1.1.4. (where σ xx ≡ σ xx ( x, y ) and the partial
derivatives are evaluated at ( x, y ) ), which is a reasonable approximation when the
element is small.

Section 1.1
∂σ ∂σ xx ∂σ xx
σ xx + Δy xx σ xx + Δx + Δy
∂y ∂x ∂y
∂σ xx
σ xx σ xx + Δx
∂x
Figure 1.1.4: linearly varying stresses acting on a differential element
The effect (resultant force) of this linear variation of stress on the plane can be replicated
by a constant stress acting over the whole plane, the size of which is the average stress.
For the left and right sides, one has, respectively,
1 ∂σ xx ∂σ xx 1 ∂σ xx
σ xx + Δy , σ xx + Δx + Δy (1.1.5)
2 ∂y ∂x 2 ∂y
One can take away the stress (1 / 2)Δy∂σ xx / ∂y from both sides without affecting the net
force acting on the element so one finally has the representation shown in Fig. 1.1.5.
∂σ xx
σ xx ( x, y ) σ xx + Δx
∂x
Figure 1.1.5: net stresses acting on a differential element
Carrying out the same procedure for the shear stresses contributing to a force in the
x −direction leads to the stresses shown in Fig. 1.1.6.
∂ σ xy
σ xy + Δy
∂y
b1 , v1
∂ σ xx
σ xx ( x, y ) Δx 2 σ + Δx
∂x
xx
Δx1
σ xy ( x, y )
Figure 1.1.6: normal and shear stresses acting on a differential element
Take a x , bx to be the average acceleration and body force, and ρ to be the average
density. Newton’s law then yields

Section 1.1
⎛ ∂σ ⎞ ⎛ ∂σ xy ⎞
− σ xx Δy + ⎜ σ xx + Δx xx ⎟Δy − σ xy Δx + ⎜⎜ σ xy + Δy ⎟⎟Δy + bx ΔxΔy = ρa x ΔxΔy
⎝ ∂x ⎠ ⎝ ∂y ⎠
(1.1.6)
which, dividing through by ΔxΔy and taking the limit, gives
∂σ xx ∂σ xy
+ + bx = ρa x (1.1.7)
∂x ∂y
A similar analysis for force components in the y −direction yields another equation and
one then has the two-dimensional equations of motion:
∂σ xx ∂σ xy
+ + b x = ρa x
∂x ∂y
2-D Equations of Motion (1.1.8)
∂σ xy ∂σ yy
+ + b y = ρa y
∂x ∂y
Three-Dimensional Equations
Similarly, one can consider a three-dimensional element, and one finds that
∂σ xx ∂σ xy ∂σ xz
+ + + bx = ρa x
∂x ∂y ∂z
∂σ yx ∂σ yy ∂σ yz
+ + + b y = ρa y 3-D Equations of Motion (1.1.9)
∂x ∂y ∂z
∂σ zx ∂σ zy ∂σ zz
+ + + bz = ρa z
∂x ∂y ∂z
These three equations express force-balance in, respectively, the x, y, z directions.

Section 1.1
Figure 1.1.7: from Cauchy’s Exercices de Mathematiques (1829)
The Equations of Equlibrium
If the material is not moving (or is moving at constant velocity) and is in static
equilibrium, then the equations of motion reduce to the equations of equilibrium,
∂σ xx ∂σ xy ∂σ xz
+ + + bx = 0
∂x ∂y ∂z
∂σ yx ∂σ yy ∂σ yz
+ + + by = 0 3-D Equations of Equilibrium (1.1.10)
∂x ∂y ∂z
∂σ zx ∂σ zy ∂σ zz
+ + + bz = 0
∂x ∂y ∂z
These equations express the force balance between surface forces and body forces in a
material. The equations of equilibrium may also be used as a good approximation in the
analysis of materials which have relatively small accelerations.
1.1.2 Problems
1. What does the one-dimensional equation of motion say about the stresses in a bar in
the absence of any body force or acceleration?
2. Does equilibrium exist for the following two dimensional stress distribution in the
absence of body forces?
σ xx = 3x 2 + 4 xy − 8 y 2
σ xy = σ yx = x 2 / 2 − 6 xy − 2 y 2
σ yy = 2 x 2 + xy + 3 y 2
σ zz = σ zx = σ xz = σ zy = σ yz = 0

Section 1.1
3. The elementary beam theory predicts that the stresses in a circular beam due to
bending are
σ xx = My / I , σ xy = σ yx = V ( R 2 − y 2 ) / 3I ( I = πR 4 / 4)
and all the other stress components are zero. Do these equations satisfy the equations
of equilibrium?
4. With respect to axes 0 xyz the stress state is given in terms of the coordinates by the
matrix
⎡σ xx σ xy σ xz ⎤ ⎡ xy y 2 0 ⎤
[ ] ⎢ ⎥ ⎢
σ ij = ⎢σ yx σ yy σ yz ⎥ = ⎢ y 2 yz z 2 ⎥
⎥
⎢σ zx σ zy σ zz ⎥ ⎢ 0 z 2 xz ⎥
⎣ ⎦ ⎣ ⎦
Determine the body force acting on the material if it is at rest.
5. What is the acceleration of a material particle of density ρ = 0.3kgm -3 , subjected to

the stress
⎡2 x 2 − x 4 2 xy 2 xz ⎤
[ ]
σ ij = ⎢
⎢
2 xy 2y − y
2 4
2 yz ⎥
⎥
⎢⎣ 2 xz 2 yz 2 z 2 − z 4 ⎥⎦
and gravity (the z axis is directed vertically upwards from the ground).
6. A fluid at rest is subjected to a hydrostatic pressure p and the force of gravity only.
(a) Write out the equations of motion for this case.
(b) A very basic formula of hydrostatics, to be found in any elementary book on fluid
mechanics, is that giving the pressure variation in a static fluid,
Δp = ρgh
where ρ is the density of the fluid, g is the acceleration due to gravity, and h is
the vertical distance between the two points in the fluid (the relative depth).
Show that this formula is but a special case of the equations of motion.

Section 1.2
1.2 The Strain-Displacement Relations

The strain was introduced in Book I: §4. The concepts examined there are now extended
to the case of strains which vary continuously throughout a material.
1.2.1 The Strain-Displacement Relations
Normal Strain
Consider a line element of length x emanating from position ( x, y ) and lying in the x -
direction, denoted by AB in Fig. 1.2.1. After deformation the line element occupies
AB  , having undergone a translation, extension and rotation.
y u x ( x  x, y )
u x ( x, y )  B

A B*
 
A B
x
x x  x
Figure 1.2.1: deformation of a line element
The particle that was originally at x has undergone a displacement u x ( x, y ) and the other
end of the line element has undergone a displacement u x ( x  x, y ) . By the definition of
(small) normal strain,
AB *  AB u x ( x  x, y )  u x ( x, y )
 xx   (1.2.1)
AB x
In the limit x  0 one has
u x
 xx  (1.2.2)
x
This partial derivative is a displacement gradient, a measure of how rapid the

displacement changes through the material, and is the strain at ( x, y ) . Physically, it
represents the (approximate) unit change in length of a line element, as indicated in Fig.
1.2.2.

Section 1.2
B
A
B*
A B x
u x
x
x x
Figure 1.2.2: unit change in length of a line element
Similarly, by considering a line element initially lying in the y direction, the strain in the y
direction can be expressed as
u y
 yy  (1.2.3)
y
Shear Strain
The particles A and B in Fig. 1.2.1 also undergo displacements in the y direction and this
is shown in Fig. 1.2.3. In this case, one has
u y
B * B  x (1.2.4)
x
y
u y ( x  x, y)
 B
u y ( x, y )

A B*
 
A B
x
x x  x
Figure 1.2.3: deformation of a line element
A similar relation can be derived by considering a line element initially lying in the y
direction. A summary is given in Fig. 1.2.4. From the figure,
u y / x u y
  tan   
1  u x / x x
provided that (i)  is small and (ii) the displacement gradient u x / x is small. A similar
expression for the angle  can be derived, and hence the shear strain can be written in
terms of displacement gradients.

Section 1.2
u x
y y
u y
y
y

y u y
y  x
x
u
x x x x
x
Figure 1.2.4: strains in terms of displacement gradients
The Small-Strain Stress-Strain Relations
In summary, one has
u x
 xx 
x
u y
 yy  2-D Strain-Displacement relations (1.2.5)
y
1  u u y 
 xy   x  
2  y x 
1.2.2 Geometrical Interpretation of Small Strain

A geometric interpretation of the strain was given in Book I: §4.1.4. This interpretation is
repeated here, only now in terms of displacement gradients.
Positive Normal Strain
Fig. 1.2.5a,
u x u y 1  u y 
 xx   0,  yy   0,  xy   x    0 (1.2.6)
x y 2  y x 
Negative Normal Strain
Fig 1.2.5b,
u x u y 1  u u y 
 xx   0,  yy   0,  xy   x  0 (1.2.7)
x y 2  y x 

Section 1.2
y u (x) u x ( x  x) ux ( y)
x
x
(a ) ( b) (c )
Figure 1.2.5: some simple deformations; (a) positive normal strain, (b) negative
normal strain, (c) simple shear
Simple Shear
Fig. 1.2.5c,
u x u y 1  u u y  1 u x
 xx   0,  yy   0,  xy   x   (1.2.8)
x y 2  y x  2 y
Pure Shear
Fig 1.2.6a,
u x u y 1  u u y  u x u y
 xx   0,  yy   0,  xy   x    (1.2.9)
x y 2  y x  y x
1.2.3 The Rotation

Consider an arbitrary deformation (omitting normal strains for ease of description), as
shown in Fig. 1.2.6. As usual, the angles  and  are small, equal to their tangents, and
  u y / x ,   ux / y .
x2
ux ( y)
 u y (x)
 x1
Figure 1.2.6: arbitrary deformation (shear and rotation)

Section 1.2
Now this arbitrary deformation can be decomposed into a pure shear and a rigid rotation
as depicted in Fig. 1.2.7. In the pure shear,      xy  12     . In the rotation, the
angle of rotation is then 1
2     .
y
u x u y
 
y x arbitrary shear strain
x
rotation
pure (no strain)
shear
 xy
 xy  z
z
1  u u   u 
 xy 
1
      x  y  z 
1
     1  y  u x 
2 2  y x  2 2  x y 
Figure 1.2.7: decomposition of a strain into a pure shear and a rotation
This leads one to define the rotation of a material particle,  z , the “z” signifying the axis
about which the element is rotating:
1  u y u 
 z    x  (1.2.10)
2  x y 
The rotation will in general vary throughout a material. When the rotation is everywhere
zero, the material is said to be irrotational.
For a pure rotation, note that
1  u u  u x u
 xy   x  y   0,  y (1.2.11)
2  y x  y x

Section 1.2
1.2.4 Fixing Displacements

The strains give information about the deformation of material particles but, since they do
not encompass translations and rotations, they do not give information about the precise
location in space of particles. To determine this, one must specify three displacement
components (in two-dimensional problems). Mathematically, this is equivalent to saying
that one cannot uniquely determine the displacements from the strain-displacement
relations 1.2.5.
Example
Consider the strain field  xx  0.01,  yy   xy  0 . The displacements can be obtained by

integrating the strain-displacement relations:
u x    xx dx  0.01x  f ( y )
(1.2.12)
u y    yy dy  g ( x)
where f and g are unknown functions of y and x respectively. Substituting the

displacement expressions into the shear strain relation gives
f ( y )   g ( x) . (1.2.13)
Any expression of the form F ( x)  G ( y ) which holds for all x and y implies that F and G
are constant1. Since f , g  are constant, one can integrate to get
f ( y )  A  Dy, g ( x)  B  Cx . From 1.2.13, C   D , and
u x  0.01x  A  Cy
u y  B  Cx (1.2.14)
There are three arbitrary constants of integration, which can be determined by specifying
three displacement components. For example, suppose that it is known that
u x (0,0)  0, u y (0,0)  0, u x (0, a )  b . (1.2.15)
In that case, A  0, B  0, C  b / a , and, finally,
u x  0.01x  (b / a ) y
u y  (b / a) x (1.2.16)
which corresponds to Fig. 1.2.8, with (b / a ) being the (tan of the small) angle by which
the element has rotated.
1
since, if this was not so, a change in x would change the left hand side of this expression but would not
change the right hand side and so the equality cannot hold

Section 1.2
y b
Figure 1.2.8: an element undergoing a normal strain and a rotation
In general, the displacement field will be of the form
u x    A  Cy
u y    B  Cx (1.2.17)
and indeed Eqn. 1.2.16 is of this form. Physically, A, B and C represent the possible rigid
body motions of the material as a whole, since they are the same for all material particles.
A corresponds to a translation in the x direction, B corresponds to a translation in the x
direction, and C corresponds to a positive (counterclockwise) rotation.

The three-dimensional stress-strain relations analogous to Eqns. 1.2.5 are
u x u y u
 xx  ,  yy  ,  zz  z
x y z
1  u x u y  1  u x u z  1  u y u z 
 xy    ,  xz    ,  yz    
2  y x  2  z x  2  z y 
3-D Stress-Strain relations (1.2.18)
The rotations are
1  u u  1  u u  1  u u 
z   y  x  ,  y   x  z  ,  x   z  y  (1.2.19)
2  x y  2  z x  2  y z 
1.2.6 Problems
1. The displacement field in a material is given by
u x  A3 x  y , u y  Axy 2
where A is a small constant.

Section 1.2
(a) Evaluate the strains. What is the rotation  z ? Sketch the deformation and any
rigid body motions of a differential element at the point (1, 1)
(b) Sketch the deformation and rigid body motions at the point (0, 2) , by using a pure
shear strain superimposed on the rotation.
2. The strains in a material are given by

 xx  x,  yy  0,  xy  
Evaluate the displacements in terms of three arbitrary constants of integration, in the
form of Eqn. 1.2.17,
u x    A  Cy
u y    B  Cx
What is the rotation?
3. The strains in a material are given by

 xx  Axy,  yy  Ay 2 ,  xy  Ax
where A is a small constant. Evaluate the displacements in terms of three arbitrary
constants of integration. What is the rotation?
4. Show that, in a state of plane strain (  zz  0 ) with zero body force,

e  z  2 u x  2 u x
2  
x y x 2 y 2
where e is the volumetric strain (dilatation), the sum of the normal strains:
e   xx   yy   zz (see Book I, §4.3).

Section 1.3
1.3 Compatibility of Strain

As seen in the previous section, the displacements can be determined from the strains
through integration, to within a rigid body motion. In the two-dimensional case, there are
three strain-displacement relations but only two displacement components. This implies
that the strains are not independent but are related in some way. The relations between
the strains are called compatibility conditions.
1.3.1 The Compatibility Relations
Differentiating the first of 1.2.5 twice with respect to y , the second twice with respect to
x and the third once each with respect to x and y yields
∂ 2 ε xx ∂ 3u x ∂ 2 ε yy ∂ 3u y ∂ 2 ε xy
1 ⎛⎜ ∂ 3 u x ∂ 3u y ⎞
= , = , = + ⎟
∂y 2 ∂x∂y 2 ∂x 2 ∂x 2 ∂y ∂x∂y 2 ⎜⎝ ∂x∂y 2 ∂x 2 ∂y ⎟⎠
It follows that
∂ 2 ε xx ∂ ε yy ∂ 2 ε xy
2
+ = 2 2-D Compatibility Equation (1.3.1)

∂y 2 ∂x 2 ∂x∂y
This compatibility condition is an equation which must be satisfied by the strains at all
material particles.
Physical Meaning of the Compatibility Condition
When all material particles in a component deform, translate and rotate, they need to meet
up again very much like the pieces of a jigsaw puzzle must fit together. Fig. 1.3.1
illustrates possible deformations and rigid body motions for three line elements in a
material. Compatibility ensures that they stay together after the deformation.
deformed
- compatibility ensured
deformed
undeformed - compatibility not satisfied
Figure 1.3.1: Deformation and Compatibility

Section 1.3
The Three Dimensional Case
There are six compatibility relations to be satisfied in the three dimensional case :
∂ 2 ε yy ∂ 2 ε zz ∂ 2 ε yz ∂ 2 ε xx ∂ ⎛ ∂ε yz ∂ε zx ∂ε xy ⎞
+ = 2 , = ⎜⎜ − + + ⎟
∂z 2 ∂y 2
∂y∂z ∂y∂z ∂x ⎝ ∂x ∂y ∂z ⎟⎠
∂ 2 ε zz ∂ 2 ε xx ∂ 2 ε zx ∂ 2 ε yy ∂ ⎛ ∂ε yz ∂ε zx ∂ε xy ⎞
+ = 2 , = ⎜⎜ + − + ⎟ (1.3.2)
∂x 2 ∂z 2
∂z∂x ∂z∂x ∂y ⎝ ∂x ∂y ∂z ⎟⎠
∂ 2 ε xx ∂ 2 ε yy ∂ 2 ε xy ∂ 2 ε zz ∂ ⎛ ∂ε yz ∂ε zx ∂ε xy ⎞
+ = 2 , = ⎜+ + − ⎟
∂y 2 ∂x 2 ∂x∂y ∂x∂y ∂z ⎜⎝ ∂x ∂y ∂z ⎟⎠
By inspection, it will be seen that these are satisfied by Eqns. 1.2.19.
1.3.2 Problems
1. The displacement field in a material is given by
u x = Axy, u y = Ay 2 ,
where A is a small constant. Determine
(a) the components of small strain
(b) the rotation
(c) the principal strains
(d) whether the compatibility condition is satisfied

7 3D Elasticity
187
188
Section 7.1
7.1 Vectors, Tensors and the Index Notation

The equations governing three dimensional mechanics problems can be quite lengthy.
For this reason, it is essential to use a short-hand notation called the index notation1.
Consider first the notation used for vectors.
7.1.1 Vectors
Vectors are used to describe physical quantities which have both a magnitude and a
direction associated with them. Geometrically, a vector is represented by an arrow; the
arrow defines the direction of the vector and the magnitude of the vector is represented by
the length of the arrow. Analytically, in what follows, vectors will be represented by
lowercase bold-face Latin letters, e.g. a, b.
The dot product of two vectors a and b is denoted by a ⋅ b and is a scalar defined by
a ⋅ b = a b cosθ . (7.1.1)
θ here is the angle between the vectors when their initial points coincide and is restricted
to the range 0 ≤ θ ≤ π .
Cartesian Coordinate System
So far the short discussion has been in symbolic notation2, that is, no reference to ‘axes’
or ‘components’ or ‘coordinates’ is made, implied or required. Vectors exist
independently of any coordinate system. The symbolic notation is very useful, but there
are many circumstances in which use of the component forms of vectors is more helpful
– or essential. To this end, introduce the vectors e1 , e 2 , e 3 having the properties
e1 ⋅ e 2 = e 2 ⋅ e 3 = e 3 ⋅ e1 = 0 , (7.1.2)
so that they are mutually perpendicular, and
e1 ⋅ e 1 = e 2 ⋅ e 2 = e 3 ⋅ e 3 = 1 , (7.1.3)
so that they are unit vectors. Such a set of orthogonal unit vectors is called an
orthonormal set, Fig. 7.1.1. This set of vectors forms a basis, by which is meant that any
other vector can be written as a linear combination of these vectors, i.e. in the form
a = a1e1 + a 2 e 2 + a3 e 3 (7.1.4)
where a1 , a 2 and a 3 are scalars, called the Cartesian components or coordinates of a

along the given three directions. The unit vectors are called base vectors when used for
1
or indicial or subscript or suffix notation
2
or absolute or invariant or direct or vector notation

Section 7.1
this purpose. The components a1 , a 2 and a 3 are measured along lines called the x1 , x2
and x3 axes, drawn through the base vectors.
a
e3
a3
e2
e1
a1
a2
Figure 7.1.1: an orthonormal set of base vectors and Cartesian coordinates
Note further that this orthonormal system {e1 , e 2 , e 3 } is right-handed, by which is meant
e1 × e 2 = e 3 (or e 2 × e 3 = e1 or e 3 × e1 = e 2 ).
In the index notation, the expression for the vector a in terms of the components a1 , a 2 , a3
and the corresponding basis vectors e1 , e 2 , e 3 is written as
3
a = a1e1 + a 2 e 2 + a3 e 3 = ∑ ai e i (7.1.5)
i =1
This can be simplified further by using Einstein’s summation convention, whereby the
summation sign is dropped and it is understood that for a repeated index (i in this case) a
summation over the range of the index (3 in this case3) is implied. Thus one writes
a = ai e i . This can be further shortened to, simply, ai .
The dot product of two vectors u and v, referred to this coordinate system, is
u ⋅ v = (u1e1 + u 2 e 2 + u 3 e 3 ) ⋅ (v1e1 + v 2 e 2 + v3 e 3 )
= u1v1 (e1 ⋅ e1 ) + u1v 2 (e1 ⋅ e 2 ) + u1v3 (e1 ⋅ e 3 )
+ u 2 v1 (e 2 ⋅ e1 ) + u 2 v 2 (e 2 ⋅ e 2 ) + u 2 v3 (e 2 ⋅ e 3 ) (7.1.6)
+ u 3 v1 (e 3 ⋅ e1 ) + u 3 v 2 (e 3 ⋅ e 2 ) + u 3 v3 (e 3 ⋅ e 3 )
= u1v1 + u 2 v 2 + u 3 v3
The dot product of two vectors written in the index notation reads
u ⋅ v = u i vi Dot Product (7.1.7)
3
2 in the case of a two-dimensional space/analysis

Section 7.1
The repeated index i is called a dummy index, because it can be replaced with any other
letter and the sum is the same; for example, this could equally well be written as
u ⋅ v = u j v j or u k v k .
Introduce next the Kronecker delta symbol δ ij , defined by
⎧0, i ≠ j
δ ij = ⎨ (7.1.8)
⎩1, i = j
Note that δ 11 = 1 but, using the index notation, δ ii = 3 . The Kronecker delta allows one
to write the expressions defining the orthonormal basis vectors (7.1.2, 7.1.3) in the
compact form
e i ⋅ e j = δ ij Orthonormal Basis Rule (7.1.9)
Example
Recall the equations of motion, Eqns. 1.1.9, which in full read
∂σ 11 ∂σ 12 ∂σ 13
+ + + b1 = ρa1
∂x1 ∂x 2 ∂x3
∂σ 21 ∂σ 22 ∂σ 23
+ + + b2 = ρa 2 (7.1.10)
∂x1 ∂x 2 ∂x3
∂σ 31 ∂σ 32 ∂σ 33
+ + + b3 = ρa3
∂x1 ∂x 2 ∂x3
The index notation for these equations is
∂σ ij
+ bi = ρai (7.1.11)
∂x j
Note the dummy index j. The index i is called a free index; if one term has a fee index i,
then, to be consistent, all terms must have it. One free index, as here, indicates three
separate equations.
7.1.2 Matrix Notation
The symbolic notation v and index notation vi e i (or simply vi ) can be used to denote a
vector. Another notation is the matrix notation: the vector v can be represented by a
3× 1 matrix (a column vector):

Section 7.1
⎡ v1 ⎤
⎢v ⎥
⎢ 2⎥
⎢⎣v3 ⎥⎦
Matrices will be denoted by square brackets, so a shorthand notation for this matrix/vector
would be [v ] . The elements of the matrix [v ] can be written in the index notation vi .
Note the distinction between a vector and a 3× 1 matrix: the former is a mathematical
object independent of any coordinate system, the latter is a representation of the vector in
a particular coordinate system – matrix notation, as with the index notation, relies on a
particular coordinate system.
As an example, the dot product can be written in the matrix notation as
⎡ v1 ⎤
[u ][v] = [u
T
1 u2 u 3 ]⎢⎢v 2 ⎥⎥
⎢⎣v3 ⎥⎦
“short”
matrix notation “full”
matrix notation
[ ]
Here, the notation u T denotes the 1× 3 matrix (the row vector). The result is a 1× 1
matrix, u i vi .
The matrix notation for the Kronecker delta δ ij is the identity matrix
⎡1 0 0 ⎤
[I ] = ⎢⎢0 1 0⎥⎥
⎢⎣0 0 1⎥⎦
Then, for example, in both index and matrix notation:
⎡1 0 0⎤ ⎡ u1 ⎤ ⎡ u1 ⎤
δ ij u j = u i [I ][u] = [u] ⎢0 1 0⎥ ⎢u ⎥ = ⎢u ⎥ (7.1.12)
⎢ ⎥⎢ 2 ⎥ ⎢ 2 ⎥
⎢⎣0 0 1⎥⎦ ⎢⎣u 3 ⎥⎦ ⎢⎣u 3 ⎥⎦
Matrix – Matrix Multiplication
When discussing vector transformation equations further below, it will be necessary to

multiply various matrices with each other (of sizes 3× 1 , 1× 3 and 3× 3 ). It will be
helpful to write these matrix multiplications in the short-hand notation.

Section 7.1
First, it has been seen that the dot product of two vectors can be represented by u T [v ] or[ ]
[ ]
u i vi . Similarly, the matrix multiplication [u ] v T gives a 3× 3 matrix with element form
u i v j or, in full,
⎡ u1v1 u1v 2 u1v3 ⎤

⎢u v u 2 v2 u 2 v3 ⎥⎥
⎢ 2 1
⎢⎣u 3 v1 u3v2 u 3 v3 ⎥⎦
This operation is called the tensor product of two vectors, written in symbolic notation
as u ⊗ v (or simply uv).
Next, the matrix multiplication
⎡Q11 Q12 Q13 ⎤ ⎡ u1 ⎤

[Q][u] ≡ ⎢⎢Q21 Q22 Q23 ⎥⎥ ⎢⎢u 2 ⎥⎥
⎢⎣Q31 Q32 Q33 ⎥⎦ ⎢⎣u 3 ⎥⎦
is a 3× 1 matrix with elements ([Q][u])i ≡ Qij u j . The elements of [Q][u ] are the same as
[ ][ ]
those of u T Q T , which can be expressed as ( u T Q T [ ][ ]) i ≡ u j Qij .
[ ]
The expression [u ][Q ] is meaningless, but u T [Q] {▲Problem 4} is a 1 × 3 matrix with
[ ]
elements ( u T [Q])i ≡ u j Q ji .
This leads to the following rule:
1. if a vector pre-multiplies a matrix [Q ] → the vector is the transpose u T [ ]

2. if a matrix [Q ] pre-multiplies the vector → the vector is [u ]
3. if summed indices are “beside each other”, as the j in u j Q ji or Qij u j
→ the matrix is [Q ]
4. if summed indices are not beside each other, as the j in u j Qij
→ the matrix is the transpose, Q T [ ]
Finally, consider the multiplication of 3× 3 matrices. Again, this follows the “beside
each other” rule for the summed index. For example, [A ][B ] gives the 3 × 3 matrix
[ ]
{▲Problem 8} ([A ][B])ij = Aik Bkj , and the multiplication A T [B ] is written as
([A ][B])
T
ij = Aki Bkj . There is also the important identity
([A][B])T [ ][ ]
= BT AT (7.1.13)
Note also the following:

Section 7.1
(i) if there is no free index, as in u i vi , there is one element

(ii) if there is one free index, as in u j Q ji , it is a 3 × 1 (or 1 × 3 ) matrix
(iii) if there are two free indices, as in Aki Bkj , it is a 3× 3 matrix
7.1.3 Vector Transformation Rule
Introduce two Cartesian coordinate systems with base vectors e i and e′i and common
origin o, Fig. 7.1.2. The vector u can then be expressed in two ways:
u = u i e i = u i′e′i (7.1.14)
x2
x′2
u1′
u2′
u
θ x1′ u2
e′2 e1′
θ
o x1
u1
Figure 7.1.2: a vector represented using two different coordinate systems
Note that the xi′ coordinate system is obtained from the xi system by a rotation of the
base vectors. Fig. 7.1.2 shows a rotation θ about the x3 axis (the sign convention for
rotations is positive counterclockwise).
Concentrating for the moment on the two dimensions x1 − x 2 , from trigonometry (refer to
Fig. 7.1.3),
u = u1e1 + u 2 e 2
= [ OB − AB ]e1 + [ BD + CP ]e 2 (7.1.15)
= [cos θ u1′ − sin θ u 2′ ]e1 + [sin θ u1′ + cos θ u 2′ ]e 2
and so
u1 = cos θ u1′ − sin θ u2′

(7.1.16)
u2 = sin θ u1′ + cos θ u2′
vector components in vector components in

first coordinate system second coordinate system
Section 7.1
x2
x2′
u1′ P
u′2
θ
x1′ u2
θ C
D
θ A B
x1
o
u1
Figure 7.1.3: geometry of the 2D coordinate transformation
In matrix form, these transformation equations can be written as
⎡ u1 ⎤ ⎡cos θ − sin θ ⎤ ⎡ u1′ ⎤

⎢u ⎥ = ⎢ sin θ cos θ ⎥⎦ ⎢⎣u ′2 ⎥⎦
(7.1.17)
⎣ 2⎦ ⎣
The 2 × 2 matrix is called the transformation matrix or rotation matrix [Q] . By pre-
multiplying both sides of these equations by the inverse of [Q] , Q −1 , one obtains the [ ]
transformation equations transforming from [u1 u 2 ] to [u1′ u 2′ ] :
T T
⎡ u1′ ⎤ ⎡ cos θ sin θ ⎤ ⎡ u1 ⎤

⎢u ′ ⎥ = ⎢− sin θ cos θ ⎥⎦ ⎢⎣u 2 ⎥⎦
(7.1.18)
⎣ 2⎦ ⎣
It can be seen that the components of [Q] are the directions cosines, i.e. the cosines of
the angles between the coordinate directions:
Qij = cos(xi , x ′j ) = e i ⋅ e′j (7.1.19)
It is straight forward to show that, in the full three dimensions, Fig. 7.1.4, the components
in the two coordinate systems are also related through
u i = Qij u ′j K [u] = [Q][u′]

[u′] = [Q T ][u]
Vector Transformation Rule (7.1.20)
u i′ = Q ji u j K

Section 7.1
x2
x2′
u
x1′
x1
x3′
x3
Figure 7.1.4: two different coordinate systems in a 3D space
Orthogonality of the Transformation Matrix [Q ]
From 7.1.20, it follows that
u i = Qij u ′j K [u] = [Q][u′]

(7.1.21)
= Qij Qkj u k K = [Q][Q T ][u ]
and so
Qij Qkj = δ ik K [Q][Q T ] = [I ] (7.1.22)
[ ] [ ]
A matrix such as this for which Q T = Q −1 is called an orthogonal matrix.
Example
Consider a Cartesian coordinate system with base vectors e i . A coordinate

transformation is carried out with the new basis given by
e1′ = a1(1) e1 + a 2(1) e 2 + a3(1) e 3

e′2 = a1( 2 ) e1 + a 2( 2) e 2 + a3( 2) e 3
e′3 = a1(3) e1 + a 2(3) e 2 + a3(3) e 3
What is the transformation matrix?
Solution
The transformation matrix consists of the direction cosines Qij = cos( xi , x ′j ) = e i ⋅ e′j , so

Section 7.1
⎡a1(1) a1( 2 ) a1(3) ⎤

[Q] = ⎢⎢a2(1) a 2( 2 )
⎥
a 2(3) ⎥
⎢a3(1) a3( 2 ) a3(3) ⎥⎦
⎣
7.1.4 Tensors
The concept of the tensor is discussed in detail in Part III, where it is indispensable for
the description of large-strain deformations. For small deformations, it is not so
necessary; the main purpose for introducing the tensor here (in a rather non-rigorous way)
is that it helps to deepen one’s understanding of the concept of stress.
A second-order tensor4 A may be defined as an operator that acts on a vector u

generating another vector v, so that T(u) = v , or
Tu = v Second-order Tensor (7.1.23)
The second-order tensor T is a linear operator, by which is meant
T(a + b ) = Ta + Tb … distributive
T(αa ) = α (Ta ) … associative
for scalar α . In a Cartesian coordinate system, the tensor T has nine components and can
be represented in the matrix form
⎡T11 T12 T13 ⎤

[T] = ⎢⎢T21 T22 T23 ⎥⎥
⎣⎢T31 T32 T33 ⎦⎥
The rule 7.1.23, which is expressed in symbolic notation, can be expressed in the index
and matrix notation when T is referred to particular axes:
⎡ u1 ⎤ ⎡T11 T12 T13 ⎤ ⎡ v1 ⎤

u i = Tij v j ⎢u ⎥ = ⎢T T23 ⎥⎥ ⎢⎢v 2 ⎥⎥ [u] = [T][v ] (7.1.24)
⎢ 2 ⎥ ⎢ 21 T22
⎢⎣u 3 ⎥⎦ ⎢⎣T31 T32 T33 ⎥⎦ ⎢⎣v3 ⎥⎦
Again, one should be careful to distinguish between a tensor such as T and particular
matrix representations of that tensor. The relation 7.1.23 is a tensor relation, relating
vectors and a tensor and is valid in all coordinate systems; the matrix representation of
this tensor relation, Eqn. 7.1.24, is to be sure valid in all coordinate systems, but the
entries in the matrices of 7.1.24 depend on the coordinate system chosen.
4
to be called simply a tensor in what follows

Section 7.1
Note also that the transformation formulae for vectors, Eqn. 7.1.20, is not a tensor
relation; although 7.1.20 looks similar to the tensor relation 7.1.24, the former relates the
components of a vector to the components of the same vector in different coordinate
systems, whereas (by definition of a tensor) the relation 7.1.24 relates the components of a
vector to those of a different vector in the same coordinate system.
For these reasons, the notation ui = Qij u ′ in Eqn. 7.1.20 is more formally called element
form, the Qij being elements of a matrix rather than components of a tensor. This
distinction between element form and index notation should be noted, but the term “index
notation” is used for both tensor and matrix-specific manipulations in these notes.
Example
Recall the strain-displacement relations, Eqns. 1.2.19, which in full read
∂u1 ∂u ∂u
ε 11 = , ε 22 = 2 , ε 33 = 3
∂x1 ∂x 2 ∂x3
(7.1.25)
1 ⎛ ∂u ∂u ⎞ 1 ⎛ ∂u ∂u ⎞ 1 ⎛ ∂u ∂u ⎞
ε 12 = ⎜⎜ 1 + 2 ⎟⎟, ε 13 = ⎜⎜ 1 + 3 ⎟⎟, ε 23 = ⎜⎜ 2 + 3 ⎟⎟
2 ⎝ ∂x 2 ∂x1 ⎠ 2 ⎝ ∂x3 ∂x1 ⎠ 2 ⎝ ∂x3 ∂x 2 ⎠
The index notation for these equations is
1 ⎛ ∂u ∂u j ⎞
ε ij = ⎜⎜ i + ⎟ (7.1.26)
2 ⎝ ∂x j ∂xi ⎟⎠
This expression has two free indices and as such indicates nine separate equations.
Further, with its two subscripts, ε ij , the strain, is a tensor. It can be expressed in the
matrix notation
⎡ ∂u1 / ∂x1 1
2
(∂u1 / ∂x2 + ∂u 2 / ∂x1 ) 1
2
(∂u1 / ∂x3 + ∂u 3 / ∂x1 ) ⎤
⎢
[ε] = ⎢ 2 (∂u 2 / ∂x1 + ∂u1 / ∂x2 )
1
∂u 2 / ∂x 2 1
(∂u 2 / ∂x3 + ∂u 3 / ∂x2 )⎥⎥
2
⎢⎣ 12 (∂u 3 / ∂x1 + ∂u1 / ∂x3 ) 1
2
(∂u3 / ∂x2 + ∂u 2 / ∂x3 ) ∂u 3 / ∂x3 ⎥⎦
7.1.5 Tensor Transformation Rule

Consider now the tensor definition 7.1.23 expressed in two different coordinate systems:
u i = Tij v j [u] = [T][v ] in {xi }

(7.1.27)
u i′ = Tij′v ′j [u′] = [T′][v ′] in {xi′}
From the vector transformation rule 7.1.20,

Section 7.1
u i′ = Q ji u j [u′] = [Q T ][u]
(7.1.28)
vi′ = Q ji v j [v ′] = [Q T ][v ]
Combining 7.1.27-28,
Q ji u j = Tij′Qkj v k [Q ][u] = [T′][Q ][v]

T T
(7.1.29)
and so
Qmi Q ji u j = Qmi Tij′Qkj v k [u] = [Q][T′][Q T ][v ] (7.1.30)
(Note that Qmi Q ji u j = δ mj u j = u m .) Comparing with 7.1.24, it follows that
Tij = Qip Q jq T pq′ K [T] = [Q][T′][Q T ]

Tensor Transformation Rule (7.1.31)
Tij′ = Q pi Qqj T pq K [T′] = [Q T ][T][Q]
7.1.6 Problems
1. Write the following in index notation: v , v ⋅ e1 , v ⋅ e k .

2. Show that δ ij ai b j is equivalent to a ⋅ b .
3. Evaluate or simplify the following expressions:
(a) δ kk (b) δ ij δ ij (c) δ ij δ jk
4. [ ]
Show that u T [Q] is a 1× 3 matrix with elements u j Q ji (write the matrices out in
full)
5. Show that ([Q][u]) = u T Q T
T
[ ][ ]
6. Are the three elements of [Q ][u ] the same as those of u T [Q] ? [ ]
7. What is the index notation for (a ⋅ b )c ?
8. Write out the 3 × 3 matrices [A ] and [B ] in full, i.e. in terms of A11 , A12 , etc. and
verify that [AB]ij = Aik Bkj for i = 2, j = 1 .
9. What is the index notation for
(a) [A ] B T[ ]
[ ] ([ ] )
(b) v T [A ][v ] (there is no ambiguity here, since v T [A ] [v ] = v T ([A ][v ]) ) [ ]
(c) [B ][A][B]
T
10. The angles between the axes in two coordinate systems are given in the table below.
x1 x2 x3
x1′ 135 o
60 o
120 o
x ′2 90 o 45 o 45 o
x3′ 45 o 60 o 120 o
Construct the corresponding transformation matrix [Q ] and verify that it is
orthogonal.

Section 7.1
11. Consider a two-dimensional problem. If the components of a vector u in one

coordinate system are
⎡ 2⎤
⎢ 3⎥
⎣ ⎦
what are they in a second coordinate system, obtained from the first by a positive
rotation of 30o? Sketch the two coordinate systems and the vector to see if your
answer makes sense.
12. Consider again a two-dimensional problem with the same change in coordinates as
in Problem 11. The components of a 2D tensor in the first system are
⎡1 − 1⎤
⎢3 2 ⎥
⎣ ⎦
What are they in the second coordinate system?

Section 7.2
7.2 Analysis of Three Dimensional Stress and Strain

The concept of traction and stress was introduced and discussed in Part I, §3.1-3.5. For
the most part, the discussion was confined to two-dimensional states of stress. Here, the
fully three dimensional stress state is examined. There will be some repetition of the
earlier analyses.
7.2.1 The Traction Vector and Stress Components

Consider a traction vector t acting on a surface element, Fig. 7.2.1. Introduce a Cartesian
coordinate system with base vectors e i so that one of the base vectors is a normal to the
surface and the origin of the coordinate system is positioned at the point at which the
traction acts. For example, in Fig. 7.1.1, the e 3 direction is taken to be normal to the
plane, and a superscript on t denotes this normal:
t ( e 3 ) = t 1e 1 + t 2 e 2 + t 3 e 3 (7.2.1)
Each of these components t i is represented by σ ij where the first subscript denotes the
direction of the normal and the second denotes the direction of the component to the
plane. Thus the three components of the traction vector shown in Fig. 7.2.1 are
σ 31 , σ 32 , σ 33 :
t (e3 ) = σ 31e1 + σ 32 e 2 + σ 33 e 3 (7.2.2)
The first two stresses, the components acting tangential to the surface, are shear stresses
whereas σ 33 , acting normal to the plane, is a normal stress.
x3
t (nˆ ) t (e3 )
σ 33
e3
e2
e1 x2
σ 32 σ 31
x1
Figure 7.2.1: components of the traction vector
Consider the three traction vectors t (e1 ) , t (e 2 ) , t (e 3 ) acting on the surface elements whose
outward normals are aligned with the three base vectors e j , Fig. 7.2.2a. The three (or six)
surfaces can be amalgamated into one diagram as in Fig. 7.2.2b.
In terms of stresses, the traction vectors are

Section 7.2
t (e1 ) = σ 11e 1 + σ 12 e 2 + σ 13 e 3
t (e 2 ) = σ 21e 1 + σ 22 e 2 + σ 23 e 3 or t (e i ) = σ ij e j (7.2.3)
t (e3 ) = σ 31e 1 + σ 32 e 2 + σ 33e 3
x3 x3 x3
t (e 2 ) (e 3 )
t
e3
t (e1 ) e2
(a ) e1 x2 x2 x2
x1 x1
x1
x1 = 0 x2 = 0 x3 = 0
x3
e3
t (e3 )
t (e 2 )
( b) e2
t (e1 )
e1 x2
x1
Figure 7.2.2: the three traction vectors acting at a point; (a) on mutually orthogonal
planes, (b) the traction vectors illustrated on a box element
The components of the three traction vectors, i.e. the stress components, can now be
displayed on a box element as in Fig. 7.2.3. Note that the stress components will vary
slightly over the surfaces of an elemental box of finite size. However, it is assumed that
the element in Fig. 7.2.3 is small enough that the stresses can be treated as constant, so
that they are the stresses acting at the origin.
x3
σ 33
σ 32
σ 31
σ 13 σ 23
σ 22
σ 11 σ 12 σ 21 x2
x1
Figure 7.2.3: the nine stress components with respect to a Cartesian coordinate
system
The nine stresses can be conveniently displayed in 3× 3 matrix form:

Section 7.2
⎡σ 11 σ 12 σ 13 ⎤
[σ ] = ⎢σ 21 σ 22 σ 23 ⎥ (7.2.4)
⎢σ ⎥
ij
⎣ 31 σ 32 σ 33 ⎦
It is important to realise that, if one were to take an element at some different orientation
to the element in Fig. 7.2.3, but at the same material particle, for example aligned with
the axes x1′ , x ′2 , x3′ shown in Fig. 7.2.4, one would then have different tractions acting and
the nine stresses would be different also. The stresses acting in this new orientation can
be represented by a new matrix:
⎡σ 11′ σ 12′ σ 13′ ⎤

[ ]
σ ij′ = ⎢σ 21
′ σ 22′ ′ ⎥
σ 23 (7.2.5)
⎢σ ′ σ ′ ′ ⎥⎦
σ 33
⎣ 31 32
x′2
′
σ 22
′
σ 21
σ 12′
′
σ 23
σ 11′
′ σ 31
σ 32 ′ x1′
σ 13′
′
σ 33
x3′
Figure 7.2.4: the stress components with respect to a Cartesian coordinate system
different to that in Fig. 7.2.3
7.2.2 Cauchy’s Law
Cauchy’s Law, which will be proved below, states that the normal to a surface, n = ni e i ,
is related to the traction vector t ( n ) = t i e i acting on that surface, according to
t i = σ ji n j (7.2.6)
Writing the traction and normal in vector form and the stress in 3× 3 matrix form,
⎡t1(n ) ⎤ ⎡σ 11 σ 12 σ 13 ⎤ ⎡ n1 ⎤
[t ]
(n )
= ⎢t 2(n ) ⎥,
⎢ (n ) ⎥
[σ ] = ⎢σ 21 σ 22
⎢σ
σ 23 ⎥, [ni ] = ⎢n2 ⎥ (7.2.7)
σ 33 ⎥⎦ ⎢n ⎥
i ij
⎢⎣t 3 ⎥⎦ ⎣ 31 σ 32 ⎣ 3⎦
and Cauchy’s law in matrix notation reads

Section 7.2
⎡t1(n ) ⎤ ⎡σ 11 σ 21 σ 31 ⎤ ⎡ n1 ⎤
⎢t (n ) ⎥ = ⎢σ σ 22 σ 32 ⎥ ⎢n2 ⎥ (7.2.8)
⎢ 2(n ) ⎥ ⎢ 12
⎣⎢t 3 ⎦⎥ ⎣ 13
σ σ 23 σ 33 ⎥⎦ ⎢⎣ n3 ⎥⎦
Note that it is the transpose stress matrix which is used in Cauchy’s law. Since the stress
matrix is symmetric, one can express Cauchy’s law in the form
t i = σ ij n j Cauchy’s Law (7.2.9)
Cauchy’s law is illustrated in Fig. 7.2.5; in this figure, positive stresses σ ij are shown.
x3
t (n )
(n )
σ 11 t3
σ 12 n
σ 21
σ 22
σ 13 t 2(n ) t1(n )
σ 23 σ 31
x2
σ 32
σ 33
x1
Figure 7.2.5: Cauchy’s Law; given the stresses and the normal to a plane, the
traction vector acting on the plane can be determined
Normal and Shear Stress
It is useful to be able to evaluate the normal stress σ N and shear stress σ S acting on any
plane, Fig. 7.2.6. For this purpose, note that the stress acting normal to a plane is the
projection of t (n ) in the direction of n ,
σ N = n ⋅ t (n ) (7.2.10)
The magnitude of the shear stress acting on the surface can then be obtained from
2
σS = t (n ) − σ N2 (7.2.11)

Section 7.2
x3
σN
t (n )
σS n
x2
x1
Figure 7.2.6: the normal and shear stress acting on an arbitrary plane through a
point
Example
The state of stress at a point with respect to a Cartesian coordinates system 0 x1 x 2 x3 is

given by:
⎡2 1 3⎤
[σ ]
ij = ⎢1 2 − 2⎥⎥
⎢
⎢⎣3 − 2 1 ⎥⎦
Determine:
(a) the traction vector acting on a plane through the point whose unit normal is
n = (1 / 3)e1 + (2 / 3)e 2 − (2 / 3)e 3
(b) the component of this traction acting perpendicular to the plane
(c) the shear component of traction on the plane
Solution
(a) From Cauchy’s law,
⎡t1(n ) ⎤ ⎡σ 11 σ 21 σ 31 ⎤ ⎡ n1 ⎤ 3 ⎤ ⎡ 1 ⎤ 1 ⎡ − 2⎤
⎢t (n ) ⎥ = ⎢σ ⎥ ⎢ ⎥ 1 ⎡2 1
σ σ n = ⎢ 1 2 − 2⎥ ⎢ 2 ⎥ = ⎢ 9 ⎥
⎢ 2(n ) ⎥ ⎢ 12 22 32
⎥ ⎢
2
⎥ 3 ⎢⎣3 − 2 1 ⎥⎦ ⎢⎣− 2⎥⎦ 3 ⎢⎣ − 3⎥⎦
⎢⎣t 3 ⎥⎦ ⎣ 13σ σ 23 σ 33 ⎦ ⎣ 3 ⎦
n
so that t ( n ) = (−2 / 3)e1 + 3e 2 − eˆ 3 .
(b) The component normal to the plane is

σ N = t (n ) ⋅ n = (−2 / 3)(1 / 3) + 3(2 / 3) + (2 / 3) = 22 / 9 ≈ 2.4.
(c) The shearing component of traction is

σS =
2
{[
t (n ) − σ N2 = (− 23 ) + (3) + (− 1) −
2 2 2
] [( ) ]}
22 2
9
1/ 2
≈ 2.1
■

Section 7.2
Proof of Cauchy’s Law
Cauchy’s law can be proved using force equilibrium of material elements. First, consider
a tetrahedral free-body, with vertex at the origin, Fig. 7.2.7. It is required to determine the
traction t in terms of the nine stress components (which are all shown positive in the
diagram).
x3
t (n )
σ 11 ΔS1
σ 12 n
σ 21 • α
σ 22 n
ΔS 2 •
σ 13 e1
σ 23 σ 31
x2
σ 32 ΔS 3
σ 33
x1
Figure 7.2.7: proof of Cauchy’s Law
The components of the unit normal, ni , are the direction cosines of the normal vector, i.e.
the cosines of the angles between the normal and each of the coordinate directions:
cos(n, e i ) = n ⋅ e i = ni (7.2.12)
Let the area of the base of the tetrahedran, with normal n, be ΔS . The area ΔS1 is then
ΔS cos α , where α is the angle between the planes, as shown to the right of Fig. 7.2.7;
this angle is the same as that between the vectors n and e1 , so ΔS1 = n1 ΔS , and similarly
for the other surfaces:
ΔS i = ni ΔS (7.2.13)
The resultant surface force on the body, acting in the xi direction, is then
∑F i = t i ΔS − σ ji ΔS j = t i ΔS − σ ji n j ΔS (7.2.14)
For equilibrium, this expression must be zero, and one arrives at Cauchy’s law.
Note:
As proved in Part III, this result holds also in the general case of accelerating material
elements in the presences of body forces.

Section 7.2
7.2.3 The Stress Tensor

Cauchy’s law 7.2.9 is of the same form as 7.1.24 and so by definition the stress is a
tensor. Denote the stress tensor in symbolic notation by σ . Cauchy’s law in symbolic
form then reads
t = σn (7.2.15)
Further, the transformation rule for stress follows the general tensor transformation rule
7.1.31:
σ ij = Qip Q jqσ ′pq K [σ ] = [Q][σ ′][Q T ]

Stress Transformation Rule (7.2.16)
σ ij′ = Q pi Qqj σ pq K [σ ′] = [Q T ][σ ][Q]
As with the normal and traction vectors, the components and hence matrix representation
of the stress changes with coordinate system, as with the two different matrix
representations 7.2.4 and 7.2.5. However, there is only one stress tensor σ at a point.
Another way of looking at this is to note that an infinite number of planes pass through a
point, and on each of these planes acts a traction vector, and each of these traction vectors
has three (stress) components. All of these traction vectors taken together define the
complete state of stress at a point.
Example
The state of stress at a point with respect to an 0 x1 x 2 x3 coordinate system is given by

1 ⎡2 0⎤
[σ ] = ⎢1
ij 3 − 2⎥
⎢⎣0 − 2 1 ⎥⎦
(a) What are the stress components with respect to axes 0 x1′ x 2′ x3′ which are obtained
from the first by a 45 o rotation (positive counterclockwise) about the x 2 axis, Fig.
7.2.8?
(b) Use Cauchy’s law to evaluate the normal and shear stress on a plane with normal
( ) ( )
n = 1 / 2 e1 + 1 / 2 e 3 and relate your result with that from (a)

Section 7.2
x3
x3′ e3
o
45
e′3
e1 x2 = x′2
e 2 = e′2
45o
x1 e1′
x1′
Figure 7.2.8: two different coordinate systems at a point
Solution
(a) The transformation matrix is
⎡ cos(x1 , x1′ ) cos( x1 , x ′2 ) cos(x1 , x3′ )⎤ ⎡ 12⎤ 0 1
⎥ ⎢
[ ]
2
Qij = ⎢⎢cos( x 2 , x1′ ) cos( x 2 , x 2′ ) cos(x 2 , x3′ )⎥⎥ = ⎢ 0
1 0⎥
⎢⎣cos(x3 , x1′ ) cos(x3 , x ′2 ) cos( x3 , x3′ )⎥⎦ ⎢− 12
0 12 ⎥⎦ ⎣
and QQ T = I as expected. The rotated stress components are therefore
⎡σ 11′ σ 12′ σ 13′ ⎤ ⎡ 12 0 − 12 ⎤ ⎡2 1 0 ⎤ ⎡ 12 0 12 ⎤
⎢σ ′ σ ′ σ ′ ⎥ = 0 1 ⎢ ⎥ ⎢ ⎥
⎢ 21 22 23 ⎥ ⎢ 0 ⎥ ⎢⎢1 3 − 2⎥⎥ ⎢ 0 1 0⎥
′ σ 32
⎢⎣σ 31 ′ σ 33′ ⎥⎦ ⎢ 12 0 1 ⎥
2 ⎦⎣
⎢0 − 2 1 ⎥⎦ ⎢− 12 0 12 ⎥
⎣ ⎣ ⎦
⎡ 32 3 1 ⎤
2
⎢ 2
⎥
= ⎢ 32 3 − 12 ⎥
⎢1 − 1 3 ⎥
⎣2 2 2 ⎦
and the new stress matrix is symmetric as expected.

(b) From Cauchy’s law, the traction vector is
⎡t1(n ) ⎤ ⎡2 1 0 ⎤ ⎡ 12 ⎤ ⎡ 2 ⎤
⎢ (n ) ⎥ ⎢ ⎥⎢ ⎥ ⎢ 1 ⎥
⎢t 2 ⎥ = ⎢1 3 − 2⎥ ⎢ 0 ⎥ = ⎢− 2 ⎥
⎢t 3(n ) ⎥ ⎢⎣0 − 2 1 ⎥⎦ ⎢ 1 ⎥ ⎢ 1 ⎥
⎣ ⎦ ⎣ 2⎦ ⎣ 2 ⎦
so that t (n ) = ( 2 )e − (1 / 2 )e + (1 / 2 )eˆ . The normal and shear stress on the
1 2 3
plane are
σ N = t (n ) ⋅ n = 3 / 2
and
2
σS = t (n ) − σ N2 = 3 − (3 / 2) 2 = 3 / 2
The normal to the plane is equal to e′3 and so σ N should be the same as σ 33
′ and it
is. The stress σ S should be equal to (σ 31′ )2 + (σ 32′ )2 and it is. The results are

Section 7.2
displayed in Fig. 7.2.9, in which the traction is represented in different ways, with
components (t1(n ) , t 2(n ) , t 3(n ) ) and (σ 31
′ , σ 32 ′ ).
′ , σ 33
σ N = σ 33
′ = 32
n = e′3
t (n)
t3(n ) = 1
σS x2 = x2′
2
′ = 12
σ 31
t2(n ) = − 1 ′ = − 32
σ 32
2
x1
t1(n ) = 2
x1′
Figure 7.2.9: traction and stresses acting on a plane
Isotropic State of Stress
Suppose the state of stress in a body is
⎡σ 0 0 0⎤
σ ij = σ 0δ ij [σ ] = ⎢⎢ 0 σ0 0 ⎥⎥ (7.2.17)
⎣⎢ 0 0 σ 0 ⎦⎥
One finds that the application of the stress tensor transformation rule yields the very same
components no matter what the new coordinate system{▲Problem 3}. In other words, no
shear stresses act, no matter what the orientation of the plane through the point. This is
termed an isotropic state of stress, or a spherical state of stress. One example of
isotropic stress is the stress arising in a fluid at rest, which cannot support shear stress, in
which case
[σ ] = − p[I ] (7.2.18)
where the scalar p is the fluid hydrostatic pressure. For this reason, an isotropic state of
stress is also referred to as a hydrostatic state of stress.
7.2.4 Principal Stresses

For certain planes through a material particle, there are traction vectors which act normal
to the plane, as in Fig. 7.2.10. In this case the traction can be expressed as a scalar
multiple of the normal vector, t ( n ) = σ n .

Section 7.2
t (n ) = σ n
no shear stress – only a normal

component to the traction
Figure 7.2.10: a purely normal traction vector
From Cauchy’s law then, for these planes,
⎡σ 11 σ 12 σ 13 ⎤ ⎡ n1 ⎤ ⎡ n1 ⎤
σ n = σ n, σ ij n j = σni , ⎢⎢σ 21 σ 22 ⎥ ⎢ ⎥
σ 23 ⎥ ⎢n2 ⎥ = σ ⎢⎢n2 ⎥⎥ (7.2.19)
⎢⎣σ 31 σ 32 σ 33 ⎥⎦ ⎢⎣ n3 ⎥⎦ ⎢⎣ n3 ⎥⎦
This is a standard eigenvalue problem from Linear Algebra: given a matrix σ ij , find [ ]
the eigenvalues σ and associated eigenvectors n such that Eqn. 7.2.19 holds.
To solve the problem, first re-write the equation in the form
⎧⎡σ 11 σ 12 σ 13 ⎤ ⎡1 0 0⎤ ⎫⎡ n1 ⎤ ⎡0⎤
(σ − σ I )n = 0, (σ ij − σδ ij )n j = 0, ⎨⎢σ 21 σ 22 σ 23 ⎥ − σ ⎢⎢0 1 0⎥⎥ ⎪⎬⎢⎢n2 ⎥⎥ = ⎢⎢0⎥⎥ (7.2.20)
⎪⎢ ⎥
⎪⎢σ ⎢⎣0 0 1⎥⎦ ⎪⎭⎢⎣ n3 ⎥⎦ ⎢⎣0⎥⎦
⎩⎣ 31 σ 32 σ 33 ⎥⎦
or
⎡σ 11 − σ σ 12 σ 13 ⎤ ⎡ n1 ⎤ ⎡0⎤
⎢ σ 21 σ 22 − σ σ 23 ⎥ ⎢n2 ⎥ = ⎢0⎥ (7.2.21)
⎢ σ σ 32 σ 33 − σ ⎥⎦ ⎢⎣ n3 ⎥⎦ ⎢⎣0⎥⎦
⎣ 31
This is a set of three homogeneous equations in three unknowns (if one treats σ as
known). From basic linear algebra, this system has a solution (apart from ni = 0 ) if and
only if the determinant of the coefficient matrix is zero, i.e. if
⎡σ 11 − σ σ 12 σ 13 ⎤
det(σ − σI) = det ⎢ σ 21 σ 22 − σ σ 23 ⎥ = 0 (7.2.22)
⎢ σ σ 32 σ 33 − σ ⎥⎦
⎣ 31
Evaluating the determinant, one has the following cubic characteristic equation of the
stress tensor σ ,
σ 3 − I 1σ 2 + I 2σ − I 3 = 0 Characteristic Equation (7.2.23)
and the principal scalar invariants of the stress tensor are

Section 7.2
I 1 = σ 11 + σ 22 + σ 33
I 2 = σ 11σ 22 + σ 22σ 33 + σ 33σ 11 − σ 122 − σ 23
2
− σ 312 (7.2.24)
I 3 = σ 11σ 22σ 33 − σ 11σ 23
2
− σ 22σ 312 − σ 33σ 122 + 2σ 12σ 23σ 31
( I 3 is the determinant of the stress matrix.) The characteristic equation 7.2.23 can now be
solved for the eigenvalues σ and then Eqn. 7.2.21 can be used to solve for the
eigenvectors n.
Now another theorem of linear algebra states that the eigenvalues of a real (that is, the
components are real), symmetric matrix (such as the stress matrix) are all real and further
that the associated eigenvectors are mutually orthogonal. This means that the three roots
of the characteristic equation are real and that the three associated eigenvectors form a
mutually orthogonal system. This is illustrated in Fig. 7.2.11; the eigenvalues are called
principal stresses and are labelled σ 1 , σ 2 , σ 3 and the three corresponding eigenvectors
are called principal directions, the directions in which the principal stresses act. The
planes on which the principal stresses act (to which the principal directions are normal)
are called the principal planes.
2
σ2
σ1
1
σ3
Figure 7.2.11: the three principal stresses acting at a point and the three associated
principal directions 1, 2 and 3
Once the principal stresses are found, as mentioned, the principal directions can be found
by solving Eqn. 7.2.21, which can be expressed as
(σ 11 − σ )n1 + σ 12 n2 + σ 13 n3 = 0
σ 21 n1 + (σ 22 − σ )n2 + σ 23 n3 = 0 (7.2.25)
σ 31 n1 + σ 32 n2 + (σ 33 − σ )n3 = 0
Each principal stress value in this equation gives rise to the three components of the
associated principal direction vector, n1 , n2 , n3 . The solution also requires that the
magnitude of the normal be specified: for a unit vector, n ⋅ n = 1 . The directions of the
normals are also chosen so that they form a right-handed set.

Section 7.2
Example
The stress at a point is given with respect to the axes Ox1 x 2 x3 by the values
0 0 ⎤ ⎡5
ij [σ ] = ⎢0
− 6 − 12⎥ .
⎢⎣0 − 12 1 ⎥⎦
Determine (a) the principal values, (b) the principal directions (and sketch them).
Solution:
(a)
The principal values are the solution to the characteristic equation
5 −σ 0 0
0 − 6 − σ − 12 = (−10 + σ )(5 − σ )(15 + σ ) = 0
0 − 12 1−σ
which yields the three principal values σ 1 = 10, σ 2 = 5, σ 3 = −15 .
(b)
The eigenvectors are now obtained from Eqn. 7.2.25. First, for σ 1 = 10 ,
− 5n1 + 0n2 + 0n3 = 0
0n1 − 16n2 − 12n3 = 0
0n1 − 12n2 − 9n3 = 0
and using also the equation n + n + n32 = 1 leads to n1 = −(3 / 5)e 2 + (4 / 5)e 3 . Similarly,
2
1
2
2
for σ 2 = 5 and σ 3 = −15 , one has, respectively,

0n + 0n + 0n = 0 20n1 + 0n2 + 0n3 = 0
1 2 3
0n1 − 11n2 − 12n3 = 0 and 0n1 + 9n2 − 12n3 = 0
0n1 − 12n2 − 4n3 = 0 0n1 − 12n2 + 16n3 = 0
which yield n 2 = e1 and n 3 = (4 / 5)e 2 + (3 / 5)e 3 . The principal directions are sketched in
Fig. 7.2.12. Note that the three components of each principal direction, n1 , n2 , n3 , are the
direction cosines: the cosines of the angles between that principal direction and the three
coordinate axes. For example, for σ 1 with n1 = 0, n2 = −3 / 5, n3 = 4 / 5 , the angles made
with the coordinate axes x1 , x 2 , x3 are, respectively, 0, 127o and 37o.
x3
n̂1
n̂3
37o
x2
x1 n̂ 2
Figure 7.2.12: principal directions

■

Section 7.2
Invariants
The principal stresses σ 1 , σ 2 , σ 3 are independent of any coordinate system; the 0 x1 x 2 x3

axes to which the stress matrix in Eqn. 7.2.19 is referred can have any orientation – the
same principal stresses will be found from the eigenvalue analysis. This is expressed by
using the symbolic notation for the problem: σ n = σ n , which is independent of any
coordinate system. Thus the principal stresses are intrinsic properties of the stress state at
a point. It follows that the functions I 1 , I 2 , I 3 in the characteristic equation Eqn. 7.2.23
are also independent of any coordinate system, and hence the name principal scalar
invariants (or simply invariants) of the stress.
The stress invariants can also be written neatly in terms of the principal stresses:
I1 = σ 1 + σ 2 + σ 3
I 2 = σ 1σ 2 + σ 2σ 3 + σ 3σ 1 (7.2.26)
I 3 = σ 1σ 2σ 3
Also, if one chooses a coordinate system to coincide with the principal directions, Fig.
7.2.12, the stress matrix takes the simple form
⎡σ 1 0 0⎤
[σ ] = ⎢ 0 σ2 0⎥ (7.2.27)
⎢0 σ 3 ⎥⎦
ij
⎣ 0
Note that when two of the principal stresses are equal, one of the principal directions will
be unique, but the other two will be arbitrary – one can choose any two principal
directions in the plane perpendicular to the uniquely determined direction, so that the
three form an orthonormal set. This stress state is called axi-symmetric. When all three
principal stresses are equal, one has an isotropic state of stress, and all directions are
principal directions – the stress matrix has the form 7.2.27 no matter what orientation the
planes through the point.
Example
The two stress matrices from the Example of §7.2.3, describing the stress state at a point
with respect to different coordinate systems, are
⎡2 1 0⎤ ⎡ 3/ 2 3/ 2 1/ 2 ⎤
[ ] ⎢
σ ij = ⎢1 3 − 2⎥ , ⎥ [ ]
⎢
σ ij′ = 3 / 2
⎢
3 − 1/ 2 ⎥
⎥
⎢⎣0 − 2 1 ⎥⎦ ⎢⎣ 1 / 2 − 1 / 2 3 / 2 ⎥⎦
The first invariant is the sum of the normal stresses, the diagonal terms, and is the same
for both as expected:
I 1 = 2 + 3 + 1 = 32 + 3 + 32 = 6
The other invariants can also be obtained from either matrix, and are
I 2 = 6, I 3 = −3
■

Section 7.2
7.2.5 Maximum and Minimum Stress Values
Normal Stresses
The three principal stresses include the maximum and minimum normal stress
components acting at a point. To prove this, first let e1 , e 2 , e 3 be unit vectors in the
principal directions. Consider next an arbitrary unit normal vector n = ni e i . From
Cauchy’s law (see Fig. 7.2.13 – the stress matrix in Cauchy’s law is now with respect to
the principal directions 1, 2 and 3), the normal stress acting on the plane with normal n is
σ N = t ( n ) ⋅ n = (σ n ) ⋅ n, σ N = σ ij n j ni (7.2.28)
σN
t (n )
principal
1 directions
Figure 7.2.13: normal stress acting on a plane defined by the unit normal n
Thus
⎧⎡σ 1 0 0 ⎤ ⎡ n1 ⎤ ⎫⎡ n1 ⎤
⎪ ⎪
σ N = ⎨⎢ 0 σ 2 0 ⎥ ⎢n2 ⎥ ⎬⎢n2 ⎥ = σ 1 n12 + σ 2 n 22 + σ 3 n32 (7.2.29)
⎪⎩⎢⎣ 0 0 σ 3 ⎥⎦ ⎢⎣ n3 ⎥⎦ ⎪⎭⎢⎣ n3 ⎥⎦
Since n12 + n 22 + n32 = 1 and, without loss of generality, taking σ 1 ≥ σ 2 ≥ σ 3 , one has
σ 1 = σ 1 (n12 + n22 + n32 ) ≥ σ 1 n12 + σ 2 n22 + σ 3 n32 = σ N (7.2.30)
Similarly,
σ N = σ 1 n12 + σ 2 n22 + σ 3 n32 ≥ σ 3 (n12 + n22 + n32 ) ≥ σ 3 (7.2.31)
Thus the maximum normal stress acting at a point is the maximum principal stress and the
minimum normal stress acting at a point is the minimum principal stress.

Section 7.2
Shear Stresses
Next, it will be shown that the maximum shearing stresses at a point act on planes
oriented at 45o to the principal planes and that they have magnitude equal to half the
difference between the principal stresses. First, again, let e1 , e 2 , e 3 be unit vectors in the
principal directions and consider an arbitrary unit normal vector n = ni e i . The normal
stress is given by Eqn. 7.2.29,
σ N = σ 1 n12 + σ 2 n22 + σ 3 n32 (7.2.32)
Cauchy’s law gives the components of the traction vector as
⎡t1(n ) ⎤ ⎡σ 1 0 0 ⎤ ⎡ n1 ⎤ ⎡ σ 1 n1 ⎤
⎢t (n ) ⎥ = ⎢ 0 σ 0 ⎥ ⎢n2 ⎥ = ⎢σ 2 n 2 ⎥ (7.2.33)
⎢ 1(n ) ⎥ ⎢ 2
0 0 σ 3 ⎥⎦ ⎢⎣ n3 ⎥⎦ ⎢⎣σ 3 n3 ⎥⎦
⎣⎢t1 ⎦⎥ ⎣
and so the shear stress on the plane is, from Eqn. 7.2.11,
σ S2 = (σ 12 n12 + σ 22 n22 + σ 32 n32 ) − (σ 1n12 + σ 2 n22 + σ 3 n32 )

2
(7.2.34)
Using the condition n12 + n22 + n32 = 1 to eliminate n3 leads to
σ S2 = (σ 12 − σ 32 )n12 + (σ 22 − σ 32 )n22 + σ 32 − [(σ 1 − σ 3 )n12 + (σ 2 − σ 3 )n22 + σ 3 ]

2
(7.2.35)
The stationary points are now obtained by equating the partial derivatives with respect to
the two variables n1 and n2 to zero:
( )
∂ σ S2
{ [
= n1 (σ 1 − σ 3 ) σ 1 − σ 3 − 2 (σ 1 − σ 3 )n12 + (σ 2 − σ 3 )n22 = 0 ]}
∂n1
( )
∂ σ S2
{ [
= n2 (σ 2 − σ 3 ) σ 2 − σ 3 − 2 (σ 1 − σ 3 )n12 + (σ 2 − σ 3 )n 22 = 0 ]}
(7.2.36)
∂n 2
One sees immediately that n1 = n2 = 0 (so that n3 = ±1 ) is a solution; this is the principal
direction e 3 and the shear stress is by definition zero on the plane with this normal. In
this calculation, the component n3 was eliminated and σ S2 was treated as a function of the
variables (n1 , n2 ) . Similarly, n1 can be eliminated with (n 2 , n3 ) treated as the variables,
leading to the solution n = e1 , and n 2 can be eliminated with (n1 , n3 ) treated as the
variables, leading to the solution n = e 2 . Thus these solutions lead to the minimum shear
stress value σ S2 = 0 .
A second solution to Eqn. 7.2.36 can be seen to be n1 = 0, n 2 = ±1 / 2 (so that

n3 = ±1 / 2 ) with corresponding shear stress values σ S2 = 1
4
(σ 2 − σ 3 )2 . Two other

Section 7.2
solutions can be obtained as described earlier, by eliminating n1 and by eliminating n 2 .

The full solution is listed below, and these are evidently the maximum (absolute value of
the) shear stresses acting at a point:
⎛ 1 1 ⎞ 1
n = ⎜ 0, ± ,± ⎟, σ S = σ 2 − σ 3
⎝ 2 2⎠ 2
⎛ 1 1 ⎞ 1
n = ⎜± , 0, ± ⎟, σ S = σ 3 − σ 1 (7.2.37)
⎝ 2 2⎠ 2
⎛ 1 1 ⎞ 1
n = ⎜± ,± , 0 ⎟, σ S = σ 1 − σ 2
⎝ 2 2 ⎠ 2
Taking σ 1 ≥ σ 2 ≥ σ 3 , the maximum shear stress at a point is
1
τ max = (σ 1 − σ 3 ) (7.2.38)
2
and acts on a plane with normal oriented at 45o to the 1 and 3 principal directions. This is
illustrated in Fig. 7.2.14.
principal
directions
τ max
1
τ max
Figure 7.2.14: maximum shear stress at a point
Example
Consider the stress state examined in the Example of §7.2.4:

⎡5 0 0 ⎤
[ ]
σ ij = ⎢0 − 6 − 12⎥
⎣⎢0 − 12 1 ⎦⎥
The principal stresses were found to be σ 1 = 10, σ 2 = 5, σ 3 = −15 and so the maximum
shear stress is
1 25
τ max = (σ 1 − σ 3 ) =
2 2
One of the planes upon which they act is shown in Fig. 7.2.15 (see Fig. 7.2.12)

Section 7.2
x3 τ max
n̂1
n̂3
37o
x2
x1 n̂ 2
Figure 7.2.15: maximum shear stress
7.2.6 Mohr’s Circles of Stress

The Mohr’s circle for 2D stress states was discussed in Part I, §3.5.4. For the 3D case,
following on from section 7.2.5, one has the conditions
σ N = σ 1 n12 + σ 2 n22 + σ 3 n32

σ + σ N2 = σ 12 n12 + σ 22 n22 + σ 32 n32
2
S (7.2.39)
n + n + n =1
2
1
2
2
2
3
Solving these equations gives
(σ N − σ 2 )(σ N − σ 3 ) + σ S2
n12 =
(σ 1 − σ 2 )(σ 1 − σ 3 )
(σ N − σ 3 )(σ N − σ 1 ) + σ S2
n2 =
2
(7.2.40)
(σ 2 − σ 3 )(σ 2 − σ 1 )
(σ − σ 1 )(σ N − σ 2 ) + σ S2
n32 = N
(σ 3 − σ 1 )(σ 3 − σ 2 )
Taking σ 1 ≥ σ 2 ≥ σ 3 , and noting that the squares of the normal components must be
positive, one has that
(σ N − σ 2 )(σ N − σ 3 ) + σ S2 ≥ 0
(σ N − σ 3 )(σ N − σ 1 ) + σ S2 ≤ 0 (7.2.41)
(σ N − σ 1 )(σ N − σ 2 ) + σ S2 ≥ 0
and these can be re-written as
σ S2 + [σ N − 12 (σ 2 + σ 3 )] ≥ [12 (σ 2 − σ 3 )]
2 2
σ S2 + [σ N − 12 (σ 1 + σ 3 )] ≤ [12 (σ 1 − σ 3 )]
2 2
(7.2.42)
σ S2 + [σ N − 12 (σ 1 + σ 2 )] ≥ [12 (σ 1 − σ 2 )]
2 2

Section 7.2
If one takes coordinates (σ N , σ S ) , the equality signs here represent circles in (σ N , σ S )

stress space, Fig. 7.2.16. Each point (σ N , σ S ) in this stress space represents the stress on
a particular plane through the material particle in question. Admissible (σ N , σ S ) pairs are
given by the conditions Eqns. 7.2.42; they must lie inside a circle of centre ( 12 (σ 1 + σ 3 ),0 )
and radius 12 (σ 1 − σ 3 ) . This is the large circle in Fig. 7.2.16. The points must lie outside
the circle with centre ( 12 (σ 2 + σ 3 ),0 ) and radius 12 (σ 2 − σ 3 ) and also outside the circle
with centre ( 12 (σ 1 + σ 2 ),0 ) and radius 12 (σ 1 − σ 2 ) ; these are the two smaller circles in the
figure. Thus the admissible points in stress space lie in the shaded region of Fig. 7.2.16.
σS
σ 3• •
σ2 •
σ1 σN
Figure 7.2.16: admissible points in stress space
The strain ε ij , in symbolic form ε , is a tensor and as such it follows the same rules as for
the stress tensor. In particular, it follows the general tensor transformation rule 7.2.16; it
has principal values ε which satisfy the characteristic equation 7.2.23 and these include
the maximum and minimum normal strain at a point. There are three principal strain
invariants given by 7.2.24 or 7.2.26 and the maximum shear strain occurs on planes
oriented at 45o to the principal directions.
7.2.8 Problems
1. The state of stress at a point with respect to a 0 x1 x 2 x3 coordinate system is given by

1 ⎡22⎤
[σ ] = ⎢1
ij0 − 1⎥
⎢⎣2 − 1 − 2⎥⎦
Use Cauchy’s law to determine the traction vector acting on a plane trough this
point whose unit normal is n = (e1 + e 2 + e 3 ) / 3 . What is the normal stress acting
on the plane? What is the shear stress acting on the plane?

⎡1 3 2 ⎤
[σ ] = ⎢3
ij 1 0⎥
⎢⎣2 0 − 2⎥⎦

Section 7.2
What are the stress components with respect to axes 0 x1′ x 2′ x3′ which are obtained
from the first by a 45 o rotation (positive counterclockwise) about the x3 axis
3. Show, in both the index and matrix notation, that the components of an isotropic
stress state remain unchanged under a coordinate transformation.
4. Consider a two-dimensional problem. The stress transformation formulae are then,

in full,
⎡σ 11′ σ 12′ ⎤ ⎡ cos θ sin θ ⎤ ⎡σ 11 σ 12 ⎤ ⎡cos θ − sin θ ⎤
⎢σ ′ σ ′ ⎥ = ⎢− sin θ cos θ ⎥ ⎢σ ⎥⎢ ⎥
⎣ 21 22 ⎦ ⎣ ⎦ ⎣ 21 σ 22 ⎦ ⎣ sin θ cos θ ⎦
Multiply the right hand side out and use the fact that the stress tensor is symmetric
( σ 12 = σ 21 - not true for all tensors). What do you get? Look familiar?

⎡ 5/ 2 − 1 / 2 0⎤
[σ ] = ⎢− 1 / 2
ij 5 / 2 0⎥
⎢⎣ 0 0 1⎥⎦
Evaluate the principal stresses and the principal directions. What is the maximum
shear stress acting at the point?

Section 7.3
7.3 Governing Equations of Three Dimensional

Elasticity
7.3.1 Hooke’s Law and Lamé’s Constants

Linear elasticity was introduced in Part I, §4.2. The three-dimensional Hooke’s law for
isotropic linear elastic solids (Part I, Eqns. 4.2.9) can be expressed in index notation as
σ ij = λδ ij ε kk + 2με ij (7.3.1)
where (see also Part I, Eqns. 6.2.21)
Eν E
λ= , μ= (7.3.2)
(1 + ν )(1 − 2ν ) 2(1 + ν )
are the Lamé constants ( μ is the Shear Modulus). Eqns. 7.3.1 can be inverted to obtain
{▲Problem 1}
1 λ
ε ij = σ ij − δ ij σ kk (7.3.3)
2μ 2 μ (3λ + 2 μ )
7.3.2 Navier’s Equations

The governing equations of elasticity are Hooke’s law (Eqn. 7.3.1), the equations of
motion, Eqn. 1.1.9 (see Eqns. 7.1.10-11),
∂σ ij
+ bi = ρai (7.3.4)
∂x j
and the strain-displacement relations, Eqn. 1.2.19 (see Eqns. 7.1.25-26),
1 ⎛⎜ ∂u i ∂u j ⎞
⎟
ε ij = + (7.3.5)
2 ⎜⎝ ∂x j ∂xi ⎟
⎠
Substituting 7.3.5 into 7.3.1 and then into 7.3.4 leads to the 3D Navier’s equations
{▲Problem 2}
∂ 2u j ∂ 2 ui
(λ + μ ) +μ + bi = ρai Navier’s Equations (7.3.6)
∂x j ∂xi ∂x j ∂x j
These reduce to the 2D plane strain Navier’s equations, Eqns. 3.1.4, by setting u 3 = 0 and
∂ / ∂x3 = 0 . They do not reduce to the plane stress equations since the latter are only an

Section 7.3
approximate solution to the equations of elasticity which are valid only in the limit as the
thickness of the thin plate of plane stress tends to zero.
7.3.3 Problems
1. Invert Eqns. 7.3.1 to get 7.3.3.
2. Derive the 3D Navier’s equations from 7.3.6 from 7.3.1, 7.3.4 and 7.3.5

Answers to Selected Problems: Part II, Chapter 1
1.1
2. Yes
3. One of the equations of equilibrium is not satisfied.
⎡− 3 x 2 ⎤
4. b = ⎢⎢ − x3 ⎥⎥
⎢⎣ − x1 ⎥⎦
⎡ 403 x1 (2 − x12 ) ⎤
⎢ ⎥
5. a = ⎢ 403 x 2 (2 − x 22 ) ⎥
⎢ 403 x3 (2 − x32 ) − 9.81⎥
⎣ ⎦
1.2
1. ε xx = 3 A, ε yy = 2 Axy, ε xy = 12 A(− 1 + y 2 ), ω z = 12 A( y 2 + 1)
2. u x = 12 αx 2 + A − Cy , u y = 2αx + B + Cx, ω z = α + C
u x = 12 Ax 2 y + C 3 − C 4 y, u y = 13 Ay 3 + Ax 2 − 16 Ax 3 + C 2 + C 4 x,
3.
ω z = Ax − 18 Ax 2 + C 4
1.3
1. ε xx = Ay, ε yy = 2 Ay, ε xy = 12 Ax, ω z = 12 Ax, A

2
[3 y ± x2 + y2 ]
A1
Answers to Selected Problems: Part II, Chapter 7
7.1
1. vi vi , v1 , v k
3. 3, 3, δ ik
6. No
7. ai bi c j
9. Aij Bkj , vi Aij v j , B ji A jk Bkl
⎡− 1 / 2 0 1/ 2 ⎤
⎢ ⎥
10. ⎢ 1/ 2 1/ 2 1/ 2 ⎥
⎢ − 1 / 2 1 / 2 − 1 / 2⎥
⎣ ⎦
⎡ 3 + 3/ 2 ⎤
11. ⎢ ⎥
⎣ − 1 + 3 3 / 2⎦
1 ⎡5 + 2 3 3−6 ⎤
12. ⎢ ⎥
4 ⎣ 3 + 10 7 − 2 3 ⎦
7.2
1. vi vi , v1 , v k
3.
Section 7.2
1. σ N = 4 / 3 , σ S ≈ 2.62
⎡ 4 0 2 ⎤
⎢ ⎥
2. ⎢0 − 2 − 2⎥
⎢ 2 − 2 −2 ⎥
⎣ ⎦
4. The 2D stress transformation equations
5. σ i = 1,2,3
1 1 1 1
n1 = e1 − e2 , n2 = e1 + e 2 , n3 = e3
2 2 2 2
τ max = 1
A7
1 Vectors & Tensors
The mathematical modeling of the physical world requires knowledge of quite a few
different mathematics subjects, such as Calculus, Differential Equations and Linear
Algebra. These topics are usually encountered in fundamental mathematics courses.
However, in a more thorough and in-depth treatment of mechanics, it is essential to
describe the physical world using the concept of the tensor, and so we begin this book
with a comprehensive chapter on the tensor.
The chapter is divided into three parts. The first part covers vectors (§1.1-1.7). The
second part is concerned with second, and higher-order, tensors (§1.8-1.15). The second
part covers much of the same ground as done in the first part, mainly generalizing the
vector concepts and expressions to tensors. The final part (§1.16-1.19) (not required in
the vast majority of applications) is concerned with generalizing the earlier work to
curvilinear coordinate systems.
The first part comprises basic vector algebra, such as the dot product and the cross
product; the mathematics of how the components of a vector transform between different
coordinate systems; the symbolic, index and matrix notations for vectors; the
differentiation of vectors, including the gradient, the divergence and the curl; the
integration of vectors, including line, double, surface and volume integrals, and the
integral theorems.
The second part comprises the definition of the tensor (and a re-definition of the vector);
dyads and dyadics; the manipulation of tensors; properties of tensors, such as the trace,
transpose, norm, determinant and principal values; special tensors, such as the spherical,
identity and orthogonal tensors; the transformation of tensor components between
different coordinate systems; the calculus of tensors, including the gradient of vectors and
higher order tensors and the divergence of higher order tensors and special fourth order
tensors.
In the first two parts, attention is restricted to rectangular Cartesian coordinates (except
for brief forays into cylindrical and spherical coordinates). In the third part, curvilinear
coordinates are introduced, including covariant and contravariant vectors and tensors, the
metric coefficients, the physical components of vectors and tensors, the metric, coordinate
transformation rules, tensor calculus, including the Christoffel symbols and covariant
differentiation, and curvilinear coordinates for curved surfaces.
1
2
Section 1.1
1.1 Vector Algebra
1.1.1 Scalars
A physical quantity which is completely described by a single real number is called a
scalar. Physically, it is something which has a magnitude, and is completely described
by this magnitude. Examples are temperature, density and mass. In the following,
lowercase (usually Greek) letters, e.g.  ,  ,  , will be used to represent scalars.
1.1.2 Vectors
The concept of the vector is used to describe physical quantities which have both a
magnitude and a direction associated with them. Examples are force, velocity,
displacement and acceleration.
Geometrically, a vector is represented by an arrow; the arrow defines the direction of the
vector and the magnitude of the vector is represented by the length of the arrow, Fig.
1.1.1a.
Analytically, vectors will be represented by lowercase bold-face Latin letters, e.g. a, r, q.
The magnitude (or length) of a vector is denoted by a or a. It is a scalar and must be

non-negative. Any vector whose length is 1 is called a unit vector; unit vectors will
usually be denoted by e.
a
b
(a) (b)
Figure 1.1.1: (a) a vector; (b) addition of vectors
1.1.3 Vector Algebra

The operations of addition, subtraction and multiplication familiar in the algebra of
numbers (or scalars) can be extended to an algebra of vectors.
Solid Mechanics Part III 3 Kelly

Section 1.1
The following definitions and properties fundamentally define the vector:

1. Sum of Vectors:
The addition of vectors a and b is a vector c formed by placing the initial point of
b on the terminal point of a and then joining the initial point of a to the terminal
point of b. The sum is written c  a  b . This definition is called the
parallelogram law for vector addition because, in a geometrical interpretation of
vector addition, c is the diagonal of a parallelogram formed by the two vectors a
and b, Fig. 1.1.1b. The following properties hold for vector addition:
ab ba … commutative law
a  b  c  a  b  c … associative law
2. The Negative Vector:

For each vector a there exists a negative vector. This vector has direction
opposite to that of vector a but has the same magnitude; it is denoted by  a . A
geometrical interpretation of the negative vector is shown in Fig. 1.1.2a.
3. Subtraction of Vectors and the Zero Vector:

The subtraction of two vectors a and b is defined by a  b  a  (b) , Fig.
1.1.2b. If a  b then a  b is defined as the zero vector (or null vector) and is
represented by the symbol o. It has zero magnitude and unspecified direction. A
proper vector is any vector other than the null vector. Thus the following
properties hold:
ao a
a   a   o
4. Scalar Multiplication:
The product of a vector a by a scalar  is a vector a with magnitude  times
the magnitude of a and with direction the same as or opposite to that of a,
according as  is positive or negative. If   0 , a is the null vector. The
following properties hold for scalar multiplication:
   a  a  a … distributive law, over addition of scalars
 a  b   a  b … distributive law, over addition of vectors
 ( a)  ( )a … associative law for scalar multiplication
a b
a
b
a
ab
a
(a) (b)
Figure 1.1.2: (a) negative of a vector; (b) subtraction of vectors

Section 1.1
Note that when two vectors a and b are equal, they have the same direction and
magnitude, regardless of the position of their initial points. Thus a  b in Fig. 1.1.3. A
particular position in space is not assigned here to a vector – it just has a magnitude and a
direction. Such vectors are called free, to distinguish them from certain special vectors to
which a particular position in space is actually assigned.
a
b
Figure 1.1.3: equal vectors
The vector as something with “magnitude and direction” and defined by the above rules is
an element of one case of the mathematical structure, the vector space. The vector space
is discussed in the next section, §1.2.
1.1.4 The Dot Product

The dot product of two vectors a and b (also called the scalar product) is denoted by
a  b . It is a scalar defined by
a  b  a b cos . (1.1.1)
 here is the angle between the vectors when their initial points coincide and is restricted
to the range 0     , Fig. 1.1.4.
a
a
b 
Figure 1.1.4: the dot product
An important property of the dot product is that if for two (proper) vectors a and b, the
relation a  b  0 , then a and b are perpendicular. The two vectors are said to be
orthogonal. Also, a  a  a a cos(0) , so that the length of a vector is a  a  a .
Another important property is that the projection of a vector u along the direction of a
unit vector e is given by u  e . This can be interpreted geometrically as in Fig. 1.1.5.

Section 1.1

e
u  e  u cos
Figure 1.1.5: the projection of a vector along the direction of a unit vector
It follows that any vector u can be decomposed into a component parallel to a (unit)
vector e and another component perpendicular to e, according to
u  u  e e  u  u  e e (1.1.2)
The dot product possesses the following properties (which can be proved using the above
definition) {▲Problem 6}:
(1) a  b  b  a (commutative)
(2) a  b  c   a  b  a  c (distributive)
(3)  a  b   a  b 
(4) a  a  0 ; and a  a  0 if and only if a  o
1.1.5 The Cross Product

The cross product of two vectors a and b (also called the vector product) is denoted by
a  b . It is a vector with magnitude
a  b  a b sin  (1.1.3)
with  defined as for the dot product. It can be seen from the figure that the magnitude
of a  b is equivalent to the area of the parallelogram determined by the two vectors a
and b.
ab
a

b
Figure 1.1.6: the magnitude of the cross product
The direction of this new vector is perpendicular to both a and b. Whether a  b points
“up” or “down” is determined from the fact that the three vectors a, b and a  b form a
right handed system. This means that if the thumb of the right hand is pointed in the

Section 1.1
direction of a  b , and the open hand is directed in the direction of a, then the curling of
the fingers of the right hand so that it closes should move the fingers through the angle  ,
0     , bringing them to b. Some examples are shown in Fig. 1.1.7.
ab
b a
 
a
b
ab
Figure 1.1.7: examples of the cross product
The cross product possesses the following properties (which can be proved using the
above definition):
(1) a  b  b  a (not commutative)
(2) a  b  c   a  b  a  c (distributive)
(3)  a  b   a  b 
(4) a  b  o if and only if a and b  o  are parallel (“linearly dependent”)
The Triple Scalar Product
The triple scalar product, or box product, of three vectors u, v, w is defined by
u  v   w  v  w   u  w  u   v Triple Scalar Product (1.1.4)
Its importance lies in the fact that, if the three vectors form a right-handed triad, then the
volume V of a parallelepiped spanned by the three vectors is equal to the box product.
To see this, let e be a unit vector in the direction of u  v , Fig. 1.1.8. Then the projection
of w on u  v is h  w  e , and
w  u  v   w   u  v e 
 u vh (1.1.5)
V

Section 1.1
v e h
u
Figure 1.1.8: the triple scalar product
Note: if the three vectors do not form a right handed triad, then the triple scalar product
yields the negative of the volume. For example, using the vectors above,
 w  v   u  V .
1.1.6 Vectors and Points

Vectors are objects which have magnitude and direction, but they do not have any
specific location in space. On the other hand, a point has a certain position in space, and
the only characteristic that distinguishes one point from another is its position. Points
cannot be “added” together like vectors. On the other hand, a vector v can be added to a
point p to give a new point q, q  v  p , Fig. 1.1.9. Similarly, the “difference” between
two points gives a vector, q  p  v . Note that the notion of point as defined here is
slightly different to the familiar point in space with axes and origin – the concept of
origin is not necessary for these points and their simple operations with vectors.
q

v
p
Figure 1.1.9: adding vectors to points
1.1.7 Problems
1. Which of the following are scalars and which are vectors?
(i) weight
(ii) specific heat
(iii) momentum
(iv) energy
(v) volume
2. Find the magnitude of the sum of three unit vectors drawn from a common vertex of
a cube along three of its sides.

Section 1.1
3. Consider two non-collinear (not parallel) vectors a and b. Show that a vector r
lying in the same plane as these vectors can be written in the form r  pa  qb ,
where p and q are scalars. [Note: one says that all the vectors r in the plane are
specified by the base vectors a and b.]
4. Show that the dot product of two vectors u and v can be interpreted as the
magnitude of u times the component of v in the direction of u.
5. The work done by a force, represented by a vector F, in moving an object a given
distance is the product of the component of force in the given direction times the
distance moved. If the vector s represents the direction and magnitude (distance)
the object is moved, show that the work done is equivalent to F  s .
6. Prove that the dot product is commutative, a  b  b  a . [Note: this is equivalent to
saying, for example, that the work done in problem 5 is also equal to the component
of s in the direction of the force, times the magnitude of the force.]
7. Sketch b  a if a and b are as shown below.
a b
2 2 2 2
8. Show that a  b  a  b  a b .
9. Suppose that a rigid body rotates about an axis O with angular speed w, as shown
below. Consider a point p in the body with position vector r. Show that the
velocity v of p is given by v  ω  r , where ω is the vector with magnitude  and
whose direction is that in which a right-handed screw would advance under the
rotation. [Note: let s be the arc-length traced out by the particle as it rotates through
an angle  on a circle of radius r , then v  v  r  (since
s  r  , ds / dt  r (d / dt ) ).]
r v
p
r
O
10. Show, geometrically, that the dot and cross in the triple scalar product can be
interchanged: a  b   c  a  b  c  .
11. Show that the triple vector product a  b   c lies in the plane spanned by the
vectors a and b.

Section 1.2
1.2 Vector Spaces

The notion of the vector presented in the previous section is here re-cast in a more formal
and abstract way, using some basic concepts of Linear Algebra and Topology. This
might seem at first to be unnecessarily complicating matters, but this approach turns out
to be helpful in unifying and bringing clarity to much of the theory which follows.
Some background theory which complements this material is given in Appendix A to this
Chapter, §1.A.
1.2.1 The Vector Space

The vectors introduced in the previous section obey certain rules, those listed in §1.1.3. It
turns out that many other mathematical objects obey the same list of rules. For that
reason, the mathematical structure defined by these rules is given a special name, the
linear space or vector space.
First, a set is any well-defined list, collection, or class of objects, which could be finite or
infinite. An example of a set might be
B  x | x  3 (1.2.1)
which reads “B is the set of objects x such that x satisfies the property x  3 ”. Members
of a set are referred to as elements.
Consider now the field1 of real numbers R. The elements of R are referred to as scalars.
Let V be a non-empty set of elements a, b, c,  with rules of addition and scalar
multiplication, that is there is a sum a  b  V for any a, b  V and a product a  V
for any a  V ,   R . Then V is called a (real)2 vector space over R if the following
eight axioms hold:
1. associative law for addition: for any a, b, c  V , one has (a  b)  c  a  (b  c)
2. zero element: there exists an element o  V , called the zero element, such that
a  o  o  a  a for every a  V
3. negative (or inverse): for each a  V there exists an element  a  V , called the
negative of a, such that a  (a)  (a)  a  0
4. commutative law for addition: for any a, b  V , one has a  b  b  a
5. distributive law, over addition of elements of V: for any a, b  V and scalar   R ,
 (a  b)  a  b
6. distributive law, over addition of scalars: for any a  V and scalars  ,   R ,
(   )a  a   a
1
A field is another mathematical structure (see Appendix A to this Chapter, §1.A). For example, the set of
complex numbers is a field. In what follows, the only field which will be used is the familiar set of real
numbers with the usual operations of addition and multiplication.
2
“real”, since the associated field is the reals. The word real will usually be omitted in what follows for
brevity.

Section 1.2
7. associative law for multiplication: for any a  V and scalars  ,   R ,

 ( a)  ( )a
8. unit multiplication: for the unit scalar 1  R , 1a  a for any a  V .
The set of vectors as objects with “magnitude and direction” discussed in the previous
section satisfy these rules and therefore form a vector space over R. However, despite the
name “vector” space, other objects, which are not the familiar geometric vectors, can also
form a vector space over R, as will be seen in a later section.
1.2.2 Inner Product Space

Just as the vector of the previous section is an element of a vector space, next is
introduced the notion that the vector dot product is one example of the more general
inner product.
First, a function (or mapping) is an assignment which assigns to each element of a set A
a unique element of a set B, and is denoted by
f :AB (1.2.2)
An ordered pair a, b  consists of two elements a and b in which one of them is
designated the first element and the other is designated the second element The product
set (or Cartesian product) A  B consists of all ordered pairs a, b  where a  A and
bB :
A  B  a, b  | a  A, b  B (1.2.3)
Now let V be a real vector space. An inner product (or scalar product) on V is a
mapping that associates to each ordered pair of elements x, y, a scalar, denoted by x, y ,
, : V V  R (1.2.4)
that satisfies the following properties, for x, y, z  V ,   R :
1. additivity: x  y , z  x, z  y , z
2. homogeneity: x, y   x, y
3. symmetry: x, y  y , x
4. positive definiteness: x, x  0 when x  o
From these properties, it follows that, if x, y  0 for all y  V , then x  0
A vector space with an associated inner product is called an inner product space.
Two elements of an inner product space are said to be orthogonal if

Section 1.2
x, y  0 (1.2.5)
and a set of elements of V, x, y , z,  , are said to form an orthogonal set if every
element in the set is orthogonal to every other element:
x, y  0, x, z  0, y , z  0, etc. (1.2.6)
The above properties are those listed in §1.1.4, and so the set of vectors with the
associated dot product forms an inner product space.
Euclidean Vector Space
The set of real triplets  x1 , x 2 , x3  under the usual rules of addition and multiplication
forms a vector space R 3 . With the inner product defined by
x, y  x1 y1  x 2 y 2  x3 y 3
one has the inner product space known as (three dimensional) Euclidean vector space,
and denoted by E. This inner product allows one to take distances (and angles) between
elements of E through the norm (length) and metric (distance) concepts discussed next.
1.2.3 Normed Space

Let V be a real vector space. A norm on V is a real-valued function,
:V  R (1.2.7)
that satisfies the following properties, for x, y  V ,   R :
1. positivity: x  0
2. triangle inequality: x  y  x  y
3. homogeneity: x   x
4. positive definiteness: x  0 if and only if x  o
A vector space with an associated norm is called a normed vector space. Many different
norms can be defined on a given vector space, each one giving a different normed linear
space. A natural norm for the inner product space is
x  x, x (1.2.8)
It can be seen that this norm indeed satisfies the defining properties. When the inner
product is the vector dot product, the norm defined by 1.2.8 is the familiar vector
“length”.

Section 1.2
One important consequence of the definitions of inner product and norm is the Schwarz
inequality, which states that
x, y  x y (1.2.9)
One can now define the angle between two elements of V to be
 x, y 
 : V  V  R,  x, y   cos 1  
 (1.2.10)
 x y 
The quantity inside the curved brackets here is necessarily between  1 and  1 , by the
Schwarz inequality, and hence the angle  is indeed a real number.
1.2.4 Metric Spaces

Metric spaces are built on the concept of “distance” between objects. This is a
generalization of the familiar distance between two points on the real line.
Consider a set X. A metric is a real valued function,
d  ,  : X  X  R (1.2.11)
that satisfies the following properties, for x, y  X :
1. positive: d (x, y )  0 and d (x, x)  0 , for all x, y  X

2. strictly positive: if d (x, y )  0 then x  y , for all x, y  X
3. symmetry: d (x, y )  d (y , x) , for all x, y  X
4. triangle inequality: d (x, y )  d (x, z )  d (z, y ) , for all x, y, z  X
A set X with an associated metric is called a metric space. The set X can have more than
one metric defined on it, with different metrics producing different metric spaces.
Consider now a normed vector space. This space naturally has a metric defined on it:
d x, y   x  y (1.2.12)
and thus the normed vector space is a metric space. For the set of vectors with the dot
product, this gives the “distance” between two vectors x, y .

Section 1.2
1.2.5 The Affine Space

Consider a set P, the elements of which are called points. Consider also an associated
vector space V. An affine space consists of the set P, the set V, and two operations which
connect P and V:
(i) given two points p, q  P , one can define a difference, q  p which is a unique
element v of V, i.e. v  q  p  V
(ii) given a point p  P and v  V , one can define the sum v  p which is a unique point
q of P, i.e. q  v  p  P
and for which the following property holds, for p, q, r  P : q  r   r  p   q  p  .
From the above, one has for the affine space that p  p  o and q  p  p  q  , for all
p, q  P .
Note that one can take the sum of vectors, according to the structure of the vector space,
but one cannot take the sum of points, only the difference between two points. Further,
there is no notion of origin in the affine space. One can choose some fixed o  P to be
an origin. In that case, v  p  o is called the position vector of p relative to o.
Suppose now that the associated vector space is a Euclidean vector space, i.e. an inner
product space. Define the distance between two points through the inner product
associated with V,
d p, q   q  p  q  p, q  p (1.2.13)
It can be shown that this mapping d : P  P  R is a metric, i.e. it satisfies the metric
properties, and thus P is a metric space (although it is not a vector space). In this case, P
is referred to as Euclidean point space, Euclidean affine space or, simply, Euclidean
space. Whereas in Euclidean vector space there is a zero element, the origin (0,0,0) , in
Euclidean point space there is none – apart from that, the two spaces are the same and,
apart from certain special cases, one does not need to distinguish between them.

Section 1.3
1.3 Cartesian Vectors

So far the discussion has been in symbolic notation1, that is, no reference to ‘axes’ or
‘components’ or ‘coordinates’ is made, implied or required. The vectors exist
independently of any coordinate system. It turns out that much of vector (tensor)
mathematics is more concise and easier to manipulate in such notation than in terms of
corresponding component notations. However, there are many circumstances in which
use of the component forms of vectors (and tensors) is more helpful – or essential. In this
section, vectors are discussed in terms of components – component form.
1.3.1 The Cartesian Basis

Consider three dimensional (Euclidean) space. In this space, consider the three unit
vectors e1 , e 2 , e 3 having the properties
e1  e 2  e 2  e 3  e 3  e1  0 , (1.3.1)
so that they are mutually perpendicular (mutually orthogonal), and
e1  e1  e 2  e 2  e 3  e 3  1 , (1.3.2)
so that they are unit vectors. Such a set of orthogonal unit vectors is called an
orthonormal set, Fig. 1.3.1. Note further that this orthonormal system e1 , e 2 , e 3  is
right-handed, by which is meant e1  e 2  e 3 (or e 2  e 3  e1 or e 3  e1  e 2 ).
This set of vectors e1 , e 2 , e 3  forms a basis, by which is meant that any other vector can
be written as a linear combination of these vectors, i.e. in the form
a  a1e1  a 2 e 2  a3e 3 (1.3.3)
a
a3  a  e3
e3
e2 a2  a  e 2
e1
a1  a  e1
Figure 1.3.1: an orthonormal set of base vectors and Cartesian components
1
or absolute or invariant or direct or vector notation

Section 1.3
By repeated application of Eqn. 1.1.2 to a vector a, and using 1.3.2, the scalars in 1.3.3
can be expressed as (see Fig. 1.3.1)
a1  a  e1 , a2  a  e 2 , a3  a  e 3 (1.3.4)
The scalars a1 , a 2 and a 3 are called the Cartesian components of a in the given basis
e1 , e 2 , e 3 . The unit vectors are called base vectors when used for this purpose.
Note that it is not necessary to have three mutually orthogonal vectors, or vectors of unit
size, or a right-handed system, to form a basis – only that the three vectors are not co-
planar. The right-handed orthonormal set is often the easiest basis to use in practice, but
this is not always the case – for example, when one wants to describe a body with curved
boundaries (e.g., see §1.6.10).
The dot product of two vectors u and v, referred to the above basis, can be written as
u  v  u1e1  u 2 e 2  u 3e 3   v1e1  v 2 e 2  v3e 3 

 u1v1 e1  e1   u1v 2 e1  e 2   u1v3 e1  e 3 
 u 2 v1 e 2  e1   u 2 v 2 e 2  e 2   u 2 v3 e 2  e 3  (1.3.5)
 u 3 v1 e 3  e1   u 3 v 2 e 3  e 2   u 3 v3 e 3  e 3 
 u1v1  u 2 v 2  u 3 v3
Similarly, the cross product is
u  v  u1e1  u 2 e 2  u 3e 3   v1e1  v 2 e 2  v3e 3 

 u1v1 e1  e1   u1v 2 e1  e 2   u1v3 e1  e 3 
 u 2 v1 e 2  e1   u 2 v 2 e 2  e 2   u 2 v3 e 2  e 3  (1.3.6)
 u 3 v1 e 3  e1   u 3 v 2 e 3  e 2   u 3 v3 e 3  e 3 
 u 2 v3  u 3 v 2 e1  u1v3  u 3 v1 e 2  u1v 2  u 2 v1 e 3
This is often written in the form
e1 e2 e3
u  v  u1 u2 u3 , (1.3.7)
v1 v2 v3
that is, the cross product is equal to the determinant of the 3 3 matrix
e 1 e2 e3 
u u2 u 3 
 1
 v1 v2 v3 

Section 1.3
1.3.2 The Index Notation

The expression for the cross product in terms of components, Eqn. 1.3.6, is quite lengthy
– for more complicated quantities things get unmanageably long. Thus a short-hand
notation is used for these component equations, and this index notation2 is described
here.
In the index notation, the expression for the vector a in terms of the components
a1 , a 2 , a3 and the corresponding basis vectors e1 , e 2 , e 3 is written as
3
a  a1e1  a 2 e 2  a3 e 3   ai e i (1.3.8)
i 1
This can be simplified further by using Einstein’s summation convention, whereby the
summation sign is dropped and it is understood that for a repeated index (i in this case) a
summation over the range of the index (3 in this case3) is implied. Thus one writes
a  ai e i . This can be further shortened to, simply, ai .
The dot product of two vectors written in the index notation reads
u  v  u i vi Dot Product (1.3.9)
The repeated index i is called a dummy index, because it can be replaced with any other
letter and the sum is the same; for example, this could equally well be written as
u  v  u j v j or u k v k .
For the purpose of writing the vector cross product in index notation, the permutation
symbol (or alternating symbol)  ijk can be introduced:
 1 if (i, j , k ) is an even permutation of (1,2,3)


 ijk   1 if (i, j , k ) is an odd permutation of (1,2,3) (1.3.10)
0 if two or more indices are equal

For example (see Fig. 1.3.2),
 123  1
 132  1
 122  0
2
or indicial or subscript or suffix notation
3
2 in the case of a two-dimensional space/analysis

Section 1.3
3 2
Figure 1.3.2: schematic for the permutation symbol (clockwise gives +1)
Note that
 ijk   jki   kij   jik   kji   ikj (1.3.11)
and that, in terms of the base vectors {▲Problem 7},
e i  e j   ijk e k (1.3.12)
and {▲Problem 7}
 ijk  e i  e j   e k . (1.3.13)
The cross product can now be written concisely as {▲Problem 8}
u  v   ijk u i v j e k Cross Product (1.3.14)
Introduce next the Kronecker delta symbol  ij , defined by
0, i  j
 ij   (1.3.15)
1, i  j
Note that  11  1 but, using the index notation,  ii  3 . The Kronecker delta allows one
to write the expressions defining the orthonormal basis vectors (1.3.1, 1.3.2) in the
compact form
e i  e j   ij Orthonormal Basis Rule (1.3.16)
The triple scalar product (1.1.4) can now be written as
u  v   w   ijk u i v j e k   wm e m
  ijk u i v j wm  km
  ijk u i v j wk
u1 u2 u3
 v1 v2 v3 (1.3.17)
w1 w2 w3

Section 1.3
Note that, since the determinant of a matrix is equal to the determinant of the transpose of
a matrix, this is equivalent to
u1 v1 w1
u  v   w  u 2 v 2 w2 (1.3.18)
u 3 v3 w3
Here follow some useful formulae involving the permutation and Kronecker delta symbol
{▲Problem 13}:
 ijk  kpq   ip  jq   iq  jp
(1.3.19)
 ijk  ijp  2 pk
Finally, here are some other important identities involving vectors; the third of these is
called Lagrange’s identity4 {▲Problem 15}:
a  b   a  b   a 2 b 2  a  b 2
a  b  c   a  c b  a  b c
ac b c
a  b   c  d   (1.3.20)
ad b d
a  b   c  d   a  b  d c  a  b  c d
a  b  c d  d  b  ca  a  d  c b  a  b  d c
1.3.3 Matrix Notation for Vectors
The symbolic notation v and index notation vi e i (or simply vi ) can be used to denote a
vector. Another notation is the matrix notation: the vector v can be represented by a
3 1 matrix (a column vector):
 v1 
v 
 2
v3 
Matrices will be denoted by square brackets, so a shorthand notation for this

matrix/vector would be v  . The elements of the matrix v  can be written in the element
form vi . The element form for a matrix is essentially the same as the index notation for
the vector it represents.
4
to be precise, the special case of 1.3.20c, 1.3.20a, is Lagrange’s identity

Section 1.3
Formally, a vector can be represented by the ordered triplet of real numbers, v1 , v 2 , v3  .
The set of all vectors can be represented by R 3 , the set of all ordered triplets of real
numbers:
R 3  v1 , v 2 , v3  | v1 , v 2 , v3  R (1.3.21)
It is important to note the distinction between a vector and a matrix: the former is a
mathematical object independent of any basis, the latter is a representation of the vector
with respect to a particular basis – use a different set of basis vectors and the elements of
the matrix will change, but the matrix is still describing the same vector. Said another
way, there is a difference between an element (vector) v of Euclidean vector space and an
ordered triplet vi  R 3 . This notion will be discussed more fully in the next section.
As an example, the dot product can be written in the matrix notation as

 v1 
 
u v   u1 u 2 u 3 v 2 
T
v3 
“short” “full”
matrix notation matrix notation
 
Here, the notation u T denotes the 1 3 matrix (the row vector). The result is a 1 1
matrix, i.e. a scalar, in element form u i vi .
1.3.4 Cartesian Coordinates

Thus far, the notion of an origin has not been used. Choose a point o in Euclidean (point)
space, to be called the origin. An origin together with a right-handed orthonormal basis
e i  constitutes a (rectangular) Cartesian coordinate system, Fig. 1.3.3.
v (point)
e3
v  v  o (vector)
e2
e1 o
(point)
Figure 1.3.3: a Cartesian coordinate system

Section 1.3
A second point v then defines a position vector v  o , Fig. 1.3.3. The components of the
vector v  o are called the (rectangular) Cartesian coordinates of the point v 5. For
brevity, the vector v  o is simply labelled v, that is, one uses the same symbol for both
the position vector and associated point.
1.3.5 Problems
1. Evaluate u  v where u  e1  3e 2  2e 3 , v  4e1  2e 2  4e 3 .

2. Prove that for any vector u, u  (u  e1 )e1  (u  e 2 )e 2  (u  e 3 )e 3 . [Hint: write u in
component form.]
3. Find the projection of the vector u  e1  2e 2  e 3 on the vector
v  4e1  4e 2  7e 3 .
4. Find the angle between u  3e1  2e 2  6e 3 and v  4e1  3e 2  e 3 .
5. Write down an expression for a unit vector parallel to the resultant of two vectors u
and v (in symbolic notation). Find this vector when u  2e1  4e 2  5e 3 ,
v  e1  2e 2  3e 3 (in component form). Check that your final vector is indeed a
unit vector.
6. Evaluate u  v , where u  e1  2e 2  2e 3 , v  2e1  2e 2  e 3 .
7. Verify that e i  e j   ijm e m . Hence, by dotting each side with e k , show that
 ijk  e i  e j   e k .
8. Show that u  v   ijk u i v j e k .
9. The triple scalar product is given by u  v   w   ijk u i v j wk . Expand this equation
and simplify, so as to express the triple scalar product in full (non-index) component
form.
10. Write the following in index notation: v , v  e1 , v  e k .
11. Show that  ij ai b j is equivalent to a  b .
12. Verify that  ijk  ijk  6 .
13. Verify that  ijk  kpq   ip  jq   iq  jp and hence show that  ijk  ijp  2 pk .
14. Evaluate or simplify the following expressions:
(a)  kk (b)  ij  ij (c)  ij  jk (d)  1 jk  3 j v k
15. Prove Lagrange’s identity 1.3.20c.
16. If e is a unit vector and a an arbitrary vector, show that
a  a  e e  e  a  e 
which is another representation of Eqn. 1.1.2, where a can be resolved into
components parallel and perpendicular to e.
5
That is, “components” are used for vectors and “coordinates” are used for points

Section 1.4
1.4 Matrices and Element Form
1.4.1 Matrix – Matrix Multiplication

In the next section, §1.5, regarding vector transformation equations, it will be necessary
to multiply various matrices with each other (of sizes 3 1 , 1 3 and 3 3 ). It will be
helpful to write these matrix multiplications in a short-hand element form and to develop
some short “rules” which will be beneficial right through this chapter.
 
First, it has been seen that the dot product of two vectors can be represented by u T v  , or
 
u i vi . Similarly, the matrix multiplication u  v T gives a 3 3 matrix with element form
u i v j or, in full,
 u1v1 u1v 2 u1v3 

u v u 2 v2 u 2 v3 
 2 1
u 3 v1 u3 v2 u 3 v3 
This type of matrix represents the tensor product of two vectors, written in symbolic
notation as u  v (or simply uv). Tensor products will be discussed in detail in §1.8 and
§1.9.
Next, the matrix multiplication
Q11 Q12 Q13   u1 

Qu  Q21 Q22 Q23  u 2  (1.4.1)
Q31 Q32 Q33  u 3 
is a 3  1 matrix with elements Q ui  Qij u j {▲Problem 1}. The elements of Q u 
  
are the same as those of u T Q T , which in element form reads  u T Q T   
i  u j Qij .
 
The expression u Q is meaningless, but u T Q {▲Problem 2} is a 1  3 matrix with
 
elements  u T Q i  u j Q ji .
This leads to the following rule:
1. if a vector pre-multiplies a matrix Q   it is the transpose u T  

2. if a matrix Q  pre-multiplies the vector  it is u 
3. if summed indices are “beside each other”, as the j in u j Q ji or Qij u j
 the matrix is Q 
4. if summed indices are not beside each other, as the j in u j Qij
 the matrix is the transpose, Q T  

Section 1.4
Finally, consider the multiplication of 3  3 matrices. Again, this follows the “beside
each other” rule for the summed index. For example, A B  gives the 3  3 matrix
 
{▲Problem 6} A B ij  Aik Bkj , and the multiplication A T B  is written as
A B
T
ij  Aki Bkj . There is also the important identity
ABT   
 BT AT (1.4.2)
Note also the following (which applies to both the index notation and element form):
(i) if there is no free index, as in u i vi , there is one element (representing a scalar)
(ii) if there is one free index, as in u j Q ji , it is a 3  1 (or 1  3 ) matrix
(representing a vector)
(iii) if there are two free indices, as in Aki Bkj , it is a 3  3 matrix (representing, as
will be seen later, a second-order tensor)
1.4.2 The Trace of a Matrix

Another important notation involving matrices is the trace of a matrix, defined to be the
sum of the diagonal terms, and denoted by
trA   A11  A22  A33  Aii The Trace (1.4.3)
1.4.3 Problems
1. Show that Qui  Qij u j . To do this, multiply the matrix and the vector in Eqn.
1.4.1 and write out the resulting vector in full; Show that the three elements of the
vector are Q1 j u j , Q2 j u j and Q3 j u j .
2.  
Show that u T Q is a 1  3 matrix with elements u j Q ji (write the matrices out in
full).
3.
T
  
Show that Qu   u T Q T .
4. Are the three elements of Q u  the same as those of u T Q ?  
5. What is the element form for the matrix representation of a  b c ?
6. Write out the 3  3 matrices A and B in full, i.e. in terms of A11 , A12 , etc. and verify
that ABij  Aik Bkj for i  2, j  1 .
7. What is the element form for
 
(i) A  B T
     
(ii) v T A v  (there is no ambiguity here, since  v T A v   v T A v  )
 
(iii) B T AB
8. Show that  ij Aij  trA  .
9. Show that det[ A]   ijk A1i A2 j A3k   ijk A1i A j 2 Ak 3 .

Section 1.5
1.5 Coordinate Transformation of Vector Components

Very often in practical problems, the components of a vector are known in one coordinate
system but it is necessary to find them in some other coordinate system.
For example, one might know that the force f acting “in the x1 direction” has a certain
value, Fig. 1.5.1 – this is equivalent to knowing the x1 component of the force, in an
x1  x 2 coordinate system. One might then want to know what force is “acting” in some
other direction – for example in the x1 direction shown – this is equivalent to asking what
the x1 component of the force is in a new x1  x 2 coordinate system.
x2 x1 component
of force
x2
f
x1
x1
x1 component
of force
The relationship between the components in one coordinate system and the components
in a second coordinate system are called the transformation equations. These
transformation equations are derived and discussed in what follows.
1.5.1 Rotations and Translations

Any change of Cartesian coordinate system will be due to a translation of the base
vectors and a rotation of the base vectors. A translation of the base vectors does not
change the components of a vector. Mathematically, this can be expressed by saying that
the components of a vector a are e i  a , and these three quantities do not change under a
translation of base vectors. Rotation of the base vectors is thus what one is concerned
with in what follows.
1.5.2 Components of a Vector in Different Systems

Vectors are mathematical objects which exist independently of any coordinate system.
Introducing a coordinate system for the purpose of analysis, one could choose, for
example, a certain Cartesian coordinate system with base vectors e i and origin o, Fig.

Section 1.5
1.5.2. In that case the vector can be written as u  u1e1  u 2 e 2  u 3 e 3 , and u1 , u 2 , u 3 are
its components.
Now a second coordinate system can be introduced (with the same origin), this time with
base vectors ei . In that case, the vector can be written as u  u1e1  u 2 e2  u 3 e3 , where
u1 , u 2 , u 3 are its components in this second coordinate system, as shown in the figure.
Thus the same vector can be written in more than one way:
u  u1e1  u 2 e 2  u 3e 3  u1e1  u 2 e2  u 3 e3
The first coordinate system is often referred to as “the ox1 x 2 x3 system” and the second as
“the ox1 x 2 x3 system”.
x2
x2
u1
u 2
 x1 u2
e2 e1

o x1
u1
Note that the new coordinate system is obtained from the first one by a rotation of the
base vectors. The figure shows a rotation  about the x3 axis (the sign convention for
rotations is positive counterclockwise).
Two Dimensions
Concentrating for the moment on the two dimensions x1  x 2 , from trigonometry (refer to
Fig. 1.5.3),
u  u1e1  u2e 2
 OB  AB e1   BD  CP e 2
 cos u1  sin  u2 e1  sin  u1  cos u2 e 2
and so

Section 1.5
u1  cos u1  sin  u 2

u 2  sin  u1  cos u 2
vector components in vector components in

first coordinate system second coordinate system
In matrix form, these transformation equations can be written as
 u1  cos  sin    u1 

u    sin  cos  u 2 
 2 
x2
x2
u1 P
u2

x1 u2
 C
D
 A B
x1
o
u1
Figure 1.5.3: geometry of the 2D coordinate transformation
The 2  2 matrix is called the transformation or rotation matrix Q . By pre-

multiplying both sides of these equations by the inverse of Q , Q 1 , one obtains the  
transformation equations transforming from u1 u 2  to u1 u 2  :
T T
 u1   cos sin    u1 

u     sin  cos  u 2 
 2 
An important property of the transformation matrix is that it is orthogonal, by which is

meant that
Q   Q 
1 T
Orthogonality of Transformation/Rotation Matrix (1.5.1)
Three Dimensions
It is straight forward to show that, in the full three dimensions, Fig. 1.5.4, the components
in the two coordinate systems are related through

Section 1.5
 u1   cos( x1 , x1 ) cos( x1 , x 2 ) cos( x1 , x3 )   u1 

u   cos( x , x  ) cos( x , x  ) cos( x , x  ) u  
 2  2 1 2 2 2 3  2 
u 3  cos( x3 , x1 ) cos( x3 , x 2 ) cos( x3 , x3 )  u 3 
where cos( xi , x j ) is the cosine of the angle between the xi and x j axes. These nine
quantities are called the direction cosines of the coordinate transformation. Again
denoting these by the letter Q, Q11  cos( x1 , x1 ), Q12  cos( x1 , x2 ) , etc., so that
Qij  cos( xi , x j ) , (1.5.2)
one has the matrix equations
 u1  Q11 Q12 Q13   u1 

u   Q Q23  u 2 
 2   21 Q22
u 3  Q31 Q32 Q33  u 3 
or, in element form and short-hand matrix notation,
u i  Qij u j  u  Qu (1.5.3)
x2
x2
u
x1
x1
x3
x3
Figure 1.5.4: two different coordinate systems in a 3D space
Note: some authors define the matrix of direction cosines to consist of the components
Qij  cos( xi , x j ) , so that the subscript i refers to the new coordinate system and the j to
the old coordinate system, rather than the other way around as used here.
Transformation of Cartesian Base Vectors
The direction cosines introduced above also relate the base vectors in any two Cartesian
coordinate systems. It can be seen that
e i  e j  Qij (1.5.4)

Section 1.5
This relationship is illustrated in Fig. 1.5.5 for i  1 .
e2
cos( x1 , x2 )  e1  e2
e1 cos( x1 , x1 )  e1  e1

e1
e3
cos( x1 , x3 )  e1  e3
Figure 1.5.5: direction cosines
Formal Derivation of the Transformation Equations
In the above, the transformation equations u i  Qij u j were derived geometrically. They
can also be derived algebraically using the index notation as follows: start with the
relations u  u k e k  u j ej and post-multiply both sides by e i to get (the corresponding
matrix representation is to the right (also, see Problem 3 in §1.4.3)):
u k e k  e i  u j ej  e i
 u k  ki  u j Qij
 u i  u j Qij  u   u Q 
T T T
 u i  Qij u j  u  Qu

The inverse equations are {▲Problem 3}
u i  Q ji u j  u   Q T u (1.5.5)
Orthogonality of the Transformation Matrix Q 
As in the two dimensional case, the transformation matrix is orthogonal, Q T  Q 1 .    

This follows from 1.5.3, 1.5.5.
Example
Consider a Cartesian coordinate system with base vectors e i . A coordinate

transformation is carried out with the new basis given by

Section 1.5
e1  n1(1) e1  n2(1) e 2  n3(1) e 3

e2  n1( 2) e1  n2( 2 ) e 2  n3( 2 ) e 3
e3  n1( 3) e1  n2(3) e 2  n3( 3) e 3
What is the transformation matrix?
Solution
The transformation matrix consists of the direction cosines Qij  cos( xi , x j )  e i  ej , so
 u1  n1 n1(3)   u1 

(1)
n1( 2 )
u   n (1) n2( 2 )

n2(3)  u 2 
 2  2
u 3  n3(1) n3( 2 ) n3(3)  u 3 
1.5.3 Problems
1. The angles between the axes in two coordinate systems are given in the table below.
x1 x2 x3
x1 135 o
60 o
120 o
x 2 90 o 45 o 45 o
x3 45 o 60 o 120 o
Construct the corresponding transformation matrix Q  and verify that it is
orthogonal.
2. The ox1 x 2 x3 coordinate system is obtained from the ox1 x 2 x3 coordinate system by a
positive (counterclockwise) rotation of  about the x3 axis. Find the (full three
dimensional) transformation matrix Q  . A further positive rotation  about the
x 2 axis is then made to give the ox1x 2 x3 coordinate system. Find the
corresponding transformation matrix P  . Then construct the transformation matrix
R  for the complete transformation from the ox1 x2 x3 to the ox1x2 x3 coordinate
system.
3. Beginning with the expression u j e j  ei  u k ek  ei , formally derive the relation
 
u i  Q ji u j ( u   Q T u ).

Section 1.6
1.6 Vector Calculus 1 - Differentiation

Calculus involving vectors is discussed in this section, rather intuitively at first and more
formally toward the end of this section.
1.6.1 The Ordinary Calculus

Consider a scalar-valued function of a scalar, for example the time-dependent density
of a material    (t ) . The calculus of scalar valued functions of scalars is just the
ordinary calculus. Some of the important concepts of the ordinary calculus are reviewed
in Appendix B to this Chapter, §1.B.2.
1.6.2 Vector-valued Functions of a scalar

Consider a vector-valued function of a scalar, for example the time-dependent
displacement of a particle u  u(t ) . In this case, the derivative is defined in the usual
way,
du u(t  t )  u(t )
 lim t 0 ,
dt t
which turns out to be simply the derivative of the coefficients1,
du du1 du du du
 e1  2 e 2  3 e 3  i e i
dt dt dt dt dt
Partial derivatives can also be defined in the usual way. For example, if u is a function of
the coordinates, u( x1 , x 2 , x3 ) , then
u u( x1  x1 , x 2 , x3 )  u( x1 , x 2 , x3 )
 lim x1 0
x1 x1
Differentials of vectors are also defined in the usual way, so that when u1 , u 2 , u 3 undergo
increments du1  u1, du 2  u 2 , du 3  u 3 , the differential of u is
du  du1e1  du 2 e 2  du 3 e 3
and the differential and actual increment u approach one another as

u1, u 2 , u 3  0 .
1
assuming that the base vectors do not depend on t

Section 1.6
Space Curves
The derivative of a vector can be interpreted geometrically as shown in Fig. 1.6.1: u is

the increment in u consequent upon an increment t in t. As t changes, the end-point of
the vector u(t ) traces out the dotted curve  shown – it is clear that as t  0 , u
approaches the tangent to  , so that du / dt is tangential to  . The unit vector tangent to
the curve is denoted by τ :
du / dt
τ (1.6.1)
du / dt
τ
 s
s
u x2

ds du2
u(t ) u(t  t ) 
 du1
x1
(a) (b)
Figure 1.6.1: a space curve; (a) the tangent vector, (b) increment in arc length
Let s be a measure of the length of the curve  , measured from some fixed point on  .
Let s be the increment in arc-length corresponding to increments in the coordinates,
u  u1 , u 2 , u 3  , Fig. 1.6.1b. Then, from the ordinary calculus (see Appendix
T
1.B),
ds 2  du1 2  du 2 2  du 3 2

so that
2 2 2
ds  du   du   du 
  1   2   3 
dt  dt   dt   dt 
But
du du1 du du
 e1  2 e 2  3 e 3
dt dt dt dt
so that
du ds
 (1.6.2)
dt dt

Section 1.6
Thus the unit vector tangent to the curve can be written as
du / dt du
τ  (1.6.3)
ds / dt ds
If u is interpreted as the position vector of a particle and t is interpreted as time, then

v  du / dt is the velocity vector of the particle as it moves with speed ds / dt along  .
Example (of particle motion)
A particle moves along a curve whose parametric equations are x1  2t 2 , x 2  t 2  4t ,

x3  3t  5 where t is time. Find the component of the velocity at time t  1 in the
direction a  e1  3e 2  2e 3 .
Solution
The velocity is
v
dr d

dt dt
  
2t 2 e1  t 2  4t e 2  3t  5e 3 
 4e1  2e 2  3e 3 at t  1
The component in the given direction is v  aˆ , where â is a unit vector in the direction of
a, giving 8 14 / 7 .
■
Curvature
The scalar curvature  (s) of a space curve is defined to be the magnitude of the rate of
change of the unit tangent vector:
dτ d 2u
 ( s)   (1.6.4)
ds ds 2
Note that τ is in a direction perpendicular to τ , Fig. 1.6.2. In fact, this can be proved
as follows: since τ is a unit vector, τ  τ is a constant (  1 ), and so d τ  τ  / ds  0 , but
also,
d
τ  τ   2τ  dτ
ds ds
and so τ and dτ / ds are perpendicular. The unit vector defined in this way is called the
principal normal vector:

Section 1.6
1 dτ
ν (1.6.5)
 ds
1 s
R
 (s)
ν (s )
ν ( s  ds )

s τ (s )  τ
τ ( s  ds )
Figure 1.6.2: the curvature
This can be seen geometrically in Fig. 1.6.2: from Eqn. 1.6.5, τ is a vector of
magnitude s in the direction of the vector normal to τ . The radius of curvature R is
defined as the reciprocal of the curvature; it is the radius of the circle which just touches
the curve at s, Fig. 1.6.2.
Finally, the unit vector perpendicular to both the tangent vector and the principal normal
vector is called the unit binormal vector:
b  τν (1.6.6)
The planes defined by these vectors are shown in Fig. 1.6.3; they are called the rectifying
plane, the normal plane and the osculating plane.
Osculating
Normal plane plane

τ
Rectifying
plane
b
Figure 1.6.3: the unit tangent, principal normal and binormal vectors and associated
planes

Section 1.6
Rules of Differentiation
The derivative of a vector is also a vector and the usual rules of differentiation apply,
d
u  v   du  dv
dt dt dt
(1.6.7)
d
 (t ) v    dv  v d
dt dt dt
Also, it is straight forward to show that {▲Problem 2}
d
v  a   v  da  dv  a d
v  a   v  da  dv  a (1.6.8)
dt dt dt dt dt dt
(The order of the terms in the cross-product expression is important here.)
1.6.3 Fields
In many applications of vector calculus, a scalar or vector can be associated with each
point in space x. In this case they are called scalar or vector fields. For example
 (x) temperature a scalar field (a scalar-valued function of position)

v (x) velocity a vector field (a vector valued function of position)
These quantities will in general depend also on time, so that one writes  ( x, t ) or v ( x, t ) .
Partial differentiation of scalar and vector fields with respect to the variable t is
symbolised by  / t . On the other hand, partial differentiation with respect to the
coordinates is symbolised by  / xi . The notation can be made more compact by
introducing the subscript comma to denote partial differentiation with respect to the
coordinate variables, in which case  ,i   / xi , u i , jk   2 u i / x j x k , and so on.
1.6.4 The Gradient of a Scalar Field
Let  (x) be a scalar field. The gradient of  is a vector field defined by (see Fig. 1.6.4)
  
  e1  e2  e3
x1 x 2 x3

 ei Gradient of a Scalar Field (1.6.9)
xi


x
The gradient  is of considerable importance because if one takes the dot product of
 with dx , it gives the increment in  :

Section 1.6

   dx  e i  dx j e j
xi

 dxi (1.6.10)
xi
 d
  ( x  dx )   ( dx )

 
dx
Figure 1.6.4: the gradient of a vector
If one writes dx as dx e  dxe , where e is a unit vector in the direction of dx, then
 d  d
  e     (1.6.11)
 dx  in e direction dn
This quantity is called the directional derivative of  , in the direction of e, and will be
discussed further in §1.6.11.
The gradient of a scalar field is also called the scalar gradient, to distinguish it from the
vector gradient (see later)2, and is also denoted by
grad    (1.6.12)
Example (of the Gradient of a Scalar Field)
Consider a two-dimensional temperature field   x12  x 22 . Then
  2 x1e1  2 x 2 e 2
For example, at (1, 0) ,   1 ,   2e1 and at (1,1) ,   2 ,   2e1  2e 2 , Fig. 1.6.5.

Note the following:
(i)  points in the direction normal to the curve   const.
(ii) the direction of maximum rate of change of  is in the direction of 
2
in this context, a gradient is a derivative with respect to a position vector, but the term gradient is used
more generally than this, e.g. see §1.14

Section 1.6
(iii) the direction of zero d is in the direction perpendicular to 
 (1,1)
 1  (1, 0)
 2
Figure 1.6.5: gradient of a temperature field
The curves   x1 , x2   const. are called isotherms (curves of constant temperature). In

general, they are called iso-curves (or iso-surfaces in three dimensions).
■
Many physical laws are given in terms of the gradient of a scalar field. For example,
Fourier’s law of heat conduction relates the heat flux q (the rate at which heat flows
through a surface of unit area3) to the temperature gradient through
q   k  (1.6.13)
where k is the thermal conductivity of the material, so that heat flows along the direction
normal to the isotherms.
The Normal to a Surface
In the above example, it was seen that  points in the direction normal to the curve
  const. Here it will be seen generally how and why the gradient can be used to obtain
a normal vector to a surface.
Consider a surface represented by the scalar function f ( x1 , x 2 , x3 )  c , c a constant4, and

also a space curve C lying on the surface, defined by the position vector
r  x1 (t )e1  x 2 (t )e 2  x3 (t )e 3 . The components of r must satisfy the equation of the
surface, so f ( x1 (t ), x 2 (t ), x3 (t ))  c . Differentiation gives
df f dx1 f dx 2 f dx3
   0
dt x1 dt x 2 dt x3 dt
3
the flux is the rate of flow of fluid, particles or energy through a given surface; the flux density is the flux
per unit area but, as here, this is more commonly referred to simply as the flux
4
a surface can be represented by the equation f ( x1 , x 2 , x 3 )  c ; for example, the expression
2 2 2
x1  x 2  x 3  4 is the equation for a sphere of radius 2 (with centre at the origin). Alternatively, the
2 2
surface can be written in the form x 3  g ( x1 , x 2 ) , for example x 3  4  x1  x 2

Section 1.6
which is equivalent to the equation grad f  dr / dt   0 and, as seen in §1.6.2, dr / dt is a

vector tangential to the surface. Thus grad f is normal to the tangent vector; grad f must
be normal to all the tangents to all the curves through p, so it must be normal to the plane
tangent to the surface.
Taylor’s Series
Writing  as a function of three variables (omitting time t), so that    ( x1 , x 2 , x3 ) , then

 can be expanded in a three-dimensional Taylor’s series:
    
 ( x1  dx1 , x 2  dx2 , x3  dx3 )   ( x1 , x 2 , x3 )   dx1  dx 2  dx3 
 x1 x 2 x3 
1   2 
  2 dx1   
2
2  x1 
Neglecting the higher order terms, this can be written as

 (x  dx)   (x)   dx
x
which is equivalent to 1.6.9, 1.6.10.
1.6.5 The Nabla Operator
The symbolic vector operator  is called the Nabla operator5. One can write this in
component form as
   
  e1  e2  e3  ei (1.6.14)
x1 x 2 x3 xi
One can generalise the idea of the gradient of a scalar field by defining the dot product
and the cross product of the vector operator  with a vector field  , according to the
rules
 
    e i  ,     e i   (1.6.15)
xi xi
The following terminology is used:
grad   
div u    u (1.6.16)
curl u    u
5
or del or the Gradient operator

Section 1.6
These latter two are discussed in the following sections.
1.6.6 The Divergence of a Vector Field
From the definition (1.6.15), the divergence of a vector field a(x) is the scalar field
   a
div a    a   e i   a j e j   i
 xi  xi
Divergence of a Vector Field (1.6.17)
a a a
 1 2  3
x1 x 2 x3
Differential Elements & Physical interpretations of the Divergence
Consider a flowing compressible6 material with velocity field v ( x1 , x 2 , x3 ) . Consider

now a differential element of this material, with dimensions x1 , x 2 , x3 , with bottom
left-hand corner at ( x1 , x 2 , x3 ) , fixed in space and through which the material flows7, Fig.
1.6.6.
The component of the velocity in the x1 direction, v1 , will vary over a face of the element
but, if the element is small, the velocities will vary linearly as shown; only the
components at the four corners of the face are shown for clarity.
Since [distance = time  velocity], the volume of material flowing through the right-hand
face in time t is t times the “volume” bounded by the four corner velocities (between
the right-hand face and the plane surface denoted by the dotted lines); it is straightforward
to show that this volume is equal to the volume shown to the right, Fig. 1.6.6b, with
constant velocity equal to the average velocity v ave , which occurs at the centre of the face.
Thus the volume of material flowing out is8 x 2 x3 v ave t and the volume flux, i.e. the
rate of volume flow, is x 2 x3 v ave . Now
v ave  v1 ( x1  x1 , x 2  12 x 2 , x3  12 x3 )
Using a Taylor’s series expansion, and neglecting higher order terms,
v1 1 v v
v ave  v1 ( x1 , x 2 , x3 )  x1  2 x 2 1  12 x3 1
x1 x 2 x3
6
that is, it can be compressed or expanded
7
this type of fixed volume in space, used in analysis, is called a control volume
8
the velocity will change by a small amount during the time interval t . One could use the average
velocity in the calculation, i.e. 12 v1 ( x, t )  v1 ( x, t  t )  , but in the limit as t  0 , this will reduce to
v1 ( x, t )

Section 1.6
with the partial derivatives evaluated at ( x1 , x2 , x3 ) , so the volume flux out is
 v v v 
x 2 x3 v1 ( x1 , x 2 , x3 )  x1 1  12 x 2 1  12 x3 1 
 x1 x 2 x3 
v1 ( x1   x1 , x2   x2 , x3   x3 )
x 3
v1 ( x1   x1 , x2   x2 , x3 )
x2
v ave
v1 ( x1   x1 , x2 , x3   x3 )
( x1 , x 2 , x 3 ) v1 ( x1   x1 , x2 , x3 )
 x1
(a) (b)
Figure 1.6.6: a differential element; (a) flow through a face, (b) volume of material
flowing through the face
The net volume flux out (rate of volume flow out through the right-hand face minus the
rate of volume flow in through the left-hand face) is then x1 x 2 x3 v1 / x1  and the net
volume flux per unit volume is v1 / x1 . Carrying out a similar calculation for the other
two coordinate directions leads to
v1 v 2 v3
net unit volume flux out of an elemental volume:    div v (1.6.18)
x1 x 2 x3
which is the physical meaning of the divergence of the velocity field.
If div v  0 , there is a net flow out and the density of material is decreasing. On the other
hand, if div v  0 , the inflow equals the outflow and the density remains constant – such a
material is called incompressible9. A flow which is divergence free is said to be
isochoric. A vector v for which div v  0 is said to be solenoidal.
Notes:
 The above result holds only in the limit when the element shrinks to zero size – so that
the extra terms in the Taylor series tend to zero and the velocity field varies in a linear
fashion over a face
 consider the velocity at a fixed point in space, v ( x, t ) . The velocity at a later time,
v( x, t  t ) , actually gives the velocity of a different material particle. This is shown in
Fig. 1.6.7 below: the material particles 1, 2, 3 are moving through space and whereas
v ( x, t ) represents the velocity of particle 2, v ( x, t  t ) now represents the velocity of
particle 1, which has moved into position x. This point is important in the consideration
of the kinematics of materials, to be discussed in Chapter 2
9
a liquid, such as water, is a material which is incompressible

Section 1.6
v(x, t )
time t 1 2 3
v(x, t  t )
time t  t 1 2 3
x  x x x  x
Figure 1.6.7: moving material particles
Another example would be the divergence of the heat flux vector q. This time suppose
also that there is some generator of heat inside the element (a source), generating at a rate
of r per unit volume, r being a scalar field. Again, assuming the element to be small, one
takes r to be acting at the mid-point of the element, and one considers r ( x1  12 x1 ,) .
Assume a steady-state heat flow, so that the (heat) energy within the elemental volume
remains constant with time – the law of balance of (heat) energy then requires that the net
flow of heat out must equal the heat generated within, so
q1 q q
x1 x 2 x3  x1 x 2 x3 2  x1 x 2 x3 3
x1 x 2 x3
 r 1 r 1 r 
 x1 x 2 x3 r ( x1 , x 2 , x3 )  12 x1  2 x 2  2 x3 
 x1 x 2 x3 
Dividing through by x1 x 2 x3 and taking the limit as x1 , x 2 , x3  0 , one obtains
divq  r (1.6.19)
Here, the divergence of the heat flux vector field can be interpreted as the heat generated
(or absorbed) per unit volume per unit time in a temperature field. If the divergence is
zero, there is no heat being generated (or absorbed) and the heat leaving the element is
equal to the heat entering it.
1.6.7 The Laplacian
Combining Fourier’s law of heat conduction (1.6.13), q  k  , with the energy

balance equation (1.6.19), divq  r , and assuming the conductivity is constant, leads to
 k    r . Now
      
  ij   
2
   e i  ej  
 x
xi   
 j  xi  x j  xi2
(1.6.20)
 2  2  2
 2  2  2
x1 x 2 x3

Section 1.6
This expression is called the Laplacian of  . By introducing the Laplacian

operator  2     , one has
r
 2   (1.6.21)
k
This equation governs the steady state heat flow for constant conductivity. In general, the
equation  2  a is called Poisson’s equation. When there are no heat sources (or
sinks), one has Laplace’s equation,  2  0 . Laplace’s and Poisson’s equation arise in
many other mathematical models in mechanics, electromagnetism, etc.
1.6.8 The Curl of a Vector Field
From the definition 1.6.15 and 1.6.14, the curl of a vector field a( x) is the vector field

curla    a  e i  a j e j 
xi
Curl of a Vector Field (1.6.22)
a j a j
 e i  e j   ijk ek
xi xi
It can also be expressed in the form
e1 e2 e3
  
curl a    a 
x1 x 2 x3
a1 a2 a3 (1.6.23)
a j a k a
  ijk e k   ijk e i   ijk i e j
xi x j x k
Note: the divergence and curl of a vector field are independent of any coordinate system
(for example, the divergence of a vector and the length and direction of curl a are
independent of the coordinate system in use) – these will be re-defined without reference
to any particular coordinate system when discussing tensors (see §1.14).
Physical interpretation of the Curl
Consider a particle with position vector r and moving with velocity v  ω  r , that is,
with an angular velocity  about an axis in the direction of ω . Then {▲Problem 7}
curl v    ω  r   2ω (1.6.24)
Thus the curl of a vector field is associated with rotational properties. In fact, if v is the
velocity of a moving fluid, then a small paddle wheel placed in the fluid would tend to
rotate in regions where curl v  0 , in which case the velocity field v is called a vortex

Section 1.6
field. The paddle wheel would remain stationary in regions where curl v  0 , in which
case the velocity field v is called irrotational.
1.6.9 Identities
Here are some important identities of vector calculus {▲Problem 8}:
grad     grad  grad

div u  v   div u  div v (1.6.25)
curl u  v   curl u  curl v
grad( )   grad   grad

div  u    div u  grad   u
curl  u    curl u  grad   u
div  u  v   v  curl u  u  curl v (1.6.26)
curl grad   o
divcurlu   0
div grad    2  grad  grad
1.6.10 Cylindrical and Spherical Coordinates

Cartesian coordinates have been used exclusively up to this point. In many practical
problems, it is easier to carry out an analysis in terms of cylindrical or spherical
coordinates. Differentiation in these coordinate systems is discussed in what follows10.
Cylindrical Coordinates
Cartesian and cylindrical coordinates are related through (see Fig. 1.6.8)
x  r cos  r  x2  y2
y  r sin  ,   tan 1  y / x  (1.6.27)
zz zz
Then the Cartesian partial derivatives become
 r     sin  
   cos 
x x r x  r r 
(1.6.28)
 r     cos 
   sin  
y y r y  r r 
10
this section also serves as an introduction to the more general topic of Curvilinear Coordinates covered
in §1.16-§1.19

Section 1.6
x3  z ez
x, y, z   r, , z  e
 
ez er
ey
x2  y ex
x1  x
 r
Figure 1.6.8: cylindrical coordinates
The base vectors are related through
e x  e r cos   e sin  e r  e x cos   e y sin 

e y  e r sin   e cos  , e  e x sin   e y cos  (1.6.29)
ez  ez ez  ez
so that from Eqn. 1.6.14, after some algebra, the Nabla operator in cylindrical coordinates
reads as {▲Problem 9}
 1  
  er  e  ez (1.6.30)
r r  z
which allows one to take the gradient of a scalar field in cylindrical coordinates:
 1  
  er  e  ez (1.6.31)
r r  z
Cartesian base vectors are independent of position. However, the cylindrical base
vectors, although they are always of unit magnitude, change direction with position. In
particular, the directions of the base vectors e r , e depend on  , and so these base
vectors have derivatives with respect to  : from Eqn. 1.6.29,

e r  e

(1.6.32)

e   e r

with all other derivatives of the base vectors with respect to r , , z equal to zero.
The divergence can now be evaluated:

Section 1.6
  1   
  v  er  e  e z   v r e r  v e  v z e z 
 r r  z 
(1.6.33)
v v 1 v v z
 r  r  
r r r  z
Similarly the curl of a vector and the Laplacian of a scalar are {▲Problem 10}
 1 v z v   v v  1   v 
 v    e r   r  z e    rv   r e z
 r  z   z r   r  r   (1.6.34)
  1  1    
2 2 2
 2  2   
r r r r 2  2 z 2
Spherical Coordinates
Cartesian and spherical coordinates are related through (see Fig. 1.6.9)
x  r sin  cos  r  x2  y2  z2
y  r sin  sin  ,   tan 1 x 2
 y2 / z  (1.6.35)
z  r cos    tan 1  y / x 
and the base vectors are related through
e x  e r sin  cos   e cos  cos   e  sin 

e y  e r sin  sin   e cos  sin   e  cos 
e z  e r cos   e sin 
(1.6.36)
e r  e x sin  cos   e y sin  sin   e z cos 
e  e x cos  cos   e y cos  sin   e z sin 
e   e x sin   e y cos 
z
er
x, y, z   r , ,   e
  
r ez e
y ey
ex

x
Figure 1.6.9: spherical coordinates
In this case the non-zero derivatives of the base vectors are

Section 1.6

e r  sin e 
 
e r  e
 
, e  cos e  (1.6.37)
 
e   e r
 
e    sin e r  cos e

and it can then be shown that {▲Problem 11}
 1  1 
  er  e  e
r r  r sin  
1  2  v
v  
r vr 
1

r sin  
sin v   1
r sin  
(1.6.38)
r r
2
 2 2  1  2 cot   1  2
 2     
r 2 r r r 2  2 r 2  r 2 sin 2   2
1.6.11 The Directional Derivative
Consider a function  x  . The directional derivative of  in the direction of some vector

w is the change in  in that direction. Now the difference between its values at position
x and x  w is, Fig. 1.6.10,
d   x  w    x  (1.6.39)
 (x  w )
D w (x)  ( x)
x
w  0
 1

Figure 1.6.10: the directional derivative

Section 1.6
An approximation to d can be obtained by introducing a parameter  and by

considering the function  x  w  ; one has  x  w  0   x  and
 x  w  1   x  w  .
If one treats  as a function of  , a Taylor’s series about   0 gives
d    2 d 2  
 ( )   (0)    
d    0 2 d 2   0
or, writing it as a function of x  w ,
 x  w   
d
 (x  w )   (x)  
d  0
By setting   1 , the derivative here can be seen to be a linear approximation to the

increment d , Eqn. 1.6.39. This is defined as the directional derivative of the function
 (x) at the point x in the direction of w, and is denoted by
 x  w 
d
 x [ w ]  The Directional Derivative (1.6.40)
d  0
The directional derivative is also written as D w x  .
The power of the directional derivative as defined by Eqn. 1.6.40 is its generality, as seen
in the following example.
Example (the Directional Derivative of the Determinant)
Consider the directional derivative of the determinant of the 2  2 matrix A, in the

direction of a second matrix T (the word “direction” is obviously used loosely in this
context). One has
 A det A [T]  det A  T 

d
d  0

d
 A11  T11  A22  T22    A12  T12  A21  T21 
d  0
 A11T22  A22T11  A12T21  A21T12

■
The Directional Derivative and The Gradient
Consider a scalar-valued function  of a vector z. Let z be a function of a parameter  ,

    z1  , z 2  , z 3   . Then

Section 1.6
d  dz i  dz
  
d z i d z d
Thus, with z  x  w ,
  dz  
 z    
d
 x [ w ]     w (1.6.41)
d  0  z d   0 x
which can be compared with Eqn. 1.6.11. Note that for Eqns. 1.6.11 and 1.6.41 to be
consistent definitions of the directional derivative, w here should be a unit vector.
1.6.12 Formal Treatment of Vector Calculus

The calculus of vectors is now treated more formally in what follows, following on from
the introductory section in §1.2. Consider a vector h, an element of the Euclidean vector
space E, h  E . In order to be able to speak of limits as elements become “small” or
“close” to each other in this space, one requires a norm. Here, take the standard
Euclidean norm on E, Eqn. 1.2.8,
h  h, h  h  h (1.6.42)
Consider next a scalar function f : E  R . If there is a constant M  0 such that

f h   M h as h  o , then one writes
f h   O  h  as h  o (1.6.43)
This is called the Big Oh (or Landau) notation. Eqn. 1.6.43 states that f h  goes to
zero at least as fast as h . An expression such as
f h   g h   O  h  (1.6.44)
then means that f h   g h  is smaller than h for h sufficiently close to o.
Similarly, if
f h 
 0 as h  o (1.6.45)
h
then one writes f h   o h  as h  o . This implies that f h  goes to zero faster than
h .

Section 1.6
A field is a function which is defined in a Euclidean (point) space E 3 . A scalar field is

then a function f : E 3  R . A scalar field is differentiable at a point x  E 3 if there
exists a vector Df x   E such that
f x  h   f x   Df x   h  o h  for all h  E (1.6.46)
In that case, the vector Df x  is called the derivative (or gradient) of f at x (and is given
the symbol f x  ).
Now setting h  w in 1.6.46, where w  E is a unit vector, dividing through by  and

taking the limit as   0 , one has the equivalent statement
f x   w  f x  w  for all w  E
d
(1.6.47)
d  0
which is 1.6.41. In other words, for the derivative to exist, the scalar field must have a
directional derivative in all directions at x.
Using the chain rule as in §1.6.11, Eqn. 1.6.47 can be expressed in terms of the Cartesian
basis e i  ,
f f
f x   w  wi  ei  w j e j (1.6.48)
xi xi
This must be true for all w and so, in a Cartesian basis,
f
f x   ei (1.6.49)
xi
which is Eqn. 1.6.9.
1.6.13 Problems
1. A particle moves along a curve in space defined by
    
r  t 3  4t e1  t 2  4t e 2  8t 2  3t 3 e 3 
Here, t is time. Find
(i) a unit tangent vector at t  2
(ii) the magnitudes of the tangential and normal components of acceleration at t  2
2. Use the index notation (1.3.12) to show that
d
v  a   v  da  dv  a . Verify this
dt dt dt
result for v  3te1  t e 3 , a  t e1  te 2 . [Note: the permutation symbol and the unit
2 2
vectors are independent of t; the components of the vectors are scalar functions of t
which can be differentiated in the usual way, for example by using the product rule of
differentiation.]

Section 1.6
3. The density distribution throughout a material is given by   1  x  x .

(i) what sort of function is this?
(ii) the density is given in symbolic notation - write it in index notation
(iii) evaluate the gradient of 
(iv) give a unit vector in the direction in which the density is increasing the most
(v) give a unit vector in any direction in which the density is not increasing
(vi) take any unit vector other than the base vectors and the other vectors you used
above and calculate d / dx in the direction of this unit vector
(vii) evaluate and sketch all these quantities for the point (2,1).
In parts (iii-iv), give your answer in (a) symbolic, (b) index, and (c) full notation.
4. Consider the scalar field defined by   x 2  3 yx  2 z .
(i) find the unit normal to the surface of constant  at the origin (0,0,0)
(ii) what is the maximum value of the directional derivative of  at the origin?
(iii) evaluate d / dx at the origin if dx  ds (e1  e 3 ) .
5. If u  x1 x 2 x3e1  x1 x 2 e 2  x1e 3 , determine div u and curl u .
6. Determine the constant a so that the vector
v   x1  3x 2 e1   x 2  2 x3 e 2   x1  ax3 e 3
is solenoidal.
7. Show that curl v  2ω (see also Problem 9 in §1.1).
8. Verify the identities (1.6.25-26).
9. Use (1.6.14) to derive the Nabla operator in cylindrical coordinates (1.6.30).
10. Derive Eqn. (1.6.34), the curl of a vector and the Laplacian of a scalar in the
cylindrical coordinates.
11. Derive (1.6.38), the gradient, divergence and Laplacian in spherical coordinates.
12. Show that the directional derivative D v (u) of the scalar-valued function of a vector
 (u)  u  u , in the direction v, is 2u  v .
13. Show that the directional derivative of the functional
2
1
l
 d 2v  l
U v( x)    EI  2  dx   p( x)v( x)dx
2 0  dx  0
in the direction of  (x) is given by

d 2 v ( x ) d 2 ( x )
l l
0 EI dx 2 dx 2 dx  0 p( x) ( x)dx .

Section 1.7
1.7 Vector Calculus 2 - Integration
1.7.1 Ordinary Integrals of a Vector

A vector can be integrated in the ordinary way to produce another vector, for example
 t  t e  5 15
2
1  2t 2 e 2  3e 3 dt   e1  e 2  3e 3
1
6 2
1.7.2 Line Integrals

Discussed here is the notion of a definite integral involving a vector function that
generates a scalar.
Let x  x1e1  x 2 e 2  x3 e 3 be a position vector tracing out the curve C between the points
p1 and p 2 . Let f be a vector field. Then
p2
 f  dx   f  dx    f dx
p1 C C
1 1  f 2 dx 2  f 3 dx3 
is an example of a line integral.
Example (of a Line Integral)
A particle moves along a path C from the point (0,0,0) to (1,1,1) , where C is the straight
line joining the points, Fig. 1.7.1. The particle moves in a force field given by
 
f  3x12  6 x 2 e1  14 x 2 x3 e 2  20 x1 x32 e 3
What is the work done on the particle?
dx 
Figure 1.7.1: a particle moving in a force field

Section 1.7
Solution
The work done is
W   f  dx   3x12  6 x 2 dx1  14 x 2 x3 dx 2  20 x1 x32 dx3 

C C
The straight line can be written in the parametric form x1  t , x 2  t , x3  t , so that
1

W   20t 3  11t 2  6t dt  13
3
or W  f 
dx
dt
dt   f  e1  e2  e3 dt 
13
3
0 C C
If C is a closed curve, i.e. a loop, the line integral is often denoted  v  dx .

C
Note: in fluid mechanics and aerodynamics, when v is the velocity field, this integral
C v  dx is called the circulation of v about C.
1.7.3 Conservative Fields
If for a vector f one can find a scalar  such that
f   (1.7.1)
then
p2
(1)  f  dx
p1
is independent of the path C joining p1 and p 2
(2)
C
 f  dx  0 around any closed curve C
In such a case, f is called a conservative vector field and  is its scalar potential1. For
example, the work done by a conservative force field f is

p2 p2 p2 2 p

p1
f  dx  
p1
   dx   xi
p1
dx i   d   ( p2 )   ( p1 )
p1
which clearly depends only on the values at the end-points p1 and p 2 , and not on the
path taken between them.
It can be shown that a vector f is conservative if and only if curlf  o {▲Problem 3}.
1
in general, of course, there does not exist a scalar field  such that f   ; this is not surprising since a
vector field has three scalar components whereas  is determined from just one

Section 1.7
Example (of a Conservative Force Field)
The gravitational force field f  mge 3 is an example of a conservative vector field.

Clearly, curlf  o , and the gravitational scalar potential is    mgx3 :
p2 p2
W    mge 3  dx  mg  dx3   mg x3 ( p 2 )  x3 ( p1 )    p 2     p1 

p1 p1
Example (of a Conservative Force Field)
Consider the force field
f  (2 x1 x 2  x33 )e1  x12 e 2  3x1 x32 e 3
Show that it is a conservative force field, find its scalar potential and find the work done
in moving a particle in this field from (1,2,1) to (3,1,4) .
Solution
One has
e1 e2 e3
curlf   / x1  / x 2  / x3  o
2 x1 x 2  x33 x12 3 x1 x32
so the field is conservative.
To determine the scalar potential, let
  
f 1e 1  f 2 e 2  f 3 e 3  e1  e2  e3 .
x1 x 2 x3
Equating coefficients and integrating leads to
  x12 x 2  x1 x33  p( x 2 , x3 )
  x12 x 2  q( x1 , x3 )
 x1 x33  r ( x1 , x1 )
which agree if one chooses p  0, q  x1 x33 , r  x12 x 2 , so that   x12 x 2  x1 x33 , to which
may be added a constant.
The work done is

Section 1.7
W   (3,1,4)   (1,2,1)  202

■
Helmholtz Theory
As mentioned, a conservative vector field which is irrotational, i.e. f   , implies

  f  o , and vice versa. Similarly, it can be shown that if one can find a vector a such
that f    a , where a is called the vector potential, then f is solenoidal, i.e.   f  0
{▲Problem 4}.
Helmholtz showed that a vector can always be represented in terms of a scalar potential
 and a vector potential a:2
Type of Vector Condition Representation

General f      a
Irrotational (conservative) f  o f  
Solenoidal  f  0 f  a
1.7.4 Double Integrals

The most elementary type of two-dimensional integral is that over a plane region. For
example, consider the integral over a region R in the x1  x 2 plane, Fig. 1.7.2. The
integral
 dx dx
R
1 2
then gives the area of R and, just as the one dimensional integral of a function gives the
area under the curve, the integral
 f ( x , x
R
1 2 )dx1 dx 2
gives the volume under the (in general, curved) surface x3  f ( x1 , x 2 ) . These integrals
are called double integrals.
2
this decomposition can be made unique by requiring that f  0 as x   ; in general, if one is given f,
then  and a can be obtained by solving a number of differential equations

Section 1.7
x3 x3  f ( x1 , x2 )
x2
R
x1
Figure 1.7.2: integration over a region
Change of variables in Double Integrals
To evaluate integrals of the type  R

f ( x1 , x 2 )dx1 dx 2 , it is often convenient to make a
change of variable. To do this, one must find an elemental surface area in terms of the
new variables, t1 , t 2 say, equivalent to that in the x1 , x 2 coordinate system, dS  dx1 dx 2 .
The region R over which the integration takes place is the plane surface g ( x1 , x 2 )  0 .
Just as a curve can be represented by a position vector of one single parameter t (cf.
§1.6.2), this surface can be represented by a position vector with two parameters3, t1 and
t2 :
x  x1 (t1 , t 2 )e1  x 2 (t1 , t 2 )e 2
Parameterising the plane surface in this way, one can calculate the element of surface dS
in terms of t1 , t 2 by considering curves of constant t1 , t 2 , as shown in Fig. 1.7.3. The
vectors bounding the element are
x x
dx (1)  dx t  dt1 , dx ( 2)  dx t  dt 2 (1.7.2)
2 const
t1 1 const
t 2
so the area of the element is given by
x x
dS  dx (1)  dx ( 2 )   dt1 dt 2  J dt1 dt (1.7.3)
t1 t 2
where J is the Jacobian of the transformation,
3 2 2
for example, the unit circle x1  x2  1  0 can be represented by x  t1 cos t 2 e1  t1 sin t 2 e 2 , 0t11 ,
0  t 2  2 ( t1 , t 2 being in this case the polar coordinates r,  , respectively)

Section 1.7
x1 x 2 x1 x1

t t1 t t 2
J 1 or J 1 (1.7.4)
x1 x 2 x 2 x 2
t 2 t 2 t1 t 2
The Jacobian is also often written using the notation
  x1 , x 2 
dx1 dx 2  Jdt1 dt 2 , J
 t1 , t 2 
The integral can now be written as
 f (t , t
R
1 2 ) Jdt1 dt 2
x2
dS
dx ( 2) dx (1)
t1  t1
t1
t 2  t 2
t2
x1
Figure 1.7.3: a surface element
Example
Consider a region R, the quarter unit-circle in the first quadrant, 0  x 2  1  x12 ,

0  x1  1 . The moment of inertia about the x1 – axis is defined by
I x1   x 22 dx1 dx 2
R
Transform the integral into the new coordinate system t1 , t 2 by making the substitutions4
x1  t1 cos t 2 , x 2  t1 sin t 2 . Then
x1 x1
t t 2 cos t 2  t1 sin t 2
J 1   t1
x 2 x 2 sin t 2 t1 cos t 2
t1 t 2
4
these are the polar coordinates, t1 , t 2 equal to r,  , respectively

Section 1.7
so
 /2 1

I x1   t sin 2 t 2 dt1 dt 2 
3
1
0 0
16
■
1.7.5 Surface Integrals

Up to now, double integrals over a plane region have been considered. In what follows,
consideration is given to integrals over more complex, curved, surfaces in space, such as
the surface of a sphere.
Surfaces
Again, a curved surface can be parameterized by t1 ,t 2 , now by the position vector
x  x1 (t1 , t 2 )e1  x 2 (t1 , t 2 )e 2  x3 (t1 , t 2 )e 3
One can generate a curve C on the surface S by taking t1  t1 ( s ) , t 2  t 2 ( s ) so that C has

position vector, Fig. 1.7.4,
xs   xt1 ( s ), t 2 ( s ) 
A vector tangent to C at a point p on S is, from Eqn. 1.6.3,
dx x dt1 x dt 2
 
ds t1 ds t 2 ds
x3
C
s
S
x(s)
x(t1 , t2 )
x2
x1
Figure 1.7.4: a curved surface
Many different curves C pass through p, and hence there are many different tangents,
with different corresponding values of dt1 / ds, dt 2 / ds . Thus the partial derivatives
x / t1 , x / t 2 must also both be tangential to C and so a normal to the surface at p is
given by their cross-product, and a unit normal is

Section 1.7
 x x  x x
n     /  (1.7.5)

 1t t 2  t1 t 2
In some cases, it is possible to use a non-parametric form for the surface, for example
g ( x1 , x 2 , x3 )  c , in which case the normal can be obtained simply from
n  grad g / grad g .
Example (Parametric Representation and the Normal to a Sphere)
The surface of a sphere of radius a can be parameterised as5
x  asin t1 cos t 2 e1  sin t1 sin t 2 e 2  cos t1e 3  , 0  t1   , 0  t 2  2
Here, lines of t1  const are parallel to the x1  x 2 plane (“parallels”), whereas lines of
t 2  const are “meridian” lines, Fig. 1.7.5. If one takes the simple expressions
t1  s, t 2   / 2  s , over 0  s   / 2 , one obtains a curve C1 joining (0,0,1) and (1,0,0) ,
and passing through (1 / 2,1 / 2,1 / 2 ) , as shown.
x3
n
t1   / 2
C1
x2
x1
Figure 1.7.5: a sphere
The partial derivatives with respect to the parameters are
x
 acos t1 cos t 2 e1  cos t1 sin t 2 e 2  sin t1e 3 
t1
x
 a sin t1 sin t 2 e1  sin t1 cos t 2 e 2 
t 2
so that
x x

t1 t 2

 a 2 sin 2 t1 cos t 2 e1  sin 2 t1 sin t 2 e 2  sin t1 cos t1e 3 
5
these are the spherical coordinates (see §1.6.10); t1   , t 2  

Section 1.7
and a unit normal to the spherical surface is
n  sin t1 cos t 2 e1  sin t1 sin t 2 e 2  cos t1e 3
For example, at t1  t 2   / 4 (this is on the curve C1 ), one has
n / 4,  / 4   12 e1  12 e 2  1
2
e3
and, as expected, it is in the same direction as r.

■
Surface Integrals
Consider now the integral  f dS

S
where f is a vector function and S is some curved
surface. As for the integral over the plane region,
x x
dS  dx t  dx t   dt1 dt 2 ,
2 const 1 const
t1 t 2
only now dS is not “flat” and x is three dimensional. The integral can be evaluated if
one parameterises the surface with t1 , t 2 and then writes
x x
 f t
S 1

t 2
dt1 dt 2
One way to evaluate this cross product is to use the relation (Lagrange’s identity,
Problem 15, §1.3)
a  b   c  d   a  c b  d   a  d b  c  (1.7.6)
so that
2 2
x x  x x   x x   x x  x x   x x 
                   (1.7.7)
t1 t 2  t1 t 2   t1 t 2   t1 t1  t 2 t 2   t1 t 2 
Example (Surface Area of a Sphere)
Using the parametric form for a sphere given above, one obtains
2
x x
  a 4 sin 2 t1
t1 t 2
so that

Section 1.7
2 
area   dS  a 2
  sin t dt dt
1 1 2  4a 2
S 0 0
Flux Integrals
Surface integrals often involve the normal to the surface, as in the following example.
Example
If f  4 x1 x3e1  x 22 e 2  x 2 x3 e 3 , evaluate  f  ndS , where S is the surface of the cube

S
bounded by x1  0,1; x 2  0,1; x3  0,1 , and n is the unit outward normal, Fig. 1.7.6.
x3
n  e3
n  e2
n  e 2
n  e1
x2
x1
Figure 1.7.6: the unit cube
Solution
The integral needs to be evaluated over the six faces. For the face with n  e1 , x1  1
and
1 1 1 1
 f  n dS    4 x e 
 x e  x 2 x3e 3  e1 dx 2 dx3  4   x3 dx 2 dx3  2
2
3 1 2 2
S 0 0 0 0
Similarly for the other five sides, whence  f  ndS 

S
3
2 .
Integrals of the form  f  ndS

S
are known as flux integrals and arise quite often in
applications. For example, consider a material flowing with velocity v, in particular the
flow through a small surface element dS with outward unit normal n, Fig. 1.7.7. The
volume of material flowing through the surface in time dt is equal to the volume of the
slanted cylinder shown, which is the base dS times the height. The slanted height is (=

Section 1.7
velocity × time) is v dt , and the vertical height is then v  ndt . Thus the rate of flow is
the volume flux (volume per unit time) through the surface element: v  n dS .
v  n dt v dt
n
Figure 1.7.7: flow through a surface element
The total (volume) flux out of a surface S is then6
volume flux:  v  ndS

S
(1.7.8)
Similarly, the mass flux is given by
mass flux:  v  ndS

S
(1.7.9)
For more complex surfaces, one can write using Eqn. 1.7.3, 1.7.5,
 x x 
 f  ndS   f  t
S
S 1

t 2
dt1 dt 2

Example (of a Flux Integral)
Compute the flux integral  f  ndS , where S is the parabolic cylinder represented by
S
x 2  x12 , 0  x1  2, 0  x3  3
and f  x 2 e1  2e 2  x1 x3 e 3 , Fig. 1.7.8.
Solution
Making the substitutions x1  t1 , x3  t 2 , so that x 2  t12 , the surface can be represented

by the position vector
6
if v acts in the same direction as n, i.e. pointing outward, the dot product is positive and this integral is
positive; if, on the other hand, material is flowing in through the surface, v and n are in opposite directions
and the dot product is negative, so the integral is negative

Section 1.7
x  t1e1  t12 e 2  t 2 e 3 , 0  t1  2, 0  t 2  3
Then x / t1  e1  2t1e 2 , x / t 2  e 3 and
x x
  2t1e1  e 2
t1 t 2
so the integral becomes
3 2
  t 
e  2e 2  t1t 2 e 3 2t1e1  e 2 dt1 dt 2  12
2
1 1
0 0
x3
n 
f
x2
x1
Figure 1.7.8: flux through a parabolic cylinder
Note: in this example, the value of the integral depends on the choice of n. If one chooses
 n instead of n, one would obtain  12 . The normal in the opposite direction (on the
“other side” of the surface) can be obtained by simply switching t1 and t 2 , since
x / t1  x / t 2  x / t 2  x / t1 .
■
Surface flux integrals can also be evaluated by first converting them into double integrals
over a plane region. For example, if a surface S has a projection R on the x1  x 2 plane,
then an element of surface dS is related to the projected element dx1 dx 2 through (see
Fig. 1.7.9)
cos dS  n  e 3 dS  dx1 dx 2
and so
1
 f  n dS   f  n n  e
S R
3
dx1 dx 2

Section 1.7
e3 n

x2
x1
Figure 1.7.9: projection of a surface element onto a plane region
The Normal and Surface Area Elements
It is sometimes convenient to associate a special vector dS with a differential element of

surface area dS , where
dS  n dS
so that dS is the vector with magnitude dS and direction of the unit normal to the
surface. Flux integrals can then be written as
 f  ndS   f dS
S S
1.7.6 Volume Integrals

The volume integral, or triple integral, is a generalisation of the double integral.
Change of Variable in Volume Integrals
For a volume integral, it is often convenient to make the change of variables

( x1 , x 2 , x3 )  (t1 , t 2 , t 3 ) . The volume of an element dV is given by the triple scalar
product (Eqns. 1.1.5, 1.3.17)
 x x  x
dV      dt1 dt 2 dt 3  Jdt1 dt 2 dt 3 (1.7.10)
 t1 t 2  t 3
where the Jacobian is now

Section 1.7
x1 x 2 x3 x1 x1 x1

t1 t1 t1 t1 t 2 t 3
x x 2 x3 x x 2 x 2
J 1 or J 2 (1.7.11)
t 2 t 2 t 2 t1 t 2 t 3
x1 x 2 x3 x3 x3 x3
t 3 t 3 t 3 t1 t 2 t 3
so that
 f x , x
V
1 2 , x3 dxdydz   f  x1 t1 , t 2 , t 3 , x 2 t1 , t 2 , t 3 , x3 t1 , t 2 , t 3 J dt1 dt 2 dt 3
V
1.7.7 Integral Theorems
A number of integral theorems and relations are presented here (without proof), the most
important of which is the divergence theorem. These theorems can be used to simplify
the evaluation of line, double, surface and triple integrals. They can also be used in
various proofs of other important results.
The Divergence Theorem
Consider an arbitrary differentiable vector field v ( x, t ) defined in some finite region of

physical space. Let V be a volume in this space with a closed surface S bounding the
volume, and let the outward normal to this bounding surface be n. The divergence
theorem of Gauss states that (in symbolic and index notation)
vi
 v  ndS   div v dV
S V
 v n dS   x
S
i i
V i
dV Divergence Theorem (1.7.12)
and one has the following useful identities {▲Problem 10}
  u  ndS   div ( u) dV
S V
 ndS   grad  dV
S V
(1.7.13)
 n  udS   curl u dV
S V
By applying the divergence theorem to a very small volume, one finds that
div v  lim
 v  ndS
S
V 0 V
that is, the divergence is equal to the outward flux per unit volume, the result 1.6.18.

Section 1.7
Stoke’s Theorem
Stoke’s theorem transforms line integrals into surface integrals and vice versa. It states
that
 curlf   ndS   f  τ ds
S C
(1.7.14)
Here C is the boundary of the surface S, n is the unit outward normal and τ  dr / ds is
the unit tangent vector.
As has been seen, Eqn. 1.6.24, the curl of the velocity field is a measure of how much a
fluid is rotating. The direction of this vector is along the direction of the local axis of
rotation and its magnitude measures the local angular velocity of the fluid. Stoke’s
theorem then states that the amount of rotation of a fluid can be measured by integrating
the tangential velocity around a curve (the line integral), or by integrating the amount of
vorticity “moving through” a surface bounded by the same curve.
Green’s Theorem and Related Identities
Green’s theorem relates a line integral to a double integral, and states that
  2  1 
  dx
C
1 1   2 dx 2    
R 
x1

x
dx1 dx 2 ,
2 
(1.7.15)
where R is a region in the x1  x 2 plane bounded by the curve C. In vector form, Green’s
theorem reads as
 f  dx   curlf  e dx dx
C R
3 1 2 where f   1e 1   2 e 2 (1.7.16)
from which it can be seen that Green’s theorem is a special case of Stoke’s theorem, for
the case of a plane surface (region) in the x1  x 2 plane.
It can also be shown that (this is Green’s first identity)
 n  grad dS     grad  grad dV

2
(1.7.17)
S V
Note that the term n  grad is the directional derivative of  in the direction of the
outward unit normal. This is often denoted as  / n . Green’s first identity can be
regarded as a multi-dimensional “integration by parts” – compare the rule
 udv  uv   vdu with the identity re-written as
    dV     n dS       dV
V S V
(1.7.18)

Section 1.7
or
   u dV   u  n dS      udV

V S V
(1.7.18)
One also has the relation (this is Green’s second identity)
  n  grad    n  grad dS       dV

2 2
(1.7.19)
S V
1.7.8 Problems
1. Find the work done in moving a particle in a force field given by
f  3x1 x 2 e1  5 x3 e 2  10 x1e1 along the curve x1  t 2  1 , x 2  2t 2 , x3  t 3 , from t  1
to t  2 . (Plot the curve.)
2. Show that the following vectors are conservative and find their scalar potentials:
(i) x  x1e1  x 2 e 2  x3 e 3
(ii) v  e x1x2  x 2 e1  x1e 2 
(iii) u  (1 / x 2 )e1  ( x1 / x 22 )e 2  x3 e 3
3. Show that if f   then curl f  o .
4. Show that if f    a then   f  0 .
5. Find the volume beneath the surface x12  x 22  x3  0 and above the square with
vertices (0,0) , (1,0) , (1,1) and (0,1) in the x1  x 2 plane.
6. Find the Jacobian (and sketch lines of constant t1 , t 2 ) for the rotation
x1  t1 cos   t 2 sin 
x 2  t1 sin   t 2 cos 
7. Find a unit normal to the circular cylinder with parametric representation
x(t1 , t 2 )  a cos t1e1  a sin t1e 2  t 2 e 3 , 0  t1  2 , 0  t1  1
8. Evaluate   dS
S
where   x1  x 2  x3 and S is the plane surface x3  x1  x 2 ,
0  x 2  x1 , 0  x1  1 .
9. Evaluate the flux integral  f  ndS where f  e1  2e 2  2e 3 and S is the cone
S
x3  a x  x , x3  a [Hint: first parameterise the surface with t1 , t 2 .]

2
1
2
2
10. Prove the relations in (1.7.13). [Hint: first write the expressions in index notation.]
11. Use the divergence theorem to show that
S x  ndS  3V
,
where V is the volume enclosed by S (and x is the position vector).
12. Verify the divergence theorem for v  x13e1  x 23e 2  x33 e 3 where S is the surface of the
sphere x12  x 22  x32  a 2 .
13. Interpret the divergence theorem (1.7.12) for the case when v is the velocity field.
See (1.6.18, 1.7.8). Interpret also the case of divv  0 .

Section 1.7
14. Verify Stoke’s theorem for f  x 2 e1  x3e 2  x1e 3 where S is x3  1  x12  x 22  0 (so
that C is the circle of radius 1 in the x1  x 2 plane).
15. Verify Green’s theorem for the case of  1  x12  2 x 2 ,  2  x1  x 2 , with C the unit
circle x12  x 22  1 . The following relations might be useful:
2 2 2 2

0
sin 2 d   cos 2 d   ,
0 
0
sin  cos d   sin  cos 2 d  0
0
16. Evaluate  f  dx using Green’s theorem, where f   x e  x13 e 2 and C is the circle
3
2 1
C
x  x  4.
2
1
2
2
17. Use Green’s theorem to show that the double integral of the Laplacian of p over a
region R is equivalent to the integral of p / n  gradp  n around the curve C
bounding the region:
p
R  pdx1dx2  C n ds
2
[Hint: Let  1  p / x 2 ,  2  p / x1 . Also, show that

dx dx
n  2 e1  1 e 2
ds ds
is a unit normal to C, Fig. 1.7.10]
ds
dx2
C dx1
Figure 1.7.10: projection of a surface element onto a plane region

Section 1.8
1.8 Tensors
Here the concept of the tensor is introduced. Tensors can be of different orders – zeroth-
order tensors, first-order tensors, second-order tensors, and so on. Apart from the zeroth
and first order tensors (see below), the second-order tensors are the most important
tensors from a practical point of view, being important quantities in, amongst other topics,
continuum mechanics, relativity, electromagnetism and quantum theory.
1.8.1 Zeroth and First Order Tensors
A tensor of order zero is simply another name for a scalar  .
A first-order tensor is simply another name for a vector u.
1.8.2 Second Order Tensors
Notation
Vectors: lowercase bold-face Latin letters, e.g. a, r, q

2nd order Tensors: uppercase bold-face Latin letters, e.g. F, T, S
Tensors as Linear Operators
A second-order tensor T may be defined as an operator that acts on a vector u generating

another vector v, so that T(u)  v , or1
T  u  v or Tu  v Second-order Tensor (1.8.1)
The second-order tensor T is a linear operator (or linear transformation)2, which

means that
Ta  b   Ta  Tb … distributive
Ta    Ta  … associative
This linearity can be viewed geometrically as in Fig. 1.8.1.
Note: the vector may also be defined in this way, as a mapping u that acts on a vector v,
this time generating a scalar α, u  v   . This transformation (the dot product) is linear
(see properties (2,3) in §1.1.4). Thus a first-order tensor (vector) maps a first-order tensor
into a zeroth-order tensor (scalar), whereas a second-order tensor maps a first-order tensor
into a first-order tensor. It will be seen that a third-order tensor maps a first-order tensor
into a second-order tensor, and so on.
1
both these notations for the tensor operation are used; here, the convention of omitting the “dot” will be
used
2
An operator or transformation is a special function which maps elements of one type into elements of a
similar type; here, vectors into vectors

Section 1.8
Ta  b 
a Tb
ab
Ta
b
Figure 1.8.1: Linearity of the second order tensor
Further, two tensors T and S are said to be equal if and only if
Sv  Tv
for all vectors v.
Example (of a Tensor)
Suppose that F is an operator which transforms every vector into its mirror-image with
respect to a given plane, Fig. 1.8.2. F transforms a vector into another vector and the
transformation is linear, as can be seen geometrically from the figure. Thus F is a
second-order tensor.
Fu  u
Fu u
Fu  v  uv
Fv
Figure 1.8.2: Mirror-imaging of vectors as a second order tensor mapping
Example (of a Tensor)
The combination u  linearly transforms a vector into another vector and is thus a
second-order tensor3. For example, consider a force f applied to a spanner at a distance r
from the centre of the nut, Fig. 1.8.3. Then it can be said that the tensor r  maps the
force f into the (moment/torque) vector r  f .
3 ~ to denote u 
Some authors use the notation u

Section 1.8
Figure 1.8.3: the force on a spanner

■
1.8.3 The Dyad (the tensor product)

The vector dot product and vector cross product have been considered in previous
sections. A third vector product, the tensor product (or dyadic product), is important in
the analysis of tensors of order 2 or more. The tensor product of two vectors u and v is
written as4
uv Tensor Product (1.8.2)
This tensor product is itself a tensor of order two, and is called dyad:
u  v is a scalar (a zeroth order tensor)

u  v is a vector (a first order tensor)
u  v is a dyad (a second order tensor)
It is best to define this dyad by what it does: it transforms a vector w into another vector
with the direction of u according to the rule5
(u  v ) w  u ( v  w ) The Dyad Transformation (1.8.3)
This relation defines the symbol “  ”.
The length of the new vector is u times v  w , and the new vector has the same direction
as u , Fig. 1.8.4. It can be seen that the dyad is a second order tensor, because it operates
linearly on a vector to give another vector {▲Problem 2}.
Note that the dyad is not commutative, u  v  v  u . Indeed it can be seen clearly from
the figure that u  v w  v  u w .
4
many authors omit the  and write simply uv
5
note that it is the two vectors that are beside each other (separated by a bracket) that get “dotted” together

Section 1.8
(u  v ) w
Figure 1.8.4: the dyad transformation
The following important relations follow from the above definition {▲Problem 4},
u  v w  x   v  w u  x 
(1.8.4)
uv  w   u  v w
It can be seen from these that the operation of the dyad on a vector is not commutative:
uv  w   v  w u (1.8.5)
Example (The Projection Tensor)
Consider the dyad e  e . From the definition 1.8.3, e  e u  e  u e . But e  u is the

projection of u onto a line through the unit vector e. Thus e  u e is the vector projection
of u on e. For this reason e  e is called the projection tensor. It is usually denoted by
P.
Pu e Pv
Figure 1.8.5: the projection tensor

Section 1.8
1.8.4 Dyadics
A dyadic is a linear combination of these dyads (with scalar coefficients). An example
might be
5a  b   3c  d   2e  f 
This is clearly a second-order tensor. It will be seen in §1.9 that every second-order
tensor can be represented by a dyadic, that is
T   a  b    c  d    e  f    (1.8.6)
Note: second-order tensors cannot, in general, be written as a dyad, T  a  b – when

they can, they are called simple tensors.
Example (Angular Momentum and the Moment of Inertia Tensor)
Suppose a rigid body is rotating so that every particle in the body is instantaneously
moving in a circle about some axis fixed in space, Fig. 1.8.6.
d
v
 r
Figure 1.8.6: a particle in motion about an axis
The body’s angular velocity ω is defined as the vector whose magnitude is the angular
speed  and whose direction is along the axis of rotation. Then a particle’s linear
velocity is
v  ωr
where v  wd is the linear speed, d is the distance between the axis and the particle, and r
is the position vector of the particle from a fixed point O on the axis. The particle’s
angular momentum (or moment of momentum) h about the point O is defined to be
h  mr  v
where m is the mass of the particle. The angular momentum can be written as

Section 1.8
h  Iˆω (1.8.8)
where Î , a second-order tensor, is the moment of inertia of the particle about the point
O, given by

Iˆ  m r I  r  r
2
 (1.8.9)
where I is the identity tensor, i.e. Ia  a for all vectors a.
 2

To show this, it must be shown that r  v  r I  r  r ω . First examine r  v . It is
evidently a vector perpendicular to both r and v and in the plane of r and ω ; its
magnitude is
2
r  v  r v  r ω sin 
Now (see Fig. 1.8.7)
r 2

I  r  r ω  r ω  r r  ω 
2
 r ω e   cos e r 
2
where e  and e r are unit vectors in the directions of ω and r respectively. From the
2
diagram, this is equal to r ω sin e h . Thus both expressions are equivalent, and one
can indeed write h  Iˆω with Î defined by Eqn. 1.8.9: the second-order tensor Î maps
the angular velocity vector ω into the angular momentum vector h of the particle.
cos 
e
eh
 er
sin 
 cos e r
Figure 1.8.7: geometry of unit vectors for angular momentum calculation
1.8.5 The Vector Space of Second Order Tensors

The vector space of vectors and associated spaces were discussed in §1.2. Here, spaces of
second order tensors are discussed.
As mentioned above, the second order tensor is a mapping on the vector space V,

Section 1.8
T :V  V (1.8.10)
and follows the rules
Ta  b   Ta  Tb
(1.8.11)
Ta    Ta 
for all a, b  V and   R .
Denote the set of all second order tensors by V 2 . Define then the sum of two tensors
S, T  V 2 through the relation
S  Tv  Sv  Tv (1.8.12)
and the product of a scalar   R and a tensor T  V 2 through
Tv  Tv (1.8.13)
Define an identity tensor I  V 2 through
Iv  v , for all v  V (1.8.14)
and a zero tensor O  V 2 through
Ov  o , for all v  V (1.8.15)
It follows from the definition 1.8.11 that V 2 has the structure of a real vector space, that
is, the sum S  T  V 2 , the product T  V 2 , and the following 8 axioms hold:
1. for any A, B, C  V 2 , one has ( A  B)  C  A  (B  C)

2. there exists an element O  V 2 such that T  O  O  T  T for every T  V 2
3. for each T  V 2 there exists an element  T  V 2 , called the negative of T, such that
T  (  T)  (  T)  T  0
4. for any S, T  V 2 , one has S  T  T  S
5. for any S, T  V 2 and scalar   R ,  (S  T)  S  T
6. for any T  V 2 and scalars  ,   R , (   )T  T  T
7. for any T  V 2 and scalars  ,   R ,  (  T)  ( )T
8. for the unit scalar 1  R , 1T  T for any T  V 2 .

Section 1.8
1.8.6 Problems
1. Consider the function f which transforms a vector v into a  v   . Is f a tensor (of

order one)? [Hint: test to see whether the transformation is linear, by examining
f u  v  .]
2. Show that the dyad is a linear operator, in other words, show that
(u  v )w   x    (u  v )w   (u  v )x
3. When is a  b  b  a ?
4. Prove that
(i) u  v w  x   v  w u  x  [Hint: post-“multiply” both sides of the definition
(1.8.3) by  x ; then show that u  v w   x  u  v w  x  .]
(ii) uv  w   u  v w [hint: pre “multiply” both sides by x  and use the result of
(i)]
5. Consider the dyadic (tensor) a  a  b  b . Show that this tensor orthogonally
projects every vector v onto the plane formed by a and b (sketch a diagram).
6. Draw a sketch to show the meaning of u  Pv  , where P is the projection tensor.
What is the order of the resulting tensor?
7. Prove that a  b  b  a  b  a   .

Section 1.9
1.9 Cartesian Tensors

As with the vector, a (higher order) tensor is a mathematical object which represents
many physical phenomena and which exists independently of any coordinate system. In
what follows, a Cartesian coordinate system is used to describe tensors.
1.9.1 Cartesian Tensors

A second order tensor and the vector it operates on can be described in terms of Cartesian
components. For example, (a  b)c , with a  2e1  e 2  e 3 , b  e1  2e 2  e 3 and
c  e1  e 2  e 3 , is
(a  b)c  a(b  c)  4e1  2e 2  2e 3
Example (The Unit Dyadic or Identity Tensor)
The identity tensor, or unit tensor, I, which maps every vector onto itself, has been
introduced in the previous section. The Cartesian representation of I is
e 1  e1  e 2  e 2  e 3  e 3  e i  e i (1.9.1)
This follows from
e1  e1  e 2  e 2  e 3  e 3 u  e1  e1 u  e 2  e 2 u  e 3  e 3 u
 e1 e1  u   e 2 e 2  u   e 3 e 3  u 
 u1e1  u 2 e 2  u 3 e 3
u
Note also that the identity tensor can be written as I   ij e i  e j  , in other words the
Kronecker delta gives the components of the identity tensor in a Cartesian coordinate
system.
■
Second Order Tensor as a Dyadic
In what follows, it will be shown that a second order tensor can always be written as a
dyadic involving the Cartesian base vectors ei 1.
Consider an arbitrary second-order tensor T which operates on a to produce b, T(a)  b ,

or T(ai e i )  b . From the linearity of T,
1
this can be generalised to the case of non-Cartesian base vectors, which might not be orthogonal nor of
unit magnitude (see §1.16)

Section 1.9
a1T(e1 )  a 2 T(e 2 )  a3 T(e 3 )  b
Just as T transforms a into b, it transforms the base vectors ei into some other vectors;
suppose that T(e1 )  u, T(e 2 )  v, T(e 3 )  w , then
b  a1u  a 2 v  a3 w
 a  e1 u  a  e 2 v  a  e 3 w
 u  e1 a  v  e 2 a  w  e 3 a
 u  e1  v  e 2  w  e 3 a
and so
T  u  e1  v  e 2  w  e 3 (1.9.2)
which is indeed a dyadic.
Cartesian components of a Second Order Tensor
The second order tensor T can be written in terms of components and base vectors as
follows: write the vectors u, v and w in (1.9.2) in component form, so that
T  u1e1  u 2 e 2  u 3e 3   e1    e 2    e 3

 u1e1  e1  u 2 e 2  e1  u 3 e 3  e1  
Introduce nine scalars Tij by letting u i  Ti1 , vi  Ti 2 , wi  Ti 3 , so that
T  T11e1  e1  T12 e1  e 2  T13 e1  e 3

 T21e 2  e1  T22 e 2  e 2  T23 e 2  e 3 Second-order Cartesian Tensor (1.9.3)
 T31e 3  e1  T32 e 3  e 2  T33e 3  e 3
These nine scalars Tij are the components of the second order tensor T in the Cartesian
coordinate system. In index notation,
T  Tij e i  e j 
Thus whereas a vector has three components, a second order tensor has nine components.
Similarly, whereas the three vectors e i  form a basis for the space of vectors, the nine
dyads e i  e j  from a basis for the space of tensors, i.e. all second order tensors can be
expressed as a linear combination of these basis tensors.
It can be shown that the components of a second-order tensor can be obtained directly
from {▲Problem 1}
Tij  e i Te j Components of a Tensor (1.9.4)

Section 1.9
which is the tensor expression analogous to the vector expression u i  e i  u . Note that,
in Eqn. 1.9.4, the components can be written simply as e i Te j (without a “dot”), since
e i  Te j  e i T  e j .
Example (The Stress Tensor)
Define the traction vector t acting on a surface element within a material to be the force
acting on that element2 divided by the area of the element, Fig. 1.9.1. Let n be a vector
normal to the surface. The stress σ is defined to be that second order tensor which maps
n onto t, according to
t  σn The Stress Tensor (1.9.5)
x2
n  31
 11
t t
 21
x1
e1
x3
Figure 1.9.1: stress acting on a plane
If one now considers a coordinate system with base vectors e i , then σ   ij e i  e j and,
for example,
σe1   11e1   21e 2   31e 3
Thus the components  11 ,  21 and  31 of the stress tensor are the three components of
the traction vector which acts on the plane with normal e1 .
Augustin-Louis Cauchy was the first to regard stress as a linear map of the normal vector
onto the traction vector; hence the name “tensor”, from the French for stress, tension.
2
this force would be due, for example, to intermolecular forces within the material: the particles on one side
of the surface element exert a force on the particles on the other side

Section 1.9
Higher Order Tensors
The above can be generalised to tensors of order three and higher. The following notation
will be used:
, ,  … 0th-order tensors (“scalars”)

a, b, c … 1st-order tensors (“vectors”)
A, B, C … 2nd-order tensors (“dyadics”)
A, B, C … 3rd-order tensors (“triadics”)
A, B, C … 4th-order tensors (“tetradics”)
An important third-order tensor is the permutation tensor, defined by
E   ijk e i  e j  e k (1.9.6)
whose components are those of the permutation symbol, Eqns. 1.3.10-1.3.13.
A fourth-order tensor can be written as
A  Aijkl e i  e j  e k  e l (1.9.7)
It can be seen that a zeroth-order tensor (scalar) has 3 0  1 component, a first-order tensor
has 31  3 components, a second-order tensor has 3 2  9 components, so A has 33  27
components and A has 81 components.
1.9.2 Simple Contraction

Tensor/vector operations can be written in component form, for example,
Ta  Tij e i  e j a k e k

 Tij a k e i  e j e k 
(1.9.8)
 Tij a k  jk e i
 Tij a j e i
This operation is called simple contraction, because the order of the tensors is contracted
– to begin there was a tensor of order 2 and a tensor of order 1, and to end there is a
tensor of order 1 (it is called “simple” to distinguish it from “double” contraction – see
below). This is always the case – when a tensor operates on another in this way, the order
of the result will be two less than the sum of the original orders.
An example of simple contraction of two second order tensors has already been seen in
Eqn. 1.8.4a; the tensors there were simple tensors (dyads). Here is another example:

Section 1.9
TS  Tij e i  e j S kl e k  e l 

 Tij S kl e i  e j e k  e l  
(1.9.9)
 Tij S kl  jk e i  e l 
 Tij S jl e i  e l 
From the above, the simple contraction of two second order tensors results in another
second order tensor. If one writes A  TS , then the components of the new tensor are
related to those of the original tensors through Aij  Tik S kj .
Note that, in general,
AB  BA
ABC  ABC … associative (1.9.10)
AB  C  AB  AC … distributive
The associative and distributive properties follow from the fact that a tensor is by
definition a linear operator, §1.8.2; they apply to tensors of any order, for example,
ABv  ABv  (1.9.11)
To deal with tensors of any order, all one has to remember is how simple tensors operate
on each other – the two vectors which are beside each other are the ones which are
“dotted” together:
a  b c  b  c a
a  b c  d   b  c a  d 
(1.9.12)
a  b c  d  e   b  c a  d  e 
a  b  c d  e  f   c  d a  b  e  f 
An example involving a higher order tensor is
A  E  Aijkl E mn e i  e j  e k  e l e m  e n 
 Aijkl Eln e i  e j  e k  e n 
and
uv 
AB  C
Au  v
Ab  C
AB  C
Note the relation {▲Problem 10}

Section 1.9
AB  CD  AB   CD (1.9.13)
Powers of Tensors
Integral powers of tensors are defined inductively by T 0  I , T n  T n 1T , so, for

example,
T 2  TT The Square of a Tensor (1.9.14)
T 3  TTT , etc.
1.9.3 Double Contraction

Double contraction, as the name implies, contracts the tensors twice as much a simple
contraction. Thus, where the sum of the orders of two tensors is reduced by two in the
simple contraction, the sum of the orders is reduced by four in double contraction. The
double contraction is denoted by a colon (:), e.g. T : S .
First, define the double contraction of simple tensors (dyads) through
a  b  : c  d   a  cb  d  (1.9.15)
So in double contraction, one takes the scalar product of four vectors which are adjacent
to each other, according to the following rule:
a  b  c  : d  e  f   b  d c  e a  f 
For example,
T : S  Tij e i  e j  : S kl e k  e l 
 
 Tij S kl e i  e k e j  e l  (1.9.16)
 Tij S ij
which is, as expected, a scalar.
Here is another example, the contraction of the two second order tensors I (see Eqn.
1.9.1) and u  v ,
I : u  v  e i  e i  : u  v 
 e i  u e i  v 
(1.9.17)
 u i vi
 uv

Section 1.9
so that the scalar product of two vectors can be written in the form of a double contraction
involving the Identity Tensor.
An example of double contraction involving the permutation tensor 1.9.6 is {▲Problem

11}
E : u  v   v  u (1.9.18)
It can be shown that the components of a fourth order tensor are given by (compare with
Eqn. 1.9.4)
Aijkl  e i  e j  : A : e k  e l  (1.9.19)
In summary then,
A:B  
A :b 
A:B  c
A:B  C
Note the following identities:
A  B  : C  AB : C  B : CA
A : B  C  CA : B   A : B C (1.9.20)
A  B  : C  D  B : CA  D  A  DB : C
Note: There are many operations that can be defined and performed with tensors. The
two most important operations, the ones which arise most in practice, are the simple and
double contractions defined above. Other possibilities are:
(a) double contraction with two “horizontal” dots, T S , A b , etc., which is based on
the definition of the following operation as applied to simple tensors:
 a  b  c    d  e  f    b  e  c  d  a  f 
(b) operations involving one cross   :  a  b    c  d    a  d    b  c 
(c) “double” operations involving the cross  and dot:
 a  b   c  d    a  c    b  d 
 a  b   c  d    a  c  b  d 
 a  b   c  d    a  c  b  d 
1.9.4 Index Notation

The index notation for single and double contraction of tensors of any order can easily be
remembered. From the above, a single contraction of two tensors implies that the indices

Section 1.9
“beside each other” are the same3, and a double contraction implies that a pair of indices
is repeated. Thus, for example, in both symbolic and index notation:
AB  C Aijm Bmk  Cijk

(1.9.21)
A:B  c Aijk B jk  ci
1.9.5 Matrix Notation

Here the matrix notation of §1.4 is extended to include second-order tensors4. The
Cartesian components of a second-order tensor can conveniently be written as a 3  3
matrix,
 T11 T12 T13 

T  T21 T22 T13 
T31 T32 T33 
The operations involving vectors and second-order tensors can now be written in terms of
matrices, for example,
 T11 T12 T13   u1   T11u1  T12 u 2  T13 u 3 

Tu  T u   T21 T22 T13  u 2   T21u1  T22 u 2  T23 u 3 
T31 T32 T33  u 3  T31u1  T32 u 2  T33 u 3 
symbolic “short”
notation matrix “full”
notation matrix
notation
The tensor product can be written as (see §1.4.1)
 u1v1 u1v 2 u1v3 

u  v  u v   T
 u 2 v1 u 2 v 2 u 2 v3  (1.9.22)
u 3 v1 u 3 v2 u 3 v3 
which is consistent with the definition of the dyadic transformation, Eqn. 1.8.3.
3
compare with the “beside each other rule” for matrix multiplication given in §1.4.1
4
the matrix notation cannot be used for higher-order tensors

Section 1.9
1.9.6 Problems
1. Use Eqn. 1.9.3 to show that the component T11 of a tensor T can be evaluated from
e1Te1 , and that T12  e1Te 2 (and so on, so that Tij  e i Te j ).
2. Evaluate aT using the index notation (for a Cartesian basis). What is this operation
called? Is your result equal to Ta , in other words is this operation commutative?
Now carry out this operation for two vectors, i.e. a  b . Is it commutative in this case?
3. Evaluate the simple contractions Ab and AB , with respect to a Cartesian coordinate
system (use index notation).
4. Evaluate the double contraction A : B (use index notation).
5. Show that, using a Cartesian coordinate system and the index notation, that the double
contraction A : b is a scalar. Write this scalar out in full in terms of the components
of A and b.
6. Consider the second-order tensors
D  3e1  e1  2e 2  e 2  e 2  e 3  5e 3  e 3
F  4e1  e 3  6e 2  e 2  3e 3  e 2  e 3  e 3
Compute DF and F : D .
7. Consider the second-order tensor
D  3e1  e1  4e1  e 2  2e 2  e1  e 2  e 2  e 3  e 3 .
Determine the image of the vector r  4e1  2e 2  5e 3 when D operates on it.
8. Write the following out in full – are these the components of scalars, vectors or
second order tensors?
(a) Bii
(b) C kkj
(c) Bmn
(d) ai b j Aij
9. Write a  b  : c  d  in terms of the components of the four vectors. What is the
order of the resulting tensor?
10. Verify Eqn. 1.9.13.
11. Show that E : u  v   v  u – see (1.9.6, 1.9.18). [Hint: use the definition of the
cross product in terms of the permutation symbol, (1.3.14), and the fact that
 ijk   kji .]

Section 1.10
1.10 Special Second Order Tensors & Properties of

Second Order Tensors
In this section will be examined a number of special second order tensors, and special
properties of second order tensors, which play important roles in tensor analysis. Many
of the concepts will be familiar from Linear Algebra and Matrices. The following will be
discussed:
 The Identity tensor

 Transpose of a tensor
 Trace of a tensor
 Norm of a tensor
 Determinant of a tensor
 Inverse of a tensor
 Orthogonal tensors
 Rotation Tensors
 Change of Basis Tensors
 Symmetric and Skew-symmetric tensors
 Axial vectors
 Spherical and Deviatoric tensors
 Positive Definite tensors
1.10.1 The Identity Tensor

The linear transformation which transforms every tensor into itself is called the identity
tensor. This special tensor is denoted by I so that, for example,
Ia  a for any vector a
In particular, Ie1  e1 , Ie 2  e 2 , Ie 3  e 3 , from which it follows that, for a Cartesian

coordinate system, I ij   ij . In matrix form,
1 0 0
I  0 1 0 (1.10.1)
0 0 1
1.10.2 The Transpose of a Tensor
The transpose of a second order tensor A with components Aij is the tensor A T with
components A ji ; so the transpose swaps the indices,
A  Aij e i  e j , A T  A ji e i  e j Transpose of a Second-Order Tensor (1.10.2)

Section 1.10
In matrix notation,
 A11 A12 A13   A11 A21 A31 

A   A21 A22 A23 , A 
T
  A12 A22 A32 
 A31 A32 A33   A13 A23 A33 
Some useful properties and relations involving the transpose are {▲Problem 2}:
A  T T
A
A  B   A T  B T
T
u  v T  v  u
Tu  uT T , uT  T T u
(1.10.3)
AB T  BT AT
A : B  AT : BT
(u  v ) A  u  ( A T v )
 
A : BC   B T A : C  AC T : B  
A formal definition of the transpose which does not rely on any particular coordinate
system is as follows: the transpose of a second-order tensor is that tensor which satisfies
the identity1
u  Av  v  A T u (1.10.4)
for all vectors u and v. To see that Eqn. 1.10.4 implies 1.10.2, first note that, for the
present purposes, a convenient way of writing the components Aij of the second-order
tensor A is A ij . From Eqn. 1.9.4, A ij  e i  Ae j and the components of the transpose
can be written as A T   ij  e i  A T e j . Then, from 1.10.4,
A T
ij  e i  A T e j  e j  Ae i  A  ji  A ji .
1.10.3 The Trace of a Tensor
The trace of a second order tensor A, denoted by trA , is a scalar equal to the sum of the
diagonal elements of its matrix representation. Thus (see Eqn. 1.4.3)
trA  Aii Trace (1.10.5)
A more formal definition, again not relying on any particular coordinate system, is
trA  I : A Trace (1.10.6)
1
as mentioned in §1.9, from the linearity of tensors, uA  v  u  Av and, for this reason, this expression is
usually written simply as uAv

Section 1.10
and Eqn. 1.10.5 follows from 1.10.6 {▲Problem 4}. For the dyad u  v {▲Problem 5},
tr u  v   u  v (1.10.7)
Another example is
tr(E 2 )  I : E 2
  ij e i  e j  : E pq E qr e p  e r  (1.10.8)
 Eiq E qi
This and other important traces, and functions of the trace are listed here {▲Problem 6}:
trA  Aii
trA 2  Aij A ji
trA 3  Aij A jk Aki (1.10.9)
trA 2  Aii A jj
trA 3  Aii A jj Akk
Some useful properties and relations involving the trace are {▲Problem 7}:
trA T  trA
tr AB   tr BA 
tr A  B   trA  trB (1.10.10)
tr A   trA
    
A : B  tr A T B  tr AB T  tr B T A  tr BA T   
The double contraction of two tensors was earlier defined with respect to Cartesian
coordinates, Eqn. 1.9.16. This last expression allows one to re-define the double
contraction in terms of the trace, independent of any coordinate system.
Consider again the real vector space of second order tensors V 2 introduced in §1.8.5.
The double contraction of two tensors as defined by 1.10.10e clearly satisfies the
requirements of an inner product listed in §1.2.2. Thus this scalar quantity serves as an
inner product for the space V 2 :
A, B  A : B  tr A T B   (1.10.11)
and generates an inner product space.
Just as the base vectors e i  form an orthonormal set in the inner product (vector dot
product) of the space of vectors V, so the base dyads e i  e j  form an orthonormal set in
the inner product 1.10.11 of the space of second order tensors V 2 . For example,

Section 1.10
e1  e1 , e1  e1  e1  e1  : e1  e1   1 (1.10.12)
Similarly, just as the dot product is zero for orthogonal vectors, when the double
contraction of two tensors A and B is zero, one says that the tensors are orthogonal,
A : B  tr A T B   0 , A, B orthogonal (1.10.13)
1.10.4 The Norm of a Tensor
Using 1.2.8 and 1.10.11, the norm of a second order tensor A, denoted by A (or A ), is
defined by
A  A:A (1.10.14)
This is analogous to the norm a of a vector a, aa .
1.10.5 The Determinant of a Tensor

The determinant of a second order tensor A is defined to be the determinant of the
matrix A of components of the tensor:
 A11 A12 A13 

det A  det  A21 A22 A23 
 A31 A32 A33 
  ijk Ai1 A j 2 Ak 3 (1.10.15)
  ijk A1i A2 j A3k
Some useful properties of the determinant are {▲Problem 8}
det( AB)  det A det B

det A T  det A
det(A)   3 det A
(1.10.16)
det u  v   0
 pqr det A    ijk Aip A jq Akr
Ta  Tb   Tc  det Ta  b   c
Note that det A , like trA , is independent of the choice of coordinate system / basis.

Section 1.10
1.10.6 The Inverse of a Tensor
The inverse of a second order tensor A, denoted by A 1 , is defined by
AA 1  I  A 1 A (1.10.17)
The inverse of a tensor exists only if it is non-singular (a singular tensor is one for
which det A  0 ), in which case it is said to be invertible.
Some useful properties and relations involving the inverse are:
( A 1 ) 1  A
(A) 1  (1 /  ) A 1
(1.10.18)
( AB) 1  B 1 A 1
det( A 1 )  (det A) 1
Since the inverse of the transpose is equivalent to the transpose of the inverse, the
following notation is used:
A  T  ( A 1 ) T  ( A T ) 1 (1.10.19)
1.10.7 Orthogonal Tensors

An orthogonal tensor Q is a linear vector transformation satisfying the condition
Qu  Qv  u  v (1.10.20)
for all vectors u and v. Thus u is transformed to Qu , v is transformed to Qv and the dot
product u  v is invariant under the transformation. Thus the magnitude of the vectors
and the angle between the vectors is preserved, Fig. 1.10.1.
Q Qu
 
v
Qv
Figure 1.10.1: An orthogonal tensor
Since
Qu  Qv  uQT  Qv  u   QTQ   v (1.10.21)

Section 1.10
it follows that for u  v to be preserved under the transformation, Q T Q  I , which is also

used as the definition of an orthogonal tensor. Some useful properties of orthogonal
tensors are {▲Problem 10}:
QQ T  I  Q T Q, Qik Q jk   ij  Qki Qkj

Q 1  Q T (1.10.22)
det Q  1
1.10.8 Rotation Tensors
If for an orthogonal tensor, det Q  1 , Q is said to be a proper orthogonal tensor,

corresponding to a rotation. If det Q  1 , Q is said to be an improper orthogonal
tensor, corresponding to a reflection. Proper orthogonal tensors are also called rotation
tensors.
1.10.9 Change of Basis Tensors
Consider a rotation tensor Q which rotates the base vectors e1 , e 2 , e 3 into a second set,
e1 , e2 , e3 , Fig. 1.10.2.
ei  Qe i i  1, 2, 3 (1.10.23)
Such a tensor can be termed a change of basis tensor from e i  to ei  . The transpose
QT rotates the base vectors ei back to e i and is thus change of basis tensor from ei  to
e i  . The components of Q in the e i coordinate system are, from 1.9.4, Qij  e i Qe j and
so, from 1.10.23,
Q  Qij e i  e j , Qij  e i  e j , (1.10.24)
which are the direction cosines between the axes (see Fig. 1.5.5).
e3 e3
e2
e2 Q
e1
e1
Figure 1.10.2: Rotation of a set of base vectors

Section 1.10
The change of basis tensor can also be expressed in terms of the base vectors from both
bases:
Q  ei  e i , (1.10.25)
from which the above relations can easily be derived, for example ei  Qe i , QQ T  I ,
etc.
Consider now the operation of the change of basis tensor on a vector:
Qv  vi Qe i   vi ei (1.10.26)
Thus Q transforms v into a second vector v  , but this new vector has the same
components with respect to the basis ei , as v has with respect to the basis e i , vi  vi .
Note the difference between this and the coordinate transformations of §1.5: here there
are two different vectors, v and v  .
Example
Consider the two-dimensional rotation tensor
 0  1
Q  e i  e j   ei  e i
 1 0 
which corresponds to a rotation of the base vectors through  / 2 . The vector v  1 1

T
then transforms into (see Fig. 1.10.3)
  1  1
Qv   e i   ei
 1  1
e1 e2
Qv v
e2 e1
Figure 1.10.3: a rotated vector
Similarly, for a second order tensor A, the operation
QAQ T  QAij e i  e j Q T  Aij Qe i  e j Q T   Aij Qe i  Qe j   Aij ei  ej

(1.10.27)

Section 1.10
results in a new tensor which has the same components with respect to the ei , as A has
with respect to the e i , Aij  Aij .
1.10.10 Symmetric and Skew Tensors
A tensor T is said to be symmetric if it is identical to the transposed tensor, T  T T , and

skew (antisymmetric) if T  T T .
Any tensor A can be (uniquely) decomposed into a symmetric tensor S and a skew tensor
W, where
symA  S 
1
2
A  AT  
(1.10.28)
1
skewA  W  A  A T
2
 
and
S  ST , W  W T (1.10.29)
In matrix notation one has
 S11 S12 S13   0 W12 W13 

S   S12 S 22 S 23 , W    W12 0 W23 
 (1.10.30)
 S13 S 23 S 33   W13  W23 0 
Some useful properties of symmetric and skew tensors are {▲Problem 13}:
S : B  S : B T  S : 12 B  B T  
W : B   W : B  W : 12  B T
T
 
S:W 0
(1.10.31)
tr SW   0
v  Wv  0
det W  0 has no inverse
where v and B denote any arbitrary vector and second-order tensor respectively.
Note that symmetry and skew-symmetry are tensor properties, independent of coordinate
system.

Section 1.10
1.10.11 Axial Vectors

A skew tensor W has only three independent coefficients, so it behaves “like a vector”
with three components. Indeed, a skew tensor can always be written in the form
Wu  ω  u (1.10.32)
where u is any vector and ω characterises the axial (or dual) vector of the skew tensor
W. The components of W can be obtained from the components of ω through
Wij  e i  We j  e i  ω  e j   e i   k e k  e j 
 e i   k  kjp e p    kji  k (1.10.33)
  ijk  k
If one knows the components of W, one can find the components of ω by inverting this
equation, whence {▲Problem 14}
ω  W23e1  W13e 2  W12 e 3 (1.10.34)
Example (of an Axial Vector)
Decompose the tensor
1 2 3
 
T  Tij  4 2 1
1 1 1
into its symmetric and skew parts. Also find the axial vector for the skew part. Verify
that Wa  ω  a for a  e1  e 2 .
Solution
One has
1 2 3 1 4 1  1 3 2
1
 1 
 
S  T  T  4 2 1  2 2 1   3
2
T
2
2 1 

1 1 1 3 1 1  2 1 1 
1 2 3 1 4 1   0  1 1
1
2
 1 
 
W  T  T T  4 2 1  2 2 1    1
2
0 0

1 1 1 3 1 1   1 0 0
The axial vector is
ω  W23 e1  W13 e 2  W12 e 3  e 2  e 3

Section 1.10
and it can be seen that
Wa  Wij (e i  e j )(e1  e 3 )  Wij ( j1   j 3 )e i  (Wi1  Wi 3 )e i

 (W11  W13 )e1  (W21  W23 )e 2  (W31  W33 )e 3
 e1  e 2  e 3
and
e1 e2 e3
ωa  0 1 1  e1  e 2  e 3
1 0 1
■
The Spin Tensor
The velocity of a particle rotating in a rigid body motion is given by v  ω  x , where ω

is the angular velocity vector and x is the position vector relative to the origin on the axis
of rotation (see Problem 9, §1.1). If the velocity can be written in terms of a skew-
symmetric second order tensor w , such that wx  v , then it follows from wx  ω  x
that the angular velocity vector ω is the axial vector of w . In this context, w is called
the spin tensor.
1.10.12 Spherical and Deviatoric Tensors

Every tensor A can be decomposed into its so-called spherical part and its deviatoric
part, i.e.
A  sphA  devA (1.10.35)
where
sphA  1
3
trA I
 13  A11  A22  A33  0 0 
  0 
3 A11  A22  A33
1
 0 

 0 0 1
3
 A11  A22  A 
33 
devA  A  sphA
 A11  13  A11  A22  A33  A12 A13 

 A21 A22  13  A11  A22  A33  A23 

 A31 A32 A33  13  A11  A22  A33 
(1.10.36)

Section 1.10
Any tensor of the form I is known as a spherical tensor, while devA is known as a
deviator of A, or a deviatoric tensor.
Some important properties of the spherical and deviatoric tensors are
tr (devA )  0
sph (devA )  0 (1.10.37)
devA : sphB  0
1.10.13 Positive Definite Tensors

A positive definite tensor A is one which satisfies the relation
vAv  0 , v  o (1.10.38)
The tensor is called positive semi-definite if vAv  0 .
In component form,
vi Aij v j  A11v12  A12 v1v 2  A13 v1v3  A21v 2 v1  A22 v 22   (1.10.39)
and so the diagonal elements of the matrix representation of a positive definite tensor
must always be positive.
It can be shown that the following conditions are necessary for a tensor A to be positive
definite (although they are not sufficient):
(i) the diagonal elements of A  are positive

(ii) the largest element of A  lies along the diagonal
(iii) det A  0
(iv) Aii  A jj  2 Aij for i  j (no sum over i, j )
These conditions are seen to hold for the following matrix representation of an example
positive definite tensor:
 2 2 0
A    1 4 0
 0 0 1
A necessary and sufficient condition for a tensor to be positive definite is given in the
next section, during the discussion of the eigenvalue problem.
One of the key properties of a positive definite tensor is that, since det A  0 , positive
definite tensors are always invertible.
An alternative definition of positive definiteness is the equivalent expression

Section 1.10
A:vv 0 (1.10.40)
1.10.14 Problems
1. Show that the components of the (second-order) identity tensor are given by I ij   ij .
2. Show that
(a) (u  v) A  u  ( A T v)
(b) A : BC   B T A  : C  AC T  : B
3. Use (1.10.4) to show that I T  I .
4. Show that (1.10.6) implies (1.10.5) for the trace of a tensor.
5. Show that tr u  v   u  v .
6. Formally derive the index notation for the functions
trA 2 , trA 3 , ( trA) 2 , ( trA) 3
7. Show that A : B  tr ( A T B) .
8. Prove (1.10.16f), Ta  Tb   Tc  det T a  b   c .
9. Show that ( A 1 ) T : A  3 . [Hint: one way of doing this is using the result from
Problem 7.]
10. Use 1.10.16b and 1.10.18d to prove 1.10.22c, det Q  1 .
11. Use the explicit dyadic representation of the rotation tensor, Q  ei  e i , to show that
the components of Q in the “second”, ox1 x 2 x3 , coordinate system are the same as
those in the first system [hint: use the rule Qij  ei  Qej ]
12. Consider the tensor D with components (in a certain coordinate system)
1 / 2 1 / 2  1 / 2
 
 0 1/ 2 1/ 2 
1 / 2  1 / 2 1 / 2 
 
Show that D is a rotation tensor (just show that D is proper orthogonal).
13. Show that tr SW   0 .
14. Multiply across (1.10.32), Wij   ijk  k , by  ijp to show that ω   12  ijk Wij e k . [Hint:
use the relation 1.3.19b,  ijp  ijk  2 pk .]
15. Show that 12 a  b  b  a  is a skew tensor W. Show that its axial vector is
ω  12 b  a  . [Hint: first prove that b  u a  a  u b  u  a  b   b  a   u .]
16. Find the spherical and deviatoric parts of the tensor A for which Aij  1 .

2 Kinematics
Kinematics is concerned with expressing in mathematical form the deformation and motion of
materials. In what follows, a number of important quantities, mainly vectors and second-
order tensors, are introduced. Each of these quantities, for example the velocity, deformation
gradient or rate of deformation tensor, allows one to describe a particular aspect of a
deforming material.
No consideration is given to what is causing the deformation and movement – the cause is the
action of forces on the material, and these will be discussed in the next chapter.
The first section introduces the material and spatial coordinates and descriptions. The second
and third sections discuss the strain tensors. The fourth, fifth and sixth sections deal with
rates of deformation and rates of change of kinematic quantities. The theory is specialised to
small strain deformations in section 7. The notion of objectivity and the related topic of rigid
rotations are discussed in sections 8 and 9. The final sections, 10-13, deal with kinematics
using the convected coordinate system, and include the important notions of push-forward,
pull-back and the Lie time derivative.
199
200
Section 2.1
2.1 Motion
2.1.1 The Material Body and Motion

Physical materials in the real world are modeled using an abstract mathematical entity
called a body. This body consists of an infinite number of material particles1. Shown in
Fig. 2.1.1a is a body B with material particle P. One distinguishes between this body and
the space in which it resides and through which it travels. Shown in Fig. 2.1.1b is a
certain point x in Euclidean point space E.

χ
 
P x
B
E
(a) (b) (c)
Figure 2.1.1: (a) a material particle in a body, (b) a place in space, (c) a configuration
of the body
By fixing the material particles of the body to points in space, one has a configuration of
the body χ , Fig. 2.1.1c. A configuration can be expressed as a mapping of the particles
P to the point x ,
x  χ  P (2.1.1)
A motion of the body is a family of configurations parameterised by time t,
x  χ  P, t  (2.1.2)
At any time t, Eqn. 2.1.2 gives the location in space x of the material particle P , Fig.
2.1.2.
1
these particles are not the discrete mass particles of Newtonian mechanics, rather they are very small
portions of continuous matter; the meaning of particle is made precise in the definitions which follow

Section 2.1
t2
t1

P 
P
Figure 2.1.2: a motion of material
The Reference and Current Configurations
Choose now some reference configuration, Fig. 2.1.3. The motion can then be
measured relative to this configuration. The reference configuration might be the
configuration occupied by the material at time t  0 , in which case it is often called the
initial configuration. For a solid, it might be natural to choose a configuration for which
the material is stress-free, in which case it is often called the undeformed configuration.
However, the choice of reference configuration is completely arbitrary.
Introduce a Cartesian coordinate system with base vectors E i for the reference
configuration. A material particle P in the reference configuration can then be assigned a
unique position vector X  X i E i relative to the origin of the axes. The coordinates
 X 1 , X 2 , X 3  of the particle are called material coordinates (or Lagrangian coordinates
or referential coordinates).
Some time later, say at time t, the material occupies a different configuration, which will
be called the current configuration (or deformed configuration). Introduce a second
Cartesian coordinate system with base vectors e i for the current configuration, Fig. 2.1.3.
In the current configuration, the same particle P now occupies the location x , which can
now also be assigned a position vector x  xi e i . The coordinates  x1 , x 2 , x3  are called
spatial coordinates (or Eulerian coordinates).
Each particle thus has two sets of coordinates associated with it. The particle’s material
coordinates stay with it throughout its motion. The particle’s spatial coordinates change
as it moves.

Section 2.1
reference
configuration current
χ configuration t
 
P P
X2 X x2
x
E2 e2
X1 e1
X3 E1 x3
Figure 2.1.3: reference and current configurations
In practice, the material and spatial axes are usually taken to be coincident so that the base
vectors E i and e i are the same, as in Fig. 2.1.4. Nevertheless, the use of different base
vectors E and e for the reference and current configurations is useful even when the
material and spatial axes are coincident, since it helps distinguish between quantities
associated with the reference configuration and those associated with the spatial
configuration (see later).
X 2 , x2 X
x
E2 , e2
X 1 , x1
X 3 , x 3 E 1 , e1
Figure 2.1.4: reference and current configurations with coincident axes
In terms of the position vectors, the motion 2.1.2 can be expressed as a relationship
between the material and spatial coordinates,
x  χ ( X, t ), xi   i  X 1 , X 2 , X 3 , t  Material description (2.1.3)
or the inverse relation
X  χ 1 (x, t ), X i   i1  x1 , x 2 , x3 , t  Spatial description (2.1.4)
If one knows the material coordinates of a particle then its position in the current
configuration can be determined from 2.1.3. Alternatively, if one focuses on some
location in space, in the current configuration, then the material particle occupying that
position can be determined from 2.1.4. This is illustrated in the following example.

Section 2.1
Example (Extension of a Bar)
Consider the motion
x1  3 X 1t  X 1  t , x 2  X 2 , x3  X 3 (2.1.5)
These equations are of the form 2.1.3 and say that “the particle that was originally at
position X is now, at time t, at position x”. They represent a simple translation and
uniaxial extension of material as shown in Fig. 2.1.5. Note that X  x at t  0 .
x2 χ
X2  
X x
X1 x1
configuration at configurations at
t0 t0
Figure 2.1.5: translation and extension of material
Relations of the form 2.1.4 can be obtained by inverting 2.1.5:
x1  t
X1  , X 2  x2 , X 3  x3
1  3t
These equations say that “the particle that is now, at time t, at position x was originally at
position X”.
■
Convected Coordinates
The material and spatial coordinate systems used here are fixed Cartesian systems. An
alternative method of describing a motion is to attach the material coordinate system to
the material and let it deform with the material. The motion is then described by defining
how this coordinate system changes. This is the convected coordinate system. In
general, the axes of a convected system will not remain mutually orthogonal and a
curvilinear system is required. Convected coordinates will be examined in §2.10.
2.1.2 The Material and Spatial Descriptions

Any physical property (such as density, temperature, etc.) or kinematic property (such as
displacement or velocity) of a body can be described in terms of either the material
coordinates X or the spatial coordinates x, since they can be transformed into each other
using 2.1.3-4. A material (or Lagrangian) description of events is one where the

Section 2.1
material coordinates are the independent variables. A spatial (or Eulerian) description of
events is one where the spatial coordinates are used.
Example (Temperature of a Body)
Suppose the temperature  of a body is, in material coordinates,
 ( X, t )  3 X 1  X 3 (2.1.6)
but, in the spatial description,
x1
 ( x, t )   1  x3 . (2.1.7)
t
According to the material description 2.1.6, the temperature is different for different
particles, but the temperature of each particle remains constant over time. The spatial
description 2.1.7 describes the time-dependent temperature at a specific location in space,
x, Fig. 2.1.6. Different material particles are flowing through this location over time.
x
motion of individual
material particles
Figure 2.1.6: particles flowing through space
In the material description, then, attention is focused on specific material. The piece of
matter under consideration may change shape, density, velocity, and so on, but it is
always the same piece of material. On the other hand, in the spatial description, attention
is focused on a fixed location in space. Material may pass through this location during
the motion, so different material is under consideration at different times.
The spatial description is the one most often used in Fluid Mechanics since there is no
natural reference configuration of the material as it is continuously moving. However,
both the material and spatial descriptions are used in Solid Mechanics, where the
reference configuration is usually the stress-free configuration.
2.1.3 Small Perturbations

A large number of important problems involve materials which deform only by a
relatively small amount. An example would be the steel structural columns in a building
under modest loading. In this type of problem there is virtually no distinction to be made

Section 2.1
between the two viewpoints taken above and the analysis is simplified greatly (see later,
on Small Strain Theory, §2.7).
2.1.4 Problems
1. The density of a material is given by   3 X 1  X 2 and the motion is given by the

equations X 1  x1 , X 2  x 2  t , X 3  x3  t .
(a) what kind of description is this for the density, and what kind of description is
this for the motion?
(b) re-write the density in terms of x – what is the name given to this description of
the density?
(c) is the density of any given material particle changing with time?
(d) invert the motion equations so that X is the independent variable – what is the
name given to this description of the motion?
(e) draw the line element joining the origin to (1,1,0) and sketch the position of this
element of material at times t  1 and t  2 .

Section 2.2
2.2 Deformation and Strain

A number of useful ways of describing and quantifying the deformation of a material are
discussed in this section.
Attention is restricted to the reference and current configurations. No consideration is

given to the particular sequence by which the current configuration is reached from the
reference configuration and so the deformation can be considered to be independent of
time. In what follows, particles in the reference configuration will often be termed
“undeformed” and those in the current configuration “deformed”.
In a change from Chapter 1, lower case letters will now be reserved for both vector- and
tensor- functions of the spatial coordinates x, whereas upper-case letters will be reserved
for functions of material coordinates X. There will be exceptions to this, but it should be
clear from the context what is implied.
2.2.1 The Deformation Gradient

The deformation gradient F is the fundamental measure of deformation in continuum
mechanics. It is the second order tensor which maps line elements in the reference
configuration into line elements (consisting of the same material particles) in the current
configuration.
Consider a line element dX emanating from position X in the reference configuration

which becomes dx in the current configuration, Fig. 2.2.1. Then, using 2.1.3,
dx  χ X  dX   χ X 
(2.2.1)
 Grad χ dX
A capital G is used on “Grad” to emphasise that this is a gradient with respect to the
material coordinates1, the material gradient, χ / X .
dX
dx
X x
Figure 2.2.1: the Deformation Gradient acting on a line element
1
one can have material gradients and spatial gradients of material or spatial fields – see later

Section 2.2
The motion vector-function χ in 2.1.3, 2.2.1, is a function of the variable X, but it is

customary to denote this simply by x, the value of χ at X, i.e. x  xX, t  , so that
x xi
F  Grad x, FiJ  Deformation Gradient (2.2.2)
X X J
with
dx  F dX, dxi  FiJ dX J action of F (2.2.3)
Lower case indices are used in the index notation to denote quantities associated with the
spatial basis e i  whereas upper case indices are used for quantities associated with the
material basis E I  .
Note that
x
dx  dX
X
is a differential quantity and this expression has some error associated with it; the error
(due to terms of order (dX) 2 and higher, neglected from a Taylor series) tends to zero as
the differential dX  0 . The deformation gradient (whose components are finite) thus
characterises the deformation in the neighbourhood of a point X, mapping infinitesimal
line elements dX emanating from X in the reference configuration to the infinitesimal
line elements dx emanating from x in the current configuration, Fig. 2.2.2.
before after
Figure 2.2.2: deformation of a material particle
Example
Consider the cube of material with sides of unit length illustrated by dotted lines in Fig.
2.2.3. It is deformed into the rectangular prism illustrated (this could be achieved, for
example, by a continuous rotation and stretching motion). The material and spatial
coordinate axes are coincident. The material description of the deformation is
1 1
x  χ ( X )  6 X 2e1  X 1e 2  X 3e 3
2 3
and the spatial description is

Section 2.2
1
X  χ 1 ( x )  2 x2E1  x1E2  3x3E3
6
X 2 , x2
D
D E
B B
A X 1 , x1
E
C
C
X 3 , x3
Figure 2.2.3: a deforming cube
Then
 0 6 0 
xi
F  1 / 2 0 0 
X J
 0 0 1 / 3
Once F is known, the position of any element can be determined. For example, taking a
line element dX  [da, 0, 0]T , dx  FdX  [0, da / 2,0]T .
■
Homogeneous Deformations
A homogeneous deformation is one where the deformation gradient is uniform, i.e.

independent of the coordinates, and the associated motion is termed affine. Every part of
the material deforms as the whole does, and straight parallel lines in the reference
configuration map to straight parallel lines in the current configuration, as in the above
example. Most examples to be considered in what follows will be of homogeneous
deformations; this keeps the algebra to a minimum, but homogeneous deformation
analysis is very useful in itself since most of the basic experimental testing of materials,
e.g. the uniaxial tensile test, involve homogeneous deformations.
Rigid Body Rotations and Translations
One can add a constant vector c to the motion, x  x  c , without changing the
deformation, Gradx  c   Gradx . Thus the deformation gradient does not take into
account rigid-body translations of bodies in space. If a body only translates as a rigid
body in space, then F  I , and x  X  c (again, note that F does not tell us where in
space a particle is, only how it has deformed locally). If there is no motion, then not only
is F  I , but x  X .

Section 2.2
If the body rotates as a rigid body (with no translation), then F  R , a rotation tensor
(§1.10.8). For example, for a rotation of  about the X 2 axis,
 sin  0 cos  
F 0 1 0 
 
 cos  0  sin  
Note that different particles of the same material body can be translating only, rotating
only, deforming only, or any combination of these.
The Inverse of the Deformation Gradient
The inverse deformation gradient F 1 carries the spatial line element dx to the material
line element dX. It is defined as
X X I
F 1   grad X, FIj1  Inverse Deformation Gradient (2.2.4)
x x j
so that
d X  F 1 d x , dX I  FIj1 dx j action of F 1 (2.2.5)
with (see Eqn. 1.15.2)
F 1F  FF 1  I FiM FM1j   ij (2.2.6)
Cartesian Base Vectors
Explicitly, in terms of the material and spatial base vectors (see 1.14.3),
x xi
F  EJ  ei  E J
X J X J
(2.2.7)
1 X X I
F  ej  EI  e j
x j x j
so that, for example, FdX  xi / X J e i  E J dX M E M   xi / X J dX J e i  dx .
Because F and F 1 act on vectors in one configuration to produce vectors in the other
configuration, they are termed two-point tensors. They are defined in both
configurations. This is highlighted by their having both reference and current base
vectors E and e in their Cartesian representation 2.2.7.

Section 2.2
Here follow some important relations which relate scalar-, vector- and second-order
tensor-valued functions in the material and spatial descriptions, the first two relating the
material and spatial gradients {▲Problem 1}.
grad  Grad F 1
gradv  Grad V F 1 (2.2.8)
T
diva  Grad A : F
Here,  is a scalar; V and v are the same vector, the former being a function of the
material coordinates, the material description, the latter a function of the spatial
coordinates, the spatial description. Similarly, A is a second order tensor in the material
form and a is the equivalent spatial form.
The first two of 2.2.8 relate the material gradient to the spatial gradient: the gradient of a
function is a measure of how the function changes as one moves through space; since the
material coordinates and the spatial coordinates differ, the change in a function with
respect to a unit change in the material coordinates will differ from the change in the same
function with respect to a unit change in the spatial coordinates (see also §2.2.7 below).
Example
Consider the deformation
x  2 X 2  X 3 e1   X 2 e 2   X 1  3 X 2  X 3 e 3
X   x1  5 x 2  x3 E1   x 2 E 2   x1  2 x 2 E 3
so that
0 2  1 1 5 1
F  0  1 0 , F 1 
  0  1 0
1 3 1   1  2 0
 
Consider the vector v (x)  2 x1  x 2 e1   3x 22  x3 e 2  x1  x3 e 3 which, in the
material description, is
 
V ( X)  5 X 2  2 X 3 E1  X 1  3 X 2  X 3  3 X 22 E 2   X 1  5 X 2 E 3
The material and spatial gradients are
0 5  2 2 1 0

GradV  1 3  6 X 2 1 , 
gradv  0  6 x 2 1
1 5 0  1 0 1
and it can be seen that

Section 2.2
2  1 0  2 1 0
GradV F 1
 0 6 X 2  
1   0  6 x 2 1  grad v
1 0 1 1 0 1
2.2.2 The Cauchy-Green Strain Tensors

The deformation gradient describes how a line element in the reference configuration
maps into a line element in the current configuration. It has been seen that the
deformation gradient gives information about deformation (change of shape) and rigid
body rotation, but does not encompass information about possible rigid body translations.
The deformation and rigid rotation will be separated shortly (see §2.2.5). To this end,
consider the following strain tensors; these tensors give direct information about the
deformation of the body. Specifically, the Left Cauchy-Green Strain and Right
Cauchy-Green Strain tensors give a measure of how the lengths of line elements and
angles between line elements (through the vector dot product) change between
configurations.
The Right Cauchy-Green Strain
Consider two line elements in the reference configuration dX (1) , dX ( 2 ) which are mapped
into the line elements dx (1) , dx ( 2) in the current configuration. Then, using 1.10.3d,
 
dx (1)  dx ( 2 )  FdX (1)  FdX ( 2 ) 
 
 dX (1) F T F dX ( 2 ) action of C (2.2.9)
 dX (1) CdX ( 2 )
where, by definition, C is the right Cauchy-Green Strain2
x k x k
C  F T F, C IJ  Fk I Fk J  Right Cauchy-Green Strain (2.2.10)
X I X J
It is a symmetric, positive definite (which will be clear from Eqn. 2.2.17 below), tensor,
which implies that it has real positive eigenvalues (cf. §1.11.2), and this has important
consequences (see later). Explicitly in terms of the base vectors,
 x  x  x k x k
C   k E I  e k  m e m  E J   EI  EJ . (2.2.11)
 X i  X J  X I X J
Just as the line element dX is a vector defined in and associated with the reference
configuration, C is defined in and associated with the reference configuration, acting on
vectors in the reference configuration, and so is called a material tensor.
2
“right” because F is on the right of the formula

Section 2.2
The inverse of C, C-1, is called the Piola deformation tensor.
The Left Cauchy-Green Strain
Consider now the following, using Eqn. 1.10.18c:
 
dX (1)  dX ( 2 )  F 1dx (1)  F 1dx ( 2 ) 
 
 dx (1) F T F 1 dx ( 2 ) action of b 1 (2.2.12)
 dx (1)b 1dx ( 2 )
where, by definition, b is the left Cauchy-Green Strain, also known as the Finger tensor:
xi x j
b  FF T , bij  FiK F jK  Left Cauchy-Green Strain (2.2.13)
X K X K
Again, this is a symmetric, positive definite tensor, only here, b is defined in the current
configuration and so is called a spatial tensor.
The inverse of b, b-1, is called the Cauchy deformation tensor.
It can be seen that the right and left Cauchy-Green tensors are related through
C  F -1bF, b  FCF -1 (2.2.14)
Note that tensors can be material (e.g. C), two-point (e.g. F) or spatial (e.g. b). Whatever
type they are, they can always be described using material or spatial coordinates through
the motion mapping 2.1.3, that is, using the material or spatial descriptions. Thus one
distinguishes between, for example, a spatial tensor, which is an intrinsic property of a
tensor, and the spatial description of a tensor.
The Principal Scalar Invariants of the Cauchy-Green Tensors
Using 1.10.10b,
tr C  tr F T F   tr FF T   tr b (2.2.15)
This holds also for arbitrary powers of these tensors, tr C n  tr b n , and therefore, from
Eqn. 1.11.17, the invariants of C and b are equal.
2.2.3 The Stretch
The stretch (or the stretch ratio)  is defined as the ratio of the length of a deformed
line element to the length of the corresponding undeformed line element:

Section 2.2
dx
 The Stretch (2.2.16)
dX
From the relations involving the Cauchy-Green Strains, letting dX (1)  dX ( 2)  dX ,

dx (1)  dx ( 2 )  dx , and dividing across by the square of the length of dX or dx ,
2 2
 dx   dX 
  
2
  dX

ˆ CdX
ˆ,  2
    dxˆ b 1 dxˆ
 (2.2.17)
 dX   dx 
ˆ  dX / dX and dxˆ  dx / dx are unit vectors in the directions of

Here, the quantities dX
dX and dx . Thus, through these relations, C and b determine how much a line element
stretches (and, from 2.2.17, C and b can be seen to be indeed positive definite).
One says that a line element is extended, unstretched or compressed according to   1 ,

  1 or   1 .
Stretching along the Coordinate Axes
Consider three line elements lying along the three coordinate axes3. Suppose that the
material deforms in a special way, such that these line elements undergo a pure stretch,
that is, they change length with no change in the right angles between them. If the
stretches in these directions are 1 , 2 and 3 , then
x1  1 X 1 , x2  2 X 2 , x3  3 X 3 (2.2.18)
and the deformation gradient has only diagonal elements in its matrix form:
1 0 0
F   0 2 0 , FiJ  i  iJ (no sum) (2.2.19)
 0 0 3 
Whereas material undergoes pure stretch along the coordinate directions, line elements
off-axes will in general stretch/contract and rotate relative to each other. For example, a
line element dX  [ ,  ,0]T stretches by   dXC ˆ dX ˆ    2   2  / 2 with
1 2
dx  [1 ,  2 ,0] , and rotates if 1   2 .

T
It will be shown below that, for any deformation, there are always three mutually
orthogonal directions along which material undergoes a pure stretch. These directions,
the coordinate axes in this example, are called the principal axes of the material and the
associated stretches are called the principal stretches.
3
with the material and spatial basis vectors coincident

Section 2.2
The Case of F Real and Symmetric
Consider now another special deformation, where F is a real symmetric tensor, in which
case the eigenvalues are real and the eigenvectors form an orthonormal basis (cf.
§1.11.2)4. In any given coordinate system, F will in general result in the stretching of line
elements and the changing of the angles between line elements. However, if one chooses
a coordinate set to be the eigenvectors of F, then from Eqn. 1.11.11-12 one can write5
3
1 0 0
F   i nˆ i  N
ˆ ,
i F    0 2 0  (2.2.20)
i 1
 0 0 3 
where 1 ,  2 , 3 are the eigenvalues of F. The eigenvalues are the principal stretches and
the eigenvectors are the principal axes. This indicates that as long as F is real and
symmetric, one can always find a coordinate system along whose axes the material
undergoes a pure stretch, with no rotation. This topic will be discussed more fully in
§2.2.5 below.
2.2.4 The Green-Lagrange and Euler-Almansi Strain Tensors

Whereas the left and right Cauchy-Green tensors give information about the change in
angle between line elements and the stretch of line elements, the Green-Lagrange strain
and the Euler-Almansi strain tensors directly give information about the change in the
squared length of elements.
Specifically, when the Green-Lagrange strain E operates on a line element dX, it gives
(half) the change in the squares of the undeformed and deformed lengths:
2 2
dx  dX 1
 dXCdX  dX  dX
2 2
1
 dXC  I dX action of E (2.2.21)
2
 dXEdX
where
E
1
C  I   1 F T F  I , E I J  1 C I J   I J  Green-Lagrange Strain (2.2.22)
2 2 2
It is a symmetric positive definite material tensor. Similarly, the (symmetric spatial)

Euler-Almansi strain tensor is defined through
4
in fact, F in this case will have to be positive definite, with det F  0 (see later in §2.2.8)
5
n̂ i are the eigenvectors for the basis e i , N̂ I for the basis Ê i , with n̂ i , N̂ I coincident; when the bases are
not coincident, the notion of rotating line elements becomes ambiguous – this topic will be examined later
in the context of objectivity

Section 2.2
2 2
dx  dX
 dx e dx action of e (2.2.23)
2
and
e
1
2
I  b 1   I  F T F 1 
1
2
Euler-Almansi Strain (2.2.24)
Physical Meaning of the Components of E
Take a line element in the 1-direction, dX (1)  dX 1 , 0, 0 , so that dX

ˆ  1, 0, 0 . The
T T
(1)
square of the stretch of this element is
ˆ CdX
2(1)  dX (1)
ˆ C
(1) 11  E11 
1
C11  1  1 2(1)  1
2 2
The unit extension is  dx  dX  / dX    1 . Denoting the unit extension of dX (1) by

E (1) , one has
1
E11  E (1)  E (21) (2.2.25)
2
and similarly for the other diagonal elements E 22 , E33 .
When the deformation is small, E (21) is small in comparison to E (1) , so that E11  E(1) . For
small deformations then, the diagonal terms are equivalent to the unit extensions.
Let  12 denote the angle between the deformed elements which were initially parallel to
the X 1 and X 2 axes. Then
dx (1) dx ( 2) dX (1) dX ( 2)  dX (1) dX ( 2 )  C12

cos 12     C 
dx (1) dx ( 2) dx (1) dx ( 2)  dX (1) dX ( 2 )  (1) ( 2 )
(2.2.26)
2 E12

2 E11  1 2 E 22  1
and similarly for the other off-diagonal elements. Note that if  12   / 2 , so that there is
no angle change, then E12  0 . Again, if the deformation is small, then E11 , E 22 are
small, and
  
  12  sin    12   cos 12  2 E12 (2.2.27)
2 2 

Section 2.2
In words: for small deformations, the component E12 gives half the change in the original
right angle.
2.2.5 Stretch and Rotation Tensors

The deformation gradient can always be decomposed into the product of two tensors, a
stretch tensor and a rotation tensor (in one of two different ways, material or spatial
versions). This is known as the polar decomposition, and is discussed in §1.11.7. One
has
F  RU Polar Decomposition (Material) (2.2.28)
Here, R is a proper orthogonal tensor, i.e. R T R  I with det R  1 , called the rotation
tensor. It is a measure of the local rotation at X.
The decomposition is not unique; it is made unique by choosing U to be a symmetric

tensor, called the right stretch tensor. It is a measure of the local stretching (or
contraction) of material at X. Consider a line element dX. Then
ˆ  RUdX
dxˆ  FdX ˆ (2.2.29)
and so {▲Problem 2}
ˆ U  UdX
 2  dX ˆ (2.2.30)
Thus (this is a definition of U)
U C C  UU  The Right Stretch Tensor (2.2.31)
From 2.2.30, the right Cauchy-Green strain C (and by consequence the Euler-Lagrange
strain E) only give information about the stretch of line elements; it does not give
information about the rotation that is experienced by a particle during motion. The
deformation gradient F, however, contains information about both the stretch and rotation.
It can also be seen from 2.2.30-1 that U is a material tensor.
Note that, since
dx  R UdX  ,
the undeformed line element is first stretched by U and is then rotated by R into the
deformed element dx (the element may also undergo a rigid body translation c), Fig.
2.2.4. R is a two-point tensor.

Section 2.2
principal
material final
axes configuration
R
stretched
undeformed
Figure 2.2.4: the polar decomposition
Evaluation of U
In order to evaluate U, it is necessary to evaluate C . To evaluate the square-root, C

must first be obtained in relation to its principal axes, so that it is diagonal, and then the
square root can be taken of the diagonal elements, since its eigenvalues will be positive
(see §1.11.6). Then the tensor needs to be transformed back to the original coordinate
system.
Example
Consider the motion
x1  2 X 1  2 X 2 , x 2  X 1  X 2 , x3  X 3
The (homogeneous) deformation of a unit square in the x1  x 2 plane is as shown in Fig.

2.2.5.
X 2 , x3
X 1 , x1
Figure 2.2.5: deformation of a square
One has
 2  2 0  5  3 0
F  1 1 0 basis : e i  E j ,  
C  F F   3 5 0 basis : E i  E j 
T
0 0 1  0 0 1

Section 2.2
Note that F is not symmetric, so that it might have only one real eigenvalue (in fact here it
does have complex eigenvalues), and the eigenvectors may not be orthonormal. C, on the
other hand, by its very definition, is symmetric; it is in fact positive definite and so has
positive real eigenvalues forming an orthonormal set.
To determine the principal axes of C, it is necessary to evaluate the

eigenvalues/eigenvectors of the tensor. The eigenvalues are the roots of the characteristic
equation 1.11.5,
 3  I C 2  II C  IIIC  0
and the first, second and third invariants of the tensor are given by 1.11.6 so that
 3  11 2  26  16  0 , with roots   8, 2, 1 . The three corresponding eigenvectors
are found from 1.11.8,
(C11   ) Nˆ 1  C12 Nˆ 2  C13 Nˆ 3  0 (5   ) Nˆ 1  3 Nˆ 2  0

C 21 Nˆ 1  (C 22   ) Nˆ 2  C 23 Nˆ 3  0   3 Nˆ 1  (5   ) Nˆ 2  0
C 31 Nˆ 1  C 32 Nˆ 2  (C 33   ) Nˆ 3  0 (1   ) Nˆ 3  0
Thus (normalizing the eigenvectors so that they are unit vectors, and form a right-handed
set, Fig. 2.2.6):
(i) for   8 ,  3 Nˆ 1  3 Nˆ 2  0,  3 Nˆ 1  3 Nˆ 2  0,  7 Nˆ 3  0 , ˆ 
N 1
E1  1
E2
1 2 2
(ii) for   2 , 3 Nˆ 1  3 Nˆ 2  0,  3 Nˆ 1  3 Nˆ 2  0,  Nˆ 3  0 , ˆ 
N 1
E1  1
E2
2 2 2
(iii) for   1 , 4 Nˆ 1  3 Nˆ 2  0,  3 Nˆ 1  4 Nˆ 2  0, 0 Nˆ 3  0 , ˆ E
N 3 3
X2 principal
material
N̂ 2 directions
X1
N̂1
Figure 2.2.6: deformation of a square
Thus the right Cauchy-Green strain tensor C, with respect to coordinates with base
vectors E1  N̂ 1 , E2  N̂ 2 and E3  N̂ 3 , that is, in terms of principal coordinates, is
8 0 0 
C  0 2 0 ˆ N
basis : N i
ˆ
j
0 0 1

Section 2.2
This result can be checked using the tensor transformation formulae 1.13.6,
C  Q T CQ, where Q is the transformation matrix of direction cosines (see also the
example at the end of §1.5.2),
 e1  e1 e1  e2 e1  e3        1 / 2 1 / 2 0

Qij  e 2  e1 e 2  e2 e 2  e3   N
ˆ ˆ
N ˆ     1 / 2 1 / 2 0 .
N
1 2 3  
e 3  e1 e 3  e2 e 3  e3        0 0 1
The stretch tensor U, with respect to the principal directions is
2 2 0 0 1 0 0
U   
C  0
 0
2 0   0 2 ˆ N
0  basis : N ˆ
0 1  0 
i j
 
0 3 
These eigenvalues of U (which are the square root of those of C) are the principal
stretches and, as before, they are labeled 1 ,  2 , 3 .
In the original coordinate system, using the inverse tensor transformation rule 1.13.6,
U   QU QT ,
 3/ 2  1 / 2 0
U   1 / 2 3 / 2 0 basis : E i  E j
0 0 1
 
so that
1 / 2  1 / 2 0
R   FU 1
  1 / 2
 0
1 / 2 0 basis : e i  E j
0 1
 
and it can be verified that R is a rotation tensor, i.e. is proper orthogonal.
Returning to the deformation of the unit square, the stretch and rotation are as illustrated
in Fig. 2.2.7 – the action of U is indicated by the arrows, deforming the unit square to the
dotted parallelogram, whereas R rotates the parallelogram through 45 o as a rigid body to
its final position.
Note that the line elements along the diagonals (indicated by the heavy lines) lie along the
principal directions of U and therefore undergo a pure stretch; the diagonal in the N̂1
direction has stretched but has also moved with a rigid translation.

Section 2.2
X 2 , x2
X 1 , x1
Figure 2.2.7: stretch and rotation of a square

■
Spatial Description
A polar decomposition can be made in the spatial description. In that case,
F  vR Polar Decomposition (Spatial) (2.2.32)
Here v is a symmetric, positive definite second order tensor called the left stretch tensor,
and vv  b , where b is the left Cauchy-Green tensor. R is the same rotation tensor as
appears in the material description. Thus an elemental sphere can be regarded as first
stretching into an ellipsoid, whose axes are the principal material axes (the principal axes
of U), and then rotating; or first rotating, and then stretching into an ellipsoid whose axes
are the principal spatial axes (the principal axes of v). The end result is the same.
The development in the spatial description is similar to that given above for the material
description, and one finds by analogy with 2.2.30,
2  dxˆ v 1  v 1 dxˆ (2.2.33)
In the above example, it turns out that v takes the simple diagonal form
2 2 0 0
v   0 2 0 basis : e i  e j .
0 0 1
 
so the unit square rotates first and then undergoes a pure stretch along the coordinate axes,
which are the principal spatial axes, and the sequence is now as shown in Fig. 2.2.9.

Section 2.2
X 2 , x2
X 1 , x1
Figure 2.2.8: stretch and rotation of a square in spatial description
Relationship between the Material and Spatial Decompositions
Comparing the two decompositions, one sees that the material and spatial tensors
involved are related through
v  RUR T , b  RCR T (2.2.34)
Further, suppose that U has an eigenvalue  and an eigenvector N̂ . Then UN ˆ  N

ˆ , so
that RUN  RN . But RU  vR , so v RN    
ˆ   RN ˆ . Thus v also has an eigenvalue
 , but an eigenvector nˆ  RNˆ . From this, it is seen that the rotation tensor R maps the
principal material axes into the principal spatial axes. It also follows that R and F can be
written explicitly in terms of the material and spatial principal axes (compare the first of
these with 1.10.25)6:
3 3
ˆ ,
R  nˆ i  N i F  RU  R  i N
ˆ N
i
ˆ    nˆ  N
i i i
ˆ
i (2.2.35)
i 1 i 1
and the deformation gradient acts on the principal axes base vectors according to
{▲Problem 4}
ˆ   nˆ , ˆ  1 1 ˆ ˆ
FN F T N nˆ i , F 1nˆ i  Ni , F T nˆ i  i N (2.2.36)
i i i i
i i i
The representation of F and R in terms of both material and spatial principal base vectors
in 2.3.35 highlights their two-point character.
Other Strain Measures
Some other useful measures of strain are
The Hencky strain measure: H  ln U (material) or h  ln v (spatial)
6
this is not a spectral decomposition of F (unless F happens to be symmetric, which it must be in order to
have a spectral decomposition)

Section 2.2
The Biot strain measure: B  U  I (material) or b  v  I (spatial)
The Hencky strain is evaluated by first evaluating U along the principal axes, so that the
logarithm can be taken of the diagonal elements.
The material tensors H, B , C, U and E are coaxial tensors, with the same eigenvectors
N̂ i . Similarly, the spatial tensors h, b , b, v and e are coaxial with the same eigenvectors
n̂ i . From the definitions, the spectral decompositions of these tensors are
3 3
U   i N
ˆ N
i
ˆ
i v   i nˆ i  nˆ i
i 1 i 1
3 3
C   i2 N
ˆ N
i
ˆ
i b   i2 nˆ i  nˆ i
i 1 i 1
   
3 3
E   12   1 N
ˆ N
i
ˆ
i
2
i e   12 1  1 / i2 nˆ i  nˆ i (2.2.37)
i 1 i 1
3 3
H   ln i N
ˆ N
i
ˆ
i h   ln i nˆ i  nˆ i
i 1 i 1
3 3
B   i  1N
ˆ N
i
ˆ
i b   i  1nˆ i  nˆ i
i 1 i 1
Deformation of a Circular Material Element
A circular material element will deform into an ellipse, as indicated in Figs. 2.2.2 and
2.2.4. This can be shown as follows. With respect to the principal axes, an undeformed
line element dX  dX 1N1  dX 2 N 2 has magnitude squared  dX 1    dX 2   c 2 , where c
2 2
is the radius of the circle, Fig. 2.2.9. The deformed element is dx  UdX , or
dx  1dX 1N1  2dX 2 N 2  dx1n1  dx2 n 2 . Thus dx1 / 1  dX 1 , dx2 / 2  dX 2 , which
leads to the standard equation of an ellipse with major and minor axes 1c, 2 c :
 dx1 / 1c    dx2 / 2 c   1 .
2 2
X 2 , x2 dX
X 1 , x1
dx
undeformed
Figure 2.2.9: a circular element deforming into an ellipse

Section 2.2
2.2.6 Some Simple Deformations

In this section, some elementary deformations are considered.
Pure Stretch
This deformation has already been seen, but now it can be viewed as a special case of the
polar decomposition. The motion is
x1  1 X 1 , x 2   2 X 2 , x3  3 X 3 Pure Stretch (2.2.38)
and the deformation gradient is
1 0 0 1 0 0 1 0 0
F   0 2 0   RU  0 1 0  0 2 0 
 0 0 3  0 0 1  0 0 3 
Here, R  I and there is no rotation. U  F and the principal material axes are
coincident with the material coordinate axes. 1 ,  2 , 3 , the eigenvalues of U, are the
principal stretches.
Stretch with rotation
Consider the motion
x1  X 1  kX 2 , x 2  kX 1  X 2 , x3  X 3
so that
1  k 0 cos   sin  0 sec  0 0

F  k 1 0  RU   sin 
 cos  0  0 sec  0
 0 0 1  0 0 1  0 0 1
where k  tan  . This decomposition shows that the deformation consists of material
stretching by sec ( 1  k 2 ) , the principal stretches, along each of the axes, followed
by a rigid body rotation through an angle  about the X 3  0 axis, Fig. 2.2.10. The
deformation is relatively simple because the principal material axes are aligned with the
material coordinate axes (so that U is diagonal). The deformation of the unit square is as
shown in Fig. 2.2.10.

Section 2.2
X 2 , x2

X 1 , x1
1 k 2
Figure 2.2.10: stretch with rotation
Pure Shear
Consider the motion
x1  X 1  kX 2 , x 2  kX 1  X 2 , x3  X 3 Pure Shear (2.2.39)
so that
 1 k 0 1 0 0 1 k 0
 
F  k 1 0  RU  0 1 0 k 1 0
 0 0 1 0 0 1  0 0 1
where, since F is symmetric, there is no rotation, and F  U . Since the rotation is zero,
one can work directly with U and not have to consider C. The eigenvalues of U, the
principal stretches, are 1  k , 1  k , 1 , with corresponding principal directions
Nˆ  1 E  1 E ,N ˆ   1 E  1 E and N ˆ E .
1 2 1 2 2 2 2 1 2 2 3 3
The deformation of the unit square is as shown in Fig. 2.2.11. The diagonal indicated by
the heavy line stretches by an amount 1  k whereas the other diagonal contracts by an
amount 1  k . An element of material along the diagonal will undergo a pure stretch as
indicated by the stretching of the dotted box.
X 2 , x2
k N̂1
k
X 1 , x1
Figure 2.2.11: pure shear

Section 2.2
Simple Shear
Consider the motion
x1  X 1  kX 2 , x 2  X 2 , x3  X 3 Simple Shear (2.2.40)
so that
1 k 0  1 k 0
F  0 1 0, 
C  k 1  k 2 0
0 0 1  0 0 1
The invariants of C are I C  3  k 2 , II C  3  k 2 , IIIC  1 and the characteristic equation

is 3  (3  k 2 ) (1   )  1  0 , so the principal values of C are
  1  12 k 2  12 k 4  k 2 , 1 . The principal values of U are the (positive) square-roots of
these:   12 4  k 2  12 k , 1 . These can be written as   sec  tan  , 1 by letting
tan   12 k . The corresponding eigenvectors of C are
ˆ  k ˆ  k ˆ E
N 1 E1  E 2 , N 2 E1  E 2 , N 3 3
1
2 k  k 4k
2 1
2
2 1
2 k  k 4k
2 1
2
2
or, normalizing so that they are of unit size, and writing in terms of  ,
ˆ  1  sin  E  1  sin  E , N
N ˆ   1  sin  E  1  sin  E , N
ˆ E
1 1 2 2 1 2 3 3
2 2 2 2
The transformation matrix of direction cosines is then
 1  sin   / 2  1  sin   / 2 0
Q   1  sin   / 2 1  sin   / 2 0
 0 0 1

so that, using the inverse transformation formula, U   Q U Q  , one obtains U in
T
terms of the original coordinates, and hence
1 k 0   cos  sin  0 cos  sin  0

 
F  0 1 0  RU   sin  cos  0  sin  (1  sin  ) / cos 
2
0
0 0 1  0 0 1  0 0 1

Section 2.2
The deformation of the unit square is shown in Fig. 2.2.12 (for k  0.2,   5.7 o ). The
square first undergoes a pure stretch/contraction along the principal axes, and is then
brought to its final position by a negative (clockwise) rotation of  .
For this deformation, det F  1 and, as will be shown below, this means that the simple
shear deformation is volume-preserving.
N̂1
X 2 , x2
N̂ 2
X 1 , x1

Figure 2.2.12: simple shear
2.2.7 Displacement & Displacement Gradients

The displacement of a material particle7 is the movement it undergoes in the transition
from the reference configuration to the current configuration. Thus, Fig. 2.2.13,8
U( X, t )  x( X, t )  X Displacement (Material Description) (2.2.41)
u(x, t )  x  X(x, t ) Displacement (Spatial Description) (2.2.42)
Note that U and u have the same values, they just have different arguments.
Uu
X x
Figure 2.2.13: the displacement
7
In solid mechanics, the motion and deformation are often described in terms of the displacement u. In
fluid mechanics, however, the primary field quantity describing the kinematic properties is the velocity v
(and the acceleration a  v ) – see later.
8
The material displacement U here is not to be confused with the right stretch tensor discussed earlier.

Section 2.2
Displacement Gradients
The displacement gradient in the material and spatial descriptions, U( X, t ) / X and
u(x, t ) / x , are related to the deformation gradient and the inverse deformation gradient
through
U  (x  X) U i x
Grad U    FI  i   ij
X X X j X j
(2.2.43)
u  (x  X) u i X
gradu    I  F 1   ij  i
x x x j x j
and it is clear that the displacement gradients are related through (see Eqn. 2.2.8)
gradu  Grad U F 1 (2.2.44)
The deformation can now be written in terms of either the material or spatial displacement
gradients:
dx  dX  dU( X)  dX  GradU dX
(2.2.45)
dx  dX  du(x)  dX  gradu dx
Example
Consider again the extension of the bar shown in Fig. 2.1.5. The displacement is
 t  3x1t 
U( X )   t  3 X 1t  E1 , u( x )    e1
 1  3t 
and the displacement gradients are
 3t 
GradU  3tE1 , gradu    e1
 1  3t 
The displacement is plotted in Fig. 2.2.14 for t  1 . The two gradients U1 / X 1 and
u1 / x1 have different values (see the horizontal axes on Fig. 2.2.14). In this example,
U1 / X 1  u1 / x1 – the change in displacement is not as large when “seen” from the
spatial coordinates.

Section 2.2
U1  u1
8
1 2 3 X1
1 5 9 13 x1
Figure 2.1.14: displacement and displacement gradient
Strains in terms of Displacement Gradients
The strains can be written in terms of the displacement gradients. Using 1.10.3b,
E
1 T
2

F FI 
1

 GradU  I  GradU  I   I
2
T

1
2
 T T 1  U
 GradU  GradU   GradU  GradU , E I J   I 
2  X J X I
 
U J U K U K 

X I X J 
(2.2.46a)
e
1
2

I  F T F 1 
1

 I  I  gradu  I  gradu 
2
T

1  u i u j u k u k 

1
2
gradu  gradu   gradu  gradu ,
T T
 eij    
2  x j xi xi x j 

(2.2.46b)
Small Strain
If the displacement gradients are small, then the quadratic terms, their products, are small
relative to the gradients themselves, and may be neglected. With this assumption, the
Green-Lagrange strain E (and the Euler-Almansi strain) reduces to the small-strain
tensor,
ε
1
2

GradU  GradU  ,
T
 1  U
 I J   I  J U 

2  X J X I 
(2.2.47)

Section 2.2
Since in this case the displacement gradients are small, it does not matter whether one
refers the strains to the reference or current configurations – the error is of the same order
as the quadratic terms already neglected9, so the small strain tensor can equally well be
written as
1  u u j 
ε
1
2

gradu  gradu  ,
T
  ij   i  
2  x j xi 
Small Strain Tensor (2.2.48)
2.2.8 The Deformation of Area and Volume Elements

Line elements transform between the reference and current configurations through the
deformation gradient. Here, the transformation of area and volume elements is examined.
The Jacobian Determinant
The Jacobian determinant of the deformation is defined as the determinant of the

deformation gradient,
x1 x1 x1

X 1 X 2 X 3
x 2 x 2 x 2
J ( X, t )  det F detF  The Jacobian Determinant (2.2.49)
X 1 X 2 X 3
x3 x3 x3
X 1 X 2 X 3
Equivalently, it can be considered to be the Jacobian of the transformation from material

to spatial coordinates (see Appendix 1.B.2).
From Eqn. 1.3.17, the Jacobian can also be written in the form of the triple scalar product
x  x x 
J     (2.2.50)
X 1  X 2 X 3 
Consider now a volume element in the reference configuration, a parallelepiped bounded

by the three line-elements dX (1) , dX ( 2) and dX ( 3) . The volume of the parallelepiped10 is
given by the triple scalar product (Eqns. 1.1.4):

dV  dX (1)  dX ( 2 )  dX ( 3)  (2.2.51)
After deformation, the volume element is bounded by the three vectors dx (i ) , so that the
volume of the deformed element is, using 1.10.16f,
9
although large rigid body rotations must not be allowed – see §2.7 .
10
the vectors should form a right-handed set so that the volume is positive.

Section 2.2

dv  dx (1)  dx ( 2)  dx (3) 
 FdX (1)

 FdX ( 2)
 FdX ( 3)  (2.2.52)
 det F dX (1)
 dX ( 2)
 dX ( 3)

 det F dV
Thus the scalar J is a measure of how the volume of a material element has changed with
the deformation and for this reason is often called the volume ratio.
dv  J dV Volume Ratio (2.2.53)
Since volumes cannot be negative, one must insist on physical grounds that J  0 . Also,
since F has an inverse, J  0 . Thus one has the restriction
J 0 (2.2.54)
Note that a rigid body rotation does not alter the volume, so the volume change is
completely characterised by the stretching tensor U. Three line elements lying along the
principal directions of U form an element with volume dV , and then undergo pure stretch
into new line elements defining an element of volume dv  1 2 3 dV , where i are the
principal stretches, Fig. 2.2.15. The unit change in volume is therefore also
dv  dV
 12 3  1 (2.2.55)
dV
reference current
configuration configuration
dV dv  123dV
principal material
axes
Figure 2.2.15: change in volume
For example, the volume change for pure shear is  k 2 (volume decreasing) and, for
simple shear, is zero (cf. Eqn. 2.2.40 et seq., (sec  tan  )(sec  tan  )(1)  1  0 ).
An incompressible material is one for which the volume change is zero, i.e. the
deformation is isochoric. For such a material, J  1 , and the three principal stretches are
not independent, but are constrained by
12 3  1 Incompressibility Constraint (2.2.56)

Section 2.2
Nanson’s Formula
Consider an area element in the reference configuration, with area dS , unit normal N̂ ,
and bounded by the vectors dX (1) , dX ( 2 ) , Fig. 2.2.16. Then
ˆ dS  dX (1)  dX ( 2)
N (2.2.57)
The volume of the element bounded by the vectors dX (1) , dX ( 2 ) and some arbitrary line
element dX is dV  N ˆ dS  dX . The area element is now deformed into an element of
area ds with normal n̂ and bounded by the line elements dx (1) , dx ( 2 ) . The volume of the
new element bounded by the area element and dx  FdX is then
ˆ dS  dX
dv  nˆ ds  dx  nˆ ds  FdX  JN (2.2.58)
dX
dx
N̂ dX (2)
dx ( 2 )
n̂
dX (1)
dx (1)
Figure 2.2.16: change of surface area
Thus, since dX is arbitrary, and using 1.10.3d,
ˆ dS
nˆ ds  J F  T N Nanson’s Formula (2.2.59)
Nanson’s formula shows how the vector element of area n̂ds in the current
configuration is related to the vector element of area N̂dS in the reference configuration.
2.2.9 Inextensibility and Orientation Constraints

A constraint on the principal stretches was introduced for an incompressible material,
2.2.56. Other constraints arise in practice. For example, consider a material which is
inextensible in a certain direction, defined by a unit vector Â in the reference
configuration. It follows that FA ˆ  1 and the constraint can be expressed as 2.2.17,
ˆ CA
A ˆ 1 Inextensibility Constraint (2.2.60)

Section 2.2
If there are two such directions in a plane, defined by Â and B̂ , making angles  and 
respectively with the principal material axes N ˆ ,N
ˆ , then
1 2
12 0 0  cos  
 
1  cos  sin  0 0  2
2 0   sin  
0 0 32   0 

   
and 12  22 cos 2   1  22  12  22 cos 2  . It follows that    ,      ,
     or     2 (or 1  2  1 , i.e. no deformation).
Similarly, one can have orientation constraints. For example, suppose that the direction
associated with the vector Â maintains that direction. Then
ˆ  A
FA ˆ Orientation Constraint (2.2.61)
for some scalar   0 .
2.2.10 Problems
1. In equations 2.2.8, one has from the chain rule
  X m    X 
grad  ei  ei   E j  m E m  e i   Grad F 1
xi X m xi  
 X j  xi 
Derive the other two relations.
2. Take the dot product   dxˆ     dxˆ  in Eqn. 2.2.29. Then use R T R  I , U T  U , and
1.10.3e to show that
dX dX
2  UU
dX dX
3. For the deformation
x1  X 1  2 X 3 , x 2  X 2  2 X 3 , x3  2 X 1  2 X 2  X 3
(a) Determine the Deformation Gradient and the Right Cauchy-Green tensors
(b) Consider the two line elements dX (1)  e1 , dX ( 2 )  e 2 (emanating from (0,0,0)).
Use the Right Cauchy Green tensor to determine whether these elements in the
current configuration ( dx (1) , dx ( 2) ) are perpendicular.
(c) Use the right Cauchy Green tensor to evaluate the stretch of the line element
dX  e1  e 2 , and hence determine whether the element contracts, stretches, or
stays the same length after deformation.
(d) Determine the Green-Lagrange and Eulerian strain tensors
(e) Decompose the deformation into a stretching and rotation (check that U is
symmetric and R is orthogonal). What are the principal stretches?
4. Derive Equations 2.2.36.
x1  X 1 , x 2  X 2  X 3 , x3  aX 2  X 3

Section 2.2
(a) Determine the displacement vector in both the material and spatial forms
(b) Determine the displaced location of the particles in the undeformed state which
originally comprise
(i) the plane circular surface X 1  0, X 22  X 32  1 /(1  a 2 )
(ii) the infinitesimal cube with edges along the coordinate axes of length
dX i  
Sketch the displaced configurations if a  1 / 2
x1  X 1  aX 2 , x 2  X 2  aX 3 , x3  aX 1  X 3
(a) Determine the displacement vector in both the material and spatial forms
(b) Calculate the full material (Green-Lagrange) strain tensor and the full spatial
strain tensor
(c) Calculate the infinitesimal strain tensor as derived from the material and spatial
tensors, and compare them for the case of very small a.
7. In the example given above on the polar decomposition, §2.2.5, check that the
relations Cn i  n i , i  1,2,3 are satisfied (with respect to the original axes). Check
also that the relations Cn i  n i , i  1,2,3 are satisfied (here, the eigenvectors are the
unit vectors in the second coordinate system, the principal directions of C, and C is
with respect to these axes, i.e. it is diagonal).

Section 2.3
2.3 Deformation and Strain: Further Topics
2.3.1 Volumetric and Isochoric Deformations

When analysing materials which are only slightly incompressible, it is useful to
decompose the deformation gradient multiplicatively, according to
( )
F = J 1/ 3I F = J 1/ 3 F (2.3.1)
From this definition {▲Problem 1},
det F = 1 (2.3.2)
and so F characterises a volume preserving (distortional or isochoric) deformation. The

tensor J 1 / 3 I characterises the volume-changing (dilational or volumetric) component of
the deformation, with det (J 1 / 3 I ) = det F = J .
This concept can be carried on to other kinematic tensors. For example, with C = F T F ,
C = J 2/ 3F T F ≡ J 2/3C . (2.3.3)
F and C are called the modified deformation gradient and the modified right
Cauchy-Green tensor, respectively. The square of the stretch is given by
ˆ CdX
λ2 = dX ˆ = J 2 / 3 dX{
ˆ C dX
ˆ } (2.3.4)
so that λ = J 1 / 3 λ , where λ is the modified stretch, due to the action of C . Similarly,

the modified principal stretches are
λi = J −1 / 3 λi , i = 1,2,3 (2.3.5)
with
det F = λ1λ2 λ3 = 1 (2.3.6)
The case of simple shear discussed earlier is an example of an isochoric deformation, in

which the deformation gradient and the modified deformation gradient coincide,
J 1/ 3I = I .
2.3.2 Relative Deformation
It is usual to use the configuration at ( X, t = 0) as the reference configuration, and define

quantities such as the deformation gradient relative to this reference configuration. As
mentioned, any configuration can be taken to be the reference configuration, and a new

Section 2.3
deformation gradient can be constructed with respect to this new reference configuration.
Further, the reference configuration does not have to be fixed, but could be moving also.
In many cases, it is useful to choose the current configuration (x, t ) to be the reference
configuration, for example when evaluating rates of change of kinematic quantities (see
later). To this end, introduce a third configuration: this is the configuration at some time
t = τ and the position of a material particle X here is denoted by xˆ = χ ( X,τ ) , where χ is
the motion function. The deformation at this time τ relative to the current configuration
is called the relative deformation, and is denoted by xˆ = χ (t ) (x,τ ) , as illustrated in Fig.
2.3.1.
configuration
xˆ = χ ( X,τ ) at t = τ
χ ( X, τ ) = χ ( t ) ( x, τ )
relative
X, t
deformation
χ ( t ) ( x, τ )
Ft
initial
configuration χ ( X, t ) x = χ ( X, t )
= χ ( t ) ( x, t )
current
F configuration
Figure 2.3.1: the relative deformation
The relative deformation gradient Ft is defined through
∂xˆ
dxˆ = Ft (x,τ ) dx , Ft = (2.3.7)
∂x
Also, since dx = F( X, t ) dX and dxˆ = F( X,τ ) dX , one has the relation
F( X,τ ) = Ft (x,τ ) F( X, t ) (2.3.8)
Similarly, relative strain measures can be defined, for example the relative right Cauchy-
Green strain tensor is
C t (τ ) = Ft (τ ) Ft (τ )
T
(2.3.9)
Example
Consider the two-dimensional motion

Section 2.3
x1 = X 1e t , x 2 = X 2 (t + 1)
Inverting these gives the spatial description X 1 = x1e − t , X 2 = x 2 /(t + 1) , and the relative
deformation is
xˆ1 (x,τ ) = X 1eτ = x1eτ −t

xˆ 2 (x,τ ) = X 2 (τ + 1) = x 2 (τ + 1) /(t + 1)
The deformation gradients are
∂xi
F ( X, t ) = e i ⊗ E j = e t e1 ⊗ E1 + (t + 1)e 2 ⊗ E 2
∂X j
∂xˆ
Ft (x,τ ) = i e i ⊗ e j = eτ −t e1 ⊗ e1 + (τ + 1) /(t + 1) e 2 ⊗ e 2
∂x j
■
2.3.3 Derivatives of the Stretch

In this section, some useful formulae involving the derivatives of the stretches with
respect to the Cauchy-Green strain tensors are derived.
Derivatives with respect to b
First, take the stretches to be functions of the left Cauchy-Green strain b. Write b using
the spatial principal directions n̂ i as a basis, 2.2.37, so that the total differential can be
expressed as
3
db = ∑ 2λi dλi nˆ i ⊗ nˆ i + λi2 [dnˆ i ⊗ nˆ i + nˆ i ⊗ dnˆ i ] (2.3.10)
i =1
Since nˆ i ⋅ nˆ j = δ ij , then
nˆ i dbnˆ i = 2λi dλi + λi2 [nˆ i ⋅ dnˆ i + dnˆ i ⋅ nˆ i ] = 2λi dλi (no sum over i) (2.3.11)
This last follows since the change in a vector of constant length is always orthogonal to
the vector itself (as in the curvature analysis of §1.6.2). Using the property
uTv = T : (u ⊗ v) , one has (summing over the k but not over the i; here dλ k / dλi = δ ik )
∂b 1 ∂b
db : (nˆ i ⊗ nˆ i ) ≡ dλ k : (nˆ i ⊗ nˆ i ) = 2λi dλi → : (nˆ i ⊗ nˆ i ) = 1 (2.3.12)
∂λ k 2λi ∂λi
Then, since ∂b / ∂λi : ∂λi / ∂b is also equal to 1, one has

Section 2.3
1 ∂b ∂b ∂λi ∂λi 1
: (nˆ i ⊗ nˆ i ) = : → = (nˆ i ⊗ nˆ i ) (2.3.13)
2λi ∂λi ∂λi ∂b ∂b 2λi
The chain rule then gives the second derivative.
The above analysis is for distinct principal stretches. When λ1 = λ 2 = λ3 ≡ λ , then

b = λ2 I , db = 2λdλI . Also, db = 3(∂b / ∂λ )dλ , so 3(∂b / ∂λ ) = 2λI , or
∂b ∂λ ∂λ
3 : = 2λI : (2.3.14)
∂λ ∂b ∂b
But ∂b / ∂λ : ∂λ / ∂b = 1 and 3 = I : I , and so in this case, ∂λ / ∂b = I / 2λ .
A similar calculation can be carried out for two equal eigenvalues λ1 = λ 2 = λ ≠ λ3 . In

summary,
∂λi 1
= nˆ i ⊗ nˆ i (no sum over i ) λ1 ≠ λ 2 ≠ λ3 ≠ λ1
∂b 2λi
∂λ 1
= (nˆ 1 ⊗ nˆ 1 + nˆ 2 ⊗ nˆ 2 )
∂b 2λ λ1 = λ 2 = λ ≠ λ3
∂λ3 1
= (nˆ 3 ⊗ nˆ 3 )
∂b 2λ3
∂λ 1 1
∑
3
= nˆ ⊗ ˆ
n = I λ1 = λ 2 = λ3 = λ
∂b 2λ i =1 2λ
i i
(2.3.15)
∂ λi
2
1
= − 3 nˆ i ⊗ nˆ i ⊗ nˆ i ⊗ nˆ i (no sum over i ) λ1 ≠ λ 2 ≠ λ3 ≠ λ1
∂b 2
4λi
Derivatives with respect to C
The stretch can also be considered to be a function of the right Cauchy-Green strain C.
The derivatives of the stretches with respect to C can be found in exactly the same way as
for the left Cauchy-Green strain. The results are the same as given in 2.3.15 except that,
referring to 2.2.37, b is replaced by C and n̂ is replaced by N̂ .
2.3.4 The Directional Derivative of Kinematic Quantities

The directional derivative of vectors and tensors was introduced in §1.6.11 and §1.15.4.
Taking directional derivatives of kinematic quantities is often very useful, for example in
linearising equations in order to apply numerical solution algorithms
The Deformation Gradient
First, consider the deformation gradient as a function of the current position x (or motion
χ ) and examine its value at x + a :

Section 2.3
F(x + a) = F(x) + ∂ x F[a ] + o( a ) (2.3.16)
The directional derivative ∂ x F[a] = (∂F / ∂x )a can be expressed as
∂ x F[a] = F(x + εa )
d
dε ε =0
d ∂ (x + εa )
= (2.3.17)
dε ε =0 ∂X
= Grada
= (grada )F
the last line resulting from 2.2.8b. It follows that the directional derivative of the
deformation gradient in the direction of a displacement vector u from the current
configuration is
∂ x F[u ] = (gradu )F (2.3.18)
On the other hand, consider the deformation gradient as a function of X and examine its
value at X + A :
F ( X + A) = F ( X) + ∂ X F[A ] (2.3.19)
and now
∂ X F[A ] = F(X + εA )
d
dε ε =0
∂
x(X + εA )
d
=
dε ε =0 ∂X
∂
=
d
(x + FεA ) (2.3.20)
dε ε =0 ∂X
= Grad(FA )
= Grada
where a = FA .
Other Kinematic Quantities
The directional derivative of the Green-Lagrange strain, the right and left Cauchy-Green
tensors and the Jacobian in the direction of a displacement u from the current
configuration are {▲Problem 2}

Section 2.3
∂ x E[u] = F T εF
∂ x C[u] = 2F T εF
(2.3.21)
∂ x b[u] = (gradu )b + b(gradu )
T
∂ x J [u] = Jdivu
where ε is the small-strain tensor, 2.2.48.
The directional derivative is also useful for deriving various relations between the
kinematic variables. For example, for an arbitrary vector a, using the chain rule 1.15.28,
2.3.20, 1.15.24, the trace relations 1.10.10e and 1.10.10b, and 2.2.8b, 1.14.9,
(GradJ ) ⋅ a = ∂ X J [a]
= ∂ F J [∂ X F[a]]
= ∂ F J [Grad(Fa )]
= JF −T : Grad(Fa )
(2.3.22)
= Jtr (F −1Grad(Fa ))
= Jtr (Grad(Fa )F −1 )
= Jtr (grad(Fa ))
= J div(Fa )
so that, from 1.14.16b with a constant,
GradJ = JdivF T (2.3.23)
2.3.5 Problems
1. Use 1.10.16c to show that det F = 1 .

( )
2. (a) use the relation E = 12 F T F − I , Eqn. 2.3.18, ∂ x F[u] = (gradu )F , and the product
rule of differentiation to derive 2.3.21a, ∂ x E[u] = F T εF , where ε is the small
strain tensor.
(b) evaluate ∂ x C[u] (in terms of F and ε , the small strain tensor)
(c) evaluate ∂ x b[u ] (in terms of gradu and b)
(d) evaluate ∂ J [u ] (in terms of J and divu ; use the chain rule ∂ J [u] = ∂ Jˆ [∂ F[u ]] ,
x x F x
with Jˆ (F ) = det F , ∂ x F[u] = Gradu )

Section 2.4
2.4 Material Time Derivatives

The motion is now allowed to be a function of time, x  χ X, t  , and attention is given to
time derivatives, both the material time derivative and the local time derivative.
2.4.1 Velocity & Acceleration

The velocity of a moving particle is the time rate of change of the position of the particle.
From 2.1.3, by definition,
dχ ( X, t )
V ( X, t )  (2.4.1)
dt
In the motion expression x  χ X, t  , X and t are independent variables and X is

independent of time, denoting the particle for which the velocity is being calculated. The
velocity can thus be written as χ (X, t ) / t or, denoting the motion by x( X, t ) , as
dx(X, t ) / dt or x(X, t ) / t .
The spatial description of the velocity field may be obtained from the material description
by simply replacing X with x, i.e.

v (x, t )  V χ 1 (x, t ), t  (2.4.2)
As with displacements in both descriptions, there is only one velocity, V ( X, t )  v(x, t ) –

they are just given in terms of different coordinates.
The velocity is most often expressed in the spatial description, as
dx
v (x, t )  x  velocity (2.4.3)
dt
To be precise, the right hand side here involves x which is a function of the material
coordinates, but it is understood that the substitution back to spatial coordinates, as in
2.4.2, is made (see example below).
Similarly, the acceleration is defined to be
d 2 χ ( X, t ) d 2 x dV  2 χ ( X, t )
A( X, t )   2   (2.4.4)
dt 2 dt dt t 2
Example
Consider the motion
x1  X 1  t 2 X 2 , x 2  X 2  t 2 X 1 , x3  X 3

Section 2.4
The velocity and acceleration can be evaluated through
dx d 2x
V ( X, t )   2tX 2 e1  2tX 1e 2 , A( X, t )  2  2 X 2 e1  2 X 1e 2
dt dt
One can write the motion in the spatial description by inverting the material description:
x1  t 2 x 2 x 2  t 2 x1
X1  , X2  , X 3  x3
1 t4 1 t4
Substituting in these equations then gives the spatial description of the velocity and
acceleration:
x2  t 2 x1 x1  t 2 x2
v( x, t )  V  χ 1 ( x, t ), t   2t e1  2 t e2
1  t4 1  t4
x2  t 2 x1 x1  t 2 x2
a( x, t )  A  χ ( x, t ), t   2
1
e1  2 e2
1  t4 1  t4
2.4.2 The Material Derivative

One can analyse deformation by examining the current configuration only, discounting
the reference configuration. This is the viewpoint taken in Fluid Mechanics – one focuses
on material as it flows at the current time, and does not consider “where the fluid was”.
In order to do this, quantities must be cast in terms of the velocity. Suppose that the
velocity in terms of spatial coordinates, v  v(x, t ) is known; for example, one could
have a measuring instrument which records the velocity at a specific location, but the
motion χ itself is unknown. In that case, to evaluate the acceleration, the chain rule of
differentiation must be applied:
v v dx
vx(t ), t  
d
v  
dt t x dt
or
v
a  grad v v acceleration (spatial description) (2.4.5)
t
The acceleration can now be determined, because the derivatives can be determined
(measured) without knowing the motion.
In the above, the material derivative, or total derivative, of the particle’s velocity was
taken to obtain the acceleration. In general, one can take the time derivative of any
physical or kinematic property   expressed in the spatial description:

Section 2.4
d
     grad v Material Time Derivative (2.4.6)
dt t
For example, the rate of change of the density    (x, t ) of a particle instantaneously at
x is
d 
    grad   v (2.4.7)
dt t
The Local Rate of Change
The first term,  / t , gives the local rate of change of density at x whereas the second
term v  grad  gives the change due to the particle’s motion, and is called the convective
rate of change.
Note the difference between the material derivative and the local derivative. For example,
the material derivative of the velocity, 2.4.5 (or, equivalently, dV ( X, t ) / dt in 2.4.4, with
X fixed) is not the same as the derivative v ( x, t ) / t (with x fixed). The former is the
acceleration of a material particle X. The latter is the time rate of change of the velocity
of particles at a fixed location in space; in general, different material particles will occupy
position x at different times.
The material derivative d / dt can be applied to any scalar, vector or tensor:
d 
    grad   v
dt t
da a
a    grad a v (2.4.8)
dt t
  dA  A  grad A v
A
dt t
Another notation often used for the material derivative is D / Dt :
Df df
  f (2.4.9)
Dt dt
Steady and Uniform Flows
In a steady flow, quantities are independent of time, so the local rate of change is zero
and, for example,   grad   v . In a uniform flow, quantities are independent of
position so that, for example,    / t
Example
Consider again the previous example. This time, with only the velocity v(x, t ) known,
the acceleration can be obtained through the material derivative:

Section 2.4
v
a(x, t )   grad v v
t
 2t 3 2t  x 2  t 2 x1 
 0 2t 
 1 t
4
1 t4  1 t4 
  x  t x12
x  t x 2   2t
2
2t 3 x1  t x 2 
2
  2t 2 e1  2t 1 e 2    0 2t
t  1 t 4
1 t4   1 t
4
1 t4  1 t4 
 0 0 0  0 
  
   
x 2  t 2 x1 x1  t 2 x 2
2 e 1  2 e2
1 t 4 1 t4
as before.
■
The Relationship between the Displacement and Velocity
The velocity can be derived directly from the displacement 2.2.42:
dx d (u  X) du
v   , (2.4.10)
dt dt dt
or
du u
v   grad u v (2.4.11)
dt t
When the displacement field is given in material form one has
dU
V (2.4.12)
dt
2.4.3 Problems
1. The density of a material is given by
e 2t

xx
The velocity field is given by
v1  x 2  2 x3 , v 2  x3  2 x1 , v3  x1  2 x 2
Determine the time derivative of the density (a) at a certain position x in space,
and (b) of a material particle instantaneously occupying position x.

Section 2.5
2.5 Deformation Rates

In this section, rates of change of the deformation tensors introduced earlier, F, C, E, etc.,
are evaluated, and special tensors used to measure deformation rates are discussed, for
example the velocity gradient l, the rate of deformation d and the spin tensor w.
2.5.1 The Velocity Gradient

The velocity gradient is used as a measure of the rate at which a material is deforming.
Consider two fixed neighbouring points, x and x  dx , Fig. 2.5.1. The velocities of the
material particles at these points at any given time instant are v (x) and v(x  dx) , and
v
v (x  dx)  v (x)  dx ,
x
The relative velocity between the points is
v
dv  dx  ldx (2.5.1)
x
with l defined to be the (spatial) velocity gradient,
v vi
l  grad v, lij  Spatial Velocity Gradient (2.5.2)
x x j
dv
v x  dx 
 x  dx
v x 
 dx
x
Figure 2.5.1: velocity gradient
Expression 2.5.1 emphasises the tensorial character of the spatial velocity gradient,
mapping as it does one vector into another. Its physical meaning will become clear when
it is decomposed into its symmetric and skew-symmetric parts below.
The spatial velocity gradient is commonly used in both solid and fluid mechanics. Less
commonly used is the material velocity gradient, which is related to the rate of change of
the deformation gradient:
V ( X, t )   x( X, t )    x( X, t )  
Grad V      F (2.5.3)
X X  t  t  X 

Section 2.5
and use has been made of the fact that, since X and t are independent variables, material
time derivatives and material gradients commute.
2.5.2 Material Derivatives of the Deformation Gradient

The spatial velocity gradient may be written as
v v X   x  X   x  X
      
x X x X  t  x t  X  x
or l  F F 1 so that the material derivative of F can be expressed as
F  l F Material Time Derivative of the Deformation Gradient (2.5.4)
Also, it can be shown that {▲Problem 1}
F T  F T
.
F 1  F 1l (2.5.5)
.
T
F  l T F  T
2.5.3 The Rate of Deformation and Spin Tensors
The velocity gradient can be decomposed into a symmetric tensor and a skew-symmetric
tensor as follows (see §1.10.10):
l dw (2.5.6)
where d is the rate of deformation tensor (or rate of stretching tensor) and w is the
spin tensor (or rate of rotation, or vorticity tensor), defined by
1  vi v j 
d
1
2

l  lT ,  d ij  
2  x j xi 
Rate of Deformation and Spin Tensors
1  vi v j 
w
1
2

l  lT ,  wij  
2  x j xi 
(2.5.7)
The physical meaning of these tensors is next examined.

Section 2.5
The Rate of Deformation
Consider first the rate of deformation tensor d and note that
ldx  dv 
d
dx  (2.5.8)
dt
The rate at which the square of the length of dx is changing is then
d
dt
 
dx  2 dx  dx ,
2 d
dt
(2.5.9)
d
dt
 
dx  dx  dx   2dx  dx   2dx l dx  2dx d dx
2 d
dt
d
dt
2
the last equality following from 2.5.6 and 1.10.31e. Dividing across by 2 dx , then leads
to

 nˆ dnˆ Rate of stretching per unit stretch in the direction n̂ (2.5.10)

where   dx / dX is the stretch and nˆ  dx / dx is a unit normal in the direction of dx .

Thus the rate of deformation d gives the rate of stretching of line elements. The diagonal
components of d, for example
d 11  e1de1 ,
represent unit rates of extension in the coordinate directions.
Note that these are instantaneous rates of extension, in other words, they are rates of
extensions of elements in the current configuration at the current time; they are not a
measure of the rate at which a line element in the original configuration changed into the
corresponding line element in the current configuration.
Note:
 Eqn. 2.5.10 can also be derived as follows: let N̂ be a unit normal in the direction of dX , and
n̂ be the corresponding unit normal in the direction of dx . Then nˆ dx  FN ˆ dX , or nˆ   FN ˆ .
Differentiating gives nˆ   nˆ   F N ˆ  lFN ˆ or nˆ   nˆ   lnˆ  . Contracting both sides with n̂
leads to nˆ  nˆ  nˆ  nˆ  /    nˆ lnˆ . But nˆ  nˆ  1  d (nˆ  nˆ ) dt  0 so, by the chain rule, nˆ  nˆ  0
(confirming that a vector n̂ of constant length is orthogonal to a change in that vector dn̂ ), and
the result follows
Consider now the rate of change of the angle  between two vectors dx (1) , dx ( 2) . Using
2.5.8 and 1.10.3d,

Section 2.5
d
dt

dx (1)  dx ( 2 )  d
dt
 
dx (1)  dx ( 2 )  dx (1) 
d
dt
dx ( 2 )  
 ldx  dx  dx  ldx
(1) ( 2) (1) ( 2)
(2.5.11)

 l  l dx T
 (1)
 dx ( 2)
 2 dx ddx (1) ( 2)
which reduces to 2.5.9 when dx (1)  dx ( 2) . An alternative expression for this dot product
is
d
dt

dx (1) dx ( 2 ) cos   d
dt
 
dx (1) dx ( 2) cos  
d
dt
 
dx ( 2 ) dx (1) cos   sin   dx (1) dx ( 2 )
d
 dx (1) 
d
dx ( 2)  
 
  dt (1) cos   dt ( 2 ) cos   sin    dx (1) dx ( 2 )
 dx dx 
 
 
(2.5.12)
Equating 2.5.11 and 2.5.12 leads to
   
2 nˆ 1 d nˆ 2   1  2  cos   sin   (2.5.13)
 1  2 
where i  dx ( i ) / dX (i ) is the stretch and nˆ i  dx ( i ) / dx (i ) is a unit normal in the

direction of dx (i ) .
It follows from 2.5.13 that the off-diagonal terms of the rate of deformation tensor
represent shear rates: the rate of change of the right angle between line elements aligned
with the coordinate directions. For example, taking the base vectors e1  n̂ 1 , e 2  n̂ 2 ,
2.5.13 reduces to
1
d12   12 (2.5.14)
2
where 12 is the instantaneous right angle between the axes in the current configuration.
The Spin
Consider now the spin tensor w; since it is skew-symmetric, it can be written in terms of
its axial vector ω (Eqn. 1.10.34), called the angular velocity vector:

Section 2.5
ω   w23e1  w13e 2  w12e 3

1  v3 v2  1  v1 v3  1  v2 v1 
    e1     e2     e3 (2.5.15)
2  x2 x3  2  x3 x1  2  x1 x2 
1
 curl v
2
(The vector 2ω is called the vorticity (or spin) vector.) Thus when d is zero, the motion
consists of a rotation about some axis at angular velocity   ω (cf. the end of §1.10.11),
with v  ω  r , r measured from a point on the axis, and wr  ω  r  v .
On the other hand, when l  d , w  0 , one has ω  o , and the motion is called
irrotational.
Example (Shear Flow)
Consider a simple shear flow in which the velocity profile is “triangular” as shown in
Fig. 2.5.2. This type of flow can be generated (at least approximately) in many fluids by
confining the fluid between plates a distance h apart, and by sliding the upper plate over
the lower one at constant velocity V. If the material particles adjacent to the upper plate
have velocity Ve1 , then the velocity field is v  x 2 e1 , where   V / h . This is a steady
flow ( v / t  0 ); at any given point, there is no change over time. The velocity gradient
is l  e1  e 2 and the acceleration of material particles is zero: a  lv  0 . The rate of
deformation and spin are
 0  0  0  0
1 1
d   0 0, w    0 0
2 2
 0 0 0  0 0 0
and, from 2.5.14,   12 , the rate of change of the angle shown in Fig. 2.5.2.
e2
h
12 e v  v1 ( x2 )e1
1
Figure 2.5.2: shear flow
The eigenvalues of d are   0,   / 2 ( det d  0 ) and the principal invariants, Eqn.

1.11.17, are I d  0, II d   14  2 , IIId  0 . For    / 2 , the eigenvector is
n1  1 1 0 and for    / 2 , it is n 2   1 1 0 (for   0 it is e 3 ). (The
T T
eigenvalues and eigenvectors of w are complex.) Relative to the basis of eigenvectors,

Section 2.5
 / 2 0 0

d 0   / 2 0
 0 0 0
so at 45  there is an instantaneous pure rate of stretching/contraction of material.
2.5.4 Other Rates of Strain Tensors

From 2.2.9, 2.2.22,
1 d
dx  dx   dX 1 C dX  dXE dX (2.5.16)
2 dt 2
This can also be written in terms of spatial line elements:
dXE 
 dX  dx F  T E
 F 1 dx (2.5.17)
But from 2.5.9, these also equal dxddx , which leads to expressions for the material time
derivatives of the right Cauchy-Green and Green-Lagrange strain tensors (also given here
are expressions for the time derivatives of the left Cauchy-Green and Euler-Almansi
tensors {▲Problem 3})
  2F T dF
C
  F T dF
E
(2.5.18)
b  lb  bl T
e  d  l T e  el
Note that
 E dt   dE
so that the integral of the rate of Green-Lagrange strain is path independent and, in
particular, the integral of E around any closed loop (so that the final configuration is the
same as the initial configuration) is zero. However, in general, the integral of the rate of
deformation,
 ddt
is not independent of the path – there is no universal function h such that d  dh / dt with
 ddt   dh . Thus the integral  ddt over a closed path may be non-zero, and hence the
integral of the rate of deformation is not a good measure of the total strain.

Section 2.5
The Hencky Strain
The Hencky strain is, Eqn. 2.2.37, h  i 1 ln i nˆ i  nˆ i , where n i are the principal
3
spatial axes. Thus, if the principal spatial axes do not change with time,
 
h  i 1 i / i nˆ i  nˆ i . With the left stretch v  i 1 i nˆ i  nˆ i , it follows that (and
3 3
 
  ln U  U
similarly for the corresponding material tensors), H  U 1 , h  ln v  v v 1 .
For example, consider an extension in the coordinate directions, so

F  U  v  i 1 i nˆ i  nˆ i  i 1 i N
ˆ Nˆ . The motion and velocity are
3 3
i i

xi  i X i , x i  i X i  i xi no sum 
i
so d i  i / i (no sum), and d  h . Further, h   ddt . Note that, as mentioned above,
this expression does not hold in general, but does in this case of uniform extension.
2.5.5 Material Derivatives of Line, Area and Volume Elements
The material derivative of a line element d (dx) / dt has been derived (defined) through
2.4.8. For area and volume elements, it is necessary first to evaluate the material
derivative of the Jacobian determinant J. From the chain rule, one has (see Eqns 1.15.11,
1.15.7)
J 
J   J (F)  
d
: F  J F T : F (2.5.19)
dt F
Hence {▲Problem 4}
J  J tr (l )
 J tr (grad v) (2.5.20)
 Jdiv v
Since l  d  w and trw  0 , it also follows that J  Jtr d .
As mentioned earlier, an isochoric motion is one for which the volume is constant – thus
any of the following statements characterise the necessary and sufficient conditions for an
isochoric motion:
J  1, J  0, divv  0, tr d  0, F  T : F  0 (2.5.21)
Applying Nanson’s formula 2.2.59, the material derivative of an area vector element is
{▲Problem 6}

Section 2.5
d
dt

nˆ ds   divv  l T nˆ ds  (2.5.22)
Finally, from 2.2.53, the material time derivative of a volume element is
d
dv   d JdV   JdV  div v dv (2.5.23)
dt dt
Example (Shear and Stretch)
Consider a sample of material undergoing the following motion, Fig. 2.4.3.
X 1  x1  kx2
x1  X 1  kX 2
1
x 2  X 2 , X2  x2

x3  X 3
X 3  x3
with    (t ) , k  k (t ) .
X 2 , x2
k


X 1 , x1
Figure 2.4.3: shear and stretch
The deformation gradient and material strain tensors are
1 k 0 1 k 0  0 1
2 k 0
F  0  0, C  k
  1  k 
2 2
0, E   12 k
 1
2  1  k   1
2 2
0 ,
0 0 1  0 0 1  0 0 0
the Jacobian J  det F   , and the spatial strain tensors are

Section 2.5
1  k 2 2 k2 0 0 01
k
 
2
   1 k  1  2 2
b   k2 2 0 , e   12 k 1
0
 0 
2
2 
 0 1  0 0 0

This deformation can also be expressed as a stretch followed by a simple shear:
1 k 0 1 0 0
F  0 1 0 0  0
0 0 1 0 0 1
The velocity is

 k  k X 2   
 k  k  /  x 2  
V
dx 
dt
 X 2

,

v 
 /  x 2 


 0   0 
   
The velocity gradient is
0 k  k  /    0
dv  
l  0  /  0
dx 
0 0 0

and the rate of deformation and spin are
 0 1
k  k  /   0  0 1
k  k  /   0
   
2 2
1 

d   2 k  k  /   / 
  1 
0, w   2 k  k  /   0 0

 0 0 0  0 0 0
 
Also
 0 k  k 0
C  
  2F T dF  k  k 2 kk  k 2  1    0

 0 0 0

As expected, from 2.5.20,

J  Jtr (d)  J  /    
■

Section 2.5
2.5.6 Problems
1. (a) Differentiate the relation I  FF 1 and use 2.5.4, F  l F , to derive 2.5.5b,

.
F  F 1l .
1
(b) Differentiate the relation I  F T F  T and use 2.5.4, F  l F , and 1.10.3e to derive
.
2.5.5c, F  l T F T .
T
2. For the velocity field

v1  x12 x 2 , v 2  2 x 22 x3 , v3  3x1 x 2 x3
determine the rate of stretching per unit stretch at (2,0,1) in the direction of the unit
vector
4e1  3e 2  / 5
And in the direction of e1 ?
3. (a) Derive the relation 2.5.18a, C   2F T dF directly from C  F T F
(b) Use the definitions b  FF T and e  (I  b 1 ) / 2 to derive the relations
2.5.18c,d: b  lb  bl T , e  d  l T e  el
4. Use 2.5.4, 2.5.19, 1.10.3h, 1.10.6, to derive 2.5.20.
5. For the motion x1  3 X 1t  t 2 , x 2  X 1  X 2 t , x3  tX 3 , verify that F  lF . What is
the ratio of the volume element currently occupying (1,1,1) to its volume in the
undeformed configuration? And what is the rate of change of this volume element,
per unit current volume?
6. Use Nanson’s formula 2.2.59, the product rule of differentiation, and 2.5.20, 2.5.5c,
to derive the material time derivative of a vector area element, 2.5.22 (note that N̂ ,
a unit normal in the undeformed configuration, is constant).

Section 2.6
2.6 Deformation Rates: Further Topics
2.6.1 Relationship between l, d, w and the rate of change of R

and U
Consider the polar decomposition F = RU . Since R is orthogonal, RR T = I , and a

differentiation of this equation leads to
& R T = − RR
ΩR ≡ R &T (2.6.1)
with Ω R skew-symmetric (see Eqn. 1.14.2). Using this relation, the expression l = F& F −1 ,
and the definitions of d and w, Eqn. 2.5.7, one finds that {▲Problem 1}
& U −1 R T + Ω
l = RU R
w=
2
(
1 & −1
R UU − U −1 U & RT + Ω )
R
= Rskew U [
& U −1 R T + Ω ]
R (2.6.2)
d=
2
(
1 & −1
R UU + U −1 U & RT )
[
& U −1 R T
= Rsym U ]
Note that Ω R being skew-symmetric is consistent with w being skew-symmetric, and that
both w and d involve R, and the rate of change of U.
& = 0 , and
When the motion is a rigid body rotation, then U
& RT
w = ΩR = R (2.6.3)
2.6.2 Deformation Rate Tensors and the Principal Material and

Spatial Bases
The rate of change of the stretch tensor in terms of the principal material base vectors is
3
& = ∑ λ& N
U
i =1
i {
ˆ ⊗N
i
ˆ +λ N
i i
ˆ& ⊗ N
i
ˆ + λ& N
i i i
ˆ&
ˆ ⊗N
i } (2.6.4)
Consider the case when the principal material axes stay constant, as can happen in some
& and U −1 are coaxial (see §1.11.5):
simple deformations. In that case, U
3 3
& = ∑ λ& N 1 ˆ
U ˆ ⊗N
ˆ and U −1 = ∑ ˆ
Ni ⊗ N (2.6.5)
λi
i i i i
i =1 i =1

Section 2.6
with U& U −1 = U −1 U & R T , that is, any spin is

& and, as expected, from 2.5.25b, w = Ω = R
R
due to rigid body rotation.
ˆ ⊗N
Similarly, from 2.2.37, and differentiating N ˆ =I,
i i
3
& = ∑ λ λ& N
E i i
i =1
i{
ˆ ⊗N
i 2 i
ˆ& ⊗ N
ˆ + 1 λ2 N
i
ˆ + 1 λ2 N
i 2 i
ˆ ⊗N
i
ˆ& .
i } (2.6.6)
ˆ ⋅N &ˆ ˆ
ˆ = δ leads to N ˆ &ˆ
Also, differentiating N i j ij i ⋅ N j = − N i ⋅ N j and so the expression
3
ˆ& = ∑ W N
N ˆ (2.6.7)
i im m
m =1
is valid provided Wij are the components of a skew-symmetric tensor, Wij = −W ji . This
leads to an alternative expression for the Green-Lagrange tensor:
( )
3 3
& = ∑ λ λ& N
E ˆ ⊗Nˆ + ∑ 1 ˆ ⊗N
Wmn λ2m − λ2n N ˆ (2.6.8)
i i i i 2 m n
i =1 m , n =1
m≠ n
Similarly, from 2.2.37, the left Cauchy-Green tensor can be expressed in terms of the
principal spatial base vectors:
{ }
3 3
b = ∑ λi2 nˆ i ⊗ nˆ i , b& = ∑ 2λi λ&i nˆ i ⊗ nˆ i + λi2 n&ˆ i ⊗ nˆ i + λi2 nˆ i ⊗ n&ˆ i (2.6.9)
i =1 i =1
Then, from inspection of 2.5.18c, b& = lb + bl T , the velocity gradient can be expressed as
{▲Problem 2}
3 ⎧ & ⎫ 3 ⎧ λ& ⎫
λ
l = ∑ ⎨ i nˆ i ⊗ nˆ i + nˆ& i ⊗ nˆ i ⎬ = ∑ ⎨ i nˆ i ⊗ nˆ i − nˆ i ⊗ n&ˆ i ⎬ (2.6.7)
i =1 ⎩ λi ⎭ i =1 ⎩ λi ⎭
2.6.3 Rates of Change and the Relative Deformation

Just as the material time derivative of the deformation gradient is defined as
∂ ∂ ⎛ ∂x ⎞
F& = F( X, t ) = ⎜ ⎟
∂t ∂t ⎝ ∂X ⎠
one can define the material time derivative of the relative deformation gradient, cf. §2.3.2,
the rate of change relative to the current configuration:
∂
F& t (x, t ) = Ft (x,τ ) τ =t (2.6.8)
∂τ

Section 2.6
From 2.3.8, Ft (x,τ ) = F( X,τ )F( X, t ) −1 , so taking the derivative with respect to τ (t is
now fixed) and setting τ = t gives
F& t (x, t ) = F& ( X, t )F( X, t ) −1
Then, from 2.5.4,
l = F& t (x, t ) (2.6.9)
as expected – the velocity gradient is the rate of change of deformation relative to the
current configuration. Further, using the polar decomposition,
Ft (x,τ ) = R t (x,τ )U t (x,τ )
Differentiating with respect to τ and setting τ = t then gives
F& t (x, t ) = R t (x, t )U

& (x, t ) + R
t
& (x, t )U (x, t )
t t
Relative to the current configuration, R t (x, t ) = U t (x, t ) = I , so, from 2.4.34,
& (x, t ) + R
l=U & (x, t ) (2.6.10)
t t
With U symmetric and R skew-symmetric, U& (x, t ), R

& (x, t ) are, respectively, symmetric
t t
and skew-symmetric, and it follows that
& ( x, t )
d=U t
(2.6.11)
& ( x, t )
w=R t
again, as expected – the rate of deformation is the instantaneous rate of stretching and the
spin is the instantaneous rate of rotation.
The Corotational Derivative

o
The corotational derivative of a vector a is a ≡ a& − wa . Formally, it is defined through
o 1
a = lim
Δt → 0 Δt
{a(t + Δt ) − R t (t + Δt )a(t )}
1
Δt → 0 Δt
[
= lim {a(t + Δt ) − R t (t ) + ΔtR & (t ) + L) a(t )}
t ]
1
= lim {a(t + Δt ) − [I + Δtw (t ) + L)]a(t )} (2.6.12)
Δt → 0 Δt
1
= lim {a(t + Δt ) − a(t )} − w (t )a(t )
Δt → 0 Δt
= a& − wa

Section 2.6
The definition shows that the corotational derivative involves taking a vector a in the
current configuration and rotating it with the rigid body rotation part of the motion, Fig.
2.6.1. It is this new, rotated, vector which is compared with the vector a(t + Δt ) , which
has undergone rotation and stretch.
a(t + Δt ) = Ft τ =t + Δt a(t )
a(t )
R t τ =t + Δt a(t )
Figure 2.6.1: rotation and stretch of a vector
2.6.4 Rivlin-Ericksen Tensors

The n-th Rivlin-Ericksen tensor is defined as
dn
A n (t ) = C t (τ ) , n = 0, 1, 2, L (2.6.13)
dτ n τ =t
where C t (τ ) is the relative right Cauchy-Green strain. Since Ct (τ )τ = t = I , A 0 = I . To

evaluate the next Rivlin-Ericksen tensor, one needs the derivatives of the relative
deformation gradient; from 2.5.4, 2.3.8,
d
dτ
Ft (τ ) =
d
dτ
[ ]
F(τ )F(t ) −1 = l (τ )F(τ )F(t ) −1 = l (τ )Ft (τ ) (2.6.14)
( T
)
Then, with 2.5.5a, d Ft (τ ) / dτ = Ft (τ ) l (τ ) , and
T T
[ ( )
A 1 (t ) = Ft (τ ) l (τ ) + l (τ ) Ft (τ ) τ =t
T T
]
= (l (t ) + l (t ) )
T
= 2d
Thus the tensor A 1 gives a measure of the rate of stretching of material line elements (see
Eqn. 2.5.10). Similarly, higher Rivlin-Ericksen tensors give a measure of higher order
stretch rates, λ&&, λ
&&& , and so on.

Section 2.6
2.6.5 The Directional Derivative and the Material Time

Derivative
The directional derivative of a function T(t ) in the direction of an increment in t is, by

definition (see, for example, Eqn. 1.15.27),
∂ t T[Δt ] = T(t + Δt ) − T(t ) (2.6.15)
or
dT
∂ t T[Δt ] = Δt (2.6.16)
dt
Setting Δt = 1 , and using the chain rule 1.15.28,
& = ∂ T [1]
T t
= ∂ x T [∂ t x[1]] (2.6.17)
= ∂ x T [v ]
The material time derivative is thus equivalent to the directional derivative in the direction
of the velocity vector.
2.6.6 Problems
1. Derive the relations 2.6.2.
2. Use 2.6.9 to verify 2.5.18, b& = lb + bl T .

Section 2.7
2.7 Small Strain Theory

When the deformation is small, from 2.2.43-4,
F = I + GradU
= I + (gradu )F (2.7.1)
≈ I + gradu
neglecting the product of gradu with GradU , since these are small quantities. Thus one
can take GradU = gradu and there is no distinction to be made between the undeformed
and deformed configurations. The deformation gradient is of the form F = I + α , where
α is small.
2.7.1 Decomposition of Strain

Any second order tensor can be decomposed into its symmetric and antisymmetric part
according to 1.10.28, so that
∂u 1 ⎛⎜ ∂u ⎛ ∂u ⎞ ⎞⎟ 1 ⎛⎜ ∂u ⎛ ∂u ⎞ ⎞⎟
T T
= +⎜ ⎟ + −⎜ ⎟ = ε+Ω
∂x 2 ⎜⎝ ∂x ⎝ ∂x ⎠ ⎟⎠ 2 ⎜⎝ ∂x ⎝ ∂x ⎠ ⎟⎠
(2.7.2)
∂u i 1 ⎛⎜ ∂u i ∂u j ⎞⎟ 1 ⎛⎜ ∂u i ∂u j ⎞⎟
= + + − = ε ij + Ω ij
∂x j 2 ⎜⎝ ∂x j ∂xi ⎟⎠ 2 ⎜⎝ ∂x j ∂xi ⎟⎠
where ε is the small strain tensor 2.2.48 and Ω , the anti-symmetric part of the
displacement gradient, is the small rotation tensor, so that F can be written as
F = I + ε + Ω Small Strain Decomposition of the Deformation Gradient (2.7.3)
It follows that (for the calculation of e, one can use the relation (I + δ ) ≈ I − δ for small
−1
δ)
C = b = I + 2ε
(2.7.4)
E=e=ε
Rotation
Since Ω is antisymmetric, it can be written in terms of an axial vector ω , cf. §1.10.11, so

that for any vector a,
Ωa = ω × a, ω = −Ω 23 e1 + Ω13 e1 − Ω12 e 3 (2.7.5)
The relative displacement can now be written as

Section 2.7
du = (gradu )dX
(2.7.6)
= εdX + ω × dX
The component of relative displacement given by ω × dX is perpendicular to dX , and so

represents a pure rotation of the material line element, Fig. 2.7.1.
du
dx
dX
Figure 2.7.1: a pure rotation
Principal Strains
Since ε is symmetric, it must have three mutually orthogonal eigenvectors, the principal
axes of strain, and three corresponding real eigenvalues, the principal strains,
e1 , e2 , e3 ), which can be positive or negative, cf. §1.11. The effect of ε is therefore to
deform an elemental unit sphere into an elemental ellipsoid, whose axes are the principal
axes, and whose lengths are 1 + e1 , 1 + e2 , 1 + e3 . Material fibres in these principal
directions are stretched only, in which case the deformation is called a pure deformation;
fibres in other directions will be stretched and rotated.
The term εdX in 2.7.6 therefore corresponds to a pure stretch along the principal axes.
The total deformation is the sum of a pure deformation, represented by ε , and a rigid
body rotation, represented by Ω . This result is similar to that obtained for the exact finite
strain theory, but here the decomposition is additive rather than multiplicative. Indeed,
here the corresponding small strain stretch and rotation tensors are U = I + ε and
R = I + Ω , so that
F = RU = I + ε + Ω (2.7.7)
Example
Consider the simple shear (c.f. Eqn. 2.2.40)
x1 = X 1 + kX 2 , x 2 = X 2 , x3 = X 3
where k is small. The displacement vector is u = kx 2 e1 so that
⎡0 k 0 ⎤
grad u = ⎢⎢0 0 0⎥⎥
⎢⎣0 0 0⎥⎦

Section 2.7
The deformation can be written as the additive decomposition
d u = εd X + Ω d X or d u = εd X + ω × d X
with
⎡ 0 k / 2 0⎤ ⎡ 0 k / 2 0⎤
⎢ ⎥ ⎢
ε = ⎢k / 2 0 0⎥, Ω = ⎢− k / 2 0 0⎥⎥
⎢⎣ 0 0 0⎥⎦ ⎢⎣ 0 0 0⎥⎦
and ω = −(k / 2)e 3 . For the rotation component, one can write
⎡ 1 k / 2 0⎤
⎢
R = I + Ω = ⎢− k / 2 1 0⎥⎥
⎢⎣ 0 0 1⎥⎦
which, since for small θ , cosθ ≈ 1, sin θ ≈ θ , can be seen to be a rotation through an
angle θ = − k / 2 (a clockwise rotation).
The principal values of ε are ± k / 2, 0 with corresponding principal directions

n1 = (1 / 2 )e1 + (1 / 2 )e 2 , n 2 = −(1 / 2 )e1 + (1 / 2 )e 2 and n 3 = e 3 .
Thus the simple shear with small displacements consists of a rotation through an angle
k / 2 superimposed upon a pure shear with angle k / 2 , Fig. 2.6.2.
n1
n2
+
θ
=
θ = k /2
Figure 2.6.2: simple shear
2.7.2 Rotations and Small Strain
Consider now a pure rotation about the X 3 axis (within the exact finite strain theory),
dx = RdX , with

Section 2.7
⎡cosθ − sin θ 0⎤
R = ⎢⎢ sin θ cosθ 0⎥⎥ (2.7.8)
⎢⎣ 0 0 1⎥⎦
This rotation does not change the length of line elements dX . According to the small
strain theory, however,
⎡cos θ − 1 0 0⎤ ⎡ 0 − sin θ 0⎤
⎢
ε=⎢ 0 cos θ − 1 0⎥⎥ , Ω = ⎢⎢sin θ 0 0⎥⎥
⎢⎣ 0 0 0⎥⎦ ⎢⎣ 0 0 0⎥⎦
which does predict line element length changes, but which can be neglected if θ is small.
For example, if the rotation is of the order 10 −2 rad , then ε 11 = ε 22 = 10 −4 . However, if
the rotation is large, the errors will be appreciable; in that case, rigid body rotation
introduces geometrical non-linearities which must be dealt with using the finite
deformation theory.
Thus the small strain theory is restricted to not only the case of small displacement
gradients, but also small rigid body rotations.
2.7.3 Volume Change

An elemental cube with edges of unit length in the directions of the principal axes
deforms into a cube with edges of lengths 1 + e1 , 1 + e2 , 1 + e3 , so the unit change in
volume of the cube is
dv − dV
= (1 + e1 )(1 + e2 )(1 + e3 ) − 1 = e1 + e2 + e3 + O(2) (2.7.9)
dV
Since second order quantities have already been neglected in introducing the small strain
tensor, they must be neglected here. Hence the increase in volume per unit volume, called
the dilatation (or dilation) is
δV
= e1 + e2 + e3 = eii = trε = divu Dilatation (2.7.10)
V
Since any elemental volume can be constructed out of an infinite number of such
elemental cubes, this result holds for any elemental volume irrespective of shape.
2.7.4 Rate of Deformation, Strain Rate and Spin Tensors

Take now the expressions 2.4.7 for the rate of deformation and spin tensors. Replacing v
in these expressions by u& , one has

Section 2.7
⎞
1 ⎛⎜ ∂u& i ∂u& j
d=
1
2
(l + lT , ) d ij = +⎟
⎟
2 ⎜⎝ ∂x j ∂xi
⎠
(2.7.11)
1 ⎛ ∂u& ∂u& j ⎞⎟
1
(
w = l − lT ,
2
) wij = ⎜ i −
2 ⎜⎝ ∂x j ∂xi ⎟⎠
For small strains, one can take the time derivative outside (by considering the xi to be
material coordinates independent of time):
d ⎧⎪ 1 ⎛⎜ ∂u i ∂u j ⎞⎟⎫⎪
d ij = ⎨ + ⎬
dt ⎪⎩ 2 ⎜⎝ ∂x j ∂xi ⎟⎠⎪⎭
(2.7.12)
d ⎧⎪ 1 ⎛ ∂u ∂u j ⎞⎟⎫⎪
wij = ⎨ ⎜ i − ⎬
dt ⎪⎩ 2 ⎜⎝ ∂x j ∂xi ⎟⎠⎪⎭
The rate of deformation in this context is seen to be the rate of strain, d = ε& , and the spin
&.
is seen to be the rate of rotation, w = Ω
The instantaneous motion of a material particle can hence be regarded as the sum of three
effects:
(i) a translation given by u& (so in the time interval Δt the particle has been
displaced by u& Δt )
(ii) a pure deformation given by ε&
(iii) a rigid body rotation given by Ω &
2.7.5 Compatibility Conditions
Suppose that the strains ε ij in a body are known. If the displacements are to be
determined, then the strain-displacement partial differential equations
1 ⎛ ∂u ∂u j ⎞
ε ij = ⎜⎜ i + ⎟ (2.7.13)
2 ⎝ ∂x j ∂xi ⎟⎠
need to be integrated. However, there are six independent strain components but only
three displacement components. This implies that the strains are not independent but are
related in some way. The relations between the strains are called compatibility
conditions, and it can be shown that they are given by
ε ij ,km + ε km,ij − ε ik , jm − ε jm ,ik = 0 (2.7.14)
These are 81 equations, but only six of them are distinct, and these six equations are
necessary and sufficient to evaluate the displacement field.

Section 2.9
2.9 Rigid Body Rotations of Configurations

In this section are discussed rigid body rotations to the current and reference
configurations.
2.9.1 A Rigid Body Rotation of the Current Configuration

As mentioned in §2.8.1, the circumstance of two observers, moving relative to each other
and examining a fixed configuration (the current configuration) is equivalent to one
observer taking measurements of two different configurations, moving relative to each
other1. The objectivity requirements of the various kinematic objects discussed in the
previous section can thus also be examined by considering rigid body rotations and
translations of the current configuration.
Any rigid body rotation and translation of the current configuration can be expressed in
the form
x * (X, t ) = Q(t )x(X, t ) + c(t ) (2.9.1)
where Q is a rotation tensor. This is illustrated in Fig. 2.9.5. The current configuration is
denoted by S and the rotated configuration by S * .
Just as dx = FdX , the deformation gradient for the configuration S * relative to the
reference configuration S 0 is defined through dx * = F * dX . From 2.9.1, as in §2.8.5 (see
Eqn. 2.8.23), and similarly for the right and left Cauchy-Green tensors,
F * = QF
C* = F * T F * = C (2.9.2)
b * = F * F * T = QbQ T
Thus in the deformations F : S 0 → S and F * : S 0 → S * , the right Cauchy Green tensors,

C and C* , are the same, but the left Cauchy Green tensors are different, and related
through b * = QbQ T .
All the other results obtained in the last section in the context of observer transformations,
for example for the Jacobian, stretch tensors, etc., hold also for the case of rotations to the
current configuration.
1
Although equivalent, there is a difference: in one, there are two observers who record one event (a material
particle say) as at two different points, in the other there is one observer who records two different events
(the place where the one material particle is in two different configurations)

Section 2.9
S
dx
x dX
S0
X reference
x*
configuration
Q
dx*
S* c
F*
Figure 2.9.1: a rigid body rotation and translation of the current configuration
2.9.2 A Rigid Body Rotation of the Reference Configuration

Consider now a rigid-body rotation to the reference configuration. Such rotations play an
important role in the notion of material symmetry (see Chapter 5).
The reference configuration is denoted by S 0 and the rotated/translated configuration by

S ◊ , Fig. 2.9.2. The deformation gradient for the current configuration S relative to S ◊ is
defined through dx = F ◊ dX ◊ = F ◊ QdX . But dx = FdX and so (and similarly for the
right and left Cauchy-Green tensors)
F ◊ = FQ T
C ◊ = F ◊ T F ◊ = QCQ T (2.9.3)
◊ ◊ ◊T
b =F F =b
Thus the change to the right (left) Cauchy-Green strain tensor under a rotation to the
reference configuration is the same as the change to the left (right) Cauchy-Green strain
tensor under a rotation of the current configuration.

Section 2.9
F◊ QT
S◊
Q
SS
X◊
S0
x
X reference
configuration
Figure 2.9.2: a rigid body rotation of the reference configuration

3 Stress and the
Balance Principles
Three basic laws of physics are discussed in this Chapter:
(1) The Law of Conservation of Mass

(2) The Balance of Linear Momentum
(3) The Balance of Angular Momentum
together with the conservation of mechanical energy and the principle of virtual work,
which are different versions of (2).
(2) and (3) involve the concept of stress, which allows one to describe the action of
forces in materials.
315
316
Section 3.1
3.1 Conservation of Mass
3.1.1 Mass and Density

Mass is a non-negative scalar measure of a body’s tendency to resist a change in motion.
Consider a small volume element Δv whose mass is Δm . Define the average density of
this volume element by the ratio
Δm
ρ AVE = (3.1.1)
Δv
If p is some point within the volume element, then define the spatial mass density at p to
be the limiting value of this ratio as the volume shrinks down to the point,
Δm
ρ (x, t ) = lim Δv→0 Spatial Density (3.1.2)
Δv
In a real material, the incremental volume element Δv must not actually get too small
since then the limit ρ would depend on the atomistic structure of the material; the
volume is only allowed to decrease to some minimum value which contains a large
number of molecules. The spatial mass density is a representative average obtained by
having Δv large compared to the atomic scale, but small compared to a typical length
scale of the problem under consideration.
The density, as with displacement, velocity, and other quantities, is defined for specific
particles of a continuum, and is a continuous function of coordinates and time,
ρ = ρ (x, t ) . However, the mass is not defined this way – one writes for the mass of an
infinitesimal volume of material – a mass element,
dm = ρ (x, t )dv (3.1.3)
or, for the mass of a volume v of material at time t,
m = ∫ ρ (x, t )dv (3.1.4)

v
3.1.2 Conservation of Mass

The law of conservation of mass states that mass can neither be created nor destroyed.
Consider a collection of matter located somewhere in space. This quantity of matter with
well-defined boundaries is termed a system. The law of conservation of mass then
implies that the mass of this given system remains constant,

Section 3.1
Dm
=0 Conservation of Mass (3.1.5)
Dt
The volume occupied by the matter may be changing and the density of the matter within
the system may be changing, but the mass remains constant.
Considering a differential mass element at position X in the reference configuration and

at x in the current configuration, Eqn. 3.1.5 can be rewritten as
dm( X) = dm(x, t ) (3.1.6)
The conservation of mass equation can be expressed in terms of densities. First,

introduce ρ 0 , the reference mass density (or simply the density), defined through
Δm
ρ 0 ( X) = lim ΔV →0 Density (3.1.7)
ΔV
Note that the density ρ 0 and the spatial mass density ρ are not the same quantities1.
Thus the local (or differential) form of the conservation of mass can be expressed as (see
Fig. 3.1.1)
dm = ρ 0 ( X)dV = ρ (x, t )dv = const (3.1.8)
• x
•X
dV , ρ 0 dv, ρ
reference current
configuration configuration
Figure 3.1.1: Conservation of Mass for a deforming mass element
Integration over a finite region of material gives the global (or integral) form,
m = ∫ ρ 0 ( X)dV = ∫ ρ (x, t )dv = const (3.1.9)

V v
or
dm d
dt dt ∫v
m& = = ρ (x, t )dv = 0 (3.1.10)
1
they not only are functions of different variables, but also have different values; they are not different
representations of the same thing, as were, for example, the velocities v and V. One could introduce a
material mass density, Ρ ( X , t ) = ρ ( x ( X , t ), t ) , but such a quantity is not useful in analysis

Section 3.1
3.1.3 Control Mass and Control Volume

A control mass is a fixed mass of material whose volume and density may change, and
which may move through space, Fig. 3.1.2. There is no mass transport through the
moving surface of the control mass. For such a system, Eqn. 3.1.10 holds.
m, ρ ( x(t1 ), t1 ), v (t1 ) m, ρ ( x(t 2 ), t 2 ), v (t 2 )
Figure 3.1.2: Control Mass
By definition, the derivative in 3.1.10 is the time derivative of a property (in this case
mass) of a collection of material particles as they move through space, and when they
instantaneously occupy the volume v, Fig. 3.1.3, or
d 1 ⎪⎧ ⎫⎪
∫
dt dv
ρ dv = lim Δt →0 ⎨ ∫ ρ (x, t + Δt ) dv − ∫ ρ (x, t )dv ⎬ = 0
Δt ⎪⎩v ( t + Δt ) v (t ) ⎪⎭
(3.1.11)
Alternatively, one can take the material derivative inside the integral sign:
= ∫ [ρ (x, t )dv ] = 0
dm d
(3.1.12)
dt v dt
This is now equivalent to the sum of the rates of change of mass of the mass elements
occupying the volume v.
time t
time t + Δt
Figure 3.1.3: Control Mass occupying different volumes at different times
A control volume, on the other hand, is a fixed volume (region) of space through which
material may flow, Fig. 3.1.4, and for which the mass may change. For such a system,
one has

Section 3.1
∂m ∂ ∂
= ∫ ρ (x, t )dv = ∫ [ρ (x, t )]dv ≠ 0 (3.1.13)
∂t ∂t v v
∂t
m(t ), ρ ( x, t ), v
Figure 3.1.4: Control Volume
3.1.4 The Continuity Equation (Spatial Form)

A consequence of the law of conservation of mass is the continuity equation, which (in
the spatial form) relates the density and velocity of any material particle during motion.
This equation can be derived in a number of ways:
Derivation of the Continuity Equation using a Control Volume (Global Form)
The continuity equation can be derived directly by considering a control volume - this is
the derivation appropriate to fluid mechanics. Mass inside this fixed volume cannot be
created or destroyed, so that the rate of increase of mass in the volume must equal the rate
at which mass is flowing into the volume through its bounding surface.
The rate of increase of mass inside the fixed volume v is
∂m ∂ ∂ρ
= ∫ ρ (x, t ) dv = ∫ dv (3.1.14)
∂t ∂t v v
∂t
The mass flux (rate of flow of mass) out through the surface is given by Eqn. 1.7.9,
∫ ρ v ⋅ nds,
s
∫ ρ v n ds
s
i i
where n is the unit outward normal to the surface and v is the velocity. It follows that
∂ρ ∂ρ
∫ ∂t dv + ∫ ρ v ⋅ nds = 0,
v s
∫ ∂t dv + ∫ ρ v n ds = 0
v s
i i (3.1.15)
Use of the divergence theorem 1.7.12 leads to
⎡ ∂ρ ⎤ ⎡ ∂ρ ∂ ( ρ vi ) ⎤
∫ ⎢⎣ ∂t + div( ρ v )⎥ dv = 0, ∫ ⎢⎣ ∂t + ⎥ dv = 0 (3.1.16)
v ⎦ v
∂xi ⎦

Section 3.1
leading to the continuity equation,
∂ρ ∂ρ ∂ ( ρvi )
+ div( ρ v ) = 0 + =0
∂t ∂t ∂xi
dρ dρ ∂v
+ ρ div v = 0 +ρ i =0 Continuity Equation
dt dt ∂xi
∂ρ ∂ρ ∂ρ ∂v
+ gradρ ⋅ v + ρ div v = 0 + vi + ρ i = 0
∂t ∂t ∂xi ∂xi
(3.1.17)
This is (these are) the continuity equation in spatial form. The second and third forms of
the equation are obtained by re-writing the local derivative in terms of the material
derivative 2.4.7 (see also 1.6.23b).
If the material is incompressible, so the density remains constant in the neighbourhood of

a particle as it moves, then the continuity equation reduces to
∂vi
div v = 0, = 0 Continuity Eqn. for Incompressible Material (3.1.18)
∂xi
Derivation of the Continuity Equation using a Control Mass
Here follow two ways to derive the continuity equation using a control mass.
1. Derivation using the Formal Definition
From 3.1.11, adding and subtracting a term:
d 1 ⎧⎪⎡ ⎤
∫
dt dv
ρ dv = lim
Δt →0 Δt
⎨ ⎢ ∫
⎪⎩⎢⎣v ( t + Δt )
ρ (x, t + Δt ) dv − ∫ ρ (x, t + Δt ) dv ⎥
v (t ) ⎥⎦
(3.1.19)
⎡ ⎤ ⎫⎪
+ ⎢ ∫ ρ (x, t + Δt ) dv − ∫ ρ (x, t ) dv ⎥ ⎬
⎢⎣v ( t ) v (t ) ⎥⎦ ⎪⎭
The terms in the second square bracket correspond to holding the volume v fixed and
evidently equals the local rate of change:
d ∂ρ 1
∫
dt dv
ρ dv = ∫ dv + lim
v
∂t Δt → 0 Δt v ( t + Δ∫t ) −v ( t )
ρ (x, t + Δt )dv (3.1.20)
The region v(t + Δt ) − v(t ) is swept out in time Δt . Superimposing the volumes v(t ) and
v(t + Δt ) , Fig. 3.1.5, it can be seen that a small element Δv of v(t + Δt ) − v(t ) is given by
(see the example associated with Fig. 1.7.7)
Δv = Δtv ⋅ nΔs (3.1.21)

Section 3.1
where s is the surface. Thus
1 1
lim
Δt →0 Δt ∫ ρ (x, t + Δt )dv = lim ∫ Δtρ (x, t + Δt ) v ⋅ nds = ∫ ρ (x, t ) v ⋅ nds (3.1.22)
v ( t + Δt ) − v ( t )
Δt →0 Δt
s s
and 3.1.15 is again obtained, from which the continuity equation results from use of the
divergence theorem.
v (t + Δ t )
v (t )
s (t )
s (t + Δ t )
s v
Δs n
Δv
Figure 3.1.5: Evaluation of Eqn. 3.1.22
2. Derivation by Converting to Mass Elements
This derivation requires the kinematic relation for the material time derivative of a
volume element, 2.5.23: d (dv) / dt = divv dv . One has
⎛ .
⎞
= ∫ ρ (x, t ) dv = ∫ ( ρ dv ) = ∫ ⎜⎜ ρ& dv + ρ dv ⎟⎟ = ∫ ( ρ& + div vρ )dv ≡ 0
dm d d
(3.1.23)
v⎝ ⎠ v
dt dt v v
dt
The continuity equation then follows, since this must hold for any arbitrary region of the
volume v.
Derivation of the Continuity Equation using a Control Volume (Local Form)
The continuity equation can also be derived using a differential control volume element.
This calculation is similar to that given in §1.6.6, with the velocity v replaced by ρv .
3.1.5 The Continuity Equation (Material Form)
From 3.1.9, and using 2.2.53, dv = JdV ,
∫ [ρ
V
0 ( X) − ρ (χ ( X, t ), t ) J ( X, t )]dV = 0 (3.1.24)

Section 3.1
Since V is an arbitrary region, the integrand must vanish everywhere, so that
ρ 0 ( X) = ρ (χ ( X, t ), t ) J ( X, t ) Continuity Equation (Material Form) (3.1.25)
This is known as the continuity (mass) equation in the material description. Since
ρ& 0 = 0 , the rate form of this equation is simply
d
( ρJ ) = 0 (3.1.26)
dt
The material form of the continuity equation, ρ 0 = ρ J , is an algebraic equation, unlike

the partial differential equation in the spatial form. However, the two must be equivalent,
and indeed the spatial form can be derived directly from this material form: using 2.5.20,
dJ / dt = Jdiv v ,
d
( ρJ ) = ρ& J + ρ J&
dt (3.1.27)
= J ( ρ& + ρ div v )
This is zero, and J > 0 , and the spatial continuity equation follows.
Example (of Conservation of Mass)
Consider a bar of material of length l 0 , with density in the undeformed configuration ρ 0

and spatial mass density ρ ( x, t ) , undergoing the 1-D motion X = x /(1 + At ) ,
x = X + AtX . The volume ratio (taking unit cross-sectional area) is J = 1 + At . The
continuity equation in the material form 3.1.25 specifies that
ρ 0 = ρ (1 + At )
Suppose now that
2m
ρ o ( X) = X
l 02
lo
so that the total mass of the bar is ∫
0
ρ 0 ( X)dX = m . It follows that the spatial mass
density is
ρ0 2m X 2m x
ρ= = = 2
(1 + At ) l 0 1 + At l 0 (1 + At ) 2
2
Evaluating the total mass of the bar at time t leads to
lo (1+ At ) 2m 1 lo (1+ At )
∫
0
ρ (x, t ) dx =
l 0 (1 + At ) 2
2 ∫
0
x dx

Section 3.1
which is again m, as required.
end of bar ( x = l 0 (1 + At ), X = l 0 )
end of bar ( x = X = l 0 )
at time t
at t = 0
Figure 3.1.6: a stretching bar
The density could have been derived from the equation of continuity in the spatial form:
since the velocity is
dx( X, t ) Ax
V ( X, t ) = = AX, v(x, t ) = V (χ −1 (x, t ), t ) =
dt 1 + At
one has
∂ρ ∂ρ ∂v ∂ρ Ax ∂ρ A
+v +ρ = + +ρ =0
∂t ∂x ∂x ∂t 1 + At ∂x 1 + At
Without attempting to solve this first order partial differential equation, it can be seen by
substitution that the value for ρ obtained previously satisfies the equation.
■
3.1.6 Material Derivatives of Integrals
Reynold’s Transport Theorem
In the above, the material derivative of the total mass carried by a control mass,
d
dt ∫v
ρ (x, t ) dv ,
was considered. It is quite often that one needs to evaluate material time derivatives of
similar volume (and line and surface) integrals, involving other properties, for example
momentum or energy. Thus, suppose that A(x, t ) is the distribution of some property
(per unit volume) throughout a volume v (A is taken to be a second order tensor, but what
follows applies also to vectors and scalars). Then the rate of change of the total amount
of the property carried by the mass system is

Section 3.1
d
dt ∫v
A(x, t )dv
Again, this integral can be evaluated in a number of ways. For example, one could
evaluate it using the formal definition of the material derivative, as done above for
A = ρ . Alternatively, one can evaluate it using the relation 2.5.23, d (dv) / dt = divv dv ,
through
⎡& ⎤
[ ]
.
d
∫ A ( x , t ) dv = ∫
d
[A ( x , t ) dv ] = ∫ ⎢ Adv + A dv ⎥ = ∫ A + divv A dv
& (3.1.28)
v ⎣ ⎦ v
dt v v
dt
Thus one arrives at Reynold’s transport theorem
⎧ ⎡ dA ⎤ ⎡ dAij ∂v k ⎤
⎪∫ ⎢ + divv A ⎥ dv ∫ ⎢⎣ + Aij ⎥ dv
⎪ v ⎣ dt ⎦ v
dt ∂x k ⎦
⎪ ⎡ ∂A ⎤ ⎡ ∂Aij ∂Aij ∂v k ⎤
⎪∫ ⎢ + grad A ⋅ v + divv A ⎥ dv ∫v ⎢⎣ ∂t + ∂xk vk + ∂xk Aij ⎥⎦ dv
⎪ ∂t
A(x, t ) dv = ⎨ v ⎣ ⎦
d
∫ ⎡ ∂Aij ∂ (Aij v k )⎤
dt v ⎪ ⎡ ∂A + div(A ⊗ v )⎤ dv
⎪∫ ⎢⎣ ∂t ⎥⎦ ∫v ⎢⎣ ∂t + ∂xk ⎥⎦ dv
⎪v
⎪ ∂A dv + A(v ⋅ n )ds ∂Aij
⎪⎩∫v ∂t ∫s ∫v ∂t dv + ∫s Aij vk nk ds
Reynold’s Transport Theorem (3.1.29)
The index notation is shown for the case when A is a second order tensor. In the last of
these forms2 (obtained by application of the divergence theorem), the first term represents
the amount (of A) created within the volume v whereas the second term (the flux term)
represents the (volume) rate of flow of the property through the surface. In the last three
versions, Reynold’s transport theorem gives the material derivative of the moving control
mass in terms of the derivative of the instantaneous fixed volume in space (the first term).
Of course when A = ρ , the continuity equation is recovered.
Another way to derive this result is to first convert to the reference configuration, so that
integration and differentiation commute (since dV is independent of time):
A(x, t ) dv = ∫ A( X, t ) JdV = ∫ (A( X, t ) J )dV

d d d
∫
dt v dt V V
dt
=∫ A ( ) (
∫ & + divv A JdV
& J + A J& dV = A ) (3.1.30)
V V
=∫ A (
& (x, t ) + divv A(x, t ) dv )
v
2
also known as the Leibniz formula

Section 3.1
Reynold’s Transport Theorem for Specific Properties
A property that is given per unit mass is called a specific property. For example,
specific heat is the heat per unit mass. Consider then a property B, a scalar, vector or
tensor, which is defined per unit mass through a volume. Then the rate of change of the
total amount of the property carried by the mass system is simply
dB dB
ρ B(x, t ) dv = ∫ [Bρ dv ] = ∫ [B dm] = ∫ dm = ∫ ρ
d d d
∫
dt v v
dt v
dt v
dt v
dt
dv (3.1.31)
Material Derivatives of Line and Surface Integrals
Material derivatives of line and surface integrals can also be evaluated. From 2.5.8,
d (dx) / dt = ldx ,
d
dt ∫
A(x, t ) dx = ∫ A [
& + Al dx ] (3.1.32)
( )
and, using 2.5.22, d (nˆ ds ) / dt = divv − l T nˆ ds ,
d
dt ∫s
A(x, t )nˆ ds = ∫ A[ ( )]
& + A divv − l T nˆ ds (3.1.33)
s
3.1.7 Problems
1. A motion is given by the equations
x1 = X 1 + 3 X 2 t , x 2 = − X 1t 2 + X 2 (t + 1), x3 = X 3
(a) Calculate the spatial mass density ρ in terms of the density ρ 0
(b) Derive a first order ordinary differential equation for the density ρ (in terms of
x and t only) assuming that it is independent of position x

Section 3.2
3.2 The Momentum Principles

In Parts I and II, the basic dynamics principles used were Newton’s Laws, and these are
equivalent to force equilibrium and moment equilibrium. For example, they were used to
derive the stress transformation equations in Part I, §3.4 and the Equations of Motion in
Part II, §1.1. Newton’s laws there were applied to differential material elements.
An alternative but completely equivalent set of dynamics laws are Euler’s Laws; these
are more appropriate for finite-sized collections of moving particles, and can be used to
express the force and moment equilibrium in terms of integrals. Euler’s Laws are also
called the Momentum Principles: the principle of linear momentum (Euler’s first law)
and the principle of angular momentum (Euler’s second law).
3.2.1 The Principle of Linear Momentum

Momentum is a measure of the tendency of an object to keep moving once it is set in
motion. Consider first the particle of rigid body dynamics: the (linear) momentum p is
defined to be its mass times velocity, p = mv . The rate of change of momentum p& is
dp d (mv ) dv
= =m = ma (3.2.1)
dt dt dt
and use has been made of the fact that dm / dt = 0 . Thus Newton’s second law, F = ma ,
can be rewritten as
d
F= (mv) (3.2.2)
dt
This equation, formulated by Euler, states that the rate of change of momentum is equal to
the applied force. It is called the principle of linear momentum, or balance of linear
momentum. If there are no forces applied to a system, the total momentum of the system
remains constant; the law in this case is known as the law of conservation of (linear)
momentum.
Eqn. 3.2.2 as applied to a particle can be generalized to the mechanics of a continuum in

one of two ways. One could consider a differential element of material, of mass dm and
velocity v. Alternatively, one can consider a finite portion of material, a control mass in
the current configuration with spatial mass density ρ (x, t ) and spatial velocity field
v (x, t ) . The total linear momentum of this mass of material is
L(t ) = ∫ ρ (x, t ) v(x, t )dv Linear Momentum (3.2.3)

v
The principle of linear momentum states that
L& (t ) = ∫ ρ (x, t ) v(x, t )dv = F(t )

d
(3.2.4)
dt v

Section 3.2
where F(t ) is the resultant of the forces acting on the portion of material.
Note that the volume over which the integration in Eqn. 3.2.4 takes place is not fixed; the
integral is taken over a fixed portion of material particles, and the space occupied by this
matter may change over time.
By virtue of the Transport theorem relation 3.1.31, this can be written as
L& (t ) = ∫ ρ (x, t ) v& (x, t )dv = F(t ) (3.2.5)

v
The resultant force acting on a body is due to the surface tractions t acting over surface
elements and body forces b acting on volume elements, Fig. 3.2.1:
F(t ) = ∫ t ds + ∫ b dv, Fi = ∫ t i ds + ∫ bi dv Resultant Force (3.2.6)

s v s v
and so the principle of linear momentum can be expressed as
∫ t ds + ∫ bdv = ∫ ρv& dv
s v v
Principle of Linear Momentum (3.2.7)
n t
ds
dv
b
Figure 3.2.1: surface and body forces acting on a finite volume of material
The principle of linear momentum, Eqns. 3.2.7, will be used to prove Cauchy’s Lemma
and Cauchy’s Law in the next section and, in §3.6, to derive the Equations of Motion.
3.2.2 The Principle of Angular Momentum

Considering again the mechanics of a single particle: the angular momentum is the
moment of momentum about an axis, in other words, it is the product of the linear
momentum of the particle and the perpendicular distance from the axis of its line of
action. In the notation of Fig. 3.2.2, the angular momentum h is
h = r × mv (3.2.8)
which is the vector with magnitude d × m v and perpendicular to the plane shown.

Section 3.2
o
• r
mv
Figure 3.2.2: surface and body forces acting on a finite volume of material
Consider now a collection of particles. The principle of angular momentum states that
the resultant moment of the external forces acting on the system of particles, M , equals
the rate of change of the total angular momentum of the particles:
dh
M = r×F = (3.2.9)
dt
Generalising to a continuum, the angular momentum is
H = ∫ r × ρvdv Angular Momentum (3.2.10)

v
and the principle of angular momentum is
d
∫r ×t ds + ∫ r × bdv =
dt ∫v
r × ρvdv
(n )
s v
Principle of Angular Momentum

d
∫ε
s
ijk x j t k( n ) ds + ∫ ε ijk x j bk dv =
v
dt V∫
ε ijk x j ρv k dv
(3.2.11)
The principle of angular momentum, 3.2.11, will be used, in §3.6, to deduce the
symmetry of the Cauchy stress.

Section 3.3
3.3 The Cauchy Stress Tensor
3.3.1 The Traction Vector

The traction vector was introduced in Part I, §3.3. To recall, it is the limiting value of
the ratio of force over area; for Force ΔF acting on a surface element of area ΔS , it is
ΔF
t ( n ) = lim (3.3.1)
ΔS → 0 ΔS
and n denotes the normal to the surface element. An infinite number of traction vectors
act at a point, each acting on different surfaces through the point, defined by different
normals.
3.3.2 Cauchy’s Lemma

Cauchy’s lemma states that traction vectors acting on opposite sides of a surface are
equal and opposite1. This can be expressed in vector form:
t ( n ) = −t ( − n ) Cauchy’s Lemma (3.3.2)
This can be proved by applying the principle of linear momentum to a collection of

particles of mass Δm instantaneously occupying a small box with parallel surfaces of
area Δs , thickness δ and volume Δv = δΔs , Fig. 3.3.1. The resultant surface force
acting on this matter is t ( n ) Δs + t ( − n ) Δs .
t (n )
n
Δs
thickness δ
−n
t ( −n )
Figure 3.3.1: traction acting on a small portion of material particles
The total linear momentum of the matter is ∫ΔV

ρ v dv = ∫ v dm . By the mean value
Δm
theorem (see Appendix A to Chapter 1, §1.B.1), this equals vΔm , where v is the velocity
at some interior point. Similarly, the body force acting on the matter is ∫ b dv = b Δv ,
ΔV
where b is the body force (per unit volume) acting at some interior point. The total mass
1
this is equivalent to Newton’s (third) law of action and reaction – it seems like a lot of work to prove this
seemingly obvious result but, to be consistent, it is supposed that the only fundamental dynamic laws
available here are the principles of linear and angular momentum, and not any of Newton’s laws

Section 3.3
can also be written as Δm = ∫ ρ dv = ρ Δv . From the principle of linear momentum,

ΔV
Eqn. 3.2.7, and since Δm does not change with time,
t ( n ) Δs + t ( − n ) Δs + b Δv =
d
[vΔm] = Δm dv = ρ Δv dv = ρ δΔs dv (3.3.3)
dt dt dt dt
Dividing through by Δs and taking the limit as δ → 0 , one finds that t ( n ) = −t ( − n ) .

Note that the values of t ( n ) , t ( − n ) acting on the box with finite thickness are not the same
as the final values, but approach the final values at the surface as δ → 0 .
3.3.3 Stress
In Part I, the components of the traction vector were called stress components, and it was
illustrated how there were nine stress components associated with each material particle.
Here, the stress is defined more formally,
Cauchy’s Law
Cauchy’s Law states that there exists a Cauchy stress tensor σ which maps the normal
to a surface to the traction vector acting on that surface, according to
t = σ n, t i = σ ij n j Cauchy’s Law (3.3.4)
or, in full,
t1 = σ 11 n1 + σ 12 n2 + σ 13 n3
t 2 = σ 21 n1 + σ 22 n2 + σ 23 n3 (3.3.5)
t 3 = σ 31 n1 + σ 32 n2 + σ 33 n3
Note:
• many authors define the stress tensor as t = n σ . This amounts to the definition used here
since, as mentioned in Part I, and as will be (re-)proved below, the stress tensor is symmetric,
σ = σ T , σ ij = σ ji
• the Cauchy stress refers to the current configuration, that is, it is a measure of force per unit
area acting on a surface in the current configuration.
Stress Components
Taking Cauchy’s law to be true (it is proved below), the components of the stress tensor
with respect to a Cartesian coordinate system are, from 1.9.4 and 3.3.4,
(e j )
σ ij = e i σ e j = e i ⋅ t (3.3.6)
which is the ith component of the traction vector acting on a surface with normal e j .
Note that this definition is inconsistent with that given in Part I, §3.2 – there, the first

Section 3.3
subscript denoted the direction of the normal – but, again, the two definitions are
equivalent because of the symmetry of the stress tensor.
The three traction vectors acting on the surface elements whose outward normals point in
the directions of the three base vectors e j are
t (e1 ) = σ 11e1 + σ 21e 2 + σ 31e 3

t (e2 ) = σ 12e1 + σ 22e 2 + σ 32e 3
(e j )
t = σe j , (3.3.7)
(e 3 )
t = σ 13e1 + σ 23e 2 + σ 33e 3
Eqns. 3.3.6-7 are illustrated in Fig. 3.3.2.
x3 x3
e3
σ 33
(e3 )
t σ 13 σ 23
(e 2 )
t σ 32
σ 31
e2
t (e1 ) σ 22
e1 x2 σ 11 σ 21 σ 12
x2
x1
x1
(a) (b)
Figure 3.3.2: traction acting on surfaces with normals in the coordinate directions;
(a) traction vectors, (b) stress components
Proof of Cauchy’s Law
The proof of Cauchy’s law essentially follows the same method as used in the proof of
Cauchy’s lemma.
Consider a small tetrahedral free-body, with vertex at the origin, Fig. 3.3.3. It is required
to determine the traction t in terms of the nine stress components (which are all shown
positive in the diagram).
Let the area of the base of the tetrahedran, with normal n, be Δs . The area ds1 is then
Δs cos α , where α is the angle between the planes, as shown in Fig. 3.3.3b; this angle is
the same as that between the vectors n and e1 , so Δs1 = (n ⋅ e1 )Δs = n1 Δs , and similarly
for the other surfaces: Δs 2 = n2 Δs and Δs3 = n3 Δs .

Section 3.3
x3
t (n )
σ 11 Δs1
σ 12 σ 21
n
σ 22 •
n α •
Δs2
σ 31 e1
σ 32 x2
σ 13
σ 23
σ 33 Δs3
x1
(a) (b)
Figure 3.3.3: free body diagram of a tetrahedral portion of material; (a) traction
acting on the material, (b) relationship between surface areas and normal
components
The resultant surface force on the body, acting in the x1 direction, is
t1 Δs − σ 11 n1 Δs − σ 12 n2 Δs − σ 13 n3 Δs
Again, the momentum is vΔM , the body force is b Δv and the mass is
Δm = ρ Δv = ρ (h / 3)Δs , where h is the perpendicular distance from the origin (vertex) to
the base. The principle of linear momentum then states that
dv1
t1 Δs − σ 11 n1 Δs − σ 12 n 2 Δs − σ 13 n3 Δs + b1 (h / 3)Δs = ρ (h / 3)Δs
dt
Again, the values of the traction and stress components on the faces will in general vary
over the faces, so the values used in this equation are average values over the faces.
Dividing through by Δs , and taking the limit as h → 0 , one finds that
t1 = σ 11 n1 + σ 12 n2 + σ 13 n3
and now these quantities, t1 , σ 11 , σ 12 , σ 13 , are the values at the origin. The equations for
the other two traction components can be derived in a similar way.
Normal and Shear Stress
The stress acting normal to a surface is given by
σ N = n ⋅ t (n ) (3.3.8)
The shear stress acting on the surface can then be obtained from

Section 3.3
2
σS = t ( n ) − σ N2 (3.3.9)
Example
The state of stress at a point is given in the matrix form
⎡2 1 3⎤
[σ ]ij = ⎢⎢1 2 − 2⎥⎥
⎢⎣3 − 2 1 ⎥⎦
Determine
(a) the traction vector acting on a plane through the point whose unit normal is
nˆ = (1 / 3)eˆ 1 + (2 / 3)eˆ 2 − (2 / 3)eˆ 3
(b) the component of this traction acting perpendicular to the plane
(c) the shear component of traction.
Solution
(a) The traction is
⎡t1( nˆ ) ⎤ ⎡σ 11 σ 12 σ 13 ⎤ ⎡ n1 ⎤ ⎡2 1 3 ⎤⎡ 1 ⎤ ⎡− 2⎤
⎢ ( nˆ ) ⎥ ⎢ ⎥ ⎢ ⎥ 1⎢ ⎥ ⎢ ⎥ 1⎢ ⎥
⎢t 2 ⎥ = ⎢σ 21 σ 22 σ 23 ⎥ ⎢n2 ⎥ = ⎢1 2 − 2⎥ ⎢ 2 ⎥ = ⎢ 9 ⎥
3 3
⎢t 3( nˆ ) ⎥ ⎢⎣σ 31 σ 32 σ 33 ⎥⎦ ⎢⎣ n3 ⎥⎦ ⎢⎣3 − 2 1 ⎥⎦ ⎢⎣− 2⎥⎦ ⎢⎣ − 3⎥⎦
⎣ ⎦
or t (nˆ ) = (−2 / 3)eˆ 1 + 3eˆ 2 − eˆ 3 .
(b) The component normal to the plane is the projection of t (nˆ ) in the direction of n̂ , i.e.
σ N = t (nˆ ) ⋅ nˆ = (−2 / 3)(1 / 3) + 3(2 / 3) + (2 / 3) = 22 / 9 ≈ 2.4.
(c) The shearing component of traction is
σ S = t (nˆ ) − (22 / 9)nˆ

= [[(−2 / 3) − (22 / 27)]eˆ 1 + [3 − (44 / 27)]eˆ 2 + [− 1 + (44 / 27)]eˆ 3 ]
= [(−40 / 27)eˆ 1 + (37 / 27)eˆ 2 + (17 / 27)eˆ 3 ]
i.e. of magnitude (−40 / 27) 2 + (37 / 27) 2 + (17 / 27) 2 ≈ 2.1 , which equals
2
tˆ ( nˆ ) − σ N2 .
■

Section 3.4
3.4 Properties of the Stress Tensor
3.4.1 Stress Transformation

Let the components of the Cauchy stress tensor in a coordinate system with base vectors
e i be σ ij . The components in a second coordinate system with base vectors e′j , σ ij′ , are
given by the tensor transformation rule 1.10.5:
σ ij′ = Q pi Qqjσ pq (3.4.1)
where Qij are the direction cosines, Qij = e i ⋅ e′j .
Isotropic State of Stress
Suppose the state of stress in a body is
⎡σ 0 0 0⎤
[σ ] = ⎢⎢ 0 σ0 0 ⎥⎥
⎢⎣ 0 0 σ 0 ⎥⎦
One finds that the application of the tensor transformation rule yields the very same
components no matter what the coordinate system. This is termed an isotropic state of
stress, or a spherical state of stress (see §1.13.3). One example of isotropic stress is the
stress arising in fluid at rest, which cannot support shear stress, in which case
σ = − pI (3.4.2)
where the scalar p is the fluid hydrostatic pressure. For this reason, an isotropic state of
stress is also referred to as a hydrostatic state of stress.
A note on the Transformation Formula
Using the vector transformation rule 1.5.5, the traction and normal transform according to
[ ] [ ]
[t ′] = Q T [t ], [n′] = Q T [n] . Also, Cauchy’s law transforms according to [t ′] = [σ ′][n′]
[ ] [ ]
which can be written as Q T [t ] = [σ ′] Q T [n] , so that, pre-multiplying by [Q ] , and since
{ [ ]} [ ]
[Q] is orthogonal, [t ] = [Q][σ ′] Q T [n] , so [σ] = [Q][σ ′] Q T , which is the inverse tensor
transformation rule 1.13.6a, showing the internal consistency of the theory.
In Part I, Newton’s law was applied to a material element to derive the two-dimensional
stress transformation equations, Eqn. 3.4.7 of Part I. Cauchy’s law was proved in a
similar way, using the principle of momentum. In fact, Cauchy’s law and the stress
transformation equations are equivalent. Given the stress components in one coordinate
system, the stress transformation equations give the components in a new coordinate
system; particularising this, they give the stress components, and thus the traction vector,

Section 3.4
acting on new surfaces, oriented in some way with respect to the original axes, which is
what Cauchy’s law does.
3.4.2 Principal Stresses
Since the stress σ is a symmetric tensor, it has three real eigenvalues σ 1 , σ 2 , σ 3 , called
principal stresses, and three corresponding orthonormal eigenvectors called principal
directions. The eigenvalue problem can be written as
t (n ) = σ n = σ n (3.4.3)
where n is a principal direction and σ is a scalar principal stress. Since the traction
vector is a multiple of the unit normal, σ is a normal stress component. Thus a principal
stress is a stress which acts on a plane of zero shear stress, Fig. 3.4.1.
t ( n ) = t1 e 1 + t 2 e 2 + t 3 e 3
no shear stress – only a normal

component to the traction
Figure 3.4.1: traction acting on a plane of zero shear stress
The principal stresses are the roots of the characteristic equation 1.11.5,
σ 3 − I 1σ 2 + I 2σ − I 3 = 0 (3.4.4)
where, Eqn. 1.11.6-7, 1.11.17,
I 1 = trσ
= σ 11 + σ 22 + σ 33
= σ1 + σ 2 + σ 3
I2 = 1
2
[(trσ ) 2
− trσ 2 ]
= σ 11σ 22 + σ 22σ 33 + σ 33σ 11 − σ 122 − σ 23
2
− σ 312
(3.4.5)
= σ 1σ 2 + σ 2σ 3 + σ 3σ 1
I3 = 1
3
[trσ 3
− 32 trσtrσ 2 + 12 (trσ )
3
]
= det σ
= σ 11σ 22σ 33 − σ 11σ 23
2
− σ 22σ 312 − σ 33σ 122 + 2σ 12σ 23σ 32
= σ 1σ 2σ 3

Section 3.4
The principal stresses and principal directions are properties of the stress tensor, and do
not depend on the particular axes chosen to describe the state of stress., and the stress
invariants I 1 , I 2 , I 3 are invariant under coordinate transformation. c.f. §1.11.1.
If one chooses a coordinate system to coincide with the three eigenvectors, one has the
spectral decomposition 1.11.11 and the stress matrix takes the simple form 1.11.12,
3
⎡σ 1 0 0⎤
σ = ∑ σ i nˆ i ⊗ nˆ i , [σ ] = ⎢⎢ 0 σ 2 0 ⎥⎥ (3.4.6)
i =1
⎢⎣ 0 0 σ 3 ⎥⎦
Note that when two of the principal stresses are equal, one of the principal directions will
be unique, but the other two will be arbitrary – one can choose any two principal
directions in the plane perpendicular to the uniquely determined direction, so that the
three form an orthonormal set. This stress state is called axi-symmetric. When all three
principal stresses are equal, one has an isotropic state of stress, and all directions are
principal directions.
3.4.3 Maximum Stresses

Directly from §1.11.3, the three principal stresses include the maximum and minimum
normal stress components acting at a point. This result is re-derived here, together with
results for the maximum shear stress
Normal Stresses
Let e1 , e 2 , e 3 be unit vectors in the principal directions and consider an arbitrary unit
normal vector n = n1e1 + n2 e 2 + n3e 3 , Fig. 3.4.2. From 3.3.8 and Cauchy’s law, the
normal stress acting on the plane with normal n is
σ N = t (n ) ⋅ n = (σ n ) ⋅ n (3.4.7)
σN
t (n )
principal
1 directions
Figure 3.4.2: normal stress acting on a plane defined by the unit normal n

Section 3.4
With respect to the principal stresses, using 3.4.6,
t ( n ) = σ n = σ 1 n1e1 + σ 2 n2 e 2 + σ 3 n3 e 3 (3.4.8)
and the normal stress is
σ N = σ 1 n12 + σ 2 n22 + σ 3 n32 (3.4.9)
Since n12 + n 22 + n32 = 1 and, without loss of generality, taking σ 1 ≥ σ 2 ≥ σ 3 , one has
σ 1 = σ 1 (n12 + n22 + n32 ) ≥ σ 1 n12 + σ 2 n22 + σ 3 n32 = σ N (3.4.10)
Similarly,
σ N = σ 1 n12 + σ 2 n22 + σ 3 n32 ≥ σ 3 (n12 + n22 + n32 ) ≥ σ 3 (3.4.11)
Thus the maximum normal stress acting at a point is the maximum principal stress and
the minimum normal stress acting at a point is the minimum principal stress.
Shear Stresses
Next, it will be shown that the maximum shearing stresses at a point act on planes
oriented at 45o to the principal planes and that they have magnitude equal to half the
difference between the principal stresses.
From 3.3.39, 3.4.8 and 3.4.9, the shear stress on the plane is
σ S2 = (σ 12 n12 + σ 22 n22 + σ 32 n32 ) − (σ 1 n12 + σ 2 n22 + σ 3 n32 )

2
(3.4.12)
Using the condition n12 + n22 + n32 = 1 to eliminate n3 leads to
σ S2 = (σ 12 − σ 32 )n12 + (σ 22 − σ 32 )n22 + σ 32 − [(σ 1 − σ 3 )n12 + (σ 2 − σ 3 )n22 + σ 3 ]

2
(3.4.13)
The stationary points are now obtained by equating the partial derivatives with respect to
the two variables n1 and n2 to zero:
( )
∂ σ S2
{ [
= n1 (σ 1 − σ 3 ) σ 1 − σ 3 − 2 (σ 1 − σ 3 )n12 + (σ 2 − σ 3 )n 22 = 0 ]}
∂n1
( )
∂ σ S2
{ [
= n 2 (σ 2 − σ 3 ) σ 2 − σ 3 − 2 (σ 1 − σ 3 )n12 + (σ 2 − σ 3 )n22 = 0 ]}
(3.4.14)
∂n 2
One sees immediately that n1 = n2 = 0 (so that n3 = ±1 ) is a solution; this is the principal
direction e 3 and the shear stress is by definition zero on the plane with this normal. In

Section 3.4
this calculation, the component n3 was eliminated and σ S2 was treated as a function of
the variables (n1 , n2 ) . Similarly, n1 can be eliminated with (n2 , n3 ) treated as the
variables, leading to the solution n = e1 , and n2 can be eliminated with (n1 , n3 ) treated as
the variables, leading to the solution n = e 2 . Thus these solutions lead to the minimum
shear stress value σ S2 = 0 .
A second solution to Eqn. 3.4.14 can be seen to be n1 = 0, n2 = ±1 / 2 (so that

n3 = ±1 / 2 ) with corresponding shear stress values σ S2 = 1
4
(σ 2 − σ 3 )2 . Two other
solutions can be obtained as described earlier, by eliminating n1 and by eliminating n2 .
The full solution is listed below, and these are evidently the maximum (absolute value of
the) shear stresses acting at a point:
⎛ 1 1 ⎞ 1
n = ⎜ 0, ± ,± ⎟, σ S = σ 2 − σ 3
⎝ 2 2⎠ 2
⎛ 1 1 ⎞ 1
n = ⎜± , 0, ± ⎟, σ S = σ 3 − σ 1 (3.4.15)
⎝ 2 2⎠ 2
⎛ 1 1 ⎞ 1
n = ⎜± ,± , 0 ⎟, σ S = σ 1 − σ 2
⎝ 2 2 ⎠ 2
Taking σ 1 ≥ σ 2 ≥ σ 3 , the maximum shear stress at a point is
1
τ max = (σ 1 − σ 3 ) (3.4.16)
2
and acts on a plane with normal oriented at 45o to the 1 and 3 principal directions. This is
illustrated in Fig. 3.4.3.
principal
directions
τ max
1
τ max
Figure 3.4.3: maximum shear stress at apoint
Example (maximum shear stress)
Consider the stress state

Section 3.4
⎡5 0 0 ⎤
[σ ] = ⎢0
ij − 6 − 12⎥
⎢⎣0 − 12 1 ⎥⎦
This is the same tensor considered in the example of §1.11.1. Using the results of that
example, the principal stresses are σ 1 = 10, σ 2 = 5, σ 3 = −15 and so the maximum shear
stress at that point is
1 25
τ max = (σ 1 − σ 3 ) =
2 2
The planes and direction upon which they act are shown in Fig. 3.4.4.
x3 τ max
n̂1
n̂3
37o
x2
x1 n̂ 2
Figure 3.4.4: maximum shear stress

Section 3.6
3.6 The Equations of Motion and Symmetry of Stress

In Part II, §1.1, the Equations of Motion were derived using Newton’s Law applied to a
differential material element. Here, they are derived using the principle of linear
momentum.
3.6.1 The Equations of Motion (Spatial Form)
Application of Cauchy’s law t = σn and the divergence theorem 1.14.21 to 3.2.7 leads
directly to the global form of the equations of motion
⎡ ∂σ ij ⎤
∫ [div σ + b]dv = ∫ ρ v& dv, ∫v ⎢ ∂x j ⎥⎥ dv = ∫v ρ v&i dv
⎢ + bi (3.6.1)
v v ⎣ ⎦
The corresponding local form is then
dv ∂σ ij dvi
div σ + b = ρ , + bi = ρ Equations of Motion (3.6.2)
dt ∂x j dt
The term on the right is called the inertial, or kinetic, term, representing the change in
momentum. The material time derivative of the spatial velocity field is
dv ∂v dv1 ∂v1 ⎛ ∂v1 ∂v ∂v ⎞

= + (gradv )v so = + ⎜⎜ v1 + 1 v 2 + 1 v3 ⎟⎟ , etc.
dt ∂t dt ∂t ⎝ ∂x1 ∂x 2 ∂x3 ⎠
and it can be seen that the equations of motion are non-linear in the velocities.
Equations of Equilibrium
When the acceleration is zero, the equations reduce to the equations of equilibrium,
div σ + b = 0 Equations of Equilibrium (3.6.3)
Flows
A flow is a set of quantities associated with the system of forces t and b, for example the
quantities v,σ, ρ . A flow is steady if the associated spatial quantities are independent of
time. A potential flow is one for which the velocity field can be written as the gradient
of a scalar function, v = gradφ . An irrotational flow is one for which curl v = 0 .

Section 3.6
3.6.2 The Equations of Motion (Material Form)
In the spatial form, the linear momentum of a mass element is ρvdv . In the material
form it is ρ 0 VdV . Here, V is the same velocity as v, only it is now expressed in terms of
the material coordinates X, and ρdv = ρ 0 dV . The linear momentum of a collection of
material particles occupying the volume v in the current configuration can thus be
expressed in terms of an integral over the corresponding volume V in the reference
configuration:
L(t ) = ∫ ρ 0 (X )V (X, t )dV Linear Momentum (Material Form) (3.6.4)

V
and the principle of linear momentum is now, using 3.1.31,
dV
ρ 0 (X )V (X, t )dV = ∫ ρ 0
d
∫
dt V V
dt
dV ≡ F(t ) (3.6.5)
The external forces F to be considered are those acting on the current configuration.
Suppose that the surface force acting on a surface element ds in the current configuration
is df surf = tds = TdS , where t and T are, respectively, the Cauchy traction vector and the
PK1 traction vector (Eqns. 3.5.3-4). Also, just as the PK1 stress measures the actual force
in the current configuration, but per unit surface area in the reference configuration, one
can introduce the reference body force B: this is the actual body force acting in the
current configuration, per unit volume in the reference configuration. Thus if the body
force acting on a volume element dv in the current configuration is df body , then
df body = bdv = BdV (3.6.6)
The resultant force acting on the body is then
F(t ) = ∫ TdS + ∫ B dV , Fi = ∫ Ti dS + ∫ Bi dV (3.6.7)

S V S V
Using Cauchy’s law, T = PN , where P is the PK1 stress, and the divergence theorem
1.12.21, 3.6.5 and 3.6.7 lead to
dV
V
∫ [DivP + B]dV = ∫ ρ V
0
dt
dV (3.6.8)
and the corresponding local form is
dV ∂Pij dVi
Div P + B = ρ 0 , + Bi = ρ 0
dt ∂X j dt
Equations of Motion (Material Form) (3.6.9)

Section 3.6
Derivation from the Spatial Form
The equations of motion can also be derived directly from the spatial equations. In order
to do this, one must first show that Div (JF − T ) is zero. One finds that (using the
divergence theorem, Nanson’s formula 2.2.59 and the fact that divI = 0 )
∫ Div (JF )dV = ∫ JF NdS = ∫ nds = ∫ Inds = ∫ divI dv = 0

−T −T
( )dV =
V S s s v
∂ JF ji−1 ∂δ ij (3.6.10)
∫ ∫ JF N i dS = ∫ ni ds = ∫ δ ij n j ds = ∫
−1
dv = 0
∂X j ∂xi
ji
V S s s v
This result is known as the Piola identity. Thus, with the PK1 stress related to the
Cauchy stress through 3.5.8, P = JσF − T , and using identity 1.14.16c,
DivP = Div σ JF − T ( ( ))
= σDiv(JF ) + Gradσ : (JF )
−T −T
(3.6.11)
= JGradσ : F −T
From 2.2.8c,
DivP = Jdivσ (3.6.12)
Then, with dv = JdV and 3.6.6, the equations of motion in the spatial form can now be
transformed according to
∫ [div σ + b]dv = ∫ ρ v& dv

v v
→ ∫ [DivP + B]dV = ∫ ρ V& dV
V V
0
as before.
3.6.3 Symmetry of the Cauchy Stress

It will now be shown that the principle of angular momentum leads to the requirement
that the Cauchy stress tensor is symmetric. Applying Cauchy’s law to 3.2.11,
∫ r × (σn )ds + ∫ r × bdv = dt ∫ r × ρvdv

d
s v v
(3.6.13)
d
∫s ε ijk x jσ kl nl dS + ∫v ε ijk x j bk dv = dt ∫v ε ijk x j ρvk dv
The surface integral can be converted into a volume integral using the divergence
theorem. Using the index notation, and concentrating on the integrand of the resulting
volume integral, one has, using 1.3.14 (the permutation symbol is a constant here,
∂ε ijk / ∂xl = 0 ),

Section 3.6
∂ (x j σ kl ) ⎧ ∂σ kl ⎫ ⎧ ∂σ kl ⎫
ε ijk = ε ijk ⎨ x j + σ kl δ jl ⎬ = ε ijk ⎨ x j + σ kj ⎬ ≡ r × divσ + E : σ T (3.6.14)
∂xl ⎩ ∂xl ⎭ ⎩ ∂xl ⎭
where E is the third-order permutation tensor, Eqn. 1.9.6, E = ε ijk (e i ⊗ e j ⊗ e k ) . Thus,

with the Reynold’s transport identity 3.1.31,
∫ {r × divσ + E : σ }dv + ∫ r × bdv = ∫ ρ dt (r × v )dv

T d
(3.6.15)
v v v
The material derivative of this cross product is
d
(r × v ) = r × dv + dr × v = r × dv + v × v = r × dv (3.6.16)
dt dt dt dt dt
and so
⎧ dv ⎫
∫E : σ dv + ∫ r × ⎨divσ + b − ρ ⎬dv = 0
T
(3.6.17)
v v ⎩ dt ⎭
From the equations of motion 2.6.2, the term inside the brackets is zero, so that
E : σ T = 0, ε ijk σ kj = 0 (3.6.18)
It follows, from expansion of this relation, that the matrix of stress components must be
symmetric:
σ = σT , σ ij = σ ji Symmetry of Stress (3.6.19)
3.6.4 Consequences in the Material Form

Here, the consequences of 3.6.19 on the PK1 and PK2 stresses is examined. Using the
result σ = σ T and 3.5.8, σ = J −1 PF T ,
J −1 PF T = (J −1 PF T ) = J −1FP T
T
(3.6.20)
so that
PF T = FP T , Pik F jk = Fik Pjk (3.6.21)
These equations are trivial when i = j , not providing any constraint on P. On the other
hand, when i ≠ j one has the three equations

Section 3.6
P11 F21 + P12 F22 + P13 F23 = F11 P21 + F12 P22 + F13 P23
P11 F31 + P12 F32 + P13 F33 = F11 P31 + F12 P32 + F13 P33 (3.6.22)
P21 F31 + P22 F32 + P23 F33 = F21 P31 + F22 P32 + F23 P33
Thus angular momentum considerations imposes these three constraints on the PK1 stress
(as they imposed the three constraints σ 12 = σ 21 , σ 13 = σ 31 , σ 23 = σ 32 on the Cauchy
stress).
It has already been seen that a consequence of the symmetry of the Cauchy stress is the
symmetry of the PK2 stress S; thus, formally, the symmetry of S is the result of the
angular momentum principle.

Section 3.7
3.7 Boundary Conditions and The Boundary Value

Problem
In order to solve a mechanics problem, one must specify certain conditions around the
boundary of the material under consideration. Such boundary conditions will be
discussed here, together with the resulting boundary value problem (BVP). (see Part I,
3.5.1, for a discussion of stress boundary conditions.)
3.7.1 Boundary Conditions

There are two types of boundary condition, those on displacement and those on traction.
Denote the body in the reference condition by B0 and in the current configuration by B.
Denote the boundary of the body in the reference configuration by S and in the current
configuration by s, Fig. 3.7.1.
Displacement Boundary Conditions
The position of particles may be specified over some portion of the boundary in the
current configuration. That is, x = χ (X ) is specified to be x say, over some portion s u of
s, Fig. 3.7.1, which corresponds to the portion S u of S . With u(x) = x − X(x) , or
U( X) = x( X) − X , this can be expressed as
u(x) = u (x), x ∈ su
(3.7.1)
U( X) = U ( X), X ∈ S u
These are called displacement boundary conditions. The most commonly encountered
displacement boundary condition is where some portion of the boundary is fixed, in
which case u (x ) = o .
Sσ u=u su
U=U
S 0u
T=T Sσ
t=t
B0
B
S x
s
X
Figure 3.7.1: Boundary conditions

Section 3.7
Traction Boundary Conditions
Traction t = t can be specified over a portion sσ of the boundary, Fig. 3.7.1. These
traction boundary conditions are related to the PK1 traction T = T over the
corresponding surface S σ in the reference configuration, through Eqns. 3.5.1-4,
TdS = PNdS = tds = σnds (3.7.2)
One usually knows the position of the boundary S and the normal N(X) in the reference
configuration. As deformation proceeds, the PK1 traction develops according to T = PN
with, from 3.5.8, P = JσF − T . The PK1 stress will in general depend on the motion x and
the deformation gradient F , so the traction boundary condition can be expressed in the
general form
T = T (X, x, F ) (3.7.3)
Example: Fluid Pressure
Consider the case of fluid pressure p around the boundary, t = − pn , Fig. 3.7.2. The
Cauchy traction t depends through the normal n on the new position and geometry of
the surface s σ . Also, T = − pJF − T N , which is of the general form 3.7.3.
X 1 , x1 θ
h
p t = − pn
T
X 1 , x1
E 2 ,e 2
a
E1 ,e1
Figure 3.7.2: Fluid pressure on deforming material
Consider a material under water with part of its surface deforming as shown in Fig. 3.7.2.
Referring to the figure, N = −E1 , n = − cos θe1 + sin θe 2 , σ = − pI , p = ρg (h − x 2 ) and
x1 = X 1 + a + X 2 tan θ ⎡1 tan θ 0⎤
x2 = X 2 , F = ⎢⎢0 1 0⎥⎥ , J = det F = 1
x3 = X 3 ⎢⎣0 0 1⎥⎦
The traction vectors and PK1 stress are

Section 3.7
⎡− cos θ ⎤ ⎡ −1 ⎤ ⎡ 1 0 0⎤
⎢ ⎥ ⎢ ⎥
t = − ρg (h − x 2 )⎢ sin θ ⎥, T = − ρg (h − X 2 )⎢ tan θ ⎥, P = − ρg (h − x 2 )⎢⎢− tan θ 1 0⎥⎥
⎢⎣ 0 ⎥⎦ ⎢⎣ 0 ⎥⎦ ⎢⎣ 0 0 1⎥⎦
with (note that dS / ds = cos θ ) t = p and T = p / cos θ . The traction vectors clearly
depend on both position, and the deformation through θ . In this example,
gradu = F − I = GradU = I − F −1 = tan θ e1 ⊗ e 2 and
θ (gradu ) = arctan gradu = arctan gradu : gradu

■
Dead Loading
A special case of loading is that of dead loading, where
T = T (X ) (3.7.4)
Here, the PK1 stress on the boundary does not change with the deformation and an
initially normal traction will not remain so as deformation proceeds.
For example, if one considers again the geometry of Fig. 3.7.2, this time take
⎡1⎤
T ( X) = PN = − pN = ρg (h − X 2 )⎢⎢0⎥⎥, P(X ) = − ρg (h − X 2 )I
⎢⎣0⎥⎦
Then
⎡1 ⎤ ⎡ 1 0 0⎤
t (x, θ ) = cos θ ρg (h − x 2 )⎢⎢0⎥⎥, σ(x, θ ) = − ρg (h − x 2 )⎢⎢ tan θ 1 0⎥⎥
⎢⎣0⎥⎦ ⎢⎣ 0 0 1⎥⎦
3.7.2 The Boundary Value Problem

The equations of motion 3.6.2, 3.6.9, are a set of three differential equations. In the
solution of any problem, one would have to supplement these equations with others, for
example a constitutive equation expressing a relationship between the stress and the
kinematic variables (see Part IV). This constitutive relation will typically relate the stress
to the strains, or rates of strain, for example σ = f (e, d) . Suppose then that the stresses
are known in terms of the strains and hence the displacements u. The equations of
motion are then a set of three second order differential equations in the three unknowns
u i (assuming that the body force b is a prescribed function of the problem). They need to
be subjected to certain boundary and initial conditions.

Section 3.7
Assume that the boundary conditions are such that the displacements are specified over
that part of the surface su and tractions are specified over that part s σ , with the total
surface s = s u + sσ , with s u ∩ sσ = 0 1. Thus
t = σn = t , on sσ
Boundary Conditions (3.7.5)
u = u, on su
where the overbar signifies quantities which are prescribed. Initial conditions are also
required for the displacement and velocity, so that
u(x, t ) = u 0 (x), at t = 0
Initial Conditions (3.7.6)
u& (x, t ) = u& 0 (x), at t = 0
and it is usually taken that x = X at t = 0 . Comparing 3.7.5 and 3.7.6, one also requires
⋅
that u 0 = u , u& 0 = u over s u , so that the boundary and initial conditions are compatible.
These equations together, the differential equations of motion and the boundary and
initial conditions, are called the strong form of the initial boundary value problem
(BVP):
divσ + b = ρv& = ρu
&&
t = σn = t , on sσ
u = u, on su Strong form of the Initial BVP (3.7.7)
u(x, t ) = u 0 (x), at t = 0
u& (x, t ) = u& 0 (x), at t = 0
When the problem is quasi-static, so the accelerations can be neglected, the equations of
motion reduce to the equations of equilibrium 3.6.3. In that case one does not need initial
conditions and one has a boundary value problem involving 3.7.5 only.
It is only in certain special cases and in certain simple problems that an exact solution can
be obtained to these equations. An alternative solution strategy is to convert these
equations into what is known as the weak form. The weak form, which is in the form of
integrals rather than differential equations, can then be solved approximately using a
numerical technique, for example the Finite Element Method2. The weak form is
discussed in §3.9.
1
It is possible to specify both traction and displacement over the same portion of the boundary, but not the
same components. For example, if one specified t = t 1 e 1 on a boundary, one could also specify u = u 2 e 2 ,
but not u = u1 e 1 . In that case, one could imagine the boundary to consist of two separate boundaries, one
with conditions with respect to e 1 and one with respect to e 2 , and still write s u ∩ s σ = 0 .
2
Further, it is often easier to prove results regarding the uniqueness and stability of solutions to the problem
when it is cast in the weak form

Section 3.7
In the material form, the boundary conditions are
T = PN = T, on S σ
Boundary Conditions (3.7.8)
U = U, on S u
and the initial conditions are
U( X, t ) = U 0 ( X), at t = 0
Initial Conditions (3.7.9)
& ( X, t ) = U
U & ( X), at t = 0
0
and the initial vale problem is
DivP + B = ρ 0 V
& = ρU
&&
T = PN = T, on S σ
U = U, on S u Strong form of the Initial BVP (3.7.10)
U( X, t ) = U 0 ( X), at t = 0
& ( X, t ) = U
U & ( X), at t = 0
0

Solid Mechanics

Uploaded by

Copyright:

Available Formats

Solid Mechanics

Uploaded by

Document Information

Original Title

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

Solid Mechanics

Uploaded by

Copyright:

Available Formats

What is solid mechanics and what does it study?

What is solid mechanics and what does it study?

What are some examples of the wide range of applications of solid mechanics?

What are some examples of the wide range of applications of solid mechanics?

1.1 What is Solid Mechanics?

When will this cliff How does the

How will the San Andreas fault

why does nature

Solid Mechanics Part I 3 Kelly

1.1.1 Aspects of Solid Mechanics

Solid Mechanics Part I 4 Kelly

More advanced theories of deformable solid materials include

Some other topics embraced by Solid Mechanics, are

Solid Mechanics Part I 5 Kelly

Solid Mechanics Part I 6 Kelly

1.2 What is in this Book?

I. An Introduction to Solid Mechanics

Solid Mechanics Part I 7 Kelly

Finally, in Book V, material models are described.

Solid Mechanics Part I 8 Kelly

2.1 The Fundamental Concepts and Principles of

2.1.1 The Fundamental Concepts

Figure 2.1.1: a particle in space

2.1.2 The Fundamental Principles

Newton’s First Law:

Solid Mechanics Part I 11 Kelly

Figure 2.1.2: the resultant of a system of forces acting on a particle; (a)

Example (illustrating Newton’s First Law)

Figure 2.1.3: a zero resultant force acting on a boat

Solid Mechanics Part I 12 Kelly

Newton’s Second Law:

Newton’s Third Law:

Solid Mechanics Part I 13 Kelly

2.2 The Statics of Particles

2.2.1 Equilibrium of a Particle

(a) (b) (c)

Figure 2.2.1: A force F decomposed into components Fx and Fy

Solid Mechanics Part I 14 Kelly

Consider the particle in Fig. 2.2.2, subjected to forces F1 , F2 , F3 . The particle is in

Figure 2.2.2: Calculating the resultant of three forces by

In general then, if a set of forces F1 , F2 , , Fn act on a particle, the particle is in

F x  0, F y  0, F z 0 Equations of Equilibrium (particle) (2.2.1)

Solid Mechanics Part I 15 Kelly

The equations of equilibrium for particle C are

F x   FBC  FAC cos 60  100 cos   0,

leading to FAC  46.2 N, FBC  36.9 N .

2.2.2 Rough and Smooth Surfaces

When a particle meets a smooth surface, there is no resistance to tangential movement.

Solid Mechanics Part I 16 Kelly

(a) (b) (c)

Figure 2.2.4: a particle sitting on a surface; (a) a rough surface, (b)

Solid Mechanics Part I 17 Kelly

Solid Mechanics Part I 18 Kelly

2.3 The Statics of Rigid Bodies

2.3.1 Moments, Couples and Equivalent Forces

The moment M o of a force F can be written as

Solid Mechanics Part I 19 Kelly

Figure 2.3.2: Two equivalent forces

Figure 2.3.3: A couple