Nonlinear
Solid
Mechanics
A
Continuum
Approach
for
Engineering
Gerhard
A.
Holzapfel
Graz
University
of
Technology,
Austria
JOHN
WILEY
&
SONS,
LTD
Chichester

Weinheim

New
York

Brisbane
Singapore
.
Toronto
Copyright
O
2000
by
John
Wiley
&Sons
Ltd,
Baffins
Lane,
Chichester,
West
Sussex
PO19
IUD,
England
National
01243
779777
International
(4)
1243
779777
emall
(for
orders
and customer
service
enquiries):
csbooks8wdey.co.uk
Visit
our
Home
Page
on
http://www.wiley.co.uk
or
http://www.wiley.com
Reprinted,
with
corrections,
December
200
1
All
rights
reserved.
No
part
of
this
publication
may
be
reproduced,
stored
in
a
retrieval
system,
or
transmitted,
in
any
form
or
by
any
means,
electronic,
mechanical,
photocopying,
recording,
scanning
or
otherwise,
except
under
the
terms
of
the
Copyright.
Designs
and
Patents
Act
1988
or
under
the
terms
of
a
licence
issued
by
(he
Copyright
Licensing
Agency,
90
Tottenham
Court
Road,
London,
UK
W1P
9
HE,
without
the
permission
in
writing
of
the
publisher.
Orher
Wrley
Edirorral
Oflces
John
Wlle~
&Sons,
Inc.,
605
Third
Avenue,
New
York,
NY
101580012,
USA
UlleyVCH
Verlag
GmbH,
Pappelallee
3,
D69469
Wemnhelm,
Germany
Jacaranda
Wiley
Ltd,
33
Park
Road.
Milton,
Queensland
4064,
Australla
John
W~leg
&
Sons
(Asia)
Pte
Ltd,
2
Clementl
Loop
#0201,
lm
Xlng
Dlstnpark,
Singapore
0512
John
Wiley
&
Sons
(Canada)
Ltd.
22
Worcester
Road.
Rexdale,
Ontario
M9W
1L1,
Canada
.I"
lmCI_._
.
..'
__1___1
Libray
of
Congress
Caialoging
in
Publication
Dafa
Holzapfel,
Gerhard
A.
Nonlinear
solid
mechanics
:
a
continuum
approach
for
engineering/
Gerhard
A.
Holzapfel.
p.
cm.
Includes
bibliographical
reierences
and
index.
ISBN
047182304XISBN
047182
198
1.
Continuum
mechanics.
I.
Title.
British
Library
Cataloguing
in
Publication
Data
A
cataloguefecord
for
this
book
is
available
from
the
British
Library
ISBN
047182304X
(ppc)
0471823198
(pbk)
Produced
from
cameraready
copy
supplied
by
the
author
Printed
and
bound
in
Grear
Britain
by
Antony
Roue
Ltd,
Chippenham,
Wiltshire
This
book
is
printed
on
acidfree
paper
responsibly
manufactured
from
sustanable
forestq,
in
uhich
at
least
two
trees
are
planted
for
each
one
used
for
paper
production.
Contents
Preface
Acknowledgements
ix
xiii
1
Introduction
to
Vectors
and
Tensors
1
Algebra
of
Vectors
...........................
1
Algebra
of
Tensors
...........................
9
Higherorder
Tensors
..........................
20
Eigenvalues,
Eigenvectors
of
Tensors
.................'
24
Transformation
Laws
for
Basis
Vectors
and
Components
.......
28
General
Bases
.............................
32
Scalar,
Vector,
Tensor
Functions
....................
40
Gradients
and
Related
Operators
....................
44
Integral
Theorems
...........................
52
2
Kinematics
Configurations,
and
Motions
of
Continuum
Bodies
..........
Displacement,
Velocity,
Acceleration
Fields
...............
Material,
Spatial
Derivatives
......................
.........................
Deformation
Gradient
Strain
Tensors
.............................
Rotation,
Stretch
Tensors
........................
Rates
of
Deformation
Tensors
.....................
..........................
Lie
Time
Derivatives
vi
Contents
3
The
Concept
of
Stress
..................
3.1
Traction
Vectors.
and
Stress
Tensors
109
.........................
3.2
Extremal
Stress
Values
119
......................
3.3
Examples
of
States
of
Stress
123
.....................
3.4
Alternative
Stress
Tensors

127
4
Balance
Principles
.........................
4.1
conservation
of
Mass
131
.....................
4.2
Reynolds'
Transport
Theorem
138
.....................
4.3
Momentum
Balance
Principles
141
....................
4.4
Balance
of
Mechanical
Energy
152
..........
4.5
Balance
of
Energy
in Continuum
Thermodynamics
161
......................
4.6
Entropy
Inequality
Principle
166
......................
4.7
Master
Balance
Principle
174
5
Some
Aspects
of
Objectivity
...........
5.1
Change
of
Observer.
and
Objective
Tensor
Fields
179
..................
5.2
Superimposed
Rigidbody
Motions
187
.............................
5.3
Objective
Rates
192
................
5.4
Invariance
of
Elastic
Material
Response
196
6
Hyperelastic
Materials
..............
6.1
General
Remarks
on
Constitutive
Equations
206
....................
6.2
Isotropic
Hyperelastic
Materials
212
.................
6.3
Incompressible
Hyperelastic
Materials
222
.................
6.4
Compressible
Hyperelastic
Materials
227
................
6.5
Some
Forms
of
Strainenergy
Functions
235
............................
6.6
Elasticity
Tensors
252
....................
6.7
Transversely
Isotropic
Materials
265
Contents
vii
6.8
Composite
Materials
with
Two
Families
of
Fibers
...........
272
6.9
Constitutive
Models
with
Internal
Variables
..............
278
6.10
Viscoelastic
Materials
at
Large
Strains
.................
282
6.11
Hyperelastic
Materials
with
Isotropic
Damage
............
295
7
Thermodynamics
of
Materials
305
Physical
Preliminaries
.........................
Thermoelasticity
of
Macroscopic
Networks
..............
Thermodynamic
Potentials
.......................
...............................
Calorimetry
Isothermal.
Isentropic
Elasticity
Tensors
................
Entropic
Elastic
Materials
.......................
........
ThermodynamicExtension
of
Ogden's
Material
Model
.............
Simple
Tension
of
Entropic
Elastic
Materials
...............
Thermodynamics
with Internal
Variables
8
Variational
Principles
371
8.1
Virtual
Displacements.
Variations
...................
372
8.2
Principle
of
Virtual
Work
.......................
377
8.3
Principle
of
Stationary
Potenha1
Energy
................
386
8.4
Linearization
of
the
Principle
of
Virtual
Work
............
392
....................
8.5
Twofield
Variational
Principles
402
...................
8.6
Threefield
Variahonal
Principles
409
References
Index
My
desire
in
writing
this
textbook
was
to
show
the
fascination
and
beauty
of
nonlin
ear
solid
mechanics
and
thembdynamics
from
an
engineering
computational
point
of
view.
My
primary
goal
was
not
only
to
offer
a
modem
introductory
textbook
using
the
continuum
approach
to
be
read
with
mterest,
enjoyment
and
curiosity,
but
also
to
offer
a
reference
book
that
incorporates
some
of
the
recent
developments
in
the
field.
I
wanted
to
stimulate
and
invite
the
reader
to
study
this
exciting
science
and
take
him
on
a
pleasant
journey
in
the
wonderful
world
of
nonlinear
mechanics,
which
serves
as
a
solid
basis
for
a
surprisingly
large
variety
of
problems
arising
in
practical
engineering.
Linear
theories
of
solid
mechanics
are
highly
developed
and
are
in
a
satisfactory
state
of
complehon.
Most
processes
in nature,
however,
are
hlghly
nonlinear.
The
ap
proach
taken
has
the
aim
of
providing
insight
in
the
basic
concepts
of
solid
mechanics
wlth
particular
reference
to
the
nonlinear
regime.
Once
familiar
with
the
main
ideas
the
reader
will
be
able
to
specialize
in
different
aspects
of
the
subject
matter.
I
felt
the
need
for
a
selfcontained
textbook
intended
primarily
for
beginners
who
want
to
un
derstand
the
correspondence
between
nonlinear
continuum
mechanics,
nonlinear
con
stitutive
models
and
variational
principles
as
essential
prerequisites
for
finite
element
formulations.
Of
course,
no
single
book
can
cover
all
aspects
of
the
broad
field
of
solidmechanics,
so
that
many
topics
are
not
discussed
here
at
all.
The
selection
of
the
material
for
inclusion
is
influenced
strongly
by
current
curricula,
trends
in
the
literature
and
the
author's
particular
interests
in
engineering
and
science.
Here,
a
particular
selection
and
style
was
chosen
in
order
to
highlight
some
of
the
more
inspiring
topics
in
solid
mechanics.
I
hope
that
my
choice,
which
is
of
course
subjective,
will
be
found
to
be
acceptable.
My
ultimate
intention
was
to
present
an
introduction
to
the
subject
matter
in
a
di
dactically
sound
manner
and
as
clearly
as
possible.
I
hope
that
the
text
provides
enough
insights
for
understanding
of
the
terminology
used
in
scientific
stateofthe
art
papers
and
to
find
the
'right
and
straightforward
path'
in
the
scientific
world
through
the
effec
tive
use
of
figures,
which
are
very
important
learning
tools.
They
are
designed
in
order
Preface
ct
attention
and
to
be
instructive
and
helpful
to
the
reader.
Necessary
mathe
and
physics
are
explained
in
the
text.
The
approach
used
in
each
of
the
eight
ters
will
enable
the
reader
to
work
through
the
chapters
in
order
of
appearance,
ic
being
presented
in
a
1ogical.sequence
and.
based
on
the
preceding
topics.
proper
understanding
of
the
subject
matter
requires
knowledge
of
tensor
alge
bra
and
tensor
calculus.
For
most
of
the
derivations
throughout
the
text
I
have
used
w
symbolic
notation
with
those
clear
boldfaced
symbols
which
give
the
subject
matter
F
'
a
distinguished
beauty.
However,
for
hgherorder
tensors and
for
final
results
in
most
of
the
derivations
I
have
used
index
notation,
which
provides
the
reader
with more
in
sight.
Terminology
is
printed
in
boldface
where
it
appears
for
the
first
time while
the
notation
used
in
the
text
is
defined
at
the
appropriate
point.
For
those
who
have
not
been
exposed
to
the
necessary
mathematics
I
have
included
a
chapter
on
tensor
algebra
and
tensor
calculus.
It
includes
the
essential
ideas
of
lin
earization
in
the
form
of
the
concept
of
the
directional
derivative.
Chapter
1
sumrna
rizes
elementary
properties
which
are
needed
for
the
vector
and
tensor
manipulations
performed
in
all
subsequent
chapters
and
which
are
necessary
to
many
problems
that
arise
frequently
in
engineering
and
physics.
It
is
the
prime
consideration
of
Chapter
2
to
use
tensor
analysis
for
the
descrip
tion
of
the
motion
and
finite
deformation
of
continua.
The
continuum
approach
is
introduced
along
with
the
notion
'Lagrangian'
(material)
and
'Eulerian'
(spatial)
de
scriptions.
In
a
systematic
way
the
most
important
kinematic
tensors
are
provided
and
their
physical
significance
explained.
The
pushforward
and
pullback
operations
for
material
and
spatial
quantities
and
the
concept
of
the
Lie
time
derivative
are
intro
duced.
The
concept
of
stress
is
the
main
topic
of
Chapter
3.
Cauchy's
stress
theorem
is
introduced,
along
with
the
Cauchy
and
first
PiolaKirchhoff
traction
vectors,
and
the
essential
stress
tensors are
defined
and
their
interrelationships
discussed.
In
Chap
ter
4
attention
is
focused
on
the
discussion
of
the
balance
principles.
Both
statics
and
dynamics
are
treated.
Based
upon
continuum
thermodynamics
the
entropy
inequality
principle
is
provided
and
the
general
structure
of
all
principles
is
summarized
as
the
master
balance
(inequality)
principle.
Chapter
5
deals
with
important
aspects
of
objec
tivity,
which
plays
a
crucial
part
in
nonlinear
continuum
mechanics.
A
discussion
of
change
of
observer
and
superimposed
rigidbody
motions
is
followed
by
a
development
of
objective
(stress)
rates
and
invariance
of
elastic
material
response.
Chapters
6
and
7
form
the
central
part
of
the
book
and
provide
insight
in
the
con
sbruction
of
nonlinear
constitutive
equations
for
the
description
of
the
mechanical
and
thermomechanical
behavior
of
solids.
These
two
chapters
show
the
essential
richness
of
the
field.
They
are
written
for
those
who
want
to
gain
experience
in
handling
ma
terial
models
and
deriving
stress
relations
and
the
associated
elasticity
tensors
that
are
fundamental
for
finite
element
methods.
Several
examples
and
exercises
are
aimed
at
enabling
the
reader
to
thinkin
terms
of
constitutive
models
and
to
formulate
more
com
Preface
xi
plex
material
models.
All
of
the
types
of
constitutive
equations
presented
are
accessible
for
use
within
finite
element
procedures.
The
bulk
of
Chapter
6
is
concerned
with
finite
elasticity
and
finite
viscoelasticity.
It
includes
a
discussion
of
isotropic,
incompressible
and
compressible
hyperelastic
ma
terials
and
provides
constitutive
models
for
transversely
isotropic
and
composite
ma
terials
which
are
suitable
for
a
large
number
of
applications
in
practical
engineering.
An
approach
to
inelastic
materials
with internal
variables
is
given
along
with
instruc
tive
examples
of
hyperelastic
materials
that
involve
relaxation
andlor
creep
effects
and
isotropic
damage
mechanisms
at
finite
strains.
The
main
purpose
of
Chapter
7
is
to
pro
vide
an
introduction
to
the
thermodynamics
of
materials.
This
chapter
is
devoted
not
only
to
the
foundation
of
continuum
thermodynamics
but
also
to
selected
topics
of
sta
tistical
thermodynamics.
It
starts
with
a
statistical
approach
by
summarizing
important
physical
aspects
of
the
thermoelastic
behavior
of
molecular
networks
(for
example:
amorphous
solid
polymers),
based
almost
entirely
on
an
entropy
concept,
and
con
tinues
with
a
systematic
phenomenological
approach
including
finite
thermoelasticity
and
finite
thermoviscoelasticity.
The
stressstraintemperature
response
of
socalled
entropic
elastic
materials
is
discussed
in
more
detail
and
based
on
a
representative
example
which
is
concerned
with
the
adiabatic
stretching
of
a
rubber
band.
Typical
thermomechanical
coupling
effects
are
studied.
Chapter
8
is
designed
to
cover
the
essential
features
of
the
most
important
varia
tional
principles
that
are
very
useful
in formulating
approximation
techniques
such
as
the
finite
element
method.
Although
finite
elements
are
not
treated
in
this
text,
it
is
hoped
that
this
.chapter
will
be
attractive
to
those
who
approach
the
subject
from
the
computational
side.
It
shows
the
relationship
between
the
strong
and
weak
forms
of
initial
boundaryvalue
problems,
presents
the
classical
principle
of
virtual
work
in
both
spatial
and
material
descriptions
and
its
linearized
form.
Two
and
threefield
varia
tional
principles
are
also
discussed.
The
present
text
ends
where
conventionally
a
book
on
the
finite
element
method
would
begin.
There
are
numerous
worked
examples
adjacent
to
the
relevant
text.
These
have
the
goal
of
clarifying
and
supplementing the
subject
matter.
In
many
cases
they
are
straightforward,
but
provide
an
essential
part
of
the
text.
The
symbol
W
is
used
to
de
note
the
end
of
an
exercise
or
a
proof.
The
endof
chapter
exercises
are
for
homework.
The
(almost)
200
exercises
provided
are
designed
to
supplement
the
text
and
to
con
solidate
concepts
discussed in the
text.
Most
of
them
serve
the
purpose
of
stimulating
the
reader
to
further
study
and
to
reinforce
and
develop
practical
skills
in
nonlinear
continuum
and
solid
mechanics,
towards
the
direction
of
coinputational
mechanics.
In
many
cases
the
solutions
of
selected
exercises
are
given
and
frequently
used
later
in
further
developments.
Therefore,
it
should
be
instructive
for
the
reader
to
work
through
a
reasonable
number
of
exercises.
Numerous
references
to
supplementary
material
are
suggested
and
discussed
briefly
xii
Preface
throughout
the
book.
However,
for
a
book
of
this
kind
it
is
not
possible
to
give
a
com
prehensive
bibliography
of
the
field.
Some
of
the
references
listed
serve
as
a
starting
point
for
more
advanced
studies.
The
material
in
this
book
is
based
on
a
sequence
of
courses
that
I
have
taught
at
the
University
of
Technology
in
Graz
and
Vienna.
The
mechanics
and
thermodynamics
of
solids
are
relevant
to
all
branches
of
engineering,
to
applied
mechanics,
mathemat
ics,
physics
and
material
science,
and
it
is
a
central
field
in
biomechanics.
This
book
is
primarily
addressed
to
graduate
students,
researchers
and
practitioners,
although
it
has
also
proven
to
be
of
interest
to
advanced
undergraduates.
Although
I
have
tried
to
provide
a
textbook
that
is
selfcontained
and
appropriate
for
selfinstruction,
it
is
desirable
that
the
reader
has
a
reasonable
background
in
elementary
mechanics
and
thermodynamics.
I
feel
that
Chapters
25
and
parts
of
the
remaining
chapters
are
well
suited
for
a
complete
course
on
nonlinear
continuum
mechanics
lasting
two
semesters
or
three
quarters.
A
one
semester
or
one
quarter
course
in
the
nonlinear
mechanics
of
solids
would
focus
on
Chapter
6,
while
a
one
quarter
course
in the
thermodynamics
of
solids
could
be
based
on
Chapter
7.
Chapter
1
is
the
core
for
a
course
that
provides
the
student
with
the
necessary
background
in
vector
and
tensor
analysis.
Chapter
8
is
certainly
not
designed
to
train
specialists
in
variational
principles,
but
to
form
a
basic
one
quarter
course
at
the
graduate
level.
I
believe
that
the
present
textbook,
in
providing
many
applications
to
engineer
ing
science,
is
not
too
advanced
mathematically.
Of
course,
some
of
the
results
pre
sented
may
be
derived
with
the
help
of
more
advanced
mathematics
using
theorems
and
proofs.
I
hope
that
this
book
will
help
pure
engineers
to
teach
nonlinear
continuum
mechanics
and
solid
mechanics.
Naturally,
as
the
author,
I
take
full
responsibility
for
not
doing
it
better.
Comments
and
criticisms
will
be
welcome
and
greatly
appreciated.
I
have
learnt
that
the
spirit
of
modem
continuum
mechanics
and
the
underlying
mathematics
are
as
important
to
the
design
of
powerful
finite
element
models
as
are
insights
in
the
theoretical
foun
dation
of
constitutive
models
and
variational
principles.
A
successful
transfer
of
that
combination
to
the
reader
would
indicate
that
my
objective
has
been
achieved.
Graz,
Austria,
August
1999
Gerhard
A.
Holzapfel
Preface
to
the
Second
Printing
The
focus
of
the
revision
for
this
second
printing
is
the
Insertion
of
additional
equations
on
pages
50,51,74,
and
an
additional
exercise
on
page
76.
This
has
caused
changes
in
the
numbers
of
some
equations in
parts
of
the
book
as
compared
with
the
first
printing.
I
have
also
done
some
minor
rewording
and
have
added
a
few
more
references.
Graz,
Austria,
October
2001
Gerhard
A.
Holzapfel
I
Acknowledgements
When
I
was
a
postdoctoral
student
at
Stanford
University
I
worked
with
the
late
Juan
C.
Simo,
Professor
of
Mechanical
Engineering;
to
whom
I
owe
my
deepest
thanks.
He
stimulated,
influenced
and
focused
my
study
and
writing
in
recent
years;
his
friendship,
versatility
and
dedication
to
scientific
excellence
provided
a
unique
learning
experience
for
me.
I
am
particularly
indebted
to
Ray
W
Ogden,
Professor
of
Mathematics
at
the
Uni
versity
of
Glasgow,
who
spent
a
lot
of
time
in
reading
the
entire
manuscript
and
rectify
ing
certain
ineptnesses.
His
outstanding
expertise
in
the
field
made
working
with
him
an
inspiring
pleasure.
His
detailed
scientific
criticism
and
suggestions
for
improve
ments
of
the
text
were
of
immeasurable
help.
Many
others
have
contributed
to
the
book.
Here
I
mention
my
collaborators,
whose
gentle
encouragement
and
support
during
the
course
of
the
preparation
of
the
manu
script
I
gratefully
acknowledge.
In
particular,
the
inspiring
and
detailed
comments
of
Dr
Christian
A.J.
SchukeBauer,
from
a
background
in
physics
and
medicine,
were
extremely
helpful.
He
let
me
filter
this
text
through
his
sharp
mind.
Also,
Christian
T.
Gasser,
whose
background
is
in
mechanical
engineering,
has
suggested
a
number
of
valuable
improvements
to the
substance
of
the
text.
His
profound
remarks
in
class
prevented
me
from
getting
away
with
anything.
Also,
Elisabetlz
Pernkopf,
a
mathe
matician,
deserves
special
thanks
for
her
generous
assistance.
She
gave
much
helpful
advice
on
the
preparation
of
this
text
and
offered
many
suggestions
for
improving
it.
Special
thanks
are
due
to
Michael
Stadler
for
his
productive
discussions
and
to
Mario
Ch.
Palli
for
his
patience
in
preparing
the
figures.
I
am
grateful
to
each
of
these
indi
viduals
without
whose
contributions
the
book
would
not
have
taken
this
shape.
I
want
to
thank
the
Department
of
Civil
Engineering,
Graz
University
of
Technol
ogy,
for
providing
an
environment
in which
this
project
could
be
completed.
My
grat
itude
goes
to
Gemot
Beer,
Professor
of
Civil
Engineering,
for
his
outspoken
support.
I
also
wish
to
acknowledge
the
Austrian
Science
Foundation,
which
has
influenced
my
scientific
agenda
through
the
financial
support
of
several
grants
over
the
past
eight
xiii
xiv
Acknowledgements
The
enjoyment
I
experienced
in
writing
this
textbook
would
not
have
been
the
same
without
the
moral
support
of
numerous
friends.
My
thanks
belong
to
all
of
them
who
tolerated
my
absence
when
I
disappeared
for
many
evenings
and
weekends
in
order
to
bring
the
ideas
of
this
fascinating
field
to
you,
the
reader.
1
Introduction
to
Vectors
and
Tensors
The
aim
of
this
chapter
is
to
present
the
fundamental
rules
and
standard
results
of
ten
sor
algebra
and
tensor
calculus
permanently
used
in
nonlinear
continuum
mechanics.
Some
readers
may
prefer
to
pass
directly
to
Chapter
2
leaving
the
present
part
for
ref
erence
as
needed.
Many
of
the
statements
are
given
without
proofs.
For
a
more
detailed
exposition
see
the
standard
texts
by
HALMOS
[19581,
TRUESDELL
and
NoLL
[I9921
or
the
textbooks
by
CHADWICK
[1976],
GURTIN
[1981al,
SIMMONDS
[1994],
DANIELSON
[I9971
and
OGDEN
[I9971
among
many
other
references
on
vectors
and
tensors.
To
recall
the
elements
of
linear
algebra
see,
for
example,
the
book
by
STRANG
[1988a].
In
this
text
we
use
lowercase
Greek
letters
for
scalars,
lowercase
bo1d;face
Lahn
letters
for
vectors,
uppercase
boldface
Latin
letters
for
secondorder
tensors,
upper
case
boldface
calligraphic
letters
for
thirdorder
tensors
and
uppercase
blackboard
Latin
letters
for
fourthorder
tensors;
for
example,
a,
P,
y,
.
.
(scalars)
a,
b,
c,
.
.
.
(vectors)
A,
B,
C,
.
.
.
(2ndorder
tensors)
4
B,
C,
.
.
.
(3rdorder
tensors)
1
(1.1)
A,
B,
@,
.
.
.
(4thorder
tensors)
.
For
equations
such
as
u=av=pa=yb
.
(1.2)
we
agree
that
(1.2)2
refers
to
u
=
pa.
Often
the
derivations
of
formulas
need
relations
introduced
previously.
If
this
is
the
case
we
refer
to
these
relations
in
a
particular
order,
reflecting
the
consecutive
steps
necessary
for
deriving
the
formula
in
question
1.
Algebra
of
Vectors
A
physical
quantity,
completely
described
by
a
single
real
number,
such
as
tempera
ture,
densit)!
or
mass,
is
called
a
scalar
designated
by
a,
0,
y,
.
.
.
A
vector
designated
2
1
Introduction
to
Vectors
and
Tensors
by
u,
v,
w,
.
.
.
(or
in
other
texts
frequently
designated
by
2,
2,
w,
.
.
.,
or
C,
v',
w',
.
.
.),
is
a
directed
line
element
in
space.
It
is
a
model
for
physical
quantities
having
both
di
rection
and
length,
for
example,
force,
velocity
or
acceleration.
Two
vectors
that
have
the
same
direction
and
length
are
said
to
be
equal.
The
sum
of
vectors
yields
a
new
vector,
based
on
the
parallelogram
law
of
addi
tion.
The
following
properties
u+v=v+u
;
(u+v)+w=u+(v+w)
;
u+o=u
;
ut
(u)
=
0
hold,
where
o
denotes
the
unique
zero
vector
with
unspecified
direction
and
zero
length.
Let
u
be
a
vector
and
(Y
be
a
real
number
(a scalar).
Then
the
scalar
multiplication
cru
produces
a
new
vector
with
the
same
direction
as
u
if
0
>
0
or
with
the
opposite
direction
to
u
if
a:
<
0.
Further
properties
are:
Dot
product.
The
dot
(or
scalar
or
inner)
product
of
u
and
v,
denoted
by
u
.
v
(or
(u:
v)),
is
where
B(u,
v)
is
the
angle
between
two
nonzero
vectors
u
and
v
when
their
origins
coincide.
This product
gives
a
scalar
quantity
with the
properties
U.V
=
v.u
;
(1.11)
u.0
=
0
;
(1.12)
u.(a~+pw)
=
Q(u.v)+~(u.w)
(1.13)
u.u>O
e
ufo
and
u.u=O
u=o
.
(1.14)
The
quantity
lui
(or
IluI)
is
called
the
length
(or
norm
or
magnitude)
of
a
vector
u,
which
is
a
nonnegative
real
number.
It
is
defined
by
the
square
root
of
u
.
u,
i.e.
1.1
Algebra
of
Vectors
3
A
vector
e
is
called
a
unit
vector
if
lei
=
1.
A
nonzero
vector
u
is
said
to
be
orthogonal
(or
perpendicular)
to
a
nonzero
vector
v
if
u
.
v
=
0
with
Q(u,
v)
=
7;/2
.
(1.16)
Thus,
using
(1.10)
we
find
the
projection
of
a
vector
u
along
the
direction
of
a
vector
e
with
unit
length,
i.e.
For
a
geometrical
inte~pretation
of
eq.
(1.17)
see
Figure
1.1.
Figure
1.1
Projection
of
u
along
a
unit
vector
e.
Index
notation.
So
far
algebra
has
been
presented
in
symbolic
(or
direct
or
abso
lute)
notation
exclusively
employing
boldface
letters.
It
represents
a
very
convenient
and
concise
tool
to
manipulate
most
of
the
relations
used
in continuum
mechanics.
It
will
be
the
preferred
representation
in
this
text.
However,
particularly
in
computational
mechanics,
it
is
essential
to
refer
vector
(and
tensor)
quantities
to
a
basis.
Addition
ally,
to gain
more
insight
in
some
quantities
and
to
carry
out
mathematical
operations
among
(higherorder)
tensors
more
readily
(see
next
section)
it
is
often
helpful
to
refer
to
components.
In
order
to
present
coordinate
(or
component)
expressions
relative
to
a
right
handed
(or
dextral)
and
orthonormal
system
we
introduce
a
fixed
set
of
three
basis
vectors
el;
e2;
e3
(sometimes
introduced
as
i;
j:
k),
called
a
(Cartesian)
basis,
with
properties
These
vectors
of
unit
length
which
are
mutually
orthogonal
form
a
socalled
orthonor
ma1
system.
4
1
Introduction
to
Vectors
and
Tensors
Then
any
vector
u
in
the
threedimensional
Euclidean
space
is
represented
uni
quely
by
a
linear
combination
of
the
basis
vectors
el:
ez,
e3,
i.e.
where
the
three
real
numbers
ul;
u2,
u3
are
the
uniquely
determined
Cartesian
(or
rect
angular)
components
of
vector
u
along
the
given
directions
el:
e2,
es,
respectively
(see
Figure
1.2).
The
components
of
el:
e2:
e3
are
(1;
0:
O),
(0:
1:
O),
(0:
0;
I),
respectively.
Figure
1.2
Vector
u
with
its
Cartesian
components
~1:
uz:
us.
Using
index
(or
subscript
or
suffix)
notation,
relation
(1.19)
can
be
written
as
u
=
c:=,
uie;
or,
in
an
abbreviated
form
by
leaving
out
the
summation
symbol,
simply
as
u
=
uiei
:
(sum
over
i
=
1:
2:
3)
:
(1.20)
where
we
have
adopt
the
summation
convention,
invented
by
Einstein.
The
summa
tion
convention
says
that
whenever
an
index
is
repeated
(only
once)
in
the
same
term,
then,
a
summation
over
the
range
of
this
index
is
implied
unless
otherwise
indicated.
We
consider
only
the
threedimensional
Euclidean
space,
which
we
characterize
by
means
of
Latin
indices
i:
j:
k
.
.
.
running
over
1:
2:
3.
We
denote
the
basis
vectors
by
{e,),E(i,2,3)
collectively.
Subsequently,
in
this
text,
the
braces
{a)
will
stand
for
a
fixed
set
of
basis
vectors
and
the
symbol
a
for
any
tensor
element.
The
index
i
that
is
summed
over
is
said
to
be
a
dummy
(or
summation)
index,
since
a
replacement
by
any
other
symbol
does
not
affect
the
value
of
the
sum.
An
index
that
is
not
summed
over
in
a
given
term
is
called
a
free
(or
live)
index.
Note
that
in
the
same
equation
an
index
is
either
dummy
or
free.
Thus,
relations
(1.18)
can
be
1.1
Algebra
of
Vectors
written
in
a
more
convenient
form
as
which
defines
the
Kronecker
delta
dij.
The
useful
properties
hold.
Note
that
hi,
also
serves
as
a
replacement
operator;
for
example,
the
index
on
ui
becomes
an
j
when
the
components
ui
are
multiplied
by
bu.
The
projection
of
a
vector
u
=
uiei
onto
the
basis
vectors
e,
yields
the
jth
compo
nent
of
u.
Thus,
from
eq.
(1.20)
and
properties
(1.13),
(1.21)
and
(1.22)2
we
have
Taking
the
basis
{ei)
and
eqs.
(1.13),
(1.20),
(1.21)
and
(1.22)>,
the
component
expres
sion
for
the
dot
product
(1.10)
gives
which
is
commonly
used
as
the
definition
of
the
dot
product.
Thus,
we
may
derive
the
dot
product
of
u
and
v
without
knowledge
of
the
angle
between
u
and
v.
In
an
analogous
manner,
the
component
expression
for
the
square
of
the
length
of
u,
i.e.
(1.15~)~
is
Note
that
in
eqs.
(1.24)
and
(1.25)
one
index
is
repeated,
indicating
summation
over
1,2,3.
In
symbolic
notation
this
is
indicated
by
one
dot.
Cross
product.
The
cross
(or
vector)
product
of
u
and
v,
denoted
by
u
x
v
(in
the
literature
also
u
A
v),
produces
a
new
vector.
The
cross
product
is
not
commutative.
It
is
defined
as
uxv=(vxu)
,
(1.26)
uxv=o
o
u
and
v
are
linearly
dependent
,
(1.27)
(au)
x
v
=
u
x
(av)
=
a(u
x
v)
,
(1.28)
u.(vxw)
=v.(wxu)=w.(uxv)
,
(1.29)
~~(~+~)=(ux~)+(uxw)=uXV+UXW
.
(1.30)
6
1
Introduction
to
Vectors
and
Tensors
1.1
Algebra
of
Vectors
If
relation
(1.27)
holds
with
u
and
v
assumed
to be
nonzero
vectors,
we
say
that
vector
u
is
parallel
to vector
v.
From
eq.
(1.29)
we
learn
that
The
magnitude
of
the
cross
product
is
defined
to
be
It
characterizes
the
area
of
a
parallelogram
spanned
by
the
vectors
u
and
v
(see
Fig
ure
1.3).
The
righthanded
cross
product
of
u
and
v,
i.e.
the
vector
u
x
v,
is
perpendic
ular
to
the
plane
spanned
by
u
and
v
(6
is
the
angle,between
u
and
v).
Figure
1.3
Cross
product
of
vectors
u
and
v.
In
order
to
express
the
cross
product
in
terms
of
components
we
introduce
the
permutation
(or
alternating
or
LeviCivita
E)
symbol
E~~~,
which
is
defined
as
1
,
for
even
permutations
of
(z,~,
k)
(i.e.
123,23
1,
3
12)
,
1
,
for
odd
permutations
of
(2,
j,
k)
(i.e.
132,
213,
321)
,
(1.33)
0
,
if
there
is
a
repeated
index
,
with
the,properties
~~~k
=
sjkz
=
&kij,
~~~k
=
~i~~
and
c~~~
=
qik,
respectively.
Consider
the
righthanded
and
orthonormal
basis
{ei),
then
or
in
a
more
convenient
shorthand
notation,
with
(1.33),
With
relations
(1.21),
(1.34),
(1.35)
it
is
easy
to
verify
that
cijk
may
be
expressed
as
the
determinant
of
a
matrix
6i1
bi2
Ji3
(1.36)
where
we
have
introduced
the
square
brackets
1.1
for
a
matrix.
The
product
of
the
permutation
symbols
~ijk
E~~~
is
related
to the
Kronecker
delta
by
the
relation
With
(1.22j1
and
(l.22j3
we
deduce
from
(1.37)
the
important
relations
EXAMPLE
1.1
Obtan
the
coordinate
expression
for
the
cross
product
w
=
u
x
v.
Solution.
Taking
advantage
of
eqs.
(1
20)
and
(1
35)
we
find
that
w
=
u
x
v
=
UteZ
x
2,e3
=
uzwJ(e,
x
e,)
=
E2,kUzC3ek
=
wkek
,
(1.39)
with the
three
components
~1
=
212213

~3~2
,
(
1.40)
w2
=
~3~1
ULW~
,
(1.41)
Wg
=
U1Vz

UZW~
(1
42)
Consequently,
(u
x
V)

ek
equals
&,,au,W,.
H
Now,
using
expressions
(1.36).
(1.39)a,
the
vector
product
u
x
v
relative
to
{ek)
may
be
written
as
u
x
v
=
det
[
:I
]
=
~ijku.ivjeb
Wl
Vz
v3
8
1
Introduction
to
Vectors
and
Tensors
The
triple
scalar
(or
box)
product
(u
x
V)
.
w
represents
the
volume
V
of
a
paral
lelepiped
'spanned'
by
u,
v,
w
forming
a
righthanded
triad
(see
Figure
1.4).
uxv
4
Figure
1.4
Triple
scalar
product.
By
recalling
definitions
(1.29),
(1.10)
and
(1.32),
we
have
=
lullvsine
~W~COS~
+
base
area
height
Using
index
notation,
then,
from
eqs.
(1.20)
and
(1.35)
we
find
with
(1.21)
the
volume
V
to
be
Hence,
the
triple
scalar
product
(1.45)3
can
be
written
in
the
convenient
determinant
form
'1I1
v1
w1
(u
x
v)
.
w
=
det
["
::I
"I
(1.46)
Note
that
the
vectors
u,
v,
w
are
linearly
dependent
$and
only
$
thew
triple
scalar
product
vanlshes
(the
parallelepiped
has
no
volume).
The
product
u
x
(v
x
w)
is
called
the
triple
vector
product
and
may
be
verified
with
(1.39)4,
(1.38)],
(1.22)~
and
representations
(1.20)
and
(1.24)4
1.2
Algebra
of
Tensors
9
U
X
(V
X
W)
=
~tJb~z(~mnjum~n)ek
=
Ekt3Emn3UtumWnek
=
(bkrnbzn

bkn&m)uzvrn~nek
=
unukWnek

Um'umWkek
=
(u.w)v
(U.V)W
:
(1.47)
which
is
the
socalled
'backcab'
rule
wellknown
from
vector
algebra
Similarly,
(u
x
v)
x
w
=
(U

w)v

(v
w)u
.
(1.48)
The
triple
vector
product
is,
in
general,
not
associative,
i.e.
(u
x
v)
x
w
#
u
x
(v
x
w).
1.
Use
the
properties
(1.5),
(1.6),
(1.8)
and
(1.9)
to
show
that
ao=
o
:
OU
=
o
:
(a)u
=
a(u)
.
2.
By
means
of
(1.30)
and
(1.28)
derive
the
property
(au+pv)xw=a(uxw)+P(vxw)
.
3.
Prove
the
triple
vector
product
(1.48)
and
show
that
the
vector
(u
x
V)
x
w
lies
in the
plane
spanned
by
the
vectors
u
and
v.
1.2
Algebra
of
Tensors
A
secondorder
tensor
A,
denoted
by
A,
B,
C
.
.
.
(or
in
the
literature
sometimes
writ
ten
as
A,
B,
C
.
.
.),
may
be
thought
of
as
a
linear
operator
that
acts
on
a
vector
u
generating
a
vector
v.
Thus,
we
may
write
which
defines
a
linear
transformation
that
assigns
a
vector
v
to
each vector
u.
Since
A
is
linear
we
have
for
all
vectors
u,
v
and
all
scalars
a.
Since
most
tensors
used
in
this
text
are
of
order
two,
we
shall
often
omit
the
adjective
'secondorder'
.
10
1
Introduction
to
Vectors
and
Tensors
If
A
and
B
are
two
(secondorder)
tensors,
we
can
define
the
sum
A
+
B,
the
difference
A

B
and
the
scalar
multiplication
aA
by
the
rules
where
u
denotes
an
arbitrary
vector.
The
important
secondorder
unit
(or
identity)
tensor
I
and
the
secondorder
zero
tensor
0
are
defined,
respectively,
by
the
relations
Iu
=
uI
=
u
and
Ou
=
uO
=
o
for
all
vectors
u.
Note
that
tensor
0
maps
every
u
to
the
zero
vector
o.
Tensor
product.
The
tensor
(or
direct
or
matrix)
product
or
the
dyad
of
the
vectors
u
and
v
is
denoted
by
u
@
v
(some
authors
use
the
notation
uv).
It
is
a
second
order
tensor
which
linearly
transforms
a
vector
w
into
a
vector
with the
direction
of
u
following
the
rule
(u@v)w=u(v.w)=
(v.w)u
.
(1.53)
The
dyad,
not
to
be
confused
with the
dot
or
cross
product,
has
the
linearity
property
In
adhtion,
note
the
relations
(au+8v)~w=cr(u@w)+B(v@w)
:
(1.55)
U(V@W)
=
(u.
v)w=
w(u.v)
.
(1.56)
(u@v)(w@x)
=
(v~w)u@x=u@x(v~w)
.
(1.57)
A(u@v)
=
(Au)@v
.
(1.58)
\
Generally,
the
dyad
is
not
commutative,
i.e.
u
@
v
#
v
@
u
and
(u
@
v)
(w
@
x)
#
(w
@
X)(U
@
v).
A
dyadic
is
a
linear
combination
of
dyads
with
scalar
coefficients,
for
example,
a(u
3
v)
+
p(w
El
x).
Note
further
that
no
tensor
may
be
expressed
as
a
single
tensor
product,
in
general,
A
=
u
@
v
+
w
@
x
#
y
@
z.
Any
secondorder
tensor
may
be
expressed
as
a
dyadic.
As
an
example,
the
second
order
tensor
A
may
be
represented
by
a
linear
combination
of
dyads
formed
by
the
(carte&)
basis
{e,),
i.e.
We
call
A,
which
is
resolved
along
basis
vectors
that
are
orthonormal,
a
Cartesian
tensor
of
order
two
(more
general
tensors will
be
considered
in
Section
1.6).
The
nine
1.2
Algebra
of
Tensors
11
Cartesian
components
of
A
with respect
to
{ei),
represented
by
.4ij,
form
the
entries
of
the
matrix
[A],
i.e.
412
A13
PI
=
A21
422
A23
,
[
:::
A,,,
AH
1
where
by
analogy
with
(1.22)2
holds.
Relation
(1.60)
is
known
as
the
matrix
notation
of
tensor
A.
EXAMPLE
1.2
Let
A
be
a
Cartesian
tensor
of
order
two.
Show
that
the
projection
of
A
onto
the
orthonormal
basis
vectors
ei
is
according
to
Aij
=
ei
.
Aej
,
(1.62)
where
Aij
are
the
nine
Cartesian
components
of
tensor
A.
Solution.
Using
representation
(1
59)
and
properties
(1
.53)2,
(1.21)
and
(1.61);
we
find
that
Aej
=
Alk(e1
@
ek)ej
=
Alk(ek
.
e,)el
=
AlkSkjel
=
Aljel
.
(1.63)
On
taking
the
dot
product
of
(1.63)4
with
ei,
the
nine
Cartesian
components
Aij
are
completely
determined,
namely
ei

Aej
=
ei
.
Aljel
=
Alj(ei
.
el)
=
Alibil
=
A,j
=
(Aj,j
.
(1.64)
In
(1.64)
we
have
used
the
notation
(A)ij
for
characterizing
the
components
of
A.
W
If
the
relation
v

Av
2
0
holds
for
all
vectors
v
#
o,
then
A
is
said
to
be
a
positive
semidefinite
tensor.
If
the
stTonger
condition
v
.
Av
>
0
holds
for
all
nonzero
vectors
v,
then
A
is
said
to
be
a
positive
definite
tensor.
Tensor
A
is
called
negative
semi
definite
if
v
,
Av
<
0
and
negative
definite
if
v
.
Av
<
0
for
all
vectors
v
#
o,
respectively.
The
Cartesian
components
of
the
unit
tensor
I
form
the
Kronecker
delta
symbol
introduced
in eq.
(1.21).
Thus,
I
=
bijei
g
ej
=
e,
8
ej
.
(1.65)
12
1
Jntroduction
to
Vectors
and
Tensors
We
derive
further
the
components
of
u
@J
v
along
an
orthonormal
basis
{e,).
Using
representation
(1.62)
and
properties
(1.53)
and
(1.23)3,
we
find
that
where
the
coefficients
uiuj
define
the
nine
Cartesian
components
of
u
@
v.
Writing
eq.
(1
.66)4
in the
convenient
matrix
notation
we
have
Here,
the
3
x
1
column
matrix
[ui]
and
the
1
x
3
row
matrix
[vj]
represent
the
vectors
u
and
v,
while
the
3
x
3
square
matrix
represents
the
secondorder
tensor
u
8
v.
A
scalar
would
be
represented
by
a
1
x
1
matrix.


EXAMPLE
1.3
Show
that
the
linear
transformation
(1.49)
that
maps
v
to
u
may
be
written
as
where
index
j
is
a
dummy
index.
Solution.
Using
expressions
(1.201,
(1.59)
and
rules
(1.53),
(1.21)
and
(1.61),
we
find
from
(1.49)
that
A
replacement
of
the
dummy
index
k
in
eq.
(1.69)4
by
j
gives
the
desired
result
(1.68).
.
Dot
pioduct.
The
dot
product
of
two
secondorder
tensors
A
and
B;
denoted
by
AB,
is
again
a
secondorder
tensor.
It
follows
the
requirement
for
all
vectors
u.
1.2
Algebra
of
Tensors
13
The
summation,
multiplication
by
scalars
and
dot
products
of
tensors
are
governed
mainly
by
properties
known
from
ordinary
arithmetic,
for
example,
(aA)
=
a(A)
=
aA
;
(1.75)
(AB)C
=
A(BC)
=
ABC
:
(1.76)
A2
=
AA
:
(1.77)
(A+B)C
=
AC+BC
.
(1.78)
Note
that,
in
general,
the
dot
product
of
secondorder
tensors
is
not
commutative,
i.e.
AB
#
BA
and
also
Au
#
uA.
Moreover,
relations
A3
=
0
and
Au
=
o
do
not,
in
general,
imply
that
A,
B
or
u
are
zero.
The
components
of
the
dot
product
AB
along
an
orthonormal
basis
{ei),
as
intro
duced
in
(1.18),
read,
by
means
of
representation
(1.62)
and
relations
(1.70),
(l.f~3)~,
or
equivalently
For
convenience,
we
adopt
the
convention
that
a
repetition
of
only
one
index
be
tween
a
tensor
and
a
vector
(see,
for
example,
eq.
(1.68)
with
the
dummy
index
j)
or
between
two
tensors
(see,
for
example,
eq.
(1.80)1
with
the
dummy
index
k)
will
not
be
indicated
by
a
dot
when
symbolic
notation
is
applied.
Specifically
that
means
for
v
=
Au
(v,
=
Aijuj)
and
A
=
BC
(Atj
=
BzkCk3)
we
do
not
write
v
=
A.
u
and
A
=
B
.
C,
respectively.
'kanspose
of
a
tensor.
The
unique
transpose
of
a
secondorder
tensor
A
denoted
by
is
governed
by
the
identity
for
all
vectors
u
and
v.
The
useful
properties
1
Introduction
to
Vectors
and
Tensors
(aA
+
p~)~
=
aAT
+
/3~~
:
(1.83)
(AB)~
=
B~A~
:
(1.84)
(u8~)~
=V@U
(1.85)
hold.
Hence,
from
identity
(1.81)
we
obtain
e,
.
ATej
=
ej
.
Ae,,
which
gives,
in
regard
to
(1.621,
the
important
index
relation
(AT)ij
=
Aji.
Trace
and
contraction.
The
trace
of
a
tensor
A
is
a
scalar
denoted
by
trA.
Taking,
for
example,
the
dyad
u
@
v
and
summing
up
the
diagonal
terms
of
the
matrix
form
of
that
secondorder
tensor,
we
get
the
dot
product
u
.
v
=
uivi,
which
is
called
the
trace
of
u
8
v.
We
write
tr(u
8
v)
=
u
.
v
=
uivi
(1.86)
for
all
vectors
u
and
v.
Thus,
with
representation
(1.59)
and
eqs.
(1.86)1,
(1.21),
(1.61)
the
trace
of
a
tensor
A
with
respect
to
the
orthonormal
basis
{ei)
is
given
by
trA
=
tr(Aijei
8
ej)
=
Aijtr(ei
8
ej)

.e.)
=
A..J..

23
2
3
23
32
=
Aii
:
(1.87)
or
equivalently
trA
=
Aii
=
All
+
Azz
+
A33
.
(1.88)
We
have
the
properties
of
the
trace:
tr~~
=
trA
;
(1.89)
tr(AB)
=
tr(BA)
:
(1.90)
tr(A
+
B)
=
trA
+
trB
:
(1.91)
tr(aA)
=
atrA
.
(1.92)
In
index
notation
a
contraction
means
to
identify
two
indices
and
to
sum
over
them
as
dummy
indices.
In
symbolic
notation
a
contraction
is
characterized
by
a
dot.
A
double
contraction
of
two
tensors
A
and
B,
characterized
by
two
dots,
yields
a
scalar.
It
is
defined
in
terms
of
the
trace
by
A
:
B
=
tr(~~B)
=
tr(BT~)
=
tr(ABT)
=
tr(BAT)
=B:A
or
ABBA.
.
v
~3

v
23
(1.93)
1.2
Algebra
of
Tensors
The
useful
properties
of
double
contractions
I:A=trA=A:I
,
(1.94)
A
:
(BC)
=
(B~A)
:
c
=
(AC~)
:
B
:
(1.95)
A:
(u@v)
=u.Av=(u@v):A
:
(1.96)
(u@v):
(WBX)
=
(u.w)(v.x)
:
(1.97)
(ei
8
ej)
:
(ek
@
el)
=
(ei

ek)(ej
el)
=
'Jik631
(1.98)
hold.
Note
that
if
we
have
the
relation
A
:
B
=
C
:
B,
in
general,
we
cannot
conclude
that
A
equals
C.
The
norm
of
a
tensor
A
is
denoted
by
IA
(or
IIAll).
It
is
a
nonnegative
real
number
and
is
defined
by
the
square
root
of
A
:
A,
i.e.
.
/AI
=
(A
:
=
(A~~A,~)~~~
2
0
.
(1.99)
Determinant
and
inverse
of
a
tensor.
The
determinant
of
a
tensor
A
yields
a
scalar.
It
is
defined
by
the
determinant
of
the
matrix
[A]
of
components
of
the
tensor,
i.e.
Ail
A12
A13
detA
=
det[A]
=
det
[
2;;
i:;
i;;
]
:
(1.100)
with
properties
det(AB)
=
detAdetB
:
detAT
=
detA
.
A
tensor
A
is
said
to
be
singular
if
and
only
if
detA
=
0.
We
assume
that
A
is
a
nonsingular
tensor,
i.e.
detA
#
0.
Then
there
exists
a
unique
inverse
A'
of
A
satisfying
the
reciprocal
relation
AAI
=
I
=
AIA
.
(1.103)
If
tensors
A
and
B
are
invertible,
then
the
properties
(AB)1
=
B~A~
.
(1.104)
(Al)l
=
A
:
(1.105)
@A)l
=
l/aAl
>
(1.106)
(A')~
=
(AT)l
;
(1.107)
A2
=
A1~I
.
(1.108)
det(A')
=
(detA)'
(1.109)
16
1
Introduction
to
Vectors
and
Tensors
I
1.2
Algebra
of
Tensors
hold.
Subsequently,
in
this
text,
we
use
the
abbreviation
for
notational
convenience.
Orthogonal
tensor.
An
orthogonal
tensor
Q
is
a
linear
transformation
satisfying
the
condition
for
all
vectors
u
and
v
(see
Figure
1.5).
As
can
be
seen,
the
dot
product
u
v
is
invariant
during
that
transformation,
which
means
that
both
the
angle
8
between
u
and
v
and
the
lengths
of
the
vectors,
lul,
vI,
are
preserved.
Figure
1.5
Orthogonal
tensor.
Hence,
an
orthogonal
tensor
has
the
property
Q~Q
=
QQ~
=
I,
which
means
that
Q'
=
QT.
Another
important
property
is
that
det(QTQ)
=
(detQ)2
with
detQ
=
f
1.
If
detQ
=
+l
(I),
Q
is
said
to
be
proper
(improper)
orthogonal
corresponding
to
a
rotation
(reflection),
respectively.
Symmetric
and
skew
tensors.
Any
tensor
A
can
always
be
uniquely
decomposed
into
a
symmetric
tensor,
here
denoted
by
S,
and
a
skew
(or
antisymmetric)
tensor,
here
denoted
by
W.
Hence,
A
=
S
+
W,
while
In
this
text,
we
also
use
the
notation
symA
for
S
and
skewA
for
W.
Tensors
S
and
W
are
governed
by
properties
such
as
which
in
matrix
notation
reads
I
In
accord
with
the
notation
introduced,
the
useful
properties
hold,
where
B
denotes
any
secondorder
tensor.
A
skew
tensor
with
property
W
=
WT
has
zero
diagonal
elements
and
only
three
independent scalar
quantities
as
seen
from
eq.
(1.1
14)2.
Hence,
every
skew
tensor
W
behaves
like
a
vector
with
three
components.
Indeed,
the
relation
holds:
where
u
is
any
vector
and
w
characterizes
the
axial
(or
dual)
vector
of
skew
tensor
W,
with
property
I
w
=
(I/&)
IW~
(the
proof
is
omitted).
The
components
of
W
follow
from
(1.62),
with the
help
of
(1.118),
(1.20),
(1.35):
(1.21)
and
the
properties
of
the
permutation
symbol
W,
=
ei
.
We,
=
ei
.
(w
x
ej)
=
e,
.
(wkek
x
ej)
=
ei.
(~ik~k~leh)
=
LJk&kj16il


~kjiwk
=
&ijk~k
.
(1.119)
I
Therefore,
with
definition
(1.33)
we
have
where
the
components
Milz,
WI3,
WZ3
form
the
entries
of
the
matrix
[W]
as
character
ized
in
(1.1
14)2.
The
inversion
of
(1.1
1
9)7
follows
with
relations
(1
.38)2
and
(1
.22)2
after multipli
cation
with
the
permutation
symbol
~i,,
18
1
Introduction
to
Vectors
and
Tensors
which,
after a
change
of
the
free
index,
gives
finally
1
Wk
=

5..
W..
1
2V~
23
and
w=~.W..e
2
~k
ZJ~

(1.124)
Projection,
spherical
and
deviatoric
tensors.
Consider
any
vector
u
and
a
unit
vector
e.
With
reference
to
Figure
1.6,
we
write
u
=
ull
+
ui,
with
UII
and
UL
charac
terizing
the
projection
of
u
onto
the
line
spanned
by
e
and
onto the
plane
normal
to
e,
respectively.
Figure
1.6
Projection
tensor.
With
(1.53)
we
deduce
that
ull
=
(u.e)e
=
(e@e)u
=
P!U
,
(1
125)
+
P!
u1=uul~
=u(e@e)u=(~e@e)u=piu
,
(1.126)
"
p:
where
P!
and
Pt
are
projection
tensors
of
order
two
The
projection
tensor
P!
applied
to
any
vector
u
maps
u
into
the
direction
of
e,
while
Pi
applied
to
u
gives
the
projection
of
u
onto
the
plane
normal
to
e
(see
Figure
1.6).
A
tensor
P
is
a
projection
if
P
is
symmetric
and
Pn
=
P
(n
is
a
positibe
integer),
with
the
properties
P,'+P!
=I
,
pl
pll
=
pll
ee
el
11pl
Pe
Pe

e
9
PIP,'
=o
.
1.2
Algebra
of
Tensors
19
Every
tensor
A
can
be
decomposed
into
its
socalled
spherical
part
and
its
devia
toricpart,
i.e.
A
=
a1
+
devA
or
Aij
=
ab,.
+
devAij
(1.131)
Any
tensor
of
the
form
aI,
with
a
denoting
a
scalar,
is
known
as
a
spherical
tensor,
while
devA
is
known
as
a
deviator
of
A,
or
a
deviatoric
tensor.
The
deviatoric
op
erator
dev(a)
is
denoted
by
the
shorthand
notation,
i.e.
dev(0)
=
(8)

(1/3)tr(a)I,
or
dev(a)u
=
(*)i,

(1/3)(*)kkbij.
Computing
the
trace
of
(1.131)
we
deduce
with
(1.92),
(1.87)5,
(1.22)1
and
(1.
132)3
that
Aii
=
3(Aii/3)
+
devAii.
This
relation
yields
the
important
property
which
means
that
the
trace
of
the
deviator
of
A
is
always
zero.
1.
Let
u
be
any
vector
and
e
any
unit
vector.
Use
eqs.
(1.125),
(1.126)
and
the
triple
vector
product
(1.47)
to
show
that
u
can
be
resolved
into
components
parallel,
i.e.
ull,
and
perpendicular,
i.e,
ul,
to unit
vector
e,
according
to
U=
UII
+UJ.
=
(u.e)e+e
x
(u
x
e)
.
2.
Prove
eqs.
(1.57)
and
(1.58)
and
show
that
(u@v)A
=
u@
(A~v)
.
3.
Let
A
be
a
tensor
with
matrix
[A]
and
the
transpose
of
A
with
matrix
[AIT.
Show
that
[A~]
=
[A]~.
4.
Show
by
means
of
representation
(1.59)
and
properties
(1.98)
and
(1.61)
that
the
double
contraction
of
the
two
tensors
A
and
B
yields
the
component
form
Ai,
Bij
.
5.
Starting
from
eq.
(1.100),
verify
that
detA
=
E~~~A~~A~~A~~
and
show
property
(1.101).
6.
Consider
eq.
and
relations
(1.103),
(1.88)
in
order
to
obtain
the
property
A~:A=~
.
(1.134)
20
1
Introduction
to
Vectors
and
Tensors
7.
Given
that
S
is
a
symmetric
tensor
and
W
is
antisymmetric,
prove
property
(1.117).
8.
Two
tensors
A
and
B
are
given
in
the
form
of
their
matrix
representations
Verify
that
these
matrices
are
orthogonal.
Show
that
[A]
describes
a
rotation
and
[B]
a
reflection.
Give
a
geometrical
interpretation.
9.
Find
the
axial
vector
of
the
skew
tensor
W
=
1/2(u
8
v

v
8
u).
10.
Let
A
be
a
tensor
whose
matrix
is
Find
the
spherical
and
the
deviatoric
parts
of
A.
1.3
Higherorder
Tensors
In
the
following
we
discuss
higherorder
tensors.
Any
tensor
of
order
(or
rank)
n
may
be
expressed
in
the
form
As
(1.135)
shows,
a
tensor
of
order
n
has
3n
components
Ai,i2...i,,
provided
with
n
inhces
il,
i2,.
.
.
,
2,.
In
particular,
a
tensor
of
order
zero
has
3'
=
1
component
with
0
(no)
index,
which
simply
is
a
scalar
a
(i.e.
a
single
real
number).
A
tensor
of
order
one
has
3l
=
3
components
characterizing
any
vector
v,
i.e.
vi
as
indicated
in
(1.23).
A
tensor
of
order
two,
for
example,
A,
has
32
=
9
components,
i.e.
Aij
as
indicated
in
(1.60).
Tensor
of
order
three.
A
tensor
of
order
three
we
distinguish
by
the
notation
A,
B,
C,
.
.
.
According
to
(1.135),
d
may
be
expressed
as
where
Auk
characterizes
the
33
=
27
components
of
d.
1.3
Higherorder
Tensors
21
A
particular
example
for
a
thirdorder
tensor
is
the
socalled
triadic
product
of
three
vectors
U:
v;
w,
denoted
by
u
@
v
8
w,
with
the
properties
(u@v)@w=u@v@w
:
(1.137)
(u@v@w)x
=
(w.x)u@v
:
(1.138)
(u@v@w):
(x@y)
=
(v.x)(w.y)u
:
(1.139)
(u@v@w):I=
(v.w)u
.
(1.140)
Hence,
by
analogy
with
the
procedure
which
led
to
eq.
(1.62)
we
may
determine
the
components
of
$I
i.e.
Aijk
=
(ei
@
e,)
:
.4ek
.
(1.141)
The
double
contraction
of
a
thirdorder
tensor
A
with
a
secondorder
tensor
B
produces
a
vector.
It
is
denoted
by
d
:
B
and
may
be
found
with
representations
(1.136):
(1.59)
and
properties
(l.l39),
(1.21)
and
(1.61)
as
.A
:
B
=
AjkBi,(ei
@
ej
@
ek)
:
(el
@
em)
=
AijkBim(ej
.
el)(%
.
ern)%
=
AijkBlmbjldkmei
A.
B.

jk
jkei
.
(1.142)
EXAMPLE
1.4
Let
E
denote
the
thirdorder
permutation
tensor,
which
may
be
expressed
as
E
=
sijkei
3
ej
@
eh
.
(1.143)
Here,
~ijk
=
(ei
x
ej)
.
ek
are
the
33
components
of
E
whlch
we
have
introduced
as
the
permutation
symbol
(recall
eqs.
(1.33)
and
(1.35)).
Show
that
the
double
contraction
of
E
with
the
dyad
of
any
vectors
u
and
v,
i.e.
u
@
v,
gives
the
cross
product
of
v
and
u.
Solution.
With
expressions
(E)Uk
=
&i3k
and
(u
@
v)~,
=
ulum
we
may
write
E
:
(u
8
v)
=
&ijkulv,(ei
@
ej
@
ek)
:
(ei
@
em)
.
(1.144)
Using
properties
(1.139),
(1.21)
and
(1.22)2,
we
find
that
&
:
(U
8
V)
=
~~~~~~t'~b~[6km~
=
EijkUjUkei
:
(1.145)
which
is
a
vector
with
components
~ijkujuk.
Finally,
by
means
of
the
permutation
22
1
Introduction
to
Vectors
and
Tensors
property
E~~~
=
&kji
and
expressions
(1.43)2,
(1.26)
we
obtain
E:(u@v)=vxu
:
(1.146)
which
is
the
desired
result.
Tensor
of
order
four.
Any
tensor
of
order
four,
denoted
by
A:
B:
C:
.
.
.,
has
34
=
81
components,
i.e.
Aijkl;
Bijkl;
Cijkl:
Σχόλια 0
Συνδεθείτε για να κοινοποιήσετε σχόλιο