INTRODUCTION

1.1 Tensor and Continuum Mechanics

Continuum mechanics is the branch of physical science which deals
with the deformation of continuous systems under the action of
external forces. External forces can be either applied loads or
environmental effects. The deformation field can be stress, strain etc.
The physical laws should be independent of any particular systems
used in describing them mathematically, which leads the
requirements of tensor analysis. A tensor is a mathematical entity
which has existence independent of coordinate system, however,
certain tensor quantities can be represented by coordinates systems.
Components of tensor in different coordinate system are determined
by transformation laws. Tensor has great applicability in the areas of
deferential geometry, mechanics, elasticity, hydrodynamics,
electromagnetic etc.
All vectors are, technically, tensors, but tensors are not vectors. This is
to say, tensors are a more general object that a vector (strictly speaking
though, mathematicians construct tensors through vectors). Tensors
are technically defined through two different objects:
a. Vector
b. Duals
Vectors are just objects for which you know what it means to add any
two of them (vector addition), and what it means to scale-change it
(scalar multiplication). Duals, likewise, have all the same notions,
except that they can operate on vectors and return scalars. Examples
are in order: The most prototypical examples are Euclidean vectors -
points of space. Examples of duals would be the magnetic potential
vector (It's actually not a true vector) or the gradient operator. The
2 Fundamentals of Solid Mechanics

most important property, when you add a few other suitable

assumptions, is that duals and vectors transform in particular
manners under a change of coordinates. These are the properties that
physicists and scientists are most often concerned.
In three dimensional spaces, a point is represented by set of three
independent number called coordinates. It can be determined by
frame of references called coordinates system. The components are
transformed from one rectangular system to another rectangular
system called Cartesian tensors. Cartesian tensors are used in the
following chapters unless stated otherwise.
The fundamental property of tensors is that, if a tensor is identically
equal to 0 in one coordinate system, it is equal to 0 in any coordinate
system. The reason that is important is that it means that if a tensor
equation, A= B where A and B are tensors, is true in one coordinate
system (so that A- B= 0), then it is true in any coordinate system (A-
B= 0 in any coordinate system so A= B in any coordinate system).
The tensors can be classified based upon the number of components
they have that is called rank or order. In three dimensional Euclidean
systems a tensor has 3N components in which N is order. Tensor of
zero order has only one component and is a scalar. A tensor of order
one has three components. In Euclidean space it has three dimensions
and called vector. Second order tensor has nine components and is
called dyadic. Tensor of third order has twenty seven components is
called triadics. Forth order tensor has eighty one components is called
tetradics. They appear while establishing constitutive relationships.

1.2 Basics
1.2.1 Indicial Notation
Components of tensors can be represented by indicial notations. In
tensor notation the indices either in the form of subscripts or
superscripts are appended to the generic or kernel letters to represent
the component of tensor.
 Introduction 3

Let x1, x2, x3, …………..xn are the variables which can be represented
by xi. Aij will represent 32=9. A11, A12, A13, A21, A22, A23, A31, A32, A33.
Bijk will represent 33=27. In general if a term has r number of indices
where each of these indices takes the values from 1 to n, then this
would represent nr quantities.
   
V  V1 ,V2 ,V3   v1e1  v2 e2  v3 e3 (1.1)

In 3-D space a vector has three components and it is useful to

abbreviate them by a single component with generalized index vi .

1.2.2 Free Index

The key concept in indicial notation is that of an index. Consider a
   
vector v  V1 ,V2 ,V3   v1e1  v2 e2  v3e3 . The components of the vector
are indexed in this case by the coordinate labels 1, 2 and 3. vi is
indexed by the numbers 1, 2 and 3.
When doing arithmetic or algebra on vectors, we have to do the same
thing to each component. Rather than going explicitly through the
math for each component, we can work with a generic component
indexed with a free index. The name of a free index must remain the
same throughout a calculation. It should not be changed midway
through. vi, for example, must remain vi throughout a calculation if i is
a free index. We cannot suddenly say that vi = vj, since j is another free
index which can also take on any of the three coordinate labels.
1.2.3 Dummy Indices
An index which does not appear in an equation after a summation is
carried out is called a dummy index. Because a dummy index does
not appear in the final result, we can change its name to whatever is
convenient, for example, aibi  a jb j .

1.2.4 Summation Convention

 
If A and B are two vectors then their multiplication can be done by
as following.
 
A.B  A1B1  A2 B2  A3B3  Ai Bi (1.2)
4 Fundamentals of Solid Mechanics

   
A  A1e1  A2 e2  A3e3 (1.3)
   
B  B1e1  B2 e2  B3 e3 (1.4)

Whenever a subscript occurs twice in the same term it will be

understood that the subscript is to be summed from 1 to 3. For
example
Aii  A11  A22  A33 (1.5)

Ajj  A11  A22  A33 (1.6)

 kk  11   22   33 (1.7)

The repeated subscripts are called dummy indices. The subscript

which occurs once in each terms of an expression or equation is called
free index. For example, aij x j  bi , where i is free index and j is called
dummy index. Some examples of vectors, its components and
equivalent indicial notations are shown in the table 1.1 below.
Table 1.1 Equivalent vector notations

Vectors Components Indicial notations


v v1 , v 2 , v3 vi
 
vu v1u1  v2 u2  v3u3 v i ui

   
  
y1 y 2 y 3 y i

1.2.4 Necessary Functions

There are two functions which are extremely useful in indicial
notation - the Kronecker δ function and the Levi - Civita  function.

1.2.4.1 The Kronecker δ

The Kronecker δ (delta) function is defined very simply. It is a
special matrix written as  ij and is defined as
 Introduction 5

1 0 0
 ij  0 1 0 (1.8)
0 0 1

 ij  0 if i  j and  ij  1 if i  j (1.9)

Thus, the summation convention is expressed as

 ii   jj   kk  11   22   33  1  1  1  3 . Furthermore, for a
quantity  ij A j substituting property of Kronecher delta can be
expressed by summing on one of the indices
 ij Aj   i1 A1   i 2 A2   i3 A3 (1.10)

Looking at the properties of the Kronecher delta, only the terms

of right hand side of equation 1.15 is non-zero, thus this can be
expressed by
 ij A j  Ai (1.11)

1.2.4.2 The Levi-Civita 

The Levi-Civita  (epsilon) function is rather more complicated,
since it is a function of three free indices, meaning there are 27
possible combinations. However, it can be summarized as
 ijk  0 if any of i, j or k are same
 ijk  1 if ijk is an even permutation of 123,
 ijk  1 if ijk is an odd permutation of 123.
By even permutation i mean that ijk is one of 123, 231 or 312. We
see that these three combinations require an even number of
swaps of indices to return them to 123, and that they are cyclic
permutations of each other. For example, if we swap the first two
indices in 312, we get 132. If we then swap the last two, we get
123, a total of two swaps, meaning that 312 is an even
permutation of 123. Just as there are three cases where  ijk is +1,
there are three cases where it is −1. In these cases, ijk is one of 132,
321 or 213. You should be able to convince yourself that each of
these combinations can be returned to 123 with and odd number
6 Fundamentals of Solid Mechanics

of permutations. For example, 321 → 231 → 213 → 123. These

three combinations are cyclic permutations of each other. The
property of  ijk can be written as
 ijk   ikj (1.12)

1.2.4.3    Relationship
The product of permutation symbol  ijk  mnk can be expressed
with Kronecher delta
 ijk  mnk   mj  jn   in jm (1.13)

Also interchanging sign of  ijk ,

 ijk  mnk   ijk  kmn   kij kmn   kij mnk (1.14)

This identity can be proved exhaustively, but we won't do that here

since there are 34 = 81 possible combinations of the free indices j, k, l
and m, each of which has three terms in the implied sum over i.
Instead, let's observe some features of the identity.
First of all, there is a dummy index i on the left hand side; this
index is repeated, so it's summed over and does not appear on the
right hand side. Any time you see two  symbols with the same index,
you can use this identity. It may be a little easier to use if you first put
the  symbols in the form above using cyclic permutation as shown in
equation 1.12.
Second, as a mnemonic device, notice that the indices on the deltas in
first term on the right hand side are parallel, while those on the deltas
in the second term are crossed. The diagram shown below may be
helpful on understanding this.

− δjm
δjl δkm δkl

j k j k

l m l m

The first delta uses the first free index from each epsilon; the section
 Introduction 7

uses the second free index from each epsilon. The third delta uses the
first free index from the first epsilon and the second from the second,
while the fourth uses the converse. Of course, within each term, the
delta may be written in either order, and within each delta, the indices
may be written in any order.

1.3 Properties of Tensors

1.3.1 Equality
Two tensors A and B are defined to be equal when their respective
components are equal. The condition of equality is

Aij  Bij (1.15)

1.3.2 Addition
Aij  Bij  C ij (1.16)

B and C should have to be same order or rank. Thus A also will of

same rank.

1.3.3 Multiplication
i. Vector multiplication by scalar quantity 
Aij  Cij (1.17)

ii. Scalar product

Scalar (dot) product of two vectors A i and B j is expressed by

Ai B j  C ij (1.18)

Cij  AB cos  (1.19)

For same unit vector Ai the dot product can be expressed by

Kronecher delta Ai A j   ij . It is an important property of matrix
operation in structural mechanics.
8 Fundamentals of Solid Mechanics

Furthermore, for a quantity  ij A j substituting property of

Kronecher delta can be expressed by summing on one of the
indices
 ij Aj   i1 A1   i 2 A2   i3 A3 (1.20)

Looking at the properties of the Kronecher delta, only the terms

of right hand side of equation 1.15 is non-zero, thus this can be
expressed by
 ij A j  Ai (1.21)

iii. Cross product

Vector cross product is expressed by permutation symbol.
Permutation
Permutation symbol  (epsilon) is defined by  ijk is called Levi-
Civita tensor.
 ijk  1 , if the numerical values of ijk appears as in the order of
12312, and if the numerical values of ijk appears in the reverse
order as in the sequence 32132,  ijk  1 . If the numerical values
of ijk appear in any other sequence,  ijk  0 . Where the indices

take the values i, j, k  1,2,3... Cross product of two vectors

 
A and B can be expressed permutation symbol as in the
following equation.
 
AXB   ijkAi Bj (1.22)

Also from the definition that interchange of any subscripts in

ijk cause a sign change so that  ijk   kji   kij   ikj and for
repeated subscript the values becomes zero. For example,
113   221  133   233  0
iv. Triple scalar product
Scalar triple product of three vectors can be expressed as
 Introduction 9

     
A.BXC  AXB.C  A , B,C 
  

A.BXC  Ai ei . Bj e j XC k ek 
 Ai ei . jkq BjC k eq

  jkq Ai BjC k iq

  ijk Ai BjC k (1.23-18)

v. Triple cross product

It can be expressed with the help of  and  relationship

   Relationship
The product of permutation symbol  ijk  mnk can be expressed
with Kronecher delta
 ijk  mnk   mj  jn   in jm
Also interchanging sign of  ijk ,

 ijk  mnk   ijk  kmn   kij kmn   kij mnk (1.24-19)

10 Fundamentals of Solid Mechanics

1.4 Transformations of Coordinates

Let y1y 2 y 3 and y1' y'2 y'3 are unprimed and primed systems of two
right handed Cartesian coordinates with common origin O as shown
in figure 1.1. The symbol a ij denotes the cosines of angle between ith

unprimed and jth primed coordinate systems. Thus, aij  cos yi , y'j is  
relative orientation of individual axes with respect to each other.
Direction cosines of axis y1' with respect to y1y 2 y 3 are a11 , a12 and a13 .
From the figure 1.1, the coefficients can be determined, a11  cos 11 ,
a12  cos 12 and a13  cos 13 respectively. Similarly, direction cosines of
y'2 with respect to y1y 2 y 3 are a 21 , a22 and a23 , and direction cosines of
y '3 with respect to y1y 2 y 3 are a 31 , a32 and a33 .

Consider a vector v has three components vi and v'i (where i=1, 2, 3)
in two systems of axes - primed and unprimed, which are expressed
by the following equations.

v1'  a11 v1  a21 v2  a31 v3 (1.20)

v'2  a12 v1  a22 v2  a32 v3 (1.21)

v'3  a13 v1  a23 v2  a33 v3 (1.22)

y3
y3'

33
32

31 29 y2'

13 32
y2

12
11
21
y1
y1'
 Introduction 11

Fig. 1.1 Rotation of axes

In indicial notation system, these equations can be written as

v'i  a jiv j (1.23)

The relation between two frames of references can be written as in

table 1.2
Table 1.2 Indicial notations

Axes y1' y'2 y'3

y1 a11 a12 a13

y2 a 21 a22 a23

y3 a31 a32 a33

If the roles of primed and unprimed vectors are interchanged in the

above development the similar equation can be written as

vi  aij v'j (1.24)

Here equation 1.23 and 1.24 are called transformation law of first
order tensor. In equation 1.23 free index appears second while in
equation 1.24 it appears at first. Substituting equation 1.23 into
equation 1.24 using appropriate dummy indices and can be written as
vi  aij akj vk (1.25)

The equation 1.25 yields the identity vi  vi such that a ija kj where
 ik is identical to Kronecher delta. The Kronecher delta is a special
matrix written as  ik and is defined as

1 0 0
 ik  0 1 0 (1.26)
0 0 1
12 Fundamentals of Solid Mechanics

 ik  0 if i  k and  ik  1 if i  k (1.27)

 ik acts as an operator is called substitution tensor is denoted as

aij akj   ik (1.28)

The equation 1.28 has nine equations when expanded are called
orthogonality condition of aij .

Proof of orthogonality of transformation matrices

Ai   ij A j (1.29)

From transformation laws any vector can be represented by

A'i  aij A j
(1.30)
Ar  asi A's

In order not to repeat the indices, substituting Ai by asi A's the eq. 1.30
can be written

a'si A's   ij A j (1.31)

a'si asj Aj   ij Aj (1.32)

asi asj A j   ij A j (1.33)

Thus, comparing left and right hand side

asi asj   ij (1.34)

Alternately,

A'i   ji Aj (1.35)

ais A's   ji Aj (1.36)

ais a js A j   ji Aj (1.37)

In matrix form

aaT  I  (1.38)
 Introduction 13

As a special case, direction cosine of equal angle  rotation

 cosθ sinθ 0
 
aij   sinθ cosθ 0 (1.39)
 0 0 1


1.5 Some Important Operations on Tensors

In general the partial differentiation of tensors field of any order raises
the order of tensor. Let  , u  ij be the tensor fields of order zero, one
and two respectively. Let yi , y'i be the position vector of a point with
respect to unprimed and primed system of axes.
Then

y'i  a ji y j and y i  aij y'i

Here

y i y'i
aij   (1.40)
y j y 'j

Gradient of scalar function  in unprimed coordinate system

  y k 
  aik (1.41)
y i y k y'i y k
 
Gi  aik G'k where G'k  , Gk  (1.42)
y k
'
y k

Thus, see that the gradient of a scalar function transform according to

the transformation rule of first order tensor that is vector. From vector
calculus operator can be expressed as
       
  e1  e2  e3  ek (1.43)
y1 y 2 y3 y k
i. Gradient of scalar field
       
  e1  e2  e3  ek (1.44)
y1 y 2 y 3 y k
14 Fundamentals of Solid Mechanics


   k (1.45)
k
y k
ii. Divergence
 
divA  .A
         
  e1  e2  e3 .A1e1  A2 e2  A3 e3  (1.45)

 y1 y 2 y 3 

A1 A2 A3  A

   i ,i (1.46)
y1 y 2 y 3
iii. Curl of a vector field A
e1 e2 e3
    
Curl A  XA = =  ijk Akj (1.47)
i i y1 y 2 y 3
A1 A2 A3 i

iv. Gauss divergence theorem

It states that the volume integral of the divergence of a vector

field A taken over any volume V is equal to the surface integral

of A taken over the closed surface surrounding the volume in
terms of components.

Consider an arbitrary continuously differentiable tensor A field
defined by finite region of physical space. Let V be the volume
and S be the surface area enclosed in the region as shown in
figure 1.2. The divergence theorem establishes a relationship

between the surface integral of tensor A associated with the
volume as the integrand to the volume integral for which a
coordinate derivative of tensor is an integrand.
 Introduction 15

y2 ni

dSi = nidS

V
O
y2

y1
Fig. 1.2 Volume with infinitesimal surface

  
 AdV   A.ndS (1.48)
V S

The left hand side can be expressed as

 A A A 
   1  2  3 dy1dy 2 dy3
V  y1 y 2 y3 
  A1dS1  A2 dS2  A3dS3 
S

  A1dS1  A2 dS2  A3dS3 

S

Thus,
Ai
 dV   Ai dSi
V yi S

 Ai ,i dV   Ai ni dS (1.49)
V S

v. Stroke theorem
It states that the line integral of a vector function round a closed
curve is equal to the normal surface integral of the curl of the
vector function over any surface enclosed by that curve. The
Green theorem relates the integral over its bounding surface,
16 Fundamentals of Solid Mechanics

whereas Stroke’s theorem relates an integral over an open surface

to a line integral around the bounding curve of the surface. Thus,
as shown in the figure 1.3, the cut C be the bounding space curve
to the surface S. If ni is the outward normal to the surface S, and
vi is any vector field defined on S and C, Stroke theorem can be
written as

y2
ni
C
dSi = nidS

yi

dy2
O
y2

y1

Fig. 1.3 Boundary space curve with tangential vector

  


 A.dy   n. XA dS 
 Ai dy i   ni ijk A jk dS (1.50)
S

Examples
1. Use ɛijk ɛkpq = δipδjq – δiqδjp to prove that ɛijk ɛkjp Aip is an
invariant under coordinate transformation.
Solution

We have, ɛijk ɛkjp = δijδjp – δipδjj

= δijδjp – 3 δip
= δip – 3 δip
= – 2 δip
 Introduction 17

Now, that ɛijk ɛkjp Aip = –2 δip Aip

= – 2[δi1 Ai1 + δi2 Ai2+ δi3 Ai3]
= –2 [δ11 A11 + δ22 A22+ δ33 A33]
= –2[A11+ A22+ A33]
= –2I1
Hence, ɛijk ɛkjp Aip is an invariant under coordinate transformation.

2. If the vector Vi is given in terms of base vectors a, b, c by Vi =

ijkvibjck
aai + bbi +gci show that a = ai + bbi + gci
pqrapbqcr

Solution
Given, Vi = aai + bbi +gci
Indicial notation,
V1 = aa1 + bb1+gc1
V2 = aa2 + bb2+gc2
V2 = aa3 + bb3+gc3
Using Cramer's rule let us find out a, b and g.

 v1 b1 c1
 v2 b2 c2
 v3 b3 c3
a = ,
 a1 b1 c1
 a2 b2 c2
 a3 b3 c3

a1 v1 c1 a1 b1 v1

a2 v2 c2 a2 b2 v2 
a3 v3 c3 a3 b3 v3 
b= , g=
 a1 b1 c1   a1 b1 c1 
 a2 b2 c2   a2 b2 c2 
 a3 b3 c3   a3 b3 c3 
ijkvibjck
Now, a = ,
pqrapbqcr
18 Fundamentals of Solid Mechanics

ijkaivjck
b= ,
pqrapbqcr
ijkaibjvj
g=
pqrapbqvr
ijkvibjck
Hence, a = a + bbi + gci
pqrapbqcr i
3. Prove the following identities:
a. σij σjk σki = σij' σjk' σki'
Solution
Let, p, q and t be the arbitrary assumed transformation axes.
Then fron stress transformation relation,
σij σjk σki = (ail ajp σip') (ajn akq σnq') (aks ait σst')
= ail ait σip' ajn ajp σnq' akq aks σst'
= δlt σlp δnp σnq δqs σst
= σtp σpq σqt
Here, p,q and t are arbitrary assumed axes, replacing p,q,t by
i,j,k results same so,
σij σjk σki = σij' σjk' σki'
b. ɛijk det Aij = ɛlmn Ail Ajm Akn

Solution
Let, A en = Alm (el em).en = Alm el δmn = Aln el

 a+x b c 
 a b+y c 
 
 a b c+z 
 A11 A12 A13 

(A el).[ (A em) x (A en)] = A21 A22 A23

 
 A31 A32 A33 
= det A
 Introduction 19

= det Aij
det A = A11 (Am2 em x An3 en)
= A11 Am2 An3 el.( em x en)
= ɛlmn A11 Am2 An3
= ɛlmn A1l A2m A3n
where,
ɛlmn = el.( em x en)
Now,

 A1i A1j A1k 

ɛijk det A = A2i A2j A2k 
 A3i A3j A3k 
This determinant can also be written as:

 A1l A1m A1n 

 A2l A2m A2n  = ɛlmn Ail Ajm Akn
 A3l A3m A3n 
c. ɛijk ɛilm = δjl δkm – δjm δkl

Solution

 A11 A12 A13 

Let, det A =  A21 A22 A23 
 A31 A32 A33 
An interchange of rows and column causes change in sign.

 A21 A22 A23   A12 A11 A13 

 A11 A12 A13  =  A22 A21 A23  = - det A
 A31 A32 A33   A32 A31 A33 
For an arbitrary number of i, j and k for rows interchange,

 Ai1 Ai2 Ai3 

 Aj1 Aj2 Aj3  = - ɛijk det A
 Ak1 Ak2 Ak3 
For any arbitrary value of column interchange with i, l and m,
20 Fundamentals of Solid Mechanics

 A1i A1l A1m 

 A2i A2l A2m  = - ɛilm det A
 A3i A3l A3m 
Then for both rows and column interchange by arbitrary values

 Aii Ail Aim 

 Aji Ajl Ajm  = ɛijk ɛilm det A
 Aki Akl Akm 
Setting Aij = δij then det A = 1 for R.H.S only and let put the
other term of determinant in L.H.S.

 ii il im 

ɛijk ɛilm = ji jl jm 
 ki kl km 
Expanding R.H.S,
δii (δjl δkm – δkl δjm) - δil (δji δkm – δki δjm) + δii (δji δkl – δki δjl)
Using δij δjk = δik,
ɛijk ɛilm = δjl δkm – δjm δkl

 Exercise
1. Show by direct expansion that the box product    ijk ai b j ck is
equal to the determinant
a1 a2 a3
b1 b2 b3
c1 c2 c3

1
2. Prove the following identity 6 ɛijk ɛilm Ail Ajm Akn = det Aij

??
 Introduction 21

1.1 Tensor and Continuum Mechanics .......................................... 1

1.2 Indicial Notation and Summation Convention ......................... 2

1.2.1 Summation Convention.................................................... 3

1.3 Properties of Tensors ................ Error! Bookmark not defined.

1.3.1 Equality ............................................................................. 7

1.3.2 Addition ............................................................................ 7

1.3.3 Multiplication ................................................................... 7

1.4 Transformations of Coordinates ............................................ 10

1.5 Some Important Operations on Tensors ................................ 13

Exercise ........................................................................................... 166