Lecture 4

Second order tensor

As we have discussed earlier, second order tensors also follow similar rules. Firstly, they should
contain 32 = 9 components and secondly, they should also follow some transformation rule as
follows.
 mn ’ = liml jh ij  
(4.1)
lim = cos ( i th old axis  –  mth new axis )

(4.2)
l jn =  cos   ( j th old axis  –  nth new axis )

(4.3)
m = 1, n = 1
 'mn =  11’

 11’  =  li1l j1 ij

11’ = li1l j1 ij + ........

=  li1l j1 ij {here, i and j are dummy indices}

i j

(4.4)
11, 22, 33 are the normal component of second order tensor. The second order tensor is
symmetric in nature.
 ij =  ji
(4.5)

Tensor as dyadic product of two vectors

As we have discussed earlier, the dyadic product of two vectors is a 2nd order tensor quantity.
For example, if v and w are two vectors, then
v w = (v11 + v2 2 + v3 3 ) (w11 + w2 2 + w3 3 )

= v1w111 + v1w21 2 + v1w31 3

+ v2 w1 21 + v2 w2 2 2 + v2 w3 2 3 (4.6)
+ v3 w1 31 + v3 w2 3 2 + v3 w3 3 3

We can write it in component form as

 v1w1 v1w2 v1w3 
 
 v2 w1 v2 w2 v2 w3 
v w v w v3 w3 
 3 1 3 2
v w = vi i w j j = vi w j i j

(4.7)
Transformation rule for vector
vi to vm ’ = limvi

(4.8)
w j to wn ’ = l jn w j

(4.9)
Dyadic product of both vectors (vm’ and wn’) gives the position of dyadic tensor according to
new rotated axis.
vm ’wn ’ = limvil jn w j = vi w j liml jn

(4.10)
Above equation proves that product of 2 vectors in second order tensor. Thus, we can write
any tensor  as

 =  ij i j
(4.11)

Mathematical operations for tensors

The following mathematical operations are possible for tensor quantities

Addition of tensors
Tensors of the same order can be added or subtracted as follows:
 +  =  ij i j +  ij i j

= ( ij +  ij ) i j

(4.12)
Similarly,
 −  =  ij i j −  ij i j

= ( ij −  ij ) i j

(4.13)
Here indices used in both tensors should be identical.

Multiplication of tensors
Multiplication operations are possible between two different order tensors. Some important
operations, which are used in this book, are as follows.
Multiplication of a tensor by a scalar
A scalar and a 2nd order tensor quantity can be multiplied as follows
a = (a ij ) i j

(4.14)
Multiplication of a tensor by a vector
A vector and a 2nd order tensor can be multiplied in three ways:

Dyadic product of a vector and tensor

Dyadic product of vector and tensor is a third order tensor which is not discussed here.
 v =  ij i j vk k =  ij vk i j k
(4.15)

Dot product of a vector and tensor

Dot operation reduces the order of resulting quantity by two. Hence, the dot product of a vector
and a tensor is a vector quantity. For example, if  is tensor and v is vector quantity, then

 . v =  ij i j . vk k =  ij vk i ( j . k )

(4.16)
Here, the nearest unit vectors take part in the dot operation
 . v =  ij vk i jk → Replace j by k or k by j

=  ij v j i = w = wi i Thus, wi =  ij v j

(4.17)
3
wl = t lj v j = åt lj v j
j=1

(4.18)
For example, if i=1, then
w1 = 11v1 + 12v2 + 13v3

(4.19)
Cross product of a vector and tensor
Cross operation reduces the order of resulting quantity by one. Hence, the cross product of a
vector and second order tensor is a second order tensor. For example, if  is tensor and v is
vector quantity, then
  v =  ij i j  vk k

=  ij vk  i ( j   k )

=  ij vk  i ( jkn n )

=  ij vk  jkn i k

(4.20)
This is a second order tensor quantity.

Multiplication of two tensors

Four types of multiplication operations can be performed between two second order tensor
quantities.

Dyadic Product of two tensors

Dyadic product of two second order tensors is a fourth order tensor quantity. It is not discussed
here.
 → 4th order tensor
(5.1)

Cross product of two tensors

Cross product of two second order tensors is a third order tensor quantity and is not discussed
here.
  → 3rd order
(5.2)

Dot product of tensors (Tensor product)

Dot reduces the order of resultant quantity by two. Thus, the dot product of two second order
tensors is a second order tensor quantity. If  and  are two second order tensors, then

 . =  ij i j . kl k l
 =  ij kl  i ( j . k )  l

(5.5)
Here the order of indices should not be changed and the dot product should take the two nearest
indices.
 . =  ij kl jk i l
(5.6)
Replace k by j (or you may also replace j by k) (compaction operation)
 =  . =  ij jl i j

= il i l where il =  ij jl

For example,
11 = 1 j j1 = 11 11 +  12 21 +  13 31

(5.7)

Double dot product or Scalar product of second order tensors

Double dot operation reduces the order of resultant quantity by four. Thus, the double dot
product of two second order tensors is a scalar quantity. If  and  are two second order
tensors, then
 :  =  ij i j :  kl k l

=  ij kl ( j . k )( i . l )

(5.8)
First dot operation should take place between the two nearer vectors and the next dot operation
should take place in between two remaining vectors. Hence,
 :  =  ij kl jk ( i . l ) Replace k by j

=  ij jl il Replace l by i

=  ij ji

(5.9)
This is a scalar quantity

Del operations involving tensors & dyads

Divergence of a second order tensor field

If  is a second order tensor then divergence of the tensor field is

    jk
. =   i  . ( j k jk ) = ( i . j )  k (Apply compaction operation)
 xi  xi

(5.10)
 ik
= k
xi
 ik
. =  k = w = wk  k
xi
(5.11)
 ik
wk =
xi
(5.12)

Now we will try to prove some examples of vector and tensor identities.
Examples:
(1) Prove: . ( sv ) = s.v + s.v

Order = 1 + 1 – 2 = 0  
LHS
= .( sv)
(5.13)

= i .( sv j j )
xi


= ( sv j ) i . j
xi


= ( sv j ) ij
xi


= ( svi )
xi
vi s
=s + vi
xi xi
(5.14)
RHS
= s.v + s.v
(5.15)
s 
= i .v j j + s i .v j j
xi xi

s v
= vj  ij + s j  ij
xi xi

s v
= vi +s i
xi xi
(5.16)

From Eq. (5.14) and Eq. (5.16)

 LHS = RHS

(2) . ( v w ) = v.w + w.v

LHS
= . ( v w )

(5.17)

= i .(v j j wk  k )
xi


=
xi
( v j wk ) ( i . j ) k

=
xi
( v j wk )  ij k

= ( vi wk )  k
xi

  v 
=  vi wk + wk i   k
 xi xi 

(5.18)

RHS
= v.w + w.v
(5.19)
  
= vi i . j wk  k + wk  k j .vi i
x j x j

 
= vi ij wk  k + wk  k  ji vi
x j x j

 w v 
=  vi k + wk i   k
 xi xi 

(5.20)

From Eq. (5.18) and Eq. (5.20)

 LHS = RHS

(3) .( .v) = v.(. ) +  : v if  is a symmetric second order tensor

LHS

= k .( ij i j .vl l )
xk
(5.21)

= k .( ij vl i jl )
xk


= k .( ij v j i )
xk


= .( ij v j ) ki
xk


= ( ij v j )
xi

v j  ij
=  ij + vj
xi xi
(5.22)
RHS
   
  
= vl l .   k . ij i j  +  ij i j :   k  (v  )
 xk   l l
 

xk 

(5.23)
  ij  v
= vl l .   ki j  +  ij il jk l
 xk  xk
  ij vl
= vl  lj +  ij il
xi x j

 ij vi
= vj +  ij
xi x j

 ij v j
= vj +  ji
xi xi

 ij v j
= vj +  ij (  is a symmetric tensor)
xi xi
(5.24)

From Eq. (5.22) and Eq. (5.24)

 LHS = RHS