Professional Documents
Culture Documents
T = ∑∑ Tij ê i ê j
i j
or simply
T = Tij ê i ê j
M.5
ê i ê j
The unit vector pairs are known as unit dyads.
2. Algebra of Vectors and Tensors
For an orthogonal set of unit vectors such that ei × e j = ek for cyclic i, j , k using
either LH or RH rule
A × B = e ijk A j B k ê i
M.8
e
where ijk is the alternating symbol defined as
eijk =
0 if any two of ( i, j, k ) are same
= +1 if { i, j , k } is a cyclic permutation of {1, 2, 3}
= -1 otherwise
Therefore an outer or cross product of two vectors is a vector defined as,
C = A×B
Ci = A j Bk − Ak B j i ≠ j ≠ k
, and ( i, j, k ) must be a cyclic combination M.9
A double dot product or double contraction of two second rank tensors is a scalar
defined as,
P Q = Pij Q rs δ jr δ is = Pij Q ji
M.14
3.1 Gradient r
For a scalar function φ ( r , t ) the gradient represents the information regarding the
instantaneous spatial rate of change along the coordinate directions. The gradient is
conveniently defined through the del operator given as,
∂
∇ ≡ ê i
∂ si , M.15
where si are related to the generalized coordinates xi as
dsi = hi dxi .
Physically dsi represents the components along the coordinate direction of the
differential line element dl i.e. dl = ds i ê i
∂
∇ ≡ ê i
hi∂ xi M.16
The gradient of a scalar function ϕ is defined as,
∂φ
∇φ ≡ ê i
hi∂ xi M.17
The gradient of a scalar is a vector.
The other important use is in finding the direction of the local normal to a given
surface. A surface in 3D/2D can be expressed as a scalar function relation
r
φ ( r ) = Const.
Ex. x + y + z = 1.0
2 2 2
spherical surface
The direction of local normal to the surface is given by the vector ∇φ and therefore
the local unit normal is given as,
∇φ
n̂ =
∇φ
M.19
∂
∇ × A = ê j × ( A k ê k )
h ∂x
j j or,
∂ ( A k ê k )
∇ × A = ê j ×
h ∂x
j j or,
1 ∂A k
∇×A = ( ê j × ê k ) + A k ê j × ∂ ê k
h j ∂x j h j ∂x j
. M.22
1 ∂A k
∇ × A = e ijk ê i + A k ê j × ∂ ê k
h ∂x h j ∂x j
j j M.23
∂A ∂A j
∇×A = k − ê i
∂x ∂x
j k
∂
∇A = ê i ( A j ê j )
h i ∂x i or,
∂A j ∂ê j
∇A = ê i ê j + A j
h i ∂x i h i ∂x i
or,
∂A j ∂ê j
∇A = ê i ê j + A j ê i
h i ∂x i h i ∂x i . M.24
For Cartesian Coordinates,
∂A j
∇A = ê i ê j
∂x i M.25
∂
∇ ⋅ T = ê i ⋅ ( T jk ê j ê k )
h i ∂x i or,
∂ Tjk ∂ ê j ∂ ê k
∇⋅T = (ê i ⋅ ê j )ê k + Tjk (ê i ⋅ ) + Tjk (ê i ⋅ ê j )
h i ∂x i h i ∂x i h i ∂x i M.26
∂ Tik ∂ ê j ∂ ê k
∇⋅T = ê k + Tjk (ê i ⋅ ) + Tik
h i ∂x i h i ∂x i h i ∂x i M.27
∂ Tik
∇⋅T = ê k
∂x i M.28
The divergence of the dot product of a rank 2 tensor with a vector can be expressed
as,
∂
∇ ⋅ T ⋅ A = ê i ⋅ ( T jk A k ê j )
h i ∂x i or,
∂ (T jk A k ) ∂ ê j
∇⋅T⋅A = (ê i ⋅ ê j ) + T jk A k (ê i ⋅ )
h i ∂x i h i ∂x i or,
∂ (T jk ) ∂ (A k ) ∂ ê j
∇ ⋅ T ⋅ A = Ak (ê i ⋅ ê j ) + Tjk (ê i ⋅ ê j ) + T jk A k (ê i ⋅ )
h i ∂x i h i ∂x i h i ∂x i M.29
∂ (Tik ) ∂ Ak ∂ ê j
∇ ⋅ T ⋅ A = Ak + Tik + T jk A k (ê i ⋅ )
h i ∂x i h i ∂x i h i ∂x i M.30
∂ (Tik ) ∂ Ak
∇ ⋅ T ⋅ A = Ak + Tik
∂x i ∂x i M.31