Mathematical Preliminaries
• Vectors
– Notation
– Algebra
• 2nd order Tensors
– Definition
– Notation
– Algebra
– Higher order tensors
• Vector and tensor analysis
1
Stress Tensor:
e.g. consider forces acting on a
very small 3D solid volume element.
𝑥𝑥2 𝑥𝑥2 𝑥𝑥2
𝒆𝒆2
𝑥𝑥1 𝑥𝑥1 𝑥𝑥1
𝒆𝒆1 𝒆𝒆3
𝑥𝑥3 𝑥𝑥3 𝑥𝑥3
𝒆𝒆2 𝒇𝒇 𝒇𝒇′ 𝒆𝒆2 𝒆𝒆2
𝒆𝒆1 𝒇𝒇′′ 𝒆𝒆1
𝒆𝒆1
𝒆𝒆3 𝒆𝒆3 𝒆𝒆3
2
2nd order Tensors:
I need to specify: 0th order tensor 1st order tensor 2nd order tensor
• Magnitude (scalar) (vector)
• Direction of force
• Orientation of surface
3
Dimension:
2D representation:
𝐴𝐴22
𝐴𝐴12
𝐴𝐴11 𝐴𝐴12
𝑨𝑨 =
𝐴𝐴21 𝐴𝐴21 𝐴𝐴22
𝐴𝐴11
𝒆𝒆2
𝒆𝒆1
4
2nd order Tensors: e.g. stress
𝐴𝐴11 𝒆𝒆𝟏𝟏 ⊗ 𝐞𝐞𝟏𝟏 + 𝐴𝐴12 𝒆𝒆𝟏𝟏 ⊗ 𝐞𝐞𝟐𝟐 + 𝐴𝐴13 𝒆𝒆𝟏𝟏 ⊗ 𝐞𝐞𝟑𝟑 +
1. Direct notation:
𝑨𝑨
𝐴𝐴21 𝒆𝒆𝟐𝟐 ⊗ 𝐞𝐞𝟏𝟏 + 𝐴𝐴22 𝒆𝒆𝟐𝟐 ⊗ 𝐞𝐞𝟐𝟐 + 𝐴𝐴23 𝒆𝒆𝟐𝟐 ⊗ 𝐞𝐞𝟑𝟑 +
2. Index notation
𝐴𝐴𝑖𝑖𝑖𝑖 𝒆𝒆𝒊𝒊 ⊗ 𝒆𝒆𝒋𝒋
surface
orientation
direction
magnitude 𝐴𝐴31 𝒆𝒆𝟑𝟑 ⊗ 𝐞𝐞𝟏𝟏 + 𝐴𝐴32 𝒆𝒆𝟑𝟑 ⊗ 𝐞𝐞𝟐𝟐 + 𝐴𝐴33 𝒆𝒆𝟑𝟑 ⊗ 𝐞𝐞𝟑𝟑
3. Matrix notation
𝐴𝐴11 𝐴𝐴12 𝐴𝐴13
𝐴𝐴21 𝐴𝐴22 𝐴𝐴23
𝐴𝐴31 𝐴𝐴32 𝐴𝐴33
5
2nd order Tensors (linear mapping):
𝑨𝑨𝒃𝒃 = 𝒄𝒄
𝑨𝑨 𝛼𝛼𝒖𝒖 + 𝒗𝒗 = 𝛼𝛼𝑨𝑨𝑨𝑨 + 𝑨𝑨𝑨𝑨 = 𝐜𝐜 linearity
𝑨𝑨 + 𝑩𝑩 𝒂𝒂 = 𝑨𝑨𝑨𝑨 + 𝑩𝑩𝑩𝑩 distributivity
𝑨𝑨𝑨𝑨 𝒂𝒂 = 𝐀𝐀(𝐁𝐁𝒂𝒂) Associativity (part 1)
𝛼𝛼𝑨𝑨 𝒂𝒂 = 𝛼𝛼(𝑨𝑨𝑨𝑨) Associativity (part 2)
𝑰𝑰𝑰𝑰 = 𝒂𝒂 Identity tensor
𝟎𝟎𝟎𝟎 = 𝟎𝟎 Zero tensor 6
nth order Tensor:
𝑨𝑨𝒃𝒃 = 𝒄𝒄
𝑛𝑛𝑡𝑡𝑡 order tensor: is a linear operator that maps a 𝑚𝑚th order tensor,
where 𝑚𝑚 ≤ 𝑛𝑛 to a 𝑛𝑛 − 𝑚𝑚 th order tensor.
𝒖𝒖. 𝒗𝒗 = 𝛼𝛼 1st order – 1st order 0th order
𝑨𝑨𝑨𝑨 = 𝒄𝒄 2nd order – 1st order 1st order
𝑩𝑩𝑩𝑩 = 𝑪𝑪 3rd order – 1st order 2nd order
ℒ: 𝜺𝜺 = 𝝈𝝈 4th order – 2nd order 2nd order
7
Example:
E.g. Consider a function that relates the pre-impact velocity of a perfectly elastic ball
to its post-impact velocity that is bouncing on a frictionless rigid surface.
𝒗𝒗0
𝒗𝒗f
𝒆𝒆2
𝒆𝒆3 𝒆𝒆1
𝑦𝑦1 𝑀𝑀11 𝑀𝑀12 𝑀𝑀13 𝑥𝑥1
𝑦𝑦2 = 𝑀𝑀21 𝑀𝑀22 𝑀𝑀23 𝑥𝑥2
𝑦𝑦3 𝑀𝑀31 𝑀𝑀32 𝑀𝑀33 𝑥𝑥3
𝑣𝑣1f 1 0 0 1𝑣𝑣 0
𝑣𝑣2f = 0 −1 0 𝑣𝑣20
0 0 1
𝑣𝑣 f
3 𝑣𝑣 0 3 8
Tensor (dyadic) product:
Second order tensor: can be expressed with index notation
𝑨𝑨 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝒆𝒆𝑖𝑖 ⊗ 𝐞𝐞j
Dyad: the result of the tensor product between two vectors
𝒖𝒖 ⊗ 𝒗𝒗
Linear mapping:
(𝒖𝒖 ⊗ 𝒗𝒗)𝐚𝐚 = 𝐮𝐮(𝐯𝐯. 𝐚𝐚)
Dyadic: linear combinations of dyads
𝛼𝛼(𝒖𝒖 ⊗ 𝒗𝒗) + β(𝒙𝒙 ⊗ 𝐲𝐲) + 𝛾𝛾(𝐰𝐰 ⊗ 𝐳𝐳)+….
9
Index notation:
𝑨𝑨 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝒆𝒆𝑖𝑖 ⊗ 𝒆𝒆j 𝒖𝒖 = 𝑢𝑢𝑖𝑖 𝒆𝒆i
𝐴𝐴𝑖𝑖𝑖𝑖 = 𝒆𝒆𝑖𝑖 . 𝑨𝑨𝒆𝒆𝑗𝑗 𝑢𝑢𝑖𝑖 = 𝒖𝒖. 𝒆𝒆i
10
Vector Array
𝑢𝑢1
𝒖𝒖 = 𝑢𝑢𝑖𝑖 𝒆𝒆𝒊𝒊 𝑢𝑢2
𝑢𝑢3
Components +basis Components
2nd order tensor Matrix
𝐴𝐴11 𝐴𝐴12 𝐴𝐴13
𝑨𝑨 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝒆𝒆𝒊𝒊 ⊗ 𝒆𝒆𝒋𝒋 𝐴𝐴21 𝐴𝐴22 𝐴𝐴23
𝐴𝐴31 𝐴𝐴32 𝐴𝐴33
Components +basis Components 11
Index notation rules (review):
• Any index that is not repeated is a free index and implies 𝑛𝑛
different expressions (𝑛𝑛 being the dimensions of the
space).
• In an equation, free indices on both sides should match.
• Any index that is repeated is a dummy index and implies a
summation over this index from 1..𝑛𝑛 (𝑛𝑛 being the
dimension of the space).
• No index can be repeated more than twice (in a single
term).
A free index in an equation can be changed to another
symbol granted free indices on both sides of the equation
are changed simultaneously.
Dummy indices can be represented with any letter of
choice.
12
Example:
1 0 2 1
Given 𝑺𝑺 = 0 1 2 and 𝒂𝒂 = 2
3 0 3 3
a) 𝑆𝑆𝑖𝑖𝑖𝑖
b) 𝑆𝑆𝑖𝑖𝑖𝑖 𝑆𝑆𝑖𝑖𝑖𝑖
c) 𝑆𝑆𝑗𝑗𝑗𝑗 𝑆𝑆𝑗𝑗𝑗𝑗
d) 𝑆𝑆𝑗𝑗𝑗𝑗 𝑆𝑆𝑘𝑘𝑘𝑘
e) 𝑆𝑆𝑚𝑚𝑚𝑚 𝑎𝑎𝑚𝑚 𝑎𝑎𝑛𝑛
f) 𝑆𝑆𝑛𝑛𝑛𝑛 𝑎𝑎𝑚𝑚 𝑎𝑎𝑛𝑛
13
Example:
1 0 2 1
Given 𝑺𝑺 = 0 1 2 and 𝒂𝒂 = 2
3 0 3 3
a) 𝑆𝑆𝑖𝑖𝑖𝑖 𝑆𝑆𝑖𝑖𝑖𝑖 = 5
b) 𝑆𝑆𝑖𝑖𝑖𝑖 𝑆𝑆𝑖𝑖𝑖𝑖 𝑆𝑆𝑖𝑖𝑖𝑖 𝑆𝑆𝑖𝑖𝑖𝑖 = 28
c) 𝑆𝑆𝑗𝑗𝑗𝑗 𝑆𝑆𝑗𝑗𝑗𝑗 𝑆𝑆𝑗𝑗𝑗𝑗 𝑆𝑆𝑗𝑗𝑗𝑗 = 28
d) 𝑆𝑆𝑗𝑗𝑗𝑗 𝑆𝑆𝑘𝑘𝑘𝑘 𝑆𝑆𝑗𝑗𝑗𝑗 𝑆𝑆𝑘𝑘𝑘𝑘 = 23
e) 𝑆𝑆𝑚𝑚𝑚𝑚 𝑎𝑎𝑚𝑚 𝑎𝑎𝑛𝑛 𝑆𝑆𝑚𝑚𝑚𝑚 𝑎𝑎𝑚𝑚 𝑎𝑎𝑛𝑛 = 59
f) 𝑆𝑆𝑛𝑛𝑛𝑛 𝑎𝑎𝑚𝑚 𝑎𝑎𝑛𝑛 𝑆𝑆𝑛𝑛𝑛𝑛 𝑎𝑎𝑚𝑚 𝑎𝑎𝑛𝑛 = 59
14
Tensor algebra:
1. Linear mapping:
𝑨𝑨𝑨𝑨 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝑢𝑢𝑗𝑗 𝒆𝒆𝒊𝒊
2. Inner/dot/scalar product:
𝑨𝑨𝑨𝑨 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝒆𝒆𝑖𝑖 ⊗ 𝒆𝒆𝑗𝑗 𝐵𝐵𝑘𝑘𝑘𝑘 𝒆𝒆𝑘𝑘 ⊗ 𝒆𝒆𝑙𝑙 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝐵𝐵𝑘𝑘𝑘𝑘 𝒆𝒆𝑖𝑖 ⊗ 𝒆𝒆𝑗𝑗
3. Transpose:
𝒙𝒙. 𝑨𝑨𝑨𝑨 = 𝒚𝒚. 𝑨𝑨𝑻𝑻 𝒙𝒙
𝐴𝐴𝑇𝑇𝑖𝑖𝑖𝑖 = 𝐴𝐴𝑗𝑗𝑗𝑗
4. Identity tensor:
𝑰𝑰𝑰𝑰 = 𝒙𝒙
𝐼𝐼𝑖𝑖𝑖𝑖 𝑥𝑥𝑗𝑗 = 𝑥𝑥𝑖𝑖 , so, 𝐼𝐼𝑖𝑖𝑖𝑖 = 𝛿𝛿𝑖𝑖𝑖𝑖 15
Tensor algebra:
5. Contraction:
𝑨𝑨: 𝑩𝑩 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝒆𝒆𝑖𝑖 ⊗ 𝒆𝒆𝑗𝑗 : 𝐵𝐵𝑘𝑘𝑘𝑘 𝒆𝒆𝑘𝑘 ⊗ 𝒆𝒆𝑙𝑙 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝐵𝐵𝑖𝑖𝑖𝑖
𝑨𝑨 ⋅⋅ 𝑩𝑩 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝒆𝒆𝑖𝑖 ⊗ 𝒆𝒆𝑗𝑗 ⋅⋅ 𝐵𝐵𝑘𝑘𝑘𝑘 𝒆𝒆𝑘𝑘 ⊗ 𝒆𝒆𝑙𝑙 = 𝐴𝐴𝑖𝑖𝑖𝑖 𝐵𝐵𝑗𝑗𝑗𝑗
𝑨𝑨 = 𝑨𝑨: 𝑨𝑨
6. Trace:
tr(𝑨𝑨) = 𝑰𝑰: 𝑨𝑨 = 𝐴𝐴𝑖𝑖𝑖𝑖
7. Determinant: a scalar that denotes the volume scaling factor
𝑨𝑨𝑨𝑨. 𝐀𝐀𝐀𝐀 × 𝑨𝑨𝑨𝑨
det 𝐀𝐀 = = 𝜖𝜖𝑖𝑖𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖1 𝐴𝐴𝑗𝑗2 𝐴𝐴𝑘𝑘3
𝒖𝒖. (𝒗𝒗 × 𝒘𝒘)
𝐴𝐴11 𝐴𝐴12 𝐴𝐴13 3
det(𝑨𝑨) = 𝑨𝑨 = det 𝐴𝐴21 𝐴𝐴22 𝐴𝐴23 = � 𝐴𝐴𝑖𝑖𝑖𝑖 −1 𝑖𝑖+𝑗𝑗 𝑀𝑀𝑖𝑖𝑖𝑖
𝐴𝐴31 𝐴𝐴32 𝐴𝐴33 𝑗𝑗=1 16
Tensor algebra:
8. Inverse:
𝑨𝑨−𝟏𝟏 𝑨𝑨 = 𝑨𝑨𝑨𝑨−𝟏𝟏 = 𝑰𝑰
1
det 𝑨𝑨 =
det(𝑨𝑨−𝟏𝟏 )
𝑨𝑨−𝟏𝟏 exist if and only if det 𝑨𝑨 ≠ 0
17
Tensor algebra:
9. Orthogonal tensor:
𝒖𝒖. 𝒗𝒗 = 𝑸𝑸𝑸𝑸. 𝑸𝑸𝑸𝑸 𝐐𝐐T = 𝐐𝐐−1 det 𝐐𝐐 = 1 Proper (rotation)
𝑸𝑸
𝒖𝒖 rotation
𝑸𝑸𝑸𝑸
𝜃𝜃
𝜃𝜃
𝒗𝒗 𝑸𝑸𝑸𝑸
18
Tensor algebra:
9. Orthogonal tensor:
Improper
𝒖𝒖. 𝒗𝒗 = 𝑸𝑸𝑸𝑸. 𝑸𝑸𝑸𝑸 𝐐𝐐T = 𝐐𝐐−1 det 𝐐𝐐 = −1 (reflection)
𝑸𝑸
𝒖𝒖 reflection
𝑸𝑸𝑸𝑸
𝜃𝜃 𝜃𝜃
𝑸𝑸𝑸𝑸
𝒗𝒗
19
1. Write 𝐴𝐴11 𝐵𝐵11 + 𝐴𝐴12 12
Homework
𝐵𝐵 + 𝐴𝐴 𝐵𝐵 in summation in index notation.
13 13
2. Write the following matrix multiplication equation in index notation
𝐴𝐴11 𝐴𝐴12 𝐴𝐴13 𝐵𝐵11 𝐵𝐵12 𝐵𝐵13 𝐶𝐶11 𝐶𝐶12 𝐶𝐶13
𝐴𝐴21 𝐴𝐴22 𝐴𝐴23 𝐵𝐵21 𝐵𝐵22 𝐵𝐵23 = 𝐶𝐶21 𝐶𝐶22 𝐶𝐶23
𝐴𝐴31 𝐴𝐴32 𝐴𝐴33 𝐵𝐵31 𝐵𝐵32 𝐵𝐵33 𝐶𝐶31 𝐶𝐶32 𝐶𝐶33
3. Simplify 𝛿𝛿𝑖𝑖𝑖𝑖 𝐴𝐴𝑖𝑖𝑖𝑖 .
4. Show that 𝑨𝑨𝑨𝑨 𝑇𝑇 = 𝑩𝑩𝑇𝑇 𝑨𝑨𝑇𝑇 , for second order tensors 𝑨𝑨 and 𝑩𝑩.
5. Show that tr 𝑨𝑨𝑨𝑨 = 𝑨𝑨: 𝑩𝑩T , for second order tensors 𝑨𝑨 and 𝑩𝑩.
6. Show that 𝑨𝑨𝑨𝑨 −1 = 𝑩𝑩−1 𝑨𝑨−1
20
