Tensors
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
1 / 28
Vector denition
A physical quantity with a magnitude (or length) and a direction.
Bold characters
a a = ||a||e
The length
The direction unit vector
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
2 / 28
Vector denition
A physical quantity with a magnitude (or length) and a direction.
Bold characters
a a = ||a||e
The length
The direction unit vector
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
2 / 28
Vector operations: Multiplication by a scalar
b = a
b b
Sum of two vectors
ab
c=a+b
b a
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
3 / 28
Scalar product or Inner product: a b = ||a|| ||b|| cos()
b a
ab=ba a a = ||a||2 0 The vector product or cross product ab=c
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
4 / 28
The vector product is not commutative: b a = a b The triple product of three vectors, scalar
a b c = ||a|| ||b|| sin()(n c) = ||a|| ||b|| sin()||c|| cos( )
c n b a
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
5 / 28
dyad or tensor product of two vectors ab (also notes by a b) ab p = a(b p)
p
Linear operator:
ab
(b p)a
ab (p + q) = a(b p) + a(b q) Transposed dyad: (ab)T = ba = ab
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
6 / 28
A basis {g1 , g2 , g3 } and an orthonormal basis {e1 , e2 , e3 }.
g3 g2 g1
{e1 , e2 , e3 } ij is the Kronecker delta.
e3 e2 e1
ei ej = ij = 0 if i = j 1 if i = j
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
7 / 28
Components a
ai
of a vector a
3
= a1 e1 + a2 e2 + a3 e3 =
i=1
ai ei = ai ei ,
n i=1 ai bi
Einstein summation convention, ai bi = sum over the repeated index i.
a3 e3 e e1 2
a a2
a1
ai
are the components of the vector ai = a ei
(Engineering Mechanics)
(projection)
UNIVERSITY OF TWENTE.
8 / 28
matrix notation
a (with tilde) ~ a1 a = a2 ~ a3
e1 e = e2 ~ e3
Thus we may write: a= or, equivalently: a= a1 a2 = eT a ~ ~ a3 e1 e2 = aT e ~ ~ e3
a1 a2 a3
e1 e2 e3
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
9 / 28
The relations can now be written as:
ai = a ei a = a~ e ~
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
10 / 28
A second order tensor is a linear operator mapping a vector onto another vector: Ap=q linear map: A (m + n) = A m + A n Special tensors: Op = 0 1p = p The unit tensor can be written as: 1 = e1 e1 + e2 e2 + e3 e3 = ij ei ej Note: in some chapters I instead of 1
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
11 / 28
Tensor components A = A11 e1 e1 + A12 e1 e2 + A13 e1 e3 + A21 e2 e1 + A22 e2 e2 + A23 e2 e3 + A31 e3 e1 + A32 e3 e2 + A33 e3 e3 with summation convention: A = Aij ei ej The components are found by: Aij = ei A ej
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
12 / 28
In addition the following operations apply: (ab) (cd) = a(b c)d = ab cd and (ab) : (cd) = (a c)(b d) = a cb d The trace of a dyadic product is the scalar product, denoted as tr(ab) = a b
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
13 / 28
scalar product or double dot product A : B = tr(A BT ) Some properties are: A : 1 = tr(A) A : B = AT : BT
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
14 / 28
Operations: multiply by a scalar, summed with another tensor, or multiplied by another tensor: (A) v = A (v), (A + B) v = A v + B v, (A B) v = A (B v) Note: AB=BA
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
15 / 28
We can store the components Aij in a matrix A (Abar ), also A= A A11 A12 A13 = A21 A22 A23 A31 A32 A33 (e1 A e1 ) (e1 A e2 ) (e1 A e3 ) = (e2 A e1 ) (e2 A e2 ) (e2 A e3 ) (e3 A e1 ) (e3 A e2 ) (e3 A e3 ) e1 = e2 A e1 e2 e3 e3
=~ e A~ eT
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
16 / 28
Using the matrix A we can write: A= e1 e2 e3 =~ eT A~ e
A11 A12 A13 e1 A21 A22 A23 e2 A31 A32 A33 e3
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
17 / 28
Invariants rst invariant or trace I1 (A) = A11 + A22 + A33 = tr(A)
note: I1 (1) = 3 second invariant:
2 2 I2 (A) = 1 2 {tr (A) tr(A )}
third invariant or determinant of A, : I3 (A) = det(A)
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
18 / 28
Eigenvalues and eigenvectors A n = n det(A 1) = 0 characteristic equation: 3 I1 (A)2 + I2 (A) I3 (A) = 0 with n=0 det( A 1) = 0
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
19 / 28
Symmetric tensor:
AT = A , A=A1
Aij = Aji
= A (n1 n1 + n2 n2 + n3 n3 ) = 1 n1 n1 + 2 n2 n2 + 3 n3 n3 The matrix A of A with respect to this basis is a diagonal matrix: 1 0 0 A = 0 2 0 0 0 3
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
20 / 28
Inverse tensor
A1 A = 1
A1 only exists if A is regular. The inverse of the product of two tensors : (A B)1 = B1 A1 Deviatoric part Adev = A 1 3 tr(A)1
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
21 / 28
Orthogonal tensor (A a) (A b) = a b special case b = a (A a) (A a) = a a inverse is identical to transpose: AT = A1 Aa = a a, b
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
22 / 28
Third order tensor, superscript 3
3
B = Bijk ei ej ek
transfers a vector to a second order tensor: A = 3B a or a second order tensor to a vector: c = 3B : A
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
23 / 28
Fourth order tensor, superscript 4 operations: A = 4D : C
3 4
B = 4D a
E = 4D C
etc. fourth order identity tensor
4
I = Iijkl ei ej ek el
where Iijkl = ik jl The product of 4 I and 1 satises the condition
4
I 1 = 4I
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
24 / 28
Spatial derivatives, da = dx1
gradient operator
a a a + dx2 + dx3 x1 x2 x3 a a a + (dx e2 ) + (dx e3 ) = (dx e1 ) x1 x2 x3 a a a = dx e1 + e2 + e3 x1 x2 x3 = dx (a)
The symbol is called the gradient operator: = e1 + e2 + e3 x1 x2 x3
=~ eT ~
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
25 / 28
The gradient of a vector function a is obtained similarly: da = dx1 a a a + dx2 + dx3 x1 x2 x3 a a a + (dx e2 ) + (dx e3 ) = (dx e1 ) x1 x2 x3 a a a + e2 + e3 = dx e1 x1 x2 x3 = dx (a)
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
26 / 28
an arrow can be put above the gradient operator in the direction of the target vector: a = a a = (a)T For the special case a = x we nd: x = 1
(Engineering Mechanics)
UNIVERSITY OF TWENTE.
27 / 28
Derivatives with respect to second order tensor ( ) ( ) = ei ej A Aij Derivative of a scalar (f ) f = ei ej A Aij ap (a) = ep ei ej A Aij Bpq (B) = ep eq ei ej A Aij Derivative of a product of second order tensors (A B) T A B = (4 I B ) : +A C C C
(Engineering Mechanics) UNIVERSITY OF TWENTE. 28 / 28
Derivative of a vector
Derivative of a second order tensor
