100% found this document useful (1 vote)
168 views

Tensor

1. The document contains 27 questions about vectors, tensors, and their operations including: computing norms, angles, areas, cross products, projections, tensor-vector operations, invariants, eigenvalues, directional derivatives, and more. Many questions involve computing quantities in different bases. 2. Questions involve computing scalar values, vectors, and tensors using index and matrix notation for various operations including: dot products, contractions, tensor products, and more. Basis transformations are used to express quantities in different coordinate systems. 3. Properties of orthogonal, symmetric, and skew-symmetric tensors are explored, along with invariants of tensors and directional derivatives of scalar and tensor quantities. Rodriguez' formula for rotations and the spectral decomposition

Uploaded by

Sebastiao Silva
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
168 views

Tensor

1. The document contains 27 questions about vectors, tensors, and their operations including: computing norms, angles, areas, cross products, projections, tensor-vector operations, invariants, eigenvalues, directional derivatives, and more. Many questions involve computing quantities in different bases. 2. Questions involve computing scalar values, vectors, and tensors using index and matrix notation for various operations including: dot products, contractions, tensor products, and more. Basis transformations are used to express quantities in different coordinate systems. 3. Properties of orthogonal, symmetric, and skew-symmetric tensors are explored, along with invariants of tensors and directional derivatives of scalar and tensor quantities. Rodriguez' formula for rotations and the spectral decomposition

Uploaded by

Sebastiao Silva
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Questions on Vectors and Tensors

1. Basic Operations Consider two vectors a ∼ (1, 4, 6) and b ∼ (2, 0, 4), where the
components have been expressed in a given orthonormal basis. Compute

1. kak.
2. The angle between a and b.
3. The area of the parallelogram bounded by a and b.
4. b × a.

2. Projection Denote the basis in Problem 1 as {ei }3i=1 . What is the component of a in
the direction z = e1 + e2 .

3. Basic Operations As we have seen we can express vectors in terms of their components
with respect to a particular basis. Thus we can write

v = vi ei .

When the particular basis is understood (ie. everyone reading the equation will assume
the same basis) one can safely express the components of the vector in a “column vector”;
ie. as  
v1
vi →  v2  .
v3
This convention also applies to second order tensors. In this case we have for a basis
e = {e1 , e2 , e3 }
T = Tij ei ⊗ ej .
If the basis is understood we can express the components in a 3×3 matrix. The following
ordering is conventional:  
T11 T12 T13
Tij →  T21 T22 T23 
T31 T32 T33

For the problems assume an orthonormal basis e = {e1 , e2 , e3 } and

a = 2e1 + 5e2 − 7e3 ,


b = 0e1 − 8e2 + 1e3 ,
 
4
ci →  5 ,
7
and  
1 8 2
Tij →  8 3 2  .
2 2 3

Find expressions for the following. If the operation is a scalar just report the number.
If the operation results in a vector, give it in dyadic form (component values times basis
vectors, summed over all needed basis vectors) as well as in column vector form assuming
the e basis. If the operation results in a tensor, give it in dyadic form (component values
times e1 ⊗ e1 etc. summed over all needed ei ⊗ ej ) as well as in matrix form assuming
the e basis.
(a) (a · b)c, (b) b3 a · c, (c) (a · b)a ⊗ b, (d) δii , (e) T3j δ3j , (f) Tij δij , (g) Tij Tij , (h) T c,
(i) IT , (j) IIT , (k) IIIT , (l) e1 ⊗ e2 , (m) e3 ⊗ e2 .

4. Tensor Operations Let a = 1e1 + 4e2 + 6e3 , b = 7e1 + 2e2 + 3e3 , c = 1e2 + 2e3 , and
T = 12a ⊗ b + 7a ⊗ c. Determine the following

1. T · T
2. T e1
3. T T

5. Tensor-Vector Operations [This is a two dimensional problem.] Let the components


of a 2nd order tensor A be A11 = 1, A12 = A21 = 2, A22 = 3. Let the components of a
vector v be v1 = 1, v2 = 2.

1. Compute the components of w = A · v


(a) using the relation wi = Aij vj ,
(b) using matrix vector multiplication rules.
2. Compute the scalar s = v · A · v ≡ A(v, v)
(a) using the relation s = Aij vi vj ,
(b) using matrix vector multiplication.

6. Orthogonality of Skew and Symmetric Tensors If A is a symmetric tensor and B


is skew-symmetric tensor, show tr[AT B] = Aij Bij = 0.

7. Symmetric Contraction Consider a tensor A ∈ S3 (i.e. symmetric) and an arbitrary


tensor B. Prove that A : B = A : sym[B].

8. Vector Component Extraction Given a = 3e1 + 4e2 + 5e3 find aA in the E A basis
as defined in the following figure. Do this by using our component extraction rules.

Page 2
e3, E3

E2

e2

e1
π/6 rad
E1

9. Tensor Component Extraction Consider a 2nd order tensor T whose components in


the ei basis are given by  
1 0 2
T → 0 3 5 
2 4 0
Using the component extraction rule for tensors, find T12 and T33 in the E A basis.
Assume ei and E A are as defined in Problem 8. [Do not try and assemble the rotation
tensor. It is a waste of time for this problem.]

10. Basis conversion for tensor components Consider the tensor T = Tij êi ⊗ êj , where
 
1 2
T ∼
3 4 {ê }
i

Consider a second basis {Ê A } whose relation to the basis {êi } is given by

1 3
Ê 1 = ê1 + ê2
2√ 2
3 1
Ê 2 = − ê1 + ê2 .
2 2
Find the value of T11 in the {Ê A } basis.

11. Identity Representation Consider an arbitrary vector Starting from a = a show


that the identity tensor can be expressed as 1 = ei ⊗ ei where {ei }3i=1 represents any
orthonormal basis. Not that the identity tensor is defined to be the tensor 1 such that
a = 1a for all vectors a.

12. Cross Product Let v = 1e1 + 2e2 , w = 2e3 + 2e1 + 1e2 and z = v × w. What is z3 ?
Do this problem using indicial notation methods. Do NOT use the classical technique
with the determinant.

Page 3
13. Permutation Symbol Show that eijk eijk = 6.

14. Orthogonal Tensor In two dimensions, any orthogonal tensor can be expressed as

R = cos(θ)e1 ⊗ e1 + sin(θ)e1 ⊗ e2
− sin(θ)e2 ⊗ e1 + cos(θ)e2 ⊗ e2 .

1. Show that RT R = 1; ie. prove that the inverse of an orthogonal tensor is its
transpose.
2. Show that kvk = kRvk for all v; ie. an orthogonal tensor does not change the
length of a vector.

15. Rodriguez Formula Let Λ ∈ SO(3) (i.e. a rotation tensor).


(a)Argue that Λ admits the following representation:

Λ = e3 ⊗ e3 + sin(θ)(e2 ⊗ e1 − e1 ⊗ e2 ) + cos(θ)(e1 ⊗ e1 + e2 ⊗ e2 ) (∗)

where {e1 , e2 , e3 } form an orthonormal basis. [Hint: Every rotation has a vector e3 such
that Λe3 = e3 .]
(b) Using the representation in part (a) show that one also has the following represen-
tation for Λ:
sin(θ) 1 sin2 (θ/2) 2
Λ=I+ Θ+ Θ (#)
θ 2 (θ/2)2
where Θ is the skew-symmetric tensor with axial vector Θ̌ = θe3 . [Remark: Formula
(#) is often referred to as Rodriguez’ formula. This representation plays a crucial role
in non-linear computational mechanics.]

16. Indices How many indices will a rank 3 tensor have?

17. Tensor Classification Classify Aij Bj Ci Dlmn Xmn Yl as the components of a scalar, vector,
or higher order tensor.

18. Tensor Classification Classify Ai bikl Cmnpp dln as the components of a scalar, vector, or
higher order tensor.

19. 3rd Order Tensor Let ϕ be a third order tensor that is skew-symmetric; i.e.

ϕ(u, v, w) = −ϕ(v, u, w) = −ϕ(u, w, v) = −ϕ(w, v, u)

for all vectors u, v, w. Let S be a second order tensor; show that

ϕ(Se1 , e2 , e3 ) + ϕ(e1 , Se2 , e3 ) + ϕ(e1 , e2 , Se3 ) = tr[S]ϕ(e1 , e2 , e3 )

[Hint: First show ϕ(Se1 , e2 , e3 ) = S11 ϕ(e1 , e2 , e3 ).]

Page 4
20. Components of the isotropic elasticity tensor Consider the 4th order tensor C
with components λδij δkl + µ(δik δjl + δil δjk ), where λ and µ are given constants.

1. Determine the value of C1122 = C(ê1 , ê1 , ê2 , ê2 ).


2. Determine the value of C1112 = C(ê1 , ê1 , ê1 , ê2 ).

21. Basic Field Operations Let v(x) and w(x) be two vector fields and S(x) be a tensor
field. Convert the following expressions to indicial or symbolic notation as appropriate.

1. div[S] · w
2. v ⊗ (Sw)
3. vi wj wk,k
4. wj Sij ul,i

22. Eigenvalues Consider the scalar x · Sx for symmetric S. Show that the necessary
equations that this quantity be a maximum over all unit vectors x is given by the
eigenvalue equation Sx = sx. To do this extremize x · Sx with respect to x subject to
the constraint kxk = 1 using a Lagrange multiplier method.

23. Invariants Consider the tensor S = 2(e2 ⊗ e1 + e1 ⊗ e2 ). Compute the principal


invariants of S.

24. Spectral Form Consider the tensor S = 1(e1 ⊗ e1 + e2 ⊗ e2 ) + 2(e2 ⊗ e1 + e1 ⊗ e2 ).


Express S in spectral form.

25. Directional Derivative Given a stress field σ(x) on a domain Ω ⊂ R3 , explain in words
d 3
the meaning of Dv σ(x) = dα α=0 σ(x + αv), where v ∈ R is given. In particular, your
answer should allow one to understand the meaning of the components of Dv σ(x).

26. Directional Derivative Compute DG(A) where G(A) = tr(A)A.

27. Directional Derivative of 2nd Invariant The second invariant of a rank 2 tensor A
is given by I2 (A) = 21 [Aii Ajj − Akl Alk ]. Determine the directional derivative of I2 (A)
in the direction H. Infer from this result the expression for ∂A∂ij I2 (A) by applying the
relation
∂I2 (A)
Hij = DI2 (A)[H]
∂Aij

28. Directional Derivative Consider f (x) = x1 x2 x3 + x1 x2

1. Compute the directional derivative of f in the direction ê1 + ê2 + ê3 at the point
(1, 2, 3). [Warning! Normalize!]
2. Compute the gradient of f – i.e. ∇f .

29. Directional derivative Consider f (x) = x1 x2 x3

Page 5
1. Compute the directional derivative of f in the direction of ê1 + ê2 at the point
(2, 2, 2). [Warning! Normalize!].
2. Compute the gradient of f .

30. Gradient of a vector field Consider a vector field v = vi êi , where v1 = x21 , v2 = x2 ,
and v3 = x3 x1 . Compute the components of ∇v.

31. Directional Derivative Consider g(x) = 4(x1 )2 ê1 √ + x1 x2 x3 ê2 + x3 ê3 and compute
Dg[ĥ] at the point (0, 0, 0) where ĥ = (e1 + e2 + e3 )/ 3

32. Directional Derivative The definition of the derivative of a scalar valued function of
a second order tensor is the operator ∂∂f
A such that

d ∂f
f (A + αH) = Hij .
dα α=0
∂Aij

Use this definition to show that


∂ 1 tr[σ ]
1. 3∂ σ = 31 1
∂(ε ε )
2. ∂εijklij = 2εkl

33. Gradient Consider f : S3 → R of the form f (σ) = ksk − (R + k tr[σ]), where s is the
deviatoric part of σ and R and k are constants. Find ∂∂fσ .

34. Gradient Consider f : R3 → R where f (x) = xk xi Aki + eijk xj ck di . A, c, and d are


given constant tensors and vectors.

1. Compute the components of the gradient of f (·); i.e., find ∂f /∂xi .


2. Express ∂f /∂x in symbolic form.

35. Directional Derivative Determine the rate of change of the following tensor valued
function:
1
T (x) = [5x · x + x · Ax]x ⊗ x
2

at the point x = ê1 + ê3 in the direction ĥ = (ê1 + ê2 + ê3 )/ 3, where A = ê1 ⊗ ê1 +
ê2 ⊗ ê3 + ê3 ⊗ ê2 .

36. Gradient Consider f : E3 → R where f (x) = x21 + 2x1 x22 + 3x22 x3 . Find grad(f ).

37. Gradient Consider f (x) = x21 + x1 x2 x3 . Compute ∇f

38. Directional Derivative Consider f (x) = 4x1 + x1 x2 x3 + cos(x2 ).

1. Compute the directional derivative of f in the direction e1 + e2 at the point (1,1,1).


2. Compute the gradient of f, ∇f .

Page 6
39. Directional Derivative Consider f (x) = 4(x1 )2 + x1 x2 x3 .

1. Compute the directional derivative of f in the direction e1 + e2 + e3 at the point


(0,1,1). [Warning! Normalize!]
2. Compute the gradient of f, ∇f .

40. Divergence theorem Use the Divergence Theorem to show that


Z
1
x · n̂ dA = vol[R] ,
∂R 3

where R is a simply connected region of space with boundary ∂R and outward unit
normal n̂.

41. Divergence Theorem Use the Divergence Theorem to show that


Z Z
5xi xj dV = xi xj xk n̂k dA .
R ∂R

42. Divergence Theorem Consider a ball, B, of radius 4 centered at the origin. Using the
divergence theorem, compute Z
x · n dA ,
∂B

where x is (as usual) the position vector and n is the unit outward normal to the surface
∂B.

43. Divergence theorem Compute the value of the integral


Z
n̂ dA ,
∂R

where R is a simply connected region of space with boundary ∂R and outward unit
normal n̂.

44. Divergence Theorem Consider the unit ball B = {x | kxk < 1} with boundary
∂B = S 2 = {x | kxk = 1}. Using the divergence theorem compute
Z  
x
div dV .
B kxk

Page 7

You might also like