Solutions Manual To Introduction To Continuum Mechanics: Second Edition

Solutions Manual to
INTRODUCTION TO
CONTINUUM MECHANICS
Second Edition
NOTE
This Manual is the proprietary property of J. N. Reddy and protected by copy-
right and other state and federal laws. This Manual is being provided only
to authorized professors and instructors for use in preparing for the
classes using the affiliated textbook. No other use or distribution of
this Manual is permitted. This Manual may not be sold and may not
be distributed to or used by any student or other third party. No
part of this Manual may be reproduced, displayed or distributed in
any form or by any means, electronic or otherwise, without the prior
written permission of J. N. Reddy.
-1
0 SOLUTIONS MANUAL
PREFACE
I know that most men, including those at ease with problems of the greatest complexity, can
seldom accept even the simplest and most obvious truth if it be such as would oblige them to
admit the falsity of conclusions which they have delighted in explaining to colleagues, which
they have proudly taught to others, and which they have woven, thread by thread, into the
fabric of their lives. Tolstoy
If a man is in too big a hurry to give up an error he is liable to give up some truth with it.
Wilbur Wright
This solution manual is prepared to aid the instructor in discussing the solutions to as-
signed problems in Chapters 1 through 9 from the book, Introduction to Continuum Mechanics
, 2nd edition, by J. N. Reddy.
The instructor should make an effort to review the problems before assigning them. This
allows the instructor to make comments and suggestions on the approach to be taken and
nature of the answers expected. The instructor may wish to generate additional problems
from those given in this book, especially when taught time and again from the same book.
The author appreciates receiving comments on the book and a list of errors found in the
book and this solutions manual.
J. N. Reddy
College Station
1
Chapter 1: INTRODUCTION
1.1 Newton’s second law can be expressed as

F = ma (1)
where F is the net force acting on the body, m mass of the body, and a the acceleration
of the body in the direction of the net force. Use Eq. (1) to determine the governing
equation of a free-falling body. Consider only the forces due to gravity and the air
resistance, which is assumed to be proportional to the square of the velocity of the
falling body.
Solution: From the free-body-diagram shown in Fig.P1.1 it follows that
dv
m = Fg − Fd , Fg = mg, Fd = cv 2
dt
where v is the downward velocity (m/s) of the body, Fg is the downward force (N or
kg m/s2 ) due to gravity, Fd is the upward drag force, m is the mass (kg) of the body,
g the acceleration (m/s2 ) due to gravity, and c is the proportionality constant (drag
coefficient, kg/s). The equation of motion is
FigP1-1 dv
+ αv 2 = g, α=
c
dt m
Fd cv2
v
Fg mg
Fig. P1.1
1.2 Consider steady-state heat transfer through a cylindrical bar of nonuniform cross sec-
tion. The bar is subject to a known temperature T0 (◦ C) at the left end and exposed,
both on the surface and at the right end, to a medium (such as cooling fluid or air)
at temperature T∞ . Assume that temperature is uniform at any section of the bar,
T = T (x), and neglect thermal expansion of the bar (that is, assume rigid). Use the
principle of conservation of energy (which requires that the rate of change (increase)
of internal energy is equal to the sum of heat gained by conduction, convection, and
internal heat generation) to a typical element of the bar (see Fig. P1.2) to derive the
governing equations of the problem.
Solution: If q denotes the heat flux (heat flow per unit area, W/m2 ), then [Aq]x is
the net heat flow into the volume element at x, [Aq]x+∆x is the net heat flow out
of the volume element at x + ∆x. If h denotes the film conductance [W/(m2 ·◦ C)],
βP ∆x(T∞ − T ) is the heat flow through the surface of the rod into the body, where T∞
is the temperature of the surrounding medium and P is the perimeter (m). Suppose
that there is a heat source within the rod generating energy at a rate of g (W/m3 ).
Then the energy balance gives
[Aq]x − [Aq]x+∆x + βP ∆x(T∞ − T ) + gA ∆x = 0 (1)
or, dividing throughout by ∆x,
[Aq]x+∆x − [Aq]x
− + βP (T∞ − T ) + Ag = 0
∆x
2 SOLUTIONS MANUAL
FigP1-2
and taking the limit ∆x → 0, we obtain
d
− (Aq) + βP (T∞ − T ) + Ag = 0 (2)
dx
Convection from lateral

g(x), internal heat generation
surface
Maintained at
temperature, T0
Exposed to ambient
temperature, T∞
L
Δx
g(x)
heat flow in, heat flow out,
(Aq)x (Aq)x+Δx
Δx
Fig. P1.2
1.3 The Euler–Bernoulli hypothesis concerning the kinematics of bending deformation of

a beam assumes that straight lines perpendicular to the beam axis before deformation
remain (1) straight, (2) perpendicular to the tangent line to the beam axis, and (3)
inextensible during deformation. These assumptions lead to the following displacement
field:
dv
u1 (x, y) = −y , u2 = v(x), u3 = 0, (1)
dx
where (u1 , u2 , u3 ) are the displacements of a point (x, y, z) along the x, y, and z coor-
dinates, respectively, and v is the vertical displacement of the beam at point (x, 0, 0).
Suppose that the beam is subjected to a distributed transverse load q(x). Determine
the governing equation by summing the forces and moments on an element of the beam
(see Fig. P1.3). Note that the sign conventions for the moment and shear force are
based on the definitions
Z Z
V = σxy dA, M = y σxx dA,
A A
and may not agree with the sign conventions used in some mechanics of materials books.
Solution: Summation of the forces in the transverse direction on the element of the
beam gives
(V + ∆V ) − V + q(x)∆x = 0.
Dividing throughout with ∆x and taking the limit ∆x → 0 gives
dV
+ q = 0. (2)
dx
Taking the moment of forces about the right end of the element, we obtain
X
Mz = 0 : −V ∆x − M + (M + ∆M ) + q∆x · α∆x = 0,
where α is a number 0 ≤ α ≤ 1. Again, dividing throughout with ∆x and taking the

limit ∆x → 0 gives
dM
− V = 0. (3)
dx
Figure P1.3
CHAPTER 1: INTRODUCTION 3
y, v q(x)
y
• x z •
L Beam
cross section
y q(x) q(x)
σ xy + dσ xy
σ xx σ xx + dσ xx
x •
+ •
+
M M + dM
V V + dV σ xy
dx dx
M= ò y⋅s xx dA, V = òs xy dA
A A
Fig. P1.3
Note that V and M denote the shear force and bending moment on the entire cross
section, and they have the meaning
Z
dM
M (x) = σxx y dA, V (x) = .
A dx
Here A denotes the area of cross section. The stress resultants (V, M ) can be related to
the deflection v. Using the linear elastic constitutive relation for an isotropic material
d2 v

σxx = Eεxx = E −y 2 .
dx
Substituting into the definition of M , we obtain
d2 v d2 v
Z Z
M (x) = σxx y dA = E −y 2 y dA = −EI 2 , (4)
A A dx dx
where I is the moment of inertia about the axis of bending (z−axis). Then
d2 v

d
V =− EI 2 . (5)
dx dx
Equations (2)-(5) can be combined to obtain the following fourth-order equation for v:
d2 d2 v

EI = q(x). (6)
dx2 dx2
1.4 A cylindrical storage tank of diameter D contains a liquid column height h(x, t). Liquid
is supplied to the tank at a rate of qi (m3 /day) and drained at a rate of q0 (m3 /day).
Assume that the fluid is incompressible (that is, constant mass density ρ) and use the
principle of conservation of mass to obtain a differential equation governing h(x, t).
Solution: The conservation of mass requires
time rate of change in mass = mass inflow - mass outflow.

4 SOLUTIONS MANUAL
The above statement for the problem at hand becomes

d d(Ah)
(ρAh) = ρqi − ρq0 or = qi − q0 ,
dt dt
where A is the area of cross section of the tank (A = πD2 /4) and ρ is the mass density
of the liquid.
1.5 (Surface tension). Forces develop at the interface between two immiscible liquids,
causing the interface to behave as if it were a membrane stretched over the fluid mass.
Molecules in the interior of the fluid mass are surrounded by molecules that are at-
tracted to each other, whereas molecules along the surface (that is, inside the imaginary
membrane) are subjected a net force toward the interior. This force imbalance creates a
tensile force in the membrane and is called surface tension (measured per unit length).
Let the difference between the pressure inside the drop and the external pressure be
p and the surface tension ts . Determine the relation between p and ts for a spherical
drop of radius R.
Solution: Consider the free-body-diagram of of a half drop of liquid, as shown in Fig.
P1.5. The force due to p is p(πR2 ), whereas the force in the surface is ts (2πR). The
force balance requires
2ts
p(πR2 ) = ts (2πR) ⇒ p= .
R
FigP1-5
ts
R
p
Fig. P1.5
5
Chapter 2: VECTORS AND TENSORS
FigP2-1
2.1 Find the equation of a line (or a set of lines) passing through the terminal point of a
vector A and in the direction of vector B.
r = A + β eˆ B = A + α B
B
• Line parallel
to vector B
B •
C = β eˆ B
r
•A
O A
Fig. P2.1
Solution: Let C be a vector along the line passing through the terminal point of vector
A and parallel to vector B. Let r be the position vector to an arbitrary point on the
line parallel to vector B and passing through the terminal point of vector A. Then the
desired equation of the line is (see Fig. P2.1)
B
r = A + C = A + β êB , êB = ,
|B|
where β is a real number.
FigP2-2
2.2 Obtain the equation of a plane perpendicular to a vector A and passing through the
terminal point of vector B, without using any coordinate system.
B P
 Q

A
B r
A
Fig. P2.2
Solution: Let O be the origin and B the terminal point of vector B. Draw a directed
line segment from O to Q, such that OQ is parallel to vector A and point Q is in the
plane, as shown in Fig. P2.2. Then OQ is equal to αA, where α is a scalar. Let P be
an arbitrary point on the line BQ. If the position vector of the point P is r, then the
vector connecting points B and P is
BP = r − B.
Because BP is perpendicular to OQ = αA, we must have
BP · OQ = 0 ⇒ (r − B) · A = 0,
6 SOLUTIONS MANUAL
which is the required equation of the plane.
2.3 Find the equation of a plane connecting the terminal points of vectors A, B, and C.
Assume that all three vectors are referred to a common origin.
Solution: Let r denote the position vector. The vectors connecting the terminal points
of vectors A, B, C, and r should be in the plane. Thus, for example, the scalar triple
product of the vectors B − A, C − A, and r − A should be zero in order that they are
co-planar, as shown in Fig. P2.3:
(C − A) × (B − A) · (r − A) = 0 [or eijk (Ci − Ai )(Bj − Aj )(xk − Ak ) = 0]
For example, if A = ê1 , B = ê2 , and C = ê3 , then the equation of the plane is
−x − y − z + 1 = 0 or x + y + z = 1.
FigP2-3
B A
z   rA
y A
B 
C A
r 
C
Fig. P2.3
2.4 Let A and B denote two points in space, and let these points be represented by two
vectors A and B with a common origin O, as shown in Fig. P2.4. Show that the
straight line through points A and B can be represented by the vector equation
(r − A) × (B − A) = 0.
FigP2-5
Solution: Here we use the fact that when two vectors are parallel their vector product
is zero. Because the vectors r − A and B − A are parallel, their vector product should
be zero, giving the required result.
A
 B
 P

A B
r
O
Fig. P2.4
2.5 Prove with the help of vectors that the diagonals of a parallelogram bisect each other.
Solution: Consider the parallelogram formed by points O, A, C, and B, as shown in

Fig. P2.5. Let us denote the line segment connecting O to A as vector A, O to B as
vector B, O to C as vector C, and B to A as vector D. Suppose that vectors C and
D intersect and cross at distances α C and β D. Then we have the following relations
among the four vectors:
A = α C + (1 − β)D; A = β D + (1 − α)C (1)
CHAPTER 2: VECTORS AND TENSORS 7
FigP2-6 B = α C − β D; B = (1 − α)C − (1 − β)D (2)
From each pair of equations, we obtain the same result, namely, α = β = 0.5, implying
that vectors C and D bisect each other.
(1   ) C
A   D  (1   ) C C
D
BA  D B (1   ) D
OC  C 
C A
B
A   C  (1   ) D
O
Fig. P2.5
2.6 Show that the position vector r that divides a line PQ in the ratio k : l is given by
l k
r= A+ B,
k+l k+l
where A and B are the vectors that designate points P and Q, respectively.
Solution: Let R denote the point on line PQ where the position vector divides it in the
ratio k : l, as shown in Fig. P2.6, and let ê denote the unit vector along the line. Then
we have the relations
FigP2-7 A + k ê = r, B = r + l ê.
To eliminate ê from the equations, we multiply the first one with l and the second one
with k and add the result to obtain
A   D  (1lA+
) Ckl ê + k B = (l + k)r + lk ê or r = l
A+
k
B,
k+l k+l
which is the desired result.
k
OP  A  l
P 
ê
OQ  B R ê 
A r Q
PR  k eˆ
B
RQ  l eˆ
O
Fig. P2.6
8 SOLUTIONS MANUAL
2.7 Represent a tetrahedron by the three non-coplanar vectors A, B, and C, as shown in

Fig. P2.7. Show that the vectorial sum of the areas of the tetrahedron sides is zero.
FigP2-8
Solution: Recall the fact that the cross product A × B vectorially represents the area of
the parallelogram formed by the two vectors, which is half the area of the face formed
by vectors A and B of the tetrahedron. Thus, we can write the vector sum of the areas
of the four faces (with normal coming out of each face) as
A × B + B × C + C × A + (C − A) × (B − A) ,
which is zero because terms cancel out.
B C
B BA
C A
C
ê
A
Fig. P2.7
2.8 Deduce that the vector equation for a sphere with its center located at point A and
with a radius R is given by
(r − A) · (r − A) = R2 ,
where A is the vector connecting the origin to point A and r is the position vector.
Solution: The required result follows from the fact that r − A is the radius vector,
whose magnitude is R, as shown in Fig. P2.8.
FigP2-9
Center of the sphere
r A
A Arbitrary point on the
 
surface of the sphere
r
A
Fig. P2.8
2.9 Verify that the following identity holds (without using index notation):
(A · B)2 + (A × B) · (A × B) = |A|2 |B|2 ,
where A and B are arbitrary vectors. Hint: Use Eqs. (2.2.21) and (2.2.25).
Solution: Let C = A × B, and consider the expression
(A × B) · (A × B) = C · (A × B) = A · (B × C), (1)
where we used the identity in Eq. (2.2.21). Then using Eq. (2.2.25), we can write
B × C = B × (A × B) = (B · B)A − (B · A)B. (2)

Hence, we have
A · (B × C) = A · [(B · B)A − (B · A)B] = |A|2 |B|2 − (A · B)2 ,
from which the desired identity follows.
2.10 If A, B, and C are noncoplanar vectors (that is, A, B, and C are linearly independent),
determine if the following set of vectors is linearly independent:
r1 = A − 3B + 2C, r2 = 2A − 5B + 3C, r3 = A − 5B + 4C.
Solution: Write the linear relation among the three vectors,
α r1 + β r2 + γ r3 = 0,
which gives
(α + 2β + γ)A − (3α + 5β + 5γ)B + (2α + 3β + 4γ)C = 0.
Because A, B, and C are linearly independent and yet the linear relation must hold
implies that the coefficients of the three vectors should be identically zero, giving the
following three relations among the real numbers α, β, and γ:
α + 2β + γ = 0, 3α + 5β + 5γ, 2α + 3β + 4γ = 0,
whose solution is
α = −5γ, β = 2γ,
and γ is arbitrary. Hence, the set is linearly dependent. In fact, we can write −5r1 +
2r2 + r3 = 0.
2.11 Determine whether the following set of vectors is linearly independent:
A = 2ê1 − ê2 + ê3 , B = −ê2 − ê3 , C = −ê1 + ê2 .
Here êi are orthonormal unit base vectors in <3 .
Solution: Set the linear combination of the vectors to zero
αA + βB + γC = 0,
which gives (if a vector is zero then all its components are zero)
2α − γ = 0, −α − β + γ = 0, α − β = 0,
whose solution is γ = 2α = 2β. Therefore, the linear relation is not trivial. The vectors
are linearly dependent. In fact, we can write vector C as
1
C = − (A + B) .
2
Note that the vectors
A = 2ê1 − ê2 + ê3 , B̂ = ê2 − ê3 , C = −ê1 + ê2
are linearly independent.
2.12 Let the vectors (î, ĵ, k̂) constitute an orthonormal basis. In terms of this basis, define
a cogredient basis by
e1 = −î − ĵ , e2 = î + 2 ĵ − 2 k̂ , e3 = 2 î + ĵ + k̂.
Determine
10 SOLUTIONS MANUAL
(a) the dual or reciprocal (contragredient) basis (e1 , e2 , e3 ) in terms of the orthonormal
basis (î, ĵ, k̂) ,
(b) the magnitudes (or norms) |e1 |, |e2 |, |e3 |, |e1 |, |e2 |, and |e3 |, and
(c) the cogredient components A1 , A2 , and A3 of a vector A if its contragredient components
are given by A1 = 1, A2 = 2, A3 = 3.
Solution: (a) Note that
e1 × e2 = 2 î − 2 ĵ − k̂, e2 × e3 = 4 î − 5 ĵ − 3k̂, e3 × e1 = î − ĵ − k̂, [e1 e2 e3 ] = 1.
Then
e1 = 4 î − 5 ĵ − 3k̂, e2 = î − ĵ − k̂, e3 = 2 î − 2 ĵ − k̂.
(b) The magnitudes of the base vectors are

√ √ √ √
|e1 | = 2 , |e2 | = 3 , |e3 | = 6, |e1 | = 5 2 , |e2 | = 3 , |e3 | = 3.
(c) We have A = A1 e1 + A2 e2 + A3 e3 = e1 + 2e2 + 3e3 . We obtain
A1 = (e1 + 2e2 + 3e3 ) · e1 = 2 − 2 × 3 − 3 × 3 = −13

A2 = (e1 + 2e2 + 3e3 ) · e2 = −3 + 2 × 9 + 3 × 2 = 21
A3 = (e1 + 2e2 + 3e3 ) · e3 = −3 + 2 × 2 + 3 × 6 = 19.
2.13 Using the Gram–Schmidt orthonormalization process, construct the orthonormal sets
associated with the following sets of vectors:
(a) e1 = î1 + î3 , e2 = î1 + 2î2 + 2î3 , e3 = 2î1 − î2 + î3 .
(b) e1 = 2î1 + î2 , e2 = î1 − 2î2 + î3 , e3 = −2î1 + î2 + î3 .
where (î1 , î2 , î3 ) is a an orthonormal Cartesian basis.
Solution: (a) Set

e1 1
ê1 = = √ î1 + î3
|e1 | 2
Then ê2 is constructed as follows:
3 1
e02 = e2 − (ê1 · e2 )ê1 = î1 + 2î2 + 2î3 −
î1 + î3 = −î1 + 4î2 + î3 .
2 2
0
e 1
ê2 = 02 = √ −î1 + 4î2 + î3 .
|e2 | 3 2
Similarly,
e03 = e3 − (ê1 · e3 )ê1 − (ê2 · e3 )ê2
3 5
= 2î1 − î2 + î3 − î1 + î3 + −î1 + 4î2 + î3
2 18
1
= 2î1 + î2 − 2î3 .
9
and
e03 1
ê3 = 0
= 2î1 + î2 − 2î3 .
|e3 | 3
(b) Set
e1 1
ê1 = = √ 2î1 + î2
|e1 | 5
Then ê2 is constructed as follows:
1
e02 = e2 − (ê1 · e2 )ê1 = î1 − 2î2 + î3 − 0 ⇒ ê2 = √ î1 − 2î2 + î3 .
6
Finally,
e03 = e3 − (ê1 · e3 )ê1 − (ê2 · e3 )ê2
3 1
= −2î1 + î2 + î3 + 2î1 + î2 + î1 − 2î2 + î3
5 2
3
= −î1 + 2î2 + 5î3 ,
10
and
e03 1
ê3 = 0
= √ −î1 + 2î2 + 5î3 .
|e3 | 30
2.14 Prove the following vector identity using index notation:
A × (B × C) = (A · C)B − (A · B)C.
Solution: We begin with the left side of the equality and arrive at the right side:
A × (B × C) = Ai êi × (ê` ejk` Bj Ck ) = Ai Bj Ck ejk` epi` êp

= (δjp δki − δji δkp )Ai Bj Ck êp = Ai Bj Ci êj − Ai Bi Ck êk
= (A · C)B − (A · B)C.
2.15 Prove the following vector identity using index notation:
(A · B)2 + (A × B) · (A × B) = |A|2 |B|2 .
Solution: We begin with
(A × B) · (A × B) = (Ai Bj eijk êk ) · (Am Bn emnp êp )

= Ai Bj Am Bn eijk emnk = Ai Bj Am Bn (δim δjn − δin δjm )
= Ai Ai Bj Bj − Ai Bi Aj Bj = (A · A)(B · B) − (A · B)(A · B),
from which the desired identity follows.
2.16 Use index notation and the e-δ identity to rewrite the vector expression as a sum (or
difference) of two vector expressions:
(∇ × A) × B
where A and B are vector functions.
Solution: We have
∂Aj
(∇ × A) × B = εijk êk × (Bp êp )
∂xi
∂Aj ∂Aj
= εijk εkpq Bp êq = (δip δjq − δiq δjp ) Bp êq
∂xi ∂xi
∂Aj ∂Aj
= Bi êj − Bj êi = B · ∇A − ∇A · B.
∂xi ∂xi

2.17 Simplify the vector expression ∇ · x−y ρ
, where ρ = |x − y| and y is a fixed point,
and x is the position vector of a point in a 3D space. Express the final result in terms
of ρ only.
Solution: We have

x−y ∂ n o
∇· = êj · (x − y)[(xi − yi )(xi − yi )]−1/2
ρ ∂xj
n o
= êj · ej [(xi − yi )(xi − yi )]−1/2
12 SOLUTIONS MANUAL
1
∂
+ êj · (x − y) − [(xi − yi )(xi − yi )]−3/2 [(xi − yi )(xi − yi )]
2 ∂xj

3 x−y 3 (xj − yj ) 2
= − êj · (x j − yj ) = − (x j − yj ) =
ρ ρ3 ρ ρ3 ρ
2.18 Using index notation prove the following identities among vectors A, B, C, and D:
(a) (A × B) · (B × C) × (C × A) = (A · (B × C))2 .
(b) (A × B) × (C × D) = [A · (C × D)]B − [B · (C × D]A.
Solution: (a) We have (A × B) · (B × C) × (C × A)
= (eijk Ai Bj êk ) · [(erst Br Cs êt ) × (emnp Cm An êp )]

= eijk erst emnp Ai Bj Br Cs Cm An etpq δkq
= (δit δjp − δip δjt ) erst emnp Ai Bj Br Cs Cm An
= (ersi emnj − ersj emni ) Ai Bj Br Cs Cm An
= (B × C · A) (C × A · B) − (B × C · B) (C × A · A)
= (B × C · A) (C × A · B) = (B × C · A)2
Note that B × C · A = A × B · C.
(b) We have
(A × B) × (C × D) = (eijk Ai Bj êk ) × (emnp Cm Dn êp )

= eijk ekpq emnp Ai Bj Cm Dn êq
= emnp (δip δjq − δiq δjp ) Ai Bj Cm Dn êq
= emni Ai Bj Cm Dn êj − emnj Ai Bj Cm Dn êi
= (C × D · A) B − (C × D · B) A.
2.19 Prove that

A·D A·E A·F

[ABC][DEF] = B · D B · E B · F ,
C·D C·E C·F
and from there show that
δir δis δit

eijk erst = δjr δjs δjt .
δkr δks δkt
Solution: Recall the two properties of determinants: (1) det ([S][T ]) = det [S]· det [T ]
and (2) det [S]T = det [S]. Therefore, we have
  
A1 A2 A3 D1 E1 F1

[ABC][DEF] =  B1 B2 B3   D2 E2 F2 
C1 C2 C3 D3 E3 F3

A1 D1 + A2 D2 + A3 D3 A1 E1 + A2 E2 + A3 E3 A1 F1 + A2 F2 + A3 F3

= B1 D1 + B2 D2 + B3 D3 B1 E1 + B2 E2 + B3 E3 B1 F1 + B2 F2 + B3 F3
C1 D1 + C2 D2 + C3 D3 C1 E1 + C2 E2 + C3 E3 C1 F1 + C2 F2 + C3 F3

A·D A·E A·F

= B · D B · E B · F . (1)
C·D C·E C·F
The second identity follows from Eq. (1) when we take A = D = ê1 , B = E = ê2 , and
C = F = ê3 .
2.20 Establish the following identities :

δi1 δi2 δi3 δip δiq δir

(a) eijk = δj1 δj2 δj3 . (b) eijk epqr = δjp δjq δjr .
δk1 δk2 δk3 δkp δkq δkr
(c) eijk eijk = 6. (d) eijk emnk = δim δjn − δin δjm .
Solution:
(a) Let êi = δip êp , êj = δjq êq , and êk = δkr êr . Then the determinant form of the
triple scalar product êi · êj × êk is

δi1 δi2 δi3

eijk = êi · êj × êk = δj1 δj2 δj3 . (1)
δk1 δk2 δk3
(b) This was already established in the previous problem; we show it independently of
Problem 2.19. Again, recall the two properties of determinants: (1) det ([S][T ]) = det
[S]· det [T ] and (2) det [S]T = det [S]. Therefore, we begin with Eq. (1):
  
δi1 δi2 δi3 δp1 δq1 δr1

eijk epqr =  δj1 δj2 δj3   δp2 δq2 δr2 
δk1 δk2 δk3 δp3 δq3 δr3

δi1 δp1 + δi2 δp2 + δi3 δp3 δi1 δq1 + δi2 δq2 + δi3 δq3 δi1 δr1 + δi2 δr2 + δi3 δr3

= δj1 δp1 + δj2 δp2 + δj3 δp3 δj1 δq1 + δj2 δq2 + δj3 δq3 δj1 δr1 + δj2 δr2 + δj3 δr3
δk1 δp1 + δk2 δp2 + δk3 δp3 δk1 δq1 + δk2 δq2 + δk3 δq3 δk1 δr1 + δk2 δr2 + δk3 δr3

δim δmp δim δmq δim δmr δip δiq δir

= δjm δmp δjm δmq δjm δmr = δjp δjq δjr , (2)
δkm δmp δkm δmq δkm δmr δkp δkq δkr
where we have used the identity of the form

δi1 δp1 + δi2 δp2 + δi3 δp3 = δim δmp = δip , etc.
(c) Using the e-δ identity we can write

eijk eijk = δjj δkk − δjk δjk = 9 − 3 = 6.
(d) From Part (b), we have

δim δin δik

eijk emnk = δjm δjn δjk
δkm δkn δkk
= δim (δjn δkk − δkn δjk ) − δjm (δin δkk − δkn δik ) + δkm (δin δjk − δjn δik )
= 3δim δjn − δim δjn − 3δjm δin + δjm δin + δin δjm − δjn δim
= δim δjn − δin δjm .
2.21 Consider two rectangular Cartesian coordinate systems that are translated and rotated
with respect to each other. The transformation between the two coordinate systems is
given by
x̄ = c + Lx,
where c is a constant vector and L = [ìj ] is the matrix of direction cosines
î · êj .
ìj ≡ ē
Deduce that the following orthogonality conditions hold:
L · LT = I.
14 SOLUTIONS MANUAL
That is, L is an orthogonal matrix.
Solution: We have
î = ìj êj ,
ē ˆm .
ên = `mn ē
Then
ˆj = `ji `jp êp .
êi = `ji ē
Taking dot product with êk on both sides, we obtain
δik = `ji `jp δpk = `ji `jk or δij = `ki `kj (renamed j as k and k as j) or L · LT = I.
2.22 Determine the transformation matrix relating the orthonormal basis vectors (ê1 , ê2 , ê3 )
and (ê01 , ê02 , ê03 ), when ê0i are given by
(a) ê01 is along the vector ê1 − ê2 + ê3 and ê02 is perpendicular to the plane 2x1 + 3x2 +
x3 − 5 = 0.
(b) ê01 is along the 0

√ line segment connecting point (1, −1, 3) to (2, −2, 4) and ê3 =
(−ê1 + ê2 + 2ê3 )/ 6.
Solution: (a) Let êi be the unit base vectors in the current orthogonal system, and ê0i
be the unit base vectors in the new coordinate system. The vector ê01 has the same
direction as the vector ê1 − ê2 + ê3 but its magnitude must be unity
ê1 − ê2 + ê3 1
ê01 = = √ (ê1 − ê2 + ê3 ).
|ê1 − ê2 + ê3 | 3
The vector ê02 is along the normal to the plane 2x1 + 3x2 + x3 − 5 = 0. Hence, ê02 = n̂,
the unit normal to the plane, which is given by
∇(2x1 + 3x2 + x3 − 5) 2ê1 + 3ê2 + ê3
ê02 = = p
|∇(2x1 + 3x2 + x3 − 5)| (2)2 + (3)2 + (1)2
1
= √ (2ê1 + 3ê2 + ê3 ).
14
The third basis vector in an orthonormal system is related to the other two vectors by

ê1 ê2 ê3
0 0 0
1 1 1
ê3 = ê1 × ê2 = √3 − √3 √3

√ 2 √3 √ 1
14 14 14

1 3 1 2 3 2
= ê1 − √ − √ − ê2 √ − √ + ê3 √ + √
42 42 42 42 42 42
1
= √ (−4ê1 + ê2 + 5ê3 ).
42
Thus, the two coordinate systems are related by (note the matrix of direction cosines)
 0   √1 √ −1 1
 
√
 ê1  3 3 3  ê1 
ê02 =  √214 √314 √114  ê2 .
 
 0  −4 √1
ê3 √5 ê3
 
√
42 42 42
(b) We have
vector connecting point (1, −1, 3) to point(2, −2, 4)
ê01 =
vector magnitude
(2ê1 − 2ê2 + 4ê3 ) − (ê1 − ê2 + 3ê3 ) 1
= = √ (ê1 − ê2 + ê3 )
magnitude 3
0 1
ê3 = √ (−ê1 + ê2 + 2ê3 ) (given)
6

ê1 ê2 ê3
−1 1 2
ê02 = ê03 × ê01 = √6 √6 √6

√1 √ −1 √1
3 3 3

1 2 −1 2 1 1
= ê1 √ + √ − ê2 √ − √ + ê3 √ − √
18 18 18 18 18 18
1
= √ (ê1 + ê3 ) .
2
Transformation matrix relating (ê01 , ê02 , ê03 ) to (ê1 , ê2 , ê3 ) is given by
 1 
√
3
− √13 √13
[L] =  √12 √12 0  (aij = ê0i · êj )
 
1 1 2
− √6 √6 √6
2.23 The angles between the barred and unbarred coordinate lines are given by
ê1 ê2 ê3

ˆ1
ē 60◦ 30◦ 90◦
ˆ2
ē 150◦ 60◦ 90◦
ˆ3
ē 90◦ 90◦ 0◦
Determine the direction cosines of the transformation.
Solution: Follows from the definition

 √ 
1 3
2
√
0
2
[L] =  −
 3 1 î · êj ).
0  (aij = ē

2 2
0 0 1
2.24 The angles between the barred and unbarred coordinate lines are given by
x1 x2 x3
x̄1 45◦ 90◦ 45◦
x̄2 60◦ 45◦ 120◦
x̄3 120◦ 45◦ 60◦
Determine the transformation matrix.
Solution: Follows from the definition

 1 
√ 0 √1
 12 √1 2
− 12 î · êj ).
[L] =  (ìj = ē

 2 2
− 12 √12 1
2
2.25 Write the following sets of equations in matrix form [A]{X} = {Y }:

(a) 2x1 + x2 − 2x3 = 1, (b) 2x1 + x2 − x3 = 0,
x1 − 2x2 + x3 = 5, 3x1 − x3 = 2,
3x1 + x2 − x3 = 4. x1 + x2 + x3 = 1.
16 SOLUTIONS MANUAL
Solution: The matrix representation of the linear equations is

         
2 1 −2  x1   1  2 1 −1  x1   0 
(a)  1 −2 1  x2 = 5 , (b)  3 0 −1  x2 = 2 .
3 1 −1 x3 4 1 1 1 x3 1
       
2.26 Determine the cofactors and the determinants of the coefficient matrices in Problem
2.25.
Solution: (a) First we compute the adjoints, as shown below.

−2 1
= −4, A13 = 1 −2 = 7,
1 1
A11 = = 1, A12 =
1 −1 3 −1 3 1

1 −2 2 −2 2 1
A21 = 3 1 = −1,
= 1, A 22 = = 4, A 23 =
1 −1 3 −1

1 −2
= −3, A32 = 2 −2 = 4,
2 1
A31 = A 33 = 1 −2 = −5.

−2 1 1 1
The determinant is given by (expansion by first row)

|A| = (−1)1+1 a11 A11 + (−1)1+2 a12 A12 + (−1)1+3 a13 A13
= 2 × 1 − 1 × (−4) + (−2) × 7 = −8.
Using the first column, we obtain the same result
|A| = (−1)1+1 a11 A11 + (−1)1+2 a21 A21 + (−1)1+3 a31 A31
= 2 × 1 − 1 × 1 + 3 × (−3) = −8.
(b) The adjoints are

0 −1 3 −1 3 0
A11 = = 1, A12 =
= 4, A13 = = 3,
1 1 1 1 1 1

1 −1 2 −1 2 1
A21 = = 2, A22 = 1 1 = 3,
A23 = = 1,
1 1 1 1

1 −1 2 −1 2 1
A31 = = −1, A32 =
= 1, A33 = = −3.
0 −1 3 −1 3 0
The determinant is given by (using the first row)
|A| = (−1)1+1 a11 A11 + (−1)1+2 a12 A12 + (−1)1+3 a13 A13
= 2 × 1 − 1 × 4 + (−1) × 3 = −5.
Using the second column, we obtain the same result
|A| = (−1)1+2 a12 A12 + (−1)2+2 a22 A22 + (−1)1+3 a32 A32
= −1 × 4 − 1 × 1 = −5.
2.27 Find the inverses of the coefficient matrices in Problem 2.25.
Solution: (a) The inverse is given by (Aij are the adjoints defined in the solution to
Problem 2.26)
   
A11 −A21 A31 1 −1 −3
−1 1  1
[A] = −A12 A22 −A32  = −  4 4 −4 
|A| 8
A13 −A23 A33 7 1 −5
(b) The inverse is given by
   
A11 −A21 A31 1 −2 −1
−1 1  1
[A] = −A12 A22 −A32  = −  −4 3 −1 
|A| 5
A13 −A23 A33 3 −1 −3
2.28 Determine if the following matrices are positive:

     
2 1 −2 2 1 −1 2 1 1
(a)  1 −2 1  , (b)  3 2 −1  , (c)  1 −1 2.
3 1 −1 1 1 0 3 2 1
Solution: (a) We write the quadratic form of the matrix
{X}T [A]{X} = 2x21 − 2x22 − x23 + 2 (x1 x2 + x1 x3 + x2 x3 )
which is not positive because the vector, for example, (x1 , x2 , x3 ) = (0, 1, 1) gives
{X}T [A]{X} = −1.
(b) The quadratic form
{X}T [A]{X} = 2(x1 + x2 )2 > 0
is positive, and hence [A] is positive.

(c) The quadratic form
{X}T [A]{X} = 2x21 − x22 + x23 + 2x1 x2 + 4x1 x3 + 4x2 x3
is not always positive, because (x1 , x2 , x3 ) = (−1, 1, −1) gives {X}T [A]{X} = 0, and
(x1 , x2 , x3 ) = (0, 1, −1) as well as (x1 , x2 , x3 ) = (0, −1, 1) give {X}T [A]{X} = −4.
2.29 Check to see if the following [Q] is nonsingular, and if it is, construct the positive matrix
associated with it:  
1 0 0
[Q] =  0 1 2  .
1 1 1
Solution: We have
|Q| = 1 × (1 × 1 − 1 × 2) + 1 × (0 × 2 − 1 × 0) = −1 6= 0.
Therefore, the positive matrix associated with [Q] is

    
1 0 1 1 0 0 2 1 1
T
[A] = [Q] [Q] =  0 1 1   0 1 2  =  1 2 3  .
0 2 1 1 1 1 1 3 5
2.30 Let r denote a position vector r = x = xi êi (r2 = xi xi ) and A an arbitrary constant
vector. Show that:
(a) ∇2 (rn ) = n(n + 1)rn−2 . (b) grad (r · A) = A.

(c) div (r × A) = 0. (d) curl(r × A) = −2A.
1 1
(e) div (rA) = (r · A). (f) curl (rA) = (r × A).
r r
Solution: First we establish following two identities:

∂ 1 1 1
grad(r) = êi (xj xj ) 2 = êi (xj xj ) 2 −1 2xi
∂xi 2
1 r
= êi xi (xj xj )− 2 = . (1)
r
∂ n n n
grad(rn ) = êi (xj xj ) 2 = êi (xj xj ) 2 −1 2xi
∂xi 2
n−2
= nêi xi (xj xj ) 2 = nrn−2 r. (2)
18 SOLUTIONS MANUAL
(a) We have,
∂2 ∂
∇2 (rn ) = (rn ) = nrn−2 xi

∂xi ∂xi ∂xi
n−3 ∂r xi
= n(n − 2)r xi + nrn−2 δii = n(n − 2)rn−3 xi + 3nrn−2
∂xi r
n−2 n−2
= [n(n − 2) + 3n]r = n(n + 1)r .
(b) Because A is a constant vector, we have

∂ ∂Aj
grad(r · A) = êi (xj Aj ) = êi δij Aj + xj
∂xi ∂xi
= êi (Ai + 0) = A.
(c) Carrying out the indicated operation, we obtain

∂ ∂xj ∂Ak
div(r × A) = êi · (ejk` xj Ak ê` ) = ejk` δi` Ak + xj
∂xi ∂xi ∂xi
= (0 + 0) = 0.
(d) Carrying out the indicated operation, we obtain

∂
curl(r × A) = êi × (erst xr As êt )
∂xi

∂As
= erst êi × êt δir As + xr
∂xi
= erst ejit êj (δir As + 0) = eist ejit êj As
= −2êj δsj As = −2A.
(e) Carrying out the indicated operation, we obtain

∂ ∂r xi 1
div(rA) = êi · (rAj êj ) = êi · êj Aj = Ai = r · A.
∂xi ∂xi r r
(f) Carrying out the indicated operation, we obtain

∂ √
∇ × (rA) = êi × xj xj Ak êk
∂xi
x x
i i
= êi × êk Ak + 0 = eikt êt Ak
r r
1 1
= eikt xi Ak êt = (r × A).
r r
2.31 Let A and B be vector functions of position vector x with continuous first and second
derivatives, and let F and G be scalar functions of position x with continuous first and
second derivatives. Show that:
(a) ∇ · (∇ × A) = 0.
(b) ∇ × (∇F ) = 0.
(c) ∇ · (∇F × ∇G) = 0.
(d) ∇ · (F A) = A · ∇F + F ∇ · A.
(e) ∇ × (F A) = F ∇ × A − A × ∇F .
(f) ∇(A · B) = A · ∇B + B · ∇A + A × (∇ × B) + B × (∇ × A).
(g) ∇ · (A × B) = ∇ × A · B − ∇ × B · A.
Solution:
(a) Using the index notation, we write

∂ ∂Ak
∇ · (∇ × A) = êi · ejk` ê`
∂xi ∂xj
∂ 2 Ak ∂ 2 Ak
= ejk` δi` = eijk = 0,
∂xi ∂xj ∂xi ∂xj
because of the symmetry of Ak,ij in i and j.
(b) We have
∂2F

∂ ∂F
curl(gradF ) = êi × êj = eijk êk = 0,
∂xi ∂xj ∂xi ∂xj
where the last result is arrived by virtue of the symmetry of (∂ 2 F/∂xi ∂xj ) = (∂ 2 F/∂xj ∂xi ).
∂2F ∂2G
(c) Because ∂xi ∂xj
is symmetric in i and j and ∂xi ∂xk
is symmetric in i and k, we
obtain

∂ ∂F ∂G
∇ · (∇F × ∇G) = êi · êj × êk
∂xi ∂xj ∂xk
2
∂F ∂ 2 G

∂ F ∂G
= ejk` (êi · ê` ) +
∂xi ∂xj ∂xk ∂xj ∂xi ∂xk
2 2

∂ F ∂G ∂F ∂ G
= eijk + = 0.
∂xi ∂xj ∂xk ∂xj ∂xi ∂xk
(d) Because F = F (x) and A=A(x), we have

∂ ∂Aj ∂F
∇ · (F A) = êi · (êj Aj F ) = δij F + Aj
∂xi ∂xi ∂xi
∂Ai ∂F
= F + Ai = ∇ · AF + A · ∇F.
∂xi ∂xi
(e) Because F = F (x) and A=A(x), we have

∂ ∂Aj ∂F
∇ × (F A) = êi × (êj Aj F ) = eijk êk F + Aj
∂xi ∂xi ∂xi
∂Aj ∂F
= eijk êk F − ejik êk Aj = F ∇ × A − A × ∇F.
∂xi ∂xi
(f) Because A = A(x) and B = B(x), we have

∂
∇(A · B) = êi (Aj Bj )
∂xi

∂Aj ∂Bj
= êi Bj + Aj = ∇A · B + ∇B · A. (1)
∂xi ∂xi
Next consider

∂Bk ∂Bk
A × curl B = (êi Ai ) × ejkp êp = êq eqip ejkp Ai
∂xj ∂xj
∂Bk ∂Bi ∂Bk
= êq (δqj δik − δqk δij )Ai = êj Ai − êk Ai
∂xj ∂xj ∂xi
= ∇B · A − A · ∇B. (2)
Therefore,
A × (∇ × B) + B × (∇ × A) = ∇B · A − A · ∇B + ∇A · B − B · ∇A
20 SOLUTIONS MANUAL
= ∇A · B + ∇B · A − (A · ∇B + B · ∇A) .
From Eqs. (1) and (2) the required vector identity follows.
(g) Because A = A(x) and B = B(x), we have

∂ ∂Aj ∂Bk
∇ · (A × B) = êi · (ejk` Aj Bk ê` ) = eijk Bk + Aj
∂xi ∂xi ∂xi
= curlA · B − curlB · A.
2.32 Let A and B be vector functions of position vector x with continuous first and second
derivatives, and let F and G be scalar functions of position x with continuous first and
second derivatives. Show that:
(a) ∇ × (A × B) = B · ∇A − A · ∇B + A∇ · B − B∇ · A.
(b) (∇ × A) × A = A · ∇A − ∇A · A.
(c) ∇2 (F G) = F ∇2 G + 2∇F · ∇G + G ∇2 F .
(d) ∇2 (F x) = 2∇F + x ∇2 F .
(e) A · ∇A = ∇ 21 A · A − A × ∇ × A.

(f) ∇(A · x) = A + ∇A · x.
(g) ∇2 (A · x) = 2∇ · A + x · ∇2 A.
(a) Because A = A(x) and B = B(x), we have

∂
∇ × (A × B) = êi × (Aj Bk ejkp êp )
∂xi

∂Aj ∂Bk
= ejkp eqip êq Bk + Aj
∂xi ∂xi

∂Aj ∂Bk
= (δjq δki − δji δkq )êq Bk + Aj
∂xi ∂xi

∂Aj ∂Bi ∂Ai ∂Bk
= êj Bi + Aj − êk Bk + Ai
∂xi ∂xi ∂xi ∂xi
= B · ∇A + A∇ · B − B∇ · A − A · ∇B.
(b) Because A = A(x), we have

∂Aj
(∇ × A) × A = eijk êk × (Ap êp )
∂xi
∂Aj ∂Aj
= eijk ekpq Ap êq = (δip δjq − δiq δjp ) Ap êq
∂xi ∂xi
∂Aj ∂Aj
= Ai êj − Aj êi = A · ∇A − ∇A · A.
∂xi ∂xi
(c) Because ∇2 = ∇ · ∇, we have

∂ ∂F ∂G
∇2 (F G) = êi · êj G + F êj
∂xi ∂xj ∂xj
2
∂2G

∂ F ∂F ∂G ∂F ∂G
= êi · êj G+ + êj +F
∂xi ∂xj ∂xj ∂xi ∂xi ∂xj ∂xj ∂xi
∂2F ∂F ∂G ∂2G
= êi · êj G + 2êj · êi + F êi · êj
∂xi ∂xj ∂xj ∂xi ∂xj ∂xi
∂2F ∂2G
= G + 2∇F · ∇G + F
= ∇2 F G + 2∇F · ∇G + F ∇2 G
(d) Using the Cartesian coordinate system, we can write

∂2

∂ ∂F ∂xj
∇2 (F x) = [F (xj êj )] = (xj êj ) + F êj
∂xi ∂xi ∂xi ∂xi ∂xi
2
∂ F ∂F ∂xj ∂F ∂xj
= (xj êj ) + êj + êj + 0
∂xi ∂xi ∂xi ∂xi ∂xi ∂xi
= ∇2 F x + 2∇F
(e) We begin with the right-hand side of the identity:

2
A 1 ∂ ∂As
∇ −A×∇×A= êi (Aj Aj ) − (Ai êi ) × erst êt
2 2 ∂xi ∂xr
∂Aj ∂As
= êi Aj − erst ekit êk Ai
∂xi ∂xr
∂Aj ∂As
= êi Aj − êk (δrk δsi − δri δsk ) Ai
∂xi ∂x
r
∂Aj ∂As ∂Ak
= êi Aj − êk As − Ai
∂xi ∂xk ∂xi
∂Ak ∂
= êk Ai = Ai (Ak êk ) = A · ∇A
∂xi ∂xi
(f) Because A=A(x), we have

∂ ∂Aj ∂xj
grad(A · x) = êi (Aj xj ) = êi xj + Aj
∂xi ∂xi ∂xi
= ∇A · x + A,
∂x
where Aj ∂xji = Aj δij = Ai is used in arriving at the final expression.
(g) We have
∂2
∇2 (A · x) = (Aj xj )
∂xi ∂xi

∂ ∂Aj ∂xj ∂ ∂Aj
= xj + Aj = xj + Ai
2
∂ Aj ∂Aj ∂xj ∂Ai
= xj + +
2
∂ Aj ∂Ai
= xj + 2 = x · ∇2 A + 2∇ · A.
∂xi ∂xi ∂xi
2.33 Show that

∇(Rn x) = Rn I + nRn−2 xx, R2 = x · x
using (a) index notation and (b) the spherical coordinate system. vskip2mm Solution:
(a) Using the Cartesian coordinate system, we can write

∂ ∂R ∂xj
∇(Rn x) = êi (Rn xj êj ) = êi nRn−1 xj êj + Rn êj
∂xi ∂xi ∂xi
n−2 n n−2 n
= êi nR xi x + R δij êj = nR xx + R I,
where we have used the identity ∂R/∂xi = xi /R. This is arrived at using the fact that
R2 = xj xj , and
∂R ∂ 1 ∂ 1 xi
= (xj xj )1/2 = (xk xk )−1/2 (xj xj ) = (2xi ) = .
∂xi ∂xi 2 ∂xi 2R R
22 SOLUTIONS MANUAL
(b) Using the spherical coordinate system, we can write

∂ 1 ∂ 1 ∂
∇(Rn x) = êR + êθ + êφ [Rn (RêR )]
∂R R ∂θ R sin θ ∂φ
Rn

n n ∂êR ∂êR
= (n + 1)R êR êR + R êθ + êφ
∂θ sin θ ∂φ
= [(n + 1)Rn êR êR + Rn êθ êθ + Rn êφ êφ ] = nRn−2 xx + Rn I.
2.34 Show that

∇2 (Rn x) = n(n + 3)Rn−2 x, R2 = x · x
using (a) index notation, and (b) the spherical coordinate system.
Solution: (a) Using the Cartesian coordinate system, we can write

∂ n−2
∇2 (Rn x) = ∇ · ∇(Rn x) = ∇ · nRn−2 xx + Rn I = êi · xx + Rn I

nR
∂xi

∂R ∂ ∂R
= êi · n(n − 2)Rn−3 (xx) + nRn−2 (xx) + nRn−1 I+0
∂xi ∂xi ∂xi
n−3 xi n−1 xi
h i
n−2
= êi · n(n − 2)R xx + nR (êi x + xêi ) + nR I+0
R R
xi
= n(n − 2)Rn−3 (êi · x)x + nRn−2 [(êi · êi )x + (êi · x)êi ] + nRn−2 xi (êi · I)
R
= n(n − 2)Rn−4 xi xi x + nRn−2 (3x + x) + nRn−2 x)
= [n(n − 2) + 4n + n]Rn−2 x = n(n + 3)Rn−2 x.
(b) Using the spherical coordinate system, we can write

∂ 1 ∂ 1 ∂
∇2 (Rn x) = êR · Rn I + nRn−2 xx

+ êθ + êφ
∂R R ∂θ R sin θ ∂φ
= êR · nRn−1 I + n2 Rn−1 êR êR

1 ∂êR ∂êR ∂êθ ∂êθ
+ êθ · Rn êR + êR + êθ + êθ +0
R ∂θ ∂θ ∂θ ∂θ

∂êR ∂ê R
+ nRn êR + êR
∂θ ∂θ

1 ∂êR ∂êR ∂êθ ∂êθ ∂êφ ∂êφ
+ êφ · Rn êR + êR + êθ + êθ + êφ + êφ
R sin θ ∂φ ∂φ ∂φ ∂φ ∂φ ∂φ

n ∂êR ∂êR
+ nR êR + êR
∂φ ∂φ

1
= nRn−1 êR + n2 Rn−1 êR + Rn êR − êR + nRn êR
R
1
+ [Rn (sin θêR + cos θêθ − sin θêR − cos θêθ ) + nRn sin θêR ]
R sin θ
= n(n + 3)Rn−2 x.
2.35 Show that the vector area of a closed surface is zero, that is,
I
n̂ ds = 0.
s
Solution: Use the Gradient Theorem in Eq. (2.4.45) with φ = 1 to obtain the required
result.
2.36 Show that the volume of the region Ω enclosed by a boundary surface Γ is
I I
1 1
volume = grad(r2 ) · n̂ ds = r · n̂ ds.
6 Γ 3 Γ
∂r
Solution: Note that ( ∂xi
= xi /r)
∂r
grad(r2 ) = 2rêi = 2êi xi = 2r.
∂xi
Using the Divergence Theorem in Eq. (2.4.46), we can write
I Z
r · n̂ ds = ∇ · r dx = δii V = 3V.
Γ Ω
which gives the required result.
2.37 Let φ(r) be a scalar field. Show that

Z I
∂φ
∇2 φ dx = ds.
Ω Γ ∂n
Solution: Using the Divergence Theorem in Eq. (2.4.46), we can write

Z I I
∂φ
∇ · ∇φ dx = n̂ · ∇φ dS = ds.
Ω Γ Γ ∂n
2.38 In the divergence theorem (2.4.45), set A = φ ∇ψ and A = ψ ∇φ successively and

obtain the integral forms
Z h I
i ∂ψ
(a) φ∇2 ψ + ∇φ · ∇ψ dx = φ ds,
Ω Γ ∂n
Z h I
2 2
i ∂ψ ∂φ
(b) φ∇ ψ − ψ∇ φ dx = φ −ψ ds,
Ω Γ ∂n ∂n
Z h I
i ∂ ∂φ
(c) φ∇4 ψ − ∇2 φ∇2 ψ dx = φ (∇2 ψ) − ∇2 ψ ds,
Ω Γ ∂n ∂n
where Ω denotes a (2D or 3D) region with bounding surface Γ. The first two identities
are sometimes called Green’s first and second theorems.
Solution: The integral relations are obvious. (a) The identity is obtained by substitut-
ing A = φ ∇ψ for A into Eq. (2.4.45)
Z I I
∂ψ
∇ · (φ ∇ψ) dΩ = n̂ · (φ ∇ψ)dΓ = φ dΓ.
Ω Γ Γ ∂n
The left side of the equality, as per Problem 2.17(d) with A replaced there with ∇ψ,
is equal to
Z Z Z
∇φ · ∇ψ + φ∇2 ψ dΩ.

∇ · (φ ∇ψ) dΩ = (∇φ · ∇ψ + φ∇ · ∇ψ) dΩ =
Ω Ω Ω
(b) This identity follows directly from (a) by interchanging φ and ψ and subtracting
the resulting identity from the one in (a).
(c) This identity follows from (b) by replacing ψ with ∇2 ψ.

24 SOLUTIONS MANUAL
2.39 Let V and S be smooth vector and second-order tensor fields defined in Ω and on Γ
(the closed boundary of Ω) and let n̂ be the unit outward normal to Γ. Establish the
identity I Z
V · S · n̂ ds = ∇ · (ST · V) dx.
Γ Ω
Solution: We begin with the left-hand side of the equality and use the divergence
theorem I I
V · S · n̂ ds = n̂ · ST · V ds
Γ
ZΓ
= ∇ · ST · V dx,
Ω
where we have used the identity V · S = ST · V.
The identity can also be established using the index notation,
I I Z
∂
V · S · n̂ ds = Vj Sji ni ds = (Vj Sji ) dx
Γ Γ Ω ∂xi
Z Z
∂
= (Sji Vj ) dx = ∇ · ST · V dx,
Ω ∂xi Ω
2.40 Let S be a smooth second-order tensor field defined in Ω and on Γ (the closed boundary
of Ω) and let n̂ be the unit outward normal to Γ. Use index notation to establish the
identity I Z
x × (n̂ · S) ds = [x × (∇ · S) + E : S] dx,
Γ Ω
where x is the position vector and E is the third-order permutation tensor [see Eq.
(2.5.23)].
Solution: We begin with the left-hand side of the equality and use the divergence
theorem
I I
x × (n̂ · S) ds = êl ekjl xk ni Sij ds
Γ
ZΓ Z
∂ ∂
= (êl ekjl xk Sij ) dx = êl ejkl (xj Sik ) dx
Ω ∂xi Ω ∂xi
Z
∂Sik ∂xj
= êl ejkl xj + Sik dx
Ω ∂xi ∂xi
Z
∂Sik
= êl ejkl xj + δij Sik dx
Ω ∂xi
Z
∂Sik
= êl ejkl xj + Sjk dx
Ω ∂xi
Z
= [x × (∇ · S) + E : S] dx.
Ω
2.41 Establish the following identities for a second-order tensor S:

1
(a) |S| = eijk S1i S2j S3k . (b) |S| = Sir Sjs Skt erst eijk .
6
Sim Sin Sip

(c) erst |S| = eijk Sir Sjs Skt . (d) Sjm
Sjn Sjp = eijk emnp |S|.
Skm Skn Skp
Solution: (a) Let S1 = S1i êi , S2 = S2i êi , and S3 = S3i êi . Then the determinant form
of the scalar triple product is

S11 S12 S13

S21 S22 S23 = S1 · S2 × S3 = eijk S1i S2j S3k . (1)

S31 S32 S33
(b) From Eq. (1) it follows that

Si · (Sj × Sk ) = erst Sir Sjs Skt . (2)
Multiplying both sides with eijk
eijk [Si · (Sj × Sk )] = eijk erst Sir Sjs Skt . (3)
By expanding the left side, we obtain
eijk Si · (Sj × Sk ) = e1jk S1 · (Sj × Sk ) + e2jk S2 · (Sj × Sk ) + e3jk S3 · (Sj × Sk )
= 6S1 · (S2 × S3 ) = 6|S|.
or
1 1
|S| = eijk Si · (Sj × Sk ) = eijk erst Sir Sjs Skt . (4)
6 6
(c) Multiply both sides of the Eq. (4) with erst and use the identity erst erst = 6,
erst |S| = eijk Sir Sjs Skt . (5)
(d) From Problem 2.20(a), we have

 
Si1 Si2 Si3 δm1 δn1 δp1 Sim Sin Sip

[Si · (Sj × Sk )]emnp = Sj1 Sj2 Sj3  δm2 δn2 δp2  = Sjm Sjn Sjp .
Sk1 Sk2 Sk3 δm3 δn3 δp3 Skm Skn Skp
We also have from the left side expression

[Si · (Sj × Sk )]emnp = emnp erst Sir Sjs Skt = emnp eijk |S|.
Thus the identity is proved.
2.42 Given vector A and second-order tensors S and T with the following components:
     
 2 −1 0 5 8 −1 6
{A} = −1 , [S] =  3 7 4  , [T] =  5 4 9 
4 986 −7 8 −2
 
determine
(a) tr(S). (b) S : S. (c) S : ST .
(d) A · S. (e) S · A. (f) S · T · A.
Solution: The expressions in (a)-(c) are scalars, whereas those in (d)-(f) are components
of a vector.
(a) Sii = S11 + S22 + S33 = −1 + 7 + 6 = 12.
2 2 2
(b) S : S = Sij Sji = S11 + S22 + S33 + 2 (S12 S21 + S13 S31 + S23 S32 )
2 2 2
= (−1) + (7) + (6) + 2[(0)(3) + (5)(9) + (4)(8)] = 240.
(c) S : ST = Sij Sij = S11

2 2
+ S12 2
+ S13 2
+ S21 2
+ S22 2
+ S23 2
S31 2
+ S32 2
+ S33
= (−1) + (0) + (5) + (3) + (7) + (4) + (9) + (8) + (6)2 = 281.
2 2 2 2 2 2 2 2
(d) Elements of the row matrix R ≡ A · S, where

j = 1 : R1 = Si1 Ai = S11 A1 + S21 A2 + S31 A3 = (−1)(2) + (3)(−1) + (9)(4) = 31.
j = 2 : R2 = Si2 Ai = S12 A1 + S22 A2 + S32 A3 = (0)(2) + (7)(−1) + (8)(4) = 25.
j = 3 : R3 = Si3 Ai = S13 A1 + S23 A2 + S33 A3 = (5)(2) + (4)(−1) + (6)(4) = 30.
(e) Elements of a column vector C = S · A, where

j = 1 : C1 = Si1 Aj = S11 A1 + S12 A2 + S13 A3 = (−1)(2) + (0)(−1) + (5)(4) = 18.
j = 2 : C2 = Si2 Ai = S21 A1 + S22 A2 + S23 A3 = (3)(2) + (7)(−1) + (4)(4) = 15.
j = 3 : C3 = Si3 Ai = S31 A1 + S32 A2 + S33 A3 = (9)(2) + (8)(−1) + (6)(4) = 34.
26 SOLUTIONS MANUAL
(f) The required result is a column vector, Sij Tjk Ak ≡ Bi ⇒ S · T · A = B or

        
−1 0 5 8 −1 6  2  −43 41 −16  2   −191 
{B} =  3 7 4   5 4 9  −1 =  31 57 73  −1 = 297 .
986 −7 8 −2 4 70 71 114 4 525
     
2.43 Determine the rotation transformation matrix such that the new base vector ē ˆ1 is along
ˆ2 is along the normal to the plane 2x1 + 3x2 + x3 = 5. If S is the
ê1 − ê2 + ê3 , and ē
dyadic whose components in the unbarred system are given by S11 = 1, S12 = S21 =
0, S13 = S31 = −1, S22 = 3, S23 = S32 = −2, and S33 = 0, find the components in
the barred coordinates.
Solution: See Problem 2.22(a) for the basis vectors of the barred coordinate system in
terms of the unbarred system; the matrix of direction cosines L is given there. Then
the components of the dyad in the barred coordinate system are
 
1 0 −1
[S̄] = [L][S][L]T = [L]  0 3 −2  [L]T
−1 −2 0
 1 −1 1     √1 √2 
√
3
√
3
√
3 1 0 −1 3 14
− √442
=  √214 √314 √114   0 3 −2   − √13 √314 √142 
   
−4 1 5 1 1 5
√ √ √ −1 −2 0 √ √ √
42 42 42 3 14 42
 1 −1 1   1 9

√ √ √ 0 √ − 42
√
 3 3 3 14
=  √214 √314 √114   − √53 √714 − √742 
 
−4 1 5 1 8 2
√
42
√
42
√
42
√
3
− √14 √42
 14

2 − √42 0
=  − √1442 15
− 37√ .
 
14 14 3 
37 13
0 − 14√3 14
2.44 Suppose that the new axes x̄i are obtained by rotating xi through 60◦ about the x2 -
axis. Determine the components Āi of a vector A whose components with respect to
the xi coordinates are (2, 1, 3).
Solution: The transformation matrix is given by

 √ 
1 3
0 −
 2 2
 √0 1 √0 

3
2
0 23
The components Āi are given by Āi = ìj Aj :

  1 √    √ 
0 − 23  2    1 − 323

 Ā1  2


Ā2 =  √0 1 √0  1 = 1 .
 
√

Ā3
 3 3 3  5 3
   
02 2 2

2.45 Show that the following expressions for an arbitrary tensor S are invariant: (a) Sii ,
(b) Sij Sij , and (c) Sij Sjk Ski .
Solution: We must use the tensor transformation equations to establish the results. (a)
We begin with
S̄ii = ìm ìn Smn = δmn Smn = Smm = Sii ,
where we have used the result of Problem 2.11.
(b) We have
S̄ij S̄ij = (ìm `jn Smn ) (ìp `jq Spq ) = δmp δnq Smn Spq = Smn Smn = Sij Sij .
(c) We have
S̄ij S̄jk S̄ki = (ìm `jn Smn ) (`jp `kq Spq ) (`kr ìs Srs ) = δqr δnp δms Smn Spq Srs
= Smn Snr Srm = Sij Sjk Ski .
2.46 If A and B are arbitrary vectors and S and T are arbitrary dyads, verify that:
(a) (A · S) · B = A · (S · B). (b) (S · T) · A = S · (T · A).
(c) A · (S · T) = (A · S) · T. (d) (S · A) · (T · B) = A · (ST · T) · B.
Solution: These identities are easy to prove (almost by inspection). The parentheses in
the expressions are not even required. Only one that is not obvious is one in (d). (d)
We have
(S · A) · (T · B) = (Sij êi êj · Ak êk ) · (Tpq êp êq · Br êr )
= Sij Aj Tiq Bq = A · ST · T · B.
2.47 If A is an arbitrary vector and R and S are arbitrary dyads, verify that:
(a) (I × A) · R = A × R. (b) (A × I) · R = A × R.
T T
(c) (R × A) = −A × R . (d) (R · S)T = ST · RT .
−1 −1 −1
(e) (R · S) =S ·R . (f) (R · S)−T = R−T · S−T .
Solution: (a) We have
(I × A) · R = [δij êi êj × (Ak êk )] · Rmn êm ên = Ak δij ejk` êi ê` · Rmn êm ên
= Ak δij ejk` δ`m Rmn êi ên = êj ejkm Ak Rmn ên = A × R.
(b) Following the same procedure
(A × I) · R = [(Ak êk ) × δij êi êj ] · Rmn êm ên = Ak δij eki` ê` êj · Rmn êm ên
= Ak ekm` Rmn ê` ên = A × R.
(c) We have
(R × A)T = [Rij êi êj × êk Ak ]T = Rij Ak ejk` (êi ê` )T

= Rij Ak ejk` ê` êi = −Ak ekj` φij ê` êi
= −(Ak êk ) × (Rij êj êi ) = −A × RT .
(d) We have
(R · S)T = (Rij Sjk êi êk )T

= Rij Sjk êk êi (interchanged the base vectors)
= (êk Skj )T (Rji êi )T (interchanged the subscripts)
= ST · RT .
(e) We begin with
(R · S) · (S−1 · R−1 ) = R · (S · S−1 ) · R−1 = R · R−1 = I.
Similarly,
(S−1 · R−1 ) · (R · S) = S−1 · (R−1 · R) · S = S−1 · S = I.
From the above two equalities, it is clear that S−1 · R−1 is the unique inverse of R · S,
that is,
(R · S)−1 = S−1 · R−1 .
28 SOLUTIONS MANUAL
(f) We begin with the left side

(R · S)−T = ((R · S)T )−1 = (ST · RT )−1 = (RT )−1 · (ST )−1 = R−T · S−T ,
where we made use of the results of parts (d) and (e) in arriving at the final result.
2.48 The determinant of a second-order tensor is also defined by the expression
[(S · A) × (S · B)] · (S · C)
|S| =
A×B·C
where A, B, and C are arbitrary vectors. Verify the identity in an orthonormal basis
{êi }.

[(S · A) × (S · B)] · (S · C) = Sip Ap eijk Sjq Bq Skr Cr = eijk Sip Sjq Skr Ap Bq Cr .
and the left side (after multiplying throughout by the scalar triple product A × B · C)
is given by
|S|A × B · C = |S|epqr Ap Bq Cr .
Then from the expression, we obtain
(|S|epqr − eijk Sip Sjq Skr ) Ap Bq Cr = 0.
Since, Ap Bq Cr is arbitrary, it follows that
|S|epqr − eijk Sip Sjq Skr = 0,
which, in view of Part (c) of Problem 2.41, is the required result.
2.49 For an arbitrary second-order tensor S show that ∇ · S in the cylindrical coordinate
system is given by

∂Srr 1 ∂Sθr ∂Szr 1
∇·S= + + + (Srr − Sθθ ) êr
∂r r ∂θ ∂z r

∂Srθ 1 ∂Sθθ ∂Szθ 1
+ + + + (Srθ + Sθr ) êθ
∂r r ∂θ ∂z r

∂Srz 1 ∂Sθz ∂Szz 1
+ + + + Srz êz .
∂r r ∂θ ∂z r
Solution: Using the del operator from Table 2.4.2, the divergence of S is computed as

∂ 1 ∂ ∂
êr + êθ + êz · [Srr êr êr + Srθ êr êθ + Sθr êθ êr + · · · + Szz êz êz ]
∂r r ∂θ ∂z

∂Srr ∂Srθ ∂Srz 1 ∂êr ∂êr
= êr + êθ + êz + Srr êθ · êr + Srθ êθ · êθ
∂r ∂r ∂r r ∂θ ∂θ

∂Sθθ ∂êθ ∂Sθr ∂êr ∂êr ∂Sθz
+ êθ + Sθθ + êr + Sθr + Szr êθ · êz + êz
∂θ ∂θ ∂θ ∂θ ∂θ ∂θ
∂Szr ∂Szθ ∂Szz
+ êr + êθ + êz
∂z ∂z ∂z
∂Srr ∂Srθ ∂Srz Srr Srθ 1 ∂Sθθ Sθθ
= êr + êθ + êz + êr + êθ + êθ − êr
∂r ∂r ∂r r r r ∂θ r
1 ∂Sθr Sθr Srz 1 ∂Sθz ∂Szr ∂Szθ ∂Szz
+ êr + êθ + êz + êz + êr + êθ + êz ,
r ∂θ r r r ∂θ ∂z ∂z ∂z
where the following derivatives of the base vectors are accounted for
∂êθ ∂êr
= −êr , = êθ .
∂θ ∂θ
Collecting the coefficients of êr , êθ , and êz , we obtain the required result.
2.50 For an arbitrary second-order tensor S show that ∇ × S in the cylindrical coordinate
system is given by

1 ∂Szr ∂Sθr 1 ∂Srθ ∂Szθ
∇ × S = êr êr − − Szθ + êθ êθ −
r ∂θ ∂z r ∂z ∂r

1 1 ∂Srz ∂Sθz 1 ∂Szθ ∂Sθθ 1
+ êz êz Sθz − + + êr êθ − + Szr
r r ∂θ ∂r r ∂θ ∂z r

∂Srr ∂Szr 1 ∂Szz ∂Sθz
+ êθ êr − + êr êz −
∂z ∂r r ∂θ ∂z

∂Sθr 1 ∂Srr 1 1 ∂Srz ∂Szz
+ êz êr − + Srθ + Sθr + êθ êz −
∂r r ∂θ r r ∂z ∂r

∂Sθθ 1 1 1 ∂Srθ
+ êz êθ + Sθθ − Srr − .
∂r r r r ∂θ
Solution: Using the cylindrical coordinate system, we can write ∇ × S as

∂ êθ ∂ ∂
∇ × S = êr + + êz × Srr êr êr + Sθθ êθ êθ + Szz êz êz + Srθ êr êθ
∂r r ∂θ ∂z

+ Sθr êθ êr + Srz êr êz + Szr êz êr + Sθz êθ êz + Szθ êz êθ
∂
= êr × [Srr êr êr + Sθθ êθ êθ + Szz êz êz + Srθ êr êθ + Sθr êθ êr + · · · ]
∂r
1 ∂
+ êθ × [Srr êr êr + Sθθ êθ êθ + Szz êz êz + Srθ êr êθ + Sθr êθ êr + · · · ]
r ∂θ
∂
+ êz × [Srr êr êr + Sθθ êθ êθ + Szz êz êz + Srθ êr êθ + Sθr êθ êr + · · · ]
∂z
h ∂S ∂Szz ∂Sθr ∂Szr ∂Sθz
θθ
= êr × êθ êθ + êz êz + êθ êr + êz êr + êθ êz
∂r ∂r ∂r ∂r ∂r
∂Szθ i
+ êz êθ
∂r
1 ∂Srr ∂êr ∂êθ ∂Szz
+ êθ × êr êr + Srr êr + Sθθ êθ + êz êz
r ∂θ ∂θ ∂θ ∂θ
∂Srθ ∂êθ ∂êθ ∂Srz
+ êr êθ + Srθ êr + Sθr êr + êr êz
∂θ ∂θ ∂θ ∂θ
∂Szr ∂êr ∂Szθ ∂êθ ˆ θ
∂e
+ êz êr + Szr êz + êz êθ + Sθz êz + Szθ êz
∂θ ∂θ ∂θ ∂θ ∂θ

∂Srr ∂Sθθ ∂Srθ ∂Sθr ∂Srz
+ êz × êr êr + êθ êθ + êr êθ + êθ êr + êr êz
∂z ∂z ∂z ∂z ∂z

∂Sθz
+ êθ êz
∂z

1 ∂Szr ∂Sθr 1 ∂Srθ ∂Szθ
= êr êr − − Szθ + êθ êθ −

1 1 ∂Srz ∂Sθz 1 ∂Szθ ∂Sθθ 1
+ êz êz Sθz − + + êr êθ − + Szr

∂Srr ∂Szr 1 ∂Szz ∂Sθz
+ êθ êr − + êr êz −
∂z ∂r r ∂θ ∂z

∂Sθr 1 ∂Srr 1 1 ∂Srz ∂Szz
+ êz êr − + Srθ + Sθr + êθ êz −

∂Sθθ 1 1 1 ∂Srθ
+ êz êθ + Sθθ − Srr − .
∂r r r r ∂θ
2.51 For an arbitrary second-order tensor S show that ∇ · S in the spherical coordinate
30 SOLUTIONS MANUAL
system is given by

∂SRR 1 ∂SφR 1 ∂SθR 1
∇·S= + + + [2SRR − Sφφ − Sθθ + SφR cot φ] êR
∂R R ∂φ R sin φ ∂θ R

∂SRφ 1 ∂Sφφ 1 ∂Sθφ 1
+ + + + [(Sφφ − Sθθ ) cot φ + SφR + 2SRφ ] êφ

∂SRθ 1 ∂Sφθ 1 ∂Sθθ 1
+ + + + [(Sφθ + Sθφ ) cot φ + 2SRθ + SθR ] êθ .
Solution: Using the del operator in the spherical coordinate system (see Table 2.4.2),
the divergence of the tensor S is computed as

∂ 1 ∂ 1 ∂
êR + êφ + êθ · [SRR êR êR + SRφ êR êφ + SφR êφ êR + · · · ]
∂R R ∂φ R sin φ ∂θ

∂SRR ∂SRφ ∂SRθ 1 ∂SφR
= êR + êφ + êθ + SRR êR + (SRφ + SφR )êφ + êR + SRθ êθ
∂R ∂R ∂R R ∂φ

∂Sφφ ∂Sφθ 1
+ êφ − Sφφ êR + êθ + [SRR êR + SRφ êφ + (SRθ + SθR )êθ − Sθθ êR ]
∂φ ∂φ R
cot φ
+ [SφR êR + Sφφ êφ + (Sφθ + Sθφ )êθ − Sθθ êφ ]
R
1 ∂SθR ∂Sθφ ∂Sθθ
+ êR + êφ + êθ ,
R sin φ ∂θ ∂θ ∂θ
where the following derivatives of the base vectors are used:
∂êθ ∂êR ∂êφ ∂êφ ∂êR
= − sin φ êR −cos φ êφ , = sin φ êθ , = −êR , = cos φ êθ , = êφ .
∂θ ∂θ ∂φ ∂θ ∂φ
Collecting the coefficients of êR , êφ , and êθ , we obtain the required result.
2.52 Show that ∇u in the spherical coordinate system is given by

∂uR ∂uφ ∂uθ
∇u = êR êR + êR êφ + êR êθ
∂R ∂R ∂R
1 ∂uR 1 ∂uφ 1 ∂uθ
+ − uφ êφ êR + + uR êφ êφ + êφ êθ
R ∂φ R ∂φ R ∂φ

1 ∂uR ∂uφ
+ − uθ sin φ êθ êR + − uθ cos φ êθ êφ
R sin φ ∂θ ∂θ

∂uθ
+ + uR sin φ + uφ cos φ êθ êθ .
∂θ
Solution: Using the del operator in the spherical coordinate system (see Table 2.4.2),
the gradient of the vector u is computed as

∂ 1 ∂ 1 ∂
êR + êφ + êθ (uR êR + uφ êφ + uθ êθ )

∂uR ∂uφ ∂uθ
= êR êR + êφ + êθ
∂R ∂R ∂R

1 ∂uR ∂uφ ∂uθ
+ êφ êR + uR êφ + êφ − uφ êR + êθ
R ∂φ ∂φ ∂φ
1 h ∂u
R ∂uφ ∂uθ
+ êθ êR + uR sin φ êφ + êφ + uφ cos φ êθ + êθ
R sin φ ∂θ ∂θ ∂θ
i
+ uθ (− sin φ êR − cos φ êφ ) ,

∂êθ ∂êR ∂êφ ∂êφ ∂êR
= − sin φ êR −cos φ êφ , = sin φ êθ , = −êR , = cos φ êθ , = êφ .
∂θ ∂θ ∂φ ∂θ ∂φ
Collecting the coefficients of êR êR , êR êφ , êR êθ and so on, we obtain the given expres-
sion for the gradient of a vector in the spherical coordinate system.
2.53 Prove the following identities when A and B are vectors and S, R, and T are second-
order tensors:
(a) tr(AB) = A · B. (b) tr(ST ) = tr S.

(c) tr(R · S) = R · ·S. (d) tr(RT · S) = R : S.
(e) tr(R · S) = tr(S · R). (f) tr(R · S · T) = tr(T · R · S) = tr(S · T · R).
Solution: (a) By the definition of trace of a dyad, we have tr(AB) = AB : I =

Ai Bj δij = Ai Bi = A · B.
(b) We have tr(ST ) = ST : I = Sji δij = Sii = tr S.
(c) We have tr(R · S) = (R · S) : I = Rik Skj δij = Rik Ski = R · ·S.
(d) We have tr(RT · S) = (RT · S) : I = Rki Skj δij = Rki Ski = R : S.
(e) We have tr(R · S) = (R · S) : I = Rik Skj δij = Rik Ski = Ski Rik = tr(S · R).
(f) We have tr(R·S·T) = (R·S·T) : I = Rik Skm Tmj δij = Rik Skm Tmi = Tmi Rik Skm =
tr(T · R · S). Similarly, we can show that tr(R · S · T) = tr(S · T · R).
2.54 Show that the characteristic equation for a symmetric second-order tensor Φ can be
expressed as
λ3 − I1 λ2 + I2 λ − I3 = 0,
where
1
I1 = φkk , I2 = (φii φjj − φij φji ),
2
1
I3 = (2φij φjk φki − 3φij φji φkk + φii φjj φkk ) = det (φij ).
6
are the three invariants of Φ.
Solution: From Problem 2.20(b), we have

1
det[(φij − λδij )] = eijk erst (φir − λδir )(φjs − λδjs )(φkt − λδkt )
6
1
= eijk erst φir φjs φkt
6
1
− λ[ eijk erst (φjs φkt δir + φkt φir δjs + φir φjs δkt )]
6
1
+ λ2 [ eijk erst (φkt δir δjs + φjs δir δkt + φir δjs δkt )]
6
1
− λ3 eijk erst δir δjs δkt
6
λ
= det(φij ) − [φjs φkt eijk eist + φkt φir eijk erjt + φir φjs eijk ersk ]
6
2 1 1 1 3 1
+ λ [ eijk eijt φkt + eijk eisk φjs + eijk erjk φir ] − λ eijk eijk
6 6 6 6
λ
= −λ3 + λ2 φii − (φii φkk − φjk φkj ) − det(φij ), (6)
2
where we have used the identities
eijk eijt = δjj δkt − δjt δkj = 3δkt − δkt = 2δkt
eijk eist = δjs δkt + δjt δks , eijk eijk = 6.
32 SOLUTIONS MANUAL
2.55 Find the eigenvalues and eigenvectors of the following matrices:

   √ 
4 −4 0 √2 − 3 0
(a)  −4 0 0 (b) − 3 4 0
0 0 3 0 0 4
   
1 0 0 2 −1 1
(c)  0 3 −1  (d)  −1 0 1
0 −1 3 1 1 2
Solution: (a) The characteristic polynomial is −λ3 + I1 λ2 − I2 λ + I3 = 0, which can be

expressed as
−λ3 + 7λ2 + 4λ − 48 = 0 → [−λ2 + 4λ + 16](λ − 3) = 0.
Clearly, λ1 = 3 is an eigenvalue of the matrix. The remaining two eigenvalues are

obtained from −λ2 + 4λ + 16 = 0. Thus, we have
√ √
λ1 = 3.0, λ2 = 2(1 + 5) = 6.4721, λ3 = 2(1 − 5) = −2.4721.
The eigenvector components Ai associated with λ1 = 3 are calculated from

    
4 − 3 −4 0  A1   0 
 −4 0−3 0  A2 = 0 ,
0 0 3−3 A3 0
   
which gives A1 − 4A2 = 0 and −4A1 − 3A2 = 0, or A1 = A2 = 0. Hence, A3 = 1 and

the eigenvector is Â(1) = ±(0, 0, 1).
√
The eigenvector components associated with λ2 = 2(1 + 5) are calculated from
    
4 − λ2 −4 0  A1   0 
 −4 0 − λ2 0  A2 = 0 ,
0 0 3 − λ2 A3 0
   
which gives
√
2+2 5
A1 = − A2 = −1.618A2 , A3 = 0, → Â(2) = ±(−0.8507, 0.5257, 0).
4
Similarly, we obtain
√
2+2 5
A1 = A2 = 1.618A2 , A3 = 0, → Â(3) = ±(0.5257, 0.8507, 0).
4
(b) The characteristic polynomial is −λ3 + I1 λ2 − I2 λ + I3 = 0, which can be expressed

as
−λ3 + 10λ2 − 29λ + 20 = 0 → [−λ2 + 6λ − 5](λ − 4) = 0.
The eigenvalues are (λ2 = 4 is obvious) λ1 = 1, λ2 = 4, λ3 = 5. The eigenvectors
are
Â(1) = ±(0.866, 0.5, 0); Â(2) = ±(0.0, 0.0, 1.0);
1 √
Â(3) = ± (1, − 3, 0) = ±(0.5, −0.866, 0).
2
(c) The characteristic polynomial is −λ3 + I1 λ2 − I2 λ + I3 = 0, which can be expressed

as
−λ3 + 7λ2 − 14λ + 8 = 0 → [−λ2 + 6λ − 8](λ − 1) = 0.
The eigenvalues are λ1 = 4, λ2 = 2, λ3 = 1. The eigenvectors are
1 1
Â(1) = ± √ (0, 1, −1), Â(2) = ± √ (0, 1, 1), Â(3) = ±(1, 0, 0).
2 2
(d) The characteristic polynomial is −λ3 + I1 λ2 − I2 λ + I3 = 0, which can be expressed

as
−λ3 + 4λ2 − λ − 6 = 0 → [−λ2 + 5λ − 6](λ + 1) = 0.
The eigenvalues are λ1 = 3, λ2 = 2, λ3 = −1. The eigenvectors are
1 1 1
Â(1) = ± √ (1, 0, 1), Â(2) = ± √ (−1, 1, 1), Â(3) = ± √ (−1, −2, 1).
2 3 6
Note that the eigenvectors are orthogonal to each other.
2.56 Find the eigenvalues and eigenvectors of the following matrices:
   
3 5 8 1 −1 0
(a)  5 1 0  (b)  −1 2 −1 
8 0 2 0 −1 2
   
1 2 0 3 2 0
(c)  1 −1 1  (d) 2 0 0
0 −2 1 1 0 2
Solution:
(a) The characteristic polynomial is −λ3 + I1 λ2 − I2 λ + I3 = 0, which can be expressed
as
−λ3 + 6λ2 + 78λ − 108 = 0.
The eigenvalues are λ1 = 11.8242 λ2 = 1.2848, λ3 = −7.1090. The eigenvector
associated with λi are
Â(1) = ±(0.7300, 0.3372, 0.5945); Â(2) = ±(0.4799, 0.8424, −0.5368);
Â(3) = ±(−0.6817, 0.4204, 0.5987).
(b) Form the deviatoric form of the matrix

1 − 53 −1
   2 
0 − 3 −1 0
0
[φ ] =  −1 2 − 53 −1  =  −1 13 −1 
0 −1 2 − 53 0 −1 13
and compute its invariants
" 2 2 2 #
0 1 2 1 1 14
I2 = − + + +2+2 = ,
2 3 3 3 6
" 2 #
0 2 1 1 7
I3 = − − 1 + (−1) = .
3 3 3 27
The eigenvalues are
λ1 = 3.24698, λ2 = 1.55496, λ3 = 0.19806.
The eigenvectors are
Â(1) = ±(0.328, −0.737, 0.591),
Â(2) = ±(0.591, −0.328, −0.737),
Â(3) = ±(0.737, 0.591, 0.328).
(c) The characteristic polynomial is (I1 = 1, I2 = −1, and I3 = −1) −λ3 + λ2 + λ − 1 =

0, and the eigenvalues are λ1 = −1, λ2 = 1, λ3 = 1. The eigenvectors are (not
normalized) Â(1) = (1, −1, 1); Â(2) = ±(1, 0, −1); Â(3) = ±(−1, 0, 1).
(d) The characteristic polynomial is (I1 = 5, I2 = 2, and I3 = −8) −λ3 +5λ2 −2λ−8 =
0, and the eigenvalues are λ1 = −1, λ2 = 2, λ3 = 4. The eigenvectors are (not
normalized) Â(1) = (−0.5, 1, 0.16667); Â(2) = ±(0, 0, 1); Â(3) = ±(1, 0.5, 0.5).
34 SOLUTIONS MANUAL
2.57 Find the eigenvalues and eigenvectors associated with the matrix
 
−2 2 10
[S] =  2 −11 8  .
10 8 −5
Solution: The characteristic equation is −λ3 + I1 λ2 − I2 λ + I3 = 0, with
I1 = −2 − 11 − 5 = −18, I2 = −81, I3 = 1458.
Thus, the characteristic equation is (not always possible to factor out a root)
−λ3 − 18λ2 + 81λ + 1458 = 0 or (λ − 9)(λ + 9)(λ + 18) = 0.
The roots of this cubic equation are (ordered from the largest to the smallest value)
λ1 = 9, λ2 = −9, λ3 = −18.
The eigenvector associated with λ1 = 9 is obtained from

   (1)   
−2 − 9 2 10  x1 
  0
 2 −11 − 9 8  x(1) = 0 .
 2 
10 8 −5 − 9  x(1)3
 0
(1) (1) (1) (1)

These equations give x1 = x3 and x2 = 0.5x3 . Hence
   
 1.0   2.0 
(1) 1
{X}(1) = 0.5 x3 , {X̂}(1) = 1.0 .
3
1.0 2.0
  
The eigenvector associated with λ2 = −9 is obtained from

   (2)   
−2 + 9 2 10  x1 
  0
 2 −11 + 9 8  x(2)
2 = 0 .
10 8 −5 + 9  x(2)  0
 
3
(2) (2) (2) (2)

These equations give x1 = −2x3 and x2 = 2x3 . Hence
   
 −2   −2.0 
(2) (2) (2) 1
{X} = 2 x3 , {X̂} = 2.0 .
3
1 1.0
  
The eigenvector associated with λ2 = −18 is obtained from

   (3)   
−2 + 18 2 10  x1 
  0
 2 −11 + 18 8  x(3) = 0 .
 2 
10 8 −5 + 18  x(3)
3
 0
(3) (3) (3) (3)

These equations give x1 = −0.5x3 and x2 = −x3 . Hence
   
 −0.5   −1.0 
(3) 1
{X}(3) = −1.0 x3 , {X̂}(3) = −2.0 .
3
1.0 2.0
  
2.58 If p(x) = a0 + a1 x2 + · · · + an xn , and [A] is any square matrix, we define the polynomial
in [A] by
p(A) = a0 [I] + a1 [A] + a2 [A]2 + · · · + an [A]n .
If
1 −1
[A] =
−1 1
and p(x) = 1 − 2x + x2 , compute p(A).
Solution: We have

1 0 2 −2 1 −1 1 −1
p(A) = [I] − 2[A] + [A][A] = − +
0 1 −2 2 −1 1 −1 1

1 0 2 −2 2 −2 1 0
= − + =
0 1 −2 2 −2 2 0 1
2.59 Cayley–Hamilton Theorem [see Gantmacher (1959)]. Consider a square matrix [S] of
order n. Denote by p(λ) the determinant of |[S] − λ[I]| [that is, p(λ) = p(S − λI)],
called the characteristic polynomial. Then the Cayley–Hamilton Theorem states that
p(λ) = 0. Here p(λ) is as defined in Problem 2.58. Use matrix computation to verify
the Cayley–Hamilton theorem for each of the following matrices:
 
2 −1 1
1 −1 0
(a) 1 0.
2 1
1 −2 1
Solution: (a) The characteristic polynomial is given by

1 − λ −1
p(λ) ≡ |[A] − λ[I]| =
= λ2 − 2λ + 3.
2 1−λ
Then we have

1 −1 1 −1 1 −1 10
p([A]) = −2 +3
2 1 2 1 2 1 01

−1 −2 2 −2 30 00
= − + = .
4 −1 4 2 03 00
(b) The characteristic polynomial is given by

2 − λ −1 1
p(λ) ≡ |[A]−λ[I]| = 0 1 − λ 0 = (2−λ)(1−λ)2 −(1−λ) = (1−λ)(1−3λ+λ2 ).

1 −2 1 − λ
Then we have
       
−1 1 −1 1 0 0 6 −3 3 5 −5 3
p([A]) =  0 0 0   0 1 0  −  0 3 0+0 1 0 
1 −2 0 0 0 1 3 −6 3 3 −5 2
    
−1 1 −1 0 −2 0 0 0 0
=  0 0 0   0 −1 0  =  0 0 0.
1 −2 0 0 1 0 0 0 0
2.60 Consider the matrix in Example 2.5.3:

 
2 1 0
[S] =  1 4 1.
0 1 2
Verify that the Cayley–Hamilton theorem and compute the inverse of [S].
Solution: The characteristic equation associated with the matrix [S] is
φ(λ) = λ3 − 8λ2 + 18λ − 12 = 0.
We are required to prove that
φ([S]) = [S]3 − 8[S]2 + 18[S] − 12[I] = [0].

36 SOLUTIONS MANUAL
We have
    
2 1 0 2 1 0 5 6 1
2
[S] = [S][S] =  1 4 1   1 4 1  =  6 18 6 
0 1 2 0 1 2 1 6 5
    
5 6 1 2 1 0 16 30 8
[S]3 = [S]2 [S] =  6 18 6   1 4 1  =  30 84 30 
1 6 5 0 1 2 8 30 16
Then
         
16 30 8 5 6 1 2 1 0 12 0 0 0 0 0
 30 84 30  − 8  6 18 6  + 18  1 4 1− 0 12 0  = 0 0 0,
8 30 16 1 6 5 0 1 2 0 0 12 0 0 0
which was to be verified. The inverse of [S] is given by

1
[S]−1 = 18[I] − 8[S] + [S]2
12
       
18 0 0 2 1 0 5 6 1 ! 7 −2 1
1  1  −2
= 0 18 0 − 81 4 1+6 18 6 = 4 −2  .
12 12
0 0 18 0 1 2 1 6 5 1 −2 7
Additional Problems for Chapter 2
N2.1 Find the gradient of a vector A in the (a) cylindrical and (b) spherical coordinate
systems.
Solution: (a) We begin with (see Table 2.4.2)

∂ 1 ∂ ∂
∇ = êr + êθ + êz , A = Ar êr + Aθ êθ + Az êz .
∂r r ∂θ ∂z
Then we have

∂ 1 ∂ ∂
∇u = êr + êθ + êz (Ar êr + Aθ êθ + Az êz )
∂r r ∂θ ∂z
∂Ar ∂Aθ ∂Az 1 ∂Ar
= êr êr + êr êθ + êr êz + êθ êr
∂r ∂r ∂r r ∂θ
Ar ∂êr 1 ∂Aθ Aθ ∂êθ 1 ∂Az
+ êθ + êθ êθ + êθ + êθ êz
r ∂θ r ∂θ r ∂θ r ∂θ
∂Ar ∂Aθ ∂Az
+ êz êr + êz êθ + êz êz
∂z ∂z ∂z
∂Ar ∂Aθ 1 ∂Ar
= êr êr + êr êθ + êθ êr − Aθ
∂r ∂r r ∂θ

∂Az ∂Ar 1 ∂Aθ
+ êr êz + êz êr + êθ êθ Ar +
∂r ∂z r ∂θ
1 ∂Az ∂Aθ ∂Az
+ êθ êz + êz êθ + êz êz ,
r ∂θ ∂z ∂z
∂êr ∂êθ
= êθ , = −êr .
∂θ ∂θ
(b) From Table 2.4.2, we have

∂ 1 ∂ 1 ∂
∇ = êR + êφ + êθ
and
A = AR êR + Aφ êφ + Aθ êθ .
Hence, we have

∂ 1 ∂ 1 ∂
êR + êφ + êθ (AR êR + Aφ êφ + Aθ êθ )
∂ 1 ∂
= êR (AR êR + Aφ êφ + Aθ êθ ) + êφ (AR êR + Aφ êφ + Aθ êθ )
∂R R ∂φ
1 ∂
+ êθ (AR êR + Aφ êφ + Aθ êθ )
R sin φ ∂θ

∂AR ∂Aφ ∂Aθ
= êR êR + êφ + êθ
∂R ∂R ∂R

1 ∂AR ∂Aφ ∂Aθ
+ êφ êR + AR êφ + êφ − Aφ êR + êθ
R ∂φ ∂φ ∂φ
1 h ∂AR ∂Aφ
+ êθ êR + AR sin φ êθ + êφ
∂Aθ i
+ Aφ cos φ êθ + êθ − Aθ (sin φ êR + cos φ êφ )
∂θ
∂AR ∂Aφ 1 ∂AR ∂Aθ
= êR êR + êR êφ + − Aφ êφ êR + êR êθ
∂R ∂R R ∂φ ∂R

1 ∂AR 1 ∂Aφ 1 ∂Aθ
+ − Aθ sin φ êθ êR + AR + êφ êφ + êφ êθ
R sin φ ∂θ R ∂φ R ∂φ

1 ∂Aφ 1 ∂Aθ
+ − Aθ cos φ êθ êφ + AR sin φ + Aφ cos φ + êθ êθ
R sin φ ∂θ R sin φ ∂θ
N2.2 If ê is any unit vector and A an arbitrary vector, show that

A = (A · ê)ê + ê × (A × ê).
This identity shows that a vector can resolved into a component parallel to and one
perpendicular to an arbitrary direction ê.
Solution: An arbitrary vector A can be decomposed into a vector sum of two vectors,
one along any arbitrary line (A1 ) and the other (A2 ) perpendicular to it (with mag-
nitudes |A1 | = |A| cos θ and |A2 | = |A| sin θ, respectively), as shown in the figure.
Because the unit vector along the line is ê, we have A1 = (A · ê)ê. To determine the
vector A2 perpendicular to the line along ê, consider the vector product A × ê, which
is perpendicular to the plane of both A and ê (that is, out of the page - not shown
in the figure). Then, vector ê × (A × ê) lies in the plane of vectors A and ê, and it
is perpendicular to vector A, with the magnitude |A × ê| = |A| sin θ, as shown in the
figure below. Hence, we have
ê × (A × ê)
A = |A| cos θ ê + |A| sin θ êp = (A · ê)ê + ê × (A × ê), êp = .
|A| sin θ
FigP2-4
Alternatively, we can obtain the result using Eq. (2.2.25) (which is to be established
in Problem 2.14) with A = C = ê and B = A.
Arbitrary line
A sin 
eˆ p 
A cos
A  A cos eˆ  A sin  eˆ p
ê A  ( A  eˆ ) eˆ  eˆ  ( A  eˆ )


eˆ  ( A  eˆ )
38 SOLUTIONS MANUAL
N2.3 In a rectangular Cartesian coordinate system find the length and direction cosines of
a vector A that extends from the point (1, −1, 3) to the midpoint of the line segment
from the origin to the point (6, −6, 4).
Solution: The vector is given by

1
A= (6ê1 − 6ê2 + 4ê3 ) − (ê1 − ê2 + 3ê3 ) = 2ê1 − 2ê2 − ê3 .
2
The direction cosines are (2/3, −2/3, −1/3).
N2.4 The vectors A and B are defined as follows:
A = 3î − 4k̂, B = 2î − 2ĵ + k̂,
where (î, ĵ, k̂) is an orthonormal basis.
(a) Find the orthogonal projection of A in the direction of B.

(b) Find the angle between the positive directions of the vectors.
Solution: (a) The orthogonal projection of A onto B is A · B = 6 − 4 = 2.

(b) The angle between A and B is

A·B 2
θ = cos−1 = cos−1 → θ = 82.34◦ .
|A| |B| 5×3
N2.5 Consider a particle constrained to move in a circular orbit of radius a at a constant

speed v. Determine the velocity and acceleration vectors.
Solution: Let r(t) denote the position vector at any time t. Then the velocity vector is
given by
dr v
v= = ω × r, ω = êz , r = aêr ,
dt a
where êr is the unit vector along r, êr = r/a, and êz is the unit vector perpendicular
to the plane of the orbit. The acceleration of the particle is given by (because êz does
not change its direction with time)
dv dω dr
a= = ×r+ω×
dt dt dt
v dêz
= ×r+ω×v =ω×v
a dt
v2

v
= êz × (ω × r) = [êz × (êz × êr )]
a a
v2
= [(êz · êr )êz − (êz · êz )êr ]
a
v2
= − êr .
a
Alternately, let v = vêi . Then v = êt = ω × r = v(êz × êr ) or êt = êz × êr . That is
êt is the unit vector normal to êz and êr . The vectors (êr , êt , êz ) form an orthonormal
basis. Further, a(dêr /dt) = v = vêt implies that
dêr
= ω êt = ω(êz × êr ) = ω × êr .
dt
N2.6 (Matrices of Functions) Let [A(t)] be an m × n matrix whose elements are all functions
of t. We define the derivative and integral of [A(t)] in terms of its elements:
Z b Z b
d daij (t)
[A(t)] = , [A(t)]dt = aij (t) dt .
dt dt a 0
Show that
d d d
(a) dt
([A(t)] [B(t)]) = dt
[A(t)] [B(t)] + [A(t)] dt [B(t)], where [B] is a matrix of order n × p.
(b) d
dt
(et[A] ) = [A] et[A] , where [A] is a constant matrix, and
d
Pn
(c) dt
(|A(t)|) = i=1 |Ai (t)|, where Ai (t) is obtained from [A] by differentiating the ith
row only.
Solution: (a) Let [C(t)] = [A(t)][B(t)]. Then we have

n
X
cij = aik bkj
k=1
and Xn
d dcij daik dbkj
([C(t)]) = = bkj + aik
dt dt dt dt
k=1
d d
= [A(t)] [B(t)] + [A(t)] [B(t)].
dt dt
(b) We have
d t[A] d taij
= [aij ] etaij = [A] et[A] .

(e ) = e
dt dt
(c) We have

d da1i da2j da3k
(|A(t)|) = eijk a2j a3k + a1i a3k + a1i a2j = |A1 | + |A2 | + |A3 |.
dt dt dt dt
40 SOLUTIONS MANUAL
Chapter 3: KINEMATICS OF CONTINUA
3.1 Given the motion
χ(x, t) = x = (1 + t)X1 ê1 + (1 + t)X2 ê2 + X3 ê3 , 0 ≤ t < ∞,
(a) determine the velocity and acceleration fields of the motion, and
(b) sketch deformations of the line X2 = 2X1 , for fixed X3 = 1 at t = 1, 2, and 3.
Solution: (a) The velocity is given by

dx x
v= =X= .
dt X=fixed 1+t
The acceleration is given by

dv d x 1 dx x
a= = = − = 0.
dt dt 1+t 1 + t dt (1 + t)2
Alternatively,

dv ∂v x x
a= = + v · ∇v = − +v·∇
dt ∂t (1 + t)2 1+t
x v
=− + = 0.
(1 + t)2 1+t
Also, we have
dv dX
a= = = 0.
dt dt X=fixed
(b) Deformations of the line X2 = 2X1 , for fixed X3 = 1, at t = 1, 2, and 3 are simply
stretches by the amount 1 + t along the original line.
3.2 Determine the deformation mapping that maps a unit square into the quadrilateral
shape shown
Figure P3.2 in Fig. P3.2. Assume that the mapping is a complete polynomial in X1
and X2 up to the term X1 X2 (note that the constant term is zero for this case).
x2 , X 2
4 •(3, 4)
3 (1, 3)•
Deformation mapping
x1 = a1 X1 + b1 X 2 + c1 X1 X 2
χ (X )
x2 = a2 X1 + b2 X 2 + c2 X1 X 2
2
1 • • (1, 1)
0 • • X1 • x1 , X1
2 3 4
−1 • (4, −1)
Fig. P3.2
x1 = a1 X1 + b1 X2 + c1 X1 X2 , x2 = a2 X1 + b2 X2 + c2 X1 X2 , x3 = 0
CHAPTER 3: KINEMATICS OF CONTINUA 41
and use the three points (X1 , X2 ) = (1, 0), (1,1), and (0,1) to determine the constants
a1 , a2 , b1 , b2 , c1 , and c2 as a1 = 4, a2 = −1, b1 = 1, b2 = 3, c1 = −2, and c2 = 2, so
that the mapping is
χ(x) = (4X1 + X2 − 2X1 X2 ) ê1 + (−X1 + 3X2 + 2X1 X2 ) ê2 + X3 ê3 .
3.3 Show that in the spatial description the acceleration components in the cylindrical
coordinates are
∂vr ∂vr vθ ∂vr ∂vr v2
ar = + vr + + vz − θ ,
∂t ∂r r ∂θ ∂z r
∂vθ ∂vθ vθ ∂vθ ∂vθ vr vθ
aθ = + vr + + vz + ,
∂t ∂r r ∂θ ∂z r
∂vz ∂vz vθ ∂vz ∂vz
az = + vr + + vz .
∂t ∂r r ∂θ ∂z
Solution: By definition,
Dv ∂v ∂v ∂vr ∂vθ ∂vz
a= = + v · ∇v , = êr + êθ + êz .
Dt ∂t ∂t ∂t ∂t ∂t
Using Eq. (2.5.24), we have

∂vr ∂vθ ∂vz
v · ∇v = vr êr + êθ + êz
∂r ∂r ∂r

1 ∂vr 1 ∂vθ 1 ∂vz
+ vθ êr − vθ + êθ vr + + êz
r ∂θ r ∂θ r ∂θ

∂vr ∂vθ ∂vz
+ vz êr + êθ + êz
∂z ∂z ∂z

∂vr vθ ∂vr ∂vr
= êr vr + − vθ + vz
∂r r ∂θ ∂z

∂vθ vθ ∂vθ ∂vθ
+ êθ vr + vr + + vz
∂r r ∂θ ∂z

∂vz vθ ∂vz ∂vz
+ êz vr + + vz . (1)
∂r r ∂θ ∂z
Thus, the acceleration components are equal to as given in the problem statement.
3.4 Show that in the spatial description the acceleration components in the spherical coor-
dinates are

∂vR ∂vR vφ ∂vR vθ ∂vR
aR = + vR + − vφ + − vθ sin φ ,
∂t ∂R R ∂φ R sin φ ∂θ

∂vφ ∂vφ vφ ∂vφ vθ ∂vφ
aφ = + vR + + vR + − vθ cos φ ,

∂vθ ∂vθ vφ ∂vθ vθ ∂vθ
aθ = + vR + + + vR sin φ + vφ cos φ .
Solution: We begin with the expression from Problem 2.50 (with u replaced by v)
∂vR ∂vφ ∂vθ
∇v = êR êR + êR êφ + êR êθ
∂R ∂R ∂R
1 ∂vR 1 ∂vφ 1 ∂vθ
+ − vφ êφ êR + + vR êφ êφ + êφ êθ
R ∂φ R ∂φ R ∂φ

1 ∂vR ∂vφ
+ − vθ sin φ êθ êR + − vθ cos φ êθ êφ

∂vθ
+ + vR sin φ + vφ cos φ êθ êθ
∂θ
42 SOLUTIONS MANUAL
and take the dot product from the left with v = vR êR + vφ êφ + vθ êθ and collecting the
coefficients of êR , êφ , and êθ , we obtain the required result.
3.5 The motion of a continuous medium is given by
x1 = 1 + eat X1 , x2 = 1 + e−2at X2 , x3 = X3 , 0 ≤ t < ∞,

where a is a positive constant. Determine

(a) the components of the deformation gradient F and the inverse mapping,
(b) the velocity components in the spatial description,
(c) the velocity components in the material description, and
(d) the acceleration components in the spatial description.
(e) Then verify the results of (d) by calculating first the acceleration components in the
material coordinates and then using the inverse transformation in (a) to obtain the
components in the spatial description.
Solution: (a) The matrix associated with the deformation tensor is given by
1 + eat
 
0 0
−2at
[F ] =  0 1+e 0.
0 0 1
The Jacobian of the transformation is
J = 1 + eat 1 + e−2at .

The inverse mapping is given by

−1 −1
X1 = 1 + eat x1 , X2 = 1 + e−2at x2 , X3 = x3 .
(b) The displacement components in the material and spatial descriptions are
eat
u1 = x1 − X1 = eat X1 = x1 ,
1 + eat
e−2at
u2 = x2 − X2 = e−2at X2 = x2 ,
1 + e−2at
u3 = 0.
(c) The velocity components in the material and spatial descriptions are
dx1 eat
v1 = = aeat X1 = a x1 ,
dt 1 + eat
dx2 e−2at
v2 = = −2ae−2at X2 = −2a x2 ,
dt 1 + e−2at
v3 = 0.
(d) The acceleration components in the spatial description are

∂v1 ∂v1 ∂v1 ∂v1
a1 = + v1 + v2 + v3
∂t ∂x1 ∂x2 ∂x3
2
eat eat eat eat

= a2 at
− a2
at 2
x1 + a at
x1 = a2 x1 ,
1+e (1 + e ) 1+e 1 + eat
∂v2 ∂v2 ∂v2 ∂v2
a2 = + v1 + v2 + v3
∂t ∂x1 ∂x2 ∂x3
2
e−2at e−2at e−2at e−2at

= 4a2 − 4a 2
x 1 + −2a x 2 = 4a2
x2 ,
1 + e−2at (1 + e−2at )2 1 + e−2at 1 + e−2at
a3 = 0.
(e) The acceleration components in the material description are

dv1
a1 = = a2 eat X1 ,
dt
dv2
a2 = = 4a2 e−2at X2 ,
dt
a3 = 0.
Using the inverse transformation, we can express a in the spatial description as
eat
a1 = a2 x1 ,
1 + eat
−2at
e
a2 = 4a2 x2 ,
1 + e−2at
a3 = 0,
which match with those calculated in (c).
3.6 For the deformation shown in Problem 3.2 (see Fig. P3.2), determine
(a) the components of the deformation gradient F and its inverse, and
(b) the components of the displacement vector.
Solution: (a) From Problem 3.2 we have the mapping
χ(x) = (4X1 + X2 − 2X1 X2 ) ê1 + (−X1 + 3X2 + 2X1 X2 ) ê2 + X3 ê3
Then
∂x1 ∂x1 ∂x1
F11 = = 4 − 2X2 , F12 = = 1 − 2X1 , F13 = =0
∂X1 ∂X2 ∂X3
∂x2 ∂x2 ∂x3
F21 = = −1 + 2X2 , F22 = = 3 + 2X1 , F23 = =0
∂X1 ∂X2 ∂X3
∂x3 ∂x3 ∂x3
F31 = = 0, F32 = = 0, F33 = =1
∂X1 ∂X2 ∂X3
or, in matrix form,  
4 − 2X2 1 − 2X1 0
[F ] =  −1 + 2X2 3 + 2X1 0 
0 0 1
The jacobian is equal to
J = (4 − 2X2 )((3 + 2X1 ) − (1 − 2X1 )(−1 + 2X2 ) = 13 + 6X1 − 8X2 > 0 for X ∈ κ,
where κ denotes the deformed body. Hence, The inverse of [F ] (nonlinear) is

 
3 + 2X1 −1 + 2X1 0
1
[F ]−1 =  1 − 2X2 4 − 2X1 0 .
J
0 0 J
(b) The displacement vector is
u = x − X = (3X1 + X2 − 2X1 X2 ) ê1 + (−X1 + 2X2 + 2X1 X2 ) ê2 + 0 ê3
so that the components are
u1 (X) = 3X1 + X2 − 2X1 X2 , u2 (X) = −X1 + 2X2 + 2X1 X2 , u3 (X) = 0.

44 SOLUTIONS MANUAL
3.7 The motion of a body is described by the mapping
χ(X) = (X1 + t2 X2 ) ê1 + (X2 + t2 X1 ) ê2 + X3 ê3 , 0 ≤ t < ∞.
Determine
(a) the components of the deformation gradient F and its inverse,
(b) the components of the displacement, velocity, and acceleration vectors,
(c) the position (X1 , X2 , X3 ) of the particle in undeformed configuration that occupies the
position (x1 , x2 , x3 ) = (9, 6, 1) at time t = 2 in the deformed configuration, and
(d) the location at time t = 2 of the particle that later will be located at x = (2, 3, 1) at
time t = 3.
(e) Then plot the deformed shape of a body at times t = 0, 1, 2, and 3, assuming that it is
initially a unit cube.
Solution: (a) The components of F are Fij = (∂xi /∂Xj ),

∂x1 ∂x1 ∂x1
F11 = = 1, F12 = = t2 , F13 = =0
∂X1 ∂X2 ∂X3
∂x2 ∂x2 ∂x3
F21 = = t2 , F22 = = 1, F23 = =0
∂X1 ∂X2 ∂X3
∂x3 ∂x3 ∂x3
F31 = = 0, F32 = = 0, F33 = =1
∂X1 ∂X2 ∂X3
Hence, [F ] and [F ]−1 are
1 t2 0 1 −t2 0
   
1  2
[F ] =  t2 1 0  , [F ]−1 = −t 1 0 .
1 − t4
0 0 1 0 0 1 − t4
Clearly, the inverse does not exist for t = 1, that is, the mapping is singular at t = 0.
(b) The displacement components are
u1 = x1 − X1 = t2 X2 , u2 = x2 − X2 = t2 X1 , u3 = x3 − X3 = 0.
The velocity and acceleration components are

dx1 dx2
v1 = = 2tX2 , v2 = = 2tX1 , v3 = 0.
dt dt
du1 du2
a1 = = 2X2 , a2 = = 2X1 , a3 = 0.
dt dt
(c) The position X of a particle X that occupies the position x = (9, 6, 1) at time t = 2
is given by X = F−1 x:
−1   
1 t2 0 1 −t2 0
      
 X1   x1  9
2 1 2
 −t 1
X2 =  t 1 0  x2 =  0  6
1−t 4
X3 0 0 1 t=2 x3 0 0 1 − t4 1
     
t=2
    
1 −4 0 9 1
1
= −  −4 1 0 6 = 2 .
15
0 0 −15 1 1
   
(d) The material particle in κ0 that will be at (x1 , x2 , x3 ) = (2, 3, 1) in κ at time t = 3

can be determined from     
2 1 9 0  X1 
3 =  9 1 0  X2 .
1 0 0 1 X3
   
By inverting these equations, we obtain

      
 X1  1 −9 0 2  25 
1  1
X2 = − −9 1 0 3 = 15 .
80 80  
X3 0 0 −80 1 80
   
This particle at time t = 2 will be at x ∈ κ

      
 x1  1 4 0  25   85 
1 1
x2 =  4 1 0  15 = 115 .
  80 80 
x3 0 0 1 80 80
  
(e) The deformed shape of the cube (thickness remains unchanged and hence not shown
in the two-dimensional plots) at times t = 0, 1, 2, and 3 are shown in
√ Fig. P3.7. Note
that for t = 1 the cube becomes a sheet of zero thickness, length 2 2 units, depth 1
unit, and oriented at 45◦ from
√ the x1 and x2 axes. For t 6= 1, the cube becomes a prism
of depth 1 unit and side 1 + t4 units. Thus, the mapping is not valid for t = 1, as
notedFigure
earlier.P3.6
x2 , X 2
t =1 t=2 t=3
t=0
x1 , X1
Fig. P3.7
3.8 Homogeneous stretch. Consider a body with deformation mapping of the form
χ(X) = k1 X1 ê1 + k2 X2 ê2 + k3 X3 ê3 ,
where ki 6= 0 are constants. Determine the components of

(a) the deformation gradient F, and
(b) the right and left Cauchy–Green tensors C and B.
Solution: (a) The components of F are Fij = (∂xi /∂Xj ),

∂x1 ∂x1 ∂x1
F11 = = k1 , F12 = = 0, F13 = =0
∂X1 ∂X2 ∂X3
∂x2 ∂x2 ∂x3
F21 = = 0, F22 = = k2 , F23 = =0
∂X1 ∂X2 ∂X3
∂x3 ∂x3 ∂x3
F31 = = 0, F32 = = 0, F33 = = k3 ,
∂X1 ∂X2 ∂X3
46 SOLUTIONS MANUAL
or  
k1 0 0
[F ] =  0 k2 0  .
0 0 k3
(b) The components of the right Cauchy–Green deformation tensor C are

 2 
k1 0 0
T 2
[C] = [F ] [F ] =  0 k2 0  .
0 0 k32
The components of the left Cauchy–Green deformation tensor B are
 2 
k1 0 0
[B] = [F ][F ]T =  0 k22 0  .
0 0 k32
Bye the way, the components of the Cauchy strain tensor B̃ are
 −2 
k1 0 0
−1 −2
[B̃] = [B] =  0 k2 0 .
0 0 k3−2
3.9 Homogeneous stretch followed by simple shear. Consider a body with deformation
mapping of the form
χ(X) = (k1 X1 + e0 k2 X2 ) ê1 + k2 X2 ê2 + k3 X3 ê3 ,
where ki 6= 0 and e0 are constants. Determine the components of
(a) the deformation gradient F, and
(b) the right and left Cauchy–Green tensors C and B.
(c) Then plot representative shapes of a deformed unit square (let k1 = k3 = 1) that are
achievable with this mapping; the suggested cases are (i) k2 /k1 = 1.5, e0 = 0.1; (ii)
k2 /k1 = 1.5, e0 = 0.25; (iii) k2 /k1 = 1.25, e0 = 0.5; and (iv) k2 /k1 = 1.25, e0 = 1.0.
Solution: (a) The components of the deformation gradient F are

   
k1 e0 k2 0 1/k1 −e0 /k1 0
−1
[F ] =  0 k2 0  ⇒ [F ] =  0 1/k2 0 
0 0 k3 0 0 1/k3
(b) The components of the right Green–Cauchy strain tensor C are

k12
 
e0 k1 k2 0
T 2 2
[C] = [F ] [F ] =  e0 k1 k2 (1 + e0 )k2 0  .
0 0 k32
The components of the left Cauchy–Green deformation tensor B are
 2
k1 + e20 k22 e0 k22 0

[B] = [F ][F ]T =  e0 k22 k22 0  .

0 0 k32
Also, note that the components of the Cauchy strain tensor B̃ are
k1−2 −e0 k1−2
 
0
−1 −T −1 −2 2 −2 −2
[B̃] = [B] = [F ] [F ] =  −e0 k1 e0 k1 + k2 0 .
0 0 k3−2
(c) Plots of representative shapes of a deformed unit square (let k1 = k3 = 1) are shown
in Fig. P3.8. For negative values of k2 and e0 can also be tried.
x2 , X 2 a =1.5, e0 = 0.1 a =1.5, e0 = 0.25

1.6 a =1.25, e0 = 0.5
a =1.25, e0 = 1.0
ae0 k1
1.2
0.8
a
0.4
k3 k
=1, a = 2
k1 k1
0.0 x1 , X1
0.0 0.5 1.0 1.5 2.0 2.5
Fig. P3.9
3.10 Suppose that the motion of a continuous medium is given by
x1 = X1 cos At + X2 sin At,

x2 = −X1 sin At + X2 cos At,
x3 = (1 + Bt)X3 , 0 ≤ t < ∞,
where A and B are constants. Determine the components of

(a) the displacement vector in the material description,
(b) the displacement vector in the spatial description,
(c) displacement vector components in the spatial description with respect to a cylindrical
basis, and
(d) the Green–Lagrange and Eulerian strain tensors in the Cartesian coordinate system.
Solution: (a) The displacements in the material description are
u1 (X) = x1 − X1 = X1 (cos At − 1) + X2 sin At,

u2 (X) = −X1 sin At + X2 (cos At − 1) ,
u3 (X) = Bt X3 .
(b) First invert the mapping

         
 x1  cos At sin At 0  X1   X1  cos At − sin At 0  x1 
x2 =  − sin At cos At 0  X2 ⇒ X2 =  sin At cos At 0  x2 .
1
x3 0 0 1 + Bt X3 X3 0 0 x3
       
1+Bt
Then the displacements in the spatial description are
u1 (x) = x1 (1 − cos At) + x2 sin At,

u2 (x) = −x1 sin At + x2 (1 − cos At) ,
Bt
u3 (x) = x3 .
1 + Bt
(c) The displacements in cylindrical coordinates (ur , uθ , uz ) can be written in terms of

(u1 , u2 , u3 ) as
ur = cos θ u1 + sin θ u2
= cos θ [r cos θ (1 − cos At) + r sin θ sin At] + sin θ [−r cos θ sin At + r sin θ (1 − cos At)]
= r (1 − cos At) ,
48 SOLUTIONS MANUAL
uθ == − sin θ u1 + cos θ u2
= − sin θ [r cos θ (1 − cos At) + r sin θ sin At] + cos θ [−r cos θ sin At + r sin θ (1 − cos At)]
= r sin At,
Bt
uz = u3 = z.
1 + Bt
where we have used the transformation equations
x1 = r cos θ, x2 = r sin θ, x3 = z.
(d) The components of the deformation gradient F are

 
cos At sin At 0
[F ] =  − sin At cos At 0 .
0 0 1 + Bt
The components of the Green–Lagrange strain tensor E are

 
00 0
2[E] = [F ]T [F ] − [I] =  0 0 0 .
0 0 (2 + Bt)Bt
The components of the Eulerin strain tensor e are
00 0
 
2[e] = [I] − [F ]−T [F ]−1 =  0 0 0 .

(2+Bt)Bt
00 (1+Bt)2
3.11 If the deformation mapping of a body is given by
χ(X) = (X1 + AX2 ) ê1 + (X2 + BX1 ) ê2 + X3 ê3 ,
where A and B are constants, determine

(a) the displacement components in the material description,
(b) the displacement components in the spatial description, and
(c) the components of the Green–Lagrange and Eulerian strain tensors.
Solution:(a) The displacements in the material description are
u1 (X) = x1 − X1 = AX2 , u2 (X) = x2 − X2 = BX1 , u3 (X) = x3 − X3 = 0.
(b) First invert the mapping

         
 x1  1 A 0  X1   X1  1 −A 0  x1 
1  −B
x2 =  B 1 0  X2 , X2 = 1 0  x2 .
1 − AB
x3 0 0 1 X3 X3 0 0 1 − AB x3
       
Then the displacements in the spatial description are

A B
u1 (x) = (x2 − Bx1 ) , u2 (x) = (x1 − Ax2 ) , u3 (x) = 0.
1 − AB 1 − AB
(c) The components of the deformation gradient F and its inverse are
   
1A 0 1 −A 0
−1 1  −B 1
[F ] =  B 1 0  , [F ] = 0 .
1 − AB
0 0 1 0 0 1 − AB
The components of the Green–Lagrange strain tensor E are

B2 A + B 0
 
T 2
2[E] = [F ] [F ] − [I] =  A + B A 0.
0 0 0
The components of the Euler strain tensor e are
1+B 2
 
A+B
1 − (1−AB) 2 − (1−AB)2 0
2[e] = [I] − [F ]−T [F ]−1 =  − A+B 2 1 − 1+A 2 0  .
 2 
(1−AB) (1−AB)
0 0 0
3.12 For the deformation mapping in Problem 3.2, determine the components of the Green-
Lagrange strain tensor.
Solution: From the solutions of Problems 3.2 and 3.6, we obtain

2[E] = [F ]T [F ] − [I]
    
4 − 2X2 −1 + 2X2 0 4 − 2X2 1 − 2X1 0 1 0 0
=  1 − 2X1 3 + 2X1 0   −1 + 2X2 3 + 2X1 0  −  0 1 0 
0 0 1 0 0 1 0 0 1
2
 
16 − 20X2 + 8X2 1 − 10X1 + 4X2 + 8X1 X2 0
=  1 − 10X1 + 4X2 + 8X1 X2 9 + 8X1 + 8X12 0
0 0 0
Alternatively, the Green–Lagrange strains can be computed directly from the displace-
ment field
u1 (X) = 3X1 + X2 − 2X1 X2 , u2 (X) = −X1 + 2X2 + 2X1 X2 , u3 (X) = 0.
We have
∂u1 ∂u1 ∂u1
= 3 − 2X2 , = 1 − 2X1 , = 0,
∂X1 ∂X2 ∂X3
∂u2 ∂u2 ∂u2
= −1 + 2X2 , = 2 + 2X1 , = 0,
∂X1 ∂X2 ∂X3
∂u3 ∂u2 ∂u1
= 0, = 0, = 0,
∂X1 ∂X2 ∂X3
Then the components of the Green–Lagrange strain tensor are
2 2 2
∂u1 ∂u1 ∂u2 ∂u3
2E11 = 2 + + +
∂X1 ∂X1 ∂X1 ∂X1
= 6 − 4X2 + (3 − 2X2 )2 + (−1 + 2X2 )2 + 0 = 16 − 20X2 + 8X22
2 2 2
∂u2 ∂u1 ∂u2 ∂u3
2E22 =2 + + +
∂X2 ∂X2 ∂X2 ∂X2
= 4(1 + X1 ) + (1 − 2X1 )2 + (2 + 2X1 )2 + 0 = 9 + 8X1 + 8X12
2 2 2
∂u3 ∂u1 ∂u2 ∂u3
2E33 =2 + + + =0
∂X3 ∂X3 ∂X3 ∂X3

∂u1 ∂u2 ∂u1 ∂u1 ∂u2 ∂u2 ∂u3 ∂u3
2E12 = + + + +
∂X2 ∂X1 ∂X1 ∂X2 ∂X1 ∂X2 ∂X1 ∂X2
= 1 − 2X1 − 1 + 2X2 + (3 − 2X1 )((1 − 2X1 ) + (−1 + 2X2 )(2 + 2X1 ) + 0
= 1 − 10X1 + 4X2 + 8X1 X2
∂u1 ∂u3 ∂u1 ∂u1 ∂u2 ∂u2 ∂u3 ∂u3
2E13 = + + + + =0
∂X3 ∂X1 ∂X1 ∂X3 ∂X1 ∂X3 ∂X1 ∂X3
∂u2 ∂u3 ∂u1 ∂u1 ∂u2 ∂u2 ∂u3 ∂u3
2E23 = + + + + =0
∂X3 ∂X2 ∂X2 ∂X3 ∂X2 ∂X3 ∂X2 ∂X3
50 SOLUTIONS MANUAL
3.13 For the deformation field given in Problem 3.7, determine the Green–Lagrange strain
tensor components.
Solution: Then the Green-Lagrange strain components are

2 2 2
∂u1 ∂u1 ∂u2 ∂u3
2E11 = 2 + + + = t4
∂X1 ∂X1 ∂X1 ∂X1
∂u1 ∂u2 ∂u1 ∂u1 ∂u2 ∂u2 ∂u3 ∂u3
2E12 = + + + + = 2t2
∂X2 ∂X1 ∂X1 ∂X2 ∂X1 ∂X2 ∂X1 ∂X2
2 2 2
∂u2 ∂u1 ∂u2 ∂u3
2E22 = 2 + + + = t4
∂X2 ∂X2 ∂X2 ∂X2
Alternatively, the components of E are computed directly from the definition

 4
t 2t2 0

2[E] = [F ]T [F ] − [I] =  2t2 t4 0  ,

0 0 0
where
1 t2 0
 
2
[F ] =  t 1 0  .
0 0 1
3.14 For the deformation mapping given in Problem 3.9, determine the current positions
(x1 , x2 ) of material particles that were on the circle X12 + X22 = R2 with radius R in
the undeformed body.
Solution: From Problem 3.9, the components of the inverse of the deformation gradient
are  
1/k1 −e0 /k1 0
−1
[F ] =  0 1/k2 0 .
0 0 1/k3
Then the positions of the material particles on the circle in the reference configuration
end up on a closed curve defined by
2 2
x1 − e0 x2 x2
X12 + X22 = R2 → + = R2 .
k1 k2
Note that when e0 = 0 and k1 = k2 = k, the circle of particles will also be on a circle
whose radius is kR.
3.15 The motion of a continuous medium is given by

1 1
x1 = (X1 + X2 )et + (X1 − X2 )e−t ,
2 2
1 1
x2 = (X1 + X2 )et − (X1 − X2 )e−t ,
2 2
x3 = X3 , 0 ≤ t < ∞.
Determine
(a) the velocity components in the material description,
(b) the velocity components in the spatial description, and
(c) the components of the rate of deformation and vorticity tensors.
Solution: Note that the given mapping can be expressed in alternative form as
x1 = X1 cosh t + X2 sinh t,
x2 = X1 sinh t + X2 cosh t,
x3 = X3 .
The inverse of the mapping is given by
X1 = x1 cosh t − x2 sinh t,
X2 = −x1 sinh t + x2 cosh t,
X3 = X3 .
(a) The velocity components in the material description are

∂x1 1
(X1 + X2 )et − (X1 − X2 )e−t = X1 sinh t + X2 cosh t,

v1 = =
∂t x=fixed 2

∂x2 1
(X1 + X2 )et + (X1 − X2 )e−t = X1 cosh t + X2 sinh t,

v2 = =
∂t x=fixed 2

∂x3
v3 = = 0,
∂t x=fixed
(b) The velocity components in the spatial description can be obtained by rewriting
them from Part (a) in terms of the spatial coordinates. We obtain
v1 = X1 sinh t + X2 cosh t = (x1 cosh t − x2 sinh t) sinh t + (−x1 sinh t + x2 cosh t) cosh t = x2 ,
v2 = X1 cosh t + X2 sinh t = (x1 cosh t − x2 sinh t) cosh t + (−x1 sinh t + x2 cosh t) sinh t = x1 ,
v3 = 0.
(c) The components of the deformation gradient F are Fij = (∂xi /∂Xj ),
∂x1 ∂x1 ∂x1
F11 = = cosh t, F12 = = sinh t, F13 = =0
∂X1 ∂X2 ∂X3
∂x2 ∂x2 ∂x3
F21 = = sinh t, F22 = = cosh t, F23 = =0
∂X1 ∂X2 ∂X3
∂x3 ∂x3 ∂x3
F31 = = 0, F32 = = 0, F33 = = 1,
∂X1 ∂X2 ∂X3
or  
cosh t sinh t 0
[F ] =  sinh t cosh t 0  .
0 0 1
The components of the velocity gradient tensor L = ∇v [or Lij = (∂vj /∂xi )] are
 ∂v ∂v ∂v 
1 2 3  
∂x ∂x ∂x
 ∂v1 ∂v1 ∂v1  010
1 2 3 
[L] =  ∂x2 ∂x2 ∂x2  = 1 0 0 .
 
∂v1 ∂v2 ∂v3 000
∂x3 ∂x3 ∂x3
The components of the rate of deformation tensor 2D = [∇v + (∇v)T ] are

 
010
1 T

[D] = [L] + [L] = 1 0 0.
2
000
The components of the vorticity tensor 2W = [∇v − (∇v)T ] are

 
000
1 T

[W ] = [L] − [L] = 0 0 0.
2
000
52 SOLUTIONS MANUAL
3.16 Nanson’s formula Let the differential area in the reference configuration be dA. Then
(1) (2)
N̂dA = dX(1) × dX(2) or NI dA = eIJK dXJ dXK ,
where dX(1) and dX(2) are two nonparallel differential vectors in the reference configu-
ration. The mapping from the undeformed configuration to the deformed configuration
maps dX(1) and dX(2) into dx(1) and dx(2) , respectively. Then n̂da = dx(1) × dx(2) .
Show that
n̂ da = JF−T · N̂ dA.

n̂da = dx(1) × dx(2)
and use dxi = FiJ dXJ to obtain
(1) (2)
ni da = eijk dxj dxk
(1) (2)
= eijk FjJ FkK dXJ dXK
Multiplying both sides with FiI and using the identity
elmn det[A] = eijk Ail Ajm Akn ,
we obtain
(1) (2)
FiI ni da = eijk FiI FjJ FkK dXJ dXK
(1) (2)
= eIJK det[F ]dXJ dXK = JNI dA,
from which we obtain (post-multiplying with ÊI )
ni FiI ÊI da = JNI ÊI dA, or n̂ · F da = J N̂ dA
Finally, we have
n̂ da = JF−T · N̂ dA
3.17 Consider a rectangular block of material of thickness h and sides 3b and 4b, and having
a triangular hole as shown in Fig. P3.16. If the material is subjected to the deformation
mapping given in Eq. (3.3.12),
χ(X) = (X1 + γX2 )ê1 + X2 e2 + X3 ê3 ,
determine
(a) the equation of the line BC in the undeformed and deformed configurations,
Figure P3.17
(b) the angle ABC in the undeformed and deformed configurations, and
(c) the area of the triangle ABC in the undeformed and deformed configurations.
ˆ 1 + X2 e
c( X ) = ( X1 + g X 2 )e ˆ 2 + X3 e
ˆ3
X2 x2
χ (X )
4b g 4b g
4b 4b
C C
A B A B
X1 x1
3b 3b
Fig. P3.17
Solution: The components of the deformation gradient and its inverse are
   
1 γ 0 1 −γ 0
−1
[F ] =  0 1 0  , [F ] =  0 1 0  .
0 0 1 0 0 1
The determinant of F is det F = 1, implying that there is no change in the volume of
the block.
(a) The equation of line BC in the undeformed configuration is X2 = c + mX1 , where

the intercept c and slope m are determined using the conditions
X2 = 3b when X1 = b; and X2 = b when X1 = 2b,
from which we obtain m = −2 and c = 5b. The equation of the line in the deformed
configuration is given by
5b 2
X2 = 5b − 2X1 → x2 = − x1 .
Figure P3.16b 1 − 2γ 1 − 2γ
(b) The angle ABC before deformation is equal to θ = 90 − α (cos θ = sin α), where
∇φ(X1 , X2 ) 1
cos α = N̂ · Ê1 , N̂ = = √ (Ê2 + 2Ê1 )
|∇φ| 5
where φ(X1 , X
√2 ) = X2 + 2X1 . Thus, the angle ABC before deformation is given by
α = sin−1 (1/ 5), as shown in the figure below.
C C
β
α
X2 N̂ n̂
x2
θ
X1 A B x1 A B
The angle ABC after deformation can be calculated similarly. We have

∇φ(x1 , x2 ) 1 µ
n̂ = = p (ê2 + µê1 ) , cos β = n̂ · ê1 = p
|∇φ| 1 + µ2 1 + µ2
where φ(x1 , x2 ) = x2 + µx1 and µ = [2/(2 − γ)]. The angle is given by 90 − β.
(c) The area of the deformed ABC is given by

1
AB × BC cos β,
2
where the deformed length of AB is the same as the undeformed length (b) and the
deformed length of BC can be calculated from
n̂ da = JF−T · N̂ dA.
We have     
1 0 0 2√  2 
[F ]−T {N }dA = b  −γ 0 0  1 5b = b 1 − 2γ .
0 0 1 0 0
   
Hence, we have
1 p
n̂ da = b 2ê1 + (1 − 2γ)ê2 = p 2ê1 + (1 − 2γ)ê2 b 5 + 4γ(γ − 1).
5 + 4γ(γ − 1)
54 SOLUTIONS MANUAL
p
Thus, the deformed length of BC is equal to b 5 + 4γ(γ − 1).
3.18 Consider a unit square block of material of thickness h (into the plane of the paper), as
shown in Fig. P3.18. If the block is subjected to a loading that deforms the square block
Figure
into the P3.18
shape shown (with no change in the thickness), (a) determine the deformation
mapping, assuming that it is a complete polynomial in X1 and X2 up to the term
X1 X2 , (b) compute the components of the right Cauchy–Green deformation tensor C
and Green–Lagrange strain tensor E at the point X = (1, 1, 0), and (c) compute the
principal strains and directions at X = (1, 1, 0) for γ = 1.
x2 , X 2 g
3g
1
0 x1 , X1
0 1
Fig. P3.18
Solution: (a) The mapping is the same as that given in Eq. (3.3.16) with γ1 = γ and
γ2 = 3γ. We can determine it as follows. The mapping is of the form
x1 = a1 X1 + a2 X2 + a3 X1 X2 , x2 = b1 X1 + b2 X2 + b3 X1 X2 , x3 = X3 .
Using the given data, we find that
1 = a1 + 0 + 0 → a1 = 1; 0 = b1 + 0 + 0 → b1 = 0
0 = a2 + 0 + 0 → a2 = 0; 1 = 0 + b2 + 0 → b2 = 1
1 + γ = a1 + 0 + a3 → a3 = γ; 1 + 3γ = 0 + b2 + b3 → b3 = 3γ.
Hence the mapping is
x = (X1 + γX1 X2 ) ê1 + (X2 + 3γX1 X2 ) ê2 + ê3 .
(b) From Eq. (3.3.17), the deformation gradient at point X = (1, 1, 0) is

   
1 + γX2 γX1 0 1+γ γ 0
[F ] =  3γX2 1 + 3γX1 0  =  3γ 1 + 3γ 0  .
0 0 1 X =1,X =1 0 0 1
1 2
The components of the right Cauchy–Green deformation tensor C are

  
1 + γ 3γ 0 1+γ γ 0
T
[C] = [F ] [F ] =  γ 1 + 3γ 0   3γ 1 + 3γ 0 
0 0 1 0 0 1
2 2
 
(1 + γ) + 9γ (1 + γ)γ + 3(1 + 3γ)γ 0
=  (1 + γ)γ + 3(1 + 3γ)γ γ 2 + (1 + 3γ)2 0.
0 0 1
The components of the Green–Lagrange strain tensor are
2γ + 10γ 2 4γ + 10γ 2 0
 
1
1 2 2
[E] = 2 [C] − [I] = 4γ + 10γ 6γ + 10γ 0  .
2
0 0 0
Alternatively, the Green–Lagrange strain tensor components can be computed from the
displacement field directly. We have the displacement components
u1 = γ X1 X2 , u2 = 3γ X1 X2 , u3 = 0.
Then the components of the Green–Lagrange strain tensor are
" 2 2 2 #
∂u1 1 ∂u1 ∂u2 ∂u3
E11 = +2 + +
∂X1 ∂X1 ∂X1 ∂X1
= γ X2 + 12 γ 2 X22 + 9γ 2 X22 + 0

" 2 2 2 #
∂u2 ∂u1 ∂u2 ∂u3
E22 = + 12 + +
∂X2 ∂X2 ∂X2 ∂X2
1
2 2 2 2
= 3γ X1 + 2 γ X1 + 9γ X1 + 0
" 2 2 2 #
∂u3 1 ∂u1 ∂u2 ∂u3
E33 = +2 + + =0
∂X3 ∂X3 ∂X3 ∂X3

∂u1 ∂u2 ∂u1 ∂u1 ∂u2 ∂u2 ∂u3 ∂u3
2E12 = + + + +
∂X2 ∂X1 ∂X1 ∂X2 ∂X1 ∂X2 ∂X1 ∂X2
= γ(X1 + 3X2 ) + γ 2 X1 X2 + 9γ 2 X1 X2 + 0

∂u1 ∂u3 ∂u1 ∂u1 ∂u2 ∂u2 ∂u3 ∂u3
2E13 = + + + + =0
∂X3 ∂X1 ∂X1 ∂X3 ∂X1 ∂X3 ∂X1 ∂X3

∂u2 ∂u3 ∂u1 ∂u1 ∂u2 ∂u2 ∂u3 ∂u3
2E23 = + + + + =0
∂X3 ∂X2 ∂X2 ∂X3 ∂X2 ∂X3 ∂X2 ∂X3
Evaluating EIJ at X = (1, 1, 0), we obtain
2γ + 10γ 2 4γ + 10γ 2 0
 
[E] = 21  4γ + 10γ 2 6γ + 10γ 2 0  .

0 0 0
(c) The principal strains for γ = 1 are determined from

6−λ 7 0
7 8 − λ 0 = 0 → λ2 − 14λ − 1 = 0

0 0 0
or, the principal strains are
√ √
Figure P3.19 λ1 = 7 − 5 2 = −0.071, λ2 = 7 + 5 2 = 14.071
The eigenvectors are given by
n̂1 = −0.7554 ê1 + 0.6552 ê2 , n̂2 = 0.6552 ê1 + 0.7554 ê2 .
3.19 Determine the displacements and Green–Lagrange strain tensor components for the
deformed configuration shown in Fig. P3.19. The undeformed configuration is shown
in dashed lines. Assume that the deformation mapping is a linear polynomial of X1
and X2 (note that for this case the constant terms are zero).
x2 , X 2
D e0 e0 x1 = a0 + a1 X1 + a2 X 2 + a3 X1 X 2 ,
D C C x =b +b X +b X +b X X ,
2 0 1 1 2 2 3 1 2
b x3 = X 3 .
A B x1 , X1
A a B
56 SOLUTIONS MANUAL
Fig. P3.19
Solution: By inspection, it is clear that the deformed configuration is a quadrilateral

(parallelogram). Hence, assume that a0 = b0 = a3 = b3 = 0
x1 = a1 X1 + a2 X2 , x2 = b1 X1 + b2 X2 , x3 = X3
and determine ai and bi (i = 1, 2) using the following four conditions:
X = (a, 0, X3 ), x = (a, 0, x3 ); X = (0, b, X3 ), x = (e0 , b, x3 )
We obtain a1 = 1, a2 = e0 /b, b1 = 0, and b2 = 1. Hence, the mapping becomes

e0
x1 = X1 + X2 , x2 = X2 , x3 = X3 .
b
The displacement components are given by
e0
u1 = x1 − X1 = X2 , u2 = x2 − X2 = 0, u3 = 0,
b
from which the Green–Lagrange strain components can be calculated as
e0 1 e 0 2
E11 = 0, E12 = , E22 = .
2b 2 b
All other strain components are zero.
3.20 Determine the displacements and Green–Lagrange strain components for the deformed
Figure P3.20
configuration shown in Fig. P3.20. The undeformed configuration is shown in dashed
lines. Use the suggested form of the deformation mapping, as implied by the deformed
configuration.
x2 , X 2
x1 = a1 X1 + a4 X 22 ,
C x 2 = b1 X1 + b2 X 2 ,
D e0 D e0
• • •C x3 = X 3 .
D
b
x1 = a1 X1 + a4 X 22
A B
• • x1 , X1
A a B
Fig. P3.20
Solution: By inspection, it is clear that the deformed configuration can be expressed

by the mapping
x1 = c1 X1 + c2 X22 , x2 = c3 X1 + c4 X2 , x3 = X3 .
We determine ci (i = 1, 2, 3, and 4) using the following four conditions:
X = (a, 0, X3 ), x = (a, 0, x3 ); X = (0, b, X3 ), x = (e0 , b, x3 ).

2
We obtain c1 = 1, c3 = 0, c2 = e0 /b , and c4 = 1. Therefore, the mapping is
e0 2
x1 = X1 + X2 , x2 = X2 , x3 = X3 .
b2
The displacements are given by
e0
u1 = x1 − X1 = ( )X22 , u2 = 0, u3 = 0.
b2
The Lagrange–Green strain tensor components are given by

e0 1 e0 2
E11 = 0, E12 = 2 X2 , E22 = 2X2 2 .
b 2 b
All other strain components are zero.
3.21 Determine the displacements and Green–Lagrange strains in the (x1 , x2 , x3 ) system
Figure P3.21for the deformed configuration shown in Fig. P3.21. The undeformed configuration is
shown in dashed lines. Use the suggested form of the deformation mapping (for this
case the constant terms are zero).
x2 , X 2
e1
C
D C e2 x1 = a1 X1 + a2 X 2 + a3 X1 X 2 ,
x 2 = b1 X1 + b2 X 2 + b3 X1 X 2 ,
b
x3 = X 3 .
x1 , X1
A a B
Fig. P3.21
Solution: The displacement field can be determined directly (without determining the
mapping) from the information shown in the figure. The displacement is of the form
(based on the form of the mapping given)
ui (X1 , X2 ) = ci00 + ci10 X1 + ci01 X2 + ci11 X1 X2 ,
where the constants cij are determined using the conditions at points A, B, C, and D.
We have
ui (0, 0) = 0 → ci00 = 0; ui (a, 0) = 0 → ci10 = 0; ui (0, b) = 0 → ci01 = 0,
e1 e2
u1 (a, b) = e1 → c111 = ; u2 (a, b) = e2 → c211 = ;
ab ab
X1 X2 X1 X2
u1 (X1 , X2 ) = e1 , u2 (X1 , X2 ) = e2 .
a b a b
The Green–Lagrange strain tensor components are
" 2 2 #
e21 + e22

∂u1 1 ∂u1 ∂u2 e1 X2 1 2
E11 = + + = + X2
∂X1 2 ∂X1 ∂X1 a b 2 a2 b2
" 2 2 #
e21 + e22

∂u2 1 ∂u1 ∂u2 e2 X1 1 2
E22 = + + = + X1
∂X2 2 ∂X2 ∂X2 b a 2 a2 b2
e21 + e22

∂u1 ∂u2 ∂u1 ∂u1 ∂u2 ∂u2 e1 X1 e2 X2
2E12 = + + + = + + X1 X2 .
∂X2 ∂X1 ∂X1 ∂X2 ∂X1 ∂X2 b a a b a2 b2
3.22 Determine the displacements and Green–Lagrange strains for the deformed configura-
tion shown in Fig. P3.22. The undeformed configuration is shown in dashed lines. Use
the suggested form of the deformation mapping (note that constant terms are zero for
this case).
Solution: By inspection, it is clear that the deformed configuration can be expressed

by the mapping
x1 = c1 + c2 X1 + c3 X2 + c4 X1 X2 , x2 = c5 + c6 X1 + c7 X2 + c8 X1 X2 , x3 = X3 .
58 SOLUTIONS MANUAL
We determine ci using the following pairs of four conditions:

X = (0, 0, X3 ), x = (0, 0, x3 ); X = (1, 0, X3 ), x = (0.8, 0.2, x3 );
FigureXP3.22
= (0, 1, X3 ), x = (0.5, 0.9, x3 ); X = (1, 1, X3 ), x = (1.3, 1.2, x3 ).
x2 , X 2
κ0 κ
C
D C
A: (X1 , X 2 ) = (0,0) A: (x1 , x 2 ) = (0,0)
D
B: (X1 , X 2 ) = (1,0) B: (x1 ,x 2 ) = (0.8,0.2)
κ0
1 κ C: (X1 , X 2 ) = (1,1) C: (x1 ,x 2 ) = (1.3,1.2)
D: (X1 , X 2 ) = (0,1) D: (x1 , x 2 ) = (0.5,0.9)
A B
x1 , X 1
A 1 B
x1 = a0 + a1 X1 + a2 X 2 + a3 X1 X 2 ; x2 = b0 + b1 X1 + b2 X 2 + b4 X1 X 2 ; x3 = X 3
Fig. P3.22
We obtain c1 = c4 = c5 = 0, c2 = 0.8, c3 = 0.5, c6 = 0.2, c7 = 0.9, and c8 = 0.1. Thus,

the motion is defined by
x1 = 0.8X1 + 0.5X2 , x2 = 0.2X1 + 0.9X2 + 0.1X1 X2 , x3 = X3 .
The displacements are given by
u1 = x1 − X1 = −0.2X1 + 0.5X2 , u2 = 0.2X1 − 0.1X2 + 0.1X1 X2 .
The Lagrange–Green strain tensor components are given by
E11 = −0.2 + 0.5 (−0.2)2 + (0.2 + 0.1X2 )2 ,

2E12 = 0.5 + (0.2 + 0.1X2 ) + (−0.2)(0.5) + (0.2 + 0.1X2 )(−0.1 + 0.1X1 ),

E22 = −0.1 + 0.1X1 + 0.5 (0.5)2 + (−0.1 + 0.1X1 )2 .

3.23 Given the following displacement vector in a material description using a cylindrical
coordinate system
u = Arêr + Brzêθ + C sin θêz ,
where A, B, and C are constants, determine the infinitesimal strains. Here (r, θ, z)
denote the material coordinates.
Solution: Using the strain-displacement relations in Eq. (3.5.26), we obtain

∂ur
εrr = =A
∂r
1 1 ∂ur ∂uθ uθ 1
εrθ = + − = (0 + Bz − Bz) = 0,
2 r ∂θ ∂r r 2

1 ∂ur ∂uz 1
εrz = + = (0 + 0) = 0,
2 ∂z ∂r 2
ur 1 ∂uθ
εθθ = + = A + 0 = A,
r r ∂θ
1 ∂uθ 1 ∂uz 1 C
εzθ = + = Br + cos θ ,
2 ∂z r ∂θ 2 r
∂uz
εzz = = 0.
∂z
3.24 Show that the components of the Green–Lagrange strain tensor in cylindrical coordinate
system are given by
" 2 2 2 #
∂ur 1 ∂ur ∂uθ ∂uz
Err = + + + ,
∂r 2 ∂r ∂r ∂r
1 1 ∂ur ∂uθ uθ 1 ∂ur ∂ur 1 ∂uθ ∂uθ

Erθ = + − + +
2 r ∂θ ∂r r r ∂r ∂θ r ∂r ∂θ
!
1 ∂uz ∂uz ur ∂uθ uθ ∂ur
+ + − ,
r ∂r ∂θ r ∂r r ∂r

1 ∂ur ∂uz ∂ur ∂ur ∂uθ ∂uθ ∂uz ∂uz
Erz = + + + + ,
2 ∂z ∂r ∂r ∂z ∂r ∂z ∂r ∂z
" 2 2 2
ur 1 ∂uθ 1 1 ∂ur 1 ∂uθ 1 ∂uz
Eθθ = + + + +
r r ∂θ 2 r ∂θ r ∂θ r ∂θ
#
2 ∂ur 2 ∂uθ uθ 2
ur 2

− uθ + 2 ur + + ,
r2 ∂θ r ∂θ r r
1 ∂uθ 1 ∂uz 1 ∂ur ∂ur 1 ∂uθ ∂uθ

Eθz = + + +
2 ∂z r ∂θ r ∂θ ∂z r ∂θ ∂z
!
1 ∂uz ∂uz uθ ∂ur ur ∂uθ
+ − + ,
r ∂θ ∂z r ∂z r ∂z
" 2 2 2 #
∂uz 1 ∂ur ∂uθ ∂uz
Ezz = + + + .
∂z 2 ∂z ∂z ∂z
Solution: Because the linear terms are already available in Eq. (3.5.26), we focus
attention on the nonlinear term
∇0 u · (∇0 u)T

∂ur ∂uθ 1 ∂ur ∂uz ∂ur
= êr êr + êr êθ + êθ êr − uθ + êr êz + êz êr
∂r ∂r r ∂θ ∂r ∂z

1 ∂uθ 1 ∂uz ∂uθ ∂uz
+ ur + êθ êθ + êθ êz + êz êθ + êz êz ·
r ∂θ r ∂θ ∂z ∂z

∂ur ∂uθ 1 ∂ur ∂uz ∂ur
êr êr + êθ êr + êr êθ − uθ + êz êr + êr êz
∂r ∂r r ∂θ ∂r ∂z

1 ∂uθ 1 ∂uz ∂uθ ∂uz
+ êθ êθ ur + + êz êθ + êθ êz + êz êz
r ∂θ r ∂θ ∂z ∂z

∂ur 1 ∂ur ∂ur ∂ur 1 ∂ur ∂ur
= êr + êθ − uθ + êz êr + êθ − uθ + êz
∂r r ∂θ ∂z ∂r r ∂θ ∂z

∂uθ 1 ∂uθ ∂uθ ∂uθ 1 ∂uθ ∂uθ
+ êr + êθ ur + + êz êr + êθ ur + + êz
∂r r ∂θ ∂z ∂r r ∂θ ∂z

∂uz 1 ∂uz ∂uz ∂uz 1 ∂uz ∂uz
+ êr + êθ + êz êr + êθ + êz
∂r r ∂θ ∂z ∂r r ∂θ ∂z
60 SOLUTIONS MANUAL
or
" 2 2 2 #
∂ur ∂uθ ∂uz
∇0 u · (∇0 u)T = êr êr + +
∂r ∂r ∂r
" 2 2 2 #
1 ∂ur ∂uθ ∂uz
+ 2 êθ êθ − uθ + ur + +
r ∂θ ∂θ ∂θ
" 2 2 2 #
∂ur ∂uθ ∂uz
+ êz êz + +
∂z ∂z ∂z

1 ∂ur ∂ur ∂uθ ∂uθ ∂uz ∂uz
+ êr êθ − uθ + ur + +
r ∂r ∂θ ∂r ∂θ ∂r ∂θ

∂ur ∂ur ∂uθ ∂uθ ∂uz ∂uz
+ êr êz + +
∂r ∂z ∂r ∂z ∂r ∂z

1 ∂ur ∂ur ∂uθ ∂uθ ∂uz ∂uz
+ êθ êz − uθ + ur + + + ···
r ∂θ ∂z ∂θ ∂z ∂θ ∂z
3.25 The two-dimensional displacement field in a body is given by
u1 = X1 X12 X2 + c1 2c32 + 3c22 X2 − X23 ,

3 3 2 1 3 3 2
u2 = −X2 2c2 + c2 X2 − X2 + c1 X1 X2 ,
2 4 2
where c1 and c2 are constants. Find the linear and nonlinear Green–Lagrange strains.
Solution: The nonlinear Green–Lagrange strain tensor components are

" 2 2 2 #
∂u1 1 ∂u1 ∂u2 ∂u3
E11 = + + +
∂X1 2 ∂X1 ∂X1 ∂X1
= 3X12 X2 + c1 2c32 + 3c22 X2 − X23

1 2 2 1 2
+ 3X1 X2 + c1 2c32 + 3c22 X2 − X23 + −3c1 X1 X22 ,
2 " 2
2 2 2 #
∂u2 1 ∂u1 ∂u2 ∂u3
E22 = + + +
∂X2 2 ∂X2 ∂X2 ∂X2
= − 2c32 + 3c22 X2 − X23 + 3c1 X12 X2

1 2 1 2
+ X12 X12 + c1 3c22 − 3X22 2c32 + 3c22 X2 − X23 + 3c1 X12 X2 ,

+
2 2
∂u1 ∂u2 ∂u1 ∂u1 ∂u2 ∂u2 ∂u3 ∂u3
2E12 = + + + +
∂X2 ∂X1 ∂X1 ∂X2 ∂X1 ∂X2 ∂X1 ∂X2
= X1 X12 + c1 3c22 − 3X22 − 3c1 X1 X22

+ 3X12 X2 + c1 2c32 + 3c22 X2 − X23 X1 X12 + c1 3c22 − 3X22

+ 3c1 X1 X22 2c32 + 3c22 X2 − X23 + 3c1 X12 X2 .

All other components are zero. The linear components are given by the first line of
each expression
∂u1
= 3X12 X2 + c1 2c32 + 3c22 X2 − X23 ,

ε11 =
∂X1
∂u2
= − 2c32 + 3c22 X2 − X23 + 3c1 X12 X2 ,

ε22 =
∂X2
∂u1 ∂u2
= X1 X12 + c1 3c22 − 3X22 − 3c1 X1 X22 .

2ε12 = +
∂X2 ∂X1
3.26 Find the axial strain in the diagonal element, ĀC̄, of Problem 3.19, using
(a) the basic definition of normal strain, and
(b) the strain transformation equations.
Solution: (a) Using the basic definition, we obtain

r r
ĀC̄ − AC b2 + (a + e0 )2 e2 + 2e0 a
εnn = = 2 2
− 1 = 1 + 02 −1
AC b +a b + a2
2 2
1 e0 + 2e0 a 1 e0 + 2e0 a e0 a
≈1+ −1= ≈ 2 (if e0 is small).
2 b2 + a2 2 b2 + a2 b + a2
(b) Using the strain transformation equations we have
ε0ij = ìm `jn εmn ,

ε011 = β11
2
ε11 + `212 ε22 + 2`11 `12 ε12 ,
ε012 = `11 `21 ε11 + `11 `22 ε12 + `12 `21 ε12 + `12 `22 ε22 ,
where ìj = ê0i · êj and

a b
`11 = cos θ = √ = `22 , `12 = sin θ = √ , `21 = − sin θ.
a2 + b2 a2 + b2
Using the strain components E11 = 0, E12 = e0 /2b, and E22 = e20 /2b2 from Problem
3.19, we obtain
0 e20 + 2e0 a ae0 0 (a2 − b2 ) + ae20 e0 a2 − b2

E11 (= Enn ) = ≈ 2 , E12 (= Ens ) = ≈ ( ).
2(a2 + b2 ) a + b2 2
2b(b + a )2 2b a2 + b2
3.27 The biaxial state of strain at a point is given by ε11 = 800×10−6 in./in., ε22 = 200×10−6
in./in., ε12 = 400 × 10−6 in./in. Find the principal strains and their directions.
Solution: The eigenvalue problem for the strain is given by

800 − λ 400
400 200 − λ = 0 → λ(1000 − λ) = 0.

Thus, the principal strains are ε1 = 0 and ε2 = 10−3 in./in. The principal direction
associated with ε1 = 0 is A1 = ê1 − 2ê2 and that associated with ε = 10−3 is A2 =
2ê1 + ê2 . Note that A1 · A2 = 0; that is, they are orthogonal.
3.28 Show that the invariants J1 , J2 , and J3 of the Green–Lagrange strain tensor E can be
expressed in terms of the principal values λi of E as
J1 = λ1 + λ2 + λ3 , J2 = λ1 λ2 + λ2 λ3 + λ3 λ1 , J3 = λ1 λ2 λ3 .
Of course, the above result holds for any second-order tensor.
Solution: The principal invariants of E are

1
(tr E)2 − tr E2 , J3 = det E.

J1 = tr E, J2 =
2
Because E can be expressed with respect to any basis, we choose the eigenvectors
associated with the principal values as the basis. The tensor components of E with
respect to this basis are  
λ1 0 0
[E] =  0 λ2 0  .
0 0 λ3
62 SOLUTIONS MANUAL
Therefore, by definition
J1 = Eii = λ1 + λ2 + λ3 ,
1
(λ1 + λ2 + λ3 )2 − λ21 + λ22 + λ23 = λ1 λ2 + λ2 λ3 + λ3 λ1 ,

J2 =
2
J3 = |E| = λ1 λ2 λ3 .
3.29 Given the displacement field in the cylindrical coordinate system
ur = U (r), uθ = 0, uz = 0,
where U (r) is a function of only r, determine the Green–Lagrange strain components.
Solution: The only nonzero Green–Lagrange strain components associated with the
displacement field are [see Eq. (7.2.3)]
2 2
dU 1 dU U 1 U
Err = + , Eθθ = + , Ezz = Erθ = Erz = Eθz = 0.
dr 2 dr r 2 r
3.30 Given the displacement field in the spherical coordinate system
uR = U (R), uφ = 0, uθ = 0,
where U (r) is a function of only r, determine the Green–Lagrange strain components.
Solution: The only nonzero Green–Lagrange strain components associated with the
displacement field are [see Eq. (7.2.4)]
2 2
dU 1 dU U 1 U
ERR = + , Eφφ = Eθθ = + .
dR 2 dR R 2 R
3.31 Show that the components of the spin tensor W in cylindrical coordinate system are

1 1 ∂vr vθ ∂vθ
Wθr = − − = −Wrθ ,
2 r ∂θ r ∂r

1 ∂vr ∂vz
Wzr = − = −Wrz ,
2 ∂z ∂r

1 1 ∂vz ∂vθ
Wθz = − = −Wzθ .
2 r ∂θ ∂z
Solution: From Eq. (2.5.27), the gradient of the velocity vector in the cylindrical
coordinate system is given by
∂vr ∂vθ ∂vz
∇v = êr êr + êr êθ + êr êz
∂r ∂r ∂r
1 ∂vr ∂vθ ∂vz
+ êθ êr + vr êθ êθ + êθ êθ − vθ êθ êr + êθ êz
r ∂θ ∂θ ∂θ
∂vr ∂vθ ∂vz
+ êz êr + êz êθ + êz êz
∂z ∂z ∂z
∂vr ∂vθ ∂vz
(∇v)T = êr êr + êθ êr + êz êr
∂r ∂r ∂r
1 ∂vr ∂vθ ∂vz
+ êr êθ + vr êθ êθ + êθ êθ − vθ êr êθ + êz êθ
r ∂θ ∂θ ∂θ
∂vr ∂vθ ∂vz
+ êr êz + êθ êz + êz êz
∂z ∂z ∂z
Therefore, we have
1h i
W= ∇v − (∇v)T
2
∂vθ ∂vz 1 ∂vr ∂vz
= êr êθ + êr êz + êθ êr − vθ êθ êr + êθ êz
∂r ∂r r ∂θ ∂θ
∂vr ∂vθ ∂vθ ∂vz
+ êz êr + êz êθ − êθ êr − êz êr
∂z ∂z ∂r ∂r
1 ∂vr ∂vz ∂vr ∂vθ
− êr êθ − vθ êr êθ − êz êθ − êr êz − êθ êz
r ∂θ ∂θ ∂z ∂z
from which follows the result.
3.32 If D = 0, show that

v =w×x+c (vi = eijk wj xk + ci ),
where both w (vorticity vector) and c are constant vectors.

dvi = (Dij + Wij ) dxj = Wij dxj = −eijk wk dxj
Integrating with respect to xj and noting that w is a constant vector, we arrive at
vi = −eijk wk xj + ci = eikj wk xj + ci = eijk wj xk + ci ,
where ci is a constant of integration. Thus, we have
v = w × x + c.
3.33 Show that (a)

1 1 T
Ė = Ċ = Ḟ · F + FT · Ḟ
2 2
and (b)
1
W= Ḟ · F−1 − F−T · ḞT
2
Solution: These follow very simply from the definitions in Eqs. (3.4.11), (3.6.3), and
(3.6.14).
(a) Carrying out the indicated differentiation on the right side of Eq. (3.4.11), we
obtain
dE 1 1 T
= Ċ = Ḟ · F + FT · Ḟ .
dt 2 2
(b) First we note that
∇v = F−T · ∇0 v = F−T · ḞT .
Then from Eqs. (3.6.2) and (3.6.9), we have
2W = Ḟ · F−1 − F−T · ḞT .
3.34 Verify that

∂v 1
+ grad (v · v) + 2W · v
v̇ =
∂t 2
∂v 1
= + grad (v · v) + 2w × v,
∂t 2
where W is the spin tensor and w is the vorticity vector [see Eq. (3.6.5)].
Solution: From Part (e) of Problem 2.32 it follows that (replace A with v and curl v
with 2w)
∂v
v̇ = + v · ∇v
∂t
∂v 1
= + grad (v · v) + 2w × v,
∂t 2
64 SOLUTIONS MANUAL
Next, we must prove that w × v = W · v. We have

w × v = eijk wi vj êk = (ejki wi )vj êk = −Wjk vj êk = Wkj vj êk = W · v.
3.35 Show that

DJ
= (∇ · v) J.
Dt
∂xk
Hints: Dx
Dt
i
= vi and ∂X∂vi
j
= ∂vi
∂xk ∂Xj
. Also, see the list of properties of determinants
highlighted in Section 2.3.6.
Solution: By definition, J is the determinant of the deformation tensor F. Hence, we

can write ∂x ∂x ∂x
1 1 1
∂X1 ∂X2 ∂X3
∂xi ∂xj ∂xk ∂x ∂x ∂x
J = eijk 2
= ∂X1 ∂X2 ∂X23 .
2
(1)
∂X1 ∂X2 ∂X3 ∂x3 ∂x3 ∂x3
∂X ∂X ∂X
1 2 3
We note that the reference coordinates X = (X1 , X2 , X3 ) are not a function of time t.
Using the definition of the material derivative (that is, derivative with respect to time
when X is fixed), we can write

DJ ∂vi ∂xj ∂xk ∂xi ∂vj ∂xk ∂xi ∂xj ∂vk
= eijk + + , (2)
Dt ∂X1 ∂X2 ∂X3 ∂X1 ∂X2 ∂X3 ∂X1 ∂X2 ∂X3
where we have used the identities
∂ 2 xi

∂ ∂xi ∂ ∂xi ∂vi ∂δim
= = , = = 0.
∂t ∂XK ∂XK ∂t ∂XK ∂xm ∂XK ∂XK
We also note that
∂vi ∂vi ∂xm ∂vi ∂x1 ∂vi ∂x2 ∂vi ∂x3
= = + + . (3)
∂XJ ∂xm ∂XJ ∂x1 ∂XJ ∂x2 ∂XJ ∂x3 ∂XJ
Thus, from Eq. (2), we have
∂v ∂x
1 m ∂v1 ∂xm ∂v1 ∂xm ∂x1 ∂x1 ∂x1
∂xm ∂X1 ∂xm ∂X2 ∂xm ∂X3 ∂X1 ∂X2 ∂X3
DJ ∂x
2 ∂x2 ∂x2
∂v ∂x
+ 2 m ∂v2 ∂xm ∂v2 ∂xm

= ∂X ∂X2 ∂X3 ∂xm ∂X1 ∂xm ∂X2 ∂xm ∂X3

Dt ∂x3
1
∂X ∂x3 ∂x3
∂x3
∂X ∂x3 ∂x3

1 ∂X2 ∂X3 1 ∂X2 ∂X3
∂x ∂x1 ∂x1

1
∂X1 ∂X2 ∂X3
∂x ∂x2 ∂x2

2
+ ∂X 1 ∂X2 ∂X3
.

∂v3 ∂xm ∂v3 ∂xm ∂v3 ∂xm
∂x ∂X ∂x ∂X ∂x ∂X
m 1 m 2 m 3
Now consider the first determinant

∂v ∂x
1 m ∂v1 ∂xm ∂v1 ∂xm ∂x1 ∂x1 ∂x1 ∂x2 ∂x2 ∂x2
∂xm ∂X1 ∂xm ∂X2 ∂xm ∂X3 ∂X1 ∂X2 ∂X3

∂v1 ∂X1 ∂X2 ∂X3

∂x
2 ∂x2 ∂x2

= ∂v 1 ∂x2 ∂x2 ∂x2 +
∂x2 0 0 0

∂X1 ∂X2 ∂X3 ∂x1 ∂X1 ∂X2 ∂X3

∂x3 ∂x3 ∂x3
∂x3 ∂x3 ∂x3 0
∂X
1 ∂X 2 ∂X

3
∂X1 ∂X2 ∂X3
0 0
∂x3 ∂x3 ∂x3
∂X ∂X2

∂X3
1
∂v1 ∂v1
+ 0 0 0 = J + 0 + 0.

∂x3 ∂x1
0 0 0
Similarly, the remaining two determinants give, respectively

∂v2 ∂v3
J and J.
∂x2 ∂x3
Thus, we have
DJ ∂v1 ∂v2 ∂v3
= + + J = (∇ · v)J.
Dt ∂x1 ∂x2 ∂x3
3.36 Establish the identities
dv = L · dx, and (∇0 v)T = L · F.
Solution: We begin with the definition of L

T
∂vi ∂vi
L = Lij êi êj = (∇v)T = êj êi = êi êj .
∂xj ∂xj
Taking the dot product with êk from the right, we obtain
∂vi ∂v
Lik êi = êi = ,
∂xk ∂xk
or
dv = Lik êi dxk = êi Lin δnm dxm = (êi Lin ên ) · êm dxm = L · dx = dx · LT .
Next consider
∂vi
∇0 v = ÊJ êi .
∂XJ
Using the chain rule of differentiation
∂vi ∂xk ∂vi
ÊJ êi = ÊJ êi = FkJ Lik ÊJ êi = Lik FkJ ÊJ êi = (L · F)T .
∂XJ ∂XJ ∂xk
Therefore, it follows that
(∇0 v)T = L · F.
3.37 Show that Ċ = 2FT · D · F, where C, D, and F are the right Cauchy–Green tensor,
rate of deformation tensor, and deformation gradient, respectively.
Solution: We begin with the definition of C in terms of F and carry out the indicated
differentiation with respect to time to obtain
d T
Ċ = F · F = ḞT · F + FT · Ḟ
dt
= (FT · LT ) · F + FT · (L · F)
= FT · (LT + L) · F = 2FT · D · F,
where Eq. (3.6.15), Ḟ = L · F, is made use of in arriving at the last result. The result
also can be established using 2E = C − I [see Eq. (3.4.11) and Ė = FT · D · F (see Eq.
(3.6.10)].
3.38 Show that the Eulerian strain rate is given by

ė = D − e · L + LT · e ,
and
ε̇ = D.
Solution: From Eq. (3.4.23) we have (see the discussions of Section 3.6.2)
d d d
[(ds)2 − (dS)2 ] = [(ds)2 ] = 2 [dx · e · dx]
dt dt dt
= 2 dv · e · dx + 2 dx · ė · dx + 2 dx · e · dv,
= 2 (L · dx) · e · dx + 2 dx · ė · dx + 2 dx · e · (L · dx),

= 2 dx · LT · e + ė + e · L · dx,
66 SOLUTIONS MANUAL
From Eq. (3.6.9), the left-hand side is equal to 2 dx · D · dx. Thus, we have

D = LT · e + ė + e · L ,
from which we have the required result.

We have
dε d h i h i h i
2 = ∇u + (∇u)T = ∇u̇ + (∇u̇)T = ∇v + (∇v)T = 2D.
dt dt
3.39 Use the index notation to establish the compatibility conditions in Eq. (3.7.11)
∇0 × (∇0 × ε0 )T = 0
for the infinitesimal strains. Hint: Begin with ∇0 × ε and use Eq. (3.5.15).
Solution: We begin with the curl of ε:

∂ 2 uj ∂ 2 ur

∂εjr
∇0 × ε = eijk Êk Êr = 21 eijk + Êk Êr
∂Xi ∂Xi ∂Xr ∂xi ∂xj
∂ 2 uj

= 12 eijk + 0 Êk Êr .
∂Xi ∂Xr
Using Eq. (3.5.15) we have
∂ 2 uj

∂ ∂uj ∂ωk
∇ × ε = 12 eijk Êk Êr = 1
e
2 kij
Êk Êr = Êk Êr
∂Xi ∂Xr ∂Xr ∂Xi ∂Xr
or
∂ωk ∂
(∇ × ε)T = êr êk = êr (ωk êk ) = ∇ω.
∂xr ∂xr
Because the curl of the gradient of a tensor of any order is zero [see Part (b) of Problem
2.31), we take curl of the above equation and arrive at the compatibility equation in
vector/tensor form,
∇ × (∇ × ε)T = 0.
3.40 Show that the following second-order tensor is symmetric:

S = ∇ × (∇ × ε)T .
Solution: We begin with the definition and find its transpose

h iT
ST = ∇ × (∇ × ε)T = (∇ × ε) × (∇)T
←
−
= (∇ × ε) × ∇ = ∇ × (∇ × ε)T .
Alternatively, we have
Srs = eikr ejls εij,kl = eiks ejlr εij,kl = Ssr ,
where first i is renamed as j and k is renamed as l, and then the symmetry of the strain
components and symmetry of ∂ 2 /∂xk ∂xk = ∂ 2 /∂xk ∂xk are used in arriving at the last
result.
We can also show the symmetry of S from the definition
S = ∇ × (∇ × ε)T
T
∂εjk ∂εjk
= ∇ × eijr êr êk = ∇ × eijr êk êr
∂xi ∂xi
∂ 2 εjk ∂ 2 εkj
= eijr eskp êp êr = eskp eijr êr êp
∂xs ∂xi ∂xi ∂xs
2
∂ εkj
= eskr eijp êp êr = ST ,
∂xi ∂xs
where first i is renamed as j and k is renamed as l, and then the symmetry of the strain
components and symmetry of ∂ 2 /∂xk ∂xk = ∂ 2 /∂xk ∂xk are used in arriving at the last
result.
3.41 Let [see the compatibility conditions in Eqs. (3.7.4)–(3.7.9)]

∂ 2 ε11 ∂ 2 ε22 ∂ 2 ε12
−S33 = R3 = 2
+ 2
−2 , (1)
∂X2 ∂X1 ∂X1 ∂X2
∂ 2 ε11 ∂ 2 ε33 ∂ 2 ε13
−S22 = R2 = 2
+ 2
−2 , (2)
∂X3 ∂X1 ∂X1 ∂X3
∂ 2 ε22 ∂ 2 ε33 ∂ 2 ε23
−S11 = R1 = 2
+ 2
−2 , (3)
∂X3 ∂X2 ∂X2 ∂X3
∂ 2 ε11

∂ ∂ε23 ∂ε13 ∂ε12
−S23 = U1 = − + − + + , (4)
∂X2 ∂X3 ∂X1 ∂X1 ∂X2 ∂X3
∂ 2 ε22

∂ ∂ε23 ∂ε13 ∂ε12
−S31 = U2 = − + − + , (5)
∂X1 ∂X3 ∂X2 ∂X1 ∂X2 ∂X3
∂ 2 ε33

∂ ∂ε23 ∂ε13 ∂ε12
−S12 = U3 = − + + − . (6)
∂X1 ∂X2 ∂X3 ∂X1 ∂X2 ∂X3
Show that
∂R1 ∂U3 ∂U2
+ + = 0,
∂X1 ∂X2 ∂X3
∂U3 ∂R2 ∂U1
+ + = 0, (7)
∂X1 ∂X2 ∂X3
∂U2 ∂U1 ∂R3
+ + = 0.
∂X1 ∂X2 ∂X3
These relations are known as the Bianchi formulas.
Solution: This is simple exercise of verifying the given result using the definitions Ri
and Ui . For example, consider the first equation to be verified:
∂ 3 ε22 ∂ 3 ε33 ∂ 2 ε23
2
+ 2
−2
∂X3 ∂X1 ∂X2 ∂X1 ∂X2 ∂X3 ∂X1
∂ 3 ε33 ∂2

∂ε23 ∂ε13 ∂ε12
− 2
+ + −
∂X1 ∂X2 ∂X3 ∂X2 ∂X1 ∂X2 ∂X3
3

∂ ε22 ∂ ∂ε23 ∂ε13 ∂ε12
− + − + ,
∂X1 ∂X32 ∂X2 ∂X3 ∂X1 ∂X2 ∂X3
which, because of the cancelation of terms, yields a zero.
3.42 Consider the following infinitesimal strain field:

ε11 = c1 X22 , ε22 = c1 X12 , 2ε12 = c2 X1 X2 ,
ε31 = ε32 = ε33 = 0,
where c1 and c2 are constants. Determine
(a) c1 and c2 such that there exists a continuous, single-valued displacement field that
corresponds to this strain field,
(b) the most general form of the corresponding displacement field using c1 and c2 obtained
in (a), and
(c) the constants of integration introduced in (b) for the boundary conditions u = 0 and
∂u1 ∂u2
Ω = 0 at X = 0 (that is, u1 = u2 = 0 and ∂X 2
− ∂X 1
= 0 at X1 = X2 = 0).
Solution: (a) Check for compatibility

∂ 2 ε11 ∂ 2 ε22 ∂ 2 ε12
2
+ 2
−2 = 2c1 + 2c1 − c2 = 0.
∂x2 ∂x1 ∂x1 ∂x2
68 SOLUTIONS MANUAL
Compatibility is met if c2 = 4c1 , an arbitrary constant.
(b) Using the strain-displacement relations

∂u1
= ε11 = cX22 → u1 = cX1 X22 + f1 (X2 ),
∂x1
∂u2
= ε22 = cX12 → u2 = cX12 X2 + f2 (X1 ),
∂x2
∂u1 ∂u2 df1 df2
+ = 2ε12 = 4cX1 X2 → 4cX1 X + + = 4cX1 X2 .
∂x2 ∂x1 dX2 dX1
This implies that df1 /dX2 = −df2 /dX1 and equal to a constant, K1 . This, the dis-
placement field is given by
u1 = cX1 X22 + K1 X2 + K2 , u2 = cX12 X2 − K1 X1 + K3 .
Clearly, K2 and K3 represent the rigid body translations in the X1 and X2 directions.
(c) The boundary condition u=0 at X = 0 gives K2 = K3 = 0. The boundary con-
∂u1 ∂u2
dition ∂X 2
− ∂X1
= 0 at X = 0 gives K1 = 0 so that the final displacement field
is
u1 = cX1 X22 , u2 = cX12 X2 .
3.43 Determine whether the following strain fields (under the assumption of infinitesimal
strains) are possible in a continuous body:
X3 (X12 + X22 ) 2X1 X2 X3 X3

 
(X12 + X22 ) X1 X2

2
(a) [ε] = , (b) [ε] =  2X1 X2 X3 X2 X1  .
X1 X2 X22
X3 X1 X32
Solution: (a) Check for compatibility
∂ 2 ε11 ∂ 2 ε22 ∂ 2 ε12

2
+ 2
−2 = 2 + 0 − 2(1) = 0.
∂X2 ∂X1 ∂X1 ∂X2
Compatibility is met; hence, it is a possible strain field.
(b) Check for the six compatibility conditions:
∂ 2 ε11 ∂ 2 ε22 ∂ 2 ε12

+ − 2 = 2X3 + 0 − 2(2X3 ) 6= 0,
∂X22 ∂X12 ∂X1 ∂X2
∂ 2 ε11 ∂ 2 ε33 ∂ 2 ε13
2
+ 2
−2 = 0 + 0 − 2(0) = 0
∂X3 ∂X1 ∂X1 ∂X3
∂ 2 ε22 ∂ 2 ε33 ∂ 2 ε23
2
+ 2
−2 = 0 + 0 − 2(0) = 0
∂X3 ∂X2 ∂X2 ∂X3
∂ 2 ε11 ∂ 2 ε23 ∂ 2 ε13 ∂ 2 ε12
+ 2
− − = 2X2 + 0 − 0 − 2X2 = 0
∂X2 ∂X3 ∂X1 ∂X1 ∂X2 ∂X1 ∂X3
∂ 2 ε22 ∂ 2 ε13 ∂ 2 ε23 ∂ 2 ε12
+ 2
− − =0+0−0−0=0
∂X1 ∂X3 ∂X2 ∂X1 ∂X2 ∂X2 ∂X3
∂ 2 ε33 ∂ 2 ε12 ∂ 2 ε13 ∂ 2 ε23
+ 2
− − =0+0−0−0=0
∂X1 ∂X2 ∂X3 ∂X2 ∂X3 ∂X1 ∂X3
The strain field is not compatible.
3.44 Evaluate the compatibility conditions ∇0 × (∇0 × E)T = 0 in cylindrical coordinates.
Solution: From Problem 2.50, the curl of a second-order tensor E is

1 ∂Ezr ∂Eθr 1 ∂Erθ ∂Ezθ
∇ × E = êr êr − − Ezθ + êθ êθ −

1 1 ∂Erz ∂Eθz 1 ∂Ezθ ∂Eθθ 1
+ êz êz Eθz − + + êr êθ − + Ezr

∂Err ∂Ezr 1 ∂Ezz ∂Eθz
+ êθ êr − + êr êz −
∂z ∂r r ∂θ ∂z

∂Eθr 1 ∂Err 1 1 ∂Erz ∂Ezz
+ êz êr − + Erθ + Eθr + êθ êz −

∂Eθθ 1 1 1 ∂Erθ
+ êz êθ + Eθθ − Err − .
∂r r r r ∂θ
Note that Green–Lagrange strain tensor components, by definition, are symmetric. The
transpose of ∇ × E is

1 ∂Ezr ∂Eθr 1 ∂Erθ ∂Ezθ
(∇ × E)T = êr êr − − Ezθ + êθ êθ −

1 1 ∂Erz ∂Eθz 1 ∂Ezθ ∂Eθθ 1
+ êz êz Eθz − + + êθ êr − + Ezr

∂Err ∂Ezr 1 ∂Ezz ∂Eθz
+ êr êθ − + êz êr −
∂z ∂r r ∂θ ∂z

∂Eθr 1 ∂Err 1 1 ∂Erz ∂Ezz
+ êr êz − + Erθ + Eθr + êz êθ −

∂Eθθ 1 1 1 ∂Erθ 1
+ êθ êz + Eθθ − Err − + Ezr
∂r r r r ∂θ r
≡ Frr êr êr + Fθθ êθ êθ + Fzz êz êz + Fθr êθ êr + Frθ êr êθ + Fzr êz êr + Frz êr êz
+ Fzθ êz êθ + Fθz êθ êz ,
where Frr , Fθθ , etc. are the components of a second-order tensor F (not symmetric).
Then taking ∇ × F (again using the result of Problem 2.50), we obtain

1 ∂Fzr ∂Fθr 1 ∂Frθ ∂Fzθ
∇ × F = êr êr − − Fzθ + êθ êθ −

1 1 ∂Frz ∂Fθz 1 ∂Fzθ ∂Fθθ 1
+ êz êz Fθz − + + êr êθ − + Fzr

∂Frr ∂Fzr 1 ∂Fzz ∂Fθz
+ êθ êr − + êr êz −
∂z ∂r r ∂θ ∂z

∂Fθr 1 ∂Frr 1 1 ∂Frz ∂Fzz
+ êz êr − + Frθ + Fθr + êθ êz −

∂Fθθ 1 1 1 ∂Frθ 1
+ êz êθ + Fθθ − Frr − + Fzr .
∂r r r r ∂θ r
Equating the above expression to zero gives nine relations corresponding to the re-
quirement ∇ × (∇ × E)T = 0. Out of nine only six compatibility equations will be
independent. For example, we have from the coefficient of êr êr the compatibility con-
dition
1 ∂Fzr ∂Fθr 1
− − Fzθ = 0,
r ∂θ ∂z r
which is equal to

1 ∂ 1 ∂Ezz ∂Eθz ∂ 1 ∂Ezθ ∂Eθθ 1 ∂Erz ∂Ezz
− − − − − = 0,
r ∂θ r ∂θ ∂z ∂z r ∂θ ∂z r ∂z ∂r
or
1 ∂ 2 Ezz 2 ∂ 2 Ezθ ∂ 2 Eθθ 1 ∂Erz 1 ∂Ezz
2 2
− + − + = 0.
r ∂θ r ∂θ∂z ∂z 2 r ∂z r ∂r
Similarly, from the coefficient of êr êθ we have the compatibility condition
1 ∂Fzθ ∂Fθθ 1
− + Fzr = 0,
r ∂θ ∂z r
70 SOLUTIONS MANUAL
which in terms of the strain components becomes

1 ∂ ∂Erz ∂Ezz ∂ ∂Erθ ∂Ezθ 1 1 ∂Ezz ∂Eθz
− − − + − = 0,
r ∂θ ∂z ∂r ∂z ∂z ∂r r r ∂θ ∂z
or
1 ∂ 2 Erz 1 ∂ 2 Ezz ∂ 2 Erθ ∂ 2 Ezθ 1 ∂Ezz 1 ∂Eθz
− − + + 2 − = 0.
r ∂θ∂z r ∂r∂θ ∂z 2 ∂r∂z r ∂θ r ∂z
From the coefficient of êθ êr we have the compatibility condition
∂Frr ∂Fzr
− = 0,
∂z ∂r
or
1 ∂ 2 Ezr ∂ 2 Eθr 1 ∂Ezθ 1 ∂Ezz 1 ∂ 2 Ezz ∂ 2 Eθz
− − + − + = 0,
r ∂θ∂z ∂z 2 r ∂z r2 ∂θ r ∂r∂θ ∂r∂z
which is the same (in view of the symmetry of E) as that obtained from the coefficient
of êr êθ .
3.45 Given the strain components
E11 = f (X2 , X3 ), E22 = E33 = −νf (X2 , X3 ), E12 = E13 = E23 = 0,
determine the form of f (X2 , X3 ) in order that the strain field is compatible.
Solution: Let us check the compatibility conditions

∂ 2 E11 ∂ 2 E22 ∂ 2 E12 ∂2f
2
+ 2
−2 =0 → = 0,
∂X2 ∂X1 ∂X1 ∂X2 ∂X22
∂ 2 E11 ∂ 2 E33 ∂ 2 E13 ∂2f
2
+ 2
−2 =0 → =0
∂X3 ∂X1 ∂X1 ∂X3 ∂X32
∂ 2 E22 ∂ 2 E33 ∂ 2 E23 ∂2f ∂2f

+ − 2 = −ν + =0
∂X32 ∂X22 ∂X2 ∂X3 ∂X32 ∂X22
∂ 2 E11 ∂ 2 E23 ∂ 2 E13 ∂ 2 E12 ∂2f
+ 2
− − =0 → =0
∂X2 ∂X3 ∂X1 ∂X1 ∂X2 ∂X1 ∂X3 ∂X2 ∂X3
∂ 2 E22 ∂ 2 E13 ∂ 2 E23 ∂ 2 E12
+ 2
− − =0
∂X1 ∂X3 ∂X2 ∂X1 ∂X2 ∂X2 ∂X3
∂ 2 E33 ∂ 2 E12 ∂ 2 E13 ∂ 2 E23
+ 2
− − =0
∂X1 ∂X2 ∂X3 ∂X2 ∂X3 ∂X1 ∂X3
Thus, we have
∂2f ∂2f ∂2f
2
= 2
= = 0,
∂X2 ∂X3 ∂X2 ∂X3
which imply f (X2 , X3 ) is a function of the form
f (X2 , X3 ) = A + BX2 + CX3 ,
where A, B, and C are arbitrary constants.
3.46 Given the strain tensor E = Err êr êr + Eθθ êθ êθ in an axisymmetric body (that is, Err
and Eθθ are functions of r and z only), determine the compatibility conditions on Err
and Eθθ . Hint: See Example 2.5.1.
Solution: Using the vector form of compatibility conditions, Eq. (3.7.11), we obtain
(see Example 2.5.1)

∂Eθθ Eθθ − Err ∂Err ∂Eθθ
F≡∇×E= + êz êθ + êθ êr − êr êθ ,
∂r r ∂z ∂z

∂ ∂Eθθ Eθθ − Err
∇ × (F)T = + (êr × êθ ) êz
∂r ∂r r
∂ 2 Eθθ

1 ∂Eθθ Eθθ − Err ∂êθ
− (êr × êθ ) êr + + êθ × êz
∂r∂z r ∂r r ∂θ

1 ∂Err ∂êr 1 ∂Err ∂êθ
+ êθ × êθ + (êθ × êr )
r ∂z ∂θ r ∂z ∂θ

1 ∂Eθθ ∂êθ ∂ ∂Eθθ Eθθ − Err
− êθ × êr + + (êz × êθ ) êz
r ∂z ∂θ ∂z ∂r r
∂ 2 Err ∂ 2 Eθθ ↔
+ 2
(êz × êr ) êθ − 2
(êz × êθ ) êr = 0 .
∂z ∂z
Noting that
∂êr ∂êθ
= êθ , = −êr , êr × êθ = êz , êθ × êz = êr , êz × êr = êθ ,
∂θ ∂θ
and that a tensor is zero only when all its components are zero, we obtain

∂ ∂Eθθ Eθθ − Err 1 ∂Eθθ Eθθ − Err
êz êz : + + + = 0.
∂r ∂r r r ∂r r
∂ 2 Eθθ 1 ∂
êz êr : − + (Err − Eθθ ) = 0.
∂r∂z r ∂z
∂ ∂Eθθ Eθθ − Err
êr êz : − + = 0.
∂z ∂r r
∂ 2 Err ∂ 2 Eθθ
êθ êθ : = 0, êr êr : − = 0.
∂z 2 ∂z 2
3.47 Determine the effect of the superposed rigid-body motion on the left Cauchy–Green
deformation tensor B = F · FT .
Solution: We have
B∗ = F∗ · (F∗ )T = (Q · F) · (FT · QT ) = Q · B · QT .
3.48 If Q(t) is an orthogonal tensor-valued function of a scalar t [that is, Q−1 = QT ], show
that Q̇ · QT = −(Q̇ · QT )T . That is, show that Q̇ · QT is skew symmetric.
Solution: We begin with the identity

Q · QT = I
and compute the time derivative
d
Q · QT = 0
dt
or
d d T
Q · QT = −Q · QT = − Q̇ · QT .
dt dt
3.49 Show that the spin tensor W under superposed rigid-body motion becomes
W∗ = Q · W · QT + Ω,
where Ω is the skew symmetric rotation tensor, Ω = Q̇ · QT [see also Eq. (3.8.19)].
Solution: We have the following result from Eqs. (3.6.3) and (3.8.9):
2 W∗ = L∗ − (L∗ )T = Q̇ · QT + Q · L · QT − Q · Q̇T − Q · LT · QT
= 2 Q · W · QT + Q̇ · QT − Q · Q̇T
= 2 Q · W · QT + 2 Q̇ · QT = 2 Q · W · QT + 2 Ω,
where we have used the fact that Q · Q̇T = −Q · QT .

72 SOLUTIONS MANUAL
3.50 Suppose that the second-order tensor T is objective in the sense that it satisfies the
condition T∗ = Q · T · Q, where quantities with and without an asterisk belong to two
different frames of reference. Then show that the following second-order tensor S is
objective (that is, show that S∗ = S):
S = F−1 · T · F−T ,
where Q is a proper orthogonal tensor.
Solution: We have
S∗ = J ∗ (F∗ )−1 · T∗ · (F∗ )−T = J (F−1 · RT ) · (R · T · RT ) · (R · F−T )
= J F−1 · T · F−T = S.
3.51 Prove or disprove if the following second tensor satisfies objectivity:
T = S · U + U · S,
∗ T T
where S = R · S · R , U = R · F is the right Cauchy stretch tensor, and F is the
deformation gradient.
Solution: We have (Q = R)
T∗ = S∗ · U∗ + U∗ · S∗
= S∗ · (RT · F∗ ) + (RT · F∗ ) · S∗
= S∗ · (RT · R) · F + (RT · R) · F · S∗
= S∗ · F + F · S∗ = R · S · RT · F + F · R · S · RT
= R · S · U + U · S · RT 6= S · U + U · S.
Hence, not objective.
3.52 Using the transformation rule F∗ = Q·F, show that the Euler stain tensor e transforms
according to the rule under superposed rigid body motion
e∗ = Q · e · QT .
Solution: We have
2e∗ = I − (F∗ )−T · (F∗ )−1
= I − (Q · F)−T · (Q · F)−1 = I − Q−T · F−T · F−1 · Q−1

= I − Q · F−T · F−1 · QT = Q · I − F−T · F−1 · QT = 2e
3.53 Show that the spatial gradient of a vector u(x, t) is objective, that is, prove
∇∗ u∗ (x∗ , t∗ ) = Q(t) · ∇u(x, t) · QT (t).
Solution: Because u∗ (x∗ , t∗ ) = Q(t) · u(x, t), we have
∇∗ u∗ (x∗ , t∗ ) = ∇∗ [Q(t) · u(x, t)]

∂ T ∂x ∂
= u(x, t) · Q (t) = · u(x, t) · QT (t)
∂x∗ ∂x∗ ∂x
= Q(t) · ∇u(x, t) · QT (t),
where
∂x∗ ∂
= [c(t) + Q(t) · x] = QT
∂x ∂x
and −1
∂x∗

∂x
= = Q−T (t) = Q(t).
∂x∗ ∂x
3.54 Show that the material time derivatives of objective vector and tensor fields, u and S,
are not objective.
Solution: Consider the material time derivative of an objective vector field u∗ = Q · u

u̇∗ = Q̇ · u + Q · u̇ 6= Qu̇.
Similarly, the material time derivative of an objective tensor field S∗ = Q · S · QT
Ṡ∗ = Q̇ · S · QT + Q · Ṡ · QT + Q · S · Q̇T 6= Q · Ṡ · QT .
3.55 Establish the uniqueness of the decomposition F = R · U = V · R. For example, if

F = R1 · U1 = R2 · U2 , then show that R1 = R2 and U1 = U2 .
Solution: Let
F = R 1 · U 1 = R 2 · U2 .
Then using the orthogonal property of R1 and R2 can write
(R1 · U1 )T · (R1 · U1 ) = (R2 · U2 )T · (R2 · U2 )
UT T T T
1 · (R1 · R1 ) · U1 = U2 · (R2 · R2 ) · U2
U1 · U1 = U2 · U2 .
Because U1 and U2 are positive-definite, it follows that U1 = U2 and then R1 = R2 .
Note that C = U · U.
3.56 Show that the eigenvalues of the left and right Cauchy stretch tensors U and V are the
same and that the eigenvector of V is given by R · n, where n is the eigenvector of U.
Solution: Let λ be an eigenvalue and n be the associate eigenvector of U. Then

U · n = λn → R · U · n = λR · n
Because U · U = V · R = F, we have
V · (R · n) = λ(R · n).
Thus, R · n is the eigenvector of V with eigenvalue λ.
3.57 (a) If λ is the eigenvalue and n is the eigenvector of U, show that the eigenvalue of C
is λ2 and the eigenvector is the same as that of U. (b) Show that a line element in the
principal direction n of C becomes an element in the direction of R · n in the deformed
configuration.
Solution: (a) Because U · n = λn, we have

U · U · n = λU · n = λ2 n or C · n = λ2 n.
Thus, n is the eigenvector of C with λ2 as its eigenvalue.
(b) Let dX = dSn be an element in the principal direction of C (and U). Then,
U · dX = dS(U · n) = λdSn
and
dx = F · dX = R · U · dX = λdS(R · n).
Thus, the deformed vector in is the direction of R · n.
74 SOLUTIONS MANUAL
3.58 Show that the spin tensor W can be written as

2W = 2Ṙ · RT + R · U̇ · U−1 − U−1 · U̇ · RT ,
where R is the proper orthogonal rotation tensor R−1 = RT and U is the symmetric
positive-definite right Cauchy stretch tensor. Also show that for rigid body motion,
one has W = Ṙ · RT .
Solution: First we note the following relations from Eqs. (3.6.14) and (3.8.1):
L = Ḟ · F−1 ; F = R · U; Ḟ = Ṙ · U + R · U̇.
Then we use the definition (3.6.3) to write

T
2W = L − LT = Ḟ · F−1 − Ḟ · F−1
h iT
= Ṙ · U + R · U̇ · U−1 · R−1 − Ṙ · U + R · U̇ · U−1 · R−1
h iT
= Ṙ · R−1 + R · U̇ · U−1 · RT − Ṙ · R−1 + R · U̇ · U−1 · RT
T
T −1 T T −1 T
= Ṙ · R + R · U̇ · U · R − R · Ṙ + R · U̇ · U ·R .
Because U is symmetric, and

d
R · RT = 0 ⇒ R · ṘT = −Ṙ · RT ,
dt
we obtain
2W = 2Ṙ · RT + R · U̇ · U−1 − U−1 · U̇ · RT .
For rigid body motion, one has F = R or U = I, and the above relation becomes
W = Ṙ · RT .
3.59 Prove the symmetry and positive-definiteness of the right Cauchy–Green deformation
tensor C = FT · F.
Solution: We must show that C is (a) symmetric and (b) positive-definite, that is,
AT · C · A ≥ 0 for any vector A, and AT · C · A = 0 only when A = 0. The symmetry
follows from
(CT = (FT · F)T = FT · F = C.
To show positive-definiteness, we use the polar decomposition theorem and write
AT · C · A = AT · (FT · F) · A = AT · (UT · RT · R · U) · A
= AT · (UT · U) · A ≥ 0.
Because UT · U > 0 the equality holds only when A = 0. When A is a nonzero vector,
we have
AT · (UT · U) · A = (U · A)T · (U · A) > 0.
√
3.60 Calculate C when  
3 2 0
[C] =  2 3 0  .
0 0 9
Solution: First we determine the eigenvalues and eigenvectors of [C]

3−λ 2 0

|[C] − λ[I]| = 2 3 − λ 0 = 0.
0 0 9−λ
We obtain
(9 − λ) (3 − λ)2 − 4 = 0 → λ1 = 1, λ2 = 5, λ3 = 9.


     
 1 1 0
1 1
{X̂}(1) = √ −1 , {X̂}(2) = √ 1 , {X̂}(3) = 0 .
2  0 2 0 1
 
The matrix of eigenvectors is

 
1 −1 0
1 
[Q] = √ 1 1 √0  .
2 0 0 2
Then we have 
1 0 0
T
[C̄] = [Q] [C][Q] =  0 5 0  .
0 0 9
Hence, we have  
p 1 √0 0
[ C̄] =  0 5 0.
0 0 3
Then
   
√ 1 1 0 1 √0 0 1 −1 0
T
p 1
[ C] = [Q] [ C̄][Q] = −1 1 √0   0 5 01 1 √0 
2
0 0 2 0 0 3 0 0 2
√ √ 
5 + 1 √5 − 1 0
1 √
=  5 − 1 5 + 1 0.
2
0 0 6
3.61 Given that

1 2 −5
[F ] = ,
5 11 2
determine the right and left stretch tensors.
Solution: The components of the right Cauchy–Green deformation tensor C are given
by
1 2 11 2 −5 1 125 12 5.0 0.40
[C] = [F ]T [F ] = = =
25 −5 2 11 2 25 12 29 0.4 1.16
The eigenvalues λ21 and λ22 of matrix [C] are determined by setting
|[C] − λ2 [I]| = 0 → λ21 = 5.0591, λ22 = 1.1009
so that λ1 = 2.2492 and λ2 = 1.0492. The eigenvectors are (in vector component form)

0.9925 −0.1222
{N (1) } = ± , {N (2) } = ± .
0.1222 0.9925
Hence, the right Cauchy stretch tensor can be written as
U = λ1 N̂(1) N̂(1) + λ2 N̂(2) N̂(2)

(1) (1) (1) (1)
= λ1 N1 ê1 + N2 ê2 N1 ê1 + N2 ê2

(2) (2) (2) (2)
+ λ2 N1 ê1 + N2 ê2 N1 ê1 + N2 ê2

(1) (2) (1) (2)
= λ1 [N1 ]2 + λ2 [N1 ]2 ê1 ê1 + λ1 [N2 ]2 + λ2 [N2 ]2 ê2 ê2
76 SOLUTIONS MANUAL

(1) (1) (2) (2)
+ λ1 N1 N2 + λ2 N1 N2 (ê1 ê2 + ê2 ê1 ) .
or in component matrix form

2.2313 0.1455
[U ] = .
0.1455 1.0671
Then the rotation tensor [R] in matrix form is given by

0.2425 −0.9702
[R] = [F ][U ]−1 = .
0.9702 0.2425
Finally, the left Cauchy stretch tensor components are computed in matrix form from

1.0671 0.1455
[V ] = [R][U ][R]T = .
0.1455 2.2313
3.62 Given that √


3 1 0
[F ] =  0 2 0  ,
0 0 1
determine the right and left stretch tensors.
Solution: We have  √ 
√3 3 0
[C] = [F ]T [F ] =  3 5 0.
0 0 1
The eigenvalues are
λ1 = 6, λ2 = 2, λ3 = 1,
and the matrix of eigenvectors is
 √ 
1 3 0
1 √
[Q] = − 3 1 0.
2
0 0 2
Then   √ 
6 0 0 6 √0 0
T
[C̄] = [Q][C][Q] =  0 2 0  , [Ū ] =  0 2 0
0 0 1 0 0 1
Hence we have
 √ √ 
3 + √3 3 − √3 0
1
[U ] = [Q]T [Ū ][Q] = √  3 − 3 1 + 3 3 √0 .
2 2 0 0 2 2
The inverse is given by
 √ √ 
1 + 3√3 −(3 −√ 3) 0
1 
[U ]−1 = √ −(3 − 3) 3 + 3 0 .
√
4 6 0 0 4 6
Then √ √ 
3+√ 1 √3 − 1 0
1 
[R] = [F ][U ]−1 = √ 1− 3 3+1 √0 .
2 2 0 0 2 2
Finally, we have
√ √ 
3 + 1 √3 − 1 0
T1 √
[V ] = [R][U ][R] = √ 3−1 3 + 1 √0  .
2 0 0 2
3.63 Calculate the left and right Cauchy stretch tensors U and V associated with F of
Problem 3.11 for the choice of A = 2 and B = 0.
Solution: We have
      
120 100 120 120
T
[F ] =  0 1 0  , [C] = [F ] [F ] =  2 1 0   0 1 0  =  2 5 0  .
001 001 001 001
The eigenvalues of C are determined from

1−λ 2 0
√
2 5 − λ 0 = 0 → λ1,2 = 3 ± 2 2, λ3 = 1

0 0 1−λ
√
λ1 = 3 + 2 2 : n̂ = 0.3827ê1 + 0.9238ê2
√
λ2 = 3 − 2 2 : n̂ = 0.9238ê1 − 0.3827ê2
λ3 = 1 : n̂ = ê3
Now we can express C and U with respect to the principal axes as
   
5.828 0 0 2.414 0 0
[C] =  0 0.172 0  , [U ] =  0 0.414 0  .
0 0 1.0 0 0 1.0
The inverse of [U ] with respect to the principal directions of C is

 
0.414 0 0
[U ]−1 =  0 2.414 0  .
0 0 1.0
The components of [U ] and its inverse with respect to the basis êi are
   
0.383 0.924 0 2.414 0 0 0.383 0.924 0
[U ] =  0.924 −0.383 0   0 0.414 0   0.924 −0.383 0 
0 0 1.0 0 0 1.0 0 0 1.0
 
0.707 0.707 0
=  0.707 2.121 0 
0 0 1.0
   
0.383 0.924 0 0.414 0 0 0.383 0.924 0
[U ]−1 =  0.924 −0.383 0   0 2.414 0   0.924 −0.383 0 
0 0 1.0 0 0 1.0 0 0 1.0
 
2.121 −0.707 0
=  −0.707 0.707 0  .
0 0 1.0
The components of the rotation tensor are
    
120 2.121 −0.707 0 0.707 0.707 0
−1
[R] = [F ][U ] =  0 1 0   −0.707 0.707 0  =  −0.707 0.707 0  .
001 0 0 1.0 0 0 1.0
Finally, the components of [V ] are computed from

    
120 0.707 −0.707 0 2.121 0.707 0
[V ] = [F ][R]T =  0 1 0   0.707 0.707 0  =  0.707 0.707 0  .
001 0 0 1.0 0 0 1.0
78 SOLUTIONS MANUAL
N3.1 Given the following mapping of the motion of a body,
χ(X, t) = x = 1 + At3 X1 ê1 + X2 ê2 + X3 ê3 ,

where A is a constant, determine

(a) the inverse mapping,
(b) the velocity and acceleration components in the spatial and material coordinates,
(c) the components of the deformation gradient, and
(d) the components of the displacement and velocity vectors.
Solution: (a) The inverse mapping is (Êi = êi )

−1
χ−1 (x, t) = X = 1 + At3 x1 Ê1 + x2 Ê2 + x3 Ê3 .
(b) The velocity vector in the material coordinates is

Dx ∂x
v(X, t) = = = 3At2 X1 Ê1 .
Dt ∂t
In the spatial coordinates it is equal to
3At2
v(x, t) = x1 ê1 .
1 + At3
The acceleration vector in the material coordinates is
Dv ∂v
a(X, t) = = = 6At X1 Ê1 .
Dt ∂t
In spatial coordinates, it is equal to
6At
a(x, t) = x1 ê1 .
1 + At3
Alternatively,
∂v
a(x, t) = + v · ∇v
∂t
2
9A2 t4 3At2

6At
= − x 1 ê1 + x1 ê1
1 + At3 (1 + At3 )2 1 + At3
6At
= x1 ê1 .
1 + At3
(c) The matrix of the deformation gradient is
(1 + At3 ) 0 0
 
[F ] =  0 1 0.
0 0 1
(d) The displacement vector is
At3
u = x − X = At3 X1 ê1 or u1 = At3 X1 = x1 , u2 = u3 = 0.
1 + At3
and the velocity vector is
Du ∂u 3At2
v= = = 3At2 X1 ê1 = x1 ê1 .
Dt ∂t 1 + At3
N3.2 The motion of a continuous body is characterized by the velocity field

1
3x1 ê1 + x2 ê2 + 5x23 ê3 .

v(x, t) =
1+t
Determine
(a) the deformation mapping,
(b) the velocity field in the material description, and
(c) the acceleration in the spatial and material descriptions.
Solution: (a) Integrating the velocity field with respect to time, we obtain
1
3x1 ê1 + x2 ê2 + 5x23 ê3 + c.

x=−
(1 + t)2
At t = 0, we must have x = X. Hence, we have
c = (X1 + 3x1 )ê1 + (X2 + x2 ) ê2 + (X3 + 5x23 )ê3 .
N3.3 The deformation mapping in two dimensions is given by
x(X) = (4 − 2X1 − X2 ) ê1 + (2 + 1.5X1 − 0.5X2 ) ê2 .
Determine
(a) the matrix form of the deformation gradient and its inverse, and
(b) sketch the deformation of a unit square, 0 ≤ X1 ≤ 1 and 0 ≤ X2 ≤ 1.
Solution: (a) The matrix of deformation gradient and its inverse are

−2.0 −1.0 −0.5 1.0
[F ] = , [F ]−1 = 2.5
1
.
1.5 −0.5 −1.5 −2.0
(b) A sketch of the reference square and deformed configuration of it are shown in
Fig. NP3.3. In particular, a vector A = 0.75 ê1 + 0.75 ê2 from the undeformed square
transforms to FigPN3-3
x(X) = (4 − 2 × 0.75 − 0.75) ê1 + (2 + 1.5 × 0.75 − 0.5 × 0.75) ê2 = 1.75 ê1 + 2.75 ê2 .
X 2 , x2
B
C χ (A)
°
(1.75,2.75)
(3,2)° A
b
(0.2,0.6) (0.75,0.75) D
D C
° A°
χ −1 (b )
X 1 , x1
A B
Fig. P3.23
80 SOLUTIONS MANUAL
To find the pre-image of a line from the deformed configuration in the undeformed
square, we need the inverse mapping, which can be determined from the given mapping

x1 −2.0 −1.0 X1 4
= + .
x2 1.5 −0.5 X2 2
as

X1 1 −0.5 1.0 x1 4 1 −0.5 1.0 x1 0
= − = + .
X2 2.5 −1.5 −2.0 x2 2 2.5 −1.5 −2.0 x2 4
For example, the vector b = −ê1 in the deformed body has the preimage in the
undeformed square as 0.2 ê1 + 0.6 ê2 , as shown in Fig. NP3.3.
N3.4 The deformation mapping is given by

x(X) = (X1 + BX1 X2 ) ê1 + X2 + BX22 ê2 + (X3 + BX1 X2 ) ê3 ,

where B is a constant. Determine

(a) the matrix form of the deformation gradient, and
(b) the displacement field.
Solution: (a) The matrix of deformation gradient and its inverse are
 
1 + BX2 BX1 0
[F ] =  0 1 + 2BX2 0  .
BX2 BX1 1
(b) The displacement field is
u(X)x − X = BX1 X2 ê1 + BX22 ê2 + BX1 X2 ê3 .
N3.5 Given the following displacement field

u1 = x1 − 0.25x2 , u2 = x1 + 2x2 , u3 = −3x3 ,
determine
(a) the matrix forms of the deformation gradient and its inverse, and if the deformation is
isochoric, and
(b) the Green-Lagrange strain tensor components.
Solution: (a) By definition, we have

∂uI
F−1 = I − (∇u)T or FIj
−1
= δIj − .
∂xj
Hence, the components of F−1 and F are
   
0.00 0.25 0.00 −4.00 −1.00 0.00
[F ]−1 =  −1.00 −1.00 0.00  , [F ] =  4.00 0.00 0.00  .
0.00 0.00 4.00 0.00 0.00 0.25
where the Jacobian is J = 1. Thus, it is an isochoric deformation.
(b) The Green–Lagrange strain tensor components are
    
−4.00 4.00 0.00 −4.00 −1.00 0.00 1.00 0.00 0.00
T
2[E] = [F ] [F ]−[I] =  −1.00 0.00 0.00   4.00 0.00 0.00 − 0.00 1.00 0.00  ,
0.00 0.00 0.25 0.00 0.00 0.25 0.00 0.00 1.00
or  
15.5 2.0 0.00000
[E] =  2.0 0.0 0.00000  .
0.0 0.0 −0.46875
N3.5 Consider a square block with a circular hole at the center, as shown in Fig. 3.3.9(a).
Suppose that block is of thickness h and plane dimensions 4b × 4b, and the radius of
the hole is b. Determine the change in the surface area of the cylinder formed by the
block and the inner edge of the hole when it is subjected to simple shear deformation
Figuremapping
3.3.8 of Eq. (3.3.14).
χ ( X) = ( X1 + γ X 2 )eˆ 1 + X 2 eˆ 2 + X 3 eˆ 3
χ (X )
X2 x2
h Nˆ = − (cos θ Eˆ + sin θ Eˆ )
1 2 4 bγ 4bγ
θ X1 n̂ x1
4b
ˆ = Eˆ
N 1 n̂
4b
(a) (b)
Fig. P3.24
Solution: The components of the deformation gradient and its inverse are
   
1 γ 0 1 −γ 0
−1
[F ] =  0 1 0  , [F ] =  0 1 0  .
0 0 1 0 0 1
The determinant of F is |F | = 1, implying that there is no change in the volume of the

block. Consider the edge with normal N̂ = Ê1 = ê1 in the reference configuration. By
Eq. (3.3.25), we have
n̂ da1 = (ê1 − γê2 ) dX2 dX3 .
Thus, da1 is p
da1 = (1 + γ 2 ) dX2 dX3 .
The total area of the deformed edge, as shown in Fig. 3.3.9(b), is
Z h Z 2b p
da1 = 4bh 1 + γ 2 .
0 −2b
Next, we determine the deformed area of the cylindrical surface of the hole. In this
case, the unit vector normal to the surface is in the radial direction and it is given by

N̂ = − cos θ ê1 + sin θ ê2 .
Hence, the components of the vector F−T · N̂ are given by

    
1 0 0  − cos θ   − cos θ 
 −γ 1 0  − sin θ = γ cos θ − sin θ .
0 0 1 0 0
   
Using Eq. (3.3.25), we obtain

n̂ dan = − cos θ ê1 + (γ cos θ − sin θ) ê2 b dθ dX3 .
Hence, the deformed surface area of the hole is

Z h Z 2π p
b cos2 θ + (γ cos θ − sin θ)2 dθ dX3 .
0 0
82 SOLUTIONS MANUAL
The integral can be evaluated for any given value of γ. In particular, we have
γ=0: an = 2πbh (no deformation),

√
Z 2π
γ=1: an = bh 1.5 + 0.5 cos 2θ − sin 2θ dθ ≈ 3πbh.
0
For other values of γ, the integral may be evaluated numerically.
N3.6 Given the mapping
x1 = X1 , x2 = X2 + κX3 , x3 = X3 + κX2 ,
where κ is a constant, determine the material and spatial descriptions of the displace-
ment field for this motion. Solution: The material descriptions of the motion (given)
and displacement field are
             
 x1  1 0 0  X1   u1   x1   X1  0 0 0  X1 
x2 =  0 1 κ  X2 , → u2 = x2 − X2 =  0 0 κ  X2
x3 0κ 1 X3 u3 x3 X3 0κ 0 X3
           
The spatial description of the motion is obtained by inverting the first matrix equation:
−1  
1 − κ2 0 0  x1 
     
 X1  1 0 0  x1 
1
X2 =  0 1 κ  x2 =  0 1 −κ  x2
1 − κ2
X3 0κ 1 x3 0 −κ 1 x3
     
The displacement field in the spatial description is given by

        
 u1   x1   X1  0 0 0  x1 
1  2
u2 = x2 − X2 = 0 −κ κ  x2
1−κ 2
u3 x3 X3 0 κ −κ2 x3
       
N3.7 For each of the displacement fields sketch, in the x1 x2 -plane, the deformed configuration
of the body which, in the X1 X2 -plane of the reference configuration, occupies the unit
square (0 ≤ X1 ≤ 1, 0 ≤ X2 ≤). In each case κ is a constant (presumably comparable
to 1 in magnitude).
(a) u = κX2 ê1 + κX1 ê2
(b) u = −κX2 ê1 + κX1 ê2
(c) u = κX12 ê2
Figure N3.7
Solution: Note that x = X + u. The sketches for the three cases are shown in Fig.
N3.7.
x2 x2 x2
κ κ κ κ
κ κ κ
κ x12
1 1 1
κ
κ κ x1
x1 x1
1 1 1
Fig. P3.25
83
Chapter 4: STRESS MEASURES
4.1 Suppose that tn̂1 and tn̂2 are stress vectors acting on planes with unit normals n̂1 and
n̂2 , respectively, and passing through a point with the stress state σ. Show that the
component of tn̂1 along n̂2 is equal to the component of tn̂2 along the normal n̂1 if and
only if σ is symmetric.
Solution: This is a trivial exercise. We have from Cauchy’s formula
Figure P4-2 n̂1 · σ = t(n̂1 ) , n̂2 · σ = t(n̂2 ) .
and
t(n̂1 ) · n̂2 = (n̂1 · σ) · n̂2 , t(n̂2 ) · n̂1 = (n̂2 · σ) · n̂1 .
If σ is symmetric, then both of the above expressions are the same. On the other hand,
if
n̂1 · σ · n̂2 = n̂2 · σ · n̂1 = n̂1 · σ T · n̂2 ,
then it follows that σ T = σ.
4.2 Write the stress vectors on each boundary surface in terms of the given values and base
vectors î and ĵ for the system shown in Figure P4.2.
3 kN/m2 2 kN/m2
H G D C
F E
5 kN/m2
2.5 kN/m2
A B
Fig. P4.2
Solution: We have
on BC: t = 5î, on CD: t = 0,
on DE: t = −2ĵ, on EF: t = 2ĵ,
on FG: t = 2ĵ, on EF: t = −3ĵ,
on HA: t = 0, on AB: unknown.
4.3 The components of a stress dyadic at a point, with respect to the (x1 , x2 , x3 ) system,
are (in MPa = 106 Pa= 106 N/m2 ):
     √ 
12 9 0 9 0 12 1 −3 √2
(i)  9 −12 0  , (ii)  0 −25 0  , (iii)  √ −3 √1 − 2 .

0 0 6 12 0 16 2 − 2 4
Find the following:

(a) The stress vector acting on a plane perpendicular to the vector 2ê1 − 2ê2 + ê3 .
(b) The magnitude of the stress vector and the angle between the stress vector and
the normal to the plane.
(c) The magnitudes of the normal and tangential components of the stress vector.
84 SOLUTIONS MANUAL
Solution: The unit normal vector is given by

1
n̂ = (2ê1 − 2ê2 + ê3 ) .
3
Solution for (i): (a) The three components of the stress vector are given by
       
 t1  12 9 0  2 1
1
t2 =  9 −12 0  −2 = 2 7 → tn̂ = 2(ê1 + 7ê2 + ê3 ).
3
t3 0 0 6 1 1
    
(b) The stress vector and the angle between the stress vector and the normal are
√ n̂ · tn̂

−1 −1 22
n̂
|t | = 204 = 14.28 MPa, θ = cos = cos − √ = 120.89◦ .
|tn̂ | 3 204
(c) The normal and tangential components of the stress vector are
r
22 484
q
n̂
σn = t · n̂ = − = −7.33 MPa, σs = |t | − σn = 204 −
n̂ 2 2 = 12.26 MPa.
3 9
Solution for (ii): (a) The three components of the stress vector are given by
       
 t1  9 0 12  2 3
1 10 10
t2 =  0 −25 0  −2 = 5 → tn̂ = (3ê1 + 5ê2 + 4ê3 ).
3 3   3
t3 12 0 16 1 4
  
10 √ n̂ · tn̂

|tn̂ | = 50 = 23.57 MPa, θ = cos−1 = cos−1 (0) = 90◦ .
3 |tn̂ |
q
σn = tn̂ · n̂ = 0 MPa, σs = |tn̂ |2 − σn2 = 23.57 MPa.
Solution for (iii): (a) The three components of the stress vector are given by
   √     √ 
 t1  1 −3 √2 1  2   8 + √2 
1
t2 =  √ −3 √1 − 2 3  −2  = 3  −8 − √2 

t3 2 − 2 1
 
4 4+4 2
or
tn̂ = 3.138 ê1 − 3.138 ê2 + 3.219 ê3 .
n̂ · tn̂

5.257
|tn̂ | = 5.482 MPa, θ = cos−1 = cos −1
= 16.47◦ .
|tn̂ | 5.482
q
σn = tn̂ · n̂ = 5.257 MPa, σs = |tn̂ |2 − σn2 = 1.554 MPa.
4.4 Consider a (kinematically infinitesimal) stress field whose matrix of scalar components
in the vector basis {êi } is  
1 0 2x2
 0 1 4x1  (MPa),
2x2 4x1 1
where the Cartesian coordinate variables Xi are in meters (m) and the units of stress
are MPa (106 Pa = 106 N/m2 ).
CHAPTER 4: STRESS MEASURES 85
(a) Determine the traction vector acting at point X = ê1 + 2ê2 + 3ê3 on the plane x1 +
x2 + x3 = 6.
(b) Determine the normal and projected shear tractions acting at this point on this plane.
Solution: (a) The unit normal vector is given by

1
n̂ = √ (ê1 + ê2 + ê3 ) .
3
The three components of the stress vector at point (x1 , x2 , x3 ) = (1, 2, 3) are given by
       
 t1  1 0 4 1 5
1 1 1
t2 =  0 1 4  √ 1 = √ 5 → tn̂ = √ (5ê1 + 5ê2 + 9ê3 ) MPa.
t3
 
4 4 1 3  
1 3  
9 3
The stress vector and the angle between the stress vector and the normal are
r
n̂ · tn̂

131 19
|tn̂ | = = 6.6081 MPa, θ = cos−1 = cos −1
√ = 16.58◦ .
3 |tn̂ | 393
(b) The normal and shear components of the stress vector are
r
19 131 361
q
σn = tn̂ · n̂ = = 6.3333 MPa, σs = |tn̂ |2 − σn2 = − = 1.8856 MPa.
3 3 9
4.5 The three-dimensional state of stress at a point (1, 1, −2) within a body relative to the
coordinate system (x1 , x2 , x3 ) is
 
2.0 3.5 2.5
 3.5 0.0 −1.5  MPa(= 106 Pa = 106 N/m2 ).
2.5 −1.5 1.0
Determine the normal and shear stresses at the point and on the surface of an internal
sphere whose equation is x21 + (x2 − 2)2 + x23 = 6.
Solution: The unit normal vector is given by [φ = x21 + (x2 − 2)2 + x23 ]
∇φ x1 ê1 + (x2 − 2)ê2 + x3 ê3

n̂(x) = = p .
|∇φ| x21 + (x2 − 2)2 + x23
At point x = ê1 + ê2 − 2ê3 the unit normal is

1
n̂ = √ (ê1 − ê2 − 2ê3 ) .
6
The three components of the stress vector at point (x1 , x2 , x3 ) = (1, 1, −2) are given by
       
 t1  2.0 3.5 2.5  1  −6.5 
6 1 1
t2 = 10 3.5 0.0 −1.5  √ −1 = √ 6.5 MPa,
 
t3 2.5 −1.5 1.0 6  −2  6  2.0 
or
1
tn̂ = √ (−6.5ê1 + 6.5ê2 + 4ê3 ) MPa.
6
The normal and shear components of the stress vector are
r
17 201 289
q
σn = tn̂ · n̂ = − = −2.833 MPa, σs = |tn̂ |2 − σn2 = − = 8.67 MPa.
6 12 36
86 SOLUTIONS MANUAL
4.6 The components of a stress dyadic at a point, with respect to the (x1 , x2 , x3 ) system,
are  
25 0 0
 0 −30 −60  MPa
0 −60 5
Determine (a) the stress vector acting on a plane perpendicular to the vector 2ê1 +
ê2 + 2ê3 , and (b) the magnitude of the normal and tangential components of the stress
vector. Solution: (a) The unit normal is given by n̂ = (2ê1 + ê2 + 2ê3 )/3. Hence, the
stress vector components are given by
      
 t1  25 0 0 2  50 
1 1
t2 = 0 −30 −60  1 = −150 MPa.
3 3
t3 0 −60 5 2 −50
    
Thus, t = (50ê1 − 150ê2 − 50ê3 )/3.
(b) The normal and shear components of the stress vector are
150
tn = t · n̂ = (100 − 150 − 100)/9 = − = −16.67 MPa,
9
r
p 2500 + 22500 + 2500 22500
ts = |t|2 − t2n = − = 52.7 MPa.
9 81
4.7 For the state of stress given in Problem 4.5, determine the normal and shear stresses
on a plane intersecting the point where the plane is defined by the points (0, 0, 0),
(2, −1, 3), and (−2, 0, 1).
Solution: The vectors connecting point (0, 0, 0) to points (2, −1, 3) and (−2, 0, 1) are
A = −2ê1 + ê3 , B = 2ê1 − ê2 + 3ê3 .
The unit vector normal to the plane of these vectors is given by

A×B 1
n̂ = = √ (ê1 + 8ê2 + 2ê3 ) .
|A × B| 69
The three components of the stress vector at the point are given by
       
 t1  2.0 3.5 2.5 1  35 
1 1
t2 = 106  3.5 0.0 −1.5  √ 8 = √ 0.5 MPa,
 
t3 2.5 −1.5 1.0 69  2  69  −7.5 
or
1
tn̂ = √ (35ê1 + 0.5ê2 − 7.5ê3 ) MPa.
69
The normal and shear components of the stress vector are
r
24 2563 576
q
σn = tn̂ · n̂ = = 0.3478 MPa, σs = |tn̂ |2 − σn2 = − = 4.2955 MPa.
69 138 4761
4.8 The Cauchy stress tensor components at a point P in the deformed body with respect
to the coordinate system (x1 , x2 , x3 ) are given by
 
1 4 −2
[σ] =  4 0 0  MPa.
−2 0 3
(a) Determine the Cauchy stress vector tn̂ at the point P on a plane passing through the
point and parallel to the plane 2x1 + 3x2 + x3 = 4.
(b) Find the length of tn̂ and the angle between tn̂ and the vector normal to the plane.
(c) Determine the components of the Cauchy stress tensor in a rectangular coordinate sys-
î are given in terms of the base vectors
tem (x̄1 , x̄2 , x̄3 ) whose orthonormal base vectors ē
êi of the coordinate system (x1 , x2 , x3 )
ˆ2 = 1 1
ē √
2
(ê1 − ê3 ) , ˆ
ē3 = 3
(2ê1 − ê2 + 2ê3 ) .
Solution: (a) The unit normal to the plane 2x1 + 3x2 + x3 = 4 is given by
n̂ = √1 (2ê1 + 3ê2 + ê3 ).

14
The components of the Cauchy stress vector are

      
 t1  1 4 −2  2   12 
t2 = √114  4 0 0 3 = √1
14
8 MPa,
t3 −2 0 3 1 −1
     
or
t(n̂) = √1 (12 ê1 + 8 ê2 − ê3 ) MPa.
14
(b) The length of the vector is

r
209
|t|2 = 1
14
(144 + 64 + 1), |t| = = 3.86 MPa.
14
The component of t normal to the plane at the point is
47
tnn = t(n̂) · n̂ = = 3.357 MPa,
14
The angle between the stress vector t and the normal is
t(n̂) · n̂ 3.357
cos θ = = → θ = 29.58◦ .
|t| 3.86
ˆ1 from
(c) First we determine ē
ˆ1 = ē
ē ˆ3 = − √1 (ê1 + 4ê2 + ê3 ) .
ˆ2 × ē
18
Hence the matrix of direction cosines is
 
−1 −4 −1
√1
[L] = √3 √0 −3
.
√

18
2 2− 22 2
Then we use the transformation equations
   √ 
−1 −4 −1 1 4 −2 −1 3 2 √2
T 1 
[σ̄] = [L][σ][L] = √ 3 √0 −3 √
 4 0 0   −4 0 −√ 2 
18
2 2− 22 2 −2 0 3 −1 −3 2 2
 √ 
32 −42 −28√2
1 
= −42√ √ −24 2 .
72 
18
−28 2 −24 2 −32
4.9 The Cauchy stress tensor components at a point P in the deformed body with respect
 
2 5 3
[σ] =  5 1 4  MPa.
3 4 3
(a) Determine the Cauchy stress vector t(n̂) at the point P on a plane passing through the
point whose normal is n = 3ê1 + ê2 − 2ê3 .
88 SOLUTIONS MANUAL
(b) Find the length of t(n̂) and the angle between t(n̂) and the vector normal to the plane.
(c) Find the normal and shear components of tn̂ on the plane.
Solution: The unit normal vector is
n̂ = √1 (3ê1 + ê2 − 2ê3 ) .

14

      
 t1  2 5 3  3 5
t2 = √114  5 1 4 1 = √1
14
8 MPa,
t3 3 4 3 −2 7
     
or
t(n̂) = √1 (5 ê1 + 8 ê2 + 7 ê3 ) MPa.
14
(b) The length of the vector is

r
2 1 138
|t| = 14
(25 + 64 + 49), |t| = = 3.1396 MPa.
14
The angle between the stress vector t and the normal is
t(n̂) · n̂ 3.357
cos θ = = → θ = 78.185◦ .
|t| 3.86
(c) The component of t normal to the plane at the point is

15 + 8 − 14
tnn = t(n̂) · n̂ = = 0.6429 MPa.
14
The shear stress is r
138 81
tns = − = 3.073 MPa.
14 196
4.10 Suppose that at√a point on the surface of a body the unit outward normal is n̂ =
(ê1 + ê2 − ê3 )/ 3 and the traction vector is P (ê1 + 2ê2 ), where P is a constant.
Determine (a) the normal traction vector tn and the shear traction vector tns at this
point on the surface of the body, and (b) the conditions between the stress tensor
components and the traction vector components.
Solution: (a) The normal and shear traction vectors are given by
tnn = (t · n̂)n̂ = √3 P n̂ = P (ê1 + ê2 − ê3 ),

3
tns = (t − tn = P (ê2 + ê3 ).
(b) The stress tensor at the point is related to the stress vector by Cauchy’s formula
t = n̂ · σ,
or
√1 (σ11 + σ21 − σ31 ) = P,
3
√1 (σ12 + σ22 − σ32 ) = P,
3
√1 (σ13 + σ23 − σ33 ) = 0.
3
4.11 Determine the traction free planes (defined by their unit normal vectors) passing
through a point in the body where the stress state with respect to the rectangular
Cartesian basis is  
1 2 1
[σ] =  2 σ0 0  MPa.
1 0 −3
What is the value of σ0 ?
Solution: By Cauchy’s formula we have
n1 + 2n2 + n3 = 0,
2n1 + σ0 n2 = 0,
n1 − 3n3 = 0.
Solving the equations, we obtain
n1 = 3n3 , n2 = −2n3 , n3 (σ0 − 3) = 0,
which gives σ0 = 3, because n3 cannot be zero. Using the equation n21 + n22 + n23 = 1,
we obtain
n3 = ± √114 , n1 = 3n3 , n2 = −2n3 .
Thus, the two traction free planes are defined by their normals
n̂ = ± √114 (3ê1 − 2ê2 + ê3 ) .
4.12 Use equilibrium of forces to derive the relations between the normal and shear stresses
σn and σs on a plane whose normal is n̂ = cos θê1 + sin θê2 to the stress components
σ11 , σ22 , and σ12 = σ21 on the ê1 and ê2 planes, as shown in Fig. P4.12:
σn = σ11 cos2 θ + σ22 sin2 θ + σ12 sin 2θ,

(1)
σs = − 12 (σ11 − σ22 ) sin 2θ + σ12 cos 2θ.
Note that θ is the angle measured from the positive x1 -axis to the normal to the inclined
plane (the same as that shown in Fig. 4.3.2). Then show that (a) the principal stresses
at a point in a body with two-dimensional state of stress are given by
r
σ11 + σ22 σ − σ 2
11 22 2
σp1 = σmax = + + σ12 ,
2 2
r (2)
σ11 + σ22 σ − σ 2
11 22 2
σp2 = σmin = − + σ12 ,
2 2
and that the orientation of the principal planes is given by
Figure P4-10
1 2σ12
θp = ± tan−1 , (3)
2 σ11 − σ22
and (b) the maximum shear stress is given by

σp1 − σp2
(σs )max = ± . (4)
2
Also, determine the plane on which the maximum shear stress occurs.
x2
t (nˆ ) σ = t( nˆ ) ⋅ nˆ
n
ê2 σs
n̂
θ
σ 11
σ 21 ê1
x1
σ 12
σ 22
Fig. P4.12
90 SOLUTIONS MANUAL
Solution: Identify the forces associated with the stresses on various planes (assume a
thickness of the triangle to be t and its diagonal length to be dL). The base is of length
dL sin θ and height is of length dL cos θ. Summing the forces along the normal to the
inclined plane, we obtain
σn tdL − (σ11 tdL cos θ) cos θ − (σ12 tdL cos θ) sin θ − (σ22 tdL sin θ) sin θ
− (σ12 tdL sin θ) cos θ = 0.
Dividing throughout by t dL, we obtain
σn = σ11 cos2 θ + 2σ12 cos θ sin θ + σ22 sin2 θ.
Similarly, summing the forces along the tangent to the inclined plane, we obtain
σs tdL + (σ11 tdL cos θ) sin θ − (σ12 tdL cos θ) cos θ − (σ22 tdL sin θ) cos θ
+ (σ12 tdL sin θ) sin θ = 0.
Dividing throughout by t dL, we obtain
σs = (σ22 − σ11 ) cos θ sin θ + σ12 cos2 θ − sin2 θ .

Express the equations in terms of the double angle

σ11 + σ22 σ11 − σ22
σn = + cos 2θ + σ12 sin 2θ (1),
2 2
σ11 − σ22
σs = − sin 2θ + σ12 cos 2θ. (2)
2
Then set the derivative of σn with respect to θ to zero (necessary condition for a
maximum of σn ) and determine the angle θp for which the normal stress is maximum
or minimum
2σ12
tan 2θp = ± . (3)
σ11 − σ22
There are two values of θp , differing by 90◦ , corresponding to the two principal stresses.
Substituting for cos 2θp and sin 2θp from
   
σ11 −σ22
2 σ12
cos 2θp = ±  q 2 , sin 2θp = ± q
   (4)
(σ12 )2 + σ11 −σ 2 + σ11 −σ22 2
22

2
(σ 12 ) 2
into Eq. (1), we obtain the following principal stresses:

r
σ11 + σ22 σ − σ 2
11 22
σp1 = + (σ12 )2 + , (5)
2 2
r
σ11 + σ22 σ − σ 2
11 22
σp2 = − (σ12 )2 + . (6)
2 2
Next, in order to derive the maximum shear stress, set the derivative of σs with respect
to θ to zero and determine the angle θs for which the shear stress is maximum or
minimum
σ11 − σ22
tan 2θs = ± . (7)
2σ12
Again, there are two values of θs , differing by 90◦ , corresponding to the two shear
stresses, which only differ in sign and not in the magnitude.
Substituting for cos 2θs and sin 2θs from
   
σ11 −σ22
2 σ12
tan 2θs = ±  q 2 , cos 2θs = ± q
   (8)
(σ12 )2 + σ11 −σ 2 + σ11 −σ22 2
22

2
(σ12 ) 2
into Eq. (2), we obtain the following maximum and minimum shear stresses:
r
σ − σ 2
11 22
σs1 = (σ12 )2 + , (9)
2
r
σ − σ 2
11 22
σs2 = − (σ12 )2 + . (10)
2
Figure P4-8
4.13 Determine the normal and shear stress components on the plane indicated in Figure
P4.13.
10 MPa
θ
θ = 30° 50 MPa
Fig. P4.13
Solution: Using the results of Problem 4.8, with θ = α = 30◦ , σ11 = 0, σ22 = 10 MPa,
σ12 = −50 MPa, we obtain
σn = σ11 cos2 θ + 2σ12 cos θ sin θ + σ22 sin2 θ

√
3 1
= 2(−50) + (10) = −40.80 MPa,
4 4
σs = (σ22 − σ11 ) cos θ sin θ + σ12 cos2 θ − sin2 θ

√
3 3 1
= (10) − (50) − = −20.67 MPa.
4 4 4
Figure P4-9
P4.14.
10 MPa
10 MPa
20 MPa
θ = 60°
Fig. P4.14
Solution: Using the results of Problem 4.12, with θ = 90 − α = 30◦ , σ11 = 20 MPa,
92 SOLUTIONS MANUAL
σ22 = −10 MPa, σ12 = −10 MPa, we obtain

√
3 3 1
= (20) + 2(−10) + (−10) = 3.84 MPa,
4 4 4

√
3 3 1
= (−10 − 20) + (−10) − = −17.99 MPa.
4 4 4
Figure P4-15
P4.15.
30 MPa
50 MPa
α = 45
60 MPa
Fig. P4.15
Solution: Using the results of Problem 4.12, with θ = 90 − α = 45◦ , σ11 = 60 MPa,
σ22 = 30 MPa, σ12 = 50 MPa, we obtain

1 1 1
= (60) + 2(50) + (30) = 95 MPa,
2 2 2

1 1 1
= (30 − 60) + (50) − = −15 MPa.
2 2 2
Figure P4-16
P4.16.
100 MPa
α
40 MPa
α = 60
Fig. P4.16
Solution: Using the results of Problem 4.12, with θ = α = 60◦ , σ11 = 40 MPa, σ22 = 0
MPa, σ12 = −100 MPa, we obtain
√
1 3
= (40) + 2(−100) + 0 = −76.60 MPa,
4 4

√
3 1 3
= (0 − 40) + (−100) − = 32.68 MPa.
Figure P4-17a 4 4 4
4.17 Find the values of σs and σ22 for the state of stress shown in Fig. P4.17.
30 MPa
30 MPa σs
σ0
α = 45
40 MPa
σs

α = 45
20 MPa
σ0 20 MPa
40 MPa
Fig. P4.17
Solution: Rotate the element 90◦ clockwise to see the correspondence between the
wedge here and that in Problem 4.12. Using the transformation equations of Problem
4.12, with θ = α = 45◦ , σ11 = σ0 MPa, σ22 = −40 MPa, σ12 = 20 MPa, and σn = 30
MPa, we obtain
σn = σ11 cos2 θ + σ22 sin2 θ + σ12 sin 2θ
30 = 12 σ0 + (−40) 12 + (−20)
or
σ0 = 60 + 40 + 40 = 140 MPa.
Then we have
1
σs = 2
(σ22 − σ0 ) sin 2θ + σ12 cos 2θ
1
= 2
(−40 − 140) = −90 MPa.
4.18 For the stress state given in Problem 4.4, determine
(a) the principal stresses and principal directions of stress at this point, and
(b) the maximum shear stress at the point.
Solution: (a) Setting |[σ] − λ[I]| = 0, we obtain

(1 − λ)[(1 − λ)(1 − λ) − 16] − 16(1 − λ) = 0 → [λ2 − 2λ − 31](1 − λ) = 0.
The principal stresses are
σ1 = λ1 = 6.6568 MPa, σ2 = λ2 = 1 MPa, σ3 = λ3 = −4.6568 MPa.
The plane associated with the maximum principal stress σ1 can be calculated from
    
1 − 6.6568 0 4  A1   0 
 0 1 − 6.6568 4  A2 = 0 ,
4 4 1 − 6.6568 A3 0
   
94 SOLUTIONS MANUAL
which gives
1
−5.6568A1 +4A3 = 0, −5.6568A2 +4A3 = 0, 4A1 +4A2 −5.6568A3 = 0 → A1 = A2 = √ A3 ,
2
or the plane is given by the vector (unit vector is also given)
1 √ 1 √
A(1) = ± √ ê1 + ê2 + 2ê3 ; n(1) = ± ê1 + ê2 + 2ê3 .
2 2
The plane associated with the principal stress σ2 = 1 psi is calculated from
    
1−1 0 4  A1   0 
 0 1 − 1 4  A2 = 0 ,
4 4 1−1 A3 0
   
which gives
4A3 = 0, 4A3 = 0, 4A1 + 4A2 = 0 → A1 = −A2 , A3 = 0,
or the plane is given by the vector

1
A(2) = ± (ê1 − ê2 ) ; n(2) = ± √ (ê1 − ê2 ) .
2
The third plane, being perpendicular to both A(1) and A(2) is given by
1 √ 1 √
A(3) = ± √ ê1 + ê2 − 2ê3 ; n(3) = ± ê1 + ê2 − 2ê3 .
2 2
4.19 Find the maximum and minimum normal stresses and the orientations of the principal
planes for the state of stress shown in Fig. P4.15.
Solution: From Problem 4.12 we have (with σ11 = 60 MPa, σ22 = 30 MPa, and
σ12 = 50 MPa) r
σ11 + σ22 σ − σ 2
11 22
σp1 = + (σ12 )2 +
2 2
√
= 45 + 2500 + 225 = 97.2 MPa
r
σ11 + σ22 σ − σ 2
11 22
σp2 = − (σ12 )2 +
2 2
√
= 45 − 2500 + 225 = −7.2 MPa
The principal plane is given by

1 2σ12 1 1000
θp = ± tan−1 = ± tan−1
2 σ11 − σ22 2 300
or θp1 = 36.65◦ and θp2 = 90 + 36.65 = 126.65◦ .
4.20 Find the maximum and minimum normal stresses and the orientations of the principal
planes for the state of stress shown in Fig. P4.16.
Solution: From Problem 4.12 we have (with σ11 = 40 MPa, σ22 = 0 MPa, and σ12 =
−100 MPa) r
σ11 + σ22 σ − σ 2
11 22
σp1 = + (σ12 )2 +
2 2
√
= 20 + 10000 + 400 = 121.98 MPa
r
σ11 + σ22 σ − σ 2
11 22
σp2 = − (σ12 )2 +
2 2
√
= 20 − 10000 + 400 = −81.98 MPa
The principal plane is given by

1 2σ12 1 200
θp = ± tan−1 = ± tan−1
2 σ11 − σ22 2 40
or θp1 = 39.35◦ and θp2 = 90 + 39.35 = 129.35◦ .
4.21 Find the maximum principal stress, maximum shear stress and their orientations for
the state of stress given.
   
12 9 0 3 5 8
(a) [σ] =  9 −12 0  MPa. (b) [σ] =  5 1 0  MPa.
0 0 6 8 0 2
Solution: (a) Setting |[σ] − λ[I]| = 0, we obtain
(6 − λ)[(12 − λ)(−12 − λ) − 81] = 0 → [−(144 − λ2 ) − 81](6 − λ) = 0.
Clearly, λ1 = 6 is an eigenvalue of the matrix. The remaining two eigenvalues are

obtained from λ2 − 225 = 0 → λ2 = 15 and λ3 = −15; thus the principal values of
stress are
σ1 = −15 MPa, σ2 = 6 MPa, σ3 = 15 MPa.
To find the principal plane associated with σ1 = −15 MPa, we write
(12 + 15)n1 + 9n2 = 0, 9n1 + (−12 + 15)n2 = 0, 21n3 = 0

n2 = −3n1 , n3 = 0 ⇒ n(1) = ±(1, −3, 0).
The principal plane associated with σ2 = 6 is n(2) = ±(0, 0, 1). To find the principal
plane associated with σ1 = 15 MPa, we write
(12 − 15)n1 + 9n2 = 0, 9n1 + (−12 − 15)n2 = 0, −9n3 = 0

n1 = 3n2 , n3 = 0 ⇒ n(3) = ±(3, 1, 0).
(b) The problem can be solved using the eigenvalue procedure. The principal stresses
are computed as follows:

3−λ 5 8
5 1 − λ 0 = 0 → −λ3 + 6λ2 + 78λ − 108 = 0

8 0 2−λ
If we use the alternative procedure of Section 2.5.5 (see Example 2.5.2), we obtain
 
1 58
[σ ] =  5 −1 0  ; I20 = −90, I30 = 64.
0
8 00
Then, we have
( 3/2 )
1 −1 64 3 1
α1 = cos = (78.770) = 26.2567◦
2 2 90 3
α2 = 146.2567◦ , α3 = −93.743◦
1/2
90
λ1 = 2 cos α1 = 9.824 → λ1 = 11.824 (MPa)
3
λ2 = 2 cos α2 (5.477) = −9.109 → λ2 = −7.109 (MPa)
λ3 = 2 cos α3 (5.477) = −0.715 → λ3 = 1.285 (MPa).
96 SOLUTIONS MANUAL
The maximum stress is σ1 = λ1 = 11.824 MPa. The direction (that is, the plane of
maximum stress) is given by
(3 − 11.824)n1 + 5n2 + 8n3 = 0

5n1 + (1 − 11.824)n2 = 0 → n2 = 0.462n1
8n1 + (2 − 11.824)n3 = 0 → n3 = 0.814n1
(1)
n = ±(1, 0.462, 0.814).
The unit normal is given by n̂(1) = ±(0.7300, 0.3372, 0.5945). The other two unit vectors
are
n̂(2) = ±(−0.6817, 0.4204, 0.5987), n̂(3) = ±(0.048, 0.842, −0.537).
4.22 (Spherical and deviatoric stress tensors) Let σ̃ denote the mean normal stress
1 1
σ̃ = tr σ = I1 .
3 3
Then the stress tensor can be expressed as the sum of spherical or hydrostatic stress
tensor and deviatoric stress tensor
0
σ = σ̃I + σ .
Thus, the deviatoric stress tensor is defined by

0 1
σ =σ− I1 I
3
For the state of stress given in Problem 4.18, compute the spherical and deviatoric
components of the stress tensor.
3+1+2
Solution: We have σ̃ = 3
= 2. Hence, the hydrostatic and deviatori stress tensors
are    
σ̃ 0 0 1 5 8
0
[σ̃] =  0 σ̃ 0  , [σ ] =  5 −1 0  .
0 0 σ̃ 8 0 0
0
4.23 Determine the invariants Ii and the principal deviator stresses for the following state
of stress (units are msi = 106 psi)
 
2 −1 1
[σ] =  −1 0 1  .
1 1 2
Solution: We have σ̃ = 2+0+2

3
= 4/3. Hence, the hydrostatic and deviatoric stress
tensors are
2 − 43 −1
   
1 0 0 1
4 0 4
[σ̃] =  0 1 0  , [σ ] =  −1 0 − 3 1  .
3
0 0 1 1 1 2 − 43
The principal invariants of σ 0 are [see Eq. (4.2.5)]
I10 = σii
0
= 0,

1 0 0 1 22 44 22 13
I20 = σij σij = + 2(−1)(−1) + 2(1)(1) + + 2(1)(1) + = ,
2 2 33 33 33 3
1 0 0 0 2
I30 = σij σjk σki = sum of 27 expressions = − , det(σ 0 ) = −2.593.
3 27
The principal deviatoric stresses are obtained by solving the eigenvalue problem. Set
2 − 43 − λ0
 
−1 1
0 0 4 0
0 = [σ − λ I] =  −1 0− 3 −λ 1 .
1 1 2 − 43 − λ0
We obtain

2 0 4 0 2 0 2 0 4 0
−λ − +λ −λ −1 − − λ + 1 + −1 + +λ ,
3 3 3 3 3
which simplifies to

2 4 2
− λ0 − + λ0 − λ0 − 3 = 0.
3 3 3
The principal deviatoric stresses are
2 1 5 1 7
λ01 = , λ02 = − + 2 = , λ03 = − − 2 = − .
3 3 3 3 3
The associated principal planes are
n̂(1) = −0.577ê1 + 0.577ê2 + 0.577ê3 , n̂(2) = 0.707ê1 + 0.707ê3 ,
n̂(3) = −0.4082ê1 − 0.8165ê2 + 0.4082ê3 .
4.24 Given the following state of stress at a point in a continuum,

 
7 0 14
[σ] =  0 8 0  MPa,
14 0 −4
determine the principal stresses and principal directions.
Solution: From inspection, we know that λ = 8 is an eigenvalue with ê2 as the eigen-
vector. The characteristic equation is
−λ3 + 11λ2 + 135.5λ − 1792 = 0 → (8 − λ)[−λ2 + 3λ − 168] = 0.
The principal stresses are
λ1 = −13.5416 MPa, λ2 = 8.0 MPa, λ3 = 16.5416 MPa.
The principal directions (normalized) associated with λi are
n̂(1) = −0.5632ê1 + 0.8263ê3 , n̂(2) = ê2 , n̂(3) = 0.8263ê1 + 0.5632ê3 .
The eigenvectors are mutually orthogonal.
4.25 Given the following state of stress (σij = σji ),
σ11 = −2x21 , σ12 = −7 + 4x1 x2 + x3 , σ13 = 1 + x1 − 3x2 ,

σ22 = 3x21 − 2x22 + 5x3 , σ23 = 0, σ33 = −5 + x1 + 3x2 + 3x3 ,
determine (a) the stress vector at point (x1 , x2 , x3 ) on the plane x1 +x2 +x3 = constant,
(b) the normal and shearing components of the stress vector at point (1, 1, 3), and (c)
the principal stresses and their orientation at point (1,2,1).
Solution: (a) The stress vector at point (x1 , x2 , x3 ) on the plane x1 + x2 + x3 =constant
can be computed using the equation
   
 t1   n1 
1
t2 = [σ] n2 ; n̂ = √ (ê1 + ê2 + ê3 ),
t
 
n
  3
3 3
where
√1 −6 + x1 − 3x2 + x3 + 4x1 x2 − 2x21 ,

t1 = 3
√1 5x3 + 3x21 − 2x22 ,

t2 = 3
t3 = √1 (−4 + 2x1 + 3x3 ) .

3
98 SOLUTIONS MANUAL
(b) The stress tensor at point (1, 1, 3) is given by

 
−2 (−7 + 4 + 3) (1 + 1 − 3)
[σ](1,1,3) =  (3 − 2 − 5 × 3) 0 
symm. (−5 + 1 + 3 + 3 × 3)
 
−2 0 −1
=  0 −14 0  psi
−1 0 8
   
 t1   −3 
1 1 10
t2 = √ −14 , tn = (−3 − 14 + 7) = − psi,
3 3 3
t 7
   
3 (1,1,3)
r r
1 100 762 − 100
ts = (9 + 196 + 49) − psi = = 8.576 psi
3 9 9
(c) The stress matrix at point (1,2,1) is given by

   
−2 (−7 + 8 + 1) (1 + 1 − 6) −2 2 −4
[σ](1,2,1) =  (3 − 8 − 5) 0  =  2 −10 0 
symm. −5 + 1 + 6 + 3 −4 0 5
Set the determinant of |[σ] − λ[I]| to zero:

−2 − λ 2 −4

0 = 2 −10 − λ 0
−4 0 5−λ
= −λ3 + I1 λ2 − I2 λ + I3 ,
where the stress invariants Ii are given by
I1 = (−2 − 10 + 5) = −7, I2 = 20 − 50 − 10 − (22 + 42 ) = −60,
I3 = (−2)(−5) − 2(10) − 4(−40) = 240.
Next use the transformation
1 7
λ̄i = λi + I1 = λi + → −(λ̄)3 + I¯2 λ̄ − I¯3 = 0,
3 3
where I¯i are the invariants of the deviatoric stress tensor
−2 + 73
 
2 −4
7
[σ̄] =  2 −10 + 3 0 
−4 0 5 + 73
and they are given by

1 1 529 484 1374 1
I¯2 = − + + + 40 = − = −76
2 9 9 9 18 3

1 23 22 22 23
I¯3 = − −2 2× − 4 −4 ×
3 2 3 3 3
506 88 368 2014
=− − + = = 74.59.
27 3 3 27
Then the eigenvalues of the deviatoric stress tensor are
1/2
I¯2
λ̄i = 2 cos αi .
3
Determine the αi and then λ̄i and λi as follows:
" 3/2 #
1 −1 2014 3 × 18 1
α1 = cos = (73.107) = 24.369◦ ,
3 54 1374 3
1/2
1374
λ̄1 = 1.8218 = 9.1896 → λ1 = 6.856,
18 × 3
2π
α2 = α1 + = 144.369◦ , α3 = −95.631◦ ,
3
1/2
1374
λ̄2 = −1.6256 = −8.1998 → λ2 = −10.533,
18 × 3
1/2
1374
λ̄3 = −0.196 = −.98989 → λ3 = −3.323.
18 × 3
To determine the eigenvector associated with λ1 = 6.856, set

    
−2 − λ1 2 −4  A1   0 
 2 10 − λ1 0  A2 = 0
−4 0 5 − λ1 A3 0
   
We have
−8.856A1 + 2A2 − 4A3 = 0,
2A1 − 16.856A2 = 0 → A2 = 0.11865A1 ,
4
−4A1 − 1.856A3 = 0 → A3 = − A1 = −2.155A1 .
1.856
Thus, the eigenvector associated with λ1 = 6.856 is (only components are displayed; it
is sufficient to find Ai ; it is not necessary to normalize them.)
1
Â(1) = ±(1, 0.11865, −2.155) = ±(0.42, 0.0498, −0.906).
magnitude
Similarly, to determine the eigenvector associated with λ2 = −10.533, set
(−2 + 10.533)A1 + 2A2 − 4A3 = 0,

2A1 + (−10 + 10.533)A2 = 0 → A2 = −3.752A1 ,
−4A1 + (5 + 10.533)A3 = 0 → A3 = 0.275A1 .
and obtain
1
Â(2) = ±(1, −3.752, 0.258) = ±(0.257, −0.964, 0.066).
magnitude
Lastly, the eigenvector associated with λ3 = −3.323 is calculated as
(−2 + 3.323)A1 + 2A2 − 4A3 = 0,

2A1 + (−10 + 3.323)A2 = 0 → A2 = 0.2995A1 ,
−4A1 + (5 + 3.323)A3 = 0 → A3 = 0.4806A1 ,
1
Â(3) = ±(1, 0.230976, 0.0897) = ±(0.870, 0.261, 0.418).
magnitude
4.26 The components of a stress tensor at a point P , with respect to the (x1 , x2 , x3 ) system,
are  
57 0 24
 0 50 0  MPa
24 0 43
Determine the principal stresses and principal stress directions at point P .
Solution: We set the determinant of |σ − λI| = 0 and obtain

57 − λ 0 24

0 = 0 50 − λ 0 = (57 − λ)(50 − λ)(43 − λ) + 24 [−24(50 − λ)]
24 0 43 − λ
100 SOLUTIONS MANUAL
or
√
(50 − λ) λ2 − 100λ + 1875 = 0 → λ1 = 50, λ2,3 = 50 ± 2500 − 1875 = 50 ± 25

Thus, the principal stresses are
σ1 = 25 MPa, σ2 = 50 MPa, σ3 = 75 MPa.
The principal planes are computed from the equations

   (1)   
57 − 25 0 24  n1 
  0
(1)
 0 50 − 25 0  n2 = 0
24 0 43 − 25  n(1)  0
 
3
   (2)   
57 − 50 0 24  n1 
  0
 0 50 − 50 0  n(2) = 0
 2 
24 0 43 − 50  n(2)
3
 0
We obtain
3 4
n̂(1) = ± ê1 − ê3 , n̂(2) = ±ê2 .
5 5
The third principal direction is computed using n̂(3) = n̂(1) × n̂(2) . We obtain

(3) 4 3
n̂ = ± ê1 + ê3 .
5 5

 
0 0 Ax2
[σ] =  0 0 −Bx3  MPa,
Ax2 −Bx3 0
where A and B are constants.

(a) Determine the body force vector such that the stress tensor corresponds to an equilib-
rium state.
(b) The three principal invariants of σ at the point x = Bê2 + Aê3 .
(c) Determine the principal stress components and the associated planes at the point x =
Bê2 + Aê3 .
(d) Determine the maximum shear stress and associated plane at the point x = Bê2 + Aê3 .
Solution: (a) Substituting the given stress field into the equations of equilibrium, we
obtain
∂σ11 ∂σ12 ∂σ13
+ + + ρ0 f1 = 0 + 0 + 0 + ρ0 f1 = 0,
∂x1 ∂x2 ∂x3
∂σ21 ∂σ22 ∂σ23
+ + + ρ0 f2 = 0 + 0 − B + ρ0 f2 = 0,
∂x1 ∂x2 ∂x3
∂σ31 ∂σ32 ∂σ33
+ + + ρ0 f3 = 0 + 0 + 0 + ρ0 f3 = 0.
∂x1 ∂x2 ∂x3
Thus, the body force vector should be ρf = Bê2 .
(b) The three principal invariants of the stress tensor are
I1 = σii = 0, I2 = 1
2
(σii σjj − σij σji ) = −2A2 B 2 , I3 = |[σ]| = 0.
(c) The stress tensor at the point x = Bê2 + Aê3 is

 
0 0 AB
[σ] =  0 0 −AB  MPa.
AB −AB 0
The characteristic equation for determining the principal stresses (σ) is
−λ3 + I1 λ2 − I2 λ + I3 = 0 ⇒ −λ3 + 2λ A2 B 2 = 0,
which gives the following principal stresses

√ √
σp1 = − 2AB, σp2 = 0, σp3 = 2AB.
√
The principal plane associated with σp1 = − 2AB is obtained from
√   (1)   
2AB √ 0 AB   n1   0
(1)
 0 2AB √−AB  n2 = 0 ,
AB −AB 2AB  n(1)  0
 
3
which gives
(1) (1) (1) (1)
n1 = − √12 n3 , n2 = √1 n
2 3
.
Hence, the principal plane 1 is given by (normalizing the vector)
√
{n}(1) = ± √12 −1, 1, 2 .

The principal plane associated with σp2 = −0 is obtained from

   (2)   
0 0 AB   n1   0
(2)
 0 0 −AB  n2 = 0 ,
AB −AB 0  n(2) 
 0

3
which gives
(2) (2) (2)
n1 = n2 , n3 = 0.
{n}(2) = ± √12 1, 1, 0 .

√
The principal plane associated with σp3 =
2AB is obtained from
 √   (3)   
− 2AB √ 0 AB  n1 
  0
 0 −AB  n(3)
− 2AB √ = 0 ,
 2 
AB −AB − 2AB  n(3)
3
 0
which gives
(3) (3) (1) (3)
n1 = √1 n , n2 = − √12 n3 .
2 3

√
{n}(3) = ± √12 1, −1, 2 .

(d) The maximum shear stress is given by

√
(σns )max = 12 (σp3 − σp1 ) = 2AB,
and the plane is

√
√1 √1

{n} = 2
0, 0, 2 or {n} = 2
−1, 1, 0 .
4.28 Derive the stress equilibrium equations in cylindrical coordinates by considering the
equilibrium of a typical volume element shown in Fig. P4.28. Assume that the body
force components are (not shown in the figure) ρ0 fr , ρ0 fθ , and ρ0 fz along the r, θ, and
z coordinates, respectively.
Figure P4-24
z
θ
êz
êθ ∂σ θ z ∂σ zθ
σθ z + dz σ zθ + dθ
êr ∂z ∂θ
∂σ zz
r σ zz + dz ∂σ θθ
dθ ∂z σ θθ + dθ
σ rr ∂θ
∂σ rθ
r σ rθ + dθ
σθr ∂θ
σ rθ σ rz ∂σ
σ zr σ zr + zr dr
∂r
σ θθ ∂σ
σ θ z σ zz σ θ r + θ r dr
∂r
Figure P4-24 dr σ zθ ∂σ
σ rr + rr dr
∂r
dz ∂σ rz
σ rz + dz
∂z
Fig. P4.28
Figure P4-24b
z
θ
Solution: Assumeêz that the body is subjected to a body force with components fr , fθ ,
θ θ, z) directions, respectively.
0.5d(r, dθ
and fz in the êθ ∂σ Summing the ∂σ forces in the r direction,
we obtain σ θ z + θ z dz σ zθ + zθ dθ
σ rθ
êr ∂σ r dθ ∂z
∂σ zz Area = 1∂[θrdθ + (r + dr )dθ ]dr
∂σrrσ θθ + θθr dθ σ + dz ∂σ
2 rθ
0 = σrr +θ dr (r ∂θ+ dr)dθ
dθ
zz
dz − σrr rdθ ∂z dz + σσ rθ += ∂ σ θθ dθ cos(0.5 dθ)drdz
σ θθ ∂r + ( r∂θ dθ dr )dr dθ
+ 0.5
dr θ θθ ∂θ
∂σ rθ σ rr
∂σ rθ
r σ rrθ dθ)drdz
+ dθ + σ + ∂σrz dz dr(r + 0.5dr)dθ σ rθ + − dθ
− σrθ cos(0.5 ∂θ σ θrz ∂ θ σrz dr(r + 0.5dr)dθ
r ∂z ∂σ
σ rθ σ rz σ zr + zr dr
σ zr r (r + dr )dθ

∂σθθ ∂r
− σθθ + dθ sin(0.5 dθ)drdz − σθθ sin(0.5 dθ)drdz ∂σ + fr dr(r + 0.5dr)dθdz
∂θ σ θθ σ θ r + θ r dr
σ θ z σ zz ∂r
∂σrr dr
∂σrθσ ∂σrz ∂σ ∂σθθ
= σrr + (r + dr) + z+
θ (r + 0.5dr) σ−rr σ+θθ −rr dr (0.5dθ) + rfr drdθdz,
∂r ∂θ ∂z ∂r ∂θ
dz ∂σ rz
σ rz +cos(0.5dθ)
where sin(0.5dθ) is approximated as 0.5 dθ and dz as 1 (because dθ is small).
∂z
Upon dividing throughout by rdr dθ dz and letting dr, dθ, and dz approach 0, we obtain
1 ∂σrr 1 ∂σrθ ∂σrz 1
σrr + + + − σθθ + fr = 0. (1)
Figure P4-24b r ∂r r ∂θ ∂z r
0.5dθ dθ
∂σ θθ r dθ
σ rθ σ θθ + dθ Area = 12 [ rdθ + (r + dr )dθ ]dr
∂θ
σ θθ θ = (r + 0.5 dr )dr dθ
dr θ
∂σ
r σ rθ + rθ dθ
∂θ
r (r + dr )dθ
Similarly, by setting the sum of forces in the θ-direction to zero, we obtain

∂σθr ∂σθz
0 = σθr + dr (r + dr)dθ dz − σθr rdθ dz + σθz + dz dr(r + 0.5dr)dθ
∂r ∂z

∂σrθ
− σθz dr(r + 0.5dr)dθ + σrθ + dθ sin(0.5 dθ)drdz + σrθ sin(0.5 dθ)drdz
∂θ

∂σθθ
− σθθ + dθ cos(0.5 dθ)drdz − σθθ cos(0.5 dθ)drdz + ρ0 fθ drdz
∂θ

∂σθr ∂σθz ∂σrθ ∂σθθ
= σθr + (r + dr) + (r + 0.5dr) + σrθ + 0.5 dθ + + rρ0 fθ drdθdz,
∂r ∂z ∂θ ∂θ
or
1 ∂σθr ∂σθz 1 1 ∂σθθ
σθr + + + σrθ + + ρ0 fθ = 0. (2)
r ∂r ∂z r r ∂θ
Finally, by setting the sum of forces in the z-direction to zero, we obtain

∂σzr ∂σzθ
0 = σzr + dr (r + dr)dθ dz − σzr rdθ dz + σzθ + dθ drdz
∂r ∂θ

∂σzz
− σzθ drdz + σzz + dz (r + 0.5dr)drdθ − σzz (r + 0.5dr)drdθ + ρ0 fz drdz
∂z

∂σzr ∂σθz ∂σzz
= σzr + (r + dr) + + (r + 0.5dr) + rρ0 fz drdθdz,
∂r ∂θ ∂z
or
1 ∂σzr 1 ∂σzθ ∂σzz
σzr + + + + ρ0 fz = 0. (3)
r ∂r r ∂z ∂z

 
1 0 2x2
[σ] =  0 1 4x1  MPa,
2x2 4x1 1
determine the body force vector such that the stress tensor corresponds to an equilib-
rium state.
Solution: Substituting the given stress field into the equations of equilibrium, we obtain
∂σ11 ∂σ12 ∂σ1
+ + + ρ0 f1 = 0 + 0 + 0 + ρ0 f1 = 0,
∂x1 ∂x2 ∂x3
∂σ21 ∂σ22 ∂σ23
+ + + ρ0 f2 = 0 + 0 + 0 + ρ0 f2 = 0,
∂x1 ∂x2 ∂x3
∂σ31 ∂σ32 ∂σ33
+ + + ρ0 f3 = 0 + 0 + 0 + ρ0 f3 = 0.
∂x1 ∂x2 ∂x3
Thus, the body force vector should be identically zero, ρf = 0.
5x2 x3 3x22 0
 
[σ] =  3x22 0 −x1  MPa,

0 −x1 0
determine the body force vector such that the stress tensor corresponds to an equilib-
rium state.
∂σ11 ∂σ12 ∂σ13
+ + + ρ0 f1 = 0 + 6x2 + 0 + ρ0 f1 = 0,
∂x1 ∂x2 ∂x3
∂σ21 ∂σ22 ∂σ23
+ + + ρ0 f2 = 0 + 0 + 0 + ρ0 f2 = 0,
∂x1 ∂x2 ∂x3
∂σ31 ∂σ32 ∂σ33
+ + + ρ0 f3 = 0 + 0 + 0 + ρ0 f3 = 0.
∂x1 ∂x2 ∂x3
Thus, the body force vector should be
ρ0 f = −6x2 ê1 × 106 N/m3 .


Bx21 x2
 
A(x1 − x2 ) 0
[σ] =  Bx1 x2 −A(x1 − x2 ) 0  MPa2 ,
2
0 0 0
determine the constants A and B such that the stress tensor corresponds to an equi-
librium state in the absence of body forces..
∂σ11 ∂σ12 ∂σ13
+ + = A + Bx21 + 0 = 0,
∂x1 ∂x2 ∂x3
∂σ21 ∂σ22 ∂σ23
+ + = 2BX1 X2 + A + 0 = 0,
∂x1 ∂x2 ∂x3
∂σ31 ∂σ32 ∂σ33
+ + = 0 + 0 + 0 = 0.
∂x1 ∂x2 ∂x3
Thus, there are no nonzero values of A and B that satisfy the equilibrium equations.

Ax21 x2 A(B 2 − x22 )x1
 
0
[σ] =  A(B 2 − x22 )x1 C(x22 − 3B 2 )x2 0  MPa,
0 0 2Bx23
where A, B, and C = A/3 are constants, determine the body force components neces-
sary for the body to be in equilibrium.
∂σ11 ∂σ12 ∂σ13
+ + + ρf1 = 2Ax1 x2 − 2Ax2 x1 + 0 + ρf1 = 0,
∂x1 ∂x2 ∂x3
∂σ21 ∂σ22 ∂σ23
+ + + ρf2 = A(B 2 − x22 ) + 3C(x22 − B 2 ) + 0 + ρf2 = 0,
∂x1 ∂x2 ∂x3
∂σ31 ∂σ32 ∂σ33
+ + + ρf3 = 0 + 0 + 4Bx3 + ρf3 = 0.
∂x1 ∂x2 ∂x3
Thus, the body force vector should be
ρf = −4Bx3 ê3 × 106 N/m3 .
4.33 Given the following Cauchy stress components (σij = σji ),
σ11 = −2x21 , σ12 = −7 + 4x1 x2 + x3 , σ13 = 1 + x1 − 3x2 ,

σ22 = 3x21 − 2x22 + 5x3 , σ23 = 0, σ33 = −5 + x1 + 3x2 + 3x3 ,
determine the body force components for which the stress field describes a state of
equilibrium.
Solution: The body force components are

∂σ11 ∂σ12 ∂σ13
ρf1 = − + + = −[(−4x1 ) + (4x1 ) + 0] = 0,
∂x1 ∂x2 ∂x3

∂σ12 ∂σ22 ∂σ23
ρf2 = − + + = −[(4x2 ) + (−4x2 ) + 0] = 0,
∂x1 ∂x2 ∂x3

∂σ13 ∂σ23 ∂σ33
ρf3 = − + + = −[1 + 0 + 3] = −4.
∂x1 ∂x2 ∂x3
Thus, the body force components are ρf1 = 0, ρf2 = 0, and ρf3 = −4.
4.34 Given the following stress field, expressed in terms of its components with respect to a
rectangular Cartesian basis,
σ11 = x21 x2 , σ12 = (c2 − x22 )x1 , σ13 = 0,

1 3
x2 − 3c2 x2 , σ23 = 0, σ33 = 2cx23 ,

σ22 =
3
where c is a constant, find the body-force field necessary for the stress field to be in
equilibrium.
Solution: The body force components are

∂σ11 ∂σ12 ∂σ13
ρf1 = − + + = −[(2x1 x2 ) + (−2x1 x2 ) + 0] = 0,
∂x1 ∂x2 ∂x3

∂σ12 ∂σ22 ∂σ23
ρf2 = − + + = −[(c2 − x22 ) + (x22 − c2 ) + 0] = 0,
∂x1 ∂x2 ∂x3

∂σ13 ∂σ23 ∂σ33
ρf3 = − + + = −[0 + 0 + 4cx3 ] = −4cx3 .
∂x1 ∂x2 ∂x3
Thus, the body force vector is ρf = −4cx3 ê3 .
4.35 The equilibrium configuration of a deformed body is described by the mapping
χ(X) = AX1 ê1 − BX3 ê2 + CX2 ê3 ,
where A, B, and C are constants. If the Cauchy stress tensor for this body is
 
0 0 0
[σ] =  0 0 0  MPa,
0 0 σ0
where σ0 is a constant, determine

(a) the deformation tensor and its inverse in matrix form,
(b) the matrices of the first and second Piola–Kirchhoff stress tensors, and
(c) the pseudo stress vectors associated with the first and second Piola–Kirchhoff stress
tensors on the ê3 -plane in the deformed configuration.
Solution: (a) The matrix of the deformation mapping and its inverse are (J = ABC)
   −1 
A 0 0 A 0 0
−1 −1
[F ] =  0 0 −B  , [F ] =  0 0 C .
0 C 0 0 −B −1 0
(b) Then the first and second Piola–Kirchhoff stress tensor components can be deter-
mined from P = JF−1 · σ and S = J F−1 · σ · F−T
 −1    
A 0 0 0 0 0 0 0 0
[P ] = ABC  0 0 C   0 0 0  =  0 0 ABσ0  kN/m2 ,
−1
−1
0 −B 0 0 0 σ0 0 0 0
 −1    −1   
A 0 0 0 0 0 A 0 0 0 0 0
−1 −1
[S] = ABC  0 0 C 0 0 0  0 0 −B  = ABσ0  0 1 0  MPa.
0 −B −1 0 0 0 σ0 0 C −1 0 0 0 0
(c) Consider a unit area in the deformed state in the ê3 -direction. The corresponding
undeformed area dA N̂ is given by [see Eq. (4.4.4)]
1 T C
dA N̂ = F · n̂ da = Ê2 .
J J
Thus, dA = C/J and N̂ = Ê2 . The pseudo stress vector T associated with the first
Piola–Kirchhoff stress tensor is given by Eq. (4.4.1)
T = P · N̂ = ABσ0 Ê3 .
The pseudo stress vector T̃ associated with the second Piola–Kirchhoff stress tensor is
given by Eq. (4.4.15)
AB
T̃ = N̂ · S = σ0 Ê2 .
C
4.36 A body experiences deformation characterized by the mapping

χ(X, t) = x = AX2 ê1 + BX1 ê2 + CX3 ê3 ,
where A, B, and C are constants. The Cauchy stress tensor components at certain
point of the body are given by
 
0 0 0
[σ] =  0 σ0 0  MPa,
0 0 0
where σ0 is a constant. Determine the Cauchy stress vector t and the first Piola–
Kirchhoff stress vector T on a plane whose normal in the current configuration is
n̂ = ê2 .
Solution: (a) The matrix of the deformation mapping and its inverse are (J = −ABC)
0 B −1 0
   
0 A 0
−1 −1
[F ] =  B 0 0  , [F ] =  A 0 0 .
0 0 C 0 0 C −1
(b) Then the first and second Piola–Kirchhoff stress tensor components can be deter-
mined from P = JF−1 · σ and S = J F−1 · σ · F−T
0 B −1 0
    
0 0 0 0 σ0 AC 0
[P ] = −ABC  A−1 0 0   0 σ0 0  = −  0 0 0  MPa,
0 0 C −1 0 0 0 0 0 0
−1 −1
     
0 B 0 0 0 0 0 A 0 0 0 0
[S] = −ABC  A−1 0 0   0 σ0 0   B −1 0 0  = −  0 AC
B
0  MPa.
−1 −1
0 0 C 0 0 0 0 0 C 0 0 0
(c) The unit vector in the undeformed configuration can be calculated using Nanson’s
formula,
1 1
N̂ dA = FT · n̂ da = ê2 ds
J AC
which gives dA = da/AB and N̂ = ê2 . Then the traction vectors T and T̃ are given by
T = N̂ · P = PT · N̂ = −ACσ0 ê2 MPa,
AC
T̃ = N̂ · S = S · N̂ = − σ0 ê2 MPa.
B
4.37 Express the stress equilibrium equations in Eq. (4.5.6) in terms of the stress components
and body force components in the (a) cylindrical and (b) spherical coordinate systems.
Solution: (a) We first express Eq. (4.5.6) in terms of the components of stress tensor
σ and body force vector f in the cylindrical coordinate system (see Table 2.4.2)
∂ 1 ∂ ∂
∇ = êr + êθ + êz ,
∂r r ∂θ ∂z
f = êr fr + êθ fθ + êz fz ,
σ = σrr êr êr + σθθ êθ êθ + σzz êz êz + σrθ êr êθ + σθr êθ êr
+ σrz êr êz + σzr êz êr + σθz êθ êz + σzθ êz êθ .
Taking the divergence of σ T and noting that the basis vectors êr and êθ have nonzero
derivatives with respect to θ (only) [see Eq. (2.5.29)], we arrive at the equations (see
Problem 2.49)
∂σrr 1 ∂σθr ∂σrz 1
+ + + (σrr − σθθ ) + ρ0 fr = 0, (1)
∂r r ∂θ ∂z r
∂σθr 1 ∂σθθ ∂σθz 1
+ + + (σθr + σrθ ) + ρ0 fθ = 0, (2)
∂r r ∂θ ∂z r
∂σzr 1 ∂σzθ ∂σzz 1
+ + + σzr + ρ0 fz = 0. (3)
∂r r ∂θ ∂z r
If the stress tensor σ is symmetric, then
σrθ = σθr , σrz = σzr , σzθ = σθz . (4)
(b) Using the solution of Problem 2.51, we obtain

∂σRR 1 ∂σRφ 1 ∂σRθ 1
+ + + [2σRR − σφφ − σθθ + σRφ cot φ] + ρ0 fR = 0,
∂σφR 1 ∂σφφ 1 ∂σφθ 1
+ + + [(σφφ − σθθ ) cot φ + σRφ + 2σφR ] + ρ0 fφ = 0,
∂σθR 1 ∂σθφ 1 ∂σθθ 1
+ + + [(σθφ + σφθ ) cot φ + 2σθR + σRθ ] + ρ0 fθ = 0.
4.38 Equation (4.2.7) can also be written

t = (n̂ · êi )ti = n̂ · (ê1 t1 + ê2 t2 + ê3 t3 ) . (1)
The terms in the parentheses can be defined as the stress dyadic or stress tensor T:
T ≡ ê1 t1 + ê2 t2 + ê3 t3 = êi ti . (2)
Show that T is the transpose of σ defined in Eq. (4.2.13).
Solution: From the definition of T, we have

t = n̂ · T = σ · n̂ = n̂ · σ T .
Thus T = σ T . Hence Tij is the component of stress acting on the xi -plane and in the
direction of xj -axis. When stress tensor is symmetric, the difference between the two
tensors disappears. Until the symmetry is established one should be careful in using
Cauchy’s formula.
4.39 Show that the material time derivative of the Cauchy stress tensor is not objective,
unless the superposed rigid body rotation is time-independent (that is, Q is not a
function of time); that is, show
σ̇ ∗ 6= Q · σ̇ · QT ,
unless Q is independent of time.
Solution: We have
d
σ̇ ∗ = Q · σ · QT
dt
= Q̇ · σ · QT + Q · σ̇ · QT + Q · σ · Q̇T
6= Q · σ̇ · QT .
If Q is independent of time, then we have Q̇ = 0, and

σ̇ ∗ = Q · σ̇ · QT .
4.40 Prove that if the stress tensor is real and symmetric, σij = σji , then its eigenvalues are
real. Also, prove that the eigenvectors of a real and symmetric σij are orthogonal.
Solution: Let σi be the principal stresses. Then σ1 = σ and σ2 = σ̄, where the bar
(1)
denotes the complex conjugate (so that σ1 + σ2 and σ1 σ2 are real. Now, if ni = ni
(2)
then ni = n̄i . Because σij nj = σni , we have n̄i σij nj = σn̄i ni ; similarly, σij n̄j = σ̄n̄i ,
so that ni σij n̄j = σ̄ni n̄i . However, σij = σji ; consequently, n̄i σij nj = ni σij n̄j , so that
(σ − σ̄)n̄i ni = 0. Because n̄i ni is, for any nonzero vector n, a positive real number, it
follows that σ = σ̄, that is, σ is real.
(1) (1) (2) (2) (2) (1)
Next, assume that σ1 6= σ2 ; then σij nj = σ1 ni and σij nj = σ2 ni . But ni σij nj −
(1) (2) (1) (2)
ni σij nj = 0 = (σ1 − σ2 )ni ni . Hence n(1) · n(2) = 0.
If σ1 = σ2 6= σ3 , then any vector perpendicular to n(3) is an eigenvector, so that we can
choose two that are perpendicular to each other. If σ1 = σ2 = σ3 (hydrostatic strain),
then every nonzero vector is an eigenvector; hence we can always find three mutually
perpendicular eigenvectors.
N4.1 The Cauchy stress tensor components at a point P in the deformed body with respect
 
3 1 1
[σ] =  1 1 2  MPa (1 MPa = 106 N/m2 ).
1 2 0
(a) Determine the Cauchy stress vector t(n̂) and its length at the point P on a plane
perpendicular to the vector A = ê1 − 2ê2 + 2ê3 .
(b) Find the normal and shear traction vectors and their magnitudes on the plane.
Solution: (a) The unit normal to the plane is given by
n̂ = 31 (ê1 − 2ê2 + 2ê3 ).

      
 t1  3 1 1  1  3
t2 = 31  1 1 2  −2 = 1
3
3 MPa,
t3 1 2 0 2 −3
     
or
t(n̂) = ê1 + ê2 − ê3 MPa.
The length of the stress vector is
√
|t|2 = 3, |t| = 3 = 1.732 MPa.
(b) The normal traction vector at the point is
tnn = (t(n̂) · n̂)n̂ = −n̂ = − 13 (ê1 − 2ê2 + 2ê3 ) MPa; |tnn | = −1 MPa.
The shear traction vector (projected onto the plane) at the point is
√
tns = t(n̂) − tnn = 31 (4ê1 + ê2 − ê3 ) MPa; |tns | = 2 MPa.
N4.2 The Cauchy stress tensor components at a point P in the deformed body with respect
6x1 x23 0 −2x33

 
[σ] =  0 1 2  MPa.
−2x33 2 3x21
(a) Show the the stress filed satisfies the equilibrium equations in the absence of body
forces.
(b) Determine the Cauchy stress vector t(n̂) at the point x = 2ê1 + ê2 + 2ê3 on the
plane x1 + 2x2 + 2x3 = 4 m.
(c) Find the normal and shear traction vectors on the plane.
Solution: (a) Substituting the given stress field into the equations of equilibrium, we
obtain
∂σ11 ∂σ12 ∂σ13
+ + = 6x23 + 0 − 6x23 = 0,
∂x1 ∂x2 ∂x3
∂σ21 ∂σ22 ∂σ23
+ + = 0 + 0 + 0 = 0,
∂x1 ∂x2 ∂x3
∂σ31 ∂σ32 ∂σ33
+ + = 0 + 0 + 0 = 0.
∂x1 ∂x2 ∂x3
(b) The unit normal to the plane is given by
n̂ = 31 (ê1 + 2ê2 + 2ê3 ).
The components of the Cauchy stress vector at point x = (2, 1, 2) are

      
 t1  48 0 −16  1   16 
1  1
t2 = 3 0 1 2 2 = 3 6 MPa,
t3 −16 2 12 2 12
     
or
2
t(n̂) = 3
(8ê1 + 3ê2 + 6ê3 ) MPa.
The length of the stress vector is
√
|t|2 = 72.6667, |t| = 72.6667 = 8.524 MPa.
(c) The traction vector normal to the plane at the point is
52 52 52
tnn = (t(n̂) · n̂)n̂ = 9
n̂ = 27
(ê1 + 2ê2 + 2ê3 ) , |tnn | = = 5.777 MPa.
9
The shear traction vector (projected onto the plane) at the point is
4
tns = t(n̂) − tnn = 27
(19ê1 − 14ê2 − 2ê3 ) MPa; |tns | = 3.509 MPa.
Chapter 5: CONSERVATION AND BALANCE LAWS
5.1 The acceleration of a material element in a continuum is described by

Dv ∂v
≡ + v · ∇v. (1)
Dt ∂t
where v is the velocity vector. Show by means of vector identities that the acceleration
can also be written as
2
Dv ∂v v
≡ +∇ − v × ∇ × v. (2)
Dt ∂t 2

v2
Solution: We must show that ∇ 2
− v × ∇ × v = v · ∇v. Thus, we have
v2

1 ∂ ∂vs
∇ −v×∇×v = êi (vj vj ) − (vi êi ) × εrst êt
2 2 ∂xi ∂xr
∂vj ∂vs
= êi vj − εrst εkit êk vi
∂xi ∂xr
∂vj ∂vs
= êi vj − êk (δrk δsi − δri δsk ) vi
∂xi ∂xr

∂vj ∂vs ∂vk
= êi vj − êk vs − vi
∂xi ∂xk ∂xi
∂vk ∂
= êk vi = vi (vk êk ) = v · ∇v
∂xi ∂xi
5.2 Show that the local form of the principle of conservation of mass, Eq. (5.2.22), can be
expressed as
D
(ρJ) = 0.
Dt
Solution: Equation (5.2.22), Z

D
ρ dx = 0, (5.2.22)
Dt Ω
can be written by writing dx = JdX,
Z Z
D D D
ρ dx = ρ J dX ⇒ (ρJ) = 0.
Dt Ω Dt Ω0 Dt
5.3 Use the equation

D
(ρJ) = 0
Dt
to derive the continuity equation
Dρ
+ ρ ∇ · v = 0.
Dt
Solution: Carrying out the indicated differentiation with respect to t and using the
result of Problem 3.35 (DJ/Dt = Jdiv v), we obtain

D Dρ DJ Dρ
0= (ρJ) = J +ρ =J + ρ (∇ · v) ,
Dt Dt Dt Dt
from which the required result follows.
CHAPTER 5: CONSERVATION AND BALANCE LAWS 111
Figure P5-4
5.4 Derive the continuity equation in the cylindrical coordinate system by considering a
differential volume element shown in Fig. P5.4.
( ρvz )z +Δz
Δθ ( ρvθ )θ +Δθ
( ρvr )r
x3
Δz
eˆ z ( ρvθ )θ
êθ
x2 Δr ( ρvr )r +Δr
eˆ r
θ
x1 ( ρvz )z
r
Fig. P5.4
Solution: The net mass inflow in the r-direction is

∆θ∆z(rρvr )r − ∆θ∆z(rρvr )r+∆r .
The net mass inflow in the θ-direction is
∆r∆z(ρvθ )θ − ∆r∆z(ρvθ )θ+∆θ .
The net mass inflow in the z-direction is
r∆θ∆r(ρvz )z − r∆θ∆r(ρvz )z+∆z .
The rate of increase of mass inside the differential volume element is

(ρ)t+∆t − (ρ)t
r∆θ∆r∆z .
∆t
Then by the principle of conservation of mass, we have
∆θ∆z(rρvr )r − ∆θ∆z(rρvr )r+∆r + ∆r∆z(ρvθ )θ − ∆r∆z(ρvθ )θ+∆θ + r∆θ∆r(ρvz )z

(ρ)t+∆t − (ρ)t
− r∆θ∆r(ρvz )z+∆z = r∆θ∆r∆z .
∆t
Dividing throughout by ∆r ∆θ ∆z and taking the limits ∆r → 0, ∆θ → 0, ∆z → 0,
and ∆t → 0, we obtain
∂ ∂ ∂ ∂ρ
− (rρvr ) − (ρvθ ) − r (ρvz ) = r ,
∂r ∂θ ∂z ∂t
which is the same as that in Eq. (5.2.30).
5.5 Express the continuity equation (5.2.24) in the cylindrical coordinate system (see Table
2.4.2 for various operators). The result should match the one in Eq. (5.2.30).
Solution: From Table 2.4.2 replacing A with ρv, we obtain

1 ∂ ∂ ∂
div(ρv) = (ρrvr ) + (ρvθ ) + r (ρvz ) .
r ∂r ∂θ ∂z
Therefore, the continuity equation (5.2.24) becomes

∂ρ 1 ∂(rρvr ) ∂(ρvθ ) ∂(ρvz )
0= + + +r .
∂t r ∂r ∂θ ∂z
5.6 Express the continuity equation (5.2.24) in the spherical coordinate system (see Table
2.4.2 for various operators). The result should match the one in Eq. (5.2.31).
Solution: Follows from the divergence of a vector given in Table 2.4.2 by replacing
A = ρv and noting that
AR ∂AR 1 ∂(R2 AR )
2 + = 2 .
R ∂R R ∂R
5.7 Determine if the following velocity fields for an incompressible flow satisfy the continuity
equation:
x1 x2
(a) v1 (x1 , x2 ) = − 2 , v2 (x1 , x2 ) = − 2 where r2 = x21 + x22 .
r r
r2
(b) vr = 0, vθ = 0, vz = c 1 − 2
R
where c and R are constants.
Solution: We must verify the equation

∂v1 ∂v2 ∂v3
+ + = 0.
∂x1 ∂x2 ∂x3
Also, recall from Example 2.4.1 that
∂r xi
= .
∂xi r
(a) We have (v3 = 0)
∂v1 1 x1 ∂r 1 x2
=− 2 +2 3 = − 2 + 2 41
∂x1 r r ∂x1 r r
∂v2 1 x2 ∂r 1 x22
=− 2 +2 3 =− 2 +2 4 .
∂x2 r r ∂x1 r r
The sum of the above expressions, being zero, clearly satisfies the continuity equation.
(b) Because v = v(r), we must verify

1 ∂(rρvr ) ∂(ρvθ ) ∂(ρvz )
+ +r = 0,
r ∂r ∂θ ∂z
which is trivially satisfied because vr = vθ = 0 and vz is not a function of z.
5.8 The velocity distribution between two parallel plates separated by distance b is
y y y
vx (y) = v0 − c 1− , vy = 0, vz = 0, 0 < y < b,
b b b
where y is measured from and normal to the bottom plate, x is taken along the plates,
vx is the velocity component parallel to the plates, v0 is the velocity of the top plate
in the x direction, and c is a constant. Determine if the velocity field satisfies the
continuity equation and find the volume rate of flow and the average velocity.
Solution: Because vx is only a function of y and vy = vz = 0, the continuity equation

is trivially satisfied. The volume rate of flow Q is given by
Z b
b
Q= vx (y) dy = (3v0 − c) m3 /(s.m).
0 6
The average velocity is given by
Q 1
vavg = = (3v0 − c) m/s.
b 6
Figure P5-8
5.9 Calculate the force exerted by a water (ρ = 103 kg/m3 ) jet of diameter d = 8 mm and
velocity v = 12 m/s that impinges against a smooth inclined flat plate at an angle of
45◦ to the axis of the jet, as shown in Fig. P5.9.
v
q = 45
v
F
v
Fig. P5.9
Solution: From Example 5.3.1 we have

π(8 × 10−3 )2 1
Fn = ρ Qv sin θ = ρ A v 2 sin θ = 103 × (12)2 √ = 5.118 N.
4 2
5.10 Calculate the force exerted by a water (ρ = 103 kg/m3 ) jet of diameter d = 60 mm
and velocity v = 6 m/s that impinges against a smooth inclined flat plate at an angle
of 60◦ to the axis of the jet. Also calculate the volume flow rates QL and QR .
Solution: From Example 5.3.1 we have
√
π(60 × 10−3 )2 3
Fn = ρ Qv sin θ = ρ A v 2 sin θ = 103 × (6)2 = 88.15 N.
4 2
The volume flow rates are
π(60 × 10−3 )2
QL = 0.5Q(1 + cos θ) = 0.5 6(1 + 0.5) = 0.01273 m3 /s,
4
π(60 × 10−3 )2
QR = 0.5Q(1 − cos θ) = 0.5 6(1 − 0.5) = 0.00424 m3 /s
4
5.11 A jet of air (ρ = 1.206 kg/m3 ) impinges on a smooth vane with a velocity v = 50 m/s
at the rate of Q = 0.4 m3 /s. Determine the force required to hold the plate in position
for the three different vane configurations shown in Fig. P5.11. Assume that the vane
splits the jet into two equal streams, and neglect any energy loss in the streams.
Solution: (a) From Example 5.3.1, the horizontal force is given by (θ = 90◦ )
F = ρv 2 A = ρvQ = 1.206 × 50 × 0.4 = 24.12 N.
(b) In this case θ = 60◦ and the horizontal force is given by

F = ρv 2 A (1 − cos 60◦ ) = 0.5ρvQ = 0.5 × 1.206 × 50 × 0.4 = 12.06 N.
Fig. P5.11
(c) For this case, the first equation in the solution of Example 5.3.2 gets modified to
(or replace θ there with 90 + θ)
Z
F ≡ −Fx êx + Fy êy = ρv(v· ds)
cv
= ρv (− sin θ êx + cos θ êy ) vA + ρv (−vA êx ) ,
from which we obtain (θ = 60◦ )
Fx = ρv 2 A (1 + sin θ) = 1.866ρvQ = 1.866 × 1.206 × 50 × 0.4 = 45 N.
5.12 In Example 5.3.3, determine (a) the velocity and accelerations as functions of x, and
(b) the velocity as the chain leaves the table.
Solution: The velocity and acceleration are

s r
x2 g 2
v(t) = ẋ = x0 λ sinh λt = x0 λ −1= (x − x20 ),
x20 L
g
a(t) = ẍ = x0 λ2 cosh λt = x(t).
L
The time t = t0 at which the chain leaves the table can be determined from
L
cosh λt0 = .
x0
The velocity as the chain leaves the table is
r
g 2 p
v(t0 ) = x0 λ sinh λt0 = (L − x20 ) ≈ gL when L >> x0 .
L
5.13 Using the definition of ∇, vector forms of the velocity vector, body force vector, and
the dyadic form of σ [see Eq. (5.3.23)], express the equation of motion (5.3.11) in the
cylindrical coordinate system as given in Eq. (5.3.24).
Solution: First compute the divergence of the transpose of the stress tensor:

∂ 1 ∂ ∂
êr + êθ + êz · [σrr êr êr + σrθ êr êθ + σθr êθ êr + · · · + σzz êz êz ]T .
∂r r ∂θ ∂z
Considering one term at a time and noting that the only nonzero derivatives of the
base vectors are ∂êθ /∂θ = −êr and ∂êr /∂θ = êθ , we obtain
∂ h
êr · σrr êr êr + σrθ êθ êr + σθr êr êθ + σrz êz êr + σzr êr êz
∂r i
+ σθθ êθ êθ + σθz êz êθ + σzθ êθ êz + σzz êz êz
∂σrr ∂σθr ∂σzr
= êr + êθ + êz
∂r ∂r ∂r
1 ∂ h
êθ · σrr êr êr + σrθ êθ êr + σθr êr êθ + σrz êz êr + σzr êr êz
r ∂θ i

σrr ∂êr 1 ∂σrθ σrθ ∂êr σθr ∂êr
= êθ · êr + êr + + êθ · êθ

σzr ∂êr 1 ∂σθθ σθθ ∂êθ 1 ∂σzθ
+ êθ · êz + êθ + + êz
σrr 1 ∂σrθ σrθ σθr σzr
= êr + êr + êθ + êθ + êz
r r ∂θ r r r
1 ∂σθθ σθθ 1 ∂σzθ
+ êθ − êr + êz
r ∂θ r r ∂θ
∂ h
êz · σrr êr êr + σrθ êθ êr + σθr êr êθ + σrz êz êr + σzr êr êz
∂z i
∂σrz ∂σθz ∂σzz
= êr + êθ + êz
∂z ∂z ∂z
Representing the body force and inertial vectors in the cylindrical coordinates as
ρf = ρfr êr + ρfθ êθ + ρfz êz ,

Dv ∂vr ∂vθ ∂vz
ρ =ρ êr + êθ + êz + v · ∇ · v,
Dt ∂t ∂t ∂t
where ∇v is known from Table 2.4.2, and and collecting the coefficients of êr , êθ , and êz
separately, we obtain the required equations of motion in the cylindrical coordinates.
Note that most authors show transpose of the components, because they use ∇ · σ
instead of ∇ · σ T . It does not matter when the stress tensor is is assumed to be
symmetric (in the monopolar case).
5.14 Using the definition of ∇, vectors forms of the velocity vector, body force vector, and
the dyadic form of σ [see Eq. (5.3.25)], express the equation of motion (5.3.11) in the
spherical coordinate system as given in Eq. (5.3.26).
Solution: In the spherical coordinate system (R, φ, θ), we write

∂ 1 ∂ 1 ∂
∇ = êR + êφ + êθ ,
v = êR vR + êφ vφ + êθ vθ ,
f = êR fR + êφ fφ + êθ fθ ,
σ = σRR êR êR + σRφ êR êφ + σRθ êR êθ
+ σφR êφ êR + σφφ êφ êφ + σφθ êφ êθ
+ σθR êφ êθ + σθφ êθ êφ + σθθ êθ êθ ,
Then we have

∂σRR 1 ∂σφR 1 ∂σθR
∇·σ = + +

1
+ [2σRR − σφφ − σθθ + σφR cot φ] êR
R

∂σRφ 1 ∂σφφ 1 ∂σθφ
+ + +

1
+ [(σφφ − σθθ ) cot φ + σφR + 2σRφ ] êφ
R

∂σRθ 1 ∂σφθ 1 ∂σθθ
+ + +

1
+ [(σφθ + σθφ ) cot φ + 2σRθ + σθR ] êθ .
R
Substituting these expressions and the gradient of a vector from Table 2.4.2 into Eq.
(5.3.11), we obtain the required equations of motion.
5.15 Use the continuity equation and the equation of motion to obtain the so-called conser-
vation form of the linear momentum equation
∂
(ρv) + div ρvv − σ T = ρf
∂t
Solution: Starting with the equation of motion and using the continuity equation we
arrive at the required result:

Dv ∂v
ρf = ρ − ∇ · σT = ρ + v · ∇v − ∇ · σ T
Dt ∂t
∂ ∂ρ
= (ρv) − v + ρv · ∇v − ∇ · σ T
∂t ∂t
∂
= (ρv) + v [∇ · (ρv)] + ρv · ∇v − ∇ · σ T
∂t
∂
= (ρv) + ∇ · ρvv − σ T .
∂t
5.16 Show that

v2

D
ρ = v · div σ T + ρv · f (v = |v|)
Dt 2
Solution: We have

D v · v 1 Dv Dv Dv
ρ = ρ v· + · v = ρv · (by product rule of differentiation)
Dt 2 2 Dt Dt Dt

= v · ∇ · σ T + ρf
= v · (∇ · σ T ) + ρv · f .
5.17 Deduce that

Dv Dw
∇× ≡ + w∇ · v − w · ∇v, (a)
Dt Dt
where w ≡ 12 ∇ × v is the vorticity vector. Hint: Use the result of Problem 5.1 and
the identity (you need to prove it)
∇ × (A × B) = B · ∇A − A · ∇B + A∇ · B − B∇ · A.
Solution: The proof of this identity requires the use of several other vector identities.
We need the following identities (here A is a vector and φ is a scalar function):
∇ · (∇ × A) = 0. (1)
∇ × (∇φ) = 0. (2)
v2

v · ∇v = ∇ − v × ∇ × v. (3)
2
∇ × (A × B) = B · ∇A − A · ∇B + A divB − B divA. (4)
The first two identities are well-known and easy to prove. The proof of (3) and (4) is
given here. We have
2
v 1 ∂ ∂vs
∇ −v×∇×v = êi (vj vj ) − (vi êi ) × εrst êt
2 2 ∂xi ∂xr
∂vj ∂vs
= êi vj − εrst εkit êk vi
∂xi ∂xr
∂vj ∂vs
= êi vj − êk (δrk δsi − δri δsk ) vi
∂xi ∂xr

∂vj ∂vs ∂vk
= êi vj − êk vs − vi
∂xi ∂xk ∂xi
∂vk ∂
= êk vi = vi (vk êk ) = v · grad v
∂xi ∂xi
Next consider
∂
∇ × (A × B) = êi × (Aj Bk ejkm êm )
∂xi

∂Aj ∂Bk
= eimn ejkm Bk + Aj ên
∂xi ∂xi

∂Aj ∂Bk
= (δnj δik − δnk δij ) Bk + Aj ên
∂xi ∂xi
∂Aj ∂Ai ∂Bi ∂Bk
= Bi êj − Bk êk + êj Aj − Ai êk
= B · ∇A − B∇ · A + A∇ · B − A · ∇B
Now we are ready to proceed to prove the identity. We begin with the material deriva-
tive of the velocity vector
2
Dv ∂v ∂v v
= + v · ∇v = +∇ − v × ω.
Dt ∂t ∂t 2
Taking the curl of the above expression, we obtain
2
Dv ∂w v
∇× = +∇× ∇ −v×w
Dt ∂t 2
∂w
= + ∇ × (w × v)
∂t
∂w
= + v · ∇ω − w · ∇v + w∇ · v − v∇ · w
∂t
Dw
= − w · ∇v + w∇ · v,
Dt
where we have used the identities (1)-(4).
5.18 If the stress field σ in a continuum has the following components in a rectangular
Cartesian coordinate system
x21 x2 (b2 − x22 )x1
 
0
2 2 1 2 2
[σ] = a  (b − x2 )x1 3 (x2 − 3b )x2 0  ,
0 0 2bx23
where a and b are constants, determine the body force components necessary for the
body to be in equilibrium.
Solution: Check the equilibrium equations

∂σ11 ∂σ12 ∂σ13
ρf1 = − − − = −a (2x1 x2 − 2x1 x2 + 0) = 0
∂x1 ∂x2 ∂x3
∂σ21 ∂σ22 ∂σ23
= −a (b2 − x22 ) + (x22 − b2 ) + 0 = 0

ρf2 = − − −
∂x1 ∂x2 ∂x3
∂σ31 ∂σ32 ∂σ33
ρf3 = − − − = 0 + 0 − 4abx3 .
∂x1 ∂x2 ∂x3
Thus, the equilibrium equations are satisfied only if the body forces are
ρf1 = 0, ρf2 = 0, ρf3 = −4abx3 .
5.19 If the stress field σ in a continuum has the following components in a rectangular
Cartesian coordinate system
x1 x2 x21 −x2
 
[σ] =  x21 0 0 ,
−x2 0 x21 + x22
determine the body force components necessary for the body to be in equilibrium.
Solution: Check the equilibrium equations

∂σ11 ∂σ12 ∂σ13
ρf1 = − − − = − (x2 + 0 + 0) = −x2
∂x1 ∂x2 ∂x3
∂σ12 ∂σ22 ∂σ23
ρf2 = − − − = −2x1 + 0 + 0 = −2x1
∂x1 ∂x2 ∂x3
∂σ31 ∂σ32 ∂σ33
ρf3 = − − − = 0 + 0 + 0 = 0.
∂x1 ∂x2 ∂x3
Thus, the equilibrium equations are satisfied only if the body forces are
ρf = −x2 ê1 + 2x1 ê2 .
5.20 A two-dimensional state of stress σ exists in a continuum with no body forces. The
following components of stress tensor are given (σ21 = σ12 ):
σ11 = c1 x32 + c2 x21 x2 − c3 x1 , σ22 = c4 x32 − c5 , σ12 = c6 x1 x22 + c7 x21 x2 − c8 ,
where ci are constants. Determine the conditions on the constants so that the stress
field is in equilibrium.
Solution: Check the 2-D equilibrium equations

∂σ11 ∂σ12
+ = 2c2 x1 x2 + 2c6 x1 x2 − c3 + c7 x21
∂x1 ∂x2
∂σ21 ∂σ22
+ = c6 x22 + 2c7 x1 x2 + 3c4 x22 .
∂x1 ∂x2
Thus, the equilibrium equations are satisfied only if the constants ci are
c2 + c6 = 0, c3 = 0, c7 = 0, c6 + 3c4 = 0 ⇒ c2 = −c6 = 3c4 .
All other constants are arbitrary (that is, take any values).
5.21 Given the following stress field with respect to the cylindrical coordinate system in a
body that is in equilibrium (σθr = σrθ ):

B C
σrr = 2A r + 3 − sin θ,
r r

B C
σθθ = 2A 3r − 3 − sin θ,
r r

B C
σrθ = −2A r + 3 − cos θ,
r r
where A, B, and C are constants, determine if the stress satisfies the equilibrium
equations when the body forces are zero. Assume that all other stress components are
zero.
Solution: This is a two-dimensional state of stress in the r and θ coordinates. We have

B B C
σrr − σθθ == 2A −2r + 2 3 sin θ, 2σrθ = −4A r + 3 − cos θ,
r r r
and

∂σrr B C ∂σrr B C
= 2A 1 − 3 4 + 2 sin θ, = 2A r + 3 − cos θ
∂r r r ∂θ r r

∂σθθ B C ∂σθθ B C
= 2A 3 + 3 4 + 2 sin θ, = 2A 3r − 3 − cos θ
∂r r r ∂θ r r

∂σrθ B C ∂σrθ B C
= −2A 1 − 3 4 + 2 cos θ, = 2A r + 3 − sin θ
∂r r r ∂θ r r
Substituting into the first equation of equilibrium [see Eq. (5.3.24)]

∂σrr 1 ∂σrθ ∂σrz 1
+ + + (σrr − σθθ ) = 0,
∂r r ∂θ ∂z r
we find that the first equilibrium equation is identically satisfied for any A, B, and C.
Similarly, substituting into the second equation of equilibrium
∂σθr 1 ∂σθθ ∂σθz σθr + σrθ
+ + + = 0,
∂r r ∂θ ∂z r
we find that the second equilibrium equation is also identically satisfied for any A, B,
and C.
The third equation of equilibrium
∂σzr 1 ∂σzθ ∂σzz σrz
+ + + =0
∂r r ∂θ ∂z r
is trivially satisfied.
5.22 Given the following stress field with respect to the spherical coordinate system in a
body that is in equilibrium:

B C
σRR = − A + 3 , σφφ = σθθ = − A + 3 ,
R R
where A, B, and C are constants, determine if the stress field satisfies the equilibrium
equations when the body forces are zero and all other stresses are zero.
Solution: We note that the stresses are only a function of R and σφφ = σθθ ; therefore,
the second and third equilibrium equations are trivially satisfied. The first equation
has the form
∂σRR 1
+ [2σRR − σφφ − σθθ ] = 0,
∂R R
which gives
3B 1 2B 2C
+ −2A − + 2A + = 0.
R4 R R3 R3
Thus, the equation is not satisfied unless B + 2C = 0.
5.23 For a cantilevered beam bent by a point load at the free end, for kinematically infinites-
imal deformations the bending moment M3 about the x3 -axis is given by M3 = −P x1
(see Fig. P5.22). The bending stress σ11 is given by
M3 x2 P x1 x2
σ11 = =− ,
I3 I3
where I3 is the moment of inertia of the cross section about the x3 -axis. Starting with
this equation, use the two-dimensional equilibrium equations to determine the stresses
σ22 and σ12 as functions of x1 and x2 .
Solution: From the 2D form of equilibrium equations (with zero body forces), we have
P x22
Z
∂σ11
σ12 = − dx2 + c1 (x1 ) = + c1 (x1 ),
∂x1 2I3
Z
∂σ12 dc1
σ22 = − dx2 + c2 (x1 ) = − x2 + c2 (x1 ).
∂x1 dx1
The boundary conditions that σ12 and σ22 are zero at x3 = ±h give
P h2
c1 = − , c2 = 0.
2I3
Hence, we have
P 2bh3
Figure P5-20 h2 − x22 ,

σ12 = − σ22 = 0 (I3 = ).
2I3 3
Note that the constant c1 can also be evaluated using the condition that the integral
of σ12 over the cross section be equal to −P .
x2
x2 P
P
V
2h x3 x1 σ 11
b M3
L x1
Fig. P5.23
5.24 For a cantilevered beam bent by uniformly distributed load (see Fig. P5.22), for
kinematically infinitesimal deformations the bending stress σ11 is given by [because
M3 = −q0 x21 /2]
M3 x2 q0 x21 x2
σ11 = =− ,
I3 2I3
where I3 is the moment of inertia of the cross section about the x3 -axis. Starting with
this equation, use the two-dimensional equilibrium equations to determine the stresses
σ22 and σ12 as functions of x1 and x2 .
Figure P5-21
x3 , w
q0
x1
L
Fig. P5.24
Solution: We have
q0 x1 x22
Z
∂σ11
σ12 = − dx2 + c1 (x1 ) = + c1 (x1 )
∂x1 2I3
Z 3
∂σ12 q0 x2 dc1
σ22 =− dx2 + c2 (x1 ) = − − x2 + c2 (x1 ).
∂x1 6I3 dx1
The boundary conditions that σ12 = 0 at x2 = ±h for any x1 gives
q0 x1 h2 q0 x1 h2
0= + c1 (x1 ) → c1 = − .
2I3 2I3
Thus we have
q0 x1 2 q0 x32 q0 x2 h2
h − x22 ,

σ12 = − σ22 = − + + c2 (x1 ).
2I3 6I3 2I3
The parameter c2 (x1 ) can be determined by requiring
Z L Z L
b (σ22 )x2 =−h dx1 = q0 L, b (σ22 )x2 =h dx1 = 0.
0 0
These conditions imply that c2 is a constant and equal to c2 = −q0 /2b. The stress σ22
becomes
q0 x2 q0
σ22 = 3h2 − x22 − .
6I3 2b
5.25 Given the following components of the second Piola–Kirchhoff stress tensor S and
displacement vector u in a body without body forces:
S11 = c1 X23 + c2 X12 X2 − c3 X1 , S22 = c4 X23 − c5 , S12 = c6 X1 X22 + c7 X12 X2 − c8 ,
S13 = S23 = S33 = 0, u1 = c9 X2 , u2 = c10 X1 , u3 = 0,

where ci are constants, determine the conditions on the constants so that the stress
field is in equilibrium.
Solution: Using Eq. (5.3.20), we obtain

∂S11 ∂S21 ∂S31
0= + +
∂X1 ∂X2 ∂X3

∂ ∂u1 ∂u1 ∂u1
+ S11 + S12 + S13
∂X1 ∂X1 ∂X2 ∂X3

∂ ∂u1 ∂u1 ∂u1
+ S21 + S22 + S23
∂X2 ∂X1 ∂X2 ∂X3

∂ ∂u1 ∂u1 ∂u1
+ S31 + S32 + S33
∂X3 ∂X1 ∂X2 ∂X3
= (2c2 X1 X2 − c3 ) + (2c6 X1 X2 + c7 X12 ) + c9 [c6 X22 + 2c7 X1 X2 + 3c4 X22 ],

∂S12 ∂S22 ∂S32
0= + +
∂X1 ∂X2 ∂X3

∂ ∂u2 ∂u2 ∂u2
+ S11 + S12 + S13
∂X1 ∂X1 ∂X2 ∂X3

∂ ∂u2 ∂u2 ∂u2
+ S21 + S22 + S23
∂X2 ∂X1 ∂X2 ∂X3

∂ ∂u2 ∂u2 ∂u2
+ S31 + S32 + S33
∂X3 ∂X1 ∂X2 ∂X3
= c6 X22 + 2c7 X1 X2 + 3c4 X22 + c10 [(2c2 X1 X2 − c3 ) + (2c6 X1 X2 + c7 X12 )].
Thus, the body is in equilibrium only if the following conditions hold among the con-
stants:
c3 = 0, c7 = 0, c2 + c6 = 0 if c10 6= 0; c6 + 3c4 = 0 all other constants are arbitrary.
5.26 Given the following components of the second Piola–Kirchhoff stress tensor S and
displacement vector u in a body without body forces (expressed in the cylindrical
coordinate system):
cos θ
Srr = −c1 , Srθ = Sθθ = 0,
r r r
ur = c2 log cos θ + c3 θ sin θ, uθ = −c2 log sin θ + c3 θ cos θ − c4 sin θ,
a a
where ci are constants, determine the conditions on the constants so that the stress
field is in equilibrium for (a) the linear (that is, infinitesimal deformations) case and (b)
the finite deformation case. Assume a two-dimensional state of stress and deformation
in r and θ coordinates.
Solution: Equation (5.3.19) for the present case takes the form (make use of the infor-
mation contained in Problem 2.49 and Table 2.4.2)
h i
0 = ∇ · ST + ST · ∇u

∂ur ∂uθ
= ∇ · Srr êr êr + Srr êr êr + Srr êr êθ
∂r ∂r

∂Srr ∂ ∂ur 1 ∂ ∂uθ 1 ∂ur
= + Srr + Srr + Srr + Srr êr
∂r ∂r ∂r r ∂θ ∂r r ∂r

∂ ∂uθ 2 ∂uθ
+ Srr + Srr êθ .
∂r ∂r r ∂r
Thus, we have

∂Srr ∂ ∂ur 1 ∂ ∂uθ 1 ∂ur
+ Srr + Srr + Srr + Srr = 0, (1)
∂r ∂r ∂r r ∂θ ∂r r ∂r

∂ ∂uθ 2 ∂uθ
Srr + Srr = 0. (2)
∂r ∂r r ∂r
For the linear case, Eq. (1) reduces to
∂Srr 1
+ Srr = 0, (3)
∂r r
and Eq. (2) is trivially satisfied.
Noting that
∂Srr cos θ ∂ur cos θ ∂uθ sin θ
= c1 2 , = c2 , = −c2 ,
∂r r ∂r r ∂r r
cos2 θ cos2 θ

∂ur ∂uθ sin 2θ ∂ ∂ur
Srr = −c1 c2 2 , Srr = 12 c1 c2 2 , Srr = 2c1 c2 3 ,
∂r r ∂r r ∂r ∂r r

∂ ∂uθ cos 2θ ∂ ∂uθ sin 2θ
Srr = c1 c2 2 , Srr = −c1 c2 3 .
∂θ ∂r r ∂r ∂r r
(a) Substituting the above expressions into Eqs. (3), we find that it is trivially satisfied.
Thus, the given stress field is in equilibrium (c1 is arbitrary).
(b) Substituting the above expressions into Eqs. (1) and (2), we find that Eq. (2) is
trivially satisfied while Eq. (1) gives
2 cos2 θ − sin2 θ

c1 c2 = 0 → c1 c2 = 0,
r3
and there are no restrictions on c3 and c4 . Thus, the given non-zero stress field does
not satisfy the equilibrium equations.
5.27 A sprinkler with four nozzles, each nozzle having an exit area of A = 0.25 cm2 , rotates
at a constant angular velocity of ω = 20 rad/s and distributes water (ρ = 103 kg/m3 )
at the rate of Q = 0.5 L/s (see Fig. P5.27). Determine (a) the torque T required on
the shaft of the sprinkler to maintain the given motion and (b) the angular velocity ω0
at which the sprinkler rotates when no external torque is applied.Take r = 0.1 m.
Fig. P5.26
Solution: (a) In steady state conditions, the moment of momentum gives (Liter/s=
10−3 m3 /s)
dL
= T = moment of momentum due to mass flow rate
dt
or
T = r × ρ(10−3 Q)v,
where v is the velocity of the water exiting the nozzle relative to the sprinkler, and it
is given by
10−3 Q 10−3 0.5
v= − ωr = − 20r = 5 − 20r.
4A 4 × 0.25 × 10−4
Solving the previous two equations, we obtain
T = 10−3 ρQr(5 − 20r) = 0.5 × 0.1(5 − 20 × 0.1) = 0.15 N-m.
(b) When T = 0, the angular velocity is

Q
ω0 = 10−3 = 50 rad/s = 477.5 rpm.
4Ar
5.28 Consider an unsymmetric sprinkler head shown in Fig. P5.28. If the discharge is
Q = 0.5 L/s through each nozzle, determine the angular velocity of the sprinkler.
Assume that no external torque is exerted on the system. Take A = 10−4 m2 .
Figure P5-22
Nozzle exit area, A

1 A r2 = 0.35 m
ω 2
r1 = 0.25 m
Discharge, Q = 0.5 ( L /s)
Fig. P5.28
Solution: If ω is the angular velocity of the sprinkler in steady state conditions, the net
efflux of moment of momentum is
r1 (ρQ1 )v1 + r2 (ρQ2 )v2 = 0
with Q1 = Q2 . The tangential velocities v1 and v2 of the fluid are given by
Q1 10−3 × 0.5 Q1 10−3 × 0.5

v1 = − ωr1 = − 0.25ω, v2 = − ωr2 = − 0.35ω.
A 10−4 A 10−4
From the two equations, we obtain
0.25(5 − 0.25ω) + 0.35(5 − 0.35ω) = 0 → ω = 16.21 rad/s = 154.8 rpm.
5.29 Show that for a multipolar continuum the Clausius–Duhem inequality (5.4.24) remains
unchanged.
Solution: This is a trivial exercise because the couple stress tensor M and body couple
ρc do not enter the second law of thermodynamics as the law is concerned with heat
input and entropy production.
5.30 Establish the following alternative form of the energy equation (σ T = σ):
v2

D
ρ e+ = ∇ · (σ · v) + ρf · v + ρr − ∇ · q.
Dt 2
Solution: The kinetic energy of the system is given by

Z
1
K= ρv · v dx.
2 R
The total internal energy of the system is given by
Z
U= ρe dx.
R
The power input is I Z

W = t · v ds + ρf · v dx.
s R
The total heat input is I Z
H=− q · n̂ds + ρr dx.
s R
Substituting into the principle of conservation of energy, we obtain
Z 2 I Z
D v
ρ + e dx = n̂ · [−q + (σ · v)] ds + (ρr + ρf · v) dx
Dt R 2 s R
I Z
= ∇ · [−q + (σ · v)] dx + (ρr + ρf · v) dx
R R
from which the required local form of the energy equation follows.
5.31 Establish the following thermodynamic form of the energy equation (σ T = σ):
De
ρ = ∇ · (σ · v) − v · ∇ · σ + ρr − ∇ · q.
Dt
Solution: From the solution of the above problem, we have

De Dv
ρ = −ρ + ∇ · σ + ρf · v + ∇ · (σ · v) − ∇ · σ · v + ρr − ∇ · q.
Dt Dt
5.32 The total rate of work done by the surface stresses per unit volume is given by ∇·(σ·v).
The rate of work done by the resultant of the surface stresses per unit volume is given
by v · ∇ · σ. The difference between these two terms yields the rate of work done by
the surface stresses in deforming the material particle, per unit volume. Show that this
difference can be written as σ : D, where D is the strain rate tensor defined in Eq.
(5.4.8).
Solution: We have
(∇ · σ · v) − ∇ · σ · v = ∇ · σ · v + σ : ∇v − ∇ · σ · v
= σ : ∇v = σ : (D + W)
= σ : D,
where σ : W = 0 because of the skew symmetry of W.
5.33 The rate of internal work done (power) in a continuous medium in the current config-
uration can be expressed as Z
1
W = σ : D dv (1)
2 v
where σ is the Cauchy stress tensor and D is the rate of deformation tensor (that is,
symmetric part of the velocity gradient tensor):
1h i dx
D= (∇v)T + ∇v , v = (2)
2 dt
The pair (σ, D) is said to be energetically conjugate because it produces the (strain)
energy stored in a deformable medium. Show that (a) the first Piola–Kirchhoff stress
tensor P is energetically conjugate to the rate of deformation gradient Ḟ, and (b) the
second Piola–Kirchhoff stress tensor S is energetically-conjugate to the rate of Green
strain tensor Ė. Hints: Note the following identities:
1
dx = J dX, L ≡ ∇v = Ḟ · F−1 , P = JF−1 · σ, σ = F · S · FT .
J
Solution: Using the symmetry of σ and D, and

Z Z
dx = J dV, L ≡ ∇v = I − F−1 , P = Jσ · F−T
v V
we can write
Z Z
1 1
U= σ : D dv = Jσ : ∇u dV
2 v 2 V
Z
1
P · FT : I − F−1 dV

=
2 V
Note that this product makes no sense because
P · FT = PiI FjI êi êj

while
I − F−1 = δIi − F−1 Ii ÊI êi

Z Z
1 1
U= σ : D dv = Jσ : ∇u dV
2 v 2 V
Z
1
P · FT : I − F−1 dV

=
2 V
and the double-dot product makes no sense. On the other hand, we can show that P
is power-conjugate to the rate of deformation gradient. First note that
T T T
d ∂x ∂ dx ∂v
Ḟ = = = = (∇0 v)T , ∇ = F−T · ∇0
dt ∂X ∂X dt ∂X
Then
Z Z Z
1 1 1
U̇ = σ : ė dv = Jσ : (∇v)T dV = Jσ : Ḟ · F−1 dV
2 v 2 V 2 V
Next note that

h i
σ : Ḟ · F−1 = σ : ḞiI êi ÊI · (F−1 )Kj ÊK êj = σ : ḞiI (F−1 )Ij êi êj

= σji ḞiI (F−1 )Ij = σij (F−T )jI ḞiI = σ · (F−T ) : Ḟ
Hence, we have
Z Z Z
1 1 1
U̇ = Jσ : Ḟ · F−1 dV = Jσ · F−T : Ḟ dV = P : Ḟ dV
2 V 2 V 2 V
Thus P is conjugate to the rate of the deformation gradient Ḟ.

This page is intentionally left blank

Chapter 6: CONSTITUTIVE EQUATIONS
6.1 Recall from Examples 3.4.3 and 4.3.1 that under the coordinate transformation
ê1 = cos θ êx + sin θ êy ,

ê2 = − sin θ êx + cos θ êy , (1)
ê3 = êz ,
the stress components and strain components εi and σi are given in terms of the
components σxx , σyy , · · · and εxx , εyy , · · · by [see Eqs. (3.4.33) and (4.3.7)]
cos2 θ sin2 θ 0 0 1
    
 ε1  0 2
sin 2θ  εxx 
2 2 1
   
 ε2   sin θ cos θ 0 0 0 − 2 sin 2θ  εyy 
  
 

    
 

ε3 0 0 1 0 0 0 ε
    
zz
, {ε̄} = [T θ ]{ε},

=  (2)
 ε
  4 
 
 0 0 0 cos θ − sin θ 0 
 
 ε yz 

 ε5  0 0 0 sin θ cos θ 0  εxz 

  
   
   
ε6 − sin 2θ sin 2θ 0 0 0 cos 2θ εxy
  
cos2 θ sin2 θ 0
     
 σ1  0 0 sin 2θ  σxx 
2
cos2 θ 0
   
 σ2   sin θ 0 0 − sin 2θ   σyy 
 
   
 

  
σ3 0 0 1 0 0 0  σzz 
  
, {σ̄} = [Rθ ]{σ}.

= 

 σ4 

 0 0 0 cos θ − sin θ 0   
 σyz 

σ5   0 0 0 sin θ cos θ 0  σxz 
 
   

  
  1 1
 
σ6 − 2 sin 2θ 2 sin 2θ 0 0 0 cos 2θ σxy
  
(3)
Show that
[S̄] = [T θ ][S][T θ ]T , [C̄] = [Rθ ][C][Rθ ]T , (4)
where [S̄] is the matrix of compliance coefficients and [C̄] is the matrix of stiffness coef-
ficients with respect to the (x1 , x2 , x3 ) coordinates and [S] is the matrix of compliance
coefficients and [C] is the matrix of stiffness coefficients with respect to the (x, y, z)
coordinates.
Solution: We note that
[Rθ ]−1 = [R−θ ], [T θ ]−1 = [T −θ ], [T θ ] = [R−θ ]T , [T −θ ] = [Rθ ]T . (5)
Then
{ε} = [S]{σ} = [S][Rθ ]−1 {σ̄} or [T θ ]−1 {ε̄} = [S][Rθ ]−1 {σ̄},
or
{ε̄} = [T θ ][S][Rθ ]−1 {σ̄} ⇒ [S̄] = [T θ ][S][T θ ]T . (6)
Similarly, we have
{σ} = [C]{ε} = [C][T θ ]−1 {ε̄} or [Rθ ]−1 {σ̄} = [C][T θ ]−1 {ε̄},
or
{σ̄} = [Rθ ][S][T θ ]−1 {ε̄} ⇒ [C̄] = [Rθ ][C][Rθ ]T . (7)
6.2 Under the coordinate transformation

ˆ1 = cos θ ê1 + sin θ ê2 ,
ē
ˆ2 = − sin θ ê1 + cos θ ê2 ,
ē
ˆ3 = ê3 ,
ē
determine S̄ij in terms of Sij and C̄ij in terms of Cij .

CHAPTER 6: CONSTITUTIVE EQUATIONS 129
Solution: Carrying out the matrix multiplications in Eqs. (6) and (7) for a triclinic
material, we obtain
S̄11 = S11 cos4 θ − 2S16 cos3 θ sin θ + (2S12 + S66 ) cos2 θ sin2 θ
− 2S26 cos θ sin3 θ + S22 sin4 θ
S̄12 = S12 cos4 θ + (S16 − S26 ) cos3 θ sin θ + (S11 + S22 − S66 ) cos2 θ sin2 θ
+ (S26 − S16 ) cos θ sin3 θ + S12 sin4 θ
S̄13 = S13 cos2 θ − S36 cos θ sin θ + S23 sin2 θ
S̄16 = S16 cos4 θ + (2S11 − 2S12 − S66 ) cos3 θ sin θ + 3(S26 − S16 ) cos2 θ sin2 θ
+ (S66 + 2S12 − 2S22 ) cos θ sin3 θ − S26 sin4 θ
S̄22 = S22 cos4 θ + 2S26 cos3 θ sin θ + (2S12 + S66 ) cos2 θ sin2 θ
+ 2S16 cos θ sin3 θ + S11 sin4 θ
S̄23 = S23 cos2 θ + S36 cos θ sin θ + S13 sin2 θ
S̄26 = S26 cos4 θ + (2S12 − 2S22 + S66 ) cos3 θ sin θ + 3(S16 − S26 ) cos2 θ sin2 θ
+ (2S11 − 2S12 − S66 ) cos θ sin3 θ − S16 sin4 θ
S̄33 = S33
S̄36 = 2(S13 − S23 ) cos θ sin θ + S36 (cos2 θ − sin2 θ)
S̄66 = S66 (cos2 θ − sin2 θ)2 + 4(S16 − S26 )(cos2 θ − sin2 θ) cos θ sin θ
+ 4(S11 + S22 − 2S12 ) cos2 θ sin2 θ
S̄44 = S44 cos2 θ + 2S45 cos θ sin θ + S55 sin2 θ
S̄45 = S45 (cos2 θ − sin2 θ) + (S55 − S44 ) cos θ sin θ
S̄55 = S55 cos2 θ + S44 sin2 θ − 2S45 cos θ sin θ
S̄14 = S14 cos3 θ + (S15 − S46 ) cos2 θ sin θ + (S24 − S56 ) cos θ sin2 θ + S25 sin3 θ
S̄15 = S15 cos3 θ − (S14 + S56 ) cos2 θ sin θ + (S25 + S46 ) cos θ sin2 θ − S24 sin3 θ
S̄24 = S24 cos3 θ + (S25 + S46 ) cos2 θ sin θ + (S14 + S56 ) cos θ sin2 θ + S15 sin3 θ
S̄25 = S25 cos3 θ + (−S24 + S56 ) cos2 θ sin θ + (S15 − S46 ) cos θ sin2 θ − S14 sin3 θ
S̄34 = S34 cos θ + S35 sin θ
S̄35 = S35 cos θ − S34 sin θ
S̄46 = (2S14 − 2S24 + S56 ) cos2 θ sin θ + (2S15 − 2S25 − S46 ) cos θ sin2 θ
+ S46 cos3 θ − S56 sin3 θ
S̄56 = (2S15 − 2S25 − S46 ) cos2 θ sin θ + (2S24 − 2S14 − S56 ) cos θ sin2 θ
+ S56 cos3 θ + S46 sin3 θ. (1)
C̄11 = C11 cos4 θ − 4C16 cos3 θ sin θ + 2(C12 + 2C66 ) cos2 θ sin2 θ
− 4C26 cos θ sin3 θ + C22 sin4 θ
C̄12 = C12 cos4 θ + 2(C16 − C26 ) cos3 θ sin θ + (C11 + C22 − 4C66 ) cos2 θ sin2 θ
+ 2(C26 − C16 ) cos θ sin3 θ + C12 sin4 θ
C̄13 = C13 cos2 θ − 2C36 cos θ sin θ + C23 sin2 θ
C̄16 = C16 cos4 θ + (C11 − C12 − 2C66 ) cos3 θ sin θ + 3(C26 − C16 ) cos2 θ sin2 θ
+ (2C66 + C12 − C22 ) cos θ sin3 θ − C26 sin4 θ
C̄22 = C22 cos4 θ + 4C26 cos3 θ sin θ + 2(C12 + 2C66 ) cos2 θ sin2 θ
+ 4C16 cos θ sin3 θ + C11 sin4 θ
C̄23 = C23 cos2 θ + 2C36 cos θ sin θ + C13 sin2 θ
C̄26 = C26 cos4 θ + (C12 − C22 + 2C66 ) cos3 θ sin θ + 3(C16 − C26 ) cos2 θ sin2 θ
+ (C11 − C12 − 2C66 ) cos θ sin3 θ − C16 sin4 θ
C̄33 = C33
C̄36 = (C13 − C23 ) cos θ sin θ + C36 (cos2 θ − sin2 θ)

C̄66 = 2(C16 − C26 ) cos3 θ sin θ + (C11 + C22 − 2C12 − 2C66 ) cos2 θ sin2 θ
+ 2(C26 − C16 ) cos θ sin3 θ + C66 (cos4 θ + sin4 θ)
C̄44 = C44 cos2 θ + C55 sin2 θ + 2C45 cos θ sin θ
C̄45 = C45 (cos2 θ − sin2 θ) + (C55 − C44 ) cos θ sin θ
C̄55 = C55 cos2 θ + C44 sin2 θ − 2C45 cos θ sin θ
C̄14 = C14 cos3 θ + (C15 − 2C46 ) cos2 θ sin θ + (C24 − 2C56 ) cos θ sin2 θ + C25 sin3 θ
C̄15 = C15 cos3 θ − (C14 + 2C56 ) cos2 θ sin θ + (C25 + 2C46 ) cos θ sin2 θ − C24 sin3 θ
C̄24 = C24 cos3 θ + (C25 + 2C46 ) cos2 θ sin θ + (C14 + 2C56 ) cos θ sin2 θ + C15 sin3 θ
C̄25 = C25 cos3 θ + (2C56 − C24 ) cos2 θ sin θ + (C15 − 2C46 ) cos θ sin2 θ − C14 sin3 θ
C̄34 = C34 cos θ + C35 sin θ
C̄35 = C35 cos θ − C34 sin θ
C̄46 = C46 cos3 θ + (C56 + C14 − C24 ) cos2 θ sin θ + (C15 − C25 − C46 ) cos θ sin2 θ
− C56 sin3 θ
C̄56 = C56 cos3 θ + (C15 − C25 − C46 ) cos2 θ sin θ + (C24 − C14 − C56 ) cos θ sin2 θ
+ C46 sin3 θ (2)
6.3 Given the transformation

ˆ1 = −ê1 , ē
ē ˆ2 = ê2 , ē
ˆ3 = ê3 , (1)
determine the stress components σ̄ij in terms of σij , strain components ε̄ij in terms of
εij , and the elasticity coefficients C̄ij in terms of Cij .
Solution: The transformation matrix is

 
−1 0 0
[L] =  0 1 0.
0 0 1
Therefore, we have
 
σ11 −σ12 −σ13
[σ̄] = [L][σ][L]T =  −σ12 σ22 σ23  .
−σ13 σ23 σ33
Similarly, we have  
ε11 −ε12 −ε13
T
[ε̄] = [L][ε][L] =  −ε12 ε22 ε23  .
−ε13 ε23 ε33
So, we can write
    

 σ̄1 
 1 0 0 0 0 0   σ1 

σ̄ 0 1 0 0 0 0 σ2 
   


 2 
 
 
     
σ̄3 0 0 1 0 0 0  σ3
 
= → {σ̄} ≡ [T ]{σ},
σ̄
 4
  0 0 0 1 0 0  σ 4

σ̄ 0 0 0 0 −1 0   σ5 
   
5

 
  

  
 
σ̄6 0 0 0 0 0 −1 σ6
 
where we note that [T ] = [T ]−1 . Then
[C̄]{ε̄} = [T ][C]{ε}, [C̄][T ]{ε} = [T ][C]{ε}, or [C̄] = [T ][C][T ].

Carrying out the indicated matrix multiplications, we obtain

   
C̄11 C̄12 C̄13 C̄14 C̄15 C̄16 C11 C12 C13 C14 C15 C16
 C̄21 C̄22 C̄23 C̄24 C̄25 C̄26   C21 C22 C23 C24 −C25 −C26 
   
 C̄31 C̄32 C̄33 C̄34 C̄35 C̄36   C31 C32 C33 C34 −C35 −C36 
 C̄41 C̄42 C̄43 C̄44 C̄45 C̄46   C41 C42 C43 C44 −C45 −C46  .
 = 
   
 C̄51 C̄52 C̄53 C̄54 C̄55 C̄56   −C51 −C52 −C53 −C54 C55 C56 
C̄61 C̄62 C̄63 C C̄64 C̄65 C̄66 −C61 −C62 −C63 −C64 C65 C66
6.4 Establish the following relations between the Lame’ constants µ and λ and engineering
constants E, ν, and K:
νE E E
λ= , µ=G= , K= .
(1 + ν)(1 − 2ν) 2(1 + ν) 3(1 − 2ν)
Solution: First recall that the stress–strain and strain-stress relations for an isotropic
material are given by

1 λ
σij = 2µεij + λδij εkk , εij = σij − σkk δij . (1)
2µ 2µ + 3λ
By definition of the shear modulus G and bulk modulus, we have
σ12 = G 2ε12 ⇒ µ = G, σkk = 3Kεkk ⇒ 2µ + 3λ = 3K. (2)
The total strain in an isotropic unit cube subjected to three normal stresses (σ11 , σ22 , σ33 )
is
σ11 σ22 σ33 1+ν ν
ε11 = −ν −ν = σ11 − (σ11 + σ22 + σ33 ) . (3)
E E E E E
But from Eq. (1), we have
1 λ
ε11 = σ11 − (σ11 + σ22 + σ33 ) . (4)
2µ 2µ(2µ + 3λ)
From (3) and (4), it follows that
E λ ν
µ=G= , = . (5)
2(1 + ν) 2G(2G + 3λ) E
From the last two equations, we obtain
2Gν νE 1 λE E
λ= = , K= = . (6)
1 − 2ν (1 + ν)(1 − 2ν) 3ν 2G 3(1 − 2ν)
It can also be established that
2 + 3r λ
E= µ, where r= .
1+r µ
6.5 Determine the longitudinal stress σxx and the hoop stress σyy in a thin-walled circular
cylindrical pressure vessel with closed ends; that is, establish Eq. (1) of Example 6.3.1.
Assume an internal pressure of p, internal diameter Di , and thickness h.
Solution: The discussion presented here is taken from the book by Fenner and Reddy
(2012). Let us consider a long cylinder of circular cross section, with an internal
diameter of Di and a constant wall thickness of h, as shown in Fig. 6.5.1(a), which is
subject to an internal pressure of p. In many cases, p is the gage pressure within the
cylinder, taking the external ambient pressure to be zero on the gage. By thin-walled
we mean that the thickness h is very small compared to diameter Di , and we may
quantify this by stating that the ratio h/Di of thickness to radius should be less than
about 0.05. The small piece of the wall of the cylinder which is shown shaded in Fig.
6.5.1(a) is shown in isolation in Fig. 6.5.1(b). In extracting this piece, we have cut the
wall along planes perpendicular and parallel to the axis of the cylinder, and the normal
stress components acting on the cut surfaces are the longitudinal or axial stress, σzz ,
in the axial direction, and the circumferential or hoop stress, σθθ , in the circumferential
direction, as shown. The term hoop stress derives from the fact that the metal hoops
Figure P6-5-1
which hold the wooden staves of a barrel experience this type of circumferential loading.
Collectively, the hoop and axial stress components are sometimes referred to as the
membrane stresses, because they act in the plane of the thin cylinder wall, which can
be regarded as a membrane.
h
Ri
q
z
sqq
r
(a)
szz
Fig. 6.5.1 (b) szz
sqq
Let us consider the cross-sectional view of the cylinder, looking along its axis, shown
in Fig. 6.5.2. The cylinder has been cut into two semicircular pieces, the upper and
lower halves, denoted as (a) and (b), respectively. Let us assume that the length of the
cylinder is L. (does not enter the calculation). The horizontal and vertical forces are
labeled as N and V respectively, and the moment is labeled as M . Their directions
are such that they tend to open both the upper and lower halves, and Newton’s third
law is satisfied. We must therefore conclude that N = 0. This provides an example
of a general principle: there are no shear stresses (and hence forces) on a plane of
symmetry through a body. The symmetry required for this to be true applies not only
Figure
to the geometry of theP6-5-2
body but also to the loads acting on it, and is sometimes referred
to as mirror symmetry. We will come across further examples elsewhere.
h
p
(a)
Ri M
M θ
N N
V V
Fig. 6.5.2
V V
N N
(b) M M
By symmetry about the horizontal line (when we flip the lower part, it must match
the upper one), we conclude that N = 0. The assumption which is only applicable to
thin-walled cylinders is that the hoop and axial stresses are constant through the wall
thickness. The effect of this in terms of the loads shown in Fig. P6.5.2 is that the force
V , which is the resultant of the normal hoop stress acting over the cut surface of one
wall of the cylinder, acts at the mid thickness of the wall. This also means that there
is no moment, and M = 0. Considering now the cross-sectional plane normal to the
axis cutting the cylinder to give Fig. 6.5.2, because this is also a plane of symmetry for
the cylinder as a whole, there are no shear forces in either the radial or circumferential
directions acting on the cut surfaces shown. Also, due to the uniformity of the normal
axial stress on these surfaces, there are no moments. As a result of eliminating shear
forces and bending moments in the cylinder wall, the analysis of the remaining normal
stresses becomes a statically determinate problem.
We may therefore consider the equilibrium of the piece of axial length dz of the upper
half of the cylinder shown in Fig. 6.5.3(a), and treat it as a free body. External forces
in the axial direction, which are self-balancing, are not shown. The only vertical forces
acting on the body are those due to the internal pressure, p, and the hoop stress, σθθ .
We may determine the first of these as follows. Consider the small element of shell
shown shaded in Fig. 6.5.3(a), whose position relative to the horizontal diameter is
given by the angle θ, and which subtends a small angle dθ at the axis of the cylinder.
The force on this element due to the pressure acting over its internal surface, which is a
rectangle dz long by Ri dθ wide, is pRi dθ dz, acting normal to the surface and therefore
at an angle of θ to the horizontal. The vertical component of this force is pRi dθ dz sin θ.
We may now sum over all such small elements which make up the piece of shell to find
theFigure P6-5-3
total vertical pressure force
Z π Z π
P = pRi dz sin θ dθ = pRi dz sin θ dθ = 2pRi dz. (1)
0 0
h Ri dθ
p
dθ
(a) θ
dz
θ θ
Ri
Fig. 6.5.3
(b)
θ
θ Pressure projected onto
horizontal surface
The result in Eq. (1) could have been obtained in a much more straightforward way,
without the need to integrate over the internal surface of the cylinder. Suppose that,
instead of taking our free body as just the piece of cylinder wall as in Fig. 6.5.3(a), we
also include the fluid it contains to create the free body shown in Fig. 6.5.3(b). The
internal pressure p now acts over the area of the horizontal plane of symmetry through
the axis of the cylinder which is contained within the cylinder. This rectangular area
has dimensions 2Ri (the internal diameter of the cylinder) by dz, giving a total vertical
pressure force of P = 2pRi dz, as in Eq. (1).
The other vertical forces acting on the free body shown in either Fig. 6.5.2 or Fig.
6.5.3 are those associated with the uniform hoop stress, σθθ , which acts over the two
cut surfaces of the cylinder wall, each of which has an area of hdz. The total force is
therefore 2σθθ hdz, which must balance the total pressure force, P , for the free body to
be in equilibrium. Therefore
2σθθ hdz = 2pRi dz
or
pRi pDi
σθθ = = . (2)
h 2h
The form of result for axial stress σzz depends on the type of end constraints applied
to the cylinder. The most common case, generally referred to as closed-ended, is where
the ends are closed by some form of end caps, which can be flat, hemispherical, or
some other curved shape. The important feature is that the end caps are attached to
the cylinder, so that the axial force exerted by the internal pressure on these caps is
transferred to the cylinder walls. We must bear in mind, however, that the present
formulae for hoop and axial stresses are not applicable in the regions of the cylinder
close to the end caps, where more complicated states of stress exist. Figure 6.5.3(b)
shows part of a closed cylinder cut by a plane at right angles to its axis. We may treat
this as a free body, subject to axial forces which must be in equilibrium. The total
axial pressure force is that due to internal pressure acting over the internal area of the
cylinder in the cross-sectional plane, or pπRi2 . The axial force associated with the axial
stress acting over the cross-sectional area of the wall is σz π(Ro2 − Ri2 ), where Ro is the
outer radius of the cylinder, and Ro = Ri + h. Therefore, we have
pπRi2 = σz π(Ro2 − Ri2 ) = σz π(Ro − Ri )(Ro + Rj ) = 2σz πhRm ,
where Rm = (Ro + Ri )/2 is the mean radius of the cylinder, and
pRi2
σzz = . (3)
2hRm
Because the mean and internal radii are almost identical, it is common practice to
approximate this result by
pRi pDi
σzz = = . (4)
2h 4h
6.6 Determine the stress tensor components at a point in 7075-T6 aluminum alloy body
(E = 72 GPa, and G = 27 GPa) if the strain tensor at the point has the following
components with respect to the Cartesian basis vectors êi :
 
200 100 0
[ε] =  100 300 400  × 10−6 m/m.
0 400 0
Solution: The strain-stress relations in Eq. (6.2.27) can be inverted to obtain the
stress–strain relations
    

 σ1  C11 C12 C13 0 0 0  ε1 
 
σ  C12 C22 C23 0 0 0   ε2 
  


 2 

  


 
 
σ3 C C C 0 0 0 ε
   
13 23 33 3

=  , (1)

 σ4   0 0 0 C44 0 0  
 
 ε4 
 σ5 

   0 0 0 0 C55 0  
 
 ε5 



  
 
σ6 0 0 0 0 0 C66 ε6
 
where Cij = Cji are the stiffness coefficients, which can be expressed in terms of the
engineering constants as
1 − ν23 ν23 ν21 + ν31 ν23 ν12 + ν32 ν13
C11 = , C12 = =
E2 E3 ∆ E2 E3 ∆ E1 E3 ∆
ν31 + ν21 ν32 ν13 + ν12 ν23 1 − ν13 ν31
C13 = = , C22 = ,
E2 E3 ∆ E1 E2 ∆ E1 E3 ∆
ν32 + ν12 ν31 ν23 + ν21 ν13 1 − ν12 ν21 (2)
C23 = = , C33 = ,
E1 E3 ∆ E1 E3 ∆ E1 E2 ∆
C44 = G23 C55 = G31 C66 = G12
1 − ν12 ν21 − ν23 ν32 − ν31 ν13 − 2ν21 ν32 ν13
∆= .
E1 E2 E3
First we compute Poisson’s ratio
E 72 1
ν= −1= −1= .
2G 54 3
For isotropic material, the material constants of Eq. (2) become [∆ = (1 + ν)2 (1 − 2ν)]
(1 − ν)E νE
C11 = C22 = C33 = = 108, C12 = C13 = C23 = = 54,
(1 + ν)(1 − 2ν) (1 + ν)(1 − 2ν)
E
C44 = C55 = C66 = G = = 27.
2(1 + ν)
The components of stress tensor are
    
  108 54 54 0 0 0  200  37.8 
σ11  54 108 54 0
  
 
0 0   300  43.2 

 
   
 
 σ22 

    
 
 

  54 54 108 0 0 0 

0

−6

27.0

σ23 = 109  6

 0 0 0 27 10 = 10 Pa.
0 0 
 800  21.6 
 σ13 

 
    
 
   0 0 0 0 27 0 

0 
 
 0.0 

σ12
  
 
 
 

0 0 0 0 0 27 200 5.4
   
6.7 For the state of stress and strain given in Problem 6.6, determine the stress and strain
invariants.
Solution: The stress invariants Ii are

I1 = σii = 108 MPa,
1 1
(37.8)2 + (43.2)2 + (27)2 + 2(21.6)2 + 2(5.4)2 = 2, 507.76 MPa2 ,

I2 = σij σij =
2 2
I3 = |σ| = 25, 666.67 MPa3 .
The strain invariants Ji are
J1 = εii = 500 × 10−6 ,
1 1
(37.8)2 + (43.2)2 + (27)2 + 2(21.6)2 + 2(5.4)2 = 235 × 10−9 ,

J2 = εij εij =
2 2
J3 = |ε| = −32 × 10−12 .
6.8 If the components of strain at a point in a body made of structural steel are
 
36 12 30
[ε] =  12 40 0  × 10−6 m/m.
30 0 25
Assuming that the Lamé constants for the structural steel are λ = 207 GPa (30 × 106
psi) and µ = 79.6 GPa (11.54 × 106 psi), determine the principal invariants of stress
and strain tensors.
Solution: The principal invariants of the strain are

J1 = εkk = (36 + 40 + 25)10−6 = 101 × 10−6 ,
1
J2 = 2
(εii εjj − εij εij )
1
1012 − 362 − 402 − 252 − 2 × 122 − 2 × 302 × 10−12 = 2296 × 10−12 , .

= 2
J3 = |[ε]| = [36 × (40 × 25 − 0) − 12(12 × 25 − 0) + 30(0 − 30 × 40)] × 10−18

= −3600 × 10−18
For isotropic material, the stress–strain equations are given as
σ11 = 2µε11 + λεkk = 103 (159.2 × 36 + 207 × 101) = 26.638 MPa,
σ22 = 2µε22 + λεkk = 103 (159.2 × 40 + 207 × 101) = 27.275 MPa,
σ33 = 2µε33 + λεkk = 103 (159.2 × 25 + 207 × 101) = 24.887 MPa,
σ12 = 2µε12 = 1.910 MPa, σ13 = 2µε13 = 4.776 MPa, σ23 = 0,
The stress invariants Ii are
I1 = σii = 78.8 MPa,
1 1
(26.638)2 + (27.275)2 + (24.887)2 + 2(1.91)2 + 2(4.776)2

I2 = σij σij =
2 2
= 1, 062.89 MPa2 ,
I3 = |σ| = 17, 368.75 MPa3 .
6.9 The components of a stress tensor at a point in a body made of structural steel are
 
42 12 30
[σ] =  12 15 0  MPa.
30 0 −5
Assuming that the Lamé constants for structural steel are λ = 207 GPa (30 × 106 psi)
and µ = 79.6 GPa (11.54 × 106 psi), determine the principal invariants of the strain
tensor.
Solution: The trace of the stress tensor is σkk = 42+15−5 = 52. Using the strain-stress
relations,
1 λ
εij = σij − σkk δij
2µ 2µ + 3λ
we can now compute the strains:

1 λ
ε11 = σ11 − σkk = 6.2814 × 10−6 [42 − 0.2653(52)] = 177.16 × 10−6 m/m
2µ 2µ + 3λ

1 λ
ε22 = σ22 − σkk = 6.2814 × 10−6 [15 − 0.2653(52)] = 7.56 × 10−6 m/m
2µ 2µ + 3λ

1 λ
ε33 = σ33 − σkk = 6.2814 × 10−6 [−5 − 0.2653(52)] = −118.07 × 10−6 m/m
2µ 2µ + 3λ
1
ε12 = σ12 = 6.2814 × 10−6 (12) = 75.38 × 10−6 m/m,
2µ
1
ε13 = σ13 = 6.2814 × 10−6 (30) = 188.44 × 10−6 m/m, ε23 = 0.
2µ
The strain invariants Ji are
J1 = εii = (177.16 + 7.56 − 118.07) × 10−6 = 66.65 × 10−6 ,
1 1
(177.16)2 + (7.56)2 + (−118.07)2 + 2(75.38)2 + 2(188.44)2

J2 = εij εij =
2 2
= 63, 883.2 × 10−12 ,
J3 = |ε| = 244, 236 × 10−18 .
Figure P6-10
6.10 Plane stress-reduced constitutive relations. Beginning with the strain-stress relations in
Eq. (6.3.23) for an orthotropic material in a two-dimensional case (that is, σ33 = σ13 =
σ23 = 0), determine the two-dimensional stress–strain relations.
x3
x2
σ33 = 0
x3 σ32 = 0
σ23 = 0
σ 31 = 0
x2
σ13 = 0 σ22
x1 σ12 σ21
σ11
x1
Fig. P6.10
Solution: From Eq. (6.3.23) we have

   1 − ν21 0
 
 ε1  E1 E2  σ1 
ε2 =  − νE121 E12 0 
 σ2 .

ε6 σ6
   
1
0 0 G12
Inverting the coefficient matrix, we obtain

−1  1 ν12 
1
− νE212 0 0

E1 G12 E2 E1 G12
 − ν12 1 G12 E1 E2  ν21 1

 E1 E2
0  = 
G12 E2 E1 G12
0 .
1 − ν12 ν21 
 
1 1−ν12 ν21
0 0 G12 0 0 E1 E2
Thus, the stress–strain relations are

    
 σ1  E1 ν12 E2 0  ε1 
1
σ2 =  ν21 E1 E2 0  ε2
1 − ν12 ν21
σ6 0 0 (1 − ν12 ν21 )G12 ε6
   
6.11 Given the strain energy potential

λ
Ψ(E) = (tr E)2 + µ tr(E · E),
2
determine the second Piola–Kirchhoff stress tensor S in terms of the Green strain tensor
E.
Solution: First we note that the following identities hold for any second order tensor,
R:
∂(tr R) ∂(tr R2 )
= I, = 2RT .
∂R ∂R
To prove the first identity, we begin with
∂RKK ∂(tr R)
= δIK δJK = δIJ → = I.
∂RIJ ∂R
Similarly,
∂ ∂
(RKM RM L )δKL = (RKM RM K ) = RKM δIM δKJ + δIK δJM RM K = 2RJI ,
∂RIJ ∂RIJ
or
∂(tr R2 )
= 2RT .
∂R
Then, it follows that
∂Ψ
S= = λ (tr E)I + 2µE.
∂E
6.12 Given the strain energy potential for the case of infinitesimal deformations
λ
Ψ(ε) = (tr ε)2 + µ tr(ε · ε),
2
determine the strain energy function Ψ(σ) in terms of the stress tensor σ.
Solution: From the given strain energy potential, we have
σ = 2µε + λ(tr ε)I.
Then
tr σ = (2µ + 3λ) tr ε,
and
1 λ
ε= σ− (tr σ)I .
2µ 2µ + 3λ
Hence, we have
3λ2

λ 2 1 2 2λ 2 2
Ψ(σ) = (tr σ) + tr(σ ) − (tr σ) + (tr σ)
2(2µ + 3λ)2 4µ (2µ + 3λ) (2µ + 3λ)2

1 λ 1
tr(σ 2 ) − (tr σ)2 = (1 + ν)tr(σ 2 ) − ν(tr σ)2 .

=
4µ (2µ + 3λ) 2E
6.13 Assuming that the strain energy density Ψ = U0 (σ) is positive-definite, that is, U0 ≥ 0,
with U0 = 0 if and only if σ = 0, determine the restrictions placed on the elastic
parameters E, K, and ν by considering the following stress states: (a) uniaxial stress
state with σ11 = σ; (b) pure shear stress state, σ12 = τ ; and (c) hydrostatic stress
state, σ11 = σ22 = σ33 = p.
Solution: First note that the strain energy density for an isotropic material is (see
Problem 6.12)
U0 = µεij εij + 12 λ(εkk )2
1
(1 + ν)σij σij − ν(σkk )2 .

=
2E
(a) For uniaxial stress state, we take σ11 = σ0 and σij = 0 for i = j 6= 1. Then
1 1 2
(1 + ν)σ02 − νσ02 =

U0 = σ0 .
2E 2E
Because Ψ > 0 for nonzero σ0 , we conclude that E > 0.
(b) For the pure shear case, we assume that σ12 = τ and set all other stresses to zero.
Then we have
1+ν 2 1 2
U0 = τ = τ > 0 → 1 + ν > 0 or ν > −1 and G > 0.
E 2G
(c) For hydrostatic stress, we take σ11 = σ22 = σ33 ≡ σh , and compute U0 as
1 3(1 − 2ν)
3(1 + ν)σh2 − 9νσh2 > 0 →

U0 = > 0 or ν < 0.5.
2E 2E
3(1−2ν)
Because 2E
= 1/K, it also follows that K > 0.
In summary, we have
E > 0, G > 0, K > 0, −1 < ν < 0.5.
Negative values of Poisson’s ratio correspond to materials that expand laterally under
uniaxial tension. This is not common with most materials.
6.14 A material is transversely isotropic at a point if it is symmetric with respect to an

arbitrary rotation about a given axis. Aligned fiber-reinforced composites provide ex-
amples of transversely isotropic materials (see Fig. P6.14). Take the x3 -axis as the axis
of symmetry with the transformation matrix
 
cos θ sin θ 0
[L] =  − sin θ cos θ 0  ,
0 0 1
where θ is arbitrary. Show that the stress–strain relations of a transversely isotropic
material are of the form
    

 σ1  C11 C12 C13 0 0 0  ε1 
 
σ  C12 C11 C13 0 0 0 ε2 
   


 2 

  


 
 
σ3 C C C 0 0 0 ε
   
13 13 33 3

=  .
Figure P6-14 σ4   0 0 0 C44 0 0 ε4 
  
   
 σ5 

   0 0 0 0 C44

0 
 ε5 



  
 
σ6 0 0 0 0 0 21 (C11 − C12 ) ε6
 
x3 = x3
x2
θ
x1 x2
θ
x1
Fig. P6.14
Solution: From the solution to Problem 6.2 we can establish, by changing θ to −θ in

the expressions for C̄ij , the following relations:
C16 = 0, C26 = 0, C11 + C22 − 4C66 = 0, C36 = 0,

C11 − C12 − 2C66 = 0, 2C66 + C12 − C22 = 0, C13 − C23 = 0,
C45 = 0, C55 − C44 = 0, C14 = 0, C24 − 2C56 = 0, C25 = 0,
C15 = 0, C15 − 2C46 = 0, C14 + 2C56 = 0, C25 + 2C46 = 0,
C24 = 0, C14 + 2C56 = 0, 2C56 − C24 = 0, C34 = 0, C35 = 0,
C46 = 0, C56 + C14 − C24 = 0, C15 − C25 − C46 = 0,
C56 = 0, C24 − C14 − C56 = 0.
Thus, we have
C14 = C15 = C16 = C24 = C25 = C26 = C34 = C35 = C36 = C45 = C46 = C56 = 0,
and
C22 = C11 , C23 = C13 , C55 = C44 , 2C66 = C11 − C12 .
6.15 The stress–strain relations of an isotropic material in the cylindrical coordinate system
are
σrr = 2µ εrr + λ (εrr + εθθ + εzz ) ,
σθθ = 2µ εθθ + λ (εrr + εθθ + εzz ) ,
σzz = 2µ εzz + λ (εrr + εθθ + εzz ) ,
σrθ = 2µ εrθ , σrz = 2µ εrz , σθz = 2µ εθz .
Express the relations in terms of the displacements (ur , uθ , uz ).
Solution: Using the strains-displacement relations from Eq. (3.5.26), we obtain

∂ur ∂ur ur 1 ∂uθ ∂uz
σrr = 2µ +λ + + + ,
∂r ∂r r r ∂θ ∂z

ur 1 ∂uθ ∂ur ur 1 ∂uθ ∂uz
σθθ = 2µ + +λ + + + ,
r r ∂θ ∂r r r ∂θ ∂z

∂uz ∂ur ur 1 ∂uθ ∂uz
σzz = 2µ +λ + + + ,
∂z ∂r r r ∂θ ∂z

1 ∂ur ∂uθ uθ
σrθ = µ + − ,
r ∂θ ∂r r

∂ur ∂uz
σrz = µ + ,
∂z ∂r

∂uθ 1 ∂uz
σθz = µ + .
∂z r ∂θ
6.16 Express the stress–strain relations of an isotropic material in the spherical coordinate
system and express the result in terms of the displacements (uR , uφ , uθ ).
Solution: The stress–strain relations are

σRR = 2µ εRR + λ (εRR + εθθ + εφφ ) ,
σφφ = 2µ εφφ + λ (εRR + εθθ + εφφ ) ,
σθθ = 2µ εθθ + λ (εRR + εθθ + εφφ ) ,
σRφ = 2µ εRφ , σRθ = 2µ εRθ , σφθ = 2µ εφθ .
Using the strain-displacement relations from Eq. (3.5.32), we obtain

∂uR ∂uR 1 ∂uφ 1 ∂uθ
σRR = 2µ +λ + + uR + + uR sin φ + uφ cos φ ,
∂R ∂R R ∂φ R sin φ ∂θ

1 ∂uφ ∂uR 1 ∂uφ
σφφ = 2µ + uR + λ + + uR
R ∂φ ∂R R ∂φ

1 ∂uθ
+ + uR sin φ + uφ cos φ ,
R sin φ ∂θ

1 ∂uθ ∂uR 1 ∂uφ
σθθ = 2µ + uR sin φ + uφ cos φ + λ + + uR
R sin φ ∂θ ∂R R ∂φ

1 ∂uθ
+ + uR sin φ + uφ cos φ ,
R sin φ ∂θ

1 ∂uR ∂uφ uφ
σRφ = µ + − ,
R ∂φ ∂R R

1 ∂uR ∂uθ uθ
σRθ = µ + − ,
R sin φ ∂θ ∂R R

1 1 ∂uφ ∂uθ
σφθ = µ + − uθ cot φ .
R sin φ ∂θ ∂φ
6.17 Given the displacement field in an isotropic body
ur = U (r), uθ = 0, uz = 0, (1)
where U (r) is a function of only r, determine the stress components in the cylindrical
coordinate system.
Solution: The only nonzero strains associated with the displacement field (1) are [see
Eq. (3.5.26)]
dU U
εrr = , εθθ = . (2)
dr r
The nonzero stresses are
dU U
σrr = 2µεrr + λ (εrr + εθθ ) = (2µ + λ) +λ
dr r
U dU
σθθ = 2µεθθ + λ (εrr + εθθ ) = (2µ + λ) + λ (3)
r dr
dU U
σzz = 2µεzz + λ (εrr + εθθ ) = λ + .
dr r
6.18 Given the displacement field in an isotropic body
uR = U (R), uφ = 0, uθ = 0, (1)
where U (R) is a function of only R, determine the stress components in the spherical
coordinate system.
Solution: The only nonzero strains associated with the displacement field (1) are [see
Eq. (3.5.32)]
dU 1
εRR = , εφφ = εθθ = U (R). (2)
dR R
The nonzero stresses are
dU U
σRR = 2µεRR + λ (εRR + εφφ + εθθ ) = (2µ + λ) + 2λ ,
dR R
U dU (3)
σφφ = 2µεφφ + λ (εRR + εφφ + εθθ ) = 2(µ + λ) + λ ,
R dR
σθθ = σφφ .
6.19 The Navier equations. Show that for an isotropic, incompressible solid with infinitesimal
deformations (that is, σ ≈ S and F · S ≈ S), the equation of motion (5.3.11), ∇ · σ +
ρ0 f = ρ0 ü, can be expressed as
∂2u
ρ0 = ρ0 f − ∇p + (λ + µ) ∇(∇ · u) + µ∇2 u.
∂t2
Solution: For an incompressible solid, The stress tensor σ has the form [see Eq. (6.2.16)]
σ = −pI + σ̃ with σ̃ given by
h i
σ̃ = 2µε + λtr (ε)I, ε = 12 ∇u + (∇u)T .
Using the equation of motion, we can write
∂2u
ρ0 = ∇ · σ T + ρ0 f
∂t2
.
= ∇ · (−pI + σ̃) + ρ0 f
= −∇p + ∇ · [2µε + λtr (ε)I] + ρ0 f
Noting that tr (ε) = ∇ · u and

h i
∇ · 21 ∇u + (∇u)T = 1
2
2
∇ u + ∇(∇ · u) ,
we arrive at the required result
∂2u
= −∇p + µ∇2 u + (µ + λ)∇(∇ · u) + ρ0 f .

ρ0
∂t2
6.20 Given the following motion of an isotropic continuum,
χ(X) = (X1 + kt2 X22 ) ê1 + (X2 + ktX2 ) ê2 + X3 ê3 ,
determine the components of the viscous stress tensor as a function of position and
time.
Solution: The inverse of the mapping is
kt2 x2
X1 = x1 − x22 , X2 = .
(1 + kt)2 1 + kt
The displacements are
u1 = kt2 X22 , u2 = ktX2 , u3 = 0.
Then the velocity components are

dx1 2tk dx2 k
v1 = = 2tkX22 = x22 , v2 = = kX2 = x2 , v3 = 0.
dt (1 + kt)2 dt (1 + kt)
The viscous stress tensor components are given by
∂v1
τ11 = 2µ = 0,
∂x1
∂v2 2µk
τ22 = 2µ = ,
∂x2 1 + kt

∂v1 ∂v2 4tk
τ12 =µ + =µ x2 .
∂x2 ∂x1 (1 + kt)2
6.21 Express the upper and lower convective derivatives of Eqs. (6.6.12) and (6.6.13) in
Cartesian component form.
Solution: We have
∇ ∂S
S= + v · ∇S − L · S − (L · S)T ,
∂t
∇ ∂Sij ∂Sij
S ij = + vk − Lik Skj − Ljk Ski .
∂t ∂xk
Similarly,
∆ ∂S
S= + v · ∇S + LT · S + ST · L
∂t
∆ ∂Sij ∂Sij
S ij = + vk + Lki Skj + Ski Lkj .
∂t ∂xk
6.22 Interpret the Laḿe constant µ by considering the flow field
v1 = f (x2 ), v2 = 0, v3 = 0,
where f is a known function of x1 .

Solution: For this case we have all Dij zero except D12
1 df
D12 = D21 = 2
.
dx2
Hence,
df
σ12 = σ21 = µ ,
dx2
and all other stresses are zero.Thus, µ is the proportionality constant relating the
shear stress to the velocity gradient, which is the definition of the coefficient of (shear)
viscosity of a fluid.
6.23 For viscous compressible flows (in spatial description), show that
1 Dρ
σ̃ − p = λ + 23 µ ,
ρ Dt
where σ̃ = −σii /3 is the mean stress and p is the thermodynamic pressure.
Solution: For a compressible fluid, we have
σii = (2µ + 3λ)Dii − 3p.
If σ̃ = −σii /3, we can write

2 2

σ̃ − p = − 3
µ + λ Dii = − 3
µ +λ ∇·v
1 Dρ
= 23 µ + λ ,
ρ Dt
where the continuity equation (5.2.27) is used in arriving at the last step.
6.24 The Navier–Stokes equations. Show that for a compressible fluid, the Cauchy equations
of motion (5.3.10) can be expressed as
Dv
ρ = ρf − ∇p + (λ + µ) ∇(∇ · v) + µ∇2 v.
Dt
Simplify the equation for (a) an incompressible fluid and (b) hydrostatic state of stress.
Solution: Using the result

h i
σ = −pI + µ ∇v + (∇v)T + λ(∇ · v)I,
we can write the equation of motion as

Dv
ρ = ∇ · σT + ρ f
Dt
= −∇p + µ [∇ · ∇v + ∇(∇ · v)] + λ∇(∇ · v) + ρ f ,
which, when simplified, yields the required result.
6.25 Show that for an incompressible fluid the equation of motion simplifies to
D
(ρv) = ρ f − ∇p + µ∇2 v.
Dt
Solution: First note that for an incompressible fluid, we have

Dρ
=0 ∇ · v = 0.
Dt
and h i
σ = −pI + µ ∇v + (∇v)T ,
Hence, we can write

D Dv Dρ Dv
(ρv) = ρ +v =ρ
Dt Dt Dt Dt
and the equation of motion in Eq. (5.3.10) takes the form
Dv
ρ = ∇ · σT + ρ f
Dt
= −∇p + µ [∇ · ∇v + ∇(∇ · v)] + ρ f
= ρ f − ∇p + µ∇ · ∇v
Therefore, it follows that
D
(ρv) = ρ f − ∇p + µ∇2 v.
Dt
6.26 Show that for the two-dimensional flow of an incompressible Newtonian fluid with
∇ × f = 0, where f is the body force vector (measured per unit volume), the vorticity
w [see Eq. (3.6.5)] satisfies the diffusion equation
Dw
ρ = µ∇2 w.
Dt
Solution: By definition we have w = 12 ∇ × v. For an incompressible fluid, ∇ρ = 0 and

∇ · v = 0. Then

Dw 1 D 1 Dv
ρ = ρ (∇ × v) = ∇ × ρ
Dt 2 Dt 2 Dt
1 1 n h i o
= ∇ × (∇ · σ + f ) = ∇ × ∇ · µ (∇v) + µ (∇v)T − pI + f
2 2
1 2 2
= ∇ × µ∇ v = µ∇ w ,
2
where we have used the equations of motion, the constitutive relation for an incom-
pressible, isotropic, Newtonian fluid, and the vector identity ∇ × (∇p) = 0 in arriving
at the final step:
Dv h i
ρ = ∇ · σ + f , σ = 2µD − pI = µ (∇v) + (∇v)T − pI.
Dt
6.27 Stokesian fluid. A Stokesian fluid is one in which (a) the stress tensor σ is a continuous
function of the rate of deformation tensor D and the local thermodynamic state (that
is, may depend on temperature), but independent of other kinematic variables; (b) σ
is not an explicit function of position x, (c) constitutive behavior is isotropic; and (d)
the stress is hydrostatic when the rate of deformation is zero, D = 0. Consider the
following constitutive equation for a Stokesian fluid:
σ = −p I + µD + β D · D (σij = −p δij + µDij + β Dik Dkj ).
Write the equation of motion (5.3.10) in terms of p and D for a Stokesian fluid. Note
that a linear Stokesian fluid is a Newtonian fluid.
Solution: We have
Dv
ρ = ∇ · σ + ρf
Dt
= −∇p + ∇µ · D + µ∇ · DT + ∇β · (D · D)
+ β [(∇ · D) · D + D : ∇D] ,
or

Dvi ∂p ∂µ ∂Dij ∂β ∂Dik ∂Dkj
ρ =− + Dij + µ + Dik Dkj + β Dkj + Dik ,
Dt ∂xi ∂xj ∂xj ∂xj ∂xj ∂xj
6.28 Irrotational motion. The velocity field v is said to be irrotational when the vorticity
is zero, w = 0. Then there exists a velocity potential φ(x, t) such that v = ∇φ. Show
that the Navier–Stokes equations of Problem 6.24 can be expressed in the form

∂φ
ρ∇ + 2 (∇φ) = ρ f − ∇p + (λ + 2µ)∇(∇2 φ).
1 2
∂t
Solution: We have
D(∇)φ
= ρ f − ∇p + µ ∇2 (∇φ) + ∇(∇2 φ) + λ∇(∇2 φ)

ρ
Dt
= ρ f − ∇p + (µ + λ)∇(∇2 φ) + µ∇2 (∇φ)
= ρ f − ∇p + (2µ + λ)∇(∇2 φ)
Also, note that

D(∇φ) ∂(∇φ)
= + v · ∇(∇φ)
Dt ∂t

∂(∇φ) ∂φ
= + (∇φ) · ∇(∇φ) = ∇ + 12 (∇φ)2
∂t ∂t
Thus, we have

∂φ
ρ∇ + 12 (∇φ)2 = ρ f − ∇p + (λ + 2µ)∇(∇2 φ).
∂t
6.29 Show that in the case of irrotational body force f = −∇V and when p is a function
only of ρ
∂φ 1
+ 12 (∇φ)2 + V + P (ρ) − (λ + 2µ)∇2 φ = g(t),
∂t ρ
Rp
where P (ρ) = p0 dp/ρ, p0 is a constant, and g(t) is a function of time only.
Solution: From the solution to Problem 6.28, we have

∂φ 1 1
∇ + 12 (∇φ)2 = −∇V − ∇p + (λ + 2µ)∇(∇2 φ),
∂t ρ ρ
and integrating we obtain (note that ρ is a constant)

∂φ 1
+ 21 (∇φ)2 + V + P − (λ + 2µ)(∇2 φ) = g(t),
∂t ρ
where Z p
1 1
P = dp, ∇P = ∇p.
p0 ρ ρ
Here we used the fact that
Z p Z p
1 1 1
∇ dp = ∇( )dp + ∇p,
p0 ρ p0 ρ ρ
where the first term on the right side of the equality is zero because ρ is constant.
6.30 Show that for an isotropic Newtonian fluid the energy equation can be expressed in the
form
De
ρ = ∇ · (k∇θ) − p J1 + (λ + 2µ)J12 − 4µJ2 + ρ r,
Dt
where J1 and J2 are the principal invariants of D [see Eq. (3.4.36)], and k is the
conductivity.
Solution: First note that
4µJ2 = 4µ 12 (Dii Djj − Dij Dij )
or
0 = 4µ 12 (Dii Djj − Dij Dij ) − 4µJ2 .
Then
σ : D = −pD + 2µD : D + λ(∇ · v)2
or
σij Dij = −pDii + 2µDij Dij + λDii Djj + 4µ 12 (Dii Djj − Dij Dij ) − 4µJ2
= −pJ1 + (λ + 2µ)J12 − 4µJ2 .
Thus, the energy equation (5.4.11) becomes

De
ρ = σ : D − ∇ · q + ρr
Dt
= −pJ1 + (λ + 2µ)J12 − 4µJ2 + ∇ · (k∇θ) + ρ r,
as required.
6.31 Show that for an isotropic Newtonian fluid the energy equation can be expressed in the
form
Dη
ρθ = ∇ · (k∇θ) + (λ + 2µ)J12 − 4µJ2 + ρ r,
Dt
where θ is the absolute temperature, η is the entropy, J1 and J2 are the principal
invariants of D, and k is the conductivity. Hint: θ dη = de+p d(1/ρ) and d/dt = D/Dt.
Solution: Consider the energy equation in the form

De
ρ = σ : D − ∇ · q + ρr
Dt
= −p(∇ · v) + ∇ · (k∇θ) + ρ r + D∗ ,
where
D∗ = (λ + 2µ)J12 − 4µJ2 .
We can write
De
+ p(∇ · v) = ∇ · (k∇θ) + ρ r + D∗ ,
ρ
Dt
and using the continuity equation p(∇ · v) = −(p/ρ)(Dρ/Dt), we can write
De p Dρ
ρ − = ∇ · (k∇θ) + ρ r + D∗ .
Dt ρ Dt
Now, if η is the entropy density (measured per unit mass),
p
θ dη = de + p d(1/ρ) = de − dρ,
ρ2
so that
Dη De p Dρ
ρθ =ρ − = ∇ · (k∇θ) + ρ r + D∗ ,
Dt Dt ρ Dt
or
Dη
ρθ = ∇ · (k∇θ) + (λ + 2µ)J12 − 4µJ2 + ρ r.
Dt
6.32 The thermal stress coefficients, βij , measure the increases in the stress components per
unit decrease in temperature with no change in the strain, that is,
∂σij

βij = − .
∂θ ε=const

Deduce from the above equation the result

∂η
βij = ρ .
∂εij
Solution: The use of the entropy density η as an independent state variable is not
convenient. A far more convenient thermal variable is the temperature θ, because it
is fairly easy to measure and to control. If the Helmholtz free energy per unit mass is
defined as Ψ = e − ηθ = Ψ(θ, ε), then
1
Ψ̇ = −η θ̇ + σij ε̇ij
ρ
so that η = −(∂Ψ/∂θ) and σij = ρ(∂Ψ/∂εij ). Now we have
∂σij ∂2Ψ ∂η

βij = − =− =ρ .
∂θ ε=const ∂θ∂εij ∂εij

6.33 The specific heat at constant strain is defined by

∂e

cv = .
∂θ ε=const

Deduce from the above equation the result
∂2Ψ
cv = −θ .
∂θ2
Solution: We have
∂e ∂ ∂Ψ ∂η
cv = = (Ψ + ηθ) = +θ + η.
∂θ ∂θ ∂θ ∂θ
Because (∂Ψ/∂θ) = −η, we obtain
∂e ∂η ∂2Ψ
c= =θ = −θ 2 .
∂θ ∂θ ∂θ
6.34 Consider a reference state “0” at zero strain and temperature θ0 , and expand Ψ(θ, ε)
in Taylor’s series about this state up to quadratic terms in θ and εij to derive the
constitutive equations, Eq. (6.8.20), for linear thermoelasticity. Specialize the relations
to the isotropic case.
Solution: Expanding Ψ about the “0” state, we have
∂Ψ ∂Ψ ∂2Ψ

2
Ψ(θ, ε) = Ψ0 + (θ − θ0 ) + (εij − 0) + 12 (θ − θ0 )
∂θ 0 ∂εij 0 ∂θ2 0
∂2Ψ ∂2Ψ

+ 21 (θ − θ0 )(εij − 0) + 12 (εij − 0)(εk` − 0) + · · · .
∂θ∂εij 0 ∂εij ∂εk` 0
Now we let
∂Ψ

= −η0 entropy at the reference state,
∂θ 0

∂Ψ 1 0

= σij stress at the reference state or initial stress,
∂εij 0 ρ0

∂2Ψ c0 ∂2Ψ 1 0

2
=− v , = − βij ,
∂θ 0 θ0 ∂θ∂εij 0 ρ0

∂2Ψ ∂σij

0 0
= = Cijk` = Ckìj .
∂εij ∂εk` 0 ∂εk`

Thus, the constitutions of linear thermoelasticity are
c0v 1 0
η = η0 + (θ − θ0 ) + βij εij ,
θ0 ρ0
0 0 0
σij = σij − βij (θ − θ0 ) + Cijk` εk` .
N6.1 Beginning with Eq. (6.2.41), establish Eq. (6.2.43).
Solution: We begin with ψ = ψ(B) and Eq. (6.2.41)

T " T #
∂ψ ∂ψ ∂B ∂ψ ∂ψ
ρ =ρ · =ρ ·F+ ·F
∂F ∂B ∂F ∂B ∂B
Dotting both sides with FT , we obtain

" T #
∂ψ ∂ψ ∂ψ
ρ · FT = ρ · F · FT + ρ · F · FT
∂F ∂B ∂B
∂ψ ∂ψ
= 2ρ · F · FT = 2ρ · B.
∂B ∂B

Chapter 7: LINEARIZED ELASTICITY
7.1 Define the deviatoric components of stress and strain as follows:
sij ≡ σij − 13 σkk δij , eij ≡ εij − 13 εkk δij .
Determine the constitutive relation between sij and eij for an isotropic material.
Solution: First we note that
sii = σii − 31 σkk δii = σii − σkk = 0; similarly, we have eii = 0.
Then
sij = σij − 31 σkk δij = 2µεij + λεkk δij − 13 (2µ + 3λ)εkk δij
= 2µεij + 32 µεkk δij = 2µ eij + 13 εkk δij − 23 µεkk δij

= 2µeij .
Thus, the constitutive equations between the deviatoric components of stress and strain
for an isotropic material is sij = 2µeij , and it only depends on the shear modulus,
µ = G.
Figure P7-2
7.2 For each of the displacement fields given below, sketch the displaced positions in the
x1 x2 -plane of the points initially on the sides of the square shown in Fig. P7.2.
α α
(a) u = 2
(x2 ê1 + x1 ê2 ) . (b) u = 2
(−x2 ê1 + x1 ê2 ) . (c) u = α x1 ê2 .
x2
(0,1) (1,1)
(0,0)
(1,0)
x1
Fig. P7.2
Figure P7-3
Solution: The sketches of the deformed configurations for the three cases are shown in
Fig. P7.2(a)-(c).
x2 x2 x2
0.5a 0.5a
a
(0,1) (1,1) (0,1) (1,1) (0,1) (1,1)
0.5a 0.5a a
(0,0) x1 (0,0)
(1,0)
x1 (0,0) x1
(1,0) (1,0)
(a) (b) (c)
7.3 For each of the displacement fields in Problem 7.2, determine the components of (a)
the Green–Lagrange strain tensor E, (b) the infinitesimal strain tensor ε, (c) the in-
finitesimal rotation tensor Ω, and (d) the infinitesimal rotation vector ω (see Sections
3.4 and 3.5 for the definitions).
CHAPTER 7: LINEARIZED ELASTICITY 151
Solution: For the first displacement field, we have

u1 = 12 αx2 , u2 = 12 αx1 , u3 = 0.
(a) The Green–Lagrange strain tensor components are

α 2 α 2 α α
E11 = 21 , E22 = 21 , 2E12 = 1
2
+ 1
2
= α.
2 2 2 2
(b) The infinitesimal strains are
ε11 = 0, ε22 = 0, 2ε12 = α.
(c) The rotation components are

Ω12 = 0, Ω13 = 0, Ω23 = 0 → ω = 0.
For the second displacement field, we have

u1 = − 12 αx2 , u2 = 12 αx1 , u3 = 0.

α 2 α 2
E11 = 21 , E22 = 21 , 2E12 = 0.
2 2
ε11 = 0, ε22 = 0, 2ε12 = 0.

2Ω12 = α, Ω13 = 0, Ω23 = 0 → ω3 = α.
For the third displacement field, we have

u1 = 0, u2 = αx1 , u3 = 0.

E11 = 12 α2 , E22 = 0, 2E12 = α.

ε11 = 0, ε22 = 0, 2ε12 = 0.

2Ω12 = α, Ω13 = 0, Ω23 = 0 → ω3 = α.
7.4 Similar to Cauchy’s formula for a stress tensor, one can think of similar formula for the
strain tensor,
εn = ε · n̂,
where εn represents the strain vector in the direction of the unit normal vector,
n̂. Determine the longitudinal strain corresponding to the displacement field u =
α
2
(x2 ê1 + x1 ê2 ) in the direction of the vector ê1 + ê2 .
Solution: From the solution of Problem 7.3, we have

" #
1 0 α
[ε] = .
2 α 0
Then we have
α
εn = ε · n̂ = √ (ê1 + ê2 ).
2 2
7.5 For the displacement vector given in the cylindrical coordinate system
u = Ar êr + Brz êθ + C sin θ êz ,
where A, B, and C are constants, determine the infinitesimal strain components in the
cylindrical coordinate system.
Solution: We have
ur = Ar, uθ = Brz, uz = C sin θ.
Then using the strain-displacement relations in Eq. (3.5.26), we obtain
∂ur
εrr = = A,
∂r
1 1 ∂ur ∂uθ uθ
εrθ = + − = 21 (0 + Bz − Bz) = 0,
2 r ∂θ ∂r r

1 ∂ur ∂uz
εrz = + = 0,
2 ∂z ∂r
ur 1 ∂uθ
εθθ = + = A,
r r ∂θ
1 ∂uθ 1 ∂uz 1 1
εzθ = + = Br + C cos θ ,
2 ∂z r ∂θ 2 r
∂uz
εzz = = 0.
∂z
7.6 The displacement vector at a point referred to the basis (ê1 , ê2 , ê3 ) is u = 2ê1 +2ê2 −4ê3 .
Determine ūi with ˆ ˆ ˆ ˆ
√ respect to the basis (ē1 , ē2 , ē3 ), where ē1 = (2ê1 + 2ê2 + ê3 )/3 and
ˆ2 = (ê1 − ê2 )/ 2.
ē
ˆ3 = ē
Solution: First we compute the ē ˆ1 × ē
ˆ2 . We obtain

ê1 ê2 ê3

ˆ3 = 32
2 1
1
ē 3 3 = √ (ê1 + ê2 − 4ê3 ) .

1 3 2
√2 − √12 0

Hence, the transformation matrix is

 2 2 1

3 3 3
√1 − √12
 
[L] = 
 2
0 
.
1 1 4
√
3 2
√
3 2
− 3√ 2
The components of the transformed vector are

 2 2 1

   4 
 2   3 
3 3 3
√1 − √12

{ū} = [L]{u} = 
 2
0  0
  2  =  20 .
−4 √

1 1 4
√
3 2
√
3 2
− 3√ 2
3 2
7.7 Express Navier’s equations of elasticity (7.2.17) in the cylindrical coordinates.
Solution: The vector form of the Navier’s equations of elasticity is
∂2u
µ∇2 u + (µ + λ)∇ (∇ · u) + ρ0 f = ρ0 . (1)
∂t2
From Table 2.4.2 we have

∂ 1 ∂ ∂
∇ = êr + êθ + êz ,
∂r r ∂θ ∂z
2
1 ∂
h ∂ 1 ∂ ∂2 i
∇2 u = r + + r (ur êr + uθ êθ + uz êz )
r ∂r ∂r r ∂θ2 ∂z 2
1 ∂ ∂ur 1 ∂ ∂uθ 1 ∂ ∂uz
= r êr + r êθ + r êz
r ∂r ∂r r ∂r ∂r r ∂r ∂r
1 ∂ 2 ur 1 ∂ 2 uθ 1 ∂ 2 uz
+ 2 2
êr + 2 2
êθ + 2 êz
r ∂θ r ∂θ r ∂θ2
1 ∂ur 1 ∂uθ
+ 2 2 êθ − ur êr − 2 2 êr + uθ êθ
r ∂θ r ∂θ
∂ 2 ur ∂ 2 uθ ∂ 2 uz
+ 2 r
ê + êθ + êz
∂z ∂z 2 ∂z2
1 ∂(rur ) ∂uθ ∂uz
∇·u= + +r
r ∂r ∂θ ∂z
∂ n 1 h ∂(rur ) ∂uθ ∂uz io 1 ∂ n 1 h ∂(rur ) ∂uθ ∂uz io
∇(∇ · u) = êr + +r + êθ + +r
∂r r ∂r ∂θ ∂z r ∂θ r ∂r ∂θ ∂z
∂ n 1 h ∂(rur ) ∂uθ ∂uz io
+ êz + +r
∂z r ∂r ∂θ ∂z
Substituting all these expressions into Eq. (1) and collecting the coefficients of êr , êθ ,
and êz separately, we obtain
1 ∂ 2 ur ∂ 2 ur

1 ∂ ∂ur ur 2 ∂uθ
µ r + 2 − − +
r ∂r ∂r r ∂θ2 r2 r2 ∂θ ∂z 2
∂ 2 ur
h
∂ 1 ∂(rur ) ∂uθ ∂uz i
+ (λ + µ) + +r + ρ0 f r = ρ0 ,
∂r r ∂r ∂θ ∂z ∂t2
1 ∂ 2 uθ ∂ 2 uθ

1 ∂ ∂uθ 2 ∂ur 1
µ r + 2 2
+ 2 − 2 uθ +
r ∂r ∂r r ∂θ r ∂θ r ∂z 2
∂ 2 uθ
h
1 ∂ 1 ∂(rur ) ∂uθ ∂uz i
+ (λ + µ) + +r + ρ0 f θ = ρ0 ,
r ∂θ r ∂r ∂θ ∂z ∂t2
1 ∂ 2 uz ∂ 2 uz

1 ∂ ∂uz
µ r + 2 2
+
r ∂r ∂r r ∂θ ∂z 2
∂ 2 uz
h
∂ 1 ∂(rur ) ∂uθ ∂uz i
+ (λ + µ) + +r + ρ0 f z = ρ0 .
∂z r ∂r ∂θ ∂z ∂t2
7.8 An isotropic body (E = 210 GPa and ν = 0.3) with two-dimensional state of stress
experiences the following displacement field (in mm):
u1 = 3x21 − x31 x2 + 2x32 , u2 = x31 + 2x1 x2 ,
where xi are in meters. Determine the stresses and rotation of the body at point
(x1 , x2 ) = (0.05, 0.02) m.
Solution: The linearized strains are

ε11 = 6x1 − 3x21 x2 , ε22 = 2x1 , 2ε12 = −x31 + 6x22 + 3x21 + 2x2 .
At point (x1 , x2 ) = (0.05, 0.02) they have the values
ε11 = 0.3 − 0.00015 = 0.29985 × 10−3 m/m, ε22 = 0.1 × 10−3 m/m,
2ε12 = −x31 + 6x22 + 3x21 + 2x2 = −0.000125 + 0.0024 + 0.0075 + 0.04
= 0.049775 × 10−3 m/m.
The values of µ and λ are
E 210 νE 63
µ=G= = = 80.769 GPa, λ= = = 121.15 GPa,
2(1 + ν) 2.6 (1 + ν)(1 − 2ν) (1.3)(0.4)
σ11 = 2µε11 + λεkk = 96.88 MPa,

σ22 = 2µε22 + λεkk = 64.597 MPa,
σ33 = 2µε33 + λεkk = 48.443 MPa,
σ12 = 2µε12 = 4.02 MPa, σ13 = 2µε13 = 0 MPa, σ23 = 0,
When displacements are given, there is no question of compatibility.
7.9 A two-dimensional state of stress exists in a body with the following components of
stress:
σ11 = c1 x32 + c2 x21 x2 − c3 x1 , σ22 = c4 x32 − c5 , σ12 = c6 x1 x22 + c7 x21 x2 − c8 ,
where ci are constants. Assuming that the body forces are zero, determine the con-
ditions on the constants so that the stress field is in equilibrium and satisfies the
compatibility equations.
Solution: Check equilibrium equations

∂σ11 ∂σ12
+ = 2c2 x1 x2 − c3 + 2c6 x1 x2 + c7 x21 = 0
∂x1 ∂x2
∂σ12 ∂σ22
+ = c6 x22 + 2c7 x1 x2 + 3c4 x22 = 0,
∂x1 ∂x2
which imply
c2 + c6 = 0, c3 = 0, c7 = 0, c6 + 3c4 = 0 → c6 = −c2 = −3c4 ,
and c1 and c5 are arbitrary.
7.10 Express the strain energy for a linear isotropic body in terms of the (a) strain compo-
nents and (b) stress components.
Solution: (a) The strain energy of an isotropic elastic body occupying the volume Ω
takes the form
Z Z
1 1
U= σij εij dx = (2µεij + λεkk δij ) εij dx
2 Ω 2 Ω
Z
1
= (2µ εij εij + λεii εjj ) dx.
2 Ω
(b) In terms of the stresses, we have

Z Z
1 1 1
U= σij εij dx = σij [(1 + ν)σij − νσkk δij ) dx
2 Ω 2 Ω E
Z
1 1
Figure P7-4 =
2 ΩE
[(1 + ν)σij σij − νσii σjj ] dx.
7.11 A rigid uniform member ABC of length L, pinned at A and supported by linear elastic
springs, each of stiffness k, at B and C, is shown in Fig. P7.10. Find the total strain
energy of the system when the point C is displaced vertically by the amount uC .
Rigid bar
A B k C k
L L
2 2
Fig. P7.10
Solution: The displacement at B is uc /2. The strain energy is

u 2
c
U = 12 k + 12 ku2c = 58 ku2c .
2
7.12 Repeat Problem 7.11 when the springs are nonlinearly elastic, with the force deflection
relationship, F = ku2 , where k is a constant.
Solution: When the sprngs are nonlinear, the strain energy density can be calculated
as Z u Z u
U0 = F du = ku2 du = 13 ku3 .
0 0
Hence the strain energy is
u 3
c
U = 13 k + 13 ku3c = 38 ku3c .
2
7.13 Consider the equations of motion of 2-D elasticity (in the x- and z-coordinates) in the
absence of body forces:
∂σxx ∂σxz ∂ 2 ux
+ = ρ0
∂x ∂z ∂t2
∂σxz ∂σzz ∂ 2 uz
+ = ρ0
∂x ∂z ∂t2
For a beam of uniform height h and width b, integrate the preceding equations with
respect to z from −h/2 to h/2, and express the results in terms of the stress resultants
N and V defined in Eq. (7.3.28). Use the following boundary conditions:

h h h h
b σxz (x, ) − σxz (x, − ) = f (x), σxz (x, ) + σxz (x, − ) = 0,
2 2 2 2
h h
σzz (x, y, − ) = 0, bσzz (x, y, ) = q
2 2
Next, multiply the first equation of motion with z and integrate it with respect to z
from −h/2 to h/2, and express the results in terms of the stress resultants M and V
defined in Eq. (7.3.28).
Solution: Multiplying the equations with b, integrating the equations of motion over
(− h2 , h2 ), and using the boundary conditions on the transverse stresses, we obtain
h
∂ 2 ux
Z
2 ∂σxx ∂σxz
0=b + −ρ 2 dz
−h ∂x ∂z ∂t
2
∂N ∂2u ∂3w
= + b [σxz (x, h/2) − σxz (x, −h/2)] − I0 2 + I1 2
∂x ∂t ∂t ∂x
∂N ∂2u ∂3w
= + f (x) − I0 2 + I1 2
∂x ∂t ∂t ∂x
Z h
∂ 2 uz

2 ∂σxz ∂σzz
0=b + − ρ 2 dz
−h ∂x ∂z ∂t
2
∂V ∂2w
= + b [σzz (x, h/2) − σzz (x, −h/2)] − I0 2
∂x ∂t
∂V ∂2w
= + q(x) − I0 2 ,
∂x ∂t
Next, multiplying the first equation of motion with b z, integrating over (− h2 , h2 ), and
using the boundary conditions on the transverse stresses, we obtain
Z h
∂ 2 ux

2 ∂σxx ∂σxz
0=b z + − ρ 2 dz
−h ∂x ∂z ∂t
2
∂M bh ∂2u ∂3w
= −V + [σxz (x, h/2) + σxz (x, −h/2)] − I1 2 + I2 2
∂x 2 ∂t ∂t ∂x
∂M ∂2u ∂3w
= − V − I1 2 + I2 2
∂x ∂t ∂t ∂x
Z h Z h
2 2
(N, V, M ) = b (σxx , σxz , zσxx ) dz, (I0 , I1 , I2 ) = b ρ(1, z, z 2 ) dz.
−h
2
−h
2
Figure
7.14 For the P7-12
plane elasticity problems shown in Figs. P7.14(a)-(d), write the boundary
conditions and classify them into type I, type II, or type III.
x2
τ p
θ Spherical
b
τ Core, μ1 , λ1
(a) τ (b)
a
Spherical
x1 shell, μ 2 , λ2
x2
Rigid core
τ0
τ0
(c) b (d)
b
Hollow
a
cylindrical
shaft μ , λ
τ0 τ0
x1
σ0 a
Fig. P7.14
Solution: (a) The boundary conditions are

u1 = u2 = 0 on x2 = 0; tn = 0, ts = τ on plane with normal n̂ = (cos θ ê1 − sin θ ê2 ) ,
tn = 0, ts = τ on plane with normal n̂ = ê2 ,
tn = 0, ts = τ on plane with normal n̂ = − (sin θ ê1 + cos θ ê2 ) .
The boundary value problem is of type III.
(b) The boundary conditions are
tRR = −p, tRφ = 0, tRθ = 0 at R = b; uR = uφ = uθ = 0 at R = 0.
(c) The boundary conditions are
trr = 0, trθ = τ at r = b; ur = uθ = 0 at r = 0.
(d) The boundary conditions are
u1 = u2 = 0 at x2 = 0;
t1 = t2 = 0 at x1 = a;
x2
t1 = 0, t2 = 0 at x2 = b; t1 = −σ0 1 − , t2 = 0.
b
Figure P7-15a
7.15 Consider a cantilever beam of length L, constant bending stiffness EI, and with right
end (x = L) fixed, as shown in Fig. P7.15. If the left end (x = 0) is subjected to a
moment M0 , use Clapeyron’s theorem to determine the rotation (in the direction of
the moment) at x = 0.
y,v y,v
L
M0 M0
x M (x ) = -M 0
x
q0 = -(dv / dx )x =0
Fig. P7.15
Solution: By Clapeyron’s Theorem we have

1
U= M0 θ0 ,
2
where the strain energy is given by
Z L Z L
1 1 M 2L
U= M 2 dx = (−M0 )2 dx = 0 .
2EI 0 2EI 0 2EI
Thus, we have
M0 L
. θ0 =
EI
One can verify the above result by solving the equation
d2 v dv
M (x) = −EI → −EI = −M0 x + A,
dx2 dx
where the constant integration A is determined from the condition that dv/dx = 0 at
Figure P7-16
x = L. We obtain A = M0 L. Thus,
dv M0 (L − x) M0 L
θ(x) ≡ − = → θ(0) = .
dx EI EI
7.16 Consider a cantilever beam of length L, constant bending stiffness EI, and with the
right end fixed, as shown in Fig. P7.16. If a point load F0 is applied at a distance a
from the free end, determine the deflection v(a) using Clapeyron’s theorem.
y, v F0
a
A
x
L
Fig. P7.16
Solution: We have
L L
F02
Z Z
1 1
1
F v(a)
2 0
=U = M 2 (x) dx = [−F0 (x − a)]2 dx = (L − a)3 ,
2EI 0 2EI a 6EI
or
F0
v(a) = (L − a)3 .
3EI
Figure P7-17a
7.17 Determine the deflection at the midspan of a cantilever beam subjected to uniformly
distributed load q0 throughout the span and a point load F0 at the free end, as shown
in Fig. P7.17. Use Maxwell’s theorem and superposition.
y,v y,v Fc q0
q0 Fc
F0 0.5L a b
A A
x x
L L
EI = constant
(a) (b)
Fig. P7.17
Solution: First we determine the deflection of a beam of length L with a point load
Fc at the center of the beam. We consider the general case of a point load Fc applied
at a distance a from the free end, as shown in Fig. P7.17(b). Using the bending
moment-deflection relationship, we can write
d2 v1 d2 v2
EI = 0, 0 ≤ x ≤ a; EI = −Fc (x − a), a ≤ x ≤ L. (1)
dx2 dx2
Integration of these equations gives the solutions
Fc
−EIv1 (x) = Ax+B, 0 < x ≤ a; −EIv2 (x) = − (x−a)3 +Cx+D, 0 < x ≤ a, (2)
6
and the constants of integration are determined as follows: (dv2 /dx) = 0 at x = L gives
C = Fc b2 /2, where b = L − a. Then v2 (L) = 0 gives D = −Fc b2 (3a + 2b)/6. Then the
two conditions of continuity, v1 (a) = v2 (a) and dv1 /dx = dv2 /dx at x = a give A = C
and B = −Fc b2 a/2 − Fc b3 /3. Thus the solution is
Fc b2 Fc
(x − a)3 − 3b2 (x − a) + 2b3

v1 (x) = − [3(x − a) − 2b] , v2 (x) = (3)
6EI 6EI
Then Maxwell’s reciprocity theorem says that the work done by q0 and F0 in moving
through the displacement produced by force Fc at the center of the beam is equal to
the work done by Fc in moving through the displacements produced by both q0 and F0
at x = 0. Let v(x) be the displacement produced by Fc ,
v1 (x), 0 ≤ x ≤ a,

v(x) = (4)
v2 (x), a ≤ x ≤ L.
Then by Maxwell’s Theorem, we have

Z a Z L
Fc v(L/2) = F0 v1 (0) + q0 v1 (x) dx + q0 v2 (x) dx
0 a
Fc b2 Fc b3 Fc a2 b2 Fc b3 a Fc b4

= F0 a+ + q0 + + q0 . (5)
2EI 3EI 4EI 3EI 8EI
Canceling Fc on both sides and simplifying the above expression with a = b = L/2, we
obtain the required result
5F0 L3 17q0 L4
vc = v(L/2) = + . (6)
48EI 384EI
Figure P7-18
7.18 Consider a simply supported beam of length L subjected to a concentrated load F0 at
the midspan and a bending moment M0 at the left end, as shown in Fig. P7.18. Verify
that Betti’s theorem holds.
y, v F0
L L
M0 2 B 2
A
x
L
Fig. P7.18
Solution: Let us label the left end point and the midpoint of the beam as A and B,
respectively. The deflection at point B (center) of the beam resulting from the moment
M0 applied at point A (left end) can be calculated from (RA = −M0 /L)
d2 v M0 x2 M0 x3
EI 2
= M0 + RA x → v(x) = − + Ax + B.
dx 2EI 6EIL
Using the boundary conditions v(0) = v(L) = 0, we obtain
M0 L
B = 0, A=− .
3EI
Thus, the deflection due to M0 is
M0 x2 M0 x3 M0 Lx M0 L2 M0 L2 M0 L2 M0 L2
v(x) = − − , v(L/2) = − − =− .
2EI 6EIL 3EI 8EI 48EI 6EI 16EI
M0 L2
vBA = w(L/2) = − .
16EI
We can determine the slope θ = −dv/dx due to F0 from
d2 v F0 dv F0 x2
EI 2
=− x → − = + C.
dx 2 dx 4EI
Because the slope is zero at the center, we have B = −F0 L2 /16EI. Thus, we have
dv F0 x2 F0 L2 F0 L2
θ(x) = − = − , θ(0) = θAB = − .
dx 4EI 16EI 16EI
The work done by F0 in moving through the displacement vBA is
F0 M0 L2
WBA = F0 vBA = − .
16EI
The work done by M0 in moving through the rotation θAB is
M0 F0 L2
WAB = M0 θAB = − .
16EI
Thus, WBA = WAB .
7.19 Use the reciprocity theorem to determine the deflection vc = v(0) at the center of a
simply supported circular plate under asymmetric loading (see Fig. P7.19):
r
q(r, θ) = q0 + q1 cos θ.
a
The deflection v(r) due to a point load Q0 at the center of a simply supported circular
plate is
Q0 a2 r2
r
3+ν r 2
v(r) = 1− 2 +2 log ,
16πD 1+ν a a a
where D = Eh3 /[12(1 − ν 2 )] and h is the plate thickness.
Figure P7-19
r
q(r,θ ) = q0 + q1 cosθ
a
y, v(r)
y, v (r) q0 + q1 q1
q0 − q1 q1
r q0 q0
a θ h
h
O r
r
a a
simply supported
Fig. P7.19
Solution: By Maxwell’s theorem, the work done by a point load (F0 ) at the center
of the simply supported plate due to the deflection vc at the center caused by the
distributed load q(r, θ) is equal to the work done by the distributed load q(r, θ) in
moving through the displacement v(r) caused by the point load at the center. Hence,
the center deflection of a simply supported plate under asymmetric load is
Z 2π Z a
a2 r2 r2

3+ν r
vc = q(r, θ) 1 − 2 + 2 2 log rdrdθ
16πD 0 0 1+ν a a a
q0 a2 a r2 r2
Z
3+ν r
= 1 − 2 + 2 2 log rdr
8D 0 1+ν a a a
Z a 3
q0 a2 3 + ν a2

r r
= +2 2
log dr
8D 1+ν 4 0 a a
a
q0 a2 3 + ν a2 2 r4 r4

r
= + 2 log −
8D 1+ν 4 a 4 a 16 0
2
2 2
4

q0 a 3+ν a a q0 a 5+ν
= − = .
8D 1+ν 4 8 64D 1 + ν
7.20 Use the reciprocity theorem to determine the center deflection vc = v(0) of a simply
supported circular plate under hydrostatic loading q(r) = q0 (1 − r/a).
Solution: In view of the solution of Problem 7.19, we need to calculate the deflection
at the center due to the load −q0 r/a. Thus, we have
Z 2π Z a
a2 r2 r2

3+ν r
vc = v(0) = q(r) 1 − 2 + 2 2 log rdrdθ
16πD 0 0 1+ν a a a
q0 a2 a r2 r2
Z
r 3+ν r
= 1− 1 − 2 + 2 2 log rdr
8D 0 a 1+ν a a a
q0 a4 5 + ν q0 a a r2 r2
Z
3+ν r
= − r 1 − 2 + 2 2 log rdr
64D 1 + ν 8D 0 1+ν a a a
a
q0 a4 5 + ν 3 + ν 2a3 2 r5 r5

q0 a r
= − − 2 log −
64D 1 + ν 8D 1+ν 15 a 5 a 25 0
4 4 4

q0 a 5+ν q0 a 3+ν 2 2 q0 a 5+ν 6+ν
= − − = − .
64D 1 + ν 8D 1 + ν 15 25 (1 + ν)D 64 150
7.21 Use the reciprocity theorem to determine the center deflection vc = v(0) of a clamped
circular plate under hydrostatic loading q(r) = q0 (1 − r/a). The deflection due to a
point load F0 at the center of a clamped circular plate is given in Eq. (7.4.21).
Solution: In view of the solution of Example 7.4.6, we need to calculate the deflection
at the center due to the load −q0 r/a. Thus, we have
Z 2π Z a
a2 r2

r
vc = q(r) 1 − 2 1 − 2 log rdrdθ
16πD 0 0 a a
a
q0 a2 r2
Z
r r
= 1− 1 − 2 1 − 2 log rdr
8D 0 a a a
Z a
q0 a4 r2

q0 a r
= − r 1 − 2 1 − 2 log rdr
64D 8D 0 a a
a
q0 a4 q0 a 2a3 2 r5 r5

r
= − + 2 log −
64D 8D 15 a 5 a 25 0
q0 a4 4 4

q0 a 2 2 43 q0 a
= − − = .
64D 8D 15 25 4800 D
7.22 Determine the center deflection vc = v(0) of a clamped circular plate subjected to a
point load F0 at a distance b from the center (and for some θ) using the reciprocity
theorem.
Solution: The Maxwell’s reciprocity theorem gives
F0 vbc = P vcb ,
where vbc is the deflection at a distance r = b due to point load P at the center (r = 0),
and vcb is the deflection at the center due to a point load F0 at r = b. The deflection
vbc at r = b of a clamped plate due to unit point load at the center is given by
a2 b2 b2

b
vbc = v(b) = 1 − 2 + 2 2 log
16πD a a a
Hence, we have
F0 a2 b2 b2 Q 0 b2 a2

b b
vcb = 1 − 2 + 2 2 log = 2 log + 2 − 1 .
16πD a a a 16πD a b
7.23 Rewrite Eq. (7.3.10) in a form suitable for direct integration and obtain the solution
2
given in Eq. (7.3.13). Hint: Note that dU 2U 1 d

dR
+ R
= 2
R dR
R U .
Solution: We begin with Eq. (7.3.11). First divide throughout by R and obtain
d2 U

2 dU 2U d dU U d 1 d 2
0= + − = + 2 = R U (1)
dR2 R dR R2 dR dR R dR R2 dR
Then, by successive integrations, we obtain
1 d d
R2 U = A, R2 U = AR2 ,

R2 dR dR
and
AR3 AR B c2
R2 U = + B, U (R) = + 2 = c1 R + 2 ,
3 3 R R
with c1 = A/3 and c2 = B.
7.24 Verify that the compatibility equation (3.7.4) takes the form
εαα,ββ − εαβ,αβ = 0 (α, β = 1, 2), (1)
or, in terms of stress components for the plane stress case,
∇2 σαα = −(1 + ν) fα,α . (2)
Solution: Adding the expression
∂ 2 ε11 ∂ 2 ε22
2
+
∂x1 ∂x22
to both sides of Eq. (3.7.4), we obtain

∂ 2 ε11 ∂ 2 ε22 ∂ 2 ε11 ∂ 2 ε22 ∂ 2 ε11 ∂ 2 ε22 ∂ 2 ε12
2
+ 2
+ 2
+ = + +2
∂x1 ∂x2 ∂x2 ∂x21 ∂x1 2
∂x2 2
∂x1 ∂x2
2
∂2 ∂ 2 ε11 ∂ 2 ε22 ∂ 2 ε12

∂
2
+ (ε11 + ε22 ) = + +2
∂x1 ∂x22 ∂x1 2
∂x2 2
∂x1 ∂x2
∂ 2 εαα ∂ 2 εαβ
= .
∂xβ ∂xβ ∂xα ∂xβ
Now using the strain-stress relations for the plane stress case,
1 1−ν
εαβ = [(1 + ν)σαβ − νσαα δαβ ] , εαα = σαα ,
E E
in Eq. (1), we obtain
1−ν 1+ν ν
σαα,ββ = σαβ,αβ − σαα,ββ ,
E E E
or
∇2 σαα = −(1 + ν) fα,α ,
where we have used the equilibrium equations σαβ,αβ = −fα α in arriving the final step.
7.25 Rewrite Eq. (7.5.17) in a form suitable for direct integration and obtain the solution
given in Eq. (7.5.21). Hint: Note that r1 dU d
− rU2 = dr U

dr r
.
Solution: Using the hint, we can write Eq. (7.5.17) as

d2 U

1 dU U d dU U d 1 d
−αr = + − = + = (rU ) .
dr2 r dr r2 dr dr r dr r dr
Thus, we have

d 1 d d α α A
(rU ) = −αr, (rU ) = − r3 + Ar, and rU (r) = − r4 + r2 + B.
dr r dr dr 2 8 2
Thus, we have the result in Eq. (7.5.21) with c1 = A/2 and c2 = B.
7.26 Show that the solution to the differential equation for G(r) in Eq. (6) is indeed given
by the first equation in Eq. (7) of Example 7.5.6. Hint: Note that (verify to yourself)
d2 G r2
Z
1 dG 1 d dG 1
+ = r ; r ln r dr = ln r − .
dr2 r dr r dr dr 2 2
Solution: Using the hint, we can write Eq. (6)1 of Example 7.5.7 as
d2 1 d d2 G 1 dG h 1 d d ih 1 d dG i
0= + + = r r .
dr2 r dr dr2 r dr r dr dr r dr dr
Hence, successive integrations yield
d h 1 d dG i 1 d dG d dG
r r = A, r = A ln r + B, r = Ar ln r + Br
dr r dr dr r dr dr dr dr
and
r2 r2

dG 1 dG r 1 r 1
r =A ln r − +B + C, =A ln r − +B +C .
dr 2 2 2 dr 2 2 2 r
Finally, we have
r2 i r2

1h 1
G(r) = A r2 ln r − − +B + C ln r + D = c5 + c6 ln r + c7 r2 + c8 r2 ln r,
2 2 4 4
with
3A B A
c5 = D, c6 = C , c7 = − + , c8 = .
8 4 2
7.27 Show that the solution to the differential equation for F (r) in Eq. (6) is indeed given by
the second equation in Eq. (7) of Example 7.5.7. Hint: Note that (verify to yourself)
d2 F

1 dF 4F 1 d 3 dF 2
+ − = r − 2r F .
dr2 r dr r2 r3 dr dr
Solution: Using the hint,
d2 F

1 dF 4F 1 d dF 1 d h 5 d F i
+ − 2 = 3 r3 − 2r2 F = r .
dr2 r dr r r dr dr r3 dr dr r2
Now we can write Eq. (6)2 of Example 7.5.6 as

1 d h 5 d 1 in 1 d h 5 d F io h 5 d 1 in 1 d h 5 d F io
0= 3 r r , r r = A.
r dr dr r2 r3 dr dr r2 dr r2 r3 dr dr r2
Then, integrating successively, we obtain
1 d h 5 d F i A 5 d
F A B
r = − + B, r = − r2 + r6 + C,
r5 dr dr r2 4r4 dr r2 8 6
and
F A B 2 C A Br4 C c1
= + r − + D, F (r) = + − 2 + Dr2 = 2 + c2 + c3 r2 + c4 r4 ,
r2 16r2 12 4r4 16 12 4r r
with
C A B
c1 = − , c2 = , c 3 = D , c4 = .
4 16 12
7.28 The only nonzero stress in a prismatic bar of length L, made of an isotropic material
(E and ν), is σ11 = −M0 x3 /I, where M0 is the bending moment and I is the moment
inertia about the x2 -axis, respectively. Determine the three-dimensional displacement
field. Eliminate the rigid body translations and rotations requiring that u = 0 and
Ω = 0 at x = 0.
Solution: The strains associated with the given stress field are
(1 + ν) ν σ11 M0 x3
ε11 = σ11 − σ11 = =− ,
E E E EI (1)
ν νM0 x3 ν νM0 x3
ε22 = − σ11 = , ε33 = − σ11 = .
E EI E EI
All other strains are zero.
We note that all strain compatibility conditions are trivially satisfied for this case,
because of two derivatives with respect to x3 or mixed derivatives with respect to x3
and xα . We now proceed to compute the displacement field.
Using the strain displacement relations, we have
∂u1 M 0 x3 ∂u2 νM0 x3 ∂u3 νM0 x3
=− , = , = . (2)
∂x1 EI ∂x2 EI ∂x3 EI
νM0 x3 νM0 x23

ε33 = , → u3 = + h(x1 , x2 ), (3)
EI 2EI
∂u1 ∂u3 ∂u1 ∂h
2ε13 = + = 0, → =− , (4)
∂x3 ∂x1 ∂x3 ∂x1
where h is a function to be determined. Integrating the above equation, we obtain
∂h
u1 = − x3 + g(x1 , x2 ), (5)
∂x1
where g is a function to be determined. Similarly,

∂u2 ∂u3 ∂u2 ∂h ∂h
2ε23 = + = 0, =− , u2 = − x3 + f (x1 , x2 ),
∂x3 ∂x2 ∂x3 ∂x2 ∂x2
where f is a function to be determined. Now comparing ε11 from Eq. (2) with that
computed from Eq. (5), we obtain
M0 ∂2h ∂g
− x3 = − 2 x3 + ,
EI ∂x1 ∂x1
we see that, because it must hold for any x3 ,
∂2h M0 ∂g
= , = 0 → g = G(x2 ). (6)
∂x21 EI ∂x1
Similarly, comparing ε22 from Eq. (2) with that computed from Eq. (6), we obtain
νM0 ∂2h ∂f
x3 = − 2 x3 + ,
EI ∂x2 ∂x2
we see that, because it must hold for any x3 ,
∂2h νM0 ∂f
=− , = 0 → f = F (x1 ). (7)
∂x22 EI ∂x2
From ε12 = 0, we see that

∂2h ∂g ∂f
−2 x3 + + = 0. (8)
∂x1 ∂x2 ∂x2 ∂x1
This gives the result
∂2h dG dF
= 0, + = 0 → G(x2 ) = c1 x2 + c2 , F (x1 ) = −c1 x1 + c3 . (9)
∂x1 ∂x2 dx2 dx1
Conditions in Eqs. (6)-(9) imply that h is of the form
M0
h(x1 , x2 ) = (x21 − νx22 ) + c4 x1 + c5 x2 + c6 , (10)
2EI
where ci are constants.
In summary, we have
∂h M0
u1 = − x3 + g(x1 , x2 ) = − x1 x3 − c4 x3 + c1 x2 + c2 ,
∂x1 EI
∂h νM0
u2 = − x3 + f (x1 , x2 ) = x2 x3 + c5 x3 − c1 x1 + c3 , (11)
∂x2 EI
M0 2
x1 + ν(x23 − x22 ) + c4 x1 + c5 x2 + c6 .

u3 =
2EI
The displacement boundary conditions give c2 = c3 = c6 , which correspond to the
translational rigid body motions. To remove the six rigid body rotations, we may
require
∂u1 ∂u2 ∂u2 ∂u3 ∂u1 ∂u3
= = = = = =0 at x1 = x2 = x3 = 0,
∂x2 ∂x1 ∂x3 ∂x2 ∂x3 ∂x1
which yield all other constants to be zero.
7.29 A solid circular cylindrical body of radius a and height h is placed between two rigid
plates, as shown in Fig. P7.29. The plate at B is held stationary and the plate at A is
subjected to a downward displacement of δ. Using a suitable coordinate system, write
the boundary conditions for the following two cases: (a) When the cylindrical object
is bonded to the plates at A and B. (b) When the plates at A and B are frictionless.
Figure P7-28
z
Rigid plate σ zθ σ zz
A
Lateral surface × σ zr
σ rz σ rθ
a
Cylinder
h ×
× σ rr
B
Rigid plate
r
Fig. P7.29
Solution: In general, in any elasticity, the boundary conditions involve specifying (that
is, we must know the value of) either the displacement u or the traction t at a point of
the boundary. More specifically, we must specify one element of each of the following
three pairs (for a 3D problem) at a boundary point: (u1 , t1 ), (u2 , t2 ), (u3 , t3 ), where
ti = nj σji , in a rectangular Cartesian system, and (ur , tr ), (uθ , tθ ), (uz , tz ) in a cylin-
drical coordinate system. Otherwise, the description of the boundary-value problem is
incomplete.
For the problem at hand we use the cylindrical coordinate system. We know that the
tractions tr , tθ and tz on the lateral surfaces of the cylindrical body are zero for both
parts of the problem:
σrr = 0, σrθ = 0, σrz = 0. (1)
For part (a) the top and bottom platens are in frictional contact with the cylindrical
member. Thus there will be no relative motion between the plates and the body. The
top is being pressed down by an amount δ. Thus, the boundary conditions are
At z = 0 : ur = uθ = uz = 0; At z = h : ur = uθ = 0, uz = −δ (2)
For part (b) the top is being pressed own by an amount δ, and the top and bottom
platens are in contact with the body without friction. Hence, the body is free to move in
radial as well as in circumferential (for non-axisymmetric case) as there is no resistance
offered (that is, tr and tθ are zero). Thus, the boundary conditions are
At z = 0 : σzr = σzθ = 0, uz = 0; At z = h : σzr = σzθ = 0, uz = −δ. (3)

Figure P7-27
7.30 The lateral surface of a homogeneous, isotropic, solid circular cylinder of radius a,
length L, and mass density ρ is bonded to a rigid surface. Assuming that the ends
of the cylinder at z = 0 and z = L are traction-free (see Fig. P7.30), determine the
displacement and stress fields in the cylinder due to its own weight.
z
y = x2
r
θ
x = x1
L
f = − ρ g eˆ z
Fig. P7.30
Solution: Using the cylindrical coordinate system shown in the figure above, the bound-
ary conditions of the problem can be stated as:
At z = 0, L : σzz = 0, σzr = 0, σzθ = 0,
(1)
At r = a : ur = 0, uθ = 0, uz = 0.
The body force component is fz = −ρg.
Neglecting the end effects, we can assume the following form of solution for the antiplane
strain problem:
ur = 0, uθ = 0, uz = U (r). (2)
The nonzero strains and stresses are given by
dU dU
2εrz = , σzr = µ . (3)
dr dr
Substitution into the third equilibrium equation yields (the other two equilibrium equa-
tions are identically satisfied)

dσzr 1 1 d dU ρg
+ σzr − ρg = 0 → r = . (4)
dr r r dr dr µ
The solution of the equation is given by
dU ρg r2 dU ρg r A
r = + A, or = + ,
dr µ 2 dr µ 2 r
ρg r2
U (r) = + A log r + B,
µ 4
where A and B are constants of integration that must be determined using the boundary
conditions. The boundary conditions associated with the antiplane strain problem are
that (a) U (r) is finite at r = 0, and (b) U (a) = 0. The first condition gives A = 0
and the second one leads to B = −(ρg/4µ)a2 . The first boundary condition can be
replaced by the requirement that σrz = 0 at r = 0. This will lead to the conclusion
that dU/dr = 0 at r = 0, from which we arrive at the same result (that is, A = 0). The
solution becomes
ρga2 r2

uz (r) = U (r) = − 1− 2 . (5)
4µ a
The stress field becomes
ρg
σθz = 0, σzr = r. (6)
2
Note that the boundary conditions (1) of the 3D problem are not satisfied at z = 0, L.
Hence, it is only an approximate solution.
7.31 An external hydrostatic pressure of magnitude p is applied to the surface of a spherical

body of radius b with a concentric rigid spherical inclusion of radius a, as shown in
Figure
Fig. P7-32 the displacement and stress fields in the spherical body. Using
P7.31. Determine
the stress field obtained, determine the stresses at the surface of a rigid inclusion in an
infinite elastic medium.
Rigid spherical core

b
a
Spherical shell ( μ , λ )
Fig. P7.31
Solution: We use the semi-inverse method to solve the problem. Assume that uφ =
uθ = 0 and uR = U (R). The boundary conditions are:
At r = b : σRR = −p; At R = a : uR = U (a) = 0 (1)
The solution of the Navier equations give the displacement [see Eq. (7.3.13)]
1 4µ
U (R) = AR + B, σRR (R) = (2µ + 3λ)A − B, (2)
R2 R3
where µ and λ are the Lamé constants. Using the boundary conditions, we obtain
(2µ + 3λ = 3K)
4a3 µ

U (a) = 0 → B = −a3 A; σRR (R) = 3K + A, (3)
R3
p
σRR (b) = −p → A = − . (4)
4a3 µ
3K + b3
Hence the displacement uR and stress field are given by

b3 pR a3

uR (R) = − 1 − (5)
3Kb3 + 4µa3 R3
3 3
a3

1 + 2α(a /R ) 2µ
σRR (R) = − p, α = , β = 2α 3 (6)
1+β 3K b
1 − α(a3 /R3 )

σθθ (R) = σφφ = − p. (7)
1+β
To obtain the stresses at the surface of a rigid inclusion in an infinite elastic medium,
we let b → ∞ and obtain
4µa3 2µa3

σRR = − 1 + p, σθθ = σ φφ = − 1 − p. (8)
3KR3 3KR3
At the interface of the rigid inclusion and the elastic medium (R = a), the stresses are

4µ 2µ
σRR = − 1 + p, σθθ (R) = σφφ = − 1 − p. (9)
3K 3K
7.32 Consider the concentric spheres shown in Fig. P7.32. Suppose that the core is elastic
Figure
and the outerP7-33
shell is subjected to external pressure p (both are linearly elastic). As-
suming Lamé constants of µ1 and λ1 for the core and µ2 and λ2 for the outer shell,
and that the interface is perfectly bonded at R = a, determine the displacements of
the core as well as for the shell.
p
Spherical core ( μ1 , λ1 ) b
a
Spherical shell ( μ2 , λ2)
Fig. P7.32
Solution: From Eq. (7.3.13), the displacements are of the form

(1) B1 (2) B2
uR (R) = A1 R + , uR (R) = A2 R + ,
R2 R2
where 3Ki = 2µi + 3λi (i = 1, 2), A1 , A2 , B1 , and B2 are constants. Note that the
radial stress in each part is given by
(i) 4µi
σRR = 3Ki Ai − Bi .
R3
The boundary conditions are
At R = b : σRR = −p, σRθ = 0, σRφ = 0.
The interface conditions are (continuity of the displacements and balance of forces)
(1) (2) (1) (2) (1) (2)
At R = a : σRR = σRR , σRθ = σRθ , σRφ = σRφ
(1) (2) (1) (2) (1) (2)
uR = uR , u θ = uθ , u φ = uφ .
We have
(1) B1 (1) 4µ1
uR (R) = A1 R + , σRR = 3K1 A1 − B1 (core)
R2 R3
(2) B2 (2) 4µ2
uR (R) = A2 R + 2 , σRR = 3K2 A2 − 3 B2 (shell)
R R
where 3Ki = 2µi + 3λi (i = 1, 2). The four constants can be determined using the
following four conditions:
(2) (1) (2) (1) (2)
σRR (b) = −p; σRR (a) = σRR (a); uR (a) = uR (a); B1 = 0 by symmetry.
Figure
7.33 Consider a longP7-34
hollow circular shaft with a rigid internal core (a cross section of the
shaft is shown in Fig. P7.33). Assuming that the inner surface of the shaft at r = a
is perfectly bonded to the rigid core and the outer boundary at r = b is subjected to
a uniform shearing traction of magnitude τ0 , find the displacement and stress fields in
the problem.
Rigid core
τ0
τ0
b
Long hollow
a
cylindrical
Shaft ( μ , λ )
τ0 τ0
Fig. P7.33
Solution: The problem can be treated as one a plane strain. First, write the boundary
conditions of the problem:
At r = a : ur = 0, uθ = 0; At r = b : σrr = 0, σrθ = τ0
Next, assume ur = 0 and uθ = U (r). Then from Eq. (7.2.3) we have

dU U
εrr = 0, εθθ = 0; 2εrθ = −
dr r
and from the stress–strain relations (7.2.9) we have

dU U
σrr = 0, σθθ = 0; σrθ = µ −
dr r
Substitution of the stress components into the 2nd equilibrium equation [only nonzero
terms are retained; see Eq. (7.2.7)]
∂σrθ 2σrθ
+ = 0,
∂r r
yields
d dU U 2 dU U
− + − = 0.
dr dr r r dr r
Noting the following identities
1 dU U d U dU U 1 d
− = , + = rU ,
r dr r dr r dr r r dr
we obtain
C
U (r) = Br +
r
Use of the boundary conditions
h dU Ui
U (a) = 0, µ − = τ0 ,
dr r r=b
gives
C b2 τ 0
B=− 2, C=− .
a 2µ
Hence the displacement uθ = U and stress σrθ become
τ 0 b2 r a b2 τ 0
uθ (r) = − , σrθ = 2 .
2µa a r r
7.34 For the plane stress field

σxx = cxy, σxy = 0.5c(h2 − y 2 ), σyy = 0,
where c and h are constants, (a) show that it is in equilibrium under a zero body force,
and (b) find an Airy stress function Φ(x, y) corresponding to it.
Solution: (a) First we check equilibrium conditions

∂σxx ∂σxy
+ = cy − cy = 0
∂x ∂y
∂σxy ∂σyy
+ = 0 + 0 = 0,
∂x ∂y
which that the stresses are in equilibrium.
(b) To determine the Airy stress function, we use the given stress field to determine
the possible terms of the polynomial used for Φ:
∂2Φ
= σxx = cxy → Φ = c1 xy 3
∂y 2
∂2Φ
= σyy = 0 → Φ = c2 x
∂x2
2
∂ Φ
= −σxy = −0.5c(h2 − y 2 ) → Φ = c3 xy + c4 xy 3
∂x∂y
Thus, we may select Φ(x, y) to be (note that the constant and linear terms in x and y
do not contribute to the stress field)
Φ(x, y) = Axy + Bxy 3 .
Clearly, ∇4 Φ = 0, and
∂2Φ ∂2Φ
σxx = 2
= 6Bxy → 6B = c, σyy = = 0,
∂y ∂x2
∂2Φ
σxy =− = −A − 3Bxy 2 = 0.5c(h2 − y 2 ) → A = −0.5ch2 .
∂x∂y
Hence, we have
Φ(x, y) = 16 c(−3h2 xy + xy 3 ).
7.35 In cylindrical coordinates, we assume that the body force vector f is derivable from the
scalar potential Vf (r, θ):

∂Vf 1 ∂Vf
f = −∇Vf fr = − , fθ = − , (1)
∂r r ∂θ
and define the Airy stress function Φ(r, θ) such that
1 ∂2Φ ∂2Φ

1 ∂Φ ∂ 1 ∂Φ
σrr = + 2 + Vf , σθθ = + Vf , σrθ = − . (2)
r ∂r r ∂θ2 ∂r2 ∂r r ∂θ
Show that this choice trivially satisfies the equations of equilibrium
∂σrr 1 ∂σrθ 1
+ + (σrr − σθθ ) + fr = 0 ,
∂r r ∂θ r (3)
∂σθr 1 ∂σθθ 2σrθ
+ + + fθ = 0.
∂r r ∂θ r
The tensor form of the compatibility condition in Eq. (7.5.33) is invariant.
Solution: Substituting for the stress components in terms of the Airy stress function
into the equations of equilibrium, we obtain
∂σrr 1 ∂σrθ 1 ∂Vf
+ + (σrr − σθθ ) −
∂r r ∂θ r ∂r
1 ∂2Φ

∂ 1 ∂Φ 1 ∂ ∂ 1 ∂Φ
= + 2 + Vf + −
∂r r ∂r r ∂θ2 r ∂θ ∂r r ∂θ
2
∂2Φ

1 1 ∂Φ 1 ∂ Φ ∂Vf
+ + 2 + V f − − V f −
r r ∂r r ∂θ2 ∂r2 ∂r
2 2 2

∂ 1 ∂ Φ 1 ∂ 1∂ Φ 1 ∂ Φ
= − + 3 = 0.
∂r r2 ∂θ2 r ∂r r ∂θ2 r ∂θ2
∂σθr 1 ∂σθθ 2σrθ 1 ∂Vf
+ + −
∂r r ∂θ r r ∂θ
∂ 2 1 ∂Φ
2
1 ∂ ∂ Φ 2 ∂ 1 ∂Φ 1 ∂Vf
=− 2 + + V f − −
∂r r ∂θ r ∂θ ∂r2 r ∂r r ∂θ r ∂θ
2 ∂Φ 2 ∂2Φ 1 ∂3Φ 1 ∂3Φ 2 ∂2Φ 2 ∂Φ
=− + − + − + 3 = 0.
r3 ∂θ r2 ∂r∂θ r ∂r2 ∂θ r ∂r2 ∂θ r2 ∂r∂θ r ∂θ
Figure P7-37
7.36 Interpret the Airy stress field obtained with the stress function in Eq. (7.5.42) when
all constants except c3 are zero. Use Fig. 7.5.5 to sketch the stress field.
Solution: This is a special case of what is already included in Example 7.5.3. From Eq.
(7.5.43), it is clear that the problem corresponds to that of uniform axial stress field,
σxx = 2c3 , which is nothing but a centroidally loaded uniaxial member.
2c3
Fig. P7.36
7.37 Interpret the following stress field obtained in Example 7.5.5 using the domain shown
Figure P7-38
in Fig. 7.5.6:
σxx = 6c10 xy, σyy = 0, σxy = −3c10 y 2 .
Assume that c10 is a positive constant.
Solution: The stress field is shown in Fig. P7.37.

y
σ xy = −3c10b2
σ xx = 6c10ab
b
x
b
a σ xx = −6c10ab
σ xy = −3c10b2
Fig. P7.37
7.38 Compute the stress field associated with the Airy stress function
Figure P7-39
Φ(x, y) = Ax5 + Bx4 y + Cx3 y 2 + Dx2 y 3 + Exy 4 + F y 5 .
Interpret the stress field for the case in which constants A, B, and C are zero. Use Fig.
P7.38 to sketch the stress field.
b
x
b
L
Fig. P7.38
Solution: The stresses and the derivatives of Φ required in the biharmonic equation are
∂2Φ ∂4Φ
σxx = = 2Cx3 + 6Dx2 y + 12Exy 2 + 20F y 3 , = 24Ex + 120F y,
∂y 2 ∂y 4
∂2Φ ∂4Φ
σyy = = 20Ax3 + 12Bx2 y + 6Cxy 2 + 2Dy 3 , = 120Ax + 24By,
∂x2 ∂x4
∂2Φ 4
∂ Φ
= − 4Bx3 + 6Cx2 y + 6Dxy 2 + 4Ey 3 ,

σxy =− 2 2 2 = 24Cx + 24Dy.
∂x∂y ∂x ∂y
Then, ∇4 Φ = 0 requires
1
120Ax+24By+24Cx+24Dy+24Ex+120F y = 0 ⇒ E = − (5A + C) , F = − (B + D) .
5
All other constants (A, B, C, and D) are arbitrary. The stress components become
σxx = −60Axy 2 − 4By 3 + 2C x3 − 6xy 2 + 2D 3x2 y − 2y 3 ,

σyy = 20Ax3 + 12Bx2 y + 6Cxy 2 + 2Dy 3 ,

σxy = 20Ay 3 − 4Bx3 + 2C 2y 3 − 3x2 y − 6Dxy 2 .

For A = B = C = 0 and D 6= 0 (E = 0 and F = −D/5), we have
Figure P7-39b σxx = 2D 3x2 y − 2y 3 , σyy = 2Dy 3 , σxy = −6Dxy 2 .

The stresses are shown in the figure below.
y σxy = -6 Db x
2
σ xx = 2D (3a2b - 2b3 )
σ yy = 2D b3
σ xx = 2D (3a2 y - 2 y3 ) σ xy = -6 D x y2
b
x
b
σ yy = -2Db3
σ xy = −6 Dxb2 a
Fig. P7.38
Figure P7-40
7.39 Determine the Airy stress function for the stress field of the beam shown in Fig. P7.40
and evaluate the stress field.
y q0 (force per unit area)
b
2b x
b
a
Fig. P7.39
Solution: The following discussion provides a logic for the selection of Φ. First note
that the constant and linear coefficients in x and y do not contribute to the calculation
of stress components and hence can be omitted. Then because σyy = −q0 at y = b
[see Eq. (3)] and σyy = 0 at y = −b [see Eq. (4)], Φ must only be an add function of
y(if terms containing even functions of y are included, they will drop out). This also
follows from the fact that the integral of σxx over the cross section must be zero [see
Eq. (2); integrals of even functions over −b to +b will not vanish]. Finally, because
σxx is a constant at y = ±b, Φ must not include powers of x greater than 2 (so that
the second derivative of Φ with respect to x is a constant). Thus, we select
Φ(x, y) = Axy + Bx2 + Cx2 y + Dy 3 + Exy 3 + F x2 y 3 + Gy 5 .
First, we must check to see if Φ satisfies ∇4 Φ = 0. We have
∂4Φ ∂4Φ ∂4Φ

∇4 Φ = 4
+2 2 2 + = 24F y + 120Gy
∂x ∂x ∂y ∂y 4
Thus, we must have F = −5G in order to have ∇4 Φ = 0 (we already reduced the
number of unknowns by one!).
Next, we compute the stress components using the definitions:
∂2Φ
σxx = = 6Dy + 6Exy + 6F x2 y + 20Gy 3 = 6Dy + 6Exy + 10G(−3x2 y + 2y 3 )
∂y 2
∂2Φ
σyy = = 2B + 2Cy + 2F y 3 = 2B + 2Cy − 10Gy 3
∂x2
∂2Φ
σxy =− = −(A + 2Cx + 3Ey 2 + 6F xy 2 ) = −(A + 2Cx + 3Ey 2 − 30Gxy 2 ) (1)
∂x∂y
The plane stress elasticity boundary conditions are:
x=0: σxx = 0, σxy = 0, (2)

y=b: (σyy )y=b = −q0 , (σxy )y=b = 0, (3)
y = −b : (σyy )y=−b = 0, (σxy )y=−b = 0, (4)
x=a: u = 0, v = 0. (5)
We wish to determine all of the constants using the stress boundary conditions in Eqs.
(2)-(4).
We begin with the boundary conditions in (2):
σxx (0, y) = 6Dy + 20Gy 3 = 0 → D = 0, G = 0 → F = 0,
σxy (0, y) = −(A + 3Ey 2 ) → A = 0, E = 0.

The second condition of Eq. (3) or the second condition of Eq. (4) gives the result
A + 2Cx = 0 which implies A = 0 and C = 0.
Thus, all of the constants are zero, giving zero stress field, which of course is not correct.
Therefore, an exact plane stress elasticity solution cannot be determined with the Φ
we have selected.
Then we wish to invoke Saint Venant’s principle and replace the first boundary condi-
tion in Eq. (2) with the following ones:
Z b Z b
x=0: h σxx dy = 0; h yσxx dy = 0 (6)
−b −b
and then determine the constants A, B, C, D, E and G using the boundary conditions
in Eqs. (2)-(4) and (6). If we are able to find a nontrivial solution, it is an approximate
solution to the problem. If we find that all of the constants are zero, we have to pick an-
other elasticity boundary condition and replace it with an equivalent integral boundary
condition. Thus there are several possible approximate solutions (even though there is
only one elasticity solution, but we cannot determine it with the present approach).
Proceeding with Eq. (6), we note that because σxx is an odd function of y and therefore
the first condition in (6) is automatically satisfied (without giving any relation among
the constants). The second condition in (6) gives
Z b
0=h y(6Dy + 10G2y 3 )dy → D = −2Gb2
−b
The second condition in (2) gives
0 = σxy (0, y) = −(A + 3Ey 2 ) → A = 0, E = 0.
The second condition of Eq. (3) or the second condition of Eq. (4) give the result
0 = σxy (x, ±b) = −x(2C − 30Gb2 ) → C = 15Gb2 . (7)
Next, using the first of Eq. (3) and the first of Eq. (4), we obtain
σyy (x, b) = 2B + 2Cb − 10Gb3 = −q0 , σyy (x, −b) = 2B − 2Cb + 10Gb3 = 0,
which give, in view of Eq. (7), the result

q0 3q0 q0 q0
B=− , C=− , G=− , D= . (8)
4 8b 40b3 20b
Thus, we have determined all of the constants using all of the stress boundary condi-
tions. Note that we have not used the boundary conditions at the fixed end (x = a).
Thus, the stresses are
5a2 x2 y 5 y3

3q0 y
σxx = + 2 2 − ,
10 b 2b a b 3 b3
y3

q0 y
σyy = −2 − 3 + 3 , (9)
4 b b
y2

3q0 a x
Figure P7-39b σxy = 4b a 1 − b2 .
The stress distributions are shown on the boundary of the domain in the figure below.
It would be a long process to determine the displacements from the stress compo-
nents and then use the displacement boundary conditions in Eq. (5) to determine the
constants of integration.
y σ yy ( x , b) = q0 , σ xy ( x ,b) = 0
V = q0ha
σ xx (0, y ) = 0,
σ xy (0, y ) = 0 x N =0
q ha 2
M = 0
2
σ yy ( x , −b) = 0, σ xy ( x , −b) = 0
7.40 Investigate what problem is solved by the Airy stress function

xy 3

3A B
Φ= xy − 2 + y 2 .
4b 3b 4b
Solution: Compute the stresses first and the derivatives of Φ required in the biharmonic
equation:
∂2Φ 3A B ∂4Φ
σxx = 2
= − 3 xy + , =0
∂y 2b 2b ∂y 4
2 4
∂ Φ ∂ Φ
σyy = = 0, = 0,
∂x2 ∂x4
2 2 4

∂ Φ 3A y ∂ Φ
σxy = − =− 1− 2 , 2 2 2 = 0.
∂x∂y 4b b ∂x ∂y
Clearly, ∇4 Φ = 0 is satisfied. Note that
Z b Z b Z b
hσxy dy = −Ah ≡ −F0 , hσxx dy = Bh ≡ P, hyσxx dy = −Ahx = −F0 x,
−b −b −b
where F0 is the transverse shear force at x = 0, P is the axial force and h is the
Figure P7-41
thickness of the domain. Similarly, Thus, the stress field corresponds to a cantilever
beam fixed at x = a and free at x = 0, and subjected to an upward transverse point
load F0 = Ah and axial load P = Bh at the free end (x = 0). The domain and loads
are depicted in the figure below.
y
3aA B 3A  y2 
h F0 = Ah σ xx = − y+ , σ xy = − 1 − 
2b3 2b 4b  b2 
P = Bh
2b x
Cross section a
7.41 Show that the Airy stress function

q0 1
Φ(x, y) = 3 x2 y 3 − 3b2 y + 2b3 − y 3 y 2 − 2b2

8b 5
satisfies the compatibility condition. Determine the stress field and find what problem
it corresponds to when applied to the region −b ≤ y ≤ b and x = 0, a (see Fig. P7.38).
Solution: Compute the stresses first and the derivatives of Φ required in the biharmonic
equation:
∂2Φ 12b2 ∂4Φ

q0 2 3 3q0 y
σxx = 2
= 3
6x y − 4y + y , =− 3 ,
∂y 8b 5 ∂y 4 b
∂2Φ 2q0 3 ∂ 4
Φ
= 3 y − 3b2 y + 2b3 ,

σyy = = 0,
∂x2 8b ∂x4
2 4
∂ Φ 2q0 ∂ Φ 3q0 y
= − 3 x 3y 2 − 3b2 ,

σxy = − 2 2 2 = 3 .
∂x∂y 8b ∂x ∂y b
Clearly, ∇4 Φ = 0 is satisfied. We have

Z b Z b Z b
q0 h 2
hσxy dy = −q0 xh, hσxx dy = 0, hyσxx dy = x ,
−b −b −b 2
σyy (x, b) = 0, σyy (x, −b) = q0 , σxy (x, b) = 0, σxy (x, −b) = 0, σxx (x, 0) = 0.
Figure P7-42
Clearly, it is a beam subjected to uniformly distributed load of intensity q0 (per unit
area) at the bottom of the beam. The stress distributions are indicated in the figure
below.
q0 y  6  y 
2 2
a
σ xx ( a , y ) =  + 3  − 2  
4 b 5 b
    
b

y
3q0 a  y2 
σ yy ( x ,b) = σ xy ( x ,b) = 0 σ xy ( a, y) = 1 − 
4b  b2 
σ xx (0, y ) = f0 ( y), σ xy (0, y) = 0,
b
q0 y  3  y  
2
f0 ( y ) =  −   x
2 b 5  b   b
 
a
σ yy ( x , −b) = q0 , σ xy ( x , −b) = 0
7.42 The thin cantilever beam shown in Fig. P7.42 is subjected to a uniform shearing
Figure P7-43
traction of magnitude τ0 along its upper surface. Determine if the Airy stress function
xy 2 xy 3 ay 2 ay 3

τ0
Φ(x, y) = xy − − 2 + + 2
4 b b b b
satisfies the compatibility condition and stress boundary conditions of the problem.
y
τ0
h
b
2b
b x
a
Fig. P7.42
Solution: We must check if Φ satisfies ∇4 Φ = 0 and the boundary conditions of the

problem. We have
y2 y3 ∂2Φ ∂4Φ

∂Φ τ0
= y− − 2 , 2
=0 → = 0,
∂x 4 b b ∂x ∂x4
3xy 2 3ay 2

∂Φ τ0 2xy 2ay
= x− − 2 + + 2 ,
∂y 4 b b b b
∂2Φ

τ0 2x 6xy 2a 6ay
= − − 2 + + 2 ,
∂y 2 4 b b b b
3 4

∂ Φ τ0 6x 6a ∂ Φ
= − 2 + 2 , = 0.
∂y 3 4 b b ∂y 4
Thus, we have ∇4 Φ = 0.
The stresses are given by
∂2Φ ∂2Φ

τ0 2x 6xy 2a 6ay
σxx = = − − 2 + + 2 , σyy = = 0,
∂y 2 4 b b b b ∂x2
∂2Φ 3y 2

τ0 2y
σxy =− =− 1− − 2 .
∂x∂y 4 b b
The plane stress boundary conditions of the problem are:
At y = −b : σyy = 0, σyx = 0; At y = b : σyy = 0, σyx = τ0 ,

At x = a : σxx = 0, σxy = 0.
It is clear that all boundary conditions except σxy (a, y) = 0 are satisfied by the stresses.
However, it is satisfied in the integral sense:
Z b
τ0 b 3y 2
Z
2y τ0
σxy (a, y) dy = − 1− − 2 dy = − (2b − 2b) = 0.
−b 4 −b b b 4
7.43 Consider the problem of a cantilever beam carrying a uniformly varying distributed
transverse load, as shown in Fig. P7.43. The following Airy stress function is suggested
(explain the terms to yourself):
Φ(x, y) = Axy + Bx3 + Cx3 y + Dxy 3 + Ex3 y 3 + F xy 5 .
Determine each of the constants and find the stress field.

Figure P7-43
æxö
q0 çç ÷÷÷
çè L ø
y
h
b
2b x
b
L
Fig. P7.43
Solution: We have
∂Φ
= Ay + 3Bx2 + 3Cx2 y + Dy 3 + 3Ex2 y 3 + F y 5 ,
∂x
∂2Φ
= 6Bx + 6Cxy + 6Exy 3 = σyy ,
∂x2
∂3Φ ∂4Φ
= 6(B + Cy + Ey 3 ), = 0,
∂x3 ∂x4
∂Φ
= Ax + Cx3 + 3Dxy 2 + 3Ex3 y 2 + 5F xy 4 ,
∂y
2
∂ Φ
= A + 3Cx2 + 3Dy 2 + 9Ex2 y 2 + 5F y 4 = −σxy ,
∂x∂y
∂2Φ
= 6Dxy + 6Ex3 y + 20F xy 3 = σxx ,
∂y 2
∂3Φ ∂4Φ
= 6Dx + 6Ex3 + 60F xy 2 , = 120F xy,
∂y 3 ∂y 4
4
∂ Φ
= 36Exy. (1)
∂x2 ∂y 2
Thus, to satisfy ∇4 Φ = 0, we have
∂4Φ ∂4Φ ∂4Φ
∇4 Φ = 4
+2 2 2 + = 0 → E = −(120/72)F = −(5/3)F. (2)
∂x ∂x ∂y ∂x4
The remaining constants, A, B, C, D, and E must be determined to meet the boundary
conditions.
First, we use the boundary condition on σyy at y = ±b
x q0
→ 6(B + Cb + Eb3 ) =
σyy (x, b) = q0 ,
L L (3)
3
σyy (x, −b) = 0 → 6(B − Cb − Eb ) = 0.
Thus, we have
q0 q0 5b3
B=
, Cb = + F. (4)
12L 12L 3
Next, we use the boundary condition on σxy at y = ±b
σxy (x, ±b) = 0 → A + 3Cx2 + 3Db2 + 9Ex2 b2 + 5F b4 = 0, (5)
which give
A + 3Db2 + 5F b4 = 0, C + 3Eb2 = 0 or C = 5b2 F. (6)
Following the discussion of Problem 7.39, we invoke Saint Venant’s principle and write
the boundary conditions at x = 0 with the integrals
Z b Z b Z b
h σxx (0, y) dy = 0; h yσxx (0, y) dy = 0; h σxy (0, y) dy = 0. (7)
−b −b −b
The first two conditions are trivially satisfied. The third one gives
Ab + Db3 + F b5 = 0, in view of Eq. (6) → A = b4 F, D = −2b2 F. (8)
Equating the value of C from Eq. (6) to that in Eq. (4), we obtain the value of F as
F = q0 /8bL. In summary, we obtain
q0 L q0 L q0 q0
A = b4 F = , B= , C = 5b2 F = , D = −2b2 F = − ,
40L 12 8bL 20bL (9)
5 q0 q0
E=− F =− , F = .
3 24b3 L 8bL
Thus, we have determined all of the constants using all of the stress boundary condi-
tions. Note that we have not used the boundary conditions at the fixed end (x = L).
The stresses are given by
∂2Φ x2 L2 y2

3 3 q0 x y
σxx = = 6Dxy + 6Ex y + 20F xy = −6 − 5 + 10 ,
∂y 2 20 L b L2 b2 b2
∂2Φ y3

q0 x y
σyy = 2
= 6Bx + 6Cxy + 6Exy 3 = 2+3 + 3
∂x 4 L b b
(10)
∂2Φ
σxy = −2 = −(A + 3Cx2 + 3Dy 2 + 9Ex2 y 2 + 5F y 4 )
∂x∂y
L x2 b y2 L x2 y 2 b y4

q0 b
=− + 15 − 6 − 15 + 5 .
40 L b L2 L b2 b L2 b2 L b4
7.44 Consider the curved beam shown in Fig. 7.23, which is fixed at the upper end and
subjected to a force P per unit thickness. Determine if an Airy stress function of the
form
B
Φ(r) = Ar3 +
Figure P7-44b r
+ C r log r sin θ
provides a solution to this problem. If so, solve for the values of the constants A, B,
and C.
x2
y
σrθ σ rθ = 0 σ =0
rr
σrr σθθ
r r
a σ rr = 0
b σ rθ = 0
θ
q
x P x1
P σ rθ
b b
σ θθ

a
σ rθ dr = P
σ
a
θθ dr = 0
Fig. P7.44
Solution: We have P = −P ê1 . Using the cylindrical coordinate system shown in the
figure above, we first write the boundary conditions of the plane strain/stress elasticity
problem:
At r = a, b : σrr = 0, σrθ = 0, (1)
π
At θ = : ur = 0, uθ = 0. (2)
2
Equation (2) can be replaced with equivalent force boundary conditions (which we do
not intend to use unless it is necessary to determine A, B and C). At θ = π/2, we have
Z b Z b
(σθθ )θ= π dr = −P, (σrθ )θ= π dr = 0,
2 2
a a
Z b (3)
a+b a+b
rσθθ − σrθ dr = P .
a 2 θ= π 2
2
The boundary conditions at θ = 0 are (in the sense of the Saint Venant’s principle)
Z b Z b Z b
(σθθ )θ=0 dr = 0, (σrθ )θ=0 dr = P, r (σθθ )θ=0 dr = 0. (4)
a a a
First, we must verify that the given stress function satisfies the biharmonic equation
(in the absence of the body forces):

∂Φ B
= 3Ar2 − 2 + C log r + C sin θ,
∂r r
∂2Φ

B C
= 6Ar + 2 + sin θ,
∂r2 r3 r

∂Φ B
= Ar3 + + C r log r cos θ,
∂θ r
∂2Φ

3 B
= − Ar + + C r log r sin θ,
∂θ2 r
∂2Φ 1 ∂2Φ

1 ∂Φ 2
∇2 Φ = + + = 8Ar + C sin θ,
∂r2 r ∂r r2 ∂θ2 r

∂ 2C
∇2 Φ = 8A − 2 sin θ,

∂r r
2
∂ C
∇2 Φ = 4 3 sin θ,

∂r2 r
∂ 2
∇2 Φ = 8Ar + C cos θ,

∂θ r
2

∂ 2 2
∇ Φ = − 8Ar + C sin θ,
∂θ2 r
2
1 ∂2

∂ 1 ∂ 2
∇4 Φ = + + 8Ar + C sin θ = 0.
∂r2 r ∂r r2 ∂θ2 r
Thus for any A, B and C, we have ∇4 Φ = 0. Next we wish to determine A, B and C
such that the stress boundary conditions in (1), (2) and (4) are satisfied. If these are
not sufficient to determine A, B and C then only we shall use the boundary conditions
in Eq. (3).
Compute stresses in terms of A, B and C:
1 ∂2Φ

1 ∂Φ 2B C
σrr = + 2 = 2Ar − 3 + sin θ,
r ∂r r ∂θ2 r r
∂2Φ

B C
σθθ = = 6Ar + 2 3 + sin θ, (5)
∂r2 r r

∂ 1 ∂Φ 2B C
σrθ = − = − 2Ar − 3 + cos θ.
∂r r ∂θ r r
Applying boundary conditions in (1), we obtain
2B C 2B C
2Aa − + = 0, 2Ab − + = 0,
a3 a b3 b
which give
B = −Aa2 b2 , C = −2A(a2 + b2 ) (6)
so that the stress components (5) become
a2 b2 a2 + b2

σrr = 2A r + 3 − sin θ,
r r
a2 b2 a2 + b2

σθθ = 2A 3r − 3 − sin θ, (7)
r r
a2 b2 a2 + b2

σrθ = −2A r + 3 − cos θ.
r r
Boundary conditions in the first and third equations of (4) are identically satisfied
(because sin θ = 0 for θ = 0). The second boundary condition in (4) yields

b
−2A b2 − a2 − (a2 + b2 ) log =P (8)
a
and we have
P P a2 b2 P (a2 + b2 )
A=− , B=− , C= , (9)
2∆ 2∆ ∆
where
b
∆ = b2 − a2 − (a2 + b2 ) log
. (10)
a
Thus, all of the constants are determined without using the boundary conditions at the
fixed end of the curved beam. It can be verified that the first two boundary conditions
in (3) are identically satisfied but the third condition is not satisfied. The solution
obtained is not an elasticity solution; it is an approximate solution.
7.45 Determine the stress field in a semi-infinite plate due to a normal load, f0 force/unit
length, acting on its edge, as shown in Fig. P7.45. Use the following Airy stress function
(that satisfies the compatibility condition ∇4 Φ = 0):
Φ(r, θ) = Aθ + Br2 θ + Crθ sin θ + Drθ cos θ,
Figure
where A,P7-45b
B, C, and D are constants [see Eq. (7.5.40) for the definition of stress
components in terms of the Airy stress function Φ]. Neglect the body forces (that
is, Ω = 0). Hint: Stresses must be single-valued. Determine the constants using the
boundary conditions of the problem.
bf0
f0 z
b θ
θ r dθ
x σ rθ
r bf0 σ rr
(bσ rr rdθ ) cos θ − (bσ rθ rdθ )sin θ

y
(bσ rr rdθ )sin θ + (bσ rθ rdθ ) cos θ
Fig. P7.45
Solution: It can be verified that the given stress function satisfies the biharmonic
equation. We must check if it also satisfies the stress boundary conditions. The stresses
are
1 ∂Φ 1 ∂2Φ 2
σrr = + 2 = 2Bθ + (C cos θ − D sin θ) ,
r ∂r r ∂θ2 r
∂2Φ
σθθ = = 2Bθ,
∂r2
1 ∂Φ 1 ∂2Φ A
σrθ = 2 − = 2 − B.
r ∂θ r ∂θ∂r r
The boundary conditions on the stresses for any r, except at r = 0, are (see Fig. P7.45)
At θ = 0, π : σrθ = 0, σθθ = 0, (1)
Z π
f0 b + (σrr sin θ + σrθ cos θ) br dθ = 0, (2)
Z0 π
(σrr cos θ − σrθ sin θ) br dθ = 0. (3)
0
The boundary conditions in (1) on σrθ and σθθ requires that A = 0 and B = 0. The
symmetry of σrr with respect to θ [or from (3)] yields C = 0. Thus the stresses reduce
to
2D
σrr = − sin θ, σθθ = 0, σrθ = 0. (4)
r
The boundary condition (2) on σrr gives D = f0 /π. Hence the compatible stress
distribution for the problem is
2f0
σrr = − sin θ, σθθ = 0, σrθ = 0. (5)
πr
To determine the stresses σxx , σyy , and σxy , we use the transformation equations in
Eq. (4.3.8):
2f0
σxx = σrr cos2 θ = − sin θ cos2 θ,
πr
2f0
σyy = σrr sin2 θ = − sin3 θ,
πr
2f0
σxy = σrr cos θ sin θ = − sin θ cos2 θ.
πr
7.46 Show that the resultant forces in the three coordinate directions on the end surface
(that is, z = L face) are zero. Also show that the resultant moments about the x- and
y-axes on the end surface are also zero.
Solution: The resultant force on the end surface in the x-direction is

Z Z
∂ψ
Fx = σxz dx dy = µθ − y dx dy.
Ω Ω ∂x
Because
∂2ψ ∂2ψ
2
+ = 0,
∂x ∂y 2
we can write
∂2ψ ∂2ψ

∂ ∂ψ ∂ ∂ψ ∂ψ ∂ψ
x −y + x +x = −y+x 2 +x = −y
∂x ∂x ∂y ∂y ∂x ∂x ∂y 2 ∂x
Therefore, we have
Z Z
∂ ∂ψ ∂ ∂ψ
Fx = σxz dx dy = µθ x −y + x +x dx dy
Ω Ω ∂x ∂x ∂y ∂y
Z
∂ψ ∂ψ
= µθ x nx − y + ny + x ds
Γ ∂x ∂y
Z
∂ψ ∂ψ
= µθ x nx − y + ny + x ds = 0,
Γ ∂x ∂y
where Eqs. (7.5.68) and (7.5.69) are used in arriving at the final step. Similarly, the
resultant force in the y-direction can be shown to be zero. The resultant force in the
z-direction is zero because the traction tz = 0.
The resultant moments about the x-axis and the y-axis are (because σzz = 0)
Z Z
Mx = z σzz dx dy = 0, My = (−x) σzz dx dy = 0.
Ω Ω
The resultant moment about the z-axis is

Z Z
∂ψ ∂ψ
T = Mz = (x σyz − y σxz ) dx dy = µθ x + x2 − y + y 2 dx dy ≡ µθ J,
Ω Ω ∂y ∂x
where J is the torsion constant
Z
∂ψ 2 ∂ψ 2
J= x +x −y + y dx dy.
Ω ∂y ∂x
7.47 Use the warping function ψ(x, y) = kxy, where k is a constant, to determine the cross
section for which it is the solution. Determine the value of k in terms of the geometric
parameters of the cross section and evaluate stresses in terms of these parameters and
µ.
Solution: Clearly, the function ψ = kxy satisfies the equilibrium condition, ∇2 ψ = 0.

The boundary condition in Eq. (7.5.69) becomes

∂ψ dy ∂ψ dx dy dx
0= −y − +x = (k − 1)y − (k + 1)x
∂x ds ∂y ds ds ds
1 d
(1 + k)x2 + (1 − k)y 2 ,

=−
2 ds
or
d 1−k
(1 + k)x2 + (1 − k)y 2 = 0, x2 + y 2 = a2 on Γ,

ds 1+k
where a is a constant. This is the equation for an ellipse with semiaxes a and b (a ≥ b)
x2 y2 a2 1−k
2
+ 2 = 1, = .
a b b2 1+k
The constant k can be expressed in terms of a and b
a2 − b2 a2 − b2
k=− , ψ(x, y) = − xy
a2 + b2 a2 + b2
The stresses are
2
a − b2 2µθa2

∂ψ
σxz = µθ −y
= −µθy 2 2
+ 1 =− 2 y,
∂x a +b a + b2
2
a − b2 2µθb2

∂ψ
σyz = µθ + x = µθx − 2 + 1 = x.
∂y a + b2 a2 + b2
7.48 Consider a cylindrical member with the equilateral triangular cross section shown in
Fig. P7.48. Show that the exact solution for the problem can be obtained and that the
twist per unit length θ and stresses σxz and σyz are given by
√
5 3T µθ µθ 2
θ= , σxz = (x − a)y, σyz = (x + 2ax − y 2 ).
27µa4 a 2a
Hint: First write the equations for the three sides of the triangle (that is, y = m x + c,
where m denotes the slope and c denotes the intercept), with the coordinate system
shown in the figure, and then take the product of the three equations to construct the
stress function. Also note that
Z Z a Z x+2a
√
3
F (x, y) dx dy = F (x, y) dy dx.
Ω −2a − x+2a
√
3
Figure P7-48
T 3a
Side 1
Side 3
x
x = -2a
Side 2
x=a
Fig. P7.48
Solution: The equations of various sides of the triangle are:

√
side 1: x + 2a − 3y = 0,
√
side 2: x + 2a + 3y = 0,
side 3: x − a = 0.
The stress function can be chosen to be the product of the equations of the line segments
bounding the domain
Ψ(x, y) = A x2 + 4a2 + 4ax − 3y 2 (x − a)

= A x3 + 3ax2 − 3xy 2 − 4a3 + 3ay 2 .

Computing ∇2 Ψ, we have
∂2Ψ ∂2Ψ
2
+ = 12aA.
∂x ∂y 2
Hence, A = −µθ/6a. The stresses are given by
∂Ψ µθ ∂Ψ µθ 2
x + 2ax − y 2 .

σxz = = y(x − a), σyz = − =
∂y a ∂x 2a
The applied torque T is related to the angle of twist θ by Eq. (5) of Example 7.5.8
Z
T =2 Ψ(x, y) dx dy
Ω
Z x+2a
√
2µθ a
Z
3
x3 + 3ax2 − 3xy 2 − 4a3 + 3ay 2 dydx

=−
6a −2a − x+2a √
3
Z a
µθ 3 x+2a
√
=− x y + 3ax2 y − xy 3 − 4a3 y + ay 3 x+2a3
dx
3a −2a − √
3
√ Z a
4 3µθ 4
x + 5ax3 + 6a2 x2 − 4a3 x − 8a4 dx

=−
27a −2a
√ a
4 3µθ x5 x4

=− + 5a + 2a2 x3 − 2a3 x2 − 8a4 x
27a 5 4 −2a
√
4 3µθ 243 5 27
=− (− a ) = √ µθa4 .
27a 20 5 3
Hence, the angle of twist is given by
√
5 3T
θ= .
27µa4
7.49 Consider torsion of a cylindrical member with the rectangular cross section shown in
Figure P7.49. Determine if a function of the form
2 2
x y
Ψ=A − 1 − 1 ,
Figure P7-49 a2 b2
where A is a constant, can be used as a Prandtl stress function.
y
T
a a
b
x
b
Fig. P7.49
Solution: The Prandtl stress function must satisfy Poisson’s equation ∇2 Ψ = −2µθ
inside Ω and vanish on the boundary Γ. Although the given function satisfies the
boundary condition for any m, The given stress function, upon differentiation, yields
2
1 x2

1 y
∇2 Ψ = 2A 2 − 1 + − 1
a b2 b2 a2
which is not a non-zero constant. Hence, Ψ will not satisfy the compatibility condition
point wise. A numerical method, such as the Ritz method, must be used to solve the
problem.
7.50 From Example 7.5.8, we know that for circular cylindrical members we have ψ = 0.
Use the cylindrical coordinate system to show that σzr = 0 and σzα = T r/J, where J
is the polar moment of inertia.
Solution: From the transformation equations, we have
σzα = cos α σyz − sin α σxz , σzr = sin α σyz + cos α σxz
∂ψ ∂ψ 1 ∂ψ ∂ψ ∂ψ 1 ∂ψ
= cos α − sin α = sin α + cos α
∂x ∂r r ∂α ∂y ∂r r ∂α
The stresses are

∂ψ ∂ψ 1 ∂ψ
σxz = µθ − y = µθ cos α − sin α − r sin α ,
∂x ∂r r ∂α

∂ψ ∂ψ 1 ∂ψ
σyz = µθ + x = µθ sin α + cos α + r cos α .
∂y ∂r r ∂α
From the above relations it follows that

1 ∂ψ ∂ψ
σzα = µθ +r , σzr = µθ .
r ∂α ∂r
Because ψ = 0 and T = µθJ for circular cylinders [see Eq. (3) of Example 7.5.7], we
have σzα = µθr = T r/J and σzr = 0.
7.51 Timoshenko beam theory. Consider the displacement field
u1 (x, y) = yφ(x), u2 (x, y) = v(x), u3 = 0, (1)
where v(x) is the transverse deflection and φ is the rotation about the z-axis. Follow
the developments of Section 7.3.4 and Example 7.6.1 (see Fig. 7.6.1) to develop the
total potential energy functional:
Z " 2 2 #
Figure 7-6-1 1 L

dφ dv
Π(u, v, φ) = EI + GA + φ − qv dx − F0 v(L) − M0 φ(L),
2 0 dx dx
where EI is the bending stiffness and GA is the shear stiffness (E and G are Young’s
modulus and shear modulus, respectively, A is the cross-sectional area, and I is the mo-
ment of inertia). Then derive the Euler equations and the natural boundary conditions
of the Timoshenko beam theory.
z,y,wv
QF00
q(x)
w(0)
v= =0
0
dv = 0
w'(0)
=0
dx x
M0
L
Fig. P7.50
Solution: The nonzero strains associated with the given displacement field are
∂u1 dφ ∂u1 ∂u2 dv
ε11 = εxx = = y , 2ε12 = + =φ+ . (2)
∂x1 dx ∂x2 ∂x1 dx
The total potential energy is
1 L
Z Z Z L
Π(v, φ) = (σxx yεxx + 2σxy εxy ) dAdx − qv dx − F0 v(L) − M0 φ(L)
2 0 A 0
1 L
Z Z L
= (M εxx + 2V εxy ) dx − qv dx − F0 v(L) − M0 φ(L). (3)
2 0 0
where the new variables introduced in arriving at the last expression are defined as
follows: Z Z
M= yσxx dA, V = σxy dA. (4)
A A
Using the constitutive equations, we can write

dφ dv
σxx = Eεxx = Ey , σxy = 2Gεxy =G φ+ , (5)
dx dx
we can write
Z Z
dφ dφ dv dv
M =E y2 dA = EI , V = G φ+ dA = GA φ + . (6)
A dx dx A dx dx
Hence, the total potential energy expression becomes
Z " 2 2 #
1 L
Z L
dφ dv
Π(v, φ) = EI + GA φ + dx − qv dx − F0 v(L) − M0 φ(L).
2 0 dx dx 0
(7)
Using the principle of total potential energy, δΠ = 0, we arrive at

Z L Z L
dφ dδφ dv dδv
0= EI + GA φ + δφ + dx − qδv dx − F0 δv(L) − M0 δφ(L),
0 dx dx dx dx 0
Z L
d dφ d dw dv
= − EI δφ + GA φ + δw + GA φ + δφ dx
0 dx dx dx dx dx
Z L L
dφ dv
− qδv dx − F0 δw(L) − M0 δφ(L) + EI δφ + GA φ + δv (9)
0 dx dx 0
from which we obtain the Euler equations in 0 < x < L

d dv
δv : − GA φ + −q =0 (10)
dx dx

d dφ dv
δφ : − EI + GA φ + =0 (11)
dx dx dx
The boundary expressions indicate that v and φ are the primary variables (that is,
variables whose specification will eliminate the rigid body motions) and

dφ dv
M = EI , V = GA φ + , (12)
dx dx
are secondary variables (variables that specify the forces acting in the beam) of the
problem. Thus the essential boundary conditions involve specifying w and φ, while the
natural boundary conditions involve specifying M and V .
An examination of the beam supports indicate that v(0) = φ(0) = 0; hence, δv(0) = 0
and δφ(0) = 0. Because φ(L) is not known, δφ(L) is arbitrary (and not equal to zero);
similarly, v(L) is not known, δv(L) is arbitrary (and not equal to zero). Consequently,
the boundary expressions in Eq. (9) yield

dφ dv
EI − M0 = 0, GA φ + − F0 = 0. (13)
dx x=L dx x=L
7.52 Identify the bilinear and linear forms associated with the quadratic functional of the
Timoshenko beam theory in Problem 7.51.
Solution: From Eq. (8) of the solution to Problem 7.51, we have

Z L
dφ dδφ dv dδv
B((v, φ), (δv, δφ)) = EI + GA φ + δφ + dx,
0 dx dx dx dx
Z L
L(δv, δφ) = qδv dx.
0
We note that the functional in Eq. (7) in the solution to Problem 7.51 can be obtained
from
1
Π(v, φ) = B((v, φ), (v, φ)) − L(v, φ).
2
7.53 The total potential energy functional for a membrane stretched over domain Ω ∈ <2 is
given by
Z ( " 2 2 # )
T ∂u ∂u
Π(u) = + − f u dx,
Ω 2 ∂x1 ∂x2
where u = u(x1 , x2 ) denotes the transverse deflection of the membrane, T is the ten-
sion in the membrane, and f = f (x1 , x2 ) is the transversely distributed load on the
membrane. Determine the governing differential equation and the permissible bound-
ary conditions for the problem (that is, identify the essential and natural boundary
conditions of the problem) using the principle of minimum total potential energy.
Solution: We have
( Z " 2 2 # Z )
1 ∂u ∂u
δΠ = δ T + dx1 dx2 − f u dx1 dx2
2 Ω ∂x1 ∂x2 Ω
Z Z
∂u ∂δu ∂u ∂δu
= T + dx1 dx2 − f δu dx1 dx2
Ω ∂x1 ∂x1 ∂x2 ∂x2 Ω
2
∂2u
Z I
∂ u ∂u ∂u
= −T + − f δu dx 1 dx2 + T n 1 + n2 δu ds
Ω ∂x21 ∂x22 Γ ∂x1 ∂x2
Thus, δΠ = 0 yields the following differential equation governing u:
2
∂2u

∂ u
−T 2
+ 2
− f = 0 in Ω
∂x1 ∂x2
The boundary terms indicates that u is the primary variable and

∂u ∂u
Q≡T n1 + n2
∂x1 ∂x2
is the secondary variable. Thus, the problem requires the specification of either u or Q
at a boundary point.
7.54 Use the results of Example 7.6.2 to obtain the deflection at the center of a clamped-
clamped beam (length 2L and EI = constant) under uniform load of intensity q0 and
supported at the center by a linear elastic spring (k).
Solution: Because of the symmetry about the center of the beam, we can consider just
a half span of the beam (see the figure below). Suppose that the beam is of length 2L,
and we consider the right half of the beam. Then we have
      
6 −3L −6 −3L  ∆1  6  Q1 
Figure P7-54
 
2 2
     
2EI  −3L 2L 3L L  ∆2 = q0 L −L + Q2 .
      
L 3  −6 3L 6 3L  
 ∆3 
 12  6   Q3 
  
−3L L2 3L 2L2 ∆4 L Q4
    

1
∆2 = ∆3 = ∆4 = 0, Q1 = − k∆1 , Q2 = 0.
2
q0
y, v 1
Fs = 1
kv(0) = 1
kΔ1
2 2 2
x
k EI = constant k
1 Fs
2
L
Half beam model
Note that because of the symmetry, the slope at x = 0 is zero; only half of the spring
force Fs = k∆1 is taken by the half span. The condensed equations are
     
 Q2   −3L   −L 
12EI k q0 L 2EI q0 L
+ ∆1 = , Q3 = 3 −6 ∆1 − 6 ,
L3 2 2 L  12 
Q4 −3L L
   
from which we obtain

     
4 Q2   3L   −L 
q0 L q0 L q0 L
w(0) = ∆1 = , Q3 = − 6 − 6 ,
24EI + kL3 12 + α  12 
Q4 3L L
   
Figure
whereP7-55
α = kL3 /2EI.
7.55 Use the results of Example 7.6.2 to obtain the deflection v(L) and slopes (−dv/dx)(L)
and (−dv/dx)(2L) under a point load F0 for the beam shown in Fig. P7.55. It is
sufficient to set up the three equations for the three unknowns.
y, v F0
L L
x 2L
Fig. P7.55
Solution: This problem illustrates how the result derived in Example 7.6.2 to beams
with multiples segments (called elements in the finite element method). First we rec-
ognize that the given beam can be viewed as an assemblage of two beam segments,
as shown in the figure below. The forces and generalized displacements and forces
are shown on each segment, with superscript on the variables indicating the segment
number. For each segment (e = 1, 2) we have
  (e)   (e) 
∆    Q1 

6 −3L −6 −3L   1(e) 
 
2 2   
  (e) 
2EI  −3L 2L 3L L ∆ Q

2 2
  (e) = (e) .
L3  −6 3L 6 3L    ∆3     Q3  
2 2
−3L L 3L 2L  (e)   (e) 
   
∆4 Q4

(1) (1) (2) (2)
∆1 = ∆2 = 0, ∆3 = 0, Q4 = 0.
The continuity and equilibrium conditions at x = L (at the interface of the two seg-
ments) are
(1) (2) (1) (2) (1) (2) (1) (2)
∆3 = ∆1 , ∆4 = ∆2 , Q3 + Q1 = −F0 , Q4 + Q2 = 0.
The last two conditions can be imposed only if we add the 3rd equation of segment 1 to
the 1st equation of segment 2, and the 4th equation of segment 1 to the 2nd equation
of segment 2. Thus we have the first two equations of segment 1, the last two equations
of segment 2, and two more equations coming from the addition. Thus, we have
   ∆(1)   Q1
(1) 
6 −3L −6 −3L 0 0  1(1) 
    

(1)
 −3L 2L2 2
   
3L L 0 0 
 ∆2  



 Q2 


  (1)   (1)
  (2) 
2EI  −6 3L 6 + 6 3L − 3L −6 −3L  ∆3 Q3 + Q1
  
2 2 2 2  (1) = (1) (2) .
L  −3L L 3L − 3L 2L + 2L 3L L  
3 
 ∆4   
 Q4 + Q2  
 0 0 −6 3L 6 3L   (2)  (2)
   
 ∆3  Q3

    

  
0 0 −3L L2 3L 2L2  ∆(2)  
 
(2) 
4 Q 4
Using the boundary conditions, the following condensed equations for the unknown
generalized displacements are obtained
   (1)   
12 0 −3L   ∆3   −F0 
2EI 
0 4L2 L2  ∆(1) = 0 .
L3  4 
−3L L2 2L2  ∆(2) 4
  0 
Figure P7-56
7.56 Use the results of Example 7.6.2 to obtain the deflection v(2L) and slopes at x = L and
x = 2L for the beam shown in Fig. P7.56. It is sufficient to set up the three equations
for the three unknowns.
y, v F0
q0
L
x
2L
Fig. P7.56
Solution: The forces and generalized displacements and forces are shown on each seg-
ment, with superscript on the variables indicating the segment number. For each
segment (e = 1, 2) we have
  (e)     (e) 
∆1   Q1 

6 −3L −6 −3L     6  
2 2 
 (e) 
   
  (e) 
2EI  −3L 2L 3L L ∆ q L −L Q
 
2 0 2
  (e) =− + (e)
L3  −6 3L 6 3L  
 ∆3   12   6    Q3  
−3L L2 3L 2L2  ∆(e)  L  (e) 
    
Q4

4
with q1 = q0 and q2 = 0. The boundary conditions are

(1) (1) (1) (2) (2) (2)
∆1 = ∆2 = 0, ∆3 = ∆1 = 0, Q3 = −F0 , Q4 = 0.
The continuity and equilibrium conditions at x = L (at the interface of the two seg-
ments) are
(1) (2) (1) (2) (1) (2)
∆3 = ∆1 , ∆3 = ∆1 = 0, Q4 + Q2 = 0.
We add the 3rd equation of segment 1 to the 1st equation of segment 2, and the 4th
equation of segment 1 to the 2nd equation of segment 2 and obtain
   ∆(1)     Q1
(1) 
6 −3L −6 −3L 0 0 
 1 
 (1)    6  
 

(1)
 −3L 2L2 3L L2 0
    
0 
 
 ∆ 2   

 −L 

 

 Q 2



  (1)     (1) (2) 
2EI  −6 3L 12 0 −6 −3L  ∆3 q L 6 Q + Q
     
0 3 1
 2 2 2
 (1) = − + (1) (2) .
L3  −3L L
 0 4L 3L L   
 ∆4   12  L 
 Q4 + Q2  
 0 0 −6 3L 6 3L  (2) 0 (2)
     
 ∆3  Q3
  
 
  
 

   
 
0 0 −3L L2 3L 2L2  ∆(2)  0
   
 (2) 
4 Q4
Using the boundary conditions, the following condensed equations for the unknown
generalized displacements are obtained
 2   (1) 
4L 3L L2 
   
 ∆4  L  0
2EI  q L

0
3L 6 3L  ∆(2) =− 0 + −F0 .
L3  3  12   
L2 3L 2L2  ∆(2) 0 0
 
4
7.57 Consider an arbitrary triangular, plane elastic domain Ω of thickness h and made of
orthotropic material. Suppose that the body is free of body forces but subjected to
tractions on its sides, as shown in Fig. P7.57. Use Catigliano’s theorem I and derive
a relationship between the point displacements and the corresponding forces at the
vertices of the triangle.
Solution: The strain energy and potential energy due to applied loads are
Z I
h
U= σij εij dx, V = − t̂i ui ds, (1)
2 Ω Γ
Figure P7-57
x2 , u2 u3y
u2y
ux3
3
2 ux2
Ω 1
uy
Γ 1
u1x
x1 , u1
(a) (b)
Fig. P7.57 Triangular domain with vertex displacement components.
where Γ represents the collection of line segments enclosing the domain Ω and t̂i (s)
are the components of the boundary stresses. The strain-displacement relations in Eq.
(7.2.2) and stress–strain relations in Eq. (7.2.10) can be expressed in matrix form as
 ∂ 
  0
 ε11   ∂x1
 u1
∂ 
ε22 = 0 ∂x2  or {ε} = [D]{u}, (2)
  u2
2ε12

∂ ∂
∂x2 ∂x1
    
 σ11  c11 c12 0  ε11 
σ22 =  c12 c22 0  ε22 or {σ} = [C]{ε}. (3)
σ12 0 0 c66 2ε12
   
Now suppose that the displacements (ux , uy ) in the body Ω can be expressed (often,
it is an approximation) as a linear combination of unknown values of the displacement
vector at the vertices of the triangle and known functions of position ψj (x1 , x2 )
( P3 )
uj1 ψj (x1 , x2 )

ux (x1 , x2 ) j=1
= P3 j or {u} = [Ψ]{∆}, (4)
uy (x1 , x2 ) j=1 u2 ψj (x1 , x2 )
where uji denotes the value displacement component ui at the jth vertex of the domain
[see Fig. P7.57(b)], and

ψ 1 0 ψ2 0 ψ3 0
[Ψ] = (2 × 6),
0 ψ1 0 ψ2 0 ψ3 (5)
1 1 2 2 3 3 T
{∆} = u1 u2 u1 u2 u1 u2 (6 × 1).
Substituting (4) for {u} into Eqs. (2) and (3), we obtain
{ε} = [D]{u} ⇒ {ε} = [B]{∆}; {σ} = [C]{ε} ⇒ {σ} = [C][B]{∆}, (6)
where
∂ψ1 ∂ψ2 ∂ψ3
 
∂x1
0 ∂x1
0 ∂x1
0
∂ψ1 ∂ψ2 ∂ψ3
 
 (3 × 6).
 0
[B] = [D][Ψ] =  ∂x2
0 ∂x2
0 ∂x2  (7)
∂ψ1 ∂ψ1 ∂ψ2 ∂ψ2 ∂ψ3 ∂ψ3
∂x2 ∂x1 ∂x2 ∂x1 ∂x2 ∂x1
Substituting these expressions into Eq. (1), we obtain

Z I
h
U= {∆}T [B]T [C][B]{∆} dx, V = − {∆}T [Ψ]{t̂} ds. (8)
2 Ω Γ
Now applying Castigliano’s theorem II, we obtain

∂U ∂V
=− ⇒ [K]{∆} = {F }, (9)
∂{∆} ∂{∆}
where Z I
[K] = h [B]T [C][B] dx (6 × 6), {F } = [Ψ]T {t̂} ds (6 × 1) (10)
Ω Γ
Equation (9) provides the necessary algebraic equations to solve for the unknown dis-
placement components. However, the matrix [K], known as the stiffness matrix, is
singular to begin with. After imposing the necessary boundary conditions to elimi-
nate the rigid body translations and the rotation, the condensed matrix in nonsingular.
Equation (9) is not exact unless the representation in Eq. (4) is exact, which is most
often not the case. The procedure described in Eqs. (4)–(10) is nothing but the finite
element development for a typical domain Ω. This particular triangular element is
known as the constant strain triangle, because the functions ψi are linear in x1 and x2
for a triangle with 3 vertex points where the displacement degrees of freedom are iden-
tified. Consequently, the strains are constant within the domain Ω. For more details,
the reader may consult a finite element book.
7.58 Find a two-parameter Ritz approximation of the transverse deflection of a simply sup-
ported beam (constant EI) on an elastic foundation (modulus k) that is subjected to
a uniformly distributed load, q0 . Use (a) algebraic polynomials and (b) trigonometric
polynomials.
Solution: The weak form of the problem is given by

Z L
d2 w d2 δw

0= EI 2 + kwδw − q0 δw dx.
0 dx dx2
This gives
2
X
0= Bij cj − Ri
j=1
where
L L
d 2 φi d 2 φj
Z Z
Bij = EI + kφi φj dx, Ri = q0 φi dx.
0 dx2 dx2 0
(a) Let φ1 = x(L − x), φ2 = x2 (L − x). We obtain
4EIL + kL5 /30 2EIL2 + kL6 /60 q0 L3 2

c1
=
2EIL2 + kL6 /60 4EIL3 + kL7 /105 c2 12 L
The determinant of the coefficient matrix is equal to
k2 L12 11
∆ = 12(EI)2 L4 + + EIkL8 .
25200 105
The solution becomes
q0 L3 kL7 q0 L3

3
c1 = 6EIL + , c2 = (0).
12∆ 420 12∆
πx 3πx
(b) Let φ1 = sin L
and φ2 = sin L
. Then
h (2i−1)π i4 EIL kL
+ , if i = j, 2Lq0
Bij = L 2 2 , Ri = .
0, if i 6= j (2i − 1)π
Thus the constant ci are given by

Ri
ci = for fixed i.
Bii
7.59 Establish the total potential energy functional in Eq. (2) of Example 7.6.6.
Solution: The total potential energy of the torsion problem is

Z
1
Π(Ψ) = (σxz γxz + σyz γyz ) dx dy − T θ, (1)
2 Ω
where γxz = 2εxz = σxz /G and γyz = 2εyz = σyz /G. We can express the shear stresses
and the torque T in terms of Ψ using Eqs. ((7.5.73) and Eq. (6) of Example 7.5.9
Z " 2 2 # Z
1 ∂Ψ ∂Ψ
Π(Ψ) = + − dx dy − 2θ Ψ dx dy, (2)
2G Ω ∂y ∂y Ω
which is the required result (multiply throughout by G).
7.60 Determine a one-parameter Ritz approximation U1 (x) of u(x), which is governed by

the equation (like the equation governing the Prandtl stress function over square cross
section of 2 units)
∂2u ∂2u
− 2
+ = f0 in a unit square,
∂x ∂y 2
subjected to the boundary conditions
∂u ∂u

u(1, y) = u(x, 1) = 0, = = 0.
∂x (0,y) ∂y (x,0)

Take the origin of the coordinate system at the lower left corner of the unit square.
Solution: For the choice of φi , i = 1, 2,

φ1 = (x − 1)(y − 1), φ2 = (x2 − 1)(y 2 − 1),
∂φ1 ∂φ1 ∂φ2 ∂φ2
= (y − 1), = (x − 1); = 2x(y 2 − 1), = 2y(x2 − 1)
∂x ∂y ∂x ∂y
we obtain
Z 1Z 1
2
B11 = [(y − 1)2 + (x − 1)2 ]dx dy =
0 0 3
Z 1Z 1
5
B12 = [2x(y − 1)(y 2 − 1) + 2y(x − 1)(x2 − 1)]dx dy =
0 0 6
Z 1Z 1
64
B22 = 4x2 [(y 2 − 1)2 + 4y 2 (x2 − 1)2 ]dx dy =
0 0 45
Z 1Z 1
f0
R1 = f0 (x − 1)(y − 1)dx dy =
0 0 4
Z 1Z 1
4
R2 = f0 (x2 − 1)(y 2 −)dx dy = f0
0 0 9
or
1 60 75 c1 f0 9
= .
90 75 128 c2 36 16
The one-parameter solution with φ1 = (x − 1)(y − 1) is given by
3
U1 = c1 φ1 (x, y) = f0 (x − 1)(y − 1).
8
The one-parameter solution with φ2 is given by
5f0 2
U1 = c1 φ2 (x, y) = (x − 1)(y 2 − 1).
16
The two-parameter solution with φ1 and φ2 is given by
f0
16(x − 1)(y − 1) − 95(x2 − 1)(y 2 − 1) .

U2 (x) = c1 φ1 + c2 φ2 = −
274
If the trigonometric functions are used, the coefficients Bij and Ri are given by (a =
b = 1)  2
 αi if i = j,
2
Bij = ,
 α2i α2j 2 2 2
2
(αi − αj ) if i 6
= j,
f0
Ri = ,
αi2
where αi = (2i − 1)π/2. For example, the one-parameter solution is given by
32f0 πx πy
U1 = cos cos .
π4 2 2
7.61 Find Beltrami–Michell equations for dynamic elasticity.
Solution: We begin with Eq. (7.2.24)

1 ν
σij,kk + σkk,ij = σrs,rs δij + σkj,ki + σki,kj . (1)
1+ν 1−ν
Then we use the equations of motion to compute the second derivative of the stress
components, σrs,rk = ρ0 üs,k − ρ0 fs,k . We obtain

1 νρ0 νρ0
σij,kk + σkk,ij = ük,k − fk,k δij − ρ0 (fj,i + fi,j ) + ρ0 (üj,i + üi,j ) ,
1+ν 1−ν 1−ν
(2)
or in vector form
1 νρ0 h i
∇2 σ + ∇[∇ (tr σ)] = − (∇ · f ) I − ρ0 ∇f + (∇f )T
1+ν 1−ν
νρ0 h i
+ (∇ · ü) I + ρ0 ∇ü + (∇ü)T . (3)
1−ν
7.62 Extend Clapeyron’s Theorem to the dynamic case by starting with the expression
Z T
(U − K) dx,
0
where K is the kinetic energy.
Solution: We begin with the integral

Z T
1 T
Z Z
∂u ∂u
(U − K)dt = σ : ε − ρ0 · dxdt
0 2 0 Ω ∂t ∂t
1 T
Z Z
∂ui ∂ui
= σij εij − ρ0 dxdt
2 0 Ω ∂t ∂t
Z TZ
∂ 2 ui 1 T
Z I
1
= ρ0 2 − σij,j ui dx + nj σij ui ds
2 0 Ω ∂t 2 0 Γ
I T
1 ∂ui
− ρ0 ui dx ,
2 Ω ∂t 0
where integration-by-parts with respect to time is performed. If we assume that ui = 0

at t = 0 and t = T , we obtain
Z T
1 T
Z Z I
(U − K)dt = fi ui dx + ti ui ds dt.
0 2 0 Ω Γ
Thus, the time integral of the difference of the strain energy and the kinetic energy is
equal to one-half of the time integral of the work done by external forces, provided the
displacements are zero at initial and final times.
Figure P7-63
7.63 Consider a pendulum of mass m1 with a flexible suspension, as shown in Figure P7.63.
The hinge of the pendulum is in a block of mass m2 , which can move up and down
between the frictionless guides. The block is connected by a linear spring (of spring
constant k) to an immovable support. The coordinate x is measured from the position of
the block in which the system remains stationary. Derive the Euler–Lagrange equations
of motion for the system.
Unstretched length
of the spring
k
x
m2
θ l
m1
Fig. P7.63
Solution: The Lagrangian function is

1 h i 1
L = m1 l2 θ̇2 + ẋ2 − 2lẋθ̇ sin θ + m2 ẋ2
2 2
1
+ m1 g(x − l cos θ) + m2 gx + k(x + h)2
2
where h is the elongation in the spring due to the masses
g
h= (m1 + m2 ).
k
7.64 A chain of total length L and mass m per unit length slides down from the edge of a
smooth table. Assuming that the chain is rigid, find the equation of motion governing
the chain (see Example 5.3.3).
Solution: Let x the length of the chain hanging from the edge of the table. Then the
kinetic energy associated with the chain is
1
K= ρlẋ2 .
2
The potential energy is
1
V = − ρgx2 .
2
Thus, the Lagrangian is L = K − V . According to Hamilton’s principle, we have the
Euler–Lagrange equation

d ∂L ∂L g
− = 0 → ρlẍ − ρx or ẍ − x = 0.
dt ∂ ẋ ∂x l
7.65 Consider a cantilever beam supporting a lumped mass M at its end (J is the mass
moment of inertia), as shown in Figure P7.65. Derive the equations of motion and
natural boundary conditions for the problem using the Euler–Bernoulli beam theory.
Solution: The kinetic energy expression in Eq. (2) of Example 7.7.2 for this case must
be modified to include the effect of the lumped mass, M . The additional terms are
Z ( 2 )
1 T

2 ∂ v̇
M [v̇(L)] + J dt. (1)
2 0 ∂x x=L
Figure P7-65
The equation of motion remains unchanged:
∂2 ∂2v ∂2 ∂2v

∂ ∂v
ρA + EI − ρI = q. (2)
∂t ∂t ∂x2 ∂x2 ∂x∂t ∂x∂t
x
M, J
Fig. P7.65
The natural boundary conditions become

∂3v ∂2v ∂2v

∂
ρI 2
+M 2 − EI 2 = 0, at x = L (3)
∂x∂t ∂t ∂x ∂x
∂2v ∂3v
EI +J = 0, at x = L. (4)
Figure P7-66 ∂x 2 ∂x∂t2
7.66 Derive the equations of motion of the system shown in Figure P7.35. Assume that the
mass moment of inertia of the link about its mass center is J = mΩ2 , where Ω is the
radius of gyration.
x0 x
x0 = unstretched
length
°
k l x
θ °
l
y
° F
mg
Fig. P7.66
Solution: We choose x and θ as the generalized coordinates, where x is measured from

the static unstretched length of the spring, and θ from the vertical. The kinetic and
potential energies are
2 2
m d m d J
K= (x + ` sin θ) + (` cos θ) + (θ̇)2
2 dt 2 dt 2
mh 2 2 2 2 2
i
= ẋ + 2`ẋθ̇ cos θ + ` (θ̇) + Ω (θ̇)
2
1
V = mg`(1 − cos θ) + kx2 .
2
The equations of motion are

d ∂L ∂L
− + = 0, m(ẍ + `θ̈ cos θ − `θ̇2 sin θ) + kx = F,
dt ∂ ẋ ∂x

d ∂L ∂L h i
− + = 0, m `ẍ cos θ + (`2 + Ω2 )θ̈ + mg` sin θ = 2aF cos θ.
dt ∂ θ̇ ∂θ
7.67 Derive the equations of motion of the Timoshenko beam theory, starting with the dis-
placement field (including the axial displacement, u):
u1 (x, y, t) = u(x, t) + yφ(x, t), u2 = v(x, t), u3 = 0.
Assume that the beam is subjected to distributed axial load f (x, t) and transverse load
q(x, t), and that the x-axis coincides with the geometric centroid.
Solution: From the given displacement field, we can easily compute the strains. We
have
∂u ∂u ∂φ ∂u ∂v ∂v
εxx = = +y , 2εxy = + =φ+ . (1)
∂x ∂x ∂x ∂y ∂x ∂x
All other strains are zero.
Hamilton’s principle for the case at hand is
Z T
− (δU + δV − δK) dt = 0 (2)
0
where δU is the virtual strain energy, δV the virtual work done by applied forces, and
δK is the virtual kinetic energy:
Z
δU = (σxx · δεxx + σxy · 2δεxy )dx
Ω
Z LZ
∂δu ∂δφ ∂δv
= σxx +y + σxy δφ + dAdx
0 A ∂x ∂x ∂x
Z L
= Nxx + Mxx + Qx δφ + dx (3)
0 ∂x ∂x ∂x
Z L
δV = − (f δu + qδv) dx (4)
0
Z Z L Z h i
δK = ρ (u̇δ u̇ + u̇δ u̇) dx = ρ u̇ + y φ̇ δ u̇ + yδ φ̇ + v̇δ v̇ dAdx
V 0 A
Z L
= m0 u̇δ u̇ + m2 φ̇δ φ̇ + m0 v̇δ v̇ dx, (5)
0
where
Z Z
(Nxx , Mxx , Qx ) = (σxx , yσxx , σxy ) dA, (m0 , m2 ) = ρ(1, y 2 )dA. (6)
A A
Substituting the expressions for δU , δV , and δK from Eqs. (3)-(5) into Eq. (2), we
obtain
Z T Z L
0= Nxx + Mxx + Qx δφ + − qδv
0 0 ∂x ∂x ∂x

− m0 u̇δ u̇ + m2 φ̇δ φ̇ + m0 v̇δ v̇ dx dt (7)
Integrating by parts in x and time t to relieve δu, δφ, and δv of any derivatives (with
respect to x or t), we arrive at
Z T Z L
∂Nxx ∂Mxx ∂Qx
0= − δu − δφ + Qx δφ − δv − f δu − qδv
0 0 ∂x ∂x ∂x

∂ ∂ ∂
+ m2 φ̇ δφ + (m0 u̇) δu + (m0 ẇ) δv dx dt
∂t ∂t ∂t
Z T Z T
+ [Nxx · δu]L
0 dt + [Mxx · δv + Qx · δφ]L 0 dt
0 0
Z L h iT
+ m0 φ̇ · δφ + m2 v̇ · δv dx (8)
0 0
The last expression is zero by virtue of the assumption that δu = 0, δφ = 0 and δv = 0

at t = 0 and t = T . The Euler–Lagrange equations are

∂Nxx ∂ ∂u
δu : − −f + m0 = 0,
∂x ∂t ∂t

∂Qx ∂ ∂w
δv : − −q+ m0 = 0, (9)
∂x ∂t ∂t

∂Mxx ∂ ∂φ
δφ : − + Qx + m2 = 0.
∂x ∂t ∂t
The boundary expressions indicate that u, v, and φ are the primary variables and
Nxx , Mxx , and Qx are secondary variables of the problem. Thus the natural boundary
conditions involve specifying the stress resultants Nxx , Mxx and Qx , which can be
expressed in terms of the displacements (u, v, φ) as
Z Z Z
∂u ∂φ ∂u
Nxx = σxx dA = Eεxx dA = E +y dA = EA ,
A A A ∂x ∂x ∂x
Z Z Z
∂u ∂φ ∂φ
Mxx = yσxx dA = Eyεxx dA = E y +y dA = EI , (10)
A A A ∂x ∂x ∂x
Z Z Z
∂v ∂v
Qx = σxy dA = 2Gεxy dA = G φ+ dA = GA φ + .
A A A ∂x ∂x
7.68 Derive the equations of motion of the third-order Reddy beam theory based on the
displacement field
∂v
u1 (x, y, t) = u(x, t) + yφ(x, t) − c1 y 3 φ +
∂x (1)
u2 (x, y, t) = v(x, t), u3 = 0,
where c1 = 4/(3h2 ). Assume that the beam is subjected to distributed axial load f (x, t)
and transverse load q(x, t), and that the x-axis coincides with the geometric centroid.
Solution: Because we are primarily interested in deriving the equations of motion and
the nature of the boundary conditions of the beam that experiences a displacement
field of the form in Eq. (1), we will not consider specific geometric or force bound-
ary conditions here. The procedure to obtain the equations of motion and boundary
conditions involves the following steps: (i) compute the strains, (ii) compute the vir-
tual energies required in Hamilton’s principle, and (iii) use Hamilton’s principle and
derive the Euler-Lagrange equations of motion and identify the primary and secondary
variables of the theory, which in turn help identify the form of the boundary conditions.
The linear strains associated with the displacement field are
εxx = ε(0) (1) 3 (3)

xx + yεxx + y εxx ,
(0)
(2)
γxy = γxy + y 2 γxy
(2)
,
where
∂2v

∂u ∂φ ∂φ
ε(0)
xx = (1)
, εxx = (3)
, εxx = −c1 + ,
∂x ∂x ∂x ∂x2
(3)
(0) ∂v (2) ∂v
γxy =φ+ , γxy = −c2 φ + ,
∂x ∂x
and c2 = 4/h2 . Note that γxy = 2εxy is a quadratic function of y. Hence, σxy = Gγxy
is also quadratic in y.
From Hamilton’s principle for deformable bodies we have
Z T Z LZ h i
0= σxx δε(0) (1) 3 (3)
xx + yδεxx + y δεxx
(0)
+ σxy δγxy + y 2 δγxy
(2)
dAdxdt
0 0 A
(
Z T Z L Z
∂ v̇ ∂δ v̇
− ρ u̇ + y φ̇ − c1 y 3 φ̇ + δ u̇ + yδ φ̇ − c1 y 3 δ φ̇ +
0 0 A ∂x ∂x
) Z TZ L
+ v̇δ v̇ dAdx dt − qδv dx dt
0 0
Z T Z L

= Nxx δε(0) (1) (3)
xx + Mxx δεxx + Pxx δεxx + Qx δγxy + Rx δγxy
(0) (2)
dxdt
0 0
Z T Z L
∂ v̇
− I0 u̇δ u̇ + I2 φ̇ − c1 I4 φ̇ + δ φ̇ + qδv dx dt
0 0 ∂x
Z T Z L
∂ v̇ ∂δ v̇
− −c1 I4 φ̇ − c1 I6 φ̇ + δ φ̇ + + I0 v̇δ v̇ dx dt
0 0 ∂x ∂x
Z T Z L
∂2u ∂2φ ∂3v

∂Nxx ∂ M̄xx
= − + I0 2 δu + − + Q̄x + K2 2 − c1 J4 δφ
0 0 ∂x ∂t ∂x ∂t ∂x∂t2
# )
∂ 2 Pxx ∂3φ ∂4v ∂2v

∂ Q̄x
+ −c1 − − q + c 1 J 4 − c I
1 6 + I0 δv dx dt
∂x2 ∂x ∂x∂t2 ∂x2 ∂t2 ∂t2
Z T(
∂δv
+ Nxx δu + M̄xx δφ − c1 Pxx
0 ∂x
)L
∂2φ ∂3v

∂Pxx
+ Q̄x + c1 − J4 2 + c1 I6 δv dt,
∂x ∂t ∂x∂t2
0
(4)
where all the terms involving [ · ]T0 vanish on account of the assumption that all varia-
tions and their derivatives are zero at t = 0 and t = T , and the new variables introduced
in arriving at the last expression are defined as follows:
   
 Nxx  Z h/2  1  Z h/2
Qx 1
Mxx = y σxx dy , = σxz dy, (5)
−h/2  3 Rx −h/2 y2
Pxx y
 
4 4
M̄xx = Mxx − c1 Pxx , Q̄x = Qx − c2 Rx , c1 = , c2 = 2 , (6)
3h2 h
Z h/2
J4 = I4 − c1 I6 , K2 = I2 − 2c1 I4 + c21 I6 , Ii = ρ(y)i dy. (7)
−h/2
Note that Ii are zero for odd values of i (that is, I1 = I3 = I5 = 0). Thus, the
Euler–Lagrange equations are
∂Nxx ∂2u
δu : = I0 2 , (8)
∂x ∂t
∂ 2 Pxx ∂2v ∂3φ ∂4v

∂ Q̄x
δv : + c1 + q = I0 + c1 J4 − c I
1 6 , (9)
∂x ∂x2 ∂t2 ∂x∂t2 ∂x2 ∂t2
∂ M̄xx ∂2φ ∂3v
δφ : − Q̄x = K2 2 − c1 J4 . (10)
∂x ∂t ∂x∂t2
The last line of Eq. (4) include the boundary terms, and they show that the primary
variables of the theory are (those with the variational symbol) u, v, φ, and ∂w/∂x. The
corresponding secondary variables are the coefficients of δu, δv, δφ, and ∂δw/∂x:
∂2φ ∂3v

∂Pxx
Nxx , Q̄x + c1 − J4 2 + c1 I6 , M̄xx , −c1 Pxx . (11)
∂x ∂t ∂x∂t2
7.69 Consider a uniform cross-sectional bar of length L, with the left end fixed and the right
end connected to a rigid support via a linear elastic spring (with spring constant k), as
shown in Fig. P7.69. Determine the first two natural axial frequencies of the bar using
the Ritz method. Hint: The kinetic energy K and the strain energy U associated with
the axial motion of the member are given by
Z L 2 Z L 2
ρA ∂u EA ∂u k
K= dx, U = dx + [u(L, t)]2 . (1)
0 2 ∂t 0 2 ∂x 2
Use Hamilton’s principle to obtain the variational equation, and for periodic motion
assume √
u(x, t) = u0 (x)eiωt , i = −1, (2)
where ω is the frequency of natural vibration, and u0 (x) is the amplitude, to reduce
Figure P7-70
the variational statement to
Z L
du0 dδu0
0= ρAω 2 u0 δu0 − EA dx − ku0 (L)δu0 (L). (3)
0 dx dx
E, A
k
L
x Deformable bar
Fig. P7.69
Solution: Substituting for K and U from Eq. (1), and V = 0 in Hamilton’s principle,
we obtain [δu(x, t1 ) = δu(x, t2 ) = 0 and δu(0, t) = 0]
Z t2
0= δ(K − U )dt
t1
(Z " 2 # 2 )
Z t2 L
1 ∂u ∂u
= δ ρA dx − k[u(L, t)]2 dt
− EA (4)
2 t1 0 ∂x ∂t
"Z ! #
t2
Z L
∂u ∂δu ∂u ∂δu
= ρA − EA dx − ku(L, t)δu(L, t) dt
t1 0 ∂t ∂t ∂x ∂x
Z t2 "Z L #
∂2u

∂u ∂δu
= −ρA 2 δu − EA dx − ku(L, t)δu(L, t) dt. (5)
t1 0 ∂t ∂x ∂x
Substituting Eq. (2) into Eq. (5), we obtain

Z L
du0 dδu0
0= ρAω 2 u0 δu0 − EA dx − ku0 (L)δu0 (L), (6)
0 dx dx
Rt
where (iω)2 = −ω 2 , and t12 e2iωt dt, being nonzero, is factored out. We use Eq. (6) to
determine the values of ω.
The Euler equation and natural boundary condition associated with Eq. (6) are
d du0
− EA − ρAω 2 u0 = 0, 0 < x < L,
dx dx (7)
du0
EA + ku0 = 0 at x = L.
dx
The essential boundary condition is u0 (0) = 0.
A nondimensionalization of the variables is used for simplicity:
x u0 kL ω 2 ρL2
x̄ = , ū = , α= , λ= . (8)
L L EA E
Then Eq. (6) becomes

Z 1
dū dδ ū
0= λūδ ū − dx − αū(1)δ ū(1). (9)
0 dx dx
The bar over the nondimensional variables will be omitted in the interest of brevity.
Further, in the following discussions, we shall assume that α = 1.
Substituting a N −parameter Ritz approximation (obviously, we have φ0 = 0)
N
X
ū(x) ≈ UN (x) = ci φi (x)
i=1
into Eq. (9), we obtain

N
(N Z )
1 Z 1
X X dφi dφj
0= λ φi φj dx − dx + αφi (1)φj (1) cj δci .
i=1 j=1 0 0 dx dx
Because of the independent nature of δci , we obtain

N Z 1 Z 1
X dφi dφj
0= λ φi φj dx − dx + αφi (1)φj (1) cj , (10a)
j=1 0 0 dx dx
and in matrix form

([A] − λ[M ]) {c} = {0}, (10b)
where Z 1 Z 1
dφi dφj
aij = dx + αφi (1)φj (1), mij = φi φj dx. (10c)
0 dx dx 0
Equation (7.80b) represents a matrix eigenvalue problem, and we obtain N eigenvalues,
λi , i = 1, 2, · · · , N .
For the problem at hand, the approximation functions can be taken as
φi (x) = xi . (11a)
i
Substituting φi = x into Eq. (10c), we obtain
Z 1
1
mij = φi φj dx = ,
0 i+j+1
Z 1
dφi dφj ij
aij = dx + φi (1)φj (1) = + 1. (11b)
0 dx dx i + j−1
Because we wish to determine two eigenvalues, we take N = 2 and obtain
1 1 1 7
m11 = , m12 = , m22 = , a11 = 2, a12 = 2, a22 = ,
3 4 5 3
and the matrix eigenvalue problem (10b) becomes
1 1
2 2 c1 0
7 − λ 13 41 = . (12)
2 3 4 5
c2 0
For a nontrivial solution (that is, c1 6= 0, c2 6= 0), the determinant of the coefficient
matrix in Eq. (12) is set to zero:

2− λ 2− λ
3 4 = 0,
2− λ 7 − λ
4 3 5
or
15λ2 − 640λ + 2400 = 0.
The quadratic equation has two roots
r r
2.038 E 6.206 E
λ1 = 4.1545, λ2 = 38.512 → ω1 = , ω2 = . (13)
L ρ L ρ
The eigenvectors are given by

(i) (i) (i)
U2 = c1 x + c2 x2
(i) (i)
where c1 and c2 are calculated from Eq. (12) for λ = λi , i = 1, 2.
The exact values of λ are the roots of the transcendental equation
λ + tan λ = 0, (14)
whose first two roots are (ω 2 = λ)

r r
2.02875 E 4.91318 E
ω01 = , ω02 = . (15)
L ρ L ρ
Note that the first approximate frequency is closer to the exact than the second.
7.70 Consider the equation
−∇2 u = λu, in Ω, u = 0 on Γ, (1)
where Ω is the triangular domain shown in Fig. P7.70 and Γ is its boundary. Equation
(1) describes a nondimensional form of the equation governing the natural vibration
of a triangular membrane of side a; mass density ρ; and tension T (λ = ρa2 ω 2 /T ,
Figure
ω beingP7-69
the natural frequency of vibration). Determine the fundamental frequency
(that is, determine λ) of vibration by using a one-parameter Ritz approximation of the
problem.
y
(0,1)
G
W
x
(1,0)
Fig. P7.70
Solution: The Ritz method is based on the variational statement

Z
∂φ1 ∂φ1 ∂φ1 ∂φ1
0= c1 + − λφ1 φ1 dx dy. (2)
Ω ∂x ∂x ∂y ∂y
The function φ1 (x, y) must vanish on the boundary Γ. Thus, we have
φ1 (x, 0) = 0, φ1 (0, y) = 0, φ1 (x, y) = 0 on the line x + y − 1 = 0. (3)
Hence, the choice for φ1 (x, y) is

∂φ1 ∂φ1
φ1 (x, y) = (x − 0)(y − 0)(x + y − 1), = y(2x + y − 1) = x(x + 2y − 1). (4)
∂x ∂y
Hence, we have
R 1 R 1−y ∂φ1 ∂φ1 ∂φ1 ∂φ1

0 0 ∂x ∂x
+ ∂y ∂y
dx dy
λ= R 1 R 1−y . (5)
0 0
φ1 φ1 dx dy
N7.1 –7.4 Problems on the determination of the form of the Airy stress function or given
the Airy stress function and determination of the constants in the Airy stress function
for the following beam problems are suggested. Cases (a), (c), (d) , and (e) are already
in the book in some form or the other.
y y
L L
P
x x
F0 F0
(a) (b)
y q0 y æxö
q0 çç ÷÷÷
çè L ø
A A
x x
L L
(c) (d)
y y æxö
t0 çç ÷÷÷
τ0 çè L ø
L L
x x
L L
(e) (f)
Φa (x, y) = Axy + Bxy 2 + Cxy 3 ,

Φb (x, y) = Axy + Bxy 3 + Cy 2 ,
Φc (x, y) = x2 (A + By + Cy 3 ) + Dy 3 + Dy 5 ,
Φd (x, y) = Axy + B 3 + Cx3 y + Dxy 3 + Ex3 y 3 + F xy 5 ,
Φe (x, y) = Ay 2 + By 3 + Cxy + Dxy 2 + Exy 3 ,
Φf (x, y) = Ay 2 + By 3 + Cy 4 + Dy 5 + Ex2 + F x2 y + Gx2 y 2 + Hx2 y 3 .
N7.5 Consider the equation
−∇2 u = λu, in Ω, u = 0 on Γ, (1)
where Ω is the triangular domain shown in the figure below, and Γ is its boundary.
Equation (1) describes a non-dimensional form of the equation governing the natural
vibration of a triangular membrane of side a, mass density ρ and tension T (λ =
ρa2 ω 2 /T , ω the natural frequency of vibration). Determine the fundamental frequency
(that is, determine λ) of vibration by using a one-parameter Ritz approximation of the
problem.
Figure P7-48
T 3a
Side 1
Side 3
x
x = -2a
Side 2
x=a
Solution: The choice of the function φ1 (x, y) is dictated by satisfying the requirement
that it be zero on the boundary of the triangle. Therefore, we choose
u(x, y) ≈ U1 (x, y) = c1 x2 + 4a2 + 4ax − 3y 2 (x − a)

= c1 x3 + 3ax2 − 3xy 2 − 4a3 + 3ay 2 ≡ c1 φ1 (x, y).

(1)
The variational problem associated with Eq. (1) is to find u such that
A(u, v) = λB(u, v) (2)
holds for all v, where

Z Z
∂u ∂v ∂u ∂v
A(u, v) = + dx dy, B(u, v) = uv dx dy. (3)
Ω ∂x ∂x ∂y ∂y Ω
The Ritz method gives the algebraic equation

A11
A11 − λB11 = 0 or λ = , (4)
B11
where Z Z
∂φ1 ∂φ1 ∂φ1 ∂φ1
A11 = + dx dy, B11 = φ1 φ1 dx dy. (5)
Ω ∂x ∂x ∂y ∂y Ω
N7.6 Use Castigliano’s theorem I to determine the slope (rotation) at the free end of a
cantilever beam (EI constant) with uniformly distributed load of intensity, q0 .
Solution: The problem can be solved using the procedure presented in Examples 7.6.2
and 7.6.3. We have
      
6 −3L −6 −3L  ∆1   6   Q1 
2 2
     
 −3L 2L 3L L  ∆2 = − q0 L −L + Q2 .
2EI        
(1)
L3  −6 3L 6 3L  
 ∆3 
 12 
 6   Q3 
2 2   
−3L L 3L 2L ∆4 L Q4
    
Because ∆1 = ∆2 = 0 and Q3 = Q4 = 0, we can write

2EI 6 3L ∆3 q0 L 6
2 = − , (2)
L3 3L 2L ∆4 12 L

6 6

q0 L L3 3L L q L3
∆4 = − = 0 . (3)
12 2EI 6 3L 6EI
3L 2L2
To use Castigliano’s theorem I (not included in this text), we assume a fictitious moment
M0 at the free end, and require
1 h L
Z
∂U ∂M
i
θ0 = = M dx
∂M0 M0 =0 EI ∂M0

M0 =0
1 h L q0 x2
Z i
= − − M0 (−1) dx
EI 2 M0 =0
Z L 2 3
1 q0 x q0 L
= dx = . (4)
EI 2 6EI
N7.7 Find the two-parameter Ritz approximation to determine the transverse deflection of a
cantilever beam (EI = constant) subjected to uniformly distributed load of intensity,
q0 .
Solution: The minimum total potential energy principle requires

Z L
d2 δw d2 w

δΠ(w) ≡ EI − q 0 δw dx = 0.
0 dx2 dx2
A two-parameter Ritz approximation of the type (φ0 = 0) W2 (x) = c1 x2 + c2 x3 yields

the Ritz equations
2 L L
d 2 φi d 2 φi
X Z Z
Bij cj = Ri , Bij = B(φi , φj ) = EI dx, Ri = L(φi ) = q0 φi dx.
j=1 0 dx2 dx2 0
The resulting algebraic equations are
q0 L2

4 6L c1 4
EIL 2 =
6L 12L c2 12 3L
whose solution is
q0 L4
2
x3 q0 L4

x
W2 (x) = 5 2 − 2 3 , W2 (L) = .
24EI L L 8EI
The bending moment computed from the approximate deflection is
q0 L2 x 5q0 L2
M2 (x) = 5− , M2 (0) =
12 L 12
205
Chapter 8: FLUID MECHANICS AND HEAT TRANSFER

PROBLEMS
8.1 Assume that the velocity components in an incompressible flow are independent of the
x coordinate and vz = 0 to simplify the continuity equation (8.2.5) and the equations
of motion (8.2.6)–(8.2.8).
Solution: We have
∂vy
= 0, (1)
∂y
∂ 2 vx ∂ 2 vx

∂p ∂vx ∂vx
µ 2
+ − + ρfx = ρ + vy (2)
∂y ∂z 2 ∂x ∂t ∂y
∂ 2 vy ∂ 2 vy

∂p ∂vy ∂vy
µ + − + ρfy = ρ + vy (3)
∂y 2 ∂z 2 ∂y ∂t ∂y
∂p
− + ρfz = 0 (4)
∂z
These equations further simplify, due to the condition ∂vy /∂y = 0, to
2
∂ 2 vx

∂ vx ∂p ∂vx ∂vx
µ + − + ρfx = ρ + vy , (5)
∂y 2 ∂z 2 ∂x ∂t ∂y
∂ 2 vy ∂p ∂vy ∂p
µ − + ρfy = ρ , − + ρfz = 0. (6)
∂z 2 ∂y ∂t ∂z
8.2 An engineer is to design a sea lab 4 m high, 5 m wide, and 10 m long to withstand
submersion to 120 m, measured from the surface of the sea to the top of the sea lab.
Determine the (a) pressure on the top and (b) pressure variation on the side of the
cubic structure. Assume the density of salt water to be ρ = 1020 kg/m3 .
Solution: The weight density of the salt water is
γ = ρg = 1, 020 × 9.81 = 10 kN/m3 .
(a) The pressure at the top of the sea lab (that is, at depth h = 120 m) is
p = γh = 1.2MN/m2 .
Figure P8-2
(b) The pressure variation, measured from the top of the lab downward, is (Pa=N/m2
and N= kg-m/s2 )
p(z) = γ(120 + z) = 10(120 + z) kPa, p(4) = 1.24 MPa,
where the coordinate z is measured from the top of the sea lab downward.
h = 120 m
4m
z
Fig. P8.2
8.3 Compute the pressure and density at an elevation of 1600 m for isothermal conditions.
Assume p0 = 102 kPa and ρ0 = 1.24 kg/m3 at sea level.
Solution: The pressure can be calculated from Eq. (8.2.37) (Pa=N/m2 and N= kg-
m/s2 )

z − z0 1, 600
p = p0 exp − = 105 exp − = 82.314 kPa abs.
(p0 /ρ0 g) (105 /1.24 × 9.81)
The density is calculated using Eq. (8.2.36)

p 82.314
ρ = ρ0 = 1.24 = 1.02 kg/m3 .
p0 102
8.4 Derive the pressure–temperature and density–temperature relations for an ideal gas
when the temperature varies according to θ(x3 ) = θ0 + mx3 , where m is taken to be
m = −0.0065◦ C/m up to the stratosphere, and x3 is measured upward from sea level.
Hint: Use Eq. (8.2.35) and the third equation in Eq. (8.2.32).
Solution: we have
pg Pg
dP = −ρgdx3 = − dx3 = − dx3 ,
Rθ R(θ0 + mx3 )
or
dp g dx3
=− .
Figure P8-5 p R (θ0 + mx3 )
Integrating both sides gives
Z p
g x3
Z
dp dx3 p g mx3
=− → log =− log 1 + ,
p0 p R 0 (θ0 + mx3 ) p0 mR θ0
or −g/mR −g/mR
mx3 mx3
p = p0 1 + , ρ = ρ0 1 + .
θ0 θ0
8.5 Consider the steady flow of a viscous incompressible Newtonian fluid down an inclined
surface of slope α under the action of gravity (see Fig. P8.5). The thickness of the fluid
perpendicular to the plane is h and the pressure on the free surface is p0 , a constant.
Use the semi-inverse method (that is, assume the form of the velocity field) to determine
the pressure and velocity field.
y
vx
h
x
Direction of gravity, ρ g α
Fig. P8.5
Solution: We assume that the velocity field of the form (semi-inverse method)
vx = U (x, y), vy = 0, vz = 0. (1)

CHAPTER 8: FLUID MECHANICS AND HEAT TRANSFER PROBLEMS 207
The body force vector is

ρf = ρg (sin αêx − cos αêy ) . (2)
The governing equations are simplified as follows:
Continuity equation:
∂vx ∂vy ∂vz ∂vx
+ + = + 0 + 0 = 0 → vx = U = U (y). (3)
∂x ∂y ∂z ∂x
Note that the convective terms in all three momentum equations are identically zero:
∂vx ∂vx ∂vx
vx + vy + vz = 0 + 0 + 0 = 0,
∂x ∂y ∂z
∂vy ∂vy ∂vy
vx + vy + vz = 0 + 0 + 0 = 0,
∂x ∂y ∂z
∂vz ∂vz ∂vz
vx + vy + vz =0+0+0=0
∂x ∂y ∂z
x-momentum equation:
∂p d2 U
− + µ 2 + ρg sin α = 0. (4)
∂x dy
y-momentum equation:
∂p
− − ρg cos α = 0. (5)
∂y
z-momentum equation:
∂p
− = 0. (6)
∂z
Equation (6) implies that p = p(x, y). Equation (5) gives the pressure
p(x, y) = −ρg cos α y + f (x). (7)
The condition p(x, h) = p0 yields f (x) = p0 + ρgh cos α. Thus, the pressure is
y
p(y) = p0 + ρgh cos α 1 − . (8)
h
Equation (4) yields
ρg sin α y 2
U (y) = − + Ay + B,
µ 2
where the constants A and B are determined using the boundary conditions

dU
U (0) = 0, σxy (h) = µ = 0. (9)
dy y=h
We obtain B = 0 and A = ρgh sin α/µ. Thus the velocity field is
ρgh2 sin α y2

y
U (y) = 2 − 2 . (10)
2µ h h
8.6 Two immiscible fluids are flowing in the x-direction in a horizontal channel of length
L and width 2b under the influence of a fixed pressure gradient. The fluid rates are
adjusted such that the channel is half filled with fluid I (denser phase) and half filled
with fluid II (less dense phase). Assuming that the gravity of the fluids is negligible,
determine the velocity field. Use the geometry and coordinate system shown in Fig.
P8.6.
Figure P8-6
Assume steady flow

Fixed wall
Less dense and
b y μ2 less viscous fluid
x Interface
b μ1 Denser and
more viscous fluid
Fixed wall
Fig. P8.6
Solution: Assume velocity field of the form
vx = U (x, y), vy = 0, vz = 0. (1)
The governing equations are simplified as follows (see the solution to Problem 2 for the
details of how the convective terms are zero):
Continuity equation:
∂U
= 0 → U = U (y). (2)
∂x
x-momentum equation:
∂p d2 U
− + µ 2 = 0. (3)
∂x dy
y-momentum equation:
∂P
− + ρg = 0. (4)
∂y
z-momentum equation:
∂p
= 0. − (5)
∂z
Equations (4) and (5) imply that P = P (x, y). If gravity is neglected, p = p(x) and
then Eq. (3) implies that dp/dx is a constant. Equation (3) is valid in −b < y < b but
µ has different values in the lower half and upper half of the channel. Integrating Eq.
(3) in the two regions
dU1 dp (1)
µ1 = y + A1 = σxy ,
dy dx
(6)
dU2 dp (2)
µ2 = y + A2 = σxy ,
dy dx
dp y 2 A1
U1 (y) = + y + B1 ,
dx 2µ1 µ1
(7)
dp y 2 A2
U2 (y) = + y + B2 ,
dx 2 µ2
where the constants A1 , A2 , B1 , and B2 are determined using the boundary conditions
U1 (−b) = 0, U2 (b) = 0. (8)
and interface conditions

(1) (2)
U1 (0) = U2 (0), σxy (0) = σxy (0). (9)
We obtain
(1) (2)
σxy (0) = σxy (0), → A1 = A2 ≡ A,
U1 (0) = U2 (0), → B1 = B2 ≡ B,
dp b2 A
U1 (−b) = 0, → − b + B = 0, (10)
dx 2µ1 µ1
dp b2 A
U2 (b) = 0, → + b + B = 0, (11)
dx 2µ2 µ2
Solving the equations, we obtain
b2

dp dp b µ1 − µ2
B=− , A=− . (12)
dx µ1 + µ2 dx 2 µ1 + µ2
Figure P8-7
8.7 Consider the steady flow of a viscous, incompressible fluid in the annular region between
two coaxial circular cylinders of radii R0 and αR0 , α < 1, as shown in Fig. P8.7. Take
p̄ = p + ρgz. Determine the velocity and shear stress distributions in annulus.
Velocity
distribution
α R0
R0
Fig. P8.7
Solution: Because the cylinder is long enough to neglect the end effects. Due to axisym-
metry of the problem, we have vr = vθ = 0 and vz = vz (r). The continuity equation
(8.2.9) is identically satisfied. The momentum equations in the r and θ directions give
∂p ∂P
= 0, = 0, (1)
∂r ∂θ
which give imply that p = p(z). The z-momentum equation takes the form

1 d dvz dp
µ r − − ρg = 0, (2)
r dr dr dz
where ρg is the gravitational force. Integrating Eq. (2) with respect to r gives
dvz dp̄ r2 dp̄ r2

µr = + c1 , µvz (r) = + c1 log r + c2 ,
dr dz 2 dz 4
where c1 and c2 are constants of integration. The boundary conditions of the problem
are
vz (R0 ) = 0 and vz (αR0 ) = 0.
Thus, we have
dp̄ R02 dp̄ α2 R02

+ c1 log R0 + c2 = 0, + c1 log(αR0 ) + c2 = 0.
dz 4 dz 4
Solving for c1 and c2 , we obtain
R02 dp̄ log R0 R02 dp̄

c1 log α = 1 − α2 , c2 = − 1 + (1 − α2 ) .
4 dz log α 4 dz
The velocity field becomes

" 2 #
dp̄ R02 (1 − α2 )

r r
vz (r) = − 1− − log
dz 4µ R0 log α R0
The shear stress is given by

dvz dp̄ r 1 dp̄ R0 r 2 1 R0
τrz = −µ =− + c1 = − 2 + (1 − α ) .
dr dz 2 r dz 4 R0 log α r
8.8 Consider a steady, isothermal, incompressible fluid flowing between two vertical con-
centric long circular cylinders with radii R1 and R2 . If the outer one rotating with
an angular velocity Ω, show that the Navier–Stokes equations reduce to the following
equations governing the circumferential velocity vθ and pressure P :
FigP8-8 v2 ∂p

d 1 d

∂p
ρ θ = , µ (rvθ ) = 0, 0 = − + ρg.
r ∂r dr r dr ∂z
Determine the velocity vθ and shear stress τrθ distributions.
Angular velocity of
the outer cylinder, Ω
Stationary
inner cylinder
+
αR
R
Fig. P8.8
Solution: Because the outer cylinder is rotating, the fluid moves in a circular motion
and therefore the velocity components vr and vz are zero and vθ is only a function of r.
Also, there is no pressure gradient in the θ-direction. The continuity equation (8.2.9)
is identically satisfied. The momentum equations (8.2.10)–(8.2.12) reduce to
v2

∂p d 1 d ∂p
− = −ρ θ , µ (rvθ ) = 0, − + ρg = 0.
∂r r dr r dr ∂z
First, we solve the second momentum equation for vθ :

1 d r 1
(rvθ ) = c1 , vθ (r) = c1 + c2 ,
r dr 2 r
where c1 and c2 are constants of integration. The boundary conditions are
vθ (R1 ) = ΩR1 , vθ (R2 ) = 0.
We obtain
2ΩR12 ΩR12 R22
c1 = c2 = − .
R12 − R22 R12 − R22
ΩR12 R2

vθ (r) = r− 2 .
2
R1 − R22 r
If R1 = R0 and R2 = αR0 with 0 < α < 1, we have

ΩR0 r 2 R0
vθ (r) = − α .
1 − α2 R0 r
The shear stress distribution is given by

2
α2

d vθ R0
τrθ = −µ r = −2µΩ .
dr r 1 − α2 r
If both cylinders rotate with outer cylinder at angular speed Ω1 and the inner cylinder
at angular speed Ω2 , the velocity field of the viscous fluid between the two coaxial
cylinders is obtained by determining the constants c1 and c2 in
r 1
vθ (r) = c1 + c2 .
2 r
Using the boundary conditions vθ (R1 ) = R1 Ω1 and vθ (R2 ) = R2 Ω2 , we obtain
R1 1 R2 1
r1 Ω1 = c1 + c2 , R2 Ω2 = c1 + c2 ,
2 R1 2 R2
Ω1 R12 − Ω2 R22 R12 R22 (Ω2 − Ω1 )
c1 = 2 , c2 = .
R12 − R22 R12 − R22
R12 R22

1 2 2
vθ = r Ω1 R 1 − Ω 2 R2 + (Ω1 − Ω2 ) .
R12 − R22 r
8.9 Consider an isothermal, incompressible fluid flowing radially between two concentric
porous spherical shells. Assume steady flow with vR . Simplify the continuity and
momentum equations for the problem.
Solution: The only nonzero velocity component is vR = vR (R). The continuity equation
(8.2.13) reduces to

vR dvR 1 dvR 1 d
R2 vR ,

0=2 + = 2vR + R = 2
R dR R dR R dR
which, when integrated, gives
R2 vR = c1 , a constant.
The momentum equations (8.2.14)–(8.2.16) simplify to
1 d2

2 ∂p ∂vR
µ 2 R vR − + ρfR = ρvR ,
R dR2 ∂R ∂R
1 ∂p
− + ρfφ = 0,
R ∂φ
1 ∂p
− + ρfθ = 0.
R sin φ ∂θ
8.10 A fluid of constant density and viscosity is in a cylindrical container of radius R0 and
the container is rotated about its axis with an angular velocity of Ω. Use the cylindrical
coordinate system with the z-coordinate along the cylinder axis. Let the body force
vector to be equal to ρf = −gêz . Assume that vr = u = 0 and vz = w = 0, and
vθ = vθ (r) and simplify the governing equations. Determine vθ (r) from the second
momentum equation subject to the boundary condition vθ (R0 ) = Ωr. Then evaluate
the pressure p from the remaining equations.
Solution: At steady state, the fluid velocities vr = vz = 0, and vθ is a function of

r only. The continuity equation is identically satisfied. The momentum equations
(8.2.10)–(8.2.12) reduce to
v2

∂p d 1 d ∂p
− = −ρ θ , µ (rvθ ) = 0, − − ρg = 0.
∂r r dr r dr ∂z
Solving the second momentum equation for vθ , we obtain

1 d r 1
(rvθ ) = c1 , vθ (r) = c1 + c2 ,
r dr 2 r
where c1 and c2 are constants of integration. Because vθ must be finite at r = 0, we
require c2 = 0. The boundary condition vθ (R0 ) = ωR0 yields c1 = 2Ω, and the velocity
becomes
vθ (r) = Ωr.
Then the first and third equations take the form
∂p ∂p
= ρΩ2 r, = −ρg.
∂r ∂z
Because p = p(r, z), we can write
∂p ∂p
dp = dr + dz = ρΩ2 rdr − ρgdz
∂r ∂z
and
1 2 2
p = −ρgz +ρΩ r + c,
2
where c is a constant of integration. If p = p0 at r = 0 and z = z0 , we obtain
c = p0 + ρgz0 and we have
1 2 2
p − p0 = −ρg(z − z0 ) + ρΩ r .
2
The free surface is defined by p = p0 :
1 2 2 1 2 2
−ρg(z − z0 ) + ρΩ r → z − z0 = Ω r .
2 2g
FigP8-9
Ω
Free surface
z0 z
r
R
Fig. P8.10
8.11 Consider the unsteady parallel flow on a flat plate (or plane wall). Assume that the
motion is started impulsively from rest. Take the x-coordinate along the plate and the
y-coordinate perpendicular to the wall. Assume that only nonzero velocity component
is vx = vx (y, t) and that the pressure p is a constant. Show that the Navier–Stokes
equations for this case are simplified to
∂vx ∂ 2 vx
ρ =µ , 0 < y < ∞. (1)
∂t ∂y 2
Solve Eq. (1) for vx (y) using the following boundary initial and boundary conditions:
Initial condition vx (y, 0) = 0,

(2)
Boundary conditions vx (0, t) = U0 ,
vx (∞, t) = 0.
√
Hint: Introduce a new coordinate η by assuming η = y/(2 νt), where ν is the kinematic
viscosity ν = µ/ρ, and seek a solution in the form vx (η) = U0 f (η). The solution is
obtained in terms of the complementary error function:
Z ∞ Z η
2 2 2 2
erfc η = √ e−η dη = 1 − erf η = 1 − √ e−η dη, (3)
π η π 0
where erf η is the error function.
Solution: For this problem, we have vy = vz = 0 and vx = vx (y, t). Then the second
momentum equation takes the form
∂vx ∂ 2 vx
=ν , 0 < y < ∞,
∂t ∂y 2
√
where ν = µ/ρ. Let vx (η) = U0 f (η) with η = y/(2 νt). We have
∂2 η2

∂ vx 1η 0 vx
=− f ; 2
= 2 f 00 .
∂t U0 2t ∂t U0 y
where prime denotes differentiation with respect to η. Then Eq. (1)takes the form
f 00 + 2ηf 0 = 0, (4)
which is now subjected to the boundary conditions (no slip at the wall)
f (0) = 1, f (∞) = 0. (5)
Solution of (4) for f 0 gives (solution of a first-order equation)

2
f 0 (η) = A e−η .
Integration of the above equation gives

Z η 2
f (η) = A e−ξ dξ + B. (6)
0
The lower limit is arbitrarily selected, which dictates the value of the constant B. The
constants A and B are determined using the two boundary conditions in (5). We obtain
Z ∞ −1
2
B = 1, A= e−ξ dξ
0
so that the solution becomes

R η −ξ2 Z η
e dξ 2 2
f (η) = 1 − 0∞ −ξ2
R =1− √ e−ξ dξ. (7)
0
e dξ π 0
The ratio of the integrals is recognized as the error function.

FigP8-11
y y
Fluid at rest, t < 0 Onset of fluid motion, t = 0

U0
Infinite flat plate

y
vx ( y, t ) = U 0 f (η )
Unsteady flow, t > 0
U0
Fig. P8.11
8.12 Solve Eq. (1) of Problem 8.11 for the following boundary conditions (that is, flow near
an oscillating flat plate)
Initial condition vx (y, 0) = 0,

(1)
Boundary conditions vx (0, t) = U0 cos nt, vx (∞, t) = 0.
In particular, obtain the solution

r
ρn
vx (y, t) = U0 eλy cos(nt − λy), λ = . (2)
2µ
Solution: Suppose that the velocity of the plate is of the form
vx (0, t) = U0 cos nt
We solve Eq. (1) of Problem 8.11, that is,
∂vx ∂ 2 vx
=ν , 0 < y < ∞. (3)
∂t ∂y 2
subject to the boundary condition (1). Note that Eq. (3) is similar to the unsteady
heat conduction equation in one dimension. The solution of Eq. (3) is (use separation
of variables) given by Eq. (2)
vx (y, t) = U0 e−λy cos(nt − λy),

p p
where λ = n/2ν. Let η = λy = y n/2ν to rewrite the solution as
vx (y, t) = U0 e−η cos(nt − η). (4)
8.13 Show that the components of the viscous stress tensor τ [see Eq. (6.5.7)] for an isotropic,
viscous, Newtonian fluid in cylindrical coordinates are related to the velocity gradients
by
∂vr 1 ∂vθ vr
τrr = 2µ + λ ∇ · v, τθθ = 2µ + + λ ∇ · v,
∂r r ∂θ r

∂vz ∂ vθ
1 ∂vr
τzz = 2µ + λ ∇ · v, τrθ = µ r + ,
∂z ∂r r r ∂θ

∂vθ 1 ∂vz ∂vz ∂vr
τzθ = µ + , τzr = µ + ,
∂z r ∂θ ∂r ∂z
1 ∂(rvr ) 1 ∂vθ ∂vz
∇·v = + + .
r ∂r r ∂θ ∂z
Solution: This is a trivial exercise in view of the relations in Eqs. (3.5.21)–(3.5.26).

The required results follow from the constitutive equations
τrr = 2µDrr + λ ∇ · v, τθθ = 2µDθθ + λ ∇ · v, τzz = 2µDzz + λ ∇ · v,

τrθ = 2µDrθ , τzθ = 2µDzθ , τzr = 2µDzr ,
and the components of the rate of deformation tensor D in terms of the velocity com-
ponents

∂vr 1 1 ∂vr ∂vθ vθ
Drr = , Drθ = + − ,
∂r 2 r ∂θ ∂r r

1 ∂vr ∂vz vr 1 ∂vθ
Drz = + , Dθθ = + ,
2 ∂z ∂r r r ∂θ

1 ∂vθ 1 ∂vz ∂vz
Dzθ = + , Dzz = .
2 ∂z r ∂θ ∂z
8.14 Show that the components of the viscous stress tensor τ [see Eq. (6.5.7)] for an isotropic,
viscous, Newtonian fluid in spherical coordinates are related to the velocity gradients
by

∂vR 1 ∂vθ vR
τRR = 2µ + λ ∇ · v, τφφ = 2µ + + λ∇ · v
∂R R ∂θ R

1 ∂vφ vR vφ cot φ
τθθ = 2µ + + + λ∇ · v
R sin φ ∂φ R R

∂ vφ 1 ∂vR 1 ∂vR ∂ vθ
τRφ = µ R + , τRθ = µ +R
∂R R R ∂φ r sin φ ∂θ ∂R R

sin φ ∂ vθ 1 ∂vφ
τφθ = µ +
R ∂φ sin φ R sin φ ∂θ
1 ∂(R2 vR ) 1 ∂ 1 ∂vθ
∇·v = + (vφ sin φ) + .
R2 ∂R R sin φ ∂φ R sin φ ∂θ
Solution: Use the constitutive relations

τRR = 2µDRR + λ ∇ · v, τφφ = 2µDφφ + λ ∇ · v, τθθ = 2µDθθ + λ ∇ · v,
τRφ = 2µDRφ , τRθ = 2µDRθ , τφθ = 2µDφθ ,
and the components of the rate of deformation tensor D in terms of the velocity com-
ponents [compare with Eq. (3.5.32) for strain-displacement relations in the spherical
coordinates]

∂vR 1 1 ∂vR ∂vφ vφ
DRR = , DRφ = + − ,
∂R 2 R ∂φ ∂R R

1 1 ∂vR ∂vθ vθ 1 ∂vφ
DRθ = + − , Dφφ = + vR ,
2 R sin φ ∂θ ∂R R R ∂φ

1 1 ∂vφ ∂vθ
Dφθ = + − vθ cot φ ,
2R sin φ ∂θ ∂φ

1 ∂vθ
Dθθ = + vR sin φ + vφ cos φ .
r sin φ ∂θ
8.15 Use the velocity field in Eq. (10) of Example 8.2.4 to show that the shear stress
component τrφ and pressure p are
4 2
3µV∞ R0 3µV∞ R0
τRφ = sin φ, p = p0 − ρgz − cos φ,
2R0 R 2R0 R
where p0 is the pressure in the plane z = 0 far away from the sphere and −ρgz is the
contribution of the fluid weight (hydrostatic effect).
Solution: The shear stress is given by

∂ vφ 1 ∂vR
τRφ = µ r +
∂R R R ∂φ
From Eq. (10) of Example 8.2.4 we have

" 3 #
1 ∂vR 1 3 R0 1 R0
= − V∞ 1 − + sin φ,
R ∂φ R 2 R 2 R
" 3 #
∂ vφ 1 3 R0 R0
R = − V∞ −1 + + sin φ.
∂R R R 2 R R
Hence, we obtain
4
3µV∞ R0
τRφ = sin φ.
2R0 R
The pressure is obtained using the φ-momentum equation,

1 ∂p µ ∂ ∂vφ ∂ 1 ∂ ∂vR
= 2 R2 + (vφ sin φ) + 2 − ρg cos φ.
R ∂φ R ∂R ∂R ∂φ sin φ ∂R ∂φ
Calculating the various derivatives appearing in the above expression,
3
∂vφ 3V∞ R0 R0
=− + sin φ,
∂R 4 R R
3
∂ ∂vφ 3V∞ R0
R2 = sin φ,
∂R ∂R 2 R
" 3 #
∂vR 3 R 1 R0
= −V∞ 1 − + sin φ,
∂φ 2 R 2 R
" 3 #
∂ 3 R0 1 R0
(vφ sin φ) = −2V∞ 1 − − sin φ cos φ,
∂φ 4 R 4 R
" 3 #
∂ 1 ∂ 3 R0 1 R0
(vφ sin φ) = 2V∞ 1 − − sin φ,
∂φ sin φ ∂φ 4 R 4 R
and substituting into the momentum equation gives

∂p 3µV∞ R0
= sin φ − ρg cos φ
∂φ 2R R
Integrating with respect to φ yields
2
3µV∞ R0
p = p0 − ρgR sin φ − cos φ,
2R0 R
where p0 is the pressure in the plane z = 0 far away from the sphere and ρgR sin φ = ρgz
is the contribution of the fluid weight (hydrostatic effect).
8.16 Derive the following vorticity equation for a fluid of constant density and viscosity:
∂ω
+ (v · ∇)ω = (ω · ∇)v + ν∇2 ω,
∂t
where ω = ∇ × v and ν = µ/ρ.
Solution: Taking curl of the equation of motion (in the absence of body forces), we
obtain
Dv
ρ∇ × =∇×∇·σ
Dt
= −∇ × ∇P + 2µ∇ × ∇ · D = −0 + µ∇ × ∇2 v + ∇(∇ · v)

= µ∇2 ω
where we have used the condition ∇ · v = 0. Then using Problem 5.17 result we obtain
the required equation.
8.17 Bernoulli’s equation. Consider a flow with hydrostatic pressure, σ = −pI and conser-
vative body force f = −∇φ.
(a) For steady flow, show that
v2

1
v·∇ + φ + v · grad p = 0.
2 ρ
(b) For steady and irrotational (that is, ∇ × v = 0) flow, show that
2
v 1
∇ + φ + ∇p = 0.
2 ρ
Solution: (a) For steady flow, we have (∂/∂t) = 0, and the identity from Problem 5.16
takes the form 2
v 1
v·∇ = v · div σ T + v · f
2 ρ
1
= − v · grad p − v · ∇φ,
ρ
from which the desired equality follows.
(b) From Part (a), we have
v2

1
v·∇ +φ+ p = 0.
2 ρ
If curl v=0, then there exists a scalar function Φ such that v = ∇Φ because curl of
gradient of a function is zero. In particular, we can choose
2
v 1
Φ= +φ+ p .
2 ρ
Then, we have
v2
2
1 v 1
∇ +φ+ p ·∇ + φ + p = 0.
2 ρ 2 ρ
Because this is a positive function that is zero, the function itself must be zero, giving
the desired equality:
2
v2

v 1 1
grad +φ+ p =0 ⇒ + φ + P = constant.
2 ρ 2 ρ
8.18 Use Bernoulli’s equation (which is valid for steady, frictionless, incompressible flow)
derived in Problem 8.17 to determine the velocity and discharge of the fluid at the exit
of the nozzle in the wall of the reservoir shown in Fig. P8.18.
Figure P5-17
d = 50 mm dia
h = 5m
Fig. P8.18
Solution: The Bernoulli equation can be applied to a point on the surface of the water
in the tank and a point downstream from the nozzle exit, as shown in the figure. The
potential φ of the body force is nothing but that due to the gravity: φ = ρz, where
z denotes the elevation from a datum. If we take the centerline of the nozzle as the
datum, we have
v12 p1 v2 p2
+ + z1 g = 2 + + z2 g.
2 ρ 2 ρ
Because p1 = p2 , the atmospheric pressure, z2 = 0, z1 = h, and the velocity v1 on the
surface of the reservoir is essentially zero, we have
p √
v2 = 2gh = 2 × 9.81 × 5 = 9.9 m/s.
Thus, the velocity of the fluid exiting the nozzle is equal to the velocity of free fall
from the surface of the reservoir, which is known as the Torricelli’s theorem. The
discharge Q at the exit of the nozzle is calculated from the product of the velocity and
the cross-sectional area of the nozzle
πd2
Q = v2 = 0.01945 m3 /s = 19.45 Liters/s.
4
Figure P8-*
8.19 The fan shown in Fig. P8.19 moves air (ρ = 1.23 kg/m3 ) at a mass flow rate of 0.1
kg/min. The upstream side of the fan is connected to a pipe of diameter d1 = 50
mm, the flow is laminar, the velocity distribution is parabolic, and the kinetic energy
coefficient is α = 2. The downstream of the fan is connected to a pipe of diameter
d2 = 25 mm, the flow is turbulent, the velocity profile is uniform, and the kinetic energy
coefficient is α = 1. If the rise in static pressure between upstream and downstream is
100 Pa and the fan motor draws 0.15 W, determine the loss (−Hnet ).
d2 = 25mm
Turbulent flow
Fan
e2 − e1 = 0 d1 = 50mm
P2 − P1 = 100 Pa ρQ0 Laminar flow
Fig. P8.19
Solution: Application of Eq. (8.2.50) to the control volume shown in Fig. P8.19 gives
P2 α2 v22 p1 α1 v12
+ +0= + + 0 + Hnet + Wshaft ,
ρ 2 ρ 2
or the loss is given by
p2 − p1 α2 v22 α1 v12
−Hnet = Wshaft − − + .
ρ 2 2
The values of Wshaft , v1 , and v2 are computed as follows:
(0.15)(1)
Wshaft = (60)(1) = 90 N.m/kg
0.1
(0.1/60)
v1 = = 0.69 m/s
(1.23)[π(50)2 /4]10−6
(0.1/60)
v2 = = 2.76 m/s
(1.23)[π(25)2 /4]10−6
Then the loss is
100 1.0(2.76)2 2(0.69)2
loss = 90 − − + = 5.3665 N.m/kg
1.23 2 2
8.20 Show that for a steady creeping flow of a viscous incompressible fluid in the absence of
body forces, the Navier–Stokes equations become
∇p = µ∇2 v.
Solution: From Problem 6.24 solution, we have

Dv
ρ = ρf − ∇p + (λ + µ) ∇(∇ · v) + µ∇2 v.
Dt
For steady case, we set Dv/Dt = 0; for incompressible fluid, we set ∇ · v = 0. Then,
in the absence of body forces, the Navier–Stokes equations reduce to
0 = −∇p + µ∇2 v.
8.21 In heat transfer, one often neglects the strain energy part of the internal energy e and
assumes that e depends only on the temperature θ, e = e(θ). Show that the energy
equation (5.4.11) reduces to
Dθ
ρc = ∇ · (k · ∇θ) + ρrh ,
Dt
where c = ∂e/∂θ. State the assumptions under which the equation is derived.
Solution: We begin with the energy equation

De
ρ = σ : D − ∇ · q + ρ rh ,
Dt
and set σ : D = 0; then
De ∂e Dθ Dθ
= =c .
Dt ∂θ Dt Dt
Further, by Fourier heat conduction law, we have
q = −k · ∇θ.
Then, we have
Dθ
ρc = ∇ · (k · ∇θ) + ρ rh .
Dt
8.22 Consider a long electric wire of length L and cross section with radius R0 and electrical
conductivity ke [1/(Ohm·m)]. An electric current with current density I (amps/m2 )
is passing through the wire. The transmission of an electric current is an irreversible
process in which some electrical energy is converted into thermal energy (heat). The
rate of heat production per unit volume is given by
I2
ρrh = .
Figure 8-3-2 ke
Assuming that the temperature rise in the cylinder is small enough not to affect the
thermal or electrical conductivities and heat transfer is one-dimensional along the radius
of the cylinder, derive the governing equation using balance of energy.
z
R0 θ (circumferential
coordinate)
(q )
r r+Δr
(q )
r r r + Δr
r
T0
Fig. P8.22
Solution: For the energy balance, we consider an cylindrical element of thickness ∆r

and length L (see the figure above). The contributions to the energy balance are as
follows:
Rate of thermal energy input at the inner wall of the cylindrical element: 2πrL qr
Rate of thermal energy output at the outer wall of the element: 2π(r + ∆r)L (qr + ∆qr )
Rate of production of thermal energy by electrical dissipation: (2πrL)∆r ρrh
Adding up all the energies (energy out as negative), we obtain
2πrL qr − 2π(r + ∆r)L (qr + ∆qr ) + (2πrL)∆r ρrh = 0.
Dividing throughout by 2πL∆r and taking the limit ∆r → 0 (the higher-order terms
go to zero in the limit), we obtain
(r + ∆r)(qr + ∆qr ) − rqr d
− lim + rρrh = 0 ⇒ − (rqr ) + rρrh = 0.
∆r→0 ∆r dr
8.23 Solve the equation derived in Problem 8.22 using the boundary conditions
q(0) = finite, T (R0 ) = T0 .
Solution: Integrating the equation

d
(rqr ) = rρrh
dr
we obtain
1 c1
qr = ρrh r + ,
2 r
where the constant of integration c1 is set to zero due to the requirement that qr be
finite at r = 0. Using the Fourier heat conduction law, qr = −k(dT /dr), we obtain
dT 1 ρrh 2
−k = ρrh r → T (r) = − r + c2 ,
dr 2 4k
where the constant c2 is determined by the boundary condition T (R0 ) = T0 :
ρrh R02
c2 = T0 + .
4k
Thus the solution becomes
" 2 #
ρrh R02

r
T (r) = T0 + 1− .
4k R0
8.24 A slab of length L is initially at temperature f (x). For times t > 0 the boundaries
at x = 0 and x = L are kept at temperatures T0 and TL , respectively. Obtain the
temperature distribution in the slab as a function of position x and time t.
Solution: The governing equation for this unsteady-state one-dimensional problem is

(with no internal heat generation)

∂ ∂T ∂T
k = ρcp . (1)
∂x ∂x ∂t
The equation can be written in an alternative form by a change of variable θ
T − T0
θ= . (2)
TL − T0
We have (for constant conductivity)

∂ ∂θ 1 ∂θ k
= , α= , (3)
∂x ∂x α ∂t ρcp
where α is known as the thermal diffusivity. The boundary conditions take the form
T (0, t) = T0 → θ(0, t) = 0; T (L, t) = TL → θ(L, t) = 1. (4)
We assume solution in separable form θ(t, x) = X(x)Y (t). The boundary conditions
on θ can be translated to those on X(x):
θ(0, t) = X(0)Y (t) = 0 → X(0) = 0; θ(L, t) = X(L)Y (t) = 1 → X(L) = 1.
Substitute θ(t, x) = X(x)Y (t) into the governing equation and obtain
d2 X dY
+ λ2 X = 0, + αλ2 Y = 0, (5)
dx2 dt
where λ2 > 0 is the constant introduced in the separation of variables. The choice of
the constant λ is such that the spatial variation is not an exponential decay or growth.
The solutions of the two equations in (5) are
2
X(x) = c1 cos λx + c2 sin λx, Y (t) = c3 e−αλ t
The boundary conditions yield

(2n + 1)π
X(0) = 0 → c1 = 0; X(L) = 1 → c2 sin λL = 1 → λn = . (6)
2L
The complete solution is
∞
X 2
θ(x, t) = Bn sin λn x e−αλn t .
n=1
Next, we use the initial condition to determine the constants Bn

∞
X
θ(x, 0) = f (x) = Bn sin λn x.
n=1
Multiplying both sides with sin λm x and integrating over 0 to L, we obtain

Z L Z L
Bm sin2 λm x dx = f (x) sin λm x dx. (7)
0 0
Because Z L
L
sin2 λm x dx = .
0 2
∞ Z L
X 2 2
θ(x, t) = Bn sin λn x e−αλn t , Bn = f (x) sin λn x dx. (8)
n=1
L 0
8.25 A slab of unit height, 0 ≤ x ≤ 1, is initially kept at temperature T (x, 0) = T0 (1 − x2 ) =

f (x). For times t > 0, the boundary at x = 0 is kept insulated and the boundary at
x = 1 is kept at zero temperature, T1 = 0. Determine the temperature distribution
T (x, t).
Solution: We follow the solution developed in the previous problem. The governing
equation for this unsteady-state one-dimensional problem is (with no internal heat
generation)
∂2T 1 ∂θ k
= , α= . (1)
∂x2 α ∂t ρcp
The boundary conditions are form
∂T

= 0; T (1, t) = 0. (2)
∂x (0,t)

and the initial condition is

T (x, 0) = f (x) = T0 (1 − x2 ). (3)
We assume solution in separable form T (t, x) = X(x)Y (t). The boundary conditions
on θ can be translated to those on X(x):
∂T dX dX

T (1, t) = X(1)Y (t) = 0 → X(1) = 0; = Y (t) = 0 → = 0.
∂x (0,t) dx x=0 dx x=0

(4)
Substitute T (t, x) = X(x)Y (t) into the governing equation and obtain
d2 X dY
+ λ2 X = 0, + αλ2 Y = 0, (5)
dx2 dt
where λ2 > 0 is the constant introduced in the separation of variables. The choice of
the constant λ is such that the spatial variation is not an exponential decay or growth.
The solutions of the two equations in (5) are
2
X(x) = c1 cos λx + c2 sin λx, Y (t) = c3 e−αλ t . (6)
The boundary conditions yield
dX (2n + 1)π

= 0 → c2 = 0; X(1) = 0 → c1 λ cos λ = 0 → λn = . (7)
dx x=0 2

The complete solution is

∞
X 2
T (x, t) = An cos λn x e−αλn t .
n=1
Next, we use the initial condition to determine the constants An

∞
X
T (x, 0) = f (x) = Bn cos λn x.
n=1
Multiplying both sides with cos λm x and integrating over 0 to 1, we obtain

Z 1 Z 1
Am cos2 λm x dx = f (x) cos λm x dx. (8)
0 0
Because Z 1 Z 1
1
cos2 λm x dx = , An = 2T0 (1 − x2 ) cos λn x dx.
0 2 0
Evaluating the integral (cos λn = 0)
Z 1
λ2n − 2

1 2 cos λn
An = 2T0 (1 − x2 ) cos λn x dx = 2T0 sin λn − − sin λ n
0 λn λ2n λ3n
4T0
= 3 sin λn . (9)
λn
The solution becomes [note that sin λn = (−1)n ]
∞
X sin λn 2
T (x, t) = 4T0 3
sin λn x e−αλn t . (10)
n=1
λn
8.26 Determine the steady-state temperature distribution in an isotropic hollow sphere (in-
ternal and external radii are a and b, respectively) with uniform temperatures at the
inner (Ti ) and outer (T0 ) surfaces but without internal heat generation.
Solution: Due to the solution symmetry (that is, temperature T is independent of φ
and θ in the spherical coordinate system), one can write the governing equation as [see
Eq. (8.3.6)]
1 ∂ 2 ∂T
R = 0, a < R < b, (1)
R2 ∂R ∂R
where a and b are internal and external radii, respectively, of the sphere. First we note
that
d2 T

1 ∂ 2 ∂T 2 dT
2
R = + ,
R ∂R ∂R dR2 R dR
and
1 d2 d2 T 2 dT
2
(RT ) = + .
R dR dR2 R dR
Thus, we have
1 d2
(RT ) = 0, a < R < b. (2)
R dR2
Integrating twice with respect to R, we obtain the solution
c1
T (R) = + c2 , a < R < b. (3)
R
Using the boundary conditions T (a) = Ti and T (b) = T0 , we obtain
ab T0 b − Ti a
c1 = (Ti − T0 ), c2 = , (4)
b−a b−a
and the solution becomes
T − Ti b a
= 1− . (5)
T0 − Ti b−a R
The heat flux is

dT ab Ti − T0
q(R) = −4πkR2 = 4πk (Ti − T0 ) ≡ (6)
dR b−a Rth
where Rth is the thermal resistance of a spherical shell.
8.27 Obtain the steady-state temperature distribution T (x, y) in a rectangular region, 0 ≤

x ≤ a, 0 ≤ y ≤ b for the boundary conditions
qx (0, y) = 0, qy (x, b) = 0, qx (a, y) + hT (a, y) = 0, T (x, 0) = f (x). (1)
Solution: The governing equation is

2
∂2T

∂ T
k + = 0. (2)
∂x2 ∂y 2
Using the separation of variables, we obtain [see Eq. (8.3.23)]

T (x, y) = (C1 cos λx + C2 sin λx) C3 e−λy + C4 eλy , (3)
where λ is the constant introduced in the separation-of-variables method. The constants

Ci (i = 1, 2, 3, 4) are determined using the boundary conditions in Eq. (1). We obtain

qx (0, y) =0 → −C2 kλ C3 e−λy + C4 eλy = 0 → C2 = 0,

qx (a, y) + hT (a, y) = 0 → C1 (−kλ sin λa + h cos λa) C3 e−λy + C4 eλy = 0
h
→ −kλn sin λn a + h cos λn a = 0 or λn tan λn a = . (4)
k
∞
X
T (x, y) = cos λn x C1 C3 e−λn y + C1 C4 eλn y
n=1
X∞
= cos λn x (An cosh λn y + Bn sinh λn y) , (5)
n=1
where λn is determined from the transcendental equation in (4). Next, we use the
boundary conditions at y = 0, b determine An and Bn . Using the condition qy (x, b) = 0,
qy (x, b) = 0 → An sinh λn b + Bn cosh λn b = 0 or Bn = −An tanh λn b.
The solution now can be expressed as

∞
X cos λn x
T (x, y) = An (cosh λn y cosh λn b − sinh λn y sinh λn b)
n=1
cosh λn b
∞
X cosh λn (b − y)
= An cos λn x. (6)
n=1
cosh λn b
The constants An are determined using the remaining boundary condition. We have
∞
X
T (x, 0) = f (x) = An cos λn x.
n=1
Multiplying both sides with cos λm x and integrating from 0 to a and using the orthog-
onality of the cosine functions
Z a
0, λm 6= λn
cos λn x cos λm x dx = a
0 2
, λm = λn
we obtain Z a
2
An = f (x) cos λn x dx. (7)
a 0
Hence, the final solution is given by Eq. (6) with An given by Eq. (7).
8.28 Consider the steady flow through a long, straight, horizontal circular pipe. The velocity
field is given by [see Example 8.2.2]
R2 dp r2

vr = 0, vθ = 0, vz (r) = − 0 1− 2 . (1)
4µ dr R0
If the pipe is maintained at a temperature T0 on the surface, determine the steady-state
temperature distribution in the fluid.
Solution: Seeking the temperature distribution that only depends on r, we simplify Eq.
(8.3.5) to
2
k d dT dvz
r +µ = 0. (2)
r dr dr dr
In view of Eq. (1), we have (dp/dr is constant)

k d dT 1 dp
r = µα2 r, α= .
r dr dr 2µ dr
Integrating the equation twice with respect to r, we obtain
µα2 3
T (r) = r + c1 log r + c2 .
9k
The constant c1 is set to zero on account of the requirement that the temperature
be finite at r = 0. Using the boundary condition T (R0 ) = T0 , we obtain c2 = T0 −
(µα2 R03 /9k) and the temperature distribution becomes
" 3 #
µα2 R03

r
T (r) = T0 − 1− .
9k R0
8.29 Consider the free convection problem of flow between two parallel plates of different
temperature. A fluid density with density ρ and viscosity µ is placed between two
vertical plates a distance 2a apart, as shown in Fig. P8.29. Suppose that the plate
at x = a is maintained at a temperature T1 and the plate at x = −a is maintained
at a temperature T2 , with T2 > T1 . Assuming that the plates are very long in the
y-direction and hence that the temperature and velocity fields are only a function of x,
determine the temperature T (x) and velocity vy (x). Assume that the volume rate of
flow in the upward moving stream is the same as that in the downward moving stream
and the pressure gradient is solely due to the weight of the fluid.
Solution: Because the plates are very long in the y-direction, the temperature at any
distance along y, except for the ends, will be a function of x alone. The governing
differential equation resulting from conservation of energy is
d2 T
k = 0, −a < x < a
dx2
The solution of the equation is T (x) = Ax + B, where the constants of integration A
and B are evaluated using the boundary conditions T (−a) = T2 and T (a) = T1 . The
temperature distribution becomes
T2 + T1 T2 − T1 x T2 − T1 x
T (x) = − = T0 − . (a)
2 2 a 2 a
Figure P8-29
2a
Temperature
distribution,
T ( x) ●
Cold plate
Hot plate
T2 ●
Velocity T0
T1
distribution,
vy ( x )
y
a
x
Fig. P8.28
The velocity distribution is determined from the x-momentum equation
d2 vy dp
µ = + ρg,
dx2 dy
where the viscosity µ is assumed to be constant. However, the density is a function of
temperature. Expanding ρ in Taylor’s series about some reference temperature Tr

∂ρ
ρ = ρ(Tr ) + (T − Tr ) + · · · = ρr − ρr βr (T − Tr ) + · · · ,
∂T Tr
where ρr is the density at reference temperature and βr is the coefficient of volume

expansion at the reference temperature

1 ∂V 1 ∂(1/ρ) 1 ∂ρ
β= = = − .
V ∂T p (1/ρ) ∂T
p ρ ∂T p
Hence the momentum equation becomes
d2 vy dp
µ = + ρr g − ρr βr g(T − Tr ).
dx2 dy
If the pressure gradient is solely due to the weight of the fluid, we have dp/dz = −ρr g
and the momentum equation becomes
d2 vy
µ = −ρr βr g(T − Tr ). (b)
dx2
The above equation indicates that the viscous forces are balanced by the buoyancy
forces.
Using the temperature distribution from Eq. (a) into Eq. (b), we obtain
d2 vy

1 x
µ 2 = −ρr βr g (T0 − Tr ) − (T2 − T1 ) . (b)
dx 2 a
Solving The equation needs to be solved subjected to the boundary conditions
vy (−a) = 0, vy (a) = 0.
We obtain
2 3
ρr βr ga2 (T2 − T1 )

T0 − Tr x x x
vy (x) = 6 1− + − . (c)
12µ T2 − T1 a a a
If we require that the net volume flow in the y-direction is zero,

Z a
vy dx = 0,
−a
we obtain Tr = T0 . The solution now reduces to

3
ρr βr ga2 (T2 − T1 ) x x
vy (x) = − . (d)
12µ a a
8.30 Determine the steady-state temperature distribution through an infinite slab of height
h, thickness b (see Fig. P8.30), and made of isotropic material whose conductivity
changes with the height according to the equation
Figure P8-30
n
1 y
k(y) = (kt − kb )f (y) + kb , f (y) = + , (1)
2 h
where kt and kb are the values of the conductivities of the top and the bottom surfaces,
and n is a constant. Assume that the top is maintained at temperature T0 and bottom
is maintained at temperature T1 .
y kt , T0
h
x
kb , T1
Fig. P8.29
Solution: Because the slab is very long in the x-direction, the temperature at any
distance along x, except for the ends, will be a function of y alone. The governing
differential equation resulting from conservation of energy is

d dT
k k(y) = 0, −h/2 < x < h/2, (1)
dy dy
with the boundary conditions
T (h/2) = T0 , T (−h/2) = T1 . (2)
Integrating the equation (1), we obtain

Z h/2
dT 1
k(y) = A, T (y) = A dy + B, (3)
dy −h/2 k(y)
where the constants of integration A and B determined using the boundary conditions.
However, the integral needs to be evaluated numerically.

229
Chapter 9: LINEARIZED VISCOELASTICITY
Note: Solutions of many of the elasticity problems of Chapter 7 can be reworked as viscoelastic
solutions for a given viscoelastic material. Thus, many new problems can be generated from
the examples as well as exercise problems presented in Chapter 7.
9.1 Method of partial fractions. Suppose that we have a ratio of polynomials of the type
F̄ (s)
.
Ḡ(s)
where F̄ (s) is a polynomial of degree m and Ḡ(s) is a polynomial of degree n, with

n > m. We wish to write in the form
F̄ (s) c1 c2 c3 cn
= + + + ··· + .
Ḡ(s) s + α1 s + α2 s + α3 s + αn
where ci and αi are constants to be determined using
(s + αi )F̄ (s)
ci = lim , n = 1, 2, · · · , n.
s→−αi Ḡ(s)
It is understood that Ḡ(s) is equal to the product Ḡ(s) = (s + α1 )(s + α2 ) . . . (s + αn ).

If
F̄ (s) = s2 − 6, Ḡ(s) = s3 + 4s2 + 3s,
determine ci .
Solution: First we write

Ḡ(s) = s3 + 4s2 + 3s = s3 + s2 + 3s2 + 3s = s2 (s + 1) + 3s(s + 1) = (s2 + 3s)(s + 1)
= s(s + 1)(s + 3) = (s + α1 )(s + α2 )(s + α3 ), α1 = 0, α2 = 1, α3 = 3.
We have
s(s2 − 6) (0 − 6)
c1 = lim = = −2
s→ 0 s(s + 1)(s + 3) (0 + 1)(0 + 3)
(s + 1)(s2 − 6) 1−6
c2 = lim = = 2.5
s→ −1 s(s + 1)(s + 3) (−1)(−1 + 3)
(s + 3)(s2 − 6) 9−6
c3 = lim = = 0.5
s→ −3 s(s + 1)(s + 3) (−3)(−3 + 1)
so that
s2 − 6 c1 c2 c3 2 5 1
= + + =− + + .
s3 + 4s2 + 3s s s+1 s+3 s 2(s + 1) 2(s + 3)
and its Laplace inverse is given by
−2H(t) + 2.5e−t + 0.5e−3t .
9.2 Given the following transformed function
(1 + ν)pa2 b

3Ḡ(s) r b
ūr (r, s) = + , (1)
2Ḡ(s)(b2 − a2 ) 3K̄(s) + Ḡ(s) b r
where K̄ = K, p1 , q0 , and q1 are constants, and Ḡ(s) is given by

1 q0 q 1 − q 0 p1
Ḡ(s) = + , (2)
2 s p1 (p−1
1 + s)
determine its Laplace inverse, ur (r, t).

Solution: Let us write the given equation as

3 r 1 b c2 c3
ūr (r, s) = c1 + , Ḡ(s) = + ,
3K + Ḡ(s) b Ḡ(s) r s (α + s)
where
(1 + ν)pa2 b q0 q1 − p1 q0 1
c1 = , c2 = , c3 = , α= .
2(b2 − a2 ) 2 p1 p1
We have
3 r c4 b
ūr (r, s) = c1 s(s + α) + ,
(s − λ1 )(s − λ2 ) b (s + λ3 ) r
where
1 p 1 p
λ1 = −B + B 2 − 4AC , λ2 = −B − B 2 − 4AC ,
2A 2A
A = 3K, B = 3Kα + c2 + c3 , c2 = α,
and
1 c2 α
c4 = , λ3 = .
c2 + c3 c2 + c3
Lastly, we write
(s + α) c5 c6
= + ,
(s − λ1 )(s − λ2 ) (s − λ1 ) (s − λ2 )
and c5 and c6 can be determined from the relations
c1 + c2 = 1, −α = λ1 c2 + λ2 c1 .
The final result is given in Eq. (14) of Example 9.4.6.
Figure 9-2-7 9.3 Determine the creep and relaxation responses of the three-element model (i.e., standard
linear solid) of Fig. 9.2.9(b) following the procedure used in Eqs. (9.2.33)–(9.2.42) for
the three-element model shown Fig. 9.2.9(a).
ε2 σ ε1 ε 2 σ 2
1
ε1 σ • • ε • • • ε
k2 σ η k2 σ
• η • • • k1 • •
k1
• •
σ2 • σ1•
ε2 ε
(a) (b)
Fig. P9.3

σ1 + σ2 = σ, ε1 + ε2 = ε. (1)
We begin with
σ2 σ̇2
ε̇ = ε̇1 + ε̇2 = + . (2)
η k2
σ̇1 σ1
Add k2
= (k1 /k2 )ε̇ and η
= (k1 /η)ε to both sides to obtain

k1 k1 σ σ̇
ε+ 1+ ε̇ = + . (3)
η k2 η k2
Creep response. Let σ(t) = σ0 H(t) and obtain

k1 k1 σ0 σ0
ε+ 1+ ε̇ = + δ(t). (4)
η k2 η k2
Taking the Laplace transform and assuming zero initial conditions, we obtain

k1 k1 σ0 σ0
ε̄ + s 1 + ε̄ = + .
η k2 sη k2
CHAPTER 9: LINEARIZED VISCOELASTICITY 231
or α
2
(α1 + sβ1 ) ε̄(s) = σ0 + β2 (5)
s
where
k1 1 k1 1
α1 = , α2 = , β1 = 1+ , β2 = . (6)
η η k2 k2
Using the factorizations (see Problem 9.1), we can write
" ! #
α2 1 β2 1
ε̄(s) = σ0 1 − α1 + , (7)
α1 β1
+s β1 α 1
β1
+s
where
α2 1 β2 1 β2 α2 k2
= , = , − =− . (8)
α1 k1 β1 k1 + k2 β1 α1 k1 (k1 + k2 )
The inverse Laplace transform of the equation gives

1 k2 β1 (k1 + k2 )η
ε(t) = σ0 − e−t/τ , τ = = . (9)
k1 k1 (k1 + k2 ) α1 k1 k2
Hence, the creep compliance function is
1 k2
J(t) = − e−t/τ . (10)
k1 k1 (k1 + k2 )
Relaxation response. Let ε(t) = ε0 H(t) and obtain

k1 k1 σ σ̇
ε0 + 1 + δ(t)ε0 = + . (11)
η k2 η k2

k1 k1 1 s
+ 1+ ε0 = σ̄ + . (12)
sη k2 η k2
or α
2
(α1 + sβ1 ) σ̄(s) = ε0 + β2 (13)
s
where
1 k1 1 k1
α1 = , α2 = , β1 = , β2 = 1 + . (14)
η η k2 k2
Using the factorizations, we can write
" ! #
α2 1 β2 1
σ̄(s) = ε0 1 − α1 + , (15)
α1 β1
+s β1 α
β1
1
+s
where
α2 β2 β2 α2
= k1 , = k1 + k2 , − = k2 . (16)
α1 β1 β1 α1
β1 η
σ(t) = ε0 k1 + k2 e−t/τ , τ = = . (17)
α1 k2
Hence, the relaxation function is
Y (t) = k1 + k2 e−t/τ . (18)
9.4 Derive the governing differential equation for the spring-dashpot model shown in Fig.
P9.3. Determine the creep compliance J(t) and relaxation modulus Y (t) associated
with the model.
Figure P9-3
η2 G2
ε
• σ
η1
Fig. P9.4
σ1 = σ2 , σ2 + σ3 = σ, ε1 + ε2 = ε, ε3 = ε. (1)
We begin with
σ1 σ̇2
ε̇ = ε̇1 + ε̇2 = + . (2)
η2 G2
σ̇3
Add G 2
= (η1 /G2 )ε̈ and ση23 = (η1 /η2 )ε̇ to both sides (because σ1 = σ2 and σ1 +σ3 = σ)
to obtain
η1 σ η2
ε̇ G2 + + η1 ε̈ = σ̇ + , τ2 = . (3)
τ2 τ2 G2
Creep response. Let σ = σ0 H(t) and obtain

η1 H(t)
ε̇ G2 + + η1 ε̈ = σ0 δ(t) + . (4)
τ2 τ2

η1 1
sε̄ G2 + + s2 η1 ε̄ = σ0 1 +
τ2 sτ2
or
1
1+ α

sτ2 σ0 1 + s1
ε̄(s) = σ0 = , (5)
s G2 + η1
+ s2 η1 η1 s(α2 + s)
τ2
where
1 G2 G2 G2 G2 (η1 + η2 )
α1 = = , α2 = + = . (6)
τ2 η2 η1 η2 η1 η2
Using the factorization (see Problem 9.1), we can write

σ0 1 1 1 1 1 1
ε̄(s) = − + α1 − +
η1 α2 s s + α2 s2 α2 s α2 (s + α2 )
(7)
σ0 α1 β β
= + − .
η1 α2 s2 s s + α2
where
α1 η1 η2
β =1− =1− = . (8)
α2 η1 + η2 η1 + η2
" 2 #
t 1 η2 −α2 t
ε(t) = σ0 + 1−e . (9)
η1 + η2 G2 η1 + η2
Hence, the creep compliance function is

" 2 #
t 1 η2 −α2 t
J(t) = + 1−e . (10)
η1 + η2 G2 η1 + η2
Relaxation response. The relaxation function can be determined directly using

Eq.(9.2.9),
1 η1 α2 (s + α2 )
Ȳ (s) = ¯ =
s2 J(s) α1 α2 + s(α1 + α2 β)

η1 α2 α2 − α
= 1+ ,
α1 + α2 β s+α
where
α1 α2 G2 (η1 + η2 ) G2
α= = α1 = , α2 − α1 = .
α1 + α2 β η1 η2 η1
Hence, the relaxation function is
1 η2
Y (t) = η1 δ(t) + G2 e−t/τ2 , τ2 = = .
α1 G2
9.5 Determine the relaxation modulus Y (t) of the three-element model of Fig. 9.2.9(a)
using Eq. (9.2.10) and the creep compliance in Eq. (9.2.37) [that is, verify the result
in Eq. (9.2.42)].
Solution: The creep compliance for the model is given by Eq. (9.2.37)
1 1 t
η
J(t) = + 1 − e− τ , τ = .
k1 k2 k2
The Laplace transform of J(t) is
!
¯ = 1 + 1
J(s)
1
−
1
.
sk1 k2 s s + kη2
Hence, we have
1 k1 s + kη2
Ȳ (s) = 2 ¯ =
s J(s) s s + k1 +k 2
η
k1 k2 1 k12 1
= + .
k1 + k2 s k1 + k2 s + k1 +k
η
2
Taking the inverse Laplace transform, we obtain
k1 k2 k12 k +k
− 1η 2 t k1 k2 k1 + k2
Y (t) = + e = 1 − e−λt + k1 e−λt , λ = ,
k1 + k2 k1 + k2 k1 + k2 η
Figure
which P9-5
is the same as that in Eq. (9.2.42).
9.6 Derive the governing differential equation for the mathematical model obtained by
connecting the Maxwell element in series with the Kelvin–Voigt element (see Fig. P9.6).
k1
• • μ2 k2 ε
μ1 • • • σ
•
Fig. P9.6
ε1 = ε2 , ε2 + ε3 + ε4 = ε, σ1 + σ2 = σ3 = σ4 = σ. (1)
We begin with
σ2 σ σ̇
ε̇ = ε̇2 + ε̇3 + ε̇4 = + + . (2)
µ1 µ2 k2
σ1
Add µ1
= (k1 /µ1 )ε2 to both sides to obtain
k1 σ σ σ̇
ε̇ + ε2 = + + . (3)
µ1 µ1 µ2 k2
Next, we must eliminate ε2 from the above equation. This requires adding appropriate
multiples of ε3 and ε4 to the left side and the corresponding expressions in terms of σ
to the right side of Eq. (3). However, the constitutive equation for ε3 does not exist;
we can only write for ε̇3 . This requires taking the time derivative of Eq. (3) first and
then add (k1 /µ1 )ε̇3 and (k1 /µ1 )ε̇4 to the left side and associated stresses to the right
side:
k1 σ̇ σ̇ σ̈
ε̈ + ε̇2 = + + ,
µ1 µ1 µ2 k2
k1 k1 σ k1 σ̇ σ̇ σ̇ σ̈
ε̈ + ε̇ = + + + + ,
µ1 µ1 µ2 µ1 k 2 µ1 µ2 k2
or
q1 ε̇ + q2 ε̈ = p0 σ + p1 σ̇ + p2 σ̈, (4)
where
k1 k1 1 1 1 k1
p0 = , p1 = + + , p2 = q1 = q2 = 1.
µ1 µ2 k 2 µ1 µ1 µ2 k2 µ1
9.7 Determine the creep compliance J(t) and relaxation modulus Y (t) of the four-element
model of Problem 9.6.
Solution: For creep response, we set σ(t) = σ0 H(t) and obtain
q1 ε̇ + q2 ε̈ = p0 σ0 H(t) + p1 σ0 δ(t) + p2 σ0 δ̇(t), (5)
Using Eq. (9.2.56), the creep response of the model is given by (α = q1 /q2 = k1 /µ1 )

σ0 p0 t 1 p0
1 − e−αt + p2 e−αt

ε(t) = + p1 −
q2 α α α

t µ1 k1 1 1 1 1
1 − e−αt + e−αt

= σ0 + + + −
µ2 k1 k2 µ 1 µ1 µ2 µ2 k2

1 t 1
1 − e−αt .

= σ0 + + (6)
k2 µ2 k1
Hence, the creep compliance is
1 t 1
1 − e−αt .

J(t) = + + (7)
k2 µ2 k1
¯
To compute the relaxation function, we use Eq. (9.2.9). First we compute J:

¯ = 1 + 1 + 1 1− 1
J(s) .
k2 s µ2 s2 k1 s s+α
Then
¯ = s 1 1 sα
s2 J(s) + + .
k2 µ2 k1 s+α
We have
1 k1 k2 µ2 (s + α)
Ȳ (s) = =
s2 J¯ k1 µ2 s2 + [(k1 + k2 )µ2 α + k1 k2 ] s + k1 k2 α

k1 k2 µ2 (s + α) k 1 k 2 µ2 λ 1 − α λ2 − α
= = − ,
(s + λ1 )(s + λ2 ) λ1 − λ2 s + λ1 s + λ2
where λ1 and λ2 are the roots of the polynomial

k1 µ2 s2 + [(k1 + k2 )µ2 α + k1 k2 ] + k1 k2 α = 0
or
√ √
−b + b2 − 4ac −b − b2 − 4ac
λ1 = , λ2 = , a = k1 µ2 , b = µ2 α(k1 +k2 )+k1 k2 , c = k1 k2 α.
2a 2a
Note that for all positive values of the parameters, λ1 and λ2 are negative and therefore
the solution may become unbounded for large values of time t. The relaxation function
is given by
k 1 k 2 µ2 h i
Y (t) = (λ1 − α)e−λ1 t − (λ2 − α)e−λ2 t .
λ1 − λ2
9.8 Derive the governing differential equation for the mathematical model obtained by
connecting the Maxwell element in parallel with the Kelvin-Voigt element (see Fig.
P9.8).
Figure P9-7 k2 η2
•
σ • k1 • σ
• •
η1
• •
Fig. P9.8

(1) (2) (1) (2)
ε1 = ε2 = ε3 = ε, ε3 + ε3 = ε3 , σ1 + σ2 + σ3 = σ, σ3 = σ3 = σ3 . (1)
We begin with
(1) σ̇3
(2) σ3
ε̇ = ε̇3 = ε̇3 + ε̇3 = + . (2)
k2 η2
σ2 = k1 ε, σ1 = η1 ε̇.
From the above relations, we obtain
k1 η1 k1 η1 σ̇ σ
ε̇ + ε̇ + ε̈ + ε + ε̇ = + . (3)
k2 k2 η2 η2 k2 η2
or
q0 ε + q1 ε̇ + q2 ε̈ = p0 σ + p1 σ̇, (4)
where
1 1 k1 k1 η1 η1
p0 = , p1 = , q0 = , q1 = 1 + + , q2 = .
η2 k2 η2 k2 η2 k2
Figure P9-8
9.9 Derive the governing differential equation of the four-parameter solid shown in Fig.
P9.9. Show that it degenerates into the Kelvin–Voigt solid when its components parts
are made equal.
k1 k2
• • • •
σ • η1 • • η2 • σ
• • • •
Fig. P9.9
Solution: From Eq. (9.2.22), we have
σ = k1 ε1 + η1 ε̇1 , σ = k2 ε2 + η2 ε̇2 . (1)
Taking Laplace transform of the equations (with zero initial conditions), we obtain
σ̄ = (k1 + η1 s) ε̄1 , σ̄ = (k2 + η2 s) ε̄2 .
Solving the first equation for ε̄1 and adding k2 + η2 s times ε̄ to the second equation,
we arrive at
k1 + η 1 s
1+ σ̄ = (k1 + η1 s) ε̄,
k2 + η 2 s
or
[k1 + k2 + (η1 + η2 ) s] σ̄ = k1 k2 + s (k1 η2 + k2 η1 ) + η1 η2 s2 ε̄.

(2)
Taking the Laplace inverse, we obtain
p0 σ + p1 σ̇ = q0 ε + q1 ε̇ + q2 ε̈, (3)
where
p0 = k1 + k2 , p1 = η1 + η2 , q0 = k1 k2 , q1 = k1 η2 + k2 η1 , q2 = η1 η2 .
When k1 = k2 and η1 = η2 , and replacing ε with 2ε, we obtain
kσ + η σ̇ = k2 ε + 2kη ε̇ + η 2 ε̈,
which is the same as that obtained by combining the equations of Kelvin–Voigt model:
σ = kε + η ε̇, σ̇ = kε̇ + η ε̈.
9.10 Determine the creep compliance J(t) and relaxation modulus Y (t) of the four-element
model of Problem 9.8.
Solution: The creep response can be directly obtained from Eq. (9.2.52):
(
e−αt e−βt

σ0 1
ε(t) = p0 − +
q2 αβ α(β − α) β(β − α)
−αt )
e−βt αe−αt βe−βt

e
+ p1 − + p2 − + .
(β − α) (β − α) (β − α) (β − α)
where α and β are the roots of the equation q2 s2 + q1 s + q0 = 0:

1 1
q q
α= q1 − q12 − 4q2 q0 , β = q1 + q12 − 4q2 q0 .
2q2 2q2
For the case at hand, we have
1 1 k1 k1 η1 η1
p0 = , p1 = , q0 = , q1 = 1 + + , q2 = .
η2 k2 η2 k2 η2 k2
The quantity under the radical sign is clearly positive and both α and β are positive
roots. The creep compliance is given by
(
e−αt e−βt

1 1
J(t) = p0 − +
q2 αβ α(β − α) β(β − α)
−αt )
e−βt αe−αt βe−βt

e
+ p1 − + p2 − + .
(β − α) (β − α) (β − α) (β − α)
The relaxation response is given by [see Eq. (9.2.54); α = p0 /p1 = k2 /η2 ]

ε0 h q 0 i
1 − e−αt + q1 e−αt + q2 δ(t) − αe−αt

σ(t) =
p1 α

k1 k1 η1 η1
1 − e−αt + 1 + e−αt + δ(t) − αe−αt

= k2 ε 0 +
k2 k2 η2 k2
−αt
= ε 0 k1 + k2 e + η1 δ(t) .
Hence, the relaxation modulus is
Y (t) = k1 + k2 e−αt + η1 δ(t).
9.11 If a strain of ε(t) = ε0 t is applied to the four-element model of Problem 9.8, determine
the stress σ(t) using a suitable hereditary integral [use Y (t) from Problem 9.10].
Solution: From Eq. (9.3.7) we have
dε(t0 ) 0
Z t
σ(t) = Y (t)ε0 + Y (t − t0 ) dt
0 dt0
Z th
0
i
= k1 + k2 e−αt + η1 δ(t) ε0 + ε0 k1 + k2 e−α(t−t ) + η1 δ(t − t0 ) dt0

0
−αt k2
1 − e−αt + η1 H(t) .

= k1 + k2 e + η1 δ(t) ε0 + ε0 tk1 +
α
9.12 For the three-element model of Fig. 9.2.9(b), determine the stress σ(t) when the applied
strain is ε(t) = ε0 + ε1 t, where ε0 and ε1 are constants.
Solution: From the solution of Problem 9.3, the relaxation modulus is

η
Y (t) = k1 + k2 e−t/τ , τ = .
k2
Hence, we have from Eq. (9.3.7) the result
dε(t0 ) 0
Z t
σ(t) = Y (t)ε0 + Y (t − t0 ) dt
0 dt0
Z t 0

= k1 + k2 e−t/τ ε0 + ε1 k1 + k2 e−(t−t )/τ dt0
0

−t/τ
k2
= k1 + k2 e ε 0 + ε 1 k1 t + 1 − e−t/τ .
τ
9.13 Determine expressions for the (Laplace) transformed modulus Ē(s) and Poisson’s ratio
ν̄ in terms of the transformed bulk modulus K̄(s) and transformed shear modulus Ḡ(s).
Solution: We have the following relation between the elastic modulus E, G, and K:
9KG
E= .
3K + G
Now replace the elastic material constants with the corresponding transformed vis-
coelastic functions E ∗ (s) = sĒ(s), G∗ (s) = sḠ(s), and K ∗ (s) = sK̄(s),
9K ∗ (s)G∗ (s) 9s2 K̄(s)Ḡ(s)

E ∗ (s) = → sĒ(s) =
3K ∗ (s) + G∗ (s) 3sK̄(s) + sḠ(s)
or
9K̄(s)Ḡ(s)
Ē(s) = .
3K̄(s) + Ḡ(s)
Similarly, we have the following relation between ν, K, and G:

3K − 2G
ν= .
2(3K + G)
Replacing the elastic material constants with the corresponding transformed viscoelas-
tic functions, we obtain
3K ∗ (s) − 2G∗ (s) 3K̄(s) − 2Ḡ(s)
ν ∗ (s) = → sν̄(s) =
2[3K ∗ (s) + G∗ (s)] 2[3K̄(s) + Ḡ(s)]
9.14 Evaluate the hereditary integral in Eq. (9.3.4) for the three-element model of Fig.
9.2.9(a) and stress history shown in Fig. 9.3.3.
Solution: The creep compliance for the model is given by Eq. (9.2.37)
1 1 t
η
J(t) = + 1 − e− τ , τ = .
k1 k2 k2
The stress history from Fig. 9.3.3 is

t , t ≤ t1
σ(t) =
σ1 H(t), t > t1
The strain is given by

t
dJ(t0 )
Z
ε(t) = J(0) σ(t) + 0
σ(t − t0 ) dt0
0 dt
1 t
Z
t 0
= + (t − t0 )e−t /τ dt0
k1 η 0
t τ
= + 1 − e−t/τ , for t ≤ t0
k1 k2
σ1 t
Z
σ1 0
ε(t) = t+ δ(t − t0 )e−t /τ dt0
k1 k2 0

t 1
= σ1 + e−t/τ , for t > t0 .
k1 k2
9.15 Given that the shear creep compliance of a Kelvin–Voigt viscoelastic material is
1
1 − e−t/τ ,

J(t) =
2G0
where G0 and τ are material constants, determine the following properties for this
material:
(a) shear relaxation modulus, 2G(t),
(b) the differential operators P and Q of Eq. (9.1.1),
(c) integral form of the stress–strain relation, and
(d) integral form of the strain–stress relation.
Solution: (a) The constitutive relations of the deviatoric strains and stresses in Eqs.
(9.3.19) and (9.3.21):
Z t 0 Z t 0
dσij 0 dεij
ε0ij (t) = Js (t − t0 ) dt 0
, σ 0
ij (t) = 2 G(t − t ) dt0 .
−∞ dt0 −∞ dt0
Taking the Laplace transform of the above relations, we obtain
ε̄0ij (s) = sJ(s) 0

¯ σ̄ij (s), 0
σ̄ij (s) = 2sḠ(s) ε̄0ij (s).
¯ = 1/s(1 + τ s)]
Thus, we have [J(s)
1 1 + τs
2Ḡ(s) = 2 ¯ = 2G0 → 2G(t) = 2G0 H(t) + τ δ(t) ,
s J(s) s
where H(t) and δ(t) are the Heaviside step function and Direct delta function, respec-
tively.
(b) The transformed stress–strain relations are

0
σ̄ij (s) = 2sḠ(s) ε̄0ij (s) = 2G0 (1 + τ s) ε̄0ij (s).
Taking the inverse transform, we obtain

dε0ij (t)

0
σij (t) = 2G0 τ + ε0ij (t) → P (σij
0
) = Q(ε0ij ),
dt
and thus we have
d
P = 1, Q = 2G0 τ +1 .
dt
(c) The integral form of the stress–strain relation is
dε0ij dε0ij (t0 ) 0

Z t Z t
0
σij (t) = 2 G(t − t0 ) 0 dt0 = 2G(t)ε0ij (0) + 2 G(t − t0 ) dt
−∞ dt 0 dt0
dε0ij (t0 ) 0
Z t
= 2G0 ε0ij (0) H(t) + τ δ(t) + 2G0 H(t − t0 ) + τ δ(t − t0 )

dt .
0 dt0
Because J(0) = 0, we have ε0ij (0) = 0. Hence, we have
dε0ij 0
Z t
0
H(t − t0 ) + τ δ(t − t0 )

σij (t) = 2G0 dt ,
0 dt0
which can be shown to be equal to the differential form given in Part (b):
dε0ij dε0ij (t)

Z t
0
σij (t) = 2G0 H(t − t0 ) 0 dt0 + 2G0 τ
0 dt dt
t 0
dεij (t)
= −2G0 ε0ij (t − t0 ) + 2G0 τ
0 dt
0
dε ij (t)
= 2G0 ε0ij (t) + τ
dt
(d) The strain-stress relation is obtained as follows:

Z t 0 0
1 0 dσij (t ) 0
ε0ij (t) = 1 − e−(t−t )/τ dt
2G0 0 dt0
0
(t0 ) 0 0
Z t Z t
1 dσij −(t−t0 )/τ dσij (t ) 0
= − e dt
2G0 0 dt0 0 dt0
t
1 t −(t−t0 )/τ 0 0
Z
1 0 0 0
= σij (t) − σij (0) − e−(t−t )/τ σij 0
(t0 ) − e σij (t ) dt0
2G0 0 τ 0
Z t
1 0 1 0
= −σij (0) 1 − e−(t/τ ) + e−t /τ σij0
(t − t0 ) dt0
2G0 τ 0
Z t
0 1 0
= −σij (0)J(t) + e−t /τ σij0
(t − t0 ) dt0 .
2G0 τ 0
9.16 The strain in a uniaxial viscoelastic bar whose viscoelastic modulus is E(t) = Y (t) =
E0 /(1 + t/C) is ε(t) = At, where E0 , C, and A are constants. Determine the stress
σ(t) in the bar.
Solution: We have from Eq. (9.3.7)

Z t Z t
dε 1
σ(t) = Y (t)ε(0) + E(t − τ ) dτ = ACE0 dτ
0 dt 0 C + (t − τ )
t
= ACE0 − ln(C + t − τ ) = ACE0 − ln C + ln(C + t) = ln(1 + t/C).
0
9.17 Determine the free end deflection v v (t) of a cantilever beam of length L, second moment
of inertia I, and subjected to a point load F (t) at the free end, for the cases (a)
F (t) = F0 H(t) and (b) F (t) = F0 e−αt . The material of the beam has the relaxation
modulus of E(t) = Y (t) = A + Be−αt .
Solution: The elastic deflection at the free end of a cantilever beam with a point load
F0 at the free end is given by
F0 L3
v e (L) = .
3EI
Using the elastic-viscoelastic analogy, we replace E with sĒ(s)

A B
sĒ(s) = s + .
s α+s
(a) For this case, we have F̄ (s) = F0 /s so that we have

F̄ (s)L3 F0 L3
v̄ v (L, s) = = .
3sĒ(s)I 3Is2 Ē(s)
Using the method of partial fractions, we can write
F0 L3

B 1 A+B 1
v̄ v (L, s) = A
+ .
3I(A + B) A s + A+B α A s
The inverse transform is given by (let E(0) = A + B ≡ E0 )

F0 L3

B − Aα t E0
v v (L, t) = − e E0 + H(t) .
3E0 I A A
(b) For this case, we have F̄ (s) = F0 /(s + α). Hence, the free end deflection is
F0 L3

1
v̄ v (L, s) = A
.
3I(A + B) s + A+B α
The inverse transform is given by

F0 L3 − Aα t
v v (L, t) = e E0 .
3E0 I
9.18 A cantilever beam of length L is made of a viscoelastic material that can be represented
by the three-parameter solid shown in Fig. 9.2.9(a). The beam carries a load of
F (t) = F0 H(t) at its free end. Assuming that the second moment of area of the beam
is I, determine the tip deflection.
Solution: The elastic deflection of the tip is v e (L) = F0 L3 /3EI, where E is Young’s
modulus. The viscoelastic solution for the tip deflection is
P̄ (s)L3 F0 L3 1
v̄ v (L, s) = = .
3IsĒ(s) 3I s2 Ē(s)
From Eq. (9.2.42), the relaxation modulus is given by

q0 t
q
1 t p1
E(t) = Y (t) = 1 − e− τ + e− τ , τ = ,
p0 p1 p0
where
k2 η
p0 = 1 + , p1 = , q0 = k2 , q1 = η.
k1 k1
Then
q0 1 q0 q1 1 c c + c1 s
Ē(s) = − − = 0 2 ,
p0 s p0 p1 p0
+s s(s + c2 )
p1
where
q0 q1 p0
c0 = , c1 = , c2 = .
p0 p1 p1
Then the transformed deflection of the viscoelastic beam is
F0 L3 F0 L3 1 1

1 c0 − c1 1
v̄ v (L, s) = = + c c .
3I s2 Ē(s) 3I c0 s c1 c0 s + 0c12
The inverse transform of the above expression is
F0 L3 p0

q0 p1 − q0 p1 −(q1 /p1 )t
v v (L, t) = H(t) + e .
3I q0 q1 q0
9.19 A simply supported beam of length L, second moment of area I is made from the
Kelvin–Voigt type viscoelastic material whose compliance constitutive response is

1
J(t) = 1 − e−t/τ ,
E0
where E0 and τ are material constants. The beam is loaded by a transverse distributed
load x 2
q(x, t) = q0 1 − t = f (x) g(t),
L
where q0 is the intensity of the distributed load at x = 0 and g(t) = t2 . Determine the
deflection and stress in the viscoelastic beam using the Euler–Bernoulli beam theory.
Solution: The deflection of the elastic beam under the static load f (x) is
2 4
q0 L4

x x x
v e (x) = 1− 7 − 10 1 − +3 1− .
360EI L L L
The transformed deflection of the viscoelastic beam is
2 4
ḡ(s) q0 L4

x x x
v̄ v (x, s) = 1− 7 − 10 1 − +3 1−
sĒ(s) 360I L L L
4
2 4
2 q0 L x x x
= 3 1− 7 − 10 1 − +3 1−
s (1 + τ s) 360E0 I L L L
Using the factorization
2τ 2 1

2 1 1 1 1 1
= − + 2 3 −
E0 s3 ( τ1 + s) E0 s τ s2 τ s s+ 1
τ
the deflection of the viscoelastic beam is obtained as

2 4
q0 L4

x x x
v v (x, t) = 1− 7 − 10 1 − +3 1− h(t),
360I L L L
where
2τ 2 τ 2 t h t i
h(t) = 1 − e−t/τ + −2 .
E0 E0 τ τ
The stress σ(x, t) is given by
∂ 2 vv q0 L2 y

M (x, t)y x x
σ(x, t) = = −Ey = 1 − h(t),
I ∂x2 60I L L
where y denotes the transverse distance through the beam height.
Alternative method(Alfrey’s analogy). The load is of the form
x
q(x) = f (x)g(t), where f (x) = q0 1 − , g(t) = t2 .
L
The deflection of the elastic beam under the static load f (x) is
2 4
q0 L4

e x x x
v (x) = 1− 7 − 10 1 − +3 1− .
360I L L L
The deflection of the viscoelastic beam is
v v (x, t) = v e (x) h(t),
where h(t) is a function of time such that

dh
P (g(t)) = Q(h(t)) → t2 = E0 τ +h
dt
or
dh t2 1
+ αh(t) = , α= .
dt E0 τ τ
This equation is to be solved subjected to the condition P (g(0)) = Q(h(0)) which gives
h(0) = 0. The solution is given by
2τ 2 τ 2 t h t i
h(t) = 1 − e−t/τ + −2 .
E0 E0 τ τ
Hence, the deflection of the viscoelastic beam is
2 4
q0 L4

x x x
v v (x, t) = 1− 7 − 10 1 − +3 1− h(t).
360I L L L
9.20 The pin-connected structure shown in Fig. P9.20 is made from an incompressible
viscoelastic material whose shear response can be expressed as
η d d
P =1+ , Q=η ,
µ dt dt
Figure P9-19
where η and µ are material constants. The structure is subjected to a time-dependent
vertical force F (t), as shown in Fig. P9.20. Determine the vertical load F (t) required
to produce this deflection history. Assume that member AB has a cross-sectional area
A1 = 9/16 in.2 and member BC has a cross-sectional area A2 = 125/48 in.2 .
C δ (t )
5
4 2 δ1
A1 = 9/16 in.2 3
A2 = 125/48 in.2 δ0
A 1 B 0 t
0 t0
L
F0 , δ
Fig. P9.20
Solution: The elastic solution can be obtained using the Castigliano’s theorem I. The
total complementary strain energy is
1 N12 L1 N 2 L2 1 (−3P/4)2 L (5P/4L)2 (5L/3)

U∗ = + 2 = +
2 A1 E A2 E 2 A1 E A2 E
2

1 (9/16)L (125/48)L P L
= + =
2 A1 E A2 E E
The deflection at point B is
∂U ∗e 2P L
δB =
= .
∂P E
The Laplace transformed viscoelastic solution is
2P̄ (s)L 1
δ̄ v (s) = , → P̄ (s) = sĒ(s)δ̄ v (s)
sĒ(s) 2L
The viscoelastic constitutive equation is
η 0 −µt
σ 0 (t) + σ̇ (t) = 2η ε̇0 (t) → 2G(t) = 2µe η .
µ
Because the material is incompressible, εkk = 0 (or ν = 0.5), we have
−µ t 3µ
E(t) = 3G(t) = 3µ e η , → Ē(s) = .
s + µη
The displacement history can be represented as

δ0 (δ1 − δ0 ) −t0 s
δ v (t) = δ0 H(t) + (δ1 − δ0 )H(t − t0 ) → δ̄ v (s) = + e .
s s
Thus, we have
1 1 h i 3µ
F̄ (s) = sĒ(s)δ̄ v (s) = δ0 + (δ1 − δ0 )e−t0 s .
2L 2L s + µη
The Laplace inverse is given by

3µ h −µt −
µ(t−t0 ) i
t 1 h i
F (t) = δ0 e η + (δ1 − δ0 )e η = δ0 E(t) + (δ1 − δ0 )E(t − t0 ) .
2L 2L
Figure P9-20
9.21 Consider a hallow thick-walled spherical pressure vessel composed of two different vis-
coelastic materials, as shown in Fig. P9.21. Formulate (you need not obtain complete
solution to) the boundary value problem from which the stress and displacement fields
may be determined.
External
pressure, p1 (t )
Internal
pressure,
Material 1 b c p2 (t )
Material 2 Dia. a
Fig. P9.21
Solution: The general elastic solutions of the problem for the radial displacement and
stresses are of the form (see Example 7.3.1)
Bi
ur (r) = Ai r + ,
r2
dur 2λ 4µ
σrr (r) = (2µ + λ) + ur (r) = (2µ + 3λ)Ai − 3 Bi ,
dr r r
dur 2(µ + λ) 4µ
σθθ (r) = σφφ (r) = λ + ur (r) = (2µ + 3λ)Ai + 3 Bi ,
dr r r
where µ and λ are the Lame’ constants
E Eν E
µ=G= , λ= , 2µ + 3λ = ,
2(1 + ν) (1 + ν)(1 − 2ν) 1 − 2ν
and Ai and Bi are constants of integration, with subscript i referring to the material
number (i = 1, 2). The boundary and interface conditions of the problem are
(2) (1) (1) (2)
σrr (a) = −p2 , σrr (c) = −p1 , σrr (b) = σrr (b), u(1) (2)
r (b) = ur (b).
Using the conditions on the stresses, we obtain
B̂1 B̂2 B̂1 B̂2

Â1 = −p1 + , Â2 = −p2 + 3 , Â1 − 3 = Â2 − 3 , (1)
c3 a b b
where
E 2E
Âi = (2µ + 3λ)Ai = Ai , B̂i = 4µBi = Bi .
1 − 2ν 1+ν
Solving for B̂1 in terms of B̂2 , we obtain
h i c 3 1
B̂1 = (pi − p0 )a3 b3 − (b3 − a3 )B̂2 . (2)
a (c3 − b3 )
Using the displacement continuity at the interface, we obtain
(1 − 2ν1 )b B̂1 (1 + ν1 ) (1 − 2ν2 )b B̂2 (1 + ν2 )

Â1 + = Â2 + ,
E1 2E1 b2 E2 2E2 b2
B̂1 (1 − 2ν1 )b B̂1 (1 + ν1 ) B̂2 (1 − 2ν2 )b B̂2 (1 + ν2 )
−p1 + 3
+ 2
= −p2 + 3 + ,
c E1 2E1 b a E2 2E2 b2
or
h i
B̂1 = 2b3 (1 − 2ν1 )p1 − 2b3 m(1 − 2ν2 )p2 + mB̂2 (1 + ν2 ) + 2(1 − 2ν2 )(b/a)3
h i−1
× (1 + ν1 ) + 2(1 − 2ν1 )(b/c)3 (3)
where m = E1 /E2 . Equating the expressions in Eqs. (2) and (3), we can determine
B̂2 . Consequently, B̂1 , Â1 , and Â2 (hence, Ai and Bi ) can be determined. Thus the
elastic solutions for displacements and stresses are known.
The Laplace transformed viscoelastic solutions for the displacements and stresses are
obtained from
B̄i (s)
ūr (r, s) = Āi (s)r + ,
r2
4µ
σrr (r, s) = (2µ + 3λ)Āi (s) − B̄i (s),
r3
4sµ̄(s)
σθθ (r, s) = σφφ (r, s) = [2sµ̄(s) + 3sλ̄(s)]Āi (s) + B̄i (s),
r3
where Āi (s) and B̄i (s) are the same as Ai and Bi with ν and E replaced by sν̄(s)
and sĒ(s), respectively. It remains then to obtain the Laplace inversion of the above
expressions and obtain the viscoelastic solutions.
9.22 The linear elastic solution for axial stress σxx (x, y) and transverse displacement v(y),
based on the Euler–Bernoulli beam theory, of a cantilever beam of length L, flexural
stiffness EI, and loaded at the free end with F0 , as shown in Fig. P9.22, is
F0 xy F0 L3 x x2
σxx = − , v(x) = 2−3 + 2 .
EI 6EI L L
Figure P9-22
Determine the viscoelastic counterparts of the stress and displacement using the vis-
coelastic analogy and Kelvin–Voigt model.
y
x
F0
Fig. P9.22
Solution: The stress, being independent of the material parameters, will remain the
same as the elastic solution. The displacement by viscoelastic analogy is
F0 L3 x x2
v̄(x, s) = 2−3 + 2
6sĒ(s)I L L
F0 L3 x x2
= 2−3 + 2
sȲ (s)6I L L
3
1 F0 L x x2
= 2−3 + 2 .
sη + k 6I L L
Hence,
F0 L3 x x2
v(x, t) = 2 − 3 + 2 e−kt/η .
η6I L L
¯ = 1/sȲ (s) [see Eqs. (9.2.9)and (9.2.26)], we obtain
Alternatively, using sJ(s)
F0 L3 x x2
v̄(x, s) = 2−3 + 2
6sĒ(s)I L L
3
F0 L x x2
= 2−3 + 2
sȲ (s)6I L L
3
¯ F 0 L x x2
= sJ(s) 2−3 + 2 ,
6I L L
or
F0 L3 x x2
v(x, t) = 2 − 3 + 2 e−kt/η .
η6I L L

Solutions Manual To Introduction To Continuum Mechanics: Second Edition

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Solutions Manual To Introduction To Continuum Mechanics: Second Edition

Uploaded by

Copyright:

Available Formats

Solutions Manual to

1.1 Newton’s second law can be expressed as

Convection from lateral

1.3 The Euler–Bernoulli hypothesis concerning the kinematics of bending deformation of

where α is a number 0 ≤ α ≤ 1. Again, dividing throughout with ∆x and taking the

time rate of change in mass = mass inflow - mass outflow.

The above statement for the problem at hand becomes

Chapter 2: VECTORS AND TENSORS

which is the required equation of the plane.

Solution: Consider the parallelogram formed by points O, A, C, and B, as shown in

FigP2-6 B = α C − β D; B = (1 − α)C − (1 − β)D (2)

2.7 Represent a tetrahedron by the three non-coplanar vectors A, B, and C, as shown in

which is zero because terms cancel out.

(A · B)2 + (A × B) · (A × B) = |A|2 |B|2 ,

Solution: Let C = A × B, and consider the expression

B × C = B × (A × B) = (B · B)A − (B · A)B. (2)

A · (B × C) = A · [(B · B)A − (B · A)B] = |A|2 |B|2 − (A · B)2 ,

from which the desired identity follows.

r1 = A − 3B + 2C, r2 = 2A − 5B + 3C, r3 = A − 5B + 4C.

Solution: Write the linear relation among the three vectors,

(α + 2β + γ)A − (3α + 5β + 5γ)B + (2α + 3β + 4γ)C = 0.

2.11 Determine whether the following set of vectors is linearly independent:

A = 2ê1 − ê2 + ê3 , B = −ê2 − ê3 , C = −ê1 + ê2 .

Here êi are orthonormal unit base vectors in <3 .

Solution: Set the linear combination of the vectors to zero

A = 2ê1 − ê2 + ê3 , B̂ = ê2 − ê3 , C = −ê1 + ê2

are linearly independent.

Solution: (a) Note that

e1 × e2 = 2 î − 2 ĵ − k̂, e2 × e3 = 4 î − 5 ĵ − 3k̂, e3 × e1 = î − ĵ − k̂, [e1 e2 e3 ] = 1.

(b) The magnitudes of the base vectors are

(c) We have A = A1 e1 + A2 e2 + A3 e3 = e1 + 2e2 + 3e3 . We obtain

A1 = (e1 + 2e2 + 3e3 ) · e1 = 2 − 2 × 3 − 3 × 3 = −13

Solution: (a) Set

2.14 Prove the following vector identity using index notation:

A × (B × C) = Ai êi × (ê` ejk` Bj Ck ) = Ai Bj Ck ejk` epi` êp

2.15 Prove the following vector identity using index notation:

(A · B)2 + (A × B) · (A × B) = |A|2 |B|2 .

Solution: We begin with

(A × B) · (A × B) = (Ai Bj eijk êk ) · (Am Bn emnp êp )

from which the desired identity follows.

where A and B are vector functions.

Solution: (a) We have (A × B) · (B × C) × (C × A)

= (eijk Ai Bj êk ) · [(erst Br Cs êt ) × (emnp Cm An êp )]

(A × B) × (C × D) = (eijk Ai Bj êk ) × (emnp Cm Dn êp )

2.19 Prove that

2.20 Establish the following identities :

where we have used the identity of the form

(c) Using the e-δ identity we can write

(d) From Part (b), we have

That is, L is an orthogonal matrix.

(b) ê01 is along the 0

ê1 ê2 ê3

Determine the direction cosines of the transformation.

Solution: Follows from the definition

Determine the transformation matrix.

Solution: Follows from the definition

2.25 Write the following sets of equations in matrix form [A]{X} = {Y }:

Solution: The matrix representation of the linear equations is

Solution: (a) First we compute the adjoints, as shown below.

The determinant is given by (expansion by first row)

2.27 Find the inverses of the coefficient matrices in Problem 2.25.

2.28 Determine if the following matrices are positive: