Continuum Mechrnics Vir Problems RND Exercises: Parti Theorq and Problems

CONTINUUM MECHRNICS VIR
PROBLEMS RND EXERCISES

Parti; Theorq and Problems
WORLD SCIENTIFIC SERIES ON NONLINEAR SCIENCE
Editor: Leon O. Chua

University of California, Berkeley
Series A. MONOGRAPHS AND TREATISES

Published Titles
Volume 9: Frequency-Domain Methods for Nonlinear Analysis: Theory and Applications
G. A. Leonov, D. V. Ponomarenko, and V. B. Smimova
Volume 12: Attractors of Quasiperiodically Forced Systems
T. Kapitaniak and J. Wojewoda
Volume 13: Chaos in Nonlinear Oscillations: Controlling and Synchronization
M. Lakshmanan and K. Murali
Volume 14: Impulsive Differential Equations
A. M. Samoilenko and N. A. Perestyuk
Volume 15: One-Dimensional Cellular Automata
B. Voorhees
Volume 16: Turbulence, Strange Attractors and Chaos
D. Ruelle
Volume 17: The Analysis of Complex Nonlinear Mechanical Systems: A Computer
Algebra Assisted Approach
M. Lesser
Volume 19: Continuum Mechanics via Problems and Exercises
Edited by M. E. Eglit and D. H. Hodges
Volume 20: Chaotic Dynamics
C. Mira, L Gardini, A. Bamgola andJ.-C. Cathala
Volume 21: Hopf Bifurcation Analysis: A Frequency Domain Approach
G. Chen andJ. L Moiola
Volume 23: Nonlinear Dynamics in Particle Accelerators
R. DilSo and ft Alves-Pims
Forthcoming Titles
Volume 18: Wave Propagation in Hydrodynamic Flows
A. L Fabrikant and Y. A. Stepanyants
Volume 22: Chaos and Complexity in Nonlinear Electronic Circuits
M. J, Ogorzalek
Volume 24: From Chaos to Order
G. Chen andX. Dong
Series Editor: Leon 0. Chua
CONTINUUM MECHANICS VIA

PROBLEMS UNO EXERCISES
ParH: Theorq and Problems
Editors
Margarita E. Eglit
Moscow State University
Russia
Dewey H. Hodges
Georgia institute of Technology
U.S.A.
Co-Authors
Margarita E. Eglit Andrei G. Kullkovsky
Alexander N. Golublatnikov Alexander G. Petrov
Jacov A. Kamenjarzh irlna s. Shlklna
Vladimir P. Karllkov Elena I. Sveshnlkova
Translation from Russian.- A. N. Tlatlushkln

Figures: E. N. Pascnenko
Y f o World Scientific
l r Singapore • New Jersey • London • Hong Kong
Published by
World Scientific Publishing Co. Pte. Ltd.
P O Box 128, Farrer Road, Singapore 912805
USA office: Suite IB, 1060 Main Street, River Edge, NJ 07661
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
British Library Cataloguing-ln-Publicatkm Data

A catalogue record for this book is available from the British Library.
Cover illustration: From "Torus Doubling in Four Weakly Coupled Oscillators" International Journal
of Bifurcation and Chaos, Vol. 5, No. 1, February 1995, p. 237.
CONTINUUM MECHANICS VIA PROBLEMS & EXERCISES

Parti: Theory and Problems
Copyright © 1996 by World Scientific Publishing Co. Pte. Ltd.
A11 rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and retrieval
system now known or to be invented, without written permission from the Publisher.
For photocopying of material in this volume, please pay a copying fee through the Copyright
Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to
photocopy is not required from the publisher.
ISBN 981-02-2546-6 (Set)

ISBN 981-02-2962-3 (Part I)
This book is printed on acid-free paper.
Printed in Singapore by Uto-Print

Contents
Preface ix
Nomenclature xi
I THEORY AND PROBLEMS 1

1 Basic Concepts used for Description of Motion and Deformation of
a Continuum 3
1 Lagrangian and Eulerian Description of Motion 3
2 Tensors and their Cartesian Components 9
3 Curvilinear Coordinate Systems 13
4 Deformation. Deformation Rate. Vorticity. 26
5 Principles of Symmetry and Tensor Functions 43
2 General Laws and Equations of Continuum Mechanics 49

6 Summary of General Laws and Equations 49
7 Mass Conservation Law. Continuity Equation 52
8 Stress Tensor 55
9 Differential Equations of Motion and Equilibrium 63
10 Using the Conservation Laws in Integral Form for Calculation of Forces
and Moments Acting on Bodies Moving in a Fluid (Control Surface
Method) 69
11 Angular Momentum Equations 77
3 Thermodynamics of Continua 85
12 Basic Laws and Concepts of Thermodynamics 85
13 The First Law of Thermodynamics.
Energy Equation. Perfect Gas 93
14 The Second Law of Thermodynamics. Entropy. Gibbs Identity. ... 98
v
vi CONTENTS
15 Restrictions Imposed by The Laws of Thermodynamics on the Form

of Constitutive Relationships 100
16 Thermodynamics of Media with Internal Angular Momentum 104
4 Discontinuity Surfaces in Continuum Media 109

17 Conditions across Discontinuity Surfaces 109
18 Discontinuity Surfaces in the Lagrangian Description 114
5 Fluid Mechanics 117

19 Summary of Equations of Fluid Mechanics 117
20 Statics of Fluid 119
21 Dynamics of an Incompressible Ideal Fluid 124
22 Dynamics of an Incompressible Viscous Fluid 142
23 Waves on a Surface of Heavy Imcompressible Liquid 155
24 Mechanics of a Compressible Fluid 162
6 Elasticity Theory 179

25 Models of Elasticity Theory 179
26 Linearly Elastic Solid 184
27 Nonlinear Elasticity 205
28 Couple Elasticity and Averaging in Media with Microstructure . . . . 206
7 Inelastic Solids 215

29 Plastic Flow Theory 215
30 Rate-Dependent Effects in Solids 223
8 Basic Notions of Relativistic Kinematics and Dynamics. General

Properties of Electromagnetic Field 227
31 Lorentz Transformations. Minkowski Space 227
32 Concepts of Relativistic Kinematics and Dynamics 230
33 Maxwell Equations 231
9 Models of Media Interacting with an Electromagnetic Field 237

34 Magnetohydrodynamics 237
35 Electrohydrodynamics 241
10 Dimensional Analysis and Modeling 245

36 Foundations of the Dimensionality Theory 245
37 Examples of the Application of Dimensional Analysis 250
Bibliography 259
CONTENTS vii
Index 263
This page is intentionally left blank
Preface
The purpose of this book is to help students understand the principles, methods and
applications of continuum mechanics. With this aim the authors consider more than
one thousand problems in continuum mechanics.
The subject of continuum mechanics deals with the behavior of any deformable
media under different physical conditions. Different mathematical models are appli
cable for the same medium, depending on the purpose of the study and the physical
environment. These models correspond to different fields of continuum mechanics
such as fluid and gas mechanics, hydraulics, theory of elasticity and plasticity, creep
theory, strength of materials and others. Students specializing in one of these areas
often pay attention only to that one area. However, a deeper understanding can
be achieved if students are exposed to how general principles work under different
conditions and to the various results that are given by different models for the same
conditions. These aspects the authors tried to demonstrate by collecting problems
associated with the general principles of continuum mechanics, as well as certain par
ticular problems and models. Thus, the characteristic feature of this book is that the
problems collected in it aim at exhibiting both the variety and the unity of the ideas
and methods of continuum mechanics.
Let us mention one example of this approach. There are many phenomena in
which we observe abrupt changes of the velocity, pressure and other parameters. In
continuum mechanics these abrupt changes are usually considered as discontinuities
of the parameters. To describe them one needs to use the conservation laws, the
second law of thermodynamics, the evolutionarity conditions, as well as conditions
for stability and the existence of a structure. In the book the use of all these conditions
is demonstrated not only for shock, detonation and combustion fronts in gas flows
but also for interfaces between water and rain falling on it, for shocks in elastic solids,
for discontinuity of electromagnetic field, for hydraulic jumps in channels, and even
for road traffic.
The problems included in the book are at various levels of difficulty. Some of
them are more or less standard exercises that are necessary for understanding the
subject. Others are much more realistic, requiring extensive creative work and giving
some understanding of state-of-the-art problems. Indeed non-traditional problems
ix
X Preface
in nonlinear elasticity, microinhomogeneous averaging, interaction of continua with

electromagnetic field, viscoelasticity and plasticity theory, applications of tensor anal
ysis (in particular, nonlinear tensor function theory), and creation of new continuum
models are all included.
All the problems in the book are provided with answers, most of them with hints
and solutions. Each group of problems is preceded by a corresponding theoretical
review. These sections help not only in solving the problems but also in understanding
the general structure of continuum mechanics. However, study from this book should
be supplemented by the reading of other texts.
The authors hope that the book will be useful for undergraduate and gradu
ate students, and also for engineers and researchers in mechanics, mathematics and
physics.
A few words about the contributors to this book. The book was written by the pro
fessors of the Department of Hydrodynamics at the M. V. Lomonosov Moscow State
University. The Department Chairman L. I. Sedov did a lot to unify the continuum
mechanics from the set of separate subjects. For many years the authors taught the
basic course of continuum mechanics and its various branches: hydrodynamics, gas
dynamics, elasticity theory, plasticity theory, thermodynamics and electrodynamics
of continua, application of dimensional analysis and modeling of the continuous me
dia. All the authors are also do active research in continuum mechanics. M. E. Eglit
and D. H. Hodges were editors. A. N. Tiatiushkin translated the book from Russian
into English; he also remarked and corrected some inexactitudes in the formulations
of problems. The computer graphics was made by E. N. Paschenko. V. A. Turkov
and V. A. NaUotova performed a computer typesetting of the book; this great work
required the thorough understanding of contents of the book. A. G. Kalugin, N.
I. Gvozdovskaya and V. V. Volovoi helped very much to put the manuscript into
proper shape. It is hard to estimate the help of V. G. Sutyrin and A. E. Yakubenko,
without whom the appearance of this book would be impossible. They took part
in the making of the book from the very beginning, typing preliminary versions and
giving very useful advice concerning the contents and form of the book. V. G. Sutyrin
provided the connection between Russian and American editors. V. G. Sutyrin and
E. V. Soutyrina worked very hard making the final version of the book ready to pub
lication. At the first stage of work on the book authors had very useful discussions
with V. V. Rosantzeva on its contents. At last, V. L. Berdichevsky should be noted
who had the idea of publishing the book in English, had read the first chapters of the
book and made a number of important comments.
The authors are grateful to all these persons.
Margarita E. Eglit,
Moscow, August 1, 1995
Nomenclature
It was not the aim of the authors to follow strictly the same notation throughout
the book. Different notations are sometimes used, and the same symbols may denote
different quantities in different sections. Usually the meaning of a symbol is given in
the text. Still there is some standard notation which is used almost everywhere in
the book and listed below.
Coordinates are usually numbered by the upper indices and denoted by xl (i =
1,2,3). Symbols x, y and z are often used to emphasize Cartesian coordinates.
Lagrangian coordinates are usually denoted by f*.
The symbol V* stands for the covariant derivative over coordinate x'. In Cartesian
coordinates
a - speed of sound
c - speed of light; speed of a characteristic; heat capacity per unit mass; concen
tration
Cp - specific heat capacity at constant pressure
cv - specific heat capacity at constant volume
dA - the amount of work for time interval dt
dq - the amount of heat per unit mass that enters a system during time interval dt
drf - uncompensated heat
- rate of entropy production per unit mass
da - element of area
dV - element of volume
e.i - covariant basis vectors
XI
Xll Nomenclature
e* - contravariant basis vectors

e j , e.i - basis vectors of Lagrangian coordinate system in the initial and current states,
respectively
ey - components of the strain rate tensor
g - acceleration due to gravity
g - metric tensor
ffy, g'j - components of the metric tensor
i - specific enthalpy, imaginary unit

k - wave number
k - internal angular momentum per unit mass
n - normal to the boundary
p - pressure
Pij - components of Cauchy stress tensor
pn - stress vector
q - heat flux vector
ql - component of heat flux vector in direction of i '
s - entropy per unit mass
t - time
u - internal energy per unit mass; component of a velocity vector in the direction
of Cartesian coordinate x
v - velocity; component of a velocity vector in the direction of Cartesian coordinate
y
v - velocity vector
w - component of a velocity vector in the direction of Cartesian coordinate z
w - displacement vector
x* - coordinates
Nomenclature xiii
x , y, z - Cartesian coordinates
B - magnetic induction
D - speed of a discontinuity surface
E - Young's modulus
E - electric field strength
T - free energy per unit mass
F - body force per unit mass
Qn - components of couple stress tensor

R - gas constant
S - entropy
T - temperature
U - internal energy; potential of body force
V - volume
W - complex potential
Fr - Froude number
M - Mach number
Pe - Peclet number
Re - Reynolds number
St - Strouhal number
Q - coefficient of thermal expansion;
7 - adiabatic exponent, the ratio Cp/cv

<5i;,<5' - Kronecker symbols
£ij - components of a strain tensor
*ijk - components of the Levi-Civita tensor

XIV Nomenclature
X - one of the viscosity coefficients; one of the elastic coefficients
/x - viscosity coefficient; one of the elastic coefficients

v - kinematic viscosity coefficient
Vi - covariant derivatives
p - density
7r,J - components of the Piola-Kirchhoff stress tensor
a - Poisson coefficient, electrical conductivity coefficient
T - tangential stress
T ,J - viscous stresses
<p - potential of velocity
■0 - stream function
f - Lagrangian coordinates
u) - vorticity vector; angular velocity
T - velocity circulation
Fljk - Christoffel symbols
Each problem is marked by a logo that signifies the level of the problem. The
logo © shows that the problem is not too difficult to be solved. The logo ©
is used for more difficult problems. Very difficult problems which require a deep
understanding of the material are denoted by © .
Chapter 1
Basic Concepts used for

Description of Motion and
Deformation of a Continuum
1 Lagrangian and Eulerian Description of Motion
Within the scope of classical mechanics, all particles of a continuous medium (or,
more briefly, continuum) can be distinguished each from other, i.e., can be individ
ualized. The individual particles are associated with triples of numbers (£1,6.6)-
Such a triple is called the Lagrangian coordinates of the corresponding individual par
ticle. The Lagrangian coordinates are used to indicate this particle, i.e., to serve as
its "name" (in the same manner as integers serve as "names" of particles in the case
of their discrete configuration). The coordinates of the point occupied by a particle
at the initial instant are ordinarily used as the Lagrangian coordinates of this par
ticle. The motion of a continuum and processes taking place in it are described by
fields of physical quantities (velocity, pressure, temperature, etc.). If these quantities
are considered as functions of the Lagrangian coordinates ( 6 1 6 . 6 ) a n d t i m e '. t n e
description is referred to as Lagrangian or material. From this viewpoint, events are
considered to take place in the individual particles. The function r(fi, 6 , 6 , 0 . o r
motion law, is used as a basic kinematical characteristic in the Lagrangian descrip
tion. The vector r ( 6 , 6 . 6 . t) is the position (relative to a chosen frame of reference)
of the particle ( 6 - 6 . 6 ) at the instant t. If a coordinate system is chosen in the
three-dimensional Euclidean space of vectors r (i.e., a one-to-one correspondence is
set between the vectors and triples of numbers r «-» (xu x2, £3)). then the motion law
is represented also by the functions re* = / . ( 6 . 6 . 6 . O . » = !. 2 > 3 - T h e velocity and
3
4 BASIC CONCEPTS
acceleration of the particles of a continuum are defined by the formulae
where £ = (£i>&2,60- In general, the rate of change in some quantity A in an

individual particle of a continuum is called the substantive or total derivative of this
quantity with respect to time. In the Lagrangian description this is simply the partial
derivative dA(£, t)/dt. Physical quantities characterizing motion of a continuum and
processes taking place can be considered as functions of the spatial coordinates (ii, x2,
£3) and time t. From this viewpoint, events are considered to take place at the points
of the space. Such viewpoint is referred to as Eulerian or spatial. The velocity field
v(x\, x 2 ,13, t) is used as a basic kinematical characteristic in the Eulerian description;
spatial coordinates are called Eulerian. The vector v(xi, x 2 , 13, t) is the velocity of a
particle occupying the point (xi,x 2 ,x 3 ) at the instant t. In the Eulerian description
the substantive derivative of some quantity A is denoted with dA(x, t)/dt, where
x = (xi,X2,£3), and computed by the formula
dA{x,t) dA(x,t) . A,dA{x,t) . A.dA(x,t) , ,dA(x,t)
In particular, the acceleration a(x,t) in Eulerian description is determined by the

formula
. dv(x, t) , .dv(x,t) . .dv(x,t) . .dv(x,t)
where Vi(x,t) are velocity components.

Lagrangian and Eulerian viewpoints are equivalent: if processes are described
from one of them, the description from the other is obtained with a simple proce
dure. To change over the description from Lagrangian to Eulerian, the functions
representing the motion law x, = /.(fi.&i&.Oi * = 1.2,3, should be resolved with
respect to the Lagrangian coordinates, i.e., the inverse functions £<» = ga(xi, x 2 , x 3 , t),
a = 1,2,3, should be found. Then, for any quantity, the Lagrangian description of
which ^(Oi&ifoiO is known, the Eulerian description is obtained as the composite
function A(gl(x, t),g2(x, t), g3(x, t), t). To change over the description from Eulerian
to Lagrangian, the solution of the ordinary differential equations
dxi
-jj- =v,(x 1 ,x 1 ,x 3 , t) , 1= 1,2,3,
subjected to the initial conditions
X\\t=0 ~ £l . X2\t=0 = & . X3\t=0 = £3

1. Lagrangian and Eulerian Description ot Motion 5
should be found. This solution x, = /i(£i,f2.&5.0. * = 1,2,3, when found for all
values of the parameters (^1,^2,^3)1 gives the motion law in Lagrangian description,
and (f 1, £2,61) are the Lagrangian coordinates of the particles. Then, for any quantity
B(xi,X2-,X3,t), the Eulerian description of which is known, the Lagrangian descrip
tion is obtained as the composite function B(f\(£,t), fc(£,t),f3(£,t),t). A motion
of a continuum is well, although incompletely, represented by the particle paths and
streamlines. The locus of the positions of a particle (^1,^2.^3) at all instants is called
the path of this particle. A curve, determined for a given instant to, whose tangent
at any point x is in the direction of the velocity vector v(x, to), is called a streamline.
The equation of a streamline at an instant to has the following form
dX\ dl2 d.X3
v\{x,t0) v2{x,t0) v3(x,tQ) '
In general, streamlines depend upon the instant t0 for which they are found. When
a motion is steady, the streamlines do not depend upon the instant to and coincide
with the particle paths. A motion is referred to as steady if the velocity field in the
Eulerian description does not depend upon the time t. In the problems of this sec
tion, (x1.x2.i3) and (£1,^2, £3) denote respectively spatial Cartesian and Lagrangian
coordinates.
— PROBLEMS —
0 1.1 Introduce a spatial coordinate system and Lagrangian coordinates of the

particles in the following cases: a) a rigid body executes a translatory motion in a
fixed direction at constant speed v; b) a rigid body rotates about a fixed axis at
constant angular speed u>.
0 1.2 For translatory motions of a rigid body, find the general form of the velocity
field in the Lagrangian description and that of the motion law.
0 1.3 The motion law of a medium is represented by the functions
xi = £1 + a t £2 , *2 = 62 + b t £1 , x 3 = f3 (a, b = const).
Verify that the numbers (£1,^2,^3) for an individual particle have the meaning of
coordinates X\, x2, X3 of the point occupied by this particle at the instant t = 0. Find
velocity and acceleration fields in the Lagrangian description. What particle occupies
the point (x\,x^,x%) at the instant t = 0?
0 1.4 The motion of a medium is represented by the functions
* ! = & ( ! + £), *2=&(l+2£), 1 3 = 6 (l + £ )

6 BASIC CONCEPTS
(T — const).
a) Find the velocity and acceleration fields in the Lagrangian description, b) A
particle occupied the point (a, b, c) at the instant t = r. Find its position at the
instant t = 3r.
0 1.5 Consider the functions £a = ga(xi,x2,X3,t), a = 1,2,3, inverse, at instant
t, to the functions ij = /,(î,^2,^3> 0> * = 1.2,3, representing a motion law. What is
the physical meaning of their values? Find the values of the substantive derivatives
dta/dt.
0 1.6 Find velocity and acceleration fields in the Lagrangian and Eulerian de
scriptions if the motion is a) a three-axis extension:
i i = a(t) & , x2 = b{t) & , i 3 = c(t) & ;
b) a simple shear:
*i = fI + b(t) & , x2 = & , x3 = & ;
c) a uniform deformation and simultaneous rotation of a body with a fixed point:
x, = i4 n (0€i + >M*)& + AaMk , (det\\A„\\ ± 0).
0 1.7 Introduce Lagrangian coordinates and find the motion law of a medium if
the motion is described by the velocity field
Xi 2ti 2 3<2X3 . „,
vi = —— , v2 = , v3 = (r = const > 0 .
t +T t* + T* ts + Tl
0 1.8 Introduce Lagrangian coordinates and find the motion law of a medium, the
streamlines, and the particle paths, if the motion is described by the velocity field
•> — ii??- * - irP' * - ° <" - v^>-«(«> "■") * - ^w

i = 1,2,3 {R = ijxl + xl + xl), Q{t) > 0; c) vx = -Axu v7 = Bx2, v3 =
0, A = const > 0, B = const > 0.
© 1.9 Introduce Lagrangian coordinates and find the motion law of a medium if
the motion is described by the velocity field
vi = -A(t)xi , v2 = B(t)x2 , v3 = 0 ,
A{t) > 0 , B(t) > 0 .
Find the streamlines and compare them with the streamlines for the special case
A, B = const (Problem 1.8). Give an example of the functions A(t), B(t) at which
the streamlines and particle paths do not coincide.
1. Lagrangian and Eulerian Description of Motion 7
0 1.10 a) Can the functions representing a motion law be found from the known
particle paths? b) Can a velocity field at a given instant be found from the streamlines
known for this instant?
0 1.11 Find and sketch the streamlines and particle paths if motions of a
medium are described by the velocity fields a) V\ = —uix2, v2 = u;xi, v3 =
u (UJ,U = const); b) Vi — — Ax2, v2 = Bxit v3 = 0 (A = const >
0, B = const > 0); c) V\ = —Vsinuit, v2 = Vcosuit, v3 = 0 (u>, V =
const).
© 1.12 Can the particles of a medium move with nonzero acceleration if a) the
velocities of all the particles are identical? b) the velocity at any point of the space
does not vary depending on the time?
0 1.13 The density of every individual particle of an incompressible medium is
constant. Can the density at any point of the space depend upon the time?
0 1.14 While a medium moves with the velocity field
Vi = —wx2 , v2 — CJXI , f3 = 0 (u> = const),
the temperature field
zl _ Kf£l) 2 _ K(—)2 _ {I—)
!
2
T = T0er a' b' c (T 0 ,r,a,6,c = const).
is created in the space (with the help of appropriately distributed sources). Find the
rate of change in temperature in an individual particle at an instant t0 if this particle
is situated at the point of the space with the coordinates Xi = a, x2 = b, 13 = c at
this instant.
0 1.15 A medium moves with the velocity field
vi = kx\ , t>2 = —kx2 , v3 = 0 (fc = const)
and the density field
p = po + Ax2ekt (po, A = const).
Find the rate of change in density in each of the particles of the medium.
0 1.16 The position of an individual particle (C1.C2.f3) at any instant t is given
by the relationships
*i = /i(Ci + W , & , 6 ) , t = 1,2,3, (U = const).
Show that a) the motion is steady, b) the parametric equations of the streamlines
have the form x{ = /j(r,C2.C3). * = 1.2,3, where T is a parameter, and the pair of
numbers C2. C° specifies a streamline.
8 BASIC CONCEPTS
Q 1.17 A medium moves so that the paths of all particles lie on rays originating
from a point O, and the speed v and density p depend only upon the instant t and
distance x to the point O. The mass contained, at the instant t = 0, within the sphere
with the center O passing through some particle is often used as one of Lagrangian
coordinates f of this particle to describe such a motion (referred to as spherically
symmetric). Show that the expression
X
f = f AirR2p(R,t)dR
is valid for the Lagrangian coordinate f of a particle at a distance a; from the point O
at an instant t. Show that the speed and density of the medium depends only upon
£ and t in the Lagrangian description. Find the equation for these functions v(£, t),
p~(£, t) (including also the function x(£, t)) with transforming the equation
dp dp dv pv _
dt dx dx x
expressing the mass conservation law in the Eulerian description.
0 1.18 The motion of a medium is represented by the functions
I =
xi=£u x2 = & (1 + - J , 3 ^J-T (r = const).
T
a) Find the velocity and acceleration fields, b) Find the velocity, at the instant t = IT,
of the particle situated at the point with the coordinates (a, a, a) at the instant t = r.
0 1.19 Find the velocity and acceleration fields in Lagrangian and Eulerian de
scriptions if the motion of the medium is represented by the functions
Xl = 6 + c(<)£2 , X2 = 6 + C(t)& , X3 = £3.
© 1.20 A medium moves with the velocity field
vi(x, t) = at , u 2 (i, t) = —«— , v3(x, t) = 0 (a,u = const)

Xl
and the temperature field
T = T0 (l + ^ j (To, r = const).
Find the rate of change in temperature at the instant t = r in the individual particle
situated at the point with the coordinates
Xi = if/a , i 2 = 2u 2 /a , x 3 = 3w2/a .
2. Tensors and their Cartesian Components 9
2 Tensors and their Cartesian Components

Expressions with indices. Quantities of the same type are often denoted by one
letter equipped with some number of indices. The Kronecker symbols <5y (<5y = 1 at
i = j , Sij =0 a.t i ^ j) is an example of a set of such quantities. Further, in absence
of special notes, different indices are implied to take independently each of the values
1, 2, 3.
To shorten the notation, the following summation convention is adopted: If in a
one-term expression composed of letters with indices some index appears twice, this
expression denotes the sum of the corresponding terms taken for every value of this
index. For example, <Zy bju denotes the sum ati 6i« +a i 2 hu +&i3 ksu- There may be
several such pairs of indices in a term; every pair denotes independent summation.
Tensors. For a pair of vectors a, b, their tensor product ab is introduced. Tensor

products can be added and multiplied by a number. A tensor product is linear with
respect to each of the cofactors:
(QIOI + a2a2)(/?i&i + fob2) = ai&ai&i + anfoafo + a 2 /?ia 2 6i + a202a2b2 .

All possible linear combinations of tensor products form a linear space, elements of
which are called tensors of second rank. The set of the tensor products e ^ can be
used as a basis in this space (e< is a basis of the original vector space). In particular,
if et is an orthonormal basis of the Euclidian space, a tensor of second rank t can
be written in the form t = Ujetej. The numerical coefficients ty are called the
components of the tensor t in this basis.
Let ej and e'j be two orthonormal bases, such that e* = ■Aye'j. Then t,j and t?u
are the components of a tensor of second rank t in these bases if and only if
Uj = AikAjit kt ■
This formula is referred to as the tensor transformation rule (in orthonormal bases).
Only orthonormal bases will be used further in this section.
Tensors of third, fourth, etc., ranks and their components are introduced with the
help of the tensor products abc, abed, etc. (similar to tensors of second rank). The
tensor products eêêt, eie.,efcef, etc., can be used as bases in the spaces of tensors of
third, fourth, etc., ranks. A vector is a tensor of first rank. A number (independent
of a chosen basis) is called a scalar and is a tensor of zeroth rank.
Multiplication by a number a is defined for any tensor t, the components of the
product a t being (at)ij„,m = aty...m where t„...m are the components of the tensor t.
Addition of two tensors a and b of the same rank is defined, the components of the
sum a + b being the sum of the components (a + b)y...m = Oy...m + 6y...m. Besides of
these operations, multiplication of any two tensors is determined. For example, the
10 BASIC CONCEPTS
tensor product A B of the tensors A = ^ e ^ and B = B W m e k eje m is the tensor of

rank 5
AB = AijBkimeiejekeiem .
For any tensor of rank not less than 2, the contraction with respect to a chosen
pair of indices is defined, the result also being a tensor. Components of the contracted
tensor are obtained from summation of the components of the original tensor with
equal values of indices in the chosen pair; the summation is performed for every
specified set of values of the rest indices. For example, the contraction of the tensor
Q = Qijkiî^j^kî with respect to the first and the third indices is the tensor
q = qêi, qjt = Qiju .
A contraction is often performed in a tensor product, e.g., the tensor with the
components Ca — AijBji is the result of the contraction in the tensor product with
the components AijBki. Contractions can be performed simultaneously with respect
to every of chosen pairs of indices, e.g., the tensor with the components Py = Aijkieu is
the result of the two contractions in the tensor product with the components >ly«em„.
For any second rank symmetric tensor t (ty = tji), there exists an orthonormal
basis e* in which the components t\2 = t2l, *i3 = t^x, t23 = t32 vanish. The straight
lines, along which the vectors of this basis are directed, are called the principal axes
of the tensor t. The components t*lt t22, t33 in this basis are called the principal
components or eigenvalues of the tensor t.
The eigenvalues and eigenvectors of a tensor t are defined as the numbers A and
vectors with components Vi satisfying the following system of equations
Up, = Xvi .
Therefore the eigenvalues can also be obtained from the criterion for existence of a
nonzero solution vt of this system det ||ty — A<5y|| = 0, i.e.,
A3-/1A2 + / 2 A - / 3 = 0
where
h = Ui , h = -^{Uitjj ~ UjUj) , I3 = det ||ty|| .
The numbers Ix, I2, h are independent of the orthonormal basis in which the com
ponents Uj are considered in these formulae (see Problem 2.15). They are called the
invariants of the tensor t. Any number-valued function of Uj possessing the same
property is also called an invariant of the tensor t. Any invariant of the tensor t can
be expressed in terms of the invariants I\, I2, h-
2. Tensors and their Cartesian Components 11
— PROBLEMS —
Expressions w i t h indices
0 2.1 Write out in detail the following expressions using only numerical values of
the indices (not their letter notations)
a) Uu
b) pijUj, Ujpih p^m, UiPij\
c) qijdibj, qijbjd,, b^di, a ^ f y , a.fyqy, fya^y, flyO,^;
d) Oijbij, cijibji, ciijbji, bijdji.
Indicate equal expressions.
0 2.2 Compute
a) sums 6it, 6^, tfyfybfa;
b) the same sums if all the indices run the values 1,2,..., n.
0 2.3 Write the formula for computation of the substantive derivative in Eulerian
description dA/dt in the short form using the summation convention.
Tensors
0 2.4 Let t{j be the components of a tensor in an orthonormal basis e*.
a) Show that the numbers Ty = tjt (for example, Tj2 = £21) are the components of a
certain tensor.
b) If Ui and Vj are the components of vectors, is TyUjiij = UjUiVjl Is UjUiUj = TyUjUj?
0 2.5 a) In some orthonormal basis, the components of a tensor of second rank
satisfy the relationship ty = tj{. Show that a similar relationship is valid for its com
ponents in any orthonormal basis. In this case, the tensor is referred to as symmetric.
b) Components of a tensor of second rank in some orthonormal basis satisfy the
relationship ty = —tji. Show that a similar relationship is valid for its components
in any orthonormal basis. In this case, the tensor is referred to as antisymmetric.
0 2.6 Consider the sums ayt + by* of the components of tensors a and b in every
orthonormal basis and show that they are the components of a tensor.
© 2.7 Consider the products flyju£mn of the components of tensors B and e build
in every orthonormal basis and show that they are the components of a tensor. Show
that their sums By«eti are the components of a tensor as well.
12 BASIC CONCEPTS
0 2.8 Let Sij and a« be the components of symmetric and antisymmetric tensors.
Show that Sijdij — 0 (the "complete contraction" of a symmetric and an antisymmet
ric tensors equals zero).
© 2.9 Show that any tensor of second rank can be represented as the sum of a
symmetric tensor and an antisymmetric one. Is this representation unique?
0 2.10 Show that for a symmetric tensors the contraction syUjVj can be expressed
in terms of contractions of the form SyiOji^ (ttj, t>j, Wi are the components of vectors).
In other words, the values of a symmetric bilinear form can be expressed in terms of
the values of the corresponding quadratic form.
© 2.11 Show that if a tensor of second rank t satisfies the relation UjViVj = 0 for
any vector v, it is antisymmetric.
0 2.12 The tensors t (5) and t (d) with the components
Uj = ^tkk6ij , ttj = tij - -tkk6ij
are called respectively the spherical component and the deviator of a symmetric tensor
t.
Find
a) the deviator of a spherical component (t w ) ( < i ) ;
b) the spherical component of a deviator (t^)'*'.
© 2.13 Find the general form of a tensor of second rank t if its component t12 is
known to equal zero in any orthonormal basis.
© 2.14 Find the principal components and principal axes of the tensor with the
following matrix of components in an orthogonal basis e^
/ 1 -v/3 0 \ / 1 -\/3 0 \
a) -y/3 -1 0 b) -N/3 -1 0
V 0 0 3y V 0 0 2/
© 2.15 Show that the following functions of the components ty of a second rank
symmetric tensor t are its invariants
a) J\ = ta, Jz = UjUj, J3 = Ujtjietid;
b) h = tu, h = \{tatii - 1^), I3 = det ||ty||.
© 2.16 Are the principal components of a second rank symmetric tensor its in
variants?
3. Curvilinear Coordinate Systems 13
0 2.17 Express the invariants Ii, I2, h, J\, J2, J3 (see Problem 2.15) in terms of
the principal components of the tensor t.
3 Curvilinear Coordinate Systems

Solving some problems calls for use of curvilinear coordinate systems although there
exists a Cartesian coordinate system in a Euclidean space. Normally, this depends
upon a symmetry of the problem. For example, to study an axially symmetric body
subjected to an axially symmetric load is more convenient in the cylindrical coordinate
system.
Coordinate systems, local bases. A coordinate system sets a correspondence

r <-> (x',x 2 ,x 3 ) between vectors and triples of numbers; such a triple (x',x 2 ,x 3 )
is also called a point with coordinates (xl,x2,x3). Numbers x 1 , x2 and x3 are the
components of a vector r only in special rectilinear coordinate systems. Note that we
use upper indices for coordinates in this section. If a coordinate system is given in the
space, then lines along which two coordinates are constant (coordinate lines), can be
drawn through each point. A coordinate system fixes a local basis of the Euclidean
space e< = dr/dxl corresponding to a point (x 1 ,! 2 ,! 3 ). In general, this basis is not
orthonormal. The basis e* satisfying the relationships ek ■ e< = 5k (a • 6 is the scalar
product of vectors a, b) is referred to as dual to ek. The dual basis exists, is unique,
and can be found by a standard procedure (Problem 3.1). If </y is a set of the scalar
products gjj = e-i ■ ej, and g'i is a set of the elements of the matrix \\gx*\\ reciprocal
to the matrix ||<7y||, the following relationships are valid
e* = g,kek, e, = gjkek, gv = ex ■ eP .
Vector and tensor fields. A vector-valued function on a domain (surface, curve)

is called a vector field on the domain (surface, curve) or, often, simply a vector. If a
coordinate system x' is chosen, a vector field is represented with use of local bases e*
or the dual bases e-7 in the form
v = vlei=Vjei, with vl = gtkvk, Vj = gjkvk. (3.1)
The quantities vl and Vj are called respectively the contravariant and covariant com
ponents of the vector field v in the coordinate system x\ Thus, we use upper indices
for contravariant components and lower ones for covariant components. In the case
of an orthogonal curvilinear coordinate system physical components are also often
used (see Problems 3.17, 3.19). If, besides coordinate system x', another coordinate
system x'k is used, its basis, dual basis and the components of a vector field v in it
14 BASIC CONCEPTS
(denoted with primes) are bound with the bases e^, e-* and the components of the
vector field v in the coordinate system x' by the relationships
dx dx
e'i = -en, v\ = ——-vk - the covariant transformation rule,
ox" ox' fi o\
dx" dx" ^ '
e" = 7T-re*i v" = TTT W * ~ the contravariant transformation rule.
ox" ox*
The following relationships are also valid
dx'k , dxlk ,
Ci =
-dxTek' Vi =
-dVVk>
dx' lk 4 dx*
d^C ' V=
dx*<
dx* dx'k ri da? dxk d

= } ,
O,- -r-r -z—7 = O,
0x'k dxi ' dxk dx" ]
'
A tensor-valued function on a domain (surface, curve) is called a tensor field on the
domain (surface, curve) or, often, simply a tensor. If a coordinate system x' is chosen,
a tensor field, e.g. of second-rank, is represented in the following forms
t = tij eê, = tkj e*e, = <', eê1 = tkt ekel .
The numerical coefficients tli, tju and t'J, t'j are called respectively the contravariant,
covariant and mixed components of the tensor in the coordinate system i*. The dots
in the notations tk^_, t*| are used to indicate the sequence of the indices. For example,
in the expression A1^, the index i is the first, and the index k is the second. The
components of a tensor in a coordinate system xik are bound with its components
in a coordinate system x* by the so-called tensor transformation rule: the covariant
transformation rule is used for every lower index, and the contravariant one is used
for every upper index, e.g.,
y. =d^_d_^_d^
■k 1
dxP dx» dx " r
'
For any two coordinate systems xl and x'k, tpqr and t?xjk are the components of a
tensor t in these coordinate systems if and only if they are bound by the tensor
transformation rule. The tensor of third-rank is considered here as an example; the
formulated statement is valid for a tensor of any rank.
Metric tensor. The quantities

gij = ei-ej, gi3=g3i=63, g,] = e' • e3
are the components of some tensor of second-rank g referred to as metric (see Prob
lem 3.4). The length squared of an arc element is determined by the relationship
ds2 = dr ■ dr = <fcj dx*dxk.

Contravariant, covariant and mixed components of a second-rank tensor are bound
by the following relationships
tl)=9lktkv tf' = ftt^, ty=sV*H.

Similar formulae are valid also for a tensor of any rank, e.g.,
1 l
...i —9 ki
an index is raised or lowered as if it is the only one, cf. Formulae 3.1.
Tensor operations. The product of a tensor t and a number a is the tensor with
components obtained by multiplying the components of a t by a, e.g.,
a t = a(i y ejej) = {at'ê^j = (af ê3 = ( c r f ^ e , - = (crfy)eV

A sum is defined for any two tensors a and b of rank r, and its components are
obtained by adding the components of a and b with the same location of indices. For
example, the sum of a = aê3 and b = bkiekel is
a + b = {an + bê'e3 = {aj + ^)e% = (o^ + Vê3 = {a13 + fiê,- .
A tensor product is defined for any two tensors. For example, the tensor product AB
of A = Aij&e3 and B = Bkimeke'em is a tensor of rank 5
AB = AijBktmeie3ekelem .
It can also be expressed in terms of the components with another location of indices,
e.g., AB = AijBklmeie3ekeiem. For any tensor of rank not less than two, a contrac
tion is defined with respect to a chosen pair of indices one of which must be upper,
and the other, lower. For example, the contraction of the tensor Q = Q^k\exejekel
with respect to the first and the third indices is the tensor
q = q!leiel, (f.\ = Qfi-

It makes no difference which of the two chosen indices is upper and which is lower,
since the equality Q^'lj = Q'3'it is valid (see Problem 3.25).
16 BASIC CONCEPTS
Covariant differentiation. The partial derivatives of a vector field v with respect

to coordinates xk can be expressed in terms of the components of a second rank tensor
Vv according to the formulae
0 = (vfcl/)ei = (vfc^V •
Here we use special notation for the components of the tensor Vw: (Vw)^' = V**;',
(Vu)jfcj = VfcVj. The quantities V*w*, Vfct); are called covariant derivatives of, respec
tively, contravariant and covariant components of a vector field t;. They are found
with the use of the formulae
where the quantities T)k are the coefficients in the decomposition of the derivatives
of the basis vectors into components in this basis
^ k - r'r e,
e
dx ~ " '
(leading to one more formula dej/dxk = -Tê')- The quantities I>fc are called the
Christoffel symbols. They can be expressed in terms of derivatives of the metric tensor
components
ik +
2^ [dx* dxk dx'
The partial derivatives of any tensor field t with respect to coordinates xk are ex
pressed in terms of the components of some tensor with rank one higher than t,
namely Vt. The components of Vt are called covariant derivatives of the corre
sponding components of the tensor t and found with the use of formulae similar to
(3.3),e.g., ift = «' tn n e,e m e n ,
J ^ = (V^' m n )e<e m e n ,
Fit1 n
V ,l n vv
-m- ■ p( is- n ps A n , p n A- »
m
' ~ ~8xk~ k mfc
'■ +l»klm- ■
k
In particular, if T is a scalar, VfcT = dT/dx are the components of the vector grad T:
gradT = VfcTe* = g e * .
— PROBLEMS —
Coordinate system
0 3.1 a) Show that the dual basis e* with respect to the given basis ei exists and
is unique, and that the following relationships are valid
e' = gtkek, d = gjkek
where g,k are the components of the matrix ||<7,J|| reciprocal to the matrix ||<7y||,
gij = ej • er b) Verify the relationship e' ■ e? = glk. c) Verify the relationships
u • v = u'vi = UiV1, v' = e' • v, Vi = e{- v .
d) Verify that a basis and its dual basis are expressed in terms of each other according
to the formulae
! e2 x e3 2 e3 x d 3 Jd x e2
e = —^—, *■ = — - — , e = —-—, V. = d • (e 2 x e 3 )
e2 x e3 e3 x e1 e1 x e2 , 2 ,
ei = — y — , e2 = — ^ ^ , e3 = - ^ ^ , K* = e 1 • (e 2 x e3) .
© 3.2 a) Find the basis dual to an orthonormal one. b) Is the basis dual to an
orthogonal basis of non-unit vectors orthogonal? c) A basis ej is formed by unit
vectors every two of which make the angle 7r/3. Find the basis dual to it. Find the
contravariant components of the vector v = ae1 + be2 + ce 3 .
0 3.3 Prove the statement formulated in Problem 2.5 in the case when the basis
ej is arbitrary.
© 3.4 For every coordinate system, the sets of numbers are considered: 1) g^ =
Ci • ej, 2) g1' - the components of the matrix ||<7y|| reciprocal to the matrix ||<7ij||.
Show that they are respectively covariant and contravariant components of one and
the same tensor (metric tensor). Find mixed components of this tensor.
0 3.5 Show that the following sets of relationships for components of a tensor are
equivalent a) Sy = Sji <=> skl = slk <=$■ smn = snm b) Oy = — a,* <=> a w =
-alk ■<=> a m n = — anm It means that any of the sets of relationships (a) or (b)
can be accepted as a definition of, respectively, a symmetric or antisymmetric second
rank tensor (see Problems 2.5 and 3.3).
0 3.6 Are the following equalities valid for a second rank tensor t? a) t{ • = £'•'
b) t\ = tkk
18 BASIC CONCEPTS
Figure 3.1
@ 3.7 Cylindrical coordinates x1 = r, x2 = ip, x3 = z (see Figure 3.1) are bound

with Cartesian ones x' 1 , x'2, x'3 by the relationships
;1 ;2 /3
x = r cos ip, x = r sm <p, x =z .
a) Find the basis of a cylindrical coordinate system at the points Mi (r = 5, <p =
0 ,z = 0) and Mi (r = 10, cp = 7 r / 6 , 2 = 1 ) (express in terms of the basis e\
of a Cartesian coordinate system x"). b) Find covariant, contravariant and mixed
components of the metric tensor in a cylindrical coordinate system, c) Find the dual
basis at Mi and M2.
0 3.8 Decompose the basis vectors e' 1 , e' 2 of plane Cartesian coordinates x' 1 , x' 2
over the basis of the polar coordinate system x 1 = r, x2 = (p bound with the Cartesian
coordinates by the relationships
/i a
x =rcos<p, x =rsmip.
0 3.9 A tensor of second rank p has the following components in a cylindrical

coordinate system: pn = a, p22 = b/r2, the other components equal zero. Find its
components in the Cartesian coordinate system (see Problem 3.7).
© 3.10 Spherical coordinates x1 = r, x2 = 0, x 3 = A (see Figure 3.2) are bound

with Cartesian ones x ' \ x' 2 , x' 3 by the relationships
x'1 = r sin # cos A, x' =rsin0sinA, x' ==7-cos0.
Find the basis of a spherical coordinate system (express in terms of the basis e\ of
a Cartesian system). Find covariant, contravariant and mixed components of the
metric tensor in a spherical coordinate system.
Figure 3.2
© 3.11 A vector field is given in a Cartesian coordinate system x":
x'le\ + x'2e'2 + x'3e'3

V
~ J(x>l)2 + {x>2)2 + (x'3)2'
Find its components in the spherical coordinate system (see Problem 3.10).
0 3.12 Find the length squared of the differential element dxle\ + dx2e.2 + dx3e$
in elliptic coordinates x1 = r, x2 = ip, x3 = z bound with Cartesian coordinates x, y, z
by the relationships
x = vV2 + a2 cos>p, y = rsmtp (r > 0).
© 3.13 Find the length squared of the element dx1ei + dx2e2+dx3e3 in ellipsoidal
coordinates x1 = r, x2 = 9, x3 = <p (with the coordinate surfaces r = const in the
shape of oblate ellipsoids of revolution) bound with Cartesian coordinates x, y, z by
the relationships
x = Vr2 + a2 sin 6 cos ip, y = Vr2 + a2 sin 6 simp, z = r cos 6.
© 3.14 Express in terms of the basis vectors e* or e' of a curvilinear coordinate

system x' a) a unit tangent vector to a coordinate line x 1 (i.e., to a curve x2 = const,
x 3 = const); b) an angle crossed at by coordinate lines x2 and x 3 in a given point; c)
a unit normal to a coordinate surface (x2x3) (i.e., to a surface x1 = const).
© 3.15 Sometimes, e.g., when studying a stream that flows about a body, it is
convenient to use a special curvilinear coordinate system associated with the surface
of the body. For a planar stream considered in its plane, such a coordinate system is
introduced as follows. Let, in the plane of the stream, the boundary of the body be
20 BASIC CONCEPTS
Figure 3.3
a smooth curve L given in the parametric representation r = f(s) = a(s)ex + b(s)ey

where s is the length of the arc of L, ex,ey is the basis of a Cartesian coordinate
system x, y on the plane. Then, in the vicinity of the curve, this coordinate system
associates the point r = f(s) + n(s)h with the pair of numbers (s, /i), where n(s) is
the unit normal to L, and h is the distance from L (see Figure 3.3). For the coordinate
system x 1 = s, x 2 = h, find the basis and the covariant, contravariant, and mixed
components of the metric tensor.
P h y s i c a l c o m p o n e n t s of vectors and t e n s o r s .
0 3.16 In mechanics, the length dimension is often assigned to Cartesian coor
dinates, and their basis vectors are regarded as dimensionless. In this case, for a
spherical coordinate system (see Problem 3.10), find the dimensions of a) the coordi
nates, b) the vectors of the basis and dual basis, c) the covariant and contravariant
components of the metric tensor, d) the covariant and contravariant components of a
velocity vector.
0 3.17 Components of a vector (having identical dimensions in a Cartesian co

ordinate system) may have different dimensions in a curvilinear coordinate system,
e.g., see the preceding problem. To avoid this, so-called physical components vPi of a
vector v
vP1 = t^leil, vP2 = v2\e2\, vP3 = t> 3 |e 3 |
are introduced in orthogonal curvilinear coordinate systems (e* J. e ; at i ^ j), and
the vector is expressed in the form
ei e2 e3
v = V U + w
PU—T + P21—T
V2 P 3V3
i—f •
Meil \e2\ \e3\
a) Show that
u P1 = t/i|e 1 |, vP2 = v2\e2\, t>P3 = t>3|e3|
and
ei e1 e2 e2 e3 e3
(This triple of vectors is called the physical basis associated with the considered
orthogonal coordinate system), b) Express a scalar product u ■ v in terms of physical
components of the vectors u and v. c) Similarly to physical components of a vector,
introduce physical components of a second rank tensor.
0 3.18 a) Express the physical basis er, ev, ez associated with a cylindrical coor
dinate system (see Problems 3.7, 3.17) in terms of bases e^ of the cylindrical and e'i of
the Cartesian coordinate systems, b) A body rotates around an axis with the angular
speed oj{t). Regarding the rotation axis as the coordinate line x3 of a cylindrical coor
dinate system, find the physical components of the velocity and acceleration vectors
of the body points in this system.
0 3.19 Show that the physical basis associated with a cylindrical coordinate
system x' (see Problem 3.18a) is not the local basis of any coordinate system
yk (x1, x2, x3).
Tensor transformation rule.

© 3.20 a) Show that the elements of the matrix ||6y|| of a bilinear form are the
covariant components of a tensor of second rank. (A bilinear form b is a scalar-
valued function of two vector arguments linear with respect to each of them. For
every coordinate system, a matrix ||6y || exists such that b(u, v) = 6yuV for any two
vectors u, v; this matrix is called the matrix of the bilinear form b.) b) A bilinear
form sets into correspondence to a pair of vectors their scalar product. Find the
matrix of this bilinear form.
0 3.21 a) Show that the elements of the matrix \\a'-\\ of a linear operator are
the mixed components of a tensor of second rank. (A linear operator a is a linear
vector-valued function of a vector argument. For every coordinate system a matrix
||a'j|| exists such that av = a'-v'ej for any vector v; the matrix Ha^H is called the
matrix of the linear operator a), b) A linear operator projects a vector on the plane
orthogonal to the given unit vector n = n'e^. Find the matrix of this operator.
© 3.22 Find the general form of a second-rank tensor t if its component ti 2 is

known to equal zero in any coordinate system.
22 BASIC CONCEPTS
Tensor operations
0 3.23 For two second rank tensor a and b, consider the sums a,J + 6y of their
components in every coordinate system. Show that they are not components of tensor
unless a = 0 or b = 0.
0 3.24 Find the contravariant components of the sum of the tensors a = eiei and
b = e 2 e 2 where e* is the basis of the coordinate system x*, x1 = x[ + x'2, x2 — x'2,
x3 = x'-, and x[ are Cartesian coordinates.
© 3.25 Prove the following equality for the components of a tensor Q

3ijkleiejeke'
=
Qij l Q jil ■
0 3.26 Consider the components of tensors B = Bijkie'eieke' and e = eê'e^

in every coordinate system and prove that a) sums Bijklekl are the components of a
tensor, b) the following equalities are valid
&iaeu = 5' J k 'e*, = B ^ , e t ' = Biiklekl. .
0 3.27 Consider the numbers tu, UjUji Ujtjktki determined by the components of
a second-rank tensor t in an arbitrary coordinate system. Do they depend upon this
system?
0 3.28 Consider the numbers J\ = trt, J2 = tyiy, J3 = Ujtjktki determined by

the components of a second-rank tensor t in a Cartesian coordinate system (they do
not depend upon this system, see Problems 2.15, 3.27). Find the formulae expressing
these numbers in terms of the mixed components of the tensor t in an arbitrary
coordinate system.
0 3.29 For a second-rank symmetric tensor t, the eigenvalues and eigenvectors

are defined as the numbers A and vectors with components v[ satisfying the system
of equations
<^ =H
in a Cartesian coordinate system, a) Derive the equations determining the numbers A
and components v[ in an arbitrary coordinate system in terms of (i) covariant, (ii) con
travariant and (iii) mixed components of the tensor t. b) Express the coefficients of
equation determining eigenvalues A
A3 - Â 2 + 72A - I3 = 0
in terms of the mixed components of tensor t in an arbitrary coordinate system.

Levi-Civita tensor. Calculations in curvilinear coordinate

systems
0 3.30 Show that the numbers eijk = e{ • (e^ x ek) are the components of a
third-rank tensor (called Levi-Civita tensor).
0 3.31 Verify that Levi-Civita tensor has the following properties: a) it is anti
symmetric with respect to every pair of indices, i.e., the relationships
are valid; b) only those its components do not equal zero whose values of the indices
are obtained by permutation of the values 1,2,3; the corresponding components equal
«i23 or — f 123 if the permutation is, respectively, even or odd, i.e., the relationships
£ = € €
£ 123 = 231 312, 213 = ^132 = «321 = "^123
are valid; c) the component £123 is expressed in terms of g = det ||py|| according to
the formula
{ ^g,
— y/g,
if ei, e 2 , e 3 - is a right-handed basis,
if e-i, e2, e 3 - is a left-handed basis.
0 3.32 a) Formulate the properties of contravariant components of Levi-Civita

tensor similar to those of covariant components from Problem 3.31a,b. Prove that
they are valid, b) Show that the component e123 is expressed in terms of g = det \\gij\\
according to the formula
123 _ / V\/5> if ei, e 2 , e 3 - is a right-handed basis,
c
~~ 1 - 1/y/g, ifei, e 2 , e 3 - is a left-handed basis.
© 3.33 Verify the validity of the following relationships for Levi-Civita tensor
ey*e„r = W +W + W 5 -
-ftffi - * # « * - £«}«$
eiikepqk = Sft-Sft
eMejkl = 26)
eiiketjk= 6
© 3.34 Show that a vector product a x b may be represented with the use of
Levi-Civita tensor in the form
a x b = e,jka,ibjek = epqTapbqeT
in any coordinate system. (This is one of the reasons for which Levi-Civita tensor is
convenient in calculations.)
24 BASIC CONCEPTS
Q 3.35 a) Show that the mixed product of vectors a, b, c may be represented with
the use of Levi-Civita tensor in the form
a (6 x c) = tljka'b'ck.
in any coordinate system, b) Express components of the double vector product a x
(6 x c) of vectors a, b, c in terms of the components of these vectors.
0 3.36 Show that the volume of the parallelepiped constructed on vectors a, 6, c
equals le.jja'fe'c*].
© 3.37 Consider a set of numbers ?y* defined for any right-handed basis: e 123 =
«23i = «3i2 = 1, «2i.i = «i32 = «32i = - 1 , the other ei]k = 0. They are called
Levi-Civita symbols and often used in calculations (the notation eV* is used for them
as well, a) Does a tensor exist, the components of which are P]k in every basis?
b) Express the determinant of a given matrix and the components of the matrix
reciprocal to it in terms of the components of this matrix using Levi-Civita symbols.
© 3.38 a) Derive the formula for the differential of the determinant of a matrix
d(det||a' ; ||)=det||aV||6Va' f c
where ||6*J| is the matrix reciprocal to the matrix ||aV||. b) Derive the formula for
the differential of the determinant g = det \\g,j\\
dg = gg'3dgtj .
© 3.39 a) Derive the formula for the differentials of the elements of the matrix
J]6* -1| reciprocal to a matrix \\a'A\
dbkt = - 0 : > a m n .
b) Derive the formula for the contravariant components of the metric tensor
dg'} = -Wdgu .
Covariant differentiation.
0 3.40 Do the derivatives dvl/dx3, where v' are the contravariant components of
a vector, represent the components of a tensor?
© 3.41 Are Christoffel symbols n t the components of a tensor?
© 3.42 Is the following statement valid: if the component vl of a vector field
vanishes in some coordinate system, then Vû1 = 0?
0 3.43 Find the covariant derivatives of components a) of the metric tensor, b)

of Levi-Civita tensor.
0 3.44 Find the Christoffel symbols a) for a cylindrical coordinate system (see
Problem 3.7), b) for a polar coordinate system on a plane (see Problem 3.8).
0 3.45 Find the Christoffel symbols for a spherical coordinate system (see Prob
lem 3.10).
© 3.46 Find the Christoffel symbols for the curvilinear coordinate system de
scribed in Problem 3.15.
0 3.47 Derive the formula T^ = diny/g/dx' where g = det ||<7y||.
0 3.48 The scalar field div v calculated by the formula
dlvv =
d7>
in a Cartesian coordinate system x" is called the divergence of the vector field v. Show
that the following formulae are valid in any coordinate system x': a) div v = V;v',
M V-y* = -i- a(v^.v')
0 3.49 The vector field curl v calculated by the formula
e i e 2 e3
curl v =
¥ ¥ ¥
in a right Cartesian coordinate system x" with the basis e'i is called the curl (rotor)
of the vector field v. Show that the formula
curl v = t,ikV{Vjek
is valid for all coordinate system.
© 3.50 Show that, for a symmetric tensor of second rank b, the formula
, = J_ djjgby) _ i dgu
j
' y/g dxi 2 dx>
is valid.
© 3.51 Show that the expression of ViVj - VjV{ (vi are the components of a
vector) in terms of the partial derivatives and Christoffel symbols, in fact, does not
contain Christoffel symbols.
26 BASIC CONCEPTS
© 3.52 Show that, for an antisymmetric second rank tensor LJ, a) the formula
Viu/fc = J= ■gp(y/gujtk) is valid, b) the expression e'jfcVjU>Jfc does not contain the
Christoffel symbols (similarly to the preceding problem).
® 3.53 Show that a vector field k = k^(x')ej(x'), by choice of a certain new

coordinate system x"(x'), can be reduced locally (in the vicinity of such a point at
which k does not vanish) to the form k = ei(x").
© 3.54 a) Consider a vector field v with the components
dxi dxk
in a curvilinear coordinate system xk (a and b are scalar fields). Show that the field v
is solenoidal, i.e. satisfying the condition div v = 0. b) Consider a solenoidal vector
field v, d i w = 0. Show that, in a neighborhood of a point where v does not vanish,
there exist two scalar fields a and b such that the formula
vi = y* da_ db_
dxi dxk
is valid for the components of v in any coordinate system xk. Thus, the previous
formula gives locally general solution v of the equation div v = 0.
4 Deformation. Deformation Rate. Vorticity.

Deformation. Strain tensors.
Deformation of a continuum is variation of distances between its particles. Deforma
tion causes a response of the continuum and, in particular, results in the appearance
of internal forces in it. So it is necessary to introduce quantitative deformation mea
sures. For example, longitudinal deformation of an extended rod can be characterized
by the relative elongation (/ - l0)/l0 where I and l0 are respectively the lengths of the
rod in the current state and in a state relative to which deformation is measured
(called the reference state). The absolute elongation I — l0 is not a reasonable defor
mation measure: rods with lengths IQ = 10 cm and lo = l m when extended by the
same elongation I — /0 = 1 mm are in different states. The choice of a reference state
is rather conventional but is also often naturally grounded; e.g., as the reference state
for an elastic rod, the state with zero internal forces is normally chosen. In what
follows we assume that the reference (undeformed) state, where, by definition, strain
equals zero, is realized at the initial instant t = 0. Let i i , 121X3 be a spatial coor
dinate system (Cartesian unless otherwise stipulated) with basis eit Xi(£,t) be the
functions describing the motion, and £ = (6,62.£3) be the Lagrangian coordinates
4. Deformation. Deformation Rate. Vorticity. 27
equal to the spatial coordinates of the particle in its position at the initial instant,
i.e., & = Xi(£,0). As a deformation measure the Green strain tensor
o o o 1 / dxi dxi
e-ea0eae0, ea0 - - ^ — — - 6a0
and Almansi strain tensor
,J
2 \ dxi dxj
are often used; here, €a(x, 0 are the Lagrangian coordinates of the particle situated at
the instant t at the point x. A material line element emanating from the particle f and
corresponding to the vector d£ = d£a ea is the collection of particles with Lagrangian
coordinates within the limits from (£1,62,£3) to (£1 +d£i,£2 + ^621 £3 + ^3) occupying
the infinitesimal segment. At an instant t, the position of a material line element is
determined by the position x(£, t) of its initial point and by the vector
dx~
dx = dxi Ci — -£— d£a e{
Ota
(see Figure 4.1). Each of the tensors e and e allows one to immediately express the
(£,+<*£,,£, + </£„ £, + </£,) (x^dx^x. + dx^x^dx,)
A
(^1.^2-Sj) ""
Y dx
(X„X2,X3)
the intial instant ( = 0 the current instant t
Figure 4.1
change in length squared ds2 of a material line element (one of them, in terms of d£,
the other, in terms of dx)
ds2 - dsl = 2 ea0 d£,a d^0 = 2e„ dxi dxj .
Here, dso is the length of the material line element at instant t = 0. The strain tensors
also allow one to find the relative elongation of any material line element, change in
angle of any two material line elements (see Problems 4.3-4.4) and relative change in
volume of a material solid element
dV - dV0 r ~i ~~o " V t 1
— - — - 2 = V I + 2 / i + 4 72 + 8 / 3 - l =
dV0 VI - 2/i + 4/a - 8/ 3
28 BASIC CONCEPTS
where I * and U, i = 1,2,3 are the invariants of the corresponding tensors defined
as h = £u, h = \{I\ - £ij£ij), h = det||ey|| (see Problem 2.15). The mechanical
meaning of components of Green strain tensor is illustrated by their relationship to
a) the relative elongations l\, I2, I3 reached at the current instant of the material line
elements directed, at instant t = 0, along the basis vectors ei, e.i, 63 respectively,
b) the angles ipap (a ^ /?) reached at the current instant made by these elements
(e.g., t/>23 is the angle between elements directed along e 2 and e 3 at the instant t = 0)
£<*0 = 2 K1 + U ( l + h) cos
4>a0 ~ <W]
(no summation over a, j3). In particular, for components with identical indices
(no summation over a). Similarly, components of Almansi strain tensor are de
termined by "reverse" characteristics of deformation: with the changes in length
"reached" at instant t = 0, relative to the length at instant t, of the material line
elements directed along the basis vectors es and angles made by these elements (see
Problem 4.17). Components of the strain tensors can be expressed in terms
of the displacement field in the Lagrangian description
u(Z,t) = {xa(Z,t)-£a)ea
or in Eulerian description
w(x,t) = (xi-Zi(x,t))ei
(where, certainly, u(f, t) = w(x(£,t),t)). Namely, the formulae
%o MIf t\) - I (^b. 4. {** +-u ^b. f*b\

^ ' " 2 \d& dZ" d? d&) '
dWk
e j[x( t) +\ - I (*tb.+ 4. î _ ^ * \
' ' ~2\dxj dx, dx{ dxj •
are valid. In the case when relative elongations and rotations of all the material line
elements are small, i.e., when all the derivatives dua/d£p ~ 6 <C 1 are small (or,
identically, all the derivatives dii)i/dxj ~ 6 -C 1 are small), Green and Almansi strain
tensor differ only by a value of order <52 from linearized strain tensors e'"', e(1'
I =|W + 0(S2) , e= eM+0(62),

l
%W -?(')„ o ?« _ (dUa 4. dU
A
-eye,e, , e ti - 2 ^ ^ + ^ J •
Green and Almansi strain tensors also differ by a value of order 62

°e(Z,t)=e(x(Z,tU) + 0(62),
°e®(Z,t)=e®(x(S,t),t) + 0(62).
Normally one neglects the difference between the two linearized tensors and considers
both as one linearized strain tensor. Depending on what is more convenient, its
components are calculated by either of the two sets of formulae
1 (dua dup\ 1 fdwi dwj\
+ +
2\d^ d^) ' 2[dx~ ~dx~) ■
This tensor is often called also the small strain tensor. The mechanical meanings of
its components are as follows: en is the relative elongation of a material line element
directed along e\ at instant t = 0); £22 and £33 have the similar meanings; £y when
i 7^ j is half of the decrease in angle made by material line elements directed along
ei and e^ at the instant t = 0.
Transformation of a small volume of a continuum

Consider a small volume of a continuum that contains a particle f and consists of all
material line elements emanating from f and corresponding to all possible infinitesimal
vectors d£. The positions of all these elements at instant t are determined by the
position of the particle £, i.e., by the point with coordinates Xj(£, t), and by the vectors
dx (see Figure 4.1) related to d£ by the linear transformation
Ox
dxi = ^-dtc
or, briefly,
dx = Fd£ , dxi = Fiad£a
where the notation dxi/d(,a = Fia is used. The linear transformation F is called
the distortion, its matrix F = \\Fia\\ is called the distortion matrix or deformation
gradient. The components of the strain tensors are obviously expressed in terms of
the deformation gradient
£
a0 = -z(FkaFk0 - 6ap) , £ij = -((5y — H^Hyj)
30 BASIC CONCEPTS
where \\Hyj\\ is the matrix reciprocal to F: F^H^ = <5y. According to the theorem
of polar decomposition, the matrix F can be written in the form
F = RU , Fia = RtpUpa
where R — \Rip\ is an orthogonal matrix, and U — \\Upa\\ is a symmetric positive-

definite matrix. The linear transformation d£ —> U • d£ determined by the matrix U
performs three extensions along the principal axes of the Green strain tensor, and the
linear transformation d£ —> R • d£ determined by the matrix R maps the principal
axes of the Green strain tensor to the principal axes of the Almansi strain tensor (see
Problem 4.19). So a transformation of a small volume of a continuum with center in a
particle £ can be represented as sequential performance of a) extensions along the three
mutually orthogonal directions — transformation U, b) rotation — transformation R
(see Figure 4.2); besides that, the translation turning the point £ into the point x(f, t)
must be performed. The transformation U has the meaning of pure deformation. It
the principal axes the principal axes

of Green strain tensor of Almansi strain tensor
Figure 4.2
is closely connected with the Green strain tensor (see Problem 4.19).
Deformation r a t e , vorticity, divergence of velocity.

Many continua respond not so much to the strains, i.e., to relative elongations of
material line elements, as to their rates. To describe them quantitatively, the strain
rate tensor is introduced. Components of the strain rate tensor are expressed in terms
of the velocity field v according to the formulae
_ 1 / dvi dvj\
The mechanical meanings of these components are as follows: en is the rate of relative
elongation of the material line element directed along e.\ at the current instant; e22
and 633 have the similar meanings; e^ at i ^ j is half of the rate of decrease in angle
made by the material fine elements directed along e* and e, at the current instant.
The rate of relative change in volume of a particle is equal to the first invariant of
the strain rate tensor which in turn is equal to divergence of the velocity vector
dv,
en = ^— = divw .
OXi
The distribution of velocity in a small volume of a continuum is expressed in terms

of the strain rate tensor and the vorticity vector
1 . 1 dvk
w = - c u r l « = - e 0 * — et
where e^* are the components of Levi-Civita tensor. Namely, if Vo is the velocity
of the particle situated at a point r 0 , the velocity of the particle situated at a point
r0 + p is expressed by the Cauchy-Helmholtz formula
v = v0 + — ei + uxp + O(p')
dpi
where 3> = (l/2)eyPiP J . According to this formula, rotation with the angular veloc
ity w equal to the vorticity vector is superimposed on the velocity connected with
deformation (the second term in the formula) and the velocity of translation v0.
Use of curvilinear coordinate systems

Solution of some problems is essentially simplified if appropriate curvilinear coor
dinates are used instead of Cartesian ones. For example, oscillation of a spherical
bubble in unbounded liquid is conveniently studied in the spherical coordinate sys
tem. Transition to a curvilinear coordinate system requires some modification of
formulae valid in a Cartesian coordinate system. This is connected with the fact that
the basis e{ = dr/dxi of a curvilinear coordinate system x\ in general, depends upon
the point x of the space. In particular, the basis ei(x[£, t)) at the point where a par
ticle f is situated at instant t, in general, differs from the basis ej(x(£,0)) = e;(£) at
the point where the particle was situated at instant t = 0 (here, as usual, the spatial
coordinates of a particle at the initial instant are used as its Lagrangian coordinates:
£ = x(£,0)). The displacement vector in the Lagrangian description is ordinarily
decomposed in the basis e^)
u((,t)=«»K,f)ea(C),
and, in the Eulerian description, in the basis e^x)
w(x,t) = wl(x,t)ei(x) ;
32 BASIC CONCEPTS
here, certainly, w(x(£, t),t) = u(£,t). To make notation briefer, the arguments of
functions are often omitted. Then, to avoid confusion, the notations e a (f) = e a and
e<(x) = e.i are introduced, in particular, u = ua e a, w = wxei. Then, the Green
strain tensor can be written in the form
o o ° a og o 1 / dx' dx* o \
e =ea0 e e" , ea0 = - ^——gij- ga0j
where e ^ is the basis dual to the basis e„ (see Section 3), 9a0 =ea- e0, and the
components of the metric tensor <7y = e< • e ; are calculated for the point x = x(£, t).
The Almansi strain tensor can be written in the form
<* 1/ d$a d^ c \
e = e«ee», £ij = - ^ fli , - — — 3a , j
where e m is the basis dual to the basis e„, #y = e^ • e,, and the components of the
metric tensor 9a0 =ea- e0 are calculated for the point £ = £(x, t). The expressions
of the components of the strain tensors in terms of displacement field remains the
same, but the partial derivatives are replaced by the covariant ones
o
where the covariant derivative V Q is calculated with use of the Christoffel symbols
o
r$ M (0 determined by the usual formulae (see Section 3) in terms of the components
o
of the metric tensor 9py{0- H t n e relative elongations and rotations of all material
line elements are small, the small strain tensor is used. It can be written in either of
the two forms
£< = aU0+ 0Ua e
" 2 {^ ^ ) ° ° *" '
£"> = - (ViWj + VjWi) e V .
Components of the strain rate tensor are expressed in terms of the velocity field by
the formula
In some cases, a special coordinate system, determined by the motion of the contin-
uum itself, is useful and referred to as concomitant. It, as any coordinate system, sets
correspondence between a point of the space and three numbers, namely, Lagrangian
coordinates £Q = £Q(x, t) of the particle situated at the point x at the current instant.
Thus, this system is differently determined for different instants: different triples of
numbers correspond to the same point of the space; rigorously speaking, they must
be denoted with (^m^m^m)- The coordinate lines £° at the instants ti ^ ti occupy
different positions, but pass through the same particles; this is why this coordinate
system is referred to as concomitant (with the motion of the continuum). Its basis
dr dr dxx dxx
In the concomitant coordinate system, the Almansi strain tensor can be written in
the form
e = iij&e3 , itj = - ^ - <?i;J
where e m is the basis dual to the basis e n , and #y = e, ■ e,- are the components of
the metric tensor. As in any coordinate system, the components e^ are expressed in
terms of the displacement field by the formula
i^ = - (ViWj + V > , - Vi«>*VjU>fc) .
The components of the Almansi tensor in the basis e* coincide with the components
of the Green tensor in the basis e a , ia0 =£ap-
Conditions of compatibility
In some important cases, the velocity field v of a flow in the continuum is determined
by one function <p and can be written in the form v = grad^x In this case, the vector
field v is referred to as potential, and the function <p is called its potential. By far not all
vector fields v are potential. It is clear that, to be potential, a vector field must satisfy
some conditions, since the three functions — its components — are expressed in terms
of one function — potential. The necessary condition for potentiality of a field v is
the relationship curl v = 0; it is called also the compatibility condition for components
of a potential field v. If the field w is considered in a simply connected domain, this
is the sufficient condition for existence of a single-valued potential. Similarly to the
relationship v = grad tp, the six components of a small strain tensor are expressed in
terms of the three components of the displacement w, and, consequently, cannot be
arbitrary. They satisfy the relationships
d2ekl | 6% d2eu ^ _Q
dxidxj dxkdxi dxÔxj dxidxi
(xi is a spatial Cartesian coordinate system); these relationships also are called the
compatibility conditions for strain components. These are the necessary conditions,
34 BASIC CONCEPTS
and, when the tensor field of small strains e is considered in a simply connected
domain, are the sufficient conditions for the relationships
_ 1 Idnii dwA
£iJ
~ 2 \dx~j + ~dx~j
to be valid, w being a vector field. In other words, the compatibility conditions for
strain components are the conditions of possibility, in principle, to obtain the strain as
a result of a displacement. In all the problems of this section, i* are spatial (Eulerian)
coordinates, and £a are Lagrangian coordinates. As Lagrangian coordinates of a
particle, spatial coordinates of the point, at which the particle was situated at the
initial instant (in the undeformed state), are taken. Unless otherwise stipulated, the
spatial coordinate system is Cartesian.
— PROBLEMS —
Deformation. Strain tensors.

0I
0 4.1 As a result of displacement, the particles (£1*62.6) a continuum are
situated at the points with the coordinates
zi = & + a£i , x2 = & i x3 = £3 (a = const)
with reference to a spatial Cartesian coordinate system ii. (This deformation is called
a uniform uniaxial extension along the axis ij.) What happened to the material line
elements initially situated parallel to the coordinate axis xt and those orthogonal to
this axis? Consider the two cases: a > 0 and —1 < a < 0?
0 4.2 For a uniaxial extension (see Problem 4.1), find the displacement field in
Lagrangian and Eulerian descriptions, and calculate the components of the Green
and Almansi strain tensors.
© 4.3 a) A material line element emanating from a particle £ corresponds to a

vector d£. Regarding the components eap of the Green strain tensor in this particle
as known, find the relative elongation of the material line element by deformation.
b) For a uniaxial extension (see Problem 4.1), find the relative elongations of material
line elements perpendicular to the axis x 3 and making the angles ±7r/4 with the axis
i i before deformation.
© 4.4 a) Two material line elements emanating from a particle £ correspond to

vectors d£(1) and d£ (2) . Regarding the components I a0 of the Green strain tensor
in this particle as known, find the angle made by these material line elements after
deformation, b) For a uniaxial extension (see Problem 4.1), find the angle made after
deformation by the material line elements perpendicular to the axis x3 and making
the angles ±7r/4 with the axis i j before deformation.
0 4.5 Find relative change in volume caused by a uniaxial extension (see Prob
lem 4.1).
0 4.6 As a result of displacement, the particles (£1,62, £3) of a continuum are
xi = & > *2 = - ( 1 + b)£i , x3 = £, (b = const > - 1 )
with reference to a spatial Cartesian coordinate system Xj. a) What happened as

a result of deformation to material line elements initially situated parallel to the
coordinate axes? b) Find the Green and Almansi strain tensors, c) Can the Green
and Almansi tensors be regarded as coincident at |6| -C 1? Compare with the answer
to Problem 4.2 at a «C 1.
OI a
© 4.7 As a result of displacement, the particles (&J&.&) continuum are
*\ = £1 + a sin(fc^) , x2 = £2 , ^3 = 6
(Q = const, \a\ < 1, k = const)
with reference to a spatial Cartesian coordinate system x<. Show that a uniaxial
extension takes place in a small vicinity of each particle of the medium (see Prob
lem 4.1). What is the relative elongation of the material line element emanating from
a given point f and parallel to the axis Xi before deformation? Calculate the Green
deformation tensor. Indicate the particles in a small vicinity of which there is no
deformation.
© 4.8 As a result of displacement, the particles (£1,62. £3) of a continuum are
i , = { j + ( ^ j , (i = 1,2,3) (a = const > - 1 )
with reference to a spatial Cartesian coordinate system Xj. Show that the relative
elongations of all material line elements are identical. Note that such deformation is
called a uniform extension or compression to what signs of a do they correspond?.
0 4.9 A simple shear is deformation of a continuum described by the formulae
xi = f 1 + a(t)£2 , x2 = 6 , x3 = £3
36 BASIC CONCEPTS
where Xj is a spatial Cartesian coordinate system, fQ are Lagrangian coordinates, and

a(t) is a function of time with a(0) = 0. Regarding the function a(t) as given, find
the Green and Almansi strain tensors. Find their principal components and principal
axes. Simplify the formulae for the case \a(t)\ <C 1.
0 4.10 Find the components of the displacement field for a simple shear (see
Problem 4.9) in Lagrangian and Eulerian descriptions. Determine the components of
the Green and Almansi strain tensors, expressing them in terms of the derivatives of
the displacement field. Find the small strain tensor.
0 4.11 For a simple shear (see Problem 4.9), find a) relative elongations of mate
rial line elements emanating from particle f and parallel to the axes x\, x2,13 before
deformation; b) all possible material line elements whose relative elongations at the
instant t equal zero.
0 4.12 Find the relative change in volume of a small material element during a
simple shear (see Problem 4.9) by two methods: with the use of the invariants of the
Green strain tensor and those of the Almansi strain tensor.
0 4.13 The small strain tensor at a point where small deformation takes place has
the following matrix of components in a Cartesian coordinate system
0.01 0.03 0 \
0.03 0.01 0
0 0 0.01 /
Find the maximal and minimal relative elongations of material line elements at this
point. Find the direction of the material elements whose relative elongation is a) max
imal, b) minimal. Calculate the relative change in volume at this point.
0 4.14 A double shear is deformation of a continuum described by the formulae
xi = £1 + b(t)£2 , z a = & + 6(t)6 , 13 = 6

where Xi are spatial Cartesian coordinates, £a are Lagrangian coordinates, and b(t) is
a function of time with 6(0) = 0. Treating the function b(t) as given, find the Green
and Almansi strain tensors.
0 4.15 Find the components of the displacement field for a double shear (see
Problem 4.14) in the Eulerian description. Find the small strain tensor.
0 4.16 The positions of three material line elements in a deformed state are char
acterized by the vectors dx® = dse{, i = 1,2,3 (ej are the vectors of the orthog
onal basis of a Cartesian coordinate system). Their "inverse" relative elongation
dsg /ds — 1 (where dsg is the length of the corresponding element before defor
mation) equal /*. The elements characterized by the vectors dx^ and dx^ in the
deformed state make the angle Vy before deformation. Prove the formula showing
the mechanical meaning of the components of the Almansi strain tensor
£y = 2 H i + fc)(l + h) cos^3 + 6n}
(no summation over i and j).

o
© 4.17 Prove that the eigenvalues of a Green strain tensor Aa and the eigenvalues
of an Almansi strain tensor Aj satisfy the inequalities
l+2Ao>0, l-2Aj>0.
© 4.18 Prove that for deformation with the Green strain tensor e and the Almansi
strain tensor e a) the material line element directed along a principal axis of the tensor
e before deformation is directed along a principal axis of the tensor e in the deformed
state, b) on the contrary, the material line element directed along a principal axis of
the tensor e in the deformed state was directed along a principal axis of the tensor
o °
e before deformation, c) the eigenvalues of the Green strain tensor A and of the
Almansi strain tensor A are related to each of them by
1
1 + 2A =
1-2A
© 4.19 Prove the following polar decomposition theorem. The deformation gra
dient F = \\Fia\\ (and, in general, any nonsingular matrix) can be written in the
form
F = RU , F{0 = RiaUafj
where R = \\Ria\\ ls a n orthogonal matrix, and U = \\Ua0\\ is a symmetric positive-
definite matrix. The principal axes of the matrix U coincide with the principal axes
of the Green strain tensor, and the corresponding eigenvalues ka of the matrix U are
o J o
expressed in terms of the eigenvalues AQ of the Green tensor: ka = \ \ +2 \ a . The

linear transformation determined by the matrix R rotates the principal axes of the
Green tensor into the principal axes of the Almansi tensor.
0 4.20 For a uniaxial extension (see Problem 4.1) at a > 0, find the initial posi
tions of the three mutually orthogonal material line elements which remain mutually
orthogonal after deformation. Find the directions of these elements after deforma
tion. Indicate the directions of the material line elements whose relative elongation
is maximal.
38 BASIC CONCEPTS
0 4.21 Show that the following motions cannot be reduced to each other by
imposing a rigid motion (i.e., a rotation and a translation): a) a rotation around the
axis X3, b) a uniaxial extension along the axis X\ (Problem 4.1), c) a simple shear in
the plane xi,x 2 (Problem 4.9), d) a double shear (Problem 4.14).
© 4.22 Represent the transformation of a small vicinity of a particle $ resulted
from a simple shear (see Problem 4.9) in the form of the extensions along three
mutually orthogonal directions, rotation and translation.
0 4.23 For a uniaxial extension (Problem 4.1) with the parameter a as a function
of time a = a(t) such that a(0) = 0, a) write the formula describing transition, for an
instant t, from a spatial coordinate system Xj to the concomitant coordinate system
£Q, b) draw the coordinate lines, find the basis vectors and the components of the
metric tensor for the concomitant coordinate system £a for an instant t, c) find the
components of the Green tensor for an instant t in the coordinate system x* as well
as the covariant, mixed and contravariant components of the Almansi strain tensor
for an instant t in the concomitant coordinate system £°.
0 4.24 For a uniaxial extension (Problems 4.1, 4.23), choose the coordinates Xj
of the space point at which a particle is situated at the instant t, as new Lagrangian
coordinates rf of this particle. Draw the coordinate lines, find the basis vectors and
the components of the metric tensor in the concomitant coordinate system rf at the
instant t = 0.
© 4.25 For a simple shear (Problem 4.9), find, for an instant t the coordinate
lines of the concomitant coordinate system passing through the space point with
the coordinates (0,0,0). Do they vary in time? Find also the basis vectors of the
concomitant coordinate system and the components of the metric tensor in it.
© 4.26 The axis of a cylindrical rod with circular cross section is positioned along
the axis x 3 of a spatial Cartesian coordinate system x{. The rod deforms according
to the formula
xi = fi - a(*)6£$ , x2 = & + «(*)£ 16 , x3 = £3

where £a are the Lagrangian coordinates, and a(t) is a function of time with a(0) = 0.
Regarding the function a(t) as given, a) find the position at an instant t of the
particles forming, at the instant ( = 0, a cross section of the rod, the circumference
bounding this cross section, its radius as well as a segment parallel to the axis of the
rod and lying on its surface; b) find the displacement field in Eulerian description;
c) if |Q| < < 1, find the small strain tensor, the value of maximal relative elongation
for the material line elements emanating from the point x and the direction of the
element experiencing the maximal elongation, d) write the formulae describing the
motion in the cylindrical coordinate system, taking the (cylindrical) coordinates of a

particle at the instant t = 0 as the Lagrangian coordinates.
0 4.27 A tube (thick-walled circular cylinder) expands under action of internal
pressure. Its deformation is described by the formulae
r = r0 + f{r0, t) , <p = <p0 , z = z0
where r, ip, z are the spatial cylindrical coordinate system (see Problem 3.7), ro, <p>o, ZQ
are the Lagrangian coordinates (the cylindrical coordinates of the initial position of
the particle), and /(r 0 ,0) = 0. Let the function f{r0,t) be given. Find a) the Green
strain tensor, b) the Almansi strain tensor, c) the relative elongations of the material
line elements emanating from a particle (r0, ipo, ZQ) directed along the coordinate lines
of the cylindrical coordinate system before deformation.
0 4.28 The components of Green and Almansi strain tensors can be of order 1
in the cases when the components of the linearized strain tensor are small or even
equal zero. Verify this, using as an example the displacement field Wi(x) = x 2 ,
w2(x) = -xi, w3 = 0 where x = (11,12,12) are spatial Cartesian coordinates.
0 4.29 The components of the linearized strain tensor can be of order 1 in the
case when the components of Green and Almansi strain tensors are small or even
equal zero. Verify this, using as an example the displacement field x^ = flyf j where
Xi is a spatial Cartesian coordinate system, £Q are the Lagrangian coordinates, ||fly||
is an orthogonal matrix.
Strain rate tensor, vorticity, divergence of velocity.

0 4.30 Find the velocity field and components of the strain rate tensor in Eulerian
description for a) a uniaxial extension (Problem 4.1), treating the parameter a as a
given function of time t, b) a simple shear (Problem 4.9) and c) a double shear
(Problem 4.14).
0 4.31 Calculate the components e^ of the strain rate tensor in a spatial Cartesian
coordinate system x, and the components of its deviator 4 , ' = e^ - |efcfc<5y for the
flows of a medium with the velocity fields described, in this coordinate system, by
the formulae a) vi = Axu v2 = Bx2, v3 = 0 {A = const, B = const); b) Vi = ottxu
v2 = 1,3 = 0 (Q = const); c) Vi = (3tx3, v2 = v3 = 0 (0 = const).
© 4.32 Calculate the components e,; of the strain rate tensor in a spatial Carte
sian coordinate system Xj for the flow of a medium with the velocity field described,
in this coordinate system, by the formulae
40 BASIC CONCEPTS
Does a change in volume of material solid elements take place during this motion?
0 4.33 The components of the velocity field of a medium in a spatial Cartesian

coordinate system Xi at a given instant have the form
V\ = kx2 , v2 = v3 = 0 (k = const) .
Find, for this instant, the rate of change in angle made by material line elements
emanating from the point x and situated along two straight lines making the angles
7r/4 with the axis Xi and n/2 with the axis 2:3.
0 4.34 The covariant components of a velocity field in a spatial cylindrical coor

dinate system i 1 = r, x2 = tp, x3 = z have the form
V\ = 0 , v2 = k , v3 = 0 (k = const)
everywhere except the point r = 0. a) Draw the particle paths of the medium, find
the value of the velocity of a particle and the physical components of the velocity.
b) Calculate the components of the strain rate tensor, c) Find the vorticity vector.
d) Find the principal axes of the strain rate tensor. Do they rotate in time in an
individual particle? e) What is, at an instant, the angular velocity of the material
line elements which are situated at this instant along the principal axes of the strain
rate tensor?
0 4.35 A medium experiences a uniaxial extension, i.e., its motion is described

by the formulae
xi = 6 + a(*)f 1 , *2 = 6 , x3 = £,
where Xi are spatial Cartesian coordinates, f0 are the Lagrangian coordinates, a(t) is
a function of time with o(0) = 0. Verify that the vorticity vector field is zero during
this motion. Show that there are material line elements which rotate (simultaneously
varying in length).
© 4.36 Find the vorticity vector field tt> for a simple shear (Problem 4.9). Indicate
the material line elements angular velocity of which equals u; at an instant t. Find
the angular velocity of the material line elements directed at the considered instant t
along the axes X\, x2, X3.
0 4.37 The velocity distribution in a rigid body is determined by the Euler formula
v = v0 + fl x r where Q{t) is the angular velocity, r is the radius-vector with respect
to a point 0, v0(t) is the velocity of the point O. Calculate the vorticity vector for
this velocity field.
© 4.38 Prove that, if the strain rate tensor is identical for all particles of a medium
at an instant, then the vorticity vector is also identical for all the particles at this
instant.
0 4.39 The strain rate tensor equals zero in all particles of a medium. Show
that, in this case, the velocity field is described by the Euler formula for the velocity
distribution in a rigid body v = v0 + fi x r where r is the radius-vector relative to
a point O, Vo{t) is the velocity of this point, and f2(t) is a vector independent of r
(the vector of instantaneous angular velocity).
0 4.40 Consider the second rank antisymmetric tensor wye'eJ determined by a

vector field v
a) Show that in any curvilinear coordinate system x', the following equality holds
1 (dvj b\ii
UJn = '
2 ydx* dxi
b) Show that the vorticity vector w = 1/2 curl v can be represented by the above-
introduced tensor, since the following formulae are valid
0 4.41 For a motion of a continuum with acceleration a satisfying the condition

curia = 0, consider the components of vorticity vector OP with respect to a con
comitant coordinate system. Consider also the components ujap of the second-rank
antisymmetric tensor representing the vorticity vector as indicated in the preceding
problem. Show that the following formulae are valid
a ) ^ ^ = 0, b)!(wV£) = 0
where g = det \\ga0\\
© 4.42 For a given in the Problem 4.41 motion of a continuum, consider at any
moment a vorticity line, that is, a line whose tangent is a vortex vector (at every
point of the line). Show that a vorticity line is frozen into the medium, that is it
passes through the same particles of the medium at any moment.
42 BASIC CONCEPTS
Compatibility conditions
© 4.43 Verify that the equality curlv = 0 is the necessary condition for poten
tiality of the vector field v.
0 4.44 Verify that the field velocity for a uniaxial extension (Problem 4.1) satisfies
the condition for potentiality. Find the potential of this velocity field.
0 4.45 Show that the compatibility condition for the small strain tensor can be
written, using Levi-Civita tensor, in the form a) ekip€ijq8z ^ = 0 in a Cartesian
coordinate system, b) ckipeljqVkS7ieij = 0 in any coordinate system. Here, e,jk are the
components of Levi-Civita tensor.
0 4.46 Indicate the conditions which a symmetric tensor field must satisfy to be
the strain rate tensor for a velocity field.
0 4.47 The tensor field is given by the formulae for its components in a Cartesian
coordinate system xi,x2, £3 &) £11 = Ax\, £22 = Bx\, £33 = Cx\, en = £13 = £23 = 0,
A,B,C = const; b) en = 2AxiX2, £22 = 2BxiX2, £12 = \{Ax\ + Bx%), £ i3 = e23 =
£33 = 0, A, B = const Is it the small strain tensor for a displacement field?
© 4.48 The components /y of a second-rank tensor field are given in a Cartesian

coordinate system xk. Consider the possibility of finding a vector field v for which
the relationships dvi/dxj — /y hold. Prove that this is possible if and only if the
components /y satisfy the following conditions
dxk
(eikj are the components of Levi-Civita tensor). Formulate this statement in a curvi
linear coordinate system.
© 4.49 The components ey and tJj of a second rank symmetric tensor field and
a vector field satisfying the condition divw = 0 are given in a Cartesian coordinate
system. Consider possibility to find the velocity field v for which e and w are the
strain rate and vorticity fields. Prove that this is possible if and only if the components
ey and c<y satisfy the following conditions
3ey du>i
3
dxk dxi
(eikj are the components of Levi-Civita tensor). Formulate this statement in a curvi
linear coordinate system.
5. Principles of Symmetry and Tensor Functions. 43
5 Principles of Symmetry and Tensor Functions.

Let us consider algebraic properties of tensor characteristics of a medium at a given
point of the three-dimensional Euclidian space with local basis et. Let T ] , . . . , TN be
a set of tensors among which there is the metric tensor g. The symmetry group G
of this set of tensors is the set of orthogonal transformations of the basis e\ = aê-j
preserving the values of the components of each of these tensors. For example, if for
the contravariant components of a tensor of a rank r,
or
bl,jl...birjrTjl "•»'- = Til-ir (5.1)
l
where (b j) is the matrix inverse to the matrix (a'j), then it is said that the tensor T
is invariant relative to the group G, determined by (a'j). The symmetry group of the
tensor g itself (isotropy) is the complete group of rotations and reflections represented
by orthogonal matrices. The symmetry group of a given set of tensors containing g is
a subgroup of the complete group of rotations and reflections or coincides with that
group.
A tensor function T = F ( T i , . . . , T N ) is a dependence of the components of the
tensor T upon the components of the tensors T 1 , . . . > T W which is invariant with
respect to choice of the basis ej. It means that, for a set of functions of the form
the relationships
JV»I—«r _ pil...ir(rHJl—JTl rjtjl—JTN\
are valid for any nonsingular matrix (aV,), i.e. in all coordinate systems the com
ponents of T are the same functions of the components of Ti, T 2 , ..., TN. This
condition can be written in the form
F ' " ( ^ . ..bfr^ 7 * ■"*",... ,fc* tl . . . ^ „ T*1 -*"') = 6 \ , . . . 6 ^ 7 * ' - * -

(5.2)
The equality (5.2) is a very strong restriction on the form of a function F. By use of
it one can prove, e.g., that
1. if a vector a is a function of a vector 6 only then
a = kb
where k is a scalar which may depend on |6| (in fact, the components of a metric
tensor g are needed to calculate |6|; so to be exact we should say that a is a function
of 6 and g).
44 BASIC CONCEPTS
2. if a second-rank tensor H is a function of a second-rank tensor T only (besides

of a metric tensor g which stands for a unit and is used to construct scalar invariants)
then
H = fcog + JbiT + fc2T2
where T 2 = T • T and fco, fci, £2 are scalar functions of invariants of the tensor T;
3. if a symmetric second-rank tensor H is a function of two independent symmetric
second-rank tensors T and P then
H = jfcog + fciT + k2P + A;3T2 + fc4(T • P + P • T) + k5(T2 • P + P • T 2 )
where ko,...k5 are scalar functions of invariants of the tensors T and P .

It follows from the definition of a tensor function that the symmetry group of
fixed values of the arguments of a tensor function is also the symmetry group of its
corresponding value. This property allows the determination of the general forms of
tensor functions with accuracy up to scalar coefficients without direct using of the
relation (5.2). The problem of determining the general form of a tensor function
T = F ( T i , . . . , TN) (let the rank of the tensor T be r) can be solved in the following
way: a) fix the arguments of the function F in the case of general situation, and,
with transformation of the basis e*, reduce them to the simplest possible form (e.g.,
two components of one of the vectors can be made zero, or, a second-rank symmetric
tensor can be reduced to the diagonal form); b) determine the symmetry group G of
the arguments; c) with the use of the relations (5.1) determine all r-rank tensors that
are invariant relative to G, choose among them a set of linearly independent ones
(the basis tensors), and express them in terms of the tensors T j , . . . ,TN; d) T is a
linear combination of the basis tensors with coefficients that are scalar functions of
invariants of tensors T i , . . . , TN.
Analytic tensor functions T = F(S), where T and S are second-rank tensors, are
defined by the power series of the form
T = a 0 g + aiS + a 2 S 2 + ...
which is equivalent to the series for the tensor components

Tj = FJ(S?) = oo^. + a,S3 + a2S£SJ= + ... .
Application of the Hamilton-Cayley theorem stating that any square matrix is a root
of its own characteristic polynomial, i.e.,
S 3 - /, S 2 + / 2 S - J3g = 0
obviates the use of powers of the matrix (Sp of orders more than two.
Let us extend the definition of a symmetry group of tensors at a point up to the
definition of a symmetry group of tensor fields. Let T i ( i ' ) , . . . , TN(xx), be a set of
tensor fields among which there is the metric tensor field g. The symmetry group
G of this set of tensor fields is the set of transformations of Cartesian coordinate
systems (x') preserving the form of each of these tensor fields. For example, for the
contravariant components of a tensor field of a rank r regarded as functions of the
variables x \
T'" - ( x " ) e : i - - - < = r > ^ ( x " ) e : i • • • < ,
or
bû ... V'jr T>""jV(o'i(irt - <*)) = r > • ''(x*)
on the transformations xH = Wjx'+d where (a^), {b>k) are orthogonal matrices each
of which is the inverse of the other. The tensor field is said to be invariant relative
to the group G. The symmetry group of the metric tensor field g is the complete
group of motions of the Euclidian space. The tensor field invariant relative to the
group of translations of the coordinate origin x" = x' + c* is referred to as uniform.
Let us indicate the method of construction of tensor fields invariant with respect to
a given group G of the transformations x'* = bljXj preserving the position of the
coordinate origin O. That method is based on the definition of the tensor function.
Let G be a symmetry group of a set of tensors Tu..., TN given at a point O, and
r be the radius-vector of an arbitrary point P relative to O. Compose, for the point
0, the tensor function T = F(r, Tu ..., T„) whose components satisfy obviously all
the necessary conditions of symmetry as functions of the components of the radius-
vector r. The remaining step is to translate (without variation of the components)
the tensor T to the point P.
— PROBLEMS —
0 5.1 Find the eigenvalues of the matrix of rotation by the angle <p around the
axis x
© 5.2 Determine the general form of a matrix of rotation around a unit vector n
by an angle ip.
Q 5.3 Show that any orthogonal three-by-three matrix has at least one eigenvalue
equal to 1 or - 1 . To what transformation does the matrix
correspond ?
46 BASIC CONCEPTS
© 5.4 Show that an arbitrary orthogonal matrix can be represented either as a ma
trix of rotation around an axis (Problem 5.2) or as the product of a matrix of rotation
and a matrix of reflection in the plane perpendicular to the same axis (Problem 5.3)
0 5.5 Find the eigenvalues of an antisymmetric second rank tensor. Compare with
the eigenvalues of the orthogonal matrix of Problem 5.1.
© 5.6 Using the expansions of the functions of one variable F(x) in power series
over x, where 1) F(x) = ex, 2) F(x) = ln(l + 1 ) , 3) F(x) = (1 - x)" 1 , define the
corresponding analytic tensor functions F(S) replacing x by S.
© 5.7 Determine the relationship between the eigenvalues and eigenvectors of a

tensor S and those of T = F(S) if F(S) is an analytic tensor function.
0 5.8 Show that 7 3 (e s ) = e/l<s> where / 3 (T) = det HT^H, 7,(T) = T \ .
© 5.9 Show that, if S is an antisymmetric matrix (Sijgik = -Sikgij), then e s is

orthogonal. Find such a matrix S so that the matrix e s is the matrix of rotation of
Problem 5.2.
© 5.10 Show that, if a tensor S is symmetric then any analytic tensor function of
S is also symmetric.
© 5.11 Let all the eigenvalues A< (i = 1,2,3) of a tensor S be real and different.
Show that, for any analytic tensor function F(S), the formula (called the Lagrange
interpolation polynomial)
(S - A3g) ■ (S - A lg ) (S - A,g) • (S - A2g)

W +
+ (A 2 -A 3 )(A 2 -Ax) (Aa-A.XAs-Aa) F ( A a )
is valid; here, g is the metric tensor.
© 5.12 Analyze, for the Problem 5.11, the cases of repeated eigenvalues of the
tensor S if there are three linearly independent eigenvectors.
© 5.13 Determine the symmetry group of the set of tensors (g, e) where g is the
metric tensor, e is Levi-Civita tensor.
© 5.14 Determine, for an orthonormal basis e<, the symmetry groups of the fol
lowing sets of tensors: a) g, e 3 ; b) g, c • e 3 ; c) g, e 3 e 3 (transverse isotropy with the
axes of symmetry parallel to e 3 ); d) g, e, e 3 ; e) g, 6, e 3 e 3 ; here, g is the metric
tensor, e is Levi-Civita tensor.
0 5.15 Determine the symmetry group of the tensors g, S where S is a symmetric

second-rank tensor of the general form, g is the metric tensor (orthotropy with the
planes of symmetry determined by principal axes of the tensor S). Consider the case
of coinciding eigenvalues of the tensor S.
0 5.16 a) Find all the second-rank tensors invariant relative to the symmetry
group of the tensor g. b) It is known that there are three linearly independent
fourth-rank tensors invariant relative to the complete group of rotations. Compose
their components of the components of the tensor g.
0 5.17 For the setting of Problem 5.14 find all the symmetric second-rank tensors
invariant relative to the corresponding symmetry groups. For what groups do there
exist invariant antisymmetric second-rank tensors?
© 5.18 Find all the symmetric second-rank tensors which are invariant relative to
the symmetry group of the tensors g, S (Problem 5.15), and compose them of the
tensors g, S.
© 5.19 Consider the matrix which is inverse to a nonsingular matrix of the co-
variant components of a second-rank tensor. Show that it is the matrix of the con-
travariant components of a certain new tensor, i.e., it defines a tensor function.
© 5.20 It is known that there exist ten linearly independent fourth-rank tensors
invariant relative to the group of transverse isotropy (Problem 5.14c). Compose their
components using the components of the tensors g and e3e3.
© 5.21 Determine the general form of the tensor function a) a = F(6, g) where
a, b are vectors, g is the metric tensor; b) c = F(a, 6, g) where c is also a vector.
© 5.22 Prove that the tensor function of Problem 5.21a has a scalar potential
$(b,g) such that a{ = —.
0 5.23 Let a tensor function F(S, g), where S is a symmetric second-rank tensor,
g is the metric tensor regarded as constant, have the form
F = (kl9ij + kiSij + k^'S^ye*
where the coefficients fci, k2, k3 are the functions of the invariants of the tensor S.
Show that the necessary conditions for existence of a scalar potential $ of the function
F, Fij = ^TT", can be represented in the form
dS*'
1 dka 1 dkp
a oJp p aJa
where Jx = SU, J2 = P'Sij, h = S^ShS*,.
BASIC CONCEPTS
48
0 5.24 Using the interpolation polynomial of Lagrange (Problem 5.11), determine

the form of the tensor functions e s , sin S, S 3 . Compare with the expansions in the
power series. Determine the form of the scalar coefficients of the representations
ag + 6S+cS 2 .
0 5.25 Calculate the tensor of moments of inertia relative to the center of mass
a) for a homogeneous solid sphere, using the properties of symmetry; b) for a homo
geneous ellipsoid, using the principal axes.
© 5.26 Show that the tensor of moments of inertia relative to the center of mass of
a homogeneous regular tetrahedron is spherical (proportional to the metric tensor g).
Show the same for a cube and a regular octahedron. Use the properties of symmetry.
0 5.27 Determine the general form of scalar, vector and second rank tensor fields
invariant relative to the symmetry group of Problem 5.14A Such fields are referred
to as axially symmetrical. How do the results change for the other groups of Prob
lem 5.14 ? Write the results in the cylindrical coordinate system.
9 5.28 Determine the general form of scalar, vector and second-rank tensor fields
which are invariant relative to the complete group of rotations and reflections. Such
fields are referred to as spherically symmetrical. Write the results in a spherical
coordinate system.
Chapter 2
General Laws and Equations of
Continuum Mechanics
6 Summary of General Laws and Equations

General or universal (i.e., valid for all media) physical "conservation laws" are con
sidered in this chapter. In Newtonian mechanics they are the mass conservation law,
the momentum conservation law, the angular momentum conservation law, the en
ergy conservation law (first law of thermodynamics, energy conservation principle)
and the entropy variation law (second law of thermodynamics, entropy increase prin
ciple). These laws are usually formulated (postulated) for finite material volumes of
a continuum (in integral form). They result in differential equations satisfied at each
point of a domain occupied by a medium for motions described by smooth functions.
If the functions describing a motion are discontinuous on some surfaces the jumps of
the functions across these surfaces are related by equalities called jump conditions
also resulted from the general conservation laws written in integral form. Below, the
summary of the general laws and equations is given. They are considered in detail in
Sections 7-18. The mass conservation law states that the mass of a material volume
is constant
| / ^ = 0. (6.1)
v
The momentum conservation law states that the rate of change in momentum of a
material volume equals the sum of the external forces acting on it
j t I pvdV = J pFdV + jpnda. (6.2)

V V E
The angular momentum conservation law states that the rate of change in angular
momentum of a material volume equals the sum of moments of external forces and
49
50 GENERAL LAWS AND EQ UATIONS
couples acting on it
— / p(r x v + k)dV
at J
v
= frx pFdV + frx pnda + J phdV + f Qnda. (6.3)
V £ V £
The energy conservation law states that the rate of change in energy of a material
volume equals the input of energy per unit time from the outside as work of external
forces, heat and other forms of energy
v
V '
jPF-vdV + jpn-vda- jq"nda + / p ^ f W . (6.4)

V E £ V
The entropy variation law states that the rate of change in entropy of a material
volume equals the sum of the input of entropy from the outside and the production
of entropy within the volume per unit time. The production of entropy within the
volume is not negative
Tt=JtlpsdV =
-dT + ^ -dT^°- (6 5)
-
V
In relationships (6.1)-(6.5) the following notations are used: V is a material volume,

T, is the boundary of V, t is the time, p is the density, v is the velocity, F is the
body force per unit mass, pn is the surface force acting on the surface element with
the unit normal n per unit area, r is the radius-vector of an arbitrary point, k is
the mass density of internal angular momentum, h is the mass density of external
body couples, Qn is the surface density of external surface couples, u and s are the
mass densities of internal energy and entropy, q£ and dq^^/dt are the surface and
mass densities of energy input per unit time with the work of macroscopic forces
excluded, deS/dt is the entropy input from the outside per unit time, and d\S/dt is
the entropy production per unit time. For smooth motions, these laws are equivalent
to the following differential equations: the continuity equation
~ + pd\vv = Q\ (6.6)
the momentum equation

p-£ = pF + Vjpêi- (6.7)
6. Summary of General Laws and Equations 51
the angular momentum equation

dk
P
~dt=pH + V Ql ei
i ' + ( e ' * e >)P° ; (6-8)
the energy equation
d,v2
P { + u) = (F- v)p + Vj{pijv'.) + P V'; (6.9)
Jt -2 dt
the entropy equation
ds ^Smass ^ + p
diS
P = > 0. (6.10)
dl "dt' dt -
In relationships (6.6)-(6.10) the following notations are used: p y ' are components of
the stress tensor, Vj are covariant derivatives, e* are basis vectors, Q ,J are components
of a couple stress tensor, q'^ are components of an energy flux vector (exclusive
of the work of macroscopic forces), sJ are components of an entropy flux vector,
pdesmass/dt~VjSj is the entropy input from the outside per unit volume per unit time,
dis/dt is the mass density of entropy production per unit time. As a consequence of
universal laws (6.1)-(6.5), the following conditions must be satisfied across an isolated
discontinuity surface Ej:
PiVni = m + p2vn2 (6.11)
R
+ Pnl- />lWnl«l = P„2 - PiV„2V2 (6.12)
M + Q n l - P i M n l = Q„2 ~ P2*W.2 (6-13)
W + p n l - V i - p i ( ^ / 2 + U i K i -q*nl
= Pn2 ■ v2 ~ P2(v\l2 + U2)vn2 - q'n2 (6.14)
P\Vn\S\ ~ p2Vn2S2 + Sni -Sn2 = Q (6.15)
Here, indices 1 and 2 mark the values of quantities on different sides of the disconti
nuity surface Ed and n is the unit vector normal to Ed directed to the side 1. The
conditions (6.11)-(6.15) are written for a coordinate system in which the velocity of
the given point of the surface E d at the given instant equals zero; m, R, M and
W are the surface densities of external mass, momentum, angular momentum and
energy inputs on Ej, f! is the surface density of entropy change at the expense of ex
ternal inputs of heat and mass on E d as well as at the expense of irreversibility of the
transition through the discontinuity. If there are no external heat and mass inputs on
Ed and p\vn\ ^ 0, P2^n2 ¥" 0> then, according to the second law of thermodynamics,
n>o.
52 GENERAL LAWS AND EQUATIONS
7 Mass Conservation Law. Continuity Equation.

The continuity equation is a consequence of the mass conservation law. It can be
written in different forms. In the Eulerian description, the following two forms are
used:
^ + divpu = 0, (7.1)
^ + pdivu=0. (7.2)
at
In the Lagrangian description, the following forms of the continuity equation are used:
^ + p d i v t > = 0, (7.3)
where divu is determined by derivatives with respect to concomitant Lagrangian
coordinates £J;
( dx* \
p det —r- = p0, (7.4)
where po, p are the densities, x', z' are the spatial coordinates of a material point in
the initial and current states, x* = x*(x ; ,r);
' \ / 9 = PoV3, (7.5)
where 9 = det(<?y), g = det(py), 9 ,}, gtj are the metric tensor components in a
Lagrangian coordinate system £' in the initial and current states;
P = . " (7.6)
>/l+2/1(l) + 4/2(l)+8/3(l)
P
Po = ■ (7.7)
v / l - 2 / 1 ( e ) + 4/ 2 (£)-8/ 3 (f)
where /*(£ ), h(e) are the invariants of Green and Almansi strain tensors e , e,
defined by the formulae /i(e) = e\, h{e) = \{If - e ! ^ ) , h(e) = det(c*J). When
using other kinds of strain tensors, other forms of continuity equation in Lagrangian
coordinates can be obtained. The incompressibihty condition
divr = 0,
m
det —;- = 1 , Jg = V9, etc.
\dx3J
(7.8)
is traditionally called the continuity equation for an incompressible medium. The

mass conservation law, for an incompressible medium, yields
ft=0, p(4\r) = po(0. (7.9)

7. Mass Conservation Law. Continuity Equation. 53
— PROBLEMS —
0 7.1 Derive the formula for differentiation of an integral over a moving volume
with respect to time
JtJAdV = j ^ d V + l Avnda (7.10)
V V £
where vn is the projection of velocity of points of surface E bounding the volume V
on the external normal to E.
© 7.2 Write the mass conservation law for a finite unmoving spatial volume
through which a medium flows. Use the formula (7.10).
0 7.3 a) Derive the continuity equation in Eulerian description from the mass
conservation law for a material volume, b) Write it in a Cartesian coordinate system,
expanding the expressions dp/dt and div v. c) Derive this equation again, considering
the mass balance for an infinitesimal rectangular parallelepiped with faces parallel to
the coordinate planes: the change in mass contained in a space volume equals the
quantity of mass flowing to it through its surface.
© 7.4 Prove that divw (= VJU') equals the rate of relative change in volume of
the infinitesimal vicinity of the considered point of medium moving with the velocity
v.
0 7.5 A motion is referred to as potential if there exists a function <p such that
v = grad ip. Write the continuity equation for a potential motion of a compressible
and an incompressible medium in terms of the potential.
0 7.6 Derive the continuity equation in Eulerian variables a) in a cylindrical
Z r
U r*> M o \*™\
b)
Figure 7.1
and b) in a spherical coordinate system, considering mass balance in an elementary

coordinate volume (Figure 7.1).
0 T.7 Write the continuity equation in Eulerian description in a curvilinear or

thogonal coordinate system, using the physical components of the velocity vector.
Derive the continuity equation from it a) in a cylindrical coordinate system, b) in a
spherical coordinate system.
© 7.8 Write the continuity equation for one-dimensional motions with the plane,
cylindrical and spherical symmetry (i.e., when all the quantities depend upon only
one spatial variable r and time t, and the surfaces r = const are, respectively, planes,
cylinders and spheres; besides, only the velocity component along the coordinate line
r does not vanish).
© 7.9 Using the condition that the mass of any material volume in the initial state
equals that in the final one, derive the continuity equation in Lagrangian variables in
forms (7.4), (7.5).
© 7.10 Derive the continuity equation in Lagrangian description in forms (7.6),

(7.7) relating the density variation to the invariants of the strain tensors. Show that
these two forms coincide in the linear approximation for the case of small strains.
© 7.11 A planar flow of an incompressible fluid is bounded, from one side, by
Figure 7.2
the immovable impenetrable bottom z = —h(x) and, from the other side, by the
free surface z = £(x, t) (x, y, z are Cartesian coordinates; all characteristics of the
planar flow are independent of the Cartesian coordinate y, and vy = 0). Introduce the
average over the depth (over z) velocity vx = vx(x,t). Obtain the equation for £(x, t)
and vx(x, t), applying the mass conservation law to a volume of the fluid between two
close cross sections of the flow.
Q 7.12 A weakly compressible fluid (the density p(x, t) slightly differs from p 0 =
const) flows in a tube with weakly deformable walls (the cross-sectional area of the
tube S(x,t) slightly differs from So(x) = S(x,0), x is the coordinate along the tube
axis). Regard the speed v along the tube axis as constant over a cross-section and
8. Stress Tensor 55
slightly different from v0(x) = v(x, 0) to obtain the linearized continuity equation,
considering the mass balance in a volume of the fluid between two close cross sections.
0 7.13 Prove the formula
V V
1 2 3
where x stands for i , x , x , V is a material moving volume, p is the density of the
medium, dA/dt is the total derivative with respect to t.
0 7.14 For a planar stream {vx, vy do not depend upon z, vz = 0 in Cartesian
coordinates x, y, z) the velocity field is given:
A(x2-y2) 2Axy
vx = —-—-. -, vy = —T—, vz = 0, A = const
r r*
where r = y/x2 + y2. Prove that this velocity field is a velocity field of an incompress
ible fluid.
0 7.15 In a planar stream (vx, vy do not depend upon z, vz = 0 in Cartesian co
ordinates x, y, z), the velocity component vx is given: vx = —Ay/r2, A = const, r =
V^x2 + yl. a) Find the component vy in the whole stream if the fluid is incompressible
and vv —> 0 as y —► cc for all values of x. b) Show that the flow is potential every
where except of the axis r = 0, and the potential is a harmonic function, c) Prove
that the particle paths are circles.
0 7.16 Show that, if a motion of an incompressible medium (divv = 0) in a
bounded simply connected domain is potential (v = grad <p), the velocity field is
uniquely determined by values of the normal velocity component vn\r on the bound
aries T of the domain, and the dependence of v upon vn\r is linear, i.e., the velocity
field is completely determined by kinematics and does not depend upon external
forces.
8 Stress Tensor
To describe forces acting on a surface, the surface density of these forces or stress
vector p n is introduced. The stress vector pn is the limit of the ratio of the surface
force AV acting on a surface element with the normal n to the area ACT of this
element as ACT —» 0 and the element shrinks to a point
p n = hm —— .
™ A<7-O ACT
Figure 8.1
The Cauchy stress tensor is the tensor with the components p ,J such that components
pxn of any stress vector pn at the considered point, for all n, can be calculated according
to the formula
Pli=PiJnj (8.1)
The existence of the Cauchy stress tensor follows from the momentum conservation
law. When using a Lagrangian coordinate system based on the initial state, the
nominal stress vector ir^ related to p n by the formula
•no da0 = VTA°~ (8-2)
is often used. In (8.2), do0, da and no, n are, respectively, the areas of a surface
element and its unit normal vectors before and after deformation. In other words,
TT^ is the limit of the ratio of the surface force to the initial area of the element. The
Piola-Kirchhoff stress tensor with the components IT1* is defined by the relationships
= TT'J, (8.3)
where noj are the components of the normal to the surface element in the initial state.
This tensor is often called the first Piola-Kirchhoff stress tensor. Components of the
Cauchy and Piola-Kirchhoff tensors in a Lagrangian coordinate system are related by
the formulae
-^°ei = -pkjek (8.4)
A) P
where po and p are initial and current densities, e < and e-k are basis vectors of the
Lagrangian coordinate system in the initial and current states,
e* =efc + | p = (4+ V t £*)!,, (8.5)
and w is the displacement vector. If Lagrangian coordinates are simply the initial
coordinates £< of particles in an Eulerian coordinate system x, and the latter system
8. Stress Tensor 57
is Cartesian, then the following formulae are valid
o Po dxk p dij o
Tit (* ) = — " 5 — Py (*). Py W = — -TT- *ik (x )
p dij Po d xk
The symmetry of the Caushy stress tensor
p« = p*
is derived from the angular momentum conservation law if an internal angular mo
mentum and moment interactions are absent. In general, the Piola-Kirchhoff stress
tensor is not symmetric.
— PROBLEMS —
© 8.1 A homogeneous heavy bar moves uniformly along a horizontal rough plane.
Figure 8.2
The area of the face touching the plane equals 0.02 m 2 , the mass of the bar equals 1 kg,
the coefficient of Coulomb friction between the bar and the plane equals 0.3. Find
the stress vector on the plane under the bar (do not take into account an atmospheric
pressure).
0 8.2 a) Using the momentum conservation law, obtain Formula (8.1) denning
stress tensor, b) What restrictions to the body forces and acceleration fields are
necessary to obtain Formula (8.1)? c) What is the physical meaning of the component
p 12 in a Cartesian coordinate system?
Q 8.3 Derive Formula (8.3) defining the Piola-Kirchhoff stress tensor from the
momentum conservation law.
0 8.4 a) Derive the formulae relating the components of Cauchy and Piola-
Kirchhoff stress tensors, b) What is the mechanical meaning of the components
irki if the initial Lagrangian coordinate system is Cartesian?
© 8.5 a) Show that, in an arbitrary coordinate system, the stress vectors acting
on the coordinate surfaces x% = const (let us denote them p " ) are determined by the
formula
p " = p^'ej/yg" (there is no summation over i).
b) Find physical components of the vectors p " (the components in the unit basis
ej/|e,|). c) What do p" and their physical components equal in a Cartesian coordi
nate system?
© 8.6 Derive the property of symmetry for a stress tensor from the angular mo
mentum conservation law when the internal angular moments and distributed couples
vanish.
© 8.7 The components of a stress tensor are given (in Pascals) at a point of a
body in a Cartesian coordinate system as follows
100 100 160

ij
(P ) 100 0 -150
160 -150 - 6 0
Find the components of the stress vector pn and magnitudes of the tangential stress
p nT and normal stress p n n for the surface element with the components of the normal
rii = 1/2, n 2 = 1/2, n 3 = l / \ / 2 . Find the angle 0 between p n and n.
© 8.8 The dimensionless components of the stress tensor at a point M are given
in a Cartesian coordinate system as follows:
Find the stress vector p n on the surface element with the normal n:
1 2 2
n = e i _ e 2 + e 3
3 3 3 -
© 8.9 Stress vectors pn and p n i act on the surface elements with normals n and
n i at a point M. Show that
=
Prnii Pmn
if the stress tensor is symmetric.
© 8.10 The tensor surface of a tensor p at a point M is the locus of the points
whose coordinates satisfies the condition pijXxx> = const, where p y are components of
p at the point M in a Cartesian coordinate system x\ The tensor surface of the stress
8. Stress Tensor 59
tensor at a point M is called the stress surface for the point M. Identify the form
of the stress surface for a point of a medium where the components of stress tensor
in an orthogonal Cartesian coordinate system are as follows: a) pn = P22 = P33 = P,
Pij = 0 for i ^ j (isotropic extension (p > 0) or compression (p < 0)); b) p n = p, the
other p^ = 0 (simple extension (p > 0) or compression (p < 0)); c) pu = P21 = T, the
other p^ = 0 (simple shear); d) p n = A, p22 = B, p^ = C, p{j = 0 for i ^ j (stressed
state of the most general form in the principal axes of the stress tensor). Study the
dependence of the shape of the stress surface upon the signs of A, B, C.
0 8.11 a) Show that the normal n1 to a stress surface (see Problem 8.10) in the
point of the surface with a radius-vector r is parallel to the stress vector on the
Figure 8.3
surface element with the normal n = r/\r\. b) Using geometrical properties of a

stress surface, show that, for an arbitrary stressed state at a point, there exist three
orthogonal vectors nu n 2 , n 3 , determined for this point, such that p n i || n ^ p„ 2 || n 2 ,
Pn3 II n 3 - c) Show that, if pn || n for all n at a given point, the stress surface for this
point is a sphere.
0 8.12 The principal coordinate system of a symmetric tensor p at a given

point is the Cartesian coordinate system in which p i ; = 0 for z ^ j . The axes of
this system are called the principal axes of the tensor p, and the components p u ,
P22, P33 (denoted ordinarily with px, p 2 , p3) are called its principal components or
eigenvalues (see Section 2). a) Show that, on the surface elements perpendicular to
the principal axes of a stress tensor, the stress vectors are directed along the normals
to the elements, b) Let the principal axes and principal components of a stress tensor
be known. Find the surface elements on which the values of the normal stresses pnn
are extremal. Find the values of (p„„)max and (pnn)mm themselves, c) A tensor has
components
/ 12 4 4
(p'i) = 4 0 8
V 4 8 0
in a Cartesian coordinate system at a point M. Find its principal components and

principal axes.
© 8.13 Decompose the stress tensor given in Problem 8.12c into the spherical
component and deviator.
© 8.14 The components of a symmetric stress tensor is given in a Cartesian or

thogonal coordinate system: pn — p-n = Pzz = 0, pn = P13 = -P23 = 1- (The values
of pij are related to some characteristic stress value p 0 and given in a dimensionless
form.) Find its principal components (eigenvalues), possible directions of principal
axes (eigenvectors) and write out the transformation of the coordinate system into
the principal system (see Section 2).
0 8.15 The components of a stress tensor field p in a Cartesian orthogonal coor

dinate system x 1 , x 2 , x 3 are given by the formulae
Figure 8.4
Ccosf, . 6 30 x Ccosf9 / 1 " • 30.

Pll =
^^(1"sin2smT)' P 2 2=
- ^■(l+sm- siny) ,
C cos % 9 39
Pl2 = P2l = ^r—- Sin - COS — , p,3 = p 3 l = p 2 3 = P32 = P33 = 0 ,
\/27rr 2 I
i x2
r = v ( x l ) 2 + (l2)2i ^ =
arctan —, C = const,
(These formulae are known to be the asymptotic (at small r and 9 ± 7r) expressions
for the stress tensor components near the edge of a slot extended symmetrically by
a normal load applied to its shares in the case of plane stress state, see Figure 8.4).
Find the eigenvalues pi, pi, p 3 of p and the orientation of the principal axes in space.
In particular, obtain a formula for the angle of inclination a of the principal direction
corresponding to the greatest p< with respect to the axis x 1 .
8. Stress Tensor 61
0 8.16 a) Stress vectors p n i , p n 2 , p„ 3 act on three mutually orthogonal surface

elements with the normals nu n 2 , n 3 . Show that (p n i ) 2 + (p„ 2 ) 2 + (p„ 3 ) 2 does not
depend upon the orientation of the surface elements, b) Write out I\,h,h — the
first, second and third invariants of the stress tensor
/i=P* 4 , hÛif-V^.l h = det(jp\)
in terms of principal components.

© 8.17 Obtain formulae expressing normal and tangential stresses on the surface
element with normal n at a point M in terms of the principal components of the
stress tensor at this point. Assume the projections of n on the principal axes of the
stress tensor to be known.
© 8.18 A surface element making equal angles with the principal directions of a
stress tensor is referred to as octahedral. Show that the normal component of the
stress vector acting on it equals I\/3 where I\ is the first invariant of the stress tensor
(see Problem 8.16). Show that the tangential stress on an octahedral surface element
(octahedral tangential stress) equals
g V/(PI - P2)2 + (P2 - P3)2 + (PS - pi) 2 = y - j j hd
where I2d is the second invariant of the deviator of the stress tensor, pi, p 2 , Ps are the
eigenvalues of the stress tensor.
0 8.19 a) Find the maximal tangential stresses p nT max and the normal unit vectors
to the surface elements on which they act, regarding the principal axes and principal
stresses as known, b) Solve the problem (a) for the special cases when some of the
principal stresses are equal, c) Construct the surface y(p'j) - k = 0 where k = const,
ipip'j) = Pnrmax (the Tresca yield surface for plastic materials) in the principal stresses
space. Find the distances from the origin to the planes forming the surface and to
the vertices of the surface cross-section by the plane passing through the origin and
orthogonal to the line p\ = P2 = Pa-
© 8.20 A bar had the shape of a rectangular parallelepiped measuring a x b x c
before deformation. Surface forces are applied to the opposite faces with the area
So = ab. The resultants of the forces equal P and - P and act along a line parallel to
the edges of initial length c. The other faces (lateral faces) are not loaded, and body
forces are absent. The body is deformed as shown on the Figure 8.5 (the initial form of
the bar is shown by dashed lines). Let the Cauchy and Piola-Kirchhoff stress tensors
be known not to depend upon coordinates, a) Write out the conditions satisfied by
Figure 8.5
the components of the Piola-Kirchhoff tensor on the boundary of the body, b) Find
the Piola-Kirchhoff stress tensor, c) What additional information must be known in
order to find the Cauchy stress tensor? d) Solve the problems (a), (b), (c) for the
case when the bar is a right cylinder with an arbitrarily shaped cross-section.
0 8.21 A Piola-Kirchhoff stress tensor is given in a Cartesian coordinate system:
a) b) ^ c)
Figure 8.6
7ru = a 7^ 0, the other 7Ty = 0. Figure 8.6a shows the initial state of a body, and
8.6b, the deformed state. What would the Piola-Kirchhoff tensor transform into, if the
body in the deformed state together with all stresses and deformations is rotated by
IT/2 around the line parallel to the axis £3 in the initial state as shown at Figure 8.6c?
0 8.22 Find the dependence of normal and tangential stresses on various surface
elements upon the orientation of the surface elements in the case of simple extension
of a bar (see Problem 8.20). The cross-sectional area of the deformed bar equals S,
and uniformly distributed loads with the resultant P act on the ends of the bar. Find
the maximal normal and tangential stresses and the normals to the surface elements
corresponding to them.
0 8.23 Verify the equivalence of the following statements (a) and (b) defining the
model of an ideal fluid: a) the stress tensor is spherical, i.e., p'' = -pg'j, p is a scalar;
9. Differential Equations of Motion and Equilibrium 63
b) the stress vector on any area element is parallel to the normal to this element.
Show that p'-* = — p<?,J in any fluid at rest (tangential stresses are known to vanish in
a fluid at rest).
9 Differential Equations of Motion and Equilib

rium
Differential equations of motion (or equilibrium if v = 0) can be derived from the
momentum conservation law for a domain where the state of the medium is described
by smooth functions. These equations are also called momentum equations. In either
Eulerian or Lagrangian description using the basis e< of concomitant system at the
current time (see Section 4), these equations have the form
pa> = pF* + Vkp>k, 0* = — + ^ * (9.1)

at
where F* are the components of body force per unit mass, plk are the components of
the Cauchy stress tensor, and a* are the components of acceleration vector a. In the
Lagrangian description using the basis et of a concomitant coordinate system at the
initial time (see Section 4), the equations of motion can be written in the form
d2 w' o . o ,
where po is the initial density of the medium, w ', F ' are the components of the
displacement vector and body force density in the basis e*, and 7r'* are the components
of the Piola-Kirchhoff stress tensor in the basis ej.
— PROBLEMS —
0 9.1 Derive differential equations of motion in the Eulerian description from the
momentum conservation law. Write them in terms of components in an arbitrary
coordinate system and in a Cartesian orthogonal one.
0 9.2 a) Find the relation between the components of the acceleration a1 and
those of the velocity vk in a concomitant system, b) Write out the equations of
motion in a concomitant Lagrangian coordinate system, c) Show that the expressions
Vkp,k and a' in a concomitant system contain the derivatives of the strain tensor
components.
© 9.3 Derive differential equations of motion in an initial Lagrangian coordinate

system from the momentum conservation law.
© 9.4 A stress tensor field is given in a Cartesian coordinate system by the fol
lowing matrix
( -P 0 0 \
(P,j) = 0 -p 0
V 0 0 -p j
2 2 2 2 2 2
where p = p 0 — 0.5u p(a - r ), r = x + y , p0, a, ui are constants, p is the density
of the medium. The medium is in equilibrium. What body forces act on it?
© 9.5 In a Cartesian coordinate system, pn = — pgx + f(y,z) where density

p = const, g — const, the other py equal zero. Find the body forces if the medium is
known to be in equilibrium.
© 9.6 The matrix of stress tensor components in a medium at rest with a given
density p has the form
/ a 0 0
ij
(P ) =
where a, b, c are constants. Find the body forces if p,j are the contravariant compo
nents a) in a Cartesian coordinate system, b) in a cylindrical coordinate system, c) in
a spherical coordinate system.
© 9.7 a) Derive the equation of equilibrium for a fluid using the fact that the
stress tensor in a fluid at rest is spherical: p'-* = — pgli (see Problem 8.23). b) Write
the equation of fluid equilibrium in the gravity field, c) Show that if F ■ curl F ^ 0,
no fluid can be in equilibrium in the field of body forces F.
© 9.8 A membrane is a very thin film which does not resist bending, but resists
extension. Let a membrane uniformly extended by a tension T be loaded by a vertical
force with density q (the axis z is directed vertically downward; the axes i and y and
coordinate origin are placed in the initial horizontal plane of the membrane). Derive
the equation for the vertical deflection w(x,y), assuming that the deflection is small
and that the tension T does not depend on w.
0 9.9 a) A body is in equilibrium under the action of only surface forces. Let
d2*3 d2*2 a 2 *!
12
P ~ ~a a. ' Pl3 a Q . P23 = ~ a Q
axdy axoz oyaz
where *!, * 2 , *3 are arbitrary thrice-differentiable functions. Show that the equilib
rium equations are satisfied if
a2#3 32tt2 a 2 * :i a2^ a2*2 a2*!
P 1 1 - dy 2
a*2 ',
— +' — PP22
2 ~2 -a*—2 + -^", ?„ = — + —
b) Let
a2*2
I a 2 *3
oyoz ozox P33 = dxoy
Show that the equilibrium equations are satisfied if
__i_a_ / » i a*2 a*s
2 dx \ dx dy dz
_ _ i _a_ /a*i _ a*2 a^3\
P31 +
~ 2 dy \ dx dy dz )
Pl2
~ 2 dz [ dx + dy dz
The functions \I>i, ^2, *3i in terms of which all the components of the stress tensor
can be expressed so that the equations of equilibrium are identically satisfied, are
called the stress functions.
0 9.10 Show that, if body forces are absent, equilibrium equations for a plane
stress state (in a Cartesian coordinate system only pn, P12 and p 22 d° not equal zero,
and they depend only upon x and y) allow the Airy stress function U(x,y) to be
introduced such that
d2U d2U d2U
P " = l O 2" - eP22 = "HZ5T
dy ' " dx2 '- eP12
" = ~ dxdy ■
0 9.11 Let a cylindrical rod be in equilibrium under the action of forces applied
only to its endwalls, and all components of the stress tensor except P13 = p 31 , p 23 = P32
are equal to zero at all points. The orthogonal Cartesian axis x, y, z are chosen as
shown on the Figure 9.1, external body forces are equal to zero. Show that a) there
exists a stress function T(x,y):
dT dT
P31 = P32 =
~dy-' "ax"'
introduction of which leads to satisfaction of all the equilibrium equations; b) the
condition pn = 0 on the side surface of the rod is reduced to the condition dT/ds = 0
along the contour of the cross section in a plane x, y (s is the length of the arc of the
contour); c) if the rod has the stress state mentioned above, then the resultant of all
forces acting at any cross section equals zero, and the resultant of the force moments
acting on a cross section is reduced to a torsional moment.
Figure 9.1
© 9.12 Derive the equations of motion for an ideal fluid (the Euler equations)
from the general equations of motion of a continuum. In an ideal fluid, by definition,
the stress tensor is p ,J = —pg'-*-
© 9.13 Obtain the Euler equations applying the momentum conservation law to
an individual volume V of an ideal fluid.
© 9.14 For a barotropic motion (i.e., when the density is a function only of the
pressure) in a field of potential body forces, transform the Euler equations to the form
of Lamb-Gromeka
dv v
— + grad —- + 2u> x v = grad U - grad V
where u = | c u r l v is the vortex vector, U is the potential of body forces, V =

p
/ dp/p(p) is the pressure function, po is some constant.
PO
© 9.15 Let a motion of an ideal fluid be steady, external body forces be potential
(.F = gradt/). Show that, for this motion, the Euler equations have the integral
(called the Bernoulli integral)
v2
where C = const along a line £ if £ is a streamline or vortex line; V{p,C) =

/ dp/p(p, £) is the pressure function whose values are different for different stream
lines or vortex lines if the motion is not barotropic (see Problem 9.14). Write the
Bernoulli integral along a streamline for a motion of an incompressible fluid in a field
of gravity.
0 9.16 Let a motion of an ideal fluid be potential (v = gradip), external body

forces have a potential (F = gradt/), the motion is baxotropic (density p depends
only on pressure p). Show that the Euler equations have the integral (called the
Cauchy-Lagrange integral)
where V = Jdp/p is called the pressure function, T[t) is an arbitrary function of

time.
© 9.17 Prove the validity of the following formulae for the acceleration compo
nents in an Eulerian arbitrary orthogonal coordinate system (the basis vectors of
observer's coordinate system e< do not depend on time)
(no summation over i). Here py are the metric tensor components.
© 9.18 Write the Euler equation in both cylindrical and spherical coordinate
systems.
0 9.19 In a linearly viscous isotropic fluid, the components of the stress tensor p i ;
are bound with the components of the strain rate tensor e^ by the relationships
Pij = ~P9ij + AJi(e)ffy -I- 2p£ij
where p is the pressure, gtj are the components of the metric tensor, Jê) = eugkl
is the first invariant of the strain rate tensor, A and p. are the viscosity coefficients.
Assuming A = const, \i = const, obtain the following form of the equations of motion
for a linearly viscous fluid (called the Navier-Stokes equations)
dv
p-— = pF - gradp+ (A + /i)graddivv + p.Av
at
where A is the Laplace operator.
0 9.20 At small relative displacements in a linear elastic isotropic medium, the

components of the stress tensor Py are bound with the components of the small strain
tensor eu by the relationships (called Hooke's law)
Pa = XiJi(e)gij + 2p.ieij
where Ji(e) = Ekigkl, Ai and H\ are coefficients (called Lame coefficients). Assuming
Ai = const, Hi = const, obtain the following form of the equations of motion for a
linearly elastic medium (called the Lame equations)
of
where w is the displacement vector, A is the Laplace operator.
© 9.21 Prom the equation of motion for a continuum, obtain the equation deter
mining variation of kinetic energy E^n for a material volume V of the continuum (the
kinetic energy theorem)
d£ k i n dA& dA® „ f v2 _, dA& r ._ . „, r. ,,

V V E
where dA^/dt, dA^/dt are respectively the work of external and internal forces per
unit time. Find the density of work of internal surface forces (i.e., the limit of the
ratio of the work of internal surface forces to the mass of the particle when the particle
shrinks to a point) a) for an arbitrary continuum b) for an ideal incompressible fluid.
© 9.22 An explosion happens at a point M of the unbounded space filled with
Figure 9.2
a fluid at rest. A quantity of energy EQ is instantly discharged and transmitted to

the fluid. Determine the motion of the fluid. Do not take into account the viscosity
and compressibility of the fluid, and neglect the gravitation as well as the pressure
9. Control Surface Method 69
value at large distances from the point of the explosion. Suppose that all the energy
is spent into creation of the kinetic energy of the fluid and that spreading of the fluid
from the point M results in appearance of an expanding cavity with the center at the
point M inside of which the pressure equals zero (vacuum), a) Set up the problem
mathematically, i.e., write the complete system of equations for determination of v
and p, boundary and initial conditions in the most convenient coordinate system.
b) Find the radius of the cavity and the distribution of v and p at instant t. c) What
is the maximal value of the pressure at instant t? Where is it reached? d) Use
dimensionality theory, in particular, the Il-theorem (see Section 36) to show that
the problem is self-similar i.e. is reduced to determination of functions of only one
combination of r and t of the form r/ta. Does the problem remain self-similar if the
pressure at infinity p<» does not equal zero? e) Consider the problem at p^ ^ 0. f) Is
a similar formulation of the problem possible if the explosion happens not at a point,
but on a straight line or on a plane?
10 Using the Conservation Laws in Integral Form

for Calculation of Forces and Moments Acting
on Bodies Moving in a Fluid (Control Surface
Method)
When one studies motion of bodies in a fluid or flow of a fluid around bodies, what
is often of interest is not the detailed distribution of the velocity and pressure over
all the fluid, but only integral characteristics of the motion, e.g., the total force and
moment acting on the body from the fluid. The desired quantities can be obtained,
in many cases, without differential equations, using only the appropriate laws in
integral form, i.e., written out for a certain appropriately chosen volume referred to
as a "control volume." The boundary of this volume is called the control surface. For
steady motion, the mass, momentum and angular momentum conservation laws (6.1)
- (6.3) for a fluid in volume V can be written in the form
fpvnda =0 (10.1)
J pvvn do = I pFdV + jpnda (10.2)

£ V Z
j p(r x v + k)vn dT, =

J p(rxF + h)dV + J(rxpn + Qn)da (10.3)

V E
where E is the boundary of V, vn is the normal component of the velocity of the

medium at E. The volume integrals in these relationships are the resultant and the
total moment of the external body forces and couples which are often known for any
volume V. For example, if the body forces are the gravity forces, / pFdV is simply
v
the weight of the fluid in volume V. The control volume V is chosen in such a manner
that the surface of the considered bodies be a part of its boundary, and parameters
of the medium on the other boundaries be known in the sufficient degree. The forces
and moments acting on the bodies are represented by the integrals of the densities
of the surface forces and moments over the surface of these bodies, so they can be
calculated from (10.2),(10.3) if the integrals over other parts of boundary E in (10.2),
(10.3) are known. This method of determination of the total action of the fluid on the
bodies is called the control volume method or control surface method. To solve some
of the problems of this section, besides the relationships (10.1)-(10.3), the Bernoulli
integral is necessary. For an incompressible fluid, the Bernoulli integral can be written
in the form
v2 v
T + *-U =C (10.4)
where U is the potential of body forces, C = const along the streamline. The Bernoulli
integral (10.4) is valid for a steady motion of ideal incompressible fluid in a field of
conservative body forces.
— PROBLEMS —
0 10.1 a) Show that the mass, momentum and angular momentum conservation
laws can be written in the form (10.1)-(10.2) for steady motion, b) Write the energy
conservation law in a similar form.
0 10.2 Show that, for any closed surface E, the equalities p0nda = 0 and
/ r x p0n da = 0 are valid, where n is the unit vector normal to E and p0 = const.
0 10.3 There are fixed bodies A and B in a tube over which a fluid flows; Ei
and E 2 are cross sections of the tube far ahead and behind the bodies A and B;
T is the inner surface of the tube between the cross sections Ei, E 2 (Figure 10.1).
The flow is steady, flow occurs around the bodies without separation, and there are
neither body forces and couples, nor body inputs of energy. All the parameters of
Figure 10.1
the stream in the sections Ei and E 2 are known, a) Find R, M and W where R
is the resultant vector of the forces, M is the resultant moment of the forces and
couples acting from the side of the fluid on the bodies A, B and the inner surface
of the tube T, W is the energy input from the fluid to ,4, B and T. b) How would
the formulae for R, M, W obtained in the item a) change if gravity is taken into
account? c) Simplify the formulae for R, M, W obtained in the item a), assuming
that, in sections Ei, E2, 1) the couple stresses, angular momentum of the fluid, and
inputs of non-mechanical forms of energy are equal to zero; 2) the stresses are reduced
to the pressure (pn = —pn); 3) the velocity v, density p and pressure p are constant
along the section, d) Are the formulae for R, M, W obtained in the item a) valid if
there are discontinuities (jumps) of characteristics of motion in the fluid?
0 10.4 a) A jet of an ideal (p„ = —pn) weightless (F = 0) fluid flows onto a fixed
rigid wall and spreads over it (see Figure 10.2a). Far from the region where the fluid
encounters the wall the jet cross section area is S, the pressure, velocity and density
of the fluid in the jet are constant along its cross section and equal to po, Vo, and
po respectively. The velocity Vo makes acute angle a with the wall. The pressure in
the surrounding medium equals po- The fluid is incompressible. Find the additional
force R acting on the wall in the presence of the jet (R is the difference of the force
acting on the wall in the presence of the jet and the force acting on it due to the
outer pressure po in the absence of a jet), b) Solve this problem in the case of a plane
jet of thickness /. Find, besides the magnitude of the resultant force (per unit width
of the wall), its point of application as well as the velocity and thickness of each of
the jets (far from the point O). c) Calculate the velocity of a vertical water jet of
cross-sectional area 0.01 m2 necessary to support a platform with a clown standing
on it (see Figure 10.2c). The weight of the clown and platform equals 81 kgf.
0 10.5 An axially symmetric jet of an ideal incompressible fluid with cross section
S and velocity v perpendicular to S and uniform over a cross section of the jet far
ahead of an obstacle flows onto the obstacle. The obstacle is a cone of angle 2a
coaxial with the jet. The velocity and pressure also become constant over a section
far downstream from the vertex of the cone A. The motion is steady, and the ambient
Figure 10.2
Figure 10.3
pressure po is constant. The gravity force is neglected, a) Calculate the additional

(see Problem 10.4) force acting on the cone in the presence of the jet. b) Find the
force acting on an obstacle which is a hemisphere (see Figure 10.3).
0 10.6 A small cart moves without friction along a smooth horizontal surface. A
closed vessel with vertical walls filled with an ideal incompressible fluid stands on the
cart (see Figure 10.4). The total mass of the fluid, vessel, and cart equals m. The
pressure inside the vessel above the free surface of the fluid equals pi, the outside
pressure is po < pi- A jet of cross section So flows out through a hole in the back of
the vessel. Neglecting the change in mass m, calculate the acceleration of the cart
caused by action of the outflowing jet. Neglect gravity when calculating the velocity
of the jet.
Q 10.7 An ideal weightless fluid of density p flows out from a large vessel through
a cylindrical mouthpiece immersed in the fluid (Borda mouthpiece) with cross section
S (see Figure 10.5). The pressure in the vessel far from the hole equals p x , the outside
pressure equals po < Pi; So is the cross section of the jet far from the vessel. Find the
contraction coefficient SQ/S of the jet.
/77777777777777777777777777
Figure 10.4
Po
Figure 10.5
0 10.8 a) A rigid body moves with constant velocity v in an infinite cylindrical

tube filled with an ideal incompressible fluid. The fluid is at rest far upstream and
downstream, and the body forces acting on the fluid are zero. Does a reaction force
of the fluid R act on this body? b) What can you say about the reaction force of
the fluid acting at the conditions mentioned in item a) on 1) bodies moving along
the tube with identical velocities, 2) a body if a finite cavity filled with gas, vapor,
or liquid is formed behind of it?
Q 10.9 a) An ideal fluid flows steadily along the cylindrical horizontal tube with
cross-sectional area S. A body extending downstream to infinity, where it has the
— .;CD u.
Figure 10.6
form of a cylinder with the area of its cross section S0, is in the tube (see Figure 10.6).
The characteristics of the flow are constant within the cross section and equal to Vi,
Pi, p far upstream of the body. The velocity downstream at infinity is also constant
within the cross section, b) What is the drag force if S —» oo and So = const?
c) What is the force if the body extends upstream, but not downstream, to infinity?
0 10.10 Consider an infinite system of identical long cylindrical wings with par
allel elements disposed periodically, called a grid of profiles. A steady plane stream of
an ideal incompressible fluid flows around such a grid with the velocity perpendicular
to wing elements (see Figure 10.7 where the wing's cross sections are shown). The
velocity and pressure fields are periodic with the period I (the vector of the period of
the grid is denoted with I) and constant far upstream and downstream of the grid.
a) Determine the relationship between the force R acting on one wing per unit length
Figure 10.7
and the fluid upstream and downstream velocities, b) Studying the limit at I —► oo,
obtain the Joukovski (Zhukovskii) formula for the force exerted by fluid on a single
wing (per unit length)
R = pvx x r
where P is the vector of magnitude T directed perpendicular to the plane of the flow
and T is the velocity circulation around the airfoil, (V = fv ■ ds). c) Obtain the
c
Joukovski formula, considering a single wing in the flow in the absence of any other
wings.
© 10.11 Find the drag force of a body if it moves in an incompressible fluid and
causes a concurrent stream behind itself. Assume the distribution of the velocity in
the concurrent stream to be known far from the body, where viscous stresses can be
neglected and the pressure in the stream can be taken equal to the pressure in the
ambient fluid p<x,. The velocity of the body is kept constant (by external forces).
Choose the control surface a) fixed relative to the body b) fixed relative to the fluid
at infinity.
G 10.12 Find the force acting on a rocket flying with a constant speed v and
throwing mass m into the air per unit time. Regard the distribution of the velocity in
JO. Control Surface Method 75
the stream as known far behind the rocket, where viscous forces can be neglected and
the pressure can be assumed to be equal to the pressure in the surrounding medium
Poo-
0 10.13 Derive the formula for the propulsive force of a rocket engine (see Figure
Figure 10.8
10.8) when the jet outflows steadily from its nozzle, defining this force as the sum
of surface forces acting on the inner E and outer E 0 surfaces of the engine (neglect
the momentum of fuel components entering the combustion chamber). Obtain the
formula for the propulsive force in the case of the design nozzle (i.e., when the pressure
on its cut p' equals the ambient pressure po).
0 10.14 An ideal fluid moves in a fixed curved tube (see Figure 10.9). a) Calculate
Figure 10.9
the force R exerted by the fluid on the tube if the motion is steady and pi, pi, Vi and
P2, P2, t>2 are the characteristics of the flow in the input and output cross sections
with areas Si, S2 respectively. Examine the cases that R ^ 0. In particular, does a
reaction force act on the tube if pi = P2, Pi = P2, Vi = v%, but the directions of Vi and
« 2 differ? b) Determine the point of application for R. c) Calculate the force exerted
by a fireman to hold support the fire-hose barrel from which water gushes out. The
fire-hose barrel is a narrowing tube fitted on a hose conveying water. The diameters
of input and output sections equal 8 cm and 2 cm. The velocity of water at the output
of the fire-hose barrel equals 15 m/s. Neglect the viscosity in the calculations.
0 10.15 A tube with constant cross section is bent through right angle. A stream
of ideal incompressible fluid enters one of its sections and issues form the other. Does
a reaction force act on the tube? What is the magnitude of this force?
Q 10.16 A stream of an ideal incompressible fluid flows through a cylindrical tube
A A'
- W , Ik
P, P2
Figure 10.10
of cross section S with velocity V\ and pressure p\, then it passes abruptly into a
broader cylindrical tube of cross section 52 (see Figure 10.10). At a comparatively
small distance from the place of broadening, the velocity and pressure of the fluid
become uniform and take the values v-i < vi and P2 > pi. Supposing that the stream
flows according to the pattern of Figure 10.10 and the pressure equals px over all
section .4, calculate, using the momentum conservation law for the control volume
and the Bernoulli integral, the pressure drop p2— p2 which is caused by the abruptness
of broadening of the tube (p is the density of the fluid, p 2 is the pressure which would
be developed in the broader part of the tube if the tube broadens smoothly).
0 10.17 A body of volume V\ moves with constant velocity v in a vertical suffi

ciently long and broad tube of constant cross section filled with an ideal incompressible
heavy fluid. Behind the body, a joint gas bubble of volume V2 is formed. Find the
vertical projection of the force R acting on the body and the bubble from the fluid,
supposing the fluid to be at rest far from the body. Can it be more than the weight
of the body?
© 10.18 Find the total hydrodynamic moment acting on the rotating turbine
wheel relative to the axis of the turbine (see Figure 10.11). Suppose that the motion
of the fluid around the wheel is steady on the average, the fluid is ideal, incompress
ible, and weightless; internal angular momentum in the fluid as well as distributed
body and surface couples are zero.
11. Angular Momentum Equations. 77
blading of
the turbine
Figure 10.11
11 Angular Momentum Equations.

The angular momentum conservation law for an individual volume of a continuous
medium V bounded by a surface E has the form
f p(r xv + k)dV
dt
v
= frx FpdV + Ir xpnda+ f hpdV + IQnda
where r is the radius-vector of the particles of the medium, v is the velocity vector, p
is the density, pF is the volume density of body forces, pn is the stress vector, pk is
the volume density of internal angular momentum, ph. is the volume density of body
couples, Qn is the surface density of the surface couples (couple stress vector). The
index n in pn and Qn is used to emphasize their dependence on the orientation of
surface element. This law allows introduction of the couple stress tensor Q = Qêêj
such that, on any element with normal n at a given point, Qn = Q^nêj. The
equation of total angular momentum can be written in the differential form
p—(r xv + k) = prxF + ph + V ; ( r x p> + Q>) .

dt
(where Q3 = Q' J ej), if the functions forming it are sufficiently smooth. By use of the
equations of motion for the medium
pf-pF + Ve,
(where p = p'ê.e^ is the stress tensor), the equation of total angular momentum is
transformed into the equation of internal angular momentum
tile . . . .
p^^ph + V,-Q«e, - Y,(p'> - p " ) e < x e, .
at {<j
From a microscopic point of view internal angular momentum can be nonzero be
cause a continuum consists of particles which can have angular momentum even if
the macroscopic velocity of the medium is equal to zero (see Figure 11.1). Granulated
V 05
&
^ 0$
k=0 k*0
Figure 11.1
media, colloidal suspensions, mixtures with rotating particles, media with quantum
and gyromagnetic properties, magnetic fluids, liquid crystals and others can possess
internal angular momentum. To make the system of equations for a continuum com
plete, constitutive equations for k, Qij, h, pij - p1* are needed. As in the theory of
rigid bodies, the angular velocity i2 of internal rotation is often introduced (some
times formally) so that k = I • (2 where I is a given tensor of internal moments of
inertia per unit mass. For media with gyromagnetic properties, M = 7 • k where 7 is
the tensor of gyromagnetic coefficients, M is the magnetization of the medium. The
form of the density of body couples is determined by the nature of the studied phe
nomenon. For example, for magnetizable media in a magnetic field if, ph = M x H;
for polarizing media in an electric field ph= PxE, where P is the polarization, and
E is the electric field strength. The couple stress tensor Q can be bound both with
the tensor of gradients Vfi and, when the elasticity of internal rotation is taken into
account, with Vn(i) where n^j are the three orthonormal vectors forming the orthog
onal matrix of internal orientation. Nematic liquid crystals consisting of partially
ordered long molecules can be regarded as an example of a medium with internal
orientation (see Figure 11.2). In the case of nematic liquid crystals, the preferred
direction can be characterized by one vector of internal orientation n, which depends
upon time t and the given point of continuum, or a tensor nn if properties of the
liquid crystal are the same in directions n and —n. In the theory of elasticity with
I \ \ \ - /
\ I \
I ' I \
I nn I \
solid crystal nematic liquid cristal isotropic liquid
Figure 11.2
couple stresses, the tensor Q is supposed to depend upon the second derivatives of
the displacement vector W i o , e.g., upon V curl to. There are models of continua
for which the antisymmetric part of the stress tensor p is expressed in terms of the
difference fi — u> where u> is the vortex vector of the medium. Let us note that the
vector f2 — u does not depend upon the angular velocity of the reference system.
Therefore, it can be regarded as an independent variable on which the stress tensor
depends. This hypothesis does not contradict the axiom about the constitutive equa
tions which demands that they be independent of the choice of coordinate systems
that can move as a rigid body, either inertially or noninertially. In certain models the
stress tensor is assumed to be symmetric in spite of the presence of internal angular
momentum and distributed couples as well.
— PROBLEMS —
© 11.1 Using the angular momentum conservation law for an individual volume of
a continuum, show that couple stress tensor Q can be introduced so that Qn = Q?nj,
Qi — Qvei where Qn is the surface couple vector per unit area acting on the surface
element with normal n.
0 11.2 Derive the differential equations of angular momentum from the angular
momentum conservation law for an arbitrary individual volume of a continuum.
© 11.3 Consider a rigid body with a single fixed point. Obtain the Euler equations
for angular velocity components in a coordinate system moving with the body from
the angular momentum conservation law (suppose internal angular momentum to be
absent).
© 11.4 The theory of elasticity with couple stresses deals with models for which
distributed surface couples Qn are taken into account, and internal angular momen-
turn and distributed body couples are neglected. For these media, a) using the equa
tion of angular momentum, eliminate the antisymmetric part of the stress tensor from
the equations of motion; b) show that, for a uniform state of stress, when the couple
stress tensor does not depend upon spatial coordinates, the stress tensor is symmetric.
0 11.5 A ferric non-magnetized bar is freely suspended and is at rest initially.

Let it be located in a magnetic field H, where H is parallel to its symmetry axis, and
uniformly magnetized with magnetization M (all the time M \\ H, see Figure 11.3)
file pc2s5n3.pcx or another .xxx so that the bar begins to rotate around its axis with
//////
OK. 3
M
tI-
Figure 11.3
angular velocity u; (Einstein-de Haas gyromagnetic effect). The bar mass m and its
polar moment of inertia / are known. Using the angular momentum conservation
law, find the gyromagnetic ratio 7 = 7^/3 if the tensor of gyromagnetic coefficients
7 is spherical.
Problems 11.6-11.16 deal with the derivation of the constitutive equations for media
with internal angular momentum.
0 11.6 The Jaumann derivative of a tensor T is defined by the formula

DjT = (— T^w/t' — Tûêê,- where e< is the basis of Euler coordinate sys-
dt
tern, T^' are the components of the tensor T in the basis eit u>y = -(Vji/j — VjVi)
are the components of the vortex tensor. Show that £>jl = 0 where I is the tensor of
the barycentric moments of inertia of a rigid body.
© 11.7 If the tensor of the barycentric moments of inertia of a rigid body has two
identical and one zero eigenvalues, the body is called a rotator, a) What is a rotator
from the geometrical point of view? b) Derive the formula for the angular velocity fl
of a rotator orthogonal, by the definition, to its axis n regarding the rate of change
in orientation — as known. Vectors H, n, — form a dextral triad. Verify that J?
at at
does not change if n is replaced by —n.
0 11.8 Show that the angular momentum k of a rotator (see Problem 11.7) is
proportional to its angular velocity Q (if there is no internal angular momentum).
Prove that the coefficient of proportionality does not depend upon the time.
© 11.9 Using the theory of tensor functions, determine the general form of the
magnetization M of nematic liquid crystal with the orientation vector n, as a func
tion of the magnetic field H, the tensor nn and the metric tensor g. Suppose the
dependence on H to be linear and homogeneous. Find the product M x H.
Q 11.10 Determine the general form of the couple stress tensor Q depending
upon the tensor VI7 and the metric tensor g. Suppose the dependence of Q on V / 2
to be linear and homogeneous. Find the symmetric and antisymmetric parts of Q.
0 11.11 Determine the general form of the couple stress tensor Q as a function
of the tensors Vf7, nn (where n is the orientation vector of the medium) and the
metric tensor g. Suppose the dependence on V/2 to be linear and homogeneous.
What is simplified when V / ? is changed by V n (|n| = 1)?
© 11.12 For a tensor function Q(Vn, n, e, g) that is linear and homogeneous with
respect to the first argument, find the scalar "potential" £/(Vn, nn, g) such that
where ey'fc are the components of Levi-Civita tensor e and \n\ = 1.
0 11.13 Using the theory of tensor functions, determine the general form of the
antisymmetric part of the stress tensor p assuming that it depends on the vector
fi — u>, the tensor nn, the metric tensor g and the Levi-Civita tensor e. Suppose the
dependence on fl — w to be linear and homogeneous.
© 11.14 The orientation vector n (|n| = 1) is used in the theory of nematic

liquid crystals. It has only two independent components, which are determined from
the equation of internal angular momentum. Using the equations of state (see Prob
lem 11.7-11.9, 11.13) for magnetized liquid crystal
/ dn\ dl
k = IQ = I\nx — \ , — = 0 , ph = MxH ,
y at J at
M = aH + f3n{n ■ H) , Q = 0 ,
p« - p* = 2tiik{a9kl + (mkm)(nl - J) ,
a) deduce the equation of internal angular momentum; b) show that, if a + b — 0,

the projection of the equation of internal angular momentum on the direction n is
an identity; c) express the antisymmetric part of the stress tensor p in terms of the
dn
Jaumann derivative Djn = — w x n, where u> is the vortex vector, if a + b = 0.
at
© 11.15 The body force pF acting on a unit volume of a magnetizable body has
the form pF = -{MlVHi — WVMi). Find pF for a nematic crystal with orientation
n, supposing the dependence of the magnetization M upon the magnetic field H to
be of the form M = aH + (3n{H ■ n) where a and 0 are constants.
0 11.16 a) For a nematic liquid crystal with the orientation n, find the general
form of the symmetric viscous stress tensor r as a tensor function of the strain rate
tensor e, the tensor nn, and the metric tensor g. b) What form does this function
have if the dependence on e is linear and homogeneous?
© 11.17 Let qus consider a steady shear flow of a continuous medium between
two parallel plates, one of which moves with a fixed velocity and the other of which
is immovable. The components of the velocity of this flow in a Cartesian coordinate
system x< have the form
vi = 2ex2, v2 = v3 = 0, e = const
The medium is a nematic liquid crystal with the orientation n. The system is placed
in an external magnetic field, which induces a stationary, uniform, interior magnetic
field H. Using the equation of angular momentum and the equation of state
d/n.
M = aH + f3n{H • n) , Q = 0 , ft = n x — ,
at
P = - p g + 2/xe + - £ijk (<5f - n*m)(n' - w')eV

where e is the strain rate tensor, u> is the vortex vector of the flow, and p is the
pressure, find the possible stationary uniform orientation when a) H \\ u and b)
H _L u). Calculate the effective viscosity coefficient which is defined as the ratio
Pi 2 /2e 12 . Find the orientation vector n if anisotropy does not affect the stress value.
0 11.18 For a magnetic fluid, the magnetization is bound with the magnetic field
by the equation
dM _ ., 1 ..,
— = nxM--{M-XH)
at T
where x is t n e magnetic susceptibility coefficient, r is the time of relaxation. This
equation describes the evolution of the vector M near its equilibrium value xH- The
other equations of state have the form Q = 0, p = —pg + 2fie + — ey/t (fifc — w*)e'eJ'
where p is the pressure, e is the strain rate tensor, and w is the vortex vector. Let
there be a steady flow of this magnetic fluid in homogeneous and constant magnetic
field H. The velocity components of this flow in a Cartesian coordinate system have
the form vx = 2ey, vy = vz = 0, e = const. Find the stationary uniform distribution of
internal rotation vector fi and that of magnetization vector M supposing |W|T < 1.
Calculate the effective viscosity coefficient pn/2ei2.
Chapter 3
Thermodynamics of Continua
12 B a s i c Laws a n d C o n c e p t s of T h e r m o d y n a m i c s
The first law of thermodynamics.
The first law of thermodynamics (energy conservation law) states that, for any ther-
modynamic system, there exists a single-valued function of the state of the system -
the energy £ - such that, for any elementary process in the system, the equality
d£ = dA{e) + dQ{t) + dQ" (12.1)
is valid. Here d£ is the differential of £, but dA(e), d<2(e), dQ" are not differentials
of any functions. The latter are infinitesimal quantities denoting the work of the
external forces acting on the system, the heat input to the system from the outside,
and the other possible inputs of energy from the outside, respectively. The variation
of the kinetic energy £^n of the system is described by the equality (kinetic energy
theorem)
d4i„ = dA{e) + dA(i) (12.2)
where dA^ is the work of the internal forces. Using (12.1) and (12.2), one can obtain
the equation of internal energy (wich is called also the equation of heat flow)
dU = -dA{,) + dQM + dQ" (12.3)
where, by definition, U = £ — £km is the internal energy. For an individual volume
V of a continuum bounded by the surface E
2
4™ = Jp\dV■, U = JpudV,
V V
(e)
dA = J{F-v)pdtdV+ f(pn-v)dtd(j ,
85
86 THERMODYNAMICS OF CONTINVA
dQw = jpdqatmdV- Jqndtda ,

V £
dQ" = jpdq'^dV- Jq'n'dtda .
where u is the internal energy per unit mass, dqmass, dq^^, q„, q" are the densities of
body and surface inputs of energy as heat and other forms different from the work of
macroscopic forces. Further dQ" will be reffered to as "additional" energy input (i.e.,
additional to heat and work of macroscopic forces, see examples in Problem 13.15 and
in Section 16). The first law of thermodynamics (12.1) is written in the form
It Ip (T + u) dV = I{F'v)pdV + (Pn v) d(J +

/ '
V ^ ' V £
(124
/T"-/'**/^'"-/'*' »
V V V T,
Using (12.4), one can obtain the differential equation of energy valid in a domain
where the motion is described by smooth functions
it ( T + u ) = {F v)+
' -^tf*vd+ îT ~ p~Vkq'k (12 5)
-
where dq^^ = dqmass + dq^^, q'k = qk + q"k; qk are the components of the heat
k
flux vector q, q" are the components of "additional" energy flux vector q". The
differential form of the kinetic energy equation (12.2) is
~ = (F ■ v) + -V,(p%) - V v , i ; * , (12.6)
at I p p
The differential form of the equation of internal energy (12.3) follows from (12.5),
(12.6) and has the form
du 1 t,„ do dq"
n = 7v'Vk + it + ir (12 7)
-
where dq = dgmass — V*?* dt/p is the mass density of heat input (i.e., heat input per
unit mass), dq" = dq^^ - Vkq"k dt/p. The quantity
- -pHVtvk (12.8)
is the work of internal surface forces dA^ per unit mass and per unit time (see equality
(12.2) and Problem 9.21). The ratio of amount of heat transmitted to unit mass to
12. Basic Laws and Concepts of Thermodynamics 87
the corresponding change in temperature during a process is called the specific heat
capacity (or, often, simply specific heat) and denoted by c, so c = dq/dT. The heat
capacity depends not only on the material, but also on the process taking place in it.
The Fourier heat conduction law states
q=-/cgradT (12.9)
where n is the thermal conductivity coefficient and T is the temperature. For
anisotropic media, the Fourier law has the form
where Kmk are the components of thermal conductivity tensor. A process is referred
to as adiabatic if, for all particles, the heat input equals zero, i.e., dq = 0. A process
is referred to as isothermal if the temperature of all particles does not vary, i.e.,
dT/dt = 0.
T h e Second law of thermodynamics.

The second law of thermodynamics (entropy variation law) states that, for any ther-
modynamic system, there exists a single-valued function of the state of the system -
entropy S - such that, for any elementary process in the system, the relationships
dS = deS + diS, d{S>0 (12.11)

are valid. Here, dS is the differential of S, but deS and diS are not differentials of any
functions; deS is the entropy supply from the outside, a\S is the entropy production in
the system taking place for irreversible processes. For reversible processes, d,S = 0.
A reversible process can be defined as a process for which an inverse process can
be carried out and for all parts of the original and inverse processes the amounts of
energy received by the system and its parts from the outside differs only by signs.
Reversible processes are described by the equations which are invariant relative to a
change of time t by —t. The entropy supply from the outside deS can be caused only
by the heat or mass inputs. If there is no mass exchange between the system and the
surrounding medium, and the temperatures of all parts of the system are identical,
then
<45 = dQ/T
For reversible processes in such system the second law of thermodynamics is
dS=^- (12.12)
Temperature of a medium usually varies spatially; heat input to an individual volume

V can occur either through its boundary E {—qndtdo to the surface element) or
88 THERMODYNAMICS OF CONTINUA
directly to the interior points {pdqmamdV to the mass element). So the corresponding
entropy supply is
deS = -Jqfdtdo + J^pdV.
£ V
For an individual volume V of a continuum the second law of thermodynamics is
written in the form (if there is no mass diffusion)
V E V V
where s is the mass density of entropy, d\s/dt is the mass density of entropy production
rate. Applying (12.13) to continuous motions, one can obtain the differential entropy
equation
(12 14)
i - N ^ H ^ * ' £*° -
that describes variation of entropy if d\s/dt is given. Entropy equation (12.14) can
be written in the form
Tds = dq + dq' (12.15)
where dq = dqmBSS — Vjt<7fc dt/p, dq1 = Td\s + qkV/tT dt/pT. The quantity dqi is called
the uncompensated heat. If heat conduction does not depend upon other irreversible
processes, the value of dq'/T equals the density of entropy production at the expense
of all irreversible processes except heat conduction and, according to the second law
of thermodynamics, is not negative. For reversible processes, dq1 = 0. Thus, for the
models of media in which mechanical processes are treated as reversible (e.g.,ideal
(inviscid) liquids and gases, elastic solids), the equation representing the second law
of thermodynamics is written in the form
Tds = dq (12.16)
if there are no other (in addition to heat conduction) irreversible processes (irreversible
chemical reactions, electrical conduction, etc.).
Dissipation inequality and Gibbs identity

Combining the equation (12.16) and the equation of internal energy (12.7), one can
obtain the relationship valid for reversible processes
du = - pkmVkvm dt + Tds + dq" (12.17)
that, if, in addition, dq" = 0, relates only the internal parameters of the medium
for any process; the quantities describing external actions - work of external forces
and heat input - are excluded with the use of the kinetic energy theorem and second
law of thermodynamics. In the presence of irreversible processes, the first and second
laws of thermodynamics lead to an inequality instead of the equality (12.17). Namely,
(12.7) and (12.14) yield
* _ i , ^ + «: + r£-±*w- r £. (.,.«,
According to the second law of thermodynamics, d,s > 0; consequently,
du - -pkmVkvm dt -dq"-Tds + -±-qkVkTdt < 0 . (12.19)

P PT
The inequality (12.19) is called Clausius inequality (dissipation inequality). As an
example of medium with reversible mechanical processes, consider an ideal (inviscid)
fluid. In an ideal fluid
pkm = -pgkm
by definition, where p is the pressure. The relationship (12.17) takes the form (if
dq" = 0)
du=2=dp + Tds. (12.20)
P2
For viscous liquids and gases,
pmk = -pgmk + T"1*
where Tmk = rkm are viscous stresses, which are functions of strain rate tensor com
ponents e ^ , vanishing at ekm = 0. If dq" = 0, u = u(p, s), p = p(p, s), T = T(p, s),
then the dissipation inequality is written in the form
(5-7)*+(5-r)*-r^-*+!^*so- (1221)
Considering processes for which ekm are small, V*T = 0 and taking into account that
u, p, T do not depend upon dp, ds and, besides, Tlk are small at small ekm, one
obtains that the inequality (12.21) can be satisfied only if
p_ = du(p,s) T = dujp, s) (12.22)
p2 dp ds
In fact, the equalities (12.22) are valid for all processes, i.e., not only if ekm are small
and VfcT = 0, since they relate quantities that are , by definition, independent of
the values of ekm and VkT. With the relationships (12.22) taken into account, the
dissipation inequality can be reduced to the form
- ^ * V f c T + Tm*e*m>0, (12.23)
and the equation (12.18) yields
Formula (12.24) determines the expression for the entropy production rate in viscous
heat-conducting liquids and gases at the expense of irreversible processes caused by
viscosity and heat conductivity. The equalities (12.22) can be written as the relation
ship
du = ^dp + Tds , (12.25)
P2
thus valid not only for ideal (see Formula (12.20)), but also for viscous liquids and
gases and called the Gibbs identity. The Gibbs identity states that, even in the
presence of irreversible processes, the relationships between the thermodynamic pa
rameters that do not depend explicitly upon parameters describing irreversibility,
remain the same as for reversible processes. The corresponding forms of the Gibbs
identity can be written also for other models of continua.
The Onsager principle

For models of continua with irreversible processes, the expression for dis/dt, as a
generalization of (12.24), is usually written in the form
^ = £*"*», (12-26)
d t
n
where Xn are so called the "thermodynamic forces" and xn are "thermodynamic

fluxes." Since, according to the second law of thermodynamics, diS is not nega
tive, the thermodynamic forces and fluxes cannot be independent. The relationships
relating thermodynamic forces and fluxes can be represented, for many cases, by the
linear functions
Xn = LnmXm ^ (12.27)
The relationships (12.27) are called the kinetic relationships, and the coefficients
L nm are referred to as kinetic coefficients. The Fourier heat conduction law, (12.9)
or (12.10), is an example of kinetic relationships. The Onsager principle — the
principle of symmetry of the kinetic coefficients — states that, if there are no axial
vectors among the arguments of Lnm, then
rnm _ j-mn
If the dissipative processes are determined by axial vectors (i.e., by the magnetic field
intensity H), the preceding equalities are changed by the relationships Lnm(H) =
Lmn(-H).
Free energy, enthalpy, Gibbs thermodynamic potential.

To describe thermodynamic properties of continua, besides the internal energy u,
temperature T and entropy s, other functions of state, in particular, specific (i.e.,
per unit mass) free energy T = u — Ts, enthalpy i = u + p/p for liquids and gases,
i = u — pmk£km/p for deformable solids, Gibbs thermodynamic potential $ = i — Ts
are used as well.
About mathematical models of continuum media.

The laws of thermodynamics are used in two ways. First, they are necessary in
calculation of specific processes. The equation of energy is contained in the complete
system of equations describing motion of a medium, and the second law determines
the set of possible processes (see, e.g., Problems 13.1-14.16 of this chapter as well as
problems of other chapters). In the second, the laws of thermodynamics are used to
define new models of continua. To establish a model of a continuum means to establish
a complete (closed) system of equations which allows a motion or an equilibrium of
the continuum to be described. A complete system of equations for any medium
consists a) of universal equations (valid for all media) representing the laws of mass,
momentum, angular momentum, energy conservation and entropy variation and b)
of equations determining the properties of the specific medium. These last equations
are referred to as determining or constitutive. A constitutive equation can relate
equilibrium parameters of the state; then it is called an equation of state (e.g.,
the Clapeyron equation p = RpT for a perfect gas). It can also relate parameters
describing a process; then it is often called a kinetic relationship (e.g., the relations
between viscous stresses and strain rates in a viscous fluid). To write constitutive
equations, one uses experimental data on the behavior of the considered specific
medium and accepts certain hypotheses about the form of the equations. These
hypotheses must necessarily be consistent with the laws of thermodynamics. Use
of the relationships between parameters obtained from the general laws decreases
the number of the hypotheses and experiments necessary to construct a model (see
Problems of Section 15).
Constitutive relationships for some models of continua.

Here a brief summary of relationships defining some classical models of continua is
given. These relationships are not a minimal system of equations necessary to set a
model of a continuum (see Problems 15.7-15.9, 15.13). 1. An ideal incompressible
fluid:
Pt3 = -pg'3, <V = 0 , ^ = 0, u = u(T), (12.28)
where the internal energy density u(T) is a given function of temperature T. 2. An

ideal compressible fluid:
p^ = -pgij, dq' = 0, p = p(p,T), u = u(p,T), (12.29)
where the pressure p(p, T) and internal energy density u(p, T) are given functions. 3.
A viscous incompressible fluid:
jP = -pgF+i*, ^ = 0, dq' = I T «eydt , (12.30)

at p
ri} = T « ( e H l T ) , u = u{T) , ey = ^ ( V ^ + V^,) ,
where T*i(eki, T) and u(T) are given functions; T,J' are components of a viscous stress
tensor. 4. A viscous compressible fluid:
p« = -p f f « + r i j , dq' = -Tijeij (ft; (12.31)
p = p(p,T)1 u = u(p,T), rij = Tij(ekl,T),

where p{p,T), u(p,T), and Ty'(efc»,T) are given functions. In particular, a fluid is
referred to as linearly viscous (or Newtonian) if
T« = A « " ( 7 > „
In an isotropic linearly viscous fluid,
Tij = Adivv g{j + 2petj (12.32)
(Navier-Stokes law) where A, p. are the viscosity coefficients, A = A(T), p. = p(T). An

ideal or viscous gas is referred to as perfect if
p = RpT, u = cvT + const, cv = const, R = const = Ro/m (12.33)
where Ro is the gas constant, m is the molecular weight of the gas. 5. A thermoelastic
medium:
pii=pii(e»,T), u = u(eu,T), dq' = 0. (12.34)
Here eki are the components of the strain tensor; p^(£kii T) and u(£ki,T) are given
functions. For an isotropic linearly elastic medium with small strains,
Pa = AJi(e)0y + 2p£ij - a(3A + 2p){T - T0)9ij (12.35)

12. The First Law of Thermodynamics 93
(Hooke's law with thermal stresses taken into account), and
u = —Jl + £ J2(e) + -(3A + 2/i) J,T 0 + C( ~ o ) + c(T - T0) + const (12.36)

Zp p p 11 o
,j l
where J\{e) = g £ij, Ji{e) = £ ^£ij, A, p. are the elastic coefficients which are called
also the Lame coefficients; a is the coefficient of linear thermal expansion, c is the
specific heat at constant strains. For all the enumerated models, the stress tensor is
assumed to be symmetrical. Thermodynamics of media with internal rotation (and
with the asymmetrical stress tensor) is considered separately in Section 16. In all
problems in this chapter, processes are considered for which the additional energy
input dq" is equal to zero unless otherwise stipulated.
13 The First Law of Thermodynamics.

Energy Equation. Perfect Gas.
— PROBLEMS —
0 13.1 Prove with the use of the first law of thermodynamics (12.4) that, for a
process of heat conductivity in a continuum, there exists a heat flux vector q such
that
qn = qn
where qn is the limit of the ratio of the amount of heat flowing per unit time through
the area element with the normal n to the area of this element when the area tends
to zero and the element shrinks to a point.
0 13.2 Derive the differential equation of energy (12.5) from the first law of
thermodynamics for a finite volume of a continuum (12.4).
© 13.3 Using the constitutive relationships (12.28) - (12.32) write the expression
for the density of the work of internal surface forces, i.e., for the quantity - p,jVjVi dt/p
(see the definition (12.8)) a) for an incompressible ideal fluid ; b) for a compressible
ideal fluid; c) for an incompressible viscous fluid, in particular, for an isotropic linearly
viscous one; d) for a compressible isotropic linearly viscous fluid.
© 13.4 Show that, for an elastic medium (and for any other medium in which
the stress tensor is symmetrical), the density of the work of internal forces (see the
definition (12.8)) can be expressed in the form - p*^diij/p where pYj, ev- are the
components of the stress and strain tensors in a concomitant Lagrangian coordinate
system.
0 13.5 Write the energy equation and the equation of internal energy a) for an
ideal incompressible fluid, b) for an ideal compressible fluid, c) for a viscous fluid, d)
for an elastic medium. Use the constitutive relationships (12.28) - (12.35).
0 13.6 Using the equation of internal energy, show that, for an ideal fluid,
du(p, T)/dT is the specific heat in a process with unvarying volume.
© 13.7 Show that, for an elastic medium, du(eij,T)/dT is the specific heat in a
process with unvarying strain.
0 13.8 Find the relationship between the specific heats at constant pressure (cp)
and at constant volume (cv) for an ideal perfect gas (see formulae (12.29), (12.33)).
© 13.9 Show that, in the general case, the specific heat of a viscous perfect gas in
a motion with unvarying volume of the particle does not equal c v . For what process
does the specific heat equal c v ?
© 13.10 Show that, for processes with constant pressure in an ideal fluid, the
increment of specific enthalpy i is equal to the heat input per unit mass from the
outside, and the specific heat Cp equals di(p,T)/dT.
0 13.11 Show that, for adiabatic motion of an ideal fluid, the temperature and
pressure of a particle depend upon only its density. Is this valid for a viscous fluid?
0 13.12 a) Find the relationship between temperature and density as well as be
tween pressure and density (the equation of "Poisson adiabat") for adiabatic motions
of an ideal perfect gas. b) Are the relationships of item a) valid for adiabatic motions
of a viscous perfect gas?
0 13.13 a) Find the relationship between the change in temperature and the
value of the first invariant of the strain tensor for adiabatic deformation of a linearly
thermoelastic isotropic medium with small strains (see formulae (12.35), (12.36)).
b) Compare the relationships binding the stresses and strains for adiabatic and
isothermal processes in this medium.
© 13.14 There is an extended homogeneous rod made of a linearly thermoelastic
isotropic material. Consider the two processes of heating the rod: 1) when the rod
is placed in a rigid shell; 2) when a shell is absent, and constant extending stresses p
are applied to the ends of the rod. Calculate the specific heats cE and Cp for the first
and second processes respectively. Explain the independence of Cp upon the values of
Pij-
0 13.15 Find the relation between the temperature T and magnetization M in

the process of adiabatic magnetization of a medium for which magnetization vector
M is proportional to the magnetic field intensity H: M = xH- Assume that the

medium does not deform, the internal energy does not depend explicitly upon the
magnetization at constant temperature, the specific heat at constant magnetization
cM is constant, the susceptibility x obeys the law of Curie x = A/T where A = const,
and the energy input to the medium from the magnetic field at magnetization is
given by the formula pdq" = H ■ dM. Note that in this Problem an "additional"
energy input dq" (i.e., different from the heat and the work of macroscopic forces) is
essential.
0 13.16 Small disturbances of density and pressure in an ideal gas are known
to propagate over the gas with the speed a = Jdp/dp. Sound propagation can be
regarded as an adiabatic process of propagation of small disturbances. Show that the
speed of sound in an ideal perfect gas depends upon only the temperature.
0 13.17 A uniform perfect gas of mass M performs the following closed quasi-
static process, called the Carnot cycle: first (I) isothermal expansion at temperature
7\ with change in density from p = p\ to p — pi < p\, then (II) adiabatic expansion
to p = p3 < p2, T = T2 < T\, then (III) isothermal compression at temperature T2
to p = p4 > p3, at last (IV) adiabatic compression to p = pi, T = 7\. Calculate the
change in the gas energy, the value of the work done by the gas and the quantity of
heat received by the gas during each of the four parts of the process and during the
whole process. Show that pi/p2 = P\lP3- Calculate the efficiency of the cycle defined
as the ratio of the work Aga3 done by the gas during the whole cycle to the quantity
of heat transmitted to the gas.
© 13.18 Using Fourier's law, obtain the formula for the density of heat input dq
due to heat conduction at the condition a) K = const b) K = K(T).
0 13.19 Write the energy equation and the equation of internal energy for the
following heat-conducting media (heat conduction obeys Fourier's law, K = const):
a) an ideal incompressible fluid, b) a linearly viscous perfect gas, c) a linearly ther-
moelastic solid. For what cases can the energy equation be reduced to the classical
equation of heat conduction dT/dt = x AT?
0 13.20 Derive, from the energy equation, the Bernoulli integral along a stream
line for a steady adiabatic motion of an ideal fluid in a potential field of body forces,
i.e., the relationship
v2 , p
\-i — U = const, i = u-\—
2 p
along a streamline, where i is the specific enthalpy, and U is the potential of the body
forces.
0 13.21 Find the stationary temperature distribution in a homogeneous static

layer of a continuum with constant thermal conductivity placed between two infinite
parallel plates with temperatures 7\ = const and T2 = const, respectively. The width
of the layer equals h.
0 13.22 Consider a Couette flow — a steady flow of a linearly viscous fluid be
tween two infinite parallel plates, one at rest and the other moving with a constant
speed VQ. The distance between the plates equals h, and the pressure p = const. The
fluid velocities at the plates are equal to the velocities of the plates (see Section 19 for
the boundary conditions for viscous fluid). Let the fluid be heat conducting, the co
efficients of viscosity and heat conductivity being constant. The temperatures of the
plates are constant and equal to T\ and T2 > 7\, respectively. Find the temperature
distribution in the layer of the fluid. What is the maximal temperature in the fluid
equal to? Evaluate the calculations at 7\ =T2 = 30°C, Vo = 10 m/s, ft = 1 cm for a)
water (/x « 0.01 g/cm • s, K « 6 x 104 g • cm/s 3 • K), b) lubricating oil (n « 4 g/cm • s,
K « 1.4 x 104 g • cm/s 3 • K). Compare with similar flow of a heat conducting, but
inviscid fluid.
0 13.23 Consider Problem 13.22 with changed conditions for the plates: one of
them is made of a heat insulating material, i.e., instead of the condition T = T2, the
condition that the heat flux through this plate equals zero is imposed. What is the
temperature of the fluid at the heat insulating plate? Perform the calculation with
the data from items a) and b) of Problem 13.22. c) Compute the temperature of the
fluid at the heat insulating plate if the speed of the plate is 8 km/s, the fluid is air,
the temperature of the second plate is 30°C (take fi/n = 7 x 10 - 8 s2 • K/cm 2 for air).
d) Is a stationary temperature distribution possible if both plates are made of a heat
insulating material? e) Explain why the solution of this problem gives the answer to
the question, "To what extent do indications of a lamellate thermometer (on a thin
plate set along the stream) differ from the value T in the stream in the absence of
the thermometer (assuming that the flow is laminar)?".
© 13.24 Consider a half-space filled with a homogeneous continuum at rest and

at temperature T0. At the instant t = 0, the temperature all over the boundary of
the half-space abruptly becomes, and is subsequently maintained to be, equal to Ti ^
T0. Determine the distribution of temperature in the half-space at the subsequent
instants, neglecting the variation of density caused by variation of temperature.
0 13.25 Which takes more time: to fry one frying-pan of thick cutlets or, instead,
two frying-pans of cutlets half as thick? Determine the answer, starting from the
condition that the cutlets laid on the frying-pan can be modelled as an infinite layer,
the lower boundary of which is maintained at the temperature of the stove 7\, and
the upper boundary is heat insulated. As the time of frying, take the time for which
the temperature at the upper boundary reaches the value T2. The initial temperature
equals T0.
0 13.26 Does the internal energy of air contained in a room increase if the tem
perature rises from T\ to T2 > Ti, and the pressure does not change? Assume that
air is a perfect gas.
0 13.27 Determine the length of a gun barrel necessary to accelerate the bullet
up to the maximal speed as a function of the initial state parameters of the gunpowder
burnout products (with the pressure ahead of the bullet taken into account). Neglect
the inertia of the gas and mechanisms of dissipation. Relate the initial pressure
with the combustion heat of gunpowder. Perform the calculation, assuming that
gunpowder has the density 1.6 g/cm 3 and the combustion heat 3 MJ/kg, the length
of the charge is 5 cm, the diameter of the barrel is 1 cm, 7 = Cp/cv = 3 for the
gunpowder burnout products regarded as a perfect gas, the pressure ahead of the
bullet equals 1 atm.
0 13.28 What is the mass of air contained in a room with size 20 x 20 x 3 m at

standard conditions (the temperature is 20°C, the pressure is 1 atm)? Assume air to
be a perfect gas (R = 287.042 m 2 /s 2 • K).
0 13.29 Vapor is admitted into the cylinder of an engine at constant pressure

20 atm. The stroke of the piston is 60 cm, and its diameter is 20 cm. How much
work (in joules) does the vapor perform for one cycle of the piston?
0 13.30 A stone of mass 5 kg having the temperature 200°C is immersed into a

vessel containing 9 kg of water at 5°C. The specific heat of the stone is 0.20 cal/g • K.
Find the final temperature, neglecting heat loss through the walls of the vessel.
Q 13.31 A perfect gas performs a Carnot cycle (see Problem 13.17) during which
the isothermal expansion occurs at T = 600 K, and the isothermal compression, at
T = 300 K. At what temperature must the expansion occur to double the work
performed for the cycle? Assume that the changes in volume and the temperature
at which the isothermal compression happens remain former. The sum of works
performed for adiabatic parts of the cycle in both cases equals zero.
© 13.32 Since there is surface tension, the work AA = uAE, where a is the
surface tension coefficient, must be performed to increase the area of a surface of a
liquid by the quantity AS. How much work must be expended to divide 11. of water
into droplets of diameter 10~2 mm (take a = 73 dyn/cm)? By how many degrees
can the temperature of water be raised via stirring at the expense of this quantity of
work (neglect heat loss into the surrounding medium)? Through what height can 11.
of water be lifted at the expense of this amount of work?
14 The Second Law of Thermodynamics. En

tropy. Gibbs Identity.
— PROBLEMS —
© 14.1 Calculate the change in entropy for each of the four parts of a reversible
Carnot cycle in a perfect gas (see Problem 13.17)
© 14.2 Using the Gibbs identity (the relation (12.25)), obtain the expression for
the density of entropy s of a perfect gas as a function a) of the temperature T and
density p, b) of the pressure p and density p, c) of the pressure p and temperature T,
d) of the internal energy u and density p.
© 14.3 Do the functions determining the dependence of the entropy of a viscous
gas upon the parameters of state differ from the functions determining that of the
entropy of the ideal gas corresponding to the considered viscous gas, i.e., to the gas
with the same functions p(p,T), u(p,T)l
© 14.4 Using the first and second law of thermodynamics, determine the expres
sions for the entropy and free energy densities s, T for an isotropic linear thermoelastic
solid as functions of T, ey. Determine the expresion for internal energy density u as
a function of s and e^. Regard the functions p,J(e/t;,T) and u(e«,T) as known (see
formulae (12.35), (12.36)
© 14.5 Write, for a perfect gas, expressions for the densities of internal energy u,
free energy !F, enthalpy i, and Gibbs thermodynamic potential \P as functions of a)
p, T; b) p, s; c) p, p; d) p, s; e) p, T.
© 14.6 Show that, for reversible quasi-static isothermal processes, the work of
external forces applied to the medium equals the increase in free energy of the medium.
© 14.7 Show that the entropy of each particle of an incompressible viscous fluid
moving adiabatically in general case increases, but it remains constant if the motion
is isothermal.
© 14.8 Show that, during isothermal compression of some amount of a perfect
gas, the entropy decreases for each particle of the gas as well as the whole amount.
© 14.9 Show that the temperature of the particles remains constant during an
adiabatic motion of an ideal incompressible fluid, but in general increases during that
of a viscous incompressible fluid.
14. The Second Law of Thermodynamics 99
0 14.10 Consider a stationary heat conduction process in a layer of a resting

medium (see Problem 13.21). How does the entropy of each particle and that of the
whole layer behave? What are the density of entropy production and the fluxes of
entropy and heat through the boundaries of the layer?
0 14.11 Answer the questions of Problem 14.10 for the flow of a viscous heat
conducting fluid considered in Problem 13.22. Calculate numerical values of des/dt
at conditions given in Problem 13.22.
0 14.12 A thermos is filled with a mixture of ice and water. Since the insulation
of the thermos is not ideal, the ice melts gradually. However, the melting comes
about slowly, and the temperature in the thermos remains practically constant and
equal to 0°C. Calculate the change in entropy corresponding to the melting of 500 g
of ice. The heat of melting for ice equals 79.67 cal/g. How does the entropy of 500 g
of water change during its transformation into ice at the temperature 0°C?
© 14.13 Calculate the change in entropy of 500 g of water during its evaporation
(at boiling) at 100°C. The heat of evaporation at this temperature equals 539 cal/g.
0 14.14 Calculate the change in entropy of an elastic steel rod of length 1 m
and cross-sectional area 1 cm2 a) during its isothermal extension up to 1.001 m at
the temperature 15CC, b) during its adiabatic extension up to the same final length.
What is, for case (b), the final temperature of the rod? What are the values of
the extending force for cases (a) and (b)? Assume that, for steel, Young's modulus
E = 21.6 x 101Q N/m 2 , Poisson's ratio u = 0.25, the specific heat at constant strain
c = 0.46 kJ/kg • K, the linear thermal expansion coefficient a = 12 x 10~6 1/K,
and p = 7700 kg/m 3 . Coefficients E and a are related to A and p. by the formulae
E = /*(3A + 2/z)/(A + /j.),a = A/2(A + p.).
© 14.15 Consider a rigid cylinder with heat insulating wall divided in two equal
volumes A and B by a partition. Volume A is filled with a perfect gas of quantity
1 mole at pressure po and temperature T0. The volume B is evacuated to high vacuum
so that the quantity of the residual gas and its pressure can be neglected. Then, the
partition is destroyed, and the gas begins to move inside the cylinder. After some
time, the motion of the gas attenuates so that the gas fills the whole cylinder and
rests. What are the final pressure and temperature of the gas? What are the changes
in its energy and entropy?
© 14.16 Let there be an infinite cylindrical tube filled with an ideal perfect gas at
rest with the density po and pressure po- A partition (piston) placed in a cross section,
at the instant t = 0, abruptly begins to move along the tube with constant speed v0.
Show that a formal solution of the problem about the motion of the gas, valid both
for region (I) where the piston moves in and for region (II) where it moves out, can
be the following. Domains where the gas has constant speed, density and pressure
adjoin the piston. These domains, by discontinuity surfaces (where jumps of the
parameters of the gas take place), are connected with domains where the gas rests as
before. Find this solution if p 0 = 1 atm = 1.013 x 105 N/m 2 , p0 = 1.25 x 10~3 g/cm 3 ,
cv = 0.169 kal/(g • K), 7 = 1.4, v0 = 100 m/s. Using the fact that passing through
a jump is an irreversible process, show that this solution for region (I) does not
contradict the second law of thermodynamics, but for region (II) it does. See the
jump conditions for an ideal gas in Problems 17.10, 17.11.
15 Restrictions Imposed by The Laws of Ther

modynamics on the Form of Constitutive Re
lationships.
— PROBLEMS —
0 15.1 Show that, for compressible fluids, the internal energy density u and pres
sure p, to be functions of density p and temperature T (u = u(p, T), p = p(p, T)) are
bound by the relationship
(du\ =V__T_ (dp\

\dp)T ? ? \dTJp
© 15.2 For an elastic medium, prove the formula
(a—a,
where the internal energy density u = «(£**, T), a'k = p'k/p, p'k are the components
of the stress tensor, and eifc are the components of the strain tensor in a concomitant
Lagrangian coordinate system.
© 15.3 Show that, if the Clapeyron equation p = RpT is valid for an ideal gas, the
density of internal energy u and specific heat capacity at constant volume cv of this
gas are the functions only of the temperature T and that to determine the expression
for u, it is sufficient to specify cv(T).
Q 15.4 Show that the conclusions of Problem 15.3 are valid for any ideal gas
whose equation of state p = p(p, T) has the form p = f(p)T.
15. Restrictions on the Form of Constitutive Relationships 101
0 15.5 Show that, if the density of internal energy of a gas is a function only of
the temperature: u = u(T), the equation of state of the gas has the form p = Tf(p)
where f(p) is an arbitrary function of p.
© 15.6 Assuming the specific heat capacity cv to be a known function of the
temperature, determine the expression for the densities of internal energy and entropy
of an ideal gas obeying the Van der Waals equation p = RpT/(l — bp) — ap2 where
a = const, 6 = const.
© 15.7 The definition of a model of a compressible fluid demonstrates that, in
general, instead of p, T, any of the following pairs p, s; p, s; p, T can be taken as
parameters determining the state of a particle of the fluid. Show that, to establish a
specific model, it is sufficient to set only one of the quantities u, T, i, V as a function
of the corresponding parameters (T, i, \P are specific free energy, enthalpy, Gibbs
potential, respectively; see Section 12). Namely, for fluids, show that the functions
u(p,s), T(p,T), i(p,s), *(p, T) are thermodynamic potentials, i.e., the following
equalities are valid a) p/p2 = (du/dp)„ T = (du/ds),,; b) p/p2 = (dF/dpfr, s =
-(d?/&T)p; c) \/p = (di/dp)„ T = (di/ds)p; d) \/p = (99/dp)Tt s = -(W/ffT),.
0 15.8 Show that the model of a perfect gas ordinarily introduced by the two
relationships
u = cvT, p = RpT (15.1)
is completely specified by the only one function
u(p,8) = vo(£\ e^ (15.2)
i.e., show that the relationships (15.1) follow from (15.2).

© 15.9 Show that, if, for an elastic medium with finite strains, ey and s are taken
as parameters determining its state, then a) the formulae for the components of the
stress tensor py in a concomitant Lagrangian coordinate system and the temperature
o o
_y 9u(fly,£y,s) du(gy,£y,s)
P =P T=
*£ ' *
are valid (<? y are components of Lagragian metric tensor in the undeformed state)
b) the components p y in an Eulerian coordinate system and the temperature T are
given for isotropic elastic media by the formulae (of Murnaghan)
p _p(«. fc + 2 e .t)—^- , T- - .
This means that, to set a specific model of an elastic medium, it is sufficient to set
one function u(</y, ey, s).
© 15.10 Show that, for an anisotropic linearly thermoelastic medium with p ,J =

AijklSki + B{i(T - T0), Ski are small, the maximal number of independent coefficients
of elasticity Alikl equals 21.
© 15.11 Show that the viscosity coefficients of a linearly viscous, isotropic fluid
must satisfy the inequalities A > — 2/z/3, fi > 0.
0 15.12 Write the dissipation inequality (the Clausius inequality) for fluids so
that, instead of the internal energy, a) the free energy T, b) the enthalpy i or c) the
Gibbs thermodynamic potential * is a part of it.
© 15.13 Using the dissipation inequality (the Clausius inequality), show that,
from the assumptions that the state of a particle of a viscous heat conducting fluid
is determined by the quantities s, p, ekm, qk, and the pressure p and temperature T
do not depend explicitly on the derivatives dp/dt, ds/dt, it follows that the density
of internal energy of the fluid does not depend on the parameters ekm, qk, and the
equalities
are valid.
© 15.14 As a result of processing the experimental data for a medium (liquid
with bubbles of vapor), it is determined that the state of the medium is specified by
the temperature, density and variation of density in time; also, the viscosity can be
neglected, and the relationship p = po + [icPp/dt2 where p. = const is valid for the
pressure. Supposing the deformation processes of this medium to be reversible, show
that a) the functions u and s do not depend upon d?p/dt2, b) the function u depends
upon dp/dt quadratically.
© 15.15 Show for the media in which p'j = —pg'j, p and the other parameters
of the state of the medium depend upon the temperature, the density and variation
of density in time that a) if p depends upon the first derivative of p with respect to
time and does not depend upon the higher derivative, then dynamical deformation
processes are necessarily accompanied by dissipation (i.e., dq1 > 0); b) if the free
energy depends upon the derivative of p with respect to time up to nth order, and
the deformation processes are not dissipative, then the pressure must depend upon
the derivatives of p with respect to t up to (n + l)th order, and this dependence upon
dn+lp/dtn+1 must be linear.
© 15.16 As a generalization of a model of an elastic medium defined more accu

rately, the model can be considered for which the internal energy and stresses depend
15. Restrictions on the Form of Constitutive Relationships 103
upon not only Ey and s, but also upon Vêy (dependence upon £y is dependence
upon the first derivatives of the displacements with respect to coordinates; now the
influence of the second derivatives is also taken into account). Show that, for this
model, it is necessary to take into account an additional input of energy dq". Obtain
the expression for dq'*. Derive the formulae binding the stresses with the derivatives
of the energy function (generalizing the formulae of Problem 15.9).
Q 15.17 Show that the equations of state having the form
p'J = -p9ij -f 7y , p = p(p, T) , 7y = T y + 2/iey , T y = COIlSt
contradict the second law of thermodynamics (the parameters T, p, e^ are supposed

to determine completely a state of the considered medium).
0 15.18 For an anisotropic linearly viscous fluid (r" = J4 , J '*'(T) ew), calculate the
maximal number of possible independent viscosity coefficients A1**1. Use Onsager's
principle.
0 15.19 For an anisotropic medium with heat conductivity obeying Fourier's law,
calculate the number of possible independent thermal conductivity coefficients.
© 15.20 The expression for the entropy production rate valid for irreversible dif
fusion and heat conduction processes in a mixture of two non-reacting gases (e.g., of
nitrogen and hydrogen) at equilibrium is known to be written in the form
d\s q pi - p.2
where q is the heat flux vector, J = pi{vi — v) is the diffusion flux vector (vector of
flux of the first component of the mixture), p,\ and p-2 are the chemical potentials of
the mixture components, which are known functions of the temperature T, pressure
p, and concentrations of the components C\ and c^. The concentrations are defined
as the ratios pt/p where p = p\ + pi. When the pressure gradient vanishes,
_ .AM-M2 ( d p.i-p.2\ ___jrr , ( d M i ~ M _ , „

= S[adT+ 0
V*-T~ w~T^)c,p [dc-^r-)^
where c denotes the concentration of the first component. Taking q and J as the
"fluxes" and (gradT)/!12, grad((/ii - pi)/T) as the "thermodynamic forces," and
using Onsager's principle, find the relationship between the coefficient determining the
heat flux caused by the concentration gradient c (Duffor's effect) and the coefficient
determining the diffusion flux caused by the gradient of temperature (Soret's effect).
16 Thermodynamics of Media with Internal An

gular Momentum.
The problems of this section deal mainly with establishing various constitutive equa
tions consistent with the first and second laws of thermodynamics (see Sections 12-14)
for media with internal angular momentum. As examples, media of the type of ne-
matic liquid crystal with unit orientation vector n are considered. For such media,
n determines the angular velocity of internal rotation fi = n x — (see Section 11).
The internal energy is supposed to have the form
u = £rot + u(p, s, gij, n\ Vtn*)
where £TOl = - /fi 2 = - fcjfi' is the energy of internal rotation (— = 0), k is the vector
of internal angular momentum, p is the density of the medium, s is the entropy per
unit mass. An electromagnetic field is regarded as an external factor; therefore, the
thermodynamics of magnetization and polarization processes is not considered in
the problems (see Problem 16.11). The corresponding quantities H, M and E, P
are regarded as given by certain equations. Gyromagnetic properties of the media
are supposed to be unessential. It follows from the equation of internal angular
momentum (see Section 11) that variation of internal rotational energy is described
by the equation
p°±i2± = pn™= phi& + v,-((tf n<) - p % t n * - ctfv^n* (ie.i)

at at
(eijk are the components of Levi-Civita tensor) which allows one to give the following
interpretation of the terms in the right-hand side: phfil1 + Vj(Q^fi') is the work (per
unit volume for unit time) of external body and surface couples, — p^cy^n* is the
work done by internal surface forces on variation of the orientation, —Q'JVjtP is the
work of internal surface couples. In accordance with this interpretation, the energy
equation corresponding to the first law of thermodynamics is supposed to have the
form
dq„
P j t ( y + u\ = Vîu + Q«n, - f) + pPvi + phitf + p
dt
Thus, there are additional (non-heat) body and surface energy inputs
p ^ = Ph0+ vj{Qimi).
16. Thermodynamics of Media with Interned Angular Momentum. 105
With the use of the evolution equation (16.1) for the internal rotation energy and the
kinetic energy theorem, the evolution equation for the function u is derived:
p % + QVVjSli + f!*cUi(tt -u') + PTYt~p^-

Here e is the strain rate tensor, w is the vorticity vector, values of the work done
by the symmetric and antisymmetric parts of the stress tensor p are separated, and,
dqf
according to the second law of thermodynamics, the uncompensated heat p — > 0 is
introduced. This equation allows one, in accordance with the supposition that Gibbs'
identity is valid, to determine the "reversible" parts of the tensors p and Q. With
the fact that the function u is scalar (Problems 16.4-16.5) taken into account, Gibbs'
identity has the form
+ {*.+>&+<<)'»<*-»+>{*-%)*-<>
ds
where the quantities V.v, = ei} + «„•*<*/*, V^fi, and — are independent. It is possible
to construct such functions x'(£k,t) (describing motions of a medium), n*(£k,t), and
s(£k, t), for which only one arbitrary component of the indicated quantities is nonzero,
while the others vanish. The form of "irreversible" parts of the tensors p and Q, as
well as of the heat flux vector q, is determined on the basis of the supposition that
the entropy production has the form
dis 1 dq' <?'V,T 1 ^
dt T dt pT2 pT V
For many cases, dependence of the "thermodynamic forces" Xn upon "fluxes" xn is
linear and homogeneous: Xn - Lnmxm, satisfying Onsager's principle of symmetry:
Lnm = Lmn. The quantity <x is called the dissipation function.
— PROBLEMS —
0 16.1 a) Derive the formula expressing the kinetic energy theorem in integral
form for an absolutely rigid body, b) What does the work of internal surface forces
for any absolutely rigid body equal if there is no internal angular momentum and
no distributed body and surface couples? c) Derive the expression for the rotational
kinetic energy of a rotator (see Problems 11.7-11.8).
Q 16.2 a) Using the equation of motion for a medium and the equation of inter
nal angular momentum, derive the evolution equation for the "total" kinetic energy,
including the kinetic energy of internal rotation, b) Separate in the equation of item
a) the work of internal surface forces connected with the antisymmetric part of the
stress tensor. For what cases can it vanish if the stress tensor p does not depend upon
the relative angular velocity of internal rotation ft - w? c) Determine the physical
meaning for the rest of the terms in the equation of item a).
Q 16.3 Using the equation of total energy for media with internal rotation (see the
introduction to this section) and the equation of Problem 16.2a, obtain the evolution
equation for the function u (u is the internal energy of the medium from which the
kinetic energy of internal rotation is excluded).
© 16.4 Derive the differential condition for the function u(n\ V,n',3y) to be in
variant as a scalar with respect to infinitesimal rotations of the Cartesian coordinate
system (x*) where n' are the components of the orientation vector of the medium,
V,n' are the components of its covariant derivative, gtj are the components of the
metric tensor.
G 16.5 a) Exclude the external heat input — = — V^* H ^ S fr0m the evo-
' at p at
lution equation for the function u (Problem 16.3), using the second law of thermody
namics. Supposing that the processes of entropy variation, motion, deformation and
internal rotation are independent and reversible and the stress p and couple stress Q
ds
tensors do not depend upon the tensors Vu, Vfi and the derivative —, determine
expressions for the tensors p and Q if 1) u = u(p, s), 2) u = u{p,s,n\ V^n'.^y)
where n is the orientation vector (use the answer to Problem 16.4). b) In a Cartesian
coordinate system (x*), construct orientation distributions n(xk,t) in a medium at
rest for which only one of the components of the tensor Vi?, e.g, V2O3 or V3fi3 is
nonzero, the others being equal to zero.
0 16.6 Using results from the solution of Problem 16.5a, write the Gibbs' identity
of media for which each of the tensors p and Q can be represented as the sum of
the "reversible" and "irreversible" parts. Assuming the function u(p,s,n\ Vjn'.gy)
as given, derive the formulae for the reversible parts of the tensors p and Q and for
the temperature T. Using the evolution equation for the function it (Problem 16.3),
determine the expression for the uncompensated heat p-rr- Derive the formula for
the entropy production — .

at
16. Thermodynamics of Media with Internal Angular Momentum. 107
0 16.7 a) Supposing that the irreversible parts of the stress p and couple stress
Q tensors and also the vector — — gradT depend, as thermodynamic forces, upon the
thermodynamic fluxes e, e • {fi — u>), q (e is Levi-Civita tensor) as linear and homo
geneous functions (there is no dependence on V/2) and regarding the corresponding
tensor coefficients as functions of the metric tensor g (see Problem 5.16), determine
the general form of the dissipation function a for media with internal angular mo
mentum. Apply Onsager's principle, b) How does the result of item a) change if
the quantities - (p,J + p 7, )i rrev , Pinev£y*> — ^ r are regarded as thermodynamic forces
and ei}, £lk — wi> Qi are regarded as thermodynamic fluxes, with Qmev vanishing?
c) Determine the relations between the thermodynamic forces and fluxes defined in
the item b), supposing that their coefficients depend upon the tensors g and nn
where n is the orientation vector of the medium. Use results from the solution of
Problems 5.17, 5.20.
0 16.8 For the case b) of Problem 16.7, reduce the dissipation function a to the
sum of squares of independent thermodynamic fluxes, and write the inequalities for
the coefficients in according to the second law of thermodynamics.
© 16.9 a) Show that the vector on the righthand side of the equation of internal
angular momentum with the components VjQ^j^ — pJJvCy* where
v3 = —o r V n - p — a1
Kev P k y
dVjn dpy
and QJL = —p a _ k eMni is always orthogonal to the orientation vector of the

dVjn
medium n for any sufficiently smooth scalar function u(p, s, n*, V ^ n ' , ^ ) (see Prob
lem 16.4). b) Let p|£«vey-*n* = 0 for any vector n. Determine the restrictions for
the scalar coefficients and the general form of the thermodynamic force with the
components V^^tijk satisfying the conditions of Problem 16.7c.
0 16.10 Consider a steady shear motion of a magnetizable liquid crystal between
two infinite parallel plates the distance between which equals h. The temperature of
both plates is maintained constant and equal to T0. The evolution equation for the
function u has the form
p 5 = - v «9'+P 0e «+p%*(n* - "").

at
where u = cT, the heat flux vector is q = -KgradT (c, K are constant), and the
other constitutive equations, the velocity, orientation distributions and the magnetic
field are the same as in Problem 11.17. Determine the temperature distribution T
inside the stream, and find the normal component qn on its boundaries. Determine
dq"
the energy input distribution —— (see the introduction to this section).
at
Q 16.11 Let the equation of internal energy for the model of incompressible mag
netic fluid with internal rotation, whose constitutive equations are presented in Prob-
. dq" dM\
lem 11.18, have the form [p—rr- dt = H, —j—)
dt
du „• „ dMx „ i
P-=P»Vjvi + Hi — -W
tf2 1
where u = Uoe'^c + \- - IQ2, a = -nViT, UQ, C, y, /, n are constant, a) Using
2px 2
the equation of internal angular momentum, the second law of thermodynamics and
the relationship
T= —
gs M=const '
fi=const
show that the uncompensated heat has the form
d£ 2MeyeiJ + X\H - w| 2 + r
dM_
-f2xM
' dt dt
b) Consider a steady shear flow of such a fluid between two infinite parallel plates
separated by a distance h. The temperature of the plates is maintained constant and
equal to T0. The velocity distribution, magnetic field, and distribution of the quantity
fi are the same as in Problem 11.18. Determine the distributions of temperature T
dqT
and of heat flux q. What is equal to?
dt
Chapter 4
Discontinuity Surfaces in
Continuum Media
17 Conditions across Discontinuity Surfaces.

A discontinuity surface is an isolated surface across which one or more quantities
change step-wise. If these quantities are parameters describing the motion or the
state of the medium, the surface is called the surface of strong discontinuity. If
the parameters themselves are continuous, but their derivatives change step-wise,
the surface is called the surface of weak discontinuity. The values of jumps across a
discontinuity surface cannot be arbitrary; at each point of the surface, they are bound
by relationships called jump conditions. Some of these conditions follow immediately
from the mass, momentum, angular momentum, energy coservation laws. These are
the conditions (6.11)-(6.14) presented in Section 6. The jumps of derivatives across
a weak discontinuity surface are bound because, in the vicinities adjacent to the sides
of the surface, the parameters satisfy certain differential equations. Such conditions
are called dynamic conditions. Jump conditions of another type, called kinematic
conditions, are connected neither with the conservation laws nor with the equations
satisfied in the vicinities adjacent to the sides of the surface. Kinematic conditions
across a strong tangential discontinuity represent purely kinematic relations between
the velocities (displacements) of the surface and those of medium or between velocities
of the media or of the media in the vicinities adjacent to the sides of the surface.
Kinematic conditions across a surface of weak discontinuity follow from the fact that
parameters themselves are continuous across the surface. The notation [i^j in this
chapter means the value of a jump of <p across a discontinuity surface.
109
110 DISCONTINUITY SURFACES IN CONTINUUM MEDIA
— PROBLEMS —
0 17.1 Let the equation of a surface in an Eulerian coordinate system be

f(xl,x2,x3,t) = 0. Find the component of velocity of this surface along its nor
mal directed to the side where / > 0.
0 17.2 A moving plane E is a plane of weak discontinuity for a function u(x, t),
where x is a Cartesian coordinate, and t is time. The equation of the plane is f(x, t) =
0. Write the relation connecting the jumps of du/dt and du/dx on E following from
continuity of u(x, t) on E (kinematic condition across a weak discontinuity surface).
© 17.3 A plane E (the equation of which is f{x,t) = 0) is a weak discontinuity
plane for functions «i(x, t), ..., u„(x, t). The functions iij(x, t) satisfy the equations
~dt+aij~dx+ aij = aij{ui,x,t) ,

bi = bi(ui,x,t) , i ,j = l , . . . , n
on both sides of the plane E. a) Write the relations connecting the jumps of dui/dt
and dui/dx following from the last condition (the dynamic conditions across a weak
discontinuity surface), b) Show that a discontinuity plane speed D is an eigenvalue
of the matrix a^.
0 17.4 The equations for one-dimensional adiabatic motion of ideal perfect gas
are (see Problem 24.4)
dp dpv
— + -i— = 0 ,
dt dx
dv dv 1 dp
dt dx p dx
Oj. (P/P1) +vfa (P/P1) =
° ' 7= const
where x is a Cartesian coordinate, p is the density, p is pressure and v = vx, vv =

vz = 0 are gas velocity components. Let x = X(t) be the plane of weak discontinuity
of p, p, v.. a) Write the kinematic and dynamic conditions across the plane x = X(t).
b) Express the plane propagation speed D = dX/dt in terms of values of p, p, v on
it.
0 17.5 a) Derive the jump conditions (6.11)-(6.15) (see Section 6) from the mass,
momentum, angular momentum, energy conservation and entropy variation laws.
b) Write them in the coordinate system relative to which the surface moves with a
velocity D = Dn.
17. Conditions across Discontinuity Surfaces. Ill
© 17.6 A tangential discontinuity is a strong discontinuity surface through which

matter does not flow, e.g., an interface between two media. Obtain the conditions
across a tangential discontinuity from the general conditions (6.11)-(6.15) in absence
of external actions at a discontinuity surface. Consider a particular case when the
velocity vector is continuous at the surface of discontinuity (then the latter is called
a contact discontinuity).
0 17.7 Let the equation of a surface E in an Eulerian coordinate system be

/(x', t) = 0. Write the condition relating the velocity field of the medium and the
function f{x',t) = 0 expressing the fact that the medium does not flow through this
surface.
G 17.8 Write the relations between the values of the components of the tensor
ViUp at two sides of the discontinuity surface expressing the fact that the displacement
vector w = w>ej is continuous at this surface. Relate the change in VittP to the
change in velocity of the medium for the case when the flux of mass through the
discontinuity surface is nonzero. Obtain the relationships for the components of the
tensor of small strains.
0 17.9 Write the conditions across an unknown, for the present, interface
f{x',t) = 0 between two media (tangential or contact discontinuity), following from
the conservation laws (dynamic conditions) and from the fact that the media do not
flow through the surface (kinematic conditions) if a) there are ideal fluids in the
vicinities adjacent to the both sides of this surface, b) there is an ideal fluid on one
side of the surface and a viscous fluid on the other side, c) there are viscous fluids
on the both sides of the surface, d) there is a viscous fluid on one side of the surface
and an elastic solid on the other side, e) there are elastic solids on both sides of the
surface. Write these conditions in terms of the strain rate tensor for a viscous fluid
and in terms of the strain tensor for an elastic solid, assuming that the medium is
isotropic linearly viscous or linearly elastic respectively (see Section 12).
0 17.10 Obtain the conditions across a strong discontinuity for an ideal fluid from
(6.11)-(6.15), assuming that p n = -pn, k = 0, Qn = 0, h = 0 and there are no
external inputs of mass, momentum and energy at the discontinuity surface.
© 17.11 Let an ideal perfect gas (see Section 12) flow through a discontinuity sur
face on which there are no external inputs of mass, momentum and energy. Assuming
that the heat fluxes qni and qn2 are equal to zero (adiabatic process) and that the
values p = pi, p ~ p\ at one side of the discontinuity surface are known, find a) p2
as a function of p2 where the index "2" marks the quantities at the other side of
the discontinuity surface (Hugoniot adiabat), b) the change in entropy s 2 — si as a
function of p2- c) For what cases is the relationship s2 — s\ = 0 valid?
ffi 17.12 An incompressible nonlinear elastic medium moves in such a manner that
all parameters are independent of y and z (x, y, and z are Cartesian coordinates) and
wz = 0 (to is the displacement vector). Write the conditions across a discontinuity
x = X(t) in this medium supposing the tangential stress r = ■pxv to be a function
only of 7 = dwv/dx: r = r(7). Express the velocity of the discontinuity relative to
the medium in front of it in terms of the jumps [7] = 7 2 - 7 1 and [r] = r 2 - TJ. Find
the velocity of a discontinuity with infinitesimal jumps of the parameters across it.
Write the conditions under which a finite discontinuity moves faster than infinitesimal
discontinuities moving in the same direction in front of it and more slowly than
infinitesimal discontinuities moving behind of it (the conditions for "evolutionarity"
of a discontinuity, see Section 24). Express the dependence of the internal energy of
this medium upon 7 and the entropy s in the general form. Write the condition for
the increase of entropy on the considered discontinuity.
© 17.13 Study the relationships across the surface of an "ideal propeller" in a

perfect gas, i.e., across a discontinuity surface where the external momentum is trans
mitted to the gas and the motion is adiabatic and reversible. Let J be the value of
this momentum per unit area of the surface. Show that, if the flow of the gas is
specified in the vicinity adjacent to one side of the surface, two different flows on
the other side satisfy the jump conditions, one of them is subsonic and the other is
supersonic. To choose one of these two solutions replace the discontinuity by a zone
with continuous one-dimensional flow of a gas in a field of an external force (this
continuous flow can be regarded as a sequence of infinitesimal discontinuities of the
considered type). Determine the conditions under which one of the two possible flows
behind the propeller occurs.
© 17.14 a) Examine whether a thin interior layer of a low-viscosity liquid in a

stream of a high-viscosity liquid can be modeled with a tangential discontinuity. For
this purpose, study a planar flow v = v(y)ex where v(y) is a continuous function.
Assume that the viscosity /i depends upon y as follows: fj, = /xx at y > h, /i = /x0 at
—h<y<h, and fj, = fi2&ty < —h. Pass to the limit h —> 0, /J,0 —* 0 at the conditions
Hi = const, H2 = const, shear stresses r are finite. Show that if fj,0/h —> const ^ 0, or
Mo//1 -* 0 then passing to the limit results in a tangential discontinuity. Determine
conditions relating the flow parameters at two sides of it. Show that if n„/h —» 00, the
limit flow velocity is continuous, b) Consider a steady motion of fluid caused by two
parallel planes moving with the velocities v0 in the opposite directions. A distance
between the planes is 2H. The viscosity coefficients of the fluid are given in item a),
and h -C H. Find the shear stress r on the planes and the jump of velocity at y = 0
for different relations between fi0 and h (for h —> 0).
© 17.15 Rain falls upon a surface of water. Treating the rain as a continuous
medium, write relationships on the interface £ between rain and water. Assume the
17. Conditions across Discontinuity Surfaces. 113
water to be an incompressible liquid of density p. Supposing that the velocity of rain

relative to the surface E, its (average) density, and its temperature are known, find
the velocity, pressure and temperature in the water under the surface E.
© 17.16 Rain of average density pB falls with a velocity « 0 upon an unmoving

wedge with the horizontal edge (upon a roof, see Figure 17.1). Assuming that the
flow is steady, and the velocity of particles does not change after passing through the
Figure 17.1
discontinuity, determine the form of interface between water and rain, the velocity
and the pressure in water. The angle of v0 and the edge of the roof is 90°. The
density of water p and the angle a of a pitch of the roof with v0 are known.
© 17.17 An opened barrel with a hole in the back vertical wall stands on a
cart. When it rains, the cart moves because a jet of water flows out from the hole
(see Figure 17.2). Neglecting friction, find the velocity of the cart, assuming that
Figure 17.2
the average density of rain is p0, rain droplets falls vertically with a speed v0, and
the steady level of water in the barrel is at a height of h above the hole. To find
the pressure in the barrel, use the boundary conditions obtained as the answer to
Problem 17.15. Assume that the horizontal component of the velocity of water in the
barrel decays in the upper thin layer because of the presence of vertical walls.
0 17.18 Obtain the equation of an averaged motion of cars along a road with
one-way traffic, assuming that their speed (chosen by drivers) is determined by their
density p and becomes equal to zero at a certain value p = p miX . Assuming that the
flux of cars <p{p) = pv(p) satisfies the condition ip" < 0, determine the conditions
for existence of a single-valued continuous solution at different initial (or boundary)
conditions. For the case when a single-valued solution is absent, introduce disconti
nuities with the jump condition expressing conservation of the number of cars. Solve
the problems about stop and start of a uniform traffic on traffic signal change.
18 Discontinuity Surfaces in the Lagrangian De

scription.
In this section the Lagrangian decription of motion is used. Problems are presented
about derivation of the equalities relating the jumps of various derivatives of the
function describing a motion law of a medium xl = x'(£*,£) with respect to the
Lagrangian coordinates £k and time at the discontinuity surface. The considered
jumps are jumps of these functions themselves, of their first derivatives (corresponding
to strong discontinuities in the Eulerian description) and jumps of second derivatives
of these functions (weak discontinuities in the Eulerian description). Jumps of the
Lagrangian coordinates ffc(x', t) take place if there is external input of mass on the
discontinuity surface. An Eulerian coordinate system is regarded as uniform in the
vicinity of the discontinuity surface, "coordinate" discontinuities are not considered.
- PROBLEMS —
0 18.1 Let a function <p(xl) (i = l,...,n) and all its partial derivatives be con
tinuous at x1 < 0 and x1 > 0. Let also this function and its derivatives with
respect to xa (a = 2, ...,n) have the right and left limits on the surface x 1 = 0
continuous with respect to xQ. Let us denote its jump across the surface by A:
A = limri_o+ <P ~ Hmxi_o_ >p = >P+ _ f- — M- Show that the function A(xa) is con
tinuously differentiable, and d\/dxQ = \d<p/dxa). Write this condition in an arbitrary
coordinate system bound with the system (x1) by smooth transformations.
0 18.2 Let the function <px defined in Problem 18.1 be continuous everywhere
(A = 0) and have a jump of the derivative [Jjjft] = /u(xQ) at x1 = 0. Write this
18. Discontinuity Surfaces in the Lagrangian Description. 115
condition in an arbitrary coordinate system bound with the system (x') by smooth
transformations.
© 18.3 Let a function tp{xx) (i = 1, ...,n) be twice continuously differentiate at

x1 < 0 and x1 > 0. Also, let this function and its partial derivatives | ^ and a a ' g
[a = 2,..., n) be continuous everywhere. Finally, let this function have a jump of the
derivative [§j&J = v(xa) at i 1 = 0. Write this condition in an arbitrary coordinate
system (y'), if the functions y' - F'(xk) are twice continuously differentiable.
0 18.4 Let y' = F'(xk) (i, k = 1 , . . . , n) be a sectionally smooth one-to-one trans

formation and have a jump of the normal derivative at the surface /(x*) = 0. Also, let
this transformation and its tangential derivatives be continuous everywhere. Prove
that [f^- det ( f ^ ) ] • J £ = 0 (the brackets denote the value of the jump), using the
answer to Problem 18.2.
G 18.5 A crack, during its formation in a solid, opens so that the displacement
of its sides equals to [r(fV)j = X(ua, t) where ua (a = 1,2) are the Lagrangian
coordinates on the discontinuity surface modelling the crack. Supposing that all the
necessary derivatives exist, find the jumps of the velocity vector [v] and of the basis
vectors [ea], using the answer to Problem 18.1. Prove that gf?H = ft[ea].
© 18.6 Using the answer to Problem 18.4, prove that, across a strong discontinuity,
{(vn - Dn)/y/g] = 0 where D - v is the velocity of the discontinuity surface relative
to the medium, g is the determinant of the matrix composed of the components of
the metric tensor in a Lagrangian coordinate system.
0 18.7 Let the equation of a moving surface have the form t = f(£9) in a La
grangian coordinate system. Prove the formula §h = ^ ■ Dniv where n' are the
components of the unit normal vector in an Eulerian coordinate system (x*), D — v
is the velocity of the given discontinuity surface relative to the medium.
0 18.8 Using the answer to Problems 18.6-18.7, show that, across a strong dis
continuity, [fik/ \/g\ = 0 where nfc are the components of the unit vector normal to
the discontinuity surface in a Lagrangian coordinate system.
0 18.9 Write the conditions across a discontinuity for the mass, momentum and
energy fluxes valid for p{vn - D„) =£ 0, using the equation of the discontinuity surface
in a Lagrangian coordinate system t = /(£'). Eliminate the quantity vn - Dn using
the formulae of Problems 18.7- 18.8. Study the case / = const
© 18.10 Prove the algebraic equalities [ab] = [a]{b) + [b]{a) = [a][b] + [a]^ + [6]ax
where the notations [a] = a2 - ait (a) = (ai + a 2 )/2 and the same ones for b and ab
are used.
© 18.11 Express, in a Lagrangian coordinate system, the jumps of the compo

nents of the metric tensor [pM] across a strong discontinuity in terms of the jumps
of Eulerian components of the velocity vector [vl], the components of the distortion
matrix ( | ^ ) in the state ahead of the discontinuity, and the equation of the discon
tinuity surface t = /(£')•
0 18.12 Express the jumps of the density of a medium [p] across a strong discon
tinuity in terms of the jumps of the normal component of the velocity of the medium
[vn], of the velocity of the discontinuity surface Dn and the state parameters ahead
of the discontinuity.
0 18.13 Determine the relation between the jump of the Christoffel symbol [l^,J
in a Lagrangian coordinate system across a strong discontinuity and the jumps of
Eulerian components of the acceleration of the medium [a'} and the continuous pa
rameters of the medium.
0 18.14 Express the jumps of Lagrangian components of the strain rate tensor
[eM] and vortex tensor [d>p,] across a weak discontinuity in terms of the jumps of
Eulerian components of acceleration of the medium [a1] and the continuous parameters
of the medium.
0 18.15 Let Lagrangian contravariant components of a tensor of second rank T
be continuous across a strong discontinuity. Express the jumps of the components
[T*J] in an Eulerian coordinate system in terms of the jump of the velocity of the
medium [vi], the normal speed of the discontinuity surface Dn and the parameters of
the medium ahead of the discontinuity.
© 18.16 Let the input of mass M = p2(vn2 - Dn) - pi(vnl - Dn) per unit area
for unit time take place at a discontinuity surface. Using the definitions of surface
density of mass a and that of the velocity vector v of a particle of a medium situated
on a discontinuity surface at a given instant, derive the formula for M from the
mass conservation law in the integral form. Take into account that the coefficients
Gap of the first quadratic form of the discontinuity surface ds2 = Ga0 dua du*3 -
(e Q • ep)duadut) (where ua (a — 1,2) are the surface coordinates, and ea are the
surface basis vectors) varies in time.
Chapter 5
Fluid Mechanics
19 Summary of Equations of Fluid Mechanics

Since there is no difference between liquids and gases from the viewpoint of mechanics,
these two kinds of medium are often called fluids. Fluids are described by models
for which a state of rest is impossible in the presence of tangential stresses. This
reflects the property of fluidity of liquids and gases. Consequently, at rest pn = -pn,
p*} = -pg'j (p„ is the stress vector on a surface element with normal n, p is the
pressure, pij, giJ are the components of the stress tensor and metric tensor). The
property of viscosity manifests itself in the fact that, during motion (more exactly,
during deformation) the tangential components of the stress vector are not zero in a
general case, and the components of the stress tensor can be written in the form
p<; = -pg'i + T « (19.1)
where TIJ are the components of the viscous stress tensor, depending upon components
of the strain rate tensor eu- All real fluids are viscous and compressible. However,
under certain conditions, one or both of these properties can be insignificant, which
allows the introduction of simplifications into the models under consideration. An
ideal (inviscid) fluid is a medium for which p ,J = -pg,} not only at rest but also
during motion, i.e., T*J = 0. The equations of motion for an ideal fluid (Euler
equation) have the form
p^ = -grsdp + pF (19.2)
at
where p is the density, v is the velocity and F is the mass density of the body forces.
A viscous fluid is a medium for which p,j are written in the form (19.1) where
Tl> = T»(eQ0<ga0,T),
p = p{p, T) for compressible fluid and p is an independent parameter for incompress
ible one. Note that rij can also depend upon various physical-chemical parameters,
117
118 FLUID MECHANICS
e.g., upon concentrations of components in the case of mixtures. Giving explicitly

the dependence of T y upon the arguments specifies a particular model for a viscous
fluid. The Navier-Stokes model (Newtonian or linear-viscous fluid), in which the
dependence r' J upon eap is linear, is the most widely used. For an isotropic medium,
this dependence has the form
rij = \divvgij +2/j,eij .
The viscosity coefficients A, fi are different for different media. In general, they are
not constant, but are dependent on the temperature and the pressure. However, for
many cases, the variation of A and (i is insignificant, and they are regarded as given
constants. The equations of motion for a linear-viscous isotropic fluid with constant
coefficients of shear (n) and bulk (£ = A + 2/x/3) viscosity
p — = pF - gradp + (C + M/3)grad div v + p,Av (19.3)

at
are called the Navier-Stokes equations. There exist also other models in which the
dependence r y (e Q/3 ) is nonlinear (non-Newtonian fluids). The Euler or Navier-Stokes
equations together with the continuity equation
dp
— + pdivt> = 0 ,
at
the equation of internal energy (u is the internal energy per unit mass)
dt fPdt+ p ,]
dt '
the equations of state
/(p,p,T) = 0 , u = u(p,T)
and the equation determining the heat input dq, e.g.,
dq 1 ,. K
— = —divq = - A T
dt p p
(when heat conductivity obeys the Fourier law with the thermal conductivity coef
ficient K = const) form the closed system of equations, respectively, for an ideal or
linear-viscous fluid. When solving particular problems, to obtain a specific solution
of a system of partial differential equations, boundary, initial and other conditions
must be specified. The typical boundary conditions for an ideal fluid are as follows.
If the location of the boundary dV of the fluid is known, the normal component of
the velocity is ordinarily specified at the points of dV. For example, the kinematic
"condition of impermeability"
Vn\dV = Un,
20. Statics of Fluid 119
where vn = v ■ n (n is the normal, U is the velocity of the boundary) is ordinarily

specified on the contact surface of a rigid impermeable body and fluid. If the location
of the boundary of the fluid is unknown in advance and should be found by solving
the problem, then besides the kinematic condition, a dynamic condition for the stress
vector on dV is used. In particular, the following kinematic (a) and dynamic (b)
conditions are specified on an interface between two ideal fluids:
(a) vnl = vn2 = Un; (b) pi = p2
(the indices "1" "2" mark the parameters at the different sides of the boundary;
surface tension is not taken into account in the condition (b)). If the pressure p = Po
is given in one of the fluids, the conditions on the boundary for the other fluid have
the form vn\av = Unt p\av = Po [U is the velocity of the boundary determined by
solving the problem). If p0 = const these conditions are called the conditions on a
free surface.
For a viscous fluid the order of the equations is higher than that for an ideal one and
the number of the boundary conditions is more. On an impermeable boundary, e.g.,
one of the following variants of the conditions can be specified: 1) the "conditions
of adhesion" v\gV = U (U is a given velocity of the boundary); 2) vn\av = Un,
pnT\gV = fT (Un and fT are given normal component of the velocity of the boundary
and tangential stress on it); 3) pnn\dv = fn, Pnr\av = fr (/n»», fT are given normal
and tangential stresses on dV). In the last case, to determine the normal components
of the velocity of the boundary, the relationship vn\ev = Un must be used. If, besides,
/„ = const and / T = 0, these conditions are called the conditions on a free surface.
On an interface between two viscous fluids, the kinematic conditions (a) and the
dynamic condition of continuity for the stress vector (b) must be satisfied:
(a) w„i = vn2 = Un , vTl = VT2 (b) pnl = pn2 .
If the region occupied by an ideal or viscous fluid is infinite, in the simplest cases, the
pressure and velocity of the medium are specified at infinity. When solving problems
of temperature determination in fluids, the values of temperature or heat flux are
ordinarily specified on dV. The choice of the model is determined by the aim of the
study and the required accuracy of the results.
20 Statics of Fluid
Statics of fluids (hydrostatics) studies equilibrium of liquids and gases. In equilib
rium, v = 0, and a viscous fluid does not differ from an ideal one. The equations of
equilibrium have the form
gradp = pF .
120 FLUID MECHANICS
It follows from the continuity equation that dp/dt = 0 at rest, i.e., p = p(x'). For a
homogeneous incompressible fluid (p = const) with a given density resting in a given
field of body forces, the equations of equilibrium and the boundary conditions deter
mine the pressure distribution. In the case of a compressible fluid or incompressible
inhomogeneous fluid, additional relationships or information is necessary for solution
of problems, e.g., the relation between p and p (for compressible media) or the density
distribution in the fluid (for an incompressible fluid). Boundary and other additional
conditions are necessary for specific problems to determine the solution uniquely. For
example, the pressure is specified on a free surface. If the location of the free surface
is unknown, the mass of the fluid being in equilibrium can be specified in addition.
If a body is surrounded by a fluid, the force FA
FA = -JpgdV,
v
where p is the density of the fluid, V is the volume of the body, acts on the body
(Archimedes' law). The line of action of the force F& passes through the center of
gravity of the fluid as if it were filling the volume of the body.
- PROBLEMS -
0 20.1 Show that, when an inhomogeneous fluid is in equilibrium, a) the field of

body forces F satisfies the condition F ■ curl F = 0 (and the condition F = grad U
if the fluid is homogeneous); b) if the field of body forces has a potential U, then
p = p{U), p = p(U), as well as, consequently, T = T(U) for the gas with the equation
of state f(p, p, T) = 0; c) if F = grad U, the interface between two fluids with different
densities coincides with the equipotential surface (with the surface where the values
of the function U are identical), and the pressure is constant on it.
0 20.2 A heavy homogeneous liquid is poured into a vessel with a plane wall (see
Figure 20.1). The free surface of the liquid and the vessel itself make contact with
the atmosphere, the pressure in which is p, = const. Let V be the resultant of the
pressure forces acting on a plane surface E lying on the wetted part of the plane wall,
and HQ the depth of submersion of the pressure center O (the center of application
of this resultant). Show that 17*1 = pgShc, ho = I/(hcS) where 5" is the area of E,
he is the depth of submersion of the geometrical center of E (point C) and / is the
moment of inertia of E relative to AB, the line of intersection of the free surface with
the wall.
© 20.3 Find the force V acting on a square wall of an aquarium that is full of
water. At what height H is the point of application of this force?
Figure 20.1
© 20.4 An opened heavy cap in the shape of a circular blunt-nosed cone, with
the angle a at the base and with the radius of the base R, stands on a horizontal
plane (see Figure 20.2). Of what weight G must the cap be to keep water poured
Figure 20.2
into it up to the height HI The pressure everywhere outside of the cap equals the
pressure in the atmosphere.
0 20.5 A vertical board separating two different levels of water Hi and Hz is set
in a rectangular channel (see Figure 20.3). The board can rotate around a hinge at
a point O. A stop is fixed on the side of the smaller level. At what minimal distance
H from the bottom must the hinge O be positioned in order to make the board open
when the lower water level exceeds H2 and close when it is less then Hfl
© 20.6 A rectangular channel is partitioned off by an unfixed concrete slab having

the shape of a rectangular parallelepiped (see Figure 20.4). What condition must the
width a and height h of the slab satisfy to cause the slab to be in equilibrium if the
level of the supported liquid equals HI Draw the region of equilibrium of the slab
on a plot of a/H versus h/H, assuming that the density of concrete is pc = 3 g/cm 3 ,
and the friction coefficient between the slab and the bottom is k = 0.2. The liquid is
homogeneous, its density is p = 1 g/cm3.
122 FLUID MECHANICS
mm
Ay/y^/yAW/V/WAy/ •vVXv
Figure 20.3
t—<*—t
s?//
____, _—__JC
Figure 20.4
0 20.7 A homogeneous liquid rests in a uniform field of gravity forces. A body

of arbitrary shape is submersed into the liquid and completely surrounded by it.
Derive the formula for the force acting on the body due to the liquid (Archimedes'
law). Show that the line of action for this force passes through the geometrical center
C of the volume V of the body (r c = V - 1 / rdV). Write the expression for the
v
Archimedes force acting on a body in an inhomogeneous liquid. Under what condition
is equilibrium possible for a inhomogeneous body submersed into a homogeneous
liquid? What force acts on a body partially submersed into a liquid?
0 20.8 A thin rod is attached with one end to a wall of a vessel, and submersed
into water with the other end (see Figure 20.5). The rod can rotate around the
horizontal axis of a hinge O positioned above the level of water, a) Find the density
of the material of the rod pT if half of the rod is known to be submersed into the
water at equilibrium, b) Calculate the ratio of the reaction force on the hinge R to
the weight of the rod P.
© 20.9 A moving opened tank contains a heavy liquid which is at rest relative
to the tank. Find the angle a made by the free surface with the horizon if the tank
a) moves in a horizontal plane with constant acceleration a; b) slides along a plane
Figure 20.5
inclined at the angle 8 with the horizon. The friction coefficient k is known. For what
value of k is the surface of the liquid horizontal?
© 20.10 A heavy homogeneous liquid poured into a vertical circular cylindrical

vessel of radius a (a glass) rotates around the axis of the cylinder as a rigid body with
a constant angular speed Q. a) Determine the pressure at each point of the liquid and
the shape of its free surface if the liquid is known to have the level H from the bottom
of the vessel when resting, and the pressure on the surface of the liquid is known to
equal the pressure in the atmosphere, b) Calculate the force V acting on the bottom
of the vessel, c) Formulate the principle for this case similar to Archimedes' law.
© 20.11 Answer the questions of Problem 20.10 a) and b) if there are two liquids
with different given densities p\ and p^ and masses M\ and M2 in the glass. Determine
the shape of the interface between the liquids. Consider the case when the free surface
intersects the bottom of the glass.
0 20.12 A closed vessel filled with a homogeneous heavy fluid rotates with a
constant angular speed Q relative to a horizontal axis. Show that the surfaces of
equal pressure are the coaxial circular cylinders the axis of which is situated at the
height g/£l2 above the axis of rotation.
© 20.13 Explain why a long log always floats on water so that its axis is horizon
tal. One would think that the center of gravity of the log in stable equilibrium must
be immersed to the maximal depth. Why this is not true?
0 20.14 A perfect gas (p = RpT) rests in a uniform field of gravity forces.

Assuming that the dependence of temperature on the height T(z) is known (a model
of the atmosphere), find the distributions of pressure and density p(z), p(z). Consider
the particular cases: a) T = T0 = const. At what height does the density of an
isothermal atmosphere decrease to half that at the surface of Earth, p0 = 1.293 kg/m 3 ,
p0 = 1.033 kgf/cm2? b) T = T0 - ATz/100 where T0 is the absolute temperature at
z = 0, AT = const is the temperature drop over the height 100 m, and z is measured
in meters. Show that p/p0 = {p/p0)n in this case, i.e., stratification of the atmosphere
124 FLUID MECHANICS
is polytropic. Find the value of AT corresponding to adiabatic (n = 7 = Cp/cv)

stratification (for air 7 = 1.4, R/g = 29.27 m/K).
0 20.15 Show that the condition dp/dz < 0 is a necessary condition for stable
(or, when dp/dz = 0, neutrally stable) equilibrium of a nonhomogeneous heavy fluid
(the axis z is in the opposite direction to g).
© 20.16 An atmosphere is referred to as polytropic if p(z) = p0(p(z)/'p0)n', where

n is constant. Prove that equilibrium of the atmosphere relative to adiabatic dis
placements of the particles of the gas (for which p/p1 = const, 7 = Cp/cy) is stable
at n < 7, unstable at n > 7, and neutrally stable at n = 7.
© 20.17 Write the equation of mechanical equilibrium for a gaseous star of density
p, particles of which are restrained by the forces of gravitational attraction. Take into
account that the gravitational potential U satisfies the Poisson equation
At/ = -4irGp
where G is the gravitational constant. Find the pressure distribution in the star and
its radius if the mass of the star is M and a) p = pa = const, b) p = Cp6/5. Assume
that the pressure in the center of the star is finite, and the pressure outside of the
star is equal to zero.
21 Dynamics of an Incompressible Ideal Fluid

A fluid is referred to as incompressible if the density of each individual particle is
constant. Then
dp
— = 0 (the condition of incompressibility),
and the continuity equation yields
divw = 0 .
The property of compressibility is characterized by the value of the relative change
in density of the particles in a process Ap/p; Ap/p «: 1 if the following conditions
are satisfied: 1) v < a, 2) L < t'a, 3) gL < a 2 (when the motion takes place in
the gravitational field), 4) /3AT < 1. Here, v, a = y/dp/dp, L, f, AT, /? = i §£,
and g are respectively the typical velocity of the medium, the speed of sound, length,
time, change in temperature, the thermal expansion coefficient, and the gravitational
acceleration. The inequalities l)-3), corresponding to the fact that the change in
density caused by the change in pressure is small, can be rewritten in the dimensionless
21. Dynamics of an Incompressible Ideal Fluid 125
form: 1) M < 1, 2) MSt < 1, 3) M/Fr < 1 where M = v/a is the Mach number, St =
L/(t*v) is the Strouhal number, Fr = vjyfgL is the Proude number. The condition
4) means that the change in density connected with the change in temperature is
small. If p not only depends on p and T, but also on other parameters, e.g., upon
the salinity, additional inequalities ensure satisfaction of the condition Ap/p <C 1.
Let us note that satisfaction of the conditions l)-4) does not always mean that an
incompressible fluid model can be used. When studying phenomena, the existence of
which itself is caused by the ability of the medium to change its density (e.g., sound
propagation, convection, etc.), the value of Ap/p cannot be neglected however small it
may be. For an incompressible ideal fluid, the problem of determining the mechanical
motion (i.e, calculation of v(xl, t), p(x', t), p{x%, t)) involves the solution of a system
of equations consisting of the condition of incompressibility, the continuity equation,
and the Euler equation. The equation of heat flow is used to determine afterwards
the temperature T(xl, t), if necessary. When using an ideal fluid model, problems, as
a rule, turn out to be considerably simpler than when using a viscous fluid model, not
only because the order of the system of differential equation is lower, but also because
motion of an ideal fluid, by contrast to that of a viscous fluid, is potential for many
cases. For potential flows, v = grad <p, w = | curl v = 0. According to theorems of
Kelvin and Lagrange, for motions of a homogeneous incompressible fluid in a field of
potential body forces, potentiality of the flow is conserved over time in any material
volume. In particular, all motions starting from a state of rest are potential under
the enumerated conditions.
General properties of potential flows of an incompressible

fluid
From the continuity equation div v = 0, it follows that the potential of the velocity
in an incompressible fluid satisfies the Laplace equation A<p = 0, i.e., ip is a harmonic
function. Boundary-value problems are divided into three basic types. In these
problems, it is necessary to find a harmonic function in a domain V on the boundary
dV of which the following conditions are met: 1) <p\gv = / (a Dirichlet problem);
2) d<p/dn\av = /i (a Neumann problem); 3) <p\avi = f, d<p/dn\dva = fi, dV =
dV\ U dVi where / and /i are given functions (a mixed problem). If the domain V
does not contain a point at infinity, the problems are referred to as internal; otherwise,
they are external. Additional conditions at an infinitely remote point are necessary
for external problems. It is often taken that, when going into infinity along any path,
the velocity approaches zero or some given constant. Let us note that, within the
mentioned settings, the problem of determining the velocity of the fluid is separated
from the problem of determining the pressure. Besides, when determining the velocity,
superposition of solutions can be used, since the Laplace equation as well as the
126 FLUID MECHANICS
enumerated boundary conditions are linear. In some hydrodynamic problems, the

boundary conditions for the potential can be nonlinear, e.g., in the theory of waves
(see Section 23).
— PROBLEMS —
0 21.1 Let the velocity field of an unbounded volume of an ideal incompressible

fluid be caused by motion in it of a rigid body, the form and size of which are
known, a) Formulate the boundary-value problem for the potential of the velocity
field, assuming that the motion is potential and continuous everywhere outside of the
body, and the fluid is at rest at infinity, b) Show that the velocity field of the fluid
at each instant is determined only by the velocity distribution over the surface of the
body, and does not depend, e.g., upon the acceleration of the body, c) Is this true
for the pressure? d) Enumerate the conditions under which the motion of the fluid is
potential.
0 21.2 A rigid plate in the shape of a circular disk of the radius a falls onto
the surface of a fluid filling the half-space (see Figure 21.1). The plate touches the
surface of the fluid at the instant t = 0, and instantaneously change its velocity to a
value U\. a) Set up the problem (formulate the equations and boundary conditions)
—_-_- — _ —
x^=r--3^- '"tfiSH
_T"^. —
Figure 21.1
of determining the velocity of the fluid immediately after the impact. Neglect the
viscosity of the fluid, b) Assuming the solution of the problem to be known, find the
velocity of the plate UQ at the instant immediately preceding the impact. The mass
of the plate equals m.
0 21.3 Prove that the kinetic energy of a potential flow of an incompressible
fluid in a simply connected domain is less than that of any other flow with the
same distribution of normal velocities over the boundary of the domain (the Kelvin
principle).
0 21.4 Prove uniqueness of the solution for internal Dirichlet, Neumann and
mixed problems for a simply connected domain.
0 21.5 A resting closed vessel is filled completely with a nonhomogeneous (p ^

const) ideal incompressible fluid. Thefluidis at equilibrium in field of gravity forces.
Show that, if one begins to move the vessel horizontally with acceleration, this causes
a non-potential motion of the fluid relative to the walls of the vessel. Study also the
case of a homogeneous fluid.
0 21.6 Show that, at an internal point of a domain of potential flow of an in

compressible fluid, a) neither the potential of the velocity nor any projection of the
velocity onto a Cartesian coordinate axis can reach either the maximal or the minimal
values; b) the speed cannot reach the maximal value; c) the pressure of a homogeneous
ideal fluid cannot reach the minimal value if the field of body forces F is solenoidal
(divF = 0).
0 21.7 Prove the Green identity
I (tp dip/dn - ,j> dtp/dn) dS = f(<pAip-ip A<p) dV

av v
where dV is the sectionally smooth boundary of a domain V, dS is a surface element
on dV, d/dn denotes differentiation along the outward normal at the point of the
element dS, dV is an element of volume. For what set of functions is this identity
valid?
© 21.8 Let V be a bounded space domain and V its complement. Functions <p
and if are harmonic respectively in the domains V and V with (p —> 0 at \r\ —> oo.
Prove the following identities
dV
1 d<P - d 1 \ _,„ „
dS
\r0 - r\ dn0 - v ^—
dn0 \r0 - r |■ ) ° = ° '
i
dV
i a? d I .
_ =
r 0 - f\ dn0 V "5—
dn0 1\r0 - r\
^T "ô °'
av
1 r( 1 dtp . d 1
av
/Urr^-^kT^r5^^'
AIT J \\r0- f\ dn0 dn0 \r0
where the points with the radius-vectors r, f, r 0 belong respectively to V, V, dV

(dV is the boundary of the domain V), dS0 is a surface element on dV, d/dn0 denotes
differentiation along the outward normal at the point r 0 to the element dS0.
128 FLUID MECHANICS
© 21.9 Prove that

a) the flow of an ideal incompressible fluid, caused by motion of a body of finite
dimensions surrounded by the fluid, has a velocity potential that can be decomposed
far from the body as
,_/(0+<a>+c,0£i+...,
where r = y/{x1)2 + (x 2 ) 2 + (x 3 ) 2 , x' are Cartesian coordinates, and the body is
situated at a finite distance from the origin of the coordinate system; b) the coefficients
of the decomposition are written in the form
v
dV dV '
where n is the normal external to the body, and dV is the boundary of the body.
c) Express C and C* in terms of the volume of the body V, the momentum Q =
— J fnpndS and the geometrical center r* = V~l JrdV. The second and third
dV V
terms in this decomposition are, respectively, the potentials of a source and a dipole;
they are followed by the terms corresponding the potentials of multipoles.
0 21.10 Let ip and <p be the solutions, respectively, of internal and external Neu
mann problems with identical conditions on the boundary. Prove that the potential
<p can be written in the form of the potential of a double layer with the density
fi= {(p — <p)/4ir, i.e., in the form
W^°^(R^) d 5 °-
0 21.11 Let ip and (p be the solutions, respectively, of internal and external Dirich-
let problems with identical conditions on the boundary. Prove that the potential
ip can be written in the form of the potential of a simple layer with the density
q = (d(ip — (p)/dri)lkK, i.e., in the form
J \r-rro\0\
dV ' '
Planar potential flows

A motion of a medium is referred to as planar if all the particles lying on a perpendicu
lar to certain plane (plane x, y with unit normal e) perform identical motions parallel
to this plane. It is convenient to describe the velocity field of a planar potential flow
of an incompressible fluid with a complex potential W(z) = <p(x,y) + iip(x,y), an
analytic function of the complex variable z = x + iy. Here, i>(x, y) is the stream
function, v = grad</> = curl(ê). The derivative dW/dz determines the velocity of
the flow
dW
-^ = vx-tvv.
For any curve C positioned in the domain occupied by the fluid,
J^-dz = T + iQ
c
where T = J(vxdx + vydy) is the velocity circulation over the line £, Q =
c
J(vx dy — vy dx) is the flow rate through the line C. In planar problems of flow around
£
rigid bodies, the domain offlowis not simply connected; and, to select a single solution
of a boundary-value problem, the values of velocity circulation over closed contours
which cannot be shrunk into a point (cyclic constants) must be specified. The prob
lems of this section demonstrate, in the simplest form, various methods of solution for
planar problems of flow around rigid bodies: a) the method of singular points (Prob
lems 21.16—21.18), b) the method of conformal mappings (Problems 21.19—21.20),
and c) the method of mirror reflections (Problems 21.21—21.22).
— PROBLEMS —
0 21.12 Show that, for planar flows of an incompressible fluid, a) there exists
a function ip(x,y) (the stream function) such that vx = drp/dy, vy = —dip/dx;
b) A ^ = 0 in a potential stream, and W — <p + iip is an analytic function of the
complex variable z = x + iy.
~dW
0 21.13 Show that, for any analytic function W(z), the function —r— at any point
of the complex plane z determines a vector tangent to the curve ImW(z) = const
passing through that point (an overline denotes the complex conjugate). Give the
hydrodynamic interpretation of this statement.
0 21.14 Express, in terms of the stream function, the flow rate Q through a
curvilinear arc connecting the points with the coordinates Xi, j/i, and x^, y^.
© 21.15 Express, in terms of the velocity potential ip(x, y), the circulation T over
a curve connecting the points with the coordinates xit yi, and i 2 , 2/2 •
130 FLUID MECHANICS
© 21.16 Find the potentials and stream functions for the flows determined by
the complex potentials a) W = (Q/2ir) In z (a planar source or sink); b) W =
(T/2m) In z (a point vortex); c) W = zn, n > 0 (flow in a corner). Draw the
streamlines, and study the velocity fields. What is the physical meaning of the real
constants Q, T, and n?
© 21.17 Find the singular point, represent qualitatively the pattern of stream
lines, and write the equations for contours flow to which the following potentials
correspond:
a) W = u(z + a2/z);
b) W = uz + (Q/2TT) \n({z + a)/(z - a))
where u, a, Q are real constants.
© 21.18 Verify that the complex potential of flow around an infinitely long circular
cylinder, defined by the equation x2 + y2 = a2, has the form
W = U \z+ — \-iV [z + -— In - .
V z ) \ z J 2iri a
where U, V, T are real constants. What physical meaning do U, V, and V have?
© 21.19 Let an analytic function z = /(£) represent a conformal mapping of the

exterior of a contour C on the plane £ onto the exterior of a circle of radius a on the
plane z, with /(oo) = oo and (dz/dQ^ = k. Show that the complex potential of
Problem 21.18 at z = /(£) corresponds to the flow around the contour C in the plane
C with the velocity at infinity equal to k(U + iV).
Q 21.20 Find the complex potential of the flow around a plate of length 2b with
angle of attack a (see Figure 21.2). Choose the circulation T over a contour enclosing
a
u
Figure 21.2
the plate so that the velocity of the fluid on the trailing edge at the point B would be
finite (the Joukovski-Chaplygin postulate; sometimes it is called the Kutta condition).
Q 21.21 Find the complex potential of flow from a source with flow rate Q posi
tioned at a point A(XA,VA) a) near the plane wall y = 0; b) inside the right corner
x > 0, y > 0; c) outside the unmoving circle with center ZQ and radius a. For this
case, find the solution as the sum of potentials of the given source, a source of the
same intensity positioned at the inverse point, and a sink at the center of the circle.
0 21.22 Find the complex potential of the flow from a) an infinite row of sources of
equal intensity, equidistantly positioned along the x-axis at the points (0,0), (±a, 0),
(±2a,0), ...; b) an infinite row of vortices of equal intensity positioned at the same
points; c) a source positioned at the point (a, 0) between two planes x = 0, x = b;
d) a vortex positioned the same as the source of the preceding item.
Axisymmetric potential motions

A motion is referred to as axisymmetric if it is identical in all the planes passing
through a certain axis (in the meridian planes). For such flows, the velocity field does
not depend upon the coordinate determining the orientation of the meridian plane.
If, in addition, the velocity component corresponding to that coordinate equals zero,
then such an axisymmetric flow is called a flow without swirl, and, if it does not
equal zero, then this is a flow with swirl or a spiral flow. Axisymmetric flows with a
single-valued potential are flows without swirl. Let us direct the z-axis of a Cartesian
coordinate system x, y, z along the axis of symmetry. For flows without swirl, the
velocity field can be represented in the form
v{x, y, z) = curl (eip/r) , r — y x2 + y2
where ^(x, y, z) is the Stokes stream function, e is the unit vector at the point (x, y, z)
perpendicular to the meridian plane and directed to the side of growth of the azimuth
angle e (tane = y/x, 0 < e < 2ir) determining the orientation of the meridian plane.
In coordinate systems x',x 2 ,x 3 = e, the stream function of an axisymmetric flow
does not depend upon x 3 , i.e., ip = i()(xl,x2).
— PROBLEMS —
© 21.23 Show that, for axisymmetric motion, a) existence of a Stokes stream

function follows from the continuity equation, b) the surfaces rp = const are the
stream surfaces.
© 21.24 Express the physical components of the velocity and vorticity in terms
of the stream function T/>(X\X2) a) in a right orthogonal curvilinear coordinate sys
tem x 1 , ! 2 , ! 3 = z where x \ x2 are coordinates in a meridian plane, e is the angle
132 FLUID MECHANICS
determining the orientation of the meridian plane, b) in the cylindrical coordinate

system i 1 = z, x2 = r, x 3 = e; c) in the spherical coordinate system x1 = R, x2 = 6,
x3 = e.
0 21.25 Consider an axisymmetric flow with single-valued potential f. a) Show

that the streamlines are situated on the meridian planes, i.e., this is a flow without
swirl, b) Write the equations for the velocity potential (f and the stream function V
in the coordinate systems enumerated in the preceding problem.
© 21.26 Express, in terms of the stream function, the flow rate through the
surface formed by rotation around the z-axis of the curvilinear arc connecting the
points with the cylindrical coordinates z\,r\ and z^.r^ lying on a meridian plane.
G 21.27 The potentials
where R = y/x2 + y2 + z'2, n is an integer nonnegative number, and d°//dz° = f,

describe axisymmetric flows of an incompressible fluid. The corresponding stream
functions are
Verify these statements for n = 0,1,2. The potential <p corresponds to the flow from
a source with unit flow rate positioned at the origin of coordinate system when n = 0,
and to the flow from a dipole with unit moment and axis directed along the z-axis
when n = 1.
Q 21.28 Find the potential and stream function of the flow around a rigid sphere
of radius o, having velocity t ^ relative to the sphere at infinity. Write the solution
in the coordinate system a) connected with the sphere; b) in which the fluid rests at
infinity.
© 21.29 Find the potential for the flow of a fluid contained between an unmoving
spherical shell and a rigid sphere moving inside of it (let Rt,Ri be their radii) at the
instant when the center of the sphere coincides with the center of the shell, and the
velocity of the sphere equals U.
© 21.30 Find the potential and stream function of the flow outside of an un
moving impenetrable sphere of radius o created by the following, positioned at the
distance b > a from the center a) a dipole with moment (j. and axis passing through
the center of the sphere and the dipole; b) a source with flow rate Q.
Rotational flows of an ideal incompressible fluid

A flow of a fluid is referred to as rotational if u> = | curl v ^ 0. The vector u is called
the vorticity vector (or, simply, the vorticity). The Stokes theorem determines the
relation between the circulation T round a closed contour C and the vorticity vector
flux through any surface S bounded by the contour C, which is
r = f v ■ dl = I curl v • ndS .
c s
According to the Kelvin theorem, the circulation of the velocity over a material
closed contour in an ideal incompressible homogeneous fluid, in which the body force
fields have potentials, is conserved (does not depend upon time). Under the same
conditions, the vorticity vector satisfies the Helmholtz equation
— = (LJ ■ V)v .
— PROBLEMS —
© 21.31 Derive the Helmholtz equation for the vorticity vector u>.
0 21.32 Show that, for planar flows, the vorticity vector is perpendicular to the
plane of flow, w = we, and u = —Atp/2.
0 21.33 Consider a two-dimensional planar flow of an ideal incompressible fluid
for which there exists a potential of the body forces. Prove that a) the vorticity is
conserved in each particle, i.e., the equation dw/dt = 0 is valid, b) in a steady flow
w = u(ip) where ip is the stream function.
0 21.34 a) Prove that the stream function ij? = A(x2/a? + y2/b2) determines a
planar flow inside of an ellipse with a constant vorticity w = const b) Assume that
this is a flow relative to the coordinate system rotating with an angular velocity u>.
Find the velocity and vorticity of the relative motion.
0 21.35 The law of Biot-Savart
v(x,y,z) = ^ J ^ ^ , R = y/(x-02 + (y-v)2 + (z-Cr

c
determines the velocity distribution in the unbounded volume of an incompressible
fluid created by the single vortex line C (infinitely thin vortex tube); here, T is the
velocity circulation over a contour that encloses the vortex line once (see Figure 21.3).
Show that the flow from a rectilinear vortex line C is described by the velocity field
of the planar flow of the corresponding point vortex (see Problem 21.16).
134 FLUID MECHANICS
&l\&
M(x,y,z)
Figure 21.3
© 21.36 A fluid fills the dihedral corner formed by two mutually perpendicular
plane walls. Determine the trajectory of a isolated vortex line parallel to the edge of
the corner. Assume that the conditions of the Kelvin theorem are satisfied.
© 21.37 Let us consider a flow for which curl v is large in a thin layer of thickness
6. The surface to which that layer shrinks as 6 —» 0 is referred to as a vortex sheet if
lim<5 • curlu = 2fi

«—o
where fi is nonzero, finite, and lies in the plane tangent to this surface, fi is called
the density of the surface vortex. A vortex sheet arises, for example, in flow around a
wing of an airplane. Prove that, if some surface 5 is a) a vortex sheet, or b) a surface
of tangential discontinuity of velocity, then 2fi x n = [v] on it where [v] is the jump
of the velocity component tangential to S. Consequently, a vortex sheet is a surface
of tangential discontinuity of velocity, and the inverse statement is also valid.
0 21.38 Show that a) the velocity field of an arbitrary axisymmetric spiral flow
of an incompressible fluid in a cylindrical coordinate system x1 = 2, X ' r, x3 _
can be written in the form
v = curl (etj)(z, r, t)/r) + ew(z, r, t)
where e is the unit vector of the coordinate line e directed in the direction of increasing
e, b) the physical components of the vorticity vector are expressed by the formulae
1 djrw) 1 d(rw) = #V <?V Idxp

z r £
r dr ' r dz dz2 dr2 r dr'
b) under the conditions of Kelvin's theorem the following equations are satisfied
duz dujz du2 dvz dvz

at dz or az or
dur du)r dii)T dvr dvr
0 21.39 Show that, for axisymmetric flows with swirl, a) the surfaces ip = const
are stream surfaces, b) if the flow is steady, rw = const along the particle paths (r
and w are defined in Problem 21.38).
® 21.40 Show that, for a steady spiral flow of an ideal incompressible homoge
neous fluid, when there are no body forces, a) the following integrals of the equations
of motion are valid:
z p
where H(ip), G(ip) are arbitrary functions; b) the physical components of the vorticity
are expressed in terms of the function H and G by the formulae
dG n dG n „dG dH 2
and the equation for the stream function has the form
d2ip d2ip _ 1 3V> _ 2 dE_ _ dG_
dz1 dr2 r dr d4> dip
0 21.41 Hill's spherical vortex is an axisymmetric flow without swirl inside of a
sphere. The vorticity of the flow has only one nonzero component, the azimuthal one,
which is proportional to the distance from the axis of symmetry: ue = cr/2, c = const.
Obtain the stream function of this flow. Draw the stream surface %p = const
® 21.42 A closed vessel at rest filled with a nonhomogeneous incompressible fluid
is instantaneously set in translatory motion with velocity u. a) Show that, in gen
eral, a vortex motion arises in the vessel, b) For the case when the fluid is weakly
nonhomogeneous {{pm^ - pmin)/p - 6 < 1), find the vortex vector u> in the zeroth
and first approximations with respect to 6. c) Does the answer change if the fluid
viscosity is taken into account?
Integrals of t h e equations of motion for an ideal incompress

ible fluid
To obtain integrals of the Euler equations, it is convenient to write these equations
in the form of Gromeka-Lamb
dv v2 1 , _
— + grad - + 2 w x i ; = — gradp + F
ot l p
136 FLUID MECHANICS
If the flow of an ideal incompressible homogeneous fluid is potential, v = grady?.

(This is possible only if there exists a potential of the body forces F: F = gradt/.)
Then, in the whole region of the flow, the Cauchy-Lagrange integral
where f(t) is an arbitrary function of time, is valid. If a motion is steady, and there
exists a potential of the body forces F: F = grad U, then, along a streamline L (or
along a vortex line), the Bernoulli integral
P
U
2
--U
p
= u(L)
is valid. The quantity i» is constant along a streamline L. On different streamlines,
the values of i,, in general, are different.
Remark. The Cauchy-Lagrange and Bernoulli integrals take place also for a
compressible fluid under the condition that the flow is barotropic, i.e., p = p(p). In
this case, the term p/p must be changed by the pressure function V(p) = / dp/p(p).
For special motions, other integrals can exist (see Problems 21.40, 21.46).
— PROBLEMS —
© 21.43 In a water clock (see Figure 21.4), the height of level of the liquid in the
Figure 21.4
upper vessel serves as the time indicator. The height must decrease with a constant
speed. Determine the shape of the vessel used in water clocks.
0 21.44 a) Find the velocity and pressure distribution on the boundary of an
infinite circular cylinder flowed around with zero circulation. The velocity of the
stream at infinity is constant. Calculate the force R exerted on the cylinder by the
fluid, b) Find the same for a sphere. Use the answers to Problems 21.18, 21.28.
© 21.45 The pressure usually cannot be negative in real liquids. If the dynamical
conditions are such that the pressure decreases to a critical value Pd * 0 in any
region of a flow, cavities filled with the vapor and the gases dissolved in the liquid
appear in it. This phenomenon is called cavitation. a) Find the value of ambient
velocity for the stream uniform at infinity flowing around a sphere at which cavitation
begins. Indicate the place where it begins. Assume that the flow is unseparated
before beginning of cavitation, the pressure in the stream far from the sphere equals
atmospheric pressure, and the pressure in the cavity equals zero, b) Find the same
for a circular cylinder when the flow around it is noncircular.
0 21.46 Derive the integral of the equations of motion for steady planar flow of
an ideal incompressible fluid a) with constant vorticity
v2 v
h - - U + %JI\> = const
2 p
b) with the vorticity u = Xtp, A = const
n2 v
—+ U + uip = const,
2 p
where U is a potential of the body forces, and i\> is stream function.
© 21.47 A spherical domain of radius o0 is abruptly excluded from an ideal

incompressible liquid filling the whole space. Derive the equation determining the
motion of the cavity boundary. Determine the time for which the cavity is filled with
the liquid. Regard the density of the liquid and the pressure at infinity as known.
0 21.48 Show that the velocity field with the potential
ip = c[x — a exp(ky) sm(kx)] , c, a, k are constants
describes a flow of an incompressible fluid. Find the complex potential of this flow.
With what accuracy is the curve y = a cos (kx) a streamline when ak < 1? At what
value of c does this curve describe the profile of a wave on the surface of a heavy
liquid of infinite depth? What is the physical meaning of the parameters c, a, and fc?
The y-axis is directed along — g.
© 21.49 Large-scale air bubbles rising to the surface are known to have the shape
of a spherical segment ABC. (The flow pattern around the bubble is shown on
Figure 21.5.) Show that the speed U with which this bubble rises is related to the
radius of curvature of the spherical segment R by the formula U = 2y/gR/3.
138 FLUID MECHANICS
A
Figure 21.5
Figure 21.6
© 21.50 A liquid in the shape of a "tongue" moves under action of the gravity
force over a sloping plane, immersing into a more light-weight liquid which is at rest
sufficiently far from the plane (see Figure 21.6). The flow is two-dimensional and
steady in the coordinate system moving together with the tongue front. Show that
the tangent to the interface between the two liquids at the leading point of the tongue
makes the angle a — 60° with the plane.
© 21.51 A progressive wave of a stationary shape with rectilinear crests on a free
Figure 21.7
surface of a heavy liquid moves with as large as possible amplitude. Under these
conditions, such a wave is known to have, near a crest, the shape of a wedge with its
edge situated on the crest and its two faces symmetric relative to the vertical. Show
that the angle a between the faces of this wedge equals 120° (see Figure 21.7).
Force and moment acting on a body in a s t r e a m of an ideal

incompressible fluid
The formulae for the force F and moment Af acting on a body in a stream of an
ideal fluid have the form
F= - fpndS , M = - JprxndS
s s
where S is the surface of the body, n is the normal to it external relative to the body,
r is the radius-vector of the points of the surface emanating from the point relative
to which the moment M is calculated. For potential streams, application of the
Cauchy-Lagrange integral allows these formulae to be reduced to a form convenient
for calculations (see Problems 21.52-21.54). For planar problems of flows around
contours (similar to the notion of complex velocity vx + ivv), the complex number
R = X + iY representing the force vector acting on a contour C is introduced, where
X = - §pnx dl,Y = - §jmy dl. For steady planar flows, if there are no body forces,
c c
the formulae of Blasius-Chaplygin
R = X-iY=l-£ fv2dz, L=Rel-?<fv2zdz\
c \ c /
are valid; here, L is the value of the moment of forces acting on the contour relative
to the coordinate origin, z = x + iy, and v = vz - ivv. These formulae allow the easy
calculation of the force and moment for potential flows with known complex potential
W(z). Thus, the formula of Joukovski
R = ipTVoo
where T is the velocity circulation round the contour C, v^ = ^(oo) - ivv(oo), is
valid for the force.
— PROBLEMS —
0 21.52 For flows with single-valued potential <p, a) derive the following identities
(E is an arbitrary closed material surface)
/(MH-5/- JS+ /(T'—)«•

140 FLUID MECHANICS
= — [<p(r xn)dS + f (^{r x n) - {r x v)vn) dS ,
b) prove that, for an incompressible fluid the last integrals on the right hand sides of
the formulae do not depend upon the shape of the surface E, and it can be arbitrarily
deformed in the region where the flow is regular.
0 21.53 Using the Cauchy-Lagrange integral and the identities of Problem 21.52,
derive formulae convenient for calculation of the force and moment acting on a body
in a potential stream of an ideal incompressible homogeneous fluid when there are no
body forces
dQ
+ IF, Q = -PJipndS, J> = pi ( — n-vvn\ dS;
-dT s s' ^ '
M = -— + IM, K = -pjf{rxn)dS ,
s
IM = PJ (y(»" xn)-{rx v)vn) dS
where S is the surface of the body, n is an outer normal to S, S' is an arbitrary surface
formed by a continuous deformation of S in the region where the flow is regular. Here,
the potential tp is implied to be single-valued.
0 21.54 For the case of a fluid at rest at infinity, show that a) the formulae of
Problem 21.53 take on the form F = -dQ/dt, M = -dK/dt (the vectors Q and K
are called the apparent momentum and apparent angular momentum of the motion);
b) if the body performs translatory motion with velocity U, then -Ft = HikdUk/dt
where n,k (apparent mass coefficients) depend upon only the shape of the body:
P-tk = —P i fkîdS, and ipk is the solution of the following Neumann problems:
s
Aip/t = 0 everywhere outside of the body, grad^loo = 0, {d<fk/dn)\s = nk where
S is the surface of the body, and n* are the components of the normal to S in the
coordinate system connected with the body. Here, the potential is implied to be
single-valued.
© 21.55 Prove d'Alembert's paradox: if a body performs a translatory motion at

constant velocity in an unbounded, ideal, incompressible fluid at rest at infinity, the
force acting on it equals zero when the flow is unseparated.
0 21.56 Prove that, if there are no body forces, a) the force dX + idY acting on
the element dz = dx + i dy of a contour, around which a stationary potential stream
of an ideal incompressible fluid flows, is represented in the form
/ 1 dW dW\
dX + idY = p[h--—--—]idz
\ 2 dz dz )
where h is the constant in the Bernoulli integral, W(z) is the complex potential,
the overline denotes complex conjugation; b) ^ dz is a real number on the contour
around which the flow occurs c) the formula of Blasius-Chaplygin
1 t(dW\2
x + ir = -ipf[-) dz
is valid for the force acting on the contour C.
© 21.57 Calculate the force exerted by the fluid on a unit length of an infinite
circular cylinder moving perpendicularly to its axis with a velocity it if a) u — u(t),
T = 0, b) u = const, |T| = T0 = const, (r is the velocity circulation over a contour
enclosing the cylinder). Assume the flow around the cylinder to be unseparated, and
neglect the gravity force.
0 21.58 Calculate the force acting per unit width (along the normal to the plane
of the flow) of a plane plate of a length 26 around whichflowoccurs without separation
with an angle of attack a. The velocity of the fluid at infinity equals u. Neglect the
gravity force.
0 21.59 Calculate the force exerted by the fluid on a sphere moving through it
with velocity u, if a) u = u(t), b) u = const. Assume that the fluid is at rest at
infinity, that the flow around the sphere is unseparated, and that the gravity force is
negligible.
© 21.60 Calculate the force exerted on an unmoving sphere by the fluid flowing
around it (without separation) if, at infinity, the stream is uniform, and its velocity
Uoo is a given function of time. Neglect the gravity force.
® 21.61 Prove the theorem of Lagalli: a source with the mass flow rate pq = const
acts on an unmoving body in the fluid with the force pqv' where v' is the regular part
of the velocity of the fluid at the point where the source is positioned.
0 21.62 What is the acceleration of a spherical gas bubble as it begins to rise

toward the surface of a heavy liquid?
142 FLUID MECHANICS
© 21.63 A spherical bubble is set in motion by a moving liquid. What is its

acceleration if the acceleration of the fluid far from the bubble equals to? Neglect the
gravity force.
22 Dynamics of an Incompressible Viscous Fluid

For incompressible viscous (when p. = const) fluids, the problem of the determination
of mechanical motion (i.e, calculation of v(xl, t), p(x\ t), p(xf, t)) involves a solution
of the system of equations consisting of the condition of incompressibility, the conti
nuity equation and the Navier-Stokes equations. The equation of energy is used to
determine afterwards the temperature T(xl,t) if necessary. However, if the depen
dence of p, upon the temperature is significant, the closed system of equations must
also include the equation of energy. Constructing the solutions of the systems of
equations describing a motion of a viscous fluid is, as a rule, a complicated problem.
So, various approximations of these equations have broad applications. To clarify the
meaning of the commonly used approximations, it is convenient to write the equations
in a dimensionless form, relating all the quantities in them to some of their certain
typical values. In particular, for a homogeneous incompressible linear-viscous fluid,
the system of equations in a dimensionless form is
,. - ,. dv ,_ - . _ ~. Av
divv = 0 , S t ^ + (u- V)u = - V p + — -
at Re
where i = t/t%, v = v/vt, xl = x'/L, p = p/pv^, Re = v,L/u is the Reynolds

number, St = L/t,v. is the Strouhal number, v., t., L are the typical speed, time,
length, and v = p/p is the kinematical viscosity coefficient. If Re -C 1, St Re <C 1,
then the "inertial terms" (i.e. acceleration) can be neglected compared to the terms
associated with viscosity. This is the way the equations corresponding to the Stokes
approximation are derived. If Re S> 1, then the "viscous terms" in the equations
are relatively small. However, they may not be neglected completely for the whole
region of a flow with preservation of the boundary conditions, because, in general, it
is impossible to find the solutions of the Euler equations satisfying these boundary
conditions. At high Re, there are boundary layers, ordinarily relatively thin, near the
boundaries. Inside these layers viscosity substantially affects the flow, and outside,
it may be omitted. For description of the motion in a boundary layer, the Navier-
Stokes equations can be replaced (since the layer is thin) by the simpler boundary layer
equations. Another phenomenon, arising when the Reynolds number is increased, is
the loss of flow stability. At sufficiently high Reynolds numbers, flows, as a rule, have
more complex behavior: all the characteristics chaotically pulsate on the background
of some regular values. This phenomenon is called turbulence. To describe turbulent
22. Dynamics of an Incompressible Viscous Fluid 143
motions the so-called semi-empirical equations for averaged values of parameters are
often used instead of the Navier-Stokes equations.
— PROBLEMS —
General properties of flows of incompressible viscous fluids

© 22.1 Derive the equation describing the variation of kinetic energy of a material
volume of viscous fluid d£kin = dA(&) + dA® where dA(e) and cL4w are the work of
external and internal forces, respectively. Why, for an incompressible fluid, is the
quantity Vdt= -<L4(i) called the dissipation of mechanical energy?
0 22.2 A heavy, homogeneous, viscous fluid completely filling a cavity in an
unmoving solid body is set in motion and then released. Prove that the kinetic
energy of the fluid decreases as a function of time.
0 22.3 A heavy, homogeneous, viscous fluid completely fills a cavity in a solid
body rotating with a constant angular velocity fl. The fluid is set in motion relative
to the walls of the cavity and then released. Prove that the kinetic energy of relative
motion of the fluid decreases as a function of time.
0 22.4 An unmoving heavy fluid fills a cavity in a solid body at rest. Can the
fluid be made to move relative to the walls of the cavity at the expense of arbi
trary translatory motions of the body? Consider the cases of an ideal and viscous,
homogeneous and nonhomogeneous fluids.
0 22.5 Prove the following identities: a) 2etJe,j = 2di\{dv/dt) + |curlw|2 where
eij = \(ViVj + VjVi); b) V = fi J \cur\v\2dV + 2fi J(dv/dt)ndS where V =
V dV
1 2êijeij dV is the rate of energy dissipation in a volume V with boundary dV.
v
If the flow is steady, then c) V = \i J \cui\v\2dV + 2/i J K{n ■ m)v2 dS where K,
v dv
m are respectively the curvature and the vector of principal normal to the particle
path, n is the external normal to dV\ d) V = /u j |curl v\2 dV + \i J (dv2/dn) dS +
V 3V
2/x / ((curlu x v) ■ n) dS.

dv
© 22.6 A body moves with constant velocity in a viscous fluid, and there are no
body forces. Show that the projection of the force acting on the body from the side
144 FLUID MECHANICS
of the fluid onto the direction of the velocity is negative.
0 22.7 A resting solid sphere of radius R begins to move in a viscous fluid with
a constant acceleration a. Prove that the average value over a path S of the drag
force is more than the product of the apparent mass of the sphere and its acceleration
(assume that there are no body forces), i.e.,
© 22.8 Show that, when a solid body moves in a viscous fluid, the force exerted
on an element dS of its surface by the fluid equals — (pn + fin x (curlv — 2f2)) dS
where n is the normal to dS external relative to the body, (2 is the angular velocity
of the body.
© 22.9 Derive the following expression for the tangential stress on a stream surface
with normal n: pnT = /x(((curl v x n) ■ T) + 2K \v\ (n • m)) where K, T, m are
respectively the curvature and the vectors of tangent and of principal normal to the
streamline.
0 22.10 Using the II-theorem (see Section 36), prove that the magnitude of the
drag force in a steady stream of a viscous fluid has the form F = pv^d2 C(Re) where
Re = Vood/u, v = fi/p, d is the characteristic size of the body, i>oo is the speed of the
stream at infinity, C(Re) is a dimensionless function of the Reynolds number Re and
the form of the body. Assume that there are no body forces.
U n s t e a d y flows
© 22.11 A viscous fluid fills the half-space y > 0 bounded by the plane y = 0 of
a horizontal solid plate. The plate performs the harmonic oscillation along a straight
line in its plane with a frequency fi. Find a) the velocity profile, b) the tangential
stress on the plate, c) the time-averaged value of dissipated energy.
© 22.12 A viscous fluid is contained between two infinite horizontal plates A and
B. The plate A oscillates along a straight line in its plane with a frequency fi, while
the plate B is at rest. What is the ratio of the maximal values of tangential stresses
acting on the plates A and B in the cases of small and large n? The distance between
the plates equals h.
© 22.13 For the system described in Problem 22.11, find the velocity distribution
in the fluid if the plate is at rest for t < 0 and, for t > 0, moves with the velocity
&)U = U0 = const, b) U = U(t).
0 22.14 Consider a space filled with a viscous fluid. At t = 0, let the velocity
distribution have the form vv = vz = 0, vx = U at y > 0, vx = —U at y < 0, i.e., the
plane y = 0 be the surface of tangential discontinuity, a) Find the velocity profile at
t > 0, and show that the tangential discontinuity instantaneously "spreads" in the
space, b) How does the thickness of the layer, in which the velocity changes from
vx = 0.99f/ to vx = -0.99 U, vary in time? c) Calculate the vorticity.
© 22.15 A viscous incompressible fluid occupies the half-space y < 0 and is at

rest at the initial instant. At t > 0 on its surface y = 0, a constant tangential stress
T0 acts. Determine the motion of the fluid (this is the simplest model of the flow of
water in the layer near the surface under action of a wind).
© 22.16 A viscous incompressible fluid occupies the half-space y < 0. There is no

action of surface forces on the boundary of the fluid at y = 0. At t = 0, vy = vz = 0,
vx = uoexp(-fa/ 2 ), uo, b are constant. At t > 0 find a) the velocity distribution,
b) the kinetic energy Eun(t), and the momentum Q(t) of the fluid in the layer of unit
width, c) the values of i ^ n and Q in the limit at t —► oo.
0 22.17 In a spherical layer Ri(t) < R < fl2(<) of a viscous fluid, a spherically
symmetric flow takes place under action of the pressures pi(t) = p(Ri) and P2M =
p(ft 2 ), given on the boundaries of the layer. The mass of the fluid in the layer equals
M. Write the system of equations determining the motion of the boundaries of the
layer Ri and R2. Assume that there are no body forces.
© 22.18 Describe the spherically symmetric flow of a thin spherical film of a vis
cous fluid in vacuum. Find the maximal and the minimal values of R/RQ as a function
of the initial Reynolds number RoRo/v where v = \ijp is the kinematical viscosity,
R{t) is the radius of the film, Ro = R\t=0l Ro = (dR/dt)\t=0.
© 22.19 At a point M of an unbounded space filled with a resting liquid, an

explosion takes place so that kinetic energy of level E0 is instantaneously produced
and transmitted to the liquid. Determine the rising motion of the liquid, regarding
it as linearly viscous. Neglect the compressibility of the liquid, the gravity force, and
the value of pressure at large distances from the point M. Compare with the solution
of a similar problem for an ideal liquid (Problem 9.22).
Steady flows
0 22.20 A layer of viscous liquid of width h is bounded from above by a free
surface and from below by an unmoving plane inclined at an angle a to horizon.
Under the action of a gravity force, a steady flow takes place in the layer. Find the
velocity distribution in the layer, the values of vmax and uav, the velocity maximum and
146 FLUID MECHANICS
its average over a cross section, respectively. Using the solution obtained, estimate
the values of vmax and vm when water (v — 0.01 cm 2 /s) flows in a channel of length
/, the difference between input and output heights H, and depth h if a) / = 100 m,
H = 1 cm, h = 0.5 cm; b) I = 3000 km, H = 300 m, h = 5 m (a model of Volga
river). Why does the result in the case b) obviously contradict experimental results?
0 22.21 A layer of viscous fluid is bounded by two horizontal infinite parallel

plates A and B, the distance between which is fixed at H. Find the distribution of
velocity and stress of friction forces TA, TB on the plates if a) the plate A is at rest,
while B moves with a velocity U in its plane; assume that the pressure drop along
the plates equals zero (Couette flow); b) both plates are at rest, and the motion of the
fluid is caused by a given pressure gradient along the plates (plane Potseuille flow);
c) there is both a motion of the plate B and a pressure gradient along U.
Q 22.22 Determine the steady motion of a viscous incompressible fluid in a long

horizontal cylindrical tube under action of a given longitudinal pressure drop i0 =
-dp/dx (Poiseuille flow) if the cross section of the tube is a) a circle of radius a; b) a
circular annulus with internal and external radii a and b; c) an ellipse with semiaxes
a and b. For all the cases, calculate the flow rate Q, maximum vmBX, and average
velocity over a cross section vav. For flow in tubes with elliptic cross sections, show
that Q is maximal when a = b for a given i0 and cross-sectional area.
0 22.23 For the system of Problem 22.22, derive the formulae for Q,
for a tube of an arbitrary cross section of a given area, using dimensionality theory.
0 22.24 A viscous fluid is contained between two infinitely long coaxial cylinders
rotating along the axis with constant angular speeds £li and fi2 (the rotational Cou
ette flow). The pressure gradient along the axis equals zero. The radii of the cylinders
equal Ri and R2 (R2 > Ri)- Find the velocity distribution and the moments of the
forces exerted on the cylinders by the fluid. Consider the limiting case when Cl2 = 0,
R2 —► 00. Neglect the gravity force.
© 22.25 Solve Problem 22.24 under the additional condition that there is a given
pressure gradient along the axis. Consider separately the case when the gap between
the cylinders is small.
© 22.26 Show that the necessary and sufficient condition of stability for the ro
tational Couette flow considered in Problem 22.24, relative to axisymmetric inviscid
disturbances, has the form d(y? 2 r 4 )/dr > 0 where >p is the angular speed of the fluid,
r is the distance from the rotation axis. What can one say about stability of this flow
if the cylinders rotate a) in different directions, b) in the same direction, c) one of the
cylinders remains at rest?
Stream function and vorticity

The vorticity vector w of a viscous incompressible homogeneous fluid in potential
field of body forces satisfies the Helmholtz equation
<&*> i - , \ A 1
— = (a; • V) v + v Aa; , w = - curl v .

at 2
For a planar flow and an axisymmetric flow without swirl, the velocity field of the
incompressible fluid can be expressed in terms of the scalar stream function ip. In the
first case; v = curl (eip), and in the second one, v = curl (eip/r). Here, e is the unit
vector perpendicular to the planes of the flow which are parallel for a planar flow and
have as a common line the axis of symmetry for an axisymmetric flow, and r is the
distance from the axis of symmetry. In both cases, the vorticity is directed along the
vector e, curlv = 2uie. For the problems of this subsection assume that the fluid is
homogeneous, the body forces have a potential, and there is no swirl.
© 22.27 Show that the vorticity in a planar flow of a viscous incompressible fluid
satisfies the equation
du) .
-— = vAui
dt
(a particular form of the Helmholtz equation).
© 22.28 Derive the following equations for the stream function and vorticity in a
planar flow of a viscous fluid: dwjdt + d(u, tp)/d(x, y) = uAu, w = —Aip/2 where
d(oj, TJj)/d(x, y) = —— — -T— -r—, x, y are Cartesian coordinates in the plane of the
ox ay ay ax
flow.
0 22.29 Derive the following equation for the vorticity and stream function in an
axisymmetric flow of a viscous fluid:
dw d(uj/r,ip) _ v d fr3 d$u/r)\ + d_ ( 3 9( w /r)
dt d(z, r) r2 dz \ dz I dr \ dr
where z and r are cylindrical coordinates.
0 22.30 Let the vorticity in a planar flow of a viscous fluid depend linearly upon
the stream function: 2u = \ip. The stream function at the initial instant t = 0 is
^0(^,2/)- Determine the stream function for an arbitrary instant.
© 22.31 For a viscous fluid, the tendency to equalize the vorticities of different par
ticles inside of fluid is a characteristic feature (diffusion of vorticity). As an example,
determine the vorticity and velocity distributions at t > 0 in a viscous incompressible
fluid filling the whole space if, at the instant t = 0, there is a line vortex on the axis
z = 0 with finite circulation IV
148 FLUID MECHANICS
© 22.32 Show that the vorticity cannot reach both maximal and minimal values
at an internal point of the region of a planar steady flow of a viscous incompressible
fluid.
© 22.33 Show that the quantity ui/r cannot reach both maximal and minimal
values at an internal point of the region of an axisymmetric steady flow of a viscous
incompressible fluid outside of the axis of symmetry.
© 22.34 In the linear stability theory, stability or instability of steady flows is

determined by the evolution of small disturbances in the parameters describing these
flows, a) Derive the linearized equation for small disturbances T/>I (X, y, t) of the stream
function in a steady planar flow of a viscous incompressible fluid with the velocity
profile vx = U(y), vy = 0 where vx and vv are the Cartesian components of the velocity.
b) Show that this equation has a solution of the form tpi(x,y,t) = t/>(y) exp(ifcx—iut)
where xp(y) satisfies the Orr-Sommerfeld equation
v{JT - 2k2i>" + k4i>) = ik ((u - | ) {i>" - k2i>) - U"4>) .
Formulate the boundary conditions for ip{y) corresponding to the flow between the
plane solid walls y = ±d.
Low Reynolds n u m b e r flows. Stokes approximation

When Re = vhjv <C 1, steady flows of a viscous incompressible homogeneous fluid in
a potential field of body forces are often described by the linear equations (the Stokes
equations)
divu = 0 , gradp' = /J,AV
where p' is the modified pressure including the potential U of body forces (p' = p—pU).
Q 22.35 Prove that, on the set of solenoidal velocity fields v' (divu' = 0) de
fined in a domain V with fixed values on the boundary (v'\av = f), the functional
I(v') = 2/i / e\ e"J dV reaches a minimum at the velocity field v satisfying the Stokes
v
equations (the theorem of Helmholtz).
© 22.36 Prove uniqueness of the solution of the following boundary-value prob
lems for the Stokes equations: a) the velocity is given on the boundary; b) the stress
vector is given on the boundary; c) the velocity is given on a part of the boundary,
and the stress vector, on the remaining part (mixed problem); d) the tangential stress
is given on the boundary stream surface; e) the tangential stress is given on a part of
the boundary stream surface, and the tangential velocity, on the remaining part.
0 22.37 Let F, M and F', M1 the forces and moments acting on a body when it
slowly moves in a viscous fluid with translational and angular velocities respectively
equal to U, fi and U', fi'. Prove that U ■ F' + fi ■ M' = U' ■ F + & ■ M (the
theorem of reciprocity).
0 22.38 Using the If-theorem (see Section 36), prove that the magnitude of the
drag force acting on a body performing a slow translational motion in a viscous fluid
is determined by the formula
F = Afj.lv
where v is the speed of the body, / is the characteristic size of the body, A is a
dimensionless coefficient depending upon the shape of the body and the direction of
its motion.
© 22.39 Prove that the exact magnitude of the drag force for a body moving in a
viscous fluid is not less than that calculated in the Stokes approximation.
0 22.40 Show that the formulae for the components of the force F exerted on
a body by a viscous fluid in which it is performing slow translational motion can be
written in the form Ft = Aijfiuj where v? are the velocity components of the body
in a coordinate system attached to it, and A y are the components of a symmetric
tensor independent upon the velocity of the body.
0 22.41 Prove that, on the set of solenoidal velocity fields v' (divu' = 0) defined
in a domain V, when the values of the stress on the boundary (p^n^y = /*) are
fixed, the functional /(«') = 2[i / e - e"-> dV — 2 / v•/' dS reaches a minimum at the
v dv
velocity field v satisfying the Stokes equations.
0 22.42 Let a solenoidal velocity field v (divu = 0) be defined in a domain V
and satisfy the following conditions on the boundary dV: vn\gV = 0, vT\aVl = 0,
Pnr\dv2 - / i 9V = dVi U dV2. Prove that, when / is fixed, the functional I(v) =
2/x / e u e u dV— 2 / v • pnT dS defined on this vector field reaches a minimum if the
v av
field satisfies the Stokes equations.
© 22.43 A steady flow in a domain V is caused by the action of tangential stresses
p n T given on a part of the boundary. On the rest part of the boundary, the adhesion
condition is satisfied (the velocity equals zero). The whole boundary is a stream sur
face. Prove that, if the nonlinear terms in the Navier-Stokes equations are neglected,
the resulting equations yield excessive values for the energy dissipation and the work
of surface forces.
150 FLUID MECHANICS
© 22.44 Find the velocity and pressure fields in an axisymmetric creep flow of a
viscous fluid (Re <§C 1) between two parallel planes coming together with a relative
speed 21/ for the instant when the distance between them equals 2ft. Seek a solution
of the form vr = r f(z), vz = g(z).
Q 22.45 Two circular co-axially positioned disks of the same radius R are im
mersed into a viscous fluid and slowly come together with a relative speed 2U. De
termine the drag experienced by the disks when the distance between them, 2ft, is
small.
0 22.46 The axisymmetric slow spreading of a viscous incompressible liquid in a

thin layer can be described by the approximate equations
dp = dV 9p d(rvr) d{rvz) =
dr dz2 dz dr dz
Justify this approximation. Solve Problem 22.45 using this approximation.
© 22.47 Prove that, in the Stokes approximation for a planar flow of a viscous
incompressible fluid, a) the stream function ip(x, y) satisfies the biharmonic equation
AAt/> = 0, b) ip(x, y) and the pressure p(x, y) can be expressed in terms of two analytic
functions H(z), G(z) by the formulae i>(x,y) = zH(z) + zH(z) + G(z) + 72(z), p =
4in(dH/dz — dH/dz) where z = x + iy, and the overline denotes complex conjugation.
0 22.48 Prove that, if H(x,y,z), G(x,y,z) are harmonic functions, then v =

grad H + z grad G - kG, p = 2fidG/dz + p0 (where z is a Cartesian coordinates,
k is the unit vector along the axis z, po = const) satisfy the Stokes equations, and
curl v = 2fc x grad G.
© 22.49 A slow stream (Re <S 1) of a viscous incompressible fluid flows around
a sphere of radius a. Far from the sphere, the velocity of the stream equals U, and
the pressure, po- Using the Stokes approximation and the form of the solution indi
cated in the preceding problem, find the velocity, vorticity and pressure distributions.
Calculate the dissipated energy. Derive the formula for the drag force F (the Stokes
formula). As H and G, take the functions Ad{l/R)/dz and B/R + D, where A,
B, and D are constant, R and 6 are the spherical coordinates with the origin in the
center of the sphere, and the axis z is parallel to U.
© 22.50 Show that, if the field of body forces is solenoidal (divF = 0), the
pressure at an internal point of a region of creep flow of a viscous incompressible fluid
cannot reach both the minimal and maximal values.
High Reynolds flows. Laminar b o u n d a r y layer

When Re 3> 1, the whole region of a flow can be divided into two parts: Vi — thin
boundary layers near the boundaries in which viscosity must be taken into account
and V2 — the other part where the fluid may be regarded as ideal. Since the thickness
of the boundary layer is small, it is possible a) neglect its existence when solving the
problem in the domain V2, i.e., to assume that V2 coincides with the whole region
of the flow V, and b) use approximate equations for the domain Vi instead of the
complete Navier-Stokes equations. The equation for a plane laminar boundary layer
on a surface S in an incompressible fluid (Prandtl equations) have the form
du dv _ du du du _ 1 dp d2u dp
dx dy dt dx dy p dx dy1 ' dy
where u, v are the components of the velocity in the orthogonal coordinate system
in which the surface S is given by the equation y = 0 and whose coordinate lines
x = const are directed along the normals to S. Taking into account the last equation,
the pressure in the second equation should be regarded as a known function of 1 and
t determined as the solution for the pressure on S in the corresponding problem for
the ideal fluid; p(x, t) is related to the speed U of the ideal fluid on S by
— U— = --^-
dt dx p dx
The boundary conditions for the system (22.1) are composed of the ordinary condi
tions on S, identical to those for a viscous fluid (see Section 19), and the condition
u —> U when y —> 00.
0 22.51 Using the fact that the values of the terms representing the acceleration
and the force caused by viscous stresses are of the same order, estimate the width S
of the boundary layer on a body with a characteristic size L moving with a constant
speed U (in a boundary layer, the characteristic measures along x, y are L, 6) if
a) L = 100 m, U = 36 km/h, v = 0.01 cm 2 /s (a ship on water); b) L = 50 m,
U = 900 km/h, v — 0.15 cm 2 /s (an airplane in air).
0 22.52 A steady stream with a speed U = const flows around a semi-infinite

plate, the coordinate in the long direction being oriented along the velocity (the
problem of Blasius). Show that the stream function for the flow in the boundary layer
can be written in the form rp = y/vUx f{rj) where 77 = yJu/(vx). The coordinate
origin is positioned on the leading edge of the plate, and the z-axis is directed along
the plate. Obtain the equation for /(»?) from the Prandtl equations. What boundary
conditions should be set for f(r))1
152 FLUID MECHANICS
0 22.53 Derive from the Prandtl equations the integral equation of momentum
v
dn\ r*J{ )y+
* >v=o o
+ J{(U ~u)^ + ^[u(U -u)] + ^[v(U -u)}^ dy .
0 22.54 The width of displacement 6i and the width of momentum loss 8 are
defined by the formulae 6X = / ( l - u/U) dy, 8 = f u(l - u/U)/Udy. Show that the
o o
integral equation of momentum (see Problem 22.53) can be written in the form
M = djusj) | du t d(u2o)
dy] n dt dx dx
© 22.55 The velocity profile in a boundary layer is given by the relationships

u = U sin (ay) at 0 < ay < TT/2, U = U at ay > 7r/2, where a = a(x) and
(/ = const. Find the width of displacement 6i and the width of momentum loss 9
(see Problem 22.54).
© 22.56 Use the velocity profile of Problem 22.55 and the integral equation of
momentum for approximate calculation of the magnitude of the tangential stress r
on the surface of the plate in the problem of Biasius. Compare the obtained value
of r with the value 0.332 Jp(iU3/x found by a numerical solution of the equation for
f(rj) (see Problem 22.52).
0 22.57 a) Show that, when U = Cxm, x > 0 (C > 0 and m are constant), the
stream function of the flow in a boundary layer has the form tp = \lvUx f(r]), r) =
V \JUI{UX)- b) Obtain from the Prandtl equation the equation for /(»?). c) Around
what bodies does this class of flows describe? d) What boundary conditions must be
set for the function /(T?)?
© 22.58 a) Determine the vorticity on the free surface bounding a planar steady
flow of a viscous fluid, b) Show that, in the boundary layer near the free surface, the
viscous addition u' to the speed U of flow for the ideal fluid is a small quantity of
order (//\/Re as Re —» oo (this is invalid for a boundary layer near a rigid surface).
® 22.59 Show that determination of the vorticity in the plane boundary layer near
the free surface y = 0 is reduced to solution of the following boundary-value problem:
du/dt + Udcu/dx - y{dU/dx){du/dy) = ud2u;/dy\ u = (curl») x /2, u>|v=0 = KU,
bj\v-.oo —» 0 where K is the curvature of the streamline on the free surface.
0 22.60 Using the answers to Problems 22.58 and 22.59, show that determination
of the viscous component u' in the boundary layer near the free surface y — 0 is
reduced to solution of the following boundary-value problem: du'/dt + u'dU/dx +
Udu'/dx - y(dU/dx)(du'/dy) = v&u'/dy2, (du'/dy)\v=0 = -2KU, u' -► 0 when
y —> oo. Prove that the resulting linear equation for the boundary layer near a free
surface is an asymptote for the Prandtl equation at Re —> oo.
0 22.61 A spherical gas bubble of a radius o moves in a viscous liquid with con
stant speed U. a) Calculate the dissipation of energy in the liquid, assuming the flow
to be potential, b) What is the accuracy of this approximation when Re 3> 1? c)
Calculate the viscous drag force.
Turbulent motion
If the characteristic speed v, line size L and viscosity v for a flow are such that
the Reynolds number Re = vL/u is more than a certain critical value Recr, the
flow is turbulent. The fields of all the mechanical (and thermodynamical) quantities
in a turbulent stream chaotically pulsate on the background of average values of
the characteristics. If a flow is not turbulent, it is referred to as laminar. The
precise description of turbulent flows is an excessively difficult problem. For practical
needs, it is often sufficient to calculate only the average values of the characteristics.
For this purpose, so-named semi-empirical theories of turbulence are used. These
theories are constructed ordinarily in the following way. The quantities characterizing
a flow are written in the form a = d + a' where a is the average value, and a' is the
pulsation. There are various definitions of a (average over time, over volume, in
the sense of probability, etc.), but it is supposed to have the following properties
(da/dt) = da/dt, (da/dx) = da/dx, a + b = a + b, ab = ab + a'b'. Next, the system of
equations is averaged with using these properties. The resulting system contains the
average velocity v, pressure p, density p, etc., and also, as new additional unknowns,
average products of the pulsations of these quantities. To obtain a closed system,
additional equations or relationships determining the mentioned new unknowns must
be added. These equations are postulated in a certain form, and their coefficients
then determined by comparison of the calculated quantities with those measured in
an experiment. At the present time, many various semi-empirical theories have been
proposed to describe various classes of turbulence (flows in boundary layers, in tubes,
turbulence in the ocean and atmosphere, etc.).
© 22.62 It is known that, if special conditions are not fulfilled, a flow is turbulent:
in open channels of depth h when vh/v > 500; in circular tubes of diameter d when
vd/v > 1300; in streams flowing around blunt bodies with nose radius r when vr/v >
105. Is there turbulence in the flow a) of water in a plane river (v = 1 m/s, h = 5 m,
154 FLUID MECHANICS
v = 0.01 cm 2 /s)? b) of water in a waterline tube (v = 2 m/s, d = 2.5 cm)? c) of air

near a spacecraft returning into the atmosphere when 1) v = 7 km/s at the height
from the Earth surface H = 75 km, 2) v = 2 km/s, / / = 40 km? For the standard
atmosphere, p = Po exp (-0.142//) where po = 1.225 • 10"3 g/cm 3 , # is measured
by km; v = fj,/p, p = 1.3 • 10"4 g/cm • s (corresponds to the temperature 200°K);
r = 1 m.
© 22.63 Averaging the continuity equation and motion equations for an incom
pressible homogeneous fluid, derive the Reynolds equations
divv = 0 , p-j± = -Vip + V-'fy + V}T*
where fy are the average viscous stresses, and T ? = — pv^v'j are called the turbulent
Reynolds stresses.
Q 22.64 Demonstrate the analogy of the mechanisms of turbulent viscosity and
ordinary (molecular) viscosity of gases, interpreting r? as the momentum transfer
through surface elements, moving with the velocity of the medium v, via chaotically
moving particles.
© 22.65 Assuming that T? = 2pvTci) (the hypothesis of Boussinesq) and that
the coefficient of turbulent viscosity uT is a constant, estimate its value, considering
the flow in Volga river (the difference in the heights of the source and mouth equals
300 m, the length equals 3000 km, the average velocity is of order 1 m/s, and the
depth equals 5 m.) See Problem 22.20.
0 22.66 Those who study flows in the ocean often use the model of turbulence in
which, for r? the hypothesis of Boussinesq (see Problem 22.65) with two different,
but constant, values are taken for the coefficients of turbulent viscosity Av and A^.
T? depends linearly, with the constant of proportionality Av, on the derivative of
the average velocity with respect to the vertical coordinate, and with the constant
of proportionality Ah on the derivative with respect to the horizontal coordinate.
(Ah is much greater than Av because of the stabilizing effect of the gravity force
and small vertical extension of the ocean.) Using this model, write the equation of
turbulent motion for the ocean, assuming that the fluid is ideal, incompressible, and
homogeneous.
© 22.67 Consider a turbulent stream of a homogeneous, incompressible fluid over

the infinite unmoving plane y = 0 if there are no body forces (such as wind over the
Earth's surface). All the characteristics of the flow are supposed to depend only upon
y, vy = i)t = 0, vz = ii(y). On the boundary y = 0, the tangential stress r 0 = const
is given. For T ^ , the hypothesis of Prandtl is taken so that T? = pvT{dv/dy) where
23. Waves on a Surface of Heavy Imcompressible Liquid 155
i/T = K2y2(dv/dy), and K = 0.4 is an empirical constant. Find v(y) in the region
y > 6 > 0 where the molecular viscosity can be neglected compared with the turbulent
one.
0 22.68 The logarithmic velocity distribution in turbulent flow v(y) obtained in

the preceding problem cannot satisfy the condition of adhesion on the surface y = 0.
To extend the solution to the boundary y = 0, a thin laminar underlayer 0 < y < S
located between the boundary and the region of turbulent flow is considered. In this
underlayer, the turbulent stresses are less than viscous ones, so only viscous stresses
are taken into account. Find v(y) for all y > 0, joining the solution found in the
preceding problem with the solution for the laminar underlayer at y = 6. Obtain the
quantity 6 from the condition Re|y=< = v(6) S/u = 1.
0 22.69 Find v(y) under the setting of Problem 22.67, using the theory of dimen
sionality (without the hypothesis of Prandtl).
0 22.70 Averaging the equation of internal energy for an incompressible ideal

heat-conductive fluid, derive the equation for the average characteristics. The terms
of this equation qj = pcT'v^ are called the components of a turbulent heat flux.
0 22.71 Find the profile of average temperature T(y) in the flow considered in
Problem 22.67, assuming that the heat flux q0 is given at y = 0, and the vector
of turbulent heat flux is proportional to the gradient of average temperature: qj =
kT(dT/dx'), and the coefficients of turbulent viscosity vT and heat conductivity A;T
are proportional: A:T = avr, a = const.
23 Waves on a Surface of Heavy Imcompressible

Liquid
In the suggested problems, it is implied that the liquid is bounded by a free surface
and an impenetrable immovable surface (bottom and shores). The coordinate system
is chosen so that the axis z is directed along — g, and z = 0 is the equation of
the free surface of the resting fluid. The free surface in wave motion is described
by the equation z = ((x,y,t) where £ is the function subjected to determination.
The impenetrable boundary is given by the equation z = -h(x,y), and the quantity
h is called the depth of the liquid. For most problems, unless otherwise specially
stipulated, an ideal incompressible homogeneous fluid is chosen as the model of the
medium.
156 FLUID MECHANICS
— PROBLEMS —
0 23.1 Regarding the motion as potential, write the equations, boundary and
initial conditions for the problem of gravitational waves in an ideal incompressible
homogeneous liquid caused by an initial disturbance (the problem of Cauchy-Poisson).
Assume the surface tension to be negligible.
0 23.2 Assuming that the displacements £ of the free surface from the undisturbed
surface (horizontal plane x,y) and the inclination of this surface d^/dx, d(,/dy are
small, set up the approximate linearized version of the problem of Cauchy-Poisson
for the velocity potential (p (the problem of waves of small amplitudes).
0 23.3 Show that the plane problem (vv = 0, d/dy = 0) of gravitational waves
of small amplitudes in an unbounded along x basin of a finite constant depth h has
a solution of the form £ = Re [A exp (ikx — iu>t)} where k, u> are real constants,
i = y/—l. This harmonic wave is referred to as progressive. Find this solution,
explain the meaning of its name, obtain the equation relating the wave number k
and the frequency LJ (the dispersion relation). Write the expressions for the length,
period and propagation speed (phase speed) of these waves. Consider the limiting
cases a) kh 3> 1, b) kh <SC 1.
© 23.4 a) Show that the complex potential W = c(Z — iae~lkZ), Z = x + iz at

ak -C 1 (with c, a, and k as real constants, c2 = g/k) describes the velocity fields of a
plane harmonic progressive wave of small amplitude in an unbounded along x basin
of infinite depth, b) What is the equation for the surface of the wave and its speed
relative to the liquid resting at z —» - c o ?
© 23.5 A progressive wave of small amplitude propagates on the surface of a heavy

liquid with infinite depth (see Problem 23.4). For Re = c/kv » 1, determine the
stationary (in the coordinate system connected with the wave) velocity distribution
a) inside of the boundary layer near the free surface of the liquid (calculate the normal
component of the viscous correction to the velocity on the external boundary of the
layer); b) outside of the boundary layer.
© 23.6 A heavy incompressible liquid performs a wave motion in a channel of the

depth h = const, £ = £(x, t) = £(x + A, t)(X is the length of the wave). Show that the
potential energy EfM of fluid in the region between the two vertical planes x = x\,
x = x 2 = a;i+Ais expressed by the formula Epot = — Q2dx + const.

Xl
© 23.7 a) Calculate the kinetic and potential energy of one period of a progressive
wave of small amplitude, a t < l (see Problem 23.4) on the surface of a heavy ideal
incompressible liquid with infinite depth, b) Show that, with viscosity taken into
account, the kinetic energy is determined by the potential flow of the ideal liquid
with an accuracy to terms ~ Re - 3 ' 2 at Re = c/kv 3> 1. c) Write the dependence of
total mechanical energy on time for the case of an ideal and viscous liquids.
© 23.8 Show that the dissipated energy T> per one period of a harmonic progressive
wave of small amplitude on the surface of a viscous fluid with infinite depth at Re =
cjkv » 1 has the form V = 47r/ic2(a/c)2(l + 0(l/\/Re)). The first term is calculated
using the velocity of the potential flow, and the second one is determined by v' —
the viscous correction to the velocity in the boundary layer.
© 23.9 a) Using the energy equation, find the function describing decay of the
amplitude of a progressive harmonic wave of small amplitude in a viscous incompress
ible fluid of infinite depth, b) By how many times does the amplitude of the wave
decrease during the time for travelling the distance equal to the length of the wave?
0 23.10 Find the solution of the problem of plane, harmonic-in-time waves of

small amplitude (Problem 23.3) in a basin of constant depth bounded by plane vertical
walls x = 0, x = L. The solution obtained is the standing wave which in geophysics
is called seiche.
0 23.11 Find the solution of the problem of Cauchy-Poisson for one-dimensional

waves of small amplitude (Problem 23.2) if the initial data Co(z). Co(x) — (§£) can
be written in the form of a Fourier integral.
© 23.12 The superposition of two one-dimensional waves of the form C<* =

aQ cos (kax — uat), a = 1,2 with identical amplitudes aj = a^ and close wave numbers
fci, k2 (^2 —fci= Afc is small) is called beating. Determine the shape of the envelope
of the resulting wave and the speed of this envelope (group speed).
0 23.13 If the phase speeds of waves c = u/k depend on their length A = 2n/k,
the waves are referred to as dispersing. (The notion of dispersion is common for all
wave processes.) a) Show that the group speed of dispersing waves U = dw/dk
differs from the phase speed c. b) Using the dispersion relation (Problem 23.3),
calculate c and U for waves on the surface of a liquid in a channel of finite depth.
Consider the limiting cases l ) / i « A ( a wave on shallow water), 2) h » A (a wave in
a channel of infinite depth). Which of these waves has dispersion?
© 23.14 Let a solution of the linear problem of waves has the form £(x,t) =
oo
Re / f(k) exp [ikx - i w(fc) t] dk where f(k) ^ 0 only on the small segment [ko, ko +
—oo
A/c]. This represents a group of waves with close values of wave number and is called
a modulated harmonic wave or a wave packet. Show that, for t <C 2n/ui"(Ak)2, this
158 FLUID MECHANICS
solution can be written in the form C(x,£) = HeA(x') exp(i6) where x' = x — Ut,
A(x') is the amplitude of the wave packet, i.e., a function whose value is nonzero on
a segment of the axis x, the length of which is inversely proportional to Ak, where
U = (duiI'dk)\k=ko is the group speed, 0 = k<>x - u)0t is the phase of the wave packet,
and kg and USQ = u(ko) are constants.
© 23.15 Let a solution of the linearized problem of waves on the surface of a liquid
moving in a basin of variable depth h(x) has the form £ = R e A(x, t) exp [i d(x, t)}
where A, 66/dx, dO/dt as well as the depth of the basin and the other parameters of
the undisturbed flow are slowly varying functions of x and t (i.e., slightly changing
over distances of the order 2ir/(d9/dx) and over times of the order 2ir/(66/dt); d6/dx
and 66/dt are finite). Such a solution is called a wave packet; A(x, i) is its amplitude,
6(x, t) is the phase, 66/dx = k is the local wave number, and -66/dx = w is the
local frequency. Under the suppositions mentioned above, k and w can be regarded
as satisfying the dispersion relation u = il(k, x, t) with the values of parameters of
the undisturbed flow at the point x, t. a) Show that k and w satisfy the Hamilton
equations dk/dt = —dQ./dx, dx/dt = d£l/dk; Q(k,x,t) serves as the Hamiltonian.
b) What is the physical meaning of the group speed U = dCl/dk?
© 23.16 Under the setting of Problem 23.15 and suppositions that the background
flow is uniform and steady, determine the dependence of the amplitude of the wave
packet upon time, using the energy coservation law.
© 23.17 A wave, whose length A is much greater than a characteristic depth of the
basin h (A » h), is referred to as long. When describing such waves, the acceleration
along a vertical is neglected, and the horizontal velocity is supposed to be constant
along a vertical, a) Show that, in this approximation of shallow water, the function
describing the free surface ^(x 1 ,! 2 , t) as well as the horizontal components of the
velocity ^ ( x ^ x 2 , t), satisfy the equations
!+^+^- (231)
dva dC
where x ^ x 2 are Cartesian coordinates, b) What simplified form do these equations

take, if h = const, the waves are plane (v2 = 0, d/dx2 = 0), and the amplitude of the
waves is small (£ <S h <C A)? What is the speed of propagation of these waves equal
to?
0 23.18 Waves caused only by action of the gravity force are called free gravi
tational waves. If other forces acting, the waves are referred to as forced, a) Solve
the problem of forced long waves of small amplitude (see Problem 23.17) caused by
the force Fx = A sin (fCx — at) in a basin of constant depth h, where A, K., and
a are given constants. (This is the form to which some problems of the theory of
lunar and solar tides on the Earth are reduced.) Find the difference of phases for the
acting force and waves caused by it. b) Determine the depth of the basin in which the
amplitude becomes infinite (resonance phenomenon), and, consequently, the validity
of the linearized setting is violated.
0 23.19 Show that there is an analogy between the motion of an ideal incom
pressible liquid in a shallow basin of constant depth and the adiabatic motion of an
ideal perfect gas with 7 = cP/cv = 2 (see Section 24).
0 23.20 Show that, if h = const, the system of one-dimensional (vy = 0, d/dy =

0) equations of shallow water theory (Problem 23.17) is equivalent to the system
dJt d,l+ dj- 3J_
0,
dt + c+ dx = 0 , dt + C- dx
where J± = vx ± 2Jg(h + C) are the Riemann invariants and c± = vx ± Jg(h + £)

are the speeds of characteristics (see Problems 24.26 and 24.27).
0 23.21 a) Consider the particular solutions of the system of one-dimensional
equations for shallow water at h = const in which one of the invariants of Riemann
is constant all over the stream. These solutions are called Riemann waves (see Prob
lem 24.29). b) Determine the variation in the shape of the free surface with time
and the condition of wave breaking (i.e. appearance of points where dC,/dx = 00).
c) Write the approximate equation for £(x, t) for a Riemann wave of small amplitude,
considering only the main non-linear term.
0 23.22 The system of equations of shallow water (Problem 23.17) is hyperbolic

and, consequently, allows solutions with discontinuities, a) Write the conditions on
a velocity jump and on that of stream depth in the suppositions of shallow water
theory. Such a jump is called a hydraulic jump, b) Find the propagation speed D
of a normal jump (where the velocity v is normal to the discontinuity surface) if the
depths hi and ft2 on its sides and the velocity «i are known, c) Show that a jump
moves always to the region of less depth.
© 23.23 Under the suppositions of the shallow water theory (Problem 23.17),
consider the motion of a liquid in a long channel of an arbitrary cross section, a) Derive
the equations for the velocity of the stream along the channel u and the cross-sectional
area S. b) Show that there is an analogy between the equations obtained and the
equations describing a motion of a gas (see Section 24). What dependence p(p) should
be valid in the gas for this analogy if the cross section of the channel is 1) a rectangle,
2) a triangle? c) What is the propagation speed c for waves with small amplitude?
160 FLUID MECHANICS
d) Considering Riemann waves, derive the conditions of their breaking as determined

by the shape of the cross section of the channel.
0 23.24 A solitary wave on the surface of a heavy liquid has the profile z = Q(x, t)
with horizontal asymptote {(, —► 0 as x —» ±co) and propagates in a layer of depth
h with speed c relative to the liquid resting at infinity, a) Find the momentum of
the liquid Q and its potential energy £,*,,. b) At £//i < 1, find c and the kinetic
energy of the liquid £kin neglecting az = O.and dvx/dz = 0 in the equations of motion.
Compare E^, and Eu,.
0 23.25 From the dispersion relation for gravitational waves of small amplitude
on the surface of a liquid with constant finite depth h (Problem 23.3), in the first
approximation with respect to kh <K 1, derive the dispersion relation of the linearized
shallow water theory for waves a) running in two opposite directions (expanding w2(Jk)
over powers of fc), b) running in one direction (expanding u>(&)).
Write out the differential equations corresponding to these expansions and compare
them with those of Problem 23.17.
© 23.26 Determine the next approximation with respect to kh -C 1 in the ex
pansion of Problem 23.25 and verify that the following differential equation Carre
s' d2C gh3 d*{
sponds to them: a) -^ = gh —— + —- ^-^ — the linearized Boussinesq equation,
b) — + yjgh I — + — ^ - j I = 0 — the linearized Korteweg-de Vries equation.

These equations take into account the dispersion of long waves.
0 23.27 Assuming that dispersion and non-linearity independently influence the
behavior of waves running in the same direction, and using the solutions of Prob
lems 23.21 and 23.26, obtain the Korteweg-de Vries equation
taking into account the dispersion and non-linearity of long waves (with accuracy up
to terms of second order with respect to C/h).
0 23.28 The Korteweg-de Vries equation has particular solutions in the form of
solitary waves (solitons) which propagate in a channel of constant depth h without
variation of their shape. Find this solution, assuming that the conditions vx = UQ = 0
and Q = §£ = gj£ = 0 are satisfied at infinity. Show that the speed of this wave
depends on its amplitude and is greater than Co = \/gR-
© 23.29 In liquids stratified in density, waves can propagate on the interfaces of

the layers (internal waves). Let there be two horizontal layers of liquids with different
23. Waves on a Surface of Heavy Iincompressible Liquid 161
densities p\ and p2, one lying on top of another. The thicknesses of the layers are,
respectively, h\ and h2 and are so small that the shallow water approximation can
be used (Problem 23.17). In the framework of the theory of small amplitude waves,
find the relationship between the velocities and amplitudes of the surface and internal
waves. Obtain the simplified formulae for the case of nearly equal values of p\ and
Q 23.30 In studying stability of homogeneous steady flows using a linear approxi

mation, evolution of small disturbances of these flows is considered. A linear problem
for one-dimensional disturbances has solutions of the form exp(ikx - iut) where k and
u> are bound by a dispersion relation D(w,k) = 0. Its roots u(k), if the values of k
are real, determine the evolution of the disturbances. If there exists at least one root
among all ui(k) for which Imu > 0 at k such that Imfc = 0, then the flow is unstable,
since the disturbances corresponding to these k grow as time proceeds. Consider a
flow with a tangential discontinuity separating two plane-parallel uniform streams of
an ideal incompressible fluid with densities pa and velocities va, (a = 1 for z > 0,
a = 2 for z < 0) (see Figure 23.1). Show that this flow is unstable (the instability of
Kelvin-Helmholtz).
P, "..
Figure 23.1
0 23.31 In studying waves in basins on the surface of the Earth, it is necessary

to take account of rotation of the Earth for many cases, a) Write the equations of
motion for an ideal liquid in a Cartesian coordinate system connected with the point
of the surface of the sphere rotating with angular speed fl (the axis z is directed
vertically upward, and z = 0 is the horizontal plane at the considered point). Assume
that g — g(tp) where ip is the geographical latitude of the place, so that the centrifugal
force is included in the gravity force, b) Ordinary for flows in the ocean, vz <C vz,vy.
Estimate the terms in the equations and write the approximate system of equations
for these motions.
© 23.32 Consider a plane vertical wall y = 0 in a layer of water with depth h on the
plane tangent to Northern hemisphere of the terrestrial globe at a latitude ip. In the
162 FLUID MECHANICS
approximation of long waves with small amplitudes (Problem 23.17), determine the
surface wave whose amplitude decays as the distance from the wall increases. Show
that this wave propagates along the right side of the wall relative to the direction
of propagation, and only the velocity component vz parallel to the wall does not
vanish (a wave of Kelvin). Find its speed and coefficient of amplitude attenuation
with increase of the distance from the wall. Neglect variation of the latitude. Hint:
a) consider the particular solution vy s 0; b) show that vy = 0 is the general solution.
© 23.33 Coriolis force can cause wave motion in a sufficiently extended basin
on the surface of the rotating Earth, the particle paths lying on horizontal planes
(v, = ()). Such waves describing oscillations of velocity over the whole thickness of
the fluid are observed in the atmosphere and in the ocean and called planetary waves
or waves of Rossby. Determine the shape of periodic waves of this type propagating
along the direction x of the parallel of the terrestrial globe in a basin so extended
that variation of geographical latitude if = tp(y) along the direction y of the meridian
must be taken into account (in the first approximation). The special choice of the
axes i , y (along the parallel (to the East) and meridian (to the North)) allows one
to assume that the Coriolis parameter / = 2f2 sin <p depends only on y. Taking into
account only the first approximation beyond the constant, one obtains f - fo + 0y,
where 0 = df/dy = const. Determine the magnitude and direction of the velocity of
wave propagation, show that the waves have dispersion.
24 Mechanics of a Compressible Fluid

This section concerns effects connected with compressibility of fluids, i.e., with vari
ation of their density due to variation of the pressure. For many cases even for easily
compressible media, e.g., gases, the change in density during motion is small and can
be neglected. The conditions under which fluids can be regarded as incompressible
are presented in Section 21. Here, the cases studied are those for which the change in
density is either not small or else it is small, but produces a phenomenon of interest
(e.g., propagation of sound). The mechanical equations (of continuity and momen
tum) are not sufficient to describe a motion of a compressible fluid Even if the
viscosity coefficients do not depend upon the temperature, these equations include
five unknown functions of the coordinates and time: the density p, pressure p and
three velocity components t;,. Therefore, the thermodynamic equations expressing
the first and second laws of thermodynamics, as well as equations of state for the
medium (see Chapter 3), are necessary for determining p, p and v. Usually, fluids
can be regarded as two-parameter media for which a thermodynamic state of small
particles is determined by two parameters, e.g., by the density p and temperature
T. In particular, the pressure p and density of internal energy u for each particle are
24. Mechanics of a Compressible Fluid 163
functions of p, T:
P = p(P, T) , u = u(p, T) .
These relationships are called the equations of state. For many cases, it is more
convenient to use the entropy s instead of the temperature. Under certain conditions,
the thermodynamic equations allow one to express the pressure as a function only of
the density. If this function is the same for all particles of the medium, the motion is
referred to as barotropic, and the dependence p = p(p) is used to close the system of
mechanical equations. Compressibility of the particles of a medium can be described
by the compressibility coefficient k
dp I (1 dp
k=—
dtl \p dt
The value of k is determined not only by properties of the medium, but also by the
process taking place in it. If the process is reversible and adiabatic, then the entropy
of each particle is conserved, and, consequently, the equation of state p = p(p, s) and
the condition ds/dt = 0 lead to the formula
dp dp i dp ds\ j (\ dp\ fdp\ _2

k
~ > ~ [dp dt + dsdt)/{-pdt)-p
k
UJ,=const " "°'
where
a=
(dp\ = [£
The quantity a is called the speed of sound. This is the speed with which sound,
i.e., small variations of pressure, propagates over the particles of the medium. For
barotropic processes,
. MP)
k = p-
dp
The quantity Jdp(p)/dp is also often denoted by a and is also, by convention, called
the speed of sound. For an incompressible medium, a = oo. The quantity a as well
as the Mach number M
a
are of great importance in the mechanics of compressible fluid. In particular, for
steady motions, subsonic streams (v < a, M < 1) differ qualitatively from supersonic
ones (v > a, M > 1).
164 FLUID MECHANICS
— PROBLEMS —
Equations describing the motion and state of a compressible

fluid
© 24.1 a) Write the closed system of equations describing a motion of a com
pressible viscous fluid, b) Write the system of item a) for a linearly viscous, heat-
conducting, perfect gas in a gravitational field. Assume that heat conduction obeys
Fourier's law, the coefficients of viscosity and heat conductivity are constant, and the
coefficient of bulk viscosity is zero.
0 24.2 a) Comparing the orders of magnitude of terms in the heat-flow equation

for a viscous, heat-conducting, perfect gas (see Problem 24.1), obtain the conditions
under which the terms describing the work of viscous stresses and heat influx due to
heat conduction in this equation can be neglected (and, consequently, the entropy of
every particle can be regarded as constant), b) Carry out an estimation specifically
for the steady air flow around a body with the velocity 100 m/s and with normal
temperature and pressure far from the body. The linear dimension of the body is 10 m.
Assume that, for air, p = 0.125 kg/m 3 , v = 1.45 • 10"5 m 2 /s, k = 0.025 J/(m • s • K),
3
Cv = 1Cv = io J/(kg • K), 7 = 1.403, a = 340 m/s.
0 24.3 Obtain the condition for the parameters of a stream under which the grav
ity force in the equations of motion can be neglected. Assume the gas to be ideal.
© 24.4 Write the closed system of equations for determination of the velocity,
density, and pressure in adiabatic motion of an ideal compressible fluid if there are
no body forces (this system is called the system of equations for gas dynamics).
0 24.5 a) For a perfect gas, obtain the expressions for the density of internal
energy u, entropy s, enthalpy i and speed of sound a = \J{dp/dp)a as functions of
the pressure and density, b) Write the closed system of equations for determination
of the velocity, density and pressure in an adiabatic motion of an ideal perfect gas.
0 24.6 Under what conditions is adiabatic motion of an ideal compressible fluid

barotropic?
0 24.7 Prove that, for continuous barotropic motions of an ideal compressible

fluid in a field of body forces having a potential, the circulation of the velocity over
a closed contour passing through the same particles of the fluid does not vary with
time (Kelvin's circulation theorem).
0 24.8 Prove that, 1) if the velocity field of an ideal fluid has a potential in the
whole space at an instant, 2) the motion is continuous and barotropic, 3) the body
forces have a potential, then the flow remains potential all the time (Lagrange's
theorem).
© 24.9 Write the system of equations for determination of the density and velocity
potential for a barotropic motion in a body force field having a potential.
0 24.10 For an isentropic (s = const) potential flow of a perfect gas in the field of
gravity force, a) write the integral of the equations of motion (the Cauchy-Lagrange
integral); b) obtain the equation which includes only the velocity potential.
Motion with small disturbances. Propagation of small pres

sure disturbances. Speed of sound
0 24.11 a) An ideal compressible fluid is in mechanical equilibrium (v = 0) in
the absence of body forces. The pressure and density are given by the values p —
Po = const, p = po = const, respectively. As a result of a small disturbance, a motion
arises, so that p = p0 + p'(x{, t), p = p0 + p'fo, t), v = v'(xi, t), where p', p', v' as
well as their derivatives are small. The motion is barotropic. Write the linearized
system of equations for the functions p', p', v'. Show that the functions p', p' satisfy
the wave equation. The velocity vector v of any motion can be expressed as the sum
v = Vi + w2 of a potential (vi = grad<p) and solenoidal ( divu2 = 0) vectors. Write
the equations for v\ (and, as a consequence, for (f) and for v'2. b) What equation
governs the small pressure disturbance p' in a homogeneous incompressible fluid?
© 24.12 Find the relation between fc and w for which the function <p = tp0el(- fcT
*-u")
is a solution of the wave equation
Here, ip0 = const, w = const, r is radius-vector of the observation point, fc is a

constant vector, i = y/^1. Such solution is called a monochromatic wave. If w is
real, (fo is the amplitude of the wave, Jw| is the angular frequency, fc is the wave
vector, A = 2?r/|fc| is the wavelength, / = fc • r - ujt is the phase, and / = const
is a surface of constant phase. The speed of motion of this surface in the direction
of its normal is called the phase speed, a) Find the phase speed, and determine the
relation between the direction of the wave vector and the displacement of the surface
of constant phase, b) Show that sound waves are longitudinal, i.e., the displacement
of particles of the fluid is parallel to the wave vector.
166 FLUID MECHANICS
© 24.13 For one-dimensional motions with plane waves (u„ = vz = 0, and vx =

v{x, t), where x is the Cartesian coordinate), a) find the general solution of the system
obtained in Problem 24.11 for the functions p'(x,t), p'(x,t), v'(x,t), ip(x,t). b) At
an initial instant, the disturbance of velocity v' equals zero everywhere, and the
disturbance of pressure is nonzero only in the layer |x| < I where it equals p'0 =
Pi(l - x2/l2). Find p'(x,t), p'{xyt), v'{x,t), at t = 21 /a and at t = 3//a. c) Particles
in a monochromatic plane sound wave oscillate with the amplitude 0.25 mm. The
frequency of the oscillations equals 500 Hz. Find the wavelength, maximal velocity
of particles and maximal relative change in density, if the wave propagates 1) in air
(a = 340 m/s); 2) in water (a = 1400 m/s).
0 24.14 The speed of propagation of small pressure disturbances over a gas at

rest is calculated by the formula a = ^J(dp/dp)o (see Problems 24.12, 24.13) and
determined by the function p = f{p). Find the expression for the speed of prop
agation of small disturbances in a perfect gas with given heat capacities Cp and cv
for the cases of isothermal disturbances (Newton's formula for aT) and of adiabatic
disturbances (Laplace's formula for a,). Calculate aT and at for air {cj,/cv = 1.4
and R = 287.14 m 2 /s 2 K). It is known from experiments that the speed of sound in
air at the temperature 15°C equals approximately 340 m/s. Which quantity corre
sponds to the speed of sound — aT or a,? Explain the answer using the results of
the Problem 24.2.
0 24.15 For spherically symmetric one-dimensional motions (motions with spher

ically symmetric waves), find the general solution of the linearized system obtained
in Problem 24.11 for the functions p'(r, t), p'(r, <), ip(r, t) where r is the distance from
the center of symmetry.
© 24.16 Obtain the linearized system of equations (see Problem 24.11) for one-
dimensional small disturbances depending upon x and t in a uniform stream moving
with constant finite velocity along the x-axis. Show that the frequency of oscillations
u in a monochromatic wave e'**1-'""' of a given wavelength, i.e. of a given value of k,
depends upon the velocity U of the stream.
0 24.17 A source of small monochromatic disturbances moves with constant ve

locity U < a along the x-axis in an unmoving compressible fluid. The disturbances
produced by the source have the form of plane waves with constant phase surface per
pendicular to x-axis. Show that the frequency \ui\ of oscillations in a monochromatic
wave (see Problem 24.12) in the reference frame of an unmoving observer differs from
the frequency u). in the reference frame moving with the source (the Doppler effect).
0 24.18 A uniform stream of gas moves with constant velocity U in the direction
of x-axis. At the instant t = 0 at a point O of this stream a source of small pressure
disturbances begins to act. Determine the region where the stream is perturbed at
the instants t = ti and t = t2 > h if a) U < a, b) U = a, c) U > a (a is the speed of
propagation of small disturbances over particles of the gas).
0 24.19 Small disturbances tend to damp due to viscosity. Demonstrate this

fact, considering small one-dimensional monochromatic oscillations of a viscous gas
for which p = p(p).
© 24.20 A plane sound wave propagating along x-axis is incident on an interface

between two media — the plane i = 0 parallel to the constant phase surface (normal
incidence, see Figure 24.1). The pressure disturbance in the incident wave is known:
p,° ■'
" . '
- - ^ p2° -==-t _ _
—
)X
Figure 24.1
Pio = fo(t-x/ai). Find the pressure disturbance in the reflected (p' n ) and transmitted
(p'2) waves if the values of densities pi, /^ and speeds of sound ai, a2 in both media
at the equilibrium state are given (the values of p'n/p'10 and Pj/Pio at the interface
are called respectively the reflection or refraction coefficients for the given pair of
media). Estimate the values of amplitudes of the reflected and transmitted waves at
the interface if the wave transmits a) from air to water, b) from water to air (the
speed of sound in water equals 1400 m/s).
0 24.21 A plane monochromatic sound wave propagating along the f-axis for
which V?o = Ae{^-^ is incident on an interface AB between two media (see Fig
ure 24.2) so that the direction of its propagation (the f-axis) makes an angle 6 with
the normal to the plane of the interface (oblique incidence). The densities pa and
speeds of sound aa (a = 1,2) in both media at equilibrium are known. Determine
the direction of propagation of the reflected and refracted (transmitted) waves (the
angles 6\ and 92). Show that for a2 > ai the transmitted wave does not exist for all
angles of incidence 0.
© 24.22 There is a fine mixture of two media, one of which is an incompressible

liquid, and the other is a perfect gas. There is neither heat exchange nor phase
transition between the phases. The pressures and velocities in both phases can be
168 FLUID MECHANICS
Figure 24.2
assumed to be identical. Find the speed of sound in the mixture as a function of

mass concentration of the compressible phase. Show that the speed of sound in the
mixture can be less than the speed of sound in each of the phases. Calculate the
speed of sound in the mixture consisting of
a) small water droplets in air (mist) if the mass concentration of air is a = 0.001,
pAa = 0.00125 g/cm 3 , pWAttt = 1 g/cm3;
b) small ice particles in air at a = 0.01, p^ = 0.8 g/cm3.
Motion with small disturbances. Steady flow around a thin

body
© 24.23 An airfoil is placed in a translatory stream of a gas under small angle of
attack. The airfoil is thin, so that it perturbs the stream weakly. The viscosity of the
gas and body forces can be neglected, the motion of the gas is steady, barotropic and
has a potential, a) Obtain the linearized equation for the velocity potential and the
Cauchy-Lagrange integral in linearized form, b) What boundary condition should be
satisfied on the airfoil (the flow is assumed to be non-separated)? c) Linearize the
boundary condition on the airfoil, d) Derive the equation for the potential tp, the
formula relating p" and <p, and the boundary conditions on the airfoil for a stream of
incompressible fluid. Which of these relationships differ from the corresponding ones
for a stream of compressible fluid?
© 24.24 A steady stream of gas flows about a thin cylindrical wing of infinite
span (normal to the flow) under a small angle of attack. The motion is barotropic
and potential, viscosity and body forces are neglected, and the flow is not separated.
a) Write the equations and boundary conditions for the velocity potential and pres
sure. b) Let the stream be subsonic (v0 < ao, M0 < 1). Reduce the problem of
determining the velocity potential to the problem of determining the potential of
the stream of incompressible fluid flowing around the same wing. Obtain the rela
tion between the fluid forces acting on the wing for the cases of compressible and
Figure 24.3
incompressible streams. Find the drag force of the wing.

0 24.25 A supersonic stream (M0 > 1) flows about a thin wing under the condi
tions presented in Problem 24.24. Find the velocity potential and the drag force of
the wing if it has the shape of a flat plate.
Propagation of finite disturbances in an ideal compressible

fluid
Problems 24.26-24.31 demonstrate the following effects. The front of a disturbance
of pressure and velocity moves over particles of the medium with the local speed of
sound while the motion remains continuous. The shape of the wave of a disturbance
varies during propagation. For many cases, the front of a disturbance instantly or for
finite time turns into a shock wave, i.e., a surface of strong discontinuity, propagating
over the fluid. Problems 24.26-24.31 deal with one-dimensional unsteady motions
with plane waves. That is, it is supposed that the parameters of the fluid depend
only upon the time t and one Cartesian coordinate x, and also that vy = v2 = 0.
0 24.26 a) Write the continuity equation, the equation of motion and the equation
of heat flow for an adiabatic motion with plane waves in an ideal compressible fluid.
b) A quasi-linear partial differential equation of the first order is said to have the
characteristic form if all the derivatives of the function sought for with respect to
the independent variables t and i appear in the equation only as the combination
d/dt + cd/dx (c is a function of t, x and the functions sought for). The lines on
the x-t-plane, the equation of which is dx/dt = c, are called the characteristics; c is
the velocity of a characteristic, d/dt + cd/dx is the derivative along a characteristic.
Do the equations of the item a) have the characteristic form? c) Obtain the system
of equations in the characteristic form equivalent to a system of item a), composing
linear combinations of the original equations, d) Let a motion be barotropic with
body forces neglected. Determine the quantities whose values are constant along the
characteristics (Riemann invariants).
170 FLUID MECHANICS
© 24.27 What are the velocities of the characteristics and the Rjemann invariants
(see Problem 24.26) equal to for an adiabatic barotropk: motion of a perfect gas?
0 24.28 Consider a one-dimensional motion of an ideal gas along the i-axis with
known v = v{x.t), p = p{xj), p = p(x,t), and s = s{x,t). At the instant t = t0,
as a result of some external perturbations, a disturbance of the stream arises in the
region xA < x < i B , so that u, p, p, s remain continuous (i = xA, i = r f l are
weak discontinuity surfaces at t = tD. see Chapter 3). Supposing that the motion is
continuous at t > t0, show the regions on the r,(-plane where the gas "does not fed"
the perturbations, i.e., where v = v(x, t), p = p(x, t), p = p(x, t), and s = ,s(r, t) as if
there were no perturbations.
© 24.29 a) Consider a barotropic motion of a compressible fluid with plane waves.

Find the particular solutions of the system of equations for which v, p, p depend only
upon one combination 6(x,t) of the independent variables i, t (such motions are
called Riemann waves). Neglect viscosity, body forces and heat conduction
© 24.30 Consider an ideal perfect gas with p = Po, p = Po and w = 0 in a

cylindrical tube unbounded from one side and closed by a piston from the other 6ide
(see Figure 24.4). At the instant t = 0 the piston starts to move out of the tube
with a velocity u(t) such that | u(t) \- Xt (X = const > 0) if 0 < t < tx and
| u(t) | - Xti =| Ui |= const if t > ti. There are no body forces and the motion
is adiabatic (see Figure 24.4). a) Write the equations and the boundary conditions
WWWWWWV \\\\\\\\\\\\\s\\\\\\s\\\\\\
ft
u(t) *- _\ V=0
/V/V/V/V/V^/V/vy;>J/;JSSSS//??????/;S//SS;
Figure 24.4
for determination of velocity and pressure in the gas. b) Find the velocity of the
interface T between the region where the gas is in motion and the region where it is
still at rest. Show that the motion in the region adjacent to V is a Riemann wave
(see Problem 24.29). c) Determine the distribution of velocity and pressure in the
tube, d) Find the maximum velocity of the piston at which it still does not break
away from the gas. e) Let the piston start to move immediately with the velocity
Hi = const (\u(t)\ = uj at t > 0). Find v(x, t) and p(x, t).
© 24.31 Consider Problem 24.30 in the case for which the piston moves to the
side filled with the gas. Show that, in this case, a continuous solution exists only at
t < t0 where t0 depends upon the acceleration of the piston.
© 24.32 A piston of mass m in a cylindrical tube with cross-sectional area S, able

to move without friction, separates a gas that is at rest with a constant density po
and pressure p 0 a t 'he initial instant (t = 0) from a vacuum. The gas is supposed to
be ideal and perfect, the motion of the gas is adiabatic. The tube is infinitely long.
There are no body forces, a) Determine the velocity of the piston vp(t), assuming
that vp(0) = 0. Compare the result with the velocity with which the gas flows out
into vacuum (m = 0). b) Let 7 be the adiabatic exponent. Analyse the limit as
7 - t o o corresponding to the case of incompressible fluid.
© 24.33 A piston of mass m moving without friction in a semi-infinite cylindrical

tube with cross-sectional area S separates vacuum from a gas of mass mg contained
between a fixed wall (x = 0) and the piston (xp(t) > 0) The gas is supposed to
be ideal and perfect, the motion is adiabatic. There are no body forces, a) Let
////////////////////////////(^///////////
', m, m ^S vacuum
V7777777777777777T?777777777fT77777777777
Figure 24.5
the density po of the gas at the initial instant be uniform, and the pressure po be
distributed so that v = A{t)x at t > 0, .4(0) = 0 (v is the velocity of the gas).
Find the function Po{x) and the efficiency of the device 77 = mVp(oo)/2E0 where E0
is the initial energy of the gas, and vp is the velocity of the piston, b) Analyze the
family of solutions of the similar problem with initial density distribution of the form
p0(x) = (mg/lS)(l/x)a(\ - a) (a < 1) and with the same v(x,t) as in the item a),
/ = i p (0). Show that 7] -» 1 as a -» 1.
Motion with shock waves

Solutions of many problems in the theory neglecting viscosity and heat conduction
are discontinuous: surfaces arise on which the velocity, pressure and density or their
derivatives are discontinuous. The jumps of these quantities are bound by relation
ships following from conservation laws and other conditions (see Chapter 4). Discon
tinuity surfaces propagating over particles of a medium are called shock waves. If
viscosity, heat conduction and other phenomena causing dissipation are taken into
172 FLUID MECHANICS
account, shock waves are replaced by narrow zones of continuous variation of the
parameters of the medium. The distribution of the parameters in such zones is called
the structure of the corresponding shock wave. Due to dissipation the entropy of
particles passing through a shock wave increases. In construction of solutions with
discontinuities, certain conditions should be taken into account, namely the relation
ships following from a) the mass, momentum and energy conservation laws, b) the
second law of thermodynamics, c) the conditions for discontinuities to be evolutionary,
d) the conditions for the existence of a structure of the shock wave.
The evolutionarity conditions. A discontinuity surface E d is referred to as
evolutionary if a linearized problem governing interaction between it and small am
plitude planar waves parallel to it have a unique solution. The amplitudes of waves
coming to E^ are assumed to be known in this problem, and those of waves going
away from E<< are to be calculated as well as the change in the speed of E<j itself. It
is necessary for solvability of this problem that
N, = N - 1
where Nf is the number of types of waves going away from the discontinuity surface
and N is the total number of relations between gas parameters across it (i.e., jump
conditions).
In the theory considered in this section the velocities of small disturbance waves
equal velocities of characteristics ck. A wave travelling in front of E<* (marked by ~)
goes away from E^ if its velocity c" is greater then the velocity D of E^; a wave behind
Ed (marked by + ) goes away from it if its velocity c + is less then D. Let n~, n+ be the
number of types of the characteristics in front and back regions, respectively. Then
the evolutionarity conditions for E d can be written in the form
c~_ < ... < ci < D < cl_l < ... < cj"
< < • • • <<+_„+fc+1< D < c „ \ _ „ + t < ... < c+ (24.1)
for various types of discontinuities corresponding to different values of k, i.e. k - 1
and N - k waves go away from E d in front of it and behind it respectively, so that
Nj = (k-l) + (N-k) = N-l.
© 24.34 a) Find the solution of Problem 24.31 under the condition that the piston
starts to move into the gas immediately with the velocity u = const, supposing that a
shock wave propagates over the gas with a speed D, ahead of the shock wave the gas
is at rest, and behind the shock wave there is a translatory stream with the velocity
v = u and pressure p\. Use the conditions across a shock wave (Chapter 4). Find D
and pi. b) For the solution, given in item (a), verify that 1) the entropy increases
across the shock wave and 2) the wave is evolutionary, c) Prove that a solution similar
to that of item (a) does not satisfy the conditions presented in item (b) if the piston
moves out of the gas.
© 24.35 a) Write, in Eulerian form, the equations of a spherically symmetrical

adiabatic motion of an ideal perfect gas in the spherical coordinates r, 9, tp if there
r
are no body forces, b) Assume the mass m{r, t) = / 47rp(r|, t)r\dr\ of the fluid to be
o
contained in the sphere with radius r at the time t as the Lagrangian coordinate where
p is the density of the gas; verify that dm/dt = 0. Derive the equation of motion
in Lagrangian form, c) Write the energy equation in Lagrangian form, d) Write, in
Lagrangian form, the conditions that must hold for a spherical shock wave. Show
that the condition for the mass flux to be continuous is a consequence of the fact that
the function r(m,t) is continuous.
© 24.36 As a result of an explosion at a point, some quantity of energy EQ is
transmitted to a gas at rest filling the whole space with a constant initial density
Po A spherical shock wave propagates from the center of the explosion. Propagation
of the wave is described by the function r = R{t), R{0) = 0, where r is a distance
from the point of explosion, R(t) is the radius of the shock wave. Treat the gas as
ideal and perfect and the motion, caused by the explosion, as adiabatic. There are no
body forces. The initial pressure is negligible comparing to the pressure behind the
shock wave ("strong explosion"), a) Denote by rn the mass of gas inside the sphere
with radius r. The value of m can be taken as a Lagrangian coordinate of the gas
particles which are situated on the sphere. Then the motion law of gas particles can
be written in the form r = r(m, t). The location of the shock wave relative to gas
particles is given by the function m = M(t) and the distribution of the gas entropy
is determined by the function p/p-1 = f{m). Write the general forms of the functions
r(m, t), M(t), f(m) which follow from dimensional analysis (see Sections 36, 37). Take
into account that all functions in this Problem depend on EQ, p0, 7. b) Introducing
the energy of the gas inside the sphere with radius r
*"■■'>-/(? + F?is)*"-
express the derivatives dE/dt and dE/dm 111 terms of the parameters of gas motion
and state, using the energy equation (Problem 24.35c). Show that E(M,t) = E 0 on
the shock wave and the relationship
is valid at each point ("the energy integral"), c) Using the energy integral (see the
item (b)), write the equation for the function r(m, t) in diraensionless form. Integrate
this equation in parametric form, taking the quantity u = vt/r as a parameter.
Determine the unknown constants, using the jump conditions across the shock wave
174 FLUID MECHANICS
and the relationship E(M, t) = E0. Show that, if 7 > 7, an expanding cavity is
formed in the vicinity of the center as a result of the explosion, d) Verify that, if
7 = 7, the solution of the problem of a strong point explosion is of exponential form,
and write it out explicitly.
© 24.37 A flat piston abruptly starts to move along the x-axis (perpendicular to
it) with a constant velocity vp, forming a shock wave in the gas with a variable initial
density po(x), with the total mass of gas per unit area of the piston being finite and
equal to M. There are no body forces. Assume the gas to be ideal and perfect and
the motion to be adiabatic. The initial pressure and velocity of the gas are zero. Let
the motion law for gas particles behind the shock wave have the form
X
x = v(m)(t + t() — I , v(0) = vp, where m = / pdx, ti = l/vp (24.2)
In the equation (24.2) / > 0 is a constant; v(m) is the velocity of gas particles, m is the
mass of the layer of the gas between the plane x = const and the piston at the moment
t (m plays the role of a Lagrangian coordinate; verify that dm/dt = 0). a) Determine
the general form of the distributions of pressure p and density p. b) Using the jump
conditions across the shock wave find the functions v(m), Po(x) consistent with the
equation (24.2). Find the function t = t,(m) that describes the motion of the shock
wave, c) Determine the distribution of energy density in the region behind the shock
wave. Analyze the limit as t —» 00, paying attention to the fact that the speed of the
shock wave and the magnitude of the jump in temperature, due to decrease of the
function po(x), grow to infinity. Calculate the total energy transmitted to the gas.
© 24.38 a) A plane uniform layer of a gas moving bodily in vacuum with a con
stant velocity Vo perpendicular to its boundaries bumps against an absolutely rigid
fixed wall. There are no body forces. The gas is ideal and perfect, the motion is adia
batic. Find the velocity of the free surface of the layer after the impact, b) Solve the
similar problem of the impact of an elastic solid layer against a rigid wall within the
framework of linear elasticity (see Chapter 6). The material of the layer is isotropic.
The process is isothermal. Compare the results with those obtained in item a).
© 24.39 A steady supersonic stream of an ideal perfect gas flows about an infinite
wedge with vertex angle 29 (see Figure 24.6). The velocity, pressure and density of
the gas far upstream are, respectively, v0, p0 and po! M0 = v0/ao > 1, ajj = 7Po//>o-
Show that, if 6 < #„««, the solution is of the following form: there is a shock wave
AOB which passes through the vertex O of the wedge (see Figure 24.6) ahead of the
wedge, the stream is not perturbed ahead of the shock wave, the velocity changes in
the shock wave so that it becomes equal to vector «i above the wedge and V2 below
it parallel to the surface of the wedge, and there are translatory streams with the
/ A
v
° ^*-
^ ^ r ,
\ B
Figure 24.6
velocities v\, v2 in the regions between the wedge and the shock wave. Determine
the relation between the angle of inclination <p of the shock wave to the velocity VQ
and the vertex angle of the wedge 26, the magnitude of vy and the limit angle 9max.
Explain why this solution is not valid if M0 < 1.
Detonation and slow combustion

The following problems deal with discontinuity surfaces where chemical energy pro
duction occurs; e.g., heat generates due to combustion. If ignition is caused by tem
perature increase in a shock wave, then the process is referred to as detonation. If it
is initiated by the gas heating due to heat conduction, and a flame travels with sub
sonic speed over the gas, then the process is called slow combustion. In the following
problems V = I/p.
Q 24.40 Consider a discontinuity surface moving in an ideal gas. a) By the use of
mass and momentum conservation laws alone (see equations (6.11)—(6.12) of Section 6
with m = 0, R = 0; see also Problems 17.10, 17.11) express the density of mass flux
through the discontinuity in terms of values pi, V\ in front of the discontinuity and
p 2 , V2 behind it. Deduce the relation between p2 and V2 with pi and V\ regarded as
fixed, which arises if mass flow density is set equal to a certain constant. Depict this
relation on p, V plane. Let the values of p2, V2, pi and V\ be known. What are the
values of the velocities of the shock wave relative to gas in front of it and behind it?
b) Taking the expression for internal energy in terms of p and V to be given, deduce
the relation between p 2 and V2 at pu Vi being fixed, which follows from the mass,
momentum and energy conservation laws (see equations (6.11)-(6.12) and (6.14) of
Section 6 with m = 0, R = 0, g*, = 0, <?*2 = 0, W = 0 ) by eliminating gas velocities.
The graph of p2 versus V2 at given pi, V\ is called a shock adiabat (Hugoniot's adiabat)
if the discontinuity is a shock wave, c) Write the equation of the shock adiabat for
176 FLUID MECHANICS
a perfect gas (u = pV/(7 — 1) + const). Prove that the derivative dp/dV along the
shock adiabat at the point p l t Vi is equal to -a\/V? (this is true not only for a perfect
gas).
© 24.41 a) Consider the discontinuity with a chemical energy discharge (burning)
at it. Assume that internal energy of the gas has the form u = pV/(-y - 1) + Uchemi
with uchem = C\ in front of discontinuity and uêm = C2 behind it, C\ and C2 being
known constants. The quantity Q = d - C 2 is the discharged energy. Let pu Vu p2,
V2 be the values of p, V in front of the discontinuity and behind it respectively. Depict
on the p, V plane the curve displaying relation between p 2 and V2 at fixed pu Vi and
Q. This curve is referred to as o detonation adiabat. b) Let the complete system
of jump conditions be formed only by mass, momentum and energy conservation
laws. Determine the set of points on the detonation adiabat corresponding to the
evolutionary jumps.
© 24.42 Investigate qualitatively variation of the parameters in a running wave,

which is a structure of a detonation wave; i.e., in a gas flow satisfying the following
conditions: 1) it is one-dimensional and steady in the frame of reference connected
with the wave and 2) chemical energy of the gas tiChem varies continuously in the
flow from C\ to C 2 . Neglect viscosity and thermal conductivity and assume that
the front boundary of the wave under consideration is a shock wave that burns the
gas. In front of this shock wave p = pi, V = Vi. For fixed values of parameters
Pi, Vi and Q = Ci — C2 there exist many solutions corresponding to different values
of wave velocity relative to gas in front of it. Find the set of values of p and V far
behind of such waves. Show that they lie on the detonation adiabat. What part
of the detonation adiabat do they occupy if a) uêm changes monotonically in the
wave; b) Uchem monotonically decreases from C\ to the given value C„uD and then
monotonically increases up to C2. Check if the corresponding detonation fronts are
evolutionary ones.
© 24.43 A tube is filled by gaseous fuel at rest with p = pu p = p x . At a moment

t = 0 a flat piston starts moving in it with a constant velocity. Supposing that at
the same moment the detonation front leaves the piston and begins to move over the
gas, perform a qualitative study of the gas flow with conditions of Problem 24.42.
Investigate the dependence of the solution type on the velocity of the piston in cases
a) and b) of Problem 24.42.
© 24.44 Consider the structure of slow combustion front in a perfect, ideal, heat-
conductive combustible gas; i.e., investigate the one-dimensional stationary wave with
the beginning of chemical reaction in the gas (igniting) initiated by its warming-up
at the expense of heat conductivity. Neglect diffusion of the products of burning.
Determine the variation of parameters inside the wave and its velocity. To make
calculations easier assume that: 1) the thermal conductivity coefficient is constant;

2) the rate of chemical energy generating dq/dt {q is the amount of chemical energy
generated per unit mass of the gas) for T > T, is proportional to the absolute tem
perature T and to the amount of unburned gas, which is proportional to Q — q in
case without diffusion (Q is the total chemical energy variation per unit mass)
^ = aT(Q -q) at T > T.; -^ = 0 at T < T,

dt dt
Here T, is the temperature at the start of the reaction, T. > Tb> ô is the initial gas
temperature (ahead of the wave); 3) the flame front velocity (determined from the
solution of the problem) is so small that it can be assumed that p = const in the wave
and the kinetic energy flux can be omitted in the energy flux expression in the frame
attached to the wave.
Steady motion of a compressible fluid

0 24.45 Write the Bernoulli integral over a streamline for steady, adiabatic motion
of an ideal perfect gas if there are no body forces. Determine the expressions for the
maximal velocity vmax possible on the streamline and for the critical velocity v, (i.e.,
the velocity at the point where v = a = v, = a.) in terms of the stagnation parameters
Po.A), ck>! T0 (i.e., the parameters of state for the particle on the streamline for which
v = 0). Calculate » „ , and v, for air at T0 = 15°C {cp/cv = IA,R = 287.14 m 2 /(s 2 K).
Compare t>max with the velocity of unsteady outflow into vacuum (see the solution to
Problem 24.30).
0 24.46 Estimate the effect of compressibility on the dependence of the pressure

on the velocity in a steady, adiabatic stream of a perfect gas. Compare the relations
between p/po and velocity v for a compressible fluid with those for an incompressible
one for velocities that are not too large: u/ao < 1; here, po is the stagnation pressure
(see Problem 24.45). For what velocities can the model of incompressible fluid be
used to calculate the pressure in a steady stream of air if the admissible error of the
calculation is less than or equal to 1%?
0 24.47 Find the velocity field of the adiabatic, steady flow caused by a spatial
source with mass flow rate Q = const in a perfect gas. The flow is spherically
symmetric, all parameters depend only upon r. Show that such flow can take place
only in the region outside a certain sphere with radius r m j n on which the Mach number
is M = 1.
0 24.48 For a steady, planar flow caused by a point vortex (curlv = 0, vr = 0,

Vy, = v(r) in the polar coordinate system r, ip) in a perfect gas under the condition
p = Op"1, 6 = const, determine the distributions of velocity v(r) and pressure p(r).
178 FLUID MECHANICS
Show that such flow takes place only outside a certain circumference on which M = oo,
and there exists a circumference in the region of the flow on which M = 1.
0 24.49 In study of steady, continuous motions of a fluid in weakly curved tubes
with smoothly varying shape and area of the cross section, a quasi-one-dimensional
description can be applied; i.e., only the value of pressure, density and longitudinal
velocity averaged over the section as functions of the distance x along the tube can
be considered. The true values of the parameters differ only slightly from the average
ones almost everywhere. The viscous normal stresses and heat fluxes in the cross
sections can be neglected. Considering the volume between two near cross sections
of the tube as a control volume, obtain the continuity, motion and energy equations
in the form
d(pvQ.) = 0 ,
v dv + - dp = Fxdx + f dx ,
P
— + u + -\ = Fxdx + dq
where v is the longitudinal velocity, ft is the cross-sectional area, Fx is the body force
acting along the tube (the body force acting across the tube is neglected), / is the
force of wall friction per unit mass of the fluid and per unit length of the tube, u is
the density of internal energy, dq is the heat input per unit mass of the fluid.
© 24.50 A steady, adiabatic motion takes place in a weakly curved tube with
smoothly varying shape and area of the cross section (see Problem 24.49). There are
no body forces. Determine the relation between the variation of velocity along the
tube and the variation of its cross-sectional area.
0 24.51 There is an adiabatic, steady motion of an ideal gas in a vertical cylindri
cal tube with constant cross section. How does the velocity change under the action
of gravity if the gas moves a) downward, b) upward?
© 24.52 There is a steady, adiabatic motion of a perfect gas in a horizontal cylin
drical tube with constant cross section, so long that friction should be taken into
account. There are no body forces. How does the velocity of the gas change under
the action of friction?
0 24.53 There is a steady motion of a perfect gas in a horizontal cylindrical tube
with constant cross section. Through the walls, the gas is supplied with heat, or heat
is conducted away from the gas. Friction and body forces can be neglected. How
does the velocity vary along the tube?
Chapter 6
Elasticity Theory
25 Models of Elasticity Theory

The strains in an elastic medium are uniquely determined by the stresses at constant
temperature. Such a relationship can be written in one of the forms
£ij = £ij(Pki,T) or Pki=Pu(£ij,T) . (25.1)
Here, £y are the components of the strain tensor, pkl are the components of the stress
tensor, T is the temperature. Deformation processes are regarded as reversible. To
write the closed system of equations for elastic medium, it is necessary, besides (25.1),
to specify the internal energy U (per unit mass) as a function of the corresponding
variables pki, T or e„, T (see Chapter 3). In the particular case when the dependence
(25.1) between £i} and pki is linear at T = const, the relationships
Pki = A\\eij , T = const
are called Hooke 's law . The constant term of the linear function equals zero because
it is usually taken that the strains equal zero in a non-stressed state. Instead of the
equations of state (25.1), their representation in terms of thermodynamic potentials
of elastic medium can be used. The equations of the first and second thermodynamic
laws in Lagrangian variables in the current state have the form
dU=-piidell+dqe
P
Tds = dqe
where U, s, and dqe are the internal energy, entropy, and heat input to the particle
per unit mass, respectively. Eliminating the external input dqe, one obtains the rela
tionship binding only internal parameters of the medium and called Gibbs's identity
179
180 ELASTICITY THEORY
dU = -jFtkii + Tds. (25.2)

P
If the internal energy U can be regarded as a function of the state variables £y, and
s, then the equality (25.2) is equivalent to the system of equations of state
V-(J£)
ae
. T.m a s (25.3)
P \ H I s=const \ I Ej,=const
in which the first formula is another representation of the equations of state (25.1).
Thus the internal energy given as a function of e^ and s is a thermodynamic potential
(see Problem 15.9). Apart from the internal energy, another thermodynamic potential
is often used in elasticity theory — free energy (per unit mass) T = U — sT, given
as a function of the parameters of state £i;- and T. Gibbs's identity (25.2) expressed
in terms of the function T has the form
cLF=-piideij-sdT
P
and allows one to obtain another form of the equations of state
The system of differential equations of the elasticity theory in Lagrangian coordinates

includes 1) the continuity equation
P\/9 = PQy/%, 5 = det||g 0 ||
where po and p are the densities of the medium in the initial and current states, \\gi: ||
is the matrix of components of the metric tensor; 2) the equations of motion
pa{ = VjPij + pF{
where a* are the components of the acceleration and F* are the components of the
body forces; 3) the equations of the first and second laws of thermodynamics, one of
which can be replaced by Gibbs' identity (25.2)
dU = -pijdeij + Tds
P
Tds = dqe
25. Models of Elasticity Theory 181
4) the equations of state (25.3) or (25.4) where the internal energy (or the free energy
) is a given function: U = f/(£y,s) {T = ^"(ey.T)). 5) components £y of Green
strain tensor are expressed in terms of displacement vector components W{
£ij = îWj + VjWi + ViWk VjWk)
The heat input dqe, if it is caused by heat conduction, can be written, according to
the Fourier heat conduction law, in the form
dq* = - d i v (/cgradT) dt
P
where K is the coefficient of heat conductivity of the material. It is obvious that
to use the function T is more convenient in the study of isothermal deformation
(T — const), and the function U, in the study of adiabatic deformation (s — const)
em Boundary conditions for the elasticity equations may have one of the following
forms: 1) a displacement vector w is given on the whole surface of the solid body,
or 2) a stress vector p n is given on the whole surface, or 3) a displacement vector on
one part of the body surface and a stress vector on the other part of it are given.
In linear elasticity theory it is supposed that the displacement gradients ViiUj and,
consequently, the components of the strain tensor
ey = 2 WiWj + VjWi) . (25.5)
are small. In this case, there is neither any difference between Eulerian and La-
grangian descriptions, nor between initial and current Lagrangian coordinates. The
functions £/(£„,s) and T{£ij,T) for small strains and small changes in temperature
can be expanded into power series with respect to e y , T — TQ, and s - s 0 , all the
terms of order more than 2 being neglected; then the equations of state (25.3), (25.4)
become linear. For an isotropic linear elastic medium, this expansion is naturally
expressed in terms of scalar invariants of the strain tensor
pU = \ \'Jl + i/J2 - a'J^s - s 0 ) - - b\s - s 0 ) 2 + d'(s - s0) ,
pT=\ A J\ + n h ~ aJi (T - T„) - i b(T - T 0 ) 2 + d(T - T0)
where
î = ej , J2 = £ij£,J •
Hooke's law for an isotropic material at T = const has the form
Pij = >-£kk9ij + %p£ij , t, J, fc = 1, 2,3 .

The coefficients A and fi are called the Lame coefficients.

In linear elasticity theory, the boundary conditions for the system of differential
equations are given on the initial undeformed surface. For static problems the St.-
Venant 's principle is often used, which is formulated as follows. Let a system of forces
and moments arts in a certain domain of a body. The strains and stresses caused
by this system far from the domain approximately do not depend on details of the
forces and moments distribution. They depend essentially on the resultant and the
total moment of the forces only. So to find strains and stresses in the points which
are far from some part of the boundary, the boundary conditions on this part may
be replaced by any convinient statically equivalent ones.
Since the problems are linear, the method of superposition is applicable. The
theorem of uniqueness for the solutions of static and dynamic problems is proved,
which allows use of the semi-inverse method (of guessing partially the solution). In
the linear theory, using Hooke's law (at T = const) and the expressions (25.5) for E,},
one can obtain the equations of motion for the displacements (Lame equations)
d2w
p -5-5- = (A + u.) grad div w + p.Aw + pF .
at*
If the boundary conditions are given in terms of the stresses, the problem of equilib
rium for an elastic body can be solved with respect to the stresses. In this case, the
equations of equilibrium
V,p,; + pF} = 0
are used together with the equations of Beltrami-Michel, which are derived from the
compatibility conditions for strains with Hooke's law and the equations of equilibrium
taken into account. If F3 = const and T = const, the equations of Beltrami-Michel
have the form
Ap y + 1 ^ ^ V,V}Mp) = 0 , J,(p) = Pkk ,
or, for each component of the stress tensor,
AAp,_, = 0 .
Here, A is the Laplacian, and AA is the biharmonic operator. An experiment in
which simple extention takes place (there is only one nonzero component of the
stress tensor pn) demonstrates Hooke's law for isotropic solid in its simplest form
Pn = Eeu. The proportionality coefficient E is called Young's modulus . The elon
gation £n of the specimen in the direction of the acting force is accompanied by
contractions t?i and £33 in the transverse directions. The ratio of a transverse strain
to the longitudinal one -£22/^11 is called Poisson's ratio a. E and a are expressed
in terms of A and p.. The relative change in volume © due to deformation equals
25. Models of Elasticity Theory 183
the first invariant of the strain tensor J\ (e) in the linear approximation. The coef
ficients E and a characterize the elastic properties of a medium and are regarded
as constant for every homogeneous material. Other elastic characteristics, also often
used, are introduced below: the bulk modulus K (Problem 26.5), the shear modulus
G = n (Problem 26.3), and the heat expansion coefficient a (Problem 26.10). Hooke's
law for an anisotropic medium py = Aijkl£ti contains, in general, 81 constants Aijkl.
However, due to symmetry of strain and stress tensors, only 36 of them are differ
ent. The thermodynamic relationships (25.3) reduce the number of elastic constants
to 21. Possible symmetry of the material further reduces their number. Symmetric
properties of a medium can be described in terms of vectors and tensors of special
form. Often in applied mechanics media are treated as orthotropic or transversely
isotropic. A material is referred to as orthotropic if there exist three mutually per
pendicular planes of symmetry for it (rhombic system of symmetry). To describe
symmetric properties of this medium, a symmetric second rank tensor dij giving the
properties of a specific orthotropic material is used. Then the free energy density is
represented by the function T = ^ ( e ^ , 5ij,dy,T). A transversely isotropic medium
has an axis of symmetry of infinite order, i.e., such that, on rotating around it by any
angle, the medium is transformed into itself (such symmetric properties are intrinsic
for textures). This model is suitable for many laminate and filamentary materials.
The symmetric properties are characterized by a certain direction I = {U}, only the
line of action of this vector being important, opposite directions being equivalent. To
describe this type of symmetry, the symmetric second rank tensor d^ = l{lj, where Z<
are direction cosines of the axis of symmetry, can be used.
In non-linear elasticity theory, in addition to the Cauchy stress tensor p^ (the
components of which determine the stress vector pn related to the surface element E
with the normal n) the Piola-Kirchhoff stress tensor TX^ is often used (see Section 8).
The "nominal" stress vector nna related to that surface element Eo with normal n0
with which the surface element E coincided in the initial state is determined by 7iy,
and nj). If F is the surface force on E at the current instant, then
JP JP
The tensor 7ry, is not symmetric in the general case. It follows from Gibbs' identity
that
dppU \ l dPoT
3
" \d(^wj))s U v ^ ) / T
The function p^T = $ is often called the elastic potential of the medium. In terms
of the tensor 7Ty, the equations of motion in Lagrangian variables take the form
d2u>i cfaij
it being always possible to choose the coordinate system connected with the initial
undeformed state in the most convenient manner, e.g, to choose a Cartesian rectilin
ear system or one corresponding to the symmetry of the problem. In this approach,
in place of the components of the strain tensor e^, the components VjWj of an asym
metric tensor are used. This tensor is called the distortion tensor . In linear elasticity
theory, the components 7Ty and py coincide. In the description of motion (equilib
rium) for an elastic medium in Eulerian variables, the internal (free) energy given as
a function of the strain components in the Eulerian immovable frame of reference —
[/(efj, s) — should be used. However, in the equations of state (25.3), differentiation
with respect to the components e^ in the Lagrangian basis is implied. Therefore,
in the case of nonlinear deformation, the strain increments should be converted in
obtaining the stress components py from £/(£y) (see Problem 15.9). In linear theory,
Lagrangian and Eulerian components e^ coincide.
26 Linearly Elastic Solid
— PROBLEMS —
Constitutive relations
© 26.1 A prismatic rod made of linearly elastic material is in equilibrium under
the action of extensional forces uniformly distributed on the ends, the lateral faces
being free (simple extension). Find the components of the strain tensor, regarding
the magnitude of stresses on the ends p as given. Determine the relationship between
the elastic constants of the medium E (Young's modulus) and a (Poison's ratio) and
the Lame coefficients A and fi.
0 26.2 A specimen made of a linearly elastic material is positioned between two

pairs of parallel rigid walls so that its transverse sizes cannot vary. Uniform compres-
sive stresses p act on the ends of the specimen. Find the stresses and strains in the
specimen, assuming that there is no friction between it and the walls.
0 26.3 A layer of an elastic material is positioned between two infinite planes
perpendicular to the y-axis. The layer is fixed on one of these planes y = 0, and
uniformly distributed tangential stresses p 12 = r act on the second boundary (see
Figure 26.1). Deformation arising under these conditions is called simple shear. Find
the components of the strain tensor and the quantity 7 — the change in inclination of
the material line element initially parallel to the j/-axis. The proportionality coefficient
between 7 and T is called the shear modulus. Find the expression for it in terms of
the Lame coefficients A and /x.
26. Linearly Elastic Solid 185
Figure 26.1
0 26.4 Find the relative change in volume of the medium 0 for the strains ob
tained in a) Problem 26.1, b) Problem 26.2, c) Problem 26.3. Note that the volume
does not change for simple shear (see Problem 26.3).
© 26.5 A state of stress given by a spherical stress tensor p^ = —p<7y is called
isotropic compression. Determine the components of the strain tensor and the relative
change in volume 0 . The proportionality coefficient between p and 0 is called the
bulk modulus K. Find the expression for K in terms of E and a and in terms of the
Lame coefficients A and fi.
© 26.6 Using the preceding problem, show that Poisson's ratio for an incompress
ible elastic medium is a = 1/2.
0 26.7 Express the function F, the free energy at T = const, as the sum of
the energy of volume deformation (see Problem 26.5) and the energy of shear (see
Problems 26.3, 26.4).
© 26.8 Regarding the strains £y and change in temperature T — TQ as small,
determine the most general form of the function of free energy for a) an arbitrary
linearly elastic medium, b) for an isotropic medium. Using the obtained expressions
for T and the thermodynamic relationships, find the expression for the components
of the stress tensor Pf,.
© 26.9 Execute the task of Problem 26.8 if the medium is specified by its internal
energy U(eijts).
0 26.10 To determine the physical meaning of the coefficient a of (T — TQ) in
Hooke's law for isotropic thermoelastic solid (see Problem 26.8)
Pij = AJi^j + 2//e0 - a(T - T0)<5y , Ji = ekk
solve the problem of determination of strains caused by a change in temperature in
absence of stresses. Show that the strain tensor is spherical. The proportionality
coefficient between the change in volume 0 and the relative value of AT = T -T0
is called the thermal expansion coefficient c*e; the proportionality coefficient between
AT" and the relative elongation of the segments is called the linear thermal expansion
coefficient. Find the expression for the coefficient a in terms of A, p, a or in terms of
a and K (the bulk modulus (see Problem 26.5)).
© 26.11 For an isotropic medium, the expression for an entropy can be obtained
from the expression for free energy T
Pos = -po -~= = or(3A + 2p) Ji + b(T - T0) + p0s0 .
To determine the physical meaning of the coefficient b, consider heating of an elastic

medium at fixed volume (G = J\ = 0). Show that b is expressed in terms of the heat
capacity coefficient at constant volume cv = (dg/dT)e=o-
0 26.12 For an isotropic linear-thermoelastic medium, write the expression for
the components of the strain tensor £*, in terms of the stress components py.
© 26.13 An isotropic elastic medium is specified by the free energy p^T = | Jf +
p.J-i — aJ\ (T — To) — | (T — To)2. The physical meanings of the coefficients o and b is
established in Problems 26.10 and 26.11. Find the change in temperature T — To at
the relative change in volume equal to 9 if there is no heat exchange between parts
of the medium. Write the dependence of the stresses upon the strains for the process
of adiabatic deformation.
0 26.14 Hooke's law expressing the relationship between the stresses and strains
is established for isothermal processes of deformation. In connection with this fact,
the coefficients A and p. and the bulk modulus K and shear modulus G correspond
ing to them are called the isothermal coefficients of a medium. The corresponding
proportionality coefficients between py and e^ for adiabatic deformation are intro
duced similarly (see the preceding problem). Find the relation between isothermal
and adiabatic coefficients A, p., K.
0 26.15 In an elastic medium specified by its free energy ^"(e^.T), the stresses
are determined by the formula p ,J = p(dJr/deij). Show that in linear elasticity there
exists a thermodynamic function $(pij,T) such that e^ = p (d$/dpij). Determine its
form.
Q 26.16 An elastic medium is specified by its free energy ^ ( e y , T). Determine
the condition for stability at T = const of the uniformly deformed state with ei:, = e*
under the action of time-invariant external stresses p y * = df/de'j given on the
boundary. Consider, as a particular case, the linear medium with pT = \ \J\ + pJ^
Ji = £«', J2 = en?*.
© 26.17 Write the expression for a free energy F(eij, dij, T) for a linear orthotropic
medium. How many elastic constants are needed to specify it at T = const?
0 26.18 The axis of symmetry for a transversely isotropic medium is taken as the
X3-axis of a Cartesian coordinate system xt. Write the expression for a free energy
T. Show that this medium is specified by five elastic constants. Determine the
dependence of stresses on strains. Use the results of the preceding problem (T =
const).
0 26.19 Find the components of the strain tensor for a transversely isotropic
medium isotropically compressed by the stress p. The axis of symmetry is directed
along the i 3 -axis. Use the relationships between p^ and e^ obtained in the preceding
problem.
0 26.20 Write the equations of motion in terms of the displacements for a trans
versely isotropic medium (direct the axis x3 = z along the axis of symmetry of the
material) (T = const).
© 26.21 Show that, for an isotropic elastic medium in equilibrium in the absence
of body forces, the function 8 = divw; is harmonic; and the displacement vector w
satisfies the biharmonic equation. Verify that this is also valid if uniformly distributed
body forces (F = const) act.
© 26.22 Show that, for an isotropic elastic medium in equilibrium in the absence
of body forces, the stress components p,j are biharmonic functions.
© 26.23 A solution of the equations of motion in terms of displacements (the
Lame equations) in the absence of body forces can be sought as the sum of terms
■u> = toj -f Wi one of which is a harmonic vector function: Awi = 0, and the other is
a potential vector: W2 = grades (representation of Papcovic-Niber). Show that, for
this case, the problem is reduced to determination of four harmonic functions.
Equilibrium under extension and shear

In the problems presented below, specimens are considered whose length is much more
than the transverse sizes. Therefore, due to the St.Venant principle, in determination
of the stress-strain state in the whole region except in the vicinity of the ends the
load distribution on the ends may be replaced by any statically equivalent one.
0 26.24 Determine the strain of a vertically standing rod of length I being in
stable equilibrium in the field of gravity forces g (see Figure 26.2).
0 26.25 Find the elongation 6 and change in volume AV of a rod with an initial
length I and weight P suspended vertically in the field of gravity forces.
g
1
y
" ///////////////////
Figure 26.2
© 26.26 Of what maximal length can a steel rope of diameter 2 mm be to sustain

a load of a weight 100 kgf? The strength limit of steel is 3500 kgf/cm2, the density
is 7.85 g/cm 3 .
0 26.27 What maximal extension strength can be applied to a steel prismatic rod
of cross section 5 x 5 cm if the limit allowable tensile stress is 1400 kgf/cm2, and the
limit allowable tangential stress is 600 kgf/cm2?
0 26.28 Find the expression for the elongation 6 of a cylindrical rod of an initial
length /, one of the ends of which is fixed to a shaft rotating with a constant angular
speed ui around a vertical axis (see Figure 26.3). Calculate this elongation for a
Figure 26.3
copper rod of a length 1 m if w = 30 s '. The density of copper is 8.87 g/cm ,

E = 1.05 • 106 kgf/cm2. Neglect the weight of the rod.
© 26.29 A specimen of length I with a rectangular cross section of constant thick
ness a = 1 has a longitudinal section in the shape of a trapezoid with bases b and
(>o (see Figure 26.4). One of the bases is fixed, and the other is under the action of
a force P. Find the components of the stress tensor pQ^ and the expression for the
elongation of the specimen 6 assuming that the inclination angle of the sides of the
/////// ///////////
TJ
y* "
p
Figure 26.4
trapezoid 6 is small, a = tan 6, a = (6 - b0)/l -C 1 so that a 2 can be neglected.

Neglect the weight of the rod.
© 26.30 A rod whose cross section slightly varies along its length is suspended in
a vertical plane, being under the action of gravity and a force P uniformly distributed
over the lower cross section S0 (see Figure 26.5). Of what shape should it be (find
i x
/////////
y
V
S„
P
Figure 26.5
S(x)) for the tensile stress to be identical in every cross section? (Due to small
variation of the shape, the specimen can be regarded as prismatic when considering
boundary conditions.)
© 26.31 Determine the shape b(x) of a turbine blade rotating around the axis with
an angular speed u) (see Figure 26.6) at which the stress in every rectangular cross
section perpendicular to the x-axis is identical. (The variation of the shape is small,
so that the boundary condition can be satisfied approximately as in the preceding
problem.)
Figure 26.6
© 26.32 A plate in the shape of a rectangle experiences bi-axial extension in its

plane under the action of uniform stresses pi and P2 on its ends (see Figure 26.7).
The front and back planes bounding the plate are free from stresses. Find the strain
Figure 26.7
components £y and the relative change in volume. Calculate the maximal tangential
stress in the plate.
Bending of b e a m s
We consider a body whose dimension in one direction is much greater than the di
mensions in the two other directions. If such a body is subjected mostly to bending
and stretching (or compressing), it is referred to as a beam. In pure bending of a beam
(see Figure 26.8), there exists a material line which does not experience extension or
compression, but only bends. It can be called the neutral axis. The curvature of this
line is determined by the bending moment which is constant along the beam. The
plane cross sections of the beam remain plane and orthogonal to the bent neutral
axis after deformation. This is a result of the exact solution for the problem of pure
bending. In practice bending is usually caused not by a moment, but by forces acting
transversely to the beam. For small deflections and sufficiently long and thin beams,
the following is taken as a hypothesis in problems of bending: in the presence of
transverse forces in a beam, as in pure bending, 1) there exists a neutral axis which
does not experience elongation, 2) plane cross sections of the beam remain plane
and orthogonal to the neutral axis after deformation. This statement is called the
hypothesis of plane sections. The approximate method of beam analysis based on
these suppositions is considered in Strength of materials. Of course, the St.Venant
principle is also used in this analysis. The analysis of beams in bending is based on
calculation of the bending moment in each cross section of the beam. This calculation
is carried out by consideration of equilibrium of the part of the beam adjoining the
cross section under consideration from one of the sides. The normal stresses in the
cross sections are obtained, as in the case of pure bending, from the calculated values
of the bending moment which is, however, variable along the axis of the beam in this
case. This approximate method should be used in solution of Problems 26.35-26.41.
In all these problems it is assumed that the beam has a vertical plane of symmetry
coinciding with the plane of the plot, the forces act in this plane, and bending takes
place in this plane. In some problems, to find the solution, it is necessary to determine
reactions at the place of fixation of the beam, using the condition of equilibrium.
© 26.33 A prismatic beam with constant cross section , one of the ends of which
is rigidly fixed, has lateral faces free from loads (see Figure 26.8). The axis of the
beam is directed along the x-axis and passes through the area centroids of the cross
Figure 26.8
sections. A force couple with a moment M parallel to the 2-axis is applied to the
second end of the rod. There are no other forces. This strained state is called pure
bending. Find the stresses and strains in the rod. Write the equation for the neutral
axis of the rod (the material line which does not experience extension or compression).
Show that plane cross sections remain plane and orthogonal to the bent neutral axis
after deformation.
© 26.34 Supposing that for bending of a beam the hypothesis of plane sections
is valid, determine the longitudinal stress and strain distributions, considering the
shape of a deformed element of the rod. Write the equation of the bent neutral axis.
© 26.35 A beam rigidly fixed at one of the ends is called a cantilever. A cantilever
beam of length / is bent by a force P applied to its end perpendicularly to the axis
(see Figure 26.9). The material of the beam (Young's modulus E) and the shape
/
/_
Figure 26.9
of its cross section are given. Supposing that the hypothesis of plane sections are
valid, find the stress pn in each cross section, write the equation for the neutral axis,
calculate the maximal deflection t).
0 26.36 Show that the differential equation for the deflection rj of the neutral axis
of a beam in equilibrium can be written in the form
dx* El
where q(x) is the density of transverse load distribution along the beam projected
onto the y-axis, the load is applied in the vertical plane of symmetry of the beam —
the plane z = 0; / is the moment of inertia of its cross section relative to the z-axis,
the axis of bending.
© 26.37 A horizontal beam of length /, rigidly fixed at each end, is under the
action of own weight (see Figure 26.10). Supposing that the hypothesis of plane
K
G ■e-,
Figure 26.10
sections is valid, determine the shape of the beam neutral axis »?(x), its maximal
deflection, and the stress pn in each section. The weight per unit length of the beam
equals q, and Young's modulus equals E. (Take into account that reaction forces R
and moments M arising in a rigidly fixed beam are unknown in advance.)
© 26.38 A beam is referred to as simply supported if both its ends are hinged,
and one of the hinges can freely slide in the axial direction (see Figure 26.11). Solve
the preceding problem for such a beam.
ky
Figure 26.11
© 26.39 Solve Problem 26.37 for a beam rigidly fixed on one of the ends and
simply supported on the other. Determine the position and the value of maximal
deflection.
'•y
Figure 26.12
0 26.40 A simply supported beam (see Problem 26.38) of length / with rectangu
lar cross section of a width h is heated non-uniformly along its thicknessThe temper
ature profile is linear, T\ is the temperature on the upper face, T2, on the lower one,
and T2 > 7i. Determine the bent shape of the beam and the maximal deflection if the
heat extension coefficient of the material equals a, and Young's modulus equals E.
Q 26.41 Find the expression for the tangential stresses p\2 in a cantilever due to
bending under the action of its own weight (see Figure 26.13). Determine where the
tangential stress is maximal.
Figure 26.13
© 26.42 A beam with rectangular cross section consists of two layers of different
materials connected so that it functions as whole (bimetallic beam) (see Figure 26.14).
ky
b/2.
■±kL -7-
r
M
h2
7 My
h,
Figure 26.14
Its width is b, the thickness of the layer of the material with Young's modulus E\
equals h\, the second layer has thickness h^ and Young's modulus E2. The beam
is in a state of pure bending with moment M in a vertical plane. Determine the
distribution of the stresses pn and location of the neutral axis o.
Torsion of rods
A strain state caused by force couples with moment M directed along the axis of the
rod is called torsion. If the moments are applied only on the ends of the rod, then the
displacement components are W\ = —azy, w2 = azx, W3 = f(x, y), and the constant
multiplier a is the twisting angle per unit length, z-axis is directed along the axis of
the rod. If the cross section of the rod is circular, then only shear stresses act in the
cross sections (f(x, y) = const), and the plane cross sections remain plane and rotate
around the axis of the rod as rigid disks.
Q 26.43 A rod with circular cross section of radius R is under the action of a
torque M. Determine the dependence of the twisting angle per unit length of the
rod a on the magnitude of the torque M.
© 26.44 Find the maximal tangential stress in a circular rod of radius R under
the action of a torque M. Calculate the maximal tensile stresses, and find the surface
elements on which they act.
0 26.45 A rod with circular cross section of radius R is loaded on the ends by ten
sile forces P and torques M (see Figure 26.15). Find the maximal tensile, squeezing
and tangential stresses in the rod.
© 26.46 Torques uniformly distributed over the length act on a rod of length I
with circular cross section of radius R, rigidly fixed at the end A (see Figure 26.16).
The torque distribution intensity (i.e., the magnitude of the torque per unit length
of the rod) equals q. Find the twisting angle <p of the end of the rod B.
-P
-M€E B-sr
Figure 26.15
B
■( \ \ ^ ^ ^
1/ J J J J )
Figure 26.16
0 26.47 A solid shaft of circular cross section consists of two parts with lengths
a and b and radii Ra and /?&, respectively (see Figure 26.17). The shaft is fixed at
the ends and loaded by a torque MQ at the interface section. Find the moments of
Af„
ii.
/ ■
a B
\J
Figure 26.17
reactions MA and MB arising at the sections of fixation. Find the rotation angle tp
of the section at which the torque act.
© 26.48 Two shafts, one hollow and the other solid, are made of the same material
and have the same external radius R. The internal radius of the hollow shaft equals
0.6 R. Compare their weights and maximal tangential stresses, supposing that both
shafts are loaded by the same torque.
0 26.49 A shaft consists of a rod over which a tube is slipped, and the tube is
rigidly connected to the rod. The radius of the rod is Ra, and the external radius of
the tube is Rb (see Figure 26.18). The shear moduli of the materials of the tube and
rod are respectively Ga and G<>. The shaft is under the action of a torque M. Find
the maximal tangential stress r in each of the materials and the twisting angle a per
unit length of the shaft.
Figure 26.18
0 26.50 A shaft of arbitrary cross section is in equilibrium under the action of

torques M at the end sections. Show that a stress function T can be introduced so
that P13 = df/dy, p-& = —dT/dx, and T satisfies the Poisson equation. Write the
boundary conditions for T on the cross-sectional contour of the shaft.
States of plane stress and plane strain

A problem of elasticity is called the plane problem if the equalites
Pii =Pii(x,y), Pl2=Pn{x,y), Pl3=P23 = 0

£a = en(x,y), ei2 = £u{x,y), eu = £23 = 0 (26.1)
i = l,2,3
(no summation over i) are fulfilled simultaneously. The body forces and the boundary
conditions are independent of z. The z-components of body forces are equal to zero.
© 26.51 A state of deformation is called plane strain state if besides the conditions
(26.2) the equality £33 = 0 is fulfilled. Write the expressions for the components of
the stress tensor, and determine the relationships among them for this case.
© 26.52 A state is called plane stress state if besides the conditions (26.2) there is
P33 = 0. Find the strain components and the relationships among them. Show that,
if there is no change in volume, a state of plane stress coincides with a state of plane
strain. Write the two-dimensional equation of motion for the displacements.
0 26.53 In solution of problems for stresses, besides the equations of equilibrium,
the compatibility conditions must be satisfied. Show that, for a plane problem, the
compatibility conditions are reduced to one equation for pa/3 (a, 0 = 1,2) replacing
the equations of Beltrami-Michel. Write this equation.
G 26.54 For equilibrium in absence of body forces in a state of plane stress, it is

possible to introduce the stress function of Airy <p(x,y) such that pn = d2ip/dy2,
P22 = d2<p/dx2, P12 = —îp/Oxdy. a) Derive the equation satisfied by the function
ip due to the compatibility equations, b) Show that it is also possible to introduce

the stress function in the presence of a constant body force, e.g., the gravity force
Fy = -g.
© 26.55 The solution of the two-dimensional biharmonic equation AAv? = 0 for

the stress function of Airy can be sought in the form of homogeneous polynomials
of various powers with coefficients consistent with the boundary conditions. For a
rectangular plate with the sides parallel to the axes x, y, the function ip can have the
form <p = A ax 2 -|- bxy + | cy2. Of what problem does this represent a solution?
0 26.56 Let the stress function tp satisfying the biharmonic equation have the
form ip = ax 3 + bx2y + cxy2 + dy3. Consider a rectangular plate of width 2h and of
length 21 with the sides parallel to the axes x, y (see Figure 26.19). Of what problem
does this represent a solution in the case a) a = 6 = c = 0, d ^ O , b) a = 6 = d = 0,
\y
I l
h
Figure 26.19
cÔ?
© 26.57 Express prT,Pee,Pre for plane strain in terms of the stress function (p(r, 6)
in polar coordinates. Write the equation satisfied by the function ip.
© 26.58 Find the general form of the Airy stress function for plane stresses with
circular symmetry (ip = tp(r)). Write the general expression for the stress components
Prr, Pee, Pre-
Various problems of equilibrium of elastic bodies

© 26.59 There is a spherical cavity of radius R filled with gas under pressure
Po in an unbounded elastic medium. The medium is free from stresses at infinity.
Determine the components of the stress and strain tensors.
© 26.60 Determine the strain of a hollow ball, the internal and external radii of
which are Ri and R^, that is under action of pressures p\ from inside and p^ from
outside.
Q 26.61 An unbounded elastic medium experiences isotropic uniform compression

by a pressure p. There is a cavity of radius a in the medium. The pressure in the
cavity equals zero (vacuum) (see Figure 26.20) Determine the stress distribution.
Figure 26.20
Show that the maximum stress is reached on the boundary of the cavity and exceeds
the pressure at infinity (stress concentration on a hole).
© 26.62 Determine the stresses in a circular cylindrical tube (with external radius
i?2 and internal radius Ri) that is under action of the pressure p\ from inside and the
pressure pi from outside. The tube is sufficiently long, and its ends are fixed so that
there are no displacements along the tube (a state of plane strain).
0 26.63 A circular hole of radius a is cut out in an unbounded plate (a plane

problem). The pressure inside of the hole equals zero. The plate is, at infinity, in the
state of uniform isotropic tension by stresses of magnitude p. Determine the stress
distribution. Show that the maximal stress is reached on the boundary of the cavity
and exceeds p (stress concentration on a hole). Compare with the results of solution
of Problem 26.61.
© 26.64 Determine the strain of a solid sphere with radius R under the action
of its own gravitational field; regarding the external boundary of the sphere as free
from stresses. Show that there are regions where the matter is compressed and those
where it is extended along the radius. Indicate the interface between them.
© 26.65 Find the stresses in a disk of radius a and of constant thickness rotating
with a constant angular speed w around the z-axis, perpendicular to its plane. The
surface of the disk is free from stresses, and the thickness is so small that one can
assume that pzz = 0 through the thickness.
© 26.66 Find the stresses and strains in an elastic half-plane on the boundary of
which a concentrated force F is applied (see Figure 26.21). There are no other forces
on the boundary. Use the stress function <p in polar coordinates (see Problem 26.57).
Figure 26.21
© 26.67 Solve the preceding problem in Cartesian coordinates, using the Fourier
transformation of the equilibrium equations and of the boundary conditions.
© 26.68 Determine the stress distribution in an unbounded plate with a circular

hole of radius R subjected, at infinity, to uniform extension in the direction of the
x-axis by stresses of magnitude T (see Figure 26.22).
Figure 26.22
Thermal strains
0 26.69 A rectangular plate of the thickness 2a is clamped at the ends between
two parallel planes and has free upper and lower faces (see Figure 26.23). The plate
is heated nonuniformly over thickness so that the profile of the temperature in each
section perpendicular to the x-axis is given by the function T = TQ(1 — y2/a2). De
termine the distribution of squeezing stresses pn. Assume that the strains at all the
points of the plate equal zero at T = Tav.
© 26.70 Write the equilibrium equation of an elastic medium for displacements in

the presence of thermal strains. How is this equation simplified if the displacements
/
/ /
/ /
/ -
',- a I
~-'r±
/ - /
s
's /
Figure 26.23
are a) axially symmetric for plane deformation, b) spherically symmetric, c) axially

symmetric for a plane stress.
© 26.71 The temperature in a circular thin disk of radius R and of constant

thickness varies from the center to the periphery as a function T = T(r). All the
surfaces of the disk are free from stresses, the thickness is small, so that a plane
stressed state is realized. Determine the stress in the disk caused by non-uniformity
of the temperature field. On the external boundary of the disk T(R) = 0.
© 26.72 There is a circular region of radius a, where a constant temperature T0 is

sustained, in the center of a circular elastic disk (see the previous problem). On the
external boundary of the disk (r = R), there are no stresses, and the temperature
equals zero. Find the radial stresses in the disk.
0 26.73 Determine the strain of a nonuniformly heated elastic cylinder with ax
ially symmetric temperature distribution T(r). There are no axial displacements
(plane strain). The temperature on the external boundary of the cylinder equals
zero.
© 26.74 Determine stresses in a long circular tube (plane deformation) with in
ternal and external radii a and b if the temperature equals To = const inside the tube
and T(b) = 0 outside it. The external and internal surfaces of the tube are free from
stresses.
Q 26.75 Determine the strain of a nonuniformly heated elastic ball with spheri
cally symmetric temperature distribution. Take T[R) — 0 on the external boundary
of the ball.
© 26.76 Find the stresses in an elastic ball with a cavity of radius a if the temper
ature inside the cavity equals T0 = const, and the temperature outside the ball (r = b)
equals zero. (Beforehand, determine the temperature distribution in the medium.)
The external surface of the ball and the surface of the cavity are free from stresses.
Stability of equilibrium
The existence of several solutions for a problem of equilibrium requires a choice from
among them of a solution that can be realized. One of the criteria is an analysis of the
stability of equilibrium. The simplest (static) method of this type of analysis consists
of the following: Add to the acting forces a small disturbance. If any sufficiently
small deviation from an equilibrium position requires that the additional forces per
form positive work, this equilibrium position is stable. Usually, the fact that there
exist several solutions corresponding to the same distribution of forces (bifurcation
of solutions) itself indicates instability of at least one of these solutions. Determina
tion of the critical force (conditions) resulting in bifurcation of solutions is the basis
of Euler's method of equilibrium stability analysis. The method of analysis for the
stability of rods is based on the hypothesis of plane sections, according to which the
deformation of a beam is completely determined by the shape of its neutral axis; the
shape of a section is characterized by the moment of inertia / of the section relative
to the axis of bending.
© 26.77 Show that the shape of a thin beam subjected to a compressive force
along its axis is not uniquely determined if the force takes a critical value (bifurcation
of solutions). Determine the possible shapes of the rod with length /, taking into
account nonlinear terms — the moments produced by the force P (see Figure 26.24)
due to virtual small deflections n. Find the critical value of the force P„ at which
Figure 26.24
bifurcation of solutions arises. Assume that both ends of the rod are hinged.
Q 26.78 Solve the preceding problem for a rod rigidly fixed on one of the ends
and free on the other (see Figure 26.25).
0 26.79 Analyze stability of the rectilinear shape of an axially compressed beam

(see the preceding problem). A disturbance in the form of a small transverse force Q
can be used.
0 26.80 A beam of length / is simply supported (see Problem 26.38) and has initial
small deflection % sin ?f. It is subjected to a system of transverse forces distributed
77777777777777777"
Figure 26.25
along the length of the beam and proportional to the value of deflection: q = a-q (see
Figure 26.26). (Such a force may be caused by rain water collecting in the reservoir
x
j ^ ^ î J L 1 1i \ ^
Figure 26.26
formed by a deflected roof. This problem is called the problem "of a filled reservoir.")
Find the limiting value of the coefficient a at which a beam of a given length I remains
in stable state.
0 26.81 Find the total deflection 77 at mid-span of a simply supported beam of

length / to which the concentrated load P = kr] is applied. The beam is supposed to
have an initial deflection 60 at mid-span. Under what conditions does the equilibrium
1/2 1/2
•y
Figure 26.27
of the beam remain stable?

Dynamical problems of isothermal linear elasticity theory

The following problems deal with one-dimensional unsteady motion in which all func
tions sought depend upon a single spatial variable and time. Such motions are usually
called waves. The solution for each of the considered functions <p(x, t) (since the prob
lem is linear) can be sought for in the form of a sum of the harmonic waves
i(hx±wt)
a cos (kx ±ut) = a Re e
with different k and w (here, i = \/~T. a n d k is a real number). The ratio uj/k = c
characterizes propagation of the plane of constant phase kx — cut = const along the
x-axis and is called the phase speed. Substitution of the indicated solution in the
equation of motion yields the dependence w = w(fc) which is called the dispersion
relation and allows one to find the phase speed c = c(k). The fact that c depends
upon A; (or w) is referred to as dispersion of the waves. The quantity dui/dk is called
the group speed. If there is no dispersion (i.e., if c does not depend upon k), the
solution has the form of a travelling wave ip = f(x ± ct). If the problem is reduced to
the wave equation
av av =
2
dt dx2 '
then the general solution is <p(x, t) = fl(x-ct) + f2(x + ct) where / i , f2 are arbitrary
functions. When the linear equations for displacement of the theory of elasticity
(i.e., the equations of Lame) are used, the functions sought are Wi(x, t), i = 1,2,3.
If only the displacement component along the direction of wave propagation varies:
wx = w(x, t), wy = wz = 0, then the wave is referred to as purely longitudinal; and if
wx = 0, wViZ = wy,z(x, t), then the wave is referred to as purely transverse.
0 26.82 Consider a disturbance depending only on x and t in an unbounded
isotropic elastic medium (plane waves). Find the components of the displacements
Wi{x, t) (i = 1, 2,3) and of the velocity of propagation of the waves. Show that there
is no variation of volume in a transverse wave (divtu = 0), and curltu = 0 for a
longitudinal wave.
0 26.83 An arbitrary displacement vector can be decomposed into the sum of
two vectors w = Wi + w2 such that curium = 0 and divw2 = 0. That allows
the consideration of any wave as a composition of compression waves and distortion
waves. Write the equations describing propagation of these waves (the equations for
wi and w2), and find their velocities assuming that t/>i and w2 vanish out of a certain
finite spatial domain.
Q 26.84 Find the speed of longitudinal waves propagating in a thin elastic rod, the
lateral surface of which is free from stresses. Compare it with the speed of longitudinal
waves in the unbounded medium (see Problem 26.82).
© 26.85 Find the longitudinal eigen-oscillations of an elastic rod of length I rigidly

fixed on one end and free on the other.
0 26.86 Find the equation for bending waves in a circular rod of a cross-sectional
area S. Show that these waves have dispersion. Find the phase and group speed.
© 26.87 A plane travelling wave with front perpendicular to a direction £ is in
cident on a rigid unmoving boundary x = 0 of an elastic half-space x > 0 and
reflects. The displacement vector in the incident wave is given: w0 = / 0 (f + c\t) £°
where £ = x cos a0 + y sin a 0 ; £°{cos a0, sin a 0 } is the unit vector of the direction
£i c i = v (^ + 2/x)/p is the speed of the longitudinal wave, a) Find the amplitudes
and directions of propagation of the reflected waves, b) In what case does only one
longitudinal wave reflect? c) What restrictions should be imposed in the setting of
the problem if the incident wave is transverse?
0 26.88 Write the equations satisfied by elastic plane waves (w = w(x, t)) propa
gating in a transversely isotropic medium in the i-direction perpendicular to the axis
of symmetry of the medium, the z-axis.
© 26.89 The pressure in a spherical cavity of radius R inside an unbounded elastic
medium changes at an initial instant from zero to po = const Assuming that the
displacements and velocities at the initial instant equal zero, determine the radial
displacements of the medium.
© 26.90 Study the waves running over the free boundary of an elastic half-space
(waves of Rayleigh). The elastic medium fills the domain y < 0, the boundary is free
Figure 26.28
from stresses, the wave runs parallel to the x-axis, and the displacements are W3 = 0,
w a = wa(x, y, t), a = 1,2. Find the speed of the wave which decays as y —* —00.
© 26.91 Study waves running along an infinite elastic layer of width 2h (along the
z-axis). There are no displacements along the z-axis: w3 = 0, and the components
UJX and w-i do not depend upon z. The upper and lower boundaries are free from
stresses. Consider a) long waves whose length is much more than h (A » h) and
b) short waves (A <C h). Find their speeds.
27. Nonlinear Elasticity 205
*>
Figure 26.29
27 Nonlinear Elasticity
— PROBLEMS —
0 27.1 A nonlinear elastic medium is specified by its free energy J-(iij, T). Write
the expressions of the components of the Piola-Kirchhoff stress tensor in terms of the
strains and temperature.
0 27.2 Show that the free energy of an incompressible nonlinear elastic isotropic
medium depends on only two scalar invariants of the strain tensor.
0 27.3 Nonlinear effects for strains that are not large can be taken into account
by only the first (main) nonlinear terms following the linear ones in the expression of
stresses in terms of strains written in the form of expansion into series over strains.
Write the general form of the free energy function T for such a medium if it is
isotropic, and the temperature is constant. (The function poF{iij) at T = const is
called the elastic potential.) Let us call this model a medium of Mumaghan. How
many coefficients (elastic constants) determine this medium?
0 27.4 An elastic model is specified by Murnaghan's potential po-^foj) (see Prob

lem 27.3). Write the expressions for the components of the Piola-Kirchhoff stress
tensor in terms of the components of the distortion tensor in the case for which the
displacement vector depends only upon one spatial coordinate: Wj(z, £). Consider
only the first nonlinear terms.
0 27.5 A nonlinear elastic medium is in equilibrium between two infinite planes
and is squeezed by pressure p uniformly distributed over the planes (see Figure 27.1).
Find the stresses 7ri; and strains in the medium. Use the model of the medium
of Mumaghan with five constants (see Problem 27.3), and consider only the first
nonlinear terms. Study the convexity of the graph 7rn(Viioi) as a function of the
elastic constants.
Figure 27.1
© 27.6 Write the equations describing propagation of one-dimensional nonlinear

waves in a medium of Murnaghan (see Problem 27.3). Show that purely longitudinal
waves can propagate in such a medium if there are no initial strains, and transverse
waves are accompanied by variation of the longitudinal component of the strain as
well (quasi-transverse waves).
0 27.7 Show that only purely transverse one-dimensional nonlinear waves can
propagate in an incompressible medium of Murnaghan.
© 27.8 A particular solution of the equations of one-dimensional unsteady motion
is referred to as a Riemann wave, if the dependence on x and t of all functions sought
for can be represented in the form of a dependence upon the combination of i and t
of the form 6 = 9(x, t) (see Problems 24.29 and 23.21). For the equations of nonlinear
longitudinal waves in a medium of Murnaghan (see Problem 27.6) find the solution of
the type of Riemann waves. Determine the characteristic speed and the condition of
wave breaking. Associate the condition of wave breaking with convexity of the graph
TTii(ViiUi) obtained in the problem of one-dimensional extension-compression of the
medium under consideration (Problem 27.5).
28 Couple Elasticity and Averaging in Media

with Microstructure
In the problems of this section, the models under consideration are models of media
with couple stresses the description of which leads to equations with derivatives of
higher order. In particular, let us relate to couple elasticity the models of media with
couple stresses connected with internal rotation (see Sections 11, 16). In addition, we
assume that the angular velocity of internal rotation ft coincides with the vorticity
vector UJ = \ curl t; (v is the velocity field of the medium)
n = w = - curl v
2
28. Couple Elasticity and Averaging in Media with Microstructure 207
It is assumed also that the internal energy per unit mass u (without the kinetic
energy of internal rotation £ rot = u — u) depends on the entropy s, components of
the strain tensor e, and derivatives of the internal orientation field, the dependence
on derivatives being reduced to dependence on the components of the tensor
3
K
= £ «(a)Vn (a)
where n( a ) are three orthonormal vectors such that
and tijk are the components of Levi-Civita tensor. It is assumed that the processes
are reversible, all the derivatives of the displacement vector w are small so that the
equalities V{Vj = ^ViiUj are valid. The relation £2 — u allows one to express the
determining tensor K in terms of the second spatial derivatives of the displacement
vector w:
n m
Kmtn = l y ' t W - V w )
(Problem 28.1) and reduce the theory to equations with derivatives of higher order.
To set up a model one should write the dependences of the stress tensor p and
the couple stress tensor Q on the other state parameters (constitutive equations).
The constitutive equations (equations of state) can be derived by use of the first and
second laws of thermodynamics if the internal energy u is given as a function of state
parameters (see Problems 15.9, and 15.7). But there are some difficulties on this way
which are described below.
Let the equation of internal energy in a medium with internal rotation have the
form (dtf/dt = 0)
>f-î^*'(£-i-»)- <28»i
= PT J+P°
dt
v,^ + v,(â) + P(ti- ^ a
Then, due to the equation of internal angular momentum A; (h is the mass density
of body couples)
die*
P^Tr = pV + V3Qij-t,3hp]k (28.2)
dt
the equation (28.1) takes the form
9
f = pT t+ p w^" + Q1VjQi+fiM<* ~ ^
where p(s) and p(a) are respectively the symmetric and antisymmetric parts of the
stress tensor p, e^ = \ (VjUj + VjUj). If fl = u»,
p
ft^pTft+êij+QiJViUi- (28 3)
'
The equations (28.3),(28.2) as well as the equations of motion p ^ = V,p,J + pF* are
invariant relative to transformations of the form
p'" = p « + eijkVkN, Q"j = Q y + 2?<J N
where the function N(xk, t) is arbitrary. Due to this fact it is impossible to determine
pxi and Q y uniquely by specifying the function u.
The same difficulty arises also in models of more general type with u(s, e^, V, VfcWj)
and
P^Y~ = VjiR^VkVi) + MijVjVi + pQ ■ dk/dt
where u and M,j are given, and p ,J and Rh,} are determined (R^J = (l/2)e* u Qj J '
in the particular case considered above). Here, there is an invariance relative to the
transformations (Problem 28.2)
jf* = p«> + ejklVkN'\, R""' = R*' + ^klN'\
leading to the transformations for Qxj = exk,Rk>j, where iV'( = N 6\ .

It is possible to obtain uniquely the equations of state, e.g., in the following way.
Consider a couple-type continuum model that is described by the equation of internal
energy of the form
Pjt{u(s,ey, VjA«) + Ay (Viity - A^)) =
with the same function u of the indicated arguments. Here, s, w, A, A are regarded
as independent characteristics. The given form of equation of internal energy con
tains now derivatives of lower order than in the preceding description, therefore all
ds ,-, dAu _ dAki dA'J . . . . . . .
the derivatives —, VjUj, —-r*-, v* ——, —— in this relationship can be treated as
at at at at
independent and forming a linear and homogeneous combination. This method, after
elimination of the quantities Aij = VjiUj and A'-' = — Vfc.Rijfc — M-*', yields
In particular, in the framework of couple elasticity,
du dkl
$) deij dt
Qii=pemn* . ,
the equation of internal angular momentum being satisfied automatically.

It is assumed in the problems of linear couple elasticity suggested below that there
are neither internal angular momentum (u = u) nor body couples, the processes
are isothermal (T = const), the free energy F — u — Ts is a quadratic function
of components of e and K, and the medium is isotropic and homogeneous. The
boundary conditions for problems formulated in terms of stresses require specifying
the quantities pn = ptjnjei and Qn = (J^nê* (et are the basis vectors) on the
boundary of a body with the normal n.
Let us note that, since there are derivatives of the displacement vector of higher
order, besides those of the second order, in the equations of motion, there are addi
tional constants with introducing a length scale I (usually small). This results in scale
effects, the presence of boundary layers, and phenomena of dispersion of transverse
waves, all of which are absent in classic theory. Taking into account couple stresses is
necessary in problems where there is a sufficiently small typical length L comparable
with /.
The equations with higher derivatives of displacements with respect to coordinates
arise also in averaging with various degrees of accuracy or in analysis of long-wave
asymptotics in approximate description of media with microstructure, e.g., compos
ites, or, in general, in analysis of dynamical discrete structures. Problems 28.8-28.14
deal with this approach to construction of more accurate models of continua.
— PROBLEMS —
0 28.1 Let # = | curl ii>. Show that, in linear approximation, | ^ | is the value
of rotation angle of the particle of the continuum counterclockwise around the unit
vector #/|4»|, and
3
* = £ < ) V i n ^ ) e m e ' e n « £ mns Vi$ s e ra e'e n =
Q=l
= \Vi(Vmwn -Vnwm)emelen .
0 28.2 Show that the expressions Vjplj and p^VjVi + Vj(i?t,JVfcVj) are invariant
relative to the transformations p' ij = pij + e^V/tiV*,, R'kij = Rki> + eJ'wN', where e«*
are the components of Levi-Civita tensor, Nlt are arbitrary functions of the coordi
nates. Determine the transformation rule for couple stresses Qxi = e>kmR'km^, prove
that the relationships VjQ1* = efjyP*7' are invariant relative to these transformations.
0 28.3 Let the free energy of a medium with couple stresses F = F(ey,«; m i n ,3p,)
be a scalar function quadratic and homogeneous with respect to components of the
tensors £ and K. Determine the general form of the function F. Following the method
stated in the introduction to this section, derive the formulae for the stress tensor p
and for the couple stress tensor Q. Derive the equations of motion of the medium
in terms of the displacements. Explain why some terms in Q do not contribute into
these equations. Determine the signs of the coefficients of the free energy required by
its positive definiteness with respect to e, n.
0 28.4 In the framework of linear couple stress elasticity, using the results of
solution of Problem 28.3, derive the equations for the coefficients of volume expansion
6 = Vjtu' and rotation vector # = | curl to (Problem 28.1), p = const.
© 28.5 Let x> (j = 1,2,3) be Cartesian coordinates. For the one-dimensional

case w = vP(xl,t)ej, using the equations of Problem 28.3, derive the equations for
longitudinal (w1 ^ 0, w2 = w3 = 0) and transverse (w2 ^ 0, w1 = w3 = 0) waves,
supposing that there are no body forces. Determine the dispersion relation w; = ujj(k)
for the waves of the form vP = Re(;4 J e^w>t-fcx '); cjj, k are real numbers, and A* are
complex. Analyze the superposition of two waves of the same amplitude A* with very
close wave numbers k!, and k". Show that the speed of the resulting wave with slowly
varying amplitude is ^f (the group speed). Calculate the group speed for longitudinal
and transverse waves, and compare with the phase speed w/fc.
© 28.6 Let an infinite slab occupying the domain {x1 € [— L, L}} experience con
stant loads pn, Qn on the boundary. Using the equations of Problem 28.3, determine
its static strain state for the two cases a) p n = 0, Qn = ±<?e3 and b) pn = ±pe 2 ,
Qn = —pLe.z respectively at x1 = ±L. The coordinates x* are Cartesian. It is sup
posed that the displacement field has the form w = wt(x1)ei with iu'(0) = 0. Analyze
the solution for / = Jujp, <C L.
© 28.7 In this Problem the so-called inverse problems are considered, i.e., the
solution is partly given and one should find the other part of the solution and the
problem which it satisfies to.
Solutions of problems of classical linear elasticity with quadratic dependence of wx
on Cartesian variables, obviously, satisfies also the equations of linear couple stress
elasticity (Problem 28.3), the stress tensor field p being the same for both cases. Cal
culate the rotation vector $ and the couple stress tensor Q and write the appropriate
boundary conditions for Q n , verifying conditions of equilibrium,
a) for extension of a beam with circular cross section of radius R, under the action of
its own weight:
2 2 2 2 3
^ = -g[{x -L ) atf + + z )), " = ^ P , - =^ P
where g is the acceleration of gravity, L is the length of the beam, E is Young's
modulus and a is Poisson's ratio;
b) for bending of a beam with circular cross section of radius R and with length L:
, Myx , M , , . , ,.. •> oMxz
where M is the magnitude of the bending moment, J = irR^/A is the moment of

inertia of the cross section of the beam relative to the axis of bending z\
c) for torsion of a rod with cross section of a radius R and with length L:
wl = —azy , w2 = azx , w3 = 0
where a = - ^ j , M is the torsional moment.
0 28.8 In the framework of classical linear elasticity, solve the static one-
dimensional problem with displacement field of the form w = w1(xl)ei, wÔ) = 0,
xl € [0,L] is a Cartesian coordinate, p n (0) = - p e i , pn(L) = pe^. There are no
body forces, the medium is isotropic, the Lame coefficients A(i'), ^(x1) are variable.
Determine the apparent modulus of elasticity pn/{eu)L where the angular brackets
mean averaging over the coordinate x1 on the segment [0, L]. Show that, if the func
tions A, fj, are periodic with a period /, and L = nl, n is an integer, then averaging
over [0, L] can be replaced by averaging over the segment [0,1]. Analyze the structure
of the solution for w1(x1) as e = l/L —> 0, prove that the term of order e is a periodic
function with the period /.
Q 28.9 Let propagation of a transverse wave in a classical linear elastic medium

be described by the equation
, . d2w d ( , . dw
where p and ft are periodic functions with a period /. Suppose that the solution has
the form
2
- =H +^ ( 0 ^ + £ i v 2 ( 0 ^ + o(£3)
(compare with Problem 28.8) where {w} is the average value over the period I, e = l/L
is a small parameter, TV!, N2 are periodic functions of f = x/l with period 1 such that
(Ni) = {N2) = 0, L is a length scale of variation of the wave amplitudes. Find the
functions (w)(x, t) and N\(£). To determine Ni, equate the term of order 1/e in the
original equation to zero; next, average the term of order 1. Show that (w) satisfies
the equations of classical elasticity for a homogeneous medium.
0 28.10 In the framework of classical elasticity, consider the transverse monochro

matic wave
ut kix)
w = Re(Ai e« - ),
w, and fci being real, Ai being a complex number, normally incident onto the discon
tinuity surface x = 0 of the piecewise constant functions p(x), n(x) in the equation
d2w d ( dw\
P
~dt? dx~\til)x')
=
Calculate the amplitudes of the transient and reflected waves, supposing that the
conditions across a contact discontinuity are satisfied, i.e., there are no jumps of - 5 -
dw
and fj, —— on the discontinuity surface. Compare with the solution of the problem of
ox
collision of two material points.
© 28.11 Let there be an infinite system of initially resting free material points
distributed along a straight line with masses m i , m 2 , . . . and a finite total mass M.
A material point of a mass mo runs into the system from the side of mi with veloc
ity t>o directed along the same straight line. As a result, a wave of elastic collisions
propagates over the system (see Problem 28.10). Let n be the number of the first
material point moving at an instant (the points with the masses mn+/t (k > 1) are still
at rest), and En be its kinetic energy. Let us consider the energy Ex = limn_oo En
which is transmitted over the chain as a result of passage of the collision wave. Using
the Cauchy-Bunyakovskii inequality, show that the best transmission of the energy
Ex, over the chain corresponds to the case for which the velocities of all the material
points after passage of the collision wave are equal. Such a chain will then move as a
rigid body without disintegration, and the excess of energy will be withdrawn to van
ishing masses. Obtain the appropriate mass distribution m„ (n = 1,2,...) from the
conditions of elastic collision, regarding m 0 and M as given. In the one-dimensional
setting, describe the collision process in the framework of continuum mechanics, rep
resenting it as motion of a plane discontinuity surface with variable surface density
a = p\l in a medium without stresses where pi is the initial density, and I is a constant
parameter with length dimension characterizing the distance between particles of the
discrete system. Determine the distribution of p\ for the best transmission of energy.
0 28.12 Within the setting of Problem 28.10, construct the exact solution of the
problem of passage through a laminated elastic medium of a wave with displacements
of the form
w = Re(F(x)eilu*-'kx))
where F(x) is a complex periodic function with period equal to the period / of the
piecewise constant functions p{x), p,(x) taking respectively the values p\, fi\ at x S
[—1,0) and p2, fi-2 at i £ [0, | ) . Analyze the dispersion relation f(k,uj) = 0 for the
case of constant sound speed c = Jn/p, determine the "zones of non-transparency"
where k becomes complex for a given real u. For small \k\ (long waves), write the
dispersion relation in the form w2 — c2k2(l—ak2) +0(k6) (find c and a), and compare
with the solution to Problem 28.5, paying attention to the sign — a opposite to the
sign of the corresponding coefficient in couple stress elasticity with positive energy.
0 28.13 For a one-dimensional discrete dynamic system consisting of identical

material points of a mass m connected by springs with stiffness 0, the equation
of motion of n-th point has the form mwn(t) = —0{2wn — wn+i — w n _i), n S Z.
Determine the dispersion relation ui(k) for waves of the form wn = Re(j4e' ( " t_ * on) )
where u>, k are real, A is a complex number, and a is a linear dimension bounding the
wavelength from below. Derive the equation of the long-wave approximation (\ka\ <^
1 dw 1 dw
1) taking into account the dispersion, and setting the derivatives - ——, — - -^— into
i at i ax
correspondence to the quantities u>, k in the relationship w2 « c ^ ^ l — ak2). Compare
with the equations of Problem 28.5.
0 28.14 Derive the equation of motion of an infinite one-dimensional chain of
material points with alternating masses m and M > m connected by springs with the
same stiffness P (see Problem 28.13). Let m 2n = m, m2„+i = M (n € Z), and w2n,
w2n+i be the corresponding displacements. Find the solution of the form
w2n = Re(Aei{u"-ka-2n)) , w 2n+1 = Re(ê i M -* a ( 2 n + 1 ) ) ) .
Show that there exist two branches of the dispersion curve w(/c) (o>, k are real)
corresponding to the "acoustical" and "optical" modes of the oscillations separated
by the band of forbidden frequencies with width Aw = \f2j3 f-Û - ^ J , determine
the form of these branches for small \ka\. Derive the partial differential equation of the
second order for more high-frequency, optical mode in the long-wave approximation
(as in Problem 28.13). Show that it is not a wave equation.
Chapter 7
Inelastic Solids
29 Plastic Flow Theory

Most solids behave as elastic bodies if the temperatures and loads are moderate, and
the loads do not vary too quickly, and are not applied for a very long time. Beyond
these conditions, inelastic response takes place. Plasticity, viscosity and (fading)
memory are three main types of inelastic properties. Their common peculiarity is the
fact that there is no formula expressing the current stress trough the current strain.
These properties will be considered below in the case of small strains, heat effects
being neglected. (The last assumption is reasonable if sufficiently intensive heat
conduction equalizes the temperature of the body and keeps it constant at the expense
of heat exchange with the surrounding medium.)
The word "loading" often used in formulations of the problems has two meanings:
1) applying external forces to a body, 2) variation of the stresses at a point of a body
(or in the whole body if the stresses are uniform, i.e., independent on the coordinates
of the point).
Plasticity is the property of a medium to experience residual (i.e., non-vanishing
after unloading) strains dependent on the manner of loading, but independent of its
rate.
Within the scope of the plastic flow theory, a maximally extended domain in the
stress space is distinguished within which the solid behaves as an elastic solid or, in
the limit case, as a rigid body. If this domain remains unchanged, the body is referred
to as perfectly plastic. If the domain varies, hardening is said to take place.
For simplicity, only one parameter x is considered here to describe hardening. The
domain of elastic (in the limit case, rigid) behavior in the stress space is specified by
the condition
/(py,x)<0.
At fixed x> 1) the elasticity domain is convex, 2) the zero stress lies inside it,
215
216 INELASTIC SOLIDS
/(O, x) < 0, and 3) the domain remains unchanged or expands as \ grows: df/dx > 0.
Hardening consisting of homothetic extension of the elasticity domain is referred to
as isotropic.
Stresses in a body described by the plastic flow theory always satisfy the condition
/(p,j, x) < 0. Plastic strains can vary provided the stress lies on the surface /(py, x) =
0 referred to as yield surface. The equality f(p,j,x) = 0 is referred to as yield criterion.
The best known yield criteria are the TVesca criterion
max {|pi - p 2 |, \p2 - p 3 |, |ps - Pi|} = 2r8 , (29.1)
and the Mises criterion
(Pi - V2? + (P2 - ? 3 ) 2 + (ps - Pi) 2 = 2a,2, (29.2)
where pi, p 2 , P3 are principal components of the stress tensor, TS and <7„ are constants
or functions of the hardening parameter x- I n t n e c a s e 0 I a perfectly plastic body,
stress cannot reach beyond the yield surface; admissible stresses satisfy the inequality
/(P.;) < 0.
Plastic strain is determined within the framework of plastic flow theory by the
normality flow rule which is also referred to as the associated law (i.e., associated
with the yield criterion):
^ P _ / * $ , A>0, if /(p",x) = 0
(29>3)
*'"\o, if / ( P " \ X ) < O
Here, A is considered as an indivisible symbol, the dot in X does not denote the time
differentiation. According to the normality rule, plastic flow is possible (efj ^ 0)
provided the stress stays on the yield surface. Plastic strain rate vanishes as soon as
the stress leaves the yield surface, in which case unloading is said to take place. Since
i^j = 0 during unloading, there remains residual, plastic strain at zero stresses.
Within the scope of elastic-plastic models of solids the total strain is composed of
the elastic and plastic components
ei} = ei, + <, . (29.4)

In this section, elastic strain is assumed to obey Hooke's law
whe
* = A+?/i
29. Plastic Flow Theory 217
and the superscript (d) denotes the deviatoric part of the tensor. In particular, elastic-
plastic model reduces to the rigid-plastic one at p. = oo, K = oo: elastic strain is
zero, and thus e^ = £?■.
Variation of the parameter x during isotropic hardening is normally described
by the equation x = Pij^j< a n d TS, as in the Tresca and Mises yield criteria are
nondecreasing functions of the hardening parameter x-
Equations (29.3), (29.4), (29.5) together with that for evolution of the hardening
parameter x are constitutive relations for the elastic-plastic body.
Qasistatic loadings are only considered in this section, i.e., the inertial terms in the
equations of motion are neglected. Components of vectors and tensors are considered
in the Cartesian coordinate system apart from some problems for which different
systems are suggested.
— PROBLEMS —
0 29.1 Show that the Tresca yield criterion (29.1) has the following physical mean
ing: the magnitude of the tangential stress on a surface element reaches the value
0 29.2 Show that a) values of the functions specifying the Mises and Tresca yield
criteria (29.2), (29.1) do not depend upon the spheric part of the stress tensor;
b) these values can be expressed in terms of the second and third invariants of the
stress deviator J^d) = p\f p'f, j j d ) = pjf p$pff;
c) the Mises criterion can be written as \p\jPji = k2, k = cr s /\/3, while both
invariants J% , J$ are involved in the Tresca criterion.
© 29.3 Long thin-walled tubular specimens are often used for testing mechanical
properties of materials. Consider such a specimen in equilibrium under the action of
forces applied to its ends. The resultant forces at the ends (let F be their magnitude)
extend the specimen along the axis, and the resultant moments (let M be their
magnitude) twist it around the axis (see Figure 29.1). The matrices of physical
components of the stress tensor in the cylindrical coordinate system r, ip, z can be
considered as equal for all points of the specimen sufficiently distant from its ends.
Find these stress components.
Note that the stresses do not depend upon the material of the specimen. This is what
allows to experimentally investigate a priory unknown material properties: the strain
can be measured and the stress is determined by the applied loads.
0 29.4 a) Consider the case when the only nonzero stress components in the
tubular specimen, under the loading described in Problem 29.3, are pzip — pvz (in the
8 << a « I
Figure 29.1
cylindrical coordinate system). What stretching force and torsional moment should
be applied to the specimen to result in this stressed state? (The value of pzv at which
plastic flow starts under this loading is referred to as yield stress in pure shear.)
b) What stretching force and torsional moment should be chosen to result in the
stressed state with the only nonzero component pzz ? (The value of pzz at which
plastic flow starts under this loading is referred to as yield stress in pure tension.)
c) Show that TS in the Tresca yield criterion is the yield stress in pure shear, and as
in the Mises yield criterion is the yield stress in pure tension.
0 29.5 It was revealed in experiments with a steel specimen that the yield stress
in pure shear (see Problem 29.4) equals 23 kN/cm2. Find the yield stress in pure
tension, assuming that steel obeys a) the Tresca yield criterion, b) the Mises yield
criterion.
0 29.6 Plastic flow takes place (e?- ^ 0) in an isotropic elastic-plastic body with
hardening. Express the multiplier A in the normality rule in terms of the stress rate
Pa-
Note that this is impossible in the case of a perfectly plastic body (why ?).
© 29.7 Consider a stressed state satisfying the yield criterion f(pij,x) — 0 and
varying at the rate p^. Show that
a) if -^-ptj < 0, the plastic strain does not vary, e? = 0, and the inequality 41. =
-§^-Pij + | ^ X < 0 is valid (such a process is called unloading);
b) if -§^-pij = 0, the plastic strain does not vary, e? = 0, and the equality df/dt = 0
is valid (such a process is called neutral loading);
c) if -M- pij > 0, the plastic strains vary, e^ ^ 0, and the equality df/dt = 0 is valid
(such a process is called active loading).
What can be said about analogous processes in a perfectly plastic body ?
0 29.8 Values of the function / specifying a yield criterion do not depend upon
the spheric part of the stress tensor. Show that the plastic strain rate satisfy the
incompressibility condition evkk = 0.
© 29.9 The function /(py) specifying the yield criterion for an isotropic body can
be written as f{p{]) = F(pi(pij),p2{pij),p3(Pij))> where pi(py), P2(j>ij), P3(Pij) are the
principal components of the stress tensor.
a) Show that the principal axes of the strain rate tensor coincide with the principal
axes of the stress tensor, and its principal components are ef = A dF/dpi (A is the
multiplier in the normality rule).
b) Show that, if the principal axes of the strain and stress tensors coincide at the
initial moment, and the principal axes of the strain tensor remain unchanged, then
the principal axes of the stress tensor also remain unchanged.
c) What arc the principal components of the plastic strain rate tensor in a body
with the Tresca yield criterion if the stress satisfies this criterion and the principal
components of the stress are pi, P2, Ps and p\ > P2 > P3 ?
0 29.10 For an elastic-plastic body with the Mises yield criterion, differentiate
the equality e y = e„ + e y with respect to time, use Hooke's law and the normality
rule to obtain the Prandtl-Reuss equations
1
-(<0 , ; (<fl *
© 29.11 The material of a tubular specimen is described by the relations of the

plastic flow theory with the Mises yield criterion and isotropic hardening. Loading
starts at the undeformed state and is carried out so that the only nonzero stress
components in the cylindrical coordinate system are pztp = p#z = o (see Prob
lems 29.3, 29.4).
a) Show that the only nonzero strain components are eZKp — evz. If no unloadings occur
during the process under consideration, the strain e can be expressed as a function
of the stress a (and vice versa). Find the relation between a and e, regarding the
constants and functions specifying the material properties as known.
b) Plot the diagram a ~ s in the case of the yield criterion \p\j Pj^ = k2 + a2x,
where k2 = const, a2 = const.
Note that the diagram does not depend upon the rate of the loading.
0 29.12 A copper tubular specimen is loaded only by a torsional moment (see

Problem 29.4); the loading starts at the undeformed state. A somewhat idealized
diagram a versus E (where a = pzip, e = ezip) obtained from the experiment is shown in
» a kN/cm
_^ e
0.005 0.01
Figure 29.2
the Figure 29.2. Assuming that behavior of the specimen can be described by relations
of the plastic flow theory with the Mises yield criterion and isotropic hardening, find
the elastic shear modulus, the initial yield stress in pure shear and the hardening law
(dependence of the yield limit upon the hardening parameter).
0 29.13 Simple loading is a process during which the components of stress de-
viator vary proportionally to a parameter r growing monotonically: p\f = r s ° ,
s°3 = const. Show that the following statements are valid for any simple loading of an
elastic-plastic material with the Mises yield criterion and isotropic hardening if the
loading starts at the undeformed state:
a) the spheric parts of the stress and strain tensors obey Hooke's law e^ = 577 Pi*;
b) the "direction tensors" of the stress and strain deviators are equal: e\f/ei = p\f/pi
(pi and £1 are the intensities of the stress and strain tensors: pi = (p\fp^)1/2, £\ =
(e!Mf),/a);
c) there is one-to-one correspondence of the intensities: £1 = F(p\), and this relation
is the same for all simple loadings. Find the function F assuming the constants and
functions specifying the material properties as known.
0 29.14 The relations indicated in Problem 29.13 establish a one-to-one corre

spondence between stresses and strains. These relations can be considered as consti
tutive laws of a certain model of a continuum (this model was introduced indepen
dently of the plastic flow theory and is called deformation plasticity theory). For a
simple loading, both models imply the same results as it follows from the statement
formulated in Problem 29.13.
Would these theories result in the same strains if a) after a simple loading, the stress
tensor varies at constant intensity? b) after a simple loading, unloading to the zero
stress takes place?
0 29.15 A loading of the type described in Problem 29.3 results in the stressed
state with the only nonzero stress components pzv = p^z = a (in the cylindrical
coordinate system). The stress evolution is shown in Figure 29.3. Assuming that the
material of the specimen obeys the relations of the plastic flow theory with isotropic
Figure 29.3
hardening, plot the graph of strain evolution.

0 29.16 An elastic-plastic body with the Mises yield criterion deforms under the
plane strain conditions (xi, X2)-p\8me (xi, £2 are Cartesian coordinates). The initial
state is unstressed. Show that 1) P13 = P23 = 0 during the process, 2) if the body is
incompressible, then p 33 = (pn+P22)/2 and the Mises yield criterion can be re-written
as (pn - P22)2 + 4pi2 = 4A;2 (k = <x8/\/3)-
© 29.17 An infinitely long pipe tube of the inner radius a and outer radius b is
subjected to the internal pressure p monotonically growing from zero. The material of
the tube is elastic perfectly plastic with the Mises yield criterion, and incompressible
(K = 00 in Hooke's law).
a) Find the stress and displacement distributions in the pipe when its deformation is
purely elastic.
b) Find the value of the internal pressure at which plastic deformation starts. Find
the place in the pipe where the plastic deformation starts.
c) Find the stress distribution in the plastic zone (i.e., in that part of the pipe where
the deformation is plastic).
d) All components of the stress tensor are assumed to be continuous on the interface
between the plastic and elastic zones. Since the problem is axially symmetric, the
interface is a cylinder. Assuming its radius known, find the stress distribution in the
elastic zone.
e) Using the stress continuity condition on the cylindrical interface between elastic
and plastic zones, derive the equation for its radius.
f) Show that, for any pressure within a certain interval Po<p<p, the above equation
for the radius has a unique solution (the radius between the internal and external radii
of the pipe) and it has no solution for the other values of pressure. Find p 0 and p..
Note that at p > p. the problem about a pipe under internal pressure does not have
any solution at all. This situation is usual for perfectly plastic solids: the stresses in
such a body are limited and cannot equilibrate sufficiently "large" loads. There is
a surface in the space of loads that separates "large" loads from admissible "small"
ones; points of this surface are referred to as limit loads. The internal pressure p. is
the limit load for the pipe in the case under consideration.
® 29.18 An elastic perfectly plastic sheet infinite in the 13-direction is bent under
the plane strain conditions in the plane Xi,i2- Bending is caused by the action of
forces distributed on the sides of the sheet with the density independent of x3 (see
Figure 29.4). The resultant force per unit length of the side is zero, and the resultant
~x
a« b
Figure 29.4
moment per unit length has only one nonzero component M3 = M (the moment
about the £3-axis). Consider domain sufficiently distant from the sides of the sheet.
Since the sheet is thin, a <IC b, the stress components pu, P22 in this domain are
negligible compared to the values of pn, P33), and p ^ = P23 = 0 due to the plane
strain condition.
a) Find the stress and displacement distributions in the cross-section of the sheet
when its deformation is purely elastic. What is the value of the bending moment M
at which plastic deformation starts ? Find the place of the cross-section were the
plastic deformation starts.
b) Find the stress distribution in the plastic zone.
c) All components of the stress tensor are assumed to be continuous on the interface
between the elastic and plastic zones. Find boundaries of the elastic and plastic
zones and stress distribution in the elastic zone. Plot the graph of of pu versus the
coordinate 12-
d) Find the limit load, i.e., the maximum value of the moment M that can be equi
librated by admissible stresses.
e) Assuming the velocity continuous on the interface between the elastic and plastic
30. Rate-Dependent Effects in Solids 223
zones, find velocity distribution in the plastic zone adjacent to the upper surface of
the sheet.
30 Rate-Dependent Effects in Solids

Viscosity is a property of a medium to resist deformation in different ways dependent
upon the strain rate. Viscosity in solids results in growth of stress with the strain rate.
Another exhibition of viscosity is nonzero strain rate at constant stresses (creep).
The simplest model of a solid that takes into account viscosity is specified by the
relations of visco-elastic Voigt solid. The constitutive relations of this model are
pjf = 2/xe\f + 2r,i[f , Pkk = 3Kekk .
The shear stresses py are composed of the elastic part 2/i£y and viscous part 2r)i\j
(as in a viscous fluid), where r] is the viscosity coefficient; the spheric part of the
stress tensor is purely elastic.
Constitutive relations of some continuum models relate the response of the body
not only to the current situation but also to the past. In this case the body is referred
to as body with memory. Plastic bodies possess this property; for example, the yield
surface once modified by the hardening may remain unchanged for infinitely long time
provided subsequent loadings are not active, i.e., do not result in plastic deformation.
Some continuum models possess memory of another type, fading memory . Within
the framework of such a model, the past events effect the response of the body to the
current situation, however, the longer is the time interval between a past event and
the current moment the weaker is the influence of that event.
The simplest model of a medium with fading memory is specified by the relations
of visco-elastic Maxwell solid. The constitutive relations of this model are
■ (d) 1 ,-(d) , 1 (<JK 1

e
« = 2 ^ + ; p ^ ' £
" = 3 ^ p
" -
The shear strain £y are composed of elastic strain Py /2/x and viscous strain, the
rate of the latter being p y / 2 / i T (as in a viscous fluid with the viscosity coefficient
rj = fir); the spheric part of the strain tensor is purely elastic.
Model of visco-plastic body is considered in Problem 30.6.
For all problems of this section, the loading is assumed to be quasistatic, i.e.,
inertial terms in the equations of motion are neglected. Components of vectors and
tensors are considered the in Cartesian coordinate system apart from some problems
for which different systems are suggested.
— PROBLEMS —
© 30.1 Under the loading described in Problem 29.3, the only nonzero stress
components in the tubular specimen are pZ{(, = pvz (in the cylindrical coordinate
system). The torsion strain is maintained constant. Assuming that the material of
the specimen is described by the Maxwell visco-elastic solid model, show that the
stress relaxes exponentially with the exponent t/r. (Because of that the constant r
is referred to as relaxation time in pure shear.)
Would the stresses relax if the material is described a) by the Voigt visco-elastic solid
model? b) by the model of plastic flow theory 1) with perfect plasticity, 2) with
hardening (see Section 29)?
© 30.2 Under the loading described in Problem 29.3, the only nonzero stress com
ponent in the tubular specimen is pzz (in the cylindrical coordinate system). The
elongation of the specimen is maintained constant.
a) Assuming that the material of the specimen is described by the Maxwell visco-
elastic solid model, show that the stress relaxes exponentially. (Note that the relax
ation time in tension differs from that in shear.)
b) Find the function describing the evolution of the cross-sectional area of the speci
men.
0 30.3 Under the loading described in Problem 29.3, the only nonzero stress
components in the tubular specimen are pztp = pvz (in the cylindrical coordinate
system). The load is maintained constant. Find the evolution of the strain ezip if
the material of the specimen is described by the visco-elastic model a) of Voigt b) of
Maxwell.
© 30.4 The material of a tubular specimen is described by the Maxwell model of
visco-elastic solid. Loading starts at the moment t = 0 and is carried out so that the
only nonzero stress components in the cylindrical coordinate system are pzip = pvz = a
(see Problem 29.3). Let a vary at a constant rate: a = At, A = const > 0.
Show that the only nonzero components of the strain tensor are ezv = evz = e.
Find evolution of e. Plot the diagrams a ~ e for different values of A (these diagrams
are described in the parametric form by functions a(t), e(t)). What is the fundamental
difference between this case and the case of an elastic-plastic specimen ?
0 30.5 A tubular specimen is deformed at the initial moment so that ezip = evz =
£o while the other components of the strain tensor in the cylindrical coordinate system
are zero. The specimen is only loaded by torsional moment (see Problem 29.3) which
begins to decrease rapidly until zero value is reached. No forces are then applied to
the specimen.
30. Rate-Dependent Effects in Solids 225
Plot the graphs of evolution of the strain ez{p if the material of the specimen is
described a) by the model of plastic flow, b) by the Maxwell visco-elastic solid model,
c) by the Voigt visco-elastic solid model.
© 30.6 Show that the constitutive relations of rigid perfectly plastic body (see
Section 29) with the Mises yield criterion are equivalent to the following relations:
sflk- A>0 if ?0
vf = any satisfying the condition
P{±P(d)mn < 2fc2 = 0
(eij are components of the strain rate tensor). On the basis of this model, construct
constitutive relations for incompressible rigid visco-plastic medium which 1) has the
same domain of rigid behavior, 2) takes into account additional viscous stress.
© 30.7 Consider a rectilinear plane flow along £i-axis in an infinite channel be
tween two parallel walls (see Figure 30.1). The flow rate per unit length of 13-axis is
/////////H
fg
-m~X
//////;;/_n 7777.
"777777777777
Figure 30.1
Q. Behavior of the medium is described by rigid visco-plastic constitutive relations

for (see Problem 30.6). The velocity vanishes on the walls of the channel.
a) Show that the only nonzero component of velocity depends only on X2. Express
the stress deviator in the zone where strain rate tensor does not vanish in terms of
the velocity field. Show that the pressure p = — \pkk in this zone does not depend
upon £2 and linearly depends upon xi: dpfdx\ = — A = const.
b) Show that the plane of symmetry %i = 0 lies inside a layer moving as a rigid body,
i.e., at the zero strain rate.
c) Show that, on the boundaries of the undeforming layer, the equalities pi2(xi, h) =
—k, Pi2(x\, —h) = k are valid, where k is the constant involved in the Mises yield
criterion (see Problem 29.2). Also show that
d) Consider the average density of the force acting on a cross-section of the unde-
forming layer from the left. Show that its i r component linearly depends upon Xi,
i.e.,
d 1 /
_Pll Xl,l2 dl2
d~2hJ
dx\ ( ) = ~B = c o n s t -
Express the constant B in terms of the thickness of the undeforming layer.

Within the scope the model under consideration, stresses in the undeforming layer
should satisfy the condition p\fp\f < 2k2 and also equilibrate the forces applied to its
boundaries: the pressure linearly depending upon x\ (see item (a)) and the tangential
force uniformly distributed with the density k (see item (c)). It is known that these
conditions can be satisfied together provided the pressure drop coefficient A in the
deforming layers and the coefficient B of decrease of the average force density in the
undeforming layer are equal. Thus, the pressure drop coefficient A is known as soon
as the constant B is found (although A is expressed in terms of the unknown thickness
2/i of the undeforming layer).
e) Find the velocity distribution in a cross-section of the channel and relation between
the thickness of the undeforming layer and the given flow rate.
Note that, for a rigid perfectly plastic body, in contrast to a viscous fluid, a finite pres
sure drop per unit length (the constant A) is required to maintain even infinitesimal
strain rate. This is usual situation for a material with the domain of rigid behavior
in the stress space: the body does not deform under the action of sufficiently "small"
loads.
Chapter 8
Basic Notions of Relativistic

Kinematics and Dynamics.
General Properties of
Electromagnetic Field
31 Lorentz Transformations. Minkowski Space

Electrodynamics is closely connected with Special Relativity theory. Special Relativity
theory is based on two experimentally justified postulates. These are a postulate that
all inertial reference systems are equivalent and a postulate that the velocity of light is
the same in all reference systems. For these postulates to be true the transition from
one reference system to another should be described not only by transformation of
spatial coordinates, but by that of the time coordinate as well (see Problems 31.1 -
31.8). Therewith, events simultaneous in one reference system are not simultaneous
in the other (see Problems 31.8 - 31.10), and linear dimensions of moving bodies differ
from dimensions of the same bodies at rest (see Problem 31.9).
Special relativity theory, in particular, offers the formulae for recalculation of
physical quantities for a change of reference system. In usual mechanics, the effects of
relativity theory are known to manifest themselves in terms of order i^/c 2 . However,
there are terms of order v/c in the formulae for transformation of an electromagnetic
field.
According to relativity theory, space and time form a four-dimensional Rieman-
nian space with the metric ds1 = g^dx* dx*, i,j = 1,2,3,4. Minkowski space, which
is the basis of special relativity theory, is characterized by the fact that there ex
ist inertial coordinate systems (Lorentz coordinate systems) in which everywhere
5n = 922 = 933 = - 1 , 944 = c2, 9pg = 0 if p ^ Q, x1, i 2 , i 3 being the spatial
227
228 RELATIVITY THEORY
coordinates, and x4 = t being the time (let us note for comparison that, according to
the pre-relativistic perceptions, there exist two independent metrics — one of them,
for time, the other, for space). In the problems presented below, Lorentz coordinate
systems are used everywhere unless otherwise stipulated.
Transformations of coordinates for transition from one Lorentz coordinate system
to another are called Lorentz transformations. The physical reasons for determination
of the form of these transformations, their properties as well as the basic notions of
the relativistic dynamics and electrodynamics are discussed in Section 31.
In the problems concerning relativity theory, as usual in that science, Latin indices
run the values 1, 2, 3, 4, and Greek ones, 1, 2, 3.
— PROBLEMS —
0 31.1 Let a linear homogeneous transformation of the variables x, t into x', t'
(dx and dx' denote line elements, and dt and dt' denote those of time) have the
following properties: a) a signal propagating along the i-axis with light velocity ±c
(c = const) in the system x, t holds its speed in the system x', t' (i.e., the straight lines
x = ±ct + const turn into the straight lines x' = ±ct' + const after transformation);
b) the matrix of the transformation depends on the speed v of one system relative
to the other and is identical for any pair of coordinate systems for a given v. Show
that this transformation (the particular Lorentz transformation) is x7 = Oni — a\2ct,
ct' = -ai2x + a22ct where an = <*22 = l / \ / l - v2/c2, al2 = («/c)y 1 — v2/c2.
0 31.2 Show that the particular Lorentz transformation, which is the transfor
mation denned in Problem 31.1 supplemented by the equalities j/ = y, z' = z,
belongs to the class of transformations (general Lorentz transformations) for which
c2(dt)2 - (dx)2 - (dy)2 - (dz)2 = ^(dt'f - (dx1)2 - (dy1)2 - (dz'f.
© 31.3 Determine the transformation of two variables x, t into x', t' possessing
the property (?(dt)2 — (dx)2 = c?(dt')2 — (dx')2, using the analogy with orthogonal
transformations that arises after introducing the variables £i = x, £2 = id, £i = x\
& = ict',
© 31.4 Verify that particular Lorentz transformations form a group. Express, in

terms of the relative velocity of the coordinate systems, the parameter of the group
which is additive for succesive Lorentz transformations. Find the relative velocity
corresponding to the Lorentz transformation which is the superposition of two Lorentz
transformations with relative speeds vi and v2 (the it relativistic law for addition of
velocities).
31. Lorentz Transformations. Minkowski Space 229
© 31.5 Show that, for two four-vectors ul and vl turning into u", v'1 after general
Lorentz transformations, the equality flyuV = <7yu'VJ is valid; here, gu = gn —
g33 — — 1, 544 = c2, and gpq = 0 if p ^ q. Verify that purely spatial vectors a{aa ^
0 , a4 = 0) and the time vector b (ba = 0 , 64 ^ 0) are orthogonal (i.e., gijcftp = 0).
0 31.6 Show that a general Lorentz transformation can be represented as the

product of a particular Lorentz transformation, rotations of the three-dimensional
space and, optionally, a reflection.
0 31.7 Verify that, if one defines the tensor with components <?y in a coordinate
system x1 = x, x2 = y, x3 = z, x4 = t by the equalities gu = g22 = 933 = —1>
944 = c2, g^ = 0 at p ^ q, then, in all other coordinate systems obtained from the
initial one by Lorentz transformations (in Lorentz coordinate systems), that tensor
(the metric tensor) has the same components enumerated above, i.e., the components
of the metric tensor of Minkowski space do not change after Lorentz transformations.
© 31.8 In Lorentz coordinates in the Minkowski space, a) find and depict the
transformation of basis vectors on the plane x, t after a particular Lorentz transfor
mation (see Problem 31.1); b) show that, for any vector {Ax, At} on the plane x, t,
if A x / A t / ±c, it is possible to choose such a coordinate system x', t' that either
Ax' = 0, At' = AT ^ 0 (the time-like segment, A T being called the interval of
proper time), or At' = 0, Ax' = Al ^ 0 (the space-like segment, A/ being called the
proper length); c) prove the invariance of the magnitude of a four-volume determined
as / dxdydzdt for Lorentz transformations that preserve orientation (without re-
v
w
flections). Integration is supposed to be carried out over a domain consisting of the
same points (their coordinates being changed on the transformations); d) show that
the projections of an arbitrary element of a hypersurface on coordinate hyperplanes
are the covariant components of a vector.
0 31.9 The length of a rod equals / in the proper coordinate system (connected
with the rod). In a coordinate system of an observer for an arbitrary instant, find
the distance between the beginning and the end of the rod, if it moves with constant
velocity v parallel to its axis. Find the ratio of this distance to / (the coefficient of
Lorentz contraction).
© 31.10 In a coordinate system moving with a constant speed v relative to an

observer, let the time between two events taking place at the same point equal r (T
is the interval of proper time). Find the time between the events in the coordinate
system of the observer (from the viewpoint of the observer, the events take place at
different points).
© 31.11 How much time elapses in a rocket which, from the viewpoint of an
"unmoving" observer, flies during the time T/2 with constant velocity v and then
changes the sign of the velocity and comes back with the constant velocity — v? What
is the reason for the disparity of the "unmoving" observer and the observer in the
rocket? Determine the proper time interval in the rocket if its speed varies: v = v(t).
Show that straight lines on the plane i , t, from the viewpoint of the Minkowski metric,
are geodesies, with the intervals of proper time being maximal. In solving, use the
particular Lorentz transformation (see Problem 31.1).
32 Concepts of Relativistic Kinematics and Dy

namics
As we introduce a four-dimensional metric (instead of two independent metrics of
space and time), it is necessary to formulate dynamic and kinematic relations of
mechanics in invariant four-dimensional form in terms of four-dimensional vectors
and tensors.
— PROBLEMS —
0 32.1 a) Calculate the components of the four-dimensional velocity u' = dx'/ds

and four-dimensional momentum p l = mdxl/dT = mcu1 in a Lorentz coordinate
system (dr = ds/c is the differential of proper time (see Problem 31.10) and ds2 =
c2dt2 — dx2 — dx\ — dx\) for a particle with mass (at rest) m moving with a three-
dimensional velocity v relative to the coordinate system xa.
b) Verify that the four-dimensional velocity is a unit vector in the Minkowski space.
© 32.2 Show that, for small v/c, the first three components of the four-
dimensional momentum are close to the components of the three-dimensional mo
mentum, and the fourth component is close to the total energy of the particle divided
by c2. The total energy is the sum of the kinetic energy and the energy at rest mc2.
Find the next terms of the expansion of these quantities with respect to v2/c2.
© 32.3 Let a particle of mass (at rest) m0, moving with velocity v0, get a given
increase in four-dimensional momentum. Find its new mass (at rest) and velocity.
Does the solution of the problem have always a physical sense?
© 32.4 There is an atomic bomb in an indestroyable box. The walls of the box
do not transmit any types of energy. Is it possible to determine whether the bomb
has exploded, by analyzing the inertial properties of the box with its contents?
33. Maxwell Equations 231
0 32.5 Write the equations of motion for a particle, regarding the derivative of the
four-dimensional momentum with respect to the proper time — the four-dimensional
force — as given. Interprete the components of the four-dimensional force in the
non-relativistic case as the three-dimensional force and the energy input divided by
c2. Find the derivative of the mass at rest with respect to the proper time, and relate
it to the energy input in the proper coordinate system.
© 32.6 Study, in the plane x, t, the motion of a particle, the acceleration of which
in the proper coordinate system is constant.
© 32.7 Let a matrix S**, i, k = 1 , . . . , 4, be composed for a medium so that 5°",
a,f3 = 1,2,3, is the three-dimensional tensor equal to the difference of the flux of
momentum and the stress tensor, Sa4 is the momentum density, S40 and S 44 are the
energy flux density and the energy density devided by c2. Show that Stk is a tensor
in Minkowski space (the energy-momentum tensor).
0 32.8 Write the expression for the energy-momentum tensor of a moving ideal
non-heat-conductive gas, using the expression for the energy-momentum tensor com
ponents in the proper coordinate system. The energy density of the gas in proper
coordinate system equals p(U + c2), U is the internal energy, p is the density of mass
at rest, (i.e., the number of the particles per unit volume multiplied by the mass at
rest of each particle).
0 32.9 For the case of no external actions, write the equations of momentum
and energy of continuum media as the divergence of the energy-momentum tensor
set equal to zero. To obtain a closed system add to these equations the equation of
conservation of mass at rest (if changes in mass at rest, which take place on nuclear
and chemical reactions, can be neglected). As a particular case, consider a dust
(l/ = 0,p = 0).
33 Maxwell Equations
An electromagnetic field is characterized by the fields E (electric field) and B (mag
netic induction). It manifests itself as a force (Lorentz force) acting on a particle
with a charge e (e is a scalar quantity independent of the motion of the particle) given
by
F = e (E + - x B\
where v is the speed of the particle, and c is the velocity of light. Here and below,
the Gaussian system of units is used. As one can see from that equality, E is a vector
and B is an axial vector. The relativistic revision of the force of Lorentz is considered
in Problem 33.3.
An electromagnetic field satisfies the system of Maxwell equations, the integral

form of which is
- -cjtjEndi: + JB-dl = ^ h , JEnd?: = 4irev , (33.1)
T L S
-— /B„dE + JE-dl = Q , JBnd£. = Q. (33.2)

E L S
Here, £ is an arbitrary, bounded, fixed, non-closed surface with the direction of its
normal vector chosen. E is bounded by a closed line L, dl is an element of that
line directed so that the circuit along L in the positive direction forms a "right-hand
screw" with the normal n to S. S is an arbitrary closed surface with external normal
n; En = E ■ n, Bn = B ■ n, 1^ is the electric current flowing through £ in the positive
direction of n (i.e., the sum of charges crossing E per unit time), and ev is the total
electric charge located in the volume V inside the surface S. In the case for which
the electric current density j and the charge density pe are determined, the equalities
Js = / j „ d E , e v = /p e dV are valid. Usually in equations (33.2), 7E, ev, E, B are
£ V
implied to be quantities averaged in a certain manner. However, if the discrete nature
of the charges is taken into account for the quantities 1% and ev as functions of time
and the surfaces E, S, then these equations describe the true electromagnetic field
taking into account "microscopic" fluctuations.
The integral form of the Maxwell equations allows one to obtain the relationships
on discontinuity surfaces (see Problem 33.1) as well as the differential form of Maxwell
equations in the case of differentiable vector fields E and B
, „ 1 dE 4TT . , „ A ,
curlB — = —j , d i v £ = 47rpe, 33.3)
c at c
1 dB
cw\E + --z- = 0 , divB = 0 . (33.4)
c at
Note that the equation of charge conservation follows from equations (33.3) (and
also from equations (33.1)). See Problem 33.7.
Let us stress that j and pe or 1% and ev are the total current and charge including
both free carriers of charge (electrons and ions) and the charges and currents dis
tributed inside atoms and molecules, so that j = y 4- j c , pe = pi + p\. Here j c , p%
are "connected" current and charge associated with the interior of the particle, and
j f , pi are "free" current and charge. Usually it is supposed that connected current
and charge don't transform into free ones, therefore dpcJdt + divjc = 0, and a total
connected charge associated with the considered volume of continuum media equals
zero (the existence of nonzero pce is induced by displacement of charges inside the
particles of the media). It let us to write
pce = divP , j c= ^ + c-curlM. (33.5)
Vectors P and M, which are called polarization and magnetization vectors, are
distinct from zero only inside the media. The factor c is introduced in (33.5) for
convenience. Elqualities (33.5) make it possible to write Maxwell equations in the
form
IdD 4w .,
curlff 5 - = —j' divZ? = 47rp{, (33.6)
c at c
B
curl E + - -T- = ~0
, „ 1 ° div£> = 0 , (33.7)
c at
D = E + 4?rP B = H + 4TTM. (33.8)
To use equations( 33.6 -33.8) we must know H and D (or P and M, correspond
ingly) as a function of other variables. In many cases linear relations are supposed
D = sE and B = fj.H, where e and /x are constants. Further, we shall neglect the
effects of magnetization and polarization, i.e., take D = E,H = B,jc = 0,pl = 0.
— PROBLEMS —
0 33.1 Obtain, from the integral form of Maxwell's equations, a) the differential
equations (33.3), (33.4) ; b) the relationships on a surface of discontinuity of the
vectors of electromagnetic field, supposing that E and B are finite and continu
ous everywhere, except the surface of discontinuity; pe and j are allowed to have
singularities of the type of ^-functions corresponding to a surface charge and current.
© 33.2 Is it possible, in the presence of a thin non-conductive immovable plate

(the thickness of the plate can then be regarded as tending to zero), to produce a
stationary electric field whose tangential components on the different sides of the plate
differ? Compare with results of the solution of Problem 33.1. Explain the difference
with the results obtained there.
0 33.3 Write the equations of motion and energy for a charged particle in an
electromagnetic field E, B, expressing the four-dimensional force in terms of the
antisymmetric matrix Fkj defined by the formulae F12 = B3, F23 = Bx, F31 = B2,
F4a = Ea/c. (Different locations of indices for B and E are essential only in non-
Lorentz coordinate systems. This is connected with the fact that unlike E, magnetic
induction B is an axial vector.) Show, that the matrix Fki represents a tensor (the
tensor of electromagnetic field), using the form of the equation obtained.
0 33.4 Using the fact that Fki is a tensor (see Problem 33.3), show that the
transformation of an electromagnetic field (i.e., of the components Fki) by a particular
Lorentz transformation (see Problem 31.1) has the form
E\ = £,| , E'L = (E + vx B/cU/Jl - vyc? ,

B[ = B|,, B'L = (B - v x E/c)jJl - vyc* ,
where the subscripts || and J_ denote the projections of the corresponding vectors
onto the relative velocity v and onto the plane perpendicular to it.
0 33.5 Show that, if E _L B and |J3| / \B\, then it is possible to choose a

coordinate system moving relative to the initial one, so that in that system either
E' = 0 or B' = 0. Use this fact to analyze the motion of a charged particle in
a uniform electromagnetic field. Find the angular speed of rotation for a charged
particle in a uniform magnetic field if E = 0 (Larmor frequency).
© 33.6 a) Write the differential Maxwell equations in terms of the tensor Fki (see
Problem 33.3), and verify that the left sides of these equations are components of a
four-dimensional vector.
b) Verify, independently of item a), that the quantities in the right-hand side of the
first group of the Maxwell equations are components of a four-dimensional vector (the
density of "four-current"), considering the case of several charged fluids in motion.
c) Find the formulae for transformation of four-current on particular Lorentz trans
formations and write non-relativistic analogues of these formulae. Consider the case
for which pe = 0, j ^ 0 in an "unmoving" system, and explain why (fe ^ 0 in a
moving system.
© 33.7 Derive the equation of conservation of electric charge from the Maxwell
equations.
0 33.8 Show that, if the equations div.E = 47rpe and d i v B = 0 are satisfied at
t = (0, then they follow from the rest of the Maxwell equations at all other t.
0 33.9 Starting from the expressions for the force and power acting from an elec
tromagnetic field on a charged particle, find the force and power transmitted by an
electromagnetic field to a continuum of unit volume containing charged particles.
0 33.10 Verify that, due to the Maxwell equations, the four-divergence of the
tensor S'k which is assumed in electrodynamics to be an energy-momentum tensor of
an electromagnetic field (the Maxwell energy-momentum tensor) (see Problem 32.7)
S" = --^{^Fj - iff«F„F") \\gik\\ = \\gik\\-1

equals the density of electromagnetic force peE+j x B/c with negative sign (the first
three components) and the input electromagnetic energy j ■ E divided by c2 with the
negative sign (the fourth component). Extract and express in terms of E and B the
tensor of electromagnetic stresses Sa®, the density of electromagnetic momentum Sai,
the density of electromagnetic energy flux (Poynting vector) (?Sia and the density of
electromagnetic energy c2S44 (see Problem 32.7).
© 33.11 Let the electric current density in a volume be j(t, xa) , and the tangential
component of the electric field ET = ET(x,t) be given on the boundary of that
volume. At the initial instant, the electric and magnetic fields E\t=o = Eo(xa),
B\t=o = B0(xa) are given with div.E 0 = ^Peo, divB 0 = 0 (peo is the initial density
of electric charge). Prove the uniqueness of a) the solution of the formulated initial-
boundary-value problem for the Maxwell equations, b) the solution of the Cauchy
problem.
0 33.12 Solve Problem 33.11 if not electric, but a magnetic field tangential com
ponent BT = BT(x,t) is given on the boundary of the volume.
© 33.13 Show that, when Lorentz transformations (see Problem 31.2) of coor
dinates and time are carried on, only the transformation of electromagnetic field,
following from the fact that F*-7 is a tensor, has the following properties: 1) it trans
forms the solution of Maxwell equations into the solution of the same equations with
the same four-vector of current (with components transformed by the vector rule) in
the right side; 2) it transforms the zero field into the zero field; 3) it does not depend
upon the previous history of the process.
Chapter 9
Models of Media Interacting with
an Electromagnetic Field
34 Magnetohydrodynamics
Magnetohydrodynamics deals with the model used for description of phenomena in
liquids and gases with high electric conductivity. High conductivity is provided by
the presence of a large number of particles with charges of different signs. Under con
ditions typical for application of the model of magnetohydrodynamics, the magnetic
field is considerably stronger than the electric one. The relationship of the electric
current with other characteristics of the field and motion of the medium (Ohm's law)
in the simplest variant is found in Problem 34.2. The internal energy of the medium
is assumed to be independent of the electromagnetic field (there is neither polariza
tion nor magnetization). In Problem 34.5, the conditions under which the model
of magnetohydrodynamics is valid are formulated, and the system of equations for
magnetohydrodynamics is derived.
In the case of smooth functions characterizing the motion, the model of magneto
hydrodynamics is given by the equations
dp dv 1 1
-± + pdivw = 0 , — = —gradp-f- —curl.B x B (34.1)
at at p 4w
dB c2
— - curl (v x B) = -curl (um curl B) , vm = - — (34.2)
(it Aira
where the density of internal energy U is a known function of p and entropy s: U =

U(p, s); a is electric conductivity of the medium, c is the velocity of light, vm is called
237
238 MEDIA INTERACTING WITH AN ELECTROMAGNETIC FIELD
magnetic viscosity coefficient. The equation( 34.2) is called the equation of magnetic
induction. For an incompressible fluid, the equation divv = 0 is supplemented, the
density is assumed to be constant, U = U(s) and the equation p = p2dU/dp must be
excluded from the system.
— PROBLEMS —
Q 34.1 Assuming that the internal energy of the medium does not depend on the
electromagnetic field (more precisely, that the energy of the system — the fluid and
the electromagnetic field — equals the sum of the corresponding energies of the fluid
and the field), show that the densities of entropy production and uncompensated heat
(Chapter 3) (called the Joule heat in the given case), per unit time and volume of
an electro-conductive inviscid non-heat-conductive fluid, equal j ' ■ E'/T and j ' • E',
respectively, with j ' and E' being the current and electric field in the coordinate
system moving together with the element of the medium.
© 34.2 a) Under the condition stated in Problem 34.1, using the methods of ther
modynamics for irreversible processes (Chapter 3) write the linear relation between
the components of the vectors j ' , and E' (Ohm's law) given by j[ = <TikE'k (where
a,* are components of the electric conductivity tensor of the medium). Show that, if
anisotropy of the relation between j ' and E' is determined only by the axial vector
B', then o\k = o{6ik + a b A + /^wfy) where 6, = B'J\B\ and cy/t are components of
the Levi-Civita tensor (a = (5 = 0 and aik = oSik in the isotropic case).
b) Using elementary considerations for the kinetics of conduction, estimate the coef
ficients in the expression for a^.
© 34.3 Using the equations of equilibrium for an "electron gas" and an "ion gas",
find the distribution of charge and electric field in plasma with constant temperature
near a plane wall. Electric potential of the wall is given, the potential at infinity is
assumed to be equal to zero. Determine the characteristic distance Lp from the wall
at which the electric field in plasma practically decays (Debye length).
0 34.4 For a homogeneous isotropic unmoving medium for which Ohm's law j =
aE, a = const, is valid, the distribution of electric charge at t = 0 pe|t=o = Pta(x,y,z)
is given. There is no external field. Determine the variation of pe with time. Explain
why it seems that the charge is not conserved.
© 34.5 Let Ohm's law j ' = aE', or similarly j = peE + a(E + v x B/c), be valid
for a medium.
a) Assuming that the conductivity a is large, determine under what conditions for a
the following statements, which form the basis of magnetohydrodynamics, are valid:
34. Magnetohydrodynamics 239
1) the "displacements current" dE/dt and convective current pev in the first Maxwell
equation( 33.1), with the expression for j from Ohm's law substituted, can be ne
glected as compared with the term oE as well as with the total electric current j ;
2) it can be assumed that B' = B on transition to a moving coordinate system;
3) the electric force peE = Ed\\ E/A-K can be neglected as compared with the mag
netic one j x B/c, and d) the Joule heat input j ' ■ E' can be assumed to be equal to
f/a = c2(cur\B)2/16w2a = //m(curlB)2/47r.
b) Assuming that the restrictions of item a) on the value of electric conductivity
a are satisfied, derive the system( 34.1) - ( 34.3) for a magnetic field and a highly
conductive gas (the equations of magnetohydrodynamics), eliminating E from the
original system of equations describing the electromagnetic field and medium.
© 34.6 a) Estimate the ratio of the second and third terms (magnetic Reynolds
number Rem) in the equation describing magnetic field in a medium with high con
ductivity ("equation of induction")
BB
— curl (v x B) + curl (um curl B) = 0 •
at
(i/m = c2/Ano being the "magnetic viscosity"), or, which is the same, the value of the
ratio of the terms in the expression for the electric field E = (vm curl B — v x B/c)/c
(see solution of Problem 34.5).
b) Write the equation of induction in the integral form similar to integral form of the
corresponding Maxwell equations.
© 34.7 Show that, if E = — v x B/c (the electric field in the proper coordinate
system of an element of a medium vanishes, for example, if a "perfectly conductive"
medium is considered, conductivity of which is infinitely large), then the flux of
magnetic field through any material surface is conserved ("the magnetic field is frozen
in the medium").
© 34.8 Assuming a magnetic field as frozen in the medium (see Problem 34.7),
determine the field if it is given at an initial instant, and the motion of the medium
particles is known.
© 34.9 a) Reduce the components of the density of Lorentz force to the form
1 1 8Tmi
-(j x B)' = -j-(carlB x B)1 = ^
cv ' 4?T dxm
where Tmi = BmB'/4n — 6m'B2/8TT are components of the tensor of magnetic stresses
(compare with Problem 33.10).
b) Write the expression for the total stress tensor T?1 of an perfect gas and magnetic
field. Using the results of solution of Problem 34.8, determine the dependence of
the stress tensor upon the tensor of displacement gradients in the case of perfect
conduction. Write the equations of motion for a perfectly conductive gas, using the
Piola-Kirchhoff stress tensor (see Section 6).
c) Show that, if E = — v x B/c, then the flux of electromagnetic energy through a
material surface, i.e. the integral of the normal component of the Poynting vector (see
Problem 33.10 ) over that surface, equals the sum of the work of magnetic stresses
on that surface and the flux of energy of the magnetic field frozen in the medium.
Q 34.10 Consider a surface of strong discontinuity in magnetohydrodynamics for a

perfectly conductive gas (discontinuity of the magnetic field connected with existence
of a surface electric current on the discontinuity surface is allowed). Write the jump
conditions followed from conservation laws for mass, momentum and energy and the
relation derived from the integral form of the equation of induction (see Problem 34.6).
© 34.11 Linearize the one-dimensional equations of magnetohydrodynamics for a

perfectly conductive medium (see Problem 34.5) with respect to a homogeneous state
of rest, and analyze the linear waves.
0 34.12 The half-space x > 0 is filled with an unmoving medium with conduc
tivity a = const. Let the magnetic field on the boundary x = 0 be given for
- c o < t < oo: B = B 0 sin utez. Find B(x,t) at x > 0. Assume that \B\ —► 0 as
x —* oo. Estimate the depth of penetration of 6 the magnetic field into the conductive
medium (depth of "skin layer").
© 34.13 Let the initial magnetic field in the layer 0 < x < I of an unmoving
medium with conductivity a = const be given: B = B0(x)e2, and B = 0 on the
boundaries of the layer. Estimate the characteristic time T of magnetic field decay at
large t, considering the proper eigenfunctions of the problem (verify that they form
a complete system).
© 34.14 a) Consider the steady flow of a viscous incompressible electrically con

ducting fluid in a non-conductive cylinder tube of an arbitrary cross-section in the
uniform external magnetic field Boo = B0ez, supposing vx = u(y, z), vv — vz = 0,
tp = (p(y, z), (p being the electrical potential. Assume that Ha = B0L/y/4irfii'm » 1
(Ha is the Hartman number, L is the diameter of the tube, /x is the dynamic viscosity
of the fluid, um = c2/4na is the magnetic viscosity).
b) Study the planar flow vx = u(z), vy = vz = 0 of a viscous incompressible fluid with
constant electrical conductivity between immovable walls z = ±H in the magnetic
field B = B(z), for a arbitrary value of Ha. There is a pressure gradient along the
35. Electrohydrodynamics 241
H
x-axis. The total current along the y-axis equals zero: / j v dz = 0, and the external
-H
field is uniform: B^ = Boez (Hartman problem).
0 34.15 Consider the steady flow of an incompressible perfectly conducting fluid
under the condition B = kv.
a) Determine the effect of a magnetic field on the pressure distribution in the flow of
an ideal fluid. Study the case k2 = Atp separately. Using Galilean transformations,
determine the motions (Alfuen waves) for which v = k~lB + const, and B = B0 =
const, v = 0 at infinity.
b) Determine the steady flow of a viscous fluid around a body for B ^ 0 if the flow
around it is known at an arbitrary Reynolds number for B — 0. Consider the cases
k2 < 4np and A:2 > 4irp.
35 Electrohydrodynamics
The model of electrohydrodynamics is applied to describe fluid flows for which consid
eration of the macroscopic electric charge is important. If there is no strong external
magnetic field, then the electric field turns out to be much greater than the magnetic
one (see Problem 35.1). If the medium contains only charges of an identical sign, then
the electric current turns out to be proportional to the density of electric charge (see
Problem 35.2 where the simplest "Ohm's law" for electrohydrodynamics is discussed).
The macroscopic electric charge density in a fluid can be nonzero and essential
for the process for sufficiently long time, if the electric conductivity of the medium is
sufficiently low (see Problem 35.4 where, for the simplest case, effect of electric con
ductivity is estimated). The system of differential equations of electrohydrodynamics
is derived in Problem 35.3. That system for the simplest considered case has the form
-£+pdivv = 0 , p^ = -gradp + peE , (35.1)

at at
- ^ + divj = 0 , divE = 4npe , curlJB = 0 , (35.2)
dt
ds
j = pe(v + bE), PT-=pebE2, (35.3)
v = p
-d-p'T =
Ts (35 4)
-
where b is a mobility coefficient. More complicated models, which are not considered
here, are related to the cases for which additional terms in Ohm's law should be
introduced (e.g., such terms can describe diffusion of charged particles).
— PROBLEMS —
0 35.1 a) Applying a Lorentz transformation, find the magnetic field of a uni

formly moving charge, neglecting the quantity v2/c? as compared with 1.
b) Show that, for the non-relativistic case, the magnetic field of moving charges of an
identical sign does not exceed Ev/c in order of magnitude (E and v are the typical
values of electric field and speed of motion of the charges) if the motion is stationary
or quasi-stationary, so that L/T <C c where T and L are the typical time and length
scales.
c) Verify that the electric field for the case under consideration do not change after
Lorentz transformation with an error not exceeding the quantity v2/(? in order of
magnitude: E' = E.
0 35.2 Using elementary knowledge of the kinetics of the phenomenon, estimate

the order of magnitude of the mobility coefficient 6 in Ohm's law j = pev+pebE which
is valid if there are charge carriers of an identical sign in the medium. Determine the
conditions under which the second term can be neglected as compared with the first
("the electric charge is frozen in the medium").
© 35.3 a) Determine the conditions under which the components of the density
of electromagnetic force can be expressed in the form
b) Deduce the equations (35.1 -35.3) (equations of electrohydrodynamics) in sup

position that the conditions of the item a) are satisfied, Ohm's law of the form
j = pe(v + bE) is valid, the internal energy of the medium depends upon p and
s, and viscosity and heat conduction are negligible.
© 35.4 There are charges of an identical sign with sufficiently large density of
charge (bpe(x*) » <r) at £ = 0 in an unmoving medium for which j = pebE + oE.
Estimate the time and region for which the equality j 3> oE is valid. Assume that
all the quantities depend only upon the Cartesian coordinate i and time t, and, at
x = ±oo, the electric field does not depend on time and the electric current vanishes.
0 35.5 Assuming that all the quantities depend upon one Cartesian coordinate
x, j x = Pevx + pebEx < 0, pe < 0, and, for the velocity of the fluid, v = vxex,
vx = const, analyze the function representing the electric potential ip(x) for the
following boundary conditions: <p = 0, Ex = E0 < 0 at x = 0; <p = (pi at x = I, EQ is
the electric field providing emission of electrons from the electrode. (In general, i?o is
a function of j . If j is small, then E0 can be taken to be constant.) The flow under
35. Electrohydrodynamics 243
consideration can be used to convert electrical energy into mechanical and vice versa
in a pump or in a generator of electric power. Find the efficiency of the corresponding
devices.
Chapter 10
Dimensional Analysis and
Modeling
36 Foundations of the Dimensionality Theory

Using dimensional analysis allows one, for many cases, to improve the way a problem
is posed, to determine significant features and form of the solutions sought, as well
as to ensure that physical experiments dealing with establishment of constitutive
relationships and modelling are carried out competently.
The goal of research is usually to find the relationship between numerical values
of a certain quantity a {determined parameter) and the numerical values of certain
independent of each other quantities a\, a^, ■ ■., On called determining parameters ,
i.e., to find the relationship of the form
a = f(aua2,...,an) . (36.1)
The numerical values of determined and determining parameters depend on the

choice of the basic units, i.e., the measuring units of quantities whose numerical
values are found by immediate comparison with the objects of the same nature. Such
quantities are called primary quantities. A set of basic dimensional units forms a
system of dimensional units.
In mechanics, the units of length (the symbol L), mass (M) and time (T) or the
units of length (L), force (F) and time (T) are usually used as basic units. The
systems of such basic units may be represented in the symbolic form as the systems
LMT or LFT. In the international system of units SI, besides the measuring units
of length, mass and time, the measuring units of current, luminous intensity and
temperature are basic.
For a fixed set of primary quantities, various systems of dimensional units can be
introduced which differ from each other only by the magnitudes of basic measuring
245
246 DIMENSIONAL ANALYSIS AND MODELING
units. Let us call below such a set of systems, e.g., of the type LMT or LFT, a
class of systems of dimensional units or shortly a class of systems and denote by the
symbols {LMT} or {LFT}. In the general case, if there are s primary quantities the
symbols of whose measuring units are A\, A2, ■■■, A„ the class of these systems is
denoted by {At, A2,..., A,}.
Besides the primary quantities the relationship (36.1) can include secondary quan
tities, i.e., quantities that are introduced "by definition" by means of formulae ex
pressing them in terms of the primary or of the secondary quantities introduced
previously. The same formulae are used also for determination of numerical values of
the secondary quantities. To denote the measuring unit of a secondary quantity b,
the symbol [b] is usually used.
With the use of the formula denning a secondary quantity, the relationship of
the symbol of the measuring unit of the secondary quantity with the symbols of the
basic measuring units can be easily obtained for any class of systems. The symbolic
formula expressing that relationship is called the dimension formula or dimension of
the given quantity.
Under the supposition about independence of the formula defining a secondary
quantity and of the ratio of two numerical values of this quantity upon the choice of
a system of units in the class of systems, it can be easily shown that the dimension
formula should be a power monomial. For example, in the class of systems {LMT},
it should have the form
[6] = LaM0r .
Here, a, j3, 7 are the exponents of dimension.
For the general case of a class of systems {At, A2,..., A,}, the dimension formula
of any secondary quantity 6j is written in the form
lbi\ = A°»Aa2»...A»" .
A dimension formula shows how the numerical value of the secondary quantity
changes on transition from a fixed system in the chosen class of systems to any other.
For example, consider a class of systems {A\, A2,..., A,}; if transition from a system
I to a system II is carried out, with basic units ni, n2, ... ns times less in magnitude
than those in the system I, then the relationship of the initial values b\ with the new
values b]1 has obviously the form
6!' = 6 ; n ? " n ^ . . . < » .
Quantities, the numerical values of which change on transition to some other
system of units from a given class of systems, are referred to as dimensional If no
change takes place, the quantities are referred to as dimensionless. In the latter case,
all the exponents of dimension obviously equal zero, i.e.,
On = Qi2 = . . . = Qj, = 0 .
36. Foundations of the Dimensionality Theory 247
If the dimension of any quantity in a set of dimensional quantities cannot be

expressed in terms of the dimensions of the others, those dimensional quantities are
referred to as dimensionally independent. It can easily be shown that the number of
dimensionally independent quantities cannot exceed the number of basic units in the
class of systems.
Consider now a relation of the form a = f(a.i,a2,...,an) describing a certain
physical law. The natural supposition that the form of this relation does not depend
on the choice of the system of dimensional units in the class of systems imposes certain
conditions on the character of that relation, requiring it to have a certain structure.
What allows one to find the character of that structure is the central theorem of
dimensional analysis, known as the U-theorem.
The essence of the IT-theorem consists in the fact that, if the form of a relationship
under consideration
a = /(aii •••, a*, a/b+ii •••,<*») (36.2)
does not depend upon the choice of the system of dimensional units in the class of
systems {A\, A2,..., A,}, all its arguments being independent among themselves and
k of them (e.g., the arguments a,i, a2, ••-, o-k) being dimensionally independent in
this class of systems, then this relationship can be represented as one relating dimen-
sionless parameters II, IIj which are combinations of the determined and determining
parameters, i.e., in the form
n = tp(nu...,Un.k) (36.3)
If the dimensions of the determined quantity a and the arguments a/t+i,..., a„ in
(36.2) in the class of systems {A\, A2,..., A,} are expressed in terms of the dimensions
of the quantities ai, 02, - - -, a* by the formulae
[a] = h p ' H " ' . . . ^ ,
[afc+1] = [a1]Ai[a2]A>...[afcp», (36 4)
* * * J
W = Kf'.-'fa^."-'...^]^."-* ,
then II and 11* in (36.3) have the form
a
n = af'a^...af*'
flk+i
III =
afaf'.-X" (36.5)
•'* )
On
Pl,n-k„02.n-k „Pk,n-k
a
i <h • • • ak
Reduction of the number of the arguments in (36.3) by A: as compared with (36.2)
can essentially simplify analysis of the studied phenomenon, and, for one particular
case for which fc = n, even allows one to determine the form of the initial relationship
(36.2) with the accuracy up to a dimensionless constant C.
In fact, one obtains for that case that
n=c
from which it follows that
a= Callcfc...alk
It can be shown that, in the class of systems {A\, A2l..., A,}, reduction of the
number of the arguments in (36.2) is restricted by the condition fc < s. However,
as has been noted by Huntley, it is possible, for some cases, to increase artificially
the number of primary quantities s, e.g., by introduction of different units of length
in measurement of lengths in two different directions. If the additional parameter
appearing as a result of that procedure, showing by what factor these units differ
from each other, turns out to be unessential for the studied phenomenon, then it
can be not included in the set of determining parameters. Increase in the number of
dimensionaJly independent arguments fc by 1 leads then to reduction of the number
of dimensionless arguments n —fcin the relationship (36.3). Below this procedure is
called the method of Huntley.
The Il-theorem allows one to formulate the concept of similarity of studied phe
nomena of different scales (prototype and model).
Two phenomena of different scales of the same physical nature can be regarded
as similar if they differ from each other only by the numerical values of the dimen
sional determining parameters, and if the dimensionless determining parameters (the
combinations 11;) composed of them and the arguments of the relationship (36.3) are
respectively the same.
In this case, i.e., if the following conditions, called the similarity criteria, are
satisfied
nf = nf, n£ = n», ..., n$_k = n?_ fc , (36.6)
where the superscripts "P" and "M" mark respectively the prototype and model
phenomena, it follows from the Il-theorem that n p = n M from which the relationships
between the dimensional characteristics of these phenomena is obtained, i.e.,
When a study of a prototype phenomenon is replaced by its experimental study

in another scale, i.e, in modelling a phenomenon, the similarity relations (36.6) allow
one to determine the conditions under which the experiments should be carried out
in a laboratory. The experimental data from testing the model obtained can be
transformed to their full-scale values by formula (36.6).
36. Foundations of the Dimensionality Theory 249
The sequence of operations in application of dimensional analysis to specific prob

lems consists usually in the following.
(1) On the basis of the mathematical statement of the problem or, in absence of it,
starting from available experimental data or from physical knowledge of the studied
medium and phenomenon, one writes out as complete as possible set of determining
parameters oi, a2, ..., a„ and the general form of the dependence upon them of the
parameter sought for (determined) (the formula (36.2))
a = /(ai,---,a*.o* + i,...,a„) •
(2) A class of systems {Ai,A2,..., A,} is chosen.
(3) The dimension formulae of the determined and determining parameters are
written out. The dimension of the determined parameters should be expressed in
terms of the dimensions of the determining parameters. If it is impossible, then the
system of determining parameters is not complete and should be supplemented.
(4) The maximal possible set of parameters CL\, a2, .-., a/t with independent
dimensions (there can be no more than s such parameters) is selected from the set of
the arguments of the relationship (36.2).
(5) The expressions for the dimensions of the other arguments and for the pa
rameter sought for in terms of the dimensions of the parameters aj, a2, ..., a* (the
formulae (36.4)) are written out.
(6) The dimensionless combinations II and Ili for the required parameter "a" and
arguments oi, 02, • • •, a* (the formulae (36.5)) are composed.
(7) According to the II-theorem, the relationship (36.3)
n = ^(n,,...,n B _ J k )
can be written.
(8) In modelling a phenomenon, the similarity criteria (formulae (36.6)) are estab
lished, and the data obtained in the test of the model are transformed to prototype
parameters according to the formula (36.7).
The detailed description of the indicated sequence of operations in application
of dimensional analysis to determining the form of specific relations as well as in
modelling is contained in the solutions to Problems 37.1-37.3, 37.20.
Application of the method of Huntley is demonstrated explicitly in the solution
to Problem 37.3.
In a number of cases, if there is a mathematical statement of the problem, di
mensional analysis allows one to decrease the number of independent variables in the
system of differential equations to be solved to determine the desired parameters (e.g.,
it allows one to reduce the system of partial differential equations to a system of or
dinary differential equations). One example of such a problem, the solution of which
belongs to the class of solutions known as self-similar is the classic problem of strong
point explosion, stated and solved in explicit form by L.I. Sedov in 1946. Separate
fragments of this problem are demonstrated in the solutions to Problems 37.24-37.26.
37 Examples of the Application of Dimensional

Analysis.
— PROBLEMS —
0 37.1 Find the weight flow rate G of an ideal heavy fluid per unit width of a
vertical dam with sharp edge and free stream (see Figure 37.1). The height of the
level of the fluid in the reservoir above the edge of the dam far from it equals h.
T
1
Figure 37.1
© 37.2 Find the weight flow rate G of an ideal fluid through a dam in the shape
If
Figure 37.2
of a vertical wall with an angular cut-out in it symmetrical relative to a vertical (see
Figure 37.2). The angle at the vertex of the cut-out equals a. The vertex is at a
depth h relative to the level of the fluid far from the dam.
37. Examples of the Application of Dimensional Analysis. 251
0 37.3 Neglecting air resistance and interaction of particles of a liquid jet blowing
with speed v at an angle a with the horizon, obtain the approximate formulae for the
range £ and height H of the jet (see Figure 37.3).
Figure 37.3
0 37.4 Two jets of an ideal fluid flow out horisontally from two small holes in the
wall of a large open vessel. The holes are located respectively at heights h\ and hi
__l 1, y
H 1
1
I '
i,
\
K
7/
'///. f '/////////.z^<y *
Figure 37.4
above the bottom of the vessel (see Figure 37.4). Neglecting, as in Problem 37.3,
interaction of particles in the jet and air resistance, determine the height H of the
level of the fluid above the upper hole for which the distances d and £ 2 from the
wall of the vessel to the places where the fluid reaches the level of its bottom are the
same, i.e., £i = £2 = £•
0 37.5 A vessel, out of which an ideal fluid flows through a circular hole in the
bottom with diameter d, is continuously supplied with the fluid with volume flow rate
Q (see Figure 37.5).
Determine, with accuracy up to a constant multiplier, the height of the level of the
fluid at which the flow will become steady.
\
~*Ar H
X
1
Figure 37.5
© 37.6 Consider a finite rigid body with characteristic linear dimension d and
shape determined by parameters ln/d. An infinite stream of an ideal weightless in
compressible fluid with density p, speed Vo, and pressure po at infinity, flows over the
body so that there is no flow separation. The orientation of the body in the stream
is determined by angles a* (see Figure 37.6).
Po
Figure 37.6
Show that the ratio Vi/v0 of the speed of the fluid at any point xt on the surface of
the body to the speed of the stream v0 depends only on the coordinates of the point
Xj/d, the shape of the body, and its orientation in the stream, but does not depend
upon v0 and p0.
Obtain a similar conclusion for the pressure coefficient CVi =2(p< — p0)/pVQ.
Q 37.7 Write out the similarity criteria for modelling the flow about a body of an
infinite stream of ideal weightless incompressible liquid, with speed v0 and pressure
Po at infinity. Assume that there is a developed natural cavity in its vicinity inside
which the pressure equals p d — the pressure of saturated vapor of the liquid (see
Figure 37.7).
Specify the geometric parameters of the body by indication of a characteristic linear
dimension d and dimensionless parameters li/d, and the orientation of the body with
respect to the stream, by indication of the angles a*.
37- Examples of the Application of Dimensional Analysis. 253
«„/>,
Figure 37.7
Write out the formula for calculation of the drag force X, as well as D and C using
the data from the model test.
Q 37.8 A rigid body of finite dimensions is in the flow of an infinite stream of

an ideal weightless incompressible liquid with density p, speed Vo, and pressure po at
infinity.
There are domains on the surface of the body where cavitation takes place at the
bubble stage of development. Obtain the condition for cavitation to arise on the
surface of the body, i.e., the condition CPmln = —<7N where Cp = 2(p — po)/pVg is the
pressure coefficient, <rN = 2(po — Vd)lpvo is the natural cavitation number, and p<j is
the pressure of saturated vapor of the liquid.
0 37.9 A progressive gravitational wave with amplitude much less than its length
propagates over the surface of a infinitely deep ideal liquid (see Figure 37.8). Obtain
Figure 37.8
the formula for the speed of propagation of this wave v supposing that it depends
only on the length of the wave A, the density of the liquid p and the gravitational
acceleration g.
0 37.10 A sound wave propagates in an ideal perfect gas with pressure p and
density p. Supposing that compression or expansion in the wave is adiabatic (the
adiabatic exponent equals 7), show that the speed of this wave c is proportional to
y/T where T is the absolute temperature.
Q 37.11 Determine the height h up to which the level of a heavy liquid rises under
action of surface tension force near a vertical plane wettable wall bounding the liquid
(see Figure 37.9).
Figure 37.9
The wetting angle equals 6, the specific weight of the Uquid is 7 = pg (p is density, g
is the gravitational acceleration), and the surface tension coefficient is a.
© 37.12 Determine the height H up to which a heavy Uquid rises under action of
capillary forces in a slot of width 6 between two paraUel plates (see Figure 37.10).
—• 6 ~—
f
H
♦
'—-^zr
Figure 37.10
The wetting angle equals $, the density of the fluid is p, and the surface tension
coefficient is a.
0 37.13 A viscous incompressible fluid with dynamic viscosity \i and density p
moves with constant average speed v in a smooth horizontal cyUndrical tube of circular
cross section with diameter d.
Obtain the general form of the formula for the drag force Pi experienced by the fluid
in the part of the tube of a length I. Use the fact that, according to experimental
data, the pressure varies along the tube Unearly.
Q 37.14 Obtain, without using experimental findings about linear variation of

the pressure along a pipeline, the Stokes law — the formula for the drag force Pi
(see Problem 37.13) in a laminar flow of a fluid with constant speed v in a smooth
cylindrical pipeline of circular cross section. The density of the fluid is p, the dynamic
viscosity is p., and the diameter of the pipeline is d.
© 37.15 Show that the drag coefficient A = (pi — P2)/(j *§-) of a cylindrical
pipeline, in which a viscous fluid moves with average speed v in laminar mode, is
inversely proportional to the Reynolds number Re = pvd/p. Here, p\ — pi is the
difference of pressures in sections 1 and 2 of the pipeline being at a distance / from
each other, p is the dynamic viscosity, p is the density of the fluid, d is the diameter
of the pipeline.
0 37.16 For a steady laminar flow of an incompressible fluid in a circular cylin

drical tube, the volume flow rate Q is determined by the diameter of the tube d, the
pressure drop per unit length of tube i, and the dynamic viscosity of the fluid p.
Obtain the form of the dependence of the flow rate Q upon these parameters.
0 37.17 Obtain the general form of the formula for the drag force X experienced
by a sphere of diameter d during motion with constant speed v in an unbounded
viscous fluid with density p and dynamic viscosity p.
Consider then two limiting cases of very small and very large values of the dimen-
sionless determining parameter — Reynolds number Re = pvd/p = vd/u {v is the
kinematic viscosity).
0 37.18 Obtain the formula for the constant rate of submersion v of a solid sphere
with density pi sinking in a viscous fluid with density p and dynamic viscosity p under
action of the gravity force.
© 37.19 A circular cone with apex spatial angle a penetrates at the instant t = 0
into a half-space filled with an incompressible viscous weightless liquid of a density p
and moves there with constant speed v. The axis of the cone is perpendicular to the
surface of the liquid (see Figure 37.11).
Determine the reaction force of the liquid X at an instant t for the two limiting cases
(a) very quick penetration for which viscosity forces can be neglected as compared
with inertia forces and (b) slow penetration for which inertia forces can be neglected
as compared with friction forces.
© 37.20 Consider a vessel filled with a viscous liquid of density p and dynamic
viscosity p. The initial depth of the liquid equals h0. There is a mouthpiece with
circular cross-section of a diameter d in the bottom of the vessel plugged by a stopper
(see Figure 37.12).
1Figure 37.11
Figure 37.12
Determine the time tQ for which Q liters of the liquid pour out from the vessel after
the hole is opened. The sizes and shapes of the vessel and mouthpiece are given by
the characteristic linear dimension - the diameter of the mouthpiece d and the ratios
of the other linear dimensions k and d (i.e., the parameters k/d).
This unsteady process is modeled using a vessel and mouthpiece whose linear di
mensions are one-nth as large as those of the prototype vessel. Using the similarity
criteria, determine the conditions under which an experiment should be carried out,
and write the formula for transformation of the model test experimental data to the
prototype-scale data, assuming the gravitational accelerations for the prototype <7P
and the model gM to be identical.
© 37.21 A ship with volume displacement D and speed v is designed (see Fig
ure 37.13). Analyze the possibility of determining the resistance W against motion
of the ship by modelling. Specify the size and the shape of the lines of the ship
by giving a certain linear dimension I and ratios of the other linear dimensions and
that parameter, i.e, U/l. Take into account the inertial properties of water (density
p), its viscosity (dynamic viscosity n) and its interaction with gravity (gravitational
acceleration g).
0 37.22 Obtain the similarity relations for modelling the flow in the vicinity of
Figure 37.13
a body flying with constant moderate supersonic speed v in an unbounded perfect

viscous gas with density po, pressure po and dynamic viscosity p. Assume that the
motion is adiabatic and neglect heating of the gas caused by viscous friction. Treat
the viscosity p. as independent of the temperature, i.e., as a constant quantity.
© 37.23 A jet of a viscous incompressible fluid with density p and dynamic viscos
ity p flows out in the direction of the i-axis with speed VQ from a circular cylindrical
tube with cross-sectional area equal to So into an unbounded space filled with the
same fluid. Its impulse per unit time is Jo = pu^So-
+-x
Figure 37.14
In the limiting case of jet flow from an infinitely thin tube (see Figure 37.14) but with
the same finite impulse per unit time Jo (ô —>■ 0, vo —* oo), determine the function
which describes the decrease of the speed of the fluid along the axis of the tube (i.e.,
at y = 0).
Show that the ratio of the speed at any point with coordinates (x, y) of the cross
section of that jet (i.e., by a plane perpendicular to the axis of the jet) to the speed
on the axis of the jet in this section can be represented as a universal function of the
dimensionless variable y/x.
0 37.24 An explosion takes place in a point M of a space filled with an ideal
perfect gas with pressure p0 and density po at the instant t = 0; this implies that
some finite quantity of energy E0 is discharged instantly in a point M. A spherical
shock wave with radius r^ = r2(t) arises and propagates through the gas, separating
the region where the gas is set in motion from the region where it is still at rest (see
Figure 37.15).
v =0
P.
P.
Figure 37.15
Assuming that the explosion is strong, i.e., neglecting the "counter-pressure" (pressure
ahead of the shock wave) po, and that compression of the gas in the shock wave is
adiabatic (the adiabatic exponent is 7), obtain the function r 2 = r 2 (t) describing
propagation of the shock wave.
© 37.25 Applying dimensional analysis, reduce the solution of the problem of
the strong point explosion (see Problem 37.24) to solution of a system of ordinary
differential equations, and write out the boundary conditions for it.
© 37.26 Obtain the similarity relations for the problem of point explosion with
counter-pressure taken into account. Demonstrate how, using a test under laboratory
conditions, one can determine the value of the pressure pf at a given distance from
the place of explosion rf at a fixed instant tf for the prototype.
© 37.27 It is supposed that the speed v of propagation of small disturbances in
a thin elastic rod depends only on the material density p and Young's modulus E.
Determine the form of this dependence.
© 37.28 Describe the modelling procedure for determination of the value of the
displacement h of the end of a homogeneous, massive beam of length I perpendic
ular to and rigidly embedded in a vertical wall due to action of own weight (see
Figure 37.16).
Figure 37.16
Bibliography
[1] Batchelor, G. K., An introduction to fluid dynamics, University Press, Cam

bridge, 1973.
[2] Bowen, R. M., Introduction to continuum mechanics for engineeers, Plenum

Press, New York, 1989.
[3] Brekovskikh, L. M. and Goncharov, V. V., Mechanics of continua and wave

dynamics, Springer-Verlag, Berlin; New York, 1994.
[4] Chung, T. J., Continuum mechanics, Prentice Hall, Englewood Cliffs, N. J.,
1988.
[5] Ciarlet, P., Mathematical elasticity, North Holland, New York, 1988.
[6] Cook, H., editor, Continua: with the Houston problem book, M. Dekker, New
York, 1995.
[7] Davis, J. L., Introduction to dynamics of continuous media, Macmillan, New

York, 1987.
[8] Ericksen, J., Introduction to the thermodynamics of solids, Cnapman & Hall,
London; New York, 1991.
[9] Gennes, P. G. d and Prost, J., The physics of liquid crystals, Clarendon Press,
Oxford University Press, Oxford, New York, 1993.
[10] Gere, J. M. and Timoshenko, S. P., Mechanics of materials, PWS-KENT Pub.

Co., Boston, 3 Edition, 1990.
[11] Giles, R. V., Evett, J. B., and Liu, C , Schaum's outline of theory and problems
of fluid mechanics and hydraulics, McGraw-Hill, New York, 3 Edition, 1994.
[12] Green, A. E. and Adkins, J. E., Large elastic deformations, Clarendon Press,
Oxford, 1970.
259
260 BIBLIOGRAPHY
[13] Groot, S. and Mazur, P., Non-equilibrium Thermodynamics, Interscience Pub.

Co, New York, 1962.
[14] Gurtin, M. E., An introduction to continuum mechanics, Academic Press, New

York, 1981.
[15] Hunt, J. N., Incompressible fluid dynamics, Wiley, New York, 1964.
[16] Hunter, S. C., Mechanics of continuous media, E. Horwood ; Halsted Press,

Chichester ; New York, 2 Edition, 1983.
[17] Kachanov, L. M., Foundations of the theory of plasticity, North-Holland Pub.

Co., Amsterdam, 1971.
[18] Kochin, N. E., Kibel', I. A., and Rose, N. V., Theoretical hydrodynamics, Inter
science Pub. Co., New York, 1964,1965.
[19] Kubo, R., Thermodynamics. An advanced course with problems and solutions,
Northholland Pub. Co., Amsterdam, 1968.
[20] Kulikovskiy A. G., Lyubimov G. A., Magneto-Hydrodynamics, Addison-Wesley

Reading, Massachusetts, 1965.
[21] Lai, M., Rubin, D., and Krempl, E., Introduction to continuum mechanics,
Pergamon Press, Oxford; New York, 3 Edition, 1993.
[22] Lamb, H., Hydrodynamics, University Press, Cambridge; New York, 1973.
[23] Landau, L. D. and Lifshits, E. M., Mechanics, Pergamon Press, Oxford; New
York, 2 Edition, 1976.
[24] Landau, L. D. and Lifshits, E. M., Fluid mechanics, Pergamon Press, Oxford;
New York, 3 Edition, 1983.
[25] Landau, L. D. and Lifshits, E. M., Electrodynamics of continous media, Perga

mon Press, Oxford; New York, 2 Edition, 1984.
[26] Landau, L. D. and Lifshits, E. M., Theory of elasticity, Pergamon Press, Oxford;
New York, 3 Edition, 1986.
[27] Loitsyanskii, L. G.( Mechanics of liquids and gases, Pergamon Press, 1962.
[28] Marsden, J. and Hughes, T., Mathematical foundations of elasticity, Prentice-

Hall, Englewood Cliffs, N.J., 1983.
BIBLIOGRAPHY 261
[29] Mase, G. E., Schaum's outline of theory and problems of continuum mechanics,
McGraw-Hill, New York, 1970.
[30] Maugin, G. A., The thermomechamcs of plasticity and fracture, Cambridge

University Press, New York, 1992.
[31] McConnell, A. J., Applications of tensor analysis, Dover Publications, New

York, 1957.
[32] Milne-Thompson, L. M., Theoretical hydrodynamics, Macmill&n, New York, 5

Edition, 1968.
[33] Muller, I., Thermodynamics, Pitman, Boston, 1985.
[34] Ogden, R., Non linear elastic deformations, Ellis Horwood ; Halsted Press,
Chichester; New York, 1984.
[35] Prager, W., Introduction to mechanics of continua, Ginn, Boston, 1961.
[36] Schouten, J. A., Tensor analysis for physicists, Clarendon Press, Oxford, 2
Edition, 1954.
[37] Sedov, L. I., Stmilartty and dimensional methods m mechanics, Academic Press,
1959.
[38] Sedov, L. I., Introduction to the mechanics of a continuous medium, Addison-

Welsley, Reading, 1965.
[39] Sedov, L. I., Two-dimensional problems m hydrodynamics and aerodynamics,

Interscience Pub. Co., New York, 1965.
[40] Sedov, L. I., A course in continuum mechanics, volume 1-4, Wolters-Noordhoff,

Groningen, 1971-1972.
[41] Smoller, J., Shock waves and reaction-diffusion equations, Springer-Verlag, New
York, 1983.
[42] Sokolnikoff, I. S., Mathematical theory of elasticity, McGraw-Hill, New York, 2

Edition, 1956
[43] Sokolnikoff, I. S., Tensor analysis, theory and applications to geometry and
mechanics of continua, Wiley, New York, 1964.
[44] Spenser, A. J. M., Continuum mechanics, Longman, London; New York, 1980.
262 BIBLIOGRAPHY
[45] Timoshenko, S. P. and Goodier, J. N., Theory of elasticity, Mcgrawhill, New

York, 3 Edition, 1970.
[46] Truesdell, C , A First Course in Rational Continuum Mechanics, Academic
Press, Boston, 2nd Edition, 1991.
[47] Valliappan, S., Continuum mechanics fundamentals, A. A. Balkema, Rotterdam,
1981.
[48] Vekstein, G. E., Physics of continuous media: a collection of problems with
solutions for physics students, Adam Hilger, Bristol, Philadelphia, 1992.
[49] Whitham, G. B., Linear and nonlinear waves, Wiley, New York, 1974.
Index
A boundary conditions for a viscous fluid
acceleration components in concomi 119
tant coordinate system 63 boundary conditions for an elastic solid
acceleration in the Eulerian description 181
4 boundary conditions for an ideal fluid
active loading 218 118
additional energy 86, 104 boundary conditions on a free surface
additional energy input 95, 103 119
adiabatic process 87 boundary layer 142, 151, 152, 209
airfoil 168 boundary layer equations 151
Airy stress function 65, 196, 197 Boussinesq equation 160
Alfven waves 241 Boussinesq turbulent hypothesis 154
Almansi strain tensor 27, 28, 32 bulk modulus 185
angular momentum conservation law
C
49,77
cantilever 191
angular momentum equations 51, 77
Carnot cycle 95, 98
apparent mass 140
Cauchy stress tensor 56, 63, 183
Archimedes force 122
Cauchy-Lagrange integral 67, 136, 165
Archimedes' law 120, 122
Cauchy-Poisson wave problem 156
associated law in plasticity 216
cavitation 137
averaging 209
cavitation number 253
axisymmetric potential motions 131
characteristic form of a differential
B equation 169
barotropic motion 66, 136, 163 characteristics of differential equations
basic units in dimensionality theory 245 169
beam 190 charged particle 232
bending moment 190 Christoffel symbols 16
Bernoulli integral 66, 70, 95, 136, 177 circulation of velocity 129, 133
Biot-Savart formula 133 Clapeyron equation 100
Blasius problem 151 class of systems of dimensional units
Blasius-Chaplygin formulae 139, 141 246
body couple 77, 78, 104, 207 Clausius inequality 89
body force 63, 82, 180 coefficient of bulk viscosity 118
263
coefficient of linear thermal expansion coordinate system, its local basis 13
93, 186 coordinates, Eulerian 4
coefficient of shear viscosity 118 coordinates, Lagrangian 27, 114
coefficient of thermal expansion 186 Coriolis force 162
combustion 175 Couette flow 96, 146
compatibility conditions for strain com couple elasticity 79, 206
ponents 33 couple stress tensor 77, 78, 207
complex potential 129 couple stress vector 77
composite materials 209 covariant components of a vector 13
concomitant coordinate system 33 covariant derivative 16, 32
condition of adhesion 119 covariant transformation rule 14
condition of impermeability 118 creep 223
condition of incompressibility 124 critical velocity 177
conditions across a discontinuity sur curl of a vector field 25
face 51 cylindrical coordinates 18
conditions at infinity 119, 125 cylindrical coordinates, metric tensor
conditions of evolutionarity 112 18
conservation laws 49, 69 D
constitutive equations 80, 91, 207
d'Alembert paradox 140
constitutive equations in couple elastisity Debye length 238
207 deformation plasticity theory 220
constitutive relations for Maxwell visco- determining equations (see also consti
elastic solid 223 tutive equations) 91
constitutive relations for the elastic- determining parameters in the dimen
plastic body 217 sionality theory 245
constitutive relations for Voigt visco- detonation 175
elastic solid 223 detonation adiabat 176
constitutive relations of a rigid per diffusion 103
fectly plastic solid 225 diffusion of vorticity 147
contact discontinuity 111 dimension formula 246
continuity equation 50, 52, 118, 180 dimensional quantities 246
continuity equation in cylindrical coor dimensionality theory 245
dinates 53 dimensionally independent quantities
continuity equation in spherical coordi 247, 248
nates 53 dimensionless quantities 246
contravariant components of a vector Dirichlet problem 125
13 discontinuity of electromagnetic field
contravariant transformation rule 14 233
control surface method 69 discontinuity surface 114, 233
control volume method 69 discrete structure 209
264
dispersing waves 157 energy dissipation rate 143
dispersion of waves 157, 203, 209 energy equation 51, 86, 104
dispersion relation 156, 203, 210, 213 energy-momentum tensor 231
dissipation function 105 enthalpy 91, 94, 98
dissipation inequality 89, 102 entropy 87
dissipation of mechanical energy 143 entropy equation 51, 88
distortion tensor 184 entropy increase in a shock wave 172
divergence of a vector field 25 entropy of a linear elastic body 186
divergence of a velocity vector 31 entropy of a linear elastic solid 98
Doppler effect 166 entropy of a perfect gas 98
double shear 36 entropy production 87, 88, 105, 238
drag coefficient 255 entropy production in a viscous fluid 90
drag force 74, 144, 149, 255 entropy supply 87
dynamic conditions across a tangential entropy variation law 50, 87
discontinuity 111 equation of internal angular momen
dynamic conditions across a weak dis
tum 78
continuity surface 109, 110
equation of internal energy 85, 86, 118,
E 207, 208
eigenvalues of a tensor 10, 59 equation of magnetic induction 238,
eigenvectors of a tensor 10 239
elastic medium 101, 179 equation of state 91, 118, 180
elastic strains in plastic solid 216 equations of angular momentum 77
elastic-plastic solid 216 equations of equilibrium 63
elasticity theory equations in La- equations of equilibrium for fluid 119
grangian coordinates 180 equations of motion 63, 208
electric charge 231, 232, 234 equations of motion for an ideal fluid
electric conductivity 237, 241 (Euler equations) 66, 117
electric conductivity tensor 238 equations of motion in Lagrangian vari
electric current 232 ables 183
electric field 237, 241 equations of shallow water 158
electrohydrodynamics 241 equilibrium equations 63, 64, 182
electrohydrodynamics, system of equa Euler equations 66, 117
tions 241 Euler equations in cylindrical coordi
electromagnetic energy 235 nates 67
electromagnetic field 231 Euler equations in spherical coordi
electromagnetic force 235 nates 67
electromagnetic momentum 235 Eulerian description of motion 4
electromagnetic stresses 235 evolutionarity conditions 172
energy at rest in Relativity theory 230 evolutionary discontinuities 172, 176
energy conservation law 50, 85 evolutionary shock wave 172
265
explosion at a point 68, 173, 257 Hooke's law 67, 93, 179, 181, 183
extension, uniaxial 34 Hugoniot adiabat 111, 175
hydraulic jump 159
F hydrostatics 119
fading memory 223 hypothesis of plane sections 191
first law of thermodynamics 50, 85, 86
flow around a circular cylinder 130 I
flow around a sphere 150 ideal compressible fluid, constitutive
formulae of Blasius-Chaplygin 139 relationships 92
four-dimensional metric 230 ideal fluid 89, 117
four-dimensional velocity 230 ideal incompressible fluid, constitutive
Fourier heat conduction law 87, 181 relationships 91
free energy 91, 98, 181, 209 incompressibility condition 52, 124
free energy of a linear elastic solid 98 inertial coordinate systems 227
Froude number 125 inhomogeneous fluid 120
frozen electric charge 242 instability of Kelvin-Helmholtz 161
frozen magnetic field 239 internal angular momentum 78, 104,
207
G internal energy 85, 98, 104, 162, 181,
gas dynamics, system of equations 164 207
Gibbs identity 90, 105, 180 internal orientation 78
Gibbs thermodynamic potential 91, 98 invariants of a tensor 12
Green identity 127 inviscid fluid 117
Green strain tensor 27, 28, 32 irreversible process 87, 100
Gromeka-Lamb equations 135 isothermal process 87
group speed 157, 203, 210 isotropic compression 185
gyromagnetic coefficients 78 isotropy 43
H J
Hamilton-Cayley theorem 44 Jaumann derivative 80, 82
hardening 215, 217 Joukovski formula 74, 139
hardening parameter 217 Joukovski-Chaplygin postulate 130
harmonic function 125 Joule heat 238, 239
harmonic wave 156, 203 jump conditions 109, 110
Hartman number 240 jump conditions for an ideal fluid 111
Hartman problem 241
heat capacity 87 K
heat conduction equation 95 Kelvin circulation theorem 125, 133,
heat flux vector 86, 93 164
Helmholtz equation 133, 147 Kelvin wave 162
Hill's spherical vortex 135 kinematic conditions across a disconti
homogeneous fluid 120 nuity surface 109
266
kinematic conditions across a tangen magnetic fluid 108
tial discontinuity 111 magnetic induction 231
kinematic conditions across a weak dis magnetic Reynolds number 239
continuity surface 109, 110 magnetic viscosity coefficient 238
kinematical viscosity coefficient 142 magnetizable media 78
kinetic energy theorem 68, 85, 86 magnetization 82, 233
kinetic relationships 90, 91 magnetized liquid crystal 81
Korteweg-de Vries equation 160 magnetohydrodynamics 237
Kronecker symbols 9 magnetohydrodynamics, system of equa
Kutta condition 130 tions 237
mass at rest in Relativity theory 230
L mass conservation law 49, 52
Lagrange interpolation polynomial 46 mathematical models of continua 91
Lagrange's theorem 165 Maxwell energy-momentum tensor 234
Lagrangian coordinates 3 Maxwell equations 232, 233
Lagrangian description of motion 3 Maxwell equations, integral form 232
Lame coefficients 68, 93, 182 Maxwell visco-elastic solid 223
Lame equations 68, 182 method of conformal mappings 129
laminar boundary layer 151 method of Huntley 248
laminar motion 153 method of mirror reflections 129
Laplace equation 125 method of singular points 129
Laplace operator 67 metric tensor 17
Larmor frequency 234 microstructure 209
Levi-Civita tensor 23, 31 Minkowski space 227
light velocity 228 Mises yield criterion 216
limit load 222 model of the atmosphere 123
linear couple elasticity 209, 210 momentum conservation law 49
linear elastic medium 67, 181 momentum equations 50, 63
linearly viscous fluid 67, 92, 103 monochromatic wave 165
liquid crystal 78 motion description, Eulerian 4
motion description, Lagrangian 3
logarithmic velocity distribution in tur
motion law 3, 114
bulent flow 155
Murnaghan medium 205
long waves 158
longitudinal wave 203, 210
Lorentz coordinate system 227 N
Lorentz force 231, 239 Navier-Stokes equations 67, 118, 142
Lorentz transformations 228 Navier-Stokes model 118
nematic liquid crystal 78, 104
M Neumann problem 125
Mach number 125, 163 neutral axis of bent beam 190, 191
magnetic field 82, 237, 241 neutral loading 218
267
Newtonian fluid 92, 118 potential motion 53, 125
non-Newtonian fluids 118 Poynting vector 235, 240
normality rule in plasticity 216 Prandtl boundary layer equations 151
Prandtl turbulent hypothesis 154
O Prandtl-Reuss equations 219
Ohm's law 238, 242 pressure coefficient 252
one-dimensional unsteady motion 166, pressure function 66, 136
169, 203 principal axes of a tensor 59
Onsager principle 90, 103, 105 principal components of a tensor 59
Orr-Sommerfeld equation 148 principal coordinate system of a tensor
orthotropic medium 183, 187 59
orthotropy 47 problem of Cauchy-Poisson 156
problem of Dirichlet 125
P problem of Neumann 125
paradox of d'Alembert 140 progressive wave 156
perfect gas 91, 92, 101, 123 propeller, ideal 112
perfect plasticity 215 proper coordinate system 229
phase speed 156, 157, 203 proper length in relativity theory 229
physical components 20 proper time in relativity theory 229
Pi-theorem (II-theorem) 69, 247 propulsive force of a rocket engine 75
Piola-Kirchhoff stress tensor 56, 62, 63, pure bending 190, 191
183
planar potential flows 128 Q
planar source 130 quasi-one-dimensional flow 178
plane problem of elasticty theory 196
R
plane strain state 196
Rayleigh waves 204
plane stress state 196
resonance phenomenon 159
plane waves 166
reversible process 87, 89
plastic flow theory 215, 216
Reynolds equations 154
plastic strains 216
Reynolds number 142
plasticity 215
Riemann invariants 159, 169
point vortex 130, 133, 177
Riemann wave 159, 170, 206
Poiseuille flow 146
rigid-plastic solids 217
Poisson adiabat 94
Rossby wave 162
Poisson coefficient 182
rotational flow 133
Poisson equation 124
rotor of a vector field 25
polarization 233
polarizing media 78 S
polytropic atmosphere 124 second law of thermodynamics 50, 87,
potential body force 66 88
potential flow 125 self-similar problem 69
268
shallow water approximation 158 stress concentration on a hole 198
shear modulus 184 stress function 65, 196, 197
shock adiabat 175 stress surface 59
shock wave 169, 171 stress tensor 56, 63
similarity criteria 248 stress tensor, its antisymmetric part 79,
similarity of phenomena 248 105
simple extension 184 stress vector 55
simple loading 220 Strouhal number 125, 142
simple shear 35, 36, 184 structure of a detonation wave 176
solitary wave (soliton) 160 structure of a shock wave 172
sound propagation in mixture 168 structure of slow combustion front 176
sound wave 167 substantive time derivative 4
spatial coordinates 4 summation convention 9
surface couple 77, 104
spatial source 177
surface of strong discontinuity 109, 114
special relativity theory 227
surface of weak discontinuity 109
specific heat 87
surface tension 97, 254
speed of sound 163
symmetry group of a tensor field 44
speeds of characteristics 159
symmetry group of tensors 43
spherical coordinates 18 system of dimensional units 245
spherical coordinates, metric tensor 18
spherical tensor 12 T
St.Venant principle 187, 191 tangential discontinuity 111, 145
stability of a flow 161 tensor 9
stability of a steady flow of viscous fluid tensor deviator 12
142, 146, 148 tensor function 43, 81, 107
stagnation parameters 177 tensor of electromagnetic field 233
standing wave 157 tensor of magnetic stresses 239
statics of fluids 119 tensor product 9
steady motion 5 tensor, its contravariant components 14
Stokes approximation 142, 148 tensor, its covariant components 14
Stokes equations 148 tensor, its invariants 10, 12, 52
Stokes formula for drag force 150 tensor, its mixed components 15
Stokes theorem 133 tensor, its physical components 21
strain rate tensor 30, 39 tensor, its principal axes 10
strain tensor of Almansi 27, 28, 33, 52 tensor, its principal components 10
strain tensor of Green 27, 28, 33, 52 thermal conductivity coefficient 87
strain tensor, linearized 28 thermoelastic medium, constitutive re
strain tensor, small 29, 181 lationships 92
stream function 129, 131 torsion 194
streamlines 5 transverse isotropy 47
269
transverse wave 203, 210 work of internal surface forces 68, 86,
transversely isotropic medium 183, 187 93, 104
travelling wave 203
Y
Tresca yield criterion 216, 217
turbulence 142 yield criterion 216
turbulent heat flux 155 yield stress in pure shear 218
turbulent motion 153 yield stress in pure tension 218
turbulent Reynolds stresses 154 yield surface 216
twisting angle 194 Young's modulus 182
Z
U Zhukovskii formula 74
uncompensated heat 88, 105, 238
unloading 216, 218
V
Van der Waals equation 101
velocity of light 231, 237
visco-elastic Maxwell solid 223
visco-elastic Voigt solid 223
visco-plastic solid 223
viscosity 223
viscosity coefficients 67, 92, 118
viscous compressible fluid, constitutive
relationships 92
viscous fluid 89, 117
viscous incompressible fluid, constitu
tive relationships 92
viscous stress tensor 89
Voigt visco-elastic solid 223
vortex sheet 134
vorticity 147
vorticity vector 31, 105, 133
W
wave equation 165, 203
wave packet 157
waves in stratified fluids 160
waves of small amplitudes 156
weak discontinuity 114
work of external and internal forces 68
work of internal couples 104
270

Continuum Mechrnics Vir Problems RND Exercises: Parti Theorq and Problems

Uploaded by

Document Information

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

Continuum Mechrnics Vir Problems RND Exercises: Parti Theorq and Problems

Uploaded by

Copyright:

Available Formats

CONTINUUM MECHRNICS VIR

PROBLEMS RND EXERCISES

Editor: Leon O. Chua

Series A. MONOGRAPHS AND TREATISES

CONTINUUM MECHANICS VIA

Translation from Russian.- A. N. Tlatlushkln

British Library Cataloguing-ln-Publicatkm Data

CONTINUUM MECHANICS VIA PROBLEMS & EXERCISES

ISBN 981-02-2546-6 (Set)

This book is printed on acid-free paper.

Printed in Singapore by Uto-Print

I THEORY AND PROBLEMS 1

2 General Laws and Equations of Continuum Mechanics 49

15 Restrictions Imposed by The Laws of Thermodynamics on the Form

4 Discontinuity Surfaces in Continuum Media 109

5 Fluid Mechanics 117

6 Elasticity Theory 179

7 Inelastic Solids 215

8 Basic Notions of Relativistic Kinematics and Dynamics. General

9 Models of Media Interacting with an Electromagnetic Field 237

10 Dimensional Analysis and Modeling 245

in nonlinear elasticity, microinhomogeneous averaging, interaction of continua with

cv - specific heat capacity at constant volume

dA - the amount of work for time interval dt

drf - uncompensated heat

- rate of entropy production per unit mass

e.i - covariant basis vectors

e* - contravariant basis vectors

i - specific enthalpy, imaginary unit

D - speed of a discontinuity surface

E - electric field strength

T - free energy per unit mass

F - body force per unit mass

Qn - components of couple stress tensor

U - internal energy; potential of body force

Q - coefficient of thermal expansion;

7 - adiabatic exponent, the ratio Cp/cv

£ij - components of a strain tensor

*ijk - components of the Levi-Civita tensor

X - one of the viscosity coefficients; one of the elastic coefficients

/x - viscosity coefficient; one of the elastic coefficients

a - Poisson coefficient, electrical conductivity coefficient

Basic Concepts used for

1 Lagrangian and Eulerian Description of Motion

acceleration of the particles of a continuum are defined by the formulae

where £ = (£i>&2,60- In general, the rate of change in some quantity A in an

dA{x,t) dA(x,t) . A,dA{x,t) . A.dA(x,t) , ,dA(x,t)

In particular, the acceleration a(x,t) in Eulerian description is determined by the

where Vi(x,t) are velocity components.

X\\t=0 ~ £l . X2\t=0 = & . X3\t=0 = £3

0 1.1 Introduce a spatial coordinate system and Lagrangian coordinates of the

* ! = & ( ! + £), *2=&(l+2£), 1 3 = 6 (l + £ )

•> — ii??- * - irP' * - ° <" - v^>-«(«> "■") * - ^w

© 1.20 A medium moves with the velocity field

vi(x, t) = at , u 2 (i, t) = —«— , v3(x, t) = 0 (a,u = const)

and the temperature field

2 Tensors and their Cartesian Components

Tensors. For a pair of vectors a, b, their tensor product ab is introduced. Tensor

(QIOI + a2a2)(/?i&i + fob2) = ai&ai&i + anfoafo + a 2 /?ia 2 6i + a202a2b2 .

tensor product A B of the tensors A = ^ e ^ and B = B W m e k eje m is the tensor of

q = q^ei, qjt = Qiju .

Uj = ^tkk6ij , ttj = tij - -tkk6ij

3 Curvilinear Coordinate Systems

0 3.17 Components of a vector (having identical dimensions in a Cartesian co

xi = fi - a()6£$ , x2 = & + «()£ 16 , x3 = £3