LS Srinath

r.=._...;;.
1Tala McGrawHill
Published by the Tata McGraw-Hili Publishing Company Limited,
7 West Patel Nagar, New Delhi 110008.
Copyright 02009 by Tata McGraw-Hili Publishing Company Limited.
Second reprint 2008
RCXDCDDXRCRDC
No part of this publication may be reproduced or distributed in any form or by any
means, electronic, mechanical, photocopying, recording, or otherwise or stored in a data-
base or retrieval system without the prior written permission of the publishers. The
program listings (if any) may be entered, stored and executed in a computer system, but
they may not be reproduced for publication.
This edition can be exported from India only by the publishers,
Tala McGraw-Hili Publishing Company Limited.
ISBN13: 978-0-07-013988-6
ISBN to: 0-07-013988-1
Managing Director: Ajay Shukla
General Manager: Publishing-SEM & Tech Ed: Vlbha Mahajan
Sponsoring Editor: Shuku Mukherjee
J r. Editorial Executive: Surabhi Shukla
Executive-Editorial Services: Sohini Mukherjee
Senior Production Manager: P L Pandita
General Manager: Marketing-Higher Education &School: Michael J Cruz
Product Manager: SEM &Tech Ed: Biju. Ganesan
Controller-Production: Rajender P Ghansela
Asst. General Manager-Production: B L Dogra
Information contained in this work has been obtained by Tala McGraw-Hili, from
sources believed to be reliable. However, neither Tata McGraw-Hill nor its authors
guarantee the accuracy or completeness of any information published herein, and
neither Tata McGraw-Hili nor its authors shall be responsible for any errors, omis-
sions, or damages arising out of use of this information. This work is published
with the understanding that Tala McGraw-Hili and its authors are supplying infer-
marion but are not attempting to render engineering or other professional services.
If such services are required, the assistance of an appropriate professional should
be sought.
Typeset at Astral Pre Media Pvt Ltd., A-39, Sector-58, Noida (UP) and printed at
Rashtriya Printers, M-135, Panchsheel Garden, Naveen Shahdara, Delhi 110032
Cover: Rashtriya Printers
The McGraw-Hili Companies ~.~'it~ ~~(~f::'~i ~
Copyrigtlted material
Contents
Preface v
List ofSvmbols XUI
SI' Ul1US {Systeme International d'Unit'es) xv
r_"pical Physical COI1S(Cll1ts (As all Aid to Solving Problems)
.
XVI
1. Analysisof Stress 1
'1.1 Introduction 1
1.2 Body Force, Surface Force and Stress Vector 2
13 The State of Stress at a Point 4
'1.4 'N'ormal aDd Shear Stress Component's 4
1.5 Rectangular Stress (.;oolponent.s 4
1.6 Stress Co,mponeots on an Arbitrary Plane 6
1.7 Digression on [deal Fluid 1. 1
1.8 Eqa]ity of Cross Shears 1,
1.9 A 1-1Qre(J eneral Theorenl 13
1.10 Principal Stresses 14
111 Siress Invariants /6
1.12 Pri.ncipal Planes are()rthogonal 17
1.13 (;ubic .Equation has -rhree Real Roots J 7
1.'14 Particular Cases 19
1.15 Recapitulation 19
1.16 The State ofSlress Referred toPrincipal Axes 24
1.17 Mohr's Circles for the Three-Dimensjonal State of Stress 25
1.18 Mohr's Stress Plane 26
1.'19 Planes of Maximum Shear 28
1.20 (lctahedral Stresses 29
1.21 The State of Pure Shear 31
1.22 Decomposition into Hydrostatic and .PureShear States 3 . 1
1.23 Caucby's Stress Quadric 34
1.24 LaIne' sEllipsoid 36
1.25 The Plane State of Stress 38
.1.26 'Differential Equations of ,Equilibrium 40
1.27 Equil.ibrium Equations for Plane Stress State 42
1.28 Boundary Conditions 45
1.29 Equations ofEq-uilihr.i:um in Cylindrical Coordinate-s 45
Copyrighted material
'vl Contents
1.30 Axisymmetric Case and Plane Stress Case 48
Problems 49
Append.i.x 1 Mohr's Circles 54
Appendix 2
Appendix 3
Tbe State of Pure Shear 56
Stress Quadric of Cauchy 60
2. A.nalysis of Strain
2.1 Introduction 63
22 Deformations 64
2.3 Deformation in the Neigbbourhood of a Point 65
2.4 Change in Length of a Linear Element 67
2.5 Cbange in Length of a Linear Element-Linear Components 69
2.6 Rectangular Strain CODJ pOnents 70
2,7 'fhe State of Strain at a Point 70
2,8
63
7 9
~,
Interpretation of y x y , Y y z , Y x z as Shear Strain Components
C:hange in Direction of a Linear Element 73
71
710 Cubical Dilatation 7 4
2,11 Change in the Angle between 'Two Line Elements 77
2.12 Principal Axes of Strain and Prine.ipai Strains 78
') 13 Plant~Siale 1) 1" Strain 8~
2.14 The Pri.ncipal .t\xes of Strain R.elllain Orthogonal after Strain 84
2.16 Coolpatibility C;onditions 86
2.17 Strain Deviator and its Invariants 90
Prob/ell/.I 91
Appendix on Compatibility Conditions 94
3. Stress-Strain Relations for Linearly Elastic Solids 97
3, I Introduction 97
3,2 Generalised Staiemeni:of flQoke's Law 97
3,3 Stress-Strain Relations for Isotropic Materials 98
3,4 Modulus of ,Rigidity .99
3,5 Bulk M'odulus 101
3,6 Young's M.odulus and Poisson's Rati.o 102
3.7 Relations between. the Elastic Coostant's 102
3,8 DisplacCIllcnt Equations ofEguilibri.lllll . I 04
Problems 10Z
4. '("heories of .-ailure or Yield Criteria and Introduction to
ldeaUv Plastic Solid 109
4,I '1ntroductjoo , 09
4,2 Theories of Failure 110
4,3 Si.guifi.cance of the l~he()'ri.esof Failure III
4,4 Use of Factor of Safety in Design 121
4.5 A Note on the usc of Factor of Safety 124
4.6 Mohr's Theory of Failure 129
4.7 Ideally Plastic Solid 132
4.8 Stress Space and Strain Space J 34
Contents vii
49 General Nature of the Yield Locus /35
4.10 Yield Surfaces of Tresca and Von Mises 136
4.11 Stress-Strain Relations (Plastic Flow) 137
4.12 Prandtl-Reuss Equations J 39
4.13 Saint Venant-Von Mises Equations 140
Problems 140
5. Energy Methods
5.1 Introduction 143
52 Hooke's Law and the Principle of Superposition 143
5.3 Corresponding Force and Displacement or Work-Absorbing
Component of Displacement 145
5.4 Work Done by Forces and Elastic Strain Energy Stored 146
5.5 Reciprocal Relation 147
5.6 Maxwell-Betti-Rayleigh Reciprocal Theorem 148
5.7 Generalised Forces and Displacements 149
5.8 Begg's Deformeter 152
5.9 First Theorem of Castiglia no /53
5.10 Expressions for Strain Energy 155
5.11 Fictitious Load Method /61
5.12 Superposition of Elastic Energies 163
5.13 Statically Indeterminate Structures 164
5:14 TheQrem of Virtual Work 16(i
5.15 Kirchhoff ts Theorem 169
5.16 Second Theorem of Castigliano or Menabrea's Theorem 170
5.17 Generalisation of Castiglia no's Theorem or Engesser's
Theorem 173
5.18 Maxwell-Mohr Integrals 176
Problems 181
6. Bending of Beams 189
62 Straight Beams and Asymmetrical Bending 190
6.3 Regarding Euler-Bernoulli Hypothesis 198
6.4 Shear Centre or Centre of Flexure 20J
6.5 Shear Stresses in Thin-Walled Open Sections: Shear Centre 202
6.6 Shear Centres for a Few Other Sections 208
6.7 Bending of Curved Beams (Winkler-Bach Formula) 209
6.8 Deflections of Thick Curved Bars 216
Problems 223
143
7. Torsion
72 Torsion of General Prislnatic Bars-Sol idSections 232
7.3 .Alternative Approach 236
7.4 Torsion of Circular and Elliptical Bars 240
7.5 Torsion of Equilateral Triangular Bar 243
7.6 Torsion of Rectangular Bars 245
230 \
'iill'
. .
Contents
7.7 Membrane Analogy 248
7.8 Torsion of Thin-Walled Tubes 249
7.9 Torsion of Thin-Walled Multiple-Cell Closed Sections 251
7.10 Torsion of Bars with Thin Rectangular Sections 255
7.11 Torsion of Rolled Sections 256
7.12 MUltiply Connected Sections 259
7.13 Centre of Twist and Flexural Centre 264
Problems 265
8. Axisymmetric Problems 269
8.2 Thick-Walled Cylinder Subjected to Internal and External
Pressures-Lame's Problem 27J
8.3 Stresses in Composite Tubes-Shrink Fits 280
8.4 Sphere with Purely Radial Displacements 287
8.5 Stresses Due to Gravitation 292
8.6 Rotating Disks of Uniform Thickness 294
8.7 Disks of Variable Thickness 298
8.8 Rotating Sbafts and Cylinders 300
8.9 Summary of Results for use in Problems 303
Problems 305
9. Thermal Stresses 310
9.1 Introduction 3I0
9.2 Thermoelastic Stress-Strain Relations 311
9.3 Equations of Equilibrium 311
9.4 Strain-Displacement Relations 312
9.5 SomeGeneral Results 312
9.6 Thin Circular Disk: Temperature Symmetrical about Centre 314
9.7 Long Circular Cylinder 3/6
9.8 The Problem of aSphere 320
9.9 Normal Stresses in Straight Beams due to Thermal Loading 323
9.10 Stresses in Curved Beams due to Thermal Loading 325
Problems 328
10. Elastic Stability
10.1 Euler'sBuckling Load 331
331
I. BeamColumns 335
10.2 BeamColumn 335
10.3 BeamColumn Equations 335
10.4 BeamColumn with aConcentrated Load 336
10.5 BeamColumn with Several Concentrated Loads 339
10.6 Continuous Lateral Load 340
10.7 Beam-Column with EndCouple 342
IL General Treatment of Column Stability Problems (Asan
Eigenvalue Problem) 344
10.8 General Differential Equation and Specific Examples 344
Contents "ix
109 Buckling Problem as aCharacteristic Value (Eigenvalue)
Problem 350
10.10 The Orthogonality Relations 352
m Energy Methods for Buckling Problems 355
10.11 Theorem of Stationary Potential Energy 355
10.12 Comparison with the Principle of Conservation of Energy 357
10.13 Energy and Stability Considerations 358
10.14 Application to Buckling Problems 359
10.15 The Rayleigh-Ritz Method 360
10.16 Timoshenko's Concept of Solving Buckling Problems 364
10.17 Columns with Variable Cross-Sections 366
10.18 Use of Trigonometric Series 368
Problems 371
11. Introduction to Composite Materials
11.2 Stress-Strain Relal'jons 375
. 11.3 Basic Cases of Elastic Symmetry 377
11.4 Laminates 381
11.5 Ply Stress and Ply Strain 404
11.6 Failure Criteria of Composite Materials 406
11.7 Micromechanics of Composites 411
11.8 Pressure Vessels 421
11.9 Transverse Stresses 422
Problems 424
12. Introduction to Stress Concentration and Fracture Mechanics 428
L
12.1
12.2
12.3
12.4
12.5
12.6
D.
12.7
12.8
12.9
12.10
12.11
12.12
12.13
12.14
12.15
12.16
374
Stress Concentration 428
Introduction 428
Members under Tension 429
Members under Torsion' 439
Members under Bending 443
Notch Sensitivity 445
Contact Stresses 446
Fracture Mechanics 457
Brittle Fracture 457
.
Stress Intensity Factor 458
Fracture Toughness 460
Fracture Conditions 462
Fracture Modes 464
Plane Stress and Plane Strain 468
Plastic Collapse at aNotch 47J
Experimental Determination ofK
Ic
475
Strain-Energy Release Rate 476
,
Meaning of Energy Criterion 479
- : ; i I . (
. '.
Contents
._---- ------------
12.17 Design Consideration 482
12.18 Elasto-Plastic Fracture Mechanics (EPFM) 482
12.l9 Plane Body 485
12.20 Green's Thorem 486
1221 Thel-J ntegral 486
12.22 Path Independence of the l-lntegral 487
1223 l-Integral as aFracture Criterion 489
12.24 ASTM-standard Test for llc 490
12.25 Relationships of K
c
, o.. and J 491
Problems 49J
Appendix 494
5() J
,
L
Copyri gtlted materi al
I I
Pr ef ac e
The present edition of the book isacompletely revised version of the earlier two
'editions. The second edition provided an opportunity to correct several
typographical errors and wrong answers to some problems. Also, in addition,
based on many suggestions received, achapter on composite materials was also
added and this addition was well received. Since this is a second-level course
addressed to senior level students, many suggestions were being received to
add several specialized topics. While it was difficult to accommodate all
suggestions in a book of this type, still, a few topics due to their importance
needed to be included and a new edition became necessary. As in the earlier
editions, the first five chapters deal with the genera] analysis of mechanics of
deformable solids. The contents of these chapters provide a firm foundation to
the mechanics of deformable solids which will enable the student to analyse and
solve avariety of strength-related design problems encountered in practice. The
second reason is to bring into focus the assumptions made in obtaining several
basic equations. Instances are many where equations presented in handbooks
are used to solve practical problems without examining whether the conditions
under which those equations were obtained are satisfied or not.
The treatment starts with Analysis of stress, Analysis of strain, and Stress-
Strain relations for isotropic solids. These chapters are quite exhaustive and
include materials not usually found in standard books. Chapter 4 dealing with
Theories of Failure or Yield Criteria isageneral departure from older texts. This
treatment is brought earlier because, in applying any design equation in strength
related problems, an understanding of the possible factors for failure, depending
on the material properties, is highly desirable. Mohr's theory of failure has been
considerably enlarged because of its practical application. Chapter 5 deals with
energy methods, which is one of the important topics and hence, is discussed in
great detail. The discussions in this chapter are important because of their
applicability to a wide variety of problems. The coverage is exhaustive and
discusses the theorems of Virtual Work, Castigliano, Kirchhoff, Menabria,
Engesser, and Maxwell-Mohr integrals. Several worked examples illustrate the
applications of these theorems.
xii Preface
Bendingof beams, Centreof flexure, CurvedBeams, etc., arecoveredinChapter 6.
This chapter also discusses the validity of Euler-Bernoulli hypothesis in the
derivations of beam equations. Torsion is covered in great detail in Chapter 7.
Torsion of circular, elLiptical, equilateral triangular bars, thin-walled multiple cell
sections, etc., are discussed. Another notable inclusion in this chapter is the
torsion of bars with multiply connected sections which, in spite of its importance,
is not found in standard texts. Analysis of axisymmetric problems likecomposite
tubes under internal and external pressures, rotating disks, shafts and cylinders
can be foundin Chapter 8.
Stresses and deformations caused in bodies due to thermal gradients need
special attention because of their frequent occurrences. Usually, these problems
are treated in books onThermoelasticity. Tbeanalysis of thermal stress problems
are not any more complicated than the traditional problems discussed in books
on Advanced Mechanics of Solids. Chapter 9 in this book covers thermal stress
problems.
Elastic instability problems are covered in Chapter] O.In addition totopics on
Beam Columns, this chapter exposes the student to the instability problem as an
eigenvalue problem. This isanimportant concept that astudent has to appreciate.
Energy methods as those of Rayleigh-Ritz, Timoshenko, use of trigonometric
series, etc., to solve buckling problems find their place in this chapter.
Introduction to the mechanics of composites is found in Chapter 11. Modern-
day engineering practices and manufacturing industries make use of avariety of
composites. This chapter provides a good foundation to this topic. The subject
material is a natural extension from isotropic solids to anisotropic soLids.
Orthotropic materials, off-axis loading, angle-ply and cross-ply laminates, failure
criteria for composites, effects of Poisson's ratio, etc., arecovered with adequate
number of worked examples.
Stress concentration and fracture are important considerations in engineering
design. Using the theory-of-elasticity approach, problems i.nthese aspects are
discussed in books solely devoted to these. However, a good introduction to
these important topics can be provided in abook of the present type. Chapter 12
provides a fairly good coverage with a sufficient number of worked examples.
Several practical problems can besolved with confidence based. on the treatment
provided.
While SI units areused inmost of numerical examples and problems, afew can
be found with kgf, meter and second units. This is done deliberately to make the
student conversant with the use of both sets of units since indaily life, kgf is
used for force and weight measurements. In those problems where kgf units are
used, their equivalents in SI units are also given.
The web supplements can be accessed at http://www.mhhe.com/srinathiams3e
and itcontains the following material:
For Instructors
Solution Manual
PowerPoint Lecture Slides
'~iii
Preface
For Students
MCQ's (interactive)
Model Question Papers
I am thankful to all the reviewers who took out time to review this book and
gave me their suggestions. Their names are given below.
K S R K Murthy Department of Mechanical Engineering,
Indian Institute of Technology, Guwahati, Assam
P K Sarkar Department of Mechanical Engineering,
Indian School of Mines, Dhanbad, Bihar
S R Pandey Department of Applied Mechanics,
,VIT Jamshedpur, Jharkhand
Dr Amit Kumar Department of Mechanical Engineering,
National Institute of Technology, Patna, Bihar
M K Singha Department of Applied Mechanics,
Indian Institute of Technology, New Delhi
P Venkitanarayanan Department of Applied Mechanics,
Indian Institute of Technology,
Kanpur. Uttar Pradesh
GRajesh Kumar Department of Mechanical Engineering,
Rizvi College of Engineering,
Mumbai, Maharashtra
C A Akhadkar Department of Mechanical Engineering,
SSVPS's B SDeora College of Engineering,
Dhule, Maharashtra
D Prasanna Venkatesh Department of Mechanical Engineering,
SRM Institute of Science and Technology.
Chennai, Tamil Nadu
E V M Sargunar Sree Sastha Institute ofEngineering and
Technology,
Chennai, Tamil Nadu
In addition to this, Iam also thankful to the staff at McGraw-Hill Education India,
especially MsVibha Mabajan, Ms Shu.k:tiMukherjee, MsSurabhi Shukla, Ms Sobini
Mukherjee and Mr P L Pandita, for their cooperation during thedifferent stages of this
project.
Lastly, I wish tothank lOY family members for their patience, support and lovegiven
to meduring thepreparation of tbis manuscript. .
Feedback and suggestions are always welcome at thewebsite of thebook.
L SSRlNATB
(J
F
/I
T
II
C F x.y,
l'X)I, yz, zx
n
x
' NY' n,
C F " C F
2
, C F
3
1,,1
2
, I')
(Joel
'rOCI
C F
r
, C F
o
, C F .
t.8, I f >
" e
yo
f yz, " ea.
u
x
' UY' liz
E X X ' E Y > ' s;
L ist of Sym bels
(In the order they appear
In ,the text)
normal stress
f orce
f orce vector on a plane with normal n
components of f orce vector in x, y, z directions
area of section
normal to the section
shear stress
normal stress on x-plane, y-plane, z-plane
shear stress on x-plane in j-direction, shear stress
on y-plane in a-direction, shear stress on z-plane in
x-direction
direction cosines of n in x, y, z directions
principal stresses at a point
f irst, second, third invariants of stress
normal stress on octahedral plane
shear stress on octahedral plane
normal stresses in radial, circumf erential, axial (polar)
direction .
spherical coordinates
shear stresses in polar coordinates
displacements in x, y, z directions
linear strains in x-direction, j-directiou, z-direction (with
non-linear terms)
linear strains (with linear terms only)
shear strain components (with non-linear terms)
shear strain components (with linear terms only)
rigid body rotations about x, y, z axes
cubical dilatation
principal strains at a point
f irst, second, third invariants of strain
;u- tyy. e ..
E X ) . s.; s;
Y X ) ' Y yz, Y zx
W" . WY' W
z
ll.""E a +l') .y+ E
r
:
" ~, ~
J" J
2
, J
3
Ust of Symbols
Ey. E
6
, e,
A , J . l
G = J .L
J .l .
E
K
p
1)
er y
U
U*
strains in radial, circumferential. axial directions
Lame's constants
rigidity modulus
engineering Poisson's ratio
modulus of elasticity
bulk modulus; stress intensity factor \
\
(Jul
er e.
au
blj
~. MY'.~
8
I X ' ~Y'I :
I p
/ xy ~'"
T
If'
a
Q
p
v
pressure
Poisson's ratio
yield point stress
elastic energy
distortion energy; complementary energy
ultimate stress in uniaxial tension
ultimate stress in uniaxial compression
influence coefficient; material constant
compliance component
moments about x, y, z axes
Linear deflection; generalized deflection
moments of inertia about x, y, z axes
polar moment of inertia
products of inertia about xy and yz coordinates
torque; temperature
warping function
coefficient of thermal expansion
lateral load
axial load
elastic potential
Poisson's ratio in r-direcrion due to stress inj-dire.ction
width
thickness
theoretical stress concentration factor
normal force
stream function
fller radius
radii
notch sensitivity
fracture toughness in mode /
offset yield stress
angular velocity
fracture resistance
fracture stress
boundary
f-integral
Vij
b, W
t
x,
N
f / >
P
D, d
q
K", ;
Sy
ill
R
G ;r
r
J
51 Units
,(Systeme International
,d/Unl~/es)
(a) Base Units
Q , u m t f t y ' : : . ' . . : PO Unit (Syf1iboQ
FI~~~~,,-~~.~ =~---=----~~-----4
length meter(m)
kilogram (kg)
second (s)
newton (N)
p a s c a l ( P a )
.
ma s s
.
time'
force
.
.
p r e s s u r e
force is a derived unit: kgm/S2
pressure is force per unit area: N/m
2
: kg/ms?
kilo-watt is work done per second: kNm/s
(b) Multiples
giga(G)
mega(M)
kilo(k)
milli(m)
micro(J t)
nano (n)
1000000000
1000000
] 000
0.001
0.000001
0.000 000 001
(c) Conversion Factors
. '
, -, i :.
: .. .,.... .
t o : -: -
To Cot nWt
..'_
. M, !lt ip/y by .
.'
kgf newton 9.8066
k g f / c m2
Pa 9.8066 x lif
,
k g f / c m2
kPa ' 98:066
newton kg( 0.10197
Pa
N/m2
1
kPa
k g f / c m
2
0.010197
HP leW 0.746
I W kNmls 0.746
kW kNmls 1
Typical Physical
Constants
(As an Aid to Solving Problems)
Material Ultimate Strength Yield Strength Elastic Poisson's CoefJ.
I
(NlPa) (MPa) Modulus Ratio Therl11
J
(GPa) iE:lpalls.

Tells. Comp Shear Tens or Shear Tells Shear per C

CO/liP
x 10 ~
Aluminium alloy 4 1 4 4 1 4 2 2 1 3 0 0 1 7 0 73 2& 0 . 3 3 4 2 3 . 2
Cast iron, gray 2 1 0 825 - - - S X J 4 1 0 . 2 1 1 1 0 . 4
Carbon steel ~ 690 5 5 2 4 1 5 250 2 0 0 83 0 . 2 9 2 1 1 . 7
S tainless steel 568 5 6 8 - 276 - 2 I J 7 S X J 0 . 2 9 1 1 7 . 0
For more accurate values refer to hand-books on material properties
,
Copyriqmed malerial
. .'
,
, '

I
I

..-
Analysis of Stress
. .,
,'INTRODUCTION'
1oeJI>A.I
Inthis book we shall deal with the mechanics of deformable solids. The starting
. point for discussion can beeither the analysis of stress or the analysis of strain. In
books on the theory of elasticity, one usually starts with the analysis of strain,
which deals with the geometry of deformation without considering the forces that
cause the deformation. However, one is more familiar with forces, though the
measurement of force is usually done through the measurement of deformations
caused by the force. Books on thestrength of materials, begin' with the analysis of
stress. The concept of stress has already been introduced in theelementry strength
of materials. When abar of uniform cross- section, say acircular rod of diameter
d, is subjected to a tensile force F along the axis of the bar, the- average stress
induced across any transverse section perpendicular to the axis of the bar and
away fromthe region of loading is given by
(J ' _ F _ _ 4 ; , , : . F
7
_ Area - 1!d2
It is assumed that the reader is familiar with the elementary flexural stress
and torsional stress concepts. In general, a structural member or a machine
element will not possess uniform geometry of shape or size, and the loads
acting on it will also be complex. For example, an automobile crankshaft or a
piston inside an engine cylinder or an aircraft wing are subject to loadings that
are both complex 'as well as dynamic in nature. In such cases, one will have to
introduce the concept of the state of stress at apoint and its analysis, which will
be the subject of discussion in this chapter. However, we shall not deal with
forces that vary with time. .
Itwill be assumed that the matter of the body that is being considered is
continuously distributed over its volume, so that if we consider asmall volume
element of the matter surrounding apoint and shrink this volume, in the limit we
shall not come across avoid. In reality, however, all materials are composed of- -
many discrete particles, which are often microscopic, and when an arbitrarily
selected volume element is shrunk, in the limit onemay end up in avoid. But in
our analysis, we assume that the matter is continuously distributed. Such abody
,
2 Advanced Mechanics of Solids
is called a continuous medium and the mechanics of such a body or bodies is
called continuum mechanics.
aonv FORCE, SURFACE FOR<;E
AND STRESS VECTOR
_ ..
Consider a body B occupying a region of space referred to a rectangular
coordinate system Oxyz, as shown in Fig. 1.1. In general, the body will be
subjected to two types of forces-
body forces and surface forces. The
body forces act on each volume ele-
ment of the body. Examples of this
kind of force are the gravitational
force, the inertia force and the mag-
netic force. The surface forces act
OD the surface or area elements of
the body. When the area considered
lies 00the actual boundary of the
body, the surface force distribution
is often termed surface traction. In
Fig. l.1, the surface forces F
I
, F
2
,
F3 ... F
r
, are concentrated forces,
Z . while p is a distributed force. The
Fig. 1.1 Body subjected toforces support reactions R
I
,R2 and R3 are
also surface forces. It is explicitly assumed that under the action of both body
forces and surface forces, the body is in equilibrium.
Let P be a point inside the body with coordinates (x, y. z). Let the body
be cut into t\VO parts C and D by a plane J -I passing through point P, as
shown in Fig. 1.2. If we consider the
free-body diagrams of C and D, tben
each part is in equilibrium under the
action of the externally applied forces
and the internally distributed forces
across the interface. In part D, let 6.A
be a small area surrounding the point
P. Inpart C, the corresponding area
at P' is l1A'. These two areas are distin-
guished by their outward drawn normals
I J ,
nand n . The action of part C on M at
point P can be represented by the force
I
vector l1T and the action of part Don M'
at P' canberepresented by theforcevector
l
l1T'. \Veassume that as l1A tends to zero,
I
theratio AT tends toadefinite limit, and
AA
p
y
/-----------......--~
o
x
F,
p
I
_ .-' " 1
I,
AT
c
I
AT
RI
Fig. 1.2 Free-body diagram of a
body cut into tlVO parts
Copyrlqhted matenal
Analysis of Stress 3
further, the moment of the forces acting on area L 1 A about any point within the
area vanishes inthe limit. The limiting vector is written as
I I
lim L 1 T =dT =T (1.1)
AA .....O 6A dA
Similarly, at point P', the action of part D on C as L 1 A ' tends to zero, can be
represented by a vector
1 1
lim f:.T' = dT' = .; ,
A_4'--+() L1A' dA'
(1.2)
I I
Vectors T and T' are called the stress vectors and they represent forces per
unit area acting respectively at P and P' on planes with outward drawn normals
J J I
nand n .
J
We further assume that stress vector T representing the action of Con D
J
at P is equal in magnitude and opposite in direction to stress vector T'
representing the action of D on C at corresponding point P'. This assump-
tion is similar to Newton's third law, which is applicable to particles. \Ve
thus. have
J 1
T=-T'
(1.3)
2
t.T
If the body in Fig. 1.1 is cut by a different
2
plane 2-2 with outward drawn normals nand
2
n' passing through the same point P, then
the stress vector representing the action
2
of C
2
on D2 will be represented by T
(Fig. (1.3, i.e.
2
2 L 1 T'
T= lim -
LlA-40 L 1 A
I
Ingeneral, stress vector T acting at point
1
P on a plane with outward drawn normal n
Fig. 1.3 Body cut by another plane 2
will be different from stress vector T acting
2
at the same point P, but on a plane with outward drawn normal 11. Hence the
stress at a point depends not only on the location of the point (identified by coordi-
nates x, y. z) but also on tbe plane passing through the point (identified by direc-
tion cosines 11... lipliz of the outward drawn normal).
.
4 . Advanced Mechanics of Solids
Since an infinite number of planes can be drawn through a point, we get an
infinite number of stress vectors acting at a given point, each stress vector
characterised by the corresponding plane on which it is acting. The totality of
all stress vectors acting on every possible plane passing through the point is
defined to be the state of stress at the point. It is the knowledge of this state of
stress that is of importance to a designer in determining the critical planes and
the respective critical stresses. It will be shown in Sec. 1.6 that if the stress
vectors acting on three mutually perpendicular planes passing through the point
are known, we can determine the stress vector acting on any other arbitrary
plane at that point.
. ; .NORMAL AND SHEAR STRESS C0MP0NENTS,
. . . .
". .
n
Let T be the resultant stress vector at point P acting on a plane whose outward
drawn normal is n (Fig. 1.4). This can be resolved into two components, one along
the normal n and the other perpendicular to n. The
component paraJ lel to n is called the normal stress
II
and is generally denoted by (J '. The component per-
pendicular to n is known as the tangential stress or
shear stress component and is denoted by ~. We
have, therefore, the relation:
n
(J '
n
T
n
(1.4)
where Ii - is the magnitude of the resultant stress.
II
Stress vector T can also be resolved into three
components parallel to the x, y, z axes. If these
Fig. 1.4 Resultant stress
vector, normal
and shear stress
components
n n n
components are denoted by T x, T y, T., we have
1
,,2 II II "
2 2 2
T =Tx +Ty +T
z
(1.5)
1.5, .RECTANGULAR STRESS COMPONENTS
Let the body B. shown in Fig. 1.1, be cut by a plane parallel to the yz plane. The
normal to this plane is parallel to the x axis and bence, the plane is called the x
x
plane. The resultant stress vector at P acting on this will be T. This vector can
be resolved into three components parallel to the x, y, z axes. The component
parallel to thex axis, being normal to the plane, will bedenoted by (J ',. (instead of by
x
0-). The components parallel to the y and z axes are shear stress components and
are denoted by 'fXY and Tnrespectively (Fig. 1.5).
Copyri ghted materi al
In the above designation, the first
subscript x indicates the plane on
which the stresses are acting and the
second subscript (y or z) indicates the
l 4 : : : : : : ; ; ; : : : : : - _ , , _ n direction of the component. For ex-
ax ample, '(~yis the stress component on
the x plane in y direction. Similarly,
'fxz is the stress component on the
x plane in z direction. To maintain
consistency, one should have denoted
thenormal stress component as 'f
cr
' This
would be the stress component on the
x plane in thex direction. However, to
distinguish between anormal stress and
a shear stress, the normal stress is denoted by C T and the shear stress by r,
At any point P, one can draw three mutually perpendicular planes, the x plane,
the y plane and the z plane. Following the notation mentioned above, the normal
and shear stress components on these planes are
y
- rxy
x
T
o
x
z
Fig. 1.S Stress components Oil x plane
C T.r' f,:., 'f,"! on x plane
~" ~'x'-r:v: on y plane
C T " 7":.<. 'Z)' on z plane
These components are shown acting on a 5111all rectangular element surround-
ing the point P in Fig. 1.6.
z
y
- ..
x
x
Fig. 1.6 Rectangular stress components
One should observe that the three visible faces of the rectangular element
have their outward drawn normals along the positive x, y and z axes respectively.
C onsequently, the positive stress components on these faces will also be directed
along the positive axes. The three hidden faces have their outward drawn normals
"6 Advanced Mechanics of Solids
in the negative x, y and z axes. The positive stress components on these faces
will, therefore, be directed along the negative axes. For example, the bottom face
has its outward drawn normal along the negative y axis. Hence, the positive stress
components on this face, i.e., 0:,,, ' f ) ' x and ' t ) ' z are directed respectively along the
negative y, x and z axes.
="'" STRESS COMPONENTS ON AN ARBITRARY PLANE
It was stated in Section 1.3that a knowledge of stress components acting on three
mutually perpendicular planes passing through a point will enable one to deter-
mine the stress components acting on any plane passing through that point. Let
the three mutually perpendicular planes be the x, y and z planes and let the
arbitrary plane be identified by its outward drawn normal n whose direction
cosines are n
x
' II) ' and n.
Consider asmall tetrahedron
at P with three of its faces
normal to the coordinate
axes, and the inclined face
having its normal parallel to
n. Let h be the perpendicu-
lar distance from P to the
inclined face. If the tetrahe-
dron is isolated from the
body and a free-body dia-
gram is drawn, then it will
be in equilibrium under the
action of the surface forces
and the body forces. The
free-body diagram is shown
in Fig. 1.7,
y
8
n
n
T"
A
x
c
z
Fig. 1.7 Tet rahedron at point P
Since the size of the tetrahedron considered is very small and in the limit as we
are going to make h tend to zero, we shall speak in terms of the average stresses
n
over the faces. Let T be the resultant stress vector on face ABC. This can be
II II II
resolved into components T, T', Tz, parallel to the three axes x, y and z, On the
three faces, the rectangular stress components are < T
x
' ".1)" "u' (Jy, ' f ) .., "v x' a; ""t
and "zy. If A is the area of the inclined face then
Area of BPC =projection of area ABC on the yz plane
=An
x
Area of CPA =projection of area ABC on the xz plane
=All
y
Area of APB =projection of area ABC on tbe xy plane
=AfI
z
Let the body force components inx, y and z directions berx, 1;. and h respectively,
per unit volume. The volume of the tetrahedron is equaJ to ~Air, where h is the
perpendicular distance from P to the inclined face. For equilibrium of the
Copyr ighled malerial
Analysis of Stress ' 7
tetrahedron, the sum of the forces in x, y and z directions must individually
vanish. Thus, for equilibrium in x direction
n 1
Tx
A
- C T.t An" - 1'),., An)' - 1'zxAn. +" 3 Ahrx = 0
C ancelling A,
n I
Tx =O:t nx +1')'" II)' +Tzx n
z
- 3 hrx
.Similarly, for equilibrium iny and z directions,
(1.6)
and
n I
Ty ='l'.,ll
x
+O:yn)' +1'.,1 '= - 3 hr"
n I
T= =1'dlx +'l',,J!)' +0:=,,=- 3 hr.
(1.7)
(1.8)
In the limit as h tends to zero, the oblique plane ABC will pass through
point P, and the average stress components acting on the faces will tend to their
respective values at point P acting on their corresponding planes. C onsequently,
one gets from equations (1.6)-( I.8)
n
r, = nx O:x +n)' ~vx +II, 'l'zx
n
Ty =II;: t:ty +II)' 0:)' +1 1z 1'z),
(1.9)
n
T, = nx t:tt +n)' 1 'yz+n, 0:.
Equation (1.9) is known as C auchy's stress formula. This equation shows that
the nine rectangular stress components at P will enable one to determine the stress
components on any arbitrary plane passing through point P. It will be shown in
Sec. 1.8 that among these nine rectangular stress components only six arc indepen-
dent. This is because 't'.." = 1'yX' 1'ry = 1"7 and 1'u = 1 ".a' This is known as the equaLity
of cross shears. Inanticipation of this result, one can write Eq. (1.9) as
n
1 1 =n.t 'fix +lIy 1 'iy +II, 1';, =~nj 'l'ij
J
where i andj can stand for x or y or z, and a, =1'u' 0:), =1'yy and 0:, =r
u
'
n .
If T is the resultant stress vector on plane ABC, we have
(1.10)
I
n2 ,,2 n
2
n
2
T =T." +T y +T z
If a; and 'f
n
are the normal and shear stress components, we have
(LIla)
(1.11b)
n
Since the normal stress is equal to the projection of T along the normal, it is
n n It
also equal to the sum of the projections of its components T
x
, 1'; , and T, along
n, Hence,
II 12 n
O:n =nxTx +II)'Ty +II
z
T.
(1.12a)
'8.Advanced Mechanics of Solids
It n u
Substituting for r., T y and T : : from Eq. (1.9)
CI"=n; o, +n; o, +n; CIt+2n,lly rxy +2nl'z 'fy z +2nllx T:u
(l.12b)
Equation (1.J I) can then be used to obtain the value of r,
. 1 = , _ . . $$ pe __ '_ .. __ ."" __ _...__ ....." ", , ~ _
Example 1 . 1 A rectangular steel bar having a cross-section 2 em x 3 em is
subjected to a tensile force of 6000 N (612.2 kg[). If the, axes are chosen as
shown in Fig. 1.8, determine the normal and shear stresses on a plane whose
normal has the following direction cosines:
(i) "x =n)'=J z .1 I
z
= 0
(ii)
(iii)
1
f1 =II=n=-
.Y Y x fj
y
F
n
T n
n
T
/1
/
,
,
I
,
,
,
I
I."- - r- , ,
, .
i..---j;
- ' / ,
x
(a) (b) (c)
/
z F
Fig. 1 . 8 Example 1 .1
,
Solution Area of section =2 x 3 =6 cnr', The average stress on this plane is
6000/6 -' J 000 Nfcm
2
. This is the normal stress Cl
v
' The other stress components
are zero.
(i) Using Eqs (1.9), (1.11b) and (1.12a)
n
T x =0, T
n _ 1000
Y - fi '
n
T ; =0
a =1000 =500 Nfcm
2
n 2
2 1" 12 2 2 4
T n =T - a; =250,000 N fcrn
T n =500 N/cln
2
(5] kgffcn12)
"9
Analysis of Stress: "
n
(ii) Tx =0,
n 1000
Ty=.fi'
U" =500 N'Icm
2
, and T" =500 N/cm
2
(51 kgf/cm
2
)
n
T, =0
n
(iii) r,=0,
n 1000
Ty=.fj'
n
Tz =0
Un =10~0 N/cm
2
Tn =817 N/cm
2
(83.4 kgflcm
z
)
i.III1' ..... 4~~_)...... adIiI'R!'1I"';M!I!_ I a. 5ll:,$ 'S! $$" lIIzS'li&O'oIl". 9"'1'7r~~
Example 1.2 At a point P in a body, O"; r. =10,000 Nlcrn
2
(1020 kgflcm
2
), 0" =
-5,000 Nlem
2
(-510 kg/len!2), O " z = -5,000 Nlcm
2
, 'r:r.y =fyz =1'n =10,000 Nle"m
2
.
Determine the normal and shearing stresses on aplane that is equally inclined
to all the three axes.
Solution A plane that isequally inclined toall thethree axes will have
I. 2 2 2 1
nx= ny;: n
z
;: .fjsmce nx +ny +n
z
=
From Eq. (1.12)
. 0"" =i[10000 - 5000- 5000 +20000 +20000 +20000]
=20000 N/cm
2
From Eqs (1.6}-{1.8)
Tx = ~ (10000 +10000+10000) = 10000 .fjN/cm
2
II I t:
Ty =.fj(10000 - 5000 +10000) =-5000 ....,3N/cm
2
11 J r:;
T, = .fj(10000 - 10000- 5000) =-5000 v3 N/cm
2
..
2
n
T =3 [(10
8
) +(25 X 10
6
) +(25 X 10
6
)] N
2
/cm
4
=450x 10
6
N
2
/cm
4
r2 =450 X 10
6
- 400 X 10
6
=50 X 10
6
N
2
/cm
4
n
or r, =7000 N/cm
2
(approximately)
~_iiEiIl~ n , __ a '$ 4 ua '" tlG.W!!!. ". ==.i1"... .,. di'S 'lL " ....~
Example 1.3 Figure 1.9 shows a cantilever beam in the form of a trapezium of
uniform thickness loaded by aforce P at the end. If it is assumed that the bending
stress on any vertical section of the beam is distributed according to the elementary
Copyrightedmaterial
'10 Advanced Mechanics of Solids
flex lireformula, show that the normal stress (J on a section perpendicular to
the top edge of the beam at point A is CT~ , where (J I is the flexural stress
cos f)
~c , as ShOlVII in Fig. 1.9(b).
A
0'1
~E
-
a
x ~
: : :
~
~
p
y
r-~Ar--~=j F
A
0'1
C
_jC
0'1
B
(b)
, .
B
(a)
-~1
(c)
Fig. 1.9 Example 1.3
(d)
Solution At point A, let axes x and y be chosen along and perpendicular to the
edge. On the x plane, i.e. the plane perpendicular to edge EF, the resultant stress
is along tbe normal (i.e., x axis). There is no shear stress on this plane since the "
top edge is a free surface (see Sec. 1.9). But on plane AB at point A there can
exist a shear stress. These are shown in Fig. 1.9(c) and (d). The normal to plane
AD makes an angle 9 with the x axis. Let the normal and shearing stresses on this
plane be (J I and 1'1'
We have
The direction cosines of the normal to plane AD are
IIx =cos f), n =0
z
The components of the stress vector acting on plane AD are
II
T; =CT
I
=IIx a, +lIy 't
yX
+liz 1'ty =(J cos9
n
1 '. " =/I" 1'X), +11)' CT. +II z 't"zy =0
n
T, =II" 't"xz +Ily 1'y" +flz (Jz =0
n n "
Therefore, the normal stress on plane AD =C T
n
=II x Tx +11)' T)' +liz Tz =
(J cos? 9.
'
Analysis of Stress '~"1'
Cfl Me
Cf= -
cos
2
(J Icos
2
(J
Further, theresultant stress on plane AB is
1,2 n n . , II
T =T;+T;+T;=~COS2(J
T2 = 0'2 COgle_ c r 2
"
=0'2 cos
2
e- (12 cos
4
e
. Hence
I .
T= - O'sln 26
2
_______ ..... .,_. _._._ ...... ... _'_.'_-.":. .... "#0 .,. ... _ mr '_',n__
or
: ' ; i . ~~ ) . 9IGRESSIO,N '.ON, IDEAL FLum
, .
By definition, an ideal fluid cannot sustain any shearing forces and the normal
force on any surface is compressive innature. This can be represented by
n
T=-pn, p>O
n n
The rectangular components of T are obtained by taking the projections of T
along thex, y and z axes. If n
x
, lI
y
and li z are the direction cosinesof n, then
n
Ty =-Plly'
n
Tz =-P"t
( 1.13)
Since all shear stress components are zero, one has from Eqs. (1.9),
(1.14)
Comparing Eqs (1.13) and (1.14)
o, =O' y =a, =-p
Since plane n was chosen arbitrarily, one concludes that the resultant stress
vector on any plane is normal and is equal to -p. This is t.hetype of stress that a
small sphere would experience when immersed in a liquid. Hence, the state of
stress at apoint wheretheresultant stress vector on any plane is normal to the
plane and has the same magnitude is known as abydrostatic or an isotropic state
of stress. The word isotropy means 'independent of orientation' or 'same in all
directions'. This aspect will bediscussed again in Sec. 1.14.
u;;,o;,;;..., EQUALITY OF CROSS SHEARS
We shall now show that of the ninerectangular stress components o. ; T
xy
' T
xz
' O' y,
T
yx
, T) ' ' Z 1 0' " 1':.<and T
zy
' only six are independent This is because ' l' xy =-ry,t, 1',,,, =T
zy
and 1'z .< =1' xz. These are known as cross-shears. Consider an infinitesimal rectan-
gular parallelpiped surrounding point P. Let the dimensions of the sides beI1x, l1y
andI1z (Fig. 1.10).
.l~Advanced Mechanics of Solids
z
,
Fig. 1.10 Stress components on a rectangular element
Since the element considered is small, we shall speak interms of average stresses
over the faces. The stress vectors acting on the faces are shown in the figure. On
the left x plane, the stress vectors are 'fu. t: x y and t: . . . . On the right face, the stresses
are 'rn' +A'fXl" 'fx y +A'fxy and Tn +A1",a. These changes are because the right face is
at adistance Ax from the left face.To the first order of approximation we have
.
. O r X l ' IJr x y i Jc
Ar....= t 7 X Ax , AT~y = O x Ax , Arn = 1 7 ; Ax
Similarly, the stress vectors on the top face are 'rY .l'+A'fY .l" " y x +A'r)!> and
'ry z+A-r".". where
IJr
AT = y z Ay
y z o y
On the rear and front faces, the components of stress vectors are respectively
.'f,.. 'rn' 'rzy
'f" +A 'fzz , 'fzx +A'r
zx
'fzy +A 'fzy
where
"' - IJr
V'ZX A ... = zy A_
A" r:r =~... O!.. AZ, Ll . zy ;:L u.e;
O Z V~
For equilibrium, the moments of the forces about the x, y and z axes must
vanish individually. Taking moments about the z axis, one gets
'f.u Ay Az ~ - ('rn +A'f.o) Ay Az ~ +
('rx y +A 'f; Cy ) Ay Az Ax - 'fY .l'Ax A.z Ax +
2
('fY .l'+A'fY .l') Ax Az ~'( - ,+A'ryx) Ax Az Ay +
Analysis of Stress :':l:t
.1. x . . 1 . y ( .1. x
Try Ax . 1 . y T-T",. ax . 1 . y T-T
zy
+.1.1'.),) Ax . 1 . y T+
( Tzr +.1. 1'" .y) . 1 . x. 1 . y . y ; : : : 0
Substituting for .1... .0-, .1.~l)' etc., and dividing by L l x: . 1 . y &
0. . 1 . u 0. xv 0.>" .1. lJ
~.t,,-X_2_ +r + . . Ax + / _'. I' _
o x 2 xy o x o y 2
Or),X A. _ t 7 c Zy L l x: +o r; ; x o y =0
y x - rl . . . ,-,y ~ 2 ., ?
-: (/Z o z_
Inthe limit as . 1 . x, . 1 . y and &tend to zero, ~eabove equation gives 1 'xy ; : : : Ty x'
Similarly, taking moments about the other two axes, we get 1 'y z; : : : T
zy
and Tz.<;::: 't
u
'
Thus, the cross shears are equal, and of the nine rectangular components, only six
are independent. The six independent rectangular stress components are O'x' O'y '
0' = , ';y, t'y z and t'zr.
1.9 A MOREGENERAL THEOREM
The fact that cross shears are equal can be used to prove a more general
theorem which states that if n and II' define two planes (not necessarily
orthogonal but in the limit
passing through the same point)
with corresponding stress vectors
T and T then the projection of
. ,
"
T along II' is equal to the pro-
n' n'
jecrion of T along n, i.e. T . n =
n'
~
/
/
/
/
/'
n'
T
tv
T' n
- "' - "- # "' n- - ""
----
T n'
f
/
n
. /T
,,' .
T . II (see Fig, 1:11).
Theproof is straightforward, If
, I d' th di ti
nx,l 1 y an n
z
are e tree on
cosines of n', then
" n n -'nr--
1 ' If' = T. n; +1 y n~ +T-. n;
From Eq. (1.9), substituting fOT
n n n
1 x, T; and 1, and regroup-
ing normal and shear stresses
Fig. 1.11 Pl anes with no rmal s nand n'
"
T
I , . I ,I , ,
. n = O'x n.r n x +O'y II)' n. Y +O'z l iz n z +1 '. t: v"x n y +": vx nl ' n x +
, , , ,
Ty z ny 1 1 z +1 'z), Il
z
n J ' +'f,X 1 1 , n x +t'xz Il x II z
Copy righted material
Using the result t'xy = 1'yx. 1',)'% ='Tty and t'zx =1'" ,
Similarly,
..
n' .
T
' , , (' ')
n =a, 11x 11 x +(Iy lIy 11y +(Iz liz 11z +1'x" "x 11" +II" 11 x +
t'y; ("" 11'. +liz lI'y) +1'z:c (II, n'x +11"11';)
Comparing the above two expressions, we observe
\
n n
T' n' =T' 11 (1.15)
Note: An important fact is that cross shears are equal. This can beused to prove
that a shear cannot cross a free boundary. For example, consider a beam of
rectangular cross-section as shown in Fig. 1.12.
..y
1/
L a
'-1-.-.-
L-- _ _ _ _ yll
x
(a) (b)
Fig. 1.12 (a) Element with free surface; (b) Cross shears being zero
If the top surface is a free boundary; then at point A, the vertical shear
stres component 1'xy =0 because if 'e
xy
were not zero, it would call for a
complementary shear 1'yx on the top surface. But as the top surface is ao
unloaded or a free surface, t'yx is zero and hence, t'x) ' is also zero (refer
Example 1.3).
PRINCIPAL STRESSES
. .
Wehave seen tbat the normal and shear stress components can be determined on
any plane with normal n, using Caucby's formula given by Eqs (1.9). From the
strength or failure considerations of materials, answers to the following questions
are important:
(i) Are there any planes passing through the given point on which the result-
ant stresses are wholly normal (in other words, the resultant stress vector
is along the normal)?
(ii) What isthe plane onwhich thenormal stress isamaximum and what is its
magnitude?
(iii) What is the plane on which the tangential or shear stress is a maximum
and what it is its magnitude?
Answers to these questions are very important inthe analysis of stress, and the
next few sections will deal with these. Let us assume that there is aplane n with
,
Anal
' fS ,"'4~1:"
ySlS 0 tress. 'iI'.,
direction cosines n
x
' ny and n
z
on which thestress iswholly normal. Let (J bethe
magnitude of this stress vector, Then wehave
"
T =(In
Thecomponents of this along thex, y and z axes are
(1.16)
n
r,=un
z
(1.11)
Also, fromCauchy's formula, i.e. Eqs(1.9),
It
T, =o, 11x +'t"xy ny +1'xz liz
"
Ty = 't'xy n" +(Jy lIy +'t')'% ".
It
T z =fyz nx +'t')'% lIy +(J. ".
Subtracting Eq. (1.17) from theaboveset of equations weget
(o; -:-Cf) "x +'t"xy lIy +f,u liz =0
't"xy"x +(Cf
y
- Cf) ",v +'t"J'ZII,=0
Tn n, +TY E "+(o,- (J) liz = 0
Wecan view theaboveset of equations asthreesimultaneous equations involv-
ing theunknowns "x' lI
y
and liz, Thesedirection cosines definetheplaneon which
theresultant stress is wholly normal. Equation (1.18)' is a set of homogeneous
equations: ,Thetrivial solution is"x = lIy =liz = O.For theexistence of a non-trivial
solution, thedeterminant of thecoefficients of 11", lI
y
and liz must beequal to zero, i.e,
(l.18)
(ax-a) 1'xy Txz
1"xy (O'y -0 ') 1")'%
=0
(1.19)
Txz 1')'% (O'z -0 ')
Expanding theabovedeterminant, onegets a cubic equation in (J as
(J" - (Jx +(Jy +(Jz)(J2 +(o; O'y +O'y (Jz +o; (Jx - T;y - 1";,. - .; x ) Cf-
(120)
Thethree roots of thecubic equation can bedesignated as Cf
l
,Cf
2
and Cf
3
It
will beshown subsequently that all thesethreeroots arereal. Weshall later give
a method (Example 4) to solvetheabovecubic equation. Substituting anyone of
thesethreesolutions in Eqs(1.18), wecan solvefor thecorresponding nr lIy and
liz. In order to avoid thetrivial solution, thecondition.
(1.21)
isused along with anytwo equations fromthesetof Eqs(1.18). Hence, with each
_,
(J there ",;11bean associated plane. These planes on each of which thestress
vector iswholly normal arecalled theprincipal planes, and thecorresponding
:1~' Advanced Mechanics of Solids
stresses, the principal stresses. Since the resultant stress is along the normal, the
tangential stress component on a principal plane is zero, and consequently, the
principal plane is also known as the shearless plane. The normal to a principal
plane is called the principal stress axis.
..
. .
.
~. ~
Thecoefficients of 0'2, o and the last termin the cubic Eq. (1.20) canbewritten as
follows:
(1.22)
(1.23)
cr
(1.24)
Equation (1.20) can then be written as
d'_/
1
0'2 +1
2
0'-/
3
= 0
The quantities II' 12 and 13 are known as the first, second and third invari-
ants of stress respectively. An invariant is one whose value does not change
when the frame of reference is changed. In other words if x', y', z', is
another frame of reference at the same point and with respect to this frame
of reference, the rectangular stress competence are 0". . . ,0" ,. . , O"z" 1'x ' y" '(y'"
and 'z'x" then the values of II' 12 and 1
3
, calculated as in Eqs (1.22) - (1.24),
will show that
.
i.e,
, , ,
O'x +O'y +O'z =O'x +O'y +O'z
II =Ii
andsimilarly,
Thereason for this can be explained as follows. Theprincipal stresses at apoint
depend only on the state of stress at that point and not on the frame of
reference describing the rectangular stress components. Hence, if xyz and
x'y'z' are two orthogonal frames of reference at the point.then the following
cubic equations
and
(13 -/
1
0'2 +1
2
0'-1
3
=0
0'3 -1'10'2 +1'20'-1'3 =0
Copyrlqhted matenal
Analysis of Stress : ': ; 1 '7 '
,.
must give the same solutions for <1. Since the two systems of axes were arbitrary,
the coefficients of <1
2
, and <1 and the constant terms inthe two equations must be
equal, i.e.
and
In terms of the principal stresses, the invariants are
I } =<11 +<12 +<13
1 2 = <11 <12 +<12<13 +<13<1}
1 3 =<1} <12<13
'J tc2 ;PR]NCIP AL .PLANES ARE ORTHOGONAL '
.
'. ,
,
,.
The principal planes corresponding to a given state of stress at a point can be
shown to be mutually orthogonal. To prove this, we make use of the general
theorem in Sec. 1.9. Let nand n' be the two principal planes and <1} and <1
2
, the
corresponding principal stresses. Then the projection of <11 in direction n' is
equal to the projection of <12 in direction n, i.e.
<1} n' , n =<1
2
n . n' (1 .2 5)
If nx , n; and n
z
are tbe direction cosines of n, and n'x' ,and n'. those of n', then
expanding Eq. (1.25)
a , (nxl l ; +l l y n; +1 1
z
n;)=a
2
(nt n; +ny n; +11;11;)
Since in general, <1, and <12 are not equal, the only way the above equation can
bold is
i.e. nand n' are perpendicular to eacb other, Similarly, considering two other
planes n' and n" on which the principal stresses <12 and 0:1 are acting, and
following the same argument as above, one finds that n' and n" are perpendicular
to each other. Similarly, nand n" are perpendicular to each other. Consequently,
the principal planes are mutually perpendicular.
a; '; ,'CUBIC EQUATION HAS THREE.REAL'ROOTS '
InSec. 1,10, it was stated that Eq. (1.20) has three real roots, The proof is as
follows, Dividing Eq. (1.20) by <1
2
,
1 2 1 3
a ! +---=0
1 a a2
For appropriate values of <1, the quantity on the left-hand side will be equal to
zero. For other values, the quantity will not be equal to zero and one can write the
above function as
(1.26)
SinceI}, ~ and 1 3 are finite.j'(o) canbemade positive for largepositive values of
<1. Similarly,/( (1) call bemadenegative for largenegativevalues of <1. Hence, if one
plots j{ 0') for different values of
(1as shown in Fig. 1.13, the curve
must cut the (1axis at least once as
shown by the dotted curve and for
(1 this value of (1,/(1) will beequal to
zero. Therefore, there is at least one
real root.
Let (13be this root and n the as-
sociated plane. Since the state of
stress at the point can .be
characterised by the six rectangular components referred to any orthogonal frame
of reference, let us choose aparticular one, x'y'z', where the z' axis is along nand
the other two axes, x' and y', are arbitrary. With reference to this system, the stress
matrix has the form,
:,_: Advanced Mechanics of Solids
f (a)
..,
B
Fig. 1.13 Plot all((1) versus (1
U:r' Tx'y'
0
'fX'),' U
y
'
0
0 0
u3
(1.27)
o
y'
x:
Fig. 1.14 Rectangular element
with faces normal
,,,
tox,y,z axes
Figure 1.14 shows these stress vectors
on a rectangular element, The shear stress
component's Tt,z' and Ty'z' are zero since the
z' plane is chosen to be the principal plane.
With reference to this system, Eq. (1.19)
becomes
(1x' - u) Tx'y'
0
~xy'
(U
Y
' - U)
0 =0 (1.28)
0 0
(13 - U)
Expanding
This is acubic in (1. One of the solutions is (1=0'3' The two other solutions are
obtained by solving the quadratic inside the brakets. The two solutions are
I
-
2
(1.29)
The quantity under the square root (power~) is never negative and hence,
(11and (12are also real. This means that the curve forj( 0') in Fig. 1.13 will cut the
(1axis at three points A, Band C in general. Inthe next section we shall study a
few particular cases.
Analysis of Stress :~tst
.
" . ~( ", ,. . _ .
~. 4': I~'
(i) If0'1' 0'2 and 0:1are distinct, i. e. 0'1' 0'2 and 0'3 have different values, then
the three associated principal axes n
1
, n
2
and n3 are unique and mutually
perpendicular. This follows from Eq. (1. 25) of Sec. 1. 12. Since 0'" 0'2 and
0'3 are distinct, we get three distinct axes n
J
, n
2
and n3 fromEqs (1. 18), and
being mutually perpendicular they are unique.
(it) If0'1 = 0'2 and 0'3 is distinct, the
axis of n3 is unique and every
direction perpendicular to "3 is
a principal direction associated
with 0'1 =0'2' This is shown in
Fig. 1. 15.
To prove this, let us choose a
frame of reference Ox'y'z' such
that the z' axis is along ") and
the x' and y' axes are arbitrary.
From Eq. (1. 29), if 0'1 =0'2' then the quantity under the radical must
be zero. Since this is the sum of two squared quantities, this can happen
only if
,
Fig. 1. 15 Case with 0'1 = 0'2
and 0'3 distinct
O'x =0'1 and 'C
xI
=0
But we have chosen x' and y' axes arbitrarily, and consequently the
above condition must be true for any frame of reference with the z' axis
along "3' Hence, the x' and y' planes are shearless planes, i. e. principal
planes. Therefore, every direction perpendicular to "3is aprincipal direc-
tion associated with 0'1 =0'2'
(iii) If 0'1 = 0'2 = 0'3' then every direction is a principal direction. This is
the hydrostatic or the isotropic state of stress and was discussed in
Sec. 1. 7. For proof, we can repeat the argument given in (ii). Choose a
coordinate system Ox'y'z' with the z' axis along "3corresponding to 0').
Since 0'1 =0'2 every direction perpendicular to n) is a principal direction.
Next, choose the z' axis parallel to "2 corresponding to (12' Then every
direction perpendicular to n2 is a principal direction since 0') = 0'3'
Similarly, if we choose the z' axis parallel to n
l
corresponding to 0'1'
every direction perpendicular to ", is also a principal direction. Conse-
quently, every direction is a principal direction.
Another proof could be in the manner described in Sec. 1. 7. Choosing Oxyz
coinciding with" l' n
2
and "3' the stress vector on any arbitrary plane n has value
0', the direction of 0' coinciding with n. Hence, every plane is a principal plane.
Such a state of stress is equivalent to a hydrostatic state of stress or an isotropic
state of stress.
. ~.
. The material discussed in the last few sections is very important and it is worth-
while to put it in the form of definitions and theorems.
'in' ,:Advanced Mechanics of Solids
Definition
n
For agiven stateof stress at point P, if theresultant stress vector T onany plane
n is along n having a magnitude (1, then (1 is aprincipal stress at P, n is the
principal direction associated with (1, the axis of (1 is aprincipal axis, and the
plane is aprincipal plane at P.
Theorem
Inevery state of stress there exist at least three mutually perpendicular principal
axes and at most three distinct principal' stresses. The principal stresses (11' (12
and u
3
arethe roots of thecubic equation
(13 -/
1
(12 +l-z(1-/
3
=0
where I I ' 12 and13 arethe first, second andthird invariants of stress. Theprincipal
directions associated with 0'1' (12and(13areobtained by substituting (1;(i =1,2,3)
inthefollowing equations and solving for n.ny and n
z
:
(a, - (1J n:t +1' x)' n)' +1' xz n
z
=0
1' ..,. nx +(1y - (11) " +1' yz n: =0
n; +n; +n; =I
If (11' (12 and (13 are distinct, then the axes of n
l
, n2 and n3 are unique and
mutually perpendicular. If, say (11= (12:;; cr" then the axis of n3 is unique and
every direction perpendicular to n3 is a principal direction associated with
(11= (12' If(11= (12= (13' then every direction isaprincipal direction.
Standard Method of Solution
Consider thecubic equation y3 +pi +qy +r =0, wherep, q andr areconstants.
1
Substitute y =x -- p
3
This gives ~ +ax +b =0
where a= ~ (3q_ p2 ), b=-f ;(2 p3_ 9pq+2 7r)
Put
Determine ~, and putting g =2 ~ a!3, thesolutions are
YI=g cos i :
3 3
Y2 =gcos(~ +120
0
) - ~
Y3 = g cos(~ +240
0
) - ~
Analysis of Stress :~~"
_ _ -----------.. . . . . . . . . _-----------,---,. . . . "' - _ . _ .
Example 1.4 At a point P, the rectangular stress components are
{1'x =1,(1' ), =- 2, {1' : =4, t:xy =2, r)lt =- 3, and rXl =1
all in units ofkl'a. Find the principal stresses and check for invariance.
Solution The given stress matrix is
1 2 1
['iij] =
2 - 2 - 3
1 - 3 4
From Eqs (1. 22}- {1. 24),
11=1-2+4=3
/2 =(- 2 - 4) +(- 8 - 9) +(4- 1)=- 20
1 3 =1(- 8- 9)- 2(8+3)+ 1(- - ti+2)=- 43
f(a) =a
3
- 3a
2
- 20a+ 43 =0
.
..
For this cubic, following the standard method,
y=o, p=-3, q=- 20, r=43
a =!(- 60 - 9) =- 23
3
b= J . . (- 54- 540+ 1161)=21
27
cos = - ( - )
(12i17 )1f2
=- 11940

The solutions are
.
0') =Yl =4. 25+1=5. 25 kPa
0'2 =Y2 =- 5. 2 +1=- 4. 2 kPa
a
3
=Y3 =0. 95 +1=1. 95kPa
Renaming such that 0' , >0' 2>0' 3 we have,
a
l
=5. 25kPa, 0'2=1. 95kPa, a3 =- 4. 2 kPa
The stress invariants are
11 =5. 25 +1. 95- 4. 2=3. 0
1 2 =(5. 25 x 1. 95)- (1. 95 x 4. 2) - (4. 2x 5. 25) =- 20
1 3 =- (5. 25 x 1. 95x 4. 2) =- 43
These agree with their earlier values.
Example I. S With respect to the frame of reference Oxyz, the following state of
stress exists. Determine the principal stresses and their associated directions. A/so,
check on the invariances of 1 1 '' 2' ' 3'
Copyrlqhted matenal
- >
~'" Advanced Mechanics of Solids
1 2 1
['t'ij] = 2 I I
1 I I
Solution For this state
1,=1+1+1=3
12=(1-4)+(1-1)+(1-1)=-3
1 3 =1(1-1) - 2(2- I) +1(2- 1)=-1
f(u) =U
3
-II U
2
+12u-1
3
=0
i.e., u
3
- 3ul - 3u +1=0
or (0'3 +1)-3u(u+ 1)=0
i.e., (o+1)(ul - a +I ) - 3u(u +I ) =0
or (u +1)(q2- 4u+I ) =0
Hence, one solution is U=-1. The other two solutions are obtained from the
solution of the quadratic equation, which are U=2 .fj.
:. ul=-I , U
2
=2+.fj, u
3
=2-.fj
Check on the invariance:
With the set of axes chosen along the principal axes, the stress matrix will have
the form
- 1 0
o 2+J 3
o 0
o
o
2-J3
I I =-1 +2+ .fj+2 - J 3 =3
12=(-2- .fj)+(4- 3) +(-2+.fj)=-3
1)=-1 (4-3)=-1
Directions of principal axes:
(i) For u
l
=-I , fromEqs (1.18) and (1.21)
Hence,
(1+I)n..+2ny +n
z
=0
2nx +(1+1)ny +n
z
=0
nx+ny+(1 +l)n:=O
together with
Fromthe second and third equations above, n, =O.Using this inthe third
and fourth equations and solving, Ilx = (1/ .J 2), ny = (I/ .J 2).
Hence, u
l
=-1 is in the direction (+I /.J 2,-I /.J 2,O).
I t should be noted that the plus and minus signs associated with n
x
, ny
and n
z
represent the same line.
Copyr ighted malerial
Analysis of Stress ',~"
(Ii) For 0'2=2+J 3

(-1- J 3 ) n . +2 1 1 y +n
z
=0
2 1 1 x +(-I - J 3 ) lIy +liz =0
n x + n y (-I - J 3)lI
z
=O
together with
Solving, we get
( )
1 /2
n =n = I+....!....
.r y J 3
I
11 =-_.::.__~
z (3+J 3 f / 2
(iii) For 0'3=2 - J 3
Wecan solve for n x ' n y and liz in amanner similar to the preceding oneor
get the solution fromthe condition that n
l
,n
2
and n 3 formaright-angled
triad, i.e. n 3 =n I x n
2
.
The solution is
. 1 /2
n
z
=g(1+)3)
- - - - - - - - - - - - - - - - - - _ . . - - - - - . - . ~ - - - - - - - - - - - - - - - - - - - - ~ ~ ~ - - - -
Example 1.6 For the given state of stress, determin e the prin cipal stresses an d their
direction s.
o 1 1
[fij] = 1 0 1
1 1 0
Solution
.

II =0, 1 2 = -3 ,1 3 = 2
f (G) =_G
3
+ 3 G +2=0
=(_0'3 - 1) +(3 0' +3)
=-(G+ 1 ) (G
2
-0'+ 1)+3(0'+ 1)
=(0' +I) (0' - 2) (0' +1) =0
0'1=0'2=-I and 0'3=2
Since two of the three principal stresses are equal, and G
J
is different, the axis of
0'3isunique and every direction perpendicular to O'J is aprincipal direction asso-
ciated with 0'1=0'2 ' For O:J =2
-2 n x +n y +n
z
=0
fl.. - 2 n y +": =0
''2;a: Advanced Mechanics of Solids
. .
n" +ny - 2n
z
=0
222
nx +"+n
z
= 1
Th
. I
ese give Il"= ny = fl.=.J 3
......... 7_~ .. _ _ , _ i > ._ % .._ u ,...._ , _ .... t.b_~.4. ... _ ............ $ ....... It ......
Example 1.7 The stale of stress at a point is such that
0'"=O'y= a, = 't'lJ ' = "t)" = 't'ZJ t = P
Determine the principal stresses and their directions
Solution For the given state,
I.=3p, 1 2 =0, 1 3 =0
Therefore the cubic is 0'3 - 3p0'2 = 0; the solutions are 0'. - 3p, 0'2
0'3 = O. For 0'. = 3p
(p - 3p)n" +pny +pn, = 0
pn" +(p- 3p) ny +pnz=O
pnx +pny +(p - 3p) n. = 0
or
The above equations give
-2n" +lIy +!Ir =0
nx - 2ny +n. =0
nx +ny - 21 1 . =0
WI
'th 2 2 2 I 11 '-3
nx +ny +n. = ,one gets nr =ny = liz = ,r~.
Thus, on a plane that is equaUy inclined toxyz axes, there is a tensile stress of
magnitue 3p. This is the case of a uniaxial tension, the axis of loading making
equal angles with the given xyz axes. If one denotes this loading axis by z', the
other two axes, x' and y', can be chosen arbitrarily, and the planes normal to
these, i.e. x' plane and y' plane, are stress free.
,.TEJ t::SlA.'FE:QF srasssREFEIUtED
:. .TOPIuNClPAL~S
In expressing the state of stress at a point by the six rectangular stress compo-
nents, we can choose the principal axes as the coordinate axes and refer the
rectangular stress components accordingly ..We then have for the stress matrix
0'. 0 0
[TijJ= 0 0'2 0
o 0 0'3
(1.30)
Copyri ghted materi al
Analysis of Stress <25
.
On any plane with normal n, the components of the stress vector are, from
Eq. (1.9),
(1.3J )
The resultant stress has a magnitutde
" 12 2 2 2 2 2 2
T =O'I"x +0'2"y +O'jllz
If(Y is the normal and " T the shearing stress on this plane, then
2 2 2
(Y =O'I"x +0'21ly +0'3"z
(1.32 )
(1.33)
and (1.34)
2 2 ( )2 2 2 ( )2 2 2 2
="xlly CT
I
- 0'2 +lIyll: CT2- 0'3 +":11.(0'3 - 0'1)
The stress invariants assume the form
II =(YI +(Y2 +(Y3
12 =(YI (Y2 +(Yl(Y3 +(Y3(YI
13 =(YI (Y2(Y3
(1.35)
MOHR'S CIRCLES FOR THE THREE-DIMENSIONAL
" . ' . . ~i~. . .
STATE OF STRESS .
l ,~
.
We shall now describe a geometrical construction that brings out some
important results. At a given point P, let the frame of reference Pxyz be
chosen along the principal stress axes. Consider a plane with normal n at
point P. Let (Y be the normal stress and r the shearing stress on this plane.
Take another set of axes rr and " T. In this plane we can mark a point Q with
co-ordinates (Y, r) representing the values of the normal and shearing
stress on the plane n, Fig. 1.16. For different planes passing through point
P, we get different values of (Y and T. Corresponding to each plane n, a
point Q can be located with coordinates (Y, r),The plane with the oaxis
and the t'axis is called the stress plane 1!. (No numerical value is associ-
ated with this symbol). The problem now is to determine the bounds for Q
(Y, r) for all possible directions n.
Arrange the principal stresses such that algebraically
(YI > (Y2 ~ (Y3
Mark off (YI' (Y2 and (Y3 along the (Y axis and construct three circles with
diameters (YI - (Yz), (Y2 - (Y3) and (YI - (Y3) as shown in Fig. 1.16.
Itwill be shown in Sec. 1.18 that the point Q( (Y, -r) for all possible n will lie
within the shaded area. This region is called Mohr's stress plane 1!and the three
circles are known as Mohr's circles. From Fig. 1.16, the following points can be
observed:
(i) Points A, Band C represent the three principal stresses and the associated
shear stresses are zero.
"26' Advanced Mechanics of Solids
Fig. 1.16 Mohr's stress plane
(ii) The maximum shear stress is equal to ~(0"] - 0"3) and the associated nor-
mal stress is ~(0"] +0"3)' This is indicated by point Don the outer circle.
(lit) J ust as there are three extremum values 0], 0"2 and 0"3 for the normal
stresses, there are three extremum values for the shear stresses, these
C T - C T 0'2- (j. (j.-(j.
being ] 2 3, 2 3 and J 2 2. The planes on which these shear
stresses act are called the principal shear planes. While the planes on
which the principal normal stresses act are free of shear stresses, the
principal shear planes are not free from normal stresses. The normal stresses
. d .h h .. I h . I 0'1+0'3 0'2 +0'3
associate Wit, t e pnncipa s ears are respective y 2 ' 2
and 0'1+0'2 . These are indicated by points D, E and F inFig. 1.16. It will be
2
shown in Sec. 1.19 that the principal shear planes are at 45 to the principal
normal planes. The principal shears are denoted by 1'1' 12and 1 ': 1 where
2r3 =(0'] -0'2), 21'2 =(0']-0'3)' 2r] =(0'2 -0'3) (1.36)
(iv)
circle and the shear stress on any plane will not exceed
~ (0"] - 0"2) according as 0"] =0"2 or 0"2 =0"3'
(v) When 0] = 0'2 = (13' the three circles collapse to a single point on the (1
axis and every plane is a shearless plane.
. MOHR'S STRESS PLANE
It was stated in the previous section that when points with coordinates (0", 1') for
all possible planes passing through a point are marked on the (1 - l'plane, as in
Fig. 1.16, the points are bounded by the three Mohr's circles. In this, section we
shall prove this.
Analysisof Stress '-'27
Choose the coordinate frameof reference Pxyz such that the axes are along the
"
principal axes. On any plane with normal n, the resultant stress vector T and the
DonnaI stress (J are such that fromEqs (1.32) and (1.33)
n 2
T (1.37)
(1.38)
and also (1.39)
The above three equations can be used to solve for n. ; . n; and n; yielding
"2 (0' - 0'2) (0' - 0'3) +r2
n =-'--;-----=:."--'-,-,--:..!---.,..-
x (0'1- 0'2)(0'1 - 0'3)
2 (0'- 0'3)(0'- 0'1) +?
ny = (0'2 - 0'3)( 0'2 - 0'1)
2 (0'- 0'1)(0'- 0'2)+ ?
1 1 =
Z (0'3 - 0'1)(0'3 - 0'2)
(1.40)
(1.41)
(1.42)
Since n; , n; and n; are all positive, the right-hand side expressions in the above
equations must all bepositive. Recall that we have arranged the principal stresses
such that 0'1~0'2 > 0'3' There are three cases one can consider.
Case (i) 0'1>0'2> 0'3
Case (ii) <11=0'2 >0'3
Case (iii) 0'1 = 0'2 = 0'3
We shall consider these cases individually.
Case (i) 0'1 > 0'2 > 0'3
For this case, the denominator in Eq. (1.40) is positive and hence, the numerator
must also be positive. In Eq. (1.41), the denominator being negative, the numera-
tor must also be negative. Similarly, the numerator in Eq. (1.42) must be posi-
tive. Therefore.
(0' - 0'2)( (J - (J3) +1'2~0
(0' - 0'3)(0'- 0'1) +'f,2S0
(0'- (1
1
)(<1- (1
2
)+1'2~0
The above three inequalities. can be rewritten as
According to the first of the above equations, the point (0', r) must lie on or
outside acircle of radius ~(0'2 - 0'3) with its centre at ~(0'2 +0'3) along the 0'axis
(Fig. 1.16). This is the circle with BC as diameter. The second equation indicates
that the point (0', .) must lie inside or on the circle ADC with radius .! .. (0'1 - 0'3)
2
and centre at ~(0'1 +0'3) on the 0'axis. Similarly, the last equation indicates that
the point (0', .) must lie00or outside the circle AFB with radius equal to .! .. (0'1 - 0'0
I 2
and centre at 2 (0'1 +O'~.
Hence, for this case, thepoint Q(0', r) should lieinside theshaded area of Fig. 1.l 6.
Case (ii) 0'1 = 0'2 >0'3
Following arguments similar to the ones given above, one has for this case from
Eqs (1.40)-(1.42)
From the rust two of these equations, since 0'1 =0'2' point (0', 1") must lie on the
circle with radius .! .. (0'1 - 0'3) with its centre at .! .. (0'1 +0'3)' The last equation
2 2
indicates that the point must lie outside a circle of zero radius (since 0'1 =0'2)'
.
Hence, in this case, the Mohr's circles will reduce to acircle BC and apoint circle
B. The point Q lies on the circle BEC.
Case (iii) 0'1 =0'2 =0'3
This is a trivial case since this is the isotropic or the hydrostatic state of stress.
Mohr's circles collapse to a single point on the 0'axis.
See Appendix I for the graphical determination of the normal and shear stresses
on an arbitrary plane, using Mohr's circles.
From Sec. 1.17 and also from Fig. 1.16 for the case 0'. ><12><1
3
, the maximum
shear stress is .!.(0'1 - 0'3) =1"2 and the associated normal stress is .!.(0'1 +0'3)'
2 2
Analysis of Stress {~'
Substituting these values in Eqs.(1.37)-( 1.39) inSec. 1.18, one gets nx =.J li2 ,
ny =0 and n. =l .fi. This means that the planes (there are two of them)
on which the shear stress takes on an extremum value, make angles of 45
and 135
0
with the (1) and <12 planes as shown in Fig. 1.17.
+-+-_Cf,
,
,
,
/1
)t- G ::\ :
, ,
(a)
Fig. 1.17 (a) Principal planes (b) Planes of maximum shear
If (1) = <1
2
><1
3
, then the three Mohr's circles reduce toone circle Be (Fig.I.I6)
and the maximum shear stress will be .! . (<1
2
- (1
3
) = 1') with the associated normal
2 .
stress ~(<1
2
+(13)' Substituting these values in Eqs (1.37)-(1.39), we get n.=0/0,
"i' = 0/0 and 11. = J .fi i.e. 11). and II-y are indeterminate. This means that the
planes on which 1') is acting makes angles of 45
0
and J 35
0
with the <13 axis
but remains .ndeterminate with respect to (1) and <12 axes. This is so be-
cause, since (1) =<1
2
:1; <1
3
, the axis of <13 is unique, whereas, every direction
perpendicular to (13 is a principal direction associated with (1) =<12 (Sec.
1.14). The principal shear plane will, therefore. make a fixed angle with <13
axis (45
0
or 135) but will have different values depending upon the selec-
tion of (1) and <13 axes .
. OCTAHEDRAL STRESSES
Let the frame of reference beagain chosen along (1). <12 and <13 axes. A plane that
is equally inclined tothese three axes iscalled
an octahedral plane. Such aplane will have "x =
"=liz. Since 11; +11; +fI ; =1 an octahedral
plane will be defined by IIx =Ily =liz = 1/..[3.
There are eight such planes, as shown in
Fig.LI8.
The normal and shearing stresses on these
planes arc caned the octahedral norma] stress
and octahedral shearing stress respectively.
Substituting "x =fly =II?; =11 .J3 in Eqs (1.33)
and (1.34),
Cf,
Cf2
Fig.I.IS Octahedral planes
Copyrighted materia)
:\0> Advanced Mechanics of Solids
or
O'OCI =~(0'1 +0'2 +0'3) =iII
T~ =~[(0'1 - 0'2)2 +(0'2 - 0'3}2 +(0'3 - 0'1)2]
9T:CL =2(0'1 +0'2 +0'3)2 - 6(0'10'2 +0'20'3 +0'30'1)
'f =.fi (1
2
_ 31)1/2
oct 3 I "2
(1.43)
and (1.44a)
or (1.44b)
(1.44c)
I t is important to remember that the octahedral planes are defined with respect
to the principal axes and not with reference to an arbitrary frame of reference.
Since O'OCI and "r
oct
have been expressed in terms of the stress invariants, one can
express these in terms of 0'..., O'Y ' 0;, " r x y > 1)..: and 'fzx also. Using Eqs (1.22) and (1.23),
(1.45)
9 T: C , =(o;- O 'y )2 +(O 'y - O'J 2 +(0': - O 'x )2 + 6(T.; y +T! +T; ) (1.46)
The octahedral normal stress being equal to 1/311, it may be interpreted as the
mean normal stress at a given point in a body. I f ina state of stress, the first
invariant (0'1 +0'2 +0'3) is zero, then the normal stresses on the octahedral planes
will be zero and only the shear stresses will act. This is important from the point
of view of the strength and failure of some materials (see Chapter 4).
_ _ _ _ _ _ ... '=e . ~~. _ _ _ ,. .. _ ... _ ... ........ 1 _ _ ,_',,_p'p __ ._
Example 1.8 The state of str ess at apoint is char acter ised by the components
0:. . =100MPa, O 'y =-40 MPa, o, =80MPa,
'fx y =fy z = 'tzx = 0
Deter mine the ex tr emum values of the shear str esses, their associated nor mal
str esses, the octahedr al shear str ess and its associated nor mal str ess.
Solution The given stress components are the principal stresses, since the
shears are zero. Arranging the terms such that 0'1 > 0'2 > 0'3'
0'1 =100MPa, 0'2 =80 MPa, 0'3 =-40 MPa
Hence fromEq. (1.36),
't
l
=u2 -u3 = 80+40 = 60MPa
2 2
'f3= UI-
U
2 ;100-80=10MPa
2 2
The associated normal stresses are
. ,
: ~Gf>yrightedmaterial
I
Analysis of Stress '-31'
, .
0'; = 0'3 +0'1 =-40 +100 = 30 MPa
2 2
------------"""' . . "" " " '''''_ -- ------------
~J THE STATE OF PURE SHEAR
The state of stress at a point can be characterised by the six rectangular stress
components referred to a coordinate frame of reference. The magnitudes of these
components depend on the choice of the coordinate system, If, for at least one
particular choice of the frame of reference, we find that < 1 x = < 1 y =<1. = 0, then astate
of pure shear is said to exist at point P. For such astate, with that particular choice
of coordinate system, the stress matrix will be
0
Txy Txz
[ Ti j J = Txy
0
Tyz
1';(2 Tyz
0
For this coordinate system, I) =<1x +< 1 y +<1. =O. Since IIis an invariant, this
must be true for any choice of coordinate system selected at P. Hence, the neces-
sary condition for a state of pure shear to exist is that /1 =0, It can be shown
(Appendix 2) that this is also a sufficient condition.
It was remarked in the previous section that when /1 =0, an octahedral plane
is subjected to pure shear with no normal stress. Hence, for a pure shear stress
state, the octahedral plane (remember that this plane is defined with respect to
the principal axes and not with respect to an arbitrary set of axes) is free from
normal stress.
, ~ . I1EC0MPO. SIT~ON INTO HXDROSTATIC.
AND'.PURE SJD;Aa STATES; ~
It will be shown in the present section tbat an arbitrary state of stress can be
resolved into a hydrostatic state and a state of pure shear. Let the given state
referred to a coordinate system be
0'.. Txy 1',..
[ Ti j ) =
T,\y O'y Tyz
l'
Tyz O'z
x z
Let p =1/ 3(< 1x +O'y +O'z) =1/ 3/
1
(1. 47)
The given state can be resolved into two different states, as shown:
ax 1 ' x y 1 ' x :
p 0 0
ax - p 1 ' x y 1 ' x z
't'xy
a
y
' t y z -0
P
0
+ ' t x y
a
y
- p
t y z
(1.48)
' T x z 1 ' y z a.
0 0
P T ; a T y z 0'% - P
The first state on the right-hand side of the above equation is a hydrostatic
state. [Refer Sec. t. I4(iii).]
The second state is a state of pure shear since the first invariant for this
state is
1'1 =(o, -p) +(a
y
-- p) +(<Tx- p)
=(1 +0: +<T_-3p
.{ y ...
=0 fromEq. (l,47)
If the given state isreferred to the principal axes, the decomposition into ahydro-
static state and a pure shear state can once again be done as above, i.e.
o o p 0 0
O-OpO+
a
3
0 0 P
o o
o (J .49)
where, as before, p =1/3(a
l
+a
2
-1- (
3
) =1/3 1
1
,
The pure shear state of stress isalso known as the deviatoric state of stress or
simply as stress deviator.
t.' , = ~"""""'. __ ~., _'_I o _ "",_ .. ,__ , _~._~, ~ .__ .
Example 1.9 The state a/stress characterised by 'r
jj
is given below. Resolve
the given state into a hydrostatic state and apure shear state. Determine the normal
and shearing stresses on all octahedral plane. Compare these with the <Toc, and 'rOCI
calculated for the hydrostatic and the pure s.tear states. Are the octahedral
planes for the given state, the hydrostatic state and the pure shear state the same 01'
are they different? Explain why.
[ 1 ' y ] = 4
6
10 4 6
2 8
8 6
Solution II =10+2 +6 =18,
!'I =6
3
Resolving into hydrostatic and pure shear states, Eq. (1.47),
6 0 0 4 4 6
[ 't'ij ) =
0 6 0 +
4 -4
8
0 0 6 6 8 0
Co pyr ighted malerial
Analysis of Stress .'33'
For the given state, the octahedral normal and shear stresses are:
(Joel =~II=6
FromEq. (1.44)
_.fi (2 2)1/2
T
oet
- "3 II - 3 /
2
=~ [18
2
- 3 (20- 16+12- 64+60- 36) J1I 2
=f(396)"2 =2..fi2
For the hydrostatic state, (Joel =6, since every plane is aprincipal plane with (J =6
and consequently, Toel =O.
For the pure shear state, (Joe! =0 since the first invariant of stress for the pure
shear state is zero. The value of the second invariant of stress for the pure shear
state is
1' 2 =(-16 - 16+0- 64+0- 36 ) =-132
Hence, the value of TOCI for the pure shear state is
't' = : : I i (396)112 =2 ..fi2
ocr 3
Hence, the value of (jOCI for the given state is equal to the value of (joel for the
hydrostatic state, and 1:0<:1 for the given state is equal to 'rOCI for the pure shear state.
The octahedral planes for the given state (which are identified after determining
the principal stress directions), the hydrostatic state and the pure shear state are
all identical. For the hydrostatic state, every direction is aprincipal direction, and
hence, the principal stress directions for the given state and the pure shear state
are identical. Therefore, the octahedral planes corresponding to the given state
and the pure shear state are identical.
Example 1.10 A cylindrical boiler, J 80 eln in diameter, is made ofplates J . 8em thick,
and is subjected to an internalpressure 1400 kPa. Determine the maximum shearing
stress ill the plate at point P and the plane
on which it acts.
pdl4t
Solution From elementary strength of
materials, the axial stress in the plate is
pd where p is the internal pressure, d
4t
the diameter and , the thickness. The cir-
curnferential or the hoop stress is I,d .
2t
The state of stress acting on an element
isas shown in Fig. 1.19.
pd/ 2t ~
Fig.l.19 Example 1.10
The principal stresses when arranged such that 0'( >0'2>0'3 are
pd pd
cr( =y,; 0'2 =4t; 0'3 =-p
The maximum shear stress is therefore,
'max = ~ ( O ' ( - c r 3 ) = ! P { ~ + 1 )
Substituting the values
=1400 (1.8x 100 1)=35 700 kP
ma.~ 2 2x 1.8 + ,. a
_ ..._ " . _ ....._ ,_ ............. _ . - ... r; ; c.o......_ ._ ._ . _ _ 7_ ... _ ; " ~ _ _ _ _ _ .,._ ....., 4'1'. _
1.f&'t CAUCHY'S STRESS QUADRIC
We shall now describe ageometrical description of the state of stress at apoint P.
Choose a frame of reference whose axes are aJ ong the principal axes. Let 0'1' 0'2
and 0'3 be the principal stresses. Consider a plane with normal n. The normal
stress on this plane is from Eq. (1.33),
Along the normal n to the plane, choose a point Q such that
PQ=R= 1/#
As different planes n are chosen at P, we get different values for the normal
stress 0' and correspondingly different P Qs. If such Qs are marked for every plane
passing through P, then we get a surface S. This surface determines the normal
component of stress on every plane passing through P. This surface is known as
the stress surface of Cauchy. This
has a very interesting property. Let
Q be a point on the surface,
Fig. 1.20(a). By theprevious defmi-
tion, the length P Q =R is such that
,
thenormal stress ontheplane whose
normal- is along P Q is given by
m
n
T
n
T
(b)
(a)
Fig. 1.20 ( a) Cauc hy' s str ess quadr ic
( b) Resultant str ess vec tor and
nor mal str ess c omponent
(1.50)
1
c r = - (1.51)
R2
If m is a norma] to the tangent
plane to the surface S at point Q,
then this normal m is parallel to the
n
resultant stress vectorT at P.
Since the direction of the result-
n
ant vector T isknown, and its COffi-
n
ponent 0' along the normal is known, the resultant stress vector T can be easily
determined, as shown in Fig. ] .20(b).
Analysis of Stress '35
n
We shall now show that the normal m to the surface S is parallel to T, the
resultant stress vector. Let Pxyz be the principal axes at P (Fig. l.21). n is the
normal to aparticular plane at P. The normal stress on tbis plane, as before, is
2 2 2
a =aln
x
+a2ny +a3n.
Y
I'- Q(x. Y. z)
-9
n
p
z
Fig. 1.21 Principal axes at P and
n to aplane
when (J is tensile
If the coordinates of the point Q are
(x, Y, z) and the length PQ =R, then
n =.!. (1.52)
Z R
Substituting these in the above equa-
tion for (J
(JR2 =(Jlx2 +(J21 +(J3z2
FromEq. (1.51), wehave (JR
2
= = I. The
plus sign is used when (J is tensile and the
minus sign is used when (J iscompressive.
Hence, tbe surface S has the equations
(a surface of second degree)
Substituting for n." flyand liz fromEq. (1.52)
x
(1.53a)
(1.53b)
We know from calculus that for a surface with equation F(x, y, z) =0, the
normal to the tangent plane at a point Q on the surface has direction cosines
proportional to :': and 0 : . FromFig. (1.20), mis thenormal perpendicular
to the tangent plane toS at Q. Hence, if m.., my, and "I, are the direction cosines of
m, then
o f
m=a-
r oz
From Eq. (l.53a) or Eq. (1.53b)
In, =2a(J.x, my=2a(J2Y'
where a is aconstant of proportionality.
n n n
T is the resultant stress vector on plane n and its components T x, T r and
n
Tz according toEq. (1.31), are
(1.54)
when (J is compressive
or
Substituting these inEq, (1.54)
n
m
x
=2aRTx,
n
nly =2a RTy,
n
m. =2a RTz
n n n
i.e. m... my and nlz are proportional to Tx, Ty and T z
n
Hence, m and T are parallel.
The stress surface of Cauchy, therefore, has the following properties:
(i) If Q is a point on the stress surface, then PQ =l/~ where (J is the
normal stress on aplane whose normal is PQ.
n
(it) The normal to the surface at Q is parallel to the resultant stress vector T
on the plane with normal PQ.
Therefore, the stress surface of Cauchy completely defines the state of stress
at P. It would be of interest to know the shape of the stress surface for different
states of stress. This aspect will be discussed in Appendix 3.
LAME'S ELLIPSOJ])
Let Pxyz be acoordinate frame of reference at point P, parallel to the principal
axes at P. On aplane passing through P with normal n, the resultant stress vector
n
is T and its components, according to Eq. (1.31), are
Let PQ be along the resultant stress vector and its length be equal to its
/I
magnitude, i.e. P Q=ITI. The coordinates (x, y, z) of the point Q are then
n
x=Tx,
II
y=T
y
,
II
Z =Tz
Sincen
2
+n
2
+n
2
=I, we get fromthe above two equations.
x y z
y
On
~T
Fig. 1.22 Lame's ellipsoid
(1.55)
This is the equation of an
ellipsoid referred totheprinci-
pal axes. Thisellipsoidiscalled
the stress ellipsoid or Lame's
ellipsoid. Oneof itsthreesemi-
axes is the longest, the otber
the shortest, and the third in-
between (Fig. 1.22). These are
theextermumvalues.
If two of the principal
stresses areequal, for instance
Copyrlqhted matenal
O "J =0"2' Lame's ellipsoid is an ellipsoid of revolution and the state of stress at a
given point is symmetrical with respect to the third principal axis Pz. If all the
principal stresses are equal, O "J = 0"2 = 0"3' Lame's ellipsoid becomes asphere.
Each radius vector PQ of the stress ellipsoid represents to acertain scale, the
resultant stress on one of the planes through the centre of the ellipsoid. It can be
shown (Example 1.11) that the stress represented by aradius vector of the stress
ellipsoid acts on the plane parallel to tangent plane to the surface called the
stress-director surface, defined by
(1.56)
The tangent plane tothe stress-director surface isdrawn at the point of intersec-
tion of the surface with the radius vector. Consequently, Lame's ellipsoid and the
stress-director surface together completely define the state of stress at point P.
____ ,... ........ ~_ ... _'.:roz'"','.n.r .... $e .... ~ .." _ ........'_..._._ ....
Example 1.11 Show that Lame 's ellipsoid and the stress-director surface together
completely define the state of stress at a point.
Solution If O "J ' 0"2 and 0"3 are the principal stresses at apoint P, the equation of
the ellipsoid referred to principal axes isgiven by
' ) 2 2
x- Y z
-+-+-=1
2 2 2
CT
J
C T 2 C T 3
The stress-director surface has the equation
, 22
X- +L+._= I
CTI C T Z C T 3
It is known from analytical geometry that for a surface defined by
Ftx; y, z) =0, thenormal tothetangent planeatapoint (x
o
,Yo, Zo ) hasdirectioncosines
proportional to 0;,: : ,: , evaluated at (xo, Yo, zo). Hence, at a point
(x
o
, Y o> zo) onthestressellipsoid, if m isthenormal tothetangentplane(Fig.1.23),then
Ellipsoid Surface
Stress-director
Surface
Fig. 1.23 Stress director sutface and ellipsoid surface
Consider aplane through P with normal parallel to m. On this plane, the resultant
"
stress vector will be T with components given by
Substituting for my my and In
z
m
Ty=ayo,
m
T =az o
i.e. the components of stress on the plane with normal m are proportional to the
coordinates (xo, Yo, zo). Hence thestress-director surface has the followingproperty.
Let L(xo, Yo, z o) be apoint on the stress-director surface. Let m be the normal to
the tangent plane at L. On aplane through P with normal m, the resultant stress
m
vector is T with components proportional to xo, Y o and z oo This means that the
m nJ m
components of PL are proportional to T, Ty and T.
n
PQ being an extension of PL and equal to T in magnitude, the plane having
this resultant stress will have m as its normal.
}l:u.THE PLANE STATE OF STRESS
If in a given state of stress, there exists a coordinate system Oxyz such that for
this system
o, =0, 'fu=0,
't: =
yz
(1.57)
then the state issaid to have a 'plane state of stress' parallel to thexy plane. This
state is also generally known as atwo-dimensional state of stress. All the forego-
ing discussions can be applied and the equations reduce to simpler forms as a
result of Eq. (1.57). The state of stress is shown in Fig. 1.24.
( a )
O'y
1':.y
0'1(
C 1
1''')'
C 1
I(
y
(b)
Fig. 1.24 (a) Plane state a/stress (b) C onventional representation
Consider a plane with the normal lying in the xy plane. If nx , " and n, are
the direction cosines of the normal, we have nx =cos e, n>, =sin e and n
z
=0
(Fig. 1.25).FromEq. (1.9)
Analysis of Stress ''''3~'
O'y
' f > y
,
n
,
,
,
'\~f!._
Ox
.- ,
,
,
,
,
,
,
,
' f > y
0'
n
0/
y
(a) (b)
Fig. 1.25 Normal and shear stress components on an oblique plane
n
T.. =C 1
x
cos 8 +'f,y sin 8
"
Ty =C 1
y
sin 8+fxy cos 8 {1.58}
n
T =0
The normal and shear stress components on this plane are from Eqs (1.11a)
and (l.llb)
C 1 =C 1
x
cos' 8 +C 1
y
sin
2
8 +2f<y sin 8cos 8
a +(j (j-(j
_ x y + x y cos 28+f.
y
sin 28
2 2 . .
(1.59)
and
or
(j -(j
r= x y sin 28+ 'fxy cos 28
2
(1.60)
The principal stresses are given by Eq, (1.29) as
1 /2
(1.61)
C 1 3 =0
The principal planes are given by
(i) the zplane on which C 1
3
=(j,=0and
(ii) two planes with normals in the xy plane such that
2r
tan 2= xy
(jx-(jy
The above equation gives two planes at right angles to each other.
If the principal stresses C 1
1
C 1
2
and C 1
3
are arranged such that C 1
1
~ C 1
2
~ C 1
3
, the
maximum shear stress at the point will be
(1.62)
(1.63a)
C opyrighted material
.
1 1 0 '- Advanced Mechanics of Solids
Inthe xy plane, the maximum shear stress will be
Tmax =1( 0'1 - 0"2 )
and from Eq. ( 1.61)
( 1.63b)
1 f2
( 1.64)
So far, attention bas been focussed on the state of stress at a point. I n general. the
state of stress in a body varies from point to point. One of the fundamental prob-
lems in a book of this kind is the determination of the state of stress at every point
or at any desired point in abody. One of the important sets of equations used in the
analyses of such problems deals with the conditions to be satisfied by the stress
components when they vary
from point to point. These con-
ditions will be established when
the body ( and, therefore. every
part of it) is inequillibrium, We
isolate a small element of the
body and derive the equations
of equilibrium from ito; free-
body diagram ( Fig. 1.26). A simi-
lar procedure was adopted in
Sec. 1.8 for establishing the
equality of cross shears.
Consider a small rectangu-
lar element with sides ax, l1y
and Az isolated from its parent
body. Since in the limit, we are
going to make ax, y and Az
tend to zero, we shall deal with
average values of the stress
components on each face.
These stress components are
shown in Fig. 1.27.
The faces aremarked as I., 2, 3
etc. Ontheleft hand face, i.e. face
No.1, the average stress com-
ponents are O"x, Txy and Txz. On
the right hand face, i.e. face
No.2, the average stress com-
ponents are
..... - -.. 1
,.- - ~I
I "
, I..'
'''- ''~..- - - - - ..
- - .....' ....
- -
, , 1"
(b)
(a)
Fig. 1.26 Isolated cubical element
in equilibrium
y
5/ ,
- - - - - 0'
I
r
1 /i"
Ox_-1----"'1-
,
,
2
x
z
Fig. 1.27
Oy
Variation ofstresses
iJr
T + .<Z I1x
xz ox
Analvsis of Stress . ., 1 '
o
This is because the right hand face is . 1 . x distance away fromthe left hand face.
Following a similar procedure, the stress components on the six faces of the
element are as follows:
00-. A
C T . + :) uZ,
uZ
Let the body force components per unit volume in the x , y and z directions
be r
x
' 1 ' )" and r,. For equilibrium in x direction
Face 1
C T x '
'" X Y ' '" x z
Face 2
00- OrX)'
C T
x
+ 0 ; ~x ,
'" x y + ox ~x ,
Face 3
C T y' ' "}, yz
Face 4
~
"
=,
or
yx
~y,
C T
y
+ oy ~y,
' "ft + oy
Face 5
fYz'
T
ZK
,
T z y
Face 6
Or
T + :cz ~x
.CZ ox
(
Or.r )
'rx + 0 ; ~Z ~x ~y- 'z x ~x ~y+Y x ~x ~y~z=0
Cancelling terms, dividing by . 1 . x , ~y,~ and going to the limit, we get
ro-x Oryx Orz .x 0
.... + "" + :) +Y x =
ox ay az
Similarly, equating forces in the y and z directions respectively to zero, we get
two more equations. On the basis of the fact that the cross shears are equal,
i.e.x y ='f
y
. , 'fyz ='f
z y
, 'fx z :='fz x ' we obtain the three differential equations of
equilibrium as
i}o- Or Or
x + X)' + n+y=O
ox oy oz x
i}o- iJr iJr
-:-y::. . . + x y of yz +y =0
oy ox oz y
ro- iJr Or, "
--::-~z + x z + ,. +y. =0
oz ox oy
Equations (1 .65) must be satisfied at all points throughout the volume of the
body. Itmust be recalled that the moment equilibrium conditions established the
equality of cross shears in Sec.I.S.
(1 .65)
: . a 2 ' Advanced Mechanics of Solids
I....
The plane stress has already been defined. Ifthere exists a plane stress state in
the xy plane, then C 1
z
=T z x =or)':=r. =0 and only C 1
x
C 1
y
, T x y , r.and Y y exist. The
diffemetial equations of equilibrium become
ro U r x y
0 : + oy +r, =0
r o y U r X } '
oy + ox +r y =O
(1.66)
Example 1.12 T he cr o ss-sectio n o f the wall o fa dam is sho wn in Fig. 1.28. T he
pr essur e o f water o nface OB is also sho wn. With the ax es Ox and Oy, as sho wn
in Fig. 1.28, the str esses at any po int (x , y ) ar e given by tr = specific weight
o f water and p =specific weight o f dam mater ial)
C 1
x
=-r y
C 1. - (p 2 r ) x +(r pJ y
J ' - tan f 3 - tan
3
f 3 tan
2
f 3 -
o
- - - ~x
A
"r ty =' f' y x =- ~ x
tan f 3
' t' y z =0, 1'zt =0, (1. =0
Fig. 1.28 Ex ample 1 .1 2
C heck ifthese stress components satisfy the differ-
ential equations of equilibrium. Also, verify ifthe
boundary conditions are satisfied on face DB.
and
So lutio n The equations of equilibrium are
ro t7r
-=-x "- + .<y +r =0
ox oy x
r o y Urx v
-=-~+ .+y =0
oy ox .v
Substituting and noting that Y x =0 and ')).=p, the first equation is satisifed. For
the second equation also
r _P_ r +p=O
tan
2
f 3 tan
2
f 3
On face DB, at any y , the stress components are (1. =-y y and T x y =O. Hence the
boundary conditions are also satisfied.
Example 1.13 C o nsider afunctio n (x , y ), which is called the str essfunctio n.

If the values o f (1x, (1y and Txy ar e as given belo w, sho w that these satisfy the
differ ential equatio ns o f equilibr ium in the absence o f bo dy fo r ces.
Solution Substituting in the differential equations of equilibrium
0
3
tP _ oJ tP =0
oy2 ox oy2 ox
Example 1.14 Consider the rectangular bean' shown in Fig. J. 29. A ccording to the
elementary theory of bending, the 'fibre stress' in the elastic range due to bending is
given by
My 12My
u =--=--7-
x I bh3
y
i - - I h. {2
~--- ---6
X Z I----+---t h/2
r-"-'-'-'-'-'-'-'-'-
Fig.l.29 Exmaple 1.14
where M is the bending moment which is a function of x. Assume that
(1, = = 1'% .1: ='fzy=0 and also that 'fxy =0 at the top and bottom. and further, that (1y=0
at the bottom. Using the differential equations of equilibrium, determine fAJ ' and (1y.
Compare these with the values given in the elementary strength of materials,
Solution FromEq. (L65)
or
ro- Or Or _
-=-,,-x xy xz 0
+ + =
ox oy O Z
Since 't'xz =0 and M is a function of x
_ 12y oM +O rAJ ' = 0
bll
3
ox dy
o.xy 12 oM
oy =bh
l
ox Y
I ntegrating
6 oM 2 j'()
- 3 ~ Y +C
1
x +c
2
bh C/X
Copyrighted malerial
,14' Advanced Mechanics of Solids

. . ' . ;
"
where f(x) is a function of x alone and c), c2are constants. It is given that
"" =0 at y =!!.
'.T.f 2
6 h
2
0}.1
~3 -4 ~ =-c1f(x) - c2
bl: o x
c f(x) +C =_ 3 oM
I 2 2bh ox
3 oM (4
y2
)
'f.t)' =2bh ox h2 - I
From elementary strength of materials, we have
or
V (h12 , ci A
'f.ty =lb J
y
Y
where V =- oM is the shear force. Simplifying the above expression
ox
or
1: = 3 oM (4
y2
-1)
x. v 2bh ox h
3
we get
i.e, the same as the expression obtained above.
From the next equilibrium equation, i.e. from
ib'y or xy or ~_
-::-'-+ + .' =0
oy ox O Z
3 (4
y2
) 02M
=- 2bh h2 - Iox2
..
3 02M (4
y3
)
O ' y =- 2bl 2 2 - Y +C
3
F(x) +c
4
' ox 3 h .
where F(x) is a function of x alone. It is given that O ' y =0 at y =-~.
Hence,
Substituting
Analysis of Stress '':45'
, .'
At Y =+h/2, the value of C J
y
is
Ia
2
M w
C J
y
=b t7X2 =b
where w is the intensity of loading. Since b is the width of the beam, the stress
will be wI b as obtained above.
'SA ctS#ilW ,:tCd!SS "'~f"au 4 :Mdll:S' 'Sf' ?,_~ ..q$".'I 'P d ..... $ . . . . . . 111 p t , m
BOUNDARY 'CONDI TI ONS
Equation (1.66) must be satisfied throughout the volume of the body. When the
stresses vary over the plate (i.e. the body having the pJ ane stress state), the
stress components C J
x
' C J
y
and 1'... y must be consistent with the externally applied
forces at a boundary point.
Consider the two-dimensional body shown in Fig.I .30. At aboundary point P,
the outward drawn normal is n, Let F, and F; be the components of the surface
forces per unit area at this point.
y
F
y
~
,
/n
F x
' f x y
~
G
y
(b)
1>-------"'(-_X
o
(a)
Fig. 1.30 (a ) Element nea r a bounda ry point (b) F ree body dia gra m
F. . and F y must be continuations of the stresses CJx , CJ )' and 't'xy at the boundary,
Hence, using Cauchy's equations
n
Tx =F x =C J xnx +1'xylly
/I
r, =Fy =0'),fly +'f,),IIx
Ifthe boundary of the plate happens to be parallel to y axis, as at point PI' the
boundary conditions become
P; =o, and F y =<xy
:EQUt\.l'I QNS, -Of. EQUI Lm~ I N' ~~~RI ~AL
.' ' COORBI NATES ,.,,', , ,
.
Till this section, we have been using a rectangular or the Cartesian frame of
reference for analyses, Such a frame of reference is useful if the body under
analysis happens to possess rectangular or straight boundaries. Numerous problems
exist where the bodies under discussion possess radial symmetry; for example, a
thick cylinder subjected to internal or external pressure. For the analysis of such
problems, it is more convenient to use polar or cylindrical coordinates. Inthis
section, we shall develop some equations in cylindrical coordinates .
. Consider an axisymmetric body as shown in Fig. 1.31(a). The axis of the
body is usually taken as the z axis. The two other coordinates are r and 8,
where 8is measured counter-clockwise. The rectangular stress components at
a point per, B, z) are
z
,
,
........ -_ ..-:-;--'---. . .
6.'" p
(a)
Fig. 1.31
These are shown acting on
the faces of a radial element
at point P in Fig.l.31(b).
C1" C1
8
and C1. are called the
radial, circumferential and
axial stresses respectively. If
the stresses vary from point
to point, one can derive the
appropriate differential equa-
tions of equilibrium, as in
Sec. 1.26. For this purpose,
consider a cylindrical ele-
ment having a radial length
lY with an included angle l1B
and aheight llz, isolated from
the body. The free-body dia-
gram of the element is shown
in Fig.l.32(b). Since tbe ele-
ment is very small, we work
with the average stresses
acting on each face.
The area of the faceaa' d' d
isr l1Bllz and thearea of face
bb' c' c is (r + lY) 118 llz. The
areas of faces dec'd' andabb' e'
areeach equal to lY llz.
The faces abed and a'b' c' d' have each an area ( r +~) l1BlY. The average
stresses on these faces (which are assumed to be acting at the mid point of eace
face) are
On face add' d
(b)
( a) Cylindrical coordinates of a point
( b) Stresses on an element
normal stress C1
r
tangential stresses 'f"and. 'frO
On face bb' c' c
normal stress C1
T
+
ib',lY
or
Analysis of Stress ~
z
y
c'
~ - - - - - - - - - - - - - - ~ y
x
(a)
x
tl)
Fig. 1.32 (a) Geometry of cylindrical element (b) Variation of stresses IU'7VSS faces
tangential stresses Trz ., .~ I1r
and
The changes are because the face bb' c'c isAr distance away fromthe faceaa' d' d.
On face dcc'd'
normal stress G
e
tangential stresses rre and 'fez
On face abb' a
&-
normal stress G
e
+ c : l!:..(}
tangential stresses 'f,e + ~ 110 and 'fez +
The changes in the above components are because the face abb' a is separated by
an angle A6fromthe face dcc'd',
On facea'b'c'd'
normal stress o,
tangential stresses 1'17 and foz
On face abed
normal stress o, +
. I t 7 r
tangenna stresses r+ 1%
rz OZ
Az and 'rOt +
.
Let Y r, Y o and y:: be the body force components per unit volume. If the element is
in equilibrium, the sumof forces in r, 6andz directions must vanish individuaUy,
. Equating the forces in r direction to zero,
( C T
r
+~~ Ar) (r +Ar) A(}A:z +(! n +~ AZ) (r +~) ll(}llr
You have either reached a page that is unavai lable for vi ewi ng or reached your vi ewi ng Iimit for thi s
book.
book.
S O '" Advanced Mechanics of Solids

passing through the point. Tbe equations are
n
M" =u;n" +My" ny +u;n:
n
M
z
=u;n.
t
+My: " +Mu n:
n n n R
Mx, My, M
z
are the x, y and z components of the vector M acting on
plane n,
12 A rectangular beam is subjected to apure bending momentM. The cross-
section of the beam is shown in Fig. 1.34. Using the elementary flexure
formula, determine the normal and shearing stresses at apoint (x, y) on the
plane AB shown.
y
M
A
"-
'" M
~.-.-.~~~.-.-.-.-.=.- ..j - x
-,
"8
Fig. 1.34 Problem 1.2
O J
I~).1
b
[ Ans.
13 Consider a sphere of radius R subjected to diametral compression
(Fig. 1.35). Let o; ulJand u; be the normal stresses and 1'r8' 1'~and. 1'90'the
shear stresses at apoint. At point P(o, y, z) on the surface and lying in the
yz plane, determine the rectangular normal stress components U'" u
y
and u.
in terms of the spherical stress components.
[Ans. U" =Ue; U
y
= U; cos
2
4'; a, =U; sin
2
4']
t
Fig.l.35 Problem 1.3
book.
book.
book.
_______ ~... tl ' _ ..._" __.._,,_ . r _ . . . . . . . . . , . - _ _ ai J " " ': ; _I . '_"" __ i f ; l! li "'~"'~_._1 ._U _
Appendix 1
Mohr's Circles
,
It was stated i n Sec. 1. 17that when poi nts wi th coor di nates (a, or )for all possi ble
planes passi ng thr ough a poi nt ar e mar ked on the a-or plane, as i n Fi g. 1. 16, the
poi nts ar e bounded by the thr ee Mohr 's ci r cles. The same equati ons can be used
to deter mi ne gr aphi caUy the nor mal and shear i ng str esses on any plane wi th
nor mal II. Equati ons (1. 40}- {I . 42) of Sec. 1. ISar e
2 ( 0" - 0 " 2 ) (0" - 0" 3) +'t" 2
n =~-~-.:..,....,----"'; _'..,.....-
x ( 0 " 1 - 0" 2)(O" J -0 " 3 )
(Al. I)
2 (0" - 0 " 3 ) (0" - O" t) +'t" 2
n =-'-:--....; ; ..:-.:..,....,--__; _,:..._~
y ( 0 " 2 - 0 " 3 ) ( 0 " 2 - O " t )
2' ( 0 " - 0" 1) (0" - 0" 2) +'t" 2
n =~- . . . . : . . . : . . . . !. : - - ;- - =. , : _ ~
Z ( 0 " 3 - 0 " 1 ) ( 0 " 3 - a2)
(Al. 2)
(AI 3)
For the above equati ons, the pr i nci pal axes coi nci de wi th the coor di nate
axes x, y and z. Constr uct a spher e of uni t r adi us wi th P as the centr e. PI' P2
and P
3
ar e the poles of thi s spher e (Fi g. AI. l). Consi der a poi nt N on the
sur face of the spher e. The r adi us vector PN makes angles a, p and r . r espec-
ti vely wi th the x, y and z axes. A plane thr ough P wi th PN as nor mal wi ll be
par allel to a tangent plane at N to the uni t spher e. If " x' fly and n
z
ar e the
di r ecti on cosi nes of the nor mal II to such a plane thr ough P, then nx =cos a,
ny =cos p, n: =cos r
n
o r
i
I
z
Fi g. AI. I Mohr's cirdesfor t hree-dimensionalst at e of st ress
Co pyri gtlted materi al
book.
book.
book.
.58 Advanced Mechanics of Solids
z '
y'
Now one has to show that there are at least
three mutually perpendicular generators of the
above cone. Let QI (x;, y; , I) be a point on the
cone and let S be a plane passing through P
and perpendicular to PQI' We have to show
that the plane Sintersects the cone along PQ2
and PQ3 and that these two are perpendicular
to each other.
Let Q(x', y', I) be a point in S. Then, S being
perpendicular to PQI' PQ is perpendicular to
PQI' i.e.
x'
Fig. A2.3 Cone with vertex
at P and axis
along PZ'
(A2.5)
If Q lies on the elliptic cone also, it must satisfy
Eq. (A2.4), i.e.
(A 2.6)
Multiply Eq. (A2.6) by 2y;2 and substitute for y'y; from Eq. (A2.S). This gives
a
1
x,2 y; 2 +a2(x; x' +1)2 - (a, +(2) y; 2 =0
or (A 2.7)
Similarly, multiplying Eq. (A2.6) by x[2 and substituting for x'x' , from
Eq. (A2.S), we get
(A2.8)
If Q(x', y: , I) is apoint lying inS as well as on the cone, then it must satisfy
Eqs (A2.S) and (A2.6) or equivalently Eqs (A2.7) and (A2.8). One can solve
Eq. (A2.7) for x' and Eq. (A2.8) for y'. Since these are quadratic, we get two
solutions for each. Let (xi, yi) and (xi, yj) be the solutions. Clearly
, 'f' [a2 -(a, +az)y: 2J
x2-~ = [ ,2 '2J
a2x, +a, y,
. (A 2.9)
(A 2.10)
Adding the above two equations
Copyrlqhted matenal
book.
book.
book.
~ Advanced Mechanics of Solida
This is the equation of an ellipse in xy plane. Hence, S =S, is an elliptic
cylinder.
inparticular, if 0'1=~, the elliptic cylinder becomes acircular cylinder.
(n) If0'1>0 and 0'2<0, then the equation becomes
,
O'I~-10'
2
1y2 =I
. or ~/(1/0'1)2 -y(l/O'i) =l
This describes conjugate hyperbolas in the xy plane. S, is given by a hyper-
bolic cylinder, the cross-sectional vertices of which lie on the nx axis and Sc is
given by a hyperbolic cylinder with its cross-sectional vertices lying on
the" axis.
Case (iii) If 0'1"# 0, 0'2:= 0'3=0, Eq. (1.53) reduces to
0'1Xl =1
When 0'1>0, this becomes
~=1(0')
f
or 1
This represents two straight lines parallel to the y axis and equidistant from it.
Hence, S =S, is given by two parallel planes, each perpendicular to nx and.equi-
distant from P.
\
book.
book.
book.
~iA 7Advanced Mechanics of Solids
.
I
The coordinates of P, P' and Q are
P: (x, y, z)
P': (x +u
x
' y +u
y
' z +u
z
)
Q: (x+~,y+~y,z+Az)
The displacement components at Q dif-
fer slightly from those at P since Q is away
from P by ~, ~y and Az. Consequently,
the displacements at Q are,
y
.' Q'
f P' ~
~r i
\ , I
\. pl,.. ". Q
~~ _ ...
~......_ ~...-...-.,..... -
\
\
\
.
/
~----------------~x
z
Fig.2.3 Displacements of two
neighbouring points
PandQ
u.. +~ux, u
y
+~UY' U
z
+au
z
If Q is very close to P, then to first-order

. approximation
OU ou ou
~u = x ~x + x ~y+ :r Az
x ox oy OZ
The first term on the right-band side is the rate of increase of u.. in x direction
multiplied by the distance traversed, Ax. The second term is the rate of increase of
U
x
iny direction multiplied by the distance traversed iny direction, i.e. ~y. Simi-
larly, we can also interpret the third term. For ~Uy and ~uz too, we have
(2.7a)
OU ou OU
~U = y ~x + y ~y+ y Az
y ox oy OZ
A _ ou. A ou. A OU: A _
i lU
z
- _;3 ilX+ ~ ily + ~ L>. <:
__ . _(/x (/y (/Z
(2.7b)
(2.7c)
.
Therefore, the coordinates of Q' are,
Q' =(x +~x +U:r +~x, Y +~y+u
y
+~ z +Az +'li z+~=) (2.8)
Before deformation, the segment PQ bad components ~x, ~y and ~z along the
three axes. After deformation, the segment P' Q' has components ~x+~U.T' ~y +
~Y' A z +~uz along the three axes. Terms like,
OU" ou.. oU
x
t
ox ' oy , oz ,e c.
are important in the analysis of strain. These are the gradients of the displacement
components (at a point P) in x, y and z directions. One can represent these in the
form of amatrix called the displacement-gradient matrix as
OU
x
oU
x
oU
x
ox oy oz
[ ~Ui ] =
oU
y
oU
y
oU
y
ox oy oz ax,
J
,
oU
z
o~z
ou
z
ox oy oz
book.
book.
book.
'. .
:1~' Advanced Mechanics of Solids
O U 8u
r,= x+ Y
xy oy ox'
O U r O U ,
r,= ~+ -
XZ i}z ox
(2.19)
EpQ"" EpQ =Exxn.; +E)')'II; +s.n; +Exl'n"n
y
+E)'2".vllz +E",lIxllz (2.20)
Equation 2.20 directly gives the linear strain at point P in the direction PQ with
direction cosines /lx, ny , liz. When II)C:::; 1, II)' :::; n
z
:::; 0, the lineelement PQ isparallel
to thex axis and the linear strain is
~
ou
E ""E:::; .l
x xx o x
Similarly, and
E '" :::; oU
z
z sz o z
are the linear strains iny and z directions respectively. In the subsequent analy-
ses, we will use only the linear terms in strain components and neglect squares
and products of strain components. The relations expressed by
Eqs (2.18) and (2.19) are known as the strairi displacement relations of Cauchy.
1Ij-RECTANGULAR' STR.AJ N'.'COMPONENTS
. ,
- '.'
E
XX
) Ew and u are the linear strains at apoint in x, y and z directions. It will be
shown later that Yxy , Yy z and Yx: represent shear strains in xy , y z and xz planes
respectively. Analogous to the rectangular stress components, these six strain
components are called the rectangular strain components at a point.
, _ .....
, .
; !'...-,
Knowing the six rectangular strain components at apoint P, one can calculate the
linear strain in any direction PQ, using Eq. (2.20). The tota1ityof all linear strains
in every possible direction PQ defines the state of strain at point P., This defini-
tion is similar to that of the state of stress at apoint. Since all that is required to
determine the state of strain are the six rectangular strain components, these six
components are said to define the state of strain at a point. We can write this as
y r
, x)' X2
(221) [ S i j ] :::; r.T)' e yy r}'Z
rx:z r)'2 e
zz
To maintain consistency, we could have written
e X')' ='Yx> ~ )'2 :::;rvz, ;<2 =Y .c
but as it is customary to represent the shear strain by r.we haveretained this
notation. In the theory of elasticity, 1 /2y xy is written as e xy , i.e.
,
1 _ 1( O llx O U
y
)_
'2KJ '- 2 oy+ ox -eT)'
(2.22)
If we follow the above notation and use
Copy righied malerial
book.
book.
~. . ".
Analysis of Strain ';~.,..'
__ ",. ....:MM....clXJi .. 'S"~""__ "'_;_''''': t..... ' : "' ....... _, _, _, _
Example 2.4 Consider the displacement field
u = [Ii +3yzj +(4+W)k] 10-
2
What are the rectangular strain components at the point P(/, 0, 2)? Use
only linear terms.
Solution
U
x
=y2 . 10-
2
u
y
= 3yz 10-
2
U
z
=(4+6x
2
) , 10.-
2
OU
=0
OU
y
=0 0; ; =12x 10-
2 x
ox ox
oU
x
= 2y 10-
2
OU)' =3z. 10-2 ou. = 0
oy
oy
oy
ou"
=0
OUy
=3y10-
2
ou.
=0
oz
oz
~
oz
The linear strains at (I, 0, 2) are
OU
e = x =0
xx ox '
=6x 10-
2
,
e =lJuz
zz oz
=0
The shear strains at (I, 0. 2) are
ou oU
y
r, = x +-: : : -"- =0+0=0
xy lJy ox
du ; lJu
r. = + z
yz O Z lJy
r, =du, + _ o=-u, -z =0+12X 10-
2
=12X 10-
2
xz IJz ox
=0+0=0
8I1IIM~Atn s $ : "(~i ~__ "_-"".,a .,.~... _......3SM$Ml
1lI: '.CBANGE IN~DIRECTION OF A LINEAR ELEMEN'f.,~,"
,
I It is easy to calculate the change in the orientation of a linear element resulting
, fromthe deformation of the solid body. Let PQ be the element of length ill,with
direction cosines n
x
, ny and nz' After deformation, the element becomes plQ' of
length ill',with direction cosines n'x' n~ and n~. If u" UY' U
z
are the displacement
components of point P, then the displacement components of point Q are.
U'x+AUX> u
y
+AU), ' Uz +Au.
where AUx.' I! ! .1l
y
and Au, are given by Eq. (2.7a)-(2. 7c).
From Eq. (2.12), remembering that in the linear range EpQ =EpQ>
/),s' =As (I+EpQ)
The coordinates of P, Q. P' and Q' are as follows:
P: (x, y, z)
Q: (x +Ax, Y +Ay, Z +AZ)
P: (x +u
x
' y +u
y
, Z +u
l
)
g: (x +Ax +"x +AU
x
, y +Ay +u
y
+AU
y
, z +AZ +U
z
+AU, )'
(2.28)
it.~AdvancedMechanics of Solids
Hence,
n = flx n = fly, n = Ilz
x Ss ' Y & : &
I flx +A U
x
, _ 6.y+A u)' _ Az +flll:
nx = &' , ny - &' , nz - &'
Substituting, for &'from Eq. (2.28) and for .llu
x
> flu
y
, .llu. from Eq. (2.7a)-(2.7c)
(2.29)
, 1 [OU: ou. (1 OU
z
) ]
nz =Ion" + 0 n)' + +..., n.
+&PQ x y oz
The value of EpQ is obtained using Eq. (2.20).
~CUBICALDR.ATAlfION ..: .'

.

Consider a point A with coordinates (x, y, z) and a neighbouring point B with
coordinates (x +llx,y +fly, Z +llz). After deformation, the points A and B move to
A' and B' with coordinates
A' : (x +U
x
' Y +u
y
' z +u.)
B' : (x +llx +U
x
+flu", y +fly +lIy +fllly. Z +flz +u; +flu
t
)
where u., " and u, are the components of diplacements of point A , and from Eqs
(2.7a)-(2.7c)
OU OU OU
flu = x flx + " fly + " flz
. ox oy OZ
_ oU
y
oU
y
oU
y
flu
y
- ox flx + oy fly + OZ flz
.. OU_" du, du, ,,_
aU . = aX+ - fly + - u.<.
s>: oy OZ
The displaced segement A'B' will have the following components along the x, y
and z axes:
X axis: llx +flu =( I OU,,) ,,_ Oil".. OU" A_
X + ~ L.U+0 ,-.y+ .;) =
ax y C/Z
au (iJ u. 1 81 1
y axis: fly +flu
y
=-::-~y llx + 1+ ...y) fly + y l!z
8x oy 8z
z axis: Ilz +flu, =_O \ ..".I ~z llx +O ilZ fly +(J t- O il.) Ilz
Ox oy Qz
(2.30)
Analysis of Strain : ' $ I S ~
y
,,"7
.." ,I
.,........ I ';
..... J I
,....... ... .., i
I' ...__... I I
" . . . . . . . . ,. . -;
R
' J I_..... .,. ....I,
. . . . , . S ' - - - - "
" - - AQ' 1 .t....-- .......
IV . . . .
j
' .......
I _....
6". . . . .
P'
o~ ~
x
z
Fig. 2. 5 Deformation of right parallelepiped
Consider now an infinitesimal rectangular parallelepiped with sides .1x, .1y and &
(Fig. 2. 5). When the body undergoes deformation, the right paral1eJ epiped PQRS
becomes an oblique parallelepiped P' Q' R' S ' .
Identifying PQ of Fig. 2. 5 with A. B of Eqs (2. 30), one has ~ y=&=O. Then, from
Eqs (2. 30) the projections of P' Q' will be
along x axis: ( I+~:') dx
along z axis:
along y axis:
Hence, one can successively identify AD with PQ (.1y =&=0), PR (dx =
&= 0), PS ( dx = .1y = 0) and get the components of P' Q' , P'R' and P' S ' along the
x, y and z axes as
.
z runs:
P'( ! P'R' P' S '
( I+o U
x
Idx
iJ u x ~
iJu
x
&
" ~
iJx J o y
iJz
Oily~ x
( iJ lI v )
Oil"
1+iJ ; .1y
.&
iJx iJz
du , dx Oil, .1
( 1+~; ) az
ox
o y ~
x axis:
yaxis:
The volume of the right parallelepiped before deformation is equal to V =dx ~ y
&. The volume of the deformed parallelepiped is obtained from the well-known
formula from analytic geometry as
V' =V +.1V =D .~ x .1y 6z
Copyrlqhted matenal
.
'16,' 'Advanced Mechanics of Solids
where D is the following determinant:
( 1 + ; : )
00"
O O
x
O y a z
a u
y
( 1+ ~ )
~
D=
O Ily
(2.31)
O x a z
00_ O u _
( 1 + ; ; )
-

O x O y
If we assume that the strains are small quantities such that their squares and
products can be negelected, the above determinant becomes
tJu - tJu o U
D=I+ x + y+ z
ox oy oz
=I~x., + Eyy + Eu ( 2.32)
Hence, the new volume according to the linear strain theory will be
V' =V+ 6V =(1 + e.u + Eyy + Eu ) 6x 6y 6z (2.33)
The volumetric strain is defined as
(2.34)
Therefore, according to the linear theory, the volumetric strain, also known as
cubical dilatation, is equal to the sum of three linear strains.
_.. '1_.0$. __ .... _' ... ~ i!i ...... _r.~" " .......... __ 9~.'1tn'5~am $ 597 '. 'lot 9 71'S' 9 *.. 3
Example 2.S Thejo/lo~rftJg state of strain exists at apoint P
/ t/.. .:
"" .
./ :>" '0.02 - 0.04 0
[ Eij] = - 0.04 0.06 0.02
o - 0.02 0
F ,!!If
In the direction PQ having direction cosines n" =0.6, fly =0 and n
z
= 0.8,
determine E
pQ
.
Solu tion From Eq. (2.20)
EpQ =0.02 (0.36) + 0.06 (0) + 0 (0.64) - 0.04 (0) - 0.02 (0) + 0 (0.48)
=0.007
.' .4 $ I " ' AU =La= _; ; "''',?m.o:-w" ;, . U te.... !; _
Example 2.6 In Example 2.5, what is the cu bical dilatation at point P?
.
\
Solu tion From Eq. (2.34)
6=Exx + Eyy+ E
zz
=0.02 +0.06 + 0 =0.08
,
Analysis of Strain :.~t

.:~~GE iN;IrHE ~GLE BETWEEN~:.:,_.:.
.' 'TW:e:'~LINE!ELEMENTS .:', ' ,!."
. i ,'..
.~-/' ..
."
"1 " I
, ..' . . '. '" .
Let PQ be aline element with direction cosines n
tl
nyl' n.
1
and PR be another line
element with direction cosines n~2' 1 1 ),2'n,,2' (Fig. 2.6). Let 9be the angle between
the two line elements before deformation.
y R '
p
!
r
,"\' 8' ,.0
, .>0'
I ...."'
, ,
c t ..>
R
p
O~ ~x
z
Fig. 2.6 Change in angle bet ween
t wo line segment s
After deformation; the line segments be-
come ,F'Q' and P'R' with anincluded angle
8'. We can determine 8'easily from the
results obtained in Sec. 2.9.
From analytical geometry
Th I f""" d
e va ues 0 n
xl
n)'I' n
zl
' n
x
2' ny2. an
n; 2 can be substituted from Eq. (2.29).
Neglecting squares and products of small
strain components.
cos 8'=( )1( ) [(1+2Exx) n
t
ln
t
2+(1+2e
yy
) n)'ln)'2
1 + &PQ 1+GpR
+(1+2E,,) 1l.
l
n
z
2 +Yx"Cn
x1
ny2+n ..2n"l)
+ yyz(n
yl
nz2 +"y2nzl) +Yzx(nrln
z
2 +"x2 n.
I
)] (2.35)
Inparticular, if the t"..o line segments PQ and PR are at right angles to each other
before strain, then after strain,
cos 9'=(I )ICl )[ 2E.0 ' n
xl
n
x
2 +2E.yy nyl lIy2 +2e..: n:] nz2
+ c PQ +CPR
+ Yx
y
(IIX]IlY2 +n
x2
n
yl
) +'Yy:(n
y1
fl
Z2
+ny2n: l) _
+ yzx(n
xi
n
z2
+n
x
2 IIzt)] (2.36a)
Now (90
0
- 9') represents the change inthe initial right angle. Ifthis isdenoted
by a, then
9'=90 - a
cos 9'=cos (90
0
- a) =sin ase a or
,
(2.36b)
(2.36c)
since a is small. Therefore Eq. (2.36a) gives the shear strain a between PQ and
PR. If we represent the directions of PQ and PR at P by x' and y' axes, then
Yx
y
' at P =cos 8' =expression given in Eqs (2.36a), (2.36b) and (2.36c)
4i~." :J _ ""'":U _ '.~I5i""n:.*'''~U!~''IIII!fI!?'';;M,e 'l"'l'&!'f:'I8_~"_""" _,__ '''''~_li'"''t''I''''''_7 ~
Example 2.7 The displac ement field for a body is given by
u =k(x
2
+y)i +k(y +z)j +k(:?+2z'l)k where k = 10-
3
At a point P(2, 2, 3), c onsider t wo line segment s PQ and PR having t he follow-
ing direc t ion c osines before deformat ion
Det ermine t he angle bet ween t he t wo segment s before and aft er deformat ion.
~ 1 ! r . " Advanced Mechanics of Solids
.
Solution Before deformation, the angle 9 between PQ and PR is
I I
cos 9 = n. >:1 11..
2
+n
yl
1 l
y2
+n.
I
II. 2 = J 6 +J 6 = 0.8165
9= 35.3
.
..
The strain components at P after deformation are
=2kx=4k ,
E =Ollz = 4kz = 12k
u O Z
=k , =4k
The linear strains in directions PQ and PR are from Eq. (2.20)
EpQ =k [ (4 x ~) +~+(12x ~) +(I x ~) +(I x t) +(4 x ~)] =k
EpR = k [(4 x ~) +(I x !)+(12 x 0) +(1x ~) +0+0] = 3k
After deformation, the angle beteween P' Q' and P'R' is from Eq. (2.35)
cos 8' =( 1) [ (1+8k) ~ +(1+2k) b - +0
1 +23/3k (1 +3k) ...;6 ...;6
+( } 6+*)k +(0+*)k +(
0
+)6)''.
=0.814 4 and 9 =35.5
. . , ; . . . - ' J
-, I' :. 'r 11+ .. r .
1 . '. 1 . . ',," . "
. , .. !".)o'
~. ',+ ., .4 ...
I .. ,0.' ,- ... ~
It was shown in Sec. 2.5 that when adisplacement field is defined at apoint P,the
relative extension (i.e. strain) at P in the direction PQ is given by Eq. (2.20) as
,
EpQ = EX-tn;+Ewn; +E"II; +r~~J ly +ryznyll. +r . . . . nx ll.
As the values of nx' ny and ". change, we get different values of strain EpQ' Now
we ask ourselves the following questions:
What is the direction (I I x, lIy, 11.) along which the strain is an extremum
(i.e. maximum or minimum) and what is the corresponding extremum value?
According to calculus, inorder to find the maximum or the minimum, we would
have to equate,
aEpr/dn
x
, dEpQldn
y
, dEpo!()n
z
,
to zero, if IIx, n; and liz were all independent. However, n:r > Ily and liz are not all
independent since they are related by the condition
(2.37)
book.
:. : Advanced Mechanics of Solids
Ifwe adopt the notation given in Eq. (2.22), i.e, put _
then Eqs (2.40a) can be written as
(Xt - )n x +ex!,y +exzn
z
= 0
ey:l'x +(Eyy - E)ny +eyzll
z
=0 (2.40b)
e Z A n . r +ezyn" +(zz - E)lI
z
=0
The above set of equations is homogeneous in n x ' ny and nz. For the existence of
a non-trivial solution, the determinant of its coefficient must be equal to zero, i.e.
(C
xr
-e )
exy ex:
e
yx
(c
yy
- 6)
eyz
=0
(2.41)
ezx ezy
(6
zz
- c)
Expanding the determinant, we get
~-J IE? +J
2
E-J
3
=0 (2.42)
where
J
I
=Exx +Eyy +En
J
2
=
E""
exy
Eyy
e y z
"xx
exz
+ +
e
yx
6
yy
e r y
6
zz
ezx e = z
(2.43)
(2.44)
(2.45)
It is important to observe that J
2
and J
3
involve exy' eyz and ezx. not 'Y x". Y y z and y ".
Equations (2.41)-(2.45) areall similar to Eqs (1.8), (1.9), (1.12), (1.13) and(1.14). The
problem posed and its analy sis are similar to theanaly sis of principal stress axes and
principal stresses. The results of Sec. 1.10-1.15 can be applied to thecase of strain.
For agiven state of strain at point P, ifthe relative extension (i.e. strain) e is an
extremum in adirection ft, then E is the principal strain at P and n is the principal
strain direction associated with E.
Inevery state of strain there exist at least three mutually perpendicular princi-
pal axes and at most three distinct principal strains. The principal strains 1' ~
and t;, are the roots of the cubic equation.
E3 - J
1
2 +J
2
- J
3
= 0 (2.46)
where J
I
, J
2
, J
3
are the first, second and third invariants of strain. The principal
directions associated with 1' ~ and E3 are obtained by substituting E
j
(i =1, 2, 3)
in the following equations and solving for n ", lIy and liz.
(E.... - Ej)n
x
+exyny +ex:llz = 0
exyllx +(Eyy - E;)ny +eyzll
z
= 0
n; +n; +n; =1
(2.47)
Analysis of Strain 81
If1' ~ and 3are distinct, then the axes of nl' n
2
and n3 are unique and mutually
peprendicular. If, say EI =~: t = f3' then the axis of n3 is unique and every direction
perpendicular to n) is aprincipal direction associated with 1=E : ! .
If 1=~=E ;, then every direction isaprincipal direction.
____________ '.-. __ "4 4' .. ~~
Example 2.8 The displacement field ill micro unit s for a body is given by
u = ( ~+y)i +( 3+z)j +(~+2y)k
Det ermine t he principal st rains at ( 3, 1, -2) and t he direct ion of t he minimum
.~
principal st rain.
Solut ion The displacement components in micro units are,
~=~+~~=3+z. ~=~+~.
The rectangular strain components are
au au
y
au
E x " = a: = 2x , liyy =ay =0, lizz =a; =0
aux ouy oU y oU z _ au. aux _ 2
r\ )' = 8 y + ax =1, ry% = O Z + oy =3, Y zx - ox + O Z - x
At point ( 3, 1, -2) the strain components are therefore,
f;tt =6, Ew. =0, E
zz
=0
rA Y = 1, ')5 ,. = 3, rzx = 6
The strain invariants fromE qs ( 2.43) - ( 2.45 ) are
11= E n +E yy +E
z
: = 6
6
J
0
3
6 3
-
-
1
2
=
2
+
2
+
23
=--
1
0
3
0 3 0
2
- -
2 2
Note that 12 and 13involve
1
e = - r
Y " 2 yz ;
6
1
3 -
2
1) =
I
0
3
=-9
-
2 2
3
3
0
-
2
The cubic fromE q. ( 2.46) is
8
3
- 68
2
- 238 +9=0
2
Following the standard method suggested in Sec. 1.15
a =j (-~-36) =-11
b =i7 ( -432 -621 +243) =-30
.
82. ' Advanced Mechanics of Solids
cos t P = - I = 0.684
2 x V-a3/27
t P =4648'
g =2~ a/3 =5.6
The principal strains in micro units are
E, =gcos;/3+2=+7.39
E
2
= g cos (;/3 +120)+ 2=-2
3 = g cos (;/3+240) +2=+0.61
-30
. .
As a check, the first invariant J , is
Eo + _ y y +E
zz
= , +~ +6
3
= 7.39 - 2 +0.61 =6
The second invariant J
2
is
EI~+~fJ +fJE
1
=-14.78-1.22 +4.51 =-11.49
The third invariant 13is
1~E3 = 7.39 x 2 x 0.61 =-9
These agree with the earlier values.
The minimum principal strain is-2. For this. from Eq. (2.47)
(6 +2) nx +! ny +3n
z
=0
1 3
2 /Ix +2ny +'2n, = 0
n
2
+n
2
+n
2
=1
x y z
The solutions are /Ix =0.267, n.. =0.534 and n
z
= -0.801.
Example 2.9 For t he st at e of st rain given ill Example 2.5, det ermine t h.e
principal st rains and t he direct ions of t he maximum and minimum prt ncipal
st rains.
Solut ion From the strain matrix given, the invariants are
1, = E" " t +E'y 'v +E
zz
= 0.02 +0.06 +0 = 0.08
0.02 -0.02 0.06 -0.01 0.02 0
12 = + +
-0.02 0.06 -0.01 0 0 0
=(0.0012 - 0.0004) +(- 0.0001) +0= 0.0007
0.02
13=-0.02
o
-0.02
0.06
-0.01
o
-0.01 = 0.02 (-0.000l) +0 +0=- 0.000002
o
The cubic equation is
~ - 0.08e
2
+0.0007+ 0.000002 =0
Co pyrlqhted matenal
,
Analysis of Strain '~.~
Following the standard procedure described in Sec. 1.15, one can determine the
principal strains. However, observing that the constant 13 in the cubic is very
small, one can ignore it and write the cubic as
E2 - 0.08E2+0.0007E= 0
One of the solutions obviously is E =O. For the other two solutions (s not equal
to zero), dividing by e
E2 - 0.08 e +0.0007 =0
The solutions of this quadratic equation are
E =0.4 0.035, i.e. 0.075 and 0.005
Rearranging such that E( <! ~ <! E
3
, the principal strains are
E
I
=0.07, E:!=0.01, 6.1=0
As a check:
1( =El +E:!+6.1=0.07 +0.01 =0.08
J
2
=E1E:! +~E2 +6.lE( =(0.07 x 0.01) =0.0007
(This was assumed as zero)
Hence, these values agree with their previous values. To determine the direction
of E( =0.07, fromEqs (2.47)
(0,02 - 0.07) n. - 0.02ny =0
- 0.02n. +(0.06 - 0.07) ny - O.OI nz =0
n
2
+n
2
+n
2
=I
y z
The solutions are n. =0.44, ny =-0.176 and n, =0.88.
Similarly, for 3 =0, fromEqs (2.47)
0.02nx - 0.02ny =0
-O.02n. +0.06ny - O.Oln
z
=0
112 +n
2
+n
2
=1
x y z
The solutions are "x =lIy =0.236 and n, =0.944 .
. PLANE STATE 'OF STRAlN
. .
"
I f, in agiven state of strain, there exists acoordinate system Oxyz, such that for
this system
Eu =0, r y : =0, rn=0 (2.48)
then tbe state is said to have a plane state of strain parallel to the xy plane. The
non-vanishing strain components are E
xx
' Eyy and rxy.
I f PQ is a line element in this xy plane, with direction cosines n., ny, then the
relative extension or the strain EpQ is obtained fromEq. (2.20) as
2 2
GpQ =G.uc 11" +Gy>, lIy +r xy nx lIy
or ifPQ makes an angle ) witb tbe x axis, then
2 Ll 2 Ll I . 2Ll
&PQ =G.
u
cos 17 +Gyy sI n 17 +2r.o / sin 17
(2.49)
Copyrlqhted matenal
If cI and ~ are the principal strains, then
(2.50)
Note that E : 3 =C
u
isalso aprincipal strain. The principal strain axes make anglese
and t P +90 with the x axis, such that
tan 2;= rxy (2.51)
lin-Byy
The discussions and conclusions will beidentical with the analysis of stress if we
I
use cn' Cyy' and Ct . in place of 0".., O"y and 0". respectively, and eX)' = 2 rry,
eyz = ; ryz , e
u
= ; rt x in place of't
x
)" 't}"Z and "z x respectively.
2. : 4, THE PRINCIPAL ~S ~f.: Sl'RA.lNRE?v.IAIN
. . ORTHOGONAL'. A1"TElt' ~STRA1N "
Let P Q beoneof theprincipal extensions or strain axes with direction cosines fl
xl
'
lIyl and "rl' Then according to E qs (2.40b)
(c
xx
- CI)"XI +e..yl
yl
+ed',1 =0
e x}Jl x
I
+(t;,. - 1)11),1 +eyz llz l =0
ex"'xl +eyll
yl
+(Sz! - CI)",I =0
Let "x2, "),2 and "z 2 be the direction cosines of a line P R, perpendicular to P Q
before strain. Therefore,
" x llIx2 +"YII1Y2 +"'ll1z 2 =0
Multiplying E q. (2.40b), given above, by 11x2' "),2 and "z 2 respectively and adding,
we get,
cx;r11xl
ll
x2 +c;,J lyIIl1'2 +c
z
!'z l"t 2 +exy(llxll1Y2 +"yl"x2) +e),< (lIyl",2 +"y2",)
+ ez x (lI
xl
n
z
2 +II
x2
":I) =0
Multiplying by 2 and putting
2cx'y =Y..Y'
we get
2SxxllXII1x2 +2C
y
yflYl1lY2 +2ez/l
z l
"z2 +YXy(l1,rl
1l
y2 +ll
yl
ll
x2
)
+I).z ("
yl
":
2
+lIY211%1) +Y
u
(lIxl1Iz 2 +".1",,2) =0
Comparing the above with E q. (2.3 6a), we get
cos 9' (I +epQ) (I +epR) =0
where 9' is the new angle between P Q and .P R after strain.
Since EpQ and EpR are quite general, to satisfy the equation, 8' =90, i.e. the
line segments remain perpendicular after strain also. Since P R is an arbitrary
perpendicular lineto the principal axis P Q, every lineperpendicular toP Q before
strain remains perpendicular after strain. In particular, P R can be the second
principal axis of strain.
Repeating the above steps, if P S isthe third principal axis of strain perpendicu-
lar to P Q and P R, it remains perpendicular after strain also. Therefore, at point P ,
Analysis of Strain :.'~'
wecan identify asmall rectangular element, with faces normal to theprincipal axes
of strain, that will remain rectangular after strain also.
[~t$3' PLANE STRAINS INPOLAR C00RDINATES
We now consider displacements and deformations of a two-dimensional
radial element in polar coordinates. The polar coordinates of a point a are
x rand 6. The radial and cir-
O~--~~---------------------+
,
,
,
:' l.d' c
, III ,
N"I
y
Fig. 2.7 Displacement components of a radial
element
cumferential displacements
are denoted by", and "e.
Consider an elementary ra-
dial element abed, as shown
in Fig. 2.7.
Point a with coordinates
(r, 6) gets displaced after
deformation to position a'
with coordinates (r +U,'
6 +a), The neighbouring
point b(r +lir, 6) gets
moved to b' with coordinates
(r +~I' +u ; +~; ~,O+a +: r .&)
The length of a'b' is therefore
The radial strain e; is therefore
OUr
e, =or
The circumferential strain e
8
is caused in two ways. If the element abed under-
goes apurely radial displacement, then the length ad = r ~6 changes to (I' +u ,)118.
The strain due to this radial movement alone is
u ,~O u ,
-'---:-=-
r~O I'
In addition to this, the point d moves circumferentially to d' through the distance
QUo
Uo + Of) ~O
(2.52)
Since point a moves circumferentially through Ue, the change in ad is
~: ~8. The strain due to this p,artis
oUo ~O IoUo
se rli0=-;. sa
Thetotal circumferential strain istherefore
u , I(}UO
EO =-+-~:-
r r Of)
(2.53)
Copyrlqhted matenal
f -
'M'; Advanced Mechanics of Solids
. .'
To determine the shear strain we observe the following:
The circumferential displacement of a is ue, whereas that of b is
oUo
u
8
+ !lr. The magnitude of 8
2
is
or
But
a= uo.
r
8
2
=( u
o
+ouo ~ r _ "0 _ "0 ~ ) _!_
or r ~r
oUo Uo
= --
or r
Similarly, the radial displacement of a is Ity, whereas that of dis u, +
OU, !l8.
e o
Hence,
Hence,
"
IOu,
=---=::-
rt5(}
Hence, the shear strain Y re is
(2.54)
. COMPATIBH..ITY CONDITIONS
It was observed that the displacement of a point in a solid body can be repre-
sented by adisplacement vector u, which has components,
UX ' u
Y
' uz
along thethree axesx, y andz respectively. Thedeformation at apoint isspecified
by the six. strain components,
xx' YY'~ z' rxy, ryz and rzx
The three displacement components and the six rectangular strain components
are related by the six strain displacement relations of Cauchy, given by Eqs (2.18)
and (2.19). Thedetermination of the six strain components fromthe three displace-
ment functions is straightforward since it involves only differentiation. However,
the reverse operation, i.e. determination of the three displacement functions from
the six strain components is more complicated since it involves integrating six
equations to obtain three functions. One may expect, therefore, that all the six
strain components cannot be prescribed arbitrarily and there must exist certain
relations among these. The total number of these relations are six and they fall
into two groups.
Analysis of Strain ;"if:

First group: We have
OU OU
y
y _ x+--=- ....
.ty - oy ox
Differentiate the first two of the above equations as follows:
1J2
sxx
03
ux
0
2
(OU.<)
oy2 = oxoy2 =oxoy oy
02Syy 03Uy 02 (ou
y
)
ox
2
- oyox
2
- oxoy ox
Adding these two, we get
0
2
(OU
x
+OUy) =02yXY
oxoy oy ox oxoy
i.e.
Similarly, by considering E
yy
' E
zz
and YYZ' and E.:, En and Y:X' we get two more
conditions. This leads us to the first group of conditions.
(2.55)
Second group: This group establishes the conditions am.ong the shear strains.
We have
Differentiating
O r X ) '
i J
2
u i J 2
Uy
x +
O Z i J z i J y i J z o x
O r } "
02u i J 2
y U.
-
+
o x i J x i J z i J x o y
book.
book.
book.
Analysis of Strain .:~1
when abody-is subjected to this particular state of strain, then every direction is
aprincipal strain direction, withastrain of magnitude e, according to Eq. (~.20). A
sphere subjected to this state of strain will uniformally expand or contract and
remain spherical.
Consider the invariants of the strain deviator. These are constructed in the
same way as the invariants of the stress and strain matrices with an appropriate
replacement of notations.
(i) Linear invariant iszero since
f{ = ( E . I ' - <- e) +( E y y - e) +(e
zz
- e) = 0
(ii) Quadratic invariant is
(2.60)
= - i [ ( & x x _t.'
y y
) 2 +(& Y .J ' _ & z: z. ) 2 +(& %2 - & x . ,) 2
+ 6 ( e x y + e
y x
+ e z x ) 2 ]
(iii) Cubic invariant is
(2.61)
& - e
e x y e x :
x x
J'-
e > : y
& - e
e ) 'Z 3-
Y .J '
e x : e z y
/ : : - e
zz
(2.62)
The second and third invariants of the deviatoric strain matrix describe the two
types of distortions that an isolated element undergoes when subjected to the
given strain matrix eij'
--------l>robl
ems
--------
2.t The displacement field for abody is given by
u = (xl + y); + (3+ z)j + (x
2
+ 2y)k
Write down the displacement gradient matrix at point (2, 3, 1).
4 I 0
Ans. 0 0 1
420
2.2 The displacement ueld for abody is given by
u = [(x
2
+ y 2 + 2); + (3x + 4y ~j + (2x 1+ 4z)k] I 0-4
What is the displaced position of apoint originally at (1, 2, 3)7
[ Ans. (1.0007,2.0019,3.0014)]
2.3 For the displacement field given inProblem 2.2, what are the strain compo-
nents at (1, 2, 3). Use only linear terms.
Copyrlqhted matenal
book.
book.
book.
book.
book.
book.
'$8: Advanced Mechanics of Solids
t...~
components of stress may be expressed as a linear function of the six components
of strain and vice versa.
The mathematical expressions of this statement are the six stress-strain
equations:
ax =att&xx +a
l2
b'yy +a
t
3&zz +a
l4
y. ty +atSY yz +al6 Y zr
a
y
=a~nGl:~ +a22t'yy +a23
G
zz +a24Y. ty +a2SY)'% +a
2
6Y zx
a, =a31
e
n +a32
e
Y,)' +a33
8
"" +a34r. ty +a351 yz +a
3
61 zx (3.2)
Ty. =aSIG
x
, +aS2
G
}), +a
S3
G
zz
+a54r. ty +0ssr yz +0S6Y zx
't'zx =a61&x, +a62&n' +a63
e
zz +a
64
y. ty +a6sY yz +a66Y zr
Or conversely, six strain-stress equations of the type:
En =bllu, +ht2
U
y +ht)u. +ht4'T. ty +hts'T yz +ht6t'zx
t),. =... etc
(3.3)
.
where all' a
12
, b
ll
b
12
... are constants for a given material. Solving
Eq. (3.2) as six simultaneous equations, one can get Eq. (3.3), and vice versa. For
homogeneous, linearly clastic material, the six Eqs (3.2) or (3.3) are known as
Generalised Hooke's Law. Whether we use the set given by Eq. (3.2) or tbat given
by Eq. (3.3), 36 elastic constants are apparently involved.
='-' Sll'aES$-STRAIN RELkTIONS.FOR]SOTROPJC' ,
MATERIALS' ,
I
, .
We now make a further assumption that the ideal material we are dealing with has
the same properties in r II directions so far as the stress-strain relations are con-
cerned. This means that the material we are dealing with is isotropic, i.e. it has no
directional property.
Care must be taken to distinguish between the assumption of isotropy, which
is a particular statement regarding the stress-strain properties at a given point,
and that of homogeneity, which is a statement that the stress-strain properties,
whatever they may be, are tbe same at all points. For example, timber of regular
grain is homogeneous but not isotropic.
Assuming that the material is isotropic, one can show that only two indepen-
dent elastic constants are involved in the generalised statement of Hooke's law.
InChapter I, it was shown that at any point there are three faces (or planes) on
which the resultant stresses are wholly normal, i.e. there are no shear stresses on
these planes. These planes were termed the principal planes and the stresses
on these planes the principal stresses. In Sec. 2.14, it was shown that at any point
one can identify before strain, a small rectangular parallelepiped or a box which
remains rectangular after strain. The normals to the faces of this box were called
tbe principal axes of strain. Since in an isotropic material, asmall rectangular box
the faces of which are subjected to pure normal stresses, will remain rectangular
book.
book.
book.
: '6., YOUNG'S MO:E>ULUSAND POISSON'S RATIO
FromEq. (3.13b), wehave
Ll =0') +0'2 +0'3
( 3 A +2 , u )
Substituting this in Eq. (3.4a)
0') =( J .. ) ( 0 ' ) + 0 ' 2 + 0 ' 3 ) + 2 . u & )
3J .. +2f.1
or (3.15)
Fromelementary strength of materials
where E is Young's modulus, and v is Poisson's ratio. Comparing this with
Eq. (3.15),
E = . u ( 3 A + 2 . u ) .
( J . . + , u ) ,
(3.16)
~1RELATIONS BETWHEN TIlE ELASTIC C0NSTANTS
In elementary strength of materials, we are familiar with Young's modulus E,
Poisson's ratio v, shear modulus or modulus of rigidity G and bulk modulus K.
Among these, only two are independent, and E and v are generally taken as the
independent constants. The other two, namely, G and K, are expressed as
G= E
2(I+v)'
K= E
3 ( 1 - 2 v)
(3.17)
It has been shown in this chapter, that for an isotropic material, the 36 elastic
constants involved in the Generalised Hooke's law, can be reduced to two inde-
pendent elastic constants. These two elastic constants are Lame's coefficients
;..and u. The second coefficient J . L is the same as the rigidity modulus G. In terms
of these, the other elastic constants can be expressed as
E =, u ( 3 J . . +2 . u )
- ( , , 1 + . u ) ,
K =(3,j, +2f.
i
)
3 '
, 1 . = vE
. ( 1 + v) ( 1 - 2 v) '
(3.18)
It should be observed from Eq. (3.17) that for the bulk modulus to be positive,
the value of Poisson's ratio v cannot exceed 1/2. This is the upper limit for v.
For v= 1/2,
3G =E and K =00
book.
book.
book.
1~06' Advanced Mechanics of Solids
" .
p
z
Fig. 3.1 Example 3.J
Solution
(i) Let I be the dimension of the cube. Since the cube is constrained in x and
y directions
and
Therefore
e .. . =0 and Y .l' =0
C 1 z =-p
Solving
v
1-v
C J. =-
z
v
p
I-v
If Poisson's ratio =0.5, then
< Y x =O'y =0', =-p
(ii) The extremum shear stresses are
(1- u,
,-' -
3- 2 '
If v < 0.5, then Gx and O'y are numerically less than or equal to G,. Since G,. ,
G
y
and O'z are all compressive
C J. =C J. =_ v p
, x 1-v
0i=C J.=- v p
y I-v
0') =G =-p
z .
... ,=p (1 _ v ) =I - 2v p,
I-v I-v
Ifv =0.5, the shear stresses are zero.
.
..
1 - 2v
1"2 =I-v p,
Example 3.2 A cubical element is subjected to thefollowing state ofstress.
O'x =100MPa, O'y =-20 MPa, a,=-40 Mpa, ... xy ="'yz ='rz x =.0
book.
book.
book.
Anyone of the above or some other factors might have caused the yielding.
Further, as pointed out earlier, the factor that causes a ductile material to yield
might be quite different from the factor that causes fracture in a brittle material
under the same loading conditions. Consequently, there will be many criteria or
theories of failure. Itis necessary to remember that failure may mean fracture or
yielding. Whatever may be the theory adopted, the information regarding it will
have to be obtained from a simple test, like that of a uniaxial tension or a pure
torsion test. This is so because the state of stress or strain which causes the
failure of the material concerned can easily be calculated. The critical value
obtained from this test will have to be applied for the stress or strain at apoint in
a genera] machine or astructural member so as not to initiate failure at that point.
There are six main theories of failure and these are discussed in the next sec-
tion. Another theory, called Mohr's theory, is slightly different in its approach
and will be discussed separately.
4.2' THEORIES OF FAILURE
Maximum Principal Stress Theory
This theory is generaJ ly associated with the name of Rankine. According to this
theory, the maximum principal stress in the material determines failure regardless
of what the other two principal stresses are, so long as they are algebraically
smaller. This theory is not much supported by experimental results. Most solid
materials can withstand very high hydrostatic pressures without fracture or with-
out much permanent deformation if the pressure acts uniformly from all sides as is
the case when asolid material is subjected to high fluid pressure. Materials with
a loose or porous structure such as wood, however, undergo considerable perma-
nent deformation when subjected to high hydrostatic pressures.On the other hand,
metals and other crystalline solids (including consolidated natural rocks) which
are impervious, are elastically compressed and can withstand very high hydro-
static pressures. In less compact solid materials, amarked evidence of failure has
been observed when these solids are subjected to hydrostatic pressures. Further,
it has been observed that even brittle materials, like glass bulbs, which are' subject
to high hydrostatic pressure do not fail when the pressure is acting, but fail either
during the period the pressure is being reduced or later when the pressure is
rapidly released, It is stated that the liquid could have penentrated through the
fine invisible surface cracks and when the pressure was released, the entrapped
liquid may not have been able to escape fast enough. Consequently, high pres-
sure gradients are caused on the surface of the material which tend to burst or
explode the glass. As Karman pointed out, this penentration and the consequent
failure of the material can be prevented if the latter is covered by a thin flexible
metal foil and then subjected to high hydrostatic pressures. Further noteworthy
observations on the bursting action of a liquid which is used to transmit pressure
were made by Bridgman who found that cylinders of hardened chrome-nickel steel
were not able to withstand an internal pressure well if the liquid transmitting the
pressure was mercury instead of viscous oil. It appears that small atoms of mer-
cury are able to penentrate the cracks, whereas the large molecules of oil are not
able to penentrate so easily.
book.
book.
book.
~4f. ; Advanced Mechanics of Solids
Work Done =
Area Under Triangle
:,ix
0', ---t:--e=t:__ -_- __ -__ -__ l__ . _ . . . . . . -.-+--+- 0'1

//' ,~~
, .
, ,
, ,
/' !
(b)
0'2
(a)
Fig. 4.3 (a) Principal stresses on a reaangular block
(b) Area representing work done
FromHooke's law
&\x=&1~x =t[0'1 - V (0'2 +0" ) )] ~x
~Y=&2 6 . y = i[ 0" 2 -V(O'I +0'3 )] 6 . y
&\z=&) ~z=~[ 0" 3 - V( 0" 1 +0'2)] ~z
Substituting these
~w=2~[ O '? +O " i +0'; - 2v (0'10'2 +0" 20'3 +0'30'1)] ~x ~y~z
The above work is stored as internal energy if the rate of defonnation is small,
Consequently, the energy Uper unit volume is
( 4.5)
In auniaxial test, the energy stored per unit volume at yield point or elastic limit
.
is 1/2E 0'; Hence, failure occurs when
0" : +0'; +0" ; - 2v ( 0'10'2+0'20'3 +0'3 0'1) >0'; (4. 6 )
This theory does not have much significance since it is possible for amaterial to
absorb considerable amount of enerty without failure or permanent deformation
when it is subjected to hydrostatic pressure.
Eneray of Distortion Theory
This theory is based on the work of Huber, von Mises and Hencky. According to
this theory, it is not the total energy which is the criterion for failure; in fact the
book.
book.
book.
1,1& Advanced Mechanics of Solids
,
(1
(1-T
- y
Fig. 4.4 Uniaxial and pure shear state of stress
If one uses the yield point shear stress 1')' obtained from a pure torsion test,
then the critical value associated with each theory of failure is as follows:
(i) Maximum N01'JtI4IStress Theory According to this theory, failure occurs when
the normal stress s at any point in the stressed member reaches a value
(1~fy
This is because, in a pure torsion test when yielding occurs, the maximum
normal stress s is numerically equivalent to t
y
(ii) Moximum SMa, Stt-ess TlJeory According to this theory, failure occurs when
the shear stress t at a point in the member reaches a value
l'~1')'
(iii) Moximum SlnIi" Theory According tothis theory, failure occurs when the
maximum strain at any point inthe member reaches avalue -
1
e =E [0'1 - v(0'2 +0'3)]
From Fig. 4.4, in the case of pure shear
0'1 = 0' = f, 0'2= 0, 0'3=--0' =-1'
Hence, failure occurs when the strain e at any point inthe member reaches a
value
(iv) OctaMd,al SMa, Stt-ess Theory When an element is subjected to pure shear,
the maximum and minimum normal stresses at apoint are sand -s (each numeri-
cally equal to the shear stress '(), as shown in Fig. 4.4. Corresponding to this,
from Eq. (l.44a), the octahedral shear stress is
I[ 2 2 2J I/ 2
fOCI =3 (0') - 0'2) +(0'2 - 0'3) +(0'3 - 0'1)
Observing that 0'1 =0' =f,0'2=0, 0:, =-0' = -1'
f =!(0'2 +0'2 +40'2 )1/2
OCI 3
=-160'= Il-c
3 '/3
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
1 s t Advanced Mechanics of Solids
In exactly the same way, one can show that
iJU' 8U
iJP; =02' iJFj =03,,etc.
That is to say, 'the partial differential coefficient of tbe strain energy function
with respect to F; gives thedisplacement corresponding with F/. This isCastigliano's
first theorem. In, the form derived in Eq. (5.16), the theorem is applicable to only
linearly elastic bodies, i.e. bodies satisfying Hooke's Law (see Sec. 5.15).
This theorem is extremely useful in de-
termining the displacements of structures
as well as in the solutions of many stati-
cally indeterminate structures. Several
examples will illustrate these subse-
quently. We can give an alternative proof
for this theorem as follows:
Consider an elastic system in equilib-
rium under the force F
I
> F
2
, ' . F", etc.
(Fig. 5.9). Some of these are concentrated
loads and some are couples and torques.
Let the strain energy stored be U. Now
increase one of the forces, say F", by
!iF" and as a result the strain energy in-
creases to U +I1U, where
,
,
,
Fig. 5.9 Elastic body in equilibrium
under forces FI1 F:z, etc.
I1U A C'
t1U =11F orr.
II
Now wecalculate the strain energy inadifferent manner. Let theelastic system be
free of all forces. Let 11 F" be applied first. The energy stored is
1
2 11F" t::.t5"
where AO " is the elementary displacement corresponding to 11F". This is a quan-
tity of the second order which can be neglected since A F; will be made to tend to
zero in the limit. Next, weput all theother forces, F .. F2 , ... ,etc. These forces by
themselves do an amount of work equal to U. But while these displacements are
taking place, the elementary force I 1Fn is acting all the time with full magnitude at
the point n which is undergoing a displacement D". Hence, this elementary force
does work equal to 11 F; 13 ". The total energy stored is therefore
I
U +11F" s, +2 t::.F" ss,
Equating this to the previous expression, we get
"
I1U I
U +I1F 11F" =U +I1F. s; +2
11Fn
11o"
n
Inthe limit, when I 1Fn ~ 0
iJU =t5
8F n
n
book.
book.
book.
158' Advanced Mechanics of Solids

0
(iv) Due to torque
(5.25)
(5.26)
,_ C U d .
I
..... _'l"""'(!:JO~a:;;;: ' 1 ,.-: " " " " _ ....... ,. $ r9 m S1s
Example 5.5 Determine the deflection at end A of the cantilever beamshown ill
Fig. 5.13.
Solution The bending moment
at any section x is
A1=Px
The elastic energy due to bend-
ing moment is, therefore, from
Eq. (5.24)
~ L p
Fig.5.13 Example 5.5
U _ L
J
{Px)2 dx p2J-3
1- 2EI =6E1
o
The elastic energy due to shear from Eq. (5.22) is (putting k, =1)
Lp2 P2L
U
2
=J 2AG dx =2AG
o
One can now show tbat U
2
is small as compared to UI' If the beam is of a
rectangular section
A= bd,
and 2G:; E
Substituting these
U P2L
_l. =77"~
U
I
2bdG
d
2
-
2L2
For a member to be designated as beam, the length must be fairly large com-
pared to the cross-sectional dimension. Hence, L >d and tbe above ratio is
extremely small. Consequently, one can neglect shear energy as compared to
bending energy. With
p2 J J
U = --::-::-:-
6El
we get
8U PL
3
8P - 3El =OA
which agrees with the solution fromelementary strength of materials.
book.
book.
book.
t6~~Advanced Mechanics of Solids
-
IJU ML. PL
2
8= iJM =EJ +2 ET
This gives the slope when M and P are both acting. If M is zero, the sLopedue to
P alone is
PL
2
f J =2 EI
If on the other hand, P is zero and A-f alone is acting the slope is
,
(J= ML
EI
_____ ,,__ --=e "''''_ e=a"
Example 5.11 For the member shown in Fig. 5.16, Example 5.8, determine the
ratio of L to r if the horizontal and vertical deflections ofthe loaded end A are
equal. P is the only force acting,
Solution In addition to the vertical for P at A, apply ahorizontal fictitious force
F to the right, The bending moment at section 8of the semi-circular partis
MI = Pr (I - cos 8) - Fr sin (J)
At any section x in the vertical part, the moment is
M2 =2 Pr+Fx
Hence,
..
1 11' 2 I L 2
V= 2 E[ f[Pr(l-cosB)-FrsinOJ rdO+
2 ET
f(2 Pr+Px) dx
o 0
iJU 2 J r I L
iJF =- ~J![pr(l- cosO) - Fr sinOJ sinOdO +EI !(2 Pr+Fx)xdx
and
au ,.2 J r I L
of p",o =s, =-EI ![Pr(J - cosO) sin OJ so +E1 !2 Pr xdx
= _ 2 Pr
3
+PrL
2
= Pr (-2 2 L2 )
EI EI EI r +
FromExample 5.8
(J,,= ~: (~ 1rr+4 L)
Equating (J" to (Jh
~ ; (~ 1rr +4 L) =~; (-2 r2 +L2 )
or L
2
-4 Lr-,.2 (3 ; +2)=0
Dividing by ?and putting f. =p
,.
p2 -4p -( +2)=0
book.
book.
book.
i.e,
. .
i.e,
.
.
or
To determine the increase in the diameter along the loads, one has to determine the
elastic energy and take the differential. If one considers the quarter ring, Pig. 5.20(c),
the elastic energy is
n/2 I [ p r ( 2 )J 2
U =!2l:!,j T tt - cos ( J r o
The differential of this with respect to (PI2) will give the vertical deflection of the
end B with reference to.A. Observe that in order to determine the deflection at B,
one has to take the differential with respect to the particular load that isacting at
that point, which is (P/2). Putting (P/2) =Q.
I en[ ( 2 )J 2
U = J Qr - - cos 0 r dO
2EI 0 1
Q2 3 tr12( )2
= r J l.-coso dO
2EI 0 tt
eo : Q,. 3 t r /2( 4 2 4 )
a Q = EI !tt2 +cos ( J - 1 t OOS ( J a
.
3 I;
=~(: 2 ~ +: -!)
=~; (~-~)=: ~;(~-~)
As this gives only the increase intbe radius, the increase in the diameter is twice
this quantity, i.e.
~=Pr
3
( 1 t _ 1 )
EI 4 7(
THEOREM OF VIRTUAL WORK
Consider an elastic system subjected to a number of forces (including moments)
F
l
, F
2
, . , etc. Let 0
1
, ~ , , etc. be the corresponding displacements. Remember
that these are the work absorbing components (linear and angular displacements)
in the corresponding directions of the forces (Fig. 5.21).
Let one of the displacements 0
1
be increased by a small quantity ':\0
1
, During
this additional displacement, all other displacements where forces are acting are
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
Bending of Beams 203
~~~-~_-t.;~~~/~'-~--==~~~~~-~~~ o a
x
-00(-,--, ....._.---+ a x + ox
, , , a x
: : 1' s x : :
~ , t f
I I J I X'
/ ' I __ ._._._.~
I ~ I \
t \1.
I ,. \
.;: ,..-- -, ~
\ ' .. I I
\ \... " ' r-__ __~__..__.l /' ....._~__
\. ........ I ~ "_'""
->( " . 1 . " / '
/.... ~ .."
... ...... .. ....
/ ....._-------------..-. ---....._-
rr-. ,-!.._ aM
M'J -< ; ---' M
z
+ ax
z
sx
Fig.6.12 Free-bo dy dia gra m o f a n elementa ry length o f bea m
Let M, bebending moment at section x and M, + oM, ax the bending moment at
o x
section x +lx. C T
x
and C 1
x
+ ~~. are corresponding flexural stresses at these
two sections. It is important to observe that for the moments shown the normal
stresses should be compressive and not as shown in the figure. However, the
sign of the stress will becorrectly givenby Eq.(6.8). Considering a Lengths of the
section,the unbalanced normal force is balanced by the shear stress ~t.rdistrib-
uted along the length lx.For equilibrium, therefore,
s80-
T s. . . =_.1 I0 x t ds
t, 0 x
ts isthe wall thickness at s. Observing that ! v I ) ' =0,the normal stress C 1
x
is given by
.Eq.(6.8) as
.
i.e. (6.20)
CT .=
x
yl}, - zl y: M
l~", -Iy t, Z
yly - zlyZ oM
z
i; - I y I , o x
(6.21) Hence,
Recalling fromelementary strength of materials
Eq.(6.20)
oM
z
= -Vand substituting in
a x '
(6.22) or
T he first integral on the right-band side represents the first moment of the area
between s =0and s about the z axis.T he second integral isthe'tii.rstmoment of the
same area between s =0and s about they axis.Since 'r...is the complementary
shear stress,its value at any s is also given by Eq,(6.22).
book.
book.
book.
book.
book.
book.
book.
Bending of Beams
is zero, the neutral axis gets shifted towards the centre of the curvature. For
equilibrium, we have,
[ C T . , dA = -Ero ( 1 , _ .i) I ydA =0
A I ro ArO-y
and -IC ixydA = +
Er
o( 1 . _ .i) Jy2dA=M
A I ro ArO-y
From the first equation above
I ydA =0
ArO - y
The second equation can be written as
( 6.33)
+ Ero ( 1 , _ .i) [ _ [y dA +ro [ y dA ] =M
I ro A ArO-Y.
The first integral represents the static moment of the section with respect to the
neutral axis and isequal to ( -Ae), where e isthe distance of the centroid fromthe
neutral axis NN and this moment-is negative. The second integral is zero accord-
ing to Eq. ( 6.33) . Thus,
Ero (!-.i) Ae = M
r ro
( 6.34)
But fromEq. ( 6.32)
Er. o (!_ _ I)
r r
o
C i
x
( ro - y)
=-
y
Substituting this inEq. ( 6.34)
C i
x
( 1 ' 0 - y) -M
- Ae -
y
or
M y
C 1 =- - ~"--,
x Ae( ro-y)
( 6.35)
As Eq. ( 6.35) shows, thenormal stress varies non-linearly across thedepth. The
distribution is hyperbolic and one of its asymptotes coincides with the line pass-
ingthrough the centre of curvature, as shown inFig. 6.21 ( a) . The maximum stress
may occur either at the top or at the bottom of the section, depending on its
shapes. Equation ( 6.35) isoften referred to as the Winkler-Bach formula.
. In some texts, the origin of the coordinate system istaken at the centroid of the
section instead of at the point of intersection of the neutral axis and they axis. If
. .
the origin is taken at the centroid and y' is the distance of any fibre from this
origin, then putting y =y' - e and ro = Po - e, Eq. ( 6.35) becomes
M y'-e
(J =- - ----"-----,--
x Ae Po . : e - y' +e
book.
book.
book.
Bending of Beams
p
A _. __ .
Fig. 6.27 Example 6.6
Solution For the section Po= II em, h = 6 em, b =4 em.
Po +hl2 7
log Po _ hl2 = log 4 =0.5596
.
..
From equations (6.38) and (6.39)
6
ro =0.5596 =10.73,
e =11-10.73 =0.27
From Eq. (6.35), owing to bending moment M
a' = _ A l y
x Ae (ro - y)
_ M' y
--
24 x 0.27 (10.73 - y)
For the problem
and, at D,
M =P (a +a +h12) = 19P
Y =-(e +h12) =-3.27
r= ~-e=2.73
(a') =_ 19P x (-3.27) =0.6848 P
sc 24xO.27 (10.73+3.27)
(
') 19P 2.73 1001P
a.
t
D =- 24xO.27 (10.73-2.73) =- .
The stress due to direct loading is
A t C,
Hence,
a nd
a" = - P =- L=- 0 0417 P
x A 24 .
Hence the combined stresses are
(O 'x} e =(0.6848 - 0.0417) P
=0.6431P = 129 kgf7cm
2
(12642 kPa)
OPY IQhted ate,- a
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
~ Advanced Mechanics of Solids
o~
=0
=0
0=
(c)
(a)
Fig. 7.2 (a) Figure to show that shear stress must be tangential to boundary;
(b) shear stress at the comer of a reaangular section being zero as
shown in (c). .
assumed. The strains are then determined from strain-displacement relations
[Eqs (2.18) and (2.19)]. Using Hooke's law, the stresses are then determined.
Applying the equations of equilibrium and the appropriate boundary conditions,
we try to identify the problem for which the assumed displacements and the
associated stresses are solutions. -
"
,(r)lff9~~;~P;J.JSlAP( >,B'~~~~~.~. :','.
'......T' t' ~. I -. t' It' ~ ~I r'. to' i~~f:~:l~i ..~l ..-''I' t.
f
t.,. . '1.+ t'-.' I .,1 ';,.-,,: - '/I,'t"
... , .' , ".... \...... .,_.. ...... ," .... c ... l ...,' ,"
We shall now consider the torsion of prismatic bars of any cross-section
twisted by couples at the ends. It is assumed here that tbe shaft does not
contain any holes parallel to the axis. InSec. 7.12, multiply-connected sections
will be discussed.
On the basis of the solution of circular shafts, we assume that the cross-
sections rotate about an axis; the twist per unit length being O . A section
at distance z fromthefixedendwill, therefore, rotate througb Oz. A point Pix, y)
inthis sectionwill undergo adisplacement rOz, as shown inFig. 7.3. Thecompo-
nents of tbis displacement are
U
x
=-rOz sinf3
u
y
=rOz cos f3
y
x
y
I.. z---+t~1
(a)
T
Y
\
(c)
Fig. 7.3 Prismatic bar under torsion and geometry of deformation
book.
book.
book.
~. Advanced Mechanics of Solids
according to the boundary condition Eq. (7.14). Similarly, we can show that
I I 'ry: dx dy = 0
R
Now coming to the moment, referring to Fig. 7.4(a) and Eq. (7.9)
T= I I ( ' r ) , % x- f u y ) dxdy
R
=GO I I ( x2 +y 2 + x~ - Y i J I { I ) dxd
Y
R oy ox
Writing J for the integral
J= I I ( X 2 +y 2 +X ~ - y ~ ) dx dy
R
(7.16)
we have T= GJ ) (7.17)
The above equation shows that the torque Tis proportional to the angle of twist
per unit length with a proportionality constant GJ , which is usually called the
torsional rigidity of the shaft. For a circular cross-section, the quantity J reduces
to the familiar polar moment of inertia. For non-circular shafts, the product GJ is
retained as the torsional rigidity.
..
An alternative approach proposed by Prandtl leads to a simpler boundary condi-
tion as compared to Eq. (7.14). In this method, the principal unknowns are the
stress components rather than the displacement components as in the previous
approach, Based on the result of the torsion of the circular-shaft, let the non-
vanishing stress components be fu and 1"yr. The remaining stress components C 1 " ,
C 1
y
, C 1
z
and 1"zy are assumed to be zero. In order to satisfy the equations of equilib-
rium we should have
Or zx +m y: =0
ox oy
(7.18)
If it is assumed that in the case of pure torsion, the stresses are the same in every
normal cross-section, i.e. independent of z, then the first two conditions above are
automatically satisfied. Inorder to satisfy the third condition, we assume afunction
t P (x, y) called the stress function, such that
'f - t 3 r P ' " - iJ (7 19)
zx- oy' ' ) ' Z- - Tx .
With this stress function (called Prandtl's torsion stress function), the third condition
is also satisfied. The assumed stress components, ifthey are to be proper elasticity
solutions, have to satisfy the compatibility conditions. We can substitute these
directly into thestress equations of compatibility. Alternatively, wecan determine the
strains corresponding to the assumed stresses and then apply the strain compatibility
conditions given by Eq. (2.56). The strain components from Hooke's law are
En =0, y y =0, E n =0 (7.20)
book.
book.
book.
book.
book.
,.'
These two are identical if
a
2
I-A
b
2
=1+A
b
2
_ a
2
A =-:---:-
b
2
+a
2
Therefore, the function
or
b
2
_a
2
" ' = b2 2 xy
+a
represents the warping function for an elliptic cylinder with semi-axes a and b
under torsion. The value of I, as given inEq. (7.16), is
I =I I (x
2
+y2 +Ax
2
- AI) dx dy
R
= (A + I)I I xl dx dy +(1- A) I I I dx dy
= (A +1}ly+(I-A)Ix
,
Substituting Ix =
llab
3
and 1=
4 y
Hence, from Eq. (7.17)
T a
2
+b
2
8=--~
G lla
3
b
3
The shearing stresses are given by Eq, (7.9) as
(7.32) or
I
I
I
I
!
I
I
I
I
I
2Tx
(733a)
and similarly,
(7.33b)
book.
book.
book.

2Af Advanced Mechanics of Solids
Our equations are, as before,
0
2
' 1 1 0
2
' 1 1
---'':- + - 0
O X
2
oy2 -
over the whole region R of the rectangle, and
(1; - y) nx +(~ +x) n, =0
on the boundary. Now on the boundary lines x = a or AB and CD, we have nx = I
and ny= O. Onthe boundary lines BC and AD, we have nx= 0 and n}' = l. Hence,
the boundary conditions become
iJIp =y on x = a
ox
iJIp =-x on y=b
oy
These boundary conditions can be transformed into more convenient forms if we
introduce a new function .,., such that
'II =xy - 'III
In terms of ., r- the governing equation is
iJ2 ' 1 1 . +iJ2 ' 1 1 . =0
ox
2
oy2
.
over region R, and the boundary conditions become
0V t1 = 0
ox
0V t . =2x
oy
on x=a
on y=b
I
I
I
I
It is assumed that the solution is expressed in the form of infinite series
I
I
I
where Xn and Y
n
are respectively functions of x alone and y alone. Substitution
into the Laplace equation for .,. yields two linear ordinary differential equations
with constant coefficients. Further details of the solution can be obtained by
referring to books on theory of elasticity. The final results which are important are
as foUows:
The function J is given by
J=Ka
3
b
For various bla ratios, the corresponding values of K are given in Table 7.J .
Assuming that b >a, it is shown in the detailed analysis that the maximum
I
I
I
I
Copyrigllted material
book.
book.
book.
book.
book.
book.
book.
; ! 60 Advanced Mechanics of Solids
The compatibility conditions given by Eq. (2.56) yield
0
2
; 0
2
;
-.;:- +- .:: V
2
1 f>: : a constant F
ox2 qy2
So far, the analysis is identical to that given in Sec. 7.3. Considering the boundary
'conditions, we observe that there are several boundaries and on each boundary
the conditions given by Eq. (7.1 2), Sec. 7.2 should be satisfied. Since each bound-
ary is a free boundary, we should have
II 8 t P _II 8 t P =0
x oy )'-ox
Substituting for " x and II)' from Eq. (7.1 3)
.
8 t P dy 8 ( J dx _
--+-- -0
oy ds ax ds .
(7.60)
.
i.e,
d; =0
ds
,It for C = K.
'I' t ,
(7.61 )
or
i.e. on each boundary If > is a constant. Unlike the case where the section did not
. .
contain holes, we cannot assume. that If > =0on each boundary. We can assume
that t / J =0on one boundary, say on Co, and then determine the corresponding
values of K, on each of the remaining boundaries C/ . To do this, we observe that
the displacement of the section in z direction, i.e. liz: : 8'1' (x, y), from Eq. (7.7),
must be single valued. Consequently, the value of dyt integrated around any
closed contour C, should be equal to zero, i.e.
~cd' l ' = ~c(~ dx +~ dY ) = 0
, 'ox oy
From Eq. (7.9), and using the stress function
rJ rp _ 1 1 8 ( J
ax - GO Tzx +y= GO oy +y
(7.62)
and
0V t _ I 1 8(J
- - -1 " -x=----x
oy GO yt GO ax
Hence, for the single valuedness of liz
__!_Pc ( !t dx _!t dY ) +~c ( y dx - x dy) = 0
GO 'oy iJ x' .
(7.63)
(7.64)
The second integral on the left-hand side is equal to twice the area enclosed by
the contour C; This can be seen from Fig. 7.22(a).
Pc, y dx :: f
GK
/{ y dx +f HLGy dx
: : area G' GKHH' - area H' HLGG'
: : area enclosed by C, =. A;
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
~; Advanced Mechanics of Solids
IUSE,OF 'f.R100NOMETRlC SEDES
- I
,
I."
In many instances, it wiU beuseful to represent tbe deflection curve in the form of
atrigonometric series. We have discussed in Sec. 10.11 that the functions satisfy-
ing Eqs (10.3Ia) and (10.3Ib) also satisfyorthogonality conditions. The trigono-
metric series which we shall consider now is made up of such functions. Let the
deflection curve be represented by
- KX '. 21l'x . n1l'X ( 066)
y = a
l
sin L+~sm L +..;+an SID L' +... 1 .
By properly determining the coefficients aI' ~, ..., the above series can be made to
represent any deflection curve. These coefficients may be calculated by a consider-
ation of the strain energy of the beam or-the column. The strain energy is given by .
1 L( d2 ')2
U=- El fl y) dx
2 0 dx
2
Now,
Hence, the square of the above expression will involve terms of two kinds
4 4
2n1l' 2
nKX
an 4 sm L
and
n
2
m 2 1l'4 . n1l'X . m1l'X
2a
m
a" 4 sin sin --
L L L
By direct integration it can be seen that
L nxx L L
f
. nnx . m1('X .
fsin
2
dx =-, and sin sm dx =0, for n *m
o L 2 0 L L
.
These are the orthogonality relations expressed by Eqs (10.42) and (10.43). Hence,
in the expression for strain energy, terms containing products like am an vanish
and only terms like a: remain. Then
I [ 1('4 L 241l'4 L 3
4
1('4 L ]
U=-EI al
2
.4 -+a; 4 -+a3
2
4 -+...
2 J..,2-L 2 L 2
4
E11(' (2 4 2 4.2 )
= 4I! I a
l
+2 a2 +3 a3 +...
. ,
El1('4 ~ 4 2
= Ln a
43 ,,~I "
Siltlilarly, if we consider the expression
(10.67)
book.
book.
book.
Introductionto Composite Materials - A t " ,

cement and gravel, is a particulate composite; fiberglass, containing glass fibres
embedded in a polymer, is a fibre-reinforced composite; and plywood, having
alternating layers of wood veneer, is alaminar composite. In this chapter, weshall
focus our attention mainly on fibre-reinforced composites and laminates.
!STRES8-STRAIN RELATIONS" =::
" ~' . . ~..'. ~..~, , ', .
. . . . ' . .
. " .
.. ..._', .
In Chapter 3, it was stated that for alinearly elastic body, the stresses are linearly
related to the strains and are given by
O'x =all 6
x
+a
l2
&y +a
l
3 e, +a
l4
r xy +a
is
r yz +a
l
6 r u
(11.1)
't'z:t = a
61
e, +a62 &y +a63 e, +a64 Y xy +a6S ryz +a66 rzx
Assuming that the sixth-order determinant of the coefficients a lJ s in Eq. (11. 1) is
not zero, one can solve for e x ' e
y
, . . . 'Y
zx
in terms of o; O'Y'... ' 'fu. The expressions
for the strain components will then be
e, =14 1 ax +112 a
y
+14 3 a, +14 4 1',ty +14 5 1'yz +14 6 1'.,.
ey =b } ' 1 ax +bn a
y
+~3 at +~4 Txy +~sTyz +b
26
Tzx
e, =b
31
a,t +b:l2 a
y
+~3 a
z
+b:l4 T.ty +b:ls l'yz +~ 1'.,.
r< y =b
41
ax +b
42
a
y
+b
43
a
z
+b
44
T,t}' +b
4S
Tyz +b
46
1'zx
(11.2 )
Ij. . =b
SI
ax +b
S2
a
y
+b
S3
a: +b
S4
Txy +b
ss
Tyz +b
S6
t:zx
Yzx =h61 ax +b
62
a
y
+h63 a
z
+h64 Txy +b
6S
Z"yz +b
66
T.,.
where the coefficients b rJ are related to aijs, The stress-strain relations given by
Eq. (11.1) contain altogether 36 elastic constants. However, this number can be
reduced based on the materia] properties. Let us assume tbat there exists an
elastic potential V such that
. . . , (11.3)
The physical meaning of the potential V will become clear soon. Assuming the
existence of such apotential, fromBq, (1] .3), one obtains
Ou
x
_ iJ
2
V Qa
y
Qax iJ 2
V
_ iJ r ) o/ ,
&y u&xQsy - iJ .s-
x
' Qyxy - Qs
x
Qyxy - Qsx' etc.
FromEqs. (11.4 ) and (11.1), one immediately gets
(11.4 )
Copyright ed mat erial
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
,- ,
398' Advanced Mechanics of Solids
, " .. <
The quantities All' A
22
, etc. are the integrated values (across the thickness) of the
off-axis components of moduli of the laminate.
Evaluation of In-plane ,Moduli The resultant values of the moduli components
are obtained by integrating the moduli component values across the thickness. In
practice, when different plies of finite thicknesses are bonded to get the laminate,
the values of atis change in discrete steps from ply to ply and they are not
continuous functions of z as indicated by Eqs (ll.45b and e). However, continu-
ing the integration sign as used earlier,
h/2
All = f all dz
- } , 1 2
=f ( Q , +Q2 cos 28 +Q3 cos 48) dz
=G Idz +Q2 fcos 2Bdz +Q
3
J eos 48 dz
The Qs for a laminate are constant because of our assumption that the laminate
consists of plies of the same kind having identical material constants, but in the
bonding process, the plies are put with their fibre axes oriented differently, i.e. e
changes from ply to ply, but the Qs are the same for each ply. Thus,
All = Qlh+Q2~ +Q
3
V2 (11.47a)
h/2
VI = f cos 28dz, and
-hI2
Similarly, fromEqs (11.45 band e),
A22 =Q,h - Q2"1 +Q3
V
2
Al2 =~h- Q
3
V2
I
AI4 =_2 Q2J13 - Q3
V
4
A24 =_!Q2V3 - Q3V4
2
A44 =Qsh - Q3
V
2
hl2 h/2
where V3 = I sin28 dz, and V
4
= f sin48 dz (11.47d)
-h/2 -11/2
It was assumed that the laminate consists of even number of plies and are sym-
metrically positioned, i.e. the positioning sequence from the bottom of the lami-
where
hl2
V
2
= I cos 48 dz
- h1 2
(11.47b)
(11.47c)
nate, i.e. from z = _~ to z = 0 is the same as from z = +~ to
z =0, the mid-plane. Hence the limits for the integrals can bechanged to z =0and
z =+; .and the quantities multiplied by 2. Also, itwas explicitly stated that the
orientations e change in finite steps, from ply .0 ply. So, the integration sign can
be changed to summation sign, i.e.
Copyrlqhted matenal
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
book.
A d v a n c e d M e c h a n i c s o f
Th i r d
Ed i t i o n
This comprehensive text on Mechani cs of Sol i ds provides a firm
understanding of the subject as the next step after an
introductory course on Strength of Materials. In-depth treatment
of stress and strain analysis, applications of various strain energy
theorems, thermal stresses, composites, and stress concentration
make this book unique.
Exhaust i ve t r eat ment of
.!~States of. stress ana strain at a..
point and methods to
determine their principal
values
l e Energy methods covering
Theorems of Castigliano,
Virtual work, Kirchhoff,
Menabrea, Least work, and
Engesser
.:. Torsion of multiple cell
sections and shafts with and
without holes
New Sect i ons on
URL:http://www.mhhe.com/srinath/ams3e
~ Tata McGraw-Hili
V i si t us at : w w w .tatamcgraw hl l l .com
I SBN- 13: ' ?~- O- O?- 013' ~~- b
I SBN- 10: O- O?- 013' ~a- l

LS Srinath

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

LS Srinath

Uploaded by

Copyright:

Available Formats

r.=._...;;.

Tells. Comp Shear Tens or Shear Tells Shear per C

(Ii) For 0'2=2+J 3

Example 1.13 C o nsider afunctio n (x , y ), which is called the str essfunctio n.

,14' Advanced Mechanics of Solids

If Q is very close to P, then to first-order

Analysis of Strain :.~t

Analysis of Strain ;"if:

0', ---t:--e=t:__ -_- - - l . _ . . . . . . -.-+--+- 0'1

158' Advanced Mechanics of Solids

Introductionto Composite Materials - A t " ,

You might also like