AE 227: Mechanics of Solids

Contents
1 General motivation 5

2 Deformation of bodies 8
2.1 Testing machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Force (stress)-deformation (strain) relationship . . . . . . . . . . . . 9
2.3 Material microstructures . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3.1 Single crystal . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3.2 Polycrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4 Crystal defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4.1 Point defects . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4.2 Line defects (dislocations) . . . . . . . . . . . . . . . . . . . 12
2.4.3 Surface defects . . . . . . . . . . . . . . . . . . . . . . . . . 12

3 Physical mechanisms of deformation 13

3.0.1 Elastic deformation . . . . . . . . . . . . . . . . . . . . . . . 13
3.0.2 Permanent/Inelastic deformation . . . . . . . . . . . . . . . 13
3.0.3 Deformation responses . . . . . . . . . . . . . . . . . . . . . 14

4 Deformation: Displacements and Strains 16

4.1 General deformation . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.1.1 Rigid motion [2D] . . . . . . . . . . . . . . . . . . . . . . . . 19
4.2 Strain in 2D Cartesian system: geometrical meaning . . . . . . . . . 20
4.3 Different types of strains . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3.1 Simple extension . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3.2 Extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.3.3 Pure shear . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3.4 Simple shear . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.3.5 Rigid rotation . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.3.6 Shear without rotation . . . . . . . . . . . . . . . . . . . . . 31
4.4 Strain in an arbitrary direction . . . . . . . . . . . . . . . . . . . . 32
4.4.1 Strain transformation . . . . . . . . . . . . . . . . . . . . . . 36

1
CONTENTS

4.4.2 Maximum normal strain . . . . . . . . . . . . . . . . . . . . 37

4.4.3 Maximum shear strain . . . . . . . . . . . . . . . . . . . . . 38
4.5 Principal strains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.6 Spherical and deviatoric strains . . . . . . . . . . . . . . . . . . . . 41
4.7 Strain compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.8 Strain in polar coordinates . . . . . . . . . . . . . . . . . . . . . . . 42

5 Concept of stress 46
5.1 Definition of stress . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.2 Stress tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.3 Equilibrium equation in Cartesian system . . . . . . . . . . . . . . 51
5.4 Equilibrium equation in polar system . . . . . . . . . . . . . . . . . 53

6 Stress components on arbitrary plane (2D) 55

6.1 Stress transformation . . . . . . . . . . . . . . . . . . . . . . . . . . 56
6.1.1 Some important transformations . . . . . . . . . . . . . . . . 57
6.2 Maximum normal stress . . . . . . . . . . . . . . . . . . . . . . . . 59
6.3 Maximum shear stress . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.4 Principal stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
6.5 Mohr’s circle for biaxial stresses . . . . . . . . . . . . . . . . . . . . 64

7 Constitutive relations 66
7.0.1 One plane of symmetry (monoclinic) . . . . . . . . . . . . . 67
7.0.2 Three perpendicular planes of symmetry (orthotropic) . . . 68
7.0.3 Axis of symmetry (transversely isotropic material) . . . . . . 68
7.0.4 Complete symmetry (isotropic) . . . . . . . . . . . . . . . . 68
7.1 Physical meaning of elastic moduli . . . . . . . . . . . . . . . . . . 69
7.1.1 Simple tension . . . . . . . . . . . . . . . . . . . . . . . . . . 69
7.1.2 Pure torsion . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
7.1.3 Hydrostatic compression/tension . . . . . . . . . . . . . . . 70

8 Formulation and solution strategies 71

8.1 Boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . 72
8.2 Principle of superposition . . . . . . . . . . . . . . . . . . . . . . . 72
8.3 Saint-Venant principle . . . . . . . . . . . . . . . . . . . . . . . . . 73

9 Two-dimensional formulation 73
9.1 plane strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
9.2 plane stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
9.3 Airy stress function . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
9.3.1 Plane strain . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

AE 227 2 Dr. Krishnendu Haldar

CONTENTS

9.3.2 Plane stress . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

10 Axially loaded bars 80

10.1 Deformation of axially loaded bars . . . . . . . . . . . . . . . . . . 80
10.2 Deflection of a bar due to its own weight . . . . . . . . . . . . . . . 80
10.3 Deflection of a bar due to rotation . . . . . . . . . . . . . . . . . . . 81
10.4 Deflection of two hinge-ended elastic bars . . . . . . . . . . . . . . . 82

11 Torsion 83
11.1 Member of circular cross-section . . . . . . . . . . . . . . . . . . . . 83
11.2 Hollow circular shafts . . . . . . . . . . . . . . . . . . . . . . . . . . 86
11.3 Tapered circular shafts . . . . . . . . . . . . . . . . . . . . . . . . . 87
11.4 Combined torsion and axial force . . . . . . . . . . . . . . . . . . . 88

12 Shearing forces and bending moments of beams 89

12.1 Supports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
12.2 Types of beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
12.2.1 Forces on beam section . . . . . . . . . . . . . . . . . . . . . 91
12.3 Relations between load intensity, shearing force and bending moment 92
12.4 Shearing force and bending moment diagrams . . . . . . . . . . . . 93
12.4.1 Cantilever with end-point load . . . . . . . . . . . . . . . . . 94
12.4.2 Cantilever with number of concentrated loads . . . . . . . . 94
12.4.3 Cantilever with a uniformly distributed load . . . . . . . . . 97
12.4.4 Cantilever with concentrated and distributed load . . . . . . 98
12.4.5 Simply-supported beam with a single concentrated load . . . 100
12.4.6 Simply-supported beam with number of concentrated loads . 103
12.4.7 Simply-supported beam with distributed load . . . . . . . . 105

13 Stress due to bending moment 107

13.1 Theory of simple bending . . . . . . . . . . . . . . . . . . . . . . . 107
13.2 Deformation geometry . . . . . . . . . . . . . . . . . . . . . . . . . 108
13.3 Equilibrium conditions . . . . . . . . . . . . . . . . . . . . . . . . . 110
13.4 Shear stress due to varying bending moment . . . . . . . . . . . . . 112

14 Beam deflection 114

14.1 Differential equation for deflection . . . . . . . . . . . . . . . . . . . 114
14.1.1 Cantilever with end-point load . . . . . . . . . . . . . . . . . 115
14.1.2 Cantilever with a load anywhere in the span . . . . . . . . . 116
14.1.3 Cantilever with a uniformly distributed load . . . . . . . . . 117
14.1.4 Simply-supported beam with a central concentrated load . . 118
14.1.5 Simply-supported beam with distributed load . . . . . . . . 120

AE 227 3 Dr. Krishnendu Haldar

CONTENTS

15 Strain energy 121

15.1 Strain energy due to axial load . . . . . . . . . . . . . . . . . . . . 122
15.2 Strain energy due to Torsion . . . . . . . . . . . . . . . . . . . . . . 122
15.3 Strain energy due to bending moment . . . . . . . . . . . . . . . . . 122
15.4 Strain energy due to shear stress . . . . . . . . . . . . . . . . . . . . 123
15.5 Superposition of elastic energies . . . . . . . . . . . . . . . . . . . . 123
15.5.1 Cantilever with end moment . . . . . . . . . . . . . . . . . . 123
15.5.2 Cantilever with concentrated load at free end . . . . . . . . 124
15.5.3 Cantilever with distributed load . . . . . . . . . . . . . . . . 124
15.5.4 Torsion with variable cross-section . . . . . . . . . . . . . . 124

Appendices 125

A Index notation 125

A.1 Summation convention . . . . . . . . . . . . . . . . . . . . . . . . . 125
A.2 Kronecker delta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
A.3 Permutation symbol . . . . . . . . . . . . . . . . . . . . . . . . . . 126
A.4 Dot product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
A.5 Cross product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
A.6 Dyadic product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
A.7 Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
A.7.1 skw-dual relationship . . . . . . . . . . . . . . . . . . . . . . 129
A.8 Differentiation notation . . . . . . . . . . . . . . . . . . . . . . . . . 129
A.9 Coordinate transformations . . . . . . . . . . . . . . . . . . . . . . 129

B Strain in 2D polar system: geometrical meaning 132

AE 227 4 Dr. Krishnendu Haldar

1 General motivation
Mechanics of solids is a part of continuum mechanics that deals with the changes of
place and shape of things we see every day. One aspect of mechanics represents the
mathematical construction of surrounding objects that composed of earth, water,
air, and fire. It also describes the observation in the laboratories where scientists
perform experiments through an ‘approximate’ theory of nature. Mechanics not
only describes but also predicts phenomena.
Mechanics does not study natural things directly. Instead, it considers bodies,
which are mathematical concepts of some common features of many natural things,
e.g., mass assigned to each body. A body at any instant of time occupies a set of
places or geometry (shape). The change of shape/geometry of a body from one
instant to another is called the motion of the body, and the description of motion
is called kinematics. The motion of a body is conceived as resulting by the action
of forces. Thus mechanics provides a mathematical model for certain aspects of
nature. The mathematical equations that describe the physical laws are called the
field equations.
The assignment of mass m to a body might be discrete, however, in the mechan-
ics of continuum the masses are an absolutely continuous function with volume. A
small volume ∆V enclosed within a surface ∆S possesses a mass density defined
by the following limit:
∆m
ρ = lim ≥ 0,
∆V →0 ∆V

Figure 1: Mass density versus volume.

AE 227 5 Dr. Krishnendu Haldar

 where ∆m is the total mass contained in ∆V . In order to judge the physi-
cal validity of the above mathematical definition, consider the following thought
experiment. Suppose you have a large number of pieces of different volume of a
(homogeneous) material. The mass density for each piece is then calculated from
the ration ∆m/∆V . If the resulting numbers ρ are plotted versus ∆V , one finds
a nearly constant ratio when ∆V is greater than a certain critical volume ∆V ∗
which roughly corresponds to a critical length L ≈ (∆V ∗ )1/3 , which is of course,
not precisely defined. The size of ∆V ∗ depends on the constitution of the material.
As ∆V approaches zero, these dependence becomes extreme (Fig. 1). The clas-
sical continuum theory may therefore not be a good approximation of a physical
theory in the range ∆V << ∆V ∗ . Examples of continuous description, in excel-
lent agreement with the experimental data in the range ∆V >> ∆V ∗ , are those
corresponding to the classical nonlinear and linear theory of solids and fluids.
More involved physical theories based on continuum hypothesis have also been
successful. Among many you can find: theories of magnetic solids and fluids,
piezoelectricity, photo-mechanical etc. where deformations and electromagnetic
fields play prominent roles. The concept of balance equations comes from free-body

Figure 2: Steps of mathematical modeling.

principle. The influence of the external world (by means of volume and surface
forces) on the motion of the body is expressed in terms of balance relations for

AE 227 6 Dr. Krishnendu Haldar

mass, linear momentum, and angular momentum.
The particular behavior patterns of any given material body are described by
statements that have to be formulated individually for each and every material,
whereas the balance relations are true for each and every material. Representations
of material behaviors are known as constitutive equations, which relate between the
processes of deformation and internal forces/stresses. They do not come under the
laws of nature but consists of mathematical models to reproduce typical behavior
patterns of real materials.
The mathematical modeling of material properties from the point of view of
general principles and systematic methods is one of the objectives of the Mechanics
of Solids. Such theories cover elasticity, viscoelasticity, plasticity and viscoplastic-
ity. Every mathematical model evolves from a combination of universal balance
relations with individually valid constitutive equations. The aim of each analysis is
to calculate the fields, which are expressed through systems of partial differential
equations with appropriate initial conditions and boundary conditions. The steps
are shown in Fig. 2.

AE 227 7 Dr. Krishnendu Haldar

2 Deformation of bodies
2.1 Testing machine

(a) (b)

Figure 3: (a) A typical experiment setup and (b) a typical specimen for uniaxial
testing.

The phenomenological method requires the experiments to be made on a volume

element of matter. The most common test is the simple tension test (Fig. 3a) in
which the useful part of the specimen (also known as gage length, Fig. 3b) is
subjected to a uniform uniaxial stress field. The force (or load) could be applied
by a continuous screw-nut system driven by electric motor or hydraulic system.
The specimen consists of a cylindrical shaft, end grips, and between the two,
‘shoulders’, to avoid stress concentration.
During an experiment, the change in gage length is noted as a function of the
applied force. With the same load and a longer gage length, a larger deformation
is observed, than when the gage length is small. Therefore it is more fundamental
to refer to the observed deformation per unit of the gage length. If L0 is the initial
gage length and L is the observed length under a given load, the gage elongation
∆L = L − L0 . The elongation ε per unit of the initial gage length is then given as
L − L0 ∆L
ε= = (1)
L0 L0

AE 227 8 Dr. Krishnendu Haldar

2.2 Force (stress)-deformation (strain) relationship

The above expression defines the extensional strain or normal strain. It is a di-
mensionless quantity and often it is given as a percentage (%). The quantity ε
is generally very small (0.1%) in most of the engineering applications. In this
course, we will confine ourself in this small strain regime. However, in some ap-
plications, e.g. metal forming,1 the strain may be large. For such purposes, one
defines natural or true strain ε. The strain increment dε for such a strain measure
is defined as dL/L, where L is the instantaneous length of the specimen, and dL
is the incremental change in length L. We write
Z L Z L
dL L
ε= dε = = ln = ln(1 + ε). (2)
L0 L0 L L0

2.2 Force (stress)-deformation (strain) relationship

In solid mechanics, the mechanical behavior of real materials under load is of
primary interest. Experiments provide basic information on this behavior. In the
experiments, macroscopic (overall) responses of a specimen to the applied loads
are observed to determine the empirical force-deformation relationships.
Like strain, ‘stress’ is a more significant parameter than force since the effect
on a material of an applied force P depends primarily on the cross-sectional area of
the member. Let us quickly define stress as force per unit area. The stress measure
corresponding to the actual undeformed cross-sectional area is known as nominal
stress, where true stress consider the current cross-sectional area (Fig. 4a). We’ll
study the concept of stress with more detail in the coming sections.
In the experimental study, it is customary to plot diagrams of the relationship
between stress and strain in a particular test. Such plots are, in general, inde-
pendent of the size of the specimen and its gage length. It is customary to use
the ordinate scale for the stress and abscissa for the strain. Experimentally deter-
mined stress-strain diagrams differ widely for different materials (Fig. 4b). It is
the material microstructures that are responsible for such diverse responses. We’ll
now look briefly some microstructural aspect of mechanical responses.

2.3 Material microstructures

2.3.1 Single crystal
The crystalline state is characterized by the regularity of the atomic arrangement.
An elementary pattern or unit cell (Fig. 5) repeats itself periodically in all the
three directions to form a space lattice, an infinite array of points (Fig. 6). A
1
It is a metalworking process where the workpiece is reshaped without adding or removing
material through large deformation.

AE 227 9 Dr. Krishnendu Haldar

2.3 Material microstructures

(a) (b)

Figure 4: (a) True stress vs nominal stress and (b) different material responses.

Figure 5: BCC and FCC unit cell.

crystalline solid can be a single crystal, where the entire solid consists of only one
crystal.

2.3.2 Polycrystals
Metals are generally produced in a liquid state, and their structure is formed
as they solidify when cooled. As the temperature of the liquid decreases, the
interatomic distances become smaller. The critical distance, at which bonding
occurs, is reached at several, randomly distributed sites, and constitute the first
germs or nuclei of crystal growth. The lattices are formed in the same crystalline

AE 227 10 Dr. Krishnendu Haldar

2.4 Crystal defects

Figure 6: Representation of space lattice with unit cell.

Figure 7: Randomly oriented single crystals in a polycrystalline material.

system but in random directions. A polycrystalline material is made up of several

single-crystals oriented randomly (Fig. 7).

2.4 Crystal defects

2.4.1 Point defects
As the name suggests, they are imperfect point-like regions in the crystal. Point
defects consist of atoms inserted or substituted or missing (one or two atomic
diameters) in the crystal. They result in a local distortion of the lattice.

AE 227 11 Dr. Krishnendu Haldar

2.4 Crystal defects

2.4.2 Line defects (dislocations)

Line defects, a one-dimensional imperfections in geometrical sense, are called dis-
locations. These are the defects which are responsible for plasticity of metals. Let
us consider a schematic representation of such defects in the case of simple cubic
crystals. An edge dislocation is created by the translation of the upper part of the
crystal (Fig. 8a). A screw dislocation is created by a local rotation of the upper
part of the crystal (Fig. 8b). Dislocations are created during the growth of the
crystals.

(a) (b)

Figure 8: (a) Edge dislocation and (b) screw dislocation.

2.4.3 Surface defects

Surface defects are the surfaces of separation between crystals or parts of a crystal
with different orientations. Their thickness is of the order of 4-5 atomic ‘diameters’.
For example, grain boundaries (or crystal boundaries) in polycrystals, interfaces
between two phases, twin boundaries, etc.

AE 227 12 Dr. Krishnendu Haldar

3 Physical mechanisms of deformation
3.0.1 Elastic deformation
Elastic deformations occur at the atomic level. The observed macroscopic effect
is the result of the variations in the interatomic spacing necessary to balance
the external loads, and also the reversible movement of the dislocations. These
geometrical adjustments are essentially reversible. In a purely elastic deformation,
the initial configuration of atoms is restored upon the removal of the load.

3.0.2 Permanent/Inelastic deformation

Inelastic deformations (plastic and viscoplastic) occur at the crystal level. In
addition to the elastic deformations, inelastic deformation corresponds to a relative
displacement of atoms which remains when the load is removed. Depending on
the case, the deformations are either purely intragranular (inside the grains) or
intergranular displacements.

Deformation by slip
Symmetry planes of the crystal lattice, which are also the planes of most densely
packed atoms, form the parallel planes with the greatest distance between them.
It is, therefore, these planes that slip due to shear can occur in the direction of
maximum shear stress, e.g., planes (1, 1, 1) in FCC crystals. They occur in the
form of parallel slip bands which result in steps on the exterior surface of the
sample.

Figure 9: Permanent displacement by slip.

Deformation by dislocation
The presence of dislocations considerably reduces the stability of the crystal lattice.
Their mobility is the essential cause of permanent deformations, homogeneous at

AE 227 13 Dr. Krishnendu Haldar

the macroscopic scale. An irreversible displacement occurs when an edge or screw
dislocation moves across the crystal. This slip displacement mechanism requires
the breaking of bonds only in the vicinity of the dislocation line, and successively
from one atom to the next. An edge dislocation can also move perpendicularly to
its slip plane. This is also known as climb displacement.
The rate of dislocation displacement (slip or climb) can be very low or very
high depending on the applied stress, but it can not be higher than the speed of
sound in the material under consideration.

3.0.3 Deformation responses

Figure 10: A typical stress-strain response.

OY is the elastic deformation (Fig. 10). Point Y is the elastic limit. The
elastic limit is characterized by the state of stress or strain which causes the first
irreversible movements of dislocations. The plastic deformation starts after Y.
YP is the plastic deformation zone. If the stress continues to rise, the slips can
cross and follow the grain boundaries. This phenomenon of inter-granular slip is
favored by thermal activation. This is the domain of visco-plasticity (PV). Upon
unloading, we end up with a permanent residual plastic/viscoelastic strain (OR).
If one continues loading after point V, the stress-strain slop gradually decreases
and levels off at D. This point corresponds to the ultimate tensile strength, where
the deformation becomes highly localized and the state of stress changes from
uniaxial to biaxial at the surface, and triaxial in the interior of the specimen. There
is also pronounced lateral contraction. Thus the area of cross-section decreases and
the stress response starts dropping until the specimen fractures at F.

AE 227 14 Dr. Krishnendu Haldar

 Figure 11: Brittle and ductile material response.

A brittle material breaks without significant plastic deformation (ex: cast iron,
concrete, and some glass products). Brittle materials fracture by cleavage or sepa-
ration and it breaks suddenly. Breaking is often accompanied by a snapping sound.
Ductile materials show a prominent plastic zone. They can be bent, twist or role
significantly before reaching the fracture point F. These failures occur primarily
due to the slip in shear and a typical ‘cup and cone’ fracture is detected.

AE 227 15 Dr. Krishnendu Haldar

4 Deformation: Displacements and Strains
A solid body changes shape or deforms when forces (loads) are applied. Such
deformations can be quantified by knowing the displacements of material points
within the body. Thus the continuum hypothesis establishes a displacement field.
The purpose of this section is to introduce the basic definition of displacement and
strain.

4.1 General deformation

Figure 12: Schematic of reference and deformed configuration.

Let us consider a generic point P0 inside the undeformed body, i.e., no external
forces/other stimuli are applied. The position vector of this point is denoted by
RP = xî+y ĵ +z k̂, where {î, ĵ, k̂} are the unit vectors of a Cartesian system. Now
we apply forces on the body and it deforms to a new configuration (red contour).
Let the point P0 moves to its new position P and the position vector is given by
r P = x0 î + y 0 ĵ + z 0 k̂. The displacement of point P is thus given by

uP = u(x, y, z) = u(x, y, z)î + v(x, y, z)ĵ + w(x, y, z)k̂ = r P − RP . (3)

AE 227 16 Dr. Krishnendu Haldar

4.1 General deformation

We denote:

u1 → u
u2 → v
u3 → w

Consider another point Q0 , infinitesimal close to point P0 in the undeformed con-

figuration. The coordinate of the point Q0 is (x + ∆x, y + ∆y, z + ∆z) and it moves
to point Q after deformation. The displacement of the point Q is then given by

uQ = u(x + ∆x, y + ∆y, z + ∆z) = u(x + ∆x, y + ∆y, z + ∆z)î

+ v(x + ∆x, y + ∆y, z + ∆z)ĵ
+ w(x + ∆x, y + ∆y, z + ∆z)k̂
= r Q − RQ . (4)

The change in displacement field reads ∆u = uQ − uP , or

∆u = ∆u(x, y, z)î + ∆v(x, y, z)ĵ + ∆w(x, y, z)k̂

h i
= u(x + ∆x, y + ∆y, z + ∆z) − u(x, y, z) î
h i
+ v(x + ∆x, y + ∆y, z + ∆z) − v(x, y, z) ĵ
h i
+ w(x + ∆x, y + ∆y, z + ∆z) − w(x, y, z) k̂
h ∂u ∂u ∂u i
= ∆x + ∆y + ∆z î
∂x ∂y ∂z
h ∂v ∂v ∂v i
+ ∆x + ∆y + ∆z ĵ
∂x ∂y ∂z
h ∂w ∂w ∂w i
+ ∆x + ∆y + ∆z k̂.
∂x ∂y ∂z
In the matrix notation we write:
   ∂u ∂u ∂u   
∆u ∂x ∂y ∂z ∆x
    
 ∆v  =  ∂v ∂v ∂v
    
 ∆y  (5)
   ∂x ∂y ∂z  
    
∆w ∂w ∂w ∂w ∆z
∂x ∂y ∂z

Now,

∆u = uQ − uP = (r Q − RQ ) − (r P − RP ) = (r Q − r P ) − (RQ − RP )
= P Q − P0 Q0 .

AE 227 17 Dr. Krishnendu Haldar

4.1 General deformation

This means, the change in displacement is the vector difference (P Q − P0 Q0 ).

Moreover, P0 Q0 = ∆x î + ∆y ĵ + ∆z k̂. If we denote
 ∂u ∂u ∂u 
∂x ∂y ∂z
 
     ∂v ∂v ∂v

∂j ui = ui,j = 
 ∂x ∂y ∂z

 (6)
 
∂w ∂w ∂w
∂x ∂y ∂z

The tensor [ui,j ] is known as the displacement gradient tensor that connects ∆u
to the initial line segment P0 Q0 .
The symmetric part of the displacement gradient tensor, ui,j is written as
1
εij = (ui,j + uj,i ) (7)
2
and these are the components of the strain tensor. Then,
∂u1 ∂u
εx = ε11 = =
∂x1 ∂x
∂u2 ∂v
εy = ε22 = =
∂x2 ∂y
∂u3 ∂w
εz = ε33 = =
∂x3 ∂z
1  ∂u1 ∂u2  1  ∂u ∂v 
εxy = ε12 = + = +
2 ∂x2 ∂x1 2 ∂y ∂x
1  ∂u2 ∂u3  1  ∂v ∂w 
εyz = ε23 = + = +
2 ∂x3 ∂x2 2 ∂z ∂y
1  ∂u3 ∂u1  1 ∂w ∂u 

εzx = ε31 = + = +
2 ∂x1 ∂x3 2 ∂x ∂z
The skew-symmetric part is written as:
1
ωij = (ui,j − uj,i ). (8)
2
and known as rotational tensor. However, rotation tensor does not contribute in
stretching or shearing. Now a dual vector wi can be associated with the rotation
such that wi = − 21 εijk ωjk = 12 εijk ∂j uk . Note that
 
0 −w3 w2
ωij =  w3 0 −w1  ,
−w2 w1 0

AE 227 18 Dr. Krishnendu Haldar

4.1 General deformation

then
1  ∂u2 ∂u3  1  ∂w ∂v 
wx = −ω23 = − − = − (9)
2 ∂x3 ∂x2 2 ∂y ∂z
1 ∂u1 ∂u3
  1 ∂u ∂w
 
wy = ω13 = − = − (10)
2 ∂x3 ∂x1 2 ∂z ∂x
1 ∂u1 ∂u2
  1 ∂v ∂u 

wz = −ω12 = − − = − (11)
2 ∂x2 ∂x1 2 ∂x ∂y
For small-strain theory, a rigid rotation takes place if all the strain components
are zero.

4.1.1 Rigid motion [2D]

Let a 2D body has a constant rotational velocity wz . Then
1  ∂v ∂u 
wz = − (12)
2 ∂x ∂y
For a rigid motion, all the strain components are zero. This means
∂u
εx = = 0, =⇒ u = c1 + f (y) (13)
∂x
∂v
εy = = 0, =⇒ v = c2 + g(x). (14)
∂y
Here c1 , c2 are arbitrary constants. Moreover,
1  ∂u ∂v 
εxy = + =0 (15)
2 ∂y ∂x
Plugging the relations for u and v in the above equation, we can write
df dg
+ =0 (16)
dy dx
From (12) we write:
dg df
− = 2wz (17)
dx dy
Adding (17) and (16) we get,
dg
= wz
dx
g = wz x + c3 (18)

AE 227 19 Dr. Krishnendu Haldar

4.2 Strain in 2D Cartesian system: geometrical meaning

Then

v = c2 + (wz x + c3 )
= (c2 + c3 ) + wz x
= v0 + w z x

Similarly, substracting (17) and (16) we get,

df
= −wz
dy
f = −wz y + c4 (19)

Then

u = c1 + (−wz y + c4 )
= (c1 + c4 ) − wz y
= u0 − wz y

So, the rigid body displacement fields are

u = u0 − wz y (20)
v = v0 + wz x (21)

Here u0 , v0 are arbitrary constants along x and y directions, respectively.

4.2 Strain in 2D Cartesian system: geometrical meaning

Let us now consider two infinitesimal perpendicular line segments P Q and P R
inside the material body in the undeformed configuration as shown in Fig. 13.
The contour of the body, like the previous section, is not shown for clarity. The
coordinate of P is (x, y) and that of Q and R are (x + ∆x, y) and (x, y + ∆y),
respectively. After deformation, the three points P, Q and R move to P 0 , Q0 and
R0 , respectively. Moreover, the new line segments P 0 Q0 and P 0 R0 make angles θ1
and θ2 with respect to the horizontal and vertical axes, respectively. Then we have
the following definitions of strains:

P 0 Q0 − P Q P 0 R0 − P R
εx = , εy = , γxy = π/2 − ∠Q0 P 0 R0 = θ1 + θ2 . (22)
PQ PR
εx is the normal strain in the x-direction, εy is the normal strain in the y-direction,
and γxy is the engineering shear strain, defined as the change of right angle due to
the deformation. Our goal will be to calculate the new line segments P 0 Q0 = |P 0 Q0 |

AE 227 20 Dr. Krishnendu Haldar

4.2 Strain in 2D Cartesian system: geometrical meaning

Figure 13: 2D geometric deformation.

and P 0 R0 = |P 0 R0 |. We’ll now compute them through vector approach. In this

context, let us now consider the position vectors of the six points P, Q, R, and
P 0 , Q0 , R0 as presented in Fig. 14. The displacement fields of the three points
P, Q, R are:

uP = u(x, y) = u(x, y)î + v(x, y)ĵ

uQ = u(x + ∆x, y) = u(x + ∆x, y)î + v(x + ∆x, y)ĵ
uR = u(x, y + ∆y) = u(x, y + ∆y)î + v(x, y + ∆y)ĵ. (23)

Now,

P Q = OQ − OP = ∆x î
P 0 Q0 = OQ0 − OP 0 (24)

where,

OP 0 = OP
h + PiP 0 h i
= xî + y ĵ + u(x, y)î + v(x, y)ĵ
h i h i
= x + u(x, y) î + y + v(x, y) ĵ

AE 227 21 Dr. Krishnendu Haldar

4.2 Strain in 2D Cartesian system: geometrical meaning

Figure 14: Position and displacement vectors of different points.

and

OQ0 = OQ + QQ0
h i h i
= (x + ∆x)î + y ĵ + u(x + ∆x, y)î + v(x + ∆x, y)ĵ
h i h i
= x + ∆x + u(x + ∆x, y) î + y + v(x + ∆x, y) ĵ
h ∂u i h ∂v i
≈ x + ∆x + u(x, y) + ∆x î + y + v(x, y) + ∆x ĵ
∂x ∂x

AE 227 22 Dr. Krishnendu Haldar

4.2 Strain in 2D Cartesian system: geometrical meaning

So,

P 0 Q0 = OQ0 − OP 0
h ∂u i h ∂v i
= ∆x + ∆x î + ∆x ĵ
r ∂x ∂x
 ∂u   ∂u 2  ∂v 2
P 0 Q0 = |P 0 Q0 | = 1+2 + + ∆x
r ∂x ∂x ∂x
 ∂u 
≈ 1+2 ∆x
∂x
 ∂u 
≈ 1+ ∆x
∂x
(25)

Then,
 
∂u
0 0
P Q − PQ 1+ ∂x
∆x − ∆x ∂u
εx = = = (26)
PQ ∆x ∂x
In a similar procedure,

P R = OR − OP = ∆y ĵ
P 0 R0 = OR0 − OP 0 (27)

where,

OP 0 = OP
h + PiP 0 h i
= xî + y ĵ + u(x, y)î + v(x, y)ĵ
h i h i
= x + u(x, y) î + y + v(x, y) ĵ

and

OR0 = OR
h + RR
0
i h i
= xî + (y + ∆y)ĵ + u(x, y + ∆y)î + v(x, y + ∆y)ĵ
h i h i
= x + u(x, y + ∆y) î + y + ∆y + v(x, y + ∆y) ĵ
h ∂u i h ∂v i
≈ x + u(x, y) + ∆y î + y + ∆y + v(x, y) + ∆y ĵ
∂y ∂y

AE 227 23 Dr. Krishnendu Haldar

4.2 Strain in 2D Cartesian system: geometrical meaning

So,

P 0 R0 = OR0 − OR0
h ∂u i h ∂v i
= ∆y î + ∆y + ∆y ĵ
∂y ∂y
s
 ∂v   ∂v 2  ∂u 2
P 0 R0 = |P 0 R0 | = 1+2 + + ∆y
∂y ∂y ∂y
s
 ∂v 
≈ 1+2 ∆y
∂y
 ∂v 
≈ 1+ ∆y
∂y
(28)

Then,
 
∂v
0 0
P R − PR 1+ ∂y
∆y − ∆y ∂v
εy = = = (29)
PR ∆y ∂y
For the angle θ1 , we consider

î × P 0 Q0 = |î||P 0 Q0 | sin θ1 k̂
h ∂v i  ∂u 
∆x k̂ = 1 · 1 + ∆x · θ1 k̂
∂x ∂x
h ∂v i ∂u −1 h ∂v i ∂u  ∂v
⇒ θ1 = 1+ ≈ 1− ≈ (30)
∂x ∂x ∂x ∂x ∂x
Similarly, for the angle θ2 ,

ĵ × P 0 R0 = |î||P 0 R0 | sin θ2 (−k̂)

h ∂u i  ∂v 
− ∆y k̂ = −1 · 1 + ∆y · θ2 k̂
∂y ∂y
h ∂u i ∂v −1 h ∂u i ∂v  ∂u
⇒ θ2 = 1+ ≈ 1− ≈ (31)
∂y ∂y ∂y ∂y ∂y
So, the shear strain (engineering)
 ∂u ∂v 
γxy = π/2 − ∠Q0 P 0 R0 = θ1 + θ2 = + . (32)
∂y ∂x
∂v ∂u
• Note-1: If there is no rotation, then ω3 = 0, i.e., ∂x
= ∂y
. So θ1 = θ2 .
∂v
• Note-2: If there is no strain (rigid rotation), then ω3 = ∂x
= − ∂u
∂y
.

AE 227 24 Dr. Krishnendu Haldar

4.3 Different types of strains

4.3 Different types of strains

Homogeneous strain: consider the displacement components as

u = a1 x + a2 y + a3 z
v = b1 x + b 2 y + b 3 z
w = c1 x + c2 y + c3 z.

Then the displacement gradient is given by:

 
a1 a2 a3
[ui,j ] = b1
 b2 b3  . (33)
c1 c2 c3

So the displacement gradient is independent of the coordinates (x, y, z). This

means strain components and the rotation components are independent of the
coordinates and same at every material points.

4.3.1 Simple extension

Figure 15: Simple extension.

The displacement components in a simple extension along the x− axis are

u = c1 x
v = 0
w = 0.

AE 227 25 Dr. Krishnendu Haldar

4.3 Different types of strains

c1 > 0 is the stretching/extension and c1 < 0 is the compression. The new position
of a point (x, y, z) is given by

x̃ = x + u = (1 + c1 )x
ỹ = y
z̃ = z.

The displacement gradient is given by:

 
c1 0 0
[ui,j ] =  0 0 0 (34)
0 0 0
so,
   
c1 0 0 0 0 0
[εij ] =  0 0 0 , [ωij ] = 0 0 0 (35)
0 0 0 0 0 0

This means only the axial strain εx = c1 is nonzero. There is no rotation. Similarly
the displacement components in a simple extension along the y− axis are

u = 0
v = c2 y
w = 0.

and along the z− axis are

u = 0
v = 0
w = c3 z.

4.3.2 Extension
Extension (or compression) occurs in all directions or in two directions or in one
direction (simple extension). The displacement components are

u = c1 x
v = c2 y
w = c3 z.

AE 227 26 Dr. Krishnendu Haldar

4.3 Different types of strains

Figure 16: Extension in x and y plane.

The new position of a point (x, y, z) is given by

x̃ = x + u = (1 + c1 )x
ỹ = y + v = (1 + c2 )y
z̃ = z + w = (1 + c3 )z.

The displacement gradient is given by:

 
c1 0 0
[ui,j ] =  0 c2 0  (36)
0 0 c3
so,
   
c1 0 0 0 0 0
[εij ] =  0 c2 0  , [ωij ] = 0 0 0 (37)
0 0 c3 0 0 0

This means only the axial strains εx = c1 , εy = c2 , and εz = c3 are nonzero. There
is no rotation.

AE 227 27 Dr. Krishnendu Haldar

4.3 Different types of strains

Figure 17: Pure shear.

4.3.3 Pure shear

In pure shear, the displacement components are
u = c1 x
v = c2 y
w = 0.
The new position of a point (x, y, z) is then given by
x̃ = x + u = (1 + c1 )x
ỹ = y + v = (1 + c2 )y
z̃ = z + w = z.
The displacement gradient is given by:
 
c1 0 0
[ui,j ] =  0 c2 0 (38)
0 0 0
so,
   
c1 0 0 0 0 0
[εij ] =  0 c2 0 , [ωij ] = 0 0 0 (39)
0 0 0 0 0 0
This means only the axial strains εx = c1 and εy = c2 are nonzero. There is no
rotation. However, for pure shear the total volume of the body remain constant.

AE 227 28 Dr. Krishnendu Haldar

4.3 Different types of strains

Considering unit thickness along the z-direction, the volume of the body in the
initial configuration is xy and in the final configuration x̃ỹ. Then,

(1 + c1 )(1 + c2 )xy = xy
c1 + c2 + c1 c2 = 0
−c1
c2 =
1 + c1
So, when c1 > 0, i.e., stretching in one direction, then the y dimension shrinks.

4.3.4 Simple shear

Figure 18: Simple shear.

In simple shear, the displacement components are

u = c1 y
v = 0
w = 0.

The new position of a point (x, y, z) is then given by

x̃ = x + u = x + c1 y
ỹ = y + v = y
z̃ = z + w = z.

AE 227 29 Dr. Krishnendu Haldar

4.3 Different types of strains

The displacement gradient is given by:

 
0 c1 0
[ui,j ] = 0 0 0 (40)
0 0 0
so,
   
0 c1 /2 0 0 c1 /2 0
[εij ] = 0 0 0 , [ωij ] = 0 0 0 (41)
0 0 0 0 0 0
This means only the shear strain εxy = c1 /2 is nonzero. There non-zero rotation
component is ωz = −c1 /2. However, for pure shear the total volume of the body
remain constant. Note that the volume is preserved (isochoric).

4.3.5 Rigid rotation

Figure 19: Rigid rotation.

As we derived earlier, in rigid rotation (without ant translation), the displace-

ment components are
u = −ωz y
v = ωz x
w = 0.

AE 227 30 Dr. Krishnendu Haldar

4.3 Different types of strains

The new position of a point (x, y, z) is then given by

x̃ = x + u = x − ωz y
ỹ = y + v = y + ωz x
z̃ = z + w = z.

Let us consider the coordinates of the four corner points of the bar are A = (0, 0),
B = (l, 0), C = (l, h) and D = (0, h). Then

A0 = (0, 0)
B0 = (l, ωz l)
C0 = (l − ωz h, h + ωz l)
D0 = (−ωz h, h)
(42)

4.3.6 Shear without rotation

Figure 20: Shear without rotation.

As we derived earlier, in rigid rotation (without ant translation), the displace-

AE 227 31 Dr. Krishnendu Haldar

4.4 Strain in an arbitrary direction

ment components are

u = c1 y
v = c1 x
w = 0.

The new position of a point (x, y, z) is then given by

x̃ = x + u = x + c1 y
ỹ = y + v = y + c1 x
z̃ = z + w = z.

Let us consider the coordinates of the four corner points of the bar are A = (0, 0),
B = (l, 0), C = (l, h) and D = (0, h). Then

A0 = (0, 0)
B0 = (l, c1 l)
C0 = (l + c1 h, h + c1 l)
D0 = (c1 h, h)
(43)

The displacement gradient is given by:

 
0 c1 0
[ui,j ] = c1 0 0 (44)
0 0 0
so,
   
0 c1 0 0 0 0
[εij ] = c1 0 0 , [ωij ] = 0 0 0 (45)
0 0 0 0 0 0

This means only the shear strain εxy = c1 is nonzero and there is no rotation.

4.4 Strain in an arbitrary direction

Let M be a point with coordinates (x, y) and N is a neighboring point with
coordinates (x + ∆x, y + ∆y). Let the strain components at M be εx , εy and εxy
(or γxy ). Now let the element M N suffers a deformation and moves to its new

AE 227 32 Dr. Krishnendu Haldar

4.4 Strain in an arbitrary direction

Figure 21: Strain in the arbitrary direction.

orientation M 0 N 0 . Then we write:

OM = x î + y ĵ
ON = (x + ∆x) î + (y + ∆y) ĵ
OM 0 = (x + u(x, y)) î + (y + v(x, y)) ĵ
ON 0 = (x + ∆x + u(x + ∆x, y + ∆y)) î + (y + ∆y + v(x + ∆x, y + ∆y)) ĵ
So,
M N = ON − OM = ∆x î + ∆y ĵ
h ∂u i ∂u   ∂v h ∂v i 
M 0 N 0 = ON 0 − OM 0 = 1+ ∆x + ∆y î + ∆x + 1 + ∆y ĵ
∂x ∂y ∂x ∂y
Now,

0 0 2
h ∂u i ∂u 2  ∂v h ∂v i 2
|M N | = 1+ ∆x + ∆y + ∆x + 1 + ∆y
∂x ∂y ∂x ∂y
h ∂u i2 2 h ∂u i ∂u  ∂u 2 
= 1+ ∆x + 2 1 + ∆y∆x + ∆y 2
∂x ∂x ∂y ∂y
 ∂v 2 h ∂v ∂v
i h ∂v i 2 
+ ∆x2 + 2 1 + ∆x∆y + 1 + 2 ∆y
∂x ∂y ∂x ∂y
(46)
As we are using a linearized approximation, i.e., no second order terms in a measure
of geometric quantities, we will neglect fourth order terms in |M 0 N 0 |2 . Thus we

AE 227 33 Dr. Krishnendu Haldar

4.4 Strain in an arbitrary direction

get:
h ∂u i 2 ∂u   ∂v h ∂v i 2 
|M 0 N 0 |2 ≈ 1+2 ∆x + 2 ∆y∆x + 2 ∆x∆y + 1 + 2 ∆y
∂x ∂y ∂x ∂y
|M 0 N 0 |2 h ∂u i ∆x2 h ∂u ∂v i ∆y ∆x  h ∂v i ∆y 2 
= 1+2 + 2 + + 1 + 2
|M N |2 ∂x |M N |2 ∂y ∂x |M N | |M N | ∂y |M N |2
h ∂u i 2 h ∂u ∂v i  h ∂v i 2 
= 1+2 cos θ + 2 + sin θ cos θ + 1 + 2 sin θ
∂x ∂y ∂x ∂y
 ∂u 2
h ∂u ∂v i ∂v 2

= 1 + 2 cos θ + + sin 2θ + 2 sin θ
∂x ∂y ∂x ∂y
 
2 2
= 1 + 2εx cos θ + 2εxy sin 2θ + 2εy sin θ
 h i
= 1 + 2 εx cos2 θ + εxy sin 2θ + εy sin2 θ

Then,
r
|M 0 N 0 | h
2 2
i
= 1 + 2 εx cos θ + εxy sin 2θ + εy sin θ
|M N |
h i
≈ 1 + εx cos2 θ + εxy sin 2θ + εy sin2 θ

So,
|M 0 N 0 | − |M N |
εθ = = εx cos2 θ + εxy sin 2θ + εy sin2 θ (47)
|M N |

Now if ∆θ be the angle between the lines M N and M 0 N 0 , then we can write

M N × M 0 N 0 = |M N ||M 0 N 0 | sin ∆θ k̂
MN M 0N 0 |M 0 N 0 |
× ≈ ∆θ k̂
|M N | |M N | |M N |
MN M 0N 0
× ≈ ∆θ k̂
|M N | |M N |
Now,
M 0N 0 h ∂u i ∆x ∂u ∆y   ∂v ∆x h ∂v i ∆y 
= 1+ + î + + 1+ ĵ
|M N | ∂x |M N | ∂y |M N | ∂x |M N | ∂y |M N |
h ∂u i ∂u   ∂v h ∂v i 
= 1+ cos θ + sin θ î + cos θ + 1 + sin θ ĵ
∂x ∂y ∂x ∂y
MN
= cos θ î + sin θ ĵ (48)
|M N |

AE 227 34 Dr. Krishnendu Haldar

4.4 Strain in an arbitrary direction

Then the cross product |M N × M 0 N 0 becomes:

MN| |M N |
 

 î ĵ k̂
i cos θ sin
h θ  0 
 
h   i
 1 + ∂u cos θ + ∂u sin θ ∂v
cos θ + 1 + ∂y ∂v
sin θ 0

∂x ∂y ∂x
" #
 ∂v 1 h ∂v i  h ∂u i ∂u 
= cos2 θ + 1 + sin 2θ − 1 + sin 2θ + sin2 θ k̂
∂x 2 ∂y ∂x ∂y
" #
∂v 1 h ∂v ∂u i ∂u 2
= cos2 θ + − sin 2θ − sin θ k̂
∂x 2 ∂y ∂x ∂y
(49)
We then obtain,

Figure 22: Change in right angle, i.e., γ θ = π

2
− ∠N 0 M 0 P 0 . Now ∠N 0 M 0 P 0 =
( π2 − ∆θ) + ∆φ. So, γ θ = ∆θ − ∆φ.

∂v 1 h ∂v ∂u i ∂u 2
∆θ = cos2 θ + − sin 2θ − sin θ (50)
∂x 2 ∂y ∂x ∂y
Change in angle ∆φ for a line element, perpendicular to the initial line M N is
obtained by replacing θ ≡ θ + 90◦ . Then,
∂v 1 h ∂v ∂u i ∂u
∆φ = sin2 θ − − sin 2θ − cos2 θ (51)
∂x 2 ∂y ∂x ∂y

AE 227 35 Dr. Krishnendu Haldar

4.4 Strain in an arbitrary direction

So the change in right angle is given by:

∂v h ∂v ∂u i ∂u
∆θ − ∆φ = cos 2θ + − sin 2θ + cos 2θ
∂x ∂y ∂x ∂y
 ∂v ∂u  h ∂v ∂u i
= + cos 2θ + − sin 2θ (52)
∂x ∂y ∂y ∂x
Or
γ θ = 2εxy cos 2θ + (εy − εx ) sin 2θ (53)

4.4.1 Strain transformation

Let εx , εy and γxy be the strain components at a material point with respect to a
Cartesian rectangular coordinate system Oxy, where O is the origin. Let Ox0 y 0 be
another (Cartesian) frame of reference inclined at an angle θ to Oxy. To determine
0
ε0x , ε0y and γxy in terms of εx , εy , γxy and θ, from (47) we can immediately write
εθ = ε0x = εx cos2 θ + εxy sin 2θ + εy sin2 θ (54)
◦
ε(θ+90 ) = ε0y = εx sin2 θ − εxy sin 2θ + εy cos2 θ (55)
and from (53)
0
γ θ = γxy = 2εxy cos 2θ + (εy − εx ) sin 2θ (56)
1
=⇒ ε0xy = εxy cos 2θ + (εy − εx ) sin 2θ (57)
2
Recall that εxy = 12 γxy . The transformation also follows the following standard
rules of second order tensor transformation:
ε0ij = Qpi Qqj εpq (58)
where, the rotation matrix Qmn = cos (xm , x0n ). In tensor notation we further write
ε0 = QεQt . (59)
For example, in 2D Cartesian system:
 
cos θ sin θ
[Q] = (60)
− sin θ cos θ
Then,
 0
εx ε0xy
    
cos θ sin θ εx εxy cos θ − sin θ
=
ε0xy ε0y − sin θ cos θ εxy εy sin θ cos θ
  
cos θ sin θ εx cos θ + εxy sin θ −εx sin θ + εxy cos θ
=
− sin θ cos θ εxy cos θ + εy sin θ −εxy sin θ + εy cos θ
εx cos2 θ + εxy sin 2θ + εy sin2 θ 21 (εy − εx ) sin 2θ + εxy cos 2θ
 
= 1
(ε − εx ) sin 2θ + εxy cos 2θ εx sin2 θ − εxy sin 2θ + εy cos2 θ
2 y

AE 227 36 Dr. Krishnendu Haldar

4.4 Strain in an arbitrary direction

So we obtain,

ε0x = εx cos2 θ + εxy sin 2θ + εy sin2 θ (61)

1
ε0xy = (εy − εx ) sin 2θ + εxy cos 2θ (62)
2
ε0y = εx sin2 θ − εxy sin 2θ + εy cos2 θ (63)

4.4.2 Maximum normal strain

dεθ
We are looking for the theta such that dθ
= 0. So,

dεθ
= −(εx − εy ) sin 2θ + 2εxy cos 2θ = 0 (64)
dθ
Or
2εxy
tan 2θ = . (65)
εx − εy
The value of θ defines the plane on which the normal strain attains an extremum,
i.e., a maximum or a minimum. The above equations gives two values of θ, since
tan 2θ = tan (2θ + 180◦ ). If θ1 and θ2 are these two values, we then have

θ2 = θ1 + 90◦ .

εθ is maximum on one of these angles and a minimum on the other. Consider now
the angle on which γ θ = 0. This gives
2εxy
tan 2θ = . (66)
εx − εy
In other words, the angles on which the normal strain is an extremum also have
zero shear strain. Angles (directions) on which the normal strains have maximum
or minimum values are called principal directions and the strain magnitudes on
these planes are the principal strains. We rewrite (61) and (62) in double angle
form as
1 1
εθ = (εx + εy ) + (εx − εy ) cos 2θ + εxy sin 2θ (67)
2 2
γθ = −(εx − εy ) sin 2θ + 2εxy cos 2θ (68)

From (65) we can write:

2εxy εx − εy
sin 2θ = q , cos 2θ = q (69)
(εx − εy )2 + 4ε2xy (εx − εy )2 + 4ε2xy

AE 227 37 Dr. Krishnendu Haldar

4.4 Strain in an arbitrary direction

Or
2εxy εx − εy
sin 2θ = − q , cos 2θ = − q (70)
(εx − εy )2 + 4ε2xy (εx − εy )2 + 4ε2xy

Substituting back in (67) and (68) we get

r
εx + εy εx − εy  2
ε1,2 = ± + ε2xy (71)
2 2
γ1,2 = 0 (72)

4.4.3 Maximum shear strain

Let us now determine the planes on which the shear strain reaches an extremum
value by differentiation γ φ with respect to φ and equating it to zero. Thus,
dγ φ
= −(εx − εy ) cos 2φ − 2εxy sin 2φ = 0
dφ
εx − εy
=⇒ tan 2φ = − (73)
2εxy
Note that tan 2θ · tan 2φ = −1. Then,
tan 2θ − tan 2φ
tan (2θ − 2φ) = =∞ assuming θ > φ
1 + tan 2θ · tan 2θ
=⇒ 2θ − 2φ = 90◦
=⇒ θ = φ + 45◦

So, the principal shear directions are at 45◦ to the principal directions on which
the normal strains attain extremum values. Now we have,
εx − εy 2εxy
sin 2φ = q , cos 2φ = − q (74)
(εx − εy )2 + 4ε2xy (εx − εy )2 + 4ε2xy

Or
εx − εy 2εxy
sin 2φ = − q , cos 2φ = q (75)
(εx − εy )2 + 4ε2xy (εx − εy )2 + 4ε2xy

Substituting in
1 1
εφ = (εx + εy ) + (εx − εy ) cos 2φ + εxy sin 2φ
2 2
γφ = −(εx − εy ) sin 2φ + 2εxy cos 2φ

AE 227 38 Dr. Krishnendu Haldar

4.5 Principal strains

we get
1 1
εφ = (εx + εy ) = (ε1 + ε2 ) (76)
2q 2
γφ = ± (εx − εy )2 + 4ε2xy
= ±(ε1 − ε2 ). (77)

4.5 Principal strains

We want to know if we can find an n (a unit vector) such that εn = εn. Here ε
is a real nonzero scalar and physically it scales initial vector εn along the specific
direction n. ε is called principal value and n is known as principal direction.

εij nj = εni =⇒ εij nj = εδij nj =⇒ [εij − εδij ]nj = 0 (78)

In 2D we can write the above equation in the matrix form as:

    
εx − ε εxy n1 0
= (79)
εxy εy − ε n 2 0
For the nontrivial solutions one has
 
εx − ε εxy 
 =0 (80)
 εxy εy − ε
Or

(εx − ε)(εy − ε) − ε2xy = 0 (81)

Or

ε2 − (εx + εy )ε + (εx εy − ε2xy ) = 0 (82)

The above equation is known as characteristic equation.We can then write the two
roots as:
r
εx + εy εx − εy  2
ε1 = + + ε2xy (83)
2 2
r
εx + εy εx − εy  2
ε2 = − + ε2xy (84)
2 2
From (79) we can write for ε1
(1) (1)
(εx − ε1 ) n1 + εxy n2 = 0
(1) (1)
εxy n1 + (εy − ε1 ) n2 = 0

AE 227 39 Dr. Krishnendu Haldar

4.5 Principal strains

From the first equation, we can write

(1) (1)
n1 n2
= = c (say) (85)
εxy −(εx − ε1 )
(86)

Here c is an arbitrary constant. Then,

 
" # εxy
(1)
n1  
(1) = c 
 r 2 . (87)
n2 
− εx −ε
2
y
+ εx −εy
2
+ ε 2
xy

If you consider the second equation, you will get the same results (check this).
Now for the vector n(1) , if θ is the angle with respect to the x-plane, then
p
−a + a2 + ε2xy
tan θ = (88)
εxy
εx −εy
where a = 2
. Then,
√
−a+ a2 +ε2xy
2 tan θ 2 εxy
tan 2θ = = √
1 − tan2 θ a2 −2a a2 +ε2xy +a2 +ε2xy
1− ε2xy
√
−a+ a2 +ε2xy
2 εxy
= √
−2a2 +2a a2 +ε2xy
ε2xy
εxy
=
a
2εxy
= (89)
εx − εy
which is shame as (65). Similarly, for the second eigenvalue ε2 ,
(2) (2)
(εx − ε2 ) n1 + εxy n2 = 0
(2) (2)
εxy n1 + (εy − ε2 ) n2 = 0

From the first equation, we can write

(2) (2)
n1 n2
= = c (say) (90)
εxy −(εx − ε2 )
(91)

AE 227 40 Dr. Krishnendu Haldar

4.6 Spherical and deviatoric strains

Here c is an arbitrary constant. Then,

 
" # εxy
(2)
n1  
(2) = c  r 2 . (92)
n2  ε −ε  
ε −ε
− x2 y − x
2
y
+ ε 2
xy

Note that n(1) · n(2) = 0, i.e., n(1) ⊥ n(2) and with respect to the principal axes
{n(1) , n(2) }, the strain matrix contains only diagonal terms, i.e.,
 
ε1 0
(93)
0 ε2

4.6 Spherical and deviatoric strains

Spherical strain tensor: ε̃ij = 31 εkk δij .
Deviatoric strain tensor: ε̂ij = εij − 31 εkk δij .

Note that εij = ε̃ij + ε̂ij .

4.7 Strain compatibility

If we specify continuous, single-valued displacements u, v, w, then through dif-
ferentiating them we can obtain the strain fields. However, the converse is not
necessarily true. That is given the six strain components, integrating the strain
displacement relation, does not necessarily produce continuous single valued dis-
placements since we are trying to solve six equations for three unknowns. In order
to ensure continuous, single-valued displacements, the strains must satisfy addi-
tional relations called integrability or compatibility equations.
The process to develop these equations is based on elimination the displace-
ments from the strain-displacement equations. We can then write,
1
(ui,jkl + uj,ikl )
εij,kl =
2
1
εkl,ij = (uk,lij + ul,kij )
2
1
εjl,ik = (uj,lik + ul,jik )
2
1
εik,jl = (ui,kjl + uj,ijl )
2
We can now interchange the order of differentiation on u, we can eliminate dis-
placement as
εij,kl + εkl,ij − εjl,ik − εik,jl (94)

AE 227 41 Dr. Krishnendu Haldar

4.8 Strain in polar coordinates

This leads 81 individual equations, most are either simple identities or repetitions,
and only six are meaningful. They are
∂ 2 εx ∂ 2 εy ∂ 2 εxy
+ = 2 (95)
∂y 2 ∂x2 ∂x∂y
2
∂ εy ∂ 2 εz ∂ 2 εyz
+ = 2 (96)
∂z 2 ∂y 2 ∂y∂z
2
∂ εz ∂ 2 εx ∂ 2 εzx
+ = 2 (97)
∂x2 ∂z 2 ∂z∂x
∂ 2 εx ∂  ∂εyz ∂εzx ∂εxy 
= − + + (98)
∂y∂z ∂x ∂x ∂y ∂z
∂ 2 εy ∂  ∂εzx ∂εxy ∂εyz 
= − + + (99)
∂z∂x ∂y ∂y ∂z ∂x
∂ 2 εz ∂  ∂εxy ∂εyz ∂εzx 
= − + + (100)
∂x∂y ∂z ∂z ∂x ∂y

4.8 Strain in polar coordinates

Recall that for the two vectors a = ai êi and b = bj êj , we can write the dyadic
product as

a⊗b = (a1 ê1 + a2 ê2 + a3 ê3 ) ⊗ (b1 ê1 + b2 ê2 + b3 ê3 )

= a1 b1 ê1 ⊗ ê1 + a1 b2 ê1 ⊗ ê2 + a1 b3 ê1 ⊗ ê3
+ a2 b1 ê2 ⊗ ê1 + a2 b2 ê2 ⊗ ê2 + a2 b3 ê2 ⊗ ê3
+ a3 b1 ê3 ⊗ ê1 + a3 b2 ê3 ⊗ ê2 + a3 b3 ê3 ⊗ ê3

In the matrix form:

 
a1 b 1 a1 b 2 a1 b 3
(a ⊗ b)ij = a2 b1 a2 b2 a2 b3 
a3 b 1 a3 b 2 a3 b 3
If the displace vector at a point (x, y, z) is u = u1 (x, y, z) ê1 + u2 (x, y, z) ê2 +
u3 (x, y, z) ê3 , then the displacement gradient is given by

∇ ⊗ u = (ê1 ∂1 + ê2 ∂2 + ê3 ∂3 ) ⊗ (u1 ê1 + u2 ê2 + u3 ê3 )

∂u1 ∂u2 ∂u3
= ê1 ⊗ ê1 + ê1 ⊗ ê2 + ê1 ⊗ ê3
∂x1 ∂x1 ∂x1
∂u1 ∂u2 ∂u3
+ ê2 ⊗ ê1 + ê2 ⊗ ê2 + ê2 ⊗ ê3
∂x2 ∂x2 ∂x2
∂u1 ∂u2 ∂u3
+ ê3 ⊗ ê1 + ê3 ⊗ ê2 + ê3 ⊗ ê3
∂x3 ∂x3 ∂x3

AE 227 42 Dr. Krishnendu Haldar

4.8 Strain in polar coordinates

In the matrix form:

 ∂u ∂v ∂w 
∂x ∂x ∂x
 
   ∂u ∂v ∂w 

uj,i = 
 ∂y ∂y ∂y  (101)
 
∂u ∂v ∂w
∂z ∂z ∂z

which is the transpose of the displacement gradient. In a similar way, we will

now compute ∇ ⊗ u in the polar coordinate. Note that in polar coordinate sys-
tem the Cartesian base vectors {ê1 , ê2 , ê1 } will change to the polar base vectors
{êr , êθ , êz }. In polar coordinate system we have the following transformations of
the coordinates:
p
x = r cos θ ; r = x2 + y 2
y
y = r sin θ ; θ = tan−1
x
z = z ; z=z

The unit vectors of the two systems are related as:

êr = cos θê1 + sin θê2

êθ = − sin θê1 + cos θê2
êz = ê3

Moreover,

ê1 = cos θêr − sin θêθ

ê2 = sin θêr + cos θêθ
êz = ê3

Also note that

∂r êr = 0, ∂r êθ = 0, ∂θ êr = êθ , ∂θ êθ = −êr

Consider a scalar function φ(r, θ). We write,

∂φ ∂φ ∂r ∂φ ∂θ
= +
∂x ∂r ∂x ∂θ ∂x
∂φ x ∂φ x2 y
= + 2
(− 2 )
∂r r ∂θ r x
∂φ ∂φ sin θ
= cos θ −
∂r ∂θ r

AE 227 43 Dr. Krishnendu Haldar

4.8 Strain in polar coordinates

Then,
∂ ∂ sin θ ∂
= cos θ − (102)
∂x ∂r r ∂θ
∂φ ∂φ ∂r ∂φ ∂θ
= +
∂y ∂r ∂y ∂θ ∂y
∂φ y ∂φ x2 1
= +
∂r r ∂θ r2 x
∂φ y ∂φ x
= +
∂r r ∂θ r2
∂φ ∂φ cos θ
= sin θ +
∂r ∂θ r
So we write,
∂ ∂ cos θ ∂
= sin θ + (103)
∂y ∂r r ∂θ
Now,
∇ = ê1 ∂x + ê2 ∂y + ê3 ∂z
∂ sin θ ∂
= (cos θêr − sin θêθ )(cos θ − )
∂r r ∂θ
∂ cos θ ∂
+ (sin θêr + cos θêθ )(sin θ + )
∂r r ∂θ
+ ê3 ∂z
∂ 1 ∂
= êr + êθ + ê3 ∂z
∂r r ∂θ
Then,
∂ 1 ∂
∇ ⊗ u = (êr + êθ + êz ∂z ) ⊗ (ur êr + uθ êθ + uz êz )
∂r r ∂θ
∂ur ∂uθ ∂uz
= êr ⊗ êr + êr ⊗ êθ + êr ⊗ êz
∂r ∂r ∂r
h 1 ∂(u ê ) i h 1 ∂(u ê ) i 1 ∂u
r r θ θ z
+ êθ ⊗ + êθ ⊗ + êθ ⊗ êz
r ∂θ r ∂θ r ∂θ
∂ur ∂uθ ∂uz
+ êz ⊗ êr + êz ⊗ êθ + êz ⊗ êz
∂z ∂r ∂z
As the unit vectors {êr , êθ } are functions of θ, we further write,
∂ur ∂uθ ∂uz
∇⊗u = êr ⊗ êr + êr ⊗ êθ + êr ⊗ êz
∂r ∂r ∂r
h 1 ∂u ur i h 1 ∂u uθ i 1 ∂uz
r θ
+ êθ ⊗ êr + êθ + êθ ⊗ êθ − êr + êθ ⊗ êz
r ∂θ r r ∂θ r r ∂θ
∂ur ∂uθ ∂uz
+ êz ⊗ êr + êz ⊗ êθ + êz ⊗ êz
∂z ∂r ∂z
AE 227 44 Dr. Krishnendu Haldar
4.8 Strain in polar coordinates

So, we write
∂ur ∂uθ ∂uz
∇⊗u = êr ⊗ êr + êr ⊗ êθ + êr ⊗ êz
∂r ∂r ∂r
h 1 ∂u uθ i h 1 ∂u ur i 1 ∂uz
r θ
+ − êθ ⊗ êr + + êθ ⊗ êθ + êθ ⊗ êz
r ∂θ r r ∂θ r r ∂θ
∂ur ∂uθ ∂uz
+ êz ⊗ êr + êz ⊗ êθ + êz ⊗ êz
∂z ∂r ∂z
In matrix form:
∂ur ∂uθ ∂uz
 
∂r ∂r ∂r
 
   
 1 ∂ur uθ
 
1 ∂uθ ur
 
1 ∂uz 
uj,i =  r ∂θ − r r ∂θ
+ r r ∂θ 
(104)
 
 
∂ur ∂uθ ∂ur
∂z ∂z ∂z

Then the strain components in the polar coordinate are 21 (ui,j + uj,i ) :

∂ur
εr =
∂r
1 ∂uθ
εθ =
r ∂θ
∂ur
εz =
∂z
1  ∂uθ 1 ∂ur uθ 
εrθ = + −
2 ∂r r ∂θ r
1 ∂uz ∂ur
 
εrz = +
2 ∂r ∂z
1  1 ∂uz ∂uθ 
εθz = +
2 r ∂θ ∂z
For the geometric interpretation, see Appendix B.

AE 227 45 Dr. Krishnendu Haldar

5 Concept of stress
One of the main problems of engineering mechanics of solids is the investigation
of the internal resistance of a body, that is, the nature of forces within a body to
balance the effect of the external applied forces. Consider an arbitrary body is
in static equilibrium with external forces F 1 , F 2 , ... (Fig. 23). Now an arbitrary
plane is passed through the body such that the plane separates the body into two
distinct parts. For such parts of a body, however, some of the forces should act at
the cut section to maintain the equilibrium. So for the upper half
N
X
F1 + F2 + s1 = 0
i=1

where, N is the number of points, like A, B, C, ..., through which the internal
forces s1 , s2 , s3 , ... act. In general, N is pretty large such that we get a continuous
internal force distribution. The internal forces on the other surface are equal and
opposite to s1 , s2 , s3 , ... such that the resultant force is zero when they combine.

Figure 23: Static equilibrium and internal forces.

AE 227 46 Dr. Krishnendu Haldar

5.1 Definition of stress

5.1 Definition of stress

In general, the internal forces acting on infinitesimal areas of a cut are of varying
magnitudes and directions, as shown in Fig. 23. In general, they vary from point
to point and are inclined with respect to the plane of the section. It is then advan-
tageous to resolve them parallel and perpendicular to the investigated section. As
an example, the components of a force vector ∆t acting on an area ∆A are shown
in Fig. 24. In this particular diagram, the section through the body is perpendic-

Figure 24: Sectioned view and enlarge view with components of ∆t.

ular to the x− axis, and the directions of ∆tx and of the normal to ∆A coincide.
The component parallel to the section is further resolved into components along
the y and z axes. Since the components of the intensity of force per unit area,
i.e.,stress, hold true at a point, the mathematical definition of stress is:
∆tx ∆ty ∆tz
σxx = lim , τxy = lim , τxz = lim (105)
∆A→0 ∆A ∆A→0 ∆A ∆A→0 ∆A

where, in all the three cases, the first subscript denotes the direction of the con-
sidered section, i.e., x−plane, and the second subscript denotes the direction of
the stress components. The intensity of the force normal to the section is called
the normal stress at a point. It is customary to refer to the normal stresses that
cause tension on the surface as tensile stress. On the other hand, those that are
pushing against it are compressive stress. In this course, we’ll denote the three

AE 227 47 Dr. Krishnendu Haldar

5.2 Stress tensor

normal stresses as σx , σy , and σz . The other components that act parallel to the
plane are known as shear stress. Stresses multiplied by the respective areas on
which they act give force. At an imaginary section, a vector sum of these forces,
called stress resultants, keeps a body in equilibrium.

5.2 Stress tensor

(a) (b)

Figure 25: (a) General state of stress acting on an infinitesimal element in the
initial coordinate system. (b) General state of stress acting on an infinitesimal
element defined in a rotated system of coordinate axes.

Let us imagine another plane parallel to the first (Fig. 24a) at an infinitesimal
distance away and an elementary isolated cubic slice out of that slice as shown
in Fig. 25. All stresses acting on this cube are identified on the diagram. The
directions of the stresses are positive if they coincide with the positive directions
of the axes. The nine stress components of the stress tensor are written as:
 
σx τxy τxz
[σ]ij = τyx σy τyz  (106)
τzx σzy σz
If a different set of axes are chosen, corresponding stresses are shown in Fig. 24(b).
The new stress components are not equal but are related with the old one. The

AE 227 48 Dr. Krishnendu Haldar

5.2 Stress tensor

process of changing stresses from one set of coordinate axes to another is termed
stress transformation. We’ll discuss stress transformation in detail in the next
section.
Now consider the traction vector on an oblique plane with arbitrary orientation.
The force ∆t acts on the slanted surface which has the unit vector n = nx î +
ny ĵ + nz k̂. Let us consider force vector acting on the x, y, and z planes are −∆tx ,
−∆ty , and −∆tz , respectively. The negative signs are due to the fact that the
forces are acting on the negative planes. From the force equilibrium we write for
the static equilibrium:

∆t − ∆tx − ∆ty − ∆tz = 0. (107)

If the magnitude of the differential area is ∆A, then the area vector is written as
∆A = ∆A n. The corresponding projections on the x, y, and z planes are

∆Ax = ∆A · î = ∆A n · î = nx ∆A
∆Ay = ∆A · ĵ = ∆A n · ĵ = ny ∆A
∆Az = ∆A · k̂ = ∆A n · k̂ = nz ∆A (108)

Now note that

Figure 26: Force vector on an oblique plane.

AE 227 49 Dr. Krishnendu Haldar

5.2 Stress tensor

∆tx  ∆tx
x
∆txy ∆txz 
= î + ĵ + k̂
∆Ax ∆Ax ∆Ax ∆Ax
∆ty  ∆ty
x
∆tyy ∆tyz 
= î + ĵ + k̂
∆Ay ∆Ay ∆Ay ∆Ay
∆tz  ∆tz
x
∆tzy ∆tzz 
= î + ĵ + k̂
∆Az ∆Az ∆Az ∆Az
and using the relations of (108), we further write

∆tx ∆txx ∆txy ∆txz

= nx î + nx ĵ + nx k̂
∆A ∆Ax ∆Ax ∆Ax
∆ty ∆tyx ∆tyy ∆tyz
= ny î + ny ĵ + ny k̂
∆A ∆Ay ∆Ay ∆Ay
∆tz ∆tzx ∆tzy ∆tzz
= nz î + nz ĵ + nz k̂
∆A ∆Az ∆Az ∆Az
Adding we get:
 ∆tx + ∆ty + ∆tz   ∆tx ∆tyx ∆tzx 
x
= nx + ny + nz î
∆A ∆Ax ∆Ay ∆Az
 ∆tx ∆tyy ∆tzy 
y
+ nx + ny + nz ĵ
∆Ax ∆Ay ∆Az
 ∆tx ∆tyz ∆tzz 
z
+ nx + ny + nz k̂
∆Ax ∆Ay ∆Az

Further using (107) and taking the limit we write

 ∆t   ∆txx ∆tyx ∆tzx 
lim = lim nx + lim ny + lim nz î
∆A→0 ∆A ∆Ax →0 ∆Ax ∆Ay →0 ∆Ay ∆Az →0 ∆Az
 ∆txy ∆tyy ∆tzy 
+ lim nx + lim ny + lim nz ĵ
∆Ax →0 ∆Ax ∆Ay →0 ∆Ay ∆Az →0 ∆Az
 ∆txz ∆tyz ∆tzz 
+ lim nx + lim ny + lim nz k̂
∆Ax →0 ∆Ax ∆Ay →0 ∆Ay ∆Az →0 ∆Az

This implies:
 
Tn = σx nx + τyx ny + τzx nz î
 
+ τxy nx + σy ny + τzy nz ĵ
 
+ τxz nx + τyz ny + σz nz k̂

AE 227 50 Dr. Krishnendu Haldar

5.3 Equilibrium equation in Cartesian system

In the matrix form the components are written as:

 n   
Tx σx τyx τzx nx
Tyn  = τxy σy τzy  ny  (109)
Tzn τxz σyz σz nz

and in the index notation:

Tin = σji nj =⇒ T n = σtn (110)

The surface force density T n is normally referred to traction vector. One of our
interests will be to compute the traction vector’s normal and shear components,
σn and τn , respectively. So,

σn = T n · n (111)
p n
τn = |T |2 − σn2 (112)

5.3 Equilibrium equation in Cartesian system

Figure 27: Infinitesimal element with stress and body force.

An infinitesimal element of a body must be in equilibrium. For the two dimen-

sional case the system of stresses acting on an infinitesimal element (∆x)(∆y)(1)
is shown in Fig. 27. In this derivation, the element is of unit thickness in the
direction of perpendicular to the plane of paper. The body force, such as the body

AE 227 51 Dr. Krishnendu Haldar

5.3 Equilibrium equation in Cartesian system

force or the electromagnetic effect, are designated fx and fy and are associated
with the unit volume of the material. Then,
X
Fx = 0 =⇒ σx (x + ∆x, y)(∆y · 1) + τyx (x, y + ∆y)(∆x · 1) − σx (x, y)(∆y · 1)
−τyx (x, y)(∆x · 1) + fx (x, y)(∆x · ∆y · 1) = 0 (113)
Moreover,
∂σx
σx (x + ∆x, y) = σx (x, y) + ∆x + h.o.t
∂x
∂τyx
τyx (x, y + ∆y) = τyx (x, y) + ∆y + h.o.t
∂y
Substituting back in the force balance equation we get
∂σx ∂τyx
+ + fx = 0
∂x ∂y
P
Similarly, considering Fy = 0 we can write
∂τxy ∂σy
+ + fy = 0
∂x ∂y
The moment balance equation at the center of the infinitesimal element reads
X 1 1
Mz = 0 =⇒ −τyx (x, y + ∆y)(∆x · 1)( ∆y) − τyx (x, y)(∆x · 1)( ∆y)
2 2
1 1
+τxy (x + ∆x, y)(∆y · 1)( ∆x) + τxy (x, y)(∆y · 1)( ∆x) = 0 (114)
2 2
Neglecting the higher order terms we get
τxy = τyx
Extending the above concept in 3D, we can write the force equilibrium equation
as
∂σx ∂τyx ∂τzx
+ + + fx = 0 (115)
∂x ∂y ∂z
∂τxy ∂σy ∂τzy
+ + + fy = 0 (116)
∂x ∂y ∂x
∂τxz ∂τyz ∂σz
+ + + fz = 0 (117)
∂x ∂y ∂x
and moment equilibrium equation as
τxy = τyx τyz = τzy τzx = τxz (118)

AE 227 52 Dr. Krishnendu Haldar

5.4 Equilibrium equation in polar system

In the index notation

σji,j + fi = 0 =⇒ ∇ · σt + f = 0 (119)

and

σji = σij =⇒ σ t = σ. (120)

As, σ t = σ, we can further write

σij,j + fi = 0 =⇒ ∇·σ+f =0 (121)

Tin = σij nj =⇒ T n = σn (122)

5.4 Equilibrium equation in polar system

Figure 28: Infinitesimal element with stress and body force.

We will only consider equilibrium equations for plane stress problem, i.e., there
is no σz component. The four stress components are denoted by σrr = σr , τrθ , τθr ,
and σθθ = σθ . Moreover, fr and fθ are the body forces in the êr and êθ directions.
X
Fθ = 0 =⇒ σθ (r, θ + ∆θ)(∆r · 1) cos (∆θ/2) + τθr (r, θ + ∆θ)(∆r · 1) sin (∆θ/2)
−σθ (r, θ)(∆r · 1) cos (∆θ/2) + τθr (r, θ)(∆r · 1) sin (∆θ/2)
+τrθ (r + ∆r, θ)(r + ∆r) · 1 · ∆θ − τrθ (r, θ)r · 1 · ∆θ
+fθ (r, θ)(∆r · r∆θ · 1) = 0 (123)

AE 227 53 Dr. Krishnendu Haldar

5.4 Equilibrium equation in polar system

Moreover,
∂σθ
σθ (r, θ + ∆θ) = σθ (r, θ) + ∆θ + h.o.t
∂θ
∂τθr
τθr (r, θ + ∆θ) = τθr (r, θ) + ∆θ + h.o.t
∂θ
∂τrθ
τrθ (r + ∆r, θ) = τrθ (r, θ) + ∆r + h.o.t
∂r
cos (∆θ/2) ≈ 1
sin (∆θ/2) ≈ ∆θ/2
(124)

Substituting back we get

∂σθ
(∆θ)(∆r · 1) + τθr (∆r · 1)(∆θ)
∂θ
∂τrθ
+ ∆r · 1 · r∆θ + τrθ ∆r · 1∆θ + fθ (r, θ)(∆r · r∆θ · 1) = 0
∂r
Or
1 ∂σθ ∂τrθ τθr + τrθ
+ + + fθ = 0
r ∂θ ∂r r
Similarly,
X
Fr = 0 =⇒ σr (r + ∆r, θ)(r + ∆r) · 1 · ∆θ + τθr (r, θ + ∆θ)(∆r · 1) cos (∆θ/2)
−σr (r, θ)(1 · r∆θ) − τθr (r, θ)(∆r · 1) cos (∆θ/2) − σθ (r, θ + ∆θ)(∆r · 1) sin (∆θ/2)
−σθ (r, θ)(∆r · 1) sin (∆θ/2) + fr (r, θ)(∆r · r∆θ · 1) = 0

or
 ∂σr   ∂τθr 
σr + ∆r (r + ∆r) · 1 · ∆θ + τθr + ∆θ (∆r · 1)
∂r ∂θ
 ∂σθ 
−σr (1 · r∆θ) − τθr (r, θ)(∆r · 1) − σθ + ∆θ (∆r · 1)(∆θ/2)
∂θ
−σθ (∆r · 1)(∆θ/2) + fr (∆r · r∆θ · 1) = 0

or
 ∂σr  ∂τ
θr
σr ∆r∆θ + r∆r∆θ + ∆θ∆r
∂r ∂θ
−σθ ∆r∆θ + fr r∆r∆θ = 0

AE 227 54 Dr. Krishnendu Haldar

Finally,
∂σr 1 ∂τθr σr − σθ
+ + + fr = 0
∂r r ∂θ r
Moreover, from the moment balance equation we get

τθr = τrθ (125)

Using the above relation, we can rewrite the linear momentum balance equations
as:
∂σr 1 ∂τrθ σr − σθ
+ + + fr = 0 (126)
∂r r ∂θ r
1 ∂σθ ∂τrθ 2τrθ
+ + + fθ = 0 (127)
r ∂θ ∂r r

6 Stress components on arbitrary plane (2D)

Recall from (111) and (112)

σn = T n · n (128)
p n
τn = |T |2 − σn2 (129)

where T n = σn due to the symmetry of the stress. If the normal of the plane is
n = cos θ î + sin θ ĵ, then
 n     
Tx σx τxy cos θ σx cos θ + τxy sin θ
= = (130)
Tyn τxy σy sin θ τxy cos θ + σy sin θ

So,

σn = Txn cos θ + Tyn sin θ = σx cos2 θ + τxy sin 2θ + σy sin2 θ (131)

Now

|T n |2 = (σx cos θ + τxy sin θ)2 + (τxy cos θ + σy sin θ)2

= σx2 cos2 θ + σy2 sin2 θ + τxy
2
+ τxy (σx + σy ) sin 2θ (132)

and

σn2 = σx2 cos4 θ + τxy

2
sin2 2θ + σy2 sin4 θ
+2σx τxy sin 2θ cos2 θ + 2σy τxy sin 2θ sin2 θ + 2σx σy sin2 θ cos2 θ

AE 227 55 Dr. Krishnendu Haldar

6.1 Stress transformation

Then,
 
|T n |2 − σn2 = σx2 cos2 θ sin2 θ + σy2 cos2 θ sin2 θ − 2σx σy sin2 θ cos2 θ + τxy
2
cos2 2θ
−σx τxy sin 2θ(1 − 2 cos2 θ) + σy τxy sin 2θ(2 sin2 θ − 1)
1 2  
= σy − σx sin2 2θ + τxy 2
cos2 2θ + σy − σx τxy sin 2θ cos 2θ
4
1  2
= σy − σx sin 2θ + τxy cos 2θ
2
So,
p n
τn = |T |2 − σn2
1  
= σy − σx sin 2θ + τxy cos 2θ (133)
2
In summary,

σn = σx cos2 θ + τxy sin 2θ + σy sin2 θ (134)

1 
τn = σy − σx sin 2θ + τxy cos 2θ (135)
2

6.1 Stress transformation

Let normal to the oblique plane EF be denoted by Ox0 and perpendicular to it
by Oy 0 . Then the values of σn and τn on the face EF will be σx0 and τxy
0
, while the
stress components on face F G will be σy0 and τyx0
. So we write,

σn = σx0 = σx cos2 θ + τxy sin 2θ + σy sin2 θ (136)

0 1 
τn = τxy = σy − σx sin 2θ + τxy cos 2θ (137)
2
0
The values of σy0 and τyx are obtained by putting θ ≡ θ + 90◦

σy0 = σx sin2 θ − τxy sin 2θ + σy cos2 θ (138)

0 1 
0
τyx = σx − σy sin 2θ − τxy cos 2θ = −τxy (139)
2
The transformation also follows the following standard rules of second order tensor
transformation:

σij0 = Qpi Qqj σij (140)

where, the rotation matrix Qmn = cos (x0m , xn ). In tensor notation we further write

σ 0 = QσQt . (141)

AE 227 56 Dr. Krishnendu Haldar

6.1 Stress transformation

Figure 29: Stress transformation at an arbitrary direction.

For example, in 2D Cartesian system:

 
cos θ sin θ
[Q] = (142)
− sin θ cos θ
Then,
 0 0
    
σx τxy cos θ sin θ σx τxy cos θ − sin θ
0 =
τxy σy0 − sin θ cos θ τxy σy sin θ cos θ
  
cos θ sin θ σx cos θ + τxy sin θ −σx sin θ + τxy cos θ
=
− sin θ cos θ τxy cos θ + σy sin θ −τxy sin θ + σy cos θ
σx cos2 θ + τxy sin 2θ + σy sin2 θ 12 (σyy − σxx ) sin 2θ + τxy cos 2θ
 
= 1
2
(σy − σx ) sin 2θ + τxy cos 2θ σx sin2 θ − τxy sin 2θ + σy cos2 θ
So we obtain,
σx0 = σx cos2 θ + τxy sin 2θ + σy sin2 θ (143)
0 1
τxy = (σy − σx ) sin 2θ + τxy cos 2θ (144)
2
σy0 = σx sin2 θ − τxy sin 2θ + σy cos2 θ (145)

6.1.1 Some important transformations

1. σx = σy , and τxy = 0:

AE 227 57 Dr. Krishnendu Haldar

6.1 Stress transformation

Figure 30: Isotropic state of stress.

σx0 = σx cos2 θ + σx sin2 θ = σx

0
τxy = 0
σy0 = σy sin2 θ + σy cos2 θ = σy

That is the normal stress on an arbitrary plane is equal to σx , and shear

stress on that plane is zero. This is equivalent to the statement that when
σx = σy , and τxy = 0, the rectangular stress components are the same with
reference to any arbitrary frame of reference, i.e., independent of frame of
reference.

2. σx = −σy , τxy = 0, and θ = 45◦ :

σx0 = 0
0
τxy = −σx
σy0 = 0

This means an element with faces at ±45◦ to the original element will be
subjected to pure shear with no normal stresses.

3. σx = σy = 0, and θ = 45◦ :

σx0 = τxy
0
τxy = 0
0
σy = −τxy

This is the converse of the previous case. This means a pure shear is equiv-
alent to tension and compression at an element with faces at ±45◦ .

AE 227 58 Dr. Krishnendu Haldar

6.2 Maximum normal stress

Figure 31: Pure shear at 45◦ .

Figure 32: Tension and compression at 45◦ .

6.2 Maximum normal stress

In discussing the strength or weakness of a material, one tries to estimate the
maximum normal stress which may lead to the failure or fracture of the material.

AE 227 59 Dr. Krishnendu Haldar

6.2 Maximum normal stress

∂σn
Thus we are looking for the theta such that ∂θ
= 0. So,
dσn
= −(σx − σy ) sin 2θ + 2τxy cos 2θ = 0 (146)
dθ
Or
2τxy
tan 2θ = . (147)
σx − σy
The value of θ defines the plane on which the normal stress attains an extremum,
i.e., a maximum or a minimum. The above equations gives two values of θ, since
tan 2θ = tan 2θ + 180◦ . If θ1 and θ2 are these two values, we then have
θ2 = θ1 + 90◦ .
σn is maximum on one of these planes and a minimum on the other. Consider now
the plane on which τn = 0. This gives
2τxy
tan 2θ = . (148)
σx − σy
In other words, the planes on which the normal stresses is an extremum also have
zero shear stress. Planes on which the normal stresses have maximum or minimum
values are called principal planes and the stress magnitudes on these planes are
the principal stresses. Since the shear stresses are zero, principal planes are also
called shearless planes. We rewrite (143) and (144) in double angle form as
1 1
σn = (σx + σy ) + (σx − σy ) cos 2θ + τxy sin 2θ (149)
2 2
1
τn = − (σx − σy ) sin 2θ + τxy cos 2θ (150)
2
From (147) we can write:
2τxy σx − σy
sin 2θ = q , cos 2θ = q (151)
(σx − σy )2 + 4τxy
2 (σx − σy )2 + 4τxy
2

Or
2τxy σx − σy
sin 2θ = − q , cos 2θ = − q (152)
(σx − σy )2 + 4τxy
2 (σx − σy )2 + 4τxy
2

Substituting back in (178) and (179) we get

r
σx + σy σx − σy 2 2
σ1,2 = ± + τxy (153)
2 2
τ1,2 = 0 (154)

AE 227 60 Dr. Krishnendu Haldar

6.3 Maximum shear stress

6.3 Maximum shear stress

Let us now determine the planes on which the shear stress reaches an extremum
value by differentiation τn with respect to φ and equating it to zero. Thus,
dτn
= −(σx − σy ) cos 2φ − 2τxy = 0
dφ
σx − σy
=⇒ tan 2φ = − (155)
2τxy

Note that tan 2θ · tan 2φ = −1. Then,

tan 2θ − tan 2φ
tan (2θ − 2φ) = =∞
1 + tan 2θ · tan 2θ
=⇒ 2θ − 2φ = 90◦
=⇒ θ = φ + 45◦

So, the principal shear planes are at 45◦ to the principal planes on which the
normal stresses attain extremum values. Now we have,
σx − σy 2τxy
sin 2θ = q , cos 2θ = − q (156)
(σx − σy )2 + 4τxy
2 (σx − σy )2 + 4τxy
2

Or
σx − σy 2τxy
sin 2θ = − q , cos 2θ = q (157)
(σx − σy )2 + 4τxy
2 (σx − σy )2 + 4τxy
2

Substituting in
1 1
σn = (σx + σy ) + (σx − σy ) cos 2θ + τxy sin 2θ
2 2
1
τn = − (σx − σy ) sin 2θ + τxy cos 2θ
2
we get
1
σn = (σx + σy ) (158)
2
1q
τn = ± (σx − σy )2 + 4τxy
2
2
1
= ± (σ1 − σ2 ). (159)
2

AE 227 61 Dr. Krishnendu Haldar

6.4 Principal stresses

6.4 Principal stresses

We already derived the relation T n = σn. Let us now keep changing the direction
of n and for each new n we will get a new traction vector T n . We want to know
if we can find an n such that σn = σn. Here σ is a real nonzero scalar and
physically it scales the traction vector along that specific direction n. σ is called
principal value and n is known as principal direction.

σij nj = σni =⇒ σij nj = σδij ni =⇒ [σij − σδij ]ni = 0 (160)

In 2D we can write the above equation in the matrix form as:

    
σx − σ τxy n1 0
= (161)
τxy σy − σ n2 0
For the nontrivial solutions one has
 
σ x − σ τxy

 =0 (162)
 τxy σy − σ 
Or
2
(σx − σ)(σy − σ) − τxy =0 (163)

Or

σ 2 − (σx + σy )σ + (σx σy − τxy

2
)=0 (164)

The above equation is known as characteristic equation.We can then write the two
roots as:
r
σx + σy σx − σy 2 2
σ1 = + + τxy (165)
2 2
r
σx + σy σx − σy 2 2
σ2 = − + τxy (166)
2 2
From (161) we can write for σ1
(1) (1)
(σx − σ1 ) n1 + τxy n2 = 0
(1) (1)
τxy n1 + (σy − σ1 ) n2 = 0

From the first equation, we can write

(1) (1)
n1 n2
= = c (say) (167)
τxy −(σx − σ1 )
(168)

AE 227 62 Dr. Krishnendu Haldar

6.4 Principal stresses

Here c is an arbitrary constant. Then,

 
" # τxy
(1)
n1  
(1) = c 
 r 2 . (169)
n2 σx −σy σx −σy 2

− 2 + 2
+ τxy

If you consider the second equation, you will get the same results (check this).
Now for the vector n(1) , if θ is the angle with respect to the x-plane, then
p
−a + a2 + τxy 2
tan θ = (170)
τxy
σx −σy
where a = 2
. Then,
√
−a+ a2 +τxy
2

2 tan θ 2 τxy
tan 2θ = = √
1 − tan2 θ a2 −2a a2 +τxy
2 +a2 +τ 2
xy
1− 2
τxy
√
−a+ a2 +τxy
2
2 τxy
= √
−2a2 +2a a2 +τxy
2
2
τxy
τxy
=
a
2τxy
= (171)
σx − σy
which is shame as (147). Similarly, for the second eigenvalue σ2 ,
(2) (2)
(σx − σ2 ) n1 + τxy n2 = 0
(2) (2)
τxy n1 + (σy − σ2 ) n2 = 0
From the first equation, we can write
(2) (2)
n1 n2
= = c (say) (172)
τxy −(σx − σ2 )
(173)
Here c is an arbitrary constant. Then,
 
" # τxy
(2)
n1  
(2) = c 
 r 2 . (174)
n2 σx −σy σx −σy 2

− 2 − 2
+ τ xy

AE 227 63 Dr. Krishnendu Haldar

6.5 Mohr’s circle for biaxial stresses

Note that with respect to the principal axes {n1 , n2 }, the stress matrix contains
only diagonal terms, i.e.,
 
σ1 0
(175)
0 σ2

6.5 Mohr’s circle for biaxial stresses

Figure 33: Construction of Mohr’s circle.

We rewrite (143) and (144) in double angle form as

1 1
σx0 = (σx + σy ) + (σx − σy ) cos 2θ + τxy sin 2θ (176)
2 2
0 1
τxy = − (σx − σy ) sin 2θ + τxy cos 2θ (177)
2
Let us now assume,
1 1
(σx + σy ) = a, (σx − σy ) = b
2 2
So,
σx0 − a = b cos 2θ + τxy sin 2θ (178)
0
τxy = −b sin 2θ + τxy cos 2θ (179)

AE 227 64 Dr. Krishnendu Haldar

6.5 Mohr’s circle for biaxial stresses

Squaring and adding we get,

2
(σx0 − a)2 + τ 0 xy = b2 + τxy
2
(180)
This is an equation of circle with respect to the variables {σx0 , τxy
0
}. The center of
0 σx +σy
the circle lies on the σx -axis at a distance a = 2 . The radius of the circle is
r 2
p 2 2 σx +σy 2 . The construction of the Mohr’s circle (Fig. 33) is
b + τxy = 2
+ τxy
the following:
0
1. Mark two coordinate axes σ and τ . (Rewriting {σx0 , τxy } as {σ, τ }).
2. Corresponding to the normal stresses σx and σy , mark off two points A and B
along the σ axis. Tensile stresses are marked along the +σ and compressive
stresses along the −σ axis.
3. Shear stress are marked along the τ -axis.
4. Point K, with coordinates (σx , τxy ), corresponds to the stress acting on the
x-face of the element. Point D corresponds to the y-face of the element. K is
called the key point since it represents the x-plane and angle 2θ is measured
from the x-axis.
5. Join K and D, cutting the σ axis at C.
6. With C as a center and CK (or CD) as radius, draw a circle. This is the
Mohr’s circle for a given state of stress at the point.
Points on the Mohr’s circle represent the normal and shear stresses acting on
different planes passing through the point P . Let σ1 be the stress at point E.
Then
σ1 = OC + CE
r
σx + σy σx + σy 2 2
= + + τxy (181)
2 2
The angle between the x-plane and principal plane in the Mohr’s circle is given by
2τxy
tan ∠KCA = = tan 2θ (182)
σx − σy
The important point to observe is that the angle measured in the element becomes
doubled in the Mohr’s circle diagram. Now if σ2 be the stress at point F , then
σ2 = OC − CF
r
σx + σy σx + σy 2 2
= − + τxy (183)
2 2

AE 227 65 Dr. Krishnendu Haldar

So these two end points are nothing but the principal stresses, where the shear
stress vanishes. Maximum and minimum shear stresses are at the points H and
G, respectively:
r
σx + σy 2 2
τmax, min = ± + τxy (184)
2
Now point H is 90◦ to point E. In the element diagram, therefore, the principal
shear planes will be at 45◦ to the principal planes. The normal stresses on the
principal shear planes are given by OC = σx +σ
2
y
.
To determine the normal and shear stresses on a plane whose normal makes an
angle φ with the x-axis in the element, we mark off a point J on the circle such
that ∠KCJ = 2φ. The coordinates of J then give the appropriate values.

Read page no 61-65 of the text book.

7 Constitutive relations
We will now construct a 3D generalized linear elastic constitutive laws where each
stress components are linearly related with each strain components. We write

σij = Cijkl εkl (185)

where the coefficients Cijkl are material parameters. Note that there are 3·3·3·3 =
81 material parameters. Now σij is symmetric then

σij = Cijkl εkl so σji = Cjikl εkl =⇒ Cijkl = Cjikl (186)

There are six independent ways to express when i and j are taken together and still
nine ways to express k and l taken together. So stress symmetry gives 6 · 9 = 54
independent constants. Now by definition, strain tensor is symmetric. So basically
there are six independent ways to express when k and l. Then total number of
independent constants further reduces to 6 · 6 = 36. If we consider the strain
energy (we’ll study it later) then,

U = [Cijkl εkl ]εij (187)

Now energy is invariant. If we interchange ij and kl, we get

U = [Cklij εij ]εkl = [Cijkl εkl ]εij (188)

Then, implies

Cklij = Cijkl (189)

AE 227 66 Dr. Krishnendu Haldar

Thus number of independent constants reduces to 6·(6+1)/2 = 21. We can finally
write
    
σ11 C1111 C1122 C1133 C1123 C1113 C1112 ε11
σ22  C2211 C2222 C2233 C2223 C2213 C2212  ε22 
    
σ33  C3311 C3322 C3333 C3323 C3313 C3312  ε33 
 =   (190)
σ23  C2311 C2322 C2333 C2323 C2313 C2312  ε23 
    
σ13  C1311 C1322 C1333 C1323 C1313 C1312  ε13 
σ12 C1211 C1222 C1233 C1223 C1213 C1212 ε12
In Voigt notation we can also write
    
σ1 C11 C12 C13 C14 C15 C16 ε1
σ2  C21 C22 C23 C24 C25 C26  ε2 
    
σ3  C31 C32 C33 C34 C35 C36 
 ε3 
 
 = (191)
σ4  C41 C42 C43 C44 C45 C46 
 ε4 
 
  
σ5  C51 C52 C53 C54 C55 C56  ε5 
σ6 C61 C62 C63 C64 C65 C66 ε6
We determine general anisotropic case where we have 21 independent elastic con-
stants. However, due to some inherent material symmetry of some material system,
there are some specific orientations where the material has same stress-strain re-
sponses. This is known as material symmetry. You will learn more about material
symmetry in a continuum mechanics course or mechanics of composite material
course. I’ll present some results without any detailed derivations.
Let [Q]ij be the transformation matrix between the actual and transformed
coordinate system. Then the transformation of the fourth order tensor is given by:
0
Cijkl = Qip Qjq Qkm Qln Cpqmn (192)

7.0.1 One plane of symmetry (monoclinic)

For this case x − y plane is the plane of symmetry (mirror reflection) and so
 
1 0 0
[Q]ij = 0
 1 0 (193)
0 0 −1
Substituting in (192), we get
 
C11 C12 C13 0 0 C16
 . C22 C23 0 0 C26 
 
 . . C 33 0 0 C36 
[C]ij = 
 .
 (194)
 . . C44 C45 0 
 . . . . C55 0 
. . . . . C66

AE 227 67 Dr. Krishnendu Haldar

We have 13 independent elastic moduli.

7.0.2 Three perpendicular planes of symmetry (orthotropic)

Here we have three mutual perpendicular planes of symmetry (mirror reflection).
Common examples are wood and fiber-reinforced composites. We thus obtain
 
C11 C12 C13 0 0 0
 . C22 C23 0 0 0 
 
 . . C 33 0 0 0 
[C]ij = 
 .
 (195)
 . . C44 0 0 

 . . . . C55 0 
. . . . . C66

We have 9 independent elastic moduli.

7.0.3 Axis of symmetry (transversely isotropic material)

This symmetry is with respect to rotations about an axis. We have
 
C11 C12 C13 0 0 0
 . C11 C13 0 0 0 
 
 . . C33 0 0 0 
[C]ij = 
 .
 (196)
 . . C44 0 0 

 . . . . C44 0 
. . . . . C11 − C12

We have 5 independent elastic moduli.

7.0.4 Complete symmetry (isotropic)

This is for complete symmetry and is known as isotropy. Here we have
 
C11 C12 C12 0 0 0
 . C11 C12 0 0 0 
 
 . . C 11 0 0 0 
[C]ij =   (197)
 .
 . . C11 − C12 0 0 

 . . . . C11 − C12 0 
. . . . . C11 − C12

We have 2 independent elastic moduli. A fourth-order isotropic tensor could also

be written as:

Cijkl = λδij δkl + µ(δik δjl + δil δjk ). (198)

AE 227 68 Dr. Krishnendu Haldar

7.1 Physical meaning of elastic moduli

This translates into Voigt notation as:

 
λ + 2µ λ λ 0 0 0
 . λ + 2µ λ 0 0 0
 
 . . λ + 2µ 0 0 0 
[C]ij = 
 .
 (199)
 . . 2µ 0 0  
 . . . . 2µ 0 
. . . . . 2µ

In a tensorial form we can write:

σij = λεkk δij + 2µεij (200)

or
1 λ 
εij = σij − σkk δij
2µ 3λ + 2µ
1+ν ν
= σij − σkk δij (201)
E E

where, E = µ(3λ+2µ)
λ+µ
and is called modulus of elasticity or Young modulus, and
λ
ν = 2(λ+µ) is referred to as Poisson’s ratio.

7.1 Physical meaning of elastic moduli

7.1.1 Simple tension
Consider a simple tension test along the x-direction. The state of stress is then
represented by
 
σ 0 0
σij =  0 0 0 (202)
0 0 0

The corresponding strain relationship could be written as

σ 
E
0 0
εij =  0 − Eν σ 0  (203)
0 0 − Eν σ

Therefore, E = εσx and is simply the slope of stress-strain (uniaxial) curve, while
ν = − eεxy = − eεxz is negative ratio of the transverse strain to the axial strain. Now
when εx > 0, i.e., stretching, then both εy < 0 and εz < 0. This implies ν > 0.

AE 227 69 Dr. Krishnendu Haldar

7.1 Physical meaning of elastic moduli

7.1.2 Pure torsion

If a thin-walled cylinder is subjected to a torsional loading, the state of stress on
the surface of the cylindrical sample is given by
 
0 τ 0
σij = τ 0 0
0 0 0

The corresponding strain relationship could be written as

 τ 
0 2µ 0
τ
εij =  2µ 0 0
0 0 0

and the shear modulus is given by µ = τ /2εxy = τ /γxy . It is simply the slope of
shear stress-shear strain curve.

7.1.3 Hydrostatic compression/tension

This type of test would be realizable if the sample was placed in a high-pressure
chamber. The state of the stress for this case is
 
−p 0 0
σij =  0 −p 0  = −pδij .
0 0 −p

The corresponding strain relationship could be written as

 1−2ν 
− E p 0 0
εij =  0 − 1−2ν
E
p 0 
1−2ν
0 0 − E p

The dilatation that represents the change in material volume is thus given by
ϑ = εkk = −3(1 − 2ν)p/E, which can be written as

p = −κϑ

where κ = E/3(1 − 2ν) is called the bulk modulus of elasticity. This additional
elastic constant represents the ratio of pressure to the dilatation. We can think
it as the volumetric stiffness of the material. Larger the value of κ, the material
becomes incompressible. So, we have five constants λ, µ, E, ν and κ. However only
two of them are required to characterize the material.

Read page no 81-90; 323-330 of the text book.

AE 227 70 Dr. Krishnendu Haldar

8 Formulation and solution strategies
Review of field equations:

1. Strain displacement relation

1 
εij = ui,j + uj,i
2

2. Equilibrium equations

σij,j + Fi = 0 (204)

3. Constitutive equations (Hooke’s law)

σij = λεkk δij + 2µεij (205)

1+ν ν
or, εij = σij − σkk δij (206)
E E

Recall that the compatibility equations are required when strains are specified
arbitrarily. However, if the displacements are included in the formulation, the so-
lution normally generates single valued displacements and the strain compatibility
is automatically satisfied.
So, the system involves total 15 unknowns and 15 equations and so the system
is solvable. However, particularly in 3D problem, analytical methods are normally
impossible and often further simplification is required to solve problems of interest.
Using (204) and (205) we can further write

λεkk,j δij + 2µεij,j + Fi = 0

Moreover, εij = 12 (ui,j + uj,i ), so εkk = uk,k and εij,j = 21 (ui,jj + uj,ij ). Then the
above equation is written as

λεkk,i + 2µεij,j + Fi = 0
λuk,ki + µ(ui,jj + uj,ij ) + Fi = 0

or

µui,jj + (λ + µ)uk,ki + Fi = 0 (207)

2
µ∇ u + (λ + µ)∇(∇ · u) + F = 0 (208)

AE 227 71 Dr. Krishnendu Haldar

8.1 Boundary conditions

8.1 Boundary conditions

The boundary conditions are provided either as displacements (Dirichlet) or trac-
tions (Neumann) at boundary points. So ∂B = ∂Bt ∪ ∂Bu . Boundary conditions
play a very crucial role in properly formulation and solving elasticity problems.
Improper specification results in either no solution or a solution different from the
original problem.

Read from pn 96-100.

8.2 Principle of superposition

Figure 34: Two dimensional superposition.

This technique applies to any problem that is governed by linear equations.

Under the assumptions of small deformations and linear elastic constitutive equa-
tions, all elasticity field equations are linear. Furthermore the boundary conditions
are also linear. Under these conditions principle of superposition holds (Fig. 34).
That is:
(1) (1) (1)
For a given problem domain, if the state {σij , εij , ui } is a solution to the elastic
(1) (1) (2) (2) (2)
problem with body force Fi and surface traction Ti , and the state {σij , εij , ui }
(2) (2)
is a solution to the elastic problem with body force Fi and surface traction Ti ,
(1) (2) (1) (2) (1) (2)
then the state the state {σij + σij , εij + εij , ui + ui } will be a solution to the
(1) (2) (1) (2)
elastic problem with body force Fi + Fi and surface traction Ti + Ti .

AE 227 72 Dr. Krishnendu Haldar

8.3 Saint-Venant principle

Figure 35: (a) Statically equivalent loading and (b) Saint-Venant principle.

8.3 Saint-Venant principle

This behavior can be generalized by considering an elastic solid with an arbitrary
loading T n over a boundary portion ∂B ∗ , as shown in Fig. 35(b). Experience
shows that the boundary loading will produce detailed and characteristic effects
only in the vicinity of ∂B ∗ . That is at points far away from ∂B ∗ the stresses
generally depend more on the resultant F R of the tractions rather than on the
exact distribution. Thus the characteristic signature of the generated stress, strain,
and displacement for a given boundary loading tend to disappear as we move away
from the boundary loading points. This is Saint-Venant principle. Saint-Venant
principle allows us to change the given boundary conditions to a simpler statically
equivalent statement without affecting the resulting solutions.

9 Two-dimensional formulation
9.1 plane strain
Consider a long cylindrical (prismatic) body as shown in Fig. 37(a). If the body
force and traction are independent of the z-coordinate and have no z-component,
then the deformation field within the body can be taken in the following reduced
form:

u = u(x, y), v = v(x, y), w = 0. (209)

AE 227 73 Dr. Krishnendu Haldar

9.1 plane strain

Figure 36: Plane strain problem.

Then,
∂u ∂v 1  ∂u ∂v 
εx = , εy = , εxy = + , εz = εxz = εyz = 0. (210)
∂x ∂y 2 ∂y ∂x

From Hook’s law (??) we write

1 ν
εx = σx − (σy + σz ) (211)
E E
1 ν
εy = σy − (σx + σz ) (212)
E E
1 ν
εz = σz − (σx + σy ) (213)
E E
1+ν
εxy = τxy (214)
E
1+ν
εyz = τyz (215)
E
1+ν
εzx = τzx (216)
E
For plane-strain assumptions, εyz = 0 =⇒ τyz = 0, εzx = 0 =⇒ τzx = 0 and
εz = 0 =⇒ σz = ν(σx + σy ). Substituting back the last relation in (211) and

AE 227 74 Dr. Krishnendu Haldar

9.1 plane strain

(212), we obtain

1 − ν2 ν(1 + ν)
εx = σx − σy (217)
E E
1 − ν2 ν(1 + ν)
εy = σy − σx (218)
E E
1+ν
εxy = τxy (219)
E
From the strain compatibility relation (equation (95)) we can write

∂ 2 εx ∂ 2 εy ∂ 2 εxy
+ = 2
∂y 2 ∂x2 ∂x∂y

Combining (217), (218) and (219) in the above relation we get

! !
2 2 2 2
∂ σ x ∂ σ y ∂ σy ∂ σ x ∂ 2 τxy
(1 − ν 2 ) 2 − ν(1 + ν) 2 + (1 − ν 2 ) 2 − ν(1 + ν) 2 = 2(1 + ν)
∂y ∂y ∂x ∂x ∂x∂y
! !
∂ 2 σx ∂ 2 σy ∂ 2 σy ∂ 2 σx ∂ 2 τxy
(1 − ν) 2 − ν 2 + (1 − ν) 2 − ν = 2 (220)
∂y ∂y ∂x ∂x2 ∂x∂y

Recall the equilibrium equation as:

∂σx ∂τxy
+ + Fx = 0
∂x ∂y
∂τxy ∂σx
+ + Fy = 0
∂x ∂y

We now want to eliminate τxy from (220) with the help of equilibrium equation.
Taking partial-x and partial-y with respect to the first and second equation, re-
spectively, we write

∂ 2 σx ∂ 2 τxy ∂Fx
+ + = 0
∂x2 ∂y∂x ∂x
∂ 2 τxy ∂ 2 σx ∂Fy
+ + = 0
∂x∂y ∂y 2 ∂y
Adding we write

∂ 2 σx ∂ 2 σx ∂ 2 τxy  ∂Fx ∂Fy 

+ +2 + + =0
∂x2 ∂y 2 ∂y∂x ∂x ∂y

AE 227 75 Dr. Krishnendu Haldar

9.2 plane stress

2
Now replacing 2 ∂∂y∂x
τxy
from (220) we write
! !
∂ 2 σx ∂ 2 σy ∂ 2 σy ∂ 2 σx
(1 − ν) 2 − ν 2 + (1 − ν) 2 − ν
∂y ∂y ∂x ∂x2
∂ 2 σx ∂ 2 σx  ∂Fx ∂Fy 
= − 2 − − +
∂x ∂y 2 ∂x ∂y
1  ∂Fx ∂Fy 
∇2 (σx + σy ) = − + (221)
1 − ν ∂x ∂y

9.2 plane stress

Figure 37: Plane stress problem.

In this case the domains are generally bounded by two parallel planes separated
by a distance that is small in comparison to other dimensions. In this case

σx = σx (x, y), σy = σy (x, y), τxy = τxy (x, y), σz = τxz = τyz = 0. (222)

AE 227 76 Dr. Krishnendu Haldar

9.2 plane stress

From Hook’s law (??) we write

1
εx = (σx − νσy ) (223)
E
1
εy = (σy − νσx ) (224)
E
ν
εz = − (σx + σy ) (225)
E
1+ν
εxy = τxy (226)
E
εyz = 0 (227)
εzx = 0 (228)

From the strain compatibility relation (equation (95)) we can write

∂ 2 εx ∂ 2 εy ∂ 2 εxy
+ = 2
∂y 2 ∂x2 ∂x∂y

Combining (223), (224) and (226) in the above relation we get

! !
∂ 2 σx ∂ 2 σy ∂ 2 σy ∂ 2 σx ∂ 2 τxy
− ν + − ν = 2(1 + ν) (229)
∂y 2 ∂y 2 ∂x2 ∂x2 ∂x∂y

Recall the equilibrium equation as:

∂σx ∂τxy
+ + Fx = 0
∂x ∂y
∂τxy ∂σx
+ + Fy = 0
∂x ∂y

We now want to eliminate τxy from (229) with the help of equilibrium equation.
Taking partial-x and partial-y with respect to the first and second equation, re-
spectively, we write

∂ 2 σx ∂ 2 τxy ∂Fx
+ + = 0
∂x2 ∂y∂x ∂x
∂ 2 τxy ∂ 2 σx ∂Fy
+ + = 0
∂x∂y ∂y 2 ∂y
Adding we write

∂ 2 σx ∂ 2 σx ∂ 2 τxy  ∂Fx ∂Fy 

+ + 2 + + =0
∂x2 ∂y 2 ∂y∂x ∂x ∂y

AE 227 77 Dr. Krishnendu Haldar

9.3 Airy stress function

2
Now replacing 2 ∂∂y∂x
τxy
from (229) we write
! !
∂ 2 σx ∂ 2 σy ∂ 2 σy ∂ 2 σx  ∂ 2σ
x ∂ 2 σx   ∂F
x ∂Fy 
− ν + − ν = −(1 + ν) + − (1 + ν) +
∂y 2 ∂y 2 ∂x2 ∂x2 ∂x2 ∂y 2 ∂x ∂y
 ∂F ∂Fy 
x
∇2 (σx + σy ) = −(1 + ν) + (230)
∂x ∂y

9.3 Airy stress function

Numerous solutions to plane strain and plane stress problems can be determined
through the use of particular stress function technique. The method employs the
Airy stress function and will reduce the general formula to a single governing
equation in terms of a single unknown. Let us assume that the body force could
be derived from a potential V , such that
∂V ∂V
Fx = − , Fy = −
∂x ∂y
Then the following stress functions satisfy the equilibrium equation:

∂ 2φ
σx = +V (231)
∂y 2
∂ 2φ
σy = +V (232)
∂x2
∂ 2φ
τxy = − (233)
∂x∂y

where φ = φ(x, y) is an arbitrary form called the Airy stress function.

9.3.1 Plane strain

Substituting Airy stress function and body force potential in
1  ∂Fx ∂Fy 
∇2 (σx + σy ) = − +
1 − ν ∂x ∂y
we write
1
∇2 (∇2 φ + 2V ) = ∇2 V
1−ν
1 − 2ν 2
=⇒ ∇4 φ = − ∇V (234)
1−ν

AE 227 78 Dr. Krishnendu Haldar

9.3 Airy stress function

9.3.2 Plane stress

Substituting Airy stress function and body force potential in
 ∂F ∂Fy 
2 x
∇ (σx + σy ) = −(1 + ν) +
∂x ∂y
we write

∇2 (∇2 φ + 2V ) = (1 + ν)∇2 V
=⇒ ∇4 φ = −(1 − ν)∇2 V (235)
∂2 ∂ 2 ∂ 4 ∂ 4 ∂4
Note that ∇2 = ∂x 4
2 + ∂y 2 and ∇ = ∂x4 + 2 ∂x2 ∂y 2 + ∂y 4 . Moreover, when there is

no body force, we need to solve only one equation, same for plane strain or plain
stress problem, i.e.,

∇4 φ = 0. (236)

AE 227 79 Dr. Krishnendu Haldar

10 Axially loaded bars
Bars are the members that can take only 1D axial loading (i.e. no transverse
loading) or torsion. For an axially loaded bar we can write:

du σx (x) P (x)
εx = = = (237)
dx E(x) A(x)E(x)

So,
Z L
P (x)
∆= dx (238)
0 A(x)E(x)

10.1 Deformation of axially loaded bars

Figure 38: Axially loaded bar.

For an axially loaded bar with uniform cross section and constant elastic mod-
ulus (Fig. 38) we can write the the total displacement at the end point as
Z L
P PL
∆= dx = (239)
0 AE AE

10.2 Deflection of a bar due to its own weight

Let us consider the origin at the end of the rod in the undeformed state. If γ be
the weight per unit volume, the force due to the length x on the cross-sectional
area (A) is P (x) = −γxA (Fig. 39).
L L
γL2
Z Z
P (x) γAx WL
∆= dx = − dx = − =− (240)
0 A(x)E(x) 0 AE 2E 2AE

where, W = γLA is the total weight of the body.

AE 227 80 Dr. Krishnendu Haldar

10.3 Deflection of a bar due to rotation

Figure 39: Vertical bar/rod under gravity.

Figure 40: Rotating bar.

10.3 Deflection of a bar due to rotation

Let a homogeneous bar of length 2L is rotating with a constant angular velocity ω
(Fig. 40). Then the tension due to the centrifugal force acting at a radial distance
x from the origin (center of rotation) due to the small mass element dm = γAdx/g
is
Z L Z L
2 1
P (x) = (γAdx)ω x = (γAω 2 )xdx = γAω 2 (L2 − x2 ). (241)
x x 2g

AE 227 81 Dr. Krishnendu Haldar

10.4 Deflection of two hinge-ended elastic bars

Here, γ be the weight per unit volume. Then the deflection at the end point is
L L 1 γAω 2 (L2 − x2 ) γω 2 L3
Z Z
P (x) 2g
∆= dx = dx = . (242)
0 A(x)E(x) 0 AE 3Eg

10.4 Deflection of two hinge-ended elastic bars

Figure 41: Hinged elastic bars.

If T be the tension in the string, then:

P = 2T sin θ

Now,

T L0
= (L − L0 ) = L0 (1 − cos θ), ⇒ T = AE(1 − cos θ)
AE
Then,

P = 2AE(1 − cos θ) sin θ = 2AE(2 sin2 θ) sin θ ≈ AEθ3 .

Now for small angle, CC 0 = ∆ = Lθ. Then,

r
3 P
P ≈ AE∆3 /L3 . ⇒ ∆=L
AE

AE 227 82 Dr. Krishnendu Haldar

11 Torsion

Figure 42: Cylinder with torque.

A prismatic (i.e. uniform cross-section) bar (slender rod) subjected to a twist-

ing moment is said to be a state of torsion. In this section we will analyze the
stress developed in members subjected to torsion. Let a bar AB of circular sec-
tion be subjected to a torque T as shown in Fig. 42. If we cut the member by a
section at C, the portion AC (or CB) will be in equilibrium under the externally
applied torque T and the internally distributed stresses that produce an equal and
opposite torque at C. The shear stress distribution will produce a pure moment
about the longitudinal axis without any resultant force in the axial direction. For
equilibrium, we only have a moment, and no resultant force.

11.1 Member of circular cross-section

To establish a relationship between the torque, stress and deformation of a bar of
circular section subjected to a torsion. We will make the following assumptions:
1. The material is homogeneous, linear elastic, and isotropic.
2. Plane sections perpendicular to the longitudinal axis remain plane after the
torque applied. In other words, there is no warping of transverse planes and
hence no distortion of the parallel planes normal to the axis of the member.
Consider Fig. 43. We can write
BB 0 = rφ = Lγ.
Here, r is the radius O2 B, φ is the angle of twist and γ is the shear strain. If µ is
the shear modulus, then
µrφ
τ = µγ = (243)
L
AE 227 83 Dr. Krishnendu Haldar
11.1 Member of circular cross-section

Figure 43: Cylinder with torque.

So the total torque is given by,

Figure 44: Torsion in circular shaft.

Z Z Z
µrφ µφ µφJp
T = r(τ dA) = r( )dA = r2 dA = (244)
A A L L A L
R
where Jp = A r2 dA is the polar moment of inertia. For a circular cross-section
4
Jp = πd
32
, where d is the diameter. From the above equation, we write
TL
φ=
µJp
Substituting back in (243), we write
Tr
τ= (245)
Jp

AE 227 84 Dr. Krishnendu Haldar

11.1 Member of circular cross-section

T R0
and the maximum torque, τ max = Jp
, occurs at the periphery (Fig. 45), where
Ro is the outer radius

Figure 45: Cylinder with torque.

Some remarks

(a) (b)

Figure 46: (a) Torque on cylindrical body. (b) An infinitesimal cylindrical element.

So far we assume that the shear stress is acting on the plane perpendicular to
the axis. However, to understand the problem further, an infinitesimal cylindrical
element, shown in Fig. 45, is isolated. The stress acting on the cross-sectional

AE 227 85 Dr. Krishnendu Haldar

11.2 Hollow circular shafts

plane is given by
Tr
τ=
Jp
On adjoining parallel planes of a disc like element, these stresses act in opposite
direction. However, to maintain equilibrium, equal shear stresses must act on
the axial planes (planes aef and bcg). Shear stresses acting on the axial planes
follow the same variation of intensity as do the shear stresses on the cross-sectional
plane (Fig. 45, red arrows). Recall that, shear stresses can be transformed into an
equivalent system of normal stresses acting at angles of 45◦ . Numerically, these
stresses are related to each other as: τ = σ1 = −σ2 . Therefore, if the shear strength
of a material is less than its strength in tension, a shear failure takes place on the
cross-sectional plane. This kind of failure occurs gradually and exhibits ductile
behavior. Alternatively, if the converse is true, i.e., σ < τ , a brittle fracture is
caused by the tensile stresses along a helox forming an angle 45◦ . This phenomena
is shown in

Figure 47: Potential torsional failure surfaces in ductile and brittle materials.

11.2 Hollow circular shafts

We have seen that the shear stress is maximum at the surface of a shaft and zero
at the center (Fig. 44). This means the material near the the center contributes
little to the shaft’s torque-resisting capacity. Shafts of large diameter can therefore
be provided with a concentric bore for material saving.

AE 227 86 Dr. Krishnendu Haldar

11.3 Tapered circular shafts

11.3 Tapered circular shafts

Figure 48: Tapered circular shaft.

The solid circular shaft in Fig. 48 is tapered in cross-section, with its diameter
varying from 2R at one end to 4R at the other over a length L. We like to
determine the angle of twist when the shaft is subjected to a pair of equal and
opposite torques T applied at the ends.
For the thin slice of thickness ∆x and radius R(x) at a distance x from the
smaller end. Then we write
2R − R R
R(x) = R + x = (x + L)
L L
Now every section of the shaft is subjected to the torque T . Then we can write
T dx
dφ =
µJp (x)

Now,

πR(x)4 π R4
Jp (x) = = (L + x)4
2 2 L4
Substituting back we get the total twist φ as,
Z L
2L4 T L
Z
T dx 7T L
φ= dx = 4 4
=
0 µJp (x) πR µ 0 (L + x) 12πµR4

AE 227 87 Dr. Krishnendu Haldar

11.4 Combined torsion and axial force

Figure 49: Combined torsional and axial force.

11.4 Combined torsion and axial force

Figure Fig. 49 shows a shaft subjected to a torsion T and an axial force F . The
stresses acting on an element on the surface of the shaft are τ due to the torque
and σx =(F/A). The two principal stresses are
r 
σx σx 2
σ1,2 = ± + τ2
2 2

AE 227 88 Dr. Krishnendu Haldar

12 Shearing forces and bending moments of beams
A Beam is a bar of material subjected to transverse forces (loads or couples)
which mainly cause it to bend in the axial plane. Many kinds of forces can act
on a beam, e.g., concentrated or point force, distributed (uniform or non-uniform),
etc. A beam is usually supported at its ends or intermediate points to help it carry
the applied transverse forces.

12.1 Supports

Figure 50: Different types of supports.

Different types of supports are shown in Fig. 50. (a) to (d) are simply or freely
supported where the reaction forces are normal to the support. At this point, the
the beam is free to rotate. (e) and (f) are the pinned or hinged support. Here the
reaction passes through the hinge and can act in any direction so as to maintain
equilibrium with the external forces acting on the beam. The hinged also allows
to rotate the member at the support. The fixed or built-in support is shown in (e).
It does not permit any rotation at this end. This support exerts a reaction force
as well as a moment at the end.

AE 227 89 Dr. Krishnendu Haldar

12.2 Types of beams

12.2 Types of beams

Figure 51: Different types of beams.

Beams are often classified according to the type of support they have (Fig. 51).
A cantilever beam has one end fixed and the other end free (a). If the beam is
freely supported at both ends, it is said simply supported (b) and can only carry
forces that are perpendicular to the axis. To carry some axial load, at least one
support should be hinged (c). Overhanging beams have a part that extend beyond
a support (d-e). If it is fixed at both ends (f), we call it fixed beam. If a cantilever
provided with additional simple support, it is called propped cantilever (g). A
continuous beam has more than two supports (h).

AE 227 90 Dr. Krishnendu Haldar

12.2 Types of beams

Figure 52: Normal force, shear force, and bending moment.

12.2.1 Forces on beam section

Let us consider a cantilever beam as shown in Fig. 52(a). We consider the vertical
axis positive in the downward direction as most of the cases the loads are applied
on the top of the member and acts downwards. Let us now consider the forces
on the section m − n (Fig. 52b) to maintain the equilibrium of the free body.
We present the stress resultant in Fig. 52(c) by the three quantities N, V and M ,
called, respectively, the normal force, shear force, and the bending moment, at that
section. These quantities will be considered positive when they have the directions
shown in Fig. 52(c). Such sign conventions, although arbitrary, must carefully
observed to avoid confusions. The force and moment equilibrium equations
X X X
Fx = 0, Fy = 0, Mz = 0

AE 227 91 Dr. Krishnendu Haldar

12.3 Relations between load intensity, shearing force and bending moment

gives

N = P cos θ
V = −P sin θ
M = −P sin θ · x

As shown in Fig. 53, the positive normal force is away from the face. A positive

Figure 53: Normal force, shear force and bending moment sign convention.

shear force is one that has a clockwise sense of rotation about a point inside the
free body. The bending moments M are positive when they tend to bend the
element concave upwards.

12.3 Relations between load intensity, shearing force and

bending moment

Figure 54: Relation between shear stress and bending moment.

AE 227 92 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

In most practical problems, we encounter beams which are horizontal and sub-
jected only to vertical loads. In such cases, there will be no normal force N on
any cross section. From the left figure we can say that, as there is no other forces
acting in the free body portion of length ∆x, then the right face should have equal
and opposite shear force. However, the bending moment will be changing on the
right hand side as M + ∆M . From moment balance equation we can write
∆M
−M + (M + ∆M ) − V ∆x = 0, ⇒ =V
∆x
In the limiting case ∆x → 0, we write
dM
=V (246)
dx
If there is some distributed load with intensity w (per unit length), then the
moment equilibrium gives
∆M
−M + (M + ∆M ) − V ∆x + w∆x(∆x/2) = 0, ⇒ = V − w∆x/2
∆x
n the limiting case ∆x → 0, we write
dM
=V
dx
So we get back the same previous equation. Force equilibrium gives
∆V
−V + (V + ∆V ) + w∆x = 0, ⇒ = −w
∆x
In the limiting case ∆x → 0, we write
dV
= −w (247)
dx
So we can write:
dV d2 M
= = −w (248)
dx dx2

12.4 Shearing force and bending moment diagrams

These diagrams show the variation of the shearing force and bending moment over
the whole length of a beam to some scale. Methods of determining theses diagrams,
often abbreviated SFD and BMD, will now be indicated for various types of beams
subjected to various types of loads.

AE 227 93 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

Figure 55: Cantilever with end-point load.

12.4.1 Cantilever with end-point load

Force equilibrium (Fig. 55):

V (x) = −P (249)

Then,
dM
= −P
dx
Z M (x) Z x
dM = −P dx
M (0) 0
M (x) − M (0) = −P x

Now M (0) = 0, then

M (x) = −P x (250)

Plotting (249) and (250) we have in Fig. 56

12.4.2 Cantilever with number of concentrated loads

0 < x < a1
Force equilibrium (Fig. 57):

V (x) = −P1 (251)

AE 227 94 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

Figure 56: SFD and BMD.

Figure 57: Cantilever with number of concentrated loads.

Then,
dM
= −P1
dx
Z M (x) Z x
dM = −P dx
M (0) 0
M (x) − M (0) = −P x
Now M (0) = 0, then
M (x) = −P1 x (252)

AE 227 95 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

a1 < x < a 2
Force equilibrium (Fig. 57):

V (x) = −(P1 + P2 ) (253)

Then,
dM
= −(P1 + P2 )
dx
Z M (x) Z x
dM = −(P1 + P2 ) dx
M (a1 ) a1
M (x) − M (a1 ) = −(P1 + P2 )(x − a1 )

Now M (a1 ) = −P1 a1 , (from (252)) then

M (x) = −P1 a1 − (P1 + P2 )(x − a1 )

= −P1 x − P2 (x − a1 ) (254)

a2 < x < a 3
Force equilibrium (Fig. 57):

V (x) = −(P1 + P2 + P3 ) (255)

Then,
dM
= −(P1 + P2 + P3 )
dx
Z M (x) Z x
dM = −(P1 + P2 + P3 ) dx
M (a2 ) a2
M (x) − M (a2 ) = −(P1 + P2 + P3 )(x − a2 )

Now M (a2 ) = −P1 a2 − P2 (a2 − a1 ), (from (254)) then

M (x) = −P1 a2 − P2 (a2 − a1 ) − (P1 + P2 + P3 )(x − a2 )

= −P1 x − P2 (x − a1 ) − P3 (x − a3 ) (256)

Note that M (l) = −P1 l − P2 (l − a1 ) − P3 (l − a3 )

Plotting (251), (253), (255) and (252), (254), (256) we have in Fig. 58

AE 227 96 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

Figure 58: SFD and BMD.

12.4.3 Cantilever with a uniformly distributed load

Force equilibrium (Fig. 59):

V (x) = −wx (257)

Then,
dM
= −wx
dx
Z M (x) Z x
dM = −w xdx
M (0) 0
2
M (x) − M (0) = −wx /2

Now M (0) = 0, then

M (x) = −wx2 /2 (258)

Plotting (257) and (258) we have in Fig. 60

AE 227 97 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

Figure 59: Cantilever with a uniformly distributed load.

12.4.4 Cantilever with concentrated and distributed load

0 < x < l/2
Force equilibrium (Fig. 61):
V (x) = −P (259)

AE 227 98 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

Figure 60: SFD and BMD.

Figure 61: Cantilever with concentrated and distributed load.

Then,
dM
= −P
dx
Z M (x) Z x
dM = −P dx
M (0) 0
M (x) − M (0) = −P x

AE 227 99 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

Now M (0) = 0, then

M (x) = −P x (260)

l/2 < x < l

Force equilibrium (Fig. 61):
l

V (x) = −P − w x − (261)
2
Then,
dM  l
= −P − w x −
dx 2
Z M (x) Z x Z x
l
dM = −P dx − w dx x−
M (l/2) l/2 l/2 2
Z x Z (x−l/2) 
l  l
= −P dx − w x− d x−
l/2 0 2 2
 l  w  l 2

M (x) − M (l/2) = −P x − − x−
2 2 2
Now M (l/2) = −P l/2, then
w l 2
M (x) = −P x − x− (262)
2 2
Plotting (259), (261) and (260), (262) we have in Fig. 62 Note that
w 2
M (l) = −P l − l (263)
8

12.4.5 Simply-supported beam with a single concentrated load

Force equilibrium gives:

RA + RB = P

Moment balance at point B gives:

RA l − P (l − a) = 0

So we get
P (l − a) Pa
RA = , RB =
l l
AE 227 100 Dr. Krishnendu Haldar
12.4 Shearing force and bending moment diagrams

Figure 62: SFD and BMD.

0<x<a
Force equilibrium (Fig. 63):
P (l − a)
V (x) = RA = (264)
l
Then,
dM P (l − a)
=
dx l
Z M (x)
P (l − a) x
Z
dM = dx
M (0) l 0
P (l − a)
M (x) − M (0) = x
l
Now M (0) = 0, then
P (l − a)
M (x) = x (265)
l

a<x<l
Force equilibrium (Fig. 63):
Pa
V (x) = RA − P = −RB = − (266)
l
AE 227 101 Dr. Krishnendu Haldar
12.4 Shearing force and bending moment diagrams

Figure 63: Simply-supported beam with a single concentrated load.

Then,
dM Pa
= −
dx l
Z M (x)
Pa x
Z
dM = − dx
M (a) l a
P a 
= − x−a
l
Pa  
M (x) − M (a) = − x−a
l
P a(l−a)
Now M (a) = l
, then

P a(l − a) P a  
M (x) = − x−a
l x l
= Pa 1 − (267)
l
AE 227 102 Dr. Krishnendu Haldar
12.4 Shearing force and bending moment diagrams

Plotting (264), (266) and (265), (267) we have in Fig. 64

Figure 64: SFD and BMD.

12.4.6 Simply-supported beam with number of concentrated loads

Force equilibrium gives:
RA + RB = P1 + P2 + P3
Moment balance at point B gives:
RA l − P1 (l − a1 ) − P2 (l − a2 ) − P3 (l − a3 ) = 0
So we get
P1 (l − a1 ) P2 (l − a2 ) P3 (l − a3 ) P1 a1 P2 a2 P3 a3
RA = + + , RB = + +
l l l l l l

0 < x < a1
Force equilibrium (Fig. 65):
V (x) = RA (268)
Then,
dM
= RA
dx
Z M (x) Z x
dM = RA dx
M (0) 0
M (x) − M (0) = RA x

AE 227 103 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

Figure 65: Simply-supported beam with a single concentrated load.

Now M (0) = 0, then

M (x) = RA x (269)

a1 < x < a 2
Force equilibrium (Fig. 65):
V (x) = RA − P1 (270)
and
M (x) = RA x − P1 (x − a1 ) (271)

a2 < x < a 3
Force equilibrium (Fig. 65):
V (x) = RA − P1 − P2 (272)
and
M (x) = RA x − P1 (x − a1 ) − P2 (x − a2 ) (273)

AE 227 104 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

a3 < x < a l
Force equilibrium (Fig. 65):
V (x) = RA − P1 − P2 − P3 (274)
and
M (x) = RA x − P1 (x − a1 ) − P2 (x − a2 ) − P3 (x − a3 ) (275)
Plotting (268), (270), (274), (??) and (269), (271), (275), (??) we have in Fig. 66
Here it is assumed that RA < (P1 + P2 ) so that a change in of the shear force

Figure 66: SFD and BMD.

occurs under the load P2 and consequently the bending moment at x = a2 will be
maximum.

12.4.7 Simply-supported beam with distributed load

Force equilibrium gives:
RA + RB = wl
Moment balance at point B gives:
RA l − wl2 /2 = 0

AE 227 105 Dr. Krishnendu Haldar

12.4 Shearing force and bending moment diagrams

Figure 67: Simply-supported beam with distributed load.

So we get
ql
RA = = RB
2
Force equilibrium (Fig. 67):
l
V (x) = RA − wx = −w(x − ) (276)
2
Then,
dM l
= −w(x − )
dx 2
Z M (x) Z x
l
dM = −w (x − )dx
M (0) 0 2
wl
M (x) − M (0) = −wx2 /2 + x
2
Now M (0) = 0, then
wl
M (x) = −wx2 /2 + x (277)
2
Plotting (276) and (277) we have in Fig. 68 Maximum moment means dM dx
= 0
i.e., when V = 0. In this case Maximum moment takes place at x = l/2 for which
2
Mmax = wl8

AE 227 106 Dr. Krishnendu Haldar

 Figure 68: SFD and BMD.

13 Stress due to bending moment

In the previous section we saw that the shear stresses and bending moments are
developed within the cross sectional area. These resisting shears and moments are
a sum total of the internal stress distribution across the cross section of the beam.
With suitable assumptions, it is possible to determine these internal stresses. We
will restrict ourselves in this section to the internal effect of the bending (pure
bending) only.

13.1 Theory of simple bending

Fig. 69(a) shows part of a beam symmetrical section subjected only to a bending
moment that acts in the plane of symmetry. It causes the be beam to bend
upwards, becoming concave at top in the plane of symmetry as shown in Fig. 69(b).
The bending can be visualized as causing the fibers of the of the beam on the
concave side AB to contract and those on the convex side CD to elongate. There
will be thus be a surface SF between AB and CD that neither contracts nor
elongates. i.e., neutral surface. The moment M thus induces compressive stress in
the cross section GH above the neutral surface and tensile stress below. Let the
resultant compressive force Fc and the tensile force Ft be acting at a distance a
apart. Then from statics Fc = Ft and Ft a = Fc a = M . The analysis is developed
to determine Fc , Ft and M based on simple or pure bending theory. The following

AE 227 107 Dr. Krishnendu Haldar

13.2 Deformation geometry

Figure 69: Simple bending.

assumptions are made to develop the theory:

1. The transverse cross sectional area of the beam is symmetrical about an axis
passing through the centroid. This axis is parallel to the plane of bending
(Fig. 69a).

2. These transverse sections of the beam that are plane and normal to the
longitudinal axis before the bending, remain plane and normal after it. In
other words the cross section rotates about an axis (neutral axis). Thus each
layer of the beam is free to contract or expand longitudinally or laterally
under internal stress, as if the layers are separated from each other.

3. The depth and breadth of the beam are small compared to its length.

4. The deformations of the beam are small.

13.2 Deformation geometry

At point P of the plane curve P P1 (Fig. 70), the radius of curvature is P C = ρ,
and the angle made by the tangent at P with the horizontal is θ. Consider a point
P1 along the curve such that P P1 = ∆s. The change in the angle of the tangent
over the length P P1 is ∆θ. Defining the curvature as the rate of change of the
inclination of the tangent along the length of the curve, we have
dθ ∆θ ∆θ 1
= lim = lim = (278)
ds ∆s→0 ∆s ∆s→0 ρ∆θ ρ
GH and J1 K1 are two plane transverse sections a distance ∆x apart. When
the beam is subjected to a bending moment M , it bends into the arc of a circle,

AE 227 108 Dr. Krishnendu Haldar

13.2 Deformation geometry

Figure 70: Deformation geometry.

Figure 71: Beam subjected to bending moment.

and the section J1 K1 rotates with respect to GH through an angle ∆θ (Fig. 71).
Since it is assumed that the plane sections remain plane after the bending, section
J1 K1 rotates about an axis passing through F1 . The original length GJ1 becomes
GJ and, similarly the original length HK1 becomes HK. Thus the portion on
the concave side AB has contracted and the portion on the convex side CD has
elongated. If the portion M N is at a distance y from the neutral surface, the

AE 227 109 Dr. Krishnendu Haldar

13.3 Equilibrium conditions

longitudinal strain in it is given by

N1 N M N − M N1 M N − S1 F 1 (ρ + y)∆θ − ρ∆θ y
ε= = = = = . (279)
M N1 M N1 S1 F 1 ρ∆θ ρ
Now for linear elastic material σ = Eε. So we can write
σ y σ E
= or = . (280)
E ρ y ρ
Note that the stress linearly varies with y.

13.3 Equilibrium conditions

Figure 72: Variation of longitudinal stress across the beam section and the area of
cross section symmetric with respect to the x − y plane.

Let the stress σ is acting through the differential area dA. Then the force
equilibrium gives,
Z Z Z
X Ey E
Fx = 0 ⇒ σdA = 0 ⇒ dA = 0 ⇒ ydA = 0
A A ρ ρ A

As E/ρ 6= 0, we have
Z
ydA = 0
A

This means the centroid of the cross section lies on the neutral axis. For a ho-
mogeneous and linear elastic beam, determining the centroid of the cross section
fixes the position of the neutral axis.

AE 227 110 Dr. Krishnendu Haldar

13.3 Equilibrium conditions

Since no moment is acting in the y-direction,

Z Z Z
X Ey E
My = 0 ⇒ zσdA = 0 ⇒ zdA = 0 ⇒ yzdA = 0
A A ρ ρ A
As E/ρ 6= 0, we have
Z
yzdA = 0.
A

This is possible as the cross section is symmetric with respect to the x − y plane.
This is also one of the basic assumption.
Finally,
Z Z Z
X Ey E
Mz = M ⇒ yσdA = M ⇒ zdA = M ⇒ y 2 dA = M (281)
A A ρ ρ A
Now
Z
I= y 2 dA
A

is the second moment of area (or the moment of inertia) of the section about the
neutral axis. Then the (281) gives
EI M E
=M or = (282)
ρ I ρ
Now combining with (280) we can further write:
σ E M
= = . (283)
y ρ I
The maximum stress due to bending at the fiber for which y is maximum. Thus
M
σmax = ymax . (284)
I
The quantity Z = I/ymax is a property of the cross section called the section
modulus.
In general if yt and yc be the distances from the neutral axis of the extreme fibers
on the tension and compression sides, then the maximum tensile and compressive
stresses are given by,
t M
σmax = yt
I
c M
σmax = yc
I

AE 227 111 Dr. Krishnendu Haldar

13.4 Shear stress due to varying bending moment

13.4 Shear stress due to varying bending moment

Figure 73: Variation of longitudinal stress across the beam section and the area of
cross section symmetric with respect to the x − y plane.

Let AR and BS be the two neighboring transverse sections of the beam, a

distance ∆x apart. Let b be the width of the cross section. Let the bending
moment, acting on the vertical plane of symmetry, be M and M + ∆M at these
two sections. The normal stress at a distance y from the neutral axis due to these
bending moments are σ = MI y on the left face ARR0 D and σ + ∆σ = (M +∆M I
)y
on
the right face BCSS 0 .
From the free body diagram of the element P Rrs, the force on the left and
right faces are
Z h/2 Z h/2
My (M + ∆M )y
F = dA and F + ∆F = dA.
y I y I

AE 227 112 Dr. Krishnendu Haldar

13.4 Shear stress due to varying bending moment

The unbalanced force is

Z h/2 Z h/2
∆M y ∆M
∆F = dA = ydA.
y I I y

This unbalanced force will create a shearing stress on the top face of the element
as shown in Fig. 73. If τ is the magnitude of the shear stress, then F = τ b∆x.
This gives,
Z h/2
∆M 1
τ= ydA.
∆x Ib y

and in the limit

Z h/2 Z h/2
dM 1 V VQ
τ= ydA = ydA = . (285)
dx Ib y Ib y Ib

Here,
Z h/2
Q= ydA.
y

For this particular example (rectangular cross section), we can further write,
h/2
b  h2
Z 
Q= bydy = − y2 .
y 2 4

So,

V  h2 2

τ= −y . (286)
2I 4
The variation of τ with y is parabolic. At y = h/2, τ = 0 and at y = 0, τ =
2
τmax = V8Ih . For a rectangular section, I = 12
1 3V
bh3 , then τmax = 2bh .

AE 227 113 Dr. Krishnendu Haldar

14 Beam deflection
The axis of a beam deflects from its original position when it is subjected to trans-
verse loads. Many situations require one to know how much a structure deflects, to
see if it is rigid enough to limit the deflections within the acceptable values. These
values depend on practical considerations like guarding against misalignment of
a machine part or avoiding cracks in the ceilings, etc. The deflection shape of a
beam is called an elastic line or deflection curve.

14.1 Differential equation for deflection

From Fig. 70 we can write ∆s ≈ ∆x and
dy
tan θ ≈ θ =
dx
Then
1 dθ d  dy  d  dy  d2 y
= = = = 2. (287)
ρ ds ds dx dx dx dx
This equation uses the absolute value of the curvature. Its sign is fixed as follows:
dy
If, as x increases the value of θ (or dx ) decreases, the bending moment is positive
2
d y
and dx2 will be negative. Similarly, a negative bending moment will cause the
dy d2 y
beam to bend concave side downwards, θ (or dx ) will increase with x and dx 2
d2 y
will be positive. Thus dx2 has a sign opposite to M . Moreover from the relation
1/ρ = M/(EI), we write
d2 y
EI = −M (288)
dx2
Then
d  d2 y  dM
EI 2 = − = −V (289)
dx dx dx
and
d2  d2 y  d2 M dV
2
EI 2
= − 2
=− =q (290)
dx dx dx dx
If E and I are constant, then
d2 y
EI = −M (291)
dx2
d3 y
EI 3 = −V (292)
dx
d4 y
EI 4 = q (293)
dx
AE 227 114 Dr. Krishnendu Haldar
14.1 Differential equation for deflection

and the slope

dy
θ= . (294)
dx

14.1.1 Cantilever with end-point load

Figure 74: Cantilever with end-point load.

AB is the cantilever beam with a uniform section subjected to a concentrated

load P at the free end B, as shown in Fig. 74. The bending moment at a distance
x from A is
M = −P (l − x)
Then,
d2 y
EI = −M = P (l − x) = P l − P x
dx2
So,
P lx2 P x3
EIy = − + c1 x + c2
2 6
Boundary conditions are:
dy
at x = 0, = 0 and y = 0
dx
These give c1 = c2 = 0. The equation of deflection is
P x2   dy Px  
y= 3l − x , θ= = 2l − x . (295)
6EI dx 2EI
The free end deflection and slope become:
P l3 P l2
yB = θB = . (296)
3EI 2EI
AE 227 115 Dr. Krishnendu Haldar
14.1 Differential equation for deflection

14.1.2 Cantilever with a load anywhere in the span

Figure 75: Cantilever with a load anywhere in the span.

AB is the cantilever beam with a uniform section subjected to a concentrated

load P at the mid-point B, as shown in Fig. 75.

0<x<a
The bending moment at a distance x from A is
M = −P (a − x)
Then,
d2 y
EI = −M = P (a − x) = P a − P x
dx2
So,
P ax2 P x3
EIy = − + c1 x + c2
2 6
Boundary conditions are:
dy
at x = 0, = 0 and y = 0
dx
These give c1 = c2 = 0. The equation of deflection is
P x2   dy Px  
y= 3a − x , θ= = 2a − x . (297)
6EI dx 2EI
The free end deflection and slope become:
P a3 P a2
yC = θC = . (298)
3EI 2EI
AE 227 116 Dr. Krishnendu Haldar
14.1 Differential equation for deflection

a<x<l
The bending moment at a distance x from C is

M =0

Then,

d2 y
EI =0
dx2
So,

EIy = c3 x + c4

Boundary conditions are:

dy P a2 P a3
at x = a, = and y =
dx 2EI 3EI
These give c3 = P a2 /2, c4 = −P a3 /6. The equation of deflection is

P a2   dy P a2
y= 3x − a , θ= = . (299)
6EI dx 2EI
The free end deflection and slope become:

P a2   P a2
yC = 3l − a θC = = θB . (300)
6EI 2EI

14.1.3 Cantilever with a uniformly distributed load

Figure 76: Cantilever with a uniformly distributed load.

AE 227 117 Dr. Krishnendu Haldar

14.1 Differential equation for deflection

AB is the cantilever beam with a uniform section subjected to a uniformly

distributed load q, as shown in Fig. 76. The bending moment at a distance x from
A is
(l − x)2
M = −q
2
Then,

d2 y q
EI 2
= −M = (l2 − 2lx + x2 )
dx 2
So,

dy q x3
EI = (l2 x − lx2 + ) + c1
dx 2 3
and
q x2 x3 x4
EIy = (l2 − l + ) + c1 x + c2
2 2 3 12
Boundary conditions are:
dy
at x = 0, = 0 and y = 0
dx
These give c1 = c2 = 0. The equation of deflection is

qx2  2  dy qx  2 
y= 6l − 4lx + x2 , θ= = 3l − 3lx + x2 . (301)
24EI dx 6EI
The free end deflection and slope become:

ql4 ql3
yB = θB = . (302)
8EI 6EI

14.1.4 Simply-supported beam with a central concentrated load

AB is a simply supported beam with a uniform section subjected to a central load
P , as shown in Fig. 77.

0 < x < l/2

The bending moment at a distance x from A is
Px
M=
2
AE 227 118 Dr. Krishnendu Haldar
14.1 Differential equation for deflection

Figure 77: Simply-supported beam with a central concentrated load.

Then,

d2 y Px
EI 2
= −M = −
dx 2
So,
dy P
EI = x 2 + c1
dx 4
and
P 3
EIy = x + c1 x + c2
12
Boundary conditions are:
l dy
at x = 0, y = 0 and at x = = 0 from symmetry y = 0
2 dx
P l2
These give c2 = 0 and c1 = 16
. The equation of deflection is

P x  2 4 2 dy P  2 x2 
y= l − x , θ= = l − . (303)
16EI 3 dx EI 4
The mid point deflection and initial(/end) slope are:

P l3 P l2
yB = θB = . (304)
48EI 16EI

AE 227 119 Dr. Krishnendu Haldar

14.1 Differential equation for deflection

Figure 78: Simply-supported beam with distributed load.

14.1.5 Simply-supported beam with distributed load

AB is the cantilever beam with a uniform section subjected to a uniformly dis-
tributed load q, as shown in Fig. 78. The bending moment at a distance x from
A is
qlx qx2
M= −
2 2
Then,
d2 y qlx qx2
EI 2 = −M = − +
dx 2 2
So,
dy ql q
EI = − x 2 + x 3 + c1
dx 4 6
and
ql 3 q
EIy = − x + x 4 + c1 x + c2
12 24
Boundary conditions are:
l dy
at x = 0, y = 0 and at x = = 0 from symmetry y = 0
2 dx
2
These give c2 = 0 and c1 = P16l . The equation of deflection is
qx  2  dy q 3 
y= l + x3 − 2lx2 , θ= = l + 4x3 − 2lx2 . (305)
24EI dx 24EI
The mid point deflection and initial(/end) slope are:
5ql4 ql3
yB = θB = . (306)
384EI 24EI

AE 227 120 Dr. Krishnendu Haldar

15 Strain energy
The elastic strain energy density U0 , i.e., the strain energy per unit volume for a
3D body is given by

dU0 = σij dεij

and the total strain energy thus given by

Z
U= U0 dV
V

Now recall that

σij = λεkk δij + 2µεij

1+ν ν
εij = σij − σkk δij
E E
If we want the energy expression in terms of stress, then we replace dεij = 1+ν E
dσij −
ν
E
dσ δ
kk ij and obtain:
Z σij
1+ν ν
U0 (σij ) = σij ( dσij − dσkk δij )
0 E E
1 + ν σij ν σkk
Z Z
= σij dσij − σkk dσkk
E 0 E 0
1+ν X 2 ν
= ( σij ) − (σkk )2
2E 2E
1+ν 2 ν
= (σx + σy2 + σz2 + 2τxy 2 2
+ 2τyz 2
+ 2τxz )− (σx + σy + σz )2 .
2E 2E
(307)

If we want the energy expression in terms of stress, then we replace σij = λεkk δij +
2µεij and obtain:
Z εij
U0 (εij ) = (λεkk δij + 2µεij )dεij
0
Z εij
= (λεkk dεkk + 2µεij dεij )
0
λ X
= (εkk )2 + µ( ε2ij )
2
λ
= (εx + εy + εz )2 + µ(ε2x + ε2y + ε2z + 2ε2xy + 2ε2yz + 2ε2zx ). (308)
2

AE 227 121 Dr. Krishnendu Haldar

15.1 Strain energy due to axial load

15.1 Strain energy due to axial load

For this problem
P
σx =
A
where A is the cross sectional area. Then the strain energy density from (307) is
given by
σx2 P2
U0 = =
2E 2A2 E
Then total strain energy for axial loading (from (307)):
Z l Z l
P2 P2
U= 2
Adx = dx (309)
0 2A E 0 2AE

15.2 Strain energy due to Torsion

For this problem
Tr
τxy =
Jp
where Jp is the polar moment of inertia. Then the strain energy density from (307)
is given by
2
τxy T 2 r2
U0 = =
2µ 2µJp2
Then total strain energy for axial loading (from (307)):
Z l Z l Z l 2
T2 T2
Z
2 T
U= 2
( r dA)dx = 2
Jp dx = dx (310)
0 2µJp A 0 2µJp 0 2µJp

15.3 Strain energy due to bending moment

For this problem
My
σx =
I
where I is the area moment of inertia. Then the strain energy density from (307)
is given by
σx2 M 2y2
U0 = =
2E 2EI 2
Then total strain energy for axial loading (from (307)):
Z l Z l Z l
M2 M2 M2
Z
2
U= 2
( y dA)dx = 2
Idx = dx (311)
0 2EI A 0 2EI 0 2EI

AE 227 122 Dr. Krishnendu Haldar

15.4 Strain energy due to shear stress

15.4 Strain energy due to shear stress

For a rectangular cross section of width b and height h
VQ
τxy =
Ib
where I is the area moment of inertia, V is the shear stress and
Z h/2
Q= ydA.
y

Then the strain energy density from (307) could be written as

2
τxy V 2 Q2
U0 = =
2E 2µI 2

Then total strain energy for axial loading (from (307)) take the following form
(after simplification)
l
V2
Z
U =κ dx. (312)
0 2µA
Here the factor κ depends on the cross sectional area of the beam. For a rectangular
cross-section κ = 6/5.

15.5 Superposition of elastic energies

If we have all the three kinds of forces (axial, torsional, bending) acting together,
then the total energy is the sum of the three individual energy. We can write the
total energy:
Z l Z l 2 Z l 2 Z l
P2 V T M2
Utotal = dx + κ dx + dx + dx (313)
0 2AE 0 2µA 0 2µJp 0 2EI

15.5.1 Cantilever with end moment

l
M2 M 2l
Z
U= dx = (314)
0 2EI 2EI

AE 227 123 Dr. Krishnendu Haldar

15.5 Superposition of elastic energies

15.5.2 Cantilever with concentrated load at free end

In this case

M = −P x, V =P

Then,
l Z l
V2 P 2 x2
Z
Utotal = κ dx + dx
0 2µA 0 2EI
2 Z l Z l Z l
P P2
= κ dx + x2 dx
2µA 0 0 2EI 0
2 2 3
P l P l
= κ + (315)
2µA 6EI

15.5.3 Cantilever with distributed load

In this case
1
M = qx2 , V = qx
2
Then,
l Z l 2 4
q 2 x2
Z
q x
Utotal = κ dx + dx
0 2µA 0 8EI
Z l Z l
q2 2 q2
= κ x dx + x4 dx
2µA 0 8EI 0
q 2 l3 q 2 l5
= κ + (316)
6µA 40EI

15.5.4 Torsion with variable cross-section

One end fixed and a torque T is applied at the free end. The diameter of the rod
is varying from d0 to 2d0 over the length L. Then,
d0 x
d = d0 + x = d0 (1 + )
L L

l
T2
Z
U= dx
0 2µJp

AE 227 124 Dr. Krishnendu Haldar

 πd4 πd40 πd40 (L+x)4
Now Jp = 32
= 32
(1 + x/L)4 = 32L4
. Then,
l
T2 T 2 l 32L4
Z Z
1
U = dx = 4
dx
0 2µJp 2µ 0 πd0 (L + x)4
16T 2 l4 h 1 iL 14T 2 L
= − = (317)
πd40 µ 3(L + x)3 0 3πd40 µ

APPENDIX
A Index notation
Consider the following sets of numbers:

a11 , a22 , a12 , a21

or

a11 , a22 , a12 , a21

or

a11 , a22 , a12 , a22

In the index notation, they are represented by: aij , aij and aij respectively, where
each of the indices (i, j) takes values 1 and 2.

A.1 Summation convention

Let n quantities be denoted by x1 , x2 , x3 , ..., xn . Consider the expression:
n
X
ai xi = a1 x1 + a2 x2 + ... + an xn
i

This expression will be written ai xi by introducing the summation convention of

EINSTEIN. When an index appears twice in a term, a summation is implied. The
range of the summation is known from the context. Note carefully that:

1. The convention does not apply to numerical indices. For instant a2 x2 stands
for single term.

2. The repeated index may be replaced by any other index. For instance ai xi =
at xt . For this reason, the repeated index is called a dummy index.

AE 227 125 Dr. Krishnendu Haldar

A.2 Kronecker delta

3. If an index is not dummy, it is called free index. Thus in the expression aik xi ,
k is the free index.
Some examples:
1. If n = 3, ak xk = a1 x1 + a2 x2 + a3 x3 .
∂φ ∂φ ∂φ
2. If φ is a function of x1 , x2 , x3 , ..., xn , then dφ = ∂x1
dx1 + ∂x2
dx2 +...+ ∂xn
dxn =
∂φ
∂xi
dxi
3. Let yi = αit xt and zi = βit yt . We can rewrite yt = αtk xk and now we can
substitute back as zi = βit (αtk xk ) = [βit αtk ]xk = γik xk . Here γik = βit αtk ,
where t is the dummy index and i, k are the free index.
4. Consider the summation (a1j xj )2 + (a2j xj )2 + ... + (anj xj )2 . This expression
is written as (ais xs )(ait xt ) or ais ait xs xt .
5. Akmn Bkm
l
= (A1mn B1n
l
) + (A2mn B2nl
+ ... + (AN l 1 l 1 l
mn BN n )=(Am1 B11 + Am2 B12 +
l
..) + (A2m1 B21 l
+ A2m2 B22 + ...) + ...

A.2 Kronecker delta

This is defied as:
(
1 if i = j
δij = (A-1)
0 if i 6= j
Note that,
δ1j xj = δ11 x1 + δ12 x2 + ... + δ1n xn = x1
δ2j xj = δ21 x1 + δ22 x2 + ... + δ2n xn = x2
..
.
δij xj = δi1 x1 + δi2 x2 + ... + δin xn = xi
(A-2)
Thus if we multiply xk by δik , we simply replace the index k of xk by i. Moreover,
δii = δ11 + δ22 + δ33 = 3 (A-3)

A.3 Permutation symbol


1
 if i, j, k form an even permutation
ijk = −1 if i, j, k form an odd permutation (A-4)

0 if two or more indices are equal


AE 227 126 Dr. Krishnendu Haldar

A.4 Dot product

Figure 79: Even and odd permutations.

A.4 Dot product

Let a = ai êi and b = bj êj

a · b = ai b i (A-5)

If the set {ê1 , ê2 , ê3 } contains a set of orthonormal unit vectors, then

ei · ej = δij (A-6)

For a Cartesian system, you can think as

     
1 0 0
e1 = 0
  e1 = 1
 e3 = 0

0 0 1

A.5 Cross product

 
ê1 ê2 ê3 
 
a × b = a1 a2 a3 
 b1 b2 b3 
= (a2 b3 − a3 b2 )ê1 + (a3 b1 − a1 b3 )ê2 + (a1 b2 − a2 b1 )ê3
= ijk aj bk êi (A-7)

Basically we represents a = aj êj and b = bk êk . So we can also write the compo-
nent form as:

(a × b)i = ijk aj bk (A-8)

AE 227 127 Dr. Krishnendu Haldar

A.6 Dyadic product

If we replace a and b, then

b × a = ikj bk aj êi = −ijk aj bk êi = −a × b
Let c = ar êr , then
c · (a × b) = ijk cr aj bk (êi · êr ) = ijk cr aj bk δir = ijk ci aj bk

A.6 Dyadic product

Let a = ai êi and b = bj êj , then the dyadic product is defined as
a ⊗ b = ai bj êi ⊗ êj
In the matrix notation
 
a1 b1 a1 b2 a1 b3
(a ⊗ b)ij = a2 b1 a2 b2 a2 b3 
a3 b 1 a3 b 2 a3 b 3
This operation constructs a second-order tensor from two vectors.

A.7 Matrices
Components of a second order tensor could be represented by a matrix. Let for
tensors A and B, the components are Aij and Bkl . Then
A = Aij
AT = Aji
tr(A) = Aii = A11 + A22 + A33
(AB)ik = Aij Bjk
(AB T )ik = Aij Bkj
(AT B)ik = Aji Bjk
Note that:
tr(a ⊗ b) = a1 b1 + a2 b2 + a3 b3 = a · b
Symmetric matrix: A = AT or Aij = Aji
Skew symmetric matrix: A = −AT or Aij = −Aji

Any matrix B could be decomposed as a symmetric and skew-symmetric part

as:
1  1 
Sym B = B + B T , ⇒ [Sym B]ij = Bij + Bji
2 2
1  1  
Skw B = B − B T , ⇒ [Skw B]ij = Bij − Bji
2 2
AE 227 128 Dr. Krishnendu Haldar
A.8 Differentiation notation

A.7.1 skw-dual relationship

For a skew symmetric matrix
 
0 −wz wy
[W ] =  wz 0 −wx  ,
−wy wx 0

its dual vector or axial vector is given by,

 
wx
dual [W ] = wy  .

wz

Note that for an arbitrary vector b,

W x = dual [W ] × x, → Wij xj = ijk dual [W ]j xk

A.8 Differentiation notation

Let

a = a(x1 , x2 , x3 )
ai = ai (x1 , x2 , x3 )
aij = aij (x1 , x2 , x3 )

Then
∂a
a, i = = ∂i a
∂xi
∂ai
ai, j = = ∂ j ai
∂xj
∂aij
aij, k = = ∂k aij (A-9)
∂xk

A.9 Coordinate transformations

Let a vector v is expressed with respect to Oxy system as:

v = v1 ê1 + v2 ê2 (A-10)

where, v1 , v2 are the components and ê1 , ê2 are the unit vectors . Let the com-
ponents with respect to the Ox0 y 0 system are v10 , v20 , where the unit vectors are ê01

AE 227 129 Dr. Krishnendu Haldar

A.9 Coordinate transformations

Figure 80: Coordinate transformation.

and ê02 . Since the vector does not change (only the components change), we can
write
v = v1 ê1 + v2 ê2
v = v10 ê01 + v2 ê02
Now,
ê1 = cos θ ê01 − sin θ ê02
ê2 = sin θ ê01 + cos θ ê02
Substituting back we get,
v = v1 (cos θ ê01 − sin θ ê02 ) + v2 (sin θ ê01 + cos θ ê02 )
= (v1 cos θ + v2 sin θ) ê01 + (−v1 sin θ + v2 cos θ) ê02
v = v10 ê01 + v2 ê02
This implies:
v10 = v1 cos θ + v2 sin θ
v20 = −v1 sin θ + v2 cos θ
In the matrix notation:
 0   
v1 cos θ sin θ v1
=
v20 − sin θ cos θ v2
and in the index notation
vi0 = Qij vj (A-11)

AE 227 130 Dr. Krishnendu Haldar

A.9 Coordinate transformations

where, the elements of the transformation matrix Qij = cos (xi , x0j ) denotes the
cosine of the angle between the x0i -axis and xj -axis. Similarly,
vi = Qji vj0 (A-12)
Lets start from
0
vi = Qmi vm
= Qmi (Qmj vj )
= (Qmi Qmj )vj
=⇒ δij vj = (Qmi Qmj )vj
=⇒ (δij − Qmi Qmj )vj = 0
The above equation is true for all vectors vj . Then
Qmi Qmj = δij QT Q = I (A-13)
In a similar way, starting from vi0 = Qij vj , we can show that
Qim Qjm = δij QQT = I (A-14)
Let us now consider a second order tensor in the x − y frame A = Aij and when
it operates on a vector v gives a vector u. That is
ui = Aij vj u = Av (A-15)
In the x0 − y 0 frame the above relation becomes
u0i = A0ij vj0
(Qim um ) = A0ij (Qjk vk )
Qim (Amk vk ) = A0ij (Qjk vk )
(Qim Amk )vk = (A0ij Qjk )vk
(A-16)
This implies
Qim Amk = A0ij Qjk
Qim Amk Qpk = A0ij (Qjk Qpk )
Qim Amk Qpk = A0ij δjp
Qim Amk Qpk = A0ip
Qim Amk Qjk = A0ij
or
A0ij = Qip Qjq Apq . (A-17)
In matrix (tensor) notation:
A0 = QAQT . (A-18)

AE 227 131 Dr. Krishnendu Haldar

B Strain in 2D polar system: geometrical mean-
ing

Figure 81: 2D geometric deformation in polar coordinate system.

Let us consider two infinitesimal perpendicular line segments P Q and P R

inside the material body in the undeformed configuration as shown in Fig. 81.
The coordinate of P is (r, θ) and that of Q and R are (r + ∆r, θ) and (r, θ + ∆θ),
respectively. After deformation, the three points P, Q and R move to P 0 , Q0 and
R0 , respectively. Moreover, the new line segments P 0 Q0 and P 0 R0 make angles θ1
and θ2 with respect to the radial and polar lines, respectively, as shown in the
diagram. Then we have the following definitions of strains:

P 0 Q0 − P Q P 0 R0 − P R
εr = , εθ = , γrθ = π/2 − ∠Q0 P 0 R0 = θ1 + θ2 . (B-1)
PQ PR
εr is the normal strain in the r-direction, εθ is the normal strain in the θ-direction,
and γrθ is the shear strain, defined as the change of right angle due to the defor-
mation. Our goal will be to calculate the new line segments P 0 Q0 = |P 0 Q0 | and
P 0 R0 = |P 0 R0 |. We’ll now compute them through vector approach. In this context,
let us now consider the position vectors of the six points P, Q, R, and P 0 , Q0 , R0 as
presented in Fig. 82. The displacement fields of the three points P, Q, R are:

AE 227 132 Dr. Krishnendu Haldar

 Figure 82: Position and displacement vectors of different points.

uP = u(r, θ) = ur (r, θ)êr + uθ (r, θ)êθ

uQ = u(r + ∆r, θ) = ur (r + ∆r, θ)êr + uθ (r + ∆r, θ)êθ
00 00
uR = u(r, θ + ∆θ) = ur (r, θ + ∆θ)êr + uθ (r, θ + ∆θ)êθ . (B-2)

Now,

P Q = OQ − OP = ∆r êr
P 0 Q0 = OQ0 − OP 0 (B-3)

where,

OP 0 = OP
h i+ PhP
0
i
= rêr + ur (r, θ)êr + uθ (r, θ)êθ
h i h i
= r + ur (r, θ) êr + uθ (r, θ) êθ

AE 227 133 Dr. Krishnendu Haldar

and

OQ0 = OQ + QQ0
h i h i
= (r + ∆r)êr + ur (r + ∆r, θ)êr + uθ (r + ∆r, θ)êθ
h i h i
= r + ∆r + ur (r + ∆r, θ) êr + uθ (r + ∆r, θ) êθ
h ∂ur i h ∂uθ i
≈ r + ∆r + ur (r, θ) + ∆r êr + uθ (r, θ) + ∆r êθ
∂r ∂r
So,

P 0 Q0 = OQ0 − OP 0
h ∂ur i h ∂u
θ
i
= ∆r + ∆r êr + ∆r êθ
r ∂r ∂r
 ∂u   ∂u 2  ∂u 2
r r θ
P 0 Q0 = |P 0 Q0 | = 1+2 + + ∆r
r ∂r ∂r ∂r
 ∂u 
r
≈ 1+2 ∆r
∂r
 ∂ur 
≈ 1+ ∆r
∂r
(B-4)

Then,
 
∂ur
P 0 Q0 − P Q 1+ ∂r
∆r − ∆r ∂ur
εr = = = (B-5)
PQ ∆r ∂r
In a similar procedure,
00
PR = OR − OP = r êr − r êr
00 00 00
= r êr − r (êr cos ∆θ − êθ sin ∆θ)
00 00 00
≈ r êr − r (êr − êθ ∆θ)
00
= (r∆θ) êθ
P 0 R0 = OR0 − OP 0 (B-6)

AE 227 134 Dr. Krishnendu Haldar

 0 0 00 00
Figure 83: Expressing {êr , êθ }, {êr , êθ }, and {êr , êθ } system.

where,

OP 0 = OP
h i+ PhP
0
i
= rêr + ur (r, θ)êr + uθ (r, θ)êθ
h i h i
= r + ur (r, θ) êr + uθ (r, θ) êθ
h i 00 00
h i 00 00
= r + ur (r, θ) (êr cos ∆θ − êθ sin ∆θ) + uθ (r, θ) (êθ cos ∆θ + êr sin ∆θ)
h i 00 00
h i 00 00
≈ r + ur (r, θ) (êr − êθ ∆θ) + uθ (r, θ) (êθ + êr ∆θ)
h i 00 h i 00
= r + ur (r, θ) + uθ (r, θ)∆θ êr + uθ (r, θ) − r∆θ − ur (r, θ)∆θ êθ
h i 00 h i 00
≈ r + ur (r, θ) êr + uθ (r, θ) − r∆θ − ur ∆θ êθ

and

OR0 = OR +hRR0 i
00 00 00
= rêr + ur (r, θ + ∆θ)êr + uθ (r, θ + ∆θ)êθ
h ∂ur i 00 h ∂uθ i 00
≈ r + ur (r, θ) + ∆θ êr + uθ (r, θ) + ∆θ êθ (B-7)
∂θ ∂θ

AE 227 135 Dr. Krishnendu Haldar

So,

P 0 R0 = OR0 − OR0
h ∂u i 00 h ∂u i 00
r θ
= ∆θ êr + ∆θ + r∆θ + ur ∆θ êθ
r∂θ ∂u 2
∂θ
h ∂u
r θ ∂uθ i  ∂uθ 2
P 0 R0 = |P 0 R0 | = +2 r + rur + ur + + u2r + r2 ∆θ
r ∂θ ∂θ ∂θ ∂θ
h ∂u i
θ
≈ r2 + 2 r + rur ∆θ
r ∂θ
h 1 ∂u ur i
θ
= r 1+2 + ∆θ
r r ∂θ r
h 1 ∂u ur i
θ
≈ r 1+2 + ∆θ
r ∂θ r
 1 ∂uθ ur 
≈ r 1+ + ∆θ
r ∂θ r
Then,
 
1 ∂uθ ur
0 0
P R − PR r 1+ r ∂θ
+ r
∆θ − r∆θ 1 ∂uθ ur
εθ = = = + (B-8)
PR r∆θ r ∂θ r
For the angle θ1 , we consider
0
êr × P 0 Q0 = |êr ||P 0 Q0 | sin θ1 k̂
0
Now, êr = êr cos α + êθ sin α ≈ êr + êθ α. So,
 
 êr êr k̂
α i 0  = 1 · 1 + ∂ur ∆r · θ1 k̂
  
 1
∂r
h i h
∂uθ
 ∆r + ∂u r
∆r ∆r 0 
 
∂r ∂r
h ∂u
θ
i ∂ur −1
⇒ θ1 = −α 1+
∂r ∂r
h ∂u i ∂ur  ∂uθ
θ
≈ −α 1− ≈ −α
∂r ∂r ∂r
∂uθ uθ
≈ −
∂r r
For the angle θ2 , we consider
0
êr × P 0 R0 = |êr ||P 0 R0 | sin θ2 (−k̂)

AE 227 136 Dr. Krishnendu Haldar

 0 00 00
Similarly, for the angle θ2 , we consider êθ = êr sin(∆θ − α) + êθ cos(∆θ − α) ≈
00 00
êr (∆θ − α) + êθ . So,

 ê00 00
 
r ê r k̂ 

(∆θ − α) 1 0
  1 ∂uθ ur 
= −1 · r 1 + + ∆θ · θ2 k̂

r ∂θ r
h 
 ∂ur i h ∂uθ i 
 ∂θ ∆θ ∂θ
∆θ + r∆θ + ur ∆θ 0 
1 h ∂ur i 1 ∂uθ ur −1
⇒ θ2 = 1+ +
r ∂θ r ∂θ r
1 h ∂ur i ∂ur  1 ∂ur
≈ 1− ≈ (B-9)
r ∂θ ∂r r ∂θ
So, the shear strain (engineering)

∂uθ uθ 1 ∂ur
γrθ = π/2 − ∠Q0 P 0 R0 = θ1 + θ2 = − + . (B-10)
∂r r r ∂θ

AE 227 137 Dr. Krishnendu Haldar