Strength of Material I &II Lecture Notes

Bonga University
College of Engineering and Technology
Department of Mechanical Engineering

Strength of Material- II Lecture notes

Prepared by:

Balemlay Alehegn (MSc. in Mechanical Design)

January: 2024
Bonga, Ethiopia

BU, Department of Mechanical Engineering

 Strength of Material I &II Lecture Notes

Strength of Material-II
CHAPTER - ONE
Deflection of Beam; Work and Energy Methods

Introduction

In the previous strength of Material-I course,

The relations existing between forces and deformations under various loading conditions
were considered.
The analysis was based on two fundamental concepts, the concept of stress and the concept
of strain where discussed. Now the third important concept to be introduced in this chapter
is the concept of strain energy.
Strain energy of a member is defined as the increase in energy associated with the
deformation of the member. From work energy view that the strain energy is equal to the
work done by a slowly increasing load applied to the member will be discussed.
The strain-energy density of a material is the strain energy per unit volume and it is
equal to the area under the stress-strain curve of the material.
From the stress-strain diagram of a material we shall also define the modulus of toughness
and modulus of resilience of the material.
The elastic strain energy associated with normal stresses will be discussed, First in
members under axial loading and then in members in bending. Later we shall consider the
elastic strain energy associated with shearing stresses such as in torsional loadings of
shafts and in transverse loadings of beams. Strain energy for a general state of stress will
be considered where we shall derive the maximum-distortion energy criterion for yielding.

1
Strength of Material I &II Lecture Notes

1.1. Introduction to Strain Energy

Energy is normally defined as the capacity to do work and it may exist in any of many
forms, e.g. mechanical (potential or kinetic), thermal, nuclear, chemical, etc. The potential
energy of a body is the form of energy which is stored by virtue of the work which has
previously been done on that body, e.g. in lifting it to some height above a datum. Strain
energy is a particular form of potential energy which is stored within materials which have
been subjected to strain, i.e. to some change in dimension. The material is then capable of
doing work, equivalent to the amount of strain energy stored, when it returns to its original
unstrained dimension.
Strain energy is therefore defined as the energy which is stored within a material when
work has been done on the material. Here it is assumed that the material remains elastic
whilst work is done on it so that all the energy is recoverable and no permanent deformation
occurs due to yielding of the material, i.e.
Strain energy U = work done
Thus for a gradually applied load the work done in straining the material will be given by
the shaded area under the load-extension graph of Fig. 2.1.
𝟏
𝐔 = 𝐏✿
𝟐

Fig. 2.1 Work done by a gradually applied load

2
Consider a rod BC of length L and uniform cross-sectional area A, which is attached at B
to a fixed support, and subjected at C to a slowly increasing axial load P (Fig. 1.1). By
plotting the magnitude P of the load against the deformation x of the rod, we obtain a
certain load-deformation diagram (Fig. 1.2) which is characteristic of the rod BC.

Fig.1.1 axially loaded rod Fig.1.2 load deformation diagram Fig.1.3 work due to load p
Let us now consider the work dU done by the load P as the rod elongates by a small amount
dx. This elementary work is equal to the product of the magnitude P of the load and of
the small elongation dx. We write
dU = P dx ........................................................... (1.1)
and note that the expression obtained is equal to the element of area of width dx located
under the load-deformation diagram (Fig. 1.3).
The total work U done by the load as the rod undergoes a deformation x1 is thus

And this is equal to the area under the load-deformation diagram between x = 0 and x = x1.
The work done by the load P as it is slowly applied to the rod must result in the increase of
some energy associated with the deformation of the rod. This energy is referred to as the
strain energy of the rod. We have, by definition,

Strain energy = ……………..… (1.2)

SI metric units are same as unit of work and energy and expressed in N.m
; this unit is called a joule (J).

3
In the case of a linear and elastic deformation, the portion of the load-deformation diagram
involved may be represented by a straight line of equation P = kx (Fig. 1.4). Substituting
for P in Eq. (1.2), we have

or

U= P1 x1 ............................................................................................................ (1.3)

Where P1 is the value of the load corresponding to the deformation x1.

The concept of strain energy is particularly useful in the determination of the effects of
impact loadings on structures or machine components. Consider, for example, a body of
mass m moving with a velocity V0 which strikes the end B of a rod AB (Fig. 1.5a).
• Neglecting the inertia of the elements of the rod, and
• Assuming no dissipation of energy during the impact
We find that the maximum strain energy Um acquired by the rod (Fig. 1.5b) is equal to the
original kinetic energy of the moving body.

T = mV 02

We may then determine the value Pm of the static load which would have produced the
same strain energy in the rod, and obtain the value 𝜎m of the largest stress occurring in the
rod by dividing Pm by the cross-sectional area of the rod.

Fig. 1.4 A linear and elastic deformation Fig. 1.5 A rod subjected to impact load

4
 1.2. Strain- Energy Density

We know that the load-deformation diagram for a rod BC depends upon the length L and
the cross-sectional area A of the rod.
The strain energy U defined by Eq. (1.2), therefore, it also depends upon the dimensions of
the rod. In order to eliminate the effect of size from our discussion and direct our attention
to the properties of the material, we shall consider the strain energy per unit volume.
Dividing the strain energy U by the volume V = AL of the rod (Fig. 1.1), and using Eq.
(1.2), we have

Recalling that P A represents the normal stress 𝜎x in the rod, and x L the normal stain 𝞮, we
write
𝗌
𝑼 =∫ 𝟏 𝝈 𝒅𝗌
𝑽 𝟎 𝒙 𝒙

Where 𝗌𝟏 denotes the value of the strain corresponding to the elongation x1. The strain
energy per unit volume, U V, is referred to as the strain-energy density and will be denoted
by the letter u. We have, therefore,

………….… (1.4)

The strain-energy density is expressed in J m 3 or its multiples kJ m 3 and MJ m 3.

Referring to (Fig. 1.7), we note that the strain-energy density u is equal to the area under
the stress-strain curve, measured from 𝗌𝒙= 0 to 𝗌𝒙 = 1. If the material is unloaded, the stress
returns to zero, but there is a permanent deformation represented by the strain 𝗌𝒑 and only
the portion of the strain energy per unit volume corresponding to the triangular area may
be recovered. The remainder of the energy spent in deforming the material is dissipated in
the form of heat
 Fig.1.6 strain energy Fig. 1.7 modulus of toughness

Fig. 2.8

Fig.1.8 modulus of resilience

The value of the strain-energy density obtained by setting 𝗌𝟏 = 𝗌𝒓 in Eq. (1.4), where 𝗌𝒓 is
the strain at rupture, is known as the modulus of toughness of the material. It is equal to
the area under the entire stress-strain diagram (Fig. 1.6) and represents the energy
per unit volume required to cause the material to rupture. It is clear that the toughness
of a material is related to its ductility as well as to its ultimate strength, and that the capacity
of a structure to withstand an impact load depends upon the toughness of the material used.

If the stress 𝜎x remains within the proportional limit of the material, Hooke ‘s law applies
and we may write
𝜎x = E𝗌𝒙 ............................................................. (1.5)
Substituting for 𝜎x from (1.5) into (1.4), we have

……………………. (1.6)
or, using Eq. (5) to express 𝗌𝒙in terms of the corresponding stress 𝜎x,

…………………………………….… (1.7)
The value uy of the strain-energy density obtained by setting 𝜎1 = 𝜎y in Eq. (1.7), where 𝜎y
is the yield strength, is called the modulus of resilience of the material. We have
…………………………….. (1.8)

The modulus of resilience is equal to the area under the straight-line portion OY of
the stress-strain diagram (Fig. 1.8) and represents the energy per unit volume that the
material may absorb without yielding. The capacity of a structure to withstand an impact
load without being permanently deformed clearly depends upon the resilience of the
material used.
Since the modulus of toughness and the modulus of resilience represent characteristic
values of the strain-energy density of the material considered, they are both expressed in
J/m3 or its multiples.
1.3. Elastic Strain Energy for Normal Stresses
Since the rod considered in the preceding section was subjected to uniformly distributed
stresses 𝜎x, the strain-energy density was constant throughout the rod and could be defined
as the ratio U V of the strain energy U and the volume V of the rod. In a structural element
or machine part with a non-uniform stress distribution, the strain energy density u may
be defined by considering the strain energy of a small element of material of volume 𝑑V
and writing

𝑑𝑈
Or u= ……….………………………..(1.9)
𝑑𝑉

The expression obtained for it in Sec. above in terms of 𝜎x and εx remains valid, i.e., we
still have
……………….… (1.10)
but the stress 𝜎x, the strain εx, and the strain-energy density u will generally vary from point
to point. For values of 𝜎x within the proportional limit, we may set 𝜎x = Eεx in Eq. (1.10)
and write

7
 ………………………… (1.11)

The value of the strain energy U of a body subjected to uniaxial normal stresses may be
obtained by substituting for it from Eq. (1.11) into Eq. (1.9) and integrating both members.
We have

…………………………..… (1.12)

The expression obtained is valid only for elastic deformations and is referred to as the
elastic strain energy of the body.
1.4. Strain Energy under Different Loading Conditions
a) Strain Energy under Axial Loading
We know that when a rod is subjected to a centric axial loading, the normal stresses 𝜎x
may be assumed uniformly distributed in any given transverse section. Denoting by A
the area of the section located at a distance x from the end B of the rod (Fig. 1.9), and by P
the internal force in that section, we write 𝜎x = P/A. Substituting for 𝜎x into Eq. (1.12), we
have

or, setting dV = A dx

…………………1.13 Fig.1.9 A rod with centric axial load

In the case of a rod of uniform cross section subjected at its ends to equal and opposite
forces of magnitude P (Fig. 1.10), Eq. (1.13) yields

……………………..1.14

Fig.1.10 A rod of uniform cross section under axial load

8
Illustrative Examples -1
A rod consists of two portions BC and CD of the same material and same length, but of
different cross sections (Figure below). Determine the strain energy of the rod when it is
subjected to a centric axial load P, expressing the result in terms of P, L, E, the cross-
sectional area A of portion CD, and the ratio n of the two diameters.

Solution
We use Eq. (1.14) to compute the strain energy of each of the two portions, and add the
expressions obtained:

or
………………………. (1.15)

We check that, for n = 1, we have

Which is the expression given in Eq. (1.14) for a rod of length L and uniform cross section
of area A. We also note that, for n > 1, we have Un < U for example, when n = 2, we have
5
U2 =( )U 1. Since the maximum stress occurs in portion CD of the rod and is equal to 𝜎max.
8

= P/A, it follows that, for a given allowable stress, increasing the diameter of portion BC
of the rod results in a decrease of the overall energy-absorbing capacity of the rod.
Unnecessary changes in cross-sectional area should therefore be avoided in the design
of members which may be subjected to loadings, such as impact loadings, where the
energy-absorbing capacity of the member is critical.
b) Strain Energy in Bending
Consider the element now be subjected to a constant bending moment M causing it to bend
into an arc of radius R and subtending an angle 𝒅𝜽 at the centre (Fig.). The beam willalso
have moved through an angle 𝒅𝜽.

F.g.
Strain energy = work done = ½ x moment x angle turned through (in radians)
1
u = Mdθ
2
But,
M E
ds = Rdθ and =
I R
Therefore,

ds M
dθ = = ds
R EI
Therefore,

1 M M2ds
u= M× ds =
2 EI 2EI
Total strain energy resulting from bending,

1.4. Elastic Strain Energy for Shearing Stresses

Consider the elemental bar now subjected to a shear load Q at one end causing deformation
through the angle 𝜸 (the shear strain) and a shear deflection 𝜹, as shown in Fig. 2.3.

Fig. 2.3 Shear

1 1
Strain energy U = Work done = Qδ = Qγds
2 2

Shear stress τ Q
G= = =
Shear strain γ γA
Q
γ=
AG
 1 Q Q2 ds
Q× × ds =
Shear strain energy = 2 2AG
AG

Hence, total strain energy resulting from shear is,

c) Strain Energy in Torsion

Consider a shaft BC of length L subjected to one or several twisting couples.

Fig 1.10 a) left b) right; Shaft subject to

torque.
Denoting by J the polar moment of inertia of the cross section located at a distance x from
B (Fig. 1.10a), and by T the internal torque in that section, we recall that the shearing
stresses in the section are 𝑟xy = Tr/J. Substituting for 𝑟xy into Eq. (1.18), we have
Setting dV = dA dx, where dA represents an element of the cross sectional area, and
observing that T 2/2GJ2 is a function of x alone, we write
Recalling that the integral within the parentheses represents the polar moment of inertia J
of the cross section, we have
………………………………………1.19

In the case of a shaft of uniform cross section subjected at its ends to equal and opposite
couples of magnitude T (Fig. 1.10), Eq. (1.19) yields

…………………………………………1.20

Illustrative Example-3
A circular shaft consists of two portions BC and CD of the same material and same
length, but of different cross sections (Fig. below). Determine the strain energy of the
shaft when it is subjected to a twisting couple T at end D, expressing the result in terms
of T, L, G, the polar moment of inertia J of the smaller cross section, and the ratio a of the
two diameters.

Solution
We use Eq. (1.20) to compute the strain energy of each of the two portions of shaft, and
add the expressions obtained. Noting that the polar moment of inertia of portion BC is
equal to n4J, we write

14
or
We check that, for n = 1, we have

which is the expression given in Eq. (1.20) for a shaft of length L and uniform cross section.
We also note that, for n > 1, we have Un < U1; for example, when n = 2, we have
Since the maximum shearing stress occurs in the portion CD of the shaft and is proportional
to the torque T, we note as we did earlier in the case of the axial loading of a rod that, for
a given allowable stress, increasing the diameter of portion BC of the shaft results in a
decrease of the overall energy-absorbing capacity of the shaft.
d) Strain Energy under Transverse Loading
In Sec. 1.4 we obtained an expression for the strain energy of a beam subjected to a
transverse loading. However, in deriving that expression we took into account only the
effect of the normal stresses due to bending and neglected the effect of the shearing stresses.
We shall now take into account the effect of both types of stresses.

Illustrative Example-4
Determine the strain energy of the rectangular cantilever beam AB (Fig.below), taking into
account the effect of both normal and shear stresses.

Solution
The strain energy due to the normal stresses 𝜎x is

15
16
 1.5. Strain Energy For a General State of Stress

In the preceding sections, we determined the strain energy of a body in a state of uniaxial
stress and in a state of plane shearing stress. In the case of a body in a general state of stress
characterized by the six stress components  x ,  y ,  z ,  xy , yz , and  zx , the strain-energy
density may be obtained by adding the expressions given in Eqs. (1.10) and (1.16), as well
as the four other expressions obtained through a permutation of the subscripts.
In the case of the elastic deformation of an isotropic body, each of the six stress-strain
relations involved is linear, and the strain energy density can be expressed as

Substituting for the strain components into (1.21), we have, for the most general state of
stress at a given point of an elastic isotropic body,

If the principal axes at the given point are used as coordinate axes, the shearing stresses
become zero and Eq. (1.22) reduces and the strain energy density of a three-dimensional
principal stress system is given by:

17
where σa, σb, and σc are the principal stresses at the given point.
This total strain energy can be conveniently considered as made up of two parts:
(a) The volumetric or dilatational strain energy density;
(b) The shear or distortional strain energy density.
Therefore, we write

(a) The volumetric or dilatational strain energy density

This is the strain energy density associated with a mean or hydrostatic stress of
1
( +  b +  c ) =  Acting equally in all three mutually perpendicular directions giving
3 a

rise to no distortion, merely a change in volume. And we can set also

Fig. 1.11 Element subjected to multi axial stress

Thus, the given state of stress (Fig. 1.11a) can be obtained by superposing the states of
stress shown in Fig. 1.11b and c. We note that the state of stress described in Fig. 1.11b
tends to change the volume of the element of material, but not its shape, since all the faces
of the element are subjected to the same stress (mean stress).
In this section you will examine the effect of the normal stresses 𝜎x, 𝜎y, and 𝜎z on the
volume of an element of isotropic material. Consider the element shown in Fig. 1.11. In its

18
unstressed state, it is in the shape of a cube of unit volume; and under the stresses 𝜎x, 𝜎y,
𝜎z, it deforms into a rectangular parallelepiped of volume

Since the strains 𝜖x, 𝜖 y, 𝜖 z are much smaller than unity, their products will be even
smaller and may be omitted in the expansion of the product. We have, therefore,

Denoting by e the change in volume of our element, we write

or
And from the generalized Hooke’s law for the multi axial loading of a homogeneous
isotropic material.

As we indicated earlier, the results obtained are valid only as long as the stresses do not
exceed the proportional limit, and as long as the deformations involved remain small
Since the element had originally a unit volume, the quantity e represents the change in
volume per unit volume; it is referred to as the dilatation of the material. Substituting for
𝜖x, 𝜖y, and 𝜖z from Eqs. (1.25) into (1.24), we write

19
Indeed, recalling Eq. (1.25), we note that the dilatation e (i.e., the change in volume per
unit volume) caused by this state of stress is

It follows that the portion uv of the strain-energy density corresponding to a change in

volume of the element can be obtained by substituting 𝜎 for each of the principal stresses
in Eq. (1.23).
We have

stress=normal stress take Mean

Or, recalling the mean stress equation (1 − 2 )

Volumetric strain energy density (uv)= ( +  b +  c )2  .............. 1.27

a
6E

And take the mean stress as the average of the three individual stress
(b) The shear or distortional strain energy density
In order to consider the general principal stress case it is necessary, to add to the mean

stress  in the three perpendicular directions, certain so-called deviatoric stress values to
return the stress system to values of a , b and c .
The portion of the strain-energy density corresponding to the distortion of the element is
obtained by solving U = Uv + Ud for Ud and substituting for U and Uv from Eqs. (1.23)
and (1.27), respectively. We write

20
Noting that each of the parentheses inside the bracket is a perfect square, and recalling from
Equations that the coefficient in front of the bracket is equal to 1/12G, we obtain the
following expression for the portion Ud of the strain-energy density, i.e., for the distortion
energy per unit volume, These deviatoric stresses are then associated directly with change
of shape, i.e. distortion, without change in volume and the strain energy associated with
this mechanism is given by

…………………1.28
In the case of plane stress, and assuming that the c axis is perpendicular to the plane of
stress, we have σc = 0 and Eq. (27) reduces to
…………………………..1.29

1.6. Deflection Under a Single Load by The Work-Energy Method

We saw in the preceding section that, if the deflection x1 of a structure or member under a
single concentrated load P1 is known, the corresponding strain energy U may be obtained
by writing

A similar expression may be used to obtain the strain energy of a structural member under
a single couple M1

Conversely, if the strain energy U of a structure or member subjected to a single

concentrated load P1 or couple M1 is known, the above equation may be used to determine
the corresponding deflection x1 or angle θ1.
In order to determine the deflection under a single load applied to a structure consisting of
several component parts:
First compute the strain energy of the structure by integrating the strain-energy density over
its various parts, and
Then use either of the above equation to obtain the desired deflection.
Similarly, the angle of twist of a composite shaft may be obtained by integrating the strain-
energy density over the various parts of the shaft and solving the below.

21
and, solving for the angle of twist 

It should be kept in mind that the method presented in this section may be used only if the
given structure is subjected to a single concentrated load or couple

22
 Chapter Two

Unsymmetrical bending

Introduction

The behavior of a structural element subjected to an external load applied perpendicularly to a

longitudinal axis of the element.

▪ A bending moment: is the reaction induced in a structural element when an external

force or moment is applied to the element causing the element to bend. Bending moment
acts along the axes of the member.

Zero stress exist at the centroid and the line of centroid is the neutral axes.

▪ The line of intersection of the neutral plane and any cross section of the beam is termed
the neutral axes. NA Is the axes at which strain (and consequently stress) is zero when
the beam is subjected to bending. NA is perpendicular to the plane of the load.
▪ Centroid: the point at which the entire area acts on it or the point of any geometrical
figure at which the area or volume of the section concentrates.
▪ Center of gravity: is the point at which the mass or weight of the body acts in it or the
point of any object whose mass and weight of the body concentrates.
▪ Centroidal axes: is any axes that passes through the centroid of the cross section. There
can be an infinite number of centroidal axes.

To find the centroid:

ẍ = ∫ 𝑥𝑑𝐴 = ∑ 𝐴𝑋
𝐴 𝐴

ӱ = ∫ 𝑦𝑑𝐴 = ∑ 𝐴𝑌
𝐴 𝐴

In Analysis of pure bending has been limited so far to members possessing:

➢ at least one plane of symmetry and

➢ Subjected to couples acting in that plane. Because of the symmetry of such members
and of their loadings, we concluded that the members would remain symmetric with
respect to the plane of the couples and thus bend in that plane.

23
In the figure below part a shows the cross section of a member possessing two planes of
symmetry, one vertical and one horizontal, and part b the cross section of a member with a
single, vertical plane of symmetry. In both cases the couple exerted on the section acts in the
vertical plane of symmetry of the member and is represented by the horizontal couple vector M,
and in both cases the neutral axis of the cross section is found to coincide with the axis of the
couple.

Fig. 2.1 Moment in plane of symmetry

▪ Symmetrical bending: This condition implies that the plane of loading or plane of
bending, is coincident with or parallel to a plane containing a principal centroidal axis of
inertia of the cross section of the beam.

Let us now consider situations where the bending couples do not act in a plane of symmetry
of the member, either because (Reasons for unsymmetrical bending)

1. They act in a different plane, the section is symmetric (rectangular, circular, I-section)
but the load line is inclined to both the principal axes.
2. The member does not possess any plane of symmetry, the section itself is unsymmetric
(angle section) and the load line is along any centroidal axes.

In such situations, we cannot assume that the member will bend in the plane of the couples.

▪ Unsymmetrical bending: the plane of loading or that of bending doesn’t lie in (or
parallel to) a plane that contains the principal centroidal axis of the cross section

24
Assumptions made to unsymmetrical bending:

1. The plane section of the beam remains plane after bending

2. The material of the beam is homogenous and linearly elastic
3. There is no net internal axial force

In each part of the figure below, the couple exerted on the section has again been assumed to act:

▪ In a vertical plane and

▪ Has been represented by a horizontal couple vector M. However, since the vertical plane
is not a plane of symmetry, we cannot expect the member to bend in that plane, or
the neutral axis of the section to coincide with the axis of the couple.

Fig. 2.2 Moment not in plane of symmetry

We propose to determine the precise conditions under which the neutral axis of a cross section
of arbitrary shape coincides with the axis of the couple M representing the forces acting on
that section. Such a section is shown in Fig. below, and both the couple vector M and the
neutral axis have been assumed to be directed along the z axis.

Fig. 3.3 Section with arbitrary shape

25
We recall that, if we then express that the elementary internal forces 𝜎x dA form a system
equivalent to the couple M, we obtain

Assuming the stresses to remain within the proportional limit of the material, we can substitute
𝜎x = -𝜎m (y/c) into above equation and write

▪ The integral ∫ 𝑦zdA represents the product of inertia Iyz of the cross section with
respect to the y and z axes, and will be zero if these axes are the principal centroidal
axes of the cross section. Which reduces to Ixy = Ahk for a rectangle of area A and
centroid distance h and k from the X and Y axes.
▪ We thus conclude that the neutral axis of the cross section will coincide with the axis
of the couple M representing the forces acting on that section if, and only if, the couple
vector M is directed along one of the principal centroidal axes of the cross section.

We note that:

1. The cross sections shown in Fig. 2.1 are symmetric with respect to at least one of the
coordinate axes. It follows that, in each case, the y and z axes are the principal
centroidal axes of the section. Since the couple vector M is directed along one of the
principal centroidal axes, we verify that the neutral axis will coincide with the axis of
the couple.
2. If the cross sections are rotated through 900 Fig. 2.4, the couple vector M will still be
directed along a principal centroidal axis, and the neutral axis will again coincide
with the axis of the couple, even though in case b the couple does not act in a plane of
symmetry of the member.

26
 Fig. 2.4 Moment on principal centroidal axis.

3. In Fig.2.2, on the other hand, neither of the coordinate axes is an axis of symmetry for
the sections shown, nor are the coordinate axes are not principal axes. Thus, the couple
vector M is not directed along a principal centroidal axis, and the neutral axis does
not coincide with the axis of the couple.

However, any given section possesses principal centroidal axes, even if it is unsymmetric, as the
section shown in Fig.2.2c, and these axes may be determined analytically or by using Mohr’s
circle. If the couple vector M is directed along one of the principal centroidal axes of the section,
the neutral axis will coincide with the axis of the couple (Fig.3.5) and the equations derived for
symmetric members can be used to determine the stresses in this case as well.

Fig. 2.5 Moment not on principal centroidal axis.

27
The Stress distribution can be found by:

a) The principle of superposition can be used to determine stresses in the most general
case of unsymmetric bending. Consider first a member with a vertical plane of symmetry,
which is subjected to bending couples M and M’ acting in a plane forming an angle 𝜃
with the vertical plane Fig. 2.6.

a) b) c) d)
Fig. 2.6 Unsymmetric Bending
b) The couple vector M representing the forces acting on a given cross section will form the
same angle 𝜃 with the horizontal z axis (Fig. 6b). Resolving the vector M into component
vectors Mz and My along the z and y axes, respectively, we write

c) Since the y and z axes are the principal centroidal axes of the cross section, the couple
Mz acts in a vertical plane and bends the member in that plane (Fig. 6.c).

The resulting stresses are

………………………. 2.1

Where Iz- is the moment of inertia of the section about the principal centroidal z axis.

The negative sign is due to the fact that we have compression above the xz plane (y > 0)
and tension below (y < 0).

28
 d) On the other hand, the couple My acts in a horizontal plane and bends the member in that
plane (Fig. 2.6 d). The resulting stresses are found to be

………………………………… 2.2

Where Iy- is the moment of inertia of the section about the principal centroidal y axis.

The positive sign is due to the fact that we have tension to the left of the vertical xy plane
(z > 0) and compression to its right (z < 0). The distribution of the stresses caused by the
original couple M is obtained by superposing the stress distributions defined above in
Eqs. 2.1 and 2.2. We have

………………………………… 2.3

The expression obtained can also be used to compute the stresses in an unsymmetric section,
such as the one shown in Fig. 2.7, once the principal centroidal y and z axes have been
determined. On the other hand, the above Eq. is valid only if the conditions of applicability of the
principle of superposition are met. In other words, it should not be used

▪ If the combined stresses exceed the proportional limit of the material, or

▪ If the deformations caused by one of the component couples appreciably affect the
distribution of the stresses due to the other.

Fig. 2.7 asymmetric cross section

29
The Position of neutral axes

The above 2.1 and 2.2 Equations shows that the distribution of stresses caused by unsymmetric
bending is linear. However, as we have indicated earlier, the neutral axis of the cross section will
not, in general, coincide with the axis of the bending couple. Since the normal stress is zero at
any point of the neutral axis, the equation defining that axis can be obtained by setting 𝜎x = 0.

We write

or, solving for y and substituting for Mz and My.

The equation obtained is that of a straight line of slope m = (Iz/Iy) tan𝜃. Thus, the angle ∅ that
the neutral axis forms with the z axis (Fig. 9) is defined by the relation

……………………………… 2.4
where 𝜃 -is the angle that the couple vector M forms with the same axis.

▪ Since Iz and Iy are both positive, ∅ and 𝜃 have the same sign.
▪ Furthermore, we note that ∅ > 𝜃 when Iz > Iy, and ∅ < 𝜃 when Iz < Iy. Thus, the
neutral axis is always located between the couple vector M and the principal axis
corresponding to the minimum moment of inertia.

Alternative 1:

For skew loading and other forms of bending about principal axes

30
where Mu and Mv are the components of the applied moment about the U and V axes.

Alternative 2:

Then the inclination of the N.A. to the X axis is given by:

Alternative 2:

Where M‘is the component of the applied moment about the N.A., IN.A. is determined either from
the momental ellipse or from the Mohr or Land constructions, and n is the perpendicular distance
from the point in question to the N.A.

Deflections of unsymmetrical members are found by applying standard deflection formulae to

bending about either the principal axes or the N.A. taking care to use the correct component of
load and the correct second moment of area value.

General Case of Eccentric Axial Loading

In the above sections we analyzed:

▪ The stresses produced in a member by an eccentric axial load applied in a plane of

symmetry of the member.

31
We will now study the more general case when the axial load is not applied in a plane of
symmetry. Let:

▪ We consider a straight member AB subjected to equal and opposite eccentric axial forces
P and P’ (Fig. 2.9 a).
▪ a and b denote the distances from the line of action of the forces to the principal
centroidal axes of the cross section of the member.

For Force P: The eccentric force P is statically equivalent to the system consisting of a centric
force P and of the two couples My and Mz of moments My = Pa and Mz = Pb represented in Fig.
10b.

For force P’: the eccentric force P’ is equivalent to the centric force P’ and the couples M’y and
M’z.

Fig. 2.9 Eccentric axial loading

In order to determine the distribution of stresses in a section S of the member, by virtue of Saint-
Venant’s principle, we can replace the original loading of Fig. 2.9 a by the statically equivalent
loading of Fig. 2.9 b, as long as that section is not too close to either end of the member.
Furthermore, the stresses due to the loading of Fig. 10b can be obtained by superposing the
stresses corresponding to the centric axial load P and to the bending couples

My and Mz, as long as the conditions of applicability of the principle of superposition are
satisfied. The stresses due to the centric load P are given as load per area, and the stresses due to

32
the bending couples as above, since the corresponding couple vectors are directed along the
principal centroidal axes of the section. We write, therefore,

……………………………… 2.5

Where y and z are measured from the principal centroidal axes of the section. The relation
obtained shows that the distribution of stresses across the section is linear.

▪ In computing the combined stress 𝜎x, care should be taken to correctly determine the
sign of each of the three terms in the right-hand member, since each of these terms can
be positive or negative, depending upon the sense of the loads P and P’ and the location
of their line of action with respect to the principal centroidal axes of the cross section.

Examples

1. A 1600-lb in. couple is applied to a wooden beam, of rectangular cross section 1.5 by 3.5
in., in a plane forming an angle of 30∘ with the vertical (Figure below). Determine:

(a) The maximum stress in the beam,

(b) The angle that the neutral surface forms with the horizontal plane.

(a) Maximum Stress. The components Mz and My of the couple vector are first determined

33
We also compute the moments of inertia of the cross section with respect to the z and y axes:

The largest tensile stress due to Mz occurs along AB and is

The largest tensile stress due to My occurs along AD and is

The largest tensile stress due to the combined loading, therefore, occurs
at A and is

The largest compressive stress has the same magnitude and occurs at E.
(b) Angle of Neutral Surface with Horizontal Plane. The angle f that the neutral surface forms
with the horizontal plane is shown

The distribution of the stresses across the section and angle of neutral axis.

34
 CHAPTER 3

TORSION OF NON-CIRCULAR AND THIN-WALLED

SECTIONS

Rectangular sections

For rectangular shafts, however, with longer side d and shorter side b, it can be shown
by experiment that the maximum shearing stress occurs at the centre of the longer side
and is given by

…(1)

The angle of twist per unit length is given by

…(2)

k1 and k2 being two constants, their values depending on the ratio d/b and being given
in Table 1.

Table 1: Table of k1 and k2 values for rectangular sections in torsion

Narrow rectangular sections

From Table 1 it is evident that as the ratio d/b increases, i.e. the rectangular section
becomes longer and thinner, the values of constants k1 and k2 approach 0.333. Thus,
for narrow rectangular sections in which d/b > 10 both k1 and k2 are assumed to be 1/3
and eqns. (1) and (2) reduce to

…(4)

35
 Thin-walled open sections

There are many cases, particularly in civil engineering applications, where rolled steel or
extruded alloy sections are used where some element of torsion is involved. In most
cases the sections consist of a combination of rectangles, and the relationships given in
eqns. (1) and (2) can be adapted with reasonable accuracy provided that:

(a) the sections are “open”, i.e. angles, channels, T-sections, etc., as shown in Fig. 1;
(b) the sections are thin compared with the other dimensions.

For such sections eqns. (1) and (2) may be re-written in the form

…(5)

…(6)

and for d/b ratios in excess of 10, k1 = k2 = 1/3 , so that

Thin-walled split tube

The thin-walled split tube shown in Fig. 2 is considered to be a special case of the thin-
walled open type of section considered in previous section. It is therefore treated as an
equivalent rectangle with a longer side d equal to the circumference (less the gap), and
a width b equal to the thickness.
where k1 and k2 for thin-walled tubes are usually equal to 1/3.
It should be noted here that the presence of even a very small cut or gap in a thin- walled
tube produces a torsional stiffness (torque per unit angle of twist) very much smaller than
that for a complete tube of the same dimensions.

longitudinal split.

Thin-walled closed tubes of non-circular section

Consider the thin-walled closed tube shown in Fig. 3 subjected to a torque T about the Z
axis; i.e. in a transverse plane. Both the cross-section and the wall thickness around the
periphery may be irregular as shown, but for the purposes of this simplified treatment it
must be assumed that the thickness does not vary along the length of the tube. Then, if
is the shear stress at B and ' is the shear stress at C (where the thickness has

increased to t’) then, from the equilibrium of the complementary shears on the sides AB
and CD of the element shown, it follows that

i.e. the product of the shear stress and the thickness is constant at all points on the
periphery of the tube. This constant is termed the shear flow and denoted by the symbol
q (shear force per unit length).

The quantity q is termed the shear flow because if one imagines the inner and outer
boundaries of the tube section to be those of a channel carrying a flow of water, then,
provided that the total quantity of water in the system remains constant, the quantity
flowing past any given point is also constant.

At any point, then, the shear force Q on an element of length ds is Q = ds = q ds and

the shear stress is q/t.

Consider now, therefore, the element BC subjected to the shear force Q = qds = tds.
The moment of this force about o

where p is the perpendicular distance from o to the force Q.

Therefore the moment, or torque,

for the whole section

38
where A is the area enclosed within the median line of the wall thickness.
Now, since

where t is the thickness at the point in question.

It is evident, therefore, that the maximum shear stress in such cases occurs at the point
of minimum thickness.
Consider now an axial strip of the tube, of length L, along which the thickness and
hence the shear stress is constant. The shear strain energy per unit volume is given by

Thus, with thickness t, width ds and hence V = tLds

1
But the energy stored equals the work done = T
2

The angle of twist of the tube is

therefore given by

For tubes of constant thickness this reduces to

where s is the perimeter of the median line.

The above equations must be used with care and do not apply to cases where there are
abrupt changes in thickness or re-entrant corners.

39
For closed sections which have constant thickness over specified lengths but varying
from one part of the perimeter to another:

40
\

Pressure vessel
4. Thin and thick cylinder

The pressure vessels (i.e. cylinders or tanks) are used to store fluids under pressure. The pressure
vessels are designed with great care because rupture of a pressure vessel means an explosion which
may cause loss of life and property. The material of pressure vessels may be brittle such ascast
iron, or ductile such as mild steel.

Classification of Pressure Vessels

The pressure vessels may be classified as follows:

• According to the dimensions

• Thin shell- If the wall thickness of the shell (t) is less than 1/20
of the internaldiameter of the shell (d). e.g. boilers, tanks, pipes
etc
• If the internal fluid pressure (p) is less than 1/6 of the allowable stress,
• Thick shell- if the wall thickness of the shell is greater than 1/20
of the internaldiameter of the shell. E.g. gun barrels, high
pressure cylinders.
• If the internal fluid pressure is greater than 1/6 of the allowable stress
• According to the end construction.

The pressure vessels, according to the end construction, may be classified as

i) open end- A simple cylinder with a piston, such as cylinder of a press is an example
of an open end vessel, the circumferential or hoop stresses are induced by the fluid
pressure
ii) Closed end- a tank is an example of a closed end vessel. In case of vessels having
open ends, longitudinal stresses in addition to circumferential stresses are induced.

Thin cylinder

Stresses in a Thin Cylindrical Shell due to an Internal Pressure

The analysis of stresses induced in a thin cylindrical shell are made on the following
41
\

assumptions:

1. The effect of curvature of the cylinder wall is neglected.

2. No pressure gradient across the wall (only internal pressure) and self-weight is neglected
3. The tensile stresses are uniformly distributed over the thickness of the walls.
4. Radial stress is small and can be neglected
5. The effect of the restraining action of the heads at the end of the pressure vessel is
neglected.
➢ When a thin cylindrical shell is subjected to an internal pressure, it is likely to fail in
the following two ways:

1. It may fail along the longitudinal section (i.e. circumferentially), when the force due to
pressure of the fluid acting vertically upwards and downwards on the thin cylinder, tend to
burst or splitting the cylinder into two troughs, as shown in Fig. 5.1 (a).

2. It may fail across the transverse section (i.e. longitudinally), when the force due to
pressure of the fluid acting at the end of the thin cylinder, tend to burst or splitting the
cylinder into two cylindrical shells, as shown in Fig. 5.1 (b).

(a) Failure of a cylinderical shell along the longitudinal section (b) failure of a
cylinderical shell along the transverse section

Fig. 5.1. Failure of a cylindrical shell.

Thus the wall of a thin cylindrical shell subjected to an internal pressure has to withstand
tensile stresses of the following two types:

(a) Circumferential or hoop stress, and

(b) Longitudinal stress.

42
\

A) Expression for Circumferential or Hoop Stress: A tensile stress on longitudinal section

(or on the cylindrical walls) acting in a direction tangential to the circumference.

Fig. 5.2. Circumferential or hoop stress.

Consider a thin cylindrical shell subjected to an internal pressure as shown in Fig. 5.2 (a) and
(b).

Let p = Intensity of internal pressure,

d = Internal diameter of the cylindrical shell,

l = Length of the cylindrical shell,

t = Thickness of the cylindrical shell, and

σt1 = Circumferential stress for the material of the cylindrical shell.

✓ We know that the total force acting on a longitudinal section (i.e. along the diameter
X-X) of the shell
= Intensity of pressure × Projected area = p × d × l.........................1
✓ and the total resisting force acting on the cylinder walls

= σt1 × 2t × l ....................................................................................2

From equations (1) and (2), we have

σt1 × 2t × l = p × d × l
𝑝𝑑
σt1=
2𝑡

43
\

Note: In constructing large pressure vessels like steam boilers, riveted joints or welded joints
are used in joining together the ends of steel plates. In case of riveted joints, the hoop stress
of the cylinder,
𝑝𝑑
σt1=
2𝑡𝜂

𝜂 = ηl = Efficiency of the longitudinal riveted joint.

C) Expression for Longitudinal Stress: A tensile stress acting in the direction of the axis.
In other words, it is a tensile stress acting on the *transverse or circumferential section Y-Y
(or on the ends of the vessel).

Fig. 5.3. Longitudinal stress.

Consider a closed thin cylindrical shell subjected to an internal pressure as shown in Fig. 5.3
(a) and (b)

Let σt2 = Longitudinal stress.

→ In this case, the total force acting on the transverse section (i.e. along Y-Y)

= Intensity of pressure × Cross-sectional area

= P*𝜋 (d)2 .......................................................................................................... (i)

4

→ and total resisting force = σt2 × π d.t ............................................................... (ii)

44
From equations (i) and (ii), we have

σt2 × π d.t = P*𝜋 (d)2

4

σt2 = Pd
4t

If 𝜂c is the efficiency of the circumferential joint, then

Pd
σt2 =
4t𝜂c

Note: From above we see that the longitudinal stress is half of the circumferential or hoop
stress. Therefore, the design of a pressure vessel must be based on the maximum stress i.e. hoop
stress.

C) Expression for maximum shear stress

At any point in the material of the thin cylinder shell, the two principal stress, hoop and
longitudinal are tensile and perpendicular to each other
σ1−σ2 Pd Pd Pd
Maximum shear stress = τmax = = - =
2 2t 4t 8t

Change in Dimensions of a Thin Cylindrical Shell due to an Internal Pressure

When a thin cylindrical shell is subjected to an internal pressure, there will be an increase in the
diameter as well as the length of the shell.
Let l = Length of the cylindrical shell,
d = Diameter of the cylindrical shell,
t = Thickness of the cylindrical shell,
p = Intensity of internal pressure,
E = Young’s modulus for the material of the cylindrical shell, and
μ = Poisson’s ratio.
The increase in diameter of the shell due to an internal pressure is given by,

45
The increase in length of the shell due to an internal pressure is given by,

It may be noted that the increase in diameter and length of the shell will also increase its volume.
The increase in volume of the shell due to an internal pressure is given by

Thick cylinders
Difference in treatment between thin and thick Cylinders- basic assumptions
Neither of the above assumptions in thin cylinder can be used for thick cylinders for which
the variation of hoop and radial stresses is shown in Fig. 5.5, their values being given by the Lamé
equations:

The hoop stress in case of a thick cylinder will not be uniform across the thickness. Actually
the hoop stress will vary from a maximum value at the inner circumference to a minimum value at
the outer circumference.

Development of the theory for thick cylinders is concerned with sections remote from the ends
since distribution of the stresses around the joints makes analysis at the ends particularly complex.
For central sections the applied pressure system which is normally applied to thick cylinders is
symmetrical, and all points on an annular element of the cylinder wall will be displaced by the
same amount, this amount depending on the radius of the element. Consequently there can be no
shearing stress set upon transverse planes and stresses on such planes are therefore principal
stresses. Similarly, since the radial shape of the cylinder is maintained there are no shears on radial
or tangential planes, and again stresses on such planes are principal stresses. Thus, consideration
of any element in the wall of a thick cylinder involves, in general, consideration of a mutually
perpendicular, tri-axial, principal stress system, the three stresses being termed radial, hoop
(tangential or circumferential)and longitudinal (axial) stresses.

46
Fig.5.4. Thin cylinder subjected to internal Fig.5.5. Thick cylinder subjected to internal pressure.
pressure.

Development of the Lame theory

Consider the thick cylinder shown in Fig. 5.6. The stresses acting on an element of unit
length at radius r are as shown in Fig. 5.7, the radial stress increasing from σr, to σr + dσr over the
element thickness dr (all stresses are assumed tensile),

Fig. 6.4

Fig. 5.6
Fig. 5.7
For radial equilibrium of the element:

For small angles:

47
Therefore, neglecting second-order small quantities,

... (5.1)
Assuming now that plane sections remain plane, i.e. the longitudinal strain εL is constant across
the wall of the cylinder,
Then,

It is also assumed that the longitudinal stress σL is constant across the cylinder walls at points
remote from the ends.
... (5.2)
Substituting in (5.1) for σH,

Multiplying through by r and rearranging,

Therefore, integrating,
𝑩
𝝈𝒓 = 𝑨 − 𝒓 𝟐 ... (5.3)
𝑩
And from eqn. (5.2) 𝝈𝑯 =𝑨+ ... (5.4)
𝒓𝟐

The above equations yield the radial and hoop stresses at any radius r in terms of constants A and
B. For any pressure condition there will always be two known conditions of stress (usually radial
stress) which enable the constants to be determined and the required stresses evaluated

48
Thick cylinder- internal pressure only
Consider now the thick cylinder shown in Fig. 5.8 subjected to an internal pressure P, the
external pressure being zero. The two known conditions of stress which
enable the Lamé constants A and B to be determined are:

At r = R1 σr = - P and at r = R2 σr = 0

NB. - The internal pressure is considered as a negative radial stress since it

will produce a radial compression (i.e. thinning) of the cylinder walls and
the normal stress convention takes compression as negative. Fig. 5.8. Cylinder cross-section.

These equations yield the stress distributions indicated in Fig. 5.5 with maximum values of both
σr, and σH at the inside radius.

49
 ✓ Longitudinal stress
Consider now the cross-section of a thick cylinder with closed
ends subjected to an internal pressure P1 and an external pressure
P2 (Fig. 5.9).
For horizontal equilibrium:

Where σL is the longitudinal stress set up in the cylinder walls, Fig. 5.9. Cylinder longitudinal
section.

i.e. a constant.
For combined internal and external pressures, the relationship σL = A also applies.
✓ Maximum shear stress
We that the stresses on an element at any point in the cylinder wall are principal stresses. It
follows, therefore, that the maximum shear stress at any point will be given by eqn. as

i.e. half the difference between the greatest and least principal stresses. Therefore, in the case of
the thick cylinder, normally,

Since σH is normally tensile, whilst σr is compressive and both exceed σL in magnitude.

The greatest value of  max thus normally occurs at the inside radius where r = R1.

Change of cylinder dimensions

(A) Change of diameter

The diametral strain on a cylinder is equal to the hoop or circumferential strain.

Therefore, change of diameter = diametral strain x original diameter
= circumferential strain x original diameter

50
With the principal stress system of hoop, radial and longitudinal stresses, all assumed tensile, the
circumferential strain is given by

Thus the change of diameter at any radius r of the cylinder is given by

(B) Change of length

Similarly, the change of length of the cylinder is given by

1
∆L = 𝜎 [ L- 𝜈𝜎r - 𝜈𝜎H]
E
𝑢- Poisson’s ratio for thick cylinder

Graphical treatment -Lame line

The Lame equations when plotted on stress and 1/r2 axes produce straight lines, as shown in Fig.
5.10

Fig. 5.11 Lam15 line solution for cylinder

with internal and external pressures
Fig. 5.10 Graphical representation of Lame
equations - Lame line.

51
52
 Compound cylinders
To obtain a more uniform hoop stress distribution, cylinders are often built up by shrinking one
tube on to the outside of another. When the outer tube contracts on cooling the inner tube is brought
into a state of compression. The outer tube will conversely be brought into a state of tension. If
this compound cylinder is now subjected to internal pressure the resultant hoop stresses will be the
algebraic sum of those resulting from internal pressure and those resulting from shrinkage as drawn
in Fig.5.12; thus a much smaller total fluctuation of hoop stress is obtained. A similar effect is
obtained if a cylinder is wound with wire or steel tape under tension.

Fig. 5.12 Compound cylinders - combined internal pressure and shrinkage effects.

(A) for the Same materials

The method of solution for compound cylinders constructed from similar materials is to break the
problem down into three separate effects:

(a) Shrinkage pressure only on the inside cylinder (Fig. 5.13-a);

(b) Shrinkage pressure only on the outside cylinder (Fig. 5.13-b);

53
 (c) Internal pressure only on the complete cylinder (Fig.5.13-c).

Fig. 5.13. Method of solution for compound cylinders.

For each of the resulting load conditions there are two known values of radial stress which
enable the Lamé constants to be determined in each case:

✓ i.e. condition (a) shrinkage - internal cylinder:

At r = R1, σr = 0
At r = Rc σr = - p (compressive since it tends to reduce the wall thickness)
✓ Condition (b) shrinkage - external cylinder:
At r = R2, σr = 0
At r = Rc , σr = - p
✓ Condition (c) internal pressure - compound cylinder:
At r = R2, σr = 0
At r = R1, σr = - P1
Thus for each condition the hoop and radial stresses at any radius can be evaluated and the principle
of superposition applied, i.e. the various stresses are then combined algebraically to produce the
stresses in the compound cylinder subjected to both shrinkage and internal pressure. In practice
this means that the compound cylinder is able to withstand greater internal pressures before failure
occurs or, alternatively, that a thinner compound cylinder (with the associated reduction in material
cost) may be used to withstand the same internal pressure as the single thick cylinder it replaces.

(B) For Different materials

The value of the shrinkage or interference allowance for compound cylinders constructed from
cylinders of different materials is given by eqn.

54
Total interference or shrinkage allowance

The value of the shrinkage pressure set up owing to a known amount of interference can then be
calculated as with the standard compound cylinder treatment, each component cylinder being
considered separately subject to the shrinkage pressure.

For a full solution of problems of this type it is often necessary to make use of the equality of
diametral strains at the common junction surface, i.e. to realize that for the cylinders to maintain
contact with each other the diametral strains must be equal at the common surface.

Now diametral strain = circumferential strain

Therefore at the common surface, ignoring longitudinal strains and stresses,

Where Eo and νo = Young’s modulus and Poisson’s ratio of outer cylinder,

Ei and νi = Young’s modulus and Poisson’s ratio of inner cylinder,
σr = - p = radial stress at common surface,
And σHo = (as before) the hoop stresses at the common surface

55
1. An external pressure of 10 MN/m2 is applied to a thick cylinder of internal diameter 160 mm
and external diameter 320 mm. If the maximum hoop stress permitted on the inside wall of the
cylinder is limited to 30 MN/m2, what maximum internal pressure can be applied assuming the
cylinder has closed ends? What will be the change in outside diameter when this pressure is
applied? E = 207 GN/m2,  = 0.29.

Solution

The conditions for the cylinder are:

56
2. (a) In an experiment on a thick cylinder of 100 mm external diameter and 50 mm internal
diameter the hoop and longitudinal strains as measured by strain gauges applied to the outer surface
of the cylinder were 240 x l0-6 and 60 x 10-6, respectively, for an internal pressure of 90 MN/m2,
the external pressure being zero. Determine the actual hoop and longitudinal stresses present in the
cylinder if E = 208 GN/m2 and  = 0.29. Compare the hoop stress value so obtained with the
theoretical value given by the Lame equations.

(b) Assuming that the above strain readings were obtained for a thick cylinder of 100 mm external
diameter but unknown internal diameter calculate this internal diameter.

Solution

a) Since σr = 0 at the outside surface of the cylinder for zero external pressure.

The theoretical values of σH for an internal pressure of 90 MN/m2 may be obtained from

Fig. 5.16, the boundaries of the cylinder being given by r = 0.05 and r = 0.025,

57
 Fig. 6.13

Therefore for the internal radius R1 where σr = 90 MN/m2

Internal diameter = 49.6 mm

58
It is thus possible to determine the value of 1 R 21 which will produce σr = - 90 MN/m2.

Fig.5.16

59
 CHAPTER -5
THEORIES OF ELASTIC FAILURE

Theories of Failure under Static Load

Mechanical components are subjected to several types of loads simultaneously. For example:
• A power screw is subjected to torsional moment as well as axial force.
• An overhang crank is subjected to combined bending and torsional moment.
• The bolts of the bracket are subjected to forces that cause tensile stress and shear stress.
• Crankshaft, propeller shaft, and connecting rod are examples of the components subjected
to complex loads.
When the component is subjected to several types of loads, combined stresses are induced. For
example, torsional moment induces torsional shear stress, while bending moment causes bending
stress in the transmission shaft. The failures of such components are broadly classified into two
groups; i.e. Elastic failure, yielding and fracture.
Elastic failure results in excessive elastic deformation, which makes the machine component unfit
to perform its function satisfactorily.
Yielding results in excessive plastic deformation after the yield point stress is reached, while
fracture results in breaking the component into two or more pieces. Theories of failure discussed in
this article are applicable to elastic failure of machine parts.
The design of machine parts subjected to combined loads should be related, to experimentally
determined properties of material under ‘similar’ conditions. However, it is not possible to conduct
such tests for different possible combinations of loads and obtain mechanical properties. In practice,
the mechanical properties are obtained from simple tension test. They include yield strength,
ultimate tensile strength and percentage elongation. In tension test, the specimen is axially loaded
in tension. It is not subjected to either bending moment or torsional moment or a combination of
loads. Theories of elastic failure provide a relationship between the strength of machine
component subjected to complex state of stresses with the mechanical properties obtained in
tension test. With the help of these theories, the data obtained in tension test can be used to
determine the dimensions of the component, irrespective of the nature of stresses induced in the
component due to complex several theories have been proposed, each assuming a different
hypothesis of failure.

60
The principal theories of elastic failure are as follows:
i. Maximum principal stress theory (Rankine’s theory);
ii. Maximum principal strain theory (Saint-Venant).
iii. Maximum shear stress theory (Coulomb, Tresca and Guest’s theory);
iv. Distortion energy theory (Huber von Mises and Hencky’s theory);
v. Maximum total strain energy theory (Haigh’s theory).
We will discuss the main theories in this chapter. Let us assume that 𝜎1, 𝜎 2 and 𝜎 3 as the principal
stresses induced at a point on the machine part as a result of several types of loads. We will apply
the theories of failure to obtain relationship between 𝜎 1, 𝜎 2 and 𝜎 3 on one hand and the properties
of material such as yield strength in tension (Syt) or ultimate strength in tension (Sut ) on the other.

i. Maximum principal (Normal) stress Theory (MNST)

This theory states that the failure of the mechanical component, subjected to bi-axial or tri-axial
stresses, occurs when the maximum principal stress reaches the yield or ultimate strength of
the material.
If 𝜎 1, 𝜎 2 and 𝜎 3 are the three principal stresses at point on the component and

𝜎1> 𝜎2> 𝜎3
Then according to this theory, the failure occurs whenever,
𝜎1= Syt or 𝜎1= Sut ......................................................................1

The theory considers only the maximum of principal stresses and disregards the influence of the
other principal stresses. The dimensions of the component are determined by using a factor of
safety.
For tensile stresses,
𝑺𝒖𝒕 ...........................................................................
𝜎1= OR 𝜎1= 2
𝑺𝒚𝒕
𝒇𝒔 𝒇𝒔

For compressive stresses,

𝑺𝒖𝒄 ..........................................................................
𝜎1=
OR 𝜎1= 3
𝑺𝒚𝒄 𝒇𝒔
𝒇𝒔

Where fs = Factor of safety

61
Region of Safety: The construction of a region of safety for bi-axial stresses is illustrated in Fig.
7.1. The two principal stresses 𝜎1 and 𝜎2 are plotted on X and Y-axes respectively. Tensile
stresses are considered as positive, while compressive stresses as negative. It is further assumed
that
Syc = Syt

It should be noted that,

(i) The equation of vertical line to the positive side of X-axis is (x = + a)
(ii) The equation of vertical line to the negative side of X-axis is (x = - a)
(iii) The equation of horizontal line to the positive side of Y-axis is (y = + b)
(iv) The equation of horizontal line to the negative side of Y-axis is (y = - b)
The borderline for region of safety for this theory can be constructed in the following way:
Step 1: Suppose 𝜎1 > 𝜎2. As per this theory we will consider only the maximum of principal stresses
(𝜎1) and disregard the other principal stress (𝜎2). Suppose (𝜎1) is tensile stress. The limiting value of
(𝜎 𝜎1= +Syt
A vertical line AB is constructed such that: 𝜎1 = + Syt.
Step 2: Suppose 𝜎1 > 𝜎2 and (𝜎1) is compressive stress. The limiting value of (𝜎1) is compressive is
yield stress (Syt). Therefore, the boundary line will be,
𝜎1= -Syc
A vertical line DC is constructed such that: 𝜎1 = -Syc.
Step 3: Suppose 𝜎2> 𝜎1. As per this theory we will consider only the maximum of principal stresses
(𝜎2) and disregard the other principal stresses (𝜎1). Suppose
(𝜎2) is tensile stress. The limiting value of (𝜎2) is yield stress
(Syt). Therefore, the boundary line will be,
𝜎2 = + Syt
A horizontal line CB is constructed such that: 𝜎2= + syt.
Step 4: Suppose 𝜎2> 𝜎1, and (𝜎2) is compressive stress. The
limiting value of (𝜎2) is compressive yield stress (-Syc).
Therefore, the boundary line will be,
𝜎2= -Syc Fig. 7.1 Boundary for maximum principal
A horizontal line DA is constructed such that: stress theory under bi-axial stress

𝜎2 = - Syc.

62
Now the complete region of safety is area ABCD. Since, we have assumed (Syc = Syt), ABCD is
square.
According to the maximum principal theory of failure,
❖ If a point with co-ordinates (𝜎1, 𝜎2) falls outside this square, then it indicates failure occur.
❖ If the point falls inside the square, the design is safe and failure may not occur.
Experimental investigations suggest that maximum principal stress theory gives good predictions for
brittle materials. However, it is not recommended for ductile materials.

ii. Maximum shear stress theory (MSST)

This theory states that the failure of a mechanical
component subjected to bi-axial or tri -axial stresses occurs
when the maximum shear stress at any point in the
component becomes equal to the maximum shear stress
in the standard specimen of the tension test, when
yielding starts. In tension test, the specimen is subjected
to uni-axial stress (𝜎 1) and (𝜎 2 = 0). The stress in the
specimen of tension test and the corresponding Mohr’s
circle diagram are shown in Fig. 7.2. From the figure,
𝜎1
𝑟𝑚𝑎𝑥 =
2
Fig.7.2 a) Stress in simple tension test
b) Mohr’s circle for stress
When the specimen starts yielding (𝜎1 = Syt), the above equation is written as
Syt
𝑟max =
2

Therefore, the maximum shear stress theory predicts that the yield strength in shear is half of the
yield strength in tension, i.e.
Ssy = 0.5Syt .....................................................................................................
… 4

If 𝜎1, 𝜎2 and 𝜎3 are the three principal stresses at a point on the component, the shear stresses on three
different planes are given by,

……………(a)

63
 Syt
The largest of these stresses is equated to (𝑟max) or ( ). Considering factor of safety,
2

……….5

The above relationships are used to determine the dimensions of the component.
Refer to expression (a) again and equating the largest shear stress (𝑟max ) to (Syt/2),

……………………………………(b)
Similarly,

………(d)
For compressive stress,

The above equations can be written as,

𝜎1- 𝜎2 = ±Syt 𝜎2- 𝜎3 = ±Syt 𝜎3- 𝜎1 = ±Syt (Assuming that Syc = Syt)

Region of safety: For bi-axial stresses, 𝜎3= 0

The above equations can be written as,
𝜎1- 𝜎2= ±Syt ……(h) , 𝜎2= ±Syt ….…(i) , 𝜎1= ±Syt ………(j)
It will be observed at a later stage that equations (h) are
applicable in second and fourth quadrants, while
equations (i) and (j) are applicable in first and third
quadrants of the diagram. The construction of the region
of safety is illustrated in Fig. 7.3. The two principal
stresses 𝜎 1 and 𝜎 2 are plotted on X and Y axes
respectively. Tensile stresses are considered as positive,
while compressive stresses as negative.

Fig.7.3: Boundary for Maximum Shear Stress

Theory under bi-axial stresses
It should be noted that,
(i) The equation (x - y = - a) indicates a straight line in second quadrant with (- a) and (+ a) as
intercepts on X and Y axes respectively.

64
 (ii) The equation (x - y = + a) indicates a straight line in fourth quadrant with (+ a) and (- a) as
intercepts on X and Y axes respectively.

iii. Distortion - Energy Theory (DET)

This theory states that the failure of the mechanical component subjected to bi-axial or tri-axial
stresses occurs, when the strain energy of distortion per unit volume at any point in the
component, becomes equal to the strain energy of distortion per unit volume in a standard
specimen of tension test, when yielding starts.
A unit cube subjected to the three principal 𝜎1, 𝜎2 and 𝜎3 is shown in Fig. 7.4 (a). The total
strain energy U of the cube is given by,

𝜎1, 𝜎2 and 𝜎3 are strains in respective directions.

Fig. 7.4 (a) Element with Tri-axial Stresses (b) Stress Components due to
Distortion of Element (c) Stress Components due to Change of Volume.
Also,

Substituting the above expressions in Eq. (a),

65
The total strain energy U is resolved into two components: one Uv corresponding to the change of
volume with no distortion of the element and the other Ud corresponding to the distortion of the
element with no change of volume.
Therefore,

The corresponding stresses are also resolved into two components as shown in Figs 7.4 (b) and (c).
From the figure,

The components σ1d, 𝜎2d, and 𝜎3d cause distortion of the cube, while the component

𝜎v results in volumetric change. Since the components 𝜎1d, 𝜎2d and 𝜎3d do not change the volume of
the cube,

Also,

Substituting Eq. (g) in Eq. (f),

Substituting Eq. (h) in Eq. (e),

66
The strain energy U, corresponding to the change of volume for the cube is given by,

From expressions (k) and (l),

Substituting expressions (j) in the above equation,

From expressions (c) and (n),

or

…………………………….6

In simple tension test, when the specimen starts yielding,

and
Therefore, … .(7 )
………………….7
From Eqs (6) and (7), the criterion of failure for the distortion energy theory is expressed as

or

…....(8)

Considering factor of safety,

….(9)

For bi-axial stresses (𝜎3 = 0),

…
 …(10)
A component subjected to pure shear stress and the corresponding Mohr’s circle diagram is shown
in Fig. 7.5.

Fig.7.5 (a) Element subjected to Pure Shear Stresses (b) Mohr’s circle for Shear Stresses

From the figure,

Substituting these values in eqn. (8)

Replacing ( 𝑟xy) by ( Ssy ),

4
Where J=π𝑑 for solid shaft
32

C=d/2
d=diameter of the shaft
0.557Syt
Remember: Allowable shear stress (permissible stress) =
f.s

Therefore, according to the distortion-energy theory, the yield strength in shear is 0.577 times the
yield strength in tension.
Experiments have shown that the distortion-energy theory is in better agreement for predicting the
failure of ductile component, than any other theory of failure.
68
7.2. Selection and use of failure theories
The plots of three theories of failure on 𝜎1, 𝜎2 coordinate system are shown in Fig. 7.6. While
selecting theories of failure, following points should be noted:
(i) Ductile materials typically have same tensile strength and compressive strength. Also,
yielding is the criterion of failure in ductile materials. In maximum shear stress theory
and distortion energy theory, it is assumed that the yield strength in tension (Syt) is equal
to yield strength in compression (Syc). Also, the criterion of failure is yielding.
Therefore, maximum shear stress theory and distortion energy theory are used for
ductile materials.
(ii) Distortion energy theory predicts yielding with precise accuracy in all four quadrants.
The design calculations involved in this theory are slightly complicated as compared
with those of maximum shear stress theory.
(iii) The hexagonal diagram of maximum shear stress theory is inside the ellipse of distortion
energy theory. Therefore, maximum shear stress theory gives results on the conservative
side. On the other hand, distortion energy theory is more liberal.
(iv) The graph of maximum principal stress theory is same as that of maximum shear stress
theory in first and third quadrants. However, the graph of maximum principal stress
theory is outside the ellipse of distortion energy theory in second and fourth quadrants.
Thus, it would be dangerous
to apply maximum principal
stress theory in these
regions, since it might
predict safety, when in fact
no safety exists.

Fig. 7.6 : Comparison of theories of Failure

(v) Maximum shear stress theory is used for ductile materials, if dimensions need not be
held too close and a generous factor of safety is used. The calculations involved in this
theory are easier than those of distortion energy theory.

69
(vi) y
t
h
e
o
r
y
c
a
u
s
e
o
f
f
a
i
r
e
o
f
c
o
m
p
o
n
e
n
t
i
s
b
70