m00842 Class Notes2

ii
Instituto Tecnológico y de Estudios Superiores de Monterrey Spring 2002

Campus Toluca
Mechanics of materials II
Class notes
Dr. José Carlos Miranda
c
Copyright 2002 Dr. José Carlos Miranda. All right reserved.
iv
Chapter 3 is devoted to the method of superposition. The principles of super-

position and proportionality are introduced and discussed in some examples.
The method is applied to find the reactions in statically indeterminated beams.
Continuous beams are presented and analized in chapter 4. The method of

three moments is introduced as an effective methodology to analyze this type
Preface of beams.
In chapter 5 the study of mechanical components subject to static loading is

presented. The concept of state of stress and state of strain is introduced.
Formulae to transform the states of stress and strain are derived. These for-
The following notes for the course Mechanics of materials II (M-00842) have mulae are used to obtain the principal stresses and strains acting at a given
been prepared with the objective of presenting the student with the topics of point. The relationship between stress and strain is discussed. Finally, the
the course in the easiest way possible (but not easier!). above concepts are used to present failure criteria for the design of mechanical
components subject to static loading.
This notes are not a course book. Most of the material has been taken from
the following sources. The student should revise them, Chapter 6 deals with the idea of strain energy. To start, the concept of strain
energy is introduced. Formulae to calculate the strain energy stored in me-
chanical components subject to static and impact loading are given. Some
1. Personal notes for the Mechanics of Materials II course. Dr. Karim important phenomena regarding loading and strain energy are discussed.
Heinz Muci-Küchler. 1998. In chapter 7 the principles for the analysis of columns are presented. The
concept of failure by buckling is introduced in the scope of ideal columns using
2. Timoshenko, Stephen & Gere, James. Mechanics of materials. PWS Euler’s formula. The critical load of a column is obtained and applied to
Publishing Company, Fourth Ed., 1997. their design. Critical loads for columns having different types of supports are
presented.
3. Hibbeler, Russell. Statics and Mechanics of Materials. Prentice Hall,
Fourth Ed., 1999. It is important to mention that these notes are far from finished. Each class the
students bring new doubts, point of views, bring up new aspects that should
4. Spiegel, Leonard & Limbrunner, George. Applied Strength of Materials. be addressed, or simply find errors in the manuscript. In this sense, I extend
Macmillan College Publishing, 1994. my gratitude to those students that have contributed to improve these notes.
These notes were prepared completely using free software whose programs and
computer codes can be freely shared. The text was prepared using LATEX and
The notes are divided into 7 chapters. The first one deals with the deflection the figures were done using XFig, The Gimp and XSane. The only exception was
of beams by the integration method. The concepts of elastic curve and the MuPAD which was used to plot some graphs. MuPAD’s code is propietary. The
relationship between the internal moment of the beam and its curvature are platform of development was a standard PC running GNU/Linux (Mandrake
discussed. 8.0 y 8.1). During the realization of these notes, not a single computer crash
In chapter 2 the deflection of beams is analyzed by means of the moment– was experienced.
area method. The first and second moment–area theorems are derived. The
method is discussed applying it to cantilever and simple supported beams.
c
Copyright 2002 Dr. José Carlos Miranda. All right reserved. c
vi CONTENTS
2.3.2 Beam with symmetric loading . . . . . . . . . . . . . . . 22
2.3.3 Beam with non–symmetric loading . . . . . . . . . . . . 27
3 Method of superposition 31
Contents 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2 Principle of superposition . . . . . . . . . . . . . . . . . . . . . 33
3.3 Principle of proportionality . . . . . . . . . . . . . . . . . . . . . 33
3.4 Proportionality and superposition . . . . . . . . . . . . . . . . . 33

Contents iii
3.5 Statically indeterminate beams . . . . . . . . . . . . . . . . . . 35
1 Deflection of beams by the integration method 5

4 Continuous beams 42
1.1 The elastic curve . . . . . . . . . . . . . . . . . . . . . . . . . . 5
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
1.2 Moment-curvature relationship . . . . . . . . . . . . . . . . . . . 6
4.2 Method of three moments . . . . . . . . . . . . . . . . . . . . . 43
1.3 Slope and Displacement by integration . . . . . . . . . . . . . . 7
4.2.1 Procedure for analysis . . . . . . . . . . . . . . . . . . . 46
1.3.1 Sign convention and coordinates . . . . . . . . . . . . . . 8
4.2.2 Angles of rotation . . . . . . . . . . . . . . . . . . . . . . 46
1.3.2 Boundary and continuity conditions . . . . . . . . . . . . 8
4.2.3 Beam with fixed support . . . . . . . . . . . . . . . . . . 46
1.4 Procedure for analysis . . . . . . . . . . . . . . . . . . . . . . . 13
4.2.4 Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2.5 Example 4.1 . . . . . . . . . . . . . . . . . . . . . . . . . 47

2 Deflection of beams by the moment–area method 14
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 5 Mechanical components subject to static loading 49

2.2 Moment–area theorems . . . . . . . . . . . . . . . . . . . . . . . 15 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.2.1 First moment–area theorem . . . . . . . . . . . . . . . . 17 5.2 Components of stress . . . . . . . . . . . . . . . . . . . . . . . . 50
2.2.2 Second moment–area theorem . . . . . . . . . . . . . . . 17 5.3 Stresses on inclined sections . . . . . . . . . . . . . . . . . . . . 51
2.3 Application to simple beams . . . . . . . . . . . . . . . . . . . . 20 5.4 Principal stresses and maximum share . . . . . . . . . . . . . . 53
2.3.1 Cantilever beam . . . . . . . . . . . . . . . . . . . . . . . 20 5.5 Strain transformation . . . . . . . . . . . . . . . . . . . . . . . . 55
c
CONTENTS vii viii CONTENTS
5.5.1 Principal strains . . . . . . . . . . . . . . . . . . . . . . 58 7.3 Ideal column with pin supports . . . . . . . . . . . . . . . . . . 83
5.5.2 Strain rosettes . . . . . . . . . . . . . . . . . . . . . . . . 58 7.4 Critical loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.6 Hooke’s law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 7.5 Radius of gyration . . . . . . . . . . . . . . . . . . . . . . . . . 86
5.7 Failure criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 7.6 Critical stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5.7.1 Rankine criterion . . . . . . . . . . . . . . . . . . . . . . 62 7.7 Optimum shapes of columns . . . . . . . . . . . . . . . . . . . . 88
5.7.2 Tresca criterion . . . . . . . . . . . . . . . . . . . . . . . 62 7.8 Columns with other support conditions . . . . . . . . . . . . . . 89
5.7.3 Von Mises criterion . . . . . . . . . . . . . . . . . . . . . 63 7.9 Effective length . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.7.4 Mohr-Coulomb criterion . . . . . . . . . . . . . . . . . . 64
6 Strain Energy 65
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.2 Principles of strain energy . . . . . . . . . . . . . . . . . . . . . 66
6.3 Impact loading . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.4 Strain energy density . . . . . . . . . . . . . . . . . . . . . . . . 68
6.4.1 Toughness and resilience modulus . . . . . . . . . . . . . 69
6.5 Elastic strain energy for normal stresses . . . . . . . . . . . . . 70
6.6 Strain energy in bending . . . . . . . . . . . . . . . . . . . . . . 74
6.7 Elastic strain energy for shear stress . . . . . . . . . . . . . . . . 74
6.8 Strain energy in torsion . . . . . . . . . . . . . . . . . . . . . . . 75
6.9 Deflection of beams due to impact load . . . . . . . . . . . . . . 76
7 Columns 79
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.2 Stable, neutral and unstable equilibrium . . . . . . . . . . . . . 82
c
2 Deflection of beams by the integration method
Deflection of beams by the 1798, likely formed the starting point for his theory of interference. In fact,
his interests and contributions were so legion that he made some anonymously
integration method to avoid the charge that he was neglecting his professional duties!
He was appointed foreign secretary to the Royal Society in 1802, a post that
he held to the end of his life. He resigned his professorial position at the Royal
Institution, feeling that his duties were affecting his medical career. The same
year he received the MB degree from Cambridge, and five years later, the
degree of MD.
It was during this period that Young conducted his now-famous experimen-
tal investigations on light. Young made significant contributions to physical
optics in the areas of double refraction and dispersion. The importance of
his work was not apparent to his contemporaries and his principle of interfer-
ence remained more or less obscure for another fourteen years, when it was
’rediscovered’ by Fresnel.
Young was the first to assign the term energy to the quantity mv 2 and to
put work done, which he defined as f orce × distance, proportional to energy.
Also he introduced absolute methods for determining the elastic properties of
materials - such as the Young’s modulus that relates the increase in length of a
wire to the force applied - and he developed the most comprehensive theory of
tides then available. His contributions to archeology were equally impressive,
he assisted in deciphering the Egyptian hieroglyphics inscribed on the Rosetta
Thomas Young (1773–1829)
Stone.
Thomas Young was born into “comfortable circumstances” at Milverton, Eng-

land on June 13, 1773, towards the end of a period known as the Intellectual
Revolution. Young was a precocious child who could read fluently at the age 1.1 The elastic curve
of two. By the time he was sixteen he was proficient in Greek and Latin and
was well acquainted with eight other languages, classical and modern. By the The elastic curve is the curve that passes through the centroid of every cross-
age of eighteen he was recognized as a truly accomplished scholar. sectional area of a beam. Although the integration method will allow us to
In 1792, at age nineteen, Young decided on a career in medicine. The following compute the elastic curve analytically, it is always helpful to sketch the elastic
year he presented a paper before the Royal Society in which he attributed the view beforehand since for most beams the elastic curve can be sketched without
accommodation of the eye to its muscular structure; he was elected one year much difficulty (see figure 1.1).
later to membership of the Society. After completing his medical studies at To successfully sketch the elastic curve of a beam it is important to remember
Edinburgh and Göttingen, he returned to London to practice but continued that in general:
his scholarly studies at Emmanuel College, Cambridge. He became financially
independent on the death of an uncle and that allowed him to pursue his real
interests. Some investigations on sound and light, which he carried out in • Supports that resist a force restrict displacement.
c
1.2 Moment-curvature relationship 3 4 Deflection of beams by the integration method
P1 P2 O0
B C E D PSfrag replacements
A
dθ

ρ ρ
B D
ds ds0
y dx M y dx M
A C E

Before After
P M
PSfrag replacements
Figure 1.2: Typical cross section of the beam before and after deformation.

A C

1. The beam is initially straight.

B

2. The beam is elastically deformed by load applied perpendicular to the

C beam axis.

A 3. The load is lying in the x-v plane.

B 4. All the deflection of the beam is caused by bending, i.e. the deflection

caused by shear is ignored.

Figure 1.1: Elastic curves of two sample beams.
• Supports that resist a moment restrict displacement and slope. Consider figure 1.2. If an internal moment M deforms the beam, the cross
sections of the beam will suffer a change. To quantify this change it is conve-
nient to use the angle between them. In order to do so, and using figure 1.2
the following quantities must be defined:
1.2 Moment-curvature relationship

• dθ is the angle between cross sections.
To analytically obtain the elastic curve of a beam it is useful to look to the • dx is the arc representing the neutral axis for each cross section.
relationship between the internal moment M in the beam and the radius of
curvature ρ of the elastic curve at a point. • ρ is the radius of curvature for the arc dx measured from the center of
curvature O0 .
To obtain such relationship between ρ and M we will limit the analysis using
the following considerations: • Saint-Venant’s principle applies.
c
1.2 Moment-curvature relationship 5 6 Deflection of beams by the integration method
Any arc ds in the cross section different from dx is subject to a normal strain Using the flexure formula, σ = −M y/I we can also express the curvature in
in the x direction as its length changes. Consider for example the strain in the terms of the stress in the beam:
arc ds located at a position y from the neutral axis dx:
1 σ
ds0 − ds =− (1.5)
ε= . (1.1) ρ Ey
ds
Equations (1.4) and (1.5) are valid for either small or large radii of curvature.
However, ds = dx = ρdθ and ds0 = (ρ − y)dθ giving
However, as the deflection of the beam is usually very small, ρ is usually a
very large quantity.
(ρ − y)dθ − ρdθ
ε= (1.2)
ρdθ
or 1.3 Slope and Displacement by integration
1 ε The elastic curve for a beam can be expressed mathematically as v = f (x).
=− (1.3)
ρ y To obtain this equation, we must first represent the curvature (1/ρ) in terms
of the coordinates v and x. The relationship between x, v and 1/ρ is
If the material is homogeneous and behaves in a linear-elastic manner, then
Hooke’s law ε = σ/E and the flexure formula σ = −M y/I applies. Combining
these equations and substituting into eq. (1.3) gives 1 d2 v/dx2
= (1.6)
ρ [1 + (dv/dx)2 ]3/2
1 M
=− (1.4) Substituting eq. (1.4) into eq. (1.6) gives
ρ EI
where
M d2 v/dx2
= (1.7)
EI [1 + (dv/dx)2 ]3/2
• ρ is the radius of curvature at a specific point on the elastic curve (1/ρ
is referred to as the curvature).
This equation represents a nonlinear second-order differential equations. Its
• M the internal moment in the beam at the point where ρ is to be deter- solution, which is referred as the elastica, gives the exact shape of the elastic
mined. curve assuming that beam deflections occur only due to bending. Unfortu-
nately, the solution of this equation is involved and only can be obtained for
• E is the material’s modulus of elasticity. very simple cases.
• I the beam’s moment of inertia computed about the neutral axis. In order to facilitate its solution, eq. (1.6) can be simplified taking into ac-
count that most engineering designs involve very small deflections. Therefore,
The product EI, known as the flexural rigidity is always positive. The sign for the slope of the elastic curve dv/dx will be very small and its square will
ρ depends on the direction of the moment. For a positive moment, ρ extends be negligible compared to the unity such that 1 + (dv/dx)2 ≈ 1. Using this
above the beam. simplification it is possible to write
c
1.3 Slope and Displacement by integration 7 8 Deflection of beams by the integration method
+w
d2 v M
2
= (1.8)
dx EI PSfrag replacements
which is a linear second-order differential equation. +M +V +V +M
It is also possible to write this equation in two alternative forms. Taking the
derivative of each side of eq. (1.8) with respect to x and substitute V = dM/dx Figure 1.3: Convention for positive quantities.
an equation describing the shear force V can be obtained
1.3.1 Sign convention and coordinates
2

d dv
EI 2 = V (x) (1.9) When applying eqs. (1.11) it is important to use the proper signs for M , V
dx dx
or w as established by the sign convention that was used in the derivation of
these equations. As a reminder:
Differentiating again, an equation describing the distributed load intensity
−w = dV /dx arises
• positive deflection, v is upward.
• positive slope angle θ will be measured counterclockwise from the x axis
d2 d2 v

EI 2 = −w(x) (1.10) when x is positive to the right.
dx2 dx
In order to solve any problem it is also convenient to remember that:
For most problems the flexural rigidity EI will be constant along the length of
the beam. Assuming this to be the case, the above results may be reordered • The slope of the elastic curve dv/dx is very small so the length of the
into the following sets of equations: beam and the elastic curve will be about the same.
• Therefore, point on the elastic curve are assumed to be displaced
vertically and not horizontally.
d4 v
EI = −w(x) • The slope angle θ is very small and its value in radians can be obtained
dx4
dv3 directly from θ ≈ tan θ = dv/dx.
EI 3 = V (x)
dx • If a single x coordinate cannot be used to express the equation for the
d2 v elastic curve, then an appropriate set of x coordinates should be selected.
EI 2 = M (x) (1.11)
dx
Solution of any of these equations requires successive integrations to obtain 1.3.2 Boundary and continuity conditions
the deflection v of the elastic curve. For each integration it is necessary to
introduce a “constant of integration” and then solve for all the constants to As mentioned, the determination of the elastic curve involves sucessive inte-
obtain a unique solution for a particular problem. gration and the determination of the constants of integration.
c
a b
P

B
A

v=0 v=0 θ=0 C

M =0 v=0
frag replacements Roller Roller Fixed end x2
x1
(b)
PSfrag replacements

a
v=0 v=0 V =0 b
M =0 M =0 P
Pin Pin Free end B
A C
Figure 1.4: Boundary conditions for beams.

x2 x1
The constants of integration are determined by evaluating the functions for (a)
shear, moment, slope or displacement at a particular point on the beam where
the value of the function is known. This values are known as boundary condi- Figure 1.5: Different sets of x coordinate.
tions.
Some common boundary conditions used to solve beam deflection problems In figure 1.5(b) continuity requires that:
are shown in figure 1.4.
Sometimes it is not possible to use only one x coordinate to solve a given • θ1 (a) = −θ2 (b) (x1 and x2 extend in opposite directions).
problem as show in figure 1.5 and two or more x coordinates must be define.
• v1 (a) = v2 (b) (continuity of deflection)
Since the elastic curve is physically continuous, continuity conditions are used
to match the deflection and the slope at points where the different x coordi-
nates unite. These continuity conditions must be used to evaluate some of the Example 1.1
integration constants.
In figure 1.5(a) continuity requires that: GIVEN The cantilever beam show in figure 1.6 is subject to a vertical load
P at its end.
FIND Determine the equation of the elastic curve. EI is constant.

• θ1 (a) = θ2 (b) (continuity of slope)
Solution I
• v1 (a) = v2 (b) (continuity of deflection) Elastic curve. The load tends to deflect the beam downwards as shown in the
c
figure. Since the load and restrictions are applied at the ends of the beam,
only one x coordinate is needed.
P
θ= (L2 − x2 )
Moment function. For this beam the moment is represented simply by 2EI
P
M = −P x (a) v= (−x3 + 3L2 x − 2L3 ) (e)
6EI
Slope and elastic curve. Applying Solution II

d2 v This problem can also be solve using
EI 2 = M (x) (b)
dx
d4 v
d v 2 EI = −w(x) (f)
EI = −P x dx4
dx2
dv P x2 In this example the distributed load is equal to zero all over the beam (it does
EI =− + C1
dx 2 not exist). Therefore, w(x) = 0 for 0 ≤ x ≤ L. Integrating once
3
Px
EIv = − + C1 x + C2 (c)
6
d4 v
Using the boundary conditions dv/dx = 0 at x = L and v = 0 at x = L it is EI =0
dx4
possible to obtain d3 v
EI 3 = C1 = V (g)
P L2 P L2 dx
0=− + C1 =⇒ C1 =
2 2
P L3 −P L3 The constant C1 can be evaluated at x = 0 since V (0) = −P (negative ac-
0=− + C1 L + C2 =⇒ C2 = (d)
6 3 cording to the beam sign convention). Thus, C1 = −P . Integrating again,
frag replacements
substituting these results and remembering that θ = dv/dx,
d3 v
v EI = −P
dx3
P d2 v
EI = −P x + C2 = M (h)
dx2

A B M

xv

A
θA V Since M (0) = 0, then C2 = 0 resulting in
x

x d2 v
EI = −P x (i)
L dx2
Figure 1.6: Problem 1.1. an the solution continues as before.
c
PSfrag replacements P for 0 ≤ x1 < 2a

x1 x2
d2 v 1 P
EI 2
= − x1
dx1 2
A C
dv1 P 2
B vC EI = − x1 + C 1
dx1 4

P 3
2a a EIv1 = − x1 + C1 x1 + C2 (b)
12
P for 0 ≤ x2 < a
x1 x2
d2 v 2
M1 M2 EI = −P x2
dx22
dv2 P
V1 V2 EI = − x22 + C3
dx2 2
P P 3
EIv2 = − x2 + C3 x2 + C4 (c)
2 6
Figure 1.7: Problem 1.2.
The four constant of integration can be determined from:
Example 1.2 1. v1 = 0 at x1 = 0
2. v1 = 0 at x1 = 2a
GIVEN Consider the beam shown in figure 1.7.
3. v2 = 0 at x2 = a
FIND Determine the displacement at C of the beam.
4. dv1 /dx1 = −dv2 /dx2 at x1 = 2a and x2 = a.
Solution
Elastic curve. Due to the loading, two x coordinates will be considered, 0 ≤
Applying this four conditions gives:
x1 < 2a where x1 is directed to the right and 0 ≤ x2 < a where x2 is directed
to the left. from v1 = 0 at x1 = 0
Moment functions. From the free body diagrams P 3
EIv1 = − x + C 1 x1 + C 2
12 1
0 = 0 + 0 + C2 (d)
P
M 1 = − x1 from v1 = 0 at x1 = 2a
2
M2 = −P x2 (a) P 3
EIv1 = − x + C 1 x1 + C 2
12 1
P
0 = − (2a)3 + C1 (2a) + C2 (e)
Slope and elastic curve. 12
c
PSfrag replacements
x

1 2w0
from v2 = 0 at x2 = a w0 x x
2 L
2w0
P w= x
EIv2 = − x32 + C3 x2 + C4 L
6
P 3
0 = − a + C3 a + C4 (f)
6 M

x V

from dv1 /dx1 = −dv2 /dx2 at x1 = 2a and x2 = a x
L L w0 L x
dv1 dv2 4 3
(2a) = − (a) 2 2
dx1 dx
2 Figure 1.8: Problem 1.3.
P P
− (2a)2 + C1 = − − (a)2 + C3 (g)
4 2
Moment function. The distributed load has a positive sign according to the
signs convention established. From the free-body diagram the equation for the
Solving, the above equations distributed loading is
P a2 7 2w0
C1 = C2 = 0 C 3 = P a2 C4 = −P a3 (h) w= x (a)
3 6 L
Substituting C3 and C4 in the expression for v2 , Hence
P 3 7P a2 P a3
v2 = − x2 + x2 − (i) w 0 x2 x

w0 L
6EI 6EI EI ΣM(x) = 0; M+ − (x) = 0
L 3 4
The displacement at C is determined by setting x2 = 0, −w0 x3 w0 L
M= + x (b)
3L 4
P a3
vC = − (j) Slope and Elastic curve.
EI
Example 1.3 d2 v −w0 3 w0 L

EI =M = x + x
dx2 3L 4
dv −w0 4 w0 L 2
GIVEN Consider the beam shown in figure 1.8. EI = x + x + C1
dx 12L 8
FIND Determine the maximum deflection of the beam. w0 5 w0 L 3
EIv = − x + x + C1 x + C2 (c)
60L 24
Solution
Elastic curve. Due to symmetry, only one x coordinate is needed for this The constants of integration can be obtained from v(0) = 0 and the symmetry
solution, in this case 0 ≤ x ≤ L/2. condition dv/dx = 0 at x = L/2,
c
1.4 Procedure for analysis

5w0 L3
C1 = − C2 = 0 (d)
192
From the previous examples it is clear that some defined steps are required for
Thus, determining the slope and deflection of a beam using the method of integration.
These steps are outlined in what follows.
w0 5 w0 L 3 5w0 L3 Elastic curve:
EIv = − x + x − x (e)
60L 24 192
The maximum deflection at x = L/2 is therefore • Draw an exaggerated view of the elastic curve for the beam. Recall that
fixed supports give zero slope and zero vertical displacement and that
w0 L4 pin and roller supports give zero vertical displacement.
vmax = − (f)
120EI • Establish the x coordinate axes. The x axis or axes must be parallel
to the undeflected beam and can have an origin at any point along the
Solution II
beam, with a positive direction either to the right or to the left.
Starting with the distributed loading, equation (a), and applying EId4 v/dx4 =
−w(x), it is possible to write • If several discontinuous loads or supports are present, establish x coor-
dinates that are valid for each region of the beam between the disconti-
nuities. The right choice of these coordinates may simplify the analysis.
d4 v 2w0
EI =− x
dx4 L
3
dv w0 Distributed load or Moment function:
EI 3 = V = − x2 + C3 (g)
dx L
• For each region in which there is an x coordinate, express the distributed
Since V = +w0 L/4 at x = 0, then C3 = w0 L/4. Integrating again yields loading w or the internal moment M as a function of x. Always assume
a positive moment M .
d3 v w0 w0 L
EI = V = − x2 +
dx 3 L 4 Slope and elastic curve:
d2 v w0 3 w0 L
EI 2 = M = − x + x + C4 (h)
dx 3L 4
• Provided EI is constant, apply either the distributed load equation
EId4 v/dx4 = −w(x), which requires four integrations to get v = v(x),
Here M = 0 at x = 0 so C4 = 0. Hence,
or the moment equation EId2 v/dx2 = M (x) which requires only two
integrations. For each integration it is important to include a constant
d2 v w0 w0 L of integration.
EI = M = − x3 + x (i)
dx2 3L 4
• The constants are evaluated using the boundary conditions for the sup-
which is exactly the first of equations (c). From here the solution proceeds as ports and the continuity conditions that apply to slope and displacement
before. at points where two functions meet.
c
1.4 Procedure for analysis 19
• Once the constants are evaluated and substituted back into the slope
(dv/dx) and deflection (v) equations, the slope and deflection at specific
Deflection of beams by the
points on the elastic curve can be determined.
moment–area method
• The numerical values obtained can be checked graphically by comparing
them with the sketch of the elastic curve.
Siméon Denis Poisson (1781–1840)
Originally forced to study medicine, Siméon Poisson began to study mathe-

matics in 1798 at the Ecole Polytechnique. His teachers Laplace and Lagrange
were to become friends for life. A memoir on finite differences, written when
Poisson was 18, attracted the attention of Legendre.
Poisson taught at Ecole Polytechnique from 1802 until 1808 when he became
an astronomer at Bureau des Longitudes. In 1809 he was appointed to the
chair of pure mathematics in the newly opened Faculté des Sciences.
His most important works were a series of papers on definite integrals and his
advances in Fourier series. This work was the foundation of later work in this
area by Dirichlet and Riemann.
In Recherchés sur la probabilité des jugements, an important work on proba-

bility published in 1837, the Poisson distribution first appeared. The Poisson
distribution describes the probability that a random event will occur in a time
c
2.1 Introduction 21 22 Deflection of beams by the moment–area method
or space interval under the conditions that the probability of the event oc- • Demonstrate that in many cases the areas and the moments defined by
curring is very small, but the number of trials is very large so that the event the M/EI diagram can be determined very easily if the bending moment
actually occurs a few times. diagram is drawn by parts.
He published between 300 and 400 mathematical works including applications • Cantilever beams and beams with symmetric loading will be considered
to electricity and magnetism, and astronomy. His Traité de mécanique pub- here. Because the maximum deflection of a non-symmetric loaded beam
lished in 1811 and again in 1833 was the standard work on mechanics for many does not occur at its center, a method to find the point where the tangent
years. to the elastic curve is horizontal will be discussed as this point will have
the maximum deflection.
His name is attached to a wide area of ideas, for example: Poisson’s integral,
Poisson’s equation in potential theory, Poisson brackets in differential equa-
tions, Poisson’s ratio in elasticity, and Poisson’s constant in electricity.
2.2 Moment–area theorems
2.1 Introduction Consider the following problem in Figure 2.1 where the product EI known as
bending rigidity remains constant throughout the length of the beam.
Up to now the calculation of the deflection and slope of a beam at a given point Consider that a beam AB subject to a given arbitrary load. The M/EI
has been carried out using a mathematical method based on the integration diagram representing the variation of this quantity along the length of the
of a diferential equation. The moment of the beam was represented as a beam, is obtained dividing the bending moment M by the bending rigidity
function M (x) of the distance x measured along the beam and two succesive EI. It easy to see that, with the exception of a difference in the scales, this
integrations were used to obtain expressions for dv/dx = θ(x) and v(x) that diagram is equal to the bending moment one, if the bending rigidity is constant.
represent, respectively, the slope and the deflection at any point on the beam.
From the previous analyses, it is known that
In what follows, certain geometric relationships of the elastic curve will be used
to determine the deflection and slope of the beam at a given point. Instead d2 v M M (x)
= = (2.1)
of expressing the moment of the beam as a function M (x) and integrate this dx2 EI E(x)I(x)
function analytically, a sketch of the variation of M/EI along the beam will
be used. Then, the moment–area theorems will provide a way to obtain the dv
deflection and slope at a point on the beam. = tan θ ≈ θ (2.2)
dx
The first moment–area theorem allows the calculation of the angle between two
From the previous equations:
tangents to the elastic curve at two different points. The second moment–area
theorem will be used to calculate the vertical distance between two tangent
lines to the elastic curve at two different points.
d2 v d dv dθ
= ≈ (2.3)
The following points will be covered: dx 2 dx dx dx
dθ M
= (2.4)
• Use the theorems of the moment–area method to determine the slope dx EI
and the deflection at specific points in cantilever beams and beams with M
dθ = dx (2.5)
symmetric loading. EI
c
2.2 Moment–area theorems 23 24 Deflection of beams by the moment–area method
Considering two arbitrary points C and D along the beam. Integrating both
sides of the last equation between C and D it is possible to write
θD xD
P
M
Z Z
dθ = dx (2.6) L/2 L/2
θC xC EI
or xD
M
Z
θD − θC = dx (2.7)
xC EI A C D B
where θC and θD denote, respectively, the slope at C and the slope at D.

V (x)
P/2
L
x
PSfrag replacements
−P/2
M (x)
PL
4
x
L
M (x)
E(x)I(x)
PL
4EI
L
x
Figure 2.1: Shear, moment and M/EI diagrams.
c
P
frag replacements ds
dθ = (2.8)
ρ
A B
x x
where ds is the length of the arc pp0 and ρ is the radius of curvature at p.

θ(x) Remembering that,

Figure 2.2: Slope at a given point x. 1 M
= (2.9)
ρ EI
2.2.1 First moment–area theorem
and observing that the slope at p is very small, ds is approximately equal to
The right hand side of Equation 2.7 represents the area below the M/EI the horizontal distance dx between p and p0 as in previous analyses. Therefore,
diagram between points C and D. The left hand side represents the angle
between the straight lines tangent to the elastic curve at points C and D. M
dθ = dx (2.10)
EI
• θD/C = angle between the straight lines tangent to the elastic curves at
points C and D. which is the same relationship obtained before.
• θD/C is the angle between the line tangent to the elastic curve at D with
respect to the line tangent to the elastic curve at C. 2.2.2 Second moment–area theorem
First moment–area theorem: Now, consider two points p and p0 between points C and D that are separated
by a distance dx. The tangents lines to the elastic curve at p and p0 intersect
a segment of length dt of the vertical that passes through point C (see Figure
θD/C = area below the M/EI diagram between C and D. 2.5).
As the slope θ at p and the angle dθ between the tangent lines at p and p0 are
It is convenient to point out that the angle θD/C and the area between the small quantities, it is safe to assume that dt is equal to the arc of a circle with
diagram M/EI have the same sign. In other words, a positive area correspond radius x that sweeps the angle dθ.
to a counterclockwise rotation of the line tangent to the elastic curve a the
point moves from C to D. A negative area correspond to a clockwise rotation. Therefore,
The first moment–area theorem can be established also from the following dt = x1 dθ
analysis considering Figure 2.4.
ds dx M
dθ = = = dx
The angle dθ between the two tangent lines to the elastic curve at points p and ρ ρ EI
p0 is also the angle between the respective normal lines to that curve. From M
dt = x1 dx (2.11)
Figure 2.4 EI
c
Now it is possible to integrate the last of the previuos equations from C to D Care should be taken to distinguish between the tangential deviation of C with
to obtain, respect to D, denoted by tC/D , and the tangential deviation of D with respect
to C which is denoted by tD/C (see Figure 2.6).
tC xD
M
Z Z
dt = x1 dx
tD xC EI
xD
M
Z
tC − tD = tC/D = x1 dx (2.12)
xC EI
Noting that point p can move between C and D to describe the elastic curve
between these two points, the tangent line at p sweeps the vertical between C
and E.
The left hand side of Equations (2.12) is equal to the vertical distance between
C and the tangent line at D. This distance is denoted by tC/D and receives
the name of tangential deviation of C with respect to D.
• tC/D is the tangential deviation of C with respect to D.

• tC/D is the vertical distance from C to the tangent line to the elastic
curve at D.
Now it is possible to observe that (M/EI)dx represents an area below the

M/EI diagram and x1 (M/EI)dx represents the first moment of that area
with respect to a vertical axis passing through C. Therefore, the right hand
side of Equation (2.12) represents the first moment with respect to that axis
at C of the area below the M/EI diagram between C and D.
From the previous analysis, the tangential deviation tC/D of C with respect to
D is equal to the first moment of the area below the M/EI diagram between
C and D with respect to a vertical axis at C.
Remembering that the first moment of an area with respect to an axis is equal
to the product of the area times the distance between its centroid and that
axis, it is possible to express the second moment–area theorem as:
Second moment–area theorem:
tC/D = area below the M/EI diagram between C and D × x¯1 .
c
c
M (x)
E(x)I(x)
PL ρ
4EI PSfrag replacements dθ
ρ
p0
dθ
L
x ds
p
xC xD
xC Figure 2.4: Angle between two tangent lines to the elastic curve.
x
eplacements θC

PSfrag replacements
xD
x
P xC xp x0p xD

θD x
x1

xC xD C D

θD/C

dt
dθ
E
Figure 2.3: Graphical description of the first moment–area theorem.
Figure 2.5: Vertical distance dt defined by two tangent lines to the elastic
curve.
c
M (x) The tangential deviation tD/C represents the vertical distance from D up to
E(x)I(x) the tangent line to the elastic curve at C. tD/C is obtained multiplying the area
PL below the M/EI diagram by the distance x¯2 from its centroid to the vertical
4EI axis passing through D
eplacements
tD/C = area below the M/EI diagram between C and D × x¯2 .

L
x
C x¯1 D
A point with a positive tangential deviation is located above of the respective
x¯2 tangent line. A point with a negative tangential deviation is located below the
respective tangent line.
x
C
D 2.3 Application to simple beams

0 tC/D
D
Remember that the first moment–area theorem defines the angle θD/C between
tD/C C0
M (x) the tangent lines to two points C and D of the elastic curve. Therefore, the
E(x)I(x) angle θD that the tangent in D has with the horizontal can only be obtained
PL if the slope at C is known.
eplacements 4EI Similarly, the second moment–area theorem defines the vertical distance of one
point of the elastic curve with respect to the tangent line at another point.
Therefore, the tangential deviation tD/C is useful to find the deflection at point
L D only if the tangent at C is known.
x
x¯1 C x¯2 D
The two moment–area theorems can only be applied effectively for
the determination of the slopes and deflections if and only if certain
reference tangent in the beam is known.
x
C 0
C D

tD/C 2.3.1 Cantilever beam

tC/D
D0 In the case of a cantilever beam, the tangent to the elastic curve at the support
(point A) is known and can be used as the reference tangent (see Figure 2.7).
Figure 2.6: Graphical definition of tC/D and tD/C . As θA = 0 the slope at any point D on the beam is θD = θD/A and can be
obtained from the first moment–area theorem.
c
2.3 Application to simple beams 33 34 Deflection of beams by the moment–area method
On the other side, the deflection vD at any point D on the beam is equal to
P
the tangential deviation tD/A measured from the horizontal reference tangent
and, therefore, can be obtained from the second moment–area theorem. D

A

PSfrag replacements
vD

A θref

θD

Figure 2.7: Cantilever beam and its tangent of reference.
P P
A C B

tB/C
θB/C

PSfrag replacements θref = θC
P P
vD A C D B
θD/C tD/C

θref = θC
Figure 2.8: Beam with symmetric loading.
c
2.3.2 Beam with symmetric loading
In the case of a simply supported beam AB subject to a symmetric loading, or B

A
in the case of a symmetric cantilever beam with a symmetric load, the tangent

at point C located at the center of the beam must be horizontal by symmetry

and can be used as the reference tangent.
As θC = 0, the slope at the support B es θB/C and can be obtained from the P
first moment–area theorem. Furthermore, it can be noted that vmax is equal to
the tangential deviation tB/C and, therefore, can be obtained from the second A
moment–area theorem. The slope at any other point D of the beam can be PSfrag replacements M RA = P
obtained in a similar fashion and the deflection at D can be obtained as A RA
MA = P L
vD = tB/C − tD/C (2.13)

B

A

vB
Example 2.1

GIVEN Consider a beam in cantilever like the one shown in figure 2.9.
FIND Obtain the maximum deflection of the beam using the moment–area Figure 2.9: Example 2.1.
method.
Solution
From Figure 2.9, the maximum deflection of the beam occurs at the free end
of the beam x = B. Therefore, vmax = vB . It is possible to draw the M/EI
diagram from the equation of the moment for the beam M (x) = −P (L − x)
as shown in Figure 2.10.
From the M/EI diagram, vB = tB/A that is, the deflection at B is equal to the
vertical distance from point B to the tangent of reference θA . It is important
to remember that the analysis is done on the deformed beam.
From the second moment–area theorem, tB /A is equal to the area between

points A and B multiplied by the distance from the centroid of such area to
point B (see Figure 2.11). From the M/EI diagram, the area between points
A and B is:
P L2

PL 1
− L = − (a)
EI 2 2EI
c
M (x)
EI
Example 2.2
L
x
GIVEN Consider the beam shown in Figure 2.12.
PSfrag replacements
FIND Determine the maximum deflection of the beam and the slopes at the
supports using the moment–area method.
Solution
PL
− From this figure, the maximum deflection of the beam will occur at point C
EI from where the unknown to be found is vC . The slopes at the supports are
simply the slopes at points A and B.
Figure 2.10: M/EI diagram for the cantilever beam.
Knowing that the deflection of the beam is symmetric, the tangent of reference
can be chosen as θC as θC = 0. Using this reference, the deflection at point C
M (x) can be obtained as tB/C , the vertical distance from point B to the tangent at
EI point C. On the other hand, the slopes can be obtained as θB/C and θC/A (see
L
PSfrag replacements x Figure 2.13).
Figure 2.14 represents the free-body and M/EI diagrams for the problem.
From this figure and the moment–area theorems the following relations can be
written
2L
1 L PL

PL θC/A = A1 =
− 3 2 2 4EI
EI
P L2
=− (a)
16EI
Figure 2.11: Area between A and B in the M/EI diagram and distance from
the centroid of the area to point B.
Therefore,
P L2 PSfrag replacements

2L P
tB/A = −
2EI 3 x
P L3
=− (b) A B
3EI
C
L/2 L/2

from where
P L3
vB = − (c) Figure 2.12: Problem 2.2
3EI
c
P
c
x
tB/C

P P

θref = θC 2 2
PSfrag replacements
x M (x)
EI
g replacements PL

θD/C
4EI
A
B
C x A1 A2 L

x

L L
θC/A 6 6
L
2
Figure 2.13: Graphics representation of θC , θC/A , θB/C and tB/C for the problem
under consideration. Figure 2.14: Free body and M/EI diagrams for the problem under consider-
ation.
c
To facilitate the measuring of the area between points C and E, it is recom-

P L2 mended to divide the total area A in two subareas A1 and A2 as shown in
θB/C = A2 = − (b)
16EI Figure 2.17. The first area A1 goes from point C to point D and the second
one, A2 goes from point D to point E. The first area does not represent any
problems but the second area has a non–standard shape which correspond to
P L2 a general spandrel y = kxn . The formulae to compute its area and centroid
θA = −θC/A = −
16EI are shown in Figure 2.18.
P L2
θB = (c) From the first theorem of the moment–area method,
16EI
θE = θE/C = A1 + A2 (a)
vC = −tA/C = −tB/C where

wa2

L L

L
tB/C = A2 − A1 = −
2 6 2EI 2
P L2 P L3 wa2

L

1
tB/C = = A2 = − a (b)
16EI 3 48EI 2EI 3
P L3
vC = −tB/C = − (d)
48EI
Example 2.3
GIVEN Consider the supported beam shown in Figure 2.15 loaded at its end
by means of a distributed load.
FIND Find the slope and deflection at E using the moment–area theorems.
Solution w w
PSfrag replacements
The first step to solve the problem is to draw the M/EI diagram for the beam
being analysed. Figure 2.16 shows the free body and M/EI diagrams for the B C D
beam under consideration.
A E
Now, denote the slope at point E as θE and the deflection at the same point
as vE . The deflection θE can be obtained without much problem measuring

the area below the M/EI diagram between points E and C since the angle at

point C is zero. The deflection vE cannot be obtained directly as the tangent
L L
of reference is at a certain deflection vC which is not known. Fortunately, vE a 2 2 a
can be obtained through the distances tD/C and tE/C that can be computed
using the moment–area method such that vE = tE/C − tD/C (see Figure 2.16). Figure 2.15: Problem 2.3.
c
θC = 0
D
E

θC = 0
Rw Rw B D θE/C
g replacements A E
A B C D E C

PSfrag replacements
RB RD
θC = 0
M (x) tD/C
B
EI A E
a L+a L + 2a C tE/C

PSfragx replacements

Figure 2.17: Key quantities for the problem under consideration.
wa2 wa2
− − a L/2 L/2 a
2EI 2EI
Figure 2.16: Free–body and M/EI diagram for the beam under consideration. M (x) a/4
EI L/4
A B C D E
x
A2
wa2 wa2
− A1 −
2EI 2EI
Figure 2.18: Areas and centroids.
c
Therefore, bh
wa2 PSfrag replacements Area =
θE = − (3L + 2a) (c) n+1
12EI y = kxn
h
As mentioned, in order to compute vE it is necessary to compute first tD/C and b
c=
tE/C . The tangential deviation tD/C can be easily computed as the product of n+2
A1 by L/4, distance between the centroid of the area A1 to point D
wa2 L b

L
tD/C = −
4EI 4 c
2 2
wa L
tD/C = − (d) Figure 2.19: Calculation of the area and centroid of a general spandrel.
16EI
The tangential deviation tE/C is obtained taking into account two products. Nevertheless, when a simple supported beam has a non–symmetric load, it is
First, the product of the area A1 by the distance from its centroid to point E. not possible to find by inspection the point where the tangent to the elastic
Second, the product of the area A2 by the distance from its centroid to the curve will be horizontal. In these cases, another tangent of reference should
same point E. Therefore, be used. As before, this tangent of reference must has known slope that can
be used to apply any of the two moment–area theorems.
L 3a
tE/C = A1 − + a + A2
4 4
2 2 3
Usually, the most convenient choice is to select as tangent of refer-
wa L wa L ence the slope at one of the supports of the beam.
tE/C = − −
16EI 4EI
wa4
tE/C = − (e) Consider for example the tangent at the support A of a simple supported beam
8EI
AB as shown in Figure 2.20. It is possible to find the slope at point A by means
of the tangential deviation tA/B of the support B with respect to A. To do
Remembering that yE = tE/C − tD/C so, it is necessary only to divide tB/A with respect to the distance between
supports L as shown in Figure 2.21.
wa3 L wa4
yE = − −
4EI 8EI Remembering that the tangential deviation of a point located above the tan-
wa3 gent of reference is positive, it is possible to write
y3 = − (2L + a) (f)
8EI −tB/A
θA = (2.14)
L
2.3.3 Beam with non–symmetric loading Once the slope of tangent of referente at A is known, the slope of the beam at
any point D, θD/A , can be obtained through the first moment–area theorem
Up to now, it has been shown that cantilever beams and simple supported since
beams with symmetrical loading have a tangent line that is horizontal. This θD = θA + θD/A (2.15)
horizontal line can be used as tangent of reference in a very convenient manner. as shown in Figure 2.22.
c
Doing a summation of moments with respect to point A, Mo Mo

X
MA = 0 A B
−M o + M o + RB L = 0 C D
RB = 0 (b)

L/4 L/4 L/2
Next, it is necessary to choose a tangent of reference from which PSfrag replacements
it is possible
to known its slope. In this regard, the slope at the support A can be obtained
as A C0 D0 B
tB/A
θA = − (c)
L
and the deflection vD can be computed as the difference between the distances C

D

vD = ED − ED0 . The distance ED 0 can be defined by means of similar
triangles noting that (see Figure 2.26)
tD/A
ED0 tB/A tB/A

= (d) E
L/2 L
x θA
and therefore, θref
tB/A
vD = tD/A − (e)
2
From the previous analysis it is only necessary to find out the values for tB/A
and tD/A to solve the problem. Using the moment–area method and the M/EI
Figure 2.26: Quantities involved in the solution of the problem under consid-
diagram shown in Figure 2.27, tB/A is obtained multiplying the area below the
eration using the moment–area method.
M/EI diagram between points B and A by the distance from the centroid of
that area to point B

L L 5L
tB/A =A + =A (f)
2 8 8
where tB/A 5M oL

L Mo

M oL θA = − =− (i)
A= = (g) L 32EI
4 EI 4EI
and therefore,
5M oL2
tB/A = (h) The value for tD/A can be obtained multiplying the area below the M/EI
32EI
diagram between points D and A (note that the area has the same value as
From the previous result, the value of θA is before) by the distance from the centroid of that area to point D. Thus,
c
2.3 Application to simple beams 53
eplacements M (x)
EI
Method of superposition
Mo
EI
A
x
L L L
4 2
L
8
Figure 2.27: M/EI diagram for the beam under consideration.

L M oL L
tD/A =A =
8 4EI 8
M oL2 Adhémar Jean Claude Barré de Saint-Venant (1797–1886)
tD/A = (j)
32EI
Jean Claude Saint-Venant was a student at the Ecole Polytechnique, entering
which finally yields, the school in 1813 when he was sixteen years old. He graduated in 1816 and
spent the next 27 years as a civil engineer. For the first seven of these 27 years
Saint-Venant worked for the Service des Poudres et Salpêtres, then he spent
1 the next twenty years working for the Service des Ponts et Chaussées.
vD = tD/A − tB/A
2
He taught mathematics at the Ecole des Ponts et Chaussées where he succeeded
M oL2 5M oL2 3M oL2
vD = − =− (k) Coriolis. Saint-Venant worked mainly on mechanics, elasticity, hydrostatics
32EI 64EI 64EI
and hydrodynamics. Perhaps his most remarkable work was that which he
published in 1843 in which he gave the correct derivation of the Navier-Stokes
equations. We should remark that Stokes, like Saint-Venant, correctly derived
the Navier-Stokes equations but he published the results two years after Saint-
Venant.
In the 1850s Saint-Venant derived solutions for the torsion of non-circular

cylinders. He extended Navier’s work on the bending of beams, publishing a
full account in 1864. In 1871 he derived the equations for non-steady flow in
open channels.
c
3.1 Introduction 55 56 Method of superposition
In 1868 Saint-Venant was elected to succeed Poncelet in the mechanics section P1 P2

of the Académie des Sciences. By this time he was 71 years old, but he con- θ(x)

tinued his research and lived for a further 18 years after this time. Barré de

v(x)
Saint-Venant is generally recognized as the most outstanding elastician of all

a b

time. PSfrag replacements
P1

3.1 Introduction θ1 (x)

v1 (x)

The differential equations representing the deflection curve of a beam,

P2
dv 4
EI = −w(x) θ2 (x)
dx4

dv 3 v2 (x)

EI 3 = V (x)

dx

dv 2
EI 2 = M (x) (3.1) Figure 3.1: Cantilever beam subject to two different loads.
dx
are linear differential equations, that is, all the terms that include the deflec- Therefore,
tion v and its derivatives are raised to the power of one only. Therefore, the
solutions for different loading conditions can be superposed. Then, dv 4
EI = w(x) = w1 (x) + w2 (x)
dx4
the deflection of the beam caused by various loads that act v(x) = v1 (x) + v2 (x) (3.5)
simultaneously can be obtained through the superposition
of the deflections caused by each load acting separately. Consider the problem described in Figure 3.1.
For example, if v1 represents the deflection due to a load w1 and if v2 represents In this case,
the deflection due to a load w2 , the total deflection caused by w1 and w2 acting v(x) = v1 (x) + v2 (x) (3.6)
simultaneously es v1 + v2 . where
If,
dv14 • v1 (x) is the deflection due to the load P1 acting independently on the
EI = w1 (x) (3.2)
dx4 beam.
and
dv24 • v2 (x) is the deflection due to the load P2 acting independently on the
EI = w2 (x) (3.3)
dx4 beam.
then,
d 4 dv14 dv24 • v(x) is the deflection due to the loads P1 and P2 acting simultaneously
4
EI (v + v ) = EI + EI = w1 (x) + w2 (x) (3.4) on the beam.
dx4 1 2
dx4 dx4
c
3.2 Principle of superposition 57 58 Method of superposition
At this point it is important to notice that, consider also that w2 (x) = aw1 (x) where a is a constant. If the deflection v2
wants to be known, it is possible to solve instead Equation (3.8) multiplying
v(x) = v1 (x) + v2 (x) both sides by a
dv dv1 dv2 dv 4
(x) = (x) + (x) = θ1 (x) + θ2 (x) (3.7) EIa 14 = −aw1 (x) = −w2 (x) (3.10)
dx dx dx dx
from where y2 (x) = ay1 (x).
The superposition method is an easy way to solve beam which loading con-
ditions can be decomposed on loads that produce deflections that are already
known. Therefore, it is convenient to have at hand tables where the solution 3.4 Proportionality and superposition
for deflections and slopes of beams under conventional loads are listed.
The principles of proportionality and superposition can be combined in order
to solve even more complex sets of loads. Consider for example that the
3.2 Principle of superposition deflection v(x) of a given beam that is subject to a complex load set should be
found. Furthermore, consider that the load set can be decomposed into two
different loading conditions in such a way that
The efect of a given set of loads that are acting simultaneously on a structure
can be obtained considering the efect of each one of the loads acting separately w(x) = aw1 (x) + bw2 (x) (3.11)
and adding the results together.
where a and b are constants.
To be able to apply the principle of superposition two conditions should
Knowing that
be met:
dv14
1. Each efect should be related linearly to the load that produces it. EI = −w1 (x) (3.12)
dx4
2. The deformation that results from any load is small and do not affect and
dv24
the loading conditions of the other loads. EI = −w2 (x) (3.13)
dx4
it is possible to write
dv14 dv 4
3.3 Principle of proportionality a(EI ) + b(EI 2 ) = a(−w1 (x)) + b(−w2(x))
dx dx
= −(aw1 (x) + bw2 (x))
If the efect of each load is linearly related to the deformation that produces = −w(x) (3.14)
it, another very convenient relationship can be found. Consider the following
two equations describing two different deflections caused by two different loads Using the proportionality principle,
acting on the same beam
dv 4 dv14 dv 4
EI 14 = −w1 (x) (3.8) EI(a + b 2 ) = −w(x) (3.15)
dx dx dx
and therefore,
dv 4
EI 24 = −w2 (x) (3.9) v(x) = av1 (x) + bv2 (x) (3.16)
dx
c
3.4 Proportionality and superposition 59 60 Method of superposition
Example 3.1 P1 = P P2 = P

GIVEN Consider a cantilever beam subject to two forces P1 and P2 of mag-

nitude P acting on it as shown in Figure 3.2. L/2 L/2

FIND Calculate the functions that describe the deflection and slope along
the beam.
P1 = P
Solution
The problem under consideration can be decomposed in two more simple prob-

lems as shown in Figure 3.3 in such a way that

L/2 L/2

v(x) = v1 (x) + v2 (x) (a)
and
θ(x) = θ1 (x) + θ2 (x) PSfrag replacements
(b) P2 = P

From tables, the solution of the second subproblem is the easiest one since the

solutions L

P x2
v2 (x) = − (3L − x) (c)
6EI
Px Figure 3.3: Descomposition of the beam into two different problems.
θ2 (x) = − (2L − x) (d)
2EI
are valid along all the length of the beam. Px
θ1 (x) = − (L − x) (f)
2EI
According to the tables, the solution for the first subproblem has to be divided
in two regions, 0 ≤ x ≤ L/2 and L/2 ≤ x ≤ L. For the second region L/2 ≤ x ≤ L
For the region 0 ≤ x ≤ L/2
P L2

L
2
v1 (x) = − 3x − (g)
Px 3L 24EI 2
v1 (x) = − −x (e)
6EI 2
P L2
θ1 (x) = − (h)
P1 = P P2 = P 8EI

From equations (a) and (b), for 0 ≤ x ≤ L/2

PSfrag replacements

P x2 9L

L/2 L/2 v(x) = − − 2x (i)

6EI 2
Px
Figure 3.2: Example 3.1. θ(x) = − (3L − 2x) (j)
2EI
c
3.5 Statically indeterminate beams 61 62 Method of superposition
and for L/2 ≤ x ≤ L PSfrag replacements w
P x2 P L2

L
v(x) = − (3L − x) − 3x − (k)

6EI 24EI 2

MA B

Px P L2 A

θ(x) = − (L − x) − (l)

2EI 8EI RA RB
L
3.5 Statically indeterminate beams Figure 3.4: Example 3.2.
Consider now beams that have more reactions than the ones that can be ob- If RB is selected as the redundant reaction, the primary structure becomes a
tained from the equations of static equilibrium. These beam are known as cantilever beam as shown in Figure 3.5.
statically indeterminate beams.
Now, it is possible to have three equations for the three reactions:
The number of reactions that exceed those that can be found from the equa-
tions of equilibrium is known as degree of indeterminacy. The additional sup- P
port reactions on the beam or shaft that are not needed to keep it in stable • Fy = 0 from statics.
equilibrium are called redundant. The number of additional support reactions •
P
MA = 0 from statics.
is necessarily equal to the degree of indeterminacy.
• vB = 0 from beam deflection.
The statically determinate beam that is obtained from taking away the redun-
dat supports is called primary structure. P
From Fy = 0
RA + RB − wL = 0
Example 3.2 RA + RB = wL (a)
GIVEN Consider for example the indeterminate beam presented in Figure P

3.4. From MA = 0
FIND The values of moments and reactions. wL2

− + M A + RB L = 0
2
Solution wL2
From the figure it is clear that there are M A + RB L = (b)
2
• 3 reactions: RA , RB and MA for only From vB = 0 and by superposition

P P
• 2 equations: Fy = 0 and MA = 0 vB = v1B + v2B = 0
v1B = −v2B
Therefore, the beam is statically indeterminate in the first degree. v1 (L) = −v2 (L) (c)
c
w From Equation (b)
wL2

MA = − RB L

MA B 2

A wL2 3 2

MA = − wL

RB 2 8
RA (vB = 0) wL2
PSfrag replacements MA = (i)
w 8
L
This is not the only way to solve this problem. For example, it is possible to

v1 (x)

B choose the moment of reaction MA as redundant to have as primary structure

A a simple supported beam as shown in Figure 3.6.

PSfrag replacements

L w

v2 (x) B

A MA A B

RB

Figure 3.5: Decomposition of the indeterminate beam.
RA RB
From tables: Figure 3.6: Another possible solution to Problem 3.2.
wL4
v1 (L) = − (d)
8EI
RB L 3 Example 3.3
v2 (L) = (e)
3EI
GIVEN Consider now the indeterminate beam presented in Figure 3.7.
From Equation (c)
wL4 RB L 3 FIND The values of the moments and reactions.
− =− (f)
8EI 3EI
Solution
solving for RB
3 From the figure
RB = wL (g)
8
• 4 reactions: RA , RB , MA and MB for only
From Equation (a)
P P
3 • 2 equations: Fy = 0 and MA = 0
RA = wL − RB = wL − wL
8
5
RA = wL (h)
8 Therefore, the beam is statically indeterminate in the second degree.
c
3.5 Statically indeterminate beams PSfrag
65 replacements
66 Method of superposition
L
One possible way to solve it is to choose MA and MB as the redundant reactions wo

in such a way that the primary structure becomes a simply supported beam
as shown in Figure 3.8. Thus, the problem can be decomposed in three simply
supported beams as shown in Figure 3.9.
MA A B MB
From Tables,
θA = 0 θB = 0
7woL3 wo L3

θ1A = − θ1B = (a)
360EI 45EI RA RB
MA L MA L
θ2A = θ2B =− (b) Figure 3.8: Simply supported beam with triangular distributed load.
3EI 6EI
MB L MB L
θ3A = θ3B = − (c) Substituting now the expressions for θ1B , θ2B and θ3B in Equation (e)
6EI 3EI
wo L3 MA L MB L
From the original problem, at the wall supports θ = 0. Therefore, − − =0 (i)
45EI 6EI 3EI
θA = θ1A + θ2A + θ3A = 0 (d) 6wo L2 − 45MA − 90MB = 0 (j)
2
45MA + 90MB = 6wo L (k)
θB = θ1B + θ2B + θ3B = 0 (e)
Solving simultaneously equations (h) and (k)
Substituting the expressions for θ1A , θ2A and θ3A in Equation (d)
wo L2
7woL3 MA L MB L MB = (l)
− + + =0 (f) 20
360EI 3EI 6EI
wo L2
2 MA = (m)
−7wo L + 120MA + 60MB = 0 (g) 30
120MA + 60MB = 7wo L2 (h) A and MB are known, it is possible to obtain RB and RA
Once the values of MP
from statics. Doing MA = 0
frag replacements
wo wo L2
RB L − M B + M A − =0 (n)
3
A Isolating the term RB L

B

MA MB wo L2

RB L = + MB − MA (o)
3

L

Substituing equations (l) and (m) into (o)

RA RB
wo L2 wo L2 wo L2
Figure 3.7: Example 3.3. RB L = + − (p)
3 20 30
c
w
W0 PSfrag replacements
A C
B
k

PSfrag replacements θ1 (x)

L L
L
wo
MA Figure 3.10: Example 3.4.
A B
RA θ2 (x) FIND Determine the constant of the spring k for which the bending moment
RB

at B is: MB = −wL2 /12.
Solution
MB To solve this problem it is possible to substitute the spring by its equivalent
A B
force F = −kvB where vB is the displacement of the beam at point B. Thus,
θ3 (x) the beam can be divided into two subproblems as shown in Figure 3.11.

Figure 3.9: Decomposition of the problem under consideration. From tables,

5w(2L)4 5wL4
v1B = − =− (a)
384EI 24EI
Simplifying the above equation
F (2L)3 F L3
7wo L v2B = = (b)
RB = (q) 48EI 6EI
20
From the problem, vB = v1B + v2B . Therefore,
P
From Fy = 0 5wL4 F L3
wo L vB = − + (c)
RA + R B − =0 (r) 24EI 6EI
2
Substituing the value of RB in the above expression L3
vB = (4F − 5wL) (d)
wo L 7wo L 24EI
RA = − (s)
2 20 From
P
Fy = 0
3wo L RA + RC + F − wL = 0 (e)
RA = (t)
20 P
From MC = 0
−RA (2L) − F (L) + w(2L)(L) = 0 (f)
Example 3.4
from where,
F
GIVEN Consider the beam and loading conditions shown in Figure 3.10. RA = wL − (g)
2
c
w

L
MB = RA L − wL (h)
2
A B C
Substituing Equation (g) into the above equation

F = −kvB
RA RC
PSfrag replacements F L wL2 wL2 − F L
MB = wL2 − − = (i)
L w 2 2 2
v1 (x) Since from the problem statement MB = −wL2 /12, it is possible to write
A B C
wL2 − F L −wL2

= (j)
2 12
B
v2 (x) from where
A C 7wL

F = (j)
F = −kvB

6
Figure 3.11: Decomposition of the beam under consideration into two sub- Substituing Equation (c) into Equation (j)
problems.
L3

28wL
vB = − 5wL (k)
24EI 6
Of course, the above result could have been obtained directly from the sym-
wL4
metry of the problem. vB = − (l)
72EI
In order to obtain the value of the constant k when MB = −wL2 /12, it is
necessary to obtain an expression for MB . From Figure 3.12, Once the values for F and vB are know, the last step is to obtain k from
F = −kvB
F
k=− (m)
vB
w
PSfrag replacements 7wL
504wLEI
MB k=− 6 4 = (n)
wL 6wL4
−
VB 72EI
RA
L
Finally,
84EI
Figure 3.12: Free–body diagram to obtain MB . k= (o)
L3
c
Example 3.5 v(x)

Mo
A C B
GIVEN Consider the beam shown in Figure 3.13.

RC
PSfrag replacements RA (vC = 0) RB
FIND Find the values for the reactions RA , RB y RC .
Solution
From the problem v1 (x)
P there are three
P unknown reactions RA , RB and RC for only
two equations Fy = 0 and MC = 0. Therefore, the beam is indetermi- Mo
nated in the first degree. A C B

To solve the problem, it is convenient to choose the value RC as the static
redundant to have a simple supported beam as primary structure as shown v2 (x)
in Figure 3.14. At this point is important to mention that the support at C
restrict the movement in the x direction.PNevertheless, the reaction force in

the horizontal direction HC is zero from Fx = 0.

RC
From Figure 3.14: Decomposition of the beam with three supports.
v(x) = v1 (x) + v2 (x) (a)
the deflection at point C is given by From Equation (b)

M oL2 RC L3
vC = + =0 (e)
vC = v1C + v2C (b) 16EI 48EI
From the previous equation the reaction RC can be found

From tables,
3M o
Mo L2 RC = − (f)
v1C = (c) L
16EI
RC L 3 An additional equations is obtained applying summation of forces equal to

PSfrag replacements v2C = (d)
48EI zero in the y direction
RA + R B + R C = 0 (g)
L/2 L/2 from where
3M o
A C B RA + RB = −RC = (h)
L
Mo
The last equation needed to obtain the values for the reactions RA and RB

P
can be found from MC = 0
RA RC RB
L L
Figure 3.13: Example 3.5. −RA + RB − M o = 0 (i)
2 2
c
3.5 Statically indeterminate beams PSfrag replacements
73 74 Method of superposition
Leaving the unknowns in the LHS only A

L

MA B

2M o

−RA + RB = (j)

L RB
RA
(vB = −∆)
The equations (h) and (j) can be solved simultaneoulsy to find the values of
RA and RB Figure 3.16: Representation of the deflected cantilever beam.
5M o
RB = (k)
2L For the analysis, it is convenient to consider the beam as presented in Figure
Mo 3.16.
RA = (l)
2L
From tables the deflection at B is
RB L 3
Example 3.6 vB = (a)
3EI
GIVEN Consider now a cantilever beam AB that is fixed to a wall in A. but the deflection vB is equal to −∆
Furthermore, consider that a traction at B deflects the beam down and that
RB L 3
the beam is maintained in that position by means of a cable as shown in Figure = −∆ (b)
3.15. 3EI
FIND Obtain the values for the reactions RA , RB and MA in terms of the Solving for RB
deflection ∆. 3EI∆
RB = − (c)
L3
Solution
where the negative sign indicates that RB is acting on the opposite sign as it
From Figure 3.15 it is clear that the reaction RB is acting in the opposite
was supposed.
direction as it is shown. Nevertheless, the procedure for the analysis will take
care of that as it will be shown.
P
The reaction RA can be found from Fy = 0
PSfrag replacements RA + R B = 0 (d)

L
A B x 3EI∆
RA = (e)

MA
L3

∆

P
Finally, the value of MA is obtained from MA = 0
RA
M A + RB L = 0 (f)
RB
3EI∆
MA = (g)
L2
Figure 3.15: Example 3.6.
c
76 Continuous beams
Continuous beams No portrait of Hooke is known to exist. A possible reason for this is that he
has been described as a “lean, bent and ugly man” and so he may not have
been willing to sit for a painting of his portrait.
Robert Hooke (1635–1703) 4.1 Introduction

Robert Hooke went to school in Westminster where he learnt Latin and Greek, The beams that are continuous over many supports are known as continuous
but unlike his contemporaries he never wrote in Latin. In 1653 he went to beams and are found typically in buildings, pipes, bridges and other classes
Christ College, Oxford where he won a chorister’s place. of specialized structures.
At Oxford Hooke met Boyle and in 1655 he was employed by Boyle to construct If the loads that are acting on a continuous beam are vertical and there are
his air pump. In 1660 he discovered Hooke’s law of elasticity. Hooke worked not axial deformations, then all the reactions will be vertical. To represent
on optics, simple harmonic motion and stress in stretched strings. For 30 years this behaviour schematically, it is possible to consider that one of the supports
he was professor of geometry at Gresham College, London, being appointed is a pin support and all other are roller supports as shown in figure 4.1. Since
there in 1665. the number of reactions is the same as the number of supports, the degree
The year 1665 was the one when Hooke first achieved worldwide scientific of indeterminacy will be the number of supports minus two. Thus, the beam
shown in figure 4.1 has five vertical reactions, of which three are redundant.
fame. His book Micrographia, published that year, contains beautiful pictures
of objects Hooke had studied through a microscope he had made himself. The From these suppositions,
book also contains a number of fundamental biological discoveries.
Hooke invented the conical pendulum and was the first person to build a • N is the number of supports (of which one is a pin support and the
Gregorian reflecting telescope. He made important astronomical observations others are roller supports).
including the fact that Jupiter revolves on its axis and his drawings of Mars
were later used to determine its period of rotation. In 1666 he proposed that • N is the number of reactions.
gravity could be measured using a pendulum.
• The number of static redundants is N − 2.
In addition to his post as professor of geometry at Gresham College, London
Hooke held the post of City Surveyor. He was a very competent architect and • The degree of indeterminacy is N − 2.
was chief assistant to Wren in his project to rebuild London after the Great
Fire of 1666. • All the reactions are vertical: there are not axial deformations.
In 1672 Hooke attempted to prove that the Earth moves in an ellipse round
the Sun and six years later proposed that inverse square law of gravitation Although it is possible to analyse the beam in a continuous fashion using
to explain planetary motions. Hooke wrote to Newton in 1679 asking for his any method for hyperstatic beams, the only practical way to solve them is
opinion. Hooke seemed unable to give a mathematical proof of his conjectures. using superposition. When this method is used, one possibility is to select as
However he claimed priority over the inverse square law and this led to a bitter redundant the reactions at the intermediate supports which results in having
dispute with Newton who, as a consequence, removed all references to Hooke as released structure a simple beam. This technique is satisfactory for beams
from the Principia. with only one or two spans.
c
4.1 Introduction 77 78 Continuous beams
When two or more redundants are present in the beam, is usually better to • Compatibility condition: each of the adjacent beams must have the same
select as redundant the bending moments at the intermediate supports. This angle of rotation. This condition provides the necessary extra equations
selection simplifies the calculations greatly since it leads to a system of si- needed to solve for the unknown bending moments.
multaneous equations where there are no more than three unknowns on each
equation, no matter the number of redundants.
When the bending moments are removed from the beam, the continuity of the 4.2 Method of three moments
beam is broken at the supports and the released structure consist in a serie
of simple beams. Each of those beam is subject to external forces that act Consider for example the segment of beam shown in figure 4.2. The three
on it in the normal direction together with the redundant moments acting at consecutive supports are identified as A, B, C and the lengths and moments of
the ends of the simple beam. Under the action of these forces, it is possible inertia of the two adjacent spans are denoted by La , Ia , Lb and Ib respectively.
to find the angle of rotation at the end of each simple beam. The equations Let MA , MB and MC the bending moments on the three supports. The true
of compatibility, that express the fact that at each support the two adjacent direction of these moments will depend on the loads acting on the beam,
beams in the original structure have the same angle of rotation, provide the but for the purpose of this derivation they will be supposed positive (causing
necessary equations to find the unknown bending moments. compression on the top fibers of the beam). The released structure consist of
simple beams as shown in figure 4.3.
Hence,
• The bending moments at the intermediate supports of the beam are

selected as redundant.
• This will lead to a system of simultaneous equations in which there are
no more than three unknowns per equation, independently of the number
of redundants.
• When the bending moments are released from the structure, the conti-
nuity of the beam at the supports is broken, which results in having a
series of simple beams as released structure.
• Each of the simple beams is subject to external forces that act on it
in the normal direction, together with the two bending moments at the
ends.
g replacements
1 2 3 4 5

Figure 4.1: Continuous beam.
c
4.2 Method of three moments 79 80 Continuous beams
Each span is loaded by the external loads and two redundant bending moments.
These loads produce deflections and rotations in the two simple beams. Denote
0
the angle of rotation at the support B of the left beam by θB and denote the
00
same rotation on the right beam by θB . These angles are considered positive
as shown in Figure 4.3, that is, they have the same direction as the positive
PSfrag replacements Ia Ib bending moment MB .

La Lb As the beam is continuous through the support B, the compatibility equation
is
0 00
θB = −θB (4.1)
that express the fact that the angles of rotation of the two adjacent simple
beams in the released structure must be consistent with the condition of con-
tinuity at support B.
0 00
The next step is to develop expressions for θB and θB and substitute them in the
A B C

previous equation. For that purpose, it is convenient to apply superposition.

RA RB RC The first span can be divided into subproblems, using as redundant the two
0
bending moments at the end as shown in figure 4.4 where θ1B represents the
Figure 4.2: A segment of a continuous beam. 0 0
slope due to the external forces and θ2B and θ3B represents the slopes due to
the internal moments.
Forget for a moment the slopes due to the external loadings and concentrate
MA MB MC momentarily on the slopes due to the internal moments. From tables:
A B B C 0 MA La
eplacements θ2B = (4.2)
6EIa

0 MB La
00
RA 0
RB 00
RB RC0 θ3B = (4.3)
3EIa
0
Therefore, the total slope θB is
0 0 MA La MB La
θB = θ1B + + (4.4)
6EIa 3EIa
0 MA MB
RA MB MC 0
where θ1B still has to be found but will not represent any problems since it
can be obtained from superposition as well.
00
A similar procedure can be used to find the angle of rotation θB based on figure

A B B C
0
θB 00
4.5.
θB
From Tables:
00 MB Lb
Figure 4.3: Derivation of the three–moment equation. θ2B = (4.5)
3EIb
c
MB MC
MA MB
00
0 θB
θB

00
A B
0
θ1B B C θ1B

PSfrag replacements MA MB
0
A B θ2B
PSfrag replacements 00
B C θ2B

MB
0 MC
A B θ3B

00
B C θ3B

0
Figure 4.4: Using superposition to find θB for the first span.

00
Figure 4.5: Using superposition to find θB for the second span.
00 MC Lb
θ3B = (4.6)
6EIb The equations (4.4) and (4.7) give the angle of rotation in the released struc-
tures at support B. These equations are sometimes referred as the force-
In the previous equations it has been assumed that the modulus of elasticity rotation relations.
E is the same for both spans of the beam. 0 00
The final step in the analysis is as follows. Since from equation (4.1) θB = −θB
00
Using equations (4.5) and (4.6), the total slope θB is MA La MB La MB Lb MC Lb
0 00
θ1B + + = −θ1B − − (4.8)
MB Lb MC Lb 6EIa 3EIa 3EIb 6EIb
00 00
θB = θ1B + + (4.7)
3EIb 6EIb it is possible to write

00
where, once more, θ1B is the slope or angle of rotation due to the external La La Lb Lb 0 00
MA + 2MB + + MC = −6Eθ1B − 6Eθ1B (4.9)
loading ont he second spam and has to be found. Ia Ia Ib Ib
c
PSfrag replacements
The above equation is known as the three–moment equation since it relates wo

P
three consecutive moments in the beam. M1

If all the spans of the beam have the same moment of inertia I, the three–

moment equations simplifies into:

1 I1 2 I2 3

0 00
MA La + 2MB (La + Lb ) + MC Lb = −6Eθ1B − 6Eθ1B (4.10)
R1 R2 R3
L1 L2
Furthermore, if all the spans have the same length, the equations simplifies
even more
6EI 0 6EI 00 Figure 4.6: Continuous beam with a wall support.
MA + 4MB + MC = − θ − θ (4.11)
L 1B L 1B
As previously mentioned, it has been assumed that the modulus of elasticity 4.2 has a intensity q. From tables the angle of rotation at support B due to
is the same throughout all spans of the continuous beam. this uniform load is
0 qL3A
θ1B = (4.12)
It is important to notice that when performing calculations with the three– 24EIA
0 00
moment equations, the angles θ1B and θ1B must be expressed in radians.
If a uniform load were acting on the right span, the angle of rotation at B
would be
4.2.1 Procedure for analysis 00 qL3B
θ1B = (4.13)
24EIB
The procedure to use the three–moment equation is:
If a concentrated load P acts at the center of each span, the corresponding
angles of rotation are
1. To formulate the equation once for every intermediate support to obtain
as many equations as redundant bending moments exist. 0 P L2A 00 P L2B
θ1B = θ1B = (4.14)
16EIA 16EIB
2. Solve simultaneously the equations obtained in the previous step.
as obtained from tables.
3. Knowing the bending moments at all supports, find the reactions at every
support.
4.2.3 Beam with fixed support
4.2.2 Angles of rotation
In the previous explanation it was supposed that the edges of the beam were
simple supported. If any of both edges were a wall support, as shown in figure
It is important to note that the RHS of the three–moment equations can be 4.6, the number of redundant moments increase.
evaluated from the loadings acting on the beam. These angles can be obtained
from tables giving deflections and slopes of simple beams. For example, con- One simple way to handle this case is to substitute the wall support by
sider that the uniform load acting on the left span of the beam shown in figure an additional span with an infinite moment of inertia. The effect of
c
this extra infinite span is to avoid rotation at support 1, which is the same This equation may be used at each support until all reactions are calculated.
restriction impose by the fixed wall.
The procedure described above is used at each support until all the reactions
The bending moments determined at points 1, 2 and 3 of the continuous beam are obtained. Of course, if a concentrated load acts directly over a support,
remain the same as in the original beam. The length of the new span is the load is transmitted directly to the reaction.
irrelevant (except that it should be larger than zero) as it will disappear from
the three–moment equation (see figure 4.7).
4.2.5 Example 4.1
4.2.4 Reactions GIVEN The continuous beam shown in figure 4.8 has three spans of equal
length L and constant moment of inertia I.
Once the bending moments at the supports of a continuous beam have been
determined, it is not difficult to find the reactions using the static equilibrium FIND Determine all the reactions of the beam using the three–moment equa-
equations. If RB0
and RB 00
are the reactions at B for two simple beams repre- tion.
0
senting two adjacent spans, the total reaction RB is just the sum of both RB
00
and RB .
0
The reaction RB is made up from three parts: the reaction of the simple beam
due the external loadings, the reaction due to MA that is equal to MA /LA ,
and the reaction due to MB that is equal to −MB /LA . Similarly, the reaction
00
RB is equal to the reaction of the simple beam due to the external forces, plus
MC /LB minus MB /LB . Combining all these terms, the total reaction RB is
obtained.
g replacements
The above procedure can be summarized in the following equation:
PSfrag replacements
RB = Simple-beam reaction of member AB due to the loads
+ Simple-beam reaction of member BC due to the loads
MA MB MB MC q P = qL
+ − − + (4.15)
LA LA LB LB
3L/4
wo 1 2 3 4
P
0 I=∞ 1 2 3

I1 I2

L L L

M1 L>0 L1 L2 R1 R2 R3 R4
Figure 4.7: Continuous beam with an infinite length span. Figure 4.8: Example 4.1.
c
00
Solution The angle θ1B is the rotation at the left-hand end of span 3-4. Since the load
The beam rests on four simple supports which give a total of four unknowns on span 3-4 is a concentrated load, the angle of rotation can be obtained from
and therefore the beam is statically indeterminated to the second degree. The tables also taking care of selecting the correct end of the simple beam
bending moments M2 and M3 at supports 2 and 3, respectively, are the redun-
dant quantities. 00 P ab(L + b) P (3L/4)(L/4)(L + L/4) 5P L2
θ1B = = = (f)
6LEI 6LEI 128EI
Since the moments of inertia and span lengths are the same for all three spans,
it is possible to use the simplified three-moment equation given in equation Since the load P is equal to qL, the above equation becomes
(4.11). Hence,
6EI 0 6EI 00 00 5qL3
MA + 4MB + MC = − θ − θ (a) θ1B = (g)
L 1B L 1B 128EI
Now, this equation should be written for supports 2 and 3. Susbtituing equations (e) and (g) in equation (a) gives
Support 2. The adjoining spans are 1-2 and 2-3, and therefore the moment MA 5qL3 15qL2

6EI
becomes M1 , MB becomes M2 and MC becomes M3 . Since the first support M2 + 4M3 = − =− (h)
L 128EI 64
does not support bending moments, M1 = 0 and only M2 and M3 will appear
in the equation.
Solution of equations. To find the redundant bending moments, equations (d)
The angle of rotation 0
θ1B is the rotation at the right-hand end of member and (h) can be solved simultaneously. The results are
1-2 when it is treated as a simple beam under a distributed load. Thus, from
49qL2
tables, M2 = − (i)
0 qL3 960
θ1B = (b)
24EI 11qL2
00
Similarly, the angle θ1B is the simple-beam rotation at the left-hand side of M3 = − (j)
240
span 2-3. Since this span has no external loading, the angle simply becomes
00
θ1B =0 (c) Reactions. The reactions at each support can now be obtained by using free-
body diagrams and equations of equilibrium, or by substituting directly into
equation (4.15).
Substituting equations (b) and (c) in equation (a) yields
qL3 qL2 At support 1, the terms in equation (4.15) are as follows (note that since

6EI
4M2 + M3 = − =− (d) LA = LB only five terms are different):
L 24EI 4
Support 3. At support 3 the moment MA becomes M2 , MB becomes M3 and 1. The simple-beam reaction for span AB is zero because span AB does
MC becomes M4 . Since M4 = 0, only the moments M2 and M3 will appear in not exist.
the equation.
2. The simple-beam reaction for span BC, which is span 1-2, is qL/2.
0
The angle of rotation θ1B is the angle at the right-hand end of span 2-3 when
it is treated as a simple beam. Thus, 3. The moment MA is zero because it does not exist.
0
θ1B =0 (e) 4. The moment MB is zero because it is the bending moment at support 1.
c
4.2 Method of three moments 89
5. The moment MC is the bending moment M2 , which is given by equation

(i).
Mechanical components subject
to static loading
Doing all the substitutions yields
qL 49qL2 431qL
R1 = 0 + +0−0−0− = (k)
2 960L 960
At support 2 the terms for equation (i) are as follows:
1. The simple-beam reaction for span AB, which now is span 1-2, is qL/2.
2. The simple-beam reaction for span BC, which now is span 2-3, is zero.
3. The moment MA is the bending moment at support 1 and is equal to

zero.
4. The moment MB is the bending moment M2 which is given by equation

(i).
5. The moment MC is the bending moment M3 given by equation (j).
Substituing term-by-term gives

Augustin Louis Cauchy (1789–1857)
2 2 2
qL 49qL 49qL 11qL 89qL
R2 = +0+0+ + − = (l)
2 960L 960L 240L 160 Laplace and Lagrange were visitors at the Cauchy family home and Lagrange
in particular seems to have taken an interest in young Cauchy’s mathematical
Following a similar procedure for the remaining two reactions, all reaction for education. Lagrange advised Cauchy’s father that his son should obtain a good
the beam can be obtained. The remaining two reactions are grounding in languages before starting a serious study of mathematics. In 1802
Augustin-Louis entered the Ecole Centrale du Panthéon where he spent two
93qL years studying classical languages.
R3 = (m)
320
From 1804 Cauchy attended classes in mathematics and he took the entrance
169qL examination for the Ecole Polytechnique in 1805. He was examined by Biot and
R4 = (n)
240 placed second. At the Ecole Polytechnique he attended courses by Lacroix,
de Prony and Hachette while his analysis tutor was Ampère. In 1807 he
graduated from the Ecole Polytechnique and entered the engineering school
Ecole des Ponts et Chaussées. He was an outstanding student and for his
practical work he was assigned to the Ourcq Canal project where he worked
under Pierre Girard.
c
5.1 Introduction 91 92 Mechanical components subject to static loading
In 1810 Cauchy took up his first job in Cherbourg to work on port facilities important to understand how the stresses behave, how to compute them and
for Napoleon’s English invasion fleet but in September of 1812 he returned how to apply this information to their mechanical design.
to Paris. An academic career was what Cauchy wanted and he applied for a
post in the Bureau des Longitudes. He failed to obtain this post, Legendre On this unit, the following topics will be covered:
being appointed. He also failed to be appointed to the geometry section of the
Institute, the position going to Poinsot. • Stresses acting on a point.
Other posts became vacant but one in 1814 went to Ampère and a mechanics • Find the principal strains and stresses.
vacancy at the Institute, which had occurred when Napoleon Bonaparte re-
• Establish the criteria for design of ductile materials.
signed, went to Molard. In this last election Cauchy did not receive a single
one of the 53 votes cast. • Establish the criteria for design of fragile materials.
Cauchy was not well regarded among his peer researchers, although most of • Apply the design criteria for design of components subject to combined
them agreed on his genius. A rather silly dispute with Duhamel clouded the loading.
last few years of Cauchy’s life. This dispute was over a priority claim regarding
a result on inelastic shocks. Duhamel argued with Cauchy’s claim to have been
the first to give the results in 1832. Poncelet referred to his own work of 1826
on the subject and Cauchy was shown to be wrong. However Cauchy was
5.2 Components of stress
never one to admit he was wrong.
Suppose that it is necessary to find the stresses acting at a point p. The first
Probably Cauchy’s greatest achivement in mechanics was developing the gen- step is to introduce a coordinate system and a sign convention in order to
eral theory of stress. Its deep and thorough originality is fully outlined only differentiate the different stresses.
against the background of the century of achievement by the brilliant geome-
ters who preceded, treating the special kinds and case of deformable bodies Consider figure 5.1. Once the x and y axes have been established, a normal or
by complicated and sometimes incorrect ways without ever hitting upon this shear stress component is positive provided it acts in the positive coordinate
basic idea, which immediately became and has remained the foundation of the direction on the positive face of the element, or it acsts in the negative coordi-
mechanics of gross bodies. nate direction on the negative face of the element. In figure 5.1, all the normal
and shear stress components are positive.
He produced 789 mathematics papers.
All the stress components in figure 5.1 maintain equilibrium of the element,
and because of this, knowing the direction of τxy on one face of the element
defines its direction on the other three faces. Hence, the above sign convention
5.1 Introduction can be remembered simply noting that
Under normal conditions, mechanical components are subject to states of stress positive normal stress acts outward from all faces and pos-
that cannot be fully described using just a normal or shear stress. Tipically, itive shear stress acts upward on the right–hand face of the
components will experience a combination of both, normal and shear stresses, element.
which adds a little bit of complexity to the analysis.
These complex states of stress may lead to failure of a part if the analyst This sign convention for the stresses agress with the equilibrium of the element,
or engineer is not able to predicted them accurately. Therefore, it is very because the shear stresses in opposite faces of an infinitesimal element must be
c
5.3 Stresses on inclined sections 93 94 Mechanical components subject to static loading
σy θ y
y
y1
τyx
τxy σ y1
1 x1
PSfrag replacementsσx τ y1x 1
σx τ x1y
PSfrag replacements
σ x1
τxy θ
x
τyx
x
σy
Figure 5.1: Two–dimensional stresses acting on a point. σ x1 1

τ x1y 1
τ y1x
equal in magnitude and opposite in direction. Hence, a positive stress τxy acts σ y1
upward on the positive face and downward on the negative one. In a similar Figure 5.2: Two–dimensional stresses acting on a point oriented to the x1 y1 z1
manner, the stresses τyx acting on the top and bottom faces of the element are angle.
positive although they have opposite directions.
At this point it is important to remember that shear stresses on perpendicular

planes are equal in magnitude and have directions such that both stresses direction. Associated with this new element is the coordinate system x1 y1 z1
point toward, or both point away from, the line of intersection of the faces. such that the z1 axis coincides with the z axis and the x1 y1 axes are rotated
Therefore, it can be show that counterclockwise through an angle θ with respect to the xy axes.
As shown in figure 5.2, the normal and shear stresses acting on this new element
τxy = τyx (5.1) are denoted σx1 , σy1 , τx1y1 = τy1x1 , noting that the shear stresses acting on all
four side faces of an element in plane stress are known, if the shear stress acting
on any one of those faces is known.
The stresses acting on the inclined x1 y1 element can be expressed in terms

5.3 Stresses on inclined sections of the stresses on the xy element by using equations of equilibrium. For this
purpose, consider a wedge–shaped stress element having an inclined face that
Given the state of plane stress shown in figure 5.1, the orientation of the is the same as the x1 face of the inclined element and where the two other
inclined plane on which the normal and shear stress components are to be faces are parallel to the x and y axes. This stress element is pictured in figure
determined will be defined using the angle θ assuming that σx , σy and σxy are 5.3.
known.
In order to write equations of equilibrium for the wedge, it is useful to con-
To describe the stresses acting on an inclined section, it is necessary to consider struct a free–body diagram showing the forces acting on the faces. For that
a new stress element that is located at the same point in the material as the effect, consider that the left–hand side face has an area Ao (remember that the
one in figure 5.1 but which faces are parallel and perpendicular to the inclined element is actually a volume and has therefore certain thickness). Then, since
c
PSfrag replacements y y
y1 y1
PSfrag replacements
x1 x1
τx1y1 τx1y1 Ao sec θ
θ σx1
θ
θ σ x Ao θ
σx
x σx1 Ao sec θ
x
τxy
τxy Ao
σy1 τyx
τyx Ao tan θ
τy1x1
σy σy Ao tan θ
Figure 5.3: Stresses acting on a wedge-shaped element. Figure 5.4: Forces acting on a wedge-shaped element.
σ = F/A, the normal and shear forces acting on that face are σx Ao and τxy Ao Isolating σx1 from the previous equation
as shown in the free-body diagram of figure 5.4. The area of the bottom face
is Ao tan θ, and the area of the inclined face is Ao sec θ. Thus, the normal and
sin θ sin θ

shear forces acting on these face have the magnitudes and directions shown in σx1 = σx cos θ(cos θ) − τxy sin θ cos θ + cos θ cos θ + σy sin θ cos θ
cos θ cos θ
the figure.
σx1 = σx cos2 θ + τxy (sin θ cos θ + sin θ cos θ) + σy sin2 θ
The forces acting on the left–hand and bottom faces can be resolved into σx1 = σx cos2 θ + σy sin2 θ + 2τxy sin θ cos θ (5.5)
orthogonal components acting in the x1 and y1 directions. Then, two equations
of equilibrium are obtained by summing forces in those directions.
From equation (5.3)
Summation of forces in the x1 direction gives
τx1 y1 A0 sec θ = −σx A0 sin θ + τxy A0 + cos θ
σx1 Ao sec θ − σx Ao cos θ − τxy Ao sin θ σy A0 tan θ cos θ − τxy A0 tan θ sin θ = 0 (5.6)
−σy Ao tan θ sin θ − τyx Ao tan θ cos θ = 0 (5.2)
Isolating now τx1 y1 ,

Accordingly, summation of forces in the y1 direction gives
τx1 y1 = −σx sin θ cos θ + τxy (cos2 θ − sin2 θ) + σy sin θ cos θ
τx1y1 Ao sec θ + σx Ao sin θ − τxy Ao cos θ
τx1 y1 = −(σx − σy ) sin θ cos θ + τxy (cos2 θ − sin2 θ) (5.7)
−σy Ao tan θ cos θ + τyx Ao tan θ sin θ = 0 (5.3)
Using the relationship τxy = τxy Equations (5.5) and (5.7) give the normal and shear stresses acting on the x1
plane in terms of the angle θ and the stresses σx , σy and τx y acting on the x
σx1 A0 sec θ = σx A0 cos θ − τxy (A0 sin θ + A0 tan θ cos θ + σy sin θ tan θA0 (5.4) and y planes.
c
The above equations can be written in a more convenient form by introducing One particularity of the stresses acting at a point, is that the sum of the normal
the following trigonometric identities: stresses acting on perpendicular faces is constant and independent of the angle
1 1 θ. Hence, it can be show that
cos2 (θ) = (1 + cos 2θ) sin2 (θ) = (1 − cos 2θ)
2 2 σx1 + σy1 = σx + σy (5.11)
1
sin θ cos θ = sin 2θ
2
As the sum of the normal stresses is always the same, it is known as the first
stress invariant.
From equation (5.5),
σx σy
σx1 = (1 + cos 2θ) + (1 − cos 2) + τxy sin 2θ
2 2
σx + σ y σx − σ y 5.4 Principal stresses and maximum share
σx1 = + cos 2θ + τxy sin 2θ (5.8)
2 2
From the equations,
From equation (5.7),
σx + σ y σx − σ y
σx1 = + cos 2θ + τxy cos 2θ (5.12)
σx − σ y 1 1 2 2
τ x 1 y1 = − sin 2θ + τxy (1 + cos 2θ) − (1 − cos 2θ)
2 2 2
σx − σ y 1 cos 2θ 1 σx − σ y
τ x 1 y1 = − sin 2θ + τxy + − (1 − cos 2θ) τ x 1 y1 = − sin 2θ + τxy cos 2θ (5.13)
2 2 2 2 2
σx − σ y
τ x 1 y1 = − sin 2θ + τxy cos 2θ (5.9)
2 it can be seen that the stresses vary as a function of θ. Hence, the stresses
vary continuously as the orientation of the element is changed. At certain
Equations (5.8) and (5.9) are known as the transformation equations for angles the normal stress reaches a maximum or a minimum value. Similarly,
plane stress because they transform the stress components from one set of the shear stress has a maximum and a minimum. It is useful, and sometimes
axes to another. However, it is important to have in mind that the intrinsic necessary, to find these maximum and minimum values for the stresses.
state of stress at the point under consideration is the same whether represented
by stresses acting on the xy element (figure 5.1) or by stresses acting on the To determine the maximum and minimum normal stresses it is necessary to
inclined x1 y1 element (figure 5.2). differentiate equation (5.12) with respect to θ and set the result equal to zero
Since the transformation equations were derived solely from equi-

dσx1 d σx + σ y d σx − σ y d
librium of an element, they are applicable to stresses in any kind = 2 + cos 2θ + (τxy sin 2θ) = 0
of material, whether linear or nonlinear, elastic or non-elastic. dθ dθ dθ 2 dθ

σx − σ y
− 2 sin 2θ + 2τxy cos 2θ = 0 (5.14)
2
As of now, only an expression for the normal stress σx1 has been found. An
expression to find the normal stress σy1 is obtained if the angle θ is substituted
for θ + 90o in equation (5.7). This substitution gives, The above expression can be simplified dividing it by cos θ to obtain
σx + σ y σx − σ y 2τxy
σy1 = − cos 2θ − τxy sin 2θ (5.10) tan 2θ = (5.15)
2 2 σx − σ y
c
5.4 Principal stresses and maximum share 99 100 Mechanical components subject to static loading
2y
+ τx
2 From equations (5.16), the following relations can be obtained
−σ
y
PSfrag replacements s σx 2
τxy = R2 sin2 2θp
2
τxy (σx − σy )2 = 4R2 cos2 2θp (5.19)
=
R
2θp Substituting the above relations into equation (5.18) gives
σx − σ y σx + σy 4R2 cos2 2θp R2 sin2 2θp

σ1 = + +
2 2 4R R
σx + σ y 2 2
Figure 5.5: Geometric representation of equation (5.15) σ1 = + R cos 2θp + R sin 2θp
2
σx + σ y
σ1 = +R (5.20)
Solving this equation, two roots for θ, namely θp1 and θp2 give the orientation 2
of the planes of maximum and minimum normal stress. These planes receive
the name of principal planes and their angles are 90◦ apart. Therefore, the From where the following relationship is obtained
principal stresses occur on mutually perpendicular planes. s 2
σx + σ y σx − σ y 2
σ1 = + + τxy (5.21)
The principal stresses can be calculated by substituting each of the two values 2 2
of θp into equation 5.8. If the principal stresses are determined using the
above procedure, both the principal stress, and the angle in which it acts, can
Doing a similar procedure using equation (5.10) or from the relationship σ1 +
be obtained.
σ2 = σx + σy , the following equation arises
Another way to calculate the principal stresses is to obtain general formulas s 2
starting from equation (5.15). From the triangle shown in figure 5.5, two σx + σ y σx − σ y 2
σ2 = − + τxy (5.22)
relationships can be obtained 2 2
σx − σ y τxy
cos 2θp = sin 2θp = (5.16) Now that the principal stresses and their directions are know, the next step is
2R R to find the maximum shear stress and the planes on which they act.
where A similar procedure to the one used for the principal stresses can the followed:
s 2
σx − σ y 2 d
R= + τxy (5.17) τx y = − (σx − σy ) cos 2θ − 2τxy sin 2θ = 0 (5.23)
2 dθ 1 1
which eventually yields
Substituting equations (5.16) into (5.12) yields σx − σ y
tan 2θs = − (5.24)
2τxy
σx + σ y σx − σ y σx − σ y τxy
σ1 = + + τxy The previous equation yields again two roots between 0 and 180◦, and both
2 2 2R R
σ + σy (σx − σy )2 τxy 2 values differ by 90◦ . Therefore, the maximum shear stress occur on perpen-
σ1 = + + (5.18) dicular planes. Because shear stresses on perpendicular planes are equal in
2 4R R
c
5.5 Strain transformation 101 102 Mechanical components subject to static loading
absolute value, the maximum positive and negative shear stresses differ only y
in sign. y1 θ
PSfrag replacements
It is possible to show that planes of maximum shear stress occur at 45◦
of the principal planes. x1
The plane of the maximum positive shear stress τmax is defined by the angle
θs1 for which the following equation apply θ
τxy σx − σ y x
cos 2θs1 = sin 2θs1 = − (5.25) O
R 2R
Figure 5.6: A coordinate system x1 y1 rotated and angle θ from the system xy.
Substituting these expressions into equation (5.13) yields
s 2
σx − σ y
τmax = 2
+ τxy (5.26) section, equation that relate the strains in the inclined section to the strains in
2 referenced directions will be found. These transformation equations are widely
where τmin has the same magnitude but different sign. used in laboratory and field investigations involving measurements of strains.
Another expression for τmax can be found from σ1 and σ2 expressions: Strains are usually measured by strain gauges placing them in aircrafts to mea-
sure structural behavior during flight or in buildings to measure the effects of
σ1 − σ 2
τmax = (5.27) earthquakes. Since each gauge measures the strain in one particular direction,
2 it is usually necessary to calculate the strains in other directions by means of
the transformation equations.
The planes of maximum shear stress also contain normal stresses. The normal
stress acting on the planes of maximum positive shear stress can be determined The transformation equations for plane strain, that allow to find the state of
by substituting the expressions for the angle θs1 into the equation for σx1 . The strain at a point at one particular direction x1 y1 , are obtained in a similar
resulting stress is equal to the average of the normal stresses on the x and y form as the ones for stress transformation.
planes:
σx + σ y Consider the two coordinate systems xy and x1 y1 which have the same origin
σave = (5.28)
2 but are rotated, one respect the other, an angle θ as show in figure 5.6.
To find the strain transformation equations for plain strain, it is necessesary

5.5 Strain transformation to relate the normal strains x1 and y1 and the shear strain γx1 y1 to the ones
associated with the original axes namely, x , y and γxy . As the equation for
x1 can be obtained taking into account that y1 (θ) = x1 (θ + 90◦ ), it is not
In the previus section it was explored how to convert from a state stresses neccesary to derive a different equation.
acting at a point to another state acting on the same point but at different
planes. It was also explored how to obtain the principal planes and the prin- To determine the normal strain x1 in the x1 direction, consider a small element
cipal stresses. of material selected so that the x1 axis is along a diagonal and the x and y
axes are along the sides of the element as shown in figure 5.7.
As the stresses, the strains at a point in a loaded structure vary according to
the orientation of the axes, in a manner similar to that for stresses. In this The strain x in the x direction produces an elongation in the x direction equal
c
to x dx where dx is the length of the element. As a result of this elongation. y

y dy sin θ
The diagonal of the element increases in length by an amount
PSfrag replacements x1
x dx cos θ (5.29)
y dy
frag replacements y x dx cos θ
x1 y1
y1 α2 dy
ds
θ
dy x
ds α1 O
θ dx
x Figure 5.8: Deformation of an element in plane strian due to a normal strain

O
dx x dx y .
Figure 5.7: Deformation of an element in plane strian due to a normal strain Similary, the strain y in the y direction produces an elongation in the y di-
x . rection equal to y dy, where dy is the length of the side of the element parallel
to the y axis.
As a result of this elongation, the diagonal of the element increases in length

by an amount
y dy sin θ (5.30)
as shown in figure 5.8.
Finally, consider the shear strain γxy in the xy plane. This strain produces
a distortion of the element such that the angle at the lower left corner of
the element decreses by an amount equal to the shear strain (see figure 5.9).
Consequently, the upper face of the element moves to the right by an increase
in the lenght of the diagonal equal to
γxy dy cos θ (5.31)
By the principe of superposition, the total increase ∆d in the length of the

diagonal is the sum of the three increases in length. Hence,
∆d = x dx cos θ + y dy sin θ + γxy dy cos θ (5.32)
c
frag replacements PSfrag replacements
y y
γxy dy cos θ
x1 b
y1 β
γxy dy γ x 1 y1 = α + β
a
y1
γxy α x1
dy
ds α3
θ θ
x
O
x
O
dx Figure 5.10: Shear strain γx1 y1 associated with the x1 y1 .
Figure 5.9: Deformation of an element in plane strian due to a shear strain

where α is the change in angle from y1 and β is the change in the angle x1 .
γxy .
It is necesary now to find α and β. The angle α can be found from the
The normal strain x1 in the x1 direction is equal to this increase in length deformation. The strain x produces a clockwise rotation of the diagonal of
divided by the initial length ∆d of the diagonal the elements. Denoting this angle as α1 , from figure 2 it can be found that
∆d dx dy dy x dx
x1 = = x cos θ + y sin θ + γxy cos θ (5.33) sin α1 = sin θ (5.37)
ds ds ds ds ds
Noting that dx/ds = cos θ and dy/ds = sin θ, the following expression is Since α1 is very small, sin α1 ≈ α1 and therefore
obtained for the normal strain dy
α1 = x sin θ (5.38)
x1 = x cos2 θ + y sin2 θ + γxy sin θ cos θ (5.34) ds
Similary, the strain y produces a counterclockwise rotation of the diagonal

The above expression relates the normal strain x1 to the values of x , y and through an angle α2
γxy associated with the xy axes. As mentioned previously, the normal strain dy
y1 in the y1 direction is obtained from this equation by substituting θ + 90◦ α2 = y cos θ (5.39)
ds
for θ
y1 = x cos2 θ − y sin2 θ − γxy sin θ cos θ (5.35) Finally, the strain γxy produces a clockwise rotation through an angle α3 equal
to then distance γxy dy sin θ divided by ds
To find the shear strain γx1 y1 it should be noted that this strain is equal to the
dy
decrease in angle batween lines in the material that were initially along the α3 = γxy sin θ (5.40)
x1 y1 axes. Considering figure 5.10, the shear γx1 y1 is ds
γx 1 y 1 = α + β (5.36) Therefore, the resultant counterclockwise rotation of the diagonal equal to the
c
angle α is 5.5.1 Principal strains

α = −α1 + α2 − α3 (5.41)
The principal strains exist on perpendicular planes with the principal angles
dx dy dy
= −x sin θ + y cos θ − γxy sin θ θp calculated from the equation
ds ds ds
γxy
observing that dx/ds = cos θ and dy/ds = sin θ tan 2θp = (5.49)
x − y
α = − (x − y ) sin θ cos θ − γxy sin2 θ (5.42) where the principal strains can be calculated from
s 2
The angle β can be found substituting θ for θ + 90◦ in the above equation but x + y x − y γxy 2
1,2 = ± + (5.50)
noting that the angle β runs in opposite direction 2 2 2
β = (x − y ) sin (θ + 90◦ ) + γxy sin2 (θ + 90◦ ) which correspond to the equation for plane stress.
= − (x − y ) sin θ cos θ + γxy cos2 θ sin2 θ

(5.43)
The maximum shear strains in the xy plane are associated with axes at 45◦ to
the direction of the principal strains. The algebraically maximum shear strain
The shear strain γxy is obtained adding α + β (in the xy plane) is given by the following equation
γx1 y1 = −2 (x − y ) sin θ cos θ + γxy cos2 θ − sin2 θ

(5.44) s 2
γmax x − y γxy 2
= + (5.51)
2 2 2
A more useful form of the equation is found dividing each term by two
γx 1 y1 γxy The minimum shear strain has the same magnitude but negative sign.
cos2 θ − sin2 θ

= − (x − y ) sin θ cos θ + (5.45)
2 2
The transformation equations for plain strain can be expressed in terms of the 5.5.2 Strain rosettes
following identities
One of the main applications of the equations of strain transformation is to
1 1
cos2 θ = (1 + cos 2θ) sin2 θ = (1 − cos 2θ) obtain the state of strain at certain points of stressed bodies.
2 2
1 In practice, the strains are measured by means of a strain gauge. The most
sin θ cos θ = sin 2θ
2 common type are resistance strain gauges which are based on the principle that
yielding, the electrical resistance of a wire is proportional to the square of its length (if
x + y x − y γxy volume keeps constant). Once the relationship between the strain of a given
x1 = + cos 2θ + sin 2θ (5.46)
2 2 2 wire and its electrical resistance has been established experimentally, the axial
γx 1 y1 x − y γxy strain of the wire can be determined by measuring its resistance, which can be
=− sin 2θ + cos 2θ (5.47)
2 2 2 done very accurately.
It can be verified easily that exists a first invariant also for the strain To use the calibrated wire as a strain gauge, it is bonded to the surface of an
unloaded mechanical component. When the component is loaded, its strain in
x 1 + y1 = x + y (5.48) the direction of the wire is determined by measuring the wire resistance. The
c
equation, equation (5.34), for each gauge. Hence,

a = x cos2 θa + y sin2 θa + γxy sin θa cos θa
b = x cos2 θb + y sin2 θb + γxy sin θb cos θb
c = x cos2 θc + y sin2 θc + γxy sin θc cos θc (5.52)
The values of x , y and γxy are determined by solving these three equations
simultaneously.
Figure 5.11: Typical resistance strain gauge.

Example 5.1
b GIVEN Strain rosettes are often arranged in 45◦ or 60◦ patterns. Consider
the case of a 45◦ or “rectangular” rosette.
PSfrag replacementsc FIND Obtain expressions to find the values of x , y and γxy .
Solution
Consider figure 5.13. Since the gauge a and c are directly aligned with the
θc axes x and y, it is possible to write
θb a
a = x (a)
θa c = y (b)
x
Figure 5.12: A strain rosette. Notice that the two above equation can also be obtained substituting the values
of θa = 0◦ and θc = 90◦ into the first and last expressions in equation (5.52).
To obtain the shear strain γxy , the strain-transformation equation expressed

wire is typically very thin and it is arranged in a pattern like the one shown in in the form of equation (5.46), can be used
figure 5.11 to minimize the size of the gauge and so measure the strain within x + y x − y γxy
a relatively small neighborhood of a point. x1 = + cos 2θ + sin 2θ (c)
2 2 2
One strain can measure strain in only one direction, which is usually not where for x1 = c , θ = 45◦ . Hence,
sufficient to obtain the state of strain at a given point. Fortunately, this can a + c a − c γxy
b = + cos 90◦ + sin 90◦
be done using three strains in what is known as a strain rosette. A strain rosette 2 2 2
are three strain gauges set up to measure strain in three different directions a + c γxy
b = + (d)
as shown in figure 5.12. As will be discussed next, this is enough to find the 2 2
complete state of strain at a point. Finally,
γxy = 2b − a − c (e)
In the general case, the axes of the three gauges are arranged at the angles θa , θb
and θc as shown in figure 5.12. If the reading a , b and c are taken, the strain Once more, the previous equation could be obtained using equation (5.52)
components x , y and γxy at the point by applying the strain-transformation directly.
c
y
a
◦
45
c
PSfrag replacements 45◦
PSfrag replacements
b
45◦ x b
a y
45◦ c
x
Evaluating equations (a), (b) and (c), the following values are obtained
Example 5.2 x = −360 × 10−6 y = 475 × 10−6 γxy = 385 × 10−6 (d)
GIVEN A rectangular strain rosette as the one shown in figure 5.14a is

mounted on the fuselage of an aircraft. The following readings are obtained: With these values, the principal strains can be calculated using equation (5.50)
a = 475 × 10−6 , b = 250 × 10−6 and c = −360 × 10−6 . s 2
x + y x − y γxy 2
FIND Obtain the in-plane principal strains. 1,2 = ± + (e)
2 2 2
Solution which gives
The first step is to find x , y and γxy . It is possible to simplify the problem
changing the reference coordinate system such that the rosette c is aligned 1,2 = 57.5 × 10−6 ± 429.74 × 10−6
with the x axis and the rosette a is aligned with the y axis (a clockwise 90◦ 1 = 517 × 10−6
rotation of the coordinate system) (see figure 5.14).
2 = −402 × 10−6 (f)
After the coordinate system has been changed, it is easy to see that
x = a (a)
5.6 Hooke’s law
y = c (b)
The stresses and strains are related to each other. Before going into more
Finding γxy is not difficult as it can be done using the same procedure as the detail, it is important to recognize that
previous problem as with the new orientation the rosette is the one shown in
figure 5.13. Using the results obtained,
• The strains relationships seen so far are not dependent on the material,
γxy = 2b − a − c (c) they are a kinematic description of the deformation.
c
5.6 Hooke’s law 113 114 Mechanical components subject to static loading
• The stress relationships do not depend upon the material as they are only due to a shear strain γxy ; that is, τxy will not cause other strains in the material.
a kinetic description of the deformation (for the relationships derived here Likewise, τyz and τxz will only cause shear strains γyz and γxz respectively.
a linear material was assumed).
Therefore, the following relationships can be written
To find this stress–strain relationship, it is convenient to use the principle 1

of superposition. Remembering that = σ/E, it is possible to consider the γxy = τxy (5.59)
G
separete application of each normal stress acting on a piece of material. When
x is applied, the element elongates in the x direction 1
γyz = τyz (5.60)
σx G
0x = (5.53)
E
1
γxz = τxz (5.61)
G
For σy , the element elongation in the x direction will be
where G is the shear modulus G = E/2(1 + ν).
σy
00x = −ν (5.54)
E The previous results can also be written expressing the stresses in terms of the
strains in the form
where the minus sign indicates that the material is contracting.
E
For σz a similar expression is obtained σx = [x (1 − ν) + ν(y + z )] (5.62)
(1 + ν)(1 − 2ν)
σz
000
x = −ν (5.55)
E E
σy = [y (1 − ν) + ν(x + z )] (5.63)
(1 + ν)(1 − 2ν)
When these three effects are superimposed, the normal strain x is obtained
σx σy σz E
x = −ν −ν σz = [z (1 − ν) + ν(y + x )] (5.64)
E E E (1 + ν)(1 − 2ν)
1
= (σx − ν(σy + σz )) (5.56) E
E τxy = τyx = γxy (5.65)
2(1 + ν)
The normal strains in y and z can be obtained in a similar fashion
E
1 τyz = τzy = γyz (5.66)
y = (σy − ν(σx + σz )) (5.57) 2(1 + ν)
E
E
1 τxz = τzx = γxz (5.67)
z = (σz − ν(σy + σx )) (5.58) 2(1 + ν)
E
Equations (5.56), (5.57) and (5.58) are the normal strains and do not take into Equations (5.62) to (5.67) express the Generalized Hooke’s law. This re-
consideration the shear strains. If a shear stress is applied in the material, lets lationship between stresses and strains is valid for isotropic linear materials
say, τxy , experimental observation indicate that the material will deform only only.
c
5.7 Failure criteria 115 116 Mechanical components subject to static loading
The above relationships can be written in a more convenient form as 5.7.1 Rankine criterion
σx = 2Gx + λe The Rankine or maximum principal stress criterion is probably the simplest
σy = 2Gy + λe theory. It states that yielding (or fracture) begins at a point in a member
σz = 2Gz + λe (5.68) when the greatest positive (or negative) principal stress exceeds the tensile (or
compressive) ultimate strength.
τxy = Gγxy
τyz = Gγyz Hence, the Rankine criterion states that yielding will occur when
τzx = Gγzx t
σ1 ≥ σult (σ1 > 0)
c
σ3 ≥ σult (σ3 < 0) (5.71)
where e = x + y + z and λ and G are known as Lame’s constants and are
given by t
where σult c
is the ultimate stress at uniaxial tension, σult is the ultimate stress
E c
at uniaxial compression and the negative sign in σ3 and σult have been omitted
G= (5.69)
2(1 + ν) for simplicity.
Unlike ductile materials, the failure of brittle materials occurs at relatively

Eν low strains, and there is little, or no, permanent yielding on the planes of
λ= (5.70) maximum shearing stress. For this reason elastic breakdown and failure in a
(1 + ν)(1 − 2ν)
brittle material are governed largely by the maximum principal stress.
A useful graphical representation of the Rankine criterion as applied to biaxial

5.7 Failure criteria stresses is shown in figure 5.15. In this plot, σ3 = 0 and σ1 and σ2 are the
principal biaxial stresses and each one can be either positive or negative. If a
point in the material whose in-plane state of stress is plotted on the boundary
The previous analyses have made use of a linear stress-strain relationships. or outside the quadrilateral area, will indicate failure. For brittle materials
However, when the load reaches a certain point, stress is no longer proportional this failure will come as fracture rather than yield.
to strain and permanent deformation occurs. The material is then said to
have yielded. Knowing the stress at which yielding behaviour commences, it
becomes primal to design components to withstand such stress. Under uniaxial 5.7.2 Tresca criterion
loading this is easy task, as there is only one principal stress to consider.
This criterion assumes that yielding is dependent on the maximum shear stress
Unfortunately, designing components containing complex principal stress sys-
in the material reaching a critical value. This is taken as the maximum shear
tems so the material remains elastic even under full load is rather complex. It
stress at yielding in a uniaxial tensile test. This theory was originally proposed
is therefore essential to have some criterion based on stresses, strains or related
by Coulomb and studied later in-detail by Tresca, the reason for which this
quantities in which the complex system can be related to simpler cases.
theory is usually refered as Tresca theory.
The most common criteria for yielding are the maximum shear-stress or Tresca
The maximum shear stress is given in terms of the maximum and minimum
criterion, the shear-strain energy or von Mises criterion, the maximum princi-
principal stresses
pal stress or Rankine criterion and the Mohr-Coulomb criterion. The first two σ1 − σ 3
apply to ductile materials while the last two apply to brittle materials. τmax = (5.72)
2
c
+σ2 in-plane +σ2 in-plane
t
σult σY
c
σult t
σult −σY
PSfrag replacements
−σ1 in-plane +σ1 in-plane −σ1 in-plane +σ1 in-plane
PSfrag replacements σY
c
σult −σY
−σ2 in-plane
−σ2 in-plane
Figure 5.16: Tresca theory illustrated on σ1 , σ2 coordinates (for biaxial
Figure 5.15: Rankine theory illustrated on σ1 ,sigma2 stresses).
where σ3 keeps its sign (negative if it is a compressive stress). 5.7.3 Von Mises criterion
Under uniaxial tension there is only one principal stress, σ1 (σ2 = σ3 = 0), so The shear-strain energy criterion was suggested independently by Maxwell,
the maximum shear stress is von Mises and Hencky, although is often attributed to von Mises. This crite-
rion states that yielding begins when the distortional strain energy (the strain
τmax = σ1 /2 (5.73) energy due to the change in shape) at a point equals the distortional strain
energy density at yield in uniaxial tension.
and at yield this become τY = σY /2.
The distortional energy density UD can be written in terms of the second
Therefore the Tresca criterion states that yielding will occur when deviator stress invariant J2 as
σ1 − σ 3 σY 1
= (5.74) UD = J2 (5.76)
2 2 2µ
or when the maximum principal stress difference equals the yield stress in where
1
(σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ3 − σ1 )2

simple tension J2 = 6
(5.77)
σ1 − σ 3 = σ Y (5.75)
At yield in uniaxial tension, σ1 = σY , σ2 = σ3 = 0. Then
Figure 5.16 illustrates this theory for biaxial stress (σ3 = 0). As in the previous J2 = 13 σY2 (5.78)
theory, if any point of the material is subjected to plane stress, and its in-plane
principal stresses are represented by coordinates (σ1 , σ2 ) lying on the boundary Therefore, the von Mises criterion states that yielding will occur when
or outside the hexagonal area shown, the material will yield and failure will
2
occur. σvm = σY2 (5.79)
c
+σ2 in-plane The failure function for the Mohr-Coulomb criterion can be written in terms
of the stress state and two material properties, the cohesion c and angle of
internal friction φ
σY
f = σ1 − σ3 + (σ1 + σ3 ) sin φ − 2c cos φ (5.83)
−σY
PSfrag replacements
−σ1 in-plane +σ1 in-plane where the principal stresses are in the order σ1 > σ2 > σ3 and failure will
σY t
occur when f = σult c
or f = σult .
−σY The material properties c and φ can be expressed in terms of the ultimate
t
stress at uniaxial tension σult and the ultimate stress at uniaxial compression
c
σult .
−σ2 in-plane s
t c
σult σult
Figure 5.17: Von Mises theory illustrated on σ1 , σ2 coordinates (for biaxial c = t
2 σult
stresses). s !
t
π −1 σult
φ = − 2 tan c
(5.84)
where σvm is the effective von Mises stress given by 2 σult
q p
σvm = 12 [(σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ3 − σ1 )2 ] = 3J2 (5.80)
Another graphical representation of the Mohr-Coulomb theory that can be
found in some books is shown in figure 5.19. This graphical representation is
Relative to the general axes (x1 , x2 , x3 ), J2 can be expressed in terms of the c
obtained sketching three Mohr’s circles obtained from the compression (σult ),
stress invariants I1 and I2 t
tension (σult ) and torsion (τult ) tests. The area that envelop these three circles
J2 = I2 + 13 I12 (5.81)
will mark the area of no failure.
in which case the von Mises stress can also be defined as
q
2 2 2 2 2 2
σvm = σ11 + σ22 + σ33 + 3(σ12 + σ13 + σ23 ) − σ11 σ22 − σ22 σ33 − σ11 σ33
(5.82)
In plane stress problems, equation (5.79) represents an elliptical curve as shown

in figure 5.17. Once more, if a point in the material has in-plane principal
stresses represented by coordinates (σ1 , σ2 ) which lay on the boundary or out-
side the elliptical area shown, the material will yield.
5.7.4 Mohr-Coulomb criterion
The Mohr-Coulomb theory can be seen to be a modification of the Tresca

theory. This criterion, which is illustrated in figure 5.18 for biaxial stresses, is
based upon the results of the uniaxial tension and compression tests.
c
5.7 Failure criteria 121
+σ2 in-plane
Strain Energy
t
σult
c
σult t
σult
−σ1 in-plane +σ1 in-plane
PSfrag replacements
c
σult
−σ2 in-plane
Figure 5.18: Mohr-Coulomb theory illustrated on σ1 , σ2 coordinates (for biax-

ial stresses).
Otto Christian Mohr (1835–1918)
τ Born of Holstein landowners on the coast of the North Sea, Otto Mohr became
one of Europe’s most decorated engineers of the 19th century. Early in his
career, while working for the railroads in Hannover and Oldenburg, he designed
some of the first steel trusses as well as some of the most renowned bridges
τult in Germany. During those years, Mohr also began his theoretical work in
mechanics and strength of materials.
c t
σult σult At 32, Otto Mohr became an educator, first as professor of mechanics at the
σ
PSfrag replacements Stuttgart Polytechnikum and, later, at Dresden Polytechnikum. Despite an
unpolished delivery, his lectures were well received by students because of their
simplicity, clarity, and conciseness. Being both a theoretician and practicing
engineer, Mohr knew his subject thoroughly and was always able to bring
something fresh and interesting to his students’ attention.
In addition to a lone textbook, Mohr published many research papers on the

theory of structures and strength of materials. Graphical solutions to specific
Figure 5.19: Another representation of the Mohr-Coulomb theory obtained
problems were a common theme in many of them. Borrowing upon earlier
from three circles of Mohr.
work by Karl Culmann, he expanded the graphical representation of stress
about a point to three dimensions. Later, using the ”circles of stress” with
c
6.1 Introduction 123 124 Strain energy
which his name is now commonly associated, Mohr developed the first theory B
of strength based on shearing stresses.
Following retirement from the Polytechnikum, Mohr remained in the Dresden

area where he continued his scientific work until his death.
C
B L A
PSfrag replacements
6.1 Introduction x
Strain energy is a fundamental concept in applied mechanics, and strain en-

ergy principles are widely used for determining the response of machines and C
structures to both static an dynamic loads. In this unit, the concepts of strain P
energy and strain energy density will be discussed. Furthermore, the use of the
strain energy to solve tension, torsion and bending problems will be explored.
Figure 6.1: Bar suffering an elongation due to the application of a static force
P.
6.2 Principles of strain energy vs. displacement diagram, the total work U done by the loading as the bar
is deformed is given by
Consider a bar BC of length L and a cross–sectional area A as shown in figure
6.1. The bar is fixed at its B end, and an gradually increasing axial force P is Z x1
applied at C. Such a load is called a static load because there are no dynamic U= P dx (6.2)
0
or inertial effects due to motion. The bar gradually elongates as the load is
applied, eventually reaching its maximum elongation x at the same time that
The work done by the load P done while the bar is deforming gradually must
the load reaches the maximum value of P .
result in an increment of some energy associated with the deformation of the
If the magnitude of the force P is plotted against the deformation, a diagram bar. This energy receives the name of strain energy of the bar, and by
load vs. displacement is obtained that is characteristic of the bar BC (see definition
figure 6.2). During the loading process, the load P moves slowly through the
distance x and does a certain amount of work. Z x1
Strain energy = U = P dx (6.3)
Consider the work dU done by the load P while the bar is being elongated a 0
small length dx as shown in figure 6.3. This differential of work is equal to the
magnitude of the load P times the small elongation dx: Remember that work and energy must be expressed on units obtained from
multiplying units of length by units of force. Hence, if the SI system is used,
the work and energy are expressed in N·m. This unit is the joule (J) which is
dU = P dx (6.1) equal to one newton meter. In the English system of units, work and energy
are expressed in foot–pounds (ft–lb), foot–kips (ft–k), inch–pounds (in–lb),
Since this work can be calculated as the area below the curve of the loading and inch–kips (in–k).
c
6.2 Principles of strain energy 125 126 Strain energy
P PSfrag replacementsP
P = kx
P1
PSfrag replacements x
O x1
x
Figure 6.4: Load vs. displacement diagram for a bar of linearly elastic material.
Figure 6.2: Load vs. displacement diagram for the bar BC.
linearly elastic material is given by the equation
The portion of the diagram loading vs. elongation corresponding to a linear
elastic deformation, can be represented by a straight line as shown in figure PL
x= (6.6)
6.4. The equation of this straight line is given by: EA
where E is the Young’s modulus, A is the cross-sectional area and L is the

P = kx (6.4) length of the bar.
Following the definition of strain energy given in equation 6.3, it is possible to Combining the equation above with equation (6.5), the strain energy of a
write linearly elastic bar can be expressed as
Z x1 Z x1 x
kx2 1
U = P dx = kxdx = P 2L EAx2
0 0 2 0 U= or U= (6.7)
1 2 1 2EA 2L
U = kx = P1 x1 (6.5)
2 1 2
The first equation express the strain energy as a function of the load and the
Furthermore, the relationship between the load P and the elongation x for a second expresses it as a function of the elongation. From the first equation
it can be seen that increasing the length of a bar increases the amount of
P strain energy even though the load is unchanged because more material is
P being strained by the load. On the other hand, increasing either the modulus
PSfrag replacements of elasticity or the cross–sectional area decreases the strain energy because the
strains in the bar are reduced.
P
dU
x x 6.3 Impact loading
x1
x dx U
The concept of strain energy is particularly useful in the determination of the
Figure 6.3: Differential work and total work done by the force P on the bar. effects of impact loads in structures or components.
c
6.4 Strain energy density 127 128 Strain energy
Consider, for example, a body of mass m that is moving with a velocity vo and Hence, if the volume of the bar is V = AL, it is possible to write from equation
impact the end B of a bar AB (see figure 6.5). Not taking into account the (6.3):
inertia of the bar, and assuming that there is no dissipation of energy during
the impact, the maximum strain energy Um that the bar absorbs is equal to
the original kinetic energy of the object that is moving U 1 x1
Z
1
Z x1
= P dx = P dx
V V AL 0
Z x1 0
1 U P dx
Um = T = mvo2 (6.8) = (6.9)
2 V 0 A L
where Um is the strain energy absorbed during the impact and T is the kinetic Remembering that the normal stress is defined by
energy of the object.
From above, it is possible to determine the value of the static force Pm needed P
σx = (6.10)
to create exactly the same strain energy in the bar. Thus, the maximum stress A
σm in the bar can be found dividing the force Pm by the cross-sectional area
of the bar. and the unit normal strain is
x
εx = (6.11)
L
6.4 Strain energy density
equation (6.9) can be expressed as
The load vs. deformation diagram for the bar BC depends on the length L and
the cross–sectional area A of the bar. Therefore, the strain energy U depends U
Z ε1
on the bar dimensions. To be able to eliminate the effect of the dimensions = σx dεx (6.12)
V O
and keep all the attention in the properties of the material, it is necessary to
g replacements
consider the strain energy per unit of volume.
where ε1 is the value of the unit strain that correspond to an elongation x1 .
The strain energy per unit volume, U/V is known as the strain energy den-
A A sity and is denoted by the letter u
U = Um
U =0 ε1
σ = σm
Z
σ=0 strain energy density = u = σx dεx (6.13)
0
B B
vo The strain energy density u has units of energy divided by volume. The SI
v=0
T = mv units are joules per cubic meter (J/m3 ) and the USCS units are foot–pounds
m T =0 per cubic foot, inch–pounds per cubic inch and similar units. Since all of these
units reduce to units of stress, the strain energy density can be expressed in
Figure 6.5: An object of mass m impacting a bar at its end. Pascals or psi.
c
6.4 Strain energy density 129 130 Strain energy
σx σx Lost σx Failure
eplacements
Recovered
u
PSfrag replacements
ε ε
ε1 εp ε1
Unloading ε
εf
Figure 6.6: Definition of the strain energy density and energy densities lost
and recovered after unloading. Figure 6.7: Definition of the toughness modulus.
The strain energy density u is equal to the area below the curve of the σ vs. ε
diagram, measured from εx = 0 to εx = ε1 . σx = Eεx (6.14)
If the material is unloaded, the stress return to zero but a permanent defor- Substituting the above expression into equation (6.13), gives
mation may exist represented by the unit strain εp and only a part of the
strain energy per unit volume can be recovered (see figure 6.6). The rest of ε
ε1
Eε2x 1 Eε21
Z
the energy used to deform the material is lost as heat. u= Eεx dεx = = (6.15)
0 2 O 2
Substituting ε1 = σ1 /E into equation (6.15),

6.4.1 Toughness and resilience modulus
1 σ12 1
u= = σ1 ε 1 (6.16)
The value of the strain energy given by taking ε1 = εf , the deformation at 2E 2
failure, is known as the toughness modulus of the material. As shown in
figure 6.7, this modulus is equal to all the area below the curve of the stress– The value uY of the strain energy density is obtained selecting σ1 = σY where
strain diagram. Hence, the toughness modulus represents the strain energy σY is the yield stress of the material. This particular value of u receives the
density needed to cause failure in a material. name of resilience modulus and can be expressed simply as
It should be noted that the toughness modulus is related to the ductility and
the ultimate tensile stress of the material. Also, it is important to keep in σY2
Resilience modulus = uY = (6.17)
mind that the capacity of a structure to resist an impact loading depends on 2E
the toughness of the material used.
As shown in figure 6.8, the resilience modulus is equal to the area below the
If the stress σx remains within the proportionality limit for the material, linear part of the stress–strain diagram and represents the energy per unit
Hooke’s law is still valid and it is possible to write volume that the material can absorb without yielding. Hence, the capacity of
c
6.5 Elastic strain energy for normal stresses 131 132 Strain energy
σx is still valid but the stress σx , the unit strain εx and the strain energy density
u will change from one point to the next.
Y
σY For values of σx within the proportionality limit where σx = Eεx ,
PSfrag replacements
1 1 1 σx2
u = Eε2x = σx εx = (6.20)
2 2 2E
ε
εY The value of the strain energy U of a body subject to a normal uniaxial stress
can be obtained as
Figure 6.8: Definition of the resilience modulus.
σx2
Z Z
U= udV = dV (6.21)
a structure to support an impact loading without yielding will depend on the 2E
material used.
The above expression is only valid for elastic deformations and receives the
Since the toughness and resilience modulus are characteristic values of the name of elastic strain energy of a body.
strain energy density of a material, both values are expressed in J/m3 or its
multiples.
Example 6.1
6.5 Elastic strain energy for normal stresses GIVEN Consider three bars of circular section that have the same length
L but different shapes as shown in figure 6.9. The first bar has a diameter d
As the bar considered in the previous section was subject to an uniform stress throughout its length, the second bar has this diameter d in the one quarter
sigmax , the strain energy density throughout the bar was constant and could of its length and the third bar in just one eighth. The second and third bars
be defined as the ratio U/V of the strain energy U and the volume V of the have a diameter 3d elsewhere. The three bars are subject to a tensile force P .
bar.
FIND Compare the amount of strain energy that are stored in the bars if a
In a structural element, or component of a machine with a non-uniform stress linear elastic behavior is considered.
distribution, the strain energy density can be defined considering the deforma-
Solution
tion of a small element of material of volume ∆V and writing
The strain energy density stored in a differential of a bar can be obtained from
the equation
∆U dU du = σx dx (a)
u = lim = (6.18)
∆V →0 ∆V dV
If a linear elastic behavior is considered, both quantities σx and εx are related
The expression
in this range by the Young’s Modulus E which is constant. Therefore it is
possible to write
Z εx Z x Z x
u= σx dεx (6.19) u= σx dεx = Eεdεx (b)
0 0 0
c
PSfrag replacements
1 2 3 F A F

L
PSfrag replacements 3d 3d
Figure 6.10: Bar of length L subject to a tensile force F .
d d L d L
L 4 8 bars of the problem under consideration. Nevertheless, it is possible to apply
the previous results by parts, obtaining the strain energy stored in the sections
of the bar of constant cross–sectional area and adding the results together to
3d 3d obtain the total strain energy stored.
Now that an expression to evaluate the strain energy in the bars has been
found, consider the cross–sectional areas given by the diameters d and 3d
P P P
Figure 6.9: Problem 6.1. πd2

A1 =
4
1 1 σ2 π(3d)2 πd2
u = Eε2x = σx εx = x (c) A2 = =9 = 9A1 (f)
2 2 2E 4 4
Since the interest is in the strain energy stored in the bars and not in the
strain energy density, it is necessary to integrate the strain energy density in
the bars to obtain the strain energy. This is a simple task since For the bar 1:
σx2
Z Z
P 2L
U= udV = dV (d) U1 = (g)
V V 2E 2EA1
For a bar of constant cross–sectional area A, as the one shown in figure 6.10,
the previous equation can be rewritten as: For the bar 2:
2
1 F
Z
U= dV P 2 (L/4) P 2 (3L/4)
V 2E A U2 = +
F2
Z
F 2V F 2 AL 2EA1 2EA2
U= dV = = 2
1 P L 1
2
P L
2EA2 V 2EA2 2EA2 U2 = +
2
F L 4 2EA1 12 2EA1
U= (e) 1 P 2L

1
2EA U2 = = U1
3 2EA1 3
The above expression is only valid for bars of constant cross–sectional area, 1
U2 = U1 (h)
and at first sight it may look that it cannot be applied to the second and third 3
c
For the bar 3:

P 2 (L/8) P 2 (7L/8)

U3 = +
2EA1 2EA2
x
2 2
1 P L 7 P L
U3 = +
8 2EA1 72 2EA1 L
2 P 2L

2 PSfrag replacements
U3 = = U1 dx

9 2EA1 9

2
U3 = U1 (i)
9
Px
The comparison of results show that the strain energy decreases as the volume
of the bar is increased, despite all bars having the same maximum stress.
Hence, the third bar has less capacity to absorb energy than the other two.
Therefore, only a small amount of work will increase the stress in the bar with As seen in the previous example, the strain energy can be found from
a notch. The stretcher the notch, the more the stress will increase. When the
σx2 P x2
Z Z
loads are dynamic and the capacity to absorb energy is more important, the U= dV = dV (b)
presence of notches is quite undesirable. It should be noted however, that for V 2E V 2EA2
static loads, the maximum stress is more important as design factor than the
ability to absorb energy. Since the cross–sectional A is constant, the integral over the volume can be
transformed into a line integral
Example 6.2 Z L
γ 2 A2 (L − x)2
U= A dx (c)
0 2EA2
GIVEN Consider the prismatic bar suspended from one of its ends and sub-
ject to its own weight as shown in figure 6.11.
where the expression for P found previously has been already substituted.
FIND Find the strain energy stored in the bar. Suppose a linear elastic
Carrying the integral out
behavior.
L
γ2A
Z
Solution U= (L − x)2 dx
The first step to solve this problem is to find out the magnitude of the force 2E 0
L
P that is acting on the bar. Note that the force P acting on the differential γ 2A 3

element dx is given by the weight below it. Hence, U= (L − x)
6E 0
P (x) = Px = γA(L − x) (a) γ 2 AL3
U= (d)
6E
where γ denotes the specific weight of the material, A is the cross–sectional
area of the bar and the term (L − x) represents the length of the bar below Consider now that an external load Q is acting on the bar at the same time
the differential element. pulling it downwards. In this case, the total axial force in the element P (x)
c
A B F F

β β

PSfrag replacements

β β
H
PSfrag replacements
P C
P
Figure 6.13: Free–body diagram for the problem under consideration.

by this load is equal to the strain energy U stored in the structure. This energy
will be the sum of the external force and the force due to the own weight of
is given by
the bar. Hence, Pδ
P (x) = Px = γA(L − x) + Q (e) W =U = (a)
2
Integrating the previous expression gives, where δ is the deflection through which a work is done by the force P . This
Z L
[γA(L − x) + Q]2 γ 2 AL3 γQL2 Q2 L equation gives a simple way to find the deflection δ if the strain energy U can
U= dx = + + (f) be found.
0 2EA 6E 2E 2EA
This method has the following limitations:
The first term in the previous expression is the same as the strain energy of a
bar hanging under its own weight, and the last term is the same as the strain • Only one force should act on the truss.
energy of a bar subject only to an axial force Q. However, the middle term
contains both γ and Q, showing that it depends upon both the weight of the • The only deflection that can be found is the deflection at the point where
bar and the magnitude of the applied load. Thus, this example illustrates that the load is being applied.
the strain energy of a bar subjected to two loads is not equal to the sum of
the strain energies produced by the individual loads acting separately. Consider now the free–body diagram for the problem under consideration
shown in figure 6.13. From summation of forces in the vertical direction
P
Example 6.3 2F cos β = P F = (b)
2 cos β
GIVEN Consider the truss shown in figure 6.12. where F is the force acting on each bar. If both bars have a length L,
FIND Determine the vertical deflexion δB of the node B of the truss. Note H H
cos β = L= (c)
that the only force acting on the truss is a vertical force P acting on B. Suppose L cos β
that both members have the same axial rigidity EA.
Remebering that
Solution P 2L
Consider a linearly elastic truss subject to just one force P . The work W done U= (d)
2EA
c
6.6 Strain energy in bending 139 140 Strain energy
The total strain energy is given by τxy

F 2L F 2L P 2H
U =2 = = (e)
2EA EA 4EA cos3 β
Since W = P δB /2 = U
P 2H P δB π/2 − γxy
= (f) PSfrag replacements
4EA cos3 β 2
Solving for δB
PH
δB = (g) Figure 6.15: Relation between shear strain and shear stress.
2EA cos3 β
Realizing that the area integral represents the moment of inertia of the beam
6.6 Strain energy in bending about the neutral axis, the final result can be written as
Z L
M2
Consider a general beam as shown in figure 6.15. The stress in the beam at a U= dx (6.25)
given distance y from the neutral axis can be found from 0 2EI
My
σx = − (6.22) To evaluate the strain energy, therefore, it is necessary to express the internal
I
moment as a function of its position x along the beam and then perform the
where M is the bending moment at a distance x from the end A of the beam integration over the entire length of the beam.
and I is the polar moment of inertia.
Not taking into account the shear effect it is possible to write

Z 2 6.7 Elastic strain energy for shear stress
σ M 2y2
Z
U= dV = dV (6.23)
2E 2EI 2
When a material is subject to plane shear stresses τxy (see figure 6.15, the
A differential of volume dV can be written in terms of a differential of length strain energy density at a point can be expressed as
times a differential of area as dV = dAdx. Hence,
Z L γxy
M2 M2
Z Z Z
U= 2
dAdx = 2
dx y 2 dA (6.24) u= τxy dγxy (6.26)
V 2EI 0 2EI A 0
PSfrag replacements where γxy is the unit angular strain that correspond to τxy . Note that the
strain energy density u is equal to the area below the curve of the shear stress
vs. unit shear strain diagram as shown in figure 6.16.
A B
x For values of τxy inside the proportionality limit it is possible to write
Figure 6.14: A General beam. τxy = Gγxy (6.27)
c
6.8 Strain energy in torsion 141 142 Strain energy
2
1 2 1 τxy Remembering equation (6.29),
u = Gγxy = τxy γxy = (6.28)
2 2 2G 2
τxy
Z
U= dV
where G is the shear modulus of elasticity or rigidity modulus. V 2G
Therefore, the value of the strain energy U of a body subject to shear stresses Substituting equation (6.30) into the previous equation gives
can be obtained from
T 2 ρ2
Z 2
τxy
Z
U= dV (6.29) U= 2
dV (6.31)
V 2G V 2GJ
It is important to remember that the above expression is only valid for elastic Remembering that dV can be expressed as dAdx and noting that T 2 /2GJ 2 is
strains and defines the strain energy associated with the shear strains present only a function of x it is possible to write
in the object. Z L
T2
Z
U= 2
ρ2 dAdx (6.32)
0 2GJ A
6.8 Strain energy in torsion Once more, the integral over the area is nothing more than the definition of
the polar moment of inertia J. Therefore,
Consider the axle BC of length L shown in figure 6.17. If the bar is subject Z L 2
T
to a torsional force the shear stress can be calculated as U= dx (6.33)
0 2GJ
Tρ
τxy = (6.30) In the case of an axle of constant cross–section subject to two torsional mo-
J
ments of equal magnitude and opposite directions as shown in figure 6.18, the
strain energy is given by
where ρ is the radial distance from the axis of rotation of the axle, J is the
polar moment of inertia of the cross–section at a distance x from the end B
and T is the internal torque acting on the same cross–section. T 2L
U= (6.34)
2GJ
τxy
PSfrag replacements x
PSfrag replacements u B
T
C
γxy L
Figure 6.16: Shear strain vs. shear stress diagram. Figure 6.17: Bar under torsional loading.
c
6.8 Strain energy in torsion 143 144 Strain energy
Substituting the expresion for the torsional moment into the previous equation,
T
PSfrag replacements T
L
q 2 (L − x)2
Z
U= dx
0 2GJ
Figure 6.18: Bar under torsional loading. q2
Z L
= (x − L)2 dx
2GJ 0
L
Example 6.4 q 2 (x − L)3
=
2GJ 3
0
GIVEN A circular bar AB fixed at one end and free on the other, as shown in q 2 L3
figure 6.19, is loaded with a distributed torsional moment which has a constant U= (c)
6GJ
intensity q.
FIND Obtain an expression for the strain energy stored in the bar when the 6.9 Deflection of beams due to impact load
load is applied.
Solution The dynamic deflection of a beam subject to an impact load can be calculated
The torsional moment acting at a distance x from the fixed end of the bar can under certain simplifying conditions stating that the work done by the impact
be obtained from statics, load and the strain energy in the beam are equal.
For that purpose, some suppositions have to be made:

Tx = qL − qx = q(L − x) (a)
• The weight that is striking the beam move with it after the impact.
The formula for the strain energy under torsion, equation (6.33) is given by
• There are no energy losses.
L • The beam is linearly elastic.

Tx2
Z
U= dx (b)
0 2GJ • The shape of the deformed beam is the same under a dynamic load and
under a static load.
qL
• The change of potential energy due to the change in the position of the
q beam can be neglected.
PSfrag replacements Tx
Generally speaking, these suppositions are adequate if the mass of the object
that is falling is very big compared with the mass of the beam. Otherwise,
x this approximated analysis is not valid, an a more detailed analysis is needed.
Consider the beam shown in figure 6.20 where an object of weight W hold at
Figure 6.19: Problem 6.4 an original height h, strikes a beam at its center. To carry out the analysis
c
PSfrag replacements
6.9 Deflection of beams due to impact load 145 146 Strain energy
W M (x) Peq
A B
x
h Peq Peq
2 L/2 L/2 2
A B
L/2 L/2 Figure 6.21: Free–body diagram of the beam under consideration.

PSfrag replacements
Peq
L
M2
Z
U= dx (6.37)
0 2EI

δ

L/2 L/2
The moment of the beam can be found from the free–body diagram shown in
figure 6.21.
Figure 6.20: Beam under an impact load.
For 0 ≤ x ≤ L/2:
Peq
is necessary as before to consider an equivalent static force Peq that produces M (x) = x (6.38)
the same amount of strain energy as the impact load. 2
From tables, the deflection of the beam due to a static load applied at its For L/2 ≤ x ≤ L:
center is: Peq

L

M (x) = x − Peq x − (6.39)
2 2
Peq L3
δ= (6.35)
48EI Substituting the previous expressions for the moment into equation (6.37)
yields,
Solving for the equivalent static force,
48EI Z L 2
2 P
Peq = δ (6.36) 1 eq 2
L3 U = x dx
2EI 0 4
Z L 2
1 Peq
where δ is the deflection of the beam due to the dynamic load at its center. + (x − L)2 dx (6.40)
2EI L2 4
This deflection is also the maximum deflection of the beam.
As described in a previous section, the strain energy stored in a beam can be

obtained from the following expression Carrying the integrals out
c
6.9 Deflection of beams due to impact load 147 148 Strain energy
• Kinetic energy of the object: Kf = 0 (at rest)

L
( L )
2
Peq x3 2 (x − L)3 • Potential energy of the object: Vf = −W δ
U = +
8EI 3 0 3 L
24EI 2
2
• Strain energy of the beam: Uf = δ
2
Peq L3 L3 L3

U = +
8EI 24 24
2 3
Since it is assumed that there are not energy looses, the sum of all the energies
Peq L before and after the impact should be zero. Hence,
U = (6.41)
96EI
From equation (6.36), the equivalent static force Peq is given by, Ki + Vi + U i = Kf + Vf + U f
24EI 2
48EI W h = −W δ + δ (6.44)
Peq = δ L3
L3
The above expression is a quadratic equations in terms of the deflection δ:
Substituting the above expression into equation (6.41),
24EI 2
24EI 2 δ − Wδ − Wh = 0 (6.45)
U= δ (6.42) L3
L3
Solving the above equation gives,
It should be noted that for this case, the strain energy can be obtained from
the relation U = 1/2P δ that is the work done by the equivalent static force, r
96EIW h
W± W2 +
L3
Peq δ 24EI 2 δ= (6.46)
U= = δ (6.43) 48EI
2 L3 L3
To define the problem completely, it is necessary to find the value of the deflec- Noting that the deflection due to a static load is
tion δ. The value of this quantity can be obtained doing a balance of energies
before and after the impact.
W L3
δst = (6.47)
Before the impact: 48EI
It is possible to write
• Kinetic energy of the object: Ki = 0 (at rest)
• Potential energy of the object: Vi = W h r

96EIh
• Strain energy of the beam: Ui = 0 δ = δst + δst 1+
r W L3
2h
After the impact: δ = δst + δst 1 + (6.48)
δst
c
6.9 Deflection of beams due to impact load 149
From this expression it is clear that the dynamic deflection is always greater
than the static one. If the height is equal to 0, which means that the load is
Columns
applied suddenly without free fall, the maximum deflection is double the static
one.
If h is very big compared with the static deflection, then the term containing
the h will have a value several times the others and therefore the above equation
can be simplified into
p
δ≈ 2hδst (6.49)
The deflection δ obtained by the above procedure generally represents a upper

limit since it was assumed that there were no losses of energy during the
impact. The local deformation of the contact surfaces, the tendency of the
falling mass to bounce and the mass of the beam tend to reduce the deflection.
Leonhard Euler (1707–1783)
Leonhard Euler’s father was Paul Euler. Paul Euler had studied theology
at the University of Basel and had attended Jacob Bernoulli’s lectures there.
In fact Paul Euler and Johann Bernoulli had both lived in Jacob Bernoulli’s
house while undergraduates at Basel. Leonhard Euler was born in Basel, but
the family moved to Riehen when he was one year old and it was in Riehen,
not far from Basel, that Leonard was brought up.
Leonhard was sent to school in Basel and during this time he lived with his
grandmother on his mother’s side. This school was a rather poor one, by all
accounts, and Euler learnt no mathematics at all from the school. However his
interest in mathematics had certainly been sparked by his father’s teaching,
and he read mathematics texts on his own and took some private lessons.
Euler’s father wanted his son to follow him into the church and sent him to
the University of Basel to prepare for the ministry. He entered the University
in 1720, at the age of 14, first to obtain a general education before going on
c
6.9 Deflection of beams due to impact load 151 152 Columns
to more advanced studies. Johann Bernoulli soon discovered Euler’s great Great. Through the requests of Daniel Bernoulli and Jakob Hermann, Euler
potential for mathematics in private tuition that Euler himself engineered. was appointed to the mathematical-physical division of the Academy rather
than to the physiology post he had originally been offered. Euler became
In 1723 Euler completed his Master’s degree in philosophy having compared professor of physics at the academy in 1730.
and contrasted the philosophical ideas of Descartes and Newton. He began
his study of theology in the autumn of 1723, following his father’s wishes, Daniel Bernoulli held the senior chair in mathematics at the Academy but when
but, although he was to be a devout Christian all his life, he could not find he left St Petersburg to return to Basel in 1733 it was Euler who was appointed
the enthusiasm for the study of theology, Greek and Hebrew that he found in to this senior chair of mathematics. The financial improvement which came
mathematics. Euler obtained his father’s consent to change to mathematics af- from this appointment allowed Euler to marry which he did on 7 January 1734,
ter Johann Bernoulli had used his persuasion. The fact that Euler’s father had marrying Katharina Gsell, the daughter of a painter from the St Petersburg
been a friend of Johann Bernoulli’s in their undergraduate days undoubtedly Gymnasium. Katharina, like Euler, was from a Swiss family. They had 13
made the task of persuasion much easier. children altogether although only five survived their infancy. Euler claimed
that he made some of his greatest mathematical discoveries while holding a
Euler completed his studies at the University of Basel in 1726. By 1726 Euler baby in his arms with other children playing round his feet.
had already a paper in print, a short article on isochronous curves in a resisting
medium. In 1727 he published another article on reciprocal trajectories and The publication of many articles and his book Mechanica (1736-37), which ex-
submitted an entry for the 1727 Grand Prize of the Paris Academy on the best tensively presented Newtonian dynamics in the form of mathematical analysis
arrangement of masts on a ship. for the first time, started Euler on the way to major mathematical work.
The Prize of 1727 went to Bouguer, an expert on mathematics relating to Euler’s health problems began in 1735 when he had a severe fever and almost
ships, but Euler’s essay won him second place which was a fine achievement lost his life. By 1740 Euler had a very high reputation, having won the Grand
for the young graduate. However, Euler now had to find himself an academic Prize of the Paris Academy in 1738 and 1740. On both occasions he shared
appointment and when Nicolaus (II) Bernoulli died in St Petersburg in July the first prize with others. Euler’s reputation was to bring an offer to go to
1726 creating a vacancy there, Euler was offered the post which would involve Berlin, but at first he preferred to remain in St Petersburg. However political
him in teaching applications of mathematics and mechanics to physiology. He turmoil in Russia made the position of foreigners particularly difficult and
accepted the post in November 1726 but stated that he did not want to travel contributed to Euler changing his mind. Accepting an improved offer, Euler,
to Russia until the spring of the following year. He had two reasons to delay. at the invitation of Frederick the Great, went to Berlin where an Academy of
He wanted time to study the topics relating to his new post but also he had a Science was planned to replace the Society of Sciences. He left St Petersburg
chance of a post at the University of Basel since the professor of physics there on 19 June 1741, arriving in Berlin on 25 July.
had died. Euler wrote an article on acoustics, which went on to become a
classic, in his bid for selection to the post but he was nor chosen to go forward Even while in Berlin Euler continued to receive part of his salary from Russia.
to the stage where lots were drawn to make the final decision on who would For this remuneration he bought books and instruments for the St Petersburg
fill the chair. Almost certainly his youth (he was 19 at the time) was against Academy, he continued to write scientific reports for them, and he educated
him. young Russians.
As soon as he knew he would not be appointed to the chair of physics, Euler During the twenty-five years spent in Berlin, Euler wrote around 380 articles.
left Basel on 5 April 1727. He travelled down the Rhine by boat, crossed He wrote books on the calculus of variations; on the calculation of planetary
the German states by post wagon, then by boat from Lübeck arriving in St orbits; on artillery and ballistics (extending the book by Robins); on analysis;
Petersburg on 17 May 1727. He had joined the St. Petersburg Academy of on shipbuilding and navigation; on the motion of the moon; lectures on the
Science two years after it had been founded by Catherine I the wife of Peter the differential calculus; and a popular scientific publication Letters to a Princess
c
6.9 Deflection of beams due to impact load 153 154 Columns
of Germany (3 vols., 1768-72). 7.1 Introduction

In 1759 Euler assumed the leadership of the Berlin Academy, although not
the title of President. The king was in overall charge and Euler was not now Mechanical components fail in a variety of ways: tension, compression, shear,
on good terms with Frederick despite the early good favour. Euler, who had fatigue, etc. There is another type of failure where the component bends
argued with d’Alembert on scientific matters, was disturbed when Frederick of- without breaking. Although the material does not completely fail, from the
fered d’Alembert the presidency of the Academy in 1763. However d’Alembert engineering point of view, the component is not performing as intended any-
refused to move to Berlin but Frederick’s continued interference with the run- more. This type of failure is known as buckling.
ning of the Academy made Euler decide that the time had come to leave.
In this section, this type of failure will be discussed, especially, buckling of
In 1766 Euler returned to St Petersburg and Frederick was greatly angered at columns. Columns are long, slender structural members loaded axially in
his departure. Soon after his return to Russia, Euler became almost entirely compression. If a compression member is relatively slender, it may fail by
blind after an illness. In 1771 his home was destroyed by fire and he was able bending or deflecting laterally, rather than by direct compression of the mate-
to save only himself and his mathematical manuscripts. A cataract operation rial. When lateral bending occurs, the column has buckled (see figure 7.1).
shortly after the fire, still in 1771, restored his sight for a few days but Euler
seems to have failed to take the necessary care of himself and he became totally Under an increasing axial load, the lateral deflections will increase too, and
blind. Because of his remarkable memory was able to continue with his work on eventually the column will collapse completely. Buckling is one of the major
optics, algebra, and lunar motion. Amazingly after his return to St Petersburg causes of failure in structures, and therefore the possibility of buckling should
(when Euler was 59) he produced almost half his total works despite the total always be considered in design.
blindness.
Euler is the most prolific writer of mathematics of all time. He made large
bounds forward in the study of modern analytic geometry and trigonometry
P P
where he was the first to consider sin, cos etc. as functions rather than as
chords as Ptolemy had done. B B
He made decisive and formative contributions to geometry, calculus and num-

ber theory. He integrated Leibniz’s differential calculus and Newton’s method
of fluxions into mathematical analysis. He introduced beta and gamma func-
tions, and integrating factors for differential equations. He studied continuum L
mechanics, lunar theory with Clairaut, the three body problem, elasticity,
acoustics, the wave theory of light, hydraulics, and music. He laid the founda- PSfrag replacements
tion of analytical mechanics, especially in his Theory of the Motions of Rigid
Bodies (1765).
A A
We owe to Euler the notation f (x) for a function (1734), P e for the base of
natural logs (1727), i for the square root of -1 (1777), π for pi, for summation
(1755), the notation for finite differences ∆y and ∆2 y and many others. Figure 7.1: Buckling of a column due to an axial compressive load P .
The number of books and papers written by Eulertotaled 886.
c
7.2 Stable, neutral and unstable equilibrium 155 156 Columns
7.2 Stable, neutral and unstable equilibrium Taking into account the magnitudes of the disturbing and restoring forces,
three cases arise:
In order to better understand the nature of buckling, consider a two bar struc-
ture consisting of weightless bars that are rigid and pin connected at their 1. The restoring force kθL/2 is greater than the disturbing force 2P θ
ends, and a spring of stiffness k that is connected to the pin connecting the
two bars and to a fixed wall as shown in figure 7.2. L kL
2P θ < kθ or P < (7.3)
2 4
Under perfect equilibrium conditions, the two bars will remain vertical. It is
This condition is known as stable equilibrium as the forces developed
possible to upset this state of equilibrium displacing the pin C a small amount
by the spring would be adequate to restore the bars back to their vertical
∆. As this displacement is very small, the angle produced will be very small
position.
as well and therefore it is possible to write ∆ = θ(L/2).
2. The restoring force kθL/2 is smaller than the disturbing force 2P θ
P
PSfrag replacements P P tan θ L kL
B 2P θ > kθ or P > (7.4)
2 4
L L θ
P This condition is known as unstable equilibrium as this force P will
2 2
θ cause a small displacement at C moving the structure out of equilibrium
k

C
∆

C F 3. The restoring force kθL/2 is equal to the disturbing force 2P θ

L kL

L θ 2P θ = kθ or P = (7.5)
L P 2 4
2 θ
2
This represents the case of neutral equilibrium and is the boundary
A P tan θ between stability and instability.
Figure 7.2: Buckling of an idealized structure consisting of two rigid bars and
The load P defined by the neutral equilibrium is the critical load. This
a spring.
critical load can be defined as the maximum axial load that a column can
support when it is on the verge of buckling. The critical load is generally
As the spring is connected to the pin A, it will produce a restoring force
denoted as Pcr .
F = k∆ while the bars are subject to an horizontal force P tan θ. Since
∆ = θ(L/2), Since Pcr is independent of the small displacement θ of the bars, any slight
disturbance given to the structure will not cause it to move further out of
L equilibrium. Instead, the bars will remain in the deflected position.
F = kθ (7.1)
2
These three different states of equilibrium for the idealized can be represented
As the angle is small, tan θ ≈ θ and therefore the disturbing force is graphically as shown in figure 7.3, a graph of axial load P versus angle of
rotation θ. The two heavy lines, one vertical and one horizontal, represent the
equilibrium conditions. Point B, where the equilibrium diagram branches, is
2P x = 2P θ (7.2) called the bifurcation point. The horizontal line for neutral equilibrium extends
c
7.3 Ideal column with pin supports 157 158 Columns
P Pcr Pcr
P
Unstable equilibrium B B B
Neutral equilibrium
PSfrag replacements B
PSfrag replacements
L F
Pcr
Stable equilibrium
θ A A A
O
Figure 7.4: Column under a critical load.
Figure 7.3: Equilibrium diagram for buckling of an idealized structure.
to the left and right of the vertical axis because the angle θ may be clockwise removed (see figure 7.4). Any slight reduction in the axial load P from Pcr will
or counterclockwise. The line extends only a short distance because of the allow the column to straighten out, an any slight increase in P beyond Pcr will
assumption that θ is a small angle. cause further increases in lateral deflection.
It should be understood that Pcr may not be the largest value of P that the Whether or not a column will remain stable or become unstable when subjected
structure can support. Nevertheless, in engineering design the critical load is to an axial load will depend on its ability to restore itself, which is based on
considered to be the largest load the column can support. its resistance to bending. Hence, in order to determine the critical load and
the buckled shape of the column, it is possible to use the equation that relates
the internal moment to the elastic curve (deflected shape),
7.3 Ideal column with pin supports
The easiest way to obtain a value of the critical buckling load, is to consider an d2 v
EI =M (7.6)
ideal column that is pin supported. An ideal column is one that is perfectly dx2
straight before loading, is made of homogeneous material, and upon which
the load is applied through the centroid of the cross section. It is further
assumed that the material behaves in a linear-elastic manner and that the It is important to remember that the above equation assumes that the slope
column buckles or bends in the x–y plane. of the elastic curve is small and that the deflections occur only by bending.
Since an ideal column is straight, theoretically the axial load P could be in- When the column is in its deflected position, the internal bending moment can
creased until failure occurs by either fracture or yielding of the material. How- be obtained cutting a section as shown in figure 7.5. In the free-body diagram
ever, when the critical load Pcr is reached, the column is on the verge of shown in the figure, the deflection v and the internal moment M are shown
becoming unstable. In this state, a small lateral force F will cause the column in the positive direction according to the sign convention used to establish
to deflect, and the column will remain in that position even if the force F is equation (7.6). Summing moments, M + P v = 0 and therefore, M = −P v.
c
7.3 Ideal column with pin supports 159 160 Columns
Thus, This equation is satisfied if C1 = 0 which is the trivial solution and would
imply that v = 0 meaning that the column is always straight, no matter the
d2 v magnitude of the force P . The other possible solution is for
EI 2
= −P v
dx
d2 v

P
+ v=0 (7.7) r !
dx2 EI P
sin L =0 (7.10)
EI
P P which is satisfied if r
P
L = nπ (7.11)
EI
x x or
PSfrag replacements L n2 π 2 EI
P = n = 1, 2, 3, . . . (7.12)
v 2 v L2
L vmax M This formula gives the values of P that satisfy the buckling equation and
P provide solutions (other than the trivial solution) to the differential equation.
Thus, the equation of the deflection curve is

n=1
nπx
v = C1 sin kx = C1 sin n = 1, 2, 3, . . . (7.13)
P L
F
x
Figure 7.5: Column under a critical load. 7.4 Critical loads

The above equation is a homogeneous, second order, linear differential equation
The lowest critical load for a column with pinned ends is obtained when n = 1:
with constant coefficients. It can be shown by using the methods of differential
equations that the general solution is of the form
π 2 EI
Pcr = (7.14)
r ! r ! L2
P P
v = C1 sin x + C2 cos x (7.8)
EI EI The corresponding buckled shape (sometimes called a mode shape) is
The two constants of integration are determined from the boundary conditions πx
v = C1 sin (7.15)
at the ends of the column. Since v = 0 at x = 0, then C2 = 0. Since v = 0 at L
x = L, then
as shown in figure 7.6b. The constant C1 represents the deflection at the
r ! midpoint of the column. Specific values for C1 cannot be obtained, since the
P exact deflected form for the column is unknown once it has buckled. However,
C1 sin L =0 (7.9)
EI C1 may have any small value, either positive or negative. Therefore, the part
c
7.4 Critical loads 161 162 Columns
x x
From equation (7.14) it can be seen that the critical load of the column is
PSfrag replacements π 2 EI 4π 2 EI proportional to the flexural rigidity EI and inversely proportional to the square
Pcr = Pcr =
P L2 L2 of the length. Of particular interest is the fact that the strength of the material
B B B itself, as represented by a quantity such as the yield stress, does not appear
in the equation for the critical load. Hence, increasing a strength property
does not rise the critical load of a slender column. It can only be raised by
C1 increasing the flexural rigidity, reducing the length, or providing additional
lateral support.
L C1 The flexural rigidity can be increased by using a material with larger modulus
of elasticity or by distributing the material in such a way as to increase the
C1 moment of inertia I of the cross section, just as a beam can be made stiffer
by increasing the moment of inertia. The moment of inertia is increased by
distributing the material farther from the centroid of the cross section. Hence,
y A y A a hollow tubular member is generally more economical for use as a column
A
(a) (b) (c) than a solid member having the same cross-sectional area.
Figure 7.6: Buckled shapes for an ideal column with pinned ends: (a) initially In the preceding analysis, it was assumed that buckling takes place in the x–y
straight column, (b) buckled shape for n = 1, and (c) buckled shape for n = 2. plane. The latter assumption will be met if the column has lateral supports
perpendicular to the plane of the figure, so that the column is constrained to
buckle in the x–y plane.
of the load–deflection diagram corresponding to Pcr is a horizontal straight line If the column is supported only at its ends and is free to buckle in any direction,
(see figure 7.3). Buckling of a pinned-end column in the first mode is called then bending will occur about the principal centroidal axis having the smaller
the fundamental case of column buckling. moment of inertia. For instance, consider the rectangular and wide-flange cross
sections shown in figure 7.7. In each case the moment of inertia I1 is greater
The type of buckling described above is called Euler buckling and the crit- than the moment of inertia I2 , hence the column will buckle in the 1–1 plane,
ical load described by equation 7.14 is called the Euler load. Both names and the smaller moment of inertia I2 should be used in the formula for the
are in honor of mathematician Leonhard Euler, who was the first person to critical load.
investigate the buckling of slender column and determine its critical load.
As the column will buckle about the principal axis of the cross section having
By taking higher values of the index n in equations (7.12) and (7.13), an the least moment of inertia (the weakest axis), engineers usually try to achieve
infinite number of critical loads and corresponding shape modes are obtained. a balance keeping the moments of inertia the same in all directions. Geomet-
The mode shape for n = 2 has two half-waves, as pictured in figure 7.6c. The rically, circular tubes would make excellent columns. Also, square tubes or
corresponding critical load is four times larger than the critical load for the those shapes having Ix ≈ Iy are often selected for columns.
fundamental case. The magnitudes of the critical loads are proportional to
the square of n, and the number of half-waves in the buckled shape is equal
to n. Buckled shapes for the higher modes are often of no practical interest
because the column buckles when the axial load P reaches its lowest critical
value. The only way to obtain modes of buckling higher than the first is to
provide lateral support of the column at intermediate points.
c
7.5 Radius of gyration 163 164 Columns
2 2 fundamental case of buckling (figure 7.6b), the critical stress is
Pcr π 2 EI
σcr = = (7.17)
A AL2
C C
1 1 1 1
in which I is the moment of inertia for the principal axis about which buckling
PSfrag replacements
occurs. This equation can be written in a more useful form by introducing the
notation
2 2 r
I
Figure 7.7: Cross sections of columns showing principal centroidal axes with r= (7.18)
A
I1 < I2 .
where r is the radius of gyration of the cross section in the plane of bending.
7.5 Radius of gyration Thus, the equation for the critical stress becomes
π2 E
The radius of gyration is a distance that is occasionally encountered in σcr = (7.19)
mechanics and that will be useful to define the critical stress that a column (L/r)2
can support. Radius of gyration of a plane area is defined as the square root
of the moment of inertia of the area divided by the area itself. Thus, In the previous expression, L/r is a nondimensional ratio called the slender-
ness ratio:
r r
Ix Iy L
rx = ry = (7.16) Slenderness ratio = (7.20)
A A r
where rx and ry denote the radii of gyration with respect to the x and y axes, The slenderness ratio depends only on the dimensions of the column. A column
respectively. that is long and slender will have a high slenderness ratio and therefore a low
critical stress. A column that is short and stubby will have a low slenderness
Although the radius of gyration of an area does not have an obvious physical
ratio and will buckle at a high stress. Typical values of slenderness ratio for
meaning, it may be consider to be the distance, from the reference axis, at
actual columns are between 30 and 150.
which the entire area could be concentrated and still have the same moment
of inertia as the original area. The critical stress is the average compressive stress on the cross section at the
instant the load reaches its critical value. It is possible to plot a graph of
this stress as a function of the slenderness ratio and obtain a curve known as
Euler’s curve shown in figure 7.8. The curve shown in the figure is plotted
7.6 Critical stress for a structural steel with E = 30 × 103 ksi. It should be noted that the curve
is valid only when the critical stress is less than the proportional limit of the
For purposes of design, the equation for the critical load can be written in steel (for the steel plotted 36 ksi), because the equations were derived using
a more useful form by dividing the load by the cross-sectional area. For the Hooke’s law.
c
7.6 Critical stress 165 166 Columns
Pcr
σcr (ksi) 50
3 in
Euler’s curve
37.5
24 ft
PSfrag replacements 2.75 in

25
12.5
Pcr

PSfrag replacements 0
0 50 100 150 200 250
L/r can support without buckling. It is always a good idea check the average
compressive stress in the column, to verify that the material has not yielded
and the Euler’s equation is still valid.
Figure 7.8: Graph of Euler’s curve for structural steel with E = 30 × 103 ksi. A force of 64.5 kip creates an average compressive stress in the column of
Example 7.1 Pcr 64.5

σcr = = = 14.3 ksi (b)
A π(3)2 − π(2.75)2
GIVEN A 24 ft long A–36 steel tube having the cross section shown in figure
7.9 is to be used as a pin-ended column. Since σcr < σY = 36 ksi, application of the Euler’s equation is appropriate.
FIND Determine the maximum allowable axial load the column can support
so that it does not buckle.
Example 7.2
Solution
Using equation (7.14) to obtain the critical load with E = 29 × (103 ) ksi, GIVEN The A–36 steel W 8×31 member shown in figure 7.10 is to be used
as a pin-connected column.
FIND Determine the largest axial load it can support before it either begins
π 2 EI π 2 [29(103)] 14 π(3)4 − 14 π(2.75)4

Pcr = to buckle or the steel yields.
L2 [24(12)]2
= 64.5 kip (a)
Hence, a force of 64.5 kip is the maximum allowable axial force the column
c
7.6 Critical stress 167 168 Columns
Solution
From the tables, the column’s cross-sectional area and moments of inertia are
A = 9.13 in2 , Ix = 110 in4 , and Iy = 37.1 in4 . By inspection, buckling will
occur about the y–y axis. Applying equation (7.14),
π 2 EI π 2 [29(103)](37.1)
Pcr = = = 512 kip (a)
L2 [(12)(12)]2
When fully loaded, the average compressive stress in the column is
Pcr 512
σcr = = = 56.1 ksi (b)
A 9.13
x Since this stress exceeds the yield stress (36 ksi), the load P is determined
from simple compression:
P
12 ft y y 36 ksi = P = 329 kip (c)
9.13
PSfrag replacements
x 7.7 Optimum shapes of columns
With few exceptions, columns have the same cross sections throughout their
Figure 7.10: Problem 7.2. lengths. Therefore, only prismatic columns are analyzed here. Nevertheless,
prismatic columns are not the optimum shape if minimum weight is desired
(prismatic columns are usually the most cost-effective solution). The critical
load for a certain amount of material may be increased by varying the shape
so that the column has larger cross sections in those regions where the bending
moments are larger.
Consider a column of constant solid circular cross sections with pinned ends.
A column shaped as shown in figure 7.11a will have a larger critical load than
a prismatic column made from the same amount of material. To approximate
this optimum shape, prismatic columns are sometimes reinforced over part of
their lengths as illustrated in figure 7.11b.
Consider now a prismatic column with pinned ends that is free to buckle in
any lateral direction (figure 7.11c). Also, assume that the column has a solid
c
7.8 Columns with other support conditions 169 170 Columns
P P P differential equation of the deflection curve by following the same procedure

used to analyze a pinned-end column.
B B B
Consider now the case of a column fixed at its base and free at the top as
shown in figure 7.12a. The first step towards obtaining the critical load is to
assume that the column is in the buckled state in order to obtain an expression
for the bending moment in the column. From the free-body diagram in figure
7.12b, the internal moment at an arbitrary section is M = P (δ − v).
The second step is to set up the differential equation for the deflection curve.
eplacements For the case under consideration,
d2 v
EI = P (δ − v) (7.21)
dx2
A A A
The previous equation can be rewritten in a more convenient way as
(a) (b) (c) (d)
d2 v P P
Figure 7.11: (a,b) Nonprismatic columns. (c) Prismatic column. (d) Some + v= δ (7.22)
dx2 EI EI
possible cross-sectional shapes for prismatic columns.
which is a nonhomogeneous, second-order, linear differential equation with
cross section of regular shape as shown in figure 7.11d. An interesting question constant coefficients.
arises: assuming that the critical load is calculated from the Euler formula,
The third step is to obtain the general solution of the differential equation.
which cross section gives the largest critical load?
Since in this case the differential equation is nonhomogeneous, the general
It can be demonstrated that the critical load for an equilateral triangle is higher solution has two parts: (1) the homogeneous solution, which is the solution
than the critical loads of the other shapes (it gives, for example, a critical of the homogeneous equation obtained by replacing the right-hand side with
load 21% higher than does a circular cross section). Hence, an equilateral zero, and (2) the particular solution, which is the solution of equation (7.22)
triangle is the optimum shape for a cross section based only upon theoretical that produces the term on the right-hand side.
considerations.
The homogeneous solution is the same as the solution of equation (7.3). Hence,
r ! r !
P P
7.8 Columns with other support conditions vH = C1 sin
EI
x + C2 cos
EI
x (7.23)
Buckling of a column with pinned ends is usually considered as the most basic where C1 and C2 are constant of integration. Note that when vH is substituted
case of buckling. However, in practice we encounter many other support condi- into the left-hand side of the differential equation (7.22) it produces zero.
tions, such as fixed ends, free ends and elastic supports. The critical loads for
columns with various kinds of support conditions can be determined from the The particular solution of the differential equation is
c
7.8 Columns with other support conditions 171 172 Columns
P P x P P
δ δ C2 = −δ (7.27)
x The application of the second boundary condition v 0 (0) = 0 can be also easily
v applied from
v
g replacements
M r r ! r r !
dv P P P P
L P = C1 cos x − C2 sin x (7.28)
dx EI EI EI EI
giving C1 = 0.
The final step is to substitute the values of the constants to obtain the deflec-
(a) (b) tion curve and the critical load. Therefore, here the deflection curve is

" r !#
Figure 7.12: (a) Ideal column fixed at the base and free at the top. (b) Free P
v = δ 1 − cos x (7.29)
body diagram of a section of the column. EI
The above equation gives only the shape of the deflection curve as the am-
vP = δ (7.24) plitude δ remains undefined. Thus, when the column buckles, the deflection
given by equation (7.29) may have any arbitrary magnitude, except that it
must remain small because the differential equation is based upon small de-
When vP is substituted into the left-hand side of the differential equation, it
flections.
produces the right-hand side, that is, it produces the term (P/EI)δ. With
these results, the general solution of the equation, equal to the sum of vH and A third boundary condition can be obtained since the deflection at the top of
vP is, the column is δ, that is, at x = L, v = δ. Using this condition with equation
(7.20) gives
r ! r !
P P
v = C1 sin x + C2 cos x +δ (7.25) r !
EI EI P
δ cos L =0 (7.30)
EI
The fourth step is to apply boundary conditions pertaining to the deflection
v and the slope v 0 to obtain the values of the constants. For the case under p
In order to fulfill the previous equation, either δ = 0 or cos(L P/EI) = 0.
consideration, the constants can be determined from the boundary conditions:
The trivial solution δ = 0 indicates that no buckling occurs, regardless of the
load P . The other possibility is that
dv
v(0) = 0 (0) = v 0 (0) = 0 (7.26)
dx r ! r
P P nπ
cos L =0 or L= n = 1, 3, 5, . . . (7.31)
Applying the first condition v(0) = 0 yield EI EI 2
c
eplacements7.8 Columns with other support conditions 173 174 Columns
P x Pcr x Pcr x Pcr P P

δ
L
L
5
v 3 δ
δ
L = Le
L
δ 25π 2 EI
Pcr = (a) (b)
π 2 EI 9π 2 EI 4L2
Pcr = Pcr = Le = 2L

4L2 4L2

L L
3 5
PSfrag replacements

(a) (b) (c) (d)
Figure 7.13: Different buckled shapes for an ideal column fixed at the base
and free at the top: (a) initially straight column, (b) buckled shape for n = 1,
(c) buckled shape for n = 3 and (d) buckled shape for n = 5.
Figure 7.14: Deflection curves showing the effective length Le for a column
fixed at the base and free at the top.
The smallest critical load occurs when n = 1, so that
7.9 Effective length

π 2 EI
Pcr = (7.32) The critical loads for columns with various support conditions can be related
4L2
to the critical load of a pinned-end column through the concept of effective
length.
By comparison with equation (7.14), it is seen that a column fixed-supported To demonstrate this idea, consider figure 7.14a where the deflected shape of a
at its base will carry only one-fourth the critical load than can be applied to a pinned-end column is shown. Now, consider a column that is fixed at its base
pin supported column. The critical loads and the buckled shapes for the cases and free at the top buckles in a curve that is one-quarter of a complete sine
when n = 1, n = 3 and n = 5 are shown in figure 7.13. (or cosine) wave. If its deflection curve is extended (figure 7.14b), it becomes
one half of a complete sine wave, which is the deflection curve for a pinned-end
Other type of columns are analyzed in much the same way and they will not column.
be covered in detail here. However, the critical loads for various types of
supported columns will be mentioned in what follows. The effective length Le for any column is the length of the equivalent pinned-end
c
7.9 Effective length 175 176 Columns
column. That is, it is the length of a pinned-end column having a deflection fixed-pinned column (figure 7.15d) has an inflection point at approximately
curve that exactly matches all or part of the deflection curve of the original 0.699L from its pinned end, so that Le = 0.699L.
column. Another way of expressing this idea to say that the effective length of
a column is the distance between points of inflection (points of zero moment)
in its deflection curve, assuming that the curve is extended until points of Example 7.3
inflection are reached. Thus, for a fixed-free column (figure 7.14b), the effective
length is
GIVEN The roof of a train station (figure ??) is supported by a row of
aluminum pipe columns having length L = 3.25 m and outer diameter d = 100
Le = 2L (7.33) mm. The bases of the columns are set in concrete footings and the tops of the
columns are supported laterally by the roof. The columns are being designed
Since the effective length is the length of an equivalent pinned-end column, it to support compressive loads due to the accumulation of snow of P = 100 kN.
is possible to write a general formula for critical loads as follows: FIND Determine the minimum required thickness t of the columns if a factor
of safety Sf = 3 is required with respect to Euler buckling.
π 2 EI
Pcr = (7.34) Solution
L2e
The first step towards the solution of the problem is recognize which support
conditions are holding the columns. On the bottom they are fixed as they are
Thus, if the effective length of a column is known, it is possible to substitute
set in concrete footings. On the top, since they are supported only laterally
it into the above equation and determine the critical load. For instance, in the
by the roof they can be considered to be pin-connected. Hence, a fixed-pinned
case of a fixed-free column, substituting Le = 2L gives equation (7.32).
column should be considered. From figure 7.15 the critical load is given by
Rather than specifying the effective length of a column, many design codes
provide column formulas that employ a dimensionless coefficient K called the 2.046π 2EI
effective-length factor. K is defined from Pcr = (a)
L2
where for a tubular cross section the moment of inertia I is:

Le = KL (7.35)
where L is the actual length of the column. Thus, the critical load of the π 4
d − (d − 2t)4

I= (b)
column is 64
From the requirements of the problem, the load that the column must support
π 2 EI is 100 kN with a safety factor of 3. Hence,
Pcr = (7.36)
(KL)2
Figure 7.15 shows the lowest critical loads and corresponding effective lengths Pcr = Sf (P ) = 3(100kN) = 300kN (c)
for an ideal column under four different support conditions, including the two
analyzed here. The fixed-fixed column (figure 7.15c) has inflection points or The three previous equations can be combined to get an equation which only
points of zero moment L/4 from each support. The effective length is therefore unknown is the required thickness t. Considering that for aluminum E = 72
represented by the middle half of its length, that is, Le = 0.5L. Finally, the GPa, and doing all the numerical substitutions it is possible to write
c
Solution
The buckling behavior of the column will be different about the x and y axes
2.046π 2(72 × 109 Pa) π due to the bracing, so two cases should be considered: buckling in the x–x axis
(0.1 m4 − (0.1 m − 2t)4

300, 000 N = 2
(d)
3.25 m 64 and buckling in the y–y axis.
After evaluating all the numerical terms, the above equation simplifies to Consider now the x–x case. As before, the first step is to recognize how the
column is supported. On this case there are no major problem as the column
is fixed at both ends and the bracing is not designed to prevent buckling in this
44.60 × 10−6 m4 = (0.1 m)4 − (0.1 m − 2t)4 (e) case. From figure 7.15 the effective length for this case will be (KL)x = 0.5(24)
ft. Thus, from equation (7.36) the critical load for the column is:
or, taking the fourth root on both sides and passing (0.1 m4 to the LHS,
π 2 EIx π 2 (29 × 103 ksi)29.1 in4
Pcrx = = = 401.7 kpi (a)
(KL)2x (144 in)2
0.1 m − 2t = 0.08635 m (f)
Next, consider now the y–y case. As a consequence of the bracing, which
which gives a value of t = 0.006825 m or t = 6.83 mm. forces the deflection at the mid-length of the column to be zero, this case can
be thought of as two columns that are fixed at one end and pinned at the other
As mentioned before, it is always a good idea to check for yielding. The critical (remember that the column is connected to the struts by pins). Furthermore,
stress in the column must be less than the proportional limit of the aluminum the compressive load that the column is supporting is the same before and after
(480 MPa). The critical stress for this column is: the struts which will give the same buckling behavior on both parts. Hence, it
is needed only to consider the buckling of one of those two “virtual” columns
Pcr 300 kN having half the length of the complete column.
σcr = = = 150 MPa (g)
A 1999mm2
The effective length of a column that is fixed at one end and pin supported at
Since the critical stress is less than the proportional limit, the calculation for the other is 0.699. For this case, (KL)y = 0.699(12) ft and the critical load
the column using Euler buckling theory is satisfactory. can be calculated as
π 2 EIy π 2 (29 × 103 ksi)9.32 in4

Pcry = = = 262.5 kpi (b)
Example 7.4 (KL)2y (100.7 in)2
GIVEN A W 6×15 steel column is 24 ft long and is fixed at its ends as Comparing both critical loads, buckling will occur about the y–y axis.
shown in figure ??. Its load-carrying capacity is increased by bracing it about
The area of the cross section is 4.43 in2 , so the average compressive stress in
the y–y (weak) axis using struts that are assumed to be pin-connected to its
the column will be
mid-height.
FIND Determine the load it can support so that the column does not buckle Pcr 262.5 kip
nor the material exceeds the yield stress. The material properties are E = σcr = = = 59.3 ksi (c)
A 4.43 in2
29 × 103 ksi and σY = 60 ksi. The moments of inertia for the cross section are
Ix = 29.1 in4 and Iy = 9.32 in4 . Since this stress is less than the yield stress (60 ksi), buckling will occur before
c
the material yields. Thus,
Pcr = 263 kip (d)
π 2 EI π 2 EI 4π 2 EI 2.046π 2EI
Pcr = Pcr = Pcr = Pcr =
L2 4L2 L2 L2
PSfrag replacements

L

Le = L Le = 2L Le = 0.5L Le = 0.699L
K=1 K=2 K = 0.5 K = 0.699
(a) (b) (c) (d)
Figure 7.15: Critical loads, effective lengths and effective-length factors for
ideal columns. (a) Pinned-pinned column. (b) Fixed-free column. (c) Fixed-
fixed column. (d) Fixed-pinned column.
c
7.9 Effective length 181
t
PSfrag replacements

12ft
24ft
12ft
frag replacements

x–x axis buckling y–y axis buckling

Figure 7.17: A braced column.
c

m00842 Class Notes2

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

m00842 Class Notes2

Uploaded by

Copyright:

Available Formats

ii

Instituto Tecnológico y de Estudios Superiores de Monterrey Spring 2002

Chapter 3 is devoted to the method of superposition. The principles of super-

Continuous beams are presented and analized in chapter 4. The method of

In chapter 5 the study of mechanical components subject to static loading is

2.3.2 Beam with symmetric loading . . . . . . . . . . . . . . . 22

2.3.3 Beam with non–symmetric loading . . . . . . . . . . . . 27

Contents 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3.2 Principle of superposition . . . . . . . . . . . . . . . . . . . . . 33

3.3 Principle of proportionality . . . . . . . . . . . . . . . . . . . . . 33

3.4 Proportionality and superposition . . . . . . . . . . . . . . . . . 33

1 Deflection of beams by the integration method 5

4.2.5 Example 4.1 . . . . . . . . . . . . . . . . . . . . . . . . . 47

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 5 Mechanical components subject to static loading 49

2.3.1 Cantilever beam . . . . . . . . . . . . . . . . . . . . . . . 20 5.5 Strain transformation . . . . . . . . . . . . . . . . . . . . . . . . 55

5.5.1 Principal strains . . . . . . . . . . . . . . . . . . . . . . 58 7.3 Ideal column with pin supports . . . . . . . . . . . . . . . . . . 83

5.5.2 Strain rosettes . . . . . . . . . . . . . . . . . . . . . . . . 58 7.4 Critical loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

5.6 Hooke’s law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 7.5 Radius of gyration . . . . . . . . . . . . . . . . . . . . . . . . . 86

5.7 Failure criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 7.6 Critical stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

5.7.1 Rankine criterion . . . . . . . . . . . . . . . . . . . . . . 62 7.7 Optimum shapes of columns . . . . . . . . . . . . . . . . . . . . 88

5.7.2 Tresca criterion . . . . . . . . . . . . . . . . . . . . . . . 62 7.8 Columns with other support conditions . . . . . . . . . . . . . . 89

5.7.3 Von Mises criterion . . . . . . . . . . . . . . . . . . . . . 63 7.9 Effective length . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

5.7.4 Mohr-Coulomb criterion . . . . . . . . . . . . . . . . . . 64

6.2 Principles of strain energy . . . . . . . . . . . . . . . . . . . . . 66

6.3 Impact loading . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

6.4 Strain energy density . . . . . . . . . . . . . . . . . . . . . . . . 68

6.4.1 Toughness and resilience modulus . . . . . . . . . . . . . 69

6.5 Elastic strain energy for normal stresses . . . . . . . . . . . . . 70

6.6 Strain energy in bending . . . . . . . . . . . . . . . . . . . . . . 74

6.7 Elastic strain energy for shear stress . . . . . . . . . . . . . . . . 74

6.8 Strain energy in torsion . . . . . . . . . . . . . . . . . . . . . . . 75

6.9 Deflection of beams due to impact load . . . . . . . . . . . . . . 76

7.2 Stable, neutral and unstable equilibrium . . . . . . . . . . . . . 82

Thomas Young was born into “comfortable circumstances” at Milverton, Eng-

1. The beam is initially straight.

2. The beam is elastically deformed by load applied perpendicular to the

A 3. The load is lying in the x-v plane.

caused by shear is ignored.

1.2 Moment-curvature relationship

which is a linear second-order differential equation. +M +V +V +M

FIND Determine the equation of the elastic curve. EI is constant.

Slope and elastic curve. Applying Solution II

Figure 1.6: Problem 1.1. an the solution continues as before.

PSfrag replacements P for 0 ≤ x1 < 2a

Example 1.3 d2 v −w0 3 w0 L

1.4 Procedure for analysis

Siméon Denis Poisson (1781–1840)

Originally forced to study medicine, Siméon Poisson began to study mathe-

In Recherchés sur la probabilité des jugements, an important work on proba-

where θC and θD denote, respectively, the slope at C and the slope at D.

Figure 2.1: Shear, moment and M/EI diagrams.

• tC/D is the tangential deviation of C with respect to D.

Now it is possible to observe that (M/EI)dx represents an area below the

Second moment–area theorem:

tC/D = area below the M/EI diagram between C and D × x¯1 .

tD/C = area below the M/EI diagram between C and D × x¯2 .

tD/C 2.3.1 Cantilever beam

Figure 2.8: Beam with symmetric loading.

2.3.2 Beam with symmetric loading

In the case of a simply supported beam AB subject to a symmetric loading, or B

vD = tB/C − tD/C (2.13)

From the second moment–area theorem, tB /A is equal to the area between

Since the transformation equations were derived solely from equi-

to x dx where dx is the length of the element. As a result of this elongation. y

Similary, the strain y produces a counterclockwise rotation of the diagonal