Ferdinand L.

Singer/Andrew Fyiei

Strength of Materials
SI Units
(Systeme International d'Unites)

A. SELECTED SI UNITS

Quantity Name SI Symbol

Energy joule J(l J = 1 Nm)
Force newton N(l N= 1 kg-m/s 2)
Length meter* m
Mass kilogram* kg
Moment (torque) newton meter N-m
Plane angle radian rad
o
degree
Rotational frequency revolution r/s
per second
Stress (pressure) pascal Pa (1 Pa = 1 N/m 2 )

Temperature degree Celsius °C

Time second* s

Power watt W W= (1 1 J/s)

B. COMMONLY USED SI PREFIXES

Multiplying Factor Prefix SI Symbol
9
10 giga G
10
6
mega M
3
10 kilo k
10- 3
milli m
10" 6
micro /*

10- 9
nano n

•SI h»c# unit

 ix7e>i

C. SELECTED RULES AND SUGGESTIONS FOR SI USAGE

1 Care must be taken to use the correct case for symbols, units,
and prefixes (e.g., m for meter or milli, M for mega).
2. For numbers having five or more digits, the digits should be
placed in groups of three separated by a space, counting both
to the left and to the right of the decimal point (e.g.,

61 354.982 03). The space is not required for four-digit

numbers. Spaces are used instead of commas to avoid con-
fusion —
many countries use the comma as the decimal
marker.
3. In compound units formed by multiplication, use the product
dot (e.g., N-m).
4. Division may be indicated by a solidus (m/s), or a negative
-1
exponent with a product dot (m-s ).
5. Avoid the use of prefixes in the denominator (e.g.,- km/s is

preferred over m/ms). The exception to this rule is the prefix

k in the base unit kg (kilogram).
Strength of Materials
 THIRD EDITION

Strength of Materials

Ferdinand L. Singer
EMERITUS, NEW YORK UNIVERSITY

Andrew Pytel
THE PENNSYLVANIA STATE UNIVERSITY

HARPER & ROW, PUBLISHERS, New York

Cambridge, Hagerstown, Philadelphia, San Francisco,
London, Mexico City, Sao Paulo, Sydney
1817
Sponsoring editor: Charlie Dresser
Project Editor: Celine Keating
Production Manager: Marion A. Palen
Compositor: Science Typographers Incorporated
Printer and Binder: The Maple Press Company
Art Studio: J & R Services
Cover Design: Helen Iranyi

STRENGTH OF MATERIALS
Copyright © 1980 by Ferdinand L. Singer and Andrew Pytel

All rights reserved. Printed in theUnited States of America. No part of

thisbook may be used or reproduced in any manner whatsoever without written
permission except in the case of brief quotations embodied in critical articles
and reviews. For information address Harper & Row, Publishers, Inc., 10 East
53rd Street, New York, N.Y. 10022.

Library of Congress Cataloging in Publication Data

Singer, Ferdinand Leon, Date-
Strength of materials.
Includes index.
1 . Strength of materials. I. Pytel, Andrew,
joint author. II. Title.
TA405.S45 1980 620.T1 79-18222
ISBN 0-06-046229-9
To Evelyn, Joan, Karen, Lucy,
Leslie, Lori, and John
 Contents

Preface xv
List of Symbols and
Abbreviations XIX

chapter 1 SIMPLE STRESS

1-1 INTRODUCTION 1

1-2 ANALYSIS OF INTERNAL FORCES

1-3 SIMPLE STRESS 5
1-4 SHEARING STRESS 16
1-5 BEARING STRESS 20
1-6 THIN- WALLED CYLINDERS 24

chapter 2 SIMPLE STRAIN

2-1 INTRODUCTION 33
2-2 STRESS-STRAIN DIAGRAM 33
2-3 HOOKE'S LAW: AXIAL
DEFORMATION 38
2-4 POISSON'S RATIO: BIAXIAL AND
TRIAXIAL DEFORMATIONS 47

vll
 Contents

2-5 STATICALLY INDETERMINATE

MEMBERS 50
2-6 THERMAL STRESSES 62

chapter 3 TORSION

3-1 INTRODUCTION AND

ASSUMPTIONS 74
3-2 DERIVATION OF
TORSION FORMULAS 75
3-3 FLANGED BOLT COUPLINGS 87
3-4 LONGITUDINAL SHEARING
STRESS 90
3-5 TORSION OF THIN- WALLED TUBES;
SHEAR FLOW 91
3-6 HELICAL SPRINGS 95

chapter 4 SHEAR AND MOMENT IN BEAMS

4-1 INTRODUCTION 106
4-2 SHEAR AND MOMENT 108
4-3 INTERPRETATION OF VERTICAL SHEAR
AND BENDING MOMENT 124
4-4 RELATIONS BETWEEN LOAD,
SHEAR, AND MOMENT 125
4-5 MOVING LOADS 147

chapter 5 STRESSES IN BEAMS

5-1 INTRODUCTION 153

5-2 DERIVATION OF FLEXURE
FORMULA 154
5-3 ECONOMIC SECTIONS 167
5-4 FLOOR FRAMING 172
5-5 UNSYMMETRICAL BEAMS 176
5-6 ANALYSIS OF FLEXURE ACTION 184
5-7 DERIVATION OF FORMULA
FOR HORIZONTAL
SHEARING STRESS 188
5-8 DESIGN FOR FLEXURE
AND SHEAR 200
5-9 SPACING OF RIVETS OR BOLTS
IN BUILT-UP BEAMS 206
Contents lx

chapter 6 BEAM DEFLECTIONS

6-1 INTRODUCTION 213
6-2 DOUBLE-INTEGRATION
METHOD 214
6-3 THEOREMS OF AREA-MOMENT
METHOD 227
6-4 MOMENT DIAGRAMS BY PARTS 231
6-5 DEFLECTION OF CANTILEVER
BEAMS 241
6-6 DEFLECTIONS IN SIMPLY
SUPPORTED BEAMS 248
6-7 MIDSPAN DEFLECTIONS 260
6-8 CONJUGATE-BEAM METHOD 265
6-9 DEFLECTIONS BY THE
METHOD OF SUPERPOSITION 269

chapter 7 RESTRAINED BEAMS

7-1 INTRODUCTION 283

7-2 REDUNDANT SUPPORTS IN PROPPED
AND RESTRAINED BEAMS 283
7-3 APPLICATION OF DOUBLE-
INTEGRATION AND
SUPERPOSITION METHODS 284
7-4 APPLICATION OF AREA-MOMENT
METHOD 293
7-5 RESTRAINED BEAM EQUIVALENT
TO SIMPLE BEAM
WITH END MOMENTS 303
7-6 DESIGN OF RESTRAINED BEAMS 306

chapter 8 CONTINUOUS BEAMS

8-1 INTRODUCTION 312
8-2 GENERALIZED FORM OF THE THREE-
MOMENT EQUATION 313
8-3 FACTORS FOR THE THREE-MOMENT
EQUATION 317
8-4 APPLICATION OF THE THREE-MOMENT
EQUATION 322
8-5 REACTIONS OF CONTINUOUS BEAMS;
SHEAR DIAGRAMS 328
8-6 CONTINUOUS BEAMS WITH
FIXED ENDS 333
 Contents

8-7 DEFLECTIONS DETERMINED BY THE

THREE-MOMENT EQUATION 340
8-8 MOMENT DISTRIBUTION 346

chapter 9 COMBINED STRESSES

9-1 INTRODUCTION 358
9-2 COMBINED AXIAL AND
FLEXURAL LOADS 359
9-3 KERN OF A SECTION; LOADS APPLIED
OFF AXES OF SYMMETRY 369
9-4 VARIATION OF STRESS WITH
INCLINATION OF ELEMENT 373
9-5 STRESS AT A POINT 375
9-6 VARIATION OF STRESS AT A POINT:
ANALYTICAL DERIVATION 376
9-7 MOHR'S CIRCLE 379
9-8 APPLICATIONS OF MOHR'S CIRCLE TO
COMBINED LOADINGS 389
9-9 TRANSFORMATION OF STRAIN
COMPONENTS 402
9-10 THE STRAIN ROSETTE 409
9- 1 1 RELATION BETWEEN MODULUS
OF RIGIDITY AND MODULUS
OF ELASTICITY 413

chapter 10 REINFORCED BEAMS

10-1 INTRODUCTION 416

10-2 BEAMS OF DIFFERENT
MATERIALS 417
10-3 SHEARING STRESS AND DEFLECTION
IN COMPOSITE BEAMS 423
10-4 REINFORCED CONCRETE BEAMS 424
10-5 DESIGN OF REINFORCED
CONCRETE BEAMS 430
10-6 TEE BEAMS OF REINFORCED
CONCRETE 433
10-7 SHEARING STRESS AND
BOND STRESS 436
 1

Contents

chapter 11 COLUMNS
11-1 INTRODUCTION 439
11-2 CRITICAL LOAD 44
1 1 -3 LONG COLUMNS BY EULER'S
FORMULA 442
1 1 -4 LIMITATIONS OF EULER'S
FORMULA 447
11-5 INTERMEDIATE COLUMNS.
EMPIRICAL FORMULAS 451
1 1 -6 ECCENTRICALLY LOADED
COLUMNS 464
1 1 -7 THE SECANT FORMULA 469

chapter 12 RIVETED AND WELDED CONNECTIONS

12-1 INTRODUCTION 474

12-2 TYPES OF RIVETED JOINTS:
DEFINITIONS 474
12-3 STRENGTH OF A SIMPLE
LAP JOINT 476
12-4 STRENGTH OF A COMPLEX
BUTT JOINT 478
12-5 STRESSES IN RIVETED JOINTS 485
12-6 STRUCTURAL RIVETED JOINTS 487
12-7 ECCENTRICALLY LOADED RIVETED
CONNECTIONS 490
12-8 WELDED CONNECTIONS 497
12-9 ECCENTRICALLY LOADED WELDED
CONNECTIONS 502

chapter 13 SPECIAL TOPICS

13-1 INTRODUCTION 510

13-2 REPEATED LOADING; FATIGUE 510
13-3 STRESS CONCENTRATION 512
13-4 THEORIES OF FAILURE 516
13-5 ENERGY METHODS 519
13-6 IMPACT OR DYNAMIC LOADING 528
13-7 SHEARING STRESSES IN THIN-WALLED
MEMBERS SUBJECTED TO BENDING;
SHEAR FLOW 534
xll Contents

13-8 SHEAR CENTER 537

13-9 UNSYMMETRICAL BENDING 546
13-10 CURVED BEAMS 554
13-11 THICK-WALLED CYLINDERS 564

chapter 14 INELASTIC ACTION

14-1 INTRODUCTION 570

14-2 LIMIT TORQUE 571
14-3 LIMIT MOMENT 573
14-4 RESIDUAL STRESSES 577
14-5 LIMIT ANALYSIS 585

appendix A MOMENTS OF INERTIA

A-l DEFINITION OF MOMENT

OF INERTIA 597
A-2 POLAR MOMENT OF INERTIA 599
A-3 RADIUS OF GYRATION 600
A-4 TRANSFER FORMULA FOR
MOMENT OF INERTIA 601
A-5 MOMENTS OF INERTIA
BY INTEGRATION 602
A-6 MOMENTS OF INERTIA FOR
COMPOSITE AREAS 607
A-7 PRODUCT OF INERTIA 617
A-8 PRODUCT OF INERTIA IS ZERO
WITH RESPECT
TO AXES OF SYMMETRY 618
A-9 TRANSFER FORMULA FOR PRODUCT
OF INERTIA 619
A- 10 MOMENTS OF INERTIA WITH RESPECT
TO INCLINED AXES 625
A- 1 1 MOHR'S CIRCLE FOR MOMENTS OF
INERTIA 627
A-12 MAXIMUM AND MINIMUM MOMENTS OF
INERTIA: PRINCIPAL AXES 630

appendix B TABLES

B-l AVERAGE PHYSICAL PROPERTIES OF

COMMON METALS 635
Contents xill

B-2 PROPERTIES OF WIDE-FLANGE

SECTIONS (W SHAPES) 636
B-3 PROPERTIES OF I-BEAM
SECTIONS (S SHAPES) 644
B-4 PROPERTIES OF CHANNEL
SECTIONS 646
B-5 PROPERTIES OF EQUAL
ANGLE SECTIONS 648
B-6 PROPERTIES OF UNEQUAL
ANGLE SECTIONS 650

Index 653
 Preface

Successful machine or structural design is practically impossible without

a thorough mastery of engineering mechanics and strength of materials.
In the modern engineering curricula, the importance of these subjects is
emphasized by a comprehensive study of their fundamental and
advanced concepts. This revised SI* edition attempts to explain lucidly
and rigorously the theory and application of these concepts.
During the 1960s, engineering education in this country became
increasingly mathematically oriented. In some instances, courses in
strength of materials were even replaced by courses such as continuum
mechanics and elasticity theory. The current philosophy of engineering
education recognizes the importance of design in the engineering
curriculum. Therefore, applied courses such as strength of materials
again emerge as important components in the education of engineers in
virtually all areas of speciality.
The unique feature of the third edition of this text, as compared
with the previous editions, is that it is written completely in SI units.
Although we are currently going through a transition period when the
practicing engineer may need to be familiar with more than one system
of units, the worldwide movement to SI units will probably make this
transition period short-lived. The authors agree with the philosophy that

*SI is the official abbreviation for the international system of units, Le Systeme
International d'Unites.

XV
xvl Preface

fundamental texts such as this one should be written completely in SI

units. Once the fundamentals have been thoroughly mastered in one
system of units, extensions to other systems should not be difficult.
This edition retains the general plan and features of the earlier
editions. The major emphasis is still on elastic analysis, although an
extensive coverage of inelastic analysis is included. After much consid-
was again incorporated into a coordinated
eration, inelastic analysis
continuous treatment of residual stress and limit analysis. As in the
second edition, a modernized exposition of the double-integration
method is given which greatly broadens and simplifies its application to
the deflection of beams. Also, energy methods, such as Castigliano's
theorem and virtual work, have been retained so that, with the double-
integration and area-moment methods, there is an exceptionally com-
plete discussion of deflection methods. There is also a short but effective
presentation of moment distribution embodying a modernized sign
treatment which should simplify its application. Each of these topics is
fairly independent of the others; thus one or more may be omitted if
time is limited.
Other features of this text include an expanded treatment of shear
flow; an extended analysis of the states of stress and strain with
emphasis on the application of Mohr's circle to strain gage and rosette
analysis; and relegation of riveted and welded connections to the latter
part of the book, where this subject does not interrupt the continuity of
basic principles. In addition, revisions in design codes have necessitated
updating the discussion of column theory.
These topics, as well as the others in the text, are presented in a
manner burden of detailed
that should relieve the instructor of the
explanation. Principles are developed by a consistent plan which first
relates stresses to deformations, then applies the equations of static
equilibrium, and finally satisfies the boundary conditions.
Primarily, the point of view of students has been retained and
their special problems kept in mind. We have made every effort to make
a fundamental principle perfectly understood, but in clear, concise
language. The physical significance of fundamental concepts, and the
assumptions and limitations made in developing them, are carefully
discussed so that memorization is reduced to a minimum. The
summaries appended to most chapters are intended to give the student a
concise statement of key elements which should be useful in review and
postcollege work. Rules of sign have also been simplified by assigning
the positive sense to all quantities to which the adjectives up, above, or
similar terms may be applied; for negative signs, the converse is true.
Numerous illustrative problems show in detail how principles are
applied. —
The explanations are complete nothing is taken for granted.
Throughout, the equation or principle to be applied is stated in brackets
Preface xvli

on the left side of the equation. In the solution, values are substituted in
the respective order in which the symbols appear in the equation. This
procedure enables the reader to follow readily the various steps of the
solution without continually referring to the body of the text.
The almost 1000 problems contained in this edition are either SI
versions of problems from the previous edition or completely new
problems. In either case, they have been carefully chosen so as to
illustrate the fundamental concepts without overburdening the student
with tedious numerical computation. The importance of free-body dia-
grams in the solution of problems in strength of materials continues to
be emphasized. The problems have been arranged approximately in the
order of their difficulty, and answers to two-thirds of them have been
given; the others may on occasion be used for quizzes.
The numbering plan used enables the reader to locate quickly any
cross reference. With this plan, all articles, figures, equations, tables,
and problems are preceded by the numeral of the chapter in which they
appear and are numbered consecutively throughout each chapter. Fig-
ures for assigned problems are given thejiumber of the problem to
which they refer in order to simplify correlation of a problem figure
with corresponding problem data.
The authors wish to acknowledge their indebtedness to their
colleagues all over the nation for their many valuable suggestions for
this edition. To identify them individually would make too lengthy a list

(with possibly an inadvertent omission), but individually each has

received our thanks. However, a special debt is due to Dr. Jean Landa,
whose assistance in the preparation of this edition is greatly appreciated.
Although great care was taken to eliminate errors, it is inevitable that
some will still be found. The authors appreciate being informed about
these and welcome any comments that readers may care to offer.

Ferdinand L. Singer
Andrew Pytel
 List of Symbols
and Abbreviations

A area
A' partial area of beam section
a b 9 coordinates of centroid of moment diagram caused by
simply supported loads
b breadth, width
c distance from neutral axis to extreme fiber
D,d diameter
E modulus of elasticity in tension or compression
e eccentricity, natural base of logarithms

f frequency
fc unit compressive stress in concrete

f, unit tensile stress in reinforcing steel

G modulus of rigidity (i.e., modulus of elasticity in shear)

g gravitational acceleration (9.81 m/s2)

h height, depth of beam
I moment of inertia of area
w moment of inertia with respect to neutral axis

xlx
xx Symbols and Abbreviations

I centroidal moment of inertia

J polar moment of inertia
J centroidal polar moment of inertia
K stress concentration factor
k spring constant, radius of gyration
L length
Le effective length forcolumns
M bending moment
m mass
TV normal force, factor of safety
n ratio of moduli of elasticity
P force, concentrated load, hoop tension
9 power
PCT critical load for columns
^uc Pxy products of inertia
p pressure per unit area
Q statical or first moment of area
q shear flow
R reaction, resultant force, radius
r radius, radius of gyration
S section modulus (I/c)
o unit stress, normal stress
ob unit bearing stress
oc unit compressive stress
a cr critical unit stress in column formula
af unit flexural stress
ar unit radial stress
ow allowable stress
a, unit tensile stress, unit tangential stress
ax ,oyi az unit normal stress in jc, >>, and z directions, respectively

(jyp stress at yield point

T torque, temperature
/ thickness, tangential deviation
t unit shearing stress
t^ unit shearing stress in x-y plane
u,i?,h> rectangular coordinates
V vertical shearing force
t; velocity
W total weight or load
>v weight or load per unit of length
X, y, Z orthogonal components of a force
x,j>,z rectangular coordinates
x,y,z coordinates of centroid or center of gravity
Symbols and Abbreviations xxl

y deflection of beam
a temperature coefficient of linear expansion
a,P,y- angles
y unit shearing strain
8 total elongation or contraction; deflection of beam; maxi-
mum deflection of column
5 st static deflection

€ unit tensile or compressive strain

€x ,ty,£z unit tensile or compressive strain in the x, y, and z direc-
tion, respectively
total angle of twist, slope angle for elastic curve
p radius of curvature, variable radius, mass density
v Poisson's ratio
to angular velocity

CG center of gravity
deg degrees
DF distribution factor
FS factor of safety
FEM fixed end moment
ID inner diameter
NA neutral axis
OD outer diameter
PL proportional limit
YP yield point
 Simpl

1-1 INTRODUCTION

Strength of Materials extends the study of forces that was begun in

Engineering Mechanics, but there is a sharp distinction between the two
subjects. Fundamentally, the field of mechanics covers the relations
between forces acting on rigid bodies; in statics, the bodies are in
equilibrium, whereas in dynamics, they are accelerated but can be put in
equilibrium by applying correctly placed inertia forces.
In contrast to mechanics, strength of materials deals with the
relations between externally applied loads and their internal effects on
bodies. Moreover, the bodies are no longer assumed to be ideally rigid;
the deformations, however small, are of major interest. The properties of
the material of which a structure or machine is made affect both its
choice and the dimensions that will satisfy the requirements of strength
and rigidity.
The difference between mechanics and strength of materials can
be further emphasized by the following example. It is a simple problem
in statics to determine the force required at the end of a crowbar to pry
up a given load (Fig. 1-1). A moment summation about the fulcrum
determines P. This statics solution assumes the crowbar to be both rigid
enough and strong enough to permit the desired action. In strength of
 a

Simple Stress

Figure 1-1. Crowbar must neither break nor bend excessively.

materials, however, the solution must extend further. We must investi-

gate the bar itself to be sure that it will neither break nor be so flexible
that bends without lifting the load.
it

Throughout this book we shall study the principles that govern

these two fundamental concepts of strength and rigidity. In this first
chapter we start with simple axial loadings; later we consider twisting
loads and bending loads; and finally we discuss simultaneous combina-
tions of these three basic types of loadings.

1-2 ANALYSIS OF INTERNAL FORCES

Consider a body of any shape acted upon by the forces shown in Fig.
1-2. In engineering mechanics, we would startby determining the
resultant of the applied forces to determine whether or not the body
remains at rest. If the resultant is zero, we have static equilibrium —
condition generally prevailing in structures. If the resultant is not zero,
we may apply inertia forces to bring about dynamic equilibrium. Such
cases are discussed later under dynamic loading. For the present we
shall consider only cases involving static equilibrium.
In strength of materials, we make an additional investigation of
the internal distribution of the forces. This is done by passing an

exploratory section a -a through the body and exposing the internal

Figure 1-2. Exploratory section a-a through loaded member.

1 -2 Analysis of Internal Forces

Figure 1-3. Components of internal effects on exploratory section a-a.

forces acting on the exploratory section that are necessary to maintain

equilibrium of a free-body diagram of either segment. In general, the
internal forces reduce to a force and a couple which, for convenience,
are resolved into components normal and tangent to the section, as
shown in Fig. 1-3.
The origin of the reference axes is always at the centroid which is
the key reference point of the section. Although we are not yet ready to
show why this is so, we shall prove it as we progress; in particular, we
shall prove it for normal forces in the next article. If the X axis is
normal to the section, the section is known as the X surface or, more
briefly, the X face. The orientation of the Y and Z axes in the plane of
the section is usually chosen to coincide with the principal axes of
inertia.

The notation used 1-3 identifies both the exploratory

in Fig.
section and the direction moment component. The first
of the force or
subscript denotes the face on which the component acts; the second
subscript indicates the direction of the particular component. Thus P
is the force on the X face acting in the Y direction.
Each component reflects a different effect of the applied loads on
the member and is given a special name, as follows:

Pxx Axial force. This component measures the pulling (or

pushing) action over the section. A pull represents a
tensile force which tends to elongate the member,
whereas a push is a compressive force which tends to
shorten it. It is often denoted by P.

Pxv>
xyi Pxz Shear force. These are components of the total resis-
tance to sliding the portion to one side of the explora-
tory section past the other. The resultant shear force is
usually designated by V, and its components by Vy and
Vz to identify their directions.
 1 Simple Stress

A/. Torque. This component measures the resistance to

twisting the member and is commonly given the
symbol T.

M^M X
Bending moments. These components measure the resis-

tance to bending the member about the Y ox Z axes

and are often denoted merely by My or M z
.

From the preceding discussion, it is evident that the internal effect

of a given loading depends upon the selection and orientation of the
exploratory section. In particular, if the loading acts in one plane, say
the XY plane as is frequently the case, the six components in Fig. 1-3
reduce to only three, viz., the axial force Pxx (or P), the shear force P
(or K), and the bending moment M
xz (or M
). Then, as shown in Fig.

l-4a, these components are equivalent to the single resultant force R. A

little reflection show that if the exploratory section had been
will
oriented differently, like b-b in Fig. l-4b where it is perpendicular to
/?, the shearing effect on the section would reduce to zero and the

tensile effect would be at a maximum.

The purpose of studying strength of materials is to ensure that the
structures used will be safe against the maximum internal effects that
may be produced by any combination of loading. We shall learn as our
study proceeds that it is not always possible or convenient to select an
exploratory section that is perpendicular to the resultant load; instead,
we may have to start by analyzing the effects acting on a section like
a-a in Figs. 1-2 and l-4a, and then learn how these effects combine to

^— Normal
L component

component

(a) Normal and shear (b) When exploratory section b-b is

components on arbitrary perpendicular to resultant R
section a-a. of applied loads, only normal
forces are produced.
Figure 1-4.
1 -3 Simple Stress 5

produce maximum internal effects like those on section b-b in Fig.

l-4b. This procedure we Chapter 9, which deals with
shall study later in
combined stresses. For the present, we restrict our study to conditions
of loading in which the section of maximum internal effect is evident by
inspection.

1-3 SIMPLE STRESS

One of the basic problems of the engineer is to select the proper

material and correctly use and proportion it so as to enable a structure
or machine to do most efficiently what it is designed to do. For this
purpose, it is essential to determine the strength, stiffness, and other
properties of materials. A tabulation of the average properties of com-
mon metals is given in Appendix B, Table B-l, on page 634.
Let us consider two bars of equal length but different materials,
suspended from a common support as shown in Fig. 1-5. If we knew
nothing about the bars except that they could support the indicated
maximum loads [500 N (Newtons) for bar 1 and 5000 N for bar 2], we
could not tell which material is stronger. Of course, bar 2 supports a
greater load, but we cannot compare strengths without having a com-
mon basis of comparison. In this instance, the cross-sectional areas are
needed. So let us further specify that bar 1 has a cross-sectional area of
10 mm and bar 2 has an area of 1000
2 2
mm
Now it is simple to .

compare their strengths by reducing the data to load capacity per unit
area. Here we note that the unit strength of bar 1 is

= SOON SOON
a, =50xl()6N/m2
10 mm2 10 X 10" 6 m 2

and bar 2 has a unit strength

= 5000 N 5000 N = 5 X 10
6
N/m 2
a,
1000 mm lOOOxKT'm2
2

Thus the material of bar 1 is ten times as strong as the material of bar 2.

Barl Bar 2

500 N 5000 N
Figure 1-5. Bars supporting maximum loads.
6 1 Simple Stress

The unit strength of a material is usually defined as the stress* in

the material. Stress is expressed symbolically as

-1 »-»
where o (Greek lowercase letter sigma) is the stress or force per unit
area, P is the applied load, and A is the cross-sectional area. Observe
that maximum stress in tension or compression occurs over a section
normal to the load, as indicated in Fig. l-4b. Shearing stress is dis-
cussed in the next section.
From can be seen that the units for stress are the units
Eq. (1-1) it

of force divided by the units of area. In SI (which is the official

abbreviation for the international system of units, Le Systeme Interna-
tional d'Unites), force is measured in newtons (N) and area is measured
in square meters (m 2 ). Thus the units for stress are newtons per square
meter (N/m 2 ). Frequently, one newton per square meter is referred to
as one pascal (Pa). Since the prefix M (read as "mega") refers to
6
multiples of 10 in SI, in the above example, the stress in bar 1 may be
expressed as 50 MN/m 2
(or 50 MPa) and that in bar 2 as 5 MN/m2 (or
5 MPa).
Even as simple an expression as Eq. (1-1) requires careful discus-
sion. Dividing load by area does not give the stress at all points in the
cross-sectional area; it merely determines the average stress. A more
exact definition of stress is obtained by dividing the differential load dP
by the differential area over which it acts:

°= § d- la>
Next let us see under what conditions a = P/A will accurately
define the stress at all points of the cross section. The condition under
which the stress is constant or uniform is known as simple stress. We
shall show now that a uniform stress distribution can exist only if the
resultant of the applied loads passes through the centroid of the cross
section/
Suppose that a cutting plane isolates the lower half of one of the
bars in Fig. 1-5. Then, as shown in Fig. 1-6, the resisting forces over
the cut section must balance the applied load P. A typical resisting force
is dP. Applying the conditions of equilibrium, we obtain

[2Z =0] P = fdP= fodA

[SA/V = 0] Pb = fx dP = fx(o dA)
•Some engineers use the terms stress or total stress as synonymous with
load or force, and unit stress or stress intensity when referring to the intensity of
load per unit area. In this book, stress will always denote force per unit area.
There are certain exceptions to this rule; they are caused by stress con-
centration (see p. 512) and by abrupt changes in cross section, and at points in
the vicinity of the applied loads (see p. 7).
1-3 Simple Stress

Figure 1-6. For uniform stress, P must pass through the centroid C.

If we specify that the stress distribution is to be constant over the

cut section, o may be written outside the integrals in the above equa-
tions to obtain

P= a f dA = oA
and therefore,

Pb = (oA)b = ofxdA
Then, canceling the common factor a, we obtain
, fxdA _
b = -A~= X
from which the coordinate b of the point C
is recognized as being the x

coordinate of the centroid of the section. By taking a moment summa-

tion about the X axis, we could similarly show that y defines the v
coordinate of C We conclude that a uniform stress distribution is

obtained only when the resultant of the applied loads passes through the
centroid of that surface.
It does not follow, however, that positioning the load through the
centroid of the section always results in a uniform stress distribution.
For example, in Fig. 1-7 is shown the profile of a flat bar of constant
thickness. The load P is applied at the center line of the bar. At sections
b-b and /-/, the stress distribution is uniform and illustrates the

principle discussed earlier; but at the other indicated sections the

stresses are not uniform.
At section e-e, the stress distribution is not uniform because the
line of action of P obviously does not pass through the centroid of the
Nor are the stresses uniformly distributed all across sections c-c
section.
and d-d because, although the action line of P does pass through the
 1 Simple Stress

Figure 1-7. Exceptions to uniform stress distribution occur at sections a-a, c-c,
d-d, and e-e.

centroids of these sections, here there are abrupt changes in section. At

such sections, the stresses are usually highly localized and can be
determined only by the mathematical theory of elasticity or some
experimental method such as photoelasticity. Also, the stress is not
uniform across section a-a because here the section is too close to the
point where the load is applied. Unless a section is located at a distance
from the end of the rod at least equal to the minimum width of the rod,
we will not obtain a uniform stress distribution.*
In order to visualize why sections c-c, d-d, and a-a do not have
uniform stress, imagine that the applied force P produces stress lines
which radiate out from the load and distribute themselves throughout
the body as indicated by the dashed lines in the figure. Although this
concept is not actually correct, it does indicate the existence of stress
concentration wherever the shape of the body interferes with the "free
flow" of the stress lines. The bunching of these lines about the hole in
section c-c, and around the sharp corner of section d-d, which indi-
cates stress concentration, contrasts with the relatively smooth flow of
stress around the radius between sections e-e and /-/.

•See S. Timoshenko and J. N. Goodier, Theory of Elasticity, 2nd ed.,

McGraw-Hill. New York, 1951, p. 33.
1 -3 Simple Stress

ILLUSTRATIVE PROBLEMS

101. An aluminum tube is rigidly fastened between a bronze rod

and a steel rod as shown in Fig. l-8a. Axial loads are applied at the
positions indicated. Determine the stress in each material.

Aluminum
Steel
Bronze = 1000
A mm2
A = 700 mm 2 A = 800 mm 2
20 kN 10 kN
15 kN 15 kN

500 mm-4" — 600 mm- 700 mm

(a)

20 kN

20 kN 15 kN ^

20 kN 15 kN 15 kN f-
(b)

Figure 1-8.

Solution: To calculate the stress in each section, we must first de-

termine the axial load in each section. The appropriate free-body
diagrams are shown in Fig. l-8b, from which we determine the axial
load in each section to be Pb — 20 kN (compression), Pa = 5 kN
(compression), and P5 — 10 kN (tension). The stresses in each section
are

a — Ol = 20 kN 20 x 10
3
N
A 700 mm 2
700 X 10" 6 m 2

= 28.6 X 10
6
N/m 2 = 28.6 MPa Ans.

o„ = 5 kN 5 X 10
3
N
1000 mm 2
1000 X 10~ 6 m 2

=5 X 10
6
N/m = 5 MPa 2
Ans.

= 10 kN 10 X 10*N
a,
800 mm 2
800 X 10~ 6 m 2

= 12.5 x 10
6
N/m = 2
12.5 MPa Ans.
10 1 Simple Stress

The stresses in the bronze and aluminum sections are compressive,

whereas the stress in the steel section is tensile.
Note that neither the lengths of the sections nor the materials from
which the sections are made affect the calculation of the stresses.
As you can see from this example, the first step in calculating the
stress in a member is to determine the internal force carried by the
member. This determination is accomplished by the analysis of correctly
drawn free-body diagrams. Note that in this example, it would have
been easier to determine the load in the steel section by taking the
section lying to the right of the exploratory section in the steel.

102. For the truss shown in Fig. l-9a, determine the stress in
members AC and BD. The cross-sectional area of each member is 900
mm 2
.

Solution: The three assumptions used in the elementary analysis of

trusses are as follows:

1. Weights of the members are neglected.

2. All connections are smooth pins.

3. All external loads are applied directly to the pin joints.

Using the above three assumptions, the members of the truss may be
analyzed as two-force members —
the internal force system carried by
any member reduces to simply a single force (tension or compression)
acting along the line of the member.
The free-body diagram of the entire truss is shown in Fig. l-9a.
An equilibrium analysis of this free-body diagram results in the follow-
ing values for the external reactions: Ay = 40 kN, Hy = 60 kN, and
Hx =0.
To determine the force in member AC, we pass an imaginary
cutting plane which isolates joint A (section ©, Fig. l-9a). The free-
body diagram of joint A is shown in Fig. l-9b. Here, AB and AC
represent the forces in members AB and AC, respectively. Note that
both members have been assumed to be in tension. Analyzing the
free-body diagram in Fig. l-9b,

[2 Y - 0] QA y
+\AB =
AB= -\Ay = -
-f(40) = -66.7 kN
[2X - 0] © AC +\AB =
AC = -\AB = -|(-66.7) = 53.4 kN

The minus sign indicates that the 66.7 kN force in member AB is

1-3 Simple Stress 11

® D

4 panels at 4 m = 16 m

(a)

£D

±E
*~AC

(b)

Figure 1-9.

compressive. The member AC is 53.4 kN, tension.

force in
To determine the force in member BD, we pass an imaginary
cutting plane which exposes the force in member BD (section ©, Fig.
l-9a). The free-body diagram of the portion of the truss to the left of
section © is shown in Fig. 1 -9c. (The portion of the truss to the right of

section © could also have been used.) The forces in members BD, BE,
and CE are assumed to be tensile. To calculate the force BD, we
eliminate the forces BE and CE by taking a moment summation about
their point of intersection, E, and write

[2ME = 0] -4,(8) + 30(4) -*Z)(4)-0

4BD m -SAy + 120 8(40) + 120
= -200
=
BD -50 kN

Therefore, the force in member BD is 50 kN, compression.

12 1 Simple Stress

The stresses in members AC and BD, are

— ~ 53.4 kN 53.4 X 10
3
N
a °ac
900 mm 2
900 X 10" 6 m2
- 59.3 x 10
6
N/m = 2
59.3 MPa Arts.
(tension)

" 50 kN 50 x 10
3
N
°bd
900 mm 2
900 X 10" 6
m2
= 55.6 X 10
6
N/m 2 = 55.6 MPa Ans.
(compression)

In truss analysis, the method of analyzing a single joint, as shown

in Fig. 1 —9b, is method of joints. The analysis of a
referred to as the
section of the truss composed of two or more joints, as shown in Fig.
l-9c, is called the method of sections. It must be reemphasized that the
force internal to a member of a truss lies along the line of the member
only because sufficient assumptions are made which reduce all members
to two-force members. As discussed in Art. 1-2, the internal forces for
an arbitrarily loaded member are considerably more complicated than
simply an axial force.

PROBLEMS
103. Determine the largest weight which can be supported by W
the two wires shown in Fig. P-103. The stresses in wires AB and AC are
not to exceed 100 MPa and 150 MPa, respectively. The cross-sectional
areas of the two wires are 400 mm 2
for wire AB and 200 mm2 for wire
AC.
Ans. W = 33.5 kN

Figure P-103.

104. For the truss shown in Fig. P-104, calculate the stresses in
members DF % C£, and BD. The cross-sectional area of each member is
1200 mm 2
. Indicate tension (T) or compression (C).
Ans. DF = 188 MPa (C); CE = 1 13 MPa (T);
BD = 80.1 MPa(C)
1-3 Simple Stress 13

/7V7

100 kN 200 kN
Figure P-104.

105. For the truss shown in Fig. P-105, determine the cross-sec-
tional areas of bars BE, BF, and CF so that the stresses will not exceed
100 MN/m 2
in tension or 80 MN/m 2
in compression. A reduced stress
in compression is specified to avoid the danger of buckling.
Arts. A BE = 625 mm2 ; A BF = 427 mm 2
; A CF = 656 mm 2
14 1 Simple Stress

106. The bars of the pin-connected frame in Fig. P-106 are each
30 mm by 60 mm in section. Determine the maximum load P that can
be applied without exceeding the allowable stresses specified in Problem
105.

Figure P-106.

107. A cast-iron column supports an axial compressive load of

250 kN. Determine the inside diameter of the column if its outside
diameter is 200 mm
and the limiting compressive stress is 50 MPa.
108. Determine the outside diameter of a hollow steel tube that
will carry a tensile load of 500 kN at a stress of 140 MN/m 2
. Assume
the wall thickness to be one-tenth of the outside diameter.
Arts. D — 107 mm
109. Part of the landing gear for a light plane is shown in Fig.
P-109. Determine the compressive stress in the strut AB caused by a
landing reaction R = 20 kN. Strut AB is inclined at 53.1° with BC.
Neglect weights of the members. Ans. a = 65.7 MN/m 2

Hollow strut
OD - 40 mm
ID 30 mm

1 200 mm-*
Flgure P-109.
450 mm-

110. A steel tube is rigidly attached between an aluminum rod

and a bronze rod as shown in Fig. P-l 10. Axial loads are applied at the
positions indicated. Find the maximum value of P that will not exceed a
1 -3 Simple Stress 15

stress in aluminum of 80 MPa, in steel of 150 MPa, or in bronze of 100

MPa.

Steel
Bronze
Aluminum A = 500 mm 2
A = 400 mm 2
A = 200 mm 2

k lm 2m
Figure P-110.
2.5 m

A
homogeneous 150-kgj bar AB carries a 2-kN force as
111.
shown P— 111. The bar is supported by a pin at B and a
in Fig.
10-mm-diameter cable CD. Determine the stress in the cable.

4 m

2kN
Figure P-111.

112. Determine the weight of the heaviest cylinder which can be

placed in the position shown in Fig. P— 112 without exceeding a stress of
50 MN/m2 in the cable BC. Neglect the weight of bar AB. The
cross-sectional area of cable BC is 100 mm2 .

6 m

Figure P-112.
16 1 Simple Stress

113. A 1000-kg homogeneous bar AB is suspended from two

cables AC and BD y each with cross-sectional area 400 mm 2
, as shown in
Fig. P— 1 Determine the magnitude P and location x of the largest
13.

additional force which can be applied to the bar. The stresses in the
cables AC and BD are limited to 100 MPa and 50 MPa, respectively.
Arts. P - 50.2 kN; x = 0.602 m

1.8 m 1.8 m

Figure P-113.

1-4 SHEARING STRESS

Shearing stress differs from both tensile and compressive stress in that it

is caused by forces acting along or parallel to the area resisting the

forces, whereas tensile and compressive stresses are caused by forces
perpendicular to the areas on which they act. For this reason, tensile
and compressive stresses are frequently called normal stresses, whereas a
shearing stress may be called a tangential stress.
A shearing stress is produced whenever the applied loads cause
one section of a body to tend to slide past its adjacent section. Several
examples are shown in Fig. 1-10. In (a) the rivet resists shear across its
cross-sectional area, whereas in the clevis at (b) the bolt resists shear
across two cross-sectional areas; case (a) may be called single shear and
case (b) double shear. In (c) a circular slug about to be punched out of
is

a plate; the resisting area is similar to the milled edge of a coin. In each
case, the shear occurs over an area parallel to the applied load. This
may be called direct shear in contrast to the induced shear that may
occur over sections inclined with the resultant load, as was illustrated in
Fig. l-4a.
1 -4 Shearing Stress 17

-f>-

^-
T7
(a) (b) h (c)

Figure 1-10. Examples of shear.

The discussion concerning uniform normal stresses in the preced-

ing article might lead us to conclude also that a uniform shearing stress
will exist when the resultant shearing force V passes through the
centroid of the cross section being sheared. If this were true, the
shearing stress t (Greek lowercase letter tau) could be found from

t = (1-2)

Actually, the shearing stress across a section is practically never uni-

formly distributed (e.g., see Art. 5-7), so Eq. (1-2) must be interpreted
as giving merely the average shearing stress. This does not limit the
usefulness of Eq. (1-2) provided we use an average shearing stress that
takes into account the actual nonuniform distribution. Moreover, the
shearing stress distribution does approach uniformity when both
the distance between the applied shearing loads and the depth of the
shearing area are small. These are the conditions that prevail in Fig.
1-10 and in the following problems.

PROBLEMS

114. As to be punched out of a plate

in Fig. l-10c, a hole is

having an ultimate shearing 300 MPa. (a) If the compressive

stress of
stress in the punch is limited to 400 MPa, determine the maximum
thickness of plate from which a hole 100 mm
in diameter can be
punched, (b) If the plate is 10 mm
thick, compute the smallest diameter
hole which can be punched. Arts, (a) / = 33.3 mm; (b) d — 30.0 mm
18 1 Simple Stress

115. The end chord of a timber truss is framed into the bottom
chord as shown in Fig. P— 1 15. Neglecting friction, (a) compute dimen-
sion b if the allowable shearing stress is 900 kPa; and (b) determine
dimension c so that the bearing stress does not exceed 7 MPa.
Ans. (a) b — 321 mm; (b) c = 41.2 mm

P=50kN

y^
4fe
i\ r°\
i.f-
1 ^ i

F gure P- 115.

116. In the landing gear described in Problem 109, the bolts at A

and B are in single shear and the one at C double shear. Compute
is in
the required diameter of these bolts if the allowable shearing stress is 50
MN/m 2
.

117. A 750-mm pulley, loaded as shown in Fig. P-117, is keyed

to a shaft of 50-mm diameter. Determine the width b of the 75-mm-long
key if the allowable shearing stress is 70 MPa. Ans. b = 1 1.4 mm

10 kN

\
10 mm ,y^Jb mm
r b -<-

6kN
Figure P-117.
1-4 Shearing Stress 19

118. The bell crank shown in Fig. P— 118 is in equilibrium, (a)

Determine the required diameter of the connecting rod A B if its axial
stress is limited to 100 MN/m
2
(b) Determine the shearing stress in the
.

pin at D if its diameter is 20 mm.

P-*

,j*.
/////////?/.

30 kN
Figure P-118.

119. The mass of the homogeneous bar AB shown in Fig. P-119

is2000 kg. The bar is supported by a pin at B and a smooth vertical
surface at A Determine the diameter of the smallest pin which can be
.

used at B if its shear stress is limited to 60 MPa. The detail of the pin
support at B is identical to that of the pin support at D shown in Fig.
P-118. Ans. d= 14.9 mm

S777

Figure P-119.

120. Two blocks of wood, 50 mm

wide and 20 mm
thick, are
glued together as shown in Fig. P-120. (a) Using the free-body diagram
concept illustrated in Fig. l-4a, determine the shear load and from it
20 1 Simple Stress

the shearing stress on the glued joint if P = 6000 N. (b) Generalize the
procedure of part (a) to show that the shearing stress on a plane inclined
at an angle to a transverse section of area A is r = P sin 20/2/1.

Figure P-120.

121. A rectangular piece of wood, 50 mm by 100 mm in cross

section, is used as a compression block as shown in Fig. P— 121.
Determine the maximum axial force Pwhich can be safely applied to
the block if the compressive stress in the wood is limited to 20 MN/m 2

and the shearing stress parallel to the grain is limited to 5 2

MN/m The .

grain makes an angle of 20° with the horizontal, as shown. (Hint: Use
the results of Problem 120.) Arts. P - 77.8 kN

100 mm
Figure P-121.

1-5 BEARING STRESS

Bearing stress differs from compressive stress in that the latter is the
internal stress caused by a compressive force whereas the former is a
contact pressure between separate bodies. Some examples of bearing
stress are the soil pressure beneath piers and the forces on bearing
plates. We now consider the contact pressures between an axle and its
bearing, or between a rivet or bolt and the contact surface of the plate
against which it pushes.
1 -5 Bearing Stress 21

pb
c©

Projected area of
rivet hole

Figure 1-11. Exaggerated bearing deformation of upper plate.

Pb'A b a b =(td)ab

In Fig. 1-11, the result of an excessive bearing stress is to cause

yielding of the plate or of the rivet, or both. The intensity with which
the rivet bears against the rivet hole is not constant, but actually varies
from zero at the edges of the hole to a maximum
back of the
directly in
rivet. The is avoided
difficulty inherent in a variable stress distribution
by the common practice of assuming the bearing stress ob to be
uniformly distributed over a reduced area which is the projected area of
the rivet hole. Then the bearing load is expressed by
Pb = * b °b = {«*)<** (1-3)

This result is analogous to that for a cylinder subjected to a

uniform internal pressure (see the next article, especially Fig. 1-14).
There, as we shall see, the net force is equal to the uniform pressure
multiplied by the projected area.

ILLUSTRATIVE PROBLEM

122. Figure 1-12 shows a W460 X 97 beam riveted to a W610 X

125 girder by two 100 X 90 X 10 -mm angles with 19-mm-diameter
Z3J

*££
^ , &***
li
V7777
2222-W610X 125 girder
web = 11.9 mm
W460 X 97 beam, web =11.4 mm
Figure 1-12. Strength of beam and girder connection.
22 1 Simple Stress

rivets. (Refer to Appendix B for the properties of structural sections.)

For the shop-driven rivets that attach the angles to the beam, assume
t = 80 MPa and ob = 170 MPa. For the field-driven rivets (riveted on
the job), assume r = 70 MPa and ob = 140 MPa. The web of the girder
is 11.9 mm thick, and the web of the beam is 1 1 .4 thick. Determine mm
the allowable end reaction.

Solution: At the girder, the shearing resistance is that of 8 field-driven

rivets in single shear; hence we have
2
[P -Ar] P- 8(-^)(19 x 10" 3
) (70 x 10
6
) = 159 kN
The bearing resistance at the girder depends on the minimum
thickness of the connection, which in this case is the 10-mm thickness of
the clip angle. We obtain for 8 field-driven rivets in bearing:

[P = Aob ] P= 8(19 X 10" 3 )(10 X 10" 3 )(140 X 10

6
)

= 213 kN
At the beam, there are 4 shop-driven rivets in double shear, giving
a total of 8 single-shear areas. With an allowable shearing stress of 80
MPa, this makes the shear resistance greater here than at the girder.
The bearing resistance at the beam depends on the web thickness
of the beam. Since this is smaller than the combined thickness of the
two clip angles, for the 4 rivets in bearing, we obtain
[P = Aob ] P- 4(19 X 10- 3 )(11.4 X 10" 3 )(170 X 10
6
)

« 147 kN
The safe beam reaction is the smallest of the above values, that is,
147 kN; it is limited by the bearing of the shop-driven rivets against the
W460 x 97 beam.

PROBLEMS

123. In Fig. 1-11, assume that a 20-mm-diameter rivet joins the

plates which are each 100 mm wide, (a) If the allowable stresses are 140
MN/m 2
for bearing in the plate material and 80 MN/m 2
for shearing
of the rivet, determine the minimum thickness of each plate, (b) Under
the conditions specified in part (a), what is the largest average tensile
stress in the plates. Ans. (a) 8.98 mm; (b) 35.0 MN/m 2

124. The shown in Fig. P-124 is fastened by three

lap joint
20-mm-diameter Assuming that P = 50 kN, determine (a) the
rivets.
shearing stress in each rivet, (b) the bearing stress in each plate, and (c)
the maximum average tensile stress in each plate. Assume that the
applied load P is distributed equally among the three rivets.
 —
1 -5 Bearing Stress 23

130 mm
o- <> -e

25 mm
_L_ 25 mm
C^ C^ <cx

—^y—o—C7 L^
Figures P-124 and P-125.

125.' For the lap joint in Problem 124, determine the maximum
safe load P which may be applied if the shearing stress in the rivets is
limited to 60 MPa, the bearing stress in the plates to 110 MPa, and the
average tensile stress in the plate to 140 MPa. Arts. 56.5 kN
126. In the clevis shown in Fig. 1 — 10b, on page 17, determine
the minimum bolt diameter and the minimum thickness of each yoke
that will support a load P= 55 kN without exceeding a shearing stress
of 70 MPa and a bearing stress of 140 MPa.
127. A
22.2-mm-diameter bolt having a diameter at the root of
the threads of 18.6 mm
is used to fasten two timbers as shown in Fig.

P-127. The nut is tightened to cause a tensile load in the bolt of 34 kN.
Determine (a) the shearing stress in the head of the bolt, (b) the shearing
stress in the threads, and (c) the outside diameter of the washers if their
inside diameter is 28 mm
and the bearing stress is limited to 6 MPa.

Figure P-127.
24 1 Simple Stress

128. Figure P-128 shows a roof truss and the detail of the
riveted connection at joint B. Using allowable stresses of t = 70 MPa
and ob = 140 MPa, how many 19-mm-diameter rivets are required to
fasten member BC to the gusset plate? Member BE1 What is the largest
average tensile or compressive stress in BC and BE1
Ans. For BC, 1 rivets; for BE, 5 rivets

129. Repeat Problem 128 if the rivet diameter is 22 mm and all

other data remain unchanged.

D
14 mm
gusset
B ^f F
6 m ^ plate

a/m
/\ 4
c
4 m
r
4 m
G
4
\ m /\
H

96 kN 200 kN 96 kN
75 X 75 X6 mm
(a)

Figures P-128 and P-129.

1-6 THIN-WALLED CYLINDERS

A under a pressure of p N/m 2 is

cylindrical tank carrying a gas or fluid
subjected to tensile forces which resist the bursting forces developed
across longitudinal and transverse sections. Consider first a typical
longitudinal section A- A through the pressure-loaded cylinder in Fig.
I — 13a. A free-body diagram of the half-cylinder isolated by the cutting

plane A -A is shown in Fig. 1 — 13b.

dF^pdA=pL(D2)dO
(a) (b)

Figure 1-13. Analytical determination of bursting force F.

1-6 Thin-Walled Cylinders 25

The elementary force acting normal to an element of the cylinder

located at an angle 9 from the horizontal diameter is

dF = pdA = pL^-dB

A similar force (not shown) acts on the symmetrically placed element on

the other side of the vertical center line. Since the horizontal compo-
nents of such pairs of forces cancel out, the bursting force F is the
summation of the vertical components of these elementary forces:

V
F = f{pL^ <#) sin = />Ly [ -cos B]
Q

which reduces to
F = pDL
It is apparent that the total bursting force F, acting normal to the
cutting plane A -A, is resisted by the equal forces P acting on each cut
surface of the cylinder wall. Applying a vertical summation of forces,
we obtain

[2 V= 0] F = pDL = IP — F^i^e- (1-4)

A simpler method of determining the bursting force F is indicated

in Fig. 1-14. Here the lower half of the cylinder is occupied by a fluid.
Since a fluid transmits pressure equally in all directions, the pressure
on the cylinder is the same as that in Fig. 1-13. From the
distribution
accompanying free-body diagram, it is apparent that the bursting force
Fy acting over the flat surface of the fluid, equals the pressure intensity/?
multiplied by the area DL over which it acts, or F = pDL as before.
The stress in the longitudinal section that resists the bursting force
F is obtained by dividing it by the area of the two cut surfaces. This
gives

°=A
F _pDL = pD
2tL It
_12^T^ „ ,.

(a) (b)

Figure 1-14. Direct evaluation of bursting force F.

26 1 Simple Stress

This stress is usually called the tangential stress because it acts tangent
to the surface of the cylinder; other common names are circumferential
stress, hoop stress, and girth stress. The stress computed by Eq. (1-5) is

the average stress; for cylinders having a wall thickness equal to 1/10 or
less of the inner radius, it is practically equal to the maximum stress at
the inside surface. (See Art. 13-11 for the stress distribution in thick-
walled cylinders.)
If we consider next a free-body diagram of a transverse section
(Fig. 1-15), we see that the bursting force acting over the end of the
cylinder is resisted by the resultant P of the tearing forces acting over
the transverse section. The area of a transverse section is the wall
thickness multiplied by the mean circumference, or tt(D + t)t; if / is

small compared to D, it is closely approximated by irDt. Thus we obtain

2
7TD
[r-r\ irDtOj = p

or

a, - S±L _ 3T*^> £^ fc^^ (1_6)

where a, denotes what is called the longitudinal stress because it acts

parallel to the longitudinal axis of the cylinder.

F - nD 2
P

"P^(nDt)(7 l

Figure 1-15. Bursting force on a transverse section.

Comparing Eqs. (1-5) and (1-6) shows that the longitudinal stress
is one-half the value of the tangential stress. In effect, this is equivalent
to stating that, if the pressure in a cylinder is raised to the bursting
point, failure will occur along a longitudinal section or longitudinal
seam of the cylinder. When a cylindrical tank is composed of two sheets
riveted together, as in Fig. 1-16, the strength of the longitudinal joint
should be twice the strength of the girth joint. In other words if, as is

often the case, the longitudinal joint is not twice as strong as the girth
1-6 Thin-Walled Cylinders 27

Longitudinal joint

Girth joint

Figure 1-16.

joint, the permissible internal pressure will depend on the strength of the
longitudinal joint.
Equations (1-5) and (1-6) have been developed primarily to
determine the relation stated in the above paragraph, not as equations to
be memorized. It is generally best to compute the stresses by determining
the resisting load P from a free-body diagram and then computing the
stress by using a — P/A.
For this purpose, Fig. 1-14 is replaced by the
equivalent skeleton diagram in Fig. 1-17, which also establishes the
relation IP = pDL.
F= P DL

Figure 1-17.

When the ends of the cylinder are not squared off as in Fig. 1-15,
but are rounded or dished as in Fig. 1-18; the bursting force on a
transverse section may still be computed as the product of the internal
pressure multiplied by the projected area of the transverse section. Thus,
using the concept discussed in connection with Fig. 1-14, we may
imagine the volume between the transverse section A -A and the

Dished or
concave end

Rounded or
convex end
Figure 1-18.
28 1 Simple Stress

rounded end in Fig. 1-18 to be full of a fluid. The resultant longitudinal

force will equal the product of the pressure intensity multiplied by the
shaded area of the transverse section.
As another application of the concept of a fluid to transmit
pressure, consider a pump chamber cast in several parts, with projecting
flanges that are bolted together as shown in Fig. 1-19. The bursting
force to be resisted by the bolts in section A -A is proportional to the
cross-sectional area at A -A and is expressed by F, = p(ttD ]
2
/4); simi-
larly, the bursting force resisted by the bolts in section B-B is F2 =
2
P (ttD 2 /4).

Figure 1-19.

The principles just discussed for determining the tangential stress

in thin-walled cylinders may also be applied to computing the contact
pressure exerted by hoops shrunk upon cylinders or the tensile stress
developed in a thin rotating ring. In the latter case, for example, the
bursting force is generated by the centrifugal force developed in one-
may be obtained (Fig. 1-20) by assuming the
half of the ring. Its value
mass of the half-ring concentrated at its center of gravity, whence we
have

F— mroi
1
(a)

in which <o is the angular velocity in radians per second and m is the

» F=m\

Figure 1-20. Free-body diagram of one-half of rotating ring.

1-6 Thin-Walled Cylinders 29

mass of one-half of the ring. For a thin ring, m is given by

m = pV = pnrArc

where p (Greek lowercase letter rho) is the mass per unit volume of the
ring, A is the cross-sectional area of the ring, and rc is the radius of the
mean circumference. From mechanics, the value of f for a semicircular
ring is r = 2rc / m. Substituting these values reduces Eq. (a) to
'

F= (pAwre )l ^L 2
= 2pAv 2 (b)

where v = rc co is the peripheral velocity of the ring.

From equilibrium of the free-body diagram in Fig. 1-20 we have

IP = F
Hence the stress is

P pAv 2 2 , v

Thus mass density and the square of

the stress varies directly with the
must be taken to use
the peripheral velocity. In applying Eq. (c), care
consistent units. With p in kilograms per cubic meter and v in meters
per second, o will be determined in newtons per square meter.

ILLUSTRATIVE PROBLEM

130. A large pipe, called a penstock in hydraulic work, is 1.5 m

in diameter. Here it is composed of wooden staves bound together by
steel hoops, each 300 in cross-sectional mm 2
is used to area, and
conduct water from a reservoir to a powerhouse. If the maximum tensile
stress permitted in the hoops is 130 MPa, what is the maximum spacing
between hoops under a head of water of 30 m? (The mass density of
water is 1000 kg/m 3 .)

Solution: The pressure corresponding to a head of water of 30 m is

given by

[p = pgh] p = (1000 kg/m )(9.81 m/s )(30 m)

3 2

= 294 x 103 N/m2 - 294 kPa

If the maximum spacing between hoops is denoted by L, then, as
shown in Fig. 1-21, each hoop must resist the bursting force on the
length L. Since the tensile force in a hoop is given by P= Ao, we
obtain from the free-body diagram

[pDL = IP]
(294 X 3
10 )(1.5)L = 2(300 X 10" 6 )(130 X 10
6
)

whence
L 0.177 m= 177 mm Arts.
30 1 Simple Stress

r P = A(T

P = A<T
<

k*,-+ +-L
Figure 1-21. Spacing of hoops in a penstock.

PROBLEMS
131. Show that the stress in a thin-walled spherical shell of
diameter D
and wall thickness / subjected to internal pressure p is given
by a « pD/At.
132. A cylindrical pressure vessel is fabricated from steel plates
which have a thickness of 20 mm. The diameter of the pressure vessel is
500 mm and its length is 3 m. Determine the maximum internal pressure
which can be applied if the stress in the steel is limited to 140 MPa. If
the internal pressure were increased until the vessel burst, sketch the
type of fracture which would occur. Arts. 1 1.2 MPA

133. Find the limiting peripheral velocity of a rotating steel ring

if the allowable stress is 140 MN/m 2
and the mass density of steel is
will the stress reach 200 MN/m
2
7850 kg/m 3 At what . angular velocity
if the mean radius is 250 mm? Ans. 134 m/s; 640 rad/s

134. A water tank is 8 m in diameter and 12 m high. If the tank

is to be completely filled, determine the minimum thickness of the tank
plating if the stress is limited to 40 MPa. Ans. 1 1.8 mm
135. The strength per meter of the longitudinal joint in Fig. 1-16
is 480 kN, whereas for the girth joint it is 200 kN. Determine the
maximum diameter of the cylindrical tank if the internal pressure is 1.5

MN/m 2
.

136. A pipe carrying steam at 3.5 MPa has an outside diameter

of 450 mm
and a wall thickness of 10 mm. A gasket is inserted between
the flange at one end of the pipe and a flat plate used to cap the end.
How many 40-mm diameter bolts must be used to hold the cap on if the
allowable stress in the bolts is 80 MPa, of which 55 MPa is the initial
stress? What circumferential stress is developed in the pipe? Why is it
necessary to tighten the bolts initially, and what will happen if the steam
1-6 Thin-Walled Cylinders 31

pressure should cause the stress in the bolts to be twice the value of the
initial stress? Ans. 17 bolts; 75.3 MPa
137. A spiral-riveted penstock 1.5 m in diameter is made of steel
plate 10 mm thick. The pitch of the spiral or helix m. The
is 3 spiral
seam is a single-riveted lap joint consisting of 20-mm-diameter rivets.

Using r = 70 MPa and ob = 140 MPa, determine the spacing of the

rivets along the seam for a water pressure of 1.25 MPa. Neglect end
thrust. What is the circumferential stress?
Ans. 43.7 mm; 93.8 MPa
Repeat Problem 137, using a 2-m-diameter penstock
138.
fastened with 30-mm-diameter rivets, with all other data remaining
unchanged.
139. The tank shown in Fig. P-139 is fabricated from 10-mm
steel plate. Determine the maximum longitudinal and circumferential
stresses caused by an internal pressure of 1.2 MPa.
Ans. 17.9 MPa; 60 MPa

140. The tank shown in Fig. P-140 is fabricated from steel plate.
Determine the minimum thickness of plate which may be used if the
stress is limited to 40
2
MN/m
and the internal pressure is 1.5 MN/m2 .

400 mm

Figures P-139 and P-140.

SUMMARY
Axial loads result in a uniform stress distribution that may be
determined from

a= AP (1-D

Shearing stresses and bearing stresses are also computed by divid-

ing the load by the resisting area, but the results represent average
values. In particular, the bearing area of a rivet against a plate is given
by the projected area of the rivet hole.
32 1 Simple Stress

The stresses in thin-walled cylinders subjected to internal pressure

are most readily obtained by applying the conditions of equilibrium to a

free-body diagram of either a longitudinal or a transverse section,
depending on whether circumferential or longitudinal stress is involved.
Ihe resisting forces thereby exposed are assumed to be uniformly
distributed over the corresponding cut surfaces.
 Simple Strain

2-1 INTRODUCTION

Our main concern in the preceding chapter was the strength of a

material, i.e., the relations between load, area, and stress. We now
consider the other major field of strength of materials — the changes in
shape, i.e., deformations, that accompany a loading. Although we limit
ourselves here to axially loaded bars, the principles and methods devel-
oped apply equally well to the more complex cases of twisting or
bending. In particular, we shall learn how to apply the geometric
relations between elastic deformations which, in combination with the
conditions of equilibrium and the relations between loads and deforma-
tions, will enable us to solve statically indeterminate problems.

2-2 STRESS-STRAIN DIAGRAM

The strength of a material is not the only criterion that must be con-
sidered in designing structures. The stiffness of a material is frequently
of equal importance. To a lesser degree, such properties as hardness,
toughness, and ductility determine the selection of a material. These
properties are determined by making tests on the materials and compar-
ing the results with established standards. Although a complete descrip-
tion of these tests is the province of materials testing and hence will not
be given here, one of the tests (the tension test of steel) and its results
33
34 2 Simple Strain

will be considered because it helps to develop several important basic

concepts
Ifa specimen of structural steel is gripped between the jaws of a
testing machine and the load and the extension in a specified length are
observed simultaneously, we can plot these observations on a graph on
which the ordinates represent load and the abscissae represent extension.
Figure 2- represents such a graph. Notice that we did not plot
1

load against extension; rather, unit load or stress was plotted against
unit elongation, technically known as strain. Only by reducing observed
values to a unit basis can the properties of one specimen be compared
with those of other specimens. The diagram in Fig. 2-1 is called a
stress-strain diagram, the name being taken from the coordinates.

* Actual rupture
,' strength
Stress
Ultimate strength
•x-
"i
Rupture
Yield point strength

\
*- Elastic limit
\ Proportional limit

Strain 6 =£

Figure 2-1. Stress-strain diagram.

Strain

To obtain the unit deformation or strain, would seem obvious

c, it

to divide the elongation 8 by the length L in which it was measured,

thereby obtaining

< = (2-1)
r
The strain so computed, however, measures only the average value of

M
strain. The any position is
correct expression for strain at

.
£
where d8 is the differential elongation of the differential length dL.
Thus, Eq. (2- la) determines the average strain in a length so small that
2-2 Stress-Strain Diagram 35

the strainmust be constant over that length. However, under certain

may be assumed constant and its value computed
conditions the strain
from Eq. (2-1). These conditions are as follows:

1. The specimen must be of constant cross section.

2. The material must be homogeneous.

3. The load must be axial, that is, produce uniform stress.

Finally, note that since strain represents a change in length divided by

the original length, strain is a dimensionless quantity. However, it is

common to use units of meters per meter (m/m) when referring to

~3
strain. In engineering work, strains of the order of 1.0 X 10 m/m are
frequently encountered.

Proportional limit

From the origin O to a point called the proportional limit, Fig. 2-1
shows the stress-strain diagram to be a straight line. From this we
deduce the well-known relation, first postulated by Robert Hooke* in
1678, that stress is proportional to strain. Notice carefully that this
proportionality does not extend throughout the diagram; it ends at the
proportional limit. Beyond this point, the stress is no longer propor-
tional to the strain. The proportional important because all
limit is
subsequent theory involving the behavior of elastic bodies is based upon
a stress-strain proportionality.* This assumption places an upper limit
on the usable stress a material may carry. This is also our first indica-
tion that the proportional limit, and not the ultimate strength, is the
maximum stress towhich a material may be subjected. We shall return
to this observation later when we discuss working stress and the factor
of safety.
Other concepts developed from the stress-strain curve are the
following: (1) The elastic limit, that is, the stress beyond which the
material will not return to its original shape when unloaded but will
retain a permanent deformation called permanent set. (2) Yield point, at
which there is an appreciable elongation or yielding of the material
without any corresponding increase of load; indeed, the load may
actually decrease while the yielding occurs. However, the phenomenon
of yielding is peculiar to structural steel; other grades of steels and steel

Robert Hooke's famous law Ut tensio sic vis, i.e., "As strain, so force,"
related total strain to total force and did not recognize a limit to this proportionality.
+The stress-strain diagram of many materials is actually a curve on which
there is no definite proportional limit. In such cases, the stress-strain proportionality
is assumed to exist up to a stress at which the strain increases at a rate 50% greater
than shown by the initial tangent to the stress-strain diagram.
 2 Simple Strain

High carbon steel

Cast iron

Strain

Figure 2-2. Comparative stress-strain diagrams for different materials.

alloys or other materials do not possess it, as is indicated by the typical

stress-strain curves of these materials shown in Fig. 2-2. These curves,
incidentally, are typical for a first loading of materials that contain
appreciable residual stressesproduced by manufacturing or aging
processes. After repeated loading, these residual stresses are removed
and the stress-strain curves become practically straight, as can be
demonstrated in the testing laboratory. (3) Yield strength, closely
associated with yield point. For materials which do not have a well-
defined yield point, yield strength is determined by the offset method.
This consists of drawing a line parallel to the initial tangent of the
stress-strain curve, this line being started at an arbitrary offset strain,
usually of 0.2% or 0.002 m/m. As shown in Fig. 2-3, the intersection of
this line with the stress-strain curve is called the yield strength. (4)
Ultimate stress, or ultimate strength as it is more commonly called,
which is the highest ordinate on the stress-strain curve. (5) Rupture
For structural steel it is somewhat lower
strength, or the stress at failure.
than ultimate strength because the rupture strength is computed by
dividing the rupture load by the original cross-sectional area which,
although convenient, is incorrect. The error is caused by a phenomenon
known as necking. As failure occurs, the material stretches very rapidly
and simultaneously narrows down, as shown in Fig. 2-4, so that the
rupture load is actually distributed over a smaller area.* If the rupture
area is measured after failure occurs, and divided into the rupture load,
the result is a truer value of the actual failure stress. Although this is
considerably higher than the ultimate strength, the ultimate strength is

commonly taken as the maximum stress of the material.

•For reasons that are explained in Art. 13-4, the actual failure is caused by
shear, resulting in the cuplike rupture shown.
2-2 Stress-Strain Diagram 37

^ Yield strength

J k-o. 2<fo offset

Strain

Figure 2-3. Yield strength determined by offset method.

rsr
Figure 2-4. Necking, or narrowing, of steel specimen at failure.

Working stress and factor of safety

The working stress is defined as the actual stress the material has
when under load. It is almost synonymous with allowable stress, which is
the maximum safe stress a material may carry. In actual design, the
allowable stress aw should be limited to values not exceeding the
proportional limit so as not to invalidate the stress-strain relation of
Hooke's law on which all subsequent theory is based. However, since
the proportional limit is difficult to determine accurately, it is customary
to base the allowable stress on either the yield point or the ultimate
strength, divided by a suitable number N, called the factor of safety:

o\.. = yp
or aM = , (2-2)
N,yp

The yield point is selected as the basis for determining ow in structural

steel because it is the stress at which a prohibitively large permanent set
may occur. For other materials, the allowable stress is usually based on
the ultimate strength.
Many must be considered in selecting the allowable stress.
factors
This selection should not be made by the novice, as will became
apparent in a moment; usually the allowable stress is set by a group of
experienced engineers and embodied in various building codes and
specifications. A short discussion of the factors governing the selection
of an allowable with the observation that in many materials
stress starts
the proportional limit about one-half the ultimate strength. To avoid
is

accidental overloading, an allowable stress of one-half the proportional

limit is usually specified for dead loads gradually applied. The term
38 2 Simple Strain

dead loads generally refers to the weight of the structure or to loads

which, once applied, are not removed. An allowable stress set in this
way corresponds to a factor of safety of 4 and is recommended for
materials that are known to be quite uniform and homogeneous. For
other materials, like wood, in which unpredictable nonuniformities
(such as knotholes) may occur, larger factors of safety are desirable. The
dynamic effect of suddenly applied loads also requires higher factors of
safety. As a rule, factors of safety are not directly specified; rather,
allowable stresses are set for different materials under different condi-
tions of use, and these stresses are used by the designer.

2-3 HOOKE'S LAW: AXIAL DEFORMATION

Let us return now to a consideration of the straight-line portion of the

stress-strain diagram in Fig. 2-1. The slope of that line is the ratio of
stress to strain. It is called the modulus of elasticity and is denoted by E:
o
slope of stress-strain curve = E =—

which is usually written in the form

a - Ee (2-3)

In this form it is known as Hooke's law. Originally Hooke's law

specified merely that stress was proportional to strain, but Thomas
Young in 1807 introduced a constant of proportionality that came to be
known as Young's modulus. Eventually this name was superseded by
the phrase modulus of elasticity.
From Hooke's law, Eq. (2-3), it can be seen that the units for
modulus of elasticity E are identical to the units for stress o — recall that
strain c is a dimensionless quantity. As an illustration, the modulus of
elasticity for steel is approximately 200 X 10
9
N/m2 (200 X 109 Pa).
Using the SI prefix G (read as "giga") to denote multiples of 109 this ,

can be expressed as 200 GN/m 2 (200 GPa).

A convenient variation of Hooke's law is obtained by replacing a
by its equivalent P/A and replacing c by 5/L, so that Eq. (2-3)
becomes

^=£-*
A L
or

Equation (2-4) expresses the relation among the total deformation

6, the applied load P, the length L, the cross-sectional area A, and the
2-3 Hooke's Law: Axial Deformation 39

modulus of elasticity E. The unit of deformation 8 has the same unit as

length L, since the units of o and E, being equivalent, cancel out of the
equation. Note that Eq. (2-4) is subject to all the restrictions previously
discussed in connection with the equations it combines. For con-
venience, let us restate these restrictions:

1. The load must be axial.

2. The bar must have a constant cross section and be homoge-

neous.

3. The stress must not exceed the proportional limit.

Shearing deformation

Shearing forces cause a shearing deformation, just as axial forces

cause elongations, but with an important difference. An element subject
to tension undergoes an increase in length; an element subject to shear
does not change the length of its sides, but undergoes a change in shape
from a rectangle to a parallelogram, as shown in Fig. 2-5.
The action may be visualized for the present as equivalent to the
infinitesimal sliding of infinitely thin layers past each other, thereby
resulting in the total shearing deformation 8S in the length L. The actual
action is more complex than that pictured and will be discussed more
fully in Art. 9-9.
The average shearing strain is found by dividing 8 by L. In Fig.
S

2-5, this defines tan y = 8J L. However, since the angle y is usually

very small, tan y «
y and we obtain

Y- (2-5)

More precisely, the shearing strain is defined as the angular change

between two perpendicular faces of a differential element.
The relation between shearing stress and shearing strain, assuming
Hooke's law to apply to shear, is
t - Gy (2-6)

/
r*i
/
/
/
/
/ /
/ 7 / L
/
/
/

Figure 2-5. Shear deformation.

40 2 Simple Strain

in which G represents the modulus of elasticity in shear, more com-

monly called the modulus of rigidity. The relation between the shearing
deformation and applied shearing forces is then expressed by

*' <2
"7>
' He
in which V is the shearing force acting over the shearing area A s Note
.

the similarity of this result with Eq. (2-4).

ILLUSTRATIVE PROBLEMS

201. Compute the total elongation caused by an axial load of 100

kN applied to a flat bar 20 mm thick, tapering from a width of 120 mm
to 40 mm in a length of 10 m as shown in Fig. 2-6. Assume E = 200 x
10
9
N/m 2 .

20 mm thick
20 mm 60 mm
P= 1
kN
100

Figure 2-6.

Solution: Since the cross-sectional area is not constant, Eq. (2-4) does

not apply directly. However, it may be used to find the elongation in a

differential length for which the cross-sectional area is constant. Then
the total elongation is the sum of these infinitesimal elongations.
At section m-n, the half width v (mm) at a distance x (m) from
the left end is found from geometry to be
y - 20 60-20
10
or v = (4x + 20) mm
and the area at that section is

A = 20(2>>) = (160.x -I- 800) mm :

At section m-n, in a differential length dx, the elongation may be

found from Eq. (2-4):

5 = PL dS = (100 x \(P)dx
AE (160.x -I- 800)(10~ 6 )(200 x 10
9
)

0.500 dx
160jc + 800
 .

2-3 Hooke's Law: Axial Deformation 41

from which the total elongation is

10 dx _ 0.500
r
o
2400
- (3.13 X 10" 3 ) In—
800
^= 3.44 X 10" 3 m = 3.44 mm Ans.

202. Two steel bars AB and BC support a load P - 30 kN as

of AB is 300 mm area of BC is 500 mm
2 2
shown Area
in Fig. 2-7a. ; . If
E= 200 GPa, compute the horizontal and vertical components of the
movement of B.

3 m

P= 30 kN
(a) (b)

Figure 2-7.

Solution: We begin by computing the deformations produced in each

bar by P. From statics, we obtain PAB = 50 kN (tension) and

?bc = 40 kN (compression). The corresponding deformations are

3
x
8 = PL 8ab
J
AB
— (50 10 )(5000)
AE (300 x 10" 6 )(200 x 10
9
)
= 4.17 mm lengthening
3
(40 X 10 )(4000)
8nr
'BC
~
9
(500 x 10" 6 )(200 x 10 )

= 1 .60 mm shortening
To analyze the effect of these deformations on the movement of B,
imagine first that bars AB and BC are disconnected at B so that they
undergo the deformations pictured (greatly exaggerated) in Fig. 2-7b.
To refasten the bars, rotate them about A and C to meet at B"
However, the arcs generated in these rotations are so small that they
may be effectively replaced by straight lines drawn perpendicular to AB
and BC, respectively; these lines, intersecting at B\ determine the
42 2 Simple Strain

B ^bc
=1 -60 mm

Figure 2-8.

effective final position of 5. The deformations 8AB and 55C are drawn to
a larger scale in Fig. 2-8, in which the total movement of B is the vector
BB ' or 8 directed as shown.
From Fig. 2-8 it is evident that the horizontal component of 8 is

$h
m 8B c ~ 1-60 mm rightward Arts.

However, 8h is also equal to the algebraic sum of the horizontal

components of 8AB and the unknown length x, so that
8h = x sin — 8AB cos

from which

1.60 = x(\ ) - 4.17(f), x = 8.23 mm

This value of x is used to determine y, i.e., 8V which
, is the sum of
the vertical components of 8AB and x:
8V = 8AB sin + x cos
= 4.17(f) + 8.23(f)
= 9.09 mm down Arts.

If we return to Fig. 2-7b, we may now compute the magnitude of

the angle through which bars AB and BC rotate. We obtain

X 8.23
U
<*ab
= = 1.65 x 10" 3 rad = 0.0945°
Lab 5000
and
9.09
OL D/~
l
/K
= y = 2.27 x 10~ 3 rad = 0.130°
BC 4000
2-3 Hooke's Law: Axial Deformation 43

These rotations are so small that it is justifiable to assume that the

directions of 8AB and 8BC coincide with the original directions of bars
AB and BC.

PROBLEMS

203. During a stress-strain test, the unit deformation at a stress

of 35 MN/m was observed to be 167 X 10" m/m and at a stress of
2 6

140 MN/m was 667 X 10" 6 m/m. If the proportional limit was 200
2
it

MN/m what the modulus of elasticity? What the strain corre-

2
, is is

sponding to a stress of 80 MN/m ? Would these results be valid 2

the if

proportional limit were 150 MN/m Explain.

2
.

Arts. E = 210 X 109 N/m2 € = 381 X 10~ 6 m/m ;

204. A uniform bar of length L, cross-sectional area A y

and unit
mass p suspended vertically from one end. Show that its total
is

elongation is 8 = pgL /2E. If the total mass of the bar is M, show also
2

that 8 = MgL/2AE.
205. A
rod having a cross-sectional area of 300
steel
2
and a mm
length of 150 m
is suspended vertically from one end. It supports a

tensile load of 20 kN at the lower end. If the unit mass of steel is 7850
kg/m 3 and E = 200 X 103 2
MN/m
find the total elongation of the rod.
,

{Hint: Use the results of Problem 204.) Arts. 8 = 54.3 mm

206. A steel wire 10 m long, hanging vertically supports a tensile
load of 2000 N. Neglecting the weight of the wire, determine the
required diameter if the stress is not to exceed 140 MPa and the total
elongation is not to exceed 5 mm. Assume E = 200 GPa.
207. A steel tire, 10 mm thick, 80 mm wide, and of 1500.0 mm
inside diameter, is heated and shrunk onto a steel wheel 1500.5 mm in
diameter. If the coefficient of static friction is 0.30, what torque is

required to twist the tire relative to the wheel? Neglect the deformation
of the wheel. Use E= 200 GPa. Ans. T- 75.0 kN m •

160 mm
2
208. An aluminum
bar having a cross-sectional area of
carries the axial loads at the positions shown in Fig. P-208. If E — 70
GPa, compute the total deformation of the bar. Assume that the bar is
suitably braced to prevent buckling.

35 kN 15 kN 30 kN 10 kN

L— 0.8 4* 1.0—4-0.6-J
m m m
Figures P-208 and P-209.
 *

44 2 Simple Strain

209. Solve Problem 208 if the magnitudes of the loads at the ends
are interchanged, i.e., if the load at the left end is 10 kN and that at the

right end is 35 kN. Ans. 8 = 1.61 mm (contraction)

210. An aluminum tube is fastened between a steel rod and a
bronze rod as shown in Fig. P-210. Axial loads are applied at the
positions indicated. Find the value of P that will not exceed a maximum
overall deformation of 2 mm or a stress in the steel of 140 MN/m2 in ,

MN/m 120 MN/m Assume

2 2
the aluminum of 80 , or in the bronze of .

that the assembly is suitably braced to prevent buckling and that

Es = 200 x 10
3
MN/m Ea = 2
, 70 X 10
3
MN/m 2
, and Eb = 83 X 3
10
MN/m 2
.

Bronze Aluminum Steel

A =450 mm 2 A = 600 mm 2
A - 300 mm 2

3P AP IP

-0.6 m— •1.0

Figure P-210.
m 0.8 m A

211. The rigid bars shown in Fig. P— 21 1 are separated by a roller

at and pinned at A and D. A steel rod at B helps support the load of
C
50 kN. Compute the vertical displacement of the roller at C.
Ans. 2.82 mm

///////
E = 200 X 10 9 N/m 2
A = 300 mm 2

L = 3m |50kN
2 m f 2 m
> 3 m 1.5 m rf
Figure P-211.

212. A uniform concrete slab of mass is to be attached, as M

shown in Fig. P-212, to two rods whose lower ends are initially at the
same level. Determine the ratio of the areas of the rods so that the slab
will remain level after it is attached to the rods. Ans. A a /A s = 8.57
2-3 Hooke's Law: Axial Deformation 45

/////

Aluminum
/////
E = 70 GPa
Steel
L = 6m
£ = 200 GPa lm
L = 3 m 5 m- ««-H

Figure P-212.

213. The rigid bar AB, attached to two vertical rods as shown in
Fig.P-213, is horizontal before the load P is applied. If the load P= 50
kN, determine its vertical movement.

/////

/////
Aluminum
Steel
L =4m
L =3m A = 500 mm 2
A = 300 mm 2
E = 70 GPa
£ = 200 GPa

-•-2 m-*; 3 m

Figure P-213.

214. The rigid bars AB and CD shown in Fig. P-214 are sup-
ported^by pins at A and C and the two rods. Determine the maximum
force Pwhich can be applied as shown if its vertical movement is
limited to 5 mm. Neglect the weights of all members.
Ans. P- 76.3 kN
^/Aluminum
L =2m
A = 500 mm 2
E = 70 GPa
3m 5
Steel
L =2m
A = 300 mm 2
E = 200 GPa

Figure P-214.
46 2 Simple Strain

A round bar of length L tapers uniformly from a diameter

215.
D one end to a smaller diameter d at the other. Determine the
at
elongation caused by an axial tensile load P.
Arts. 8 = APL/mEDd
216. A uniform slender rod of length L and cross-sectional area
A is rotating in a horizontal plane about a vertical axis through one end.
If the unit mass of the rod is p, and it is rotating at a constant angular
velocity of w rad/s, show that the total elongation of the rod is

f*>
2
L 3 /3E.
217. As shown in Fig. P-217, two aluminum rods AB and BC,
hinged to rigid supports, are pinned together at B to carry a vertical
load P = 20 kN. If each rod has a cross-sectional area of 400 mm 2 and
70 x 10 MN/m compute the elongation of each rod and the
3 2
E ,

horizontal and vertical displacements of point B. Assume a = 30° and

9 m 30°. Ans. 8h - 0.412 mm; 8V = 3.57 mm

Figures P-217 and P-218.

218. Solve Problem 217 if rod AB is of steel, with E=

200 X 103
MN/m 2
. Assume a = 45° and = 30°; all other data remain un-
changed.
A round bar of length L, tapering uniformly from a diame-
219.
ter Done end to a smaller diameter d at the other, is suspended
at
vertically from the large end. If p is the unit mass, find the elongation
caused by its own weight. Use this result to determine the elongation of
a cone suspended from its base.
2
Pg L (D + d) pgL 2d 2
Ans. 8 =
6E(D - d) 3ED(D - d)
 2-4 Poisson's Ratio: Biaxial and Triaxial Deformations 47

• 2-4 POISSON'S RATIO: BIAXIAL AND

TRIAXIAL DEFORMATIONS

Another type of elastic deformation is the change in transverse dimen-

sions accompanying axial tension or compression. Experiments show
that if a bar is lengthened by axial tension, there is a reduction in the
transverse dimensions. Simeon D. Poisson showed in 1811 that the ratio
of the unit deformations or strains in these directions is constant for
stresses within the proportional limit. Accordingly this ratio is named
after him; it is denoted by v and defined by

„=-^=-ii (2-8)
c c

where €x is the strain due only to stress in the X direction, and c^ and cz
are the strains induced in the perpendicular directions. The minus sign
indicates a decrease in transverse dimensions when €x is positive, as in
the case of tensile elongation.
Poisson's ratio permits us to extend Hooke's law of uniaxial stress
to the case of biaxial stress. Thus, if an element is subjected simulta-
neously to tensile stresses in the X and Y directions, the strain in the X
direction due to the tensile stress ox is ax /E. Simultaneously the tensile
stress oy will produce lateral contraction in the X direction of the
amount vay / E, so the resultant unit deformation or strain in the X
direction will be

2 " 9>
'--E-"-k <

Similarly the total strain in the Y direction is

«,=5-4 < 2 - 10 >

If desired, Eqs. (2-9) and (2-10) can be solved to express the

stresses in terms of the strains as follows:

(<x + V€y )E (€, + P€X )E

A further extension of the above discussion results in the following

expressions for strains caused by simultaneous action of triaxial tensile
stresses:

1 r
*x - *(Oy + «J]

V|k - v(o
z + "*)]
(2-12)

-v(ax +
«z '

-it- ",)]
48 2 Simple Strain

All the above equations are valid for compressive effects also; it is only
necessary to assign positive signs to elongations and tensile stresses,
and, conversely, negative signs to contractions and compressive stresses.
An important relation* between the constants £", G, and v for a
given material is expressed by

C (2 " ,3)
=2(TT7)
which is useful for computing values of v when E and G have been
determined. Common values of Poisson's ratio are 0.25 to 0.30 for steel,
approximately 0.33 for most other metals, and 0.20 for concrete.

ILLUSTRATIVE PROBLEMS

220. A specimen of any given material is subjected to a uniform

triaxial stress. Determine the theoretical maximum value of Poisson's
ratio.

Solution: If we add the three relations in Eq. (2-12), we obtain

«x + S + c
*
=
~TZ< < -
+ °y + °^ (a)

For a uniform triaxial stress, we have €x = c^ = ez = c, and ox =

o - oz - o. Hence Eq. (a) reduces to

-c-»(|)
Since both c and a must be of the same sign, it follows that
(1 - 2v) must be positive; that is,

1 - 2v £
from which

v ^\ Ans.
221. A
aluminum shaft of 80-mm diameter fits concentri-
solid
Compute the minimum internal diameter of
cally in a hollow steel tube.
the steel tube so that no contact pressure exists when the aluminum
shaft carries an axial compressive load of 400 kN. Assume v = \ and
EQ - 70 X 109 N/m 2 .

Solution: The axial compressive stress in the aluminum is

3
P 400 x 10
= ox = = -79.6 MN/m,, 2
~A 2

J(0.080)
•This relation is proved in Art. 9-11
2-4 Poisson's Ratio: Biaxial and Triaxial Deformations 49

For uniaxial stress, the transverse strain is

ax^ i/-79.6xio6 \
r
[* --*--'?] S--JI 70x10* J
- 379 x 10- 6 m/m
Therefore, the required diametrical clearance is

[5 = cL] 6, = (379 X 10" 6 )(80) - 0.0303 mm

The required internal diameter of the tube is found by adding this
clearance to the original diameter of the aluminum shaft, thus giving

D= 80 + 0.0303 = 80.0303 mm Arts.

PROBLEMS

222. A solid cylinder of diameter d carries an axial load P. Show

that its change in diameter is 4Pv /mEd.
^223^) A rectangular aluminum block is 100 mm long in the X
direction, 75 mm wide in the Y direction, and 50 mm thick in the
Z direction. It is subjected to a triaxial loading consisting of a uniformly
distributed tensile force of 200 kN in the X direction and uniformly
distributed compressive forces of 160 kN in the Y direction and 220 kN
in the Z direction. If v =\ and E = 70 GPa, determine a single
distributed loading in the X direction that would produce the same Z
deformation as the original loading. Arts. 410 kN tension
224. A welded steel cylindrical drum made of 10-mm plate has
an internal diameter of 1.20 m. By how much will the diameter be
changed by an internal pressure of 1.5 MPa? Assume that Poisson's
ratio is 0.30 and E - 200 GPa.

225. A 50-mm-diameter steel tube with a wall thickness of 2 mm

just fits in a rigid hole. Find the tangential stress if an axial compressive
load of 10 kN is applied. Assume v = 0.30 and E = 200
X 10
9
N/m2 Neglect the possibility of buckling.
.

226. A 150-mm-long bronze tube, closed at its ends, is 80 in mm

diameter and has a wall thickness of 3 mm. It fits without clearance in
an 80-mm hole in a rigid block. The tube is then subjected to an
internal pressure of 4.00 MN/m 2
Assuming v = \ and E — 83
.

X 10
3
MN/m 2
, determine the tangential stress in the tube.
Arts. 8.89 MN/m2
227. A 200-mm-long aluminum tube, closed at its ends, is 100
mm in diameter with a wall thickness of 2 mm. If the tube just fits

between two rigid walls at zero internal pressure, determine the longitu-
dinal and tangential stresses at an internal pressure of 4.00
2
MN/m .

= 70 X 10 N/m
9
Assume v \ and E = 2
.
50 2 Simple Strain

2-5 STATICALLY INDETERMINATE MEMBERS

There are certain combinations of axially loaded members in which the

equations of static equilibrium are not sufficient for a solution. This
condition exists in structures where the reactive forces or the internal
resisting forces over a cross section exceed the number of independent
equations of equilibrium. Such cases are called statically indeterminate
and require the use of additional relations which depend upon the
elastic deformations in the members. The cases are so varied that they
can best be described by sample problems illustrating the following
general principles:

1. To a free-body diagram of the structure, or a part of it, apply

the equations of static equilibrium.

2. If there are more unknowns than independent equations of

equilibrium, obtain additional equations from the geometric relations
between the elastic deformations produced by the loads. To define these
relations clearly, you will find it helpful to draw a sketch that ex-
aggerates the magnitudes of the elastic deformations.

ILLUSTRATIVE PROBLEMS

228. The short concrete post in Fig. 2-9 is reinforced axially with
six symmetrically placed steel bars, each 600 mm 2
in area. If the applied
load P is 1000 kN, compute the stress developed in each material. Use

300 mm

Bearing
N
plate

*c=<5,

Figure 2-9. Reinforced concrete post.

2-5 Statically Indeterminate Members 51

the following moduli of elasticity: for steel, E =

s
200 X 10
9
N/m2 ; for
concrete, Ec - 14 X 10
9
N/m 2
.

Solution: As shown in Fig. 2-9, the applied load and the resisting
forces on any transverse section m-n form a collinear system. From
equilibrium, we have
[2F=0] P, + Pc =lxl0 N 6
(a)

Since no other equations of static equilibrium are available to

indicate in what proportion the load is distributed to each material, we
consider the elastic deformation of the structure. By symmetry, it is

evident that the bearing plate causes the steel and concrete to deform
equally. Hence, applying 8 = oL/E to these equal deformations, we
obtain

»-« (fi-(f),
from which, by canceling out the equal lengths of steel and concrete and
substituting the moduli of elasticity, we have the following relation
between the stresses:

°s - y° c
=
l4"
ac
= c O
Equation (b) may be called the governing relation for stress. Note
that depends only upon the fact that both materials deform equally.
it

This relation is independent of the loads or areas and remains valid

provided the proportional limit of neither material is exceeded.
We may now use P = aA to rewrite Eq. (a):
a,[6(600 X 10" 6 )] + ac {[(300 X 300) - 6(600)] X 10" 6 } = 1 X 10
6

a,(3.6 X 10" 3 ) + ac (86.4 X 10~ 3 ) = 1 X 10

6

whence, replacing as by 14.3ac from Eq. (b), we obtain

14.3ac (3.6 X 10" 3 ) + ac (86.4 X 10" 3 ) = 1 X 10

6

ac = 7.25 X 106 N/m2 - 7.25 MN/m2 Arts.

and from Eq. (b)

as = 14.3ac = 104 MN/m 2

Arts.

229. In the preceding problem, assume the allowable stresses to

be as = 120 MN/m 2
and oc = 6 MN/m 2
. Compute the maximum safe
axial load P which may be applied.

Solution: The unwary student may substitute the allowable stresses

only in the equation of static equilibrium. This would be incorrect
because would not consider the equal deformations of the materials.
it

From Eq. (b) in the preceding problem, we saw that equal deformations
52 2 Simple Strain

produce the following governing relation between the stresses:

os - 14.3af

From this relation we note that if the concrete were stressed to its

limit of 6 MN/m 2
, the corresponding stress in the steel would be

os - (14.3)6 = 85.8 MN/m 2

Therefore, the steel could not be stressed to its limit of 120 MN/m2
without overstressing the concrete. The actual working stresses are
thereby determined to be ac = 6 MN/m 2
and os = 85.8 MN/m 2
. These
values are substituted in the equation of static equilibrium [Eq. (a) in
the preceding problem] to obtain

/>=/>, + />
c
= as A s + oc A c
6 3
= (85.8 x 6
10 )(3.6 x 1(T 3 ) + (6x 10 )(86.4 x 1(T )
= 827 kN Ans.
230. A copper rod is inserted into a hollow aluminum cylinder.
The copper rod projects 0.130 mm as shown in Fig. 2-10. What
maximum load P may be applied to the bearing plate? Use the data in
the accompanying table.

COPPER ALUMINUM
2
Area (mm ) 1200 1800
E (GPa) 120 70
Allowable stress (MPa) 140 70

Bearing
plate
I
^ ~
1*"

250 mm
Aluminum
cylinder

AV^vA '

Figure 2-10.

Solution: To find the governing relation between the stresses, we

consider the elastic deformations, which are shown exaggerated in
Fig. 2-10. We obtain

[Se = 8a 4- (0.130 x I0" 3 )m]

(fHf). + «"*>* 10" 3

)

of (0.25) o (0.25)
x 10" 3 )
fl

9
+ (0.130
120 x 10 70 x 109
2-5
—
Statically Indeterminate Members 53

from which
oc = \J\aa + (62.4 X 10
6
)
(a)

From this stresses, we note that using aa = 70

governing relation for the
MPa copper to 182 MPa. Therefore copper governs,
will overstress the
and the corresponding stress in the aluminum is determined from Eq.
(a) to be

(140 X 10
6
) - 1.71aa + (62.4 X 10
6
) aa = 45.4 MPa
The total safe load is given by
P=P C + Pa = oc A c + oa A a
whence, substituting the working stresses just determined, we obtain
6
P- (140 x 6
10 )(1200 X 10" 6 ) + (45.4 x 10 )(1800 X 10" 6 )
= 250 kN Arts.

231. A horizontal bar of negligible mass, hinged at A in

Fig. 2-1 la and assumed rigid, is supported by a bronze rod 2 m long

//s/y//

LU
Bronze 2m

Steel lm
0.6 mn 1.0 m 0.8 m
r

50 kN
(a)

Original
Ps ^b f position
1.0 m 0.8 mi
1
i. r 1" I

Final
position 50 kN

(b)
Figure 2-11.
54 2 Simple Strain

and a steel rod 1 m long. Using the data in the accompanying table,
compute the stress in each rod.

STEEL BRONZE
2
Area(mm ) 600 300
E (GN/m 2 ) 200 83
Proportional limit (MN/m 2 ) 240 140

Solution: The free-body diagram of the bar in Fig. 2- lib shows it to

be statically indeterminate to the first degree; that is, there is one more
unknown force than can be found from the equations of static
equilibrium. A moment summation about A gives one relation between
the forces in the rods:

[LMA - 0] 0.6P5 + \.6Pb = 2.4(50 x 10

3
) (a)

Another relation between these forces is obtained from the elastic

deformations of the rods. From the similar triangles formed in Fig.
2-1 lb (recall that the horizontal bar is assumed rigid), we obtain

A.
0.6
4
1.6
or —(
0.6\
—
AE]
\
S
=-L(Ik\
\.6\AE) b
1 ^(1) 1 ^(2)
0.6 (600)(200) 1.6 (300)(83)
whence
J»,-3.61/»„ (b)
Note that the areasneed not be expressed in square meters, since the
~
conversion factor of 10 6 appears on both sides of the equation and
thus will cancel out. Similarly the factor 109 in the moduli of elasticity is
not shown in the foregoing equation since it also cancels.
Solving Eq. (a) and (Z>), we obtain
P =
s
115 kN
Pb = 31.9 kN

Computing the stresses,

^
*s
= J15 ii 10^ =192xl06
600 x 10" 6

= 192 MN/m 2
Arts.

_£_2!£xjo> xloSN/m2
At 300 X 10" 6
= 106 MN/m 2
Ans.

Since both stresses are less than the proportional limits, the
answers may be accepted. If the steel stress, for example, had exceeded
the proportional limit, the results would not have been valid and a
 -

2-5 Statically Indeterminate Members 55

redesign would be required. Perhaps the simplest redesign would be to

making it less rigid. It may be
increase the length of the steel rod, thus
noted here that generally the most rigid parts of an indeterminate
structure carry the most load. This is a fundamental principle in the
theory of indeterminate structures and is known as the principle of
ridigities.*

PROBLEMS

A steel bar 50 mm in diameter and 2 m long is surrounded

by a shell of cast iron 5 mm thick. Compute the load that will compress
the combined bar a total of 1 mm in the length of 2 m. For steel,
E = 200 X 109 N/m2 and for cast iron, E = 100 X 109 N/m2
, .

Ans. P= 240 kN
233. A reinforced concrete column 250 mm in diameter is de-
signed to carry an axial compressive load of 400 kN. Using allowable
stresses of ac = 6 MPa and os = 120 MPa, determine the required area
of reinforcing steel. Assume that Ec ^ 14 GPa and Es = 200 GPa.
Ans. A s = 1320 mm
2

234. A timber block 250 mm square is reinforced on each side by

a steel plate 250 mm wide and mm thick. Determine the thickness
t / so
that the assembly will support an axial load of 1200 kN without
exceeding a maximum timber stress of 8 MN/m2 or a maximum steel
stress of 140 MN/m 2
. For timber, E= 10 X 10
3
MN/m 2
; for steel,
E= 200 X 10
3
MN/m 2
.

L235S A rigid block of mass M is supported by three symmetri-

cally spaced rods as shown in Fig. P-235. Each copper rod has an area
of 900 mm 2
E = 120 GPa; and the allowable stress is 70 MPa. The
;

steel rod has an area of 1200

2
mm
E — 200 GPa; and the allowable
;

stress is 140 MPa. Determine the largest mass M which can be sup-
ported. Ans. M = 22.3 X 10 kg 3

Copper Steel Copper

160 mm 240 mm 160 mm

7777777V

Figures P-235 and P-236.

See J. I. Parcel and G. A. Maney, Statically Indeterminate Stresses, 2nd ed.

Wiley, New York, 1936, p. 109.

 •

56 2 Simple Strain

236. In Problem 235, how should the length of the steel rod be
changed so that each material will be stressed to its allowable limit?
The lower ends of the three bars in Fig. P-237 are at the
237.
same homogeneous 18 Mg block is attached. Each
level before the rigid

steel bar has an area of 600 mm and E = 200 GN/m

2 2
For the bronze .

bar, the area is 900 mm and E = 83 GN/m

2 2
Find the stress developed.

in each bar. Arts. os - 124 MN/m 2

; ob = 32.0 MN/m 2

//////

•//( s/s. s
N CO
B c B
o
u *0>

CD
p PQ 1 p
•—
ii
GO I
•^ ^

18 Mg
Figure P-237.

The rigid platform in Fig. P-238 has negligible mass and

238.
rests on two aluminum bars, each 250.00 mm long. The center bar is
steel and is 249.90 mm long. Find the stress in the steel bar after the

center load P — 400 kN is applied. Each aluminum bar has an area of

1200 mm and E = 70 GPa. The steel bar has an area of 2400 mm and
2 2

E = 200 GPa.

S77
Figure P-238.

Three steel eye-bars, each 100

239. by 25 mm
in section, are mm
to be assembled by driving 20-mm-diameter drift pins through holes
drilled in the ends of the bars. The center-line spacing between the holes
is 10 m in the outer two bars but is 1.25 mm shorter in the middle bar.
2-5 Statically Indeterminate Members 57

Find the shearing stress developed in the drift pins. Neglect local
deformation at the holes and use Es = 200 GPa.
Ans. r = 66.2 MPa
240. As shown P-240, three steel wires, each 30
in Fig. in mm 2

area, are used to lift a mass M. Their unstretched lengths are 19.994 m,
19.997 m, and 20.000 m. (a) If M
= 600 kg, what stress exists in the
longest wire? (b) If M
= 200 kg, determine the stress in the shortest
wire. Use E = 200 GN/m
2
.

/////////;

M
Figure P-240.

( 2AX~ The assembly P-241 consists of a rigid bar AB

in Fig.
(having negligible mass) pinned at O
and attached to the aluminum rod
and the steel rod. In the position shown, the bar AB is horizontal and
there is a gap A = 4 mm between the lower end of the aluminum rod
and its pin support at D. Find the stress in the steel rod when the lower
end of the aluminum rod is pinned to the support at D.
Ans. ar = 174 MPa

rO.b m—
*--* 1.2 m
A C B
7^7
Steel Aluminum
A = 300 mm2 A = 400 mm 2
E = 200 GPa E = 70 GPa
L= 1.5 m

DS. A = 4 mm

Figure P-241.
 V

58 2 Simple Strain

242. A
homogeneous rod of constant cross section is attached to
unyielding supports.It carries an axial load P applied as shown in Fig.

P-242. Prove that the reactions are given by R = Pb/L and R 2 = l

Pa/ L. (Note that these reactions are equivalent to those of a simply

supported beam carrying a concentrated load.)

/ /

/ ^2
/
/ • •
/ r _ <
3"" " '

Figure P-242.

243. A homogeneous bar with a cross-sectional area of 500 mm2

is attached to rigid supports. It carries the axial loads P = x
25 kN and
P2 = 50 kN, applied as shown in Fig. P-243. Determine the stress in the

segment BC. (Hint: Use the results of Problem 242, and compute the
reactions caused by P, and P2 acting separately. Then use the principle
of superposition to compute the reactions when both loads are applied.)

B P x C P2

*0.60 EQ; 1.20 m 0.90 m-

Flgure P-243.

244. The bar shown in Fig. P-244 is firmly attached to unyield-

ing supports. Find the stress caused in each material by applying an
axial load P = 200 kN.
Ans. oa = 62.8 MN/m2 ; as = 120 MN/m 2

245. Refer to Problem 244. What maximum load P can be

applied without exceeding an allowable stress of 70 MPa for aluminum
or 120 MPa for steel? Can a larger load P be carried if the length of the
aluminum rod changed, the length of the
is steel portion being kept the
same? If so, determine this length.

200 mm-»4- 300 mm-

Aluminum
Steel
E 70 GPa E - 200 GPa
A 900 mm 2
A = 1200 mm 2
Figures P-244 and P-245.
2-5 Statically Indeterminate Members 59

246. A rod is composed of three segments shown in Fig. P-246

and carries the axial loads P =x
120 kN and P2 = 50 kN. Determine the
stress in each material if the walls are rigid.
Ans. as - 122 MN/m2
247. Solve Problem 246 if the left wall yields 0.60 mm.

<*- 600 mm 400 mm 300 mm*

Bronze Aluminum Steel

A = 2400 mm 2 A = 1200 mm2 A = 600 mm 2

E= 83 GPa E = 70 GPa E = 200 GPa

Figures P-246 and P-247.

248. A steel tube 2.5 mm thick just fits over an aluminum tube
2.5 mm thick. If the contact diameter is 100 mm, determine the contact
pressure and tangential stresses when the outward radial pressure on the
aluminum tube is p = 4 MN/m 2
. Here, E = s
200 X 10
9
N/m 2 , and
Ea = 70 X 109 N/m 2 .

Ans. pc = 2.96 MN/m 2

; os = 59.2 MN/m2 ; aa = 20.8 MN/m2
249. In Problem 248, assume that there is a radial clearance of
0.01 mm between the tubes before the internal pressure of 4 MN/m 2
is

applied to the aluminum tube. Solve for the contact pressure and
tangential stresses.

250. In the assembly of the bronze tube and steel bolt shown in
Fig. P-250, the pitch of the bolt thread is 0.80 and the cross-sec- mm
tional area of the bronze tube is 900
2
and of the steel bolt is 450 mm
mm 2
The nut is turned until there is a compressive stress of 30
.
2
MN/m
in the bronze tube. Find the stress in the bronze tube if the nut is then
given one additional turn. How many turns of the nut will reduce this
stress to zero? Take E as in Problem 246.
Ans. ab = 75.4 MN/m 2
; 1.66 turns

Figure P-250.
60 2 Simple Strain

As shown in Fig. P-251, a rigid beam with negligible mass

251.
is pinned at O and supported by two rods, identical except for length.

Determine the load in each rod \\ P — 30 kN.

Ans. PA = 9.10 kN; PB = 11.94kN

-* -2m •+• \5 m- — *~ -•-1.5 m—+•

i o )
777770
L- 1.5 in
\r L-
A
77777
B
/7? //
Figure P-251.

252. As shown in Fig. P-252, a rigid beam with negligible mass

ispinned at one end and supported by two rods. The beam was initially
horizontal before the load P was applied. Find the vertical movement of
P if P = 120 kN. Ans. 2.92 mm

\\\w

\\\\\
Steel

A = 600 mm 2 Aluminum
E = 200 GPa A = 900 mm 2
L = 4m E = 70 GPa
L = 3m

>
3 m 2 m
1 m

Figure P-252.

253. A rigid bar of negligible mass, pinned at one end, is sup-

ported by a steel rod and a bronze rod as shown in Fig. P-253. What
maximum load P can be applied without exceeding a stress in the steel
of 120 MN/m 2
or in the bronze of 70 MN/m 2
?
2-5 Statically Indeterminate Members 61

Steel / // / /
A = 900 mm 2
Bronze
E = 200 GPa A = 300 mm 2
L =3m E = 83 GPa
L = 2m

lml
U— 2 m—* 3 m-
^1
Figure P-253.

254. Shown in Fig. P-254 is a section through a balcony. The

total uniform load of 600 kN is supported by three rods of the same

area and material. Compute the load in each rod. Assume the floor to
be rigid, but note that it does not necessarily remain horizontal.
Arts. PB = 183 kN
zz zy
///// B

6 m 6 m
5 m
2m
4 m

k— 3 m
*
600 kN
Figure P-254.
3 m-

255. Three rods, each with an area of 300

2
jointly support mm ,

the load of 10 kN, as shown in Fig. P-255. Assuming there was no slack
or stress in the rods before the load was applied, find the stress in each
rod. Here, Es = 200 X 10 N/m and Eb = 83 X 10 N/m
9 2 9 2
.

Figure P-255.
62 2 Simple Strain

Three bars, AB, AC, and AD, are pinned together to sup-
256.
port a load P = 20 kN as shown in Fig. P-256. Horizontal movement is
prevented at joint A by the short horizontal strut AE. Determine the
stress in each bar and the force in the strut AE. For the steel bar,
A - 200 mm 2 and E - 200 GPa. For each aluminum bar, A = 400
mm 2
and E= 70 GPa. Ans. PAE =\80N
257. Refer to the data in Problem 256, and determine the maxi-
mum value of P that will not exceed an aluminum stress of 40 MPa or a
steel stress of 120 MPa.

AJuminum\30 Aluminum

Figures P-256 and P-257.

2-6 THERMAL STRESSES

It is well known that changes in temperature cause bodies to expand or

contract, the amount of the linear deformation, 5 r being expressed by
,

the relation

6T = aL(lT) (2-14)

in which a is the coefficient of linear expansion, usually expressed in

units of meters per meter per degree of temperature change, L is the
length, and AT is the temperature change. Substituting these units in
Eq. (2-14) shows the dimensional unit of 8 T to be the same as that of
the length L.
If a temperature deformation is permitted to occur freely, as by
the use of expansion joints, no load or
stress will be induced in the
structure.But in some cases it may not be feasible to permit these
temperature deformations; the result is that internal forces are created
which resist them. The stresses caused by these internal forces are
known as thermal stresses.
A general procedure for computing the loads and stresses caused
when temperature deformation is prevented is outlined in these steps:

Imagine the structure relieved of all applied loads and con-

1.

straints so thattemperature deformations can occur freely. Represent

these deformations on a sketch, and exaggerate their effect.
2-6 Thermal Stresses 63

2. Now imagine sufficient loads applied to the structure to res-

tore it to the specified conditions of restraint. Represent these loads and
corresponding load deformations on the sketch for step 1.

3. The geometric relations between the temperature and load

deformations on the sketch give equations which, together with the
equations of static equilibrium, may be solved for all unknown quanti-
ties.

The following examples illustrate these steps applied in several

different types of problems.

ILLUSTRATIVE PROBLEMS

258. A steel rod 2.5 m long is secured between two walls. If the
load on the rod is zero at 20 °C, compute the stress when the tempera-
ture drops to -20°C. The cross-sectional area of the rod is 1200 mm2 ,

a = 11.7 /tm/(m °C), and E = 200 GN/m2 Solve, assuming (a) that
•
.

the walls are rigid and (b) that the walls spring together a total distance
of 0.500 mm as the temperature drops.

v-
3-^-P
[•-4
Figure 2-12. Rigid walls.

Solution:

Imagine the rod is disconnected from the right wall.

Part a.

Temperature deformations can then freely occur. A temperature drop

causes the contraction represented by 8 T in Fig. 2-12. To reattach the
rod to the wall will evidently require a pull P to produce the load
deformation 8P . From the sketch of deformations, we see that 8 T = 8P ,
or, in equivalent terms,

/Arr7W PL oL

whence
a = Ea(AT) = (200 X \(f)(UJ X 10" 6 )(40) = 93.6 X 10
6
N/m2
= 93.6 MN/m2 Arts.

Note that L cancels out of the above equation, indicating that the stress
is independent of the length of the rod.
 2 Simple Strain

r
Yield-'
Figure 2-13. Nonrigid walls.

Part b. When the walls spring together, Fig. 2-13 shows that the
free temperature contraction is equal to the sum of the load deformation
and the yield of the walls. Hence
8T = 8P -I- yield

whence, replacing the deformations by equivalent terms, we obtain

aL(AT) - —+
oL
yield

or

g(2 5)
'

(11.7
V
x 1(T 6 A A ; =
)(2.5)(40) Q
+ (0.5
V
x 10" 3
)
'
9
200 X 10

from which we obtain

o = 53.6 MN/m 2
Arts.

Notice that the yield of the walls reduces the stress considerably,
and also that the length of the rod does not cancel out as in Part a.

259. A rigid block having a mass of 5 Mg is supported by three

rods symmetrically placed, as shown in Fig. 2-14. Determine the stress
in each rod after a temperature rise of 40° C. The lower ends of the rods
are assumed to have been at the same level before the block was

"_"

Steel Steel
L = 0.5 m L = 0.5m

Figure 2-14. Free-body diagram.

2-6 Thermal Stresses 65

attached and the temperature changed. Note that symmetry dictates

that the block will remain horizontal. Use the data in the accompanying
table.

EACH STEEL ROD BRONZE ROD

Area(mm2) 500 900
9 9
E (N/m2) 200 X 10 83 X 10
a [/xm/(m* >C)] 11.7 18.9

Solution: With the block detached, the original lower position of the
rods is as shown in Fig. 2-15. With the rods free of any constraint, a
temperature cause the temperature deformations 8 T and 8 T in
rise will
the steel and bronze, respectively. When the rods are attached to the
rigid block after the temperature change has occurred, assume their
final horizontal level to be as shown. To attach them to the block, it will
be necessary to pull their expanded ends through the load deformations
8P and 8P by means of the loads Ps and Pb in the steel and bronze,
respectively. The free-body diagram of the block in Fig. 2-14 represents
the equal and opposite effects of the forces exerted by the rods upon the
block.
From the deformations shown in Fig. 2-15, we obtain the follow-
ing geometric relation between the deformations:

$ts + 8pM " 8n + $pb

or

(« la
Mt§L=^ a
Mi§),

Original

'T b
T
6P
"P<
Vi±
mm
Ps Final
/ level
Figure 2-15. Deformations.
66 2 Simple Strain

whence, substituting the given data, we have

6
(11.7 X 1(T )(0.5)(40)
y
(500 X 10" 6 )(200 X 10 )

P (\)
= (18.9 x 10" 6 )(1)(40) +
(900 x 10" 6 )(83 x 10
9
)

Simplifying the above equation, we obtain

Ps - 2.6SPb = 104 X 10
3
N (a)

Another relation between Ps and Pb is obtained from the free-

body diagram in Fig. 2-14.

[2 Y = 0] 2PS + Pb = (5000)(9.81) = 49.05 X 103 N (b)

Solving Eqs. (a) and (b) yields

Ps = 37.0 kN
Pb = -25.0 kN
The negative sign for Pb means that the load Pb acts oppositely to that
assumed; that is, the bronze rod is actually in compression and suitable
provision must be made to prevent buckling.
The stresses are

P 37.0 X 103
°*
500 X 10" 6
= 74.0 MN/m 2
(tension) Ans.
3
25.0 X 10
900 x 10~ 6
= 27.8 MN/m 2
(compression) Ans.

260. Using the data in Problem 259, determine the temperature

rise necessary to cause all the applied load to be supported by the steel
rods.

Solution: Instead of trying to use the results obtained in the solution

of Problem 259, we apply the three steps outlined in Art. 2-6. Imagine
the rods disconnected from the block and hanging freely, as in
Fig. 2-16. A temperature rise causes the temperature deformations 5 r
and 8 T .

Since the bronze rod is to carry no load, the final level of the steel
rods must coincide with the unstressed expanded length of the bronze.
If the rods are to be at the same final level, the steel rods must go

through a load deformation 8 P caused by the pulls Ps each of which ,

must be equal to one-half of the weight, or j(5000)(9.81) = 24.53 kN.

2-6 Thermal Stresses 67

Bronze

Steel Steel

. Original
/ level
1""
IP. $T ^p^ Final
8P */
;
level

Figure 2-16. Bronze rod supporting no load.

From Fig. 2-16, the geometric relation between the deformations

IS

8T = 6> 4- 8i

which can be written as

(aLAr) t = (aLAr) s +(^[-

whence

(18.9 X l(T 6 )(l)(Ar) = (11.7 x 10- 6 )(0.5)(Ar)

(24.53 X \tf){0.5)
10" 6 )(200 9
(500 x x 10 )

Ar=9.4°C Ans.
It is evident that a greater temperature rise will cause the bronze to
push against the rigid block, thereby causing compression of the bronze.
This confirms the result obtained in Problem 259, where the tempera-
ture rise was given to be 40° C.

PROBLEMS

261. A rod with a cross-sectional area of 150

steel
2
is mm
stretched between two fixed points. The tensile load at 20°C is 5000 N.
What will be the stress at — 20°C? At what temperature will the stress
be zero? Assume a - 11.7 jum/(m- °C) and E= 200 X 10
9
N/m 2 .

Ans. a = 127 MN/m T = 2

; 34.2°C
262. A steel rod is stretched between two rigid walls and carries a
tensile load of 5000 N at 20° C. If the allowable stress is not to exceed
130 MN/m 2
-20°C, what is the minimum diameter of
at the rod?
Assume a = 11.7 jum/(m- °C) and E = 200 GPa.
68 2 Simple Strain

263. Steel railroad rails 10 m long are laid with a clearance of

3 mm at a temperature of 15°C. At what temperature will the rails just

touch? What stress would be induced in the rails at that temperature if

there were no initial clearance? Assume a= 11.7 jLtm/(m-°C) and

E - 200 GPa.
264.At a temperature of 90°C, a steel tire 10 mm thick and 75
mm wide that is to be shrunk onto a locomotive driving wheel 1.8 m in

diameter just fits over the wheel, which is at a temperature of 20°C.

Determine the contact pressure between the tire and wheel after the
assembly cools to 20° C. Neglect the deformation of the wheel caused by
the pressure of the tire. Assume a=
11.7 jum/(m-°C) and E =
200 x 10 N/m
9 2
.

265. At 130°C, a bronze hoop 20 mm thick whose inside diame-

ter is 600 mm just fits snugly over a steel hoop 15 mm thick. Both hoops
are 100 mm wide. Compute the contact pressure between the hoops
when the temperature drops to 20 °C. Neglect the possibility that the
inner ring may buckle. For steel, E= 200 GPa and a = 11.7 jitm/
(m- °C). For bronze, E- 83 GPa and a = 19 jum/(m •
°C).
Ans. p - 2.86 MN/m 2

266. At 20 °C, a rigid slab having a mass of 55 Mg is placed upon

two bronze rods and one steel rod as shown in Fig. P-266. At what
temperature will the stress in the steel rod be zero? For the steel rod,
A = 6000 mm 2 E = 200 X 109 N/m2 and a = 11.7 jum/(m-°C). For
, ,

each bronze rod, A = 6000

2
mm
E - 83 X 109 N/m2 and a =
, ,

o
19.0/im/(m- C). Ans. T = 129°C

55 Mg

a
2
250 mm
PQ

;/?/;//, (7777 rrr

^rT7
50 mm
T
Figure P-266.

267.At 20° C, there is a gap A = 0.2 mm between the lower end

of the bronze bar and the rigid slab supported by two steel bars, as
shown in Fig. P-267. Neglecting the mass of the slab, determine the
stress in each rod when the temperature of the assembly is increased to
100°C. For the bronze rod, A = 600 2
mm
E = 83 x 109 N/m 2 and , ,
2-6 Thermal Stresses 69

800 mm

Figure P-267.

a = 18.9 /xm/(m- °C). For each steel rod, A = 400 mm E =2

, 200 X 10
9

N/m 2
, and a - 1 1.7 /xm/(m °C). •

268.An aluminum cylinder and a bronze cylinder are centered

and secured between two rigid slabs by tightening two steel bolts, as
shown in Fig. P-268. At 10°C no axial load exists in the assembly. Find
the stress in each material at 90 °C. For the aluminum cylinder, A =
1200 mm 2 E = 70 X 10 N/m and a = 23 /im/(m°C). For the
9 2
, ,

bronze cylinder, A = 1800 mm E = 83 X 10 N/m2 and a = 19.0

2 9
, ,

/im/(m- °C). For each steel bolt, A = 500 mm £ = 200 X 109 N/m2
2
, ,

and a - 1 1.7 jum/(m- °C). Ans. as = 33.7 MN/m

2

269. Resolve Problem 268 assuming there is a 0.05 gap mm

between the right end of the bronze cylinder and the rigid slab at 10°C.

20 20 mm
-75 mm*- —100 mm-*

C"
^

Alum.
Bronze
C=: = J-
Figures P-268 and P-269.

270. A steel cylinder is enclosed in a bronze sleeve; both simulta-

neously support a vertical compressive load of 250 kN which is applied
to the assembly through a horizontal bearing plate. The lengths of the
cylinder and sleeve are equal. Compute (a) the temperature change
that will cause a zero load in the steel, and (b) the temperature change
that will cause a zero load in the bronze. For the steel cylinder,
70 2 Simple Strain

A = 7200 mm 2
, E = 200 GPa, and a = /xm/(m°C). For the
11.7

bronze sleeve, A = 12 X 10
3
mm 2
, £= 83 GPa, and a = 19.0 /im/(m-

271. A bronze sleeve is slipped over a steel bolt and is held in

place by a nut that is tightened "finger-tight." Compute the temperature
change which will cause the stress in the bronze to be 20 MPa. For the
steel bolt, A = 450 mm E 2
,
- 200 GPa, and a -
/xm/(m- °C). For
1 1.7

the bronze sleeve, A = 900 mm E = 2

, 83 GPa, and a = 19.0 jLtm/(m-
°C).

272. For the sleeve-bolt assembly described in Problem 271,

6
assume the nut is tightened to produce an initial stress of 15 X 10
N/m in the bronze sleeve. Find the stress in the bronze sleeve after a
2

temperature rise of 70°C. Ans. 38.2 MN/m 2

273. The composite bar shown in Fig. P-273 is firmly attached

to unyielding supports. An axial load P = 200 kN is applied at 20°C.
Find the stress in each material at 60°C. Assume a = 11.7 fim/(m-°C)
for steel and 23.0 jnm/(m °C) •
for aluminum.
Ans. oa = 18.7 MN/m 2
; os = 181 MN/m 2

200 mm-H 300 mm

Aluminum
Steel
E = 70 X 10
9
N/m2 E= 200 X 10
9
N/m 2
A = 900 mm 2
A = 1200 mm 2

Figures P-273 and P-274.

274. At what temperature will the aluminum and steel segments

in Problem 273 have numerically equal stresses?

275. A composed of the three segments shown in

rod is

Fig. P-275. If P and P2 are each zero, compute the

the axial loads x

stress induced in each material by a temperature drop of 30 °C if (a) the

walls are rigid and (b) the walls spring together by 0.300 mm. Assume
a = 18.9 jLtm/(m°C) for bronze, 23.0 /xm/(m-°C) for aluminum, and
11.7fim/(m • °C) for steel.

Ans. (a) ^=118 MPa; (b) oa = 40.0 MPa

800 mm 500 mm- 400 mm-*-'

Bronze Aluminum Steel

A = 2400 mm* A = 1200 mm 2
A = 600 mm 2

E 83 X 10
9
N/m 2 E -
70 X 10
9
N/m 2 E - 200 X 10
9
N/m 2
Figures P-275 and P-276.
2-6 Thermal Stresses 71

276. Solve Problem 275 if P and P2

l
each equal 50 kN and the
walls yeld 0.300 mm when the temperature drops 50°C.

(_27Ls The bar AB is pinned at O and connected to two rods

rigid
as shown in P-277. If the bar AB is horizontal at a given
Fig.
temperature, determine the ratio of the areas of the two rods so that the
bar AB will be horizontal at any temperature. Neglect the mass of bar
AB. Arts. A s /A a - 0.516

3m 4 m
O,
Ai )JB
s?77?7
-t Sp-&r>St,
Aluminum Steel
E = 70 GPa £ = 200 GPa
o
a = 23.0/im/(m. C) a =11.7Mm/(m«°C)
L= 8m L= 8m

Figure P-277.

278. A rigid horizontal bar of negligible mass is connected to two

rods as shown in Fig. P-278. If the system is initially stress-free,

determine the temperature change that will cause a tensile stress of

60 MPa in the steel rod.

/////

Steel

3m A = 900 mm 2
E = 200
GPa
a =11.7Mm/(m-°C)

2 m 3m

P Bronze

2m A = 1200 2mm
E = 83 GPa
a = 18.9 Mm/(m • °C)
r
7T7
Figure P-278.
72 2 Simple Strain

279. For the assembly shown in Fig. P-279, determine the stress
in each of the two vertical rods if the temperature rises 40°C after the
load P = 50 kN is applied. Neglect the deformation and mass of the
horizontal bar A B.
Ans. os = 134 MPa (tension); oa = 11.3 MPa (compression)

L£ t^y

/////
Steel
Aluminum 600 mm 2
m A
=
4
A =900 mm 2 E 200 X 10 9 N/m 2
=
E = 70 X 10 9 N/m 2 3 m
a = 11.7/im/(m.°C)
a =23.0/um/(m-°C)

3 m

Figure P-279.
H-* 3 m
l
P= 50 kN

The lower ends of the three steel rods shown in Fig. P-280
280.
are at the same level before the force P = 600 kN is applied to the
horizontal rigid slab. For each rod, A = 2000 mm a = 11.7 /xm/
2
,

(m- °C), and E = 200 X 10 N/m Determine the relationship between

9 2
.

the force in rod C and the change in temperature AT, measured in

degrees Celsius. Neglect the mass of the rigid slab.

// /// // ///
B C
// U/
A
6 m 6 m
5 m
2m

3 m 3m
P = 600 kN
Figure P-280.

281. Four steel bars jointly support a mass of 15 Mg as shown in

Fig. P-281. Each bar has a cross-sectional area of 600 2
Find the mm .

load carried by each bar after a temperature rise of 50°C. Assume

a = .7 /im/(m °C) and E = 200 X 10 N/m
•
9 2
1 1 .

Ans. PA = PD = 21.5 kN; PB = Pc = 67.3 kN

2-6 Thermal Stresses 73

Figures P-281 and P-282.

282. Solve Problem 28 1 if bars A and D are steel and bars B and
C are aluminum. For aluminum, a — 23.0 /im/(m- °C) and E = 70 X
10
9
N/m 2 .

SUMMARY

Axial loads cause uniform stress- distribution that is computed by

*= (i-i)
£
and elongations determined from

«-§ (2-4)

Remember that Eq. (2-4) is valid only for homogeneous materials of

constant cross section, axially loaded to stresses below the proportional
limit.

Axial loads cause changes in lateral dimensions which are de-

termined by Poisson's ratio. For cases in which such lateral deforma-
tions are restricted, Hooke's law for biaxial or triaxial stresses must be
used.
Structures that are composed of two or more materials or are
statically indeterminate are solved by applying the equations of static
equilibrium in combination with additional equations obtained from the
geometric relations between the elastic deformations.
Thermal stresses are computed by determining the relations be-
tween the thermal deformations

8T m aL(±T) (2-14)

and the elastic deformations which are used, in combination with the
equations of static equilibrium, to solve the various problems that may
be encountered.
 .

Torsion

3-1 INTRODUCTION AND ASSUMPTIONS

In this chapter we shall consider the derivation and application of the

twisting or torsion problem only in connection with circular shafts or
closed thin-walled tubes. The twisting of noncircular shafts is so com-
plex that we will only state the formulas that are used.
Torsion is our introduction to the problems of variablestress.
Although the general theory of these problems is complex, its applica-
tion consists of little more than substituting given values in the formulas
soon to be derived, and is fairly simple.
The general technique used in all cases of nonuniform stress
distribution is outlined in these steps:

1. From a study of the elastic deformations produced by a

specified load, plus the application of Hooke's law, determine the
relations between stresses that are compatible with the deformations.
Such relations are known as the equations of compatibility

2. By applying the conditions of equilibrium to a free-body

diagram of a portion of the body, obtain additional relations between
the stresses. These relations, resulting from a study of the equilibrium
between externally applied loads and the internal resisting forces over
an exploratory section, are called the equations of equilibrium.
3. Be sure that the solution of the equations in steps 1 and 2 is

consistent with the loading conditions at the surface of the body. This is

known as satisfying the boundary conditions.

74
3-2 Derivation of Torsion Formulas 75

In the theory of elasticity, it is shown that a solution satisfying

these three steps is unique; that is, it is the only possible solution.
we make the following assump-
In deriving the torsion formulas,
These assumptions may be proved mathematically, and some may
tions.
be demonstrated experimentally. The first two apply only to shafts of
circular section.

1. Circular sections remain circular.

2. Plane sections remain plane and do not warp.

3. The projection upon a transverse section of straight radial
lines in the section remains straight.

4. Shaft is loaded by twisting couples in planes that are per-

pendicular to the axis of the shaft.
5. Stresses do not exceed the proportional limit.

3-2 DERIVATION OF TORSION FORMULAS

Figure 3-1 shows two views of a solid circular shaft. If a torque T is

applied at the ends of the shaft, a fiber AB on the outside surface, which
is originally straight, will be twisted into a helix AC as the shaft is

twisted through the angle 0. This helix is formed as follows:

Figure 3-1. Deformation of circular shaft.

76 3 Torsion

Imagine the shaft to consist of innumerable thin slices, each of

which is rigid and joined to adjacent slices by elastic fibers. Slice will ©
rotate past slice (T) until the elastic fibers joining them are deformed
enough to create a resisting torque which balances the applied torque.
When this happens, slices (T) and © will act as a rigid unit and
transmit the torque to slice ©; this slice will rotate enough so that the
elastic fibers joining it and slice © develop a resisting torque equal to
the applied torque. This type of deformation proceeds throughout the
length L of the shaft. The helix AC is the line joining the original
reference line AB on these slices as they become infinitely thin. This
description of the twisting action is idealized, but the resulting helix is

accurately described; actually, all such slices start rotating simulta-

neously relative to each other as soon as the torque is applied, the angle
of rotation becoming larger as the applied torque is increased.
Consider now any internal fiber located a radial distance p from
the axis of the shaft. From assumption 3 in Art. 3-1, the radius of such
a fiber also rotates through the angle 0, causing a total shearing
deformation 8S equal to DE. The length of deformation is the arc of
this
a circle whose radius p and which
is is subtended by the angle of 9
radians; the length is given by

8S = DE = P
V~^T x(iLj
(a)

The unit deformation of this fiber is

The shearing stress at this typical fiber is determined from Hooke's law
to be
G0\
r = Gy

Equation (c) may be called the equation of compatibility, since the

stresses expressed by it are compatible with the elastic deformations.
Note that each of the terms in the parentheses in this equation is a
constant which does not depend upon the particular internal fiber
chosen for analysis; the product of these terms represents a constant.
Therefore we conclude that the shearing stress at any internal fiber is
determined by the product of a constant and a variable radial distance;
i.e., the stress distribution along any radius varies linearly with the radial
distance from the axis of the shaft. Figure 3-1 illustrates the stress
variation along the radius OB; the maximum stress occurs at the outside
fiber and is denoted by max. r.
In line with the general procedure outlined in Art. 3-1, the shaft is

divided into two segments by a cutting plane M-N. Figure 3-2 shows
the free-body diagram of the left-hand portion.
 .

3-2 Derivation of Torsion Formulas 77

Section M-N

dP= TdA

Figure 3-2. Free-body diagram of Fig. 3-1

A differential area of section M-N at a radial distance p from the

axis of the shaft carries the differential resisting load dP = t dA. By
considering an area infinitesimally small, we may assume the stress to
be uniform over such an area. Since the function of this resisting load
dP is to produce resistance to the applied torque T, the load must be

directed perpendicular to the radius p in order to produce the maximum

effect. It is true, but difficult to prove here, that in circular sections dP is
directed perpendicular to p; nevertheless, we may take it as axiomatic
that loads always distribute themselves as efficiently as possible. It is

this variation of the principle of the conservation of energy that de-

termines the direction of dP as perpendicular to p so that it produces
maximum torsional resistance.
To satisfy the condition of static equilibrium, we apply SM = 0,
or the fact that the applied torque T equals the resisting torque Tr The
.

resisting torque Tr is the sum of the resisting torques developed by all

differential loads dP:

T= Tr =fpdP=fp(rdA)
Replacing t by its value from Eq. (c) gives

GO
dA

or, since
2
/p dA = /, the polar moment of intertia of the cross section,

This is usually written,*

•-§ CM)

*Note the similarity of Eq. (3-1) and the equation for linear deformation
8 = PL/ AE. This similarity will make the two equations easier to remember.
78 3 Torsion

In order for 9 to be in the proper units of radians, T must be in

N m •
and L in m; J of course is in m4 , and G is in N/m 2
. If we wish to
express in degrees, we multiply the right-hand member of Eq. (3-1) by
the unit fraction, 180 deg/7r rad = 57.3 deg/rad.
By replacing the product GO/ L in Eq. (c) by its equivalent value
T/J from Eq. (3-1), we obtain

r-^ (3-2)

This is called the torsion formula. The formula that determines the
maximum shearing stress is a more common form of the torsion
formula. It is obtained by replacing p by the radius r of the shaft:

Max. t « -= (3-2 a)
j
Note Hooke's law was used in deriving these equations,
that since
the stresses must not exceed the shearing proportional limit*; also, these
formulas are applicable only to circular shafts, either solid or hollow/
The values of polar moments of inertia for circular shafts are given
in Fig. 3-3. Using these values, we obtain the following modifications of
the torsion formula:

Solid shaft: Max. t = ^-3 = -^ (3-2b)

arr ltd

Hollow shaft: Max. t = 2

4
™ 4
- =
X*
4
TD -
(3-2c)
v(R - r ) tt(D - d 4)
In many practical applications, shafts are used to transmit power.
From dynamics, it is known that the power 9 transmitted by a constant
torque T rotating at a constant angular speed co is given by

9 - 7w
where to is measured in radians per unit time. If the shaft is rotating
with a frequency of / revolutions per unit time, co = 2 77/, and we have
<3> = T2<rrf
Thus the torque can be expressed as

•Equation (3-2a) is sometimes used to determine the shearing stress at

rupture. Although the proportional limit is exceeded, the fictitious shearing stress
so obtained is called the torsional modulus of rupture. It is used to compare the
ultimate strengths of specimens of various materials and diameters.
' A satisfactory formula for determining the maximum shearing stress in
rectangular shafts is

ab<2
1b V
^(3a)
+ 1.8*)
where a is the long side and b the short side of the rectangular section.
3-2 Derivation of Torsion Formulas 79

Hollow shaft

Solid shaft

A
_nr _7Td 4
~ 2 ~ 32

Figure 3-3. Polar moments of inertia.

With 9 measured in watts (1 W= 1 N-m/s) and /in revolutions per

second (r/s), the above equation will detennine the torque T in
newton-meters. This value of T may be used in Eq. (3-2) to obtain the
maximum shearing stress and in Eq. (3-1) to determine the angle of
twist.

ILLUSTRATIVE PROBLEMS

301. A solid shaft in a rolling mill transmits 20 kW at 2 r/s.

Determine the diameter of the shaft if the shearing stress is not to
exceed 40 MN/m
2
and the angle of twist is limited to 6° in a length of 3
m. Use G = 83 GN/m2 .

Solution: This problem illustrates a design that must possess sufficient

strength as well as rigidity. We start by applying Eq. (3-3) to determine
the torque:

3
X
r=
2tt/
r=
20
2tt(2)
10
= 1590 Nm
To satisfy the condition of strength, we apply the torsion formula,
Eq. (3-2b):

T = 167 40 X 10° =, 16(1590)

3
TTd md'

from which
-6
d 3 = 202 X 10 m = 3
202 X 10
3
mm 3
and d= 58.7 mm
We next apply the angle of twist relation, Eq. (3-1), to determine
the diameter necessary to satisfy the requirement of rigidity. In degrees,
 —
3 Torsion

this is

* = §X57.3 or y = §x57.3
whence

md* 1590(3)(57.3)
=
==
9
32 (6)(83 x 10 )

From this

-6
dA = 5.59 X 10 m4 = 5.59 x 10
6
mm 4 and d = 48.6 mm
The larger diameter, d = 58.7 mm, will satisfy both strength and
rigidity.

302. Two solid shafts of different materials are rigidly fastened

together and attached to rigid supports as shown in Fig. 3-4. The
aluminum segment is 75 mm in diameter, and Ga — 28 X 109 N/m 2 .

The steel segment has a diameter of 50 mm and Gs = 83 X 109 N/m 2 .

The torque, T= 1000 N-m, is applied at the junction of the two

segments. Compute the maximum shearing stress developed in the
assembly.

3 m »+*
*jr< 1.5 m +•

Aluminum
*c
\75mmdia. T=1.0kN-m
Ga = 28GN/m2 5 0mmdia'
Gs = 83 GN/m 2

Figure 3-4. Statically indeterminate composite shaft.

Solution: This problem is statically indeterminate in that we do not

know how the applied torque is apportioned to each segment. The
procedure we follow is exactly the same as that discussed in Art. 2-5 for
statically indeterminate axially loaded members. Applying the condi-
tions of static equilibrium and of geometric compatibility, we obtain
first

[ZA/ = o] 7; + 7; = r= iooo (a)

Another relation between Ts and Ta is obtained from the condition

that eachsegment has the same angular deformation, so that 9S = a .

Applying Eq. (3-1) gives

1 JG Js [ JG )a\
r,(1.5) Ta (3)

^(83X10 9
)
^r (
28xl09 )
3-2 Derivation of Torsion Formulas 81

from which
Ts = 1.177; (b)

Solving Eqs. (a) and (b), we obtain

Ta = 461 Nm and Ts - 539 N -m
Applying the torsion formula, we find the stresses to be

T = \6T = J6(4611 = 557xl()6

3 3
TTd tt(0.075)

= 5.57 MN/m 2

16(539)
T„ — = 22.0 X 10
6
N/m 2
3
tt(0.050)

= 22.0 MN/m 2

303. A steel shaft with a constant diameter of 50 mm is loaded as

shown in Fig.3-5 by torques applied to gears fastened to it. Using
G= 83 X
10
3
MN/m2
compute in degrees the relative angle of rota-
,

tion between gears A and D.

700 N-m

1200 N-m

1300 N-m

700 N-m

D C B A
TCD = S00 TBC --b00 TAB =700
Figure 3-5. Angular deformations.

Solution: The use of double-headed vectors to denote torques, as

shown in the lower portion of Fig. 3-5, simplifies determining the
torque transmitted by each segment of the shaft. All we need do is pass
a section through each segment and apply the conditions of equilibrium
to a free body between that section and one end, say, D. Then, relative
to D, the torques transmitted by each segment are TAB = 700 N-m
82 3 Torsion

clockwise, TBC = 500 N m•

counterclockwise, and TCD = 800 N m
•

clockwise.
The total angular deformation is the algebraic sum of the defor-
mation each segment. Arbitrarily denoting clockwise deformation as
in
positive, we apply Eq. (3-1), remembering to convert radians to degrees.
Doing this, and recognizing the common value of JG, gives

A/D
f
Zu r^
JG

A/D =^
JG
27XX57.3
57.3
[700(3) - 500(1.5) + 800(2)]
4
7r(0.050) 9
(83 X 10 )
32
which gives
f
A/D = 3.32* Ans.

The positive result indicates that the net rotation of A relative to D is

clockwise.

PROBLEMS

304. What is the minimum diameter of a solid steel shaft that

will not twist through more than 3° in a 6-m length when subjected to a
torque of 14 kN m? What maximum
•
shearing stress is developed? Use
G - 83 GN/m 2 . Ans. d= 1 18 mm; r = 43.4 MN/m 2

305. A solid steel shaft 5 m long is stressed to 60 MPa when

twisted through 4°. Using G— 83 GPa, compute the shaft diameter.
What power can be transmitted by the shaft at 20 r/s?
Ans. d = 104 mm; 9 = 1.67 MW
306. Determine the length of the shortest 2-mm-diameter bronze
wire which can be twisted through two complete turns without exceed-
ing a shearing stress of 70 MPa. Use G = 35 GPa.
307. A steel marine propeller is to transmit 4.5 MW at 3 r/s
without exceeding a shearing stress of 50 MN/m 2
or twisting through
more than 1
° in a length of 25 diameters. Compute the proper diameter
if C = 83 GN/m 2
.

308. Show that a hollow circular shaft whose inner diameter is

half the outer diameter has a torsional strength equal to of that of a

-|f

solid shaft of the same outside diameter.

3-2 Derivation of Torsion Formulas 83

309. A steel shaft with a constant diameter of 60 mm is loaded

by torques applied to gears attached to it as shown in Fig. P-309. Using

G= 83 GN/m 2 , determine the relative angle of twist of gear D relative
to gear ,4. Arts. BD/A = 2.17°

800 N-m

1000 N-m

1200 N-m

1000 N-m

Figure P-309.

310. Determine the maximum torque that can be applied to a

hollow circular steel shaft of 100-mm outside diameter and 70-mm
inside diameter without exceeding a shearing stress of 60 X 10 N/m
6 2

or a twist of 0.5 deg/m. Use G= 83 X 10

9
N/m2 .

311. A stepped steel shaft consists of a hollow shaft 2 m long,

with an outside diameter of 100 mm and an inside diameter of 70 mm,
rigidly attached to a solid shaft 1.5 m long, and 70 mm in diameter.
Determine the maximum torque that can be applied without exceeding
a shearing stress of 70 MN/m 2 or a twist of 2.5 deg in the 3.5-m length.
Use G - 83 GN/m 2
. Arts. T = 4.01 kN m •

312. A flexible shaft consists of a 5-mm-diameter steel wire

encased in a stationary tube that fits closely enough to impose a
frictional torque of 2 N
m/m. Determine the maximum length of the
•

shaft if the shearing stress is not to exceed 140 MPa. What will be the
angular rotation of one end relative to the other end? Use G = 83 GPa.
Ans. L= 1.72 m; = 33.3°
84 3 Torsion

313. The steel shaft shown in Fig. P-313 rotates at 3 r/s with 30
kW taken off at /*, 15 kW removed at B and 45 y
kW applied at C. Using
G - 83 X 10
9
N/m 2
, find the maximum shearing stress and the angle of
rotation of gear A relative to gear C.
Ans. Max. r = 64.9 MN/m 2
; 9 = 8.23°

Figure P-313.

314. A solid steel shaft is loaded as shown in Fig. P-314. Using

G = 83 GN/m 2
, determine the required diameter of the shaft if the
shearing stress is limited to 60
2
MN/m
and the angle of rotation at the
free end is not to exceed 4 deg.

500 N-m 1000 N-m

tir -ft
r 2 m-

Figure P-314.
3 m J

315. 5-m steel shaft rotating at 2 r/s has 70 kW applied at a

A
gear that m from the left end where 20 kW are removed. At the
is 2
right end, 30 kW are removed and another 20 kW leaves the shaft at 1.5
m from the right end. (a) Find the uniform shaft diameter so that the
shearing stress will not exceed 60 MN/m 2
. (b) If a uniform shaft
diameter of 100 mm is specified, determine the angle by which one end
of the shaft lags behind the other end. Use G= 83 GN/m2 .

Ans. d = 69.6 mm; 9 = 0.448°

316. round steel shaft 3 m long tapers uniformly from a

A
60-mm diameter at one end to a 30-mm diameter at the other end.
Assuming that no significant discontinuity results from applying Eq.
(3-1) over each infinitesimal length, compute the angular twist for the
entire length when the shaft is transmitting a torque of 170 N-m. Use
G = 83 X 10
3
MN/m 2
. Ans. 9 = 1.29°

317. A hollow bronze shaft of 75 mm outer diameter and 50 mm

inner diameter is slipped over a solid steel shaft 50 mm in diameter and
of the same length as the hollow shaft. The two shafts are then fastened
rigidly together at their ends. Determine the maximum shearing stress
3-2 Derivation of Torsion Formulas 85

developed in each material by end torques of 3 kN • m. For bronze,

G - 35 GN/m
2
for steel, G = 83
;
2
GN/m .

Ans. rb = 28.5 MN/m2 ; rs - 45.1 MN/m2

QUA. A solid compound shaft is made of three different materials
and is subjected to two applied torques as shown in Fig. P-318. (a)
Determine the maximum shearing stress developed in each material, (b)
Find the angle of rotation of the free end of the shaft. Use Ga = 28
GN/m2 Gs = 83 GN/m 2 and Gb - 35 GN/m2
, , .

100 mm dia. 75 mm dia.

4.0kN-m 1.5kN-m
/
Aluminum

3 m —>U— 2 m—»4-
Bronze
>\
1.5 m
Figure P-318.

(_3J5* The compound

shaft shown in Fig. P-319 is attached to
rigid supports. For the bronze segment AB, the diameter is 75 mm,
t < 60 MN/m and G = 35 GN/m 2 For the steel segment BC, the
2
, .

diameter is 50 mm, t < 80 MN/m and G = 83 GN/m If a = 2 m

2 2
, .

and 6 = 1.5 m, compute the maximum torque T that can be applied.

C;

S
Figures P-319 and P-320.

320. In Problem 319, determine the ratio of lengths b/a so that

each material will be stressed to its permissible limit. What T is
torque
required? Ans. b/a = 1.19; T = 6.93 kN-m
321. A compound shaft consisting of an aluminum segment and
a steel segment is acted upon by two torques as shown in Fig. P-321.
Determine the maximum permissible value of T subject to the following
conditions: ts < 100 MPa, ra< 70 MPa, and the angle of rotation of
the free end limited to 12°. Use G = 83 GPa and G„ = 28 GPa.
r

75 mm dia. 50 mm dia
T I 2T
/
'Aluminum

1.5 m-

Flgure P-321.
 J

3 Torsion

322. A torque T is applied, as shown in Fig. P-322, to a solid

shaft with built-in ends. Prove that the resisting torques at the walls are
T x
Tb/ L and T2 = Ta/ L. How would these values be changed if the
shaft were hollow?
T

-.c t
Figure P-322.

323. A shaft 100 mm in diameter and 3 m long, with built-in

ends, is subjected to a clockwise torque of 4 kN •m applied m from 1

the left end, and to another clockwise torque of 16 kN m applied 2 m

from the left end. Compute the maximum shearing stress developed in
each segment of the shaft. {Hint: Use the results of Problem 322 and
apply superposition.)
324. A shaftcomposed of segments AC, CD, and DB is fastened
to rigid supports and loaded as shown in Fig. P-324. For steel, G = 83
GN/m 2 aluminum, G = 28 GN/m and for bronze, G = 35
2
; for ;

GN/m 2 . Determine the maximum shearing stress developed in each

segment. Ans. TB = 472 N-m; t^ - 9.3 MN/m 2

T r = 300 N-m 700 N-m

B
/?C D /£
G 25
Steel

mm dia. Vi/
Aluminum
^ 50 mm dia.
I

VK
Bronze
25 mm dia. C
2m — -lm *

Figure P-324.

325.
125. The two steel shafts shown in Fig. P-325, each with one end
built into a rigid support, have flanges rigidly attached to their free
ends. The shafts are to be bolted together at their flanges. However,
6°

"-
— 40 mm dia

/
/
s /
1 /
, /
/
\ ~sc
>

50 mm dia. /

— '

2 m — 1 m—
Figure P-325.
3-3 Flanged Bolt Couplings 87

initially there is a 6° mismatch in the location of the bolt holes, as

shown in the figure. Determine the maximum shearing stress in each
shaft after the shafts are bolted together. Use G= 83 GN/m2 and
neglect deformations of the bolts and flanges.

3-3 FLANGED BOLT COUPLINGS

A commonly used connection between two shafts is a flanged bolt

coupling. It consists of flanges rigidly attached to the ends of the shafts
and bolted together, as in Fig. 3-6. The torque is transmitted by the
shearing force P created in the bolts.

P=At = -4~t
6=

T T

Figure 3-6. Flanged bolt coupling.

Assuming that the stress is uniformly distributed, the load in any

given by the simple stress equation P = Ar and equals (ird /4)r.
2
bolt is

It acts through the center of the bolt and tangent to the bolt circle. The

torque resistance of one bolt is PR, where R is the radius of the bolt
circle. Therefore, for any number of bolts, /i, the torque capacity of the

coupling is expressed by

2
ird
T= PRn = -?-TRn (3-4)

Occasionally a coupling has two concentric rows of bolts, as in

Fig. 3-7. Letting the subscript 1 refer to bolts on the outer circle and
subscript 2 refer to bolts on the inner circle, the torque capacity of the
coupling is

T = P R n + P2R 2n 2
l l l (3-5)

The relation between P and P2

x
can be determined from the fact
that the comparatively rigid flanges cause shear deformations in the
bolts which are proportional to their radial distances from the shaft axis.
 A

3 Torsion

Figure 3-7. Coupling with two concentric bolt circles.

Thus, the shearing strains are related by

{a)

Using Hooke's law for shear, G= r/y, we have

PJ t
P2 /A 2
(b)
or
G,R t
G2 R 2 G,/?, G2 R 2
If the bolts on the two circles have the same area, A = A 2 and if the x
,

bolts are made of the same material, G = G 2 the relation between P

x
,
x

and P 2 reduces to

P P
(3 " 6)
£ £ =
This is the case shown in Fig. 3-7. Using the relation between P x
and
P2 , Eq. (3-5) will determine the torque capacity of the coupling.
A similar procedure may be used for three or more concentric bolt
circles. As we shall see in Chapter 12, this situation occurs in eccentri-
cally loaded riveted connections.

PROBLEMS

326. A flanged bolt coupling consists of eight steel 20-mm-diam-

eter boltsspaced evenly around a bolt circle 300 in diameter. mm
Determine the torque capacity of the coupling if the allowable shearing
stress in the bolts is 40 MN/m' Ans. T= 15.1 kN-m
327. A flanged bolt coupling is used to connect a solid shaft 90
mm in diameter to a hollow shaft 100 mm
in outside diameter and 90
mm in inside diameter. If the allowable shearing stress in the shafts and
the bolts is 60 MN/m 2
, how many 10-mm-diameter steel bolts must be
used on a 200-mm-diameter bolt circle so that the coupling will be as
strong as the weaker shaft?
3-3 Flanged Bolt Couplings 89

328. A flanged bolt coupling consists of six 10-mm-diameter steel

bolts on a bolt circle 300 mm in diameter, and four 10-mm-diameter
steel bolts on a concentric bolt circle 200 mm in diameter, as shown in
Fig. 3-7. What torque can be applied without exceeding a shearing
stress of 60 MPa in the bolts? Arts. T= 5.50 kN • m
329. Determine the number of 10-mm-diameter steel bolts that
must be used on the 300-mm bolt circle of the coupling described in
Problem 328 to increase the torque capacity to 8 kN m. Ans. 10 bolts •

330. Solve Problem 328 if the diameter of the bolts used on the
200-mm bolt circle is changed to 20 mm.
331. In a rivet group subjected to a twisting couple T, show that
the torsion formula r = Tp/J
can be used to find the shearing stress t
the center of any rivet. Let J = *2Ap where A is the area of a rivet
2
at ,

at the radial distance p from the centroid of the rivet group.

332. A plate is fastened to a fixed member by four 20-mm

diameter rivets arranged as shown in Fig. P-332. Compute the maxi-

mum and minimum shearing stress developed. (Hint: Use the results of
Problem 331.)

16 kN

100 a 100 a 100

t mm t

16 kN
Figure P-332.

Six 20-mm-diameter rivets fasten the plate in Fig. P-333 to

333.
the fixed member. Using the results of Problem 331, determine the
average shearing stress caused in each rivet by the 40-kN loads. What

40 kN
I

7_5
mmn
TX mm
§1
itiiriix:

40 kN
Figure P-333.
90 3 Torsion

additional loads P can be applied before the average shearing stress in

any rivet exceeds 60 MN/m 2
?

Ans. Max. t = 45.9 MN/m P = 2

; 55.4 kN
334. The plate shown in Fig. P-334 is fastened to the fixed
member by three 10-mm-diameter rivets. Compute the value of the
loads P so that the average shearing stress in any rivet does not exceed
70 MPa. (Hint: Use the results of Problem 331.)
Ans. P = 7.12kN

Figure P-334.

335. A flanged
bolt coupling consists of six 10-mm-diameter steel
bolts evenly spacedaround a bolt circle 300 mm
in diameter, and four
20-mm-diameter aluminum bolts on a concentric bolt circle 200 in mm
diameter. What torque can be applied without exceeding a shearing
stress of 60 MN/m in the steel or 40 MN/m in the aluminum? Use
2 2

Gs = 83 GN/m and Ga = 28 GN/m' Ans. T = 5.94 kN m

2

3-4 LONGITUDINAL SHEARING STRESS

So far in our discussion of torsional stress, we have considered only the

shearing stress on transverse sections. However, a longitudinal shearing
stress is also induced which is perpendicular and numerically equal to
the transverse torsional shearing stress. As we shall see again in Art.
5-7, this fact illustrates the general principle that a shearing stress
acting on one face of an element always accompanied by a numeri-
is

cally equal shearing stress acting on a perpendicular face.

To demonstrate the existence of a longitudinal shearing stress,
consider the element isolated by two transverse planes, two longitudinal
planes through the axis,and two surfaces at different radii as shown in
Taking moments about the axis gh of the enlarged free-body
Fig. 3 -8a.
diagram of this element shown in Fig. 3-8b, we see that equilibrium is
possible only if a longitudinal shearing stress r' acts in addition to the
torsional stress r. Multiplying these stresses by the areas of the faces
3-5 Torsion of Thin-Walled Tubes; Shear Flow 91

Figure 3-8. Equivalence of longitudinal and torsional shearing stress.

over which they act, we obtain

[2Mgh = 0] (t dr r dO) dx - (r' dr dx)r dO =

Canceling out the common product r dO dr dx, we see that

r'- T

A pictorial view illustrating this equivalence of longitudinal and

torsional shearing stresses is shown in Fig. 3 -8c, where a portion of the
shaft has been removed to clarify the concept.

3-5 TORSION OF THIN-WALLED TUBES; SHEAR FLOW

Although the torsion of noncircular shafts requires advanced methods, a

simple approximate solution is possible for the special case of thin-
walled tubes. In Fig. 3 -9a we consider a tube of arbitrary shape with a
variable wall thickness t which is relatively small compared with the
dimensions of the cross section. Figure 3-9b shows, enlarged, a free
body of a typical element of length AL cut from this tube. The torsional
stress r x across the thickness t
x
induces a numerically equal longitudinal
stress, as was proved in the preceding article. Similarly, across the
thickness t
2 , a different torsional shearing stress t 2 is accompanied by a
numerically equal longitudinal stress.
92 3 Torsion

F2 = q2 AL

Fl =q AL'
1

(a) (b)

Figure 3-9. Shear flow in a thin-walled tube.

The resultants of these longitudinal shearing stresses are

F, = q x
AL and F2 = q2 AL (a)

where the symbol q represents /'/?/ 2 t <#. The term # is called the j/jetfr
ytow and is a convenient concept in cases where the precise distribution
of shearing stress across a thickness is unknown or unimportant. Con-
sidering longitudinal equilibrium of the element, we have

q l
AL = q2 AL or q l
— q2 (b)

The equivalence of the shear flow at two arbitrary positions means that
the shear flow must be constant around the cross section of the tube. In
fact, the name shear flow is based on the mathematical analogy between

shear flow and the obviously constant flow of an incompressible fluid

around a closed channel whose boundaries are the inner and outer walls
of the tube.
To relate shear flow to the applied torque T, consider Fig. 3-10.
Over any infinitesimal length dL, the tangential force is q dL, and its
contribution toward resisting the torque is measured by its moment
r(q dL) about any convenient center O. Since the moment of the
torsional couple T is independent of a moment center, on equating T to
the summation of such contributions, we have

-/, dL (c)

Instead of carrying out this integration, we note that r dL is twice the

area of the shaded triangle whose base is dL and whose altitude is r.

Consequently, since q is constant, the value of the integral is q times

twice the area A enclosed by the center line of the tube wall, or
3-5 Torsion of Thin-Walled Tubes; Shear Flow 93

q dL\. \/

Figure 3-10.

T - 2Aq (3-7)
Finally, the average shearing stress across any thickness / is given
by

(3-8)
t 2At

ILLUSTRATIVE PROBLEM

336. A tube has the semicircular shape shown in Fig. 3-11. If

stress concentration at the corners is neglected, what torque will cause a

shearing stress of 40 MN/m 2

?

t = 2 mm

Figure 3-11.

Solution: In applying Eq. (3-8), observe that A is the area enclosed by

the centerline of the tube. Thus we obtain

[T-lAtr] T^ll^y
= 2[|(0.025) 2 (0.002)(40 X 10
6
)

= 157N-m Ans.
94 3 Torsion

PROBLEMS

337. A torque of 600 N m is applied to the rectangular section

•

shown in Fig. P-337. Determine the wall thickness / so as not to exceed

a shear stress of 60 MPa. What is the shear stress in the short sides?
Neglect stress concentration at the corners.

30 mm

Figure P-337.

338. A tube 3 mm thick has the elliptical shape shown in Fig.

P-338. What torque will cause a shearing stress of 60 MN/m2 ?
Ans. r=3.18kNm

75 mm

Figure' P-338.

339. A tube 3 mm thick has the shape shown in Fig. P-339. Find
the shearing stress caused by a torque of 700 N m •dimension a = 75
if

mm.

Figures P-339 and P-340.

340. Find dimension a in Problem 339 if a torque of 600 N m

•

causes a shearing stress of 70 MN/m 2

. Ans. a = 55.7 mm
3-6 Helical Springs 95

341. Derive the torsion formula t = Tp/J for a solid circular

section by assuming the section is composed of a series of concentric
thin circular tubes. Assume that the shearing stress at any point is
proportional to its radial distance.

3-6 HELICAL SPRINGS

The close-coiled helical spring in Fig. 3-12 is elongated by an axial load

P. The spring is composed of a wire or round rod of diameter d wound
into a helix of mean radius R. The helix angle is small, so that any coil
of the spring may be considered as lying approximately in a plane
perpendicular to the axis of the spring.

Figure 3-12. Helical spring.

To determine the stresses produced by P, we follow the general

procedure of passing an exploratory cutting plane m-n through any
typical section as shown and then determining the resisting forces
required for equilibrium. We then analyze the stress distribution that
creates these resisting forces.
Figure 3- 13a shows the free-body diagram of the upper half of the
spring. To balance the applied axial load P, the exposed shaded cross
section of the spring Pr equal to P. The free
must provide the resistance
body is now and horizontal summa-
in equilibrium as far as a vertical
tion of forces is concerned. To complete equilibrium, however, a mo-
ment summation must also equal zero. It is evident that P and Pr being ,

equal, opposite, and parallel, create a couple of magnitude PR which

96 3 Torsion

2jr=Pi?

(b) Magnified view of

pr =p
spring cross-section m-n
(a) Free-body diagram

Figure 3-13. Analysis of helical spring

can be balanced only by an opposite couple. This resisting couple is

created by a torsional shearing stress distributed over the cross section
of the spring; it is represented by T = PR.
The magnified view of the cross section in Fig. 3- 13b shows the
stress distribution that created the resisting forces. Two types of shear-
ing stress are produced: (1) direct shearing stresses like t„ uniformly
distributed over the spring section and creating the resisting load Pr
that
passes through the centroid of the section; and (2) variable torsional
shearing stresses like r 2 caused by the twisting couple T — PR. The
torsional stresses t 2 vary in magnitude with their radial distance from
the centroid and are directed perpendicular to the radius, as at A The
.

resultant shearing stress is the vector sum of the direct and torsional
shearing stresses. At B, the stresses are oppositely directed, and the
resultant stress is the difference between t 2 and r,. At the inside fiber C,
however, the two stresses are collinear and in the same sense; their sum
produces the maximum stress in the section. The maximum stress
always occurs at the inside element of the spring wire. Is there any

position on the diameter BC at which the shearing stress is zero? If so,

how can you locate it?
To summarize the foregoing discussion, the maximum shearing
stress occurs at the inside element and is given by the sum of the direct
shearing stress t, = P/A and the maximum value of the torsional
shearing stress t 2 — Tr/J, or
4P \6(PR)
T = T, -I- T2 =
md' md'
This may be written

= 16PR
t
ltd'
(£) (3-9)
 :

3-6 Helical Springs 97

(a) (b)

Figure 3-14. Torsion of straight and of curved segments.

Examination of Eq. (3-9) shows that the ratio d/4R is small for a
spring composed of a wire of relatively small diameter wound on a
spring with a large radius; this indicates that in such cases the maximum
stress is caused primarily by torsion of the spring wire. On the other
hand, heavy coil springs, such as those used on railroad cars, are made
of wire with a relatively large diameter d in comparison with R, the
mean radius of the spring; in these springs the effect of direct shearing
stress is 14% or more of the total stress and cannot be disregarded.
It should be noted that the above discussion contains an error
because the torsion formula derived for use with straight bars was
applied to a curved bar. This error is of significance in heavy springs
ind is explained in Fig. 3-14. In the straight bar in Fig. 3- 14a, torsion
produces the same shearing deformation 8S on fibers AB and CD. The
shearing strain, y = 5,/L, is the same at B and Z>, since the elements
AB and CD have the same original length. A different situation,
however, exists in the curved bar in Fig. 3- 14b. Although fibers AB and
CD undergo the same shearing deformation, the shearing strain at B on
the inside element is greater than at D on the outside element because
of the shorter initial length of AB. Therefore, since stress is proportional
to strain, the shearing stress on the inner fibers of a curved bar is greater
than on the outer fibers. This fact is not taken into account in Eq. (3-9).
Of course, the importance of this error depends upon how greatly
elements AB and CD differ in original length. Evidently this difference
depends on how sharply curved the spring wire is, i.e., upon the ratio of
d to R. A.M. Wahl has developed the following formula that takes
account of the initial curvature of the spring wire*

Max. = \6PR ( 4m 1 + 0.615

t (3-10)
nd 3 V 4m — 4 m
where m — 2R/ d = D/d, the ratio of the mean diameter of the spring
to the diameter of the spring wire. In light springs, where the ratio m is
large, the first term in the parentheses approaches unity. Compare with

*See A. M. Wahl, Stresses in heavy closely coiled helical springs. Trans.

A.S.M.E. 51, paper No. APM-51-17.
98 3 Torsion

Eq. (3-9), which may be rewritten in the following form:

= \6PR
Max. t
?(«¥)
md
(3-9a)

For heavy springs which are sharply curved and in which m is not so
(3-10) emphasizes and corrects the error in Eq. (3-9).
large, Eq.
Factors 0.5 and 0.615 differ in Eqs. (3-9a) and (3-10) largely
because the direct shearing stress is not actually distributed uniformly
over the cross section. We shall see later (Art. 5-7), when discussing
horizontal shearing stress in beams, that for a circular cross section the
maximum shearing stress produced is approximately \ times the average
shearing stress and varies from 1.23 at the outside edges to 1.38 at the
center. The factor 0.615 in Eq. (3-10) results from multiplying 0.5 by
1.23.

Note that springs are made of special steels and bronzes in which
the allowable shearing stresses range from 200 to 800 MPa.

Spring deflection

Practically all the spring elongation, measured along its axis, is

caused by torsional deformation of the spring wire. If we temporarily
assume all the spring in Fig. 3-15 to be rigid except the small length dL,
the end A will rotate to D through the small angle dO. Because dO is

Figure 3-15. Deflection of helical spring.

3-6 Helical Springs 99

small, the arc AD = AB

dO may be considered as a straight line
•

perpendicular to AB, whence, from the similarity of triangles ADE and

BAC, we obtain

AE = BC
AD AB
or

</5 R
AB •
40 ^5
whence
d8 = RdO (a)

Applying Eq. (3-1), we may replace dO by its equivalent value in

terms of the torque and length:

which
«=
is
*^
integrated to give the total elongation contributed by
w
all

elements of the spring:

.-&
Replacing L by 2mRn, which is the length of n coils of radius R, and /
4
by 7rd /32, we obtain

6 = ^^ Gd 4
(3-11)

This expression for spring deflection neglects the deformations

caused by direct shear:

„ PL P(2irRn) SPRn (3-12)

A ,G *r</
2
_ Gd
4
This latter deformation, however, is generally negligible compared to

the value of 8 given by Eq. (3-11) and consequently is usually ignored.

Equation (3-11) is also used to compute the deflection in compression
springs provided the coils are not spaced so closely that they touch
when the load is applied.

ILLUSTRATIVE PROBLEM

342.A load P is supported by two steel springs arranged in

series asshown in Fig. 3-16. The upper spring has 20 turns of 20-mm-
diameter wire on a mean diameter of 150 mm. The lower spring consists
of 15 turns of 10-mm-diameter wire on a mean diameter of 130 mm.
100 3 Torsion

Determine the maximum shearing stress in each spring if the total

deflection is 80 mm and G = 83 GN/m2 .

Solution: The total deflection is the sum of the deflection in each

spring. By applying Eq. (3-11), we find the load P to be

64PR 3n
* = 2 Gd 4
3 3
64P (0.075) (20) (0.065) (15)
0.080 =
9 4 4
83 x 10 (0.020) (0.010)
P = 233 N
Knowing P, we can now find the stresses. For the upper spring,
m = 2R/d = 2(0.075)/0.020 = 7.5; Am - 30.0. Applying Wahl's for-
mula, Eq. (3-10), we obtain
- 1
Max. t = \6PR ( 4m 0.615 \

m
|

7rd
3
I 4w - 4 )

16(223)(0.075) 30 - 1 0.615
Max. t = / \

tt(0.020)
3
V 30 - 4 7.5

= 12.7 MN/m 2
Am.
Similarly for the lower spring where m = 2(0.065)/ 0.010 = 13
and 4m = 52, we find

,16(223)(0.065)^52_-l 0.615 \
Max>T
tt(0.010)
j
V 52 - 4 13 )

= 81.9 MN/m 2
Ans.
If we had used Eq. (3-9) to compute these maximum shearing
stresses, the results would have been 11.4 MN/m 2 in the upper spring
and 76.7 MN/m 2
in the lower spring. Thus, the approximate formula
gives results that are 10.2% and 6.35% lower than the more precise Wahl
formula.
3-6 Helical Springs 101

PROBLEMS

343. Determine the maximum shearing stress and elongation in a

helical steel spring composed of 20 turns of 20-mm-diameter wire on a
mean radius of 80 mm when the spring is supporting a load of 2 kN.
Use Eq. (3-10) and G= 83 GN/m2 .

Ans. Max. r = 121 MN/m 2

; 8 = 98.7 mm
344. What is the maximum elongation of the spring in Problem
343 if the spring is made of phosphor bronze for which G= 42 GN/m2
and is stressed to 140 MN/m ? 2
Use Eq. (3-10).

345. A helical spring is made by wrapping steel wire 20 mm in

diameter around a forming cylinder 150 mm in diameter. Compute the
number of turns required to permit an elongation of 100 mm without
exceeding a shearing stress of 140 MPa. Use Eq. (3-9) and G= 83 GPa.
Ans. n = 17.9 turns

346. Compute the maximum shearing stress developed in a phos-

phor bronze spring having a mean diameter of 200 and consisting mm
of 24 turns of 20-mm-diameter wire when the spring is stretched 100
mm. Use Eq. (3-10) and G = 42 GN/m2 .

347. A clutch is activated by six helical springs symmetrically

spaced. Each spring consists of 12 turns of steel wire 10 mm in diameter
and has a mean diameter of 50 mm. Determine the load exerted against
the clutch plate by a contraction of 40 mm in the springs. What is the
maximum shearing stress in the springs? Use Eq. (3-9) and G = 83
GN/m2 .

348. Two steel springs arranged in series as shown in Fig. P-348

support a load P. The upper spring has 12 turns of 25-mm-diameter
wire on a mean radius of 100 mm. The lower spring consists of 10 turns
of 20-mm-diameter wire on a mean radius of 75 mm. If the maximum
spring must not exceed 200 MN/m compute
2
shearing stress in either ,

P
Figure P-348.
1 02 3 Torsion

the maximum value of P and the total elongation of the assembly. Use
Eq. (3-10) and G - 83 GN/m 2
. Compute the equivalent spring con-
stant by dividing the load by the total elongation.

349. A load P is supported by two concentric steel springs

arranged as shown P-349. The inner spring consists of 30 turns
in Fig.

of 20-mm-diameter wire on a mean diameter of 150 mm; the outer

spring has 20 turns of 30-mm wire on a mean diameter of 200 mm.
Compute the maximum load that will not exceed a shearing stress of
140 M Pa in either spring. Use Eq. (3-9) and G = 83 GPa.
Ans. P= 9.05 kN

Figures P-349 and P-350.

350. If the inner spring in Problem 349 is made of phosphor

bronze, with G= 42 GN/m 2
compute the maximum shearing stress in
,

each spring resulting from a load P — 5 kN. Use Eq. (3-10).

351. A rigid plate of negligible mass rests on the central spring in
Fig. P-351 which is 20 mm higher than the symmetrically located outer
springs. Each of the outer springs consists of 18 turns of 10-mm wire on
a mean diameter of 100 mm. The central spring has 24 turns of 20-mm
wire on a mean diameter of 150 mm. If a load P = 5 kN is now applied
to the plate, determine the maximum shearing stress in each spring. Use
Eq. (3-9) and G - 83 GN/m2 .

Ans. Central spring: Max. t = 170 MN/m 2

3-6 Helical Springs 103

352. Solve Problem 351 if the outer springs are made of phos-
phor bronze for which G = 42 GN/m2 Can you predict qualitatively.

the effect of this change upon the stresses?

353. A rigid bar, hinged at one end,

is supported by two identical

springs as shown P-353. Each spring consists of 20 turns of

in Fig.
10-mm wire having a mean diameter of 150 mm. Compute the maxi-
mum shearing stress in the springs, using Eq. (3-9). Neglect the mass of
the rigid bar. Ans . Max. t = 46.5 MN/m2

/ II II

S3
|-*-2 m »|« 2 m »|« 2 m— *-

10 kg

Figures P-353 and P-354.

354. If each spring in Problem 353 consists of 16 turns of 10-mm

wire on a mean diameter of 160 mm, determine the largest mass that
may be supported at the free end without exceeding a shearing stress of
140 MN/m 2
in either spring. Use Eq. (3-9).

355. As shown in Fig. P-355, a homogeneous 50-kg rigid block is

suspended by three springs whose lower ends were originally at the
same level. Each steel spring has 24 turns of 10-mm-diameter wire on a
mean diameter of 100 mm, and G = 83 GN/m 2 The bronze spring has .

48 turns of 20-mm-diameter wire on a mean diameter of 150 mm, and

G = 42 GN/m2 Compute the maximum shearing stress in each spring
.

using Eq. (3-9). Ans. For bronze, Max. t - 9.93 MN/m 2

Phosphor
Bronze
Steel

Figure P-355.
1 04 3 Torsion

SUMMARY

The discussion of torsion in this chapter is limited to circular

sections, solid or hollow. The shearing stress varies directly with the
radial distance from the center of the cross section and is expressed by

r-S (3-2)

The maximum shearing stress in solid shafts of diameter d becomes

r-^J
*nd
(3-2b)

In hollow shafts of external diameter D and internal diameter d, it is

1672)
t = (3-2c)
v(D 4 -d 4 )
The angular deformation in a length L is expressed in radians by

TL

which is converted to degrees by multiplying by 180/77- = 57.3. Equa-

tion (3-1) is useful not only in determining angular deformations, but
also as a basis for solving statically indeterminate problems in torsion.
The relation between the torque T and power 9 transmitted by a
shaft rotating at a frequency / is

T (3 - 3)
=h
The study of flanged bolt couplings (Art. 3-3) is in effect the
torsion formula reduced to a finite number of elements subjected to
shear.
The existence of longitudinal shear stress (Art. 3-4) induced by
torsional shear stress served to prove that the shear flow q is constant
along the cross section of any thin-walled tube (Art. 3-5). In terms of
the area A enclosed by the center line of the tube wall, its value is

.-£ (3-7)

from which the average shearing stress across any thickness t is

T
= 7 = 3 " 8)
2^ <

In close-coiled helical springs (Art. 3-6), the maximum shearing

stress is expressed fairly accurately by
3-6 Helical Springs 105

and more exactly by

vd 3 V 4m - 4 m I

where m = 2R/d.
The elongation of the spring generally neglects the effect of direct
shearing deformation and is given by

9 = ^^
Gd*
(3-11)
 Shear and
Moment
in Beams

4-1 INTRODUCTION

The basic problem in strength of materials is to determine the relations

between the stresses and deformations caused by loads applied to any
structure. In axial or torsional loadings, we had little trouble in applying
the stress and deformation relations because in the majority of cases the
loading either remains constant over the entire structure or is distributed
in definite amounts to the component parts.
The study of bending loads, however, is complicated by the fact
that the loading effects vary from section to section of the beam. These
loading effects take the form of a shearing force and a bending moment,
sometimes referred to as shear and moment. These terms will be defined
in the next article. be shown in Chapter 5 that two kinds of stress
It will

act over the transverse section of a beam: (1) a bending stress, which
varies directly with the bending moment, and (2) a shearing stress,
which varies directly with the shear. As a preliminary to the study of

106
4-1 Introduction 107

w N/m
1
//// rf.

l*i t
(a) Simple beam (b) Cantilever beam

Pi w N/m

I 1

(c) Overhanging beam

Figure 4-1. Statically determinate beams.

stresses inbeams, therefore, this chapter is concerned with the variation

in shear and bending moment in beams subjected to various combina-
tions of loadings under different conditions of support, particularly the
determination of the maximum values of shear and moment. Beam
deflections will be discussed in Chapter 6.
Methods of supporting some types of beams are shown in Fig.
4-1. A simple beam is supported by a hinged reaction at one end and a
roller support at the other, but is not otherwise restrained. A cantilever
beam supported at one end only, with a suitable restraint to prevent
is

rotation of that end. An overhanging beam is supported by a hinge and

a roller reaction, with either or both ends extending beyond the sup-
ports.These beams are all statically determinate; their reactions can be
determined directly from the equations of static equilibrium.
Other methods of supporting beams are shown in Fig. 4-2. The
propped beam, the fixed-ended or restrained beam, and the continuous
beam all have at least one more reactive element than is absolutely
necessary to support them. Such beams are statically indeterminate; the
presence of excess supports requires the use of additional equations
obtained from considering the elastic deformations of the beam. Their
solution is discussed in Chapters 7 and 8.

A concentrated load is one that acts over so small a distance that it

can be assumed to act 4- la. In contrast, a

at a point, as in Fig.
distributed load acts over a considerable length of the beam. It may be
distributed uniformly over the entire length, as in Fig. 4- lb, or over
part of the length as in Fig. 4-lc. Distributed loads may also be
uniformly varying or nonuniform. In a uniformly varying or triangular
108 4 Shear and Moment in Beams

w N/m w N/m

(a) Propped beam (b) Fixed or restrained beam

w2 N/m

T*! |J? 2 1*3 1*4

(c) Continuous beam

Figure 4-2. Statically indeterminate beams.

load, the intensity of loading increases or decreases at a constant rate,

as in 4-2a and 4-2b; this condition might result from water
Fig.
pressure acting on the face of a dam or from the dumping of a pile of
sand. The trapezoidal loading in the right segment of Fig. 4-2c is a
combination of a uniform and a uniformly varying load. The loading
may also be nonuniform, as in the left segment of Fig. 4-2c; this may
result from the haphazard piling of sand bags.

4-2 SHEAR AND MOMENT

Figure 4- 3a shows a simple beam that carries a concentrated load P

and is R and R v For the time
held in equilibrium by the reactions x

being, neglect the mass of the beam

and consider only the effect of
itself

the load P. Assume that a cutting plane a-a at a distance x from R x

divides the beam into two segments. The free-body diagram of the left
segment in Fig. 4- 3b shows that the externally applied load is R v To
maintain equilibrium in this segment of the beam, the fibers in the
exploratory section a-a must supply the resisting forces necessary to
satisfy the conditions of static equilibrium. In this case, the external
load is vertical, so the condition HX = (the X axis is horizontal) is

automatically satisfied.
To satisfy SX = 0, the vertical unbalance caused by R x
requires
the fibers in section a-a to create a resisting force. This is shown as Kr ,

and is called the resisting shearing force. For the loading shown, Vr is
numerically equal to /?,; but if additional loads had been applied
4-2 Shear and Moment 109

-* X >
a r
i
a
—L
i

R *
t
>
(a)

Vri ,

\f Mf
I
r r

x >-
L-x
*1
(b) (c)

Figure 4-3. Equilibrium of segments to left and right of any exploratory section
a-a.

between R and section a-a (as in Figs. 4-5 and 4-6), the net vertical
x

unbalance (which is equal but oppositely directed to the resisting

shearing force) would be found from the summation of their vertical
components. We define this net vertical unbalance as the shearing force
in the beam. It is denoted by V and may be determined from the
summation of the vertical components* of the external loads acting on
either side of the section. However, for the reason given in the footnote
on page 1 1 2, it is simpler to restrict this summation to the loads that act
on the segment to the left of the section. This definition of shearing
force (also called vertical shear or just shear) may be expressed mathe-
matically as

V=GY) L (4-1)

the subscript L emphasizing that the vertical summation includes only

the external loads acting on the beam segment to the left of the section
being considered.
The resisting shear Vr set up by the fibers in any section is always
equal but oppositely directed to the shearing force. V. In computing V>
upward acting forces or loads are considered as positive. This rule of
sign produces the effect shown in Fig. 4-4, in which a positive shearing
force tends to move the left segment upward with respect to the right,
and vice versa.

•The beam is assumed to be horizontal. With the beam in any other position,
the shearing force is computed from the summation of the components parallel to
the exploratory section.
110 4 Shear and Moment in Beams

Positive shear Negative shear

Figure 4-4. Relative movements corresponding to signs of shearing force.

For complete equilibrium of the free-body diagram in Fig. 4-3b,

the summation of moments must also balance. In this discussion, R x

and Vr are equal, thereby producing a couple that is equal to R x M x

and is called the bending moment because it tends to bend the beam. The
fibers in the exploratory section must create a numerically equal resist-
ing moment, A/r that acts as shown.* In most beams, the free-body
,

diagram carries a number of loads, as shown in Fig. 4-5; hence a more

complete definition of bending moment is necessary.

r r t

• B

«i

Figure 4-5.

Definition of bending moment

Bending moment is defined as the summation of moments about

the centroidal axis of any selected section of all the loads acting either
to the left or to the right side of the section, and is expressed mathemati-
cally as

M = (2M) L = (SM), (4-2)

the subscript L indicating that the bending moment is computed in

terms of the loads acting to the left of the section, and the subscript R
referring to loads to the right of the section.
Why the centroidal axis of the exploratory section must be chosen
as the axis of bending moment may
not be clear at this point; however,
the reason is explained in Art. 5-2. Actually, in Fig. 4-5, where the

•Art. 4-3 shows that the bending moment, and hence the resisting moment,
is always a couple.
4-2 Shear and Moment 111

loads are perpendicular to the beam, the axis of bending moment may
be at point A, or B, or anywhere in the exploratory section, without
changing the moment arms of the applied loads. But if the applied loads
are inclined to the beam as shown in Fig. 4-6, the moment arms of the
applied loads are unspecified unless the moment axis is at a definite
location in the exploratory section. Such inclined loads cause combined
axial and bending effects which are discussed in Art. 9-2.

A
\ / M r

v
Figure 4-6.

Sign of bending moment

To many bending moment is positive if it produces

engineers,
bending of the beam concave upward, as in Fig. 4-7. We prefer to use
an equivalent convention which states that upward acting external forces
cause positive bending moments with respect to any section; downward
forces cause negative bending moments. In so far as the left segment of a
beam is concerned (Fig. 4-3b), this is equivalent to taking clockwise
moments about the bending axis as positive, as indicated by the mo-
ment sense of R v With respect to the right segment of a beam (Fig.
4-3c), this convention means that the moment sense of the upward
reaction R 2 is positive in a counterclockwise direction. This convention
has the advantage of permitting bending moment to be computed,
without any confusion in sign, in terms of the forces to either the left or
the right of a section, depending on which requires the least arithmetical
work. We never need think about whether a moment is clockwise or
counterclockwise; upward acting forces always cause positive bending
moments regardless of whether they act to the left or the right of the
exploratory section.

Positive Bending Negative Bending

Figure 4-7. Curvatures corresponding to signs of bending moment.

112 4 Shear and Moment in Beams

The definition of shearing force and bending moment may be

summarized mathematically as

V=(1Y) L (4-1)

M = &M) L = (2A/)* (4-2)

in which positive effects are produced by upward forces and negative

effects by downward forces. This rule of sign* will be used exclusively
hereafter, and it will be further extended to give a positive sign to any
quantity or expression in which such adjectives as "up" or "above" are
used, and vice versa for negative signs. Remember that the subscripts L
and R refer to the beam segment lying respectively to the left and right
of the exploratory section.

ILLUSTRATIVE PROBLEM

401. Write shear and moment equations for the beam loaded as
shown in Fig. 4- 10a and sketch the shear and moment diagrams.

Solutions: Begin by computing the reactions. Applying *ZMR =

gives /?, =63 kN, and SM^ = yields R2 — 67 kN. A check of these
values is given by 2^ = 0. The sections in the beam at which the
loading conditions change are called change of load points and are
designated by the letters A, B, C, and D.
If a section a -a is taken through the beam anywhere between A

and B, the external loads on it appear as in Fig. 4-8. Applying the

definitions of vertical shear and bending moment, and noting that they
apply only to external loads, we obtain

[K = (2 7) L ] VAB = (63 - 20*) kN (a)

[M = (1M) L ]
MAB = 63* -
(20*)f
= (63* - 10*
2
)kNm (b)

These equations are valid only for values of * between and 5,

that between points A and B. To obtain shear and moment equations
is,

between B and C, assume another exploratory section b-b taken

anywhere between B and C. Note that the location of section b-b is still
defined in terms of * as measured from the left end of the beam,
although * now ranges between the limits of 5 and 10. The effects of the

•To avoid conflict with this rule, it is necessary to compute vertical shear in
terms of the forces lying to the left of the exploratory section. If the forces to the
right of the section were used, it would be necessary to take downward forces as
positive so as to agree with the sign convention shown in Fig. 4-4.
 (20 x) kN 100 kN
1 1
1

-*2.5 m^+-*- x-\ >i.O

wmm
FT

1
'//////At//

i i
y&
1
\

-X . 1

R = x
63 kN R = 63 kN
L

Figure 4-8. Figure 4-9.

-* x 30 kN

20 kN/m £
16
k I
X
Load
diagram
(a)

-5 m- 5 m- 4m >
i?!=63kN = 67 kN
2

'63kN

30 kN
(b)
Shear
diagram
-*=3.15m-|

199.23 kN-m

(c)
Moment
diagram

(d)
Elastic-
curve

Figure 4-10.
113
114 4 Shear and Moment in Beams

external forces on this section are determined by applying the defini-

tions of shear and moment to Fig. 4-9.
[v-Vrh] VBC = 63 100= -37 kN (c)

[M - (2A/) L ] MBC = 63* - 100(jc - 2.5)

= (-37jc + 250)kN-m (d)

100 kN
I

2.5 m > \
« x-2.5

10 m-
/? 9 =67 kNl
nI
x -(x-10j

/? L = 63kN

Figure 4-11.

The shear and moment equations for segment CD are obtained

similarly by passing a section c-c anywhere between C and D. The
external loads acting on the beam to the left of this section are shown in
Fig. 4- 1 1 whence we obtain
,

[V - (2 Y) L VCD = 63 - 100 + 67 = +30 kN (e)

]

[M - (2Af) L ] ^cz> - 63jc - 100(jc - 2.5) + 61(x - 10)

= (30jc -420)kN-m (/)
A simpler method of obtaining MCD is to consider the forces lying
to the right of section c-c as shown in Fig. 4-12, from which noting
that downward forces produce negative bending moment, we also
obtain

[M - (2M) R ] MCD - -30(14 - x)

= (30.x -420)kN-m (/')

Summarizing, we have computed V by considering only the ex-

any exploratory section, whereas
ternal forces lying to the left of may M
be computed by taking moments about the exploratory section caused

30 kN
<|^-M-*-J

Figure 4-12.
4-2 Shear and Moment 115

by the external loads which lie either to the left or to the right of the
section. We have been careful to assign plus signs to V and caused M
by upward acting loads, and minus signs to V and caused by M
downward acting loads. We shall be consistent in assigning a plus sign
to any upward quantity and a minus sign to any quantity associated
with the word "down" or its equivalent.
Note further that Figs. 4-8, 4-9, 4-11, and 4-12 have been used
only for explanation; you will soon learn to visualize such diagrams
directly from the original beam loading.

Shear and moment diagrams

Shear and moment diagrams are merely the graphical visualization

of the shear and moment equations plotted on V-x and M-x axes,
usually located below the loading diagram, as in parts (b) and (c) of Fig.
4-10.
The diagram (Fig. 4- 10b) are joined by
discontinuities in the shear
vertical lines drawn up or down to represent the abrupt changes in shear
caused respectively by upward or downward concentrated loads. Art.
4-4 (page 134) shows why this is correct. A final point to be observed
at this time is that the highest and lowest points on the moment diagram
(Fig. 4- 10c) always correspond to sections of zero shear. This observa-
tion is also discussed in Art. 4-4, but it should be noted now that the
value of x making MAB maximum can be found by differentiating MAB
with respect to x and equating the result to zero. This result will be the
shear equation VAB . Thus we see that maximum moment corresponds to
the section of zero shear.
Shear and moment at change of load points can be computed by
substituting appropriate values of x in the foregoing V and M
equations
{a to /), but it is simpler and more direct to compute these numerical
values by applying the fundamental definitions of and to specific V M
For example, the section of zero shear between A and B c^curs
sections.
because the downward force due to x meters of load applied at 20
kN/m must balance the vertical shear of 63 kN at A. Hence we have
= 20* or x = 3.15 m
63
The moment at this section of zero shear is computed by taking
moments of the forces to the left of the section. These forces consist of
the upward reaction R — 63 kN and the downward load of 63 kN
Y

caused by the length of the uniformly distributed load necessary to

cause zero shear. From the definition of bending moment, we obtain

[M = C2M) L ]

at jc = 3.15, M = (63)(3.15) -63(^-) =99.23 kNm

116 4 Shear and Moment in Beams

A brought out in Fig. 4-10d, which shows

final point of interest is
the shape taken by the beam under the given loading, assuming the
beam to be quite flexible. The beam between A and E is concave up,
and between E and D it is concave down. Since anything associated
with up has a positive sign, it is not surprising that the moment diagram
has positive values corresponding to the region AE, while for the
portion ED, where the beam is concave down, the moment diagram has
negative values. Sketching the shape of the beam therefore provides a
check of the sign of bending moment.
At point E, where the beam changes its shape from concave up to
concave down, we have what is called a point of inflection; it corre-
sponds to the section of zero bending moment. Its position may be
calculated by setting Eq. (d) equal to zero, which yields

[MBC - 0] -37x + 250 = 0; x = 6.76 m

ILLUSTRATIVE PROBLEM

402. Write the shear and moment equations for the cantilever
beam carrying the uniformly varying load and concentrated load shown
in Fig. 4-13. Also sketch the shear and moment diagrams.

6 kN/nw 20 _kN

Load diagram

Shear diagram

3rd degree curve

-36kN-m Moment diagram
-72 kN-m

148 kN-m
Figure 4-13.
4-2 Shear and Moment 117

Solution: Shear and moment calculations for a cantilever beam are

simplified by drawing the load diagram with the restraining wall at the
right end. Drawing diagrams similar to Fig. 4-8 or 4-9 where necessary,
we find the shear and moment equations between each change of load
position by applying Eqs. (4-1) and (4-2). For the region AB, in which
x varies from to 6, we have (see Fig. 4-14),

[K-(Zr) L ]
K,„=
AB -4^kN
[M-(2M) L ] A/, B =-^(|)=-^kN-m

Force

lx
18 kN
4 m — I

»-f-**-4*-
6kN/m

Figure 4-14. Flgure 4-15.

After passing B, the resultant force due to the triangular load

(equal to its area) is constant at value of j(6)(6) = 18 kN acting through
the centroid of the triangular load diagram at 4 m from A. For the
region BC y
in which x varies between 6 and 8, we therefore obtain (see
Fig. 4-15)
[V-&Y) L ]
nr = - 18 kN
VBC
[M - (2M) L ] MBC = -18(x - 4) = (-18.x + 72) kN-m
For a section between C and D (Fig. 4-16) in which x varies from
8 to 10, we obtain

[V-&Y) L ]
VCD = -18 - 20= -38 kN
[M = (2A/) L ] MCD = - 18(jc - 4) - 20(.x - 8)
= (-38.x + 232)kN-m

18 kN
< 4 m H"*— x-4
i
20 kN
i

8 m

Figure 4-16.
118 4 Shear and Moment in Beams

The shear and moment equations are graphed as shown in the

shear and moment diagrams in Fig. 4- 13. At the wall D the diagrams y

are returned to zero by means of the shear and moment reactions

exerted by the wall. Observe that the maximum shear and the maximum
bending moment always occur at the restrained end of a cantilever
beam. An exception to this rule is possible if some of the loads are
upward and the other loads downward. Note also that the moment
equation M
CD is not necessary in computing the bending moment at D;
this is found by a direct application of Eq. (4-2):

M= (2M) L = - 18(6) - 20(2) - - 148 kN-m

PROBLEMS
Write shear and moment equations for the beams in the following
problems. Also draw shear and moment diagrams, specifying values at
allchange of loading positions and at all points of zero shear. Neglect
the mass of the beam in each problem.

403. Beam loaded as shown in Fig. P-403.

Ans. VCD = 20 kN; MCD = 20(x - 140) kN-m

50
)U K
k 20 kN
!m-»

1
V 6 m % m

Figure P-403.

404. Beam loaded as shown in Fig. P-404.

Ans. MCD = (-4jc + 28) kN-m
10 kN
3m 2 m
t2m-*
&
40 kN-m

Figure P-404.
4-2 Shear and Moment 119

405,, Beam loaded as shown in Fig. P-405.

Arts. MBC = (-5x 2
+ 44x + 60)kN-m

30 kN

H 10 kN/m

10 m

Figure P-405.

406. Beam loaded as shown in Fig. P-406.

20 kN 40 kN

[-—2 i

"1c
20 kN/m

B D
L
i

l-«— 2 m—+Y*
\ 4 m *-

Figure P-406.

407. Beam loaded as shown in Fig. P-407.

Ans. Max. M = 57.6 kN m
•

r+ — 2m-—> ^ 2 m *-

B 30 kN/m C
A 1

I
O 111

Figure P-407.
120 4 Shear and Moment in Beams

408. Beam loaded as shown in Fig. P-408.

Arts. Max. M = 83.33 kNm
2m -

B
30 kN/m
15 kN/m
A
I i

Figure P-408.

409. Cantilever beam loaded as shown in Fig. P-409.

H>N/m B
A L L
2 2

Figure P-409.

410. Cantilever beam carrying the uniformly varying load shown

in Fig. P-410.

Figure P-410.

411. Cantilever beam carrying a distributed load varying from w

N/m at the free end to zero at the wall, as shown in Fig. P-41 1.
Arts. M= (wx 3 /6L) - (wx 2 /2)

w N/m

Figure P-411
 — —
4-2 Shear and Moment 121

412. Beam loaded as shown in Fig. P-412.

Ans. Max. M = 25 kN m
2m—Hfi
10 kN/m

c -6 m

Figure P-412.

413. Beam loaded as shown in Fig. P-413.

25kN-m
10 kN/m

f# C

«,
1 m '1 m
+ 3
=P 2 m

Figure P-413.

414. Cantilever beam carrying the loads shown in Fig. P-414.

2 kN/m ^
yyy( 5 kP
I

B C
-+— 2m + -* 3 m
Figure P-414.

415. Cantilever beam loaded as shown in Fig. P-415.

Ans. MBC = (-4x 2
+ 20.x - 40) kN-m

////}
8 kN/m

1 2m LB 3m C
W)
20 kN
Figure P-415.
 *

122 4 Shear and Moment in Beams

416. Beam carrying uniformly varying load shown in Fig. P-416.

Ans. Max. M
= wL 2 /9V3

Figure P-416.

417. Beam carrying the triangular loading shown in Fig. P-417.

Ans. Max. M - wL 2
/\2

Figure P-417.

418. Cantilever beam loaded as shown in Fig. P-418.

5kN/m

2 m — 2 m
Figure P-418.

419. Beam loaded as shown in Fig. P-419.

Ans. Max. M = 27.89 kN-m

Figure P-419.
4-2 Shear and Moment 123

420. A total distributed load of 60 kN supported by a uniformly

distributed reaction as shown in Fig. P-420.

60 kN

H"™
-*-2 m -*-«— 4 m
r

»
TT
-*-2m-^
Figure P-420.

421. Write the shear and moment equations for the built-in
circular bar shown in Fig. P-421 if (a) the load P is vertical as shown
and (b) the load P is horizontal to the left.

Ans. (a) V = -P cos 9: M= -Pi? sin

Figure P-421.

422. Write the shear and moment equations for the semicircular
arch shown in Fig. P-422 if (a) the load P is vertical as shown and (b)
the load P is applied horizontally to the left at the top of the arch.
Ans. (a) MAB =\PR(\ - cos 9); MBC =\PR(l + cos 9)

IP

\b
:

^^^S\
^f^
\
\
ft/
R
\
I 99 9
///////
•Xv> 1
7//////

Figure P- 422.
 — -

124 4 Shear and Moment in Beams

4-3 INTERPRETATION OF VERTICAL SHEAR

AND BENDING MOMENT

The beam in Fig. 4-l7a carries a uniformly distributed load as well as

concentrated loads. The external effects of each load acting to the left of
section b-b are shown separately and (d). In each of
in parts (b), (c),
these figures, the effect of the applied load has been transferred to the
exploratory section by adding a pair of equal but oppositely directed
forces at that section which, as shown in the right side of the figures,
reduce to a force at that section plus a couple. The moment of the
couple is equal to the bending moment of the load. Hence, as is shown
in the composite figure (e), the effect of the loads at one side of an
exploratory section reduces to a system of forces whose vertical summa-
tion is the vertical shear and a system of couples whose algebraic

H
summation is the bending moment.

1
6
**
w N/m 1
(a)

-« .t >
R, 1

(b)

,f
* X
A«i

=&
wx

(c)

-« X-
Jwx
n p'i
-^(x-a)~i •

(d)
X 9-| h« X-

1* ?
1
V=(XY)L
Ri

(e)
r—^NT^n
r
Ri
x — X
wx
»
M=(ZM) r

PJ
Figure 4-17. Shear and moment are resultant effects of loads acting to one side
of exploratory section.
 d

4-3 Interpretation of Vertical Shear and Bending Moment 125

r2

1
w N/m w N/m
) M, M;
c,
——
*i
(a) (b)

Figure 4-18. Free-body diagrams of segments of Fig. 4-17 in terms of shear and
moment.

We may therefore conclude that the resultant effect of the forces

at one side of an exploratory section reduces to a single force and a
couple which are respectively the vertical shear and the bending mo-
ment at that section. Applying these forces to a free-body diagram of a
beam segment produces equilibrium of that segment. Thus, in Fig. 4-18,
the segments to the left and right of section b-b in Fig. 4-17 are held in
equilibrium by the shear and moment at section b-b. As far as the left
segment is concerned, these equilibrants are resisting shear and resisting
moment; but insofar as the right segment is concerned, they represent
actual shear and bending moment. In other words, a beam may be cut
at any section and the effect of the loads acting to one side of the
section be replaced by the shear and moment at that section.
One application of this concept is that we can compute shear and
moment at any section in terms of the shear and moment at another
section. Thus in Fig. 4- 18b, the moment at section c-c will be

,2
wz
M c
= Mb + Vb z -

4-4 RELATIONS BETWEEN LOAD, SHEAR, AND MOMENT

In this article we between the loads,

shall discuss the relations existing
shears, and bending moments any beam. These relations provide a
in
method of constructing shear and moment diagrams without writing
shear and moment equations. The relations are not independent of the
basic definitions of shear and moment; instead they supplement them
and are used in conjunction with them.
We begin by considering the beam in Fig. 4- 19a, which is
assumed to carry any general loading. The free-body diagram of a
segment of this beam of length dx is shown magnified in Fig. 4- 19b. As
we saw in the preceding article, the effect of the loads to the left of this
segment reduces to the shear V and the moment M
and the loads to the
,

right of this segment produce the slightly different values of shear and
126 4 Shear and Moment in Beams

\+dx
'

T W I T I l\ I

*i ^W
(a)

Figure 4-19.

moment V + dV and M+ </A/. Although the loading is variable, it may

be assumed constant at the intensity of w N/m over the small length dx,
thereby producing the upward load w dx which completes the free-body
diagram.
Applying the conditions of static equilibrium to Fig. 4- 19b, a
summation of vertical forces yields

[1Y = 0] K+ wdx - (V + dV) =

which reduces to
dV= wdx (a)

From a moment summation about point B we have

[2MB = 0] M+ V dx + (w^)-y- (A/+ </A/) =0

The third term in this equation is the square of a differential that is

negligible in comparison with the other terms; hence the equation

reduces to

dM = V dx (*)
Integrating Eq. (a), we obtain
/'V2 rx 2
w dx

in which the limits are the shear V at position x and the shear V, at x x

positionx 2 The left-hand term is easily integrable; it reduces to V2 -

.

K, and represents the change in shearing force between sections x 2 and

Xj. We this change in shear as AK. In the right-hand term the
denote
product dx represents the area of an element of the load diagram like
h>
x
that shown shaded in Fig. 4- 19a. Hence the definite integral j x 2 w dx,
which means the summation of such terms, represents the area under
the load diagram between positions x and x 2 Therefore the integration x
.

of Eq. (a) yields

V, = W= (Area),^ (4-3)
4-4 Relations Between Load, Shear, and Moment 127

Similarly the integration of Eq. (b) gives

dM = / V dx

This reduces to

M 2
- M x
» AM = (Area) shear (4-4)

inasmuch as the product V dx in the right-hand integral represents the

area of an element under the shear diagram. Therefore the integral itself

is equivalent to the area under the shear diagram between positions x l

and x 2 Expressed in words, Eq. (4-4) shows that the change in bending
.

moment AM
between any two sections is equal to the area of the shear
diagram for this interval.
Positive shearing forces are plotted upward from the X axis; hence
positive shear areas are those which lie above the X axis and represent
increases in the bending moment. The load diagram, however, is usually
drawn with the loads on top of the beam because this is their natural
position; as a consequence, the area of such downward acting loads is

considered negative and represents decreases in the shearing force.

Equations (4-3) and (4-4) provide a convenient means of comput-
ing the changes in shear and moment and also the numerical values of
shear and moment at any section, as will be demonstrated in the
illustrativeexamples below. Of almost equal importance are the follow-
ing variations of Eqs. (a) and (b\ which enable us to sketch the shapes
of the shear and moment diagrams:

w= —
dV =
slope of shear diagram (4-5)

V = —— = slope of moment diagram (4-6)

dx
As an application of these principles, consider the simply sup-
ported beam carrying the variable loading shown in Fig. 4-20a. Since
positive slopes are directed up to the right and negative slopes are
directed down to the right, i.e.,

+ slope =

— slope =

we observe from Eq. (4-5) that the shear diagram in Fig. 4-20b must
slope continuously down to the right. The inclination varies directly
with the corresponding ordinate of the load diagram, being steepest
where the load ordinate is maximum, and horizontal (or of zero slope)
at the ends where the intensity of loading is zero.
128 4 Shear and Moment in Beams

(a)
Load
diagram

(b)
Shear
diagram A V = shaded area of (a)
x

AM =net shaded area of (b)

(c)
Moment
diagram

Figure 4-20. Relations between load, shear, and moment diagrams.

Similarly, by means of Eq. (4-6), the slope and shape of the

moment diagram in Fig. 4-20c are determined by the corresponding
ordinates of the shear diagram, which, being positive but decreasing in
magnitude moment diagram
for the left portion, determine that the
slopes continuously up to the right with decreasing inclination. The
slope becomes zero where the shear ordinate is zero. Note also that after
the shear ordinates change sign and become increasingly larger nega-
tively, the moment diagram slopes correspondingly more steeply down

to the right. These conditions establish a maximum moment at the

ordinate of zero shear.
The changes in shear (AK) and moment (AM) defined by Eqs.
(4-3) and (4-4) are indicated in Figs. 4-20b and 4-20c. The shaded
negative area of the load diagram determines AK to be negative or
directed downward; in the shear diagram, the excess of positive over
negative area in the region between jcj and x 2 determines the positive
change moment AM, which is therefore directed upward.
in

A summary of the principles presented here and in Art. 4-2

suggests the following procedure for the construction of shear and
moment diagrams:

1. Compute the reactions.

2. Compute values of shear at the change of load points, using

either K= (2 v) L or AK = (area) load .

3. Sketch the shear diagram, determining the shape from Eq.

(4-5); i.e., the intensity of the load ordinate equals the slope at the
corresponding ordinate of the shear diagram.
4-4 Relations Between Load, Shear, and Moment 129

4. Locate the points of zero shear.

5. Compute values of bending moment at the change of load
points and at the points of zero shear, using either M = (2M) L =
(ZM)R or AM = (area) shear , whichever is more convenient.
6. Sketch the moment diagram through the ordinates of the
bending moments computed in step 5. The shape of the diagram is
determined from Eq. (4-6); i.e., the intensity of the shear ordinate
equals the slope at the corresponding ordinate of the moment diagram.

ILLUSTRATIVE PROBLEMS

423. Using the semigraphical method described in this article,

sketch shear and moment diagrams for the beam shown in Fig. 4-21,
computing the values at all change of loading points and the maximum
shear and maximum moment.

P=8kN

Shear diagram

12 kN

Moment diagram

9kN-m
Figure 4-21. Load, shear, and moment diagrams.
130 4 Shear and Moment in Beams

Solution: The reactions are determined by equating to zero a moment

summation about R 2 and then about /?,; this yields respectively /?, = 20
kN and R 2 = 12 kN.
We next determine the values of vertical shear at the change of
load positions. At A, the shear is zero. At the left of B, applying
V = (ZY) L gives the shear as -6 kN caused by the downward re-
sultant of the distributed load of 2 kN/m applied for 3 m. The same
result may
be obtained with Eq. (4-3), which indicates that the change
in shear between A and B caused by the downward or negative
uniformly distributed load is equal to the area of the load diagram in
this interval, that is, AV = —2x3= —6 kN. Hence the shear ordinate
to the left of B has decreased 6 kN from the zero shear ordinate at A , to
yield a net value of — 6 kN. The concentrated load reaction at B causes
the shear at B to increase abruptly by 20 kN to a net positive shear
ordinate of 14 kN at the right of B.
Between B and C, the area of the load diagram is —2x9 =
— 18, which by Eq. (4-3) represents the change in shear between B and
C. The net shear ordinate at the left of C is therefore Vc = VB + A V =
14 — 18 = —4 kN. At C, the concentrated load of 8 kN changes the
shear ordinate to —12 kN at the right of C. The shear ordinate stays
constant at this value between C and D, since there is no load in this
interval; at D, the upward reaction of 12 kN reduces the shear
ordinate to zero.
The shape of the diagram connecting these shear ordinates is
determined from Eq. (4-5), which shows that the slope is equal to the
corresponding ordinates of the load diagram. Thus between A and B,
the load intensity is constant and downward (or negative); hence the
slope of the shear diagram in this interval is constant and down to the
right. Similarly, between B and C, the load intensity is constant and
negative; therefore the slope of the shear diagram here also is constant
and down to the right. The slopes in the intervals AB and BC are
because they are each equal to the same load intensity. Finally
parallel
between C and Z>, the intensity of loading is zero and the corresponding
slope of the shear diagram is zero (a horizontal line).

We may conclude therefore that the shear diagram consists of

straight horizontal lines for intervals in which the load intensity is zero,
and of straight inclined lines for intervals of uniform load intensity.
The shear diagram passes through zero at B, where x = 3 m, and
also at E. The position of E is determined from the fact that the shear at
the right of B is 14 kN, which is reduced to zero in the interval BE at
the rate of 2 kN/m. Hence BE = d = 14/2 = 7 m.
As a preliminary to computing the bending moments, we de-
termine the areas of the shear diagram marked A A 2 A 3 and A 4 x
, , , .

A, -I(3)(-6) = -9kNm
4-4 Relations Between Load, Shear, and Moment 131

^2 =I(7)(14) = + 49 kN-m
^3-l(2)(-4)= -4 kN-m
A 4 = 3(-12) = -36 kN-m
According to Eq. (4-4), the change in bending moment between
any two sections equals the corresponding area of the shear diagram;
hence, since the bending moment is zero at A (there are no loads to the
left of A to cause a bending moment), the bending moment at B is given

by A v or B = M— 9 kN-m.
Similarly, the bending moment at E is

ME = mb + AM = A x
+ A 2 = -9 + 49 - +40 kN-m
The bending moment at C can also be computed as the sum of the
areas A l9 A 2 and A 3 giving
, , Mc = 36 kN-m; but small errors arising
from neglecting sufficient significant figures in computing these areas
may cause a cumulative error. Hence at sections near the right end of
the beam, it is usually preferable to use the shear area to the right of
M = (£M) R whence in
such sections, or to apply the basic definition ;

terms of the loads acting to the right of C we find Mc = 12 X 3 = 36

kN-m. The correlation between this result and the area A 4 is evident if
we observe that A 4 represents the amount by which the moment at C
changes to become zero at D; that is,
MD = Mc + AM = Mc + A 4

or

= Mc - 36 and Mc = 36 kN-m
Whenever the change in bending moment between the ends of a
beam is zero, as in this problem, the net area under the shear diagram is

also zero; in other words, there must be as much positive as there is

negative shear area. This provides a useful check on the accuracy of all
intermediate values of bending moments computed from the area of a
shear diagram.
After these bendingmoments are plotted as ordinates, the shape
of the moment diagram connecting them is determined from Eq. (4-6);
that is, the intensity of the shear ordinate equals the slope at the
corresponding ordinate of the moment diagram. Thus as the shear
ordinates between A and B change linearly from zero to —6 kN, the
slopes at corresponding ordinates of the moment diagram change from
horizontal or zero slope at A to increasingly steeper negative slopes as
we pass from A producing a second-degree curve concave
to B, thereby
downward, as shown. In other words, the tangents to the moment curve
have increasingly steeper slopes directed downward to the right.
The sudden change in shear at from —6 to 14 kN causes the
B
moment curve there to slope abruptly upward to the right; it becomes
132 4 Shear and Moment in Beams

less steep, and eventually horizontal, as we move from B to E, because

the corresponding positive shear ordinates decrease uniformly to zero.
From E to C, however, the slope is again increasingly steeper and
directed downward to the right as the corresponding shear ordinates
change their sign to negative.
At C, the shear ordinate changes abruptly from —4 to — 12 kN, at
which it remains until D. The slope of the moment diagram is corre-
spondingly abruptly steeper and constant, resulting in the straight line.

From above discussion, it is apparent that an abrupt change in

the
shear creates an abrupt change in the slope of the moment curve as at
C, and that if the shear changes sign abruptly, there is a cusp in the
moment diagram as at B. Also, where the shear diagram is constant
because only concentrated loads are involved, as in the region CD, the
moment diagram consists of straight lines, whereas for an interval in
which the shear diagram varies uniformly because of a uniformly
distributed load, the moment diagram is a parabolic arc with a vertical
axis of symmetry at the section of zero shear, as at E.

424. Without writing shear or moment equations, sketch shear

and moment diagrams for the beam in Fig. 4-22, and compute values at
all change of loading points and the values of maximum shear and

maximum moment.

Solution: We begin by computing the reactions. Replacing the uni-

formly distributed reaction between C and D
by its resultant R 2 and
equating moments about R {
to zero, we R 2 = 42 kN. We then set
obtain
moments about R 2 equal to zero, whence R = 24 kN. Dividing R 2 by
x

the length of 4 m, over which it is assumed to be uniformly distributed,

gives the upward intensity of this reaction as 10.5 kN/m.
The shear diagram starts with an abrupt change in shear of 24 kN
caused by the reaction R Applying AK= (area) load between A and 2?,
}
.

the change in shear is equal to the area of the triangular load diagram,
j X 9 X (-12) - - 54 kN; this reduces the vertical shear at B to -30
kN. The shape of the shear diagram from A to B is determined from Eq.
(4-5), which shows that the shape must vary from zero slope to
increasingly steeper negative slopes corresponding to the increasingly
greater downward intensity of the loading.
Between B and C, the intensity of loading is zero; hence, from Eq.
(4-5), the slope of the corresponding portion of the shear diagram is
zero, that is, horizontal.
From C to D, the loading is at a constant upward rate of 10.5
kN/m, creating a positive change of shear of 42 kN equal to the area of
the load diagram in this interval, and a constant upward slope of the
shear diagram as shown. This shear change of +42 kN added to the
shear of -30 kN at C produces the shear of + 12 kN at D. From D to
4-4 Relations Between Load, Shear, and Moment 133

^=24=^
12 kN

jTTTT r~4 w = —= 10.5 kN/m

Shear
diagram
(values inkN)

Moment
diagram
(values in kN- m)

- 18.84

Figure 4-22. Load, shear, and moment diagrams.

E, the loading is zero, which means that the slope is horizontal. The
concentrated load of 12 kN at E reduces the shear abruptly to zero.
Before we locate the positions of zero shear at F and G on the
shear diagram, consider the effect of narrowing the distance over which
the reaction R2 is distributed. If points C and D are moved an equal
amount toward each other, this narrowing will not change the magni-
tude or position of However, a reduction in the distance of CD will
R2 .

increase the intensity at which R 2 is distributed and cause a correspond-

ing increase in the slope of the shear diagram in the narrowed interval
134 4 Shear and Moment in Beams

CD. For the extreme case in which the reaction is distributed over an
infinitesimal width — that is, becomes a concentrated force — the inten-
sity of loading is infinitelyupward, and the corresponding slope of the
shear diagram is vertically upward, as is the case at the reaction R x
.

Similarly for a concentrated downward load at E y

the intensity of
loading is infinitely downward, and the corresponding slope of the shear
diagram is vertically downward. This explains why a concentrated load
causes an abrupt change in the vertical shear.
Let us now locate the sections of zero shear. The vertical shear of
24 kN at A is reduced to zero at due to the load F by the force P x

diagram (shown shaded) applied over the interval AF. Evidently the
magnitude of P is also 24 kN and is equal to the area \xy, where y
x

denotes the intensity of loading at the section F. Hence

24 = \xy (a)

Another relation between x and y is obtained from similar trian-

gles in the load diagram:

y 12 n
7= y <* y = Yx it.\
(b)

which is substituted in Eq. (a) to obtain

l X2
1A 2*'~9~*

from which

*-&£-» and x= 6m
The section of zero shear at G is found from the fact that the
upward reaction applied over the interval CG must total 30 kN in order
to reduce the shear of — 30 kN at C to zero at G. Since the reaction is

distributed at 10.5 kN/m,

30 - \0.5d or d= 2.86 m
or, in terms of x measured from the left end of the beam,
x = 10+ d= 12.86 m v

Bending Moment The moment at F where x = 6 is found by m

applying the definition of bending moment.* Note that the shaded part
of the triangular load applied between A and F totals 24 kN and acts at
the centroid of the triangular area, that is, at j of 6 m from F. Thus we
•It is better not to use Eq. (4-4) when the shear diagram is curved, since

then the areas under the diagram are not too easily computed, especially in the
interval FB. The computation of such areas is discussed in Chapter 6, especially on
p. 231 and in Fig. 6-13. For the present, when the shear diagram is curved, as
over the region AB, compute the bending moment by applying (ZM) L or (ZM) R
rather than using the area of the shear diagram.
4-4 Relations Between Load, Shear, and Moment 135

obtain

[M = (2M) L ] MF = 24(6) - 24(f) = 96 kN-m

Similarly, the moment at B where x = 9 m is found to be

[M - (2A/) L ] MB - 24(9) - 54(f) = 54 kN-m

The moment at C can also be computed from this basic definition,
but whenever the shear diagram consists of straight lines (either hori-
zontal or inclined) it is usually simpler to apply Eq. (4-4); that is, the
change in bending moment between any two sections equals the corre-
sponding area of the shear diagram. For example, between B and C, the
area of the shear diagram is a rectangle and equals AAf = — 30 X 1 =
— 30 kN-m. Hence, if the bending moment changes by —30 kN-m
between B and C, the bending moment at C is

[Mc = Mc = 54 - 30 = 24 kN-m
MB + AM]
The bending moment at D may be found from MD — (2M) R =
— 12xl = — 12 kN-m; the negative sign comes from the downward
load. However, to emphasize the significance of the shear diagram, note
that the area of the shear diagram between D and E equals 12 X 1 = 12
kN-m. Since this change in moment is positive, it follows that to
produce zero moment at the free end E, the moment at D must be — 12
kN-m.
Similarly, the area of the shear diagram between G and D equals
\ X 12 X 1.14 = 6.84kN-m. Since this is a positive increase in bend-
ing moment, the moment at G must be smaller by this amount than the
moment at £>, or

MG = - 12 - 6.84 = - 18.84 kN-m

Sketching the bent beam, as in Fig. 4-10d in Problem 401, shows it to
be concave downward at G and D, which is further verification of the
negative bending moment signs at these sections.

Shape of the Moment Diagram. After plotting the values of the

bending moment, we consider next the shape of the moment curve
connecting these points. Applying Eq. (4-6), we notice that since the
shear ordinates are positive and decrease to zero as we move from A to
F, the moment curve has correspondingly positive slopes (that is,

directed upward to the right) that decrease to zero slope at F.

Between F and B, the vertical shear is increasingly negative, and
hence the corresponding slopes of the moment curve become increas-
ingly steeper downward to the right until the steepest slope is reached at
section B. Between B and C, the shear stays constant; therefore the
slope of the moment curve is constant, being represented by the straight
line that joins sections B and C
136 4 Shear and Moment in Beams

Between sections C and G, the shear is negative, becoming zero at

G; hence the slope of the moment curve is negative and gradually
reduces to zero (that is, horizontal) at G. Similarly the increasing
positive shear between G and D results in an increasingly positive slope
(directed upward to the right) up to D. Between D and E the slope
remains constant because the vertical shear is constant between D and
E.
The moment curve between C and D is a symmetrical parabola
with its vertex at G, because at equal distances to either side of G the
shear is numerically equal but of opposite sign, thereby producing equal
slopes oppositely directed. The moment curve between A and B, how-
ever, is not symmetrical about F because the shear ordinates do not
have equal values at equal distances on either side of F; here the
moment curve is actually a third-degree parabola.

PROBLEMS

Without writing shear and moment equations, draw shear and mo-
ment diagrams for the beams specified in the following problems. Give
numerical values at all change of loading positions and at all points of zero
shear. (Note to instructor: Problems 403 to 420 may also be assigned for
solution by the semigraphical method described in this article.)

425. Beam loaded as shown in Fig. P-425.

30 kN 24 kN

2m f 3m { 1

f )
Figure P-425.

426. Cantilever beam acted upon by two forces and one couple
as shown in Fig. P-426.

50 kN
60 kN • m
t
=e=i
U-l m - 1 m-*
Figure P-426.
4-4 Relations Between Load, Shear, and Moment 137

427. Beam loaded as shown in Fig. P-427.

Arts. Max.A/ = 13.8 kNm

10 kN
2m
1 m 1 m 10 kN/m 2m

Figure P-427.

428. Beam loaded as shown in Fig. P-428.

Arts. Max. M= - 60kNm

60 kN

30 kN
1 m{
5 kN/m

4 m I 2m

Figure P-428.

429. Beam loaded as shown in Fig. P-429.

20 kN

20 kN/m
10 kN/m

-«— 2 m »
tLlm»

^i
^ 4 m-
=f R.
2m

Figure P-429.
138 4 Shear and Moment in Beams

430. overhanging beam shown in Fig. P-430, determine

In the P
so that the moment over each support equals the moment at midspan.
Ans. P= 8.75 kN

5kN/m

1 m 6 m

#1
f
#2

Figure P-430.

431. Beam loaded as shown in Fig. P-431.

50 kN 40 kN

m
2
tm
11 20
10
kN/m
kN/m

-7 m- -J-* 3 m —
Flgure P-431

432. A distributed load is supported by two distributed loads as

shown in Fig. P-432.

30 kN/m
3 m 2 D I

5 n

kN/m 10 2 kN/m
Figure P-432.
4-4 Relations Between Load, Shear, and Moment 139

433. Overhanging beam loaded by a force and a couple as shown

in Fig. P-433. Arts. Max. M« - 160 kN m

50 kN
200 kN-m

p 2 m —
3
^»4«« 3 m t
»4-« — 2 m * i

i?i

Figure P-433.

434. Beam loaded as shown in Fig. P-434.

Ans. Max. M= -36 kN-m

30 kN

60 kN-m

lm 4
20 kN/m

3m «
a [-• *-l

*i

Figure P-434.

435. Beam loaded and supported as shown in Fig. P-435.

20 kN 40 kN
lm lm
2m « 2 m- 2 m-

10 kN/m

irrii ii i ii i

T w kN/m

Figure P-435.
140 4 Shear and Moment in Beams

436. Cantilever beam loaded as shown in Fig. P-436.

Arts. Max. M = 55 kN m •

10 kN

10 kN/m

Figure P-436.

437. Cantilever beam loaded as shown in Fig. P-437.

20 kN
2m
^^
I '
2 m
1
y 1 m 15 kN/m lm n
'/yyA
40kN-m

Figure P-437.

438. A propped cantilever beam loaded as shown in Fig. P-438

consists of two segments joined by a frictionless hinge at which the
bending moment is zero.

Hinge rz

1 m
-+
15 kN/m
t ^
m 2m
3
A
Figure P-438.
4-4 Relations Between Load, Shear, and Moment 141

439. A beam
supported on three reactions as shown in Fig.
P-439 consists of two segments joined at a frictionless hinge at which
the bending moment is zero. Ans. Max. = 57.6 kN M • m

40 kN

2m -
Hinge

20 kN/m 1 m 1 m 20 kN/m
,

V* 4 m -4— 6 m
i

Figure P-439.

440. A frame ABCD, with rigid corners at B and C, supports the

concentrated load as shown in Fig. P-440. (Draw shear and moment
diagrams for each of the three parts of the frame.)

&

Figure P-440.

441. A beam ABCD is supported by a hinge at A and a roller at

D. It is subjected to the loads shown in Fig. P-441, which act at the

5kN

2m
2m B 2m 3m
\D
£5 2m
?m59
14 kN
Figure P-441.
142 4 Shear and Moment in Beams

ends of the vertical members BE and CF. These vertical members are
rigidly attached to the beam at B and C. (Draw shear and moment
diagrams for the beam A BCD only.) Ans ^ Max M= - 22 kN -m

442. Beam carrying the unformly varying load shown in Fig.

P-422. Arts. Max. M= wL 2 /9V3 at x = L/V3

Figure P-442.

443. Beam carrying the triangular loads shown in Fig. P-443.

Arts. Max. M
= wL 2 /\2

Figure P-443.

444. Beam loaded as shown in Fig. P-444.

w N/m "s s^ w N/m

i*_ L | L _J
* 2 2 *

Figure P-444.
4-4 Relations Between Load, Shear, and Moment 143

445. Beam loaded as shown in Fig. P-445.

Ans. Max. M= -80kN-m

80 kN/m
40 kN/m lm
2m I 2 m

Figure P-445.

446. Cantilever beam carrying the loads shown in Fig. P-446.

30 kN/m
20 kN

Figure P-446.

447. Beam loaded as shown in Fig. P-447.

Ans. Max. M= -80kN-m

20 kN

Figure P-447.
144 4 Shear and Moment in Beams

448. Beam carrying the loads shown in Fig. P-448.

Ans. Max. M= l37.5kN-m

20 kN/m 80 kN/m

Figure P-448.

449. A beam carrying the triangular load shown in Fig. P-449 is

supported on a uniformly distributed reaction.

Ans. Max. M — — 45kNm

"* O „-
O 111 " 9.
o m
m
-
60 kN/m

I I I I i i

*1 m+- *1 m-^
Figure P-449.

450. Beam loaded and supported as shown in Fig. P-450.

50 kN 50 kN

4 m
1 m 20 kN/m 1 m

Figure P-450.
4-4 Relations Between Load, Shear, and Moment 145

451. Beam loaded as shown in Fig. P-451.

Ans. Max. M= — 45kN-m

12 kN/m 12kN/m

452. Beam loaded as shown in Fig. P-452.

Ans. Max. M = 49.5 kN m •

12 kN/m 18 kN/m

Figure P-452.

453. A uniformly varying load is supported on two distributed

reactions as shown in Fig. P-453.

Figure P-453.

In the following problems, draw moment and load diagrams corre-

sponding to the given shear diagrams where values are given in kilonew-
tons. Specify values at all change of load positions and at all points of zero
shear.
146 4 Shear and Moment in Beams

454. Shear diagram as shown in Fig. P-454.

Ans. Max. M - 30kNm
2 m 2 m lm lm

L5 15

-5

Figure P-454.

455. Shear diagram as shown in Fig. P-455.

10

1 m 2 m
2m
-10

15

Figure P-455.

456. Shear diagram as shown in Fig. P-456.

Ans. Max. A/ = -120kN-m
-4 m- -2 m- ^2m-

10

<2 -80
Figure P-456.
4-5 Moving Loads 147

457. Shear diagram as shown in Fig. P-457.

Arts. M = 22.5 kN-m
0.5 m

458. Shear diagram as shown in Fig. P-458.

2nd-degree curve

lm lm
Figure P-458.

4-5 MOVING LOADS

A truck or other vehicle rolling across a beam or girder constitutes a

system of concentrated loads at fixed distances from one another. For
beams carrying only concentrated loads, the maximum bending mo-
ment occurs under one of the loads. Therefore, the problem here is to
determine the bending moment under each load when each load is in a
position to cause a maximum moment to occur under it. The largest of
these various values is the maximum moment that governs the design of
the beam.
In Figure 4-23, P v P2 P3 and P4
, , represent a system of loads at
fixed distances a, b, and c from one another; the loads move as a unit
across the simply supported beam with span L. Let us locate the
position of P2 when the bending moment under this load is maximum.
If we denote the resultant of the loads on the span by R and its position
from P2 by e, the value of the left reaction is

R =j-(L-e-x)
x
148 4 Shear and Moment in Beams

Pi P2 P, P.

,r-~i——rt~l
\j ^ ——32c? i

R,
L
\
Figure 4-23. Moving loads.

The bending moment under P 2 is then

[M - (1M) L ] M 2
= * (L - e - x)(x) - P x
a

To compute the value of x which will give maximum A/2 we , set

the derivative of A/2 with respect to x equal to zero:

from which

x ---- (4-7)
2 2

This value of x is independent of the number of loads to the left of P 2 ,

since the derivative of all terms of the form P a with respect to x will be
x

zero.
Equation (4-7) may be expressed in terms of the following rule:
The bending moment under a particular load is a maximum when the
center of the beam is midway between that load and the resultant of all
loads then on the span.With this rule we locate the position of each load
when moment at that load is a maximum and compute the value of
the
each such maximum moment.
The maximum shearing force occurs at, and is equal to, the
maximum reaction. The maximum reaction for a group of moving loads
on a span occurs either at the left reaction, when the leftmost
load is over that reaction, or at the right reaction, when the rightmost
load is over it. In other words, the maximum reaction is the reaction to
which the resultant load is nearest.

ILLUSTRATIVE PROBLEM

459. A truck and

combination having the axle loads
trailer
shown in Fig. 4-24a rolls across the simply supported span of 12 m.
Compute the maximum bending moment and the maximum shearing
force.
4-5 Moving Loads 149

15 kN 25 kN R = 90 kN 50 kN
I

3 m »+« 3 m »i

2m- 6m

L = 12m
*T
(a) Original loading

R = 90 kN

— — —+
I

m m m m
|

3.5 * < 2.5 *+« 2.5 3.5

2m 6 m
± i

(b) Position of loads for maximum moment at A

R = 90 kN
1.5 m
— 1.5 m I
,
m
I

4.5 m n-* »-| 4.5

-• 6 m
t2m—5JL i
R,

(c) Position of loads for maximum moment at B

R' = lbkN
I lm lm
— |

5 m *{+ ++« »
5 m-

2m 6 m
ti ±

(d) Position of loads for maximum moment at C

with only B and C on span
Figure 4-24. Moving loads.
150 4 Shear and Moment in Beams

Solution: The resultant of the three loads is R = 90 kN and is located

as shown in Fig. 4-24a. The position of the loads that will cause the
bending moment to be maximum under A is shown in Fig. 4-24b, in
accordance with the rule expressed by Eq. (4-7) that the center line of
the beam is midway between A and R. Taking moments about R 2 equal
to zero, we find /?, to be

[2A/* = 0]
2
12/?, = 90(3.5) /?, = 26.25 kN
whence the bending moment at A is

[M = (2M) L ] MA = 26.25(3.5) = 91.9 kN-m

We next consider Fig. 4-24c, where the loads are so located that
the center line of the beam is midway between B and R. Setting
moments about R 2 equal to zero, we find that the value of R {
for this
position of the loads is

[2MRi = 0] 12/?, = 90(4.5) R =

x
33.75 kN
whence the bending moment at B is

[M = (2M) L ] MB = 33.75(4.5) - 15(2) = 122 kN-m

If we now position the loads so that the center line of the beam is

midway between C and /?, in order to have the bending moment a

maximum under C, we find that load A comes off the span, which is
contrary to the assumption that all three loads are on the span. This
indicates the possibility of a maximum bending moment under C when
B and C are on the span.
only loads
When only loads B and C are on the span, their resultant is

R' = 75 kN at 2 m from C. This position of the loads to cause

maximum bending moment under C is shown in Fig. 4-24d, in which
the center line of the beam is midway between R' and C. Setting
moments about R equal x
to zero, we find R 2 for this condition to be
[2A/*, = 0] \2R 2 = 75(5) R2 = 31.25 kN
whence the bending moment at C is computed to be
[M - (ZM) R ] Mc = 31.25(5) = 156 kN-m
It is left as an exercise for the reader to show that the maximum

bending moments under A and B, when only loads A and B are on the
span are, respectively, 96.3 kN-m and 105 kN-m, and that with only
C on the span, the maximum moment occurs with C at midspan and
equals 150 kN-m.
A comparison of the above
results shows that the most dangerous
bending moment is 156 kN-m, occurring under C when only loads B
and C are on the span.
4-5 Moving Loads 151

Maximum Shearing Force. If all three loads are on the span, the
resultant load R is 3 m from R 2 when C is over R2 ; 5 m from R
it is x

when A is over R v Evidently the maximum reaction, and consequently

the maximum shearing force, is at R2 , since it is nearer the resultant
load. By setting moments about R equal x
to zero, the value of R2 is

found to be

[2MRi = 0] \2R 2 = 90(12 - 3) R2 = Max. V- 67.5 kN

We must also investigate the possibility of the maximum shearing
force occurring B and C are on the span. The
when only loads
maximum reaction in this case will be at R when B is over R and the l9 Y

resultant load R' = 75 kN is 4 m from R v Its value will be R^ =

yf (12
— 4) = 50 kN. The condition when only A and B are on the span
need not be checked, because their resultant load of 40 kN is less than
the reaction R 2 = 67.5 kN found above.

PROBLEMS

460. A truck with axle loads of 40 kN and 60 kN on a wheel base

of 5 m rolls across a 10-m span. Compute the maximum bending
moment and the maximum shearing force.
Arts. Max. M= 160 kN-m; Max. V - 80 kN
461.Repeat Problem 460 using axle loads of 30 kN and 50 kN
on a wheel base of 4 m crossing an 8-m span.
462. A tractor with axle loads of 4 kN and 8 kN has a wheel
base of 3 m. Compute the maximum moment and maximum shearing
force when crossing a 6-m span.
Arts. Max. M= 12.5 kN-m; Max. V= 10 kN
463. Three equal wheel loads of 30 kN each, separated by 2 m
between each load, roll as a unit across a 12-m span. Determine the
maximum moment and maximum shear.
Arts. Max. M = 210 kN-m; Max. V= 75 kN
464. Three wheel loads roll as a unit across a 16-m span. The
loads are A - 10 kN; B = 20 kN, 2 m
to the right of A; and C - 40
kN, 4 m to the right of B. Determine the maximum moment and
maximum shear in the simply supported span.

465. A truck and trailer combination crossing a 12-m span has

axle loads of 10 kN, 20 kN, and 30 kN separated respectively by
distances of 3 m and 5 m. Compute the maximum moment and
maximum shear developed in the span.

Ans. Max. M= 104 kN-m; Max. V = 45 kN

152 4 Shear and Moment in Beams

SUMMARY
The fundamental definitons of shear and bending moment are
expressed by

V=VY)l (4-1)
and
M= (ZA/) £ = (2M), (4-2)

in which upward-acting forces or loads cause positive effects. The

shearing force V should be computed only in terms of the forces to the
left of the section being considered; the bending moment may be M
computed in terms of the forces to either the left or the right of the
section, depending on which requires less arithmetical work.
Relations between load, shear, and moment are given by

and

V= ^
ax
(4-6)

These relations are amplified in Art. 4-4 to provide a seimgraphical

method of computing shear and moment which supplements Eqs. (4-1)
and (4-2). We obtain

V2 - V^lV-iare*)^ (4-3)

and
A/2 -M, = AM = (area)^ (4-4)

which provide alternate methods of computing shear and moment. The

relations (4-5) and (4-6), expressed in the form

intensity of load = corresponding slope of shear diagram

and
intensity of shear = corresponding slope of moment diagram
enable us to sketch the proper shapes of the shear and moment
diagrams rapidly and correctly.
When systems of wheel loads move as a unit across a beam, the
bending moment is a maximum under one of the loads. To determine
the position of the loads when the moment is maximum under a
particular load, the system of loads must be in such a position that the
center line of the span is midway betweenand the resultant of
that load
all on the span. With the loads in this position, the
the loads then
reactions are computed and Eq. (4-2) is applied to compute the bending
moment in the beam under the particular load.
 in Beams

5-1 INTRODUCTION

In this chapter we derive the relations between the bending moment and
the flexure stresses it causes, and between the vertical shear and the
shearing stresses. In deriving these relations, the following assumptions
are made:

1. Plane sections of the beam, originally plane, remain plane.

2. The material in the beam is homogeneous and obeys Hooke's
law.

3. The moduli of elasticity for tension and compression are

equal.

4. The beam is initially straight and of constant cross section.

5. The plane of loading must contain a principal axis of the

beam cross sectionand the loads must be perpendicular to the longitu-
dinal axis of the beam.

The application and limits of these assumptions will be discussed

in succeeding articles as the reason for them becomes apparent.

153
154 5 Stresses in Beams

5-2 DERIVATION OF FLEXURE FORMULA

The stresses caused by the bending moment are known as bending or

flexure stresses, and the relation between these stresses and the bending
moment is expressed by the flexure formula. The derivation of this
relation follows the procedure developed in deriving the torsion formula
(see Art. 3-2); that is, the elastic deformations plus Hooke's law
determine the manner of stress variation, after which the conditions of
equilibrium then establish the relation between stress and load.
Figure 5- la shows two adjacent sections, ab and cd, separated by
the distance dx. Because of the bending caused by load P, sections ab
and cd rotate relative to each other by the amount dO, as shown in Fig.
5- lb, but remain straight and undistorted in accordance with assump-
tion 1 of the preceding article.
is shortened, and fiber bd at the bottom is
Fiber ac at the top
lengthened. Somewhere between them is located fiber ef whose length is
unchanged. Drawing the line c'd' through / parallel to ab shows that
fiber ac is shortened an amount cc' and is in compression, and that fiber
bd is lengthened by an amount d'd and is in tension.
The plane containing fibers like ef is called the neutral surface
because such fibers remain unchanged in length and hence carry no
stress. It will be shown shortly that this neutral surface contains the

centroids of all transverse sections.

Consider now
the deformation of a typical fiber gh located y units
from the neutral surface. Its elongation hk is the arc of a circle of radius
y subtended by the angle dO and is given by
8 = hk - y dO
The strain is found by dividing the deformation by the original length ef

Neutral surface
dx
a : (

I) cI

«I
(a)

Figure 5-1. Deformations.

5-2 Derivation of Flexure Formula 155

of the fiber:

8 y dO

If we denote the radius of curvature of the neutral surface by p, the

curved length ef is equal to p d0\ whence the strain becomes

C = yd9 = y
pd9 p
Assuming that the material is homogeneous and obeys Hooke's law
(assumption 2), the stress in fiber gh is given by

a = £e = ( — \y (a)

Equation (a) indicates that the stress in any fiber varies directly with its

location y from the neutral surface, since assumed that the modulus
it is

of elasticity E is equal in tension and compression (assumption 3) and

the radius of curvature p of the neutral surface is independent of the
location y of the fiber. However, the stresses must not exceed the
proportional limit, for this would invalidate Hooke's law on which this
stress variation is based.
To complete the derivation of the flexure formula, we apply the
conditions of equilibrium. As we saw in Art. 4-3, the external loads that
act to one side of an exploratory section are balanced by the resisting
shear V r
M
and the resisting moment r To create this balance, a typical
.

element in the exploratory section is subjected to the forces shown in

the pictorial sketch* in Fig. 5-2. The line of intersection between the
neutral surface and the transverse exploratory section is called the
neutral axis, abbreviated NA.
To satisfy the conditions that the external loads have no X
components (assumption 5), we must have
[2* = 0] fax dA=0
where ox is equivalent to a in Eq. (a). On replacing ox by Ey/p, this
becomes

fydA=0
P
The constant ratio E/p is written outside the integral sign. Since y dA is

the moment of the differential area dA about the neutral axis, the
integral fy dA is the total moment of area. Hence

— Ay =
P
The cross section is drawn as rectangular only for ease of representation; it

may have any shape.

156 5 Stresses in Beams

Longitudinal plane
of bending loads

*<&
*?»
>S& &

o„dA

Figure 5-2. Forces acting on a typical element of the cross section of a beam.

However, since only y in this relation can be zero, we conclude that the
distance from the neutral axis (which is the reference axis) to the
centroid of the cross-sectional area must be zero; i.e., the neutral axis
must contain the centroid of the cross-sectional area.
The condition that 27 = 0, resulting in V= V r,
leads to the shear
stress formula, the derivation of which is postponed until later (Art.
5-7). It should be observed here that the resisting shear Vr is the
summation of the shearing forces r^ dA; that is, Vr = fr dA.
The condition 2Z = leads to frxz dA = 0. Since the loading has
no Z components, the system of shear forces txz dA must be self-balanc-
ing. We examine this in greater detail in Art. 13-8, where the plane of
loading may be offset from the XY plane but remains parallel to it. In
those cases, the loading causes a moment about the X axis which is
balanced by fy(rxz dA)-fz(T^ d4) in order to satisfy the condition
2MX = 0. This condition is automatically satisfied for sections that are
symmetrical about the Y axis because then the element under discussion
has a symmetrically placed counterpart so the integrals are equal to
zero. As a result, the plane of loading for sections symmetrical about the
Y axis must coincide with the XY plane, or the beam will twist.
We consider next the condition SAf^ = 0. The external loads have
no moment about the Y axis nor do the internal forces t w dA and
dA Therefore
r rT .

[2A/, = 0] (z(ox dA) =

5-2 Derivation of Flexure Formula 157

Again replacing ax by Ey/p, we have

— Jfzy dA =
p

The integral fzy dA is the product of inertia Pzy which is zero only if Y
,

or Z is an axis of symmetry or a principal axis. This is the justification

for assumption 5.

The final condition of equilibrium, 2Af2 =0, requires that the

bending moment be balanced by the resisting moment; that is, — r M M .

The resisting moment about the neutral axis of a typical element being
y(ax dA), this condition requires that

M = fy(a x dA)

which, by replacing ox by Ey/p from Eq. (a), becomes

M = —p Jfy 2
dA

2
Since fy dA is defined as /, the moment of inertia* of the area about a
reference axis, which here is the neutral axis (equivalent to the
centroidal axis), we finally obtain

M=— (b)
P

Observe now that it was necessary in Art. 4-2 to specify the centroidal
about which bending moment
axis of the exploratory section as the axis
is computed in order to obtain a common axis for computing and
equating Mand r M .

The usual form of writing Eq. (b) is

\ M , ,

which we shall use in Arts. 6-2 and 6-3 as the basis for determining
deflections in beams. Because curvature is equal to the reciprocal of the
radius of curvature, Eq. (5-1) indicates that curvature is directly propor-
tional to bending moment, an observation which we have already used
(page 116) in checking the sign of bending moment with the shape of
the deflected beam-positive curvature, which is concave upward, corre-
lating with positive bending moment, and vice versa.

*A complete discussion of moment of inertia is given in Appendix A.

 —

158 5 Stresses in Beams

Equating the ratio E/p from Eq. (5-1) with its value from Eq. (a),

we have

E- —
ply - = -

which leads directly to the flexure formula

«= & (5-2)

This formula indicates that the flexure stress in any section varies
directly with the distance of the section from the neutral axis. In a more
common form of the flexure formula yis replaced by the distance c,

which is defined as the distance from the neutral axis to the remotest
element. With this change, the maximum flexure stress in any section is

given by

Max. a = — (5-2a)

If I/ c is called the section modulus and denoted by S> another

common variation of the flexure formula is

Max.a = -^ = ^ (5-2b)

This variation is useful for beams of constant cross section, as it shows

that maximum flexure stress occurs at the section of maximum bending
moment. Various values of section modulus for common cross sections
are listed in Table 5-1.
An interesting analysis, similar to that we shall use later in
analyzing reinforced concrete beams (Art. 10-4), is to consider the
variation in flexure stress over a rectangular cross section, as shown in
Fig. 5-3.
Because the horizontal summation of forces over a section must be
zero, the total compressive force C in the upper half of the cross section
is equal to the total tensile force T in the lower half. Thus the resisting

M
moment r consists of the couple composed of the equal, oppositely
directed forces C and T. The value of each of these forces is equal to the
product of the average stress multiplied by the area. Therefore, since the
average stress in a linear stress distribution is one-half the maximum
stress,

7-=C=K ve .)(Area)
= ({a)(6|)

•Note the similarity between the torsion formula t - Tp/J and the flexure
formula o - My/ 1. This similarity makes them easy to remember.
5-2 Derivation of Flexure Formula 159

TABLE 5-1. Section Moduli

Rectangle Solid circle

3 3
bW 77T _ 7rd
S = S =

Tube Triangle

5= ^(R* - r*)

Area in

Area in tension

Figure 5-3. Resisting moment is equivalent to the couple created by the resultant
compressive and tensile forces.
1 60 5 Stresses in Beams

The forces C and T act through the centroid of the triangular load
distribution at a distance k from NA. Since k = §c = \(h/2\ the
moment arm of the resisting couple is e — 2k =\h. Equating bending
moment to resisting moment, we have
M = Mr = Ce = Te
2
bh
°-6

which agrees with Eq. (5 -2b) for a rectangular section.

Modulus of rupture

Equation (5 -2a) may be used to compute the flexure stress in a

beam loaded to rupture in a testing machine. Because the proportional
limit of the material is then exceeded, the stress determined in this

manner is not a true stress; nevertheless, the fictitious stress so obtained

is called the modulus of rupture. It is used to compare the ultimate
strengths of beams of various sizes and materials.

ILLUSTRATIVE PROBLEMS

501. A beam 150 mm

wide by 250 mm
deep supports the loads
shown in Fig. 5-4. Determine the maximum flexural stress.

Solution: We begin by computing the maximum bending moment.

The shear diagram shows that zero shear occurs at x = 2 m. Using the
area of this diagram to compute the bending moment, we have at
x = 2 m,
*
[AM = (area) K ] Max. M= [ j(2) = 16 kN m
It is unnecessary to draw the moment diagram.
We now apply the flexure formula, being careful to use consistent
units for the various quantities. From Table 5-1 we find that the section
modulus is S = bh 2 /6, so

M _6M Max. a =
6(16 X 103 )
S " bh 2 (0.150)(0.250)
2

= 10.24 MPa Ans.

502. A timber beam 100 mm wide by 300 mm high and 8 m long

carries the loading shown in Fig. 5-5. If the maximum flexural stress is

9 MPa, for what maximum value of w will the shear be zero under P,
and what is the value of PI
5-2 Derivation of Flexure Formula 161

15 kN
150 mm
2m | lm
6kN/m 250 mm

-* 3 m — *>

R x
= 14 kN R2 = 19 kN

Shear diagram

-19

Figure 5-4.

100 mm
6 m 2m V.

w N/m
300 mm

8m

Note: Any steeper slope, i.e., larger w,

shifts pt. of zero shear to the left. Any
smaller slope keeps pt. of zero shear at P.
1 62 5 Stresses in Beams

Solution: To satisfy the given conditions, the shear diagram must

appear as shown. The maximum value of w to reduce the shear to zero
P is determined from Eq. (4-3):

[AK = (area) load ] 4w + —p = 6w

which determines the following relation between P and w:

P = 8>v (a)

The maximum bending moment occurs under P and is

[AM = (area) K ] Max. M - £(6)(6w) = 18w N m•

Applying the flexure formula, we obtain

:

)'"
(0.100)(0.300)
M= o— =
c
a
6
18w = (9X 106
6
w = 750 N/m Ans
whence from relation (a), the value of P is

P = 8w = 8(750) = 6000 N ^/w.

PROBLEMS

503. A cantilever beam, 60 mm wide

by 200 mm high and 6 m
long, carries a load that varies uniformly from zero at the free end to
1000 N/m at the wall, (a) Compute the magnitude and location of the
maximum flexural stress, (b) Determine the type and magnitude of the
stress in a fiber 40 mm from the top of the beam at a section 3 m from
the free end. Ans. (b) o = 1.13 MPa
504. A simply supported beam, 60 mm wide by 100 mm high and
4 m long, is a point 1 m
subjected to a concentrated load of 800 N at
from one of the supports. Determine the maximum fiber stress and the
stress in a fiber located 10 mm
from the top of the beam at midspan.
505. A high-strength steel band saw, 20 mm wide by 0.80 mm
thick, runs over pulleys 600 mm in diameter. What maximum flexural
stress is developed? What minimum diameter pulleys can be used
without exceeding a flexural stress of 400 MPa? Assume E = 200 GPa.
Ans. o = 267 MPa
506. A flat steel bar, 25 wide by 6 mm
thick and 1 long, is mm m
bent by couples applied at the ends so that the midpoint deflection is
20 mm. Compute the maximum stress in the bar and the magnitude of
the couples. Use E= 200 GN/m 2
.

Ans. o = 95.8 MPa; M= 14.4 N m

•
5-2 Derivation of Flexure Formula 163

507. In a laboratory test of a beam loaded by end couples, the

fibers at layer AB in Fig. P-507 are found to increase 30 X 10" 3 mm
while those at CD decrease 90 X 10~ 3 mm in the 200-mm-gage length.
Using E= 100 GPa, determine the flexural stress in the top and bottom
fibers.

— 200 mm*
i
50 mmj
"T
100 mm^
_JL
75 mm;

Figure P-507.

508. Determine the minimum width b of the beam shown in Fig.

•508 if is not to exceed 10 MPa.
the flexural stress
Ans. b = 75.0 mm

5000 N

| lm
2000 N/m 200 mm
1ml
1
3m \

w b

R 1
R2
Figure P-508

509. A box section used in aircraft is constructed of tubes

connected by thin webs as shown in Fig. P-509. Each tube has a
cross-sectional area of 130
2
mm
If the average stress in the tubes is not
.

to exceed 70 MPa, determine the total uniformly distributed load that

can be supported on a simple span 4 m
long. Neglect the effects of the
webs.

100 mm-. 100 mm

<§>
9nn
200
—mm <§>
-.
©
150 mm

Figure P-509.
164 5 Stresses in Beams

510. A 40-mm diameter bar is used as a simply supported beam

2 m long. Determine the largest uniformly distributed load which can be
applied over the right half of the beam if the flexural stress is limited to
60 MN/m 2
. Arts, w = 1340 N/m
511. A simply supported rectangular beam, 50 mm wide by
100 mm deep, carries a uniformly distributed load of 1200 N/m over its
entire length. What is the maximum length of the beam if the flexural
stress is limited to 20 MPa?
512. The circular bar 20 mm in diameter shown in Fig. P-512 is

bent into a semicircle with a mean radius of 600 mm. If P= 2000 N

and F= 1000 N, compute the maximum flexural stress developed in
section a -a. Neglect the deformation of the bar.
Ans. o = 331 MPa

Figure P-512.

513. A rectangular steel beam, 50 mm wide by 80 mm deep, is

loaded as shown in Fig. P-513. Determine the magnitude and location

of the maximum flexural stress.

2kN
M = 5kN-m
1 m J
I 1 m y 1 m

Figure P-513.

514. The right-angled frame shown in Fig. P-514 carries a uni-

formly distributed loading equivalent to 200 N for each horizontal
projected meter of the frame; that is, the total load is 1000 N. Compute
the maximum flexural stress at section a-a if the cross section is 50 mm
square. Ans. o = 30.0 MPa
515. Repeat Problem 514 to find the maximum flexural stress at
section b-b.
5-2 Derivation of Flexure Formula 165

Figures P-514 and P-515.

516. A rectangular steel bar, 20 wide by 40 mm high and mm

4 m long, is simply supported at its ends. If the density of steel is
7850 kg/m3 determine the maximum bending stress caused by the
,

weight of the bar.

517. A simply supported beam 4 m long is composed of two
C230 X 30 channels riveted back to back. What uniformly distributed
load can be carried, in addition to the weight of the beam, without
exceeding a flexural stress of 140 MN/m2 if (a) the webs are vertical
and (b) the webs are horizontal. Refer to Appendix B for channel
properties.

518. A beam with a S3 80 X 74 section is simply supported at the

ends. It supports a central concentrated load of 40 kN and a uniformly
distributed load of 15 kN/m over its entire length, including the weight
of the beam. Determine the maximum length of the beam if the flexural
stress is not to exceed 140 MPa. Refer to Appendix B for properties of S
shapes.

519. 10 m long is simply supported 1 m from each end.

A beam
It is made C380 X 50 channels (see Appendix B) riveted back to
of two
back and used with the webs vertical. Determine the total uniformly
distributed load that can be carried along its entire length without
exceeding a flexural stress of 120 MPa. Ans. w = 22.0 kN/m
520. A beam with a W200 X 27 section (see Appendix B) is used
as a cantilever beam 6 m long. Find the maximum uniformly distributed
load which can be applied over the entire length of the beam, in
addition to the weight of the beam, if the flexural stress is not to exceed
140 MN/m2 .

521. Repeat Problem 520 using a 4-m cantilever beam with a

W250 X 67 section.
522. In the portion of a butt joint shown in Fig. P-522, a 28-mm
rivet secures 14-mm cover plates to a 20-mm main plate. Assuming that
the loads are uniformly distributed along the rivet, determine the
maximum bending stress in the rivet.
166 5 Stresses in Beams

90 kN

30 kN
Figure P-522.

523. A square timber beam used as a railroad tie is supported by

a uniformly distributed reaction and carries two uniformly distributed
loads each totaling 48 kN as shown in Fig. P-523. Determine the size of
the section if the maximum stress is limited to 8 MPa.

0.2 m 0.2 m
0.5 m-H 1 m •-0.5 m-H

TV=48 kN W=48kN
I
'TTTTTTTTTTrn
Figure P-523.
n
524. A wooden beam 150 mm wide by 300 mm deep loaded as is

8 MN/m find
2
shown in Fig. P-524. If the maximum flexural stress is ,

the maximum values of w and P that can be applied simultaneously.

Arts, w = 9kN/m;/> = 18 kN
525. In Problem 524, if the load on the overhang is 10 kN/m
and the overhang is x meters long, find the maximum values of P and x
that can be used simultaneously.

w N/m

3m 3 m 4 2m

R x
R.
"2
Figures P-524 and P-525

526. A rectangular beam, 120 mm wide by 400 mm deep, is

loaded as shown in Fig. P-526. If w = 3 kN/m, find P to cause a

maximum bending stress of 10 MPa. Ans. P— 32.5 kN
527. Solve Problem 526 if w - 6 kN/m.
5-3 Economic Sections 167

4 m f lm
w N/m

i 5 m -

#1 #2
Figures P-526 and P-527.

5-3 ECONOMIC SECTIONS

In a beam having a rectangular or circular cross section, the fibers near

the neutral axis are understressed compared with those at the top or
bottom. The fact that a large portion of the cross section is thus
understressed makes it inefficient for resisting flexure.
The flexure formula, M— ol/c, shows that if the area of a beam
of rectangular section (Fig. 5 -6a) could be rearranged so as to keep the
same overall depth buthave the shape shown in Fig. 5 -6b, the moment
of inertia would be greatly increased, resulting in a greater resisting
moment. Physically, the increase in resisting moment is due to more
fibers being located at a greater distance from the NA, for such fibers
carry a greater stress and have a larger moment arm about the NA to
resist the applied bending moment. However, the section in Fig. 5 -6b is
not practicable; the two parts of it would collapse together. It is
necessary to use some of the area to fix these parts in place relative to
each other, as in Fig. 5 -6c. We see later (page 197) that this web area
transmits practically all the vertical shear, and we shall learn how to
compute its dimensions.
Figure 5 -6c represents a wide-flange beam (referred to as a W
shape). This is one of the most efficient structural shapes manufactured

Ezzzzzzzzzzzza mzzzzzzzzm ZZZ7ZZD

A
y
(a) (b) (c) (d)
Wide flange I beam
(W Shape) (S Shape)

Figure 5-6.
168 5 Stresses in Beams

because it not only provides great flexural strength with minimum

weight of material but is when used as a column (see
highly efficient
Chapter 1). Another structural shape is the I beam (referred to as an S
1

shape) in Fig. 5-6d; it preceded the wide flange and because it is not as
efficient has been largely replaced by the wide flange beam. Properties
of both these sections are given in Appendix B. A beam of either type is

specified by stating its nominal depth in millimeters and its nominal

mass per unit length in W610 X
kilograms per meter. The designation
140, for example, indicates a wide-flangebeam with nominal depth 610
mm and nominal mass per unit length of 140 kg/m. The tables in
Appendix B indicate that the actual depth of this beam is 617 mm and
its theoretical mass per unit length is 140.1 kg/m.* The structural tables

give the dimensions and other properties of the cross-sectional area,

such as moment of inertia (/), section modulus (S), and radius of
gyration* (r) for each principal axis of the section.
In selecting a structural section to be used as a beam, it is obvious
that the resisting r
moment M — al/c = aS
must be equal to or greater
than the applied bending moment M. This may be expressed as

5 > — (5-3)

Equation (5-3) indicates that a beam must be selected whose section

modulus is equal to or greater than the ratio of bending moment to
allowable stress. Illustrative Problem 528 demonstrates the necessary
procedure and cautions.

Lateral deflection of beams.

The compression flanges of beams tend to buckle horizontally

sideways if the beam is too long. This buckling is a column effect.
(Columns will be discussed in Chapter 1 1 .) When this lateral deflection
is prevented by the floor system or by bracing the compression flanges
at proper intervals, the full allowable stresses may be used. Otherwise,
the stresses should be reduced. Formulas for the reduction of the
allowable stress are given in the specifications of the American Institute
of Steel Construction. Hereafter, we assume that all beams are properly
braced against lateral deflection.

•Many designs are based on the nominal mass per meter. However, to
illustratethe use of the tables, we shall use the theoretical mass per meter.
t
The use of r for radius of gyration conforms with American Institute of Steel
Construction (AISC) notation. Be careful not to confuse this term with the r which is
frequently used to denote the radius of a circle.
5-3 Economic Sections 169

ILLUSTRATIVE PROBLEM

528. What is the lightest W

shape beam that will support the
load shown in Fig. 5-7 without exceeding a flexural stress of 120 MPa?
Determine the actual stress in the beam. Assume the beam is properly
braced against lateral deflection.

45 kN

4m I 2m

f
R1 = 15 kN R2 = 30 kN

15 kN

Shear diagram
-30 kN
Figure 5-7. Live load. (Applied load exclusive of weight of beam.)

Solution: We begin by computing the reactions and sketching the

shear diagram. The maximum moment occurs under the load and
equals 15 X 4 = 60 kN • m. Applying Eq. (5-3), we have

60 X 103
S > 6
= 500 x 10" 6 m 3

120 x 10

> 500 x 10
3
mm 3

Referring to the table of properties of W shapes (Appendix B) and

starting at the bottom, first beam whose section
we find that the
modulus is X 103 mm3 is W200 X 52 with S = 512 X
greater than 500
10 mm In the W250 group we find a W250 X 45 with S = 534 X 103
3 3
.

mm which is also satisfactory, as well as being lighter. The W310

3
,

group lists a W310 X 39 beam with S = 549 X 103 mm 3 This one is the .

best one, because the lightest suitable beam in the remaining groups
have a mass per meter which is greater than 38.7 kg/m, the mass per
meter of the W310 X 39 beam.
The reader may wonder why more than one size beam is manufac-
tured with approximately the same section modulus. The explanation is
that although the lightest beam is the cheapest on the basis of weight
alone, frequently headroom clearances require a beam of less depth
than the lightest one.
1 70 5 Stresses in Beams

The selection of the beam is not complete until a check calculation

ismade that includes the weight of the beam.* The beam's resisting
M
moment B must be equal to or greater than the sum of the live load
M
moment L caused by the applied loads and the dead load moment D M
caused by the dead weight of the beam:

MB > ML + MD
Dividing each term of this equation by the stress a gives

o a o
whence, replacing M/ a by the section modulus S, we obtain the
governing equation for design:

The weight of the beam in this example is 38.7 kg/m X 9.81

m/s 2 = 380 N/m, which is not sufficient to change the location of the
maximum moment resulting from the combined live and dead loads.
Hence, we compute the dead load moment D at x = 4 m (Fig. 5-8). M
From the definition of bending moment, = (ZM) R we have M ,

MD (1.14)(2) (0.380 X 2)(1) 1.52 kN m

= - = •

or, from the shaded area of the shear diagram,

(1.14 + 0.38),^ ,^
MD = * - L
(2) = 1XT
1.52 kN m •

Therefore, the section modulus required to support the dead

weight of the beam is

S =—2-
M, L52x^ =127xl0 _ 6m3
Sd 6
o 120 x 10
= 12.7 X 10
3
mm 3

Applying the governing equation,

3 3
[SB > SL + SD ] 549 X 10 > (500 + 12.7) X 10

we see that a W310 x 39 beam is satisfactory.

Frequently, the steel beam is encased in concrete for fireproofing or to form

part of a concrete floor. The concrete may be assumed to extend at least 50 mm
beyond the dimensions of the beam. For example, a W310 X 39 with a depth of 310
mm and a flange width of 165 mm will be encased in concrete having overall
dimensions of 410 mm by 265 mm. Computed at a mass density of 2400 kg/m 3 the ,

encasement adds an extra load of approximately 2.56 kN/m, which should be

included in the weight of the beam. In subsequent problems, the weight of the
encasement will be neglected. In actual practice, its weight is usually included in an
estimate of the dead loads.
5-3 Economic Sections 171

0.380 kN/m

6 m
i2 2 = 1.14kN

"^
9 m
^-^_
-0.38 kN mi -1.14 kN
Figure 5-8. Dead load. (Due to weight of beam.)

The actual beam stress is easily determined from the ratio of the
beam modulus to the design modulus, viz.,

[M-o>SB -o(SL + SD )]
a'(549 X 10" 6 ) - (120 X 6
10 )[(500 + 12.7) X 10" 6
]

from which
(*'= 112 MPa Ans.

PROBLEMS

Assume that the beams in the following problems are properly

braced against lateral deflection. Be sure to include the weight of the
beam itself.

529. A 10-m beam simply supported at the ends carries a uni-

formly distributed load of 16 kN/m over its entire length. What is the
lightest W shape beam that will not exceed a flexural stress of 120 MPa?
What is the actual stress in the beam selected?
Ans. W610X 82; 113 MPa
530. Repeat Problem 529 if the distributed load is 12 kN/m and
the length of the beam is 8 m.
531. A concentrated load of 90 kN is applied at the center of a
simply supported beam 8 m long. Select the lightest suitable W shape
section using an allowable stress of 120 MN/m 2
. Ans. W530 X 74
532. Solve Problem 531 if the length of the beam is changed to
12 m.

533. A beam
simply supported on a 12-m span carries a uni-
formly distributed load of 30 kN/m over the middle 6 m. Using an
172 5 Stresses in Beams

allowable stress of 140 MPa, determine the lightest suitable W shape

beam. What is the actual maximum stress in the selected beam?
Ans. W610 x 125; 133 MPa
534. Repeat Problem 533 if the uniformly distributed load is

changed to 80 kN/m.
535. A simply supported beam 16 m long carries a uniformly
distributed load of 20 kN/m over the right half of the beam. Select the
lightest suitable W shape beam if the allowable stress is 120 MN/m 2
.

536. A simply supported beam 10 m long carries a uniformly

distributed load of 20 kN/m over its entire length and a concentrated
load of 40 kN atmidspan. If the allowable stress is 120 MPa, determine
the lightest W shape beam which can be used.
Ans. W610 X 125

5-4 FLOOR FRAMING

Probably the most common structural use of beams is to provide

support for the floors and frameworks of buildings. Figure 5-9
illustrates a typical detail in a home. The subfloor is supported by floor

Subfloor

Column

Figure 5-9. Floor framing.

5-4 Floor Framing 173

|
W=paL

w=paN/m

Figure 5-10. Loading on a floor joist.

beams in steel construction). The floor joists are

joists (called floor
assumed to act as simply supported beams. They are supported by
heavier beams called girders, which in turn are supported by columns
that transmit the loads to the foundation.
The floor load is specified as p N/m2 and varies from 2.5 kN/m2
for homes to as high as 25 kN/m for industrial buildings. If the floor
2

joists are L meters long and spaced a meters apart on centers, each joist
is assumed to support the loading of an area aL m
2
shown shaded in ,

Fig. 5-9. The loading on a typical floor joist therefore is the uniformly
distributed load shown in Fig. 5-10. The total load equals the load/? W
N/m2 acting over the area aL. This may be divided by the length L to
give a loading per meter of w = ^ N/m.
In steel construction, the same general plan is followed except that
the floor beams are usually riveted or bolted to the webs of the girders
as shown in Fig. 1-12 (page 21). The figures in the following illustrative
problem represent a small building and show how to construct the
loading diagrams for the various beams.

ILLUSTRATIVE PROBLEM

537. Determine the loading diagrams for beams B-l, G-l, B-2,
and G-2 for the building whose partial floor plan is shown in Fig. 5-11.
The loading for each bay is indicated.

Solution: Beams supporting only floor loads are designated B-\, B-2,
etc. Beams beams are called girders
that support the reactions of floor
and are denoted by G-l, G-2, At beam B-l, the loading is 5 kN/m2
etc.
unif ormiy distributed over a length of 4 m and a width of 2 m, resulting
in the loading diagram shown in Fig. 5-12.
Beam G-l is a girder used to support one end of beams B-\ and
framed into beams B-2. It is loaded by the end reactions of beams B-l,
as shown in Fig. 5-13.

Beam B-2 supports the end reaction of beam G-l, as well as half
the loadings in the bays adjacent to it. Its loading diagram therefore is
 —

174 5 Stresses in Beams

-*— 2 m 2m—+ 2 m + 2 m 2 m

(B-3) (B-3) (B-3) (B-3) (B-3) (B-3)

10 kN/m 2 10 kN/m 2 10 kN/m 2 10 kN/m 2 10 kN/m 2

Col. Col.

Tt
(G-2)

Stair well 2m

(G-l)

6 m
(B-2) (B-l) (B-l) (B-2)

10 kN/m 2 5 kN/m 2 5 kN/m 2 5 kN/m 2 10 kN/m 2

Figure 5-11. Floor plan and loading.

W= 5(2 X 4) = 40 kN
I

4 m
i?=20kN «=20kN
Figure 5-12. Beam (B-1).

20 kN 20 kN

2m 1 2m f 2m

C
K=20kN
6 m 3
R=20kN
Figure 5-13. Girder (G-1).
5-4 Floor Framing 175

20 kN

'

4 m
2 m
10 kN m 15 kN m

U 111

Figure 5-14. Beam (B-2).

60 kN 60 kN

2 m UJ 2 m

m
2 m

F 6
3
Figure 5-15. Girder (G-2).

as shown in Fig. 5-14. For the first 2 m, beam B-2 supports a total
load of 10 kN/m 2 over a floor area 2 m x 1 m, equivalent to 20 kN
applied at 10 kN/m. The reaction of beam G-l is shown as a con-
centrated load of 20 kN. For the rest of the beam, the loading is 15
kN/m computed as the sum of the loadings per meter extending for 1 m
into the10-kN/m 2 bay and the 5-kN/m2 bay.
The girder G-2 is loaded by the reactions of beams B-3 only, as
shown in Fig. 5-15. Verify that the reaction of beam B-3 is 60 kN.

PROBLEMS

538. Floor joists 50 mm

wide by 200 high, simply supported mm
2
on a 4-m span, carry a floor loaded at 5 kN/m Compute the center- .

line spacing between joists to develop a bending stress of 8 MPa. What

safe floor load could be carried on a center-line spacing of 0.40 m?
Ans. 0.267 m; 3.34 kN/m 2
539. Timbers 300 x 300 mm, spaced 0.90 m apart on centers, are
driven into the ground and act as cantilever beams to back up the sheet
piling of a coffer dam. What is the maximum safe height of water
3
behind the dam if the density of water is 1000 kg/m and the bending
MN/m ?
stress is limited to 8
2
Ans. h = 2.90 m
540. Timbers 200 mm wide by 300 mm deep and 5 m long,
simply supported at top and bottom, back up a dam restraining water
176 5 Stresses in Beams

3 m deep. The density of water is I000kg/m 3 . (a) Determine the center-

line spacing of the timbers if the flexural stress is 8 MPa. (b) Will this
spacing be safe if the maximum flexural stress is limited to 12 MPa and
the water reaches its maximum depth of 5 m? Ans. (a) 0.939 m
541. The 6-m long floor beams in a certain building are simply
supported at their ends and carry a floor load of 4 kN/m 2 . If the beams
have W250 X
45 sections, determine the center-line spacing using an
allowable flexural stress of 120 MPa.

542. Select the lightest W shape sections which can be used for
the beams and girders in Illustrative Problem 537 if the allowable stress
is 120 MPa. Neglect the weights of the members.
Ans. B-\: W250 X 18; B-2: W410 X 46;
CM: W310 x 28; G-2: W410 x 60
543. A portion of the floor plan of a certain building is shown in
Fig. P-543. The total loading (including live and dead loads) in each
bay is as shown. Select the lightest suitable W shape beams if the
allowable flexural stress is 120 MPa. Assume the beams are adequately
braced.

y^\N all

15 kN/m 2 15 kN/m 2 9 kN/m 2 9 kN/m 2 15 kN/m 2 15 kN/m 2

7 m
(fi-3) (B-2) (B-l) (B-2) (B-S)

(G-l)

t
Stair well 2 m
>

Figure P-543.

2
544. Repeat Problem 543 if the 15 kN/m loading is changed to
and the 9 kN/m loading is changed to 12 kN/m
2
24 kN/m 2 2
.

5-5 UNSYMMETRICAL BEAMS

All the beams discussed so far have been symmetrical with respect to
the neutral axis. Because flexure stresses vary directly with distance
5-5 Unsym metrical Beams 177

from the neutral axis — which is the centroidal axis — such beam sections
are desirable for materials that are equally strong in tension and
compression. However, for materials relatively weak in tension and
strong in compression, such as cast iron, it is desirable to use beams that
are unsymmetrical with respect to the neutral axis. With such a cross
section, the stronger fibers can be located at a greater distance from the
neutral axis than the weaker fibers. The ideal treatment for such
materials is to locate the centroidal or neutral axis in such a position
that the ratio of the distances from it and in
to the fibers in tension
compression is exactly thesame as the ratio of the allowable stresses in
tension and in compression. The allowable stresses thus reach their
permitted values simultaneously.

ILLUSTRATIVE PROBLEMS

545. A cast-iron beam carries a uniformly distributed load on a

simple span. Compute the flange width b of the inverted T section (Fig.
5-16) so that the allowable stresses ot = 30 MPa and oc = 90 MPa
reach their limits simultaneously.

» -*20 mn i

"7
wN/m
120 mm 1

NA
i I
i
T > '

Li

20 mm * X
)*

Figure 5-16.

Solution: The beam is bent concave upward so that the uppermost

fibers are in compression and the lowermost fibers are in tension. As

discussed in Art. 5-2, flexure stresses vary directly with their distance
from the neutral axis. Therefore to cause a, and ac to reach their limits
simultaneously, we have

yj_ yj_
30
yc 90

or

yc = 3y, (a)
178 5 Stresses in Beams

Figure 5- 16 shows that another relation between v, and>>c is

v, + yc = 140 mm (b)

Solving relations (a) and (6), we obtain

y = 35
t
mm and y c = 105 mm
Now consider the T section to consist of the two shaded rectan-
gles. Since the neutral axis coincides with the centroidal axis, we take
moments of areas with respect to an X axis through the base of the
flange and obtain
[Ay = Say]
(120 x 20 + b x 20)v, = (120 x 20)(20 + 60) + (ftx 20)(10)

In this is substituted the value y —

t
35, which gives
b = 216 mm Arts.

546. Compute the maximum tensile and compressive stresses

developed in the beam that is loaded and has the cross-sectional
properties shown in Fig. 5-17.

8kN

r
120 mm
NA
50 mm
i W777f/fr777777\
|*-x = 1.8m-H
I NA = 20 X 10 6 mm 4

Shear diagram

Solution: Sections of zero shear are at x = 1.8 m and x = 4 m. The

bending moments at these sections are A/ L8 = 16.2 kN-m and M 4
=
-8 kN-m. Check these values.
The positive moment at x = 1.8 m indicates curvature concave
upward (see Art. 4-2); hence the upper fibers are in compression and
the lower fibers are in tension. Applying Eq. (5-2) shows the flexure
5-5 Unsymmetrical Beams 179

stresses to be

= My
3
= (16.2 X 10 )(0.120)
a o„ = 97.2 MPa
I 20 X 10" 6

20 x 10" 6
Note that for the units to be consistent, is expressed in N M
m, y in m, •

and / in m4 .

At x = 4 m, the negative bending moment is interpreted as curva-

ture concave downward, so that the upper fibers are in tension and the
lower ones in compression. Having thus interpreted the sign of the
bending moment, we substitute the numerical value of the bending
moment in Eq. (5-2) and obtain the following flexure stresses:
3

a — My (8 x 10 )(0.120)
= 48.0 MPa
20 x 10~ 6

(8xl03)(0.050)
© ^--\-\,C>f^
=2()0Mpa
20 x 10" 6

Hence the maximum compressive stress is 97.2 MPa, occurring at

x — 1.8 m, and the maximum tensile stress is 48.0 MPa, occurring at
x = 4 m. In an unsymmetrical section having a reversal in curvature,
the maximum stresses need not both occur at the section of maximum
moment. The stresses at each section of zero shear must be investigated.

547. The overhanging beam in Fig. 5-18 is made of cast iron, for
which the allowable stresses are ot
= 40 MPa and ac = 100 MPa. If the

w N/m

-1.2m-4» 3 m- 4«-1.2m — NA
/J 80 mm
f
1.5 w' 180
1.2 m;

7 NA = 50 X 10 6 mm 4
-1.2 m;
1.5 m;

Shear diagram

Figure 5-18.
180 5 Stresses in Beams

properties of the cross section are as shown, determine the maximum

uniformly distributed load that can be supported.

Solution: At x = 1.2 m, the bending moment -0.72w N- m, the

is

negative sign indicating tension in the upper fibers. Using Eq. (5-2), we
find that the safe resisting moments in tension and compression are

6
(40 x 10 )(50x 10" 6 ) ,.- 1M
M.. -
= ~
O080
^ = 25.0kN. m
y
6 10- 6
(100 x 10 )(50 X
K ~_ O180
) _ 278 kN m
" '

Evidently tension governs, since the safe resisting moment is the lower
value. Equating this to the bending moment, we have

[M - Mr]
0.72w = 25.0 X 103 w = 34.7 kN/m
Before concluding that this is the safe load, we must also investi-
gate the other section of zero shear. At x = 2.7 m, M= + 0.405 w
N •
m. Although this is lower than the moment at x = 1.2 m, the
curvature is reversed, being concave upward and placing the upper
fibers in compression and the lower ones in tension. From Eq. (5-2), the
safe resisting moment is

M -± (100xlO-)(5 xlO-^ 625kN m

r Mc= U.UoU
,

y
6 1Q- 6
(40 x 10 )(50 x
M
,,
-
O180
)
=H.lkNm
Equating the lower resisting moment to the bending moment, we obtain

[M = Mr]
0.405 w = 11.1 X 10
3
w = 27.4 kN/m
The maximum safe load is the lower of the values obtained at
x = 1.2m and x = 2.7 m, that is, kN/m. Why is it unnecessary to
27.4
investigate the section of zero shear at x = 4.2 m? Show that inverting
the beam section will reduce the allowable load to 15.4 kN/m.

PROBLEMS

The inverted T section of a 4-m simply supported beam has

548.
the properties shown in Fig. P-548. The beam carries a uniformly
distributed load w over its entire length. Determine w if a, < 30
MN/m 2 and oc < 70 MN/m2 Ans. w = 3750 N/m
.
5-5 Unsymmetrical Beams 181

n
160 mm

I NA

80 mm

I NA = 20 X 10 6 mm 4
Figure P-548.

549
(549. Determine the maximum tensile and compressive bending
stresses developed in the beam shown in Fig. P-549. The cross section
has the given properties. Ans. o t
— 20.0 MPa; ac = 10.0 MPa

5kN
M = 5kN-m ZZZZZZZZZ2 60 mm
lm 2m 1 2m NA

pc
|

200 mm
1
#2 / N A=40 X 10 6 mm 4
Figure P-549.

550. maximum tensile and compressive

Find the flexure stresses
for the cantilever beam shown in Fig. P-550.

TZZZZZZA~^ mm
4kN/m
Wl NA
130

200 mm
6m
^^ 7 NA =
\ZZZ2
100X 10 6 mm 4
10 kN
Figure P-550.

551. A beam with the cross section shown in Fig. P-551 is

loaded in such a way that the maximum moments are +1.5Z
>
and N m
•

-2.2P N •
m, where P is the applied load in newtons. Determine the
maximum safe value of P if the working stresses are 30 MPa in tension
and 70 MPa in compression. Ans. P = 16.0 kN
182 5 Stresses in Beams

60
T
mm
J NA
r

100 mm

I
/ NA = 80 X 10 6 mm 4
Figure P-551.

552. Resolve Problem 551 if the maximum moments are +3.2P

N- m and -5.8P N • m.
553. Determine the maximum safe value of W that can be
carried by the beam shown in Fig. P-553 if a, < 20 MN/m 2
and
ac < 60 MN/m 2
. Arts. W= 3 kN

u w
200 mm
6W NA
I 80 mm
2m 8m 2m YZZZZZZZA—.\
7 NA = 60 X 10
6
mm 4

Figure P-553.

554. What safe value of W can be applied to the beam loaded as

shown in Fig. P-554 if a, < 60 MPa and ac < 100 MPa?

W
^80 mm
2W 4 m 4 m
2m NA
120 mm
*1
.A
7 NA =30 X 10
6
mm 4
Figure P-554.

555. A cast-iron beam carries the loads shown in Fig. P-555. If

a, < 20 MN/m 2
and oc < 80 MN/m 2
, compute the permissible limits
of the overhang. Ans. x = 2.0 to 2.5 m
5-5 Unsymmetrical Beams 183

4kN 4kN
80 mm
16 kN NA
^=1-:
200 mm
6 m
T 6
mm 4
NA = 40 X 10
/ma
Figure P-555.

556. AT beam supports the three concentrated loads shown in

Fig. P-556. Show that the NA is 70 mm from the top and that
7NA = 15.52 X 106 mm 4
. Then use these values to determine the maxi-
mum value of P so that a, < 30 MPa and oc < 70 MPa.
Ans. P= 1.41 kN

80 mm

20 mm

I 2m \ 2m \

1mA 160 mm
r
«i
VA
20 mm
Figure P-556.

557. A cast-iron beam 10 m long and supported as shown in Fig.

P-557 carries a uniformly distributed load of w N/m (including its own
weight). The allowable stresses are a, < 20 MN/m2 and a < 80 c
MN/m2 . Determine the maximum safe value of w = m.
if x 1

Ans. w = 1.92 kN/m

558. In Problem 557, find the values of x and w so that w is a
maximum. Ans. x = 1.59 m; w = 3.16 kN/m

10 m T
wN/m / 180 mm
y
NA V,

t T V7/777\3P mm
NA = 36 X 10
(

/ma
Figures P-557 and P-558.
 —
184 5 Stresses in Beams

5-6 ANALYSIS OF FLEXURE ACTION

If a beam were composed of many thin layers placed on each other,

bending would produce the effect shown in Fig. 5-19. The separate
layers would slide past each other and the total strength of the beam
would be the sum of the strengths of the various layers. Such a built-up
beam would be considerably weaker than a solid beam of equivalent
dimensions. For a demonstration of this, flex a deck of playing cards
between the fingers holding them rather loosely so that the cards can
slide past one another as they are bent. Then grip the ends of them
tightly, so that they cannot slip —
thus approximating a solid section
and try to flex them. You will discover that considerably more effort is

required.

Figure 5-19. Sliding between layers of a built-up beam.

Figure 5 -20a will aid in understanding this action. The figure is a

pictorial representation of the flexural stress distribution over the por-
tion to the left of the exploratory section m-n of the solid beam in Fig.
5-20b.

L
\n

(b)

(a)

Figure 5-20. Distribution of compressive and tensile forces.

5-6 Analysis of Flexure Action 185

If we add the horizontal forces acting over the entire depth of the
section, the compressive forces will exactly balance the tensile forces, as
is required by the equilibrium condition2^ = (Art. 5-2). However, if
we summation of horizontal forces over a partial depth of the
take a
section, say from the top elements a-b to those at c-d, the total
compressive force C, over the area abed (equal to the area abed
multiplied by the average of the stresses oa and od ) can be balanced only
by a shear resistance developed at the horizontal layer dee. Such shear
resistance is available in a solid beam but not in a built-up beam of
unconnected layers.
If we extend summation of horizontal forces down to layer fg,
the
is increased by C
the resultant compressive force 2 which is the average
,

of the stresses ad and o multiplied by the area cdfg. Thus a larger shear
f
resistance must be developed over the horizontal layer at fg than at dee.
Of course, the total compressive force C plus C2 acting over the area
x

abgf may also be computed as the average of the stresses oa and oy

multiplied by the area abgf. However, the first method indicates the
decreasing magnitude of the increase in the total compressive force as
we descend by equal intervals from the top; i.e., although the total
compressive force increases as we descend by equal intervals from the
top, itdoes so by smaller increments.
This analysis shows that the maximum unbalanced horizontal
force exists at the neutral axis. This unbalanced force decreases gradu-
ally to zero as the effects of layers below the neutral axis are included.
This is so because the horizontal effect of the compressive forces is

increasingly offset by the neutralizing effect of the tensile forces, until

finally complete balance is attained and 2X = over the entire section.
This analysis also indicates that layers equidistant from the neutral
axis, such as fg and hk, are subject to the same net horizontal unbal-
ance, because in adding the horizontal forces from the top to these
layers the equal compressive forces C3 and T3 cancel out. We conclude
that equal shear resistances are developed at layers fg and hk. However,
this requires that the areas from the neutral axis to the equidistant layers
be symmetrical with respect to the neutral axis. The conclusion would
not hold, for example, if the beam section were a triangle with its base
horizontal.

PROBLEMS

559. composed of 6 planks, each 100 mm wide and 20

A beam is
mm thick, piled loosely on each other to an overall dimension of 100
mm wide by 120 mm high, (a) Compare the strength of such a beam
with that of a solid beam of equal overall dimensions, (b) What would
186 5 Stresses in Beams

be the ratio if the built-up beam consisted of 12 planks each 100 mm

wide by 10 mm thick? Arts, (a) 1 to 6; (b) 1 to 12

560. The wide-flange beam shown in Fig. P-560 is strengthened

by riveting two cover plates 160 mm by 20 mm to the top and bottom
flanges. If the maximum flexure stress is 110 MPa, compute the total
force (a) each cover plate and (b) in each flange. Neglect the
in
weakening effect of the rivet holes. Ans. (a) 336 kN; (b) 304 kN

M— lbU mm — »-|

-H-
20 mm
20 mm

360 mm

H- n- 20 mm
-^r 20 mm

Figure P-560.

561. AT section has the dimensions given in Fig. P-561. Show

that the neutral axis is 60 mm below the top and that /NA = 26.67 X 106
mm 4 . If the tensile stress at the bottom of the flange is 10 MN/m2 ,

determine (a) the total tensile force in the flange and (b) the total

120 mm
T mm
40

160 mm
20 mm

Figure P-561.
5-6 Analysis of Flexure Action 187

compressive force in the cross section. Also determine (c) the moment
of the total compressive force and (d) the moment of the total tensile
force about the NA. (e) How does the sum of (c) and (d) compare with
the total applied bending moment as computed from the flexure for-
mula?
Arts, (a) 96.0 kN; (b) 98.0 kN; (c) 9.15 kN m;
• (d) 4.19 kN •
m
562. In any beam section having a maximum stress a, show that
the force on any partial shaded area A' in Fig. P-562 is given by
=
F (o/c)A'y' where y' is the centroidal coordinate of A' Also show
y
.

that the moment of this force about the NA is

where /' is the moment of inertia of the shaded area about the NA.

Figure P-562.

563. A box beam is made from 50-mm by 150-mm pieces

screwed together as shown in Fig. P-563. If the maximum flexure stress
is 8 MPa, determine the force acting on the shaded piece and the

moment of this force about the NA. (Hint: Use the results of Problem
562.) Ans. F = 45.0 kN; = 3.50 kN M • m

7//////////A
•tu W//////A /
jT '

4AA <W ttrf

Figures P-563 and P-564.

564. Resolve Problem 563 using one of the vertical pieces of

wood instead of the shaded piece.
188 5 Stresses in Beams

5-7 DERIVATION OF FORMULA FOR HORIZONTAL

SHEARING STRESS

Consider two adjacent sections, (l) and (2), in a beam separated by the
distance dx, as shown in Fig. 5-21, and let the shaded part between
them be isolated as a free body. Figure 5-22 is a pictorial representation
of this part, the beam from which it is taken being shown in dashed
outline.
Assume the bending moment at section (2) to be larger than that
at section (1), thus causing larger flexural stresses on section (2) than on
section (1). Therefore, the resultant horizontal thrust H
2 caused by the
compressive forces on section (2) will be greater than the resultant
horizontal thrust H
on section (1). This difference between 2 and
l
H H x

can be balanced only by the resisting shear force dF acting on the

bottom face of the free body, since no external force acts on the top or
side faces of the free body.
Since H2 — H is the summation of the differences in thrusts
l

o2 dA and o dA on the ends of all elements contained in the part shown

x

Section (1)

Figure 5-21

0\dA

2 dA

Section
A
(1)
Section (2)

r*^
<
Figure 5-22.
5-7 Derivation of Formula for Horizontal Shearing Stress 189

in Fig. 5-22, a horizontal summation of forces gives

[2/7 = 0] dF=H -H 2 x

= I o 2 dA — I o dA
x
Jy\ '
Jy\

whence, replacing the flexural stress a by its equivalent My/I, we

obtain

dF = -+[ydA--±fydA
J 1 J J
= —?- l
- f y dA
J
y\ y\ * y\

From Fig. 5-21 we note that dF = rb dx, where t is the average

shearing stress over the differential area of width b and length dx; also
that M 2
— M x
represents the differential change in bending moment dM
in the distance dx; hence the above relation is rewritten as

T = dM CydA
Ibdx J
y\

From Art. 4-4 we recall that dM/dx~ V, the vertical shear; so we

obtain for the horizontal shearing stress,
<NC .

We have replaced the integral t y dA, which means the sum of the

moments of the differential areas dA about the neutral axis, by its

equivalent A'y, where A' is the partial area of the section above the
layer at which the shearing stress is being computed and y is the
moment arm A' is
of this area with respect to the neutral axis;
the shaded area in the end view of Fig. 5-21. A variation of the product
A'y is the symbol Q, which frequently is used to represent the static
moment of area.

Shear flow

If the shearing stress t is multiplied by the width b, we obtain a

quantity q, known as shear flow, which represents the longitudinal force
per unit length transmitted across the section at the level y v It is
analogous to the shear flow discussed previously in the torsion of
thin- walled tubes (see page 91). Using Eq. (5-4), we find that its value is
given by

q = rb = jQ (5-4a)
190 5 Stresses in Beams

One application of this relation is discussed in Art. 5-9; another is given

in Illustrative Problem 1321 (page 541).

Relation between horizontal and

vertical shearing stresses

Most students are surprised to find the term vertical shear (V)
appearing in the stress (th ). However, as
formula for horizontal shearing
we show presently, a horizontal shearing stress is always accom-
shall
panied by an equal vertical shearing stress. It is this vertical shearing
stress tc shown in Fig. 5-23, that forms the resisting vertical shear
,

Vr — /t dA which balances the vertical shear V. Since it is not feasible

to determine tv directly, we have resorted to deriving the numerically
equal value of rh .

To prove the equivalence of rh and tv consider

, their effect on a
free-body diagram of a typical element in Fig. 5-23. A pictorial view of
this element is shown in Fig. 5-24a; a front view, in Fig. 5-24b. For
equilibrium of this element, the shearing stress rh on the bottom face
requires an equal balancing shearing stress on the top face. The forces
causing these shearing stresses (Fig. 5 -23c) form a counterclockwise
couple, which requires a clockwise couple to insure balance. The forces
of this clockwise couple induce the shearing stresses rv on the vertical
faces of the element as shown.
By taking moments about an axis through A (Fig. 5 -24c), we
obtain

[2MA - 0] (rh dx dz)dy - (rv ay dz)dx =

Figure 5-23. Horizontal and vertical shearing stresses.

5-7 Derivation of Formula for Horizontal Shearing Stress 191

1
dy

> r r v dydz

h
T h dx dz

(a) (b) Stresses (c) Forces

Figure 5-24. Shearing stresses on a typical element.

from which the constant product dx dy dz is canceled to yield

r* = t„ (5-5)

We conclude therefore that a shearing stress acting on one face of an

element is always accompanied by a numerically equal shearing stress
acting on a perpendicular face.

Application to rectangular section

The distribution of shearing stresses in a rectangular section can

be obtained by applying Eq. (5-4) to Fig. 5-25. For a layer at a
distance^ from the neutral axis, we have
V At _ V\ Jh \ir \ I h \

which reduces to
2
VI h
-ffi-4
This shows that the shearing stress is distributed parabolically across the
depth of the section.

NA

Figure 5-25. Shearing stress is distributed parabolically across a rectangular

section.
 1 92 5 Stresses in Beams

—^» The maximum shearing stress occurs at the neutral axis and is

found by substituting the dimensions of the rectangle in Eq. (5-4), as

follows:

_ V ry- = V bh h \
7 ( \(
lb (bh'/\2)b\ 2 )\4)

which reduces to

Max T= (5 ~ 6)
-
!iri7
* This indicates that the maximum shearing stress in a rectangular section
is 50% greater than the average shear stress.

Assumptions and limitations of formula

We have assumed, without saying so implicitly, that the shearing

stress is uniform across the width of the cross section. Although this
assumption does not hold rigorously, it is sufficiently accurate for
sections in which the flexure forces are evenly distributed over a
horizontal layer.
This condition is present in a rectangular section and in the
wide-flange section shown in Fig. 5 -26a, where the flexure forces on the
both shaded and unshaded, are evenly distributed across
vertical strips,
any horizontal layer. But this condition does not exist in the triangular
section in Fig. 5 -26b, where the shearing stress is maximum at the left
edge of the neutral axis, diminishing to zero at the right edge. Even here,
however, Eq. (5-4) can be used to compute the average value of
shearing stress across any layer. Another exception is a circular cross
section (Fig. 5-26c). It can be shown that the stress at the edge of any
layer must be tangent to the surface, as in the right half of the figure;
but the direction of shearing stresses at interior points is unknown,
although they are assumed to pass through a common center C as
shown. The vertical components of these shearing stresses are usually
assumed to be uniform across any layer, as in the left half of the figure,
and are computed by means of Eq. (5-4). With this assumption, the
maximum shearing stress across the neutral axis is \{P/irr 2 ). A more
elaborate study* shows that shearing stress actually varies at the neutral
2
axis from \.liP/irr 2 at the edges to \.3SP/irr at the center.

•See S. Timoshenko and J. N. Goodier, Theory of Elasticity, 2nd ed.,

McGraw-Hill, New York, 1951, p. 321.
5-7 Derivation of Formula for Horizontal Shearing Stress 193

NA

(a) (b) (c)

Figure 5-26.

ILLUSTRATIVE PROBLEMS

565. A simply supported beam 120 mm wide, 180 mm deep, and

6 m long carries a uniformly distributed load of 4 kN/m, as shown in
Fig. 5-27. (a) Compute the shearing stress developed at horizontal
layers 30 mm apart from top to bottom for a section 1 m from the left

end. (b) Compute the maximum shearing stress developed in the beam.

Solution:

Part a. As shown on
the shear diagram (Fig. 5 -27a), the defini-
tion of vertical shear V=
(2 Y) L gives V = 8 kN at x = 1 m.
The moment of inertia about the neutral axis is

bh 3 1
120(180)
/ = ^NA
~~ = 58.32 x 10
6
mm4
12 12

= 58.32 X -6
10"° _4
m
Applying Eq. (5-4) to a layer 30 mm from the top (Fig. 5 -27b),
we find that the shearing stress is

r - %A>y
8000
T = _6 (0.120 X 0.030)(0.075)
(58.32 X 10 )(0.120)
= 309 kPa

Note that for consistency of units, the shear force is expressed in

newtons, the moment of inertia in m 4 and the distances in m.
194 5 Stresses in Beams

At 60 mm from the top (Fig. 5 -27c), the shearing stress is

V
r = -A'y
8000
T = (0.120 x 0.060)(0.060)
(58.32 x 10" 6 )(0.120)
= 494 kPa
The shearing 60
stress at mm from the top can also be computed
from Fig. 5-27d, in which the area A' is resolved into two strips 30 mm
thick. Since a moment of area equals the sum of the moments of area of
its parts (that is, A'y = Soy), an identical result is obtained as follows:

T = |2>
8000
T = [(0.120 x 0.030)(0.075)
(58.32 x 10~ 6 )(0.120)
-I- (0.120 x 0.030)(0.045)]
- 494 kPa
 —
5-7 Derivation of Formula for Horizontal Shearing Stress 195

Although this computation is admittedly more complex than the preced-

ing one, it indicates the procedure to be followed when the area A' is

more complex, as in the case of a wide-flange beam.

At the neutral axis, or at 90 mm from the top (Fig. 5-27e), the
shearing stress is

= Tb A '>

8000
T = (0.120 x 0.090)(0.045)
(58.32 X 10- 6 )(0.120)
= 555 kPa

If desired, Eq. (5-6) may be used. As noted on page 192, this equation
determines the maximum shearing stress on any rectangular section.

_ 3 8000
7
2 bh
T " 2 (0.120)(0.180)
" 5 ^ kFa
The shearing stress at the 120-mm layer and the 150-mm layer are
determined similarly to be 494 and 309 kPa, respectively.
Note that equal values of r are obtained at layers equidistant from
the NA
in any beam symmetrical about the neutral axis. Physically, this
is was said on page 185, the compressive and tensile
true because, as
between these layers cancel each other. Analytically it is
flexure forces
true because the neutral axis is the centroidal axis, and hence the
moment of area A'y computed for a partial area A' located above the
NA equals that for a symmetrically placed area below the NA. Further,
since the total moment of area is zero with respect to a centroidal axis, it

follows that the moment of area about the NA of the area above any
layer equals that of the area below that layer. Stated differently, in
computing A'y we may use either the area above or that below any
layer, depending u^ jn which is easier to use.

Part b. The maximum shearing stress occurs at the NA of the

section of maximum shear. The shear diagram shows that maximum
shear occurs at either end, and hence from Eq. (5-6) the maximum
shearing stress is

T
3
= T ~r H *„
Max. r = — ——
3 12 x 10
3

^ <f> „. = „ 1T^
833 kPa Arts.
2 A J
2 (0.120 X 0.180)

566. A beam has the wide-flange section shown in Fig. 5-28a. At

a section where the vertical shear is V= 70 kN, compute (a) the
maximum shearing stress and (b) the shearing stress at the junction of
the flangeand the web. (c) Plot the shearing stress distribution in the
web, and determine the percentage of shear carried by the web alone.
 H

196 5 Stresses in Beams

h< — 160 mm —
n_j
•17.4 MPa
i
20 mm f y///////A 20 mm
W^ t
100 mm
NA ^
200 mm

-*20 mm
2.60 MPa-*J
20 mm -*-13.5 MPa^-]
7"ave.

(a) (b) (c)

Figure 5-28.

Solution: The moment of inertia is found by resolving the section into

a large rectangle from which is subtracted the two shaded rectangles.
We obtain

/-2 12
:

160(240) 70(200);
~~
^NA 12 12

- 91.0 X 10
6
mm 4
= 91.0 x 10" 6 m4
The maximum shearing stress occurs at the neutral axis. In apply-
ing Eq. (5-4),compute A'y as the sum of the moments of area of the
rectangles shaded in Fig. 5-28b.

r = Ay
Tb
3
= 70 X 10
Max. t
(91.0 X 10~ 6 )(0.020)
x [(0.160 x 0.020)(0.110) + (0.020 x 0.1 00) (0.050)]

- 17.4 MPa
At the junction of the web and flange, there is a discontinuity in
the shearing stress because the width b — 160 mm when computing the
shearing stress in the flange whereas b — 20 mm when considering
stresses in the web. Then, at the junction, the shearing stress in the
web is

3
V 70 x 10
T = Ay
a>- T = (0.160 X0.020)(0.110)
Tb (91.0 x 10~ 6 )(0.020)
= 13.5 MPa
 .

5-7 Derivation of Formula for Horizontal Shearing Stress 197

Show that the shearing stress at the junction in the flange is 1.69 MPa.
The shearing stresses in the web vary parabolically from top to
bottom, as shown in Fig. 5-28c. The average height of the parabolic
segment is f (17.4 — 13.5) = 2.60 MPa. Therefore the average shear
stress in the web is

The shearing force in the web is

[P = Aa &ve ] Kweb = (0.200 X 0.020)(16.1 X 10

6
) = 64.4 kN
whence the percentage of shear carried by the web alone is

64 4
%Vwcb = 5-g- x 100 = 92.0%

This shows that the flanges are almost i neffective in resisting the vertical
shear. If it is assumed that the total vertical shear is carried by the web
alone, the average shearing stress will be very close to the maximum
stress as computed from Eq. (5-4). Thus

T = T = 70Xl(fi
=17.5 MPa <3NWeS
^web (0.200)(0.020) Ovi^S

This isvery close to the computed maximum, 17.4 MPa.

This method gives results that closely approximate the actual
maximum r. In most design specifications or codes, however, the height
of the web is not taken as the distance between flanges but is assumed
to be the depth of the beam. This procedure is not as accurate as
total
the method just given, but lower allowable shearing stresses are usually
specified in order to compensate.

PROBLEMS

567. A timber beam 90 mm wide by 160 mm high is subjected to

a vertical shear V= 20 kN. Determine the shearing stress developed at
layers 20 mm apart from top to bottom of the section.

568. Show that the shearing stress developed at the neutral axis
of a beam with circular cross section is r =\(V/irr 2). Assume that the
shearing stress is uniformly distributed across the neutral axis.

569. Show that the maximum shearing stress in a beam having a

thin- walled tubular section of net area A is t = 2 V/ A
570. A beam 4 m long has the cross section
simply supported
shown in Fig. P-570. Determine the maximum uniformly distributed
load which can be applied over the entire length of the beam if the
shearing stress is limited to 1.2 MPa. A ns. w = 4.60 kN/m
198 5 Stresses in Beams

150 mm-*-

Flgure P-570.

571. The T section shown in Fig. P-571 is the cross section of a

beam formed by joining two rectangular pieces of wood together. The
beam is subjected to a maximum shearing force of 60 kN. Show that the
NA is 34 mm from the top and that 7NA = 10.57 x 106 mm4 Using .

these values, determine the shearing stress (a) at the neutral axis and
(b) at the junction between the two pieces of wood.
Am. (a) 3.28 MPa; (b) 3.18 MPa, 31.8 MPa

40 mm

100 mm

20 mm
Figure P-571.

572. if P=
kN, compute the shearing stress at
In Fig. P-572, 5
horizontal layers 20 mm apart from bottom to top at the section of
maximum V. The NA is 70 mm from the top and 7NA = 15.52 X
10
6
mm 4 .
5-7 Derivation of Formula for Horizontal Shearing Stress 199

80 mm

^
I*
1
20 mm

I 2m \ 2m I

1 mi 160 mm

*1
f
\lA

20 mm
Figure P-572.

573. The cross section of a beam is an isosceles triangle with

vertex uppermost, of altitude h and base b. If V is the vertical shear,
show that the maximum shearing stress is 3V/bh located at the mid-
point of the altitude.

574. In the beam section shown in Fig. P-574, prove that the
maximum horizontal shearing stress occurs at a layer h/ 8 above or
below the NA.

Figure P-574.

575. Determine the maximum and minimum shearing stress in

the web of the wide-flange section in Fig. P-575
100 kN. Also if V=
compute the percentage of vertical shear carried only by the web of the
beam. Arts. Max. t = 30.5 MPa; Min. t = 23.5 MPa; 90.2%
576. Rework Problem 575 assuming that the web is 200 mm
instead of 160 mm.
200 5 Stresses in Beams

|-t— 120 mm->j ,

§ 160 mm
20
iimm
20 mm
*
Figures P-575 and P-576.

577. A plywood beam is built up of 6-mm strips separated by

blocks as shown in Fig. P-577. What shearing force V will cause a
maximum shearing stress of 1.4 MPa?

50 50
mm mm

s
1t
6 20

|
y

Yy 20
'/, mm
V A
T
Figure P-577.

5-8 DESIGN FOR FLEXURE AND SHEAR

In this article we consider the determination of load capacity or the size

of beam and
section that will satisfy allowable stresses in both flexure
shear. No beyond those already developed.
principles are required
In heavily loaded short beams the design is usually governed by
the shearing stress (which varies with V); but in longer beams the
bending moment varies with
flexure stress generally governs because the
both load and length of beam. Shearing is more important in timber
beams than in steel beams because of the low shearing strength of
wood.
5-8 Design for Flexure and Shear 201

ILLUSTRATIVE PROBLEMS

578. A rectangular beam carries a distributed load of w N/m on

a simply supported span of L meters. Determine the critical length at
which the shearing stress t and the flexure stress a reach their allowable
values simultaneously.

Solution: As shown in Fig. 5-29, max. V = W/2, where Wis the

total distributed load. The maximum load as limited by the allowable

shearing stress is determined from Eq. (5-6):
3 V 3 W/2
Max. T«7T7
2 bh
T = 2* bh
W = -bhr
Note that W is independent of the length.
At the point of zero shear, the maximum bending moment, com-
puted from the area of the shear diagram, is

»-um)-™«~
Substituting this value in the flexure formula, Eq. (5 -2a), we obtain

M=— a£
= obh 2 ] WL obh
2

c 6 8 6

Replacing W by its value in terms of the shear stress, we have

2
obh
(Mj)-
W=wL
I

wN/m

Max. V=t
Max. M - -y
Cross -section

Flgure 5-29.
 202 5 Stresses in Beams

which reduces to

For values larger than this critical length, flexure governs the design; for
shorter values, shear governs.

579. A box beam supports the loads shown in Fig. 5-30. Com-
pute the maximum value of P that will not exceed a flexural stress a =
8 MPa or a shearing stress r = 1.2 MPa for sections between the
supports.

Solution: We start by computing / for the net section, which is the

difference between two rectangles. Hence
3
bh 160(200)' 120(160)-
/-2 12
=
" 12 12
= 65.7 X 6
10 mm4
= 65.7 x 10~ 6 m4
Determining the reactions from statics gives the shear diagram
shown in Fig. 5-30. In terms of P, the maximum V is — (\P + 2000). If
the area of- the cross section above the NA, where r is a maximum, is

resolved into the three rectangles shown, the static moment of the area,
Q, is

[Q - Say] Q= (160 X 20)(90) + 2(80 X 20)(40)

= 416 x 10
3
mm 3
= 416 x 10" 6 m 3
If desired, the area could also be resolved into the difference

between the outer 100-mm by 160-mm rectangle and the inner 80-mm
by 120-mm rectangle. This gives the same value of Q, viz.,

[ Q = lay ] Q = (160 X 100)(50) - (120 X 80)(40)

= 416 X 10
3
mm 3

4000 N

2m f 2 m 2 m 160 mm- 160 mm

J
20 mm
f T=^+6000 -120^H /, 80 mm
/J^f-2000 i? 2
mm NA
200 mm 20* 20 mm
(•f
- 2000 V 160 mm '/
mm
WM )

V//M
4000

^)////;;///,
/,

lift
-(|+2000)
Figure 5-30.
5-8 Design for Flexure and Shear 203

We now substitute the absolute values of V and Q into Eq. (5-4)

to obtain

-P +
'-*« 1.2 x 6
10 = —2000 (416 x 10" 6
)
(65.7 x 10" 6 )(0.040)

from which
P= 11.2 kN
The maximum moment between the supports in terms of P is at
x = 2 m and has the value
M = (\P - 2000)(2) = (P - 4000) N m
Applying the flexure formula, we have
6
= X 10 )(65.7 X 10~ 6
P - 4000
(8 )

C 0.100
P- 9.26 kN
The maximum safe value of P is the smaller of the above values,
namely, P = 9.26 kN.

PROBLEMS

(^580^ A rectangular beam of width b meters and height h meters

carries a central concentrated load P on a simply supported span of
length L meters. Express the maximum t in terms of the maximum af .

Ans. t = afh/2L
581. A laminated beam is composed of three planks, each
150 mm by 60 mm, glued together to form a section 150 mm wide by
180 mm high. The allowable shear stress in the glue 600 kPa, the is

wood is 900 kPa, and the allowable flexure

allowable shear stress in the
stress in the wood is 8 MPa. Determine the maximum uniformly
distributed load which can be carried by the beam on a 2-m simple
span. Ans. w = 12.2 kN/m
582. Find the cross-sectional dimensions of the smallest square
beam which can be loaded as shown in Fig. P-582 if r < 900 kPa and
a < 8 MPa.
4kN
M
W = 5kN-m
lm lm | 1

ur
Figure P-582.
204 5 Stresses in Beams

583. A wide-flange section having the dimensions shown in Fig.

P-583 supports a central concentrated load P on a simple span L
meters long. Determine the ratio of the maximum flexure stress to the
maximum shear stress.

584. A built-up timber beam having the same cross section as

that in Problem 583 is used to support a concentrated load Pon a
simply supported span 8 m long. Determine P and its location that
would cause simultaneously a maximum flexural stress of 8 MPa and a
maximum shearing stress of 1.2 MPa.

i 25 mm
UZZZZ7*7777A

250 mm
/— 25 mm
T777T//77777\ 25 mm
[^200 mm^
Figures P-583 and P-584.

585. A simply supported beam L meters long carries a uniformly

distributed load of 16 kN/m over its and has the cross
entire length
section shown in Fig. P-585. Find L to cause a maximum flexural stress
of 40 MPa. What maximum shearing stress is then developed?
Ans. L = 1.77 m; t = 5.55 MPa

140 mm

y/////////////A 2o mm

160 mm

JL20 mm
Figure P-585.

586. A simply supported beam 6 m long carries a uniformly

varying load which varies from zero at one end to w N/m at the other.
5-8 Design for Flexure and Shear 205

The beam section is the same as that in Fig. P-577. Find the maximum
safe value of w if of < 10 MPa and t < 800 kPa.
587. The wide flange beam shown in Fig. P-587 supports the
concentrated load W
and a total uniformly distributed load of 2W.
Determine the maximum safe value of W if of < 10 MPa and t < 1.4
MPa. AnS. W = 2.62 kN

20 mm
VI p 200 mm »-j

2m - I

V///////////77\
2W 1 m { 2 m
20 mm +- 180
7mm

E
20
1mm
Figure P-587.

588. The in Fig. P-588 is supported by a

distributed load shown
box beam with the given dimensions. Determine the maximum value of
w that will not exceed a flexural stress of 14 2
or a shearing stress MN/m
of 1.2 MN/m 2
.

300 mm

lm w N/m
mm
T 250 mm

=P
150
2m
J
200 mm
Figure P-588.

589. A
channel section carries two concentrated loads and a W
total uniformly distributed load of 8 W, as shown in Fig. P-589. Verify
that the NA is 50 mm above the bottom and that 7NA = 15.96 X
10
6
mm 4
. Then use these values to determine the maximum value of W
that will not exceed allowable stresses in tension of 30 MPa, in compres-
sion of 70 MPa, or in shear of 20 MPa. Arts. W = 3.19 kN
 —

206 5 Stresses in Beams

20 mm —*- *— 140 mm +• 20 mm
H W

sw
r
NA 140 mm
1 111

r^\ 1 m
tZz.

20
J

mm
w
-

Figure P-589.

590. A rectangular beam, 150 mm wide by 250 mm high, carries

a uniformly distributed load of 8 kN/m and a concentrated load P as
shown in Fig. P-590. Determine the maximum safe value of P if a <
lOMPaandr < 1.2 MPa.

1 m 2m
8 kN/m

Figure P-590.

5-9 SPACING OF RIVETS OR BOLTS IN BUILT-UP BEAMS

In our analysis of flexure action (Art. 5-6) we showed that the various
elements composing a built-up beam tend to slide past one another. We
shall now consider the size and spacing of rivets or bolts in a built-up
beam to resist this sliding action. The first step is to calculate the force
to be resisted by such rivets.
Figure 5-31 shows a beam composed of three planks bolted
together by two rows of bolts spaced e apart. Equation (5-4) gives the
shearing stress at the contact surface between the two upper planks as

where Q is the static moment about the NA of the shaded area in the
end view. Multiplying this shearing stress by the shaded area eb in the
top view gives the force F to be resisted in a length e:
5-9 Spacing of Rivets or Bolts in Built-up Beams 207

Top view

n Ol JZX
NA
-H
-H
— t+£
h+
ii M
O C7
Front view
^C7
End view
Figure 5-31.

The same result can be obtained more directly by using the

concept of shear flow, which is the longitudinal shearing force devel-
oped per unit length. Thus in the length e, Eq. (5 -4a) determines the
shear force to be

as before.
Friction being neglected, this force is resisted by the shearing or
bearing strength R of the bolts, whichever is smaller. Equating R to F
gives

R= (5-7)

If the vertical shear varies in a beam, V is the average vertical

shear in the interval e; but it is usually taken as the maximum V in this
interval, especially in built-up steel girders where the length of the
interval is taken as a panel length equal to the depth of the girder. In
this case, Eq. (5-7) gives the rivet pitch in each panel length.

ILLUSTRATIVE PROBLEM

591. A plate and angle girder is fabricated by attaching the short

legs of four 125 mm angles to a web plate 1100 mm by
X 90 x 13
10 mm to form a section 1120 mm deep, as shown in Fig. 5-32. The
moment of inertia* about the NA / = 4140 X 10 mm4 At a section
is
6
.

•Current specifications of AISC call for no deduction for rivet holes in

computing /, provided that the rivet hole area does not exceed 15% of the gross
flange area. If it does, only the area in excess of 15% need be considered in
modifying / to deduct for rivet holes.
208 5 Stresses in Beams

23.7 mm-. A= 2630 mm'

\ a>

560 - 23.7 = 536.3 mm

ip NA
1120 mm -

-Web plate 1 100 mm X 10 mm

,
<f =t
Figure 5-32.

where V= 450 kN, determine the spacing between 19-mm rivets that
fasten the angles to the web plate. Use t = 100 MPa; in bearing, use
ob = 220 MPa for rivets in single shear and ob = 280 MPa for rivets in
double shear.

Solution: The rivets must resist the longitudinal force tending to slide
the two flanges past the web. Hence it is the static moment of area of
these two flange angles that must be used in Eq. (5-7). Referring to Fig.
5-32, we obtain

Q - 2(2630)(536.3) - 2820 x 103 mm 3

= 2820 x 10" 6 m 3

The shearing resistance of a 19-mm rivet in double shear is

2
Rs = (A s t)(2) - j(0.019) (100 X 6
10 )(2) = 56.7 kN
The bearing resistance against the web plate is

Rb = {dt)ob = (0.019)(0.010)(280 x 10
6
) = 53.2 kN
Using the lower of these values in Eq. (5-7), we get the required
rivet pitch

x 3
x 10" 6
e = —RI
—= (53.2
^
10 )(4140
-^ -r
)
= A tnA
0.174 m
VQ (450 X 3
10 )(2820 x 10" 6
)
= 174 mm Ans.

PROBLEMS

592. A wide-flange section is formed by bolting together three

planks, each 80 mm by 200 mm,
arranged as shown in Fig. P-592. If

each bolt can withstand a shearing force of 8 kN, determine the pitch if
5-9 Spacing of Rivets or Bolts in Built-up Beams 209

the beam is loaded so as to cause a maximum shearing stress of

1.2 MPa. Ans. e - 98.2 mm

A
*r 80 mm X 200 mm

Figure P-592.

593. A box beam, built up as shown in Fig. P-593. is secured by

screws spaced 100 mm apart. The beam supports a concentrated load P
at the third point of a simply supported span 3 m long. Determine the
maximum value of P that will not exceed a shearing stress of 800 kPa in
the beam or a total shearing force of 1200 N in the screws. What is the
maximum flexural stress in the beam?

— 160 mm-
^\\\\\\\^
120
mm
160 mm 200 mm

4^\s\sN\^W
Figure P-593.

594. A distributed load of w N/m is applied over the entire

length of a simply supported beam 4 m long. The beam section is that of
Problem 593, but used here so that the 160-mm dimension is vertical.
Determine the maximum value of w if oy < 10 MPa, t < 800 kPa, and
the screws have a shear strength of 800 N and a pitch of 50 mm.
Ans. w = 2.05 kN/m
595. A concentrated load P is carried at midspan of a simply
supported beam 6 m long. The beam is made of 50-mm by 150-mm
pieces of wood, screwed together as shown in Fig. P-595. If the
maximum flexural stressdeveloped is 9 MN/m 2
, find the pitch of the
screws if each screw can resist 800 N.
210 5 Stresses in Beams

M V^ JU>

4JJ$tazQ

Figure P-595.

596. Three planks 100 mm by 150 mm, arranged as shown in

Fig. P-596 and supported by bolts 0.4 m apart, are used to support a
concentrated load P at the center of a simply supported span 6 m long.
If P causes a maximum flexural stress of 12 MPa, determine the bolt
diameters, assuming the shear between the planks is transmitted by
friction only. The bolts are tightened to a tension of 140 MPa, and the
coefficient of friction between the planks is 0.40. Ans. d = 19.1 mm

150 mm

300 mm

Figure P-596.

597. A plate and angle girder similar to that shown in Fig. 5-32
is fabricated by riveting the short legs of four 125 X 75 X mm angles
13
to a web plate 1000 mm by 10 mm to form a section 1020 mm deep.
Cover plates, each 300 mm by 10 mm, are then riveted to the flange
angles making the overall height 1040 mm. The moment of inertia of the
entire section about the NA is / = 4770 X 106 mm4 Using the allow-
.

able stresses specified in Illustrative Problem 591, determine the rivet

pitch for 22-mm rivets attaching the angles to the web plate at a section
where V- 450 kN.
598.As shown in Fig. P-598, two C380 X 60 channels are
riveted together by pairs of 19-mm rivets spaced 200 mm apart along
the length of the beam. What maximum vertical shear V can be applied
to the section without exceeding the stresses given in Illustrative Prob-
lem 591? Ans . 25.9 kN
 -

5-9 Spacing of Rivets or Bolts in Built-up Beams 211

Jl
=9=

Figure P-598.

599. A beam is formed by riveting together two W250 X 73

sections as shown in Fig. P-599. It used to support a uniformly
is

distributed load of 30 kN/m (including the weight of the beam) on a

simply supported span of 8 m. Compute the maximum flexural stress
and the pitch between rivets that have a shearing strength of 26 kN.

NA , O '
i
f> ,

Figure P-599.

SUMMARY

For homogeneous beams, originally straight, carrying transverse

loads in the plane of symmetry, the bending moment creates flexural
stresses expressed by

•-* (5-2)

The flexural stresses vary directly with their distance y from the neutral
axis, which coincides with the centroidal axis of the cross section.
Maximum flexural stresses occur at the section of maximum
bending moment at the extreme fibers of the section. The distance from
the NAto the extreme fibers being denoted by c, the flexure formula
becomes

Max. a =
iwt
Mc
— — =—
M (5-2a, b)
5
in which S = I/ c represents the section modulus of the beam. For
geometric shapes, values of S are tabulated in Table 5-1 (page 159); for
structural shapes, the values are given in Appendix B.
212 5 Stresses in Beams

The up numerically equal shearing stresses on

vertical shear sets
longitudinaland transverse sections (Eq. 5-5, page 191), which are
determined from

in which A' is the partial area of the cross section above a line drawn
through the point at which the shearing stress is desired. Q= A'y is the
static moment about the NA of this area (or of the area below this line).
Maximum shearing stresses occur at the section of maximum V
and usually at the NA. For rectangular beams, the maximum shearing
stress is

Max.r=|£ (5-6)

In wide-flange beams, a very close approximation is

= V
Max. r
^web
where A web is the web area between the flanges.
The rivet pitch in built-up beams is given by

-% < 5-7>

where R is the rivet resistance in the pitch length e, / is the moment of

inertia of the gross section about the NA, V is the maximum vertical
shear in the interval e, and Q is the moment of area about the NA of
the elements whose sliding is resisted by the rivets.
 Beam
Deflections

6-1 INTRODUCTION

In this chapter we consider the rigidity of beams. Frequently the design

of a beam determined by its rigidity rather than by its strength. For
is

example, in designing metalworking equipment for precision work, such

as lathes, milling machines, or grinders, the deformations must be kept
below the permissible tolerances of the work being machined. Again,
floor beams carrying plastered ceilings beneath them are usually re-
stricted to a maximum deflection of 1/360 of their length in order to
avoid cracks in the plaster. One of the most important applications of
beam deflections is to obtain equations with which, in combination with
the conditions of static equilibrium, statically indeterminate beams can
be analyzed. (See Chapters 7 and 8.)
Several methods are available for determining beam deflections.
Although based on the same principles, they differ in technique and in
their immediate objective. We shall consider first a modernization of the
double-integration method which greatly broadens and simplifies its
application. Another method, the area-moment method, is thought to be
the most direct of any, especially when the deflection at a specific
location is desired. After a preliminary discussion (Art. 6-4), it will be

213
214 6 Beam Deflections

found to be not only simple, but extremely rapid to apply. A variation

of it, which we will take up in Art. 8-7, is also rapid and easy to use.

Other methods are the conjugate-beam method and the method of

superposition. The conjugate-beam method is a variation of the area-
moment method but differs from it in technique. The method of
superposition is not an independent method; it uses the deflection
formulas for certain fundamental types of loadings to obtain results for
loadings that consist of combinations of these fundamental types.

6-2 DOUBLE-INTEGRATION METHOD

The edge view of the neutral surface of a deflected beam is called the
elastic curve of the beam. It is shown greatly exaggerated in Fig. 6-1.
This article shows how to determine the equation of this curve, i.e., how
to determine the vertical displacement y of any point in terms of its x
coordinate.
Select the left end of the beam as the origin of an X axis directed
along the original undeflected position of the beam, and a Y axis
directed positive upward. The
assumed to be so small
deflections are
that there is no appreciable difference between the original length of
the beam and the projection of its deflected length. Consequently, the
elastic curve is very flat and its slope at any point is very small. The
value of the slope, tan = ay/dx, may therefore with only small error
be set equal to 0; hence

. dy
= (a)
dx

and

d9__ d^y
(b)
dx 2
dx

i Y A
\

1 \ Elastic curve -v.

o ""-^^ t

y ^^O^ L>^ X
»* x *

*dx* ^-Segment of
beam
Flgun I 6-1. Elastic curve.
6-2 Double-Integration Method 215

If we now consider the variation in in a differential length ds

caused by bending in the beam, it is evident that
ds = pdB (c)

where p is the radius of curvature over the arc length ds. Because the
elastic curve is very flat, ds is practically equivalent to dx; so from Eqs.
(c) and (b) we obtain

«—
2
-= 1
—
dO
ds
dO
dx
or
1 dy
- = -^-
2
,.v
(d)
P p dx
In deriving the flexure formula in Art. 5-2, we obtained on page
157 the relation

p
" EI (5-1)

Equating the values of 1/p from Eqs. (d) and (5-1), we have

EI ±± = M (6 _ 1}
dx
This is known as the differential equation of the elastic curve of a beam.
The product EI, called the flexural rigidity of the beam, is usually
constant along the beam.
The approximations we have made do not seriously invalidate Eq.
(6-1), for if we replace 1/p by its exact value as found in any calculus

text, we have, from Eq. (5-1),

£l
dx 1 _ M
213/2 EI
1 + (£)'
'

Since dy/dx is very small, its square is negligible compared with unity,
and hence we obtain

d^y M_
dx 2 EI

which is the same as Eq. (6-1).

If Eq. (6-1) is now integrated, assuming EI constant, we obtain

EI^ =JMdx+C l (6-2)

This is the slope equation specifying the slope or value of dy/dx at any
point. Note that here M represents the moment equation expressed in
terms of *, and C x
is a constant to be evaluated from the given
conditions of loading.
216 6 Beam Deflections

We now integrate Eq. (6-2) to obtain

Ely = f ( Mdxdx + C,jc + C2 (6-3)

This is the required deflection equation of the elastic curve specifying

y for any value of x\ C 2 is another constant of integration
the value of
which must be evaluated from the given conditions of the beam and its
loading.
If the loading conditions change along the beam, there is a
corresponding change in the moment equation. This requires that a
separate moment equation be written between each change of load
point and that two integrations of Eq. (6-1) be made for each such
moment equation. Evaluation of the constants introduced by each
integration can become very involved. Fortunately, these complications
can be avoided by writing a single moment equation in such a way that
it becomes continuous for the entire length of the beam in spite of the

discontinuity of loading.
For example, consider the beam shown in Fig. 6-2. Using the
definition M = (2M) L discussed in Art. 4-2, we find that the moment
equations between the change of load points are

MAB = 480jc N-m

MBC = [480* - 500(x - 2)] N-m
MCD = 480x - 500(x - 2) - ^- ( x ~ 3)
2
N-m

Observe that the equation for M

CD will also be valid for both AB and M
MBC provided that the terms (x — 2) and (x — 3) are neglected for
2

values of x less than 2 and 3, respectively. In other words, the terms

(x — 2) and (x — 3) are nonexistent for values of x for which the terms
2

in parentheses are negative.

As a reminder of these restrictions, we adopt a notation in which
the usual form of parentheses is replaced by pointed brackets, viz.,

< > . With this change in notation, we obtain a single moment equation

M= (480* - 500<x - 2> - ^O - 3>

2
)
N- m

500 N

450 N/m
X
2 m 1 m 2m

480 N 920 N
Figure 6-2.
 X

6-2 Double-Integration Method 217

600 N

600 N
J
400 N/m 400 N/m
B C D El
I—
lm 3m 2ml2m
J
1
™ ^ l...D^_ J
2m I 2 m
R2 = 1300N

(a) (b)

Figure 6-3. Technique of establishing continuity of loading.

which is valid for the entire beam if we postulate that the terms between
the pointed brackets do not exist for negative values; otherwise the term
is be treated like
to any ordinary expression.*
As another example, consider the beam in Fig. 6-3a. Here the
distributed load extends only over the segment BC. We can create
continuity, however, by assuming that the distributed load extends
beyond C and adding an equal upward-distributed load to cancel its
effect beyond C, as shown in Fig. 6-3b. The general moment equation,
written for the last segment DE in our new notation using pointed
brackets, is

M m- 500.x — 400/
<jc- fv2
2
1> +
400/ „ v2
<x-4> 2
—
+ 1300<x -6>)n- m

As before, we specify that the terms between the pointed brackets do

not exist for negative values. Notice that all loadings are automatically
included in the general moment equation by writing it for the last
segment of the beam.

ILLUSTRATIVE PROBLEMS

601. A concentrated load of 300 N is supported as shown in Fig.

6-4. Determine the equations of the elastic curve between each change
of load point, and the maximum deflection in the beam.

*The justification for ignoring negative values of the terms in the pointed
brackets depends on the fact that the general moment equation is written using the
definition M
* (2M) L which means that we consider the effects of loads lying only
,

to the left of an exploratory section. A negative value of the terms in a pointed

bracket indicates a loading that is to the right of an exploratory section, whereas a
zero value merely indicates the start of a loading.
218 6 Beam Deflections

Y\ 300 N

2 m b\ 1 m C y

r
Z^-IOON
^A>

Figure 6-4.
^
/? 2 =200N

Solution: Writing the general moment equation for the last segment
BC of the beam, applying the differential equation of the elastic curve,
and integrating twice, we obtain the following slope and deflection
equations:

EI^ = M = (lOOx - 300<a: - 2» Nm (a)

El^- - (50.x
2
- 150<x - 2> 2 + C,) N-m 2 (b)

= -y x - 50<x - 2> + C x + C2 ) N-m3

3 3
Ely f
x
(c)

To evaluate the two constants of integration which are physically

equivalent to slopeand deflection at the origin, we apply the following
boundary conditions:

1. At A where x = 0, the deflection v = 0. Substituting these

values in Eq. (c), we find that C2 = 0. Remember that (x — 2>
3
is to be
ignored for negative values.

2. At the other support where x = 3, the deflection y is also zero.

Knowing that C2 = and substituting these values in the deflection
equation (c), we obtain

3 3
- = -
- y (3) - 50(3 2) + 3C, or C, 133 N-m2

Having thus evaluated the constants of integration, we return to

Eqs. (b) and (c) to rewrite the slope and deflection equations in the
conventional form shown in the tabulation on page 219.
Continuing the solution, we assume that the maximum deflection
will occur in the segment A B. Its location may be found by differenti-

ating Eq. (e) with respect to x and setting the derivative equal to zero,
or, what amounts to the same thing, setting the slope equation (d) equal

to zero and solving for the point of zero slope. We obtain

50a:
2
- 133 = or x = 1.63 m
6-2 Double-Integration Method 219

SEGMENT AB (0 <x < 2)

ElQ- -(50x 2 - 133) Nm 2

(«) £fy =
^ x3 _ i3 3x
\
Nm . 3

SEGMENT BC (2 < x < 3)

50x 2 - 150(x - 2)
2
- 133] N m2
(g) £7y = ^jc 3 -50(x-2) 3 - 133* Nm 3

Since this value of x is valid for the segment AB, our assumption
that the maximum deflection occurs in this region is confirmed. Hence,
to determine the maximum deflection, we substitute x = 1.63 in Eq. (e),
which yields
Max. Ely = -145 N-m3
The negative value obtained indicates that the deflection v is downward
from the X axis. Frequently only the magnitude of the deflection,
without regard to sign, is desired; this is denoted by 5, the use of v being
reserved to indicate a directed value of deflection.
The unit of the product N m This follows from integrat-
Ely is
3
.

ing Eq. (6-1) twice. With M in units of N m, the first integration gives •

N-m as the unit of the slope equation. A second integration results in

2

N-m 3 as the unit of the deflection equation. For consistent units, E

must be in units of N/m and / in units of m Then the deflection v
2 4
.

will be determined in meters. For example, if E — 10 X 10 N/m and

9 2

/ = 1.5 X 10 6 4
mm
= 1.5 X 10" 6
m the value of y is
4
,

(10 X 9
10 )(1.5 X 10" 6 )v 145
whence
y = -9.67 x 10"
3
m = -9.67 mm
602. Find the value of Ely at the position midway between the
supports and at the overhanging end for the beam shown in Fig. 6-5.

600 N
lm "* o m ** l m • 9 m

400 N/m
B C D 1

X
i "*
^----— — -"" ^^-
1\

/?! = 500 N
Figure 6-5.
220 6 Beam Deflections

Solution: This is the same beam for which we determined the general
moment equation on page 217. Applying the differential equation of the
elastic curve, and integrating twice, we obtain

„£.*.(»>, -Me -* + ««« -4?

+ 300<jc
13(XX;c -6>^Nm
- 6>)

£l^-(250^ .™<*-|>3 + »°<*-4>>

H-650<x -6> 2 + C,)n m2

Ely- /250 r3
fV-f<*-l>< +
f<*-4><
650
<x - 6> 3 + C,x + C2 N )
3
To determine C2 we note that Ely =
, at x = 0, which gives
C2 = 0. Note that we ignore the negative terms in the pointed brackets.

Next we use the condition that Ely - at the right support where
x = 6. This gives

0-^(6)3 -y(5)4 + y(2)4 + 6C 1

or C^-BOSN-m 2

midspan deflection, we substitute x = 3 in

Finally, to obtain the
the deflection equation for segment BC obtained by ignoring negative
values of the bracketed terms <x - 4> and <x — 6> We obtain
4 3
.

3 4
- = -
Ely = ^(3) - y (2) 1308(3) 1941 N-m 3 Ans.

Also, at the overhanging end where x = 8, we have

4 3
= ^0(8) 3 - ^(2) -
Ely
f (1? + f (4) + 1308(8)

= -1814 N-m 3 ^aw.

603. A simply supported beam carries the triangularly distrib-

uted load shown in Fig. 6-6a. Determine the deflection equation and
the magnitude of the maximum deflection.

Solution: Because of symmetry, each reaction is one-half of the total

load of {wL, or R = R 2 =\wL. In this example, we take further
l

advantage of symmetry to note that the deflection curve from A to B is

the mirror image of that from C to B. The conditions of zero deflection
at A and of zero slope at B do not require the use of a general moment
equation. Only the moment equation for segment AB is needed, and this
is easily found with the aid of Fig. 6-6b.
6-2 Double-Integration Method 221

—X

Ro = ™k

Flgure 6-6.

Applying the differential equation of the elastic curve to segment

AB and integrating twice, we obtain
2

EI-^z2 = MAB = —4-x L

- x
dx
2

EI
dy _ wLx
lTx-~T~
Ely = wLx
24
To evaluate the constants of integration, we note that at the
support A , v = at x = 0. Hence from Eq. (c), we have C2 = 0. Also,
because of symmetry, the slope dy / dx = at midspan where x = L/2.
Substituting these conditions in Eq. (b) yields

= wL( L\
2
^ w L \4 5>vL
:

n
x(y) -i2iT +c I
'
or C, =
192
Hence the deflection equation from ^4 to 5 (and also from C to 5
because of symmetry) becomes

= wLx
5
5wL 3x
Ely
24 60L 192
which reduces to
wx
£/y - (25L 4 - 40L 2x 2 + 16x
4
)
960L
The maximum deflection at midspan, where x — L/2, is then
found to be
wL4 _ WL 3

y 120
" 60
where W —\wL is the total load.

604. Determine the equation of the elastic curve of a cantilever

beam supporting a uniformly distributed load of w N/m over part of its

length as shown in Fig. 6-7.

222 6 Beam Deflections

M--£(L -a M 2 2
) .

V-w(L-a)
Figure 6-7.

Solution: At the wall, equilibrium conditions determine the shear and

moment to be

V=w(L-a) and M= - w(L - a)(a + ^-^)

= -t(^ 2 -« 2
)

In terms of the general moment equation, the differential equation of

the elastic curve becomes

EI^- = w(L - a)x - -(L 2 - a

2
)
- Tz <x - a)' (a)
dx z

Integrating Eq. (a\ we obtain the slope equation:

^ T dy x' w .2

Elfx = w{
/ r
L-a)^-- (L
x 2
- a )x
2
- ^(x - ay + C, (b)

However, the slope dy/dx is zero at x = 0, so C, = 0. We may now

integrate the slope equation (with C, = 0) and obtain the deflection
equation:

Ely = w(L - a)£ - ^(L 2 - a 2 )*

2
- ^<x - a)* + C2 (c)

Since v = at x = 0, we find that C2 = also. Observe that selecting

the origin of axes at the perfectly restrained wall where the slope and
deflection are zeromakes the constants of integration also zero.
The value of the maximum deflection, which occurs at the free
end, is denoted by 8. Evidently 8 = — v; so on substituting x = L and
simplifying, we obtain

w{L -
EIS=
a)
[L> + L 2a + La 2 -^
One important variation of this result occurs when a = 0. Then
the entire length of the beam is uniformly loaded and the maximum
deflection is given by

EI8 = WL
8 8
6-2 Double-Integration Method 223

PROBLEMS

605. Determine the maximum deflection 8 in a simply supported

beam of length L carrying a concentrated load P at midspan.
Ans. 8 = PL 3 /4SEI
606. Determine the maximum deflection 8 in a simply supported
beam of length L carrying a uniformly distributed load of w N/m
applied over its entire length.
Ans. 8= (5/3S4)(wL 4 /EI) = 3
(5/384)(JfT / EI)
607. Determine the maximum value of Ely for the cantilever
beam loaded as shown in Fig. P-607. Take the origin at the wall.

Figure P-607.

608. Find the equation of the' elastic curve for the cantilever
beam shown in Fig. P-608; it carries a load that varies from zero at the
wall to w N/m at the free end. Take the origin at the wall.

w N/m-

Figure P-608.

609. As shown P-609, a simply supported beam carries

in Fig.
two symmetrically placed concentrated loads. Compute the maximum
deflection 8 and compare one-half this result with the midspan 8 to case
7, Table 6-2, page 270. Check your answer by letting a = L/2 and

comparing it with the answer to Problem 605.

Ans. 8 = (Pa/24EI)(3L 2 - 4a 2 )

]_L
F \
Figure P-609.
224 6 Beam Deflections

610. The simply supported beam shown in Fig. P-610 carries a

uniform load of w N/m symmetrically distributed over part of its
length. Determine the maximum deflection 8 and check your result by
letting a = and comparing with the answer to Problem 606.

L 26—*j-
N/m |

K,

Figure P-610.

611. Compute the value of EI8 at midspan for the beam loaded
as shown in Fig. P— 61 1. If E — 10 GN/m 2 what value of / is required ,

to limit the midspan deflection to 1/360 of the span?

Arts. EI8 = 500 3
/ Nm ;
= 4.50 X 10
6
mm4
300 N/m
2 m 2 m

*i *2
Figure P-611.

o
612. Compute the midspan value of EI8 for the beam loaded as
shown in Fig. P-612. Ans. EIS = 657 Nm 3

2 m
1.5 m 300 N/m 0.5 m

s 4

Figure P-612.
m X
613. Compute the value of Ely at the right end of the beam
loaded as shown in Fig. P-613. Ans. Ely = 195 Nm 3

400 N
3 m
400 N/m 1 m lm |

4 m

Ri
Figure P-613.
6-2 Double-Integration Method 225

614. For the beam loaded as shown in Fig. P-614, (a) compute
the slope of the elastic curve over the right support and (b) determine
the maximum deflection between the supports.

800 N
800 N/m

1 m 2 m

Ri
r
R2
Figure P-614.

61^ Compute the value of Ely midway between the supports for
the overhanging beam shown in Fig. P-615.
Ans. Ely = -4.66kNm 3

41

2kN/m 2kN/m
2m 2m t i 2m

*1
Figure P-615.

616.For the beam loaded as shown in Fig. P-616, determine (a)

the deflection and slope under the load P and (b) the maximum
deflection between the supports.
Ans. (b) Max. Ely = Pa b/9V3
2

*— b
r

Figures P-616 and P-617.

617. Replace the load P by a clockwise couple applied at the M

right end and determine the slope and deflection at the right end.
Ans. EI dy/dx = - (M/3)(L + 2b); EI8 = (Mb/6)(2L + b)
226 6 Beam Deflections

618. A simply supported beam carries a couple M applied as

shown in Fig.P-618. Determine the equation of the elastic curve and the
deflection at the point of application of the couple. Then, letting a — L
and a = 0, compare your solution of the elastic curve with cases 1 1 and
12 in Table 6-2 on page 270.
Ans. Ely = (Ma/3L)(L 2 - 3La + 2a 2)
M
"

00 777777"
777779

Figure P-618.

619. Determine the midspan value of Ely for the beam loaded as
shown in Fig. P-619. {Hint: Take advantage of symmetry to note that
the slope is zero at midspan.) Ans. Ely = — 3.33 kN-m 3

400 N/m 400 N/m

k
2m 2m 2m ik

*i R2
Figure P-619.

620. Find the midspan deflection 8 for the beam shown in Fig.
P-620, carrying two triangularly distributed loads. {Hint: For con-
venience, select the origin of the axes at the midspan position of the
elastic curve.) Ans. 8 = (9/1920)(wL 4 / EI)

w N/m \>^^

U- <
L -J * L
2 2 '

K, R,
Figure P-620.

Determine the value of EI8 midway between the supports

621.
for the in Fig. P-621. Check your result by letting a =
beam shown
and comparing with Problem 606. (Apply the hint given in Problem
620.)
6-3 Theorems of Area-Moment Method 227

N/m
W IN/
a L la
R x R2
Figure P-621.

6-3 THEOREMS OF AREA-MOMENT METHOD

A useful and simple method of determining slopes and deflections in

beams involves the area of the moment diagram and also the moment of
that area-the area-moment method. We discuss first the two basic
theorems of the method; then, after showing how to compute the area
and moment of area of the moment diagram, we shall apply the method
to several types of problems. The method is especially useful in directly
determining the slope or deflection at a specified position. Depending as
itdoes upon the geometry of the elastic curve, the area-moment method
emphasizes the physical significance of slope and deflection.
The area-moment method is subject to the same limitations as the
double-integration method; but in order to present it in its entirety as a
completely independent alternative method, we repeat a small portion
of the preceding article. Figure6-8a shows a simple beam that supports
any type of loading. The elastic curve is the edge view of the neutral
surface and is shown, with greatly exaggerated deflections, in Fig. 6- 8b;
the moment diagram is assumed to be as in Fig. 6-8c.
As we saw in the derivation of the flexure formula, Art. 5-2, two
adjacent plane sections of an originally straight beam will rotate
through the angle dO relative to each other. This is demonstrated in the
enlarged detail of Fig. 6- 8b, in which it is also apparent that the arc
distance ds measured along the elastic curve between these two sections
equals p d9 where p is the radius of curvature of the elastic curve at the
9

given position. From Eq. (5-1) we have

" EI
p

and since ds — p d0, we now write

p EI ds
or

<»-%* (a)

In most practical cases the elastic curve is so flat that no serious

error is made in assuming the length ds to equal its projection dx. With
228 6 Beam Deflections

Area-moment theorems.

this assumption, we obtain

(b)
EI
It is drawn to the elastic curve at C and D in
evident that tangents
Fig. 6-8b are separated by same angle dO by which sections OC and
the
OD (in the enlarged detail) rotate relative to each other. Hence the
change in slope between tangents drawn to the elastic curve at any two
points A and B will equal the sum of such small angles:

rO B 1 r xB
(c)

Note also in Fig. 6- 8b that the distance from B on the elastic

curve (measured perpendicular to the original position of the beam) that
will intersect a tangent drawn to this curve at any other point A is the
sum of the intecepts dt created by tangents to the curve at adjacent
points. Each of these intercepts may be considered as the arc of a circle
of radius x subtended by the angle dO:

dt - xdO
6-3 Theorems of Area-Moment Method 229

Reference tangent at A
Reference tangent at B
Figure 6-9. Inequality of tA/B and tB/A .

Hence
t
B/A - Sdt - fx dO

Replacing dO by the value in Eq. (b\ we obtain

1 r*BB
t
B /A=-Eil x(Mdx) (d)
x

The length tB / A is known as the deviation of B from a tangent

drawn at A or as the , tangential deviation of B with respect to ^4 . The
subscript indicates that the deviation measured from B relative to a
is

reference tangent drawn at A. Figure 6-9 illustrates the difference

between the deviation tB/A of B from a reference tangent at A, and the
deviation tA/B of A from a reference tangent at B. In general, such
deviations are unequal.
The geometric and (d) gives rise to the two
significance of Eqs. (c)
basic theorems of the area-moment method. From the moment diagram
in Fig. 6- 8c we see that M
dx is the area of the shaded element located
a distance jc from the ordinate through B. Since j dx means a M
summation of such elements, Eq. (c) may be expressed as

r
AB = — (area)^ (6-4)

This is the algebraic expression of Theorem I, which is stated as follows:

Theorem I: between tangents drawn to the elastic

The change in slope
curve at any two points A and B is equal to the product of I /EI
multiplied by the area of the moment diagram between these two points.
Figure 6-8c shows that the expression x(M dx) which appears
under the integral sign in Eq. (d) is the moment of area of the shaded
element about the ordinate at B. Hence the geometric significance of
the integral fx(M dx) is that the integral is equivalent to the moment of
area about the ordinate at B of that part of the moment diagram
230 6 Beam Deflections

between A and B. Thus we obtain the algebraic form of Theorem II

>*/A = —
£/
(area) BA *B (6-5)

This is stated more formally as:

Theorem II: of any point B relative to a tangent drawn

The deviation
to the elastic curve atany other point A in a direction perpendicular to
,

the original position of the beam, is equal to the product of \/EI

multiplied by the moment of area about B of that part of the moment
diagram between points A and B.
The product EI is called flexural rigidity. Note that we have tacitly
assumed that E and / remain constant throughout the length of the
beam; this is usually the case. If they vary, however, they cannot be
written outside the integral sign, and the manner of their variation with
x must be known. Such variations are usually taken care of by dividing
each moment ordinate by EI to obtain an M/EI diagram which is used
in place of the moment diagram in the foregoing theorems.
In the two theorems, (area)^ is the area of the moment diagram
between points A and B, and xB is the moment arm of this area
measured from B. When the area of the moment diagram is com-
posed of several parts (this is explained in Art. 6-4), the expression
(area)^ xB includes the moment of area of all such parts. The moment
•

of area is always taken about an ordinate through the point at which the
deviation is being computed. An automatic method of using the correct
axis for moments is to give x the same subscript, for example, B
(meaning that moment arms are to be measured from this point), as
appears in the numerator of the subscript to / (i.e., B/A).
One rule of sign is very important: The deviation at any point is
positive if the point lies above the reference tangent from which the
deviation is measured, and negative if the point lies below the reference
tangent. Positive and negative deviations are shown in Fig. 6-10.
Conversely, a computed positive value for deviation means that the
point must lie above the reference tangent.

(a) Positive deviation; B located (b) Negative deviation; B located

above reference tangent below reference tangent
Figure 6-10. Signs of deviations.
6-4 Moment Diagrams by Parts 231

(a) Positive change of slope; 9AB (b) Negative change of slope; AB

iscounterclockwise from left is clockwise from left tangent
tangent
Figure 6-11. Signs of change of slope

Another rule of sign that concerns slopes is shown in Fig. 6-11. A

positive value for thechange in slope 9AB means that the tangent at the
rightmost point B is measured in a counterclockwise direction from the
tangent at the leftmost point, and vice versa.

6-4 MOMENT DIAGRAMS BY PARTS

In order to apply the theorems of the area-moment method, we should

be able to compute easily and accurately the area under any part of a
moment diagram, and also the moment of such an area about any axis.
A method of doing this from calculus is to integrate the two expressions
jM dx and fx(M dx) between proper limits, noting that the bending
moment M must be expressed as a function of x.
Our purpose here, however, method of dividing
is to discuss a
moment diagrams whose areas and centroids are known; this
into parts
permits simple numerical calculations to replace integrations. The first
step is to learn how to draw moment effects of each separate loading
(hereafter called moment diagram by parts) instead of a conventional
moment diagram.
The construction of moment diagrams by parts depends on two
basic principles:

1. The resultant bending moment at any section caused by any

load system is the algebraic sum of the bending moments at that section
caused by each load acting separately. This statement is expressed
algebraically by
M - (2M) L = (I M) R I (4-2)
where (2 M) L indicates the sum of the moments caused by all the forces
to the left of the section, and (2M) R is the sum of the moments caused
by all the forces to the right of the section.
2. The moment effect of any single specified loading is always
some variation of the general equation

y = kx
n
(a)
232 6 Beam Deflections

Figure 6-12. Figure 6-13.

The graph of this equation is shown in Fig. 6-12. The shaded area
and the location of its centroid are easily shown by calculus to be

-• bh (b)
n + 1

1
x = (c)
n + 2

where b is the base and h is the height.

In computing the area under the curve between positions like A
and B in Fig. 6-13, Eqs. (b) and (c) refer to the shaded area between
the curve, the ordinate at B and the tangent at A To this area must be .

added, of course, the shaded trapezoidal area between the tangent and
the X axis.
Table 6-1 demonstrates the truth of the second basic principle
stated above, viz., that the moment effect of any load is some variation

of the equation v — kx
n
This table gives data on four cantilever beams,
.

each loaded differently with increasingly complex loads.

Note that a cantilever loaded by a couple C has a moment
equation of the type v = kx n in which k = — C and n is zero, viz.,
M — — Cx°. In other words, a couple type of loading produces a
moment equation of zero degree. Similarly, a concentrated load pro-
duces a moment equation of the first degree; a uniform load produces a
moment equation of the second degree, etc.

In the column headed "Area," the area of the moment diagram is

expressed in terms of a factor multiplied by the general base distance b

and the maximum height h of the moment diagram. The position of the
centroid of each moment diagram from the maximum ordinate of the
diagram is defined as a factor of the base distance. These factors or
coefficients, which increase very simply, are obtained from Eqs. (b) and
(c) by assigning to n the value of the degree of each equation, i.e.,
n — 0, n = 1, n = 2, etc.
 \

•Q <i «c O
'K
-|«N -|<-> ~N- -|<n

<
w
•*:
•q
<:
-q
«
-Ci
-s:
«o
-^ l<^ ~|<N ~|co -|Tf
3
O
i 1h ah
1 II
i 1

< kVVVVNJT i vwvv f

— ii

t
II

^\Ny
S
H it

•o 1
^ IK

ii
»<

1
v/^
*'

H
# S^ IH

§ 1
/—
o
OF 3
QUATION
o '

o
N
VIOMENT
EGREE

*
Q" w

Z
o
H >> <s
o
3 S ~ X
1

«
|

a.2 II
1

II
sect
II

MENT
(mome
* II

Si
O
S

*
W
i
TT>-er>-^^«

^1 Mp^ i
Pi I
MPI
\
\/
X
/ i i

£s
(0
^ S
c
CO
r
X
3 is
3
3
2 \
3

5 U
L 7 TX
1 H
I
C
II
* > r r
1

\ 1, '

CO
O •g
-> >>

°S §1
i

(O WQ 3
UJ o o
CD
c o
H j
t3
O D3 1
D
>
< j
r-

233
234 6 Beam Deflections

An example will illustrate how Table 6- 1 is used to draw moment

diagrams by parts. The simple beam in Fig. 6-14 is 3 long and m
supports a uniformly distributed load of 300 N/m over the right 2 of m
the span.
At any section a-a between A and B, the moment effect defined
by M = (2M) L is caused only by /?,. Also, at any section b-b between
B and C, the moment effect will be due to /?, and to the portion of the
uniformly distributed load included between B and b-b. Note that
defining the bending moment in terms of the forces to the left of the
section means that the uniformly distributed load has no moment effect
on segment AB. Actually, the moment effect of R at any section of the x

beam is equivalent to the cantilever loading at (a), whereas the moment

effect of the uniform loading on any section of the beam is equivalent to
the cantilever loading at (b).
By referring to Table 6-1, we can plot the moment diagrams of
beams (a) and (b) on a common base line (the line of zero moment), as
shown at (c) in the figure. That the algebraic sum of the shaded areas of
(c) will yield the resultant or conventional moment diagram is evident
from the fact that the moment at any section of the original beam is
equal to the sum of the moments at that section caused by the individ-

"•"Hi
m
(a)
C 3

Moment diagram
-
300 N/m P
yy,
by parts (300 X 2) X 1 (b)

= -600 N-rn
1 m 2 m
M= -(300X2) XI
= -600 N-m
Moment caused .

Equivalent cantilever
by U D L
loadings

Resultant
moment

Conventional moment diagram

Figure 6-14. Moment diagram by parts.
6-4 Moment Diagrams by Parts 235

ual loads (basic principle 1, above). Hence, if the triangular area is

revolved about the line of zero moment as an axis, we obtain diagram

(d). The shaded area of diagram (d) is evidently equal to the area of the
conventional moment diagram (e) obtained by plotting the resultant
moment any section as ordinates to a horizontal base line. Hence the
at
conventional moment diagram may be replaced by an equivalent mo-
ment diagram constructed of parts, as in Fig. 6- 14c, whose areas and
centroids can be easily computed from the data in Table 6-1.
To compute the moment of area of the moment diagram, we
observe that the moment of the area of the conventional moment
diagram (Fig. 6-14e) is equivalent to the sum of the moments of area of
its parts, as 6- 14c. Hence, noting that each such area is
drawn in Fig.
the product of a coefficient listed in Table 6-1 multiplied by the
dimensions of the circumscribing rectangle, and that its centroidal
location is the product of a similar coefficient times the base length of
this rectangle, we obtain as the moment of area with respect to the right
end C
[(area)^^ = Sax]
[(area^c-Xc

= 700 N-m3 Arts.

The symbol xc means that moment arms are to be measured from C.

From Theorem II of the area-moment method, this result represents the
product EItc/A where tc/A is the deviation of point
, C from a tangent
drawn to the elastic curve at point A .

One final observation: It is usually unnecessary to compute the

beam bending moment at C must be
reactions. Since the conventional
zero (because no loads act to the right of C to cause bending moment),
the moment of the reaction R at C (equal to 3R ) must balance the
x X

bending moment of —600 N-m at C caused by the loading.

By now the reader may realize that the technique of drawing
moment diagrams by parts is really a graphical interpretation of the
method of writing a general moment equation described on page 217.
Thus for the beam shown in Fig. 6-15, the general moment equation is

M - R x + M (x - a
t x x
}° - -^<x - a2 >
2
- g<x - *3 > 3

If each of the terms in pointed brackets is replaced by a new

variable that starts at the position where each loading begins, we see
that each term in the general moment equation is the moment equation
of a cantilever type given in Table 6-1. Thus replacing <x — a } by
x
w,
<x — a 2 } by t>, and <x — a 3 > by z, we obtain

M= Rx +
x
M u° - -^-v
x
2
2
- ttz 3
6b
236 6 Beam Deflections

u
P <*3
«
* a
o »
*

^ 02 ^
M>2

N/m
u>i N/m
i

Figure 6-15.

ILLUSTRATIVE PROBLEMS

622. For the beam shown in Fig. 6- 16a, compute the moment of
area of the moment diagram about the left end.

Solution: From statics, the reactions are computed to be /?, = 300 N

and R2 = 600 N. The moment diagram by parts is constructed by
applying the definition M — (LM) L
to segments AB and BC, and the
definition M = (2M) R segment CD. Observe that it is simpler to
to
compute the bending moment for an exploratory section in segment CD
by taking moments of forces to the right of the section rather than to
the left. Computing bending moment by applying either — (ZM) L or M

600 N/m
5 m
M = 300 X 5
A^OON =1500N-m

1 m
= 3 X600 600 N/m
W
(cK

M = -900 X 1
= -900Nm

2 m "\3 r

3rd degree curve

(b)
M = 600
= 600N-m
nJF\ fl
2
= 600N

Equivalent cantilever loadings

Figure 6-16.
6-4 Moment Diagrams by Parts 237

M = (2M)R will generally indicate the simplest manner of drawing the

moment diagram by parts.
The equivalent cantilever loadings are shown in Fig. 6- 16c. The
bending moment diagram in Fig. 6- 16b is checked by the fact that the
moment at C in terms of the forces to the left of C is 1500 - 900 = 600
N-m, which equals, as it should, the moment at C expressed in terms
of the forces acting to the right of C.
The moment of the area of the M
diagram about the left end A is
now computed as equal to the sum of the moments of area of its parts.
Referring to Table 6-1 for the coefficients of area and of centroidal
distances,we obtain:
[(area)^-*, - I, ax]

,
(area),.
, _
xA =
/

—
1500 x 5\/2
(— - X )(
v
_\
s) +
/600X1W
(-j—ft
+
i

yXl

-FP^M
= 11.13kN-m 3 Ans.
What value of Eh does this result represent?

623. For the overhanging beam shown in Fig. 6- 17a, compute

the moment of area about C of the moment diagram included between
the supports at A and C.

Solution: At any section between A and C, the conventional bending

moment computed more easily by applying
is — (2M) L whereas M ,

between C and D it is simpler to apply = (ZM) R The momentM .

diagram by parts shown in Fig. 6- 17b is therefore constructed by

combining the cantilever loadings in Fig. 6- 17c.

3 m

M = 3R 1

2m
lm 800 N/m
<>

M = - (800 X 2) X 1
= -1600 N-m
800 N/m
M = -(800X 1) X0.5
= -400 N-m
2nd degree
curve lm
-1600 N m
•
Equivalent cantilever loadings
(b)

Figure 6-17.
238 6 Beam Deflections

In this problem, the value of the reactions need not be computed.

The moment at C caused by R
found from the fact that the bending
l
is

moment at C of all forces to the left of C must equal the bending

moment of all forces to the right of C, which is in accord with the
fundamental definition M = (1 M) L = (2M) R In other words,
I . 3/?, —
1600 = - 400; hence 3/?, = 1200 Nm.
We obtain the moment of area of the M diagram between A and
C about C by applying

[(area),4 C = Sax]
xc

——
•

x
(arca)^-
_
xc =
/3xl200\/3\
-) -
/2xl600\/2\
—3—4
= 1270 Nm- Ans.

PROBLEMS

For each of the beams in Problems 624 to 629, compute the

moment of area of the M
diagram between the reactions about both the
left and the right reaction.

624. Beam loaded as shown in Fig. P-624.

Arts, (area)^ • xA = 2500 N m•
3

900 N
1Af = 600N-im
2m m

~ji
R.

Figure P-624.

625. Beam loaded as shown in Fig. P-625. (Hint: Draw the

moment diagram by parts from right to left.)

600 N

3 m
300 N/m
I
2m
A B

Ri
Figure P-625.
6-4 Moment Diagrams by Parts 239

626. Beam loaded as shown in Fig. P-626.

Arts. (area)^ •
xB = 8.25 kN-m3

400 N/m
A
* k 1 m 3 m 1 m i

*1
Figure P-626.

627. Beam loaded as shown in Fig. P-627. (Hint: Resolve the

trapezoidal loading into a uniformly distributed load and a uniformly
varying load.)

200 N/m

628. Beam loaded with a uniformly varying load and a couple as

shown in Fig. P-628. Arts, (area)^ xA • = 2.13 kN-m 3

lkN/m
M = 2kN-m
|T*1

Figures P-628 and P-629.

629. Solve Problem 628 if the sense of the couple is counterclock-

wise instead of clockwise as shown in Fig. P-628.
240 6 Beam Deflections

630. For the beam loaded as shown in Fig. P-630, compute the
value of (area)^ •
xA From this result determine whether the tangent
.

drawn to the elastic curve at B slopes up or down to the right. {Hint:

Refer to Eq. (6-5) and Fig. 6-10.)
A ns. (area)^ xA = - 463 3
slope is down to right
' Nm ;

200 N
400 N/m
B

1 m 3 m 2m

*i R2
Figure P-630.

631. Determine the value of P for the beam loaded as shown in

Fig.P-631 so that the moment of area about A of the diagram M
between A and B will be zero. What is the physical significance of this
result?

800 N/m

lm k 3m

*i
Figure P-631.

632. For the beam loaded as shown in Fig. P-632, compute the
value of (area)^ xA From this result, is the tangent drawn to the
•
.

elastic curve at B directed up or down to the right? {Hint: Refer to Eq.

(6-5) and Fig. 6-10.) Ans. (area)^ xA = 1.27 kN-m 3 up to right •
;

200 N

800 N/m
1
2m 1 m kB 1

Figure P-632.
6-5 Deflection of Cantilever Beams 241

IP

Figure 6-18.

6-5 DEFLECTION OF CANTILEVER BEAMS

It will be recalled that the tangential deviation at any point is the

distance from the point on the elastic curve to a tangent drawn to the
curve at some other point (Art. 6-3 and Fig. 6-8). As a consequence,
the tangential deviation is generally not equal to the deflection. In
cantilever beams, however, the wall is usually assumed to be perfectly
fixed, and hence the tangent drawn to the elastic curve at the wall will
be horizontal, as in Fig. 6-18. Therefore, if the tangential deviation at A
is measured from a tangent drawn at B, the deviation t
A/B will equal the
deflection 8A dXA.
Several examples will illustrate iiow area-moment principles are
used to determine slope and deflection in cantilever beams. Other types
of beams are considered later.

ILLUSTRATIVE PROBLEMS

633. For the cantilever beam in Fig. 6-19, it is assumed that

E= 12 GN/m 2 , / = 10 X 6
10 mm 4 . What value of P will cause a
20-mm deflection at the free end?

400 N

-2PNm

1200 Nm
Figure 6-19.
242 6 Beam Deflections

Solution: The moment diagram by parts is drawn as shown. Evidently

the deflection 8 at A is numerically equal to the deviation of A from a
tangent drawn at C. Since the deviation at A is negative because A lies
below the tangent, we have from Theorem II,

1
1
a/c = ^(area^c
/ x
xA •

1
f
/2X2PW 2 \ /3X1200W2 \l

whence
£75 = (4.667P+ 3600) Nm 3

When substituting numerical values into the above result, it must

be remembered that, for dimensional homogeneity, E must be expressed
in N/m 2 , / in m4 , and 8 in m. Therefore, upon substitution, the above
equation becomes
(12 X 9
10 )(10 X 10- 6 )(20 X 10" 3 ) = 4.667 P + 3600
whence
/>= -257 N Arts.

The minus sign of P indicates that the direction of P must be opposite

to that originally assumed; that is, P must act upward.

634. Compute the maximum slope and deflection for the canti-
lever beam shown in Fig. 6-20a which carries a load varying uniformly
from zero at the wall to w N/m at the free end.
Solution: Although the given cantilever loading is not one of the types
given in Table 6-1, it is easily transformed into them by replacing the
given loading in Fig. 6-20a by those shown in Fig. 6-20c, i.e., superpos-
ing a downward uniformly distributed load and an upward uniformly

8 = -t

3rd degree
wL2
curve x 6

L (0
2nd degree
curve (b)^\
T wL 2

F Igure 6-20
 .

6-5 Deflection of Cantilever Beams 243

varying load. Thus we obtain the parts of the moment diagram shown in
Fig. 6-20b.
The elastic curve in Fig. 6-20a shows that the maximum slope and
deflection occur at the free end A. The angle 9A is clearly equal to the
change in slope 9AB measured between the tangents drawn to the elastic
curve at A and B. Then by Theorem I we obtain
= 1 wL 2 1 wL 2 '

*A = 9AB = area ^*
lF7 4* L " 3* L
itf( 6 2

wL 3 -

SEI
According to Fig. 6-11 on page 23 1 the minus sign means that the angle
is measured in a clockwise sense from the left tangent at A to the right

tangent at B; hence the slope at A is up to the right as shown.

The maximum deflection 8 at A is numerically equal to tA / B9
which is the deviation of A from a tangent drawn at B. Since the
deviation at A is negative because A lies below the tangent, we have
from Theorem II

1
( TQ \
~~ X
gj \& &)aB A
t
A/B '

which reduces to

Remember that the symbol xA means that the moment arms of the
areas are to be measured from A

635. For the cantilever beam loaded as shown in Fig. 6-2 la,
compute the value of EI8 at A and at B.

Solution: The deflections at A and at B are numerically equal to the

deviations of A and B from the horizontal tangent drawn to the elastic
curve at C. Since A and B are both below this tangent, the deviations
A/c and tB/c are both negative. Applying Theorem
t II, we have

- *a " *a/c " ^(area^c xA • (a)

and

~ 8b = f
B/c = ^(area)^ xB • (b)

Observe that in Eq. (a) we shall use the diagram between A and M
C, whereas in Eq. (b) we need only the diagram between B and C. M
 244 6 Beam Deflections

( k>\M= -300X3 =-900 N-m

5, f
V=300N
2 m
% 4 m
N-m ^f
M=300X4 = |V= 300 N
y* 1200 N-m

ft 4 m
-900 N-m
K)N.m

900 N-m ^ N/m

150

300 X 1 = ^2m 2 m
300 N-m M = -300xl - -300 N-m

2 m Bs- -300 N-m

M=2x300^
E 2 m
= 600 N-m' Tv= 300 N
'//

S& 2m
-900 N-m
3f=-900N.m
150 N/m

2 m 2

itf=(150X2)Xl =
m 1
300 N-m
150 N/m
iUUIUUg
4m "R^
M= -(150 X 4)X 2 =
-1200 N-m -1200 N-m
Figure 6-21. Variations of moment diagram by parts.

Let us examine first the various ways in which the M diagram by parts
can be drawn so that the simplest diagram be used.
First, the M
diagram in Fig. 6-2 lb is drawn by expressing the
bending moment at any section in terms of the loads to the right of the
section in accordance with the basic definition M
— (I,M) R The equiv- .

alent cantilever loadings are also shown.

Second, the M
diagram can also be drawn by expressing the
bending moment between A and B by the definition = (2A/) L and M ,
6-5 Deflection of Cantilever Beams 245

that between B and C by M

= (EM) R This results in the . M diagram
and equivalent cantilever loadings shown in Fig. 6-2 lc.
Third, still another method is to replace the given loading by
superposing the loadings in Fig. 6-2 Id. This gives what is probably the
simplest M
diagram by parts.
Although identical results will be obtained by using any of the
aforementioned diagrams, we shall apply Eq. (a) to the third one (Fig.
6-2 Id). Noting that xA means to take the moment of area about A, we
obtain

-&A- J_
EI
which reduces to
mt-H-i^HH
EI8A = 4100 Nm 3
Ans.
Since the deflection at B is expressed in terms of (area) BC either ,

of the M diagrams in part (b) or (c) of Fig. 6-21 may be used, for they
are identical for segment BC. Taking moments about B as indicated by
xB in Eq. (b), we obtain

- s
»=Ti
from which we have
B^XH - < 2
* ™>(H
EI8B m 400 - 1800 or EI8B = 1400 Nm 3
Ans.

Although not shown in fig. 6-21, a fourth way to draw the M

diagram by parts from B to C is to use the concept discussed in Art. 4-3
on page 125: a beam may be cut at any section and the effect of the
loads to one side of the section replaced by the shear and moment at the
cut section. At B, the load causes a shear force of —300 N and a
bending moment couple of —300 N-m whose effects from B to C
produce a triangular and a rectangular moment diagram. The student
may well sketch these diagrams and use them to compute 8B as a check.

PROBLEMS

636. The beam shown in Fig. P-636 has a rectangular

cantilever
cross section 50 mm wide by h mm
high. Find the height h if the
deflection is not to exceed 10 mm. Use E — 10 X 10 N/m
9 2
maximum .

Ans. h = 328 mm
2kN KIN

1.5 m

Figure P-636.
246 6 Beam Deflections

637. For the beam loaded as shown in Fig. P-637, determine the
deflection at 2 m from the wall. Use E = 10 X 109 N/m 2 and / = 20 X
10
6
mm 4 . Ans. 8 = 22.7 mm
800 N/m
3 m
Figure P-637.

638. For the cantilever beam shown in Fig. P-638, determine the
value of EI8 at the left end. Is this deflection upward or downward?
Ans. EI8 = 6.67 kN-m 3 upward ;

250 N
M=2kN-m
2m \\
2m

•IT
Figure P-638.

639. A distributed load and an upward-concentrated load act as

shown on the cantilever beam in Fig. P-639. Compute the amount the
free end deflects upward or downward given that E = 10 X 10 N/m
9 2

and / = 60 X 106 mm4 .

3kN/m
1 m 4 3 m

4kN
Figures P-639 and P-640.

640. Compute the value of 8 at the concentrated load in Problem

639. Is the deflection upward or downward?
Ans. 8 = 46.9 mm; downward
641. For the cantilever beam shown in Fig. P-641, what value of
P will cause zero deflection at Al Ans. P= 150 N

Af=400N.m
^TLB CI
1.5 m mrr
Figure P-641.
6-5 Deflection of Cantilever Beams 247

642. Find the maximum deflection for the cantilever beam

loaded as shown in Fig. P-642 if the cross section is 50 mm wide by 150
mm high. Use E = 69 GN/m 2 . Arts. 8 = 28.0 mm
4kN/m
2m lm
Figure P-642.

643. Find the maximum value of EI8 for the cantilever beam
shown in Fig. P-643. Arts. EI8 = 2
(Pa /6)(3L - a)

\P
a
I
Figure P-643.

644. Determine the maximum deflection for the beam loaded as

shown in Fig. P-644.

w N, m
L
2

Figure P-644.

645. Compute the deflection and slope at a section 2 m from the

wall for the beam shown in Fig. P-645. Assume that E = 10 GN/m2
and / = 30 X 10
6
mm 4
. Arts. 8 = 16.4 mm; = 0.739°

1200 N/m

646. For the beam shown in Fig. P-646, determine the value of /
that will limit the maximum deflection to 20 mm. Use E = 10 X 10
9

N/m 2
.

600 N/m

1 m 2 m 1 m

Figure P-646.
248 6 Beam Deflections

647. Find the maximum value of E18 for the beam shown in Fig.
647. Ans. EI8 = (121/1920)wL 4
w N/m

L
2
Figure P-647.

648. For the cantilever beam loaded as shown in Fig. P-648,

determine the deflection at any section x meters from the support.
Ans. EI8 = (wx 2 /120L)(10L 3 - 10L 2 * + 5Lx 2 *
3
)

u/N/m

6-6 DEFLECTIONS IN SIMPLY SUPPORTED BEAMS

Deflections in cantilever beams were simplified by the fact that the

tangent to the elastic curve at the fixed end was known to be horizontal.
In simply supported beams, the position at which a tangent to the
elastic curve will be horizontal is usually unknown, and therefore a

different method must be used. This method may seem devious, but
actually it is simple and rapid. It is illustrated by Fig. 6-22, which shows
only the elastic curve of a simple beam. The loads and moment
diagrams have been omitted for clarity.

C/A

Figure 6-22. Geometry of area-moment method applied to simple beams.

6-6 Deflections in Simply Supported Beams 249

The problem is to determine the value of the deflection 8 at some

position B. If a tangent to the elastic curve is drawn at A, the deviation
tB/A at B from this tangent is evidently not the required deflection 8.
However, the sum of 8 and tB/A constitutes the distance EF; and if both
EF and tB/A were known, 8 could easily be found. Hence the distance
EF must also be found. This is done by noting that the triangle AEF is
similar to the triangle ACD, of which the leg CD equals the deviation

*c/a °f C from the reference tangent drawn at A .

The proper procedure to apply is obtained by reversing the steps

of the above analysis into the following order:

1. Compute tc/A using , the relation

1
( \
t
C/A ~ ~£yA area ,Jc,4 * XC

2. From the relations between similar triangles, determine EF in

terms of tc/A . We obtain

EF-j-tc/A

3. Compute tB/A from the relation

1 .
t
t
B/A " -£j\* Te *)BA ' XB

4. Since EF is the sum of 8 and tB/A the value of 8

, is given by

8 = EF- t
B/A

As mentioned previously, this procedure may seem long, but

actually it is rapid. Several examples will demonstrate the method. Only
simple loadings are used, since this concentrates attention upon basic
ideas.For more complex loadings the basic method is unchanged, the
more complicated M
diagram by parts being constructed as indicated in
Art. 6-4.

ILLUSTRATIVE PROBLEMS

649. The simple beam in Fig. 6-23 supports a concentrated load

of 300 N at 2 m from the left support. Compute the value of EI8 at B,
which is 1 m from the left support.
Solution: We begin by sketching the dashed outline of the elastic
curve and drawing the M diagram by parts from left to right. Following
250 6 Beam Deflections

300 N

lm I 1

300 N- m
Figure 6-23.

the procedure just discussed, we first obtain tc/A :

'c/a =^7(area) c ^-xc

j

(3 X 300\/ 1 A / 1 X 300\/ 1 ,\ 400

f
C/A ~ £I EI
(a)

Since triangle ABD is similar to triangle ACE,

1 400
BD
r>r,
= -X t r/A
C/A = ^r-— (b)
3 3 EI

The deviation tB/A is next obtained from

I 1 X 100W 1 A
XB X1
h/A ~ £.^(
area )fi/l * l
B,A ~ Ej
(^)(3 )

100
6EI (c)

Finally, the value of 8 is given by

400 100
[8=BD = t
B/A ] 8 = J_
EI 3 6

EI8 = 116.7 Nm 3
Ans.
6-6 Deflections in Simply Supported Beams 251

650. Compute the slope of the elastic curve at the left reaction
for the beam discussed in the preceding problem and shown in Fig.
6-23.

Solution: The A is given by tan A where

slope of the elastic curve at ,

9A is and the tangent drawn to the

the angle between the horizontal
elastic curve at A (Remember that the deflections and slopes in beams
.

are assumed to be very small compared to the length of the beam.)

Hence, tan A is practically equivalent to A expressed in radians.
From Fig. 6-23, therefore, we obtain

'A ~ "*" VA
AC AC
whence, substituting the value t
c/A = 400/ EI from Eq. (a) in the
preceding problem, we have
400/£7 400
•a 3— ~JeI Am '

651. For the beam described in Illustrative Problem 649, locate

the position of maximum deflection and compute the maximum EI8.

Solution: The required values can be determined by either of two

methods. Both methods should be mastered because sometimes one is
easier than the other.

Method I. We begin by computing the deflection at any position

B located x meters from the left reaction. Using the technique applied
in Illustrative Problem 649, we obtain from Fig. 6-24,

1 /
*B/A =^7( area)/M-*J *B/A
-i[(i-H(!)]
IOO'jt
6EI
Using the value of tc/A given in Eq. (a) of that problem, we also
have in Fig. 6-24,
x
nn = — 400 400jc
**>-J'*c/a BD
3 EI 3EI
The value of the deflection 8 is the difference between BD and
t
BfA Hence the equation of the elastic curve, with deflections 8 consid-
.

ered positive downward, is

400x _ l00x3
[8-BD-tB/A ] 5 = _
EI 3 6
or

EI8
_
= —
400* —
100* 3
(a)

This is valid for any position between the left reaction and the load
252 6 Beam Deflections

300 N
2 m »4-* 1 m -

1* \

/?, = 100 N

300 Nm

-300N-m
Figure 6-24.

(that is, between x — and x — 2). The slope equation for this portion
of the beam is found by differentiating the deflection equation, Eq. (a):

rj
EI
d8
Tx=
— 400 300x 2
—r— , LS
K b)

From the principle of maxima and minima developed in calculus,

setting the first derivative of Eq. (a) equal to zero will determine the
position of maximum deflection. This is equivalent to setting Eq. (b)
equal to zero, thus determining the position of zero slope in the beam.
We obtain
2
400 300.x
= or x = 1.63 m Arts.
3 6
Substituting this value of x in Eq. (a) determines the maximum deflec-
tion.

Max. EI8 = ^(1.63) - ^(1.63) = 145

3
Nm 3
Ans.

It is instructive to compare the maximum value of EIS with the

value of EI8 at midspan. At midspan, x = 1.5 m, and substitution of
this value of x in Eq. (a) gives

Midspan EI8 = ^(1.5) - ^(1.5) 3 = 144 N-m 3

3 6
6-6 Deflections in Simply Supported Beams 253

This is about 1% less than the maximum value. This difference is so

negligible in comparison with possible variations in the given data that
for all practical purposes we may assume the midspan deflection to be
equivalent to the actual maximum deflection. Indeed, it may be shown
that, for a concentrated load located anywhere on a simple span, the
maximum difference between midspan and maximum deflection is only
2.6%. In the next article we method for computing
discuss a simplified
midspan deflection.

Method II. At the position of maximum deflection, the tangent

to the elastic curve will be horizontal. As shown in Fig. 6-24, the
change in slope between tangents at this position B and at A (i.e., AB ) is
equal to the slope 9A at A, since for small angles the radian measure and
the tangent of the angle are practically equivalent, that is,
A — tan A .

From Theorem I of the area-moment method, we obtain

50x'
°AB = area ^* 100 *)"
£/( ^-ir(f*- EI
which, on being equated to A = 400/3 EI obtained in Problem 650,
gives

50x 2 400
or x = 1.63 m Arts.
EI 3EI
Computing the value of Eld at this position gives maximum
EI8 = 145 N-m3 , as was obtained with Method I.

652. Determine the value of the deflection at D of the beam

shown in Fig. 6-25a.

Solution: This problem brings out the importance of correctly inter-

preting the meaning of positive and negative deviation, particularly as it

affects thegeometry of the elastic curve. For illustrative purposes, this

—
problem will be solved in two ways first, by drawing the reference
tangent at C; and second, by drawing the reference tangent at A.
Before drawing the reference tangent at C, we compute the devia-
tion of A from this reference tangent. Correctly interpreting the sign of
t
A/c wm< indicate the direction in which the reference tangent slopes.
The geometrically correct position of the reference tangent can then be
drawn. Thus

A/C ~~
£T\ ai * AJAC *A

-^[(£X3X900)(§X3)
2 Jil 367
-/Ix2x lOOoVl +|x2Jl = EI
254 6 Beam Deflections

D/C

-1000N-m
(b)

Figure 6-25.

The positive value of tA C means that A on the elastic curve lies above
/
the reference tangent at C. Hence the reference tangent at C slopes
down to the left, as shown.
From the similar triangles ACE and CDF, we obtain

DF
13 D A/C A/C 367
l

or DF =
3 3EI
The deviation of from the reference tangent at C is

1X100W2 \1
~ area )z>c XD xi
1
d/c ~F?
EI (
'
D/C
-M-( -iH(3 )J

100
3EI
The minus sign for t
D/c means that D' on the elastic curve is below the
 "

6-6 Deflections in Simply Supported Beams 255

reference tangent. Also, since DF is numerically greater than the ab-

solute magnitude of tD/c , it is now apparent that D is deflected upward
from its original position. Therefore, the elastic curve between C and D
is sketched as shown.
The deflection 8 is obtained from
367 100 j*9
8 = DF-\tD/C\ ~ "TTT Ans.
3EI 3EI EI
If the reference tangent had been drawn at A and the elastic curve
had been assumed to have the shape shown in Fig. 6-26, we would have
obtained the following results:

*C/A ~~ I \& TQ &)CA ' XC

EI
j? T

*D/A ~"
1

EI
1
miH-^xn
3 X 900

area )zM * XD
683
EI

gj (
3 X900
-Mi 2
100
)(-H-F^)("H
-(
1000
KM]
EI
The similar triangles ACE and ADF give
DF l
C/A ^
DF
4 4/683\ 911
4
or
-V c'A -l\El)- EI
Finally, using the elastic curve sketched in Fig. 6-26, we obtain the
deflection
89
8 = DF- ^=-(911-1000)=--
t
1

which, except for the minus sign, is the value obtained previously. Here
the minus sign indicates that the deflection at D is opposite to the
direction sketched in Fig. 6-26; i.e., it is directed upward as before. This
is checked by the fact that tD/A is numerically larger than DF.

Figure 6-26. Elastic curve of Problem 652 with reference tangent drawn at A.
256 6 Beam Deflections

PROBLEMS

653. Compute the midspan value of EI8 for the beam shown in
Fig. P-653. (Hint: Draw the M
diagram by parts, starting from mid-
span toward the ends. Also take advantage of symmetry to note that the
tangent drawn to the elastic curve at midspan is horizontal.)
Ans. EI8 = 3350 N-m 3

600 N/m 600 N/m

2 m lm 2m

«1
Figure P-653.

654. For the beam shown in Fig. P-654, find the value of EI8 at
1.0 m from R 2. (Hint: Draw the reference tangent to the elastic curve at
R2 .)

600 N/m

2m 7m A

*i
Figure P-654.

655. Find the value of EI8 under each concentrated load of the
Ans. 608 N-m 850 N-m3
3
beam shown in Fig. P-655. ;

200 N 500 N

lm { lm { 2 m

A'.
*i
Figure P-655.

656. Find the value of EI8 at the point of application of the 100
3
N-m couple in Fig. P-656. Ans. EI8 - 342 N-m

500 N
M =100N-m
1 m 2 m 1 m

•n
Figure P-656.
6-6 Deflections in Simply Supported Beams 257

657. Determine the midspan value of EI8 for the beam shown in
Fig. P-657.

800 N/m

3m
P
Figure P-657.

658. For the beam shown in Fig. P-658, find the value of EI8 at
the point of application of the couple.
Arts. Eld - (Ma/3L)(L 2 - 3La + 2a 2)

M
/5=
yf777t~
T7777T

Figure P-658.

659. A simple beam supports a concentrated load placed any-

where on the span, as shown in Fig. P-659. Measuring x fromv4, show
that the maximum deflection occurs at

-b-
-f
I
B
I J
Figure P-659.

660. A simply supported beam is loaded by a couple M

at its
right end, as shown in Fig. P-660. Show that the maximum deflection
occurs at x = 0.577 L.

M
ABB m?
Figure P-660.
258 6 Beam Deflections

Compute
661. the midspan
deflection of the symmetrically
loaded beam shown P-661. Check your answer by letting
in Fig.
a — L/2 and comparing with the answer to case 6 in Table 6-2 on page
270. Also compare one-half your answer with the midspan deflection of
case 7 in Table 6-2.

1_1
F
J?!
1
R2
Figure P-661.

662. Determine the maximum deflection of the beam shown in

Fig. P-662. Check your result by letting a = L/2 and comparing with
case 8 in Table 6-2. Also use your result to check the answer to
Problem 653.

w N/m |
\w N/m
U-a-*| \*-a-*
Lr _

Figure P-662.

663. Determine the maximum deflection of the beam carrying a

uniformly distributed load over the middle portion, as shown in Fig.
P-663. Check your answer by letting 2b — L and comparing with case
8 in Table 6-2. Arts. EI8 - (wb/24)(L 3 - 2Lb 2 + b
3
)

* a-**j-» 26—*f*- a ~^-

I - N/m |

\ i

^ r
XJ

Figure P-663.
 6-6 Deflections in Simply Supported Beams 259

664. The middle half of the beam shown in Fig. P-664 has a
moment of inertia twice that of the rest of the beam. Find the midspan
deflection. {Hint: Convert the M diagram into an M/ EI diagram.)
Arts. 8=\(Pa 3 /EI)

i 21

Pr^Tf
Figure P-664 and P-665.

665. Replace the concentrated load in Problem 664 by a uni-

formly distributed load of w N/m acting over the middle half of the
beam. Find the maximum deflection.

666. Determine the value of EI8 at the right end of the overhang-
ing beam shown in Fig. P-666. Ans. EIS = (wb 3 /24)(4a + 36)

w N/m
/}//}/ OOP

Figure P-666.

667. Determine the value of EIS at the right end of the overhang-
ing beam shown in Fig. P-667. Is the deflection up or down?

200 N
400 N/m
B
3 m 2m
F
Figure P-667.

668. For the beam shown in Fig. P-668, compute the value of P
that will cause the tangent to the elastic curve over support R2 to be
horizontal. Ans. P=1350N
260 6 Beam Deflections

300 N

1 m 2 m 2 m
i

T
Figure P-668.

669. Compute the value of EI8 at the left end of the beam shown
in Fig. P-669.

2kN/m 2kN/m

2m 2m 2m

Figure P-669.

670. Determine the value of EI8 at the left end of the overhang-
ing beam shown in Fig. P-670. Arts. EIS = 428 N -m 3 down

M = 600N-m
900 N/m

6-7 MIDSPAN DEFLECTIONS

In a symmetrically loaded simple beam, the tangent drawn to the elastic

curve at midspan is horizontal and parallel to the unloaded beam. In
such beams, the deviation at either support from the midspan tangent is

equal to the midspan deflection.

For simple beams that are unsymmetrically loaded, the midspan
deflection can be found as easily as for a symmetrically loaded beam.
All that need be done is to add a symmetrically placed load for each
load actually acting on the beam. The effect of this transformation to
symmetry is to double the actual midspan deflection. In other words,
6-7 Midspan Deflections 261

the actual midspan deflection is equal to one-half the midspan deflec-

tion of the transformed symmetrically loaded beam. Note that there is
so little difference between midspan deflection and the actual maximum
deflection that practically the two values may be considered equivalent.

ILLUSTRATIVE PROBLEMS

671. A simply supported beam L meters long carries a load that

varies uniformly from zero at the left end to w N/m at the right end, as
shown in Fig. 6 -27a. Determine the midspan deflection.

Solution: Create symmetry, as shown in Fig. 6-27b, by adding a load

that varies uniformly from zero at the right end to N/m at the left
w
end. The result is a uniformly distributed load of w N/m over the entire
span. The deviation at C from the midspan tangent drawn at B is equal
to 28 or twice the actual midspan deflection in Fig. 6-27a.
Since we are considering the deviation of C from the midspan
tangent drawn at B, we need the moment diagram of only half the
beam. This M
diagram may be drawn by parts from C to B as shown in
Fig. 6-27c or from B to C as shown in Fig. 6-27d. To facilitate
understanding of Fig. 6-27d, the free-body diagram of the right half of
the beam is also shown adjacent to it. Note that the midspan moment M
is found from the fact that its effect at the right end must be equal and

w N/m- w N/m

(a)

wL
2

w
k
N/i
M\

V-0 wL

Figure 6-27. Transformation to symmetry.

 6 Beam Deflections

opposite to that of the load in order to give a resultant bending moment

of zero at the end of the beam.
We shall apply the area-moment theorem to the M diagram in
part (d), leaving the student to verify that using part (c) yields the same
result. We obtain

whence

EI8 = -—wL 4 Ans.

/DO

672. Determine the midspan value of EIS caused by the loads

shown in Fig. 6-28a.

Solution: The transformation to symmetry is shown in Fig. 6-28b.

Since we shall need the M diagram for only half of the beam, this is
shown for the left half in Fig. 6-28c, together with the corresponding

400 N

600 N/m
(a) Original loading
3m 1 m 2 m

*i
800 N

(b) Transformation
to symmetry

2400 N-m

-1200 N-m

M= 2400 N-m
1200 N-m
1600 N
(c)

Figure 6-28. Midspan deflection.

6-7 Midspan Deflections 263

free-body diagram. Note that the midspan moment is found as de-

scribed in the preceding problem. Also note that the reaction in the
transformed beam is equal to one-half the symmetrically applied loads;
that is, it is equivalent to the sum of the original loads.
deviation tA/B of A from the horizontal tangent drawn at B is
The
equal to twice the midspan deflection of the original loading. We obtain

[EItA/B = (area)^-**]

2£/8 = (2400x3)(ix3)-(^00)(ix2)
(3 X 1200 \( A

which reduces to
-HHU x3
1

EI8 = 4300 Nm 3
Ans.

PROBLEMS

673. For the beam shown in Fig, P-673, show that the midspan
deflection is 8 = (Pb/4SEI)(3L 2 - 4b 2).

a f
f b

rj* L nR>
Figure P-673.

674. Find the deflection midway between the supports for the
overhanging beam shown in Fig. P-674.

_ a
1

Figure P-674.

675. Repeat Problem 674 for the overhanging beam shown in

Fig. P-675. Ans. 8 = wa 2L 2 /32EI

w N/m
a jfrdf L
Figure P-675.
264 6 Beam Deflections

676. Determine the midspan deflection for the simply supported

beam loaded by the couple shown in Fig. P-676.

3L
4

Figure P-676.

677. Determine the midspan deflection for the beam loaded as

shown in Fig. P-677.

Figure P-677.

678. Determine the midspan value of EI8 for the beam shown in
Fig. P-678. Arts. EI8 = 3100 N-m3

900 N 600 N

3m I 1 ™!*

Figure P-678.

679. Determine the midspan value of EI8 for the beam shown in
Fig. P-679 which carries a uniformly varying load over part of the span.
Arts. EI8 = 2940 N-m 3

900 N/m^^/
2rm
i
R x
3 m
6 m
2

3 R2
Figure P-679.
6-8 Conjugate-Beam Method 265

680. Determine the midspan value of EI8 for the beam loaded as
shown in Fig. P-680.

400 N
M = 600N-m
lml 2m
2 m 1 i

Figure P-680.

2
Show that the midspan value of EI8 is (wb/4$)(L - 2Lb
3
681.
3
+ b ) for the beam in part (a) of Fig. P-681. Then use this result to find
the midspan EI8 of the loading in part (b) by assuming the loading to
extend over two separate intervals that start from midspan and adding
the results. Ans. EI8 = 9280
3
Nm

w N/m 800 N/m

2m 3 m lm 1

f 3Ro f
R
*2 i

(a) (b)

Figure P-681.

6-8 CONJUGATE-BEAM METHOD

Successive differentiation of the deflection equation discloses the follow-

ing relations:

Ely - deflection

EI^- = slope
dx

EI
d^ = moment = M
dx 2
3
^r d y = = rV, =
dM
EI—*z 3
.
shear —dxr-
dx
rT . dV = d M
—d*y— = loadA = -r- -
2

EI A
dx 4 dx dx 2
266 6 Beam Deflections

Area = -u\r

V//////////A
1
-^
«,N /m
B

Figure 6-29.

It is evident that the relations among deflection, slope, and moment are
the same as those among moment, shear, and load. This suggests that
the area-moment method can be used to determine bending moment
from the load diagram, just as deflection was obtained from the mo-
ment diagram. For example, in the load diagram in Fig. 6-29, the
bending moment at B should equal (area of load diagram)^ B x B That •
.

it does is seen from

wx
(-^)(2 X ) =
Thus we could apply area-moment principles to compute bending
moment, although this is impractical because better methods are availa-
ble.
Nevertheless, the similarity of relations among load, shear, and
moment, and among moment, slope, and deflection suggests that the
relations among moment, slope, and deflection can be found by using
the methods developed in Chapter 4 for computing shear and moment
from load diagrams. We need merely assume that a beam is loaded, not
with the actual loads, but with the M/EI diagram corresponding to
these loads. Treating this M/EI
diagram as a fictitious loading, we
compute the shear and moment at any point caused by this loading.
These fictitious shears and moments correspond to the actual slopes and
deflections in the beam at corresponding points. This technique is

known as the conjugate -beam method and sometimes as the method of

elastic weights.
Applying the principles of shear and moment to a beam loaded
with an M/EI diagram, we conclude that
1. The actual slope = the fictitious shear (6-6)
2. The actual deflection = the fictitious moment (6-7)

The method is especially useful for simply supported beams. For other

beams, such as cantilevers or overhanging beams, artificial constraints

must be applied; they are discussed later.
To evaluate the conjugate-beam method, let us compare it and the
area-moment method when applied to a simple beam. Only in simply
supported beams can the conjugate-beam method be applied directly
without using artificial constraints.
6-8 Conjugate-Beam Method 267

-« L *-

w N/m B (b) Moment diagram (by parts)

\2EI 2/1^3^
I
-« x H ^^ — *-

(a) Actual loading

(c) Conjugate-beam loading

Figure 6-30. Comparison of conjugate-beam and area-moment methods.

Figure 6-30a shows a simply supported beam carrying a distrib-

uted load of w N/m. The moment diagram (drawn by for this loading
parts in Fig. 6-30b) by \/EI and used as the conjugate-
is multiplied
beam loading shown simply supported on the span L in Fig. 6-30c. The
reaction R of this conjugate beam is found by setting moments of the
x

fictitious loads about B equal to zero. We obtain

The right-hand member of Eq. (a) will be recognized as (1/ EI)(a.rea) BA

•xB that is, tB/A Obviously, solving for R is equivalent to tB/A /L,
, .
x

which is the actual slope at A this fact is evident from the geometry of
;

the elastic curve in Fig. 6-30a. Nevertheless, this is confirmation of

rule 1 of the conjugate-beam method: The fictitious shear equals the
actual slope at the corresponding point in the actual beam.
To obtain the deflection at any point on the actual beam, we apply
the definition of bending moment to the conjugate loads:

Deflection v = (EM) L = R x - A x x j + A2 j
x x
= R.x -< AA l
A2 (*)
3 4

However, in terms of the moment diagram in Fig. 6-30b, [A^x/3) —

^2(*/4)] equals (l/jE7)(area) C/< • x c which equals
, t
c/A on the elastic
268 6 Beam Deflections

curve in Fig. 6-30a. Hence Eq. (b) may be rewritten as

y = R x - t c/A
x
(c)

which, since R x = 6x =
t
(tB/A /L)x, is equivalent to the following
area-moment relation:

y = (B/a(j) - 'C/A (d)

This is the result previously obtained in Art. 6-6 for deflections in

simple beams by the area-moment method.
Thus the conjugate-beam method, which uses the fictitious shears
and moments of an M/EI loading to determine actual slopes and
deflections, involves precisely the same computations as the area-
moment method but has the disadvantage of obscuring the physical
significances of the computations. This disadvantage is even more
pronounced when the method is used with cantilever and overhanging
beams, where certain artificial constraints must be applied. Neverthe-
less, the conjugate-beam method offers an occasional advantage in

certain routine work, in that it permits direct application of the defini-

tions of shear and moment to the fictitious loading to find slope and
deflection without any need of an elastic curve.
Now for a word about the need for artificial constraints in certain
cases. For the cantilever beam frr> Fig. 6-3 la, the M/EI diagram
appears as in Fig. 6-3 lb. This diagram cannot be applied directly as a
fictitious load to a cantilever with the wall at the right end C because
the fictitious shear and moment at B would be zero whereas the actual
slope and deflection at B are not zero. Therefore, the diagram of
fictitious loads must be modified, as in Fig. 6-3 lc, so that the fictitious

w N/m
B

(a) Actual loading and elastic curve 2EI

(b)
si diagram

£
_ wL
2EI
(c) Conjugate-beam loading
Figure 6-31. Constraints required to solve cantilever beams by the conjugate-
beam method.
6-9 Deflections by the Method of Superposition 269

loading will correspond to the actual slope and deflection at the free
end.
The reason for supplying artificial constraints when solving canti-
lever problems should now be clear. To produce an actual zero slope at
C in the original cantilever, the fictitious shear must be zero at C;
therefore

[V-(2Y) L ] 0= V- A

From this we see that the fictitious shear of the conjugate beam at B
must equal the area A M / EI diagram. Also, to produce a zero
of the
fictitious moment at C, we must calculate the restraint M from
fictitious

[Mc = (£M) L ]
= M+ VL- A j
Only after the artificial constraints M
and V have been found can
the fictitious shear and moment (corresponding to actual slope and
deflection) be computed. Hence, cantilever problems can be solved
more simply and more directly by the area-moment method.

PROBLEMS

Problems 653 to 665 inclusive and cases 6 through 12 in Table 6-2

(page 270) may be assigned for solution by the conjugate-beam method.

6-9 DEFLECTIONS BY THE METHOD OF SUPERPOSITION

In a supplementary method of determining slopes and deflections, the

results of a few simple loadings are used to obtain those for more
complicated loadings. This procedure, called the method of superposition,
determines the slope or deflection at any point in a beam as the
resultant of the slopes and deflections at that point caused by each of
the loads acting separately. The only restriction on this method is that
the effect produced by each load must be independent of that produced
by the other loads; i.e., each separate load must not cause an excessive
change in the original shape or length of the beam.
The technique of superposition is advantageous primarily for
loadings that combine the types in Table 6-2. For partially distributed
loads, the method of superposition requires integration (see Illustrative
Problem 683). In such cases, the double-integration method is prefer-
able; if the deflection at only one specific location is desired, the
area-moment method is generally best.
 Sroo
X
3

5 «N
MS
^2 I

V
z H
v
o

Z «
o I I

a;
Q ii

5 hs
£ z tt]

slla
O u ft, <N

ish^l Uh°*

t-
i n
i I

i I
I*
I
:: /
I •o 1
^ 2: n ^ *
/

/
/
/
I

H- •1

1 m
1

* rg§
I

h
o
<2

5Z
 -o

<N
1 m
l^J "5
1
?
I

II

X
OO "J
f> ON Ms
«J
o ^ A
2 ^ UJ

(S
\ kj II
«> II

•g
•C
5 to -
>
I

5 <
On
-©
ft,

H
<*5

cm cm

xH
V I

X
V
o
*J|-c I
|m IS
II

ii

^-v / V

<1 *q
| rf I

IN CM
s i < ho < m
£
ft.
9
ft.

5^5
On <N
II II II H

If
11 —
H -3

/
i

Eg: #s-
272 6 Beam Deflections

ILLUSTRATIVE PROBLEMS

682. Using the method of superposition, compute the midspan

value of EI8 for the beam carrying two concentrated loads, shown in
Fig. 6-32a.

Solution: From case 7 of Table 6-2, the midspan deflection of an

eccentrically placed concentrated load is EI8 = (Pb/4S)(3L 2 - 4b 2 )
where b is the smaller of the two segments into which the beam is
divided by the load. Resolving the given loading into those shown in (b)
and (c) in Fig. 6-32, we find that the midspan deflection of (a) is equal

300 N

1 m { 3m 1 2m

m
6
i
*i

(a)

300 N

4m I 2m

R\

(b)

600 N

1 m | 5 m

R'{ Ro
(c)

Figure 6-32.
6-9 Deflections by the Method of Superposition 273

to the sum of the midspan deflections of (b) and (c). Thus

Pb
EI8 = 2^(lL 2
-4b 2 )

= «[3(6) -4(2)V™[3(6) 2 2
-4(lf]
= 2450Nm 3
Ans.

683. A simply supported beam carries a uniformly distributed

load over part of its length, as shown in Fig. 6-33. Compute the
midspan value of EI8.

w = 600
P = wdx

N/m
i

/
m
= 600 dx

1 2m 4 m i

*1

Figure 6-33.

Solution: The continuous load may be considered as a series of

concentrated elemental loads, each of P = w dx = 600 dx and located
at a distance x from the end. At midspan, therefore, applying the result
for case 7, we have

En = 2 ^(3^ 2 - 4* 2 )

_ (-3 (600 dx)(x) 2 3 (600 dx)(x) 2

[3(6) -4x 2 [3(6) - 4* 2 ]
-i 48 ]+J 48

= 5063 + 2562 = 7625 N m •

3
Ans.

Two integrations are necessary —

one from to 3 for the right half
of the beam, and the other from 2 to 3 for the left half. This is true
because x, which replaces 6, must be the length of the smaller segment
into which P divides the beam.
If the given loading is divided into halves and each half is replaced

by its resultant of 1200 N acting as shown in Fig. 6-34, the sum of the
midspan deflections of these loads will be a good approximation of the
274 6 Beam Deflections

1200N 1200N
3 m—

M
1 1
+ ^% 1
l
i

7/
^M^^
I
'

b rti

R, R2
Figure 6-34.

actual midspan deflection. Applying the result for case 7 gives

- 8000 N-m 3
A closer approximation could be obtained by subdividing the
given loading into three or more parts. However, even with only two
subdivisions, the result is only about 5% larger than the correct value,
7625 N-m3 .

684. The overhanging beam in Fig. 6-35 carries a concentrated

load P at its end. Determine the deflection under P.

Solution: The tangent to the elastic curve at R2 is inclined at a very

small angle 9 with the horizontal, but the figure shows it greatly
exaggerated for convenience in representation. Imagine the original
position of the beam to coincide with this tangent and to be clamped in
this inclined position at R2 . Application of the loads R
and P will
x

produce the actual elastic curve. The deflections 5, and 8 2 produced by

these loads are similar to those in case 1. Theoretically 5, (and 8 2 also)
should be directed perpendicularly to the clamped tangent. However,
this perpendicular distance 6, cos 9 « 5, because 9 is extremely small.
In other words, there is no essential difference between a perpendicular
to the tangent and the vertical deflection 8 shown.
X

From the geometry of the elastic curve, we have

5, - 9a\ hence
a
Therefore the deflection at P is

v = 9b + 82 = 6,- 4- 82

whence, substituting 8 — PL 3 / 3 EI from case 1, we obtain

Pb 3
(« 2
Pb 1 ,
- Pb L
)
b Ptf_ LX
y = + (a + b)
3EI
Ans.
3 EI a 3EI 3 EI
6-9 Deflections by the Method of Superposition 275

Figure 6-35.

Actually 8 and 8 2 are deviations from a tangent drawn at

Y
R 2 and
,

thisprocedure duplicates that described in Illustrative Problem 652. The

difference is that here we used a formula from Table 6-2 instead of
computing deviations by using a moment diagram.

Alternate Solution:The rotation of the beam at the support R 2 may

alsobe found by dividing the beam in Fig. 6-35 into the two parts
shown in Fig. 6-36. The action of the overhang upon the portion
between the supports is replaced by a shear force P and a moment
M= Pb. The shear force is transmitted directly to the reaction R2 ,

whence the couple produces the effect of case 11. The rotation of the
beam at R 2 is given by
ML = (Pb)a
3EI 3EI

The deflection at the end of the overhang may now be found as in

the case of a cantilever beam (case 1) built in with an initial inclination
9. The total deflection is

Pb 3
v = 9b + 82 = Pba -b +
3EI 3EI

which, as before, reduces to

Pb 2 ,
LX Pb 2L
3EI 3EI
276 6 Beam Deflections

The deflection equation for any point between the supports, as

found from case 1 1 , is

Pbax
y = 6EI
K)
and, in terms of x measured from /? 2 , the deflection equation for the
overhanging portion is

Pba Px 2
y = 3EI x + 6EI (3b - x)

685. Two cantilever beams, having the same cross section and
made of the same material, jointly support a distributed load of w N/m
as shown in Fig. 6-37. Determine the force P at the roller between
them.

w N/m
=¥*

tP
Figure 6-37.

Solution: The force P may be determined by the condition that at B

both cantilevers have the same deflection. The deflection at B for the
lower cantilever is found from case 1 to be

6 = Pa
3EI

The upper cantileverloaded with a combination of cases 3 and

is

2, the resultant deflection at B being

* = i£i« L2 + °
2

-^-ih
Equating these deflections gives, for P,

w
P = -r^-(6L
2
4- a
2
- 4La) Arts.
16a
6-9 Deflections by the Method of Superposition 277

PROBLEMS
In solving the following problems, use Table 6-2.
j
686. Determine the value of EI8 under each concentrated load in
Fig. P-686. Ans. 575 N-m3 ; 767 N-m3

400 N 300 N

lm f lm } 2m

*i
Figure P-686.

687. Determine the midspan deflection of the beam shown in

Fig. P-687 if E = 10 X 109 N/m2 and/ = 20 X 106
4
mm .

Ans. 25.8 mm
2kN

lm f 3 m
lkN/m

*± in

Figure P-687.

688. Determine the value of EI8 at the left end of the beam
shown in Fig. P-688.

400 N/m

2m /^ 7
Figure P-688.
4^ ^^
278 6 Beam Deflections

689. The beam shown in Fig. P-689 has a rectangular cross

section 100 mm wide by 200 mm deep. Compute the value of P that will
limit the midspan deflection to 40 mm. Use E = 10 X 10 N/m
9 2
.

Arts. P = 7.07 kN

4 kN/m

1 ml 6 m
F
Figure P-689.

690. The beam shown in Fig. P-690 has a rectangular cross

section 50 mm wide. Determine the proper depth d of the beam if the
midspan deflection of the beam is not to exceed 20 mm and the flexural
stress is limited to 10 MN/m2 . Use E- 10 GN/m2 .

400 N 600 N

lm } lm \ 3 m

Figure P-690.

691. Determine the midspan deflection for the beam shown in

Fig. P-691. (Hint: Apply case 7 and integrate.)
Ans. Eld = (h*7 /48)(3L - 2a )
2 2 2

w N/m 1 [ w N/m
^a-+| \*-a -w
~
Lt

Ri
Figure P-691.
6-9 Deflections by the Method of Superposition 279

692. Determine the value of EI8 midway between the supports

for the beam in pig. P-692. (Hint: Combine case 11 and one-half of
case 8.)

1000 N/m 1000 N/m

2 m 2m 2m

*1
Figure P-692.

693.Determine the value of EI8 at the right end of the overhang-

ing beam in Fig. P-693. Arts. EI8 = 680 N m 3 downward ,

900 N
M = 600N-m
_JT
2m 3m 2m
zu
*i
Figure P-693.

694. The frame shown in Fig. P-694 is of constant cross section

and is perfectly restrained at its lower end. Compute the vertical
deflection caused by the couple M. Ans 8 =
^a (h -\- —\
EI 2

^i
c

Figures P-694 and P-695.

695. Solve Problem 694 if the couple is replaced by a vertical

downward load P.
Am 5= -
^ r(6 +
f)
280 6 Beam Deflections

696. In Fig. P-696, determine the value of P for which the

deflection under P be zero.
will Arts. P — 400 N

800 N/m
I
1.5 m 3 m

Figure P-696.

697. Two identical cantilever beams in contact at their ends

support a distributed load over one of them as shown in Fig. P-697.
Determine the restraining moment at each wall.
Ans. MA = - 3 wL 2 / 16; MB = - 5wL 2 /\6

Figure P-697.

698. The beam in Fig. P-698 is supported at the left end by a

spring which has a spring constant of 60 kN/m. For the beam, E= 10
X 10
9
N/m 2
and / = 60 X 10
6
mm4 . Compute the deflection of the
spring. Ans. 8 — 13.6 mm
4 m-

800 N/m

k= 60 kN/m

Figure P-698.

699. Two timber beams are mounted at right angles and in

contact with each other at their midpoints. The upper beam A is 50 mm
wide by 200 mm deep and simply supported on an 3-m span; the lower
beam B is 80 mm wide by 200 mm deep and simply supported on a 4-m
span. At their crossover point, they jointly support a load P= 10 kN.
Determine the maximum flexural stress in the assembly.
6-9 Deflections by the Method of Superposition 281

SUMMARY

Starting with the relation 1/p = M/EI developed in Art. 5-2, two
separate methods of determining slopes and deflections are discussed.
The first one, the double-integration method, is primarily mathematical.
Before the deflection at a particular point can be found, complete slope
and deflection equations must be determined; this is easily accom-
plished by using the concept of a general moment equation developed
on page 217. The constants of integration will become zero if the origin
of axes is selected at a position where the slope and deflection are
known to be zero, as at a perfectly restrained end or at the center of a
symmetrically loaded beam.
The area-moment method is generally more direct than the dou-
ble-integration method, especially when the deflection at a particular
position is desired. Depending as it does upon the geometry of the
elastic curve, the area-moment method emphasizes the physical signifi-
cance of the computations. The two basic theorems of this method,
developed in Art. 6-3, are summarized by

and

h/A - jj(**™)ba-Xb (6-5)

The any point is positive

tangential deviation at if that point lies
above the reference tangent from which the deviation is measured; the
converse is true for negative deviation. A positive value for the change
in slope means that the tangent at the right point is rotated in a
counterclockwise direction relative to the tangent at the left point.

The use of the area-moment theorems requires that the areaand

the moment bending-moment diagrams be readily calculated.
of area of
Therefore, beginning in Art. 6-4, and discussed more fully in succeed-
ing articles, there is developed a method of drawing moment diagrams
by parts (i.e., in terms of equivalent cantilever loadings) which is
equivalent to, and replaces, the conventional moment diagram.
Deflections in cantilever beams (Art. 6-5) are easily found from
the fact that the deflection at any point is equal to the deviation of that
point from a reference tangent drawn at the fixed end.
The deflections in simply supported beams are determined by
using a reference tangent to the elastic curve drawn at either reaction.
The complete procedure is outlined in the four steps on page 249. Since
midspan deflections are practically equivalent to maximum deflections,
an easy method of finding them is developed in Art. 6-7.
282 6 Beam Deflections

The conjugate-beam method (Art. 6-8) shows how the definitions

of shear and moment may be applied to a fictitious loading (in terms of
the M/ El diagram of the original loading) to determine the slope and
deflection at any point.
When beam loadings are combinations of the types listed in Table
6-2, deflections and slopes are obtained most easily by superposing the
results listed there. See Art. 6-9 for further details.
 Restrained
Beams

7-1 INTRODUCTION

Our study of simple stresses and torsion has shown that statically
indeterminate problems require relations between the elastic deforma-
tions in addition to the equations of static equilibrium. Similarly, for our
present study of indeterminate beams, additional relations must be
found from the geometry of the elastic curves of the beams. Such
relations are obtained from our study of the deflections in statically
determinate beams.
Three techniques are discussed: (1) double-integration; (2) the
method of superposition, which uses the general solutions in Tables 6-2
and 7-1; (3) the area-moment method, which deals directly with the
shape of the elastic curve. As we shall see, sometimes the deflection at a
particular position is required, sometimes a relation between the slopes
at two positions, sometimes a combination of these concepts.

7-2 REDUNDANT SUPPORTS IN PROPPED

AND RESTRAINED BEAMS

A cantilever beam is supported by two reactive elements, the shear V

and the moment M at the wall, as shown in Fig. 7- la. Since these

283
284 7 Restrained Beams

w N/m w N/m
M 4\ y*
v.*
(a) \V
I "(b)
y&
iv
Figure 7-1. Determinate and indeterminate beams.

Figure 7-2. Perfectly restrained beam.

values are readily computed from and 27 = 2M

= 0, the cantilever
beam is statically determinate. Propping up the beam at some other
position, as in Fig. 7- lb, introduces an additional reaction but does not
increase the equations of static equilibrium, so the beam has one
redundant support.
In other words, if any arbitrary value is assumed for R, values of
V and M may be computed that will satisfy the equations of static
equilibrium. Determination of the correct combination of R, V, and M
therefore requires a condition in addition to those found from static
equilibrium. Usually themost convenient condition is that the deflec-
tionunder R is either zero or some known value. Another condition
sometimes used, generally in the method of superposition, is that the
slope at the wall is zero.
A beam restrained at both ends, as in Fig. 7-2, has four reactive
elements. Because only two equations of static equilibrium are available
— namely, 27 = and SM = — the beam has two redundant sup-
ports. It is usually best to consider that the redundant supports are the
shear and moment at the same wall, i.e., VA and MA ; but sometimes, as
we shall see in Art. 7-5, the two end moments are taken as the
redundancies.

7-3 APPLICATION OF DOUBLE-INTEGRATION AND

SUPERPOSITION METHODS

The double-integration method may be applied to propped or restrained

beams in exactly the same manner as was described in Art. 6-2. In
applying it, we select the origin of the reference axes at a fixed end
where both the slope and deflection are zero, thereby causing the
constants of integration to be zero. However, the general moment
equation and its subsequent integrations will contain the unknown
values of shear and moment at the fixed end. These are easily evaluated
7-3 Application of Double-Integration and Superposition Methods 285

TABLE 7-1 Slope and Deflection at Free End

ML ML 2
EIO - — r^r
n+1
; Eld = -rr^
n+ 2
; M= moment at wall

LOADING EI9 EIS

ML ML 2
M 1
(•

ML ML'
2

u>N/m ML ML'
3

u;N/m
ML ML'
4

by substituting the boundary conditions at the other support into the

slope and the deflection equations. The details are explained in the
illustrative problems that follow.
The method of superposition may be applied by using the general
results listed in Table 6-2, but it is more convenient to use the summary
of slope and deflection at the free end of cantilevers under various
loadings that is given in Table 7-1. In this table the results are
expressed in terms of the moment M at the fixed end. Notice that the
coefficients are exactly the same as those previously developed for area
and for location of centroid in Table 6-1 (page 233). It is simple to
verify that the general results are

EI9 = ML and EIS = ML 1

n + 1 n + 2
where n is the degree of the moment curve.

ILLUSTRATIVE PROBLEMS

701. Solve for the reactive elements of the propped beam in Fig.
7-3a by two methods: (1) the method of superposition, considering RA
as the redundant support; (2) the double-integration method.
286 7 Restrained Beams

3 m

N
f M-3R
400 Ra
(b)

400 N

lm |
*2 .^
1X0 ' M=-800N.m
(0

Figure 7-3. Propped beam solved by superposition.

Solution:

Method of Superposition, The propped beam may be duplicated

by superposing the cantilever loadings shown in Fig. 7-3b and 7-3c so
that the resultant deflection at A is zero; that is, 6, = 5 2 + 8 3 Observe .

that 8 2 is the deflection under the 400-N load and that 8 3 is caused by
rotation of the 1.0-m unloaded segment through the slope angle 0. Using
Table 7-1, we combine the free end deflections of these loadings as
follows to give the zero deflection at A in the original beam:
2 :
(3RA )(3) (800)(2) (800)(2)
[1EI8 = 0] - 1 x =
from which we obtain
RA - 207 N
With RA known, a vertical summation of forces determines Vc :

[2 Y = 0] Vc + 207 - 400 = Vc = 193 N

From the definition of bending moment at C, which is equivalent to
2A/C = but simpler to apply, we obtain

[M = (2A/) L ] Mc = 3(207) - 2(400) = -179 N-m

An alternate method would be to let
of superposition c be the M
redundant support. Its value could then be determined by replacing the
given loading by a combination of cases 7 and 1 1 of Table 6-2 so
that the sum of the right end slopes is zero. Check Mc using this
procedure.
7-3 Application of Double-Integration and Superposition Methods 287

400 N

C 2 m Bf m A 1
^"

Figure 7-4. Solution by superposition.

Double-Integration Method. Returning to Fig. 7-3a, select the

origin of axes at the fixed end C. For convenience, we redraw the beam
as in Fig. 7-4 so that the fixed end is at the left. At this fixed end, the
slope and deflection are both zero, so the constants of integration C,
and C2 be zero since these constants are physically equivalent
will also
to slope and deflection at the origin. Setting up the differential equation
of the elastic curve in terms of the general moment equation, and
integrating twice, we obtain

Mc + vc x ~ 400 < JC ~ 2> (a)

dx 2
Vr x 2

»#-
dx ^c* + -\ 200 <* ~ 2>
~

+ C^~=
x
o
(*)

To
Ely -

evaluate
Y Mc and
+ ^_200 <x _ 2>3 + ^=0
Vc we, note that the boundary conditions at
to

A, where x = 3, are that the moment and the deflection are both zero.
Hence substituting x = 3 in Eqs. (a) and (c), we obtain

MC + 3VC -400(1) =

|(3f4(3f-f(.) -0
5

Solving these simultaneously, we have, as before

Vc = 193 N and Mc - - 179 N-m Arts.

One interesting aspect of this double-integration solution is that, if

the given beam had been perfectly restrained at both ends, the redun-
dants could be found by merely changing the boundary conditions at A
to zero slope and zero deflection. Then by using Eqs. (b) and (c), in

which to substitute x = 3, we would obtain Vc — 104 N and Mc =

-89 N-m.
702. A beam 4 m long and perfectly restrained at the ends carries
a uniformly distributed load over part of its length, as shown in Fig.
7-5. Use the double-integration method to compute the end shears and
288 7 Restrained Beams

Figure 7-5.

end moments, and then check these results using the method of super-
position.

Solution: The moment at A is shown acting in a positive or clockwise

direction, although we know from the downward concavity of the elastic
curve that this moment is actually negative or counterclockwise. By
making this deliberate error in the direction of the vector quantity MA,

our solution will determine not only the correct numerical value of MA
but also specify itscorrect negative sign.
The elastic curve in Fig. 7-5 shows that the slope and deflection at
A are both zero. Choosing the origin of axes at A will make the
constants of integration C x
and C2 both zero since they represent
respectively the slope and deflection at the origin. Setting up the
differential equation of the elastic curve in terms of the general moment
equation, and integrating twice, we obtain
900,
El MA + VA x 1>' {a)
dx
dy VA x-
EI-j- = M, x + - 150<x - 1>
3
+ €[ (b)
dx
MA x- VA x
Ely =
150
<*- 1>
4
+ c (c)

To evaluate MA and VA we, note that at the other restrained end

B where x = 4, the slope and deflection are also zero. Hence
y
substitut-
ing x — 4 in Eqs. (b) and (c), we obtain

AM* + (4)^ 150(3)

3
= (d)

(4)
2
^ + (4)
s
-£ ^(3)<-0 (e)

Solving these equations simultaneously yields

yA = 949 n and MA - -886 Nm Ans.

The negative sign for MA indicates that the direction of MA was

incorrectly assumed, as was stated earlier. The moment is actually
7-3 Application of Double-Integration and Superposition Methods 289

negative, and there should now be no confusion about sign in its

subsequent use.
M
Having determined VA and A we apply the conditions of static
,

equilibrium to the free-body diagram in Fig. 7-5 to compute the shear

and moment at B. A vertical summation of forces gives

[27 = 0] VB + 949 - 900(3) = VB = 1750 N Arts.

We can now determine MB from the condition SA/g = 0, but it is

simpler (and less confusing with respect to sign) to apply the definition
of bending moment:

[M = (2M) L ] MB = 4VA + MA - (900 X

3)(|)
= 4(949) - 886 - 4050
= -1140Nm Am.
As was shown in the preceding problem, this solution may also be
adapted to a beam propped at B by changing the
instead of fixed
boundary conditions at B to zero moment and zero deflection. Then by
substituting x = 4 in Eqs. (a) and (c), we could determine A and VA if M
this beam were propped at B.
To apply the method of superposition, we replace the given
loading repeated in Fig. 7-6a by the equivalent cantilever loadings in

Figure 7-6. Solution by superposition.

290 7 Restrained Beams

Fig. 7-6b, c, and d. We now apply the conditions that 0, + 2

- 93 =
and 6, + 82 - 83 - (1 x 3)
= 0.

Using Table 7-1, we then obtain

^
and

[1EIS = 0] + <1^_(4«3)!
_ lx (4050)0)
=Q
The solution of these equations yields, as before,

VA = 949 N and MA = -886 N-m Check

What would the answers be if the beam had been propped at A instead
of being fixed?

PROBLEMS

In solving the following problems, use superposition or double-

integration as directed by your instructor. Unless otherwise stated, the
supports are assumed to remain at the same level. Additional problems

may be selected from those in Art. 7-4.

703. For the propped beam shown in Fig. P-703, find R and
draw the shear and moment diagrams. Ans. R = (wb 3 /SL 3 )(4L — b)

w N/m -m
i

*
L

Figure P-703.

704. Compute the reaction R and sketch the shear and moment
diagrams for the propped beam shown in Fig. P-704.
Ans. R = wL/10

Figure P-704.
7-3 Application of Double-Integration and Superposition Methods 291

705. For the propped beam shown in Fig. P-705, determine the
reaction R and sketch the shear and moment diagrams.

Figure P-705.

706. A couple M is applied at the propped end of the beam

shown in Fig. P-706. Compute R at the propped end and also the wall
restraining moment. Arts. R = 3M/2L
M
&
Figure P-706.

707. Determine the reaction R and sketch the shear and moment
diagrams for the propped beam shown in Fig. P-707.

4kN/m
S TmT 3 m

R
Figure P-707.

708. Determine the reaction R for the propped beam shown in

Fig. P-708.

2kN/m

Figure P-708.
292 7 Restrained Beams

709. Determine the end moments for the restrained beam shown
in Fig. P-709. Ans. MA - - wL 2 /30; MB = - wL 2 /20

Figure P-709.

710. Solve for the end moments in the restrained beam loaded as
shown in Fig. P-710.
Ans. MA = -Pab 2 /L 2 \ Mc = -Pa 2b/L 2

y <
ZA t
Figure P-710.

711. There is a small initial clearance A between the left end of

the beam shown in Fig. P— 71 1 and the roller support. Determine the
reaction at the roller support after the uniformly distributed load is

applied.

w N/m
1

Figure P-711

712. Compute the end moments for the restrained beam shown
in Fig. P-712. Ans. MA = - 575 N-m; Mc = - 1525 N-m

W 1

A B m
w
///\Z
2m

Figure P-712.
3m C
w,
7-4 Application of Area-Moment Method 293

713. Determine the end moment and midspan value of EI8 for
the restrainedbeam shown in Fig. P-713. (Hint: Because of symmetry,
the end shears are equal and the slope is zero at midspan. Let the
redundant be the moment at midspan.)
Arts. M
= - (2/3)Pa; EI8 = (5/24)Pa 3
XP^
LA
Figure P-713.

714. For the restrained beam shown in Fig. P-714, compute the
end moment and maximum EI8. (Hint: Use equivalent cantilevers fixed
at midspan and extending toward either end.)
Ans. = - 2750 N-m M
lm 3kN/m lm
4 m
Figure P-714.

715. Determine the end moment and maximum EI8 for the
restrainedbeam shown in Fig. P-715. (Hint: Let the redundants be the
shear and moment at midspan. Also note that the midspan shear is zero.
Why?)

900 N/m 900 N/m

2 m 4 m 2 m
Figure P-715.

7-4 APPLICATION OF AREA-MOMENT METHOD

The method of superposition, described in the preceding article, requires

the use of tabulated values of slopes and deflections as listed in Table
6_2 or 7-1. Instead of relying on such tabulations, it is usually prefer-
able to use the area-moment theorems which develop the equations
needed to determine the redundants directly from the moment diagram.
For propped beams we use the condition that the deviation of the
support from the tangent drawn to the elastic curve at the wall is zero or
some known value. Specific details are explained in Illustrative Problem
294 7 Restrained Beams

Figure 7-7.

716, below. For beams perfectly restrained at both ends, as in Fig. 7-7,
the tangents to the elastic curve at the ends remain horizontal. There-
fore, since there is no change in slope between the ends, AB = 0. In
addition, if the ends A and B are at the same level, the deviation at B
from a tangent drawn at A is zero; i.e. tB/A = 0. Also, the deviation at A
from a tangent drawn at B is zero, or tA/B = 0. Applying the theorems
of the area-moment method, we may put these conditions in the form:

EI9AB - (area)^ = (a)

EItB/A = (area)^ xB = • (b)

EItA/B = (area),,* xA = •
(c)

These three equations are not independent; any two may be used
together with the equations of statics to determine the four reactive
elements. As a rule, it is best to use Eq. (a) and either Eq. (b) or (c),
depending on whether it is simpler to compute the moment of area of a
particular moment diagram about the right or the left end. Which to
choose will be apparent in the discussion of Illustrative Problem 717,
below.

ILLUSTRATIVE PROBLEMS

716. Figure 7-8 shows a cantilever beam carrying a uniformly

distributed load of w N/m. The beam is propped up at the free end by
an unyielding support. Compute the reactions, and sketch the shear and
moment diagrams.

Solution: The wall is assumed to provide absolute resistance to rota-

tion of the beam, so that at B (Fig. 7-8a) a tangent to the elastic curve
is horizontal and passes through A Therefore, the deviation of A from
.

the tangent at B is zero. Expressing this in terms of the moment diagram

by parts shown in Fig. 7-8b, we obtain

[
EItA/B " (&™&) AB -xA = 0]

(L)(RA L)(2 (L)/wL 2 \/3

(H
r \

K^H
7-4 Application of Area-Moment Method 295

w N/m X ._
B
(a) Load diagram and
• deflection curve

I Elastic curve

RA L

(b) M diagram by parts

(c) Shear diagram

(d) Moment diagram

vulf
l

B- 8

Figure 7-8. Propped beam solved by area-moment method.

or

A vertical summation of forces determines V at the wall:

[27 = 0] RA + V - wL =
whence, substituting the value of RA we
, obtain

F= wL -\wL =fwL
From the definition of bending moment at /?, which is equivalent
to 2Af5 = but simpler to apply, we obtain

wLl
uB -
M r
The shear diagram being drawn as in Fig. 7-8c, the location of
zero shear is determined from

[K-(2 7) t = 0] fwL-wjt = x=|L

296 7 Restrained Beams

whence themaximum positive moment is

[AA/ = (area),] M = j(|wL)(|L) = ^wL 2

The moment diagram in Fig. 7-8d shows that the maximum

moment is at the wall.

717. A beam 4 m long and perfectly restrained at the ends carries

a uniformly distributed load over part of its length as shown in Fig.
7-9a. Compute the end shears and end moments.

Solution: This problem was solved on page 288 by double integration.

Its by the area-moment method shows the close similarity
solution here
of the two methods and demonstrates the direct use of the moment
diagram. We begin by drawing the moment diagram by parts from left
to right, as in Fig. 7-9b. Although we know from the downward
concavity of the elastic curve that the moment at A should be negative,
we nevertheless show it as positive, that is, clockwise. Because of this
deliberate error in the direction of the vector quantity MA our solution
,

will determine not only the correct numerical value of MA but also its

correct negative sign.

The elastic curve in Fig. 7-9a shows that the change in slope
between tangents drawn at A and B is zero. Applying the first theorem
of the area-moment method, we obtain
4(4KJ 3(4050)
[EI$AB - (area)^ - 0] + 4M,
"'* = (a)
2 '
3
The deviation of B from a tangent drawn at A being zero, we
obtain from the second theorem of area-moment:

[EItB/A - (area)*,, xB = 0]

(*)

-*1 mi -3 m

MAVc m 900 N/m

yj M
A
(a) Load diagram and
/,

^ elastic curve

MA (b) M diagram by parts

(900 X 3)(3/2) = -4050N-m

Figure 7-9. Restrained beam.
7-4 Application of Area-Moment Method 297

Solving Eqs. (a) and (b) simultaneously yields

VA = 949 N M = -886 Nm Ans.

and A

The negative sign for MA confirms that the direction of MA was

incorrectly assumed. It actually is a negative moment, and there should
now be no confusion regarding sign in its subsequent use.
The deviation A from a tangent drawn at B is also zero, so that
at
we could have used EItA/B — (area)^ xA = 0. A glance at the moment
•

diagram (Fig. 7-9b) shows that it is simpler in this problem to take

moments of area about B rather than about A. For this reason, we used
EItB/A = rather than EItA/B = 0.

Comparison with the solution by double integration on page 288

discloses that Eqs. (d)and (e) used there are identical with Eqs. (a) and
(b) here. On the other hand, comparison with the superposition solution
on page 290 discloses that its second equation is equivalent to EItA / B =
0, which we could have used here in place of EItB/A = 0.
Having determined VA and M
A we compute the shear and mo-
,

ment at B in the identical manner as was used on page 289. It is

needless to repeat it here.

718. Compute the end moment and maximum deflection for the
symmetrically loaded and perfectly restrained beam shown in Fig.
7- 10a.

VA =1800N VC =1800N

lb)
M,
%
VB -0
m
900 N/m
2 m

(0 - 1800 X 1 >

-1800N.m
Figure 7-10. Symmetrical loading.
298 7 Restrained Beams

Solution: Because of symmetry, the end shears are equal to each other,
and each equals one-half the applied loads. The end moments also equal
each other but are unknown. The simplest condition that determines
the unknown end moment is that in a symmetrically loaded beam the
tangent to the elastic curve at midspan is horizontal, and therefore the
change in slope is zero between this tangent and a tangent drawn at
either end.
In this problem, instead of solving directly for the end moment, it
mpler to solve first for the moment M
B at midspan and then apply
the definition of bending moment to determine the end moment. We
begin by drawing the free-body diagram of the segment BC as in Fig.
7- 10b. Since the end shears each equal 1800 N, at midspan the shear
VB 0; consequently the M diagram by parts is drawn as in Fig.
7- Because there is no change in slope between the tangents at the
10c.
midpoint B and the end C, we obtain

EI9BC = (area), c = 0] 3MB -

im x 2 =
[
^
MB - 400 N-m
Applying the definition of bending moment, we find the end
moment to be

[Mc - (2M) L ] Mc = MB - 1800 = 400 - 1800

= -1400 N-m Ans.
The maximum deflection occurs at midspan and is numerically
equal to the deviation of B from the horizontal tangent at C. Using the
now known value of MB = 400 N-m, we obtain

[
EIt B/c " (*™a) BC 'XB ]

£„. (w *3,(§)-(i^)(, + 2 x2
)

= -1200 N-m 3
Ans.

The minus sign indicates that the deflection y is directed downward as

shown.

PROBLEMS

Unless otherwise stated, assume unyielding supports. Additional

problems may be selected from those in Art. 7-3.

719. For the propped beam shown in Fig. P-719, determine the
reaction R and the midspan value of EI8.
Ans. R = (5/16)/>; EI8 = (7/76S)PL 3
7-4 Application of Area-Moment Method 299

i
4 . . 4
f
Figure P-719.

720. Compute the reaction R and sketch shear and moment

diagrams for the propped beam shown in Fig. P-720.

Figure P-720.

721. For the propped beam shown in Fig. P-721, determine the
reaction R and sketch the shear and moment diagrams.
Arts. * = (ll/40)wL

Figure P-721.

722. For the beam shown in Fig. P-722, compute the reaction R
at the propped end and the moment at the wall. Check your results by
letting b — L and comparing with the results in Problem 706.

/T7t77
^tj\

Figure P-722.
300 7 Restrained Beams

723.Find the reaction R and the moment at the wall for the
propped beam shown in Fig. P-723.

w N/m
L
2
I. L
2
& t2Z

Figure P-723.

724. The beam shown in Fig. P-724 is only partially restrained

at the wall so that, after the uniformly distributed load is applied, the
slope at the wall is wL 3 /4%EI upward to the right. If the supports
remain at the same level, determine R. Ans. R = (7/16)wL

N/m

Figures P-724 and P-725.

725. If the support under the propped beam in Problem 724

settles an amount 6, show that the prop reaction decreases by 3EI8/L 3 .

726. A beam L meters long, perfectly restrained at both ends,

supports a concentrated load P at midspan. Determine the end moment
and maximum deflection.

727. Repeat Problem 726, assuming that the concentrated load is

replaced by a uniformly distributed load of w N/m over the entire

Ans. M = - wL / 12; EI8 = wL /384
2 4
length.

728. Determine the end moment and maximum deflection for a

perfectly restrained beam loaded as shown in Fig. P-728.

Figure P-728.
7-4 Application of Area-Moment Method 301

729. For the restrained beam shown in Fig. P-729, compute the
end moment and maximum EI8.
Ans. M= - 1830 N-m; EI8 = 1080 N-m 3

lm 2kN/m lm

4 m

Figure P-729.

730. Determine the end moment and maximum deflection for

beam loaded as shown in Fig. P-730.
the perfectly restrained
Ans. M = - (wa 2 /6L)(3L - 2a); Ely = - (wa 3 /24)(L - a)

w N/m| w N/m
<M
Figure P-730.

731. The beam shown in Fig. P-731 is connected to a vertical

rod. If the beam is horizontal at a certain temperature, determine the
increase in stress in the rod if the temperature of the rod drops 50 °C.
Both the beam and the rod are made of steel with E= 200 X 10
9

N/m2 For the beam, use / = 60 X 106

.
4
mm .

Ans. o = 95.8 MN/m 2

/////

L = 10 m
-^ A = 50 mm 2
a = 11.7 M m/(m.°C)

2m
Figure P-731.
302 7 Restrained Beams

732. The midpoint of the steel beam in Fig. P-732 is connected

to the vertical aluminum rod. Determine the maximum value of P if the
stress in the rod is not to exceed 150 MN/m 2
.

Aluminum
/////
L = 5 m
~L A- 40 mm 2
~T E = 70 X 10 9 N/m 2

2m 2m

Steel

/ = 50 X 10 6 mm 4
£=200 X 10 9 N/m 2

Figures P-732 and P-733.

733. The load P in Problem 732 is replaced by a counterclock-

wise couple M. Determine the maximum value of M if the stress in the
vertical rod is not to exceed 100 MN/m 2
. Arts. M — 41.1 kN-m
734. Determine the end moments for the restrained beam shown
in Fig. P-734.
Arts. MA = - (5/192)wL 2 ; MB = - (11/192)>vL
2

wN/m

2 '
2

Figure P-734.

735. The beam shown in Fig. P-735 is perfectly restrained at A

but only partially restrained at B, where the slope is wL 3 /4SEI directed
up to the right. Solve for the end moments.

w N/m

Figure P-735.

736. For the restrained beam shown in Fig. P-736, compute the
end shears and end moments and sketch the shear and moment dia-
grams.
7-4 Application of Area-Moment Method

737. In the perfectly restrained beam shown in Fig. P-737,

support B has settled a distance M
A below support A. Show that B =
- MA - 6EIA/L 2 .

Figure P-737.

738. A perfectly restrained beam is loaded by a couple M ap-

plied where shown in Fig. P-738. Determine the end moments.

Ans. MA =—
Mb
T (
A
-l); *—¥(*-)
I 3a

^r
B m

Figure P-738.

7-5 RESTRAINED BEAM EQUIVALENT TO

SIMPLE BEAM WITH END MOMENTS

Usually the redundant elements in a restrained beam are most easily

determined by applying the method which considers the redundant
supports to be the shear and moment at one wall. However, it is
sometimes desirable to treat the end moments as the redundant sup-
ports. To do this, the restrained beam is considered equivalent to a
304 7 Restrained Beams

simple beam acted on not only by the given loading but also by end
moments sufficient to rotate the ends of the beam until the slopes at the
ends correspond to the slopes at the ends of the restrained beam. Thus
the restrained beam in Fig. 7-1 la may be considered equivalent to
superposing the loadings in Figs. 7-1 lb and 7-1 lc.
It is evident that the unsymmetrical loading in Fig. 7-1 lb causes a
greater slope 9 2 at the right end than the slope 0, at the left end. To
cancel these end slopes by adding the loading in Fig. 7-1 lc, we must
have 0, = 6[ and 9 2 = 9 2 which requires that
, MB be greater than A in M
order for 9 2 to be greater than 9[. In other words, the larger end moment
acts at the wall that is closer to the resultant of any single load.
The end moments
difference between the B and A is balanced M M
by the couple R'L, consisting of the forces R' applied at the ends of the
beam in Fig. 7-1 lc. By superposing the reactions in Figs. 7-1 lb and
7- lie, we obtain VA = R - R' and VB = R 2 + R'. If the loading
x

were symmetrical, the end slopes 0, and 2 would be equal, which would
require equal end moments A and M M
B In this case, there would be no
.

couple reaction R ', so the end shears would equal the end reactions of a
similarly loaded simple beam. This conclusion agrees with the observa-
tion in Illustrative Problem 718 (page 297) on symmetrical loading.
To consider the end moments as the redundant supports, there-
fore, the moment diagram by parts for the beam in Fig. 7- 11a will be
drawn to correspond to the loadings in Figs. 7-1 lb and 7- lie and will
appear as in Figs. 7- 12a and 7- 12b. Applying EI9AB = and EItB/A =
0, we can solve directly for A and B as M
the redundant M
supports,
whence VA and VB are obtained by applying the equations of static
equilibrium. This procedure is not generally as simple as treating the
shear and moment at one end as the redundants.

-*-a^>~-< b *-

L w N/m

r>—jvi
o if
wN/m
,„- B
\}mb =
(b) Load simply supported

M •„
A

1 'a Vb
(a) Restrained beam

(c) End couple loading

Figure 7-11. Restrained beam resolved into simple beam loadings.

7-5 Restrained Beam Equivalent to Simple Beam with End Moments 305

(a)
^^^ wb _

L
(b) MA *=^~~
M f

Figure 7-12. Moment diagrams for beams in Fig. 7-1 1b and 7-11c.

PROBLEMS

Use the end moments as the redundants in solving these problems.

739. Determine the wall moment in the propped beam described

in Problem 705 (page 291).
740. Solve for the wall moment in the propped beam shown in
Fig. P-740. Arts. M
= - 3200 N-m

900 N/m

2m 4 m

Figure P-740.

741. Compute the moment at the restrained end of the propped

beam shown in Fig. P-741.

m
3m lm
m
Figure P-741.
306 7 Restrained Beams

742. Determine the end moments in the restrained beam de-

scribed in Problem 710 (page 292).
743. Compute the end moments in the restrained beam described
in Problem 712 (page 292).
744. Determine the end moments in the perfectly restrained
beam shown in Fig. P-744.

Arts. MA = - 1380 N-m; MR = - 1820 N-m

800 N/m
2m 3 m lm

Figure P-744.

745. The restrained beam in Fig.P-745 carries a uniformly

distributed load over part of the span and a couple. Compute the end
moments. Arts. MA = - 2700 N-m; MB = - 2190 N-m

Figure P-745.

7-6 DESIGN OF RESTRAINED BEAMS

By applying the methods described in the preceding articles, we can

determine general values for end moments and deflections in perfectly
restrained beams that carry various general loadings. These values are
summarized in Table 7-2. In the following problems we show how these
general values are superposed to solve design problems involving vari-
ous combinations of loads.

ILLUSTRATIVE PROBLEMS

746. Determine the section modulus required for a beam to

support the loads shown in Fig. 7-13 without exceeding a flexure stress
of 80MN/m 2
.
 7-6 Design of Restrained Beams 307

TABLE 7-2. Restrained Beam Loadings

CASE VALUES OF Ely

NO. TYPE OF LOAD END MOMENTS (y is positive downward)

Pab 2
2
Pb
L2 Midspan Ely - ^-
IT(3L - 4b)
Pa 2b
Note: only for a >b
L2

PL PI?
MA =MR = Max. EIy =
192
~m
WL wL 4
u/N/m MA = M t
wL
12
2

12
Max. Ely -^-^ WL?

N/m
si Midspan £/y=
w w
wL 4
=
WL 3

loN/rn "X
wL 2
A/,
30 15 WL 3
Midspan £/y
wL 2
WL 768 384
M K
20 10

u;N/m 5wL 2 5WL 7h>L 4

= 7IFL
3

JL_
M A = M,
96 48
Max. £7y =
3840 1920

M "<-^(?-)
W -^^-l)
lU
6£/A
L2
6£7A
MB =
308 7 Restrained Beams

1=4000 Nl |P2 =6000N

2 m |lm y lm \yy^
u>=2000N/m
L=4 m
Figure 7-13.

Solution: The end moments due to the distributed load and the central
load P ]
are equal because of symmetry. As discussed in Art. 7-5, the
eccentric load P2 causes a larger end moment at the nearer wall — the
right wall, in this case. From Table 7-2, the maximum bending moment
occurring at the right end is given by

wL' P L P2 a 2b
Max. M — —
~I2
X

2
2000(4)' 4000(4) 6000(3) (1)
12 8 2
(4)

Max. A/= -8040 N-m

The negative sign of moment indicates a tensile stress at the top
fibers. In an unsymmetrical section like a T beam, this would be
important; but in a symmetrical section, only the numerical value of
moment need be used. From the flexure formula, the section modulus
required is therefore

5 = ^o S =
8040
6
= 100.5 x 10" 6 m 3

80 x 10

= 100.5 x 10
3
mm 3
Arts.

747. Select a suitable wide flange beam to support the loads

shown in Fig. 7-14 without exceeding a flexural stress of 120 MPa.
Compute the midspan deflection of this beam. Neglect the mass of the
beam. Use E - 200 GPa.

P x
= 25 kN P2 = 40 kN

1 m | 1 .8 m

L = 4m
Figure 7-14.
7-6 Design of Restrained Beams 309

Solution: Because of the unsymmetrical loading, the end at which the

maximum moment occurs is not evident and computations for moments
at each end must therefore be made. At the left end, we obtain
2
P ab P2 ab 2
M r
- x

L2 L2
2 2
(25)(1)(3)
2
(40)(2.8)(1.2)
2
_ ^„m m
(4) (4)

At the right end, the moment is

P,a 2b P,a 2b
M D = -

= _
L2
(25)(ir(3)

(4)
k

2
2/o\
_
/Af\\/~> o\2
(40)(2.8r(1.2)

(4)
2
= ^^
Substituting the larger numerical value of end moment into the flexure
formula, we find that the required section modulus is

28 21 X
= 235 x lO" 6 m3 = 235 X 103 mm3
'

S - *f6
a 120 x 10 -

A suitable beam is a W200 X 27 with S = 249 X 103 mm 3 and

/ - 25.8 X 10
6
mm4 = 25.8 X 10" 6 m4 Table 7-2 gives the midspan
.

deflection* as
Ph 2

Efy

£/v
- !®p.
= 12.81 kN-m3
m - 4(1)] +&>££ [3(4) -4(1.2)]
Substituting the numerical values for E and /, we obtain

(200 X 9
10 )(25.8 x 10" 6 )v = (12.81 x 10
3
)

from which

y - 2.48 X 10" 3 m= 2.48 mm Arts.

PROBLEMS

748. A restrained beam 6 m long supports a concentrated load of

30 kN at 2 m from the left end and another concentrated load of 50 kN
at 1.5 m from the right end. Select the lightest wide flange beam that
will support these loads without exceeding a flexural stress of 120 MPa.

•In computing midspan deflection for a concentrated load, the term b is the
smaller of the two segments into which the load divides the length of the beam.
310 7 Restrained Beams

Neglect the mass of the beam. Compute the midspan deflection of this

beam ifE = 200 GN/m 2 . Ans. W360 X 33; S = 3.21 mm

749. A timber beam 150 mm wide by 300 mm deep and 6 m long
is perfectly restrained at both ends. It supports a uniformly distributed
load of 4 kN/m over its entire length and a concentrated load P at 2.5
m from the left end. Determine P so as not to exceed a flexural stress of
10 MN/m or a 2
midspan deflection of 1/360 of the span. Assume that
E - 10 GN/m 2
.

750. A W200 x 36 steel beam 5 m long is perfectly restrained at

both ends. It carries a concentrated load of 20 kN at 1 from the left m
end and another concentrated load of 30 kN at 2 from the right end. m
Compute the maximum flexural stress and the midspan deflection.
Neglect the mass of the beam. Use E = 200 GPa.
751. A timber beam with a square cross section supports the
loads shown in Fig. P-751. Determine the cross-sectional dimensions if

the allowable flexural stress is 10 MN/m 2

. What is the maximum
shearing stress developed in the beam?

2 kN 4 kN

lm |lm{ 2m
3 kN/m

4 m

Figure P-751.

752. Using Table 7-2, check the values of end moment and
midspan deflection for the restrained beam in Problem 713 (page 293).
753. A timber beam 100 mm
wide by 150 deep supports the mm
loads shown in Fig. P-753. Determine the maximum shearing stress
developed.

4kN

1 m Im 4 kN/m
4 m

Figure* P-753 and P-754.

754. In Problem 753, compute the maximum flexural stress if the

right end settles 20 mm relative to the left end. Assume that the beam is
 1

7-6 Design of Restrained Beams 31

perfectly restrained against rotation at its ends and that E= 10 X 10

9

N/m 2 . Arts, a = 16.1 MN/m2

755. An SI 30 X 22 steel beam 4 m long carries a load varying
uniformly from zero at the left end to 15 kN/m at the right end. The
beam is perfectly restrained against rotation at its ends, but the right

end settles 10 mm relative to the left end. Determine the ratio of the
maximum flexural stress to the flexural stress if no settlement had
occurred. Use E= 200 GPa. Ans. 1.06

SUMMARY
The principles of beam deflections studied in Chapter 6 are
applied here to obtain additional equations which can be combined with
the equations of static equilibrium to solve problems involving statically
indeterminate beams.
In propped beams, we generally use the fact that the deflection
under the redundant support is zero (if the support does not settle) or
some known value (if For beams perfectly
the support does settle).
restrained at the ends, the elastic curve is no change in
is such that there
slope between the ends and the deflection of one end relative to the
other end is zero.
The method of superposition is usually the easiest way to de-
termine the redundant support in propped beams. It is also the best
method of determining the end moments in restrained beams subjected
to loadings of the types listed in Table 7-2.
With the double-integration method or the area-moment method,
we generally take the shear and moment at one end as the redundant
supports. Either method is as simple to apply as the other. The double-
integration method is essentially mathematical and automatically de-
termines deflections as well as redundant supports. The area-moment
method, by emphasizing the geometric relations between the elastic
curve and the moment diagram, is perhaps more direct in obtaining the
equations that determine the redundant supports, but additional work is
needed if deflections are required.
In Art. 7-5, the reduction of a restrained beam to the combination
of a simply supported beam carrying the given loading and another
simply supported beam subjected to end couples is valuable in visualiz-
ing which restrained end carries the larger moment. This concept is
particularly advantageous in applying the area-moment method to con-
tinuous beams, as will be explained in the next chapter. In addition, it

provides the basis for a rapid method of drawing the shear diagram,
also explained in the next chapter.
 Continuous
Beams

8-1 INTRODUCTION

In this chapter we consider beams that are continuous over two or more
spans, thereby having one or more redundant supports. It is possible to
determine these redundancies by applying the deflection relations devel-
oped in Chapter 6, but a more convenient method is to consider the
unknown bending moments at the supports of the beam as the re-
dundancies. After these bending moments are found, it is comparatively
simple to determine the reactions, as we shall show in Art. 8-5.
We present two methods of solving for the moments at the
supports, either or both of which may be studied, depending on the
available time. In the first method it is necessary to find a general
relation between the bending moments at any three sections in a beam.
This relation is known as the three-moment equation and is easily derived
by applying the area-moment theorems. We shall show how this equa-
tion is used to determine deflections as well as redundancies in any type

of beam. Actually, the three-moment equation can be used to solve all

the problems in Chapters 6 and 7; however, in some instances, it is best
used in combination with the area-moment or the double-integration
method. Such combinations of techniques will also be discussed.

312
8-2 Generalized Form of the Three-Moment Equation 313

An alternate method of solving continuous beams is moment

distribution, which isdescribed in Art. 8-8. This method is completely
independent of the three-moment equation, but Art. 8-5 for determin-
ing shear diagrams is common to both. Before the moment-distribution
method may be applied, however, each span of the continuous beam
must be assumed to be perfectly restrained at the supports and the fixed
end moments must be computed. Usually the span loadings are such
that the fixed end moments are readily obtained by superposition of the
general results tabulated in Table 7-2 on page 307. However, this table
does not list results for distributed loads over part of a span; for these
loadings the three-moment equation is preferable if a more general list is
not available.

8-2 GENERALIZED FORM OF THE

THREE-MOMENT EQUATION

A portion of a beam that is loaded and supported in any manner is

shown in Fig. 8- la. At any three points 1, 2, and 3, pass cutting sections
and replace the effects of the loads to^ the left or right of these sections
by the proper values of vertical shear and bending moment. Thus the
beam segments between points 1 and 2 and between points 2 and 3
(hereafter referred to as spans 1 and 2, respectively) may be isolated by
means of the free-body diagrams in Fig. 8- lb. The lengths of the spans
(or segments) are L, and L 2 and the bending moments at points 1, 2,
,

and 3 are M v Af2 3 , M

the vertical shears at these points are V v V_ 2
;

Any loading j

1 2 3

-« Lx A* L2 *
(a) Loading diagram

Load on Load on
span 1 span 2
MX t
)m 2 Af2
Q ;>
v. v_2

(b) Free-body diagrams of beam segments

Figure 8-1. General loading on any beam.
 314 8 Continuous Beams

(just to the left of point 2), V2 (just to the right of point 2), and V_ 3
just to the left of point 3.
The technique discussed in Art. 7-5 enables us to resolve the
free-body diagrams of the beam segments into simply supported spans
that carry the actual beam loading, and spans loaded only by the
bending moments and held in equilibrium by the couple reactions R[ on
span and by R 2 on span 2. This equivalent loading is shown in Figs.
I
'

8-2a and 8-2b, respectively. When these loadings are superposed, they
produce the free-body diagrams in Fig. 8- lb. Hence the vertical shears
at points 1, 2, and 3 are equal to the algebraic sum of the simple beam

reaction and couple reaction at these respective points.

In this manner, the moment diagram of each original beam
segment is resolved into the moment diagram of the loads assumed to
be carried on a simply supported span and the trapezoidal moment
diagram caused by the bending moments in the original beam at the
selected points 1, 2, and 3. These diagrams are shown in Figs. 8-2c and
8-2d, respectively.
For clarity, the elastic curve of the beam has been drawn sep-
arately in Fig. 8-3. The deflection of the curve is greatly exaggerated in
order to show the geometric relations. Note that points 1, 2, and 3 lie on
it.

A tangent drawn to the elastic curve at point 2 determines the

tangential deviations r
1/2 at point 1 and f
3/2 at point 3. Another line

Load (a) Loads on simply

supported spans
Lx

(assume M >M x 2)
(assume M >M 3 2)
/ i i \ (b) Loading by end / ,X
L \ 'M
'M2 moments and j>\ L2 \\
7
III | i
balancing couple
j I
y
reactions
m

(c) Moment diagrams

of loads on simply
supported spans

(d) Moment diagrams

M 2
of loading by end jg t

L,
moments L, 2
l

Figure 8-2. Analysis of original loading.

8-2 Generalized Form of the Three-Moment Equation 315

'3/2

=-3/2 "3

\t v

Figure 8-3. Elastic curve of any beam.

drawn through point 2 parallel to the initial position of the unloaded

beam (which has been assumed horizontal for convenience) determines
the heights of points 1 and
above point 2 to be h and h 3 There are
3 x
.

formed the shaded similar triangles having the bases L and L2 and the x

altitudes (h - t x/2 ) anc* (ty 2 ~ ^)-

x

From the proportions between similar triangles, it is evident that

h x
-t 1/2 !

3/2
- h.

which reduces to

1/2 3/2
(«)

The values of the tangential deviations are found from

'*/ 2
= area )i-2**i
£y(
and

3/2
= -£j(are 3-2 • X-

where (area)^ x is the moment of area about point 1 of the moment

•
x

diagram between points 1 and 2. As was said previously, this moment

diagram has been resolved into the area A (see Fig. 8-2c) and the two x

triangular areas into which the trapezoidal diagram for the end mo-
ments is divided (see Fig. 8-2d). Likewise, (area) 3 _ 2 x 3 is the moment
about point 3 of the area of the moment diagram between points 2 and
3, as represented by area A 2 and the trapezoidal diagram for the end

moments which has been subdivided into two triangles.

316 8 Continuous Beams

We can therefore express the tangential deviation /

I/2 at 1 from a
tangent to the elastic curve drawn at 2 as

'" 2 " EI
1

/*,*, +-M L X X
X-L, + - A/2 L, X -L, w
and the tangential deviation /
3/2 at 3 from the same tangent drawn at 2
as

'
3 /2
- A 2 b 2 + - M2 L 2 X - L 2 + - A/3 L 2 x j L i (c)
El
Substituting these values of /
1/2 and t
3/2 in Eq. (a) gives

6A b
ML + 2A/2 (L, + L 2 ) + ML +
6 A.
— p- a.
1 + — 2 2
-*-*

M
X X 3 2

= 6EI (8-1)

This equation expresses a general relation among moments at any three

points in a beam, and hence is known as the three-moment equation.
When 1, 2, and 3 are on the same level in the deflected
points
beam, the heights h and h 3 in Fig. 8-3 become zero and so does the
x

right-hand term in Eq. (8-1). This is the usual condition in which the
three-moment equation is applied. The three points selected in applying
the equation to continuous beams are the points at the supports (usually
assumed as rigid or else as settling the same amount); the equation is
used to determine the bending moments in the beam over the supports.
If the three-moment equation is used for deflections, two of the

points are selected over supports and the third is chosen at the point
whose deflection is desired. Evidently the moments at the three points
must first be known in order to compute deflections. We shall expand
this application of the three-moment equation in Art. 8-7.

Rules of sign

Equation (8-1) was derived under the assumption that the bend-
ing moments
at the selected points were positive and that points 1 and 3
were above point 2. Hence, heights h and h 3 must be considered x

positive when measured upward from point 2. If the moment at any

point is actually negative, the negative sign must be used when substitut-
ing its value in Eq. (8-1). Conversely, if an unknown moment is actually
negative at any point, Eq. (8-1) will give a negative value for that
moment; in other words, the sign of the moment at that point is

automatically opposite to the positive value assumed in the derivation of

the three-moment equation.
8-3 Factors for the Three-Moment Equation 317

8-3 FACTORS FOR THE THREE-MOMENT EQUATION

The usefulness of the three-moment equation depends on the ease with

which the expressions 6Aa/ L and 6Ab/L in it can be found. As was
said earlier, these expressions refer to the moment of area of the
moment diagram resulting from carrying the applied loads on a simple
span of the same length as the equivalent beam segment. The general
expressions in Table 8-1 were obtained by the following procedure.

Case 3: Uniformly varying load

The loading over a span L in a continuous beam varies uniformly

over the span. If this loading is assumed to be supported on a simple
span, the moment diagram is drawn by parts from left to right, as in
Fig. 8-4. The moment of area of this moment diagram about the right
end is given by

Ab

_ wL
wL*(l 4
1 \ 7
W 4
•L
m
6 \ 6 20/ 360 <

Multiplying this by 6/L, we obtain the following general value for this
type of loading:

6Ab = 7 ,,
L W WL Arts.

wV
wN/m ^^ 6

4^^ L
t

' 3rd degree—^^V.

|1 curve ^V
wV
Moment diagram 6
(a) Load on simple span (b)

Figure 8-4. Uniformly varying load.

Special loadings

For cases not listed in Table 8-1, or if the table is not available,
the following example may be helpful.
Assume a continuous beam loaded as in Fig. 8-5; we wish to
evaluate6A 2 b2 /L2 for span 2. Take the loading on span 2 as if it were
simply supported on a 4-m span, and draw the moment diagram by
 318 8 Continuous Beams

TABLE 8-1 . Values of SAa L and SAb L

CASE TYPE OF LOADING 6Ad 6Ab

NO ON SPAN L L

?C*-* ™(L -b 2 2
)

r^=\
N/m
wL 3 = WL^ wL 3 _ WL 2
4 4 4 4

w N/m

_8_
60
h,L
3
= 4
30
**X 2
60 ^ 30
WL

N/m

60 ^ 30
WL —
60 ^
w/ 3
= —
30
WL
™L 4

*«-a-*

wN/m w
[6
2
(2L 2 - 6 2 ) - a 2 (2L
2
- a 2 )] \d\2L 2 - d 2 ) - c\2L 2 - c
4L 4L
* o *
i4

N/m

^wL
32
3
= ^WL
16 32 ^ 16
WL

prq ^(3a 2 - L*) + ^(3* 2 -Z. 2 )

8-3 Factors for the Three-Moment Equation 319

(a) Continuous beam

900 N-m
600 N/m

3rd degree curve

-900 N-m
(b) Load on span 2 simply supported (c) Moment diagram by parts

Figure 8-5. Evaluation of $A 2 b 2 /L 2 for special loading.

partsfrom right to left (this is more convenient here). Since 6Ab/L

means multiplying 6/L by the moment of area of the moment diagram,
moments being taken about the right end, we have

6A,b
2"2 /900X4W2 A /900x3\/ 4 A
x4 t
+ x3
(^)(3 H^r-)( l
5 )

= ^(4800 2295) = 3758 N-m2 Arts.

If it had been necessary to evaluate 6A 2 a 2 /L2 for span 2, the

symbol a 2 would have told us to take the moment of area about the left
end, since, as Fig. 8 -2c shows, the symbols a and b refer to moment
arms measured respectively from the left and right ends of a span.

PROBLEMS

If the span loadings on a continuous beam reduce to the simply

supported loads shown jn each of the following problems, evaluate the

factors 6Aa/L and 6Ab/L.
801. See Fig. P-801. Check your result by letting a = L/2 and
comparing with case 2 of Table 8-1. Ans. 2
wa (3L — 2a)
320 8 Continuous Beams

wN/m | 1 wN/m
^.-| \—a -w
Li ~

Ri
Figure P-801

802. See Fig. P-802. When b = L/2, how does your result
compare with case 2 of Table 8-1?

•*- a -*\« — 26 —«4»- a

1
wN/m 1

Ri X.
Figure P-802.

803. See Fig. P-803. Ans. ^vvL 3

Figure P-803.

804. See Fig. P-804. Check your result by subtracting the answer
for Problem 803 from case 2 of Table 8-1.

wN/m ^\^^ ^s^ w N/m

c
L >r L
2 2 *

Figure P-804.
8-3 Factors for the Three-Moment Equation 321

805. See Fig. P-805. The roller support may resist upward or
downward reaction.
Ans. 6Aa/L = - (M/L)(3a 2 - L 2); 6Ab/L = (M/L)(3b 2 - L 2)

M
Id
^r^f BOB
777777"

Figure P-805.

806. See Fig. P-806.

Ans. 6AZ/L = 5436 N m2 6Ab/L =
•
; 4014 N m •
2

600 N/m

Figure P-806.

807. See Fig. P-807. Solve by combining the results for Problems
805 and 806.

600 N/m
M = 500N-m

o 3m
Figures P-807 and P-808
////)/

808. Solve Problem 807 if the couple is applied in a counter-

clockwise sense.

809. See Fig. P-809. Ans. 6Aa/L = 6648 N m2

•
322 8 Continuous Beams

800 N/m

Figures P-809 and P-810.

810. Solve Problem 809 if the couple is applied in a counter-

clockwise sense.

8-4 APPLICATION OF THE THREE-MOMENT EQUATION

We now see how the three-moment equation may be applied to de-

termine the moments over the supports in various types of continuous
beams. Later articles will show how these moments are used to de-
termine the reactions of continuous beams and will describe a speedy
method of drawing shear and moment diagrams.

ILLUSTRATIVE PROBLEMS

For the continuous beam in Fig. 8-6, determine the values

811.
of the moments overthe supports. The supports are assumed to be rigid
or, what amounts to the same thing, to have equal deformations. This

assumption applies to all problems unless stated otherwise.

500 N

Figure 8-6.

Solution: Apply the three-moment equation to points over the sup-

ports. Since the supports remain at the same level, heights h and h 3 are x

zero, and the equation reduces to

M L + 2M 2 (L + L + 2)
ML 3 2
+ —
6A*a, 6A b
-^ + —
L
2
=*-*
2
= (a)
X X ]
L x 2
8-4 Application of the Three-Moment Equation 323

We begin by noting that the bending moment at support 1, caused

by the load to the left of R l9
is M x
= - 200 X 1.5 = - 300 N-m,
whereas over support R 3 the moment 3
is zero since no loads act to M
the right of R 3 Observe that the negative sign of
. must be retained M x

when substituting its value in the three-moment equation.*

The factors in Eq. (a) are found by using Table 8-1. The load on
span 1 is that of case 3; hence we have

=
^ wL = £ = N
'
(800)(3)3 2880 '
m2 (b)
"lT
For span 2, the factor is found by adding the results listed for cases 1

and 5; hence we obtain

=Mi) [ (4f-( 1
)V^[2(4f-(3)^]
= 1875 -I- 5175 = 7050 N m •
2
(c)

Substituting these results in Eq. (a) determines M 2, which is now the

only unknown. We obtain
- 300(3) + 2M2 (3 + 4) + 2880 + 7050 =
from which
9030
M 2
*
= — —
14
N-m
*vl«VT
= -645 Ans.

812. Determine the moments over the supports in the continuous

beam shown in Fig. 8-8.

*A common error involves applying the three-moment equation between the

overhang and span 1 as shown in Fig. 8-7, thus completely forgetting that the
200-N load causes the overhang to deflect downward an unknown distance h.
Under these conditions, the right-hand part of the general three-moment equation
contains the unknown h and is not zero. However, after M, and M2 have been
found, the general three-moment equation may be applied between spans and 1
to solve for the deflection h. This procedure will be discussed in Illustrative Problem
858 on p. 342.

200 N 800 N/m

1 I ^
h k V
\ *- Span 1 >
1 h
F Igure 8-7.
324 8 Continuous Beams

600 N 700 N

400 N/m 800 N/m 2m

3 m lml
* Span 1=4 m +\+- Span 2 = 3 m *{•• Span 3 = 4 m *|

Figure 8-8.

Preliminary: Writing the three-moment equation between spans 1 and

2, and between spans 2 and 3, we have

A/,L, + 2A/2 (L, + L 2 ) -I- ML 3 2

+ —f-± + —
6/1, J,6A
f-±
= 7 b>> ,
(a)
k

L x
L2

A/ 2 L 2 + 2M 3
(L 2 + L 3 ) + ML 4 3 + —-^ —j-± = -I- (6)

From the definition of bending moment, both Af, and M 4 are zero.
Hence Eqs. (a) and (6) are a pair of simultaneous equationsin M 2 and
M which can be solved when the values of 6Aa/L and 6Ab/L for the
3,

given loadings are known. Using Table 8-1, we compute these values as
follows:

^1 = ^(2L> - ft*) = «! [2(4)

2
- (3)
2
]
= 5175 N • m>

^
6i4,a
2"2
=
|^ 3
= >00)(3)> = 2880N. m >

= -- wL 3 = --(800)(3) J = 2520 N •
m'
60 60

=«[(4) 2
-(3)
2
]
+ «[(4) 2
-(2)
2
]

= 3150 + 4200 = 7350 N m •

2

Solution: The values just computed are substituted in Eqs. (a) and (b):

2A/ 2 (4 + 3) + 3M 3 + 5175 + 2880 =

3M + 2 2A/ 3 (3 + 4) + 2520 + 7350 =
or

14A/ 2 -I- 3A/ 3 + 8055 =

3M +2
14A/ 3 + 9870 = id)
8-4 Application of the Three-Moment Equation 325

Solving Eqs. (c) and (d) simultaneously for the two unknowns M 2 and
M 3
gives

M 2
= -445 N -m and M =-610Nm
3 Arts.

PROBLEMS

Unless otherwise stated, the continuous beams in the following

problems are supported on rigid foundations which are at the same
level. In each problem, determine the bending moments in the beam

over the supports.

813. See Fig. P-813.

900 N 1200 N/m -^

600 N/m 1 m 1 m
y

i 2m 3m
1
Rx R2 -R 3

Figure P-813.

814. See Fig. P-814. Ans. M 2

= - 262 N m •

1200 N/m ^
700 N

Figure P-814.

815. Determine the lengths of the overhangs in Fig. P-815 so

that the moments over the supports will be equal. Ans. x = L/V6

wN/m
T I J l T~

i?! R2 #3
Figures P-815 and P-816.
 J
326 8 Continuous Beams

816. Solve Problem 815 if one span is three-fourths the length

of the other span.

817. See Fig. P-817.

600 N 800 N/m

M=1200N-m

Figures P-817 and P-818.

818. In Problem 817, determine the changed value of the applied

couple that will cause M 2 to become zero.
Ans. M= 105 N m •
clockwise

819. See Fig. P-819.

Pa(L 2 - a 2 - + p)
2(«
Ans. M 2
— )

L 2
4(1 + a)(a + p) - a2
Pa(L 2 - a 2 )
M 3
= +
4(1 4- a)(a + p) - a'

alb
T* L »I-«-aL-J-. — |3L—
Figures P-819 and P-820.

820. Solve Problem 819 if the concentrated load is replaced by a

uniformly distributed load of w N/m over the first span.

wL 2 + p)
2(a
Ans. M 2
=
4 + a)(a + P) - a 2
4(1
2
wL a
M 3
= +
4 + a2
4(1 «)(« + fi)-
821. See Fig. P-8211.

3 PL +2j8
Ans. M 2
=
8 4(1 f «)(1
1

+ - 1
/8)

3 PL 1 +2o
M.3 =

8 4(1 -h a)(l + P) - 1
8-4 Application of the Three-Moment Equation 327

L/2 fL/2

U-aL J* L JL— (3L—J

Ri R2 Rs R4
Figures P-821 and P-822.

822. Solve Problem 821 if the concentrated load is replaced by a

uniformly distributed load of w N/m over the middle span.

wL +20
Arts. M2 - 4 4(1 -1-
1

a)(l + P) - 1

— wL 2 1 +2a
M 3
4 4(1 + a)(l + p) - 1

823. A continuous beam simply supported over three 4-m spans

carries a concentrated load of 2 kN at the center of the first span, a
concentrated load of 3 kN at the center of the third span, and a
uniformly distributed load of 900 N/m over the middle span. Solve for
the moments over the supports and check your answers using the results
obtained for Problems 819 and 822.
824. The span of a simply supported continuous beam is 4
first

m long, the second span is 2 m

long, and the third span is 4 long. m
Over the first span there is a uniformly distributed load of 2 kN/m, and
over the third span there is a uniformly distributed load of 4 kN/m. At
the midpoint of the second span, there is a concentrated load of 10 kN.
Solve for the moments over the supports and check your answers, using
Problems 820 and 821.
Ans. M 2
= - 2900 N • m; M 3
= - 6100 N m
•

825. See Fig. P-825.

6 kN 3 kN
900 N/m

2m 2m 5 m
I f
Figure P-825.
328 8 Continuous Beams

826. See Fig. P-826.

Ans. M 2
= - 1690 N •
m; A/ 3 = - 3230 N m •

6kN
8kN-m

jn 2kN/m
1
2m 2m m m m
vU 4 2

*i
f R4
Figure P-826.

827. See Fig. P-827.

8kN ^10kN/m

6kN/m
1
2m A 4m 1 2m 2ml 6 m

*4
Figure P-827.

8-5 REACTIONS OF CONTINUOUS BEAMS;

SHEAR DIAGRAMS

The major reason for computing the reactions of continuous beams is to

be able to draw the shear diagram. Two methods of computing reac-
tions are available: in one, reactions are computed by using the defini-
tion of bending moment; in the other, the reaction is divided into parts
from which the shear diagram can be drawn easily. The second method
is preferred for reasons which will be given later. In both methods the

moments over the supports must first be determined.

As an example of the first method, consider the beam in Fig. 8-9
whose moments over the supports were found in Illustrative Problem
812 to be 2
M
= - 445 N-m and 3 = - 610 M Applying the Nm.
definition of bending moment, we express 2 in terms of M
the moments

600 N 700 N

400 N/m 2 m
3 m
Span 1 = 4 m 15 Span 2 = 3 m I Span 3 = 4 m- 3
Flgure 8-9.
8-5 Reactions of Continuous Beams; Shear Diagrams 329

about R2 of all loads to the left ofR 2 and obtain

[M2 - (2M) L ] M 2
= -445 = 4R - X
(400 X 3) X 2.5

whence
R - x
639 N Ans.

To determine R 2 we apply , the definition of M 3 to moments about

R3 of all loads to the left of R3 , as follows:

[M3 -(2M) L ]
- 610 - 1R - *°°* 3
X
(400 x 3) x 5.5 + 3R 2 - I \ X| x 3

Substituting in this relation the now known value of /?, = 639 N, we

find that

R2 = 1306 N Ans.

The value of R4 is also obtained from the value of M3 by

expressing M 3
in terms of the moments about R 3 of all loads to the right
of R3 :

[M3 = (2M) R ]

- 610 = 4R 4 - 700 X 2 - 600 X 1

whence
r4 = 348 N Ans.

The value of R3 can now be found by taking a vertical summation

of all forces acting on the entire beam. This gives
[zr = o]
800 X ^
R + R2 + R3 +
Y
fl 4 = 400 X 3 + +600 + 700

639 + 1306 + fl, + 348 = 1200 + 1200 + 600 + 700

whence
R3 = 1407 N Ans.

It is method carries through any numerical error

evident that this
and is also tedious more than three spans to the beam. An
if there are
alternate method eliminates both of these objections and presents the
results in a form suitable for drawing the shear diagram rapidly. This
alternate method depends on isolating each span and determining the
supporting shear forces at the end of each one.
In Art. 8-2 and Fig. 8-1 we saw that any span can be isolated as a
free body by applying to it the proper values of end moments and
shears. The isolated span can then be resolved into a simply supported
beam carrying the given loads, plus another beam loaded only by the
end moments and couple reactions. Span 2 of Fig. 8-9 is thus resolved
330 8 Continuous Beams

W-1200N
800 N/m

Af3 « (a) Free-body diagram

of span 2
610
N-m

W- 1200 N
I

(b) Load on span 2

simply supported

800 N 400 N

AW^F
445 y 610
3
=
End moments and
— L2 = 3 m J m N
(c)
•
couple reactions

fl'=55N #' = 55 N
Figure 8-10. Component loadings on span 2.

into its component parts in Fig. 8-10. Because the end moments 2 M
and M 3
are negative, they act as shown, and their absolute magnitudes
may be used. The term V2 denotes the vertical shear in the beam to the
right of R2 , and V3 is numerically equivalent to the vertical shear in the
beam to the left of R 3 The minus
. sign in the subscript of V_ 3 indicates
that it acts opposite to the actual vertical shear in order to create
equilibrium in span 2.

Since parts (b) and (c) of Fig. 8-10 are superposed to form part
(a), it follows that the actual end shears V2 and V_ 3 are the algebraic
sum of the equivalent simple beam reactions and the couple reactions
R In this example,
'.
3 is M
numerically larger than M
2 hence there is an
\

unbalanced clockwise couple acting on part (c) of magnitude 3

— 2 M M .

It can be balanced only by the counterclockwise moment caused by the

couple reactions R' acting at the supports and having a moment arm
equal to the length L 2 of the span. Evidently the numerical value of R
'

is given by

R'L 2 = M 3
- M 2

or
M 3
- M 2 610 - 445
R' m = 55 N
The couple reaction R acts upward at the larger moment 3
and
' M
downward at the smaller moment 2
In the algebraic summation of M .

reactions referred to previously, we take the upward direction as posi-

tive and the downward direction as negative.
 8-5 Reactions of Continuous Beams; Shear Diagrams 331

M,
c
P
V )*l '
M( 8
s
^ f
I
|AT#

R' R' #' R'

(a) Negative end moments (b) Negative moment Mj and
positive moment s M
Figure 8-11. Couple reaction acts upward at the end having the larger negative
moment.

Generalizing this discussion, we may state that the couple reaction

R' on any span is given by

R = ^ZJL (8-2)

where M t
is end moment on the span,
the larger absolute value of s
is M
the smaller absolute value of end moment, and L is the length of the
span. As a rule, the couple reaction R' acts upward at the end of the
span having the larger absolute value, and downward at the other end.
This assumes that negative moments act over the supports. If one
support moment is actually positive in sign, the negative moment is
taken as the larger absolute value and its numerical value is used as M,
in Eq. (8-2). The proof of these statements is shown by the free-body
diagrams in Fig. 8-11.
A convenient way of arranging the values of simple beam and
couple reactions is shown in Fig. 8- 12a. The couple reactions were

80( )N/m^^ 600 N 700 N

400 N/m ^^ lm{lm{ 2m
1
3m 1 m > 3m 4 m
Simple beam 450 150
reaction 750 450 800 400 350 350
Couple reaction -111 111 - 55 55 152 -152
Vertical shear 639 561 745 455 952 348
(a)

952

639

Figure 8-12. A second method of computing reactions and drawing shear dia-
gram.
332 8 Continuous Beams

computed as follows, the numerical subscripts referring to the span on

which they act. The couple reaction acts upward (or is given a positive
sign) at the end having the larger numerical negative end moment.

M, - M. 445 -
R' = *l
= 111 N
4

-
*J
= 610 445 = 55 >
3

610 -
*3 = A
152 N

The shear diagram in Fig. 8- 12b may now easily be plotted.

Remember that the values of vertical shear in Fig. 8- 12a which act to
the left of the supports are equal but opposite to the actual vertical
shear in the beam. This accounts for the minus values in the shear
diagram. If the values of the reaction are desired, they may be obtained
by adding the vertical shears acting at the reaction. Thus

R x
= 639 N
R2 = 561 + 745 = 1306 N
R3 = 455 + 952 - 1407 N
R4 - 348 N
which agree with the values determined by the first method.

PROBLEMS

In the following problems, determine the reactions and sketch the

shear diagrams. Then compute maximum vertical shear V
the values of
and maximum positive bending moment M. In solving the problems,
use the moments determined in the reference problems unless otherwise
instructed.

828. A continuous beam carries a uniform load over two equal

spans as shown in Fig. P-828.
Arts. M 2
= - wL 2 /S; R - R 3 =|wL; R 2 =\wL
{

w N/m

t
*i «2
Figure P-828.
8-6 Continuous Beams with Fixed Ends 333

829. A uniform load is carried over three equal spans as shown

in Fig. P-829.
wL 2
Arts. M 2
= M 3
= - -r7r R = R 4 = OAwL; R 2 = R 3 =
10
;
x
l.lwL

u/N/m
^ L i
\ t
R\ R% R$ R4
Figure P-829.

830. Refer to Problem 814.

831. Refer to Problem 817 for which M2
= + 156 N • m.
Ans. R2 = 109 N; max. +M = 452 N m •

832. Refer to Problem 824.

833. Refer to Problem 825 for which M 2
= - 2.04kN-m and
M 3
- - 2.81 kN • m.
834. Refer to Problem 826.
Ans. R2 = 6.04 kN; *3 = 9.46 kN; max. + M = 3.15 kN m •

835. Refer to Problem 827 for which M 2

= - 1.895 kN m • and
M 3
= - 16.42 kN •
m.
For the continuous beam loaded as shown in Fig. P-815 on
836.
page 325 determine the length x of the overhangs that will cause equal
,

reactions. Ans. x = 0.44L

8-6 CONTINUOUS BEAMS WITH FIXED ENDS

For continuous beams with fixed ends, assume the fixed end to be
equivalent to an imaginary span with an imaginary loading. The three-
moment equation, when applied to the beam, includes this imaginary
span; however, all the terms that refer to the imaginary span have zero
values.
The foregoing statement is easily proved by using the last span of
a continuous beam as shown in Fig. 8- 13a. The moment at V is M x x

due to the loads on the beam that he to the left of V The right end at B x
.

is assumed to be perfectly fixed, that is, a tangent to the elastic curve at

B will be perfectly horizontal. The end may
effect of the perfectly fixed
be duplicated by adding the reflection of the loads (that assuming is,

that the wall at B is a mirror), as shown in Fig. 8- 13b. Because of the

symmetry of loading thus obtained, the tangent to the elastic curve at B
will be horizontal, which is the effect given by a perfectly fixed end.
334 8 Continuous Beams

Perfectly
P fixed end r P
! M x

G I
^"-"I"^'
(Span 1) (Span 2)
*2 VI

(a) (b)

Figure 8-13. A fixed end is equivalent to an imaginary span.

Applying the three-moment equation to spans 1 and 2 of Fig.

8- 13b gives

A/,L, + 2A/ 2 (L, -I- L2 ) -I- A/ 3 L 2 + —

6A,a*
f-± +
—
6Ajbj
-^- =
.

(u)
.

whence, on substituting values corresponding to those in the figure, we

obtain
6A a 6A a
ML X
+ 2M 2
(L + L) + ML+ X
x x

L
x ]

=
x

or
6^4, a
m
2A/,L + 4M L + 2X
2
i i

Dividing by 2, we obtain

A/,L + 2M L +2
—j-*- - (*)

This would have been obtained from Eq. (a) at once if zero had been
substituted for all the terms referring to the imaginary span (span 2 in
this example). The principle that a fixed end is equivalent to an
imaginary span has thus been proved and will now be applied to several
examples.

ILLUSTRATIVE PROBLEMS

837. Find the moments over the supports for the propped beam
in Fig. 8-14. The right end is assumed to be perfectly fixed.

Solution: This problem canbe solved by the basic area-moment

method, but can be solved more easily and more quickly by consider-
it

ing the fixed end to be equivalent to an imaginary span.

The three-moment equation is applied between spans 1 and 2,
whose supports are at the same level, so we obtain
6A
A#,L, -I- 2A# 2 (L, + L2 ) + A/ 3 L 2 + —
6A
-*-±
}
a,
+ — ib-,
-2-1 -
8-6 Continuous Beams with Fixed Ends 335

w,
400 N/m (3)
%
lmf\ (1) 4m (2)
Sk
(Spanl) (Span 2)
(Imaginary)
1R
3

Figure 8-14.

The moment at R }
due to the overhang is

M l
= - (400 X1)X{= -200 N-m

and zero is substituted for all terms referring to span 2. Table 8- 1 shows
that 6A a l l
/L = wL 3 /4,
l
so the three-moment equation reduces to

400 X4 = 3
- 200 X 4 + 2M, X4 +
from which
A/2 = -700 N-m Ans.

838. Find the moments over the supports for the continuous
beam in Fig. 8-15. Both ends of the beam are assumed to be perfectly
fixed.

Solution: The perfectly fixed ends are considered equivalent to the

imaginary spans and 3. Writing the three-moment equation for spans
and 1, for spans 1 and 2, and for spans 2 and 3, we have

ML + 2M (L + }
Z,,) + ML 2 t
+
6A a
— p- n
2 + ^=
6Aib,
— («)

y-^- + —^ =
6A*a, 6A
ML X X
+ 2M2 (L + L2 + l
) ML 3 2 + — 1 b-y
(*)

ML 2 2 + 2M (L 2 + L + 3)
A/4 L 3 + —
6A
y-± + —f-± =
1 a-, 6A->b*
3
L2 £3

400 N
** fe
2m I 2m 600 N/m f/^
« 4 m >\ 3m /T^
\
(0)
^ (1) (2) (3) (4) J

! (Span 0) ^ (Span 3)
K*
R (Imaginary) VY (Spanl) *2 (Span 2) V3 (Imaginary)

Figure 8-15.
336 8 Continuous Beams

In applung these equations, we neglect any terms referring to the

imaginary spans.
Using Table 8-1, we next compute the following values:

6/!,/?,
j^ (L 2 _ 400 x2 - =
b
2
}
m (i 6 4) 2400 N m2•

6A a l ]

>'--- 4
1(16 - 4) = 2400 Nm 2

3
6A 2 a 2 6A?b
u wL' 600 X 3
2 2
= 4050 N m2

Substituting these values in the three-moment equations gives

From Eq. (a) 8A/, + 4A/2 -I- 2400 = (d)

From Eq. (b) 4A/, -I- 14M2 3M3 + 6450
-I- = (e)

From Eq. (c) 3M 2

+ 6A/3 + 4050 = (/)
Solving Eqs. (d), (e), and (/) simultaneously gives

A/ ]== -147Nm, M 2
= -307 N-m, M 3
= -522 N-m Ans.

PROBLEMS

In the following problems, the ends of the beams are assumed to

be perfectly fixed by the walls against rotation. All supports are
assumed to remain at the same level.

839. Determine the prop reaction for the beam in Fig. P-839.

4TPT4
Figure P-839.

840. For the propped beam shown in Fig. P-840, determine the
prop reaction and the maximum positive bending moment.
Ans. R = 2.96 kN; M = 696 N •
m

1200 N/m

1 m i 2m 2m

Figure P-840.
8-6 Continuous Beams with Fixed Ends 337

841. Determine the wall moment and prop reaction for the beam
shown in Fig. P-841. Arts. M= - 1.35 kN m; R = 2.33 kN

2kN/m

Figure P-841.

842. For the propped beam shown in Fig. P-842, determine the
wall moment and the reaction of the prop support.

2 kN/m

843. For the propped beam shown in Fig. P-843, determine the
wall moment and the prop reaction.
Arts. M= - 1566 N •
m; R= 359 N
2 kN/m

Figure P-843.

844. In the propped beam shown in Fig. P-844, determine the

prop reaction.

6kN
Af = 4kN-m
4 kN/m

Figure P-844.
338 8 Continuous Beams

845. Compute moments over the supports

the for the beam
shown in Fig. P-845 and then draw the shear diagram.

BOON 900 N

600 N/m
[
2m 4 m 2m 2m
T A
«i
Figure P-845.

846. Sketch the shear diagram for the continuous beam shown in

Fig. P-846. Ans. M 2

= -2.52 kN-m; M 3

900 N

2m
600 N/m

m 2m 2m
6

m
Figure P-846.

847. Compute the moments over the supports and sketch the
shear diagram for the continuous beam shown in Fig. P-847.
3 kN 5 kN/m

Figure P-847.

848. Determine the support moments and reactions for the beam
shown in Fig. P-848.
9kN
4 kN/m

Figure P-848.
8-6 Continuous Beams with Fixed Ends 339

849. Find the moments over the supports for the beam shown in
Fig. P-849.
Arts. M x
= - 300 Nm; A/2 = - 1500 N •
m; M3
= -2700N-m
4kN/m

850. Determine the moments over the supports for the beam
loaded as shown in Fig. P-850.
wL 2
— 2 + 3a wL 2 2
.

Ans. M, =
8
•
^
3 + 3a
; MX4 2
=
8 3 + 3a '

M, = +
wU
8 3 + 3a

u;N/m
L i i
ah

Figures P-850 and P-851.

851. Replace the distributed load in Problem 850 by a con-

centrated load P at midspan and solve for the moments over the
supports. Ans .
Mi= _^. 16
2^3a
3 + 3a
. _ 3PL
16 3 + 3a
'

3PL 1
M, = +
16 3 + 3a
852. Use the results of Problems 850 and 851 to check the
answers to Illustrative Problem 838.
For the continuous beam shown in Fig. P-853, determine
853.
the moments over
the supports. Also draw the shear diagram and
compute the maximum positive bending moment. (Hint: Take advan-
tage of symmetry.)

2kN/m

3m 4 m 3 m

Vi
Figure P-853.
340 8 Continuous Beams

854. Solve for the moments over the supports in the beam loaded
as shown in Fig. P-854.
wL wL'
Arts. A/, = M A = +
12 2 + a
; M 2
= M 3
= -
12 2 + a

wN/m
erL crL

i?2 *3
Figures P-854 and P-855.

855. If the distributed load in Problem 854 is replaced by a

concentrated load P at midspan, determine the moments over the
supports.

856. For the beam shown in Fig. P-856, determine the moments
over the supports. Also draw the shear diagram and compute the
maximum positive bending moment.
Ans. M 2
= - 4460 N- m; M 3
- - 661 Nm; M 4
= - 3270 Nm
2kN 2kN 6kN/m

4kN/m lm 2m
1 m i 4 m 4 m

Figure P-856.

8-7 DEFLECTIONS DETERMINED BY THE

THREE-MOMENT EQUATION

Before discussing the use of the general three-moment equation to find

deflections let us review a few facts. The three-moment equation de-
termines the relation among moments at any three points in any
the
beam. These three points determine two segments of the beam, and the
terms 6A a /L and 6A 2 b 2 / 2 of the three-moment equation refer to
] l l

the moment diagram resulting from the loads acting on these segments.
The loads are assumed to be simply supported on spans that are as long
as the segments. Heights h
]
and h 3 refer to the heights of points 1 and 3
relative to point 2 (see Fig. 8-3, page 315); the heights are considered
positive if above point 2 and negative if below it.
8-7 Deflections Determined by the Three-Moment Equation 341

The general method for determining deflections by means of the

three-moment equation is to select points 1, 2, and 3 so that either (or
both) of the heights h x
and h 3 is equal to the desired deflection. This
occurs when two of the points are selected over supports and the third is

chosen at the location where the deflection be determined. The

is to
values of the moments at points 1, 2, and 3 must first be known or
computed. This method will now be illustrated.

ILLUSTRATIVE PROBLEMS

857. Use the three-moment equation to determine the value of

EI8 at 1 m from the left support of the beam loaded as shown in Fig.
8-16.

Solution: The exaggerated position of the deflection curve is shown by

the dashed line. Selecting point 2 at the position of the desired deflec-
tion and points 1 and h and h 3 each equal
3 at the reactions will make x

to the desired deflection. Also, since above the horizontal 1 and 3 are
line through 2, heights h and h 3 will be positive. The segments into
x

which the beam is divided by the three points are designated as span 1
and span 2.
The general three-moment equation is
6A a 6A 2 b-
ML X X
-I- 2M2 (L + L2 +X ) ML 3 2 + x x

= 6EI
(M)
Since span 1 in Fig. 8-16 is unloaded, 6A a /L
x x x
is zero. For span 2,
case 5 of Table 8- 1 gives

6A,b
2 u2 wd 2 ,w^2
L
(2L>-^)=«-(2x9-4) = 4200 Nm 2

By taking moments about R2 we , determine that R =

450 N. Having
x

found R l9
the bending moment at point 2 is 2 M450 X 1 = 450
—
N m.
• Since M x
= M 3
= and h x
= h3 = 5, substituting the above

(i) 2m 2 m
900 N/m T
h --^ i5

(2)-
Ri — lm — 3 m
(Span 1) (Span 2)

Figure 8-16.
342 8 Continuous Beams

values in the three-moment equation yields

+ 2(450)(1 + 3) + + + 4200 = 6£/(| +

j)
which reduces to

Eld - 975 N m •
3
Ans.

858. Determine the value of EI8 under the 200 -N load of the
continuous beam shown in Fig. 8-17.

Solution: This beam is the one for which the support moments were
determined in Illustrative Problem 811. Here we select points 0, 1, and
2, as shown, between which to write the three-moment equation. Point

on the elastic curve is below point 1 hence h = — 8 and h 2 = 0. Using ;

the results of Illustrative Problem 811, we find that M = 0, M =

-300 Nm, and M 2
- - 645Nm.
Applying the three-moment equation, we write

6 A b*
ML + 2A/,(L -I- L,) + M L, + —
2
6A a
f-± +
n n }

unloaded and hence 6A a /L

= 6EI
Uo M
In this instance span is is zero. Case 3
of Table 8-1 gives for span 1

6i4,£,
—r- 1
7
= ttt^L 3 =
60
*
—
7
60
X 800 X 27 = 2520 N m 2

Substituting these values in the three-moment equation, with careful

note of the minus signs of M x
and M 2, we obtain

2(- 300)0.5 + 3) + (-645) x3 + 2520 = ^Hjf + o)

whence
EI8 = 529 N m •
3
Ans.

500 N
200 N

Figure 8-17.
8-7 Deflections Determined by the Three-Moment Equation 343

The positive value of the result indicates that the deflection is down-
ward as assumed.

PROBLEMS

859. Determine the value of EI8 under P in Fig. P-859. What is

the result if P isreplaced by a clockwise couple M?

Arts.
^ = -^—
EI8
PLb 2

;
^ ro
Eld = MbilL
^-z
+ b)
L
3 6

>+*— 6-J

Figure P-859.

860. Determine the value of EI8 at the end of the overhang and
midway between the supports for the beam shown in Fig. P-860.

w N/m
77777T OOP
>frf/f

Figure P-860.

861. For the beam shown in Fig. P-861, determine the value of
EI8 at 1 m and 3 m from the left support. Ans. 900
3
N m
767 N- 3 •
; m
800 N
M = 600N-m
lm } 2m lm

it
Figure P-861.

862. Determine the value of EI8 at B for the beam shown in Fig.

P-862. Ans. EI8 = ^(4L - 3a)(L - a)

B w N/m
a ,

f Ri
Figure P-862.
344 8 Continuous Beams

863. For the beam shown in Fig. P-863, determine the value of
EI8 midway between the supports and at the left end.

4kN

2kN/m
i
2m A2m 2m
I
Figure P-863.

864. A 6-m beam, simply supported

at 1 from each end,m
800 N/m over its entire length.
carries a uniformly distributed load of
Compute the value of EI8 at the middle and at the ends.
865. For the beam shown in Fig. P-865, compute the value of
EI8 at x = 3m and at the end of the overhang.
Arts. At overhang, EI8 = 813 N m 3 down •

800 N/m

Figure P-865.

866. Determine the midspan value of EI8 for the beam shown in
Fig. P-866.

1200 N/m
Af = 4600N-m

n
p=ic 3 m

Figure P-866.
8-7 Deflections Determined by the Three-Moment Equation 345

867. For the beam in Fig. P-867, compute the value of P that
will cause a zero deflection under P. Ans. P= 489 N

900 N/m

Figure P-867.

868. Determine the values of EI8 at midspan and at the ends of

the beam loaded as shown in Fig. P-868.
Ans. At ends, EI8 = 10.9 kN • m 3
up

4kN/m

R l
R2
Figure P-868.

869. Find the value of EI8 at the center of the first span of the
continuous beam in Fig. P-869 if it is known that M 2
= — 2040 N m •

andM3 = -2810N-m Ans. EI8 = 120Nm3 up

6 kN 3 kN
0N/msy
N^lmjlmylm 800 N/m

i 2m 2 5 m -

"t t
R4
Figure P-869.
346 8 Continuous Beams

870. Compute the value of EI8 at the overhanging end of the

beam in Fig.P-870 if it is known that the wall moment is + 1 100 N m. •

2 kN/m

Figure P-870.

871. The continuous beam in Fig. P-871 is supported at its left

end by a spring whose constant is 50 kN/m. For the beam, E= 10 X

10 N/m and / = 40 X 10
9 2 6
mm4 . Compute the deflection of the spring.

Arts. 8 - 79.4 mm
872. Repeat Problem 871 assuming that the loadings on the
spans are interchanged.

6kN

3 kN/m 1 2m f 2m
4 m

Figures P-871 and P-872.

8-8 MOMENT DISTRIBUTION

Modern techniques of designing continuous structures are based on a
method of successive approximations popularized by Hardy Cross.*
This method, which is widely known as the moment-distribution method,
is applicable to all types of rigid-frame analysis. Its application to
continuous beams will serve to introduce this powerful tool of the
structural engineer.
Several preliminary concepts are necessary. The first, the carry-
over moment, is defined as the moment induced at the fixed end of a
beam by the action of a moment applied at the other end. Thus consider
the beam in Fig. 8-l8a which is perfectly fixed at A and hinged at B. A
M
moment B applied at B flexes the beam as shown and induces the wall

•See Cross's papers, Continuity as a factor in reinforced concrete design.

Proc. A.C.I, pp. 669-711 (1929); Simplified rigid frame design. Proc. A.C.I. 26,
170 183 (1930); Analysis of continuous frames by distributing fixed end moments.
Trans. A.S.C E. 96, 1-156 (1932). See also H. Cross and N. D. Morgan, Continuous
Frames of Reinforced Concrete, Wiley, New York, 1932.
8-8 Moment Distribution 347

(a)

Figure 8-18. Carry-over moment and beam stiffness.

moment M A The moment diagram, which is drawn as described in Art

.

7-5, is M
shown in Fig. 8- 18b. Although A is actually negative (because
of the downward curvature of the elastic curve at A), it is convenient to
consider it positive as shown; consequently the solution will determine
not only its absolute value but also the correct sign (negative).
The deviation at B from a reference tangent drawn at A is zero
because of the perfect wall constraint at A Hence .

[EItB/A = (area)^ xB ] ~{\MA L){\L) + (\MB L)(±L)

whence

MA = \MB (8-3)

This result means that a moment applied at the hinged end B "carries
over" to the fixed end A a moment that is half the amount and of
opposite sign.
A second concept needed for the moment-distribution method is
beam Beam stiffness is the moment required at the simply
stiffness.

supported end of a beam to produce unit rotation of that end, the other
end being rigidly fixed. Note that this definition implies no relative
linear displacement of the two ends of the beam.
The slope at B in Fig. 8- 18a is found from the first theorem of the
area-moment method and is expressed in terms of the moment diagram
in Fig. 8-18b:

[
ei9ab (area)^] EIO =\MA L +\MB L
Replacing MA by — \ MB from Eq. (8-3) gives

MB = 4EI9
As was said above, the value of MB when equals 1 radian is known as
the beam stiffness. It varies with the ratio I / L as well as with E. It is
348 8 Continuous Beams

(a) Actual loading

Y/ (b) Fixed end moments (FEMe)

caused by P and Q assuming
support B locked against ro-
tation

(c) Unlocking B distributes difference

in FEM at B according to relative
stiffness of AB
and BC
and causes
carry-over moments as shown.
(Note L mu8t equal R to ensure
continuity)

(d) Final moments are algebraic

sum of values in (b) and (c)
-30 •70 -105

Figure 8-19 Qualitative description of moment-distribution procedure.

denoted by the symbol K\ hence

4£7
absolute K= —=— (8-4)

However, in many structures E remains constant, so only a relative

measure of resistance to rotation is required. This may be called relative
beam stiffness and is expressed by

relative K= — (8-5)

We are now ready to describe qualitatively the moment-distribution

procedure.
The continuous beam in Fig. 8- 19a is perfectly restrained at A
and C and simply supported at B. Assume that at B the beam is

temporarily locked or rigidified against the rotation caused by the loads

P and Q. Under these conditions, segments AB and BC will act as fixed
ended beams subjected to the fixed end moments caused by loads P and
Q. These fixed end moments (hereafter abbreviated as FEM) are
assumed to have the values shown in Fig. 8- 19b.
If the support at B is now released or unlocked, the difference in
FEM between sections to the left and right of B creates an unbalanced
moment of 30 N m which causes the beam at B to rotate, as shown in
•

Fig. 8- 19c, until the moments at B are balanced. Obviously, the

8-8 Moment Distribution 349

moment to the left of B

be increased by some amount, say 20
will
N-m; and B
the moment will be decreased by the
to the right of
remaining 10 N -m of the 30 N m
difference between the FEM at B
in Fig. 8- 19b. Thus the unbalanced moment is distributed at the
unlocked support. The rotation of B caused by these distributed mo-
ments induces, at A and C, carry-over moments of half the amount and
of opposite sign. These carry-over moments are indicated by the arrows
in Fig. 8- 19c.
The unbalanced moment at B is fixed
ratio of distribution of the
by the two beams must rotate through the same angle at B.
fact that the
This means that the unbalanced moment must be distributed in the
ratio of the stiffness factors of the adjacent beams. The ratio of
distribution to any beam is called a distribution factor, DF, and is
defined by

where K is the stiffness factor for that beam and SA' is the sum of the
stiffness factors for adjacent beams. If the beams are of the same
material (as is generally the case), only relative K need be used. Further,
if they are of the same cross section, relative K (that beam stiffness)
is,

is inversely proportional to the length (see Eq. 8-5). In distributing the

FEMs the object is to secure balance at the unlocked support.
The final moments in Fig. 8-19d are obtained by superposition of
the FEMs in Fig. 8- 19b and of the distributed moments and carry-over
moments in Fig. 8- 19c.

Sign Convention

In the preceding discussion, signs were based on conventional

bending moments. This convention requires the carry-over moment to
be of opposite sign and often leads to confusion regarding the magni-
tude and sign of the unbalanced moment to be distributed, especially
when more than two members frame into a common joint. Computa-
tional accuracy is increased and confusion eliminated by using an
alternate sign convention based upon the sense of rotation of the end
moments.
With this convention, counterclockwise moments
acting on the
beam are considered to be positive, and clockwise end moments are
negative. As a result of this alternate convention, several minor modifi-
cations occur. The first is that carry-over moments are of the same sign.
The second is that in distributing the unbalanced moments at each
support, the distributed moments are of the same sign and are so applied
350 8 Continuous Beams

-M A -MB +MA -Af B

(a) Signs based on conventional (b) Signs based on rotational
restraining moments sense of restraining moments
Figure 8-20. Difference between sign conventions.

as to make the algebraic sum moments at a support or joint equal

of the
to zero. Finally, in a fixed-ended beam carrying downward loads as in
Fig. 8-20, the fixed end moments will be positive at the left end and
negative at the right end.
Sometimes the alternate sign convention is based upon the rota-
tion of the end joint rather than the member. Then a moment tending to
rotate a joint clockwise is considered positive. However, since the action
of the member upon the joint is equal but opposite to that of the joint
upon the member, this convention based on joint rotation is exactly
equivalent to that of moment on the member.
rotation
The moment-distribution method may be summarized in the
following steps:

1. Assume that all supports are fixed or locked and compute

fixed end moments for each span considered separate from every other
span. Table 7-2 (page 307) will be helpful in computing these FEMs.
Unlock each support and distribute the unbalanced moment at
2.

each one to each adjacent span by means of Eq. (8-6). Then relock each
support.

3. After distributing the unbalanced moment to each adjacent

span, carry over one-half this amount, with the same sign, to the other
end of each span.

This completes one cycle of distribution. Steps 2 and 3 must be

repeated because of the new unbalance caused by the carry-over mo-
ments. Such repetitions are made until the carry-over moments become
zero or negligibly small. The process may be stopped when any distribu-
tion is completed, the accuracy of the final results depending on the
number of cycles. As a rule, no more than four cycles are necessary,
since the unbalance caused by the carry-over moments usually de-
creases rapidly to zero.
The following illustrative problems show the method of recording
results, as well as some suggested modifications or shortcuts.
8-8 Moment Distribution 351

ILLUSTRATIVE PROBLEMS

873. The continuous beam of constant cross section and material

shown in Fig. 8-21 is perfectly restrained at the ends. Compute the
moments over the supports.

Solution: Although / is not specified, it is convenient to take / as a

common multiple of the span lengths, that is, 12 units. Then the values
of relative stiffness K= I/L are shown, and the distribution factors
(DF) are computed from Eq. (8-6) and also listed.
Assuming all supports locked, and using Table 7-2 (page 307), we
compute values for the fixed end moments (FEM):
:

Pab'
Span AB: MA = - 4500(4)(2)
= -2000 N-m
L2 (6)
2

^ .2

L2
= _ 4500(4r(2)
(6)
2
= _ 400()N>m

Span^C: MB = Mc = - ^ - - ^^t = -8000N- m

Using the rotation sign convention, these FEMs are inserted in the table
in Fig. 8-21 with a "plus" sign at the left end of each span and a
"minus" sign at the right end.
With B unlocked, the unbalanced moment is the numerical dif-
ference between the FEM at B, or 8000 - 4000 = 4000 N m. Using •

4500 N

4m I 2m 6000 N/m

m
I
6 m
% 4
%
"i T=© ?-@
DF 0.4 0.6

FEM + 2000 -4000 +8000 -8 000

— 1600 -2400^
Carry-over -800^ ^-1200
Summation or -5600
+ 1200 + 5600 -9200
moments
final

Figure 8-21.
352 8 Continuous Beams

the values of DF, we distribute part of this unbalanced moment to the

left of B as 0.4(4000) = 1600 N m • and the remainder to the right of B
as 0.6(4000) = 2400 N m. As•
indicated in the table in Fig. 8-21, the
signs of these distributed moments are chosen to be minus so that the
sum of the moments at B will become zero. One-half the values of these
distributed moments are now carried over with the same sign. Thus
- 1600 applied to the left of B is carried over as -800 to A, and -2400
applied to the right of B is carried over as — 1200 to C.
Since A and C and are specified as remaining
are locked or fixed
so, they absorb these carry-over moments and the distribution is com-

pleted. The final values of the moment at each support are obtained by
algebraic summations of each vertical column, giving the results shown.
If desired, the moments in the final summation are easily converted

back to conventional bending moments by merely changing the sign at

the left end of each span.
The continuous beam in Fig. 8-22 carries
874. the same loads as
the beam Problem 873, but the ends
in Illustrative at A and C are
simply supported. Compute the support moments.

Solution: Values of K and DF are computed and listed as in Illustra-

tive Problem 873. Assuming all supports locked, the FEMs are also

4500 N

4 m 2m 6000 N/m
I

6 m
? 4 m

K=
12
6
® 12
4 ®
DF 0.4 0.6

FEM 2000 4 000 8000 -8000

1st distribution

Carry-over
2000 1600

1000
2400

4 000
^ + 8000

1200
2nd distribution

Carry-over
3rd distribution
1200

400
400
1800

+ 600
-600
X 1200

900
900

C.irrv-over 300 450

>e 300
4th distribution 300 450 300

7 800 7 800

Figure 8-22.
 8-8 Moment Distribution 353

computed and listed, with "plus" signs at the left end of each span and
"minus" signs at the right end.
All the supports are now unlocked, which restores the beam to the
specified conditions at each support. The unbalanced moment at each
support set up by unlocking the supports must now be distributed. At B
the distribution is as described in Illustrative Problem 873; but releasing

6000 N
9000 N/m

6000 N/m 2m } 2

4 nv
I 4 m-
% 4 m

K -L
® ® ©
DF 1 0.5 0.5 0.5 0.5 i

FEM +8000 -8000 + 3 000 -3000 +4 800 -7 200

1st distribution -8000 ^^, +2 500 + 2 500 ^^ -900 -900
x^ + 7 200

Carry-over + 1250 > ^ -4 000 -450 >^V + 1250 + 3 600 -450

2nd distribution -1250 ^^ +2 225 + 2 225 \/ -2425 -2425 +450

Carry-over + 1113 ^^ -625 -1213 J

<*^ + 1113 +225 -1213
3rd distribution -1113 +919 +919 -669 -669 + 1213

+460
\s
**< -557 -335 <*^*<
:xr +460 +607 y^ -335
Carry-over
4th distribution -460 +446 +446 -533 -534 + 335

2 -7 092 + 7 092 -4 704 +4 704

Solution (a) Regular procedure

*-i
3
4
xi -<D ® 3
4
*>-©
3 4 4 3
DF 1
7 7 7 7
1

FEM +8000 -8000 + 3000 -3000 +4 800 -7 200

Release A &D -8000 -4 000 + 3 600 -+ +7 200

Adjusted FEM -12 000 + 3 000 -3000 + 8400

1st distribution -3 086 -2 314

+ 3857 + 5 143
\^
Carry-over -1543 + 2572
2nd distribution
+661 +882 -1470 -1102

Carry-over
-735 +441
3rd distribution +420 -189
+ 315 -252

Carry-over
-126 s^ + 210
4th distribution + 54 + 72 -120 -90

2 -7113 + 7113 -4 705 +4 705

Solution (b) Shortcut procedure

Figure 8-23.
354 8 Continuous Beams

A and C is equivalent to adding moments of —2000 and +8000,

respectively, so as to cause a final moment of zero at these free ends.
When this distribution is completed and all the supports are relocked,
the distributed moments cause by the
the carry-over effects indicated
arrows,and this again introduces unbalance at the locked supports.
However, this unbalance is appreciably smaller.
We continue to unlock each support, distribute the unbalanced
moment, and relock each support, thus completing another cycle of
distribution, until the carry-overmoments become negligibly small or,
as here in cycle 3, until the sum of the carry-over moment and the
distributed moment is zero on each side of a support. Further cycles of
distribution will then produce no effect, as cycle 4 shows. Note that the
analysis must end with a distribution and not with a carry-over.

875. Apply the moment-distribution method to the continuous

beam of three spans with free ends shown in Fig. 8-23.

Solution: There are two solutions. The first, in (a), involves the same
procedure that was described in Illustrative Problem 874 for free ends
on two spans. inconvenient to treat a free end as fixed, carry
It is

moment over toand then release it again. It is simpler to use a

it,

modification in which the free end, initially assumed fixed, is released

only once and has no moment carried over to it for further distribution.
To understand this modification, we shall show how the moment that is
distributed to the left of B in the first distribution of solution (a) is

carried through the computations. This moment, denoted by M in Fig.

8-24, carries over to A as +\M if A is locked. If A is a free end,
releasing it causes the distributed moment —\M at A, which then
carries over to B the moment \{— \M) = — \M. A summation of these
values gives zero at A (which is now freely hinged) and \M at B.
If the initial distribution at B had been modified so that only \M
were distributed to B, these results would have been obtained directly,

n
with no moment carried over to A In other words, if the stiffness factor
.

Distributed
moment to left
of B only
Carry-over to A
Distribution at A
Carry-over to B

Figure 8-24. Modification of distribution at B to avoid carry-over to free end A.

8-8 Moment Distribution 355

for AB is multiplied by |, the distribution at B is modified so that no

moment need be carried over to the free end.
This shortcut is applied in the second solution (b) shown in Fig.
8-23. As a start, reduce A and D
ends by releasing them, to free
applying the balancing moments of - 8000 at A and + 7200 at Z>, and
carrying over half these amounts with the same signs to B and C as

4000 N

K L f
X2 = @ © ©
2 3
DF 1 3 5 I

FEM -6000 +8000 -8000 +3 000 -3 000 +4 800 -7 200

Release A -2000 -lOOO

Adjusted FEM -6000 +6000 -9 000 + 3 000 -3 000 +4 800 -7 200

1st distribution
+ 2000 +4000 ^ -1080 -720 .
N
Carry-over -540 '^ +2 000 ^ -360
2nd distribution + 180 +360 ^^ +1 200 -800 v
Carry-over
-600 *^ +180 -400
3rd distribution + 200 +400 ^^ -108 -72 v
Carry-over
-54 ^^ +200 X -36
4th distribution -80
+ 18 + 36 -120

Z -6000 +6000 -6 602 +6 602 -3128 +3128 -7 996

Solution (a) Regu ar procedure

2 3
DF i 3 I o
5

FEM -6000 + 8000 -8 000 +3 000 -3000 +4 800 -7200

1st distribution
-2000^ +1667 +3 333 ^^-1080 -720
^
,.

Carry-over
^-1000 -540 ^^ +1667 -360
2nd distribution +513 + 1027 ^^-1000 -667.

Carry-over
-500 ^^ +514 ^ -334
3rd distribution + 167 + 333
^^ -308 -206^
Carry-over -154 ^ +167 ^ -103
4th distribution +51 +103 -100 -67

2 -6 000 +6000 -6602 +6602 -3140 + 3140 -7 997

Solution (b) Alternate procedure

Figure 8-25.
 8 Continuous Beams

shown. Now distribute the FEMs at B and C, using the distribution

factors obtained by modifying the stiffness of AB and CD as described
above. Note that no moment is carried over to the free ends under these
conditions. Note also that we are approaching the final result more
rapidly in the fourth distribution here than in the corresponding fourth
distribution in the first solution.

876. The loading in Illustrative Problem 875 is modified in Fig.

8-25 by adding an overhang at A and fixing the end at D. The moments
of inertia of segments AB and CD are each equal to 20 units, but that of
BC is increased to 30 units. The relative stiffnesses are therefore
KAB = 2, KBC - 3, and KCD = 2. Compute the moments over the sup-
ports.

Solution: This problem illustrates two additional concepts: (1) The

cross section may vary from segment to segment and is taken into
account by computing the I / L ratio for each segment.* (2) The over-
hanging end offers no resistance to rotation. Hence, when A is released,
the unbalance of 2000 N m • must be distributed as zero to the left of A
and as -2000 N m•
to the right of A. When - 1000 N m
•
is carried
over to B, the adjusted fixed end moments are as shown below the first
double line. Also, to avoid carrying moments back to A, we multiply
KAB by \, and the shortcut procedure described in Illustrative Problem
875 is then applied, as shown in Fig. 8-25. Note that the fixed end at D
absorbs the carry-over moments from C but does not transfer any back
to C.
An by some engineers gives sub-
alternate distribution preferred
stantially the same results and is also shown in Fig. 8-25. Here, instead
of starting with an adjusted FEM at A and 2?, the unbalanced moment
at all supports is first distributed at each support and then carried over

as shown.

PROBLEMS

By means of the moment-distribution method, solve for the mo-

ments over the supports in the continuous beams referred to below.

877. See Problem 814 (page 325).

878. See Problem 826.

879. See Problem 827.

880. See Problem 845 (page 338).

881. See Problem 846.

•More complex cases, in which the cross section varies along the segment,
are treated by H. Cross and N. D. Morgan, Continuous Frames of Reinforced
Concrete, Wiley, New York, 1932.
8-8 Moment Distribution 357

882. See Problem 849.

883. See Problem 853.
884. See Problem 856.
885. Solve Problem 856 if the moment of inertia varies from span
tospan so that the relative stiffness for span 1 is 2, that for span 2 is 1.5,
and that for span 3 is 1.
886. Solve for the support moments in Problem 825 (page 327) if

the ends are perfectly fixed instead of simply supported.

Arts. M x
- - 73 Nm; M 2
- - 2100 Nm; M 3
- - 2380 Nm;
M 4
= - 1310Nm

SUMMARY

The general form of the three-moment equation is

MLX X
+ 2M2 (L X
+ L2 ) + ML 3 2 + ^1 + ^^
L x
L 2

The factors in this equation are tabulated for various span loadings in
Table 8-1 on page 318.
For continuous beams whose supports are on the same level, the
terms h and h 3 reduce to zero and the three-moment equation readily
x

determines the moments at the supports. If one end of a continuous

beam is fixed, it may be treated as though it were an imaginary span.
When determining deflections, the three-moment equation is writ-
ten between three points, 1, 2, and 3, so that either (or both) of the
heights h and h 3 is equal to the desired deflection. Generally, two of the
x

points are at the supports, whereas the third point is at the location of
desired deflection. The values of the moments at these three points must
first be known or computed.
An alternate method of determining support moments in continu-
ous beams (as well as in more general frames) is moment distribution.
To apply it, each span is first considered to be fixed at the supports, and
the fixed end moments are computed from the general values listed in
Table 7-2 on page 307. Each support is then released, and the resulting
unbalanced moments are distributed at each support and also carried
over to adjacent supports in accordance with the details given in Art.
8-8. As explained in Art. 8-5, once the support moments have been
found, it is a simple matter to determine the shear diagram, from which,
as explained in Chapter 4, the maximum shear and bending moment
can be found.
 Combined

9-1 INTRODUCTION

In preceding chapters we studied three basic types of loading: axial,

torsional, and flexural. Each of these types was discussed on the
assumption that only one of these loadings was acting on a structure at
a time. The present chapter is concerned with cases in which two or
more of these loadings act simultaneously upon a structure. The three
basic types of loading and the corresponding stress formula may be
summarized as follows:

p
Axial loading aa = —
T
Torsional loading t = ~-
j

Flexural loading of = —j-

There are four possible combinations of these loadings: (1) axial

and flexural; (2) axial and torsional; (3) torsional and flexural; and (4)
axial, torsional, and flexural, acting simultaneously. We shall consider

358
9-2 Combined Axial and Flexural Loads 359

the axial and combination first because it combines only

flexural
normal and is therefore the simplest. All the others combine
stresses
shearing and normal stresses and require a preliminary discussion (see
Arts. 9-4 to 9-7) before they can be considered.

9-2 COMBINED AXIAL AND FLEXURAL LOADS

The simply supported beam in Fig. 9- la carries a concentrated load Q.

The supports are hinged to the beam at its centroidal surface. At point
A, the flexural stress af = My /I. It is tensile and is directed normal to
the surface of the cross section, as shown. The force exerted on the
element at A is af dA .

If the same beam supported in the same way is loaded only with

an axial load P (Fig. 9- lb), the axial stresses are uniformly distributed
across any transverse section (Art. 1-3). Their magnitude is aa = P/A;
they are tensile and directed normal to the cross section. The force
exerted on the element at aa dA A is .

If both loads act simultaneously on the beam (Fig. 9-lc), the

resultant stress at A is equal to the superposition of the two separate
effects. Thus, the resultant force at A is the vector sum of the collinear
forces aa dA and of dA Dividing . this by the area dA gives the resultant
stress a = a o f
+ a directed normal to the cross section.
B in the same section, also at a distance v
Similarly, at a point
from the neutral axis but above it, the resultant stress is the difference
between the axial and flexural stresses. If tensile stress is denoted by a
positive sign and compressive stress by a negative sign, the resultant
stress at any point of the beam is given by the algebraic sum of the axial
and flexural stresses at that point:
<*>>\
L^
V ^~\

or
,^ L

Note that the axial stress may be compressive; this is the reason
for the circled and © signs before P/A The circling of these signs is
.

a reminder that the axial stress is uniform and of the same type all over
a cross section whereas the magnitude and type of the flexural stress
vary with position.
In Eq. (9-1) we used the method of superposition. One note of
caution is necessary, as the following makes clear: Fig. 9-2 shows, in
exaggerated form, the flexing effect of Q. If P is tensile, as in Fig. 9-2a,
the bending moment of P at any section, i.e., P8, tends to reduce the
bending moment due to Q and hence slightly reduces the flexural stress.
360 9 Combined Stresses

KT> ^ tfr

*j
(a) Flexure stress <M
Section mn °/'

p —*-
P
77T7TT

A' >
n
(b) Axial stress a„</A
Section m-n

>' ^^c°^A
y, J a.rfA
OfdA
«,
v Section m-n
o n dA
(c) Axial and flexure stress combined
Note shift in position of line of zero stress
Figure 9-1.

S=^
Rt R2
(a) Axial tension (b) Axial compression
Figure 9-2.
9-2 Combined Axial and Flexurai Loads 361

The opposite effect occurs when the axial load is compressive, as in Fig.
9-2b, where the additional bending moment PS slightly increases the
flexurai stress. In other words, the values given by Eq. (9-1) are slightly
high when P is tensile and low when P is compressive. These
slightly
effects are negligible in the case of most structural members, which are
usually so stiff that stresses produced by bending moments like P8 can
be neglected. But in long slender members or columns, the effect is
significant and more exact methods must be used. (Columns are dis-
cussed in Chapter 11.)

ILLUSTRATIVE PROBLEM

901. A cantilever beam (Fig. 9-3) has the profile shown so that it

will provide sufficient clearances for large pulleys mounted on the line
shaft it supports. The reaction of the line shaft is a load P= 25 kN.
Determine the resultant normal stresses at A and B at the wall.

Solution: We begin by determining the bending moment of P. This is

computed in terms of its components Px = 20 kN and Py = 15 kN by

taking moments about the centroidal axis of section AB:

M- - (15 X ltf)(0A50) + (20 X lO^O.^O) 3750 N-m

Since Py
acts down, its moment effect is negative (Art. 4-2) and
the opposite moment effect of Px must be positive. The negative sign of
the bending moment at AB indicates that the beam curvature at section
A -B is concave downward (Art. 4-2), thereby causing tension at A and
compression at B. Having thus interpreted the sign of the bending
moment, we use only its absolute value in applying Eq. (9-1).
However, it may not as yet be obvious that the axial tensile effect
is due solely to P . Use the principles of mechanics to convert the given

150 mm - r

20 kN

Py =15 kN P=25kN
Figure 9-3.
362 9 Combined Stresses

load into either of the equivalent loadings in Fig. 9-4. It is evident from
the principle of transmissibility that the entire is due to moment effect
P 9-4b and therefore the axial effect is caused by Px alone. Or,
in Fig.
we may add a pair of collinear forces each equal to Px as in Fig. 9-4c, ,

thereby reducing the system to that shown in Fig. 9-4d. Once again we
see that the axial effect is caused by Pxf since the bending moment,
which consists of —0A50Py plus the couple 0.1 50/^, is equivalent to the
bending moment as computed above.
We are now ready to compute the resultant stresses by applying
Eq. (9-1). At A we obtain

P I Mc 6A/V

3
20 x 10 6(3750)
Oj =
2
(0.050)(0.150) (0.050)(0.150)
6
- (2.67 x 10
6
) + (20.00 x 10 ) = 22.67 MPa Ans.

At B, where the flexural stress is compressive, we have

= P t Mc = 6M\]
a-(— -^)\
20 x 10
3
6(3750)
oD =
2
(0.50)(0.150) (0.050)(0.150)
= (2.67 X 10
6
)
- (20.00 X 10
6
) 17.33 MPa Ans.

The signs indicate that the stress is tensile at A and compressive at B.

M= (0.150 Px )N-m

/
150 mm

450 mm
(c)

Figure 9-4.
-450

(d)
mm-
\
9-2 Combined Axial and Flexural Loads 363

PROBLEMS

902. Compare the maximum stress in a bent rod 10 mm square,

where the load P is 10 mm off center as shown in Fig. P-902, with the
maximum stress were the rod straight and the load applied axially.
{Note: This problem illustrates why lateral deflection in columns is so
dangerous.) Am. 7 to 1

10 mm
Figure P-902.

903. A cast iron link is 40 mm wide by 200 mm high by 500 mm

long. The allowable stresses are 40 MN/m2 in tension and 80 MN/m2
in compression. Compute the largest compressive load that can be
applied to the ends of the link along a longitudinal axis that is located
150 mm above the bottom of the link.
$04s To avoid interference, a link in a certain machine is de-
signed so that its cross-sectional area is reduced one-half at section
A-B, as shown in Fig. P-904. Compute the maximum tensile stress
developed across section A-B if (a) the cross section of the link is 160
mm square and (b) the cross section is a circle 160 in diameter. mm
Ans. (a) 75.0 MPa; (b) 85.3 MPa

P = 240kN

Figure P-904.

305. A wooden beam 100 mm by 200 mm, supported as shown in

P-905, carries a load P. What
Fig. is the largest safe value of P if the
maximum stress is not to exceed 10 MPa?

^S^| 100 mm

200 mm
TTnrr / / / //

2m -lm

Figure P-905.
364 9 Combined Stresses

906. The bent steel bar shown in Fig. P-906 is 200 mm square.
Determine the normal stresses at A and B.
Arts. aA = - 29.2 MPa; oB = 4.2 MPa

500 kN

:
250 mm

Figure P-906.

907. Determine the largest load P that can be supported by the

platform of the cast-iron bracket shown in Fig. P-907 if a, < 30

MN/m 2
and oc < 70 MN/m2 . Ans. P= 32.9 kN

< < /./ f < /

50 mm-*- L— 150-^
I mm
0i ZZZZZZZ3
NA
Section A-B
Area = 8000 mm 2

7 NA =20Xl0 6 mm 4

Figure P-907.
9-2 Combined Axial and Flexural Loads 365

908. A punch press has the cast-steel frame shown in Fig. P-908.
Determine the largest force P that can be exerted at the jaws of the
punch without exceeding a normal stress of 120 MPa at section A-B.
The properties of the area are as shown where 1-1 is the centroidal axis.

ZZ2ZZZZZ

2///Y////fc

200 -300-*]
mm mm
Section A-B
;////////////////////;//////;////;///' /i.!=1600 X 10
6
mm4
Area= 80 X 10 3 mm 2

Figure P-908.

909) A rectangular beam, 100

J

mm
wide by 400 deep, is mm
pinned at A, supported by a cable CD, and carries a load P, as shown in
Fig. P-909. Determine the maximum value of P that will not exceed a
normal stress of 120 MPa. Neglect the possibility of buckling.
Arts. P = 457kN

Figure P-909.

(^$5b The inclined beam in Fig. P-910 is supported by a pin at A

and a roller at C. The cross section is 100 mm by 300 mm. Determine
the maximum compressive stress developed in the beam.
Arts. o\ = 70.7 MPa
366 9 Combined Stresses

4 m

h 2 m

Figure P-910.

911. P -
If 100 kN for the bracket shown in Fig. P-911, com-
pute the maximum tensile and compressive stresses developed at section
A-B.

1 = 50 X 1Q6 mm4
NA
Area = 8000 mm 2

200 mm /,

Figures P-911 and P-912.

Determine the maximum safe load P that can be applied to

912.
the bracket in Fig. P-911 if the allowable stresses at section A-B are
o, < 8 MPa and oc < 12 MPa. Ans. P = 12.1 kN
913. Compute the stresses at A and B for the link loaded as
shown in Fig. P-913. Ans. aA= - 14.1 MPa; aB = 5.10 MPa
9-2 Combined Axial and Flexural Loads 367

20 kN
\ 200 mm * 400 mm

200 mm 200 mm 100 mm

mm
£ m
//////
200

j
100 mm 80 mm
Cross-
section

100 kN
Figure P-913.

914. A timber 100 beam AD,

thick by 300 mm high and mm
loaded as shown in Fig. P-914, pinned at its lower end and supported
is

by a horizontal cable CE. Determine the maximum compressive stress

in the beam. Arts, a = 14.3 MPa

10 kN

Figure P-914.

915. A dam has the profile shown in Fig. P-915. If the

concrete
density of concrete 2400 kg/m3 and that of water is 1000 kg/m3
is ,

determine the maximum compressive stress on section m-n if the depth

of the water behind the dam is h = 15 m.
368 9 Combined Stresses

i -*-3 m
•
i /• -

/ e> '.

----- J i

25 m /: .p
'':'.

p'0+ A
p
J.

1°. - : '° •"•*•

\
'
mr '. '•'*. '
-'.V; n !

v//. vy.
f-*-9
m -^
Figure P-915.

916. For the pin-connected frame shown in Fig. P-916, de-

termine the maximum normal stress in member BD if its cross section is
100 mm wide by 400 mm
deep. Neglect the weights of the members.
Ans. a = 25.4 MPa

4 m

Figure P-916.

ST
917. The shown in Fig. P-917 is hinged to fixed sup-
structure
ports at A and E. Compute the maximum compressive stress developed
in bar BDE if its cross section is 200 mm square. Neglect the weights of
the members.
9-3 Kern of a Section; Loads Applied off Axes of Symmetry 369

20 kN

Figure P-917.

9-3 KERN OF A SECTION; LOADS APPLIED

OFF AXES OF SYMMETRY

A special case of combined axial and flexural loads is illustrated in Fig.

9-5a, in which a short strut* carries a compressive load P applied with
an eccentricity e along one of the principal axes 1 of the "
section. The
addition of a pair of forces P and P2 each of magnitude
l
, and acting P
at the centroid of the section, causes the equivalent loading shown in
Fig. 9-5b. The stresses across any typical section m-n are the result of
the superposition of the direct compressive stress (oa = P/A) in Fig.
9-5c and the flexural stress (oy - Mc/I = (Pe)c/I) in Fig. 9-5d. If the
maximum flexural stress is larger than the direct compressive stress,
the resultant stress appears as in Fig. 9-5e. The point of zero stress TV is

the new location of the neutral axis and is easily found by computing
the distance a at which the tensile flexural stress equals the direct
compressive stress:

P = My = (Pe)a
A I I
whence

a = (9-2)
Ae
strut is one whose length is no more than ten times its smallest
*A short
lateraldimension; the flexural deflection is so small that its effect can be neglected.
The eccentric loading of long bars is discussed in Art. 1 1 -6.
+ The
principal axes are the axes of maximum and minimum moments of
inertia.
370 9 Combined Stresses

P« p P.
^-e—J \
' > M=Pe
4f

4 i

1
(a) (b)

m m
n
(c)
ehh]
-.-i

« jp
^ntv*
(

V7s* v////////\ '/////////, V/A

°f+°c
(e)

Figure 9-5.

It is evident that there will be no tensile stress anywhere over the

section if the direct compressive stress equals or exceeds the maximum
flexural stress. Thus, for a rectangular section of dimensions b and h,
with P applied at an eccentricity e (Fig. 9-6), we obtain

Pe
P = Mc (!)
A I

12

The maximum eccentricity to avoid tension is thus

h
(9-3)

This formula is the basis of the well-known rule that, in designing

masonry or other structures weak in tension, the resultant load should

fall in the middle third of the section.

Figure 9-6.
9-3 Kern of a Section; Loads Applied off Axes of Symmetry 371

Figure 9-7.

We now consider the general case* in which the load P is applied

at any point with respect to the principal axes X and Y, as in Fig. 9-7.
If ex and e represent the eccentricities of P, the moments of P with
y
respect to the X and Y axes are respectively Pey and Pex By superposi- .

tion, the stress at any point of the cross section whose coordinates are x
and>> is

P— (Pex )x (Pey )y
a= r (9-4)

To determine the neutral axis or line of zero stress in the cross section,
we set a — 0. Using Iy = Ary and Ix = Arx , where r
y
and rx are
respectively the radii of gyration* relative to the Y and X axes, we
obtain

0= 1 («)
r rx
y

The intercepts u and v of the neutral axis with the X and Y axes
respectively are found by substituting first y = and then x = in Eq.
(a). This gives
2 2
«=- 2
r
and d=-- (b)

The neutral axis passes through the quadrant which is opposite to

that containing P, and in general it is not perpendicular to the direction
OP. For example, the stress distribution on a rectangular section caused
by a load P not on the principal axes (Fig. 9-8a) is shown in Fig. 9-8b.
If the stresses at A, B, and C are computed, the intersection of the

•Actually, this is an application of unsymmetrical bending, which is dis-

cussed at length in Art. 13-9.
^Here we are using r to denote radius of gyration to conform to AISC
notation. Be careful not to confuse this r with the r that is frequently used to denote
the radius of a circle.
372 9 Combined Stresses

D
^ n\
*
J/7?b
fcs.
4 x
n
B c
h
6
-
(c) (d)

Figure 9-8. Neutral axis for load P eccentrically applied and kern of rectangular
section.

neutral axis with AB and BC (or their extensions) can be easily

determined by proportion.
Let us now determine the coordinates ex and ey of the load P, for
which the neutral axis passes through the corner 2?, as in Fig. 9- 8c.
Substituting o = 0, x = - /j/2, and>> = — b/2 in Eq. (9-4), we obtain

0= - <MS) (
M!)
bh

12 12

or

V6+J76- 1 W
This is the equation of the straight line /wai in Fig. 9-8d; it intersects the
X and y axes at h/6 and ^?/6, respectively. This line is the locus of
points of application of P y
for which corner B has zero stress. Any
compressive load above and to the right of this line causes tension at B.
Similarly the linem n is the locus of loads that cause zero stress in
l l

corner C. Continuing this procedure indicates that evidently no corner

and no part of the cross section will be in tension if the resultant
9-4 Variation of Stress with Inclination of Element 373

compressive load lies on or within the diamond-shaped figure. This

shaded area is known as the kern of the cross section.
Show that the kern of a circular section is a circle whose diameter
is one-quarter the diameter of the section.

PROBLEMS

918. A compressive load P— 80 kN is applied, as shown in Fig.

9- 8a, at a point mm to the right and 60 mm above the centroid of a
40
rectangular section for which h = 400 mm and b = 200 mm. Compute
the stress at each comer and the location of the neutral axis. Illustrate
your answers with a sketch similar to Fig. 9 -8b.
919. From the data inProblem 918, what additional load applied
at the centroid is necessary so that no tensile stress exists anywhere on
the cross section? Ans. 112 kN
920. A compressive load P— 100 kN is applied, as shown in Fig.
9-8a, at a point 70 mm to the left and 30 mm above the centroid of a
rectangular section for which h = 300 mm and b — 150 mm. What
additional load, acting normal to the cross section at its centroid, will
eliminate tensile stress anywhere over the cross section? Ans. 160 kN
921. Calculate and sketch the kern of a W360 X 122 section.
Ans.A diamond-shaped figure having coordinates on the
X axis of ± 30.9 mm and on the Y axis of ±130 mm
922. Calculate and sketch the kern of a W3 10 X 500 section.

9-4 VARIATION OF STRESS WITH *

INCLINATION OF ELEMENT

In Art. 1-2 we saw that the magnitude and type of stress depend on the
inclination of an element. As a review of that discussion, consider that
the body in Fig. 9-9a is acted upon by the given forces which are in

Figure 9-9. Stress at a point varies with inclination of plane through point.
374 9 Combined Stresses

equilibrium. Pass two exploratory sections a-a and b-b through the
body, section a-a perpendicular to the resultant R of P and P2 x

as in Fig. 9-9b, and section b-b inclined to R as in Fig. 9-9c. The

element in 9-9b is subjected only to a normal stress, but
Fig.
9-9c is subjected to both normal and shearing stress
the element in Fig.
caused by the N and T components of the equilibrant E. Thus at the
same position in a stressed body (located here at the intersection of
sections a-a and b-b), the stresses on an element vary with the
orientation of the element.
In the following articles we discuss the manner in which these
stresses vary with the orientation of the element. Our purpose will be to
determine the orientation of the element on which the maximum normal
stress exists, and magnitude; also the orientation of the element on
its

which maximum shearing stress exists, and its magnitude.

In general, it is not possible to compute directly the stresses that
exist on any arbitrarily chosen surface. In beams, for example, the
flexure formula determines stresses only on planes normal to the longi-
tudinal axis of the beam. So also for torsional shearing stresses: the
torsion formula determines shearing stresses only on sections normal to
the longitudinal axis of a twisted bar. Thus, in a bar subjected to
simultaneous bending and twisting as in Fig. 9-10, we can compute the
flexural and shearing stresses only for elements in the position shown.
But the discussion of Fig. 9-9 indicates that if the element in Fig. 9-10
were rotated about the axis shown, there would be a particular position
at which maximum normal stress would exist.
There are two methods of determining this position of the element
and computing the maximum stresses to which it is subjected: one is
analytical; the other is graphical, based on Mohr's circle. The analytical

Top view
of element

Figure 9-10. Stresses caused by simultaneous flexure and torsion.

9-5 Stress at a Point 375

9-6 is given primarily to demonstrate the construction

discussion in Art.
and Mohr's circle, which is discussed at length in Art. 9-7.
validity of
(The formulas in Art. 9-6 should not be memorized.)

9-5 STRESS AT A POINT <

The average stress over an area is obtained by dividing the force by the
area over which it acts. If the average stress is constant over the area,
the stress is said to be uniform. If the stress is not uniform, the stress at
any point is found by permitting the area enclosing the point to
approach zero as a limit. In other words, stress at a point really defines
the uniform stress distributed over a differential area. In Fig. 9-11, for
example, the normal stress in an X direction acting at the point whose
coordinates are x, y and z means the uniform stress acting over the
differential area ay dz.
When the stress at a point is defined by components acting in
several directions, the stresses may be represented as acting on the
differential element volume) enclosing the point. For example, let
(i.e.,

the stresses at a point be ox o and t', ,Fig. 9- 12a shows these stresses
;
y
as they act on the differential element enclosing the point. The element
is usually represented by its front view, as in Fig. 9- 12b. Note that there

dy

/9
Figure 9-11. Stress at a point.

(b)

Figure 9-12. Stress components.

376 9 Combined Stresses

is a shearing stress ryx acting on the Y face in the direction. This is X

due to the fact that a shearing stress on any plane induces an equal
shearing stress on a plane perpendicular to the first one. (See Art. 5-7,
page 188.)
The notation used here defines a normal stress by means of a
single subscript corresponding to the face on which it acts. A face takes
the name of the axis normal to it, for example, the X face is perpendicu-
lar to the X axis. A shearing stress is denoted by a double subscript, the
first letter corresponding to the face on which the shearing stress acts
and the second indicating the direction in which it acts. Thus the
shearing stress on the X face acting in the Y direction is denoted by rxy9
and the shearing stress on the Y face acting in the X direction is

denoted by ryx Of course . rxy = r since the shearing

VJC
, stresses on
perpendicular planes are equal.
In this and succeeding articles, we consider only plane stress in
which the stresses act parallel to a single plane such as the XY plane. In
a triaxial stress, the Z
an element may be subject to the normal
face of
stress o2 as well as to shearing stresses rzx and r^. These shearing
stresses then induce the numerically equal shearing stresses rxz and ryz
which act, respectively, on the X and Y faces.

9-6 VARIATION OF STRESS AT A POINT:

ANALYTICAL DERIVATION

The stress acting at a point is represented by the stresses acting on the

faces of the element enclosing the point. As we saw in Art. 9-4, the
stresseschange with the inclination of the planes passing through that
point; that is, the stresses on the faces of the element vary as the angular

position of the element changes.

In determining these stress variations analytically, a plane is
passed that cuts the original element into two parts and the conditions
of equilibrium are applied to either part. Figure 9- 13b shows the
normal and shearing stress components acting on the plane whose
normal TV makes an angle with the X axis (see Fig. 9- 13a). The
triangular element in Fig. 9- 13b is in equilibrium under the action of
the forces arising from the stresses that act over its faces. The area of
the inclined face being denoted by A, these forces are shown in the
free-body diagram in Fig. 9- 13c. The point diagram of these forces is
shown in Fig. 9- 13d.
Applying the conditions of equilibrium to axes chosen as in Fig.
9- 13d, we obtain
[IN = 0] Ao = (ox A cos 9) cos 9 + (ay A sin 9) sin 9
— {rxy A cos 9) sin 9 — (ryx A sin 9) cos 9 (a)
9-6 Variation of Stress at a Point: Analytical Derivation 377

Y\

(a) Original state of stress (b) Stresses acting on wedge

Area of wedge face

taken as A
Txy A cos Q
A X =A cos 6,

T yx A sin $

A sin Q
Area (A y )=A sin $

(c) Free-body diagram of forces on wedge (d) Point diagram of forces

Figure 9-13. Variation of stress components.

and
[2T = 0] At = (ax A cos 9) sin 9 - (ay A sin 9) cos 9

+ {rxy A cos 9) cos - (ryx A sin 0) sin 9 (b)

Since the common term A can be canceled and since t„_

yx is

numerically equal to txj,, we use the relations

2 . 1 + cos 20 2 „ 1 - cos 29
snr 9 =
.
z
cos = ,
-

sin 9 cos 9 = — sin 20

to reduce Eqs. (a) and (6) to

<r, + cr, a, — a,
a= ~—^cos 20 - txj sin 20
, (9-5)

and
<T — <T
*
t = * sin 20 + rxy cos 20 (9-6)
378 9 Combined Stresses

The planes defining maximum or minimum normal stresses are

found by differentiating Eq. (9-5) with respect to 9 and setting the

derivative equal to zero, whence

tan 20 = - -^- (9-7)

Similarly, the planes of maximum shearing stress are defined by

tan 20, = ^V^ (9-8)

Equation (9-7) gives two values of 29 that differ by 180°. Hence, the
planes on which maximum and minimum normal stresses occur are 90°
apart. Similarly, from Eq. (9-8) the planes on which the maximum
shearing stress occurs are also found to be 90° apart.
The planes of zero shearing stress may be determined by setting t
equal to zero in Eq. (9-6); this gives

tan 29 =
2txv
~
—
°x °y

which is identical with Eq. (9-7). Hence maximum and minimum normal
stresses occur on planes of zero shearing stress. The maximum and
minimum normal stresses are called the principal stresses, sometimes
referred to as the p and q stresses.
Equation (9-8) is also the negative reciprocal of Eq. (9-7). This
means that the values of 29 defined by Eqs. (9-7) and (9-8) differ by
90°. In other words, the planes of maximum shearing stress are at 45°
with the planes of principal stress.
Substituting values of 29 from Eqs. (9-7) and (9-8) respectively in
Eqs. (9-5) and (9-6), we obtain the following expressions for maximum
stresses:

wx - -V" * V(^"^) +
^ (9_9)

If a similar analysis is made for a plane whose normal is per-

pendicular to N, the stress components on this perpendicular plane will

be
°x + <*y ox - ay
2
cos 29 + r*y
xy sin 29 (9-5<z)

y
T = sin 29 -I-
'
t'*y cos 29 {9-6a)
2

Adding Eqs. (9-5) and (9-5ar) shows that the sum of the normal stresses
9-7 Mohr's Circle 379

on any two perpendicular planes is a constant equal to ax + ay Also, .

comparison of Eqs. (9-6) and (9-6a) confirms the equivalence of

shearing stress on perpendicular planes.

9-7 MOHR'S CIRCLE

The formulas developed in the preceding article may be used for any
case of two-dimensional stress. A visual interpretation of them, devised
by the German engineer Otto Mohr in 1882, eliminates the necessity for
remembering them.* In this interpretation a circle is used; accordingly,
the construction is called Mohr's circle. If this construction is plotted to
scale, the results can be obtained graphically; usually, however, only a
rough sketch is drawn, analytical results being obtained from it by
following the rules given later.
We can easily show that Eqs. (9-5) and (9-6) define a circle by
first rewriting them as follows:

a - y
= y
cos 29 - rxy sin 29 (a)
2
x
t - -"
sin 29 + rxy cos 29 (b)

By squaring both these equations, adding the results, and simplifying,

we obtain

Recall that ox > o

y , and t^ are known constants defining the specified
state of stress, whereas a and t are variables. Consequently (ox + oy )/2
is a constant, say C, and the right-hand member of Eq. (c) is another
constant, say R. Using these substitutions, we transform Eq. (c) into

(a - Cf + r
2
= R2 (d)

Equations (9-5) and (9-6), as well as the succeeding variations of them, are
identical to the equations that express the variations in moments of inertia with
respect to U and V axes inclined at an angle to the reference axes X and Y.
Replacing normal stress by the moment of inertia I and the shearing stress by the
product of inertia P, we obtain
Ix +Iy Ix -L
Iu - JL
-
y^ + 2
cos 20 - Pxy sin 20

and

Puv - —^-^sin 20 + Pxy cos 20

A Mohr's circle treatment of these equations is described in detail in Appendix A, p.

627.
380 9 Combined Stresses

yf">'

-f H
T v.v

Figure 9-14. Mohr's circle for general state of stress.

This, being of the form (x — C) 2 + y 2 = R 2 , is readily recognized as a

circle of radius

whose center
'-Vm is
+ (rxy f
offset rightward a distance

from the origin.

Figure 9-14 represents Mohr's circle for the state of stress which
was analyzed in the preceding article. The center C is the average of the
normal stresses, and the radius

»-#^ + (O xy<

is the hypotenuse of the right triangle CDA. How do the coordinates of

points £, Ft
and G compare with the expressions derived in Eqs. (9-9)
and (9-10)? We Mohr's circle is a graphic visualization of
shall see that
by Eqs. (9-5) and (9-6). The following rules
the stress variation given
summarize the construction of Mohr's circle.

Rules for applying Mohr's circle to combined stresses

1. On rectangular o—r axes, plot points having the coordinates

(ox , rxy ) and (oy ryx ). These points represent the normal and shearing
,

stresses acting on the X and Y faces of an element for which the stresses
 —
9-7 Mohr's Circle 381

are known. In plotting these points, assume tension as plus, comp ressio n
as minus, and shearing stress as plus when its moment about the center
of the element is clockwise.* «*
2. Join the points just plotted by a straight line. This line is the
diameter of a circle whose center is on the a axis.

3. As different planes are passed through the selected point in a

stressed body, the normal and shearing stress components on these
planes are represented by the coordinates of points whose position shifts
around the circumference of Mohr's circle.

4. The radius of the circle to any point on its circumference

represents the axis directed normal to the plane whose stress compo-
nents are given by the coordinates of that point.

5. The angle between the radii to selected points on Mohr's circle

is twice the angle between the normals to the actual planes represented
by these points, or to twice the space angularity between the planes so
represented. The rotational sense of this angle corresponds to the
rotational sense of the actual angle between the normals to the planes;
i.e., if the N axis is actually at a counterclockwise angle from the X
axis, then on Mohr's circle the TV radius- is laid off at a counterclockwise
angle 20 from the X radius.

ILLUSTRATIVE PROBLEMS

923. At a certain point in a stressed body, the principal stresses

are ax — 80 and ay = — 40 MPa. Determine a and t on the planes
MPa
whose normals are at +30° and + 120° with the X axis. Show your
results on a sketch of a differential element.

Solution: The given shown in Fig. 9- 15a. Following

state of stress is
the rules given above, draw a set of rectangular axes and label them a
and t as shown in Fig. 9- 15b. (Note that, for convenience, the stresses
are plotted in units of MPa.) Since the normal stress component on the
X face is 80 MPa and the shear stress on that face is zero, these
components are represented by point A which has the coordinates
(80, 0). Similarly, the stress components on the Y face are represented
by point B (-40, 0).
According to rule 2, the diameter of Mohr's circle is AB. Its center
C, lying midway between A and B, is 20 MPa from the origin O. The

*This special rule of sign for shearing stress makes r^ = — r^ in Mohr's

circle. From here on, we shall use this rule to designate positive shearing stress.
However, the mathematical theory of elasticity uses the convention that shearing
stress is positive when directed in the positive coordinate direction on a positive
face of an element, i.e., when acting upward on the right face or rightward on the
upper face. This other rule makes r^ - r^, which is convenient for mathematical
work but confusing when applied to Mohr's circle.
382 9 Combined Stresses

r(MPa)-*

ov =-40MPa

4— a(MPa)
o =80MPa

(a)

(b)

Figure 9-15.

radius of the circle is the distance CA — 80 — 20 = 60 MPa. From rule

4, the radius CA represents the X axis. In accordance with rules 4 and 5,

point D components on the face whose normal is

represents the stress
inclined at +30° to the X axis, and point E represents the stress
components on the perpendicular face. Observe that positive angles on
the circle are plotted in a counterclockwise direction from the X axis
and are double the angles between actual planes.
From rule 3, the coordinates of point D represent the required
stress components on the 30° face. From the geometry of Mohr's circle,
these values are

a = OF = OC + CF = 20 + 60 cos 60° - 50 MPa

T = DF = 60 sin 60° = 52.0 MPa

Figure 9-16.
9-7 Mohr's Circle 383

On the perpendicular 120° face we have

a' . OG = OC - CG = 20 - 60 cos 60° = - 10 MPa
T ' . GE - -60 sin 60° = -52.0 MPa
Both sets of the above stress components are shown on the differential
element in Fig. 9-16. Observe the clockwise and counterclockwise
moments of t and t', respectively, relative to the center of the element
(see rule 1). Finally, note that a complete sketch of a differential
element shows the stress components acting on all four faces of the
element and the angle at which the element is inclined.

924. A 9- 17a. Determine the

state of stress is specified in Fig.
normal and shearing on (a) the principal planes, (b) the planes
stresses
of maximum shearing stress, and (c) the planes whose normals are at
+ 36.8° and + 126.8° with the X axis. Show the results of parts a and
b on complete sketches of differential elements.

Solution: Mohr's circle for the given state of stress is shown in Fig.
9- 17b. The stresses on the X face are represented by point A, which has
an abscissa of 32 and a negative ordinate of 20. rxy is considered
negative because its moment sense is counterclockwise about the center
of the element in Fig. 9- 17a. The stresses on the Y face are given by
point B, which has an abscissa of - 10 (negative because o
y
is compres-
sive) + 20 (positive because the moment sense of ryx
and an ordinate of
is A and B gives the diameter of Mohr's circle, its
clockwise). Joining
center C being midway between the abscissae of A and B, or at 11 MPa
from the origin O. Hence the radius R is computed from the right
triangle whose sides are 21 and 20; the radius is 29.0 MPa.

j) a (MPa)

Max. a = 40

(b)

Figure 9-17.
384 9 Combined Stresses

The principal stresses are represented by points D and £, where

the shearing stress coordinates are zero. From the geometry of the
circle, we obtain
Max. o = OD - 1 1 -I- 29 = +40 MPa
Min. a = OE = 1 1 - 29 = - 18 MPa
The radius to D makes a counterclockwise angle 20 measured from the
radius CA, which denotes the X axis. From the circle, we see that
tan 20 = 20/21 = 0.952, and hence 20 = 43.6° and = 21.8°. The
principal stresses and principal planes are as shown on the differential
element in Fig. 9- 18a. There is, of course, no shearing stress on the
principal planes.
The on the planes of maximum shearing stress are given
stresses
by the coordinates of points F and G, the values being max. t = 29
MPa and min. t = - 29 MPa; the normal stress on each plane is -I- 1
MPa. The radius CF is 90° counterclockwise from CD, so the normal to
the plane of maximum shearing stress is 45° counterclockwise from the
maximum principal plane, or at 45° + 21.8° = 66.8° with the X axis.
These results are shown on the element in Fig. 9- 18b.
To complete the solution, the stresses on the plane whose normal
is at +36.8° with the X axis are represented by point located at the H y

intersection of the radius CH with

Mohr's circle (see rule 3). From rule
5, the angle between the normals to any two faces is laid off double size

on the circle; hence, angle ACH = 2 X 36.8 = 73.6° and angle HCD =
73.6 — 43.6 = 30°. Therefore the coordinates of point H are
a - 11 + 29 cos 30° - 36.1 MPa
r = 29 sin 30° = 14.5 MPa
The on the plane whose normal is at + 126.8° with the X axis
stresses
are represented by point /. Points H
and / are 180° apart on the circle
since the planes they represent are actually 90° apart. The coordinates
of point / are

a' - 11 - 29 cos 30° = - 14.1 MPa

t' = -29 sin 30° - - 14.5 MPa

sj***,*
7

(a) Principal stresses (b) Maximum shear stresses

Figure 9-18.
9-7 Mohr's Circle 385

The student should show the above stress components on a complete

sketch of a differential element.

PROBLEMS
(915. Two wooden joists 50 X 100 mmare glued together mm
along the joint AB as, shown in Fig. P-925. Determine the normal stress
and shearing stress in the glue if P= 100 kN. Ans. a — 15 MPa

60°

Figure P-925.

926. A short 50-mm-diameter circular bar is made of a material

for which the maximum safe compressive stress 80
is MN/m 2
and the
maximum 30
safe shearing stress is MN/m 2
maximum . Determine the
axial compressive force which can be safely applied to the bar.

An element is subjected
4*27: to the principal stresses ax = — 50
MPa and ay = 30 MPa. Compute components on planes
the stress
whose normals are at +30° and +120° with the X axis. Show your
answers on a complete sketch of a differential element.

928. A small block is 50 mm long, 30 mm high, and 10 mm thick.

The block is subjected to uniformly distributed tensile forces having the
resultant values shown in Fig. P-928. Determine the stress components
developed along the diagonal AB.
Ans. a = 83.9 MPa; t = - 39.7 MPa
929. Solve Problem 928 assuming that the 30 kN forces are
reversed to act in compression.

30 kN

1 B
~7\

30mm S
45 kN /
/ 45 kN
/
•yS 50 mm

I30 kN
Igures P-928 and P-929.
 9 Combined Stresses

930. A closed cylindrical tank is fabricated from 10-mm plate

and subjected to an internal pressure of 1400 kPa. Determine the

maximum diameter if the maximum shearing stress is limited to 30
MPa. [Hint: In a closed-end thin-walled cylindrical pressure vessel, the
longitudinal stress is given by pD /At and the circumferential stress is
pD/2t (see Art. 1-6). Also, refer to Problem 941.]
931. For the state of stress shown in Fig. P-931, determine the
principal stresses and the maximum shearing stress. Show all results on
complete sketches of differential elements.
Arts. Max. o = 90 MPa at B = - 18.4°

80 MPa
X

30 MPa
Figure P-931.

932. If a point is subjected to the state of stress shown in Fig.

P-932, determine the principal stresses and the maximum shearing
stress. Show all results on complete sketches of differential elements.
Ans. Max. a = 58.1 MPa at B = 19.9°; -I-

Min. o = - 98.1 MPa at B = 4- 109.9°

50 MPa

40
—
MPa

Figure P-932.
9-7 Mohr's Circle 387

933. For the state of stress shown in Fig. P-933, determine

normal and shearing stresses on the planes whose normals are at +60°
and +150° with the X axis. Show these stresses on a sketch of the
element. Arts. For 9 - 60°, a - 7.32 MPa and t - 10 MPa

lOMPa
L - 20 MPa

10 MPa

20 MPa

Figure P-933.

934. If an element is subjected to the state of stress shown in Fig.

P-934, find the principal stresses and the maximum shearing stresses.
Also determine the stress components on planes whose normals are at
45° and 135° with the X axis. Show all results on complete sketches of
the appropriate elements.

i k
60 MPa

40 MPa

60 MP
X
'

40 MPa

I
Figure P-934.
388 9 Combined Stresses

935. For the element shown in Fig. P-935, determine the values
of ax and ay if the principal stresses are known to be 20 MPa and - 80
MPa.

Y
i

iiOy

30 MPa

°x
1

30 MPa

T
Figure P-935.

936. A tube with an external diameter of 150 mm is fabricated

from 10-mm plate. Theweld used makes an angle of +30° with
spiral
the longitudinal axis. Determine the maximum torque that can be
applied if the shearing stress along the weld is limited to 30
2
MN/m .

Ans. T= 17.3 kN-m

937. A closed-end cylindrical pressure vessel having an outside
diameter of 600 mm
and fabricated from 10-mm steel plate is subjected
to an internal pressure of 1400 kPa. Determine the normal and shearing
stresses on the spiral weld used which makes an angle of + 30° with the
longitudinal axis.

938. The state of stress at a point is the result of two separate

actions: one produces the pure shear of 35 MPa shown in part (a) of
Fig. P-938, and the other produces the pure shear of 30 MPa shown in
part (b). Find the resultant stress by rotating the state of stress in part

35 MPa
(a)

Figures P-938 and P-939.

9-8 Applications of Mohr's Circle to Combined Loadings 389

(b) to coincide with that in part (a) so that the stresses can be super-
posed and added directly. Then determine the principal stresses and
principal planes. Ans. Max. a - 32.9 MPa at = - 18.8°

939. Solve Problem 938 assuming that the directions of the 30

MPa shearing stresses in Fig. P-938b are reversed.

940. The state of stress at a point is the result of three separate

actions that produce the three states of stress shown in Fig. P-940.
Determine the principal stresses and principal planes caused by the
superposition of these three stress states.
Ans. Max. a = 37.8 MPa at 9 = 26.2°

40 MPa 20 MPa

-
30 MPa
X

(a) (b)

Figure P-940.

941. The principal stresses on an element are x , oyy and az

<t .

Assuming ax > ay >

a2 show that the maximum
, shearing stress on any
plane through the element is equal to \(ax — oz ). {Hint: Consider all
orientations of the element by drawing three separate Mohr's circles,
each representing a rotation about one of the three principal axes.)
942. A state of plane stress is defined by ax = 20 MN/m2 ,

ay = 40 MN/m 2
20
, and rxy = MN/m
Determine the maximum 2
.

shearing stress on any plane through the stressed point. (Hint: First find
the principal stresses and then apply the results of Problem 941.)
Ans. Max. r = 26.2 MN/m2

9-8 APPLICATIONS OF MOHR'S CIRCLE

TO COMBINED LOADINGS

The most important use of combined stresses is in the design of

members subjected to combined loadings, or the determination of safe
loads. Here Mohr's circle makes possible a visualization of conditions
that is superior to mere analytical manipulation. The usual procedure is
to consider an element on which the effect of the three fundamental
loadings —
axial, torsional, and flexural —
can be computed. A study of
Mohr's circle for this element indicates the design criteria. The illustra-
tive problems at the end of this article are typical of the procedures
involved.
390 9 Combined Stresses

Stress trajectories

An element on the surface of the cylinder in Fig. 9- 19a is

subjected to the indicated torsional shearing stress. Fig. 9- 19b shows

Mohr's circle for this state of stress. The radius OA specifies the X axis.
The maximum denoted by point D whose radius OD is
tensile stress is y

90° clockwise from OA. Hence the normal to the plane of maximum
tensile stress is 45° clockwise from the X axis, as shown in Fig. 9- 19a.
The lines in Fig. 9-20 that follow the directions of the principal stresses
are called stress trajectories. For torsion, they are 45° helices. If the
material is weakest in tension, as is common for brittle materials, failure
occurs along a 45° helix such as AB. This may be confirmed experimen-
tally by twisting a piece of chalk until it breaks.
Further visualization of the tensile and compressive stresses in-
duced by pure shear is provided in Fig. 9-21. The distorted appearance
of the element A BCD, originally rectangular, indicates that the diagonal
AC has been lengthened and BD shortened. These deformations agree
with the directions of the tensile and compressive stresses previously
obtained.
In beams, the directions of the principal stresses vary with the
intensities of the flexural stress af and the horizontal shearing stress t.

For example, at point A of the cantilever beam in Fig. 9-22, Mohr's

circle shows that the direction of the principal compressive stress makes
a clockwise angle 9 with the X axis; the principal tensile stress is at right

Torsional shearing Equivalent

stress (a) principal stresses

Figure 9-19. Cylinder subjected to torsion

Figure 9-20. Stress trajectories for torsion.

9-8 Applications of Mohr's Circle to Combined Loadings 391

X
Figure 9-21.

-4 F Max. o f

State of stress
at point A

Mohr's circle for

stress at point A
Figure 9-22.

angles. The value of 9 varies with the ratio r/af (i.e., tan 29 = 2r/af).
At the extreme fibers m and n of the section through A, t is zero and the
principal stress directions are horizontal and vertical. At the neutral
plane where oy is zero, the principal stresses are at 45° to the X axis.
The solid and dashed lines on the beam represent the stress
trajectories. They consist of two systems of orthogonal curves whose
tangents at each point are in the direction of the principal stresses at
that point. The solid lines indicate the direction of the maximum
compressive stresses, and the dashed lines indicate the direction of the
maximum tensile stresses. Be careful not to confuse stress trajectories
 392 9 Combined Stresses

with lines of constant stress. Stress trajectories are lines of principal stress
direction but of variable stress intensity.

ILLUSTRATIVE PROBLEMS

943. A shaft 100 mm in diameter that rotates at 30 r/s is

subjected to bending loads that produce a maximum bending moment

of 25007T Nm. Determine the torque that can also act simultaneously
on the shaft without exceeding a resultant shearing stress t = 80 MPa
or a resultant normal stress a = 100 MPa. What is the maximum power
that can be transmitted by the shaft?

Solution: The bending moment produces a maximum flexural stress at

the top or bottom of the shaft. Its value is

Mc 4M_ _ 4(2500^ = 80Mpa

I 3
tt(0.050)

An applied torque T, as yet undetermined, produces a torsional

shearing maximum at the periphery of the shaft, which also acts
stress,

on the element at the top or bottom of the shaft.* This state of stress is
shown in Fig. 9-23a. Although the torsional shearing stress (denoted by
t, to distinguish it from the maximum resultant shearing stress t which

the shaft can withstand) is not yet known, Mohr's circle can be drawn in
terms of it, as in Fig. 9-23b.

Max. r=80MPa = # r

oy = 80MPa

0^ = 80 MPa
Max. o = 100 MPa
(a) = 40 + R a

(b)

Figure 9-23.

•At the extremities of the neutral axis, the torsional shearing stress is added
to the shearing stress caused by the vertical shear, thereby forming a resultant
shearing stress In a short heavily loaded shaft, this can be the maximum resultant
shearing stress that limits the value of the torsional shearing stress. However, this
possibility is ignored here.
9-8 Applications of Mohr's Circle to Combined Loadings 393

To produce the maximum permitted resultant shearing stress, the

radius of the circle must be /^. =
80 MPa. However, the radius that will
produce the maximum permitted normal stress must satisfy the condi-
tion a = 100 - OC + Ra = 40 + Ra whence Ra = 60 MPa. ,

It should be clear that the proper radius of Mohr's circle is R^. or

i^, whichever is smaller, so that the allowable values of max. t or
max. a will not be exceeded. Having thus determined the proper radius,
that is, R =
60 MPa, we compute, from the shaded triangle in Fig.
9-23b, the maximum torsional shearing stress t, that can be combined
with the flexural stress. Hence
2 2
Tf
2
= R2 - (40) = (60) - (40)
:

or

T( = 44.7 MPa

The torsion formula shows that the torque required to produce this
torsional shearing stress is

6 3
rJ rirr (44.7 X 10 )7r(0.050)
r 2
= 8780N-m Ans.
Finally, the maximum power that can be transmitted by the shaft
is given by

[<3> = 2tt/T] <$ = 2tt(30)(8750) - 1650 kW Ans.

944. A solid shaft is subjected to simultaneous twisting and

bending due to a torque T and a maximum bending moment M.
Express the maximum resultant shearing stress r and the maximum
resultantnormal stress a in terms of T, Af, and the radius r of the shaft.
Using these relations, determine the proper diameter of a solid shaft to
carry simultaneously T = 1200 and Nm
= 900 Nm, if o < 100 M
MPa and t < 70 MPa.

Solution: The simultaneous bending and twisting in this problem is

commonly encountered in designing shafts. The formulas that will be

developed are very useful, but their use should be limited to cases in
which both M
and T are known. Under any other conditions, Mohr's
circle should be used.
The state of stress of an element subjected to simultaneous flexure
and torsion shown in Fig. 9 -24a, and the equivalent Mohr's circle in
is

Fig. 9-24b. The maximum resultant shearing stress t is equal to the

radius R, which, from the shaded right triangle, is

2 2
Max.r=i?=^/(ior/) + (T,) (a)
394 9 Combined Stresses

7 f Max. t = R

\ T'

'V^—f \7+~°t
1
° H ]/
^§§?c Of
\

I
Max o =

o
\o f + R

§p™
(a)
T
'\
(b)

Figure 9-24.

The following variations of the flexure and torsion formulas as

applied to a circular shaft are used:

of = —
AM
- and t = —IT (A)

Substituting these values in Eq. (a) yields

Max.

which reduces to
t
-#H r\
3 J

Max. t = -K VM 2 + T 2
7rr

Letting 7e - Va/ 2 + T 2 we , obtain finally

271
Max. t = (9-11)
*nr

The similarity between Eq. (9-11) and the torsion formula in Eq. (b)
suggests equivalent torque as a suitable name for Te .

An equation for maximum resultant normal stress that is similar to

the flexure formula but involves an equivalent moment Me
is obtained as
follows: In Fig. 9-24b, the maximum and resultant normal stress is
max. o =\of + R. Substituting o = 4M/7rr
f
3
and R = 2Te /7rr 3 we ,

have

Max. a = —
vr
(M + T e)

Multiplying and dividing the right side by 2 gives

4Af
Max. a =
-
(9-12)
•zrr
9-8 Applications of Mohr's Circle to Combined Loadings 395

which is equivalent to the flexure formula in Eq. (b) if the equivalent

moment M e is expressed by

3/,-±(M + Te )
It is not necessary to memorize Eqs. (9-11) and (9-12) because
they are so similar to the torsion and flexure formulas. In using them,
however, the following definitions of equivalent torque and equivalent
moment must be remembered:

Te = VM
+ T2 2
(9-13)
M e
=\{M+ Te ) (9-14)

For the numerical data given in this problem, the equivalent

torque and equivalent moment are

Te =V M + T 2 =^/(900) 2 + (1200) 2 =
2
1500 Nm
M e
- \{M + Te ) = ±(900 + 1500) = 1200 Nm
The shaft radius required so that the maximum shearing stress will
not be exceeded is found from Eq. (9-11):
27;

2 ( 15 Q°)
70 X 10
6 =
3
or r = 23.9 X 10" 3 m = 23.9 mm
77T

The radius that will avoid exceeding the maximum normal stress
is, from Eq. (9-12),
4A/
a =
vr*
4(1200)
100 X 10
6
= v ;
,
3
or r = 24.8 X 10" 3 m= 24.8 mm
Trr

The larger of these two radii will satisfy both stress conditions; hence
the proper diameter is

rf = 2x 24.8 = 49.6 mm Ans.

945.Design a solid shaft to carry the loads shown in Fig. 9-25, if

max. t < 70 MPa and max. a < 120 MPa. The belt pulls on pulleys B
and C are vertical, and those on pulley E are horizontal. Neglect the
masses of the pulleys and shaft.

Solution: The given loading produces bending in both the vertical and
horizontal planes. The bending moment diagrams in these planes are
given in Fig. 9-25b, c. The resultant moment at any section is given

by M =yAf A
2
+ M 2
. Therefore the resultant bending moments at
396 9 Combined Stresses

500 N

500 N 2000 N
^=1250 N
(a) Loading on line shaft

3500 N 500 N
A u = 3625 N
3625 N-m
2875 N-m

5000 N-m
(c) Bending moments in horizontal plane

1500 N- m
750 N-m

(d) Torque distribution in shaft

Figure 9-25.

B y C, and D are MB = 3834 N-m, Mc = 4725 N-m, and MD =

5000 N
m. Combining these values with the torque distribution in the
•

shaft (Fig.9-25d) shows that the dangerous sections are at C and D.

Since both moment and torque values are known, it is advanta-
geous to use the method in Illustrative Problem 944. Applying Eqs.
(9-13) and (9-14), we find the equivalent torque and equivalent mo-
9-8 Applications of Mohr's Circle to Combined Loadings 397

ment to be

at C: Te = V M 2 + T 2 = y(4725) 2 + (1500) 2 = 4957 N-m

Me = \{M + Te ) = |(4725 + 4957) = 4841 N-m
at D: Te = V M 2 + T 2 = y(5000) 2 + (750) 2 = 5056 N-m
Me = \(M + Te ) = £(5000 + 5056) = 5028 N-m
The larger value of Te is applied to Eq. (9-11), and the larger value of
M e
is applied to Eq. (9-12). In this problem, both max. Te and max. M e
occur at D. Hence, we obtain

2Te
T =
m>
ikj /\ iv - r = 35.8 X 10" 3 m = 35.8 mm
n>
r 4m c i

"
L° - 3
J

120X10'=
4
^
•nr*
r = 37.7 X 10" 3 m= 37.7 mm

The larger value of r determines the proper radius. Therefore the

required diameter of the shaft is d= 2 X 37.7 =
75.4 mm. In view
of the standard sizes of shafting, a shaft of 80-mm diameter would be
selected.

PROBLEMS
946. Explain why the stress trajectories in Fig. 9-22 tend to
become horizontal as they approach the wall. Where are they exactly
horizontal? What are the stress trajectories for axial tension or compres-
sion?

The solid shaft in a small hydraulic turbine is 100 mm in

947.
diameter and supports an axial compressive load of 1407T kN. De-
termine the maximum power that be developed at 4 r/s without
exceeding a maximum shearing stress of 70 MN/m 2
or a maximum
normal stress of 90 MN/m 2
. Arts. 273 kW
948. A solid shaft 100 mm in diameter is subjected simulta-
neously to an axial compressive force of 600 kN and to a torque which
398 9 Combined Stresses

twists the shaft through an angle of 1.5° in a length of 8 m. If

G 80 x 10
9
N/m 2 , compute the maximum normal and shearing
stresses in the shaft. Ans. r = 40.4 MN/m 2

949. A solid shaft 100 mm in diameter carries simultaneously an

axial tensile load of SOtt kN, a maximum bending moment of Itt
kN-m, and a torque of 37r kN-m. Compute the maximum tensile,
compressive, and shearing stresses produced in the shaft.
Ans. o, = 106 MPa; ac = 74.8 MPa; r = 63.8 MPa
950. Repeat Problem 949 assuming that the axial load is changed
to a compressive load of 4077- kN.
951. A solid shaft 80 mm
in diameter is subjected to a maximum
bending moment of 800tt N-m and an axial tensile load of 4077 kN.
Find the maximum torque which can be safely applied if max. o < 100
MN/m 2
and max. t < 80 2
MN/m .

952. A solid circular shaft is used to transmit simultaneously a

torque of 2600 N-m and a maximum bending moment of 2000 N-m.
Determine the radius of the smallest shaft which can be used if max. a
< 80 MPa and max. t < 60 MPa. Ans. r = 34.8 mm
953. A shaft 80 mm in diameter carries a maximum bending
moment of 3 kN-m. What torque can also be applied without exceed-
ing a maximum shearing stress of 80 MN/m or a maximum normal
2

stress of 120 MN/m ? 2

954. A closed-end cylindrical pressure vessel has an external

diameter of 400 mm and a wall thickness of 20 mm. The vessel carries
simultaneously an internal pressure of 4 MPa, a torque of 80 kN-m,
and a bending moment of 20 kN-m. Determine the maximum tensile
stress in the walls of the vessel. Neglect the possibility of buckling.
Ans. Max. a, = 53.9 MPa
955. A closed-end cylindrical pressure vessel has an external
diameter of 300 mm
and is fabricated from 10-mm steel plate. If the
tank is subjected to an internal pressure of 6 2
find the maxi- MN/m ,

mum torque which can also be applied if the normal stress in the walls
of the vessel is limited to 100 MN/m 2
. Neglect the possibility of
buckling.

956. Compute principal stresses and maximum shearing

the
stress at point P-956 at the section x = 250 mm. The beam is
A in Fig.
rectangular, 20 mm wide by 120 mm deep, and point A is 20 mm above
the centerline of the beam. Assume the 50-kN load acts at the centroid
of the cross section. Show answers on complete sketches of appropriate
9-8 Applications of Mohr's Circle to Combined Loadings 399

differential elements. {Hint: Be sure to include the shearing stress

caused by the applied load.)
Ans. Max. a = 72.6 MPa at = - 12.9°

-* X *-

'/A

A'
I
,

1
Bj^O'
J

^ 50 kN

Figures P-956 and P-957.

957. For the 20 by 120 mm mm

beam described in Problem 956,
determine the stress components on the 30° plane at point B. Assume
that x — 300 mm and that B is 20 mm below the centerline of the beam.
Show answers on a complete sketch of a differential element.

958. A 50-mm diameter bracket, securely fastened to the wall as

shown in Fig. P-958, carries the given horizontal and vertical loads.
Find the principal stresses and maximum shearing stress at point A on
Ans. Max. a = 41.2 MN/m
2
the upper surface of the shaft.

4000 N
3600 N
Figures P-958 and P-959.
400 9 Combined Stresses

959. Repeat Problem 958 to find the principal stresses and maxi-
mum shearing stress at point B on the front surface of the shaft.

960. A 100-mm diameter shaft carries an axial load P and a

torque T, as shown P-960. Determine the normal and shearing
in Fig.
stresses on the spiral weld which makes an angle of 30° with the axis of
the shaft. Arts, a = 18.4 MPa; t = - 16.5 MPa

P = 160 kN
T=3kN-m
Figure P-960.

961. Twenty kilowatts are transmitted through a speed reducer.

At one part of the machine, as shown in Fig. P-961, the pinion drives
the gear A of shaft AB at 6 r/s. Determine the minimum diameter of
shaft AB max. a < 80
if MN/m2 and max. t < 60 2
MN/m
Consider
.

only torsional and bending stresses in the shaft.

Gear
250 mm P.D
Figure P-961.

962. A line shaft50 mm

in diameter is subjected to the loads
shown in Fig. P-962. The belt pulls on pulley A are horizontal, and
those on pulley B are vertical. Calculate the maximum normal and
shearing stresses developed in the shaft.
9-8 Applications of Mohr's Circle to Combined Loadings 401

600 mm

600 mm
-. ^£ £
600 mm
SL mm dia

vV
-250
150 N/^
i ^shr *
li4
/2700N 300 N

$L 1350 N ^^ 500 mm dia.

Figure P-962 m

963. Design a solid shaft to carry the loads shown in Fig. P-963
if max. t < 60 MPa and max. a < 80 MPa. The belt pulls at A and C
are horizontal, and those on pulley E are vertical.
Arts, d= 68.4 mm

5000 N

0.80 m dia

200 N 2200 N
Figure P-963.
402 9 Combined Stresses

9-9 TRANSFORMATION OF STRAIN COMPONENTS

Most of the problems encountered in engineering design involve a

combination of axial, torsional, and flexural loads applied to homoge-
neous materials of a prismatic shape. In such cases, the stresses may be
computed as described in the preceding articles and maximum resultant
stresses used as a criterion of design. Occasionally, however, irregulari-
ties in a structure, or conditions which violate the basic assumptions of
the torsion or flexure theories, require us to resort to experimental
methods of determining stresses. Since stress is a mathematical concept
that represents the intensity of force on a it cannot be
unit area,
measured directly. Nevertheless, the stress-strain relations defined by
Hooke's law permit us to estimate stresses from strains which can be
measured. In this article we study the transformation of a given set of
strains into principal strains. In the next article we consider the applica-
tion of strainmeasurements and their conversion into stresses.
Consider now an element subjected to the general state of plane
stress shown in Fig. 9-26a. The normal stresses (assumed to be tensile)
elongate the element in the X and Y directions, and the shearing stresses
distort the element through the shearing strain yxv as shown in Fig.
9-26b. The effect of these strains on any line element OA in Fig. 9-26a

(b)

y"y dy
A'

— K
/ —
r € dx
x
dy

*ydy

(d)

Figure 9-26.
 .

9-9 Transformation of Strain Components 403

is shown in Fig. 9-26c, where OA is elongated to OA' and also changes

its angular position by a very small amount that we call ft. The
movement of A to A ' is the vector sum of the elongation €x dx in the X
direction, e
y
Y direction, and
dy in the the shear distortion yxy dy in the
X direction. These movements of A are exaggerated in Fig. 9-26d to
show how their vector sum may be resolved into two components
parallel and perpendicular to OA. The parallel component A" A' repre-
sents the increase in length of OA, whereas the perpendicular compo-
nent AA " causes the change /? in the angular position of OA
The magnitude of A" A' found by , projecting c^ dx, ty dy, and
y^ dy upon the direction of OA, is
A" A' = €x dx cos 9 + e
y
dy sin 9 - yxy dy cos 9 (a)

This increase in length of OA divided by its original length ds is defined

as the strain ea in the direction OA :

A" A' _ €x dx cos 9 t dy sin 9 yxy dy cos 9

Cfl
__
~ ~ds ds
+ y
ds ds
^
But from Fig. 9-26c, dx/ ds = cos 9 and dy / ds = sin 9. With these
relations, Eq. (b) reduces to

cfl = €x cos 9
2
+ e
y
sin
2
9 — yxy sin 9 cos 9 (c)

2 2
On substituting for cos 9 and sin 9 their equivalents in terms of 29,
this becomes

«« - €jL
^ + ^pcos 20 - \y xy sin 20 (9-15)

The angular deviation of OA (i.e., /?) is determined by dividing the

perpendicular component A A" by the original length ds of OA. Project-
ing the movement of point A upon the direction perpendicular to OA,
we obtain, from Fig. 9-26d,

AA " = cx dx sin 9 - €
y
dy cos 9 - yxy dy sin 9 (d)

whence
dy cos yx, 4> sin 9
_ AA" _ cx </* sin €^

ds ds ds ds
= €x sin 9 cos 9 - e
y
sin 9 cos 9 - yxy sin
2
(e)

For the line element at right angles to OA, the angular deviation
fi'may be found by substituting 9 + 90° for 9. Since sin(0 + 90°) =
cos 9 and cos(0 + 90°) = - sin 9, we obtain

ft' — -ex sin 9 cos 9 + e^ sin 9 cos 9 - yxy cos

2
(/)
404 9 Combined Stresses

Since fi and /?' rotate in opposite directions,* their absolute sum is

equal to their algebraic difference.Hence the total change in the right

angle between OA and its normal OB, which defines the shearing strain
for an element located 0° from the X- Y axes, becomes

yab = p - /?' = €jc (2 sin cos 9) - c (2 sin

y
cos 0)

+ yxy (cos 2
9 - sin
2
0)

which, in terms of the double angle 20, reduces to

\y* -
(<X ^ sin 20 + \y*y «" 20 ( 9 ~ 16 )
2

A comparison of Eqs. (9-15) and (9-16) with the normal and

shearing stress transformation given by Eqs. (9-5) and (9-6) shows that
they are identical in form.We conclude that the normal and shearing
strainscan be represented by a Mohr's circle for strain, constructed in
the same manner as Mohr's circle for stress except that half values of
shearing strain are plotted instead of shear stress.
In applying Mohr's circle for strain, we use the following rules of
sign. Extensional strains are considered positive, compressive strains
negative,and shearing strains positive when they increase the original
right angle of an unstrained element.* A more general rule of sign for
shearing strain is obtained by denoting the strain between two per-
pendicular directions OA and OB by yab the first subscript indicating ,

the direction OA associated with the angle 0. The shearing strain is

considered as positive if the directional line associated with the first

subscript (OA in this case) moves clockwise relative to the other

directional line, and vice versa. This rule of sign makes y^ = — y^.
This agrees with our convention for shearing stress, that rxy = — ryx , as
well as for products of inertia, that P^ = - Pyx .

The form between stress and strain components is

similarity in
further exemplified by the following statement: A Mohr's circle for
strain can be transformed into a concentric Mohr's circle for stress by
means of the scale transformations

(9 " 17)
*--*«T§7
(OC)„ = (OC) tT
^ (9-18)

•If ft and /?' were in the same direction, there would be no change in the right
angle, i.e., no shearing as is easily verified by adding fi and /?'.
strain,
t
This is contrary to the convention adopted in the mathematical theory of
elasticity where the closing of a right angle is taken as positive shearing strain. The
convention adopted in this book not only achieves consistency between Mohr's
circles of inertia, stress, and strain but avoids the artificiality of plotting values of
positive shearing strain in a negative or downward direction.
9-9 Transformation of Strain Components 405

2 y ^~~ ""*^\ ^~ Strain circle

Stress circle

Figure 9-27. Transformation of Mohr's circle of strains to Mohr's circle of stress.

in which R^ and R€ are respectively the radii of the stress and strain
circles in Fig.9-27 and (OC) a and (OC\ are respectively the stress and
strain coordinates of the centers of the concentric circles. Also E is the
modulus of elasticity and v is Poisson's ratio. The proof of these
relations is called for in Problem 965. Their application is discussed in
the following illustrative problem.

ILLUSTRATIVE PROBLEM

964. In a body subjected to plane strain, there act at a certain

point ex -
800 X 10" 6 m/m, €>,
= 200 X 10" 6 m/m, and yxy = 600 X
10" 6 rad. Compute (a) the principal strains and the principal strain
axes; also (b) the strain ca in a direction of 60° with the X axis, the
strain tb perpendicular to ca , and the shearing strain y^. Further, if

E= 200 GPa and v = 0.30, determine the principal stresses and the
normal and shearing stresses on the element rotated 60° from the X
axis.

Solution: Mohr's circle for the given state of strain is shown in Fig.
6
9-28. The factor 10" being omitted, the coordinates of point A are
^ — 800 and \yxy = 300; and the coordinates of B are c y = 200 and
\yyx = — 300. The X axis is represented by the radius CA, and the Y
axis by the radius CB. The radius of the circle is computed to be 424,
whence the maximum principal strain c,, denoted by point D, equals
500 + 424 = 924 X 10" 6 The minimum principal strain € 2 denoted by
. ,

point E, equals 500 — 424 = 76 X 10"

6
The angle between the maxi-
.

mum strain axis and the X axis is one-half the angle ACD, or 22.5°, in a
clockwise direction from the X axis, as shown in Fig. 9-29.
To determine the strain ca , lay off the radius CF at twice 60°, or
120°, counterclockwise from the X
denoted by CA. The strain
axis
of point F are ea = 500 - 424 cos 15° = 90 X 10~
6
coordinates ,

and \yab = 424 sin 15° = 110, whence y^ = 220 X 10" 6

radians.
406 9 Combined Stresses

Figure 9-28. Strain circle.

Figure 9-29.

Laying off the 90° angle between the directions of ca and eb to double
scale locates G as diametrically opposite from F; hence tb =
500 424 cos 15° = 910 x
-I-
10" 6 .

Except for changing the symbols and plotting only half values of
shearing strain, the procedure is the same as that given in Art. 9-7
describing Mohr's circle for stress.
The stress components may now be found by applying Hooke's
law to the strain components, as will be shown later; but the most
convenient method is to transform the strain circle into the stress circle.
The only results needed from the strain circle are the radius and the
location of the center. These are transformed into corresponding values
9-9 Transformation of Strain Components 407

for the stress circle by means of Eqs. (9-17) and (9-18). We obtain

E
R= K 1 -I- v

*
*a = (424x W- 6 )
200
1 + 0.30
^
1

E
(OC) a = {OC\-
9
200 x 10
(OC) a = (500 x 10-*)
_ - 143 MPa
t 30

Using these values, we plot Mohr's circle of stress as shown in Fig.

9-30. The points are labeled to correspond with those in Fig. 9-28.*
Reading from the circle, we see that the principal stresses at D and E
are respectively

at D: Max. o = 143 -I- 65.2 = 208 MPa

at E: Min. a - 143 - 65.2 - 77.8 MPa
For the element rotated 60° from the X axis, the stress compo-
nents are given by points F and G, and the values are

at F: o= 143 - 65.2 cos 15° = 80.0 MPa

t = 65.2 sin 15° = 16.9 MPa
at G: a = 143 + 65.2 cos 15° = 206 MPa
Instead of using the transformed circle of stress, we can find the
stresses directly from the strains, using Hooke's law for biaxial stress

a (MPa)

Figure 9-30. Stress circle.

*A separate figure is unnecessary because the circles could be plotted

it would be confusing to do so here.
concentrically, but
408 9 Combined Stresses

(sec page 47) as expressed by the following equations:

£(£, + !>£,) £(€, + vtx )

y
1 - v
1 '
X-v 1

Txy - Gy. IV
2(1 + v)

On substituting the principal strains of 924 X 10" 6 and 76 X 10" 6

found from the strain circle in Fig. 9-28, the principal stresses are
computed to be

Max. o =
(200 x.0-)(924

1 -
+ 0.30

(0.30)
2
X76X.0-) = ^ Mpa
9 6
(200 x 10 )(76
" -I- 0.30 x 924)(10"
-
Mm. o = J :
2
) '- = 77.6 MPa
1 - (0.30)

These results agree with those found previously; consequently if a

Mohr's circle of stress is plotted using these principal stresses, the radius
and center will have the values shown in Fig. 9-30.
In a similar manner, the normal and shearing stresses on the
element at 60° from the X axis can be determined from the correspond-
ing strain components of ca = 90 X 10"
6
cb = 910 X 10"
6
and yab = , ,

6
220 x 10~ Applying Hooke's law, we obtain
.

6
_ (200 x 1Q>)(90 + 0.30 X 910)(1Q- ) _
a ?? g Mpa
1 - (0.30) 2
(200xl0-)(220xl0--)
ab
= 169Mpa
2(1 + 0.30)

A comparison of these two methods of computing stress compo-

nents should convince you of the advantages of transforming the strain
circle to the stress circle.

PROBLEMS

965. As shown in Fig. 9-27, prove that Eqs. (9-17) and (9-18)
will transform a strain circle into a stress circle.
966. Starting with an element subjected only to principal stresses,
show that the angular deviation
fi of a line element such as OA in Fig.
9-26 is equal to one-half the shearing strain y^.
967. A state of strain is defined by cx = - 400 X 10" 6 , c^ = 200
x 10" 6 and = 800 x 10" 6 If = -
,
yx> , . E 200 GPa and v 0.30, de-
9-1 The Strain Rosette 409

termine the principal stresses and stress components on the face whose
normal is at + 40° from the axis. X
Ans. MPa; min. a = - 106 MPa;
Max. a = 48.3
a = - 97.2 MPa; r = - 34.8 MPa

968. A state of strain is defined by ex = 600 X 10" e = - 400

6
,
y
X 10" 6
and yxy - -600 X
,
10" 6
If E = 200 x 10 N/m and
.
9 2 - i>

0.30, determine the principal stresses and maximum shearing stress.

969. The strain components at a given point are ex = — 800 X

lO"
6
,€y
- 200 X 10" 6 and yxy = - 800 x 10" 6 If E = 200 GN/m2
, .

and v = 0.30, find the stress components on the face whose normal is at
+ 20° from the X axis. Ans. a = - 105 MPa; r = - 96.6 MPa

9-10 THE STRAIN ROSETTE

The stress in a bar subjected to uniaxial stress can be determined

experimentally by attaching a strain gage oriented in the direction of the
stress. The stress is then computed, in terms of the strain, from a = Ee.
The strain is generally small (under 1 part in 1000); hence sensitive
instruments are required for measuring it. Originally, strain gages were
mechanical or optical, but these have now been almost completely
replaced by electrical gages. This type of gage contains a wire or foil
element whose electrical resistance varies with its deformation. The gage
is cemented to the test specimen, the strain in the specimen being
measured as a function of the change in the electrical resistance of the
gage element. The wire-type of strain gage has been brought to a high
state of perfection by the Baldwin Southwark Division of the Baldwin
Locomotive Works. These gages are marketed under the well-known
SR-4 trademark of the Baldwin-Lima-Hamilton Company.
As was said above, a single strain gage oriented in the direction of
a uniaxial stress is sufficient for computing the stress. For biaxial stress,
we might suppose that two strain gages would be sufficient; this would
be true if the directions of the principal stresses were known, but this is
not usually the case. To determine the direction of the principal stresses
in addition to their magnitudes, three values of strain are required. In
the preceding article we showed how the values cx €
,
y
and yxy can be
,

used for this purpose. Unfortunately, there is no equipment that con-

veniently gives a direct measurement of the shearing strain yxy so other
,

methods are needed.

We will now show that a state of strain is uniquely determined by
the measurement of three linear strains €a c6 and €c acting in three , ,

arbitrary directions a 9b and 9C at the same point as in Fig. 9-31. By

, ,

substituting these strains in Eq. (9-15), we obtain the following set of

 410

(a) General strain rosette (b) 45° strain rosette (c) 60° strain rosette

Figure 9-31. Strain rosettes.

simultaneous equations:

£„ + €., Cv — C.
* _ y
cos 20 sin 20,

*x + S
-cos2^ -^sin206 («)

** + *y
<c
"_ y^cos20c --^sin20 c

The solution of these equations determines the required strain compo-

nents cx , eyy and yxr
As a matter of practical convenience, the linear strains are ob-
tained by using either of two combinations of three resistance strain
gages: (l) three gages set with their axes at 45° with each other, or (2)
three gages whose axes are at 60° with each other as in Fig. 9-3 lb and
c. These combinations are known as strain rosettes. The three gages are
electrically insulatedfrom each other and are used to determine the
strain at the surface of the structure to which they are attached. We
shall now construct a Mohr's circle of strain for each of these rosettes.

The 45° or rectangular strain rosette

By substituting 9a = 0°, Bb = 45°, and 6C = 90° in Eq. (a) and

solving, we obtain

=
C+ €.
<x <«>
2^ xy (9-19)

thus defining a state of strain from which the strain circle and the stress
circle may be constructed as explained in the preceding article.
These results can also be obtained by constructing a Mohr's circle
directly from ca c6 and cc as follows: In Fig. 9-32, the three strains ca
, ,
,

V and c r 45° apart, will be represented by three radii CA, CB, and CD,
,
9-10 The Strain Rosette 411

Figure 9-32. Mohr's circle for 45° strain rosette.

90° apart, as shown in the figure. The center C is midway between G

and E, so one side (CE) of triangle ACE is known. To construct the
circle, the other side AE must be computed. Evidently, triangles CBF
and CAF are congruent, so that AE = CF. From the geometry of the
circle, CF = OC - OF; hence

AE = i±S - eb (b)

Also
€„ — €„
CE = (c)

Therefore the radius R = CA is determined from

R = ^(CE) 2 + (AE) 2 (d)

The student should correlate these results with a strain circle con-
structed from the strain components given by Eq. (9-19).

The 60° or equiangular rosette

In the 60° rosette, the reference angles are a

= 0°, Bb = 60°, and
C
— 120°. On substituting these values in Eq. (a) and solving, we obtain
c_ = c.

S= 3^ + c-0 2€
(9-20)
1 X
t ^
412 9 Combined Stresses

From these results, the strain circle and the stress circle can be con-
structed as described in the preceding article.

PROBLEMS

970. For the 60° strain rosette, show that the expressions in Eq.
(9-20) are correct.
971. Show that in a 60° strain rosette, the principal strains are

(. + Ci, + c 2 ~
r z r z r*

*max.
min.
* 5
J
± TV
J
i
Ca( €a ~ «*) + *b(*b ~ O+ *c(*c " O
and the direction of the maximum principal strain is defined by

tan 29 =
V3 (€» - Q
2 *a ~ €* ~ Cc

in which a positive value of 9 is measured in a counterclockwise

direction from the direction of ea .

972. Show that in a 45° strain rosette, the principal strains are

ma v —
«„ + *r 1 f7~ .1
+
~T
*
+ -^-V^ - e* )2 (e
* - 'J
2 V2
and the direction of the maximum principal strain is defined by

€
'
+ €<- 2€*
tan2fl=
*a ~ *c

973. The three readings on a 45° strain rosette are ea = 400 X

10" 6 , eb = - 200 X 10" 6 and
, cc = - 100 X 10" 6 . If E= 200 GPa and
v = 0.30, determine the principal stresses and their directions.
Ans. Max. a = 109 MPa at = - 27.2°
974. Repeat Problem 973 assuming that the strain readings are
€a = 300 x 10" 6 t = 600 X 10" 6 and c = 100 X 10" 6 .
b ,
c ,

975. The measured on a 60° strain rosette are ca = 300 X

strains
10" 6 , €b = -400 x and cc = 100 X 10" 6 If E = 200 GN/m2
10" 6 , .

and v = 0.3, compute the principal stresses and their directions.

Ans. Max. o - 64.2 MN/m at = - 22.0°
2

A
60° strain rosette attached to the aluminum skin of an
976.
airplane fuselage measures the following strains: ea = 200 X 10" 6 tb = ,

200 X 10" 6 and e = 400 X 10" 6 If E = 70 GPa and v -

,
r compute .
\,
the principal stresses and their directions.

977. Repeat Problem 976 assuming that the strain readings are
ta = - 100 x 10~ 6 e b = 200 x 10"
,
6
and cc = -400 X 10~ 6 , .

Ans. Max. a - 7.7 MPa at 9 = 45°

9-1 1 Relation Between Modulus of Rigidity and Modulus of Elasticity 413

9-11 RELATION BETWEEN MODULUS OF RIGIDITY

AND MODULUS OF ELASTICITY

In Art. 2-4 we said that G = E/2(\ + v)\ we are now ready to prove
this relation. The state of stress shown in Fig. 9-33a consists of a tensile
stress ox and a compressive stress a of the same magnitude. Mohr's
y
circle for this9-33b) indicates that an element rotated 45°
(Fig.
counterclockwise as in Fig. 9-33c in subjected to pure shear in which t
is numerically equal to ox and o These shearing stresses deform the
.
y
element abed to the dashed outhne a'b'c'd' in Fig. 9-33a. The right
angle at a has decreased to 90° - y, where y is the shearing strain.
Simultaneously, the right angle at b has increased to 90° + y.
Consider now the deformation of the right triangle aob that has
equal legs. Since ax = — ay = r, we obtain, with Eqs. (2-9) and (2-10),
(see page 47), the normal strains

=
t(1 + v)
= t(1 -I- v)
€x
E S E

Hence the deformed lengths of sides oa and ob respectively are

=
t(1 + v) t(1 + v)
oa oa 1 + ob' = ob 1
E

Figure 9-33. Pure shear and shearing strain.

414 9 Combined Stresses

Therefore, from the right triangle oa'b', we have

r(l -I- v)
l
-
ob
tan oa '*' = tan(45°-f)=^ = -iL- (fl)

From trigonometry, the expanded form of the tangent is

tan 45° - tan I 1 -

^
tan(45°
v
-
\)
2 '
=
1 + tan 45° tan
\
Jr
=
1
\
+ £
(b)

2 2

since for small angles y like those that occur with shearing strain,
tan(y/2) is practically equivalent to y/2 expressed in radians. Equating
this expanded value of tan[45° — (y/2)] to the right side of Eq. (a) gives

2 E

z E
which, with a little algebra, reduces to

y
7 = —
2t(1
E
+ v)
or —t = E
+
y 2(1 v)

Replacing r/y by G, as specified in Hooke's law for shear (Eq.

2-6), we obtain finally

c = " 13
<2 >
2(Tf7)

which expresses the desired relation between the three elastic constants
G, £, and v.

SUMMARY

The normal stresses caused by a combination of axial and flexural

loads are determined from

® P My «,
/gx

'-©i^-r cm)
The positive sign refers to tension and the negative sign to compression.
The circled signs indicate that the axial stress is uniform and of the
same type all over the cross section whereas the magnitude and type of
the flexural stress vary with position.
 5

9-1 1 Relation Between Modulus of Rigidity and Modulus of Elasticity 41

9-3) is the part of a cross section"?

TheJTcernj, of a section (Art.
through which the resultant compressive force must pass if no tensile X
stress is to be developed over the section.
For bodies subjected to other than axial and flexural combinations
of loading, the elements of the body are subject to both shearing and
normal stresses. The stresses on such elements (in fact, on any element)
vary with the angular position of the element and are expressed by the
following equations which were derived in Art. 9-6.

y x y
a- * + cos 20 - rxy sin 20 (9-5)

T - sin 20 + r
*y
cos 20 (9-6)
2
However, the rules for Mohr's circle (page 380) make it unneces-
sary to remember these equations and
more important conditions
the
which determine the maximum resultant normal and shearing stresses.
A drawing of the circle provides all the information needed to compute
the variations in stress at any element. Further applications of Mohr's
circle in practical designing are given in Art. 9-8.
The use of Mohr's circle of strain is described in Art. 9-9. The
procedure is similar to that with the circle of stress, except that half
values of shearing strain are plotted as ordinates. A strain circle is

readily transformed into a stress circle by means of

K = *<TT7 <9- 17>

(OC)„ = (OC),^ (9-18)

whence the principal stresses may be easily determined. For experimen-

tal determination of stress, the strain rosette (Art. 9-10) is especially
useful. The normal strains inany three predetermined directions estab-
lish a state of strain from which the strain circle is constructed and
converted into the stress circle.
 Reinforced
Beams

10-1 INTRODUCTION

It was once common to strengthen timber beams by bolting strips of

steel to them. With increasingly lower prices for steel, this practice has
ceased except where timber is and the cost of transporting steel
plentiful
to the construction site is high. The most common type of reinforced
beam used today is the concrete beam reinforced with steel rods.
The theory of flexure does not apply to composite beams because
it was based on the assumption that the beam was homogeneous and

that plane transverse sections remained plane, whence the strains varied
directly with their distance from the neutral axis. In investigating the
bending of composite beams, only one assumption is retained: that
plane sections remain plane, that is, the strains vary directly with their
distance from the neutral axis.
The most common method of dealing with a nonhomogeneous
beam is to transform it into an equivalent homogeneous beam to which

the flexure formula may be applied. The basic principle involved here is
that the strains and load capacities must remain unchanged. We shall
discuss first the general case of beams composed of different materials
and then reinforced concrete beams, applying to the latter a procedure

416
10-2 Beams of Different Materials 417

more fundamental and more commonly used than a modification of the

flexure formula.

10-2 BEAMS OF DIFFERENT MATERIALS

The timber beam in Fig. 10- la is reinforced with a steel strip, the steel
being securely fastened to the timber so that no slip occurs between
them as the beam is bent. From page 153 we see that the ordinary
theory of flexure is restricted to beams of homogeneous material and
hence does not apply to the beam under consideration. However, by
suitable modifications we can obtain an equivalent section in terms of
one material to which the theory can be applied.
To obtain an equivalent section, consider a longitudinal steel fiber
of the beam at A. Since the steel and wood are assumed to be firmly
bolted together, the strains of the steel and wood fibers at A must be
equal, that is, €, = €„,. Expressing this relation in terms of the stresses
and moduli of elasticity, we have

5- - -^ (a)

Furthermore, in order to be equivalent, the loads carried by the

steel fiber and the equivalent wood fiber must be equal, so
P. - P.,

or, in terms of the areas,

A s°s " A w°w w

Combining Eqs. (a) and (b), we obtain

A w aw
(4K-

(a) Timber and (b) Equivalent wood section (c) Equivalent

steel section steel section

Figure 10-1. Equivalent sections.

418 10 Reinforced Beams

from which, by canceling out aw and denoting the ratio of the moduli of
EJ Ew by n we finally have
elasticity y

Aw - nA, (10-1)

This indicates that the area of the equivalent wood is n times the
area of the steel. The location of the equivalent area is governed by the

condition that the equivalent wood fibers must be at the same distance
from the neutral axis as the steel fibers they replace in order to satisfy
the criterion of equal deformations in Eq. (a). In other words, the
equivalent wood area is n times as wide as the steel it replaces; the
equivalent wood section is shown in Fig. 10- lb. If desired, an equiv-
alent steel section can be set up by replacing the original wood by steel

1/n as wide, as shown in Fig. 10-lc.

The flexure formula can now be applied directly to either the
equivalent wood section or the equivalent steel section. With the equiv-
alent wood section, the actual steel stress is n times the stress in the
equivalent wood; with the equivalent steel section, the actual wood
stress is \/n times the stress in the equivalent steel.
A similar procedure is used with beams composed of other combi-
nations of materials. For example, consider the experimental section
shown in Fig. 10-2a consisting of an aluminum core to which plates of
steel and bronze are securely attached. By using the ratio of the moduli

of elasticity of the steel and bronze to that of the aluminum, the

composite section is transformed into an equivalent section of
aluminum (Fig. 10-2b) to which the flexure formula may be applied. In
all such cases the neutral axis passes through the centroid of the
equivalent section, and the moment of inertia used is the value com-
puted about the centroidal axis of the transformed section.

Steel

Aluminum

Bronze

(a) Composite section (b)

I— f?*—
Equivalent aluminum section

Figure 10-2.
1 0-2 Beams of Different Materials 419

ILLUSTRATIVE PROBLEM

1001. A timber beam 150 mm by 300 mm reinforced, on the

is

bottom only, with a steel strip 75 mm wide by 10 mm thick. Determine

the maximum resisting moment if the allowable stresses are as < 120
MPa and aw < 8 MPa. Assume Es /Ew = n = 20.

Solution: Although only rarely is steel used to reinforce one side of a

timber beam, this problem illustrates many of the concepts encountered
later in reinforced concrete beams. The first of these involves the
location of the neutral axis. Since the neutral axis coincides with
the centroidal axis of the equivalent section shown in Fig. 10-3, the
moments of area about an axis through the base gives

[Ay - "Lay]

)]^ =
3 3 3 3
[(45 X 10 ) -I- (15 x lO (45 x 10 )(160) + (15 X 10 )(5)

y — 121 mm
First finding the moment of inertia about an axis through the top
of the flange and then using the transfer formula, we compute the
moment of inertia about the neutral axis:

150(300)^ L^O)'
3 3
[1=1- Ad 2 ]

INA = (1350 x 10
6
)
- (60 x lO^lll) 2 = 611 X 10
6
mm4

150 mm 150 mm

189 mm
300 mm

y = 121 mm
TZZZZ 10 mm CZZZ "7771 I

-J C~f
75 mm
(a) Original section (b) Equivalent wood section

Figure 10-3.
420 10 Reinforced Beams

The resisting moment in terms of the maximum wood stress is

M ( 8 xl^ll X |Qi =259Wm

y 189 x 10" 3

In the wood equivalent of the steel, the maximum stress is

whence the resisting moment that will not exceed the permissible steel
stress is

6 6
(6 X 10 )(611 X IP"
M r
10~ 3
)
= 30.3 kNm
y 121 X
The smaller resisting moment (i.e., M
w = 25.9 kN-m) is the safe
resisting moment. In this case there is an excess of steel; hence the beam
may be said to be over-reinforced. (The reader will find it instructive to
solve this problem using an equivalent steel section.)

PROBLEMS
1002. A timber beam is reinforced with steel plates rigidly
attached at the top and bottom as shown in Fig. P-1002. By what
amount is the moment increased by the reinforcement if n = 15 and the
allowable stresses in the wood and steel are 8 MPa and 120 MPa,
respectively? Ans. 52.2 kN m •

pmm n {
V/////////1 ,_ lQ
I. mm

300 mm

Ezzzzzzzzn 10 mm
L_150_J
250 mm"
Figures P-1002, P-1003, P-1004.

1003. A simply supported beam 4 m

long has the cross section
shown in Fig. P-1002. It carries a uniformly distributed load of 20
kN/m over the middle half of the span. If n = 15, compute the
maximum stresses in the wood and the steel.
1 0-2 Beams of Different Materials 421

Repeat Problem 1002 assuming that the reinforcement

1004.
consists of aluminum plates for which the allowable stress is 80
MN/m 2
. Use n = 5.

1005. A timber beam 150 mm by 250 mm is reinforced at the

bottom only by a steel plate, as shown
P-1005. Determine the
in Fig.
concentrated load that can be applied at the center of a simply sup-
ported span 6 m
long if n = 20, as < 120 MPa, and aw < 8 MPa. Show
that the neutral axis is 170.2 mm
below the top and that 7NA = 416 X
10
6
mm4 . Arts. P- 13.1 kN

250 mm

10 mm
mm
Figures P-1005, P-1006, P-1007.

1006. Determine the width b of the 10-mm steel plate fastened to

the bottom of the beam in Problem 1005 that will simultaneously stress
the wood and the steel to their permissible limits of 8 MN/m and 120
2

MN/m2 , respectively.
Ans. b = 42.5 mm
1007. A uniformly distributed load of 4 kN/m (including the
weight of the beam) is simply supported on a 6-m span. The cross
section of the beam is as described in Problem 1005. If n = 20, de-
termine the maximum stresses produced in the wood and the steel.
1008. A timber beam 150 mm wide by 250 mm deep to be is

reinforced at the top and bottom by steel plates 10 mm thick. How wide
should the steel plates be if the beam is to resist a moment of 50
kN-m? Assume that n = 15 and the allowable stresses in the wood and
steel are 8 MPa and 1 10 MPa, respectively. Ans. 6 = 143 mm
1009. beam 150 mm wide by 200 mm deep is to be
A timber
reinforced at the topand bottom by aluminum plates 6 mm thick.
Determine the width of the aluminum plates if the beam is to resist a
moment of 16 kN-m. Assume n = 5 and take the allowable stresses as
8 MN/m and 70 MN/m in the wood and aluminum, respectively.
2 2
422 10 Reinforced Beams

1010. AC250 X 30 steel channels are securely bolted to a

pair of
wood beam 200 mm
by 254 mm, as shown in Fig. P-1010. (From Table
B 2 in Appendix B, the depth of the channel is also 254 mm.) If
bending occurs about the axis 1-1, determine the safe resisting moment
if the allowable stresses are os = 120 MPa and aw = 8 MPa. Assume

n - 20. Arts. A/ = 74.7kN-m

1011. In Problem 1010, determine the safe resisting moment if

bending occurs about axis 2-2.

|2 ^C250 X 30

254 mm

Figures P-1010 and P-1011.

1012. An aluminum beam having the properties of a W200 X 46

section is reinforced by bolting 6-mm steel plates 203 mm wide to the
flanges, as shown in Fig. P-1012. The allowable stresses in the steel and
aluminum are 140 MPa and 100 MPa, respectively, and the ratio
EJ a
= 3. Determine (a) the percentage change in the strength of the
original unreinforced aluminum section and (b) the percentage change
in the flexural rigidity EI. Arts, (a) +21.4%; (b) + 175%

203 mm
It mm
Eg 3 6

Eg 6 mm
Figure P-1012.

1013. A solid steel beam 50 mm in diameter protected against

is

corrosion by a shell of aluminum 6 mm thick firmly bonded to it.

Compute the maximum moment the composite section can resist if

o, < 120 MPa and oa < 100 MPa. Assume EJ a

= 3.
10-3 Shearing Stress and Deflection in Composite Beams 423

1014. A rectangular section 150 mm wide by 250 mm deep resists

a bending moment of 140 kN-m. The material of the beam is noniso-

tropic with the tensile modulus of elasticity twice the compressive
modulus. Determine the maximum tensile and compressive stresses in
the section. Ans. a, = 108 MPa; ac = 76.5 MPa
1015. Solve Problem 1014 if the compressive modulus of elastic-
ity is 1.5 times the tensile value.

1016. An experimental beam is composed of the three materials

shown P-1016. The materials are firmly fastened together so that
in Fig.
there is no relative movement between them. Determine the safe resist-
ing moment if as < 120
2
MN/m
aa < 80 ,
2
ow < 10 MN/m 2
, MN/m ,

and Es = 200 GN/m

2
Ea - 70
, GN/m
2
Ew = 10 GN/m 2
, .

Ans. M = 33.8 kN-m

1017. In a section similar to that in Fig. 10-2a, on page 418, the
width b is mm for each material. The vertical dimensions are 20 mm
140
for steel, 150 mm for aluminum, and 50 mm for bronze. Assuming that
bonded together, determine the maximum stress
the materials are firmly
ineach material when the section is resisting a bending moment of 70
kN-m if Es - 200 GPa, Ea = 70 GPa y and Eb = 80 GPa.

80 mm

Steel 20 mm

Wood 150 mm

Aluminum 50 mm

Figure P-1016.

10-3 SHEARING STRESS AND DEFLECTION IN

COMPOSITE BEAMS

The formula for horizontal shearing stess [Eq. (5-4), page 189] devel-
oped for homogeneous beams applies equally well to the equivalent
section of a composite beam because its derivation was based on the
difference in normal forces between two adjacent sections. Since the
forces on the original composite section and on the equivalent section
are the same (see page 417), Eq. (5-4) is valid for either section.
424 10 Reinforced Beams

Deflections in composite beams can also be computed, as in

homogeneous beams, by using the flexural rigidity EI of the equivalent
section. This is true because the deflection is the result of the changes in
length of the fibers in the beam, and one of the basic principles of
composite beams is that the strains of fibers are identical at correspond-
ing points in the original and equivalent beams.

PROBLEMS
1018. Compute the allowable total vertical shear for a beam
having the same cross section as in Problem 1005, if n = 20 and the
maximum shear stress is 800 kN/m 2 . Arts. V= 23.0 kN
1019. In the beam section in Problem 1010, assume that the
channels are bolted to the wood by two rows
bolts spaced 300
of 20-mm
mm apart and located 75 mm
above and below axis 1-1. Assuming
n = 20, compute the shearing stress in the bolts caused by a central load
of 80 kN applied to a simply supported span 3 m long, if bending takes
place (a) about axis 1-1 and (b) about axis 2-2.
Ans. (b) t = 74.8 MPa
1020. The beam in Problem 1002 carries a uniformly distributed
kN/m
load of 30 on a simply supported span 5 m long. If Es = 200
GN/m 2 and Ew = 10 GN/m compute the midspan deflection.
2
,

Ans. 8 = 12.2 mm
1021. Problem 1016, determine the shear flow developed
In
between the and wood and between the wood and aluminum.
steel
Express the results as a function of the vertical shear V measured in
newtons. Ans. 5.09 V N/m; 5. 12 V N/m

10-4 REINFORCED CONCRETE BEAMS

Concrete is an excellent building material because it is cheap and

fireproofand does not rust or rot. It has about the same strength in
compression as soft wood, but its tensile strength is practically zero. For
this reason, the tensile side beams is reinforced with steel
of concrete
bars. Ideally, these steel bars
should follow the tensile stress trajectories,
but practically they are placed in one layer on the tensile side. For-
tunately there is a natural adhesion or bond between concrete and steel;
hence no slipping occurs between them during bending,* and the

•Sufficient bond is developed in long beams to permit the steel bars to be

laid straight, but in short beams the ends are usually bent over to anchor the steel
more securely in the concrete.
 E

10-4 Reinforced Concrete Beams 425

principles developed in the preceding article can be used here. For-

tunately also, both concrete and steel have about the same coefficient of
temperature expansion.
It is usually assumed that the concrete carries no tensile stress, the
tensile side of the concrete beam serving merely to position the steel
which carries the entire tensile load. The steel isassumed to be uni-
formly stressed (since it is all at nearly the same distance from the
neutral axis), so the line of action of the tensile force acts through the
center of the reinforcing steel. The compressive stress in the concrete is
assumed to vary linearly from the neutral axis, which places the re-
sultant compressive force at the centroid of the compressive stress
triangle (see Fig. 10-4a). The value of E for concrete is usually assumed
to be between -^ and \ that for steel, depending upon the quality of the
concrete.
The portion of a reinforced concrete beam in Fig. 10-4a has the
cross section shown in Fig. 10-4b. The equivalent section in terms of
concrete, shown in Fig. 10-4c, is obtained by using Eq. (10-1) to
transform the steel area A s into the equivalent concrete area nA s As .

before, n is the ratio of the moduli of elasticity, i.e., n = EJ c The .

shaded portions of Fig. 10-4c indicate the areas that are effective in
resisting bending. The distance from the top of the beam to the center
of the reinforcing steel is conventionally denoted by the symbol d, and
the location of the neutral axis is specified as being a fractional part k of
this distance.
If the quantities b, d, A s and
, n are known, the neutral axis (the
centroidal axis of the shaded areas in Fig. 10-4c) is located by applying
the principle that the moment of area above the neutral axis equals the

Allowance for fireproofing;

usually about 50 mm
Figure 10-4. Reinforced concrete beam and equivalent section.
426 10 Reinforced Beams

moment of area below this axis:

(b kd)(") = nA,(d - kd) (10-2)

This is a quadratic equation in terms of the distance kd. Sometimes the

equation is solved directly for k, but it is more useful to determine kd.
The resultant compressive force C in the concrete acts at the
centroid of the stress triangle in Fig. 10-4a, and is therefore at a
distance \kd from the top of the beam.* The resisting couple, composed
of the equal compressive and tensile forces C and T, therefore has a
moment arm jd equal to

jd = d-\(kd) (10-3)

The notation for stresses in reinforced concrete differs from that

used elsewhere in this book, it being conventional to denote the stress in
the concrete by fc and the stress in the steel by fs The neutral
. axis
having been located, the moment of inertia of the equivalent section
may be computed and the flexure formula applied as in Art. 10-3.
However, more it is direct to compute the resisting moment from the
couple composed of the compressive force C and the tensile force T;
this is C(jd) or T(Jd).
value
According to this concept, the average compressive stress in the
concrete is \fc where fc is the maximum compressive stress. The
,

compressive force C in the concrete is the product of the average

compressive stress multiplied by b kd (the area under compression):

C = \fc (b kd) (10-4)

The resisting moment based on the maximum compressive stress is

therefore

M c
= C(jd) = \fc {b kd)(jd) (10-5)

The tensile force T in the steel is the product of the steel area As
multiplied by the steel stress^; hence the resisting moment in terms of
the steel is

H- T(Jd)=fs A s Ud) (10-6)

The safe bending moment is the lower of the two values M c
and M s
.

In solving problems, it is better to follow numerically the steps in

the derivations above rather than to use these equations directly. The
first step is which the moment arm of the
to locate the neutral axis, after
resisting couple is easily computed. The resisting moment is then de-
termined by means of the concept of force (C or T) times the moment
arm jd.
•This statement and the procedure which follows should be compared with
the discussion of a homogeneous rectangular section on pages 158 and 160.
10-4 Reinforced Concrete Beams 427

ILLUSTRATIVE PROBLEMS

1022. = 300 mm, d = 500 mm,

In a reinforced concrete beam, b
As = 1500 mm 2
—
Determine the maximum stresses in the concrete
, n 8.

and steel produced by a bending moment of 70 kN-m.

Solution: The equivalent section of the beam is shown in Fig. 10-5.

Since the total moment of area about the neutral axis is zero,

[Lay - 0] 300^^) = (12 X l(P)(5O0 - kd)

which reduces to
2
(kd) + SOkd - (40 x 10
3
) =
from which
kd - 164 mm = 0.164 m
The moment arm of the resisting moment is

[jd = d- \kd] jd = 500 - ^(164) = 445 mm = 0.445 m

In terms of the concrete, the resisting moment is

[
Af = ifc (b kd)(jd)] 70 x 10
3
= ±/c (0.300)(0.164)(0.445)

from which the maximum compressive stress is

fc = 6.39 MPa Ans.

In terms of the steel, the resisting moment is

[M = f AJd] s
70 X 10
3
= /,(1500 X 10" 6 )(0.445)

from which the steel stress is

fs = 105 MPa Ans.

500 mm
500 -kd

nA, = Sx 1500 = 12 X 10 3

Figure 10-5.
428 10 Reinforced Beams

1023. In a reinforced concrete beam, b = 250 mm, d == 400 mm,

As - 1000 mm 2
, and n = 8. If fc < 12 MPa
the allowable stresses are
and/, < 140 MPa, determine the maximum bending moment that may
be applied. Is the beam over- or under-reinforced?

Solution: The equivalent section of this beam is shown in Fig. 10-6.

Proceeding as in Illustrative Problem 1022, we begin by computing the
factors kd and jd:

[lay = 0] 250fc/(-y ) = (8 x 3
10 )(400 - kd)

whence
kd = 131 mm = 0.131 m
[jd = d - \kd] jd = 400 - j(131) = 356 mm = 0.356 m
To stress the concrete to its maximum limit will require a bending
moment
[Me -\fe (bkd)(jd)]
M c =|(12 X 6
10 )(0.250)(0. 13 1)(0.356)
= 70.0kNm
To stress the steel to its limit, the required bending moment is

[Ms = fs AJd] M s
= (140
6
X 10 )(1000 X 10" 6 )(0.356)
= 49.8 kN-m

The safe bending moment is therefore 49.8 kN-m. Since the steel
governs, we conclude there is not enough steel; hence the beam is
under-reinforced.

400 mm
400 -kd

nA s = SX 1000 = 8 X 10 3 mm :

Figure 10-6.
10-4 Reinforced Concrete Beams 429

PROBLEMS

1024. In a reinforced concrete beam, b = 200 mm, d = 400 mm,

As - 1400 mm 2
. Find the values of kd and jd if (a) n = 6 and (b)
n = 10. Ans. (b) kd = 177 mm;y</ = 341 mm
1025. = 250 mm, d = 450 mm,
In a reinforced concrete beam, b
and n = 10.The actual maximum stresses developed are fc = 6 MPa
and fs = 120 MPa. Determine the applied bending moment and the
steel area required. Ans. M
= 45.0 kN-m; As = 938 mm2
1026. Repeat Problem 1025 if d = 540 mm, all other data re-
maining unchanged.
1027. maximum stresses produced in the concrete
Determine the
and steel of beam by a bending moment of 70 kN-m if
a reinforced
b - 300 mm, d = 500 mm, A s = 1200 mm 2 and n = 8. ,

Ans. fc = 6.91 MN/m2 ;/, = 130MN/m2

1028. In a reinforced concrete beam, b = 500 mm, d — 750 mm,
As = 6000 mm2 and n = 10. What are the maximum stresses developed
,

in the concrete and the steel by a bending moment of 270 kN-m?

1029. The dimensions of a reinforced concrete beam are b = 300

mm, d = 450 mm, A s = 1400 mm2 and n = 8. If the allowable stresses
,

are fc < 12 MN/m and fs < 140 MN/m

2 2
determine the maximum ,

bending moment that may be applied. In what state of reinforcement is

the beam? Ans. M
= 78.4 kN-m; under-reinforced
1030. = 250 mm, d = 450 mm, A s —
In a reinforced beam, b
1400 mm 2
and n , Determine =
the safe uniformly distributed load
8.

that can be carried on a simply supported span 4 long if fc < 12 MPa m

and fs < 140 MPa. Assume 50 mm of concrete below the reinforcing
steel, and include the weight of the beam. The density of concrete is

2400 kg/m3 .

1031. 300 mm, d — 600 mm,

In a reinforced concrete beam, b =
and n =
9. If a maximum stress of 5 MPa is developed in the concrete

when resisting a moment of 80 kN-m, what stress is developed in the

steel? What area of reinforcing steel is required?
fs = 90 MPa; A s = 1670 mm
2
Ans.
1032. Solve Problem 1031 if the bending moment is 70 kN-m,
all other data remaining unchanged.

1033. Solve Problem 1027 by computing the moment of inertia

of the transformed section and then applying the flexure formula
according to the procedure in Art. 10-2. The distance of the equivalent
concrete area from the neutral axis of the transformed section can be
taken as its radius of gyration with respect to this axis.
430 10 Reinforced Beams

1034. Solve Problem 1029, using the procedure outlined in Prob-

lem 1033.

10-5 DESIGN OF REINFORCED CONCRETE BEAMS

In the preceding article the dimensions of the reinforced beam were

specified. This fixed the location of the neutral axis. Inasmuch as the
stresses vary directly with their distance from the neutral axis, the
applied bending moment may stress the concrete to its permissible limit
while leaving the steel understressed — a condition known as over-rein-
forcement. The opposite condition, under-reinforcement, may occur
when the steel reaches its permissible limit first, the concrete remaining
understressed. For maximum economy, both materials should reach
their limiting stresses simultaneously — a condition known as balanced-
stress reinforcement.
In designing a concrete beam with balanced-stress reinforcement,
therefore, we start with the assumption that the position of the neutral
axis is such that the maximum fc and the maximum stress jj n in the
equivalent concrete occur simultaneously; this is shown in the stress
diagram in Fig. 10-7. From this, by the proportional relations between
the triangles ABC and ADE, we obtain
kd fc
d
L + f<
n
or

fc
k = (10-7)

c
n

ir*-u
•*—/ |
kd

NA
B
/
7 \c\

\E V

1—4^k*-J
Figure 10-7. Stress distribution.
10-5 Design of Reinforced Concrete Beams 431

Having computed k in terms of the allowable stresses, we obtain

the value of j by canceling out the term d in Eq. (10-3):

/-I -I* (10-8)

Once the values of k and j are determined, Eq. (10-5) is used to

compute the quantity bd 2 The deeper the beam, the greater will be the
.

moment arm of the resisting couple and the smaller the force. A deep
beam therefore requires less concrete and steel than a shallow beam.
Because of practical limits to the depth, however, d is usually made
2
about 1.5 &. From this and the now computed value of bd the dimen- ,

sions b and d are found.

As the final step, the area of the reinforcing steel is computed
from Eq. (10-6) or, preferably, from the condition that C = T — AJS '
,

in which A s is now the only unknown. However, the reinforcing rods

generally available are not of such size as to equal precisely the steel
area required. As a consequence, balanced-stress reinforcement can be
only closely approximated.
For most well-designed rectangular beams, the values of k andy
k = § andy —\. If these values are used, dimensions
are very close to
may be rapidly estimated by substituting them in Eq. (10-5), thereby
giving*

bd- =
^Jc
(10-9)

Having thus found bd 2 and assigning values to b and d, we compute the

tensile force in the steel and the steel area from

T = ^- and As =j (10-10)

ILLUSTRATIVE PROBLEM

1035. Design a concrete beam with balanced-stress reinforce-

ment to resist a bending moment of 90 kN-m. The allowable stresses
are/c = 12 MPa,X - 140 MPa, and n - 8.

Solution: With balanced-stress reinforcement, the stresses in the con-

crete and the concrete equivalent of the steel have the values shown in
Fig. 10-8. From the proportional relations between the similar triangles
ABD and AEF, the value of k is found to be
kd 12
k = 0.407
d (140/8) + 12

The similarity of this result to the flexure formula for rectangular beams,
a - 6M/bh 2 makes
, it simple to remember.
432 10 Reinforced Beams

C- 1 fe bkd

F T = A s fs

Figure 10-8.

from which the value of j is

y . 1 - £* • 1 _ 1(0.407) = 0.864

In terms of the concrete, the resisting moment is C •

(jd), so

[Mc -C{jd) = {\fc bkd){jd)]

3 6 2
90 x 10 =1(12 x 10 )(&/ )(0.407)(0.864)

bd 1 = 0.0427 m = 3
42.7X 10
6
mm 3
(a)

Assuming that d = 1.56, we now obtain from Eq. (a), b = 267 mm

and d = 400 mm.
The area of reinforcing steel is now the only unknown. Since the
compressive force C in the concrete equals the tensile force T in the
steel, we obtain,

[\fc bkd=AJs ]
6 6
1(12 x 10 )(0.267)(0.407)(0.400) - ,4,(140 X 10 )

whence

A, = 1.86 x 10" 3 m =2
1860 mm 2

Usually the available stock sizes of reinforcing steel do not produce

exactly this area of steel, so the final design only closely approximates
balanced-stress reinforcement.

PROBLEMS

1036. A reinforced concrete beam is designed to reach fc = 12

MPa and fs = 140 MPa simultaneously. If n = 8 and d = 450 mm,
10-6 Tee Beams of Reinforced Concrete 433

compute the moment arm of the resisting couple.

Arts, jd = 389 mm
1037. In a reinforced beam, d = 600 and n = 9. Find the mm
dimensions b and A s that will resist a bending moment of 80 kN-m
with balanced-stress reinforcement, if fc = 9 X 106 N/m 2 and/5 = 140
X 10
6
N/m2 .

1038. In a reinforced beam, b = 250 mm, d = 450 mm, and

n = 9; the allowable stresses are fc = 10 MPa and ^ = 140 MPa.
Determine /* 5 for balanced-stress design and the safe resisting moment.
1039. Design a reinforced concrete beam with balanced-stress
reinforcement that will resist a bending moment of 140 kN-m, assum-

ing d- 1.56, fc = 12 MPa,/, = 160 MPa, and n - 8.

,4/w. 6 = 316 mm; A s = 2110
2
mm
1040. Solve Problem 1039 if b = \d.
1041. A beam 6
simply supported m
long is designed to carry a
concentrated load of 80 kN
midspan. Compute b and A s for a depth
at
</ = 600 mm
using balanced-stress reinforcement with fc = 8
2
MN/m ,

^ = 120 2
MN/m
and « = 10. Allow 50
, mm
of concrete below the steel
and include the weight of the beam, assuming the density of concrete is
2400 kg/m 3 (Hint: Assume an initial weight per meter for the beam
.

and check this assumption after you have found dimensions.)

Ans. b = 278 mm; A s = 2220
2
mm
1042. A reinforced concrete beam 6 m long and perfectly re-
strained at the ends is to carry a live load of 20 kN/m in addition to its

weight. Assuming d = 600 mm, design a beam with balanced-stress

reinforcement, using fc = 6 MPa, fs — 120 MPa, and n = 10. Allow 50
mm of concrete below the reinforcing steel. The density of concrete is
2400 kg/m 3 (Use the hint in Problem 1041.)
.

Ans. b = 219 mm; A s = 1100 mm 2

Design a reinforced concrete beam with balanced-stress

1043.
reinforcement to carry a live load of 80 kN/m over a simple span 4 m
long. Use/C = 12
2
MN/m
and/, = 140
2
MN/m
and n = 8. Allow 50 ,

mm of concrete below the steel and include the weight of the beam,
assuming the density of concrete is 2400 kg/m Assume b = 200 mm.
3
.

(Use the hint in Problem 1041.)

10-6 TEE BEAMS OF REINFORCED CONCRETE

The method used for rectangular reinforced concrete beams becomes

quite involved when applied to T beams. Because of the flange of the T,
the centroid of the compressive area is no longer \ kd from the neutral
434 10 Reinforced Beams

axis, nor is the line of action of the resultant compressive force \kd
from the top of the beam. As a consequence, it is cumbersome to use
the basic procedure described in Art. 10-4, although textbooks on
reinforced concrete do develop equations in terms of k and j. It is
preferable for the beginner to apply the flexure formula directly to an
equivalent section, as indicated in Art. 10-2 and as illustrated in the
following problem.

ILLUSTRATIVE PROBLEM

1044. The T beam in Fig. 10-9 is reinforced with 2400 mm 2 of

steel. Assuming n = 8,/c < 12 MPa,/5 < 140 MPa, determine the maxi-
mum safe resisting moment.

Solution: Denoting the distance from the bottom of the flange by v,

we compute the moments of area about the neutral axis:

[2ay = 0]

(750 X 100)(v + 50) + (400j>)(^) - (19.2 X 3

10 )(500 - v) =

This reduces to

y
2
-I- 471>> - (29.25 X 10
3
) =
from which

y — 55.6 mm
The moment of inertia about the neutral axis is computed by
resolving the compressive area into a rectangle 750 mm by 155.6 mm

750 mm
1
100 mm
t»^m^l
^^ i 155.6 mm
NA
y = 55.6 mm
500 mm 500 -y = 444.4 mm

W77Z s y s s s / s / / / / 77777
;yyyyyyy7777 VVVV/

U-400mm-^l ^nA s
= 8(2400)
= 19.2 X 10 3 mm 2
Figure 10-9.
1 0-6 Tee Beams of Reinforced Concrete 435

reduced by a rectangle 350 mm by 55.6 mm. Thus we have

/ = ™op£ - *°f£ + (19 2 . x mmAf
= 4714 X 10 mm4 = 4714 x 10 -6 m4
6

The distance 444.4 mm may be taken as the radius of gyration of the

concrete equivalent of the steel.

In terms of the permissible concrete stress, the flexure formula

gives

M c
= ^
(12 X 6
10 )(4714
^ X 10" 6 ) '- = „ kNm
XT
364A ,

y 155.6 x 10" 3

In terms of the concrete equivalent of the steel, the permissible

limit is fs /n = 140/8 = 17.5 MPa, and the flexure formula gives
6
ol _ (17.5 X 10 )(4714 X 10~ 6 )
M= = 186 kN-m
444.4 X 10" 3
This, being smaller than M c,
is the maximum safe resisting moment.

PROBLEMS

1045. In the reinforced T beam in Fig. P-1045, b = 500 mm, x

h x
= 150 mm, b - 250 mm, h = 500 mm, A s - 3000 mm 2 and n = 10. ,

Compute the maximum stresses produced in the concrete and the steel
by a bending moment of 140 kN-m.
Ans. /c = 4.45 MPa; X = 80.5 MPa

Figures P-1045, P-1046, P-1047.

1046. The dimensions of the reinforced concrete T beam in Fig.

P-1045 are b = 750 mm, h = 100 mm, b = 300 mm, and h = 450 mm.
x x

If n — 8 and A s = 3300 mm
2
determine the maximum bending moment
,

that may be applied without exceeding fc = 12 MN/m and fs = 140

2

MN/m 2
. Ans. M = 231 kN-m
436 10 Reinforced Beams

In the reinforced concrete T beam in Fig. P-1045, b =

1047. x

900 mm, /?, = 80 mm, b = 300 mm, h = 520 mm, and n - 9. Find A s
and the maximum resisting moment for a balanced-stress design, using
fc - 9 MPa and/, = 160 MPa.

10-7 SHEARING STRESS AND BOND STRESS

In the bending of reinforced concrete beams, the steel is prevented from

sliding by the grip of the enveloping concrete. The stress developed by
dividing this gripping force by the surface area of the reinforcing bars
per linear meter is the bond stress. The bond stress is analogous to the
shearing stress in a homogeneous beam; it may be computed by
applying Eq. (5-4) (page 189) to the equivalent section of concrete
shown in Fig. 10-10. Thus we obtain
V
r = j^(nA s )(d kd) (a)

where b' is the effective width of the steel bars, equivalent to the sum of
the perimeters of the steel bars, usually expressed by 2#.
The moment of inertia of the transformed section is found from
the flexure formula. Thus, at the concrete equivalent of the steel, the
stress is fs /n y so

= My fs = M(d - kd)
a (b)
I n~ I

But M= T(Jd) = A s fs jd; hence Eq. (b) reduces to

/ = nA (d s
- kd)Ud) (c)

Since nA (d - kd) =
s
(bkd)(kd/2), as we saw in Eq. 10-2, the moment
of inertia can also be expressed by

I=\b(kd)\jd) (d)

\ i

nA 9
it f t { < i ri zz 1
(b) Shear stress
(a) Equivalent section distribution

Figure 10-10.
10-7 Shearing Stress and Bond Stress 437

Substituting in Eq. (a) the value of / from Eq. (c) gives

V(nA s )(d - kd)

[nA s (d - kd)(jd)]b'

from which the bond stress is given by

^ih = iko < 10

- n)

The shearing stress at the neutral axis is similarly

~
T = Tt A y
'
Max.r = j-{nA s )(d- kd)

By using the value of / defined in Eq. (c), we reduce this to

Max. 7 = ^-
ja b
(10-12)

Further examination of Eq. (5-4) shows that the shearing stress in

the compressive portion of the concrete varies parabolically as in a
homogeneous rectangular section, whereas it remains constant below
the neutral axis, as shown in Fig. 10- 10b. Comparison of Eqs. (10-11)
and (10-12) shows that the bond stress becomes larger than the shear-
ing stress on the neutral surface if 2 o is smaller than the width b of the
beam. To increase 2tf and keep A s constant, more reinforcing bars of
smaller diameter may be used.

PROBLEMS

1048. The reinforced beam in Problem 1028 is subjected to a

vertical shear V= 120 kN. Calculate the maximum shearing stress and
bond stress if the reinforcement consists of six bars 20 mm square.
Ans. 373 kPa; 389 kPa
1049.Determine the vertical shear that can be sustained by the
beam Problem 1027 if the reinforcing consists of four bars 10 mm
in
2
square. Assume that the allowable shearing stress is 350 kN/rh and the
allowable bond stress is 550 kN/m
2
Ans.
. V = 39.6 kN

SUMMARY

Nonhomogeneous beams of two materials, for example, wood

and steel, may be transformed into equivalent sections of one material
by using the relation
Aw = nA, (10-1)

where n is the ratio of the moduli of elasticity of the transformed

438 10 Reinforced Beams

material to that of the equivalent material. The flexure formula may

then be applied directly to the transformed section as shown in Arts.
10-2 and 10-6.
The basic procedure in Art. 10-4 is usually used with reinforced
concrete beams. The principles outlined in deriving the following equa-
tions should be followed numerically, rather than slavishly applying the
equations themselves:

(bkd)(™) = nA,(d- kd) (10-2)

jd= d-\(kd) (10-3)

The stresses are found from the following equations. If the allow-
able stresses are specified, the safe resisting moment of the beam is the
smaller of the two values:

K - C(jd) = \f c (b kd)(jd) (10-5)

M, - Hjd) = fs A,(jd) (10-6)

In designing a concrete beam with balanced-stress reinforcement,

the dimensions of the section are unknown, so the above equations
cannot be applied. Nevertheless, a stress diagram can be drawn in terms
of the specified maximum stresses, as in Fig. 10-7 in Art. 10-5, whence
the position of the neutral axis is defined in terms of k as follows:

* = -/*— (10-7)

Equation (10-7) is based on the assumption of balanced-stress

reinforcement, i.e., that both concrete and steel reach their allowable
stresses simultaneously. This relation between the stresses does not exist
when the dimensions of a beam are specified; hence Eq. (10-7) cannot
be used as a shortcut substitute for Eq. (10-2) in locating the neutral
axis.

The maximum shearing stress in reinforced rectangular concrete

beams is given by

Max. T = -r£r (10-12)

ja d

and the bond stress by

V
(10-11)
jd?,o
where So is the sum of the perimeters of the reinforcing steel bars.
 Columns

11-1 INTRODUCTION

A column is a compression member that compared to its

is so slender
length that under gradually increasing loads by buckling at loads
it fails

considerably less than those required to cause failure by crushing. In

this respect it differs from a short compression member, which, even if
eccentrically loaded, undergoes negligible lateral deflection. Although
there is no sharp line of demarcation between short compression mem-
bers and columns, a compression member is generally considered to be
a column when its unsupported length is more than 10 times its least
lateral dimension.
Columns are usually subdivided into two groups: long and
%
inter-
mediate, sometimes the short compression block is considered to be a
third group. The distinction between the three is determined by their
behavior. Long columns fail by buckling or excessive lateral bending;
intermediate columns, by a combination of crushing and buckling;
short compression blocks, by crushing. We shall now examine these
differences in detail.
An ideal column is assumed to be a homogeneous member of
constant cross section that is initially straight and is subjected to axial
compressive loads. However, actual columns always have small imper-
fections of material and fabrication, as well as unavoidable accidental

439
440 11 Columns

eccentricities which produce the effect shown, greatly ex-

of load,
aggerated, in Fig. 1 The initial crookedness of the column, together
1-1.
with the placement of the load, causes an indeterminate eccentricity e
with respect to the centroid of a typical section m-n. The loading on
on an eccentrically loaded short strut (Art.
this section is similar to that

9-3, page 369), and the resultant stress is due to a combination of a

direct compressive stress and a flexure stress.
If the eccentricity is small and the member short, the lateral

deflection is negligible and the flexural stress is insignificant compared

with the direct compressive stress. A long member, however, is quite
flexible, because deflection is proportional to the cube of the length;

hence a relatively low value of P may cause a large flexural stress

accompanied by a negligible direct compressive stress. Thus, at the two
extremes, a short column carries principally direct compressive stress,
and a long column is subjected primarily to flexural stress. As the length
of a column increases, the importance of the direct compressive stress
decreases, and that of the flexural stress increases. Unfortunately, in the
intermediate column range it has not been possible to determine the
rates of change in these stresses or the proportion of each stress that
constitutes the resultant stress. It is this indeterminateness that gives rise
to the many formulas for intermediate columns; these are discussed in
Art. 11-5.
For the present, we have deliberately refrained from establishing
any between long and intermediate columns,
criterion for the difference
except the fact that the long column is subjected principally to flexural
stress and the intermediate column to a combination of direct and

Unavoidable or
accidental
eccentricity

Actual center line due

to initial crookedness
(greatly exaggerated)

e- eccentricity
of Pat a
typical section m-n

Mathematically
straight center line

Figure 11-1. Factors contributing to eccentricity of loads in columns.

 H*

11-2 Critical Load 441

flexural stress. The distinction between them in terms of actual length

can be discussed intelligently only after we have studied the action in a
long column.

11-2 CRITICAL LOAD

A long beam is mounted vertically and hinged at the ends so that it is

free to bend in any direction. A central horizontal load H is applied to

cause bending in its most limber plane, as shown in Fig. 11 -2a. Since
flexural stress is proportional to deflection, there will be no change in
stress if an axial load P is added at each end, as in Fig. 1 1 —2b, H being
simultaneously decreased as P increases so that the midspan deflection
8 is unaltered. The midspan bending moment is then

M
which becomes
-m + P8

M = (P CT )8

when H has been reduced to zero. Here, as shown in Fig. 11 -2c, PCT is

the critical load required to maintain the column in its deflected

position without any side thrust. Any increase in P beyond this value
increases the deflection 6\ thereby increasing M, thence 8, etc., until the
column buckles or fails. On the other hand, if P is decreased slightly
below this critical value, the deflection is decreased, thereby decreasing
the bending moment, thence the deflection, etc., and the column

P [Pa-

i ,2 — *f
r t

/ /
/
l

I 5 F
/
6 —H
i
-*- 6 *"-H -

\\ t
<

^2H
'

— ^2 \
P Per

(a) (b) (c)

Figure 11-2. Beam and column sustaining equal deflections.

442 11 Columns

straightens out. A critical load, therefore, can be interpreted as the

maximum load to which a column can be subjected and still
axial
remain straight, although in such an unstable condition that a slight
sideways thrust will cause it to bow out, as in Fig. 11 -2c. The next
article discusses the computation of this critical load.

11-3 LONG COLUMNS BY EULER'S FORMULA

A theoretical analysis of the critical load for long columns was made by
the great Swiss mathematician Leonhard Euler
in 1757. His analysis is
/
based on the differential equation of the elastic curve ElidPy dx )
2
'
=
M. As we know now, such an analysis is valid only up to the stress at
the proportional limit. In Euler's time, neither the concept of stress nor
the limiting stress at the proportional limit had been formulated; hence
he did not emphasize the concept of an upper limit to the critical load
P. This upper limit is considered in Art. 11-4.
Figure 11-3 shows the center line of a column in equilibrium
under the action of its critical load P. The column is assumed to have
hinged ends (sometimes called round, pivoted, or pinned) restrained
against lateral movement. The maximum deflection 8 is so small that
there isno appreciable difference between the original length of the
column and its projection on a vertical plane. Under these conditions
the slope dy/dx is so small that we may apply the approximate
differential equation of the elastic curve of a beam, viz.,

EI&-M-P(-y)= -Py (a)

dx*

O "
I J [
t /
x 1
1 /
•

]i
x
\
J

'

\
6 I r it

1
\ 7r

!
' \ f

1 1

P
F "Igure 11- 3.
 —
11-3 Long Columns by Euler's Formula 443

M is negative because in Fig. 11-3 the deflection v is negative. If the

column should deflect in the opposite direction so that y is positive, M
would still be negative because of the sign convention adopted for
bending moment in Art. 4-2.
Equation (a) cannot be integrated directly, as was done in Art.
6-2, because here M
is not a function of x. However, we present two

methods of solving it. Students who are familiar with dynamics will find
Eq. (a) similar to the equation of a simple vibrating body:

m —1- — —kx
2
d x
dt

for which the general solution is

* =c sin +c > c°s

>

('V^) ('Vin
Hence, by analogy, the solution of Eq. (a) can be written at once as

y =

Substituting v
C, sinjx

= at
y-J^
x =
)
+ C2 cos|x

in Eq. (b) gives

^C2 —
j

0. If we apply y =
(

at
b)

x — L, we obtain

0=C,sin(L^")
This is satisfied if C = 1
(in which case there is no bending of the
column), or by

L-yJ^j =nir (n = 0,1,2,3,...)

from which

P= n2 — (c)
Li

Students not familiar with dynamics can solve Eq. (a) by rewriting
it in the form

S/ A(±.\ m
£l£)=-*
After multiplying this by 2 dy to obtain perfect differentials, we get, by
integration,
444 1 1 Columns

Since, according to Fig. 1 1-3,>> = 8 when dy/dx = 0, substitution

in Eq. (d) gives C, = P8 2
y
whence Eq. (d) becomes

or

dx
Separating the varia bles, we obtain

dy

which is integrated to yield

To evaluate C2 we use the relationship y — at x = 0; hence

C2 = 0, so

-- \
This indicates that the column has the shape of a sine curve. Setting
v at x = L in Eq. (e) gives

sin L
( V^)=°
or

from which
v^ = nm (h - 0, 1, 2, 3, ... )

/> = «^ L
(/)

This agrees with the value found previously in Eq. (c).

The value n = is meaningless because then the load P is zero.

For the other values of n, the column bends into the shapes shown in
Fig. 11-4. Of these, the most important is (a); the others occur with
larger loads and are possible only if the column is braced at the middle
or third points respectively.* The critical load for a hinged-ended

•Bracing Fig. 1 1 -4b at the midpoint reduces it to the shape in Fig. 1 1 -4a with
an equivalent length \ L. Substituting ^L in place of L in Eq. (11-1) increases the
critical load 4 times, which checks Eq. (f) for n - 2.
11-3 Long Columns by Euler's Formula 445

EIn P2 =4P X
P3 =9P 1

(a) n=l (b)n = 2; (c)n = 3;

midpoint third point
bracing bracing

Figure 11-4. Effects of n on loads.

column is therefore

/> = EW (11-1)

The critical load for columns with other end conditions can be
expressed in terms of the critical load for a hinged column, which is

taken as the fundamental case. Thus, from symmetry, the column with
fixed ends in Fig. 1 1 -5a has inflection points at the quarter points of its

unsupported length. Since the bending moment is zero at a point of

inflection, the free-body diagrams show that the middle half of the
fixed-ended column is equivalent to a hinged column having an effec-
tive length Le = L/2. If this is substituted in Eq. (11-1), the critical
load on a fixed-ended column is

EIm A
EItt' EIv1
P= = 4 (11-2)
^n/ ^ %&&
\ *=±±&J>

(i)
This is four times the strength of the column if its ends were hinged.
Figure also provides a means of determining the load
11 -5a
capacity of a column built in at one end and free at the other the —
flagpole type of column. The critical loads on it (Fig. 1 1— 5b) and on the
fixed-ended column (Fig. 11 -5a) are equal, provided the fixed-ended
column is four times as long as the flagpole type. In other words, by
substituting in Eq. (11-2) an equivalent length Le that is four times its
 11 Columns

(c)

p iP
(a) (b)

Figure 11-5. Built-in column and free-body diagrams.

actual length, the critical load on a flagpole column is given by

4EItt
2
4EItt
:
1 EIv 1 PRtJZ- X F iX-fcA?
P= (11-3)
(4L)
:
4 L2
This load is one-quarter the critical load on a hinged column of the
same length.
One other type of one end and built in at the
column is hinged at
other, as in Fig. 1 1 -6. For it, the point ofcan be shown to be
inflection
nearly 0.7 L from the hinged end. Hence substituting an effective length

L e =0.7L

Figure 11-6. Column hinged at one end and built in at the other.
11-4 Limitations of Euler's Formula 447

TV = NUMBER OF
TIMES STRENGTH OF Le =
END CONDITION HINGED COLUMNS EFFECTIVE LENGTH
Fixed ends 4 \L
One end fixed, the other hinged 2 0.7 L
Both ends hinged 1 L
One end fixed, the other free \ 2L

Le - 0.7 L in Eq. (11-1) gives

2
x^_ _=
EhP_ Elm 2
x^v, Elm
p= = f^|_
~m 2 L 2 (yery neady) (n _ 4)
L] (0.7 Lf

The effect of end conditions on the critical load can therefore be

expressed in terms of the critical load for the fundamental type of
hinged column of the same length. All we need do is apply Eq. (11-1)
multiplied by a factor N
which varies with end conditions as
summarized in the above table, or, preferably, replace L in Eq. (11-1)
by the tabulated value of an effective or modified length Le that is, ;

Elm 2 Elm 2
L, Li

11-4 LIMITATIONS OF EULER'S FORMULA

A column always tends to buckle in its most limber direction. For this "\
reason, and since flexural resistance varies with moment of inertia, the i
value of / in the column formulas is always the least moment of inertia \
of the cross section. Any tendency to buckle therefore occurs about the '
least axis of inertia of the cross section.
Euler's formula also shows that the critical load which causes
buckling depends not upon the strength of the material but only upon
its dimensions and modulus of elasticity. For this reason, two dimen-^
one of high-strength steel and the other /
sionally identical slender struts,
of ordinary structural steel, will buckle under the same critical load
because, although their strengths are different, they have the same
modulus of elasticity. Good design also requires that a section have as
large a moment of inertia as possible. Hence, for a given area, the
material should be distributed as far as possible from the centroid and
in such a way that the moments of inertia about the principal axes are
equal or as nearly equal as possible.
In order for Euler's formula to be applicable, the stress accompa-
nying the bending which occurs during buckling must not exceed the
 11 Columns

proportional limit. This stress may be found by replacing in Euler's

formula the moment of inertia / by its equivalent Ar 2 where A
, is the
cross-sectional area and r the least radius of gyration.* This being done
for the fundamental case of a hinged column, Eq. (11-1) becomes

£
A
- -AL (H-5)
(L/r) 1

For other end conditions, substitute in this equation the equivalent

column from the table
length of a hinged in the preceding article.

Here P/A is the average stress in the column when carrying its

critical load. This stress is often called the critical stress. Its limiting

value is the stress at the proportional limit. The ratio L/r is called the
slenderness ratio of the column. Since an axially loaded column tends to
buckle about the axis of least moment of inertia, the least radius of
gyration should be used to determine the slenderness ratio.
Conventionally, we define long columns as those for which Euler's
formula applies. The limiting slenderness ratio that fixes the lower limit
for Euler's formula is easily found by substituting in Eq. (11-5) the
known values of the proportional limit and the modulus of elasticity of

the specified material. This limiting ratio varies with different materials
and even with different grades of the same material.
As an example, for steel that has a proportional limit of 200 MPa
and for which E = 200 GPa, the limiting slenderness ratio is

er-
(
200Xl0
200 X 10
y ~10
6
000 or ^100
r

Below this value, as shown -7 by the dashed portion of Euler's

in Fig. 1 1

curve, the Euler unit load exceeds the proportional limit. Hence for
L/r < 100, Euler's formula is not valid, and the proportional limit is
taken as the critical stress. The curve also shows that the critical or
allowable stress on a column decreases rapidly as the slenderness ratio
increases; hence it is good design to keep the slenderness ratio as small
as possible.
Finally, remember that Euler's formulas determine critical loads,
not working loads. therefore necessary to divide the right side of
It is

each formula by a suitable factor of safety usually 2 to 3, depending —

on the material — in order to obtain practical allowable values.

•Here, we are using r to denote radius of gyration to conform to AISC

notation Be careful not to confuse this r with the rthat is frequently used to denote
the radius of a circle.
11-4 Limitations of Euler's Formula 449

\
\
\
\
\
\
\ Euler's curve
200 MPa
V, P_ En 2
(P.L.)
X A (L/r) 2

L/r
100
Figure 11-7. Critical or allowable stress is given by the solid line. Dashed portion
of Euler's curve is not valid.

ILLUSTRATIVE PROBLEM

1101. Select the lightest W shape that can be used as a column

7 m long to support an axial load of 450 kN with a factor of safety of 3.
Assume (a) both ends hinged and (b) one end fixed and the other
hinged. Use aPL = 200 MPa and E = 200 GPa.
Solution:

Part a. For steel with a proportional limit of 200 MPa, the

specifications for Euler's formula with hinged ends require that L/r >
100. If L/r < 100, the limiting stress is the proportional limit.
The specified working load, when multiplied by the factor of
safety, gives a critical Euler load of 1350 kN. Applying Euler's formula
and solving for /, we obtain

2 2

P=
Elm'
I = PL = (1350 X \(fi)(7)

Em 1 " (200 X 10
9 2
)(tt )

- 33.5 x 10" 6 m4 - 33.5 x 10

6
mm4
Also, the slenderaess ratio L/r > 100, from which the least r%is

L 7000
r <
100 100
= 70.0 mm
These criteria establish that the section must have a least / > 33.5
X 106 mm4 and a least r < 70.0 mm. This is satisfied by choosing a
W250 X 73 section with a least / = 38.8 X 106 mm4 and a least r = 64.6
mm.
If the selection were based on the proportional limit, the section

must have a minimum area of 6750 2

mm
(obtained by dividing the load
of 1350 kN by the proportional limit of 200 MPa) and a least r greater
450 11 Columns

than 70.0 mm. These conditions are satisfied by a W310 X 97 section

300 mm = mm.
2
with A = 12 and least r 76.9
The lightest section therefore is the W250 X 73 section.

Part b. The critical Euler load is 1350 kN, as before. With one
end fixed and the other hinged, the effective length of an equivalent
hinged column is 0.7 L = 0.7(7) = 4.9 m. Using this effective length in
place of the actual length, we find that the criteria for Euler's formula
are

/> ^
E<n
2
= (1 350iO0^)(4
(200 x ioV
;
9f =
2
16 _ 4xl0 - 6m,
4
16.4 X lO'mm
and
L
r <
100 r-«°-
The lightest section that satisfies these conditions is the W360 X 64 with
least / - 18.8 x 10
6
mm4 and least r = 48.1 mm.
The other set of criteria based on the proportional limit are

A > ±±£lLSL-t}L = 6>75 x 10" 3 m 2

= 6750 mm 2
and
200 x 106
r > 49.0 mm
which the lightest section available is the W250
for X 58 with A = 7420
mm 2 and r = 50.4 mm. Comparing the two sets of criteria, we see that
the proper section is the W250 X 58.
The unwary reader might be tempted to base his section only on /
without checking r and thereby choose a W200 X 52 section with
least / = 17.8 x 10
6
mm4
However, this section has a least r = 51.8
.

mm and an area of 6660 mm 2 which results in a stress exceeding the

,

proportional limit of 200 MPa. It is therefore not acceptable because it

violates the stress-strain proportionality on which Euler's formula is
based.
This problem demonstrates the importance of the slenderness ratio
in column analysis. In Part a, the selection is governed by elastic
stability (i.e., the use of Euler's formula), whereas in Part b the selection
is determined by the proportional limit.

PROBLEMS

1102. A 50-mm by 100-mm timber is used as a column with fixed

ends. Determine the minimum length at which Euler's formula can be
11-5 Intermediate Columns. Empirical Formulas 451

used if E= 10 GPa and the proportional limit is 30 MPa. What central

load can be carried with a factor of safety of 2 if the length is 2.5 m?
Arts. L - 1.66 m; P= 32.9 kN
1103. An aluminum strut 2 m long has a rectangular section 20
mm by 50 mm. A bolt through each end secures the strut so that it acts
as a hinged column about an axis perpendicular to the 50-mm dimen-
sion and as a fixed-ended column about an axis perpendicular to the
20-mm dimension. Determine the safe central load, using a factor of
safety of 2.5 and E= 70 GPa. Ans. P= 9.24 kN
/llO^ A
square aluminum bar is to support a load of 40 kN on a
lengtlfof 3 m. Assuming pinned ends, determine the length of each side.
Use E- 70 GPa. Ans. 50.0 mm
1105. Repeat Problem 1 104 assuming that the column is made of
wood for which E = 10 GPa.
/il06i Two C310 X 45 channels are latticed together so they have
equal^moments of inertia about the principal axes. Determine the
minimum length of a column having this section, assuming pinned ends,
E - 200 GPa, and a proportional limit of 240 MPa. What safe load will
the column carry for a length of 12 m with a factor of safety of 2.5?
Ans. L = 9.89 m; P = 742 kN
1107. Repeat Problem 1106 assuming that one end is fixed and
the other hinged.

1108. Select the lightest W shape that will act as a column 8 m

long with hinged ends and support an axial load of 270 kN with a factor
of safety of 2.5. Assume that the proportional limit is 200 MPa and
E- 200 GPa. Ans. W310 X 74
1109. Select the lightest W shape that will act as a column 12 m
long with fixed ends and support an axial load of 700 kN with a factor
of safety of 2.0. Assume that the proportional limit is 200 MPa and
E= 200 GPa.

11 -5 INTERMEDIATE COLUMNS. EMPIRICAL FORMULAS

The preceding discussion showed that long columns can be treated by

Euler's formula provided that the slenderness ratio is larger than the
value at which the average stress reaches the proportional limit. For
hinged steel columns, this limit is L/r w 100 at 200 MPa. Euler's
formula is not valid for smaller slenderness ratios. The definition of a
short column one whose length does not exceed 10 times the least
as
lateral sets the upper limit of the slenderness ratio at about
dimension
30 for a rectangular section. For practical purposes, the limiting stress
452 11 Columns

Tangent-modulus curve

on a short column has been found to be the stress at the yield point;
extreme care is required to prevent buckling when stressed to this point.
Figure 11-8 shows these conditions for steel having a yield point of 280
MPa and a proportional limit of 200 MPa.
Various methods have been proposed for bridging the gap between
the short column range and the long column range. However, none of
them has been accepted universally for intermediate columns, partly
because of their departure from the stress-strain relationship when the
stresses exceed the proportional limit, and partly because of their
indeterminate mixture of direct and flexural stresses when loads are
reduced by an appropriate factor of safety to bring the stresses below
the proportional limit.
Most empirical formulas for intermediate columns have been
developed for steel because it is such a common structural material. We
shall discuss these first and then indicate their extension to other
structural materials.
—
one proposed method that of the tangent-modulus theory
In
the Euler formula is extended to intermediate columns stressed above

the proportional limit by replacing the constant modulus £ by a

11-5 Intermediate Columns. Empirical Formulas 453

reduced modulus E, viz.,

i. = J^- (11 _ 6)
A (L/r)
1

The reduced modulus £,_also called the effective or tangent modulus, is

obtained by using for E the slope of the tangent to the stress-strain

diagram at the point corresponding to the average stress in the column.
This yields a curve that connects the curves in Fig. 11-8 representing
the short and long column formulas. Although method is empirical
this

because it violates the stress-strain proportionality assumed in the

derivation of Euler's formula, actual tests show close agreement with the
theoretical curve.*
Other methods are frankly empirical. One of the simplest, pro-
prosed by T. H. Johnson in 1886, consists of drawing a straight line
through the average of the test points obtained by plotting values of
P/A (when failure due to buckling appeared imminent) against the
corresponding values of L/r. The general equation of this straight-line
formula is

—P = - r L
A
a C— r

where a is the intercept for L/r = and C is the magnitude of the

slope.
The results obtained by Tetmajer and Bauschinger with tests on
structural steel bars with hinged ends have been widely used. Their
results gave for the critical load*:

4
A
- 330 - 1.45-
r
MPa (11-7)

This equation is shown in Fig. 11-8. As mentioned previously, the yield

point is the practical limit to P/A ; this is recognized in the formula by
setting a lower limit to L/r that corresponds to 35 for a yield point of
280 MPa.
Equation (11-7), when divided by a factor of safety of 3; gives an
equation for safe working loads that was once widely used in building
codes. It is so conservative, however, that it has been largely superseded
by others that will be discussed later. We hereafter refer to it in the
following form as the straight-line formula:

4
A
= HO - 0.483-
r
MPa (11-8)

*For a discussion of the tangent-modulus method, see F. R. Shanley,

Strength of Materials, McGraw-Hill, New York, 1957, pp. 582-588.
Numerical values for empirical column formulas given here are approximate
conversions to SI units.
 1 1 Columns

This equation is limited to a slenderness range of 30 < L/r < 120 for

main members but may be extended to 150 for secondary members used
for bracing. Below L/r = 30, P/A - 96.5 MPa is to be used.
Another widely used empirical column formula is the Rankine-
Gordon formula developed about 1860. It assumes that the maximum
2
deflection in a column varies with L /c; that is, 6 max = <f>L 2 /c, where <f>

is a constant of proportionality depending upon end conditions. Then

the maximum stress in a column is

O — —P -r —Mc
— P
= —- +
(PS m&x )c P
A I A Ar 2 A '-(tJ
whence the formula for average stress is given by

L=" °
2
* 1 + <t>(L/r)

A commonly used form, which we hereafter designate as the Rankine-

Gordon formula, is
P 124
MPa (11-9)
a
. 18 x 103 m
This which includes a factor of safety, is valid for main
formula,
members with L/r between 60 and 120 and for secondary members
with L/r up to 200. Below L/r = 60, a working stress of P/A = 103
MPa is specified.
Still another variation of the intermediate column formula is the
parabolic type proposed i* 1
S92 by Professor J. B. Johnson (not related
to T. H. Johnson of straight-line fame). This formula has the general
form

A • - <t)
in which a is the stress at the yield point and C is a constant chosen to
make the parabola tangent to Euler's curve.
The American Institute of Steel Construction (AISC) defines the
limit between intermediate and long columns to be the value of the
slenderness ratio Cc given by
-\rt\fT g^v/t^ U

in
;
-#
which E is the modulus of elasticity (200 GPa, for most grades of
steel)and a^ is the yield stress for the particular grade of steel being
used. For columns of effective length L and minimum radius of
e
11-5 Intermediate Columns. Empirical Formulas 455

gyration r, AISC specifies that for L€ /r > Cc , the working stress, aw , is

given by

2S(LJrf
(Note that this is Euler's formula with a factor of safety of 23/12 =
1.92.) For Le /r < Cc , AISC specifies the parabolic formula
k2
(Ljry
= (u - n)
a~. t
2c; n
where the factor of safety, FS, is given by

Observe that the factor of safety is 1.92 at Le /r = Cc and

, it becomes
smaller for larger values of the slenderness ratio.The variation of aw
with Le /r for several shown in Fig. 11-9.
grades of steel is

Most column formulas are based on columns with hinged ends.

Fixity of the ends increases the load capacity (see Art. 11-3): However,
structural columns, which compose the great majority of intermediate
columns, practically never have completely rigid ends. Hence it is good

practice to assume hinged ends, even though the column is actually

riveted or otherwise rigidified at its ends. The effective length may be
used for economical column design in the case of partial or complete
end restraints.
We should mention one formula that will be developed in Art.
11-7. This formula, called the secant formula, assumes a definite
eccentricity of load and is theoretically correct but exceedingly
cumbersome to use. It is given by
P a
(11-13)

+
1
tM^VS")
where a max is the maximum stress developed by a load P having a
known eccentricity e. The term c isfrom the
the perpendicular distance
axis of bending to the extreme and ec/r 2 is the eccentricity ratio.*
fiber,

We now consider column formulas for some materials other than

steel. The Aluminum Association, Inc., lists column specifications for

each of the various types of aluminum alloys. In these specifications, the

allowable stress is a constant for short columns, a straight-line relation

2
•It is equivalent to 8 m9x..c/ r in the discussion of the Rankine-Gordon for-
mula on p. 454.
456 1 1 Columns

400 \ huier s curv B

En 2
\

300
\ \

\
\
\
\
^#0 \
200
\
^9o \ 1

\
^T^25CT^ \ /
—— X Euler's curve

100
/^\

25 50 75 100 125 150 175 200

Figure 11-9. Working stress for columns (AISC specifications) for several grades
of steel.

approximating the tangent-modulus formula is used for intermediate

columns, and Euler's formula is used for long columns. For example,
the specifications for 2014-T6 aluminum alloy are*

a w = 193 MPa ^<12 (11-14)

r

a w = 212 - 1.59- MPa 12<- <55 (11-15)

r r

ow =
(L/rf
—
372 x 10 3
MPa ->55 (11-16)

*The numerical values quoted here are approximate SI conversions of values

found inSpecifications for Aluminum Structures, 3rd ed., Construction Manual
Series. Sec 1, Aluminum Association, New York, April 1976, p. 21.
11-5 Intermediate Columns. Empirical Formulas 457

The column length L in the specifications is defined as the "length

of the compression member between points of lateral support, or twice
the length of a cantilever column (except where analysis shows that a
shorter length can be used)."
For wood columns, the National Lumber Manufacturers Associa-
tion* recommends Euler's formula in the following form:

= ^E ^—,
3.619£
«« ,2
- (H-17)
2.727(L/r) {L/rf
Of course, load duration and moisture content adjustments must also be
made. For rectangular columns with least lateral dimension d,
r = Vd/\2 and Eq. (11-17) reduces to

2
(L/d)
In addition to the empirical column formulas just presented, a
great many other equations are used. However, it is unnecessary to
memorize them, for the formula to be used must always be specified. All
the equations have one feature in common: they reduce the safe
working load as the slenderness ratio increases, although in varying
.

proportions. Depending on the formula specified, the same column may

therefore support any one of several safe legal loads.

ILLUSTRATIVE PROBLEMS

1110.Using AISC column specifications, determine the safe

axial loads on a W360 x 122 section used as a column under the
following conditions: (a) hinged ends and a length of 9 m; (b) built-in
ends and an unsupported length of 10 m; (c) built-in ends and a length
of 10 m braced at the midpoint. Use o^ = 380 MPa.

Solution: Table B-2, Appendix B, gives, for a W360 X 122 section, an

area A = 15 500 mm
2
and a least r of 63.0 mm.

Part a. For ayp = 380 MPa, the limiting slenderness ratio is

9
2 ^(200 x 10 )
c =1Q2
-iW\ -V 380 x 10
6

Here, the slenderness ratio is Le /r = 9000/63.0 = 143 which is greater

^National Design Specification, National Lumber Manufacturers Association,

Washington, D.C., 1962.
458 11 Columns

than Cc . Therefore, the working stress is given by

2
=
\2tt E m 12^(200
x 10>)
m
a...
2
5()A Mpa
L\ 2
23(143)

and the safe axial load is

6
[P = oA] P= (50.4 x 10 )(15 500 x 10" 6 ) = 781 kN A ns.

Part Using the concept of effective length, we find that a

b.

column with ends is equivalent to a hinged column of

built-in or fixed
half the actual length. Hence, with Le - 0.5 L = 0.5(10) = 5 m, the
slenderaess ratio is Le /r = 5000/63.0 = 79.4, which is less than Cc =
102. Therefore, the working stress is determined as follows:

5 3(Lc /0 {LJrf
pc ,

FS "3 + 8C 8c?

3
5 3(79.4) (79 A) _ L9°
+ 3
3 8(102) 8(102 )

whence

1
- (Ljr)'
yp
2C?
o~, =
FS

2 2
(l-[(79.4) /2(102) ]}
a„, = (380 X 10
6
) = 139 MPa
1.90

Finally, the safe axial load is

[P = oA] P = (139 x 6
10 )(15 500x 10~ 6
)

= 2150kN Arts.

Part c. Braced at the midpoint, the column is equivalent to one

having a length of 5 m, fixed at one end and hinged at the other.
11-5 Intermediate Columns. Empirical Formulas 459

The is L = 0.7L = 0.7(5) = 3.5 m, whence Ljr =

effective lengthe
3500/63.0 = 55.6, which is less than Cc = 102. Proceeding as in Part b,
we find FS = 1.85 and aw — 175 MPa. Hence, the safe axial load is

[P = oA] P= (175 X 6
10 )(15 500 X 10" 6)
= 2710kN Arts.

This problem illustrates the increased strength of a column whose

ends can be perfectly rigidified. Since this condition is never realized in
practice, it is better when determining allowable loads always to assume
hinged ends, or to be more realistic in selecting the effective length with
fixed ends as about 0.75 L instead of 0.5 L.

1111. Select the lightest W shape that will support an axial load
of 360 kN on an effective length of 4.6 m. Use AISC column specifica-
tions with a^ 250 MPa.

Solution: Since both the area A and the least radius of gyration r are
unknown and no convenient relation between them can be set up, the
selection of the lightest W shape involves a trial-and-error procedure.
The steps are (1) assume a working stress; (2) calculate the area
required; (3) select the lightest appropriate section based on the area
required; and (4) for the section selected, calculate the allowable load
based on the column specifications. If the allowable load equals (or is

slightly larger than) the applied load, the section selected is the ap-
propriate one. If the allowable load is less than the applied load, a
heavier section must be selected and the procedure repeated. Clearly,
the number of trials which must be attempted before the correct section
isdetermined depends on how close the initial assumed stress is to the
actual stress. One suggestion is to assume an initial working stress of
80% of the stress at L/r = determined from the column specifica-
tions.
For steel with a yp = 250 MPa, the limiting slenderness ratio is

C =^ / =1/ .
2
/2?r (200 X 10
t-2
9
)
- ^ %

found to be
-iW-t 250 x 106
126

First try. At Ljr = 0, FS = f and aw - a yp /FS = 250/f = 150 ,

MPa. Assuming an initial stress of 0.80(150) = 120 MPa, the required

area is

36 ° X 10 *
A = -- 6
= 3 X 10- 3 m2 = 3000mm2
° 120 X 10
460 1 1 Columns

Therefore, from Table B-2, Appendix B, we select a W200 X 27 with

A 3390 mm 2
and least r = 31.2 mm. For this section, the slenderness
ratio is LJr = 4600/31.2 = 147 which is greater than Cc - 126. There-
fore, the working stress for this section is

2 2 9
\2tt E = 127T (200 x 10 )
-
CL.
2
47.7 MPa
23(Le /rf 23(147)
-6 =
Then, the allowable load is P = oA - (47.7 X 10
6
) (3390 X 10 )

162 kN. Since this is less than the applied load of 360 kN, the section
is inadequate.

Second try. Next, we select a W200 X 36, which has a larger area
and a larger least r. For this section, A — 4580 mm and least r =
2

40.9 mm. The slenderness ratio is Le /r = 4600/40.9 = 112, which is

less than Cc = 126. The working stress for this section is determined as

follows:

(Ljrf
™ 3 + 8C 8CC3
3
(112)
^ 3
+
8(126) 8(126)'
- 1.91

whence

yp
2C 2
a., =
FS

2
(H2)
:

2(126)
a., - (250 x 10
6
) = 79.2 MPa
1.91

Then, the allowable load for this section is

p = oA = (79.2 x 6
10 )(4580 x 10" 6 ) = 363 kN
Since this load only slightly larger than the applied load of 360 kN,
is

the W200 X 36
the appropriate section.
is

The procedure of selecting a section is greatly simplified by using

tables giving the allowable axial loads for different sections of various
lengths. Such tables are found in a steel handbook such as that pub-
11-5 Intermediate Columns. Empirical Formulas 461

by the AISC. However, this problem illustrates the

lished trial-and-error
method which arises frequently in structural design.

PROBLEMS
1112. Determine the slenderness ratio of a 4-m column with
built-inends if its cross-section is (a) circular with a radius of 50 mm
and (b) 40 mm square. Use the concept of effective length.
Arts, (a) 80.0; (b) 173

1113. Using AISC column specifications, determine the maxi-

mum length of a W360 X 122 section if it is used as a hinged-end
column to support a load of 1200 kN. Use o w = 450 MPa.
Arts. L = 7.25 m
1114. Determine the maximum length of a W250 X 167 section
used as a hinged-end column to support a load of 1600 kN. Use AISC
specifications with o^ = 380 MPa.

1115. What factor of safety should be used with Euler's column

formula so that it will give the same load capacity for steel as is given by
the upper limit for main members of (a) the straight-line equation, Eq.
(11-8); (b) the Rankine-Gordon equation, Eq. (11-9).

1116. A W360 X 134 section is used as a column with hinged

ends. Using AISC specifications, determine the maximum load which
can be applied if (a)L = 9m;(b)L= 15 m. Use aw = 290 MPa.
Ans. (a)1740kN;(b)687kN
1117. A W200 X 100 section is used as a column 9 m long with
built-in ends, (a) Using AISC specifications, compute the safe load that
can be applied if the effective length is three-quarters the given length,
(b) What is the safe load if the column is also braced at its midpoint?
Use a yp - 380 MPa. Ans. (a) 837 kN; (b) 2400 kN
1118. Repeat Problem 1117 assuming that the length of the
column is 14 m with a W310 X 500 section.

1119. column with an effective length of 10 m is fabri-

A steel
cated from two C250 X 45 channels latticed together so that the section
has equal moments of inertia about the principal axes. Determine the
safe load using AISC specifications. Use a
w = 380 MPa.
Ans. P= 883 kN
1120. Four 100 X 100 X 10 mm
angles are latticed together to
form the column section shown in Fig. P-1120. Using AISC specifica-
tions with a^ = 290 MPa, determine the maximum effective length at
which a 500-kN load can be safely supported. What should be the
spacing between lattice bars if the slenderness ratio of each separate
angle is not to exceed three-fourths of that of the fabricated section?
462 11 Columns

Figure P-1120.

1121. In the bridge truss shown in Fig. P— 1 121, the end chord
AB is composed of two C230 X
30 channels latticed together so that the
fabricated section has equal moments of inertia about the axes of
symmetry. If the safe load P on the truss is governed by the strength of
member AB, determine P using AISC specifications with aw = 290
MPa. Ans. P = 210kN

1122. Select the lightest W shape that can be used as a column to

support an axial load of 420 kN on an effective length of 4 m. Use
AISC specifications with a
yp
= 250 MPa. Ans. W200 X 36
1123. Select the lightest W shape, according to AISC specifica-
tions, that can be used as a column to support an axial load of 700 kN
on an effective length of 5.5 m. Assume o^ = 250 MPa.
1124. Repeat Problem 1123 assuming that the axial load is 690
kN and o^ = 345 MPa. .

1125. A
hinged-end steel column 10 m
long is fabricated from a
W200 X 46 beam and two C310 X 45 channels arranged as shown in
Fig. P— 1 125. Determine the safe axial load using AISC specifications
with o = 250 MN/m' Ans. P = 1440 kN
 )

11-5 Intermediate Columns. Empirical Formulas 463

0=^
W200X46- C310 X 45

[L

Figure P-1125.

1126. The connecting rod of an engine has a cross section with

the following properties: area = 300 mm2 rx — 3.00 mm; and ry =
;

1.40 mm. The wrist pin and crank pin cause the rod to act as a hinged
column about the X axis and as a fixed-ended column about the Y axis.
Applying the concept of effective length, find the safe load for a length
of 250 mm, using the Rankine-Gordon equation, Eq. (11-9).
Arts. P = 25.8kN
1127. Derive a parabolic formula of the general type P/A =
a — C(L/r) that will be applicable to aluminum alloy columns with
2

hinged ends. Assume that the parabolic formula will be tangent to an

Euler formula with a factor of safety of 2. Use o = 110 MPa and
E= 70 GPa. {Hint: For the two formulas, equate their unit loads and
also equate their derivatives with respect to the slenderness ratio.)
2
Ans. -j - (110 X 10
6
) - 8760[ — for -< 79.3

1128.Four 100 X 100 X 13 mm angles are bolted back to back

as shown P— 1 128. Determine the safe load when they are used as
in Fig.
a hinged-end column 4 m long. Use AISC specifications with o^ =
250 MPa.

^
ZS

1
Figure P-1128.

1129. Determine the safe axial load which can be applied to a

2014-T6 aluminum alloy column if its length is (a) 1 and (b) 3 m. m
Assume the geometric properties of the cross section are identical to
those of an S3 10 X 52 steel section. .
, . _ _
.
T „. ,__
,
,^T
Ans. (a) 984 kN; (b) 172 kN
464 11 Columns

1130. Repeat Problem 1 129 assuming that the geometric proper-

ties of the cross section are identical to those of an S250 x 52 steel

section.

1131. Determine the safe axial loads on an oak column 150 mm

by 200 mm if the length is (a) 2 m and (b) 4 m. Use E = 1 1.5 GPa.
Arts. (a)582kN;(b) 146 kN
1132. Repeat Problem 1131 for a pine column 50 mm by 200 mm
for which E- 11.2 GPa.

1 1 -6 ECCENTRICALLY LOADED COLUMNS

Columns are usually designed to support axial loads, and the preceding
formulas have been presented with this in mind. Under certain condi-
tions, however, columns are subject to loads having a definite eccentric-
ity. This occurs, for example, in the case of a beam connected to the
column flange in a building. The secant formula derived in the next
article is especially adapted to such cases, but it is so unwieldy that
several simplified procedures are currently used.
In the maximum stress approach, the eccentrically loaded column is

treated as if it were an eccentrically loaded short strut (see Art. 9-3).

However, to eliminate the possibility of buckling so that the effect of
deflection on the moment arm of eccentric loads may be neglected, the
maximum compressive stress is limited to the unit load as computed
from a specified column formula. This approach is valid only for
moderate slenderness ratios.
Applying this procedure to the column in Fig. 11-10 which
supports an axial load PQ and a load P at an eccentricity e, we find the
design criterion to be

Here o is the unit load computed from the specified column formula
(always use the least radius of gyration to determine the slenderness
ratio), / is the moment of inertia with respect to the axis about which
the eccentric load causes bending (axis X-X in Fig. 11-10), and S is the
section modulus with respect to that axis.
Modern design criteria have refined the maximum stress approach
to include the moments, called secondary moments, which are also
introduced because the neutral axis is deflected (the so-called P-8
effect). These criteriamost often take the form of interaction equations
which attempt to "weigh in" the relative importance of the axial stress
and the bending stress.
11-6 Eccentrically Loaded Columns 465

Figure 11-10. Axial load P and eccentric load P on column.

For example, AISC*recommends when

computed axial
that, the
stress^ is less than 15% of the actual stress Fa
would be permitted
that
were only axial stress acting, the secondary moments may be neglected
and the member must satisfy the following criterion:

L + ^L + :§L < i.o (a)

bx by

When^ > 0.15Fa secondary moment

, effects cannot be neglected.
In these cases, AISC requires the following formulas to be satisfied:

^ (-£>
'
Cmxfl
mxJbx ^ my J by
< 1.0 (*)

('-iK
+ * (c)
0.60F, Fbx Fby

In Eqs. (a), (b), and (c), the various terms are as follows:

fa = computed axial stress;

Fa = allowable axial stress if axial force alone were acting;

fbx = computed bending stress about the major axis disre-

garding the secondary moment;

ftp
= computed bending stress about the minor axis disre-
garding the secondary moment;

The notation used here is that used in Manual for Steel Construction,
American Institute of Steel Construction, New York, 1970.
466 1 1 Columns

fbx m allowable compressive bending stress about the major

axis if moment alone were acting;

F « allowable compressive bending stress about the minor

axis if moment alone were acting;

F'ex — Euler buckling stress for buckling about major axis;

F' — Euler buckling stress for buckling about minor axis;

C^ C m reduction factors to correct for overconservatism in

some cases of the amplification factor [1 - (/a /Fe')].

For compression members in frames subject to joint translation, or

sidesway,Cm may be taken as 0.85.
For compression members in frames braced against sidesway and
subject to end moments (but not transverse loads between supports), use
Cm m 0.6 - 0A(M l
/M
2 ) > 0.4,
where 2 is M /M
the ratio of the
x

smaller to the larger end moment. This ratio is positive when the
member is bent in reverse curvature and negative when it is bent in
single curvature.
For compression members in frames braced against sidesway in
the plane of the loading and subjected
to transverse loading between the
supports, Cm may
be taken as 0.85 for members with restrained ends
and as unity for members with no end restraints; Cm may also be
determined by rational analysis in this case.
AISC specifications also include formulas for determining the
allowable bending stresses Fb as a fraction of the yield stress. The value
of Fb depends upon the width/thickness ratio of the section and the
bracing intervals.
Interaction formulas similar to the AISC formula have been
adopted for other structural materials such as wood and aluminum.
The design of members to carry both axial and bending loads is
essentially an iteration procedure. An assumed section is checked for
adequacy using the appropriate criteria. This procedure is greatly sim-
plified by the many tables and graphs which are available to assist the
designer. Computer programs are also available which will assist the
designer in selecting the optimum section that satisfies the interaction
formulas.
The problems which follow will illustrate the application of the
maximum stress approach. For applications of the interaction equa-
tions, the reader is referred to any modern structural design text.*

•See. for example. L. A. Hill, Jr., Fundamentals of Structural Design: Steel,

Concrete, and Timber, Intext, New York, 1975.
11-6 Eccentrically Loaded Columns 467

ILLUSTRATIVE PROBLEM

1133. A W360 x 134 section is used as a column with an effec-

tive length of 7 m to support one track of a traveling crane in a factory.
Determine the permissible reaction P if the column also
maximum
carries a load of 400 kN
from an upper floor, as shown in Fig. 11-11.
Use the maximum stress approach (Eq. 11-19) and the AISC column
specifications. Assume a
w = 250 MPa.
Solution: Table B-2, Appendix B, gives the properties of a W360 X
134 section as A = 17 100 2
mm
Sx = 2330 X 103 ,
3
and least mm ,

r = 94.0 mm. The slenderness ratio is Ljr = 7000/94.0 = 74.5. For

<7yp = 250 MPa, the critical slenderness ratio is

W-V
2 9
27T (200 x 10 )
= 126
250 x 106

Since Le / r < Cc , the appropriate AISC formula (Eq. 11-11) determines

the working stress as follows
3
3(LJ r) (LjrY k

8CC SCC 3
3
3(74.5) (74.5)
FS=3 + ;
= 1.86
8(126) 8( i26)

400 kN

W360 X 134

Figure 11-11.
 11 Columns

whence

l
-
yp
2C 2
o~. =
FS
2

- (74.5)
1
:

2(126)
6
(250 x 10 ) = 111 MPa
1.86
Using the maximum stress approach, we consider the column to
act as an eccentrically loaded short compression member limited to this
maximum stress of 111 MPa. Applying Eq. (11-19), we obtain
SP M
A S
3

111 x 10
6 = ( 40Q x 1q3 + p) + 0.125P - 0.075(400 x 10 )
" -6
(17 100 x 10" 6 ) 2330 X 10

whence

P - 896 x 10
3
N= 896 kN Am.

PROBLEMS
In the following problems, use the maximum stress approach and
the AISC column specifications unless otherwise directed.

1134. A W360 X
122 section is used as a column with an effec-
tive length of 10m. Determine the maximum load that can be carried at
an eccentricity of 300 mm. Should the load be placed on the X or Y
axis? Assume o^ = 290 MPa. Ans. P = 190 kN
1135. Repeat Problem 1 134 for a column with an effective length
of 4.5 m.

1136. A by 75
steel column 50
in section has an mm mm
effective length of m. Compute the maximum load that can be
1.5
carried at an eccentricity of 120 from the geometric axis. Assume mm
250 MN/m The column also carries an axial load of 50 kN.
2
o
yp
= .

Ans. P- 25.8 kN
1137. A steel pipe 2.5 m long, built in at its lower end and free at
itsupper end, supports a sign whose center of gravity is 0.6 from the m
axis of the pipe. Applying the concept of effective length, determine the
maximum weight of the sign. The outside diameter of the pipe is
140 mm, its area is 2800mm 2
, and its moment of inertia is 6.32 X 6
10
mm 4
. Use o = 250 MN/m 2
.
1 1 -7 The Secant Formula 469

1138. A W360 X 134 section used as a column whose effective

is

length is 6 m. The column carries axial load of 260 kN and an

an
eccentric load of 220 kN applied on the minor axis. Determine e, the
maximum eccentricity of the load, using a
w= 250 MPa.
Ans. e = 952 mm
1139. A C310 X 45 channel is used as a hinged-end column 2.2
m long. How far off center can a load of 50 kN be placed on the X axis?
Assume o^ = 380 MPa and that the tensile stress is to be limited to 140
MN/m 2
. On which side of the Y axis must the load be placed.
Ans. 100 mm
1140. Repeat Problem 1139 using a C3 10 X 31 channel.

1141. A W360 X
134 section is to be used as a column with a
length of 9 m. The column supports an axial load of 260 kN and an
eccentric load of 360 kN acting on the Y axis. Determine the maximum
eccentricity of the 360-kN load using the maximum stress method and
the straight-line formula, Eq. (11-8). Ans. 178 mm
1142. Repeat Problem 1141 using a W360 X 347 section.

11-7 THE SECANT FORMULA

A theoretically correct formula for eccentrically loaded columns can be

obtained by extending Euler's analysis in the following manner. Figure
11-12 shows the center line of a column carrying a load P with an
eccentricity eon a length L. If this column is extended as indicated by
the dashed lines, it becomes equivalent to a hinged Euler column
having a length X. The value of P shown is the critical load for this
unknown length X. Such a column has the shape of a sine curve whose
equation with respect to an origin at one end was shown in Eq. (e), Art.
11-3, to be

=8sin
(*V50
Since from Eq. (11-1) VP/EI = ir/L for the fundamental shape of a
hinged column, we obtain

y-Ssin(^) (a)

If the origin is taken at the center, Eq. (a), in terms of the

equivalent but unknown length A, becomes:

>> = Scos(^) (b)

470 11 Columns

J k

r '

L
2

Y + — \^6+
f-J° '

i
)

J
i

X) i

3
<

Ae ><-

\ p
i i

>

Figure 11-12. Eccentrically loaded column.

Applying the condition that y — e at x = L/2 gives

e = o cos

from which the value of S

m is obtained for substitution in Eq. (b), thereby
yielding

C0S
(f)
y - e (c)
COS
(£)
The value of X is found by applying Euler's formula in Eq. (11-1)
with a length \ so that

P-_, or A
T '
and
x
=
V^
When this is substituted in Eq. (c) we obtain the following equation for
the column in Fig. 11-12:

v = e (d)

cos
11-7 The Secant Formula 471

The curvature is found by differentiating Eq. (d) twice, whence

d^y
dx 2
^

bending moment at x = is
eP_

EI
COS

cos
M
(*\/l0

Hence, from the differential equation of the elastic curve, the maximum

cos
(2 y-Ei J

(«)

The maximum stress in the eccentrically loaded column is com-

posed of a direct stress and a flexural stress as in a short strut, so

P Mc

whence, using I = Ar 2 and the value of M from Eq. (e), we obtain

P
1 +

This equation is known as the secant formula. The buckling load P

specified in it is converted to a working load Pw by replacing P by fPw ,

where / is the factor of safety. Doing this and taking the maximum
stress as the yield point we reduce Eq. (11-20) to

J
°yp- A 1 + ^sec \^-\ -prr (H-21)

Trial- and- error methods are necessary in using these equations.

Their use is facilitated by calculating the values of L/r corresponding
to assumed values of P/A for various values of the eccentricity ratio
ec/r 2 such as 0.2, 0.4, and so forth, to 1.0. This procedure gives the data
in the Table 11-1, from which design curves like those in Fig. 11-13
may be potted.
It is interesting to note that as the slenderness ratio approaches
zero, the value of the secant term in Eq. (11-20) approaches unity.
Hence in the limit, Eq. (11-20) reduces to

P/ ec\ P Mc

which is the equation used for eccentrically loaded short struts.

 472 11 Columns

TABLE 11-1. Design Data for Eq. (11-21) Using

<r„- 290 MPa, / - 2\, and E - 200 GPa

p L
A r

ec
(MPa) 77-0.2 0.4 0.6 0.8 1.0

20 1 93 188 183 178 172

25 I7l 165 159 153 146
30 155 148 140 133 125
35 142 134 125 116 107
40 131 122 112 102 90.9
45 122 HI 100 88.6 74.6
50 113 101 87.9 73.6 56.9
55 106 91.7 76.2 58.4 34.1
60 98.9 82.4 63.7 39.5 —
65 92.1 72.7 49.1 — —
70 85.3 62.0 28.4 — —
75 78.1 49.0 — — —
80 70.2 30.1 — — —
85 60.8 — — — —
90 48.1 — — — —

a.l*c

25 .50 75 100 125 150 175 200

Slenderness ratio,

Figure 11-13. Design curves for secant formula with factor of safety — 2\.
1 1 -7 The Secant Formula 473

SUMMARY
Long slender columns are solved by Euler's column formula. For
columns with hinged ends it is

n« 2
p_ 2 (li-D
L
or
2
P Eir
(H-5)
A {L/rf
For other end conditions, L in these formulas is replaced by an effective
length Le values of which are tabulated on page 447.
,

Euler's formulas are theoretically correct, provided the stress does

not exceed the proportional limit. The lower limit of L/ r for which they
are valid may be obtained by assuming that P/A represents the actual
stress in a straight, axially loaded column, and replacing P/A in Eq.
(11-5) by the value of the proportional limit.

Columns with a slenderness ratio less than the lower limit for
Euler's formulas are known as intermediate columns. No theoretically
correct formulas have yet been developed; the closest approach, so far,
is the secant formula (Eq. 11-13). However, this formula is too un-
wieldy; various empirical formulas are used instead. These empirical
formulas are specified by the building codes of various communities,
and the legal specifications of these codes must be adhered to.
Eccentrically loaded columns are analyzed by using either the
maximum stress approach or interaction equations. In the maximum
stress approach, the columns are treated as eccentrically loaded short
struts (see Art. 9-3), except that the value of the working stress is
obtained by using a specified column formula. Interaction equations
attempt to "weigh in" the relative importance of the axial and bending
stresses.
 Riveted and
Welded
Connections

12-1 INTRODUCTION

The analysis of riveted and welded connections involves so many

indeterminate factors that an exact solution is impossible. Nevertheless,
by making certain simplifying assumptions, practical solutions can be
readily obtained. The most significant of these assumptions is that when
the applied load passes through the centroid of the rivet group, each
rivet transmits a load equal to its shear or bearing capacity, depending
on which is lower. This assumption, in conjunction with the one that the
joint is made from a ductile material, permits us to consider riveted
joints as examples of uniform stress distribution.

12-2 TYPES OF RIVETED JOINTS: DEFINITIONS

There are two types of riveted joints: lap joints and butt joints. In a lap
joint,the plates to be connected are lapped over one another and

474
12-2 Types of Riveted Joints: Definitions 475

fastened together by one or more rows of rivets, as in Fig. 12-1. In a

butt joint, the plates are butted together and joined by two cover plates
riveted to each of the main plates. (Occasionally only one cover plate is

used.)The number of rows of rivets used to fasten the cover plates to

each main plate identifies the joint as single-riveted, double-riveted, etc.
(Fig. 12-2).Frequently the outer cover plate in a boiler joint is narrower
than the inner cover plate, as in Parts (c) and (d) of Fig. 12-2, the outer
plate being wide enough to include only the rivet row in which the rivets
are most closely spaced. This type of connection is called a pressure

-0- -0- 0- (::)

fy <& fy (: ::)

Figure 12-1. Lap joints.

•0- -0--$- -0 ")

(:

(b) Double-riveted butt joint

Repeating Repeating^,
section section
«« >

~<
Row 1

TV >>^\] Rowl C #, Row 2 C--

Row 2 C Row3(
Row3C ) Row 4C
/•K <4\ A\ A\ As Row3C Row4C
vl/ \U \1/ \1/ \i/
/TV /TV /TV /T\ /TV
Row 2 C"- ~\I/ \i/ \i/ \iy \i/ Row3<
$ Row 1 (: O- -^ Row
Rowl
2 (::>
(::)

(c) Triple-riveted butt joint — (d) Quadruple-riveted butt-joint"-

pressure type pressure type

Figure 12-2. Butt joints.

476 12 Riveted and Welded Connections

joint. Caulking along the edge of the outer cover plate to prevent
leakage is more effective in this type.

The spacing between the rivets in a given row is called the pitch.
When the spacing varies in different rows, as in a quadruple-riveted
joint, the smallest spacing is known as the short pitch, the next as the
intermediate pitch, and the greatest as the long pitch. The spacing
between consecutive rows of rivets is called the back pitch. When the
rivets in consecutive rows are staggered, the distance between the
centers of the rivets is the diagonal pitch.
In determining the strength of a riveted joint, computations are
usually made for a length of joint corresponding to a repeating pattern
of rivets. The length of the repeating pattern (more commonly called a
repeating section) is equal to the long pitch.
Sometimes confusion arises in deciding how many rivets belong in
a repeating section. Study of the repeating section in Fig. 12-2c shows
that there are five rivets effective in each half of the triple-riveted butt
two half rivets in row
joint: 1, two whole rivets in row 2, and one whole
and two half rivets in row 3. Similarly there are 11 rivets effective in
each half of the repeating section of the quadruple-riveted joint in Fig.
12-2d. Inspection of the enclosed repeating rivet groups confirms these
statements.
The efficiency of a riveted joint indicates how well the joint has
been designed. Itcompares the strength of the joint with that of the
unriveted plate and is defined by

= ^ngth of joint
efficiency
strength of solid plate

The rivet holes in boiler joints are usually drilled or subpunched

and reamed out to a diameter 1.5 mm
larger than the rivet; however, the
rivet is assumed to be driven so tightly that it fills the hole completely.
In calculations, therefore, the diameter of the driven rivet is considered
equal to that of the rivet hole.

12-3 STRENGTH OF A SIMPLE LAP JOINT

Riveted joints may be considered to be examples of uniform stress

governed by the equation P = Ao. The application of this equation to
the elemental types of failure is easily understood by considering a
single-riveted lap joint. The failure of a riveted boiler joint is equivalent
to any relative movement of the main plates of the joint, because this
will destroy its function, which is to maintain a rigid and leakproof
connection.
1 2-3 Strength of a Simple Lap Joint 477

In Fig. 12-3, shear of the rivet evidently permits the main plates to
separate. The failure load in shear is given by

P. ~ A.T = —— T (12-2)

where d represents the diameter of both the rivet hole and the driven
rivet.

Figure 12-4 represents a failure caused by the tearing of the main

plate. This failure occurs on a section through the rivet hole, this section
evidently having minimum tearing resistance. If p is the width of the
plate or the length of a repeating section, the resisting area is the
product of the net width of the plate (p — d) and the thickness t.

The failure load in tension is

P = A a = (p- d)ta
t t t t
(12-3)

A third type of failure, called a bearing failure, is shown in Fig.

12-5. In this case, relative movement between the main plates may
result from a permanent deformation or enlargement of the rivet hole
caused by excessive bearing pressure. The rivet itself may also possibly
be crushed.
Actually, the intensity with which the rivet bears against the rivet
hole is not constant but varies from zero at the edges of the hole to a

wzzz
^ \\\\\\^

Figure 12-3. Shear failure.

-*rWM7<

Figure 12-4. Tear of plate at section through rivet hole.

Pt -A o -(p-d)to
t t t
478 12 Riveted and Welded Connections

c©

7777Z& 3 Projected area of

rivet hole

Figure 12-5. Exaggerated bearing deformation of upper plate.

Pt" Ab ab "(td)ab

(a) Tear out (b) Shear behind rivet

Figure 12-6. Possible types of failure if rivet hole is too close to edge of plate.

maximum value directly in back of the rivet. It is common practice,

however, to use a bearing stress ab that is assumed to be uniformly
distributed over the projected area of the rivet hole. Then the failure
load in bearing is expressed by

Pb = A b ak = (td)ob (12-4)

Other types of failure are possible but will not occur in a properly
designed joint. Among them are tearing of the edge of a plate back of a
rivet hole, as shown in Fig. 12-6a, shear failure of the plate behind a
rivet hole, as shown in Fig. 12-6b, or a combination of both. Such
failures are unlikely to occur if the distance from the edge of the plate to
the center of the rivet is l| to 2 times the diameter of the rivet. In the
problems we shall assume that this distance is great enough to prevent
this type of failure.

12-4 STRENGTH OF A COMPLEX BUTT JOINT

The strength of a joint is limited by the capacity of the rivets to transmit

load between the plates or by the tearing resistance of the plates
themselves, depending on which is smaller. The calculations are divided
12-4 Strength of a Complex Butt Joint 479

into two steps: (1) preliminary calculations to determine the load that
can be transmitted by one rivet in shear or bearing, and (2) calculations
to determine possible methods of failure. The procedure and reasoning
are explained in the following illustrative problem.

ILLUSTRATIVE PROBLEM

1201. A repeating section 180 mm long of a triple-riveted butt

joint of the pressure type is illustrated in Fig. 12-7. The rivet hole
diameter is d= 20.5 mm, the thickness of the main plates is t — 14 mm,
the thickness of each cover plate is ^ = 10 mm. The ultimate stresses in
shear, bearing, and tension are respectively r 300 MPa, ab = 650
=
MPa, and a t
— 400 MPa. Using a factor of safety of 5, determine the
strength of a repeating section, the efficiency of the joint, and the
maximum internal pressure that can be carried in a 1.5-m-diameter
boiler for which this joint is the longitudinal seam.

Solution: The use of ultimate stresses will determine the ultimate load,
which is then divided by a suitable factor of safety to determine the safe
working load. An equivalent but preferable procedure is to use allow-
able stresses to determine the safe working load directly. This has the
advantage of involving smaller numbers. Thus, dividing the ultimate
stresses by the specified factor of safety, we find that the allowable
stresses in shear, bearing, and tension respectively are t = 60 MPa,
ab = 130 MPa, and a, = 80 MPa. These values are used in the following
computations.

«
I

II
< 180

It mm
i

U
I
I ^
*i
CO lit!
ij
!
CO **

t'= 10 mm t = 14 mm
Figure 12-7.
480 12 Riveted and Welded Connections

Preliminary Calculations. To single-shear one rivet:

p «
'
— 4
r = -(20.5 x
4
1(T
3 2
) (60 x 10
6
) = 19.8 kN

To double-shear one rivet:

Ps - 2 x 19.8 = 39.6 kN
To crush one rivet in the main plate:

Ph = (td)ob = (14 x 3
1(T )(20.5 X 3
1(T )(130 x 10
6
) = 37.3 kN
To crush one rivet in one cover plate:

P'h = {t
f 3 3 6
d)ob - (10 x 10" )(20.5 x 10- )(130 x 10 ) = 26.7 kN

Methods of Failure. Generally, there are only two basic

Possible
methods of These are determined by (a) capacity of the rivets to
failure.

transmit load and (b) the tearing resistance of the plates.

(a) Rivet capacity.The strength of the single rivet in row 1 in a

repeating section determined by the lowest value of the load that will
is

single-shear the rivet, crush it in the main plate, or crush it in one cover
plate. On the basis of the values derived from the preliminary calcula-
tions, this value is 19.8 kN.
The strength of each of the row 2 depends on the
two rivets in
lowest value required to double-shear the rivet, crush it in the main

plate, or crush it in both cover plates. The preliminary calculations show

this value to be 37.3 kN per rivet, or 74.6 kN for both rivets in row 2.

Each of the two rivets in a repeating section in row 3 transmits the

load between the main plate and the cover plates in the same manner as
those in row 2, and hence the strength of the two rivets in row 3 is also
74.6 kN.
The total rivet capacity equals the sum of the rivet strengths in all

rows:

Pr = 19.8 + 74.6 + 74.6 = 169.0 kN (a)

(b) Tearing capacity. The external load applied to the joint acts
directly to tear the main plate at row 1, and a failure would be similar to
that shown in Fig. 12-4. The load that will tear the main plate at row 1
is given by
/>, = (p - d)ta t

6
= [(180 x 10" 3 ) - (20.5 X 10" 3 )](14 X 10" 3 )(80 X 10 )

- 178.6 kN (b)

The external load applied to the joint does not act directly to tear
the main plate at row 2 because part of the load is absorbed or
1 2-4 Strength of a Complex Butt Joint 481

transmitted by the rivet in row from the main plate to the cover plate.
1

Hence, if the main row 2, the external load must be

plate is to tear at
the sum of the tearing resistance of the main plate at row 2, plus the
load transmitted by the rivet in row 1 from the main plate to the cover
plate. This statement is illustrated by the free-body diagram in Fig.
12-8 and is further clarified by Fig. 12-9, which shows how the failure
may actually occur.
The load transmitted by the rivet in row 1 is the first term on the
right-hand side of Eq. (a) and is the value of the rivet strength in row 1.
The external load to tear the main plate at row 2 must include this
value; it is given by

^2 Z=Z
{P~ 2d)ta t
+ rivet strength in row 1

= [(180 x 10" 3 ) - 2(20.5 x 10" 3 )](14 x 10" 3 )(80 x 10

6
)
3
+ (19.8 X 10 )

- (155.7 x 10
3
) + (19.8 x 10
3
) = 175.5 kN (c)

Similarly, the external load required to tear the main plate at row
3 must include the rivet resistance in rows 1 and 2, or

P3 = [(180 x 10" 3 ) - 2(20.5 x 10f )](14

3
x 10" 3 )(80 x 10
6
)

+ (19.8 x 10
3
) + (74.6 x 10
3
) = 250.1 kN (d)

It is obvious now that this computation need not be made because,

since the tearing resistance of the main plates at rows 2 and 3 is equal, it
gives a larger value than that of Eq. (c). However, this computation

Tearing resistance of
main plate at row 2

^mz^hA z:

Load transmitted to cover

plate by rivet in row 1

Figure 12-8.

Row 1

Row 2

Figure 12-9. Failure by shear of rivet in row 1 plus tear of main plate in row 2.
482 12 Riveted and Welded Connections

illustrates the procedure to be used when a greater number of rivet holes

in row 3 than row 2 reduces the tearing resistance of row 3 below that of
row 2.
At row 3, the tearing of the cover plates is resisted by the tensile
strength of the reduced section at that row (see Fig. 12-8). The tensile
strength of one cover plate is given by

Pc * [(180 x 10" 3 ) - 2(20.5 x 10" 3 )](10 x 10~ 3 )(80 x 10

6
)

= 111.2 kN
In an ordinary butt joint, the tensile capacity of both cover plates is
twice this value. In a pressure joint, however, where one cover plate is
shorter than the other, the load capacity of the shorter plate must be
compared with the rivet load transmitted to it. In this example, the

upper cover plate transmits the rivet load of four rivets in single shear,
or 4 X 19.8 = 79.2 kN, which is less than its tear capacity of 111.2 kN.
Hence the load capacity of both cover plates becomes
Pc = 79.2+ 111.2= 190.4 kN (e)

determined by rivet shear in the upper plate and by tension at row 3 in

the lower plate.
The safe load is the lowest value of these several possible methods
of failure. Its value is

P - 169.0 kN
determined by the capacity of the rivets to transmit load.
The efficiency is

Eff.= - safe load

strengtn ol solid plate
16Q0 v ID3
6
X 100
(180 X 10" 3 )(14 x 10" 3 )(80 X 10 )
= 83.8% Ans.
The maximum internal pressure is found by applying the safe load
of 169.0 kN in a repeating length of 180 mm:

[IP = pDL] 2(169.0 X 10

3
= ) /?(1.5)(180 X 10" 3
)

p = 1.25 MPa Ans.

Observations Concerning Riveted Joints. We are now ready to

appreciate the significance of the following observations. Since the
rivets are driven when hot and contract as they cool, there are devel-
oped normal forces which press the plates of the joint tightly together.
Because of these normal forces, there will be a frictional resistance to
any motion of the plates past one another. This frictional resistance
1 2-4 Strength of a Complex Butt Joint 483

must be overcome before there is sufficient deformation of the plates to

permit the rivets to bear against the rivet holes. Thus there is an extra
margin of strength in the riveted connection. However, because it is

difficult to estimate this frictional resistance accurately, it is not consid-

ered when computing the strength of a riveted connection.
assumed further that each rivet, when driven, expands to fill
It is

the rivet hole completely. Only when this is true will all the rivets
transmit the load simultaneously. If some of the rivets fill the holes only
partially, these rivets will not begin to bear against the plates until there
has been sufficient deformation in the remaining rivets and/ or plates to
take up the slack in the rivet holes.
We have also assumed that the rivets do not bend but remain
essentially straight. This is possible only if there are equal elastic
deformations of the main and cover plates between adjacent rows of
rivets. From the preceding discussion concerning tearing of the main
plate at row 2, it is evident that the main plate between rows 1 and 2
carries substantially more load than the cover plate. Hence the physical
requirement of equal deformations cannot be true, even if we allow for
the thinner thickness of the cover plates. But since the plates are usually
specified to be ductile, equal plastic deformations , can occur as the
stresses approach the yield point.*
Because of these and other reasons, an exact analysis of a riveted
connection cannot be made. The procedures used here for boiler joints
(and those described later for structural joints) give usable values
determined by comparatively simple methods.

PROBLEMS
Unless otherwise stated, assume the allowable stresses in the
following problems to be r = 60 MPa, ob = 130 MPa, and a, = 80 MPa.
1202. The longitudinal joint of a boiler having 14-mm plates has
a strength of 350 kN in a pitch length of 400 mm. The efficiency of the
girth joint 45% and the allowable tensile
is stress is 80 MPa. Determine
the maximum diameter of the boiler if it is designed to operate at a
pressure of 1.4 MPa. Ans. D— 1.25 m
1203. A double-riveted lap joint forms the girth seam of a boiler
1.5 m in diameter. Pitch of the rivets mm; diameter of the rivet
is 80
holes is 17.5 mm; thickness of the plate mm. Find the strength of a
is 12
repeating section, the efficiency, and maximum internal pressure.
Ans. p = 962 kPa

*Other examples of inelastic action are discussed in Chapter 14.

484 12 Riveted and Welded Connections

1204. The longitudinal seam of a

boiler is a triple-riveted lap
rows equal to 140
joint, with the pitch in the outer and that in the mm
middle row equal to 70 mm. Diameter of the rivet holes is 23.5 mm;
thickness of the plate is 12 mm. Determine the strength of a repeating
section and the efficiency.

1205. The dimensions of a double-riveted butt joint like that in

Fig. 12-2b are: diameter of the rivet holes, 23.5 mm; long pitch, 140
mm; short pitch, 70 mm; thickness of main plate, 14 mm, and of cover
plates, 10 mm. Compute the strength of a repeating section and the
efficiency.

1206. A double-riveted butt joint is of the pressure type in which

the upper cover plate extends over only the inner rows of whereas
rivets,

the lower cover plate extends over all rows. Its dimensions are: diameter
of the rivet holes, 23.5 mm; thickness of main plate, 14 mm; thickness
of each cover plate, 10 mm; long pitch, 140 mm; short pitch, 70 mm.
Determine the strength of a repeating section and the efficiency of the
joint. Ans. P = 1 1 1.6 kN; 71.2% efficient
1207. If the cover plates in Problem 1206 were each 8 mm thick,
determine the method of failure and the efficiency of the joint.
1208. In a double-riveted butt joint of the pressure type in which
the upper cover plate extends over only the inner rows of rivets, the
thickness of the main plates is 14 mm, that of the shorter upper cover
plate is 6 mm, that of the longer lower cover plate is 10 mm. Diameter
of rivet holes is 20.5 mm; long pitch is 100 mm; short pitch is 50 mm.

Compute the strength of a repeating section. Ans. P = 75.5 kN

1209. A triple-riveted butt joint like that in Fig. 12-2c has a long
pitch of 200 mm and a short pitch of 100 mm. Diameter of rivet holes is
26.5mm; thickness of main plate is 16 mm, and of each cover plate, 12
mm. Find the strength of a repeating section and the efficiency.
Ans. Efficiency = 86.4%
1210. A quadruple-riveted joint similar to that in Fig. 12-2d has
a long pitch of 350 mm. Diameter of rivet holes is 20.5 mm; thickness of
main plate, 10 mm; thickness of each cover plate, 8 mm. Determine the
strength of a repeating section and the efficiency.

1211. A quadruple-riveted butt joint like that in Fig. 12-2d has a

long pitch of 430 mm. Diameter of rivet holes is 32.5 mm; thickness of
main plate is 20 mm, and of each cover plate, 14 mm. Compute the
strength of a repeating section, using a factor of safety of 4, based on
the ultimate stresses of t = 300 MPa in single shear and t = 520 MPa
in double shear, ob = 660 MPa, and a, = 400 MPa. If this joint is the
longitudinal seam of a boiler carrying an internal pressure of 1.8 MPa,
and the girth joint is 50% efficient, what is the maximum allowable
boiler diameter? Ans. D= 2.03 m
12-5 Stresses in Riveted Joints 485

12-5 STRESSES IN RIVETED JOINTS

Sometimes it is necessary to investigate a joint and determine the

stresses caused by a given loading. The usual assumption is that each

rivet carries a load proportional to its resisting shear area. With this

assumption, the tensile load acting across any interior row of rivets is

found by subtracting the shear load transmitted by preceding rivet rows

from the applied load.

ILLUSTRATIVE PROBLEM

1212. A load of 144 kN acts on the repeating section of the

triple-riveted butt joint in Fig. 12-10. Length of sectionis 200 mm;

diameter of rivet holes is 23.5 mm; thickness of main plate is 14 mm,

and of each cover plate is 10 mm. Determine the shearing, bearing, and
tensile stresses developed in the joint.

Solution: The shear resisting area is that of one rivet in single shear
and four rivets in double shear, for a total of nine shear areas. The
average load transmitted by one shear area is therefore ^(144) =
16.0 kN. The average shearing stress is

3
/\ 16.0 X 10
T = = 36.9 MPa Arts.
2 2
ttJ /4 tt(23.5 X 10" 3 /4
)

Sketching a free-body diagram of any one of the rivets in rows 2

or 3, where the rivets are in double shear, will disclose that the shear
areas cause 32.0 kN main plate and that 16.0 kN acts
to act against the
against each cover plate. Since the combined thickness of two cover
plates is larger than that of the main plate, the maximum average

144 kN
14 mm t = 10 mm
Figure 12-10.
486

bearing stress occurs in the main plate and is given by

Ph 32.0 X 10
3

= 97.3 MPa Arts.

td (14 x 3
1(T )(23.5 x 1(T 3 )
The tensile stress in the main plate at row 1 is caused by the entire
loading acting across the net section of the plate:
3
Py 144 X 10
°' " - d)t
" x 10~ 3 ) - x 10" 3 )](14 x 10" 3
(p [(200 (23.5 )

= 58.3 MPa
The external load, reduced by the amount already transmitted by
the single shear area of the rivet in row 1, acts across the net section of
the main plate at row 2. The tensile stress at row 2 therefore is

P2 (144 - 16) X 10
3

°'~ - 2d)t " 10" 3 ) - 10" 3 )](14

(p [(200 X 2(23.5 x X 10" 3 )
- 59.8 MPa
The tensile stress at row 3 in the main plate need not be computed
because the net load acting there has been still further reducedby the
load transmitted by the rivets in row 2, and the net tensile area at row 3
is the same as that at row 2.
At row 3 in the cover plates, the lower cover plate carries the
greater load — that transmitted to it by 5 shear areas. Its average tensile
stress is

3
P 5(16 X 10 )

(p - 2d)t' [(200 x 10" 3 ) - 2(23.5 X 10" 3 )](10 X 10" 3 )

- 52.3 MPa
The greatest tensile stress in the joint therefore occurs in the main
plate at row 2 and is 59.8 MPa.

PROBLEMS

Compute the maximum shearing, bearing, and tensile stresses

developed in the following riveted joints under the action of the indi-
cated loads.

1213. Double-riveted lap joint in Problem 1203: load = 350 kN

per meter of length.

1214. Double-riveted butt joint in Problem 1205: load - 700 kN

per meter of length.
Ans. t = 37.7 MPa; ab = 99.3 MPa; a, = 60.1 MPa
1215. Triple-riveted lap joint in Problem 1204: load on a repeat-
ing section = 90 kN.
12-6 Structural Riveted Joints 487

1216. Double-riveted butt joint in Problem 1206: load on a

repeating section = 90 kN.
Arts, t = 41.5 MN/m 2
; ab = 109 MN/m2 ; a
t
= 58.1 MN/m2
1217. Triple-riveted butt joint in Problem 1209: load on a repeat-
ing section = 200 kN.
1218. Quadruple-riveted joint in Problem 1210: load on a repeat-
ing section = 220 kN.

12-6 STRUCTURAL RIVETED JOINTS

Structural riveted joints differ in several ways from those for pressure
vessels. The most important differences are as follows: (1) The entire
joint is considered because there generally no repeating pattern of
is

rivets. (2) Cover or splice plates are usually the same length because
there no need to shorten one plate to permit tighter caulking against
is

leakage. (3) Each rivet is assumed to transmit its proportional share of

the applied load. (4) The diameter of the rivet hole is taken as 3 mm
larger than the diameter of the undriven rivet.
For condition (3) to be true, it is necessary that the applied load
pass through the centroid of the rivet group (see page 7). Eccentric
loading of riveted connections is discussed in the next article. Condition
(4) results from the the parts to be joined are punched
fact that
separately, the diameter of the punched hole being 1.5 mm larger than
the rivet diameter. Another 1.5 mm is added to compensate for the
probable damage done to the metal around the hole by punching. For
shearing and bearing calculations, however, the diameter of the un-
driven rivet is used, because the rivet holes in plates that are punched
separately are unlikely to match perfectly and consequently the cross
section of the rivet will be less than that of the rivet hole.

ILLUSTRATIVE PROBLEM

1219. Using the structural method of assuming all rivets to carry

a proportional share of the load, compute the tensile stress in the main
plate at row 3 of the quadruple-riveted butt connection shown in Fig.
12-11. The connection transmits a load P = 360 kN across a plate
width p = 250 mm.
Also compute the width of the cover or strap plates
at row 2 if is not to exceed 100 MPa. Rivet diameter is
the tensile stress
19 mm; thickness of main plate is 14 mm, and of each cover plate, 8
mm.
Solution: Figure 12-11 shows the entire connection; there is no re-
peating pattern of rivets as in a boiler joint. As there are 10 rivets to
488 12 Riveted and Welded Connections

Figure 12-11.

-«-ef : Ipr 10
jj»4#frj
10^
Figure 12-12. Free-body diagram of section between rows 2 and 3.

carry the load, each rivet may be assumed to carry -£ of the load. Since

the single rivet in row 1 transmits -£ of the load to the cover plates, and
the two row 2 transmit -^ of the load, -£ of the load is left to
rivets in
tear the main row 3. This is shown in the free-body diagram of
plate at
a section between rows 2 and 3 in Fig. 12-12. Thus we obtain for the
tensile stress in the main plate at row 3, where 3 rivet holes reduce the
net section (note that the rivet hole diameter = rivet diameter 3 mm), -I-

[p = Ao = (p - 2>d)to ] t

-£(360 X 10 ) = [(250 X 10" ) - 3(22 x 10~ )](14 x 10" )a,

3 3 3 3

a, = 97.8 MPa Ans.

Figure 12-12 also shows that the cover plates at row 2 are subject
to jq of the applied load. Hence, the width of these cover plates
required to develop the maximum tensile stress permitted is

[P - Ac - (p - 2d)(2f)at ]
^(360 x 10 ) =[p - 2(22 X 10" 3 )][2(8 X 10" )](100 X
3 3 6
10 )

whence
p - 0.112 m = 112 mm Ans.

PROBLEMS

1220. Determine the safe load on the butt connection in Fig.

12-11 if the allowable stresses are t = 100 MN/m 2
, o, = 140 MN/m 2
,
12-6 Structural Riveted Joints 489

and ob = 280 MN/m 2

. Use 19-mm rivets; plate width p = 280 mm;
thickness of main plate = 14 mm and of each strap plate = 10 mm.
Ans. P= 506 kN
1221. In the joint in Illustrative Problem 12 19, if the rivet at row
1 is omitted, determine the maximum stresses in shear, bearing, and
tension in the main plate for a load of 260 kN. Also compute the
minimum width of the strap plates in rows 2 and 3 if the tensile stress is
limited to 100 MPa.

1222. Omitting row 4 of the butt connection shown in Fig.

if the allowable stresses are t = 90 MPa,
12-11, compute the safe load
a, MPa, and ab = 190 MPa. Use 25-mm rivets; plate width
120
p = 230 mm; thickness of main plate = 14 mm and of each strap
plate = mm.
10

1223. Two plates are joined by four rivets 25 in diameter, as mm

shown in Fig. P-1223. Find the allowable load P if the working stresses
are r - 70 MN/m 2
,a, = 100
2
MN/m
and ab = 140
,
2
MN/m .

Ans. P= 122 kN

4-
130
mm 150 mm
+
V 14 mm
Figures P-1223 and P-1224.
tt:
x = 10 mm
^2

1224. Repeat Problem 1223 assuming that the allowable stresses

are t = 100 MN/m 2
a, = 140
,
2
MN/m
and ob = 220 ,
2
MN/m .

1225. Find the safe load on the lap connection shown in Fig.
P-1225 if the rivets are of 19-mm diameter and the plates are 8 mm
thick. Use allowable stresses of t = 95 MPa, o = 140 MPa, and ab = t

220 MPa. Ans. P= 242 kN

Tt? ^ K&r

Figures P-1225 and P-1226.

490 12 Riveted and Welded Connections

1226. Repeat Problem 1225 assuming that the rivets are of 22-
mm nominal diameter and the plates are 10 thick. mm
1227. Two wide and 20
plates 250 mm thick are to be mm
connected by a riveted butt joint with two strap plates using rivets of
22-mm nominal diameter. The axial tensile load on the joint is 400 kN.
If the allowable stresses are r = 70 MPa, a, = 110 MPa, and ob = 130

MPa, determine: (a) the minimum number of rivets; (b) the minimum
rows of rivets and the best distribution of rivets in each row; (c) the

minimum thickness of each strap plate, using the distribution in part

(b). Arts, (c) / = 10.4 mm
1228. Solve Problem 1227, using 19-mm nominal diameter rivets
and allowable stresses of t = 1 10 MPa, o = 140 MPa, and ob = 220
t

MPa.

1 2-7 ECCENTRICALLY LOADED RIVETED CONNECTIONS

Occasionally it is impossible to load a riveted connection so that the

load passes through the centroid of the rivet group. Such a condition is

called eccentric loading; the load is not distributed equally over all the
rivets (see Fig.12- 13a). However, by adding a pair of equal, oppositely
directed, and collinear forces of magnitude P (shown dashed) at the
centroid of the rivet group, the applied eccentric load P is replaced by a
central load P and the torsional couple T= Pe, as shown in Fig.
12-13b.
The P is resisted equally by the direct
effect of the central load
load Pd = P/n on each of the n rivets, as shown by the free-body
acting
diagram of the plate in Fig. 12- 14a. The torsional couple T is resisted
by torsional loads P (Fig. 12- 14b), which act perpendicular to the
t

radius p from the centroid of the rivet group and vary directly with the
distance of the rivets from the centroid. To determine the torsional load
on any rivet, we may consider that the connection is equivalent to a
flanged coupling consisting of three concentric circles of rivets, and use

(a) (b)

Figure 12-13. Eccentrically loaded riveted connection.

12-7 Eccentrically Loaded Riveted Connections 491

^V sN- ^V

X v3

(a) Equal direct rivet loads (b) Distribution of torsional loads

(c) Resultant rivet loads

Figure 12-14. Analysis of eccentrically loaded connection.

the method outlined in Art. 3-3 (see page 87). Then the resultant load
on any rivet is the vector sum of the direct and torsional loads on that
rivet, and appears as shown in Fig. 12- 14c.

A better method of determining the torsional load is to apply the

torsion formula t = Tp/J. Here r represents the average shearing stress
on any rivet, p is the radial distance from its center to the centroid of
the rivet group, and / may be expressed as

J = 2V 2
(a)

Since all same area A, and since p may be expressed

the rivets have the
in terms of the x and y coordinates of any rivet so that p 2 = x 2 + y 2
(see Fig. 12- 14b), we may rewrite Eq. (a) as

/ = A&x 2
+ 2y 2 ) (b)

whence the torsion formula becomes

T = (c)
A(2x 2 + 2y 2 )
Transposing A to the left side of this equation determines the torsional
492 12 Riveted and Welded Connections

load P, on any rivet from P =

t
At, so we finally obtain

(12-5)

The resultant load on a typical rivet is obtained as the vector sum

of Pd and Pr See Fig. 12- 14c. This vector addition is easily performed
analytically by resolving Pd and P into x and y components. The
t

components Pd and Pd of the direct load are constant for all rivets. The
components of the torsional load P are obtained by observing from Fig.
t

12- 14b that the angle a between the radius p and the X axis equals the
angle between P, and the Y axis; hence

= P t
sin a = P-
t
P
and
x
— P, cos a = P— t

P
since sin a = y/p and cos a = x/p. Replacing P t
in these relations by
its value from Eq. (12-5), we obtain

Z* 2 +2/ (12-6)
T
Ix 1 +2/
where x and y are the coordinates of a rivet measured from the centroid
of the rivet group.
The maximum load on any rivet occurs when Pdx and maximum
Ptx
Pd and maximum P, are additive, as at the upper right
as well as
corner; whence the resultant rivet load is found from

'^=V^ + p,y (^ + pj (12-7)

The use of these equations is illustrated in the following illustrative

problem.

ILLUSTRATIVE PROBLEM

1229. On the connection of 12 rivets shown in Fig. 12-15, the

load P = 200 kN passes through the center of rivet C and has a slope of
4/3. Determine the resultant load on the most heavily loaded rivet.

Solution: The effect of the applied load is equivalent to an equal

central load acting through the centroid of the rivet group plus a
12-7 Eccentrically Loaded Riveted Connections 493

p,+p
y 'y
i

i i
V
r
rh .
Pd + i>
^J TA

100
mm "

'
t
Px = 120 kN
, rh t
, i
V
ciT 9 <j
**
X
1
K
mm
100
\
\
•

t
Py = 160 kN 3^
'

V r ~\ 7~ -X7- -XT- \\
P=200 kN
-<-80 mm -> *- 80 mm-*- *- 80 mm-*-

Figure 12-15.

torsional couple equal to the moment of P about the centroid of the

rivet group. Replacing P by its components Px = 120 kN and i^ =
160 kN, and noting that the moment of P is equal to the moment sum
of its components, we find that the torsional couple is

-3
r- (160 x \&)(\20 x 10 ) = 19.2 kN-m
Before applying Eq. (12-6) we compute the value of 2x 2 and 2>> 2 .

There are six rivets whose x coordinate is 40 mm and six rivets whose x
coordinate is 120 mm. Also there are eight rivets whose y coordinate is

100 mm. Therefore,

2 2 2
Sx 2 + 2>>
2
- [6(40) + 6(120) ] + 8(100)
= 0.176 x 10
6
mm 2 = 0.176 m2
Applying Eq. (12-6) gives the maximum components of the tor-
sional load as
3
= 19.2 x 10
*<.- P. (100 x 10" 3 )
2x 2 + 2r 0.176
= 10.9 kN
3
= 19.2 x 10
P. (120 x 10" 3 )
2
* 2* 2 + 2>> 0.176
- 13.1 kN
The x and y components of the direct load on any rivet are found
by dividing the x and y components of the applied load by the number
494 12 Riveted and Welded Connections

of rivets. Thus

,< _£_J 2 x»>_ lft< , kN

and
J
160 X 10
r4;--*- = 13.3 kN
12

The most heavily loaded rivet is at A, where the maximum

components of the direct and torsional loads are additive as shown.
Applying Eq. (12-7), we have

2
rr-^(r<. + r, )
M
+ (*<, + *
2 2
= y(iaO + 10.9) + (13.3 + 13.1) = 33.7 kN Ans.

PROBLEMS

1230. Compute the resultant load on the least loaded rivet in

Illustrative Problem 1229.
1231. A gusset plate is riveted to a larger plate by four 22-mm
rivets arranged and loaded as shown in Fig. P— 1231. Determine the
maximum and minimum shear stress developed in the rivets.
Ans. Max. r = 47.4 MPa; min. r = 29.4 MPa

P-40kN
Figure P-1231.

1232. In the gusset plate connection shown in Fig. P-1232, each

rivet has a cross-sectional area of 300 mm 2
.The allowable load P was
designed for a shearing stress of 70 MPa. What will be the maximum
shearing stress in the rivets if the rivet at A is improperly driven so that
it cannot carry any load?
12-7 Eccentrically Loaded Riveted Connections 495

80 ^u 80 80 80
mm »,. mrn_.
i„ „|«
nmniTmm 1

Figure P-1232.

1233. If the maximum load permitted on any single rivet in the

connection shown in Fig. P-1233 is 15 kN, compute the safe value of P.
Arts. P = 37.9kN

Figures P-1233 and P-1234.

1234. Repeat Problem 1233 assuming that the upper left rivet is

improperly driven so that it does not carry any load.

1235. Rivets 22 mm
in diameter are used in the connection
shown in Fig.P-1235. If P = 90 kN, what thickness of plate is required
so as not to exceed a bearing stress of 140 MPa? Ans. t = 9.90 mm

Figures P-1235, P-1236, and P-1237.

.-
496 12 Riveted and Welded Connections

1236. For the riveted connection in Problem 1235, compute the

load P which will cause a maximum rivet load of 20 kN if the lower
right rivet is removed. Ans. P = 55.6 kN
1237. Resolve Problem 1235 assuming that the load P is replaced
by a 90-kN load which passes through the center of the upper right rivet
with a slope of 3 vertically upward to 4 horizontally to the right.

1238. In the gusset plate connection shown in Fig. P-1238, if

P 60 kN, determine the shearing stress in the most heavily loaded of

the four 22-mm rivets. Ans. r = 70.0 MN/m 2

Figure P-1238.

1239. For the connection shown in Fig. P-1239, determine the

shearing stress in the most heavily loaded of the three 22-mm rivets.

Ans. t = 159 MPa

80 mm 80 mm
X^jC

90 mm
o-
100 mm

60 kN ^
^4
Figure P-1239.

1240. For the riveted connection in Fig. P-1240, determine the

allowable load P if the shearing stress in the 25-mm rivets is limited to
140 MN/m 2
.
 12-8 Welded Connections 497

120 mm

Figure P-1240.

12-8 WELDED CONNECTIONS

The reliability of welded connections has increased to the point where

they are used extensively to supplement or replace riveted or bolted
connections in structural and machine design. It is frequently more
economical to fabricate a member by welding simple component parts
together than to use a complicated casting.
Welding is a method of joining metals by fusion. With heat from
either an electric arc or an oxy acetylene torch, the metal at the joint is
melted and fuses with additional metal from a welding rod. When cool,
the weld material and the base metal form a continuous and almost
homogeneous joint. To protect the weld from excessive oxidation, a
heavily coated welding rod is used which releases an inert gas that
envelopes the arc stream; this technique is called the shielded arc
process*
The two and fillet welds
principal types of welds are butt welds
(see Fig. 12-16). The strength of a butt weld is equal to the allowable
stress multiplied by the product of the length of the weld times the
thickness of the thinner plate of the joint. The allowable stress is taken
to be the same as that of the base metal.
The strength of side or transverse fillet welds is assumed to be
determined by the shearing resistance of the throat of the weld regard-
less of the direction of theapplied load. In the 45° fillet weld in Fig.
12-17, with leg equal to t mm, the shearing area through the throat is
the length L mmof the weld times the throat depth, or A — Lt sin 45°
= L(0.707f) mm
2
The allowable stresses for fillet welds specified by the
.

*A complete description of the welding process may be found in most books

on structural design. See, for example, B. Bressler, T. Y. Lin, and J. B. Scalzi,
Design of Steel Structures, 2nd ed., Wiley, New York, 1968.
498 12 Riveted and Welded Connections

P P ^ P

(a) Single-V butt weld (b) Double-V butt weld

Tr-

(d) Transverse or end

fc) Side fillet weld fillet weld

Figure 12-16. Types of welds.

Throaty
t^ uf

Figure 12-17.

AISC (based upon recommendations of the American Welding Society)

depend upon the electrode used in the welding process and upon the
grade of steel being welded. For example, if E-70 electrodes are used to
weld A36 steel (one of the more common grades of structural steel used
today), the allowable shearing stress is 145 MPa.* For this case, the
strength of the 45° fillet weld, in newtons, is

P m TA - (145 X 6
10 )(0.707fL X 10" 6 ) - 103 tL

Usually, however, the strength of a fillet weld is expressed in terms of

the allowable force q per millimeter of weld length, given by
p
q = -r= 103/ N/mm
where, to repeat, t is expressed in millimeters.
As a rule, special precautions are necessary to ensure that the
length of the leg of a fillet weld along an edge is actually equal to the
thickness of the edge. One reason is that edges of rolled shapes are
rounded and the length of the leg would be less than the nominal
thickness of the shape. Another reason is that, during welding, the
corner of the edge may melt into the weld, which would reduce the

•Numerical values for stresses and lengths cited in this section are ap-
proximate SI conversions of the specifications in Manual of Steel Construction,
American Institute of Steel Construction, New York, 1973.
12-8 Welded Connections 499

fci f-2

(a) t < 6 mm (b) t > 6 mm

Figure 12-18. Maximum size of fillets.

length of the leg. Therefore, AISC specifications require that the maxi-
mum size of a fillet weld should be 2 mm less than the material
thickness along edges 6 mm or more thick. For edges less than 6 mm
thick, the maximum weld may equal the edge thickness.
size of the
These specifications are illustrated in Fig. 12-18. Weld sizes may exceed
these specifications if the designer stipulates that the weld is to be built
out to obtain full throat thickness.

ILLUSTRATIVE PROBLEM

1241. A 100 X 100 X 10 mm angle is to be welded to a gusset

plate. The angle carries a load of 190 kN applied along its centroidal
axis, (a) Determine the lengths of side fillet welds required at the heel
and toe of the angle. The weld at the heel is to be the same size as the
maximum permissible weld at the toe. (b) Determine the lengths of
the side fillet welds if a transverse fillet weld is added at the end of the
angle. Assume that the allowable shearing stress through the throat of
each weld is 145 MPa.

Solution:

Part a. Figure 12-19 shows the forces that keep the angle in
equilibrium: P and P2
x
are the resisting forces exerted by the welds at
the heel and toe, respectively. assumed to They are act along the edges
of the angle. Taking moments about a center on the line of action of P2 ,

-> L2
r i

p*
71.3 mm
1 190 kN

—r mm
28.7

h i

<-L -**

Figure 12-19.
500 12 Riveted and Welded Connections

wc obtain
[1MPi - 0] 100/*, = 190(71.3) />, = 135.5 kN
With respect to a moment center on the line of action of P v we have

[ZA/,, - 0] 100/> 2 = 190(28.7) P2 = 54.5 kN

Since the thickness of the angle is 10 mm (which is greater than
6 mm), the maximum size of the fillet weld at the toe of the angle is

10 - 2 = 8 mm. The strength per millimeter of this size weld with

t = 145 MPa is q = 103/ = 103(8) - 824 N/mm. Hence the required
lengths of weld are
3
T
135.5 x 10
o24

and
3
54.5 x 10
r
2
= 894 = mm
These values should be increased by a small amount to provide for
starting and stopping of the weld.

Part b. If a transverse fillet weld is added along the edge at the

end of the angle, its maximum size is again 8 mm. This weld should be
symmetrically applied about the action line of the applied load in order
to avoid eccentricity of loading. As shown in Fig. 12-20, this limits the
transverse weld to twice the dimension of 28.7 mm, and this length will
sustain a force of qL — (824)(2 X 28.7) = 47.3 kN acting at its mid-
point.*
Taking moments first about a center on the line of action of P2
and then about a center on the line of action of P v we obtain

[2MPi = 0] \O0P = (190

x
- 47.3)(71.3) P, = 101.7 kN
[2MPi = 0] 100^ = (190 - 47.3)(28.7) P2 = 41.0 kN

Thus, the lengths of weld required at the heel and toe are respectively:
3
? 101.7 x 10
L = L — x
^r-
824
= 123 mm Ans.
1
3

L2 =
41.0
—X 10
= 49.8 mm Ans.

A larger size of weld may be used at the heel if it is necessary to reduce

the length L x
.

•It is not uncommon, however, to let the transverse fillet weld run completely

across the leg of the angle and ignore the resulting eccentricity of loading.
Eccentrically loaded welded connections are discussed in the next article.
12-8 Welded Connections 501

28.7 mm !

47.3 kN -^ 190 kN
28.7 mm

4-T

PROBLEMS
1242. A plate 150 mm wide and
14 mm
thick is lapped over and
welded to a gusset Determine the minimum length of an 8-mm
plate.
side fillet weld that
will be necessary if the plate is subjected to an axial
stress of 140 MPa. Use an allowable shearing stress through the throat
of the weld of 145 MPa. Ans. 179 on each side mm
1243. Solve Problem 1242 using the maximum size of welds
permitted.

1244. A 150 X 100 X 13 mm angle is to be welded to a gusset

plate with the 150-mm leg against the' plate. If the angle carries a
centroidal load of 400 kN, what lengths of 8-mm side fillet welds will be
required along the toe and heel of the angle. Assume the allowable
shearing stress through the throats of the welds is 145 MPa.
Ans. 161 mm and 324 mm
1245. Solve Problem 1244 using a 12-mm weld at the heel of the
angle and the maximum permissable size at the toe.

1246. A 16-mm plate is lapped over and secured, as shown in

Fig. P-1246, by transverse fillet welds on the inside and outside to form
a penstock 1.5 m
in diameter. Determine the safe internal pressure,
assuming allowable stresses of a, = 160
2
MN/m
for the plate and
r = 120 MN/m 2
through the throats of the welds. Use the maximum
size of welds permitted. Ans. p = 3.17
2
MN/m

Figure P-1246.
502 12 Riveted and Welded Connections

1247. A tank is fabricated by welding two caps, as shown in Fig.

P-1247, to the ends of a cylinder 1.2 m in diameter. If the caps and
cylinder are 10 mm
thick, determine the safe internal pressure that will
not exceed a shearing stress of 110 MPa in the throat of the maximum
size of fillet weld around the entire circumference.
Arts, p = 2.07 MPa
\\v^\\\\\^v^^yv,v^^v^YrT^q
f^ ^
1.2m

Figure P-1247.

1 2-9 ECCENTRICALLY LOADED WELDED CONNECTIONS

In the preceding article the methods of static equilibrium were used to

determine the force to be resisted by a weld. This analysis was based on
the concept that the weld is uniformly loaded along its length. This
assumption is reasonable if all welds are of the same size and if the
applied load passes through the centroid of the weld lines. If the
resultant load P does not pass through this centroid, the welds are not
uniformly loaded per millimeter of length. This causes a variable elastic
deformation to exist in the welds. The following simplified analysis
shows how to determine the maximum intensity of loading per millime-
ter of weld on which the size of the weld is based.

Proceeding as in the case of eccentrically loaded riveted connec-

tions, we add a pair of equal, oppositely directed, and collinear forces P
(shown dashed in Fig. 12-2 la) at the centroid C, of the weld lines,
thereby reducing the eccentric load P to a central load P and a torsional
couple T= Pe. In Fig. 12-2 lb, the central load P is resisted by the
direct force qd per millimeter of weld. This direct force is uniformly
distributed over all the welds and is given by

<ld
= P (12-8)
2Z.

where 2L is the total length of all the welds.

In Fig. 12-21c, the torsional couple is resisted by a variable
torsional force qt per millimeter of weld. Assuming elastic action of the
welds, but that the plate is rigid and twists about the centroid C, we
may find the intensity of this torsional force by applying the torsion
formula with a modified value of J.
12-9 Eccentrically Loaded Welded Connections 503

W
(a)

/I Pe

(b) Direct force q \ (c) Torsional force q

t

Figure 12-21. Analysis of eccentrically loaded welded connections. Part (a) is the
vector sum of parts (b) and (c).

To determine /, consider the arrangement of straight weldsshown

in Fig. 12-22. For any weld of length L, the centroidal value / is the
sum of the rectangular moments of inertia with respect to axes through
its center directed along and perpendicular to its length. These values
are respectively zero and -^L?. Applying the transfer formula, we
obtain, with respect to the centroid C of the weld group,
[J = J + Ld 2 ]

J = ^L 3
-I- Lp 2
- ^L 3 + L(x 2 + y 2 )

Repeating this computation for every weld in the connection and

adding the results, we find that the modified J of the torsion formula
becomes

/ = 2L(-^L 2 + x 2
+y 2
)

A-
^y
^ X I

Figure 12-22. Evaluation of J.

504 12 Riveted and Welded Connections

Applying the torsion formula gives the torsional force q acting t

perpendicular to the radial location p of any point on a weld as

Tp
(12-9)
2L(±L + 2
x 2
+ y 2)
More useful are the following expressions for the components of qn
obtained as in the analysis of riveted connections:

Ty
Vt.
-
2L(^L + 2
x 2 + y 2)
(12-10)
Tx
**>
2L{±L 2 + x 2 + y 2)

inwhich x and y are the coordinates of any selected point of any weld.
The maximum intensity of the weld force occurs at the point
where qdx and maximum q as well as q^ and maximum q, t
,are ,

additive. Combining these values vectorially gives

*=Vte* + *.
)2 + ta* + **) (12-H)
Frequently this value of q used to determine the size of all welds, but
is

occasionally the size of each individual weld is based on the value of the
highest stressed point in that weld.

ILLUSTRATIVE PROBLEM

1248. A plate is attached to the frame of a machine by two side

fillet welds as shown in Fig. 12-23a. Determine the size of the welds to
resist a vertical load of 40 kN. Use an allowable shearing stress of 145
MPa through the throats of the welds.

P=40kN
<lt
P
[100 100 mm -^
D \

T 1
100
60 mm
Imm -£]___ 4-
^_ 85 1_ 65 !
40 mm
j
mm r mm -I
B 150 mm B\ A

(a) (b)

Figure 12-23.
 :

12-9 Eccentrically Loaded Welded Connections 505

Solution: The centroid of the weld lines, with respect to an origin at A,

is found to be

[Lx = S/jc] (150 + 100)3c = 150(75) + 100(50) x = mm

65

[Ly = 2fy] 250^ = 100(100) y = 40 mm

Using these values, we locate the centroid C as shown in Fig. 12-23a.
The moment of P about this centroid determines the torsional couple to
be

T = Pe = 40(100 + 65) - 6600 kN • mm

The modified / for the weld group is the sum of the / values for
each weld. Remembering that x and y are the coordinates of the center
of each weld relative to the common centroid C, we obtain

[/-L(aL» + **+**)]
(150)'
•/.»
1

AB = 150
12
+ (10)
2
+ (40)
:
= 0.536 X 10
6
mm 3

2
poo)
nr = 100
JDE
12
,

(15)
2
+ (60)
:
- 0.466 X 10
6
mm 3

Their sum determines the modified J to be

/ = 2/ = (0.536 -I- 0.466) x 10

6
- 1.00 x 10
6
mm 3

The components of the direct load are

3

to.
" P 40 X 10
= 160 N/mm T and qd =
2L 250

These values are to be combined with the components of the torsional

A and E. These are the highest stressed points in the welds AB
forces at
and DE as revealed by an inspection of Fig. 12-23b.
Applying Eq. (12-10), we obtain
3
Ty (6600 X 10 )(60)
«t.
= AtE: 4k
"
6
1.00 x 10

- 396 N/mm <-

3
(6600 x 10 )(40)
At A 1u
"
6
1.00 X 10

= 264 N/mm
3
Tjc (6600 x 10 )(65)
* - At £ and ^ * " 6
1.00 x 10

- 429 N/mm t
506 12 Riveted and Welded Connections

Combining the direct and the torsional force components shows

that the highest stressed values in the welds are

2 2
-V(z*J + (H)
g£ »^(396)
2
+ (160 + 429)
2
= N/mm
710

qA ~^/(264)
2
+ (160 + 429)
2
= 645 N/mm
We now apply the AISC specification that the allowable force per
millimeter of weld for r = 145 MPa is 103/ (regardless of the direction
of the force), where / is the length of the leg of the weld in millimeters.
Hence the size of the welds, based on the highest stressed point, is

[q - 103/] qE - 710 = 103/ / - 6.89 mm

which requires the use of 7-mm welds.
If desired, a slightly smaller weld may be used in AB, based on
qA - 645 N/mm.

PROBLEMS

1249. A bracket is welded to the frame of a machine as shown in

Fig. P-1249. Determine the size of fillet weld to be specified to the
nearest millimeter. Use t = 145 MPa through the throats of the welds.
Arts. / = 5 mm

100 mm ^ P=80kN

r A.
r
£
X
X
X

X
X
B X
1
'
o
m
X

CM
X

L X
X

Figure P-1249.

1250. A
plate is lapped over and welded to a gusset plate as
shown P-1250. Determine the size of fillet welds to be specified
in Fig.
to the nearest millimeter. What maximum value of P can be applied
vertically as shown if 8-mm welds are used? Use r = 145 MPa through
the throats of the welds.
 12-9 Eccentrically Loaded Welded Connections 507

P-90kN

) U-150 mm— — 100^

T
1

150
mm
1
B
F
^i^-4^150
mm i
mm- A
Figures P-1250, P-1251, and P-1252.

1251. Referring to Fig. P-1250, find the maximum force per

millimeter of weld if a transverse fillet weld is added along the entire
150-mm width of the plate at A. Ans. q = 531 N/mm
1252. In Fig. P-1250, transverse fillet welds are added along the
entire 150-mm width of the plate at both A and F. Determine the
maximum force developed per millimeter of weld.

1253. An angle is welded to a plate to resist a load P acting

through its centroid. The specified lengths of 8-mm welds are as shown
at (a) in Fig.P-1253, but a welder applies them as shown in (b). Using
the design load determined from (a), find the maximum load per
millimeter of weld in (b), assuming elastic action to occur only in the
welds. Use t = 145 MPa through the throats of the welds.

U75*
|
mm r U-150 mm—
j
t
100 mm \ 100 mm

50 mm 50 mm
|*150 mm— T \
1-150 mm- t

(a) (b)
Figure P-1253.

1254. Solve Illustrative Problem 1248 assuming that a transverse

weld is added along the entire 100-mm width of the plate at A.
Ans. Max. q = 391 N/mm; use 4-mm welds

SUMMARY

There are two classes of riveted joints: those used for boilers and
those used in structures. In the former, the diameter of the rivet hole
 12 Riveted and Welded Connections

determines the diameter used for shearing and bearing calculations; in

the latter, the actual rivet diameter is used for shearing and bearing, the
rivet hole being assumed to be 3 mm larger.

Further differences between these two types of riveted joints are

given by the methods used to compute their strength. For boiler joints,
it is preferable to begin with the strength of a single rivet in shear and
bearing, depending upon the conditions in each row of rivets,
and then,
compute the governing load per rivet in each row. The sum of these
governing loads is the total rivet strength. The tensile strength of the
main any row is assumed to be increased by the strength of the
plate in
rivets in the rows between that row and the externally applied load. In
structural joints, all the rivets are assumed to carry their proportional
share of the load. In either type of joint, the effect of frictional
resistance is neglected.
Occasionally it is impossible to have the load on a riveted connec-
tion pass through the centroid of the rivet group. If the eccentric load is

replaced by a central load P and a torsional couple T— Pe, the

resultant load on any rivet is the vector sum of the direct rivet load
Pd - P I n (n being the number of rivets) and a torsional rivet load Pt

expressed by

2jt +z,y
However, the resultant rivet load Pr is more conveniently determined by
combining the components of the direct rivet load and the torsional
load. The components Pd and Pd are constant for all rivets, and the
components of the torsional load are given by

(12-6)

P = I
''
2x 2 +2/
where x and y are the coordinates of a rivet measured from the centroid
of the rivet group.
The maximum rivet load is found by combining the components
of the direct load with the maximum additive components of the
torsional couple in accordance with

/ 2 2
-7>
< + pJ + + /v)
/>
(^ < 12
^=V (
12-9 Eccentrically Loaded Welded Connections 509

The length of a weld determined by computing the resisting

is

force required to satisfy the conditions of static equilibrium,and then

dividing the resisting load by the strength of the weld in newtons per
millimeter. AISC specifications for weld strength depend upon the
electrode used in the welding process and upon the grade of steel being
welded. The maximum size of a fillet weld is 2 less than the mm
material thickness along edges 6 mm or more thick. For edges less than
6 mm thick, the maximum size of the weld may equal the edge
thickness.
Eccentrically loaded weld groups, in which the applied load does
not pass through the centroid of the weld lines, are treated in a similar
manner as are eccentrically loaded riveted connections. The eccentric
load is replaced by a central load P passing through the centroid of the
weld lines and a torsional couple T equal to the moment of P about this
centroid. The intensity of loading is then determined as the vector sum
of the uniform direct loading qd = P/2L
(2L being the total length of
allwelds) and a torsional loading qt expressed by

(,2 -'»
•-iZSIiTFT?) .

in which p is the radial distance from the centroid of the weld lines to
any point on the weld group, and x and y are the centroidal coordinates
of the center of each weld line.
Usually it is more convenient to determine the components of the
torsional loading as expressed by

** 2L(£L 2 + x 2 + y 2)
(12-10)
Tx
2L(±L 2 + x 2 + y 2)

inwhich x and y are the coordinates of any selected point of any weld.
The maximum intensity of the weld force occurs at the point
where qd and maximum q as well as qd and maximum q.
t
, are ,

additive. Combining these values vectorially gives

- !1 >
« - V(«t + 9'f + <«<-
+ I'f ( 12
 Special
Topics

13-1 INTRODUCTION

The preceding chapters have dealt with topics that comprise the usual
undergraduate course in strength of materials, but the subject does not
end there. We now consider briefly some additional topics* that prop-
erly belong in an advanced course in strength of materials; some, such
as photoelasticity and finite element methods, constitute a field of study
in themselves. Each of the topics is complete in itself so that all or any
of them may be studied independently of the others.

13-2 REPEATED LOADING; FATIGUE

Many machine parts are subjected to varying stresses caused by

repeated loading and unloading. Parts subjected to such loading

•For extended discussions of these topics, see one or more of the following
books: A. P. Bores, O. N. Sidebottom, F. B. Seely, and J. O. Smith, Advanced
Mechanics of Materials, 3rd ed., Wiley, New York, 1978; S. Timoshenko, Strength of
Materials, Vols Iand II, Van Nostrand Reinhold, New York; 1955, G. Murphy,
Mechanics of Materials, McGraw-Hill, New York, 1948.

510
13-2 Repeated Loading; Fatigue 511

frequently fail at a stress much smaller than the ultimate strength de-
termined by a static tensile test. Failures of this type are known as
fatigue failures. In order to properly design members that are subjected
to stress reversals, it is necessary to know the stress that can safely be
carried an indefinite number of times (or a somewhat higher stress that
can be carried safely for a limited number of reversals, as when a
machine is used only occasionally and may therefore have a long life).
Testing to determine these values is called fatigue testing. The
simplest method involves reversed bending. In it, a round specimen S is
mounted in bearings, as shown in Fig. 13-1, and subjected to bending
couples by the load W. As the motor M
rotates the specimen, a fiber
that was originally on top passes from compression to tension and back
to compression, thereby undergoing a complete reversal of stress for

each revolution. A revolution counter, R, registers the number of

revolutions until failure occurs. The motor then stops automatically. In
testing a given material, about a dozen identical specimens are prepared
and rotated machine, each at a different load W, until failure
in the
occurs or until four to five million cycles have been recorded. A typical
result obtained in this way is shown in the o—N diagram in Fig. 13-2,
where stress versus number of cycles is plotted with semilogarithmic

M
K2
vv;;>;///////v;;/
SzzzzzzzzS s/;;;;;/777i ] V/////////7Z. [
I

W
Figure 13-1. Rotating beam fatigue-testing machine.

290

260

230
A
200

CG
170

140
3 4
10 10 10 5 10 6 10-
Cycles of stress reversals for failure (N)

Figure 13-2. a-N diagram.

512 13 Special Topics

scales. The point at which this diagram* levels off is called the en-
durance limit, denoted by ae Although no definite relation exists be-
.

tween the endurance limit and the ultimate strength, tests show that the
endurance limit is between 40 and 50% of the ultimate strength.
When a ductile steel specimen is subjected to a gradually increas-
ing load, yielding of the specimen is evident considerably before actual
failure occurs. But a specimen of the same material that is subjected to
stress reversals fails suddenly, without any plastic deformation or any
other warning. Thus the fatigue failure of ductile steel is similar to the
static failure of a brittle material.
At first it was thought that repeated applications of the load
changed the crystalline structure of the material, but we now know that
this is not true. is explained more satisfactorily by the
Fatigue failure
localized stress theory, which is based on the stress concentrations that
occur (a) inside a material because of discontinuities in the material
itself and (b) at the surface of a material because of abrupt changes in

section. These stress concentrations are not serious when a ductile

material is subjected to a static load; but when the load is repeatedly
applied, they cause minute cracks which spread with each repetition
until the member suddenly fractures.
For example, a flat steel bar with a small central hole is loaded by
axial tensile forces. As shown in the next article, the stress at the edge of
the hole is three times the stress across the full section. However, if this
bar is subjected to a gradually increasing static tensile load, it will
sustain practically the same ultimate load as a similar bar without a
hole. The apparent insignificance of the stress concentration at the hole
is due to the fact that although the material there does yield at one-third
of the load at which the whole bar yields, this yielding is purely local
and the stress at the edge of the hole remains constant during it. With
increasing load, the whole bar finally yields; then the effect of a small
hole on the strength of the bar negligible. But if two bars, one solid
is

and the other with a small were subjected to a tensile fatigue test,
hole,
the failure load of the bar with the hole would actually be about
one-third that of the solid bar.

13-3 STRESS CONCENTRATION

As we said in the preceding article, the effect of stress concentration on

ductile steel subjected to repeated loading is similar to its effect on a
brittle material subjected to static loading. We consider now the effect
of abrupt change in section upon the stress distribution. In Fig. 13-3, a

•A log-log plot may be used for this diagram, but a Cartesian plot does not
show the endurance limit so clearly.
13-3 Stress Concentration 513

'
tttfttttttltt

JJIlJHHJIi*
Figure 13-3.

small circular hole in a rectangular plate subjected to a uniform tensile

stress a causes a stress distribution across a section through the hole as
shown by the shaded area. This stress distribution is expressed by

, W. d1 3 d4 \
(a)
2 \ 4r 2 16 r 4
)

where d is the diameter of the hole and r is the distance from the center
of the hole.* From Eq. (a) we find the stresses at points m and n to be
3a. Because of bending action around the hole, compressive stresses of
magnitude a are created at the top and bottom points of the hole.
A similar stress concentration is caused by the small elliptical hole
shown in Fig. 13-4. The maximum stress at the ends of the horizontal

tlUlHHllija

WWWW °

Figure 13-4.

See S. Timoshenko and J. N. Goodier, Theory of Elasticity, 2nd ed.

McGraw-Hill, New York, 1951, p. 81.
514 13 Special Topics

axis of the hole is given by

.-M
This stress increases with the ratio b/a; hence a very high stress
(b)

concentration is produced by a narrow hole or crack perpendicular to

the direction of the tensile stress and therefore such cracks tend to
spread. This spreading may be stopped by drilling small holes at the
ends of the crack, thus replacing a high stress concentration by a
relatively smaller one.
Small semicircular grooves in a plate (Fig. 13-5) produce stress
concentrations at point m and n that are about three times the average
stress o applied at the ends of the plate.

ttmttmtH *

Figure 13-5.

Values* of stress concentration factors for several other cases of

abrupt change in section are listed in Table 13-1. If A: denotes the factor
of stress concentration, the maximum stresses for axial, torsional, and
flexural loads are given by

= Mc
a = A:—, °f
k
A /

Factors of stress concentration for repeated loading are sometimes

but not always lower than the theoretical values for static loading given
in Table 13-1. In alloy steels and quenched-carbon steels, these theore-
tical values are recommended for repeated loading; somewhat lower

•These values are taken from a more complete tabulation for these and other
cases on pp. 382-406 of R. J. Roark's Formulas for Stress and Strain, 4th ed.,
McGraw-Hill, New York, 1965. For general expressions of these stress concentra-
tion factors, see pp. 590-603 of the 5th ed. of the same book by R. J. Roark and W.
C. Young (1975).
 1 3-3 Stress Concentration 515

TABLE 13-1. Stress Concentration Factors

I. Square shoulder with fillet in rectangular bar

¥
L
r/d
(a) Tension h/r 0.05 0.10 0.20 0.27 0.50 1.0

0.5 1.70 1.60 1.53 1.47 1.39 1.21

1.0 1.93 1.78 1.67 1.59 1.42 1.22
1.5 1.89 1.72 1.65 1.43 1.23
2.0 1.95 1.80 1.70 1.44 1.23

- r/d
(b) Bending h/r 0.05 0.10 0.20 0.27 0.50 1.0

0.5 1.61 1.49 1.39 1.34 1.22 1.07

1.0 1.91 1.70 1.48 1.38 1.22 1.08
1.5 2.00 1.73 1.50 1.39 1.23 1.08
2.0 1.74 1.52 1.39 1.23 1.09

II. Square sho ulder wit l fillet in cii cular s haft

A i )— » ^5:
\'/yj —y r

(a) Tension: Approximately same as Case 1(a)

(b) Bending: Approximately same as Case 1(b)

r/d
(c) Torsion D/d 0.005 0.01 0.02 0.03 0.04 0.10

2.00 — 3.0 2.25 2.00 1.82 1.44

1.33 — 2.7 2.16 1.91 1.76 1.40
1.20 3.00 2.5 2.00 1.75 1.62 1.34
1.09 2.20 1.88 1.53 1.40 1.30 1.15

(Continued)
516 13 Special Topics

TABLE 13-1. (Continued)

III. Semicircular notch in circular shaft.

(a) Tension 0.05 0.15 0.30 0.40 0.52 0.75

D- 2r
k 2.57 2.16 1.81 1.65 1.51 1.36

r
(b) Bending 0.05 0.10 0.20 0.30 0.50 0.75
D - 2r
k 2.20 1.86 1.59 1.45 1.30 1.18

= 2D
(c) Torsion: k
D+ 2r

values are permissible for carbon steels not quenched. The fatigue factor
of stress concentration also varies with the size and type of material.*

13-4 THEORIES OF FAILURE

Various theories of failure have been proposed, their purpose being to

from the behavior of a material subjected to simple tension or
establish,
compression tests, the point at which failure will occur under any type
of combined loading. By failure we mean either yielding (resulting in
excessive permanent deformation) or actual rupture, whichever occurs

•A more complete discussion of fatigue will be found in Prevention of

Fatigue of Metals, Battelle Memorial Institute, published by Wiley. See also
Timoshenko's Strength of Materials, Vol. II, Arts. 78 to 81, for a discussion of
fatigue, and his Chap 8 for stress concentration. A fairly complete bibliography on
the subject of stress concentration is given in Roark and Young's Formulas for
Stress and Strain, 5th ed., pp. 604-606.
13-4 Theories of Failure 517

first. Failure resulting from local crippling or elastic instability is not

considered here.
The beginning of plastic flow (i.e., yielding) is indicated in a
uniaxial tensile test by the deviation from proportionality of stress to
strain. Practically, yielding begins at the yield strength at which plastic
deformation becomes significant. When several components of stress
occur, yielding depends on some combination of these components.
Although no theoretical method has been devised which correlates
yielding in the uniaxial tensile test with yielding in more complex
loadings, several theories of failure, based on predicting the onset of
yielding, have been proposed.

The maximum stress theory

The maximum stress theory proposed by Rankine is the oldest as

well as the simplest of all the theories.It is based on the assumption that

failure occurs when the maximum principal stress on an element reaches

a limiting value, the limit being the yield point in a simple tension test
(or ultimate strength, if the material is brittle). The theory disregards the
effect of possible other principal stresses and of the shearing stresses on
other planes through the element. For example, Mohr's circles for the
pure shear and pure tension in parts (a) and (b) in Fig. 13-6 show that
although both are equally strong according to the maximum principal
stress theory, part (a) has twice the maximum shearing stress of part (b).

This indicates that the maximum tensile or compressive stress alone is

lfot sufficient to define yielding. Nevertheless, this theory does give

results that agree well with test results from brittle materials.

(a) Pure shear (b) Pure tension

Figure 13-6. Although principal stresses are equal in parts (a) and (b), (a) has
twice the shearing stress of (b).
518 13 Special Topics

The maximum strain theory

According to the maximum strain theory, which is credited to

B.de Saint Venant, a ductile material begins to yield when the maxi-
mum principal strain c reaches the strain at which yielding occurs in
simple tension or when the minimum principal strain (i.e., the compres-
sive strain) equals the yield point strain in simple compression. How-
ever, if we examine Hooke's law for triaxial stress expressed by the
following equations:

(2-12)

we see that when ox = — ay = — o2 , the maximum strain is

(1 + 2v)a/E. On the other hand, when —ox =—oy =—o 2, as in

hydrostatic compression, the maximum strain is (1 —2v)a/E. Thus,
different strains may appear with the same maximum stress.

The maximum shear theory

Sometimes called Guest's theory, the maximum shear theory

assumes that yielding begins when the maximum shearing stress equals
the maximum shearing stress developed at yielding in simple tension.
Since the maximum shearing stress is equal to one-half the difference
between the principal stresses, the condition for yielding is

Tw " 2( a max.
_a = Ia
min.) yp

The Mlses yield theory

The Mises yield theory, also known as the maximum shear distor-
tion theory, assumes that yielding can occur in a general three-dimen-
sional state of stress when the root mean square of the differences
between the principal stresses is equal to the same value in a tensile test.
If a, > a
2 > a 3 are the principal stresses and a w is the yield strength in
simple tension, this concept gives
2 2 2
f[(*l
" o2 ) + (o 2 - o3 ) + (a 3 - a,)
]

2
= ^[(a yp -0) + (0-0) + 2
(0-a yp ) 2 ]
a
3 yp
13-5 Energy Methods 519

from which we obtain

2<J
2 = (
a ~ a 2f + (°2 ~ a a)
2
+ (°3 ~ a \f
yp i

Summary
Of work shows best
these several theories of failure, experimental
agreement with the Mises yield theory when applied to ductile materials.
For such materials the maximum shear theory also gives good agree-
ment. For rupture in brittle materials, such as cast iron, the maximum
stress theory is generally preferred.

13-5 ENERGY METHODS

By equating the external work done by applied loads as they deform an

elasticbody to the internal strain energy stored in the body, we obtain a
method of determining deflections that is based on the principle of
conservation of energy. This method will be shown to be extremely
versatile. We begin by obtaining expressions for the strain energy U
stored in a body under various loadings.

Axial loading

As an axial load is gradually applied to a bar of constant cross

section, the load P increases from zero to its applied value as the bar
undergoes a deformation 8. (The case of suddenly applied loads is
considered in the next article.) The work done, stored as strain energy in
the body, is the product of the average force \P and the deformation
8 = PL/AE. Hence we obtain

V= \% (13-1)

If the bar varies in cross section, this result may be applied to segments
of length dx and integrated over the length of the bar to obtain

Torsional loading

For a circular bar of constant cross section, the strain energy

stored in the body is equal to the product of the average torque and the
520 13 Special Topics

angular deformation, that is,

V =2 TB =2 T\-Jg)=2-JG (13 " 2)

When the torque varies, this result may be applied to segments of length
dx and integrated over the length of the bar to obtain
1
/•l T dx

Flexural loading

In the portion of a beam shown in Fig. 13-7, consider the

differentialelement isolated by two transverse sections a distance dx
apart. Treating this element as an axially loaded bar, where P — oy dA
= (My/ 1) dA, the energy stored in it is

t/ = Ilk P 2 dx My 'idAf
2,,
dx M 2
dx
-y 2 dA
2AE 2AE I* 2(dA)E 2EI 2
Summing this result to include all elements across the depth of the beam
gives the energy stored in the differential length dx of the beam as

dU =
M 1
dx r ,..
2
dA =
M 2
dx
2EI 2 fy
>' ~ 2EI
whence for the entire length of the beam we finally obtain

M 1
dx
v.J- 1EI
(13-3)

y X Neutral
>>">>>> iJ+»P=o dA surface
f

Figure 13-7. Energy stored in a differential element of a beam:

P- a, dA - (My/ 1) dA

Deflections

Instead of directly equating external work to the foregoing expres-

sions for strain energy, considerable simplification is obtained by apply-
ing Castigliano* s theorem or its close equivalent, the virtual-work method.
13-5 Energy Methods 521

Castigliano's theorem states that the deflection caused by any external

force is equal to the partial derivative of the strain energy with respect
to that force. To prove the theorem, consider the elastic body in Fig.
13-8 supported on rollers at A and B and by a hinge at C. As the loads
P, Q, F, are each applied independently to the body, they cause a
. . .

strain energy U to be stored in the body that is exactly equivalent to the

work done on the body by these loads. Note that the reactions do no
work, because the roller reactions are perpendicular to the displace-
ments of the rollers and the hinge reaction is fixed and cannot move.
Assume now that one of the loads, Q, is increased by a small
amount dQ. The increase in strain energy will be

However, since only Q is assumed to get an increment, dP — dF

= . = 0, and the change in U therefore is
. .

On the assumption that the principle of superposition applies, if

dQ had been acting initially on the body and P, Q, F, had been . . .

subsequently applied, the body would contain the same strain energy as
before. If 8 is the displacement of the point of application of Q in the
direction of g, then the increment dQ rides through the distance 8 as
the actual loads are applied and produces the external work increment
dU = 8 dQ. On equating work to the internal work given
this external

by Eq. (a), we obtain

dU-8dQ-jgdQ

Figure 13-8.
522 13 Special Topics

or

-g <->
This is interpreted as follows: The partial derivative of the strain energy
with respect to one of the external loads equals the displacement of the
point of application of that load in the direction of that load.
In applying Castigliano's theorem to a bar in which the strain
energy is the sum of the following integrals,

_
U~
r P 2 dx
+
r T 2 dx
+
r M 2
dx
J 2AE J 2JG J 2EI

the deflection, 8 = dU/dQ, is best evaluated by differentiating inside

the integral sign before integrating. This procedure is permissible be-
cause Q is not a function of x. With this simplification, the deflection at
Q in the direction of Q is given by

D dP.
/P-r-zz OX ~ x ar.
T^r dx M-,3a/.
~—- dx
«!£_ + rje- + r_»e_ ( i3- 5)
AE J JG J EI

If no load is acting at the point where the deflection is desired, a

dummy load Q in the direction of the desired deflection may be added
at that point. Then, after differentiating but before integrating, the
dummy load is set equal to zero.
In the virtual-work method, a dummy load of 1.0 N is used in
place of a dummy load Q. Then the derivatives dP/dQ and dM/dQ are
designated respectively as u and m. It is convenient to think of u and m
as the axial force and bending moment caused by the unit dummy load.
The resulting inconsistency in units is eliminated by multiplying 8 by

1.0 N thereby

Pu dx r Mm dx
1.0 N x
«=/^*/tt <>'-«>

This result can be obtained by a different process of reasoning

based on equating total external work to total internal work. For brevity
we have developed it here as a special case of Castigliano's theorem
which enables us to show the equivalence of the methods. The following
illustrative problems show how both methods are used and also give a
physical interpretation to the integral /(Mm/ EI) dx which simplifies
the computations.
13-5 Energy Methods 523

ILLUSTRATIVE PROBLEMS

1301. A bar is bent into a circular arc of radius R and held in a

horizontal plane as shown in Fig. 13-9a. Find the deflection at A
caused by a vertical load P applied there.

ds-Rdd

Moment
axis -j*

B //>. \
Torque ;4fl_!£*
axis-^y -A

(a) (b)

Figure 13-9. Deflection of free end of bar bent into a circular arc in a horizontal
plane.

Solution: The vertical load causes both twisting and bending over the
element of length ds = R dO. As shown by the top view in Fig. 13-9b,
the moment arm about the torque axis BC is AC = BD = OB - OD —
R — R cos 9; and the moment arm about the moment axis BO is
AD = R sin 9. Hence we have
T= PR(\ - cos 9) and
|| «*(!- cos 9)

dM =
M = PR sin 9 and
dp
R sin 9

Applying Eq. (13-5), we have

iif M—ds
=
s
7df T % ds +

=
JG
y' lpR ^ ~ cos9 R ( -
)[
l cos 9)]Rd9
/2
+ 4?
EI Jq
r PR sin 9{R sin 9)R d9

which is evaluated to give

- PR 3
= PR
:
/ 3?r 8 \ /7T\
+
8
JG \ 4 ) EI U )
Ans.

1302. A rigid frame, loaded as shown in Fig. 13- 10a, is hinged at

A and supported on rollers at D. Assuming EI is constant throughout
 524 13 Special Topics

m = 3+ 2"
o xc
c
\ *c 1600 N-m
800 N/m 2m
B
2 m
2 m 1600 N-m
m = jc
3m / 3 m
4 m Im "

m 7^
D
1 N

/~TTT7 1 N/^7 / / / //
\
fl D = 800 N
|n
!
RA - 800 N
N

(a) Actual loading (b) Dummy loading (c) Moment diagram

of actual loading

Figure 13-10.

the frame, determine the horizontal deflection at the roller D. Neglect

the effect of axial deformations.

Solution: In this example, we use the virtual work relation*

6
Mm dx
-i EI
Because discontinuities in the moment equation occur at the change of
loading points, we rewrite the integral in the following form:

b Mm dx c Mm dx d Mm dx
«B - f
EI f
J B EI i EI
(a)

The various values of m are found from the free-body diagram of

the frame (Fig. 13- 10b) in which a unit dummy load has been applied
at D in the direction of the desired deflection. The conditions of static
equilibrium determine the reactions to this unit load to be as shown.
The values of m shown are now determined from the definition of
bending moment, which specifies that we may take moments about an
exploratory section of the forces lying to one side of the section.

•The coefficient 1.0N that should precede S is usually omitted but is

understood to be included so as to insure dimensional homogeneity.
13-5 Energy Methods 525

Substituting these values of m in Eq. (a) and noting that EI is constant

we obtain

EI8D =
j
B
xA M dx +/ C (3 + ^xc \Mdx + j°xD M dx (b)

These integrals have the same physical meaning that we encoun-

tered in our earlier study of the area-moment method (page 229), and
hence we may express them in the equivalent form

EI8D - xA + 3(area)* c + ^(area) CB x c

(area),,* • •

+ (area)^ xD •
(c)

in which, for example, (area)^ xA represents the moment of area

•

about A of the actual moment diagram

for the segment AB. Observe
that moments of area are taken about the point from which x was
measured in finding the various values of m in Fig. 13- 10b.
The moment diagram by parts of the actual loading is now drawn
as in Fig. 13- 10c. Observe that no moment exists in the legs AB and
CD; hence in Eq. (c) the first and last terms on the right side are zero.
Evaluating Eq. (c), we now obtain

1600 x 2 1600 x 2
EI8 D = 3
2 3
1
+
2 (^)(§)-(^)(!)]
from which

EI8 D = 1867 N-m 3 Ans.

If this result had been negative, it would merely mean that the deflec-
tion at D was opposite to the direction of the unit load at D.

PROBLEMS

1303. Determine the midspan value of EI8 for a simply sup-

ported beam L meters long carrying a uniformly distributed load of
w N/m over its right half.

1304. As shown in Fig. P-1304, two aluminum rods AB and BC,

hinged to rigid supports, are pinned together at B to carry a vertical
load of 20 kN. If each rod has a cross-sectional area of 400
2
and mm
E= 70 GPa, compute the horizontal and vertical deflections of point B.
Assume a = 30° and = 30°. Ans. 8h = 0.412 mm; 8V = 3.57 mm
526 13 Special Topics

Figures P-1304 and P-1305.

1305. Solve Problem 1304 if rod AB is of steel, with E=

200 GPa. Assume a = 45° and = 30°; all other data remain un-
changed. Ans. 8h — 0.417 mm; 8V = 1.37 mm
1306. A bent into the shape of a quarter-ring and
circular bar is

supported in a vertical plane as shown in Fig. P-1306. Compute the

horizontal and vertical displacements of point A

1307. A circular bar is bent into the shape of a half-ring and

supported in a vertical plane as shown in Fig. P-1307. Determine the
horizontal movement of point C and the vertical movement of point B.
3 3

Ans. «a = PR 7T
; $b
= PR
2EI 2EI

Figures P-1307, P-1308, P-1309.

1308. Repeat Problem 1307 assuming that the load P is applied

vertically downward at C
13-5 Energy Methods 527

1309. In Problem 1307, let the load P be applied at C and

perpendicular to the plane of ABC. What will be the displacement of C
t_ : r
in the direction of the load?
i. i _.o
Arts.
PR 3 (3— \
I tt
J
H —
PR 3 /7T\
I
— I

1310. A P is applied to the rigid cantilever frame

vertical load
shown P-1310. Assuming EI to be constant throughout the
in Fig.
frame, determine the horizontal and vertical displacements of points B
and C. Neglect axial deformations. z
= Pba PP
Arts. 8B = 0; 8„ = 8^
2EI EI K)
B
T

Figures P-1310 and P-1311.

1311. In Problem 1310, let the load P be applied at C and

perpendicular to the plane of ABC. Find the displacements of points B
and C in the direction of the load. Ans.

1312. The rigid frame shown in Fig. P-1312 is supported by a

hinge at A and and carries
a roller at D the triangularly distributed load.
Assuming constant EI, compute EI8 at the roller D. Neglect axial
deformation. Ans. £/S = 2970N-m3
900 N/m

3 m
2m

4 m HD

/77Z7
Figure P-1312.
528 13 Special Topics

1313. The frame in Fig. P— 1313 is subjected to a horizontal

rigid
load and shown. Assuming constant EI and neglecting
a vertical load as
axial deformations, determine the value of EI8 at the roller support D.
Ans. EI8 = 58.2 kN m 3 •

900 N

2 m

3m

5 m

/Z7Z7
Figure P-1313.

13-6 IMPACT OR DYNAMIC LOADING

The deformations produced in elastic bodies by impact loads cause

them to act as springs, although that is not their designed function. If
the equivalent spring constant for such members is defined as the load
required to cause a unit deformation, the spring constant in each case
can be determined from our earlier study of deformations. Actually,
however, as we shall see, it is unnecessary to determine the equivalent
spring constant. For the present, we may consider the problem of
impact as analogous to that of a falling body stopped by a spring (Fig.
13-11). The mass m has zero velocity when first dropped, and also
when the spring is deflected through the maximum dynamic deflection
8. Equating the resultant work done on mass m to the zero change in
kinetic energy, we therefore obtain*

mg(h + 8) -\k8 2 = (a)

2
where mg(h + 8) is the work done by gravity on the body, and \k8 is

the resisting work done by the equivalent spring.

•The mass is assumed to remain in contact with the spring. Also, some
energy is dissipated by the impact, so the actual deflection is always less than that
given by Eq. (a).
 )

13-6 Impact or Dynamic Loading 529

m
t
h
Spring *
constan t
*N/i
m $
"*"
^M
////////////////.
Figure 13- 11.

If Eq. (a) is rearranged in the form

k k
and mg/k is replaced by 8sn which is the static deformation produced
by a gradual application of the weight mg, the following general value
of 8 is obtained:

2
8 = «„ +y(fi s ,) + 2« st /, (6)

Two extreme cases are of interest. If h is large compared with 5,

we may neglect the work mg8 in Eq. (a), which then reduces to

-V2^-h =V2SJ, (c)

In the other extreme case, h — (i.e., the load is suddenly applied)

and Eq. (a) reduces to

8 = 2-^ = 28 st (d)

Because of a suddenly applied load, the deflection and consequently the

stress which is directly proportional to it are therefore twice as great as
that caused by the same load gradually applied.
The ratio of the maximum dynamic deformation 8 to the static
deformation 5 st gives a value which may be called the impact factor. This
is easily determined by rearranging Eq. (b) in the form

8 =«..*v^-4 +A^f
Hence the impact factor is

= 1+
"
y f

Multiplying mg by this factor gives an equivalent impact load P, which

may be used in the formulas for static loading to compute the maximum
 530 13 Special Topics

stress and deflection. Or, if preferred, the static stress due to a

gradual
application of mg may be multiplied by the impact factor to give the
maximum stress:

<W=",t(l+yi+^ (13-8)
J

We now apply these results to various types of impact loading.

Tension

The most usual type of impact loading is shown in Fig. 13-12. A

mass m drops freely through a height h before striking a stop on the end
of the rod, thereby producing the dynamic deflection 8. Assuming that
the stresses remain within the elastic range and that 8 is negligible
compared with /*, we replace 8 st in Eq. (c) by its value from the
deformation equation 5 st = mgL/AE, whence we obtain

1L_
mgh (e)
-iAE
The corresponding stress in the rod is

= 8 „ I2E
a E =*\l^^mgh (/)
L V AL
-i.
By replacing mgh by the kinetic energy \mv 2 , this may also be used to
determine the shock stress caused by the sudden stopping of a mass m
that is moving with a velocity v:

mv
(g)
V AL

Figure 13-12. Impact loading of a rod.

 I

13-6 Impact or Dynamic Loading 531

These equations show that the stress due to impact can be reduced
by using a material with a lower value of E or by increasing the area A
or the length L of the rod. This is quite different from static tension
where the stress is independent of E or L.
The foregoing discussion assumes that the stress remains below the
proportional limit. When the stress is above the proportional limit the
problem is more complex, because the elongation is no longer propor-
tional to the load. Nevertheless, we can still find a basis for determining
actual rupture due to impact. Thus, it being assumed that the shape of a
tensile test diagram does not depend upon the speed with which the bar
is strained, the area OABC in Fig. 13-13 represents the work done upon

the bar to produce an elongation 8; this must be equal to the work

mg(h + 8) done by the falling weight mg. When mg(h + S) equals or
exceeds the total area OADE, the falling weight will rupture the bar.
A bar's resistance impact also depends upon the ductility of the
to
material. Figure 13-14 shows the tensile test diagram of a high-strength
steel of low ductility superimposed upon the diagram for a steel of lower
strength but high ductility. The horizontally shaded area A is much x

larger than the vertically shaded area A 2 showing that the more ductile
,

steel will absorb more energy before rupture than the less ductile steel.
For this reason, ductile materials are usually selected for members
subject to impact or shock loading.
In connection with the above, the total area of a stress-strain
diagram is called the modulus of toughness; it represents the energy
absorbed per unit volume. Its value is equal approximately to
a. + a..
u =
r
(h)

where ay and au are respectively the yield point and the ultimate
strength, and e u is the ultimate strain. The partial area of a stress-strain
diagram up to the stress oe at the elastic limit is called the modulus of
resilience; it represents the energy that can be absorbed per unit volume

^*—
Figure 13-13. Figure 13-14.
 532 13 Special Topics

without creating I permanent distortion. If the stress-strain relation is

•
its value is

(i)
IE

Flexure

13-15 a simply supported beam is subjected to the impact

In Fig.
of a mass m
falling freely through the height h before striking the
midpoint of the beam and causing the dynamic deflection 8. If the
proportional limit is not exceeded, Eqs. (b), (c), and (d) apply here also.
Assuming that h is large compared with S st we obtain ,

8 = V2h8 %x

Hence the impact factor for a centrally loaded simple beam where
8 tX mgL 3 /48£7 (Case 6, Table 6-2, page 270) becomes

_8_ /IT _ bdElh

mgL 2
The static stress found from the flexure formula is

=
Mc mgL c
a.,
/ 4 /

and the maximum stress is

6
J 6mghEc'
LI U)

..p:
Figure 13-15. Impact loading of a beam.
I
Limitations

preceding discussions assumed that the work done by a falling

body or the kinetic energy of a moving body can be stored in the
13-6 Impact or Dynamic Loading 533

resisting member in the form of strain energy. This assumption can

never be realized for the following reasons: If the velocity of impact is

high, the deceleration of the moving body may be so rapid as to

transform the kinetic energy into partly heat and partly local deforma-
tions, largely inelastic, of both the moving body and the resisting
member. Even if the velocity of impact is low, the resisting member may
have great stiffness, causing the same results. Finally, if the mass of the
resisting member is large compared to that of the moving body, the
inertia of the resisting member may also cause the same result.

PROBLEMS

1314. A 50-kg mass falls through 2 m and is then caught on the

end of a wire rope 30 m long having a cross-sectional area of 250 mm 2
.

Compute the maximum stress in the rope. Assume E= 100 GPa.

Ans. a = 164 MPa
An elevator having a mass of 2 Mg is being lowered at the
1315.
rate of 2 m/s. The hoisting drum is stopped suddenly when 30 of m
cable have been unwrapped. If the cross-sectional area of the cable is
600 mm
2
and E = 100 GN/m 2 compute the maximum stress in the
,

cable. Neglect the weight of the cable.

1316. A 6-kg mass falls 0.8 m and strikes the head of the steel
bolt shown in Fig. P— 1316. Assuming all the energy is absorbed by the
bolt, compute the required thickness / of its head if the shearing stress
on the cylindrical surface through the head is not to exceed 80 MN/m2 .

Assume E = 200 GN/m2 . Ans. t = 12.5 mm

A,

m
1.5 m

20-mm dia
steel bolt-
0.8 m

=-f,
Figure P-1316.
534 13 Special Topics

1317. A simply supported rectangular beam of length L and

ss-sectional area A is struck at the center by a mass m falling through
a height h. Show that the maximum bending stress is given by a 2 -
\%mgkE/Ak
1318. Compute the impact factor for a simply supported beam 3
m long subject to an impact by a 900-kg mass that drops from a height
of 2.5 m onto a point 1.0 m
from one end. Assume the beam section is
rectangular, 40 mm
wide and 90 mm
deep, and E = 200 X 10
9 2
N/m .

Neglect the mass of the beam. Ans. 25.9

1319. A rectangular beam 60

100 mm wide and
deep is used mm
I cantilever 2 m long. A 40-kg mass
through a height of 0.2
falls m
before striking the free end. Compute the maximum bending stress and
deflection caused by the impact. Neglect the mass of the beam and
assume that the 40-kg mass remains in contact with the beam. Assume
£«200GPa. Ans. o = 162 MPa; 8 = 21.5 mm
1320. A freight car having a mass of 12 Mg is moving at 1.2 m/s
when it is stopped by a nest of eight bumper springs. Each spring is 10
turns of 25-mm-diameter steel wire on a mean coil radius of 100 mm.
Use the Wahl formula [Eq. (3-10) on page 97] to determine the maxi-
mum stress developed in the springs. Assume G— 80 GPa.

1 3-7 SHEARING STRESSES IN THIN-WALLED MEMBERS

SUBJECTED TO BENDING; SHEAR FLOW

The formula developed in Art. 5-7 (page 188) for determining the
shearing stress induced by flexure can also be used to determine the
shearing stress across the flanges of wide flange beams or channels or
other sections. The existence of this shearing stress is explained in Fig.
13- which shows the free-body diagrams of portions of the flanges
16,

cut out by two adjacent sections. The dashed lines indicate a phantom
view of the cantilever beam from which these sections were cut. When
the external vertical shear acts downward, the upper flange is in tension,
the tensile force T2 being larger than T because the bending moment is
l

greater at section 2 than at section 1. For equilibrium of the upper

flange segment, a longitudinal shearing force F must act as shown,
s
thereby inducing the lateral shear force H x
. The direction of H l
de-
termines the directions of the shearing stresses in the upper flange.
Similarly, the compressive forces C2
and C, developed in the lower
flange segment required a leftward lateral shear // and hence the ,
2
shearing stresses are here directed leftward. Because of tension in the
upper flange and compression in the lower flange, the shearing stresses
in these flanges are in opposite directions.
13-7 Shearing Stresses in Thin-Walled Members 535

Figure 13-16. Lateral shear forces H, and H2 in flanges of cantilever beam.

External vertical shear acts downward.

The magnitude of the longitudinal shearing stress across the flange

is given by Eq. (5-4) (page 189) if the flange is assumed to be relatively
thin so that, as was done in deriving this equation, the shearing stress
may without serious error be considered uniform across the thickness of
the flange. Even more convenient is the concept of shear flow developed
in Art. 3-5 and again in Art. 5-7 as Eq. (5-4a). Thus at a distance z
from the free edge of the flange in either part (a) or (b) of Fig. 13-17,
Eq. (5 -4a) gives

(a)

This shows that the shear flow in the flanges varies linearly with the
distance from the free edge.
The variation in shear flow and its direction are shown graphically
in Fig. 13- 18a and b, in which the external vertical shear V is assumed
to act downward.
Similarly, for the split tube in Fig. 13- 17c, the shear flow across
any radial section defined by </> is

V V /•<*>
m Vtr
4 = ~jQ = -[
J
rsin *("" d<>) —j-0 ~ cos *) (*)
536 13 Special Topics

1 * —

J.
(a) (b) (c)

Figure 13-17.

-4?^
2q
I— Tvf— I [\

f^^^i [/
1
^L
(a) (b)

Figure 13-18. Variation in shear flow. In parts (a) and (b), external vertical shear V
is down; in parts (c) and (d), external V is up.
13-8 Shear Center 537

When the external vertical shear acts upward, Fig. 13- 18c shows how
the shear flow varies from zero at the split to a maximum value at a
section opposite the split. As an exercise, verify the variation in shear
flow for the solid tube in Fig. 13— 18d.

13-8 SHEAR CENTER

We are now ready to consider the bending of members that have only
one axis of symmetry, the loading being such as to cause this axis to
become the neutral axis. On any section of a beam subjected to other
than pure bending, there exist shearing stresses. These stresses create
internal shearing forces whose resultant must be equal, opposite, and
collinear to the external shear; otherwise the bending is accompanied by
twisting of the beam. Bending without twisting occurs only when the
resultant of the shearing forces passes through the shear center (also
called the center of twist and sometimes the flexural center). The shear
center is defined as the point in the cross section of a beam through
which the plane of the transverse bending loads must pass so that the
beam will bend without twisting.
We begin by considering a channel section used as a cantilever;
the free-body diagram is shown in Fig. 13-19. The resisting forces
consist of the resisting vertical shear Vr considered as acting through the
web of the channel (as shown for a wide flange beam in Illustrative
Problem 566), the resisting couple M r
composed of the tensile and
compressive forces T and C (shown for convenience as acting through
and normal to the flanges), and the horizontal flange forces H, which

Figure 13-19. Free-body diagram of channel section.

538 1 3 Special Topics

arc the resultants of the shearing stresses in the flanges computed as

shown in the preceding article. It may seem surprising that the load P
does not act through the longitudinal centroidal plane of the section;
hut the equations of equilibrium show why P must act as shown. Later
we shall compute its position. The six equations of equilibrium, and the
reasons they are satisfied, are as follows:

L *£X = 0, by balance between the equal and oppo-

satisfied

sitely directed tensile and compressive forces T and C.

2. 2Y = 0, satisfied by the resisting vertical shear Vr balancing
the vertical shear V caused by P.
3. 2Z = 0, satisfied by the balance of the equal and oppositely
directed flange forces H.

4. *EMy = 0, satisfied because vertical loads cause no moment

about the Y axis, and the moments of the horizontal forces T and C
about Y cancel each other, as do the flange forces H.
5. 2 Mz = 0, satisfied because the applied bending moment M is

balanced by the resisting moment Mr

supplied by T and C.

6. SA/jr = 0; this condition must be satisfied to prevent the beam

from twisting as it bends. It can be satisfied only if the moment of the
applied load balances the moments of the shearing forces developed
over the section. Selecting the X axis through the web eliminates the
moment of Vr In the end view
. of Fig. 13-19, shown in Fig. 13-20, we
set a moment summation about O equal to zero and obtain

[LM = 0] Ve = Hh (a)

The value of the flange force H is the product of the average shear
flow in the flange multiplied by the length of the flange. Using Eq. (a)
of Art. 13-7, we have

2
__ (\ Vhtb\, Vhtb

This value of H may now be substituted in Eq. (a), yielding

2 2
Hh h b t

Point C on the neutral axis (NA) located a distance e from the center of
the web is the shear center for the channel.
wide flange beam or an I beam, flexure loads develop a
In a
lateralshear in each outstanding flange as in a channel section, but the
shearing forces in each flange are equal and oppositely directed, as in
Fig. 13-21. (See also Fig. 13- 18a for the direction of the shear flow.)
Hence, moments of the shearing forces about the centroid of the section
balance out, and the plane of the loads must also contain the centroid if
13-8 Shear Center 539

h-b

r
H «
/ i

i
—
i/
*- i
-* —H i

Figure 13-21. In a section having two axes of symmetry, the shear center C
coincides with the centroid O.

twisting is to be avoided. We conclude that the shear center coincides

with the centroid of the wide flange section. In general, the shear center
is located at the intersection of two axes of symmetry, if they exist.

For an unequal flanged H section with relatively thin flanges and

web; like that in Fig. 13-22, the bending resistance of the web may be
assumed to be negligible and the total vertical shear V may be assumed
to be resisted by the internal shears V and V2
x
acting along the center
lines of the flanges. To prevent twisting, the resultant Vr of the flange
shears must be equal, opposite, and collinear with the external shear V.
Hence, taking moments about the shear center C, we have
V x
b x
= V2 b 2 (b)

Another relation between V and V2 may be found from the fact that
x

the two flanges bend as though they were separate beams which have
identical radii of curvature. Hence, applying p = EI/M gives

(c)
M,
But the bending moments M x
and M 2 at any section along the beam are
equal respectively to the product of the vertical shears V and V2
x
in the
S40 13 Special Topics

Figure 13-22.

flanges and the distance to the section. Hence Eq. (c) becomes

(d)

which when combined with Eq. (b) yields

(13-10)

Therefore, since the ratio b x

/b 2 and the sum of b x
and b 2 are known, the
position of the shear center is easily located. It lies between the centroid
of the section and the centroid of the flange that has the larger moment
of inertia.
When there is only one flange, as in the T section in Fig. 13-23, if
the bending resistance of the web is again assumed to be negligible, the
shear center coincides with the centroid of the flange. In general, for
any section composed of two narrow rectangles where the shear flow is

along the longer center lines of the rectangles, the shear center is at the
intersection of these center lines, as shown in Fig. 13-24.

V,

Figure 13-23. Figure 13-24.

13-8 Shear Center 541

H
< i

(r

^+=A I *
u V

H
(a) (b) (c)

Figure 13-25. Shear center of a Z section coincides with the centroid.

The procedure for a Z section is the same as for a channel section;

the shear flow is shown in Fig. 13-25a, and the resultant shear forces in

Fig. 13 -25b. The resultant of the two flange forces is 2H acting through
the centroid of the section. Combining this resultant with the shear
force in the web gives the resultant shear force R in the section.
Evidently the shear center C coincides with the centroid O. The fact
that the plane of loading must coincide with R requires that the section
be oriented as shown in Fig. 13-25c if we expect a vertical loading to
cause bending in the vertical plane. In the next article we explore more
completely the general relation between the plane of loading and the
plane of bending.

ILLUSTRATIVE PROBLEM

1321. If the vertical shearing force acting on the thin-walled

channel section in Fig. 13-26 is 2000 N, compute and illustrate the
shear flow and determine the shear center.

Solution: The moment of inertia about the neutral axis is computed

from
3
(2.5)(250)
=
-2(f ^) + /
12
2(2.5)(100)(125)'

= 11.07 x 10 mm6 4

= 11.07 X 10" m 4 6

Because of the thin wall, the shear flow may be assumed to act
along the center line ABCDE of the section, as shown in Fig. 13-26b.
At A, the shear flow is zero; at B, Eq. (5-4a) gives

V 2000
10" 3 )(100 10" 3
**' QaB ~ (2.5 x x )
I 11.07 X10- 6
x(125 x 10" 3 )
= 5646 N/m
542 13 Special Topics

gB =5646 N/m
tf=282N

250 mm
g c = 9175N/mJ

(•)

The shear flow at any other point C may be found directly from
qc — (V/I)QA However, since QAC is the moment of area from A to
c
C, which is equivalent to the sum of the moment of area from A to B
plus that from B to C, we may write qc — (V/ I)(QAB + QBC \ which
then reduces to the more convenient form qc = qB + {V/ I)QBC Thus .

the shear flow at C is

V 2000
*C
<7c -
^
1b + "7 ^c
C*c " 5646 + ^^ r( 25
6V x 10
"3
)
^
7 [ 11.07 X10"
^- x 10" 3 \
/i^c 3 x/125
x(125 x 10 )l I

- 5646 + 3529 = 9175 N/m

As shown in Fig. 13-26b, the shear flow from A to B varies
directly with the distance from A ; but from B to C to D it varies along a
parabolic arc. The average shear flow in the web is therefore 5646
-I-
f(3529)
= 8000 N/m. The shear force in the web is Vr = ?ave L = •

8000(0.250) = 2000 N, which agrees with the applied vertical shear

V 2000 N. The shear force in each flange is H = q &ve L = •

(j x 5646)(0.100) = 282 N.
To avoid twisting of the section, the external shear V must lie a
distance e to the left of O (Fig. 13-26c) so that the twisting moments
exerted by the internal shear forces will be balanced. Hence,

[2M = 0] 2000 e = 282(250) e = 35.3 mm

The value of e can be computed more easily from Eq. (13-9):
.2l2
_ hVt _ (250r(100r(2.5) =
47
35.3 mm
4(11.07 X 106 )
13-8 Shear Center 543

However, the above numerical computations are presented in order to

emphasize the principle of shear flow and to indicate its extension to the
more complex problems that follow.

PROBLEMS

1322. Locate the shear center for the section shown in Fig. 13-22
if f, — t
2
— t
3
= 10 mm, h — 150 mm, h 2 = 100 mm, h 3 = 180 mm.
x

1323. Determine the position of the shear center for a section

composed of a thin-walled cylinder of thickness / and mean radius r
which is split along one longitudinal element, as in Fig. 13- 17c.
Ans. e — 2r measured along the axis of symmetry
from the center of the cylinder in a direction
opposite to the split element.

1324. Show that the position of the shear center for the semi-
circular thin ring in Fig. P- 1324 is e = 4r/w to the left of O.

Figure P-1324.

1325. The thin-walled section shown in Fig. P-1325 consists of a

semicircular ring of mean radius r and two straight pieces of length r.
4
Show that the shear center is e = (tr / I)(it + 3) to the left of 0, and

hence for r = 50 mm and t = 2.5 mm, that e = 86.0 mm. Need the
value of / be specified?

Figure P-1325.
544 13 Special Topics

1326. If the vertical shear on the section shown in Fig. P-1326 is

3600 N. construct a shear flow diagram and locate the shear center.
4ns. q = 3 and 6 kN/m at the junction of flange
and web; e = 18.8 mm to the left of the web
center

ft
50 .,
mm 100 mm ^

300 mm

*«4 mm

Figure P-1326.

1327. If the vertical shear on the section shown in Fig. P-1327 is

3000 N, construct a shear flow diagram and locate the shear center.
Arts. qB — qE = 3.01 kN/m; qc = qD = 8.14
kN/m; e = 62.8 mm to the left of web center

f*-100-*|
mm ' _
c-!
'
B
80 mm
U_i

t — 8 mm 300 mm

F
1mm
80

i)

Figure P-1327.

1328. Locate the shear center for the thin- walled section shown
in Fig. P- 1328. Ans. 27.5 mm to the left of the web
 i

13-8 Shear Center 545

r — 150 mm
<
150 mm
*
U
!!
80 mm
II

NA
80 mm
y
il

80 mm
y
it

80 mm
V

Figure P-1328.

1329. Show that the shear center for the thin-walled split square
section shown in Fig. P-1329 is b/(lV2) to the left of the comer
opposite the split.

Figure P-1329.

1330. A thin-walled section has the shape shown in Fig. P-1330.

The outstanding legs have a slope of 3 vertical to 4 horizontal. Locate
the shear center. Arts. 16.3 mm to the left of the web

.A

60 mm
NA
\
60 mm

Figure P-1330.
 13 Special Topics

13-9 UNSYMMETRICAL BENDING

The theory of flexure developed in Chapter 5 was restricted to loads

King in a plane that contained an axis of symmetry of the cross section.

With this restriction, the neutral axis passes through the centroid of the
section and is perpendicular to the plane of loading. The preceding
articleextended the application of the flexure formula to sections with
only one axis of symmetry that were loaded so that this axis became the
neutral axis. In either case, bending without twist is possible only if the
plane of loading contains the shear center, a requirement that is auto-
matically satisfied when the axis of symmetry coincides with the plane
of loading.
There is a further restriction which so far has been adhered to,

although not emphasized. The plane of loading must be parallel to or

contain a principal axis of inertia of the beam cross section. We first

consider the case in which the plane of loading contains an axis of

symmetry, such as the Y axis in Fig. 13-27. In deriving the flexure
formula (Art. 5-2), we applied the condition of equilibrium that the
applied bending moment about the X axis is balanced by the resisting
moment exerted by the flexure stresses, i.e., M
x
— fy(a dA).
If the bending loads are restricted so that they lie in the longitudi-
nal plane containing the Y axis, the external moment My must be zero.*
However, the flexure force o dA on a typical element of the section has
a moment x(o dA) about the Y axis. If the Y axis is an axis of
symmetry, this moment about Y is neutralized by an identical force (not
shown) acting through the point of symmetry. For sections that do not
have an axis of symmetry, the resultant moment of the flexure forces
about the Y axis is

M y
= jx(odA) = jx(— y\dA = — j xy dA

Therefore M y
will be zero and equilibrium satisfied only if the integral
fxy dA is zero. This integral is the product of inertia P^ which is zero 9

only if X
and Y are the principal axes of inertia of the section. We
conclude that the flexure formula may be applied only if the bending
loads act in a longitudinal plane parallel to or containing one of the
principal axes of the section. These planes are called the principal
planes of bending.
We are now ready to discuss unsymmetrical bending, which is
defined as bending caused by loads that are inclined to the principal
planes of bending. Examples of unsymmetrical bending are roof purlins

•When the section has only one axis of symmetry which becomes the neutral
axis, the plane of loading must be offset from, but parallel to, the longitudinal
centroidal plane so as to pass through the shear center; but even then is zero. Mv
13-9 Unsymmetrical Bending 547

Longitudinal plane
of bending loads

odA=± ydA

Figure 13-27. Flexure stress causes resisting moment about Y axis as well as
about X axis. My «= if Y is an axis of symmetry (or a principal axis if there is no axis

of symmetry).

that, because of the inclination of the roof, are subjected to loads whose
planes make large angles with the principal axes of inertia of the
section; and beams in structures and machines that are subjected to
loads which, because of deformation or design, are inclined to the
principal planes.
We consider first the case shown in Fig. 13-28, in which a
symmetrical section is subjected to loads inclined to the axes of symme-
try. Resolving the loading into horizontal and vertical components, we
obtain the two loading conditions shown in parts (b) and (c), which can
each be solved directly by the flexure formula. In part (b), the X axis is

N
^ n
/
PcosO

Psin0
+
TrlN

(a; (c)

Figure 13-28. Unsymmetrical bending resolved into symmetrical bending about X

and Vaxes.
548 1 3 Special Topics

the neutral axis, whereas in part (c) the Y axis becomes the neutral axis,

i id) of these conditions produces flexure stresses that are normal to the
cross section; hence the resultant stress at any point is the algebraic sum
of the stresses at that point caused by each case considered separately,

My Mx x
(13-11)
h h
where M x bending moment about the X axis caused by P cos 0,
is the
and M
Y
bending
is the moment about the Y axis due to P sin 9. In terms
of the total bending moment A/, it is evident that x
= cos 9 and M M
M y
Msin 0, so Eq. (13-11) can also be written as

9 m
(Mc^)y
'x
+
(Msjn«)x
*y
^^
In applying the algebraic summations of stress indicated by Eq. (13-11)
or (13-1 la), tabulating the stresses as in Illustrative Problem 1331 below
will avoid confusion regarding signs.
To return to Fig. 13 -28a, the neutral axis has beenshown at an
inclination a with the X axis. To determine this inclination, we locate a
point A on it by the condition that the neutral axis is the locus of points
of zero stress. For the given inclination of the load, these points of zero
stress must be in the second and fourth quadrants because only there
will the flexural stresses of parts (b) and (c) be of opposite sign. In these
quadrants, the coordinates of a point A on the neutral axis will be — xA
and +yA or + xA and —yA Substituting either set in Eq. 13-1 la, we
, .

obtain

(M cos 0)(-yA ) (Msin9)(xA )

oA + t

=
x J
y

from which, by canceling the common term M and rearranging, we get

y± mm
h t
sinfl
xA I
y
cos 9

which is equivalent to

tana = ytan0 (13-12)

y

From this, we see that unless lx = / or tan 9 = or oo, the

neutral axis is not perpendicular to the load. Also observe, as shown in
Fig. 13-28a, that the neutral axis from the X axis in the same
is inclined
angular sense as the plane of loading is from the Y axis.
To determine the deflection, we combine vectorially the deflec-
tions caused by the X and Y components of the load. These deflection
13-9 Unsymmetrical Bending 549

Figure 13-29. Unsymmetrical bending produced in a nonsymmetrical section.

Axes 1-1 and 2-2 are principal axes.

components are respectively 8X = Px L 3 /kEIy and 6^ = Py L 3 /kEIx ,

where A: is a factor which depends on how the beam is supported. The

inclination of the total deflection with the Y axis is expressed by

6\ PJK Psin0 L= /:
-f tan
PJIX Pcos0 J

This result is the same as tan a in Eq. (13-12). Since a was previously
defined with respect to the X axis and here is measured from the Y axis,
we conclude that the total deflection is perpendicular to the neutral axis.
We now consider the nonsymmetrical section like the vertically
loaded Z section in Fig. 13-29. For this section, the principal axes are
1-1 and 2-2; hence the load P is inclined at the angle with one of the
principal planes of bending. This loading, therefore, also causes unsym-
metrical bending. The stresses may be determined, as shown below in
Illustrative Problem 1332, by resolving P into components directed
along the 1-1 and 2-2 axes and then applying Eq. (13-11). For reference,
we direct attention back to Fig. 13-25c, where the resultant load R will
cause a vertical deflection if R coincides with the principal axis 1-1. This

will cause 9 to be zero in Eq. (13-12) and axis 2-2 will become the
neutral axis.

ILLUSTRATIVE PROBLEMS

1331. A W250 x 33 section is used as a cantilever beam to

support the given loads inclined to the Y axis as shown in Fig. 13-30.
Compute the stresses at the corners A, B, C, and D at the wall section.
What is the inclination of the neutral axis at the wall?
550 13 Special Topics

6000N T 8000N

8000 N
JUfN 6000
t>uuu N

2m 1 2 m
?.

Figure 13-30.

Solution: We begin by resolving the loads into their X and Y compo-

nents and computing M x and My :

M x
= - (6000 cos 45°)4 - (8000 cos 30°)2 = -30.8 kNm
The negative sign for M x indicates downward curvature at the wall, and
hence the bending causes tension at A and B and compression at C
and D.
Taking moments of the X components of the loads about a Y axis
at the wall, we obtain
My
= (6000 sin 45°)4 - (8000 sin 30°)2 - 8.97 kNm
This bending moment causes tension at A and D and compression at B
and C.
From Appendix B, Table B-2, we find Sx = 379 X 103 mm3 = 379
x 10" 6 m 3 and Sy - 64.7 x 103 mm3 = 64.7 X 10" 6 m3 The stresses .

caused by M x and My are therefore

or =
M x 30.8 x 10
3
= 81.3 MPa
379 X 10" 6
3
A/., 8.97 x 10
6
- 138.6 MPa
64.7 x 10"

As indicated by Eq. (13-11), these combined alge- stresses are

braically. For this purpose, tensile stresses in the accompanying table
are denoted by a positive sign and compressive stresses by a negative
sign.

Stress in MPa
due to: A B C D
+ 81.3 + 81.3 -81.3 -81.3
+ 138.6 -138.6 - 138.6 + 138.6

2 + 219.9 -57.3 -219.9 + 57.3

13-9 Unsymmetrical Bending 551

It is not convenient in this example to orient the neutral axis by

means of Eq. (13-12) since the plane of loading is not constant along
the length of the beam. Instead, we return to Eq. (13-11) and by setting
the stress equal to zero for a point on the neutral axis, we obtain

whence, replacing v / x by tan a and substituting the values of Ix and Iy

Table B-2 (Appendix B), we have
listed in

48.9 8.97
x = 3.01 or a = 71.6°
4.73 -30.8

In using this method, it is preferable to ignore signs completely

and merely substitute numerical quantities. The neutral axis is then
oriented so as to be consistent with the signs of the stresses tabulated
above. In this instance, the neutral axis is directed up to the right at
71.6° with the X axis.
1332. A cantilever beam 2 m long carries a vertical load P=
900 N at the free end, as shown in Fig. 13-31. Compute the maximum
stress at the corner A. The properties of the structural Z section are
/, - 8.00 X 10
6
mm 4 , I2 - 3.75 X 10
6
mm4 , and Iy = 1.23 X 10
6
mm4 .

Solution: We begin by determining the direction and values of the

principal moments of inertia. The given properties enable us to plot a
Mohr's circle of inertia, as shown in Fig. 13-32, from which we obtain
Ix = 10.53 X lO
6
mm 4 and0 = 31.4°.
From Fig. 13-31, the coordinates of point A are

xA = 76.2 cos 31.4° - 63.5 sin 31.4° = 32.0 mm

yA = 76.2 sin 31.4° + 63.5 cos 31.4° = 93.9 mm

P /Y
76.2 mm-

<4
Z* —
P=900N
I >t
2 m-
63.5 mm
7
.1
fc -^ /^=31.4'
7
-4-12.7 mm

Figure 13-31.
552 13 Special Topics

i mm 4

(IX 10- 6 )mm4

IX X 10" 6 = 5.88+4.65
"6
= 10.53
/V X 10 = 1.23

cos 20 = 2JJ*
4.65
6 = 31.4
C

Figure 13-32.

and the components of P are

Px = P sin 9 = 900 sin 31.4° = 469 N

Py = P cos 9 = 900 cos 31.4° = 768 N
Hence the components of the maximum bending moment are

M x
- Py L = 768(2) = 1536 N-m
My = PX L = 469(2) = 938 N-m
We now apply Eq. (13-11), noting that Mx produces tension and
M produces compression at A . Substituting the values previously com-
puted, we obtain
_ M^y_ _ M^ = 1536(0.0939) 938(0.0320)
/, ly ' 10.53 x 10" 6 1.23 x 10" 6
= (13.7 - 24.4) x 10
6
= - 10.7 MPa Ans.

The negative sign indicates that the stress at A is compressive.

PROBLEMS

In all these problems the loading passes through the shear center.

1333. Compute the horizontal and vertical components of deflec-

tion at the free end of the cantilever beam in Illustrative Problem 1331.
Use E= 200 GPa.
1334. A beam simply supported at the ends has the cross section
and is loaded with a concentrated load P as shown in Fig. P-1334. If
themaximum bending stress is not to exceed 120 MPa, determine the
maximum safe value of P. Ans. P= 36.2 kN
13-9 Unsymmetrical Bending S53

30° P

20 mm [
r

2m L 150 mm
20 mm

20 mm [ =1
f 150 mm
Figure P-1334.

1335. The T section shown in Fig. P-1335 is the cross section of

a simply supported beam 5 m long that carries a central concentrated
load inclined at 60° to the Y axis. The X axis 72.9 mm
centroidal is

below the top of the section; Ix = 34.18 X mm4 and Iy = 5.76 X 106
10
6

mm 4
If ac < 80
. MN/m
2
and a, < 40 MN/m what the maximum
2
, is

load that will not overstress the beam? Arts. P= 3.05 kN

^60°

20 mm

200 mm V
150 mm

—H U— 20 mm
Figure P-1335.

1336. A beam 3 m long with the same T section as in

cantilever
Problem 1335 two concentrated loads applied as shown in Fig.
carries
P-1336. Compute the inclination of the neutral axis at the wall, and the
maximum compressive and tensile bending stresses.
Ans. a = 46.6°; max. oc = 61.3 MPa; max. a, = 59.1 MPa
554 13 Special Topics

4kN 3kN
30 45

3kN 1 kN

1 m 2m

Figure P-1336.

1337. The Z beamProblem 1332 is used as a roof

in Illustrative
purlin, P-1337. It carries a vertical, uniformly
as indicated in Fig.
distributed load of 3000 N/m on a simply supported span 4 long. If m
the slope of the roof is 1 to 4, compute the stress at corner A if (a) leg
AB points up and (b) leg AB points down. Which design is better?

Figure P-1337.

1338. A 150 x 100 X 16 mm angle is used as a cantilever beam

2 m long with the 150- mm leg vertical. It supports a load of 4000 N
applied at the free end of the beam. Compute the maximum stress.

( Hint: Compute the inclination of the neutral axis to determine the

point of maximum stress.) Ans. 128 MPa

13-10 CURVED BEAMS

Members subjected to bending are not always straight; sometimes, as in

the case of crane hooks, they are curved before a bending moment is

applied. If the member is sharply curved, the stress distribution is

markedly different from that given by the flexure formula a = Mc/I,

which was derived for beams that are initially straight.
For example, a sharply curved beam is subjected to bending
couples, as shown in Fig. 13-33. It is usually assumed that plane radial
13-10 Curved Beams 555

Straight ^L
beam "**y r^^
J/.
Curved
beam

I ^s^/ " \ x Neutral

j^HT \ |
surface

Figure 13-33.

sections remain plane after bending. Although not strictly accurate, this
assumption gives results that agree closely with actual strain measure-
ments. In accordance with this assumption, bending causes section cd to
rotate, relative to section ab, to thedashed position. Consequently, the
two fibers e and /, each equidistant from the neutral surface, will have
equal deformations, i.e., 8e = fy. Applying Hooke's law, 8 = oL/ E, we
have

oL L
°f f

From Fig. 13-33 it is evident that the length Le of fiber e is greater than
the length Lf of fiber /, the difference in length depending on how
sharply the beam is curved initially. Consequently, oe is less than oy, and

the result is the nonlinear stress distribution shown.

As a consequence of the nonlinear stress distribution, there can be
no balance between the tensile and compressive forces over the section
if the neutral surface passes through the centroid of the section; the
neutral surface must
from the centroid of the section toward the
shift
axis of curvature O. The dashed
linear stress distribution shows not only
this shift but also the relatively increased stress at the inner fibers and
the decreased stress at the outer fibers, in comparison with the stresses
computed from the flexure formula.
To determine the shift in position of the neutral axis and to
express the stress at any fiber in terms of the applied bending moment,
we proceed as follows: In Fig. 13-34, ab and cd represent two adjacent
sections of a curved beam. Let dO be the angle between these sections
before bending and dy be the angle of rotation of cd relative to ab
caused by bending. Let y denote the coordinate of a typical element dA
with respect to the neutral axis, which is at an as yet undetermined
distance e from the centroid of the section; R represents the radius of
curvature of the centroidal axis.
556 13 Special Topics

Centroidal axis

Axis of curvature

Figure 13-34.

The total elongation of a fiber at a distance v from the neutral axis

is y The original length of this
dtp. fiber is (R — e + y) dQ. Hence, the
unit elongation or strain is

y d<p
C (a)
L (R •
+ y)dO
and from Hooke's law, the stress is

n
O -
— F,
£L€ - Ed(f
= — — y
W
dO R - e +y
If the beam is loaded in pure bending, the conditions of
equilibrium require that the sum of the normal forces over a cross
section be equal to zero and that the moment of these normal forces
balance the applied bending moment. In accordance with the first of
these conditions, a force summation over the entire area yields

Since
ia

E
dA =
E
dO

d<p/d9 cannot be zero,

d<p

h
J
r
R -
y dA
e +y
we
=

obtain
(c)

y dA
= (d)
/ R e +y
in which e is the only unknown. Its value may be found from Eq. (d) by
letting v denote the distance from the axis of curvature to the element
dA. Then v = v - (R - e), and Eq. (d) is rewritten as

M
y dA v - (R - e) dA
J
f
R - e + v J
f dA
-h -(*->/?-
13-10 Curved Beams 557

from which we obtain

e = R - -4— (13-13)
/OA
oA
V
Equating the applied bending moment to the resisting moment
gives

M-fyoU-^f
J' dO
* 2dA
R - +y J e
w
(e)

This integral is simplifiedby adding and subtracting (R — e) to one of

the two >>'s in the numerator so that y = (R — e + y) — (R — e). The
integral is then rewritten as

y u_ =
y" {R . e) c^J±
2
ydA
dA .
f ydA
=fy -(R-e)( (/)'
Jf R - e + yv J*
J R-
D - e + v
'J R
') v
y
x

The first integral on the right side of Eq. (/) is the moment of the
entire cross-sectional area about the neutral axis and equals Ae. The
second integral, from Eq. (d), equals zero. Eq. (e) can now be rewritten
as

Edy = M
d9 Ae
This value of E dy/dO is then substituted in Eq. (b) to yield finally

a= ^._^
Ae R - e
—+y
= ^.z
Ae v
(13 _ 14 ')

Equations (13-13) and (13-14) are theoretically adequate to de-

termine the stresses in curved beams but are limited in usefulness by the
difficulty ofcomputing the value of e. This difficulty may be avoided by
means made by B. J. Wilson and J. F. Quereau.* These
of a study
investigators computed the extreme fiber stresses in curved beams of
various cross sections with the curved beam theory and with the
ordinary flexure formula. From a comparison of these results Wilson
and Quereau determined values of a correction factor K by which
stresses computed with the flexure formula can be multiplied to give the
actual stress in a curved beam. A modified equation for computing the
extreme fiber stresses in curved beams is therefore

a= K^f (13-15)

Values of K in Eq. (13-15) vary with the ratio R/c> where R is the
radius of curvature of the centroidal axis and c is the distance from the

*See "A Simple Method of Determining Stresses in Curved Beams," Circular

16, Engineering Experiment Station, University of Illinois, Urbana, III., 1928.
558 13 Special Topics

centroidal axis to the inner fiber. As Fig. 13-33 shows, these stress

correction factors are greater than unity for the inner fibers and less

than unity for the outer fibers. At values of R/c greater than 20, these
factors approach unity and the flexure formula may be applied directly
to such slightly curved beams. Table 13-2 lists correction factors for
various cross sections.
For beams subjected to other than pure bending, as in Fig. 13-35,
the system of coplanar forces acting in the plane of curvature is reduced
to a single force R acting at the centroid of the section, plus a bending
couple M. The moment of this couple is equivalent to the summation of
moments about the centroidal axis of the external forces acting to one
side of the cross section. The normal stresses produced by this couple
are found as in pure bending.
The force R is resolved into two components: a shearing force V
in the plane of the cross section, and a normal force TV perpendicular to
the plane of the cross section. The normal force acting at the centroid of
the section produces tensile or compressive stresses, uniformly distrib-
uted over the section, of the magnitude a = N/A, where A is the

TABLE 13-2. Correction Factors KJor Curved Beams for Use In Eq. (13-15)

Ht
y////^
Cross t

section C Vy///
WzZs^ R 1 t
IprR
A
OTHER SECTIONS
CIRCLE OR ELLIPSE RECTANGLE (AVERAGE VALUES)
R/c Inside Outside Inside Outside Inside Outside

1.2 3.41 0.54 2.89 0.57 When section is unsym-

1.4 2.40 0.60 2.13 0.63 metrical, R/c refers to
1.6 1.96 0.65 1.79 0.67 the inside fiber.
1.8 1.75 0.68 1.63 0.70 — —
2.0 1.62 0.71 1.52 0.73 1.63 0.74
3.0 1.33 0.79 1.30 0.81 1.36 0.81
4.0 1.23 0.84 1.20 0.85 1.25 0.86
6.0 1.14 0.89 1.12 0.90 1.16 0.90
8.0 1.10 0.91 1.09 0.92 1.12 0.93
10.0 1.08 0.93 1.07 0.94 1.10 0.94
20.0 1.03 0.97 1.04 0.96 1.05 0.95
 — -

13-10 Curved Beams 559

Figure 13-35. Reduction of applied forces to a single force R and a couple M.

cross-sectional area. The total normal stress is found by superposing this

uniform stress algebraically upon the stresses produced by the bending
couple. The shearing stresses produced by the transverse shear V may
be calculated as for a straight beam, using t = (V/Ib)Q.

ILLUSTRATIVE PROBLEM

1339. The circular link shown in Fig. 13-36 has a rectangular

section 100 mm wide by 50 mm
thick. Using Eq. (13-15) and values of
K from Table 13-2, compute the stresses at A and B and at C and D.
Check the stresses at A and B by means of Eqs. (13-13) and (13-14).

Solution: The radius of curvature of the centroidal axis is R = 100 +

50 150 mm. The ratio R/c = 150/50 = 3; hence, from Table 13-2,
=
K 1.30 and K = 0.81. The bending moment at section AB is caused
t
—
by the moment of P about the centroidal axis of the section. Hence
M = 0.150P = (0.150)(50 X 10
3
) = 7500 Nm. Applying Eq. (13-15),
we obtain

a = A — = A—
I bh 2

A
= 1.30
2
=117 MPa compression
(0.050)(0.100)
6(7500)
a„ - 0.81 = 72.9 MPa tension
:

(0.050)(0.100)

Adding these results (Fig. 13-37a) to the uniform axial stress

3
= P - (50 x 10
= - innAym
a- -7 = (0.050)(0.100)
)
ia ° MPa

gives the final values aA — - 127 MPa and aB = +62.9 MPa.

560 13 Special Topics

P= 50 kN

H—ioo—H
mm
60 mm
Era Cross-section

P= 50 kN
Figure 13-36.

At section CD, the bending moment = P(0.150 cos 30°) = M

6495 N-m. The component of P normal to CD is N = P cos 30° =
(50 x 10 )(0.866) = 43.3 kN. Hence, the stresses at C and D (Fig.
3

13-37b) are

N 6M
a - - -r ± K-—2
A bh .

3
43.3 x 10 6(6495)
1.30
2
(0.050)(0.100) (0.050)(0. 100)
- (8.66 x 106 ) - (101 x 10
6
) = -110 MPa
3
43.3 x 10 6(6495)
On - + 0.81
2
(0.050)(0.100) (0.050)(0.100)
= - (8.66 x 10
6
) + (63.1 X 10
6
) = +54.4 MPa
For a more precise solution, we apply Eq. (13-13) to Fig. 13-38 to
obtain

.-*-(y/f)-*-(*//;i£)-*-(y*s)
whence, substituting numerical values,

100 100
150 = 150- = 150 - 144.3
log,(200/100) 0.6931
= 5.7 mm
13-10 Curved Beams 561

K ^£=-H7MPa
.

72.9MPa=tf ^fpF'

10.0 MPa =o

Figure 13-37.

Having determined the value of e, we are now ready to apply Eq.

(13-14), which shows the flexure stresses at A and B to be

a = M i

Ae
7500 (0.050 - 0.0057)
A
(0.050 X 0.100)(0.0057) 0.100
= 117 MPa compression
7500 (0.050 + 0.0057)
* (0.050 x 0.100)(0.0057) 0.200
= 73.3 MPa tension
To these values we must add
the axial stress, 10.0 MPa, giving finally
aA = - 127 and aB = + 63.3 MPa.
MPa
Thus Eq. (13-15) and Table 13-2 give results that agree closely
with Eqs. (13-13) and (13-14), and in addition are simpler to use.
562 13 Special Topics

b •

V//////////,
r d,

_t_ Centroidal axis

e

T
<< ,,

NA

" Axis of curvature

Figure 13-38.

PROBLEMS

1340. A sharply curved beam of rectangular section is 10 mm

thick by 50 mm deep. If the radius of curvature R = 50 mm, compute
the stress in terms of the bending moment M at a point 40 mm from the
outer surface.

1341. For the hook of circular section shown in Fig. P— 1341,

(a) determine the maximum load P that may be supported without
exceeding a stress of 120 MN/m 2
. (b) What stress then exists at 5?
Ans. (a)P = 46.1kN

100 mm dia.

Figures P-1341 and P-1342.

13-10 Curved Beams 563

1342. Repeat Problem 1341 assuming that the hook has a circu-
lar section 75 mm
in diameter. Obtain the values of K and from a t
K
graph of K and K plotted against R/c.
t

1343. Determine the diameter of a round steel rod that is to be

used as a hook to lift a 10-kN load acting through the center of
curvature of the axis of the hook. Assume that R/c = 4 and the
maximum stress permitted is 1 10 MPa. Ans. d — 49.0 mm
1344. A crane hook has a cross section that is approximated by
the trapezoidal section shown in Fig. P-1344. What is the maximum
load P that will not exceed a stress of 80 MPa?

Figure P-1344.

1345. The cross section of a ring is the T section shown in Fig.

P-1345. The inside diameter of the ring is 366 mm. Determine the value
of P that will cause a maximum stress of 120 MN/m2 .

Ans. P= 80.2 kN

20 mm 150 mm
r
100
r
mm fy////////,<A
-
i
20 mm

Section AB

v////////777?A.
Figure P-1345.
564 13 Special Topics

13-11 THICK-WALLED CYLINDERS

In the analysis of thin-walled cylinders in Art. 1-6, the forces trans-

mitted across a longitudinal section were determined from a free-body
diagram of the section (see Fig. 1-11, page 21). A
similar procedure can

be used to determine the force transmitted across the longitudinal

section of a thick-walled cylinder. Dividing this force by the area over
which it acts determines the average tangential stress for either a thin- or
a thick-walled cylinder. The two cases differ in that in a thin-walled
cylinder (the wall thickness being equal to or less than one-twentieth of
the internal diameter) this average stress is practically equal to the
maximum tangential stress, whereas in a thick-walled cylinder it is much
smaller than the maximum tangential stress; moreover, the stress dis-
tribution is nonlinear.
The problem of determining the tangential stress a, and the radial
stress or at any point on a thick-walled cylinder in terms of the applied
pressures and the dimensions was solved by the French elastician
Gabriel Lame in 1833. The cylinder shown in Fig. 13-39 has radii a and
b and is subjected to both a uniformly distributed internal pressure of p t

N/m 2 and an external pressure of p N/m2 This cylinder may be

.

assumed composed of thin shells. Figure 13-40 shows a half-section of a

typical shell, the radius of which is r, the thickness dr, and the length
unity. The tangential stress in this shell is a,; the radial stress on the
inner surface is or and that on the outer surface is ar + dar where dor ,

is the increment in a due to the variation of pressure across the cylinder

r

wall. The radial stresses are assumed (incorrectly) to be tensile, so a

negative result for ar will denote compression.

Figure 13-39. Thick-walled cylinder subjected to uniform internal pressure p, and

uniform external pressure p .
13-11 Thick-Walled Cylinders 565

o r +do r

Unit length

Figure 13-40. Stresses on half-shell in Fig. 13-39.

This shell may be treated as a thin cylinder; hence, for equi-

librium, a vertical summation of forces must equal zero. Thus
(or + dor ) • 2(r + dr) - ar (2r) - 2a, dr =
The product dr • dar being neglected because it is very small compared
to the other quantities, this reduces to.

r
^+ °' o, = («)

Another relation between ar and a, is obtained from the assump-

tion that plane cross sections remain plane, and hence that the longitu-
dinal strain cz is constant for all fibers. Applying Hooke's law for
triaxial stress (see page 47), we have

Since cz E, az and v are all constant, it follows that or + a, is a constant

, ,

throughout the cross section. Let this constant be 2A, so that

or + o t
= 2A (b)

An equation involving only or can now be set up by adding Eqs.

(a) and (b):

do
r
r-r + 2(i
r
= 2A
dr
or

*,)

whence, separating the variables, we obtain

da.
— 2±
<x. r
566 13 Special Topics

Integration gives

- \og (A
e
- O,) -2 log, r+ C = log, r
2
+ C
or

\o% t (A - ar )r 2 - - C
and

(/I - o> 2
= e" c = £
c Solving or we
where B is a more convenient constant than e~ . for

finally obtain

,. = ,-4 (c)

Substituting this value of or in Eq. (b) gives

B
', -A +3 (d)
r

The values of the constants ,4 and 5 are determined by substitut-

ing in Eq. (c) the known values of or at the inner and outer surfaces of
the cylinder. These values are
°r = ~Pi at r=a\
°r = -Po at r " b J

The minus sign for ar indicates a compressive stress.

Applying these values, we obtain

B
-P, = A

which, when solved simultaneously for A and B, produce the following

values:

A =
b
2
-a 2
2
a b\ Pl - Po )
B m
b - a
2 2

Substituting these values of A and B in Eqs. (c) and (d) gives the
following general expressions for or and a, at any point:

-
= «V *V
«V(j*-p.)
o.
*
2
-a 2
(*
2
-aV (13-16)
= «V>, ~ ^o aV(A ~ A,)

V
,
a.
2
* - a2 (b
2
- a
1 3-1 1 Thick-Walled Cylinders 567

SPECIAL CASES: MAXIMUM STRESSES

Case I: Internal pressure only

If the internal pressure is p and

t
the external pressure is zero
(Po
= 0)» Eq. (13-16) reduces to

...
b1
**(-?)
(13-17)

Note that or
b ^K)
is always a compressive stress, and that a, is always a
tensile stress. Obviously a, is always larger than or and is maximum at
the inside surface of the cylinder, where

By representing the ratio b/a by #, Eq. (13-18) may be written:

(
v
)
//max.
= *l±i
^2 "'
i

The average tangential stress may be found by the method used

for thin cylinders in Art. 1-6. Its value is

( \
aP * P >

Wave b- a K-\
Hence the ratio of the maximum to the average tangential stress is

(",)««. K2 + 1
(13-19)

Thus, for a wall thickness equal to one-twentieth of the internal

diameter, or K= b/a = maximum a, is only about 5%
1.1, the larger
than the average ar Hence we may assume, without appreciable error,
that where the wall thickness is one-twentieth of the internal diameter or
less, the tangential stresses are uniformly distributed, which justifies the
procedure used in Art. 1-6.
As a Mohr's circle of stress shows, the shearing stress is one-half
the difference of the principal stresses, and since the failure of ductile
materials like steel (of which most cylinders are made) is assumed to be
governed by the maximum shear stress theory (page 518), the value of
the maximum shearing stress is important in design. It is maximum at
the inner surface of the cylinder where at and ar are both maximum and
is given by

W 5 Ta— 3* (13 ~ 20)

568 13 Special Topics

Case II: External pressure only

If the external pressure is p and the internal pressure/?, = 0, Eq.

(13-16) reduces to

(13-21)
P.

In this case, both or and o, are always compressive, and a, is always

larger than ar The maximum compressive
. stress (a,) occurs at the inner

surface of the cylinder (at this position or is zero) and is given by

( ) = _ J^L_ (13 _22)

(o,) max approaches the value — 2p as b becomes very large compared

with a, as in a cylinder with a small central hole.

PROBLEMS

1346. The cylinder for a hydraulic press has an inside diameter

of 300 mm. Determine the wall thickness required if the cylinder is to
withstand an internal pressure of 40 MPa without exceeding a shearing
stress of 80 MPa. Arts, t = 62.1 mm
1347. Plot a curve showing the percentage increase in maximum
o, over average a, for ratios of thickness to inside radius of thick-walled
cylinders varying from to 3.

1348. A
hoop shrunk onto a hollow steel tube exerts a
steel
MN/m 2 on the tube. An internal pressure of
contact pressure of 20
70 MN/m is then applied to the tube. The inner and outer radii of the
2

tube are 40 mm and 60 mm, and 60 mm and 100 mm for the hoop.
Determine the maximum tangential stress in the tube (a) before and
(b) after the internal pressure is applied- (c) Find the maximum stress in
the hoop due only to the original contact pressure.
Ans. (a) - 72.0 MN/m 2
;
(b) 24.7 MN/m 2
;
(c) 42.5 MN/m 2

1349. A thick-walled cylinder is built up by shrinking an

aluminum tube 20 mm thickupon a hollow aluminum cylinder having
an outside diameter of 150 mm and an inside diameter of 100 mm,
thereby causing a contact pressure of 20 MPa. What is the largest
internal pressure that can be applied to the assembly without exceeding
a tangential stress of 100 MPa at the inner surface? Ans. p = 91 A MPa
13-11 Thick-Walled Cylinders 559

1350. The inner and outer radii of a hollow steel shaft are 50 mm
and 100 mm, respectively. The hub of a steel gear wheel that is shrunk
onto the hollow shaft has an outer radius of 150 mm. The maximum
allowable tangential stress developed by shrinking the gear wheel onto
the hollow shaft is 200 MPa. The length of hub parallel to the shaft axis
is 200 mm, and the coefficient of static friction between the hub and

shaft is 0.40. Determine the maximum torque that may be transmitted

by the gear wheel without slipping on the shaft.
Ans. r=377kN-m
 Inelastic
Action

14-1 INTRODUCTION
The preceding chapters were devoted to analyses in which the stresses
were all in the elastic range; i.e., stress strain. The
was proportional to
maximum permissible stresswas the proportional limit, even though the
ultimate stress of the material was higher. We now consider a class of
problems in which loads producing inelastic strains may safely be
applied, even though doing so creates permanent deformations in the
material when the loads are removed. We revise our concept of safety to
one involving loads that produce deformations which may be large
compared with elastic deformations, but not so large that they produce
collapse of the structure. In other words, we now consider loads that
may be designated as limit loads. A limit load is defined as the
maximum load that can be applied to a structure before collapse begins.
With a ductile material, this condition will not occur until the yield
point has been reached throughout the most highly stressed section in
statically determinate structures. The application of limit loads to inde-
terminate structures is called limit analysis and is considered in Art.
14- 5. It must be emphasized that none of the concepts discussed in this
chapter is applicable to brittle materials; some ductility must always be
present.

570
14-2 Limit Torque 571

C=

(a) Elastic-plastic (b) Elastic-perfectly plastic

Figure 14-1. Idealized stress-strain diagrams.

The stress-strain relation of ductile materials may be approxi-

mated by the idealized diagram shown in Fig. 14- la. The elastic portion
of the diagram is a straight line whose slope is £, the modulus of
elasticity of the material. The plastic portion is also a straight line
beginning at the yield a^, and having a slope C. Slope C is
stress,

smaller than slope E\ therefore the increment of stress required to

produce a specified increment of strain is less within the plastic region
than it is within the elastic region. Such a material is said to strain-
harden', it does not permit an increase in strain without an increase in
stress. A material for which C is zero is called elastic-perfectly plastic,

for such a material, indefinite plastic flow can occur with no increase in
stress beyond the yield point. It has the idealized stress-strain diagram
shown in Fig. 14- lb. In our subsequent discussion, the material will be
assumed to be of this elastic-perfectly plastic type.

14-2 LIMIT TORQUE

In considering the torsion of circular bars stressed into the plastic range,
the only change in the conditions presented in Art. 3-1 is that now the
strains may exceed the shear yield y yp strain, . A transverse section still

remains rigid and does not warp; consequently the shearing strain y
remains proportional to its radial distance from the center of the bar.
We now discuss what happens to a circular bar made of an
elastic-perfectly plastic material that is twisted progressively through
the elastic into the fully plastic range. Until the shear yield point t^ is

reached, the bar is elastic and has the stress distribution shown in Fig.
14- 2a. At the beginning of yielding, the torque is given by
.3
irr
Typ = yp
(a)
572 14 Inelastic Action

If we twist the shaft beyond this point, the shearing strains

continue to increase but the yield stress remains constant, as shown in
Fig. 14- lb. Thus at some intermediate radius ri in the partly plastic case
shown in Fig. 14- 2b, the outer portion will be subjected to the constant
yield stress T
yp
; the inner core remains elastic. The torque carried by the
clastic core is

J, irrt
vp vp

For the plastic portion, it is

T„ - f'p(T„ dA) - T„£pQ«P dp) = ^(r>- n 3 } yp

The total torque is their sum

2m
^~
2
T
yp
+ 1"C 3
n )r yp

which reduces to

r-ZMh (14-1)

The fully plastic case shown in Fig. 14-2c cannot be reached

because it requires an infinite angle of twist. The torque required to
produce it is called the limit torque and is denoted by TL . Its value can
be found by setting r, = in Eq. (14-1). This gives
7 2 3
(*)

which is one-third more than the maximum elastic torque. For future
reference, we express this relation as

h- 4jT„ (14-2)

Elastic- plastic
boundary

yp yp
(b) Partly plastic (c) Fully plastic

Figure 14-2. Shear stress distribution as torque is increased.

14-3 Limit Moment 573

PROBLEMS

1401. A solid circular shaft 80 mm in diameter subjected to ais

torque T. If the yield stress T

yp
= 140 MN/m determine the maximum
2
,

elastic torque and the limit torque. If T — 16 kNm, to what radius

does the elastic action extend? Arts, r, = 33.5 mm
1402. Determine the ratio of the limit torque to the yield torque
in a hollow circular shaft whose outer radius is twice the inner radius.
Arts. 1.24

1403. At what fraction of the maximum elastic torque will the

elastic region extend to three-quarters of the outer radius in a solid
circular shaft? Arts. 1.19

14-3 LIMIT MOMENT

In considering inelastic action in flexure, the assumptions made in Art.

5-1 still apply, except that stresses need not be proportional to strain.
Plane transverse sections are still assumed to remain plane; hence
strains are proportional to their distance from the neutral axis. How-
ever, if the beam is made of an elastic-perfectly plastic material, the
stresses remain constant at the yield stress a^ wherever the strain
exceeds the yield strain c .
yp
The shaded area in Fig. 14-3a denotes that part of the beam that
has been strained into the plastic range as the load P on the rectangular

(b) Section a-a: (c) Section b>b: (d) Section c<:

elastic partly plastic fully plastic

Figure 14-3.
574 14 Inelastic Action

cantilever beam is increased. At section a-a, the stresses on the outside

Fiben have jutt reached the yield stress, but the stress distribution is still

clastic, as shown in part (b). Applying the flexure formula, we find that
the refitting moment at this section is

2
bh , x
Myr " "yp-T {a)

At section b-b, the section is elastic over the depth 2y but plastic t
,

outside this depth, as shown by the stress distribution in part (c). The
stress is constant at o over the plastic portion and varies over the
yp
elastic region. The resisting moment of the elastic region as determined
b\ the flexure formula is

y,

where /, is the moment of inertia of the elastic region about the neutral
axis.For the plastic region, which here is symmetrical about the neutral
axis, the resisting moment is

h/2 . . rh/2
= y(a yp dA) = 2a yp y dA = lo^Q
2f f
where Q is the moment of area of one of the plastic regions about the
neutral axis. The total resisting moment over a partly plastic symmetri-
cal section therefore is

M= ^ + T*„Q (14-3)

At section c-c, the beam is fully plastic; the stress distribution is

constant at the value a
yp
over the tensile and compressive portions of
the section as shown in part (d). The resisting moment that causes this
stress distribution is called the limit moment and is given by

^-2aypG -2«yp(f.|) = a^ (»)

Comparing Eqs. (b) and (a), we find that

ML -\Mn (14-4)

The ratio ML / Myp , here equal to |, varies with the shape of the cross
section. Some values of this ratio for various shapes are listed in Table
14-1. These ratios indicate that the limit moment for rectangular and
circular sections 70% over the yield moment, whereas structural
is 50 to
sections will be only about 10% stronger if inelastic action is permitted.
For beams that are unsymmetrical in section, such as the T beam
shown in Fig. 14-4, the neutral axis changes its location as the section
enters the plastic range. In the fully plastic case where the yield stress is
 M -

14-3 Limit Moment 575

TABLE 14-1. Ratio of Limit Moment

to Yield Moment
CROSS SECTION MJM.
Solid rectangle 1 .5

Solid circle 1.7

Thin- walled circular tube 1.27
Typical wide-flange beam

—^
1.1

100 mm
yp
i r
20 mm
45 mm

Plastic
150 mm NA
109 mm 125 mm

20 mm- yp

Elastic Fully plastic

Figure 14-4. Shift in neutral axis in unsymmetricai section.

constant over the section, the equilibrium condition that the total axial
forceon the section be zero requires that the areas subjected to tension
and compression be equal, i.e.,
[T-C] 0^ = 0^ or A = Ac
t
(14-5)

Apply this condition to Fig. 14-4 and show that, for the given
dimensions, the neutral axis for the fully plastic case is 16 mm above the
neutral axis for the elastic case.

PROBLEMS

1404. Verify the ratios ML / yp

specified in Table 14-1 for the
solid circle and thin-walled tube.
1405. Compute the ratio of the limit moment to the yield mo-
ment for a W200 X 100 beam. Ans. 1.15

1406. Repeat Problem 1405 for a W360 X 91 beam.

1407. Repeat Problem 1405 for the T section shown in Fig. 14-4.
Ans. 1.77

1408. The centroidal axis of the section shown in Fig. P-1408 is

202 mm above the bottom, and the moment of inertia about this
576 14 Inelastic Action

cenimkl.il axis is 260 X 10

6
mm 4
. Determine the ratio of the limit
moment to the yield moment for a beam having this section. Arts. 1.38

Figure P-1408.

1409. A rectangular beam 50 mm wide and 160 mm deep is

made of an elastic-perfectly plastic material for which o^ = 300 MPa.
Compute the bending moment that will cause one-half of the section to
be in the plastic range. Arts. = 88.0 kN- m M
1410. In Problem 1409, determine the bending moment that will
cause the middle three-fourths of the section to be in the elastic range.

1411. If a = 270 MPa, compute the limit moment for the

section shown in Fig. P— 1411. Arts. ML = 177kN-m
80 mm

120 mm 200 mm

Figure P-1411 and P-1412.

14-4 Residual Stresses 577

1412. If a yp = 270 MN/m 2

, compute the bending moment that
will cause the elastic region to extend 40 mm from the neutral axis of
the section shown in Fig. P— 141 1. Ans. M— 171 kN-m
1413. A rectangular section is loaded with a bending moment M
such that M yp
< M < ML . If k denotes the fractional part of the depth
that remains elastic, determine k in terms of M and ML .

Ans. k2 = 3[1 - (M/ML )]

14-4 RESIDUAL STRESSES

Experiments indicate that a ductile material loaded into the plastic

range (curveOAB in Fig. 14-5a) unloads elastically following a path
BC that is essentially parallel to the initial elastic path OA. Upon
reloading, a slight hysteresis loop is formed; but the material now
remains elastic up to the previously strained point B, after which it

again becomes plastic (curve CBD). For an idealized elastic-perfectly

plastic material to which our treatment is limited, this loading, unload-
ing, and reloading cycle appears as shown in Fig. 14-5b.
The principal effect of unloading a material strained into the
plastic range is to create a permanent set. If this permanent set is

restricted in any manner, there is created a system of self-balancing

internal stresses known as residual stresses. The magnitude and distribu-
tion of these residual stresses may be determined by combining the
stress pattern (partly or fully plastic) caused by the given loading with
the stress pattern created by a load equal to but of opposite sense to the
original load. The effect of applying a load equal but opposite to the
given load is equivalent to unloading the structure. This unloading
creates a stress pattern that is assumed to be fully elastic, as indicated in

Fig. 14-5. However, this method of superposition cannot be used if the

residual stresses thereby obtained exceed the yield stress.

Figure 14-5. Unloading and reloading of (a) actual ductile material and (b)
elastic-perfectly plastic material. In part (b) the unloading and reloading lines
actually coincide but are shown slightly separated for better comparison with part
(a).
578 14 Inelastic Action

-T
T-0

>•

m
m
m
*
to
to
to
to
•»

yp

(a) Plastic loading (b) Elastic unloading (c) Residual stresses

Figure 14-6. Residual stresses in torsion.

As a example of residual stress, we consider a circular bar

first

strained into the fully plastic stateby the limit torque. As we saw in Art.
14-2, the limit torque is \ the yield torque, and the stress distribution is
as shown in Fig. 14-6a. To unload the bar, we now apply an opposite
torque as shown in Fig. 14-6b. Recalling that the unloading is assumed
to be elastic, we obtain the stress distribution shown. Superimposing the
loadings and stress patterns of parts (a) and (b), we obtain the unloaded
bar with the residual stress distribution shown in part (c).

An interesting phenomenon of residual stresses is that the bar now

behaves elastically if the original limit torque is now reapplied, as shown
in Fig. 14-7. Combining parts (a) and (b) of Fig. 14-7, we obtain the
original plastic state shown in part (c). On the other hand, with reversed
reloading, the residual stresses are unfavorable and no more than
two-thirds of the yield torque can be reapplied in the opposite sense
before the bar reaches its plastic limit; this is shown in Fig. 14-8.
Evidently, further plastic yielding occurs if the sum of the original and
reversed loading exceeds twice the value of the maximum yield torque.
Under these conditions, the additional yield that occurs after each cycle
soon we know from bending a strip of metal back
results in rupture, as
and forth plastically for afew times.*
As a second example of residual stress, we consider a rectangular
section of a beam that is strained into the fully plastic state by the limit
moment. For the rectangular section, the limit moment is \ the yield
moment and produces the stress distribution shown in Fig. 14-9a.
Releasing the load is equivalent to adding the equal but reversed
loading of part (b) to part (a); this results in an unloaded bar having the
residual stress distribution shown in part (c). Remember that the un-
loading is assumed to be elastic, which produces the linear stress

•For further detail, see J. A. Van den Broek, Theory of Limit Design, Wiley,
New York. 1948, especially pp 23 25.
14-4 Residual Stresses 579

r=o ±L 3 **>

*
*
-*
*
«
»
*•
>
>
*

Typ

(a) Elastic stresses (b) Residual stresses (c) Fully plastic

Figure 14-7. Original torque reapplied to part (c) of Fig. 14-6.

T=0

'yp

(a) Elastic stresses (b) Residual stresses (c) Start of yielding

Figure 14-8. Effect of reversed reloading on part (c) in Fig. 14-6.

M 3
2
Myp M--\Myp M=0

*
n/2
+

|
(
^3 ^
* h/2
f ^
~yp -\°yp 2°yp

(a) Fully plastic (b) Elastic unloading (c) Residual stresses

Figure 14-9. Residual stresses in flexure.

580 14 Inelastic Action

distribution shown Observe that although residual stresses

in part (b).

are self-balancing, if some is removed, an unbalance is

of the material
created. This explains why members which are cold-formed distort after
machining.
As we saw in the case of torsion, a beam that has been unloaded
from the fully plastic state may be reloaded in the same sense, the beam
BOH remaining elastic until the limit moment is reached. For reversed
reloading, the beam also remains elastic, the only restriction being that
the sum of the original and reversed loads must not exceed twice the
maximum elastic moment Myp if further yielding is not to occur. Since
the limit moment is \ M w this restriction makes the maximum reversed
,

moment \ A/ yp for rectangular sections.

As a final example of residual stress, we consider the effect of
bending a straight rectangular bar about a circular die as shown in Fig.
14- 10a. When the bar is released, it springs back through the angle 5 as
shown in Fig. 14- 10b. This springback angle is of great importance in
metal-forming operations. Its value and the relation between the form-
ing radius of curvature the final radius of curvature R can be
R and f
found by combining the plastic strain caused by loading with the elastic
strain caused by unloading. This procedure duplicates that used to
determine residual stresses.
As we saw in Art. 5-2, the strain in bending is c = y/p; hence at
an outside element of the bar, the initial plastic strain in part (a) is

' Ho
and the final residual strain in part (b) is

<=^
' */

Unloading the bar is equivalent to applying a moment opposite to the

limit moment to the deformed bar in part (a). As we saw in Fig. 14-9b,
the maximum stress in this unloading is \o w and hence the correspond-

Flgure 14-10. Springback.

14-4 Residual Stresses 581

ing elastic strain is

a yp
= —= 2

E E
Superimposing these strains gives the residual strain as
= c " ee
V «

or

h/2 = h/2 \ Qyp

Rf R E
which reduces to
3cr
1 w (14-6)
Rf R Eh

The angular change f

caused by the final radius of curvature Rf is
found by applying ds = Rf d9, rewritten as

Rf
and then integrated to give

Since 1//^ is constant and the length 5 of the bend as shown in part (a)
is s = Ro o, we obtain

The springback angle may now be evaluated as

from which, by using Eq. (14-6), we finally obtain

(14 " 7)
•-••Miir)
This result indicates that the relative amount of springback may
be reduced by using a smaller forming radius, or thicker bars, or
material having a low yield strain o^/E. It also indicates the amount
by which the forming angle o must be modified to produce a final bend
of a specified amount.
In circular bars twisted into the plastic state, springback also
occurs after the torque is removed. In this case, the elastic springback is

equal to the angle of twist caused by elastic unloading.

582 14 Inelastic Action

ILLUSTRATIVE PROBLEM

1414. The outer bars in Fig. 14-11 are of 2024-T4 aluminum

alios for which o yp = 330 MPa; the center bar is of steel for which
a^ = 290 MPa. The cross-sectional area of each
is 600 aluminum bar
mm 2
;
900
for the steel bar, mm 2
If P is the
it limit
is load, determine .

the residual stress after P is removed. Assume that the bars are securely
attached to the rigid end plates and that Ea — 70 GPa and E5 = 200
GPa.

Solution: The limit load is the load at which yielding begins in both
materials. Its value is

[P = %Ao yp ] PL = 2(600 X 10" 6 )(330 X 10

6
)

10~ 6 )(290 6
+ (900 x x 10 )

= [2(198) + 261 ]
x 10
3
= 657 kN
Applying an equal opposite load of this amount is equivalent to remov-

ing the load. Assuming this action to be elastic, we apply the procedure
discussed in Art. 2-5 for statically indeterminate members. One relation
between the bar loads is found from the free-body diagram, viz.,

[ZF = 0] 2/>a + />, = 657 kN (a)

Another relation is found from the fact that the bars undergo equal
deformations:

f —
\AE)a ) "(\AEjs ) —
Pa (250) Ps (350)
or P. = 3.06P (*)
(600)(70) (900)(200)

Solving Eqs. (a) and (b\ we find the unloading forces to be

K - 130 kN and P. - 398 kN

657 kN

Al. St. Al. 250 mm Pa P. Pa

Free-body diagram
of unloading
t y > f f ft

100 mm
T*
Flgur* 14-11.
14-4 Residual Stresses 583

An algebraic summation of these results and the original loads de-

termines the residual forces to be

Pa = 130- 198 = -68 kN

P. = 398 - 261 = +137kN

and the residual stresses are

P 68 x 10
3

a„ = — = -113 MPa Arts.

°=A 600 X 10" 6

137 X 103
= +152 MPa Ans.
900 X 10 -6
Observe that after this prestressing, the system remains fully
elastic as any load value up to the original limit load is reapplied.

PROBLEMS
1415. In Illustrative Problem 1414, let the area of the steel bar be
changed to 1200 2
mmIf a load P =7= 600 kN is applied and then
.

removed, determine the residual force in each bar.

Ans. Ps = 54.7 kN
1416. The bar shown in Fig. P-1416 is firmly attached to rigid
supports. The yield strengths for steel and aluminum are, respectively,
290 and 330 MPa. Determine the residual stresses if the limit load is
applied at P and then removed.

300 mm 500 mm

Aluminum Steel
E= 70 GPa
E = 200 GPa
A = 600 mm 2 A = 900 mm 2
Figure P-1416 and P-1417.

1417. Solve Problem 1416 if a load P= 350 kN is applied and

then removed. Ans. oa = - 15.0 MN/m2 ; as - - 10.0 MN/m2
1418. A sandwich beam is made by bonding 4-mm strips of
2024-T3 aluminum alloy between layers of foam plastic to form the
section shown in Fig. P-1418. The foam plastic acts only to separate the
aluminum strips; its effect on bending resistance is negligible. A positive
bending moment of 16 kN-m is applied and then removed. Determine
the residual stresses ifa yp = 300 MPa.
Ans. + 20 MPa on outer strips; ± 40 MPa on inner strips
584 14 Inelastic Action

50 mm

tM 50 mm
&&
'•i>
Tmm
100

1
'•*'!&: 50 mm

Figure P 1418.

1419. The outer diameter of a hollow circular shaft is twice the

inner diameter. Determine the residual stress pattern after the limit
torque has been applied and removed.

1420. The torque applied to a solid circular bar of radius r causes

the elastic region to extend to one-half the radius. Determine the
residual stress pattern after the torque is removed. What will be the
n Jjj
residual angle of twist? Ans. ar = —
1

—-^
24 Gr
1421. A 150-mm length at each end of a straight shaft 1 m long
and 10 mm in diameter is bent at right angles to the shaft. Determine
the angle through which one of the bent ends must be twisted relative to
the other so that they will be exactly 90° apart after the twist load is

removed. Assume that r^ = 140 MN/m 2

and G= 80 GN/m2 .

Ans. 108.7°

1422. A rectangular bar 50 mm wide by 90 mm deep is subjected

to a bending moment makes two-thirds of the bar plastic. If
that
Oyp = 260 MN/m 2
, determine the residual stress pattern after the mo-
ment is removed.
1423. A rectangular bar 30 mm wide by 60 mm deep is loaded by
a bending of 6 kN m, which is then removed. If o
moment •

w = 280
MN/m 2
determine the residual stress at 20
, from the neutral axis. mm
(Hint: Refer to Problem 1413.) Ans. ar - 15.1 2
MN/m
1424. If a beam of unsymmetrical section is loaded into the fully
plastic state, show that any release of the load will cause a residual
stress in excess of the yield stress which, being impossible, means that
the theory of elastic unloading cannot be applied to beams of unsym-
metrical section.

1425. A beam with the cross-section shown in Fig. P-1425 is

loaded with a bending moment that causes the elastic region to extend
for 60 mm
from the neutral axis. If a = 270 MPa, determine the
w
residual stress pattern after this moment is removed.
14-5 Limit Analysis 585

80 mm

120 mm 200 mm

Figure P-1425.

1426. A sheet of steel 10 mm thick is bent over a 90° arc of a

circular die 100 mm
in radius. If a
yp
= 270 MPa and E = 200 GPa,
determine the residual radius of curvature. A ns. R = 10 mm
1

1427. Determine the angle of contact with the circular die so that
the sheet described in Problem 1426 will have a permanent bend angle
of 90°.

1428. A circular die with a radius of 250 mm

is used to bend a

2024-T4 aluminum alloy plate 10 mm Determine the angle of

thick.
contact so that the plate will have a permanent bend angle of 180°.
Assume that <x
yp
= 330 MN/m 2
and E = 70 GN/m2 Arts. 9o - 282° .

14-5 LIMIT ANALYSIS

We now consider the application of limit load, limit torque, and limit
moment to the analysis of statically indeterminate structures. This
procedure, known as limit analysis* is the method of determining the
loading that causes actual collapse of the structure to impend or results
in excessively large deformations. It is applicable only to ductile materi-
als, which in this simplified presentation are assumed to be elastic-
perfectly plastic (see Fig. 14- lb). The method is surprisingly simple,
since it consists of only two steps. The first step is a geometric study of
the structure to determine what part or parts of it must become fully
plastic to permit the structure as a whole to undergo large deformations.
The second step is an equilibrium analysis to determine the external
loading that creates such localized fully plastic parts.

*See J. A. Van den Broek, Theory of Limit Design, Wiley, New York, 1948, for
a concise justification of the principles and an extended application to redundant
beams and other structures.
586 14 Inelastic Action

L-3 m

L- 2m B

fc
h
Y\ 2m
t
2m
t
1ml
1 m
W
Figure 14-12.

One example of limit analysis applied to axially loaded members

has already been presented in Illustrative Problem 14 14. As another
example, consider the rigid beam in Fig. 14-12 which is supported by
two steel rods of different lengths. In the elastic solution, the deforma-
tions of the rods are proportional to the distances of the rods from the
hinge. This condition results in one relation between PA and PB after ,

which the equation of static equilibrium 2A/^ = can be applied to

determine the maximum load that will not overstress either rod.
In limit analysis, however, the capacity of each rod is determined
by the load which yielding begins, i.e., P = Ao w Thus if we assume
at .

the areas of the steel rods A and B to be 300 mm and 400

2 2
mm ,

respectively, and the yield stresses to be 330 MPa and 290 MPa,
respectively, a moment summation about the hinge determines the
maximum value of to beW
[2*ii-0]
_6 6
5W = 2(300 x 10 )(330 x 10
6
) + 4(400 X 10" 6 )(290 x 10 )

W = 132 kN
Observe that yield deformations will not become excessive until this
value of W
is reached. Usually this limit load is divided by a suitable

factor of safety to guard against excessive deformation. For comparison,

if both steel rods must remain in the elastic range, the maximum load

will be reduced to 119 kN.

As an example of limit analysis in torsion, consider the compound
shaft attached to rigid supports shown in Fig. 14-13. The problem is to
determine the maximum torque T that can be applied at the junction of
thealuminum and steel segments before uncontrolled rotation begins.
Since we are assuming the materials to be elastic-perfectly plastic,
remember that yielding is assumed to progress unimpeded after the
is reached. While either segment remains in the elastic range,
yield stress
angular deformations will remain small. Hence excessive yielding will
 14-5 Limit Analysis 587

1.5 m -2 m-
Aluminum ^^ Steel

eJ
70 mm dia. \S 50 mm dia.
=£
G= G= 80 GPa
28 GPa
r = 160 MPa
r
yp
= 140 MPa
yp

Figure 14-13.

not occur until both segments have reached their limit torques. The limit
torque is

TL =*Typ -*(?£\ Typ

3 3\ 2 /

Hence a moment summation about the axis of the shaft gives

[27* -0]
T = |(|)(0.035) 3 (160 x 10
6
) + !(|)(0.025) 3 (140 X 10
6
)

= 18.9 kN-m Ans.

For comparison, both materials must remain elastic, the maximum

if

torque will be reduced to 9.60 kN m. •

Finally we discuss the limit analysis of beams. If we reconsider the

cantilever beam in Fig. 14-3 (page 573), the most highly stressed
section at the wall changes successively from fully elastic to partly
plastic and finally to fully plastic as the load is increased. Other
sections, such as b-b, become partly plastic; to the left of section a-a
the beam remains fully elastic. Only the wall section approaches the
fully plastic state because, although ductile yielding has begun between
sections a-a and c-c, the end of the beam will not deflect uncontrolla-
bly as long as the wall section can absorb an increase in bending
moment. Once a section becomes fully plastic, all its fibers yield without
further increase of stress, thereby permitting the parts of the beam on
either side of this section to rotate relative to each other.* For this
reason, a fully plastic section is called a plastic hinge, and the bending
moment which creates it is assumed to be the limit moment ML .

Collapse of a statically determinate beam is considered to be

synonymous with the formation of a plastic hinge. The outline of a
beam when uncontrolled deformation can occur is called a collapse
mechanism. Several examples are shown in Fig. 14-14. In each case the
dashed outlines represent the collapse mechanism.

*We exclude strain-hardening by assuming the material to be elastic-

perfectly plastic.
588 14 Inelastic Action

f #1 1

1 „

i
Hi
Figure 14-14. Plastic hinges H form at sections of maximum moment.

In general, a plastic hinge will form at a section of zero shear, that

is,where the bending moment is a maximum. The location of plastic
hinges is therefore obvious for beams subjected to concentrated loads
and reactions. For indeterminate beams carrying distributed loads, the
location of plastic hinges is more complex. Sometimes more than one
collapse mechanism is possible, in which case we must compute the
limit load for each possibility and then use the smallest limit load. These
concepts are discussed in the following illustrative problems.

ILLUSTRATIVE PROBLEMS

1429. A beam perfectly restrained at the ends carries a uniformly

distributed load of w N/m over its length, as shown in Fig. 14-15.
Determine the limit load and compare it with the maximum elastic load.

Solution: From symmetry, it is evident that sections of zero shear and

of maximum momentoccur at midspan and at the ends. The collapse
mechanism shown by dashed lines will occur when plastic hinges form
at A, B and C as the moments at these positions each approach the
s

limit moment ML . The redundant supports VA and MA can now be

14-5 Limit Analysis 589

sN^I w N/m N^
^fl
A B B^
L L
vA
2 t
- 1-
2
vB
Figure 14-15.

found by applying the equations of static equilibrium to this collapse

mechanism. However, an equivalent and preferable procedure is to
apply the definition of bending moment.
Since the values of MA MB , , and Mc become M but the sign
all Li
convention of bending moment makes MA and MB negative, we obtain*
[MC = (2M) L ]
*-1t-M?Xt) (a)

[MB = <$M) L ]
-ML = VA L - ML - (wL)(y) (b)

from which

v.-f »nd M^ = 16
(c)

The elastic solution for this case, previously solved in Chapter 7

and Table 7-2, gives
listed in A M = — wL 2 /\2. Ignoring the minus sign
(which merely means that the end moment was originally assumed to be
positive) and letting MA be the maximum elastic moment M yp
and we be
the elastic load, we obtain

wL 2

^yP = 12
(d)

On dividing (c) by (d), the ratio between the limit load and the elastic
load is

M,
Ans.
M yp

The ratio L yp
M /M
(listed in Table 14-1) is appreciable for rectan-
gular or circular sections; but for structural sections it is so close to
unity that ML is usually taken as equal to M^
1430. A propped beam carries a distributed load of w N/m over
its length, asshown in Fig. 14- 16a. Determine the relation between the
limit load and the limit moment.

•It is unfortunate that there are so many L's in Eqs. (a) and (6). Because of
the plan of using the first letter of a word as a symbol, L here has three meanings: it
stands for left in (SM) L for limit in
, L M
and for length of beam. We hope this
,

explanation will eliminate confusion.

f
See Van den Broek, Theory of Limit Design, p. 39.
590 14 Inelastic Action

w N/m

(a)

(b)

w N/m
I C-
X

V-0 v=o
fc)

Figure 14-16.

Solution: The mechanism is shown in Fig. 14- 16b. The loca-

collapse
C is unknown, but it may be found from the
tion of the plastic hinge at
condition that when the moment at C is a maximum, the vertical shear
is zero. This condition was developed from statics (see Art. 4-4) and is

valid whether or not the stresses or strains are in the elastic range.
Drawing the free-body diagrams of the beam segments as in Fig.
14- 16c, where the moments at C and B are the limit moments, we have
from the segment AC
wx.2
[MA -(2M) R ] 0=M (a)

and from the segment CB

ML = ML f» *)' (b)

Solving these equations, we obtain

x = 0.414L and ML = 0.0858wL :

Ans.

1431. Two cantilever beams separated by a roller jointly support

a uniformly distributed load as shown in Fig. 14-17. Determine the
limit load.
14-5 Limit Analysis 591

(a)

B w N/m

(b)

Solution: In this variation of the preceding example, the prop support

is replaced by a cantilever support. This introduces the possibility of a
plastic hinge forming at A. hinge does not form, we have the
If the

situation discussed previously in which M

L = 0.0858 wL
2
If the hinge
.

forms, collapse is possible by rotation about plastic hinges at A and D

as shown in part (c).
To determine we draw the
the limit load for this possibility,
free-body diagrams as shown in part (b\ where the wall moments
become the limit M
moments L Expressing these moments in terms of
.

the common contact force P, we have

[MA -($M) R ]
ML = -P-
wL'
[MD = (2M) L ]
ML = PL
From this, by eliminating P we obtain

wL 2
ML = = 0.167wL 2
592 14 Inelastic Action

Since this gives from L -

a smaller limit load than that found M
0.0858vW \ it determines the desired limit load. Note that two collapse
—
mechanisms were possible in this problem one occurring as described
above, and the other occurring from collapse of BD while A B acts as a
pu^p support. The actual collapse mechanism is determined by the
smaller limit load. It can be shown that the minimum length of AB at
which collapse must occur as shown in Fig. 14- 17c is 0.207 L. (See
Problem 1435.)

PROBLEMS
1432. A bracket is fastened to a rigid wall by three identical bolts
as shown in Fig. P-1432. The cross-sectional area of each bolt is 150
mm 2
. If a^ = 300 MPa, compute the maximum moment M that can be
applied to the bracket. Assume the bracket to be rigid so that the
deformations of the bolts is caused by rotation of the bracket about O.
Determine the ratio of the limit moment to the yield moment.
Arts. M = 13.5 kNm; 1.29

)M
50 mm N

50 mm X

50 mm X

Figure P-1432.

Three rods, each 300

1433.
2
mm
in area, jointly support a load W
as shown in Fig. P-1433. Assume there is no slack or stress in the rods
before the load is applied. Determine the ratio of the limit load to the
maximum elastic load. For bronze, assume that aw = 140 2
and MN/m
E = 83 GN/m For steel, assume that a vp = 240
2
.
2
and E = 200 MN/m
GN/m 2
. Ans. 1.31
14-5 Limit Analysis 593

1434. Determine the maximum torque that can be applied at 1.0

m from the rightend of the shaft in Fig. 14-13 on page 587 without
causing excessive rotation.
1435. Prove the statement made at the end of Illustrative Prob-
lem 1431 that the minimum length of AB at which collapse must occur
as shown in Fig. 14- 17c is 0.207 L.

1436. Determine the limit load for the propped beam loaded as
shown in Fig. P-1436. Ans. PL =—±(L + a)
ab

F
Figure P-1436.

1437. Two cantilever beams are separated by a roller as shown in

Fig. P 1437. Ifboth beams have the same limit moment ML , determine
the load P at which collapse impends. Ans. P = 2ML /a

3a 2a-

Flgure P-1437.

1438. Two
beams separated by a roller support a load
cantilever
P as shown P-1438. Each beam is rectangular and has the cross
in Fig.
section shown. Determine the limit load that can be applied if o
w = 300
MN/m 2
. What is the collapse mechanism if the beam cross sections are
interchanged? What is the situation if both beams have the same cross
section? Ans. P= 31.6 kN

50 mm B 1
100 mm §\

Section
2 m-*4* — 2 m—*[*— 2 m through BD
through AB
Figure P-1438.
594 14 Inelastic Action

1439. A load P is supported by a cantilever resting on a simple

beam as shown in Fig. P-1439. If the limit moment of the simple beam

is three-quarters that of the cantilever beam, determine the load P at

which collapse impends. Ans. P = 3ML /a

Figure P-1439.

1440. Two steel beams are mounted at right angles and in

contact with each other at their midpoints. The upper beam is a
W200 x 27 section simply supported on a 3-m span; the lower beam is
a W250 X 33 section simply supported on a 4-m span. At their cross-
over point they jointly support a load P. If o^ = 290 MPa, determine
the load at which collapse impends. Assume ML = M^.
Ans. P= 206 kN
1441. A restrained beam is loaded as shown in Fig. P-1441. UK
denotes the ratio of the limit moment to the yield moment, determine
the ratio of the limit load to the maximum load at the beginning of
yielding. (Hint: Refer to Problem 713 on page 293.) Ans. \K

Figure P-1441.

1442. Repeat Problem 1441, using the restrained beam loading of

case 6 of Table 7-2 on page 307. Ans. \K
1443. Determine the load P in terms of the limit moment at
which collapse will impend for the restrained beam loaded as shown in
Fig. P- 1443. Ans. P= 2ML /3
2P P

1
2m \ 3 i
if
t
i

Figure P-1443.
m
14-5 Limit Analysis 595

1444. A continuous beam is simply supported over two spans

each of length L. It carries a uniformly distributed load of w N/m over
its entire length, as shown in Fig. P-1444. Determine w in terms of the

limit moment.

w N/m

i?!
?
R2 i?3

Figure P-1444 and P-1445.

1445. Repeat Problem 1444 assuming that both ends of the

continuous beam are perfectly restrained.

1446. Determine the load P in terms of the limit moment for the
continuous beam loaded as shown in Fig. P-1446.
Ans. P = 2.5ML /a

J 1
q^zpzzq
i?2 R3
Figure P-1446.

SUMMARY
Inelastic action is applicable only to ductile materials. In this
introductory presentation, the material is restricted to the elastic-
perfectly plastic type so that strain-hardening effects are not considered.
For solid circular shafts twisted into the plastic range, the torque is

where
"SKh.
r, defines the elastic-plastic boundary. For the fully plastic case,
the limit torque is

n-jT„ (14-2)

For symmetrical beams bent into the plastic range, the bending
moment is

A/ = + 2o^Q (14-3)
yp
596 14 Inelastic Action

where v defines the elastic-plastic boundary, /, is the moment of inertia

of the elastic core, and Q is the moment of area about the neutral axis
of one of the plastic regions.
In a rectangular section, the limit moment for the fully plastic case

II

ML -\M„ (14-4)

For other sections, the ratio of the limit moment to the yield moment
changes as listed in Table 14-1. For structural sections, limit moment
is essentially equal to yield moment.
For unsymmetrical sections, the neutral axis changes its location as
the section enters the plastic range. For the fully plastic case, the
position of the neutral axis is determined by the condition that the
tensile and compressive areas of the section must be equal.
Residual stresses remain in a structure after it is released from
being loaded into the plastic range. They are computed by combining
the stress pattern caused by the actual loading with a stress pattern,
assumed to be fully elastic, thatis caused by an equal opposite load.

The net effect of these two loadings is equivalent to unloading the

structure. Of special importance is the fact that this process will not
apply if the residual stresses thus obtained exceed the yield stress.
In metal-forming operations involving rectangular sections, elastic
springback causes a residual radius of curvature expressed by

R, R Eh
and a springback angle given by

in which h is the depth of the section and R and Rf are the radii to the
neutral axis.
Limit analysis is the process by which we determine the loading
that causes actual collapse of the structure to impend or results in
excessively large deformations. Two steps are involved. The first de-
termines what part or parts of a structure must become fully plastic in
order to permit the structure as a whole to undergo large deformations.
The second applies the conditions of static equilibrium to determine the
external loading that causes the localized fully plastic sections to occur.
 Appendix A

Moments
of Inertia

A-1 DEFINITION OF MOMENT OF INERTIA

Many engineering formulas, such as those relating to strength of beams,

columns, and deflection of beams, involve the use of a mathematical
2
expression of the form Jp dA, where p is the perpendicular distance
from dA to the axis of inertia. This integral appears so frequently that it
has been named moment of inertia.* Moment of inertia applied to areas
has no real meaning when examined by itself; it is merely a mathemati-
cal expression usually denoted by the symbol /. However, when used in

•The term moment of inertia is derived as follows: force is related to the

mass inertia) of a body and its acceleration by the equation F = Ma. The
(i.e.,

equation relating applied forces to the angular acceleration a of rotating bodies is

Fd = [ fp 2 dM] a. If the first equation is stated as force equals inertia times
acceleration, then by analogy the second equation may be stated as moment of
force equals moment of inertia times acceleration. By comparison of the above
statements, the expression f p 2 dM is termed moment of inetia. Similarly, for areas,
the expression / p 2 dA is known as the moment of inertia of the area.

597
598 Appendix A Moments of Inertia

Figure A-1.

combination with other terms, as in the flexure formula for beam

stresses, o = Afc/7, it begins to have significance.

The mathematical definition of moment of inertia, / = /p

2
dA ,

indicates that an area is divided into small parts such as dA, and each
area is multiplied by the square of its moment arm about the reference
axis. Thus, as shown in Fig. A-1, if the coordinates of the center of the
differential area dA are (x, y), the moment of inertia about the axis is X
the summation of the product of each area dA by the square of its
moment arm y. This gives
2
Ix = fy dA (A-1)
Similarly, the moment of inertia about the Y axis is given by

Iy = fx 2 dA (A-2)

The moment of inertia (of area) is sometimes called the second

moment of area because each differential area multiplied by its moment
arm gives the moment of area; when multiplied a second time by its
moment arm it gives the moment of inertia. The term second moment of
area is preferable to the expression moment of inertia; the latter is
confusing when applied to an area having no inertia. The term moment
of inertia, however, is long established and is not likely to be superseded
by the other.

Units and signs

Examination of the integral fp 2dA shows it to be a fourth-dimen-

sional term because it is composed of a distance squared multiplied by
an area. Thus if L is the unit of distance, the unit of / is L 4 A .

convenient unit of L is millimeters; this gives quartic millimeters (mm )

as the dimensional unit of /.
 x

A-2 Polar Moment of Inertia 599

The sign of / is obviously independent of the sign of the moment

arm p (since if p is minus, squaring it makes it plus); it depends entirely
on the sign of the area. We shall define a positive area as one which
adds to the area of a figure, and a negative area as one which reduces
the area of the figure. For a net area, the moment of inertia must always
be positive.

A-2 POLAR MOMENT OF INERTIA

The moment of inertia for an area relative to a line or axis perpendicu-

lar to the plane of the area is called the polar moment of inertia and is
denoted by the symbol /. In Fig. A-2 the moment of inertia of an area
in the XY plane with respect to the Z axis is
2
/ = dA] Jz = fr 2 dA - f(x 2 + v
2
) dA
fp

=j 2
dA + fy 2 dA

Whence from Eqs. (A-l) and (A -2) we finally obtain

h- /, + h (A-3)

Expressed in words, this equation states that the polar moment of

inertia for an area with respect to an axis perpendicular to its plane is

equal to the sum of the moments of inertia about any two mutually
perpendicular axes in its plane which intersect on the polar axis.

Figure A-2.
 Appendix A Moments of Inertia

A-3 RADIUS OF GYRATION

The term radius of gyration is used to describe another mathematical

expression and appears most frequently in column formulas. Radius of
gyration is usually denoted either by the symbol k or by the symbol r
and is defined by the relation

VT or / = Ak 2 (A-4)

where / is the moment of inertia and A the cross-sectional area.*

The following is a geometric interpretation of this relation.
Assume the area of Fig. A- 1 to be squeezed into a long narrow strip as
shown in Fig. A-3. Each differential element of area dA will then be
the same distance k from the axis of inertia. The moment of inertia is
given by

/ - jp 2 dA = k 2 (dA = Ak :

because each differential element has the same moment arm. The strip
may be placed on either side of the reference axis, since if k is minus,
squaring it will automatically make it plus. Or part of the strip may be
at a distance k from one side of the reference axis, and the remainder of
the strip at an equal distance k from the other side of the axis.
In view of this discussion, the radius of gyration is frequently
considered to be the uniform distance from the reference axis at which
the entire areamay be assumed to be distributed. For an area whose
dimensions perpendicular to a reference axis are negligibly small com-
pared with its distance from that axis, the radius of gyration is practi-
cally equivalent to the centroidal location of the area.

/-Area squeezed into a strip

t
k

Figure A-3. Concept of radius of gyration.

•In this appendix, radius of gyration will be denoted by k. In structural

applications, such as in column theory, it is denoted by/\
A-4 Transfer Formula for Moment of Inertia 601

A-4 TRANSFER FORMULA FOR MOMENT OF INERTIA

It is often necessary to transfer the moment of inertia from one axis to

another parallel axis. The method of doing
transfer formula affords a
this without further integration. For example, in Fig. A-4, the moment
of inertia with respect to a centroidal X axis (X ) is given by the
expression Ix — Jy dA. The moment of inertia for the same area with
2

respect to a parallel axis (X) located a distance d from the centroidal

axis is given by the equation

/ - fp 2 dA Ix = f( y + df dA

= jy 2 dA + 2d jy dA + d 2 j dA (a)

The d is written outside the integral sign because it is a constant that

represents the distance separating the axes. The second of the right-

hand terms in Eq. (a) becomes zero because I y dA = A -y, where y

represents the distance from the reference axis X to the centroid. In this
instance y has the value of zero because X passes through the centroid.
We obtain finally,

Ix = Ix + Ad 2 (A-5)
Put into words, this equation states that for any area the moment
of inertia with respect to any axis in the plane of the area is equal to the
moment of inertia with respect to a parallel centroidal axis plus a
transfer term composed of the product of the area multiplied by the
square of the distance between the axes. Evidently the least moment of
inertia for any given direction of an axis is the centroidal moment of
inertia. Note carefully that the centroidal axis involved in the transfer
formula is always the centroidal axis of the area used in the transfer
term Ad 2 .

A similar relation exists between the radii of gyration with respect

2
to parallel axes, one of which is a centroidal axis. Replacing Ix by Ak

Figure A-4. Moments of inertia between parallel axes.

602 Appendix A Moments of Inertia

2
and /, by Ak in Eq. (A-5), we obtain

Ak 2
= Ak 2
+ Ad 2
whence

h,
1 -£,*+** (A-6)
In like fashion, for polar moments of inertia and polar radii of
gyration, we obtain the following analogous relations between any axis
and a parallel centroidal axis:

/-/-KM*] (A _ 7)

A-5 MOMENTS OF INERTIA BY INTEGRATION

In determining the moment of inertia by integration, it is desirable to

choose the differential area so that either (1) all parts of the differential
area are at the same distance from the reference axis* or (2) the moment
of inertia of the differential area with respect to the reference axis is

known (the moment of inertia of the area is then the summation of the
moments of inertia of its elements).
As in the case of centroids, the moment of inertia of a composite
figure may be found by combining the moments of inertia of its parts.
When the evaluation of / for a particular part is known, the transfer
formula (Art. A-4) is used to refer the moments of inertia of the various
parts of the figure to a common reference axis.

ILLUSTRATIVE PROBLEMS

Al. Determine the moment of inertia for a rectangle of base b

and depth h with respect to (a) a centroidal axis parallel to the base and
(b) an axis coinciding with the base.

Solution:

Centroidal Axis. Select the differential element as shown in Fig.

A-5. All parts of the element are at the same distance from the
centroidal X axis. Applying Eq. (A-l), we find the centroidal moment

•When all parts of an element are at the same distance from an axis, this
distance is really the radius of gyration for the element.See Fig. A-3.
 A-5 Moments of Inertia by Integration 603

Figure A-5.

of inertia to be

h/2 h/2
-,-/;
2
dA Ix = f"~y 2bdy = b Ans.
J -h/2 3 12
-h/2

Axis Coinciding with the Base. The preceding result can be trans-
ferred through the distance h/2 to the parallel base axis by applying the
transfer formula as follows:

=^+ w (!)
2
bh>_
[/-/+.«*] /x = Ans.
3
The moment of inertia of the parallelogram in Fig. A-6 has the
same values as for a rectangle because the elemental strips composing
the parallelogram have merely shifted their position laterally from the
dashed rectangle of corresponding dimensions but have not altered their
distances from the corresponding axes of inertia.

c
bh bh<
l
*o~ 12 Ix ~
'
3
Figure A-6.

A2. Determine the moment of inertia for a triangle of base b and

altitude h with respect to (a)an axis coinciding with its base and (b) a
centroidal axis parallel to its base.
 Appendix A Moments of Inertia
604

Solution:

Axis Coinciding with the Base. Select the differential element, as

shown in Fig. A-7. From similar triangles, the length x = (b/h)(h - y).

The moment of inertia with respect to the X axis is obtained from

2 2
!
M -fy
2
dA K - / y xdy = j y -^( h - y) dy

b\ V y_
h[ 3 4
Jo

L= 12
Ans.

Centroidal Axis. To obtain the centroidal moment of inertia Ix ,

we transfer the known value of Ix from the base axis to the parallelX
centroidal axis X . Since the transfer distance is h/3 as shown in Fig.

A -8, we obtain

['.-'. + "*] =I* +

-n \-2h)
b>S
L= 36
t4/w.

Figure A-7. Figure A-8.

A3. Determine the moment of inertia of a circular area of radius

r with respect to a diametral axis.

Solution: Using polar coordinates, select the differential element as

shown in Fig. A-9. From the figure, y = p sin 0. The moment of inertia
A-5 Moments of Inertia by Integration 605

dA - pdBdp

Figure A-9. Figure A- 10.

with respect to the diameter is

-/H t-n
2tt
2 2
p sin pdO dp
o •'o
2
= (' f "p 3 dp -sin 2 0dO
4 4
rim
= —
r
/ sin
,
2
0d0 = ^-7r
r

77T
/ = ^4aw.

An alternate and simpler solution is to use Fig. A- 10 in which the

differential element is taken as the shaded ring of area dA = (27rp)(dp).
The polar moment of inertia is

77T
J = fp 2 dA J = f'p 2 =
•
2-np dp
T
The rectangular moments of inertia Ix and Iy are obviously equal
because of symmetry so that applying Eq. (A- 3) yields

[/-/, + 4] ^- 7 + / - -
or 7^ =
^ Check

PROBLEMS
A4. Determine the moment of inertia of a triangle of base b and
an axis through the apex parallel to the base.
altitude h with respect to
Use the transfer formula and the results of Illustrative Problem A2.
Ans. I = 3
bh /4
606 Appendix A Moments of Inertia

A5. Determine the moment of inertia of the quarter circle shown

in Fig. P- A5 with respect to the given axes.
Ans. Ix = 1= 7rr
4
/16

Figure P-A5.

A6. Determine the moment of inertia of the semicircle shown in

Fig. P-A6 with respect to the given axes. Ans. Ix — Iy = 7rr
4
/S
A7. Show that the moment of inertia of a semicircle of radius r is
4
O.llr with respect to a centroidal axis parallel to the diameter.

A8. Determine the moment of inertia for the quarter circle

shown in Fig. P-A5 with respect to a centroidal X axis.
4
Ans. Ix - 0.055r
A9. Determine the moment of inertia with respect to the X axis
2 2 2
for the area enclosed by the ellipse whose equation is (x /a ) + (y 2 /b )
1 . Also determine the radius of gyration.
Ans. Ix = 7rab
3
/4; kx = b/2
A 10. Determine the moment of inertia and the radius of gyration,
with respect to the Y axis, of the area cut from the first quadrant by the
curve v = 100 — 0.04.x
2
, where x saidy are in millimeters.
Ans. I
y
= 1.667 X 10
6
mm4 ; ky = 22.4 mm
All. Determine the moment of inertia with respect to the X axis
of the shaded parabolic area shown in Fig. P-Al 1.
3
Ans. Ix =jiab

Figures P-A11 and P-A12.

A 12. Determine l
y
for the shaded parabolic area in Fig. P-All.
Ans. I = 2
;a b
3
y
A-6 Moments of Inertia for Composite Areas 607

A-6 MOMENTS OF INERTIA FOR COMPOSITE AREAS

When a composite area can be divided into geometric elements (rectan-

gles, triangles, etc.) for which the moments of inertia are known, the
moment of inertia for the composite area is thesum of the moments of
inertia for the separate elements. Before the moments of inertia of the
elements can be added, however, they must all be found with respect to
the same axis.
In the illustrative problems which follow, the values of the mo-
ments of inertia for geometric elements can be taken from the results of
the problems in Art. A-5; these are summarized in Table A-l. The
properties of a geometric shape that is the cross section of a structural
element such as a W shape beam or an angle or channel are given in the
tables in Appendix B.

ILLUSTRATIVE PROBLEMS

A13. Determine the moments of inertia with respect to the

centroidal X and Y axes of the wide-flange beam section shown in Fig.
A-ll.

Solution: The moment of inertia of a composite area is the sum of the

moments of inertia of the various parts of the area, all the moments of
inertia being referred to the same axis of inertia before the addition is

made.

20 mm Y///////AV///777X

260 mm

20 mm

Figure A-11.
608 Appendix A Moments of Inertia

TABLE A 1 Moments of Inertia for Geometric Shapes

SNAP! MOMENT OF INERTIA RADIUS OF GYRATION

Rectangle

L = 12
K = --
VI2
x
4- bh 3
L = k =
V3

Any triangle

/*- K =
36 Vl8

k b —aV* L = Ml
12
k =
Vo~

Circle
mr
=
/"
t-i
irr
/ = k, =
V2

/
*
- / - 2- K - ky -
y 8 2

Ir = 0.11r
4
kx = 0.264r
U— d=2r^l
Quarter circle

/
*
-2-
16

4
Ix = Iy = 0.055r kx = ky= 0.264r

3
/. = mab *. =

3
irba
h- 4
5-
A-6 Moments of Inertia for Composite Areas 609

With respect to the X axis, the simplest subdivision of the given

area is to resolve it into a large rectangle 200 by 300 mm from which
two smaller rectangles, each 90 by 260 mm, are subtracted. The
centroidal axis for each of these parts coincides with the X axis of the
figure; hence the transfer formula need not be used. Using the result
listed in Table A- 1 we therefore obtain
,

3
bh
L- 12

200 00 )'
200- by 300-mm rectangle: Ix =
^
= 450.0 x 10
6
mm4
90(260)'
Two 90- by 260-mm rectangles: 4= 2
12

= 263.6 x 10
6
mm4
Hence for shaded area: Ix = (450.0 - 263.6) x 10
6

= 186.4 X 10
6
mm4 Ans.

With respect to the Y axis, assume the figure to be composed of a

20- by 260-mm rectangle and two 20- by 200-mm rectangles. The Y
axis is also the centroidal axis for each of these rectangles, so this
subdivision of the area again eliminates the need for using the transfer
formula. Referring again to Table A-l, we have
3
hb
h= 12
3
260(20)
20- by 260-mm rectangle: *- 12

= 0.173 x 10
6
mm4
J
20(200)
Two 20- by 200-mm rectangles: / = 2
12

= 26.67 X 10
6
mm4
Hence for shaded area: Iy = (0.173 + 26.67) X 10
6

= 26.84 X 10
6
mm4 Ans.
A14. Compute the moment of inertia for the composite area
shown in Fig. A- 12 with respect to the indicated X axis.
Solution: The area is composed of a semicircle (S) whose radius is
50 mm, a 100- by 240-mm rectangle (R), and a 75- by 240-mm triangle
610 Appendix A Moments of Inertia

240 mm

Figure A- 12.

( T). With respect to the X axis, the moment of inertia for the area is the
sum of the moments of inertia of these elements, each moment of inertia
being referred to the X axis before addition:

Ix = IR + Is + It (a)

Expressing the moment of inertia of each element in terms of its

centroidal moment of inertia plus a transfer term, we obtain

IK -IK + (Ad 2
)R

2
1S = IS + (Ad )s

IT = IT + (Ad 2
) T
Adding the left- and right-hand members of these equations results in

Equation (b) indicates that the moment of inertia of a composite

figure summation of the centroidal moments of inertia of the
is the
elements plus the summation of the transfer terms for these elements.
This equation is readily adapted to tabular computation, as shown
below. From Table A-l, the values of / for each of these elements are
A-6 Moments of Inertia for Composite Areas 611

IR - bh /\2, Is = O.llr
4
3
, and IT - bh 3 /36. The transfer distances are
indicated on Fig. A- 12.

/ AREA d Ad 2
ITEM (10
6
mm 4
) (10
3
mm 2
) (mm) (10
6
mm4 )

Rectangle 115.20 24.00 70.00 117.6

Semicircle 0.69 3.93 71.22 19.9
Triangle 28.80 9.00 110.00 108.9

Totals 144.69 246.4

Taking the summations from the above table and substituting in

Eq. {b\ we obtain

['* = 2 ' + 2 A<1 2

] h= (144.69 + 246.4) X 10
6

= 391.1 X 10
6
mm4 Arts.

A15. A composed
girder is mm angles of four 150 X 150 X 13
connected to a web plate 600 mm by 20 mm, plus two flange plates each
460 mm by 20 mm, as shown in Fig. A- 13. The properties of the angle
are Ix = I = 8.05 X 10 mm the area = 3730 mm and x = y = 42.3
6 4 2
; ;
y
mm. Compute the moment of inertia with respect to the centroidal X
axis.

Solution: The tabular computation used in Illustrative Problem A 14 is

well suited for cases in which there are many elements. In the present
problem, the elements are symmetrically placed, so a table is hardly
justified.

460 mm-»-j
1 20 mm

150 X 150 X 13 mm
610 mm
*o
600 mm X 20 mm y

] 20 mm

Figure A- 13.
612 Appendix A Moments of Inertia

Direct application of the transfer formula 1 = 1 + Ad 2 to each

element gives
3
20(600)
2 °(^°)
For web plate: / = + (20X600)(0) 2 - 360 X 106 mm 4

For two flange plates: 7 = 2 «1 + (460 x 20)(3.5)

2

= 1830x 10
6
mm 4
For four angles: / = 4[(8.05 X 10
6
) -I- (3730)(305 - 42.3)
2
]

= 1060x 10
6
mm4
6
For entire figure: Ix = (360 -h 1830 H- 1060) x 10

= 3250 X 10
6
mm4 Ans.

PROBLEMS

A 16. Determine the moment of inertia of the T section shown in

Fig. P-A16 with respect to its centroidal X axis.
Ans. y = 87.5 mm; /_ = 113.5 x 10
6
mm4

50 mm

Figure P-A16.

A17. Determine the moment of inertia of the area shown in Fig.

P-A17 with respect to its centroidal axes.
Ans. y = 202 mm; Ix = 260 X 10
6
mm4 ;
/" - 50.8 X 10
6
mm 4
A-6 Moments of Inertia for Composite Areas 613

300 mm
20
Lt
mm

20 mm

300 mm

20 mm
150 mm
Figure P-A17.

A18. The base b of an equilateral triangle is horizontal. Show

that the centroidal moments of inertia with respect to horizontal and
vertical axes are equal.

A 19. Compute the moment of inertia with respect to an axis

passing through two opposite apexes of a regular hexagon of side a.
Ans. / = (5V3/16)a 4
A20. Compute the moment of inertia of the 200- by 300-mm
rectangle shown in Fig. P-A20 about the X axis to which it is inclined
_,
atan angle = sin (4/5). {Hint: Resolve the figure into parts A, B,
and C.) Ans. Ix = 576 X 6
10 mm4
614 Appendix A Moments of Inertia

A21. The cross section shown in Fig. P-A21 is that of a struc-

tural member known as a Z section. Determine the values of Ix and I
y
.

Arts. I - 17.55 x
6
10 mm 4
I = 6.91 X 10
;
6 4
area - 5800 mm ; mm 2

90 mm !

20 mm
20 mm
150 mm *-— X,

I 20 mm

90 mm
Figure P-A21.

A22. Two C200 X 28 channels are latticed together to form the

section shown in Fig. P-A22. Determine how far apart the channels
should be placed so that Ix y
= I for the section. (Neglect the lattice bars
which are indicated by the dashed lines.) Arts, d= 1 1 1 mm

y.

\-
Figure P-A22.

A23. The area of the shaded section shown in Fig. P-A23 is

40 x 103 mm 2
. If / = 250 x 10
6
mm 4 , determine /*2
Ans. Lx -
2
550 x 10
6
mm4
A- 6 Moments of Inertia for Composite Areas 615

Figure P-A23.

A24. The short legs of four 150 x 100 X 13 mm angles are

connected to a web plate 600 by 8 mm mm to form the plate and angle
girder shown in Fig. P-A24. Compute the value of Ix .

610 mm

3-1

Figure P-A24.

A25. A plate and angle column is composed of four 200 x 100

X 13mm angles with the short legs connected to a web plate 350 mm
by 20 mm plus two flange plates, each 460 mm by 60 mm, as shown in
Fig. P-A25. Determine the values of Ix and Iy .

Ans. L- 2910 X 10
6
mm4 ; / = 1140 X 10
6
mm4

360 mm x„

Figure P-A25.
616 Appendix A Moments of Inertia

A26. Determine the centroidal moments of inertia of the built-up

column section shown in Fig. P-A26. It is composed of two 400- by

20-mm plates connected to two C310 X 31 channels.

Y
I

400 mm

100 mm 100 mm

2 - C310 X 31

400 mm
Figure P-A26.

A27. Four Z bars, each having the size and properties de-
termined Problem A21, are riveted to a 300- by 20-mm plate to form
in
the section shown in Fig. P-A27. Determine the centroidal moments of
inertia. Ans. L= 527 X 10
6
mm4 ; /, = 238 X 10
6
mm4

Figure P-A27.
A-7 Product of Inertia 617

A28. A C250 X 23 channel is welded to the top of a W360 X 57

beam as shown in Fig. P-A28. Compute y and the moment of inertia
about the centroidal X axis.

r!~*
Figure P-A28.

A29. Two C250 X 23 channels are welded together as shown in

Fig. P-A29. Compute y and Ix .

Arts, y - 185.6 mm; Ix = 48.5 X 10 mm

6 4

=P
Figure P-A29.

A-7 PRODUCT OF INERTIA

The product of inertia is a mathematical expression of the form fxydA
and is denoted by the symbol P. The product of inertia is not used as
often as the moment of inertia but is needed in such problems as
determining maximum and minimum moments of inertia, unsymmetri-
cal bending of beams, and structural analysis of indeterminate frames.

Units and Signs: The unit of the product of inertia is of the same form
4
as that of the moment of inertia, namely, (length) . Unlike the moment
of inertia, however, the sign for the product of inertia depends upon the
location of the area relative to the axes, being positive if the area lies
618 Appendix A Moments of Inertia

X' Y

Y'

Figure A- 14.

principally in the first or third quadrants and negative if the area lies
principally in the second or fourth quadrants. For example, the area in
Fig. A- 14 lies in the first quadrant of the XY axes and P^ = fxy dA is
positive because all x and
v coordinates of each differential area are
positive.However, with respect to a new set of axes, marked X' and Y'
and rotated 90° counterclockwise from the original set of axes, the area
is in the fourth quadrant. The new coordinates of dA are x' = y and
y' = - x, so that with respect to the new axes the product of inertia is

Pxy = fxY dA = jy(-x)dA = -jxydA = - xy

Not only does confirm the rule of sign stated previously,

this result
but it indicates that during the rotation of the axes there will occurone
critical position at which the product of inertia changes sign and will
have a zero value. When in this position, the axes are known as the
principal axes of the area. Their application is discussed in Art. A- 12.

A-8 PRODUCT OF INERTIA IS ZERO WITH

RESPECT TO AXES OF SYMMETRY

If an area has an axis of symmetry, this axis together with any axis
perpendicular to it will form a set of axes for which the product of

inertia is zero. Consideration of the symmetrical T section shown in Fig.

A- 15 will disclose that, for any differential area like that at A, there is a
symmetrically placed equal differential area at B. With respect to the Y
axis of symmetry, the x coordinates of A and B are equal but of
opposite sign, whereas their y coordinates are equal and of the same
sign regardless of the position of the X axis. Hence the sum of the
products xy dA for each such pair of symmetrically placed elements as
A and B will be zero. It follows, therefore, that the value of fxy dA for
A-9 Transfer Formula for Product of Inertia 619

Y
-X ->

Figure A- 15.

the entire area will be zero if either or both reference axes are axes of
symmetry.

A-9 TRANSFER FORMULA FOR PRODUCT OF INERTIA

Consider any irregular area, such as that in Fig. A- 16, whose cross-
sectional area is A and whose product of inertia relative to the
centroidal axes is denoted by P . Let a parallel set of axes X
and Y be
located so that the coordinates of the centroid of the given irregular
area are xandy relative to these axes.
From the fundamental definition of product of inertia we have,
with respect to centroidal axes,

P^-fx'y'dA (a)

Figure A-16. Products of inertia between parallel sets of axes.

620 Appendix A Moments of Inertia

and with respect to any parallel set of X and Y axes,

x)(y'+y)dA
/\,
-JV+ (b)

Expanding Eq. (b) gives

x»
= jx'y' dA + xfy' dA + y j x' dA + xy j dA (c)

Note that the two middle terms represent the moment of area relative to
the centroidal axes multiplied respectively by the constants x and y.
Since the moment of area relative to centroidal axes is zero, Eq. (c)
finally reduces to

r,y = K + A*y < A -8)

This equation, which is known as the transfer formula for products
of inertia, forms the basis of the method of computing products of
inertia for areas composed of simple geometric shapes. The signs of 3c

and y in this equation may be taken either as the coordinates of the

centroid relative to the X-Y axes or as the coordinates of O with
respect to the centroidal X -Y axes. If the former, x andy in Fig. A- 16
are both plus; if the latter, x andy are both minus: in either case their
product is the same.

ILLUSTRATIVE PROBLEMS

A30. Determine the product of inertia of the right triangle shown

in Fig. A- 17 with respect to the X and Y axes.
Solution: In applying the definition of product of inertia, P= fxy dA,
observe that x and v represent the coordinates of the centroid of the
differential area dA For
. the right triangle illustrated, select the differen-

Flgure A-17.
A-9 Transfer Formula for Product of Inertia 621

tial area as the shaded strip parallel to the base. The area of this strip is
dA — x dy, and the coordinates of its centroid are \x and>>.
From consideration of similar triangles, it is evident that

x = -(h-y)
Hence

dA = x dy = -{h - y) dy

Applying the definition of product of inertia, we obtain

h\ 1 b
[P-fvdA] j^-jf --(A-,) •y (h-y)dy

-£-j\h>y-2hy 2 +y 3 )dy
2h* o

b2 r
AV 2hy 3 y4
2A 2 ,2 3 4 Jo

2 2
A
^-+2 Arts.

A31. Determine the product of inertia of the angle section shown

in Fig. A- 18 with respect to the indicated X and Y axes.

Solution: The angle section can be considered composed of a 100- by

20-mm rectangle plus a 160- by 20-mm rectangle. For the first rectangle,
the centroidal axes parallel to the X and Y axes are axes of symmetry;
hence, from Art. A-8, P^ for this rectangle equals zero. The situation is

similar for the other rectangle. Hence for the composite area, we

I*
20 mm-H

100 mm
120 mm

]
20 mm x
160 mm
Figure A-18.
622 Appendix A Moments of Inertia

obtain

100 X 20 mm rectangle: P^ = (100 x 20) x 10 x 70

= 1.40 X 10
6
mm 4

160 X 20 mm rectangle: P^ = (160 X 20) X 80 X 10

= 2.56 x 10
6
mm 4
For the composite area: P^ = (1.40 -I- 2.56) X 10
6

= 3.96 X 10
6
mm 4 Ans.

If the angle is rotated 90° counterclockwise to the position shown

in Fig. A- 19, the same value of P^ but with opposite sign will be
obtained; that is, Pxy = - 3.96 X 10
6
mm4
This result is equivalent to a.

90° rotation of the axes in Fig. A- 18, thereby showing that P =

- Pxy The . negative sign resulting from the interchange of subscripts is

caused by choosing the Y second axis is

axis as the first axis; then the
the negative part of the X axis lying 90° counterclockwise from Y. In
this respect, products of inertia are analogous to shearing stresses on

perpendicular planes; i.e., t^ = — ryx , as discussed in the footnote on

page 381.
\Y

Figure A-19.

A32. From the answer to Illustrative Problem A30, use the transfer
formula to obtain the product of inertia of the right triangle shown in Fig. A-20
with respect to the indicated centroidal axes.

Solution: From the answer to Illustrative Problem A30, we have

P « 2 2
b h / 24. Applying the transfer formula, we obtain
2l2
Z
-
b h
/% = Pxy + Axy] = P™
xy +
—bh- b h
x — x —
24 2 3 3
2 2
b h
= /% +
18
2l2
b 2l2
2 2 Z
b'h b h h
Pxy = Ans.
24 18 72
A-9 Transfer Formula for Product of Inertia 623

Figure A-20.

Note the minus sign carefully. It confirms the rule of sign stated on
page 617 because here most of the area lies in the second and fourth
quadrants of the centroidal axes. If the_triangle is rotated through 90°
from the position shown, the sign of P^ will change to plus, but its
magnitude will still be b 2h 2 /12.

PROBLEMS

A33. For the angle section of Illustrative Problem A31 (Fig.

A- 18), determine the product of inertia with respect to centroidal axes
parallel to the X and Y axes. Arts. P^ = - 5.17 X 106 mm4
A34. Determine the product of inertia of the Z section shown in
Fig. P-A34 with respect to the centroidal X and_y axes.
Arts. Pxv = - 8.19 X 10
6
mm4

90 mm

20 mm
20 mm
150 mm —
A
1 20 mm

90 mm
Figure P-A34.
624 Appendix A Moments of Inertia

A35. Compute the product of inertia of the triangular area

shown in Fig. P-A35 with respect to the X and Y axes.
Ans. P = 242 X 10
6
mm 4

90 mm

X
240 mm
Figures P-A35 and P-A36.

A36. Compute the product of inertia of the triangular area of

Prob. A3 5 with respect to centroidal axes parallel to the given X and Y
axes. Ans. P^ = - 11.4 X 10
6
mm4
A37. Determine the product of inertia of the quarter circular
area shown in Fig. P-A37 with respect to the given X and Y axes.
Am. P„-r4/9

Figure P-A37.
A-10 Moments of Inertia with Respect to Inclined Axes 625

A38. Usethe result of Problem A3 7 to determine the product of

inertia of the shaded area described in Fig. P-A38 with respect to the
given X and Y axes. Arts. P = r
4
/\2

Figure P-A38.

A-10 MOMENTS OF INERTIA WITH RESPECT

TO INCLINED AXES

In some cases, it is necessary to determine the moment of inertia with

respect to axes which are inclined to the usual axes. The moment of
inertia in such cases can be obtained by formal integration, but a
general formula is usually easier to use.
The problem may be stated as follows: Assuming the values of Ix ,

Iyf and P with respect to the X and Y axes to be known, determine the
values of Iu Iv and Puv with respect to the U and V axes inclined at an
, ,

angle a with the X and Y axes, as shown in Fig. A-21.

Figure A-21. Moments of inertia with respect to inclined axes.

626 Appendix A Moments of Inertia

The coordinates for a typical differential area dA are given by x

and v with respect to the X and Y axes, and by u and v relative to the U
and V axes. The relations between these coordinates can be obtained by
projecting x and y upon the U and V axes. This gives

v ™y — x sin a
cos a )

(«)
u =ysin a + x cos a j
By definition (/ = /p dA),
2
the values of 7M and Iv are

2
Iu
=fv dA (b)

2
Iv
=fu dA (c)

Replacing t; in Eq. (b) by its value from Eq. (a), we obtain

Iu m f(y 2 cos a
2 - 2xy sin a cos a + x1 sin
2
a) <//*

Since

Ix =(y 2
dA, I
y =fx
2
dA, and Pxy =jxydA
this reduces to

7U = Ix cos
2
a + />, sin
2
a — P^ sin 2a (<i)

If the relations

1 + cos 2a 1 — cos 2a
cos a
2
z = and
, .

sin
2
a = ,
(e)
x

are substituted in Eq. (d), the result is

lx + h* Jx~ h^
/„ = + cos 2a - P^ sin 2a (A-9)

Similarly, replacing m in Eq. (c) by its value from Eq. (a) gives

2 2
I9
J (>> sin a + 2xy sin a cos a + x2 2
cos a) dL4

This reduces to
2 2
Iv - /, sin a + /^ cos a + /^ sin 2a (/)

The relations in Eq. (e) transform Eq. (/) into

-
lx + /,
-
t- A,
+ /^
/r * ' cos 2a sin 2a (A-10)
2 2
When values of Ix Iyf and P^ are known, Eqs. (A-9) and (A-10)
,

permit the values of lu and 7C with respect to the and V axes inclined
, U
at an angle a to the * and Y axes, to be determined without further
integration. In a sense, these equations do for inclined axes what the
transfer formula does for parallel axes.
A-1 1 Mohr's Circle for Moments of Inertia 627

Adding Eqs. (A-9) and (A- 10) gives the relation

h+h-h+ Iy

which shows that the sum of the moments of inertia with respect to any
set of rectangular axesthrough the same point is a constant quantity.
This conclusion could also have been obtained from Art. A-2, which
shows that the polar moment of inertia Jz is the sum of the moments of
inertia with respect to rectangular axes passing through the polar axis.
Hence, since Jz is a constant, we obtain as before

To determine the product of inertia relative to the U and V axes,

we note that Puv is defined as

Puv =fuvdA (g)

Substituting the values of u and v given in Eq. (a), we have

2 2
Puv = I (y sin a cos a + xy cos a

— xy sin
2
a — x2 sin a cos a) dA

— -y sin 2a + P^ cos a
2 — P^ sin
2
a —^ sin 2a (h)

whence by using the relation cos

2
a — sin
2
a = cos 2a, we obtain

Pm = — r-^sin 2a + P cos 2a (A-ll)

The angles defining maximum and minimum moments of inertia

may be found by differentiating Eq. (A-9) with respect to a and setting
the derivative equal to zero. For these values of a, it will be found that
the product of inertia is zero and the maximum and minimum moments
of inertia are:

4± 2+(/>j
^f
'

(a_,2)
= -
V(^)
A-11 MOHR'S CIRCLE FOR MOMENTS OF INERTIA

Except for a change of symbols, Eqs. (A-9) and (A-ll) are identical
with Eqs. (9-5) and (9-6), which express the variation in normal and
shearing stress. It follows that the method described in Art. 9-7 for
Mohr's circle of stress may be similarly applied to obtain a Mohr's
circle of inertia. This construction will give a visual representation of all
the possible values of / and P with respect to all axes passing through a
528 Appendix A Moments of Inertia

specified point inan area. For convenience, the rules in Art. 9-7 are
rephrased as follows to apply to moments of inertia:
1. On a set of rectangular coordinate axes choose one axis on
which to plot numerical values of moments of inertia and the other on
which to plot products of inertia; call these the / and P axes. Plot points
having the coordinates ( /x P^) and (Iyf -P^). The values for Ix Iy
, , ,

and P are assumed to be already known. Note carefully that the value
for Pxy with its real sign is associated with the value of Ix and that the
value of P with the opposite sign is associated with I
y
. Actually Pyx
should be paired with I , but Pyx — — P^, as we saw in Problem A31.
2. Join the points just plotted by a straight line. This line is the
diameter of Mohr's circle having its center on the / coordinate axis.

Draw the circle.

3. As different axes are passed through the selected point in the

given area, the values of / and P relative to these axes are represented
by the coordinates of points whose positions shift around the circum-
ference of Mohr's circle.

4. The radius of the circle to any point on its circumference

represents the axis of inertia corresponding to the / coordinate of that
point.

The angle between any two radii on Mohr's circle is double

5.

the actual angle between the two axes of inertia represented by these
two radii. The rotational sense of this angle corresponds to the rota-
tional sense of the actual angle between the axes; that is, if the U axis of
inertia is located at a counterclockwise angle a relative to the X axis of
inertia, then on Mohr's circle the U radius is laid off at a counterclock-
wise angle 2a from the X radius.

ILLUSTRATIVE PROBLEM
For the rectangle shown in Fig. A-22, compute the values
A39.
of /u , and Puv with respect to the U and V axes. These axes are
Iv ,

inclined 30° counterclockwise to the X and Y axes.

Solution: The moments of inertia and the product of inertia are first

found with respect to the X and Y axes, as follows:

150(300):
/ -
12 12
= 3375X10W

Iy
hb 3 1
Iy = —y-'
300(1 50)
3

=84.4X 10
6
mm 4
P^ = because X and Y are axes of symmetry
 ~

A-1 1 Mohr's Circle for Moments of Inertia 629

U-150 mm^-
Flgure A-22.

Following the rules given above, draw a set of rectangular coordi-

nate axes and label them / and P, as shown in Fig. A-23. Using the
values of 7X , Iy , and P , plot points A and B whose coordinates are
(337.5, 0)and (84.4, 0).
According to Rule 2, the diameter of Mohr's circle is AB. Its
center C is midway between A and B. The / coordinate of C is 211.0.
The radius of the circle is the distance CA — 337.5 — 211.0 = 126.5.
From Rule 4, the radius CA represents the axis of inertia corre-
sponding to the / coordinate of A in this case the X axis. Applying Rule
,

5, we find that the U axis of inertia is represented by the radius CD laid

PX 10

6
84.4 3 37.5 I X 10
A

Figure A-23. Application of Mohr's circle.

630 Appendix A Moments of Inertia

off 60° counterclockwise from the X axis (CA). Also, since V is actually
90° from U, the V axis (C£) is laid off 180° (i.e., double scale) from the
V axis; Z), C, and E form a straight line.

From Rule 3, the coordinates of D represent /„ and Pw \ the

coordinates of E represent Iv and Puv with the opposite sign. Accord-
ingly, from the diagram we obtain

[/„ x 10" 6 - OC -I- CD cos 60°]

/„ = (21 1.0 -I- 126.5 cos 60°) x 10

6
- 274.3 X 10
6
mm4 Arts.

-6
[ /,
x 10 = OC - CE cos 60° ]

7t = (211.0 - 126.5 cos 60°) X 106 = 147.8 X 10

6
mm4 Arts.

[Puv X 10" = 6
CD sin 60°]

P = (126.5 sin 60°) x 10

6
= 109.6 X 10
6
mm4 Ans.

A-12 MAXIMUM AND MINIMUM MOMENTS

OF INERTIA: PRINCIPAL AXES

An inspection of Mohr's circle will show that the points whose coordi-
nates indicate maximum and minimum moments of inertia are located
on the Iaxis and have a zero product of inertia. Conversely, axes which
have a zero product of inertia must be axes of maximum or minimum
inertia. Such axes are called principal axes.
As we have already seen (Art. A- 8), the products of inertia
relative to axes of symmetry are zero. Hence we conclude that axes of
symmetry must be principal axes because they always yield values of
maximum and minimum moments of inertia. But many figures do not
have axes of symmetry, although they do have principal axes with
respect to which the product of inertia is zero. Axes of symmetry are
always principal axes, but the converse is not necessarily true.

ILLUSTRATIVE PROBLEM

A40. A is found to have the following values with

certain area
and Y axes: Ix = 100 X 106 mm4 Iy = 60 X 10 mm
6 4
respect to the X , ,

and Pxy = 15 X 10 mm Determine the maximum and minimum mo-

6 4
.

ments of inertia and illustrate the position of the principal axes relative
to the X and Y axes.
A-12 Maximum and Minimum Moments of Inertia: Principal Axes 631

IX 10

Figure A-24. Maximum and minimum moments of inertia.

Solution: On a set of I and P axes, as shown in Fig. A-24, plot points

having the following coordinates.

Ix = 100 x 10
6
I
y
- -60 x 10
6

Pxy - 15 X 10
6
-P = -15 x 10
6

Note that the given value of Pxy is associated with Ix and that the value
of Pxy with the opposite sign is associated with I If Pw had been
negative originally, this value would have been associated with Ix and
;

the positive value of P^ with Iy .

Plotting these points gives two points on Mohr's circle. Joining

them gives the diameter of the circle shown in Fig. A-24. Obviously the
radius of the circle is CA = y(20) 2 -4-
(15) =25. The maximum and
minimum moments of inertia are located at B and D; hence
[Max./ X 10" 6 = 0C+ CB]
Max. / = (80 + 25) X 106 - 105 X 10
6
mm4 Aw.
[Min. / X 10" 6 = OC - CD]
Min. / = (80 - 25) x 106 = 55 X 10
6
mm4 Arts.

To go from the X axis to the axis of maximum inertia CB, we must

rotate clockwise through an angle 2a. From the diagram

tan 2a = AE tan 2a = —= 0.75

CE
2a = 36.9° and a - 18.45° Ans.
Angle a, which locates the axis of the maximum moment of inertia
(the U axis), is also rotated clockwise on the original reference axes; this
gives the position shown in Fig. A-25. The axis of minimum moment of
inertia (i.e., the V axis) is at 90° to the U axis.
632 Appendix A Moments of Inertia

Figure A-25. Location of U and V axes of maximum and minimum moments of

inertia.

PROBLEMS

A41. For a certain area it is known that Ix — 80 X 10

6
mm4 ,

Ly - 40 x 10
6
mm 4
, and w =
P*y 0. Find the moment of inertia of this
area with respect to a U axis which is rotated 30° counterclockwise from
the X axis. Ans. Iu = 70 X 10
6
mm4
certain area has the following properties: Ix = 40 x
A42. A
mm4 100 X 10 mm P^ = 40 X 106 mm4 Determine the
6 4
10
6
y
= ; I ; .

maximum and minimum moments of inertia, and also the angle that the
axis of maximum moment of inertia makes with the X axis. Illustrate by
Ans. Max. / = 120 X 106 mm min. / =
4
a diagram. ;

20 X lO mm ;0 = 63.4°
6 4

A43. A right triangle has a base of 300 mm and an altitude of

600 mm. Determine the maximum and minimum moments of inertia
with respect to principal axes passing through the centroid.
Ans. Max. / = 1936 X 106 mm4 ; min. / = 314 X 10
6
mm4
A44. The properties of a given area are A = 8000 mm Ix = 16.0
2
,

x 10
6
mm 4
, /
y
= 340 x 10
6
mm 4
, and P„
xy
- - 14.0 X 10 mm Com-
6 4
.

pute the minimum radius of gyration for axes through the centroid.

A45. Determine the maximum and minimum moments of inertia

of the Z section described in Problem A21 on page 614 with respect to
the principal_axes passing through the centroid. From Problem A34 it is

known that P^ = - 8.19 X 10

6
mm 4 .

Ans. Max. / = 22.0 X 106 mm4 ; min. / - 2.46 X 10

6
mm4
A46. Show that the moment of inertia for the area of any regular
polygon is a constant with respect to all axes in the plane of the area
which pass through its centroid.
A-12 Maximum and Minimum Moments of Inertia: Principal Axes 633

A47. Show that the moment of inertia for the area of a quarter
circle with respect to its axis of symmetry is {it — 4
2)r /\6, where r is the
radius.

A48. The figure for Problem A20 is redrawn as shown in Fig.

P-A48. Check the_ answer to Problem A20 (i.e., find Ix ) by first
computing Iu and Iv then using Mohr's circle to find Ix and finally
, ,

transferring this value to the X axis.

A49. Use the method discussed in Problem A48 to compute the

value of Ix for the area shown in Fig. P-A49.
Ans. T = 1040 X 10
6
mm4

150 mm— ^-J

Figure P-A49.
 Appendix B
Tables

B-l. Average Physical Properties of Common Metals

B-2. Properties of Wide-Flange Sections (W Shapes)

B-3. Properties of I-Beam Sections (S Shapes)

B-4. Properties of Channel Sections

B-5. Properties of Equal Angle Sections

B-6. Properties of Unequal Angle Sections

Acknowledgment

Data for Tables B-2 through B-6

are taken from Metric Struc-
tural Steel Design Data, 1978,by permission of the Canadian Institute
of Steel Construction. Wide-flange (W shapes) and angle sections are
designated by the Canadian Standards Association (CSA) Standard
CAN3-G312.3-M78.

634
 -

w s
.a

1
1
ill
w w 5 en
OO «r> © "§Q oo
•Js - ««o
"> 8 1 S2 ^
1
c/5
1
1o '€
s (4 I
O O
^ a
«-i

o
o o
OO
O
00
o SSS o o CO CO
Q
^ ««0
CO ? f fl
CO U §
J c % §
14)

8 s 8 8 8
<N — — —
Prt © OO a
~* 2
<N <s <N -*
,o &

-a
O O
<N
O
«r>
O
co "Q n o o
tt r- es
°
m 1 i
1 (A

CO ^ «r> r* CO <N <N CO '

CO
«2

3
•Q -C -O c o «»0
•c -c
CO
c
1 t
h
© o 8 o
en
O© © £8 O O o
00 CO 00 i
2
1>
5 *r\ VO OO - n n en CO <N CO
5 S3
-3

O © o 8 o a
cm CN fN
"*
§ § g § 2?

2 H ^ a
d
SSI o
J3
o o o o S3
o ?
(N
a Tt
OO
«3F <N <N <N Els <N
8
a &
p* I
Ph
I 13
c
3 < o a .23 >»
1
PS 55
u £ I
g
«~o rs a^.
*
~
OO 0\
© ~ H
P-4»* r* ooo 'o
4)
«9
—
_! co co co 00 00 NO
c ~
4>
5
«J O > 13 cu a
> < •4-1
c2
o
CA
o
00
S
<4Z S
3
3 ^t §«
O o
»«o
o o
IT) 888 ««o © 888 3 •&§
|
OO OO oo OO rs cm r^ vo r-» •n (S oo s w
^« 4>
S3
TO J3
r- r^ h» r- r^ r>~ r^ <N <N oo oo oo
T3
C
Ol 8 >
a'S
12 3
o ^_,
n. «« &
2 H I ^2a
3 Q..3

8 d "8 a
CO
6 -
c«
a
o
I
1
1 11 O
*• >»
1 1 c
o 1° .a ^2
s 2 3 1

H
O
*o 1 •bog a a
<n
M 1 §1 — a
© d a a 2 D. t 3 d 1-S
c/f a ^•8
O a o S
VD
*1
S J
£ 33 OQ
ro
pq
o
U

635
 noK oo ri vD -<r rs oo en O r- r* (N N ^ »» « «
1
l^f r^
fl\ ff»
in «n
^ ff> £ SO 33 i»4
SO 3
i
mm
<N <N <N <N
— ©oo — oosoooor--
— — —
on r*» m t*» t^ *© «o

g-n-oo>o«ft^nNTt
^OiflNjhVMtnM-* » a <n ^f n o »
00 Tf — — o ^ r~
«/> m ^- ^ m — — — — — — o m m — —

^ o -
*«ftfSO
oo oo oo oo
r*"i ro c*"> Ci cj fO c>
$ m
m f>
o\ r* "n
«/"> «n m n o
Tf ^ m*o
rn
ro c> c«"> c«"» rn f*"> c>

© © © ©
<n oo r- ©
N iO » (S (S
On oo 00 l*-

OOOOOQOOQOO
— r^fNTtor^f^or-<N
oo
O
— O © ©
«n ^ ©
0\ ro r- ^
O
— O Q
— oo ^
r^ ^
iftin ^ n n nn

— —
^ p
r^ <n ^; r-»
^rc>i
<N <N <N <N
— ©o\ — osodr^voviin
— <N — — — — — —
-' o\ oo <> vi #

^ :

r-;a\vornp«nppa\.ONON — VO ^ (N 00 ^ -
h; OO
wS rj o\ »o ^-' oo
W n (S (S (S M -

C-><NOCT\OO^OOr^^"£>33
fSNrs - — ooooooo ss ** en <N <N
^\ ^\
M (N
^
ON
N
<S

r>oo^>0(SMt^nairi-n

N«ovq^r«;iq^h;n<sn ^ n 'O oo in q
Ko6r^^rNi<No6-^m'o6^ —
t* — oo <) Tt - oor-mmrNO
:
On On
N <) ©
fi vO
- O
5 TcnmmfNfNfS<N(N<N *
C) <N
(N
<N <N — h
—

— cnoom —
r-r-v-><Ncoo\
«oo«^-oo>mnfSO On 0\ On
«n <n o\ <n
^ O
— ro
o ^O
r»
^mmrorn<N<N<N<N<N«N
XXXXXXXXXXX X X X X X X X
 ' «

oo p
co ~ *n r-# v> ~* OO N hri ^O "O a •^f-vor-ONcot^r^r-- O O
06 06 r»* r»* v© «/S tj-* co cn cn
» ^*n co cn
00 00 »n «n »n <n
^ r^vo«r>^"cooONoor^
r-r^r-.r--r^mTtTtTr
«/->
co co
^>

O O O co <r> r- o> o - - O (N CO QOn ^ VO

r~ — »-
^ co vo vo —
«n «n
vn ov
•

co ^ cn g> r^ vo
— On 00 r-
Tt O«T> NO CO
co CO CN ~ —

r«. ~ r-. r-; o\ (N OO h - ~* CO CO «o ^ ~*

VOOO^^OOQCN
•

~*oo«noor^v©vow->
CO CN CN
CO O 5 s

r*-w->cogvoo«ncNoo Tj-CN — O
Ov VO CN Tf-CNOt^vOOONVOcO^-CN
>OvOvo<nvN>o^'t^'f fin
CNCNCNOOOOOn
cocococococococn
On On On 00 P"* t*"* P*
M (S N M N N N CNCNCNCNCNCNCNCNCNCNCN

O Q O O Q Q O O O Q O O O O
cn *n ^ — 00 00
000Q 00000000
00 r^oocNcofSooco^tr-
t--00 CN 00 ^ On ^ CN ^t r- ON co Ov v-> r^ ^ r^<Nv©<Noo«n^oo
9\ 00 vo m m ^ ^ 00 r^ vo ^ rf co co VO vo 1^

00000 ssgSvig: OOOO0O«n«r>^

*/">~oor^ON<Noor^vo s
(N 0O t ^
(N vo co r- v> co »-* ^Onvo^-cN'— ONoor^
^- CO CO CN CN CN CN CN ~ —

r^ovovoovTtoocs ^ 00 ^- »r> ~* -t r- ON*n^pr^-^ON<N«oovp

^oovotn^^-rin 00 vo <n ^ n m -^

"*. *"?^ ^ P
""! ^. **1 cn Tt 00 vq -^ on co Or^^-vOOCNvOcOOvOOO
coor^irico^ONr^
cococncncncn — — n^
n« ^n
r*
(S
06 vd
(S -* -

^(S^-oor*^vo>n
oooooovovovovovo ovoor-u-i^-oovooooosoo
cNcNcNCNCNcocNcNcNr^r^
COCOCOCNCNCNCNCN COCOCOCOCOCNCNCNCN^ —

VOOVCOOVOCNOOCO vp ~- *n co 00 ^- 00 »O00CNvo-*r^CN00C0C0Ov
»i^h-i^vovo"n«o O O ON On 00 00 t^
r-~ r-~ vo vo vo vo vo vovovovovovovovovovo>/^

888888
- ^
-H (N OO --
888
t- vo r-
88888
00 00 On CN r^
O r-» Tt <N on r-
CO CN CN CN — —

^Osvqoqoqvo^;^ 00 on -- 00 vo r^ on vq CO on ~ -* * r~ co on
^cor^vo^coor^ * Ov r-~ Ov CN Ov «n ~ — ON
r-' «/S
^oo«no\oor^-voTt vO co •— vo «n co cn
CN CN CN
•

~* ^ ~* — '
^
CN CN —

Tj-Ttr-vo^co-.r-
-^oo«noNoot^-vo^-
«£
vO * - O
O
r» cn
~ m
|^
O «n
— *
— (S
~
xxxxxxxx CN <N CN
X X X X X X X xxxxxxxxxxx
o
s
I I

637
 r^^^ONCOCN — On<NOnOnv©
til
mm — ©<NONt-~oom«r>c*NON
co^on<n«o«/->c<-> — o\ s
I - in «n
I
B cO co <N <N <N — — — —
10 ©

hxiflO\oon*otv) t«fnn-MO\»o*oio(s

co— — o on — ©ONONr-Tt<Noo«£
,

c->cocOCO<N<N<N— — — — — ©©
:| <NCNrS<N<N<N<N<NCN<N<N<N<NCN

s
OOOOQOOOOOOO
© — —
oo »n oo m oo r*» ^p rsi
— oo-<rco©oooo«nco
r*.
c*->

COCO<N<N<N<N — — — —

O © © © © — — j^r^ »/-> — — ©v©v©r«-NOoO«n©©cO©ON

— » n in «n » 5 - r^mo«n
mcors — ©oor-N©NO St ^ co Tf rf rr co m co <N
h

co «n <n p r-_ r-; — \0 9\ (S V} M ^ O; v© p no <s oq vo rt «0 On «/"> © —

oo ^o «r! ^ <N Tt rS — O O On O On 00 no «/S ci c4 ©
<n —
©on'onoooo

<Nm^<Nmvo<Noo'*tvom io « ^ onon — NOcoNoor^o«nr-oo

r» co NO CO — No'cocNONr-'cJoNr^NOTrcNio
szl «/*>

OONO«ncO<NTtfS — ©ONONN©N©«n NO
00
Q 1 (N (N <N <N <N

CS^OfSr-COONVOCOOr*-^-©
Mt^t^NONONONONOvoioinin
mift^in/i^^iniriin^>oinin

8888888888888R
Ono — —«<NNor^ONONOoinoo«oco oooooo©op«r>oo^
NO-^^t^«Om^^^NOh
I s io n - On r-~ »r> m <N — OOOnoo <N©00VOTtCO<N — ©
(SCS — — — — — — —
On 00 f-

ON«nr^coNqcofNpxt«ri^-r-r^r-- cni*^NO^ — oqvocnONfS — on

»>£> — NQ©o6c«Sg> — CNifNiTt^*/-;
— ONooNO«nco<NOOONoooor--No
— — — — .— ^ _» _
hV ^
r^«r»^-fs
od n
— «n no o\ ^ ^
©ONoooor-^oN©© oo"

cn

OnvOCSnOOoOcOOn— 5 M n qq
-0\x<omnNOOM(Svi^« q j^ CO NO

xxxxxxxxxxxxxx —xxxxxxxxxxxx
o
— — — — — ONoooor~>vovo
en
00 Tfr ON
in ri
co co co
d On r- NO ^t (N ~ o s B8

8 O O © © © Q O© ©
oo~coooo\r-««/->co~r^
no r- « ^co oo «n co u->
*C (S o\
Tfr OO <N
(S o\ NO CO *-"
oo r^ 'O^OVi^^^nwnNNJSN-

^ O; n r»^r<N«novoooo -2S OnM"A^O-^O-«00

o\ i^ \d «0 T*
s ©
MM© © O
h >o <0
©
m rs

r^ co oo tt
oo r- r^ r- r- O O ON ON
<N <N -« ~

O O
On <N
<N ~* ON
co
-«fr
OOOOOOOO^fOTT
r-roofN^moNP<Nr«-m
ooin<NON«nro<NOONr-NO
8QOOOOOOO
Ttr-oo^ON^"mr^
ooON^fN«or^^u->o
v oo r- so vo «o «r>

t^ «n <n ONOO«NOO»n«nvONONOvor~-
On «r» -^
CN <N <N
NOONO<N<NONTtONO«r>00^«nOON
ON^ir>ONTtONt^«n<NOOONOTtfS^ON

*- O
no On co vo © On r-; oo r-^ >n p -^ poNON«o*ocNvqpp^oooqoq<NONNO
on oo r^ Tt CO ~ O © On o6^«nc>«o^r^«nr4oN«nfNONr--"'^ fN
r-r^vONO»n«0^'TtTtCOCOCO<N<NfNfN
:

^ co oo O <N CO On O ^" oo On (S oo on «n ^ co no r^ ^ \o ,
o
r-» no vo
u-i co © uS CN On NO
(N <N ^ - -
<
n©
— _ "*t CN
,* _-
© —
^
° 00 » - h
oo oo r-»
(S h NI^ (N OO CO ON NO
r- no no «r> «n ^t ^ n n

«oco— O^OONOOr->QQ
vovovovooooor-r-r-^^ ^OOCNr^CSOO^t-^OONOCSONNp^^ON
«n^^-coco<NCN<N~'r--<pppooN
(S M<N (S

on «n
m «n «n
o ~*</->©mr^co©r«-cocoON
cocncs^ — — ^QQQOn On O -« ^ oo
no «n co ^h on
tj- m «r> \o »n v*» vo
^ CO (N ~*
r*~
© On co
ON ON
Tf Tt tj- »r> «n «n «n ^- ^ Tt

88888— oo «r>

^O O »n O

r)Hinioo*«nin^MM ONO<NOcooocoNONo«r>oor-TtO'^-vo
voon r-^cNON^-t^^tcNOONCs^csr^cor^
oooNO^"n-r-cooN«oovo<Noo'^-— a
^t co <
OONONoor-vovo^min^Tfcorococs

O ON <N "<* O oo© ooTfrr--^"<N-*ONco~*

^^i^mo\"nONorj cn t^ <* r^
^ —
OO CM
so vo «r> -Hrt-MrtOOh-NOvOin^W ° &
^ On ootv»NONOininiri^,,t
OO OO
CO CO CO <N
X X X xxxxxxxxxxx
o
X X XXXXXXXXXXXXXX
8
I

639
 T _,
«n<Noomooor-<NONvn oo oo ^ ^
K SO

r- so oo
OOOOOO^t—
o
s© i^— OO — <n
oo«or>msoovoo\r^^noo
h n a <n n « oo <n — o\ m

v>. r^ so oo <n ^t
OOmCTvr-sor^- — — vi o 3;'^ — oo — Os oo ro <N
— oo<NOooso«nso«r>«/->^-<N<N — —
mro«N<N<N — — —
«/->

ooooooooooooooor-so — o^t
so ^ oo«fiMO»^>oN"-
— — o oo r- so vn tj-

^^nnriNrSNnc-~

•^ oo <N s© *r> so ro — <N r^ — oo — —

^\ oo —
mi^- so — — in
«r>
sorooso<NOr~so^-rNO

on ts o> «n oq
— oor^vou-ifncsi — co — ©oson r^ r^ so so «ri

ro <N r^ <Nasooooor^a\m^"oo — «/-> — so oo

cS o
r^ rn — Os

ONOsa\r^t^-r^r~so«r>«r>«n«riOOot^t^r^(N<N
rororofOromrom<N<N<N<StS<NfS — — — — —

r-o«n<Noo^ovOfnor^m^-or»>oo«n<NfnoN
oooor^r--sososo«nsoso«n«r>»nm^«/">inin«n'^'
!

88888
—
8QQ0000000
oo — ^<Nvnmoor-
Osso — — — <NTtr-<7\ —
oo so oo «/->.

oo h m <N — OOsoOt^-so»nTfTt

r^cnro«rjcsip«opr-;<Ncsiooco«nost-* sop — oo
'- o — o o> — rS so o
#

risdsdsdo\(Nr^^ :
o> <s
soro — or-so^-r^csi — OOSr~-r^so«r>»n'^"com
«/•>

tNr-sosoa\<Nr-Tt<so—
* _ —
O- Osr-sOTfr^<N — — Os^f^^O __,
—
<N<N<N — — — — — — — OZT^J^5k
r- «n as
m
,

•n «n ^"
xxxxxxxxxxxxxxx X X X X
S

640
©oo©oo<NfON©voq<N«no\vopvo<NO\vqpf* fn — rj » Tt Tt <j\ in -.
^a
^ n<N-<'-doso6o6o6^^»onwos9lo\o\odo6^oo\o\
#

o6«o»o^ : :

oooooo(X>oooooooooooor^r~r^r^r^r^r-N©vo'<tTi-'<rcococo<N<N--~ 2,

— no © on «n
©
— ©© ©^ ©oo 88 ©©©©0\«r>o\<NoO'— oo^^^nonnooi^-
— ONinooM^iO'-
m — ©ONOor«-NO«o«r>^"coco<N<N~*—
co on «n <n©oo<n©oococo<n cs on
— -
t\
On no co © mco 55,
~*
N N N
(S

m r^ <n»r> 00 so 00
<n <n on «n on tj- r^ rofn«n<NON«rj^ON
^•\o~*co©r"-NO«noNN©^-«nco©©~*<N^ON co©oo©'odr^--~-^©

©©©©©©©©©©©©©©©©
^<MWN\OvOO\'-n(NiOOONO<ri^«nO\
ON^^-r^<Nt^cooo^"©NOro — ONr~-«r»
OOOa\0\^^o\m- ©
CO <N <N 3
r- r- no «/> *r> ^- ^ MmMNNn---'-'-'--
(N — © co r- ©
©©©©©NOt^<N o©v>vooooo«noo(NONr^»n«ooNOOON
,

ONoo©cn'-ONOOooONfS^oo^'©r^^"<NONr->.NO'^'«N^ON
NO ,
<frco-*©oor~N©«n«n^'cococo<N<N<N — — — < — — —«

— en On ^ sq © On ^ lO U-> no VO OO no C> vo -;
•o^'od^fNONb^-tN
#
"©00wS^'cO'-©OnOn ONodr^r^vdinvovo «n in
Tf^-corOcoco<N<N<N

^^fsrtyqco-NONOoo-^ONNO^q^nNqM\q^rj(Njh;ooo\h;^
«no6rNir»'Ho6^os'/l^od«nHdodr>>'in\d^
r»\ONO«Oir>^-^"COCOCO<NfNfN<N — -^ — ^ — ^ — ^-^^
© 06 no nS
vd'tpSw-- on — :

©
^ NO
CO
0»«ri(S9\^>nn02!99t^!i5^^CSC^
co <N <N <N •—
S QQSSS
— -« — 00000<r>'nOOONONONOOOOO v ,

CO CO
»-<
cocococococococococococococo<N<N<N<N<N~- — ^-, .^ — — —

t^"nn^(S^inNOoo
c\|-H©ONOor^NO«nTt^-
— cor-cooo"^"^oo©Np©Npcor^co©cogN
^ ^ ^ CO CO CO CO CO cO cO COCOCOCOCOCOCOCOCOCOCOCOCOCOCOCOCOCO

88 88888 80©©©©©©©
ON-HONr-ON^oo^
-^ir>invovoo\- -no
©ONoor" »No«n^ ^'co, ,

^©^cococo©ONOONOoo^^No«nONoq^ON«r»oqNocoNor-- 00 ^ 00 —
— r-^-^oo«ncM©r^wo^<N — ©ONoor^r^NO«n«r>^'
ir>

"n^tnwwMNrN)(S------ nm^ w nn n^ n^
8^-irjvnfNcococoNoeNONOocoONOor^
u^r-r^Tr — oo<nrsor~«nTf(N — O in
o\ 00
rs
xxxxxxxxxxxxxxxxxxxxxxxxxx XXX

641
 T ^
til
- » h oo in q -

to *£>

ooism — <A^ — OOMM — co

o -9- ro
On r-
r^ «/-> ^ — -* — ©

:|

ootJ-— — ^©oooN©ONr--cov->.r"-©N©c\ir''~
O M ^ ^ (N -« <7noOOOV©«/->«otJ-cOC><n<N —
* «b CM — ~- — ~« —

vq — -^ on Q (N ON
8 ON -* r-» © ~J © 00
9
n M n

^ CO ^
^ S3 ? <N m ^ «o on t-~ tj-

g on r- «o m — o ONooooodr^r^vo^ONONduS^

U co &
^od^(Nio\t^"n^inw^fn^o!ci

«nro^ONr^vo«n^-5ro<Noor^
NONO^o«nm«n»r>«nOOOTj-Tt

oooor~-v©vONO»o«rimvn'Ttv© o«nNO«n«r>w"> v

rs|c^r\icvJr>JCNioir>jfSrsrM<N(N<N<NrNl<N<N

88 ^ <N (N «/"> Tf <N ON — (N OO M

— ©
' On 00 r- NO CO <N <N

< CO O6
U co — — O ^
EhX $<«<
uj On < 00 <N NO ~* On ON «/->. co On

h- uj
vo tj- m -^ o 00 00 r~ no $3 06 c4 06
r)
«->
N <n
is -
r»'

r~ on
NO^-ro
— «n —
— OONOror-ooON«r>ONrooOw->«Noo
xxxxxxxxxxxxxxxxxx

642
 1

ooenooovooesesovOesenvoTj- r- en ON Tt «n o NO —
oo oo VO en <n es <N
«n «n <n «n en en en es en en

vq —« vq oq vq ^ oc vo On OO r- "fr vo o
ON
O en a oo ON
«n 5
en en es ~* — -^ — t--

oo
i^ *o i^
© \o n -
vq *; en Tt ^ oo <n oq oo es o\ oq p-;
;

en *-« — « O r^ *A en *-* — ^
© en en

«nv©t--o\© — r-«-r>-voenv©r~-oo «n en -^ en en O en -* On
oo r^ in so en en en
VO VO VO VO VO VO <n in

On en
oo «n
ONeNeNooONeNONON^mr^
©00-^^ONTfrON^ONNOeN •^ On On OO
t~~ ~* «n VO eS On
O ^
On oo CS es —* -h ^

vo r*
t-; vq _< r-» «n on oo © vo r- es es ~ ~
Tf oo O 00 £
Tf VO -*'
-h On r- VO
»n © ~ ~
<N <N
vo es es r^ eN en On vo
<N — — ^ — 00 <*

«n©es~o>eseseSTtoqesoqen — vq oq vq oq en On *«
•^*
en c> on r-' K K vo vo «n vd vS ^ oo vd «n vo vi Tt vd vd

r- vo -<t <n vq©oqeses^©<r}es ON ^

es~^©'©ododvd»/S ~ on vd © r»" «n © ON*

es es es es es es
SS8

ONeNvoovoen«n^or->vpmQ
N M - - ©©©©-*©©©©
(Nr)MNMN(S(N(S(NN(SM 8

OO^vovovO — ooQOnvoooQ © © © ©
en On
© ©
VO On en
t^ ,— —Hiovooom«nomoo ^-ON
v*>
,
p- oo © eN r»
r-~ «n o
tt en es en es ~* en en
6

«nr^«n^enpr- ONTtvqTi;Ttp -* oo en © © vo
3
:
ONvd^ONeNvd-^«ri^-'vdeNON«n
O\oof»'n«n#'^'eneneses^^ es <N —

8
^* 8
oo
ON es vo es VO
r^ «n
—
t^ <N On «n
en en es es — r- oes
en en es a oo
^ oo
es 3
X X X X X X X X X X X X X X X X X X X X X
8 o
<n
o
en

643
 ro ^r »n co ro — r- no no

^1 a Q
co ^
(N
co co CO
M5 Tt* "^ ON
co co <n co
O no r-'

m x » J -
Tf <n — o o
O oo Tt r~
co — «n -*
CO CO <N <N <N (N CM — —
G*

- » - o\ a O «n «n r-« co r-~ S =
in ri
co co <n
6 —
oo r^
»— <N — — —
-^ On <N
— © oo NO NO

^ <n -
(S o\ ^t
^^
1^ n
<N
^r Nn ^
<N <N <N <N

Q©
^ © ©r^ 8 o o
- r^
OONO r>-
NO Tf — On
. 8
^ co ro co CN <N CS —

So—
(N
r- oo on
oo on co r^
On o\ oo
Tt CO <N Tf
t~- co co On
NO NO m t
r-
00 CO
CO CO
m co r-
O —
<N
oo

« ^™ co On ©
oo F~^ co no co r^ O rt
© «n On «n rs 00 — ©
£xz <<t

o o — ~ — CO co -^ «— NO NO OO 00
co co O O
(NfNCNCN — — •n «n
~*~*

SOO ^ — oo
oo oo r^

</>
o o o o o 00 oo oo oo
O O O O «n »n
NO NO NO NO NO •n «n in «n

88888 8888
— 88^ rs co «n co
© «n
«n —
* 1 ^ M O
< O OO NO co o\ oo

2^1
^
o\ prs On mm
oo oo On CO ON
r- >n CO
00 OS ON oo
PB4 NO *4 r-'
Tf CM •^ ON © oo
H UJ <* ^
K
Z
o NO ON
CO
—
<
On r-' On
f* —
00
r- «n <* CO •—
~*
s
—
r» (N
"* CM — r-'
*— <— w* On
32
— oo
z X X X X X X X X X X X x x
o o o
55
W E
Q
 .'

~H O «n — « co Tf CO — o\ p co
r- «r>.
«S fS (N
»/->. o — wj vo co co

vO On «0 <— «0 <N <N — 0O vO <S ON VO "fr <*1 <N

VO 00
CO co
VO CO VO CO © On"

O 00 SB CM *-
— OO©
vO r~ — On s vo *r\
VO
CO CO
CO On
<N—
vo »/-> ^
#

co CO (N ^ — -m _ © O © O OO O O

—- vo On co r^ r^ co VO -* r^ On
vo - o n vO co OO (N 00 <N r- cm r- cm On ~ 0\ -m
~* <N <N CM ON O r» oo vo r- v->vo co ^t

vo r» on r-
n^ n 0O CM co On On O <r> On -* r»
33 g — — On OO «o ^ co CM
,
o\ on r- Tt
oo r^ vo «a CM CM ~ ~ .
-*

co ^
— — CM
00 -« VO Tf © O oo o
29
co 00 «/-> s
00 "^ «n ~-
vo v->. CM CM
O On vo v> CM <N

^ f; On On — On CM On ^t "<t oo OO k-> CO CO oo
#

r-' *« ©
oo «r>. r- -« vo —, vo ~- «n cm «/"> OO Tf oo ^t

oo oo oo oo oo oo © o -* — < co co so vo
cm cm © © © d vo vo

8 g

o o tn
«r> m
v->
o o
ft f) CI fl cm cm

8 cm co
8
on r^ vo vo

VO 00 CM "t co oo r- r- «r> ^ ,
On oo
—
Tt ©
cm r^
r- vo «r> -^
cm r^
«n co
On <n v> oo
S3 * oo

^- o m
<n r-
^>
<N oo
tn CO
Tf t^
CO <s
o
CO
<N
<N
vO ON
fS —
<s *n
<N
,
^- ^
^M rN oo
r^ <o
X X X X X X X X X X X X X X X X X X
o O 8 s o o
CO 8
rs <s CO
8 w oo CO CO CO CO

645
 m r^ © p — «n f*1 CO On
I

o o\ o
CN — <N
r»' p* r-»* no <o vS «n

o »n o m oo — oo —
«n o\
n© r^ r-' r--'
si
^1 u
"t "0 ^
«
vq o\ <s
n
oo co «n oo
ri in
no in <n m d
m oo
rs
NO It
M (N (N
OO - -

s O Q no
NO ^t —
<N
ON

on Tt (s q
o\ en r» no on en oo
n WTt OO OO On On

— o
i

OO NO oo
t- (Sn -
^ On «n
r- On r- On
m on »n —
II

CO
OO 1^ NO **m o n N M M

0\*/~> OO ON I — OO
f^ On co
no «n «n

<N fS <N — . Tf NO ^
od ri 6 r>" ro o\ no

"1 "1 ""}

NO NO NO
^ ^
fS <N
•"":

<N

NO CO On «n
r-- r- no no

Tt "t Tf
Ci co c*>
m
M
<<t
*r> *r>
(S (S (N
*r>

OO f- en
"^ «n ^-
£ o o
Ont"» no r- r- oo
on r- no «n tt cn m
A< co o
^- -^ «n r* — oo «n m no no
XH $ w>
^t os o * on r>i
r--'
r-» «n «n 5 3- en fN <N

m o
m
r*- co
<n
XXX X X XXX
u
646
 1 ' '

oq on — O «o cn oq
«/-> On ooq On CN "<t OO On
^' Tf »/S CO co co CN CN CN •-5 CN ^ O O

n oo oo CN oq fl fl oo (S n CN «0 m — —
0*0 «/S
«/-}

«r* «/->
iri
co co co CN CN odd

co oq v© vo On oo M t n
On vo IO VO CO CN CN ~ ©

.82 rs
5
«n r»
n ^ OO v©
NO£ © «T> m ^ On
cn «n r-
CN
On ? CN
co
co no r-
© oo r- oo NO w-> ^
•^ Tf Tf CO CN CN 2 8 o
— o o
> O O O o o o o o o O O oO o o o

on vo vo no oq oq «n v> co r-«- Tf CO V©
^ vo oo —' «o od
$3 r» On
co co
r- oo on
CN CN CN

f- On no NO NO *" NO co r-» no
cn vo co
MOOh 25 co -* O r- oo CN On r-
— —
es ^ —
On oo r- co co r^

NO <S (S NO
COO OO - (S n NO O ON \D
— O OO r-* <6 v" — — i o o o

^ CN ON NO O CO — o —
. CO oo cn r~ O vo CO
»-« t> «n O oo m — oo «r> oo ^» 00 TT On NO Tt

«n «r> «r> On On On co co co r- r- t-» — ~ o\ o\ o\

o o o On On On On On On NO NO NO

3 9 §

On On On
CN CN CN
C4 M
<N
s s s
CN CN CN

o o o
oo — m
O O O 8S O CO CO
CN CO NO
r- co oo ^ On «0 r- cn CO o ~ On r-
CO CN CN CN CN -^

OO CO OO On rt O ON CN V> cn «n -* co On oo o oo co O
ON (N ON
cn cn ~*
r-
CN CN
© —
r^ — 00
CN ~-
^
— On ir> CN CO On O OO oo r- vo

o
CO
CN
CN CN
o oo wm r»
CN CN
CN oo v>
CN >— f*
On NO CN
— 1-*
CO
— o _4 oo
•-^
ON r- NO

X X X X X X X X X X X X X X X X X X X
o
CO 8 o
oo
o 8 8
CN
V
CN
u u u u u u
647
 o — c> v> i*- on n ^ ^o oo Tf «n r- oo «r» «r> r^ oo on
no no f- On
On On On On Tt tt ^r tr On On On On On
On On
M Cl
On On On On
fl fl f> f) <N <N <N (N fS (S (N (S — « ^ _ _« r«: t«^ r^: ^

oo
— On CN oo oo
© p rNi
CN CN in
>< o ^ 60.9 59.2 57.4
cn
3 CO
Tf
Cn|
<<t Tf
r^ no
ro co
On oo oo
CO CN CN CN Cn! r-*' NO
<N CN

vO q O Tf On oo O ^ 00 en
60.3 60.9 61.6
<N rsi en 00 00 oo On' d o d CN NO 00

II NO NO NO to co COm co co m r-> cn r-'

CN CN
r-'
CnI
oo
CN

— CN no
«/"> co On Cn) no r-
* h
II
p r- <n «n no no
On ^ O NO m O 2g£ - - Q M
NO V* ^ CO
X -
co co
ifl
cn
O «r
— «n
(S VN
d — —
no cn
Co CN CN — — —Cn|
cn|
rsj <N

oo r- r- w-n
co «n r-~
no no co © -
^ Tt M NO
«n no On
in t O * ^
no cni oo
oo
— O On r- <N
n© cn © oo
0O CO On
<N (N
m
- - — On OO no «/-> ^- co cn CN cn — — — — — — d

8o» p ^p oM p ss—
— m OnO — O On
o § co
r- On
p O Tf On
Tf CO
On ^t On

< s — On r^ NO "^ n <r>

«/">

-<fr co cn CO CN — MM- i/%

— - — r-~ co o

-J
A< co -^
— no r- <n »n no p r-^cn oo rf cn oo cn _ co oo o
©
F- CO On 0O On
OO r» vo -f CO CO
O 3
^ «0 0> (N
fO CNl fN|
On Tf OO wS
CN CN — — K n O od
cn

CO ©
co
»r>
cn cn
© no
— co
— «
©
— o
CN
NO co © NO
— o
m oo — m
no
— O
— OO NO
i X X X X X X X X X X X X X X X X X X X oo NO
WQ o X X X X
* p 8 CN 8
y S CN
X X X X
E 8 O
«n
W-N
CN 8

648
 i ;'

v© NO r^ OO ON r- r-» oo on r* r^ oo oo os o
NO On <N f- CO
t^ t^ OO 00 OS
~ CO NO
oo oo oo o<
— r- r- on
oo oo oo
Tf Tt T* Tt Tt (N <N rs| <N o o O O o — OO OO OO 00 oo NO NO NO NO Tf Tf Tt

no — m
<n ^ hO
— ~
«
O
On <N «n — Tt P- o
NO <N .ON <N "* « f- «fr
NO
ON NO <N OO
o o O
o r* en
CO <N
<N fN <N <N <N
On On OO OO r» no NO <n »n
^ 2 m en cj fN On oo r- r-

^» OM^ NO» « (S Tt NO ON O
«-« en -<t r- oo On _ «n NO t-» oo
On Oen
en «n no
<n <n en en en
CN CN <N fN <N
O O
ON ON
— — <N (N
NO NO NO f- r- r- en m m en T* o o o o r^ r^ r^

— r- ^t >n ro on ^
Tt ON (N
^ ON W-N
(S ON NO
« — w p -
odd^
•

00 in oo oo on «n «n t- Tt <N On r^ «n
r^ n - oo t^ O oo no vn r^ «n ^
en en r-i n ri ri ri — — — d
'

m m <s o\ oo
On <N O no On
ON mom
o
oo
<n
oo «n r- fN «n no
S m 2 s
NO rs
-ggg—
oo Tt r- <N
«n
CS no O Tt
r* en en rs
<No 00
oo r- no -^ en «•n n m
l/">

CM fS o o o o o o o 8
o o o o d o O O o o o o o o o o o o o o o o o o o o o

no in
_ r- tt rs
^ NO Tf Irt
— -H
<N <N (N <N
^ no
>n
^
«n
o ^ —
cs "q- no
^ «n «j
oo <N no
—
O
fS On r- no oo no in ^ n no «n ^t en cs en en <N <N

fS OO ON CN NO Tf IN «n —« O —
rs en <N m
— no ^ O o
«n — «n r- oo
o o on r^ no ^fr no OO On oo -^ on en «n On en r- O >n o«n 5 =
^ - oo no in on r» «n ^ ,
•

r- no Tt -^ en rs «n en en cn es MM (S -

m o OO NO «n
o OO NO «n
o oo vo NT) ^ en oo NO m no «n -^ en «n
X X X X X X X X X X X X X X X X X X X X X X X X X X X
m «n
NO
«n
»n
«n «n
CO
«n

X X X X X X
>n «n «n «n «n «n
NO m en

649
 NO cm
7,
CM
sO
1 CM
co
I SI
O d o © d d d d d d d d
J
O .

M «/i co ^ no oo no r- ON o
N r-i CN
m CO cm rs cm CM cm CM sri

^M tf\

+
ir>
od
— O
in fn co
co oo r- »n
cm
CM r*4 cm 3
ON o
oo
ro ro
CM cm CM CM
O

cs) oo CO r* CM sO O
Tt *"}r- P- CO

^1 9 rn
33 no no r-'
CM rM cm CM
od oc ON ON
CM cm cm CM

OO NO oo oo r*» Tt ^ r- NO NO
ii
B ON >r%
ON
d 5 ro CM ? H
rn
NO
cm cm
—
co 'd

t-» CM CM oo ro co r-
^_ r- m oo cm r- 8 O NO
in rsi d 00 CO CO CM CM CO cm* CM —

ro m ,

o CO oq f; «n ON ON 00 O
NO
NO i ro CM
NO NO
CM »« d d ON od od
»n rr

fl
ro © m o
CM fO ro
NO NO NO 3 co
NO n
NO CM no rm

NO
Tt OO rs «rj
r-" od od

OO 00 rs
NO ro OO o O
oo
r-— On oo r-
d uo 3
ii

CO fc
rn ON
cm 3 2

Q co oo wo
NO CM NO OS
— no r-'^ —
no r^ no co
cm od «n cm
^ p
od r^
*f) «n
«n ^
CO CM CM

^ NO CO CO
§ § Tf OO
r- o Q
^Tf
o\
OO NO «/-> Tf «n Tf co cm CO CO CM —

OO OO O ^t O r- co oo ^t cm oq cm
On tT od no
5 M <N -> -

— ~ O
NO co
— oo

wQ XXX XXX X X X X
HZ
CO <

650
 1 > 1 " 1

ON «n
ONO in <n
«n
NO en r- CN
m NO NO
CO co CO
NO
Ov 88
8
00 00 OO «nm 3
1
Os
m «n
•n
ON en I- Ov
00 00 OO
NO NO NO NO VO
8 cn 00
8
«n in in in
8
3
© © d O d d O d d d O d d d d d d O d O d d d d d C
C
cn co >n NO CN en "T vo ^ IT)NO t-» ^
cn CO NO r- 00 ON Os O rs CN
VO NO NO VO vo NO NO vO «n in in in NO co
ON ON ON ON 06 OO OO 00
^ rr
2

r-- r^ NO 00 ON 0O co ^ O CO WJ Ov OO co O 00 r» O co Os co NO ON m
Tf CO CN
CN CN cn cn
06 r-' K NO in en
CN CN CN CN 3 Os Os 00
CN 8 g ON 06 r-' NO in in

N© p
Tf NO
«n VO NO
r- p en NO OO CN «n 00
vd
«n ON CN
CN
Tfr OS rs <n 00 ON
CN CN
r^ ON CN "*
06 OO ON Os
CN CN CN 8 8 cn fS CN CN cn cn CN CN CN cn a CN CN H a

CO ON CN cn 00
o r* CN m r- CN Os m00 OO en r- OO NO OO <n NO NO
CN VO
co cn CN
K 06 TT cn* Ov «n
CN ti
NO* CN 06
2 zz
ON K Tf
^ 06 r-' d 06 NO in

9 O CO NO
m NO NO m O CN O —
S£^ 00 3 ON
s 00 VO in « <o
en
«n
00
r--
ON
ON
NO «n I NO CO
NO
ON CN
in
CN co 00
CO CN
cn ^ ^* d d d d d d d d d d d O d d d d d

cn cn _ CO Os 00 CO_ l en «n "t CO «n 00 CO CN in 00 'Nf 00 _- -vf O

§ ON
co
CO cn CN
<* 5 CO en
d ON 06
CN
. fO CN
CO CO CO CO
d ON OO
CN
NO VO
CN CN CN
Os ON 06 06
CN CN CN CN

«t 00 ^ NO O CO NO r^ p^ Tf r- ^^ in 00 ^ NO OCO NO r-; co in OO O
OS ON ON
CO CO en § Os
co if? d
CO en en CO
#

CO CO en co
06 00 OO 00
CN CN CNl CN CN
00 00 00 ON
CN CN CN CN

m no CN ^ ^^ rm _ © Tf Os en «n NO CS t-; l
00 r- en "t
l

CN r-; CN~
00* 00 OO
CO CO
r^ r-" CO
Tt CO CO CNl CO CN
«nd O
CO 3
ON »n
CN
Os NO CN Os in CN O

r-m no CN «n CNO r- Tt co ~ OO CO Os
no «n 'vf cn
O CN NO 00 «n on — r- ^ — P NO CO p O r»
•n cn — on r- no
^r CO CN CO CO CN ~* CN — — — ~ ~ ~ d d ~ d d d

0000 O 00 NO o
0000
~ ~ O O O «fr Q
o o o tt o
00 co »n no
^ NO NOO ^ O
CO
os
Tt NO
«n — n O ^ - — *r>it
O
NO CO
00 «n no
on »n cn
«n
on 00
«n 00 on «n
Tt— 00 r-
CO CN CN ^ CN — — — ^ ^ ^

NO VO Os 00
q so ^ o ~ ~ Tt sq ; «n o
«n on •n cn 00 Tf cn ^(S o
•n
CN
d vd co
CN — ^
ON Tf CN Os OO Tt -^ OO vd co r»*d «n" cn on r-* no « o! K

NO CO
^^ "" 00
co
»« •— 00 NO
CO
00 VO
co •—
00 NO
CO >- 00 NO «n O 00 NO m
X X X X X X X X X X X X X X X X X X X X X X X X X
g •n
s «n «n •n
NO
X X X X X X
in «n
CN CN 8 8 g S

651
 n
t ^
T «n
w>
x fN o m
OO OO
r>
OO <£ o — no
<n
m
on ^n
m m ^
«r»
5 5 5 F-
m ^ <§
^ Tf TT Tf Tf •*
dodo
»r> »n «/-> «r>
d d d dodo dodo dodo
\r\ «/i «/-> CO

5 r- r- —
^
OO On O O O
OO On n© r^ oo oc Tf Tt «o
*n «n co co o o O o o o O r- r- r» r-> NO NO NO »0

«n oo —
r-~ O <N On oo — r- "T
so «r> vn ^- CO <N
J2
m co <n rj on oo oo f*- oo r- r- r>-

vn oo On — -^ «n 1 r- On © r» oo <n O
W^ NO 0O M
r»" r-" r-" r ^ ^ w-> o\ o\ <d <5 oo oo oo oo

m\%t a O r-
—
<N co no On
On co no r-
no
© t-~»n r-» no
On OO <N NO
«/->
OO
OO On On
«r> <N On
<n
— m
cr)
<*n
On r-
I §
oo ^
in ^ «/"> co co ^t n nri — — — d — — do
co o

oo
» ^ -
rj- r-> © <N <N ©
OO Tt <N O
oo — tj- r- On m — no
—
n mm
(N

d d d d
N ifl

O O O d d d d dodo
<N CnI
o o o o
<5 <5 d <5
CnJ

«n oo —
r«; CO N© <N ON © r- CO On f-; "^ p NO
NO U0 «/S Tt"
<N <N CM <N
— O O On
(N <N <N -^
On OO 00 t** »n iri iri
^'

0\ (N m VO r- o — (S "^ t^ OO wo t-» oo O <N m W-n

rn Tt rr
CN <N (N
© © © ©
<N CM (N (N
r-^ r»' r-^ ^r "^ ^ ^'

!
i
E
MM - -
II

I «o o v~> <N On OO CO VO WO Tt ^ co <N <N

00
O O CM CO voO On — vo «/o CM CMO (NO ^ OO
r- CM cm
^ co vo r^ co On co
«/0
§ ^
oo r- w-n
o o o o
i
1
oo no
o o o o O O O
ro <N <N
o o o o O O O o dodo
I
<
sss «r> NO "<T Q NO T O *
— ° 3: *>
—
"5
3
r-
NO
CO
On r- no
WO V~> v->
00 NO in H" I 3
wo co CM Tt
«n
c->
oo
rs «N

DI
O
A< c/5 <p On — O WO © — (S n <n O wo m nO
% cm cm co co CO * N i*-. ~ ro g r- o
ifl
rj h; (S F;
IH >5 60 d oo no vS n© wo ^ ro <S r>i --
I
i
— oo no «r>
OO NO Tl 00 NO NT) ^t NO m rr CO no m Tt n
x x X X XXX X X X X X X X X X X X X
o
o 8 CO
X X X X
»n
«r>
r^ NO
<r>
«n 3

652
Index
Allowable stress, 37 limit analysis of, 585
Analysis of flexure action, 184 stress due to, 5, 50, 62
of internal forces, 2
Angle sections, properties of, Beam design for flexure and shear,
equal 648
legs, 200
unequal legs, 650 critical length, 201
Angular deformation, 77 Beam stiffness for moment distribu-
Area-moment method, 227 tion, 348
cantilever beams, 241 Beams, of two materials. See Rein-
change in slope by, 229 forced beams
moment diagrams by parts, spacing of rivets in built-up,
231 206
propped and restrained beams, types of support for, 107
295, 296 Bearing stress, 20
230
rules of sign for, in riveted connections, 478,
simply supported beams, 248, 486
260 Bending, combined with axial
tangential deviation, 229 loads, 359
Axial loads, combined with flexure, combined with torsion, 392-397
359 Bending moment, 1 10
combined with torsion, 397 and shearing force, relation
deformation due to, 38 between, 125

655
656 Index

Bending moment (Continued) structural, 21, 487

change in. 127 welded, 497, 502
diagrams. 15, 128, by parts,
1 Constants of integration, evaluation
231 of, 218, 220
equations, 1 12 Continuous beams, 312. See also
sign of, 1 1 Three-moment equation
Biaxial stress-strain relations, by moment distribution, 346
47 deflection in, 340
Bond stress,436 reactions for, 328
Built-up beams, 184, 206, 417, shear diagrams for, 331
424 simply supported, 322
Bursting force in cylinder, 25 with fixed ends, 333
Butt joint, 475, 479 Correction factors for curved
beams, 557
Carry-over moment, 347 table of, 558
Castigliano's theorem, 520 Couplings, flanged bolt, 87
applied, 523 Critical length in beams, 201
Center of twist. See Shear center Critical load, in columns, 441
Channel sections, properties of, stress, 448
646 Curvature, radius of, in beam,
Circumferential stress, 26 157
Collapse mechanism, 587 related to bending moment,
Columns, 439 116
critical load, 441 Curved beams, 557
critical stress, 448 correction factors for, 558
eccentrically loaded, 464, 469 shift of neutral axis in, 556
empirical formulas for aluminum, Cylinders, thick-walled, 564
456 thin-walled, 24
steel, 453-^55
wood, 457
Euler's formula, 442 Dead load, 171
limitations of, 447 Deflection, due to impact, 521,
factor of safety for, 448, 455 532
interaction formulas for, 465 formulas, summary of, 270,

slenderness ratio, 448 307

tangent-modulus method, 453 of helical spring, 98
Combined axial and flexure loads, lateral, inbeams, 168
359 methods beams, 213
for
flexure and 392-397torsion, area-moment, 227
Combined 359
stress, conjugate beam, 265
equations for, 377 double-integration, 214
Mohr's circle for, 379 elastic energy, 519
notation for, 375 superposition, 269, 309
Components, of strain, 402 three-moment equation, 340
of stress, 377 midspan, by transformation
Compression flange of beam, to symmetry, 260
buckling of, 168 of reinforced beams, 424
Concrete, reinforced beams of, Deformation, angular, 77. See also
424. 430, 433 Strain
reinforced struts of, 50 axial, 38
Conjugate beam method, 265 shearing, 39
artifical constraints for, 268 Deviation, tangential, 230
Connections, riveted, 474, 487, Direct shear, 16
490 Distribution factor, 349
Index 657

Double-integration method, 214 Flanged bolt couplings, 87

applied to restrained beams, Flexural action, analysis of, 184
284, 287, 288 Flexural center. See Shear center
general moment equation for, Flexural rigidity, 215
217 Flexure, limit analysis of, 586
Double shear, 16 Flexure formula, 158
Dynamic loading. See Impact assumptions and limitations,
loading 153
Floor framing, 172
Eccentricity, maximum, to avoid Force, notation for internal, 3
tensile stress, 369, 372
Eccentricity ratio, 455 General moment equation, 217
Eccentric loading, of columns, Girth joint, 27
464, 469 Girth stress, 26
of riveted connections, 490
of short strut, 369 Helical spring, 95
general case, 371 deflection of, 98
location of neutral axis, 372 Wahl formula for, 97
of welded connections, 502 Hooke's law, 38
Economic section for beams, 167 for biaxial stress, 47
Efficiency of riveted joint, 476 for triaxial stress, 47
Elastic curve, differential equation Hoop stress, 26
of, 215 Horizontal shearing stress, 188
shape of, 116 applied to rectangular section,
Elastic limit, 35 191, 194
Elastic-plastic boundary, in flexure, applied to wide flange section,
574 196, approximate formula
in torsion, 572 for, 197
Elastic-plastic material, 571 assumptions and limitations
stress-strain diagram for, 571 of formula for, 192
Elastic weights, method of, 266
Endurance limit, 512 I-beam sections, properties of,
Energy methods for deflection, 644
519 Impact loading, 528
Equation of three moments. See deflection due to, 529, 532
Three-moment equation impact factor, 529, 532
Equivalent sections in reinforced limitations of, 532
beams, 417, 425, 434 maximum stress due to, 530
Euler's column formulas, 442 Indeterminate, axially loaded
effect of end conditions, 447 bars, 50
equivalent length, 447 beams. See Continuous beams;
limitations of, 447 Propped beams; and Restrained
Exploratory section, 2 beams
in beams, 109, 124 shafts, 80
Induced shear, 16
Factor of safety, 37 Inelastic action, 570
in columns, 448, 455 in flexure, 573
Factors for three-moment equation, in torsion, 571
317 limit analysis, 585
special loadings, 317 residual stress, 577
table of, 318 springback, 580
Failure, theories of, 516 Inflection, point of, 116
Fatigue, 510 Interaction formulas for columns,
testing, 511 465
658 Index

Internal forces, analysis of, 2 Moment diagrams, 1 15. See also

Interpretation of vertical shear Bending moment
and bending moment, 124 by parts, 231
construction of, 128
Joints. See Riveted joints; Welded table of cantilever loadings,
connections 233
Moment distribution method,
Kern of a 369
section, 346
circular section, 373 beam stiffness, 348
rectangular section, 372 carry-over moment, 347
wide flange section, 373 distribution factor, 349
qualitative description of, 348
Lap joint, 475 shortcut procedures, 353, 355
Lateral deflection of beams, 168 sign convention, 349
Limit analysis, 585 summarized, 350
collapse mechanism, 588 Moment of inertia for areas, 597
in axial loading, 586 by integration, 602
in flexure, 587 by tabular computation, 61
in torsion, 586 for composite areas, 607
plastic hinge, 588 for geometric shapes, table
Limit load, defined, 570 of, 608
for axial loading, 582 maximum and minimum, 630
for flexure, 573 Mohr's circle for, 627
for torsion, 571 polar, 599
Limit moment, 573 principal axes of, 630
Limit torque, 571 radius of gyration, 600
Live load, 169 transfer formula for, 601
Longitudinal joint, 27 units and signs, 598
stresses in, 485 with respect to inclined axes,
Longitudinal stress in cylinder, 26 625
Moving loads, 147
Materials, average physical properties
of, 635
Members of two materials. See National Lumber Manufacturers
Reinforced beams; Reinforced Association column formula,
struts 457
Modulus, of elasticity, 38, related Neutral axis, location of, 156
to modulus of rigidity, 413 in curved beam, 555
of resilience, 531 in eccentrically loaded short
of rigidity, 40 strut, 372
of rupture, in flexure, 160, in in plastic action, 574
torsion, 78 in reinforced beams, 419, 425,
of toughness, 531 430, 434
Mohr's circle for moments of in unsymmetrical bending,
inertia, 627 548
Mohr's circle for strain, 404 Neutral surface, 154
transformation into circle for Nonhomogeneous beams, 418
stress. 404 Normal stress, 16, 375
Mohr's circle for stress, 379 combined with flexure stress,
application 389
of, 360
in design, 392 combined with shearing stress,
rnbined torsion and 377, 380
flexure, 393 in flexure, 155
rules for, 380 maximum and minimum, 378
Index 659

Parabolic column formula, 454 over-reinforced, 420, 430

Penstock, spacing of hoops for, under-reinforced, 428, 430
30 Relations between load, shear,
Permanent set, 35,577 and moment, 125
Pitch, in built-up beams, 206 Repeated loading, 510
in riveted joints, 476 Repeating section of riveted joint,
Plastic analysis. See Inelastic 475
action Residual stresses, 577
Plastic hinge, 588 in flexure, 578
Poisson's ratio, 47 in torsion, 578
Polar moment of inertia, 79, 599 springback, 580
Power-torque relation, 78 Resilience, modulus of, 531
Pressure joint, 476 Restrained beams, 283
strength of, 478 by area-moment, 293
Principal axes of inertia, 630 by double-integration, 284,
Principal planes of bending, 546 287, 288
Principal strains, 405 by superposition, 284, 286, 289,
Principal stresses, 378 309
by Mohr's circle,383 by three-moment equation,
Principle of rigidities, 55 333
Product of inertia, 617 deflection in, 297
transfer formula for, 619 design of, 306
units and signs, 617 simple beam equivalents, 303
Properties, of angle sections, 648, table of end moments and deflec-
650 tions, 307
of channel sections, 646 Rigidity, flexural, 215
of I beams, 644 modulus of, 40
of S shapes, 644 Riveted joints, assumptions for,
of W shapes, 636 482
of wide flange beams, 636 definitions, 474
Proportional limit, 35 eccentrically loaded, 490
Propped beams, 283 elemental failure of, 476
by area-moment, 293 for pressure vessels, 479
by double-integration, 287 for structures, 487
by superposition, 286 stresses in, 485
by three-moment equation, Rivet spacing in built-up beams,
334 206
Rosette, strain, 409
Radius of gyration, 600 Rotating ring, stress in, 28
Rankine-Gordon column formula, Rupture, modulus of, 78, 160
454 Rupture strength, 36
Reinforced beams, basic principle,
416 S shapes, properties of, 644
bond stress, 436 Secant column formula, 455, 469
of timber, 417, 419 design curves for, 472
deflection of, 424 Second moment of area. See Moment
design of, 430 of inertia for areas
equivalent sections, 417, 425, Section modulus, 158
434 for geometric sections, 159
shearing stress in, 423, 436 Shear center, 537
of concrete, 425 for angle, 540
Reinforced strut, 50 for channel, 539, 541
Reinforcement, balanced-stress, for T section, 540
430, 431 for unequal flanged H section, 540
660 Index

Shear center (Continued) Strain energy, of axial loading,

for wide flange section, 539 519
for Z section, 541 of flexure, 520
Shear deformation, 39 of torsion, 519
Shear diagrams, 115, 128 Strain gage, 409
construction of, 128, 129, 132 Strain-hardening, 571
for continuous beams, 328 Strain rosette, 409
Shear flow, in beams, 189, 207, equiangular or 60°, 41
534 rectangular or 45°, 410
in thin-walled tubes, 91 Strength, of butt joint, 478, possible
Shearing force in beams, 1 12 methods of failure of, 480
diagrams, 115, 128, 328 of lap joint, 476
related to bending moment, theories of, 516
125 ultimate, 36
Shear modulus of elasticity, 40 Stress, 6
Shear strain, 402, 404 at a point, 375
Shear stress, 16 biaxial, 47
in beams. See Horizontal shearing circumferential, 26
stress combined. See Combined stress
in reinforced beams, 423, 436 concentration, 8, 512, table
on perpendicular planes, 91, of, 515
191 conditions for uniform, 7
Shrink fit, pressure due to, 68, flexure, 158
568 horizontal shear, 188
Simple stress, 5 impact due to, 530
Single shear, 16 longitudinal shear, 90
Slenderness 448
ratio, Mohr's circle for, 379
Slip in built-up beams, 184 normal, 16, 156, 375
Slope in beams, by area-moment, notation, 375
229, 242, 250 principal, 378
by double-integration, 215 shear. See Shear stress
Spacing of rivets or bolts in built-up tangential, 16
beams, 206 thermal, 62
Spring, helical, 95 torsional shearing, 78
deflection of, 98 triaxial, 47
Springback, 580 variation of, 373; analytical,
Statically indeterminate beams. 376
See Continuous beams; Propped working, 37
beams; Restrained beams Stress-strain diagram, 34
Statically indeterminate members, for elastic-plastic material,
axially loaded, 50 571
in torsion, 80 Stress trajectories, 390
Straight-line column formula, for beams, 391
453 for torsion, 390
Strain, 34 Structural riveted connections,
biaxial, 47 22, 487, 490
components of, 402 Summary, column theory, 373
conditions for constant, 35 combined stresses, 414
Mohr's circle for, 404 continuous beams, 357
normal, 34, 402 deflection methods, 281
principal, 405 inelastic action, 595
shear, 39, 402, 404 reinforced beams, 437
tnaxial, 47 restrained beams, 3 1
Index 661

riveted and welded connections, Torsion, 74

507 combined with axial load, 397
shear and moment, 152 combined with flexure, 392-397
simple strain, 73 formula, 78
simple stress, 31 assumptions for, 75
stresses in beams, 211 for rectangular section, 78
torsion, 104 limit analysis in, 586
Superposition, deflection by, 269, of thin- walled tubes, 91
table for, 270 shear flow in, 92
in restrained beams, 284, 286, Toughness, modulus of, 531
289, 309 Trajectories, stress, 390, 391
table for, 307 Triaxial stress, 47
Twist, angle of, 77
Table center of. See Shear center
cantilever loadings, 233, 285
correction factors for curved Ultimate strength, 36
beams, 558 Uniform stress, 7
deflection formulas for beams, exceptions to, 8
270 Unsymmetrical bending, 546
factors for three-moment equation, defined, 546
318 inclination of neutral axis for,
moments of inertia, 608 548
restrained beam loadings, 307 due to, 548
stresses
section moduli, 159 Unsymmetrical sections subjected
stress concentration factors, to flexure, 176
515
Tangential deviation, 230 Vertical shear, 109
Tangential stress, 16 change in, 126
in thick-walled cylinder, 564 definition of, 109
in thin-walled cylinder, 26 diagrams, 115, 128, 331
Tee beams, reinforced concrete, equations, 112
433 related to bending moment,
Tension test of steel, 34 125
Theories of failure, 516 sign of, 110
maximum shear theory, 518 Virtual work method, 522
maximum strain theory,518 applied, 524
maximum stress theory, 517
Mises yield theory, 518 W shapes, properties of, 636
Thermal stress, 62 Wahl formula for helical springs,
Thick-walled cylinders, 564 97
external pressure only, 568 Welded connections, 497
general case, stresses for, 566 eccentrically loaded, 502
internal pressure only, 567 Welds, types of, 498
Thin-walled cylinders, 24 maximum size of, 499
stress in, 25 strength of, 497
Three-moment equation, 313 Wide flange sections, properties
applied to restrained beams, of, 636
333 Working stress, 37
applied to simply supported
beams, 322 Yield point, 35
deflections by, 340 Yield strength, 36
factors for, 317 Young's modulus, 38. See also
rules of sign, 316 Modulus of elasticity
81 82 83 9 8 7 6 5 4 3
SI Units
(Systeme International d'Unites)

A. SELECTED SI UNITS

Quantity Name SI Symbol

Energy joule J(l J = 1 Nm)
Force newton N(l N = 1 kg-m/s 2)
Length meter* m
Mass kilogram* kg
Moment (torque) newton meter Nm
Plane angle radian rad
o
degree
Rotational frequency revolution r/s
per second
Stress (pressure) pascal Pa(l Pa= 1 N/m 2
)

Temperature degree Celsius °C

Time second* s

Power watt W(l W= 1 J/s)

B. COMMONLY USED SI PREFIXES

Multiplying Factor Prefix SI Symbol
9
10 giga G
10
6
mega M
3
10 kilo k
10- 3 milli m
10" 6 micro M
10" 9 nano n

•SI base unit

 .

C. SELECTED RULES AND SUGGESTIONS FOR SI USAGE

1 Care must be taken to use the correct case for symbols, units,
and prefixes (e.g., m for meter or milli, M for mega).
2. For numbers having five or more digits, the digits should be
placed in groups of three separated by a space, counting both
to the left and to the right of the decimal point (e.g.,
61 354.982 03). The space is not required for four-digit
numbers. Spaces are used instead of commas to avoid con-
fusion —many countries use the comma as the decimal
marker.
3. In compound units formed by multiplication, use the product
dot (e.g., N-m).
4. Division may be indicated by a solidus (m/s), or a negative
_1
exponent with a product dot (ms ).
5. Avoid the use of prefixes in the denominator (e.g., km/s is

preferred over m/ms). The exception to this rule is the prefix

k in the base unit kg (kilogram).
ISBN 0-06-046229-9