Volumen 1.1

r.
McGRAW-HILL
i^^S?liiftk^\i
ENGINEERING
SCIENCi!^^''^^'^^
Digitized by the Internet Archive
in 2010
http://www.archive.org/details/principlesofmech01yehh
civD aaiUBB^
V(^
t^^4
Principles of Mechanics of Solids and Fluids
VOLUME 1. PARTICLE AND RIGID-BODY MECHANICS

McGRAW-HILL SERIES IN ENGINEERING SCIENCES
Stephen H. Craxdall and Paul yi. Xaghdi, Consulting Editors
Craxdall and Dahl •

An Introduction to the INIechanics of Solids
Frankel •
Principles of the Properties of Materials
Hodge •
Plastic Analysis of Structures
Karplus and Soroka •

Analog Methods
Yeh and Abrams •
Principles of IMechanics of Solids and Fluids
PRINCIPLES OF
Mechanics of Solids and Fluids
\ O L U M E 1
PARTICLE AND RIGID-BODY MECHANICS
•IMFACafiaa^lBRARYf
Hsuan Yeh
PROFESSOR OF MECHANIC AL KXGINKKRIXG
DIRECTOR OF TOWNE SCHOOL OF CIVIL
AND MECHANICAL ENGINEERING
UNIVERSITY OF PENNSYLVANIA
Joel I. Abrams
ASSISTANT PROFESSOR OF CIVIL KNGINEKRING
YALE UNIVERSITY
McGPvAW-IIITJ. ROOK ("O.MI'AXV, INC.
1960 New York Toronto London

3 Iff MEZDE
Guar
^Wi :bi=~ tk :
Preface
This book concerned with the basic principles of the mechanics of

is
Stands and It is an outgro\N"th of a set of mimeographed not<is

fluids.
prepared, since 1952, by the senior author for use in a junior course at
The Johns Hopkins University. The central theme of the book, as
originally conceived, may be stîted as the integration of the mathe-
matical subjects which are otherwise abstract for many students \\ith the
realities of a mechanics course and the demonstration of the \miversality
(^f the mathematical expressions which portniy the physical phenomena
encountered in the several phases of mechanics. It is the authors' hope
that the completed l-K>ok fulfills these aims.
We have made an attempt to present the fundamentals of mechanics
in a thoroughly mature manner to persons \Aith no (or little) previous
training in the subject, other than a first course in general physics and
the completion of two yeîrs of college-level mathematics. In doing this
we are reaffirming our strong belief that the professional engineer of
tomorrow should receive an early training, in w hich the strongest empha-
sis should l>e placed on mathematics, fimdamental science, and other
b;vsic disciplinessuch as mechanics. More specifically, it is the purpose

of this text to present the mechanics of rigid and doformable bodies in a
wider scojx* than usual so that many diverse types of applications cnn bo
adequately described.
This bo<ik is divided into two vohmies. Volume 1 is concerneii essen-
tially with the mechanics of particles and rigid bodies; Volume 2 deals
mainly xKith the mechanics of deformable bodies, the continumn. As a
result of this subdixision the first volume deals mainly \\ith vector
meth<xis; the second vohnne extends also into tensor methods; integra-
ti<^n of the two isaccomplished through both mathematical and physical
reasoning. Since it is felt that the student may not possess sufliicient
background in either of thest^ mathematical discipliiies. some effort has

V
vi Preface
been expended to fill that gap. Other mathematical techni(nies are

developed as required. Future engineers will undoubtedly consider the
study of mechanics without some vector or tensor analysis a luxury few
can afford in this scientifically fast-growing age.
In an effort to present a unified approach to solid and fluid mechanics
a choice had to be made, within the confines of reasonable length, to
limit the treatment on specific applicationsand to emphasize solely the
basic principles: the analytical aspects and physical interpretation of
mechanics starting, as a point of departure, with the classic form of
Newton's laws. For example, in statics many interesting topics such as
special methods for the analysis of trussed structures are omitted.
is present, however, to allow the individual instructor
Sufficient flexibility
to supplement the book by the inclusion of this and other topics at the
appropriate places. It is not our intention to produce an expert tech-
nician, but rather to enable the student to pursue expeditiously further
studies in order to achieve professional competence.
Again, following the general aim of integration and unification, and
also for pedagogical reasons, we have taken certain liberties. Whenever
the opportunity presents itself we introduce some ideas, secondary to the
immediate discussion, but which are anticipatory in nature, and thus
gently ease the student into the more advanced concepts. While this
procedure may be uneconomical of eff"ort because of some unnecessary
duplication, it is considered to be extremely worthwhile didactically.
Also, we utilize every possible opportunity to interject different, but
related, material aside from the central core of the customary "mechanics
"
tcjpics, again demonstrating the universality of the mathematical

approach. Occasionally, for pedagogic reasons, we have deferred (or
entirely omitted) proofs of certain statements until a later time in order
to achieve desired continuity.
After teaching engineering undergraduates for some years, the authors
became aware that the best means to incite or motivate a student's
interest to a theoretical or mathematical analysis is to show its practical
application. The viewpoint that intellectual curiosity alone should be a
sufficient motive can only be developed slowly. For this reason the
majority —
although by necessity not all ôf the examples and problems
in lliis book are int(Mul(>(l to illustrate the theory by means of practical
:ipplic;ilioii, ('\(mi though the main discussion is on a more fundamental
level.
Moi-(>over, the o\aini)les are dcsigncMl

and ulilized to supi)lenuMit, and
ainpliiy main stream of the t(>\tuaJ matter. As such, the examples
I lie
and problems form an integral jârt of the book, and careful consideration
should be extended to tluMU. At this point the instructor has ample
opportunity to discuss soni(> of the rainilicat ions and practical aj^plica-
Preface vii
tioiis of Problems involving substitu-

the theory not otherwise covered.
tion ofnumerical values into previously developed relationships are
studiously avoided. Except in a f6w isolated cases, it seems worthwhile
to gain some idea of physical magnitude. In all there are GO worked-out
examples and some 270 problems, of which about three-quarters are
supplied with answers.
As a textbook, the first volume has been found to be suitable for a
one-year course in undergraduate mechanics, and the second volume for
either one or two semesters at either the undergraduate or graduate level.
Although continuity in thought (as well as in notation) is maintained in
the two volumes, the material in Volume 2 is also made sufficiently self-
contained so that advanced students may study Volume 2 directly, once
they have acquired some knowledge of rigid-body mechanics.
The authors take pleasure in acknowledging their indebtedness to
many of their teachers, friends, colleagues, and former students for their
interest and encouragement. In the early phase of the work at Johns
Hopkins, much support was given to the senior author by Professor
J. Trueman Thompson, Professor George F. Wislicenus, and Dean
Robert H. Roy during that difficult period when the importance of

engineering science was not yet widely acclaimed. Thanks are also due
to the mechanics committee at Johns Hopkins that initially outlined the
scope of the course that was later to become this book. Among the
senior author's assistants and students at that time, special thanks are
due to Dr. Dominic Edelen and Messrs. Lester L. Jones, Jr., and Howard
E. Holland for their help in various ways. Professor Maurice A. Brull
of the University of Pennsylvania kindly reviewed the final manuscript
and offered many valuable suggestions and contributions.
Hsuan Yeh
JoelT. Abrams
Contents
Preface V
Chapter 1. Vector Algebra 1
1-1. Scalars and Vectors 1
1-2. Geometrical Representations of Vectors 2

1-3. Equality of Vectors 3
1-4. Product of a Vector and Scalar 3
1-5. Addition and Subtraction of Vectors 3
1-6. Unit Vectors 5
1-7. Components of a Vector 6
1-8. Addition of Vectors by the Method of Components 7
Problems 9
Chapter 2. Newton's Laws; Concurrent Force Systems; P^quilibriuni of

a Particle 12
2-1. Newton's Laws 12

2-2. Discussion of Newton's Laws 13
2-3. Statics and the Equations of Equilibrium of a Particle ..... 15
2-4. Simple Structures 17
2-5. Friction 21
2-6. Procedure of Solving Prol)lems in Statics of a Particle 24
Problems 25
Chapter 3. Vector Products 29
3-1. Introduction 29
3-2. Scalar Product 29
3-3. Vector Product 33
3-4. Physical Significance of Scalar Product; Work; Line Integrals ... 36
3-5. Physical Significance of Vector Product; Moment 39
3-6. Moment of a Vector 43
3-7. Velocity Due to Rotation; Angular-velocity Vector 45
3-8. Triple Vector Products 46
Problems 49
ix
X Contents
Chapter 4. Statics of Systems of Particles and of Rigid Bodies ... 52
4-1. Equilibrium of a System of Particles 52

4-2. Constraint of a Rigid Body in Space 54
4-3. Equilibrium of a Rigid Body in Space 56
4-4. Procedures of Solving Problems in Statics of Rigid Bodies .... 57
4-5. Examples of Statics of Rigid Bodies 60
4-6. Couple 65
4-7. Equipollence of Force Systems in Space 66
4-8. Reduction of Force Systems; Wrench 68
Problems 72
Chapter 5. More Applications of Statics 77
5-1. Systems of Parallel Forces 77

5-2. Center of Gravity; Center of Mass; Centroid 79
5-3. Pressure in a Static Fluid 85
5-4. Forces Due to Fluid Static Pressure: Plane Surfaces 89
5-5. Forces Due to Fluid Static Pressure: Curved Surfaces 94
5-6. Buoyant Force Stability of Floating Bodies
; 96
Problems 99
Chapter 6, Derivatives of Vectors; Kinematics of a Particle .... 104
6-1. Derivative of a Vector 104

6-2. Derivative of Sums and Products 105
6-3. Curves in Space; Principal Normals and Binormals 107
6-4. Displacement, Velocity, and Acceleration of a Particle 110
6-5. Kinematics of a Particle: Rectangular Coordinates Ill
6-6. Kinematics of a Particle: Tangential and Normal Components . . 113
6-7. Kinematics of a Particle: Cylindrical Coordinates 114
6-8. Kinematics of a Particle: Spherical Coordinates 116
Problems 119
Chapter 7. Kinematics of a Rigid Body; Relative Motions .... 123
7-1. Displacements of a Rigid Body 123

7-2. Kinematics of a Rigid Body 126
7-3. The Eulerian Angles 131
7-4. Motion of a Point of a Rigid Body 133
7-5. Four-bar Linkages 136
7-6. Rate of ('hange of a Vector in a Moving Frame 139
7-7. Motion Referred to a Moving Coordinate System 141
I'roblenis 147
Chapter 8. Dynamics of a Particle 153
8-1. General Considerations 153

8-2. Simple Harmonic Motion 157
8-3. Motion in a Resisting Medium 161
8-4. Central Forces; Planetary and Satellite Motions 162
8-5. Example: Satellite ()rl)its 167
8-6. Work and Kinetic lOnergy 170
8-7. I'otciitial iMiergy 171
8-8. Examples of Conservative Force Fields; Gravitation 173
8-9. lOxample: The Simple Pendulum 176
Contents xi
8-10. Motion of a Particle in a Moving Coordinate System 181

rrohlenis 184
Chapter 9. ^ ihruting Systems . .'

188
9-1. ("iciicralized Coordinates and Degrees of Freedom 188
9-2. Vibrations 189
9-3. Differential Equations of One-degree-of-freedoin Vil)rating Systems . 191
9-4. Free Vibrations without Damping 193
9-5. Free Vibrations with Damping 197
9-6. The Case of Negligible Mass; The Differential Equation tx + x = 200
9-7. Forced Vibrations without Damping 202
9-8. Forced Vibrations with Viscous Damping 207
9-9. Vibration Isolation 210
9-10. Torsional Vibration 213
9-11. Vibrations with Xon}x>riodic Forces 214
9-12. More Complicated Vibrating Systems; Nonlinear Vibrations
Problems
... 215
219
Chapter 10. Dynamics of Systems of Particles 223

10-1. Motion of the Mass Center 223
10-2. Impulse and Momentum of a System of Particles 224
10-3. Angular Momentum of a System of Particles 227
10-4. Systems with Variable .Mass; Rocket and Jet Pro{)ulsion 233
10-5. Kinetic Energy of a System of Particles 236
10-6. The Virial Theorem 238
10-7. Elementary Kinetic Theory of Gases 238
10-8. Viscosity of a Gas 241
10-9. Statistical Mechanics 244
Problems 245
Chapter 11. Moments and Products of Inertia 249

11-1. Introduction and Definitions 249
11-2. Translation of Coordinate Axes 252
11-3. Moments and Products of Inertia of Some Simple Bodies 253
11-4. Rotation of Coordinate Axes 256
11-5. Orthogonal Transformations 258
11-6. Moments and Products of Inertia with Respect to Rotated Coordinates 260
11-7. Cauchy's Inertia Ellipsoid 262
1 1-8. Principal Moments of Inertia and Principal Axes 263
11-9. Plane Rotation of Axes; Mohr's-circle Representation 268
Problems 274
Chapter 12, Dynamics of Kigid Bodies 277

12-1. Euler's Equations of Motion of a Rigid Body 277
12-2. Work-and-<>nergy Ecjuation of a Rigid Body 282
12-3. Example: Motion of an Air-borne Vehicle 286
12-4. Stability of the Rotational Motion of a Rigid Body 290
12-5. Types of Motion of a Rigid Body 293
12-6. Rigid Body under No Forces; Poinsot's Representation 294
12-7. Free Motion of a Symmetrical Gyroscope 298
12-8. Symmetrical Gyro.scope under External Moments 303
12-9. The Gyrocompass 307
xii Contents
12-10. Rigid Body in Translation 310

12-11. Rotation about a Fixed Axis; Dynamic Balancing 311
12-12. Plane Motion of a Rigid Body .316
Problems 319
Chapter 13. Lagrange's Equations 326
13-1. Introduction 326

13-2. Virtual Displacements; Stability of Equilibrium 328
13-3. Lagrange's Equation for a Particle 333
13-4. Lagrange's Equation for a System of Particles 338
13-5. Small Free Vibrations of Coupled Systems 340
13-6. Small Vibrations of Two-degree-of-freedom Systems 342
13-7. Free and Forced Vibrations of n-degree-of-freedom Systems; Normal
Coordinates 346
Problems 356
Appendix A. Properties of Plane Geometric Shapes Al

Appendix B. Properties of Solid Homogeneous Bodies . A5
Answers to Selected Problems A9
Index A17
CHAPTER 1
Vector Algebra
1-1. Soalars and Vectors

Many of our common physical quantities fall into one of the two
classes: scalar quantities (or scalars) and vector (juantities (or vectors).
A scalar (juantity has magnitude, but does not have direction. It
therefore can be represented by a number which gives the
single real
magnitude of the quantity as a ratio to a specific unit of the same quantity.
Thus, we speak of a mass of 10 lb or a volume of 3 ft^, mass and volume
being scalar quantities. Other common scalar quantities are density,
temperature, energy, work, electric charge, etc.
A vector quantity,! on the other hand, requires for its specification not
only a number but also a direction. A

displacement or a velocity can be
completely specified, for example, only by (1) a number representing the
magnitude of the quantity as a ratio to a unit fjuantity, as for example a
displacement of 3 ft or a speed of 5 fps, and (2) a direction along which the
displacement or the velocity is to be oriented. Such quantities are called
vectors. Other common vector (juantities are force, acceleration,
moment, angular velocity, angular acceleration, and electric and magnetic
field intensities.
In addition to possessing a magnitude and direction, a vector quantity

must, as part of its definition, also obey a specified operation, the paral-
lelogram law of addition (Sec. 1-5). Finite rotations, although they
possess both magnitude and direction, cannot be classified as vector
quantities .since they do not follow the parallelogram law. A more
complete definition of a vector based on its transformation properties will
be given in Sec. 14-1, and a more precise illustration of the fact that finite
rotations are not vector quantities will be given in Sec. 7-1.
t It might be instructive for the student interested in motivation to read at this

point the last paragraph of Sec. 14-1 of Vol. 2.
1
2 Vector algebra
It willbe shown later in the text that there are other physical quantities
of a still more complicated nature than vectors, and they therefore require
more specifications than those of a vector. However, we shall not con-
cern ourselves with such quantities until we reach Chap. 11.
Vectors are often denoted in print by a Gothic or a boldface letter, A.
In manuscript they are usually denoted by an arrow placed above, as, for
example, A . A plain letter A or the vector enclosed by the absolute-value
sign, |A|, shall mean the magnitude of the vector A.
1-2. Geometrical Representations of Vectors

Any vector A
can be represented geometrically by a directed line seg-
ment or arrow (in space) whose length corresponds to |A|, the magnitude
of the vector, and whose direction is the same
as the direction of the vector A (Fig. 1-1).
Since we have not specified the point of origin
for the vector A, each one of the directed line
,y^ segments in Fig. 1-1 represents equally well
•
This does not mean that
the given vector A.
all vector quantities can be shifted around
with regard to their origin under all circum-

stances. One must always inspect the nature
of the physical problem to determine whether a shift of origin of the vector
in question is permissible.
Vectors may be classified into three general types that categorize the
permissible shifting. These are the free vector, the sliding vector, and
the bound vector. may have its origin shifted anywhere
The free vector
in space, whereas the sliding vector may be shifted anywhere along its line
of action. On the other hand, a bound vector is fixed to one origin or
point of action and may not be moved from that position. As we have
previously noted, the physical problem determines the type of vector with
which we are dealing. In Fig. 1-1 the vector A as illustrated is a free
vector. Unless noted to the contrary, we shall use the word vector in this
text to denote a free vector.
Example 1-1. If two equal and opposite forces —A and A are applied
to a \nivPR, with their points of application P and Q such that the line PQ
has the same direction as that of the two forces, as shown in Fig. l-2a, the
body is said to be in equilibrium. (See Chaps. 2 and 4 for the exact mean-
ing of ('(|iiilil)rium.) Now, if we shift the point of application of A from Q
to R along its line of application (Fig. 1-2/^), the body is again in equilib-
rium. M()reo\-cr, if we shift the vc(;tor A to position (Fig. l-2c) and if
we imagine flic line ()(J to represent a perfectly rigid and weightless con-
ned ion lot he bar /V»', it is easy to visualize that the eciuilibrium of the bar
+ Q__A \
A
4 Vector algebra
in Sec. 1-2 and Fig. 1-3 next ; we draw the vector B from the terminal of A.
The new vector (in this case vector C) that joins the origin or beginning of
A with the terminal of B is the sum of A and B. It can be seen that the
same resultant C is obtained if we first draw B, then draw A at the terminal
of B, and finally join the beginning of B with the terminal of A. In either
case the sum C (represented by oc) of the two vectors A (represented by oa
or he) and B (represented by oh or ac)
is the diagonal of a parallelogram
(oach)formed by having A and B as
the two neighboring sides in the man-
ner shown in Fig. 1-3. In many engi-
neering applications we term the sum
Fig. 1-3
two vectors A and B the result-
of the
two vectors.
ant of the
The magnitude C of the vector C may be related to A and B, the magni-
tudes of the vectors A and B, by use of the cosine law of trigonometry:
C^ = A^ + B'- -^ 2AB cos 6
where 6 is the angle between the positive directions of the two vectors A
and B. By suitable trigonometric relations, the direction of C in Fig. 1-3
may be easily determined.
From the above graphical construction, it is clear that the addition of
vectors is commutative; that is,
A 4-B = B -1- (1-1)
It can also be shown that the addi-

tion of vectors is associative; i.e.,
(A -h B) -f C = A -f (B + C) (1-2)
To prove this, we refer to Fig. 1-4.

Following the graphical construction,
the sum of A -f B is represented by oh,
and therefore (A -f B) -|- C is repre-
sented by oc. On the other hand,
(B + C) is represented by ac, and there-
fore A 4- (B + C) is again represented
Fig. 1-4
by oc. Hence Eq. (1-2) is true. It
thcroforc is permissible to drop the parentheses on either side of Eq. (1-2)
;iii(l write the sum of three vectors simply as
A +B+C
Figure 1-1, roprosoutiug the addition of vectors, is called a vector polygon,
iiud the j?,-;,])jii(.;,l niclhod of oonstructiou is termed the polygon rule.
Sec. 1-6] Unit vectors 5
Once we define the addition of vectors, the subtraction of xectors

immediately follows as the inverse operation: The vector B to be sub-
tracted is reversed, and the rule of addition again applies. Thus
A-B = A + (-B) (1-3)
An obvious fact is that if vector B is equal to vector A, then by Eq. (1-3)

A - A = (1-4)
On the other hand, A -f A gives, following the rule of addition, a new

vector having the same direction as A
but twice as large in magnitude;
therefore we write A -t- A = 2A. This is seen to be consistent with the
definition of the product of a vector and a scalar of Sec. 1-4.
It also follows from the definition of the product of a vector and a scalar
and the definition of addition of vectors that
nA -\- mA = (n -\- m)A (l-5a)

and that w(A + B) = nA + nB (1-56)
Hence the multiplication of a vector by a scalar is both distributive and

associative.
The can be extended to any numV)er of vectors
results of this section
oriented anywhere in space, since by a process of continuation we can
consider the vectors in successive pairs. In particular, we note that the
addition or subtraction of two vec-
tors yields a third vector that is t
coplanar with the first two; i.e., the

sum or difference of any two vectors
lies in the plane of the original two
vectors.
1-6. Unit Vectors
A vector whose magnitude is unity

is called a unit vector. A unit vector
may have any specified direction.
Any vector can be considered as the
product of its magnitude and a unit
vector along the same direction.
Thus, if n is a unit vector having
the same direction as A, we can write
that A = An. Unit vectors having
directions along three mutually perpcndi(ûlar coordinate axes are called
înit orthogonal vectors. If we choose the cartesian coordinate system with
the three axes x, y, and z, the three unit vectors are denoted by i, j , k, with i
along X, j along //, and k along z (Fig. 1-5).

6 Vccl.or (il.(j('.})ra
Thr; curtfiHJjui c.aovd'mu.U', Hy.sinm wo have choHcn f;f)nH),iiu<,f;H a rij^lii-
hurifJcoordinate Hy.stem .such thai if we rotaU; t,he jixi.s iril,o iJie yy ./; axi.s, a
ri^hl-hiUKJ screw would advance; alon^ the posit/ive dinsciion of ifie ^; axin.
'rhree-(Jifn(!nHional I<]iiclideiJ,n ,sf>n,cc! iw well ref)resenl,e(l hy this orUiojônal
coonJin!i,l,e .syHtern. Ah hucJi, ilie l,hr(!e orJ,fio}ônal unit vecl/ons i, j, an(i k

aHHUrne j^rciU, itn(jofl,;U)(;f;.
1-7. (l«>rri|M»iictils of a V<'«l<>r
liy r(;ver,sin}^ tlie ruh; of Uie {^rii-pliicM,! a<J(htioii of vcjfitorM, it (;aii bo soon
tFial/ a vector A can b(; (Jocoinf)OHed into any ruimbor of component vectorH
Sec. 1-8] Add il ion of veclors by the method of components 7
along the three coordinate axes. The projected lengths A^, Ay, A^
represent the three components of A, or as they are sometimes called, the
three scalar components of A. From the geometry of Fig. 1-6, it is
evident that
A = VAJ + /!,2 + A,2 (1-6)
B}' the rule of vector addition we see that
A = AA + .4,j + AM (1-7)
The above equation expresses any vector A as a sum of its three orthogo-
nal components, or vector components, and one of the most frequent is
expressions in vector algebra. Equation (1-7) can be considered as a

standard form for expressing a vector. Any vector can always be written
in this form.
If the angles measured to A from the positive x, y, and z axes are a, 0,
and 7, respectively, then the three components of A are
Ax — A cos a
Ay = A cos & (1-8)
A^, = A cos 7
The three cosines cos a, cos ^, and cos 7 are called direction cosines of A.
They are related by the following equation:
cos- a + cos- /3 -f cos- 7 = 1 (1-9)
obtained by substituting Eqs. (1-8) into Eq. (1-6).

Because of the relationship between the three direction cosines [Eq.
(1-9)], only two of the three direction cosines are independent. The third
isdependent and is obtained by a direct application of Eq. (1-9). Thus
the complete specification of the direction of a vector (or any line for that
matter) requires only two direction cosines.
1-8. Addition of Vectors by the Method of Components
If we have two vectors A and B, we can write, following Va\. (1-7),
A = ^î Ay} + AM
-I-
B = 5,1 By'} BM
-1- -f-
Using the associative and commutative laws of vector addition and

Eqs, (1-5), we have
A -1- B = (^, -h B,)i -f {Ay -f By)} -h {A, A- B,)k (1-10)
where this plane intersects the x axis, say at point c, determines the projection of A
along the x axis. This projected length Oc is usually denoted by the symbol Ax. It
should further be noted that Ax is an algebraic scalar quantity.
8 Vector algebra
If we let A+B = C and write
C = C.i + C,j + CM
then Ci = Ax + x>x
Ly = Ay -\- By (1-11)
C. = A, + B,
We have thus proved a very important fact: The component along any
direction of the sum of vectors is equal to the sum of the components of the
vectors along that direction. Although Eqs. (1-11) are derived for the sum
of onlytwo vectors, they can be easily
extended to the sum of any number of
vectors.
The above fact can be conveniently
used to calculate the resultant (i.e.,
sum) of any number of vectors. The
procedure is decom-
as follows : First
pose each vector into its components
along the x, y, and z axes; next form
the algebraic sum of all the scalar com-
ponents along X and similarly for all
components along y and z; and then
finally form the resultant by vector
addition of the three resultant vec-
tor components. This is called the
method of components.
It was pointed out in Sec. 1-3 that
two vectors are equal if, and only if,
they have the same magnitude and
direction. Certainly, then, expressed
from the point of view of components,
Fig. 1-7
two vectors are equal if, and only if,
they have identical scalar components in any three noncoplanar direc-
tions.! The converse of this statement is obviously true also. From this
it follows that if a vector is equal to zero, each and every one of its scalar
components must necessarily be equal to zero. In general, then, a vector
equation A = is equivalent to the three scalar equations
A, = Ay = .4, =
Example 1-2. Given two forces A and B having magnitudes
A = 2 -\/lO lb and B = 15 lb, respectively, and directions as indicated
in Fig. 1-7. Find the resultant force R = (A -1- B).
t It is most usual to take the orthogonal directions as the tliree noncoi)lanar

directions.
Problems 9
Solution. According to Eqs. (1-8), the scalar components of the

vectors A and B are found to be
A, = -^
Vio
X 2 VTO = 2 A, = -^
Vio
X 2 VTO = 6 .1, =
fix = By = 3.^ X 15 = 9 B, = +5 X 15 = 12
lOxpressing the vectors in the standard form of E(i. (1-7),
A = 2i + 6j + Ok lb
B = Oi 4- 9j -I- 12k lb
Xow, using the results of Eqs. (1-11), we have

i?^ = 2 Ry = 15 R,= 12
so that R = 2i -f- 15j -|- 12k lb
The magnitude of R, from Eq. (1-6), is
R = \/2- + (15)2 + (12)2 = y'373 i^
and the direction cosines of the resultant force are, from Eqs. (1-8),
2 15 12
cos a = cos
,
/3
= cos 7
\/373 -\/373 V'373
PROBLEMS
1-1. ^^'hat is the significanco of A/,l?
1-2. Show that |A| + 1B| > \\ + B\.
1-3. Determine the sum and difference of the two vectors A and B:
A = 6i - 4j - 6k
B = 4i - 2j - 8k
What is the magnitude of the resulting vectors?

1-4. For the following vectors find their magnitude and direction cosines:
A = 4i + 5j + 3k
B = 4i - 5j + 3k
C = 3i + 5j - 4k
1-5. A body is acted on by two plane forces, A, B, as

shown.
.4 = 10 lb li = 30 lb
Find the magnitude and direction of the resultant force

by two methods:
(a) Construct the parallelogram and solve by trigono-
Fio. P 1-5
metric relations.
(b) Decompose each force into its components and solve by Eq. (1-10).
10 Vector algebra
1-6. Given two forces Fi and F2:

Fi = 7 lb
Fo = 2 VT3 lb
Express Fi and F2 in the form of Eq. (1-7). Also find the magnitude and the direction
cosines of
(a) Fi + F2.
(b) F, - F2.
(c) F2 - Fj.
Fig. P 1-6
1-7. Referring to Fig. 1-6, find a, /3, 7 in terms of Aa A„, A,

1-8. Derive Eq. (1-9).
1-9. If the components of A
along x and y axes are Ax and Ay, respectively, what
are the magnitudes of the components of A along an inclined pair of axes x' and y' if
the angle of inclination between x' and x is 6? Assume the two pairs of axes lie in
the same plane.
ky
1-10. Extending the above problem to a three-dimensional space, we assume that

the three cartesian components of a vector A are Ax, Ay, and A^. Find the component
of A along an arbitrary direction (in space) whose direction cosines with respect to
the coordinate axes are ax, ay, a^.
1-11. A reciprocating engine consisting of crank, connecting rod, piston, and
cylinder assunuvs at a certain instant the position shown. The pressure at the back
Fui. P 1-11
Problems 11
of the piston exerts a horizontal force of 300 lb on the right end of tlie connecting rod.
Find the components of this force along the connecting rod and along a direction
perpendicular to the connecting rod.
1-12. An airplane climbs at a speed of 200 mph and at an angle of climb of 30°, both
relative to the air. If there is a vertically upwartl gust of 30 mph, what is the absolute
velocity of the airplane and its absolute angle of climb? (By absolute, we mean here
relative to the earth.)
200 mph
Fin. P 1-12
1-13. If n is a unit vector with direction cosines Ox, ay, and a., show that its scalar
components are simply equal to its direction cosines. Express n incomponent form.
1-14. A unit vector makes equal angles with each of the three coordinate axes.
Express this vector in terms of its vector components.
1-15. What is the locus of the extremity of the vector formed by A + A:B? Write
the equation of this locus if A and B are given in the form of Eq. (1-7) and A; is a real
number between — =o and + « (Assume that the origin of the above vector sum
.
coincides with the origin of the cartesian coordinate axes.)

CHAPTER 2
Newton's Laws; Concurrent Force

Systems; Equilibrium of a Particle
2-1. Newton's Laws

The basic laws of classical mechanics f were stated by Ne^î:on in 1687
and were formulated for a particle. They are, expressed in a slightly

revised form,
First Law. Every body continues in its state of rest, or of uniform
motion in a straight line, unless compelled to change that state by forces
impressed upon it.
Second Law. The rate of change of momentum is proportional to the
impressed force and occurs in the direction of the straight line in which
that force is impressed. Expressed in mathematical form, this means
F= (mv) (2-1)
J
where F = force
V = velocity
m = mass
mv = momentum
For an ordinary body, the mass is a constant, and we can write, alterna-
tively,
F = ma (2-2)
where a = d\/dt = acceleration. J

t By classical, or Newtonian, mechanics, we want to distinguish from quantum and
rclativisticmechanics. The latter is necessary to exphxin many atomic phenomena
(for example, the lack of determinacy), but is not required in deahng with ordinary
problems that can be adequately explained by classical mechanics. On the other
hand, problems dealing with bodies traveling with speeds approaching the speed of
light (some pr()l)lems in celestial mechanics, for exami)le) need he examined from the
point of view of rclalivistic incchanics.
X The derivative of a vector (such as v) with icsjx'ct it) a scalar (su(di as t) will be
discussed in Sec. (i-1.
12
Sec. 2-2] Discussion of iS'ewlons laws 13
Third Law. To every action there is always opposed an equal reaction;

or the mutual actions of two bodies upon each other are always equal and
opposite in direction.
These are the basic laws that serve as a foundation for our study of
mechanics. We shall proceed in the next section to discuss briefly some
of the ramifications of these laws.
2-2. Discussion of Newton's Laws
Although Newton's laws are familiar to readers of this text, there are
scA'eral aspects of these laws that need to be emphasized again. Past
experience shows that the following points are often not sufficiently clear.
1. A bod}' can be said to be in motion or at rest only when compared
with another body which serves as the frame of reference. In Newton's

laws, the frame of reference might be a fixed star when the laws are
applied to astronomical problems, but in daily engineering applications,
sufficient accuracy is obtained by taking the surface of the earth as the
frame of reference. Motions relative to the surface of the earth are often
called absolute motions in engineering practice, for the sake of brevitj\
(The surface of the earth obviously would not be a suitable absolute frame
of reference when the rotation of the earth exerts a strong influence on the
problem. Such problems are to be found, for example, in high-altitude-
trajectory studies, meteorology, and oceanography.)
2. Newton's laws in their simplest form are concerned only with a
particle. A particle is assumed to possess a mass, but analogous to a

geometrical point, it has no dimension. The application of these laws to
a system of particles (e.g., rigid or deformable bodies) requires some
further rigorous mathematical treatment. Too often the beginner takes
too many things for granted because in elementary physics he was led
from particle mechanics to the mechanics of rigid bodies w'ithout too much
effort. Actually, such simplicity is possil)le only if the motion of bodies
is restricted to a few simple cases. It is the purpose of this text to present
the mechanics of rigid and deformable bodies in a somewhat wider
perspective so that many more types of applications can be adequately
described. Although Newton's laws are still the only fundamental laws
that we shall use, the development from there on often demands our
careful consideration, especially if the problem is three-dimensional in
nature.
3. Although the third law states that action and reaction are always
equal and opposite, this docs not mean that these two forces always
cancel out, because these two forces do not act on the same particle. (Two
equal and opposite forces acting on the same particle always cancel out.)
Thus, if we take both particles as our system, the action and reaction are
14 Newton s laws; concurrent force systems; equilibrium
internal forces and their influences do cancel out. On the other hand, if
we take only one of the two particles as our system, then only one of the
two forces is present.
The paragraph also illustrates the importance of clearly defining
last
one's system. Picking a system (or in the language of applied mechanics,
choosing afree body) is a prerequisite to the solution of almost all problems
in mechanics and sometimes requires rather careful thinking. Once a
free body (at present the particle) is chosen, all the forces that act on that
particle should be clearly indicated. The forces that need be considered
generally are the gravity force, or the weight of the body which is usually
assumed to act vertically downward (or more accurately, but still not
exactly, toward the mass center of the earth), all reactions and friction
forces due to the constraining of the particle, and finally all other forces
due to the mutual action of neighboring bodies and external agencies.
This concept will be illustrated by a number of examples in subsequent
discussions.
Clearly, what we choose as our free body depends to a large extent on
the idealizations we employ. The depend on the
idealizations themselves
type and accuracy of the information we desire. For example, if we are
interested in the motion of the earth around the sun, we might idealize
the earth as a particle traveling around another particle, the sun, fixed in
space. On the other hand, if we were interested in earthquake phenom-
ena, we might idealize the earth as an elastic sphere, fixed in space.
Neither of these models or idealizations exactly depicts the earth; the
validity of the idealization must be confirmed or rejected by experiment.
We are taking the view in this text that Newton's three laws are
fundamental postulates and thus serve as our basis for the study of
mechanics. Certainly this does not imply that these laws will remain in
their present or unaltered form forever. A little reflection on their
comparative newness in the historical sense would indicate that it
required many centuries for the enumeration of these principles in their
present form, and it is entirely possible that in centuries to come addi-
tional knowledge in the field of mechanics could lead to their modifica-
tion. Any future modification of such basic postulates is more than
likely to include the present laws as a particular case. This is so because
the present laws seem to apply so very well to most situations; therefore
one would expect any future discovery to be generalization rather than
contradiction.
Our discussion, then, will be centered around the use and interpretation
of these basic laws. Any discussion relative to the more philosophical
aspects of these laws will l)e omitted. Sufiico to say, then, that the
physical quantities, specifically force and mass, related by the laws exist
and that our knowledge of them is intuitive. Velocity and acceleration,
Sec. 2-3] Sialics and Ihe equations of equilibrium of a parlicle 15
of course, will he defined in terms of the spatial motion of a point, that is,
in geometrical terms. By momentum of a particle we mean (without

being too explicit for the present) the product of its mass and velocity.
2-3. Statics and tlie E({uatioiis of

Equilibriiini of a I*arlicle
A particle is said to he in a state of static efjidlihrium, or simply equilib-

rium, when it is either at rest or in uniform motion along some straight
line. The branch of mechanics that deals with particles (or systems of
particles) under such a state of affairs is called statics. According to
Newton's law, a condition as just described may be achieved if, and
first
only if, the total, or resultant, force acting on the particle is ecjual to zero.
It follows, then, that a particle is in equilibrium when the resultant of the
forces acting on the particle is zero.
Thus, if Fi, F2, . . . ,
F„ are the forces acting on the particle, we may
express analytically the neces.sary and sufficient conditions for the etjui-
librium of a particle as
2 Fi = F, + F, + • • •
+ F„ = (2-3)
t = i
n
or, introducing the notation F^ = ^ F = > F,- for brevity,

1 = 1
F, = 2F = (2-4)
where F, is the resultant force, or sum of all forces, acting on the particle.
The above vector equation can also be expressed in the form of three
scalar equations if we deal with the components of the vectors. Let us
write, for example,
Fi = F,A + F,J -I- F,,k
Then Eq. (2-3) or (2-4) is eciuivalent to the following three scalar ecjua-
tions, which, following Sec. 1-8, are both necessary and sufficient condi-
tions for the equilibrium of a particle:
Fr. = 2F, = Fu + F2, -i- • • •

+ F„, =
Fr, = ZFy = F,„ -h F2„ + • • •
+ Fn„ = (2-5)
F,, = 2F. = Fu + F,, + • • •
4- /^„. =
Equations (2-3) to (2-5) are called equations of equilibrium of a particle.
It be shown in Chap. 4 that these same equations, although
will
formulated for a particle in equilibrium, also apply to and are ncccssanj
for a finite-size body in equilibrium. It is this extension that enables us
to use the equations of equilibrium of a particle to solve problems in the
equilibrium of the finite-size bodies that we find in practice. As long as

the forces acting on the body are concurrent (i.e., meet at one point), we
may treat or idealize the body as if it were a particle, a mathematical
abstraction.
Equilibrium of a particle, according to Eq. (2-4), requires that the
resultant force acting on the particle vanish. Expressed geometrically,
this statement may be interpreted to mean that the system of forces acting
on the particle forms a closed polygon. Thus, for example, it is clear that
if a particle in equilibrium is acted upon by two forces, they must neces-
sarily be equal in magnitude and opposite in direction and have the same
line of action, or be coUnear. On the other hand, if a particle in equilib-
rium is acted upon by three forces, the force polygon becomes a force
triangle, and the three forces must lie in one plane; i.e., they must be
coplanar.
Example 2-1. An electric lamp of weight W
is suspended at the end
of a string AB, the upper end of which is attached to the ceiling by three
(a) (b) (c)
Fig. 2-1
strings of equal length /. The points of attachment CDE of the strings

make an eciuilateral triangle of sides o in the plane of the ceiling. Assum-
ing the strings are weightless and perfectly flexible, find the force in the
strings (Fig. 2- la).
Solution. First we isolate the electric lamp (considered as the free
body) (Fig. 2-16). There are two forces acting on it: the gravity force IT
and the force 7'i due to the action of the string AB. Since the lamp is in
equilibrium, we nuist have, according to Eqs. (2-5), ^Fy = Ti — = 0, W
or 7'i = IF. (In prol)lcms of statics, this conclusion is often stated
immediately without having to go through any analysis.)
Next we considcf lli(> joint .1 as the free body (Fig. 2-lr). By reason of
Sec. 2-1] Simple slruclures 17
-ymmetry it is clear that the forces in the three strings AC, AD, and AE
are the same. Let us denote them by T2. Because of the ecjuiUbrium of
the joint A, we have, according to Eqs. (2-5),
where we have used the fact that the cosine of the angle between the
three strings and the vertical is
1 - —
V 3/2
The argument of symmetry is often a convenient one. By symmetry

we mean that the three strings are all oriented similarly with respect to
the force Ti. As a result of this lack of preferred direction, the forces in
the strings have no preferred magnitude and thus are all equal. It would
1)6 instructive for the student to verify the results of this example without
lecourse to the symmetry properties by using each of Eqs. (2-5) (see

I'rob. 2-5).
Again because of symmetry, the remaining two equations of equilibrium
[Eqs. (2-5)] are automatically satisfied. We thus obtain the tension in
each of the three supporting strings:
3 Vl - a^SP
It should be noticed that in this problem the physical consideration
enabling us to obtain a solution was the fact that the entire system, in this
instance composed of the lamp and all the connecting strings, is in a state
of static equilibrium. Here we have tacitly assumed that
system in for a
equilibrium each and ever}' particle comprising that sj^stem must also be
in equilibrium. That this assumption is valid can be easily verified by a
consideration of the definition of equilibrium. The directions of the
forces in the strings are indicated in Fig. 2-lc. Remembering that Fig.
2-lc indicates the action of the strings on the particle .1, the particle -I in
turn reacts in an equal and opposite fashion on the strings. Therefore
the force acting on the string tends to elongate the string. We term this
type of force a tensile force; the string is in tension. Forces that tend to
contract or shorten a body are known as compressive forces; the body is in
compression. One characteristic of a flexible string is that it can sustain
only tensile forces.
2-4. Simple Structures

In the preceding example we assumed that the string is flexible and
weightless; such an ideal string (or rope, or cable) can sustain only tensile
forces in the direction of its center line. On the other hand,

rigid " bar a '^
can sustain both compressive and tensile forces. Rigid bars are therefore
essential elements of all engineering structures. In the analysis of
structures, it is usually permissible to assume that the bars are hinged
without friction at their two ends. In actual practice the bars may be
riveted or welded together, often with addition of gusset plates; but the
assumption of frictionless, hinged joints gives a sufficiently good answer as
a preliminary analysis, requiring only a fraction of the labor needed for a
more detailed analysis. Bars used in structures are usually slender in
cross section when compared Avith their lengths and are usually considered
weightless in comparison with the applied forces. A weightless, friction-

less,hinged bar under no lateral loads between its two ends can transmit
only compressive or tensile forces along its center line (Fig. 2-2). This is
intuitively convincing, but a rigorous proof will be given in Chap. 4.
Example 2-2. A portable crane loaded as shown.

is The vertical
mast AB of the crane is rigidly supported by two guy wires BD and BE.
The boom AC
can swing around the mast. For the case under investiga-
tion, the vertical plane ABC is in such a position that AF, the line of inter-
section of the vertical plane ABC with the horizontal plane ADE, forms
an angle a with AD and an angle (3 with AE. Assuming all hinges are
frictionless, find the tension in the two guy wires (Fig. 2-3a).
Solution. First we isolate point C as a free body and investigate its
e(iuilil)rium. Certainly, if the entire structure is in equilibrium, point

C is also in equilibrium. The forces on AC act only at its two ends; hence
we can assume that the reaction Fi in AC is along the axis AC. The free-
body diagram of point C is therefore as shown in I'îg. 2-36, which gives a
view in the vertical plane ABC. We choose for our coordinate system the
X, y, z axes as indicated. The force Fi in AC and the force Fo in BC are
Fi = -Fi sin 45° i - Fi cos 45° j -t- Ok

Fo = —Fn sin T i — F... cos 7 j 4- Ok
Sec. 2-11 Sim/)le sirticlures 19
iy
(c)
(6)
Fio. 2-3
Therefore, using the equations of etjuilibrium [Eqs. (2-5)], we find
vF^ = -Fi sin 45° - F^ sin 7 =

1T„ = -Fi cos 45° - F2 cos 7 - ir =
The third equation, 2Fj = 0, is identically zero, since none of the forces
has a z component. Solving the above simultaneous equations, we find
\/2 IV sin 7
Fi= - — cos 7
sni 7
F, =
sm 7 — cos y
20 Newton's laws; concurrent force systems; equilibrium
Since 7 is larger than 45°, Fi turns out to be a negative quantity and Fo a

positive quantity. In turn, this means our assumed directions of Fi and
F2 in Fig. 2-36 are, respectively, incorrect and correct. Therefore the bar
CA is in compression, the string CB in tension.
Next we take the point B as the free body and consider its equilibrium.
Again we assume that the reaction in AB is along AB because the forces
on AB act only at the ends. We let the reactions in the two guy wires
BD and BE be F3 and F4 and the reaction in the mast AB be F5, acting in
the negative y direction. The horizontal projection of the forces on B is
shown in Fig. 2-3c. (The reaction in AB or F5 has no horizontal com-

ponents.) The various forces at B may be expressed as
= Fo sin y i -\- F2 cos 7 j -|- Ok

= -F3 cos 40° cos a i - F3 sin 40° j - Fs cos 40° sin a k
= -Fi cos 40° cos - /3 i F^ sin 40° j + F4 cos 40° sin /3 k
= -F,\
'5J
Writing the equations of equilibrium,
2F:, = F2 sin y - Fs cos 40° cos a - F^ cos 40° cos ^ =

XFy = F2 cos 7 - ^3 sin 40° - F^ sin 40° - ^^5 =
XF, = Fi cos 40° sin (3 - Fs cos 40° sin a =
The first and last equations can be solved simultaneously for F^ and F4.
Combined with the expression iorFo obtained before, we find the value of
Fz and Fi in terms of the weight W.
For example, if a = /?, then
W
2 cos 40° cos a (I — cot 7)
After we find F3 and Fi, the equation of equilibrium 2Fy = jdelds the
reaction F5.
Remarks. In the case of joint C, all the forces lie in one plane. This is
called a plane problem. In a plane problem only two of the three equa-
tions of equilibrium are nontrivial; hence we are able to solve for two
unknowns, magnitudes of the two forces (Fi and Fo). In
in this case the
the case of joint B the forces do not all lie in one plane. This is a space,
or three-dimensional, problem. In a space problem we have at our dis-
posal all three ecjuations of e(iuiHbrium; hence we are able to solve for
three unknowns, in this case the magnitudes of F3, F4, and Fa.
We also observe from the preceding solution that the initial choice of
sense of Fi and F2 is completely arbitrary. Whether this choice is correct
or not will be completely determined by the ensuing solution, a positive
rcsulf indicating the original selection was correct. However, once this
choice has been made, it is iinp(M-at i\(> that in all other free bodies that may
Sec. 2-5] Friclion 21
contain this force, the direction he consistent with the initial choice.
Remenil)ering that Fi. for example, acting at joint C in Fig. 2-3^, repre-
sents the action of member CA on -joint C and that in turn joint C,
according to Newton's third law, reacts on member CA in an efjual and
opposite fashion, we see that joint C "pushes" on CA, and thus CA is in
compression. It is usually not wise to attempt to guess beforehand
whether the force in the various members is tensile or compressive. A
better approach would be to solve the problem in a straightforward
manner, assuming that all the forces are tensile, and then subsequently
checking the results with physical intuition. In this fashion a sign
convention is established, positive indicating tension, negative indicating
compression.
It can be seen that if we had started by taking joint B instead of
joint C as our free body, we would have too many unknown magnitudes
{F i,P z,F i,F h) in our three equations of equilibrium. The proper sequence
of taking free bodies in a problem is therefore of importance if we wish
to solve a problem in the most convenient fashion. Occasionally a situ-
ation arises such that, regardless of the sequence of isolating joints, there
are more unknowns than there are available e<iuati()ns of e(iuilibrium.
Obviousl}^, then, a unique solution is not possible. We term these prob-
lems statically indeterminate (not de-
terminable by methods of statics
alone). They require for their solu-
tion additional information that
usually can be obtained by a con-
sideration of the deformation of
the structure caused by the various
forces.
2-5. Friction
When two bodies are in contact,
the forces of action or reaction ex-

erted by one body on the other can be
Fk;. 2-4
resolved into two components: one
which is normal and one which is tangential to the surface of contact. In
Fig. 2-4, R is the total force exerted by body B on
body .1 By Newton's .
third law, an ecjual but opposite force (not shown) is exerted by .1 on B.

The figure shows a sectional view in the plane formed by R and the normal
to the surface element of contact. Resolving R in this plane, we ol)tain a
normal component N and a tangential component F. The force N is
called the normal force, and F is called the friction force, or simply the
friction. In reality, the two contacting bodies may not have a single
: .
point of contact or even a single line of contact, but rather may have a
surface of contact. The contact forces, then, may be distributed over
some finite area. In this event the normal force and friction force we are
considering here are simply the resultant of the distributed forces and
represent a convenient idealization that enables us to achieve a reasonable
solution for problems involving friction. More will be said of distributed
forces in later chapters (Chaps. 5 and 15).
When a body is in contact with a perfectly smooth surface (certainly

such a condition constitutes an ideal mathematical model) the reaction is
always along the normal to the surface. This fact serves as our definition
of a perfectly smooth constraining surface.
The law of friction between dry surfaces was first formulated by
Coulomb and is now referred to as Coulomb's law of dry friction. This
approach represents the classical approach to a subject, which in reality
is not yet completely understood. The law states the following
1. When two contacting surfaces are at rest relative to each other, the
friction force reaches a maximum when sliding or slipping between the

two surfaces is imminent. This maximum friction is proportional to the
normal force, but is independent of the magnitude of the area of contact.
Expressed mathematically, this says
F = iXsN (2-6)
The constant of proportionality Ms is called the coefficient of static friction

The quantity Ms depends on the roughness and the material of the two
contacting surfaces and may vary from zero for "perfectly" smooth
surfaces to 1 or above for rough surfaces.
2. At low velocities of slipping the friction is again proportional to the
normal force and is independent of both the magnitude of the area of

contact and the velocity. This means that we can express the friction
again by the formula
F = n,N (2-7)
where ma.- is called the coefficient of kinetic friction. The coefficient of

kinetic friction is usually smaller than the coefficient of static friction;
hence maintaining a slipping recjuires a smaller force than inciting slipping.
It might he added that if the two surfaces are separated by a thin film
of fluid (such as in the case of a lubricated bearing), we have a case of

viscous friction, which follows entirely different laws. The amount of
viscous friction is proportional to the velocity of slipping and the area of
contact, but is independent of the normal force. (This fact is of sig-
nificantimportance in Chap. 9.)
Hetuniing to the laws of (hy friction, it should he emphasized that the
formula A' = ju.V holds only if there is imminent or actual slipi)ing. If
Sec. 2-5] Friction 23
slipping is impending. F = n^N, whereas if slipping actually occurs.

F = fik^- When the two contacting Ijodies arc merely at rest with
respect to each other and if slipping is-not imminent, thenF is not equal to
/i^.V. but rather/'"" is less than n^X. The following example will illustrate
this point.
Example 2-3. A flat, rectangular block of weight IT rests on an
inclined plane which makes an angle with the horizontal (l''ig. 2-oa).
</>
The coefficient of static friction between the block and the inclined plane
l"iG. 2-5
is n,, and we assume that ju., > tan </>. (Unless Hs > tan 0. the block will
not stay at rest on the inclined plane.) Find (a) the horizontal force P
(applied to the center of the block in a plane parallel to the inclined plane)
which will just cause the block to begin to slip, and (h) the friction force F
if P than the value required to start the motion.
is less
Solution. We take the block as the free body and consider its cciuilib-
rium (Fig. 2-56). The forces acting on the block, referred to the coordi-
nate axes shown, are
W = IF sin i - IF cos j
N = iVj
P = Pk
F = FJ + /'\k
\\'riting the equations of eciuilibiiuni, we find that
Zf'^ = IF sin + F^ - Fx = — IF sin <f>
^F, = -W cos + A^ =
</) N = W cos <t>
ZF, = P + F, = F, = -P
Hence F = \/FJT~F? = VlF^ .sin^ + P^
24 Newton s laws; concurrent Jorce systems; equilibrium
Let us now consider the two cases separately.

(a) When slipping is imminent, then /^ = iisN = n^W cos <^. Hence
P = W a/ms^ cos^ </)

— sin^ (f)
(6) If P is less than the above value, then slipping is not imminent, and
F = VW^ sin2 + P2
where P can be any value smaller than given in (a) . It can be seen that
here F ?^ iXsN and that F < jisN.
2-6. Procedure of Solving Problems in

Statics of a Particle
The three examples of this chapter illustrate the procedure of solving

problems body in which all the
of statics of a particle, of a joint, or of a
forces pass through one point (so-called concurrent forces). The pro-
cedure can be outlined as follows:
1. Choose a free body which may consist of one particle, one joint, or
one component part of the given system. A suitable free body is one
which is acted on by only two concurrent forces of known direction but
unknown magnitudes in a plane problem or by only three concurrent
forces of known direction but unknown magnitudes in a space problem.
Alternatively, in a plane problem the unknowns could consist of one
unknown magnitude and one unknown direction angle, and in a space
problem two unknown magnitudes and one unknown direction angle or
one unknown magnitude and two unknown direction angles, etc. The
essential feature is that there be a total of tivo unknowns in a plane
problem and a total of three unknowns in a space problem. (The reason
for such a choice was given in Sec. 2-4.)
2. Sketch the free body and show all the forces acting on it. The forces
will generally include all the known, or active, forces as well as all the
unknown, or reactive-type, forces that we ultimately desire to determine.

3. Choose and draw a convenient set of coordinate axes. (Items 2 and
3 constitute what is known as w Jrvc-hody diagram.)
4. Resolve each force into its components. A convenient representa-
tion isto write each force vector in the form
A = .Ui + .IJ + .l,k
5. Write the ('((uations of ('(|uilihriuin
i:/'', = 17'', = IF. =

(In a piano pr()l)lom one of the al)o\-p eriuntions is trivial.)
Sec. 2-6] Procedure of solving problems in sialics of a parlicle 25
6. Solve for the unknown magnitudes, direction angles, or scalar

components.
7. If the problem requires more than the analysis of one free body
(such as the example of Sec. 2-4), repeat the above procedure for the next
free body system is solved.
until the entire
The problems end of this chapter are examples of a great variety
at the
of problems that can be completely sohed bj- following through this
procedure.
PROBLEMS
2-1. .\. "particlp" at .1 is .suspended by two strings, .1/^ and AC. If the particle
weighs 100 lb, find the force in each of the strings.
Fi
tlOOIb
Fu;. V 2-1 Fig. P 2-2
2-2. Three coplanar forces F,, Fa,and F3 are acting on a particle in equilil)rinm.
If Fiand a are known quantities, determine F2, F3, and such that F3 is a niininium.
2-3. A simple structure consists of a bar AB and an ideal string CB pin-connected
at their ends as shown. This structure carries a vertical force of 1,000 lb at end B.
Find the forces in each of the members.
1 1000 lb
Fig. P 2-3 Fio. P 2-4
2-4. A small bead of weight W can slide on a vertical circular wire hoop of radius a.
The bead is attached to point yl by a and spring con-
linear spring of initial length a
stant k. The coefficient of static friction between the bead and the wire is pi,. Find
the position of equilibrium. By spring con.stant A- we mean that the force in the spring
is given as F = k8, where 5 is the change in length of the spring.
2-5. Verify the results of Example 2-1 by the use of all three of the equations of
equilibrium for the particle at A.
2-6. A weight W
is supported by the structure shown consisting of three hinged
bars AB, AC, and AD. Find the reactions in each of the bars.
Qw
Fig. P 2-6 Fig. P 2-7
2-7. A truss in the shape of a tetrahedron and consisting of six bars of equal length
is loaded by a vertical force P as shown, with the base triangle horizontal. Find the
reactions in all bars.
2-8. A
simple structure consists of three hinged bars AB, AC, and AD
oi equal
length 2a. Find the reaction in the bars due to a vertical load W
applied at the
joint A.
Fui. P 2-8
Problems 27
2-9. In the truss shown, A BCD and A'B'C'D' are two squares witli sides of lengths
(I and Ad, resiiectively. The vertical distance between the two squares is 2«. Find
the forces in the bars .1/^, H(\ HIV, C(", Cli', CD' due to tiie four vertical loads P.
Fic. V 2-9
2-10. Twelve bars are hinged togetlier to form a hexagon. One of the radial bars
isecjuipped with a turnbuckle so that the tension in the bar can be set at any value,
say T. Find the tension (or compression) in all bars.
Fui. r 2-10
2-11. Tworollers A and li are hinged at their centers to a rigid weightless bar .1/^
The rollerscan move without friction along the two inclined wails as shown. If the
weight of .1 is 10 lb and the weight of li is 40 lb, find the angle a between AB and the
iiorizoiital for the system to be in equilibrium.
Fig. P 2-11
2-12. A circular cylinder A rests on top of two half circular cylinders B and C, all
having the same radius r. TF and that of B and C is W/2 each.
The weight of A is
Assume that the coefficient of static friction between the flat surfaces of the half
cylinders and the horizontal table top is /xs- Determine the maximum distance d
between the centers of the half cylinders to maintain equilibrium.
V ////////// ^/ 77/77
Fig. P 2-12
2-13. What horizontal force P must be applied to the weightless wedge in order to
raise the 1,000-lb weight? Assume that the coefficient of static friction between all
contacting surfaces is 0.1.
Fic. I* •2-\:] Vui. V 2-14

2-14. A small l)l(ick on a plane making an angle of 'M)° with the liorizontal.
.1 rests
There is a cocflicieiit of static friction between th(> block and tlio planeof 0.4. Block .1
is connected to a bUxrk H l)y an ideal, perfectly flexible string, i)!issing (>\'er a frictiim-
less pulley. Determine what is necessary to maintain eiiuilihrium if

(a) \Va = 1,000 11), Wu = 200 Uk
{h) Wa = 1,000 1b, Wi, = COO 11..
(c) Wa = 1,000 11), Wb = 1,000 11..

CHAPTER 3
Vector Products
3-1. Introduction
The addition and subtraction of vectors defined in Chap. 1 appear to

us both natural and obvious. Our daily experience, for example, tells us
that a velocity due east plus a velocity of eciual magnitude due north
results in a velocity due northeast. This is indeed the basis of our defini-
tion of addition of vectors, considering in this case velocities as vectors.
The multiplication of vectors, on the other hand, is not something that
comes to us as obvious. " Multiplying " a velocity due east by a velocity
due north, for example, does not make sense to anyone, including mathe-
maticians. In fact, even the word "multiplying" is without meaning
here, since we have not yet defined such an operation for vectors. And
yet, for many physical problems that we shall presently examine, the
product of two vectors does have a real significance, if the product is
defined in a special way that almost appears peculiar at first. As a matter
of fact, there are two such products that are in common use, each defined
differently. If we were to consider, for a moment, two vectors and their
orthogonal components, we could immediately conceive of a great number
of linearcombinations of the nine (juadratic products {AjcBi, A^By, etc.)
of the respectivecomponents. Of all these possible combinations, two
combinations prove to be particularly useful. f They are the scalar (or
dot) product and the vector (or cross) product.
3-2. Scalar Product
Definition. The scalar product of two vectors A and B (written A B, •
hence also the term dot product) is a scalar algebraic number having the
t Not only are vector products of importance in mechanics, hut they have great
significance in other physical sciences such as optics, electricity, and magnetism.
29
30 Vector products
value of AB cos 6, where 6 is the smaller enclosed angle between the

positive directions of the two vectors, both vectors having a common
origin (Vig. 3-1).
By this definition, and because the magni-
tude of a vector is always positive, we see
that the scalar product of two nonzero vectors
can be positive, zero, or negative, depending
on whether the angle 6 is less than, equal to,
or larger than 90°. In particular, Ave see
that A A = A\•
Applying the above statements to the

orthogonal unit vectors i, j, k, we find that
ij=jk = k.i =
Ji=jj =kk=l (3-i;
Fk!. 3-1
Since, as illustrated in Fig. 3-2, A cos 6 is
the scalar component of A projected along B and B cos 6 is the scalar

componeni of li projot^ted along A, the scalar product may be interpreted
Fkj. 3-2
gcoiiicl ric.'illy as
A •
it .|(/) cos 0) . 1 (coiiipoiiciit (if It oil A)
or A •
It li{.\ COS 0) - -
/>(C0111|)011("llt ol' A (til It)
Plxprcsscd in words, this result may bestatiHl as follows: Tlu^ scalar prod-
uct of two \(>clors is \\\v pro(hict of the inngnitudc of I he lirst times the
magnifiidc of the second project(Ml on the lirst, or \i('(> \-ersa.
:
Sec. 3-2] Scalar prodiicl 31
Fig. 3-;i
Again from the definition, or from the geometric interpretation, we see

that the scaUir product is
Commutative: A •B = B A •
Associative: mX • ;jB= mn\ B •

(3-2)
Distributive: C • (A -f- B) = C A + C B • •
The validity of the last statement can be demonstrated as follows.

Using the geometrical interpretation of scalar products, we can write that
C • (A + B) = C[component of (A + B) on C]
Since the component of (A + B) is equal to the sum of the components of

A and B (Fig. 3-3),
C (A + B) = C(,component of A
• on C + compt)neiit of B on C)
= C A + C B • •
An important consequence of all this is the following Let
A = .l,i + -4 J + -l.k
B = BA+ BJ + BM
Then. A- B = (.4,i -H-4 J
-H-4,k) -{BA -hBJ +B.-k) c
Since the dot product obeys the dis-
tributive law, we can expand the
right-hand side and obtain, with the
help of Eqs. (3-1).
Fin. .-^4
A B= • .4 ,B, -\- A ,B, -h .4 ,/^, (3-3)
Hence the scalar product of two rectors is equal to the sum of the products of
their respective orthogonal components. In particular, we see that
A-A = .4,24.^4^24.^4^2 = ^42,
Example 3-1. Let C = A B -|- (Fig. 3-4). Then
C C = (A
•
-h B) • (A + B) = A • A + 2A • B -j- B B
•
or C- = A^ -\- 2AB cos 6 -{- B-

the well-known cosine law in trigonometry.
A
32 Vector products
Example 3-2. Let
A = 3i - 2j + 6k
B = 4i + Oj - 3k (Fig. 3-5)
Then A B =
• 3 •
4 + (-2) + •
6 •
(-3) =
Component of A on B = —^— = —^
Component of B on A = —— ^ "^ ~7~
cos = AB = -^
-6
e = 99°52'
A B 35
Example 3-3. Find the shortest distance from a point a(3,l, — 1) to

the line joining the points 6(2,3,0) and c( — 1,2,4). Figure 3-6 is a view of
Fic. 3-5 Fig. 3-6
the plane determined by the three points. Let the directed line segments
be, ab, and ac be denoted by the vectors A, B, and C, respectively. Then
A - -3i - + j 4k
B - -i -F 2j + k
C = -4i +j + 5k
Tlic projection of 15 on A, th(> distanci^ b(\ is given as
= B- 3 - 2 -f 4
be
'^
\/<) + 1 +16 V26
and finally, the perpendicular distance from point a to line t)c is
(/•= = IP - {Inr- - (1 +4 -I- 1) ^^2(; = '^}2%

psAf AilWtMV imm Vector product
Sec. 3-3] 33
The reader should verify this result by conipatiuji; the distance ec in a

similar fashion.
3-3. Vector Product
Definition. The vector product of two vectors A and B (written A X B,

hence also the term cross product) is a vector having the direction per-
pendicular to the plane containing A and B and having the magnitude of
AB sin 6. Again, d is the smaller angle between the positive directions of
the two vectors drawn at a common origin. The sense of A X B is
obtained by the right-hand-screw ruleif the positive sense is along the

advancement of a right-hand screw
which is rotated with the same sense AAXB
as the rotation from the positive direc-
tions ofA to B through an angle less
than 180° (Fig. 3-7). Then the cross
product may be written in explicit
form as A X B = .4B sin n, where
n is a unit vector perpendicular to the
plane of A and B and having direction
as specified above.
According to this definition we see
that the vector product of two non- BxA
zero vectors can be a zero vector only Fio. :?-7
when d = 0°, i.e., when the two vectors

are parallel. If, on the other hand, A X B = 0. then there are three
distinct possibilities:
A = B A and B are parallel
In particular, it follows that A X A = 0.
Again, from the previous definition,
A X B (B X A)
Hence the vector product is not commutative, and the order of forming the
vector product is of great importance.
Also from the definition, it can be seen that the magnitude of A X JJ is
equal to the area of the parallelogram having A and B as its conterminous
sides (Fig. 3-7). This geometrical interpretation often proves to be
convenient.
t A more
convenient application of the right-hand-screw nile would be to let the
hand curl in the direction of the indicated rotation; then the thumb
fingers of the right
of the right hand points in the direction of the positive sen.se of the cross product.
34 Vector products
To prove that the vector product is distributive, i.e.,
CX(A + B)=CXA + CXB (3-4)
we draw any three vectors A, B, and form the sum A + B. For

C and
convenience in illustration, we take C Using the triangle
to be vertical.
formed by A, B, and A + B as a base, we construct a triangular prism
(Fig. 3-8a) with the parallel edges in the direction of C. Finally, through
Fig. 3-8
the origin 0, we construct a plane normal to C; this plane cuts through

the extension of the triangular prism, thereby forming another triangle
Oab. From the definition of the vector product or from its geometrical
interpretation,
|C X (A + B)| = C(Oh)
|C X A| = C{Oa)
|C X B| = C{ab)
Furthermore, the above three vector ijrodiicts lie in the i)lanc Oah, since
they are all perpendicular to C. Viewed in the plane Oab (Fig. o-Sb), we
find that the magnitude of the three vector products is proportional to the
lengths of the three sides of the triangle Oab and that the orientation of
these vector pi-odnds is iioniinl to Ihc corresponding sides of the triangle
Sec. 3-3] Vector product 35
Oab. The three vector products therefore form a closed triangle O'a'b'
(Fig. 3-8c) which is similar to the triangle Oah but rotated 90°. Hence
C X (A + B) = C X A + C: X B
and Eq. (3-4) is proved.
Applying the definition of cross product to the orthogonal miit vectors
i.j. k,
iXi=jX.j = kxk =
i X j = -j X i = k
(3-5)
j X k = -k X j = i
kXi= -iXk=j
Equations (3-5) lead us to the calculation of the cross product in terms
of their components:
A X B = (.4,i + .1,J + .l,k)
X (^xi + b;} + BM)

Expanding terms (made possible by
the distributive law) and making use
of Eqs. (3-5),
A X B = {AyB, - A,B,)x + (A,B,

- A,B,)j -[- (A^B, - A„B,)k (3-G)
This result can be written as a determinant for easier remembrance:
A X B = LI, .1,, .1, (3-7)

36 Vector products
ExaiTiple 3-5. Suppose we again solve Example 3-3 by employing the

results of this section. Referring to Fig. 3-6, we find that
1 i
k
A X B -3 4 -9i+j - 7k
-4 5
AB sin d A X B VT31
|A X B
Therefore d =B sin d =
A
= V 13L 26
The student should verify the fact that the same result would have beeu
obtained had we used the product A X C rather than A X B. This
result can be generalized if we let r represent a vector from a point P to
any arbitrary point along the line of action of a vector A. Then the
shortest distance from the point P to the line of action of A is given as
= X A[
=
d r sm
A
That r may have its terminal anywhere along the line of action of vector A
is amply demonstrated by Prob. 3-15.
Example 3-6. As a final example of vector products we propose to

determine the shortest distance between two nonintersecting lines in
space. The shortest distance between the two lines is the length of their
common perpendicular. Let the two vectors A and B, respectively, have
the same lines of action as the two lines in question. Then the vector
C = A X B depicts the line of action of their common normal. Further,
if we be any vector having its end points on both lines, the projection
let r
of r on the common normal C is the required shortest distance. The

shortest distance d can then be expressed as
d =
C
3-4. Physical Significance of Scalar Product; Work;

Line Integrals
In physics avc learn that the work done by a force acting on a particle
is defined as the produc^t of the ('omi)onent of the force in the (.lirectiou
of the displacement of the particle and the displacement. In Fig. 3-10,
V is a varial)le forcef acting on a particle P which travels along a spatial
t To help in visuuliziiifz; this oxaiiiple, wc iin}j;ht eoii.sider P as a bead stninfj; along

a curved wire C and pulled liy a variable force F. We can also l)e more {j;eneral and
ini;if!;iiie tlint exists not otdy aloiijz; a i);irl icul;ir path, hut also e\-ery\\here (dse in
!•'
:
>^,r. 3-1] Significance of scalar product; ivork: line integrals 37
lurved path C. Consider an infinitesimal displacement
dr = {d.v)i + (<///) j + {dz)k
The work done by F, moving through the infinitesimal displacement dr,
is defined as
d\V = Fidr) cos d
where 9 is the angle between F and dr. We see, however, that the Avork
d\V thus defined is just the scalar product (and conse(iuently a scalar
Fic. 3-10
fjuantitAM
dW = F . dr = Fx dx -f F„ dy -H F, dz
According to this definition, the work done by F may equally well be

considered as the product of the force and the component of displacement
in the direction of the force. The scalar quantity, the work done by the
force F, may be positive, negative, or zero, depending on the angle between
the force and the displacement. If the components of force and displace-
ment have the same work done is positive; if the components
sense, the
have an opposite sense, the work done is negative; whereas if the force
and displacement are mutually perpendicular, the work done is zero.
The total work, when P moves through a finite displacement from Pi to
space, so that F is a continuously varying function of the space coordinates. This

gives rise to the concept of a vector field in which a vector quantity exists at every point
and varies continuously from point to point. Typical examples arc the distribution
of velocities in a moving body of fluid or the forces on a charged particle in an electric
field.
r +
38 Vector products
P2 along a prescribed path, is therefore
W = r^ F 'dr =
jl^^'
F, dx +Fydy + F, dz (3-8)
Thus we note that the total work done by the force F while moving from
Pi to Po is simply the sum of the work done by each component of the
force
W = TF. + T7, + W, d
where Wx = \ F^dx etc.
J XI
In general, if F is any vector quantity of interest (not necessarily force),

the integral
^'F-rfr (3-9)
/.
is called the tine integral of F. The line integral is taken along a specified
path between the two points Pi and Po and is itself a scalar quantity.
Certainly, the line integral, in general, depends upon the path as well as
the end points of the path. A conservative force field is a particular type of
force field defined such that the work done does not depend on the path,
but rather depends onhj on the end points.
The determination of the total work done by a force F moving between
the points Pi andPo, or the evaluation of the associated line integral, may
be accomplished by the use of Eq. (3-8). Suppose that the path traveled
is expressed in parametric form, such that
X = x{s) y = y{s) z = z{s)
Assuming F is a continuous function of the space coordinates, we can

write Eq. (3-8) in the form
''
=
^'(''
I + ^"('' t ^•(•^)
I ds (3-10)
I
so that the line integral has been converted into an ordinary type of
integral.
Example 3-7. Determine the work done by the force F = y-i — x^j
as it travels along the parabolic path y = x- from Pi(0,0) to P2(2,4).
Letting x = s, we have dx = ds, dy = 2s ds, and
F = .sî - .s'^
Substituting these values into Ef|. (3-10),
^^' = jjô ^•'' - -''^ '^' = -'^

As an alternative procedure we can express I']f|. (3-8) in terms of one
variable in order to convert the line integral into an ordinary integral.
Sec. 3-5] Physical significance of vector prodiicl: ntoment 39
This may he achieved since the path is fixed, tlius establishing a definite
relationship between the variables. Choosing, for example, .r as our
variable, we have
W =
f^'
{X' -2x^)dx = -«5
The student should verify tliis result by selecting y as the one variable of
integration.
3-5. Physical Significance of Sector Product; Monient

One of the more important applications of the vector product in the
field of mechanics is its use in connection with the concept of the moment
FCQSd
: .
40 Vector products
the bolt, then Fig. 3-1 la is a view in a plane parallel to the xy plane.
At the start we assume that the force F has no z component. Let P

shall
be the point of application of the force F and let I be the shortest distance
from the point P to the z axis (l is roughly the length of the wrench handle)
It is common experience that the effectiveness of turning the bolt depends
on the magnitude of / and the force component F sin 6, which is perpen-
dicular to both I and the z axis. In physics the measure of the turning
effect of a force about an axis is called the moment of the force about the
axis. In the present example the moment of F about the z axis is
M, = IF sin 6 (3-11)
where the subscript z denotes the axis about which the moment is taken.
It can be seen in Fig. 3-1 la that sin 6 represents the perpendicular
/
distance between the z axis and the F vector (which is in a plane normal
its extensions; this distance is called the moment arm of F.
to the z axis) or
Thus we may consider the moment of a force about an axis as the product
of the force and its moment arm. Furthermore, the value of M^ given
by Eq. (3-11) can also be considered as the sum of the moments of the
two components of F, namely, Fx and Fy. In other words, we can also
write
M, = xFy - ijFx (3-12)
where x and y, the components of the distance I, are the moment arms of
Fy and Fx about the z axis, respectively (Fig. 3-116). That the two
expressions for M^ as given by Eqs. (3-11) and (3-12) are identical can
readily be proved as follows
xFy - yFx = (I cos (t))F sin (0 4- e) - (/ sin 4>)F cos (0 + ^) = IF sin 6
Let us now, however, examine the problem in a more general manner.

For this purpose we draw Fig. 3-1 la once more, but this time in a three-
dimensional view (Fig. 3-12). The applied force F may in general have
a z component, but we see that F^ does not have any tendency to rotate
about the z axis and does not, therefore, contribute to the moment about
the Hence the moment of F about the z axis remains the same as
z axis.
that given in Eq. (3-12). However, Fx and F^ also contribute to a rota-

tion about the axis. Similarly, F„ and /^z contribute to a rotation about
,//
the X axis. Analogous to Eq. (3-12), we can form the moments about the
x and the y axes as follows:
Mx = yFz - zFy
My = zFx - xF, (3-13)
M, = xF„ - yFx
The last of Eqs. (3-13) is merely a repetition of E(|. (3-12). In these
expressions for niomenis, W(> ha\-e used llic lii^lil-lumd-scrcw rule: The
1
S.H'. 3-5] Physical signijlcarwe of rector product: momeril 1
moment is considered positive if its direction of rotation is that of a right-

hand screw which is adx'ancing at the same time along the positive direc-
tion of the axis of rotation.
Let us again define a moment vector M of a force F about a point
such that the three components of M are the moments of F about the
Fk;. 3-12
three axes through 0, i.e.,
M - MA + M,j + MM
Substituting Eqs. (3-18) into the above expression, we find
i J k
M
42 Vector products
we see that M ( = r X F) is along a direction perpendicular to the plane

of r and F or perpendicular to the plane containing and F, i.e., the
plane of Fig. 3-13. Furthermore, the magnitude of is, by definition, M
|M| = |r X F| = rF sin 9. But since r sin 6 is just the moment arm of F
about [or to the line of action of F
the shortest distance of the point
(Example moment of F about
3-5)], the is numerically equal to the
product of F and its moment arm. We see that this is in complete agree-
ment with our usual concept of "moment" as given in introductory
Fig. 3-13 Fig. 3-14
physics, t We also see that the axis of rotation due to the action of F
about the point is just the direction of the newly defined INI vector.
Once we have determined the moment of a force about a point, it is an

easy matter to determine the component of that moment vector in the
direction of any line passing through the point. In Fig. 3-14, let M
represent the moment of the force F about the point and let L be some
line passing through 0. Then the line L can be described by a unit vector
n, and according to Sec. 3.2, the component of ]M on the lineL is a scalar
quantity given as
ilf L = n . INI = n . (r X F) (3-1 3a)
In particular, if we let the line L coincide, respectively, with the .r, y, and
z axes, we determine immediately that the components of projected in M
these directions arc the same as those given by Eqs. (3-13). For example,
il/, = i . M = !jp\ - zF,
Thus we have that the moments of a force about the three orthogonal
axes passing through a point are simply the scalar components of the
t See, for example, F. W. Sears and I\I. \\'. Zemansky, "College Physics," p. 52,
Acldison-We.sley PiiMishinf!; Company, Readiiifj;, Mass., l'.>52.
Sec-. 3-6] MonienI of a vector 43
moment vector at
point. In many practical applications the scalar
tliat
methods [Eqs. nuiy offer a more direct solution than the more
(.3-1^^)1
fioneral vector methods [Eq. (3-15)].

We can now summarize our results as follows:
1.Moment about a Point. CJiven a force F acting at a point P(x,y,z);
its moment about another point is a vector, denoted l)y M and expressed ,
as
M = r X F
where r is the position vector of P relative to 0. The direction of M is the
axis al)out which there is a tendency for rotation to occur because of the
action of F, and the magnitude of M is a measure of the effectiveness of
turning about that axis.
2. Moment about an Axis. The component of the above-defined M
along any axis through is the moment of F about that axis passing
through [FCq. (8-1 3a)]. The sense of the moment is determined l)y tiie
right-hand-screw rule.
Example 3-8. A force vector F = i -|- 3j — k acts at a point
P{S, —I, 4). Find the moment of F about an axis passing through the
point Q(l,l,2) and having the direction of the \ector A = 4i -f- 3j -\- Ok.
Since the vector r from Q to P is 2i — 2j + 2k, the moment of F about
point (J is
i
J k
= X F - -2 = -4i
I
AI r 2 2 -f 4j + 8k
I
1 3-1
Hence the moment of F about the axis thiough (J along A is
M A ^ 44
A ~ V<iT
3-6. Moment of a Vector
We ha\e in the foregoing defined the

moment of a force F. In general, we
can define the moment of any vector
(juantity, not necessarily force, in a
similar way. Consider a vector A act-
ing at a point P{x,y,z) (Fig. 3-15). The
moment of A al)out another point, say
point 0, is defined as
Fio. :i-i5
M = r X A (3-1(3)
where r is the position xcctor of /' rc^lative to 0.

The moment of a vect(H- about a point is a constant with regard to any
sliding of the vector along its line of action. This can be easily shown by
44 Vector products
the property of a vector product. Any sliding of the vector A (Fig. 3-15)
would result in a change of the position vector from r to r + A;A. The
resulting moment is, by virtue of the fact that A X A = 0,
(r + AA) X A = r X A
Hence the moment of a vector is unchanged by sliding the vector along
its own line of action. When the vector in question is a force, this con-
cept sometimes referred to as the transmissihility of a force. It also
is
verifies a statement made previously that the position vector r can be

taken from the point to any point on the line of action of the vector and
the moment of the vector will remain unchanged.
Fig. 3-16 Fig. 3-17
On the other hand, the moment is seriousl}^ altered by a change of line

of action of the vector. Thus, the moment of a vector is a bound vector
in the sense that it is associated with the particular point about which the
moment is taken and its value changes for different moment centers.
Suppose we have two equal and opposite vectors with a common line of
action (Fig. 3-16).Then, from Eq. (3-16), the moment of these vectors
about point is given as
M= ri X A + r.> X (-A)
Because the moment is~ unchanged by a sliding of the vectors, this can
also be written as
M ri X A +.r, X (-A) = ri X (A A) =
Hence the vector sum of moments about an}' point of two equal, opposite,
and colinear vectors is equal to zero. This statement, which is almost
intuitive in nature, assumes importance when we consider the forces of
interaction between particles, which, because of Newton's third law, are
equal, opposite, and colinear.
Now suppose we have a system of vectors acting at a point P (Fig. 3-1 7).
Si'c. 3-7] Vehcily due lo rotation: angular-velocity vector 45
The moment of the.se vectors alnnit point is
M = r X F, + r X F, + + r X F„
and because of the distril)Uti\e property of vector cross products,
.AI = r X (Fi + F, + • • •
+ FJ
M = r X F,
where Fr is the sum Thus we have

of the vectors acting at point P.
I'arignon's theorem, which states that the moment
system of vectors of a
at a point is equal to the moment of the resultant of the system of vectors.
It is easy to show that the converse of this statement is also true.
Many times in applications it becomes convenient to use these general

properties in order to find the moment of a force about some point; that
is to say, we can decompose the force into any convenient set of compo-
nents, and moreover, we can consider these components to act at any

convenient position along the line of action of the force.
3-7. Velocity Due to Rotation; Angular- velocity Vector f
Another example of the physical application of vector products is the

velocity of a point on a rotating body. Let P be a point in a rigid body
Fic;. 3-1
which is rotating with an angular velocity of magnitude w about a fixed

axis L We can represent the angular velocity by a vector <<>
(Fig. 3-18).
with magnitude equal to w and direction oriented along the axis of rota-
tion L. (The sense of o> is determined by the right-hand-screw rule.)
Under these circumstances the point /^ travels in a circular path of radius
a, the plane of which is perpendicular to the axis of rotation. Recalling
from physics (or e(iually well from our physical experience), the velocity
V of point/'* has magnitude proportional to the product of w and a and has
direction tangent to and lying in the plane of the circular path, and sense
t What we present here in an intuitive fa.shion will he {hs(Mis.se(l more completely

in Sec. 7-2.
46 Vector products
as indicated by the direction of rotation. More exactly, we can write

V = coa.
Let us select any point on g> and let the position vector of P relative to
be represented by r. Then the velocity v of P due to rotation is given as
V = to X r (3-17)
In order to show that this definition agrees with our previously described
concept of velocity due to rotation, we observe that to X r is a vector having
direction perpendicular to r and w and a sense as determined by the right-
hand-screw rule and therefore has the same direction and sense as
previously described. Moreover, we have that
|v| = |(o X r| = cor sin 6 = coa
Thus the two vectors defining the velocity of point P have the same
magnitude and direction and by definition are equal. Equation (3-17)
isthereby verified.
According to Sec. 3-6, we can interpret the velocity v due to rotation
about a fixed line as the negative of the moment of to about the point P.
From this it follows that the location of point along w is immaterial, for
a new location 0' along o> merely changes r to r -1- ^w and therefore does
not modify v.
3-8. Triple Vector Products
Products involving three vectors can be formed according to each of

the following three types: (A B)C, A (B X C), and (A X B) X C.
• •
The first type, (A B)C, is merely the product of a scalar (A B) and a

• •
vector C. The product is therefore a vector along C with the magnitude

of ABC cos 6, where 6 is the angle between A and B.
The second type, A • (B X C), is the scalar product of A and (B X C).
Since the product is a scalar, it is called the scalar triple product. Refer-
ring to Eqs. (3-7) and (3-3), we readily see that
A • (B X C) = A
Sec. 3-8] Triple vector products 47
111 particular, since the dot product of two vectors is coininutatixe, we

see from Va\. (3-19) tliat the dot and the cross in a scalar triple product can
be interchanged.
The geometrical (B X C)
interpretation of the scalar triple product A •
is as follows. As shown
magnitude of (IJ X C) is equal to
in Sec. o-o. the
the area of the parallelogram having B and C as its conterminous sides
and the direction of (B X C) is perpendicular to the plane of B and C (Fig.
3-19). Since A (B X C) has a value ecjual to the product of the
•
magnitude of (B X C) and the projection of \ on (B X C), it therefore is

numerically eciual to the volume of the parallelepiped having A. B. and C
as its conterminous edges. The
sign of the product is positive or
negative according to whether A
and (B X C) lie on the same side of
the plane of B and C or not. Since
B (C X A) and C (A X B) also
• •
represent the volume of the same I
parallelepiped, the truth of Eqs.

(3-19) and (3-20) is again verified.
From this geometrical interpreta-
tion we deduce the following state-
ments: If the three vectors A, B, and Fig. 3-19
C are coplanar, the volume of the
parallelepiped and consequently the scalar triple product vanish. More-
over, if any two of the three vectors are equal, again the volume and the
scalar triple product vanish.
The' third type of triple product, or
A X (B X C)
is a vector perpendicular to A and is a (B X C). Since the product

vector, it is C) is perpendicular
called the vector triple product. But (B X
to B and C. Therefore A X (B X C) lies in the plane of B and C and
can thus be expres.sed as a linear combination of B and C. This means
A X (B X C) = aB -f /3C
where a and /? numbers to be determined as follows.

are scalar
Let us select, with complete generality, a coordinate system such that
its .r axis is along B, the /y axis is perpendicular to B in the plane of B and
C, and the z axis is perpendicular to the plane of B and C Then
A = AA -t- Ayj + AM
B = BA
C = Cxi -H C'J
48 Vector products
Hence B X C = 5.C^k
and A X (B X C) = A,B£yi - A^.Cj

The right-hand side of the above equation, however, is equal to
(A . C)B - (A B)C •
since A C =
• A^Cx + AyCy
and A B =
• A^B,
Hence A X (B X C) = (A • C)B - (A B)C (3-21)
The form of the above expressions shows clearly that the vector triple
product does not remain the same by any interchange of its elements.
Example 3-9. We have already seen evidence of the utility of triple
vector products where in Sec. 3-5 we used the scalar triple product to
b
Fig. 3-20
define the component of a moment vector along the direction of some line
L. As a finalexample of triple vector products we shall consider two
geometric problems.
First, let us determine the piercing point (point of intersection) of a
lineand a plane. Let the line ah intersect the plane at point P (Fig. 3-20).
Suppose we choose any three points, c, d, and e, lying in the plane.
Denote the directed line segments ab, aP, ac, cP, ce, and cd by the vectors
A, B, C, P, E, and D, respectively. Then, because the three vectors P,
E, and D are coplanar, we have
P . (E X D) =
Further, we can write that
P = B - C = ;//A - C
where m is some scalai' luiinlxM- (o be determined. Substituting this into
the previous expression,
(;/;A - C) . (E X n) =
Prohlenis 19
Upon solving this vector e(iuation,
_ C • (E X D)
"" ~ A . (E' X D)
which can readily be evaluated according to Eq. (3-18). Once the value
of m is known. P can he readily determined and in turn the point of inter-
section P can be located.
As a second illustration, determine the vector projection of vector A on
some plane defined by its normal n (Prob. 3-10). The vector projection
of A on the normal n is, from Sec. 3-2,
(n . A)n
Now. considering the triangle formed
by the vector A, its vector projection on
the normal, and its vector projection B
on the plane itself (Fig. 3-21), we im-
mediately^ determine the vector projec-
tion on the plane as
B = A - (n- A)n
Using Eq. (3-21), this becomes
B = n X (A X n) Fig. 3-21
PROBLEMS
If d is the angle between two vectors A and B, derive an expression for ros 6
3-1.
in terms of the components of A and B.
3-2. Two line segments OP and OQ are formed by joining the origin O with the
points P(3, —5,2) and Q{ — 1,2,0), respectively. Find (a) the projection of OP on OQ,
(6) the angle between OP and OQ, (c) the unit vector perpendicular to the plane of
OP and OQ.
3-3. Given the force vector F whose components are
Fx = 3 lb Fy = -3 lb F^ = 1 lb
Find the component of F along the direction of the line x — z = y = 2z.

(a)
{b) F is acting at a point of coordinates ( —2 ft, 1 ft, 2 ft), find the moment of F
If
about the origin, i.e., the magnitude of the moment vector and its three components.
3-4. Show that (A + B) (C + I)) = A C -|- A D -|- B C 4- B D.
• • • • •
3-5. Prove that any two lines drawn from each extremity of the diameter of a circle
to a point on the circumference are mutually perpendicular.
3-6. Find the angle between the body diagonal and one edge of a unit cube, both
intersecting at a point.
3-7. Find the shortest distance from the line joining points a(3,l, — 1) and />(2,3,0)
to the origin O.
3-8. Show that A X A = by expanding the cross product in terms of its
components.
3-9. Using the method of Sec. 3-3, prove that the vector product is distributive for
the special case of three coplanar vectors.
50 Vector products
3-10. Given a plane whose cartesian equation is Ax + By + Cz +D = 0, show

that the normal n to this plane can be represented as
n = -4i + Bj + Ck
Hint: A normal to a plane is perpendicular to any two nonparallel lines in the plane.
3-11. IfC is normal to the vectors A and B, show that C is normal to A + B and
A — B. Whsbt does this result indicate about lines perpendicular to planes?
3-12. A force vector field is represented by
F = (if- + z)i + 2xyj + xk
where F is the force acting on a particle located at point {x,y,z). Find the amount of
work done on the particle when the particle is moved along a straight line from the
origin to the point (2,1, —2).
3-13. A force vector F = 3i + j — k acts on the point (3,-2,3). Find the
moment of F about an axis through the origin and having the direction of the line
x-2,=2y + l=z.
3-14. A bar OP is 10 units long. a vertical wall through two guy
It is fastened to
wires at S and P and a ball joint at 0. The load W
is attached to the bar at T.
Assuming that the reactions in the two guy wires at S and P are Fi and F2, respec-
tively, find the moments due to the forces about the ball joint 0; i.e., find (a) moment
Ml due to Fi in terms of F,, {h) moment M2 due to Fo in terms of Ft, (c) moment M3
due to W.
Fig. P 3-14
Prohlems 51
3-13. Prove tliat the cross product of a vector A and any other vector that tcrini-
iiates on a hue jiarallel to A is constant; tliat is, show that A X H = A X C. This
ri'sult will lie extrenielv useful to us.
Fi(,. P :M5
3-16. A body rotates about an axis x = // = r with a constant angular speed
riîd
of 1,000 rpm. Find the velocity and its components along x, y, and z axes of a parti-
cle at (1 in., —1 in., 3 in.) on the body.
3-17. Prove (A X B) (C X D) = (A C)(B D) - (A n)(B C).
• • • • •
3-18. Find the shortest distance between the two nonintersecting lines, one of
which is the body diagonal and tiie other of which is a face diagonal of a unit cube.
3-19. The force field of Prob. 3-12 acts on a i)article as it niak<'s one circuit around
a square in the xi/ plane of sides of length a. Find the work done on the particle.
Is the force field conservative?
CHAPTER 4
Statics of Systems of Particles

and of Rigid Bodies
4-1. Equilibrium of a System of Particles
In Chap. 2, we learned that the equation of equihbrium of a particle is
F, = 2F = (2-4)
where Fr = 2F represents the sum, or resultant, of all forces acting on the

particle. When considering a system of particles, it is natural, then, to
define equilibrium as meaning that each and every particle comprising
that system is in a state of static equilibrium, or more explicitly that
each and every particle is in a state either of rest or of uniform motion.
Let us now consider a system that consists of a number of particles.
The total force acting on any one of the particles can be considered as the
sum of a total external force F and a total internal force f acting on that
particle. The external forces are exerted by mediums that are outside of
the system under consideration, while the internal forces are the mutual
actions and reactions between particles within the system. If this
system of particles is in static equilibrium, then each and exexy particle
is in efiuilihriuin, and thus for a typical, say the iih., particle, the equa-
tion of eciuilibrium is
F, + f, = (4-1)
where the subscript i denotes the particle under consideration. Natu-

rally,we can write one such equation for each particle. Hence, if there
are a total number of n particles, we can write n equations by letting
i = \,2, . n, successively.
. . , We now add all eciuations and obtain /(
n H n
^ (F, + f,) = 2 F, -}-

2 f, = (4-2)
Sec. 1-1] E(fnHi})riuni of a system of purl ides 53
Xow l)vXewton's third law of action and reaction, ail the internal forces
occur in e<iualand opposite pains, and thus we see that all the internal
forces must atld up to zero. Hence Ecj. (4-2) reduces to
2 F, =
i= 1
Or, introducing the notation F, = ^F = ^ F, for brevity, the above
ccjuation may be written as

F, = 2F = (4-3)
where F^ = I'F is the sum, or resultant, of all the external forces acting on
the system.
We now consider the moments of the forces. Let be any fixed point
in space and let the position \ector from to the ^th particle be r,. From
Eqs. (4-1) and (3-4), we obtain
r, X (F, + f,) - r, X F, + r, X f, =
We recognize that each of the two terms in the ai)o\c equation is a
moment (cf. Sec. 3-5). Adding n such equations for all /; particles, we
obtain
n n
2 r,- X F, + ^ r, X f,- = (4-4)

1=1 1=1
Xow the internal forces f^ always occur in equal, opposite, and colinear
pairs. Let us, for call them fi2 and fî,
example, pick one such pair and
so that fi2 is the force acting on particle
due to particle 2 and f-ii is the 1
force acting on particle 2 due to particle L Then, from Xewton's third

law,
and ri X fi2 = —r-> X fsi
The last relation is true because fi2 and £21 are colinear and act along the
line joining the two particles, although ri 7^ Vo (Sec. 3-G). The total
moment about (), of all the internal forces, is the sum of all such pairs of
moments and is therefore zero. Hence
y r, X f ,
= (4-5)
and P>|. (4-4) reduces to
^ r. X F. = (4-6)
54 Statics of systems of particles and of rigid bodies
Upon introducing the notation M^ = y -NI = ^ r, X Fi for brevity, the

i = l
above equation may be written as
TVI,. = 2M = (4-7)
where Mr = 2M the sum, or resultant, of the external moments,

is i.e.,
the moments of the external forces, about anj^ point 0.

To summarize, if ever}^ individual particle in a s^^stem of particles is in
equilibrium, we say that the system is in equilibrium and that the sum
of the external forces is zero and the sum of the external moments about
any point is zero; i.e.,
2F = 2IM = (4-8)
We can also deal "oith the components of F and M; the above two
vector equations are equivalent to the following six scalar equations:
2F, = Sil/,= Z(^F, - sF,) =

i:Fy = ^My = -^{zF, - xF,) = (4-9)
2F, = 23/, = Z{xFy - yF.) =
It must be recognized that Eqs. (4-8) or (4-9) are the necessary but not
sufficient conditions for the equilibrium of all particles in a system in
general, for they do not ensure that (Fi -\- f,) = and r, X (F,- -\- f,) =
for each particle individualh'. For example, consider a system of two
particles acted upon by the colinear forces Fi, — F2 and — Fi, Fo. respec-
tivel^^ If Fi ^ Fo, each individual particle is not in equilibrium,
whereas considering the system of both particles, Eqs. (4-8) and (4-9) are
both satisfied. It will be shown in Sec. 4-3 that these equations are both
necessary and sufficient for the equilibrium of a rigid body.
4-2. Constraint of a Rigid Body in Space
A rigid body is a system of particles having the property that the

mutual distances between all particles remain constant. The most
general displacement of a rigid body can always be described by a transla-
tion of some point in the body and a rotation about some axis passing
through the point. We state here this general result without actually
proving it in a rigorous fashion, f Suppose, then, that a rigid body is
displaced from some initial position in space to some new configuration.
We describe this displacement by specifying the spatial location of a point
in the body (requiring three coordinates) and then by specifying the
rotation of a line passing through the point. The line itself requires two
t A more complete discussion of this point will lie found in Sec. 7-1.
sec. 1-2] Constraint of a rigid Ixxiy in space 55
independent direction cosines for its definition; one additional quantity is

needed to define the amount of rotation about that fine. Therefore a
rigid body requires six independent quantities to describe a generalized
displacement and thus possesses six degrees of freedom.
Now let us examine the minimum number of constraints that are
necessary to completely restrain a rigid body in space. In order to
accomplish this we must completely prevent an}- motion associated with
the six degrees of freedom.
//////^ ////yf
\y
TTT^^^TTTT TTTTTTT^T-
FiG. 4-1
In Fig. 4-1, let us fix any -4, of a rigid body to the

point, say point
foundation b}' This t^'pe of joint offers force
a frictionle.ss ball joint.
along all three mutually perpendicular directions it therefore fixes .4 but
;
still allows the body to rotate about any line through .4. The ball joint
at .4 is termed a reaction and accommodates anj' force necessarj- to
maintain point .4 in a fixed position. We next pick a second point B,

and we now need only to attach B to the foundation by a frictionless ring
guide, a device that does not exert a force along the axis of the guide, in
this case the direction AB. Now the body can still move, but only as a
rotation about the line AB. We finally pick a third point C and fix with it
a frictionless hinged bar which offers a force onl^' along one direction,
namely, along the axis of the bar. If the axis of the bar at C does not
intersect AB. rotation about AB is prevented and the body is completely
constrained. can be replaced by three hinged

It is clear that a ball joint
bars not all in the same plane and that a
ring guide can be replaced by
two hinged bars in a plane normal to the axis of the ring guide but not
both along the same direction. Therefore a rigid body can be completely
constrained by six hinged bars or their equivalents.
It is fairly obvious, however, that six arbitrary hinged bars do not
always completely constrain a rigid body. For example, if the six bars
(or their equivalents) are such that (1) they can be intersected by one
straight line, (2) they all lie in one plane or parallel planes, or (3) four or
more bars intersect in one point, then the body is not completely con-
strained and is termed unstable.
A rigid body that is completely constrained by six hinged bars (or
their equivalents) is said to be statically determinate.Only under such
conditions can the reactions in the supports be completely determined by
the laws of statics. Any additional support is called a redundant support
and would make the problem statically indeterminate, the solution of which
requires additional information based on the deformation properties of the
body.
4-3. Equilibrium of a Rigid Body in Space
In Sec. 4-1 it was proved that a system of particles in equilibrium

necessarily satisfies the following equations:
SF = SM = (4-8)
where 2F = sum of external forces

SM = sum of external moments, i.e., moments of external forces
about any point
Since a rigid body is a special case of a system of particles, a rigid body
under equilibrium also satisfies Eqs. (4-8).
It remains to be proved, however, that if Eqs. (4-8) are satisfied, the
rigid body is always in equilibrium. Let us assume that the external
forces acting on the rigid body satisfy both Eqs. (4-8) but that the body is
not in equili))rium. The body, however, can always be put in equilibrium
by introducing the kind of supports shown in Fig. 4-1, namelj^ a ball
joint at A, a ring guide at B, and a hinged bar at C such that it becomes
fully constrained. Let us represent the reactions of the supports at
.1, and C by A, B, and C, respectively; then
B, for the coordinate system
as shown in Fig. 4-1 we have
A = A A + .1,J + .l,k
B = Byj+ BM
C= C,j
S(H'. 1-1] Procedures of solving problems in sialics of riqid bodies 57
We assume that B has no x component and C has no x and z components

because these components are not needed for the complete constraint of
the body. The body is now in e(iuiUl)rium upon introduction of A, B,
and C.
Let us denote the original system of forces and moments (taken about
some point, say .1) by 22F and i^M and the additional system of forces
and moments (again taken al)out point .1), reciuired to comj)letely con-
:;train the body, by ISF' and ^^\' Since the body is now in eiiuilibrium,
.
:^¥ + 2F' =
i:M + 2M' =
Because the original system satisfies Eqs. (4-8),
vF.= o = A-|-B-fC
vM' = = Cy{AD)x - B,{AB)i -f B„(.4B)k
These vector ecjuations reduce to six scalar equations whose only solutions
are
Ax = Ay = A^ — By = B; — Cy =
We leave it Therefore the imposed con-
to the student to verify this.
straints offerno resistance and the body is in eciuilibrium without any
reactions at the supports A, B, and C. The supports can therefore be
removed without any effect on the state of the body. This proves that
the Eqs. (4-8), or their equivalent Eqs. (4-9), are both necessary and
sufficient for the equilibrium of a rigid body.
It now becomes apparent that the scope of statics is not unduly
restricted by the fact that Newton's laws are applicable, in a narrow
sense, to particles alone. Equations (4-8) or (4-9) indicate that a body of
finite size in equilibrium may be treated as a particle provided 2M = 0,
which in turn may be achieved if the forces are all concurrent. The tacit
assumptions made in Chap. 2 are thereby justified, and the scope of
particle mechanics is thus increased from a mathematical idealization to a
much more accurate physical reality.
4-4. Procedures of Solving Problems in Statics

of Rigid Bodies
The procedure outlined in Sec. 2-0 for solving problems of the statics of
a particle or a joint applies etiually well to statics of rigid bodies in general.
Only one modification is recjuired. In the case of a particle or a joint,
the lines of action of all the forces intersect in one point (i.e., they are
concurrent), which is the particle or the joint in (juestion. With respect
to this point, none of the forces contributes any moment. The moment
equation is identically zero, and hence does not yield any information.
The problem of the statics of a particle or a joint is therefore solved by

making use of only the equations of equilibrium of the forces; i.e.,
Si^. J^l J,
iF. =
For a rigid body, on the other hand, the lines of action of the external
forces do not in general intersect at the same point. As can be seen
from the discussion of Sec. 4-2, there are (in response to any given load-
ing) six unknown reactions in the six supporting hinged bars located, for
example, at points A, B, and C of Fig. 4-1. To find these six reactions, we
have to make use of all six equations of equilibrium:
2F, = SF, = 2F. =

(4-9)
SM, = SM = „ SM, =
This modifies step 5 of the procedural outline of Sec. 2-6. The remaining
steps are still the same.
TTTTt^TTT"
(a) (6)
Fig. 4-2
In counting the number of unknowns in a problem, it is to be noted that

a reaction that is unknown both in its magnitude and in its direction is
counted as three unknowns: one unknown magnitude and two unknown

direction cosines. Or we may substitute such an unknown reaction by
its three rectangular components, which again constitute three unknowns.
Following this reasoning, we see that a reaction that lies in a given plane,
but is otherwise unknown in magnitude and in direction, is equivalent to
two unknowns. Further, a reaction of unknown magnitude but known
direction constitutes one unknown.
If we have a problem in which all the acting forces lie in one plane, say
the xy plane, our useful equations are reduced from the six of Eqs. (4-9) to
only three:
^ZV., = :^F, = ^M, = (4-10)
Hence, for every plane free-body problem, we are generally able to solve
for three unknowns. The complete constraint of a rigid body in a plane
by means of three hinged bars is shown in Fig. 4-2a. The two hinged
Sec. 4-1] Procedures of solving problems in sialics of rigid bodies 59
liars at A can be substituted by a frictionless hinge support, which is a

ring guide (Sec. 4-2) with its axis normal to the figure shown. The
hinged bar at B can be substituted by- a frictionless roller support. Thus
tiie complete constraint of a rigid body in a plane can also be accomplished
by a hinge and a roller support (Fig. 4-2/>).

There are other special situations when some of Eqs. (4-9) are satisfied
identically. If all the forces are parallel to a plane, say the .r// plane,
then the eciuation ZF, = is useless and the number of useful etjuations
(and unknowns) is reduced to five. Again, if all the forces are parallel to
some line, the number of useful independent eciuations is reduced to three,
whereas if all the forces intersect some line, there are only five independent
equations of equilibrium.
Although we have specified the form of the three scalar equations of
equilibrium [Eqs. (4-10)] for plane problems, it is not necessary that they
be u.sed explicitly. Actually, any set of three independent .scalar equa-

tions satisfying Eqs. (4-8), in conjunction with the plane problem, are
suitable. One set of suitable efjuations would be
ZM,' = 2.1/,« = SF, =

where the two moment equations are taken about z axes passing through
the two points .4 and B, respectively, and the s direction is any direction
in the .r^ plane not perpendicular to the line AB. Still another appropriate
set would be
SM,-^ = 23// = 2il/,^ =

where the moment equations are taken about z axes passing through the
points .4 B, and C, respectively, provided the three points do not lie on
,
the same straight line. We leave it to the reader to verify that these
alternative sets of equilibrium efiuations do indeed satisfy Ecjs. (4-8) in
any plane problem.
In a similar fashion, it is possible to enumerate alternative sets of six
equations of etiuilibrium, satisfying Ecjs. (4-8), for three-dimensional
equilibrium problems. However, to explicitly describe these equations
and their restrictions more involved; the number of different sets is
is
much more than the two alternatives of two-dimensional eciuilibrium.

Furthermore, there is not much advantage to be gained by doing this.
We simply remark, therefore, that any set of six equations of eciuilibrium
which .satisfies Eqs. (4-8) would be suitable. In any given problem
capable of solution, the suitability is adetiuately determined by the fact
that a .solution is attainable. It must be noted, of course, that it is possi-
ble to write mon; than six ('(luations of cfiuiiiWriuin, l)ut at most only six
(three for plane ])r<)i)lems) will be independent in the ca.se of three-
dimensional ('(luilibrium. The adxantagc; of the a\ailability of (he alter-
native sets of equations of equilibrium lies in the fact that now we have
the opportunity to select our equations such that as few unknowns as
possible are involved in any one equation. Thereby we materially reduce
the effort required in the solution' of simultaneous algebraic equations.
As an extreme, if we are able to select six equations of equilibrium such
that each equation contained only one unknown, no simultaneous solution
would be necessary.
4-5. Examples of Statics of Rigid Bodies

Example 4-1. If a rigid body is in equilibrium under two forces, the
two forces must be equal in magnitude, but opposite in sense, and their
lines of action must be colinear.
Fig. 4-3
Let the two forces be Fi and F2 (Fig. 4-3). By the equation

Proof.
2F = 0, find that Fi = — F2.
we Let us now take the moment about
any point, say A, on the line of action of one of the forces, say Fi. The
moment of Fi about A is zero. By the equation SM = 0, the moment of
F2 about A must also be zero. Hence A also lies on the line of action of
F2. The forces Fi and F2 are therefore colinear. This proves the state-
ment (made in the first paragraph of Sec. 2-4) that a weightless hinged
bar under no lateral loads between its two ends can transmit forces only
along its center line.
Example 4-2. If a rigid body is in equilibrium under the action of

three nonparallel forces, the three forces must lie in one plane (proof left
to the reader) and their lines of acti(m must intersect in one point.
Proof. Let us draw the plane of action of the three forces Fi, F2, and F3
(Fig. 4-4). Extend the lines of action of Fi and F2 so that they intersect
at a point 0. Writing 2SM = with as the moment center, we find
that, since F'l and F2 do not produce any moment about 0, the moment of
F;t about must also be zero. Hence the line of action of Fa also passes
through (). As a coi'ollary to this statement we may say that three non-
zero forces acting on a rigid body nvc in cfiuilihriuiii (iiil.\- if (he forces lie in
the same plane and are either conciiiirnt oi' paiallcl.
Sec. 1-3] Examples of sialics of rigid Ixtdies 61
Examnle t-3. Tlie simple plane st riictiire shown in V\^. \-'m supports
a load of 3.000 lb. If the weijrht of the bars is neglected, find (a) the
reactions at the supports .1 and H and {h\ the reactions at the pin joint C.
Assume all joints are frictioidess.
12501b
62501b
(a) (6) (c) id)
Fic. 4-5
Solution, (a) First we take the entire structure as our free body. The
free-body diagram is shown in Fig. 4-56. The unknown reactions at the
supports .4 and B are
A = .l,i + .IJ at .4
B = BA at B
Applying the e(|uations of eriuilibrium.
i:F. =
SM,-' =
It is noted that the reaction at D must be along ED because bar ED is

under the action of only two forces at its ends. Applying the equations
of equilibrium,
SF. = C. + HD =
ZFy = Cy + V5D - 3,000 =
2M/ = 9i%D) - 15 X 3,000 =
Solving these equations,
D = 6,250 lb C. = -3,750 lb Cy = -2,000 lb
The directions of Cx and Cy are therefore opposite to those shown in Fig.

4-5c.
1^
Fig. 4-6
(c) If we finally take bar AB as a free body, we can verify that all the
equations of equilibrium are satisfied. In the free-body diagram for AB

(Fig. 4-5d), Newton's third hiw on action and reaction is used in deciding
the sense of the newly found reactions at the joints C and E.
Example 4-4. A hp from the driving pulley A to
shaft transmits 50
the driven pulley B. The shaft is supported b.v bearings at C and D. It
rotates at 300 rpm. The diameter and the weight of the two pulleys are
30 in. and 80 lb for pulley A and 42 in. and 150 lb for pulley B. The belt
on pulley .1 is inclined ;^0° wilh the horizontal, as shown, while that on
Sec. 4-5] Examples of sialics of rujid bodies 63
pulley B is vertical. Assume that the belt tension on the tight side of the
pulley is three times that on the loose side. Find the reactions in the two
supporting bearings, assuming that bearing 1) alone prevents motion in the
X directi(m (Fig. 4-0).
To transmit 50 hp at 800 rpm requires a tortjuc of
,, 50 X 33,000 X 12 ,r. -,.,.. ,,

^^' = = '^'"^^ "'-^'^
M300)
Hence {1\ - r,)15 = 10,500
(T3 - T4)21 = 10,500
T T-
Assuming T^n-
= rrr = 3
1 2 J 4
we find that
7'i = 1 .050 11) r. = 350 lb
1\ = 750 lb 7'4 = 250 lb
Let us now take the complete shaft-pulley as.sembly as our free body.
The forces acting on this free body are
Ti = Ti(- sin 30° j + cos 30° k)

T, = 7'-,(- sin 30° j -\- cos 30° k)
T3 = -r:j
T4 = -T4
C = C',J -f CM
u = /;.i -f- L>j + DM
Writing the ecjuations of eciuilibrium,
ZMy - :L{zt\ - xh\) = -3(ri -f r.) cos 30° - 10/). =

l)^ = -364 lb
Z^h = 2(xF„ - yF,) = -3(7'i -\- 1\) sin 30° - 3 X 80

- 7(7', + 7'4 + 150) + 10/)„ =
/;„ = 1 ,039 lb
ZFy = (-7\ + 7',) sin 30° - 80
- (7', + 7'4) - 150 + />>„ + C'„ =
Cy = 891 lb
2F, = (7', + 7^2) cos 30° + Ih + C. =
a = -846 lb
It is obvious that /)i = since there are no forces applied along the axis of
the shaft.
Example 4-5. A uniform bar AB is fixed by a ball joint at one end
A and leans against a vertical wall at its other end B. The distance from
A to the vertical wall is a. The length of the V)ar is /, and its weight IF.
: ,
Fig. 4-7
If the coefficient of static friction between the end B and the wall is /is,
find the angle a at which the bar will just start to slide down (Fig. 4-7).
When the bar just starts to slide down, the frictional force at B tends to
stop the motion and therefore tangent to the circular arc DBE. Taking
is
the bar AB as our free body, and considering the equilibrium of moments
about joint A, we have the following forces that contribute to the moment
about A :
Normal force: N = iVi

Friction force: Y = F sin a } —F cos a. k
Weight W = - Hj
The position vectors at the points of action of these forces are
2rii- = r// = r.v = — ni + s/l- — a'- cos a j + \/F — a- sin a k
Hence the moments about .t of the N'arious forces are
M/.- = r/,- X F = — ^l'^ — a- — aF

/''
\ cos a j — aF sin ak
M.v = r.v X N = iV V^^ - a^ tîn aj - N Vl^ - a^ cos ak
M,r =
T»/i
r„ X
«-
W = W /t:.
- «"
., .
«
.
+
,
a IF
k
-2" V^" •'^m i .^^
S(v. 1-6] Couple 6.')
Applying -M.i = 0, we (ihtain
—F +
W sin
— rt =
— aF cos a + X \/l' — a- sin « =
aF sin a — .V y/l- — a- cos a + "V =
Solving the first two cfpuitions (the thii'd eciuation ;ul(ls no new informa-
tion in this jiroblem),
^V
rJ? = — sni a
a\V cos a
A^ =
Hence, for a less than a certain valne, and /''

X have values given above.
When motion impends, then/'' = n^X and
the limit of a is
Ustt
tan a =
Vl' - a'
This represents the maxinuim a for

e(iuilil)rium.
4-6. Couple
A couple is defined as a pair of ecjual and y^^^, 4.^
opposite forces whose lines of action do not
coincide. Let F and — F he the pair of forces passing through points .1
and B (Fig. 4-8). The moment of this couple, f al)out an arbitrary point
0, is
M = -r X F + (r + u) X F
or M = a X F (4-11)
where a is the vector connecting the two arbitrary jôints on the lines of
action of the two forces B and A. Equation (4-11) shows that the result-
ant of a couple is independent of the location of the moment center. A
couple vector, therefore, has the same moment about all points in space
and thus is a free vector. Furthermore, since
|a X F| = /'' (projection of a ± F)
= /•' (perpendicular distance between F and — F)
= /''(/
customary, for simplicity, to speak nf the inoniciit of a cdiiplc as tlic ((inj)!!-

t It is
vector, or even more hriefly as the couple.
the moment of a couple is a vector of magnitude equal to the product of

the force comprising the couple and the perpendicular distance between
the pair of equal and opposite forces. The direction of the couple vector
is perpendicular to the plane of the pair of forces with sense as determined
by the right-hand-screw rule. Accordingly, we can change the individual
values of r and F without changing the moment of the couple provided
(1) the product oi F and the perpendicular distance between F and — F
is kept constant, (2) the plane containing r and F remains in the same
orientation, and (3) the sense of the couple is maintained.

Clearly then, since the moment of a couple, or simply the couple, is a
vector quantity, all the permissible vector operations are applicable. If
we consider a couple as a force system, it has the peculiar property that its
resultant force is zero although its resultant moment is a X and there-

F,
fore nonzero. Hence, if a couple were applied to a rigid body, it would
exhibit a tendency to rotate the body without any tendency to translate
the body.
4-7. Equipollence of Force Systems in Space
Two force systems are said to be equipollent^ if they have (1) the same
total force and (2) the same total moment about the same moment center.
The significance of the concept of equipollence of force systems lies in the
following statement: Equipollent force systems acting on a rigid body
influence the equilibrium or motion of the body in exactly the same way.
To prove the above statement, let us assume that a system of forces S
consisting of Fi, F2, . . . , F„ acts on a rigid body (Fig. 4-9a). For each
of these forces, we can calculate its moment with respect to a common,
arbitrarily assigned moment center 0. We then form
2F = Fi + F2 -H • • •
+ F„
SM = Ml -1- M2 + • • •
+ M„
The graphical construction of SF and 2M is indicated in Fig. 4-9/>, in
which we define
^'- " ^^ (4-12)'
M, = SM ^
In general, F^ and Mr do not vanish, and hence the rigid body is not in
equilibrium.
t We use the term (MiuipoUont, rather tlian the term eeiuivnleiit or even mechanically
equivalent, because of its by its definition. If
precise technical meaninjr as supplied
we consider a defornuible bar acted upon by two and colinear tensile
equal, opposite,
forces, this force system would be e(juiiwllent to a force system composed of two com-
pressive forces, or even equipollent to a zero force system. However, the three force
systems would not be equivalent; certainly their effect on the deformable bar would be
different in each case.
Sec. 4-7] 'Eqnipolletice of force syslems in space 67
We now
add to the force system S a pair of ecjual and opposite forces
Fr and F„+i and a pair of equal and opposite moments Mr and M,.-i-i.
(To simplify our demonstration, we may let Fr and F„+i act at the point
so that neither force contributes any moment about 0. We also let Mr
and M„+i be two couples so that neither of them contributes any net
force.) Now the state of e(iuilil)rium or motion of the body is obviously
not changed by the addition of such e(iual and opposite pairs of forces and
pairs of moments. But now the sum of forces Fi + Fo + -f F„ • • •
(b)
Fi(i. 4-0
-}-F„+i and the sum of moments Mi M) + + + M„ + M^+i are

• • •
both zero. They actually satisfy the etiuations of equilibrium, and hence
can be removed from our consideration. Therefore the cfiuiiibrium or
motion of the body is due solely to the influence of a single force Fr acting
at and a couple Mr. Since the force system that consists of Fr at and
Mr is, by definition, equipollent to the original force system S, this proves
that ecjuipollent force systems exert the same influence on the equilibrium
or motion of a rigid body. In terms of force systems, it now becomes
apparent that the neces.sary and sufficient condition for the c(iuilibrium
of a rigid body is expressed by the fact that the total external force system
be equipollent to zero; i.e., Fr = Mr = 0.
It can be proved that if two force systems are equipollent with respect
to some moment center 0, they are etiuipoUent with respect to any other
moment center, say C/. The first part of the condition of equipollencc,
namely, that the resultant force is the same for both systems, is obviously
not influenced l)y any shift of the moment center. The second part of
the condition of equipollence states that if two systems are equipollent

with respect to the moment center 0, the resultant moment about 0, or
2Mo = S(r X F)
is the same for both systems. If we now shift the moment center from
to 0', the new resultant moment will be
SMo' = S(r' X F) = S(r + a) X F

(4-13)
= SMo + a X SF
where a is the position vector from 0' to 0.
Since SF, SlMo, and a are
the same for both systems that are equipollent with respect to 0, this
proves that SMo' is also the same for both systems, and hence the two
systems are also equipollent with respect to any other moment center 0'.
4-8. Reduction of Force Systems; Wrench

Let us suppose that a given force system S consisting of many forces has
itsresultant force represented by 2F and its resultant moment about an
arbitrary point represented by SM. As has just been shown, this force
system is equipollent to a single force
f; = F,
¥r acting at and a single couple
having the moment M^, where
SF and M, = SM
This is called the reduction of the
force system to a force and a couple.
In particular, a single force F atP can
always be reduced to a force and
couple acting at some other point 0.
Here the equipollent system would
be
Fk;. 1-10
F, = F M, = r X F
where r is the position vector from to some point on the line of action
of F. As we shall see shortly, the converse is only valid when the force
and couple vectors are mutually perpendicular.

Although the couple M, can be located anywhere in space, its particular
value is associated with the choice of the point about which liiM is taken
and on which the single force F, acts. \'\n- this i-oason, we alsodraw INT^
at the point (Fig. 4-10). It is conveincnt to use a doul)le arrowhead to
denote couples M as distinguisiied from a single arrowhead for forces F.
We can now resolve M^ into a longitudinal (i.e., along F^) component M^l
and a cross (i.e., perpendicular to F,) componont Mrc as shown in Fig. 4-10.
Sec. 1-8] Reduction of force syslems: wrench 69
Now Mrc and Fr at is F', = F^ located at a

eciuipolhMit to a sin«rlo force
di.-^taiice r from 0, so that M^c. r X F^ =
The single force F^ must
certainly be located in a plane perpendicular to M^c passing through the
point 0. This is always possible because M^r is perpendicular to F,. In
fact, from the property of vector products we see that r may assume an
infinite number of values. However, the distance from to F^ is given
by d = Mrc, I'\ (Example 3-5). The locus of r is a straight line parallel
to Fr. The S is therefore reduced to a single force
original force .system
Fr at r fiom and a single couple MrL, both having the same direction.
()
A force system consisting of a force and a couple, with the same direc-
tion for both, is called a wrench or a screw. The original force .system is
therefore reduced to a wrench. The location of the wrench is obxiously

not arbitrary, since the line of action of the wrench is fixed at a distance r
from 0. If we such a way that
shift the location of 0, r al.so ciianges in
the line of action of the equipollent wrench remains at the same place
with respect to the original force system. For exami)le. if the location of
point were shifted to .some new point 0' along the line of action of the
wrench, the original force system would be eijuipollent to
F, = 2F and M; = ^M^.
where Ml represents the vector component of each individual moment in
the direction of Fr.
To conclude: A general force system can he reduced to (l) a force and a
couple in an infinite number of ways, depending on the location of the point
on which the force acts, or (2) a wrench in only one way.
For the special case when a force system consists of parallel forces
only, Mr is always perpendicular to Fr, and hence consists of cross com-
ponents only. From the previous discussion, it is .seen that such a force
system can be reduced to (1) a force and a couple in an infiiiile number of
ways; the force itself must be Fr = 2)F, but the value of the couple
depends on the position of the moment center; or (2) a single force in
one way, where again the single force must be Fr = 2F and its line; of
action is determined uniciuely by the perpendicular distance d fioni an
arbitrary point to the line of action of F^, d = Mr/Fr. One important
applicaticMi of such a force .system is when the parallel forces are the
weights of the particles of a system; then the location of the e(|uipollent
single force becomes the center of gravity of the .system and the single
force becomes the total weight (gravity force) of the .system (body).
This will be discussed in some detail in the next chapter.
()th(!r special forcesystems merit some comment. Concurrent force
systems (Mr =
about the point of concurrency) are clearly (!(|uipollent
to their resultant pas.sing through the point of application and may be
reduced to a force couple system in an infinite uMinbcr of ways and a
single force (the resultant itself) in one way.

Force systems consisting
of couples alone are equipollent to their resultant couple and cannot be
further reduced. Two-dimensional, or planar, force systems constitute
a system where Mr is always perpendicular to Fr, and thus the previous
discussion of parallel force systems is especially pertinent. In general,
then, reduction to single force is possible only when M^ and Fr are mutu-
ally perpendicular.
Example 4-6. Given a system of coplanar forces Fi, F2, and F3 acting
as shown (Fig. 4-11). Find (a) an equipollent force system acting at the
point and (b) an equipollent force system consisting of a single force.
Fi=100lb
F2 = 80 lb
F, = 60 lb
Fig. 4-11
Solution. Expressing the forces and position vectors in component

form,
Fi - 80i + 60j _ ri = 4i + 6j
F2 = 40i + 40 V'Sj r2 = 8i + 4j
F3 = -60i r3 = 8i +j
Then, according to Eqs. (4-12),
F,. = Fi + F2 + F3 = 60i + 69.3J

Mr = ri X Fi -h r. X F. + rs X F3 214.2k
It should noted that the resultant moment of any coplanar force sys-
})e
tem about any point in the plane consists only of a component perpen-
dicular to the plane. The original force sj^stem is now reduced to an
equipollent system consisting of a single force Fr acting at and a single
couple Mr.
To further reduce this system consisting of a resultant force Fr and a
resultant couple Mr to an equipollent system consisting of a single force,
we note that M,. is jiorixûlicular to F,. and thus Mrc = Mr. Therefore
Sec. 1-8] Reduclion of force systems; wrench 71
the location of the single force F' = F^ is defined by
d = ^lr ^ 214. 2
=
Fr
~ 91.7
2:A in.
where d is the perpendicular distance from to F^ The sense of (/ may

be determined by inspection so as to
maintain eciuipoUence with respect
to the .sense of the couple M, (Fig.
4-12).
In a more formal manner the same
result may be obtained directly' from
the definition of eciuipollence. Thus
we require that
F; = F. = fiOi + r)9..3j
aiu r X F' = Mk = Mr
Letting r = .)vi + //,j + 2rkandsolv 1.16x,-3.57
ing, we find that

4-12
Zr = 09.8.(v - 00//, = 214.2
Fir,.
The equation of the line of application of F, is then given as
yr = I.IG.JV — 3.57
Example 4-7. Given a general system of forces Fi, F2, and F3.
Reduce this system to an equipollent system of forces at the point and
further reduce the .system to a wrench.
Solution. The given forces and their respective position vectors are
Fi = 8i - 4j + 2k r, = Oj
F2 = 3i + 2j + 3k r, - Si + 2k
F3 = i + Oj - Ok rs = i — k
The given system is equipollent to a force couple system consisting of
Fr = 2F = 7i + 4j - i
Mr = 2M = 2r X F = I4i -f- 2k - Ok
Referring now to Fig. 4-10, we find that
_ _ Mr • F, Fr 50 ,_. ,
,
. , .
and Mrc = Mr - Mrt = 2..^3(35i " 79j - 71k)
The equation of the line of action of the ecjuipollent wrench may now be
determined by letting
r X F' = MrC
where r is a position vector of that hne of action and is expressed as

r = Xri -\-
Vrj +
Zrk. Solving the resulting vector equation,
Xr = KVr ^>66 = -7; i^M3

The perpendicular distance from the origin to the line of action of the
wrench is
d = Mr 2_ 12,507
33 66
PROBLEMS
4-1. A
"simple" beam is supported at one end by a fixed hinge and at the other
end by (The rollers are such that complete restraint is offered in the
"rollers."
vertical direction, both upward and downward, but no restraint is offered in the hori-
zontal direction.) The beam is loaded as shown. Find the reactions at each end.
20001b
10001b
'777777' %:2400fMb ooO

3' 3' 3'
12'
Fig. P 4-1
4-2. A beam, fixed in a wall so that no rotation or translation can occur, is loaded
as indicated. Determine the reactions necessary to maintain equilibrium.
20001b
12001b
|2000lb
2' I
4'
18'
Fig. P ^-2
A small truss-type highway bridge has a span length of 80 ft. The trusses,
4-3.
of which there are two, are assumed to consist of weightless bars connected together
with frictionless hinges. A loaded vehicle is positioned in panel he as shown. The
weight distril)ution of the vehicle is 2,800 lb on the rear axle and 2,200 lb on the front
axle. Assuming that half of the vehicle weiglit is distributed to each truss and that
the forces acting on tiie panel points h and c are carried through the floor system by
simple beam action, determine the reactions on a single truss. Find the forces in the
members /(; and U.i. Hint: Since llic truss is in (Mpiililiriuni, each portion is in equi-
Problems 73
librium. Hcncp, solcct a portion as a fnM> Ixxly (consitlcnnl also as a r\^,'u\ \mdy) that
will I'liabie a dctormination of the nMiuirt'd fort-cs.
Fig. P 4-3
4-4. A piano truss sliown is liiiipiod at .1 and siiiii)ly supported <ni rollers at D. The
bars CD and AB make G0° with the horizontal, and the bars (\\ and lilJ make 30°
with the horizontal. A total load of 3P is distributed at the two bottom joints as
shown. Find the reaetions in the bars AH, CD, and EF.
40001b
'.jX
1000 lb
Fi(i. P 4-4 Fk;. P 4-5
4-5. Determine the reactions re(iuircd for efiuililtriuni shownfor the structure
Neglect the weights of the mend)ers.

4-6. Find the force in the two guy wires of Prol). 3-11 if the bar 01' is to be in
equilibrium under its own weight.
4-7. A .shaft carries a gear .1 and a pinion B. The diameters are 48 in. for .1 and
6 in. for B. The tangential force F, acting at the rim of .1 is found to be 50 lb in the
direction .shown. The force F2 acting at the rim of B is vertical. Find the reactions
in thetwo bearings C and D. If the shaft rotates at 350 rpm, how much is the horse-
power transmitted between A and B?
Fig. P 4-7
4-8. A boom AB, 12 ft long, is held at one end A by a frictionless ball joint. It
carries a load of 1,000 lb attached to the middle. Besides the joint A, the boom is
also supported by two wires attached to the boom at B and C as shown. Find the
reactions in the wires and at the ball joint.
100 lb
Fig. P 4 I'^Ki. P 4-9
4-9. A space structure is supported at A by a ball-and-socket joint, at B by a

hirigod bar, and at C by a slip ring. Find the reactions necessary to maintain equi-
libriuni. Assume that all the bars are woishtless.
i-lO. An ('(luilatcral triangular tal)le toj) of si(l(>s (t is .supixjrtcd by throe vertical
bars and throe inclinod bars hiiigod to its corners as sliown. Tho throo inclined bars
make an angle of 'M)° with tho horizontnl. Dotorniiiio tlio forces produced in those
bars duo to a couple of niomont .1/ acting in tho ]il:iiio of tlio table top.
Problems 75
Fic. P 4-10
i-11. A uiiifonn box lid of wfiight IF with hinges at .1 and li is lifted :}()° from the
horizontal by a rope attached at one lifted corner C. Find the reactions of the
hinges if (a) the rope is horizontal and parallel to the x axis, (b) the rope still lies in the
xij plane but makes 30° with x.
ABC lies in the xy plane and triangle DEF lies
4-12. In the truss shown, triangle
plane parallel to the xa plane. AB = AC = DE = DF = EH = DA =
in a vertical
FC = 1 ft. The truss is supported by a ball joint at A, a, ring guide at D, and a roller
support at C. A vertical load \V is api)lied at the joint F, and a horizontal load F is
applied at the joint E. Assume all joints are frictionless. (a) Find the reacitions at
the three supports (some reactions may be negative), (h) Find the reaction on the
l)ar AF. State whether it is tension or coinprcssion.
37501b
llOlb
Fi«;. 4-12 Fi<i. V 4-i;i
4-13. The forces indicated are acting on an automobile. Find an equipollent sys-
tem of forces acting at 0. Further, reduce this system to a single equipollent force.
4-14. Determine the relation between the magnitude of the forces shown, so that
they may be reduced (i.e., are equipollent) to a single force.
Fig. P 4-14 Fig. P 4-15
4-15. A force system consists of two forces Fi and Fo and a couple C acting in the
plane of abed.
Fi = 5\h F. = V2 lb
The magnitude of the couple is 10 ft-lb.
Oa = 4 ft 06 = 3 ft 6c = 3 ft
Find the equipollent force system consisting of one force through

(a) and one
couple. (6) Find the equipollent wrench, (c) Can the given force system be
reduced to a single force?
4-16. Three forces of equal magnitude act on a cube of sides a as shown. Find (a)
the equipollent system of a force at the origin O and a couple; (6) the equipollent
wrench and its location.
Fig. P 4-16 Fi(!. lM-17
4-17. What is the necessary relation betweenn, b, and c so that the three noninter-
secting forces (of equal magnitude P) acting as shown may lie nMluced to a single
force?
4-18. (live ail example of a body or system that satisfies I'^js. (4-S) and yet is not
in equilibrium.
4-19. tShow that Mrc and Fr are equipollent to a single force V, located at a position
r from O as defined by r = F, X Mrc/Fr. Show also that this results in the location
of Fr at a perpendi(Mdar distance d = Mrc/Fr-
CHAPTER 5
\
More Applications of Statics
j
5-1. Systems of Parallel Forces
j
It was discussed in Sec. 4-7 that a sj'stem of forces, in ^fncral, is
I equipollent to a single force F^, acting at an arbitrarily chosen point 0,

I together with a couple Mr. The magnitudes and directions of F, and .Mr
are determined by
Fr = ZF
Mr = i:M
where -F and 2M are, respectively, the sum of the forces of the given
system and the sum of the moments about due to these forces. It is
clear that Fr is independent of the choice of the point O l)ut Mr varies both
in magnitude and in direction with a shift of point (). At one parti(adar
location of the point 0, the vector direction of .Mr can be made parallel to
the direction of Fr (i.e., the plane of action of the couple is perpendicular
to Fr) ;
this particular eciuipollent force system is called a wrench.
The statements made above are fjuite general and apply to any system
of forces. Sometimes, however, the given system of forces is such that
one or more of the following situations may arise:
1. Fr = 0.
2. Mr = about a particular point 0.

3. Mr about a particular point O is perpendicular to Fr.
If Fr = 0, the given system of forces is (M|uipollent to a couple Mr only.
If Mr = 0, the gi\'en system can be reduced to a single force Fr acting at
point 0. If both Mr and Fr are /(^ro, the gi\(Mi system of forc(!s is in
ecjuilibrium. I'inaliy, if Mr about is perpendicular to Fr, the system
can further be reduced to a single force; Fr ac^ting at such a distance away

Ironi that its moment about is Mr.
It is this last situation that applies when all the forc(^s in the system are
parallel, since then .M, is always perpendicular to Fr. A system of
78 More applications of statics
parallel forces, therefore,can always be reduced to a single equipollent

force unless SF = 0. The
location of this single force is such that the
moment M^ of the single force taken about some arbitrary point is equal
to SM taken about the same point. If F^ = 0, the given system is
equipollent to a couple M^. And if both F^ and M^ are zero, the given
system is equipollent to zero (i.e., the system is in equilibrium).
Let us now consider a system of parallel forces and investigate the
location of the equipollent single force. (We assume that SF ?^ 0.)
For this purpose we orientate our coordinate axes so that the y axis is
along the direction of the parallel forces
(Fig. 5-1). The forces in the given system
can therefore be represented by
Fi = Fij
F2 = F2J
etc. Hence the equipollent resultant force

Fi is
F, = SF - (Fi + F2
+ •
)j = (SF)j (5-1:
Thus the scalar component of the resultant

Fig. 5-1
force is simply the algebraic sum of the
scalar components of the forces Fi, F2, etc., and
F, = SF
The location of F^ must be such that
M, = SM (5-2)
for equipollence. But with all the forces having the y components only,
M, = -z,Fri + XrFM
and SM = i:{-zFi -\- xFk)
where x and z represent the coordinates of the lines of action of the forces.
Hence Eq. (5-2) is equivalent to the following two scalar equations:
XxF
I
SF
(5-3)
Zr = S2F
SF
Equations (5-3) give the x and z coordinates of the line of action of the
single resultant force equipollent to the given system of i)arallcl forces.
It is also clear that the total moment S!M of the gi\-ensystem of forces
vanishes about any center that lies on the liiu^ of action of the equipollent
single force. Thus a singl(>, force F^ = SF acting at coordinates x — Xr
Sec. 0-2] Center of gravily: center of mass: ceniroid 79
. and ij = ih on a rigid lH)dy would produce the same ellect as that caused
'
by the original system of parallel forces.
5-2. Cent<M- of Gravity; Center of Mass; Cenlroidf
An important application of the previous discussion regarding systems

of parallel forces is when the parallel forces are taken as the gra\itational
forces acting on the various particles of a body. The center of yraritij of a
j
body is defined as the point (denoted by C) of action of the single, resul-
i tant force equipollent to the system of gravitational forces. Considering
the gravitational forces as comprising a system of parallel forces is valid
only when the distances between the various particles is sufficiently small
so that all particles lie in an essentially constant gravitational field.
Fortunately, the vast majority of bodies we deal with in engineering
applications satisfy this condition. Then Eqs. (5-3) can be written in
the following form for the calculation of the coordinates {.vcUc^c) of the
center of gravity:
xc
^VxiVi Sr.-w;.-
Sz,-ty»
zc =
where x„ r„ and iVi refer to the x coordinate, position vector, and weight
of the ith particle, respectively.
In the above we have extended E(is. (5-8) to obtain not only .re and Zc,
but also an equation for ijc. This can be done by merely rotating the
body as well as the coordinate axes so that either the z or x axis becomes
vertical, and hence parallel to the gravitational forces. The .same
reasoning that was used in deriving Eqs. (5-3) then gives a similar
equation for ijc.
If we consider an actual body as consisting of a continuous distribution

of weight, the numerators and the denominators in Eqs. (5-4) must be the
summations of an infinite number of terms, iw, in each term representing
the ma.ss within an infinitesimally small volume dV. Carrying this to the
limit according to the concept of integral calculus, Eqs. (5-4) should now
be written
f X dw I y dw I z dw
xc = . yc = ^ .dw Zc - r
- (.o-oj
\dw \ I dw
t Values of the centroid of some plane geometric shapes and the center of mass of
some homogeneous bodies are given in Appendixes .\ and B.
where the integration is extended over the entire volume of the body.
Equations (5-5) may be written in the alternative form
/ r dw
re
/.-
dw
This concept of the center of gravity depends, for its definition, on the
existence of a gravitational field. Another concept not so dependent
may be defined. This is the idea of center of mass of a body, where we
define the center of mass as that point (denoted by C) having coordinates
/ X dm /
^ y dm j ^ z dm
XC = —r VC = —r 2c = —r
/ dm, / , dm / ^
dm
(5-6)
I
^ r dm
or re
dm
/.
The increment of mass dm is equal to p dV, where p is the mass density of

the body; the indicated integration in Eqs. (5-6) is to be taken over the
entire volume of the body. The denominator / dm is the total mass of
the body. The numerators are often referred to as they^rs^ moment of the
mass of the body with respect to a plane. For example, / ^ x dm is
the first moment mass with respect to the yz plane. Equations

of the
(5-6) can thus be interpretedin the following manner: The first moment of
the total mass about any plane
is equal to the product of the total mass and the
distance of the center of mass to that plane.
Thus any plane containing the center of mass is determined so that the
first moment of the mass to one side of the plane is equal and opposite to
the first moment of that portion of the mass to the other side of the plane.
In particular, then, the center of mass must lie in a plane or planes of
symmetry.
By analogy with Eqs. (5-5) or (5-6) , it can be defined that the centroid
of a volume V is located at the point (denoted by C) with coordinates
._/v xdV
X= ^, (5-7a)
with similar expressions for ?/ and z. Again, the centroid of a surface is

> Sec. 0-2] Center of grarily: eenler of mass; eenlroid 81
defined as the point witli the coordinates located by
/ x'dA
with simihir expressions for // and z. Finally, the centroid of a line seg-
ment L may be defined to be located at the point with the coordinates
! xdL
^ - 'r (5-7c)
again with similar expressions for // and z. The intcj^rals { x dW

I
X dA, or / X dL, for example, are the (irst moments of the xolume, of
the area, or of the line segment about theijz plane and are to be taken over
the entire volume, area, or length of the respective body.

When we consider bodies in the vicinity of the surface of the earth, the
aecelerati(ni of gravity g ma}^ be considered to be essentially constant.
Since dw = g dm = 7 dV = pg d\\ and comparing Eqs. (5-5) and (5-G),
we see that for constant g the center of mass and center of gravity
coincide. Moreover, according to Ecjs. (5-6) and (5-7a) it is clear that for
a homogeneous body, one whose mass density p is uniform throughout,
the mass center and centroid of the volume coincide. If the body is not
only of uniform material, but also of uniform thickness, which is small
compared with the other dimensions of the body (a thin plate, the area of
which, however, need not be in one plane), the center of mass of the body
approximately falls on the centroid of the middle surface of th(! body [cf.
Eqs. (5-6) and (5-76)]. When the body possessing uniform material and
uniform thickness is prismatical (all its generators ai-e parallel), the center
of mass coincides exactly with thecentroidof tluMniddlesurface. Finally,
if the area is small compared with the length of the body (i.e., the body
has the shape of a slender wire, not necessarily in one plane), the center of
ma.ss of the body approximately coincides with the centroid of the line
segment which forms the spatial axis of the body [cf. Eqs. (5-6) and (5-7c)].
When this body possessing uniform material has a constant cross-section
area, the center of mass coincides exactly with tlu; centroid of the
spatial axis.
Although the previous discussion refers directly to continuous bodies,
these ideas may easily be (extended to include: cases of discont iimous bodies
of a finite number of j)ortions. For this event, relationships similar to
Eqs. (5-4) are obtained.
This entire discussion may be generalized by introducing the concept

of a weighting function. Suppose we consider a system of points in space,
and to each point of the set is associated a particular value, or weight.
This weight may —
be any desired property mass, gravity weight, volume,
charge, etc. — and is denoted by the symbol pi. We define the centroid
of the given set of weighted points as
- = ^' (5-8)
where Vi and pi represent the position vector and weight of each point,
respectively. If instead of having a discrete number of points in the set
we were to have a continuous distribution, the centroid would take the
form
_ jrdp .
^^-^^
""'-Jdp
where p now is a weighting function. Equations (5-8) and (5-9) can, as
before, be written in terms of the rectangular coordinates of the centroid.
It is noted that this more general definition of centroid is in perfect
agreement with our previous development, if we let the weight or weighting
function be the particular quantity of interest. In a certain sense, then,
the centroid represents the average position of the weighted set of points.
We also note from Eqs. (5-8) and (5-9) that the position of the centroid is
not altered if we replace the weighting numbers pi by a set of numbers

proportional to them.
Further, it can be shown that there is only one point C that satisfies
our definition [Eq. (5-8)]. To do this let us choose our origin, from which
the position vectors r^ are measured, at the centroid C. Then re = 0,
and Eq. (5-8) becomes, provided Spi y^ 0,
Sripi =
Suppose there is another point D which satisfies Eq. (5-8). Then
Sr^pi =
where r •
is the position vector measured from an origin at D. Subtracting
these two expressions,
Sp,(r. - rO = (Zpî =
where d is the position vector from (' to D. The Inst st(>p follows since d
isa constant, independent of each individual point. T'rom this it follows,
provided Sp, 9^ 0, that d == 0. Thus the points (' and D coincide, and
there is a unique centroid as defined by Eq. (5-8).
The reader should refer to a book on integral calculus and review the
Sec Center of (fravily: center of muss: cenlroid 83
techni(iue for finding the centers of mass and centroids ol a xohiinc, an

area, or a line segment by means of calculus. In particular, we would
like to mention the theorem of Pappus, \vhich states:
1. The area of the surface generated by rotating any plane curve about
an axis in its plane is equal to the product of the length of the generating
curve and the distance of travel of its centroid during the rotation, if the
generating curve lies only on one side of the axis.
2. The volume of the solid generated by rotating any plane area about
an axis in its plane is equal to the product of the generating area and the
distance of travel of its centroid during the rotation, if the generating area
lies only on one side of the axis.
The proof of this theorem can be found
in any standard text on calculus. The
following examples show one applica-
tion of this theorem.
Example 5-1. Determine the /y
coordinate of the centroid of the area

enclosed between the straight line and
the parabola shown in Fig. 5-2. Ac-
cording to Eq. {5-7b), the desired cen-
troidal distance is obtained by evaluat-
ing the following expression:
y = -
Fig. 5-2
The equations of the straight line and the parabola are given, respectively,
as
y 6-" //
= p.T-
Choosing now as the elemental area dA the horizontal strip of variable

width x' and height dij, we find that dA = x' dij and that
dy
y =
-iri -lU dy
Performing this integration, the centroidal distance is
y = Hh
Example 5-2. Knowing that the volume V of a sphere of radius r is
V = %Trr^
:
and that the surface area is *S = Airr'^, use the theorem of Pappus to
determine the centroid of a semicircle and of a semicircular arc.

Since a sphere is generated by rotating a semicircle about its own
diameter, we have (Fig. 5-3), by the second theorem of Pappus,
4 _ Trr^
V - 7rr=* = 2'n-yA -^
4r
Hence Va
3^
The surface area of a sphere is generated by rotating a semicircular arc
about its own diameter. Using the first theorem of Pappus,
S = 47rr2 = 2TryL
2r
and therefore
It is important to notice that the centroid of a line, as evidenced by this

example, does not necessarily have to lie
on the physical confines of the geometri-
cal line. Similar statements apply to
areas and volumes.
Often we are faced with the problem

of finding the center ofmass of a body
consisting of several portions. If the
individual centers of mass of each portion
-r^
_ „ are known, the center of mass of the
Fig. 5-3 .
'
composite body can be found by a simple

application of operations similar to Eqs. (5-4)
re (5-10)
J
Here the distributed ma>ss of each portion is considered to be equipollent
to having the total mass of the portion concentrated at a mass point,
namely, its own center of mass. We leave the proof of this statement as
an exercise (Prob. 5-25). The composite body can therefore be considered
as consisting merely of a finite number of mass points, and Eq. (5-10) can
be directly applied. A similar procedure can be used to find the centroid
of composite vokimes, of composite areas, or of composite line segments or
the center of gravity of a composite body.
Exaiiii>le 5-3. A crank-and-shaft assembly has the dimension and
shape as shown in Fig. 5-4. Assinning the assenU:»ly is made of uniform
material, find its center of gravity.
Since tlu; assembly is made of nnifoi'm niatcM'ial, the center of gravity
>t'C. -3] Pressure ht a sialic Ihiid 83
coincides with the centroid of the vohiine. Let us. for convenience, phice
our coorilinate system so that the indiviihial centroids of the five portions
all lie in tlie .r/y phme. Thus zc ^ 0, and the problem therefore is reduced
to a two-dimensional one. We also arbitrarily place the origin at the
left end of the assemblv and at the axis of the main shaft. The volumes
what from its A fluid, on the other hand, is unable to

unstressed state.
withstand any shear no matter how small, without being con-
stress,
tinuously deformed by a relative motion between its layers.
With the absence of shear stress, or shear force, the force in the interior
of a fluid at rest is always normal to the surface on which the force acts.
The normal force per unit area is called the pressure, or the pressure
intensity. The pressure at a point is the limit of the ratio of normal force
to area as the area at the point approaches zero as a limit. It can be
shown t that when shear stresses are absent, the normal stresses become a
constant in all directions. The consequence here is that at any given point
in a fluid at rest, the pressure on a surface is a constant regardless of the
orientation of the surface. An ele-
mentary proof of this is given as follows.
In a fluid at rest, let us consider a
small triangular prism of fluid of unit
dimension perpendicular to the paper
(Fig. 5-5). As explained before, the
on the three surfaces of the
forces acting
prism are along the directions normal to
the surfaces. Let px, Py, and ps be the
average pressures on the three surfaces
and let 7 be the specific weight (i.e., weight per unit volume) of the fluid.
Considering the fluid in the triangular prism as our free body, we can
write the equations of equilibrium along the .1; and the y directions.
2Fx = Pxdy - {ps ds) sin =

2F, = pydx - (ps ds) cos e - lî^ =
I
Since (rfs)sin 6 = dy, (ds) cos 6 = dx, the above equations may be
reduced to
Px = Ps
Pv - Ps - ^
y dy
= „
When we consider the pressure at a point, we let the triangular prism

shrink to zero size as a limit. As a result of this limiting process, the last
term of the second equation approaches zero; it thus becomes negligibly
small compared with the other terms and may be neglected. Hence
Px = Py = Ps
Since 6 is arbitrarily chosen, the direction of p^ makes no difference to

the above equality. Hence the pressure in a fluid at rest is a constant in
all directions.
t The proof, Imscd on tho p;onoral consifloration of stross tensors, w ill ho p;ivcn in
Chap. 15.
Sec. 5-3] Pressure in a sialic fluid 87
Let us now consider the ef|uilihriuni of a horizontal tube of fluid iiaving

the shape of a cylinder (Fig. o-Oa). The cross section of the tui)e is dA,
and the axis of the tube is arbitrarily assigned to be our x axis. Consider
the fluid in the tube as a free body: we find that the only forces acting
|P2dA-(p+ dp)dA
PidA p^dA dA
^ dA
dy
ydydA
(a) ib) 1. dA = pdA

Fic. 5-0
along the x direction are the pressure forces on the two ends of the tube.
The equation of eciuilibrium 2Fi = gives
pi dA - P-,
dA =
or Pi = P2
Hence the pressure in a continuous fluid at rest is constant in each horizontal

plane.
Next consider the equilibrium of a vertical column of fluid of cross
sectiondA and height dij (Fig. o-()/^). Writing the ecjuation of e(juilib-
rium ZFy = 0, for the vertical column of fluid as a free body, we obtain
p dA — (p + dp) dA = 7 dij dA
or dp = -y dy (5-11)
The negative sign indicates that the pressure decreases with increasing
altitude. Equation (o-ll) is referred to as the pressure-altitude relation
of a static fluid.
actual calculation of the pressure difference between two points
The
separated by a vertical distance in a compressil)le fluid re(|uires the
integration of Eq. (5-1 1). In order to do this, we nuist possess a relation
between 7 and p. For example, if the fluid under consideration is a
perfect gas under constant temperature, we have the well-known iso-
thermal relation
p = constant X 7
This constant can be evaluated if we know the pressure and the specific
weight pi and 71 at a point of elevation yi. Integrating Eq. (5-11) then
gives the pressure p2 at some other point of elevation i/2 as
y2
- Â
^— = exp (5-12)
Pi 7>i/7i /
As another example, consider a gas that follows a polytropic relation
p = constant X 7"
where the exponent n may have different values depending on the problem.
Integrating Eq. (5-11) gives
n pi /pA("-r)ln-
y2 - iji
— 1 (5-13)
n 1 7i
For liquids, which can be considered incompressible for practical

purposes, 7 is a constant.Equation (5-11) then gives
P2 Pi = -yiu-i - iji (5-U)
An example of this type of pressure distribution is that existing in a

body of liquid at rest (hence the
name hydrostatic pressure, after the
most usual liquid) . FromEq. (5-14),
it is clear that, at the free surface of
a liquid (the interface with the atmos-
phere), the hydrostatic pressure,
measured from ambient atmospheric
pressure as a base, vanishes. Thus
hydrostatic pressure varies linearly
with depth below the free surface,
increasing from zero at a rate of 7
units of pressure per unit of depth.
Equation (5-14) is the basis for the measurement of pressure differences
by a device called a manometer. Figure 5-7 shows a simple type of
manometer. The pressure pi in a vessel containing liquid of specific
weight 7i is indicated by the difference in elevations of the two legs of
/)•)
the manometer li(iuid with specific weight 72. The pressure at point A,
where the two liquids come into contact (the two licjuids must be immisci-
ble), is the same as the pressure at point B, since the two points have the
same altitude and are connected continuously by the same licûid 72 in
the tube ACB. Hence
V\ + 7i/'i = V-i + 7i.'/'2
or V\ — P-i =" Tj/'j — 7i/'i (5-15)

1 Sec. 5-4] Forres due lo fluid shilir jtreasnre: plane snrfares 89
Thus we obtain the j^ressure at I'roin the known ])ressure at 2 and the
1
manometer readings hi and Other types of manometers can he more

/;_..
compUcated than this, l)ut tlie underlying principle is the same.

The engineer is faced with a nuilf iplicity of units for the measurement
of pressures. Since pressure is the noiinal force per unit area, the English
system u.ses the unit pounds per siiuare inch (psi) or i)ounds per square
foot (psf ). On the other hand, pressure can also he expressed in terms of
a column of licjuid, in which case it is interpreted as the force per unit area
at the base of the column. Thus we can express pressure in feet or inches
of water or in inches of mercury (in. Hg). Finally, all the above vuiits
can be either a gauge pressure (e.g., psig) which is measured above the
atmospheric pressure or an absolute pressure (e.g., psia). The specific
weight 7 for water is 02.4 lb ft''. Hence
02 4
1 ft of water = ^^T ~ 0.488 psi
441
12
1 ft of water = ,., -. = 0.884 in. Hg
since the specific gravity of mercury is 18.0. The standard atmospheric

pressure at .sea level is 14.7 psi. or 84 ft of water, or 21).i)2 in. Hg.
5-4. Forces Due to Fluid Statie Pressure: Plane Surfaces
Consider the problem of the pressure force acting on a Jlat plate

immersed in a fluid. Since the pressure is normal to the plate, the plate
is acted upon by a system of parallel forces. According to Sec. o-l,
the ecjuipollent resultant force is ecjual to Fr = SF and is located at
Xr = SxF/SF, ijr = I,!jF/I,F. Considering an element of area A.l of the
immersed plate, the force acting on it is p A. Therefore 1 .
p\ = ZF = rCp A.l
We now let AA approach zero as a limit while at the same time the number
of forcesapproaches infinity. The summation in the above e(}uation
becomes an integration taken over the entire area of the plate; i.e.,
Fr = p dA
f^
I
xp dA
Similarly, Xr = ^ (•''>1G)
f^pdA
UP dA
Vr = f^
It is recalled that only the x and y coordinates enter into the problem if the
direction of forces is along z.
Let us now apply Eqs. (5-16) to the case of a fiat surface immersed in a
liquid for Avhich the simple pressure-altitude relation Eq. (5-14) holds.
Figure 5-8 shows a flat surface AB submerged in a liquid and inclined at
an angle d with the horizontal. We take the intersection of the plane of
AB with the free surface as our x axis. The y axis is taken to lie in the
plane AB. This means that the plane AB becomes our xy plane.
Fig. 5-8
To compute the resultant pressure force Fr due to the liquid on one side
of AB and its location Xr and yr,
we proceed as follows. Take an element
of areadA formed by two horizontal lines on AB at a distance dy apart.
The pressure on dA due to the liquid is
p = yh = yy sin 6
Substituting into Eqs. (5-16),
Fr = y sin By dA = yy sin 9 A (5-17)
where y is the y coordinate of the centroid C of the given area. The

pressure at the centroid C is
Pc = yy yin 6 = yhc
Hence we can also write
Fr = pcA (5-18)
Thus the magnitude of the resultant pressure force on one side of aflat surface
is the product of the area of the s)irfa<-(' and the pressure at the centroid.
Sco. .i-1] Forces due lo Jhiid sialic pressure: plane surfaces 91
Since p = yy sin 6, Eq. (o-17) can also ho interpreted as the total volnme
under the pressure diagram acting on the plane area.
The location of the resultant pressure force Fr is described by the
coordinates x,, ijr of the piercing point of Fr with the given area. This
point is called the center of pressure. iMjuations (o-lOj yield
dA
Ja''^ III
Xr
yA
f.ydA
(5-19)
l.y^dA h
yr = -,
yA
lûdA
Here we have defined Iiy = j xy dA as the product of imrtia of the
given area with respect to xy axes and Iz — \ y' dA as the moment of
inertia of the given area with respect to the x axis.f

What seems at this point to be simply a mathematical abstraction,
moments and products of inertia, turns out to be an important physical
quantity appearing frecjuently in the study of mechanics. We digress
here to point out some of the features of moments and products of inertia
of plane areas. Later, in Chap. 11, we shall discuss the subject in
considerable detail.
The moment and the product of inertia of a given area can be computed
in general through the use of integral calculus. Their values for a few of
the simple cases are given as:
Rectangle Circle Triangle
r 6r T
'''" M!
^"v - 12 12
bhH2a + b\
K 2A
Values for otlicr geoniotri(r shajx's will \)v, found in A|)p('n(lix A.
t Tliis .soiiicwliat restrictive definition is part of a more neiK'ial flelinition to he dis-

If Ix and Ixy of a given area A are known with respect to a set of axes
X and y, we can compute the moment of inertia
Ix' and the product of
inertia Ix'y of A
with respect to a new; set of axes x' and ?/' which are
parallel to and displaced from the x and y axes (Fig. 5-9). The coordi-
nates of a small element of area dA are x, y, when referred to the original
t''
Sec. 5-1] Forces due lo fluid .sialic pressure: plane surfaces 93
and // =
f^
//'- dA =
j ^ (// + h)- dA
= 1^ if dA + 'Ih dA + Ir
j^ dA
f^,!,
= /. + 2byA + h-A (5-21)
Consider now the special case in which the original axes .c// are the
centroidal axes, meaning that their origin coincides with the centroid of
the given area. For this case x = y = 0, and Ecjs. (5-20) and (5-21) are
reduced to
x'y' = Izryr + dljA (5-22)
h' -^ I.r + h'A
where the subscript C is used to emphasize the fact that the original axes
can be only the centroidal axes if the above etjuations are to hold. Once
the original axes are designated by
xcyc, ^ve can afford to drop the prime
on the new axes x'y' and designate them
simply by xy (Fig. 5-10). Then, if x, y
are the coordinates of the centroid C
with respect to the new axes xy, we see
that X, y are just the (luantities a, h in
Eqs. (5-22). Hence Ecjs. (5-22) can be

rewritten
1 xu i xcyc I
XyA.
(5-23) Fic. 5-10
h = he + r-'^
Equations (5-23) are known as the parallel-axes theorems for the moment
and the product of inertia of an area. Sub.stituting Eqs. (5-23) into Eqs.
(5-19),
Xr
yA
+ x
(5-24)
yAf + ^
Three things might be mentioned in connec^tion with the above

equations:
1. If either xc or yc is an axis of symnietrv.
J zryr O and Xr = X
2. always positive and nonzero, //r is always larger than y.

Since I^r is
The center of pressure is therefore; always below the centroid of the

area.
3. The coordinates Xr and //r of the center of pressure may also be
interpreted as the centroid of the volume under the pressure diagram.
Example 5-4. Find the location of the resultant force on one side
of a rectangular plate submerged ver-
- - / tically in a liquid. The upper edge of
the plate is at the free surface (Fig.
5-11). Let the height of the plate be

h and its width be unity. Then
/
_ h
y = n
Sec. 5-51 Forces due lo JUiid sialic pressure: curved surfaces 95
of this vertical component passes tlinui^li the ccntioiil of tlu> al)Ove

described column of licjuid.
The first statement can readily be proved by considering a horizontal

tube of fluidbounded at the two ends by a small element of the curved
surface and its vertical projection (Fig. 5-12). If the axis of the small
tube is along x, the projected area is
parallel to or in the ijz plane. The condi-

tion of equilibrium along the x direction
requires that the x component of the pres-
sure force on the element of the curved
surface be equal and opposite to the pres-
sure force on the projected area of the same
element. This holds true for every other
element of the curved surface. Hence the
first statement is proved.
The .second statement can be proved by

considering the equilibrium of a vertical
Fi(i. rt-V.i
column above an element of the
of licjuid
curved surface and extending to the free surface (Fig. 5-18). The vertical
force on the element must be equal to the weight of the licjuid column.
Since this applies to every other element of area of the curviMl surface, the
second .âtement is proved.
Example A
sluice gate .1/^ ha\-ing the shape of a (luarter
5-5.
cylindrical surface hinged at one edge B (Fig. 5-14). Find the force
is
recjuired at the other edge -I to hold the gate in place.
Solution. Considering a unit width of the gate, the \-ertical forces on

the gate are equal to the weight of the column of li(|ui(l al)ove the gate.
(Wh(*ther the lifpiid is actually there or not is immateriai.) Therefore
the total vertical force is composed of the two forces (i*rob. 5-15),
Py^ = 'Yr{h — r), acting at a horizontal distance r/2 to right of B and

due to fluid bounded by OBCD
Fy^ = 77rrV4, acting at a horizontal distance r — 4r/37r to right of B
and due to fluid bounded by OB A
The horizontal force on the gate is
,.2
r
Fx = y h — -^] r, a.t a vertical distance .,„,, y^ + ^ below B
{
\ 2/ 12(/i — r/2) 2
Hence M^ about an axis through B due to the pressure forces is
('-0 12(/i - r/2)

"^
2
= 7r2 ( /i — r +
The force P required at A to hold the gate in place is
P = — - =yrlh — r-{-jr\ per unit width of gate
5-6. Buoyant Force; Stability of Floating Bodies
When a body is completely or partially submerged in a fluid, the

resultant of all the pressure forces acting on the surface of contact of the
body with the fluid is an upward force called the buoyant force. The
magnitude of the buoyant force can be obtained by the following simple
reasoning. Imagine that the submerged portion of the body is replaced
by the same fluid as in the surrounding medium. The fluid that now
occupies the space of the submerged body is in equilibrium with the rest
of the fluid. Hence the substituted fluid will experience a resultant pres-
sure force equal to its own weight. This resultant must also act through
the centroid of the volume of the substituted fluid for equilibrium. Now
the pressure forces acting on the surface of the substituted fluid are the
same as the pressure forces acting on the surfaces of the submerged body,
since the material inside these surfaces does not influence the pressure
forces acting on them. Hence we arrive at Archimedes' principle: The
buoyant force on a submerged body is equal to the weight of the displaced
fluid; it acts through the centroid of the displaced volume, which point is also
called the center of buoyancy.
A floating body is stable if its center of gravity is below the center of
buoyancy. Before defining the stability of a floating body we first
describe in a brief fashion the general concept of stability. By stable, or
more accurately stable equilibrium, we mean that if a slight disturbance is
imparted to a system in stable equilibrium, the system will return to its

Sec. 5-6] Buoyant force; slahilify of Jloaliiuj bodies 97
original configuration when the disturbance is removed. The sy.stem will

not return to its initial configuration if the eciuilihriuni i.s not stable. The
necessary condition (stated here without proof) is that the potential
energy be a minimum. f A usual example of these states of eiiuilibrium
is the cone. Supported on its ba.se it is in a condition of stable eciuilib-
rium, whereas when resting on its apex it is in a condition of unstable
equilibrium, for then a slight disturbance will cause it to tip over. On the
vy
G*I
B*
ib)
other hand, if the cone is rested on its side (along one of its generators),
the cone is said to be in a position of nrulral ((luilihrium.
By .stable, then, we imply that any slight departure from the neutral
position of the body willcause a net force, or moment, that tends to
restore the body to its neutral position. A floating body may or may not
be stable if its center of gravity is above the center of buoyancy. The
following analysis will show the reason why.
LetG be the center of gravity of a floating body of weight \V (Fig. 5-15)
and B its center of buoyancy. The body is in e(iuilibrium if undisturbed.
The buoyant force Fn is equal in magnitude to the weight of the body.
Now if the body is disturbed so that it rotates about its longitudinal axis z
t We .shall discuss this point further in See. Ki-2.
by a small angle dd, the displaced volume will increase by the additional
submergence on one side of the z axis and decrease by the same amount on
the other side. This results in an increase of buoyant force on one side
and a decrease on the other. Hence a restoring couple is formed which
tends to counteract the overturning moment caused by the lateral dis-
placement of the center of gravity. To evaluate this couple, let us
consider a small element of volume dV in the two wedges that represent an
addition of submergence on one side and a reduction on the other. The
element of volume is
dV = (x dd) dA
where dA is an element of area on the plane of the free surface. The

buoyant force on dV is 7 dV The moment of this buoyant force about
.
the longitudinal axis z is 7a; dV . The restoring couple is therefore
M, = f 7 {dd)x-' dA = yidd)I
where I is the moment of inertia about the z axis of the entire cross-
section area of thebody at the free surface. This restoring moment M,
combined with the original buoyant force Fb forms an equipollent single
force F'g, which may be called the new buoyant force. The location of
the new buoyant force is determined by the distance BB' (Fig. 5-15),
where
, M. M. y{dd)I I
^^ ^ V ^^
-f;-~w- ~w~
where V is the displaced volume of the floating body. The intersection
of the new buoyant force with the y axis is the point AI, called the meta
center. The distance BM is
sin do
For small dd, sin dO = dd. Hence
BM =
y (5-25)
Equation (5-25) gives the distance of the meta center from the center
of buoyancy. If BM
is larger than BG, the restoring moment is larger
than the overturning moment and the body is stable. On the other hand,
if BM is smaller than BG, the body is unstable. Finally, if coincides M
with G, the body is in neutral equilibrium. It should be emphasized that
/ in Eq. (5-25) is the moment of inertia of the cross-section area of the
floating body at the free surface. This cross-section area is, in general, not
the deck area.
Problems 99
FKOBLEMS
5-1. Determine the mass center of a ho niogenco us pyramid of base dimensions a
anil '' and height h.
Fig. P5-1 Vn.. 1
5-2. Determine the eentroid of the area inrhided helween the i)arahola tj =
(h/b")!", the X axis,and the Hne x = 6.
5-3. A small cylinder of radius r and height is made of material of uniform {h'nsity.
/*
However, there is a large empty spherical cavity w hose radius is r/2 located as showu.
Locate the center of mass.
I-k;. 1' o-:i
o-i. A beam is loaded by a distributed load, the intensity of load (force per unit
length) varying in a parabolic fashion. Uctermine the beam reactions. (For
Fi(i. I' Â
problems involving equilibrium of a rigid body, a distributed force may be replaced

by an equipollent force system.)
5-5. The mass of the earth is approximately 80 times the mass of the inoon. The
distance from the center of the earth to the center of the moon 240,000 miles. The
is
radius of the earth is 4,000 miles. Where is the center of mass of the earth-moon sys-
tem relative to the surface of the earth?
5-6. Orange juice is held in an essentially weightless conical Dixie cup of height h
and base radius r. After some time there is some sedimentation, so that the density
of the orange juice is given by 7 = 1 + ky, where A; is a constant and y is the distance
from the top of the cup. Locate the center of gravity of the juice.
5-7. A uniform circular plate is supported in a horizontal position by three vertical
bars symmetrically located around the circumference of the plate. The plate itself
weighs Wi. A second load W2 is applied at the mid-point of the line joining the cen-
ter of the plate to one of the supports. Find the reactions in the supports.
Fig. P 5-7
5-8. Knowing that the surface area of a sphere of radius r is 47rr^ find the location
of the centroid of a semicircular arc by the theorem of Pappus.
5-9. Find the center of gravity of a slender wire OABCD consisting of three straight
portions OA, CB, and CD, each of length a, and a fourth portion in the shape of a
quarter circle of radius a. Assume the wire is of homogeneous material.
Fk;. I' 5-<)

Prohlenn 101
5-10. A somicylindrical shell of uniform wall thickness weighs wi Ih/lin ft along the
direction of the axis of the shell. It carries along its one edge a linearly distrihuted
load of Wi lb/ft. ?'ind the relation between the ratio «'i/h'j and the angle of inclina-
tion a when the shell rests on a horizontal table.
/////////7///////////
Fig. P 5-10
5-11. A semispherical shell of uniform wall thickness weighs H'l lb. It carries at
one point of its periphery a concentrated load of i lb. W

Find the relation between the
ratio IFi/U'o and the angle a when the shell rests on a horizontal table.
of inclination
5-12. The atmospheric pressure is 14.7 psia, and the specific weight of air is ().U70o
lb/ft^ both at sea level. Calculate the pressure at 10,000 ft elevation assuming that
(a) the atmosphere is isothermal; (6) the atmosphere follows the poly tropic relation,
with the exponent n equal to 1.4. Compare these values with the value that one
would get if one made the rough assumption that the air is incompressible and has a
constant 7 equal to 7 at sea level.
5-13. The with water is to be measured by a simple
pres.sure in a pipeline filled
mercury manometer. When the left meni.scus is 3 in. below A, the scale reads zero
and p^ = 5 psig. Find the reading of the right meniscus for p.4 = 10 psi if the U
tube and the scale are both left undisturbed.
Fk;. 1' 5-i;i
5-14. A pipeline contains natural gas, 7 = 0.045 lb/ft'. At one point a simple
water manometer shows the pressure to be 6 in. of water. What is the pressure in
inches of water at a iM)int in the line 500 ft higher than the first pf)int? Take 7 for
air as 0.07G5 Ib/ft^ and neglect the compres.sibility of air and gas for the relatively
small difference in elevation in this problem.
5-15. Verify the magnitude and location of the pressure forces acting on the sluice
gate of Example 5-5.

5-16. A rectangular gate AB
separates two tanks of water. The level in the left
tank is 10 ft above the top of the gate, and the level in the right tank is 8 ft above the
gate. The gate is 8 ft high and 1 ft wide perpendicular to the paper and hinged at
the top A. Calculate the force (magnitude and direction) that must be applied at the
bottom B to hold the gate shut.
10'
8'
I
Fig. P 5-16 Fig. P 5-17
5-17. Locate the resultant force on each end of the cylindrical drum of 3-ft diameter
and 5-ft length immersed water as shown. Also find their respective magnitudes.
in
5-18. At one section of Grand Coulee Dam on the Columbia River, the dimensions
are approximately as follows: The area ABD is 13,750 ft 2, and its centroid Ci is 22.3 ft
to the right of B. The vertical distance DB is 430 ft. The area of the cross section
of the concrete dam is 98,000 ft^; its centroid, C2. (a) Compute the position where the
resultant force exerted by the ground on the base of the dam crosses the base, (b) If
the base BE of the dam is 400 ft, find the maximum and minimum stresses (intensity
of pressure) at the base (assume a linear stress distribution along the base).
430'
Problems 103
5-19. A plane gate woiphing; 500 lb/ft of length perpendicular to the pai)er is
hinged at one end O. The width of the gate is 10 ft, and its center of gravity is in
the middle. Find the (l(>pth li of water as a function of d for e(|uilii)riuni of the gate.
What is the range of e for stable equilibrium?

5-20. A in. in diameter, sp gr 0.60, has a 50-lb metal piece attached to its
pole 4
submerged end. Neglecting the volume of the metal piece in the problem, determine
the range of lengths of the pole for it to float vertically in stable e(|uilibrium.
5-21. A pontoon bridge is to be constructed from ga.soline drums 2 ft in diameter,
with axes vertical. The drums are spaced on 5-ft centers along each edge of the
l)ridge and are submerged 2.5 ft when the l)ridge is loaded. W hat is the minimum
distance d required for stability if G is 5 ft above B'?
'
1 1 1
-
CHAPTER 6
Derivatives of Vectors;
Kinematics of a Particle
6-1. Derivative of a Vector

Avariable y can be a function of another variable x, expressed by
y = = y{x), where ordinarily both y and x are scalar quantities.
f{x) or y
Similarly, a vector variable A can be a function of a scalar variable t,
meaning that A can change in magnitude and direction when t changes.
At a given value of t the function X{t) is said to be continuous if A(i A^ +
-^ K{t) as M
-^ 0. In Fig. 6-1, the vector
A is plotted as a position vector from the
origin. For successive values of t, the
terminal of A traces a continuous curve C
in space. Let us now pick two values of
A: A(t) and A(^ A^). +
Because of the
difference A^ in the values of t, the two
values of A also differ by an amount which
AU+ At) we shall call A A. Thus
AA = A{t + At) - A{t)
Fig. 6-1
The ratio AA/At is the average change in
A per unit of t during the increment At. As At approaches zero, we assume
that AA/At approaches a limit. (A is then said to be differentiable with
respect to t.) We use the notation
dA
-7-
,.AA
= hm — (6-1)
at Ai-»o At
and call it the derivative of A with respect to the scalar variable t. It

can be seen that dA, being the limit of the chord AA, is along the tangent
to the curve C. Hence dA/dt is also a vector and is tangent to the curve
104
Sec. 6-2] Derivative of sums and products 105
C in the direction of increasing t. Since dX UU is again a vector function

of the scalar variable assuming d\/dt is a ditTerential)le vector function,
t,
we can form the derivative of dX/dt and obtain the second derivative of A,
which again is a vector function of the scalar variable t.
d (dX\ _ d'-X
dt\~dt) - w^ (^-2)
Similarly, thirdand higher derivatives can be formed.

Thus the deri\ative of a vector function with respect to some scalar
variable measures the rate of change of both magnitude and direction of
the vector function with change of scalar variable, ^'ariation of either
magnitude or direction of the vector function is non-
sufficient to yield a
zero derivative. Accordingly, should be noted that a vector is a
it
constant, and hence its derivative is zero, only when both its magnitude
and direction are constant.
6-2. Derivative of Sums and Products

Let C be the sum of two vectors A and B, both of which are differentia-
ble vector functions of t, When changes by an amount
a scalar variable. /
At, then A, B, and C change by an amount AA, AB, and AC, respectively.
The new value of C is therefore
C + AC = (A + AA) + (B + AB)
Hence AC = AA + AB
Dividing each term by the scalar cjuantity At and passing to the limitf as
AtÔ yields
dC ^ d{X + B ^dX
) dB
dt dt ~ dt ^ dt ^ '
This result can obviously be extended to the sum of any number of

vectors. Hence the differentiation of vectors is distributive, just as is the
differentiation of scalars.
Next let f{t)X{t) be the product of a differentiable scalar and vector
function of t. When
/ changes by an amount At, then /, A, and the
product /A change by the amount A/, AA, and A(/A), respectively. The
new value of /A is therefore
/A + A(/A) = (/+A/)(A + AA)
Hence A (/A) = /AA + A/A -\- A/ A A
t It Ls asHumc'd here, and in what follows, that the vector fuiictioiis are of such a
form as to ensure that the hniit of a .sum is equal to the sum of the limits and that the
limit of a product i.s ecjual to the product of the limits.
:
106 Derivatives of vectors; kinematics of a particle
Dividing by A^, passing to the limit At -^ 0, and neglecting the second-

order term, we obtain
This is again similar to the differentiation of a product of two scalar

functions.
An important application of Eqs. (6-3) and (6-4) is the differentiation of
a vector when the vector is expressed by its components in rectangular
coordinates
A = ^xi + Ayj -\- A^k
Here A is the sum of three vector components, which are in turn the
product of a scalar and a vector. Performing the differentiation of A,
we obtain, according to Eqs. (6-3) and (6-4),
dA _
~ dAx d\ dAy d^ dA^ dk
It 'dt^'^^^'dtîir^
• \ A .
^ ^''dt^~dr^^ ^'H
_, a >
%^ _, *
Now, since the unit vectors i, j , k are constant in magnitude as well as in

direction, their derivatives are zero. Hence the derivative of a vector
with respect to some scalar variable, when the vector is expressed in terms
of its rectangular components, becomes
dA dAx . . dAy . dA^ ,
or the components in rectangular coordinates of the derivatives of a vector are

equal to the derivatives of the components of the vector.
The same procedure used in obtaining Eq. (6-4) can be followed to
show that
|(A.B) = A.f + §.B (6-6)
The proof of the above depends on the distributive law for the dot and the
cross products of vectors. In particular, if B = A,
= 2A.!^ = 2.1^ (6-8)

S(^-^'-J<(-^'>
. dX , dA
Moreover, if A is a vector of constant magnitude,
A-^
dt
=0
which implies either that dA/dt = or that A is perpendicular to dA/dt.
Soc. 6-3] Curves in space: princl/ml iiorniuls <in(i hinorntals 107
Kxaniple 6-1. Let r = a cu.s pt i + b sin />/ j, where a, h. and p are

all scalar constants. Then
(Ir
dt
ap sin pt i + hp cos p^ j
d'-r
-Tj-o = —ap- cos pt i — bp- sin pt j
dt
From this we see that
p-r
dt'-
which implies that r = a cos pt i -\- b sin pt j is a solution of the vector

equation
differential
6-3. Curves in Space; Principal Normals and Binormals
Let C be any curve in space and Po a reference point on C. Let P be any

other point on C and r = .ri + + zk the position vector from the origin
//j
Fic. ()-2
to P. Let the numerical value of the distance along C from Po to P be s

(Fig. 6-2) hence s is a scalar (juantity. If we arc given the curve C and
;
the reference point Po, the location of F is a unicjue function of s; hence

r = r{s). The derivative of r, according to Eq. (G-5), is
dr _ dx . dij . , dz .
(6-9)
ds ds ds ds
We see that dx/ds, dy/ds, and dz/ds are the direction cosines of the tangent
to the curve at the point P; hence dr/ds is along the tangent to the curve in
the direction of increasing s (as was also pointed out in Sec. 6-1, where a
general scalar variable t was used instead of the particular variable s).
Furthermore, since
ds = -y/dx^ + dy^ + dz^
the magnitude of dr/ds, given by Eq. (6-9), is unity. Thus the derivative
of a position vector to a space curve, with respect to the distance s along the
curve, is a unit vector tangent to the space curve in the direction of increasing s.
We denote this unit tangent vector by e^. Thus
Since e^ • e^ = 1, a constant, differentiation with respect to s yields
det ^
ds
Hence the vector det/ds = d-r/ds^ is perpendicular to the unit tangent

vector Ct. The magnitude of det/ds represents the rate at which the unit
tangent vector changes its direction with respect to the distance along the
curve. This is the curvature of the curve. The reciprocal of this value is
the radius of curvature, denoted by p. Hence
P = -en
ds p
(6-ll)t
where e„ is a unit vector along the direction of det/ds, called the direction
of the principal normal.
The direction of the principal normal can be visualized as follows : The
vector det/ds is zero only when e< is a constant, in which case the curve C
is a straight line. In general for nonzero det/ds, we can pick three points
on C and call themP, Pi, P2, where Pi, P2 are two points
at each side and in
the immediate neighborhood of P. These three points determine a plane.
When we let Pi and P2 approach P as a limit, the plane becomes the
osculating plane of C at P. The unit tangent e^ and its derivative det/ds
both lie in the osculating plane. The direction of the principal normal is
thus fixed as follows It is perpendicular to the tangent of the curve and
:
lies in the osculating plane. It should also be noted that the unit principal
normal e„ points to the center of curvature (Fig. 6-2).
t Equation (6-11) is the first of three equations called Frenet's formulas, which are
of great importance in differential geometry. The other two of Frenet's formulas will
be developed in Example 6-2.
Sec. 6-3] Curves in space: principal nornuils and hinorfnals 109
The vector product

et X e„ = eft (6-12)
defines a unit vector perpendicular to both e^ and e„. The unit vector Cb
is called the unit binormal. The three mutually perpendicular vectors e^,
Cn. and Cfc form a moving trihedral, "moving" because all three continu-
ouslj^ change their directions along the curve.
Example 6-2. Since e5 is a unit vector, dcb/ds is a vector that is
perpendicular to Cb and thus lies in the plane of e^ and e„, the osculating
plane. Furthermore, because e^ • ej = 0,
dcb dci
-y-
ds
•
Ct + .
e6 • -y-
ds
=
Using Eq. (6-1 1) and the statement e^ • e„ = 0, this becomes
dcb ,
1 dcb f.
ds p ds
Therefore dcb/ds is also perpendicular to ct and must necessarily be a

vector parallel to e„. We can now write
^=
ds
re„ (6-13)
where t is the magnitude of dcb/ds and is called the torsion of the space
curve C.
An examination of Fig. 0-2 reveals that
e„ = Cft X et
from which it follows that
de„
= ^ dni . dcb
X
-5— Cb X -. \-
-J- Ct
ds ds ds
-7- = - efc X <',. + re„ X Cj

ds p
Thus, noting that e^ X e„ = — e^ and e„ X e^ = — Cb,
de„
ds
-(^ê. + re.j (0-14)
Equations (6-11), (0-13), and (0-14) constitute the complete set of

Frenet's formulas.
Example 6-3. Given .some plane curve C and a position vector
defined by
r = x{t) i + y{t) j
Find an expression for the curvature 1/p.

Solution. From Eq. (6-10) we find that
dr , dt
where the superscript dots indicate differentiation with respect to the

scalar variable
t. The quantity dt/ds can be evaluated by recognizing
that
e,.e,= l = (i^ + r)(|y

from which i- = (ac^+ ^^)~''^^
as
Then Ci = (ii + 2/j)(x2 + 2/2)-^^
Now, using Eq. (6-11), we obtain after some simplification
^^ 1 ^ y{xy — xy)\ -h x{yx — p)j
The curvature 1/p, which represents the magnitude of det/ds, can now be
found to be
1 _ yx — xy
In particular, if the plane curve C is expressed a,s y = f(x), we can let

X = t, from which it follows that
. dy dy dx dy
X = I y
.
= -17 = -T- -T7 = -r

dt dx dt dx
dx _ d^y
dt \dx/ dx \dxj dt dx^
The curvature is then given as
1 d^y/dx'~
p [1 -1- idy/dxyf'
6-4. Displacement, Velocity, and Acceleration of a Particle
Let r be the position vector of a particle and t be the time. The

extremity of r traces a curve which is the path of the particle. The
velocity of the particle is defined as the rate of change of the position
vector with time; hence
V = (6-15)
^;
Velocity, therefore, is a vector quantity whose magnitude, the speed, is
given by \dr/dt\ and whose direction, according to Sec. 6-1, is tangent to

Sec. 6-5] Kinematics of a parlicle: rectangular coordinates 111
the path of the particle. The acceleration oi the particle is defined as the
rate of change of velocity with time; hence
dy d-r
In general, the direction of the acceleration vector is inclined to the

tangent to the path of the particle.
It cannot be overemphasized that r, v, and a, representing the position,
velocity, and acceleration, are all vector quantities having both a magni-
tude and a direction. A change of either magnitude or direction of a
vector is make the derivative of the vector nonzero.

sufficient to
Following conventions, we shall use a dot on top of a quantity to
represent its derivative with respect to time. Having two dots on top
of a quantity means the second derivative of the cjuantity with re.spect
to time.
6-5. Kinematics of a Particle: Rectangular Coordinates
The position, velocity, and acceleration of a particle in the rectangular

coordinate system are [cf. Eq. (6-5)]
r = x'i + ;/j + zk
V = r = ii + ?/j + ik (6-17)
a = V = r == .ri -\- i/j -\- zk
What we have achieved by expressing the velocity v and acceleration

a in terms of rectangular components [Eqs. (6-17)] is very illuminating.
We note that, for example, x, x, and x represent the scalar x components
of the position, velocity, and acceleration, respectively. In turn, the.se
quantities represent the motion of the particle projected on the .r axis
and as such may be conceived to be a rectilinear motion, i.e., a motion
along a straight line. Thus we have succeeded in taking a complicated
three-dimensional motion and decomposing it into a system of three
much simpler rectilinear motions, each in the direction of a coordinate
axis, the vector sum of which represents the complete spatial motion.
Much general information relative to the type of motion may be
obtained from these expressions [Eqs. (6-17)] without recourse to supply-
Consider a typical rectilinear component,
ing explicit analytical details.
say along the x axis. Here we have y^ = x = dx/dl and a^ = dvjdi.
Interpreted geometrically, these expressions tell us that the speed /'^ is
efjual to the slope of the x-t (position-time) curve and that the scalar
acceleration a^ is equal to the slope of the vA (speed-time) curve.
Rewriting these relations, we have
dx = Vj, dt and dv^ = a^ dt
Integrating between times ti and ^2, we now obtain
X2 — Xi = Vxdt
I
Vx, — Vx, = /
"
ttx dt
These expressions, interpreted geometrically, indicate that the change of

position (displacement) in any time interval equals the area under the
Vx-t curve in that same time interval,
and that the change of speed in any

time interval is equal to the area
under the ax-t curve during the time
Pii.i)
interval. Similar relations also ap-
ply in the y and z directions. These
\
propositions enable us to construct
\
\
the so-called motion curves when
\ the original data are of a numerical
\ rather than an analytical nature.
Example 6-4. A particle moves
o
with a constant speed v along the
Fig. 6-3
parabola ij = x'^ (Fig. 6-3). Find
the acceleration of the particle when it passes through point P(l,l).
Differentiating the given equation y = x~ with respect to time t,
y = 2xx (a)
But x^ -{- y^ = v^
Hence
1 -\- 4x2
Differentiating with respect to t and rearranging terms,

.. _ -4.^2
^ ~ (1 -1- 4.r;2)2
Differentiating Eq. (a) and substituting for x and x by the results just
obtained,
y =
= 'Zxx -{- 'Zx- = -j—
(i
At point P,
Hence
Sec. 6-6] kinematics of a parlicle: tangential, normal components 113
The magnitude of a
is 2v-/{b \/5) and the direction of a makes an angle
of tan~i _i..^ =
153°26' from the positive x direction. As an exercise,
the student should verify that the direction of v is tangent to the path
of the particle whereas the direction of a is not tangent.
6-6. Kinematics of a Particle: Tangential and

Normal Components
Referring to Fig. 6-2, let a particle travel from P» to P along a given
curve C. If s is the distance from Po to P, it was shown in Eq. (G-10) that
rfr
_
ds
where cj is the unit vector along the tangent to the curve C. Hence the
velocity of the particle is
dr dr ds . ,,. ,„.
'' = 'dt
= lsJt = '^^ ^^-^^)
or the velociiij is a vector along the tatigeut with the magnitude of .<-.
The acceleration of the particle is obtained from Eq. (G-18) by the use
of Eqs. (6-16) and (6-4). Thus
a
Since the particle travels with a constant speed v,
s — V
s =
For the given curve y = x^, dy/dx = 2x, and d^y/dx- = 2. The curva-
ture is, from Example 6-3,
1 _ d'^y/dx'^ 2
p
~ [1 + {dy/dxYf^
~ (1 + 4a;2)?
2v^
Hence ^"
(1 + 2vy2
4x2)^
At the given point P, a; = 1 and
2v'
5 V5
The direction of a is therefore along the principal normal. Since the
slope of the curve at P or tan 63°26', the direction of the principal
is 2,
normal makes an angle of 90° +

63°26' = 153°26' with the positive
X direction. This completely checks the results of Example 6-4.
6-7. Kinematics of a Particle: Cylindrical Coordinates

The cylindrical coordinate system consists of the following three
mutually perpendicular directions: one parallel to the z axis, one along
the radial direction from the z axis in
a plane parallel to the xy plane, and
one along the direction of increasing </>
in a plane parallel to the xy plane (Fig.

6-4). We shall use Bz, Cr, and e^ to
denote the unit vectors along these
three directions. In the xy plane
alone, the cylindrical coordinates be-
come 'polar coordinates. For a plane
circular path, cylindrical coordinates
coincide with the tangent normal coor-
dinates in direction alone and polar
coordinates in both sense and direction.
One readily sees that among the
p „ . three unit vectors, only e^ is a con-
stant; the other two unit vectors e^
and e^ change in their directions with changes in <^. Although one may
5
Sec. 6-7] Kinemafics of a particle: cylindrical coordinates 1 1
use K(is. (6-10) and (()-ll) to ol)tain the derivatives of Cr and e^.f the
following direct method is perhaps more ilkistrative. From Fig. (i-o,
which shows the eê^ plane, it can be s-een that lêcause of a dilTerence of
A0. the unit vectors e^ and e.^ change by an amount
Acr = (A0)e^ Ae^ = -(A</))er
Dividing by A<^ and taking the limit as A0 approaches zero,
dcr de^
— Cr (0-21)
d^ ~d^
We now consider the change of Cr and e^ with respect to time. Applying

the operator c^/rf^ = {d/d){d4>/d{) = 4>{d/d4>) to Eqs. (6-21), the result is
Cr — (j>e^ e^ = —4>er (6-22)
In order to represent the kinematics of a particle in the cylindrical

coordinate system, we start with the
position vector of a particle
rCr + ZCz (6-23)
Differentiating Eq. (6-23) with respect

to time to obtain velocity, we obtain,
after making use of Eqs. (6-3), (6-4),
and (6-22),
V = rer + r0e^ + ie^ (6-24)
We see that the velocity vector con-

sists component r, a tan-
of a radial
gential component and an axial com-
r4>,
Fi(i. (i-o
ponent z. The scalar quantity ^ is the
angular speed of P about the z axis. Differentiating (0-24) with respect
to time, we obtain the acceleration of the particle:
a = fcr + rê.<t, + '<ê^ +r<ê<(, — r^-Cr + zcz

= (r - r4>-)er + + 2i-<j>)e^ + ze,
(r(/> (6-25)
Hence the acceleration in the cylindrical coordinate system consists of
t Considering the unil circle of Fig. G-5 as the curve of Fig. G-2, we see that here
e^ e„ = — Cr = — r ds = d<t> p = 1
Hence -7-^
d<t>
=
dr
'
-r
ds
= c<t> hy Eq. (0-10)]
J^*
= ^= -c. [by Kq. (O-Il)]
These results are the same as Eqs. (6-21).

: :
the following scalar components
Radial: f — r4>^
Tangential: r^ + 2r4> (6-26)
Axial z
The tangential component of the acceleration can also be written as

(l/r){d/dt)(r-4>), the significance of which will be discussed again in
Chap. 10.
Example 6-6. A particle moves
with constant speed v along a helix.
The equations of a helix (Fig. 6-6) in
cylindrical coordinates are
r — constant
where p, the pitch of the helix, is again

a constant. Calculate the compo-
nents of velocity and acceleration.
Solution. The velocity is obtained
from Eq. (6-24) as
r(j)e4
Fig. 6-6 ^27r
where <j), the angular speed of the particle, is given by
The acceleration is obtained from Eq. (6-25) or (6-26) as
V'r
a = —rcp-Cr = —
r2 + (p/27r)2
It is noted that in this problem r = f = (^ = s = 0.
6-8. Kinematics of a Particle: Spherical Coordinates

The mutu-
spherical coordinate system consists of the following three
ally perpendicular coordinates: the radius r measured from the origin 0,
the longitude measured from the xz plane, and the colatitude 6 meas-
</>
ured from the z axis. Positive directions of these coordinates are shown
in Fig. 6-7. It is easy to verify that any point in space is explicitly
described by this coordinate system. We shall use the unit vectors
Cr, e^, and e^ to denote the three orthogonal directions of the coordinates.
Sec. 6-8] hinenialics of a particle: spherical coordinales 117
Fig. 6-7
Fig. 0-8
.
Since the directions of the unit vectors are not fixed in space, it is
necessary to determine their true rate of change before defining the

first
kinematics of a particle. There are several ways of accomplishing this;

we choose the more instructive method and leave two other methods as
exercises for the reader (Probs. 6-2 and 6-3).
Suppose we first orient the three unit vectors on a unit sphere (Fig. 6-8)
Now consider the effect of an incremental change in colatitude, Ad.
Because of this change, the unit vectors e^ and ee undergo a change in
the meridional plane Oah. An examination of the meridional plane (Fig.
6-9) indicates the magnitude of these
changes, which can be expressed as
Ae^ = Ad ee
Ae0 = —Ader
Next, consider the change in unit
vectors due to an incremental change in
longitude A<^. As indicated in Fig. 6-8,
all three unit vectors suffer a change of
direction; Cr moves along a horizontal
circle of radius sin 6, ee moves along a
horizontal circle of radius cos 6, and e^

displaces along a horizontal circle of
radius unity, the equator. From the
geometry of Fig. 6-8, we immediately
find that
Acr = A(f) sin 6 e^
Aee — A0 cos d e^
Fig. 6-9
The change of e^, Ae0, has components
in the directions of both e^ and e Looking along e^ [i.e., at the meridional
plane (Fig. 6-9)], it can be expressed as
Ae^ = —A(f) sin ^ e^ — A0 cos 6 ee
Adding each of these separate effects together, we obtain the total

change in the unit vectors:
Acr = A0 ee + Acf) sin 6 e^

Aee ^ -Ad er + A(t> cos 6 e^ (6-27)
Ae^ = — A0 sin 6 er — A0 cos 6 ee
Now, dividing by the scalar variable At and passing to the limit as

At —^ 0, in the usual fashion, we obtain
e^ = dee + $ sin 6 e^
eg = — 6er + cos 6 e^
4> (6-28)
e^ = — cos 6 ee — sin
(j) 4> 6 e.
Problems 119
In order to define the kinematics of a particle in the spherical coordi-

nate system, we start with the position vector of a particle,
r = rcr (6-29)
Differentiating Eq. (6-29) with respect to time to obtain velocity and

making use of Eqs. (6-4) and (6-28), we find
V = i-Cr -\- rdeo -\- r4> sin d e^ (6-30)
In a similar fashion, upon diifcrentiating the velocity witli respect to

time, we obtain the acceleration
a = (/• — rd- — >•(}>'- sin- d)er
-I- (2/-^ -\- rd - r<j>'^ sin 6 cos d)ee
-f (2r0 sin 6 + r4> sin 6 -{- 2r^ 6 cos 0)e^ (6-31)
PROBLEMS
6-1. Dorivo the compononts of acceleration in the cj'liiulrical coonnnate system
without using vectors. I'roceed as follows:
(a) PVoni the transformation equations, x = r cos 0, ?/ = sin find x and ij in /•
</>,
terms of f, r, and
ij), .
(b) Find the components of x and y along Cr and c^. This gives Or and a,p, which
should check expressions (6-26).
6-2. The transformation equations from spherical to rectangular coordinates are
given as X = r sin d cos y — r sin d sin z = r cos d.
(j>, Proceeding as in Prob. 6-1,
<f>,
find the velocity and acceleration components in the spherical coordinate system.
6-3. Verify Eqs. (6-28) by first expressing the three unit vectors of the spherical
coordinate system in terms of three cartesian unit vectors and then determining their
rate of change.
6-4. If r = a cosh pt \ — b sinh pt j and a, b, and p are .scalar constants, find the
first and second derivatives of r with respect to the scalar varial)le I. What differen-
tial equation is satified by r?
6-5. Find a vector differential equation that is satisfied by
(a) r = Ae'-î (b) r = Ae'"i + He-"'}
6-6. Given a plane curve ?/

= 1 -(-2x4- '-ix^ + Ij-'. Determine tiie curvature of
this curve at the point /'(1, 10). liocate also the center of curvature.
6-7. A particle is traveling along the x axis so that its distance from a fixed origin
as a function of time is given as x = a + bl + cl^. Find the velocity and acceleration

of the particle. Plot the motion curves.
6-8. the motion of a particle traveling along the y axis is such that y = —ae''\
If
determine the velocity and position of the particle. At t = 0, y = yo and y — v^.

6-9. F'or the two separate motion curves shown, determine the corresponding
remaining motion curves. In case a the particle starts from rest.
(fps^) 200
20 25 10
35 45 t (sec) 6 t (sec)
40
(a) (6)
250
Fig. P 6-9
6-10. A particle is traveling around the parabolic curve of Example 6-4 with a
constant x component of speed, I'x. Determine the velocity and acceleration of the
particle. Use both the rectangular and tangent normal coordinate systems.
6-11. A rocket-propelled sled travels along a straight track. It starts from rest at
t = 0, and at any subsequent time its speed is given as z; = at^ -\- bt. At time to,
"burnout" (propulsion ceases) occurs. Subsequently, brakes are applied to stop the
sled, the brakes producing a constant deceleration of a = — c. How far has the
rocket traveled during propulsion, what is its speed at burnout, and how much
farther must the rocket travel in order to come to rest?
6-12. An automobile is traveling along a highway whose profile may be approxi-
mated by the If the vehicle is traveling with constant speed, what
sine curve shown.
is its and acceleration? Suppose the vehicle has a constant x component of
velocity
speed; what is its velocity and acceleration? Assume that the vehicle remains in
contact with the surface.
2 I
2 I
Fig. P 6-12 Fig. P 6-13
6-13. A rod AB of length / slides down in the xy plane while keeping its two ends
A and B on the x and y axes, respectively. If the lower end A travels to the right
with a constant velocity va, find the acceleration of the upper end B and of the mid-
point C. What is the path of C?
6-14. A particle moves with constant velocity along a straight line parallel to the
X axis. Find (a) r and ^ in cylindrical coordinates; (6) the components of acceleration
in cylindrical coordinates.
Problems
Fi<i. P G-14 Fk;. 1' (Vl;

I
6-15. A particle travels with constant speed r along an Archimedean spiral r = a<t>.
Find the normal and tangential components of acceleration and

(a) (h) the acceler-
ation in cj'lindrical coordinates.
6-16. Find the velocity and acceleration of the piston in a reciprocating engine.
Assume the crank rotates with the constant angular speed 6 = u. The crank is of
length a, and the connecting rod of length /. (Hint: First find the geometrical rela-
tion between the two variables, the piston displacement x and the crank angle 6.)
Fi<;. P G-16
6-17. In the above problem, if / is much larger than a, find an

approximate solution
for the velocity and acceleration. (Hint: Make
use of the binomial series expansion.)
6-18. A special type of reciprocating mechanism is the "Scotch crank." Find the
velocityand the acceleration of the piston for a constant angular crank speed oi 6 = u.
Explain how this mechanism is related to the general case of Prob. ()-16.
Fig. P 6-18
6-19. A quick-return mechanism consists of a crank A, an oscillating arm B, a

connecting link C, and a sliding block D. The dimensions are as shown. The crank
rotates at a constant angular velocity d = w.
(a) Find the two crank angles 9 for the two extreme positions of D.
(b) Find the ratio of the time during which D travels to the left (return stroke) to
the time during which D travels to the right (advance stroke).

Fig. P 6-19
(c) Form the relation between the displacement of D and the crank angle B. (This
relation need not be
(d)
an
Find the velocity of
explicit one.)
Z) as a function of 6. For 6 = 30° and w = 120 rpm, what
I
is the velocity of D?
(e) Indicate the means of finding the acceleration of D.
CHAPIER 7
Kinematics of a Rigid Body;

Relative Motions
7-1. Displacements of a Rigid Body
The most rigorous treatment of the kinematics of a rigid body can be

quite complicated. In the next section, an attempt willbe made to
derive the important conclusions without too much mathematical com-
plexity. These conclusions are stated in four theorems. The reader is
advised to familiarize himself with these theorems and their significance,
for they give us the means for solving, analytically or graphically, the
motion of even the most complicated mechanisms in engineering. How-
ever, before embarking on this program it is well to describe briefly the
concept of the displacements of a rigid body.
As was stated previously (Sec. 4-2), a rigid body is a continuous dis-
tribution of particles constrained in such a fashion that the distance
between any two of these particles remains unaltered regardless of the
change of position of the body as a whole. In order to completely specify
the position of the rigid body, we need at most to describe the coordinates
oi any three noncolinear points. Because of the rigid-body constraint,
however, these coordinates are not all independent, and the number of
independent quantities necessary to specify the position of the rigid
body is thereby reduced to six.
Before we go into the description of the most general displacement,
i.e., change of position of a rigid body, let us con.sider first the following
special types of rigid-body displacement that are relatively simple to

visualize. Translation is that type of displacement in which each point
undergoes the same vector displacement. Physically, then, if a straight
line were scribed on the body, during translation the straight line would
move parallel to itself. Rotation about a fixed axis is that type of dis-
placement in which points on the line of rotation remain fixed and all
123
124 Kinematics of a rigid body; relative motions
other points travel circular paths with centers on the axis and in planes
perpendicular to the axis of rotation. Finally, if there is one point in
the body, say point 0, that remains fixed in position, any displacement
of the body
is said to be a rotation about the point 0.
It is a remarkable fact that if a rigid body is displaced, keeping one

point fixed in position, all points on some line undergo zero displacement,
meaning that they either remain fixed or else return to their original
position. This fact is expressed by Euler's theorem, which can be stated
as follows The most general rotation of a rigid
:
body about a point is equivalent to a rotation

about some axis through that point. Accord-
ing to this theorem, although the actual
movement of the rigid body may consist of
a number of different rotations about the
fixed point, the displacement of the body
from its initial position to its final position
(whatever that final position may be) can
always be accomplished by a rotation about
a particular axis passing through the fixed
point.
The proof of Euler's theorem is as follows:
Fig. 7-1
In Fig. 7-1 let be the fixed point and let A
and B be any other two points of the rigid body before displacement.
After displacement A moves to A' and B to B'. The three points A', B',
and completely define the displaced position of the body. (In Fig. 7-1,
A and A' are connected by a circular arc with as center. Similarly, B
and B' are connected by another circular arc. The arcs A A' and BB' are
drawn merely for convenience. They do not necessarily represent the
actual paths of movement of points A and B.) Obviously, for a rigid
body OA = OA', OB = OB', and AB = A'B'.
Now erect a plane (plane aOC) perpendicular to plane AOA' and
bisecting the angle AOA'. Similarly, erect a plane (plane bOC) per-
pendicular to plane BOB' and bisecting the angle BOB' These two new
.
planes intersect along a line which is represented by OC. (The line OC is

not necessarily perpendicular to the planes AOA' or BOB'.) Since C lies
on the bisecting plane AOC, the angles AOC and A'OC are equal. Simi-
larly, since C also lies on the bisecting plane bOC, the angles BOC and
B'OC are equal. Thus the relative positions of the four points ABOC are
identical with the relative positions of the four points A' B'OC.
Hence,
with OC fixed in space, a rotation aliout OC by an appropriate angle will
displace .1 to A' and B to B'. The appropriate angle is, of course, the
angle AOA'. Therefore the original displacement can be effected by a
rotation about OC alone. This proves Euler's theorem.
See. 7-1] Displacemenis of a rhjid body 12'
An immediate and almost theorem

self-evident corollary of the Euler
is the followinjj; theorem due to Chasles: The most general finite din place-
ment of a rigid hodij is equiralent to a'translation of some point {called base
point) plus a rotation about an axis through that point. Suppose we choose
a base point .4 on the rigid body. By a translation we can bring the
original position of A into coincidence with displaced position A'. Then,
by Euler's theorem, we can obtain any subsc(iucnt ilisplaccmcnl we
desire by a suital)le rotation about some line through .1'.
The choice of the base point is completely
arbitrary. For each different choice of base

point the translation part of the total dis-
placement will differ. However, since the
change of orientation of any line in the body
is determined solely by rotation, the rotation
part of the total displacement will remain

unaltered with respect to both the orienta-
tion of the axis of rotation and the angle
through which the rotation takes place.
(Although the rotation involved is the same
in the way just described, the location of the
axis of rotation will be different with each
different choice of base point, since by defini-
tion the axis must pass through the base
point.)
Suppose now we consider an infinitesimal
rotation of pointP about .some line through Fic. 7-2
a fixed point 0. During the infinitesimal
rotation through an angle Ad, the point P travels in a plane perpendicular
to the axis of rotation (Fig. 7-2) to .some new position P'. The magnitude
of the displacement is, then.
lArl = Irl sin Ad
Since Ar is normal to the plane OPQ,
Ar = An X r (7-1)
where we have used the vector An (to be justified immediately) to rei)re-

sent the infinitesimal angle of rotation such that the magnitude of An is
numerically ecjual to Ad, the direction of An is the direction of the axis
of rotation, and the sense of An is determined by the right-hand-screw
rule.
Although we have seen that we can represent an iiifinitcsinial rotation
by a vector (juantity, this alone does not justify tlu; fact that it is a
vector. Indeed, to fully qualify as vectors, infinitesimal rotations must

combine according to the parallelogram rule. In other words, we must
show that if the point P suffered two infinitesimal rotations, Arii and An2,
with both axes through the fixed point 0, then the total displacement
would be
Ar = (Ani + Ans) X r
which may be accomplished by a single rotation An = Ani + An2.

To show this, we again let r be the position vector of P with respect
to 0. After the first rotation, Arii, the position vector of P becomes, by
Eq. (7-1), r + Ani X r. After the second rotation, An2, the position
vector of P will be r + Ani X r + An2 X (r + Ani X r). Now, if Ani
and An2 are product can be neglected. Hence the
infinitesimal, their
total displacement of P is Ani X r + An2 X r = (Ani An2) X r. +
This proves the previous statement, and thus infinitesimal rotations of a
rigid body are vectors.
the contrary, however, finite displacements of a rigid body are
On
not additive according to the parallelogram rule and thus may not be
considered as vector quantities. In order to verify this statement, the
student should consider an ordinary right-hand coordinate system and
successively apply two finite rotations (say of 90° each), noting the
final position. Upon performing the same two rotations in reverse
order now, the difference in final positions is quite clear. Thus the order
of performing the operations is essential, and finite rotations do not
obey the parallelogram rule of addition.
Following Chasles' theorem on finite displacements and letting the
displacements be small, we can also state that a general infinitesimal
displacement is equivalent to a translation of a base point and a rota-
tion about some line through the base point. Thus, for infinitesimal
displacements, the displacement of a point P (whose position vector
relative to the base point is r) is given as
Ar = As + An X r (7-2)
where As denotes the displacement due to the translation part of the

motion and is the same as the displacement of the base point. Again
the choice of base point is immaterial; a different base point simply
revises the translation As, leaving the infinitesimal rotation An unaltered.
7-2. Kinematics of a Rigid Body
We are now in a position to proceed with our original aim, the discus-
sion of the kinematics of a rigid body and, in particular, the basic theo-
rems to which wc previously referred.
Sec. 7-2] Kinenmlics of a rigid body 127
Suppose that in Fig. 7-2, a particle occupied position/^ at some time t
and. during some infinitesimal time interval At, moved to position /*'
because of an infinitesimal rotation An about line L passing through the

fixed point 0. Then, from Eq. (7-1), the velocity of the particle is
Ar
= hm An
dr .. ,.
V = ^77 = lim —- —- X r
ai A/-.0 ^t :u-ô At
or V = 0) X r (7-3)
The cjuantitv to = lim (An /A/) = dn/dt is called the angular-velocitv
vector. Its magnitude is = dd/dt, and its direction coincides with

|(>)|
the instantaneous direction of the infinitesimal rotation vector An and

is known as the instantaneous axis of rotation. We see that these
remarks are in complete agreement with the discussion of Sec. 3-7.
The magnitude of or dd/dt, is also
|fc>|, called the speed of rotation; it is
commonly represented by the units of radians per second or revolutions
per minute (rpm).
Theorem I. Angular-velocity vectors of a rigid body about a common base
point can be compounded vectoriaUy to form one resultant angular-
velocity vector passing through 0.
As we have just seen, an angular-velocity vec^tor of a l)ody is a vector
whose line of action is the instantaneous axis about which the body is
turning, whose magnitude is the speed of turning, and whose sense is as
determined by the right-hand-screw rule. Since a pure rotation about
one line is not the same as a pure rotation of the same magnitude about a
parallel line, angular-velocity vectors must be regarded as localized along
a definite line. However, an angular-velocity vector can be shifted
along its own line of action without causing any change in the motion of
the body (Sec. 3-G). Any point on the line of action of an angular-
velocity vector can be assigned as a base point. Thus a rotation about a
certain base point shall be taken to mean a rotation about an axis which
passes through that point.
A rigid body may have several rotations, each about a different axis
in space, all at the same time. Theorem I then tells us that if the axes
of rotation are concurrent, a complicated motion invohing se\eral
simultaneous rotations can be reduced, i.e., rendered etjuivalent, to a
single rotation about one axis passing through the common point of
intersection of the axes of the given rotations. To prove this theorem,
let 0)1 and W2 be the two angular-velocity vectors describing the two
simultaneous rotations of a rigid body. Assume that these two vectors
intersect, and let their point of intersection be 0. The point is there-
fore a common base point for the two rotations. Let P be another point
on the rigid body with a position vector r from (Fig. 7-3). The
velocities of P due to the two rotations are, from Eq. (7-3),
Vi = o>i X r V2 = a>2 X r
Since velocities can be compounded vectorially, the resultant velocity

of Pis
Vr = Vi + V2 = (0)1 + Wo) X r
If we let the above Vr be the velocity due to a single resultant angular-

velocity vector o)r passing through 0,
i.e., if we let
Vr = o)r X r
then it follows from the above two

equations that
0)r = 0>l + 0)2 (7-4)
This proves Theorem I. From the

above it is also clear that in common
with other vector quantities, the ad-
dition of angular-velocity vectors is
commutative; i.e.,
Fig. 7-3 (Oi -f- 0)2 = "2 + t«>i (7-5)
This result of course could have been anticipated since we have already
demonstrated that infinitesimal rotations are vector quantities.
Theorem The motion of a rigid body, due to a rotation of o about a
II.
base 'point and a translation, is identical with the motion due to a rotation
of the same to about any other base point 0' and a new, different translation.
The most general infinitesimal displacement of a rigid body occurring
in atime interval A^ is given by Eq. (7-2). Then the total velocity of a
point P, located by the position vector r from an arbitrarily chosen base
point 0, can be obtained by dividing Eq. (7-2) by and subsequently M
passing to the limit as A^ — > 0. Thus
dr d^ ,
It
= * + "'<
or vp = vo + to X r (7-6)
where vo is the total velocity of the base point 0, and hence is the velocity
of the translation part of the motion common to all points in the rigid
body (Pig. 7-4).
now suppose that the total velocity of P can also be considered as
Let us
due to a rotation of to' about a different base point 0' and a translation.
According to this consideration, the velocity of P is
vp = Vo' + w' X r' (7-7)

Sec. A inertialirs of <i ritjid body 12<)
where \o- is the total velocity of 0', and hence is also the velocity of the
translation part of the motion, the position vector oi P nieasureil
and r' is
from ()' . If the position vector from to 0' is represented by a,

tlu'ii, using K(|. (7-(>),
vo' = vo + w X a (7-8)
Sul)stituting r = r' + a into Va\. (7-()).
v/' = Vo + w X (r' + a)
But from E(is. (7-7) and (.7-8)
v/' = Vo + w' X r' + w X a

Hence CO X r' = o>' X r'
Since this is true for a arl)itraiily chosen Aalues of r'. it necessarily

follows that
(7-9)
Therefore Theorem II is proved.

We note that although the motion
of a rigid body can be considered as
the combination of a rotation of the
same o> about any base point and a
translation, the translation part of
the motion in general varies both in
magnitude and in direction when we
change the base point.
Theorem II also serves as a demonstration of the assertions we made
previously regarding change of ba.se point for infinitesimal rotations.
The change of the translation part of the motion with a change of base
point is given by Eq. (7-8), or
Vo' — Vo = to X a (7-10)
This equation can be expressed by the following theorem:

Theorem III. The relative vclocilij of any -point 0' with respect to any
other point in a rigid body is equal to the velocity due to rotation of u> of the
body with as the base point.
Another important consequence of Eq. (7-10) is the following: Suppose
the motion of a rigid body consists of a rotation of u> about the base point
and a translation of vq. Let us resolve vo into its two components, a
longitudinal component \oi, along the direction of to and a rro.s.s component
yoc perpendicular to w. We can always find another point 0' such that
(o X a = ^'uc
where a is the position vector from to ()'. If we choose this particu-

lar point O' as our new l)iis(' point, Ef|. (7-10) then shows that vo' is etjual
to the longitudinal component of vo, yoL, only. Since yol is given as
yoL = — ,—
CO"
<»>
the point 0' is some point whose velocity has magnitude

Vo • «
Vo'
to
and whose direction is parallel to gj. Thus the point 0' is oriented with
respect to so that a satisfies the equation
Vo • W w
CO''
^— o) = Vo +I
to X a
The motion of the rigid body therefore becomes that of a rotation of to

about the new base point 0' and a translation of vo' parallel to to. This
type of motion is the motion of a screw, the axis of the screw being parallel
to a. We also remark, without proof, that the process of determining the
screw is a unique one. We can state this result by another theorem.
Theorem IV. The motion of a rigid body at any instant can be considered
as that of a screw about a particular axis.
The reader will notice the similarity between the above theorem and
the result of Sec. 4-8, which states that a general system of forces acting on
a rigid body is equipollent to a wrench along a particular line of action.
The force in Sec. 4-8 is similar to the rotational velocity in this discussion,
and the moment there is similar to the translational velocity here. No
further physical significance, however, should be attached to this simi-
larity. Specifically, this similarity does not imply any dynamic relations
between forces and moments on the one hand and motions on the other.
It should be noted that the most general motion of a rigid body may
consist of several rotations about different nonintersecting axes and
several translations. To reduce such a motion to a simpler but equivalent
form, one can first use Theorem II to shift the axes of rotation so that they
all pass through an arbitrarily assigned common base point. Each shift
of the rotational axis is accompanied by a change of translation according
to Eq. (7-10). We next apply Theorem I to sum vectorially the several
rotations into one rotation. We certainly can sum vectorially all the
translations into one translation. We end up, therefore, with either (1) a
rotation and a translation in an infinite number of ways, depending on the
choice of the base point, or (2) a screw motion in only one way. Either
of these two motions completely describes the most general motion of a
rigid body at any instant.
If the motion of a rigid body is confined to two-diinonsioiial, i.e., plane
motion, there cannot exist any velocity components along o, or alterna-
tively, o> is normal to the plane of the motion. Here vo/, = or vo to =0, •
Sec. 7-3] The Eulerian angles l.'U
so that the location of a point on the screw axis is jîven hy the expression
vo = — X a. Theorem l\ can
u) l)e stated for this special case:
Theorem Ha. The plane motion of a rigid body at ang instant ran alumgs
be eonsidered a.s o pure rotation about a partieular axis. (The intersection
of this axis with the plane of motion is often called the instantaneous center
of rotation, or the velocity pole.)
The instantaneous center, then, is a point of zero velocity. Its position
in space varies with time and depends on the time variation of u> and v.
Moreover, the instantaneous center need not lie within the physical
confines of the rigid body, hut merely associated with it. The locus of the
instantaneous center with respect to a fixed set of axes (space axes) is
usually referred to as the spcLce centrode, while the locus of the instan-
taneous center with respect to a set of axes fixed in the hody and mo\in}»;
with it (body axes) is called the body centrode. With respect to the
instantaneous center the velocity of any point is given as v/. = o> X b.
where b is the position vector of the point /^ relative to the instantaneous
center.
Many future misunderstandings can beavoidid il the reader takes note
that:
1. A base point "belongs" to the rigid body and tluMcforc nioxcs with
it. The velocity of the base point is the velocity of the translation |)art of
the motion.
2. Although a base point is a point of the body, it need not be in the
body. If a base point is not in the body, we can imagine that it lies in an
imaginary rigid extension of the rigid body.
3. The ba.se point can be any arbitrarily cho.sen point of the l)o<ly.
4. The motion of a body may change completely in nature from one
instant to the next.
7-3. The Eulerian Angles

The body in space is completely determined by
position of a rigid
specifying the coordinates and z of any three points of the body, if the
x, y,
three points do not lie on the same line. This means that we have to
specify nine variables, three for each of the three points. However, since
the three mutual distances between these three points are fixed for a rigid
body, we have three equations that relate the nine variables. We there-
fore are left with only six independent variables to completely specify the
position of a rigid body.
In.stead of specifying the coordinates of three points, the position of a
rigidbody can more conveniently be described by the following method:
We specify the position of an arbitrary point
first of the body by its
coordinates Xo, yo, and Zq. The body is now still free to rotate about 0.
The orientation of the body is then specified by three more independent

variables called the Eulerian angles. (The total number of independent
variables is therefore still six.)
Let OX, OY, OZ be a set of absolute coordinate axes in space (space axes)
and Ox, Oy, Oz a set of coordinate axes that are fixed in the rigid body and
therefore move with it (body axes) (Fig. 7-5). Let the two sets of axes be
cut by a sphere of unit radius and center 0. The first Eulerian angle 6 is
Fig. 7-5
the angle between OZ and Oz. The second

the angle between angle is
the meridian plane ZOz and the plane and measured in the XY
ZOX is
plane. The third angle ^p is the angle between the meridian plane ZOz
and the plane zO.r and is measured in the .vi/ plane. It can be seen that
the first two angles 6 and determine the orientation of the moving axes
(j)
Oz; the third angle determines the amount of rotation about Oz. The
three Eulerian angles,! therefore, completely specify the orientation of
the rigid l)()dy.
t It is unrortuiialc (lial. tlic second and tliiid I'lnlcriaii antj;l('s arc defined din'crcntiy
by did'erenl anthors. Some British autliors use xp second anf^le and
for the for the <t>
third. Still others use tiiid \p that arc dilfcrent by 90° from the ones used here.
Sec. 7-1] Motion of a point of n riijid txtdy \\V.\
Let the angular velocity of tlio body at any iiisiunt l)e
o) = ojî + ooj + a,'..k
where i, j. k are unit Nectors aloufj; the connliiiate axes .r, //, z fixed on the
body. We now seek to find to or its coniponents when the motion of the
body is specified by its Eulerian angles 6, 0, and \^ as functions of time.
Let us extend the great circle /»/*, lying in the xij plane, to intersect the
meridian st at B. .Mso draw the radius ()V in the plane of the great circle
but perpendicular to the meridian. It is clear that the angular-velocity
vector 6, cfjrresponding to the infinitesimal rotation of magnitude A0.

has the direction of Oi' in the xij plane. Since the lines OH and ()(' are
perpendicular and both lie in the xij plane, the angles BOx antl COij are
both equal to \p. The comjjonents of 6 along the .r. //, and z axes are
d sin ^, d cos xp, and zero, respectively. Further, the remaining two
angular-velocity vectors <^ and t^, corresponding to the infinitesimal
rotations of magnitudes Ac^ and Ai^, are clearly directed along the Z and
z axes, respectively. The component of <j). along the z axis, is 4> cos 6; its
component in the xij plane is <^ sin d, which can further he resolved into
two components — <^ sin B cos ^ and sin 6 sin ^ along tlie x and y axes,
respect i\ely. The angular-xelocity vector has a component in the ij;
2 direction only. Hence
03^ — 6 sin \p — <i)

sin 6 cos \j/ (7-11)
Uy = 6 cos -{- \J/
(j) sin d sin \p (7-12)
u, = ^ cos e -{- \l/
(7-i:i)
Api)lication of these relations will hegixen in Chap. 12 on the Dynamics

of Rigid Bodies.
7-1. Motion of a Point of a Ki«;i<l Body

In Sec. 7-2, we noted that the general motion of a rigid body can be
considered as the combination of a rotation about an arbitrary base point
and a translation (Theorem II). CVmsidering any point /^ fixed in the
rigid body with the position vector r from (), we can write the velocity of
P [Kq. (7-Oj] as
v/' = vo -i- (.> X r (7-14)
We have seen that the two terms on the right, \o and m X r. represent the
velocities due to translation and due to rotation about (K respectively.
The acceleration of/'' is obtained by dilTercntiating Kq. (7-14) with
respect to time;
a/' = v/' = Vo + w X r -}- o> X r
where o> is the time rate of change <if t.> and is calleil the niKjiitar arrrfrrn-
lion. Now r is the relative velocity of P with respect to and, according

to Theorem III of Sec. 7-2, is <o X r. Hence the above equation becomes
ap = vp = vo + o> X r + G) X (co X r) (7-15)
In the above expression o, or the acceleration of the base point,

'
represents the acceleration due to translation, while w X r and w X (<o X

r) represent the acceleration due to rotation about the base point. From
the definition of a vector product, we see that the product o) X r is
(c)
Fig. 7-
perpendicular to the plane containing to and r. Imagine a plane passing

through P and normal to (b. Let C be the point at which w pierces this
plane (Fig. 7-6a). Let us now draw a circle in this plane with C as center
and CP, the perpendicular distance from P to (b, as radius. We see that
to X r is always in the direction of a tangent to this circle. The quantity
w X r is therefore called the tangential acceleration of rotation. From
the discussions of a vector triple product (Sec. 3-8) we see that the product
0) X
(o X r) lies in the plane containing o) and r and is perpendicular to
(o. This product is therefore along the direction of the perpendicular
from /-* to (o, always pointing toward o (Fig. 7-6?>). The quantity o> X
:
Sec. 7-t] Molion of a j)(>inl of a rl(fi<l Ixxly 13")
(o X r) is culletl the rndinl, or the cittlripdaL uccek'rat ion of rotation.

The respective magnitudes of these acceleration components are given as
follows
Tangential acceleration: |(b X r] = wr sin </>i
Radial acceleration: |o> X (a> X d = u'-r sin <^2
In plane motion, the same interpretation holds, with the exception

(and simplification) that the direction of w and cb coincide, both being
normal to the plane of the motion. For the special ca.se, then, of the i)a.se
point and the point P
same plane, the plane of the motion,
lying in the
sin (t>
= sin î = sin 02 =
and the above magnitudes simplify accord-
1
ingly. Figure 7-6c summarizes these

considerations in plane motion for
this special case; all the vectors
whose magnitudes are shown lie in
the plane.
Thus the acceleration of a point fixed
in a rigid body is the sinn of the acceler-
ation of the base point and the accelera-
tion due to rotation about the base point.
The latter, in turn, is the sum of a
tangential acceleration d X r and a
radial acceleration w X (<o X r).
Example 7-1. A plane wheel of radius R rolls down a slope without

slipping. Find the velocity and acceleration of a point in the wheel.
Let us for convenience choose the center of the wheel as the base point
and erect the coordinate axes x and y, as shown in Fig. 7-7. Consider any
point P at a distance of r from the origin. Let the angle between OP and
the X axis be 0. We can consider the motion of the wheel as a rotation
about the base point and a translation. The relation between the
amount of rotation and the amount of translation is fixed by the condition
that the wheel does not slip. This condition gives
vo = w/?i
vo = cjRi
The position vector of any point /* on the wheel from the l)ase point is
r — r cos i + / sin </>

j
The angular velocity and the angular acceleration of the wheel at any
instant are assumed to be
= — wk = — cjk
By Eq. (7-14), the velocity of point P is
vp = vo + w X r = w(R + r sin </>)i — cor cos (/>

j
By Eq. (7-15), the acceleration of point P is
vp = vo -H " X r -t- to X (o> X r) = (oiR -\- cbr sin (/> — wV cos (/>)i
— (cor cos (f)

-\- co-r sin </>)j
Since we body in plane motion, the concept

are dealing with a rigid
of the instantaneous center be used profitably. We immediately
may
notice that the point of contact 0' of the rolling wheel and the surface
over which it is rolling without slipping is a point of zero velocity, and
thus is the instantaneous center. (The point of contact is always the
instantaneous center for the situation of rolling without slipping along a
fixed surface.) The position vector of P with respect to 0' is
b = (r cos (f))i -\- (R -\- r sin (/))j
Accordingly, then,
vp = G) X b = w{R -{- r sin <^)i — cor cos <^ j
Notice that this concept cannot necessarily be extended for the deter-
mination of the acceleration since, in general, the instantaneous center
is a moving point in space.
7-5. Four-bar Linkages
The discussion in the previous section can be directly applied to the

plane motion of an important class of mechanisms grouped together
Fig. 7-8
under the name of four-bar linkages. A typical four-bar linkage (Fig.

7-8) consists of a driving crank AB, a connecting link BC, a driven
linkCD, and a fixed link AD, which is usually the foundation itself.
There are four hinges, one at each of the four joints A, B,C, and D. The
link AB executes complete rotations about the fixed center .4, whereas
the link CD may either rotate completely or oscillate back and forth
about the fixed center D.
Sec. 7-5] Four-bar linkages 137
A reciprocating-engine mechanism (Fig. 7-9) is merely a special case

of the four-bar linkage if the link CD is of infinite length. Also, one or
more of the links of a four-bar linkage ireed not have the actual shape of a
bar. as is illustrated by Fig. 7-10.
Given a four-bar linkage and. at any instant, the angular velocity
w^B and the angular acceleration w.ib of the driving crank AB, we can
find the motion of the other members of the linkage by our knowledge
of the motion of a rigid body (Sec. 7-4). In the following we let the
plane of motion be the xy plane. The linear velocities and accelerations
then have only i and j components. The angular velocities and accelera-
tions are along k. Considering the link BC of Fig. 7-8 as a rigid body,
the velocity of C is, by Eq. (7-1 -i),
Vc = V/i + u>/ir X Tfic
where r/,c is the position vector from B to C. Considering the link CD

as a rigid body, we have
\c = a>r/> X r/;c
whereas, for link AB.

V/i = Ui.Mi X r.t/;
For plane motions |a>/<c X r/,c| = omcBC, \i^ci) X rnc\ = c^cdDC, and
w.i/i X r,t/}| = u.niAB. Equating the two expressions for vc and sub-
stituting in the known value of vb,
w.-iB X Tab + oiBc X Tbc = <^cd X Tdc
This vector equation is equivalent to two scalar equations along the
directions of and j, respectively. Hence we can solve for the two

i
unknowns wbc and ojcd, either one of which will in turn give us vc.
The above discussion can be illustrated graphically by Fig. 7-11.
This method consists essentially of the graphical solution of the vector
equations just developed. At joint C let us draw v^, which we can easily
achieve once (o^b is specified. At the terminal of \b we draw a line
ubc >^ EC
Fig. 7-11
perpendicular to BC. This line represents wcc X tbc, which is as yet of

unknown magnitude. The terminal of vc must therefore lie along this
line. Now we draw a line from C perpendicular to DC, again of unknown
length and representing vc = toc» X vdc- Its intersection with the
previous line then determines the quantity vc as well as the quantities
G)c/) and (^bc- It is also helpful to note that the instantaneous center of
rotation (Theorem IVa) of link BC is the point E, the intersection of the
lines perpendicular to \b and vc, respectively. Hence
Vb Vc
BE CE
After determining the velocities of a four-bar linkage, its accelerations
are to be found in the following way. Considering the link BC (Fig. 7-8)
as a rigid body, the acceleration of C is, from Eq. (7-15),
a.c = SLB -\- o>Bc X Vnc + w/jc X (co/jr X Tbc)

Sec. 7-6] Rale of cliaiufe of a vector in a tnoviiuj frame 139
Considering the link CD as a rigid l)()dy, we have

Uf — del) X r/,c -|- G>c/j X (ioc/> X rue)
while for link .1/^ we ha\e

'<îi — W.4K X r..,« + o>,4rt X (a>.i« X t_\n)
It is noted that for plane motions X jw/yc fbcI = uhcBC,
|<Jc/> X r/jc| = iocoDC \u>Hc X (o>cc X r/,f)| = uuc-BC
and IwcD X (wr/) X rr)c)\ = ojcd-DC, etc.
Substituting tlu^ known \alue
of ae, and then eciuating the two expressions for ac, we again obtain one
vector equation which is equivalent to two scalar equations along and j. i
tic/} X DC
Fic. 7-12
We are thus able to solve for the two unknowns umc and wcd, either one
of which will in turn give us ac.
The above consideration on accelerations can be illustrated graphically
by Fig. 7-12. At C we draw the vector an, which can be easily obtained
once (JiAB and wîc are known. At the head of an we add the vector
<>>Bc X (<OBc X Tnc), which is in the direction from C to B (Sec. 7-4).
At the head of the la.st vector we draw a line perpendicular to CB. This
is line I, representing C^bc X rnc, which is yet of unknown magnitude.
The terminal of ac lies somewhere along this line. Now, again at C,

we draw the vector iocd X (wcd X roc), which is in the direction from
C to D. At the head of the last vector we draw a line perpendicular
to DC. This is line II, representing o>cd X r/>c, which is yet of unknown
magnitude. Somewhere along this line lies the terminal of ac. Then
the intersection of lines I and II determines all three quantities ac, ubcBC,
and ojcdDC. Hence we obtain ac, w/<cs and ojcd-
7-6. Rate of Change of a Vector in a Moving Frame

In many problems of dynamics, it is convenient to refer basic (luanti-
ties such as velocity and acceleration to a moving reference frame. Let
X, y, be a set of rectangular coordinate axes which are in motion.

and z
The axes and z therefore constitute a moving frame of reference, or a

x, y,
moving coordinate system. Let i, j, and k be the unit vectors directed

along the axes x, y, and z, respectively. We would first like to find the
rate of change of these unit vectors with respect to a fixed set of axes,
the absolute, or Newtonian, frame of reference.
For the sake of brevity we shall use the term vehicle to denote the
moving frame of reference. (We might visualize that the vehicle is an
''imaginary" rigid body firmly attached to the moving axes and there-
fore moving with them.) Depending
on the nature of the problem, the
vehicle, representing the moving frame,
may or may not coincide with the actual
moving rigid body under consideration.
Since the general motion of the vehi-
cle, according to Sec. 7-2, consists of a
translation and a rotation about any
base point, we let the base point coin-
cide with the origin 0' of the unit

vectors so that the angular-velocity
vector (0/ of the vehicle passes through
0' (Fig. 7-13). For an infinitesimal
Fig. 7-13
rotation about 0', the change of the unit
vector i is the displacement made by the terminal of i. Hence the rate
of change of i, or i, is the velocity of the terminal of i and is therefore to/ X i.
Thus
1= o)/ X 1
Similarly, J = "/ X j (7-16)
k = 0)/ X k
It is magnitude nor the direction of the unit vectors
clear that neither the
i,j, and k
changed by the translation part of the motion of the vehicle.
is
Equations (7-16) therefore represent the total rate of change of the unit
vectors fixed in a vehicle due to the motion of the vehicle.
Let us next consider any vector quantity A. We can represent A, at
any instant, by its components along the moving coordinate axes x, y,
and z and write
A = .4,i + .4 J + .Lk
To an observer in the absolute, or Newtonian, frame of reference, the
rate of change of A is, from Eq. (6-4),
dA
-y- = A = i.i + A,j + i,k 4- A^ -f- .4j + .4,k (7-17)
dt
1
Sec. 7-7] Motion referred lu a nioriny voordinale syslrni I 1
The first three terms of the ri<j;ht-hand side of the above ecjuation repre-
sent the rate ofchange of A relative to the vehicle (i.e., as if the vehicle
were not moving, or as detected by an observer at 0' traveling with the
vehicle). Let us use the svmbol
-^\^
\dt).,,
= AA + .l,j + AM (7-18)
to denote the relative rate of change of A.

The la.st three terms of Eq. (7-17) become, from Imjs. (7-lG) and the
distrit)utive law of vector products,
u)/ X (.hi + -IJ + -Uk) = to/ X A

Hence Eq. (7-17) can be written
f/A /f/A\ , .
/-, ,^x
It = \ii),., + "' ^ '^ <'-'»>
The significance of the three terms in the above equation is as follows:
dA/dt is the rate of change of A with respect to an absolute frame of

reference.
(dA/dt)rei is the relative rate of change of A (relative to the xchicle).
10/ X A is the rate of change of A due to its getting a "ride" on the
vehicle.
be emphasized again that, in Eq. (7-19), A can be ani/ vector

It .should
ciuantity with no restriction whatsoever as to its nature. Thus E(i. (7-19)

should be viewed more properly as an equation between two types of
operators, namely, the time derivatives with respect to moving and fixed
coordinate systems. This point of view may be emphasized by writing
Eq. (7-19) in the following form:
It
= (jt)r.^ + "^ "^
where the three operators are to be applied to any gi\en \ector.

Since dX/dt is also a \ector. it may be represented by its components
along any convenient set of fi.xed coordinate axes in space. The fixed
coordinate axes may, at the particular instant under consideration, ha\'e
the same directions as the moving .set of coordinates. This is often done
intentionally for convenience, as the examples in the rest of the chapter
will show.
7-7. Motion Referred to a Moving Coordinate System

bet .\)'Z be the cooidinate axes in an absolutes frame of reference
and xi/z the ones in a inoxing fianie of reference. (We can imagine, for
example, that xijz axes are inscribed on a moving vehicle.) A point P

moves with respect to the vehicle. We want to find the absolute motion
of P. This is important because Newton's law applies only to an absolute
frame of reference (Sec. 2-2).
Let r be the position vector of P measured in the absolute, or XYZ,
system and p the position vector of P measured in the relative, or xyz,
system (Fig. 7-14). Let R be the position vector of the origin 0' of the
relative system measured with
respect to the absolute system,
i.e., the vector from to 0'. We
shall call the unit vectors along the
relative coordinate axes i,
j , and k.
From the above definitions, we see
that the absolute displacement r of
the point P is
r = R-1- e
The absolute velocity of P is
dR dg
^ dt~ ~dt ^t
Fig. 7-14
Since 9 is the position vector of
point P relative to a moving system of coordinates, we have, applying
Eq. (7-19),
Using the symbolism R = dR/dt and pr = {dg/dt)rei, the velocity of

point P becomes
V = R -F e, -f 0)/ X p (7-20)
In Eq. (7-20) «/ is the angular velocity of the vehicle, and
%,. = *i +M+ *
is, by Eq. (7-18), the relative velocity of P
with respect to the vehicle.
We also see from Eq. (7-14) that the two terms R w/ X p in Eq. (7-20)
+
are the vehicle velocity at P. Equation (7-20) therefore tells us the
following: The absolute velocity is the sum of the relative velocity and the
vehicle velocity.
The absolute velocity v represents the velocity of point P that an
observer at the origin of the fixed XYZ coordinate system would
detect. The relative velocity pr is that velocity seen by an observer
at 0' traveling with the moving axes and is sometimes referred to as the
Sec. 7-7 Motion referred to a nioring coordumle svsleni 1 13
apparent velocity. Vehicle velocity is composed of two parts: the first,

R, represents the absolute velocity of 0' as seen by an observer fixed
at 0, whereas the second (w/ Xp ) repl-esents the velocity of/' due to the
rotation of the vehicle alone, as seen by an observer stationed at ()' but
not rotatin<j; with the vehicle.
Example 7-2. As an example, we consider the followiug pr()l)iem.
A thin wheel of radius a is rotating about a horizontal shaft with angular
speed p (Fig. 7-15). The horizontal shaft is itself rotating a))out a

vertical axis with an angular speed S2.
The distance from the vertical axis to
the center of the wheel is L. Find the
absolute velocity of a point on the rim
of the wheel, first, by the methods of
Sec. 7-4 and, second, by the methods
of Eq. (7-20).
We a moving coordinate
assign
system, attached to the horizontal
shaft at its point of intersection with C..>n
the wheel and oriented as shown in Fig.
7-15. As our point in question we
select a point on the rim of the wheel
at an angle 6 counterclockwise from
the moving z axis.
according to Eq. (7-14), the

First,
velocity of the point P is composed of the translational velocity of the base
point plus the velocity due to the rotation about the base point. Letting
0', the origin of the moving coordinate system, be our base point, its veloc-
ity, by virtue of the fact that it is a point on a rotating rigid body, is given
as
vo' = (iik) X (Li) = «Lj
The velocity due to the rotation about ()' is given by w X p, whore w

is the total angular velocity of the wheel and can be expressed as
(>) = flk -\- pi
and p is the position vector of the point P relative to the base point and
can be expressed as
p = —a sin 6 } -\- a cos 6 k
Thus, upon performing the required cross multiplication, we find that
the velocity due to rotation is given as
(0 X p = iia sin d i — pa cos j — pa sin k

Hence the absolute velocity of point P becomes vp = vo' + w X p, or
vp = Qa sin 6 i -\- (12L — pa cos d)j — pa sin 6 k

Wenext use the method of Eq. (7-20) to solve the same problem.
According to this method the absolute velocity is the sum of the relative
velocity and the vehicle velocity. The relative velocity, being the
velocity of point P relative to the moving frame, by virtue of its rotation
in that frame, is given by pr = w X p, where now to = pi and p is as

previously defined. Thus
Qr = —pa cos 6 j — pa sin d k
The vehicle velocity is vo' + w/ X p, where w/ is the angular velocity of
the rotating frame and is equal to w/ = 12k. Thus we have, for the vehicle
velocity, using the previously determined vo' = OLj and o/ X p = Oa
sin 6 i,
Vo' + 0)/ X p = â sin 6 i -\- 12Lj
The absolute velocity now becomes vp = p^ + \o' + o/ X p, or
Vp = Ua sin 6 i -\- (12L — pa cos d)j — pa sin 6 k

which agrees with the result of the previous method.
We have thus obtained \p by two methods: first, by the use of Sec. 7-4,
which entails kinematic relationships and, second, by the use of Eq. (7-20),
which is the result of differentiation of a vector expressed in terms of a
rotating coordinate system. The more convenient of the two methods
will depend on the nature of the particular problem at hand. It is
important in problems of this type to clearly distinguish between the
angular velocity of the body or bodies under consideration and the
angular velocity of the rotating frame of reference.
We next consider the absolute acceleration of point P as referred to the

fixed coordinate system at 0. Thus, from Eq. (7-20),
+ g^
(o>,) X 9 + .o, X j^ (5)
Again, applying Eq. (7-19), since we are forming derivatives with respect
to quantities referred to a moving coordinate system, this becomes
a = R -|- {Qr)rcl -h O/ X pr + {<^f)rel X p + (tO/ X W/) X p

-h to/ X (p)r<; + to/ X (to/ X p)
Noting that --r. (to/) = (tb/) = (oif)rci -f to/ X to/ = (to/),,;
and that (&r)rei = 9r = -Ti + VJ + ^k

Sec. 7-7] Motion referred to a moviruj coord'uiale sysletn 1 13
the acceleration of the point P finally becomes
= R + Pr + W/ X p + w,.X X p) + 2w, X pr (7-21)
We now examine the various terms comprisinfj; E(|. (7-21). The

absolute acceleration a of the point P represents the acceleration as it
would appear to an observer at in the fixed coordinate system. The

term pr is the relative acceleration of P with respect to the vehicle and
represents the acceleration that an observer, at 0' and moving with the
vehicle, would detect. This acceleration is sometimes referred to as the
apparent acceleration. Next the terms R -f o>/ X p + to/ X (o>/ X p)
are. by virtue of Eq. (7-15), the vehicle acceleration atP as if/-* were not
moving relative to the vehicle. Individually, the various terms have
the following significance: R represents the absolute acceleration of the
moving origin 0' as viewed by an ob.server stationed at 0; 6/ X p and
0)/ X (w/ X p) represent the tangential and centripetal acceleration due
to the angular acceleration and angular velocity of the moving axes,

respectively, and represent the corresponding accelerations measured
by an observer stationed at 0', but not rotating with the vehicle, and
due to the angular acceleration and velocity of the vehicle alone. Finally,
the term 2a>/ X pr is called the CorioHs acceleration^ after G. Coriolis
(1792-1843), who first discovered it. Clearly, the Coriolis acceleration
vanishes when the point in question is at rest relative to the moving
axes or if the angular velocity of the
moving axes is zero. Equation (7-21) X.x
can then be stated as follows: The abso-
lute acceleration is the sum of the relative
acceleration, the vehicle acceleration, and
the Coriolis acceleration.
Example 7-3. A disk rotating about
an axis normal to its plane and passing
through its center has a radial groove
milled in it (Fig. 7-lG). A ball bearing
travels in the groove, and the distance b
from the center of the disk to the ball
bearing is given as a function of time, b — b{t). If the disk has an angular
speed of w and angular acceleration a, find the absolute acceleration of the
bearing at any instant.
For convenience, at the instant under consideration, \v.t tlu; absolute
and relative coordinates and the groove coincide; as shown in Fig. 7-10.
The various quantities of interest then are as follows:
R =
Now the various terms comprising the absolute acceleration a of Eq.

(7-21) are computed as follows:
R =
<b/ X e = ahj
CO/ X (o>/ X e) = — co"5i
Qr = bi
2o>/ X Pr = 2co6j
The total acceleration, being the sum of those terms, now becomes
a = (6 - w-6)i + (2co6 + ab)j
Unlike the procedure employed in Example 7-2, where we had two

alternativemethods available, here we have no other recourse than to
use Eq. (7-21) alone. A direct use of
the kinematic relationships (Sec. 7-4)
would result in the omission of the
Coriolis acceleration. Thus, the ac-
celeration of a point in a moving
frame of reference generally cannot
be described by the acceleration of a
point on a rigid body.
Example 7-4. Because of changes
in angular speed of a steam-engine
governor, the balls of the governor
move up or down (Fig. 7-17). Find
the acceleration of the ball.
Let the absolute coordinates XFZ
Fig. 7-17
and the relative coordinates xyz be
located as shown. We assume that at the instant under consideration,
the two sets of axes happen to coincide. Let the governor be rotating
about its vertical axis with the angular speed co and the angular accelera-
tion w. The inclined bars of the governor make (instantaneously) an
angle with the vertical axis. We assume that at the instant under
<t>
consideration, both ^ and 4> are known. We see that
4) = 0k ^ = (/)k
e = {a -\- L sin — L cos (/>)! (f) j
Qr = (t)L cos -f <^L
(/»sini
j <^
The various terms that make up the acceleration of the ball P are
obtained from E(i. (7-21) as follows:
: :
Problems 147
Vehicle acceleration
K =
w/ X p = — a)(a + L sin <^)k
to/ X (lo/ X o) = — a>-(a + /. sill </))i
Relative acceleration (Imi. (7-15) and Kxaniple 7-lJ:
e, = <) X r.,/' + (j) X (<> X r,,/0

= (<^/> cos — 0- L sin <^)i + (<^L sin </> + <^'- L cos 0)j
(The same pr can also l)e obtained hy {liffeifntiating p^, assnniinji; con-
stant and i j.)
Coriolis acceleration
2o)/ X p, = 2^Z/ - —24>Lo: COS <f>

k
cos 4> sin
Hence the acceleration of P is letting h fl + /> sin 0,
a/' = ( — w'-6 + 0L cos — <^- L sin 0)i

+ (<^/> sin + 0- L cos 0)j — (a)/> -(- 20Lw cos 0)k
The reader is advised to check each term of the above expression and
to familiarize himself with the significance of the various terms.
Although in this example we have chosen the origin 0' of the moving
axes to coincide with the origin of the absolute axes, this need not be so.
The location of 0' is entirely arbitrary. A change of the location of 0'
merely redistributes the individual values of the three terms that make
up the vehicle acceleration. But the vehicle acceleration, which is
the sum of the three terms, is not affected. Nor do the relative accelera-
tion and Coriolis acceleration undergo any change by a shift of ()'.
PKOBLKMS
7-1. A unit cube undergoes an infinitesimal rotation of magnitude Ad about its
body diagonal OA. Find the displacement of the point A. Verify this result by
considering the rotation to be the sum of three infinitesimal rotations taken about the
three axes at O and also by considering the rotation to l)e the sum of three infinitesimal
rotations taken about the three face diagonals meeting at O.
7-2. Show that in a rigid body in motion, there is in general no point with zero
velocity but that there is, in general, one point with zero acceleration. (Under
special conditions there can be a ix)int or jjoints with zero velocit}' and no jjoint with
zero acceleration.)
7-3. Two points on a rigid body, .1 (1 in., 2 in., in.) and B (2 in., —',i in., in.),
have, at the instant under con.sideration, the following velocities: v^ = Gi + 4j in. /sec
and vb = ( — 12i 8j)c in. /sec.+ Locate the instantaneous center of rotation.
j
Devise a graphical procedure for locating the instantaneous center under similar (but
more general) circumstances. The constant c is some value to be determined.
7-4. Two points on a rigid body, .4(1,2,0) and 5(3,6,0), have, at the instant under
consideration, the following velocities: v^ has magnitude 4 and direction perpendicular
to hne AB, while vb has magnitude 10 and direction
perpendicular to line AB. All
coordinate distances are expressed in inches, and speed in inches per second. Locate
the instantaneous center of rotation. For this type of situation devise a graphicaj
procedure for locating the instantaneous center.
7-5. A wheel of radius a mounted on a shaft of radius b rolls on horizontal tracks
without slipping. If at a certain instant the wheel has an angular velocity of w and
an angular acceleration of w, find the velocity and acceleration of the following points
on the wheel: point A, or the point of contact; point B, or the highest point; point C
on the periphery of the wheel with OC inclined at angle with horizontal. (j)
y/^/////Ay/////////
Fig. P 7-5
7-6. Two points A and 5 of a rigid body are located 4 in. apart along the x axis.
At the instant under consideration.
VA = 3i — j + 2k in. /sec
Vb = 3i + 3j — k in. /sec
VA = i + 3k in. /sec 2
ti = i +2 radians /sec 2
It is further understood that cox = 0. Find the acceleration of B.
Fi(i. P 7-(l Fig. P 7-:
7-7. Two plane wheels, rotating about perpoiidicidar axes through their centers,
are connected by a link C. Wheel A is rotating about a fixed center, while wheel B
Problems 149
is rolling without slipping along the periphery of wheel .1. Find a relationship
between the angular speeds of the two wheels and the link.
7-8. A wheel in plane motion is rolling without slipping inside a fixed eircular track
ol radius R. The radius of the wheel is r, and its angular speed and angular acceler-
ation are given as w and a. Determine the velocitj- and acceleration of tlic center of
the wheel.
Fig. P 7
7-9. A rocker is rolling without slipping along a plane surface. .\t the instant
shown, determine the velocity and acceleration of its center of mass. Tlie radius of
the rocker surface is R, and the center of mass of the rocker is located on the axis of
symmetry a distance a from the rocker surface.
•7-7-7y-r777-rz777'>y, y77-r7-r7-rT7-7-7T7777-
Fiti. P 7-i»
7-10. A wheel rolls without slipping on a track. Given the velocity of .1, find the
velocity of R by graphical construction.
TTTTT
Fio. P 7-10
7-11. For the mechanism shown with a given velocity of A, find the velocity of P
by graphical construction.
7-12. A four-bar linkage consists of a driving crank AB = ft, a driven crank

1
CD = 2 ft, and a connecting link BC = 2 ft. Crank AB rotates with a constant

angular speed câb = 10 radians/sec in the counterclockwise direction. At a certain
instant, the two cranks are parallel and both are perpendicular to the connecting link.
(a) Find the angular acceleration of crank CD at that instant.
(b) Find the angular acceleration of link BC at that instant.
(c) Locate the point, on the rigid body BC or its extensions, at which the acceler-
ation is zero.
Fig. P 7-12
7-13. A motor-driven pump mechanism consists of a crank AB, a connecting link

BC, and a piston and piston rod CD. It can be seen with this construction that CD is
always vertical. AB = 4 in., and BC = 12 in. AB rotates steadily at 1,200 rpm.
At the instant shown, AB makes an angle of 30° with the horizontal, while BC is
perpendicular to AB. Find:
Problems 151
(a) Velocity of point C.
(b) Angular velocit}- of rigid boch' BC.
(f) Acceleration of point C.
Angular acceleration of rigid body BC'
(d)
(e) point, on BC or its extensions, at which the velocity is zero.
The
(/) jx)int, on BC or its extensions, at which the acceleration is zero.
The
7-14. Determine an expression for the time rate of change of a vector A fixed in
-pace and expressed in terms of a moving coordinate system of angular velocity w/.
7-15. The f in a moving rigid body measured from a base
position vector of a point
point in the body A
coordinate .system is chosen to be fixed in the body at the
is r.
liase point. Find the time rate of change of r if the body has an angular velocity w(,.
7-16. .\ wheel without slipping along a horizontal track with an angular
is rolling
-peed w. We define a fixed reference frame
i, j, and k and a moving reference frame
fixed in the wheel ei, e-i, and Cj. At the instant under consideration e^ makes an angle
e with j. The weight force of the wheel is W. Find the rate of change of relative W
to the moving reference system (a) using the results of Prob. 7-15 and (b) by express-
ing W in terms of the moving reference .sjstem.
Fk;. P 7-IG
7-17. .\n automobile is traveling along a meridian of the earth at a speed of GO mph
relative to the earth. What is the acceleration of the automobile at a location of
30''X latitude? If the automobile were traveling along a circle of constant latitude,
what would be its acceleration? Assume that the earth is a perfect sphere and rotates
with constant angular speed and that its mass center is fixed in space.
7-18. A bead slides down a circular hoop with a constant speed of 2 fps wliile tiie
hoop rotates about its vertical diameter with a constant angular speed of 1 radian/
sec. The radius of the hoop is 1 ft. Find the acceleration of the bead when the bead
is at the position shown.
Fig. P 7-18
7-19. A helicopter blade oscillates in such a manner that the angle <j) varies with
time t according to
<j> = (t>o
-\-
<l>' sin pt
where <Ao is the neutral position of the blade and <j)' is the amplitude of oscillation.
The blade rotates about the vertical axis with the constant angular speed w. Find
the velocity and the acceleration of a point at the tip of the blade for any value of (j). I
7-20. A centrifugal pump runner rotates at 1,000 rpm. Just before leaving the
impeller, a particle of water P travels at the speed of 100 fps with the speed decreasing
at the rate of 5,000 ft/sec^, all relative to the impeller. Find the acceleration of P.
Fig. P 7-20
:
CHAPTER 8
Dynamics of a Particle
8-1. General Considerations
For a particle with constant mass m, Newton's second law gives the
relation between the motion of the particle and the force acting on it:
F = wa (8-1)
where F is the resultant force and a the acceleration of the particle.

Expressing the above vector equation in terms of the components in a
rectangular coordinate system, we have the following scalar equations
of motion:
Fx = mx Fy = my F^ = ml (8-2)
We can also write the equations of motion in terms of the normal and the
tangential components (Sec. 6-G)
Fr. = '— F, = ms (8-3)

P
If we use the cylindrical coordinate system, the scalar equations of motion

become (Sec. ()-7)
Fr = m{r - r<p-)
F, = + =
m(r4> 2r<j>)
^ | (r^) (8-4)
F2 — m'z
Although the above equations of motion are, after Newton's second law,
specified to apply only to a particle, be proved subseciuently that
it will
they are sufficient to solve the motion of a body if the rotational elTect of
the body can be neglected. In many prol^lems, the body under considera-
tion is so small compared with its range of motion that the motion can be
con.sidered as a pure translation. The dynamics of the body then follows
15:5
154 Dynamics of a particle
that of a single particle having the mass of the body and located at the
center of mass of the body. Even if the rotational effect is considerable
and must be taken into account, the above equations, as will be shown
later, still define the motion of the mass center of the body.
In applying the above equations to the motion of a particle, if must be
F depends, in general, not only on time t but also on
realized that the force
position r and velocity v. Thus
F = F(r,v,0 (8-5)
Take, for example, the special case of a one-dimensional motion along the
X axis. We have, in general,
mx = F{x,x,t) (8-6)
This is a second-order differential equation of x as a function of t. The

general solution of such an equation contains two arbitrary constants of
integration. In the majority of cases these two constants are determined
by the two initialconditions of the motion, namely,
X = Xo when t =
X = xo when t =
where Xo = initial displacement
±0 = speed
initial
In general, any two independent conditions are sufficient to enable one to
evaluate the undetermined constants of integration.
Let us consider some examples of one-dimensional motion. The
simplest case is when F, and consequently the acceleration x, is a constant.
After two successive integrations and corresponding evaluations of the
constants of integration according to the initial conditions, Eq. (8-6) then
yields
X = Xo + xot -^ I-
2 m
t'~
(8-7)
The reader easily recognizes that the above equation applies, for example,
to the displacement of a falling body under no other force except that due
to gravity (so-called "free" falling body) provided the gravity is assumed
to be a constant and air resistance is neglected.
The next simplest case is when the force is a function of time only, or
F = F{t). Proceeding again in a similar fashion, Eq. (8-6) gives
X = Xo + xd +^ /' /' -^(î) ^'l dh (8-8)
Section 8-2 will illustrate these methods.

Still another situation is when the force is a function of velocity only;
Sec. 8-1] General cons ide rations 155
i.e., F = F{x). In this ease the e(iuation of motion may he written,
mx = F{x) = F{v) (8-6a)

Often the substitution
.. _ dv
^ ~ dt
or the substitution
X = V andJ
..
n; = —
dv
= _ _=
dv dx
y
dv
dt dx dl dx
) renders Eq. (8-G) amenable to solution by direct integration. Which

substitution to employ, of course, is dictated by the form of the function
I
F{x) and or the type of solution desired. The examples and problems will
[
serve to illustrate this point.
I From the first substitution we obtain
j^ m dv
and, since v = dx/dt,
, _ mv dv
"^"^ ~ IW
Integrating these expressions and evaluating the constants of integration,
t - f{v) X = g{v)
from which x = h(t) can be obtained by eliminating r between the two

expressions.
Using the second sul)stituti()n, we obtain directly the eciuation
mv dv
^•' = F{v)
which can be solved, yielding
X ^ g{v) or v = g'{x)
dx
Then dt =
which again can be integrated subject to the initial conditions, 3'ielding
t = h'{x) or .1- = h(t)
(In the above solution the supcrscii])! piinic denotes anotiici- function.)
When the force is a function of displacement only, the ecjuation of
motion takes the form
mx = F(x) (8-6b)
: :
and the substitution x = v{dv/dx) will enable us to obtain a solution

conveniently. Thus, we have, from Eq. (8-66),
F(x) dx
V dv =
in
from which we obtain =

Together with the statement that
v g\x).
V — dx/dt, the solution follows immediately after another integration,
subject of course to the initial conditions. In this type of problem it is
especially convenient to use the concept of work and energy (Sec. 8-6).
////////////// It should be emphasized that this discussion is

intended to present general methods rather than
k, c
"formulas" for substitution.
Equilibrium A vast number of problems can often be re-
position duced, or made analogous, to a spring-sus-
pended mass (Fig. 8-1), which is acted on by the
following four forces
F^{t)i W 1. A variable force/î = iî(0 exerted by some

outside agent
Fig. 8-1
2. A spring force {
— ks), where k is the spring
constant and s is the total change in length of the spring
3. A resistance force (
— ex) if the mass is moving in a viscous medium,
where c is the coefficient of viscous damping
4. The gravity force W
Let x be the displacement of the mass from its equilibrium (i.e., static)
position, which is clearly at a distance 5 = W/k below the unstretched
position of the spring. The change in length of the spring at any
total
instant is therefore s = 8 Taking into account all the forces acting
-{- x.
on the mass m as a free body, we obtain the following equation of motion
mx = W -\- Fi(t) - k(8 -f- x) - ex
Upon simplification, the differential equation of motion, expressed in

terms of the displacement of the mass from its equilibrium position,
becomes
mx + ex -\- kx = Fi(t) (8-9)
The gravity force W has disappeared from the equation of motion by

virtue of our agreement to measure x from the equilibrium position.
Stated differently, then, we may say that the motion of the mass is
influenced by the unbalanced forces only.
Equation (8-9) is a linear differential equation with constant coeffi-
cients, provided c and k do not vary. A complete solution of this problem
will be given in Chap. 9. If c and k are functions of time, the differential
Sec. 8-2] Simple harmonic molion 137
equation is still linear, although

are not constant. The
its coefficients
situation becomes more complicated

the coefficients c and k are functions
if
of .r or x. For example, the spring in the system may be such that its
deflection bejônd a certain range is not proportional to the applied force;
or the resistance may be proportional to the square of velocity instead of
to velocity. For such cases the ditYerential equation becomes nonlinear
and no general analytic solution is available.
What has been said above about one-dimensional motion holds equally
well for throe-dimensional motion. The thro(^dimonsional motion, how-
ever, has this added complication: the force component along anij one
direction may now depend on the position and the velocity components in
all three directions and, of course, time. This means, in general, that
mx = F^{x,ij,z,x,y,z,t)
my = Fyix,y,z,x,y,z,t)
mz = F,{x,y,z,x,y,z,t)
The three equations of motion become three simultaneous, second-order

differential etiuations, and a general analytical solution is (juite hopeless.
Yet it is true that the motion of a particle is uniquely determined if its
initial conditions (i.e., initial position and velocity) and the forces acting
on it are known. f This, however, offers us small comfort, since it is one
thing to know that a solution exists and a (}uite different thing to actually
find the solution. It should be added, however, that solutions to many
practical problems can be obtained by numerical integration, even if no
analytic solution is possible. The use of modern computing machines
opens up the possibility of obtaining numerical solutions to many compli-
cated problems which would otherwise remain unsolved owing to the
tremendous amount of computation involved.
8-2. Simple Harmonic Motion
Suppose a particle of mass m is acted on by a rectilinear force /''

that
varies sinusoidally with time,
F = nsin iot (8-10) J
t This led Laplace to make the following rcrnark.s (1820): "An intclliKPnt being . . .
could expre.ss in the same e(juation the motion.s of the largest bodies of the universe
and the motion of the smallest atoms. Nothing would be uncertain for him, and he
would see the future as well as the past in one glance." This concept is disclaimed
in quantum mechanics, in which only the probability of motions is considered.
X We shall denote the amplitude of an alternating force by Fo rather than Fm-

F is thus alternately positive and negative (Fig. 8-2). Since
sm = sm
coi CO
(-1)
[ t -\
F repeats itself, i.e., goes through a complete cycle, after a time interval of
27r/co. means that during a unit time, F goes through
It also co/27r cycles.
We introduce the following notations and definitions:
T = —
CO
= period, sec/cycle
f = 7=: = T— = frequency, cycles/sec

I Itt
27r
CO = 27r/ = "m" = circular frequency, radians/sec
The maximum value of F is Fo, which is called the amplitude of F.
1^
Fig. 8-2
Letting F be acting parallel to the x axis, we have, from Newton's law,
mx = Fq sin wt (8-11)
Integrating once and evaluating the constant of integration by the initial
speed i; = io at i = 0, we obtain
X = Xq -\ (1 — COS (Jit) (8-12)

co?n
Integrating once more, we find
x = Xa -\- xot -\—— F,

5
cohn
(co^ — sin wt) (8-13)
where a:o is the initial displacement at t = 0. This result can also be

obtained by a direct application of Eq. (8-8).
The variations of x and x with time are plotted in Fig. 8-3. It can be
seen that this motion is a combination of two motions:
Sec. 8-2J Simple Imrnwnic molion 159
Constant-velocity motion
xo + wm
X = Xu + I i(j H
Simple harmonic motion:
X = ^
m sin cot
X = — —
Fo
com
cos COf,
X = 5— Sm Oil = :,
OJ^W co-
lt is convenient to let .t,„ l)e the junplitude of x in the simple harmonic

motion. Then Xm = —Fo/m'^m, and the above e(iuations can be written
X = — cj-.r,„ sin co^
X = cos oit
aj.f„, (8-14)
X = Xm sin oil
f
It can be seen from the first and third equations of (8-14) that
* + co^x = (8-15)
Thus a solution of the differential equation x -f w^x = Oisa simple harmonic

motion, described by Eqs. (8-14).
As an example, if in Eq. (8-9) c and Fi{t) are both zero, the differential
equation becomes
mx -\- kx = (8-15a)
This the same equation as Eq. (8-15) if we let co = \/k/m.

is Hence a
mass m
suspended at the end of a spring (Fig. 8-1) of spring constant k
may experience a simple harmonic motion. The period of the motion is
2'K/\/k/m, its frequency \//c/w/27r, and its circular frequency -y/k/m.
Other examples of simple harmonic motions will be given in Chap. 9,

which is concerned solely with vibrating systems. However, to further
illustrate the methods of Sec. 8-1, let us consider again the system of Fig.
8-1, such that the equation of motion is given by Eq. (8-15a) and the
initial conditions are .r = 0, i = Vo at ^ = 0. Using the substitution
X = V dv/dx and letting co^ = k/m, Eq. (8-1 5a) can be written in the form
V dv = — w'^x dx
Integrating, we find that
and, using the initial conditions, this becomes
V = -y/vo^ — (xi^x^
After rewriting this expression in the form
dx
dt
another integration yields
t = - sin~^ —X -\- B
CO Vo
t The complete solution of Eq. (8-15) is given as
X = A cos wt +B sin ait
which, ii X = Xo and x = xo Sii t = 0, becomes
Xo ,
X = Xo cos
.
cof
,
i
.
sin wt
Sec. 8-3] Motion in a resist i tig ntediiint Id I
from which it cleveh)p8 that B = 0. Solving for x, we immediately fiiul
that
= Vo
—X ' .
X sill cot
03
This is in theform of Eqs. (8-14), noting that here the amplitude of the
f
resulting motion is Xm = yo/w.
8-3. Motion in a Resisting Medium

Consider the plane motion of a small object in a viscous medium which
offers a resistance proportional to the velocity of the body (Fig. 8-4). The
Pîfi. 8-4
equations of motion along the horizontal and vertical directions are
mx = —kv cos = —kx

my = —mg — kv sin = —mg — </> ky
where </> represents the angle between the velocity vector and the hori-
zontal. Integrating the above e(iuations, using the substitutions
X = dx/dt and y = dy/dt, we obtain
log X = t -^ A
m
log B
h + ^) m
t -f
If xo and ijn are the initial speeds at / = 0, we have, after evaluating the
constants of integration in the above equations,
X = ioe-^*/""' (8-16)
= mg
y + (- + tO g-{klm)t (8-17)
:
Integrating again and assuming zero initial displacements,
Xom
(1 -e — (klm)t (8-18)
mot ni mg -
+ ^
,
?/o )
(1 e-(^^/-)0 (8-19)
We can, if we wish, eliminate t in the two above equations and obtain an

equation for the path (Fig. 8-4), It is interesting to note the following
facts
1. The horizontal displacement reaches Xom/k asymptotically as
i —> 00 , as can be seen from Eq. (8-18) and Fig. 8-4.
The vertical speed
2. reaches —mg/k asymptotically as ^^ oo, as
shown by Eq. (8-17).
This final vertical speed is called the terminal velocity. At this speed,
the weight and the resistance balance each other; hence the acceleration
vanishes and the speed cannot increase any further. This is true
independent of the amount of the initial yo.
8-4. Central Forces; Planetary and Satellite Motions
In this section we consider a very important class of problems, namely,

the motion of a particle under the action of a central force. A central
force is defined as that type of force
whose of action always passes
line
through a fixed point, called the
center of force. The gravitational
attraction of the sun on the planets,
r + dr
or of the earth on an artificial satel-
lite, falls into this category; another
example is the force exerted on a
particle by a spring with a fixed end.
Also, in electrostatics, the attraction
or repulsion on a point charge due to
another point charge located at a
Fig. 8-5
fixed point is a central force.
We first examine some general properties of motion under the action of
a central force. We consider a particle, shown at point A in Fig. 8-5,
moving under the action of a force F, the line of action of which always
passes through the fixed point 0. We choose 0, the force center, as the
origin and let r be the position vector of the particle from 0. Since F acts
along the same direction as r,
F
F = -r
r
See. 8-1] Central forces: planetary and satellite niolions 163
The equation of motion, mr — F. of the partiele may thus he written as
r = —
at
x= —
mr
r (8-20)
Now consider the (juantity
d , s dv ,
d\
The first term on the right is v X v. which vanishes identically. The

second term on the right, from Ecj. (8-20), also vanishes. Hence
^^(r X v) = (8-21)
which may be integrated to give
r X V = h = constant (8-22)
Equation (8-22) represents a fundamental property of central-force

motion, and some important characteristics can be deduced from it.
First, we note that since the constant vector h (constant in both magni-
tude and direction) is perpendicular to the plane containing both r and v,
the latter two vectors are always constrained in one fixed plane. Thus
central-force motion is always planar, or two-dimensional. Second, we
examine the physical meaning of the magnitude of the constant vector
h. For this purpose let us consider a small time interval dt during which
the particle travels a distance dr = v dt (Fig. 8-5). The area dA "swept
out" by the position vector r during time dt (represented by the cross-
hatched area in Fig. 8-5) is numerically equal to
dA = lilr X rfr| = jirv dt sin ^
where the angle between the position and the velocity vectors.
\p is
The
integrated area A over a finite time interval is known as the sectorial area.
The time rate of change of .1 is
^^krv
dt 2
sin tA = ;^|rXv|=U
2 ^
[ironi Vai (8-22)]
The quantity h/2 called the sectorial, or areal, speed and. according to
is
Eq. (8-22), is constant for central-force systems. have shown, there- We

mider the action of a central
fore, that the position vector of a particle
motion When ai)plied to the
force sweeps out ecjual areas in equal times.
statement becomes Kepler's second law
of the planets around the sun, this
The radius vector from the sun to a planet sweeps out
of planetary motion :
We have seen that both the magnitude and

equal areas in equal times.
central-force
the direction of the constant vector h are basic quantities in
motion. In addition, it will be seen later that the vector rnh = r X (mv)
is the angular-momentum vector of the particle of mass m about (Sec.
10-3).
Tocontinue our discussion of central-force motion, we now consider a
specific problem, namely, the motion of a planet around the sun or that of
a satellite around the earth. We consider a system consisting of two
bodies, the sunand one of the planets, or the earth and an artificial satel-
lite. According to Newton's law of gravitation, the mutual force of
attraction between two particles acts along the line joining the two parti-
cles and may be expressed by the so-called inverse-square law as follows:
F=-«ê, (8-23)
where G = universal gravitation constant

M, m = masses of the two particles
r = distance between M
and m
Cr =a unit vector along line joining the two particles (Fig. 8-6)
In the following discussion, we shall assume that the two bodies under
examination (i.e., sun and planet or earth and satellite) may be considered
to be particles.! We further assume
that the larger of the two bodies (sun
in the sun-planet problem and earth in
the earth-satellite problem) is fixed in
space, and we attach our coordinate
system to this fixed body. It should be
noted that this last assumption is justi-
fied when the mass of one of the bodies
is much larger than that of the other,
because in such a case the smaller body

will have a negligible influence on the
YiG, 8-6 motion of the larger one. With these

assumptions and using Newton's law of
motion together with Eq. (8-23), we obtain the equation of motion for the
smaller particle, i.e., of mass m:
m d\
GMm
dt r^
which simplifies to
^-"Tê. (8-24)
Since the particle under consideration is under the action of a central

force, Eq. (8-22) must also be satisfied. Remembering that h is a
t See Sec. 8-8 for justification of this point.

Sec. 8-1] Cenlral forces; planclury and sulci I He niolions 163
constant vector, we proceed to form
gj(vXh)=^^Xh + vX^=^Xh
which, from Eq. (8-2-4), becomes
_d (v
,
X ,
h)
,
= -
GM
— - (e. ,
X h) (8-25)
It is clear that Eq. (8-25) could be integrated if its right-hand side could
be expressed as the time derivative of some function. We therefore

consider the right-hand side and note that
Cr X h - e, X (r X vj = r[er X (c, X v)]
Using the expansion rule of a triple- vector product [Eq. (.'3-21)], this
becomes
e, X h = r[(er • v)er — (e^ • er)v] = r[(cr • v)er — v]
Now the velocitj^ vector v may be expressed as
dr d , . dcr , dr
''^'dt^d^'^'^^'irt^-dt^'
and thus we have
_ ddr ,
dr _ dr
^
dt dt dt
the last step resulting since Cr isof constant magnitude of unity, and thus
er'(der/dt) =0 (Sec. 6-2). Substituting the expressions just found
into the expression for (cr X h),
/dr \ „ dcr
Cr X h =
,
'' = -T, e
'-{11'''-'')
r I
'-'-lit
This result can now be substituted into Ecj. (8-25), with the result that
It is clear that the above ecjuation can be integrated once to yield
V X h = GM(er + t) (8-26)
where e = a constant vector of integration, to be determined by the

velocity and the position vectors at a given instant.
A great deal of information on the trajectory of the particle can be
extracted from Eq. (8-26). For example, to determine the eciuatioii of
the trajectory, we note that
r • (v X h) = (r X v) •
h = h h = •
/i2
Hence, making use of Eq. (8-26),
h' = GM(r . e. + r • e) = rGM{l + e cos (/>)
u^here (j) is the angle between the position vector r and the constant vector
E. The last expression may be written
= hyOM
r (8-27)
I -\- e cos
which may be recognized as the equation, in polar form, of a conic section

with eccentricity e (Fig. 8-7). We have therefore proved Kepler's first
law The trajectory {or orbit) of a particle
:
under the action of an inverse-square central

force field is a conic section having the center
of force as one of its foci.
The type of orbit represented by Eq.
(8-27) depends on the value of e in the
following manner:
Orbit type
-r COS
:
Sec. 8-5] Example: salelliie orhils 167
where L = h-fCM is known as the snnilaius rcclum. In analytic geome-

try shown that all geometrical properties of an ellipse are determined
it is
if any two independent parameters are given. Thus the semimajor

axes a and semiminor axes b are related to e and L in the following way:
^ Tg + r^ _ Jj
2 ~ I - e2 (8_2(j^
b = a{l - e-y- = L(l - 6^)-^^
The area of the ellipse is
A = irah = TrTJ{\ - €2)-^'i = 7ra2(l - e2)u.
Since the sectorial speed dA/dt = h/2 is constant, the period of an elliptic
orbit (i.e., the time required for a complete revolution) is
2^ _ 27raHl - e^)^^
h h
Substituting h = \/GML,
T = —7=^ or -r- = ;^^^ (8-30)
which, expre.ssed in words, leads to Kepler's third law of planetary motion

The square of the periods of revolution of the planets is proportional to the
cubes of the semimajor axes of their orbits.

It is a remarkable historical fact that these laws of planetary motion
were enumerated by Kepler in 1()09 and 1619, years before the science
of mechanics was placed on a firm analytical foundation. Ecjually
remarkable is the fact that Newton deduced the law of gravitation from
these seemingly diverse experimental laws.
8-5. Example: Satellite Orbits
We now apply the results just obtained to the motion of artificial earth
satellites. To this end it is convenient to reformulate the basic eîuations
in terms of the gravitational acceleration at earth's surface rather than
the universal gravitation constant and the mass of the earth. We note
that the weight of a particle at the earth's surface is given by
mge =
GMm
—y- I e
where Ve = radius of earth = 4,000 miles, approximately

Qe = gravitation acceleration at earth's surface
=
32.2 ft/sec^, approximately
This expression can be solved ior GM, yielding
GM = Qere^ (8-31)
We wish now to develop an expression relating the eccentricity of the

trajectory to the initial conditions, i.e., the velocity and position of the
artificial satellite vehicle at the instant when the propulsion system is
shut down so that the gravitational force is the only force acting. (Note :
Any other point of the trajectory subsequent to propulsion shutdown

can also be used as an initial condition if the velocity and position at
that point are known.) Using the subscript to denote the initial
condition, we write, from Eqs. (8-26) and (8-31),
1
£ = 5
/
(Vo X h) — CrO
Since our interest is the magnitude of t, we form
e2 = E • E = ^^ (vo X h) . (vo X h) ^^ (ô X h) • e.o + 1 (8-32)
We recall that h is given by
h = rXv = roXvo
so that Vo X h = Vo X (ro X vo) = i'o^ro — (vo • ro)vo
We may then compute the following:
(vo X h) • (vo X h) = yoVo^ — WoVo^ cos^ xp^ = WoVo^ sin^ \]/q
(vq X h) • €,-0 = Vq^Vq — Va^Q cos^ \pQ = foVo sin^ i/ô
where i/'o is the angle between ro and vo. Substituting into Eq. (8-32)
gives the relation we seek:
woVo^ 2yoVo /o ooN

e^ = -^r^. sm2
.
i/'o ^ sm-
• , ,
iô + , 1
1 (8-33)
It will be useful now to consider the very special case of a circular orbit
at sea level, which of course cannot be attained because of the atmospheric
resistance but will give an estimate of the required speeds for a low-
altitude satellite. For such a case, we take
ro = Te "^0 = 2
^ =
and Eq. (8-33) gives
,2
=
\9ere }
where, in order to avoid confusion with the general case, the satellite
speed Wo in this special case has been denoted by Ve. Thus
Ve= VgeTe (8-34)
Substituting g^ = 32.2 ft/sec^ and r« = (4,000) (5,280) ft,
Ve = 25,930 fps = 17,600 mph (8-35)

Sec. 8-5] Example: salellile orhils IM
We may now substitute Eq. (8-34) in E(j. (8-33) to obtain a dimension-
less form of this equation. The result is
""=*"+' <*-^""
''= fey fey ^""*"-'K^') fey
Given the initial conditions ro, Vo, ô, at any instant after rocket burnout,
Eq. (8-36) enables us to compute the eccentricity e. From e and
h- _ ro-;'u- si n- iô
7^1 ~
^ = GM = zr:n
g,re
= ^^ (
- L;) )
-î"'- "Ao (8-37)
we may compute the apogee and perigee of the elliptic orbit by Eqs.
(8-28). The period of the satellite follows immediately from Eq. (8-30).
As a special case, consider the orbital speed for circular orbits at any
altitude. For such trajectories we take ô = tt/'Z and c = 0, so that
Eq. (8-36) gives
Vo = V.
^P (8-38)
Thus the orbital speed for circular orbits is inversely proportional to the
square root of the altitude.
Equation (8-36) maj'^ also be used to solve for the so-called escape
velocity,which is really the stfiallcst speed recjuired to achieve an open
orbit. It is clear from Eq. (8-27) that this corresponds to a parabolic
orbit with e = 1, so that for the case of ô = ir/2, Efj. (8-36) reduces to
fe"yfey-fe)fe"y-
This gives, letting ro/r,, = 1,
ye,cape = V^ Ve = 36,650 fps = 24,900 mph

where rc»oape is the escape velocity for a satellite projected from the earth's
surface. For speeds above this value, it is possible then to obtain a
hyperbolic trajectory. P'igure 8-8 illustrates how the elliptical orbit
becomes increasingly- eccentric, and finally the orbit becomes parabolic,
and then hyperbolic, as the launching speed at rocket burnout is progres-
sively increased while keeping ô constant at 90°.
It might be noted that lunar and interplanetary missions may be
accomplished by any one of the three trajectories, elliptical, parabolic,
or hyperbolic. As has been shown, the elliptical trajectory requires the
least launching speed. Obviously, the ellipse will have to be very
elongated (so-called "transfer ellipse"), with its perigee point close to
earth and its apogee point in the neighborhood of the target moon or
planet. It is also clear that the trajectory near the target will be influ-
Hyperbolic orbit
Parabolic orbit
Elliptical orbit
Circular orbit
Fig. 8-8
enced by the gravitational attraction of the target; this additional influ-
ence has not been included in our discussion.
8-6. Work and Kinetic Energy
Suppose we take the equation of motion
F = mr
and form its dot product with the displacement dr:
F ' dr = mr • dr
From Eq. (6-8),
d .. .^ d , ^. .
dr
d&"'^==d&~^^^-'"='-''dt
Hence F • dr = -m-r (?;-) dt
Integrating each side between the interval ri at ti and r2 at t2,
or F ' dr = }'2^nv2" — M'^nvi'^ (8-39)

/:
Sec. 8-7] Poleiilial enerqy 171
I'rom the :M, we have seen that the Hue integral on

discu.s.sioii in Sec.
the left side of Ecj. defined as the work done iiy the force F
(8-39) is
traveling between /^ and P.. The term ^-imv-] is called the kimlic
cncryij of the particle. Hence E(i. (8-39) tells us that the tvork done on a
particle is equal to thechange of kinetic energy of the particle.
We have said before that F in general depends on time, position, and
velocity. It is clear that the evaluation of the
work integral for the
general case and we would have no advantage in deri\ing
is difficult,
Eq. (8-39). However, if we assume that F is a function of position

only, i.e.,
F = F(x,!j,z)
then there is hope that the work integral can l)e evaluated l)y analytical
means, as in Example 3-7. The use of work and energy is therefore

limited to the forces that are either constant or functions of position
only. ]Many problems that we encounter in practice involve forces that
do no work, workless constraints, such as, for example, the normal reac-
tion between a particle and a smooth surface over which it is sliding.
In these cases, the method of work energy pos.sesses some advantage
since we need not consider the wôrkless constraints if we require to know
the speed of the particle. If, on the other hand, the particle is gliding
over a rough surface, the friction force between the two surfaces in
contact must necessarily be considered in the ecjuation of work and
energy.
Summarizing what we have accomplished, then We started with the :
equation of motion of a particle and integrated it in a certain fashion

to obtain a particularly useful result, namel}'. the cfiuation of work and
energy. If we had proceeded in a similar fashion except that the integra-
tion was performed in a different manner, we would obtain still another
useful result, namely, the impulse-momentum theorem and its counter-
part, a theorem on angular momentum. We shall not discuss these
aspects further in this chapter. Ample coverage will be provided,
however, in Chap. 10.
8-7. Potential P^nergy
We define a potential energy V{x,y,z) as follows: Let Pi(.vijju2\) and

P'i{x2,y2,Z2) be two points in space. The woik doiu; by F in inoxiiig a
mass from Pi io P-i is the line integral
''
F . dr
/
t The kinotio pnorgy f.-iri be oxplicitly expressed in terms of any convenient coordi-
nate system. See I'rob. 8-17.
which will in general depend not only upon the positions of the two end
points Pi, Pi, but also upon the path of travel of the particle. However,
for a particular type of force field, called the conservative force field, the
force ¥{x,y,z) such a function of position that the above line integral
is
depends only on the end points and is independent of the path. As a

consequence, the line integral taken around any closed path in a con-
servative force field must vanish. For such a case, the negative of the
line integral is defined as the change of potential energy (denoted by V)
from Pi to Pi] i.e.,
V{P,) - F(Pi) = - f^^F'dr (8-40)
This definition of potential energy is consistent with the more usual

definition that the potential energy is the work done by the force as it
moves from some configuration Pi to some new configuration Po. Thus
-
V{Pi) =
P
-
F . dr ViP,) = j^' F . dr
and F(P2) V{Pi) = F • dr

jl^^
From this somewhat us examine, first, whether

artificial definition, let
such a conservative force can actually exist and, next, what is the
field
necessary condition for a force field to be conservative. After this we

shall show the significance of potential energy, which we have just
defined.
Let us consider the change of potential energy for a small displacement
whose components are dx, dy, and dz. From Eq. (8-40),
dV = -F 'dr = -F, dx -Fydy - F, dz (8-40a)
But since F is a function of position, i.e., V = V{x,y,z), we can form the

total differential of V as
dV = T— dx -f T— dy -\- —- dz
dx dy dz
Comparing with Eq. (8-40a), this means that
---^ ---I ---I' («-«)t
Hence the potential energy exists and the force field is conservative if
Eqs. (8-41) are satisfied.
From Eqs. (8-41) we can form the following partial derivatives:
dF^ ^ dW dFy _ dW
dy dy dx dx dx dy
t As an alternative derivation of Eqs. (8-41) see Prob. 8-18.

Seo. 8-8] E.ramples of consermlive Jurre Jields: urnrilaiion 173
For functions that are continuous and have continuous derivatives, the
order of differentiation is immaterial. Hence
^ - ^^ =
dx dij
dF BF
_?___«
Similarly. = (8-42)
dz dx
It can be shown that cither Eqs. (8-41) or Kqs. (8-42) are the necessary
and sufficient conditions for the existence of a potential energy.

These ideas can be convenienth^ expressed if we consider the concept
of a vector operator. Consider an operator V such that
dx dy dz
in terms of this operator we immediately see that Eqs. (8-41) can be

expressed (for a conservative force field) as
F = -TV (8-41a)
and Eqs. (8-42) may be written

V X F = (8-42a)
Thus we may say that F is derival)le from a potential function as illus-

trated by Eq. (8-4 la) and that the condition that a force field be con-
servative is given by Eq. (8-42a). Here TV is called the gradient of V
and V X F is called the curl of F. What is briefly presented now is
simply some results taken from the general study of vector fields and
will be covered more adefiuately in Chap. 15.
Combining Eqs. (8-39) and (8-40),
V2 - Vi = - f'F'dr = -{yzmv,' - l-jmvi'-)

J' (8-43)
1 1 + I'-zmvi^ = + y-zmvi-
1 2
Ilcnce, in a conservative force field, the sum of the potential energy and the
kinetic energy is a constant. This is the principle of the conservation of
mechanical energy.
8-8. Examples of Conservative Force Fields; Gravitation
A most fre(juent example of a c<)n.servati\'(; foice field is the assumed

uniform field of gravitation. If we confine our attention to mass particles
near the surface of the earth and if we neglec^t the small elTect of the
earth's rotation, the force of gravity is essentially constant and equal to
:
the weight of the particle; i.e., W

= mg. Moreover, the direction of
this force may be considered to be vertical. Thus we have a uniform
force field and
F" = —W (8-44)
The above equations clearly satisfy Eqs. (8-42); hence the uniform
gravitational field is a conservative force field. By Eq. (8-40), we find
that the difference of potential energy between two points at elevations
y\ and iji is
V, - Fi = - !"'
i-W) dy = Wy, - Wy, (8-45)
J yi
Hence we have the equation of the conservation of energy:
Wy -\- }4.'^v'^ = constant (8-46)
for a particle in a uniform gravitational field with no other force except

that due to gravity acting on it.
According to our definition of center of mass (or center of gravity in a

uniform field) (Sec. 5-2) it is clear that Eq. (8-45) also applies to a body
(or system of particles) provided W
is the total weight of the body and
the y distances refer to the center of the mass of the body (or system of
particles).
Another example of a conservative force field is the central force field
of Sec. 8-4. If we choose the central point as the origin, then the force
in a central field consists only of a radial component which is a function
of the radius only. This means that in the spherical coordinate system
(Sec. 6-8),
(«-*^)
I: : I'lt
It can be shown that this again satisfies Eqs. (8-42) ; hence this type of
force field is also conservative. The change of potential energy in this
case is
F2 - Fi = - jyr{r)dr (8-48)
One example of the central field is the force of attraction of two mass
particles mi and m-i, respectively, separated by a distance r (Sec. 8-4)
F.= -G'^ (8-49)
This is the well-known inverse-square law, Newton's law of gravitation.

Sec. 8-8] Eraniples of conserralirr force fields: (jnirlldliitn 17.')
For this case, the indefinite integral on the right side of Eq. (8-48) can
be easily evahiated to obtain
,r ^ m\m->
= -G —^-' ,
V -I- constant (8-50)
The constant in Eq. (8-50) can be taken as zero for a given datum.
Without modification, Ecj. (8-49) gives the vector force of attraction
between any two particles. In order to find the attracting force between
two bodies of finite size it then becomes necessary to add vectorially the
interaction between each particle of each body. That this procedure
is a cumbersome one is (juite evident. However, if we make use of the
concept of potential energy, we can determine the attracting force
between two finite bodies by scalar methods alone. For convenience,
let us further define a poictitial which is equal to the potential energy for
a unit mass. For example, the potential at any point P in a gravitational
field due to a particle of mass mi at a distance r from P, by Eq. (8-50), is
<f>
= —Gnii/r. Subsequently, the force components acting on a unit
mass particle at P due to wi can be found by the use of Eqs. (8-41).
Suppose we desire to determine the potential at a distance r away from
a homogeneous spherical body of radius a, such that r > a. Using the
results of Prob. 8-19, the potential at a distance r from the center of a
spherical shell of radius a is = —Gnii/r, where nii is the total ma.ss of
<t)
the This important result states that the potential at a point

shell.
situated outside a spherical shell is the same as if the spherical shell were
a particle concentrated at its center. Since the homogeneous sphere
may be regarded as a large number of concentric spherical shells, the
potential at a point outside a homogeneous sphere is <^ = —GM/r, where
now M
is the total mass of the sphere, and thus the force acting on a
unit mass at the point P is F/m = —{GM/r-)eT. This result has been
used previously in Sec. 8-4, where we idealized both the sun and its planet
as homogeneous spherical bodies and suV)sofiuently as particles.
On the point at which the potential is to be computed
the other hand, if
lies within the spherical shell, the results of Prob. 8-19 indicate that the
potential is constant and can be taken as zero. Thus, if we recjuire the

potential at some point interior to a homogeneous sphere, only the mass
located between the spherical surface of the same radius as the point and
the center of the sphere is of consequence. Again the force of attraction
is given as F/m = —(GM /r-)er = —fsirf/jrer, where M
is the total mass
in the sphere interior to the point in question.

A second example of the central force field is the force acting on a mass
particle due to a ma.ssless, elastic spring. The spring force is Fr = —At,
where k is the spring constant and r is the deflection (or change in length)
of the spring. Substituting into P]q. (8-48), we obtain the potential
energy due to an elastic spring:
V = Ukr^ (8-51)
We have above described three examples of conservative force fields.

The most obvious example of a nonconservative force field is that due to
friction, either sliding or viscous friction.
8-9. Example: The Simple Pendulum

In addition to Sees. 8-2, 8-3, and 8-4, illustrating the application of the
equations of motion, we shall give here another example involving also
the use of the principle of work and
energy. Let us consider a simple
pendulum consisting of a heavy bob
of mass m suspended by an inex-
tensible string of length a (Fig. 8-9).
This problem serves also to illustrate
what is termed constrained motion,
which implies that the particle is
compelled to travel along some fixed
path. In this case the path in ques-
tion is a circular arc of radius equal
Fig. 8-9
to the length of the string. The
constraining force, then, is the force
in the string. The upper end
of the string is fixed at a point 0. The
pendulum with the amplitude a, which represents the maximum
oscillates
angle of swing, the motion taking place in a vertical plane.
Let us take the potential energy to be zero when the particle is at its
neutral position, i.e., when the string is vertical. For any angle 6
between the string and the vertical, the kinetic energy is 'jy^ma^d- and
the potential energy is mga(l — cos 6). At the extreme position, or
when 6 =^ a, the kinetic energy is zero and the potential energy is mga
(1 — cos a).
The principle of conservation of energy therefore gives
3--2Wza-^- -f- mga(l — cos ff) = mga{l — cos a)
or e- = 2- (cos d - cos a) (8-52)

a
We shall use the notation

Sec. 8-9] Eraiuple: the .sim/)le penduluni 177
for convenience. Then \-a\. (8-52) eiin l)e written
at ^ —
M tig
p .
V 2(co's 6 — cos a)
Integrating and assuming that d = when t = 0, we get
"^^
pt = r^ —
(8-52a)
Jo \/2(cos 6 cos a)
Thus we have expressed the location of the penduhnn as a function of
time. Unfortunately, however, the integral involved functional in this
relationship cannot l)e e\aluated in terms of elementary functions. In

order to proceed, then, suppose we let
k = sm
.a- .6
sui = k
.
^m
.
:^ <t>
Then cos a = 1 - 2k- cos 0=1- 2A'- sin- <f>
and V2(cos 6 — cos a) = 2A: cos <t>
du = — 2k cos
,
d<l>
Vl - A'^ sm2 (t>
Hence Kq. (8-52a) becomes, using these sul)stitutions,
pt- r /r-4^^-T^ (8-53)
where the upper limit ol integration is the value of corresponding to the

particular value of 6 and is given by
.
_, sin (6/2)
sm {a/2)
The on the right side of Im}. (8-53) is Legendre's elliptic

integral
integral t of the first kind and is usually denoted by F{k,(i>), where
k is
called the modulus. Thus

pt = Fik.4>) (8-54)
The same functional relation between and / can also he e.\i)res.sed by
sin = sn (pt)
or sin .^
= sin - sn (pt) (8-55)
where sn (pt) is referred to as the first Jacobian elliptic function. It can

be shown that sn (pt) is a periodic function of Hence the angle 6
p(.
of the pendulum is a periodic function of /, as one would have expected.
t Values of this function are tabulated in E. Jahnke and F. Enide. "Tal)leof Func-
tions," Dover Publications, New York, 1945.
The T can be determined from Eq. (8-53) if we

period of the motion
observe that the period merely four times the interval of swing from
is
the neutral position to an extreme position. Furthermore, at an extreme

position Q = a and = 7r/2. Thus
'o a/I — ^^ sin^
or vT = 4F (k, = 4K(fc) (8-56)
where K{k) is called the complete elliptic integral of the first kind.t
Without going into details of elliptic functions, we can obtain a great
deal of information about the behavior of a simple pendulum in the
following manner:
1. If the amplitude a is so small that k^ can be neglected when com-
pared with unity, Eq. (8-53) becomes simply
pf =
But sin ^ = sin (0/2)/sin (a/2) ~ d/a for small a, and hence small 6.
This means
e =- a sin 'pt (8-57)
Hence the motion is simple harmonic with the circular frequency of

-p = ^'gja and the period of T = 27r \/a/g. This is the result usually
given in elementary physics.
2. For a better approximation, let us assume that a is small, but let us
keep the first-order small term. In other words, let us assume that
k « a/2 is a small enough quantity so that only the /c^ term is retained,
whereas higher powers of k {k^ and greater) can be neglected when com-
pared with unity. The integral on the right side of Eq. (8-53) can then
be expanded by the binomial theorem to become
f
*
(1 + y,k-- sin^ 0) d* = * + m- (I
- '^^
In order to find the period of the motion, we substitute (j) = 7r/2 (corre-
sponding to = a) in the above expression:
2 V ^ 4
This represents the amount of -pt for the pendulum to swing y^ cycle,
t Values of this function are tabulated in V.. Jahnke and F. Emde, "Table of Func-
tions," Dover Publications, New York, 1945.
Sec. 8-9] Example: the s'nnjAe petidnlum 179
from 6 — () io d — a. The period is this vahie nuiltipHed by 4/p, or
since k- = sin^ (a/2) ~ a-/4 for small \alues of a. Thus the period
increaseswith the amplitude. I'or an amplitude of 0.4 radian, or
approximately 22.9°, the error in the period computed from the
elementary, or first-order, solution of Efj. (8-57) is shown by Ecj. (8-58)
to be per cent.
1 Moreover, the error involved here in the second-order
approximation compared with the exact solution of Eq. (8-56) is negligi-
bly small. On the other hand, if the amplitude is 1 radian or more, even
the second-order solution of E(|. (8-58) becomes inaccurate, and still
higher-order approximations, or the use of Eq. (8-50), may become
necessary.
So far we have solved the problem through the use of the conservation
of energy. It is also possible to solve the problem by the equations of
motion. We shall use Newton's law expressed in terms of the normal
and the tangential components [Eqs. (8-3)]. It is most natural to employ
these coordinates when dealing with constrained motion, and it is par-
ticularly advantageous when the path is circular. Referring to Fig. 8-9,
we notice that at any angle 6,
Fn = -^ T - mg cos d = md'-a (8-59)

P
Ft = wi.s —mg sin 6 = mad (8-60)
In Eq. (8.59) and for the remainder of this section, T stands for the
tension in the string.

Integration of the second eciuation gives
ad- = 2^ (cos 6 + C)
Evaluating the constant of integration C by using the condition that

^ = for 9 = a. we obtain
ad- = 2^(cos d — cos a)
This is exactly Eci. (8-52). The solution from here on is therefore the
same as given before.
Let us consider once more the first-order approximation for small 6.
Assuming sin 6 — 6, Eq. (8-()0) becomes
61 + ^0 =
a
Letting p^ = g/a,
d -{- p^6 =
:
The solution is, according to Sec. (8-2) and the prescribed initial
conditions,
d = a sin pt
where a is the amplitude of 6. We see that this is exactly the result of

Eq. (8-57) and that the motion is simple harmonic motion.
The equations of motion, however, can tell us something else that the
method of energy cannot do. If we want to investigate the tension in
the string at any instant, we have to use Eqs. (8-59), which have not been
used so far. Substituting Eq. (8-52) into Eqs. (8-59), we get
T = mg{3 cos d — 2 cos a)
The minimum value of T occurs at the end of the swing when 6 = a:
T^min = '>ng cos a
The maximum value of T occurs when 6 = 0:
râx = mg{^ — 2 cos a)
For small a, cos a ~ 1 — a^/2; then
Tmin ~ mg ll -
^j
Tmax « mgil -\- a2)
The is advised that, in the discussion of this section, what is

reader
important not the knowledge of the simple pendulum; rather it is the
is
application of the equations of motion and the method of energy. In

particular, the following points should be noted
1. The energy method does not yield anything new that the equations
of motion cannot tell, since the method is merely an integrated form
(but an extremely useful one) of the equations of motion.
2. The energy method, however, is convenient for finding the speed.
Because we avoid having to find a first integral, a direct application of

work-energy principles yields the speed.
3. The equations of motion are used to find force or acceleration,
especially the force component normal to the motion, since this component
does not appear in the work-and-energy equation. In this respect it is
worthwhile noting that a combination method consisting of the determina-
tion of the speed by energy methods and the normal force by the equations
of motion is particularly efficient.
4. Instead of an exact solution, one can sometimes obtain approximate
solutions to a problem. The approximate solutions can furthermore be a

first-order approximation, or a second-order or higher one, depending on
the degree of accuracy needed and information desired.
See. 8-10] Motion of a jHirlicle in a nioring coordinale syslem 181
'). The concept of fiXMiueucy, circular t're(|uency. and period ol' an

oscillating motion, togetlier with their physical significance.
8-10. IMotion of a Particle in a Moving; (Coordinate System

There are many instances when it becomes desirable to express the
eciuations of motion with respect to a moving coordinate system. Parti-
cles traveling near the surface of the earth, for e.xample, are most usually
Fi<;. 8-10
described with respect to a coordinate system fixed on the surface of the

earth and rotating with it. If we denote F as the force acting on tho
particle, the Newtonian e()uations of motion referred to an absolute
coordinate system, by virtue of Eq. (7-21), become
F = m[R + e,
-j- w/ X p + w/ X (w, X o) + 2o)/ X p,] (8-01)
Rewriting this in terms of the apparent acceleration in the observable

coordinate system,
m'Qr F - w[R -f- 0)/ X p + o)/ X (u)/ X p) -f 2oi/ X prl (8-62)
Suppose we consider the motion of a particle in a coordinate system

fixed on the surface of the earth (idealized as a perfect sphere) at a
particular location. Furthermore, let us assume that a set of axes fixed
at the center of the earth can be considered as fixed and that the earth's
rotation w is constant so that <o/ ^ w and w/ = (l"'ig. 8-10).
From Eq. (7-15),
R = o X (o> X R)
and Eq. (8-62) now becomes
niQr = F - m{o> X [(0 X (R + 9)] + 2a) X pr} (8-63)
The fictitious forces — mw X [o> X (R + e)] and — 2mo) X Pr are usually

referred to as the centrifugal and Coriolis forces, respectively.
Thus the motion of the particle exhibits a slight deviation from that
predicted by the less realistic idealization of the earth as a nonrotating
sphere. For ordinary problems, the deviation is of little consequence
because of the smallness of |to|, but in some instances, such as high-altitude
ballistic missiles, this correction takes on greater importance.
Further, suppose that we have a "stationary" particle at 0', acted
upon by a gravitational force and a reactive force N. For this case, the
particle has zero acceleration with respect to the coordinate system on the
surface of the earth. Hence Eq. (8-63)
hY becomes, setting Pr = 9r = e = 0,
= GMm R mw X (w X R) +N
or,writing this in terms of the coordinate
system of Fig. 8-10,
N
GMm «'• mw'^h
R^
Here we have taken the gravitational

and b
force to be in the radial direction
to be the position vector of the particle
measured from the axis of rotation.
Thus the effect of the earth's rotation is
Fig. 8-11 and magni-
to slightly alter the direction
tude of the reaction force acting on a
particle at the earth's surface. Vertical is usually taken to mean the direc-
tion of the reactive force N. If we were to take as our model of the earth
the more realistic oblate spheroid, the gravitational force would deviate
slightlyfrom the radial direction e^ and, in turn, the reactive force N would
have a further slight deviation from the radial direction.
Example 8-1. We now consider the effect of the earth's rotation on
a freely falling particle. In Fig. 8-11, let the direction of the unit vectors
be specified as follows k : is and is positive
in the direction of the vertical
upward, i is in the horizontal plane (perpendicular to k) and points
southward, while J is in the horizontal plane and points eastward. If we
—
Sec. 8-10] Motion of a particle in a moving coordinate system 183
limit oursehes to a consideration of particles near the earth's surface, the

gravitational force may he taken as constant and acting in the vertical
direction, i.e.. F = — mgk. Furthermore, we have the approximation
that R +p =^ R. I'nder the.se conditions, Eq. (8-03) becomes
wjpr = —nigk — //([to X (co X R) + 2u) X Pr]
This equation may be further simplified by taking advantage of the

smallness of |a>| and neglecting the term —wo X (to X R), so that
m'Qr = — '«^k — 2///to X pr
where to = aj(— cos X i -f- sin X k)

X = latitude
Xow the scalar equations of motion for the freely falling particle take
the form
mx = 2m<x)ij sin X
mij = — 2mco(i cos X -j- i" sin X)
m'z — 2tncoy cos X — mg
Integrating the first and third of these and using the initial conditions
t = 0, x = ]j
= z = x = y = z = 0,
X = 2co/y sin X
i = 2oiy cos X — gt
Substituting these into the second of the e(iuations of motion and neglect-
ing terms in w-,
y = 2cogt cos X
Two integrations yield
y = ojg — COS X
whereupon, substituting this expre.ssion into the expressions on the speeds

X and i and neglecting the terms in w'-, sub.se(iuent integration yields
a; =
It is therefore clear that a freely' falling body suffers a deviation to the

east due to the effect of the earth's rotation. For a free fall from a height
h, the time of fall is th = \^2h/g and tlu; diniation becomes
y = —
2o}h cos X
5 J2h
PROBLEMS
8-1. A mass in rectilinear motion is acted upon by a force such that the displace-
ment is given asx = A cosh wt + B sinh ut. Describe the nature of the applied force.
8-2. Determine the motion of a mass resulting from the action of a repulsive force
proportional to its displacement. The initial conditions for the motion are, at t = 0,
X = xo, X = 0.
8-3. A projectile is shot vertically upward with an initial velocity of vo. If the air
resistance has the magnitude of kv^, find (a) the height reached by the projectile; (6)
the velocity of the projectile when it returns to the ground.
8-4. The resistance of the air to the fall of a raindrop can be considered propor-
tional to the velocity of fall. If the terminal velocity of the raindrop is found to be
20 fps, find the distance of the fall for the speed to reach 99.9 per cent of the terminal
velocity. (Assume that the initial velocity of the raindrop is zero.)
8-5. A particle weighing 100 lb travels horizontally in a viscous medium. The
drag force is proportional to the velocity. It is found that when the speed is 40 fps
the drag force is 60 lb.
(a) Find the distance of travel when the speed slows down from 40 to 10 fps because
of the viscous medium.
(b) Find the time elapsed for part a.
8-6. If a rocket for space travel does not possess any means of propulsion of its
own after it is projected from the ground, find the speed of projection required so that
the rocket can escape from the earth's gravitation. The gravitational force is
inversely proportional to the square of the distance between the object and the cen-
ter of the earth. The radius of the earth is 4,000 miles.
8-7. A particle falls
from rest in a medium that possesses a velocity type of resist-
ance of the form —kx^. Determine the terminal velocity of the particle.
8-8. A shot putter is capable of throwing the shot at an initial speed Vo. His hand
is at a height h above the ground upon release of the shot. Determine the angle <^
from the horizontal at which he should throw the shot in order to achieve maximum
range. Calculate this maximum range. (Neglect air resistance.)
8-9. A particle of weight W is traveling around a parabolic curve as specified in
Example 6-4. Find the reaction between the particle and the surface at its lowest
point of travel.
8-10. A particle traveling in the xy plane is acted upon by an attractive force (or
spring) proportional to the distance of the particle from the origin. Using the rec-
tangular coordinate system, determine the path of the particle. The initial condi-
tions are, al t = 0, x = Xo, y = 0, x = 0, and y = Vq.
8-11. A satellite is launched under the following initial conditions: ro, Vo, vô = 7r/2.
Classify the various possible orbits as functions of the launching speed.
8-12. The apogee and perigee (measured from the earth's surface) of the following
three artificial satellites are estimated to be as follows:
Sputnik I Explorer I Vanguard I
Apogee, miles. 59 1,585 2,460

Perigee, miles 145 219 404
Compute tlic period of cacli of the tlircc satellite Also coinjjute the orbital speed of
each at the perigee point.
Problems 185
8-13. An artificial satellite is designed to reach a circular orbit at an altitude of
550 miles. During launching the designed orbital speed is met at the re<iuired alti-
tude but the launching angle at rocket burnout is not exactly tangential to the earth's
surface as designed. Instead of a circular ()rbit, the satellite will follow an elliptical
orbit. Calculate the variation of a|K)gec and perigee with launching angle, using the
following values for the deviation of the launcliing angle: 1°, 2°, and '.i°.
8-11. ABCa smooth rail in the form of a vertical semicircle of l-ft radius.
is .\
particle weighing 50 lb slides along the rail. /-* is

a force of 150 lb alwajs inclined at
:iO° to the horizontal. Q is a force of 40 lb always directed along the tangent. Com-
pute the work done on the particle in nioving it from .1 to H. If the initial speed
at A is 10 fps, find the velocitj- of the particle when it reaches li.
P 8-14 Fk;. V 8-15
8-15. The force acting on a particle is given by
F = {Ax"- + liyYi + r'//2j
Show that the work done moving the particle around the triangular
b^- this force in
path OabO is —y^Bab. system conservative?
Is this force
8-16. A force system is given by F = yzH + {xz"^ — 1 j + 'iixijz — l)k. Show )
that the force is conservative and find the potential energy.

8-17. Express the kinetic energy of a particle in terms of the cartesian, cylindrical,
and spherical coordinate systems.
8-18. Show that the neces.sary and suflicient condition for the existence of a poten-
tialenergy as expressed by Eqs. (8-41) can also be obtained if the right side of Eq.
(8-40a) can be expre.s.sed as a perfect differential.
8-19. Find the potential for a point located a distance from the center of a liomo- /
geneous .spherical shell of radius a. The point is located outside the shell, i.e., r > a.
If r < a, what is the potential?
8-20. For a simple pendulum compute the period of oscillation according to the
methods of Eqs. (8-50) to (8-58) for amplitudi's of a = 30°, 75°, and !)0°. Compare
the results.
8-21. \ small bead of ma.ss in slideswithout friction along a vertical circular curve
of radius a. The l)ea<l starts at from an initial position, nuikiiig an angle of do
re.st
with the vertical axis. P'ind the reaction between the beatl and the wire. Fiiwl the
jM-riod of oscillation of the bead.
8-22. If the mass at the end of a pendulum travels around a circle in the hori-
zontal plane, the device ispendulum. Letting the radius of the circle
called a conical
of travel be r, the length of the string a, and the mass m, find (a) the tension in the
.string; (b) the angular speed 0; (r) the frecpiency and period of the motion.
Fig. P 8-22
8-23. A particle sitting at the top of a hemispherical dome is permitted to slide

down the surface without friction. At what angle </> will the particle leave the surface
of the dome?
Fig. P 8-2:^
8-24. A device called "centrifugal pendulum" has been used for the reduction of
torsional vibrations in aircraft engines. can be illustrated by a simple
Its principle
pendulum AB with one end A attached to an eccentric point that rotates with the
shaft and the other end B holding a mass m. Assume the distances OA = r and
AB = a.
(a) If the pendulum is deflected from the radial direction by a small angle it will (j),
oscillate back and forth about the radial direction. Set up the differential equation
of (f)
for small values of 4>.
(b) Find the frequency of oscillation of the pendulum.
8-25. A stretch of the Mississippi River at latitvide 40° during flood conditions is
1.5 miles wide. It is flowing to the south with a speed of 10 mph. Calculate the
difference in level between the two lianks of the river. Wliich bank is higher?
Prublenis 187
8-26. A heavy bead of mass m is free to move on a smooth c-ireular wire of radius a
w hich rotates with constant angular speed w about a fixed vertical diameter.
(a) Find the neutral position (i.e., the angle 6) of the bead for w' > g/a.
{b) If oj- > (j/a, find the period of small crscillations about the neutral position.
Fic. P 8-'_>()
8-27. .\ heavy particle moves with constant speed along a horizontal circle upon
tlicinner wall of an inverted right circular cone who.se axis is vertical. The half-cone
angle is a, and the circle of travel of the particle is at a height /( from the vertex of
the cone. Friction is negligible.
(a)Find the velocity of the particle.
(6) If through some slight disturbance the particle is shifted slightly in elevation
along the wall of the cone, show by differential equation that the particle will oscillate
up and down along the walls. Find the frequency of such small oscillations.
Fic. P 8-27
8-28. A particle at rest is dropped from a height of 100 ft. Find tiic deviation of
full due to the earth's rotation.
CHAPTER 9
Vibrating Systems
9-1. Generalized Coordinates and Degrees of Freedom
The independent variables used to completely specify the configuration

of a system are called generalized coordinates. The number of generalized
coordinates is called the number of degrees of freedom of a system.
To illustrate the above statements, consider a system consisting of
n particles. Since one particle requires, in general, three coordinates,
say X, y, and z, to completely specify its position, one might at first
conclude that 3n coordinates would be required to specify the configura-
tion of the entire system. This is indeed true if there are no constraints
whatsoever that limit the motion of the system. In the majority of
applications, however, constraints in one way or another exist. Thus, as
we have already seen in the case of a rigid body, the constraint that
distances between particles remain unchanged reduces the total number
of independent variables to only six for the complete specification of the
position of the rigid body. " Another example is the case of a bead con-
strained to slide along a given wire. In this case we need only one
variable, namely, the distance along the wire from a fixed point to the
bead, for the specification of its position. We thus see that constraints
impose certain relations between the otherwise independent variables
and therefore reduce the number of independent variables of the system.
Constraints also imply that there exist certain internal or external forces
which, although not given a priori, must produce a net influence on the
system to confine its motion in a prescribed manner.
The following outline ilkistrates the previous discussion by means of
nine examples:
188
Sec. 9-2] 1 ihnilions 189
(ieiieraliz«Ml
190 Vibrating systems
general periodic motion. The displacement repeats itself after a time

interval T, which is defined in Sec. 8-2 as the period of the motion.
Vibrations of mechanical systems are encountered by engineers in
many problems, of which the most obvious ones are to be found, for
example, in rotating machinery and in structures under dynamic loads.
In addition, there are countless phenomena in other engineering or
technical fields that are governed by similar mathematical models.
Fig. 9-2
The simplest type of periodic motion is the simple harmonic motion

of the one-degree-of-freedom system described in Sec. 8-2.
X = Xm sin o)t
X = X^CO COS wt (9-1)

X = —x,nCo- sin c>)t = —co'^x
The period of this motion is T = 2x/co, and the natural frequency is

/ = o)/2w. The quantity w is called the natural circular frequency, and
Xm is termed the amplitude of the motion. Figure 9-2 shows the variations
of displacement, velocity, and acceleration with respect to time, for the
simple harmonic motion of Eqs. (9-1).
It is obvious that the simple harmonic motion represented by
X = Xm COS wt
is identical with that of Fig. 9-2, except for a phase shift of one-quarter
of a period along the abscissa.
motions in general, however, are not necessarilj^ simple
Periodic
harmonic. The motion shown in Fig. 9-1, for example, is periodic but
not simple harmonic. It is a great mathematical discovery that any
periodic motion can be considered as the sum of a series of simple har-
Sec. 9-3] One-degree-of-freedom vibrating systems 191
monic motions. Consider any periodic function /(/) with the period T
so that /(/ + 7')
/(/). =
Subject to certain restrictions, which are of no
concern in the present discussion, one can write
f{t) = ao + Oi sin wt + a-, sin 2u)t + as sin Soit + •

•
+ bi cos (Jit + 6-2 cos 2oit + bs cos 'Scjt }-••• (9-2)
where w = 'Iw/T. Alternatively, one can also write (Prob. 9-1)
fit) = .4o + .4i sin (ut + <î) + .4 2 sin i2u)t -\-

<t>-2)
-\- A 3 sin (3a)/ -\-

(t>,) + (9-3)
The infinite series represented by Eq. (9-2) or (9-3) is called a Fourier

series. The terms Oi sin mt and 61 cos ut in Kq. (9-2) and the term .1 1 sin
(oit -\- <f>i) in E(\. (9-3) are called the fundamental, or the first harmonic,
of /(<)• In general, the terms a„ sin nut and 6„ cos no}t in Eq. (9-2) and
the term .4„ sin {ncot + 0„) are called the nth harmonic of /(/). The
and An in Ec}. (9-3) represent the amplitude
coefficients a„, 6„ in Ec}. (9-2)
of the various harmonics, and the angles 0„ in Eq. (9-3) are called the
phase angles. The method of determining a„ and 6„ (or equivalently,
-4„ and 0„) for a given f{t) is called a Fourier analysis. It might be
added that the constant term Oo or ,4o represents the average value of
f{t). In the event that/(0 is not periodic over the range of interest, the
Fourier-series expansion may still be utilized by extending the range and
defining f{t) such that it is periodic within the new range.
9-3. I)i(T"erenlial Ec|iiations of One-degree-of-freedom

Vibrating Systems
The motion of a mass constrained to move along a straight line has

one degree of freedom. A single-degree-of-
freedom vibrating system can be idealized by
a mass suspended at the end of a spring (Fig.
9-3 ) . Let us assume that the mass is acted on
by ( 1 ) a linear spring whose spring force is pro-
portional to its change in length, (2) a viscous b
resistance proportional to the velocity, and (3)
an external, or exciting, force. We found in
Sec. 8-1 that the differential equation of motion
of the sy.stem, expressed in
placement .T measured from the neutral or
terms of the dis-
S
equilibrium position, is J Fit) 1x
Vw.. \\-.\
wx -I- ci -I- kx = F{i) (9-4)
Let us now review the significance of the various ([uantities appearing

in the above e(iuation:
— : ;
1. The restoring spring force acting on the mass is —kx, where k is the
spring constant, i.e., the number of units of force required for a unit
deflection of the spring. The negative sign of the spring force means
that the force always opposite to the displacement. This force is
is
therefore a restoring force. In the following discussion, the spring itself

is assumed to be massless. We note that only the restoring spring force
enters into the equation of motion, since we are measuring the displace-
ment of the system from the equilibrium configuration. The total
spring force is, of course, the restoring force plus the static force in the
spring.
2. The viscous resistance is —ex, where c is the coefficient of viscous
damping, i.e., the number of force units of resistance per unit velocity
of motion (cf. Sec. 2-5). The negative sign of this resistance shows that
the resistance is always opposite to the velocity. This resistance is also
called the damping force. We often purposely introduce a damping
force in a vibrating system by means of, for example, a dashpot filled
with oil.
3. The external force F(t), also called the exciting force, is in general
some function of time. The most important kind of exciting force is a
sinusoidal function of time, given before as Eq. (8-10),
F(t) = iô sin o:t (9-5)
where Fo = amplitude
CO = circular frequency
An exciting force that is periodic but not sinusoidal can be written as a
Fourier series of an infinite number of terms, or harmonics (Sec. 9-2)
each term, or harmonic, is a sinusoidal function of time of the type of
Eq. (9-5). Since the differential equation (9-4) is linear, its solution
for a general periodic exciting force can be obtained by adding the solu-
tions for each harmonic of the exciting force.
For the type of exciting force represented by Eq. (9-5), the differential
equation (9-4) may be written in the following standard form
X -{- 2nx + plT

^l{t)_
mm ^Fo g.^^ ^^ ^g_g^
where p- = — and n = ^^ (9-7)

m, 2m,
The reason for introducing the new parameters p and n will be made
clear very soon.
The problem of the idealized mechanical vibrating system described
above one of the more important problems of mechanics, partly because
is
the idealized system possesses all the essential features of more compli-
cated vibrating systems, and partly because the differential equation (9-4)
:
Sec. 9-1] Free vibraliotis wiUioul daniphig 193
can also arise in other entirely ditTerent physical problems, for example,
in the theory of electrical circuits. Let us consider a series circuit
(Fig. 0-4) consisting of a variable voltîgc source E = K(t), a resistance R
(ohms), inductance L (hcnrys), and a capacitance C (farads). If i is the
instantaneous value of the current (amperes) in the circuit, then the
voltage drops (volts) across the three elements are as follows:
Across R : Ri
di
Across L :
J
Across C
where Q = ji dt is the charge (coulombs). E(iuating the applied voltage

source to the total voltage drop across the three elements, we have, since
i = Q,
LQ + 7?0 + iQ = E{t) (9-8)
This is the differential etiuation of an RLC circuit. It can lie seen that
Ecj. (9-8) can be recast into the form
of Eq. (9-f)) if we let
'iZ
p-
LC
ajid ^ = ^ (9-9) Eit)
The of such an electrical

solution
system is therefore identical with that
Fi< '.t-t
of a mechanical vibrating system.

The complete analogy between an electric circuit on the one hand and a
mechanical vibrating system on the other is the basis of the modern
electrical analogue computers, which indicate the solutions to mechanical
problems from the response of an electric circuit consisting of suitable
combinations of electrical elements. Many other examples of physical
phenomena defined by an equation of the form of Ecj. (9-()) are available
in the technical literature.
9-4. F'ree \ ihration.s without Damping

When there is no exciting force in Eq. (9-()j, th(^ vibration is said to be
free. The differential equation, from Efj. (9-0), becomes
X -^2nx -\- p^x = (9-10)
If in the system there is neither exciting force nor damping, the differ-
ential equation is further reduced to
X -^ pH = (9-11)
This is the same type of equation that we discussed in Sees. 8-2, 8-9,
and 9-2. The solution is a simple harmonic motion represented, in
general, by
X == A cos pt -\- B sin pi (9-12)
or alternatively (Prob. 9-1),
X = C sin (pt + (j>)
The general solution of Eq. (9-11) can be obtained most conveniently

and generally by the method of trial functions. Suppose we try a solution
of Eq. (9-11) in the form
X = De'^
Substituting this trial function into Eq. (9-11), there results, after simpli-
fication, the following characteristic equation:
s2 -f p2 =
Thus the trial function constitutes a solution for values of s as given by
Si,2 = dzip, where i = \/— 1. Since we are dealing with a linear dif-
ferential equation, the total solution is composed of the sum of the two
individual solutions multiplied by arbitrary constants. Therefore

X = Die+^p^ -f- D2e~^p^
It is convenient to transform this solution by means of the relations
e*^ = cos y -f i sin y

Q-iy — pQg y _ ^ g^-^
y
The total solution now becomes

X = A cos pt -\- B sin pt (9-12)
where the two constants of integration A and B (or C and <f>) are to be
determined by the two initial conditions. If at ^= 0, a; = Xo and
X — ±0, we have then
A = Xo 5 =:= -
P
The solution is therefore
X = Xo cos pt -\ sin pt (9-13)
Consider the case of a mass at the end of a spring being displaced away
from its neutral position by an initial distance Xo and then released. The
initial conditions for this case are x = Xq and i = at the instant of
Sec. *-)-\] tree vibrations ivilfwul damping 1*).')
release, at which instant we assume t = 0. Equation (9-13) then

l)ecomes
X = xo cos pt (9-14)
The initial displacement .ro in this case is also the maximum displacement
.r„,. i.e., the amplitude of the resulting vibration.
The period of the free vibration represented by Imj. (9-13) or (9-14) is
r = | = 2jf (9-,5)
The period is therefore independent of the amplitude.

It is convenient to call p = vA'/m the natural circular frequency.
The natural frequency of an undamped vibration is
= 1
= -L /I
/
•^
r f
27r 27r \m (9.10)
^
It can be seen from Eq. (9-13) or (9-14) that an undamped free xibration
lasts forever.
A different method, based on the energy eciuation, can also be applied
to find the period of a free undamped vibration. Assuming that the
motion is harmonic with amplitude ecjual to Xm, we can write that
X = x„, sin o}t (9-17)
Differentiating once with respect to time gives
X = x„, (ij cos n:t (9-18)
where the circular frequency oj is as yet unknown. The net force acting
on the mass, in a free undamped vibration about an eciuilil)rium position,
is the spring force —kx, which is conservative. The potential energy
due to the spring force is, according to Eq. (8-51), V = kx-/2. The
kinetic energy of the mass is mx'^/2. The principle of the conservation
of mechanical energy then gives
3-^mi-'- -\- y-^kx- = constant (9-19)
Consider the following two positions of the mass: the neutral position at
which X = and x = xô} and the extreme position at which x = Xm
and i = 0. According to Eq. (9-19),
j^niXmW = l-2kx„^
Hence o, = J- = p (9-20)
The frofiuency is therefore the same as that found by solving the differ-
ential equation (9-11).
It should be noted that the energy method described above gives the
exact frequency of vibration only if the vibration is truly simple harmonic.
Many practical problems are so complicated in nature that the solution
of the differential equation becomes impossible. For such cases, the
vibration can be assumed as simple harmonic, just as we did in
still
Eq. (9-17), and an approximate frequency can be found by the energy

method. The frequency so found, however, is only approximate because
the true motion may not be simple harmonic in general. This generalized
energy method for estimating the approximate frequency of vibration is
called Rayleigh's method.
Example 9-1. There are many examples of oscillating mechanical
systems that can be idealized as shown in Fig. 9-3 or
as defined by Eq. (9-11). The small oscillations of a
pendulum were one such example. As an
(Sec. 8-9)
additional example consider the following. A small
weight W is suspended between two tightly stretched
wires whose tension at the equilibrium position is To

.W
(Fig. 9-5). Assuming that the force in the wire is pro-
portional to the change in length of the wire, we have,
at any arbitrary position of the mass defined by x, its
displacement from the neutral position, that the force in
each wire is
^777777777:^
Fig. 9-5
To +U 1 1 +
where k is the spring constant of each wire, Further, the horizontal
component of the force is
r. = To +n
z vi + ix/iy
If we limit ourselves to consideration of small oscillations of the mass

about its neutral position, we have x/l <<C 1, so that {xlt)- may be neg-
lected when compared with unity. Thus
T. = Tc
I
The equation of motion of the mass now becomes
—X = -2Tof
9 I
or X -p IT/ 7
^
Comparing this result with Eq. (9-11), we see that the mass under-
Sec. 9-3] Free vibralions trilli (hunpiurj 197
goes a simple h armonic motion with a natural circular fre(|ucncy of
9-5. Free ^ ibrations with Damping

The differential equation describing a free vibration with damping
has been given in the previous section [Eq. (9-10)]:
.f -|- 2nx -\- p-x =
To solve this equation, let us again use the trial function
X = Ce-" (9-21)
The differential etiuation is thereby reduced to the algebraic ecjuation
s- -\- 2/(.s + p- =
whose two roots of .s are
si,,. = -n ± \/n^ - p- (y-22)
Hence the total solution is
X = Cie'^' + Cê'-^' (9-23)
where Ci and C2 are two constants to be determined l)y the two initial
conditions.
The physical significance of the
above solution depends on the relative 1.0
magnitudes of n- and p'. We there-
fore distinguish the following three
cases.
Case I. Ifn- > p-, then both Si and
s-2 are real and negative and are given
as
Si,2 = —n±r
Fi«i. 9-G
where r = [x/n^ — p-| is a real, posi-
tive number such that n > r. The two functions e"'' and e'-' are therefore
exponentially decreasing functions of time (Fig. 9-G). The complete solu-
tion of Ecj. (9-28) represents, in this case, a creeping baeli to the neutral
position. The details of the motion can be easily worked out. Suppose
that the initial conditions are such that at = 0, x = .ro and x = 0. <
Writing the solution [K([. (9-23)] in the form
and solving for the constants of integration according to the initial
conditions,
X = P [(n -f r)e-("-'^' - (n - /•)c-^"+^^'] (9-23a)

2r
.
For other initial conditions the response can be obtained in a similar

fashion.
Three typical cases of such a motion are shown in Fig. 9-7, in which
curve I is for a;o = and curve II is for Xo = 0. When neither xo nor
±0 is 0, the motion may be of the nature of either curve II or curve III.
Because of the large amount of damping (indicated by a large n), the

mass in such cases does not have enough kinetic energy left after creeping
back to its neutral position to allow itself to oscillate around the neutral
position. Such a system is therefore said to be overdamped. It can be
Fig. 9-7
shown (Prob. 9-11), in general, that once the mass passes through the
equilibrium position (zero displacement), it can never pass through the
zero position again and can only approach zero as / °o ^
Case II. A system is critically damped if n^ = p^ or c = 2 \/mk. We
shall call this value of c the critical damping coefficient Cc', i.e..
Cc = 2 -s/nik = 2mp (9-24)
The solution of a critically damped system is, from the theory of linear
differential equations for repeated roots of the characteristic equation,
X = {Ci + CaOe""' (9-25)
where Ci and C2 are arbitrary constants, to be determined by the initial
conditions.
The nature of the motion in the critically damped case is somewhat
similar to the motion indicated in Fig. 9-7 for the overdamped case. In
other words, the mass tends to creep hack to the neutral position. This
case represents the transition between the overdamped creeping-back
behavior of Case I and the underdamped vibratory behavior of Case III.
Case III. If n^ < p'-, the system is said to be underdamped. We can,
from Eq. (9-22), write
Si,2 = —n ± i \/p- — n- = —n ± iq
1
>'<• ^^-j] Free vibrations ihmi/flny

ivilli 199
where i = v—
q = \-\/p^ — n^\
It can be seen that for an underdamped system, q is a positive real nuinl)er.

Sul)stituting into Vai. (9-23), the complete solution for this case is
X = e~"'(Cie'"' + C-iC-"")
Using the relations
e'" = cos y -{- i sin y
g-iy — gQg y — i sin y
we can write the sohition in the following form:
.(: = e""'(.l cos ql +B sin qt) (9-20)

or alternatively,
X = Ce~"' sin {qt + 0) (9-27)
where the constants A and B in Eq. (9-26) or the constants C and in

Eq. (9-27) are to be found by the initial conditions. For example, if the
X
the period of oscillation by decreasing the circular frequency from p to q.
We call
q = Vp' - n^ - V Vl - {c/coY (9-28)
the damped natural circular frequency.

Considering now any two neighboring peaks a and b differing by one
period, as shown in Fig. 9-8, we find that the ratio of the two amplitudes is
Xa _ e~"'i(A cos qti -\- B sin qti)

Xb
~ e-"(«i+r)[A cos q{ti -\- T) -{- B sin qiti -\- T)]
= e"^ = e^^"/" (9-29)
The natural logarithm of this ratio is called the logarithmic decrement; i.e.,
a = log £.
= 2™
Xb q
The logarithmic decrement is an indication of the amount of damping

in the system and is equal to the difference of the logarithms of any two
successive positive maximum displacements. It can be seen that since
Xa/xb is a constant, depending on the properties of the system, the
successive positive amplitudes of oscillation decrease in a geometric
series.
9-6. The Case of Negligible Mass;

The Differential Equation tx -\- x =
In the system of Fig. 9-3, consider the special case of a small negligible
mass at the end of a stiff spring and under large viscous damping. If
the mass in this case is released from an initial displacement Xo, what
is its subsequent motion? This question can be answered by solving
the differential equation
ex -\- kx = (9-31)
which is obtained by omitting the terms mx and F(t) in Eq. (9-4).

Write Eq. (9-31) in the following form:
ri + a: = (9-32) t
where t = c/k. For reasons that will be made clear shortly, r is called
the relaxation time, or time constant. Thus
r^-^dt =
x
t The student should solve this differential equation by the method of trial
functions.
Sec. 9-6] Tfie case of negiujible mass; (he eqnalion tx }- x = 201
Integration gives
t = —T log X -\- constant
or X = Cie,-"^
The constant of integration (\ is evalualcHl hy tlie initial condition x- Xo
at = 0. The solution is
<
X = x^-'ir (9_33)
The motion of Eq. (9-3.'^) is shown by the curve in Fig. <)-<). We see that
the relaxation time, or time constant,
has the following significance: (1) it is
the time at the end of which the deflec-
tion reduces to \/e of its initial value;
and (2) it is the intercept on the ab-
scissa of the tangent to the curve
Sitt = 0.
We can also pose our original (jues-

tion in a slightly different way. Let
the mass be initially at rest in its neu- Fi(i. 9-9
tral position. A constant force F is
applied to the mass and is kept there indefinitely. What is the resulting
motion of the mass? We see that the differential ecpiat ion goxcniing the
motion of the mass, for t > 0, is
F
TX -\- X = (9-:54)
The solution of the abo\e e(|uation for constant F//c is
X = Cie-"' +^ (9-35)
We notice that the first term on the right-hand side is the general solu-
tion, or the solution to the homoge-
neous equation tx -\- x = 0, and the
second term is the particular solution,
or that value of x such that tx -f x
equals F/k. The evaluation of the
constant of integration by the initial
condition x = at t = yields the
solution
X = y(] - e-"') (9-36)
Fk;. 9-10
The motion of K({. (9-36) is shown in
Fig. 9-10. The deflection is shown to approach its final value/*'/// asymp-
totically. The significance of the time constant t is again demonstrated.
The problem of a negligible mass discussed abovean is in itself not

important one. However, it problems which
illustrates a large number of
are characterized by the fact that the rate of change of the variable in
the problem is proportional to the variable itself. Such problems are
described by the differential equation (9-32) or (9-34), in which r may
in some cases be negative. To cite only a few examples:
1. A mass particle traveling in a viscous medium (Sec. 8-3). The
speeds x and y in this problem correspond to the variable x in Eqs. (9-32)
and (9-34), respectively. The speed x therefore follows the curve of
Fig. 9-9, and the speed y follows the curve of Fig. 9-10.
2. The closing of a switch S of an electric circuit consisting of a con-
stant-voltage source E, sua inductance L, and a resistance R (Fig. 9-11).

The differential equation describing the
current i in the circuit is
e:=:
Li + Ri = E
The increase of current with time after
R *S is closed follows the curve of Fig. 9-10.
FiG. 9-11
The time constant in this case is L/R.
3. If in place of the inductance L in
Fig. 9-11 we have a capacitance (capacitor) C, the closing of the switch
represents charging of the capacitor. The charge Q of the capacitor is
described by the differential equation
If Q = at i = 0, the increase of Q with time is again represented by

Fig. 9-10. The time constant is, RC.
in this case,
4. The cooling of an object in a moving stream of air. The rate
of cooling is proportional to T — Ta, the difference of temperature
between the object and the air. The differential equation for T is
t + k(T - To) ^ and hence is of the type of Eq. (9-34).
5. The growth of population, if the rate of growth is proportional to
the population. The differential equation here is i = kx, again of the

type of Eq. (9-32), where k is a constant of proportionality.
Other examples are too numerous to mention,
9-7. Forced Vibrations without Damping
We now come to the case where an exciting force is applied to the

system. The resulting motion of the system is called a forced vibration.
Let us assume that the exciting force is a sinusoidal function of time.
We also assume at first that there is no damping. The differential
Sec. 9-7] Forced vibrations wilhout damping 203
equation descril)iiig the motion, after setting n = in Imj. (U-0), becomes
Fo
X -^ p- = ^sm coi (9-37)
The
general solution of Eq. (9-37) consists of a homogeneous solution,
i.e., a solution of x +
p-x = 0, plus a particular solution of the non-
homogeneous eciuation. In order to oiitain the particular solution, let
us try the solution .r = x„, sin u)t. Upon substitution of the trial function
into Eq. (9-37), we find that
= ^'
ni{p'^ — 0}^)
Since p- = k/tn, we can write the particular solution as
Hence the total solution of the ditTerential e(iuation (9-37) is
F /k
X = A cos pt -\- H sin pt -f -,
1 - —
-. ;-—
(aj/p)2
;: sin mt (9-39)
provided that u) 9^ p. The constants of integration A and B are to be

determined by the appropriate initial conditions.
It is recalled that the first two terms on the right side of the above
equation constitute the solution to the homogeneous differential eciuation
X -|- p'-x = (Sec. 9-4). The last term is the particular solution of the
complete differential equation (9-37). The combined motion represented
by Eq. (9-39) is therefore the sum of two simple-harmonic motions: an
undamped free rihration with fre(juency p and an undamped forced
vibration with frecjuency oj etjual to the forced or impressed frequency.
It might be noted that this combined motion is in general not periodic
unless (1) A and B are zero or (2) the ratio i)etween p and w is a rational
number.
Considering general starting conditions, such that at t = 0, x = Xo
and X = Xo, the motion of the ma.ss is given by
X = Xq cos pt -i sin pt + ,
—— —'wVp , -, ( sin ut sin pt ] (l)-39a)
p 1
\ P /
and consists of both a free and a forccui \ibration. i^ven it the ma.ss
started from rest so that xo = Xo = 0, the solution is
and still consists of both a free and a forced \ibration.

.
now only the forced vibrations. The response amplitude due

Consider
to the forcing function (the amplitude of the particular solution) is,
from Eq. (9-38),
Fo/k
Xm =
1 — coVp^
and is seen to be a function of the frequency ratio oo/p. Since Fo/k = 8st
is the static displacement of the mass under the action of a slowly applied
1
|AF
FJk l-(coVp^)
2 \
\
\ V.
1
-1
-2
-3
Fig. 9-12
constant force Fo, the response amplitude may be written
x^ = 5.,(AF)
where AF is the amplitude factor (or amplification factor)
1
AF =
1 coVp^
and represents the ratio of the dynamic to static displacements. Its

value depends only on the frequency ratio w/p. A plot of AF against
shown
oo/p is in Fig. 9-12. When co/p is larger than 1, AF is negative.
The negative sign is not of great practical importance; it merely means
that X is 180° out of phase withF. Our chief interest is in the magnitude
of AF. For this reason the magnitude AF is plotted as a dashed curve
for co/p > 1.
One can from Fig. 9-12 that for w/p close to 0, x,„/(Fo/k) is approxi-
see
mately equal to unity. Physically, this means that for extremely slow
variations, F(t) can be considered to be a static force and the displacement
Sec. 9-7] Forced vihrulions ivilltoiil (hinij)liu/ 20.3
of the mass is merely' the static deflection corresponding to F; i.e.,
./ ~ F/k. At the exciting force oscillates with a frecjuency

oj/p close to 1,
that is nearly the same as the natural, frequency of the system and the
response amplitude .r,„ becomes very large. The system is then said to
he in resonance. Resonance indicates physically that if the disturbing
force acts on the mass at the proper time and in the proper direction, the
displacements of the sj'stem ma}' become dangerou.sly large. Finally,
when co/p is very large, Xm becomes very small and approaches zero as a
limit. Physically, this means that when the exciting force changes
>ign too rapidl}', the mass refuses to follow, and therefore moves only
-lightly, although with the same frequency.
If u = p, the differential equation (9-37) becomes
X -{- p~ = — sm pt
m
The particular solution to this differential efjuation is not periodic as
is the case for o: ^ p. Instead, we have the following particular solution:
X = — ^
2mp
—
Fot
cos pt
and the total solution becomes
Fot
X — A cos pt +B sin })t — ,^—^ cos pt
2mp
Evaluating the constants of integration for general starting conditions,
we find that the total solution is
X = Xo cos pt
, ,
-\
io
p
sm
.
pt
,
+ Imp
,
,^;
^0
— —/sin
I
\ p
/;/
— t cos pi
\
I
The same result would have been oljtained if we had evaluated Eq.
(9-39a) for the condition of = p. The (valuation leads to an indeter-
o>
minate form whose limit may be determined by using L'llopital's rule.

We see that the amplitude of this motion increases continuously with
time. Thus resonance does not occur instantaneously, but rather
refjuires an infinite amount of time. However, after a short period of
time the displacement grows to excessive limits. It is for this reason
that the resonance of a mechanical .system acted upon by a ])eriodic force
should be avoided.
Example 9-2. One useful application of the foregoing theory lies in
the field of vibration measurement. Consider a mass-spring system
suspended from one end of a rigid enclosure, the other end of which is
attached to some vibrating body (Fig. 9-13). Suppose the vibrating

body were undergoing a displacement X
= JD sin c^t and an acceleration
X = — w^D sin oot. Then the equation of motion of the suspended
mass becomes, since the enclosure and body have the same displacement,
mx + k(x — X)
where x is the absolute displacement of the mass from m its equilibrium
position. This equation can be rewritten in the form
X + p-x = p^D sin cat
Assuming now that the free vibrations are damped out, the forced vibra-
tion, from Eq. (9-38), becomes
X = D—sin ost
1 co^/p^
Now, u = x — Xhe the rela-

letting
tive displacement between the vibrat-
ing mass and the rigid enclosure.
1X = Ds\n ut
»>>»»»: *
"7777777777777-
u = D sin wt
Yp2
Fig. 9-13
1 — coVp^
The relative displacement u may be recorded by attaching a stylus to the

vibrating mass and allowing it to scribe a record on recording paper
attached to the enclosure.

Consider the special case where oj^ p. In this event the relative dis-
placement is approximately equal to the negative of the impressed
displacement; i.e.,
u ^ —D sin wt = —X
and the recording device indicates the motion of the vibrating body. It
should also be pointed out that under these conditions the vibrating mass
suffers only negligible absolute displacement; i.e., it remains essentially
motionless.
On the other hand, suppose p co. ^
Here the relative displacement is
approximately proportional to the negative acceleration of the vibrating
body; i.e.,
w^D sin wt X
p. V
Thus the recording device now presents a record of the acceleration of the
supporting body. The preceding condensed discussion serves as the
theoretical basis for many of the present-day vibration-measurement
devices and accelerometers.
:
"îc. 9-8] Forced lihralioiis ivUfi viscous damping 207
9-8. Forced Vibrations Mith Viscous Damping
After the preliniinary work done in the i)revious sections, we are ready
now to handle the complete problem of a forced vibration with \ iscous
(lamping. Let us consider the complete ecjuation (li-O)
Fo
X + 2nx + p2x = — sin w< (9-6)
?«
A particular solution to this differential eciuation can l)e found l)y letting
X = ^1/ cos o)t -\- N sin cot
Substituting the above expression into the differential efjuation,
m
2noiN + (/;- - <^-)M =
Solving for M and A'^ and substituting into the trial expression for x,
Fo/m . , >\ , o .1
= + 4nw f^P" ~ ~
'•" -'^"
'^
ip-^
- c.r-
""'^ "' '''' '^'^
The above equation can also be written (Prob. 9-1)
X = Xm sin {ut - (f>) (9-40)
is the forced response amplitude, and
« = tan-(-f) = tan-..,2n<^ (9^2)
is the phase angle.

The complete solution of the differential equation (9-6) is the sum of the
general solution (9-23) of the homogeneous equation and the particular
solution (9-40) of the nonhomogeneous ecjuation just derived. Thus
X = CiC^' -\- C-rC'^' -\- .c„, sin (oj/ - 0) (9-43)
where S\ and S2by Eq. (9-22). The first two terms on the right
are given
side of this equation are thehomogeneous solution and represent a free
vibration with damping, as discussed in Sec. 9-5. With damping in the
system the free vibration eventually dies out. For this reason the free-
vibration part of the complete solution is called the transient. For an
underdamped or vibratory system the transient has a frequency of q
[Eq. (9-28)]. The last term of the complete solution, i.e., the particular
:
solution of Eq. (9-40), persists as long as the exciting force is present.

This part of the vibration is therefore called a steady -state vibration. The
steady-state motion has the same frequency co as the exciting force.
c/c^=0
Fig. 9-14
The response amplitude x„, (Eq. 9-41) of the steady-state vibration can
also be written in the following form
(9-44)
Fo/k {(1 - coVp2)2 + [2{n/p)(c./p)Yy^^
Thus the amplification factor AF = Xm/{Fo/k) is a function of the fre-

quency ratio w/p and the damping ratio n/p = c/cc and does not depend
on the initial conditions. Equation (9-44), showing the amplitude factor,
is plotted in Fig. 9-14.
can be seen in Fig. 9-14 that for c = 0, the AF curve becomes identi-
It
cal with that given in Fig. 9-12 for an undamped forced vibration. As
the amount of damping increases, the damping ratio c/cc grows larger and
:
! Sec. 9-8] Forced vihralions wilh viscous damping 200
f
the amplitude factor decreases. Hence danii)inj!; reduces the response
ampHtude. Near the resonant fre(iuency the rethiction is particularly
pronounced, indicating that the severit-y of the resonance condition may
' be materially reduced by the introduction of some damnint:; The I 1-1
I
i .
j!
resonant frequency itself becomes snudler as the damping ratio c/cc
decreases. Differentiating Eq. (9-44) with respect to w and setting the
derivative to zero, we can find the resonant frequency:
}r = y/p'^ 2n'- = p. 1
- (9-45)
KO'
For small damping, as is common in the majority of problems, /('-
is much
smaller than p- and cor ~ p. For such cases, the maximum amplification
factor, from Eq. (9-44), is
Cc
(9-40)
2n 2c
It can also be seen from Fig. 9-14 that for very small or \-ery large
values of u/p, the respective amplification factor is nearly or 1. For
such values of co/p the influence of damping becomes insignificant.
Applying the above results to an electrical RLC circuit (Fig. 9-4), one
can write the transient and the steady-state responses to an a-c source,
E = Eo sin ot, by merely substituting Ecjs. (9-9) into Eqs. (9-43), (9-22),
(9-26), (9-41), and (9-42). Thus, for the underdamped case, we have
Transient
1/coC and coL are, respectively, the capacitive and the inductive reactances.
The resonant frequency is defined as the frequency at which the steady-
state current is largest. (Notice the difference of definitions of the
resonant frequency in mechanical and in electrical systems.) From
Eq. (9-52) we see that the frequency for maximum im, i.e., the resonant
frequency, is
^^'^^^
EC
at which frequency the capacitive and the inductive reactances are equal.
Thus the resonant current is E/R. It is limited only by the resistance of
the circuit.
Fq sin ut
9-9. Vibration Isolation
In the foregoing we have found the displacement of

the mass of a vibrating system under the influence of an
exciting force acting on the mass. It is also of great
importance to find the force transmitted to the founda-
tion. In a practical problem, the mass may be an elec-
tric motor. Any unbalanced material in the motor
gives rise to an alternating exciting force F = Fo sin ut,
acting on the motor. To reduce the vibration, the
Fig. 9-15
motor may be mounted on steel springs, w^hich, besides
supplying the spring constant k, introduce also a slight amount of damping.
Sometimes greater damping is supplied by rubber or cork paddings. The
system can be symbolized by Fig. 9-15, which is in no way different from
the system shown in Fig. 9-3. The force transmitted to the foundation is
the sum of the spring and the damping forces. According to Eq. (9-40),
the steady-state motion of the mass is
X = Xm sin (cot —
Hence x = x^co cos (ut
Sec. 9-9] Vibralion isolalion 211
Substituting for x,„ of tho above eciuution ])y Fa\. (9-44),
1 + (2 " "Y
(9-56)
i:-^-m\
The ratio Pm/Fo of the transmitted force to the applied force is called the
transmissibility and is a measure of the effectiveness of the isolation

mounting. A plot of Fa\. (9-56) is shown in Fig. 9-16. We notice that
in theabsence of damping, the transmissibility is the same as the ampli-
tude factor.
It can be seen that, although Fig. 9-16 is similar to Fig. 9-14 in many
major respects, there is the following important difference. For the
smallest possible transmissibility, damping is desirable only if oj/p is less
than y/2. For w/p larger than y/2, damping actually increases the
transmissibility. Since it is our desire to reduce the transmissibility to a
value much smaller than 1 (otherwise we might as well use a rigid support,
with transmissibility equal to 1), we want always to operate normally in
the range where u/p > \/2. Damping is therefore undesirable at nor-
mal operating speeds. Some damping, however, is provided for the pur-
pose of avoiding excessive vibration when the motor is near the resonance
speed, which might occur on starting or stopping, at which time damping
is then highly useful.
Example 9-3. Consider now a spring-and-dashpot-niountod motor,
rotating with a constant angular speed co and having an unbalanced weight
of w. The motor of weight W is constrained to oscillate in the vertical

direction only (Fig. 9-17). Pertinent data for this problem are as follows:
W = 200 lb c = 3.94 lb sec/in.

m; = 2 lb CO = 120 rpm
/e = 30 lb/in. r = 6 in.
Because of the unbalanced mass, the vertical exciting force may be

expressed as
F = — mrco^ sin uit
First let us consider the undamped motion of the system. The natural
Fig. 9-17
and forced frequencies are
30 X 12 X 32.2
7.61 radians/sec
200
120 X 27r
= 12.57 radians/sec
60
12.57
= 1.65
7.61
From Eq. (9-39), the amplitude factor is AF = 0.58, and

Fig. 9-12 or
from Eq. (9-56), the transmissibility is equal to AF. Thus the
Fig. 9-16 or
steady-state amplitude [Eq. (9-38)] becomes
and the fonê transmitted to ground [Eq. (9-56)] is
P.n =
2 X 6 X (12.57)'-^
X 0.58 = 2.84 lb
32.2 X 12
Sec. 9-10] TorsioiKil rihnilion 213
Now let us examine the effect of damping. I'lu^ ilaniping ratio is
c
^ _c_ ^ 3.94 X 3 2.2 X 12 ^ .
''
Cc 2Mp 2 X 200- X 7.61
Therefore, from Fig. 9-14 or Eq. (9-44). the amplitude factor r(Mluces to
AF = 0.42. while from Fig. 9-l() or Va\. (9-r)0), the transmissil)ility
increases to a value of 0.81. The resonant frefjuency, according to Eq.
(9-45), decreases to
ojr = 7.61 Vl - 2(0.5)2 = 5.38 radians/sec
and the amplitude factor at resonance, which may occur when starting or
stopping the motor, decreases to a value of .\F = 1.15.
9-10. Torsional Mbratioii
The proper mounting (Fig. 9-18) of most rotating machines (such as

electric motors or internal-combustion engines) recpiires also an investiga-
tion of their torsional vibration.
The torque acting on such machines
is, for almost all cases, nonuniform
with respect to time. We can con-
sider the torcjue as consisting of a
periodically alternating component
superimposed on an average value.
The ratio of the alternating tor(iuc to
the average torque is of course differ-
Fi(i. ".(-IS
ent for different machines. On
account of this alternating tonjue, the machine and its casing vibrate in a
rotational motion about an axis that is approximately along the axis of the
shaft. The differential equation describing this motion (called torsional
vibration) is
/</) + r<^ + /.(^ = 7'o sin ut (9-57)
where / = moment of inertia

<f)
= angular displacement
To sin Oil = alternating portion of torque
The (|uantities —c4> and —k4> are the damping torque and the spring
torque, respectively. Since Eq. (9-57) has the same form as Eci. (9-4),
all the conclusions (i.e., amplification factor, transmissibility, resonance,
etc.) derived for a rectilinear vibration apply also to torsional \ibi-ations,
after a proper substitution of ecjuivalent c(jefficients.
It might be added that although coil springs are shown in tlu* symbol-
ized Fig. 9-18, leaf springs or elastic paddings are also often used in
practice.
9-11. Vibrations with Nonperiodic Forces
There are many instances when the force applied to the system under
consideration is not sinusoidal or even periodic. Blast- and earthquake-
type forces are well-known examples. Since a general forcing function
can be expressed in terms of a trigonometric series, according to the dis-
cussion of Sec. 9-2, the response due a general forcing function may be
obtained by an extension of the previous methods together with use of the
principle of superposition. However, it is sometimes more convenient to
employ an alternative approach.
Consider a single-degree-of-freedom undamped system acted upon by
some general force. Referring to Eq. (9-6), the equation of motion for
this system is
f^ + plT = -F(r) <t<h

where F{t) = disturbing force
T = a "dummy variable" representing time
^1 = duration of disturbing force
Multiplying this expression by sin pr and integrating between the limits of
T = and T ^ t,
-7-^ sm VT dr -\- v'x sm vt dr = -^— pr dr

sni ^
/
Jo dr^ "^
jo
I I
jo m
Now, integrating the left side by parts, f and after substituting in the
proper limits,
dx
sm
. ,
pt — ,. ©.To
f + f
px cos pi = nF(r) sm .
pr dr
,
-TT
dt Jo
/
m
Performing a similar series of operations on the equation of motion, except
thatnow we start with the multiplying factor cos pr,
dx F(t)
-rj cos pt — Xo + px sin pt = f'
/ —'^— cos pr dr
dt Jo m
Solving these resulting expressions for x by multiplying each equation
by and
cos pt sin pt, respectively, subtracting and simplifying, the response
is given by
Xn 1 f'
X ^ Xo cos pt -\ sin pt H / F(t) sin p(t — r) dr (9-58)
p mp Jo
t Note that
I
Jo dr^
pT
-r-^ sin ^ dr = -7- sin
X^dr
vt \
jO
—piJO -r cos pr dr
dr
p'^x sin pr dr = —[px COS pt]q + P I cos pr dr

I -f
stv. 9-12] More complicated vihraling syslems 215
For the usual case of blast loads, the system initially starts from rest, i.e.,
at t = 0, Xo = io = 0, and
^' = ~ j/i^) -"î"

P(^ - r) dr (9-59)
This result
is usually referred to as the convolution integral and repre-
sents the superposition of the free and forced vibration components of the
response of the system initially at rest and acted upon by some general
forcing function. easy to verify that the term (1/wp) sin p{t — t)
It is
in Eq. (9-59) represents the response of the system at time t to a unit
impulse acting at time r, for i > t (Prob. 9-21).
Example 9-4. In order to illustrate this method let us consider the
case of the forcing function having the form F(t) = Fo sin cot. Assuming
that the system starts from rest, the response as given by Eq. (9-59) is
X = —
mp
1 p
/
/o
Fo sm cor .sin p(t
^' — r) dr
A trigonometrical substitution recasts this into the form
[t
Fo
X = ;z 1 [cos (iOT -\- pr — pt) — COS (cor — pr -{- pt)] dr
which can then be integrated to give
Fo/k
X = (sin oit sin pt
1 — wVP" P /
j
This result, of course, conforms to that given in Eq. (9-396). which was
obtained by another method.
9-12. More Complicated Vibrating Systems;

Nonlinear Vibrations
We have discussed in this chapter the theory of the free and forced
vibrations of a single-degree-of-freedom system with linear spring and
damping characteristics. Although such a sy.stem is only a somewhat
idealized model of actual situations, the theory developed can describe to
us the phenomena of resonance, the calculation of natural frequencies, the
reason for the creeping or the o.scillatory motions of transients, and the
concept of vibration isolation. Without any additional theory, many
practical problems can and have been solved in this manner. Other
problems, however, cannot be so idealized without losing the essential
character of the motion. Time does not permit us to delve into such more
complicated problems in this book, but in order to broaden our view, we
;
should enumerate some of the problems that prove to be important in

engineering applications.
1. Problems Involving More Than One Degree of Freedom. The simplest
example for a two-degree-of-freedom system is shown in Fig. 9-19. It con-
sists of two masses and two springs. It can be shown that for a two-degree-
of-freedom undamped system there exist two natural frequencies corres-
ponding to two distinct types of free vibrations, called the natural modes of
vibrations. A discussion of this subject will be given in Chap. 13. In
general, it can also be shown that an n-degree-of-freedom system without
damping can vibrate freely in n natural modes, each with its own fre-
quency. Systems with more than one degree of freedom are also called
coupled systems. An n-degree-of-freedom system can
////////////////^
^^ idealized by n different masses coupled together and
to the foundation by n springs. The idealization of a
^
vibrating system as a discrete system of n springs and
masses may not be realistic in many instances. Some
vibrating bodies consist of an infinite number of parti-
cles,all coupled together elastically (or plastically)
such bodies (strings, bars, plates, turbine blades, for

example) therefore have an infinite number of degrees
of freedom and, correspondingly, an infinite number of
Fig 9-19 natural frequencies and modes of vibration. Fortu-

nately, only the lowest frequency and sometimes a few
higher frequencies (maybe even 10 or 20) are of practical importance.
The calculation of the natural frequencies and mode shapes of vibrating
bodies, and of the motion of such bodies acted upon by exciting forces,
constitutes a major effort in the more advanced treatment of mechanical
vibrations.
2. Problems with Variable or Nonlinear Characteristics. A vibrating
system may have a variable spring stiffness or a variable damping
coefficient. If k and c in the differential equation (9-6) are functions of
time only, the differential equation is still linear. A special class of prob-
lems involves a spring stiffness that varies sinusoidally with time. The
differential equation for this type of motion is expressed by
mx + (a + b sin cct).v = (9-60)
The above equation is equivalent to Eq. (9-10) if, in Eq. (9-10), we let
c = and k = a -\- b sin oot. Equation (9-58) is Mathieu's equation, the

solution of which has been extensively investigated.
A still more complicated problem is the case where the coefficients c and
k are functions of x or x. For such problems, the differential equation
becomes nonlinear. Two typical examples of nonlinear spring stiffness are
shown in Figs. 9-20 and 9-21.
Sec. 9-12] More coniplicaled vibruling sysleins 217
Figure 0-20a shows a flat spring supporting a mass m When deflected,
.
the spring contacts a curved guide which decreases the effective length of
the spring. Thus the stiffness of the spring increases with deflection,
i'he variation of the spring force with the deflection is shown in Fig.
Force
Fi(i. 0-21
9-206, which clearly indicates that the spring force is not proportional to
the deflection. Figure 9-2 la shows a mass m supported by one set of
springs for small deflections of m. When the deflection of m exceeds a
certain value, other sets of springs come into action, thereby increasing
the total spring stiffness. The variation of the spring force with deflec-
tion is shown in Fig. 9-216. Both cases shown here are examples of hard
springs, i.e., springs whose stiffness increases with the deflection. A
Belleville spring (Fig. 9-22), on the other hand, is an example of a soft
spring whose stiffness decreases with deflection. This device is used as a
spring washer in electronic equipment.
It hasbeen previously mentioned that at high velocities the resistance
of an object immersed in a fluid is roughly proportional to the square of
the velocity instead of to the velocity. On the other hand, the resistance
due to dry friction (Coulomb friction)
is independent of the velocity (Sec.
2-5). In either case we have a non-

linear damping.
From the theory of differential
equations we know that the principle
of superposition holds for the solu-
tions of linear differential equations
(with constant or variable coeffi-
Fig. 9-22
but does not hold for the solu-
cients)
tions of nonlinear equations. This
principle states that if Xi is a solution of mx -{- ex -\- kx = Fi{t) and :r2 is
a solution of mx + ex -\- kx = F^^t), then Xi -\- X2 is a solution of
mx -\- ex -\- kx = Fi(t) + /^2(0- The validity of the principle of super-
position rests entirely on the fact that
kxi + kx2 = k{xi + X2) exi -\- ex2 = c(xi -\- ±2)
which is true only if k and c are independent of x or its derivatives.
On
account of the principle of superposition, the complete motion of a
linear system under excitation is a combination of a free and a forced
vibration. This is not true for nonlinear systems. Furthermore, the
natural frequency (of free vibrations) of a linear system is independent
of the amplitude of vibration. With nonlinear systems, however, the
frequency of free vibrations changes with amplitude. We have seen this
phenomenon occurring in the case of a simple pendulum with large
amplitudes (Sec. 8-8). When a nonlinear system is under the action of a
sinusoidal exciting force F = Fo sin oot, the system may exhibit the follow-
ing typically nonlinear characteristics.
1. The amplitude of the response may suddenly jump up or down as
the exciting frequency 00 increases or decreases. The variation of the
response amplitude with co may appear as Fig. 9-23a for a hard spring or
Fig. 9-236 for a soft spring. Both curves are drawn for a constant Fo.
With increasing co, both curves follow the path 1-2-3-4. With decreas-
ing both curves follow the path 4-5-6-1. Thus, in the region 2-3-5-6,
CO,
the amplitude may have two different values for the same co (i.e., ampli-
Problems 219
tude not a single-valued function of uj), depending on the past history

is
of Figure 9-23a and b should be compared with Fig. 9-12 to show the
o).
contrast between nonlinear and linear S3\stems.

2. The may have freciuencies lower than the exciting
forced vibration
frequency This phenomenon is known as subharmonic resonance.
oj.
Although the detail of nonlinear vibrations is beyond the scope of this

book, engineers should be aware of the possibility of having essentially
Amplitude
Fig. 9-23
nonlinear characteristics in their problems and of consequent departures

from the linear solutions given in this chapter.
PROBLEMS
9-1. Show that the sum a„ sin iiwt + h„ co.s m^t can also be written A„ sin (nut +
0„). Find A„ and </>„ in terms of a„ and b,,-
9-2. A ma,ss is .supported by two springs of spring constants Ai and h> as shown.
If the spring.s are connected (a) in .series, (/;) in parallel, and (r) in parallel as shown,
determine the resultant spring constant and the freciuency of vibration.
2a
Fig. P 9-3
9-3. Two cylindrical rollers with distance 2a between the two centers roll
widi the
anRular speed w in opposite directions as .shown. .\ uniform bar of length / and weight
\y rests on top of the two rollers. The coefhcient of kinetic friction between tlie bar
and the roller is ma- If the bar is pulled slightly away from its middle position, it will
oscillate back and forth when relea.sed. Find the circular fretiuency of the resulting
oscillation.
9-4. AU tube is partially filled with water. The total length of the water column
is I. the water in one arm is depressed by a distance x, the water in the other arm
If
will rise by the same distance x. The system is then released. Write the equations
of motion of the water column and show that the motion is periodic. Find the period
of the motion.
Fig. P 9-4 Fig. P 9-5
9-5. Amass m slides on a curved surface having the shape of a parabola

particle of
y = Find the equation of motion of the particle and the frequency of small
ax"^.
oscillations about the lowest point 0.

9-6. A solid right circular cylinder of weight W
and cross-section area A is sus-
pended from a spring of spring constant k and hangs partially submerged in fluid of
weight density y. Find the frequency of small oscillations in the vertical direction.
/////////
/////////////////y///y/^^y/y///y///
'
\
^-^
Vo.Po ô-Po
^V/////////////////^////A//////y//A
Fig. P 9-6 Fig. P 9-7
9-7. A cylinder contains gas under pressure po and a piston .4 of weight The W .
enclosed gas obeys Boyle's law; i.e., pF = constant. The initial volume of each half
of the cylinder is Fo; the piston moves freely within the cylinder without allowing any
leakage between compartments. Derive the equation of motion of the piston, assum-
ing it was given some initial displacement. What are the frequency and period of
small oscillations about the position of equilibrium?
9-8. A mass-spring-dashpot system is started in motion with the initial conditions
at i = 0, a; = Xo, =
the system is overdamped, find the resulting dis-
x •'Co- (a) If
placement. (6) underdampcd, find the resulting displacement.

If the system is
9-9. A drop-hammer machine weighing 20 tons is supported on the floor by means

of very stiff springs which also supply some damping. It is estimated that the spring
constant is /c = 10 tons/in. and the damping coefficient is c = 8 ton sec /ft. A 1-ton
Problems 221
liummer falls on the niacliim- without rolioiiiul, aiul tlio two dfscoiul together after
the impact with an initial velocity of 2 fps. Fiiul
(a) The maximum displacement of the machine due to the impact.
(6) The displacement of the machine at the end of three complete cycles after the
maximum displacement.
(f) The minimum damping to prevent tlie machine from oscillatinK after impact.
9-10. Sliow that for the case of overdampeil free vibrations, a hody never pa.sses
tiiroush the equilibrium position if (a) it is released with initial velocity but no initial
displacement, or {h) it is released with initial displacement but no initial velocity.
9-11. Demonstrate that for the ca.se of overdamped free vibrations with general
starting conditions, once the mass passes through the equilibrium position it can never
pass through the equililirium position again, but can only approach it as < w. ^
9-12. A body weighing 20 lb is suspended by a spring and a dashpot that produces a
resistance of 0.02 lb at a speed of 1 in. /sec. In what ratio is the amplitude of vil)ra-
tion reduced after 10 cycles, if tlie spring constant is k = 5 lb/in.?
9-13. A damped vibrating system consists of a weight of 10 lb and a spring with

A; = 5 lb/in. If, at the end of 4 cycles, the amplitude is reduced to one-half of its
initial value, find
(a) The damping coefficient c.
(6) The
frequency of oscillation.
(c) The
resonant frequency of the system.
(d) The
amplitude of the force at resonant freciuency necessary to cause the mass
to vibrate at 1 in. amplitude.
9-14. A simple undamped vibrating .system is initially at rest. If a constant force
F is applied to the ma.ss for a time interval t', find the motion of the mass after F is
removed.
9-15. A motor has an unbalanced weight of }4 lb at a distance of C) in. from the axis
of rotation. The rpm of the motor is 1,200. The motor weighs 500 lb and is mounted
at the middle of a platform. It is estinuited that a load of 10,000 lb acting at the mid-
dle of the platform will cau.se a deflection of 1 in. under the load. Assume a damping
ratio c/cc of 0.1.
(a) Find the amplitude of the motion of the motor.
(6) Find the amplitude of the force transmitted to the ground.
(c) How much should the stiffness of the platform be reduced so that the di.splace-
nient of the motor is limited to one-tenth of the value given in part a?
Fi(i. P 9-15
9-16. A capacitor is charged by a constant-voltage .source K in a circuit which also

includes a resistance R and an inductutue L. At time / = 0, the switch is closed
<S'
^C
R
I AAV-
Fic. P '.)-IG
and the charging begins. Write the differential equation for the charge, and find the
expression of the charge as a function of time.
9-17. The inductance of a coil is measured by adjusting a variable capacitor in
series with it The resistance of
to produce resonance at a frequency of 0.5 Mc/sec.
the circuit is 4 ohms, and the capacitance is 0.001 /if (microfarad). The amplitude
of the current at resonance is found by an ammeter to be 2.00 amp. The impressed
voltage and capacitance are now kept constant, while the frequency of the impressed
voltage is changed to 2 Mc/sec. Assuming that the resistance and the inductance do
not change with frequency, find the amplitude of the current.
9-18. Determine the complete solution for general starting conditions, for the case
of forced damped The
forcing function is of the form Fq sin ut.
vibrations.
9-19. A mass of weight W
suspended from a spring of spring constant k and is
is
acted upon by a vertical force Fo sin cot. Determine the range of values of w for which
the amplitude of steady-state vibrations exceeds twice the static deflection caused by
the force Fq.
9-20. A mass-spring vibrating system starts from rest and has applied to it forcing
functions of the form (a) F{t) = ior t = 0, Fit) = Foiort > and (fe) F(t) = F^t/ti
for < t < ti, F{t) = Fo for t > ti. Determine the responses of the system using
the method of Sec. 9-11.
9-21. Show that the response of a simple mass-spring system initially at rest and
acted upon by a unit impulse at time t is x(t) = (l/mp) sin p(t — t), iort > t. A unit
impulse is defined as the product of force F and increment of time dr such that its
resulting magnitude is unity. According to Newton's law of motion, F dr = m dv;
i.e., a unit impulse at t causes a unit change of momentum at t.
CHAPTER 10
Dynamics of Systems of Particles
10-1. jMotioii of the Mass Center

In Chap. 4 we discussed the statics of a system of particles. We shall
now undertake a similar approach to dev'elop the dynamics of such a
system. In this chapter we shall treat the (jcncrnl case for which the
particles in the given system may or may not be rigidlj^ connected. For
such a general case, several general relations can be derived. Certainly,
for the special case of a single particle, such relations are applicable and
have been discussed in Chap. 8. Other relations which apply only for
the special case of a rigidbody will be discussed in Chap. 12.
Using the same notation as in Sec. 4-1, we consider the ith particle
whose mass is rrii and whose location is represented by the position
vector r,. The force acting on the iih. particle consists of a total external
force F, (exerted by mediums from outside the system) and a total inter-
nal force fi (due to interaction between particles within the system).
Applying Newton's second law f(jr the iih. particle,
F, + f. = w.r. = ^ (m.r.) (10-1)
We can write one such ecjuation for each of the particles in tlu; system.
Adding up all such equations, we obtain
t'--l'-i\t i>'i,r,]
where all the summations are cairicd from i = 1 to i = n, if ii is the

total numlx'r of particles in For simplicity, we shall omit
the system.
writing the indices of the summations in the further development.
223
:
224 Dynamics of systems of particles
According to Newton's third law, the internal forces occur in equal and
opposite pairs, thus:
2f, =
Furthermore, by Eqs. (5-6) (written in summation rather than in integral
form)
S(miri) = (2?ni)rc
where re is the position vector of the mass center or the center of gravity
for a constant gravitational field. We thus have, for the usual case of
time-independent mass,
2F,- = (2mi)rc
For brevity, let us use the notation F = SF^- for the resultant external
forceand ?n = linii for the total mass of the system. We therefore can
write
F = mrc (10-2)
This is the important equation that describes the motion of the mass
center of a system of particles and can be expressed in the following form
The motion mass center of a system of particles is identical with that
of the
of a particle ofmass m (the total mass of the system) concentrated at the
mass center and acted upon by a force F (the total external force on the
system). This explains why particle dynamics of Chaps. 8 and 9 is of
great practical importance, for it describes not only the motion of a
particle but also the motion of the mass center of a system, irrespective
of whether the system is a rigid body or not. It is important to under-
stand, however, that this result will not in general enable us to specify
the motion of each particle of the system.
An artillery shell may explode in mid-air into thousands of fragments
whose individual paths are difficult to trace, and yet the mass center of
all the fragments will follow a parabolic path (if air resistance is neglected
and the acceleration of gravity can be considered as constant) before as
well as after the explosion, until one of the fragments hits another system.
10-2. Impulse and Momentum of a System of Particles

For a single particle, the equation of motion
F = (mr)
I
can be integrated with respect to time and becomes
[^"'Fdt ^ mi-. - mi-i (10-3)
[The subscripts 1 and 2 on the right side of Eq. (10-3) denote the values
of velocity at time /i and ^., for the particle.] We call I

'
F dt the
Sec. 10-2J Impulse and niomeiihini of a sy.slern of particles 225
impulse (or linear impulse) and mir the mmnentum (or linear momentum).
Hence, for a particle, impulse is e(ptal to theThis change of momentum.
theorem, called the theorem of impuljê and momentum, actually offers
no new information, .since E(|. (10-3) is no more and no less than Newton's
law in a somewhat different form, the tlilTerent form, of course, heiufj; a
first integral. Both impulse and momentum are vector ([uantities;
Eq. (10-3) thereby is a vector eîuation that can i)e expressed i)y three
scalar equations.
Con.sidering now a sy.stem of particles, we can likewise integrate the
equation of motion [E(\. (10-2)]:
'
F dt = mrc-2 - mbci (10-4)
/,:
where /
'
F dt is the impulse of the resultant external force acting on
the system and mtc is the momentum of the total mass concentrated at
the mass center. Since
2(??i,r,) = mrc
we have, upon differentiating everj- term with respect to time,
Zimii-i) = mrc (10-5)
where 2(m,r,) represents the total momentum

Hence, of the system.
the total momentum momentutn of a
of a system of particles is equal to the
particle having mass equal to the total mass of the sijstcm and moving with
the mass center of the system. Then we can express Ecj. (10-4) by the
following theorem: For a system of particles, the impulse due to the total
external force is equal to the change of the total tnomcntum of the entire
syste7n. Alternatively, we may expre.ss this theorem as follows: The
rate of change of the total momentum of a system of particles (or particle)
is equal to the resultant externallv applied force.
The scalar form of Ecj. (10-4J can be written in rectangular coordinates
as follows:
^
F^dt = m{xc2 — ici)
/:
Fydt ^ m{yc2 - yci)
_
/:
F, dt = m{zci — ici)
As a special case of Ecj. (10-4). if the total external force is zero along
any one direction, the linear momentum must be constant in that direc-
tion and the ma.ss center of the .system travels with constant velocity
along that same direction. When the resultant external force vanishes
in all three orthogonal directions, F = 0, we ol)tain the conservation
i.e.,
theorem of momentum, which states that with the absence of external forces,
the total momentum is conserved. In this event the mass center of the
system travels along a straight line with constant speed, and if the mass
center is initially remain so. The recoil of guns firing
at rest, it will
horizontally, for example, such that the mass center of the system
is
consisting of the gun and the projectile remains at rest immediately after
the firing. This means that the velocities of the gun and the projectile
must be of opposite direction and of magnitude inversely proportional
to their masses.
Impulsive forces are defined as those forces approaching infinite mag-
nitude and acting on the system (or particle) for an infinitesimal time,
such that their product approaches a definite limit. Thus an impulsive
force Fjr can be defined as
Fz= lim r+^^Fdf
This definition of impulsive forces enables us to operate with them as

if they behaved similarly to our more usual concept of force. In particu-
lar, we note [Eq. (10-4)] that a sudden or instantaneous change of veloc-
ity, and consequently change of momentum, can only be caused by an
impulsive force. Impulsive forces are of considerable interest in colli-

sions and explosions. When the impulsive forces are internal, there is
no sudden change of velocity of the mass center of the system. Thus a
sudden change of momentum of a system is equal to the total external
impulse.
Example 10-1. A gun of mass m fires a shell of mass mo with a muzzle
speed of Vo at an angle of a with the horizontal. The gun is free to move
horizontally. Assuming that there is no recoil mechanism and that it
takes the shell a small time interval At to leave the barrel, find the recoil
velocity of the gun and the reaction between the gun and the ground
surface.
Solution. Assigning the instant of time just before firing as ti and the
instant when the shell leaves the barrel as t^, we can form Eq. (10-4).
The various forces that must be considered are the weight of the gun
and the shell and the vertical reaction R between the gun and the ground
surface averaged over the time interval At. Then, letting Vc be the
recoil velocity,
(R — mg — mog) Atj= —mvci -f moVo cos a i + moVo sin a j
Solving,
mov sin a
R„ = mg ,
-\- m.og +
,
At
moVo cos a
Vc =
Sec. 10-3] Angular niunienhini of a syslem of particles 227
10-3. Angular ^Momentum of a System of Parli<les

The angular momentum. Lo of a particle al)()ut an arbitrary point O is
; defined as
If Lo = p X wr (lO-Oa)
where p = position vector of particle relative to

m = mass of particle
r = position vector of particle relative to some fixed point and r
isthen absolute velocity of particle (Fig. 10-1)
Sincemr is the momentum of the particle, the angular momentum is also
known as the moment of momentum (Sec.
i
3-6).
According to the of Eq.
definition
,
(10-Oa), the origin may
be either fixed
I
or moving. However, to avoid future
confusion, we shall restrict Eq. (10-6a) to
the case where the origin is fixed (in the
Newtonian frame of reference). Then

Eq. (10-6a) may be written in the form
P (fixed point)
Lo = r X mr (10-66)
where r is now the position vector relative

Ki(.. 10-1
to the fixed origin 0.
Further, we define the relative angular nionientiun Lo of a particle
about a moving origin 0' as
Lo' = p X /Np (10-6c)
where p = position vector of particle relative to moving origin ()'
Q = velocity of particle relative to moving origin

Thus, in this convention, unprimed ([uantities refer to tliose taken about a
fixed origin whereas primed quantities are taken with respect to a moving
origin unless otherwise defined.
Let us consider at first the case where the origin is lixed in the New-
tonian frame. From Newton's law,
F = ^ (mr)
we obtain, after forming the cross product with r,
r X F = r X jA^^)
Considering the vector identity [{Va{. 6-7)1,
d d
(r X mr) = r X mr + r X -
.. (mi)
dt
The first term on the right-hand side is obviously zero. Hence
r X F = g (r X mr) = _
But r X F is the moment Mo. We thus have for a particle
Mo = ho (10-7)
Equation (10-7) states that the moment of the forces acting on a particle
about a fixed point is equal to the time rate of change of the angular momen-
tum of the particle about 0.
The same statement expressed in scalar form says that the moment
of the external forces acting on the particle and taken about some line
is equal to the time rate of change of the angular momentum of the
particle about the same line, the line in question passing through some
fixed point. The reader is reminded of the equation of motion of a par-
ticle along the tangential direction in cylindrical coordinates [Eqs. (8-4)],
F^r = in -r {r'^4>)
which is the component of Eq. (10-7) along the axial direction.

The above statements about the angular momentum of a particle can
again be used to derive a relation for a system of particles. As in Sec.
10-1, let us consider the tth particle:
Moi = r, X F, + r,: X f .• = |, Ui
at
Summing up all such equations for all the particles in the given system,
y Mo, = y (r, X F,) + y (r, X f,) =

^ y Lo.
Now 2(ri X fi) = (Sec. 4-1). Let us for brevity use the notation
Mo = S(ri X Fi) for the total external moment about and
Lo = 2Loi = S(r,: X ?n.r,:)
for the total angular moment about 0. We thus have for a system of
particles
Mo = Lo (10-8)
This important theorem states: For a system of particles, the total moment
of the external forces about a fixed origin is equal to the time rate of change
momentum about the same origin. As a special case,
of the total angular
we have the conservation theorem of angular momentum: with the absence
of external moments about a fixed origin 0, the total angular momentum
about of a system is conserved. Notice that the above theorems hold
•. 10-3] Angular momenliini of a sysleni of purlicles 229
lor any fixed origin O about which the moment and the angular momentum
are computed. The moment and the anguhir momentum of course
(hange in vakie when we shift the origin.
For the important special case of a particle, or system of particles,
acted upon by external forces directed toward a fixed origin -the central-
force problem —
the moment about the fixed origin vanishes and angular
momentum is conserved. Consecpiently, the path of the particle or
system is confined to one plane (Sec. 8-4).
It can be shown that the previous theo-
rem does not apply in general to a moving
origin. However, under special circum-
stances and, in particular, if the moving
origin is the mass center of the system, the
previous theorem again holds. To prove
this let r' and r, be the position vectors of
an arbitrary moving origin 0' and the /th
particle, respectively, both referred to a
fixed, or Newtonian, frame of reference
(Fig. 10-2). Let o, be the position vector
Fk:. 10-2
from the origin 0' to the ith. particle; p, is
therefore the position vector relative to 0'. Taking moments about 0'
of the forces acting on the ith particle and summing for all particles,
Mo = ^Qi X (F. + f.) (10-9)
Referring to Fig. 10-2, we note that r, = r' -|-

9,, r, = r' -f- p,, and from
Efj. (10-1), F, -|- f,- = {d/dt){miri). Substituting these relations, Eq.
(10-9) becomes
Mo' =-
y P. X |[m.(r'4-p.)]
Expanding this,
i\Io' = -r' X f > m,o\ + \ p, X ^ (w.6.) (10-9a)
Recognizing that
y p, X /N,p, = > p, X m,p, -h y Oi y,

Yf
(îQi) = 7 V- X If
(niiQi)
Eq. (10-9aj now becomes
Mo' = -r' X ly m,o,\ -I-

'
y p, X »?.e. (10-9/;)
The term on the right side of Eq. (10-9/>) can be rewritten as

first
— r' X
moc, where qc is the position vector from the point 0' to the mass
center of the system; the second term represents the time rate of change
of the relative angular momentum with respect to the arbitrary point 0'
and is denoted as Lo' [Eq. (10-6c)]. Hence we obtain, finally,
Mo' = to' - r X mgc (10-10)
Thus Eq. (10-8) is modified for a moving origin by the introduction

of an additional term. There are three situations when Eq. (10-10)
may be recast into the form of Eq. (10-8), i.e., when the additional term
— r' X niQc vanishes. First, if the point 0' is a fixed point or has a
constant velocity, the acceleration r' = 0. Second, if the acceleration r'
is such that its line of action passes through the center of mass, r' and qc
are parallel and the additional term vanishes. Finally, the third case
occurs when the arbitrary origin 0' coincides with the moving mass
center C, for then pc = 0. It is this last case that enables us to write
the important result
Mc = Lc (10-11)
where Mc = momentof all external forces about moving mass center

Lc = time rate of change of relative angular momentum about
moving mass center due to motion relative to mass center
Equations (10-8) and (10-11) can be summarized then as follows: For
a system of particles, the equation of moment of momentum may be written
with respect any arbitrary fixed point or with respect to the moving mass
to
center of the system. For these two points, the total moment of all
external forces is equal to the time rate of change of the total angular
momentum and relative angular momentum, respectively.
It is interesting to note that when the moving origin 0' is the mass
center of the system,
Spi X miQi - Spi X mii-i (10-12)
i.e., the angular momentum about th£ moving mass center due to the
relative velocities is the same as that due to the absolute velocities. We
leave the proof as an exercise (Prob. 10-7).
another interesting result that appears as a by-product of the
Still
previous discussion is the following. In Fig. 10-2, suppose we let 0'

coincide with the moving mass center C. Then the total angular momen-
tum about the fixed origin is
Lo = 2ri X mii-i = 2rc X m,fi -1- Sp,- X m^-ri
But the last term can be rewritten according to Eci. (10-12). Further,
2rc X wijTi = re X ^niii-i = re X mi-c
Plence we have the result
Lo = I'c X mi-c + 2pi X m^pi

i,
Sec. 10-3] Angular rnomenlum of a syslem of particles 2:U
In terms of the previous nomenclature, this l)ecomes
Lo = rc X nivc -\- Lc (10-13)
Thus, the angular 7uo)ncntHm of a ,si/stcm of parli'rlrs about an arbitrary

fixed point is the sum of the angular momentum of the mass center about 0,
imagining the entire mass of the system concentrated at the mass center and
moving with it, and the angular momentum of the system about the mass
center due to the relative motion with respect to the mass center.
The vector equation (10-8) or (10-11) can also be written in terms of
its three components, in general form, as ff)llo\vs:
Along r: y {yF\ - zF,) = \ m{yz - zy)
Along y :
y {zF. - xF.) = y m(zx — xz)
Along z: Ul's) = \ m(xy - yx)
where the coordinates x, y, z are measured with respect to either a fixed
origin or the mass center of the system, depending on the equation to be

expressed. For example, in Eq. (10-8) .t, y, and z are measured from the
fixed origin 0; in Eq. (10-11), x, y, and z are measured from the moving
mass center C.
In a fashion similar to that employed in discussing impulsive forces,
we can define an impulsive moment as a large moment acting on the
system for a short period of time.
Thus
M, = lim
li / M dt
-Vgain we note by an examination of
Eq. (10-8) or (10-11) that a sudden
change of angular momentum must
be accompanied by an impulsive
moment. Therefore the sudden or
instantaneous change of angular
Fi<i. 10-:?
momentum about a fixed point or the
moving ma.ss center is equal to the total external impulsive moments about
those same points.
Example 10-2. A small particle of mass m is attached to the end of an
inextensil)le string and travels a horizontal circular path of radius ro
with angular velocity wo (Fig. 10-8). If the radius of travel is reduced
to .some new value ri by pulling on the string, find the new angular
velocity wi.
Solution. Writing Eq. (10-7) with respect to the point at 0,
= [r X w((o X r)] (mr-co)k

J
Integrating, between the limits of ô and ti,
P' d
(mr-co) dt = mr^co
jto dt
= m(r rcoi — ro-coo)
Example 10-3. A weightless rigid rod of length 21 is constrained to

rotate about a vertical axis in such a fashion that it makes a constant
angle with the vertical (Fig. 10-4).
Two masses m are connected to the
rod at its extremities. Find the
angular momentum of the system
and the constraining moment.
Solution. Suppose we form a co-
ordinate system fixed in the system,
rotating with the system, and pass-
ing through the fixed point in
space. Let us choose the y axis to
be directed along the axis of rota-
tion of the bar, the x axis to be in
the plane of the y axis and the bar,
Fig. 10-4
and finally the z axis to be perpen-
dicular to the xy plane. Then the angular velocity is
and the velocity of each mass point is
ri-= (0 X r, = + wl sin 6 k
where r,-= + (Z sin i + cos / j). It follows, then, that the total
angular momentum becomes
ho = 2ri X mr-i = 2mri X r,- = 2??7paj sin d {
— cos 6 i -\- sin 6 j)
We immediately note that the angular-momentum vector ho is per-

pendicular to the bar and travels Math the rotating bar; moreover,
ho does not have the same direction as o).
In order to determine the constraining moment necessary to maintain
the type of motion described, we make use of Eq. (10-8). Accordingly,
we must form the time rate of change of the angular-momentum vector
Sec. 10-1 J
Syslenis wilh variable mass; rockel and jrl ftrofjulsioii 233
ho. Since Lo is expres^^sed in terms of a rotating coordinate system of

angular velocity to, the derivative, from Eq. (7-19), is given as
Lo = (-7t) +oiXLo = '2mi- sin d ( —w cos d i -\- u sin d j
-\- <j}- cos 6 k)
Therefore the constraining moment is
Mo = 2ml- sin 6 (
— 6) cos i + cj sin j -f co- cos d k)
10-1. Systems with Variable Mass;

Rocket and Jet Propulsion
I'p to this point, we have discussed the motion of particles and systems
of particles assuming that the mass of the system remained constant.
A' m K„
masses, we write, according to Eq. (10-4),
[(m + Ki At){y + Av) + Ko At(y + Av+ U^)]

- [(m + Ko At)y + Ki At{y + U,-)] = F A^ (10-14)
The bracketed term on the left represents the total linear momentum
first
at the time t -\- At; the second bracketed term on the left represents the
total linear momentum at the time t; their difference represents the
change of the total momentum during the time interval At. We have
assumed above formulation that after entering or before leaving,
in the
the additional mass in the system takes on the velocity of the system.
(This assumption is made merely for convenience. It is not a necessary
assumption.)
Expanding Eq. (10-14) and neglecting higher-order terms, we obtain,
after passing to the limit as A^ -^ 0, the following expression for the
motion of the system with variable mass:
7n ^ + KoVo -
at
K^Vi = F (10-15)
For those cases in which the velocity vectors in Eq. (10-15) are approxi-
mately parallel (i.e., rectilinear motion), the vector equation can be
reduced to a scalar equation in the direction of the motion:
m^ -F KoUo - K,U. = F (10-16)

at
In order to illustrate the foregoing ideas, let us consider first a rocket

engine. In this case Ki = 0. Letting /x be the rate of mass ejection of
the discharging gas particles and Ve their exit speed relative to the rocket,
then
Ko = M Uo = —Ve
and the equation of motion of the rocket [Eq. (10-16)] becomes
m~=F + ixve (10-17)
where F is the total-external-force component acting on the rocket in

its rectilinear motion.
the direction of
This external force consists of the aerodynamic forces, the gravity
forces, and the pressure forces. (The separation of the pressure forces
from the aerodynamic forces is quite arbitrary.) Letting A be the
discharge area, pe the discharge pressure measured above the ambient
atmospheric pressure, and p,. the mass density of the discharged gas
particles, then y. = peAve, and Eq. (10-17) now becomes
mj^= F' + {pA + PeAve') (10-18)

Sec. 10-1] Systems with variable mass: rorkel and jet propulsion 23.')
where F' is the resultant extenuil-force coinponont. exchuliiifj; the pressure

force. The (luantity in tiio parcntiiescs on the riji;ht side ot" Imj. (10-18)
is commonly referred to as tiie thrust. t Tims, for a rocicet, the thrust is
given as
T - prA + p,.1jv2 (10-19)
If the rate of mass ejection m is constant durinj; the period of rocket

action, the mass m of the rocket at any instant is given by
m= mo — fit = nU) — p,Avet
where nio is the initial gross weight of rocket including full load of fuel
and is measured from the instant of firing. Further, if we neglect the
t
aerodynamic forces, and thereby consider the gravitational force as the

only external force, then F' = —my during the vertical ascent of the
rocket. Equation (10-18) then gives
dv ,
T
dt= -^^m
or. substituting the expression for m and rearranging,
dv ^(-g-\- ^^ dt (10-20)
Assuming that n, T, and g are all con.stant during the period of rocket
operation, Eq. (10-20) can be easily integrated to yield the speed of the
rocket at any instant t after firing:
y = - log - gt
n mo — ,
fit
In particular, the maximum speed is achieved at rocket burnout, denoted

by the subscript h, and is expressed as
v, = I'log^'-^/. (10-20a)
Thus, in order to increase vt, we have to increase both ratios+ T/n and
mo/mfe, the latter being the ratio of the rocket weight with fuel to its
weight without fuel (i.e., the structure weight plus payload). With the
present chemical fuels and rocket-structure design, it has been found that
Vb does not quite reach the .satellite speed (Sec. 8-5) in a single-stage
rocket. Multistaging therefore becomes necessary for satellite launching
at present.
t The thru.st may be easily determined by placitiu the

rocket in a test stand and
performing a static test.
X The dimensional ratio T/fxg is eommonly
referred to as the specific impulse.
As a second example, consider the case of jet propulsion, say a jet plane
of mass m flying at a speed v. In this case air enters and leaves the jet
engine with Ko larger than Ki by a small percentage, the difference due to
the fuel added. If we neglect this small difference and again let A be the
discharge area of the jet nozzle, pe the discharge pressure, and Ve the dis-
charge jet speed relative to the engine, then
Ui = —V Uo = —Ve Ki = Ko = peAVe
Hence Eq. (10-16) yields
mj^ = F'^ [peA + peA{ve - v)Ve] (10-21)
where F' has the same significance as in Eq. (10-18). The quantity in
I
the brackets of Eq. (10-21) is commonly referred to as the thrust. Thus,
for jet propulsion, the thrust is given as
T = PeA -\- PeA{Ve - v)Ve (10-22)
10-5. Kinetic Energy of a System of Particles

The equation of work and kinetic energy for a system of particles can be
obtained by summing up all such equations [Eq. (8-39)] for each of the
individual particles in the system. Thus
2
j' (F, + f,) . dvi = 2 y^m^Vi' I'
(10-23)
The left side of this equation represents the total work, and the right side
represents the change of the total kinetic energy.
Consider the total work on the left side of Eq. (10-23), written with
respect to the arbitrary moving point 0' (Fig. 10-2).
= r (I ''O
•
^^'
+ /i' ll ^0 •
^^' + A' I ^^'^ + ^^^ • ^'^
I
Because 2fi = 0, the second term vanishes. The total work therefore
becomes
W = f'F' dr + f' y^
(F, + fd .
dgi
I
where F = SF^ represents the resultant external force acting on the sys-
tem. In particular, if the point 0' coincides with the mass center of the
system, the total work becomes
W = y^' F . drc + 1^' 2(F, + fi) .

dQi (10-24)
:
Sec. 10-5] Kinelic energy of a sysleiii of /Hirlicles 237
Equation (10-24) expresses the fact that the total work done by the
forces acting on a system of particles is etiual to work done by the external
forces due to the displacement of the mass ccMitcr, plus the work done by
all the forces (internal and external) due to the displacements relative to
the mass center and due to rotation about tiie mass center.
Consider now the total kinetic-energy term on the right side of Eq.
(10-23)
T = i:}2miv:- = 2i2m.(r' + p.) (r'
• -\-
p.)
= i.ym.r'^ -\- f' • 2w,p. + r>._,'".pr
Again, if the point 0'
is chosen coincident with the center of mass C, p, is
measured from the mass center, 2/«,p, and its derivative wm,p, are both
zero, and the second term therefore vanishes. Thus the total kinetic
energy can be written
T = ^hzm.vr- = ^-^mrc- + î.im.p.^ (10-25)
Hence the total kinetic energy of a system of particles is the sum of the kinetic
energy of the mass center {as if the total mass of the syste77i were concentrated
there) and the total kinetic energy of the particles due to the relative motion
with respect to the mass center.
Equation (10-28) can now be written
j^' F'drc-\- j'

2 (F' + fO • ^/p. = ^ 2>nrc2 1'
+^ ' 2'"'^''
li
^^^'^^^
This efiuation can be separated into two ecjuations. Taking the dot
product of E(i. (10-2) with drc and integrating (see Sec. 8-0),
r/rc = y2mfc'^ |' (10-27)

/,
We see that this is one part of Ec]. (10-2()). vSubtracting this equation
from Eq. (10-2G), we obtain the other part of Eq. (10-26):
ji 2 ^^' + ^'^ • ^^P' = Z ^^'^'^•' li

^^^"^^^
Ecjuation (10-27) describes the motion of the mass center of a system of

particles — this motion is influenced only by the external forces —and
states that the total work of all the external forces due to the displacement
of the mass center is etjual to the change of kinetic energy of the mass as if
the entire mass were concentrated there. On the other hand, Eq. (10-28)
describes the relative motion with respect to the mass center; the relative
motion is influenced both by the external and by the internal forces.
Therefore the total work of both internal and external forces acting on
the .system due to motions relative to the mass center is ecjual to the
change of kinetic energy of the .system due to motion relative to the mass
:
center. Internal forces generally contribute to the equation of work and

energy, whereas internal forces are always eliminated in momentum
principles when applied to the system as a whole.
10-6. The Virial Theorem

Considering an arbitrary, say the iih, particle of a system of particles,
we can form the scalar quantity
The derivative of this quantity with respect to time is
^- (miA- •
i-i) = rriiV^ + r^ • (F,- + f,) (10-29)
where we have used Newton's law F^ +

= miii. The above equation
£j
is of course true at all times. Let us now

take the time average of this
equation. By definition, the time average (denoted by a bar above) of
any quantity, say Q, over a time interval r is
Q = - \
Qdt (10-30)
Applying this definition to Eq. (10-29) and taking the time average
over a very long time (i.e., r ^^ oo), we find that the left side of Eq.
(10-29) averages to zero if the particle moves in a bounded region for
which r and r are finite at all times. Hence
TYliVi^ = -Xi' (Fi + fi)
Let us divide the above equation by 2 and sum up over all particles in the
system
HZ;^^ = -y^-Zvi' (F, + fO (10-31)
(Notice that, since the time average is a distributive operation, the

summation and the time averaging can be interchanged.) The quantity
on the right side of Eq; (10-31) was called by Clausius, a nineteenth-
century scientist, the virial of the forces. Equation (10-26) demonstrates
the virial theorem, which states: The average total kinetic energy of a system
of particles moving in a bounded region is equal to the virial of the forces
acting on the particles.
10-7. Elementary Kinetic Theory of Gases

One of the most important applications of the dynamics of a system of
particles is to be found in the kinetic theory of gases. In its elementary
form, the kinetic theory considers a gas to be composed of a large number
Sec. 10-7] Elementary kinelic theory of gaaes 239
of iniiuite. moving particles, called the molecules, separated most of the

time In' distances that are large compared with the size of the molecules
themselves. A molecule moves ahoît in straigiit Hues until it hits
another molecule or the walls of the container; upon collision it bounces
off into new directions, with a possible interchange of kinetic energies
between molecules, but with no loss of total energy of the system. \u

ideal, or perfect, gas is one in which the molecules are located so far ai)art
(compared with their own size) most of the time that they c.xert negligil)le
forces upon each other except when they come into contact during
collision. The frequent collision of molecules with the walls of the
container is responsible for the pressure on the walls. The temperature
of the gas is interpreted as the mean kinetic energy per molecule. Thus
heat energy of a gas is identified as the mechanical energy of the molecules.
In the following, w'e shall neglect the kinetic energy of rotation of a
molecule and of vibration within a molecule and consider only the kinetic
energy of translation of the molecules. '
We now seek to apply the virial theorem [Eq. (10-31)] to a perfect gas.
Since the interaction between molecules that are not in direct contact is
neglected in a perfect gas, the virial of the system becomes that due to the
external forces alone. To be sure, there exist also internal forces of action
and reaction at the instant two molecules, but such a pair of
of collision of
forces, besidesbeing e(iual and opposite, act at the same position r antl
therefore contribute nothing to the virial. Hence
2r. • f, =
and E([. (10-31) becomes
1
2^m,rr = - ' 2i>, • F, (10-32)
or >^2^^ = - '2^(-'-./''.x + !/J'\u + 2./'\-.) (10-33)
where Xi, yi, Zi = rectangular (H)mponents of ri

Fix, Fiy, Fi^ = rectangular components of F,
Let us consider, for simplicity, the gas enclosed in a cubic container of
sides a (Fig. 10-6). The average external forces F^, Fy, F, are due to the
pressure between the gas and the walls of the container. Take, for
example, F^, which acts at the two walls A and B. Since Xi = on wall .1
and Xi = a on wall B,
ZxJ'\^ = a X average total force on wall B

^ —a(pa'-) = —pV
where V = total volume of container
p = pressure exerted
Applying this procedure to the other walls, we can readily see that
^XiFi^ = 2(/./'\v = ^ZiFiz = -pV

I
Hence Eq. (10-33) is thereby reduced to
}''2'2'miVi^ V2VV (10-34)
It should be remarked that Eq. (10-34), although derived for a cubic I
container, can be shown to hold also for gases enclosed in a container of {
any size or shape. j
Assuming that the total kinetic energy of the molecules is proportional I
to the absolute temperature.
y2^miVi'' = %RT (10-35)
where T is the absolute temperature and R is called the gas constant.

Equation (10-24) now becomes
pF = RT (10-36)
This is the well-known equation of state

of a perfect gas. It has been experi-
mentally verified for gases like hydrogen,
oxygen, and nitrogen at room tempera-
tures if the pressure is not very high. A
special case of Eq. (10-36) is when the
temperature is held constant; then the
volume of a gas is inversely proportional
to its pressure. This is Boyle's law.
Fig. 10-6
j^^^ ^g assume that in a gas in equili-
brium, all the molecules have the same mass m and the same y^ One
consequence of this assumption is that the molecules have the same aver-
age kinetic energy j^^my^, an assumption referred to as the theorem of
equipartition of energy. Thus
where A^ is the total number of molecules in volume V. Substituting this

into Eq. (10-24),
_
mNiT- = 3pF (10-37)
We now assume that the average kinetic energy of a molecule is directly

related to the temperature so that
mv 2 = 3kT
Thus pV = NkT (10-38)
or alternatively, p = nkT (10-39)
where n = N/V is the number of molecules per unit volume. Comparing

Eqs. (10-38) and (10-36), we see that k = R/N. The quantity k is
therefore the gas constant for one molecule. It is a universal constant for
Sec. 10-8] Viscosity of a gas 241
many gases and is called the BoUzmann constant, possessing a numerical
value of
k = 1.381 X 10-'« cm dyne/°C
It can also be seen from Eq. (10-37) that
where p = mN/V is the density of the gas. We thus are able to find the
root-mcan-square (rms) velocity of the molecules:
y. = Vv-' = J'^-? = V3ft. (10-40)
where /?, = R/mN is the gas constant per unit mass.f We shallhave
occasion to refer to this ecjuation when we discuss the velocity of sound
in a gas.
10-8. Viscosity of a Gas

One of the most important phenomena in the mechanics of fluids is
viscosity. We shall take this opportunity to analyze this phenomenon
Fig. 10-7
for a gas on the basis of elementary kinetic theory of gases. The practical
aspect of the problem (for example, the influence of vi.scosity on fluid
motion) will be discussed in Volume 2.
A fluid will deform continuously when subjected to a shear stress (Sec.

5-3). Consider, for example, two parallel plates with the space between
filled with some fluid (Fig. 10-7). When we apply a steady force F on
the upper plate so that the upper plate moves with a constant speed U
t Once we introduce li^, the equation of .state (lO-IiO) can l>e written
^ = li.T
p
The 8ub.script s is often omitted in thermodynamics with tlic unthTstanding that the
gas constant in the c(|uation is that for a unit mas.s.
while the lower plate is held stationary, the speed of the fluid will vary
uniformly from at the lower plate to U at the upper plate. This is so

because the fluid in immediate contact with a solid boundary adheres to it.
Thus the fluid between the two plates is continuously being deformed an ;
element ahcd of the fluid at one instant will assume the position ah'c'd
some time later. Let A be the area of one plate and y be the distance
between the two plates. It is found experimentally that for most com-
FiG. 10-8
mon fluids the shear stress FyA is proportional to the velocity gradient
U/y; i.e.,
F U (10-41)
A "^
y
This is Newton's law of viscosity. The factor of proportionality n is
called the coefficient of viscosity, or simply viscosity.

The above example helps us to visualize the phenomenon of viscosity,
but the presence of the upper and lower solid boundaries is not necessary
for the existence of shear stress in a fluid. Consider the case of a fluid
moving with horizontal velocities of magnitude U along the x direction.
In general, U may vary in magnitude in a perpendicular direction y (Fig.
10-8). In other words, U = U{y). Since the velocity gradient at any
station of y is dU/dy, the shear stress acting there on the fluid, following
Newton's law of viscosity, is
dU
(10-42)
dy
The direction of r is always such as to tend to equalize the fluid velocities.

Thus the shear stress exerted b}'- the fluid above y on the fluid below y in
Sec. 10-8] Viscosi'ly of a gas 243
Fig. 10-8 is directed toward -\-x. By Newton's third law, the fhiid helow
// exerts an equal but opposite shear force, and conse(iuently shear stress,
to the fluid above ij.
The viscosity of liquids can perhaps be explain(>d by the cohesion of

their molecules. No
such explanation is possible with an idval gas, for
which the intermolecular forces are assumed to be negligible since the
distance between molecules is proved to be large compared with the size
of the molecules themselves. A satisfactory explanation of the viscosity
of gases can be made. howe\-er, on the basis of elementary concepts of
kinetic theory.
Considering Fig. 10-8 again, let .1 be the area of a horizontal surface
located at a vertical distance y above the origin. Since dU/dy is positive
lor the region under consideration, the average x momentum of the
molecules is larger at larger values of y. The molecules cro.ssing .4 from
helow to above, therefore, arrive with less x momentum than the proper
amount at the new position. This means that the gas above .1 suffers a
loss of momentum because of the arrival of the molecules from l)elow.
The net effect is equivalent to the action of a shear force toward — .r on
the gas above .4. Similarly, the molecules crossing .4 from above to
l)elow arrive with greater x momentum
than the proper amount at the
new po.sition. The momentum of the gas below A
resulting increase of
is ecjuivalent to the action on it of a shear force toward -\-x.
In order to arrive at some quantitative results, we have to introduce the

concept of a mean free path, denoted by L and defined as the average dis-
tance through which a molecule may travel before colliding with another
molecule. The number of molecules crossing .4 from above, a.ssuming
all molecules have the same molecular (rms) velocity i^ and exhibit an
equal tendency to travel in all six directions (i.e., ±.r, ±y, ±z), is nAvJQ>
per second. These molecules come on the average from a region which
is at a distance of L above A and at which the fluid velocity is greater than
that at .4 by the amount L{dU/dy). The molecules crossing .4 from

above therefore carry with them an excess x momentum of
-r-^ z- per unit of time (10-43)

b dy
This rate of change of x momentum is equivalent to the following pair of

forces: (1) a shear force along +x on the fluid below .4 and (2) a shear
force along —x on the fluid above .4; both shear forces, of cour.se, have
the magnitude given by Eq. (10-43). A similar reasoning shows that the
molecules cro.ssing from below carry with them a deficit x momentum
.4
of the (10-43); the resulting change of momentum is

same amount as Eq.
equivalent to the action of another pair of shear forces which is the same
as the previous pair. The total shear force f at A is therefore twice the
value of Eq. (10-43), or, since nm = p, 1|
F = eA^f
3 dy
(10-44)
'
The shear stress is therefore

F pVgL dU
(10-45)
A 3 dy
Comparing (10-44) with (10-42). we obtain the viscosity of a gas
pVsL
(10-46)
This formula was obtained by ^laxwell in 1860. Because of the various

assumptions and simplifications introduced in the derivation, the factor 3^^
in Eq. (10-46) needs some correction. Various investigators have arrived
at different factors, but Eq. (10-46) remains essentially correct. To give
an idea of the order of magnitude of the various quantities used in the
above discussion, it has been found that for oxygen at 15°C:
Viscosity, = 2 X 10~^ dyne sec/cm
jjl
Mean path at atmospheric pressure, L = 6.8

free X 10~^ cm
Elastic-sphere equivalent diameter of molecule = 3.6 X 10~* cm
Rms molecular velocity, Vs = 47,300 cm sec
Mass of a molecule, m = 5.32 X 10~-^
g
One of the early and spectacular achievements of the kinetic theorj' of
gases is the prediction that the viscosity of a gas increases with tempera-
ture but independent of pressure and density. This can be demon-
is
strated by considering the various terms in Eq. (10-46). If the density

of the gas is reduced to half of its original value, the distance between
molecules is doubled. A molecule therefore can travel on the average
twice as far before hitting another molecule. Hence L increases roughly
by a factor of 2 so that the product pL remains approximately constant.
The molecular velocity Vs is independent of density; it increases only with
temperature. The viscosity, therefore, also increases with temperature
and is a function of temperature only. The exception to this rule occurs
when the pressure is either vanishingly low (such as air at extremeh' high
altitudes) or very high (100 atm or more). For such extreme pressures
Eq. (10-46) is no longer valid.
10-9. Statistical Mechanics

The analyses in the three sections, Sees. 10-6 to 10-8. although based on
the same classical laws of dynamics, i.e., Newiion's laws, introduce an
entirely new concept, namely, that of dealing with averaged quantities.
Sec. lO-t)] Statistical mechanics 215
The reader will have noted that in Chaps. 8 and 9 we looked for and
()l)tained detailed information on the motion of a particle or of particles
liy solving the differential ecjuation describing the motion. The .><ame
approach, when applied to a system containing a large numl)er of par-
ticles, would require the solution of the .same large number of differential
cciuations. one for each particle. Furthermore, in order to solve these

differential equations, we would have to know all the internal and
also
external forces acting on the system and the initial conditions for each
one of the particles. Such a requirement obviously cannot be met.
1 ortunately, our main interest in systems containing an enormous num-
ber of particles usually lies only in finding certain broad features over long
periods of time. By necessity and by choice, we deliberately abandon
the attempt to describe the motion of each particle at every instant and
content ourselves with seeking the laws relating the averages and the
of the significant properties of the system.
probabilities The large
number of particles becomes an advantage in.stead of an obstacle in this
new approach. When we deal with averages and probabilities, the
method that we use is known as statistical. The branch of physics that
analyzes dynamic systems by means of statistical methods is referred to
as statistical mechanics.
The elementary kinetic theory discussed in the previous three .sections
applies statistical concepts in introducing averaged quantities, but the
application there is quite crude on at least two accounts. In the first
place, too many and too drastic assumptions are made in connection
with the structure of the particles and the forces of interaction between
them. Secondly, the elementary theory is not based on a rigorous
mathematical foundation. It is the purpose of statistical mechanics to
reduce the number of assumptions to as few as possible and to formulate
and apply the mathematical theory of probability in a precise manner.
For example, one of the fundamental problems in statistical mechanics is
the ergodic problem, i.e., the problem of determining the necessary and
sufficient conditions under which a time average of a single system can be
replaced by an ensemble average of a huge collection (or ensemble) of
replicas of the system. Another problem is the computation, based on
the so-called limit theorems, of the ultimate statistical properties of a
system possessing a large number of particles or degrees of freedom. It
can be appreciated that mathematical subtlety and precision are neces-
sities in this branch of science.
PROBLEMS
10-1. A .shell explodes at the hif^hest point of its path into two .splinters of equal
mass. It is observed that one .splinter falls down vertically. W hat is the path of the
other .splinter? Neglect air resistance in this prol)lem.
10-2. Neglecting friction and assuming massless pulleys and cord, find the acceler-
ation of the two weights Wi and W2 and the tension in the cord.
10-3. A 2-ton automobile traveling due east at
60 mph collides with a 5-ton truck traveling due
south at 40 mph. If the two vehicles lock together
after collision, find
(a) The velocity after collision.
(6) The velocity of the mass center before and
after collision.
(c) The kinetic energy of the two vehicles before
and after collision.
10-4. The impact of two bodies whose initial
velocities are along the same straight line is called
direct impact. the line of action of the impact
If
force is normal to the contact surface and passes

through the mass centers of both bodies, the impact
is For the direct, central impact
said to be central.
of two bodies, Newton defined a coefficient of restitu-
tion as the ratio of their relative speed after the
Fig. P 10-2
impact to the relative speed before the impact. If
vi and Vo (see figure) are the speeds of the two bodies before the impact and v^ and v^
are the speeds after the impact, then the coefficient of restitution is
Vi — V2
Thus, for purely elastic bodies, e = 1, and for purely inelastic bodies, e = 0. If two
bodies of masses Wi and nio travel with speeds Vi, v-i along the same straight line as
shown, with Vi > v-i:
(a) Find the speeds Vj and v^ of the two
bodies after impact, using e as the coeffi-
cient of restitution.
(6) Find the loss of total kinetic energy
of the two bodies due to the impact.
(c) Evaluate the solution of cases a and Fig. P 10-4
h for nil = m-i = to and e = and e = 1.
{d) Evaluate the solution of cases a and h for TO2 = <» and e = and e = 1.
10-5. A pile 25 ft long, weighing 40 lb /ft, is being driven into the ground by a ham-
mer of 3,000 lb weight. With the last blow, the hammer is dropped freely from a
height of 16 ft and the pile penetrates 3 in. deeper into the ground. The coefficient
of restitution between the hammer and the pile is 0.5. Calculate the carrying capacity
of the pile. (Assume that the resistance from the ground to the pile is a constant dur-
ing the 3-in. movement after the last blow.)
10-6. A acted upon by a force which is always directed toward a fixed
particle is
point. Show that the particle moves in a plane.

10-7. Verify Eq. (10-12); i.e., show that the angular momentum about a moving
mass center due to the relative velocities is the same as the angular momentum about
the mass center due to the absolute velocities.
10-8. Find the frequency of small oscillations of an inverted pendulum as shown.
The pendulum is restrained by a sufficiently stiff spring of spring constant k. Assume
that the entire mass of the pendulum is located at a distance I from the bottom
support.
Problems 247
/ —wvw
Fig. P 10-8 Kiii. r lU-'J
10-9. particles of masses m and 2m, respectively, are connected l)y a rigid bar
Two
of negligible mass and of length /. The system is initially at rest in the xy plane. A
force F of constant magnitude is applied to the mass m as shown; F lies in the xi/ plane
and is always normal to the bar. Find the angular velocitj' and the angular displace-
ment of the bar after a time interval t.
10-10. Two particles of masses m and 2m are moving with con.stant speeds v in
the xz plane and -\/5 v in the xy plane, respectively. At the instant under consider-
ation, m is located at point (2,0,2) and 2m is at point (1,1,0).
(a) Find the velocity of the mass center of the sj'stein.
(6) Find the angular momentum of the system about the origin.
(c) Find the angular momentum about the mass center (due to velocities relative
to the mass center).
(d) What is the physical significance of the dif'l'erence of {!>) and (c)?
C Jo
mi
a
10-12. Two particles of mass nii and m2, respectively, move about so that at a cer-
tain instant the relative velocity between the two is Vr and the velocity of their mass
center is Find the total kinetic energy of the system.
vc.
10-13. A pendulum (for determining the velocity of a bullet) consists of a
ballistic
bag of sand of weight W
suspended from a string of length I. The bullet, of weight w,
is fired horizontally into the sand. Find the velocity of the bullet if the pendulum
swings out to an angle a.
V >
W
Fig. P 10-13
10-14. Suppose, in Prob. 10-13, that the bullet is fired downward into the bag of
sand at an angle (3 with the horizontal. Find the speed of the bullet if the pendulum
swings out to angle a.
10-15. A box of sand slides down a plane of inclination a. with the horizontal. If
the mass of the box and its contents is equal to m at some time t and the sand spills
out of the box and continues down the inclined plane, find the equation of motion of
the center of mass of the box of sand. Assume that friction is negligible.
10-16. Suppose, in Prob. 10-15, that the sand is released in such a fashion that it
remains at rest after release. In this situation write the equation of motion of the
center of mass of the box of sand.
10-17. In Prob. 10-16, the mass of the box of sand and its contents at time I =
is mo. The sand escapes at rate k. If the box starts from rest, find its speed at some
later time t.
10-18. Given a rocket. Explain how you would conduct an experiment to measure
its thrust.
10-19. Under the conditions specified for Eq. (10-20), what physical conditions
should be striven for in order that a given rocket achieve maximum possible speed?
For this ideal condition, and assuming that the propellant has a specific impulse of
200 sec, determine the ratio of payload to weight of fuel necessary to achieve satellite
speed and escape velocity.
10-20. Estimate the rms velocity of hydrogen molecules at 15°C, considering the
density of hydrogen as one-sixteenth that of oxygen.
10-21. Suggest the dependence of the mean free path of an ideal gas on the pres-
sure and temperature. Estimate the mean free path of oxygen at 100°C and a pres-
sure of 0.001 atm.
CHAPTER 11
Moments and Products of Inertia
11-1. Introduction and Definitions
Before we discuss the dynamics of rigid bodies, it is convenient first to

introduce in this chapter certain quantities which are important in
defining the dynamic properties of a
body. These quantities are the mass
moments of inertia and the mass products
of inertia of the body. The role that
these quantities play in the dynamics of
a rigid body will be made clear in the
next chapter.
Consider a body (Fig. 11-1) and an
arbitrary set of cartesian coordinate
axes X, y, z fixed wdth respect to the
body. Thus the quantity that we shall
subsequently define will be referred to
these axes and therefore will be time-
Fic. 11-1
independent. Let dm
be the mass of a
small element of the body. The position vector of this small element dm.
with respect to the given coordinate axes will be represented by
r = x-i 4- //j + 2k
The mass moments of inertia of the body with respect to the axes x, y,
and z are defined asf
= /^,
(//- + 2-) dm 3=
1^,
(r2 - .t2) dm
= 1^,
{z'- -f x'-) dm = j^,
(r^ - //2) dm (11-1)
= f (.t:2 -H //2) dm = f .
(r2 - z^-) dm
t Note the slight difTerence of nomenclature used here compared with that used in
Sec. 5-4. Here we use double subscripts to denote moment of inertia, whereas before
we used the single-subscript notation. The reason for the double-subscript notation
will become clear when we discuss rotation of coordinate axes in Sec. 1 1-4.
249
250 Moments and products of inertia
where the integration is carried through the entire volume of the body.
The quantity Ixx is termed the mass moment of inertia about the x axis,
lyy the mass moment of inertia about the y axis, etc. In the expressions
of Eqs. (11-1), we see that the integrands (y^ + z^), (z^ + x^), (x^ + y^)
are the distances of the mass element dm from the x, y, and z axes,
respectively. The mass moment of inertia of a body about an arbitrary
axis is therefore the summation (or integration) of the products of the
constituent masses of the body and the square of their distances to that
axis and thus can be thought of as a second moment of the mass about the
axis in question. According to this definition, we immediately notice
that the mass moment
of inertia must always be a positive quantity
greater than zero regardless of the shape of the body, its mass distribution,
or the location of the axis. It might be added that our definition of mass
moment of inertia is in agreement with what is usually given in elementary
physics, in which I alone often stands for /^^ here.f
The mass products of inertia of the body with respect to the coordinate
axes X, y, and z are defined as
Ixy = j^ xy dm
(11-2)
Again, the integrations are to be extended over the entire volume of the
body. The quantity Ixy is called the mass product of inertia of the body
with respect to the x and y axes (or more precisely to the two planes whose
normals are the x and y axes, respectively) and is simply the summation
(or integration) of the products of the constituent masses and their x and
y coordinates. Similar statements hold for ly^, etc. Unlike mass
moment of inertia, mass product of inertia is an algebraic quantity which
may be greater than, equal to, or less than zero.
can be seen from the definitions of Eqs. (11-2) that Ixy = lyx,
It
lyz =
Izy, Izx = I xz- The three mass moments of inertia and the six mass
products of inertia can be conveniently written in an array as follows:
Sec- 11-1] Introduction and dejiniliona 251
have any particular meaning; it is simply, for the present at least, a

convenient method of arranging and presenting nine (|uantities.
In Eqs. (11-1) and (11-2), dm can be replaced by pdV, where p is the
mass density and dV = dx dy dz is the volume of the small clement. If
the body is homogeneous so that p remains a constant at all i)()ints of the
body, then Eqs. (11-1) and (11-2) can be written in the form
Ixz = p (//- + 2-) dV (1 1-la)

Jj.
/xv == p J^^xydV (ll-2a)
/xx
k) of the body about any axis is defined in such a way that mJc^ is equal to
the moment body about that axis. Thus the moment of
of inertia of the
inertia of a body remains the same if we replace the body by a particle
having the same total mass as the body and located at a distance k from
the axis.
11-2. Translation of Coordinate Axes
Let us suppose that with respect to a given set of axes xyz, all the
moments and products of inertia hx, hy, etc., of a given body are known.
-, We wish to obtain the moments and
products of inertia I'^^, I'^y, etc., with
'^
respect to a new set of axes x'y'z' ob-
tained by a translation of xyz (Fig.
11-2). Let the origin 0' of the new
system be located at {a,b,c) in the xyz
system. We thus have the trans-
formation equations of the two coor-
dinate systems:
X = X —a
y' = y -b
z' = z — c
Fig. 11-2
By definition [Eqs. (11-1)], the
moment of inertia about the x' axis therefore becomes
/L = /^ {])" + ^'^) dm = [{y - hr -\-iz- cy] dm

f^
= I XX + (6^ -\- c^)m — 2bycm — 2czcm (11-4)
where yc and zc are the coordinates of the mass center in the xyz system
and are given by Eq.s (5-6). Similarly,
r^y = dm = j^(x - a)(y - 6) dm

f^^'y'
= Ixy + (ab)m — ay cm — bxcm (11-5)
Our task of finding /^^, I'^^, etc., is thus completed.

As a special but extremely important and useful case, let us suppose
that the origin of the xyz system falls on the mass center of the body.
We then have xc = yc = zc = 0, and Eqs. (11-4) and (11-5) are reduced
to
IL = Itx + {h^ + c2)w = I XX + d'm (11-6)
I'xy = I*xy+{ab)m (11-7)
where d is the distance between the two parallel axes x and x'. Equations
(11-6) and (11-7) are sometimes referred to as the parallel-axes theorem of
Sec. 11-3] Moments and producis of inerlia of some simple bodies 253
the moments and products of inertia. The special case of a plane area
has been treated by Eqs. (5-22) and (5-23) in Chap. 5. It must be
emphasized that Eqs. (11-6) and (ll-J) are valid only xijz axes
if the
are the center-of-mass axes, i.e., axes having their origin at the mass
center. Accordingly, moments of inertia taken with respect to axes
passing through the center of mass (or centroid) are indicated by an
asterisk in Eqs. (11-6) and (11-7), thus emphasizing this point. Ecjuation
(11-6) can be expressed by the following statement: The moment of
inertia of abody with respect to any axis is equal to the sum of (1) the
moment of inertia with respect to a parallel center-of-mass axis and (2)
the product of the mass of the body and the square of the distance between
the two axes. A similar statement holds for Eq. (11-7). Since each
term on the right side of Eq. (11-6) is positive, we can conclude that of
all parallel axes, the moment of inertia taken about an axis passing through
the center of mass must be a minimum. A similar statement cannot be

made about the product of inertia.
11-3. Moments and Products of Inertia of

Some Simple Bodies
In this section we shall derive, as examples, the moments of inertia of a
rectangular parallelepiped, a cylinder,
and a sphere. In all three cases the
body is assumed to be homogeneous.
Evaluation of the integrals involved
may be accomplished in any permis-
sible fashion; for purposes of illustra-
tion we choose different methods and
usually the most expeditious.
Rectangular Parallelepiped {Fig.
1 1-3). Let the origin be the center of
mass of the body; then x, y, z are the
center-of-mass axes and, since the
body homogeneous, are also the
is
Kic. 11-3
centroidal axes for the volume. Con-
sider, for example, 7*^. From Eq. (11-1 a),
-b/2 Jr"
n.- pT"
^ J -c/2 Jr"
{<f+z^)dxd,dz '
-a/2^-^
Evaluating the triple integral, we find
^ pabc{b'^ -f c^)
12
mjb^ -H c")
Hence I* =
12
where m = pabc is the total mass. Similarly,
I*yy - 7n{a^ + c^)

n^^ _
mâ"^ + h^)
^
12 12 II
For the special case of a thin rectangular plate in the xy plane, we take
c = 0. Then
I* = mb^ ^^
/* =
ma"^ ,^ ni(a'^ + h^)
12 ~Y1 12
Circidar Cylinder {Fig. ll-4a). Let the radius of the cylinder be R and
the length be I. Again we choose the x, y, and z axes to be the centroidal
(b)
Fig. 11-4
axes. To calculate I*^, we proceed as follows: Consider an elemental

disk of radius R and thickness dz, at a distance z along the z axis from Cx.
The mass of this disk is pirR"^ dz, and its mass per unit area is o- = p dz.
The moment about an axis in the plane of the disk
of inertia of this disk
and parallel to Cx is therefore awR^/A = prR'^ dz/4: [Sec. 5-4 and Eqs.
(11-16)]. By the parallel-axes theorem, the moment of inertia of the ele-
mental disk about Cx is
dl*^ = P]^^^ + p^jiiî dz
is the sum of the moments of inertia of

Since I*^ of the cylinder all such
elemental disks extending from z = — 1/2 to z = -\- 1/2,
1/2 (J^l \ mR^ ml^
r PTT.i?2 dz
T' ^ 12
where m= pirRH is the total mass of the cylinder. The above evaluation
Sec. 11-3] -Monienls and products of inertia of some simple bodies 255
of 7*^, due to symmetry, clearly applies to I*y as well. Hence

I*
* J/l/
= I*XX
•"
\ would of course give rise to the same

direct application of Eqs. (11-1)
results; the method employed here, however, proves to he more efficient.
To calculate /*, we divide the cylinder into thin cylindrical shells of
radius r, length /, and thickness dr (Fig. 11-46). Thus r- = x- -\- y- in
this case. The mass of one such shell is dm = 2irrlp dr, and its moment of
inertia with respect to Oz is
(//* = (.1-2
+ ,^2) fifn ^ 2iTrHp dr
The moment of inertia of the entire cylinder with respect to Oz is therefore
/* = 27r/p / r' dr =
^
Sphere. Again we choose the center of mass, in this case the center
of the sphere, as the origin of our coordinate axes. By definition, or
Eqs. (11-1), we have
/* 4. /* 4- I* = 2 [ r- dm
where r- = x- + //" + z-. In the case of a sphere, all three axes x, y, and z
are diameters. Hence
= = =
-^'AIr-'dm
/* /* /* 2,
Dividing the sphere into thin spherical shells of radius /• and thickness dr,
the mass of one such shell is dm = -iirr'-p dr. For the entire sphere of
radius R, we obtain
/ r- dm = 47rp / /•* dr = -îpnrR^
Hence I*^ = I*, = /* = H^pivR' = HmR^

where m = %fyirR^ is Moments and products of
the mass of the sphere.
inertia of some additional geometric shapes and homogeneous bodies are
given in Appendixes A and B.
The products of inertia of the above three cases with respect to the
given center-of-mass axes are all zero. This can be shown in the following
way. Consider first the xy plane, which is seen to be a plane of sym-
metry. For each dm at -\-z we have a corresponding dm, of the same x
and y, at —z. The summation of xz dm and yz dm therefore vanishes, or
Izz = lyz = 0. Similarly, since xz is a plane of symmetry, I xy = lyz = 0.
Thus all products of inertia are zero with respect to the particularly
chosen axes. Generalizing this, we can see that the products of inertia
1^
(mass, volume, or area) vanishwhen taken with respect to any two axes
whose symmetry, provided one of the two axes is
origin lies in a plane of
perpendicular to that plane. For a nonhomogeneous body, a plane of
geometrical (volume) symmetry is not necessarily a plane of symmetry
with respect to mass. For plane areas, these statements imply that when
one axis about which the product of inertia is to be found is an axis of
symmetry, the product of inertia vanishes.
Often in practice, we are faced with the problem of determining the
moments and products of inertia of irregular bodies, bodies composed of
several portions. In this event one simple integral will not suffice to find
the moments and products of inertia. However, a method of composite
bodies will enable us to solve the problem. According to this method,
the moments and products of inertia of a body composed of n portions
may be determined by expressions of the form
n n
-i XX / J-
-t XX;
xxi -l
-^ xy 7
/ -* xv: ' ' '
where /ix, and moments and products of inertia of each portion,

Ixyt are
respectively. The validity of this method follows directly from the
definitions of Eqs. (11-1) and (11-2). In the use of this method it is
immaterial whether Ixxi and Ixy, are computed directly by integration or
their values taken from the tables of Appendix A or B. If the tabular
values are used, might be necessary also to utilize the parallel-axes
it
theorems in order to have the moments and products of inertia taken

about the proper axes. An illustration of this method will be given in
Example 11-2.
11-4. Rotation of Coordinate Axes
In Sec. 11-2 we discussed the change to the moments and products of

inertia due to a translation of the coordinate axes. In this section we
shall consider the change due to a rotation of axes. We ask the following
question: If the moments and products of inertia of a body are known
with respect to a set of axes xyz, what are the moments and products of
inertia of the body with respect to any new set of axes x'y'z' that are
inclined to the original xyz axes? In such a transformation, only the
coordinate axes have been rotated; the body itself is supposed to remain
fixed. (If the body and the axes rotate as a unit, then the moments and
products of inertia of the body will remain the same with respect to the
same set of axes and therefore we have no problem.)
In the study of transformation of axes, it is convenient to adopt the
following notation. The original axes will be called x-^x<2.Xi (instead of
Sec. 11-1] Rotalioii of coordinale a.res 257
xyz). The unit vectors along these axes will be called ii, io, ij (instead of
the usual i, j, k). Similarly, the new rotated set of axes will be called
whose unit vectors are ij.
x'lXlr's, ij if,, (Fig. 11-5). In this new notation
the array of Eq. (11-3) is rewritten as
In
- (/) (11-8)
, etc. I'roni the definitions of
ky.xz
values of the complete set of Iki's, where the indices i, j, Jc, I can be any one
of the three numbers 1, 2, or 3. Before undertaking this task, we shall
first consider the simpler but more basic problem of describing the
components of a vector in terms of rotated coordinates.
11-5. Orthogonal Transformations
In this section we shall consider the following problem : Given the three
rectangular components Xi, X2, Xz of a fixed vector r. What are its
components x'l, x'2, x'^ when the coordinate axes are rotated?
Let us use the symbol {x\,x^ for the angle between the positive direc-
tions of the axes x\ and Xj. We also adopt the notation = cos {x'i,x^.
a,-,-
Thus the nine a^-'s represent the nine direction cosines between the two
sets of axes x']_,x'2,x'^ and x^^x^.x^. They can be written
an =
or, in general,
Sec. 11-5] Orlltugonal IransJ'urnialions 259
etc., from which
»i = 2"^'»' iJ = 1. ^. 3 (11-17)
j
The
nine elements of the tran.sformation matrix of Eq. (11-18) are not
allindependent because the unit vectors in each set of coordinate systems
are mutually perpendicular and have unit magnitude. In other words,
we must have
i[ ' i'o — io • i'n = i'-, • i[ —
with similar relations for ii, U, 1,^. Substituting Va\. (11-15), or its
e(iuivalent, Eqs. (11-14), into the six equations (11-18), we obtain the
following six equations:
011^21 -|- 012022 + 013023 =

O21O31 -|- 022032 + O23O33 =
O31O11 + 032012 + O33O13 =
On- -f 012- + oi.r = 1
021- + 022- + 023- = 1
O3I- -f- 032- + O33- = 1
The first three equations of (11-19) are called orthogonality relations since
they come from the fact that i^, i2, is are mutually perpendicular. The
la.st three equations of (11-19) are called normalization conditions; they
come from the fact that ij, i2, is are each of unit magnitude. All six
equations can be represented by the following simple ecjuationit
N o,A.o>fc = bij (11-20)

k
t After the reader is accustomed to this type of notation, it is easy to derive Eq.
(11-20) without having to go through the process of writing out Eqs. (11-19). Let
us take Eq. (11-15):
"i = ) ««;«j
Similarly, \[ = y Okiii
Forming tlie dot ])n)d\ict of the two, rcinciiilx-riiig that the dot product is associative,
we obtain
i'i
•
i'k
= 5,t = ) ) (lijUiuij ' ij
but ij • i; = 8ji
Hence 5ik = / (lijUk}
J
which is the same as Eq. (11-20).
Similarly, if we substitute Eq. (11-17) into the six equations analogous

to Eqs. (11-18) but for ii, i2, is, we obtain
y^ttkiakj = 8ij (11-21)

k
It is possible to show that Eqs. (11-20) and (11-21) are equivalent. The
nine elements ajj, therefore, are related by the six equations of (11-20) or
(11-21), and only three elements out
of the nine are independent. This,
with our previous knowledge that the orientation of a
of course, agrees
rigid body with one point fixed is determined by three independent
variables (Sec. 7-3).
Let us now return to the original problem we have proposed to solve.
Consider any vector r whose components are Xi, x^, Xz in the unprimed,
or original, coordinate system and x[, X2, x'^ in the primed, or rotated,
coordinate system. Forming the dot product of r with Eq. (11-15) and
noticing that r • i'- = x'^ and r = Xj, we obtain
• i_,
x'i = yaijXj (11-22)

y
Similarly, forming the dot product of r with Eq. (11-17),
Xi = yajix'j (11-23)
Equations (11-22) and (11-23) represent a special type of linear trans-

formation, called an orthogonal transformation. They are a generalization
of the results of Prob. 1-10. In an orthogonal transformation, the trans-
formation elements of Eq. (11-13) satisfy the relations of Eq. (11-20) or
(11-21). Through such a transformation the components of a fixed
vector along any rotated orthogonal coordinate system can be found when
its components are known with respect to an original, or unrotated,
orthogonal coordinate system.
11-6. Moments and Products of Inertia with

Respect to Rotated Coordinates
Now that we have investigated the orthogonal transformation of

components of a vector in rotated coordinate systems, we may attack the
problem posed in Sec. 11-4 with ease. As stated before, our definitions
of moments and products of inertia can be represented by Eq. (11-11) i.e., ;
I'ii = Rr^^îj - x'ix'j) dm

Substituting for .r' and Xj their values as given by Eq. (11-22), with a
Sec. 11-6] Momenls and producls of inertia; rotated coordinates 261
necessary renaming of some indices, i.e.,
the result is
I'ij = îj r- dm - ^ y a^kQji j x^xi di

J
k I
However, using our definitions of moments and products of inertia

[Eq. (11-9)], i.e.,
Iki = (r-dki — XkXi) dm

I
we obtain /-^ = 8,j

J
r- dm + Y X aa-ajiihi - 5*/ f /•-"
dm)
k I
But the coefficient of the last term can be reduced in the follow ing manner:
) } a.kOji 8u = ) n,kajk = 5,j
k I
(11 -24a)
(11-246)
This completes our task of finding the moments and products of inertia
with respect to a rotated coordinate system. I^(iuations (ll-24a) and
(11-246) are sometimes called tensor transformations; they provide us
with a means of computing the moments and products of inertia with
respect to a rotated set of axes in terms of the moments and products of
inertia with respect to an unrotated set of axes, or vice versa.
To Eq. (ll-24a), we shall write out the complete
illustrate the use of
erjuation for the following two elements of /'y:
^11 = O'li'În + ai2^/22 + 013^/33 + 2aiiai2/i2 + 2anai:j/i3 + âxâuhz

I'll = dnaiJii + ai2«22/22 + dvid-iJ :i^ + (011022 + a,via-n)I vi
+ {a\\(ir.i + aufltzO/i.t {nioa->z + '''i3«22)/23

-f-
Referring back to the usual xyz and x'y'z' notation and recalling that
J\i = I XX, 1 12 = —Jxy, etc., we can write out the moment of inertia /'„
:
and the product of inertia I'„j with respect to the rotated coordinates as
follows
/^^ -- aiiÎxx + aîHyy + ai3^/.2 — 2aiiai2lxv

- 2aiiaiJxz — 2a-i_2ai3ly, (11-25)
— I xy — 0,l\0'2llxx ~\~ a^^diilyy H" dud^S^zz + ^12^^21)-'
(Ctlia22 x;/
— (aiiaos + aua2i)Ixz — {ai2aiz + ana22)Iyz (11-26)
Similarly, we can write out the remaining moments and products of

inertia with respect to the rotated axes. Our adoption of the new index
notation enables us to have all nine such equations represented by a
single equation, namely, Eq. (ll-24a), thus establishing the efficiency
of the new notation.
11-7. Cauchy's Inertia Ellipsoid
Suppose we have a body whose moments and products of inertia with

unprimed coordinate system at a point
respect to the. are known.
Fig. 11-6
Equation (11-25) tells us that the moment of inertia 7^^ of the body
about any inclined axis x' passing through that point is dependent on
the three direction cosines of x', i.e., an, ai2, and au. Let us now lay off
on the inclined axis x' (which passes through the same origin 0) a distance
OP which is numerically equal to l/\/7^. If the three coordinates of P
with respect to the unprimed system are denoted by x, y, and z (Fig. 11-6),
then by this construction
x = aiiOP = an/Î'xx
y = av£>P = aii/Vl'xx
z = auOP = au/\/Pxx
Solving the above three equations for an, Oio, an and substituting into
Sec. 11-8J Principal monienis of inertia and principal axes 263
Eq. (11-25) yields

I
'
/...r- + /,,,//- + I,,z- - 21,, XII - 2h,xz - 2Iy,ijz = 1 (11-27)
,
This is the eciuation of a (luadiic surlWe (Fig. Il-(i6). It is, in fact, the
equation of an ellipsoid about the center since the coefHeients of the
I
first three terms are positive here. The locus of the point P, therefore,
is an ellipsoid, called the inertia ellipsoid by Cauchy (1827). It is also
known as the momental The geometrical interpretation of the
ellipsoid.
ellipsoid is useful and enlightening. Thus the moment of inertia about
any line L passing through is numerically equal to the reciprocal of the
square of the distance from the origin to the point where the line L
intersects the ellipsoidal surface.
The concept
of the inertia ellipsoid is an important one. (As a matter
of the geometrical representation of a physical quantity by a
fact,
quadric surface is a useful device appearing frecjuently in various portions
of mechanics.) Our original unprimed axes xyz may, after all, be ori-
ented arbitrarily with respect to the body. With each different orienta-
tion of X, y, and z, we shall have different values of Ixx, Ix,,, etc. Thus
Eq. (11-27) will look different, since all its coefficients will be different.
Yet all such different-looking equations will describe the same ellipsoid
with respect to the body. In other words, the inertia ellipsoid, represent-
ing the inertial properties of the body at 0, is an invariant of the body and
is independent of the orientations of the coordinate axes, provided the origin
is unchanged. (The inertia ellipsoid changes, however, Avhen the origin

is shifted to a different point of the body.) It will be shown in the next
chapter that two rigid bodies having the same mass and the same inertia
ellipsoid about their mass centers will exhibit identical dynamical
behavior.
11-8. Principal Moments of Inertia and Principal Axes
Every has at least three principal diameters. One of these

ellipsoid
is clearly the largest of all possible diameters, while another is the smallest
of all possible diameters. Suppose we denote the three principal diam-

eters (principal axes) of the inertia ellipsoid by ^, rj, and f. Then the
ecfuation of the ellipsoid with respect to the ^, 17, T coordinate system is
I^^^^-{-I^^r,' + I^,r~ = 1 (11-28)
where /j^, /,„ and /j-j- are the moments of inertia with respe(;t to the
principal axes. They are called the principal moments of inertia. The
planes formed by the principal axes are termed principal planes. Since
the moment of inertia about any axis is numerically equal to the reciprocal
of the s(iuare of the radius along that axis, from the origin to its point
:
of intersection with the ellipsoid, the moment of inertia becomes a

maximum if the radius of the ellipsoid is minimum, and vice versa.
Hence, the targest of the three principal ynoments of inertia is also the
maximum among moments of inertia of the body at 0. Simi-
all the possible
larly, the smallest principal moment of inertia is also the minimum among
all the possible moments of inertia of the body at 0. Every rigid body
therefore possesses three principal axes, corresponding to the three
principal diameters of the inertia ellipsoid. For the special cases where
the inertia ellipsoid becomes an ellipse of revolution or a sphere, there are
more than three principal axes (and principal moments of inertia), each
corresponding to one of the principal diameters of the inertia ellipsoid.
Comparing Eq. (11-28) with Eq. (11-27), we see that with respect to
the principal axes ^, rj, and f, all six products of inertia (/j„ etc.) vanish.
The moments and products of inertia with respect to the principal axes
therefore appear as
//,, \
/„ (11-29)
\0 /,,/
We can thus define the principal axes of moments and products of

inertia in any one of the following three ways
1. They are the principal diameters (principal axes) of the inertia
ellipsoid.
2. They by two axes about which the moments of inertia
are formed
are, respectively, a a minimum and a third axis perpen-
maximum and
dicular to the first two. Combining this result with what we have
previously determined relative to the translation of axes, we immediately
observe that of all possible axes the minimum moment of inertia occurs
about a principal axis passing through the mass center of the body.
3. They are the axes about which all the products of inertia vanish.
A substantial reduction in the number of terms appearing in Eqs.

(11-25) and (11-26) may be achieved by referring the rotated system to
the principal axes. Thus the moments and products of inertia with
respect to a set of coordinate axes x'y'z' that are inclined to the prin-
cipal axes ^, rj, and f are, by Eqs. (11-25) and (11-26),
/;, = anV^^ + a,,'~I,, + au'In (11-30)

— I'^y = aiittoi/a + aiidi'ilnn + aiâ^Jn (11-31)
etc., where an, etc., are the direction cosines between the x'y'z' and the
^77f axes.
We see that Eqs. (11-30) and (11-31) are considerably simpler than
Eqs. (11-25) and (11-26). The computation of moments and products
of inertia about rotated axes can therefore be simplified if the unrotated
Sec. 11-8] Principal moments of inertia and principal axes 263
axes are the priiioipal axes. It is theicl'ore useful to locate the principal
axes and to compute the moments of inertia. To locate the
principal
principal axes of the ellipsoid such as the one defined l)y 1m i. (1 1-27) is a
problem of geometry. The derivation is somewhat lengthy, and we shall
merely state the result in the following theorem.!
Theorem. If a, 13, and y are the direction cosines of any one of the
three principal axes with respect to the given xyz coordinate system, the
following three equations must then all he satisfied:
(/xx -
- I, - /,,7 =
\)a ,13
-Iy,a - \)I3 - /,,,7 =

-f (/,„ (11-32)
-h,a - /,,/5 -f (/,, - X)7 =
The above three eciuations can have a nontrivial solution for a, /3. and 7
only if the determinant of the coefficients vanish; i.e.,
^xx X ~I XI, — I IZ
^ i/X I III/
A 1 y, = (11-33)
i zi i zii i zz — A
This a cubic equation for X.J Solving this equation, we obtain, in

is
general, three roots of X. It can be shown that the three roots of X Xi, —
X2, and X3 — areand are just the three principal moments of
all real
inertia Ijj, I,,,, and Iff. For each value of X, say Xj, we can solve for
the ratios î/at and 7,/ai by using any two of the three Eqs. (1 1-32). It
is noted here that onl}^ two of these three equations are independent by
virtue of Eq. (11-33). These two ratios together with the equation
tti^ + iSi^ + 7i~ = 1 are sufficient to give us one set of values of a„ /3,, 7,.
The direction cosines thus obtained are the ones belonging to that par-
ticular principal axis with respect to which the moment of inertia has the
assigned value of Xj. Thus, if we substitute Xi = /^^ in Eqs. (11-32),
the resulting ai, |8i, 71 are the direction cosines of the principal axis |.
Taking successively the three values of X, we can find all three sets of
direction cosines belonging to the three principal axes. As stated before,
the three principal axes are mutually perpendicular.
We thus have reached the following objective: Starting with a given
set of axes xyz and the array
/xx
(/)
=
t The discu.s.sioii in this .section, including thi.s theorem, turns out to be the con-
sequence of some more general theorems on symmetrical UiiHors. A rigorous proof of
this theorem will therefore be given in its proper place in Chap. 14 (Volume 2).
JSee Prob. 11-7.
;
at a point 0, we can compute the three principal moments of inertia and

the orientations of the three principal axes through by means of
Eqs. (11-32) and (11-33).
In practice, the task of locating the principal axes can often be simpli-
fied by the following considerations. These considerations are based on
the fact that the moment of inertia about any axis determines uniquely
the radius along that axis of the inertia ellipsoid. If a body has a plane
of symmetry and if the origin lies in this plane, the inertia ellipsoid
with the center at is also symmetrical about the plane. (The reverse
is not necessarily true.) This means that two of the principal axes
through must lie in the plane of symmetry
and the third one perpendicular to it. A thin
fiatplate whose generators are all parallel
would constitute an example of this situation,
the middle surface of the plate being a plane
of symmetry. Since our interest in the iner-
tial properties of a body is usually confined
to the properties at the mass center, which
necessarily lies in the plane of symmetry if
there is such a plane, the above consideration
is often very useful.
If the body has two perpendicular planes of
Fig. 11-7 symmetry, its principal axes, at a point on the
the two planes, accord-
line of intersection of
ing to the above consideration, can be determined without any calculation
at all : the first principal axis is the intersecting line of the two planes of
symmetry; the second axis lies in one plane of symmetry perpendicular
to the first axis; and the third axis lies in the second plane of symmetry
perpendicular to both of the first two axes. Because of the symmetry
here, it is easy to visualize that the products of inertia, about the three
axes described, all vanish, and again it follows that these axes are prin-
cipal axes.
A special case of the above is where the body is a volume of revolution
the inertia ellipsoid at a point on the axis of revolution is then an ellipsoid
of revolution. One such example is a conical top (Fig. 11-7) for which
the principal axes are the longitudinal axis of the top and any line in the
equatorial plane, i.e., the plane normal to the longitudinal axis at the
point in question. Again, considerations of symmetry would lead to the
same any plane passing through the longitudinal axis is a plane
result;
of symmetry. The inertia ellipsoid at a point on the longitudinal
centroidal axis also reduces to an ellipsoid of revolution for a homogeneous
prismatical body if the cross section of the body is an equilateral polygon
or a circle. Extending this consideration still further, it becomes obvious
I
S.r. 11-8] Principal ni(tnieiils of inerlia and princijnil axes 267
tliat the inertia ellipsoid at the eentroid of a hoiiiogeiieous cuhic Ixuly

(ie^t'iierates into a sphere (Fig. 11-8). for whicli any axis is a prineipal
;i\is. Likewise, the inertia ellipsoid at the centroid of a homogeneous
-phere another sphere; every radius is a principal axis.
is
Exaini>le 11-1. In order to illustrate the general theory' just di.s-

cussed, let us consider the following example. Given a eube of dimen-
sion a as .shown in Fig. 11-9. We desire to find the moment of inertia
ahout the diagonal axis pa.ssing through O and al.so to determine the
j)rineipal axes and the corresponding princi|)al moments of inertia at ().
Fic. 11-8 Fic. 11-0
Using the results of Sec. 11-3, we find the moments of inertia about
the center-cjf-mass axes as
' xz ' ' zz

yy
IT
' zy ' zz ' yz "
According t(j the parallel-axes theorems [Eqs. (11-6) and (11-7)],
Ixx = In = Izz = Hma^

Ixy = Ixz = lyz = ^^0^
Noting that the direction cosines between axis ;/ and the axes x, ?/, and
are given as
1
flux CIuu "ut

V3
we obtain, from K(|. (1 1-25),
/,^„ = 2/â2 - i-^wa2 = i.^n?fl2
In order to determine the princii)al axes and principal moments of

inertia we let c = ma-/l2 and form the determinant of Va\. (11-33).
Thus
8c - X -3r -3r
-3c Sc - \ -'So =
-3c -3c 8c - X
Expanding this determinant, we obtain the following cubic equation in X:
X^ - 24cX2 + leSc^X - 2426=* =

whose roots are
Xi = 2c X. = lie X3 = lie
Therefore the three principal moments of inertia are
Substituting the value of Xi into the first two of Eqs. (11-32) and solving,
we obtain the following ratios of direction cosines:
^ = 1 :^ = i
Using these values in conjunction with the equation
«i- + (3r + 71- = 1
the three direction cosines of the first principal axes ^ are found to be
„, = ^, = ^, = -L
Thus the principal axis ^ coincides with the body diagonal u of the cube.
Since the other two principal moments of inertia are equal, it is clear
that the inertia ellipsoid an ellipsoid of revolution whose axis is the
is
body diagonal. Therefore any axis at lying in the plane normal to the
^ axis is a principal axis. From symmetry considerations alone, the
diagonal plane passing through is a plane of symmetry, and thus two
principal axes lie in this plane; the third principal axis is perpendicular
to it.
11-9. Plane Rotation of Axes; Mohr's-circle Representation
If a body possesses a plane of symmetry, an axis perpendicular to that

plane is a principal axis at its point of intersection with the plane. Then,
to determine the remaining principal axes, we consider a rotation of
coordinate axes taking place in the plane of symmetry of the body, and
we have an especially simple case. Let the plane of symmetry be the
xy plane. In a plane rotation of axes, the rotated axes .(•'//' lie in the
original xij plane. In other words, the z and z' axes coincide and are
principal axes.
From the discussion of Sec. 11-3, we see that hz = /-x = /,/.- = hu = 0.
Let the angle of rotation from xy to .1'//' be <^ (Fig- U-IO). Then
ttii = a22 = cos (f), ai2 = — ttoi = sin 0, and a.u = n-^-i = 0^ = 023 = 0.
Sec. 11-9] Plane rotation of uses: Mohr's-virde representation 269
With respect to the rotateil axes .c'//'. we have. Irmu Imjs. (11-2")) and
(11-20) or the more jj;eMeial Imis. (11-24).
I'zi = I XI COS- -|- /y„ sin- 0.— 2/x„ cos sin
/^„ = Ixs sin- -I- /„„ cos- -I- 2/,„ cos sin (1 1-84)
J't^ = i^xx - I„u) P"!^ <t> ^=ii> + (f'os- - sin- 0)/^„
F^nipioyino; a trigonometric sul)stitution. these e(iuations can he rewritten

in a more convenient form as foUows:
/' = -^-^
' I'!'
I f_
/xx
" yy
ny =
If we now let the unrotated axes be the principal axes from which the ]
angle of rotation 0p is measured, we have, as special cases of Eqs. (11-35),
hi + hv I
hi —A
r =
hy-
— /
^ff. 1 1-*>] Plane ratal ion of (ues: Mohr's-vircle re/)resenlal{on 271
The above method l)y graphical construction gives us the moments
;incl products of inertia with respect to a pair of axes that are ()l)tained
l>y a phine rotation from the principal axes ^ and t;. By a reverse pro-
(cdure, the same method can also give us the orientation of the principal
;i\esand the principal moments of inertia from a given set of /„, I^y, /,„.
A little deliberation of the geometry of Fig. 11-11 will show that
Eqs. (11-3(3) and (11-37) come out as natural conse(iueMces in the method
of Mohr's circle.
A more formal proof of the Mohr's-cirde construction may be obtained
as follows. From Efjs. (11-35), we have
/;. - ^" ^ ^""

= ^" ~ ^""
cos 20 - /,, sin 20
/;. = ^"^ ~ ^"^

sin 20 + /.„ cos 20
Squaring both sides of these ccjuations and adding the results, we obtain,
after some simplification,
(J,
/..
I XX ~r
^
' ii!i\'
)
j^
+
jfo
^'^ ^
V ——
I i xz ^ i yu\' I
J 2
This is the equation of a circle in the /^^, /^^ plane. The center of the
circle located on the abscissa a distance of (/„
is + lyy)/'^ from the origin,
and the radius is equal to \[{Ixx — Iyy)/2]- + Ixv^\^'-. It corresponds
exactly to the construction previously described.
Since the accuracy of such a graphical construction is usuall}' poor
unless great care is perhaps lies
exercised, the value of the Mohr's circle
in enabling us to visualize the over-all behavior of the problem and to
write out Eqs. (11-36) to (11-38) with little need of memorization.

For example, several of the properties of the moment of inertia previ-
ously discussed are evident from this construction. To cite only two,
we see immediately that the principal axes corresponding to the maxi-
mum and minimum moments of inertia in the plane are axes of zero
products of inertia and that the two principal axes are normal to each
other; i.e., the angle between the ^ and r; axes is = 7r/2. Other prop-
erties are equally obvious.
It might be mentioned in closing that the method of Mohr's circle is
based on the transformation characteristics of a carlesian tensor under a

plane rotation of axes. The subject will be discussed once moic and wit h
greater generality in Chap. 14.
Example 11-2. An illustration of these methods now follows. Sup-
pose we have a plane area composed of a rectangle and triangle as shown
in Fig. 11-12. At the point 0, we desire to locate the principal axes,
determine the magnitudes of the principal moments of inertia, and find
the moments and products of inertia for axes x' and y', rotated through
a positive or counterclockwise angle of 30° with the x, y axes at the
point 0. Clearly the plane rotation of axes applies in this case; we
therefore propose to use the Mohr's-circle construction.
We first compute the moments and products
of inertia with respect to
the unrotated axes at 0. This may
be accomplished by the use of the
method of composite areas (Sec. 11-3), the parallel-axes theorems, and
the tabular values given in Appendix A. Suppose we divide the figure
.
Sec. 11-9] Plane rolalion of axes; Mohr's-circlr rrf)resenlali()n 273
The principal moments of inertia, corresponding to the abscissas of

points .1 and B, are then
/ij= 3,780 + 2A:ir, = 0,2 1.-) in.*
/„ = 3.780 - 2.43.") = 1.34.") in."
The location of the major principal axis ^ determined by the angle 20 is
defined by
tan 20 = ":'1{,- = 4.40

.)40
or 20 = 77°12'
and is measnred negatively from the x axis. The minor principal axis 77
4320 I
2376
Fi(i. ii-i:i
i.s perpendicnlar to the ^ axis. The moments and prodncts of inertia
for the axes x' ,

//' are defined by the points /•'
and (1, where the \\m\ FG
makes an angle of 20p = (>0°0' + 77° 12' = 137° 12' with th(; principal
axis ^. Observing the geometry indicated in I'ig. I 1-13, we obtain
/'„ = 3,780 - 2,435 cos 42°48' = 1. *.»'.».") in.'
/;„ = 3,780 + 2,435 cos 42°48' = 5,5()5 in.'

/' = 2,435 sin 42°28' = 1,()55 in."
Clearly, we could have obtained these same results by a direct application

of Eqs. (11-36) to (11-38).
PROBLEMS
11-1. Show that III = I [r^ — (r • n)^] dm, where L is any line in space passing
through the origin and defined by the unit vector n, and r isthe position vector of an
element of mass relative to the origin. Verify Eq. (11-25) by use of this relationship.
11-2. direct integration find the mass moment of inertia of a homogeneous cone
By
of radius rand altitude h about the principal axes passing through its center of mass.
11-3. Determine the mass moment and products of inertia of a homogeneous
paraboloid of revolution about center-of-mass axes.
11-4. A composite body consists of a homogeneous hemisphere of radius r and mass
density pi atop a right circular cylinder of radius r and length I and mass density po.
Find the mass moment and product of inertia of the body about axes whose origin lies
at the base of the cylinder. Explain how you would obtain the mass moment and
products of inertia about center-of-mass axes.
11-5. From the result of Sec. 11-2, find the moment of inertia of a solid hemisphere
about a diametral axis A A. Then by means of the parallel-axes theorem, find the
moment of inertia about an axis BB tangent to the North Pole.
Fig. P 11-5
11-6. Write out the complete set of Eqs. (11-22) and (11-23) for a two-dimensional
orthogonal transformation assuming <p is the angle from x[ to Xi. Also write out the
orthogonality relations, Eqs. (11-20) and (11-21).
Fig. P 11-6
11-7. On the basis of Eq. (11-33), write out the cubic equation of X.
Problems 27.')
11-8. A uniform cylinder is of radius /?, lenRth /, and mass //(. Find tlir
solid
moment of inertia /^ând the produet of inertia /^^ about the inclined axes x'y'
shown. (The x' axis n)akes an enclosed angle of 'M° in the vertical plane witli the
base of the cylinder.)
Fig. 1' 11-8 Fio. r 11-10
With respect to a certain set of cartesian coordinate axes, the moments of

11-9.
body are 3, 4, 5 and the products of inertia vanish. Find the maximum
inertia of a
and the minimum moments of inertia of the body with respe<d to all j)o.ssible axes
through the origin.
11-10. Find the moments and products of inertia of a homogeneous, solid parallele-
piped with respect to the coordinate axes at a corner. How would you locate the
principal axes through O?
11-11. Show that the sum In + I22 + hs is invariant; i.e., the sum of the tiiree
moments of inertia is con.stant and independent of any rotation of axes.
11-12. Specialize the result of Prob. 11-11 for the case of plane rotation; i.e..
Ill + lyv is invariant. Verify this result by the use of Ecis. (1 \-'.\\) and by the Molir's-
cirde construction.
11-13. If the inertia ellipsoid at the mass center of a homogeneous circular cylinder
is a sphere, find the length of the cylinder as a ratio to the radius.
11-14. Calcidatc the moment of inertia of a homogeneous right circular cone of
altitude h and radius of ba.se a with respect to a generator.
11-15. Derive Ecj. (11-24^).
1 1-16. Using \'a\. (1 1-24«), write out tiie complete exj)re.ssions for I .,., and Z,,. Al.so
write the complete expressions for /^^ and /,,.

11-17. In Fig. 11-10, we can see that
x' = I cos 4> -f- v sin <^
y' = // (;os 

— J- sin <^
Using these two equations, derive once more Kqs. ( 1 l-iiS) for the ca.se of a plane rota-
tion of axes without going through the general Eqs. (11-24) to (11-20).
11-18. For a plane rotation of axes, find the orientation of axes about which the
product of inertia is zero. Also determine the orientation of axes about which the
product of inertia is a maximum or minimum. How are these axes oriented with
respect to the principal axes?
11-19. Verify the results of Prob. 11-18 by the method of the Mohr's-circle con-
struction.
11-20. Find the area moment of inertia and the area product of inertia of the
Z-shaped area shown in the figure.
-h.
Fig. P 11-20
11-21. Determine the moment and product of inertia of the area included between
the X axis, the curve y = (fe/a'')x", and the line x = a with respect to the xy coordi-
nate system.
11-22. Find the area moment of inertia and the area product of inertia with
respect to the centroidal axes of a standard unequal angle section shown. Find the
principal axes and the principal moments of inertia. Draw Mohr's circle to check
the calculations.
-
CHAFIER 12
Dynamics of Rigid Bodies
12-1. Killer's E<|iiatioiis of Motion of a Hiîd Body

III Chap. 10, we discussed the motion of a system of particles. Since
a rigid body a special case of a system of particles,
is it obeys the follow-
ing e(iuations of motion derived in Chap. 10:
F = mrc (10-2)
Mo = Lo i.e.. V (r, X F,) = V

^^
(r, X w.r.) (10-8)
Mc = Lc i.e.. y (e, X F.) = ^ V (p, X m.9.) (10-1 Ij
The reader is reminded of the fact that the first e(|uation de.scribes the
translation of the ma.ss center, the second e(iuati()n describes the rotation
about an arbitrary fixed point, and the last equation describes the rota-
tion al)out the mass center. Whereas the first ecjuation is always needed
in finding the translatory part of the general motion of a rigid body,
either the second or the third equation may be used to obtain the rotatory
part of the motion of the body. The choice between these latter two is
ba.sed on the following consideration. For a body that moves "freely,"

meaning that no point of the body is permanently fixed, we prefer using
Eq. (10-11). On the other hand, a rigid body that is pivoted (meaning
fixed) at one point is more conveniently described by K(i. (10-8), in which
both the moment and the angular momentum are taken about the pivot;
the advantage of this choice lies in the fact that the unknown external
forces acting on the iôdy at the pivot do not appear in the e(|uation since
the.se forces do not contribute any moment about the pivot. It should
of course be pointed out that Eq. (10-11) retains its general form for
moving points other than the mass center provided these moving points
meet the restrictions specified in Sec. \i)-'.i. In general, we shall restri(^t
ourselves to the u.se of K(\. (10-1 1) as applied to the moving ma.ss center.
f
278 Dynamics of rigid bodies
In order to apply Eq. (10-8) or (10-11) to a rigid body, we shall need

an expression for the angular momentum Lo, taken about the pivot in
the case of a pivoted body, and the relative angular momentum Lc,
taken about the mass center in the case of a freely moving body. It
turns out that both of these angular momenta can be represented by the
same expression, if in each case the pomt about which the angular
momentum is taken is considered as the origin of the coordinates. (This
does not mean that Lo is equal to Lc. On the contrary, it means that Lo
and Lc are in general different because of a difference in the two origins.)
Let us first consider Lc, i.e., the relative angular momentum of a rigid
body about the mass center of the body. We assign a set of body axes
xyz, fixed in the rigid body and therefore moving with it, and let the
origin of this coordinate system coincide at all times with the mass
center C. The position vector of any point of the body relative to C is
therefore
p = .Ti + yj + zh (12-1)
and the angular velocity of the body is denoted by to. The velocity of a
point of the body relative to C is, by Eq. (7-10),
By definition, Lc = 2(ei X wiip,)
For a continuous distribution of mass in a rigid body, the summation in

the above equation becomes an integration. Hence
Lc = /^ (9 X e) dm (12-2)
The integrand of Eq. (12-2) can, with the help of Eq. (3-21), be written
(e X e)
= 9 X (w X e)
= (e •
e)<o - (p • w)?
= {x- -|- 7/2 -f- z-)io — (xcc^ + !jw,^ -t- zco,)q
= [{y- -f 2;-)cox — xyuy — xza),]i + — [ //.rco^ -|- (.r- -f z-)oo,i — yzu^,]j
-{- [
— zxw, — zyoiy + (.1-- + /y-)co,]k
Substituting this evaluation into Eq. (12-2) and applying the definitions
of moments and products of inertia [Eqs. (11-1) and (11-2)],
Lc = {Ixx^x — Ixuiû — h:0:,)i -\- {

— Lix^x + — /,/;W,-)j/-///CO,,/
+ — /--xCOx — Iz!,CO„ LzOO,)^

( -\- (12-3)
t In this soction and in the remaining portion of lliis cliaptor, for simplicity, wc
shall not \isc the asterisk notation to denote moments and ])roducts of inertia taken
with respe(rt to center-of-mass axes uidess such notation is warranted for clarity. The
context of the equation should indicate clearly which origin wc are using.
Sec. 1 2- 1] hauler's equal ions of molioii of a rigid (tody 279
Comparing tlir lij^lit side of this eciuatioii witli the array of i;(i. (1 l-:i), i.e.
we notice iininediately a similarity in form. In fact, if we revert to tlie

notation of indexing l)y nnml)ers (Sees. 11-4 to 11-6), we find that
Eq. (12-3) can be most conveniently and compactly represented by
Lc. = V/.ôj,. ij = 1, •_>. :] (12-4)

J
Or, eiiuivaiently. Lc = 2^ } /,yavi, (12-4a)
Before we proceed any further, let us pause for a moment to consider

the angular momentum Lo body about a particular
of a pivoted rigid
fixed point, namely, the pivot of the body. As before, we assign a set
of body axes xyz, fixed in the rigid body and moving with it. However,
we now let the origin of the coordinates coincide with the pivot. The
position vector of any point of the body is
r = xi + //j + zk
where .r. //, and z are the coordinates of any point relative to the fixed
point. The velocity of any point of the body is in this case due to a pure
rotation about the pivot. Hence
r = o> X r
Substituting these expressions int(j the prcxiously defined form of Lo. i.e.,
Lo = > (r, X ifi.ir,) = /. (r X r) iliii
we find that Lo assumes exactly the .same form as K(\. (12-3). Hence,
tnomentutn of anij rigid body ahoid its ynass center and the
(hr atujular
angular momentum of a pivoted rigid body about its pivot can both he
expressed by Eq. (12-3), if in each case the point abont which the angular
momentum is taken is considered as the origin of the coordinates. In
addition, the angular momentum of any rigid body about .some arbitrary
moving point 0' can also be expre.s.sed by K(\. (12-3), provided 0' fulfills
the restrictions specified in Sec. 10-3. Furthermore, in both cases the
external moment taken about the proper point is etjual to the rate of
change of angular momentum. It is for this reason that we shall hence-
forth u.se the .same treatment for both cases. In particular, we shall
omit the subscripts C and and rewrite Eq. (12-3) as follows:
L = L4 + L,j + L,k (12-5)

where L^ = IxxCjOx — Ixy^y — Ixz^J^z
J-Jy= — -Lyx^x ~r J-yy'J^y — iyz'J^z (12-6)
i^z = — J-zx<^x — J^zy^y zz^z I J-
Since these components of L

are oriented along the axes xyz, which
are fixed in the body and are themselves in motion, we have to apply
Eq. (7-19) to determine L. Hence, using w/ = w,
L = -77 = Lxi + Ly] + Lz^ + (0 X L
and the equation of motion becomes
M = L = (Lx + WyLz — o)zLy)i -\- (Ly -\- w,Lx — w^L^)^

+ (A -f co.L, - co,L,)k (12-7)
This is the general vector equation of motion of a rigid body written for
axes fixed in the body. Its three scalar components are
Mx = Lx -f COyL,
(12-8)
where Lx, Ly, L^ are given by Eqs. (12-6).

These equations can be greatly simplified if we pick the xyz axes to be
the principal axes of the body, for then I^x, lyy, and Izz are the three prin-
cipal moments of inertia and all the products of inertia vanish. Clearly,
such a choice is possible, for principal axes have been shown to exist at
every point of a body and we have the freedom of selecting our coordinate
system to coincide with these axes. As a consequence, then, Eqs. (12-6)
are reduced to
Lx = IxxO^x
Ly = lyyC^y (12-9)
Lz = IzzWz
Substituting these simplified values of the angular-momentum compo-

nents into Eqs. (12-8),
Mx = JxxOOx — {Jyy — I zz)<^yWz

My = LjyOiy " {hz " Ixx)Wz<^X (12-10)
Mz = Izz<^z — {I XX — Iy,,)iOxO}y
These are the equations of motion of a rigid body if the xi/z axes are the
principal axes fixed in the rigid ))od3^ Equations (12-10) are called
Sec. 12-1] Elder's equations of nwlion of a rigid body 281
Philrr's equations. As stated before, tlie origin of coordinates in these
e<iuations c-an be either the moving mass center or, in the case of a
pivoted body, tlie pixot.
In some instances we may choose axes fixed in direction in space with
origin at the fixed point O or the moving mass center C Here the
derivative of the anguhir momentnm with respect to time is revised by
the omission of the term w X L. whicli is recjuired when (hnVrenliating
with respect to a rotating coordinate system. Tlie ecjuation of motion
now becomes
M = L = L,i + LJ + L,k (12-11)
where L is given by Eqs. (12-5) and One serious disadvantage

(Ti-f)).
of this formulation is the fact that now moments of inertia depicted

the
in Eqs. (12-6) are time-dependent. This method is therefore seldom
used.
In some practical situations it becomes extremely convenient to
select coordinate systems which are neither fixed in space nor completely
fixed in the body; however, the origin is still taken at the fixed point or
the moving mass center. A.ssuming that the moving coordinate system
(the moving triad) has angular velocity a>/, the scalar equations of motion
now become, from E(i. (7-19),
Mr = Lx -{ (JifyLi — wfzLy
My = Ly -{ u3f,L, - w/xL, (12-12)
where Lx, Ly, L^ are given by Eris. (12-0) and are referred to the moving
triad. This method is justified only if we can simplify Eqs. (12-12) by
.selecting axes that are permanently principal axes although they may
not all be fixed in the body.Assuming that this may be achieved, and
it may be, provided that at least two of the three principal moments of
inertia are the same (i.e.. the inertia ellipsoid of the body under considera-
tion is an ellipsoid of revolution;, Eqs. (12-12j may now be rewritten
Mi = IxxO^X + IziOlzUify — Iyy(j)y(j)f:
My = I„i/<û + Jxx^xi^fz — IitOitOifx (12-i;i|

Mi = I zzl^Z + IyyU}yO}/x " Ixz(^x<^fu
Once again, the origin of the moving triad may be cither at the fixed
point, the moving mass center, or any other moving point that satisfies
the conditions descril)ed in Sec. 10-3.
Example 12-1. Suppose we solve again Example 10-3, this time i)y
the u.ê of Euler's equations. Let us choo.se a set of principal axes fixed
in the body such that the x axis coincides with the axis of the bar, the
y axis is perpendicular to the bar and lies in the plane of the bar and the
vertical axis of rotation, and the z
axis is perpendicular to the xy plane

(Fig. 12-1). For this problem
(o = oj cos i -f- CO sin 6 j

(b = CO cos i -j- CO sin 6 j
Ixx = lyy = Izz = 2mZ2
Substituting these values into Eqs.

(12-10), we determine the scalar com-
ponents of the restraining moment:
M, =
My = 2ml^6i sin 6
M, = 2'mlw' sin d cos 6
Fig. 12-1 which agrees with our previous result.
12-2. Work-and- Energy Equation of a Rigid Body

From the equations of motion of systems of particles, we have derived
the equations of motion of a rigid body. Similarly, we can derive the
work-and-energy equations for a rigid body on the basis of the following
energy equations given in Sec. 10-3 for systems of particles:
1' (F, + f,) . dr, = 2 }im,v^

^ I
Sec. 12-2] ^^ ork-and-energy eqnalion of a rigid lH)dy 283
Let us treat first a rigid Ixxly in general ami then the special ease of a
pivoted body.
Rigid Body in General. Consider Iî. (12-1.")). The left side, repre-
senting the portion of work done hy the external forces due to the motion
relati\e to the mass center, is
V j' F. . do, = y j'^ F, . 6, dt =

Y^ j^ F. . (a, X p,) dt
= X
^ /,' ^ • Cp. P.) dt
Thela.st step in the reduction accords with Eq. (3-19). Hut p, x F, = M,

is the moment of the external force V, about the mass center. Letting
M = 2M, be the total moment about the ma.ss center, the left side of
Eq. (12-15) becomes, after interchanging the integration and summation
operations,
2
/,"' F, • rfp. = j^ CO • M dl (12-lG)
This .shows that the integrand w is the rate of doing work. • M

To evaluate the right side of Eq. (12-15), which is the kinetic energy
due to the velocities relative to the mass center, we assign the body
axes xyz fixed in the rigid body Avith the origin at the mass center. We
then have, for the continuous distribution of mass in the rigid body,
^ ^2"'tPi' = / .
^
ip- dyn = '
2 / . (w X p) • (to X p) dm
= '-i / [('^)/2 — o),!j)- 4- (coj.r — u}j,z)- + (oixU — oj„.r)-] dm
— 2/j„ajâj„ — 2IyzUi,iOi: — 2/,âj.-aJ,) (12-17)
It is easy to .see that considerable simplicity can again be accomplished

by choosing the xnz axes to be the principal axes. With such a choice of
axes, we can write the kinetic energy due to relative motion:
= M<^-L (12-18)
The last step follows from the expression for L (p]cjs. 12-9). Substitut-
ing E(is. (12-10) and (12-18) into E(|. (12-15), we obtain for a rigid body
f"(.) •
M dt "._„o. iJ^ (12-19)
The total kineti(; energy (denoted li\- 7'), howcxcr, is given by Im|.
(10-25):
T = 2:>2WJ.vr = Vzmvc- + 3 2" • I- (12-20)

:
We can interpret }y^mvc'^ as the kinetic energy of translation and 3'^(o L •
as the kinetic energy of rotation about the mass center. This means
that for a rigid body in general we can write the total work-and-energy
equation (12-14) as follows:
^ f^ F, . dn = t\1= ^mvc' \\
+ 3'^co • L
I'
(12-21)
This equation can be interpreted as the sum of the two equations (10-27)
and (12-19), namely,
2 2
F • drc = yimvc"^ (10-27)
/.^
/,
^
(0 • M d^ = Ho> L •
(12-19)
Of these two equations, the first is the work-and-energy equation for the
translatory part of the motion and the second is the Avork-and-energy
equation for the rotatory part of the motion about the mass center.
If the forces acting on the body are all conservative, the total work
Fi on the left side of Eq. (12-21) becomes

y I
•
dvi
'^jy.'dv, = Vr-V,
where V is the potential energy of the rigid body. Hence Eq. (12-21)
becomes, for the special case of conservative forces,
T -\-V ^ constant (12-22)
where T is given by Eq. (12-20). This is the equation of conservation

of mechanical energy applied to a rigid body. In particular, the poten-
tial energy of a rigid body due to gravitational forces is easy to calculate
^ y^ Fi • dvi = Y,]i~ '^'^ ^^' ^ 2 '"^^^^^" ~ ^'^''
or F = mgijc + constant (12-23)
Hence the potential energy of a rigid body due to gravitational forces is
merely the product of the weight of the body and the elevation of the
mass center above some arbitrarily chosen datum.
Pivoted Body. We now consider the rotation of a pivoted body for
which the pivot is taken as the origin of the principal axes xiiz fixed in
the body. Since the origin is now permanently fixed with respect to an
absolute frame, the appropriate equation to be used in this case is
Eq. (12-14). Following the procedure used in the derivation of Eq.
Sec. 12-2] ^\ ork-and-enenjy equalion of (I rigid body 283
(12-10), we can write the lel't side of E(|. (12-14) as follows:
F. . (/r. - y 1^' F, (CO X r.) dt = V 1^-' ^^ X

2 J," F.) dt
. .
(r,
= j'^ a> • .M dt
Following the procedure used in the derivation of K(|. (12-18), we can

write the right side of E(|. (12-14) as follows:
y yirtiiVi- = \> f (10 X r) • (to X r) dm = '

^^ •
L
Equation (12-14) therefore yields, for a pi\oted body,
j\- Mdt = Vâ)-L|^ (12-24)
which is seen to have the same form as Ef]. (12-19). The difference
between the two ecjuations (12-19) and (12-24) lies in the diiTerence of
the two origins. It can be seen that for a pivoted body [Eq. (12-24)]
using the pivot as the origin is more convenient, since IVI about the pivot
docs not involve the (usually unknown) external forces at the pivot.
For external forces that are conservative, Eq. (12-24) can be written
r -f 7' = constant (12-24a)
where T = kinetic energy = ^^w • L for a pivoted body

V = potential energy
The potential energy V due to the gravitational forces alone is again
given by E({. (12-23).
Since the equation of work and energy for a particle is a result of the
integration of the eciuations of motion, we would expect a similar result
to be forthcoming in the case of a rigid body. To verify this, let us
multiply each of the three of Euler's equations [E(is. (12-10)] by w^, u>y,
and w^, respectively. Adding,
But this can be rewritten
M • (O = -
^^
(/xxCOx- + I,„(û- + Izz^^z"^) = :j ^ (to • I.)
On integration, this yields the efjuation of work and energy of a rigid

body for rotation about the mass center:
12
/2 M -tor// = l.^to- L I
: :
12-3. Example: Motion of an Air-borne Vehicle

To illustrate the application of the previously derived equations of
motion of a rigid body, we consider the problem of the general motion
of an air-borne vehicle —
an airplane or a missile. It is customary in
dealing with such problems to assign a set of body axes fixed in the
vehicle in the following manner (Fig. 12-2)
X = longitudinal axis (of vehicle), positive sense forward

y = transverse axis, perpendicular to x and pointing in direction of right
wing or fin
z = normal axis, perpendicular to x and y
The origin of this set of axes is the mass center C. For missiles with
cruciform fins, the y and z axes may be interchangeable. It might be
added that the x axis as defined
is in general not the direction of
flight.
It is necessary next to determine

the orientation at any instant of
the body axes with respect to a
fixed reference. For this purpose
a set of fixed coordinate axes XYZ
are defined such that the Z axis is
vertically downward (i.e., along
the gravity vector). The X and
Y axes are horizontal ; the X axis
some arbitrarily assigned
points to
direction, say north. The orientation of the vehicle or its body axes
with respect to these fixed axes are then specified by the following three
angles (Fig. 12-3)
6 = attitude
r/' = heading

= bank
The angle 6 is defined as the angle between the x axis and its own projec-
tion on the XY plane, and defined as the angle between the projection
\p is
of the X axis on the XF plane and X axis. Thus d and jointly determine rp
the orientation of the x axis. The third angle represents the amount
of roll or spin about the x axis, defined as the angle between the y axis
and the line of intersection (CD in Fig. 12-3) of the yz plane with the
XY plane. It is somewhat unfortunate that these three customarily
used angles in the analysis of air-borne vehicles are not the Eulorian
angles used in dynamics. For this reason we shall limit the usage of the
Sec. 12-3] l'^.r(uii/)l(': niolion of an (lir-hornc iTliicIf 287
symbols d, 4'. and in the inaiiiicr just (ie(iiu'i.l lor the reinainder of this
section only. Elsewhere in this chapter the same symbols represent
the Eulcrian anj^lcs introduced in Sec. 7-3.
The time rate of change of the thred angles 6. \p, and are represented </>
by the angular- velocity vectors 6, and whose directions, shown in ifr, <{>,
Fig. \'2-'-\. are perpendicular to the i)lancs in which each of the three
A' y plane
yz plane
angles measured. Each angular velocity can be decomposed into

is
three components along the body axes .r, y, and z. Thus
6 = ^ cos <f) j — sin k

x\t - \p sin i -\- 4/ cos d sin j + i^ cos d cos k
4> = 0i
where i, j. k are the unit vectors along .r, //, z. The svuii of 6. t[r. and is
<j>
the total angular velocity to = wî + wj + w,k of the vehicle.! Hence

cox = — sin d t^
(j}y = 6 cos + 1/' cos d sin (12-25)

oi^ = —6 sin + 1/' cos 6 cos
Alternatively, we may solve for 6, yp, and in the above e<iuati(ins in

terms of co,, w„, and w,. The result is
6 — w„ cos — Wj sin
^ = (a)„ sin -f- Wi cos 0) sec 6 (12-20)

<^ = ojx + (w,/ sin + ujj cos 0) tan 6
t In current practice the symbols p, </, r are often used instead of tux, w„, w,.
For the special case of an airplane in level flight, 6 and are both (/>
small. The following is then approximately true: d = Wy, xp = co^,

(}> = co^c.
Equations (12-26) may be integrated numerically to obtain d, and 0; \J/,
hence the orientation of the vehicle, at all instants. In order to do this,

however, we must have information on co^, coy, and co^ and their time
derivatives. This is supplied by the equations of motion.
Before applying the equations of motion, let us first examine the forces
and moments acting on an air-borne vehicle. First there is the gravity
force mg. Since this force has the direction of Z, its three components
along the body axes are (Fig. 12-3)
Along x: —"^W sin 6

Along y: mg cos 6 sin </>
Along 2: mg cos 9 cos
The gravity force acts through the mass center C of the vehicle; hence
its moment about C is zero.
In addition, we also have the propulsive forces and aerodynamic
forces. The propulsive force, or thrust, is due to the action of a jet of
expelled air or gas behind a propeller, a jet engine, or a rocket (Sec. 10-4).
Although the main function of the propulsive force is to move the vehicle
forward, its moment about the mass center may also be used to adjust
the orientation of the vehicle. For convenience, we shall lump the

propulsive force with the aerodynamic forces and represent the total
force by F = Fî + Fyj + F^h and their total moment about the mass
center by M
= Mî + ilfj + M,k.
The equation of motion [Eq. (10-2)] describing the translation of the
mass center now becomes
F + wg = m.rc (12-27)
The mass center re on the right-hand side of Eq.

acceleration of the
(12-27) equal to dyc/dt, where vc is the absolute velocity of the mass
is
center. It is convenient to represent vc by its components along the

directions of the body axes; i.e.,
Vc = Wii + Woj + nsk
Then r'c = -rr = Uii + W2J + w^k + co X vc
[see Eq. (7-19)]. Thus the three scalar equations of Eq. (12-27) describ-
ing the motion of the mass center are
Fx — W(/ sin 6 = m{ui + ii-^cjOy — n-ioiz)
Fy -\- mg cos 6 sin = m{u2 + Uio^z — Ws^x) (12-28)

Fz + mg cos 6 cos = m{u;i -{- UiuO:^ — uiWy)
Sec. 12-3] Example: niollon of an air-horne vehicle 289
Next consider the motion dne to rotiition al)(»ui the mass center.
H(tuations (12-8), with the help of Ecjs. (12-G), give
Mx = I xxOix — {lyy " I !z)^yOi: — /./.(COJ,- — CO:')

+ /j„(wja)^ — —
6),,) /„(ajâ)„ -f oj,)
+ I„:{u)xOiy — CO.) - Ly{oiyi>), 6)x) -\-
Ml = I,,6i, — (/„ — Iyy)Oij,Uiy — I^yiu}^- — Oiy-)

-f f,s(0}y(j}z — Wjc) — lyzioiiOJx + a)„)
The aerodynamic forces and their moments appearing in these ecjua-

tions are functions of a Uirge number of variables. Firstly, they are
functions of the velocity of the vehicle and to a lesser extent the accelera-
tion of the vehicle. Next, they are functions of the angular velocity
and to a lesser extent the angular acceleration of the vehicle. Thej^ also
depend on the altitude Z, \vhi(;h determines the atmospheric conditions,
such as air density and temperature. And lastly, they are functions of
the control-surface (aileron, elevator, rudder, etc.) settings 8u d>, . . .
,
and their derivatives 5i, 82, ... Thus, in the most general form,
.
Fx = FxiUl,U2,U3,Ui,U2,U3,i>}x,OJy,i>}z,<hx,Oiy,0}z,Z, 81,82, . . . ,81,82, . . .)
with analogous expressions for the other force and moment components.
Assuming such functional relationships are known, Eqs. (12-2(j), (12-28),
and (12-29) can then be used to calculate the orientation of the vehicle
at every instant of its flight. The same equations can also be uised to
predict the general behavior and the stability characteristics (i.e., the
response to a disturbance) of the vehicle. Furthermore, these equations
can be used to calculate the location of the vehicle at every instant.
To do this, the following auxiliary ecjuations are needed in relating
the velocity components Ui, U2, Us along the body axes x, y, and z to the
velocity components Ui, U2, U3 along the fixed axes X, )', and Z. Hy
the geometry of Fig. 12-3, it can be seen that
i'l = III cos 6 cos ^ -f ii2(— cos <t>

sin \l/
-^ sin sin d cos i/')
-|- ?/3(sin 4> iîii 'A + cos </> sin 6 cos xj/)
('2 = Wi cos d sin \l/ -f ?f2(cos <(> cos ^ -j- sin sin d sin \f/) (12-30)
+ U3(— sin <f> cos ^ -j- cos sin sin ^)
Us = Wi(— sin 6) + U2 sin cos 6 -{- 113 cos 6 cos
The position vector re of the vehicle at any instant with respect to a fixed
/
reference, say its initial position, can then l;e obtained by integration.
rex
(12-31)
Because of the complexity of the relations involved, especially the aero-

dynamic forces and moments, such calculations are usually performed
on large-scale computing machines.
In practice, the equations of motion (12-28) and (12-29) can often be
simplified, on the basis of the following considerations:
1. Simplification based on the symmetry of mass distribution. If the
mass distribution one plane of symmetry, say the xz
of the vehicle has
plane (this is approximately the case for an airplane), then Ixy = lyz =
and the only remaining product of inertia is Ixz- Equations (12-29) can
thus be simplified considerably. In the case of missiles, it is sometimes
permissible to assume that the mass distribution is symmetrical with
respect to both the xy and the xz planes. Under such conditions, all
three products of inertia vanish. It should be pointed out that such
assumptions of symmetry can be invalidated by the nonsymmetrical
burning of fuel or by the antisymmetrical deflections of wings or fins.
2. Simplification based on a consideration of order of magnitudes. A
quite different type of simplification of the dynamical equations can be
achieved if the angular velocities w^, coy, w^ are so small that their quadratic
products can be neglected in the equations, f Furthermore, the linear
speeds W2 and u^ are often much smaller than Ui. Under such conditions
the six equations of motion can be simplified to
Fx — 'mg sin 6 = mui

Fy -\- mg cos sin = mui -\- Uiw^ (12-28ffl)
Fz + ing cos 6 cos </> = miii — Wicoy
Mx = — Ixx<^x IxyOiy — I zxOiz

My = lyyOiy — I yz^Z " Ixy(^X (12-29o)
Mz = IzzOiz — IzxO^x — lyzOiy
There are, of course, cases where both simplifications 1 and 2 as given
above can be applied, and the equations can be further simplified.
12-4. Stability of the Rotational Motion of a Rigid Body
We have previously seen (Chap. 11) that the principal moments of

inertia of a rigid body at its mass center include the maximum and mini-
mum moments of inertia of the body with respect to all possible axes
through the mass center. One may then wonder whether the relative
magnitudes of these principal moments of inertia have an effect on the
character of the motion of the rigid body. We shall presently see that
they have indeed a pronounced influence; to illustrate this point we
consider here the stability of the rotational motions of a moment-free
t For a spinning missile, cox may be quite large and the quadratic terms involving
Wx may not be negligible in all six equations of motion.
Sec. 12-4] Slabilily of (he rolalional molion of o r'ujid f)ody 291
rigid body. Specifically, we consider a rigid hoily with principal moments

lit" inertia /„, lyy. I :z and an angular velocity to. and we investigate
whether a small disturbance in the ongular velocity to will leail to a
"tumbling" of the body.
For purposes of this di.scu.ssion we shall a.ssume that /„ is the minimum
principal moment of inertia, while /„ is the maximum principal moment
of inertia, so that
/« < /,„v < hz (12-32)
Further, we let Q be the total angular \elocity, .so that
where i, j, and k are unit vectors oriented in the directions of the prin-
cipal inertia axes at the mass center. We also assume that ii is of the
form
il = CO + p (12-.S4)
where to is a specified initial angular velocity, while p = a perturbation

angular velocity due to the action of some disturbance acting on the
body and is assumed such that |p| <K |<o|.
The total angular velocity H must of cour.se sati.sfy Euler's ecjuations
(12-10), for the case of M = 0. Thus
hA - {lyu - hzH^h =
Iyytl„ - (/,, - /„)S2.12. = (12-35)
hzih - (/.. - i„u)iuXj =

The initial angular velocity to must also .satisfy Fcjs. (12-.S5) when Ux,
u3,„ and a'j are substituted for Six, ^y, and 12j, respectively.
We now examine conditions under which p will remain bounded (stable
motion) and under which it will tend to grow and ultimately lead to
tumbling (instability). Substituting for A from Ef]. (12-34) into Eqs.
(12-35) and neglecting higher-order terms in the perturbation velocities
yields
lzx{0)x + Pz) — (/,/,/ — lzz){0}„0}^ -\- W„Pj -1- OJ^p,;) =

hui<^ll + Py) — Uzz — Izz)(oizO}x + 0)zPz + OixPz) = (12-36)
Izzioiz + Pz) — Uzz — I„„){03x0},, + +OixP,, ûPx) =
Piccalling that the initial velocity vectors also satisfied Euler's ecpiations
of the form of E(is. (12-35), E(is. (r2-3r)) then simplify to the following
perturbation e(| nations:
IzxPx - (/,/,/ - Izz)(0)uPz + C^zPu) =

///»//>// - (fzz - Ixx){03,px + WrPz) = (12-37)
IzzPz — (I XX - /,/,/) (WxP,/ + 03„Pz) =
Equations (12-37) may now be used to investigate the stability of

specific motions. To this end we shall consider three particular cases.
Case a. Let coj, = co^ = 0, co^ = constant ; it is easily verified that these
values satisfy Euler's equations. With these values for the initial
angular velocity, Eqs. (12-37) become
IxxPx =
lyvPy - {hz — hx)cOxPz = (12-38)
IzzPz — (I XX — Ivv)(^x'Pv =
The Eqs. (12-38) indicates that px will be constant. Now, to
first of
solve for the remaining components of the disturbance angular velocities,

we eliminate pz by differentiating the second of Eqs. (12-38) and sub-
stituting for Pz from the third of Eqs. (12-38); the resulting equation
for py is
{J^ ixx)\ixx n /10

~.
Vv — .
zz
j—f
J-iiy) 2^
^=^Vy = r>(\\
(12-39)
^ VV^ zz
A similar equation can be obtained in terms of p^ alone; Eq. (12-39) is
sufficient, however, for our purpose. From Eq. (12-39) we immediately

note that if the quantity {hz — Ixx){Ixx — lyy) is positive, the solution
for py will contain a positive exponential and will therefore grow without
limit, while if this quantity is negative, the solution will be a harmonic
which we consider stable.
oscillation
Considering the inequality of Eq. (12-32), we can see that {hz — hx)
> and (7xx — hy) < 0, so that the quantity (hz — hx) {hx — ly,) < 0.
Thus the motion o> = coî can be considered stable since the disturbance p
remains bounded.
Case h. Now we assume an initial motion of the form co^ = co^ = 0,
coy = constant. Proceeding exactly as in Case a, we obtain the perturba-

tion equations
Ixxpx — (lyy — Izz)(^yPz =
lyypy = (12-40)
IzzPz — (I XX — Ljy)0}yPx =
In this case the second of Eqs. (12-40) yields py = constant, while com-
bining the first and third of Eqs. (12-40) leads to
.^ _ (lyy - IzMIxx - lyy) ^ô^^

_ q (j2-41)
I xx-i zz
As Case o, we see that if the quantity (/,,;;/ — ^z^ (J^x — L,,) is posi-
in
tive, the motion will be unstable; i.e., px will tend to "grow" without
bound. l^]xamining again the inequality of Eq. (12-32), we see that
(/,„ - /,,) < 0, (/.. - /,,) < 0. Therefore (/,„ - Lz){Ixx - I.J > 0,
and thus the motion to = co,J is considered unstable.

Sec. 12-.')] lypes of motion of (I rigid body 293
Case
c. If the initial motion is assumed of the form oj, = a)„ = 0,
CO, =
constant, and proceeding in a fashion as in Cases a and b, we find
that pi — constant aiul that
Px j—f oj^'Px = (12-42)
It is readily seen from Eq. (TJ-iVJ) that (/„„ — /„)(/j. — /„) is negative.
Therefore px remains hounded and the motion to = oj^k is considered
stable.
We may therefore draw the following important conclusions concerning
rotational motions of a rigid body: Rotations about the axes correspond-
ing to the maximum and minimum principal moments of inertia are
stable motions, whereas rotation about the axis corresponding to the
intermediate principal moment f)f inertia is an unstable motion.
It should be noted that instability as defined here does not mean that
the disturbance would continue growing without limit. While the
exponential solution would tend to indi(;ate this, we must not forget that
the solution was obtained utilizing the assumption that /> 1^. As p «
tends to grow, this assumption would be violated, and the exact solution
should be determined from the iMiler eciuations without recourse to the
simplification. The important point we wish to make, however, is that
rotation about the axes of the intermediate principal moment of inertia
cannot be maintained without external forces (i.e., some form of control)
and will eventually lead to tumbling. The practical im])lications of this
characteristic are clear; for exami)le, any spinning i)rojectile must be
designed so that the spin axis corresponds to the maximum or minimum
principal moments of inertia and not the intermediate one. Numerous
other applications will be encountered in various fields of engineering.
One further remark should be made; the magnitudes of the dilTerences
of the moments of inertia (/„ — lyy), {I,,,, — Iii), {Izz — In) determine
the rate at which the disturbance will grow for unstable motions or the
fretpiency of the disturbance oscillation for stal)le motions. Iôw-
frequency oscillations will usually lead to relatively large excursions
away from the desired initial motion, (lenerally speaking, then, it is
desirable, for stable motions, to distribute the mass .so that the dilTer-
ences of the piiiicipal UKjments of inertia are relatively large.
12-5. Types of Motion of a Kigid Hotly

Although the efpiations derived in Sees. 12-1 and 12-2 constitute
all the ecjuations that are needed to .solve any problem in rigid-body
dynamics, the most general solution of such problems is still very diffi-
cult, as exemplified by the discussion of Sec. 12-8 on just one of such
problems. Instead of attempting to solve the general problem, we shall,

problems and indicate
in the following, consider several classes of special
methods of approach that prove to be of use. In most of these cases
we shall content ourselves with the determination of the rotational
aspects of the motion. From Eq. (10-2), the translational portion of
the motion can be determined as if the rigid body were a particle con-
centrated at the mass center; thus our previous methods (Chap. 8) are
applicable. This does not at all imply that the rotational and transla-
tional portions of the general motion may be treated independently.
On the contrary, they may be seriously coupled by the presence of
external forces, including constraining forces, which are in general
influenced by both types of motions, thereby materially increasing the
complexity of the solution. We begin (Sees. 12-6 and 12-7) with the
motion of a rigid body under no external forces or moments. For such a
motion a geometrical interpretation by Poinsot is of great value. We
next (Sec. 12-8) consider a special class of rigid body, called symmetrical
body, under the influence of external forces or moments. After these
we consider three more simple types of motions that are of practical
importance: translation (Sec. 12-10), rotation about a fixed axis (Sec.
12-11), and plane motion (Sec. 12-12).
12-6. Rigid Body under No Forces; Poinsot's Representation

The problem we shall consider in this section is the determination of
motionof a rigid body under no external forces, or more precisely under
external forces that do not form a resultant the mass moment about
center. Such a problem may be realized in practice by the devices shown
in Fig. 12-4. Figure 12-4a shows a spinning top with the pivot (i.e.,
point of support) at the mass center of the top; Fig. 12-46 shows a gyro-
scope mounted in gimbals (also called Cardan's suspension'\) so that the
mass center is fixed with respect to the absolute frame and all rotation
is about the mass center. It is usually assumed that the mass of the
gimbals is negligible compared with the mass of the gyroscope itself and
that the friction is negligible. In both devices, then, there are no external
moments about the mass center.
Let us select a coordinate system coinciding with the principal axes
and fixed in the body at the mass center. From Euler's equations
(12-10), we find that, for the case of 71/^ = il/, = M, = 0,
IxxOlx = (//,,,/ — fzz)î,^z
Ill>J<^!l
="
U — ^z /xx)cO,COj (12-43)
i zz<j>z = \i XX J nil) '^^^ II
t A11lH)Uf>;li wo \v(M-c told that Cardan (1501-1575) was "unal)lo to assign a use
for it."
Sec. 12-6] Biyid lunly under no forces; Poinsol's repreaenialiou 20r>
Since M = L. the uhsence of M also niean.s tliat L is constant. ( )ne

conse(iuence of this is [E(is. (12-9)]
L- = Lj- -f L,r -\- L;- = /„-Wx"- + Ii'i!,-^,,' + I zz'^:' = constant (12-11)
We alst) note tliat L possesses a lixed direction in space. Furthermore,

we have, after substituting M = in the work-and-energy e(|uation
(12-19),
0) • L = /„co,2 -\- I^,û},- -H Lzior = con.stant = IT (12-45)
It should he added that Ecjs. (12-43) to (12-45) also apply in principle
to the motion of a rigid body which is pivoted at one point other than
the mass center provided there is no external force, or at least no external
(a)
Fir;. 12-4
moment about the pivot. This situation is, lunvever, rather diflicult
to realize in practice. Unless the pivot is the mass center, as in the cases
described in I'ig. 12-4, we would need at least one external force whose
moment about the pivot always counteracts that due to the gravitational
force.
Before we attempt to obtain an analytical solution to the problem
posed at the beginning of this section, we shall consider an elegant
geometrical representation of this problem due to Poinsot. From
Ec|s. (12-44) and (12-45). we .see that, with respect to the rigid body and
the principal axes x\fz, rigidly attached to the body, the angular-velocity
vector <o in a given body with initially specified motion is no longer
arbitrary. If we let the terminal of w be located at the point ha\ing
coordinates ix,ij,z) so that
fa)r = X ?/ Oi3z = Z (12-4G)

then from Eqs. (12-44) and (12-45),
L.'x^ + lyyhf -f IJz'~ = L2 (12-47)

h.x' + lyyy' + L.Z' = 2T (12-48)
where both L^ and 2T are constants in a given problem. Each of the

above two equations describes an eUipsoid rigidly attached to the body.
In order to satisfy both equations, the terminal of the vector w must
always fall on the curve of intersection of these two ellipsoids. To an
observer fixed in the body and moving with it, the extremity of w describes
this curve of intersection. This curve, representing the locus of the
terminal of o with respect to the body, is called the polhode. The
ellipsoid described by Eq. (12-48) is called the Poinsot ellipsoid. It can
be seen that the Poinsot ellipsoid represented by Eq. (12-48) and the
inertia ellipsoid represented by Eq. (11-28), namely,
differ only by a scale factor.

We shall now show that at the terminal of any given w (hence a given
point on the surface of the Poinsot ellipsoid) the tangent plane to the
Poinsot ellipsoid is normal to the angular-momentum vector L. To see
this, we let the terminal of any given w be (a:;o,^/o,2o). At this point, the
tangent plane f to the Poinsot ellipsoid is represented by
IxxX^X -I- lyylj^y + IzzZiiZ = 2T

This plane is normal to the line
X y z
JxxXo -^yi/yo i zzZQi
But this line is just the line of action of the angular-momentum vector
L corresponding to the given o> [Eqs. (12-9) and (12-46)]. Hence the
statement made at the beginning of this paragraph is proved.
Since L is a constant both in magnitude and in direction when = 0, M
the direction of L can be represented by a fixed line in space, called the
t The equation of the tangent plane to a quadric surface,
F{x,y,z) = axix"^ + 2ai2xy + 2auxz + 022?/^ + 2a23?y2 + a^^z^

+ aux + any + a^z -{- an =
at the point (xo,yo.zo), is given by
dF_
{x - Xo) + T— I (2/ - ?/o) + -T- (z - zo) =
dx dy |o dz
At the same point the equation of the normal to the surface is
X — xo y — yo z — zo
dF\ dF\ dF
dx |o dy \o dx
Sec. 12-6] Rigid body under no forces: Poinsol's represenlalion 297
invariable line (Fig. 12-5). At a distance of u> L/L from tlie mass
•
center C along the invariable we draw a plane normal to the invari-

line,
able line; this distance is constant in a given problem [cf. Eqs. (12-44)
and (12-45), from which the distance' is 2T/L] and represents the projec-
tion of o> on L. The plane so drawn is again fixed in space, and is called
the invariable plane.
By
this construction, the terminal of o> must lie on the invariable plane.
But we have just proved that the plane passing through the terminal
of o> and tangent to the Poinsot ellipsoid is normal to L, the invariable
line. Hence the invariable plane is always tangent to the Poinsot
ellipsoid. At the point of tangency, since it lies on the angular velocity
(0, the Poinsot ellipsoid has zero \elocity. The hodij therefore moves in
nvariable plane
Poinsot ellipsoid
Fk;. 12-5
such a manner that its rigidly attached Poinsot ellipsoid rolls, with C as
pivot, on the invariable plane. The angular-velocity vector at any
instant is represented by the vector from C to the point of contact of the
Poinsot ellipsoid and the invariable plane. The locus of the point of
contact on the invariable plane is called the herpolhode. We thus can
make the following statement: The polhode on the Poinsot ellipsoid rolls
without slipping on the herpolhode in the invariable plane. The so-called
space cone is obtained by joining the fixed point C to the herj)olhode,
while the body cone is formed by joining C to the polhode. f The angular-
velocity vector to therefore traces out the space cone in the ab.solute frame
of reference; it traces out at the same time a body cone with respect to
the body. Poinsot's representation also means that the body cone rolls
on the space cone.
Poinsot's representation of the motion of a force-free (or, more appro-
priately, moment-free) body clearly demonstrates the following impor-
t In general, both the body and space cones need not be riglit circuhir cones.
tant conclusion. Although the angular moment L is constantwhen

M = 0, the angular velocity o> is in general not a constant. Thus the
body, under no external forces, will rotate about a different axis in space
at different times.This phenomenon is called precession, of which more
will be said in the next section.
The only exception is when the
initial rotation happens to be a rotation
about only one of the principal axes. Then o> and L are constant and
have the same direction and the motion will remain a pure rotation
about the same principal axis as if the axis were fixed. The body cone
and the space cone degenerate, in this special case, into a straight line
along (0 or L. Thus, only if motion is initiated by a rotation about a
principal axis will a steady rotation ensue.
12-7. Free Motion of a Symmetrical Gyroscope

For a rigid body having the shape of a body of revolution, called a
symmetrical body, or symmetrical gyroscope, the Poinsot ellipsoid becomes
an ellipsoid of revolution (Sec. 11-8). The polhode becomes a circle
about the axis of symmetry (denoted by Cz) of the body. The herpolhode
also becomes a circle in the invariable plane (Fig. 12-6). Thus both the
space cone and the body cone are right circular cones, with the body cone
rolling on the space cone, according to Poinsot's representation. This
immediately suggests three things:
1. G) is of constant magnitude and makes a constant angle a with L.
2. The axis of symmetry Cz makes a constant angle 6 with L. This

angle 6 is the same as the first Eulerian angle defined in Sec. 7-3.
3. L, fa), and the axis of symmetry Cz always
lie in one plane, which
rotates with L as the axis of rotation. This rotation about L is called

precession, and its angular velocity (denoted by 4>) is called the velocity
of precession. It is seen that 
is the same as defined in Sec. 7-3.
Since the direction of L is fixed in space for a given problem, we let
this direction be the absolute axis CZ. In addition to the (resultant)
angular velocity fa) and the velocity of precession ^, we define another
angular velocity, which is the velocity of spin of the gyroscope with
respect to its own axis Cz. This angular velocity is the same as t[r in
Sec. 7-3. Since co, a, and 6 are all constants, it follows that ^ and ijf are
also of constant magnitude, for the of and must yield sum <{) tjf fa).
Let the moment body about its own axis be denoted

of inertia Izz of the
simply by /, and the equatorial moments of inertia /^x and I,,,/ be denoted
by /i (I XX = I mi for a symmetrical body). We shall now show that the
motion deduced from Poinsot's representation, namely, a motion con-
sisting of a rotation of the body about its own axis of symmetry and a
precession of this axis in such a manner that co, a, d, 0, and \p are all con-
Sec. 12-7] Free niolion of a symnielrical yyroscope 299
stants,is indeed the motion that satisfies Euler's ecjuations (12-4.*}) for a
moment-free, sj'mmetrical body.
To show this, let us choose, in addition to the z axis already assi}»;netl
along the axis of synunetry of the l)ody. the xij axes such as to form a
right-hand system (Fig. 12-()). This coordinate system is fixed in the
body at C and rotates with it. Clearly this situation fulfills all the
conditions necessary to use Euler's ecjuations, for the .r// plane is per-
pendicular to the axis of .-^ymnu'try and thus the .r and ij axes are per-
Herpolhode
Polhode
Space cone
Fic. 12-6
manently principal axes. Kulcr's eciuations (12-43), for the case under
consideration, now become
I\6}x = — (/ — I \)0iy(jiz
/ici,, - (/ - /i)co,co. (12-49)
/w. =
From the last of Eqs. (12-49) we immediately determine that w, and
= /a;, are both con.stants. In view of this the first two of Eqs.
/.. (12-49)
become
/ - /.
=
ojjo;,/
/l
h- I OJjCJz =
Eliminating one of the variables from these, for example, cox,
^y + n'-ooy = (12-50)
where w^ = {
—— j
co^
j
For n^ a positive quantity, Eq. (12-50) represents the equation of simple

harmonic motion whose solution can be written
coy = coo sin (nt -\-

y) (12-51)
where coo =
amplitude
7 =
phase angle
It also follows in a similar fashion that I
cox = Wo cos (nt + 7) (12-52)
Since co^ and coo are constants, we see immediately that u is also con-
stant in magnitude. Moreover, from Eqs. (12-51) and (12-52),
cox" + ooy" = ojo" = constant
Thus the extremity of o> travels around a circular path with center on the
z axis ofsymmetry, the polhode; the body cone is a right circular cone,
and the angular velocity o> precesses about the axis of symmetry with a
constant angular velocity of precession of magnitude equal to
PT- = ^ r "2 per unit time

revolutions ^
27r 2irli
It should be clear that this velocity of precession is with respect to the

body which are themselves rotating with angular speed co^.
axes,
Applying this result to the earth as a moment-free symmetrical rotating
body, we immediately deduce that the axis of rotation of the earth
precesses about the axis of symmetry (the north-south line). This
precession, with respect to the body, describes the polhode, a circle of
diameter approximately equal to only 30 ft. These considerations lead
us to the conclusion that the earth differs only slightly from a homo-
geneous spherical body rotating about a principal axis.
Up to this point our analysis has been with respect to axes fixed in
the body. Information relative to these axes still does not completely
define the motion of the body. In order to accomplish this new aim we
employ a set of coordinate axes XYZ, fixed in space with origin at C.
In terms of the Eulerian angles [Sec. 7-3, Eqs. (7-11 to 7-13)], the angular-
velocity components can be written
Wx = d sin \p — (j) sin 6 cos \p
coy = 6 cos \f/

-\- (j) sin d sin \p (12-53)
Sei'. 12-7] Free nioKoii of a synimelriral gyroscope 301
Sul)stituting Eqs. (12-58) into Killer's o(|iiati()Ms (12-4U). wo obtain a

set of three equations who.se .solution explicitly desrrihes the motion of
the body.
A .solution of these equations, omitting .some of the details, may he
found as follows: From the third of K(is. (12-49) we find that w, is con-
stant. Using this information, one solution of the third of Fa\s. (12-53)
is that 0. xp, and 6 are all constants. Then, substituting the first two
of Eqs. (12-5.3) into the first two of Eqs. (12-49j, we find that both of
Euler's first two equations are satisfied if
-A = - ^~^ CO, (12-54)

« 1
Einally, using Eq. (12-54) and the last of Eqs. (12-53),
Id' I oi.
(/ — /i) cos d I\ COS 6
This is the precessional velocity of the symmetry axis about the fixed
line in space repre.sented by L or the fixed axis CZ.
The magnitude of the angular-velocity vector o can be obtained from
lOqs. (12-53) and is eciual to
and, remembering that ^ = 0, this becomes
CO- = (^2 -t- i/'^ -^ 24>^p cos d (12-5())
Thus the angular-velocit}^ vector is constant in magnitude and can be

.seen (l-ig. 12-()) to be the sum of the two vectors and
<}) Therefore o>
t['.
is coplanar with L and the axis of symmetry.

As the angular velocity processes about the axis CZ it descrilx's on the
invariable plane a circle, the herpolhode. (In the case of the earth, the
diameter of the herpolhode is approximately only 1 in.)
Hence the motion predicted by Poinsot's representation satisfies the
three Euler's e<iuations and represents the true solution of the motion
of a moment-free, symmetrical body if the three (luantities 4>, \p and 6
sati.sfy Ec|. (12-55).
In addition to Eq. (12-55), another relation describing the behavior

of thismotion can be deduced. Let us resolve the angular velocity to
into components along Cz and jierpendicular to Cz. Let the direction
its
perpendicular to Cz at the instant under consideration be Cx. As .seen

from Fig. 12-G,
Wi = to sin {6 — a)
CO, = CO cos {6 — a)
Since x and z are the principal axes, the angular momentum along these
two axes is
Lx = Iioix = /ico sin {d — a)
Lg = loiz = /co cos {d — a)
But the vector L is along the invariable line or the axis CZ. Hence
Lx =L sin 6
L^ =L cos d
/i
Thus tan d tan (6 — a) (12-57)
I
We therefore can distinguish the following two cases.

Case 1. Ii > I. This is the case of a long slender rod. Equation
(12-57) shows that d > 6 — a. The motion is the same as described in
Body cone
Fig. 12-7
Fig. 12-6, in which the precession ^ and the spin about the axis of sym-
metry, t|, are both counterclockwise for an observer looking from the top
down. This motion is called direct, or progressive, precession.
Case 2. I > lx. This is the case of a flattened gyroscope such as
the one shown in Fig. 12-4a. Here 6 < d —
a, and thus to has to be on
the other side of L (Fig. 12-7). By Eq. (12-55) we find that ^ becomes
negative if 4> '•'^ positive. That t[r should be negative is also obvious
from a geometrical consideration since if and must combine vectorially t|r
to form 0). Viewed from the top down, the precession 4> is counterclock-
wise but the spin about the axis of symmetry is clockwise. This
t[r
See. 12-8] Symmelricul (jyroscope under exlernul moments 303
niotion is called rclriH/radc prccamiun. The free motion of the earth is
an example of such a motion.
12-8. Symmetrical Gyroscope under External Moments

In the previous sections, we discussed the free (i.e., force- or moment-
free)motion of a symmetrical gyroscope. We shall now consider the
motion of a symmetrical gyro.scope (or top) under external moments.
Our derivation will again be based on Euler's equations. It is therefore
understood (Sec. 12-1) that the followiiig discussion, unless otherwise
Fk;. 12-8
specified, applies to either of the two cases: (1) the gyroscope has one
point fixed (pivoted) in .space, and the origin is at that point; or (2) the
gyroscope has no fixed point, and the origin is the ma.ss center.
To describe the general motion of a gyroscope, let us assign a set of
ab.solute coordinates XYZ and a set of moving coordinates xijz (Fig. 12-8).
We let the z axis coincide with the axis of symmetry of the gyro.scope.
Furthermore we select the x axis to lie in the meridional plane ZOz,
perpendicular to the z axis. Thus the y axis is horizontal and lies in
the XY plane. This choice of x and // axes is possible because, for a

.symmetrical body, any axis in the plane normal to the axis of symmetry
Oz is a principal axis. Euler's etjuations are therefore applical)le.
However, the coordinate system we have selected is fixed neither in
space nor in the body, although the z axis is fixed in the body. The
appropriate etjuations to u.se are, then, Efis. (\'2-\'.i).
:
We first compute the angular velocity of the movmg triad. With our
choice of axes, the third Eulerian angle \p (Sec. 7-3) is zero, and Eqs.
(7-11) to (7-13) become
co/x = —4> sin 6
o^fy = e (12-58)
CO/2 = ^ cos d
The total angular velocity of the gyroscope differs from this only by an
amount equal to the spin of the gyroscope about its own axis 4^. Thus
cox = — sin <^ 6
w, = d (12-59)
w^ = cos 6
4> -\- xj/
I
Substituting these values of angular speeds into Eqs. (12-13), we obtain
the following three scalar equations of motion
Mx = -Ii{4> sin d -f 2d<j) cos 6) + ie{^ cos 6 ^- yp)
My = Ii(e - 02 COS d sin d) + I4> sin 6 {4> cos + i^)

,.^ ^^.
M, = I
j^
{cj> cose + rP)
The same set of equations of motion can also be derived if, instead of
adopting the moving frames of reference just described, we adopt a set of
body axes xyz which are fixed in the body and moving with it. As
before, the z axis is the axis of symmetry of the body. .4^ the instant
under consideration, we let the x axis lie in the meridional plane ZOz, per-
pendicular to the z axis. This appears to be the same as before, except
that here the x axis assumes such a position only for an instant. The
y axis is again horizontal and lies in the plane at the same instant. XY
With such a position of the body axes, = at the instant under con- i/'
sideration, and Eqs. (7-11) to (7-13) yield expressions for Wx, c^y, coz which
are identical to Eqs. (12-59). In order to apply Euler's equations [Eqs.
(12-10)], we also have to evaluate Wx, 03 y, and co,. They are obtained by
differentiating Eqs. (7-11) to (7-13) with respect to time before setting
1/' = 0. (Since xp = only instantaneously yp is not necessarily zero.)
Thus
u^ = d\p — — {(!> sii^ ^)
6:y = d -\- (j)\p sin 6
«2 = ^ (<^ cos d + xp)
Now substituting Wx, co,,, w.-, Wx, co,;, and co, into Eqs. (12-10), the same set
of equations, Ecjs. obtained again. These are the general
(12-59), is
equations of motion of a symmetrical body. (5ivon the moment com-

Sec. 12-8] Symmetrical gyroscope under esterniil inoineiils MV
poiuMits Mi, My, M: and the initial conditions, we should l)e able, at U'ast
in principle, to solve for the three unknowns 4>, and ^ as functions of
6,
time. Today, with the advent of high-speed computing devices, the

recjuirednumerical solution usually 'can he obtained rcti;arcness of the
complexity of the governing eijuations. Still, an analytical solution is
most desirable; general rigid-body motion is sufficiently complicated so
that a numerical solution may not reveal all features of the motion. The
analytical solution is, however, (juite complicated, even with the absence
of one or two of the moment components. We see that our previous
solution in Sec. 12-7 is for the special case when all three moment com-
ponents are zero. We shall consider now two more special cases.
Special Case 1. The motion for
Steadij Precession of a Gyroscope.
this ca.se is body and consists of a spin
similar to that for a moment-free
of constant about Oz and a steady precession about OZ. By steady
yj/
precession we mean that the velocity of precession 4> and the half-cone
angle ^ are bothcon.stant. With such a motion, Eqs. (r2-()0) are reduced to
il/. =
M^ = [I\P + (/ - Ii) (j) cos e\ 4> sin (iLMUj
il/. =
Thus a steady precession is possible if an external moment .1/,, ol the
magnitude given by the second equation above is supplied. If .1/,, = 0,
the .same etjuation shows a relation
between ^, <^, and 6. This relation
is just Eq. (12-55) of Sec. 12-7.
For 6 = 90°. the axis of symmetry

of the gyroscope is normal to the
axis of precession (Fig. 12-0). The
moment nece.s.sary to maintain this
motion is, from Eqs. (12-61),
M = My = 1x1^4, (12-02)
This result seems contrary to our

intuition; for Eq. (12-62) tells us that
a moment applied about the y axis ^
is re(|uired to produce a rotation or

[)recession about the Z axis; i.e., the
plane of the applied couple is per-
p<'ndicul;ir to the plane of precession, The scalar moment —M is
commonly known as the gyroscopic moment. In practice, M is the

moment that must be supplied through the bearings to the shaft of the
gyroscope. This means that the gyroscope will exert a moment —M on
the bearings dining such a motion. If the retjuired .1/ is not supplied
to the shaft, the gyroscope will rise or fall to a new 6. This characteristic
of a gyroscope is the basis of many engineering devices, such as the gyro-
scopic compass, the gyroscopic stabilizer for ships and torpedoes, the
artificial horizon for airplanes, etc. It is also the cause of undesirable
shaft and bearing stresses in ship turbines, airplane jet turbines, airplane
propellers, etc., when the ship or the airplane rolls, turns, or pitches
(Probs. 12-6 and 12-7).
In the case of very large spin velocity i/- compared with <j) and 6, Eqs.
(12-61) can be approximated by
My ~ I^4> sin e (12-63)
This equation is sometimes derived in the elementary theory of gyro-

scopes by considering the angular-momentum vector as along the axis
of symmetryof the gyroscope and having the magnitude Is. The rate
of change of such an angular-momentum vector then gives directly
Eq. (12-63). Such a treatment, however, is not always satisfactory,
since we know from previous discussions that the angular-momentum
vector L, the angular-velocity vector and the axis of symmetry are in
i,i,
general along three different directions. To show that Eq. (12-63) can
sometimes be in gross error, we merely have to recall that in the free
motion of a gyroscope, 0, and d are all nonzero but the moment is zero,
i/-,
whereas Eq. (12-63) would erroneously show a nonzero My. For d = 90°,
Eq. (12-63) becomes the exact equation (12-62).
Special Case 2. Steady Precession Due to Gravity of a Top Pivoted at
One End (Fig. 12-10). Let the mass center C be located at distance a
along Oz from 0. The moment about due to the weight of the top is
M = ak X — mgK.) = mga
( sin 6 j
Comparing this equation with Eqs. (12-61), we find that a steady preces-
sion is possible under gravity alone if
mga = [Ixp + (/ - Ii)(j) cos 6]^ (12-64)
For a given top, this equation relates the three quantities i/', (j), and d.
Equation (12-64) is valid only when the motion is a steady precession.

U d = 90°, Eq. (12-64) reduces to
I\}/(i) = mga (12-65)
We see that the gyroscopic moment is in this case balanced by the moment
due to the weight of the top. Furthermore, we see from Eq. (12-64) that
the larger the spin ;/-, the smaller the precession For very large i/, and
.
hence very small 4>, Eq. (12-64) can be approximated by Eq. (12-65)
even for 6 different from 90°.
Sec. 12-0] The gyrocompass 307
As mentioned previously, the general motion of the symmetrical gyro-

scope extremely complicated. We shall not pursue this sui)ject any
is
further except to mention a few other important phenomena. Aside

from the precession of the top, whicli may not in general he a steady
precession, there is motion known as nutation. By nuta-
also a type of
tion we mean the bobbing of the top as evidenced by the rate of charge
of the angle d. Nutation may occur with and without precession. In
addition, the stability of a spinning gyroscope or top is another interesting
prolilem. (lenerally we find that these bodies are stable for a sulliciently
large value of spin velocity. K\idence of this fact is of cour.^ê familiar;
Fio. 12-10
consider the spinning projectile, the rolling hoop, the bicycle, etc. Many
of these phenomena may be deduced from the general efjuations of
motion (12-()0) used in conjunction with the work-energy principles.
12-9. The Gyrocompass

( )ne of the more useful applications of the preceding theory is a device
called a gyrocompass, which is a gyroscope (a body with an axis of rota-
tional symmetry mounted in a Cardan's suspension) that utilizes the
earth's rotation to determine true north. In order to discuss the perti-
nent features of this device, we shall limit ourselves to an idealized and
simplified version. An actual g^'rocompass much more
is, of course, a
complex apparatus.
Consider the mass center of the gyroscope mounted at some point C
on the earth's surface and defined by the latitude X. Further suppose
that the axis of the gyroscope is constrained to remain in the horizontal

plane (the tangent plane at C). This may
be accomplished by suitable
restraint of one of the bearings of the Cardan's suspension. We take
as our coordinate system a right-hand triad oriented so that the z axis
is directed along the axis of the gyroscope, the x axis is perpendicular to
the z axis and lies in the horizontal plane, and the y axis is perpendicular
Tangent (horizontal)
plane
Fig. 12-11
to the horizontal, or xz, plane. We also take a Z axis parallel to the

axis of the earth's rotation. Thus the x and z axes lie in the horizontal
plane and the y and Z axes lie in the meridional plane OCN. The xyz
coordinate system is fixed neither in space nor in the body (Fig. 12-11).
Denote the angular velocities of the earth, the xyz triad, and the
gyroscope by G>e, w/, and to, respectively, and the angle between the
meridional plane and the z axis by 6. Then the angular velocity of the
triad is composed of the angular velocity of the earth and angular velocity
S(>c. I2-^)J The yyroconi/Hiss 309
tlue tt) the c-!iaiigo ol' the aiiji;li' 0. Thus

ti)/ = a'fk + ^j
Since
K = — cos X sin 6 i + sin X j + cos X cos 6 k

we have
CO/ = — cof COS X sin 6 i + (co, sin X + ^)j + co^ cos X cos d k (12-06)
The anguhir velocity of the gyroscope differs from this only because of
the spin about its own axis z, so that
xj/
CO = — oje cos X sin & i + (cog sin X + ^)j + (w^ cos X cos d -\- 4')k (12-07)
The only external moment due to the constraint supplied by the bearings
is about the x axis; thus M = Mi. Substituting these values of co/, co.
and M into the last two motion (12-13), and letting

of the equations of
/ = /j^ and 1 = I XX = I„u be the axial and transverse moments of
1
inertia, respectively, the scalar equations of motion in the // and z direc-

tions become
1 id — Ii(j)J cos- X sin 6 cos 6 + /co^ cos X sin 6 (cog cos X cos d -\-
\l/)
=0
(12-()8)
1(4' - coe cos X sin 6 6) = (12-09)
Assuming now that the spin is large compared with w,. and neglecting i/'
terms of the order of co^- when compared with terms of the order of co.i/',
Eq. (12-68) becomes

Ii6 + /oje^ cos X sin 9 = (12-70)
A solution of Eq. (12-09) is given by
\p + CO,, COS X COS 6 = constant
This shows that for ip coc, the spin » ip is approximately constant. Equa-
tion (12-70) can now be rewritten
6 + /(•-
sin 6 = (12-71)
where n^ = — , — = constant
Since Kq. (12-71) is the ecjuation of motion of a siini)le pendulum

[Eqs. (8-00)], we find that the gyroscope oscillates in the horizontal
plane about the neutral position 6 = 0. The neutral position, then,
defines the direction of true north.
Suppose now that the axis of the gyroscope were confined to the
meridional plane OCN. We select a coordinate system such that the
2 axis is along the axis of the gyroscope and the x and y axes lie in the hori-
zontal plane and the meridional plane, respectively, and we let be the
angle between the axis of the gyroscope and the vertical axis Z.
V
Proceeding in a similar fashion, we would find that the gyroscope

oscillates in the meridional planeabout the vertical axis Z as a neutral
position. Thus the latitude is determined to be the angle between this
neutral position and the horizontal. The details of this analysis are
omitted here and should be completed by the student (Prob. 12-10).
12-10. Rigid Body in Translation

By pure translation of a rigid body we mean that the motion of the
body such that if lines were initially scribed on the body, they would
is
always remain parallel to themselves during any subsequent motion.

Pure translation may be subdivided into two main categories, rectilinear
translation, where each point on the body travels a straight-line path,
and curvilinear translation, where each point on the body travels along
some curved path in space. The motion of a rigid body in pure transla-
tion is described by
F = mrc (10-2)
The angular velocity is zero in a pure translation. This means
COi = COj; = CO2 =

and Eq. (12-4) then gives he = 0. Therefore Eq. (10-11) requires that
there be no resultant external moment about the mass center; i.e.,
Mc = (12-72)
We can also write these two vector equations (10-2) and (12-72) in terms
of their cartesian components:
F^ = mxc Fy = mijc F^ = mzc n9 "q^
^^^"'"^^
Mc. = Mc, = Mc^ =
These six scalar equations completely determine the motion of the mass
center of the body and the external forces required for the prevention
of rotation of the body.
Since a rigid body in translation does not possess a pivot point, Eq.
(12-72) is valid in general only with respect to the mass center. Thus L
and M are not zero with respect to any fixed point. They are zero only with
respect to the mass center of the body.
It is of some interest (mainly academic) to examine the form of the
moment equations when they are taken with respect to points other than
the mass center. For some fixed point 0, and for some arbitrary moving
point 0', these equations take the following form:
Mo = j^
r X r dm (10-8)
Mo' = ^ / e X 9 dm - r X mgc (10-10)

at I
Sec. 12-11] Botalion ahoul a fired ajis; dynaniic Unkuic'uuj '.\\ I
Here we have iutroduced the iiitej^ration over the entire vohiine ot the
bodj' for a continuous mass cHstriWutiou.
Equation (10-8) becomes, upon performing the iniUcated (hlTcrentia-
tion. using Va\. (o-O) and recognizing that in translation the acceleration
of every point on the l)t)dy is the same.
Mo = mrc X r = mrc X re (12-74)
In a similar fashion, Va\. (10-10) becomes

iVIo' = —fu'r X pc = —mrc X pc (12-75)
This reduction is possible since, for a rigid body in translation, p = 0.

Both Eqs. (12-74) and (12-75) graphically illustrate again that the
moment eciuations in general vanish only when taken with respect to the
mass center. The scalar form of Eqs. (12-74) and (12-75) can be easily
written, and special cases deduced therefrom. We leave this as an
exercise for the reader (Probs. 12-12 and 12-13).
12-11. Rotation about a Fixed Axis; Dynamic Balancing
A type of problem often met in practice is the rotation of a rigid body,

not necessarily symmetrical, about a fixed axis. Such a problem is an
especially simple case of the more general
problem of rotation of abody about a fixed
point (i.e., the pivot). When a body rotates
about a fixed axis, every point on the axis of
the body is a fixed point.
Considering ani) point on the fixed axis as
the origin of a coordinate system fixed in the
body and letting the z axis be the axis of
rotation (I'îg. 12-12), we immediately have
the simplification that
CO:, = OJ,, =
But the z axis thus assigned may not be the j.^,, y^.vi
principal axis. We therefore cannot use the
Euler's ecjuations (12-10). In.stead, we have to start with the more
general treatment [Eqs. (12-6)], which becomes
Lz = — hzo:
L,j = —lyzO}
Lz = /«w
Substituting these components of the angular momentum into Ecjs. (12-8),
My = -hz^-' - /„,ci (12-76)

M, = /z.W
The third equation completely determines the rotation. The first two
equations determine the external moment required to keep the axis (i.e.,
the shaft) fixed. These external moments are usually supplied by bear-
ing reactions, i.e., forces exerted by the bearings to the shaft.
In the elementary treatment of rigid-body dynamics, it is often assumed
that the body is symmetrical with respect to either the xy plane or the
axis of rotation or both. When the body is so restricted,
I- yz -t xz O
(Sec. 11-8) and the angular-momentum vector L coincides with the direc-
tion of fa). Hence
M, =
Aly = (12-77)
These equations apply, for example, to a flywheel, which is symmetrical

about the shaft. But they also apply to a body of whatever cross section
if the body is symmetrical with respect to a plane normal to its axis of
rotation, since the products of inertia also vanish for such a case if the
origin chosen to lie in the plane of symmetry.
is
In addition to the moment equations (12-76), the equation of motion

(10-2) for the mass center, i.e.,
F = mrc
is, of course, also valid. This equation becomes, for the case of rotation
about a fixed axis,
Fa: = mxc = m{ — uyc — co^xc)

Fy = myc = m(ooxc — oi'^yc) (12-78)
F. =
where xc and yc are the components of the distance from the axis of rota-
tion to the mass center.
The procedure in applying Eqs. (12-76) and (12-78) to solve a problem
is as follows: We first use the third of Eqs. (12-76) to determine co. The
first two of Eqs. (12-76) and the first two of Eqs. (12-78) are then used to
find the four bearing reactions, i.e., the reactions along x and y for each of
the two bearings.
In some cases, depending on the body, it may prove to be advantageous
to select a fixed coordinate system with origin on the axis of rotation and
oriented in the directions of the principal axes. Under these conditions
Euler's equations may be used directly.
When the mass center of a rotor on the axis of rotation, the rotor
falls
is said to be statically balanced. For a statically balanced rotor, the
Sec. 12-11] Rolalion ahoul a fixed art's: dynamic halancing 3i:i
sum of the l)earing reactions must he zero on account of lujs. (12-78).t

However, this docs not mean that the reaction of each in(li\idual hearinjj;
is zero.
Unless the products of inertia are also zero, a statically balanced rotor
still recjuires tlie external couples M^ and My from the hearing reactions
as shown by Eqs. (12-7G). The resulting bearing reactions vary periodi-
cally in direction, the period being the time of one revolution of the body.
As we have seen in Chap. 9. this .setsup a forced vibration of the support-
ing structure. To avoid this, the rotor must be dynamicaUy balanced.
Dynamic balancing recjuires not only static balancing but al.so other bal-
ancing weights so situated that /x^ = /^^ = in addition to xc = yc = 0.
When a rotor is dynamically balanced, its bearings will experience no

dynamic loads. This permits quiet operation and less wear. If a rigid
body were dynamically balanced and rotating about a line fixed in
frictionless bearings and if there were no applied moments about the
axis of rotation, upon removal of the bearings the body would continue
to rotate about the same axis. The criteria for this to occur are that
the axis of rotation contain the center of mass and be a principal axis.
The ecjuation of work and energy (12-24) yields, in the case of rotation
about a fixed axis,
M.i^dt = i.^/..a;2p (12-79)

/.
The work (if any) done by the gravitational forces belongs to the left
side of Ef). (12-79); this portion of the work is equal to mg{yc\ — ya)
[Eq. (12-23)].
Example 12-2. Given a thin rectangular plate rotating about a
diagonal axis with angular speed of w and angular acceleration of co.
We desire to determine the couple exerted on the bearing supports.

Assigning a coordinate system fixed with respect to the plate with
origin at the center ofmass as shown in Fig. 12-13, we find that the various
moments of inertia are
_ 2m a^b^
L. " ~ T a2 -j-b^
, m , a"^ — b"^
3 a/ -j- b-
ly^ =
Substituting the.se values into Ecjs. (12-70), we immediately determine
the following recjuired couples:
t We are concerned here with dynamic loads only. The .static load (due to the
weight of the rotor, etc.) on the bearings is obtained by statics (Chap. 4) and is to be
added to the dynamic loads.
6 a'- -\- h~
'\n a'
My = co^-ab
3 a'
^
+ , ,
0- .,
,, 2w . a-6-
3 a^ + O'^
If instead of using this approach we desire to use Euler's equations,

we select the primed coordinate shown in Fig. 12-13 coinciding with the
Fig. 12-13
principal axes. For this coordinate system,
(ai' -\- 6k')
CO
(ai' -F 6k')
Va' + b^
and
^;. - -3 («^ + ^>^)
Substituting into Eqs. (12-10), we have the following required couples:
^^, ^ m wab~
3 Va' + t'
?w a6(a- — 6^)
il/.' = - ^''" 2
fl^ + 6-^
a'b
M' =
3 V«' + ^'^
Comparing the sets of results for the different axes, we immediately

note agreement. Vox a symmetrical rotating body, there is a slight
advantage in the use of l*]ulei-'s (Miuation, since the necessity of computing
Sec. 12-11] Boladon about a fixed (i.ris: dyiiuniic iHildncitKj ^1.")
the nioiiUMits ami products of inertia for inclined axes is circumvented.

On the other hand, our jiractical interest is likely to he centered on .1/,,
My, and M- rather than .1/^. .1/;^. and M[. We also note that in the
absence of any anjûlar acceleration a couple is still iêiiuired to maintain
the motion. This couple always acts in the plane of the plate; it rotates
with the plate, thereby producinjj; alternating!; reactions at the bearings.
Example 12-3. Show that the dynamic balancinj^ of a rotor can
always be achiexcd by adtling two correction weights in two arbitrarily
assigned planes, called correction planes, normal to the axis of rotation.
Fig. 12-14
Let the mass of an unbalanced rotor be m. The rotor is statically

unbalanced if one or both of its xc, yc are nonzero. It is dynamically
unbalanced if it is statically unbalanced or if one or both of its products
of inertia Ix^, lyz are nonzero. Let the two correction planes A and B
be located at 2 = Za and z = zn, respectively (l-'ig. 12-14). Let mA be
the ma.ss of the correcting weight in plane .1 , and let mu be the corre-
sponding (piantity in plane B. The magnitudes and their loca-
/«,i, ?««
tions (.r,i,//,i), i.rii,i//i) are determined by the following relations, which
ensure that, for the entire system including the correcting weights, the
center of mass lies at and that the z axis is a principal axis:
niAXA + m,fXi{ = —nixc

mAl/A + miiijn = —mi/c
itiaXaZa + uihXhZii = — /x.
mAilAZA + nijiiniZii = —ûz
These four equations enable us to solve for the four unknowns niAXA,
inAVA, tnuXB, and mBi/B. It can be seen that with the correction weights
and their locations so determined, the total assembly consisting of the
rotor and the correction weights is dynamically balanced. Furthermore,
since only the four products niAXA, etc., are significant, we can even
arbitrarily assign values for rriA and m^ or, alternatively, their distances
from the axis of the rotating shaft (Prob. 12-21).
12-12. Plane Motion of a Rigid Body

The three-dimensional motion of a rigid body, even for the special
casesshown in Sees. 12-6 to 12-9, is more or less complicated. Fortu-
nately, many engineering problems are concerned only with the motion
in a plane. This simplification allows a great variety of problems to be
solved. Plane motion of a body occurs when the body is constrained to
move parallel to some reference plane.
The equations of motion are again given by Eqs. (10-2) and (12-8).
We let the motion be in the xy plane so that the z axis is perpendicular
to the motion. body may not have a fixed point, we let the
Since the
origin of our coordinate system be fixed at the mass center by previous
agreement (Sec. 12-1). The z axis so assigned, of course, may not be the
principal axis, but it is a nonrotating axis; it always remains normal to
the plane of the motion, the xy plane. Since oox = oiy = z = for plane
motions, Eqs. (12-8) give, with the help of Eqs. (12-6),
Mx = Iyz<^^ — Jxz<J^
My = -Ixz<^~ - Iy.<^ (12-80)

M, = /,,«
where co = w^ and cb = (b^. We also have Eqs. (10-2), which yield
Fx = mxc
Fy = myc (12-81)
F, =
We find that the equations of motion (12-80) and (12-81) are of the
same form as Eqs. (12-76) and (12-78). The only difference is that the
moment equations (12-80), for plane motion, are taken with respect to
the mass center as origin, whereas the moment equations (12-76), for
rotation about a fixed axis, are taken with respect to a point on the axis
of rotation as origin.
The procedure in applying Eqs. (12-80) and (12-81) to solve a problem
of plane motion is not as clear-cut as the procedure for solving problems
of rotation about a fixed axis presented in the previous section. This is
because the unknown reactions do often contribute moment about the
.
Sec. 12-12] Plane motion of a rigid body 317
mass center, and therefore the third of Ecjs. (12-80) alone is often not
sntlicient to solve for w. In general, we have to solve the simultaneous
efiuations consisting of all of Eqs. (12-80) and the first two of Ecjs. (12-81 ).
If there a kinematic relation in the problem relating .re, j/c, oiz< and w.-.
is
that relation will certainly be necessary in solving the problem. The

student should therefore always be cognizant of such possible kinematic
relations in a given problem.
In the elementary theory of dynamics, it is often assumed that the
xy plane is a plane of symmetry of the body. Then /x.- = /„.- = 0. The
products of inertia also vanish if the z axis is an axis of symmetry of the
body. When the body is so restricted, the pertinent ecjuations of motion
are reduced to only three in number:
Fx = rnxc
Fy = myc (12-82)
The equation of work and energy for a body in plane motion is often
also of interest. The total kinetic energy of the body is, by K(\. (12-20),
T = Yimvc' -F >2/..'^' (12-83)
This equation can be used in connection with Ya\. (12-21) or, for conserva-
tive external forces, with PJq. (12-22).
Example 12-4. A nonhomogeneous
circular cylinder with its mass center C at
a distance a from its geometrical center .1
(Fig. 12-15) is initially at rest with AC

horizontal (i.e., =0). It is then allowed
to roll without slipping on a rough hori-
zontal plane. At any subsequent position
of 0, find (a) the angular velocity w; (/>)
the angular acceleration w and the reac-

Fk;. 12-1.-)
tions at the point of contact B.
(a) At the initial position = 0, the kinetic energy T is zero.
</> We
also assume that the potential energy V = 0.
At any subsequent the conservation of energy reciuircs
<f>,
tliat
r + ]' = constant = 0. Now

V = -mga .sin <f>
(.seeEq. (12-2:^)]
T = \imvc- + 3^/«c.2 [see Ec). (12-8:^)1
/„ in the preceding expre.s.sifin is of course taken about the z axis passing

through V
The velocity of C is given l)y Example 7-1, from which wc get
V(P- = w-[{U — a sin 0)'- -\- a- cos- 0] =- (ji-Cli-

This result is to be expected, since, as far as velocity is concerned, the

motion is equivalent to a pure rotation about the instantaneous center B,
atwhich point the velocity is zero. Substituting fc^ into the expression
for T,
T = Kco2(/3, + mCB'-) = yils^-
where Ib = Izz + mCB- is the moment of inertia of the cylinder about

an axis passing through B and parallel to the axis of the cylinder. The
kinetic energy is seen to be that due to a pure rotation about B. From
T -{-V = 0, we obtain
2mga sin </>
(6) The forces acting on the body consist of the weight force mg
acting at C and of the normal force N
and the friction force F acting at B
(Fig. 12-16). This friction force is not
a dissipative type of friction force; i.e.,
it is not a sliding friction force, for point
B is the instantaneous center and has

zero On the contrary, it is
velocity.
more nature of a constraining
of the
force, a force required to maintain the
specified type of motion. The acceler-
ation of C in terms of the angular
velocity and angular acceleration can be
N obtained from Example 7-1.
Fig. 12-16
Using the appropriate forces (Fig.
12-16)and accelerations (Example 7-1), the equations of motionf (12-82)
become
Fx = mxc F = m(coR — coa sin — co-a cos </>)
Fy = myc N — mg = m{ — ua cos + co-a sin </>) (12-84)

M^ = I,M Na cos — F{R — a sin 0) = /,,co
4>
Substituting for F and N in the third equation their values as deter-

mined by the first two equations, we obtain an equation consisting of
only one unknown, co. After a few algebraic steps which are somewhat
lengthy, the result is
w = ma cos
g + CO-/?
In
t We take the nonlioinoRonoous eyliiuler to be synunetriral with respect to the

xy plane.
Problans 319
iSubstituting the \alue of a;- obtained from part a,
w = mga cos , I li + 'ImRa

j—
.sin <6
^
In
This result can also be obtained by ditTorontiatinfj; the expression of a;
in part a with respect to time. (It will be left as an exerci.se. I'rob.
12-25.) After we have found co and co, the unknown reactions .V and /•'
are given by the first two of Eqs. (12-84). The ratio F/N must be less
than the coefficient of static friction if our ori<j;inal as.sumption of rolling
without slipi)ing is to hold. If our original as.sumption proves to be
invalid, then a new solution must be obtained for a sliding cylinder. In
this event the friction force is a kinetic friction force; it has direction
opposing motion and magnitude ecjual to aia-.V, where ma- i^^ thecoeflicient
of kinetic friction; and the kinematic relationships pertaining to rolling
without slipping are no longer applicable.
PKOBLEMS
12-1. A
square plate is fixed at its center. At a certain in.stant, the anmilar-
velocity vector makes an angle of 30° with the normal to tiie plate. Find tiie angle
between the angular-momentum vector and the normal.
12-2. A circular disk is rigidly attached to a rotating shaft. The axis of .symmetry
of the disk makes an angle of '.M)° with the axis of the shaft. The disk is 1 ft in diame-
ter and ;J in. in thickness and weighs 100 lb. The shaft rotates at a uniform speed of
1,200 rpm in the direction shown.
(a) Find the orientation and magnitude of the angular-momentum vector.
(6) From the rate of change of the angular-momentum vector, find the moment
supplied bj' bearing react ioii.s.
(c) Find the kinetic energj- of the disk.
Fk;. V \1-1
l2-.'{. .\ body in the shape of a s(iu:irc prism of sides n, a, 'In is j)i voted at its center.
.\t a certain instant the body .spins with angular velocity u iibout ;i diagonal. There
are no external forces acting on the body.
(a) Find the time refpiired for the longitudinal axis of .symmetry of the body to
make one complete movement arfiund the surface of a cone.
(/>) Find the half-<'one angle.
(c) Find the kinetic energy of the body if its ma.ss is m.
Fig. P 12-3
12-4. Express the kinetic energy of a symmetrical body rotating about a fixed
point in terms of the Eulerian angular velocities.
12-5. A homogeneous body having the shape of a solid of revolution moves about
a point. Its motion is represented by the uniform roUing of a cone of half-cone angle a
fixed in the body upon an equal cone fixed in space, the axis of the former being the axis
of the body. Show that the couple necessary to maintain the motion is of magnitude
}i co2 tan a[I + (I - /i) cos 2a]
where w is the resultant angular speed, and that the couple lies in the plane of the
axes of the cones.
12-6. A 2,000-lb auxiliary Laval turbine rotor mounted on a destroyer rotates at
10,000 rpm. The rotor withblades can be assumed, for the purpose of calculating
its
moment of inertia, equivalent to a uniform circular cylinder of 10-in. radius. The
Fig. P 12-6
destroyer rolls inrough weather with an amplitude of 40° from the vertical, and it
takes a period of 10 sec to complete one cycle of roll. The distance between the two
bearings of the rotor shaft is 1 ft. Find the maximum reactions at the bearings.
12-7. The propeller of an airplane rotates at 1,200 rpm. It weighs 600 lb and has
a radius of gyration of 3 ft. If the airplane is traveling at 400 mph in a horizontal
circle of radius 3,000 ft, find the gyroscopic moment on the propeller shaft bearings.
12-8. Prove that a top cannot exhibit steady precession with its axis of symmetry
inclined at 6 with the vertical unless it has a sufliciently large angular velocity about
its axis of symmetry such that
w, > 4 -T^ 7)i(ja cos 9

Pruhlenii' \V1\
12-9. A unifonn .s(iuaro plate rotates about its vertical diagonal IH) witli the
angular velocity w. The corner A is then suddenly fixed so that the plate now rotates
about O'O'. What is the new angular velocity?
Fig. P 12-9 i"i(i. r 12-11
12-10. Indicate how a gyrocompass can be used to defeniiine latitude.

12-11. A disk 2 ft in diameter rolls on a circular track of radius 10 ft. The center
of the disk has a speed of 20 fps. The disk weighs 041 II). Find the tension T an<l
the horizontal pressure P.
12-12. Write out the cartesian scalar form of the equations of motion represented
by Eqs. (12-74) and (12-75).
12-13. Write out the tangent normal scalar form of the (•(niations of motion rei)re-
sented by Eqs. (12-74) and (12-75).
12-14. A sliding garage door is i)ushed by a force /'' at the bottom edge. It is
as.sumed that the door is of \miform material and that there is no friction. Find the
value of F for which one of the wheels lifts off. (Which wheel?)
(^
HDk I I B
Fu;. P 12-14 Fig. P 12-15

12-15. two uniform bars weighing 322 lb and 64.4 lb each. They
AB and CD are
are connected C and are hinged to the carriage at A and D. The weight
by a pin at
E of 32.2 lb is attached by a string to the bar AB at B. If the entire system moves to
the left with an acceleration of magnitude 20 ft/sec^, find the horizontal and vertical
components of the forces acting on each bar Sit A, C, and D.
12-16. A uniform stick is pivoted at one end and supported at a fairly small angle a.
To its end are attached two very Hght cups, the outer one of which contains a marble.
When the bar is allowed to fall to the horizontal position on a felt pad, it is found that
the marble is no longer in the outer cup but has jumped to the inner one. Explain
this experiment.
Fig. P 12-16
Fig. P 12-17 Fig. P 12-18
12-17.A body hinged about a horizontal axis is allowed to oscillate. The center
of mass C is at distance a from 0. If W
is the weight of the body and /o is the moment
of inertia of the body about 0, find the frequency of small oscillations. Find the
length of an equivalent simple pendulum of the same mass that has the same frequency
of small vibrations.
12-18. A connecting rod AB is suspended on a horizontal knife edge at A and made
to oscillate. The frequency found to be 60 cycles /min. Then the rod is suspended
is
at B and oscillates at 54 cycles/min. The distance AB is 14 in. Find the radius of

gyration about the center of gravity C and location of C.
12-19. The compound pendulum of Prob. 12-17 is started in motion by imparting
to it an angular speed of wo at its equilibrium position,
initial (a) Using the conser-
vation of energy, find the subsequent angular speed as a function of 0. (b) What is
Problems 323
the amplitude of the motion? (r) Under wliat conditions will tlu- |>endulnm not
oscillate?
12-20. A 2- by 4-ft rectangular steel plate of 2 in. thickness is mounted in hearings
to rotate about a diagonal as shown. The weight of 2-in. -thick steel plates is SO psf.
The plate rotates at 1,200 rpm, and the distance between bearings is o ft. Calculate
the bearing reactions.
12-21. A horizontal shaft carries three unbalanced weights H'l = lb, Wi = 1 2 lb,
and H'3 = 3 lb. The distances of these weights from the shaft are ri = 8 in., r^ =
4 in., /-3 = 8 in. At each end of the shaft is a circular steel disk (,1 and li) of 1-in.
thickness. The entire as.sembly is by drilling one hole at a radius of
to be balanced
8 in. in each of the two disks. P'ind thediameters d of the two holes and the angles of
inclination of their radii. (The specific weight of steel is 0.283 lb/in.')
Fi(i. V 12-21
12-22. In order to retrieve a sjKml of thread by ])ulling at the thn>ad, what is the
limit on the angle a? Assume that there is sufficient friction between the floor and
the spool for rolling without slipping to occur.
W
A homogeneous circular cylinder of weight has a thread wrapped around
12-23.
itsmiddle. One end of the thread is fixed at A. The cylinder is allowed to fall.
Find the tension in the thread and the acceleration of the center of the cylinder.
Fig. P 12-23 Fig. P 12-24
12-24.homogeneous cylinder of weight Wa rolls down the face of a triangular

A
prism of weight Wb- Assume that the rolling is without slipping and the prism is
free to move horizontally on a smooth table top. Find the acceleration of the prism.
12-25. Find the angular acceleration of the cylinder in Example 12-4 by differenti-
ating the angular velocity with respect to time.
12-26. A rope of negligible mass passes over a frictionless pulley of mass M. The
moment of inertia of the pulleyabout its own axis is I. Two unequal masses mi
and OT2 are attached to the two ends of the rope. Find the acceleration of the system.
Fig. P 12-26
12-27. A r and mass )n is started in motion down an inclined plane

sphere of radius
of inclinationa by some initial translational speed vo. (a) )etormine tlie acceleration
I
of the sphere. (6) At what time does rolling witliout .slij)ping occur? (c) What is
the subsequent acceleration of the sphere?

12-28. Find the equation of motion of the wheel of mass /;( in Prob. 7-8. Write
the equation of motion directly, and also use the energy method. Exju-ess tlie equa-
Problems 325
tion of motion in terms of the angle <j> that the center of the wluu'l makes witli the
vertical through the center of the circular track.
12-29. A bar AH of weight U'l lias at one end li a cylindrical roller, of weight \V>
and radius r, that rolls without slipping along a circular track of radius /{. Har and
roller are made of uniform niaterial. Initially AH makes an angle ^n with the verti-
cal. Assume no friction at the hinges .1 and H.
(a) Find the angular velocity of the har .1^ when it falls to the vertical position.
(b) Find the angidar acceleration of the har at the inilial posilion.
(c) What is the frequency of o.scill.itiiin if the initial angle <)>» is small?
Fi(i. r 12-29 Fi(i. P r_'-;it)
12-30. A concentrated mass m is attached to one end of a uniform har of length I

and mass w/2 and hinged at its upper end. At what point can the har he struck hy
a perj)endicular force so that there is no initial acceleration at the concentrated mass?
12-31. The thin bar of Proh. 6-13 has weight IK and slides without friction because
of the action of gravity. Find the reaction between the bar and the s\irfaces over
which it is sliding.
CHAPTER 13
Lagrange's Equations
13-1. Introduction
In Sec. 9-1 we introduced the concept of generalized coordinates, which

were defined there as the independent variables used to specify the con-
figuration of a system. The configuration of a free particle, for example,
can be specified by three generalized coordinates, which may be x, y, and z
in the rectangular system or r, and z in the cylindrical system or, in
</>,
fact, three coordinates in any curvilinear system. In solving problems

in the previous chapters, we have seen that the most convenient coordi-
nate system to be used in a given problem depends on the nature of the
problem. Quite often the motion of a particle is constrained by some
physical conditions imposed in the problem so that with a proper choice
of coordinates, the generalized coordinates of a particle become less than
three in number. Thus the motion of the mass particle of a simple
pendulum oscillating in a vertical plane is described by only one gen-
eralized coordinate, namely, the angle of the pendulum from its neutral
position. Likewise, a particle constrained to move on the surface of a
fixed sphere has two generalized coordinates, the two angles representing,
respectively, the latitudeand the longitude. The number of generalized
coordinates is called (Sec. 9-1) the number of degrees of freedom. Thus
we see that constraint in a system is accompanied by a reduction in the
number of degrees of freedom of the system, f
t Our consideration in this chapter will be limited to the so-called "holonomic con-
straints" for which this statement is true. Holonomic constraints are those con-
straints which impose certain definite functional relationships between the coordinates
of the particles in the system. (Such relationships are referred to as integrable
constraint conditions.) For a system with holonomic constraints the generalized
coordinates necessary for the complete specification of the configuration of the system
can be varied individually and arbitrarily without violating the constraints.
326
Sec. 13-1] Inlrodudion 327
For a system consisting of .V particles, the numlxT of generalized
coordinates re(iuired to desc-rihe all possible configurations of the system
would be 'SN none of the particles were constrained. (Juite often, and
if
fortunately so, we do not have to consider such an extremely general

state of affairs. Depending on the nature of the physical system umler
consideration, certain idealizations and simplifications are often permis-
sible. Thus, if the system is considered a rigid body, the condition that
the distances between particles must remain constant reduces the number
of generalized coordinates to only six, three coordinates of the base point
(which can be any point of the rigid body) and three Eulerian angles
about the base point. For other examples illustrating the generalized
coordinates and degrees of freedom of systems under various constraints,
the reader is referred back to the outline in Sec. 9-1.
If we can describe the configuration of a system most efficiently by
its generalized coordinates, it is only natural, then, that we wish to write

the ecjuations of motion of the system also in terms of such coordinates.
By unknowns of the problem,
treating the generalized coordinates as the
we reduce the number unknowns to a minimum. The solution, in
of
turn, requires a minimum number of equations, since the number of
ecjuations required is ecjual to the number of unknowns.
Now it is true that Xewton's ecjuations written in any coordinates can
be transformed to any other coordinates. For example, in rectangular
coordinates Newton's equations for a particle in a plane are
F^ = mx Fy = my (8-2)
Transformed to cylindrical coordinates, the same e(iuations becôme
Fr = m{f - r<j>-') F^ = (r^) (8-4)

^ j^
Such a transformation was made in Chap. (>. There is, however, a

simpler and more general way of achieving this transformation, by means
of what is known as Lagrange's equations, which allow us to write directly
motion in terms of any desired generalized coordinates.
the equations of
Although historically Lagrange's equations were derived from Newton's
laws (and this is what we shall also do in Sec. 13-3), they can also be
derived from a more general principle, known as Hamilton's principle.
A discussion of Hamilton's principle is beyond the scope of this book;
we might merely mention that it represents a basic postulate which takes
the place of Newton's laws. Starting from such a postulate, it is possi-
ble to derive Lagrange's ecjuations,which in turn would yield Newton's

laws. Conceptually, therefore, one might consider Lagrange's equations
together with Hamilton's principle as being more general and funda-
mental than Newton's laws. For instance, Lagrange's e(iuations can be
328 Lagrange s equations
applied to nonmechanical systems such as electric circuits, which obvi-

ously cannot be described by Newton's laws.
As a practical tool in solving problems of dynamics, Lagrange's equa-
tions offer nothing that is basically new from what can be obtained by
Newton's laws. However, as mentioned previously, Lagrange's equa-

tions prove to be so powerful in the analysis of dynamical systems in
terms of generalized coordinates that it is entirely possible future students
will be taught right at the beginning to use Lagrange's equations to solve
all problems in dynamics. For the elementary type of problems, it is
true that Lagrange's equation is not as simple and as physically clear as
Newton's equation. But as the types of problems that we have to handle

grow rapidly in this modern age, before long we simply cannot afford the
luxury of dealing with each individual type of problem by separate
means. Instead, we may be forced to adopt a more general method
that enables us to solve a wider range of problems in a uniform manner
and with the least mental effort. In dynamical systems, Lagrange's
equation fulfills such a need.
Before actually deriving Lagrange's equations it will be instructive for
us to discuss briefly two related topics, virtual displacements and the
stability of equilibrium.
13-2. Virtual Displacements; Stability of Equilibrium

Let us consider a particle in equilibrium. Suppose we were to allow
this particle to suffer a virtual displacement. By virtual displacement
we mean any arbitrary infinitesimal displacement consistent with the
constraints imposed on the particle. If r is the position vector of the
particle relative to some fixed origin, we denote the virtual displacement
by 5r to distinguish it from an actual displacement dr that may occur in
some time interval dt. For example, if the particle were constrained to
remain in contact with a plane, the virtual displacement could be any
displacement in that plane.
Suppose the resultant force acting on the particle is denoted by
SFi = F. This resultant force may consist of forces exerted on the
particle by the conditions of restraint, known as reactive forces, as con-
trasted with all known as applied, or
other forces exerted on the particle
active, forces. The total virtual work done by the various forces acting
on the particle moving through the virtual displacement is, from Sec. 3-4^
8W = 2(Fi • 5r) = (SF,-) 5r = F • 5r
Now the necessary and sufficient condition for the equilibrium of a

particle (Sec. 2-3) is that F = 0. Therefore, since 5r is arbitrary and
cannot vanish,
8W = (13-1)
Sec. 13-2] Virludl displacenienls: stabilily of equillhrium 329
Equation (13-1) states that the total virtual work due to both active
and reactive forces is zero if the particle is in etiuilihrium. Now, if we
restrict ourselves to systems in which friction is not involved, then the
total reactive force at the constraint is perpendicular to the virtual dis-
placement. (For example, is constrained by its contact
if the particle
with a smooth surface, the reactive force at the contact is perpendicular
to the surface, whereas the virtual displacement satisfying the constraint
must be tangent to the surface.) Hence the reactive forces contribute
nothing to the virtual work. (Such constraints are known as workless
constraints.) Thus the principle of virtual work may be expressed as
follows: // a particle with workless constraints is in equilibrium, the total
virtualwork due to the applied forces must vanish under any arbitrary dis-
placement satisfying the constraints. The converse of this statement is
also true.
A more complete statement of the principle of virtual work con-
still
tains both a necessary and a sufficient condition. We have in etTect

proved the necessity condition; we leave it to the reader to supply the
sufficiency condition.
Moreover, it becomes a straightforward extension in order that the
principle be applicable to systems of particles, rigid bodies, and s^'stems
of rigid bodies. The statement remains the same; the
of the principle
subject of the principle is simply modified to cover the specific system in
question.
Now consider the case where the forces acting on the particle in
equilibrium are con.servative. According to Eqs. (8-41), the forces may
be derived from a potential-energy function and may be expressed as
F = -^
dx
Fy =
so that F= — ^-«
d.C
— -T~J
Oij
',
dz
-
The virtual displacement, for a particle free to move in three-dimensional

space, may be expressed as
5r = 5.ri -f- 5//j -|- 8zk
and the principle of virtual work [Eq. (13-1)] now becomes
5IK = F • 6r = - ^
dx
dx - ^
dy
5ij - "^
dz
bz = (13-2)
Since the virtual displacements are entirely arbitrary and independent,

Eq. (13-2) requires for efjuilibrium that
|L = o
dx
^= dy
^' =
dz
(13-3)
330 Lagrange's equations
Equations (13-3) serve as an alternative formulation of the necessary

and sufficient conditions for the equilibrium of a particle. It may be
expressed as follows:! A in equilibrium under the action of a
'particle is
conservative force system if, and only if, the variation of the ^potential energy
about the equilibrium configuration vanishes. If the constraints imposed
upon the given were such as to establish some relationship

particle
between the virtual displacements bx, by, and bz, then the condition of
equilibrium would assume a slightly different form from that of Eqs.
(13-3) (see Prob. 13-1). On the other hand, if there were more than
three independent virtual displacements (i.e., more than three generalized
coordinates which might occur for a system more complicated than a
single particle), we would have
y = V(qi,q2, . . . ,qn)
and bV = ^-
dqi
bqi
^
+ ^bq2+
dqi ^
• • •
-f- t-
dqn
5g„ =
Thus the equilibrium conditions would now become
f^f^...^f = Q
dqi dq2 dqn
(13-4)
where qi, q2, . . .

, qn are the generalized coordinates of the system.
From a mathematical point of view the equilibrium conditions of
Eq. (13-4) are just the necessary conditions to establish the extremum,
or stationary, values of the potential-energy function. In turn, these
values may be a relative maximum, a relative minimum, or another type
of stationary value. Thus we may state that a position of equilibrium
exists if, and only if, the potential energy has a stationary value.
According to our previous discussion of Sec. 5-6, equilibrium can take
on more than one form, namely, stable, unstable, and neutral equilibrium.
We now look for some method to categorize this situation. An examina-
tion of the physical problem discussed in Sec. 5-6 leads us to the intui-
tively convincing notion that a position of equilibrium is stable if, and
only if, the potential energy is a minimum at the equilibrium configuration.
Still another example of these considerations is the case of a ball
classic
resting, respectively, on the horizontal portion of a concave, convex, and
fiat surface. Here again we are led to the same conclusion.
We may further clarify these considerations from a slightly different
point of view. Suppose we consider the case where F is a function of one
generalized coordinate, say x. Letting the subscript represent the
t This conclusion could luivc also been reached by observing from Sec. 8-7 that
SV = —8W and tliat tlu'rcfore, for ('(luilibriuui. 8V = 0.
/
Sec. 13-2] Virlual displacements; slabilUy of equilibrium 331

equilibrium position, a Maclaurin's expansion of V al)out yields
\<l.v)/ '^\d.v^),2\
If we expand T in the immediate vicinity of the equilibrium position,

then we can let x be some infinitesimal virtual displacement dx. Thus
Since is a position of eciuilibrium, wc have, from lv|. (\'A-\),
m. "
I'pon neglecting terms that are of a higher order than dx'-, we deduce that
the change in potential energy is
/d-V\ 8x
AV = r - r., . . „.
\dx'-)o .,,
2!
Now in order for the potential energy to be a minimum at the equilib-

rium position, and hence to have stable equilibrium, we must have
d^'
It will be immediately recognized that Eq. (13-5) is just the mathematical

condition necessary for the existence of a relative minimum value of the
potential-energy function 1'. If d-V/dx- < 0, unstable eciuilibrium will
occur; if d^'/dx^ = 0, higher-order derivatives must be examined in
order to test for the type of equilibrium.
If r were a function of several variables, the criterion defining the
types of stationary values l)ecomes increasingly complex. For example,

if V were a function of two variables, say x and //, the mathematical con-
ditions for defining the stationary values of a function of two variables,

and thus the e(|uilibriuni conditions, could be summarized as follows:
Equilibrium; -r— = t— =
dx ay
Stable equihbnum; (
- ~- I — ô ^r^ < ^
\dx dij dx^ dy-
dx- dy-
Unstable equilibrium :
f . . ) — ^V ^-r < ^'
dx' dy^
dx^ dy^
Example 13-1. In order to illustrate the foregoing ideas, let us con-

sider the following example (Fig. 13-1). A perfectly frictionless wheel
of radius r has attached to it a rigid, weightless bar of length I and a
spring of spring constant k. Attached to the bar is a flexible string from
which hangs a weight W. When 6 = 0, the spring is in its unstretched
position. Determine the equilibrium position of the system and classify
the types of equilibrium.
The system we are dealing with is acted upon by conservative forces;
we may therefore profitably use the potential-energy formulation to
Fig. 13-1
achieve a solution. Taking the position ^ = as our datum for com-

puting the potential energy, we have, at any position 6, that the potential
energy of the weight W is (Sec. 8-7)
TV = -117(1 - cos 9)
The change of length of the spring at this position is rd; therefore the
potential energy of the spring is (Sec. 8-7)
The total potential energy of the system now becomes
V = Vw + Vs = Hh-r'-d'' - Wl{l - cos 6)
According to Eq. (13-4), the equilibrium position is defined by
^=
ad
= h-d - 117 sin e
and equilibrium can take place for angles 6, satisfying the transcendental
equation
e = T^ sm d
Icr-
S«'c. 13-3] Lagrange s equation for a jHirCirle 333
One = 0; other positions can he deter-

ol)\i()Us e(niilil)i-ium position
is
mined remaining roots of the transcendental eciuation.

l)y finding th(>
In order to determine tlie type of e(piiHl)rium according to Eq. (13-5),

we form
, 2
= /»'*- - ^yi t'os e
For 6 — 0. this becomes
w= ''' - "''
Thus e(iuilil)rium is stable at a position of 6 = for /.•/•-'

> 117 and
unstable if At- < 117. At other positions of eciuilibrium, corresponding
to the other roots of the transcendental equation, the type of eciuilibrium
can be investigated similarl3^
13-3. Lagrange''s K(|iiatioii for a Particle
Let Xi, .To, Xz be the rectangular coordinates of a particle and 71, 70,
. . . , 7n the indepeiidcni generalized coordinates of the sj'stem to which

the particle under consideration belongs. It is, of course, possil)le that
the system consists of onl}^ one particle, in whicii case the problem
becomes relatively simple. The following discussion, however, is valid,
irrespective of the number of the other particles in the system.
Since the motion of the particle, and hence its rectangular coordinates,
are completely determined l)y spec-ifying the generalized coordinates
of the system, there must exist the following functional relationships
between the two sets of coordinates if
X, = x.ifjijf., q„) (13-7)
where the index i may be any one of the three nunibeis I, 2, 3. 'i'he
scalarcomponents of the \elocities of the particle can therefore be written
in terms of the generalized cooidinates by differentiating the above
expressions with respect to time and are given as
Thus any f, is a function of the \ariables dXi/d(/j and (jj for all values
of j from 1 to n. Considering the variation of i-, with respect to a par-
ticular ('/j while kccpitu/ (dl other ('//s and all ('ix,/(l{//s eonstont. we have,
t We restrict our {liscu.s.sion here to tho.sc .systems where flic nLitimis hefween tlie
two coordinate systems <lo not involve time explieitly.

from Eq. (13-8),
^=P
dqj dqj
(13-9)
Consider now the kinetic energy of the particle

3
mx{
(13-10)
2
1 = 1
From Eqs. (13-7) and (13-8), Xi is seen to be a function of the generalized

coordinates qj and their time derivatives gy. Hence we can form
3
dqj
/ ^î^Z
L^ dqj
(13-11)
i=l
3 3
dT
and -r-^ / mxi~ = / mxi^ (13-12)
dqj Zv dqj Zv dqj
i=l 1=1
where we have used Eq. (13-9). Since dT/dXi = mxi is the momentum
by analogy we
of the particle, call dT /dqj the generalized momentum.
Differentiating the preceding expression with respect to time, we obtain
3
dt
1=1
But ^(|5) = y-£!|^,,.^(V|^,;.)

dt \dqjj Zv dqj dqu dqj \L^ dqu /
/fc=l k=l
Here the index k is simply a dummy index used in connection with the
operation of summing from 1 to n. Furthermore, on account of Eq.
(13-8), this becomes
d /dx\ _ dXi
dt\dqjj dqj
Hence the last term on the right side of Eq. (13-13) reduces to
3 3
L( ^^' dt \dqjj
~ /^ ^-^
' ^q. Qq.
1=1 1=1
where we have used Eq. (13-1 1). Substituting this result into Eq. (13-13)
and taking into consideration that mXi = Fi by Newton's law,
3
d'^
dt \dqj/ dqj L^ dqj
Sec. 13-3] Lagrange's equation for a parlirle 335
The work done by the forces/-', for a siiuill (lisplaceineiit of (lie particle
is rSec. 3-4)
dW = y I\dxi (13-11)
If dx, is considered to he a virtual displacement, then l'](i. (IM-Jl) repre-

sents the \iitual work. Since .r, is a function of r/'s, by 1m|. (i;i-7) we
can form
3
dW
dqj
1=1
and call it Qj. Thus
This is Lagrange^ s equation of motion.

The significance of Qjis also important. By definition,
dW
Qj =
^ (13-16)
Comparing this with
F =^-E
'
dXi
which is obtained from Ya\. (13-14), we find the following analogy. The
work done upon a small displacement rf.r, is b\ dx„ while the work done
upon a small displacement dqj is Qj dqj. We therefore call f^/s the
generalized forces corresponding to the generalized coordinates qj. The
(luantity Qj may he thought of as the work-absorbing force along q,.
Since the generalized coordinates qj may not have the dimension of a

length, the generalized forces Qj may not have the dimension of a force.
For example, if qj is an angle, (Jj will have the dimension of force-length,
the dimension of a moment. In any event, the product of a generalized
force and a generalized displacement will ha\'e the dimensions of work.
Again, if dqj is considered to be a \irtual displacement, then dW = (Jj dqj
may be thought of as the virtual woik.
Lagrange's ecpiation (13-15) can he further simplified if the forces
acting on the particle are conservative, for then the work done is the
negative of the increment of potential energy 1' [K(\. (8-40)]. Thus
Eq. (13-10) can be written
Qj = -^ dqj
(13-17)
and Eq. (13-15) now becomes

d /dT\ _dT ,dV ^Q
dt\dqj) dqj dqj
Let us introduce a function L, called the Lagrangian function, such that
L = T -V (13-18)
Remembering that for a conservative force field F is a function of posi-

tion onlyand is hence independent of the velocities qj (Sec. 8-7), we can
write Lagrange's equation for a particle acted upon by conservative
forces as
should be noted that in the Lagrange equation (13-15) or (13-19),

It
qj may
be any one of the generalized coordinates in the problem, so that
we can write one Lagrangian equation for each of the g's. We therefore
always have available the same number of equations as the number of
unknowns, which are "picked beforehand to describe most conveniently the
configuration of the system. We thereby automatically obtain a neat
formulation of the problem and in many cases introduce considerable
simplification. It must be realized, however, that Lagrange's equation
helps us only to formulate the problem, i.e., it supplies us with the
differential equations, but that it does not tell us how to solve them.
Another important advantage of Lagrange's equations, so far as the
motion of the system is concerned, is that the forces exerted by the con-
straints do not appear in the equations since they do not contribute to
the work done, W. These forces are generally unknown, and their
presence in Newton's laws sometimes requires additional manipulations
(see, for instance, part b of Example
In addition, the kinematic
12-4).
relations necessary to deal with acceleration are not required;we need
only concern ourselves with the velocities involved in the formulation
of the kinetic energy.
We shall now illustrate the use of Lagrange's equations by two exam-
ples. The first example is completely trivial. The second one shows
the neatness of this new method.
Example 13-2. A particle is constrained to travel along a straight
line which may be considered as along the direction of x. The particle
has one degree of freedom; its position is specified by the single general-
ized coordinate q = x, measured from some fixed point along the line.
Therefore
T = — 2
r
Sec. 13-3] Lagrange s equal ion for a /xirlirle 337
dT dT .
(I /dT\
^ = 5^ =
dq dx
•
r/U = I' ^(Ix Q = —- = —- = Fx

dq ax
Lagrange's equation (13-15) therefore gives f'^ = mx.

Example 13-3. Mud tlic e(|uati()ns of motion of a particle in spherical
coordinates (Sec. G-8).
Let the spherical coordinates r, 6, he the generalized coordinates
'/h 92, qi- Then, according to Fa]. (()-30). the kinetic energy becomes
T = >^ ;/*('•- + r-d- + r-0- sin- 6)

dT dT d (dT\
,
and, for c
i = -i
\, ^— = ^^ = wr
.
rn t^ = ""'
dqi dr (ft \dqj
( I
dT = dT ...
= m{rd' , ...
+,
np' sni^
.
, „,
^)
dq, dr
Hence Lagrange's e<ination (18-15) gives the first equation of motion as
m{r - rd- - r4>- sin-' 6) = (^
Similar!}', the second and third of the ecjuations of motion become
m -7- (r-d) — /•-0- sin 6 cos Q:
m -r {r-4> sin- 6) = Q^
Now the work done in a small general displacement
dr = dr Qr -\' r dd V'o -\- r sin d e^

ist d\V = V dr
or dW = I\dr -\- I'\r dd -\- l'\r sin d4>
Hence Qi = -r- = I'
t The small general displacement dr may Ix- fouiid liy forniiiiK tlic (liffcrciitial of
r = rv:r and usinp; the results of See. (i-8.

:
Therefore the Lagrangian equations yield the following equations of

motion
Fr = w(f — rd"^ — r^~ sin^ d)
Fe
Sec. 13-1] Lagrange's eqnalion for o sysleni of /xirlirles 330
Also, the potential energy, using the initial position of the cylinder as a
datum, becomes
V = mgijc
Hence Lagrange's eciuation (Ki-lOj gives
^^ ^ 1 + hjmr-
Exaniple 13-5. For a symmetrical gyroscope pi\<)ted at a point
(_l"ig. 12-8), the kinetic energy is (Proh. 12-4)
r = ].2\h{e- + - sin'-^ d) + /(i^ -h 4> cos ey-\
Considering 6, . and \p as the generalized coordinates, we can lorm the

generalized momenta
dd
^ = /i0 sin2 d -\- I cos d {\P

-\-
(i>
cos 0)
^ = /(^ -f 4, eo.s 0)
Hence the Lagrange ('(juation (lo-).")) becomes
he - /i0- sin d cos 6 -\- I<i) sin 6 {\p -\- 

cos 0) = Qs
[/,0 sin- -h / cos 6 {i -\- 4> cos ^j] = C>0

J
^ [/(,A 4- (^ cos d)] = Q^
The generalized forces can he obtained by considering the work done

(Sec. 12-2):
Making use of Eqs. (12-59), we obtain
d\V = -.U, sin f/0 + -U„ r/9 -|- il/,(cos 6 d -\- d^)
Hence Qe = -^ = Alu
dtp
With the above values for f/s. Lagrange's efpiations in this case are
identical with Kqs. (12-00) as derived in Sec. 12-8.
13-5. Small Free Vibrations of Coupled Systems

In this section we shall demonstrate the wide application of Lagrange's
equations by considering the free vibration of a system of particles which
are coupled; i.e., interaction exists between the motion of the particles
such that the motion of one particle affects the motion of some other
particle or particles. Physically, this may be accomplished by connect-
ing the particles together with linear restoring forces. The restoring
forces are linear if their magnitudes are proportional to the displacements
of the particles from their respective neutral positions. As an example
of linear coupling, we imagine that the particles are connected to one
another by Many mechanical systems can be considered
elastic springs.
elastic, having linear restoring forces for sufficiently small displace-
i.e.,
ments of the constituent particles. We shall restrict our discussion to

those systems in which the dissipative, or damping, forces can be ignored.
In other words, our system is assumed to be conservative, and hence a
potential energy V exists. Dissipative forces may be treated equally
well following the general method employed here.
Let Qi, q2, . . .
, Qn be the generalized coordinates describing the
configuration of our system and of every particle in it. We so choose
these generalized coordinates that they all vanish when the particles
are at their respective neutral (or equilibrium) positions. The potential
energy of the system is a function of coordinates only; hence we can write
the potential energy in a Maclaurin's expansion of a function of several
variables about the equilibrium position as follows:
where the subscript denotes the value at qi = q2 = = ?« = 0. • • •
In the preceding equation, the first term on the right, Tô, can be arbi-
trarily chosen to be zero if we agree to measure the potential energy of
the system from the neutral position. Hence
Fo =
FromEq. (13-17) ^=
oQi
-Qi
At qi = q^ = = r/„ =
the forces acting on the system are in
0,
equilibrium, and hence the component of the total force along any
Sec. 13-.')] Small frci' vihralions of conpled sy.slenis 311
direction is zoio. iSiiu-c the generulizecl loice.s Qi ure the work-ahsorhing

forces along Qi they must also he zero at qi = qz = • • • = q„ = 0.
Thus
Alternati\ely. wo could \u\\v reached tliis sanu> coMclusion Ironi Va\.

(18-4).
Equation (13-21) now heconies
= (13-22)
» >2
2 2^.^.
t=i y=i
where we haxc used the notation
bn =
for the sake of l)revity. It is seen that because of our assunii)tion of

small displacements, terms in Eq. (13-21) of degrees higher than (juad-
ratic have been neglected. f We also notice that 6,j = 6>, by definition
and that they are constants to be evaluated at the equilibrium position.
The potential energy V in ¥a\. (13-22) is a quadratic form. Since 7, is
measured from a position of minimum To (the neutral position), then

h,j is a positive constant for all i and j (Eq. 13-5) and V is thus a positive
quadratic fmiction.
To an expression for the kinetic energy of the system, we start
find
with one particle, say the />th particle, of the system, which may consist
of a total of N particles. The kinetic energy of the Ath particle is, by

Eq. (13-10),
3
i = \
where two subscripts are now necessary for ./, the first subscript k specify-
ing the particle under consideration. Substituting for its value as ./•
given by Ya\. (13-8), the kinetic energy of the Ath particle is
1 = 1 t = 1 j = \
Summing over all the .V particles of the system, we can write tiie kinetic
energy of the system in the following form:
II n
^ = ^^
I I "'>'?'''> (13-23)
.=1 >=i
t This statomont also .serves to rlarify what we iricaii hy ".small (iisjjlaccincnts."

342 Lagrange' s equations
where the new symbol a^y stands for

N 6
dXki dXki
;=1 1=1
By this definition, a,j = a^,. In general, the ai/s are functions of

Qi, qo, . . .
, qn- However, for small ^'s near the equilibrium position
of the system, we may consider a,j as positive constants and equal to
their values at 51 = ^2 = " ' • =0; i.e.,
A^ 3
L^ L^ \dqi/o\dqj/c
k=l 1=1
Since T is inherently a positive quantity, it also is a positive quadratic
form.
We see by Eqs. (13-22) and (13-23) that V and T are quadratic forms of
the generalized coordinates q and generalized velocities q, respectively.
Substituting these two equations into Lagrange's equation (13-19),
n n
2 a^r + ^ hjiq, = (13-24)

1=1 i=l
This is the equation of motion for small, free, undamped vibrations of

mechanical systems. For an n-degree-of-freedom system, we have n
such equations by letting j = I, 2, n, consecutively. The . . .
,
interpretation and solution of such equations can best be illustrated by

a two-degree-of -freedom system. Generalization to an n-degree-of-
freedom system follows a similar pattern.
13-6. Small Vibrations of Two-degree-of-freedom Systems

For n = 2, Eq. (13-24) becomes
aiiqi + ai292 + b^qi -|- 612^2 =

C13-25)
a2i9i + 0-22(12 + &2igi + ^2292 =
The system defined by Eqs. and dynamic
(13-25) contains both static
coupling; there is static coupling between the restoring forces due to the
presence of the terms 612 = 621, as well as dynamic coupling between the
accelerations by virtue of the terms a^ = ci2i- Assuming now that

9i and <72 are sinusoidal functions of time, we choose as our trial functions
for the solution of the two simultaneous linear homogeneous ditferential
equations (13-25),
qi = Ai sin {pt ^ <t>) ,
52 = A2^\n{pl-V )
^ ^ ^
Sec. 13-6] Small vlhrdlitins of lu'o-degree-of-freedoni syslenis 313
Sul)stitutiiig these trial solutions into Va\s. {\'.\-2i)),
{bn - Oup'-)Ai + (/>!•• - a, ./>-).!. = . _.

^''^"'''^
(621 - a,,p-)A, + (/^>, - a,,//^).l, =
I'roin the theory of linear aljêhraic eiiuations, we reco^nizc^ tiiat .4 j
aiui A
can have nonzero
. (iiontrivial) solutions only it the determinant
vanishes, i.e.. onlv if
^11 - aup- 612 - ai^p"^

= (13-28)
&21 — a-iiP' ^22 — a22P*
Expanding this determinant leads to an alf2;el)raic ecjuation of the

second degree for the unknown p-. Since p is a frecjuency, this algebraic
efjuation is called the frcquenc]! equation, or characteristic equation. It
can he shown that roots of the frequency equation are all positive and
real provided the potential energy T is a positive real function. f We
shall denote the two positive roots as pi^, p-^. These two natural
circular frecjuencies are sometimes referred to as the normal frequencies.
In this discussion we
omit from further consideration the special
shall
cases of zero roots or repeated roots of the frefjuency eciuation. Zero
roots occur when it is possible to have displacements without changing
the potential energy of the system, and as such correspond physically
to a rigid-body motion of the system. Repeated roots indicate that the
system vibrates harmonically in two modes with the .same frequency but
with different amplitude ratios. Further di.scu.ssion will clarily this
point.
Corresponding to each of the normal frequencies pi and p^, we have a
solution of the type (take pi, for example)
qi = An sin (pit + 0i) 72 = A->i sin (p^t + 0i)
(The second subscript to .1 refers to the normal frequency pi.) Tlie

general solution, then, is obtained by the superposition of the .solutions
corresponding to each of the normal frequencies and is given as
qi = An sin {pit + 0i) + A v> sin (/;•-/ + <j)->)

(13-'><J)
92 = ^l2i sin (/>,/ + </)i) + -122 sin {pj + <!)•}
The determined by each root or fre(|uen('y of the characteri.stic

oscillation
ecjuation called a principal mode, or normal mode, of vibration.
is
We notice that the coefficients A,j appearing in Eqs. (l.V2()) are not
independent; they must .satisfy Iv|s. (1:^27) and are thereby related
by virtue of tho.se expressions. For example, substituting pr into the
t Since we have shown that T is a positive quadratic form, we indeed obtain posi-
tive real roots, the fn'(|uencies, from the characteri.stic c(|Uiiti()ii. W'c (h), however,
omit the proof of the above statement.
firstf of Eqs. (13-27), we obtain a distinct ratio between A 21 and An.

The ratio A21/A11 — X21 is called the amplitude ratio and is determined
to be
_ 611 - aiipi^
A2I — — T
— 7,
O12 ai2Pi-
In a similar fashion, the amplitude ratio A22A412 = X22 can be formed.

It should be noted that the amplitude ratios are definite numbers depend-
ent only on the physical properties of the system.
In terms of amplitude ratios, the general solution can now be written
qi = All sin (pit + 0i) -1- .4 12 sin (p2t -f ^2)

(13-30)
92 = A 11X21 sin {pit + <î) 4- -.412X22 sin (p2t -H ^2)
The general solution (13-30) contains four arbitrary constants of integra-

tion. All, A12, 4>i, and 02, which may be determined from the initial
conditions of the problem. Clearly, there exist two initial conditions for
each generalized coordinate or degree of freedom.
We observe that in each of the principal oscillations each coordinate
executes simple harmonic motion in the same frequency and in the same
phase. Moreover, each coordinate reaches its extreme and neutral
position simultaneously. The amplitudes of vibration of each coordinate
are not the same; however, their relative magnitude remains constant
throughout time and is given by the amplitude ratios. Thus the most
general form of free oscillation of a conservative system with two degrees
of freedom consists of the superposition of two simple harmonic motions;
the most general form for a system of n degrees of freedom consists of
the superposition of n simple harmonic motions.
This general theory will be illustrated by a typical example. In order
to achieve some economy of effort, we shall make use of the general
expressions previously developed. It should be noted, however, that the
solution could also be obtained without recourse to the general equations,
by a straightforward application of the principles. %
Example 13-6. Consider a two-degree-of-freedom system consisting
of two particles of equal mass m suspended from two springs of the same
stiffness ki and statically coupled by a third spring of stiffness A-2 (Fig.
13-2). If the two masses are constrained to move in a vertical direction
only, the configuration of the system is completely determined by the
vertical displacements Xi and X2 of the two masses from their respective
t Either of the two Eqs. (13-27) would he satisfactory, since only n — 1 of these
equations are independent by virtue of the frequency equations (13-28). It is easily
verified that the second of Eqs. (13-27) yields identical results.
} For purposes of better understanding of the theory, when solving problems the
student should follow the alternative approach ratlier than the one employed here.
Sec. 13-6) Sntall vlhrulions of I iro-deyrec-of-freedom syslenis 3 15
neutral positions. The.se coordinates can then l)e cho.sen as the general-
ized coordinates of the .system.
For the entire system, tiie kinetic and potential energies are
V - Uk.xr + ^k,x^ + Uk,{x, - .»•.)-
Comparing this with Imjs. (i;i-22; and (13-23),

we find that
012 = 021 =
bn = bri = A-i + k. hvi = b.n = -k
Hence Ecjs. (13-27) give
{ki -h Ao - w//).li - A-2,l, =

A-o.li + (/,-, -h /,-2 - mp-')A, =
Also E(is. (13-28) yield

the two masses therefore remains at rest. This motion is equivalent to

that of a single-degree-of-freedom system consisting of a mass m sus-
pended from a spring of stiffness fci + 2k2. The natural frequency is
therefore a/(/ci + 2/c2)/m. The general free motion two masses

of the
is a superposition of the two modes of vibration as indicated by Eqs.
(13-30). Hence
xi = All sin (pit + (j)i) + ^12 sin (p2t + </>2)
X2 = All sin (pit + 4>i) — A 12 sin (^2^ + ^2)
The four constants An, A12, 0i, 02 are determined by the initial con-
ditions, i.e., the values of Xi, ±1, X2, ±2 at t = 0. Although each mode of
vibration is sinusoidal with time, the superposition of the two modes is
not sinusoidal. In fact, it is not even periodic unless the ratio of the
two natural frequencies is a rational number. If the two natural fre-
quencies differ by only a slight amount, the combined motion will
exhibit the phenomenon of beats.
It should be noted that our solution of X21 = +1, X22 = — 1 in the
above example is due to the fact that we have chosen equal mass for the
two particles and the same stiffness ki for the top and the bottom springs.
For a more general two-degree-of-freedom system, the procedure out-
lined at the beginning of this section should be followed to determine
pi, P2 and the corresponding values of the two amplitude ratios. The
two amplitude ratios will not in general have values of +1 and —1.
13-7. Free and Forced Vibrations of

n-degree-of- freedom Systems; Normal Coordinates
The previous discussions of a two-degree-of-freedom system can now
be generalized to an n-degree-of-freedom system.
As stated in Sec. 13-5, the potential energy and kinetic energy of an
n-degree-of-freedom system can be expressed, respectively, by Eqs.
(13-22) and (13-23).
n n
Potential energy: V = l-i / / ôMy (13-31)
n n
Kinetic energy: T = }4 / / «uMj (13-32)

1=1 >=i
The equations of motion for an n-degree-of-freedom system are therefore

given by Eq. (13-24), which can be written
n n
2 Oi/h + 2 b^jQj = (13-33)
There are n such equations by letting t = 1, 2, . . . ,

n.
Sec. 13-7] n-degree-of-freedom systems; normal coordinates 347
As in the two-degree-of-frcoiloin system of Sec. 13-G, we substitute

the following .solution for the unknown 7^:
g> = C.l, .sill (p/ + 0) (13-34)
Here the constant C is introduced as a scale factor for rea.sons that will
he clarified .shortly. Suhstitutinji; E(|. (13-34) into Va\. (13-33). we obtain
the following /( algebraic eciuations for the n unknown amplitudes .{,:
2 {baAj - v'aijî) =0 t = 1, 2, . (13-35)

y=i
From the theory of linear homogeneous algebraic ecjuations. we

recognize that there can be nontrivial solutions for Aj only if the deter-
minant of the coefficients vanishes, i.e.,
îH - ;/-V/i„
621 — p-a-.'i h-n — p''a22
631 — p-'asi • •
•
= (13-3G)
h„\ — p-a„i • • • • • •
h„n
f>„„ — p-a„„
Expanding this determinantal equation yields an algebraic ecjuation

of the n{\\ degree for p-. The roots of this e(|uationt are the /; \alues
for p^, each p representing a normal frequency . We shall identify these n
values by a subscript, for in.stance, pk^, k = 1,2, . . . ,
/;.
Corresponding to each pa--, we can calculate from E{i. (13-35) the

relative magnitude of the unknown amplitudes Aj, expre.ssed in n — 1
ratios between any n —

amplitudes and the one remaining amplitude.
1
The vibration corresponding to such a set of A/^ for each pk is a prin-

cipal mode, or normnl mode, of vibration. [Since there is one complete
set of amplitudes for each pk, we need two subscripts to identify the
amplitudes; for instance, Ajk, the first subscript, refers to the coordinate
Qj as indicated by Ecj. (13-34), and the second subscript refers to the

normal frefjuency p/t.] For an n-degree-of-freedom system, there are
therefore normal modes, each mode with its own amplitude, fretjuency,
)i
and pha.se angle.

The most general vibration for such a .system consists of a linear com-
bination of all the n normal modes. Hence
Qj = y (\Ajk sin {pkt + 0jt) (13-37)

k=\
The magnitude Ck and the pha.se angles 0a are determined by the 2n

initial conditions, i.e., the given r/'s and ^'s at time / = 0.
ha stated previou.sly, the roots are all real and positive as a roiise(nienr-e of posi-
t
tive values for V and T and a,/ = a,,, fc,, = /'/,.
1
All this discussion must sound familiar to the reader, because it is more
or less a repetition of precisely the same concepts already described in
Sec. 13-6 for a two-degree-of -freedom system. In the following, how-
ever, we shall bring forth yet another concept, the concept of normal
coordinates, which proves to be essential for a deeper understanding of
the problem.
Let us consider any two of the n normal modes, corresponding to two
normal frequencies pk and pi, respectively. The amplitudes Ajk and Aji
for these two modes satisfy the following equations, which are merely
Eq. (13-35):
n ft
^ bijAjk = 2 VkâijAû (13-38)

7=1 y=i
n n
^ bijAji = ^ pMjAji (13-39)

y=i y=i
We notice that each of the above two equations represents n equations I
fori = 1,2, . . . ,n. Now multiplying Eq. (13-38) by .4^7 and summing |
over all i, !
n n n n I
hijAjkAu = 2 ^ pk'a.jAjkAu (13-40)

2 ^
i=iy=l =
i=l j 1
j
Similarly, multiplying Eq. (13-39) by Aik and summing over all i,

j
n n n n i
Y y hijAjiAik = X A praijAjiAik
i=l j = 1 i= 1 j =1
Since i and j are dummy indices, they can be interchanged in the last
expression, which becomes
n n n n
h,,AaAik = ^ ^ pra,;AaAik (13-41)

i
2
=
2
= i i=\ j=l
1
Remembering that a,^ = aî and 6,j = ^ji (Sec. 13-5), we see that the left
sides of Eqs. (13-40) and (13-41) are identical. Taking the difference
of these two equations,
n n
- pr) a„.l,...l,, =
(Pa-
^ 2=
» = 1 i 1
Sec. 13-7] n-degree-of-freedoin syslenis; normal courdinules 349
Since, in general, pk'- 9^ pr il A- ^ I, we must huve

ri II
a.jAaAjk = k (13-42)
2 ^ 9^ I
On tlie other hand, il k = /. the expression
n n
,=1 y=i
is, from E{[S. (13-32) and (13-34), proportional to the magnitude of the
kinetic energy due to the mode pk. Since the magnitude of kinetic
energy cannot be zero, the last expression does not vanish. Now it has
been mentioned several times before that Eq. (13-35) can only determine
the ratios of Ajk and not their absolute values. Hence there will be no
loss of generality if we set
n n
a^jA^kAjk = (13-43)
^ 2
1=1 >=i
1
When the amplitudes are made to satisfy Eq. (13-43), they are said
to be normalized. It is now clear why we need a scale factor C in Ya\.
(13-34) or Ck in Etj. (13-37) (the subscript k is to identify C with the
particular p^-^). We need Ck to fix the magnitude of each normal mode
to fulfill the given initial if the amplitudes Aik are normalized.
conditions
Equations (13-42) and (13-43) can now be coniljined to form
ri )i
t = 1 j = l
where 8ki is the Kronecker delta. Equation (13-44) represents the rela-
tions that must exist between the n- normalized amplitudes, referred to
as orthogonality conditions.
The orthogonality conditions can be used to advantage in determining
the magnitude Ck and the phase angles </>*.- of the various normal modes
from the given 2n initial conditions. At / = 0, we have, from Eq.
(13-37),
n
9.(0) = Y,
^*'-^'*^^" '^'^ (l-*^--*^)
k=\
11
7.(0) = C'^-'1'*P* co« k (13-46)

Y,
k = \
Assuming ^,(0) and r/,(0) are given, we can then multiply each side of
the above ecjuations by ttijAji and sum over both i and^. On account of
:
the orthogonality conditions, the results are
n n
Ci sin 0^ = y y [q,(0)]aijAji (13-47)

1=1 y=i
n n
Ci ^^^ î^ -
1 Vy V [qi{0)]aijAji (13-48)
/^
1=1 y=i
Thus Ci and 0; can be evaluated in a straightforward manner.

Let us now define a new set of coordinates q'j, related to the original
coordinates q^ in the following manner:
q, = 2 Ai,q', (13-49)
On account of the orthogonality equation, this is equivalent to
n n
lk= ^^
iî yî
a-u-AMi (13-50)
Equation (13-49) or (13-50) constitutes a linear transformation between

the two coordinates qi and qi.
In terms of the new coordinates qi, the potential energy of Eq. (13-31)
becomes
n n
V = }i J^
X ^''^'^'
n n n n
^ ^^ 2 Z X 2 ^'i^'^î'^k^'i
i=li=\ k=\ 1=1
n n n n
= H X X ( 2 2 ^îî^î) ^'^<i'i
k=\i=\ i=i j=\
Substituting into this expression the following identity from Eq. (13-35)
n n
y hijAji = ) pi-aijAii
y=i y=i
we obtain, with the help of the orthogonality conditions (13-44),
y = y2^ Pk'q^ = Hiprq? + P^q? + " ' •

?>«V„^) (13-51)
k=l
Sec. i:i-T] n-degree-of-freedom systems: normal coordituitcs 351
Furtliennore. .since Kq. (13-4*)) also means
n
9. = A,,q[
Yi
k=\
the kinetic energy of Kf|. (13-32) becomes, follow injj; an analogous

procedure.
ft
T = H 2"^^ ^ ^2(7;' + V^' + • • •

+ O (13-02)
Thus we reach the remarkable result that, in terms of the new coordi-
nates ql-.
the expressions for potential enerjj;y and
kinetic ener}z;y are both
simplified to such an extent that all cross-product terms disappear.
These new coordinates which render this simplification possible are
known as normal coordinates of the system. The procedure involved in
such a simplification, i.e., in obtaining Fa\s. (13-o1) and (l3-.')2). is called
the reduction, or diagonalization, of quadratic forms.
In terms of normal coordinates, the Lagrange ecjuation
di \d<rj a,;
now yields the equation of motion for free \il)rations as
9', + PrÂ =
92 + P-r(/2 = (13-53)
7: + P«Vy'„ =
The solutions to these eciuations are, clearly,
q', = C, sin {p,t + 0,) (13-54)
Substituting these solutions into K(\. (13-49), we get the linal solution
of the generalized coordinates as
Qi = y. ^'*-**'* '"î" (P''f + <^*) (13-55)

jt=i
which is Again we see that the solution

just Iv|. (13-37j obtained before.
for any generalized coordinate simply the superposition of // harmonic
is
motions, each one of which corresponds to a normal mode.

The use of normal coordinates becomes esi)ecially advantageous when
dealing with forced vibrations. Suppose we let F,{t) be the forces acting
on the system and corresponding to the generalized coordinate q,. Then
the work done by the forces due to small displacements of the generalized
coordinates is
dW = ^ Fi dqi
By the use of Eq. (13-49), this work can also be expressed in terms of
the normal coordinates, becoming
dW = y y FiAij dq'j
y=ii=]
Now, according to Eq. (13-16), the generalized forces corresponding to

the normal coordinates become
n
^'^ =
Wr 1(^'^''
1 = 1
(13-56)
Hence Lagrange's equations (13-15) yield, for the case of the forced
vibration of the n-degree-of-freedom system, the following set of n
independent, linear, nonhomogeneous differential equations:
n
?i + VH = Qj= 2 F,A,j (13-57)
Following now the solution given by Eq. (9-58), we see that solutions of
the n equations of Eq. (13-57) are given as
q'j = Cj sin (p,t + 0.) + :^ / y F,(r)A,j sin p,{t - r) dr (13-58)

i=l
Once the normal coordinates are determined as functions of time (which

implies that constants of integration Cj and 0^ must be evaluated), the
solution for the generalized coordinates as functions of time can be
readily found by a direct application of Eq. (13-49).
It is now readily apparent that the method of normal coordinates has
great merit, for by its we obviate the necessity of solution of n
application
simultaneous linear nonhomogenous differential equations. Instead we
solve n independent equations, the total solution consisting of the super-
position of these n individual solutions, one for each normal mode. For
this reason this method is sometimes referred to as the method of modal
superposition. Additionally, it might be added that the use of normal
coordinates is of great utility for continuous systems where the number
of degrees of freedom (and consequently the number of generalized
coordinates) becomes infinite.
Sec. 13-71 n-degree-oj-jreedom systems: normal coordinates 353
Example 13-7. As an illustration of the free vibration of a multi-
degree-of-freodom system ineliulinji; the use of normal coordinates, we
consider the following problem. Figure 13-3a shows a diagrammatic
sketch of a three-story building. We desire to determine the vibration
characteristics of this building.
we make the following assump-
In order to proceed with the solution
tions. We
consider the building as a shcar-tijpc building; i.e., the dis-
placements that occur are horizontal translations only; rotations are
considered negligible. Furthermore, we assume that all mass may be
concentrated, or lumped, at the story levels; the conc(Mit rated masses are
m, = 2.5 ksec/m.
..<5
;!!i
= 8,000 k/ in. *i
I
Jt2= 10,000 k/ in.
= 4
m^Q
/
/
*3 = 3,000 k/ in. /
/
/
/
X
I
t
/
•
}^//
1 kip=lk = l,0001b
(a) (6)
Ki<;. \\\-\\
connected together by weightless shear springs. \'alues of tiie concen-

trated masses and spring constants are indicated in I'ig. \\\-.\a. The
spring constants represent the forces recjuired to cause a unit relative
displacement between any two adjacent stories.
Under these conditions we consider the building as a three-degrec-of-
freedom system which can be idealized as shown in Fig. V.VM). We take
as our generalized coordinates the three horizontal displacements of each
lumped mass, Xi, x^, and .<:i. The total kinetic energy of the system
now becomes
T = >i^(2.r,.r,2 + 3.r,2 -^ Ax,'')
and the total potential energy of the system (stored in the shear springs)
can be expressed as
V = H[S(x, - x^y + 10(:r2 - x^V + ;-5.r:,2]IO^

or I' = >^(8x,2 - HJX1X2 + 18x2- - 20.r.,X3 + 13x3-) 10»
Comparing these expressions for the kinetic and potential energies with
those given by Eqs. (13-31) and (13-32), we immediately find thatf
an = 2.5 a22 = 3 ass = 4

tti2 = a2i = ai3 = asi = a23 = a32 =
and 6ii = 8,000 622 = 18,000 633 = 13,000
612 = ^21 = -8,000 613 = 631 = 623 = 632 = -10,000

It should be noted that this particular system contains static coupling
only.
With these values the determinantal equation (13-36) becomes
5,000 - 2.57>2 -8,000

-8,000 18,000 - 3p2 -10,000 =0
-1,000 13,000 - 4p2
Expanding this determinantal equation yields a cubic algebraic equation

in 7>2, the frequency equation, which after some simplification can be
expressed as
7)6 - 12.45 X lO'p' + 32.33 X 10«p2 - 8 X 10^ =

Now following some fairly lengthy numerical calculations, we determine
the three frequency numbers (the three roots of the frequency equation)
as
pi^ = 2.77 X 102 ps' = 32.23 X 10^ p^^ = 90.55 X 10^
all expressed in units of radians per second squared.

Substitution J of the normal frequencies into Eq. (13-35) and again
some lengthy calculation lead to the amplitude ratios Ajk. For example,
for fc = 1 and pk~ = Pi" = 2.77 X 10-, and using the first n — 1, or 2,
of the three equations of Eq. (13-35),
- 2.77 X 102 X 2.5) .4 n - 8,000.4 21 -
(8,000
-8,000^11 + (18,000 - 2.77 X 10^ X 3).42i - 10,000.431 =
Since the amplitude ratios are only relative values, we may arbitrarily
set Ai\ = 1.000, whereupon, solving the above two equations,
Aii= 1.000 A 21 = 0.913 .431 = 0.768
t The coefficients ai, and hij, when arranged in an array, arc usually referred to,
by analogy with the single-degree-of-freedom case, as the ?/;c.s,s, or inertia, matrix and
the stiffness matrix, respectively.
X Actually, there are more efficient methods for the calculation of the normal fre-
(|uencies and mode shapes than the straightforward apjilications of the general princi-
ples used here. These numerical methods are usually iterative in nature and often
produce the frequencies and mode shapes sinuiltaiieously. It can be appreciated
that these procedures constitute a major cITort in flie study of inulti-degree-of-free-
dom systems.
Sec. 13-7] n-degree-()f-freedom syslenis; nornuil cnnrdinales 355
111 a similar fashion, by letting /.• = 2 and '.i, respectively, we determine

.4 12 = -1.000 A 22 = 0.009 A,.> = ().S()7
.4,3= 1 .000 .4 = -
,,.H
'
1 .
7.-) 1 A r, = 0. 7.31
These mode shapes are repres(Mit(>(l in I'iir. i;^-4.
F = -
-=7
L
-:i t
I
\
..} --•I
"; "7
I
/ I
/
/ / /
I
/
1st mode 2d mode 3d mode

V\y\. 13-4
For convenience, we now normalize the mode shapes according to

i:q. (i;-5-43):
.4„ = 0.;^()8 .421 = 0.336 A 31 = 0.283

.4 12 = -0.443 .422 = 0.004 .432 = 0.3o8
.4 13 = 0.2()8 .4 23 = -0.4(19 Ax, = 0.202
Thu.s the general solution for free vibrations following E(j. (13-37) can
be written
xi = 0.3()8r', sin (pi« -f 0i) - 0.443C2 sin {'p-il -f 02)

-f 0.2()8C3 sin {vzl + 03)
X2 = 0.3.3()r, sin (p,/ + 0,) + 0.004r.. sin (p.< -|- 0,.)
- 0.4()9C3 sin (p:,f -f- 03)

xz = 0.283Ci sin {jpxl + 0i) + 0.358C2 sin {-p-il + 02)
At O.202C3 sin {jp-î + 03)
For any given set of six initial conditions the six undetermined con-
stants C'l, C2, C's, 01, 02. and 03 can be easily determined by lOcjs. (13-47)
and (13-48). Suppo.se, for example, that Xi = ro; all other initial veloc-
ities and all initial di.splacements vanish. Then

1
C\ cos 01 = Wo (2.5 X 0.308) Ci sin 0i =
\/2.77 X 102
C2 cos 02 = To (2..") X 0.443) C2 sill 0., =

\/'^2.23 X 10'
C3 cos 03 = —7==
X
V90.55 102
j'o (2.5 X 0.2()8) C'3 sin 03 =
a solution of which is
4>i = <l>2
= 4>3 =
Ci = 0.054 C2 = 0.0195 C3 = 0.007
The final solution for free vibrations therefore becomes
.Ti = O.O2OV0 sin Pit - 0.0086^0 sin p^t + 0.00192^0 sin pst
X2 = 0.018^0 sin p^t + 0.00008vo sin p2t - 0.0033ro sin p-^t
Xz = 0.015^0 sin pit + O.OOTOi'o sin p2t + 0.0014yo sin p^t
We shall now illustrate the same solution in terms of the normal

coordinates. From Eq. (13-54), the normal coordinate solutions are
x[ = Ci sin {pit + 0i)

X2 = C2 sin {p2t + </>2)
x'^ = Cz sin {p^t + 03)
and from the transformation equations of Eq. (13-50) the initial condi-
tions become
x'M = xm = x'M =
x[(0) = 2.5 X 0.3682;o
i:'2(0) = -2.5 X 0.443yo
^'3(0) = 2.5 X 0.268^0
Hence the constants of integration become
</>l = </>2 = 03 =
C\ = 0.054yo C2 = 0.0195yo C3 = O.OOTi'o
and the normal coordinates are expressed as
x'l = 0. 054^0 sin p\t
x'l = 0.0195^0 sin p2t

x'z = 0.007^0 sin pst
The total solution for free vibrations can now be determined by the
use of Eq. (13-49). It is easily verified that exactly the same solution
as before will be obtained.
PROBLEMS
13-1. Suppose a particle were constrained to remain in contact with the surface
f{x,y,z) = constant. Derive the extremum conditions on the potential energy for
the equilibrium of the particle.
13-2. For the system shown, determine the force P necessary to maintain equi-
librium.
Problems 357
In.. V i;}-2
l.'}-3. In Example 13-1, suppose that kr^/Wl = 'Ki. What arc the possible ecpii-
libriiiiii eonfinurations and their corresponding classifications?
13-4. Two weightless bars of length / are connected together hy a frictionless pin.
The system issuj)ported hy a spring at the center and two frictionless rollers at the
ends. P'ind the largest value of F for stable equilibrium. The si)ring is undeformed
at the position shown.
^:r
FiG. P 13-1
13-5. Using Lagrange's efpiat ions, find the acceleration in I'rob. i'J-'Jl.
13-6. Using Lagrange's equations, find the acceleration in I'rob. 12-25.

13-7. Using Lagrange's equations, find the acceleration in Prob. 12-29.
13-8. Using Lagrange's equations, derive the equations of motion in cylindrical
coordinates.
13-9. A
double i)endulum consists of two equal masses attached to the ends of two
mas-sless strings of equal length. Using Lagrange's equations, derive the dilTerentiai
equations for small oscillations of the pendulum. Find the natural frequencies and
amplitude ratios of the two modes of vibrations.
Fk;. P 13-'.»
13-10. A
two-degree-of-freedom system consists of a mass ini suspended from a
A;i and a second mass mo coupled to mi by a spring of stiffness kz.
spring of stiffness
Find the natural frequencies and amplitude ratios of the two modes of vibrations.
m2
Fig. P 13-10
13-11. A uniform, slender bar of mass m and length I is supported by two springs
of equal spring constant k. The springs are situated symmetrically at a distance a
from the mid-point of the bar. Taking the vertical displacement of the mass center
of the bar and the angle of inclination of the bar as the two generalized coordinates,
write the differential equations by Lagrange's method and find the natural frequencies
of the vibration.
Fi<i. P 13-11
13-12. In the figure shown, body A, of mass nii, is free to slide in a horizontal
groove. A weightle-ss, rigid bar AB

hinged without friction at its one end at »ii.
is
The other end B of the bar is attached to a second mass ?«>. A spring of spring con-
stant k is connected to bar AB at C, the other end of the spring being fixed. If
AB = I and AC = a, write the differential equations governing tlie oscillation.
13-13. A double iK'iululum is confined fo move in a plane. It consists of a rigid
body .1 hinged at <) and a rigid body B hinged to A at 0'. Tiie mass center of body
.1 is at ('a, and that of liody B at ('b. OCa = a, O'Cb = b, 00' = /. Tiu^ innineiit
of inertia of A with respect to is /i, and that of B with respect to O' is l-i. Both /'s
Problems 359
are about axes perpendicular to the plane of motion. Tiie masses of A and li are hi,
and mi, rospeetively. Derive tlie differential equations govertiinR small oscillations
of the system by means of LagranRian etiuations. (Hint: Use the two angles and ^
as the generalized coordinates.)
Flc. P liMli l-i(i. P 13-14
13-1 1. For the system shown, determine the e()Uatioiis of motion for small oscil-
lations. The mass of the compound pendulum is m>; its moment of inertia about the
center of ma.ss is 1 2- The car of mass itii is constrained to move horizontally.
13-15. In Prob. 13-10, suppose A-, = 100 lb/in., ki = 200 lb/in., w, = 8 lb secVft,
and mo = 12 Find the natural frecjuencies. the amplitude ratios, the
lb sec^/ft.
normalized mode shapes, the normal coordinates, and the respon.se of the system if
motion is initiated by giving ni-, an initial (î^spla(•(ment of 1 in.
13-16. Verifj' that the mode shapes of Example 13-6 are orthogonal. What are
the normal coordinates in this ca.se?
13-17. Verify that the mode shapes and the normalized mode shapes of Example
13-7 are orthogonal. Suppose in this ca.se motion were initiated by imparting an
initial speed Vo to the coordinate x>. What would be the resjwn.se of the .system?
APPENDIX A
Properties of Plane Geometrie Shapes
The following nomenclature will be used in this appendix:
X, y, z = axes to which shape is referred

C = centroid of area
.4 = area of geometric shape
X, y, z = coordinates of centroid of area referred tox, y, z system
Iz, ly = area moments taken with respect to x, y axes

of inertia
Ixy = area product of inertia taken with respect to x, y axes
Jxc, Jyc = area moments of inertia taken with respect to axes parallel
to X, y axes and passing through centroid
Ixa/c = area product of inertia taken with respect to axes parallel to
X, y axes and passing through centroid
Al
A2 Properties of plane geometric shapes
Shape
ProjH'rhrs of phu\e (fconirlric shaf)es A3
Shape Area and centroid
*y
-I =
I = —
4r
y = z
Quarter circle
/ith degree parabola
b ^\/i
nth-degree parabola
A4 Properties of plane geometric shapes
Shape Area and centroid Moment and product of inertia
he =- sin 261
he = -^ (902 + 96* sin d cos
16 sin2 e)
^XCVC "
Circular segment
.4r2 2 sin^ 6 cos d

h =
n 2e\
3U -
^r^ 2 sin^ 6 cos d

h =
4
1 H
sin 2e
Circular sector
^xci/c ~ U
APPENDIX B
Properties of Solid Homogeneous Bodies
The following nomenclature will be used in this appendix:
X, y, z ^ axes to which body is referred

C = center of mass of body
V = v(jlume of body
m = pV = mass of body, where p = mass density
Xc, yc, Zc = coordinates of mass center of body referred tox, y, z system
Izz, lyy, Jzz = mass moments of inertia taiên with respect to x, y, z axes
fzy. I x:. I„2 = mass products of inertia taken with respect to x, y, z axes
^*x' ^tu^ i*z — mass moments of inertia taken with respect to axes pass-
ing through mass center and parallel to x, y, z axes
^ty, t*zz' Jlz = rnass products of inertia taken with respect to axes passing
through mass center and parallel to x, y, z axes
A5
A6 Properties of solid homogeneous bodies
Volume and center

Body Moment and product of inertia
of mass
i^y
2
?/c = 12
zc = /* = /* =0
rz Thin rod
mr^ — sin Q cos 9

/.. =
r sin Q
mr^ Q + sin 6 cos B
2/c =
2C =
Ixy = /x. = /.. =
Thin circular rod
• c
'
'^s b
Rectangular prism
Right rectangular pyramid
y
Properties of solid homogeneous bodies A7
Volume and coiitcr

Body Miinicnt and |)r<)diict of iiicrti
of niiiss
V = ttH.
xc =
L
yc = - /* = mil
zc = I*
'xy = /*
'it = /*
'III
=0'-'
Right circular cylinder
4xr'
V =
xc = Q
yc =0
zc =
* Sphere
27rr'
V =
xc
A8 Properties of solid homogeneous bodies
Volume and center

Body
Answers to Selected Problems
CHAPTER 1
1-3. A + B = lOi - Gj - 14k, A - B = 2i - 2j + '2k, A = ii.Gl, li = 18.22

1-5. |A + B| = 34.7 lb; angle between the x axis and resultant force = 84°20'
1-6. F, = Gi - 2j + 3k, Fo = 4j - 6k
(a) |F, + F,] = 7 lb, cos a = ^f, cos = ^j, cos 7 = - H /3
(6) |F, - F2I = \/l53 lb, cos a = 6/\/l53, cos = -G/y/TEs, /3
cos 7 = 9/\/l53
(c) IF2 - F,| = \/T53 lb, cos a = -6/V153, cos = G/Vl-^î. /3
cos 7 = -9/\/l53
1-7. a = cos-' 4^.

-4
/3 = cos-' ^>y
A
= cos"' 4"''
-4
^ = V/iTT^VTX*
1-9. A'^ = i4x cos + i4„ sin 0, A[ = —Ax sin + --li/ cos 6
1-10. .4a = ^lOi AyOy AtOt + +
1-11. Force component along rod =289 lb; force component perpendicular to
rod =
80.4 lb
1-12. 216.5 mph, 36°52'
1-14. n = (l/>/3)(i -j - k)
1 IX ^ ~ ^» _ y - ^v _ 2 - ^' _ K
CIIAI'TKK 2
2-1. AH = 51.8 lb, AC = 73.4 lb

2-3. AH = 173.2 lb (compr.), BC = 200 lb (tensO
2-6. AB = \/2 JK (compr.), ^C = AD = W/V-i (tens.)

2-7. = P/\/G (compr.); ba.se bars = P/'i \/G (ten.s.)
Ix.'KS
2-8. AB = AD = 0.87GIK (compr.), AC = IA2\V (tens.)
2-9. AB = BC = P/2 (compr.), ZfZ^' = CC = ^ (compr.), CD' = V^ Cii' =
2-10, Radial bars = T (tens.); circumferential bars = T (coinpr.)
2-11. a = G°35'
2-12. d = 8M.r/vTTW
2-13. P = 496 1b
AO
AlO Answers to selected problems
CHAPTER 3
3-1. COS e = A A+ AyBy + A^,

VAJ + Ay^ + A,2 a/Sx^ + By^ + 5,2
3-2. (a) -13/V5, (b) cos-i (-13/V190), (c) (l/V21)(-4i - 2j + k)
3-3. (a) 4/aA4 lb, (6) Vl22 ft lb
3-12. -2
3-13. 2%
3-14. Ml = (2Fi/V29)(8i - 6j + 3k)
M2 = M^2(-4i +3j + 10k)
Ms = -%Wh _
3-16. V = (4,0007r/\/3) (2i - j - k), w = 5,6507r in./min
CHAPTER 4
4-1. Vl = 982 lb, Vr = 1,750 lb, //z. = 1,000 lb

4-2. Vl = 1,200 lb. Ml = -3,200 ft lb, Hl =
4-3. fc = 271 lb (tens.), fg = 2,365 lb (compr.)
4-4. AB = 2CD_= 4P/(5 VS) (compr.), ^F = V3 P/5 (tens.)
4-6. = 9 V29 W/92, F2 = 21TF/46
ii'i
4-7. Cy = -131.2 lb, a

= -32.5 lb, Dy = -243.8 lb, D^ = -10.8 lb, 6.66 hp
4-8. Tb = 1,300 lb, Tc = 3,290 lb, A^ = -2,112 lb. Ay = -206 lb, A, = 2,487 lb
4-10. Inclined bars = 4Af/3a (tens.), vertical bars = 2ilf/3a (compr.)
4-11. (a) A^ = 0.866PF, Ay = W/2, B^ = 0, By = W/2
(b) A^ = 0.433 PF, Ay = W/4, B^ = 0, By = Tr/2
4-12. (a) Ax = 0, 4j, = -(TF +P), A^ = 0, D, = -P, />, = W, Cy = W +P
(6) AP = \/2 (TF + P) (compr.)
4-15. (a) Fr = -4i + j + 4k, Mr = -8i - 12j - 3k
(b) MrL = %3(-4i + j + 4k), x= -4y - ^îs = 2 + 103,^3
(c) No
4-16. (a) Fr = P(0.707i + 0.293J + k), M^ = Fa{-j - 0.707k)
(6) MrL = Pa(-0.446i - 0.185J - 0.631k)
X = 2Aliy - 0.1075a) = 0.707(2 + 1.15a)
4-17. a = b +c
CHAPTER 5
5-1. yc = h/4 (from base)
5-2. X =
n
—^
-\-2
b, y = ^
2«, +j—12^
r
5-4. Vl = -Wol/S, Vr = 9u;o//8, Hl =

5-7. Ta = Tb = {2Wi + 1^2)76, Tc = {Wi + 2W2)/'3
5-8. y = 2r/ir (from diameter)
.5-9. X = 0.328a, y = ().548a, z = 0.328a
5-10. W1/W2 = (7r/2) cot a
5-11. W,/W2 = 2 cot a
5-12. (a) 10.25 psi, {!,) lO.O psi, (r) 9.37 psi
5-13. 15.68 in.
5-14. U.03 in. water

5-16. 499 lb toward left and horizontal
An.sivers lo seleclcd f)r()hl('rus Al 1
5-17. Fr = 3,535 Ih (upper end), Ft = 5,4-15 Ih (lower end)

5-18. (a) 21'.) ft rinht of H, (h) P„„„ = 2<t,4(K) psf, f„», = 48,400 psf
5-19. /i' = 240 cos e sin- 0. 13.5° < 6 < .>1.G°
5-20. 27G in. < < li33 in. /
5-21. (/ > 7 ft
5-22. 0..=><)25 < sp Kr < 1
5-23. («) Fb = >2(»-r*L)-), {/>) .1^ = -Hr'Ly
CIIAPTKK 6
6-4. if - p*r =
6-5. (a) if - p=r = 0, (h) r - p=r =
6-6. 1/p = 30/(401)^2
6-7. 1-^=6+ 2ct, a, = 2f
6-8. y = a/bHl - e*') + (j'o + a/b)t + i/o
6-10. V = vA + 2xv4, a = 2r^2j, v = i-^ v'l + 4t* e,

ivs'x ,
2r.'
6-1 i. (a) f = (I'Vr) sin' (i>, 4> = {'2v^/r-) sin </> eo.s <^
(b) a =
6-1... (a) a, = 0. «,. = ^,)..^
^^, _^
O COS g \I
/; lA vp
6-16. = — aa> fl/i
sin ^ ( 1 H
V V ,
- a' sin' 0/
a r <* cos 9
= "
, 1 "I
H
I
Of a'co' sni' d
- "" ''''
"^
(/' a' sin' g)'^ (/'-a'sin'e)!'J
6-17. I'/. = — a)a[sin — {a/21) sin 26], ap = — a>-rt[c()s d + (a/t) cos 2S]
6-18. Vp = — aw sin g. ap = — ow' cos 5
6-19. (a) 01. = 19y°28', Or = 340°32'; (b) 0.045
rilVPTKK 7
7-3. r, = i%,r„ H,
7-4. r, = -}4,r, = -%
7-5. v^ = 0, a.4 = wVy, vfl = —uia + b)i, an = — w(o + b)i — w'oj
vc = — w(a sin + />)i + ua cos j </> </>
7-6. as (21/4)i - 5k in./.sec'

7-7. rAUA + fBuiB = (rA + rB)(^c
7-8. Vr = rw, at = ra, Ur = rû^/ili — r)
7-12. (a) wcD = 0, (b) o)Bc = 25 radians/sec', (r) 2 ft rij^lit of (' on the extension
of BC
7-13. (a) Vc = 34,850 in./min, (6) «Bf = 1,4.50 radians/niin
(c) ac = -28,000.000 in./min», (d) ubc = -20,000,000 radians/min»
(e) r^ = 18 in., r„ = ti \/3 in. (relative to li)
7-14. (d.\/d()„i = -w/ X A
7-1.1. t\r/(lt = U6 X r
7-16. (d\i'/dl)„i = irwCcos tf ei + sin d)
A 12 Answers to selected problems
4
7-18. as = -5/2i - 2 V3j + 2 V3 k ft/sec^, as = ^M ft/sec^
7-19. \ = —4>b sin <!> i + 4>b cos j — a)(a + b (^ cos 0)k
a = — co-(a
[ + 6 cos <^) — (p</>' cos pi)^^ cos (j> + p^^' sin pt b sin 0]i
— [ip4>' cos 7)<)^& sin <^ + p'^(t>' sin pi5 6 cos </)]j
+ i2o:p4>' cos pi 6 sin 0)k, where ^ = </>o + </>' sin pi

7-20. 12,060i - 23,310j
CHAPTER 8
8-1. Repulsive force proportional to displacement

8-2. (xo/w^) cosh o:t, where w^ = c/m
8-3. (a) {m/2k) log (1 + kv^lmg), (6) vj^\ + A:2;oVw3
8-4. 73.5 ft
8-5. (a) 62.2 ft, (&) 2.865 sec '
8-6. 7 miles/sec • ,
8-7. {mg/kyi- \
8-9. -y
W (2?;2 + ^)
i
8-10. xVô^ + ?/Â;2>o^ = 1 I
8-14. W= 872 ft lb
I
8-15. No j
8-16. V = —xyz'^ -\- y -\- 2z constant

-{- i
8-19. r > a, = 4> —Gm\/r; r < a, = constant

<l>
8-22. (a) ,
"^ , (6) (^- ^V^'\ (c) / = i- f ,
^
V^ |
r = 2.(^^ÎHZZ«!)^^
8-23. cos <^ = %
8-24. (a) ^ + ^% = 0, (6) / = f \/^
8-25. West bank is 0.337 ft higher
8-26. (a) cos 6 = gf/wâ, (6) T = 27r/(co sin 6)
8-27. (a) y = \/gh., (b) f = I-^tt y/'ig/h cos a
CHAPTER 9
9-1. An = VaJ + bn\ n = tan"! (fe„/a„)

9-2. {a)k=_ kiki/iki + ki), (b) k = ki k-., + (c) k = ki + A'a
9-3. -y/nkg/a
9-4. r = 27r V//2j7
9-5. (1 + 4a2x2).c + 4o2a:.t2 + 2ar/.r = 0, / = \/2^/2-n-
9-6. p = yf^^S
9.7. p = J^M^
V'olF
9-9. (a) 0.084 ft, (/>) O.OOOOOOO ft, (r) 17.67 ton sec/ft
9-13. (a) 0.01985 lb sec/in., (6) 13.9 radians/sec, (c) 13.88 radians/sec
(d) 0.2765 lb
—
Ansioers (o srlrrled j}r()hlenis Al3
= Fv — , . Fp
«>-! i. .t -— (\ cos pi ) sin pl\ + -p sin pi cos pli
F — —F ,
= r-
. .
X (1 t'os pt
,
) cos pl\ -\- sin pi sm />/i
where t = t' -\- t\ ami / > /'
y-l.-,. (a) 0.0110 in., (h) 115 lb
9-16. Q + X ^ + ^' = X'

^ = -ii'C'e-*'"*'-" (cos <ji
+^ «i" '7^) + EC
./I «*
9-19. VHp < w < vli P

9-20. (a) ^" (1 - cos pi), (b) ^(^ - ^^P') f..r I < <,
+ - - sin pt\\ for >

^ 1 1 -^ [sin pî /,) / /,
CIIM'I IK 10
10-1. T = 2fo7/, y = y^ — y^gr- + /i (/i = lici^lit of shell iit hish point; io = hori-
zontal speed just before explosion: ?/o = vertical sjx'cd imparted to each
splinter by explosion)
10-2. a, = -2(/V2n-, - H'2)/(1F., + AWi), a-. = {/(2ir, - ir,)/(ir-. + AWx)
T = iWxWJiWo + 4H',)
10-3. (a) V = 25. li - 41.9j fps, {b) same as (a), (c) before impact, 7'e„r = 240
ton ft, 7',ruck = 208 ton ft; after impact, Tear = 74 ton ft, 7'truck = 185 ton ft
— — + nioVi 7»ie(t;i —
,
10-4. (a)
, ,
Vi
/
= niiVi -H m-fa
1
mêivx Vi)
> i'2
'
= niiVi
——+ -[—
Vi)
,.> 1 — e* mim2 , ,,
2 7ni -|- 7/12
10-5. 81,000 lb
10-8. p = y/ka^/ml^ - ^/Z
10-9. 
= F//h(/, <^ = ^2Fiytul
10-10. (a) i.î;(4i + 2j + k), (/*) 2)nv( - j - k), (r) 2.;„„.(i - 5j - lik)
10-11. xc = a, yc = -Hyf^ + a^»)^. ^ = ô^, L.c = (wia- + mJ>'^)do
Lo = Lc — vigal uia^du -\-
10-12. ^(m, + m2)t.c'' + '^"^"if

/Hi -J" "l2
"''
H^ +UJ
10-1.3. V = -^^—^^-=;
\/2ff/(l - cos a)
10-15. m (/t;/d< = wgr sin a

10-16. mdv/dt — vdm/dl = »/</ sin a
10-17. — ^^
rr~ H ^m a
nio — kI
10-20. 0,200 fps

10-21. L = 0.0088 cm
CIIAPTKH 11
11-2. /:, = /*.= -^ (.^ + ;i^),

C= ^^
mr^
11-3. /:. = ^ ^^ /:„ = /:. = ^ CJ/-^ + a^)

A14 A nswers to selected problems
11-5. Iaa = %nir'^, Ibb = ^Homr'^
11-8. /:. = - 771/
(5E^ + U^), /:, = ——. Yilt
(4/^ - ZR^)
lo 16 V3
11-9. /max = 5, /min = 3
/xy = /yx = ahm/A, ly^ = hy = —hem/ 4:, hx = Ixz = —acm/A

11-13. I = \/^R
6/1^
11-14. Mowa' a2 + /i2
11-20. 7xx = b(±I^ + bh ^^^, _ g^^ ^ 3^,^
= -^^^^^^' - = - - -
/..
^"^
+^ (4^,2 66/ii + 3/^1^), /x,
^ (6 /iOCa h)
11-21. /xx = -... ,. > /„„ =

3(3n + .
1) ^'' n + 3
11-22. /xx = 80.8 in.*, lyy = 38.8 in.«
/xy = -32.3 in. = -28°29' 4, cl>
/j^ = 98.6 in.^ /,, = 21.0 in.*

11-23. (a) /xx = 2,957 in.*, Tyy = 223 in.*
(6) /;, = 2,274 in.*, l'^^ = 906 in.*
CHAPTER 12
12-1. cos-i V^^ = 16°7'

17°22' with normal to face of disk
12-2. (a) 44.3 ft lb sec,
(6) 1,215 ft lb, (c) 2,700 ft lb
12-3. (a) t = 10x/(co Vll)> (fc) « = 25°12', (c) ?«a2wV8

12-4. T = ^^[/iC^^ 02 sin2 + 0) 4- /(.^ cos 6 + ip)^
12-6. 9,910 lb
12-7. 4,125 ft lb
12-9. w' = w/7
12-11. T = 1,200 lb, P = 400 lb
12-14. F = Wd/h, left wheel
12-15. Ax = 80 lb, ^„ = 66.4 lb, Cx = 300 lb, Cy = 287.8 lb
Dx = -340 lb, Dy = 352.2 lb
12-16. Show that vertical acceleration of cup is equal to ^2f7
12-17. p = VWa/h
12-18. Radi\is of gyration = 4.85 in., BC = 9.66 in.
Aniga 1
12-19. (a) <>2 = ^^2 ~^sin2-
.
,
</>, (6) sin-i (a,o y/h/mga),
y 2
(f) Wo- > -inuja/lo
12-20. Max. reaction = 25,180 lb
12-21. (Ia = 10.25 in., = 179°18'
(j>a
(1,1 = 3.11 in., 0a = -36°35'

12-22. a < cos-i (d,A/o)
12-23. T = W^/3, ^c = -Ha
W^.*(/«in2a
12-24. x^= -
3(TF^ + Wb) - '2Wa cos2 a
.
Answers lo sricrlcd pmhlcins Al.')
(wi — mi)g
12-26. -Cl =
Ttii + nii + I/R*
12-27. (b) (c) sin a

gCAl^k cos a — sin a)
•(J
12-2«. 4> + HiR9 - r)

sin <^ =
> . */6fl(l —
- cos
CO <t>o){mi -h 'im.)
12-29.
(/2 - r)(2»Hi + y/Hj)
=
— 3(; sin </>o (»ii + 2ot2)
(6) <i.
(72 - r)(2m, + Owj)
^'"'
' 2^ \(/e - r)(2m, + Oj/Ij)
12-.'il, /e^ = IL (li eos <^ - lyjiB = -r sin <^ (9 cos <^ - G)
4 4
CIIAPTKIt U
I.J-I.
_dfdv ^dVdf ^^^
_dfdV_^dVdf^^
dz i>x dz di '
dz dy dz dy
i:i-2. 02.5 U)
i.j-.i. e = 0, unstal)le; 6 = 73°3', stable
i:{-4. P = A-//2
W aQ sin 2a
3(»rx + H^b) - 2Pf^cos«a
l.{-6. 6} = mâ cos </>
/fl + 2mRa sin
7^^
— 3g sin <t>
(nil + 2?na)
13-7.
(R - r)(2ffi, + !)/«.)
13-9. = (2 _ ^/OA 9,
13-1(1. ,..2 = 1 (kL+J^ +h.) + \/l /'^lIA^ _|. ^y _ Ml

'
2\ ~
mi nii/ ^'4\ /«i mj/ JMimj
"
\-li/i.2 2 2»(j '-IV »ii »f2/ niimj
1 3- 1 I mx + 2A-J- 0, /(> + = 2Aa20 =

13-12. (wi + m>)x + m>l4> kx ka(t> = -\- -{-
molx + m-il^^ + kax + + mi(jl<t> = ka'^<t>
13-13. (/, + mP) + mjilyi/ + (wiiC + = »(,./)(/</)
mM + /2^ + = /"•.-'^f/iA
13-14. (/, + w,a*)<ji + (m-.(/a + - klx = kl'')<t>
TOiX + A-x — A;/0 =

1
Judex
Al.soliitc iucclcration. i:?, 1-41, 145 Airplane, 2S()

AKsolutc cnordinatcs, i:i, \:V2, 14U, 181 in level flight, 288
[Set also Frames of reference) Alternating fonre, 157
Absolute pressure, 81) Alternating tor(|Ue, 213
Al)s<)lute teiuperature. 240 \V (amplitude factor), 204, 208, 209
Absolute velocity. V2H, 142 Amplification factor (.vr Amplitude
Aeeeleration, 12, 14, 111 factor)
absolute, i:i, 144, 145 Amplitude, of alternating force, 157//., 158
angular, 133 in nonlinear vibrations, 218
apparent, 145 normalized, 349
centripetal, 135 of simple harmonic motion. 159
of Coriolis, 145, 147 of simple jM-ndulum. 17S, 179
in cvlindrieal coordinates, 114 of steady-state current, 209
of gravity, 81 of vil)rations, forced d;impetl, 205
of jjoint on rigid l)ody, radial, 135 forced undamix-d, 204
tangential, 134 free damped, 199, 200
in rectangular coordinates, 1 1 free undamped, 195, 343 349
referred to moving reference frame, resonant, 209
141-147 steady-state, 208
relative, 145, 147 (See also Vibration)
in spliericalcoordinates, IK) 11!) Amiiiitude factor (AK). 204, 20S. 209
in tangential and normal coordinates, Am|)litude ratio, 344-34(5
113 .Analogue computer, 19:5
of vehicle, 145, 147 .\nalogy between mechanical and
Acceleration-time curve, 100 electrical systems, 193
A<<-eleronieter, 20(i Angular acceleration, 133
Action and reaction, law, 13, 53, 224 .\iigular displacement . inlinitesinuil, 125
Addition of vectors, 3 Angular momentinn, of |)article, 104, 227
law for, associative, 4 principle, 228
commutative, 4 of rigid body, 278
parallelogram, 3 H'lative, to fixed point, 279
by method of components, 7 to imi.ss center, 278, 279
polygon rule for, 4 of system of particles, 228
Aerodynamic forces, 234, 2H8, 28'J cons<Tvation theorem, 228
Air, resistance, 184 principle, 228
siH'cilic weight, 101 relative, to arbitrary |)oint, 230
(Sep nlxo (las) to mass center, 230
p
Air-borne vehicle, 280-290 total, 228, 229, 231
A17
A18 Index
Angular speed, 45, 115, 127 Bars, in frame, 61
Angular velocity, 45 hinged, 18, 55, 56
of earth, 181, 308 lateral load on, 18, 60
perturbation, 291 rigid, 18
in plane motion, 137 in simple structure, 17-20
of rigid body, 127, 133 Base point, 125-131
of rotating frame, 144 change, 126, 129
{See also Kinematics of rigid body) (See also Kinematics of rigid body)
Apogee, 166, 169, 184, 185 Beam, 72, 99
Apparent acceleration, 145 Bearings, 62, 63, 73
Apparent velocity, 143 reactions in, dynamic, 313-316
Applied force (see External forces) static, 62, 63, 313«..
Approximate solution, of air-borne Beats, 346
290
vehicle, Belleville spring, 218
of gyrocompass, 309 Belt tension, 63
of gyroscope, 306 Binomial theorem, 178
of simple pendulum, 178-180 Binormal, 109
Archimedes, principle of, 96, 103 {See also Curves in space)
Areal speed, 163, 167 Blast force, 214
Array, 250, 257, 279 Body axes, 132, 278-286
elements, 2b7n. (<See also Moving frames of reference)
Artificial satellite (see Satellite) Body centrode, 131
Associative property, of scalar product, Body cone, 297
31 (See also Kinematics of rigid body)
of vector addition, 4 Boltzmann constant, 241
Atmospheric pressure, 89 Bound vector, 2
Attraction, electrostatic, 162 Boyle's law, 220, 240
gravitational, 162 (See also Gas)
Newton's law, 164, 174 Buoyancy, 96
potential for, 175 Buoyant force, 96, 97
Average ensemble, 245
Average kinetic energy of molecule, 240
Average quantity, 244, 245 Capacitance, 193, 202
Average time, 238 Capacitive reactance, 210
Average total kinetic energy, 238 Capacitor, 202
Axes, body, 132, 278-286 Cardan, J., 294
of inertia, principal, 263-264, 280 Cardan's suspension, 294, 307
location of, 265 Cartesian coordinates, 5, 111
instantaneous, 127 Cartesian tensor, 271
of rotation, 45, 123-125 Cauchy, A. L., 263
rotation of, 256-260 Cauchy's inertia ellipsoid, 262-267
of screw, 130 Center, of buoyancy, 96
in plane motion, 131 of curvature, 108, 113, 119
of wrench, 69 of force, 162
{See also Coordinates; Fixed axes; of gravity, 79, 81, 224
Moving frames of reference) of mass, 80, 81, 224
Axial components of velocity and acceler- of pressure, 91, 93
ation, 115 of rotation, 131, 136
of system of parallel forces, 77
Central force, general, 162, 229
Balancing, dynamic, 313, 315 proportional, to distance, 184
of rotating bodies, 312-316 to inverse sciuare of distance, 164,
static, 312, 315 174
{See also Bearings, reactions in) {See also Gravitational attraction)
Ball-and-socket joint, 55, 63 Central force field, 1()2, 174, 175
Ballistic pendulum, 248 {See also Central force)
Bars, forces in, 21 Central impact, 246
Index A 19
CtMitrifiinal tone, 18J ConJiK)nents, of velocity, 1 1-1 19
1
Centrifugal pi'iKluluiii, 180 {See also E(iuations. of e(|uilibriuni,

Centrifugal j)unip runner. lo2 of motion)
Centripetal aeceleration, i:io Composite bodies, centers of mass, 81, 85
Cent rode, body, liil io:i
space, I'M centroids, 81
Centroids. of composite l)odie.s, 8 1. 85 moments of inertia, 25('), 272
determination by integration. 79 Comjjosition, of angidar velocities, 127
of lines. 81 of couples, 00
of plane geometric shapes, A1-A4 of finite rotations, 1, 120
of surfaces. 80 of infinitesimal rotations, 120
of volumes, 80 (Set also Addition of vectors)
Change of base point, IJCi, l'2\) Compound pendulum, 322. 359
Characteristic e(iuation, H(4 Compression, 17
(iSVf aluo Fre<iuency ('(piation) (.Sec also Korc<'s)
Charge, in electric circuit, liCJ Computing maciiine, 157, 220, 305
in electrostatics, 162 analogue, l!»3
Chasles, M., V2o Concurrent angular velocities, 127
Chasles' theorem, 125, 12(> Concurrent for<«'s, Ki, 24, 57, (i9
Circular cylinder, moments of inertia, 254 Condenser (capacitor), 202
rolling, :il7, MS Cone, body. 297
Circular fre<iuency, 158, 100, lUO, l'J5 inertia ellipsoid for, 200
dampeil, 200 space, 297
Circular orbit, 108, 109 stability, 97
orbital speed for, 109 Configuration of system, 188, 320
Clausius, R. J. E., 2:i8 (<SVe also Degrees of freedom)
Coefficient, of critical damping, 198 Conic orbit, 100
of friction, kinetic, 22 Conic section, 100
static, 22 Connecting rod, forces on, 10
of restitution, 240 motion, 121, 130
variable, for vii)rating body, 210 (See also Four-bar linkage)
of viscosity, 242 Conservation, of mechanical energy, i)ar-
of viscous damping, 150, 192 ticle, 173-170
Colatitude, 118 rigid body, general, 2Sl
[See also Coordinates) I)iv()te(l, 285
Colinear forces, 18, 00 in plane motion, 317
Collisions, of molecules, 239 of momentum, angular, 228
of particles, 220 linear, 225
(See also Direct impact) Conservative force fielil, :<S. 172 17<»
Commutative |)roperty, of scalar examples, 173-170
product, :il potential energy for, 172
of vector addition, 4 Conservative forces, 281, 329
of vector product, iil Constant of gravitation, 104
CVjmi)()neiit motions. 111 Constant vector, 105, 103, 105
Comj)oneiits, of acceleration, 111 119 Constants of integration, 154, 155
addition by method of, 7, 8 (See also IO(|Uations, of motion)
of angular momentum, 231, 280 Cxjiistrained motion, 170
of angular velocity, i:W Constraints, 188
of derivatives of a vector, 100 forces due to, 170
of force, 15 holonomic, 320
of moment, 41 of rigid body, 50
of momentum, 225 workless, 171, 329
rectangular (see Hect angular Contact, forces, 21, 22
components) rolling, 135, 130, 318
of unit vectors, II, 258 rough, 22
of vectors, 0, 7 smooth, 22
relative to rotated axes, 258 Continuous vector function, 101
A20 Index
Convolution integral, 215 Current, 193
Cooling of object in air, 202 steady-state, 209, 210
Coordinate axes, in rectangular com- Curvature, center, 108, 113, 119
ponents, 258-260 for plane curves, 110
rotation, 256, 258 radius, 108
Coordinates, absolute, 13, 132, 140, 181 for space curves, 108
cartesian, 5, 111 Curves in space, 107-110
cylindrical, 114-116 binormal, 109
generalized, 188, 189, 326 curvature, 108, 109
moving, 139-141 direction cosines, 108
normal, 348, 351 Frenet's formulas for, 108, 109
polar, 114 osculating plane for, 108
rectangular, 5, 111 principal normal, 108
rotating, 144 tangent to, 108
spherical, 116-119 torsion, 109
tangential and normal, 113, 114 Curvilinear motion (see Kinematics)
transformation equations for, 119 Cycle, 158
Coplanar forces, equilibrium, 16, 20, 24, of vibration, 189-191
58, 59 (See also Simple harmonic motion)
equipollence, 70, 71 Cylinder, moment
of inertia, 254
for three-force body, 60 rolling, 317-318
Coplanar vector, 5 Cylindrical coordinates, 114-116

Coriolis, G., 145 transformation equations for, 119
Coriolis acceleration, 145, 147
Coriolis force, 182
Correction planes, 315 Damped harmonic function, 198
(See also Dynamic balancing)
Damped vibrations, forced, 207-210
free, 197-200
Cosine law of trigonometry, 4, 31
Coulomb, C. A., 22
Damping, coefficient, 192
critical, 198
Coulomb friction, in dynamic system, 22,
over-, 197, 198, 221
318, 319
under-, 198, 199
laws, 22
variable, 216, 218
in static system, 22-24
viscous, 156, 161, 192, 207
in vibrating system, 218
Damping ratio, 208
Couple, 65
Dashpot, 192, 211
in equipollent force systems, 67-70
Decomposition method, 6, 7
moment, 65
Decrement, logarithmic, 200
as vector, 66
Degrees of freedom, 55, 188, 189, 326
Coupled pendulums, 357, 359 n, 346-352
Coupled systems, 216 single, 190-191
forced vibrations, 351, 352
two, 216, 342-344
free vibrations, 340-351
Density, 81, 251
Coupling, 340 Derivative of vec^tor, 104, 105
dynamic, 342 components, 106
static, 342 distributive law foi-, 105
Creeping back in damped free vil)ration, in moving frame, 140, 141
197, 198 position, 108
Critically damped vil)rations, 198 and scalar, 105
Cross component, of iiionicnt , 68 unit, 108, 109, 115, 118
of velocity, 129 Deterniinaiifs, 46, 265, 3-13, 347
Cross product, 33-:i() Determinate problems, 21, 56
(iSVe aluo Vector product) {See also l'](iuat ions, of eiinilihrium)
Cubic (Hiuation, for dctcrininnl ion, of l)('\iations due to cnrth's rotation, 182,
freciuency, 35 I
183, 187
of principal axes of inertia, 265, 274 Diagonalization (reduction) of ciuadratic
Ctul of vector, J 73 forms, 351
1 1
Index \21
DiaKrnin, arccl«>ration-tiiiu'. 112 Earth, angular velocity. 181

frcH'-lMwIy. 14, Hi. IS. 24. oS, 5J» attraction, 102, 104
po.sitioii-tiiiic. 1 1 mass, 1(H)
siKH'il-timr. 1 1 mod.ls. 14. 104, 175, 181, 300
DitTomitiahlr v»M(or fuiK-tioii. MM ' motion, particle rel.ative to, 181
DilTcniitial, iHTfect, 185 .satellite around, l(>4
total. 172 precession. 300, 303
DilTcrcntial ('(juatioii, lioniom-iu'iuis. 201. rotation, to determine latitude. 310
203. 207 to determine north, 307
of motion, 15;i-157 EA'centricity, of conic orbit, KHi
nonlinear. 218 of .satellite trajectory, lt»8
particular solution. 201, 20;i, 207 Elastic .systems, viliration, 21(1. 352
of KW circuit, li>3 Electric circ\iit, 193. 202
trial .solution, 194, 197 respon.se, 209
vector, 107 {See also Charge; Indui-tance; Resist-
of vibration, 191, 194 ance)
{See also Equation.s, of motion; N'ilira- Electrical analogue computer, 193
tion) Electrical mechanical analogy, 19:{
Differential geometry, 108 Ellip.soid, inertia, 203 207, 290

Differentiation of vectors {see Derivative of I'oinsot, 295-298
of vector) Elliptic function, Ja<;ol)ian, 177
Direct impac^t, 246 Elliptic integral, complet<', 178
Direct precession, 302 Ix'gendre's, 178
Directed line .segment, 2 Elliptical orbit, 100, 1(59
Direction cosines, 7, 10, 11 Energy, conservation, 173, 170, 284
between sets of axes, 258 kineti(\ of molecule, 239
of tangent to space curve, 108 of particle, 171
Discontinuous motion, 22(j, 231 quadratic form, 341
Disk, moment of inertia, 251, 254 of rigid body, 282-285, 313, 317
rolling. 135 of system of particles, 230-238
Displacement. finit<', 120 loss due to impact, 240
of particle, 110 potential, of jiarticle. 171 175
of rigid body, Cliasles' theorem for, 125 quadratic form, 340
Euler's theorem for, 124 of rigid body, 284
finite, 120 of system of particles, 174, 340
infinitesimal, 37, 125, 127, 337 tl:eorem of equipartition. 240
virtual, 328 in vibration problems, 195
(See aluo Kinematics of rigid body) (See also Principle, of work and energy)
Distributed forces, 22, 89, 99 En.semble average, 245
Distributive property, of derivative of E(|Ual vectors, 3
vector, 105 Etjuation, of state, 240
of scalar product, 31 of work and energy (ser I'riticiple. of
of time averages, 238 work and energy)
of vector jjroduct, 34 Erjuations, of e<|uilibriurn. for |)arti(!c. 15
Disturbing force, 214 for rigid body. 50-(»0
(.SVf also Kxciting force) for sy.stems of particles, 53. 54
Dot product, 29-31 (See also Equilibrium)
(.SVf also Scalar })roduct) of motion, Lagrange's (see Lagrange's
Double f)endulum, 357, 359 e(iuations)
Drag force, 184 Newton's, 12, 153
(.SVe also Vis<'OU8 medium) of particle, in cylindrical coordi-
1 )rop hammer, 220 nates, 153
Dry Coulond) friction)
friction (see in normal and tangential coordi-
iJynamic balancing, 312-313, 315 1.53, 179
nat<-s,
Dynamic coupling, '.iA'2 one-dimensional, 151 150
Dynamics iscr Kcpiations, of motion; in re<"tangui.ir co(»rdinates, 1.53,
Particle; Uigid liody; Vibration) 180
A22 Index
Equations, of motion, of particle, rela- Field, electrostatic, 162
tive to rotating frame, 181 force, 38, 172
of rigid body, Euler's, 280 central, 162, 174
moment-free, 294, 295 conservative, 38, 172-176, 282-285
in plane motion, 316-317 gravitational, 162, 164, 174, 175
rotating about fixed axis, 311, 312 uniform, 79, 173, 174
of symmetrical body, 304 vector, 37
in translation, 310 Finite displacements of rigid body,
of small vibrations, forced, 352 123-126
free, 342, 346 Chasles' theorem for, 125
of spring-suspended mass, 156 Euler's theorem for, 124
for systems of particles, about fixed Finite rotation, 1, 126
point, 228 First moments, of area, 81
about mass center, 230 of mass, 80
of mass center, 224 of volume, 81
Equatorial moments of inertia, 298 (See also Center, of mass; Centroids)
Equatorial plane, 266 Fixed axis, equations of motion for, 311,
Equilibrium, of particle, 15, 17, 20, 330 312
procedures in, 24, 25 rotation about, velocity due to, 45, 46,
of rigid body, 56-60 127
necessary and sufficient conditions Fixed vectors, 2
for, 57, 328-331 Flexible string, 17
under three forces, 60 Floating bodies, buoyant force on, 96, 97
under two forces, 60 center of buoyancy for, 97
stability, 96, 97, 330, 331 meta center, 98
of systems of particles, 52-54 stability, 96-98
(See also Equations, of equilibrium) Fluid, 85
Equilibrium position, 156, 187, 191, 196, incompressible, 88
198 pressure in, 86, 87
Equimomental system, inertia ellipsoid forces due to, 89-91, 94, 95
for, 267 pressure-altitude relation for, 87
Equipartition-of-energy theorem, 240 shear stress in, 85, 241, 242
Equipollent force systems, concurrent, 69 viscosity, 241
couples, 70 Fluid friction (viscous friction), 22, 176
definition, 66 Foci of conic section, 166
for equilibrium, 67 Force, aerodynamic, 234, 288, 289
general, 69 alternating, 157
parallel, 69, 78 applied, 328
planar, 70 attractive, 162, 164
reduction to, 68-70 blast, 214
wrench for, 69 buoyant, 96, 97
Ergodic problem, 245 central, 162, 174
Escape from attracting body, 166 centrifugal, 181
Escape velocity, 169, 184 colinear, 16
Euler, L., 124 compressive, 17
Eulerian angles, 131-133 concvu-rent, 16, 24, 57, 69
angular velocity in terms of, 133 conservative, 138, 172-174, 282-285
Euler's equations of motion, 280, 281, constraining, 76, 329
299-304 coplanar, 16, 70
Euler's theorem, 124 Coriolis, 181
Exciting force, 191, 192 damping, 156, 161, 191, 192
External forces, 52, 53, 223, 236 distrilnited, 22
Extremum, 330 disturbing, 214
drag, 184
exciting, 191, 192
Fi('(itious forces, 182 external, 52, 223
JMcid, electric, 37 field, 38, 172
Index A23
lort'C, fictitiovis, 181 Free vibrations, of /( degrees of freetlom,
friction, 21 346-351
generalized, Sio, '•io'2 of two degrees of freedom, 342-34ti
of gravity, 15G un(lanij)ed, 193-197
impulsive, 22li 'Freedom, degrees of, 55, 188. 189, 326
interniolecular, 242, 243 Freely moving air-borne vehicle, 286-290
internal, 52, 53, 223, 282 Freely moving rigid body, 277
nonperioilie, 214 Frenet fornuilas, lOS. 109
norniul, 21 Frequency, definition, 158
periodic, 157, li>2, 202, 207 forcing, 203
pressure, 8(), 8«.», IM, 234 natural, 1.58, 190, 195, 216
propulsive, 288 natural circular, 158, 190, 195
reactive, 328 ilamped, 200
rectilinear, 157 normal, 343, 345, 347
repulsive, lti2, 184 resonant, 205. 209
restoring, H»2, 340 of rotating rigid body, 293
resultant, 15, 08, 153 of simple harmonic motion, 100
shear, 85 of simple pendulum, 178-179
spring, 15(1, I'Jl, 192 Fre(iuency equation, 343, 347
tensile, 17 repeated roots, 343
variable, 156 zero roots, 343
viscous, 150, 191 Friction, 21-23
weight, 14 coefficient, 22
F'orce-couple system, 67-69 Coulondj, 22, 218
Force field, 38^ 172 dry, 218
central, 174
lt)2, fluid, 22
conservative, 38, 172-176 force, 21, 318
uniform, 79, 173, 174 kinetic, 22, 176
Force systems, equipollent, 66-70 laws, 22
parallel, 69,77-79 static, 22
reduction, 68, 69 viscous, 22, 176
F"orced vibrations, damped, 207-210 (See also Damping)
of n degrees of freedom, 351, 352 PVictionless bars, 18
undamiM'd, 202-206 Frictionless bodies, 55, (i3
Ff)rces, parallelogram law, 3 I'Victionless surface, 21
polygon law, 16 Fundamental haniionic of Fourier series,
transmissihility, 44 191
workless, 329 I'vindamental modi' of \ii>ratioii, 216
Forcing fretiuencj", 203
Four-bar linkage, 136-139
Fourier analysis, 191 Gas, Boyle's law for, 240
Fourier series, 191, 192 constant, 240
PVames (see Structures) equation of state, 240
Frames of reference, 13 ideal, 87, 239, 243
absolute, 13, 140-142 incompressible, 101
astronomical, 13 isothermal relation for, 87
moving, 139-142, 181 kinetic theory, 238 24 1
Newtonian, 140 poly tropic relation for, 88

rotating, 144 temjH'rature, 239
(See also Coordinates) Gauge pressure, 89
Free-body fliagram, particle, 14, 1(5, 24 Generalized coordinates, 188, 189, 326
rigid body, 58, 59, 61 Generalized forces, 335, 3.52
PVee falling body, 154 (Jeometric series, 200
Fre«> vector, 2 Geometrical rei)re.sentation, of Mohr's
Free vibrations, of coupled systems, circle, 268-271
340-342 of moment of inertia, inertia ellipsoid,
damped, 197-200 262. 263
A24 Index
Geometrical representation, of motion of Homogeneous differential equation, 201,
rigid body, Poinsot ellipsoid, 295- 203, 207
298 Horizontal plane on rotating earth, 182,
of vector, 2 308
Gimbals, 294 Horsepower, 62
Governor, 146 Hydrogen, 248
Gradient vector, 173 Hydrostatic pressure, 88
Gravitation constant, 164 Hyperbolic orbit, 166, 169
Gravitational acceleration, 167
Gravitational attraction, 162
field, 174, 175 Ideal gas, 87, 239, 243
Newton's law, 164, 174 Idealization, 14
potential, 175 of earth, 14, 164, 175, 181, 300
Gravity center, 79, 81, 224 of mechanical vibrating system, 156,
Gravity force, 14 192
{See also Gravitational acceleration; Impact, direct, central, 246
Gravitational attraction) Impedance, 209
Gyration radius, 251 Impeller, 152
Gyrocompass, 307-310 Impulse, linear, 225
Gyroscope, 294 unit, 220
flattened, motion, 302 Impulse and momentum principle, 225
motion, under external moments, Impulsive force, 226
303-307 Impulsive moment, 231
free, 298-303 Impulsive motion, 226, 231
nutation, 307 Inclined axes, moment of inertia, 257, 261
precession, direct, 30i Independent variables, 188
retrograde, 303 Indeterminate problems (see Determi-
steady, 305 nate problems)
Gyroscopic moment, 305 Inductance, 193, 202
Gyroscopic stabilizer, 306 Inductive reactance, 210
Inertia (see Moments; Products)
Inertia ellipsoid, 263-267, 296
Hamilton, W. R., 327 principal diameters, 263
Hamilton's principle, 327 of revolution, 266
Hard springs, 218 Inertia tensor, 250
Harmonic motion, graphs, 190 Infinitesimal displacement, of particle, 37
simple, 157-161, 189, 190 of rigid body, 126
amplitude, 159 Infinitesimal rotation, 125-128
differential equation for, 160 vector character, 125
frequency, 158 Initial conditions, 154, 155, 194, 197,
circular, 158 203, 208
period, 158 Instability (see Stability of rotating
phase shift, 190 rigid body)
for simple pendulum, 178' Instantaneous axis of rotation, 127
(See also Vibration) Instantaneous center of rotation, 131
sum, 190, 191 Instantaneous change, of angular mc-
superposition, 344, 351 mentum, 231
Harmonic oscillator, 156, 191, 192 of velocit}', 226
Harmonics, fundamental, 191 Integrable constraint conditions, 326/;.
nth, 191 Integration of eqiuitions of motion, by
Helicopter blade, vibration, 152 computing macliines, 157, 220, 305
Herpolhodc, 297 rectilinear,154-156
Hinged l)ars, 55 by approximations, 178-180
scries
Hinges, 18, 59 {See also Motion, of rigid body)
Hc)l()iK)mi(t constraints, 326n. Interinolecular forces, 242, 243
Homogeneous body, 81, 251 Internal forces, 52, 53, 223, 282
properties, A5-A8 Interplanetary mission, 169
hide J \2:)
InvariaMe line, 297 Kinematics, of rigid body, translation

Iiivarial)le plane, '2".)7 123
Invariant, inertia eliijisoid as, 203 velocity. 128
ln\('rse-s(|uare law, l(i4, 174 cro.ss component, 129
I.solation, iiKiuntin^, 211 longitudinal c(>m|>onent. 129
vihration, 210 relativ*', 129
Isothcrmai relation, 87, 101 Kinetic energy, of molecule, 239
of particle, 171
quadratic form, 342
Jacobian elliptic functions, 177 relative to ma.ss center, 237
Jet enpine, thrust, 23ti of rigid body, general, 283, 284
Jet propulsion. 2:^3-236 of mass center, 284
Joints, e(|uilihrivini, 58 relative to mass center, 283
liinncd, 18, 59 total, 284
Jump phenomena, nonlinear vibration, pivoted, 284, 285
218 in plane motion, 317
rotating about fixed axis, 313
of system of particles, 236-238
Kepler, J., 167 of ma.ss center, 2:i7
Kepler's laws, 163, 166, 167 total, 237
Kinematics, of particle, 110-119 average, in bounded region. 238
acceleration. 111 Kinetic friction, 22, 176
in cylindrical coordinates, 114-116 Kinetic theory of gases, 238-241
in rectangular coordinates. 111. 112 KroMcckcr delta, 257, 349
simple harmonic motion, 159
in spherical coordinates, 116-1 19
velocity, 110 Lagrange, J. L., 327
of rigid body, 123-145 Lagrange's eciuations, 333-339
acceleration, 133-135 applicat ions, 33()-339
angular velocity, 45, 127 introduction to, 327, 328
base point for, 125-130 for particle, 333-:i3()
baiiic theorems for, 126-131 for system of particles, 338
Chiusles' theorem, 125 Lagrangian function, 336
dL-placement, finite, 124 126 Latitude, 183, 310
infinitesimal, 125, 126 Launching speed, lt)9
Eulerian angles for, 132-133 Laws, Boyle's, 240
Euler's theorem, 124 C'Oulomb's, of friction, 22
four-bar linkages, 13(y-139 Kepler's, of planetary motion. 163. 166,
instantaneous axes, 127 167
in moving coordinate system, Newtim's (nee Newton's law)
141-145 IjCgendre's elliptic integral, 177
acceleration, absolute, 144 L'Hopital's rule, 205
apparent, 145 Limit of sums and products. 105
Coriolis, 145 Line of action, concurrent forces, 57
relative, 145 force, sliding along, 44
vehicle, 145 screw, 130
velocity, absolute, 142 vector, shifting, 2, 3, 36
apparent, 143 wrench, 69

relative, 142 Line integral, 38, 171
vehicle, 143 Linear impulse, 225
plane motion, 130, 131, 135 Linear momentum, of p.irticle, 12, 15,
centrodes, 131 225
instantaneous center for. 131 principle, 225
reduction to screw, 130 unit change. 222
roiling wiieel, 135 of system of |)articlo8, 225
rotation, about fixed axis, 123 conservation theorem, 225, 226
about fixed point, 124 principle, 225
A26 Index
Linear restoring- force, 340 Mechanics, quantum, I57n.
Linear transformation, 350 relativistic, V2n.
{See also Orthogonal transformation) statistical, 245
Linear velocity, 137 Mechanism, 136
Linkages, four-bar, 136-139 four-bar linkages, 136-139
Liquids, hydrostatic pressure in, 88 pump, 150
manometer for, 88, 89 quick return, 121
viscosity, 243 reciprocating engine, 121, 137
(See also Fluid) Scotch crank, 121
Load {see Force) Mercury, 89
Logarithmic decrement, 200 Meridional plane, 118, 132
Longitude, 116 Meta center, 98
Longitudinal component, of moment, 68 Minimum of potential energy, 97, 330,
of velocity, 129 331
Lunar mission, 169 Minimum moment of inertia, about all
253
parallel axes,
at point, 264, 269
Maclaurin's expansion, 331, 340 (»See moments of inertia)
also Principal
Magnification factor (amplitude factor), Modal superposition, 352
204, 208, 209 Model, mathematical, 14
Magnitude of vector, 2 {See also Idealization)
Manometer, 88 Modes of vibration, natural, 216
Mass, definition, 12, 14 normal or principal, 343, 345, 347
of earth, 100 amplitude ratio for, 344
of moon, 100 Modulus, Legendre's elliptic integral, 177
of oxygen molecule, 244 Mohr, O., 270
variable, 233 Mohr's circle, 268-271
Mass center, 80, 81 Molecular velocity, 241, 244
angular momentum relative to, 230, Molecules, collisions, 239
278 gas constant for, 240
of composite bodies, 84 mean free path, 243, 244
determined, by integration, 83 mean kinetic energy, 239
by symmetry, 80 properties, hydrogen, 248
of homogeneous bodies, A5-A8 oxygen, 244
kinetic energy, 237, 283 velocity, 241, 244
motion, 223, 224, 277, 294 vibration, 239
Mass matrix, 354n. Moment, constraining, 231
Mass moments of inertia {see Moments of couple, 65
of inertia) first, of area, line, mass, and volume,
Mass point, 84 80
Mass products of inertia {see Products of of force, 39-43
inertia) components, 41
Mathieu's equation, 216 about line, 40, 42, 43
Matrix, mass, 354ri. about point, 39, 41, 43
stiffness, 354n. as vector product, 39, 41
Maximum moment of inertia, 264, 269 of momentum (.st'(^ Angular momentum)
{See also Principal moments of inertia) resultant, 05
Maxwell, J. C, 244 component, cross, 68
Maxwell's formula for viscosity of gas, longitudinal, 68
244 of vector, 43-45
Mean free path, 243, 244 Momental ellipsoid {see Inertia ellii)soi(l)
Mean kinetic energy, 239 Moment arm, 40
Mechanical electrical analogies, 193 Moments of inertia, of areas, 91-9)}, 251
Mechanical energy conservation {see parallel-axes theorem for, 93
Principle, of conservation) tabulated, A1-A4
Mechanics, classical, \'2n.. of mass, 249-256
Newtonian, 12^. array, 250
1
Index A27
Moiiu-nts of inertia, of mass, of com- Natural fre(]uencv, of coupled systems,
posite bodies, 25(> 2 If)
determined by inleumtion, jr»:<-JA5 of sim|)le harmonic motion, 190

elTeet on stnhility of rotating IhxIv, of undamped free vibration, 195
2«H)-J(»3 Natural modes of vii)ration, 21()
of homogeneous bodies, Ao AS Normal modes of vibration)
(iSVc (ilso
inertia ellipsoid, '2iV2, '2(V.\ Necessary and sufficient conditions of
parallel-axes theorem for, 252, 25'.\ ecpiilibrivun, 57, 328, 3;il
prineipal axes, 203-205 Negligible mass, 200
relative to principal axes, 2()4 Neutral etpiilibrium, 97, 3:iO, 331
relative to rotated axes, 257, 201 Neutral position (equilibrium position),
as tensor, 250 15t>, 1S7. 191, 190, 198
maximum, 201, 20*) Newton, I., 12

mininmm, 25;},204, 209 Newtonian frame of reference, ]:\, 1 10,
Mohr's circle for, 270 227, 229
principal, 20!i, 204 Newtonian mechanics, discussion, {'A. II
Momentum {see Anpular momentum; interpretation, 13
Linear momentum) laws, 12
Moon, data on, 100 Newton's law, of gravitational attract um,
Motion, of particle, 12, 15:?, 154 104, 174
absolute, 12 of motion, first, 12
constant-velocity, 159 second, 12
constrained, 170 third, 13
general, 154, 157 of restitution, 240
rectilinear, 154 of viscosity, 242
relative to movins frame of ref- Nonconservative force field, 176
erence, 141-147 Nonlinear differential ecjuation, 218
simple harmonic, 159 Nonperiodic force, 214
of rifîd body, air-borne vehicle, Nornuil components, of acceleration, 1 13
280-290 of contact force, 21
gyroscope, free, 298-30:5 Normal coordinates of vibration, 348, 351
with moment, 30:}-;}07 Normal force, 21
Poinsot's representation, 294-298 Normal freciuencies, 343, 347
plane, 310-319 Nornuil modes of vibration, 343, :?45, 317
rotation about fixed axis, 311-313 Normal reaction, :i29
translation, 310, 311 Norn)al tangential coordinates, 113
{See also Ecjuations, of motion) Normal vector, principal, 108
Motion curves, 1 1 Normalization conditions, 259
Moving frames of reference, 13, i:i9- Normalized amplitudes, :{49
141 North, determination, ;}07
motion of particle in, 181, 182 Null vector, 3
motion referred to, 141-145 force system eciuipollent to, 07
(See also Hody axes) Nutation of top, 307
Moving trihedral, 109, 281
Mult ij)licat ion, of vector and scalar, 3
of vectors, scalar product, 29-31 Orbit, in central force field, 102 104
scalar triple i)roduct, 40, 47 Kepler's law for, 10;i
vector product, 33-35 j)lanar charaiiter, Hi:}

vector triple product, 47, 48 in gravitational field, 104-107
eccentricity, 100
Kej)ler's laws for, 100, 107
Natural circular frefpiency, of d;uiiped type, 100
free vibration, 200 I)lanetary, 100, 107
by Kayleigh's method, 195 satellite, H;7-170
of simj)le harmonic motion, 190 circular, 108, 109
of two-degree-of-freedom systems, :{43 eccentricity, H)8
of undam|M"d free vibration, 190 Orbital speed, 108, 109, 184
A28 Index
Order of magnitude, simplifications based Period, of simple harmonic motion, 160,
on, 290, 309 190
Orthogonal transformation, 2G0, 274 of simple pendulum, 178, 179
Orthogonal triad (moving trihedral), 109, of undamped free vibration, 195
281 Periodic force, 157, 192, 202, 207
Orthogonality conditions, 349 Periodic motion, 189, 190, 203
Orthogonality relations, 259 (See also Simple harmonic motion;
Oscillation, helicopter blade, 152 Vibration)
(See also Vibration) Perpetual-motion machine, 103
Oscillator, harmonic (spring-suspended Perturbation, angular velocity, 291
mass), 156 equations, 291
Osculating plane, 108 Phase angle, 191, 207
Overdamped vibrations, 198 Phase shift, 190
Oxygen properties, 244, 248 Piercing point, 48
Pivoted rigid body, 284, 295
Planar motion in central force field, 163
Pappus theorem, 83 Plane, invariable, 297
Parabolic orbit, 166, 169 osculating, 168
Parallel-axes theorem, for moments of tangent, to earth, 303
inertia, of area, 93 to Poinsot ellipsoid, 296
of mass, 252 Plane body, center of mass, 81
for products of inertia, of area, 93 moment of inertia, 251
of mass, 252 Plane motion of rigid body, 130
Parallel forces, force equipollent to, 69, 78 equations, 316-319
line of action, 78, 79 kinematics, acceleration, 135
Parallel springs, 219 velocity, 130, 131
Parallelepiped, moment of inertia, 253 kinetic energy, 317
Parallelogram law of addition, 3 Plane rotation of axes, 268-271
Particle, 13 Planet, motion around sun, 164
acceleration. 111 Planetary orbit, 166
angular momentum, 227 apogee, 167
under central force, 162-167 Kepler's laws for, 166, 167
dynamics, 153-183 perigee, 167
equilibrium, 15-24 period, 167
kinematics, 110-119 Poinsot, L., 294
kinetic energy, 171 Poinsot ellipsoid, 295-298
Lagrange's equation for, 335 Point of application of force, 2
linear momentum, 225 Polar coordinates, 114
Newton's law for, 12 Pole, velocity, 131
position vector for, 111 Polhode, 296, 297
potential energy, 172 Polygon rule for vector addition, 4
on rotating earth, 181 Polytropic relation, 88
velocity. 111 Population growth, 202
Particular solution of differential equa- Position-time curve. 111
tions, 201, 203, 207 Position vector, 36, 39, 43, 46
Past and future, determination, I57n. Potential, gravitational, 175
Pendulum, ballisti(!, 248 Potential energy, 171-173
centrifugal, 186 for central force field, 174
compound, 322, 359 for conservative force field, 172
double, 357, 359 datum for, 175
sinipUi, 176-181 of elastic spring, 175
Perfect differential, 185 for inverse-S(iuare law of attraction,
Perfect gas (ideal gas), 87, 239, 243 175
Perigee, 166, 169, 181, 185 mininunn for stability, 97, 330
Period, of alternating fonu', 158 necessary and sufficient conditions for,
of damped free vibration, 200 173
of planets, 167 quadratic form, 341, 346
Index A29
Potential pnorgy, of ripid hody. _>H4 Principle, of work and energy, for rigid
stationary values, :VM) body, 282-285
of uniform forro fieltl, 174 rotating about fixed axis, 313
Procession, 21)8 for system of particles, 236
direct, 302 Probability, of motions, 157/t.

of earth. 300 of significant properties of system, 245
progressive, Ii02 Product, of vector and scalar, 3
retroRrade, 'M'.i of vectors, scalar, 29-31
steady, 305, 30G, 320 scalar triple, 46, 47
velocity, 298 vector, 33-35
vector triple, 47, 48
Pressure, absolute, 85)
rrixlucts, of inertia, of areius, 91-93, 251
atmospheric, aml)i»'nt, 88
standard, 89 |)arallel-axes theorem for, 93
center, 1)1, '.)4
tabulated, Al-.\4
continuous fluid at rest, 87 of mass, 249-256
in
8G array, 250
in fluid at rest,
of composite bodies, 256
in gas, 239
hydrostatic, 88
determination l>y synniietry, 2.55,
units of measurement, 89
256
parallel-axes theorem for, 252, 253
Pressure-altitude relation, 87
relative, to principal axes, 2()4
Pressure forces, 89
to rotated axes, 257, 261
on curved surface, 94, 95
as tensor, 250
on flat plate, 89, 94
Mohr's circle for, 270
resultant, 90, 94
Progressive precêssion, 'M)'2
on rocket, 234
Projectile with resistance, 184
Pressure gauge, 89
Projection of vector, on some line, 6
Pressure intensity {see Pressure)
on some plane, 49
Principal axes of inertia, 203 Propulsion shutdown, U)8
location, 2<')5
(.SV^ also Rocket burnout)
rotated moments of inertia referred to,
Propulsive force, 2S8
204 Pump, impeller, 152
by symmetry, 266, 267 mechanism, 150
Principal modes of vibration (see Normal
modes of vibration)
Principal luonients of inertia, 203, 2r)4 Quadratic forms, of kinetic energy, 342
Princii)al normal, 108 of potential energy, 341
Principal planes of inertia, 203 rethiction, 351
Principle, of angular momentum, for (Quadratic surface for inertia elii|)soid, 263
particle, 228 Quantum mechanics, 157//.
for syst<'m of particles, about fixed (ûick-return mechanism, 121
origin, 228
about moving mass center, 230
of Archimedes, 9(), 103 liadial components, of acct-leration, of
of conservation, of mechanical energy, particle, 115, 119
for particle, 173, 174 of rigiil body, 135
for rigid body, general, 284 of central force, 174
pivoted, 2S5 of velocity of particle, 115, 119
of momentum, angular, 228 Radius, of curvature, 108
linear, 225 of gyration, 251
of impuls*' and momentum, for Rate of change of vector (see Derivative
particrle, 225 of vector)
for system t)f particles, 225 IJayleigh's method, 196
of superi)osition, 214, 21S Reactions, of beam, 72
of virtual work, 329 of body in e(iuilibrium, 20, 55
of work and energy, for i)article, 171, of constraint, 22, 329
180 normal, 22, 329
A30 Index
Reactions, periodic, in bearing, 313 Resistance, of air, 184
redundant, 56 of series circuit, 193
for rotating rigid body, 313 Resisting medium, 161, 184
of rough contact, 22 Resonance, 205
of smooth contact, 22 210
in electric circuit,
workless, 329 subharmonic, 219
Reciprocating engine mechanism, 121, Resonant amplitude, 209
137 Resonant frequency, 205, 209
Recoil of guns, 226 Rest, 13
Rectangular components, of acceleration, Restitution, coefRcient, 246
111 Restoring force, 192, 340
of angular momentum, 280 Resultant, angular velocity, 126
of force, 15 external force, 53, 56, 236
of moment, 41 external moment, 54, 56
of position, 41, 111 of finite rotations, 126
of vector, 7 force, general, 68, 153
derivative, 106 parallel, 78
of velocity. 111 of infinitesimal rotations, 126
(See also Equations, of equilibrium, of Resultant moment, 68
motion) Resultant pressure force, 90, 94
Rectangular parallelepiped, moments of Retrograde precession, 303
inertia, 253 Revolution, body of, moment of inertia,
Rectilinear force, 157 266
Rectilinear motion, of particle, 154-157, motion, 298
184 Right-hand coordinate system, 6
constant velocity, 159 Right-hand-screw rule, 33, 40
free falling body, 154 Rigid bars, 18
simple harmonic, 159 colinearity of forces on, 60
of rigid body, 310 Rigid body, 54, 123
of system of particles, 234 angular momentum, about mass center,
Reduction, of angular velocities, 127 278, 279
of concurrent force systems, 69, 70 about pivot, 279
of general force system, to force and angular velocit}-, 45, 127
couple, 69 balancing, 312, 313
to wrench, 69 constraint, 54-56
of parallel force systems, 69, 78, 79 displacement, 54, 123
of quadratic forms, 351 Chasles' theorem for, 125
of rigid body motion, to rotation and Euler's theorem for, 124
translation, 130 equations of motion for, Euler's, 280,
to screw, 130 281
Redundant reaction, 56 general, 280
Reference frame (see Frames of reference) for moving coordinate axes, 281
Relative acceleration in rotating coordi- Eulerian angles for, 131-133
nate system, 145 freely moving, 277
Relative angular momentum, of rigid kinematics, point on, acceleration,
body, 278, 279 133-135
of system of particles, 230, 231 velocity, 45, 133
Relative kinetic energy, for rigid Ixxiy, theorems for, 127-131
283, 284 kinetic energy, general, 283, 284
for system of particles, 237 for pivoted, 284, 285
Relative velocity, of point on riîil body, plane motion, equations, 316-317
129 kinematics, 130, 131, 135
in rotating coordinate system, I 12 Poinsot's representation for motion,
Relaxation time, 200, 201" 29-4-298
ReiK'ated roots of frc(iuency e<iu:iti()n, potential energy, 284
343 rotation about fixed axis, 311, 313
Re])ulsive force, 162, 184 rotational stability, 290-293, 307
Index WW
Rigid liody, stable, oti Scalar product, gcoinelrical iulerprcla-
translation. 3ia-:ill tion, 30
typos of motion. J'.ll S<'.alar triple jiroduct, 46
work and oncrKy for, "JSJ 285 S<-alar variable. 104
(See also Air-borne vcliicle; Gyroscope) "Scotch crank," 121
Ring guide, 55 Screw, 130
/^Lr eireiiit, 193, 209 Second moment of m;is.<. 250
Roeker, 149 (See also Moments of inertia)
Rocket, engine, 234 Sections method. 73
escape velocity, 184 St'ctorial speed (areal speed), 163, 1()7
propulsion, 233-236 Semilatus rectum, 167
sf>ecific impulse, 235n. Series circuit, 19:}
thrust, 235 Series spring, 219
Rocket burnout, 168, 169 Shear force, 85
Rod motion, 302 Shear 244
stress, 85. 211.
Roller. 59 Shortest distance, from point to line. .32,
Rolling without slipping, 136, 317 36
Rolling contact. 13(5 between two nonintersecting lines, 36
lioiling cylinder. 317 Simple harmonic motion, 159, 160, 194
Rolling wheel, 135, 148 amplitude, 159
Hoot-mean-square velocity, 241, 244 fretiuency, circular, 160
Rotating frame of reference, 139, 140 period, 160
ilerivative of vector in, 140, 141 Simple pendulum, 176-180
motion of particle in, 181, 182 amplitude, 179
motion referred to, acceleration, 144, period, 179
145 as simple harmonic motion, 180
velocity, 142. 143 Simple structures, 17-21, 61
Rotation, of coordinate axes, 256-258 Simplifications, based, on order of mag-
of earth, 181 nitude, 183, 290, 309
about fixed axis, rigifl Ixxly, ((iiiations on symmetry, 26(), 290
of motion, 311, 312 Simultaneous dilTereiitial ecpiations, 157,
velocity, 45 342, 352
instantaneous axes, 127 Sine law of trigonometry, 35
instantaneous center, 131 Sliding friction (kinetic friction), 22. 176
plane, of axes, 208 Sliding vector, 2
Molir's circle for, 270, 271 Smooth contact, 329
Rotations, finite, 1, 126 Smooth surface, 22
infinitesimal, 125-127 Soft springs, 218
resultant, 127 Space axes, 132
Rough contact, 22 Space centrode, 131
Sj)ac(; cone, 297
Space structure, 74
235
Satellite, launching, Si)ecific gravity, 89
motion around earth, 164, 184. 185 Specific impulse, 2;{5«.
orbits, 1()7-170 Specific weight, 88
Scalar, examjjle, 1 of air, 101
Scalar components, of acceleration, of water, 89
111-119 Speed, 110
of vector, 7 S|)eed-time curve, 1 1 I
of velocity, 111-119 Sphere, moment of inertia, 255

Scalar ec|uation.s (see Equations, of e((iii- jjotential for, 175
librium, of motion) surface area, 83
S<ralar protluct, associative law for, 31 volvnne, 83
commutative law for, 31 Spherical coordinates, 116-119
in component form, 31 transformation e(|Uations for, 119
definition, 29 Spherical slieli. moment of inertia, 255
distributive law for, 31 I)otential, 175, 185
A32 Index
Spin, gyrocompass, 309 Symmetry, axis, used to find area product
gyroscope, 298 of inertia, 93
Spring, constant, 25 plane, used to find, centroids and mass
force, 156, 191, 192 centers, 81
potential energy, 175 mass products of inertia, 256
Spring-suspended mass, 156 principal axes, 266
Springs, Belleville, 218 simplifications of motion based on, 290
hard, 218 System of particles, angular momentum,
linear, 156, 160, 191 228
in parallel, 219 equilibrium, 52
in series, 219 kinetic energy, 236
variable, 216 Lagrange's equations for, 338
Stability of rotating rigid body, 290-293, linear momentum, 225
307 motion, 223-245
Stable equilibrium, 96, 97, 330, 331 potential energy, 284
Stable rigid bodies, 56 work done on, 236
Standard atmospheric pressure, 89 Systems of forces, equipollent, 66-70
Standard form of vector, 7 parallel, 69, 77-79
Starting conditions (see Initial conditions) reduction, 68, 69
Static coupling, 342
Static equilibrium of particle, 15
plane problems, 20 Tangent to curve in space, 108
procedures, 24 Tangent plane, to earth, 303
space problems, 20 to Poinsot ellipsoid, 296
Static-friction coefficient, 22 Tangential components, of acceleration,
Static position (see Equilibrium position) particle, 113, 115
Statically balanced rigid body, 312, 315 point on rigid body, 134, 135
Statically determinate particle, 21 of velocity, particle, 113, 115
Statically determinate rigid body, 56 point on rigid body, 45
Statically indeterminate particle, 21 Temperature of gas, 239
Statically indeterminate rigid body, 56 absolute, 240
Statics, 15 Tensile forces, 17, 21
Stationary values, 330 Tension, belt, 63
Statistical mechanics, 245 in string, 17
Steady precession, 305, 306, 320 Tensor, cartesian, 271
Steady-state current, 209, 210 250
of inertia,
Steady-state vibration, 208 symmetrical, 265n.
Stiffness matrix, 354n. transformation, 261
String, flexible, 17 Terminal velocity, 162, 184
Structures, analysis, 18 Theorem of equipartition of energy, 240
simple, 17-21, 61 Three-dimensional equilibrium, of
space, 74 particle, 20
Subharmonic resonance, 219 of rigid body, 57-60
Submerged surfaces, forces on, curved, {See also Equations, of eciuilibrium)
94, 95 Three-force body, 60
plane, 89-91 Thru'jt, jet, 236
Subtraction of vectors, 5 rocket, 235
Sun, attraction, 162 Time average, distributive projiorty, 238
motion of earth around, 164 Time constant, 200-202
Superposition, modal, 352 Top (see Gyroscope)
princtiple, 214, 218 Toniue, alternating, 213
Surfatre, rough, 22 Torsion of spac-e curve, 109
smooth, 22 Torsional vibrations, 213
sul)ni('rgo(l, forces on, 89-95 Total angular momentum of system of
Suit ell, 202 particles, al)out fixed origin, 228, 231
Symmetricâl l)0(ly, motion, 298 about mass center, 229
Symmetrical tensor, 265/t. Total dilTerential, 172
hide J A33
Total force, 07 T\vo-degree-of free<l()m systems, vibra-
extornal. 5.i, 50, 224 tion, 210, 342 344
intiTiml, '2'2'.i Two-dimensional e(iuilibriuiii, of particle,
Total iiii pulse, 225 20
Total kinetic enerRV, of molecule, 240 of rigid body, 58, 59
of rigid body, 2S;i (See also Ecjuations, of e(|iiilii)rium)
of system of particles, 237 Two-force body, 00
Total linear momentum of system of
particles, 225
Total moment, 07 r tube, vibration, 220
of external forces, 54, 50, 228, 230 I'nbalanced motor, vibrations due to, 210
Total work, and energv, for rijîd liodv, I'nderdamped vibrations, forced, 207
284 free, 198-200
for system of particles, 2ii7 Uniform field of force, gravitational, 79,
of rigid body, 282 173, 174
of system of particles, 230 potential energy for, 174
Trajectory, in central force field, 100 I'niform rectilinear motion, 159
in resisting medium, 101, lt)2 Uniformly accelerated rectilinear motion
Transfer-axes theorem (see Parallel-axes (free falling body), 154
theorem) Unit circle, 1 \5n.
Transfer ellipse, 109 Unit impulse, 222
Transformation, of axes, 256 Unit sphere, 118
linear, 200, 350 Unit trihedral, 109
matrix, 25'.) Unit vectors, 5
of moments of inertia, rotation of axes, binormal, 109
200-202 cartesian, 5
translation of axes, H2, 1)3, 252, 253 in cylindrical coordinates, 1 14
orthogonal, 200, 274 orthogonal, 5
tensor, 201 principal normal, 108
Transient vibrations, 207 in splierical coordinates, 110
Translation, of coordinate axes, for area tangent, 108
moments and products of inertia, in tangent normal coordinates, 113
92, 93 Universal gravitation constant, 104
for mass moments and products of Unstable e(|uilibrium, 97, 331
inertia, 252, 253 Unstable rigid body, 50
of rigid body, definition, 123, 310
equations of motion for, 310, 311
kinematics, acceleration, 134 Variable, independent, 188
velocity, 128-130 scalar, 104
kinetic energy, 284 vector, 104
of system of particles, angular mo- Variable force, 30rt.
mentum, 231 Variable mass, system with, 233, 234
kinetic energy, 237 Variable spring, 2 Hi
momentum, 225 Varignon's theorem, 45
motion of mass center, 224 Vector, acceleration, 12, 14
Transmissibility, of force, 44 angular, 133
of vibrations, 211 angular momentum, 104, 227
Triad, moving, 281 angular velocity, 127
unit orthogonal, 10!) binormal, 109
Trial functions for solution of dilTerential bound, 2
•(Illations, 194, 197, 203, 342 constant, 105, 103, 105
Triangle rule of addition (polygon rule), 4 e(|Ual, 3
Triple vector products, scalar, 40 fixed, 2
ve<tor, 47 force, 12, 14
Truss, plane, 72, 73 free, 2
space, 27, 75 gradient, 173
Tumbling of rigid body, 291 infinitesimal rotation, 125
A34 Index
Vector, infmitesimal translation, 126 Velocity, referred to moving reference
linear impulse, 225 frame, 142, 143
momentum, 12, 15, 225 relative, 129, 142
linear, 225 of sound in gas, 241
negative, 3 in spherical coordinates, 116-119
position, 36, 43, 108 in tangential and normal coordinates,
principal normal, 108 113
sliding, 2 terminal, 162, 184
unit, 5 of vehicle, 142
variable, 104 Velocity gradient, 242
velocity, 12, 14 Velocity pole, 131
zero, 3 Velocity-time curve, 111, 190
Vector addition {see Addition of vectors) Vertical on rotating earth, 182
Vector components, 6, 7 Vibration, amplitude (see Amplitude)
Vector decomposition, 6 amplitude factor, 204, 208, 209
Vector differential equation, 107 amplitude ratio, 344
Vector differentiation (see Derivative of of coupled systems, 216, 340-342
vector) creeping back, 197, 198
Vector equality, 3 equation for, 191
differential
in terms of components, 8 energy method for, 196
Vector field, 37n., 173 equilibrium position for, 191
Vector function, 37«.., 104 forced, damped, 207-210
Vector multiplication (see Multiplication) undamped, 202-210
Vector operator, 173 free, damped, 197-200
Vector product, commutative law for, 33 critically, 198
in component form, 35 over-, 197
definition, 33 under-, 198
as determinant, 35 undamped, 193-197
distributive law for, 34 frequency, forced, 203
geometrical interpretation, 33 natural, 195
Vector quantities, 1 natural circular, 190, 195, 200
Vector triple product, 47 normal, 343, 347
Vectors, example, 1 resonant, 205, 209
function, 104 isolation, 210
geometrical representation, 2 logarithmic decrement, 200
magnitude, 2 of mechanical systems, 190
moment, 43 modes, natural, 216
projection, 6 normal, 343, 345, 347
specification, 1 of n-degree-of-freedom system, forced,
standard form, 7 351, 352
Vehicle, 140 free, 346-351
acceleration, 147 nonlinear, 218
air-borne, 286-290 with nonperiodic forces, 214
velocity, 142 normalized amplitudes, 349
Velocity, 12, 14, 110 period, 190, 195, 200
absolute, 13, 142 resonance, 205
angular (see Angular velocity) of single-degree-of-freedom system,
apparent, 143 190, 191
areal, 163, 167 steady-state, 208
in cylindrical coordinates, 114 torsional, 213
escape, 169, 184 transient, 207
instantaneous ('liango, 226 transmissibility, 2] I
lin(>ar, 137 of two-degree-()f-freo(loin system, 216,

of point on rigid body, 45, 128 342-346
in j)lane motion, 131 with variable cliaracteristics, 216
of precession, 298 ViriaJ theorem, 238
in rectangular coordinates, HI Virtual displacement, 328
Imlf \:v.
\irtual work, :V_»8, .V.ia Weight, force, 14

princii)lt\ '.V2\) j)roperty, 82
Viscosity, 241 specific, 88
242
c-oefru'ient, of air, 101
Newton's law. 242 of water, 89
Viscous d:iini)iiiK, lot), l'.t2 Weiglited point, 82
coefticient, 15t>, 1"J2 Weighting function, 82
Visrous force, l*.t2 Wlieel, roiling, 135, 148, 149
N'iscous friction, 22, I7t) Work, ilone hy coujjle, 283, 285
N'iscous niediuiu, loti. ltd. 202 done hy force, 36, 171
N'iscous resistance, I'.ll done hy forces, on rigid hody, 284
N'oltage drops, 1U3 on systems of particles, 236
N'oltafie source, an-, 209 as line integral. 37, 38
constant, 202 virtual. 328
variable, WV.i Work and energy priiuii)lc {so Princi-
\'oluine, centroid, 80 ple)
moment of inertia, 251 Workless constraints, 171, 329
of revolution, 266 Wrench, 06, 76, 77
of sphere, 83
Zero force, e(|uip()llent to, 67

Wc.ln,., 28 Zero roots of fretjuency e(iuatioii, .'M3
Weight, in center of gravity, 79 Zero vector, 3
'/I

Volumen 1.1

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Volumen 1.1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Volumen 1.1

Uploaded by

Copyright:

Available Formats

r.

VOLUME 1. PARTICLE AND RIGID-BODY MECHANICS

Craxdall and Dahl •

Karplus and Soroka •

PARTICLE AND RIGID-BODY MECHANICS

McGPvAW-IIITJ. ROOK ("O.MI'AXV, INC.

1960 New York Toronto London

This book concerned with the basic principles of the mechanics of

Stands and It is an outgro\N"th of a set of mimeographed not<is

b;vsic disciplinessuch as mechanics. More specifically, it is the purpose

background in either of thest^ mathematical discipliiies. some effort has

been expended to fill that gap. Other mathematical techni(nies are

tcjpics, again demonstrating the universality of the mathematical

:ipplic;ilioii, ('\(mi though the main discussion is on a more fundamental

Moi-(>over, the o\aini)les are dcsigncMl

tioiis of Problems involving substitu-

Robert H. Roy during that difficult period when the importance of

Chapter 1. Vector Algebra 1

1-1. Scalars and Vectors 1

1-2. Geometrical Representations of Vectors 2

Chapter 2. Newton's Laws; Concurrent Force Systems; P^quilibriuni of

2-1. Newton's Laws 12

Chapter 3. Vector Products 29

Chapter 4. Statics of Systems of Particles and of Rigid Bodies ... 52

4-1. Equilibrium of a System of Particles 52

Chapter 5. More Applications of Statics 77

5-1. Systems of Parallel Forces 77

Chapter 6, Derivatives of Vectors; Kinematics of a Particle .... 104

6-1. Derivative of a Vector 104

Chapter 7. Kinematics of a Rigid Body; Relative Motions .... 123

7-1. Displacements of a Rigid Body 123

Chapter 8. Dynamics of a Particle 153

8-1. General Considerations 153

8-10. Motion of a Particle in a Moving Coordinate System 181

Chapter 9. ^ ihruting Systems . .'

Chapter 10. Dynamics of Systems of Particles 223

Chapter 11. Moments and Products of Inertia 249

Chapter 12, Dynamics of Kigid Bodies 277

12-10. Rigid Body in Translation 310

Chapter 13. Lagrange's Equations 326

13-1. Introduction 326

Appendix A. Properties of Plane Geometric Shapes Al

1-1. Soalars and Vectors

only a number but also a direction. A

In addition to possessing a magnitude and direction, a vector quantity

t It might be instructive for the student interested in motivation to read at this

1-2. Geometrical Representations of Vectors

with regard to their origin under all circum-

C^ = A^ + B'- -^ 2AB cos 6

A 4-B = B -1- (1-1)

It can also be shown that the addi-

To prove this, we refer to Fig. 1-4.

Once we define the addition of vectors, the subtraction of xectors

An obvious fact is that if vector B is equal to vector A, then by Eq. (1-3)

On the other hand, A -f A gives, following the rule of addition, a new

nA -\- mA = (n -\- m)A (l-5a)

Hence the multiplication of a vector by a scalar is both distributive and