Mechanics Book

Maithripala D. H. S.
,
Dept. of Mechanical and Manufacturing Engineering,
Faculty of Engineering,
University of Ruhuna,
Sri Lanka.
Introduction to Mechanics, Dynamics and

Mechanisms
Class notes for ME2204, ME3305, ME4308
February 14, 2009
D.
c H. S. Maithripala, Dept. of Mechanical and Manufacturing Engineering, Univeristy of Ruhuna, Hapugala, Galle, Sri Lanka.
Sep 2008.
Preface
This is a compilation of notes used by me as class notes for ME2204 Engineering Mechanics, ME3305 Dy-
namics and Vibrations and ME4308 Theory of machines at the department of Mechanical and Manufacturing
Engineering, Faculty of Engineering, University of Ruhuna during the period of 2006 to 2008. They are far from
complete and I will be frequently updating them as time permits. I am deeply indebted to all the students who
have suffered through them and have tried out the exercises and have provided me with valuable input. I am sure
there are many errata and will appreciate greatly if you can please bring them to my notice by sending an e-mail
to mugalan at yahoo.com.
Hapugala, Galle, Sri Lanka, D. H. S. Maithripala

February 14, 2009
3
Contents
1 Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1.1 Configuration Space, Co-ordinates and Degrees of Freedom . . . . . . . . . . . . . . . . . . . . . . . . 8
1.1.2 Holonomic Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2 Particle Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.1 Kinematics of a Particle Moving in Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.2 Kinetics of a Particle Moving in Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.2.3 Relative Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3 Particle Motion in two-dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.4 Kinetic Energy of a Point Particle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.5 Rigid Body Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.5.1 Rotational Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.5.2 General Rigid Body Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.5.3 Kinetic Energy of a Rigid Body . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
1.5.4 Euler Angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
1.5.5 The Axi-symmetric Heavy Top . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.5.6 Forced Gyroscope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.6.1 Exercises on Particle Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.6.2 Exercises on Rigid Body Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.7 Answers to Selected Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2 Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.2 Linear Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.2.1 RC Circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
2.2.2 Spring Mass Damper System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.3 Nonlinear Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.3.1 Simple Pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
2.3.2 van der Pol Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.3.3 Double Pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.3.4 Lorenz Weather Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.3.5 Rigid Body Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.4 Lattice Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.4.1 The Linear Lattice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5
6 Contents
3 Input Output Representation of Linear Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

3.1 Graphical Description of a Input Output System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3.2 Mathematical Description of a System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3.2.1 First Order Linear Time Invariant System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.2.2 Second Order Linear Time Invariant System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.3 Transfer Function Model of Linear Time Invariant Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.3.1 Forced response of the nth order system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.4 Poles and Zeros of an Input Output System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.4.1 The Effect of the Poles of the System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.5 Bounded Input Bounded Output Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4 Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.2 Degrees of Freedom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.3 Linkage Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.3.1 Four Bar Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.3.2 Instantaneous Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.4 Cam Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.5 Gear Trains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Chapter 1
Mechanics
Fig. 1.1 Mechanics gone wrong: Tacoma Narrows bridge minutes before collapse. Figure curtesy of Wikipedia.
1.1 Introduction
Historically it is the the study of mechanics that evolved into what we call Engineering today. The following
quote by Wikipedia exemplifies this belief: “Important aspects of the fields of mechanical engineering, aerospace
engineering, civil engineering, structural engineering, materials engineering, biomedical engineering and biome-
chanics were spawned from the study of mechanics”. Mechanics deals with the scientific description of the world
as we perceive it. The language of mechanics is mathematics. There are two major branches of mechanics: classi-
cal mechanics and quantum mechanics. The latter has only a history of about hundred years while the former goes
beyond the period of written history. Classical Mechanics is concerned with describing macroscopic phenomena
while quantum mechanics is mainly concerned with describing microscopic phenomena. It is widely believed
that at “large” scales quantum mechanics approximates classical mechanics. Since at this stage of undergraduate
Engineering we are mostly interested in macroscopic systems we will only consider classical mechanics.
7
8 1 Mechanics
The subject of classical mechanics is generally divided into three branches; rigid-body mechanics, deformable-
body mechanics and fluid mechanics. In this class we will concentrate on learning the basics of rigid-body
mechanics while elsewhere in the mechanical Engineering program we will learn the basics of the other two
branches. The study of rigid-body mechanics begins with describing the motion of a mathematically abstracted
point particle that does not occupy any volume in space. A general rigid body is considered to consist of a large
number of such point particles where the distance between each of the particles remain fixed. The geometric
description of the motion of these objects is what is generally known as Kinematics while the study of the cause
of motion is referred to as Kinetics.
In classical mechanics, to describe the motion of point particles we need to define and accept three fundamen-
tal concepts; mass, time and space. The quantity called mass is a fundamental property that in a certain sense
describes the resistance to change of motion. Time is the quantity that measures the distance between two sepa-
rate events and is assumed to have an independent meaning. Finally it is also assumed that space where particle
motion occurs is a three-dimensional Euclidean space, that is a space where Euclidean geometry holds.
1.1.1 Configuration Space, Co-ordinates and Degrees of Freedom
The first step in describing the motion of a system of point particles is the complete specification of each of their
positions in space. This specification is referred to as the configuration of the system. The set of all possible
configurations is called the configuration space, Q. Given a certain configuration, that is a point in configuration
space, we need to describe it mathematically. The process of doing this is what we call co-ordinates. A little
more specifically a one-to-one correspondence between portions of the configuration space and portions of Rn
for some n is called a choice of co-ordinates for the configuration space1 . In general there can be many co-
ordinate choices for a given configuration space. The relationship between two different sets of co-ordinates is
called a co-ordinate transformation. An important fact about co-ordinates is that if the configuration space is
“smooth” then what ever the choice of co-ordinates we use, the number n in the identification is a constant2 . This
number is referred to as the dimension of the configuration space. The degrees of freedom (DOF) of a system
of particles is defined to be the dimension n of the configuration space. In the following we will explore these
notions in bit more detail using several examples.
A geometrically unconstrained point particle can occupy any point in three-dimensional Euclidean space.
Thus the configuration space of a geometrically unconstrained point particle can be identified with three-
dimensional Euclidean space. Let us first see how we can mathematically describe points in three-dimensional
Euclidean space, or in other words how to prescribe co-ordinates for three dimensional Euclidian space. Eu-
clidean space is a space where Euclidean geometric notions such as points, lines, angles and distances have the
usual meaning. To describe a point P in this space we begin by picking a point O and setting three mutually
perpendicular unit length axis at O. Such a set of axis is called an ortho-normal frame of reference. Labelling the
axis e1 , e2 , e3 to give a right handed orientation we can symbolically represent the frame as a matrix e = [e1 e2 e3 ]
where the e1 , e2 and e3 are to be taken as symbols and nothing more. Using such a frame any point P in Euclidian
space can be uniquely described using only the three measurements (numbers) x1 , x2 and x3 . These three numbers
describe respectively the distance to the point along the e1 , e2 , and e3 axis. Symbolically we describe this as
 
x1
OP , x1 e1 + x2 e2 + x3 e3 = e1 e2 e3  x2  = e x.
x3
In these notes the matrix  

x1
x =  x2  ,
x3
1 The precise mathematical definition requires that these portions be open sets and that sufficiently many such correspondences will
be needed to cover the entire configuration space. These details are beyond the scope of this class
2 In fact configuration spaces that satisfy this hypothesis can be called “smooth”
1.1 Introduction 9
will be referred to as the Euclidean representation matrix, or simply the representation matrix, of the position of
the point P with respect to the frame e. The components will be referred to as the position components of P with
respect to e.
If Q is another point in space and if its representation y with respect to e is given by
 
y1
y =  y2  ,
y3
then the distance between P and Q is assumed to be given by,

q
d(P, Q) , ||x − y|| = (x1 − y1 )2 + (x2 − y2 )2 + (x3 − y3 )2 .
This notion of distance is essentially what it means to be a Euclidean space. It allows us to mathematically define
what a right angle is and also when two straight lines are parallel. Convince yourself that this statement is true.
Fig. 1.2 Representation of a point P in two differen Euclidean frames e and b. The two frames are parallel to each other.
Observe that the introduction of the ortho-normal frame has allowed us to make a one-to-one correspondence
between points of three dimensional Euclidean space and the set of ordered triple of numbers, R3 . Namely by
identifying each point P with the uniquely corresponding ordered triple (x1 , x2 , x3 ). It has to be kept in mind
that this identification depends entirely on the choice of the ortho-normal frame e. For instance consider figure
1.2. The position of the particle P is described by the two different ordered triple of real numbers (x1 , x2 , x3 )
and (X1 , X2 , X3 ). Orhto-normal frames are also not the only means of identifying Euclidean space with R3 .
Curvilinear frames can also be used for this identification. For example, as shown in figure 1.3, we may use
the quantities (r, θ , φ ) to describe a point3 . All three sets of ordered triples (x1 , x2 , x3 ), (X1 , X2 , X3 ) and (r, θ , φ )
describe the same point but note that they are in general different. Thus the three descriptions provide three
different identifications with R3 . A particular identification of Euclidian space with R3 is what we refer to as a
set of co-ordinates for the Euclidian space. Therefore all three sets of ordered triples (x1 , x2 , x3 ), (X1 , X2 , X3 ) and
(r, θ , φ ) provide three different choices of co-ordinates for Euclidian space. However all identifications are with
3 Observe that not every point in three dimensional Euclidian space can be uniquely represented using these co-ordinates. For
instance none of the points on the e3 axis has a unique representation. Such points are called co-ordinate singularities of the co-
ordinate system.
10 1 Mechanics
R3 (have three numbers or components) and thus we say that the dimension of the configuration space of the
unconstrained particle is three.
When the co-ordinates are expressed using an ortho-normal frame they are called rectangular co-ordinates.
The relationship between two different co-ordinates are called a co-ordinate transformations. The rectangular
co-ordinates of the point P, (x1 , x2 , x3 ) are related to the spherical polar co-ordinates (r, θ , φ ) by
Fig. 1.3 Description of the point P using Spherical-Polar Co-ordinates. Figure copied from [11].
x1 = r sin φ cos θ ,
x2 = r sin φ sin θ ,
x3 = r cos φ .
Thus using spherical polar co-ordinates we can express the Euclidean representation matrix x as,
   
x1 r sin φ cos θ
x =  x2  =  r sin φ sin θ  .
x3 r cos φ
One important thing that we have to keep in mind is that what ever the co-ordinates we use to identify Euclidean
space with R3 we will always use the Euclidean representation matrix to represent a point in space.
When a particle is geometrically constrained the configuration space of the particle will no longer be the
three-dimensional Euclidean space. For instance if the particle is constrained to move such that its distance to
the origin is always fixed then the configuration space can be identified with the sphere and if it is additionally
constrained to move in a plane then the configuration space can be identified with the circle. A little less obvious
example is the planar double pendulum discussed in the following example.
Example 1.1. Consider the planar double pendulum shown in figure 1.4. The configuration space of mass m1 can
clearly be identified with the circle with origin at the fixed point and radius L1 . With respect to m1 the possible
configurations of m2 lie on the circle with origin at m1 and radius L2 . Thus if the configuration of m1 (that is
the location of m1 on the circle of radius L1 and origin at the fixed point) is known then the configuration of m2
can be uniquely specified if we specify the position of m2 on the circle with origin at m1 and radius L2 . Thus the
configuration space of the two particle system is the torus S × S.
1.1 Introduction 11
Fig. 1.4 The Double Planar Pendulum
The idea of co-ordinates for a general configuration space, such as a sphere, a circle or a torus, is developed
in a fashion similar to what we have done so far for Euclidean space. Roughly, we try to associate points
of the configuration space with an ordered n-tuple of numbers in an essentially unique way. That is we
identify either the entire configuration space, or a significant portion of it with with some portion of Rn .
Such an identification is called a choice of co-ordinates for the configuration space4 . If a particular co-
ordinate choice is incapable of covering all the points of a configuration space we may require several
other choices to represent these points as well. Then by “patching” them together we can “cover” all the
points in the configuration space.
Example 1.2. If the configuration space is a circle, as is the case for a simple blob pendulum, then the single
quantity θ is sufficient to uniquely describe every point except one point on the circle5 . Thus prescribing θ
allows us to uniquely identify almost all points on the circle with R and provides a choice of co-ordinates for the
circle. Since the identification has only one component or in other words is with R the circle is one dimensional.
To account for the problematic point we can start from a different point on the circle to get a different angle
correspondence. That gives us another choice of co-ordinates for the circle. Collectively both these co-ordinates
will cover all the points on the circle.
Example 1.3. Let us consider the sphere. Referring to figure 1.3 we can see that the two independent quantities
θ and φ uniquely describe every a point on the sphere except the North and South poles of the sphere6 . Thus
(θ , φ ) provide us with a choice of co-ordinates for the sphere and covers almost all points on the sphere. That is
we have prescribed a way of uniquely identifying any point, except the poles, on the sphere with a point in R2 .
Since the identification is with R2 the dimension of the sphere is two.
1.1.2 Holonomic Constraints
In general when a point particle is geometrically constrained to move in three-dimensional Euclidean space
the three quantities x1 , x2 , x3 , comprising the components of the Euclidean representation matrix x, will not be
independent. If the motion of the particle is geometrically constrained so that the quantities (x1 , x2 , x3 ) have
to satisfy some scalar expression h(x1 , x2 , x3 ) = 0 then the particle is said to be Holonomically constrained.
Holonomic constraints reduce the degrees of freedom. A single constraint of the form h(x1 , x2 , x3 ) = 0 reduces
the DOF by one, from three to two. If the number of such constraints are two then the DOF of the particle
reduces to one. For instance when the particle is constrained to move on a sphere of radius equal to one then
the constraint equation in the spherical-polar co-ordinates is h(r, θ , φ ) = r − 1 = 0 and the position matrix x will
become [sin φ cos θ sin φ sin θ cos φ ]T . Notice that the two independent measurements (θ , φ ) are sufficient to
5 The points described by θ = 0 and θ = 2π is the same point hence the uniqueness fails at this point.
6 Try to reason why this is so.
12 1 Mechanics
describe the Euclidean position uniquely when φ 6= 0, π . If the particle is constrained to move in a circle of radius
one that lies in the e1 , e2 plane then there are two constraints, h1 (r, θ , φ ) = r − 1 = 0, h2 (r, θ , φ ) = φ − π /2 = 0
and the position matrix x will become [cos θ sin θ 0]T . Once again notice that the single measurement θ is
sufficient to describe the Euclidean position uniquely. The general expression
n = 3− f, (1.1)
relates the DOF, n, of a particle moving in three dimensional space to the number of holonomic constraints, f ,
imposed on that particle. Also recall that this is equal to the dimension of the configuration space of the particle.
If N points are moving in space then the total degrees of freedom of the system of points (or the dimension of
the configuration space) is
n = 3N − f , (1.2)
where now f is the total number of holonomic constraints of the system of points. In this class what ever the
co-ordinates we use to describe the configuration of a particle we will always use a Euclidean representation
matrix to represent the point in space and analyze its motion in the underlying Euclidean space. The components
of the representation matrix will be functions of the co-ordinates.
1.2 Particle Motion
1.2.1 Kinematics of a Particle Moving in Space
Fig. 1.5 The Velocity of a Particle described with respect to a frame e.
Consider a point particle moving in space. The position P(t) of the particle at a given time t is expressed by
the representation matrix x(t) with respect to some frame e. Since the position of the particle is changing with
1.2 Particle Motion 13
time x(t) is a function of time and describes a curve in e that is parameterized by time t. The velocity of the point
P with respect to the e frame is defined to be the infinitesimal change of position with respect to e. That is
 
ẋ1
x(t + δ t) − x(t)
v(t) , lim = ẋ(t) =  ẋ2  .
δ t→0 δt ẋ3
Thus by definition v(t) gives the tangent to the curve at P(t) (Refer tofigure 1.5). The components
xi (t + δ t) − xi (t)
vi (t) = lim i = 1, 2, 3
δ t→0 δt
represent the infinitesimal change of position in the ei direction. Thus the correct way to visualize v(t) is to
consider it as the description of a point in a frame that is parallel to e but with origin at P(t) (Refer tofigure 1.5).
The magnitude of the velocity in the e frame is defined to be
q
v , ||v|| = v21 + v22 + v23 .
The acceleration measured in the e frame is defined to be the infinitesimal change of velocity in the e frame,
 
ẍ1
v(t + δ t) − v(t)
a(t) , lim = v̇(t) = ẍ(t) =  ẍ2  .
δ t→0 δt ẍ3
In general, since the position representation are different from frame to frame the velocities and accelerations
expressed in one frame will be different form those expressed with respect to another. Thus it is important to
always specify the frame with which they are expressed.
Example 1.4. Consider a particle constrained to move on a sphere of radius r (refer to figure 1.3). The repre-
sentation of P using an ortho-normal frame e fixed at the center of the sphere can be expressed using the polar
co-ordinates (θ , φ ) as  
sin φ cos θ
x = r  sin φ sin θ  .
cos φ
Then the velocity in the e frame expressed using the polar co-ordinates is
 
φ̇ cos φ cos θ − θ̇ sin φ sin θ
v = ẋ = r  φ̇ cos φ sin θ + θ̇ sin φ cos θ  .
−φ̇ sin φ
and the acceleration in the e frame expressed using the polar co-ordinates is
 
aφ̈ cos φ cos θ − φ̇ 2 sin φ cos θ − 2φ̇ θ̇ cos φ sin θ − θ̇ 2 sin φ cos θ − θ̈ sin φ sin θ
a = v̇ = r  φ̈ cos φ sin θ − φ̇ 2 sin φ sin θ + 2φ̇ θ̇ cos φ cos θ − θ̇ 2 sin φ sin θ + θ̈ sin φ cos θ  .
−φ̈ sin φ − φ̇ 2 cos φ
1.2.2 Kinetics of a Particle Moving in Space
Kinetics deal with the apparent causes of motion. In this section we investigate the cause of motion in classical
mechanics. Given the present “state” of a particle, the primary interest in mechanics is to be able to completely
predict the future “state” of it. More precisely we need to know what measurements are required at the present
14 1 Mechanics
time in order to be able to uniquely describe the position of a point of mass m for all future times. For macro-
scopic motions, the answer to this deep and profound question was provided by Newton. He postulated that for
macroscopic motions the acceleration a(t) of a given particle was not an independent (arbitrary) quantity but that
it was completely specified. More precisely in his second law he stated that in a given frame e
mẍ = m a(t) = f (t), (1.3)
where f (t) was completely known. For macroscopic motion this notion has been verified experimentally. In this
expression the mass of the point is taken to be a fundamental property of the point and is further assumed to be
a constant. From a mathematical perspective equation (1.3) describes a second order differential equation and
solving it for x(t) will enable us to predict the position for all future time. To solve this equation we need to know
the right hand side at every time instant t as well as know the initial conditions x(t0 ) = x0 and v(t0 ) = v0 . Thus
knowing the right hand side turned out to be a crucial step in describing the position and as such was termed
to be the cause of motion. Newton termed this constraint on mass times acceleration, the Force acting on the
particle in the ortho-normal frame e. Thus the force “felt” in a frame e is nothing but mass times the acceleration
in that frame. The force can not be directly measured and can only be inferred indirectly by resorting to other
physical reasoning and justifications.
Example 1.5. Consider the problem of a horizontal spring with one end fixed to a support and the other end fixed
to point mass P, of mass m. If we give an initial horizontal displacement to the point mass we know empirically
that the point mass will exhibit a simple harmonic motion if the air and other resistances on the particle is
negligible and the motion is small. That is if x(t) is the displacement of the mass from the unstretched position
and if the air resistance on the particle is negligible and the motion is small the position x(t) of the particle P at
a given time t is described by the second order differential equation
mẍ(t) = −k x(t).
Observing this expression it is evident that the mass times the acceleration of the particle is constrained and is
equal to −k x(t). Thus we would call −k x(t) the force exerted by the spring on the particle. Observe that we can
not directly measure this instead we can only infer it by measuring the deflection of the spring.
1.2.3 Relative Motion
In the previous sections it was pointed out that the position, velocity and acceleration of a point particle are
expressed with respect to a given reference frame. Choosing different frames of reference result in different
representations. Recalling that the force felt in a frame is nothing but mass times acceleration in that frame we
see that the forces felt in the two frames will also be different. In this section we will see how these different
descriptions with respect to different frames are related to each other.
1.2.3.1 Parallely Translating Frames
Let e be an ortho-normal frame of reference and b(t) be another ortho-normal frame of reference that is parallely
translating with respect to e. We will also refer to b(t) as a moving frame. Let the origin O0 (t) of the translating
frame have the representation o(t) with respect to the frame e. Consider a point P(t) moving in Euclidian space.
This point can be expressed in two different ways using the two different frames. Let x(t) and X(t) be the
representations of the point P(t) in the e and b(t) frames respectively (refer to figure 1.6). In Euclidian space the
meaning of b(t) being parallel to e is that the two representations x(t) and X(t) of P are related by,
x(t) = o(t) + X(t). (1.4)
Observing that symbolically this means
OP(t) = e x(t) = e o(t) + b(t) X(t)

Fig. 1.6 Frame b(t) is parallely translating with respect to frame e.
you will notice that this is what you traditionally wrote down as
OP(t) = OO0 (t) + O0 P(t),
for addition of “vectors”.
The velocity of the point P(t) measured in frame e is
v(t) = ẋ(t),
and measured in the moving frame is

V (t) = Ẋ(t).
Differentiating (1.4) we see that the two measured velocities are related by
v(t) = ẋ(t) = ȯ(t) + Ẋ(t) = ȯ(t) +V (t). (1.5)
The acceleration of the point P(t) measured in the e frame is
a(t) = v̇(t) = ẍ(t),
and measured in the moving frame is

A(t) = V̇ (t) = Ẍ(t).
Differentiating (1.5) we see that the two measured accelerations are related by
a(t) = v̇(t) = ö(t) + V̇ (t) = ö(t) + A(t). (1.6)
From (1.6) and (1.3) we have

f (t) = m a(t) = m(ö(t) + A(t)),
and hence that the total force experienced in the moving frame is
16 1 Mechanics
F b (t) = m A(t) = −m ö(t) + f (t). (1.7)
That is an observer moving with the translating frame will, in addition to the force f (t) felt in the e frame, also
experience another apparent force
Fapp = −m ö(t). (1.8)
Recall that force felt in a frame is simply the mass times the acceleration of the particle in that frame. This
particular type of apparent forces that arise as a consequence of the translational accelerating motion of the
reference frame are called Einstein forces. If either b(t) is fixed with respect to e or is moving at a constant
velocity with respect to e then ö = 0 and therefore the forces felt in both frames are the same. Such frames are
said to be inertial with respect to e.
Do exercise 1.6 at this moment. In exercise 1.6 you are asked to explain why a person standing on a scale
inside an elevator sees his or her weight doubled as the elevator accelerates up at a rate of g and sees the weight
reduced to zero if the elevator decelerates at a rate of g where g is the gravitational acceleration. You are also
asked to show that if, for some reason, the gravitational force field vanished and that the elevator was moving up
at an acceleration of g then the scale would show the correct weight of the person. This last observation shows
that a person inside the elevator can not distinguish between the following two cases:
a.) Gravity is present and the elevator is standing still (or moving at constant velocity).
b.) Gravity is absent and the elevator is accelerating upwards at a rate of g.
It is this observation that led Einstein to the conclusions of General Relativity and in particular that gravity is an
apparent force !!!
1.2.3.2 Rotating Frames
Fig. 1.7 Frame b(t) is rotating with respect to frame e about the coinciding origin O.
Let e be an ortho-normal frame of reference and b(t) be an ortho-normal frame of reference that is rotating
with respect to e. We will consider the case where both origins coincide at the point O in space (rotating frames).
Then the two frames are related by
b(t) = e R(t), (1.9)
where R(t) is a special orthogonal matrix (ie RT R = I and det(R) = 1). For 3-dimensional Euclidian motions
e = [e1 e2 e3 ], b(t) = [b1 (t) b2 (t) b3 (t)] and R(t) is a 3 × 3 matrix (refer to figure 1.7).
The point P(t) can be expressed in both fames as
OP(t) = e x(t) = b(t) X(t) = e R(t)X(t), (1.10)
where x(t) is the representation of the point P(t) with respect to the frame e and X(t) is the representation of the
point P(t) with respect to the moving frame b(t). Thus the two representations are related by
x(t) = R(t)X(t). (1.11)
The velocity of the point P(t) measured in the e frame is v(t) = ẋ(t), and measured in the moving frame b(t) is
V (t) = Ẋ(t).
Now differentiating (1.11) we have that

v(t) = ẋ(t) = Ṙ(t)X(t) + R(t)Ẋ(t) = R(t) RT (t)Ṙ(t)X(t) +V (t) = R(t)Vb (t). (1.12)
The quantity Vb (t) is given by

Vb (t) = RT (t)Ṙ(t)X(t) +V (t). (1.13)
Observe that Vb is the velocity v(t) expressed using the moving frame b(t). This is only a derived quantity as
compared to the other measured quantities. Since RT (t)R(t) = I, it follows that
ṘT R + RT Ṙ = 0,
and hence that

RT Ṙ = −(RT Ṙ)T = Ω̂ ,
where Ω̂ is a skew symmetric matrix.
In exercise 1.8 you are asked to show that the space of 3 × 3 skew-symmetric matrices can be identified with,
the set of ordered triple of real numbers, R3 , using the identification Ω 7→ Ω̂ where
 
0 −Ω 3 Ω 2
Ω̂ =  Ω3 0 −Ω1  , (1.14)
−Ω2 Ω1 0
and Ω = (Ω1 , Ω2 , Ω3 ). In exercise 1.10 you are also asked to show that for 3 × 3 skew symmetric matrices
Ω̂ X = Ω × X.
It can be shown that in the case of 3-dimensional Euclidian motion the quantity Ω corresponds to an instan-
taneous rotation of the b(t) frame about the axis Ω through O by an amount equal to the magnitude ||Ω ||. Thus
Ω is defined to be the angular velocity of points that appear to be fixed in the b(t) frame.
Substituting RT Ṙ = −(RT Ṙ)T = Ω̂ in (1.13) the velocity v expressed using the moving frame b(t) is given by
RT (t)v(t) = Vb (t) = Ω̂ (t)X(t) +V (t). (1.15)
Therefore the two velocities, v(t) and V (t) are explicitly related by

v(t) = R(t) Ω̂ (t)X(t) +V (t) . (1.16)
The acceleration of the point P(t) measured in the e frame is a(t) = v̇(t) = ẍ(t), and measured in the moving
frame is A(t) = V̇ (t) = Ẍ(t). Differentiating (1.12) we have that

a(t) = Ṙ(t)Vb (t) + R(t)V̇b (t) = R(t) RT (t)Ṙ(t)Vb (t) + V̇b (t) ,

= R(t) Ω̂ (t)Vb (t) + V̇b (t) = R(t)Ab (t), (1.17)
18 1 Mechanics
where
Ab (t) = Ω̂ (t)Vb (t) + V̇b (t),

= Ω̂ 2 (t)X(t) + 2 Ω̂ (t)V (t) + Ω̂˙ (t)X(t) + A(t). (1.18)
Observe that Ab is the acceleration of the particle in the e frame expressed in the moving frame b(t). This too is
only a derived quantity as compared to the other measured quantities.
From (1.17) and (1.18) the two accelerations, a(t) and A(t) are explicitly related by

a(t) = R(t) Ω̂ 2 (t)X(t) + 2 Ω̂ (t)V (t) + Ω̂˙ (t)X(t) + A(t) . (1.19)
From (1.18) and (1.3) we have that

f (t) = m a(t) = m(R(t)Ab (t)) = m R(t) Ω̂ 2 (t)X(t) + 2 Ω̂ (t)V (t) + Ω̂˙ (t)X(t) + A(t) , (1.20)
and hence in the rotating moving frame we have that
F b (t) = m A(t) = −m Ω̂ 2 (t)X(t) − 2m Ω̂ (t)V (t) − m Ω̂˙ (t)X(t) + RT f (t). (1.21)
The quantity RT f (t) is the force felt in the e frame but expressed using the b(t) frame. Thus an observer moving
with the rotating frame will, in addition to the rotating frame version of the force felt in the e frame, feel the
effect of the apparent force:
Fapp = −m Ω̂ 2 (t)X(t) − 2m Ω̂ (t)V (t) − m Ω̂˙ (t)X(t). (1.22)
The first term −m Ω̂ 2 (t)X(t) is known as the Centrifugal force, the second term −2m Ω̂ (t)V (t) is known as the
Coriolis force and the third term −m Ω̂˙ (t)X(t) is known as the Euler force.
Using these expression we can explain many physical effects. In the following we also use these equations to
show the effects of Earths rotation on gravity as well as on the formation of hurricanes and the motion of a long
pendulum known as the Foucault’s pendulum.
Effects of Earth’s Rotation about its Axis
Let us consider the effect that the Earth’s rotation about its axis has on the motion of a particle as observed in an
Earth fixed frame at a point O0 (with latitude α ) on the surface of the Earth. Consider figure 1.8. Let c(t) be an
ortho-normal frame fixed on Earth with center at O0 . The frame is oriented such that c2 is aligned perpendicular
to the Earth (vertical direction), c3 is aligned in the South - North direction (along the latitude), and c1 is aligned
in the East - West direction (along the longitude). Let b be a frame that is parallel to c and fixed on Earth with
origin at the center of the Earth O. Let a be a frame with origin at O and a3 aligns along the axis of rotation of
the Earth, a1 ≡ b1 and O0 lies in the a2 − a3 plane. Thus if b(t) = a(t)R1
 
1 0 0
R1 =  0 cos α − sin α  ,
0 sin α cos α
where α is the latitude angle and is a constant. Let e be a frame with origin, O, at the center of Earth and parallel
to a frame fixed on the sun. The frame is oriented such that the e3 direction coincides with the axis of rotation of
the Earth, ie a3 ≡ e3 and a(t) = eR2 (t)
 
cos θ − sin θ 0
R2 =  sin θ cos θ 0  ,
0 0 1
Fig. 1.8 Effects of Earth’s Rotation about its Axis: Frame c is fixed on the surface of the Earth with origin coinciding at O0 . Frame
b is paralle to c with center O coinciding with the center of the earth. Frame a is fixed on the earth with origin at O and a3 aligned
along the axis of rotation of the earth. The origin of frame e coincides with O and e3 is always aligned along a3 .
and θ is the angle of rotation of the Earth about its axis.
We are interested in analyzing the motion of a particle P as observed in the frame c(t). Let the representa-
tion of the point P in the c(t) frame be Xp (t), ie. O0 P = c(t)X p (t). Since OO0 = b(t)o and c is parallel to b we
have that OP = OO0 + O0 P = b(t)(o + X p ). Since by construction O0 is a distance r (r is the radius of the Earth)
away on the b2 axis, o = [0 r 0]T and is a constant.
Thus if x(t) is the representation of P in e we have
e x(t) = b(t) (o + X p (t)) = eR2 R1 (o + X p (t)) = eRX.
Thus
x = RX(t)
where R = R2 R1 and X = o + X p . From Newton’s equations in the rotating frame, (1.21), we have
m Ẍ(t) = −m Ω̂ 2 (t)X(t) − 2m Ω̂ (t)Ẋ(t) − m Ω̂˙ (t)X(t) + RT f (t). (1.23)
Let us decompose the total force in to two components f = f g + f e where f g is the gravitational force and f e is
the additional external forces. Gravity acts in the OP direction. Thus
mg mg
fg = − x=− RX,
||x|| ||X||
and hence
mg
RT f g = − X.
||X||
Substituting this in (1.23) and noting that the Earth is rotating at a constant rate (hence Ω̂˙ = 0) we have
20 1 Mechanics
X 1
Ẍ = −Ω̂ 2 X − 2Ω̂ Ẋ − g + RT f e , (1.24)
||X|| m
X 1
= −2Ω̂ Ẋ − gI3×3 + ||X||Ω̂ 2 + RT f e . (1.25)
||X|| m
Observe that these equations describe the motion of a particle as observed in a frame fixed on Earth with origin
coinciding that with the center of the Earth. From (1.24) it can be seen that the motion of the particle in the Earth
fixed frame, in addition to the gravity and external forces, is also influenced by the Centrifugal and Coriolis terms
that arise due to the rotation of the Earth. In equation (1.25) the Centrifugal term has been combined with the
gravity term. This allows one to see that the effective gravity felt by an observer will change with the latitude of
the location of the observer. We will explain this in a bit more detail at the end of this section.
Recall X = o + X p where o is a constant and X p is the representation of the point P in the Earth fixed frame c
fixed on the surface of the Earth at O0 . Then Ẋ = Ẋ p and Ẍ = Ẍ p thus from (1.25) we have that
Xp + o 1
Ẍ p = −2Ω̂ Ẋ p − gI3×3 + ||X p + o||Ω̂ 2 + RT f e , (1.26)
||X p + o|| m
describes the motion of a point particle m as observed in the Earth fixed frame with origin O0 on the surface of the
Earth. This is the case that applies to us when we observe particle motion. Since compared to the motion of the
particle o is very large (since the r is very large) we can approximate (X p + o) ≈ o and then (1.26) approximates
to
1
Ẍ p = −2Ω̂ Ẋ p − gI3×3 + rΩ̂ 2 χ + RT f e , (1.27)
m
where χ = [0 1 0]T . These equations can be used to describe many natural phenomena. For instance it explains
why a Hurricane formed in the Nothern hemisphere rotates in a counter–clockwise direction and a Hurricane
formed in the Southern hemisphere rotates in a clockwise direction (see figure 1.9). You are asked to show this
in exercise 1.22. It can also be used to show the shift of the oscillating plane in the Foucault’s pendulum.
(a) A Southern Hurricane (Catarina). (b) A Northern Polar Hurricane.
Fig. 1.9 Figures show the clockwise and anti-clockwise rotation of respectively a southern hemisphere and northern hemisphere
formed hurricane. Figures are courtesy of Wikipedia.
To compute Ω̂ consider Ṙ = RΩ̂ . Since R = R2 R1 differentiating we have that
Ṙ = Ṙ2 R1 = R2 Ω̂e R1 = R2 R1 RT1 Ω̂e R1 = RΩ̂ ,
and hence that

Ω̂ = RT1 Ω̂e R1 .
Now from Ṙ2 = R2 Ω̂e we have that  
0 −Ω e 0
Ω̂e =  Ωe 0 0  ,
0 0 0
where Ωe = θ̇ is the angular velocity of Earth about its axis of rotation. Thus we have
 
0 −Ωe cos α Ωe sin α
Ω̂ =  Ωe cos α 0 0 ,
−Ωe sin α 0 0
and  
1 0 0
Ω̂ 2 = −Ωe2  0 cos2 α − cos α sin α  .
0 − cos α sin α sin2 α
Thus explicitly written down equation (1.27) reads
1
Ẍ p1 = 2Ωe cos α Ẋ p2 − 2Ωe sin α Ẋ p3 + F1 , (1.28)
m
1
Ẍ p2 = −2Ωe cos α Ẋ p1 + (rΩe2 cos2 α − g) + F2 , (1.29)
m
1
Ẍ p3 = 2Ωe sin α Ẋ p1 − rΩe2 cos α sin α + F3 , (1.30)
m
where RT f e = F = [F1 F2 F3 ]T . From equation (1.30) it can be seen that the gravity felt at the point O0 on
the surface of the Earth is in fact rΩe2 cos2 α − g and depends on the latitude of the point O0 . At the poles
α = π /2, −π /2 and cos α = 0. Thus “true” gravity is only felt at the South and North poles. At the equator
α = 0 and thus the gravity felt is a minimum at the equator. This also tells us that away from the equator we feel
slightly heavier than at the equator!!!
π
The rotational velocity of the Earth, Ωe = 23h 256m −5
4s = 7.292 × 10 rad/s, is very small and thus for most
applications these effects can be neglected and then the equations (1.25) reduce to the usual
1 T e
Ẍ p = −gχ + R f . (1.31)
m
Explicitly written down they are:
1
Ẍp1 = F1 , (1.32)
m
1
Ẍp2 = −g + F2 , (1.33)
m
1
Ẍp3 = F3 . (1.34)
m
Example 1.6. Consider the following problem. At a point on Earth on the equator a cannon is dropped from a
vertical height of h with zero velocity. Where will the cannon land?
At the equator the latitude is zero, α = 0. Since no other external forces are present F = 0. Furthermore let
Ẋpi = Vpi . Then (1.28) — (1.30) reduce to
V̇p1 = 2ΩeVp2 ,
V̇p2 = −2ΩeVp1 + (rΩe2 − g),
V̇p3 = 0.
Integrating the third equation twice we have
X p3 (t) = Vp3 (0)t + X p3 (0).
From second equation since compared to g the other two terms are very small we can approximate the third
equation by
V̇p2 = −g.
Thus approximately we have
22 1 Mechanics
Vp2 (t) = −gt +Vp2 (0),

g
X p2 (t) = − t 2 +Vp2 (0)t + X p2 (0).
2
Now integrating the V̇p1 equation we have
Vp1 (t) = −Ωe gt 2 + 2ΩeVp2 (0)t +Vp1 (0),

Ωe g 3
X p1 (t) = − t + ΩeVp2 (0)t 2 +Vp1 (0)t + X p1 (0).
3
In the case of the cannon being dropped from a vertical height, h, we have that all the initial conditions except
for X p2 (0) = h is zero. Thus we have
Ωe g 3
X p1 (t) = − t ,
3
g
X p2 (t) = − t 2 + h,
2
X p3 (t) = 0.
2
pthe time that the cannon falls back on to the ground we have that X p2 (T ) = −gT /2 + h = 0 and thus that
If T is
T = 2h/g. Thus s
Ωe g 3 2hΩe 2h
X p1 (T ) = − T =− .
3 3 g
a cannon dropped from a vertical height,h, from a point O0 on the equator will land at a distance
That is,p
2hΩe /3 2h/g to the East from O0 . Approximately if h = 100 m then the cannon will fall 1 cm towards the
East.
1.2.3.3 General Moving Frames
Fig. 1.10 Frame b(t) is is translating and rotating with respect to e.
Let e be an ortho-normal frame of reference and b(t) be a moving ortho-normal frame of reference with
origins O and O0 respectively. In this section we consider the case where the frame b(t) is rotating with respect to
the frame e as well as translating with respect to e (ie. the origin O0 of the b(t) frame is also moving with respect
to the frame e). Then a point P(t) can be expressed in both fames as
1.3 Particle Motion in two-dimensions 23
OP(t) = e x(t) = e o(t) + b(t) X(t) = e (o(t) + R(t)X(t)) , (1.35)
where x(t) and o(t) are the representations of the points P(t) and O0 (t) in the frame e while X(t) is the represen-
tation of P(t) in the moving frame b(t). Thus the two representations of the point P are related by
x(t) = o(t) + R(t)X(t). (1.36)
The velocity of the point P(t) measured in the frame e is v(t) = ẋ(t), and measured in the moving frame is
V (t) = Ẋ(t).
Now differentiating (1.36) we have that

v(t) = R(t) RT (t)ȯ(t) + Ω̂ (t)X(t) +V (t) = R(t)Vb (t), (1.37)
where the quantity Vb (t)

Vb (t) = RT (t)ȯ(t) + Ω̂ (t)X(t) +V (t), (1.38)
is the velocity in e expressed with respect to the moving frame b(t). Let Vo (t) = RT (t)ȯ(t) be the velocity of the
point O0 (t) represented using b(t).
The acceleration of the point P(t) measured in the frame e is a(t) = v̇(t) = ẍ(t), and measured in the mov-
ing frame is A(t) = V̇ (t) = Ẍ(t). Differentiating (1.37) we have that

a(t) = Ṙ(t)Vb (t) + R(t)V̇b (t) = R(t) Ω̂ (t)Vb (t) + V̇b (t) ,
= R(t)Ab (t), (1.39)
where
Ab (t) = V̇o (t) + Ω̂ (t)Vo (t) + Ω̂ 2 (t)X(t) + 2 Ω̂ (t)V (t) + Ω̂˙ (t)X(t) + A(t), (1.40)
is the acceleration of the particle in e expressed in the moving frame b(t).

f (t) = m a(t) = m R(t) V̇o (t) + Ω̂ (t)Vo (t) + Ω̂ 2 (t)X(t) + 2 Ω̂ (t)V (t) + Ω̂˙ (t)X(t) + A(t) , (1.41)
and hence in the general moving frame we have that
F b (t) = m A(t) = −m V̇o (t) − m Ω̂ (t)Vo (t) − m Ω̂ 2 (t)X(t) − 2m Ω̂ (t)V (t) − m Ω̂˙ (t)X(t) + RT f (t). (1.42)
1.3 Particle Motion in two-dimensions
In this section we specialize the results derived in the previous sections to motion restricted to two dimensional
Euclidean space. First let us consider the case of rotating 2-D frames. Without loss of generality let us assume
that the motion takes place in the e1 , e2 plane and that the frame b(t) is rotated about e3 by an angle θ (refer to
figure 1.11). That is b3 (t) ≡ e3 and if b(t) = eR(t) the rotation matrix R(t) takes the form
 
cos θ − sin θ 0
R(t) =  sin θ cos θ 0  . (1.43)
0 0 1
Let
cos θ − sin θ
R2×2 = .
sin θ cos θ
24 1 Mechanics
Fig. 1.11 Frame b is rotated about the e3 axis.
Then in this compact notation

R2×2 0
R(t) = , (1.44)
0 1
where the zero in the first row is a 2 × 1 zero matrix and the zero in the second row is a 1 × 2 zero matrix. Observe
that R2×2 is a 2 × 2 special orthogonal matrix.
The representation of a point P in the e1 , e2 plane takes the form

 
x1
x
x =  x2  = 2×1
0
0
where
x1
x2×1 =
x2
in the e frame and  
X1
X
X =  X2  = 2×1
0
0
where
X
X2×1 = 1
X2
in the b(t) frame.
In these notations the relationship between the representations (1.11) reduces to
x2×1 (t) = R2×2 (t)X2×1 (t). (1.45)
Now we also have that

Ṙ2×2 0 R2×2 Ω̂2×2 0
Ṙ(t) = = , (1.46)
0 0 0 0
where
0 −Ω
Ω̂2×2 = ,
Ω 0
1.4 Kinetic Energy of a Point Particle 25
with Ω = θ̇ and thus

Ṙ2×2 = R2×2 Ω̂2×2 . (1.47)
The force felt in the e frame and represented in the e frame can be expressed as
 
f1
  f2×1
f = f2 =
f3
f3
where
f1
f2×1 = ,
f2
and in the b(t) frame as 
F1
T   F2×1
F = R f = F2 =
f3
F3
where
F1
F2×1 = .
F2
Finally using all of this we will have that the Newton’s equation (1.21) reduces to
mẌ2×1 (t) = −mΩ̂2×2

2
(t)X2×1 (t) − 2mΩ̂2×2 (t)Ẋ2×1 (t) − mΩ̂˙ 2×2 (t)X2×1 (t) + F2×1 (t), (1.48)
0 = f3 . (1.49)
Observe that equation (1.49) simply states that the force felt in the e3 direction should be zero. Also observe that
(1.48) is identical in form to (1.21). Thus from the beginning if we had simply neglected the e3 direction and
used e = [e1 e2 ] and b = [b1 b2 ] then b(t) = eR2×2 (t). Now following the same procedure of applying Newton’s
equations in the b(t) frame we would arrive at (1.48).
The two dimensional restriction for general moving frames follows exactly the same way with matrices ap-
propriately restricted to their 2 × 2 versions as described above, and therefor we omit it.
1.4 Kinetic Energy of a Point Particle
The kinetic energy KE p of a particle of mass m is defined with respect to an inertial frame. An inertial frame
is a frame that is either fixed with respect to the distant galaxies or is translating without rotations at a constant
velocity with respect to the distant galaxies. Let e be such an inertial frame. Then the kinetic energy of a point
particle of mass m is defined to be
1
KE p , m||v(t)||2 , (1.50)
2
where || ∙ || is the Euclidian norm in R3 and is defined to be ||X||2 = X T X = X ∙ X (physically this gives the length
of the vector X).
Since ||RX|| = ||X|| with respect to a frame b(t) rotating purely with respect to the inertial frame, e, we also
have that
1 1 1
KE p = m||v(t)||2 = m||RT (t)v(t)||2 = m||Vb (t)||2 .
2 2 2
Hence from (1.15) we have that,
1 1
KE p = m||Vb (t)||2 = m −X T Ω̂ 2 X + 2V T Ω̂ X + ||V ||2 . (1.51)
2 2
If the particle P is fixed with respect to the purely rotating frame then V (t) = Ẋ(t) = 0 and hence
26 1 Mechanics
1 1
KE p = m||Vb (t)||2 = m −X T Ω̂ 2 X . (1.52)
2 2
1.5 Rigid Body Motion 27
1.5 Rigid Body Motion
(a) (b)
(c) (d)
Fig. 1.12 Two examples of rigid body motion. Figure (a) is a picture of a Segway, figure (b) is a model Vertical Take Off and
Landing (VTOL) Vehicle, figure (c) is a model of the Mars rover land robot and figure (d) is a picture of a mobile land robot with a
manipulator.
1.5.1 Rotational Motion
First let us consider a rigid body that is free to rotate about a fixed point O. Let e be a “fixed” ortho-normal
frame and let b(t) be a body-fixed ortho-normal frame (which we will call the body frame). Both frames have
coinciding origins at O. The position of a point P in the body, at a time t, is given by the x(t) with respect to the
frame e and by X with respect to the body frame b(t). Observe that since the frame b(t) is fixed in the body, the
representation X is independent of time. At t = 0, the two frames coincide (ie. b(0) = e). We have already seen
that the two frames are related by
b(t) = e R(t)
and hence that
x(t) = R(t)X,
28 1 Mechanics
where R(t) is a special orthogonal matrix.Observe that the rest of the points are fixed with respect to P and that
therefore specifying P amounts to specifying the configuration of the rigid body. Since X is a constant specify-
ing R(t) amounts to the specification of P. In this fashion any given configuration of the body can be uniquely
identified with a special orthogonal matrix R. Thus the configuration space of a rigid body rotating about a fixed
point is the space of 3 × 3 special orthogonal matrices. This space can be shown to be a three dimensional space.
Thus a rigid body with one point fixed has 3 DOF. There are many ways of choosing co-ordinates for this space.
One common way of doing it is to use Euler angles.
Recall that,
Ṙ = R Ω̂ ,
is given by (1.14) for 3-dimensional Euclidian motions. It can be shown that in the case of 3-dimensional Euclid-
ian motion the quantity Ω corresponds to an instantaneous rotation of the body about the axis Ω by an amount
equal to the magnitude ||Ω ||. Thus in both cases Ω is defined to be the body angular velocity of the rigid body.
To analyze the kinematics of a rotating rigid body we consider it to be made up of a very large number of
point masses that are fixed with respect to each other. Let us analyze the mechanics of one of these point masses,
P. The velocity and acceleration of the point P at time t as measured in the inertial frame e is v(t) = R(t)Vb (t)
and a(t) = R(t)Ab (t) respectively where from (1.15)
Vb (t) = Ω̂ (t)X, (1.53)
and from (1.18)
Ab (t) = Ω̂ 2 (t)X + Ω̂˙ (t)X, (1.54)
are their respective body representations. Let
f p (t) = f pr (t) + f pe (t)
be the expression of the total force acting on the point mass P with respect to the fixed frame e. Here f pr is
the force that maintains rigidity and arises due to the interaction of the neighboring particles and f pe (t) is the
resultant external force expressed in the inertial frame. Their respective representations in the body frame are
RT (t) f pr (t) = Fpr (t) and RT (t) f pe (t) = Fpe (t). Due to the assumption of rigidity it follows that Fpr , the interaction
force as seen in the body, is a constant. Applying Newton’s equations we have that

f p (t) = m a(t) = m R(t)Ab = m R(t) Ω̂ 2 (t)X + Ω̂˙ (t)X ,

RT (t) f p (t) = m Ω̂ 2 (t)X + Ω̂˙ (t)X . (1.55)
Let us pre-multiply both sides of (1.55) by the 3 × 3 skew symmetric matrix X̂ and sum over all the points in the
body
∑ m X̂ Ω̂ 2
(t)X + ∑ p p
˙ (t)X = X̂(F r + F e (t)).
Ω̂ (1.56)
Consider individually each of the terms in the above equation. The internal constraint forces that maintain rigidity
are of the nature that they satisfy Newton’t third law, action equals reaction. Thus
∑ X̂Fpr = ∑ X × Fpr = 0.
Let
∑ X̂Fpe (t) = ∑ X × Fpe (t) = T (t), (1.57)
where X̂Fpe (t) = X × Fpe (t) is defined to be the moment of the force Fpe (t) about the fixed point of rotation O and
therefore T (t) as the resultant moment of the external forces about O. In the exercises you are asked to show that
∑ m X̂ Ω̂˙ (t)X = I Ω̇ (t),

and
∑ m X̂ Ω̂ 2 (t)X = − I Ω (t) × Ω (t),
where
I = ∑ m ||X||2 I3×3 − XX T , (1.58)
is defined to be the inertia tensor of the body about the fixed point O. Then we have that (1.114) reduces to
I Ω̇ = I Ω × Ω + T.
Thus the complete motion of a rigid body rotating about a fixed point in 3-dimensions is governed by
Ṙ = R Ω̂ , (1.59)
I Ω̇ = I Ω × Ω + T. (1.60)
The space of special orthogonal matrices is of dimension three. Thus at least three parameters are required to
parameterize R and hence (1.59) can be expressed by three first order ODE’s. Euler angles are a common choice
of parameterization of R. Clearly (1.60) gives three more first order ODEs.
Recall that Ω (t) was defined to be the body angular velocity. The body angular velocity expressed in the e
frame ω = R Ω is called the spatial angular velocity. The quantity Π (t) = I Ω (t) is defined to be the body an-
gular momentum. The body angular momentum expressed in the e frame π = R Π is called the spatial angular
momentum. Later we will show that when T = 0 the spatial angular momentum, π , is a constant.
In terms of the body angular momentum the equations of motion (1.59)–(1.60) take the form
Ṙ = R I[
−1 Π , (1.61)
Π̇ = Π × I −1 Π + I −1 T. (1.62)
The inertia tensor (1.58) is a symmetric matrix and it can be shown that it is positive definite. Consequently
it is always possible to find a body frame such that the Inertia tensor is diagonalized. For instance in an axi-
symmetric rigid body if the body frame b(t) is aligned along the axes of symmetry then it can be shown that
the inertia matrix I is diagonal. The diagonal elements are called the principle moments of inertia. Let the body
frame b(t) be chosen such that the inertia tensor is diagonalized
 
I1 0 0
I =  0 I2 0  . (1.63)
0 0 I3
The principle moments of inertia are thus given by I1 , I2 , I3 . In this frame (1.60) takes the form
I1 Ω̇1 = (I2 − I3 )Ω2 Ω3 + T1 , (1.64)

I2 Ω̇2 = (I3 − I1 )Ω3 Ω1 + T2 , (1.65)
I3 Ω̇3 = (I1 − I2 )Ω1 Ω2 + T3 , (1.66)
and (1.60) takes the form
(I2 − I3 )
Π̇1 = Π2 Π3 + T1 , (1.67)
I2 I3
(I3 − I1 )
Π̇2 = Π3 Π1 + T2 , (1.68)
I3 I1
(I1 − I2 )
Π̇3 = Π1 Π2 + T3 . (1.69)
I1 I2
30 1 Mechanics
Later it will be useful to know the Rate of change of spatial angular momentum and the magnitude of the
body angular momentum along a rigid body motion. Let us calculate these.
The rate of change of the Spatial Angular Momentum is:
d d d
π (t) = (R(t)Π (t)) = (R(t)I Ω (t)) = ṘI Ω + RI Ω̇ = RΩ̂ I Ω + R(I Ω × Ω + T )
dt dt dt

= R(t) Ω̂ I Ω + I Ω × Ω + T = R(t) (Ω × I Ω + I Ω × Ω + T ) = R(t)T. (1.70)
The rate of change of the Magnitude of the Body Angular Momentum is:
d d√ I Ω ∙ I Ω̇ I Ω ∙ (I Ω × Ω + T ) IΩ ∙ T
||Π (t)|| = IΩ ∙ IΩ = √ = √ =√ , (1.71)
dt dt IΩ ∙ IΩ IΩ ∙ IΩ IΩ ∙ IΩ
where once again the last equality follows from A ∙ (A × B) = 0.
1.5.1.1 Two-dimensional Rigid Body Rotational Motion
In this section we specialize the results derived in the previous sections to rigid body motion restricted to two
dimensional Euclidean space. That is rigid bodies restricted to rotate in the e1 , e2 plane. From the last part of
section 1.3 it follows that R(t) is a 2 × 2 special orthogonal matrix and Ω̂ is a 2 × 2 skew-symmetric matrix

0 −Ω
Ω̂ = , (1.72)
Ω 0
where Ω = θ̇ . Thus the quantity Ω corresponds to an instantaneous rotation of the rigid body about the fixed
point O by an amount equal to Ω .
Since Ω̂ 2 (t) = −Ω 2 (t)I, Newton’s equation (1.55) in two dimension reduces to

m −Ω 2 (t)X + Ω̂˙ (t)X = RT (t) f p (t) = Fpr + Fpe (t). (1.73)
Now if we pre-multiply both sides by X̃ where for X = [ X1 X2 ]T

X̃ = −X2 X1 ,
and sum over all the points in the body we obtain

∑ m − Ω 2
(t)X̃X + X̃ Ω̂ ∑ p p
˙ (t)X = X̃(F r + F e (t)).
This reduces to
Ω̇ (t) ∑ m ||X||2 = ∑ X̃(Fpr + Fpe (t)),
because it can be easily shown that the first term becomes zero and X̃ Ω̂˙ (t)X = Ω̇ ||X||2 .
Let us individually consider each of these terms. The internal constraint forces that maintain rigidity are of
the nature that they satisfy Newton’t third law, action equals reaction. Thus
∑ X̃Fpr = 0.
Let
∑ X̃Fpe (t) = T (t). (1.74)
If we define
I = ∑ m ||X||2 , (1.75)
as the Moment of Inertia of the 2-dimensional body about the axis of rotation then the Kinetics of a 2-dimensional
rigid body is given by
I Ω̇ = T (t).
Thus the complete motion of the rigid body in 2-dimensions is governed by
Ṙ = R Ω̂ , (1.76)
I Ω̇ = T (t). (1.77)
These equations are referred to as Euler’s 2-dimensional rigid body equations. Let us look at the T appearing in
(1.77) a bit more closely and show that it is the resultant force moment of the external forces about the axis of
rotation and hence corresponds to the same notion adopted for three dimensional rigid body motions. To show
this it suffices to show that X̃Fpe (t) is the moment of the force Fpe (t) about the axis of rotation. Considering the
e1 , e2 plane in which the rigid body is moving to be embedded in the 3-dimensional Euclidean space with a frame
e = [e1 e2 e3 ] and the body frame b(t) = [b1 (t) b2 (t) b3 (t)] with b3 (t) ≡ e3 , we see that the external forces at P
expressed in the body frame b(t) take the form
 e 
Fp 1
Fp =  Fpe 2 
0
and the representation of the point P in the body frame b(t) takes the form
 
X1
X p =  X2  .
0
Then it can be easily shown that the force moment of Fpe at P about O is
   
0 0
Tp = X̂ p Fp =  0 = 0 
−Fpe 1 X2 + Fpe 2 X1 X̃Fpe .
1.5.2 General Rigid Body Motion
In this section we consider the case where the body describes a general three dimensional Euclidean motion.
Chose a body frame b(t) and an inertial frame e. From (1.42) a point P fixed in the moving frame satisfies the
following relationship.
m V̇o (t) + m Ω̂ (t)Vo (t) + m Ω̂ 2 (t)X(t) + m Ω̂˙ (t)X(t) = Fpr + Fpe (t). (1.78)
First let us sum (1.78) over all the points in the body.

∑m V̇o (t) + ∑m Ω̂ (t)Vo (t) + Ω̂ 2 (t) ∑ m X(t) + Ω̂˙ (t) ∑ m X(t) = ∑ Fpr + ∑ Fpe (t). (1.79)
Now if we choose the b(t) frame so that its origin O coincides with the center of mass of the rigid body, then the
last two terms on the left hand side of (1.79) become zero and we have
M V̇o (t) + M Ω̂ (t)Vo (t) = F(t), (1.80)
where M = ∑ m, F(t) = ∑ Fpe (t) and we have set ∑ Fpr = 0 because the constraint forces occur in equal and
opposite pairs.
32 1 Mechanics
Let us now pre-multiply both sides of (1.78) by the 3 × 3 skew symmetric matrix X̂ and sum over all the
points in the body
∑ m X̂ V̇o (t) + ∑ m X̂ Ω̂ (t)Vo (t) + ∑ m X̂ Ω̂ 2 (t)X(t) + ∑ m X̂ Ω̂˙ (t)X(t) = ∑ X̂Fpr + ∑ X̂Fpe (t). (1.81)
If the body frame b(t) is fixed at the center of mass of the body then the fist two terms become zero and we
end up with the rigid body equations (1.60) for a purely rotational rigid body derived in section 1.5.1, Thus the
complete equations of motion for general rigid body motion where the body frame b(t) is fixed at the center of
mass of the body are
Ṙ(t) = R(t)Ω̂ (t), (1.82)

ȯ(t) = R(t)V0 (t), (1.83)
M V̇o (t) = −M Ω̂ (t)Vo (t) + F(t), (1.84)
I Ω̇ = I Ω × Ω + T. (1.85)
1.5.3 Kinetic Energy of a Rigid Body
1.5.3.1 Rotating Rigid Body
In a rigid body all the point particles that make up the body are fixed with respect to the body frame b(t). Thus
to compute the total energy of a purely rotating rigid body we sum up the kinetic energy of all the particles given
by (1.52) to obtain
1 1
KE = ∑ Tp = ∑ m||Vb (t)||2 = ∑ m −X T Ω̂ 2 X . (1.86)
2 2
In 2-dimensions it is easy to show that
1 1
∑ −m 2 X T Ω̂ 2 X = 2 I Ω 2 ,
and hence that
1
KE = I Ω 2 . (1.87)
2
In exercise 1.26 you are asked to show that in 3-dimensions
1 1
∑ −m 2 X T Ω̂ 2 X = 2 Ω T I Ω ,
and hence that
1 1
KE = Ω T I Ω = Ω ∙ I Ω . (1.88)
2 2
Later we will need the rate of change of kinetic energy given by

d 1 T Ω̇ T I Ω + Ω T I Ω̇
Ω IΩ = = Ω T I Ω̇ = Ω ∙ I Ω̇ = Ω ∙ (I Ω × Ω + T ) = Ω ∙ T. (1.89)
dt 2 2
Where the last equality Ω ∙ (I Ω × Ω + T ) = Ω ∙ T follows from the easily verifiable property of cross products
and dot products A ∙ (A × B) = 0, and the second equality follows from
Ω̇ T I Ω = (Ω̇ T I Ω )T = Ω T I T Ω̇ = Ω T I Ω̇ ,
since I is symmetric.
1.5.3.2 General Rigid Body
For a general rotating and translating body where the body frame is fixed at the center of mass of the body we
can show that the kinetic energy is given by
M 1
KE = ||V0 ||2 + Ω T I Ω . (1.90)
2 2
1.5.4 Euler Angles
Up to this point we have not considered how to parameterize the rotation matrix R. That is how one can introduce
co-ordinates for the space of 3 × 3 special orthogonal matrices.
Consider two frames e and b. Let us first see what the rotation matrix R that relates the two frames is for basic
simple rotations such rotations about each axis. First let b be a frame that is rotated about the e3 axis by an angle
θ . That is b3 ≡ e3 . Then the two frames related by b = e R3 (θ ) where
 
cos θ − sin θ 0
R3 (θ ) =  sin θ cos θ 0  ,
0 0 1
and θ is the angle of rotation of b about the fixed e3 , b3 axis.

Let R2 be the special orthogonal matrix that represents frame b where b is now a frame obtained by rotating
about the e2 axis. That is b2 ≡ e2 . Then b = e R2 (θ ) where
 
cos θ 0 sin θ
R2 (θ ) =  0 1 0  ,
− sin θ 0 cos θ
and θ is the angle of rotation of b about the b2 , e2 axis. Finally let R1 be the special orthogonal matrix that
represents frame b where b is now obtained by rotating about the e1 axis. That is b1 ≡ e1 . Then b = e R1 (θ )
where  
1 0 0
R1 (θ ) =  0 cos θ − sin θ  ,
0 sin θ cos θ
and θ is the angle of rotation of d about the b1 , e1 axis.
The matrix R obtained by
R = Ri (α )R j (β )Rk (γ ) (1.91)
is once again a special orthogonal matrix. It can be shown that any special orthogonal matrix R can be written in
this fashion for some angles (α , β , γ ). This allows us to parameterize R using the three angles (α , β , γ ). These
are called the i − j − k Euler angle parameterization of R (not all i, j, k are equal). It is important to note that for
certain angles this correspondence is not unique, meaning that there exists certain R such that the corresponding
choice for (α , β , γ ) is not unique. An example of the Euler angle parameterization of R using the three angles
(θ1 , θ2 , θ3 ) is  
c1 c3 − c2 s1 s3 c2 c3 s1 + c1 s3 s1 s2
R = R3 (θ1 )R1 (θ2 )R3 (θ3 ) =  −c3 s1 − c1 c2 s3 c1 c2 c3 − s1 s3 c1 s2  , (1.92)
s2 s3 −c3 s2 c2
where we have used the notation ci = cos θi , si = sin θi . This is called the 3-1-3 Euler angle parameterization of
R. In this parameterization θ1 is called the angle of precession, θ2 is called the angle of nutation, and θ3 is called
the angle of spin.
In terms of the 3-1-3 Euler angles the body angular velocities are
34 1 Mechanics
Ω1 = θ̇1 sin θ2 sin θ3 + θ̇2 cos θ3 , (1.93)

Ω2 = θ̇1 sin θ2 cos θ3 − θ̇2 sin θ3 , (1.94)
Ω3 = θ̇1 cos θ2 + θ̇3 , (1.95)
and the spacial angular velocities are
ω1 = θ̇3 sin θ2 sin θ1 + θ̇2 cos θ1 , (1.96)

ω2 = −θ̇3 sin θ2 cos θ1 + θ̇2 sin θ1 , (1.97)
ω3 = θ̇3 cos θ2 + θ̇1 . (1.98)
1.5.5 The Axi-symmetric Heavy Top
In this section we consider the axi-symmetric top in a gravitational field. The pivot point is along the axis of
symmetry. The distance from the pivot point to the center of mass is l. Then the force moment due to gravity is
T = mgl (RT ϒ ) × ϒ where ϒ is the direction of gravity as observed in the e frame. Thus the equations of motion
for the heavy top are
Ṙ = R Ω̂ , (1.99)
I Ω̇ = I Ω × Ω + mgl (R ϒ ) × ϒ .
T
(1.100)
To implement (1.101) and (1.102) we need to parameterize the rotation matrix R(t). Let a be a fixed frame and
Fig. 1.13 The Axi-symmetric Heavy Top. Figure copied from [11].
b be a frame fixed on the top as shown in figure 1.13. The configuration of the top is given by the rotation matrix
R where b(t) = a R(t). Using intermediate frames a0 (t) and f (t) we can parameterize R(t) using the 3-1-3 Euler
angles as follows. As shown in the figure we have that
a0 (t) = a R3 (φ ), f (t) = a0 (t) R1 (θ ), b(t) = f (t) R3 (ψ ),

Then if b(t) = a R(t)

R(t) = R3 (φ )R1 (θ )R3 (ψ ).
In this parameterization φ is called the angle of precession, θ is called the angle of nutation, and ψ is called the
angle of spin.
It is interesting to note that if we let Γ (t) = RT ϒ (the direction of gravity as seen in the body frame) and
differentiate it we can replace (1.101) and (1.102) by the equivalent set of equations,
Γ̇ = Γ × Ω , (1.101)
I Ω̇ = I Ω × Ω + mgl Γ × ϒ . (1.102)
Observe that they remove the need for Euler angles.
1.5.6 Forced Gyroscope
Fig. 1.14
For the axi-symmetric gyroscope, I1 = I2 = Il , Iz and the rigid body equations take the form
Il Ω̇1 = (Il − Iz )Ω2 Ω3 + T1 , (1.103)

Il Ω̇2 = (Iz − Il )Ω3 Ω1 + T2 , (1.104)
Iz Ω̇3 = T3 . (1.105)
Let us parameterize the rotation matrix from the frame e to d using the 3-1-3 Euler angles. We will need this
to express the force moment  
T1
T =  T2  .
T3
Let e be an Earth fixed frame with origin, O. A frame b is fixed to the outer gimbal of the Gyroscope, a frame
c is fixed to the inner gimbal and a frame d is fixed on the disc. Let d(t) = e R(t). The b(t) frame is fixed on the
outer gimbal so that b3 (t) ≡ e3 (t). Then b(t) = e R3 (θ1 (t)) where θ1 is the angle of rotation of b about the fixed
e3 , b3 axis. The frame c(t) is fixed on the inner gimbal so that c1 (t) ≡ b1 (t). Then c(t) = b(t) R1 (θ2 (t)) where θ2
is the angle of rotation of c about the c1 , b1 axis. The frame d(t) is fixed on the disc so that d3 (t) ≡ c3 (t). Then
d(t) = c(t) R3 (θ3 (t)) where θ3 is the angle of rotation of c about the c3 , d3 axis.
Now if d(t) = e R(t) we have that
36 1 Mechanics
 
c1 c3 − c2 s1 s3 c2 c3 s1 + c1 s3 s1 s2
R = R3 (θ1 )R1 (θ2 )R3 (θ3 ) =  −c3 s1 − c1 c2 s3 c1 c2 c3 − s1 s3 c1 s2  (1.106)
s2 s3 −c3 s2 c2
Neglecting the gimbal inertia, the force moments acting on the disc are calculated as follows. The moments
acting on the outer gimbal due to the fixed end.
 1
T1
T f = e T 1 = e  T21 
u1
The moments acting on the inner Gimbal due to the outer Gimbal.
 
u2
T I = b T 2 = b  T22 
T32
The moments acting on the disc due to the inner Gimbal.

 
T13
T D = c T 3 = c  T23  = d T
u3
Neglecting the inertia of the outer and inner gimbals we have
e T 1 = −b T 2 , b T 2 = −c T 3
and hence
T 1 = −R3 (θ1 ) T 2 , T 2 = −R1 (θ2 ) T 3 , T 3 = R3 (θ3 ) T, (1.107)
and
T 1 = R3 (θ1 )R1 (θ2 )R3 (θ3 ) T = R T.
From the first set of equations in (1.107) we have that T32 = −u1 and from the second set of equations in
(1.107) we have that T13 = −u2 . Substituting these in the second set of equations in (1.107) we have
   
u2 u2
 T 2  = R2  −T 3  .
2 2
−u1 −u3
The last equation in this set of equations give
(u1 − cos θ2 u3 )
T23 = .
sin θ2
Now since T 3 = R3 T the force moments on the disc expressed in d is given by

     
−u2 cos θ3 sin θ3 0 0 −1 0 u1
(u1 −cos θ2 u3 )  =  − sin θ3 cos θ3 0   1 0 − cot θ2   u2  .
T = RT3  sin θ2 sin θ2
u3 0 0 1 0 0 1 u3
  
sin θ3
sin θ2 − cos θ3 − sin θ3 cot θ2 u1
 cos θ3 
T = sin θ2 sin θ3 − cos θ3 cot θ2   u2  .
0 0 1 u3
Here u1 , u2 and u3 are the external moments applied about the rotation axis of the oute gimbal, the rotation axis
of the inner gimbal and the rotation axis of the disc.
Observe that the above expression for the force moment becomes singular when θ2 = 0 or θ2 = π . At these
angels the external moment directions on the outer gimbal and the disc coincides. This situation is commonly
referred to as gimbal locking.
38 1 Mechanics
1.6 Exercises
1.6.1 Exercises on Particle Motion
Exercise 1.1. Two point masses P1 and P2 are moving in three-dimensional Euclidean space such that P1 moves
on a sphere and the distance between P1 and P2 remains fixed. Specify the configuration space and a suitable set
of coordinates for the system. What is the DOF of the system.
Exercise 1.2. A bead is constrained to move on a frictionless wire that lies on a horizontal plane. The wire is bent
to a shape of a parabola (ie. y = x2 ). Find the constraint forces that keep the bead on the wire and the equation of
motion of the bead.
Exercise 1.3. For each of the systems shown in figures 1.15 to 1.17, derive the governing differential equations
using Newton’s law.
Fig. 1.15
Fig. 1.16
Exercise 1.4. Consider the 1-DOF electrostatic MEMS mirror model shown in figure 1.18. Find the governing
differential equations of the system. Note that the capacitance between two parallel plate capacitors are given by
q(t)2
c(l(t)) = l(t)
ε A and the attractive coulomb force between the two plates are given by f e (t) = 2ε A where A is the
cross sectional area of the plates.
Exercise 1.5. A particle P of mass m is constrained to move on the surface of a sphere of radius r and origin
O (refer to figure 1.19). Except for the constraints no other external forces act on the particle. Let frame e be a
frame with origin coinciding with O and fixed with respect to the sphere. The representation of the point P with
respect to e is x. Answer the following:
1. Show that xT x = r2
2. Show that the velocity of the particle is always tangential to the sphere (orthogonal to the position x, ie.
xT ẋ = 0).
1.6 Exercises 39
Fig. 1.17
Fig. 1.18 1-DOF Electrostatic MEMS model.
3. Differentiating the constraint xT ẋ = 0 show that the motion of the particle in the e frame is described by
1 ||ẋ||2
ẍ = fc (t) = − 2 x,
m r
where fc (t) is the representation of the constraint force in the e frame (Hint: The constraint that keeps a
particle on the surface of a sphere should act in the radial direction).
4. Explicitly write down the differential equation derived in item 3 above using spherical polar co-ordinates
(Hint: Use the results of Example 1.4).
Exercise 1.6. — Einstein’s box experiment: Explain why a person standing on a scale inside an elevator sees
his or her weight doubled as the elevator accelerates up at a rate of g and sees the weight reduced to zero if
the elevator decelerates at a rate of g. Also show that if, for some reason, the gravitational force field vanished
and the elevator was moving up at an acceleration of g then the scale would still show the correct weight of the
person.
Exercise 1.7. For the system shown in figure 1.20,
1. Derive the equations of motion as observed by an observer in a frame fixed to the surface on which the box is
moving.
2. Derive the equations of motion as observed by an observer moving with the box (ie. in the moving frame
co-ordinates).
Exercise 1.8. The space of 3 × 3 skew-symmetric matrices can be identified with R3 using the identification
Ω 7→ Ω̂ where  
0 −Ω 3 Ω 2
Ω̂ =  Ω3 0 −Ω1  , (1.108)
−Ω 2 Ω 1 0
40 1 Mechanics
Fig. 1.19
Fig. 1.20 Spring Mass Damper System in a Moving Box
and Ω = [Ω1 Ω2 Ω3 ]T .
Exercise 1.9. Show that the space of 2 × 2 skew-symmetric matrices can be identified with R using the identifi-
cation Ω 7→ Ω̂ where
0 −Ω
Ω̂ = . (1.109)
Ω 0
Using this identification show that Ω̂ 2 = −Ω 2 I2×2 where I2×2 is the 2 × 2 identity matrix.
Exercise 1.10. Show that for 3 × 3 skew-symmetric matrices
Ω̂ X = Ω × X. (1.110)
Exercise 1.11. The three rotating frames a(t), b(t), c(t) are related to a fixed frame e as shown in figure 1.21. Let
a(t) = eR1 (ψ ), b(t) = eR2 (φ ), and c(t) = eR3 (θ ). Show the following
     
1 0 0 cos φ 0 sin φ cos θ − sin θ 0
R1 (ψ ) =  0 cos ψ − sin ψ  , R2 (φ ) =  0 1 0  , R3 (θ ) =  sin θ cos θ 0 
0 sin ψ cos ψ − sin φ 0 cos φ 0 0 1
and
1.6 Exercises 41
     
0 0 0 0 0 φ̇ 0 −θ̇ 0
RT1 Ṙ1 = Ω̂1 =  0 0 −ψ̇  , RT2 Ṙ2 = Ω̂2 =  0 0 0  , RT3 Ṙ3 = Ω̂3 =  θ̇ 0 0 
0 ψ̇ 0 −φ̇ 0 0 0 0 0
Fig. 1.21
Exercise 1.12. Now let us write down Newton’s equations for the particle described in exercise 1.5 but using a
frame b(t) with respect to which the particle appears to be fixed. Without loss of generality we may align the
b3 axis along OP (refer to figure 1.35). Let b(t) = e R(t) and X be the representation of the point P in b(t) (In
this case since P is aligned along b3 , X = [0 0 r]T ). Since P appears fixed in b, X is a constant. Answer the
following:
Fig. 1.22
1. Show that ||ẋ|| = ||Ω̂ X|| where Ṙ = RΩ̂ (Hint: use the fact for special orthogonal matrices R are length
preserving transformations, ie ||RY || = ||Y ||).
2. Show that Newton’s equations in the b(t) frame correspond to
42 1 Mechanics
||Ω̂ X||
2
Ω̂˙ X = −Ω̂ 2 X − X.
r2
3. Write down R in terms of the Spherical Polar co-ordinates and evaluate Ω̂ in terms of them (Hint: Use results
of exercise 1.11). Convince yourself that writing down the Newton’s equations in the above form is much
more convenient than writing them down using the e frame as derived in exercise 1.5.
Exercise 1.13. Consider the particle described in exercise 1.12. Recall, Ṙ = RΩ̂ , and that Ω corresponds to the
angular velocity of any point that appears to be fixed in b. Thus the angular velocity of P with respect to e is Ω .
Expressed in the e frame the angular velocity has the the representation ω = RΩ . Show that ẋ = ω × x and hence
that ω ∙ ẋ = ω T ẋ = 0. This shows that ω , x, ẋ are mutually perpendicular to each other and form a right handed
axis system.
Exercise 1.14. Consider 2-dimensional Euclidean motion considered in section 1.3. In this case, explicitly write
down the Newton’s equations given by (1.48).
Exercise 1.15. Particle motion in a circle: Consider a particle constrained to move in a circle of radius r. Let e
be a frame fixed on the circle and x be the representation of the particle with respect to e.
1. Using the results of exercise 1.5 show that the velocity of the particle ẋ is tangential to the circle.
2. Using a frame b(t) that is moving with the particle (ie. the particle appears fixed in b(t)), show that ||ẋ|| = rΩ
where Ω is the angular velocity of the particle.
3. Show that the constraint force is radial and is given by fc (t) = −m Ω 2 x and that the motion of the particle in
the e frame is described by
ẍ = −Ω 2 x.
4. Show using Newton’s equations in the b(t) frame that the angular velocity, Ω , of the particle is a constant.
5. Show that the position components x1 and x2 behave sinusoidally (ie. x1 (t) = r sin (Ω t + φ1 ) and x2 (t) =
r sin (Ω t + φ2 ) ).
Exercise 1.16. — 2-dimensional particle motion in polar co-ordinates: For a particle moving in 2-dimensions,
write down explicitly the motion of the particle as observed in the moving frame b(t) = [er (t) eθ (t)] (refer to
figure 1.23). Also write down the apparent forces acting on the particle as observed by an observer in the moving
frame. If the particle is constrained to move in a circle write down the equations of motion and the constraint
forces. Compare the results of this with those obtained for exercise 1.15.
Fig. 1.23
Exercise 1.17. Consider a ball of mass m constrained to move as shown in figure 1.24. The disk is rotating at
a constant angular rate of Ω . Write down the equations of motion of the ball and also the constraint forces.
Simulate the motion using MATLAB (you may use R = 1, Ω = 1, m = 1, k = 2). When the mass is sinusoidally
1.6 Exercises 43
forced in the direction along the slot the constraint force acting on the walls, if measured, can be used to find
the angular velocity of the disk. Explain this. This is essentially the principle used in many commercially used
vibrating MEMS gyroscopes (angular rate sensors).
Fig. 1.24
Exercise 1.18. Consider the setup shown in figure 1.25. The disk can rotate in a horizontal plane. C is a parallel
plate capacitor. The fixed plate is rigidly attached to the center of the disk O while the movable plate is rigidly
attached to the mass P. The end of the spring of stiffness kl , (the point O0 ) can freely move inside the circular
slot. When the mass is at a distance l0 away from the origin the spring kl is unstretched. It can be assumed that
OPO0 remains a straight line. Assuming small values of θ the force exerted by the spring kθ can be assumed to
equal kθ l θ and act in a direction perpendicular to OP and in the plane of the disc where l is the distance OP
(this spring is a simplified representation of a torsional spring and hence assume that it does physically obstruct
the angular motion of OPO0 ). When a voltage is applied across the capacitor plates the resulting Coulomb forces
exert a force, fc (t) on the particle. The voltage across the capacitor is varied such that fc (t) = m sin (ω t).
1. If the disc is rotating at a constant angular rate of Ω derive the equations of motion of the particle.
2. Simulate the motion of the particle using MATLAB. Use r = 2, l0 = 1, kl /m = (2π × 100)2 , kθ /m = (2π ×
10)2 , ω = 2π × 11, Ω = 300 r.p.m and initial conditions l(0) = l0 , l(0)˙ = 0, θ (0) = 0, θ̇ (0) = 0. Show all
necessary graphs and attach all MATLAB work that is needed to produce these graphs.
3. Leaving all other parameters constant and using the same initial conditions as above simulate the motion for
three different values of Ω in the range of 100 r.p.m to 1000 r.p.m
4. Using above results discuss how this device may be used to measure the angular rate of change of the disk.
This captures the basic operating principle of a vibrating MEMS gyroscope.
5. Discuss what modifications you would do to change the operating range.
6. Write a two page introduction to MEMS gyroscopes. The introduction should include methods of fabrication,
device types and usages and the current state of the art.
Exercise 1.19. At a point on the equator a cannon is fired towards the West with an initial velocity of 100ms and
a firing angle of 45o with the equator. Taking into consideration the rotation of the Earth. Find the horizontal
distance to the landing point of the cannon. Compare these results with those obtained by neglecting the effects
of rotation of the Earth.
Exercise 1.20. Prove that the Earth’s orbit around the sun is an ellipse (Kepler’s First Law). You may assume
that the Earth and the Sun are point particles of mass ME and MS respectively and that the Gravitational attraction
force between two masses is of magnitude GMrS2ME where G is the universal gravitational constant and r is the
distance between the Earth and the Sun. Recall that the equation of an ellipse is given by
1+e
r = r0
1 − e cos θ
44 1 Mechanics
Fig. 1.25 The Basic Operating Principle of a Vibrating MEMS Gyroscope
Fig. 1.26
r1 −r0 b
√
where e = r1 +r0 is the eccentricity and a = 1 − e2 (refer to figure 1.26).
Exercise 1.21. Consider the case of two observers E and B. Observer E is positioned in a 2-dimensional ortho-
normal frame, e = [e1 e2 ] and the observer B is positioned in a rotating ortho-normal frame b(t) = [b1 (t) b2 (t)].
Both frames have coinciding origins. Discuss the following for the case where the frame b(t) with observer B is
rotating at a constant angular rate of θ̇ (t) = Ω in the counter clockwise direction.
1. The observer E sees a particle P to be fixed with respect to him. Show that B observes the particle to be
moving in a circle in a clockwise direction.
2. The observer B sees a particle P to be fixed with respect to him. Show that E observes the particle to be
moving in a circle in a counter-clockwise direction.
3. At time t = 0 observer B takes a particle of mass m and slowly releases it at a point long the b1 (t) axis that
is at distance l0 away from the origin. Neglecting friction between the plane and the particle, simulate using
MATLAB, the motion of the particle as seen by B. You may use m = 1, l0 = 1 and Ω = 1.
4. Observer B takes a particle of mass m and constraints it so that it can move only along the b1 (t) axis and
it is connected to the origin with a spring of un-stretched length l and spring constant k. He then slowly
releases it at time t = 0 a distance l away from the origin (ie. at the un-stretched position of the spring).
Find the constraint force at a time t and simulate the motion as observed by B using MATLAB. You may use
m = 1, k = 2, Ω = 1 and l = 1. If the frame was fixed and not rotating (ie. Ω = 0) what would be the expected
motion of the particle.
1.6 Exercises 45
5. Simulate the motion of the particle in item 4 above as observed by E.

6. Assume that a central inward force field is acting in the e frame where the intensity at point P is equal to r12
Newton’s per unit mass, where r is the distance from the origin of the b(t) frame to the point P. A particle
of unit mass is released in the rotating frame at a distance of 0.5 units away from the origin and with zero
velocity. Write the equations that describe the behavior of the particle as observed in the rotating frame and
using MATLAB simulate the motion of the particle as observed in the rotating frame b(t).
Exercise 1.22. Explain using equation (1.25) why a Hurricane that has formed in the Northern hemisphere has a
counter clockwise rotation and a Hurricane that has formed in the Southern hemisphere has a clockwise rotation
(see figure 1.9). Compare your results with the simulation results obtained in exercise 1.21.6.
Exercise 1.23. — Self study on vibrating MEMS Gyroscopes: The operation of a vibrating MEMS Gyro-
scope relies on the presence Coriolis forces. Write a two page report on vibrating MEMS gyroscopes. Include in
the report an explanation of its working principle, its use, and methods of fabrication.
1.6.2 Exercises on Rigid Body Motion
Exercise 1.24. Prove the following properties of cross products in R3

1. A × (B ×C) = (A ∙C)B − (B ∙ A)C
2. A ∙ (B ×C) = C ∙ (A × B) = B ∙ (C × A)
3. (Ω × X) ∙ (ω × X) = (Ω ∙ ω )||X||2 − (Ω ∙ X)(ω ∙ X)
4. Ω ∙ (Ω × X) = 0 = X ∙ (Ω × X)
Here we use the notation A ∙ B = AT B where A, B are both column matrices. Note that the last property tells us
that (Ω × X) is perpendicular to both X and Ω .
Exercise 1.25. Show that x(t) × f pe = R (X × Fpe ) where f pe = R Fpe . (Hint: First show using the properties of cross
product that R(a × b) = Ra × Rb)
Exercise 1.26. Using the properties proven in exercise 1.24, show that
1 1
∑ −m 2 X T Ω̂ 2 X = 2 Ω T I Ω ,
where
I = ∑ m ||X||2 I3×3 − XX T ,
is defined to be the inertia tensor of the body about the fixed point O, and
∑ m X̂ Ω̂˙ (t)X = I Ω̇ ,
and
∑ m X̂ Ω̂ 2 (t)X = − I Ω × Ω .
Exercise 1.27. Calculate the moment of Inertia tensor I of each of the solid figures shown in figure 1.27.
Exercise 1.28. Prove the parallel axis theorem of moments of inertia.
Exercise 1.29. Prove the perpendicular axis theorem of moments of inertia for planar rigid bodies.
Exercise 1.30. Find the moment of inertia tensor of the shaft and arm system shown in figure 1.30. The radius
of the shaft is rs , the length of the shaft is ls and the radius of the arm is ra .
Fig:RotatingSpringPendulum
46 1 Mechanics
Fig. 1.27 Calculate the inertia tensor of each of these solid shapes.
Exercise 1.31. Consider the mechanical system shown in the figure 1.28. The moment of inertia of the hoop
about the vertical axis of rotation is I. The radius of the hoop is r. The mass of the bead on the hoop is m. A
viscous frictional torque acts about the vertical shaft of rotation. The magnitude of this torque is proportional
to the rotational speed about the vertical axis and C is the constant of proportionality. Answer the following
questions. Derive the equations of motion of the system. For I = 1, m = 1, r = 1,C = 1, g = 1 simulate using
MATLAB the behaviour of the system for the case where initially the mass point P is displaced by an angle
7π /6 from the vertical.
Fig. 1.28 Bead on a Hoop. Figure copied from [11].
Exercise 1.32. Figure 1.29 shows a schematic representation of the Free/Forced vibration apparatus in the ap-
plied mechanics lab. The apparatus consists of a beam pivoted at O. A damper is attached to the point A and a
spring is attached to the point B. The other end of the spring, P, is constrained to move vertically. When the beam
is horizontal and the point P has zero displacement with respect to the reference, that is when θ = 0 and y = 0,
the spring is at its natural length (unstretched or uncompressed). The spring constant is k the damping constant
is C and the moment of inertia of the beam about O is I. Let OA = Lc and OB = Lk . Neglecting the thickness
of the beam and assuming that the angle of deflection of the beam is small find the governing equations of the
motion.
Exercise 1.33. Consider the mechanical system shown in figure 1.30. The moment of inertia of the shaft plus the
arm about the vertical axis of rotation is I. The mass of the point P is m. The un-stretched length of the spring is
L0 . Neglecting friction and the moment of inertia of the spring derive the equations of motion of the system and
simulate them using MATLAB. Also find the Kinetic Energy of the system.
1.6 Exercises 47
Fig. 1.29 A schematic representation of the Free/Forced Vibration Apparatus in the Applied Mechanics Lab.
Fig. 1.30 Rotating Spring Pendulum. Figure copied from [11].
Exercise 1.34. Consider the ball inside the swinging hoop shown in figure 1.31. The hoop has a mass distribution
of ρ per unit length and a radius of r. Neglecting friction between the ball and the hoop write down the constraint
forces acting on the ball. Write down Euler’s rigid body equations for the hoop and using this expression elimi-
nate the constraint forces appearing in the equations for the ball and find the equations that govern the motion of
the entire system.
Exercise 1.35. Consider the Gyroscope shown in figure 1.32. Using the results of section 1.5.6 derive the equa-
tions of motion.
Exercise 1.36. For the single bar pendulum shown in figure 1.33 derive the equations of motion using Euler’s
2-dimensional rigid body equations and also write down the kinetic energy of the system.
Exercise 1.37. For the double bar pendulum shown in figure 1.34 write down the kinetic energy of the system.
48 1 Mechanics
Fig. 1.31 Ball on Swinging Hoop. Figure copied from [11].
Fig. 1.32 The Gimbal Gyroscope
Fig. 1.33 Simple Pendulum
1.7 Answers to Selected Exercises
Answer to Exercise 1.5
In a Euclidean frame e, with no external force fields present, a particle P is constrained to move on a sphere.
Since the particle is constrained to move on a sphere of radius r we have that
r2 = x12 + x22 + x32 = ||x||2 = xT x = constant,

1.7 Answers to Selected Exercises 49
Fig. 1.34 Double Pendulum
where x is the representation of the particle in the e frame. Differentiating the constraint xT x = constant we have
that
2xT ẋ = 0,
thus the velocity in the e frame ẋ is orthogonal to x and hence lies tangent to the sphere. Differentiating the
constraint xT ẋ = 0 we have that
xT ẍ + ẋT ẋ = 0.
Since the particle is constrained to move on a sphere we see that the constraint force acting on the particle should
be normal to the surface (ie. lies in the radial direction) and hence that its representation in the e frame is of the
form
fc = α x.
Thus from Newton’s equations in the e frame
1 α
ẍ = fc = x.
m m
Hence we have that
α T
x x + ẋT ẋ = 0,
m
and that
||ẋ||2
α = −m ,
r2
2
and the constraint force is fc = −m ||ẋ||
r2
x. Now the motion of the particle is described by
1 ||ẋ||2
ẍ = fc = − 2 x.
m r
Let f be the total force felt in a frame e. Then if F is the force felt in a parallely translating frame b(t)then we
know that
F = −mö + f ,
where o is the co-ordinates of the origin of b(t) with respect to e. Newton’s equations in the e frame are
1
ẍ = f,
m
50 1 Mechanics
Newton’s equations in the moving frame are

1
Ẍ = F
m
For the problem of the elevator let e be the frame fixed on the Earth and let b(t) be the frame fixed on the
elevator. The weight measured by a scale is the total force exerted by the scale W . The total force acting on the
person as observed from the e frame is
f = −mg +W.
When the elevator is at rest,

ö = 0,
and hence from the first equation
F = f,
Applying Newton’s equations in the elevator we have
mẌ = F = f = −mg +W,
The person is at rest with respect to the elevator and therefore X = const and Ẍ = 0 and hence
0 = −mg +W,
and
W = mg.
Thus when the elevator is at rest the weight measured by the elevator is the correct weight of the person.
When the elevator is moving at a constant acceleration of g,
ö = g,
and hence from the first equation,

F = −mg + f ,
Applying Newton’s equations in the elevator
mẌ = F = −mg + f = W − 2mg,
0 = W − 2mg,
and
W = 2mg.
Thus when the elevator is accelerating constantly at g, the weight measured by the scale is twice as large as the
correct weight of the person.
When the elevator is moving at a constant acceleration of g but gravity is absent, Then
f = W.
and
ö = g,
and since
F = −mg + f ,
Applying Newton’s equations in the translating frame
mẌ = F = −mg + f = −mg −W,
0 = −mg +W,
and
W = mg.
Thus when the elevator is moving at a constant acceleration of g but gravity is absent the weight read by the
scale is the same as the correct weight of the person.
Fig. 1.35
Recall that Ṙ = RΩ̂ . Differentiating x = RX we have that
ẋ = ṘX + RẊ = RΩ̂ X.
where we have used the fact that Ẋ = 0 since P appears to be fixed in b(t). Thus
||ẋ|| = ||RΩ̂ X|| = ||Ω̂ X|| = ||Ω × X||.
Newton’s Equations in the b frame are
mẌ = −mΩ̂ 2 X − 2mΩ̂ ẋ − mΩ̂˙ X + RT fc ,

2
where fc = −m ||ẋ||
r2
x. Since x = RX and X is a constant we have
52 1 Mechanics
||Ω̂ X|| 2
0 = −mΩ̂ 2 X − mΩ̂˙ X − m X,
r2
and that
||Ω̂ X|| 2
Ω̂˙ X = −Ω̂ 2 X − X.
r2
Now let us explicitly write down these equations. By construction of the frames X = [0 0 r]T . Using results
of exercise 1.11 we see that a = eR3 (θ ) and b = aR2 (φ ). Thus b = eR = eR3 (θ )R2 (φ ) and R = R3 (θ )R2 (φ ).
Differentiating we have
Ṙ = Ṙ3 (θ )R2 (φ ) + R3 (θ )Ṙ2 (φ ).
Now
Ω̂ = RT Ṙ = RT2 (φ )RT3 (θ )Ṙ3 (θ )R2 (φ ) + RT2 (φ )RT3 (θ )R3 (θ )Ṙ2 (φ ) = RT2 (φ )Ω̂3 R2 (φ ) + Ω̂2 .
Evaluating this we have  
0 −θ̇ cos φ φ̇
Ω̂ =  θ̇ cos φ 0 θ̇ sin φ  .
−φ̇ −θ̇ sin φ 0
Then  
0 −θ̈ cos φ + θ̇ φ̇ sin φ φ̈
Ω̂˙ =  θ̈ cos φ − θ̇ φ̇ sin φ 0 θ̈ sin φ + θ̇ φ̇ cos φ  .
−φ̈ −θ̈ sin φ − θ̇ φ̇ cos φ 0
and  
−(φ̇ 2 + θ̇ 2 cos2 φ ) −θ̇ φ̇ sin φ −θ̇ 2 cos φ sin φ
Ω̂ 2 =  −θ̇ φ̇ sin φ −θ̇ 2 θ̇ φ̇ cos φ .
−θ̇ cos φ sin φ θ̇ φ̇ cos φ −(φ̇ 2 + θ̇ 2 sin2 φ )
2
Substituting these in the Newton’s equations
||Ω̂ X|| 2
Ω̂˙ X = −Ω̂ 2 X − X.
r2
we have    2   
φ̈ θ̇ cos φ sin φ 0
r  θ̈ sin φ + θ̇ φ̇ cos φ  = r  −θ̇ φ̇ cos φ  − (φ̇ 2 + θ̇ 2 sin2 φ )  0  .
0 (φ̇ 2 + θ̇ 2 sin2 φ ) r
Which gives us
φ̈ = θ̇ 2 cos φ sin φ ,
θ̈ = −2θ̇ φ̇ cot φ .
The particle P(t) has the representation

X (t)
X(t) = 1
X2 (t)
in the rotating frame b(t).
The velocity of the particle P(t) measured in the rotating frame is

V (t)
V (t) = 1 .
V2 (t)
The acceleration of the particle P(t) measured in the rotating frame is


A1 (t)
A(t) = .
A2 (t)
It can be shown that

b(t) = b1 (t) b2 (t) = e1 e2 R(t) = e R(t).
where,
cos θ − sin θ
R(t) = .
sin θ cos θ
Observe that R(t) is a 2 × 2 special orthogonal matrix and that it can be parameterized by the single angle
co-ordinate θ . Furthermore now
RT Ṙ = −(RT Ṙ)T = Ω̂
and in this case it turns out that

0 −1
Ω̂ (t) = Ω (t) ,
1 0
where Ω (t) = θ̇ (t). In this case an explicit calculation shows that Ω̂ 2 = −Ω 2 I2×2 .
Let f be the representation of the force acting on the particle in the e frame. Then Newton’s equation (1.14)
expressed in the inertial frame is
F b = mẌ = −m Ω̂ 2 (t)X(t) − 2m Ω̂ (t)V (t) − m Ω̂˙ (t)X(t) + RT (t) f
and specializing to the 2D case

Ẍ1 2 −X1 −V2 −X2 F
b
F =m = −mΩ − 2mΩ − mΩ̇ + 1
Ẍ2 −X2 V1 X1 F2
where
F1
RT f I = =F
F2
is the force experienced in the e frame expressed in the rotating frame b.
The first three terms from left to right are the Centrifugal, Coriolis, and Euler forces respectively. If the
rotating frame is rotating at a constant rate then Ω̇ = 0 and hence

Ẍ1 2 −X1 (t) −V2 (t) F1 (t)
b
F =m = −mΩ − 2mΩ + .
Ẍ2 −X2 (t) V1 (t) F2 (t)
In the [er eθ ] frame the representation of P is

r
X=
0
and hence
ṙ r̈
Ẋ = Ẍ = .
0 0
From exercise 1.14 Newton’s equations in 2D are

r̈ −r 0 0 F
m = −mθ̇ 2 − 2mθ̇ − mθ̈ + r .
0 0 ṙ r Fθ

mr̈ mrθ̇ 2 0 0 F
= + + + r .
0 0 −2mθ̇ ṙ −mrθ̈ Fθ
54 1 Mechanics
Observe that the Centrifugal force is mrθ̇ 2 and is in the er direction, the Coriolis force is 2mθ̇ ṙ in the −eθ
direction and the Euler force is mrθ̈ in the −eθ direction.
Simplifying the above equations we have that
mr̈ = mrθ̇ 2 + Fr ,
mrθ̈ = −2mṙθ̇ + Fθ .
Observe that if the particle is constrained to move in a circle then r = constant and hence
Fr = −mrθ̇ 2 , mrθ̈ = Fθ .
When the rotating frame is rotating at a constant angular rate of Ω̇ = θ̈ = 0 from answers to exercise 1.14
Newton’s equation in the rotating 2D-frame b(t) for a point particle of mass m reduce to

Ẍ −X1 (t) −V2 (t) F (t)
m 1 = −mΩ 2 − 2mΩ + 1 .
Ẍ2 −X2 (t) V1 (t) F2 (t)
These equations describe the motion of the particle in the b(t) frame.
Part 1.21.3
Since friction is neglected f I = RT F = 0 and hence F = [F1 F2 ]T = 0. The motion of the particle in the b(t)
frame is given by,
Ẍ(t) = − Ω̂ 2 X(t) − 2 Ω̂ V (t).
In the case of 2D motion
Ẍ1 X −Ẋ2
= Ω 2 1 − 2Ω .
Ẍ2 X2 Ẋ1
The initial conditions that correspond to releasing a mass at a unit distance away along the b1 (t) axis is X(0) =
[1 0]T and Ẋ(0) = [0 0]T . The MATLAB code to simulate the solutions for the case where θ̇ = Ω = 3.
---------------------------------------------------------------------
T=0:0.1:5;
X0=[1 0 0 0];
[T,Y]=ode23s(’MotnRotatnFrame’,T,X0);
n=length(T);
for i=1:n
figure(1)
plot(Y(i,1),Y(i,3),’*’)
axis([-15 15 -15 15])
hold on
pause(0.1)
end
---------------------------------------------------------------------
function xdot=MotnRotatnFrame(t,x)
X1=x(1);
V1=x(2);
X2=x(3);
V2=x(4);
omega=3;
m=1;
F1=0;
F2=0;
X1dot=V1;
V1dot=omegaˆ2*X1+2*omega*V2+F1/m;
X2dot=V2;
V2dot=omegaˆ2*X2-2*omega*V1+F2/m;
Xdot=[X1dot;V1dot;X2dot;V2dot];
---------------------------------------------------------------------
Fig. 1.36
Simulation results of the path of the particle is shown in figure 1.36.
Part 1.21.4
If the particle is constrained to move along the b1 (t) axis. Then

y(t)
X(t) = ,
0
where y(t) is the distance to the particle from the origin along the b1 (t) axis. The force due to the constraints
expressed in the rotating frame is
T I 0
R f (t) = F(t) = .
F2
Thus from Newton’s equations
56 1 Mechanics

ÿ(t) y(t) 0 0
= Ω2 − 2Ω + .
0 0 ẏ(t) F2
From this we have that
ÿ = Ω 2 y,
F2 = 2Ω ẏ.
Part 1.21.6
If an inward radial force field of intensity 1/r is present then, expressed in the rotating frame, the force due
to the field is
1 1 X1 (t)
RT f I (t) = F(t) = − X(t) = − .
||X(t)||2 ||X(t)||2 X2 (t)
Applying Newton’s equations in the rotating frame we obtain

Ẍ1 X −Ẋ2 1 X1
= Ω 2 1 − 2Ω − .
Ẍ2 X2 Ẋ1 m||X(t)||2 X2
The initial conditions that correspond to releasing a mass at a distance 0.5 away along the b1 (t) axis with zero
velocity is X(0) = [0.5 0]T and Ẋ(0) = [0 0]T . The MATLAB code to simulate the solutions for the case where
θ̇ = Ω = 1.
---------------------------------------------------------------------
T=0:0.01:.8;
X0=[0.5 0 0 0];
[T,Y]=ode23s(’MotnRotatnFrame’,T,X0);
n=length(T);
for i=1:n
figure(1)
plot(Y(i,1),Y(i,3),’*’)
axis([-.1 .6 -.3 .3])
hold on
pause(0.1)
end
---------------------------------------------------------------------
function xdot=MotnRotatnFrame(t,x)
X1=x(1);
V1=x(2);
X2=x(3);
V2=x(4);
omega=1;
m=1;
r2=(X1ˆ2+X2ˆ2);
F1=-X1/r2;
F2=-X2/r2;
X1dot=V1;
V1dot=omegaˆ2*X1+2*omega*V2+F1/m;
X2dot=V2;
V2dot=omegaˆ2*X2-2*omega*V1+F2/m;
Xdot=[X1dot;V1dot;X2dot;V2dot];
---------------------------------------------------------------------
Fig. 1.37
Simulation results of the path of the particle is shown in figure 1.37.
Fix a frame b(t) on the shaft and the arm such that its origin, O, is at the point of intersection of the arm and
the shaft with b1 (t) aligned along the arm, b3 (t) aligned along the shaft and pointing upwards and b2 (t) aligned
so that the frame is right hand oriented. Let e be a fixed frame such that the origin coincides with O and e3 is
aligned along b3 . Let φ be the counter clockwise angle between e1 and b1 (t), θ be the counter clockwise angle
between the spring and the vertical and l(t) be the length of the spring. The variables (l(t), θ (t), φ (t)) are what
are referred to as the configuration variables or co-ordinates of the system.
Then b(t) = e R(t) where  
cos φ − sin φ 0
R(t) =  sin φ cos φ 0 
0 0 1
and Ṙ = R Ω̂ where  
0 −φ̇ 0
Ω̂ =  φ̇ 0 0 
0 0 0
Consider the system to be a coupled system. Where a point mass is coupled to a rigid body, the arm and the
shaft. Lets consider the two systems separately. First let us consider the point mass. The representation of the
point mass P in e is xP (t) and in b(t) is XP (t). Then xP (t) = R(t) XP (t) and it easily seen that
58 1 Mechanics
 
d
XP (t) =  l sin θ  .
−l cos θ
Differentiating we have  
0
ẊP (t) =  l˙sin θ + l θ̇ cos θ 
−l˙cos θ + l θ̇ sin θ
and  
0
ẌP (t) =  l¨sin θ + 2l˙θ̇ cos θ − l θ̇ 2 sin θ + l θ̈ cos θ  .
−l¨cos θ + 2l˙θ̇ sin θ + l θ̇ 2 cos θ + l θ̈ sin θ
Recall that Newton’s equations for the particle are
F e = RT f e = m ẍP = m(Ω̂ 2 XP + 2Ω̂ ẊP + Ω̂˙ XP + ẌP ), (1.111)
where the matrix F e is the external force acting on the particle represented in the b(t) frame. Considering the
forces on the particle  
N
F e =  −k(l − l0 ) sin θ  .
k(l − l0 ) cos θ − mg
where N is the constraint force on the particle in the b1 direction and constraints the particle to only move in the
plane perpendicular to the shaft. Now simplifying the Newton’s equations (1.111) we have
k
l¨ + (l − l0 ) + (φ̇ 2 sin2 θ − θ̇ 2 )l + φ̈ sin θ d = g cos θ , (1.112)
m
l θ̈ + 2l˙θ̇ − φ̇ 2 sin θ cos θ l + φ̈ cos θ d = −g sin θ . (1.113)
Now lets consider the motion of the rigid body. The body rotates about the e3 axis and hence can be considered
as a 2D rigid body motion in the e1 , e2 plane. Then euler’s 2D rigid body equations are
I φ̈ = T e , (1.114)
where Te is the force moment about the e3 axis. The force acting at the tip of the arm in the e1 and e2 plane is

e
N
= [b1 b2 ]Fplane = b1 b2
k(l − l0 ) sin θ
The representation of the end point, A, of the arm is

d
OA = [b1 b2 ]XA = b1 b2 ,
0
Then the force moment about the e3 axis is

N
T e = X̃A Fplane
e
= 0d = kd(l − l0 ) sin θ .
k(l − l0 ) sin θ
Thus the rigid body equations (1.115) are
I φ̈ = kd(l − l0 ) sin θ . (1.115)
Thus the complete governing equations for the system are (1.112),(1.113) and (1.115). These are three coupled
second order equations of the three configuration variables.
The kinetic energy of the system is
KE = KErigid body + KE particle .
where the kinetic energy of the rigid body is given by

I
KErigid body = φ̇ 2 ,
2
and the kinetic energy of the particle is given by
m m m
KE particle = ||ẋ||2 = ||R(Ω̂ XP + ẊP )||2 = ||Ω̂ XP + ẊP ||2 .
2 2 2
Upon simplification we have that
m
KE particle = −XPT Ω̂ 2 XP + 2ẊP Ω̂ XP + ẊPT ẊP
2
(I + md 2 + ml 2 sin2 θ ) 2 m ˙2 ml 2 2
= φ̇ + l + θ̇ + 2d sin θ φ̇ l˙ + 2d cos θ l φ̇ θ̇
2 2 2
This is a quadratic form in the velocities.
Fig. 1.38 A schematic representation of a free vibration apparatus.

7
Consider figure 1.38. Assume that when z ≡ 0 and θ ≡ 0 the system is in equilibrium. Let the deflection of
the spring at this condition be Δ . By applying Euler’s 3D rigid body equations,
Ṙ = RΩ̂ (1.116)
60 1 Mechanics
I Ω̇ = I Ω × Ω + T. (1.117)
Here  
cos θ − sin θ 0
R =  sin θ cos θ 0  ,
0 0 1
and hence    
0 −θ̇ 0 0
Ω̂ =  θ̇ 0 0  , Ω =  0  .
0 0 0 θ̇
The inertia tensor of the beam is  
I1 0 0
I =  0 I2 0  ,
0 0 I3
where I1 , I2 , I3 are the inertia components of the body about the b1 , b2 , b3 axis respectively. Thus it is easily seen
that I Ω × Ω = 0 and hence (1.117) reduces to:
I Ω̇ = T, (1.118)
The external TorqueT = Tc + Tk + Tg , where Tc and Tk are the force moments (torques) expressed in the b frame
due to the damper force and spring force respectively and Tg is the force moment due to the gravitational forces.
For small angles sin θ ≈ θ ≈ 0 and cos θ ≈ 1. For small deflections the above small angle approximations
are valid and in the b frame the spring force has the representation,
 
0
Fk =  −K(Lk θ + Δ + z)  (1.119)
0
The point at which this force acts has the representation in the b frame given by
 
Lk
Xk =  0 
0
Thus the torque in the b frame due to the spring force is

 
0
Tk = Xk × Fk =  0  (1.120)
−K(Lk θ + Δ + z)Lk
In the b frame the damper force has the representation,

 
0
Fc =  −CLc θ̇  (1.121)
0
The representation in the b frame of the point of action of the damper force is
 
Lc
Xc =  0  .
0
The the force moment due to the damper force is

 
0
Tc = Xc × Fc =  0  . (1.122)
−CLc2 θ̇
In the b frame the gravitational force force has the representation,

 
0
Fg =  Mg  (1.123)
0
where M is the total mass of the beam and L is the distance from O to the center of mass of the beam. The
representation in the b frame of the point of action of the damper force is
 
L
Xg =  0  .
0
The the force moment due to the damper force is

 
0
Tg = Xg × Fg =  0  . (1.124)
MgL
Also  
0
I Ω̇ =  0  .
I3 θ̈
Thus Euler’s rigid body equations (1.117) gives,
I3 θ̈ = −CLc2 θ̇ − KLk2 θ + (MgL − KLk Δ ) − KLk z. (1.125)

Applying rigid body equations at equilibrium conditions we have (MgL − KLk Δ ) = 0 and hence for small de-
flections we have the governing differential equation given below.
I3 θ̈ +CLc2 θ̇ + KLk2 θ = −KLk z. (1.126)

Chapter 2
Dynamics
2.1 Introduction
In mechanics we have seen how the application of physical principles such as Newton’s equations or Euler’s rigid
body equations are used to mathematically describe certain physical systems. This process is generally referred
to as mathematical modelling. Typically the mathematical object so derived is a set of differential equations. The
sense in which these differential equations describe the physical system is that the solutions of the differential
equations correspond to certain behaviour of the physical system1 . Thus the study of the behaviour of the solu-
tions of a system of ODEs is of paramount importance in understanding a behaviour of the physical system. The
study of the behaviour of differential equations is referred to as Dynamics.
We will first introduce some terminology involved in dealing with dynamics. Mathematically a dynamic
system consists of the following ingredients: A space X called the state space, and a first order differential
equation defined on X by
d x(t)
= ẋ(t) = f (x(t)). (2.1)
dt
Here x is called the state of the system and f (x) is called a vector field. The fundamental theorem of differ-
ential equations tells us that there always exists a solution2 of (2.1) for any given initial condition x(0) and if
f (x) satisfies certain continuity conditions, this solution is unique. Most of the physical systems give rise to
dynamic systems that satisfies the said continuity condition and hence desirably the solutions are always unique.
Furthermore the solutions depend continuously on the initial data x(0). What that means is that the solutions
corresponding to two “nearby” initial conditions say “close” to each other for a “while”.
A solution x(t) of (2.1) can be visualized as a curve on X that is parameterized by t. In this context we
refer to x(t) as a trajectory of the system that begins at x(0) at t = 0. Observe that the tangent vector at any
point x(t) on the trajectory coincides with the value of the vector field at x(t). Having this picture in mind
is extremely useful in dynamics.
In most of the examples considered in this class the state space X = Rn for some n or is an open subset of
Rn .In this case the state x is an ordered n-tuple of variables (x1 , x2 , ∙ ∙ ∙ , xn ) and f (x) is a set of ordered n-tuple
of functions ( f1 (x), f2 (x), ∙ ∙ ∙ , fn (x)) where each fi is a real valued function defined on Rn (that is fi : Rn 7→ R).
Writing the ordered n-tuples in a column matrix we can then write the dynamic system (2.1) as
   
x1 (t) f1 (x1 (t), x2 (t), ∙ ∙ ∙ , xn (t))
d   
 x2 (t)   f2 (x1 (t), x2 (t), ∙ ∙ ∙ , xn (t)) 

 ..  =  .. , (2.2)
dt  .   . 
xn (t) fn (x1 (t), x2 (t), ∙ ∙ ∙ , xn (t))
1 Remember that this correspondence is approximate.

2 Note that this solution may not be defined for all time t
63
64 2 Dynamics
Thus a dynamic system on Rn is nothing but a collection of n number of coupled first order ODEs. To solve this
system we know that n initial conditions x(0) = (x1 (0), x2 (0), ∙ ∙ ∙ , xn (0)) need to be specified.
Let us now look at several types of behaviour of a dynamical system. These behaviour correspond to special
types of solutions, or trajectories, of (2.1). The most common behaviour that one sees in Engineering systems is
steady state (or equilibrium) behaviour, periodic behaviour and chaotic behaviour. They correspond respectively
to steady state, periodic and chaotic solutions of the corresponding dynamic system (2.1).
Steady State Solutions
A solution x(t) is said to be an steady state or equilibrium solution of the dynamic system (2.1) if it remains
a constant for all time. That is x(t) ≡ x.
ˉ Substituting this equilibrium solution in (2.1) it can be seen that such
equilibria are solutions of f (x)
ˉ = 0.
Periodic Solutions
A solution x(t) is said to be a periodic solution of the dynamic system (2.1) if x(t + T ) = x(t) for all t for which
the solution is defined and some T > 0 called the period of the solution3 .
Chaotic Solutions
Chaotic solutions are characterized by their extreme sensitivity to initial conditions. Although the solutions
depend continuously on the initial conditions they move apart at an exponential rate if they are chaotic.
2.2 Linear Systems
When the vector field f (x) is linear, that is when f (x) = Ax, where A is a constant n × n matrix, we say that the
dynamic system is a linear system:
ẋ = Ax. (2.3)
Linearity Property of Solutions: It can be easily verified that if x1 (t) is a solution with initial condition
x1 (0) and x2 (t) is a solution with initial condition x2 (0) then x(t) = x1 (t) + x2 (t) is a solution with initial
condition x1 (0) + x2 (0) and x(t) = λ x1 (t) is a solution with initial condition λ x1 (0) for any λ ∈ R. This in
fact can be used as the characterizing property of linear systems. Systems that do not satisfy this property
are called nonlinear systems.
Linear systems have the simplest possible behaviour. We shall see that the eigenvalues of A completely char-
acterize the entire behaviour of the system. For linear systems (2.3) the origin xˉ = 0 is always an equilibrium
solution and no other equilibria exist if A has no zero eigenvalue4 . The equilibrium xˉ = 0 is said to be Hyperbolic
if none of the the eigenvalues of A are on the imaginary axis while it is called Non-Hyperbolic otherwise.
By direct verification we can see that a general solution of (2.3) is given by
x(t) = eAt x(0). (2.4)
Here eX is the matrix exponential function given by the infinite series
3 The smallest such value is chosen.

4 Reason out why this is true.
2.2 Linear Systems 65
1 2 1 3
eX = I + X + X + X +∙∙∙ ,
2! 3!
where X is a n × n matrix.
We will see how one may calculate this solution for a second order system. For a second order system x ∈ R2
and A is a 2 × 2 constatnt matrix. Let the eigenvalues of A be λ1 and λ2 . From linear algebra we know that A is
similar to one of the following normal forms:

λ1 0 λ 0 λ 1 σ −ω
A1 = , A2 = , A3 = , A4 = .
0 λ2 0λ 0λ ω σ
That is there exists a 2 × 2 invertible matrix Ti such that A = Ti Ai Ti−1 for some Ai given above. The similarity
between A and Ai is denoted symbolically as A ∼ Ai . The particular normal form depends on the eigenvalues of
A and their multiplicities. When λ1 and λ2 are real and distinct A ∼ A1 , when λ1 = λ2 = λ and their geometric
multiplicity is two then A ∼ A2 , when λ1 = λ2 = λ and their geometric multiplicity is one then A ∼ A3 , and
finally if λ1 = σ + iω and λ2 = σ − iω then A ∼ A4 . It is easy to show that
λt λt λt λt σt
e 1 0 e 0 e te e cos ω t eσ t sin ω t
e A1 t = , e A2 t
= , e A3 t
= , e A4 t
= .
0 eλ2 t 0 eλ t 0 eλ t eσ t sin ω t eσ t cos ω t
You are asked to show this in exercise 2.2. Let A ∼ Ai then since
−1
eTi Ai Ti = Ti eAi Ti−1 ,
we have that −1
x(t) = eAt x(0) = etTi Ai Ti x(0) = Ti etAi Ti−1 x(0)
Thus we can show that if λ1 and λ2 are real and distinct the solution takes the form

x1 (t) c eλ1 t + c12 eλ2 t
= 11 λ1 t
x2 (t) c21 e + c22 eλ2 t
where all the ci j are constants that depend on the initial condition x(0). Note that not all four of these constants
are independent and only two of them are. If λ1 = λ2 = λ we can show that a solution either takes the form

x1 (t) c eλ t
= 11 λ t
x2 (t) c12 e
or
x1 (t) (c11 + c12 t)eλ t
=
x2 (t) (c21 + c22 t)eλ t
are independent and only two of them are. Finally if λ1 = σ + iω and λ2 = σ − iω we can show that a solution
takes the form σt
x1 (t) e (c11 sin ω t + c12 cos ω t)
= σt
are independent and only two of them are.
From this we see that if all the eigenvalues of A are in the strict left half of the complex plane then
all solutions asymptotically converge to the equilibrium solution xˉ = 0. That is limt→∞ x(t) → 0 for all
initial conditions x(0). In this case we say that the equilibrium solution xˉ = 0 is globally asymptotically
stable. On the other hand if both eigenvalues are in the strict right half plane then all solutions, except the
66 2 Dynamics
equilibrium solution, diverges to infinity. That is limt→∞ x(t) → ∞ for all initial conditions x(0) 6= 0. If only
one eigenvalue is in the strict right half plane then almost all solutions diverge to infinity. If the eigenvalues
are on the imaginary axis and are distinct then all solutions are periodic.
A linear system with forcing is given by

ẋ = Ax + u(t). (2.5)
The general solution of this can be written down explicitly. You asked in exercise 2.1 to show that the general
solution is given by the Variation of Parameters formula:
Z t
x(t) = eAt x(0) + eA(t−τ ) u(τ ) d τ . (2.6)
0
2.2.1 RC Circuit
Fig. 2.1 RLC Circuit
If the inductance L = 0 then the circuit shown in figure 2.1 reduces to a RC circuit. Applying Kirchoff’s
voltage law we obtain
1 1
q̇ = − q + v(t), (2.7)
RC R
where q is the charge on the capacitor. Consider the case where the external voltage v(t) ≡ 0. Then the equation
(2.7) becomes
1
q̇ = − q = −a q, (2.8)
RC
where we have set a = RC 1
. This is a first order linear system where the state space is X = R. The equilibrium
solution of the system is qˉ = 0. The equilibrium is a hyperbolic equilibrium as long as a 6= 0 and is non-hyperbolic
when a = 0. In our elementary ODE class we have seen that the general solution of the system is
q(t) = e−at q(0).
It can be seen that, when a > 0 limt→∞ q(t) = 0 and that when a < 0 limt→∞ q(t) = 0. This also shows that the
equilibrium is globally asymptotically stable if a > 0 and is unstable when a < 0.
When the forcing v(t) is non zero we have
1
q̇ = − q = −a q + u(t), (2.9)
RC
where u(t) = R1 v(t). The solution of this system can also be written down using the variation of parameters
formula
Z t
q(t) = e−at q(0) + e−a(t−τ ) u(τ )d τ .
0
2.2.2 Spring Mass Damper System
Fig. 2.2 Spring Mass Damper System.
Assuming the viscous damping force is fv = −Cẋ Newton’s equations yield the following ODE model of the
system.
M ẍ(t) +Cẋ(t) + Kx(t) = f (t). (2.10)
For the purpose of this section we will consider the case where there is not external force; f (t) ≡ 0. Dividing by
M and setting ωn2 = MK
and 2ζ ωn = M
C
in (3.19) we have
ẍ(t) + 2ζ ωn ẋ(t) + ωn2 x(t) = 0. (2.11)
For small displacements of the mass, the solutions of this equation describes the motion quite accurately. Let
us try to see what conclusions we can draw about the behaviour of the system by studying the behaviour of the
solutions of the differential equation (3.20). By using the variable transformation x1 = x, x2 = ẋ we can write
(3.20) as a set of two first order ODE’s. We write them in concise form using matrices as

ẋ1 0 1 x1
= . (2.12)
ẋ2 −ωn2 −2ζ ωn x2
We can also write this as ẏ = Ay where

x1 0 1
y= , A= .
x2 −ωn2 −2ζ ωn
This is a second order linear system. Recall that the eigenvalues of A determine the complete behaviour of the
system. The eigenvalues of A are the roots of the characteristic polynomial
λ 2 + 2ζ ωn λ + ωn2 = 0.
The roots are

λ1 , λ2 = −ζ ωn ± iωd ,
p
where ωd = ωn 1 − ζ 2 . Thus depending on ζ the eigenvalues will be real or complex.
68 2 Dynamics
The origin yˉ = [0 0]T is an equilibrium solution. It is hyperbolic when none of the eigenvalues are on the
imaginary axis, that is ζ 6= 0, and is non-hyperbolic when at least one of them is on the imaginary axis, that is
ζ = 0.
Observe that the parameters ωn , ωd , and ζ , completely determine the eigenvalues. Thus we have special names
for these parameters.
ζ - Damping ratio.
ωn - Natural frequency.
ωd - Damped natural frequency.
The meaning will be clear after we look at the solutions a little more closely. Depending on the damping ratio
ζ the eigenvalues will be real and distinct, real and coinciding, complex and conjugate or purely imaginary. We
will consider each of these cases separately.
(a) ζ = 0 (b) ζ = 0.1
(c) ζ = 1 (d) ζ = 2
Fig. 2.3 The displacement x1 (t) Vs t for different values of ζ .
Un-Damped: ζ = 0
When ζ = 0 the eigenvalues are

λ1 , λ2 = ±iωn ,
and are purely imaginary and conjugate. Then form the results of section 2.2 the solution takes the form

x1 (t) c sin ωnt + c12 cos ωnt
= 11 .
x2 (t) c21 sin ωnt + c22 cos ωnt
where all the ci j are constants that depend on the initial condition x(0). Note that c22 = ωn c11 and c21 = −ωn c12 .
A typical such solution is shown in figure 2.3 (a). We see that the solution is periodic with period 2π /ωn for all
initial conditions. The frequency of oscillation is ωn . ωn is called the undamped natural frequency of the system.
Under Damped: 0 < ζ < 1
When 0 < ζ < 1 the eigenvalues are

λ1 , λ2 = −ζ ωn ± iωd ,
and are complex and conjugate. Figure 2.4(a) shows the location of two typical such eigenvalues. Then form the
results of section 2.2 the solution takes the form
−ζ ω t
x1 (t) 11 sin ωd t + c12 cos ωd t)
e n (c
= −ζ ωn t .
x2 (t) e (c21 sin ωd t + c22 cos ωd t)
where all the ci j are constants that depend on the initial condition x(0). Note that c21 = −ζ ωn c11 − ωd c12 and
c22 = −ζ ωn c12 + ωd c11 . A typical solution for x1 (t) is shown in figure 2.3(b). We see that the solution converges
to [0 0]T . However the convergence is oscillatory. The frequency of oscillation is ωd and the rate of decay of
the amplitude is e−ζ ωn t . The frequency ωd is called the damped natural frequency. Figure 2.4(b) shows a typical
underdamped response.
(a) Eigenvalue location. (b) A typical solution.
Fig. 2.4 Figure (a) shows the location of the eigenvalues λ1 and λ2 when 0 < ζ < 1. Figure (b) shows a typical underdamped
response of x1 (t).
Critically Damped: ζ = 1
When ζ = 1 the eigenvalues are

λ1 , λ2 = −ωn , −ωn ,
and are real coincident and negative. Then form the results of section 2.2 the solution takes the form
−ω t
x1 (t) e n (c + c12 t)
= −ωn t 11
x2 (t) e (c21 + c22 t)
where all the ci j are constants that depend on the initial condition x(0). Note that c21 = −ωn c11 + c12 and
c22 = −ωn c12 . The solution can be shown to converge to the origin [0 0]T monotonically. A typical solution for
x1 (t) is shown in figure 2.3(d).
70 2 Dynamics
Over Damped: ζ > 1
When ζ > 1 the eigenvalues are p

λ1 , λ2 = ωn (−ζ ± 1 − ζ 2 ),
and are distinct and negative. Then form the results of section 2.2 the solution takes the form

x1 (t) c11 eλ1 t + c12 eλ2 t
=
x2 (t) λ1 c11 eλ1 t + λ2 c12 eλ2 t
where all the ci j are constants that depend on the initial condition x(0). Since both λ1 and λ2 are real and negative
we see that for any initial condition x(0) the solution converges to [0 0]T monotonically as well. However the
convergence is slower than for the critically damped case. A typical solution for x1 (t) is shown in figure 2.3(d).
For comparison purposes the response x1 (t) for various different values of ζ are plotted in the same figure in
2.5. Observe that for 0 ≤ ζ ≤ 1 oscillations decay out faster when the damping ratio increases (see figure 2.5(a)).
Also observe that when ζ ≥ 1 the solution is monotonic and that the convergence becomes slower as ζ increases
(see figure 2.5(b)). Figure 2.6 shows the phase plot of the solution that begins at [π /36 0]T for several different
values of ζ .
(a) 0 ≤ ζ ≤ 1 (b) 1 ≤ ζ
Fig. 2.5 The displacement x1 (t) Vs t for different values of ζ .
2.3 Nonlinear Systems
For general nonlinear systems the situation is more complicated. First of all linear super position does not work.
That is, in general the linearity property described in the gray box at the beginning of section 2.2 does not hold.
Second it is very rarely that one can explicitly write down the solution. Nonlinear systems for which explicit
solutions can be written down are called completely integrable systems. The simple pendulum system and the
rigid body system to be discussed subsequently are two examples of such systems. On the other extreme we
have chaotic systems. The double pendulum and the Henon weather model are two examples that exhibit such
behaviour.
In Engineering the local behavior of a system near an equilibrium is of great importance. An equilibrium
xˉ is said to be stable if solutions that begin “sufficiently near” the equilibrium stay “sufficiently near” the
equilibrium for all future time. If not the equilibrium is said to be unstable. If all solutions that begin
“sufficiently near” the equilibrium “asymptotically” converge to the equilibrium (ie. as t → ∞) then the
equilibrium is said to be locally asymptotically stable.
2.3 Nonlinear Systems 71
Fig. 2.6 Phase plot, ẋ1 (t) Vs x1 (t), for different values of ζ .
In order to investigate the local behavior of solutions near an equilibrium xˉ we can consider the perturbation
x = xˉ + δ x where δ x is a very small quantity. Substituting this in (2.1) and using the Taylor series expansion of
f (x) we have
∂ f
δ x = f (x)
˙ ˉ + δ x + H.O.T. (2.13)
∂x xˉ
For small motions near the equilibrium the higher order terms H.O.T are negligible compared to the δ x term
and recalling that f (x)
ˉ = 0 we have the perturbed dynamic system
δ˙x = A(x)
ˉ δ x, (2.14)

∂f
ˉ =
where A(x) ∂ x xˉ . This is referred to as the linearization of (2.1) about the equilibrium x.
ˉ
The local behavior of the solutions of (2.17) about the equilibrium xˉ is equivalent to the behavior of the
solutions of the linearized system (2.14) provided that none of the eigenvalues of A(x) ˉ lie on the imaginary
axis (that is when the equilibrium xˉ is hyperbolic). The behavior of solutions of a linear system (2.14) are
completely characterized by the eigenvalues of A(x). ˉ Thus the local behavior of the solutions of (2.1) near
hyperbolic equilibria xˉ are completely determined by the eigenvalues of A(x). ˉ If all the eigenvalues of A(x)
ˉ
are in the strict left half complex plane then the equilibrium xˉ is locally asymptotically stable and if at least
one eigenvalue of A(x) ˉ is in the strict right half complex plane then the equilibrium xˉ is unstable. When the
equilibrium xˉ is non-hyperbolic the local stability of xˉ can not be determined by the stability properties of
the linearized system.
The above procedure works when one needs to consider the stability of more general solutions than equilib-
rium solutions. However then the linearized system will in general be a time varying system and the eigenvalue
analysis does not hold.
72 2 Dynamics
2.3.1 Simple Pendulum
Fig. 2.7 A simple pendulum of length L and total mass M.
Consider the simple pendulum shown in figure 2.7. Assume that a viscous frictional torque, due to air re-
sistance and bearing resistance, acts on the pendulum. Let the torque constant be C N/ms−1 . Using rigid body
Euler’s equations the governing equation of the system is given by,
3C 3g
θ̈ + 2
θ̇ + sin θ = 0. (2.15)
ML 2L
Setting α1 = θ and α2 = α̇1 = θ̇ the state space form of (2.15) is

d α1 α2
= . (2.16)
dt α2 2 α2 − 2L sin α1
3C 3g
− ML
This is also known as the dynamic system model of the system. Using the notation

α1 α2
α= , f (α ) =
α2 3C
− ML 2 α2 − 2L sin α1
3g
(2.16) can be written succinctly as

α̇ = f (α ). (2.17)
The right hand side of (2.17) is referred to as the vectorfield of the dynamic system. A solution α (t) is said to
be an equilibrium solution of the dynamic system (2.17) if it remains a constant for all time. That is α (t) ≡ αˉ .
Substituting this equilibrium solution in (2.17) it can be seen that such equilibria are solutions of f (αˉ ) = 0. For
the simple pendulum the equilibria are given by αˉ = [0 0]T and αˉ = [π 0]T .
In order to investigate the local behavior of solutions near an equilibrium αˉ we can consider the linearized
system
δ˙α = A(αˉ )δ α , (2.18)
where
∂ f 0 1
A(αˉ ) = = .
∂ α α =αˉ − 2L cos αˉ − ML
3g 3C
2
For the simple pendulum it can be shown that the eigenvalues of A(αˉ ) are in the strict left half complex plane
for the equilibrium αˉ = [0 0]T and that at least one eigenvalue is in the strict right half complex plane when
the equilibrium αˉ = [π 0]T . Thus the vertically downward equilibrium αˉ = [0 0]T is ‘locally’ stable while the
vertically upward equilibrium αˉ = [π 0]T is unstable.
2.3.1.1 Analytical Solution of the Undamped Simple Pendulum
If the frictional torque is negligible then C ' 0 and the governing equation reduces to
θ̈ + ω 2 sin θ = 0, (2.19)
where ω 2 = 3g
2L The vectorfield of the dynamic system model is

α2
f (α ) = .
−ω 2 sin α1
The total energy of the system is

1 θ
T (θ , θ̇ ) = θ̇ 2 + 2ω 2 sin2 . (2.20)
2 2
Differentiating T along the trajectories of the system (2.19) we have that
d
T = θ̇ (θ̈ + ω 2 sin θ ) = 0,
dt
and hence that T (θ , θ̇ ) = 12 θ̇ 2 + 2ω 2 sin2 θ2 = E is a constant along the solutions of (2.19). This fact allows us to
find the solutions of (2.19) analytically since we can write
θ
θ̇ 2 = 2E − 4ω 2 sin2 (2.21)
r 2
d θ
θ = 2E − 4ω 2 sin2 (2.22)
dt 2
The last of the above two expressions can be integrated using quadrature as follows:
Z Z
1
q dθ = dt.
2E − 4ω 2 sin2 θ2
Large Total Energy: 2ω 2 ≤ E
A change of variables u = sin θ2 reduces the above integral to

Z Z
1
p du = Ω dt, (2.23)
(1 − u2 )(1 − ku2 )
where k2 = 2ω 2 /E and Ω 2 = E/2. For total Energy 2ω 2 ≤ E it can be seen that 0 ≤ k ≤ 1. Then the left hand
side is known as a Jacobi Elliptic integral of a first kind. Specifically a sn function of modulus k.
Z
1
sn−1 (u) = p du. (2.24)
(1 − u2 )(1 − ku2 )
Thus (2.23) becomes

sn−1 (u) = Ω t + c,
and hence
u(t) = sn(Ω t + c).
Thus the complete analytic solution of the simple pendulum equation is
θ (t) = 2 arcsin u(t) = 2 arcsin (sn(Ω t + c)). (2.25)
From (2.22) we also have that

θ̇ (t) = 2Ω dn(Ω t + c). (2.26)
Two interesting limiting behaviors of the pendulum result when E is very large, (that is when k → 0) and when
74 2 Dynamics
E = 2ω 2 (that is when k → 1).
In the limit k → 0 the integral (2.24) reduces to

Z
−1 1
lim sn (u) = √ du = arcsin u
k→0 1 − u2
and sn to the trigonometric sin function. It can be shown form properties of Elliptic functions that cn reduces to
cos and dn = 1.
On the other hand in the limit k → 1 the integral (2.24) reduces to
Z
1 1 1+u
lim sn−1 (u) = 2
du = ln = arctanh(u).
k→1 1−u 2 1−u
It can be shown that in this case cn reduces to sech and dn = sech.
Thus for large total energy E (>> 2ω 2 ) the motion is given by
θ (t) = 2(Ω t + c), (2.27)

θ̇ (t) = 2Ω . (2.28)
Similarly when E = 2ω 2 (ie. k = 1) the motion is given by

θ (t) = 2 arcsin (tanh(ω t + c)) = π − 4 arctan e−(ω t+c) , (2.29)
θ̇ (t) = 2ω sech(ω t + c). (2.30)
Small Total Energy: E ≤ 2ω 2
For total energy such that E ≤ 2ω 2 from equation (2.22) we have

r
d E θ
θ = 2ω − sin2 .
dt 2ω 2 2
Now setting k2 = E
2ω 2
and sin θ /2 = k sin ψ we have
q
d
ψ =ω 1 − k2 sin2 ψ ,
dt
where 0 ≤ k ≤ 1. Now setting u = sin ψ we have that for small energy E ≤ 2ω 2 the solutions of the pendulum is
given by the Elliptic integral of the first kind,
Z Z
1
p du = ω dt. (2.31)
(1 − u )(1 − ku2 )
2
Thus as before we have
θ (t) = 2 arcsin ku(t) = 2 arcsin (k sn(ω t + c)), (2.32)

√
θ̇ (t) = 2E cn(ω t + c). (2.33)
In summary we have shown that all the solutions of the simple pendulum can be completely written down using
Jacobi Elliptic functions. For Energy E ≥ 2ω 2 the solutions are given by (2.25) and (2.26), while for energy
E ≤ 2ω 2 the solutions are given by (2.32) and (2.33). Jacobi Elliptic sn and cn functions are periodic functions
of period 4K where
Z 1 Z π /2
1 1
K(k) = sn−1 (1) = p du = q dψ .
0 (1 − u2 )(1 − k2 u2 ) 0
1 − k2 sin2 ψ
Thus the motion of the pendulum is periodic of periodicity T (k) = 4K/Ω for E ≥ 2ω 2 and T (k) = 4K/ω for
E ≤ 2ω 2 . Notice that it is a function of k and that k is a function of the energy. Thus the period of periodicity is
a function of energy and hence the initial conditions.
For k <<< 1 using the binomial expansion the right hand side of the last integral can be approximated to
give,
Z π /2 Z π /2
1 k2 k2 π
K= q dψ = 1 + (1 − cos 2ψ ) + H.O.T d ψ ≈ 1 + .
4 4 2
0
1 − k2 sin2 ψ 0
Thus the period of the solution is

2π k 2 π 2π E
T≈ + = 1+ .
ω 2ω ω 8ω 2
For small Energy k ≈ 0 and θ is very small and sin θ ≈ θ . For k ≈ 0 the Jacobi Elliptic functions approximate
the trigonometric functions, sn ≈ sin, cn ≈ cos and dn ≈ 1. Thus for very small Energy (2.32) and (2.33) reduce
to
√
2E
θ (t) ≈ sin (ω t + c), (2.34)
√ω
θ̇ (t) ≈ 2E cos (ω t + c). (2.35)
These correspond to the solutions of the linearized system
θ̈ + ω 2 θ = 0.
The period of the solution is in this case s

2π 2L
T≈ = 2π ,
ω 3g
since k ≈ 0.
2.3.2 van der Pol Oscillator
ẋ1 = −x2 − x1 (x12 + x22 − 1), (2.36)

ẋ2 = x1 − x2 (x12 + x22 − 1). (2.37)
The state space of the system is X = R2 . Show that the origin [0 0]T is a hyperbolic equilibrium and that it
is unstable. Then show that the unit circle in the state space R2 is a periodic orbit. Figure 2.8 shows the phase
portrait of the van da Pol oscillator where we have plotted a selected few trajectories.
2.3.3 Double Pendulum
2.3.4 Lorenz Weather Model

76 2 Dynamics
Fig. 2.8 The phase portrait of the van da Pol oscillator.
ẋ1 = σ (x2 − x1 ), (2.38)

ẋ2 = rx1 − x2 − x1 x3 , (2.39)
ẋ3 = x1 x2 − bx3 . (2.40)
The state space of the system is R3 .
Fig. 2.9 The Lorenz Attractor.
Figure 2.9 shows a trajectory of the attractor that begins at [1 0 0]. The parameters are chosen such that
σ = 10, r = 28, b = 8/3. c
Fig. 2.10 x1 (t) Vs t and x2 (t) Vs t for very close initial conditions. The figures in the top row correspond to an initial condition
[1 0 0] while the figures in the top row correspond to an initial condition [1.001 0 0]. The solutions diverge rapidly around twenty
seconds.
2.3.5 Rigid Body Dynamics
2.3.5.1 Free Rotational Rigid Body
In this section we will analyze the dynamics (the evolution over time) of a free rotating rigid body. What we mean
by free is that no external force moments are present. That is T = 0. The free rotational rigid body equations are
Ṙ = R Ω̂ , (2.41)
I Ω̇ = I Ω × Ω , (2.42)
or alternatively
Ṙ = R I[
−1 Π , (2.43)
Π̇ = Π × I −1
Π. (2.44)
By direct calculation we will show that the Kinetic Energy, the spatial angular momentum and the magnitude
of the body angular momentum are conserved. That is KE = Ω ∙ I Ω /2, π and ||Π || are conserved. Setting T = 0
in (1.70),(1.71), and (1.89) we see that along solutions
1 1
KE = Π (t) ∙ I −1 Π (t) = Ω (t) ∙ I Ω (t) = constant, (2.45)
2 2
π (t) = R(t) I Ω (t) = constant, (2.46)
||Π (t)|| = ||I Ω (t)|| = constant. (2.47)
When the body frame is chosen so that I is diagonal (in particular for an axi-symmetric rigid body with the body
frame aligned along the axes of symmetry) then the free rigid body equations (2.44) are
78 2 Dynamics
(I2 − I3 )
Π̇1 = Π 2 Π3 , (2.48)
I2 I3
(I3 − I1 )
Π̇2 = Π 3 Π1 , (2.49)
I3 I1
(I1 − I2 )
Π̇3 = Π1 Π2 . (2.50)
I1 I2
and

1 1 Π12 Π22 Π32
KE = (I1 Ω1 + I2 Ω2 + I3 Ω3 ) =
2 2 2
+ + = E = constant, (2.51)
2 2 I1 I2 I3
   
I1 Ω1 Π1
π (t) = R(t)  I2 Ω2  = R(t)  Π2  = constant 3 × 1 matrix, (2.52)
I3 Ω3 Π3
||Π (t)|| = I12 Ω12 + I22 Ω22 + I32 Ω32 = Π12 + Π22 + Π32 = h = constant. (2.53)
We will consider three distinct cases:

Symmetric Rigid Body: I1 = I2 = I3 .
Asymmetric Rigid Body: I1 > I2 > I3 .
Axi-Symmetric Rigid Body: I1 = I2 > I3 .
The dynamics are trivial in the case of a symmetric rigid body. In this case the right hand sides of (2.48) —
(2.50) reduce to zero and all solutions remain constant for all time t.
Fig. 2.11 The figure shows several intersection curves of different energy ellipsoid with a fixed constant angular momentum sphere.
Let us consider the asymmetric rigid body, I1 > I2 > I3 : Much of the qualitative behaviour of the rigid body
can be inferred from the above conservation laws. Equation (2.51) defines a ellipsoid in the body angular momen-
tum space and (2.53) defines the surface of a sphere in body angular momentum space. What the conservation
of these quantities mean is that the solutions of the rigid body equation (2.44), in the body angular momentum
space Π , are restricted to lie on each of the surfaces and hence on the intersection of these surfaces. Figure 2.11
shows several intersection curves of different energy ellipsoid with a fixed constant angular momentum sphere.
How the constant energy ellipsoid intersects with the constant angular momentum sphere completely determines
the qualitative behaviour of the rigid body. In general, as we can see in figure 2.11, this intersection is a closed
curve. Thus the solutions in Π space correspond to closed curves and hence periodic motion in angular mo-
mentum space. As we can see in the figure 2.11 there are three special cases that occur when the intersection
de-generates to a point. Such de-generacies occur at the points where the principle axis meet the surface of the
sphere. They are explicitly of the form [0, 0, c3 ], [0, c2 , 0] and [c1 , 0, 0] and we see that they are equilibria of (2.48)
— (2.50). Since they respectively correspond to steady rotations of the body about the b3 , b2 and b1 axis they
are called relative equilibria. From figure 2.11 we can see that the relative equilibria [c1 , 0, 0] and [0, 0, c3 ] are
stable while the relative equilibria [0, c2 , 0] are unstable5 . It should be pointed out that a solution being periodic
in angular momentum Π space does not imply that the corresponding motion of the rigid body is also periodic.
To see this consider the following. Let T be the periodicity of the solution in Π space. That is Π (T ) = Π (0).
Then from the conservation of spatial angular momentum we have that RT (0)μ = Π (0) = Π (T ) = RT (T )μ
(= constant) and hence μ = R(T )RT (0)μ . Thus we see that R(T )RT (0) = Rμ is a rotation about the μ axis6 .
Thus in general R(T ) 6= R(0). The rotation angle of Rμ about μ is called the phase of the solution.
In the case of the Axi-symmetric rigid body, I1 = I2 > I3 : We see that any point on the equator lying in the
Π1 − Π2 plane of the constant angular momentum sphere is a relative equilibrium. All these relative equilibria
are unstable while the relative equilibria of the form [0, 0, c3 ] are stable.
Analytic Solution of the Asymmetric Rigid body
The conservation laws (2.51) and (2.53) implies that the rigid body equations (2.48)-(2.50) are completely inte-
grable. Meaning that the solution can be explicitly written down. In the following we show how to do this. From
(2.51) and (2.53) we have that
Π12 = α21 − β21 Π22 , (2.54)

Π32 = α23 − β23 Π22 , (2.55)
where
I1 (h − 2EI3 ) I3 (h − 2EI1 )
α21 = , α23 =
I1 − I3 I3 − I1
and
I1 (I2 − I3 ) I3 (I2 − I1 )
β21 = , β23 = .
I2 (I1 − I3 ) I2 (I3 − I1 )
Let 2EI3 < h < 2EI1 . Then both α21 > 0 and α23 > 0. If we assume I1 > I2 > I3 then β21 > 0 and β23 > 0.
Using these (2.49) can be written as
q
(I1 − I3 )
Π̇2 = (α21 − β21 Π22 )(α23 − β23 Π22 ).
I1 I3
Using the variable transformation s
β21
u= Π2 ,
α21
p s
I1 − I3 (I2 − I3 )(2EI1 − h)
ω = β21 α23 =
I1 I3 I1 I2 I3
and
α21 β23 (I1 − I2 )(h − 2EI3 )
k2 = =
α 23β21 (I2 − I3 )(2EI1 − h)
this reduces to q
u̇ = ω (1 − u2 )(1 − k2 u2 ).
This can be explicitly integrated using quadrature.
5 Recall that for stability “near by” solutions should stay “near by” for all time t
6 Recall that if R is a rotation about the axis μ in the spatial frame then Rμ = μ .
80 2 Dynamics
Z Z
1
p du = ω dt = ω t + c.
(1 − u2 )(1 − k2 u2 )
The left hand side of the above integral is the inverse of the Jacobi elliptic function sn of modulus k where
0 ≤ k ≤ 1, Z u
1
sn−1 (u) = p du.
0 (1 − u )(1 − k2 u2 )
2
Thus we have
u(t) = sn(ω t + c),
and hence s
I2 (h − 2EI3 )
Π2 (t) = sn(ω t + c).
(I2 − I3 )
Π12 = α21 (1 − u2 ) = α21 (1 − sn2 (ω t + c)) = α21 cn2 (ω t + c), (2.56)

Π32 = α23 (1 − k2 u2 ) = α23 (1 − k2 sn2 (ω t + c)) = α23 dn2 (ω t + c), (2.57)
and hence
s
I1 (h − 2EI3 )
Π1 (t) = cn(ω t + c), (2.58)
I1 − I3
s
I3 (2EI1 − h)
Π3 (t) = dn(ω t + c). (2.59)
I1 − I3
Jacobi Elliptic functions sn and cn are periodic functions of period 4K and dn is periodic of period 2K, where
Z 1 Z π /2
1 1 k2 π
K(k) = sn−1 (1) = p du = q dψ ≈ 1 + .
(1 − u2 )(1 − k2 u2 ) 4 2
0 0
1 − k2 sin2 ψ
Thus Π1 (t), Π2 (t), Π3 (t) are periodic of period

s
2π k2 I1 I2 I3 (I1 − I2 )(h − 2EI3 )
T = 4K(k)/ω = 1 + + H.O.T. ≈ 2π 1+ .
ω 4 (2EI1 − h)(I2 − I3 ) 4(I2 − I3 )(2EI1 − h)
Thus we see that Π (t) is periodic of period T . However what may be surprising is that the motion of the rigid
body is not periodic!!!
Figure 2.12 shows the simulation results of for an asymmetric rigid body of I = diag[3, 2, 1].
The angle between the body angular momentum and the minor axis of symmetry of the body, as seen in the
body frame, is given by,
s !
ϒ ∙ Π Π3 I3 (2EI 1 − h)
α (t) = cos−1 = cos−1 = cos−1 dn(ω t + c) ,
h h h2 (I1 − I3 )
where ϒ = [0 0 1] is the representation of the minor axis in the body frame (we have assumed that the body
frame is aligned along the principle axis of the body). The angle α (t)is defined as the angle of nutation of angular
momentum about the minor axis of the rigid body. Recall that the body angular momentum as observed in the
spatially fixed fame e is constant, that is RΠ (t) = π (t) = M = constant. Since π ∙ Rϒ = RΠ ∙ Rϒ = Π ∙ ϒ we see
(a) (b)
Fig. 2.12 Components of the angular momentum vector Π (t) for the Non-Aix-Symmetric Rigid Body. I = diag[3, 2, 1]. Figure (a)
corresponds to an initial condition of Π (0) = [0.6, −1, 2] thus a k = 0.2951 and T = 5.4. Figure (b) corresponds to an initial condition
of Π (0) = [3, −1, 2] thus k = 0.88 and T = 7.34.
that, when observed in the spatially fixed frame e, the angle that the minor axis of symmetry of the rigid body
makes with the constant spatial angular momentum vector M is also α (t).
Figure 2.13 shows the motion of the tip of Π (t) in the angular momentum space for a rigid body of I =
diag[3, 2, 1] and an initial condition of Π (0) = [0.6, −1, 2].
Fig. 2.13 The motion of the tip of the angular momentum vector as observed in the body co-ordinates for the asymmetric Rigid
Body. I = diag[3, 2, 1] and an initial condition of Π (0) = [0.6, −1, 2].
Analytic Solution of the Axi-symmetric Rigid Body
Let I1 = I2 > I3 . Then k = 0 and the elliptic functions reduce to trigonometric functions (sn → sin, cn →
cos, dn → 1) and we have s
(2EI1 − h)(I2 − I3 )
ω= ,
I1 I2 I3
and
82 2 Dynamics
s
I1 (h − 2EI3 )
Π1 (t) = cos(ω t + c), (2.60)
I1 − I3
s
I2 (h − 2EI3 )
Π2 (t) = sin(ω t + c), (2.61)
(I2 − I3 )
s
I3 (2EI1 − h)
Π3 (t) = = constant. (2.62)
I1 − I3
The body angular momentum Π (t) is periodic of period
Fig. 2.14 Components of the angular momentum vector Π (t) for the Symmetric Rigid Body. I = diag[2, 2, 1] and Π (0) =
[0.4, −1, 2]. The period of oscillation is T = 6.23.
s
I1 I2 I3
T = 2π .
(2EI1 − h)(I2 − I3 )
Consider the minor axis of symmetry of the body ϒ = [0 0 1]T (as represented in the body frame). The angle
between the body angular momentum and the minor axis of symmetry of the body (angle of nutation of the body
angular momentum) is given by,
s !
−1 ϒ ∙ Π −1 Π3 I3 (2EI1 − h)
α = cos = cos = cos −1
= constant
h h h2 (I1 − I3 )
and is a constant.
Figure 2.14 shows the simulation results of an symmetric rigid body of I = diag[2, 2, 1] and an initial condition
of Π (0) = [0.6, −1, 2]. The motion of the tip of the angular momentum vector as observed in the body co-
ordinates for the symmetric Rigid Body is shown in figure 2.15.
The nutation angle of motion of the body is the angle between ϒ = [0 0 1] and Rϒ . That is, as seen in the e
frame, the angle between the third axis of the spatially fixed e frame and the axis of symmetry of the body.
2.3.5.2 Heavy Top
Consider the heavy top equations (1.101) and (1.102).

2.4 Lattice Dynamics 83
Fig. 2.15 The motion of the tip of the angular momentum vector as observed in the body co-ordinates for the Symmetric Rigid
Body. I = diag[2, 2, 1] and an initial condition of Π (0) = [0.6, −1, 2].
Γ̇ = Γ × Ω , (2.63)
I Ω̇ = I Ω × Ω + mgl Γ × ϒ . (2.64)
It can be shown that the following quantities are conserved along the solutions of the system.

1 1 Π12 Π22 Π32
T E = (I1 Ω12 + I2 Ω22 + I3 Ω32 ) + mgl Γ T ϒ = + + + mgl Γ3 = E, (2.65)
2 2 I1 I2 I3
Π T Γ = Π1Γ1 + Π2Γ2 + Π3Γ3 = c, (2.66)
||Γ ||2 = Γ12 + Γ22 + Γ32 = a. (2.67)
A heavy √top for which I1 = I2 is called a Lagrange top. An upright spinning Lagrange top is stable if and only
if ||Ω || > 2 MglI1 /I3 (see Section 15.10 of [4] for further details).
2.4 Lattice Dynamics
Consider the one-dimensional, lattice with dynamics described by
mq̈n =−φ 0 (qn − qn−1 )+φ 0 (qn+1 − qn ), (2.68)
where m is the mass of the nth particle, qn is the displacement of the nth particle from its equilibrium, φ (rn ) is the
interaction potential between the nth and n − 1th particle (rn = qn − qn−1 ) and the prime denotes differentiation.
There are two types of boundary conditions under which this lattice is considered: Fixed end, and Periodic. In the
case of the fixed end boundary conditions it is customary to consider a finite lattice of N + 1 particles numbered
from 0 to N with q0 ≡ 0 and qN ≡ 0. In the case of a N-periodic lattice we assume qn+N = qn for all n where
the lattice is infinite in length. We may also consider this as a N number of particles where the Nth particle is
coupled back to the 1st particle (N particles connected in a circle).
2.4.1 The Linear Lattice

αβ 2
Consider the N-periodic harmonic lattice resulting from φ (rn ) = 2 rn , where α , β are constants such that
αβ > 0, and ω02 = αβ
m .
84 2 Dynamics
q̈n = ω02 (qn+1 − 2qn + qn−1 ) (2.69)
Letting a be the lattice spacing one seeks a travelling wave solution, of wavelength λ , of the form

2π a
qn (t) = C exp i ±ω t + n (2.70)
λ
It clearly satisfies the N-periodic boundary conditions for

Na
= M = 0, 1, 2, 3, ∙ ∙ ∙ ,
λ
where the physically reasonable limits of M are determined by the condition 2a ≤ λ ≤ Na which results in
1 ≤ M ≤ N2 . Noting that for (2.70) qn+1 + qn−1 = 2 cos (2π a/λ ) qn we see that (2.70) is a solution of (2.69) if

and only if ω 2 (λ ) = 4ω02 sin2 πλa . It is also customary to write this relationship as

ka
ω (k) = 2ω0 sin , (2.71)
2
where k = 2π /λ is known as the wave number. This relationship is referred to as the dispersion law of the lattice.
From 2a ≤ λ ≤ Na we see that the physically reasonable wave numbers k satisfy 0 ≤ k ≤ πa and are distributed
2π
with a uniform spacing of Δ k = Na .
Fig. 2.16 Dispersion curves for the lattice with nearest neighbor linear interactions.
The dispersion relationship satisfies the periodic condition ω (k) = ω (k + 2aπ ). It is customary to plot this
curve in the interval, known as the Brillouin zone, −π /a ≤ k ≤ π /a. This is shown in figure 2.16. Thus it can
be seen that physically the positive k numbers correspond to a wave length of 2a ≤ λ ≤ Na while negative k
numbers correspond the wave lengths of a ≤ λ ≤ 2a which due to the discretizing effect give rise to identical
wave shapes.
In terms of the wave number the travelling wave (2.70) takes the form
qn (t) = Ck exp i (±ω (k)t + k a n) (2.72)
However from (2.70) and (2.71) we see that k and −k gives a traveling wave of the same wavelengths but
traveling to the left and right respectively. Thus the right half of the dispersion curve corresponds to left moving
traveling waves and the left half corresponds to the identical but right moving travelling waves. A general solution
of the lattice is a linear super position of the above modes:
π /a
qn (t) = ∑ (Ck exp i (ω (k)t + k a n) + Dk exp −i (ω (k)t − k a n)) (2.73)
k=−π /a
2.4.1.1 Dispersion
The phase velocity of a given mode shape is defined to be the velocity of the traveling wave
ωλ sin (π a/λ ) sin (ka/2)

v(λ ) = = aω0 = aω0 . (2.74)
2π (π a/λ ) (ka/2)
For waves with very large wave length compared to the lattice spacing, that is for wave numbers near the Brillouin
zone center, sin (ka/2) ≈ (ka/2) and the phase velocity of the travelling wave is a maximum and is equal to
v = aω0 . This is termed to be the speed of sound in the lattice. For such waves the phase velocity is independent
of the wavelength. For waves with wavelength comparable to the lattice constant this is not true with the velocity
taking a minimum v = 0 at the Brillouin zone edges. Typically we are interested in how a local disturbance would
propagate in the lattice. From equation (2.73) it is clear that any initial perturbation will breakup into a train of
left moving waves and a train of right moving waves. Such a train of waves is called a wave packet. The waves
that form the packet consists of sinusoidal waves of different wave lengths and hence travel at different speeds.
Thus the wave packet deforms as it propagate. This process is called dispersion. Since it arises out of the ω (λ )
dependence on wavelength the relationship (2.71) is usually called the dispersion curve and is plotted in figure
2.16. It is plotted against the wave number k = 2π /λ .
One measure of how well a packet of waves move through space is captured by the notion of group velocity
πa
ka dω
u = v(λ ) − λ v0 (λ ) = aω0 cos = aω0 cos = .
λ 2 dk
From figure 2.16 it can be seen that for wave packets consisting of a group of low k (long wavelength) waves,
the group velocity is nearly a constant (equal to limk→0 u(k) = aω0 = c) and therefore the packets travel with
relatively low distortion. On the other hand for large k (small wavelength) the group velocity is almost zero and
wave packets consisting of a group of small wavelength waves travel at very low speeds. The long wavelength
modes are usually referred to as zone center acoustic phonons while the short wavelength modes are referred to
as zone edge acoustic phonons. Initial perturbations that give rise to zone edge phonons persist for longer and
contribute towards local heating. This is shown in the numerical simulation of a 128 element lattice where an
initial perturbation is provided to the middle 16 elements of the lattice. Figures 2.17 shows the time evolution of
the lattice and figure 2.18 shows the total energy in the middle 16 elements of the lattice. Figure 2.17(b) show
how the energy created by the fast moving zone center phonons disperse to the left and right and figure 2.18(b)
clearly shows the energy in the middle 16 elements dissipating rapidly. Contrastingly figure 2.18(a) shows that
the energy created by the zone edge phonons dissipates quite slowly.
(a) Zone edge phonon (b) Zone center phonon
Fig. 2.17 The time evolution of a zone edge and zone center acoustic phonon generated in the middle 16 elements of a 128 element
lattice.
In this note we are primarily interested in using feedback control to breakup the slow zone edge modes and
transform them into fast zone center modes in an energy conserving manner. Specifically the control problem
that we are interested in is that for given any initial condition, while conserving the total energy of the lattice,
the solution of the lattice should asymptotically made to converge to a solution with the following properties.
86 2 Dynamics
(a) Zone edge phonon (b) Zone center phonon
Fig. 2.18 The total energy in the middle 16 elements of the lattice.
1. The energies contained in all the last m modes are zero.

2. The energy initially contained in all the last m modes is completely transformed in to the jth mode (1 ≤ j ≤
N − m). That is H j = H j (0) + ∑Nk=N−m+1 Hk (0).
Such a solution lies on a (N − m)-Liouville torus
 
H j = H j (0) + ∑N
 k=N−m+1 Hk (0), 
M˜E = (qn , q̇n ) ∈ Rn × Rn Hk = Hk (0) ∀ k = 1, ∙ ∙ ∙ , N − m and i 6= j, .
 Hk = 0 ∀ k = N − m + 1, ∙ ∙ ∙ , N 
We will call this the problem of controlled decay of slow phonons.
2.4.1.2 Modal Decomposition
The lattice equation (2.69) can be written as

q̈ = Ω q. (2.75)
where q = [q1 , q2 , ∙ ∙ ∙ , qn ]T , and
 
−2 1 0 ∙∙∙ 0 0 1
 1 −2 1 ∙∙∙ 0  0 0
 
 0 1 −2 ∙∙∙ 0  0 0
 
2  .. .. .. .. ..  .. ..
Ω = ω0  . . . . .  . .
 
 0 0 0 ∙ ∙ ∙ −2 1 0 
 
 0 0 0 ∙ ∙ ∙ 1 −2 1 
1 0 0 ∙ ∙ ∙ 0 1 −2
This is a periodic Jacobi matrix and can be diagonalized. The modal matrix T of Ω is explicitly given by
1
Tn,1 = √ , for n = 1, 2, ∙ ∙ ∙ , N
N
r
2
Tn,2i = sin (2π in/N) for i = 1, 2, ∙ ∙ ∙ , N/2 − 1, n = 1, 2, ∙ ∙ ∙ , N
N
r
2
Tn,2i+1 = cos (2π in/N) for i = 1, 2, ∙ ∙ ∙ , N/2 − 1, n = 1, 2, ∙ ∙ ∙ , N
N
1
Tn,N = √ cos (π n) for n = 1, 2, ∙ ∙ ∙ , N
N
Let T T Ω T = D = diag([−ω12 , −ω22 , ∙ ∙ ∙ , −ωN2 ]) where we have numbered the modes in increasing order. In the
transformed co-ordinates q = T z the equation (2.75) reduces to the completely uncoupled form
z̈ = Dz. (2.76)
Explicitly writing it out we have the modal equations
z̈k = −ωk2 zk . (2.77)
This also shows that if one needs interaction between modes then one necessarily needs nonlinear feedback. The
energy in the kth mode is given by
1
Hk = (ż2k + ωk2 z2k ). (2.78)
2
The variables given by
Hk ω k zk
(Ik , θk ) = , arctan for k = 1, 2, ∙ ∙ ∙ , N (2.79)
ωk żk
are the so called action-angle variables of the system. In these co-ordinates the linear lattice (2.69) takes the form
I˙k = 0, (2.80)
θ̇k = ωk , (2.81)
for k = 1, 2, ∙ ∙ ∙ , N. From this it is easy to see that, each of the Ik and hence the Hk are invariants of the motion of
the lattice. In fact it shows that the sets
M = {(qn , q̇n ) ∈ Rn × Rn |Hk = ck ∈ R, for k = 1, 2, ∙ ∙ ∙ , N} .
are invariant under the flow of the lattice and that the flow linearizes on them. It can be easily seen that M is
generically a N-dimensional torus and that such Tori foliate the state space R2n . When certain ck are zero the tori
become de-generate.
Equations (2.80) and (2.81) can be easily integrated. Thus the solution in action angle co-ordinates can be
written down explicitly as Ik = ck , θk = ωk t + φk . Transferring back to the co-ordinates (zk , żk ) we have
zk = 2ck sin ωk t + φk , żk = 2ck cos ωk t + φk .
Now using the transformation, q = T z we can explicitly write down the original co-ordinates. Well in this case
this not surprising since we already know that we can explicitly construct the solution since the system is linear.
Completely integrable nonlinear systems are nonlinear systems that admit a local action-angle co-ordinate system
so that the dynamics take a form like (2.80) – (2.81) thereby allowing the solution to be explicitly constructed.
88 2 Dynamics
2.5 Exercises
Exercise 2.1. A linear system with forcing is given by
ẋ = Ax + u(t).
Show that the general solution of this system can be written down explicitly by the Variation of Parameters
formula: Z t
x(t) = e x(0) + eA(t−τ ) u(τ ) d τ .
At
0
Exercise 2.2. If
λ1 0 λ 0 λ 1 σ −ω
A1 = , A2 = , A3 = , A4 = .
0 λ2 0λ 0λ ω σ
show that
λt λt λt σt
A1 t eλ1 t 0 A2 t e 0 A3 t e te A4 t e cos ω t eσ t sin ω t
e = , e = , e = , e = σt .
0 eλ2 t 0 eλ t 0 eλ t e sin ω t eσ t cos ω t
Exercise 2.3. Show that

−1
eT XT = TeX T −1 .
Exercise 2.4. Consider the second order linear dynamic system
ẋ = Ax.
Let λ1 and λ2 be the eigenvalues of A.

1. Show that if λ1 and λ2 are real and distinct the solution takes the form

x1 (t) c eλ1 t + c12 eλ2 t
= 11 λ1 t
x2 (t) c21 e + c22 eλ2 t
2. If λ1 = λ2 = λ show that a solution either takes the form

x1 (t) c eλ t
= 11 λ t
x2 (t) c12 e
or
x1 (t) (c11 + c12 t)eλ t
=
x2 (t) (c21 + c22 t)eλ t
3. Show that if λ1 = σ + iω and λ2 = σ − iω a solution takes the form
σt
= σt
Exercise 2.5. Consider the RC circuit obtained by setting the inductance L = 0 in figure 2.19. Let R = 1,C = 2.
For the two cases (a) v(t) ≡ 0 and q(0) = 1, (b) v(t) ≡ 1 and q(0) = 0, do the following
1. Using ODE45 in MATLAB find q(t) numerically and plot q(t) Vs t.
2. Find the explicit (analytic) expression for q(t).
2.5 Exercises 89
Fig. 2.19 RLC Circuit
Exercise 2.6. Consider the RLC circuit shown in figure 2.19.

1. Find the governing equations of the system.
2. Find the undamped natural frequency ωn , damping ratio ζ and the damped natural frequency, ωd , of the
system in terms of the system parameters L, R and C.
Exercise 2.7. Consider the spring mass damper system shown in figure 2.20 where f (t) ≡ 0 and the damping
force fv = −Cẋ. Let M = 1, K = 1 and C = 2ζ . Consider the four cases (a) ζ = 0, (b) ζ = 0.5, (c) ζ = 1 and
(d) ζ = 1.5 that correspond respectively to an un-damped, under damped, critically damped, and over damped
system. For the initial condition that corresponds to an initial displacement of one unit and initial velocity of
zero, do the following for each of the above four cases:
1. Using ODE45 in MATLAB, numerically find the solution and obtain the following three plots:
(a) x(t) Vs t,
(b) ẋ(t) Vs t,
(c) ẋ(t) Vs x(t),
2. Find the analytical expression of the solution.
Fig. 2.20 Spring Mass Damper System.
Exercise 2.8. The figure 2.21 shows the response of a suspension system. Estimate using the graph the damping
ratio and the natural frequency of the system.
Exercise 2.9. Consider the van da Pol equations:
ẋ1 = −x2 − x1 (x12 + x22 − 1),

ẋ2 = x1 − x2 (x12 + x22 − 1).
90 2 Dynamics
Fig. 2.21 Displacement Vs Time Results of a suspension system.
1. Show that the origin [0 0]T is a hyperbolic equilibrium and that it is unstable.
2. Show that the unit circle in the state space R2 is a periodic orbit.
3. Simulate the system, using MATLAB, and obtain the solutions through the initial conditions, x(0) = [.001 0]
and x(0) = [2 0] and obtain the plots
(a) x1 (t) Vs t.
(b) x2 (t) Vs t.
(c) x2 (t) Vs x1 (t).
4. Obtain the phase portrait for several different solutions.
Exercise 2.10. Consider the Lorenz weather model:
ẋ1 = σ (x2 − x1 ),
ẋ2 = rx1 − x2 − x1 x3 ,
ẋ3 = x1 x2 − bx3 ,
with parameters σ = 10, r = 28, b = 8/3.

1. Find the equilibrium solutions of the system.
2. Determine the local stability of each of the equilibria.
3. Simulate the system, using MATLAB, and plot the solution in state space that begins at [1 0 0].
Exercise 2.11. Consider the 1-DOF electrostatic MEMS mirror model shown in figure 2.22. Write down the
dynamic system form of the governing equations. Let v(t) ≡ v.
ˉ
1. Show that there are three equilibria of the system for vˉ less than a certain critical value and that there is only
one equilibrium when vˉ greater than that critical value.
2. Determine the local stability of each of the equilibria.
3. Numerically simulating the system for ε = 1, A = 1, m = 1, k = 1, l0 = 3, and b = 0.02 obtain l(t) Vs t and
q(t) Vs t for ..........................
Exercise 2.12. Consider the simple pendulum shown in figure 2.7.

2.5 Exercises 91
Fig. 2.22 1-DOF Electrostatic MEMS model.
1. Derive the state space form (2.16) of the governing equations.

2. Show that the perfectly vertical downward configuration at rest and the perfectly vertical upward configuration
at rest are the only two equilibria of the system.
3. Derive analytical expressions for small motions of the pendulum near the equilibria for each equilibrium.
4. Show, using the linearized dynamics derived in 3, that the vertically downward configuration is locally asymp-
totically stable while the vertically upward configuration is unstable.
5. Verify your answers for question 4 using MATLAB simulations.
6. Obtain the phase portrait of the system.
7. Obtain the phase portrait of the system if the frictional torque is neglected (ie. C = 0). Using this plot explain
in detail the expected physical behavior of the simple pendulum when C = 0.
8. Analytically calculate the period of oscillations of the simple pendulum for small oscillations.
9. Using the experimental setup in the applied mechanics lab, calculate the period of oscillation for small mo-
tions and compare with the analytical results.
10. Discuss discrepancies between the theoretical and experimental period of small oscillations.
11. Show experimentally that the period of oscillation depends on the amplitude for large oscillations.
Fig. 2.23 A simple pendulum of length L and total mass M.
Exercise 2.13. For the free rigid body show that 2EI2 < h < 2EI1 when I1 > I3 > I2 . Then show that
s
(h − 2EI2 )(2EI1 − h)
ω2 =
I1 I2
and
(I1 − I3 )(h − 2EI2 )
k22 =
(I3 − I2 )(2EI1 − h)
92 2 Dynamics
s
I1 (h − 2EI2 )
Π1 (t) = cn(ω2t + c), (2.82)
I1 − I2
s
I2 (2EI1 − h)
Π2 (t) = dn(ω2t + c). (2.83)
I1 − I2
s
I3 (h − 2EI2 )
Π3 (t) = sn(ω2t + c). (2.84)
(I3 − I2 )
s
2π k2 I1 I2 (I1 − I3 )(h − 2EI2 )
T2 = 4K(k2 )/ω ≈ 1+ 2 = 2π 1+ .
ω2 4 (h − 2EI2 )(2EI1 − h) 4(I3 − I2 )(2EI1 − h)
Also when 2EI2 < h < 2EI3 and I3 > I1 > I2 then
s
(h − 2EI2 )(2EI3 − h)
ω3 =
I3 I2
and
(I3 − I1 )(h − 2EI2 )
k32 =
(I1 − I2 )(2EI3 − h)
s
I3 (h − 2EI2 )
Π1 (t) = cn(ω3t + c), (2.85)
I3 − I2
s
I2 (2EI3 − h)
Π2 (t) = dn(ω3t + c). (2.86)
I3 − I2
s
I1 (h − 2EI2 )
Π3 (t) = sn(ω3t + c). (2.87)
(I1 − I2 )
s
2π k2 I3 I2 (I3 − I1 )(h − 2EI2 )
T2 = 4K(k3 )/ω ≈ 1+ 3 = 2π 1+ .
ω3 4 (h − 2EI2 )(2EI3 − h) 4(I1 − I2 )(2EI3 − h)
Exercise 2.14. Consider the unforced Gyroscope shown in figure 2.24. If the system is started with an initial
angular velocity that has a nonzero component only about the spin axis, show that:
1. The spin axis direction will remain a constant If no external torques are applied on the system.
2. An external torque applied about the rotation axis of the outer gimbal will induce a motion about the axis of
rotation of the inner gimbal. Verify your answers using MATLAB and then using the Gyroscopic experimental
setup in the applied mechanics lab.
2.5 Exercises 93
Fig. 2.24 The Gimbal Gyroscope

Chapter 3
Input Output Representation of Linear Systems
A system is a collection of objects that interact with each other as well as with the environment that they are con-
tained in. Examples are mechanical systems, electrical systems, thermodynamical systems, biological systems
etc. Two types of quantities are required to describe the evolution of a system. Quantities that do not change
with time, called system parameters and quantities that vary with time (quantities that are a function of time),
called system variables. The system variables that are directly influenced by the interaction of the system with
its environment (external surrounding) are called the inputs of the system while the output of the system is some
function (or in the case of multiple outputs, a collection of functions) of the system variables.
3.1 Graphical Description of a Input Output System
In this class we will graphically represent a system by a box with inputs coming in and outputs leaving the box.
This allows us to think of a system as a black box that processes an input signal to give an output signal. A
system with a single input and a single output is called a single input single output (SISO) system while a system
with multiple inputs and multiple outputs is called a multi input multi output (MIMO) system. In this class we
will consider only SISO systems.
Fig. 3.1 Block diagram representation of a SISO system.
3.2 Mathematical Description of a System
Let U be the set of all possible signals (That is all possible functions of time). Then we can think of a system,
denoted by S , as a function that takes an input function, u(t), from the set U as the argument and gives the
output function, y(t), as its result. That is,
95
96 3 Input Output Representation of Linear Systems
Fig. 3.2 Block diagram representation of a three input two output MIMO system.
y(t) = S (u(t)). (3.1)
This is consistent with our desire to graphically represent a system as a box that processes the input to give us
the output. Let y1 (t) = S (u1 (t)) and y2 (t) = S (u2 (t)). That is y1 (t) is the output of the system to the input
u1 (t) while y2 (t) is the output of the system to the input u2 (t). Then, If the system satisfies the following two
properties,
S (u1 (t) + u2 (t)) = y1 (t) + y2 (t), (3.2)

S (k u1 (t)) = k y1 (t) (3.3)
it is referred to as a Linear system. A system is said to be time invariant if the system response to any given input
is exactly the same irrespective of the time at which the input was applied. For instance the behavior (response)
of a time invariant system to a sinusoidal input should be exactly the same whether you apply the input at time
t = 0 or t = 10 seconds. In this class we will consider only Linear Time Invariant (LTI) systems. Usually LTI
systems result in constant coefficient linear ordinary differential equation (ODE) models. For instance consider
a simple forced spring mass damper system shown in figure 3.4.
Let the displacement x(t) from its equilibrium be the desired output and f (t) be the input. That is y(t) = x(t),
u(t) = f (t). Applying Newton’s equations yield the following ODE model of the system in terms of the input
and output variables.
M ÿ +Cẏ + Ky = u(t). (3.4)
Fig. 3.3 The Block Diagram Representation of the Spring Mass Damper System.
However we would like to represent this system as in figure 3.3. How do we make this mathematical model
fit into our desired graphical representation? or more precisely, how can we write down the function S (•)
explicitly. The mathematical tool required for this is Laplace transforms. In the next couple of sections we will
briefly and quickly review this mathematical background. We will do so using two typical examples: a first order
system and a second order system. Any linear system can be thought of as higher order coupling of first order
and second order systems.
3.2 Mathematical Description of a System 97
3.2.1 First Order Linear Time Invariant System
Consider the single water tank system with an inflow and an outflow. Let the height of the water level of the tank
be h(t) and let it be the desired output, the cross sectional area of the tank is A, the inflow rate is qi (t). Assuming
that the outflow rate, qo (t), is a linear function of the height, that is q0 (t) = rh(t) where r is a positive constant.
Applying the principle of conservation of mass, the governing equation, ODE model of the system is,
Aḣ(t) + rh(t) = qi (t). (3.5)

r
Dividing by A and setting p1 = A and u(t) = A1 qi (t) in (3.5) we have
ḣ(t) + p1 h(t) = u(t). (3.6)
For a given u(t) that does not grow faster than an exponential function the output will also not grow faster than
an exponential function (this is a consequence of linearity of the system). Thus taking Laplace transform of both
sides of (3.6) we have

L ḣ(t) + p1 h(t) = L {u(t)}, (3.7)
(s + p1 )H(s) − h(0) = U(s). (3.8)
This yields
1 1
H(s) = h(0) + U(s). (3.9)
s + p1 s + p1
The polynomial Δ (s) = s + p1 is the characteristic polynomial of the system and
Δ (s) = s + p1 = 0, (3.10)
is the characteristic equation. The output h(t) can be calculated by taking the inverse Laplace transform of H(s).
That is

1 1
h(t) = L −1 {H(s)} = L −1 h(0) + U(s) , (3.11)
s + p1 s + p1

1
= h(0)e−p1 t 1(t) + L −1 U(s) , (3.12)
s + p1
The first part of the output

hIC (t) = h(0)e−p1 t 1(t) (3.13)
is called the initial condition response, since it depends only on the initial conditions and not on the input.
Observe that p1 = Ar depends only on the system parameters. The part

−1 1
h f (t) = L U(s) (3.14)
s + p1
is called the forced response, since it depends only on the input and has nothing to do with the initial conditions.
In this notation the complete response is
h(t) = hIC (t) + h f (t).
Note that if we set the initial condition to zero h(0) = 0 then the response of the system is simply the forced
response. Thus let us look at the forced response a little more closely.
Most of the inputs of interest to control Engineers result in a Laplace transform that is a proper rational
function of the complex variable s. That is U(s) is of the form
N(s)
U(s) = (3.15)
(s + β1 )(s + β2 ) ∙ ∙ ∙ (s + βk )
where β1 , β2 , ∙ ∙ ∙ βk are real or complex constants and N(s) is a polynomial of the complex variable s. For example
if u(t) = e−3t 1(t) then U(s) = s+3 1
and if u(t) = e−2t cos (5t) 1(t) then U(s) = (s+2)
s+2 s+2
2 +52 = (s+2+5 j)(s+2−5 j) .
Let us substitute (3.15) in the forced response (3.14) and take the partial fraction expansion of it under the
assumption that all p1 , β1 , ∙ ∙ ∙ , βk are distinct,

1 α1 r1 rk
h f (t) = L −1 U(s) = L −1 + +∙∙∙+
s + p1 s + p1 s + β1 s + βk
h f (t) = α1 e−p1 t 1(t) + r1 e−β1 t 1(t) + ∙ ∙ ∙ + rk e−βk t 1(t), (3.16)
where α1 , r1 , ∙ ∙ ∙ rk are constants determined in the process of partial fraction expansion,

α1 = U(s) s=−p1 ,
U(s)
ri = (s + βi ) .
(s + p1 ) s=−βi
The forced response given by (3.16) can further be broken up into two parts,
h(t) = htr (t) + hss (t).
Where htr (t), given by,

htr (t) = α1 e−p1 t 1(t), (3.17)
is called the the transient response. Observe that the transient response htr (t) is qualitatively similar to the initial
condition response hIC . The hss (t), given by,
hss (t) = r1 e−β1 t 1(t) + ∙ ∙ ∙ + rk e−βk t 1(t), (3.18)
is referred to as the steady state response. To illustrate why we use these names consider the following. If the
real part of p1 is positive then
lim htr (t) = lim α1 e−p1 t 1(t) = 0
t→∞ t→∞
and thus it decays to zero. Hence the name transient response. Also because of this
lim h(t) = lim htr (t) + lim hss (t) = lim hss (t).
t→∞ t→∞ t→∞ t→∞
and thus hss (t) is the response of the system when t becomes very large or in other words is the steady state
response.
3.2.2 Second Order Linear Time Invariant System
Consider a simple forced spring mass damper system shown in figure 3.4. Let the displacement x(t) from its
equilibrium be the desired output and f (t) be the input. That is y(t) = x(t) and u(t) = y(t). Newton’s equations
yield the following ODE model of the system.
M ÿ(t) +Cẏ(t) + Ky(t) = f (t). (3.19)
Dividing by M and setting ωn2 = K

M and 2ζ ωn = C
M in (3.19) we have
1
ÿ(t) + 2ζ ωn ẏ(t) + ωn2 y(t) = u(t). (3.20)
M
For a given u(t) that does not grow faster than an exponential function the output will also not grow faster than
an exponential function (this is a consequence of linearity of the system). Thus taking Laplace transform of both
sides of (3.20) we have
Fig. 3.4 Forced Spring Mass Damper System
1
L ÿ(t) + 2ζ ωn ẏ(t) + ωn2 y(t) = L { u(t)}, (3.21)
M
1
(s2 + 2ζ ωn s + ωn2 )Y (s) − ẏ(0) − (s + 2ζ ωn )y(0) = U(s). (3.22)
M
This yields
1 s + 2 ζ ωn 1/M
Y (s) = ẏ(0) + y(0) + 2 U(s). (3.23)
s2 + 2ζ ω n s + ωn
2 s2 + 2ζ ωn s + ωn
2 s + 2ζ ωn s + ωn2
The polynomial Δ (s) = s2 + 2ζ ωn s + ωn2 is the characteristic polynomial of the system and
Δ (s) = s2 + 2ζ ωn s + ωn2 = (s + p1 )(s + p2 ) = 0, (3.24)
is the characteristic equation where −p1 , −p2 are the roots of the characteristic equation polynomial Δ (s). Ex-
plicitly p
−p1 = (−ζ + ζ 2 − 1)ωn , (3.25)
p
−p2 = (−ζ − ζ 2 − 1)ωn (3.26)
and they will be either real and distinct, real and repeated, complex conjugate or purely imaginary and conjugate
depending on ζ .
Over Damped If ζ > 1 then both −p1 , −p2 will be real and distinct.
Under Damped If 0 < ζ < 1 p they are complex conjugate of each other and −p1 = −ζ ωn + iωd , and −p2 =
−ζ ωn − iωd , where ωd = ωn 1 − ζ 2 is called the damped natural frequency of the system.
Un Damped If ζ = 0 then −p1 = iωn , −p2 = −iωn and they are purely imaginary and conjugate of each
other.
Critically Damped If ζ = 1 then −p1 = −p2 = −ζ ωn and they are real and repeated.
Using (3.24)
1 s + 2ζ ωn 1/M
Y (s) = ẏ(0) + y(0) + U(s). (3.27)
(s + p1 )(s + p2 ) (s + p1 )(s + p2 ) (s + p1 )(s + p2 )
The output y(t) can be calculated by taking the inverse Laplace transform of Y (s). That is
Fig. 3.5 The location of the system poles when 0 < ζ < 1.

−1 −1 1 s + 2ζ ωn 1/M
y(t) = L {Y (s)} = L ẏ(0) + y(0) + U(s) , (3.28)
(s + p1 )(s + p2 ) (s + p1 )(s + p2 ) (s + p1 )(s + p2 )

−1 1 −1 s + 2 ζ ωn 1 −1 1
= ẏ(0)L + y(0)L + L U(s) ,
(s + p1 )(s + p2 ) (s + p1 )(s + p2 ) M (s + p1 )(s + p2 )
(3.29)
The part of this output that depends only on the initial conditions is called the initial condition response, yIC ,

1 s + 2ζ ω n
yIC (t) = ẏ(0)L −1 + y(0)L −1 (3.30)
(s + p1 )(s + p2 ) (s + p1 )(s + p2 )
and the last part of the output

1 −1 1
y f (t) = L U(s) (3.31)
M (s + p1 )(s + p2 )
is called the forced response, since its behavior depends only on the input.
3.2.2.1 Initial Condition Response
Let us look a bit closely at the initial condition response. The partial fraction expansion of a rational function
takes distinct forms depending on whether the poles are distinct or not. Thus we will have to consider these two
cases separately.
All roots of the denominator of Y (s) are distinct
Observe that when ζ 6= 1 the roots −p1 , −p2 of the characteristic polynomial are distinct. In cases I,II, and III
we have that ζ 6= 1. In these cases from partial fraction expansions

1/(p2 − p1 ) 1/(p1 − p2 )
yIC (t) = ẏ(0)L −1 +
(s + p1 ) (s + p2 )

(2 ζ ω n − p 1 )/(p 2 − p1 ) (2ζ ωn − p2 )/(p1 − p2 )
+y(0)L −1 + .
(s + p1 ) (s + p2 )
Explicitly;

1 −p1 t 1 −p2 t (2ζ ωn − p1 ) −p1 t (2ζ ωn − p2 ) −p2 t
yIC (t) = ẏ(0) e + e 1(t)+y(0) e + e 1(t).
(p2 − p1 ) (p1 − p2 ) (p2 − p1 ) (p1 − p2 )
Now let us investigate the initial condition response for each of the three cases I,II and III separately.
Over damped: ζ > 1

p p
When ζ > 1 then both
p −p1 , −p2 are real and distinct and p2 − p1 = ωn ζ 2 − 1, 2ζ ωn − p1 = (ζ + ζ 2 − 1)ωn ,
2ζ ωn − p1 = (ζ − ζ 2 − 1)ωn . Then from (3.32)
!
1 1
yIC (t) = ẏ(0) p e−p1 t + p e−p2 t 1(t) +
ωn ζ 2 − 1 −ωn ζ 2 − 1
p p !
(ζ + ζ 2 − 1)ωn −p1 t (ζ − ζ 2 − 1)ωn −p2 t
y(0) p e + p e 1(t).
ωn ζ 2 − 1 −ωn ζ 2 − 1
Under damped: ζ > 1
p 0 < ζ < 1 then −p1 and −p2 are complex conjugate of each other and p2 − p1 = iωd , where ωd =
When
ωn 1 − ζ 2 is the damped natural frequency of the system, 2ζ ωn − p1 = (ζ ωn + iωd ), 2ζ ωn − p1 = (ζ ωn − iωd ).
Then from (3.32)

1 −p1 t 1 −p2 t (ζ ωn + iωd ) −p1 t (ζ ωn − iωd ) −p2 t
yIC (t) = ẏ(0) e + e 1(t) + y(0) e + e 1(t).
iωd −iωd iωd −iωd
Observe that the partial fraction expansion coefficients are complex conjugate of each other. Now plugging in
−p1 = −ζ ωn + iωd and −p2 = −ζ ωn − iωd in the above equation we have that

1 iωd t 1 −iωd t
yIC (t) = ẏ(0)e−ζ ωn t e + e 1(t)
iωd −iωd

(ζ ωn + iωd ) iωd t (ζ ωn − iωd ) −iωd t
+y(0)e−ζ ωn t e + e 1(t). (3.32)
i ωd −iωd
1 −iπ /2 (ζ ωn +iωd ) ζ
Now observing that 1
iωd = ωd e and that iωd = reiφ where r = √ 1 and φ = arctan √1− ζ2
equation
1−ζ 2
(3.32) reduces to

ẏ(0) −ζ ωn t −ζ ω n t
yIC (t) = e cos (ωd t − π /2) + y(0)re cos (ωd t + φ ) 1(t). (3.33)
ωd
Undamped: ζ = 0
In this case ζ = 0 and the corresponding initial condition response can be derived from (3.33) by setting ζ = 0
in it. Then ωd = ωn , r = 1, and φ = 0. This gives,

ẏ(0)
yIC (t) = cos (ωnt − π /2) + y(0) cos (ωnt) 1(t). (3.34)
ωn
All roots of the denominator of Y (s) are coinciding
Coinciding roots occur when ζ = 1. That is when the system is critically damped, case IV. In this case let
−p1 = −p2 = −p = −ζ ωn . Therefore the partial fraction expansion yields

−1 1 −1 1 (2ζ ωn − p)
yIC (t) = ẏ(0)L + y(0)L + .
(s + p)2 (s + p) (s + p)2
Explicitly;

yIC (t) = ẏ(0) te−pt 1(t) + y(0) e−pt + (2ζ ωn − p) te−pt 1(t),
= e−ζ ωn t (ẏ(0) t + y(0)(1 + ζ ωn t)) 1(t). (3.35)
where the last equality follows from plugging in −p = −ζ ωn .
3.2.2.2 The Forced Response
In this section we will look at the forced response a bit more closely. If we choose all the initial conditions of
the system to be zero, that is y(0) = 0 and ẏ(0) = 0 then the initial condition response is identically equal to zero
and the response of the system is only the forced part of the response.
Most of the inputs of interest to control Engineers result in a Laplace transform that is a proper rational
function of the complex variable s. That is U(s) is of the form
N(s)
U(s) = (3.36)
(s + β1 )(s + β2 ) ∙ ∙ ∙ (s + βk )
where β1 , β2 , ∙ ∙ ∙ βk are real or complex constants and N(s) is a polynomial of the complex variable s. For example
if u(t) = e−3t 1(t) then U(s) = s+3 1
and if u(t) = e−2t cos (5t) 1(t) then U(s) = (s+2) s+2 s+2
2 +52 = (s+2+5 j)(s+2−5 j) . Refer
to table A.1 on page 813 in the text book for more examples on standard inputs and their Laplace transforms.
Substituting (3.36) in (3.31) gives

−1 1/M N(s)
y f (t) = L . (3.37)
(s + p1 )(s + p2 ) (s + β1 )(s + β2 ) ∙ ∙ ∙ (s + βk )
To evaluate this we need to take the partial fraction expansion of the right hand side. The form of the partial
fraction expansion differs depending on if the denominator contains repeated factors or not.
All roots of the denominator of Y (s) are distinct
Under this assumption the partial fraction expansion gives

−1 −1 −1 1/M N(s)
y f (t) = L {Y (s)} = L {G(s)U(s)} = L
(s + p1 )(s + p2 ) (s + β1 )(s + β2 ) ∙ ∙ ∙ (s + βk )

−1 α1 α2 r1 rk
=L + + +∙∙∙+
(s + p1 ) (s + p2 ) (s + β1 ) (s + βk )
1 1 1 1
= α1 L −1 { } + α2 L −1 { } + r1 L −1 { } + ∙ ∙ ∙ + rk L −1 { }
(s + p1 ) (s + p2 ) (s + β1 ) (s + βk )

= α1 e−p1 t 1(t) + α2 e−p2 t 1(t) + r1 e−β1 t 1(t) + ∙ ∙ ∙ + rk e−βk t 1(t) . (3.38)

α1 = (s + p1 )Y (s) s=−p1 ,

α2 = (s + p2 )Y (s) s=−p2 ,

ri = (s + βi )Y (s) s=−βi .
y(t) = ytr (t) + yss (t).
The transient response, ytr (t), is given by,

ytr (t) = α1 e−p1 t + α2 e−p2 t 1(t), (3.39)
and the steady state response, yss (t), is given by,

yss (t) = r1 e−β1 t + ∙ ∙ ∙ + rk e−βk t 1(t). (3.40)
To illustrate why we use these names consider the following. If the real parts of −p1 , −p2 are negative (observe
that this is always the case when you have nonzero positive damping, ie ζ > 0) then
lim ytr (t) = lim (α1 e−p1 t + α2 e−p2 t ) 1(t) = 0

t→∞ t→∞
and thus it decays to zero. Hence the name transient response. Also because of this
lim y(t) = lim ytr (t) + lim yss (t) = lim hss (t).
t→∞ t→∞ t→∞ t→∞
and thus yss (t) is the response of the system when t becomes very large or in other words is the steady state
response.
Let us now consider in detail the three cases that correspond to distinct roots. That is when ζ 6= 1. Specifically
in the three cases where ζ > 1 (over damped), 0 < ζ < 1 (under damped) and ζ = 0 (undamped) the transient
response will be given by (3.39). Specifically,
Over damped: ζ > 1
When ζ > 1 then both −p1 , −p2 are real and distinct. Then from (3.39)

ytr (t) = α1 e−p1 t + α2 e−p2 t 1(t), (3.41)
Under damped: 0 < ζ < 1
When 0 < ζ < 1 then −p1 and −p2 are complex conjugate of each other and −p1 = −ζ ωn + iωd , −p2 =
−ζ ωn − iωd . Then from (3.39)
ytr (t) = α1 e−p1 t + α2 e−p2 t 1(t). (3.42)
Recall that α1 and α1 are complex conjugate of each other. Then setting α1 = reiφ , where r = |α1 | and φ is the
angle of the complex number α1 , we will have that
ytr (t) = re−ζ ωn t cos (ωd t + φ ) 1(t). (3.43)
Please refer to section 3.2.2.1 for further similar details of this calculation. Also compare this with the initial
condition response (3.33) and observe the similarity in behavior.
Undamped: ζ = 0
This case is similar to case II above the only difference is that now since ζ = 0 ωd = ωn , and therefore 3.43
gives,
ytr (t) = cos (ωd t + φ ) 1(t). (3.44)
Of the roots of the denominator of Y (s), Roots −p1 and −p2 coincide but all the others are distinct
Let −p1 = −p2 = −p and −p, −β1 , ∙ ∙ ∙ , −βk are all distinct then the partial fraction expansion gives

−1 −1 −1 1/M N(s)
y f (t) = L {Y (s)} = L {G(s)U(s)} = L
(s + p)2 (s + β1 )(s + β2 ) ∙ ∙ ∙ (s + βk )

−1 α1 α2 r1 rk
=L + + +∙∙∙+
(s + p) (s + p)2 (s + β1 ) (s + βk )
1 1 1 1
= α1 L {−1
} + α2 L {
−1
} + r1 L −1 { } + ∙ ∙ ∙ + rk L −1 { }
(s + p) (s + p)2 (s + β1 ) (s + βk )

= α1 e−pt 1(t) + α2te−pt 1(t) + r1 e−β1 t 1(t) + ∙ ∙ ∙ + rk e−βk t 1(t) . (3.45)
d
α1 = (s + p)2Y (s) s=−p ,
ds

α2 = (s + p)2Y (s) s=−p ,

ri = (s + βi )Y (s) s=−βi .
y f (t) = ytr (t) + yss (t).
The transient response, ytr (t), is given by,
ytr (t) = e−pt (α1 + α2t) 1(t), (3.46)
and the steady state response, yss (t), is given by,

yss (t) = r1 e−β1 t + ∙ ∙ ∙ + rk e−βk t 1(t). (3.47)
Critically damped: ζ = 1
In this case −p1 = −p2 = −p = −ζ ωn and then
ytr (t) = e−ζ ωn t (α1 + α2t) 1(t). (3.48)
Compare this with the corresponding initial condition response (3.35) and observe the similarity in the behavior.
3.3 Transfer Function Model of Linear Time Invariant Systems
In the previous two examples of the first order system and the second order system we have seen that the initial
condition response and the transient response both have the same qualitative behaviour. Therefore analysing one
suffices. Therefore we loose no information by neglecting the initial condition response. This can be naturally
done if we set all the initial conditions to zero in our analysis. In that case the response of the system is only
the forced response part. In addition to providing us with a simplification of analysis it also gives us a way of
representing linear time invariant (LTI) systems in a way that will fit into the framework discussed in section 3.2.
3.3 Transfer Function Model of Linear Time Invariant Systems 105
In general the modelling process of a linear nth order time invariant SISO system without time delays will
result in a single nth order linear constant coefficient ODE or a set of such coupled ODEs. For convenience of
introducing ideas lets only consider the former type:
dn y d n−1 y dm y d m−1 y
+ an−1 + ∙ ∙ ∙ + a 0 y = b m + b m−1 + ∙ ∙ ∙ + b0 u, (3.49)
dt n dt n−1 dt m dt m−1
where n ≥ m.
Note that a0 , a1 , ∙ ∙ ∙ an−1 , b0 , b1 , ∙ ∙ ∙ bm are constants and depend only on the system parameters. For a given
u(t) that does not grow faster than an exponential function the output will also not grow faster than an exponential
function (this is a consequence of linearity of the system). Thus taking Laplace transform of both sides of (3.49)
and setting the initial conditions to zero we have
n
d y d n−1 y dm u d m−1 u
L + a n−1 + ∙ ∙ ∙ + a 0 y = L b m + b m−1 + ∙ ∙ ∙ + b 0 u , (3.50)
dt n dt n−1 dt m dt m−1
(sn + an−1 sn−1 + ∙ ∙ ∙ + a0 )Y (s) = (bm sm + bm−1 sm−1 + ∙ ∙ ∙ + b0 )U(s) (3.51)
This yields
bm sm + bm−1 sm−1 + ∙ ∙ ∙ + b0
Y (s) = U(s). (3.52)
sn + an−1 sn−1 + ∙ ∙ ∙ + a0
Let
bm sm + bm−1 sm−1 + ∙ ∙ ∙ + b0
G(s) = ,
sn + an−1 sn−1 + ∙ ∙ ∙ + a0
and now we have that
Y (s) = G(s)U(s). (3.53)
The polynomial Δ (s) = sn + an−1 sn−1 + ∙ ∙ ∙ + a0 is the characteristic polynomial of the system and
Δ (s) = sn + an−1 sn−1 + ∙ ∙ ∙ + a0 = 0, (3.54)
is the characteristic equation.

Looking closely at equation (3.53) we see that we have arrived at a representation of LTI systems that fit the
block diagram representation described in section 3.2. Where now what the block does is to simply multiply the
Laplace of the input U(s) by the rational function G(s) to yield the Laplace transform of the output Y (s). The
rational function G(s) is referred to as the transfer function model of the system. Remember that this rational
function is a function of the complex variable s.
Revisiting the two previous examples we see that when we set all initial conditions to zero we obtain for the
first order tank system
1
H(s) = U(s), (3.55)
s + p1
and for the second order spring mass damper system
1/M
Y (s) = U(s). (3.56)
(s + p1 )(s + p2 )
Comparing with equation (3.53) we see that the transfer function model of the single tank system is
1
G(s) = , (3.57)
s + p1
and the transfer function model of the second order SMD system is
1/M
G(s) = . (3.58)
(s + p1 )(s + p2 )
3.3.1 Forced response of the nth order system
In the absence of the initial conditions y(t) = y f (t). Thus the Laplace transform of the forced response of the
system is what is given by (3.52). For completeness lets look at this in detail.
Let −p1 , −p2 , ∙ ∙ ∙ , −pn be the roots of the characteristic polynomial. Thus
Δ (s) = sn + an−1 sn−1 + ∙ ∙ ∙ + a0 = (s + p1 )(s + p2 ) ∙ ∙ ∙ (s + pn ), (3.59)
In this notation (3.52) becomes
bm sm + bm−1 sm−1 + ∙ ∙ ∙ + b0
Y (s) = U(s). (3.60)
(s + p1 )(s + p2 ) ∙ ∙ ∙ (s + pn )
The output y(t) can be calculated by taking the inverse Laplace transform of Y (s). That is

bm sm + bm−1 sm−1 + ∙ ∙ ∙ + b0
y(t) = L −1 {Y (s)} = L U(s) , (3.61)
(s + p1 )(s + p2 ) ∙ ∙ ∙ (s + pn )
Most of the inputs of interest to control Engineers result in a Laplace transform that is a proper rational function
of the complex variable s. That is U(s) is of the form
N(s)
U(s) = (3.62)
(s + β1 )(s + β2 ) ∙ ∙ ∙ (s + βk )
where β1 , β2 , ∙ ∙ ∙ βk are real or complex constants and N(s) is a polynomial of the complex variable s1 . Let
us substitute (3.62) in (3.61) and take the partial fraction expansion of it under the assumption that all
p1 , ∙ ∙ ∙ , pn , β1 , ∙ ∙ ∙ , βk are distinct,

bm sm + bm−1 sm−1 + ∙ ∙ ∙ + b0 N(s)
y(t) = L −1 {Y (s)} = L −1 ,
(s + p1 )(s + p2 ) ∙ ∙ ∙ (s + pn ) (s + β1 )(s + β2 ) ∙ ∙ ∙ (s + βk )
( )
n
αi k
ri
=L −1
∑ +∑
i=1 s + pi i=1 s + βi
n k
= ∑ αi e−pit 1(t) + ∑ ri e−βit 1(t). (3.63)
i=1 i=1
The constants α1 , α2 , ∙ ∙ ∙ , αn and r1 , rk , ∙ ∙ ∙ , rk are determined in the process of partial fraction expansion as
follows,
αi = (s + pi )Y (s) s=−pi ,

ri = (s + βi )Y (s) s=−βi ,
Let
y(t) = ytr (t) + yss (t),
where
n
ytr (t) = ∑ αi e−pi t 1(t),
i=1
is the transient response of the system and

n
yss (t) = ∑ ri e−βi t 1(t),
i=1
is the steady state response of the system.
1 For example if u(t) = e−3t 1(t) then U(s) = 1

s+3 and if u(t) = e−2t cos (5t) 1(t) then U(s) = s+2
(s+2)2 +52
= s+2
(s+2+5 j)(s+2−5 j) .
3.5 Bounded Input Bounded Output Stability 107
3.4 Poles and Zeros of an Input Output System
The transfer functions of all real physical systems are strictly proper or proper rational functions. The poles are
the poles of it’s transfer function and the zeros of the system are the zeros of it’s transfer function. Recall that by
definition the number of zeros and number of poles of a transfer function are the same2 .
Thus the system given by the transfer function
bm sm + bm−1 sm−1 + ∙ ∙ ∙ + b0 bm sm + bm−1 sm−1 + ∙ ∙ ∙ + b0

G(s) = = ,
sn + an−1 sn−1 + ∙ ∙ ∙ + a0 (s + p1 )(s + p2 ) ∙ ∙ ∙ (s + pn )
where n ≥ m has n poles and zeros. The poles are −p1 , −p2 , ∙ ∙ ∙ , −pn and the system has m finite zeros, since
the numerator polynomial of G(s) is of order m. The remaining n − m zeros are at infinity.
For example in the case of the single tank system,
1
G(s) = ,
s + p1
the system has one pole and one zero. The pole is finite and has value −p1 while the zero is infinite. For the
spring mass damper system,
1/M
G(s) = ,
(s + p1 )(s + p2 )
the system has two poles and two zeros. The poles are finite and have values −p1 , −p2 while the two zeros are
at infinity.
3.4.1 The Effect of the Poles of the System
Observe that if all the poles of the system −p1 , −p2 , ∙ ∙ ∙ , −pn are in the strict left half complex plane, then
lim ytr (t) = 0,

t→∞
and
lim y(t) = lim yss (t).
t→∞ t→∞
On the other hand if any one of the −pi are in the right half complex plane then
lim ytr (t) = ∞,

t→∞
and
lim y(t) = ∞.
t→∞
(For instance if −p3 = 2 − 3 j, then limt→∞ e−p3 t = limt→∞ e(2−3 j)t = limt→∞ e2t e−3 jt = ∞.)
Thus the long term behavior (the behavior as limt → ∞) of the transient response depends entirely on the
location of the poles the system. If any of the poles of the system are in the right half plane then the transient
response of the system will grow unboundedly as limt → ∞.
3.5 Bounded Input Bounded Output Stability
If u(t) is bounded then all −β1 , −β2 , ∙ ∙ ∙ , −βk are in the strict left half complex plane or on the imaginary axis
(repeated values on the imaginary axis are not allowed). For example u(t) = 1(t) is bounded U(s) = 1s , u(t) =
2 See note by Benjamin C. Kuo and Farid Golnaraghi

sin (ω t) 1(t) is bounded and U(s) = s2 +ωω 2 , u(t) = e−t 1(t) is bounded and U(s) = s+1
1
. Note that u(t) = t 1(t) is
1
not bounded and U(s) = s2 , thus repeated poles of U(s) on the imaginary axis do not occur if u(t) is bounded.
Now consider what happens when a bounded input is applied to a system. For simplicity assume all
p1 , p2 , ∙ ∙ ∙ , pn , β1 , β2 , ∙ ∙ ∙ , βk are distinct. Then from (3.63)
n k
y(t) = ytr (t) + yss (t) = ∑ αi e−pi t 1(t) + ∑ ri e−βi t 1(t).
i=1 i=1
Since all −β1 , −β2 , ∙ ∙ ∙ , −βk are on either the strict left half complex plane or on the imaginary axis with no
repeated values on the imaginary axis the last k terms
k
∑ ri e−βit 1(t) = yss (t),
i=1
remains bounded as limt → ∞. Thus the whole response will remain bounded only if the transient response,
n
∑ αi e−pit 1(t) = ytr (t)
i=1
remains bounded. This will happen if all −p1 , −p2 , ∙ ∙ ∙ , −pn lie in the left half complex plane.
Therefore if the system response to any given bounded input is bounded then the poles of the system,
−p1 , −p2 , ∙ ∙ ∙ , −pn , will necessarily have to be in the strict left half complex plane. This notion of stability
is called Bounded Input Bounded Output (BIBO) stability.
Definition 3.1. A system is said to be Bounded Input Bounded Output (BIBO) stable if the output of the system
to all bounded inputs remain bounded.
Theorem 3.1. A system is BIBO stable if and only if all the poles of the system lie in the strict left half complex
plane.
s−1 s−1
Examples: The system G(s) = (s+1)(s+2)(s+7) is BIBO stable while the systems G(s) = (s+1)(s−2)(s+7) , G(s) =
s+1 1
s(s+1)(s+2) , G(s) = (s2 +1)
are not BIBO stable.
3.6 Exercises
Exercise 3.1. Identify the, system parameters, system variables, inputs and possible outputs of: a) a forced spring
mass damper system. b) a heat Exchanger. c) a water tank system.
Exercise 3.2. Consider the spring mass damper system shown in figure 2.20 with the external force f (t) = sin ω t
and M = 1. The governing equations of this system in the normal form is given by
ẍ + 2ζ ωn ẋ + ωn2 x = sin ω t.
1. Find analytically x(t) for initial conditions x(0) = 0 and ẋ(0) = 0.

2. Using the above results show that
lim x(t) = M(ω ) sin (ω t + φ (ω ))

t→∞
where
1 2ζ ωn ω
M(ω ) = p , φ (ω ) = tan −1
(ωn2 − ω 2 )2 + 4ζ 2 ωn2 ω 2 ωn2 − ω 2
3. For ωn = 1 plot M(ω ) Vs ω and φ (ω ) Vs ω for the three cases,
(a) Over damped system (1 > ζ )
(b) Under damped system (0 < ζ < 1)
(c) Critically damped system (ζ = 1)
3.6 Exercises 109
4. Numerically simulating the system for ωn = 1, and ζ = 0.05, plot the response x(t) for each of the cases
ω = 0.01, 0.1, 0.9, 1, 1.1, 10, 100. Compare your results with the graph obtained 3(b) above.
Exercise 3.3. Find the transfer function model of the system system shown in figure 3.6 from input u(t) = f (t)
to output y(t) = x2 (t).
Fig. 3.6 A 2-DOF spring mass damper system. Assume that the damping force fv is proportional to the separation velocity with
proportionality constant C.
Chapter 4
Mechanisms
4.1 Introduction
A coupling of rigid bodies through various types of connections is called a kinematic chain. The various types
of connections are called kinematic pairs while each constituent rigid body is called a link. A kinematic chain
is said to be open if at least one link is coupled to the others only through one kinematic pair and is said to be
closed otherwise.
A kinematic chain is said to be a structure if the couplings are such that no mobility remains in the chain. A
kinematic chain is said to be a mechanism if at least one link is “fixed” and is not a structure. A mechanism is
capable of transmitting force and/or motion from one place to another. The fixed link of a mechanism is usually
referred to as the frame and the motion of the rest of the links are expressed with respect to this link. By fixing
different links of a given chain we can obtain different mechanisms. This is called kinematic inversion. It can be
seen that the relative motion of all kinematic inversions are identical.
There are several types of mechanisms: linkage mechanisms, cam and follower mechanisms and gear trains.
4.2 Degrees of Freedom
A mechanism is a collection of rigid bodies with couplings between certain rigid bodies. The space of all possible
configurations of the mechanism is called the configuration space. The dimension of the configuration space is
called the degrees of freedom (DOF) of the system. If N rigid bodies make up a system and there is no coupling
between any of the bodies then we know that the configuration space of the system is of dimension 6N or in
other words has 6-DOF. Each coupling between rigid bodies imposes a holonomic constraint and reduces the
dimension of the configuration space (and hence the DOF). In general we know that if f number of constraints
exist then the dimension n of the configuration space is n = 6N − f .
For a given pair of rigid bodies the total possible unconstrained relative movability is six. When the two
bodies are coupled we say that they are joined by a kinematic pair. Kinematic pairs are classified by their degree
of movability s (ie. the degrees of freedom of the relative motion of the two links). Thus the number of constraints
imposed by a kinematic pair is 6 − s. Thus for a given mechanism of N total number of links with ns number of
pairs of movability s, where s = 0, 1, 2, 3, 4, 5 we have that the DOF of the system is given by the formula
5
n = 6N − ∑ (6 − s)ns . (4.1)
s=0
In this convention we consider the frame of the mechanism as one of the links and that it is coupled to the fixed
frame with a kinematic pair of zero movability. Equation (4.1) is commonly known as the Gruebler’s equation.
Revolute, prismatic and screw are examples of s = 1 while cylindric and roll-slide pairs are examples of s = 2.
Higher pairs of movability are spheric and planar — s = 3, cylinder-plane — s = 4, sphere-plane — s = 5.
When the chain is constrained to move in a two-dimensional Euclidian plane then (4.1) reduces to
111
112 4 Mechanisms
2
n = 3N − ∑ (3 − s)ns . (4.2)
s=0
In this course we will mainly consider planar mechanisms.

The first step in analysis of a mechanism is to sketch an equivalent skeleton or kinematic diagram. Conven-
tionally in this diagram the links are numbered and the joints are lettered with the fixed link always numbered as
1. The next step is to represent the mechanism using a graph whose vertices correspond to links and whose edges
correspond to kinematic pairs where the number of edges joining two nodes correspond to the number of DOF,
ns , of the kinematic pair that joins the two links represented by the nodes. These graphs enable one to easily
calculate the number of DOF and furthermore allows the use of graph theory to design and analyze mechanisms.
Example 4.1. Consider the Slotted Bell Crank shown in figure 4.1a. This can be considered as a five bar mecha-
nism. Number the links as follows. The frame — 1, disc — 2, the angle link — 3, the shaft — 4, the sleeve —
5. The corresponding graph is shown in figure 4.1b. The links 1 and 2 are connected through a revolute joint of
(a) (b)
Fig. 4.1 Slotted Bell Crank
movability one. Link 2 is connected to link 3 through a roll-slide contact of movability two. Link 3 is connected
to 4 through a revolute joint of movability one. Link 3 is also connected to 1 through a revolute joint of mov-
ability one. Link 4 is connected to 5 through a prismatic joint of movability one. Finally link 5 is connected to 1
through a revolute joint of movability one. Thus total links of movability zero is one, n0 = 1, movability one is
5, n1 = 5 and movability two is 1, n2 = 1. Thus from (4.2) the total degrees of freedom of the mechanism is
f = 3 × 5 − ((3 − 0)1 + (3 − 1)5 + (3 − 2)1) = 15 − 3 − 10 − 1 = 1.
The next step in the analysis process is the problem of geometric analysis. If a given mechanism has n DOF
then there exists n number of independent quantities q = (q1 , q2 , ∙ ∙ ∙ , qn ) called configuration co-ordinates that
uniquely describe the configuration of the mechanism. The output of a mechanism are a m number of functions
yk = φ (q) k = 1, 2, ∙ ∙ ∙ , m of the configuration co-ordinates q = (q1 , q2 , ∙ ∙ ∙ , qn ). It is only rarely that one can find
these functions in explicit form. Typically one finds m number of constraint equations between the qk ’s and the
yk ’s that need to be solved analytically, graphically or numerically. Sometimes the inverse problem of finding
the possible qk for a given yk is the primary concern. The next step is to obtain the corresponding relationships
between the instantaneous velocities of various points of interest in the mechanism. The last step is to find the
relationship between the generalized input forces and the generalized output forces. This is typically obtained by
neglecting the dynamics of the mechanism and by considering energy conservation. Dynamic analysis will be
4.3 Linkage Mechanisms 113
required when one needs to consider the stresses that develop in each part of a mechanism. In such an analysis
one writes down using Rigid body equations the relationships between accelerations and forces and solve them,
typically using numerical methods.
4.3 Linkage Mechanisms
In this section we look at analytical methods of analysis for linkage mechanisms. We will use the four-bar
mechanism as a prototype example. The approach applied is valid for different mechanisms as well.
4.3.1 Four Bar Mechanisms
Fig. 4.2 Four Bar Mechanism Nomenclature
As the name implies a four-bar mechanism has four links. Figure 4.2 shows a typical four-bar mechanism and
terminology used in describing such a mechanism. It can be easily seen that the mechanism has only one DOF.
Thus specifying a single quantity amounts to knowing the exact configuration of the mechanism. For instance
if the angular position of the input link is given then the angular position of the rest of the links are completely
determined up to an Inversion. That is given any four-bar mechanism there exists two distinct configurations that
correspond to a given input link position. Such configurations are called Geometric Inversions. A Dead Center
condition is said to occur when the coupler and follower links are in line. To move from one Geometric Inversion
to another one has to either dis-assemble and re-assemble the mechanism or has to move through a dead center
condition.
Fixing different links in the mechanism does not change the relative motion between them. Thus different
mechanisms obtained by fixing different links are called kinematic inversions. Figure 4.3.1 show the four differ-
ent kinematic inversions of the four-bar mechanism.
Depending on the range of motion of the two links connected to the frame link, four bar mechanisms can
be classified into four groups: crank-rocker, double-crank, double-rocker and parallelogram linkage. These are
shown in figure 4.3.1. A given four-bar mechanisms becomes one of these types depending on the length ratios
114 4 Mechanisms
Fig. 4.3 Different Kinematic Inversions of the Four-Bar Mechanism
Crank-Rocker Double-Crank
Double-Rocker Parallelogram
Fig. 4.4 Four-Bar Types
of the links. Grashof’s law states that if the sum of the shortest and longest links is less than or equal to the sum
of the other two links then there exists continuous relative rotation between two links.
Let us now try to find the relationship between the input and desired outputs of a four-bar mechanism. Con-
sider figure 4.5. Link A0 A is the input link, link B0 B is the output link, and link AB is the follower link. The
lengths of the links are A0 B0 = l1 , A0 A = l2 , AB = l3 and B0 B = l4 .
Fig. 4.5 Four-Bar Skeleton Diagram
Fix a frame e on the fixed link 1 with origin coinciding with A0 such that e1 is aligned along A0 B0 . Fix frame
a on the input link (link 2) with origin at A0 and a1 aligned along A0 A. Let frame b be fixed on the output link
(link 4) with origin at B0 and b1 aligned along B0 B. Let c be a frame on the coupler or follower link (link 3) with
origin at B and c1 aligned along BA. All frames are right handed oriented. Let
a(t) = eRφ , b(t) = eRψ , c(t) = b(t)Rγ
where we use the subscript to denote the right handed relative rotational angle between the two frames. For
instance
cos φ − sin φ
Rφ = .
sin φ cos φ
Now let
l1 l l l
L1 = , L2 = 2 , L3 = 3 , L4 = 4 .
0 0 0 0
Then we can write
A0 A = aL2 = eL1 + bL4 + cL3 ,

eRφ L2 = eL1 + eRψ L4 + eRφ Rγ L3 .
and hence that

Rφ L2 = L1 + Rψ L4 + Rψ +γ L3 . (4.3)
These can be thought of as constraint equations. Any configuration of the mechanism should satisfy them. Sim-
plifying this we have
l2 cos φ = l1 + l4 cos ψ + l3 cos λ , (4.4)

l2 sin φ = l4 sin ψ + l3 sin λ , (4.5)
where we have set λ = (ψ + γ ) and correspond to the angle that the coupler makes with the frame fixed on the
fixed link. From (4.4) and (4.5) we have,
2 2 2 2
l + l − l − l − 2l1 l2 cos φ
γ = arccos 1 2 3 4 . (4.6)
2l3 l4
From (4.5) we have after a tedious calculation that

p
ab ± a2 b2 − 4(b2 + c2 )(a2 − c2 )
sin ψ = , (4.7)
2(b2 + c2 )
where
a = l2 sin φ , b = (l4 − l3 cos γ ), c = l3 sin γ .
116 4 Mechanisms
Knowing ψ to knowing the entire configuration of the mechanism. That is, the position of any point on any of
the links is completely determined. However observe that there exists two possible solutions for ψ . These two
correspond to the two geometric inversions of the mechanism.
Typically in a four-bar mechanism one is either interested in the position of the output link or some point
fixed with respect to the coupler (or follower) link (a tracer point). We have seen that the angle ψ of the output
link is completely known as a function of the input variable φ and is given by the solutions of (4.7). If P is any
point fixed with respect to the coupler link (we call it a tracer point) then we can write
A0 P = ex = eL1 + bL4 + cX,
where x is the representation of P with respect to e and X is the representation of P with respect to c. Simplifying
we have,
x = L1 + Rψ L4 + Rλ X. (4.8)
From this expression given X we can calculate the position of the couple point with respect to the fixed frame e.
Now let us see how to find the relationship between the instantaneous velocities of the various points of
interest in the mechanism. Differentiating the constraint equation (4.3) we have,
Rφ φ̇ˆ L2 = Rψ ψ̇ˆ L4 + Rλ λ̇ˆ L3 .
Simplifying we have
l2 φ̇ sin φ = l4 ψ̇ sin ψ + l3 λ̇ sin λ , (4.9)

l2 φ̇ cos φ = l4 ψ̇ cos ψ + l3 λ̇ cos λ . (4.10)
Solving this set of equations for ψ̇ and λ̇ we have
l2 sin (λ − φ )
ψ̇ = φ̇ , (4.11)
l4 sin (λ − ψ )
l2 sin (ψ − φ )
λ̇ = φ̇ . (4.12)
l3 sin (ψ − λ )
Finally to find the instantaneous velocity of any point P that is fixed with respect to the coupler link (a tracer
point) we differentiate (4.8) and obtain,
ẋ = Rψ ψ̇ˆ L4 + Rλ λ̇ˆ X. (4.13)
To calculate the force transmission we neglect the dynamics and assume that there is no energy loss in the
mechanism. Thus the work in is equal to the work out and if Ti is the torque on the input link and To is the torque
on the output link we have that
Ti φ̇ = To ψ̇
and hence that
φ̇ l4 sin (λ − ψ )
To = Ti = Ti (4.14)
ψ̇ l2 sin (λ − φ )
In summary: equations (4.7), (4.8) describe the complete configuration analysis while equations (4.11), (4.12),
and (4.13) describe the complete instantaneous velocity analysis of a four-bar mechanism. The force transmission
in the absence of energy loss is given by (4.14). Once these equations are programmed using a desired software
one can easily analyze any four-bar mechanism that only has revolute joints.
The slider crank mechanism shown in figure 4.6 is also a four-bar mechanism. It is not a mechanism with
only revolute joints. It has a prismatic joint between the slider and the fixed link. In the following example we
will derive the equations that describe the configuration and the instantaneous velocity of the various points on
the mechanism.
Example 4.2. Consider the slider-crank shown in figure 4.6. Let r = AB. Fixing frames appropriately we can
show as before that

r sin θ
φ = arcsin , (4.15)
L
p
x = r cos θ + L cos φ = r cos θ + L2 − r2 sin2 θ , (4.16)
r cos θ r cos θ
φ̇ = θ̇ = p θ̇ , (4.17)
L cos φ L − r2 sin2 θ
2
!
r sin (θ + φ ) r cos θ
ẋ = − θ̇ = −r sin θ 1+ p θ̇ . (4.18)
cos φ L2 − r2 sin2 θ
Fig. 4.6 Slider Crank Mechanism
4.3.2 Instantaneous Center
Fig. 4.7 Instantaneous Center
A useful tool in the velocity analysis of planer mechanisms is the use of instantaneous centers. In this section
we will look at this. Let B1 and B2 be two rigid bodies moving in 2-dimensional Euclidean space. There exists
a point in both bodies with zero relative velocity. That is viewed from either body there is a point in the other
118 4 Mechanisms
that appears to be instantaneously fixed . Such a is called an instantaneous center. Let us first prove this fact.
Consider the frames e and b fixed on B1 and B2 respectively. O1 is the origin of e and O2 is the origin of b.
Now let us consider points on B2 as viewed from B1 . A point P on B2 will have representations x and X with
respect to e and b respectively. Let O1 O2 = e y and b = e Rφ then
x = y + Rφ X, (4.19)
ẋ = ẏ + R φ̇ˆ X.
φ (4.20)
If there exists a point O21 P on B2 that appears to be fixed in B1 then there exists XI such that ẋI = 0. If we can
show the existence of such a XI then we have shown the existence of an instantaneous center. If ẋI = 0, from
(4.20) it follows that
φ̇ˆ XI = −RTφ ẏ, (4.21)

1
XI = 2 φ̇ˆ RTφ ẏ. (4.22)
φ̇
That is the point in B2 with representation XI with respect to b and given by (4.22) has zero velocity with respect
to B1 . The point O21 is said to be the instantaneous center of B2 with respect to B1 . With respect to B1 , every
point on B2 appears to instantaneously rotate about P with angular velocity φ̇ . You are asked to show this in
the exercises. In addition if the bodies B1 and B2 are two convex shaped rigid bodies that are moving relative
to each other such that at each time instance their surfaces are in contact only at one point P then from the
assumption that the bodies are rigid it follows that the relative velocity of the contact point should lie along the
surface (that is perpendicular to the common surface normal through the contact point) . Thus the instantaneous
center of rotation of B2 with respect to B1 (or visa versa) lies along the common normal to the surfaces at the
contact point. We will explicitly show this in the following simple example of a roll-slide contact.
Example 4.3. A vertical disc of radius r rolling with slip on a horizontal surface. The surface can be considered
as B1 and the disc can be considered as B2 . Fix e (with origin O1 ) on B1 and fix b on B2 such that its origin O2
coincides with the center of the disc (refer to figure 4.8). Let b = e Rφ and O1 O2 = e y where y = [x r]T . Then
from (4.22) we have that the representation of the instantaneous center in the b frame is given by
" #
0
Rφ X = ẋ .
φ̇
Thus it can be seen that the instantaneous center lies on the line joining the center of the disc and the contact point
and that it is at a distance ẋ/φ̇ away from the center of the disc. Now if in addition we impose no slip conditions
(rolling without slipping) then the instantaneous center should be at the contact point. Hence ẋ/φ̇ = −r and
therefore ẋ = −rφ̇ as expected.
4.3.2.1 Kennedy’s Theorem
An important result applicable for two-dimensional mechanisms is what is known as Kennedy’s theorem. This
states that three rigid bodies in relative motion with each other will have their respective instantaneous centers
lying in a straight line.
To prove Kennedy’s theorem consider three rigid bodies B1 , B2 and B3 moving in 2-dimensional Euclidean
space. Let e, b and c be frames fixed on B1 , B2 and B3 respectively. At a given particular time instant, without
loss of generality, we pick the origins of e and b to coincide with the instantaneous center O21 of B2 with respect
to B1 and the origin c to coincide with the instantaneous center O31 of B3 with respect to B1 . What we need to
show is that the instantaneous center O32 of B3 with respect to B2 lies on the line joining O21 to O31 . To do this
we observe points on B3 with reference to B2 . Let P be a point on B3 . The representation of P with respect to
e is x, with respect to b is Xb and with respect to c is Xc and let b = e Rθ , c = e Rφ . Let O21 O31 = e y. Then
x = Rθ Xb = y + Rφ Xc . (4.23)
Fig. 4.8 A Rolling Disc
Fig. 4.9 Kennedy’s Theorem
Now differentiating (4.23) we have
ẋ = Rθ Ẋb + Rθ θ̇ˆ Xb = ẏ + Rφ Ẋc + Rφ φ̇ˆ Xc . (4.24)
Let P = O23 . The instantaneous velocity of O32 with respect to B2 (and hence the b frame) and B3 (and hence
the c frame) is zero. Thus Ẋb = 0 = Ẋc . Since O31 is the instantaneous center of B3 with respect to B1 the point
O31 has zero instantaneous velocity with respect to B1 . Thus ẏ = 0. Substituting these in (4.24) we have
ẋ = Rθ θ̇ˆ Xb = Rφ φ̇ˆ Xc . (4.25)
Using (4.23) we then have

θ̇ˆ x = φ̇ˆ (x − y). (4.26)
and finally that
120 4 Mechanisms
φ̇
x= y. (4.27)
(φ̇ − θ̇ )
From (4.27) we see that O32 lies on the line joining the instantaneous centers O21 and O31 and we have proved
Kennedy’s theorem that the instantaneous centers of three rigid bodies moving in 2-dimensional Euclidean space
lie on the same line. Furthermore (4.27) gives the ratio of the distances to the instantaneous centers as a function
of the relative angular rotations of the two bodies. Explicitly
φ̇
O21 O23 = O21 O31 . (4.28)
(φ̇ − θ̇ )
By substituting O21 O31 = O21 O23 + O23 O31 in this we also have
O21 O23 θ̇ + O31 O23 φ̇ = 0. (4.29)
Example 4.4. Consider two convex shaped bodies B2 and B3 coupled to a frame. The frame can be considered
to be another body B1 and to be fixed. The two bodies are pin joined to the frame such that B2 is pinned at O21
and B3 is pinned at O31 . Both bodies move in a plane with their free ends in contact with each other at a point P.
This device can be considered to be a 3-bar mechanism. If this mechanism is to transmit the rotary motion of B2
to a rotary motion of B3 then from (4.2) we see that the contact between B2 and B3 should have two degrees of
freedom. That is the contact should both allow rolling and sliding (a roll-slide contact). In the following to find
the relationship between the rotational velocities of the two links B2 and B3 we will use Kennedy’s theorem
and the constraint that the contact surfaces are rigid.
It is clear that the instantaneous center of B1 and B2 is O21 and that of B1 and B3 is O31 . Now according
to Kennedy’s theorem the instantaneous center, O32 , of B2 with respect to B2 lies on the straight line O21 O31
(that is O21 , O32 , O31 are co-linear). However it does not tell us the exact location of O32 . To determine this we
take into consideration the constraint that the contact surfaces of B2 and B3 are rigid and thus the velocity of
the contact point P on B3 with respect to B2 should lie in the direction along the tangent to the surface (that
is perpendicular to the common normal to the contact surfaces through P). Now since O32 is the instantaneous
center of B3 with respect to B2 the velocity of P with respect to B2 is perpendicular to O32 P. Thus O32 P
should lie along the common normal to the surfaces. Hence O32 is the intersection point between O21 O31 and
the common normal to the contact surfaces through P.
Let ||O21 O32 || = ||x|| = l1 and ||O21 O31 || = ||y|| = l and the angular velocity of B2 about O21 in the counter
clockwise direction be θ̇ . Then from (4.27) we have that the angular velocity of B3 about O31 in the counter
clockwise direction φ̇ is given by
−l1
φ̇ = θ̇
(l − l1 )
From this relationship, it is interesting to note that, when θ̇ is a constant for φ̇ to be a constant l1 should remain
to be a constant. We will see later that this plays a crucial role in the design of gear tooth profiles.
Example 4.5. Given the mechanism shown in figure 4.10 let us use Kennedy’s theorem and the assumption that
the disc rolls inside the cylinder without slip.
Consider the three rigid bodies: the frame, the bar and disc and name them B1 , B2 , B3 respectively. It
can be seen that the instantaneous centers of B1 and B2 is O21 = O, B2 and B3 is O23 = O0 . Let O31 be the
instantaneous center of B1 and B3 . Then OO0 = O21 O23 = l and if O31 O23 = O31 O0 = z we have from (4.29)
that
θ̇
z = l. (4.30)
φ̇
Recalling that in the proof of the Kennedy’s theorem equation (4.29) was derived assuming that O23 is between
O21 and O31 (see figure 4.9). Thus the velocity of the contact point, P, on the disc with respect to the frame
is O31 Pφ̇ = (r − z)φ̇ = rφ̇ − l θ̇ in the direction perpendicular to OO0 . Now since the disc is rolling inside the
cylinder without slipping the velocity of the contact point on the cylinder is equal to the velocity of the contact
point on the disc. Thus O21 Pψ̇ = O31 Pφ̇
(l + r)ψ̇ = rφ̇ − l θ̇ ,
and hence
4.4 Cam Mechanisms 121
Fig. 4.10 Disc rolling inside a cylinder
r l
ψ̇ = φ̇ − θ̇ (4.31)
(l + r) (l + r)
4.4 Cam Mechanisms
Fig. 4.11 Cam Nomenclature

122 4 Mechanisms
4.5 Gear Trains
Fig. 4.12 Gear Types

4.6 Exercises 123
Fig. 4.13 Law of Gear Meshing
4.6 Exercises
1. Obtain the corresponding graph of the mechanisms shown in figure 4.20, 4.19 and find their number of DOF.
2. Find the degrees of freedom of the digger shown in figure 4.20
3. Using the expressions derived in example 4.2 find the length ratios of the slider crank mechanism shown in
figure 4.21 so that the point P moves in a vertical straight line.
4. Considers two bodies moving relative to each other. Show that every point in one body, when viewed in the
other, appears to rotate instantaneously about the instantaneous center.
5. Consider the centrifugal governor shown in figure 4.22. Find the relationship between the input wheel angular
velocity and the velocity of the point D.
6. Consider the Hood opening mechanism shown in figure 4.23. The link pivoted at A is the driving link and the
one pivoted at B is the driven link.
(i.) Find the relationship between the input angle θ and the outputs φ and l3 .
(ii.) Assuming that the disc is attached firmly to the input link, find the angular velocity, φ̇ , of the output
link.
(iii.) Replacing the assumption that the disc is firmly attached to input link by the assumption that the disc
is pivoted to the input link, find the angular velocity of the disc and the angular velocity of the output link.
Observe that this change of assumption does not affect the calculation of the angular velocity of the output
link.
124 4 Mechanisms
Fig. 4.14 Gear Nomenclature
References
1. Haim Baruh, Analytical Dynamics McGraw-Hill, 1999.

2. R. Abraham and J. E. Marsden, Foundations of Mechanics, Second Ed. Westview, 1978.
3. V. I. Arnold, Mathematical Methods of Classical Mechanics, Second Ed., Springer-Verlag, New York 1989.
4. J. E. Marsden and T. S. Ratiu, Introduction to Mechanics and Symmetry, Second Ed. Springer-Verlag, New York 1999.
5. F. Bullo and A. D. Lewis, Geometric Control of Mechanical Systems: Modeling, Analysis, and Design for Simple Mechanical
Control Systems, Springer-Verlag, New York 2004.
6. J. J. Uicker, J. E. Shigley, and J. R. Pennock, Theory of Machines and Mechanisms, Oxford University Press, 2003.
7. A. G. Erdmam, G. N. Sandor, and S. Kota, Mechanical Design Analysis and Synthesis: Vol. I, Prentice Hall, 1996.
8. N. Chironis and S. Sclater, Mechanisms and Mechanical Devices Source Book, McGraw-Hill, 3rd edition, 2001.
9. Y. Zhang and S. Finger, Introduction to Mechanisms, available online at http://www.cs.cmu.edu/ rapid-
proto/mechanisms/tablecontents.html
10. G. J. Sussman and J. Wisdom, Introduction to Classical Mechanics, (available online at http://mitpress.mit.edu/SICM/), MIT
Press, 2001.
11. D. T. Greenwood, Advanced Dynamics, Cambridge University Press, 2003.
References 125
Fig. 4.15 Gear Design Parameters
(a) (b)
Fig. 4.16 Epicyclic Gear Train at the Applied Mechanics Lab

126 4 Mechanisms
Fig. 4.17 Epicyclic Gear Train Example 2
Fig. 4.18 Epicyclic Gear Example 3

References 127
Aircraft Door Hood Opening Mechanism
Windshield Wiper Plier
Quick Return Mechanism Slider Crank Higher Pair
Fig. 4.19 Linkage Mechanisms

128 4 Mechanisms
Fig. 4.20 Digger Mechanism
Fig. 4.21 Slider-Crank Mechanism

References 129
Fig. 4.22 Centrifugal Governor
Fig. 4.23 A hood opening mechanism

Mechanics Book

Uploaded by

Copyright:

Available Formats

Mechanics Book

Uploaded by

Document Information

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

Mechanics Book

Uploaded by

Copyright:

Available Formats

Maithripala D. H. S.

Introduction to Mechanics, Dynamics and

February 14, 2009

Hapugala, Galle, Sri Lanka, D. H. S. Maithripala

3 Input Output Representation of Linear Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

1.1.1 Configuration Space, Co-ordinates and Degrees of Freedom

In these notes the matrix  

then the distance between P and Q is assumed to be given by,

Fig. 1.4 The Double Planar Pendulum

1.1.2 Holonomic Constraints

1.2 Particle Motion

1.2.1 Kinematics of a Particle Moving in Space

Fig. 1.5 The Velocity of a Particle described with respect to a frame e.

1.2.2 Kinetics of a Particle Moving in Space

mẍ = m a(t) = f (t), (1.3)

1.2.3 Relative Motion

1.2.3.1 Parallely Translating Frames

x(t) = o(t) + X(t). (1.4)

Observing that symbolically this means

OP(t) = e x(t) = e o(t) + b(t) X(t)

Fig. 1.6 Frame b(t) is parallely translating with respect to frame e.

OP(t) = OO0 (t) + O0 P(t),

for addition of “vectors”.

The velocity of the point P(t) measured in frame e is

and measured in the moving frame is

v(t) = ẋ(t) = ȯ(t) + Ẋ(t) = ȯ(t) +V (t). (1.5)

The acceleration of the point P(t) measured in the e frame is

a(t) = v̇(t) = ẍ(t),

and measured in the moving frame is

a(t) = v̇(t) = ö(t) + V̇ (t) = ö(t) + A(t). (1.6)

From (1.6) and (1.3) we have

F b (t) = m A(t) = −m ö(t) + f (t). (1.7)

1.2.3.2 Rotating Frames

The point P(t) can be expressed in both fames as

OP(t) = e x(t) = b(t) X(t) = e R(t)X(t), (1.10)

x(t) = R(t)X(t). (1.11)

Now differentiating (1.11) we have that

The quantity Vb (t) is given by

and hence that

RT (t)v(t) = Vb (t) = Ω̂ (t)X(t) +V (t). (1.15)

Ab (t) = Ω̂ (t)Vb (t) + V̇b (t),

From (1.18) and (1.3) we have that

and hence in the rotating moving frame we have that

F b (t) = m A(t) = −m Ω̂ 2 (t)X(t) − 2m Ω̂ (t)V (t) − m Ω̂˙ (t)X(t) + RT f (t). (1.21)

Fapp = −m Ω̂ 2 (t)X(t) − 2m Ω̂ (t)V (t) − m Ω̂˙ (t)X(t). (1.22)

Effects of Earth’s Rotation about its Axis

and θ is the angle of rotation of the Earth about its axis.

e x(t) = b(t) (o + X p (t)) = eR2 R1 (o + X p (t)) = eRX.

m Ẍ(t) = −m Ω̂ 2 (t)X(t) − 2m Ω̂ (t)Ẋ(t) − m Ω̂˙ (t)X(t) + RT f (t). (1.23)

(a) A Southern Hurricane (Catarina). (b) A Northern Polar Hurricane.

To compute Ω̂ consider Ṙ = RΩ̂ . Since R = R2 R1 differentiating we have that

Ṙ = Ṙ2 R1 = R2 Ω̂e R1 = R2 R1 RT1 Ω̂e R1 = RΩ̂ ,

and hence that

Thus explicitly written down equation (1.27) reads

Integrating the third equation twice we have

X p3 (t) = Vp3 (0)t + X p3 (0).

Vp2 (t) = −gt +Vp2 (0),

Vp1 (t) = −Ωe gt 2 + 2ΩeVp2 (0)t +Vp1 (0),

1.2.3.3 General Moving Frames

Fig. 1.10 Frame b(t) is is translating and rotating with respect to e.

Then in this compact notation

Now we also have that