You are on page 1of 265

UNIVERSITY OF WISCONSIN – MADISON

Department of Engineering Physics

EMA 542
Advanced Dynamics
EMA 542
Advanced Dynamics

Copyright  1974
Alois L. Schlack, Jr.
Philip G. Kessel

Revised 2002
Daniel C. Kammer
Department of Engineering Physics
University of Wisconsin – Madison
Engineering Research Building • Room 539
1500 Engineering Dr, Madison, WI 53706
Phone 608.262.5724 • Fax 608.262.6707
Table of Contents
1.1 Motion of a Point: Position, Displacement, Velocity and Acceleration Vectors ....................................... 1

1.2 Motion of a Line: Angular Displacement, Angular Velocity Vector and Angular Acceleration Vector 3

1.3 Derivative of a Vector – Basic Concept ........................................................................................................ 7


Example 1.1 Second Derivative of a Vector .......................................................................................................... 9
Example 1.2 Angular Velocity and Acceleration of Motor on Platform ............................................................. 10

1.4 Derivative of Position Vectors Expressed in Various Cartesian Coordinate Systems ........................... 12
(A) Rectangular Coordinates ............................................................................................................................. 12
(B) Cylindrical Coordinates .............................................................................................................................. 12
(C) Spherical Coordinates ................................................................................................................................. 14
(D) Normal and Tangential Coordinates ........................................................................................................... 15

1.5 Examples of Particle Motion ....................................................................................................................... 19


Example 1.3 Radius of Curvature ........................................................................................................................ 19
Example 1.4 Circular to Rectilinear Motion Mechanism .................................................................................... 21
Example 1.5 Oscillating Particle in Groove......................................................................................................... 23
Example 1.6 Slotted Guide .................................................................................................................................. 25
Example 1.7 Helical Path..................................................................................................................................... 27

1.6 Derivative of a General Vector that is Referred to a Rotating Coordinate System ............................... 32
Example 1.8 Angular Velocity and Acceleration of Motor on Platform 2 .......................................................... 36
Example 1.9 Particle Relative to Rotating Platform ............................................................................................ 40

1.7 Motion of a Particle Relative to a Moving Frame of Reference ............................................................... 43

1.8 Examples of Rotating Coordinate Systems and Relative Motion ............................................................ 46


Example 1.10 Particle on Platform Using Rotating Coordinate System.............................................................. 46
Example 1.11 Pendulum on Circular Platform .................................................................................................... 48
Example 1.11 Cutting Tool Mechanism .............................................................................................................. 51
Example 1.13 Projectile Fired from Slewing Gun Turret .................................................................................... 53

1.9 Multiple Moving Reference Frames ........................................................................................................... 56


Example 1.14 Projectile Fired from Slewing Gun Turret 2 ................................................................................. 58

1.10 The Coriolis Effect for Motion Relative to the Rotating Earth................................................................ 61
Example 1.15 Projectile with Coriolis Effects..................................................................................................... 66

1.11 Rigid Body Kinematics ................................................................................................................................ 67


A. Euler’s Theorem............................................................................................................................................. 70

i
B. Chasles’ Theorem .......................................................................................................................................... 71
C. Instantaneous Centers for Plane Motion ........................................................................................................ 72
Example 1.16 Instantaneous Center of Gears ...................................................................................................... 75
Example 1.17 Instantaneous Center of Crank Mechanism .................................................................................. 78

1.12 Euler Angles ................................................................................................................................................. 82

1.13 General Euler Angle Transformations....................................................................................................... 86

KINETICS OF PARTICLES .................................................................................................. 89

2.1 Direct Integration of Newton’s Second Law .............................................................................................. 91


I. Constant acceleration (constant force) ........................................................................................................... 91
II. Force as a function of velocity ....................................................................................................................... 92
III. Force as a function of time.......................................................................................................................... 92
IV. Force as a function of displacement............................................................................................................ 93

2.2 Examples of Direct Integration of Newton’s Law ..................................................................................... 94


Example 2.1 Projectile Motion with Laminar Viscosity...................................................................................... 94
Example 2.2 Projectile Motion with Turbulent Viscosity.................................................................................... 96
Example 2.3 Elasticity Effect on Crane Locking Mechanism ........................................................................... 100
Example 2.4 External Force on Crane Locking Mechanism.............................................................................. 106
Example 2.5 Mass on Spring in Rotating, Laminar Viscous Slot ...................................................................... 108

2.3 Newton’s Second Law for a System of Particles...................................................................................... 114


Example 2.6 Force on Weather Vane Shaft ....................................................................................................... 117
Example 2.7 Force of Falling Chain .................................................................................................................. 119

WORK-ENERGY METHODS FOR PARTICLES ..................................................................... 123

3.1 Work-Energy Principle for a Single Particle........................................................................................... 123


Example 3.1 Crate Stopped by Bumper............................................................................................................. 126

3.2 Conservative Force Systems...................................................................................................................... 131

3.3 Potential Energy and Conservation of Energy ........................................................................................ 136


Example 3.2 Conservative Force Systems ......................................................................................................... 139
Example 3.3 Potential Energy of Springs, Weights and Gravitation ................................................................. 141
Example 3.4 Natural Frequency of Mathematical Pendulum ............................................................................ 145

3.4 Work-Energy Methods for a System of Particles.................................................................................... 149


Example 3.5 Work Done by Gravity ................................................................................................................. 155
Example 3.6 Energy of Cart on Circular Platform............................................................................................. 156
Example 3.7 Hammer Stamping Mechanism..................................................................................................... 159
Example 3.8 Vibrating Beam............................................................................................................................. 161

IMPULSE-MOMENTUM METHODS FOR PARTICLES ............................................................. 165

4.1 Linear Impulse-Momentum Methods for a Single Particle.................................................................... 165


A. Superposition Integral Methods ................................................................................................................... 166
Example 4.1 Displacement of Particle under Transient Force........................................................................... 173
Example 4.2 Spring-Mass System under Transient Force ................................................................................. 177
b) Application of Fourier Series Techniques.................................................................................................... 181

ii
Example 4.3 Fourier Series of Forcing Function ............................................................................................... 183

4.2 Linear Impulse-Momentum Methods for a System of Particles ............................................................ 184

4.3 Conservation of Linear Momentum ......................................................................................................... 186


Example 4.4 Linear Momentum of Man and Boat ............................................................................................ 187

4.4 Collision Problems ..................................................................................................................................... 189


Example 4.5 Impacting Cars.............................................................................................................................. 194

4.5 Angular Momentum Methods for a Particle ........................................................................................... 196


Example 4.6 Motion of Pendulum on Cart using Angular Momentum ............................................................. 199

4.6 Angular Impulse-Momentum Principle for a Particle............................................................................ 208

4.7 Angular Momentum Methods for a System of Particles......................................................................... 209


Example 4.7 Model of a Spring-Lever System.................................................................................................. 214
Example 4.8 Elastic Mechanism on Rotating Platform ..................................................................................... 216

4.8 Angular Impulse Momentum Principle for a System of Particles ......................................................... 220
Example 4.9 Astronauts Walking in Spinning Space Station ............................................................................ 223
Example 4.10 Projectile Striking a Pendulum ................................................................................................... 225

KINETICS OF RIGID BODIES ............................................................................................ 228

5.1 Translational Equations of Motion........................................................................................................... 228

5.2 Angular Momentum of a Rigid Body ....................................................................................................... 230

5.3 Rotational Equations of Motion for Rigid Bodies ................................................................................... 233


Example 5.1 Force of Rolling Disc constrained by Rod.................................................................................... 242

5.4 Euler’s Equations of Motion Using Euler Angles.................................................................................... 246

MOMENT OF INERTIA ...................................................................................................... 249

PRODUCT OF INERTIA .................................................................................................... 251

PARALLEL AXIS THEOREMS ........................................................................................... 252

TRANSFER OF MOMENTS OF INERTIA FOR INCLINED AXES ................................................ 254

SUMMARY OF TRANSFER FORMULAS FOR PRINCIPAL AXES .............................................. 256

INERTIA TENSOR ........................................................................................................... 257

SUMMARY OF GENERAL INERTIA TERMS FOR INCLINED AXIS............................................. 259

iii
1
Chapter

Kinematics of Particles and Rigid Bodies

K inematics is the branch of dynamics concerned with the motion of particles and rigid bodies
without regard to the forces causing the motion. A thorough mastery of this phase of
dynamics is essential for a solid understanding of kinetics, which is the portion of dynamics relating
motion to the forces causing the motion.

The study of kinematics is primarily geometrical and mathematical in nature and does not depend
directly on the physical laws employed in the study of kinetics. Thus, whenever one establishes a
coordinate system as a frame of reference in order to locate and orient a particle or rigid body as a
function of time, it is immaterial in kinematics whether the coordinate system is an inertial or a
relative frame of reference. An inertial frame of reference, frequently referred to as an absolute
coordinate system, is required only when one introduces Newton’s law of motion into the analysis.
Consequently, in the following development of kinematical relationships, one may interpret all
coordinate systems as either relative or inertial. Usually however we will imply an inertial reference
frame so that the subsequent applications of these kinematical relationships will be consistent with
Newtonian mechanics.

1.1 Motion of a Point: Position, Displacement, Velocity and


Acceleration Vectors
Point P shown in Figure 1.1 is located relative to an x, y, z reference frame at time t by the position
r
vector r directed from the origin O to point P.

1
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

C
r
v P r
v
r
∆r
P′
r
r r r
r r v + ∆v
r + ∆r
O x

Figure 1.1: Motion along path C.

r
If point P moves along the path C to P’, undergoing a displacement ∆ r in the time interval ∆ t , the
r
r ∆r
velocity vector v is found by taking the liming value of as ∆ t approaches zero given by
∆t

r r
r ∆r dr r&
v = lim = =r (1.1)
∆t →0 ∆t dt

r r
The direction of ∆ v is tangent to the path of motion P and v is the speed.

r r r r
Similarly, letting the velocity be v at time t and v + ∆ v at time t + ∆ t , the acceleration vector a is
r
∆v
found by taking the limiting value of as ∆ t approaches zero yielding
∆t

r r r
r ∆v dv r& d 2 r &r&
a = lim = =v = 2 =r (1.2)
∆t →0 ∆t dt dt

r
If one considers the velocity v as a free vector originating from the origin O’ of the reference frame
shown in Figure 1.2, then the locus of the tip of the velocity vector as a function of time is known as
r
the hodograph. The acceleration a is tangent to the hodograph at any time t as shown.

2
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

vy
r
v t
r
a
r
∆v t + ∆t
r
v
r r
v + ∆v
O′
vx

vz
Figure 1.2: A hodograph; locus of the velocity vector tip through time.

1.2 Motion of a Line: Angular Displacement, Angular Velocity


Vector and Angular Acceleration Vector
In order to describe the kinematical properties of particle motion, it is convenient to consider the
particle to be sufficiently small so that its motion is characterized by the motion of a reference point on
the particle. However, in addition to the linear displacements, velocities and accelerations discussed
in the previous section, it is often useful to describe the motion of a point P in terms of the motion of
the line OP in Figure 1.1, for example. Furthermore, for rigid body motion it becomes necessary to
consider the rotational motion of reference lines on the body in addition to the linear displacement,
velocity and acceleration of an appropriate reference point on the body.

Rotational displacements used for describing the motion of a line are considerably more difficult to
deal with than linear displacements, since it can be readily shown that finite angular displacements are
not vectors. This can be done by recalling that in order for a quantity to be a vector it must have
magnitude and direction and obey the laws of vector algebra. Unfortunately, finite angular
displacements do not obey the commutative law of vector addition, which can easily be demonstrated
by the following example.

Referring to Figure 1.3, the line segment OA, which is initially oriented along the y axis, is subjected
to a series of two consecutive 90º rotations, called θx and θz, about the x and z axes respectively. If the
rotation sequence of θx followed by θz is executed, the line OA ends up along the positive z axis as
shown in Figure 1.3 as line OA’. However, if the sequence θz followed by θx is carried out, line OA

3
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

ends up oriented along the negative x axis, shown as line OA’’. Thus, the final orientation of the line
OA is dependent on the order in which the displacements occur so that the commutative law of vector
addition would not be satisfied if one were to attempt to describe θx and θz as vectors, although it is
easy to assign a magnitude and direction to each.

A′′ O
x

θ x = 90°
A′
z θ z = 90°
Figure 1.3: Finite angular displacements.

It is clear from the previous discussion that in order to describe the angular displacement of a line (on
a rigid body, for example) for a given application, a definite rotational sequence must be defined and
strictly adhered to for that application. Of course, any appropriate rotational sequence may be selected
for a given problem as long as it is used consistently throughout.

Fortunately, however, infinitesimal angular displacements do obey the commutative law of vector
addition, as well as other laws of vector algebra, and consequently can be treated as vectors. In order
to show that infinitesimal angular displacements obey the commutative law, consider the line segment
OA shown in Figure 1.4 to be displaced from its original orientation by the following two rotational
sequences:

1. Angular displacement θx followed by θz

2. Angular displacement θz followed by θx

The angular displacements are considered positive according to the right hand rule applied to torques,
realizing that with point O held fixed the coordinates of point A appropriately describe the orientation

4
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

of line A. Further consider that the y axis is selected in the direction of the original orientation of OA
which is not necessarily vertical as drawn.

O
x

θx
θz
z

Figure 1.4: Infinitesimal angular displacements.

Sequence 1: θx followed by θz

After an angular displacement of θx about the x axis, the direction cosines l’, m’, and n’ of line OA are
given by

l′ = 0
m′ = cos(θ x )
n′ = sin(θ x )

Then after a further rotation from this orientation through an angle θz about the z axis, the final
direction cosines l1, m2, and n1 of line OA for sequence 1 are

l1 = − cos(θ x ) sin(θ z )
m1 = cos(θ x ) cos(θ z ) (1.3)
n1 = sin(θ x )

Sequence 2: θz followed by θx

Similarly, the direction cosines after the angular displacements θz followed by θx are

5
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

l2 = − sin(θ z )
m2 = cos(θ z ) cos(θ x ) (1.4)
n2 = cos(θ z ) sin(θ x )

Comparing the results of Eqs. 1.3 and 1.4 we see as expected that the final orientations of line OA and
consequently the final positions of point A relative to O are different for each sequence of finite
rotations. However, if we consider the limit as the angular displacements become small for which θx
= dθx and θz = dθz we find that both sets of direction cosines approach the same values, neglecting
higher order terms, which are given by

l = − dθ z
m =1 (1.5)
n = dθ x

Thus, infinitesimal angular displacements add up vectorially and are treated accordingly as vectors.

r
This fact is especially important since we may now define the angular velocity vector ω in a way
r
r ∆θ
similar to that used for linear velocity v by taking the limiting value of as ∆ t approaches zero
∆t
yields

r r
r ∆θ dθ r&
ω = lim = =θ (1.6)
∆t → 0 ∆t dt

Recalling that the orientation of a line in space is independent of the reference point selected in any
r r
application, the angular velocity of ω is a free vector. The direction of ω is perpendicular to the
r
place containing the angle dθ and the sense is given by the customary right hand rule customarily
applied to torques.

r r r
In addition, let the angular velocity be ω at time t and ω + ∆ ω at time t + ∆ t so that the definition of
angular acceleration follows as

r r r
r& ∆ω dω r& d 2θ &r&
ω = lim = =ω = 2 =θ (1.7)
∆t → 0 ∆t dt dt

6
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

1.3 Derivative of a Vector – Basic Concept


One of the more difficult concepts to grasp in the study of dynamics, and kinematics in particular, is
that of determining the derivative of a vector. Although the formal definition for the derivatives of
vectors and scalars are similar, the actual evaluation of these derivatives usually differs considerably.

In vector differentiation one must account for both changes in magnitude and direction. As a result,
the evaluation of a vector derivative depends primarily on the type of components used to describe the
vector.

r
A most instructive method for evaluating the time rate of change of vector A is in terms of
r
components parallel and normal to the initial vector A (commonly referred to as elongation-rotation
r
components), which represent changes in the magnitude and direction of A respectively.

r r
In Figure 1.5(a), consider that the vector A changes to | A + ∆ A in the time interval ∆ t. Noting that
r r r r
vectors A and A + ∆ A define a plane, let ∆θ be the angular change of vector A measured in this
r
plane. Note that in the limit the rotation ∆θ is represented by a vector dθ normal to this plane with its
r r
direction given by the right hand rule. Further, define a set of unit vectors e A and en parallel and
r r r r
normal respectively to A and in the plane defined by A and A + ∆ A .

r r r
eA eA + ∆eA
r
∆A r
r
r r
en + ∆en r ∆eA
en eA r r
eA + ∆eA
r r r ∆θ
A r
A + ∆A en
∆θ r r r
en + ∆en ∆en

(a) (b)

Figure 1.5: A vector (a) shifting, with (b) its unit vectors changing directions.

r r
The derivative of A = A e A becomes

7
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
r& dA d r r r
A= = ( A e A ) = A& eA + A e& A (1.8)
dt dt

However, by definition we may write

r
r& de A [(erA + ∆erA ) − erA ] r
∆e A
eA = = lim = lim (1.9)
dt ∆t → 0 ∆t ∆t →0 ∆t

r r r
Referring to Figure 1.5(b) and noting also by definition that e A = eA + ∆ e A = 1 , we may write

r r
∆eA ≈ eA ∆θ = ∆θ

r r
Furthermore, noting that in the limit ∆ e A will have a direction normal to e A , we may write

r  ∆θ  r r
e&A = lim   en = θ& en (1.10)
∆t →0 ∆t
 

Moreover in Eq. 1.10 we recognize that we may write

r r r
θ& en = ω A × eA

r r
where ω A is the total instantaneous angular velocity of vector A with respect to the x, y, z reference
space. Thus Eq. 1.8 becomes

r& r r r
A = A& eA + ω A × A (1.11)

In Eq. 1.11 we have the change of vector | A conveniently represented in terms of a magnitude change
r r r r
A& eA and a direction change ω A × A . Vector A of course may be represented in any chosen system
r
of coordinates and its derivative taken with respect to any reference space as long as ω A is the total
r
angular velocity of A relative to that reference space.

Several important conclusions may be drawn from Eq. 1.11. First, if a vector undergoes no rotational
motion, its derivative is determined in a manner similar to a scalar, whereby

8
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r& r
A = A& eA (Constant Direction Vector) (1.12)

Second, if a vector has constant magnitude and undergoes only a rotational change, its derivative is
given by

r& r r
A = ωA × A (Constant Magnitude Vector) (1.13)

Note also that if a constant magnitude vector is simply displaced parallel to itself no rate of change
r&
actually occurs and A = 0 .

r
Finally, any unit vector e A , which by definition has constant magnitude, has a time derivative given
by

r r r
e& = ωe × e (Unit Vectors) (1.14)

r r
where ωe is the angular rate at which the unit vector e rotates.

In order that we may further advance these basic notions, we will now illustrate this procedure in the
following examples.

Example 1.1 Second Derivative of a Vector


r
Determine the second derivative of a general vector A in terms of elongation and rotational
components.

r
Solution: The second derivative of A with respect to time according to Eq. 1.11 is

r&
&r& dA d & r
( )
r r
A= = A eA + ω A × A (a)
dt dt

Applying conventional product rule differentiation techniques we obtain

&r& && r r r r r r&


A= A eA + A& e&A + ω& A × A + ω A × A (b)

9
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Utilizing Eqs. 1.11, 1.13, and 1.14,

&r& && r
( )
r r r r r r r r
A= A eA + A& (ω A × eA ) + ω& A × A + ω A × A& ⋅ eA + ω A × A (c)

Collecting terms we may write

&r& && r
( )
r r r
A= A
r r
( r
) r
eA + 2 ω A × A& eA + ω& A × A + ω A × ω A × A (d)

Example 1.2 Angular Velocity and Acceleration of Motor on Platform

A circular platform is spinning relative to the ground about a fixed vertical axis with an instantaneous
angular velocity of Ω = 5 rads which is increasing at a rate of 10 rad
s2
. Mounted on the platform is a

motor rotating at a rate φ& = 3 rads which is increasing at a rate of 6 rad


s2
, both relative to the platform.

At this instant the motor axis is located as shown in the xy plane. What are the angular velocity and
acceleration vectors of the motor with respect to an xyz frame attached to the ground as shown?

z
r
eΩ

O
y

x
r
φ& eφ
45º

Figure 1.6: Motor mounted to platform.

r
Solution: It is clear that the total angular velocity of the motor ω m relative to the ground is given by
the motion of the platform relative to the ground plus the motion of the motor relative to the platform
expressed by

10
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r r r
ωm = Ω eΩ + φ& eφ (a)

r r r r r
For convenience, eΩ and eφ may be expressed in terms of a conventional i , j , k system of unit

vectors attached to the xyz frame. Thus we obtain

3 r 3 r r
r
ωm = i+ j + 5k [ ]
rad
s (b)
2 2

In order to determine the angular acceleration of the motor we can differentiate Eq. (a). Hence, using
product rule differentiation

r & er + Ω er& + φ&& er + φ& er


ω& m = Ω Ω Ω φ φ (c)

r r
Utilizing Eq. 1.14, the derivatives of the unit vectors eΩ and eφ are given by

r r r
e&Ω = ωeΩ × eΩ = 0
r r r r r (d)
e&φ = ωeφ × eφ = Ω eΩ × eφ

r r r r
where ωeΩ and ωeφ are the angular rates at which the eΩ and eφ vectors rotate relative to the xyz

frame respectively. Returning to Eq. (c) we obtain

r & e + φ&& e + φ& (Ω e × e )


ω& m = Ω
r r r r
(e)
Ω φ Ω φ

r r r r r
Finally, expressing the unit vectors eΩ and eφ in terms of the i , j , k system we find

r
ω& m = −
9 r 21 r
2
i+
2
r
j + 10 k [ ]
rad
s2
(f)

11
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

1.4 Derivative of Position Vectors Expressed in Various


Cartesian Coordinate Systems
A significant step in the solution of any kinematics problem is to select the most appropriate type of
coordinates to describe the problem. Four commonly used Cartesian frames of reference are
presented below.

(A) Rectangular Coordinates

The most direct method of evaluating the velocity and acceleration vectors for a particle is through the
use of rectangular coordinates. In this description we may write

r r r r r r r
r = x(t ) i + y (t ) j + z (t ) k = x + y + z (1.15)

r
where each component of r is a vector of variable magnitude with constant direction. Applying Eq.
r
1.12 to each component of r the velocity is given by

r r r r
r dr
v= = x& (t ) i + y& (t ) j + z& (t ) k (1.16)
dt

Similarly, the acceleration is given by

r r r r
r dv
a= = &x&(t ) i + &y&(t ) j + &z&(t ) k (1.17)
dt

It is clear that in order to use a rectangular description one must either know or want to determine x, y,
and z as functions of t.

(B) Cylindrical Coordinates

In cylindrical coordinates the particle is located by a system of coordinates R, θ, z as shown in Figure


1.7. The directions of the coordinates R, θ, z are specified respectively by an orthogonal triad of unit
r r r r r r r
vectors e R , eθ , e z , the permutation of which satisfies e R × eθ = e z . The direction of e z is constant
r r
while e R and eθ undergo directional changes in a plane parallel to the xy plane.

12
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
ez
P
r

r
r
r r eR
z
y
r
R

x
θ

Figure 1.7: Cylindrical coordinates.

The position of point P is then specified by

r r r
r = R eR + z ez (1.18)

Differentiating Eq. 1.18 with respect to time we obtain

r r r r r r
v = r& = R& eR + R e&R + z& ez + z e&z (1.19)

r r r
The angular velocities of the unit vectors e R and eθ with respect to the xyz frame are given by θ& ez
r
while the angular velocity of e z is zero. Thus, from Eq. 1.14

r r r r
e&R = θ& ez × eR = θ& eθ
r r r r
e&θ = θ& ez × eθ = −θ& eR (1.20)
r
e&z = 0

Substitution of the results of Eq. 1.20 into Eq. 1.19 yields

r r r r
v = R& eR + R θ& eθ + z& ez (1.21)

A second differentiation with respect to time yields the acceleration vector

r && r r r r r r
a=R eR + R& e&R + R& θ& eθ + R θ&& eθ + &z& ez + z& e&z (1.22)

13
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Again utilizing Eq. 1.20 and collecting terms, Eq. 1.22 becomes

r &&
a= R ( r
) ( r r
− R θ& 2 eR + R θ&& + 2 R& θ& eθ + &z& ez ) (1.23)

Obviously we can obtain the equations for motion described in polar coordinates by simply taking z ,
z& , and &z& equal to zero in Eqs. 1.18, 1.21, and 1.23 respectively.

(C) Spherical Coordinates


r r
In spherical coordinates, the particle P is located directly by a space position vector r = r er whose
r r r
changes are measured by an orthogonal triad of unit vectors er , eφ , eθ , the permutation of which
r r r
satisfies er × eφ = eθ (see Figure 1.8).

r
P er
r

r
φ r r

y

x
θ

Figure 1.8: Spherical coordinates.

The position vector is written as

r r
r = r er (1.24)

Differentiation of Eq. 1.24 gives the velocity as

r rr r
v = r& er + r e&r (1.25)

14
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Referring to Figure 1.8, if the coordinates θ and φ undergo infinitesimal changes, the angular
velocity corresponding to these infinitesimal changes is

r r r r r
ω = φ& eθ + θ& ez = θ& cos(φ ) eφ + φ& eθ (1.26)

r r r
The angular velocity of the triad of unit vectors er , eφ , eθ is given by Eq. 1.26 and the derivatives of

these unit vectors according to Eq. 1.14 become

r r r r r
e&r = ω × er = φ& eφ + θ& sin (φ ) eφ
r r r r r
e&φ = ω × eφ = −φ& er + θ& cos(φ ) eθ (1.27)
r r r r r
e&θ = ω × eθ = −θ& sin (φ ) er − θ& cos(φ ) eφ

Substitution of the first of Eq. 1.27 into Eq. 1.25 yields the velocity vector

r r r r r
v = r& = r& er + r φ& eφ + r θ& sin (φ ) eθ (1.28)

The acceleration is found by a second differentiation with respect to time giving

r r r r r r r
a = v& = r&& er + r& e&r + r& φ& eφ + r φ&& eφ + r& θ& sin (φ ) eθ
r r r (1.29)
+ r θ&&sin (φ ) eθ + r θ& cos(φ )φ& eθ + r θ& sin (φ ) e&θ

Utilizing Eq. 1.27, Eq. 1.29 becomes

r r
( r
a = v& = r&& − r φ&2 − r θ& 2 sin 2 (φ ) er )
( r
+ r φ&& + 2 r& φ& − r θ& 2 sin (φ ) cos(φ ) e ) φ (1.30)
+ (r θ&& sin (φ ) + 2 r& θ& sin (φ ) + 2 r φ& θ& cos(φ ))e
r
θ

(D) Normal and Tangential Coordinates

In this description, also known as natural or path coordinates, the velocity and acceleration are
described in terms of unit vectors that are tangent and normal to the path of motion. The position
r
vector is simply described as r and the transition to path variables is accomplished by differentiation
r
of r . This system of coordinates is shown in Figure 1.9 with the particle P located by the position

15
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
vector r at time t as it moves along the path. The scalar distance measured along the path from an
origin O’ given by s at time t. Referring to Figure 1.9, the following definitions are made:

r
1. et is a unit vector tangent to the instantaneous path of motion,
r
2. en is a unit vector normal to the instantaneous path of motion directed toward the center of
curvature Q and is termed the principal normal,
r r
3. the plane formed by et and en is referred to as the osculating plane or instantaneous plane of
motion and it contains the instantaneous center of curvature Q,

4. the normal to the osculating plane is termed the binormal and is defined by the unit vector
r r r
eb = et × en .

∆s
r
et
P
r P′
r ∆r
O′ r r
en r r et + ∆et
r en + ∆en
r
ρ
r r
O
r + ∆r
x
∆θ
Q – Center of Curvature

z Osculating Plane

Figure 1.9: Normal and tangential coordinates.

The velocity at time t is defined by


r r
r dr  ∆r ∆s  r
v= = lim   = s& et (1.31)
dt ∆t → 0
 ∆ s ∆t 

r
∆ r& r
since in the limit becomes the unit vector et . Note that s& is simply the magnitude of velocity
∆s
(called speed) as measured along the path by a speedometer for example.

The acceleration is determined by differentiation of Eq. 1.31 which yields

16
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r r r r
a = v& = &s&et + s& e&t (1.32)

r
The derivative e&t is given by

r r r r r r
e& = ωet × et = θ& eb × et = θ& en (1.33)

Utilizing Eq. 1.33, the acceleration is thus written as

r r r
a = &s&et + s&θ& en (1.34)

The first term on the right side of Eq. 1.34, referred to as tangential acceleration, depends only on the
rate of change of speed along the path while the second term, called normal acceleration, depends on
the product of the magnitude of velocity times the angular rate at which the velocity vector rotates
with respect to the xyz frame due to its motion along a curved path. The direction of the latter term is
the same as the direction in which the tip of the velocity vector tends to move due to its rotation,
namely directed toward the center of curvature Q in Figure 1.9.

The normal acceleration can be written in various other ways by noting that according to Figure 1.9,
∆ s ≈ ρ ∆ θ so that in the limiting process

∆s  ∆θ 
s& = lim   = lim   = ρ θ& (1.35)
∆t → 0 ∆ t ∆t → 0 ∆ t 
   

where ρ is the instantaneous radius of curvature of the path at time t. Accordingly, the acceleration
given by Eq. 1.34 can also be written as

r r r r s& 2 r
a = &s&et + ρ θ& 2 en = &s&et + en (1.36)
ρ

Recalling some well known results of analytic geometry, we can further describe the geometrical
nature of this problem by noting that in terms of direction cosines

r dx r dy r dz r
et = i + j+ k (1.37)
ds ds ds

17
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

In addition, the curvature of the path is given by

2 2 2
 d 2x   d 2 y   d 2z 
κ =  2  +  2  +  2  (1.38)
 ds   ds   ds 

with the radius of curvature ρ defined by

1
ρ= (1.39)
κ

It is left as an exercise to show that for two dimensional motion

x& &y& − &x& y&


κ= (1.40)
(x& 2
+ y& 2 ) 3
2

where x = x(τ ) , y = y (τ ) , and x& =


dx
. Typically, if τ = x we obtain a more familiar two-

dimensional expression for curvature given by

d2y
κ= dx 2 (1.41)
3
  dy  2 
2

1 +   
  dx  

The distance s measured along the path can be determined by integrating along the path from O’ as
follows

2 2 2
 d 2x   d 2 y   d 2z 
s = ∫ ds = ∫  2  +  2  +  2  dτ (1.42)
 dτ   dτ   dτ 

Thus, for example, if τ = x , we have

2 2
 d 2 y   d 2z 
s=∫ 1 +  2  +  2  dx (1.43)
 dx   dx 

18
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r r r
Finally, the rates of change of the unit vectors et , en , eb as functions of time or position along the
path can be readily determined by considering, for example, that

r r r r r r
e&t = ωet × et = θ& eb × et = θ& en (1.44)

Therefore, referring to Eq. 1.35, we may write

r r
det det dt θ& r 1r r
= = en = en = κ en (1.45)
ds dt ds s& ρ

Similarly, we may immediately show that

r r r r
( r r
)
r r
e&n = ωe n × en = β& et + θ& eb × et = −θ& et + β& eb (1.46)

from which

r
r β& r
ds s&
(
den 1 & r & r
s&
) r r
= − θ et + β eb = −κ et + eb = −κ et + kτ eb (1.47)

In some applications of path variables, it is desirable to obtain the rate of rotation of the osculating
plane as a function of position along the path. This rotation is defined by the torsion of the curve kτ ,
which has already been used in Eq. 1.47. As in the previous work, it can then be shown that

r
deb r
= −kτ en
ds (1.48)
r& r
eb = − s& kτ en

1.5 Examples of Particle Motion

Example 1.3 Radius of Curvature

The position vector of a particle is given by

r  t5 r r t6 r
r =  + 2 t  i + 3 t 2 j − k [ ft ] (a)
5  3

19
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Evaluate the radius of curvature of the path at t = 1 second .

Solution: The velocity and acceleration are found by differentiation to be

r
r r
( r
) r
v = r& = t 4 + 2 i + 6 t j − 2 t 5 k (b)

r r r r r
a = v& = 4 t 3 i + 6 j − 10 t 4 k (c)

which for t = 1 second become

[]
r r r r
v = 3i + 6 j − 2 k ft
s (d)

[]
r r r r
a = 4 i + 6 j − 10 k ft
s2
(e)

From Eq. (d), a unit vector tangent to the path is given by

r r r r
r v 3i + 6 j − 2 k r r r
et = = = 0.428 i + 0.856 j − 0.286 k (f)
s& 32 + 6 2 + 2 2

Thus, the component of acceleration tangent to the path is

[]
r r r r r r r
at = (a ⋅ et )et = 4.15 i + 8.31 j − 2.78 k ft
s2
(g)

Accordingly,

[]
r r r r r r
a n = a − at = −0.150 i − 2.31 j − 7.22 k ft
s2
(h)

Considering normal and tangential coordinates,

r s& 2
an = (i)
ρ

so that the radius of curvature is given by

20
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

s& 2 32 + 6 2 + 2 2
ρ= r = = 6.46 [ ft ] (j)
an 0.150 2 + 2.312 + 7.22 2

Example 1.4 Circular to Rectilinear Motion Mechanism

The mechanism shown in Figure 1.10, which is used for converting circular motion into rectilinear
motion, is driven by a pin P moving in a circular path at a constant angular velocity ω . Determine the
position, velocity, and acceleration of point Q considering that θ = 0 at t = 0 .

Tool Linkage
ω P

r
O

θ Q x

Figure 1.10: Mechanism to convert circular motion to rectilinear.

Solution: The position of Q measured in the xy coordinate system is governed by the rotating vector
r
r and is given by

r r r
rQ = (r cos(θ ) + l )i = (r cos(ω t ) + l )i (a)

Differentiating Eq. (a) we obtain

r r
vQ = −r ω sin (ω t )i (b)

and

r r
aQ = −r ω 2 cos(ω t )i (c)

21
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Thus, we see that the position, velocity and acceleration of Q are all harmonic with the period


T= [sec] (d)
ω

where ω is called the circular frequency. This term is used extensively in dynamics for describing all
r
types of oscillatory motion and its relationship to a rotating vector r is readily apparent through this
example.

The amplitudes of motion are all constant for this class of motion and are given by r, r ω , and r ω 2
for position, velocity, and acceleration respectively. Note that the amplitudes of velocity and
acceleration are very much dependent on the circular frequency ω as well as the radius r.

Furthermore, since l is constant in Eq. (a), the actual motion of point Q from its initial position in
space is given by

x = r cos(ω t )
 π
x& = −r ω sin (ω t ) = r ω cos ω t +  (e)
 2
&x& = −r ω cos(ω t ) = r ω cos(ω t + π )
2 2

It is observed that for the case of rectilinear harmonic motion of a point, the velocity leads the
displacement by 90 degrees, and the acceleration leads the displacement by 180 degrees and the
velocity by 90 degrees. Thus, the effect of differentiating a harmonic function is to advance the phase
90 degrees and multiply the amplitude by the circular frequency.

Substituting from the first of Eqs. (e) into the third we may write

&x& + ω 2 x = 0 (f)

Equation (f) is the general form of the differential equation that describes rectilinear harmonic motion
of a point, where we note that the coefficient of the dependent variable is the square of the circular
frequency.

22
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Example 1.5 Oscillating Particle in Groove

A particle P is forced to oscillate sinusoidally along a radial groove in a rotating platform with an
amplitude A and circular frequency β about point O as shown in Figure 1.11. Considering that the
rad
angular velocity of the platform increases linearly at the rate of K s in the counterclockwise
direction, determine:

1. the position, velocity, and acceleration of the particle with respect to the inertial xy frame using
polar coordinates and

π
2. the radius of curvature of the particle’s inertial path at the instant t = .

θ& r
eθ r
eR
R
θ
P
O

Figure 1.11: Particle oscillating in radial groove.

Solution:

1. The position of P in terms of polar coordinates (R, θ ) is given by

R = A sin (β t )
1 (a)
θ = K t2
2

where t = 0 , θ = R = 0 .

23
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r r r
The velocity of P in polar coordinates according to Eq. 1.21 for z = 0 is v P = R& eR + R θ& eθ where

R& = A β cos(β t )
θ& = K t

r r
Therefore, in polar coordinates where the directions of eR and eθ are orientated with respect to the

1 2
inertial xy axes by the angle θ = K t at any time t, we have
2

r r r
v P = A β cos(β t )eR + A K t sin (β t )eθ (b)

Similarly, according to Eq. 1.23 for which

&& = − A β 2 sin (β t )
R
θ&& = K

we have

a P = −(β 2 + K 2 t 2 ) A sin (β t )eR + K A (sin (β t ) + 2 β t cos(β t ))eθ


r r r
(c)

π
2. At the time t = , cos(β t ) = 0 and sin (β t ) = 1 . Therefore, from Eqs. (b) and (c) above,

r π AK r
vP = eθ

(d)
r  2 π2 K2 r r
a P = − A  β +  e + K A eθ
2  R
 4β 

r
Since the radial component of velocity is zero at this instant, the tangent to the path is the eθ direction

π2 K
given by the angle θ = .
8β 2

r r r r
Thus, by making use of path coordinate concepts, namely Eq. 1.36, where et = eθ and en = −eR for
this instant, we may write the normal acceleration an as

24
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

s& 2  π2 K2 
an = = A ⋅  β 2 +  (e)
ρ  4 β 2 

π AK π
where s& = . Solving Eqs. (e) for ρ , we obtain for the instant t =
2β 2β

A
ρ= (f)
4β 4
1+ 2 2
π K

r
The center of curvature is located in the negative eR direction.

Example 1.6 Slotted Guide

A mechanism consisting of a slotted guide in a circular arc and a yoke arm are constrained to move
together by a cylindrical pin of negligible size. At the instant shown in Figure 1.12 the arm is moving
r
with a constant velocity v A to the left. The magnitude of the acceleration of the pin for the position
ft
shown is 36 s2
. Determine the velocity of the pin at this instant.

Solution: The position of the pin in rectangular coordinates is given by

x = r cos(θ )
(a)
y = r sin (θ )

Differentiating Eq. (a) for the velocity and acceleration, we have

r r r
v = −r sin (θ )θ& i + r cos(θ )θ& j (b)

r
( r
) ( r
a = − r sin (θ )θ&& − r cos(θ )θ& 2 i + r cos(θ )θ&& − r sin (θ )θ& 2 j ) (c)

From the statement of the problem however, vx is constant. Thus from Eq. (c)

− r sin (θ )θ&& − r cos(θ )θ& 2 = 0 (d)

25
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
r = 2 ft. Pin
varm

O
θ = 30° x

Figure 1.12: Slotted guide in circular arc with yoke arm.

r
In addition a = 36 , which yields from Eq. (c)

r cos(θ )θ&& − r sin (θ )θ& 2 = 36 (e)

Solution of Eqs. (d) and (e) for θ = 30° yields

r r
θ& = 3 k [rads ]
(f)
[ ]
r r
θ&& = −9 3 k rad
s2

for which

r r r
v = −3 i + 3 3 j []
ft
s (g)

A more direct solution can be realized by noting that since the arm has a constant velocity, the
resultant acceleration of the pin must be vertical. Further, since the normal component of acceleration
must be directed toward the center of curvature, one may construct Figure 1.13.

26
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

30°
r
an

r
a
r
at

Figure 1.13: Acceleration vectors for pin

for which

r
an = 36 cos(60°)
= 18 []ft
s2
(h)

Therefore, since

r s& 2
an = (i)
ρ

we have that

s& = 6 (j)

from which

 1r 3 r
r
v = 6  − i +
r r
j  = −3 i + 3 3 j []
ft
s (k)
 2 2 

Example 1.7 Helical Path

A particle P shown in Figure 1.14 moves along a helical path described by the equations

27
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

x = a cos(Ω t )
y = a sin (Ω t ) (a)
z = bt

r
r

Figure 1.14: Helical path of particle P.

Determine the velocity and acceleration of the particle P.

Solution:

1. Rectangular Coordinates

For rectangular coordinates we may write directly from Eq. (a) that

r r r r
r = a cos(Ω t ) i + a sin (Ω t ) j + b t k (b)

r r r r r
v = r& = − a Ω sin (Ω t ) i + a Ω cos(Ω t ) j + b k (c)

r r r r
a = v& = − a Ω 2 cos(Ω t ) i − a Ω 2 sin (Ω t ) j (d)

from which

r
v = x& 2 + y& 2 + z& 2 = a 2 Ω 2 + b 2 (e)

28
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
a = &x&2 + &y&2 + &z&2 = a Ω 2 (f)

2. Cylindrical Coordinates

From Eq. (a) we have according to cylindrical coordinates that

R = a ; R& = R && = 0
θ = Ω t ; θ& = Ω ; θ&& = 0 (g)
z = b t ; z& = b ; &z& = 0

from which the velocity and acceleration are given by

r r& r r r r r
v = R eR + R θ& eθ + z& ez = a Ω eθ + b e z (h)

r &&
a= R ( r
) ( r r
) r
− r θ& 2 eR + 2 R& θ& + R θ&& eθ + &z& ez = − a Ω 2 eR (i)

The magnitudes of these quantities are again

r
v = a2 Ω2 + b2 (j)

r
a = a Ω2 (k)

3. Normal and Tangential Coordinates

The velocity is given in path coordinates by

r r
v = s& et (l)

Directly by use of Eqs. (a) we obtain

2 2 2
ds  dx   dy   dz 
s& = =   +   +   = a2 Ω2 + b2
dt  dt   dt   dt  (m)
d 2s
&s& = 2 = 0
dt

29
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
The unit vector et is defined by

r r r r
r v − a Ω sin (Ω t )i + a Ω cos(Ω t ) j + b k
e= = (n)
s& a2 Ω + b2

The acceleration is of the form

r r s& 2 r
a = s et + en
&& (o)
ρ

where from Eqs. 1.33 and 1.35

r s& r
e&t = en (p)
ρ

Therefore, differentiating Eq. (n) with respect to time we have

r r
− a Ω 2 cos(Ω t )i − a Ω 2 sin (Ω t ) j s& r
= en (q)
a2 Ω + b2 ρ

Considering only the magnitude of both sides of Eq. (q) we obtain

s& a Ω2
= (r)
ρ a2 Ω + b2

Moreover, from a consideration of Eqs. (q) and (r)

r r r
en = − cos(Ω t )i − sin (Ω t ) j (s)

Substitution of these computed values into Eqs. (l) and (o) yields

r r r r
v = − a Ω sin (Ω t )i + a Ω cos(Ω t ) j + b k
r r r (t)
a = − a Ω 2 cos(Ω t )i − a Ω 2 sin (Ω t ) j

r r
Thus, the magnitude of v and a agree once again with the results obtained by the use of the previous
two coordinate systems.

30
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

4. Spherical Coordinates

The velocity in spherical coordinates is given by

r r r r
v = r& er + r φ& eφ + r θ& sin (φ )eθ
(u)

Now

r
r = x2 + y 2 + z 2 = a2 + b2 t 2
(v)

and

tan (φ ) =
a
bt (w)

from which

sin (φ ) = cos(φ ) =
a bt
;
a +b t
2 2 2
a + b2 t 2
2
(a-a)

From Eq. (w), by differentiation we obtain

ab 1 ab 1 ab
φ& = − =− 2 2 =− 2
b t sec (φ )
2 2 2
b t 1 + tan (φ )
2
(
a + b2 t 2 ) (b-b)

Further

r r r r
er = sin (φ )cos(Ω t )i + sin (φ )sin (Ω t ) j + cos(φ ) k
r r r r
eφ = cos(φ )cos(Ω t )i + cos(φ )sin (Ω t ) j − sin (φ ) k
r r r
eθ = − sin (Ω t )i + cos(Ω t ) j (c-c)

Substitution of these results into Eq. (u) yields

31
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r 2b2 t  r r r
 cos(Ω t )i + sin (Ω t ) j + k
a a bt
v=
2 a 2 + b 2 t 2  a 2 + b 2 t 2 a2 + b2 t 2 a 2 + b 2 t 2 

− a b  r r r
+ a2 + b2 t 2 2

bt
cos (Ω t ) i +
bt
sin (Ω t ) j −
a
k 
a +b t a + b t 
2 2
 a +b t a +b t
2 2 2 2 2 2 2 2 2

aΩ
+ a2 + b2 t 2 ( r r
− sin (Ω t )i + cos(Ω t ) j )
a2 + b2 t 2

which upon simplification yields

r r r r
v = − a Ω sin (Ω t )i + a Ω cos(Ω t ) j + b k (d-d)

In a similar manner, one may show that the acceleration in spherical coordinates (Eq. 1.29) is

r r r
a = − a Ω 2 cos(Ω t )i − a Ω 2 sin(Ω t ) j (e-e)

r r
These results for v and a of course agree with the prior results. It should be very clear from this
example that a wise selection of coordinates in a given problem will greatly expedite a solution.

1.6 Derivative of a General Vector that is Referred to a


Rotating Coordinate System
r
In the previous sections we dealt with a vector A that was expressed in terms of various sets of
r
components referred to the xyz coordinate system and discussed the time derivative of A with respect
to that reference frame. If we now consider that the xyz frame is in motion relative to another frame,
r
denoted by XYZ, the change in the orientation of A with respect to the XYZ frame depends not only
on its time rate of change relative to the xyz axes, but also on its time rate of change due to the motion
of the xyz frame relative to the XYZ coordinate system.

Applications of kinematics for which it is desirable to talk about relative motion and rotating
coordinate systems are numerous, such as, the motion of an airplane taking off relative to an aircraft
carrier deck that is rolling, pitching, and yawing in the sea, for example. Also, we will see later that in

32
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

describing events relative to the surface of the Earth, we occasionally must include consideration of
the fact that the Earth is experiencing rotational motion in space.

For initial convenience consider that the origins of the XYZ and xyz coordinate systems are
coincident as indicated in Figure 1.15. Hence, only rotational relative motion occurs between the xyz
and XYZ reference frames.

r
z A

O
Y

r
ω
X x

Figure 1.15: Vectors relative to rotating reference frame.

r
Vector A can be expressed in any suitable system of components (i.e., rectangular, polar, path
r
coordinates, etc.) measured in the xyz system with ω representing the angular velocity of the xyz
r
frame relative to the XYZ frame. Further, let the total angular velocity of vector A relative to the
XYZ coordinate system be written as

r r r
ω A = ω + ωr (1.49)

r r
where ωr is the angular velocity of A relative to xyz. Then from Eq. 1.11

r& r r r
AR = A& eA + ω A × A
r r r r r (1.50)
= A& e + ω × A + ω × A
A r

r r&
where the derivative of A with respect to the XYZ frame is denoted by AR .

33
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
Now since A is measured in components in the xyz frame, the first two terms on the right hand side of
r r&
Eq. 1.50 simply represent the derivative of A with respect to the xyz frame, called AR , such as
r
discussed in Sections 1.3 and 1.4. This may also be further demonstrated by letting ω = 0 .

Thus, define

r& r r r
AR = A& e A + ω r × A (1.51)

from which Eq. 1.50 may be written as

r& r& r r
AR = Ar + ω × A (1.52)

In order to further understand this concept it is instructive at this point to apply Eq. 1.52 to the special,
r
but very important case of rectangular coordinates. Consider that the vector A is expressed in
r
rectangular components along the xyz axes which rotate with the angular velocity ω with respect to
r
the XYZ frame. The ω vector can also be resolved into components along the rotating xyz axes, as
shown in Figure 1.16, since any vector can be expressed in terms of components along any set of axes.

Y y
r
ωy ω r
Ay A
x
Ax
ωx
r r
j i
r r
R O k
O′ X ωz
Az
z

Figure 1.16: Vectors in relative coordinate system.

Similar to the discussion in Section 1.4 (A), we may in general write

r r r r r r r
A = Ax i + Ay j + Az k = Ax + Ay + Az (1.53)

34
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
Considering that A is the sum of three components that are each vectors in their own right, we can
r
first determine the time rate of change of A relative to the xyz rotating frame by applying Eq. 1.51 to
each vector component yielding

r& r r r
Ar = A& x i + A& y j + A& z k (1.54)

r
since each vector component is nonrotating (i.e., ωr = 0 ) relative to the xyz axes. This fact is the one
r r r
that characteristically makes rectangular coordinates easy to apply, since the unit vectors i , j , and k
have fixed directions relative to the xyz frame.

r
However, since the xyz frame is rotating relative to the XYZ frame, the vector A and each of its
rectangular vector components rotates relative to this frame resulting in a further time rate of change
r r r
given by ω × A . Thus, according to Eq. 1.52, the total derivative of A with respect to the XYZ frame
is given by

r& r r r r r
AR = A& x i + A& y j + A& z k + ω × A (1.55)

r
The developments are similar if the vector A is expressed in any other set of components (i.e.,
cylindrical, path coordinates, etc.) referred to the xyz frame. In particular, if the vector under
consideration is a position vector, the velocities and accelerations presented for the various types of
coordinate systems in Section 1.4 now become the velocities and accelerations relative to the rotating
xyz frame of reference. This is a very important concept to thoroughly understand. Referring back to
r
Eq. 1.52, the following conclusions can also be drawn. First, the derivative of a vector A depends on
r
the frame of reference from which the changes in A are observed, so that even if a vector is fixed in a
rotating xyz frame, it has a derivative with respect to an XYZ frame as given by

r& r r
AR = ω × A (1.56)

r
where ω is the relative angular velocity between the two frames. Equation 1.56 is really just a
r r
restatement of Eq. 1.13, since the angular rate of vector A , called ω A in Eq. 1.13, is simply the
r&
angular velocity of the coordinate system if Ar is zero.

35
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r r
Finally, a very important fact may be noted from Eq. 1.52 if we consider the special case of A = ω for
which we have

r r
ω& R = ω& r (1.57)

Equation 1.57 shows that for the angular velocity vector one may find its derivative in either of two
r
ways. One may either express ω in components referred to a rotating frame of reference and simply
r r
differentiate the scalar components as indicated by ω& R or one may differentiate ω& R in accord with Eq.
1.11. The former method is often most easily applied.

For example, if the angular velocity of the rotating xyz coordinate system is expressed in rectangular
coordinates as shown in Figure 1.16, we have

r r r r
ω = ωx i + ω y j + ωz k (1.58)

Differentiating according to Eq. 1.55 we see immediately that the angular acceleration of the rotating
coordinate system is given by

r r r r r
ω& R = ω& x i + ω& y j + ω& z k = ω& r

or simply

r r r r
ω& = ω& x i + ω& y j + ω& z k (1.59)

In order to further clarify these concepts it is instructive to reconsider Example 1.2 at this point.

Example 1.8 Angular Velocity and Acceleration of Motor on Platform 2

Solve Example 1.2 by the methods of Section 1.6.

Solution:

1. Application of Eq. 1.52

36
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Referring to the statement of Example 1.2 along with Figure 1.16, let the xyz body axes attached to the
platform be instantaneously aligned with the XYZ inertial axis as schematically shown in Figure 1.17.

Z z
r

r
k r Y
O j
r
i Motor
y

r&
x
φ
X 45º

Figure 1.17: Motor mounted to platform

The total angular velocity of the motor is given as before by

r r r 1 &r 1 & r r
ω m = Ω k + φ& eφ = φi + φ j + Ωk (a)
2 2

r
Equation (a) expresses the vector ω m in terms of rectangular components referred to the rotating xyz
body axes. Thus, the special form of Eq. 1.52 for rectangular coordinates given by Eq. 1.55 in the
preceding discussion can be directly applied.

r
Since the angular velocity of the xyz coordinate system to which ω m is referred is given by

r r
ω = Ωk (b)

r r
we obtain from Eq. 1.55 for A = ωm

r 1 && r 1 && r & r r r


ω& m = φi + φ j + Ω k + Ω k × ωm
2 2

or

37
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r  1 && 1 &  r  1 && 1 &  r & r


ω& m =  φ− φ Ω i +  φ+ φ Ω j + Ω k (c)
 2 2   2 2 

Substituting the numerical values from Example 1.2 corresponding to φ& = 3 rads , φ&& = 6 rad
s2
, Ω=5 rad
s ,
& = 10 rad2 , we obtain
and Ω s

r 3 r 3 r r
ωm = i+ j + 5k [rads ] (d)
2 2

[ ]
r 9 r 21 r r
ω& m = − i+ j + 10 k rad
s2
(e)
2 2

which agrees with Eqs. (b) and (f) of Example 1.2.

2. Application of Eq. 1.57

In order to apply Eq. 1.57, which is written for rectangular coordinates in Eq. 1.59, we must realize
r r
from Eq. 1.52 that A = ω implies that we are seeking the derivative of the angular velocity of the
rotating xyz coordinate system itself. Bearing in mind that we are interested in determining the
angular acceleration of the motor with respect to the XYZ reference system, we must therefore attach
our rotating coordinate system xyz directly to the motor in applying Eq. 1.59.

Further, since we are interested in obtaining the angular acceleration at a specific instant of time, it is
then necessary to express the angular velocity of the motor as a general function of time in order to
evaluate its derivative. Referring to Figure 1.17 and Figure 1.18, the angle β represents the rotation
of the motor due to the platform rotation in the XY plane and the angle φ is the angular motion of the
motor about its own axis, denoted here by the x axis. Thus, the inertial directions X, Y, Z are denoted
r r r
by i ′ , j ′ , and k ′ respectively for this solution and the coordinate axes xyz attached directly to the
r r r
rotor are denoted here by i , j , k respectively as shown schematically in Figure 1.18.

38
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

z φ
r r
Ω = β& k ′
y

r
r k r φ
k′ j′
rO r β Y
i r′ j
i
r&
φ
X
β x

Figure 1.18: Rotation of relative coordinate system for motor

We may then write

r r r r r
ω R = ω r = φ& i + β& sin (φ ) j + β& cos(φ ) k (f)

and hence through differentiation we obtain

r r r r r r r
ω& R = ω& r = φ&& i + β&& sin (φ ) j + φ& β& cos(φ ) j + β&& cos(φ ) k − φ& β& sin (φ ) k (g)

Thus, for the special orientation for which φ = 0 , we obtain

r r r r r
ω& R = ω& r = φ&& i + φ& β& j + β&& k (h)

Further, for β = 45° , we have from geometry

r 1 r 1 r r 1 r 1 r r r
i = i′+ j′ ; j =− i′+ j′ ; k = k ′ (i)
2 2 2 2

from which

[ ]
r 9 r 21 r r
ω& r = − i′+ j ′ + 10 k ′ rad
s2
(j)
2 2

which, of course, agrees with our prior results from Solution 1.

39
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Example 1.9 Particle Relative to Rotating Platform

A particle P shown in Figure 1.19 moves with a constant speed v around the edge of a circular
platform, where v is measured relative to the platform. In addition, the platform rotates at a constant
angular velocity θ& as shown. Determine the velocity and acceleration of P with respect to the surface
of the Earth represented by the X, Y, Z axes by direct application of Eqs. 1.51 and 1.52.

Y
y
r
r v
j
r φ& r &
eφ eρ θ
r
P φ x
ρ r
O i θ

Figure 1.19: Particle on edge of circular platform.

Solution:

1. Velocity Solution

Let the xyz coordinate axes be attached to the platform as shown in Figure 1.19 and consider first that
r r
A = ρ in Eqs. 1.51 and 1.52 given by

r& r& r r
AR = Ar + ω × A (1.52)

where

r& r r r
AR = A& e A + ω r × A (1.51)

The following terms may be defined and identified in these equations:

40
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r& r r r
• AR = ρ& R = v = time derivative of ρ relative to the XYZ frame

r& r r r
• Ar = ρ& r = vr = time derivative of ρ relative to the xyz frame

r r r
• A = ρ = ρ eρ = position of P expressed by components in the xyz frame

r r
• ω = θ& k = angular velocity of the xyz frame relative to XYZ

r& r r
• A = 0 = ρ& = rate of change of magnitude of ρ referred to the xyz frame

r r r
• e A = eρ = unit vector in direction of ρ

r r r
• ω r = φ& k = angular velocity of ρ vector relative to xyz frame

Therefore, from Eq. 1.51

r& r r r
Ar = φ& k × ρ eρ = φ& ρ eφ (a)

and from Eq. 1.52

r& r r r r
AR = v = φ& ρ eφ + θ& k × ρ eρ (b)

r
(r
∴ v = ρ ⋅ φ& + θ& eφ ) (c)

This is as expected if one notes that P is simply moving in a circular path relative to the XYZ frame at
a total angular rate of φ& + θ& . ( )
2. Acceleration Solution

r r
( )
The velocity v = ρ ⋅ θ& + φ& eφ is given here in terms of components in the xyz frame, so again
r r
referring to Eqs. 1.51 and 1.52, let A = v . Consequently, we have for this case:

r& r r r
• AR = v&R = a = time derivative of v relative to the XYZ frame

41
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r& r r r
• Ar = v&r = ar = time derivative of v relative to the xyz frame

r r
• (
r
)
A = v = ρ& ⋅ θ& + φ& eφ = total velocity of P expressed by components by components in the xyz

frame

r r
• ω = θ& k (same as before)

r& r
• ( ) r
A = v& = ρ ⋅ θ&& + φ&& = 0 = rate of change of magnitude of v referred to xyz axes

r r r
• e A = eφ = unit vector in direction of v

r r r
• ω r = φ& k = angular velocity of v vector relative to the xyz frame

Referring to Eq. 1.52, we may write

r
( ) ( )
r& r r
Ar = φ& k × ρ ⋅ θ& + φ& eφ = − ρ φ& ⋅ θ& + φ& eρ (d)

and from Eq. 1.51

r
( ) ( )
r& r r r
AR = a = − ρ φ& ⋅ θ& + φ& eρ + θ& k × ρ ⋅ θ& + φ& eφ (e)

r
( 2 r
a = − ρ ⋅ θ& + φ& eρ ) (f)

Again, this result can be checked by physical logic, since particle P moves in a circular path of radius
ρ at a constant total angular rate of (θ& + φ&) relative to the XYZ frame yielding a normal acceleration

( )
of ρ ⋅ θ& + φ& directed toward the center of curvature.
2

42
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

1.7 Motion of a Particle Relative to a Moving Frame of


Reference
As indicated in the previous section, it is frequently convenient to evaluate the derivative of a vector
r
A in terms of components measured in a moving frame of reference, which in many instances may
also be translating in addition to rotating. This is particularly true of position vectors whenever the use
of moving reference frames leads to mathematical simplifications or helps produce more meaningful
physical results. To clarify this point further, consider the motion of a piston relative to a moving
vehicle. In order to determine the appropriate loadings imposed by the piston on the cylinder walls,
one needs to know its total acceleration with respect to an inertial frame. However, to meaningfully
design such a system, one needs to know these loadings described in a reference frame attached to the
vehicle. Thus, both the kinematical analysis and design problem are simplified by the introduction of
a moving frame of reference, since the motion of the piston relative to this frame is rectilinear and thus
easy to describe.

Consider the system shown in Figure 1.20 where the XYZ frame constitutes a primary frame of
reference which is not necessarily an inertial frame. The origin of the xyz system moves with a linear
r& r
velocity R and the xyz frame has a rotational velocity of ω , both relative to the XYZ system. Further,
r
the point P, whose motion is to be described, is located relative to the xyz frame by ρ while the origin
r
O of the xyz frame is located by R relative to O’ of the XYZ frame.

From Figure 1.20, we may write

r r r
r = R+ρ (1.60)

The velocity of P with respect to XYZ is then

r r r& r
v = r& = R + ρ& (1.61)

r
Now since ρ is measured in the xyz system, from Eq. 1.52 we have

r r r r
ρ& = ρ& r + ω × ρ (1.62)

43
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

and Eq. 1.61 becomes

r r& r r r
v = R + ω × ρ + ρ& r (1.63)

Y y

x
C
r r
r ρ
r
R O r
ω
O′ X

r& z
R
Z

Figure 1.20: Motion of a particle relative to a moving frame of reference.

In Eq. 1.63, we note the following:

r&
1. R represents the velocity of the origin O of the xyz frame with respect to the XYZ frame,

r
2. ρ& r represents the velocity of point P relative to the xyz system as viewed by an observer
attached to the xyz frame, and

r& r r
3. R + ω × ρ represents the velocity with respect to XYZ of a point attached to the xyz system
that is instantaneously coincident with P. In other words, this coincident point is the point of the xyz
reference frame past which the particle is instantaneously moving.

The acceleration is found by a second differentiation and is given by

r r &r& d r&
a = v& = R +
dt
( )
r r r r
ρ r + ω& × ρ + ω × ρ& (1.64)

Again, by Eq. 1.52

44
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

d r&
dt
( ) r r r
ρ r = ρ&&r + ω × ρ& r (1.65)

Substituting Eqs. 1.62 and 1.65 into Eq. 1.64, we find that

r &r& r r r r r r r r
a = R + ω& × ρ + ω × (ω × ρ ) + ρ&&r + 2 ω × ρ& r (1.66)

In Eq. 1.66, we may further identify the various terms as:

&r&
1. R represents the acceleration of the origin O of the xyz frame with respect to the XYZ frame,

r
2. ρ&&r represents the acceleration of point P relative to the xyz system as viewed by an observer
moving with the xyz frame,

&r& r r r r r
3. R + ω& × ρ + ω × (ω × ρ ) represents the acceleration with respect to XYZ of a point attached to
the xyz system and coincident with P at any given time t. This may be readily noted by taking the
r r r r r r
relative motion terms ρ& r and ρ&&r equal to zero. Further, ω& × ρ and ω × (ω × ρ ) are the instantaneous
tangential and normal accelerations of the coincident point of the xyz coordinate system as it moves in
r
its spherical path about O at a radius ρ , and

r r
4. 2 ω × ρ& r is referred to as the Coriolis acceleration in honor of G. Coriolis, the French
engineer, who in 1829 first pointed it out. The Coriolis acceleration, as may be noted from Eq. 1.64,
r r r
arises from two effects; (1) the change of the direction of ρ& r due to ω and (2) the effect of ω on the
r
change in magnitude of ρ relative to xyz.

The Coriolis acceleration, in addition to obviously representing an inertial term, produces a variety of
interesting effects in many different types of problems, ranging for example from ballistics to
oceanography. Some of these will be examined later.

In applying the rotating coordinate system and relative motion concepts presented in this section, it is
important to keep in mind that coordinate systems are really a tool devised by analysts to aid in
describing, understanding, and solving physical problems. Consequently, when one chooses an

45
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

appropriate rotating coordinate system for a particular application, one must keep in mind that it
should be an intermediate coordinate system that is selected according to the following criteria:

1. The motion of its origin should be relatively easy to describe,

2. Its angular velocity and angular acceleration should be relatively easy to describe in terms of
components referred to the rotating coordinate system, and

3. The motion relative to the rotating frame should be considerably simpler to describe than the
total motion. It is very important to understand that the “relative problem” often becomes a problem
much like those discussed in Sections 1.4 and 1.5 where we were determining velocities and
accelerations relative to the xyz frame that could have been rotating. Thus, the process of using a
r r
rotating coordinate system and thinking in terms of a “relative problem” for ρ& r and ρ&&r becomes
primarily a matter of choosing a rotating frame that has a “relative problem” that one is able to readily
r r
identify and describe. The associated relative motion vectors ρ& r and ρ&&r are the values used in the
general Eqs. 1.63 and 1.66.

1.8 Examples of Rotating Coordinate Systems and Relative


Motion

Example 1.10 Particle on Platform Using Rotating Coordinate System

In order to compare methods, solve Example 1.9 by considering the motion of particle P relative to a
rotating coordinate system using Eqs. 1.63 and 1.66.

Solution: Choose a rotating coordinate system fixed to the rotating platform with origin at 0. Then,

r r
ω = θ& k
r r r (a)
ω& = 0 ; R& = 0 ; R&& = 0

46
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Y
y
r
r v
j
r φ& r
eφ eρ θ&
r
P φ x
ρ r
O i θ

Figure 1.21: Particle on edge of circular platform.

r
Also, the vector ρ drawn from the origin of the rotating coordinates to the instantaneous position of P
is given by

r r
ρ = ρ eρ (b)

The “relative problem” is very simple to describe for this example if one forgets all else and first
considers the motion of P relative to the xyz coordinate system. It is easy to see that P is simply
moving at constant speed around a circular path relative to the platform. Thus, considering only the
simple case of circular motion relative to the xyz frame similar to the discussions of Sections 1.4 and
1.5, we see that

r r r
ρ& r = v eφ = ρ φ& eφ
r& v2 r r (c)
ρ r = − eρ = − ρ φ& 2 eρ
ρ

Thus, the total velocity of P becomes

r r& r r r
v P = R + ω × ρ + ρ& r
r r r
= 0 + θ& k × ρ e + ρ φ& e
ρ φ

or

47
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
( r
v P = ρ ⋅ θ& + φ& eφ) (d)

Similarly,

r &r& r r r r r r r r
a P = RO + ω × (ω × ρ ) + ω& × ρ + ρ&&r + 2 ω × ρ& r
r r
( r
) r r
= 0 + θ& k × θ& ρ eφ + 0 − ρ φ& 2 eρ + 2θ& k × ρ φ& eφ

or

r
( 2 r
a P = − ρ ⋅ θ& + φ& eρ ) (e)

It is interesting to compare this method with Example 1.9, noting that the terms come from somewhat
different computations. The important difference is that usually it is much easier to apply the methods
of Section 1.7 because the problem has essentially been separated into a consideration of a “relative
problem” and the motion of the associated coordinate system.

Example 1.11 Pendulum on Circular Platform

The circular platform of radius a shown in Figure 1.22 rotates about bearings A and B at a constant
r
angular velocity ω . A simple pendulum is hinged at point Q and constrained to rotate in the yz plane
where the axes xyz are platform body axes. Determine the velocity and acceleration of particle m for
a general position θ .

48
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Z
r
ω

A z′
r
k
Q r
O a j
θ&,θ&& Y , y′
l
θ
r x′ r
X i B eθ
m P
r
eR
Figure 1.22: Rotating platform with pendulum.

Solution (A): Consider a rotating xyz coordinate system with origin at O fixed to the platform. Thus,
the relative problem is simply a circular motion of m about Q and we have

r r& r r r
v P = R + ω × ρ + ρ& r

where

r&
R=0
r r r
[ r r
] r
ω × ρ = ω k × (a + l sin (θ )) j − l cos(θ ) k = −ω ⋅ (a + l sin (θ )) i
r
r r r
ρ& = l θ& e = l θ& cos(θ ) j + l θ& sin (θ ) k
r θ

Therefore,

r r r r
v P = −ω ⋅ (a + l sin (θ )) i + l θ& cos(θ ) j + l θ& sin (θ ) k (a)

The acceleration of point P is given by

r &r& r r r r r r r r
a P = R + ω × (ω × ρ ) + ω& × ρ + ρ&&r + 2 ω × ρ& r

where

49
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

&r&
R=0
r
r r r
[ r
]
ω × (ω × ρ ) = ω k × − ω ⋅ (a + l sin (θ ))i = −ω 2 (a + l sin (θ )) j
r
r r r
ω& × ρ = 0 × ρ = 0
r
r r r
( r
) (
ρ&& = l θ& eθ − lθ& 2 eR = l θ&&cos(θ ) − l θ& 2 sin (θ ) j + l θ&&sin (θ ) − l θ& 2 cos(θ ) k )
( )
r r r r r r
2 ω × ρ& r = 2 ω k × l θ& cos(θ ) j − l θ& sin (θ ) k = −2ω l θ& cos(θ )i

Thus, gathering terms we have

r r
[ ( r
a P = −2 ω l θ& cos(θ )i + − ω 2 (a + l sin (θ )) + lθ&& cos θ − l θ& 2 sin (θ ) j
r
)]
( )
(b)
+ l θ&&sin (θ ) + l θ& 2 cos(θ ) k

Solution (B): One may also have been tempted to select a rotating coordinate system x ′y ′z ′ fixed to
the platform, but with origin at Q rather than O. For this case, the relative problem is still a simple
circular motion of m about Q exactly as before. However, the velocity of P is now given by

r r& r r r
v P = R + ω × ρ + ρ& r

where

r& r
R = −ω a i
r r r
( r r r
ω × ρ = ω k × l sin (θ ) j − l cos(θ ) k = −ω l sin (θ )i
r
)
r r
ρ& = l θ& cos(θ ) j + l θ& sin (θ ) k
r

Gathering terms, we once again obtain

r r r r
v P = −ω ⋅ (a + l sin (θ )) i + l θ& cos(θ ) j + l θ& sin (θ ) k (c)

Similarly, the acceleration of P is

r &r& r r r r r r r r
a = R + ω& × ρ + ω × (ω × ρ ) + ρ&&r + 2 ω × ρ& r

where

50
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

&r& r
R = −a ω 2 j
r
ω × (ω × ρ ) = ω k × (− ω l sin (θ )i ) = −ω 2 l sin (θ ) j
r r r r r
r r
ω& × ρ = 0
r
r
( r
) (
ρ&&r = l θ&&cos(θ ) − l θ& 2 sin (θ ) j + l θ&&sin (θ ) + l θ& 2 cos(θ ) k )
r r r
2 ω × ρ& = −2 ω l θ& cos(θ )i
r

Therefore,

r r
[ ( r
a P = −2 ω l θ& cos(θ )i + − ω 2 (a + l sin (θ )) + lθ&& cos θ − l θ& 2 sin (θ ) j
r
)]
( )
(d)
+ l θ&&sin (θ ) + l θ& 2 cos(θ ) k

As expected, the answers are the same for both solutions, but the various terms appear in somewhat
different ways.

Example 1.11 Cutting Tool Mechanism

The pin P is attached to the rim of a disc that rotates at a constant angular velocity of 10 rads
counterclockwise, as shown in Figure 1.23. Pin P is used to control a cutting tool as it moves along
the slotted rod OD. Determine the angular velocity and acceleration of rod OD for the instant shown.

Y D

P Tool Linkage

5”
O
Q

X
7” 5”

ω = 10

Figure 1.23: Mechanism for cutting tool.

51
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Solution: Since P is a point on the rim of the disc, its velocity and acceleration are easily determined
to be

r r r
v P = −50 e X + 50 eY (a)

r r r
a P = −500 e X − 500 eY (b)

It is also apparent that pin P moves relative to the rotating rod OD as it slides along the slot. Thus, it is
convenient to attach a rotating coordinate system xyz to rod OD as shown in Figure 1.24.

x
Y r
D i
y
r r
r ρ& r , ρ&&r
r r ρ P
j eY

O r φ = tan−1 (125 )
r ω&
ωOD OD
r
eX X

Figure 1.24: Analysis of cutting tool mechanism with a rotating coordinate system

Accordingly, we may write

r r& r r r
v = R + ω × ρ + ρ& r

where

r&
R=0
r r r r r r r
ω × ρ = ωOD k ×13 i = 13ωOD j = −5ωOD e X + 12 ωOD eY
r r 12 r 5 r
ρ& r = ρ& r i = ρ& r e X + ρ& r eY
13 13

Thus

52
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r  12  r  5 r
v P =  − 5ωOD + ρ& r  e X + 12 ωOD + ρ& r  eY (c)
 13   13 

Equating Eqs. (a) and (c), we obtain

r r
ρ& r = −26.6 i [ins ]
r (d)
r
ωOD = 5.08 k [rads ]

Continuing to the determination of acceleration,

r &r& r r r r r r r r
a = R + ω& × ρ + ω × (ω × ρ ) + ρ&&r + 2 ω × ρ& r

where

&r&
R=0
r r r r r
ω × (ω × ρ ) = −310 e X − 129 eY
r r r r r r
ω& × ρ = ω& OD k ×13 i = −5ω& OD e X + 12 ω& OD eY
r r 12 r 5 r
ρ&&r = ρ&&r i = ρ&&r e X + ρ&&r eY
13 13
r
2 ω × ρ r = 10.16 k × (− 26.6 i ) = −270 j = 104 e X − 249 eY
r r& r r r r

r  12  r  5 r
a P =  − 206 − 5 ω& OD + ρ&&r  e X +  − 378 + 12 ω& OD + ρ&&r  eY (e)
 13   13 

Thus, equating Eqs. (b) and (e), we have

r
ρ&&r = −318 i
r
[] in
s2

[ ]
r (f)
ω& OD ≈ 0 rad
s2

Example 1.13 Projectile Fired from Slewing Gun Turret


r r
A projectile is fired from a gun turret with a muzzle velocity v and acceleration a at the instant the
shell leaves the gun. At the instant of firing, the gun is elevating relative to the turret at a rate of α&
which is increasing at a rate of α&& . Further, the turret is in rotation about a vertical axis relative to the

53
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

XYZ reference as shown at a rate β& which is increasing at a rate β&& . Determine the velocity and
acceleration of the shell relative to the XYZ frame.

Z, z

r
r& &r& P r a
β, β v

α
O Y, y

L
r r
α& ,α&&
X, x

Figure 1.25: Gun turret.

Solution: In this problem, there are two angular velocities involved and one faces the choice of how
to best define the rotating coordinate system. The choice is between fixing a rotating coordinate
system to the turret so that the gun barrel rotates relative to it or fixing it directly to the gun barrel so
that the only relative motion involves the projectile motion relative to the gun barrel. Note that the
first choice suggested means that the relative problem comprises a rotating gun barrel with a projectile
moving relative to it. This relative problem really involves a second rotating coordinate system with
its own associated Coriolis effects. This type of problem, coming under the heading of multiple
rotating coordinate systems, is discussed in the next section, where this problem is resolved.

Consequently, in order for the relative motion to be easily described, it will be necessary to attach the
rotating xyz coordinate system directly to the gun barrel with the gun barrel at a constant angle α
from the x-axis. Therefore consider that the xyz frame shown in Figure 1.25 is instantaneously aligned
with the XYZ frame for the position shown at time t. Thus, as time goes on, the angle α remains
constant, since the xyz axes are fixed to the gun barrel.

54
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r&
However, the angular velocity vector of the turret, given by β , must rotate relative to the xyz frame at
r r&
the rate − α& i since β is defined as a rotation about the vertical Z axis.1

The angular velocity of the xyz coordinate system relative to XYZ is given for the position shown at
time t by

r r r
ω = α& i + β& k (a)

Thus, according to Eqs. 1.57 and 1.51, the angular acceleration is

( )
r r
ω& R = ω& = α&& i + β&& k + (− α& i )× α& i + β& k
r r r r r

or

r r r r
ω& = α&& i + α& β& j + β&& k (b)

The velocity of projectile P relative to the XYZ frame is given by

r r& r r r
v P = R + ω × ρ + ρ& r

where

r&
R=0
r r r
( r
)( r r
ω × ρ = α& i + β& k × l cos(α ) j + l sin (α ) k
r
)
r r
ρ& = v cos(α ) j + v sin (α ) k
r

Therefore,

r r r r
v P = −l β& cos(α )i + (v cos(α ) − l α& sin (α )) j + (v sin (α ) + l α& cos(α )) k (c)

Similarly, the acceleration of projectile P relative to the XYZ frame is given by

1 This is precisely the complication that can be avoided by using multiple coordinate systems as discussed in the next section. It

becomes especially important if one were to include further rotational effects such as the XYZ being attached to a rolling, pitching,
and yawing ship.

55
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r &r& r r r r r r r r
a P = R + ω& × ρ + ω × (ω × ρ ) + ρ&&r + 2 ω × ρ& r

where the previously undefined terms are:

&r&
R=0
r r r
ρ&& = a cos(α ) j + a sin (α ) k
r

Carrying out the indicated algebra, we find

r
( r
a P = 2 l α& β& sin (α ) − l β&& cos(α ) − 2 v β& cos(α ) i )
( r
− l β& 2 cos(α ) + l α& 2 cos(α ) + l α&& sin (α ) + 2 v α& sin (α ) − a cos(α ) j )
r
(
+ l α&& cos(α ) + 2 v α& cos(α ) − l α& 2 sin (α ) + a sin (α ) k )

1.9 Multiple Moving Reference Frames


In addition to the use of a single moving reference frame, it is important to extend this concept to
multiple moving reference frames, since in many complicated kinematic situations such an approach
may be highly desirable.

Consider the system shown in Figure 1.26, where although only three frames are shown, this
procedure may be readily extended to any number of frames. In Figure 1.26, x0y0z0 is considered the
basic, or primary, reference frame, while x1y1z1 and x2y2z2 represent two auxiliary frames that are in
r
motion relative to each other and also with respect to the x0y0z0 frame. The point P is located by ρ 2
r r
with respect to the x2y2z2 frame. In addition, the notation R2 and R1 is introduced to locate the
consecutive origins.

56
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Z0
Z1 Y2
P
Y1
r
ρ1 r
O1 ρ2
r r X2
R1 R2
O2
X1
O
Y0 Z2

X0

Figure 1.26: Multiple moving reference frames.

Initially considering only the x1y1z1 and x2y2z2 frames, we may write the velocity of P relative to frame
1 from Eq. 1.63 as

r r& r r r
v1 = R2 + ω 2 / 1 × ρ 2 + ρ& 2 ( ) r (1.67)

r
where ω 2 / 1 is the angular velocity of frame 2 with respect to frame 1. Considering next the x0y0z0 and
x1y1z1 frames, we see upon reapplying Eq. 1.63 that

r r& r r r
v0 = R1 + ω1/ 0 × ρ1 + ρ&1 ( ) r (1.68)

where

(ρr& )
1 r
r
= v1 (1.69)

In a similar fashion, one may extend this chain rule concept to any number of frames.

Proceeding in an analogous way, the acceleration relative to frame 1 from Eq. 1.66 is:

r &r& r r r r r r
( ) r r
a1 = R2 + ω 2 / 1 × (ω 2 / 1 × ρ 2 ) + ω& 2 / 1 × ρ 2 + ρ&&2 r + 2 ω 2 / 1 × ρ& 2 ( ) r (1.70)

Reapplying Eq. 1.66, the acceleration of P relative to the x0y0z0 frame is

57
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r &r& r r r r r r r
( ) r
a0 = R1 + ω1 / 0 × (ω1/ 0 × ρ1 ) + ω&1/ 0 × ρ1 + ρ&&1 r + 2 ω1/ 0 × ρ&1 ( ) r (1.71)

where

(ρ&r& )
1 r
r
= a1 (1.72)

and again

(ρr& )
1 r
r
= v1 (1.73)

This procedure applied to Eqs. 1.70 and 1.71 can be extended as a chain rule to any number of frames.
Perhaps the most orderly, though not necessarily the most expedient, approach is to introduce a
rotating frame for each angular velocity involved in a problem. Then the question of directional
derivatives of angular velocities will never become an issue.

Example 1.14 Projectile Fired from Slewing Gun Turret 2

Apply the concept of multiple moving reference frames to check the results obtained for example
1.13.

Solution: Referring to Figure 1.25 and Figure 1.26, we will define the following frames of reference

• x0y0z0 = XYZ inertial frame in Fig. Figure 1.25

• x1y1z1 = reference frame fixed to the turret

• x2y2z2 = reference frame fixed to the gun barrel, which was the xyz frame used in Example
1.13 that is instantaneously coincident with the XYZ frame.

Using the notation of Section 1.9, we have

58
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r r
ω1 / 0 = β& k
r r
ω 2 / 1 = α& i
r r (a)
ω&1 / 0 = β&& k
r r
ω& 2 / 1 = α&& i

Therefore, the velocity of the projectile P relative to the turret (frame 1) according to Eq. 1.67 is

r r& r r r
v1 = R2 + ω 2 / 1 × ρ 2 + ρ& 2 ( ) r

where:

r&
R2 = 0
r r
ω 2 / 1 = α& i
( )
r r r
ρ 2 = l cos(α ) j + l sin (α ) k
( )
r r r
ρ& = v cos(α ) j + v sin (α ) k
2 r

Thus

r r r
v1 = (v cos(α ) − l α& sin (α )) j + (v sin (α ) + l α& cos(α )) k (b)

From Eq. 1.68, we may then write the velocity of P relative to the XYZ frame as

r r& r r r
v0 = R1 + ω1/ 0 × ρ1 + ρ&1 ( ) r

where

r&
R1 = 0
r r
ω1 / 0 = β& k
r r r
ρ1 = l cos(α ) j + l sin (α ) k
( )
r r r
ρ& = (v cos(α ) − l α& sin (α )) j + (v sin (α ) + l α& cos(α )) k
1 r

Therefore,

r r r r
v0 = −l β& cos(α )i + (v cos(α ) − l α& sin (α )) j + (v sin (α ) + l α& cos(α )) k (c)

59
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

which agrees with Eq. (c) of Example 1.11.

Similarly, the acceleration of P relative to the turret (frame 1) is given by Eq. 1.70 as

r &r& r r r r r r
( ) r r
a1 = R2 + ω 2 / 1 × (ω 2 / 1 × ρ 2 ) + ω& 2 / 1 × ρ 2 + ρ&&2 r + 2 ω 2 / 1 × ρ& 2 ( ) r

where the previously unidentified terms are:

&r&
R2 = 0
( )r r r
ρ&&2 r = a cos(α ) j + a sin (α ) k

Upon carrying out the mathematical operations, we have

r
( r
a1 = a cos(α ) − l α&& sin (α ) − l α& 2 cos(α ) − 2 v α& sin (α ) j )
r
( )
(d)
+ a sin (α ) + l α&& cos(α ) − l α& 2 sin (α ) + 2 v α& cos(α ) k

Finally, the acceleration of P relative to the XYZ frame according to Eq. 1.71 and 1.72 is given by

r &r& r r r r r r
( )
r r
a0 = R1 + ω1/ 0 × (ω1/ 0 × ρ1 ) + ω&1/ 0 × ρ1 + ρ&&1 r + 2 ω1/ 0 × ρ&1 ( ) r

where

&r&
R1 = 0
( )r r
ρ&& = a
1 r 1

Therefore, after substitution, we obtain as before

r
( r
a0 = 2 l α& β& sin (α ) − l β&& cos(α ) − 2 v β& cos(α ) i )
( r
− l β& 2 cos(α ) + l α& 2 cos(α ) + l α&& sin (α ) + 2 v α& sin (α ) − a cos(α ) j ) (e)
r
(
+ l α&& cos(α ) + 2 v α& cos(α ) − l α& 2 sin (α ) + a sin (α ) k )

60
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Although this procedure may lead to a greater number of steps in a particular analysis, when dealing
with systems requiring more than two rotating reference frames, each step is relatively straight
forward. Furthermore, the procedure can be readily applied as a chain rule to problems of any degree
of complexity.

1.10 The Coriolis Effect for Motion Relative to the Rotating


Earth
It is interesting to investigate the effect and importance of Coriolis acceleration in problems involving
motion relative to the Earth, which is actually a rotating reference frame. In most practical
applications of kinematics, our basis of reference is usually selected as a frame attached to the Earth’s
surface. The Earth, however, has a rather complex motion in inertial space. Neglecting for a moment
the spin, nutation, and precessional motions of the Earth, we have such effects as the motion of the
Earth with respect to the sun, the motion of the sun with respect to the Milky Way galaxy, and the
motion of the galaxy with respect to the universe (See Figure 1.27).

The motion of the Milky Way galaxy within the universe is not well known, although we may
postulate that it is small. The motion of the sun with respect to the center of the galaxy is at a speed of
r
mi
180 sec at a distance of about ρ s = 33,000 light years in a spiral path. The distance conversion is

r
ρ s = 33,000 yr ×186,000 sec
mi
× 5280 mift × 3600 sec
hr × 24 day × 365 yr = 1.022 × 10
hr day 21
ft

which gives rise to an angular velocity of approximately

r mi
180 sec × 5280 mift
ωS / G = r = 9.30 × 10 −16 rad
ρs s

The magnitude of the associated normal acceleration is about

r s& 2
aN = = 8.84 × 10 −10 ft

ρ s2

which as a percentage of g is about 0.27 × 10 −8 %.

61
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

The motion of Earth with respect to the sun is of course one revolution per year in a slightly elliptic
r
orbit with a mean radius of about ρ E = 93,000,000 mi . This gives rise to an angular velocity of

about

r
ω E / S = 1 rev
yr = 1.99 × 10
−7 rad
s

The magnitude of this normal acceleration is about

r r r 2
a N = ρ E ω E / S = 1.95 × 10 −2 ft
s2

which is about 0.06 % of g. Another acceleration which occurs in the motion of the Earth about the
sun is a small tangential acceleration due to the slightly elliptic orbit of the Earth about the sun. This,
however, is extremely small.

It is clear from the above discussion that for engineering work one may certainly neglect the
r r
accelerations associated with ω S / G and ω E / S compared to the acceleration of gravity, g. Turning
attention now to the Earth’s precession and spin, we find these angular velocities to be approximately

r 1 rev
ω precession = = 7.65 × 10 −12 rad
s
26,000 yr
r
ω spin = 1 day
rev
= 7.27 ×10 −5 rads

Clearly, we may neglect the Earth’s precessional motion compared to its spin.

In order to investigate the importance of the Earth’s daily rotation, we will determine the normal
acceleration associated with a point on the Earth’s surface as it moves in a circular path about the spin
axis. Thus, we will define the XYZ axes as an equatorial geocentric coordinate system with the origin
at the Earth’s center, the X axis directed toward the vernal equinox (line of intersection of the ecliptic
and equatorial planes), the Z axis along the Earth’s spin axis with the XY plane being the equatorial
plane as shown in Figure 1.28.

In addition, we will define a topocentric coordinate system xyz with its origin at the surface point O at
latitude λ so that the axes xyz are fixed to the Earth with the x axis directed east, the y axis north, and

62
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

the z axis radially outward, as shown in Figure 1.28. In this way, the xyz coordinate system rotates
about the Z axis at the rate ω e = 1 day
rev
relative to the XYZ frame.

The velocity and acceleration of the origin O of the xyz frame relative to the XYZ frame, considering
the angular velocity of the Earth to be constant, is approximately given by

Earth Center of Sun’s Galaxy

G
E r
ρE r
ρS
S

Sun

Universe Sun’s Galaxy

Figure 1.27: Relative location of approximate inertial frames.

r r r r
vO = ω e × R = ωe R cos(λ )i
r r
( )
r
( )
r r r (1.74)
aO = ωe × ωe × R = R ω e cos(λ ) sin (λ ) j − cos(λ ) k
2

For the radius of the Earth of approximately 3960 miles and ω e = 7.27 ×10 −5 rad
s , we have

r r
vO = 1520 cos(λ )i ft
s []
r
(
r r
aO = 0.112 cos(λ ) sin (λ ) j − cos(λ ) k ) [] ft
s2
(1.75)

with maximum values obviously occurring at the equator for λ = 0 . Thus, the maximum normal
acceleration due to the spin of the Earth is approximately 0.35 % of the acceleration of gravity,
g = 32.2 sft2 , which is usually a negligible effect in engineering applications. It is also interesting to

note from the first of Eqs. 1.75 the advantage gained from the Earth’s rotation in launching an
artificial satellite eastward as near to the equator as feasible.

63
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

In general, the acceleration of a particle P, shown in Figure 1.28 referred to the xyz frame by the
r
position vector ρ , is given by

r r r r r r r r r r
a P = aO + ωe × (ωe × ρ ) + ω& e × ρ + ρ&&r + 2 ω e × ρ& r (1.76)

where

r r r r r r
(
aO = ωe × ωe × R + ω& e × R )
r
However, since ω e is very nearly constant for the Earth, we have approximately that

( ( ))
r r r r r r r r
a P = ω e × ω e × R + ρ + ρ&&r + 2 ω e × ρ& r (1.77)

r r
If we consider that the only external force on the particle is gravity, for which a P = − g k , we may
write

ρ&&r = − g k − ω e × (ω e × (R + ρ )) − 2 ωe × ρ& r
r r r r r r r r

r r r
For example, if ρ , ρ& r , and ρ&&r are expressed in rectangular components as

r r r r
ρ = xi + y j + z k
r r r r
ρ& = x& i + y& j + z& k (1.79)
r r r r
ρ&& = &x&i + &y& j + &z& k

we may write Eq. 1.78 in component form as the following set of linear differential equations:

&x& = 2 ω e ⋅ ( y& sin (λ ) − z& cos(λ )) + x ωe 2


&y& = −2 ω e x& sin (λ ) − R ω e 2 sin (λ )cos(λ ) − ωe 2 sin (λ )( z cos(λ ) − y sin (λ )) (1.80)
&z& = − g + 2 ω e x& cos(λ ) + R ω e cos (λ ) + ωe cos(λ )( z cos(λ ) − y sin (λ ))
2 2 2

r r
For most situations, for which ρ << R and the ωe terms are small, Eqs. 1.80 can be simplified to
2

64
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

&x& = 2 ω e ⋅ ( y& sin (λ ) − z& cos(λ )) (a) 



&y& = −2 ωe x& sin (λ ) (b) (1.81)
&z& = − g + 2 ω e x& cos(λ ) (c) 

where the only terms remaining are the acceleration of gravity g and the Coriolis terms.

A first integration of Eqs. 1.81 with respect to time yields

x& = 2 ω e ⋅ ( y sin (λ ) − z cos(λ )) + C x (a) 



y& = −2 ωe x sin (λ ) + C y (b) (1.82)
z& = − g t + 2 ωe x cos(λ ) + C z (c) 

Substituting Eqs. 1.82b and 1.82c into Eq. 1.81 a we find

&x& = 2 ω e ⋅ (− 2 ω e x sin (λ ) + C y )sin (λ ) − 2 ω e ⋅ (− g t + 2 ωe x cos(λ ) + C z )cos(λ ) (1.83)

Neglecting terms of order ωe and considering the initial conditions at t = 0 to be x = x0 , y = y0 ,


2

z = z 0 , x& = x& 0 , y& = y& 0 , and z& = z&0 , then two integrations of Eq. 1.83 with respect to time yields

ωe g t 3
x= cos(λ ) − ωe t 2 z&0 cos(λ ) + ωe t 2 y& 0 sin (λ ) + x&0 t + x0 (1.84)
3

Substituting Eq. 1.84 into Eqs. 1.82a and 1.82b, neglecting terms of order ωe , and integrating with
2

respect to time, yields the governing set of equations given by

ωe g t 3
x= cos(λ ) + ω e t 2 ( y& 0 sin (λ ) − z&0 cos(λ )) + x& 0 t + x0
3 (a) 

y = −ω e t x& 0 sin (λ ) + y& 0 t + y0
2
(b) (1.85)
gt2 (c) 
z=− + z&0 t + z 0 + ω e x& 0 t 2 cos(λ )
2

It is readily apparent that Eqs. 1.85 reduce to the more familiar form from elementary dynamics if we
let ωe = 0 . It is also apparent from Eqs. 1.85 that x, y, z depend on ω e and t 2 which indirectly

65
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

involves distance traveled. Although ωe is small, if t is large, then substantial deviations will occur in

the x, y, z positions due to the Coriolis drift, which led to all terms in Eqs. 1.85 containing ω e .

Example 1.15 Projectile with Coriolis Effects

A projectile is fired due north from the Earth’s surface at the equator with initial velocity components
x&0 = 0 , y& 0 = 4000 fts , z&0 = 3000 fts . Determine where it passes the local xy horizontal plane upon

descent.

Solution: Let us set the origin of the xyz frame at the launching position so that x0 = y0 = z 0 = 0 .

Then directly from Eq. 1.85c, when z = 0 , we have

gt2
0=− + z&0 t (a)
2

or

2 z&0
t= = 186.3 s (b)
g

From Eq. 1.85b,

y = y& 0 t = 745,200 ft = 141.1 mi (c)

and thus from Eq. 1.85a

ωe g t 3
x= cos(λ ) + ω e t 2 ( y& 0 sin (λ ) − z&0 cos(λ ))
3
 32.2 
= 7.27 × 10 −5  186.33 − 186.3 2 ⋅ 3000 
 3 

x = −5120 ft (d)

66
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Thus, the projectile has drifted 5120 feet westerly so that upon descent when z = 0 , x = −5120 ft
and y = 745,000 ft .

r
ω y
r
ρ
z
O
x
r
R
Y
λ
Equator
ψ
X

Figure 1.28: Relative reference frame for projectile

1.11 Rigid Body Kinematics


The concept of a rigid body is a convenient way of modeling bodies whose deformations are
negligible. The analytical advantage of treating a body as being rigid results from the fact that each
point of a rigid body remains at a fixed distance from every other point on the body. Thus, by
attaching a coordinate system directly to a rigid body, shown in Figure 1.29 as an xyz frame, we can
easily describe the motion of any point B on the body relative to any other point A by using the general
r
kinematical equations for rotating reference frames. Note that the angular velocity ω and angular
r
acceleration ω& of the xyz coordinate system and of the rigid body are identical. The xyz axes are body
axes (i.e., attached directly to the body).

67
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

z
r
Y r ω& r
ω vA
r
aA
B y
r
r ρ A/ B
rB
r A
rA

O
X x

Figure 1.29: A point on a rigid body relative to a reference frame

Referring to Figure 1.29 and Eqs. 1.63 and 1.66, we may describe the velocity and acceleration of
point B in terms of the motion of point A on the body and the body’s angular motion as follows:

r r r r
vB = v A + ω × ρ A / B (1.86)

and

r r r r r r r
a B = a A + ω& × ρ A / B + ω × (ω × ρ A / B ) (1.87)

r r
The schematic diagram shown in Figure 1.30 shows that the ω × ρ A / B term in Eq. 1.86 represents a
r r r
circular motion of point B about the ω vector drawn through point A. The vector ω × ρ A / B shown in
r
Figure 1.30 is tangent to the circular path at B and perpendicular to the plane formed by the ω and
r
ρ A / B vectors. Thus, point B has a velocity equal to the velocity of point A plus a component of
motion relative to point A which can be thought of as motion along a circular path centered along the
r
ω vector. Similarly, the last two terms in Eq. 1.87 can be physically described in terms of circular
motion, which is left as an exercise for the student.

68
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
ω

r r
B ω × ρA/ B
r
ρA/ B
r
vA

A
Figure 1.30: Visualization of relative velocity equation.

Furthermore, if we consider plane motion, this concept can be further illustrated as shown in Figure
r r
1.31. For this case, the angular velocity vector ω and angular acceleration vector ω& are
perpendicular to the xy plane as shown.

r r r r r
ω × ρA/ B ω × (ω × ρ A / B )
r
ρA/ B
y r
A
r
vA ω&
x
r
ω

Figure 1.31: Plane motion projection of relative velocity and a relative acceleration component.

r
For plane motion, we may conclude that the general velocity v B of any point B on a rigid body is
r
given by the sum of the velocity v A of a reference point A plus the velocity of point B relative to A as
r r
it moves in a circular path of radius ρ A / B about point A shown in Figure 1.31 as the vector ω × ρ A / B .
This is a fairly obvious result since point B always remains at a fixed distance from point A.

69
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Similarly, since point B moves in a circular path relative to A, it has the characteristic acceleration
r r
components of circular motion—namely tangential acceleration ω& × ρ A / B and normal acceleration
r r r r
ω × (ω × ρ A / B ) , both relative to the motion of point A. Therefore, the general acceleration a B of any
r
point B on a rigid body in plane motion is equal to the sum of the acceleration a A of the reference
point A plus the acceleration of B as it moves in a circular path of radius ρ A / B about point A which in
r r r r r
this case consists of acceleration components ω& × ρ A / B and ω × (ω × ρ A / B ) as shown in Figure 1.31.

There are three additional important concepts that have been known for many years concerning basic
rigid body motion, which will be discussed in the following sections, namely

A. Euler’s Theorem

B. Chasles’ Theorem

C. Instantaneous Centers for Plane Motion

A. Euler’s Theorem

The most general displacement of a rigid body with one point fixed is equivalent to a single rotation
about some axis through that point.

In order to establish this theorem, refer to Figure 1.29 and let point A coincide with the origin O of
r r r
reference frame XYZ at all times t. Then, from Eq. 1.86, since now v A = 0 and rB = ρ A / B , we have

r r
r drB r r dθ r
vB = = ω × ρ A/ B = × rB (1.88)
dt dt

and writing Eq. 1.88 in differential form it reduces to

r r r
drB = dθ × rB (1.89)

Thus, from Eq. 1.89, the only possible motion of an arbitrary point on a rigid body with some point
fixed is a rotation.

70
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Furthermore, the most general displacement of point B can be accomplished by an infinite number of
infinitesimal displacements written as

r ∞
r ∞ r r
drB = ∑ drBi = ∑ dθ i × rB (1.90)
i =1 i =1

r
From prior proof in Section 1.2 that infinitesimal rotation dθ i are vectors and may thus be added
r
vectorially to yield a resultant vector dθ , we may reduce Eq. 1.90 to

r ∞
r  ∞ r r r r
drB = ∑ drBi =  ∑ dθ i  × rB = dθ × rB (1.91)
i =1  i =1 

Accordingly, from Eq. 1.92 the most general displacement of any point B on a rigid body consists of a
resultant rotation about an axis through O, thus proving Euler’s Theorem.

B. Chasles’ Theorem

The most general displacement of a rigid body with no fixed points is equivalent to a translation of an
arbitrary reference point on the body plus a rotation about an axis through that reference point.

Consider again Eq. 1.86 with both points A and B in Figure 1.29 free to move with respect to the XYZ
coordinate system. We may then write

r r r r
vB = v A + ω × ρ A / B (1.92)

or

r r r
drB drA dθ r
= + × ρ A/ B (1.93)
dt dt dt

which in differential form becomes

r r r r
drB = drA + dθ × ρ A / B (1.94)

71
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r r
Since we have already established that dθ × ρ A / B represents a rotation of the rigid body about point A
r
and drA represents a translation of point A, we may conclude that Eq. 1.94 establishes the validity of
Chasles’ Theorem. Furthermore, since three translational and three rotational scalar parameters are
sufficient for describing the position and orientation of a rigid body, we refer to a rigid body as having
six degrees of freedom; three translational and three rotational.

C. Instantaneous Centers for Plane Motion

The instantaneous velocity of any point on a rigid body moving in plane motion can be determined by
considering the rigid body to be undergoing pure rotation about an instantaneous center, which is
defined as a point on the rigid body or on a rigid extension of the body with zero velocity at the instant
of consideration.

Basically, the concept of instantaneous centers can be discussed in a general way only for plane
motion. Furthermore, this concept only applies to velocity, not acceleration. The basic question
suggested is whether or not it is always possible to determine some point on the rigid body or a rigid
extension of it that has zero velocity at an arbitrary instant of time. Surely, the point satisfying this
condition at time t1 and at a different time t 2 will in general be a different point. Then, it remains to
be shown that the velocity of any point on the rigid body can be described by analytically treating the
entire rigid body as rotating about this instantaneous center, with each point of the rigid body
instantaneously moving in a circular path about the instantaneous center at time t.

r r
Consider the rigid body shown in Figure 1.32 with points A and B having velocities v A and vB
respectively at time t. The rigid body shown moves in plane motion and rotates with the angular
r
velocity ω , which would be represented by a vector directed normal to the plane of motion.

72
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
r A j
vB r
B vA r
i

C
r
ρ O′ / C

r r
O′ ω =ωk

Figure 1.32: Intantaneous center.

r r
Construct perpendiculars to each of the velocity vectors v A and v B as shown and denote the
intersection of these lines as point O′ . Consider that point O′ denotes a point on the rigid body or on
an extension of the body, as shown by the dashed lines for this case. Therefore, using point O′ as a
reference point on the body, we may describe the motion of each point A and B in terms of the motion
of point O′ and their motion relative to O′ as follows

r r r r
v A = vO ′ + ω × ρ O ′ / A (1.95)

and

r r r r
vB = vO′ + ω × ρ O′ / B (1.96)

r r
Since the vector ρ O′ / A is perpendicular to the vector ω , which is directly outward in Figure 1.32, we
r r r
note that the vector ω × ρ O′ / A is parallel to vector v A . Therefore, in order for the equality described
r r
by Eq. 1.95 to be satisfied, we may conclude that the velocity vector vO′ is parallel to vector v A .

r r r
Similarly, referring to Eq. 1.96 we note that the vector ω × ρ O′ / B is parallel to vector v B and therefore
r r
vO′ must also be parallel to vector v B . In conclusion, we see that the only way both Eqs. 1.95 and
r
1.96 can be satisfied is for the velocity vO′ of point O′ to be equal to zero. Consequently, point O′ is
called the instantaneous center, which is a point which instantaneously has zero velocity.

73
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

It follows directly that once the instantaneous center O′ is located, then the velocity of any point C on
r
the rigid body located from point O′ by the position vector ρ O′ / C can be readily determined by
writing

r r r r
vC = vO′ + ω × ρ O′ / C

r
or since vO′ = 0 , we have simply

r r r
vC = ω × ρ O′ / C (1.97)

Recall that Eq. 1.97 shows that at the instant under consideration, the velocity of any point C may be
determined by considering it as instantaneously moving in a circular path of radius ρ O′ / C centered at
O′ at an angular rate ω . Frequently, in velocity problems involving a rigid body or systems of rigid
bodies, the instantaneous center method is much more easily applied than other standard vector
approaches.

It is not intended to imply from the previous discussion that the only way to locate the instantaneous
center O′ is to be given the velocities of any two points on the rigid body at time t. In fact, in the
discussion associated with Figure 1.32, it was only necessary to be given the direction of the velocities
of any two points A and B in order to uniquely locate the instantaneous center O′ by constructing
r r
perpendicular lines to the velocity vectors v A and v B through the points A and B. Use was therefore
made of only two items of information at time t, namely the instantaneous directions of the velocities
of points A and B.

Similarly, if several other combinations of instantaneous motion parameters are available, the
instantaneous center can be readily located. It is left as an exercise for the student to show that the
instantaneous center can be located, for example, given the direction of the velocity of one point on
r
the body along with the angular velocity vector ω . Furthermore, the student should prove that if the
r r
velocity vectors v A and v B in Figure 1.32 are parallel, then the instantaneous angular velocity ω of
r
the rigid body is zero. The motion of a rigid body for which ω = 0 is called translation.

74
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Example 1.16 Instantaneous Center of Gears

In Figure 1.33 is shown a system of three gears, A, B, and C. Gear A is attached to a shaft and rotates
about axis O while gear B is constrained to move about gear A by arm OD which is rotating at a rate
rad
of 60 s counterclockwise. Determine the angular velocity of gear A and the instantaneous center of
gear B if

a) gear C is fixed

rad
b) gear C rotates clockwise about axis O at 40 s

r F D
O E
r x

A B
C

Figure 1.33: Gear assembly.

Solution:

a) Gear C fixed

The velocity of point D at the center of gear B is given by

r r r
vD = ω × ρ O / D

or

75
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

( )
r r r
v D = 60 k × (2 r i ) = 120 r j
r
(a)

The instantaneous center of gear B by inspection is the point of contact between gears B and C, since
this is a point of zero velocity on gear B. Designating this point as point E and the contact point
between gears A and B as F, we may write

r r r
v F = 2 v D = 240 r j (b)

Thus, the angular velocity of gear a is given by

vF
ωA = = 240 rads (c)
OF

or

r
r
ω A = 240 k [rads ] (d)

b) Gear C rotates

r r
The velocity of point D is still written v D = 120 r j . We also note that the velocity of point E is now
given by

( )
r r r
vE = − 40 k × (3 r i ) = −120 j
r
(e)

76
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
vD = 120 r

F D E
O′ x

r
vE = 120 r

Figure 1.34: Intantaneous center of gear B.

A sketch of gear B is shown in Figure 1.34 from which it is apparent that the instantaneous center of
gear B is located midway between points D and E. Therefore, using the concept of the instantaneous
center, the angular velocity of gear B is related to the velocity at E by the equation

r
ωB = v E = 120 r (f)
2

from which we have

r
ω B = 240 k [rads ] (g)

Consequently,

r 3 r
v F = r ω B = 360 r j (h)
2

r
The magnitude of ω A becomes

vF
ωA = = 360 rads (i)
OF

or

77
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r
r
ω A = 360 k [rads ] (j)

Example 1.17 Instantaneous Center of Crank Mechanism

A crank mechanism system from an internal combustion engine is shown in Figure 1.35. In the
rad
position shown, the crank arm AB has an instantaneous angular velocity of 3 s clockwise and an
rad
angular acceleration of 5 s2
counterclockwise.

a) Using the method of instantaneous centers, determine the velocity of the piston P for
the instant shown.

b) Check the answer to part (a) by using a vector method.

c) Determine the acceleration of the piston for the instant shown.

B
10"
5"

4"
ω& = 5
ω=3 23.6°
A
P x
3"

Figure 1.35: Crank mechanism system from internal combustion engine.

Solution:

a) Velocity by Instantaneous Centers Method

The schematic diagram shown in Figure 1.36 shows the direction and magnitude of the velocity of
point B and the direction of the velocity of the piston P. Thus, using the method of instantaneous

78
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

centers for the connecting rod BP, the entire rod BP can be thought of as instantaneously rotating
about the instant center O ′ .

O′
y
ω BP
q1

q2

B
5" r ω BP
vB
P r
A ω=3 vP
x
Figure 1.36: Instantaneous center of crank mechanism.

Referring to Figure 1.36, we have

vB
ωB / P = (a)
q1

and

vP = ω B / P q2 (b)

Thus, from geometry, q1 = 15.3 in and q 2 = 16.2 in from which we find

15
ωB / P = = 0.98
15.3

or

r
r
ω B / P = 0.98 k [rads ] (c)

and

79
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

r r
v P = 0.98 ⋅16.2 i = 15.9 ins (d)

b) Velocity by Vector Method

Repeating part (a) by a vector method, we may apply Eq. 1.86 to describe the motion of point P of the
connecting rod relative to the crank pin at point B. Thus, we have for rod BP

r r r r
vP = vB + ω B / P × ρ B / P (e)

where

r r
vP = vP i
r r r
v B = 12 i − 9 j
r r
ωB / P = ωB / P k
r r r
ρ B / P = 9.16 i − 4 j

Substituting Eq. (e) and carrying out the indicated vector operation we have

r r r r r
v P i = 12 i − 9 j + 9.16 ω B / P j + 4 ω B / P i (f)

r
Equating the i components we have

v P = 12 + 4 ω B / P (g)

r
and equating the j components yields

0 = −9 + 9.16 ω B / P (h)

Therefore, solving Eqs. (h) and (g) we have, as before

r
r
ω B / P = 0.98 k [rads ] (i)

and

r
r
v P = 15.9 i [ins ] (j)

80
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

c) Acceleration

In a way similar to part (b), we may write the acceleration of point P in terms of the acceleration of
point B plus its acceleration relative to point B as follows

r r r r r r r
a P = a B + ω& B / P × ρ B / P + ω B / P × (ω B / P × ρ B / P ) (k)

where

r r
aP = aP i
r r r r r r r r
a B = ω& A / B × ρ A / B + ω A / B × (ω A / B × ρ A / B ) = −47 i − 21 j
r r
ω& A / B = 5 k
r r r
ρ A/ B = 3i + 4 j
r r
ω A / B = −3 k
r r
ω& B / P = ω& B / P k
r r r
ρ B / P = 9.16 i − 4 j
r r
ω B / P = 0.98 k

Substituting these quantities into Eq. (k) and carrying out all vector operations yields the following
vector equation

r r r r r r r
a P i = −47 i − 21 j + (4 ω& B / P i + 9.16 ω& B / P j ) + (− 8.8 i + 3.84 j ) (l)

r r
Therefore, equating the i components of Eq. (l) as well as the j components, we obtain

a P = −48.3 sin2

or

r r
a P = −48.3 i []
in
s2
(m)

and

ω& B / P = 1.87 rad


s 2

81
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

or

[ ]
r r
ω& B / P = 1.87 k rad
s2
(n)

1.12 Euler Angles


A particularly useful and important way of describing the general angular orientation of a rigid body
involves the use of Euler angles. In Figure 1.37, point O is considered to be any reference point on a
rigid body or hypothetical extension of that body. Such a point, for example, may well be the center
of mass of the body or a fixed point. Axis Oz represents a geometric reference axis of the body, and
frame XYZ is a reference system whose origin coincides with point O for convenience. The body has
undergone the following three rotations in order to reach the general position shown. With the body
reference axis Oz initially aligned with the Z axis, the body is rotated through an angle ψ about the
OZ axis. Then, the geometric reference axis Oz is displaced an angle θ from the reference axis OZ,
and finally, the body is rotated through an angle φ about its geometric axis Oz. These Euler angles,
ψ , θ , and φ are called the angles of precession, nutation, and spin respectively.

Z
ψ&
θ
z

φ&

O Y

ψ
θ&
X
Figure 1.37: Rigid body in xyz frame after rotating the frame by Euler angles

This system of rotations may be more exactly defined in the following manner. Consider a coordinate
system x, y, z attached to the body with origin at point O and initially coinciding with the XYZ frame.

82
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

The xyz system first undergoes a positive rotation ψ about the Z axis to the position x1 , y1 , z1 shown
in Figure 1.38. Positive rotations are considered counterclockwise when viewed along a given axis
toward the origin O. Next, the xyz system is given a positive rotation θ about the Ox1 axis to the x2 ,
y 2 , z 2 position. The axis Ox1 is often referred to as the nodal line. Finally, the body undergoes a
positive rotation φ about the Oz2 axis to arrive at the general position of the x, y, z axes shown in
Figure 1.38.

Precession Axis
Z , z1
y
θ y2
z2 , z
φ
θ
Spin Axis
y1
ψ
Y
x
φ Nutation Axis
ψ (Line of Nodes)

X x1 , x2
Figure 1.38: Definition of Euler angle axis rotations.

It should be pointed out that other systems of Euler angles are in use, differing primarily in the order in
which the rotations are accomplished. The system presented here is the most commonly used one. In
flight mechanics problems, the Euler angles of precession ψ , nutation θ , and spin φ are directly
related to the angle of yaw, the pitch angle, and the angle of roll respectively.

A common example of the application of Euler angles is illustrated in Figure 1.39 in connection with
describing the motions of Earth in space. In addition to it movement about the sun in an elliptic orbit
in the ecliptic plane, the planet Earth experiences precession, nutation, and spin. The spin of Earth is
simply its daily rotation which has a period of 24 hours. The angle of nutation is the inclination of the
Earth’s spin axis with respect to the normal to the ecliptic plane, which is an angle of about 23.27º.

Furthermore, the spin axis of the Earth precesses westerly along the surface of a hypothetical cone
whose apex angle is 23.27º approximately once every 26,000 years. Thus, the line of intersection of

83
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

the equatorial and ecliptic planes, called the equinox line which is an important reference direction in
space, precesses westerly with the period of 26,000 years. This phenomenon is known as the
precession of the equinoxes and is mainly dependent on the moment exerted on the Earth by the sun
and moon due to the Earth’s oblateness, which refers primarily to the fact that the Earth is flattened at
its poles.

z (spin) Z
ψ (precession)
Earth’s Spin Axis
Spin Period = 24 hrs
φ
Precessional Period = 26,000 years

North Pole
23.27° Autumnal Equinox
θ

θ (nutation)
O Y
Equatorial Plane

Ecliptic Plane
θ&
Vernal Equinox
23.27°
South Pole
X,x
Figure 1.39: Motion of the Earth with respect to the Sun.

In Figure 1.37, the xy plane is the Earth’s equatorial plane and the XY plane is the ecliptic plane. Note
that the spin and precessional motions are opposite directions. Such motion is termed retrograde
precession. If the spin and precessional motions are in the same sense, it is called direct precession.

Recalling that infinitesimal angular changes add up vectorially, we may readily consider the
infinitesimal changes in the Euler angles, namely dψ , dθ , and dφ , in terms of their limits taken
with respect to time. The resulting angular velocity components as shown in Figure 1.40 are denoted
by ψ& (precession), θ& (nutation), and φ& (spin) respectively. These angular velocity components are
probably most familiar to the reader in reference to the motions of a top. The precession refers to the
top’s relatively slow coning action of its axis of radial symmetry, nutation refers to the angle of this
space cone (which is often nearly constant until friction dissipates the initial energy of the top) and the
spin refers to the top’s angular velocity about its axis of radial symmetry.

84
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

Frequently, in rigid body dynamic analysis, it becomes necessary (as will occur in Chapter 5) to
describe the components of angular velocity of the rigid body in terms of components along the x, y , z
r r r
body axes in the form ω = ω x i + ω y j + ω z k or along the precessing and nutating, but nonspinning

x2, y2, z2 axes. These components of angular velocity can be quite easily described from Figure 1.40
by purely geometrical considerations.

Z , z1 r
(Precession) ψ& A
y
y2
z2 , z
φ
θ
φ& (spin) θ
ψ& sin(θ ) y1

ψ& cos(θ ) ψ
Y
x
φ
ψ
θ& (nutation)
X x1 , x2 Line of Nodes

Figure 1.40: Vector projections in Euler angles

First, the components of precession, nutation, and spin along the x2, y2, z2 axes can be written in matrix
form as

ω x2   0 1 0 ψ& 
    & 
ω y2  =  sin (θ ) 0 0 θ  (1.98)
ω z  cos(θ ) 0 1   φ& 
 2

Similarly, the components of ψ& , θ& , and φ& along the x, y, z body axes shown in Figure 1.40 are
written in matrix form as

ω x   sin (θ )sin (φ ) cos(φ ) 0 ψ& 


ω  = sin (θ )cos(φ ) − sin (φ ) 0 θ&  (1.99)
 y   
ω z   cos(θ ) 0 1  φ& 

85
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

1.13 General Euler Angle Transformations


r r r r
Any vector A = AX i + AY j + AZ k whose components are given with respect to the X, Y, Z reference
axes shown in Figure 1.40 may be resolved into components along any other set of axes by
appropriate coordinate system transformations. A very important set of transformation equations are
r
the Euler angle transformations which are used for determining the components of a general vector A
along various sets of coordinate axes oriented by means of Euler angles.

r
Referring to Figure 1.40, any general vector A described in terms of components along the axes x1,
y1, z1 is given by

 Ax1   cos(ψ ) sin (ψ ) 0  AX 


  
A = − sin (
ψ ) cos (
ψ ) 0   A  = [ψ ] Ar (1.100)
 1 
y  Y 
 Az   0 0 1  AZ 
 1

r
Similarly, the vector A in terms of its components along the axes x2, y2, z2 is written

 Ax 2  1 0 0   Ax1   Ax1 
        r
A
 2 
y = 0 cos (θ ) sin (θ ) A
 1 
y = [θ ] A
 1
y = [θ ][
ψ ] A (1.101)
 Az  0 − sin (θ ) cos(θ )  Az   Az 
 2  1  1

r
and finally, the vector A in terms of its components along the x, y, z body axes becomes

 A   cos(φ ) sin (φ ) 0  Ax 2   Ax2 


r  x       r
A =  Ay  = − sin (φ ) cos(φ ) 0  Ay 2  = [φ ] Ay 2  = [φ ][θ ][ψ ] A (1.102)
 Az   0 0 1  Az 2   Az 
 2

where [ψ ] , [θ ] , [φ ] are called rotation matrices defined by Eqs. 1.100, 1.101, and 1.102 respectively.

r r r r
As an example, the position vector = X e X + Y eY + Z eZ , which locates a point P in the X, Y, Z
R
r r r r
reference frame relative to an origin O, is equivalent to the position vector r = x i + y j + z k
locating the same point P in the x, y, z frame relative to the same origin where the components are
related to one another by the equation

86
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

x X 
r
r =  y  = [φ ][θ ][ψ ] Y  = [φ ][θ ][ψ ] R
r  
(1.103)
 z   Z 

In later analyses, it will occasionally be convenient to allow a rigid body that has an axis of radial
symmetry to spin relative to the primary reference axes, shown as axes x2, y2, and z2 in Figure 1.40. in
these analyses, the following transformation equations will prove to be very useful:

 Ax2   AX   AX 
     
 Ay2  = [θ ][ψ ] AY  = [θ ,ψ ] AY  (1.104)
 Az   AZ   AZ 
 2

where

 cos(ψ ) sin (ψ ) 0 
[θ ,ψ ] = [θ ][ψ ] = − cos(θ )sin (ψ ) cos(θ )cos(ψ ) sin (θ ) (1.105)
 sin (θ )sin (ψ ) − sin (θ )cos(ψ ) cos(θ )

Taking the inverse transformation, we also have that

 AX   Ax2 
 A  = [θ ,ψ ]  A 
−1
(1.106)
 Y  y2 
 AZ   Az 
 2

where

cos(ψ ) − cos(θ )sin (ψ ) sin (θ )sin (ψ ) 


[θ ,ψ ] = [ψ ] [θ ] =  sin (ψ ) cos(θ )cos(ψ ) − sin (θ )cos(ψ )
−1 −1 −1
(1.107)
 0 sin (θ ) cos(θ ) 

Note that [θ ,ψ ] = [θ ,ψ ] since the matrices involved are all orthogonal.


−1 T

In addition, we will encounter a wide variety of applications in studying rigid body dynamics for
which it will be convenient to use a set of axes attached to the rigid body, called body axes and

87
C H A P T E R 1 : K I N E M A T I C S O F P A R T I C L E S A N D R I G I D B O D I E S

denoted in Figure 1.40 as axes x, y, and z. For these applications, the following Euler angle
transformation will be useful:

 Ax   AX   AX 
 A  = [φ ][θ ][ψ ] A  = [φ ,θ ,ψ ] A  (1.108)
 y  Y  Y
 Az   AZ   AZ 

where

[φ ,θ ,ψ ] =
 (cos(φ )cos(ψ ) − sin (φ )cos(θ )sin (ψ )) (cos(φ )sin(ψ ) + sin(φ )cos(θ )cos(ψ )) sin (φ )sin (θ )
(− sin (φ )cos(ψ ) − cos(φ )cos(θ )sin (ψ )) (− sin (φ )sin (ψ ) + cos(φ )cos(θ )cos(ψ )) cos(φ )sin (θ )

 sin (θ )sin (ψ ) − sin (θ )cos(ψ ) cos(θ ) 

(1.109)

Similarly, taken in the inverse order, we have

 AX   Ax 
 A  = [φ ,θ ,ψ ]−1  A  (1.110)
 Y  y
 AZ   Az 

where

[φ ,θ ,ψ ]−1 =
(cos(φ )cos(ψ ) − sin (φ )cos(θ )sin (ψ )) (− sin (φ )cos(ψ ) − cos(φ )cos(θ )sin (ψ )) sin (θ )sin (ψ ) 
(cos(φ )sin (ψ ) + sin (φ )cos(θ )cos(ψ )) (− sin (φ )sin (ψ ) + cos(φ )cos(θ )cos(ψ )) − sin (θ )cos(ψ )

 sin (φ )sin (θ ) cos(φ )sin (θ ) cos(θ ) 

(1.111)

As before, since all matrices involved are orthogonal, we have simply that [φ , θ ,ψ ] = [φ , θ ,ψ ] .
−1 T

88
2
Chapter

Kinetics of Particles

I n his work a particle is considered to be of infinitesimal size compared to all other macroscopic
dimensions of the problem so that its motion can be appropriately characterized by the motion of
its mass center, with the rotational effects about its mass center assumed negligible. For example, in
considering the orbital motion of the Earth about the Sun, the Earth may be considered of negligible
size compared to its distance from the Sun. However, in treating the spin and precessional motion of
the Earth itself, it becomes necessary to include consideration of its rotational inertia, which depends
on its mass distribution as well as its total mass.

By considering a particle to have infinitesimal size we are not implying that it must be treated as a
purely mathematical point mass, although this idealization occasionally proves useful, since this
would preclude consideration of such important physical concepts as stress or fluid drag which by
their very nature require a surface area. Thus, we include elemental volumes dV of a rigid body or
fluid continuum and their associated elemental masses dm in our concept of a particle to which
Newton’s laws apply.

In this and the following chapters we will study the relationships between the motion and the forces
causing the motion for both single particles and systems of particles. Furthermore, it will later be
shown that we can develop the equations of motion for rigid and elastic bodies by considering them to
be infinite sets of particles at constrained distances with respect to one another. The study of kinetics
of a general system of particles is one of the most important phases of Newtonian mechanics.

r d (m vr )
This leads us directly to consideration of Newton’s second law F = , for which problems fall
dt
into essentially two major classifications

89
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

r d (m vr )
A. Variable mass problems F=
dt

r r
B. Constant mass problems F = ma

Moreover, each of the categories (A) and (B) has the following subset of problems:

r
a) The motion is known, find the force F that causes the motion,

r
b) F is known, find the resulting motion,

c) Mixed value problems.

It is convenient to initially turn our attention to constant mass systems, since they are most frequently
encountered in engineering applications and the related analysis is considerably easier than for
variable mass systems which are discussed later. With respect to subset (a), the solution usually
follows a standard technique. We draw a free body diagram, write the equations of motion, use the
r
kinematic relationships developed in Chapter 1 to describe the acceleration a and solve the equations
r
directly for the applied force F .

r r r r r
In subset (b) the problem becomes m a = F (r , v , t ) where the force F is some given function of
position, velocity, time or possibly all of these. In any of these types of problems however, once the
principles of dynamics are mastered, the difficulty is not in writing the equations of motion, but rather
in solving them.

r r
It is important to note that while all problems in subset (b) may be directly solved by using F = m a ,
in many cases the most expedient method will entail an alternate principal developed from Newton’s
second law, such as the work-energy or impulse-momentum concepts that will be discussed later in
the text. For example, if:

r r r
1) F = F (r ) , (i.e., the force is a function of position such as gravitational attraction or a spring
force) in many instances one will find the work-energy relationships most convenient,

90
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

r r
2) F = F (t ) , (i.e., the force is a function of time) which would commonly occur in variable
speed mechanisms, the impulse-momentum relationships often prove to be the most useful,

r r r r r
3) F = F (v ) , (i.e., the force is a function of velocity), either a direct application of F = m a or
the work-energy methods prove to be best.

The point being emphasized here is that one should keep all of these principals in mind when
attacking a given problem, weighing the merits of each method for each particular case.

2.1 Direct Integration of Newton’s Second Law


I. Constant acceleration (constant force)

In this case, we may write

r
r F
a= (2.1)
m

from which we have in rectangular coordinates, for example,

Fx Fy Fz
&x& = ; &y& = ; &z& = (2.2)
m m m

r r r r
Integration of Eq. 2.2 subject to the initial conditions that r = r0 and v = v0 at t = 0 yields

Fx t 2
x= + x&0 t + x0
2m
Fy t 2
y= + y& 0 t + y0 (2.3)
2m
Fz t 2
z= + z&0 t + z0
2m

Perhaps the most notable example of the use of these equations is a freely falling particle for which the
only force is gravity, i.e. Fz = −m g , Fx = Fy = 0 .

91
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

II. Force as a function of velocity

In this instance, we write

r r r
F (v ) = m
dv
(2.4)
dt

for which the scalar equations in rectangular coordinates become

dvi
∫ dt = m ∫ F (v ) + C
i i
i i = x, y , z (2.5)

r
The solution of Eq. 2.5 will yield v (t ) , which can in turn be integrated in the form

r r r
∫ ∫
d r = v dt + C 2 (2.6)

r
to yield r (t ) . One of the more common problems encountered involving forces that are functions of
velocity is the case of fluid friction. Typically laminar viscous friction may be described as
r r r r
F = −C s& et , while turbulent viscous friction is given by F = −C s& s& et . It is interesting to note that

most friction forces as suggested by Newton can be described by

r 2
rn
F = ∑ Cn v (2.7)
n =0

for which the term n = 0 corresponds to Coulomb friction, n = 1 to laminar, and n = 2 to turbulent
viscous damping.

III. Force as a function of time

In this case, we have

r 1 r r
∫ m∫
d v = F (t ) dt + C1 (2.8)

A second integration yields

92
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

r r r
∫ ∫
d r = v dt + C 2 (2.9)

Time varying forcing functions are encountered in all variable speed mechanisms and in forced
vibrations problems.

IV. Force as a function of displacement

In order to show a general solution procedure for this case, it is necessary to consider the scalar
r r
equations of F = m a . Accordingly, we may write (i = x, y, z)

Fi (ri ) = m
dvi dv dr dv
= m i i = m vi i (2.10)
dt dri dt dri

from which we have

∫ F (r )dr = m ∫ v dv
i i i i i + Ci1 (2.11)

or

∫ F (r ) dr
1
= m vi + Ci1
2
i i i (2.12)
2

Equation 2.12 yields vi (ri ) , the velocity components in terms of the displacements, from which we
may write

dri
∫ dt = ∫ v (r ) + C
i i
i2 (2.13)

A common example of systems involving forces which are functions of displacements occurs in
studying vibratory phenomena in which the restoring mechanisms, such as elastic springs or buoyant
forces, depend on displacement.

While the function solution procedure is relatively simple in the above four cases, the evaluation of
these integrals may be quite a different story, with the exception of Case I (constant forces). With the

93
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

aid of computers, however, a numerical solution can always be accomplished if direct integration
becomes intractable.

2.2 Examples of Direct Integration of Newton’s Law

Example 2.1 Projectile Motion with Laminar Viscosity


r r
A projectile is fired with initial velocity v0 at t = 0 from the position r0 . Assuming air resistance to
be laminar viscous, determine the motion of the projectile as a function of time.

r r
v = s& et
r r m
F = c s& et

Figure 2.1: Projectile’s trajectory.

Solution: Referring to Figure 2.1, the scalar equations of motion are

m &x& = Fx = −C x& (a)

and

m &y& = Fy = −C y& − m g (b)

From Eq. (a) we have

dx& C
= − x& (c)
dt m

Separating variables and integrating, we have

94
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

dx&
x& C t

x&0 x&
= − ∫ dt
m 0
(d)

which yields

 x&  C
ln  = − t (e)
 x&0  m

or

C
− t
x& = x&0 e m
(f)

A further integration of Eq. (f) with respect to time gives the following equation for the horizontal
position of the projectile,

C
m − t
x=− x&0 e m + C1 (g)
C

m
For the given initial conditions, C1 = x0 + x&0 from which we conclude that
C

m  −C t 
x = x0 − x&0 ⋅  e m − 1 (h)
C  

Returning to Eq. (b), we have

dy& C
= − y& − g (i)
dt m

Separating variables and integrating as before, we find

t m y& dy&
∫ dt = − C ∫
0 y& 0 mg
(j)
y& +
C

or

95
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

 & mg 
y+
m  
t = − ln C  (k)
C  y& + m g 
 0 
 C 

Solving Eq. (k) for y& , we obtain

 mg mg
C
− t
y& = e m
 y& 0 + − (l)
 C  C

A final integration with respect to time yields

m − t mg mg
C
y = − e m  y& 0 + − t + C2 (m)
C  C  C

m mg
For the initial condition of y = y0 at t = 0 , we find C2 = y0 +  y& 0 +  from which we
C C 
conclude that

m g  − m t 
C
mg m
y = y0 − t −  y& 0 +  e − 1 (n)
C C C  

It is interesting to note from Eq. (l) for a freely falling body with an initial velocity y& 0 that as t → ∞ ,

mg
y& → . Thus, the velocity approaches a definite value known as the terminal velocity. This is, of
C
course, a considerably different result than would occur if we neglected air resistance in Eq. (b). In
that case, we would have found &y& = − g and y& = − g t + y0 , so that y& would be unbounded in time.

Example 2.2 Projectile Motion with Turbulent Viscosity

Determine the motion of the projectile in Example 2.1 considering the drag force to be turbulent
viscous drag, which is a more realistic approach for typical velocities encountered in ballistics
problems.

96
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

Solution: For this case, the scalar equations of motion are given by

( r r
m &x& = −C ⋅ x& 2 + y& 2 (et ⋅ i ) ) (a)

and

m &x& = −C ⋅ (x& 2 + y& 2 )(et ⋅ j ) − m g


r r
(b)

r r x& r r y&
Noting that et ⋅ i = and et ⋅ j = , Eqs. (a) and (b) become
x& + y&
2 2
x& + y& 2
2

2
 y& 
m &x& = −C x& 2
1+   (c)
 x& 

and

2
 y& 
m &y& = −C x& y& 1 +   − m g (d)
 x& 

It is immediately evident from equations (c) and (d) that ballistics problems very quickly lead to non-
linear differential equations for realistic drag forces. However, if we make the assumption that the
trajectory is shallow, we may make the approximation that

 y& 
  << 1 (e)
 x& 

Thus, Eqs. (c) and (d) may be written as

m &x& = −C x& 2 (f)

and

m &y& = −C x& y& − m g (g)

C
Separating variables in Eq. (f) where for simplicity we let K =
m

97
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

dx&
∫ x& 2
= − K ∫ dt + C1 (h)

or

1
− = − K t + C1 (i)
x&

1
which for x& = x&0 at t = 0 we have C1 = − . Equation (i) thus becomes
x&0

x&0
x& = (j)
K x&0 t + 1

A second separation of variables yields

x&0 dt
∫ dx = ∫ K x& 0 t +1
+ C2 (k)

from which we obtain

ln (K x&0 t + 1) + C2
1
x= (l)
K

For x = x0 at t = 0 , we have C2 = x0 . Thus, Eq. (l) becomes

m  C x&0 t 
x = x0 + ln + 1 (m)
C  m 

Returning to Eq. (g) into which we substitute Eq. (j), we obtain

K x&0 y&
&y& + = −g (n)
K x&o t + 1

In order to solve the differential equation, Eq. (n), it is convenient to make the change of variable
y& = p from which we obtain

98
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

K x&0 p
p& + = −g (o)
K x&o t + 1

Considering the complementary solution of Eq. (o), we may write

dp K x&0 dt
∫ p
= −∫
K x&0 t + 1
+ C3 (p)

or

ln( p ) = − ln(K x&0 t + 1) + ln(C3 ) (q)

From Eq. (q), the complimentary solution is given by

C3
y& c = (r)
K x&0 t + 1

For a particular solution of Eq. (n) we assume

C 4 ⋅ (K x&0 t + 1)
y& p = (s)
K x&0

Substituting Eq. (s) into Eq. (n), we find

C4
K x&0
+
K x&0
C4
(K x&0 t + 1) = − g (t)
K x&0 K x&0 t + 1 K x&0

or

g
C4 = − (u)
2

Accordingly, we have

C3 g (K x&0 t + 1)
y& = − (v)
(K x&0 t + 1) 2 K x&0

99
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

g
where at t = 0 , y& = y& 0 and then C3 = y& 0 + . Equation (v) then becomes
2 K x&0

gm
y& 0 +
2 C x&0 m g  C x&0 t 
y& = −  + 1 (w)
C x&0 t 2 C &0  m
x 
+1
m

A final integration of Eq. (w) with respect to time, subject to the initial condition that y = y0 at t = 0
yields

 gm  m  C x&0 t  m g t  C x&0 t 
y =  y& 0 +  ln + 1 −  + 1 + y0 (x)
 2 C x&0  C x&0  m  2 C x&0  2m 

Example 2.3 Elasticity Effect on Crane Locking Mechanism

An electromagnet supported by a massless cable of length L is used to haul scrap iron, with the total
suspended mass denoted in Figure 2.2 as m. The crane carriage O is considered to be massless and is
held near position A by a locking mechanism that is not perfectly rigid, causing the carriage to deviate
a distance q from A as the mass m swings back and forth. The elastic spring constant of the locking
mechanism is denoted by k and its restoring force is zero for q equal zero. Neglecting friction,
determine the motion of the pendulum mass m as a function of time, noting in particular how the
elasticity of the carriage locking mechanism influences the motion.

100
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

q
A

r x

O
k r
er
L

Figure 2.2: Electromagnetic crane system.

Solution: In terms of rectangular coordinates, the position of m is given by

x = q + L sin (θ ) (a)

and

y = − L cos(θ ) (b)

Two differentiations of Eqs. (a) and (b) with respect to time yield the components of acceleration of m
given by

&x& = q&& + L θ&& cos(θ ) − L θ& 2 sin (θ ) (c)

and

&y& = L θ&& sin (θ ) − L θ& 2 cos(θ ) (d)

The free body diagrams for the carriage and mass m are shown in Figure 2.3 and Figure 2.4.

101
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

T
θ
r

mg

Figure 2.3: Free body diagram for the suspended mass.

kq
θ T

Figure 2.4: Free body diagram for the carriage.

A brief reflection on these figures readily shows that it is advantageous to write the equations of
r r r r
motion in the i and eθ directions rather than the i and j directions when considering the motion of
m.

Thus, since

r r
aθ = eθ ⋅ (&x& i + &y& j ) = L θ&& + q&& cos(θ )
r r
(e)

the equations of motion become

Fx = m &x&
( )
(f)
− T sin (θ ) = m ⋅ q&& + L θ&&cos(θ ) − L θ& 2 sin (θ )

and

102
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

Fθ = m aθ
( )
(g)
− m g sin (θ ) = m ⋅ L θ&& + q&& cos(θ )

Noting from Figure 2.4 that T sin (θ ) = k q since the cart is massless, we may simplify Eqs. (f) and (g)
as follows

q&& + L θ&&cos(θ ) − L θ& 2 sin (θ ) + q = 0


k
(h)
m

and

q&& cos(θ ) + L θ&& + g sin (θ ) = 0 (i)

A simultaneous solution of Eqs. (h) and (i) is in general very difficult since they are non-linear
differential equations. However, if we restrict our attention to small motions, then we may set
sin (θ ) = θ and cos(θ ) = 1 , neglecting higher order terms. Thus, we find

k
q&& + L θ&& + q = 0 (j)
m

and

q&& + L θ&& + g θ = 0 (k)

Subtracting Eqs. (j) and (k), we obtain

k
q = gθ (l)
m

Therefore, Eq. (k) becomes

g
θ&& + θ =0 (m)
mg
L+
k

A solution of Eq. (m) may be sought in the form θ = A e p t . Substituting θ and its derivative into Eq.
(m) we have

103
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

(
Ae pt p2 + Ω2 = 0 ) (n)

where

g
Ω2 = (o)
mg
L+
k

From Eq. (n), we find the roots p1, 2 = ±i Ω and the solution for θ becomes

θ = A1 e i Ω t + A2 e −i Ω t (p)

or using Euler’s identity, e i x = cos( x ) + i sin ( x ) , we find

θ = C1 cos(Ω t ) + C 2 sin (Ω t ) (q)

The constants C1 and C 2 may be evaluated from initial conditions which in general are θ = θ 0 and

θ&0
θ& = θ&0 at t = 0 . For these conditions, we find C1 = θ 0 and C 2 = so that the general solution for

θ becomes

θ&0
θ = θ 0 cos(Ω t ) + sin (Ω t ) (r)

It is clear from Eq. ® that the motion of θ is harmonic with an amplitude which depends on the initial
displacement θ 0 ,initial velocity θ&0 , and Ω , which is a physical constant of the system. Further, we

see that Ω is the circular frequency of motion ( rads ) for which the natural frequency and period of
motion are

f =

=
1
2π 2π
g
mg
[ ]
cycles
s (s)
L+
k

and

104
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

T=
1 2π
f
=

[ ]
s
cycle (t)

Note that by taking the spring constant k = ∞ , we have in effect reduced the problem to that of an
ideal pendulum oscillating about a fixed pivot point for which the natural frequency is well known to

1 g cycles
be and independent of the mass m. Thus, the main effect of the elasticity in the carriage

s
l
locking mechanism is to reduce the natural frequency of the system according to Eq. (s).

Two final points should be noted from the example. First, Eq. (r) may be written as

θ = Q sin (Ω t + φ ) (u)

where

θ&0
Q = θ0 + (v)

and

 θ0 
φ = tan −1  
()
 θ&0
 Ω


(w)

which may readily be verified by expansion. Equation (u) really verified that the solution given by
Eq. (r) is harmonic (i.e., the sum of two harmonics of the same frequency is harmonic). This is not
true for the sum of two harmonics of different frequencies.

Finally, we note that if a motion is harmonic (i.e., X = X 0 sin (λ t ) ), then

X&& = − X 0 λ2 sin (λ t ) = −λ2 X (x)

or

X&& + λ2 X = 0 (y)

105
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

Equation (y) tells us that any equation of the form given by Eq. (y) will yield a harmonic solution,
called harmonic motion, which has a circular frequency λ rads . Thus, to determine the natural
frequency of any undamped single degree of freedom system, we need only to put the equation of
motion in standard form and note the coefficient of X in Eq. (y) whereby the natural frequency of the
λ
system is given by f = .

Example 2.4 External Force on Crane Locking Mechanism

Determine the motion of the system discussed in Example 2.3 if the carriage is subjected to an
r r
external force F = F0 sin (ω t ) i while in the locked position.

q
A

r x

O F0 sin (ω t )
k r
er
L

Figure 2.5: Electromagnetic crane system with forcing applied.

Solution: It is clear that Eqs. (f) and (g) from Example 2.3 will again result after applying Newton’s
law to m, yielding as before

(
− T sin (θ ) = m ⋅ q&& + L θ&&cos(θ ) − L θ& 2 sin (θ ) ) (a)

and

(
− m g sin (θ ) = m ⋅ L θ&& + q&& cos(θ ) ) (b)

106
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

In this case however, since the carriage is massless compared to m, we have


T sin (θ ) = k q − F0 sin (ω t ) . Substituting T sin (θ ) into Eqs. (a) and (b) and again considering small
oscillations we find

q&& + L θ&& + q = 0 sin (ω t )


k F
(c)
m m

and

q&& + L θ&& + g θ = 0 (d)

Subtracting Eqs. (c) and (d), we have

q − g θ = 0 sin (ω t )
k F
(e)
m m

Utilizing Eq. (e) and its derivatives we find

θ&& + Ω 2 θ = E sin (ω t ) (f)

where

ω 2 F0 g
E= ; Ω2 = (g)
 mg 
L+
mg
k ⋅ L + 
 k  k

The complimentary solution of Eq. (g) as given before is

θ = C1 cos(Ω t ) + C 2 sin (Ω t ) (h)

For a particular solution, we try θ = A sin (ω t ) since the right hand side of Eq. (f) is harmonic.
Equation (f) thus becomes

− Aω 2 sin (ω t ) + Ω 2 A sin (ω t ) = E sin (ω t ) (i)

from which we find

107
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

E E
A= =
Ω −ω2
2
 ω2 
Ω 2 1 − 2 
 Ω 
ω 2 F0
=
 mg  g  ω2 
k ⋅ L + ⋅ ⋅ 1 − 2 
 k   mg   Ω 
L+ 
 k 

ω 2 F0
A= (j)
 ω2 
k g ⋅ 1 − 2 
 Ω 

Therefore, the general solution of Eq. (f) is given by

ω 2 F0
θ = C1 cos(Ω t ) + C 2 sin (Ω t ) + sin (ω t ) (k)
 ω2 
k g ⋅ 1 − 2 
 Ω 

It is clear from Eq. (k) that if ω = Ω then θ is unbounded, which is called resonance of the system,
and is independent of the initial conditions. Therefore, the general equation of motion for any
harmonically excited, undamped single degree of freedom system is given by

&& + Ω 2 Q = F sin (ω t + φ )
Q (l)
0 0

from which it follows that resonance occurs whenever the external forcing frequency ω equals the
natural frequency Ω , where physically ω is the frequency at which the system would vibrate if
plucked and allowed to vibrate freely.

Example 2.5 Mass on Spring in Rotating, Laminar Viscous Slot


r r
The machine shown in Figure 2.6 rotates at an angular velocity ω and angular acceleration ω& about a
vertical axis at O. The accelerometer mass m is constrained to move in a slot parallel to the y body
axis and is acted on by a spring of spring constant k which is unstretched for y = 0 and by laminar

108
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

viscous damping of damping constant c. Determine (a) the equation of motion of m in the y direction
and discuss its solution for ω& = 0 , and (b) the normal force acting on m in the x direction.

m ω&
O y
x
k

Figure 2.6: Mass in rotating slot.

Solution:

a) Equation of Motion

The velocity of mass m referred to the rotating x, y, z frame is

r r r r r
v = v0 + ω × ρ + ρ& r
r
( )
= 0 + ω k × (a i + y j ) + y& j
r r r

or

r r r
v = − w y i + (ω a + y& ) j (a)

Similarly, the acceleration of m is given by

r r r r r r r r r r
a = a 0 + ω × (ω × ρ ) + ω& × ρ + ρ&& r + 2 ω × ρ& r
r
( ) r r r r r r r
( )
= 0 + ω k × (− ω y i + ω a j ) + ω& k × (a i + y j ) + &y& j + 2 ω k × ( y& j )
r

or

109
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

( ) ( )
r r r
a = − ω 2 a − ω& y − 2 ω y& i + − ω 2 y + ω& a + &y& j (b)

Thus, applying Newton’s Law to the component of motion in the y direction, we have

(
− k y − c y& = m ⋅ − ω 2 y + ω& a + &y& ) (c)

which reduces to

(
m &y& + c y& + k − m ω 2 y = −m a ω& ) (d)

Note that the angular acceleration ω& acts essentially as a forcing function on the spring mass system
and that the m ω 2 term reduces the effective spring constant of the accelerometer system.

The solution of Eq. (d) is very complex unless ω& = 0 for which it reduces to a simple homogeneous
differential equation with constant coefficients given by

(
m &y& + c y& + k − m ω 2 y = 0 ) (e)

From mathematics we recall that the homogeneous solution for any differential equation with constant
coefficients is of the form

y = Ae pt (f)

Thus, substituting Eq. (f) into (e), we obtain

[m p 2
(
+ c p + k − mω 2 A e p t = 0 )] (g)

from which we can conclude that for nontrivial solutions it is necessary that

(
m p 2 + c p + k − mω 2 = 0 ) (h)

which is called the characteristic equation.

The characteristic equation is a polynomial equation of second order with two characteristic roots p1
and p2 given by

110
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

p1, 2 =
− C ± C 2 − 4 m ⋅ k − mω 2 ( ) (i)
2m

C k
Letting γ = and ω n = , Eq. (i) becomes
2

2m m

p1, 2 = −γ ± γ 2 − ω n − ω 2 ( 2
) (j)

Thus, the general solution for the motion of the accelerometer mass for ω& = 0 is

y = e −γ t  A1 e
(
γ 2 − ωn 2 −ω 2 t ) + A e− ( )
γ 2 − ωn 2 −ω 2 t
(k)
 2


The form of solution for Eq. (k) depends primarily on the sign of the quantity under the square root
( )
sign; namely whether γ 2 − ω n − ω 2 is greater, less than, or equal to zero.
2

( )
Case I: ωn − ω 2 > γ 2 ; Underdamped System
2

For this case, the radical is imaginary and p1, 2 can be written as

p1, 2 = −γ ± i (ω n
2
)
− ω2 − γ 2 (l)

Recalling also that eiθ = cos(θ ) + i sin (θ ) , we may write the solution for y as

 
( 
)
y = e −γ t  B1 sin  ω n − ω 2 − γ 2 ⋅ t  + B2 cos ωn − ω 2 − γ 2 ⋅ t 
2


2


( ) (m)

or in terms of an amplitude C and a phase angle φ as


2
(
y = C e −γ t sin ωn − ω 2 − γ 2 ⋅ t + φ 

) (n)

which is schematically graphed in Figure 2.7.

111
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

C e −γ t
C
C sin(θ )

Figure 2.7: Underdamped response of the mass.

( )
Case II: ω n − ω 2 < γ 2 ; Overdamped System
2

For this case, we see that both p1 and p2 are real, negative quantities denoted by α1 and α 2 where

(
α1 = γ − γ 2 − ω n 2 − ω 2 ) (o)
α2 = γ + γ − (ω −ω )
2 2 2
n

Thus, the solution y for the overdamped case is

y = A1 e −α1 t + A2 e −α 2 t (p)

which is graphed in Figure 2.8.

112
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

y
y = A1 e−α1 t + A2 e−α 2 t
A2 e−α 2 t
A1 e−α1 t

Figure 2.8: Overdamped response of the mass.

( )
Case III: ωn − ω 2 = γ 2 ; Critical Damping
2

For the case of critical damping, we see that the two roots p1 and p2 become the same, which
according to the theory of differential equations, requires the form of the solution to Eq. (e) to be

y = A1 e p t + A2 t e p t (q)

for which we have

y = A1 e −γ t + A2 t e −γ t (r)

Critical damping represents the smallest possible value of the damping parameter γ for an aperiodic
solution (i.e., y always ≥ 0 ). The shape of the solution curve is thus similar in form to Figure 2.8.

b) Normal Force

The acceleration of mass m in the rotating x direction is given in Eq. (b) as

a x = −ω 2 a − ω& y − 2 ω y& (s)

r
Therefore, the force N exerted on m due to the slot is

113
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

r
N = − m ⋅ (ω 2 a + ω& y + 2 ω y& )i
r
(t)

For example, if we set ω& = 0 , we have from part (a) that

y = A1 e p1 t + A2 e p2 t (u)

and

y& = A1 p1 e p1 t + A2 p2 e p2 t (v)

r
Thus, for constant angular velocity ω , the normal force N becomes

( )
r
N = − m ⋅ ω 2 a + 2 ω A1 p1 e p1 t + 2 ω A2 p2 e p2 t (w)

r
It is important to note that the time dependent component of N , which is important in the wearing
characteristics of the slot surface, is directly due to Coriolis acceleration.

2.3 Newton’s Second Law for a System of Particles


Next consider the application of Newton’s laws to a system of particles. Let n be the number of
particles in the system and let mi be any representative particle. Referring to Figure 2.9, we designate
r r
Fi as the resultant external force on mi and f i , j as the internal force on mi due to any other particle

of the system m j ( j ≠ i ) . The internal forces may result, for example, from gravitational attraction,

electromagnetic attraction, contact forces, surface stresses, etc. Also, according to Newton’s third law,
r r
let f j ,i = − f i , j , which states that the internal forces occur in equal and opposite pairs. However, this
r r
does not necessarily imply that f i , j and f j ,i are collinear, since there are some unusual exceptions

such as certain electromagnetic forces.

r
Each particle mi is located by a position vector ri with respect to an inertial xyz frame as shown in
r
Figure 2.9. Furthermore, the mass center of the system is located by the vector rc where

114
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

r n
r
m rc = ∑ mi ri (2.14)
i =1

n
for which m = ∑ mi . Point P is considered to be an arbitrary reference point for the system and each
i =1
r
particle is located by the position vector ρ i relative to P as shown in Figure 2.9. In addition, the
r
vector ρ c locates the mass center C relative to point P.

y
r
r fj
r r f j ,i
fi , j mj
fi
mi r
rj
r
r r ρj
ri ρi
r P r
rP r ρc
rc C (mass center)
x

mn

Figure 2.9: Vector definitions for forces acting on arbitrary masses.

Restricting attention to a constant mi , we can apply Newton’s Law to each particle of the system to
form n equations of motion given by

r n r
r
Fi + ∑ f i , j = mi &r&i ; i = 1,2, K , n (2.15)
j =1
j ≠i

In general, this results in n simultaneous differential equations of second order which can be solved
either analytically or numerically to describe the motion of each individual particle of the system.

115
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

r r
However, since the internal forces f i , j = − f j ,i , it is easy to recognize that by adding all n of these Eqs.

2.15 together, a considerable simplification results, since the internal forces in the sum cancel in equal
and opposite pairs. The resulting equation is

n r n
r
∑ F = ∑ m &r&
i =1
i
i =1
i i (2.16)

r
Noting that the term on the left side of Eq. 2.16 is the resultant FR of all external forces acting on the
system of n particles, we may rewrite Eq. 2.16 as

r n
r
FR = ∑ mi &r&i (2.17)
i =1

Furthermore, the right side of Eq. 2.17 has a striking similarity to the definition of the mass center
given by Eq. 2.14. By restricting attention to constant mass systems, Eq. 2.14 can be differentiated
twice to yield

r n
r
m r&&c = ∑ mi &r&i (2.18)
i =1

Substituting Eq. 2.18 into 2.17, we obtain the well known principle for the motion of the mass center
of any system of particles written as

r r
FR = m &r&c (2.19)

Thus, by analysis, we have extended Newton’s Laws for a single particle to a system of particles,
resulting in the physical conclusion that the acceleration of the mass center of any general system of
r
constant mass particles is proportional to the resultant external force FR acting on the system. An
interesting result of this law applied to the universe which is considered to consist entirely of celestial
bodies governed only by their mutual gravitational attraction for one another is that the mass center of
the universe (wherever it is) moves through space with a constant velocity. Unfortunately however,
this tells us nothing about the motion of each individual celestial body; only the motion of the

116
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

system’s mass center. In order to describe the motion of each celestial body, one must refer back to a
system of equations as given by Eqs. 2.15.

It is of further importance that we can also write Eq. 2.18 as

r r r r
( r
) r
n n n
m &r&c = ∑ mi &r&i = ∑ mi &r&P + ρ&&i = m &r&P + ∑ mi ρ&&i (2.20)
i =1 i =1 i =1

r r r
since ri = rP + ρ i . Thus, we can rewrite Eq. 2.19 in the alternate form of

r r n
r
FR = m &r&P + ∑ mi ρ&&i (2.21)
i =1

r
where &r&P is the acceleration of some conveniently selected reference point whose motion is either
r r r
( )
known or required in the analysis and ρ&&i = r&&i − &r&P represents the inertial (or total) acceleration of

particle mi relative to point P. An important precaution worth noting is that all differentiations are
taken with respect to inertial space which is represented by the xyz frame in this analysis.

Example 2.6 Force on Weather Vane Shaft

A weather vane, which may be dynamically modeled by a system of four masses on the ends of
r
weightless rigid bars, as shown in Figure 2.10, is spinning with an angular velocity ω1 relative to a
main frame assembly about axis A-A. The main frame assembly shown schematically as a circular
r
platform is in turn spinning about a fixed vertical axis with an angular velocity ω 2 . Determine the
resultant force transmitted by the weather vane shaft A-A for

a) Case I; m1 = m2 = m3 = m4 = m

b) Case II: m1 = m ; m2 = 2 m ; m3 = 3 m ; m4 = 4 m

117
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

z m1
A
m2 r
r r
r
ωz r m4
m3 r
ω1 D

O
y
A

R
Figure 2.10: Weather vane on rotating platform.

Solution:

a) It is clear in Case I that point B is the center of mass of the weather vane and thus Eq. 2.17 is
certainly most convenient. Accordingly, using the x, y, z body axes, we write

r r r r r
Fx i + Fy j + Fz k − 4 m g k = 4 m &r&B (a)

r r r r
where F = Fx i + Fy j + Fz k is the force exerted by shaft A-A on the weather vane. Noting that

B 2
r
2
r
[r r
( 2
r
&rr& = ω k × ω k × R j + D k = − R ω 2 j )] (b)

then the components of force exerted on axis A-A are

Fx = 0
Fy = 4 m R ω 2
2
(c)
Fz = −4 m g

b) In Case II however, the center of mass, which of course could not be located is not
immediately apparent. By using Eq. 2.21 and selecting B as a reference point, we may write

r r r r 2 r r r r r
Fx i + Fy j + Fz k − 10 m g k = −10 m R ω 2 j + m1 ρ&&1 + m2 ρ&&2 + m3 ρ&&3 + m4 ρ&&4 (d)

118
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

From kinematics, we can determine the difference between inertial acceleration of each mass mi and
r r r
the inertial acceleration of the reference point B (i.e., ρ&&i = &r&i − &r&B ) to obtain

r r
ρ&&1 = r ⋅ (ω 1 + ω 2 )2 i
r r
ρ&& 2 = r ⋅ (ω 1 + ω 2 )2 j
r r (e)
ρ&& 3 = − r ⋅ (ω 1 + ω 2 )2 i
r r
ρ&& 4 = − r ⋅ (ω 1 + ω 2 )2 j

Thus, from Eq. (d) we have

r r r r 2 r
Fx i + Fy j + Fz k − 10 m g k = −10 m R ω 2 j
2 r 2 r
+ r ⋅ (ω1 + ω 2 ) i + 2 r ⋅ (ω1 + ω 2 ) j (f)
2 r 2 r
− 3 r ⋅ (ω1 + ω 2 ) i − 4 r ⋅ (ω1 + ω 2 ) j

Therefore, the force components exerted on shaft A-A for this case are

Fx = 2 m r ⋅ (ω1 + ω 2 )
2

Fy = 2 m r ⋅ (ω1 + ω2 ) + 10 m R ω2
2 2
(g)
Fz = −10 m g

This is a good example for the student to verify by directly applying the principle of the motion of the
mass center given by Eq. 2.19.

Example 2.7 Force of Falling Chain

A chain of length l and total mass m as shown in Figure 2.11 is dropped from rest from the initial
position for which the lower end of the chain just barely touches the platform of a scale. Determine
the reading on the scale dial as a function of y as the chain falls. Assume no rebound of the chain links
and assume that the height of the mass center of the chain links that have come to rest on the scale is
negligible compared to the length of the chain.

119
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

Figure 2.11: Chain falling onto scale platform.

Solution: The height of the mass center C of the entire chain above the scale platform is given by

m 
m yc =  y (0 ) + (l − y )
m (l − y ) (a)
l  l 2

∴ yc =
(l − y)
2
(b)
2l

Consequently, the motion of the chain’s mass center is given by

− (l − y ) y&
y& c = (c)
l

and

&y&c = −
(l − y ) &y& + y& 2 (d)
l l

A free body diagram of the entire chain is shown in Figure 2.12 where N is the resultant reaction of
the scale platform on the chain. Consequently, N is the reading of the scale dial.

120
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

y& c , &y&c
C
W = mg
yc

Figure 2.12: Free body diagram of the chain.

Thus, from the principle of the motion of the mass center given by Eq. 2.19, we have

 (l − y ) &y& y& 2 
N − m g = m yc = m ⋅  −
&& +  (e)
 l l 

It is also apparent from Figure 2.12 that the force N acts only on those links that have already come to
rest plus the individual link that is instantaneously being brought to rest by striking the platform.
Therefore, the effect of N is not felt on the links that are still in free fall, since a chain cannot transmit
compression through itself. For example, considering the uppermost link taken as mass m which
falls at the rate y& and &y& acted upon only by its own weight, we may write Newton’s Second Law for
this single particle yielding

m &y& = m g (f)

or

&y& = g (g)

Eq. (g) simply indicates that the links of the chain are in free fall, accelerating at the acceleration of
gravity, which may have been obvious to some.

Thus, integrating Eq. (g), we have

121
C H A P T E R 2 : K I N E T I C S O F P A R T I C L E S

y& = g t

and

1 2
y= gt (h)
2

Consequently,

y& = 2 g y (i)

Substituting for &y& and y& in terms of y into Eq. (e), we obtain

 (l − y ) g 2 g y 
N = m g + m ⋅ − + (j)
 l l 

which reduces to

m
N =3 g y = 3w (k)
l

where w is the weight of the portion of the chain that has come to rest on the scale at any time t.

122
3
Chapter

Work-Energy Methods for Particles

A lthough one can in general proceed to solve any problem by directly applying Newton’s laws
to relate forces to the resulting accelerations of a system it is nevertheless very useful to
consider two alternate forms of this basic method. First, work-energy principles will be developed in
this chapter which are particularly useful for problems in which one wants to relate a change in
velocity to a change in position of the system. Then in the next chapter we will develop impulse-
momentum methods that prove to be especially useful when it is desired to relate a change in velocity
to the time interval involved. These latter methods usually are much simpler to apply to their
respective classes of problems than would be possible by directly applying Newton’s second law.

3.1 Work-Energy Principle for a Single Particle


Referring to Figure 3.1, let the path of a particle of mass m be denoted by curve C and consider it to be
r r
acted upon by the resultant force F at its instantaneous inertial position r .

123
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

m r r r
1
r
et v = r&
r v1 r
r1 r dr
r r 2
F
r
x v2
O
x

Figure 3.1: Particle acted on by a force.

From Newton’s Second Law we have

r r
r dv dr&
F =m =m (3.1)
dt dt

r
We can advantageously modify the form of Eq. 3.1 by taking the dot product of each side by dr and
integrating as the particle moves along its path from position 1 to 2. Therefore,

r r r
2 dr&
2 r
∫1
F ⋅ dr = m ∫
1 dt
⋅ dr (3.2)

r r
or since dr = r& dt , we have that

∫1
2 r r 2 r r m 2 r r m 2
F ⋅ dr = m ∫ dr& ⋅ r& = ∫ d r& ⋅ r& = ∫ d r& 2
1 2 1 2 1
( ) ( )

Thus,

∫1
2 r r 1
2
2 1
2
2 1
F ⋅ dr = m r&2 − m r&1 = m ⋅ v2 − v1
2
2 2
( ) (3.3)

124
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

The left side of Eq. 3.3 is defined as the work done, Wk , by the resultant force acting on m as the

r r 1
particle moves from r1 to r2 along path C. The quantity m v 2 is defined as the kinetic energy of the
2
particle and is denoted by T. Thus, Eq. 3.3, called the work-energy principle, may be written

Wk = T2 − T1 = ∆T (3.4)

which states that the work done by the forces acting on a particle it moves from position 1 to 2 equals
the change in its kinetic energy associated with those positions.

Two important points should be particularly noted with reference to Eq. 3.4. First, the work done,
r r
Wk , in general depends on the path taken from r1 to r2 and is referred to as a path function, while the
r r
change in kinetic energy depends only on the instantaneous values of velocity at r1 and r2 , and is thus
r r
a point function independent of the path taken between r1 and r2 . Furthermore, both work and kinetic
energy are scalar quantities.

Finally, we can conveniently show that Eq. 3.4 can be written in component form for rectangular
components, although this is not possible in general for more complex coordinate systems (i.e., polar,
r
path, or spherical, etc.). To show this, consider the dot product of each side of Eq. 3.1 by dx i ,
written as

r r r r
dv
F ⋅ dx i = m ⋅ dx i (3.5)
dt

r r r r r r r r
Thus, writing F = Fx i + Fy j + Fz k and v = x& i + y& j + z& k , we obtain

dx&
Fx dx = m dx
dt

and since dx = x& dt , we have

Fx dx = m x& dx& (3.6)

125
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Integrating Eq. 3.6 between positions 1 and 2, we can conclude for the x components, and similarly
for the y and z components, that


1
2 1
Fx dx =
2
m⋅ (x& 2
2
− x&1
2
)
2 1
∫1 Fy dy = 2 m ⋅ (y& 2
2
− y&1
2
) (3.7)
2 1
∫1 Fz dz = 2 m ⋅ (z& 2
2
− z&1
2
)
The significance of Eqs. 3.7 should not be underestimated, since they state that one can apply the
work-energy principle either by setting the total work done equal to the total change in kinetic energy
of the particle, or by simply treating the x, y, z components of these quantities. It becomes apparent,
for example, that if the component of force Fx is zero during the entire state of motion, then the x
component of velocity, x& , remains constant. However, this does not imply that the work-energy
principle is a vector equation, since these relationships are not possible for most sets of coordinates
and are quite uniquely useful for rectangular coordinates only.

Example 3.1 Crate Stopped by Bumper


ft
The 322 lb. crate is pushed down a ramp with an initial velocity of 10 s . The crate is to be stopped
by a mechanical bumper which is initially compressed by cables which produce an initial compressive
force of 500 lbs. in the bumper spring. Consider the bumper to be of negligible mass and take the
coefficient of friction as 0.2 from A to B and zero below point B.

a) Design the spring constant k so that the maximum bumper deflection is limited to 1 foot for
this case.

b) Determine the force acting in the normal direction on the crate immediately before striking the
bumper.

126
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

y r
i′ x′

A
r
y′ j
r 20'
j′
45°

Cables
1'
10'
D B
r
k C i
x

Figure 3.2: Crate sliding towards bumper.

Solution:

a) Determine k

The principle of Wk = ∆T according to Eq. 3.4 can be applied between the initial position A and the
final position D, where the system has come to rest after compressing the spring 1 foot. The change in
kinetic energy is given by

∆T = TD − TA = 0 −
1
2
(10 slugs ) 10 ( ft 2
s ) = −500 ft ⋅ lb (a)

The total work done is the sum of the work done by each force acting on the crate during the motion
r r r
of the system. First, the weight W = −W j = −322 j does work equal to

r r r r r
dWk w
= F ⋅ dr = (− 322 j )⋅ (dx i + dy j ) = −322 dy (b)

127
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

(− 322)dy = 322( y A − yD )
yD
∴Wk w
=∫ (c)
yA

where

y A = 20(0.707 ) + 10(1 − cos(45°)) = 17.1 ft


yD = 0 ft (d)
∴Wk w
= (322 lb )(17.1 ft ) = 5500 ft ⋅ lb

The work done by the normal force exerted on the crate by the ramp is zero, since this normal force is
r r
always at right angles to the notion of the crate. However, the frictional force F = µ N i ′ does work
on the crate from A to B. This component of work is given by

r r
= ∫ (µ N i ′)⋅ dx′ i ′ = ∫ µ N dx′
0 ft
Wk f
(e)
20 ft

where N = (322 lb )(0.707 ) = 228 lb and µ = 0.2 from A to B. Therefore,

Wk f
= −(0.2 )(228 lb )(20 ft ) = −910 ft ⋅ lb (f)

Finally, the work done by the spring stopping the crate can be determined by integrating the spring
force times the differential displacement of the spring. It is convenient here to introduce the auxiliary
variable ξ here as shown in Figure 3.3.

1'
ξ
D
C x
Fs = 500 + k ξ

Figure 3.3: Deflected system

The spring force Fs acting on the crate is given by

128
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Fs = 500 lb + k ξ (g)

so that the differential work is given by

r r r r
F ⋅ dr = (500 lb + k ξ ) i ⋅ (− dξ ) i = −(500 lb + k ξ ) dξ (h)

Therefore, the work done by Fs as the crate moves from position C to D is given by

(500 lb + k ξ ) dξ = − 500 ft ⋅ lb + 1 k 


1 ft
Wk s = − ∫ (i)
0 ft
 2 

Therefore, the total work done by all forces acting on the crate is given by

1
Wk = Wk w
+ Wk f
+ Wk s = 4900 − k (j)
2

Equating Wk = ∆T = −500 , we obtain the spring constant

k = 9180 lbft (k)

b) Determine normal force

In order to determine the normal force exerted by the ramp on the crate at c, it becomes necessary to
determine the velocity of the crate at c to determine its normal acceleration at that point. Referring to
Figure 3.4, we see by summing forces normal to the path at point C that

2
vc
N − W = m &y& = m (l)
ρ

∴ N = 322 + vc
2
(m)

129
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

2
vc
&y& = an =
ρ = 10′ 10′

W = 322lb.
r
vc
x

Figure 3.4: Force diagram of sliding crate at point C, prior to impact.

The velocity vc at point C can be determined by the work-energy principle as before. Therefore, since
the work done by friction and the weight are the same as in part (a), we may write

Wk w
+ Wk f
= TC − TA (n)

Inserting the appropriate data into Eq. (n), we obtain

5500 − 910 =
1
2
(10 slugs )(vc )2 − 1 (10 slugs ) 10
2
( s )
ft 2
(o)

Solving for vc , we find

vc = 31.9 ft
s (p)

Therefore, according to Eq. (m), the normal force acting on the crate immediately before it strikes the
bumper is

N = 1340 lb (q)

130
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

3.2 Conservative Force Systems


In discussing the work-energy principle in the previous section, it was noted that the change in kinetic
energy depends only on its values at the end points of the displacement and was thus called a point
function. For certain types of force systems, it can be shown that work done is also independent of the
path taken from position 1 to 2 in the field, and therefore is a point function of the limits of integration
only. For this class of force systems, called conservative forces, the work-energy principle can be
shown to depend only on the values of the system parameters evaluated at the endpoints of motion,
which considerably simplifies certain analyses by means of conservation of energy concepts.

Now in order for the work integral

r r
ωk = ∫ F ⋅ dr
2
(3.8)
1

to depend only on the limits of integration (i.e., to be independent of the path taken from 1 to 2) it is
r r
necessary and sufficient that F ⋅ dr be an exact differential, which we will denote by dφ .
Mathematically, this may be written as

r r
F ⋅ dr = dφ (3.9)

which when substituted into Eq. 3.8 yields

r r
ωk = ∫ F ⋅ dr = ∫ dφ = φ2 − φ1
2 2
(3.10)
1 1

Thus, for this class of force systems, called conservative forces, the work done is a scalar function of
the values of φ , called a potential (or work) function, evaluated at the limits of integration. In this
way, work equals the change in the potential function associated with positions 1 and 2 where it is
further noted that φ is a function of position only.

It is also important to note that the work done by a conservative force evaluated around a closed path
is zero and is written

r r
ω k = ∫ F ⋅ dr = ∫ dφ = 0 (3.11)

131
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

It is seen from Eq. 3.10 that in physical applications only, the change in potential, φ2 − φ1 = ∆φ , is of
importance and as such we can only determine the potential function itself within an arbitrary
constant. One convenient method for treating this constant is to consider the potential function φ to
be measured relative to some arbitrary reference point. Then, for various selections of reference
points, the potential functions φ associated with each will differ, but only by a constant value
dependent on the alternate choices of reference position.

This can be explained mathematically by referring to Figure 3.5, and noting that for a conservative
force system we may write

r r r r r r
ωk = ∫ F ⋅ dr = ∫ F ⋅ dr + ∫ F ⋅ dr = φ2 − φ1
2 R 2
(3.12)
1 1 R

where R denotes the reference position. Interchanging the limits of integration, we obtain

1
P
R
r 2
r r1 r
rR r
R′ r
r2
r
rR′
O y

Figure 3.5: Path through a potential.

2 r r 1r r

R
F ⋅ dr − ∫ F ⋅ dr = φ2 − φ1
R
(3.13)

Thus, we can define the potential at any point P in the field as the work done in going from the
reference position R to point P, written mathematically as

r r
φ = ∫ F ⋅ dr
P
(3.14)
R

132
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Note that if point R′ were selected as a reference point rather than R, then the new potential function
becomes

r r
φ ′ = ∫ ′ F ⋅ dr
P
(3.15)
R

This second description of the potential function is related to the first by the equation

φ ′ = φ + C1 (3.16)

R r r
where the arbitrary constant C1 = ∫ F ⋅ dr .
R′

r r
In order for F ⋅ dr in Eq. 3.9 to be an exact differential dφ , it can be shown that it is necessary and
r r r r r
sufficient that ∇ × F = 0 . Or, in other words, a force F is conservative if and only if ∇ × F = 0 .

Proof:

1) Consider to be an exact differential. Then working in rectangular coordinates, we may write

∂φ ∂φ ∂φ
( )
r r r r
F ⋅ dr = dφ = dx + dy + dz = ∇φ ⋅ dr (3.17)
∂x ∂y ∂z

r ∂ r ∂ r ∂ r
where ∇ = i + j + k . Thus, we see from Eq. 3.17 that
∂x ∂y ∂z

r r
F = ∇φ = grad (φ ) (3.18)

r r r
Finally, we can conclude from Eq. 3.18 that if F is conservative, (i.e., if F ⋅ dr = dφ is a perfect
differential) then

( )
r r r r
∇× F = ∇×∇ φ = 0 (3.19)

r r r
2) Conversely, consider that for a particular force F it has been shown that ∇ × F = 0 . Then,
according to Stokes theorem,

133
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

( )
r r r r
∫ F ⋅ dr = ∫∫ × F ⋅ dS = 0

S
(3.20)

Referring to Figure 3.6, and considering two different paths from position 1 to 2, denoted by C and
C ′ , we may write

r r 2 r r
∫ ⋅ dr = ∫ F ⋅ dr
F
1
( ) − ∫ (Fr ⋅ drr )
C
2

1
C′ =0

2 r
( ) = ∫ (F ⋅ drr )
r r
∴ ∫ F ⋅ dr
2
C C′ (3.21)
1 1

Since the work done in going from position 1 to 2 is independent of path, we may conclude that the
r r r
integrand F ⋅ dr = dφ is a perfect differential. Consequently, we conclude that F is a conservative
r r
force if it can be shown that ∇ × F = 0 .

C′ C

Figure 3.6: Two different paths.

We can recast the work-energy principle for conservative force systems by stating that the change in
kinetic energy of a particle equals the change in the potential function in moving from position 1 to 2
in the field, which is written mathematically as

φ2 − φ1 = T2 − T1 (3.22)

or simply

∆φ = ∆T (3.23)

134
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

r
Furthermore, we note that for a conservative force F , we can conveniently relate the components of
force to the potential function φ , which mathematically describes the force field. By referring to
rectangular coordinates, for example, we may write

r r
F ⋅ dr = Fx dx + Fy dy + Fz dz = dφ (3.24)

In addition, considering the potential function φ to be a function of x, y, and z, we may express its
differential as

∂φ ∂φ ∂φ
dφ = dx + dy + dz (3.25)
∂x ∂y ∂z

Thus, we may express the rectangular components of force as

∂φ
Fx =
∂x
∂φ
Fy = (3.26)
∂y
∂φ
Fz =
∂z

or simply

r r
F = ∇φ = grad (φ ) (3.27)

We thus conclude that for a conservative system, the force is expressible as the gradient of a scalar
potential function, φ . Therefore, the force is a function of spatial coordinates only. This concept of a
force potential in many instances affords a convenient method for analyzing certain important classes
of engineering problems. It offers the advantage of analyzing force systems without necessitating a
consideration of the mechanism causing the force. In addition, in some cases this concept leads to the
Laplace equation, ∇ 2φ = 0 , which is one of the more important equations of mathematical physics.

135
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

3.3 Potential Energy and Conservation of Energy


With the concept of a conservative force available, we may now define the change in potential energy
of a conservative force as the negative of the work done by the force in moving from position 1 to 2 in
the field, written mathematically as

2 r r
V2 − V1 = − ∫ F ⋅ dr (3.28)
1

Change in potential energy is defined only for conservative forces and is seen from Eq. 3.28 to be a
scalar point function dependent only on the limits of integration. In fact, it looks very much like what
we discussed in the previous section under the concept of a potential function.

Comparing Eq. 3.28 with Eq. 3.10, we see that

V2 −V1 = φ1 − φ2 (3.29)

or simply

∆V = −∆φ (3.30)

Thus, the change in potential energy equals the negative of the change in the potential function
associated with positions 1 and 2. In differential form, this becomes dV = − dφ , which upon
integration can be written as

V = −φ + C (3.31)

Consequently, the potential energy function V and the negative of the potential function φ differ only
by an arbitrary constant.

Since we are restricting our attention to conservative forces when discussing both potential energy and
r r
the potential function, we are in essence redefining the perfect differential F ⋅ dr , discussed in Eq.
3.9, as minus dV, rather than plus dφ . This seems like a trivial thing to do, until we take a look at the
form of the work-energy principle in terms of potential energy, which becomes

ωk = V1 − V2 = T2 − T1 (3.32)

136
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

or simply

T1 + V1 = T2 + V2 = C (3.33)

Stated in words, we may now say that the sum of kinetic energy plus potential energy is constant for a
conservative mechanical system. This is the well known law of conservation of energy, which we
must stress is applicable only to conservative systems which have, for example, no energy input or
energy dissipation due to friction. The importance of this method lies in the fact that one can simply
equate the sum of kinetic energy plus potential energy at any instantaneous configuration of the
system to that at any other instant.

Just as we developed relationships in the previous section relating the components of force to the
potential function, we can make similar statements related to potential energy. Considering
rectangular coordinates, for example, the potential energy differential dV can be written as

∂V ∂V ∂V
dV = dx + dy + dz (3.34)
∂x ∂y ∂z

r r
Thus, recalling F ⋅ dr and dφ from Eqs. 3.24 and 3.25, we may write

∂φ ∂V
Fx = =−
∂x ∂x
∂φ ∂V
Fy = =− (3.35)
∂y ∂y
∂φ ∂V
Fz = =−
∂z ∂z

or simply

r
F = grad (φ ) = − grad (V ) (3.36)

Similar to the discussion on potential functions, the potential energy V can also be determined only
within an arbitrary constant. Therefore, as before, it is convenient to define potential energy relative to
a reference or datum position, R, as shown in Figure 3.5. Since the work done by a conservative force
is given by

137
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

r r
ωk = ∫ F ⋅ dr = V1 − V2
2
(3.37)
1

we may write

R r r 2 r r R r r R r r
ωk = ∫ F ⋅ dr + ∫ F ⋅ dr = ∫ F ⋅ dr − ∫ F ⋅ dr = V1 − V2 (3.38)
1 R 1 2

Thus, comparing Eqs. 3.37 and 3.38, we may conveniently denote the potential energy V associated
with a general position P in the field as the work done by a conservative force in going from a general
position P to a reference position, R, written mathematically as

R r r
V = ∫ F ⋅ dr (3.39)
P

Similarly, if reference point R′ is used in Figure 3.5 rather than R, we have

R′ r r
V ′ = ∫ F ⋅ dr (3.40)
P

so that

V ′ = V + C2 (3.41)

where

R′ r r
C2 = ∫ F ⋅ dr (3.42)
R

These results can be readily compared to Eqs. 3.12 through 3.16, where it can be noted that V and φ
are exactly the negative of one another if they are measured relative to the same reference position, R.
If not, they differ by an arbitrary constant as stated in Eq. 3.41.

138
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Example 3.2 Conservative Force Systems

Given a force system characterized by Fx = 2 x y + y , and Fy = x 2 − y 2 + x , is this a conservative

system? If so, find the potential function φ . Also, evaluate the work done around a closed path
described by straight lines from (0,0) to (2,0) to (2,3) to (0,0).

Solution: If this is a conservative system, then

r r
∇× F = 0 (a)

which yields the requirement that

∂Fy ∂Fx
− =0 (b)
∂x ∂y

Substitution of Fx and Fy into Eq. (b) yields

∂ 2 ∂
∂x
( ∂y
)
x − y 2 + x − (2 x y + y ) = 2 x + 1 − 2 x − 1 = 0 (c)

r r r
Therefore, since ∇ × F = 0 , then φ exists and F is conservative. Now

∂φ ∂φ
Fx = ; Fy = (d)
∂x ∂y

so that

φ = ∫ Fx dx + f ( y ) (e)

and

φ = ∫ Fy dy + f ( x) (f)

From Eqs. (e) and (f), we find

φ = x2 y + x y + f ( y) (g)

139
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

and

y3
φ = x 2 y + x y + f (x ) − (h)
3

y3
Accordingly, we find f ( x ) = C1 , f ( y ) = C1 − and then
3

y3
φ = x2 y + x y − + C1 (i)
3

r
Since F is conservative, we know immediately that

W = ∫ dφ = 0 (j)

However, let us also show this by definition of

r r
W = ∫ F ⋅ dr (k)

Thus we write

W = ∫ (Fx dx + Fy dy + Fz dz )

= ∫ (2 x y + y ) dx
2

0
3
(
+ ∫ x 2 − y 2 + x dy
0
)
y =0 x=2

+ ∫ (2 x y + y ) dx
0

2 3
0
(
+ ∫ x 2 − y 2 + x dy
3
) 2
(l)
y= x x= y
2 3
3 0 0
 y3   3   4 3 y3 y2 
= 0 +  4 y − + 2 y  +  x 3 + x 2  +  y − + 
 3 0  4  2  27 3 3 3
= (12 − 9 + 6 ) − (8 + 3) − (4 − 9 + 3) = 0

Finally, we note that

y3
V = −x2 y − x y + + C2 (m)
3

140
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Example 3.3 Potential Energy of Springs, Weights and Gravitation

Three types of conservative forces that frequently occur in engineering problems are (1) linear spring
forces, (2) weights, and (3) gravitational attractions. Determine the potential energy and the potential
function for each, based on a suitably selected reference position.

Solution:

1) A linear spring force is shown acting on a mass m in Figure 3.7, where the spring is stretched
a distance x from its natural length.

k Fs = k x
m m

Figure 3.7: Mass-spring system.

The force acting on mass m is given by

r r
Fs = − K x i (a)

where K is the elastic spring constant.

The usual reference position for a spring force is at x = 0 corresponding to the unstretched
configuration of the spring. Thus, the potential function φ for a spring stretched a distance x
according to Eq. 3.14 is given by

r r
φ = ∫ (− K x′ i )⋅ (dx′ i ) = − K x 2
x 1
(b)
0 2

where x′ is a dummy integration variable.

Similarly, the potential energy according to Eq. 3.39 becomes

141
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

r r 1
V = ∫ (− K x′ i )⋅ (dx′ i ) = K x
0
(c)
x 2

Since both φ and V can only be calculated within an arbitrary constant dependent on the choice of
reference position, we can in general write the potential function and potential energy of a spring
respectively as

1
φ = − K x 2 + C1 (d)
2

1
V= K x 2 + C2 (e)
2

2) A weight W is expressed as a force vector by the relationship

r r r
W = −W k = −m g k (f)

r
where g is the local acceleration of gravity and the unit vector k is taken as vertically upward. Thus,
the potential function for a weight W at a height z above the reference position taken at z = 0 is given
according to Eq. 3.24 as

( r
φ = ∫ − W k ⋅ dz′ k = −W z
0
z
)( r
) (g)

Similarly, from Eq. 3.39, we have

0 r
( r
V = ∫ − W k ⋅ dz ′ k = W z
z
)( ) (h)

or in general, for arbitrary reference positions

φ = −W z + C1 (i)

and

V = W z + C2 (j)

142
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Frequently, z is simply measured from the Earth’s surface.

3) Gravitational force is given according to Newton’s Law of Gravitation by

r G m1 m2 r
F =− er (k)
r2

where G is the universal gravitational constant, m1 and m2 are the masses of two attracting bodies
and r is the distance between their mass centers as shown in Figure 3.8.

r m2
F
r
−F
r r
m1 r = r er

Figure 3.8: Mutual gravitation.

For problems involving gravitational attraction such as orbiting satellite problems, for example, it is
most common to choose infinity as the reference position. Accordingly, the potential function
associated with a pair of masses m1 and m2 a distance r apart is determined by calculating the work
done according to Eq. 3.24 in bringing the mass m2 from infinity to within a distance r of mass m1 as
r
shown. The simplest path from infinity is along the radial direction er , so we may write

r
 G m1 m2 r  r
er (dr er ) = − ∫
r Gm m G m1 m2 G m1 m2
φ = ∫ −
r
1 2
dr = = (l)
 
∞ 2 ∞ 2
r r r ∞ r

Similarly, the potential energy is given by

∞ G m1 m2 G m1 m2
φ = −∫ 2
dr = − (m)
r r r

143
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Thus, by selecting the reference point at infinity, the potential energy for an orbiting satellite or
celestial body is negative.

It is very interesting and instructive to note that weight is really the direct result of gravitational
attraction and therefore the potential energy of a weight W should be very similar in form to the
potential energy associated with gravitational attraction. We see by comparing Eqs. (h) and (m) that
the similarity is not obvious. This can be resolved by using the surface of the Earth as the reference
position for each case and reevaluating the potential energy for gravitational attraction. Thus, Eq. (m)
becomes

re G me m  1 1
V = −∫ 2
dr = G me m ⋅  −  (n)
r r  re r 

where re is the radius of the Earth, me is the Earth’s mass, and m is the mass of a body near the
Earth’s surface. Considering m to be at a height z above the Earth’s surface, we may replace r by
re + z yielding

G me m  r − re  G me m z
V=  = (o)
re  r  re r

Assuming that z is small compared to re , we may write

G me m
V≈ 2
z = g0 m z (p)
re

where the acceleration of gravity, g 0 , at the surface of the Earth is given by

G me
g0 = 2
(q)
re

Thus, the potential energy of the weight W as presented in part (2) of this solution in effect neglects
the variation in the acceleration of gravity g with altitude.

144
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

In closing, it is interesting to note that the weight W0 at the Earth’s surface and the weight W at any
distance r from the Earth’s center are written respectively as

r r
W0 = −m g 0 k (r)

r r
W = −m g k (s)

Also, according to Eq. (k), we see that

r G me m r re
2 r
W =− er = − g 0 m k (t)
r2 r2

Therefore, we can relate the acceleration of gravity at any position to the value at the Earth’s surface
by the equation

2
re
g = g0 (u)
r2

or in terms of relating the weights we have

2
r
W = W0 e2 (v)
r

Example 3.4 Natural Frequency of Mathematical Pendulum

Determine the natural frequency of the mathematical pendulum shown in Figure 3.9 for small
oscillations about the equilibrium position shown as θ 0 .

145
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

O x
θ0
C l

k
m

Figure 3.9: Mathematical pendulum.

Solution: We note that this system is a conservative system since only conservative forces are acting,
so let us reconsider Eq. 3.33 for a moment.

V1 + T1 = V2 + T2 = const. (a)

Equation (a), of course, states that the energy E of the system is constant as a function of time, which
is shown graphically in Figure 3.10.

constant

E V

t
Figure 3.10: Conservation of energy principle.

Also, recall that for harmonic motion of a particle, we have

q = q0 sin ( p t − φ ) (b)

146
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

and thus

 π
q& = q0 p cos( p t − φ ) = q0 p sin  p t − φ +  (c)
 a

where q0 is the amplitude of motion and p the circular frequency. We see from Eq. (c) that q& is

exactly 90° out of phase with q. That is, the velocity function leads the displacement function by
90° . Further, we note that

q = qmax = q0  π
 at ( p t − φ ) = (d)
q& = 0  2

q& = q&max = q0 p 
 at ( p t − φ ) = 0 (e)
q=0 

Accordingly, we may write

q&max = qmax p (f)

Now, since the energy is constant and depends directly on q and q& which vary harmonically, and
since qmax and q&max are 90° out of phase, then Figure 3.10 may be redrawn as Figure 3.11.

constant

V V
E

T T T

Figure 3.11: Harmonically varying energy.

147
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Thus, we conclude from Figure 3.11 that for a single degree of freedom, undamped, free vibratory
motion

Vmax = Tmax

Returning to our problem now, we have for equilibrium of the pendulum shown in Figure 3.9

r
M0 = 0 (g)

or

F0 C = W l cos(θ 0 ) (h)

where F0 is the spring force for the equilibrium state. If we displace the pendulum a small clockwise

angle α from equilibrium, which is considered the reference position for zero potential energy, we
have

T=
1
m v2 =
1W
(l α& )2 (i)
2 2 g

and

V = ∫ [F0 + k α C ]d (α C ) − ∫ W l cos(θ 0 + α ) dα
α α
(j)
0 0

Substituting cos(θ 0 + α ) = cos(θ 0 ) cos(α ) − sin (θ 0 )sin (α ) into Eq. (j) and performing the required
integration, we obtain

k C2 α 2
V = F0 C α + − W l cos(θ 0 ) cos(α ) + W l sin (θ 0 )(1 − cos(α )) (k)
2

α2
However, for small oscillations, we may make the approximations sin (α ) = α and cos(α ) = 1 −
2
which with the aid of Eq. (h) reduce Eq. (k) to

148
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

k C2 α 2 α2
V= + W l sin (θ 0 ) (l)
2 2

Note that one must be careful in making small angle assumptions, since all terms of the lowest order
must be retained. Note how Eq. (l) would differ if we had simply replaced cos(α ) by unity.

Finally, equating Vmax = Tmax (where α max = Q and α& max = Q p ) we obtain

l Q p = k C 2 Q 2 + W l sin (θ 0 )Q 2
1W 2 2 2 1 1
(m)
2 g 2 2

Thus, the natural circular frequency of oscillation p is given by

k C 2 + W l sin (θ 0 )
p2 = rad
s2
(n)
ml2

and the natural frequency of oscillation f is

p 1 k C 2 + W l sin (θ 0 )
f = = cycles
(o)
2π 2π
s
ml2

3.4 Work-Energy Methods for a System of Particles


Now we are in a position to extend the work-energy relationships to a system of particles. Again, we
let n be the number of particles in the system and let mi be any representative particle. Further, as
r r
previously discussed in Section 2.3, Fi represents the resultant external force on mi , and f i , j the

internal force exerted on mi by m j . Referring to Figure 3.12, mi is located with respect to the xyz
r r r
inertial space by ri and with respect to an arbitrary point P by ρ i . Further, rc locates the center of
r
mass of the system in inertial space and ρ c locates the mass center C relative to point P. All of this
notation used here is consistent with that used in Section 2.3, even to the extent that Figure 3.9 is the
same as Figure 2.5.

149
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

y
r
r fj
r r f j ,i
fi , j mj
fi
mi r
rj
r
r r ρj
ri ρ
r i P ρr
rP r c
rc C (mass center)
x

mn

Figure 3.12: Vector definitions for forces acting on arbitrary masses.

The equation of motion for the particle mi according to Newton’s Law is

r
r n r dr&i r
Fi + ∑ f i , j = mi = mi &r&i (3.43)
j =1 dt
j ≠i

Now, if we take the dot product of Eq. 3.43 by dri and integrate from position 1 to position 2 in the
field, we have again the work-energy principal for each particle. If we further take the sum of these n
equations for all particles of the system, we have

r n r  r
n
  r n 2 dr&i r
∑∫  Fi + ∑ f i , j  ⋅ dri = ∑ ∫1 mi
2
⋅ dri (3.44)
i =1
1
 j =1  i =1 dt
 j ≠i 

Since the summation in Eq. 3.44 is independent of the integration, Eq. 3.44 may be written

150
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

r
2 n r r 2 n n r r n 2 dr&i r&

1

i =1
Fi ⋅ dri + ∫ ∑∑ f i , j ⋅ dri = ∑ ∫ mi
1
i =1 j =1 i =1
1 dt
⋅ ri dt
j ≠i
n
r r
= ∑ ∫ mi r&i dr&i
2

1
i =1
2
n 1 r r
= ∑ mi ⋅ r&i ⋅ r&i ( )
 i=1 2 1
2
n 1 r2 
= ∑ mi ⋅ r&i 
 i=1 2 1

or simply

2
2 n r r 2 n n r r  n 1 r2 

1

i =1
Fi ⋅ dri + ∫1
∑∑
i =1 j =1
f i, j ⋅ dr i = ∑
 i =1 2
mi ⋅ r&i 
1
(3.45)
j ≠i

The first and second terms on the left side of Eq. 3.45 represent the work done by the external and
internal forces respectively as the system moves from position (or configuration) 1 to 2. Also, by
defining the kinetic energy of the system as the sum of the kinetic energies of the individual particles
making up the system, written as

n
1
T = ∑ mi r&i
2
(3.46)
i =1 2

we may state that the total work done by all forces acting on any system of particles equals the change
in kinetic energy of the system. This statement takes the form

Wk = Wk E
+ Wk I = ∆T (3.47)

where Wk E
and Wk I , the work done by the external and internal forces of the system, are written

respectively as

r r
=∫ ∑ F ⋅ dr
2
Wk E i i (3.48)
1

and

151
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

2 n n r r
Wk I = ∫
1
∑∑ f
i =1 j =1
i, j ⋅ dri (3.49)
j ≠i

r
Frequently, it is more desirable to describe the position of each particle mi by a position vector ρ i
measured from some convenient reference point P (Figure 3.12) rather than using the inertial position
r r r r
vectors ri . This is easily done by noting in Figure 3.12 that ri = rP + ρ i which also implies that
r r r r r r
dri = drP + dρ i and r&i = r&P + ρ& i . Substituting these expressions into Eq. 3.45, we obtain

r r 2 n r n r n r
2 n r 2 n r 2 n r

1
∑ i P ∫∑
F ⋅
i =1
d r + F ⋅ d ρ i + ∫ ∑∑ i , j P ∫ ∑∑ fi , j ⋅ dρi
1
f ⋅ d r
i =1
+
1
i =1 j =1
1
i =1 j =1
j ≠i j ≠i
2
(3.50)
 1 r r r r 
r r
(
= ∑ mi ⋅ r&P ⋅ r&P + 2 r&P ⋅ ρ& i + ρ& i ⋅ ρ& i  )
 2 1

In Eq. 3.50 we make the following observations

2 n r r r r
∫ ∑ F ⋅ dr = ∫
2
P FR ⋅ drP
1 1
i =1

2 n n r r
∫ ∑∑
1
i =1 j =1
f i , j ⋅ drP = 0
j ≠i (3.51)

∑ 2 m ⋅ (r& ) ( )
n
1 r r& 1 r r 1
P ⋅ rP = m ⋅ r&P ⋅ r&P = m r&P
2
i
i =1 2 2

∑ 2 m ⋅ (2 r& )
n
1 r r r r
i P ⋅ ρ& i = m r& ⋅ ρ& c
i =1

r
where FR is the resultant external force acting on the system.

Substitution of Eq. 3.51 into Eq. 3.50 yields

r 2 n r n r
r r 2 n r
∫ FR ⋅ drP + ∫ ∑ Fi ⋅ dρ i + ∫ ∑∑ f i , j ⋅ dρ i
2

1 1 1
i =1 i =1 j =1
j ≠i
(3.52)
2
1 r 2 1 r r  n
=  m rP + ∑ mi ρ& i + m r&P ⋅ ρ& c 
2

2 2 i=1 1

152
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

r
where ρ& c is the velocity of the system’s mass center C relative to the general reference point P. Eq.
3.52 states that the total work done by the system of internal and external forces acting on the system
of particles from configuration 1 to configuration 2 is equal to the change in total kinetic energy of the
system.

Note in Eq. 3.52 that the total kinetic energy of the system is expressed as the kinetic energy of the
total mass of the system moving with the velocity of our reference point P, plus the kinetic energy of
the system relative to P, plus the linear momentum of the mass of the system at point P dotted with the
velocity of the center of mass with respect to P. Recalling Eq. 3.46, we conclude that the kinetic
energy of a system of n particles can be written in either of the following forms

n
1 1 1 n r r
T = ∑ mi r&i = m r&P + ∑ mi ρ& i + m r&P ⋅ ρ& c
2 2 2
(3.53)
i =1 2 2 2 i=1

where r&i is the magnitude of the inertial velocity of mi relative to point P. Note also that if the
r r
reference point P is selected as the system’s mass center C for which ρ c = ρ& c = 0 , Eq. 3.53 reduces to

1 1
T= m r&c + ∑ mi ρ& i
2 2
(3.54)
2 2

where ρ& i is now the speed of mi relative to the mass center C.

Furthermore, by selecting the mass center C as the reference point P it is also worth noting that Eq.
3.52 takes the form

2
r 2 n r n r
r r 2 n r 1 1 n 
∫ FR ⋅ drc + ∫ ∑ Fi ⋅ dρ i + ∫ ∑∑ f i , j ⋅ dρ i =  m r&c + ∑ mi ρ& i 
2 2
(3.55)
1 1
i =1
1
i =1 j =1 2 2 i=1 1
j ≠i

r
where ρ i is measured from the mass center C to each particle mi .

With the work-energy principle written in terms of the motion of the mass center as done in Eq. 3.55,
it is also very interesting to recall the principle of the motion of the mass center from Section 2.3
which is written as

153
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

r
r dr&c
FR = m (3.56)
dt

r
Taking the dot product of Eq. 3.56 by drc and integrating this result from position 1 to 2, we obtain

r
r r dr&c r

2
FR ⋅ drc = ∫ m ⋅ drc (3.57)
1 dt

which upon integration yields

r
dr&c r&
2 2


2 r r 2
FR ⋅ drc = ∫ m
2r r&  1 r& r&  1
( 2
⋅ rc dt = m ∫ rc ⋅ drc =  m ⋅ rc ⋅ rc  =  m r&c 
& )
1 1 dt 1
2 1  2 1

or simply

2
2 r r 1 2

1
FR ⋅ drc =  m r&c 
2 1
(3.58)

A comparison of Eqs. 3.55 and 3.58 reveals in fact that Eq. 3.55 actually contains two separate
equations,

2
r r n r
2 n 2 n r 1 n 

1

i =1
Fi ⋅ dρ i + ∫1
∑∑
i =1 j =1
f i, j ⋅ dρ i = ∑
 2 i=1
mi ρ& i 
1
(3.59)
j ≠i

and, of course, Eq. 3.58. Therefore, in dealing with the work-energy principle for a system of particles
based on the motion of the center of mass of the system, one has the choice of three equations to use,
namely either Eq. 3.55, Eq. 3.58, or Eq. 3.59. One must remember, of course, that neither Eq. 3.58
nor Eq. 3.59 contains the total work or energy of the system. Thus, if one is concerned with those
quantities, then Eq. 3.55 must be used. Finally, we note that if both internal and external forces are
conservative, then any or all of Eqs. 3.55, 3.58, and 3.59 may be case in the form

∆T + ∆V = 0 (3.60)

154
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Example 3.5 Work Done by Gravity

Prove that the work done by gravity acting on any system of particles equals the total weight of the
system multiplied by the change in vertical position of its mass center.

r
Solution: Let the vertically upward direction be denoted by the unit vector k . Thus, the weight of
mass mi is given by

r r
Wi = − g mi k (a)

r
where g is the acceleration of gravity and is assumed to be constant. The work done by Wi during a
r r r r
displacement of mi by an amount dri = dxi i + dyi j + dzi k is

r r
dωk = Wi ⋅ d ri = − g mi dzi (b)

Consequently, the work done by all components of weight acting on each particle of the system as the
system moves from configuration 1 to 2 is given by

2
2 nr r 2 n n
ωk = ∫
1
∑Wi ⋅ dri = − g ∫ ∑ mi dzi = − g ∑ mi zi
i =1
1
i =1 i =1
(c)
1

Finally, by the definition of the position of the mass center in the vertical direction given by

n
m zc = ∑ mi zi (d)
i =1

n
where m = ∑ mi , we conclude that
i =1

ωk = − m g zc 1 = −W ⋅ [( zc )2 − (zc )1 ] = W ∆h
2
(e)

In Eq. (e), the quantity ∆h is the change in vertical height of the mass center of any general system of
particles and is considered positive for a downward displacement.

155
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Example 3.6 Energy of Cart on Circular Platform

A cart of mass m1 rolls in a slotted track on a circular platform which is rotating about a vertical axis
r
with an angular velocity ω . The cart is constrained to move on the platform by a spring of spring
constant k. Suspended from the cart by a 2-dimensinoal hinge at point A is a mathematical pendulum
of length l and mass m2 . If the axis of the hinge is normal to the zy body plane, determine the total
kinetic energy of the system, treating both the cart and pendulum.

r r
ω1 = ω1 k
q

m1
O
y

l
x
θ m

Figure 3.13: Cart on spinning platform with pendulum hanging through slot.

Solution: The kinetic energy of this system can be calculated in either of three ways:

1. by Eq. 3.46 using the inertial velocity of each mass,

2. by Eq. 3.53 using velocities relative to some convenient reference point P, or

3. by Eq. 3.54 using velocities relative to the mass center of the system.

In order to demonstrate these various approaches, this solution will be carried out by both methods (1)
and (2) above, leaving method (3) as an exercise for the student.

156
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Method (1) In using Eq. 3.46, it is necessary to first determine the inertial velocities of each mass
in the system. Reviewing the methods of kinematics for rotating coordinate systems from Chapter 1,
r r
it is convenient to employ an x, y, z frame attached to the platform which rotates at ω = ω k

Therefore, the velocity of the cart is given by

r r r r r r r r
r&1 = r&0 + ω × ρ + ρ& r = 0 + ω k × q j + q j

r r r
∴ r&1 = −ω q i + q& j (a)

Similarly, for the pendulum mass m2 we have

r r
[ r r
] [( r
)
r
r&2 = 0 + ω k × (q + l sin (θ )) j − l cos(θ ) k + q& + l θ& cos(θ ) j + l θ& sin (θ ) k ]
r
r r
( r
)
∴ r&2 = −ω ⋅ (q + l sin (θ )) i + q& + l θ& cos(θ ) j + l θ& sin (θ ) k (b)

Consequently, the kinetic energy of the system according to Eq. 3.46 is

T=
1
2
(
r r 1
) r r
m1 ⋅ r&1 ⋅ r&1 + m2 r&2 ⋅ r&2
2
( ) (c)

which yields

T=
1
2
(
m1 ⋅ ω 2 q 2 + q& 2 )
(d)
(
+ m2 ⋅ q 2 ω 2 + 2 ω 2 q l sin (θ ) + ω 2 l 2 sin 2 (θ ) + q& 2 + 2 q& l θ& cos(θ ) + l 2 θ& 2
1
2
)

Method (2) Hinge point A is a convenient reference point for determining the system’s kinetic
energy according to Eq. 3.53, which becomes

(m1 + m2 ) r&A 2 + 1 ∑ mi ρ& i 2 + m rr&A ⋅ ρr& c


2
1
T= (e)
2 2 i=1

r r
where r&A = r&1 is given by Eq. (a).

157
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

In addition, the velocities of m1 , m2 , and the mass center C relative to reference point A are given
respectively by

r
ρ&1 = 0 (f)

( )( )
r r r r r
ρ& 2 = θ& i + ω k × l sin (θ ) j − l cos(θ ) k
r r r (g)
= −ω l sin (θ ) i + l θ& cos(θ ) j + l θ& sin (θ ) k

r
( r r
ρ& c = θ& i + ω k ×
l m2
m1 + m2
) (sin(θ ) rj − cos(θ ) kr )
(h)
ω l m2 r θ& l m2 r θ& l m2 r
=− sin (θ ) i + cos(θ ) j + sin (θ ) k
m1 + m2 m1 + m2 m1 + m2

Substituting Eqs. (a), (f), (g), and (h) into Eq. (e) yields the kinetic energy given by

T=
1
2
( ) (
(m1 + m2 ) ω 2 q 2 + q& 2 + 1 m2 ⋅ ω 2 l 2 sin 2 (θ ) + l 2 θ& 2
2
)
(i)
 ω q l m2
2
q& θ& l m2 
+ (m1 + m2 ) sin (θ ) + cos(θ )
 m1 + m2 m1 + m2 

which simplifies to Eq. (d) as expected.

It is very important for the student to understand exactly what is meant by the relative velocity terms
r
that appear in Eq. 3.53. Thus, it is worth noting here, for example, that ρ& 2 as given by Eq. (g) in
physical terms is really the difference between the inertial (total) velocity of mass m2 and the inertial
velocity of point A. This can be easily demonstrated by recalling from geometry with the aid of
Figure 3.12 that

r r r
ρ 2 = r2 − rA (j)

Therefore,

r r r
ρ& 2 = r&2 − r&A (k)

158
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

which could be conveniently formed for this example by subtracting Eq. (a) from Eq. (b) yielding the
same result as Eq. (g). This method is probably the easiest way of understanding the physical
r
meaning of a term like ρ& 2

Example 3.7 Hammer Stamping Mechanism

The hammer mechanism shown in Figure 3.14 is released from rest in the vertical position and strikes
the horizontal stamping surface at A. The hammer mechanism is driven by a torsional spring of spring
ft ⋅lb
constant k rad in addition to gravity. The hammer is made up of a uniform rod of mass m and length

l with a concentrated mass 2 m at its end. Determine the speed at which the hammer head strikes the
π
stamping plate A considering the torsional spring unstressed for θ = .
2

y
2m

l
m
θ&
θ

k
x
A
Figure 3.14: Spring and gravity-driven hammer.

Solution: Since this is essentially a problem relating a change in velocity to a change in position of
the system, the work energy principle is well suited. Consequently, we must calculate the kinetic
energy of the hammer mechanism which consists of a particle of mass 2 m plus a slender rod of mass
m.

The slender rod can be thought of analytically as made up of a system of infinitesimal particles, each
m
of length ds and mass dm = ds as shown at Q in Figure 3.15.
l

159
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

dS

Q
S
θ&
l

Figure 3.15: Infinitesimal representation of the slender rod.

m
In this way we will determine the kinetic energy of each elemental particle of mass dm = ds at
l
position s along the rod and then sum over the entire rod by integration, which mathematically has
been defined as the limit of a summation as the element size approaches zero. Accordingly, the
kinetic energy of the elemental mass at Q is written

dT =
1
2
2 1m 
dm vQ =  ds  s θ&
2 l 
( ) 2
(a)

Thus, the kinetic energy of the entire slender rod is given by

1 m &2 l 2 1
T = ∫ dT = θ ∫ s ds = m l 2 θ& 2 (b)
2 l 0 6

The kinetic energy of the hammer mechanism is then

1 7
T = m l 2 θ& 2 + m l 2 θ& 2 = m l 2 θ& 2 (c)
6 6

The work done as the hammer falls from the vertical to the horizontal position due to gravity and the
torsional spring respectively is

π  k
2
l
Wk = m g + 2 m g l +  
2 2 2

160
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

or

5 π2
Wk = mgl + k (d)
2 8

Therefore, equating work to the corresponding change in kinetic energy yields

5 π2 7
mgl + k = m l 2 θ& 2 (e)
2 8 6

Consequently, the angular velocity of the hammer mechanism immediately before striking the
stamping platform A is

1 15 3π 2 k
θ& = gl + (f)
l 7 28 m

and the corresponding linear velocity of the hammer head is

15 3π 2 k
v= gl + (g)
7 28 m

Example 3.8 Vibrating Beam

A uniform beam of total mass M and flexural rigidity EI is supported on equal springs with total
π x 
vertical stiffness of k lb
in . Assuming the deflected shape of the beam is given by y = b + sin  ,
 l 
determine the fundamental frequency of vibration of the beam.

161
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

k k
2 2
x

Figure 3.16: Rigid beam supported by springs.

Solution: It is clear that we have an infinite system of elemental particles in this case and must
therefore deal with an integration rather than summation of energies. Further, on the basis of Example
3.4, we must determine the maximum values of kinetic and potential energies of the system.

Therefore, we assume that the beam vibrates harmonically with its shape given by

  π x 
y =  b + sin   sin (ω t ) (a)
  l 

The kinetic energy of the beam is thus given by

1 lM 2
2 ∫0 l
T= y& dx

 π x   π x 
ω cos 2 (ω t )b 2 + 2 b sin 
1M l 2
=
2 l 0 ∫   l 
 + sin 2 
 l 
 dx

 l l
cos 2 (ω t )b 2 l + 4 b + 
1 M
= ω2
2 l  π 2

 4b 1 
M ω 2 cos 2 (ω t )b 2 +
1
T= + (b)
2  π 2 

It follows from Eq. (b) that Tmax is given by

1  4b 1 
Tmax = M ω 2 b 2 + + (c)
2  π 2 

162
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

The potential energy of the system consists of the potential energy of the springs plus the elastic strain
energy stored in the beam due to its deformation. For the springs, we have

1k  1k 
k b sin 2 (ω t ) + k b 2 sin (ω t )
1 2 1
Vs =   y 2 +   y2 = (d)
2 2 x =0
2 2 x =l
4 4

with the maximum value

1 2
Vsmax = kb (e)
2

σε
Recalling from strength of materials that the strain energy per unit volume is given by , we have
2
for the entire beam

σε l A σ2
VB = ∫ dV = ∫ ∫ dA dx (f)
V
2 0 0 2E

where by Hooks law, σ = E ε and A is the cross sectional area of the beam. Recalling further that

My d2y M
σ= , I = ∫ y 2 dA , and = , Eq. (f) becomes
I dx 2 E I

2
l A M 2 y2 1 1 lM2 1 l d2y
VB = ∫ ∫ dA dx = ∫ dx = ∫ E I ⋅  2  dx (g)
0 0 I2 2E 2 0 EI 2 0  dx 

Thus, we find

 π  4 2  π x  2  E I  l  π 
4

∫0  l  sin  l  sin (ω t ) dx = 2  2  l  sin (ω t )


EI l
VB = 2
(h)
2  

and

E I  l  π 
4

VBmax =    (i)
2  2  l 

163
C H A P T E R 3 : W O R K - E N E R G Y M E T H O D S F O R P A R T I C L E S

Finally, equating Vmax to Tmax , we find

 l  π 
4
1 2 1 1  4b 1 
k b + E I ⋅    = M ω 2  b 2 + +  (j)
2 2  2  l  2  π 2

Solving for the circular frequency ω , we get

E I π 
4

  l +kb
2

2 l
ω2 = (k)
1 4b 
M ⋅ + + b2 
2 π 

from which the natural frequency f is

E I π 
4

  l +kb
2

ω 1 2 l
f = = (l)
2π 2π 1 4b 
M ⋅ + + b2 
2 π 

164
4
Chapter

Impulse-Momentum Methods for Particles

I n Chapter 3 various forms of the energy principles were discussed for a single particle and
systems of particles. These methods were found to be particularly useful in treating forces given
as functions of displacements and in these cases they provided a first integral of Newton’s second law
for determining velocities. In particular, the work-energy methods were shown to be very valuable for
relating changes in velocity of the particles to changes in their positions.

However, if it is desirable to relate changes in velocity to the time interval involved, it will usually
prove advantageous to use the impulse-momentum methods as discussed in this chapter. Also, in
r r
dealing with time dependent forces, F = F (t ) , or impulsive forces, which are forces exerted on the
particles over very short time intervals, the impulse-momentum methods will prove to be most
appropriate. Impulse-momentum methods for both linear and angular motion will be presented in this
chapter, with the initial discussion dealing with linear impulse-momentum concepts.

4.1 Linear Impulse-Momentum Methods for a Single Particle


If we refer to Newton’s second law

r r
r dr& dv
F =m =m (4.1)
dt dt

and integrate this elementary differential equation by the method of separation of variables, we obtain

t2 r r
v2 r

t1
F dt = ∫r m dv
v1
(4.2)

or simply

165
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r r r
F dt = m ⋅ (v 2 − v1 )
t2
∫ t1
(4.3)

r t2 r
The quantity m v is defined as the linear momentum of the particle while the integral ∫
t1
F dt is

defined as the linear impulse acting on the particles during the time interval ∆t = (t 2 − t1 ) .
Accordingly, Eq. 4.3 states that the change in linear momentum of a particle during a time interval ∆t
is equal to the linear impulse acting on the particle during this time interval. Symbolically we write

r r
I = ∆P (4.4)

r t2 r r r
where we define the impulse I = ∫ F dt and the linear momentum P = m v . It is clear from Eq. 4.4
t1

that the linear impulse-momentum equation is a vector equation and thus has the scalar components
I x = ∆Px , I y = ∆Py , I z = ∆Pz or

Fx dt = m ⋅ (v2 − v1 )x
t2
∫ t1

Fy dt = m ⋅ (v 2 − v1 ) y
t2
∫ t1
(4.4a)

Fz dt = m ⋅ (v2 − v1 )z
t2
∫ t1

r r
It is also clear from Eq. 4.4 that if the impulse I = 0 then the linear momentum P = constant, and we
simply have a restatement of Newton’s first law.

A. Superposition Integral Methods

In many instances, we encounter forces which act for only a very short duration of time as indicated in
Figure 4.1.

166
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r
F

I r
=F

τ t

Figure 4.1: Impulse function.

In this case, we designate those forces as impulsive forces. As dt approaches zero, we consider that
r
the force F tends to infinity in such a way that the impulse, represented by the area under the force-
r r
time curve, during this time interval is finite and designated by I . In particular, when I = 1 , the

impulse is called a unit impulse and is symbolically designated by the Dirac-delta function δ . The
basic properties of the δ function are described below:

0 for t ≠ τ
δ (t − τ ) =  (4.5)
∞ for t = τ


∫ δ (t − τ )dt = 1 ,
−∞
− ∞ <τ < ∞ (4.6)

It follows directly from Eqs. 4.5 and 4.6 that

F (t )δ (t − τ ) = F (τ )δ (t − τ ) , − ∞ <τ < ∞ (4.7)

and furthermore, we may conclude that

F (t ) δ (t − τ ) dt = F (τ ) ∫ δ (t − τ ) dt = F (τ ) ,
∞ ∞
∫ −∞ −∞
− ∞ <τ < ∞ (4.8)

Also, since we are concerned only with t > 0 , Eqs. 4.6 and 4.8 become

167
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

∫ δ (t − τ )dt = 1 ,

0 <τ < ∞ (4.9)
0

∫ F (t )δ (t − τ )dt = F (τ ) ,

0 <τ < ∞ (4.10)
0

Not to be confused with the δ function is the unit step function u (t − τ ) , shown in Figure 4.2.

u (t − τ )

τ t

Figure 4.2: Unit step function.

The unit step function u (t − τ ) has the properties

0 for t < τ
δ (t − τ ) =  (4.11)
1 for t ≥ τ

It is also important to note by comparing Eqs. 4.5 and 4.11 that the Dirac-delta function is zero
everywhere except at t = τ , where it is infinite, which corresponds exactly to the slope of the unit step
function shown in Figure 4.2. Thus, we may relate the time derivative of the unit step function to the
Dirac-delta function as follows

u& (t − τ ) = δ (t − τ ) (4.12)

In order to illustrate the use of the unit impulse and unit step functions, let us first consider the
response of a free particle of mass m subjected to a unit impulse. For simplicity, consider the force to
be acting in the positive x-direction with the initial displacement of the particle equal to zero. Thus,
the particle will be considered to be “struck” with a unit impulse while at rest in its equilibrium

168
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

position at time t = 0 and we will then analytically describe its displacement response resulting from
this unit impulse.

Since I = Fx dt = m dv x , we may conclude that an impulse I acting on the mass will produce a sudden
change in velocity in the x-direction with no appreciable change in displacement. Thus, we have
essentially imparted an initial velocity to the mass m in the x- direction given by

I
x& 0 = (4.13)
m

while at the initial position x0 = 0 . Upon integrating Eq. 4.13, we obtain the resulting displacement
response of mass m as a function of time given by

I 
x =  t (4.14)
m

Thus, if we designate the displacement response of mass m to a unit impulse applied at t = 0 as g (t ) ,


we have from Eq. 4.14 that

1
g (t ) =   t (4.15)
m

from which it follows that the displacement response due to an impulse I acting on mass m can be
written as

x = I {g (t )} (4.16)

We may proceed in a similar manner for an arbitrary forcing function F (t ) applied to mass m. In
order to proceed, let us consider F (t ) as shown in Figure 4.3 to be divided into a series of impulses of
magnitude F (τ ) dτ each acting for a time duration dτ . Each impulse F (τ ) dτ acting on the particle
F (τ ) dτ
at time τ can be thought of as changing its instantaneous velocity by an amount as the
m
motion progresses.

169
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

F (t )

F (t )

τ t

t
Figure 4.3: A series of impulses.

However, if we restrict our attention to linear systems, we may employ the principle of superposition
and consider the resultant dynamic response as a superposition of responses due to a series of
impulses. From Eq. 4.16, the contribution to the total displacement at a time t due to an impulse of
magnitude F (τ ) dτ applied to mass m at time (t − τ ) , is given by

dx = F (τ ) g (t − τ ) dτ (4.17)

and the total displacement by superposition is thus given by

x = ∫ F (τ ) g (t − τ ) dτ
t
(4.18)
0

As an important example, the displacement response of a system to a unit step function u (t ) applied at
t = 0 , designated as h(t ) , may be obtained directly from Eq. 4.18 and given by

h(t ) = ∫ g (t − τ ) dt
t
(4.19)
0

We must emphasize that Eq. 4.18 is based on initial conditions of zero displacement and velocity
called homogeneous initial conditions. Eq. 4.18 is essentially a general particular solution to the
equation of motion of a particle, where F (t ) denotes the applied external force acting on the system
and g (t − τ ) describes its displacement response to a unit impulse, which is a function of the

170
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

constraints that may be acting on the particle. This description will become more apparent in the
example problems.

Furthermore, since the complete solution to any differential equation of motion consists of the
homogeneous (transient) solution plus any particular (steady state) solution, it is still necessary to add
a transient solution to Eq. 4.18, which can be used to satisfy non-homogeneous initial conditions by an
appropriate selection of the arbitrary constants. Thus, the steady state solution deals entirely with the
forcing function F (t ) and satisfies homogeneous initial conditions, while the transient solution
describes the free response of the system and must be used to account for the possibility of non-
homogeneous initial conditions. Eq. 4.18 is referred to as either the convolution integral, the
superposition integral, or Duhamel’s integral in honor of the French Mathematician J. M. C. Duhamel
(1797-1872) who is credited with its development.

Eq. 4.18 can also be written in an alternate form by essentially interchanging the role of τ and t − τ .
Accordingly, Figure 4.3 may be redrawn as Figure 4.4 resulting in an expression for the displacement
response to a forcing function F (t ) given analytically by

x = ∫ F (t − τ ) g (t ) dτ
t
(4.20)
0

F (t )

F (t )

τ t

t

Figure 4.4: Interchanged roll of t and t −τ .

171
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

It is also possible to express Duhamel’s integral given by Eq. 4.18 in terms of a superposition integral
based on the response of a system to a unit step function. In Figure 4.5, we again consider a general
forcing function F (t ) , only this time separated into horizontal elements dF (τ ) .

F (t )

dF (τ )

F (0)

τ t

t
Figure 4.5: General forcing function, with horizontal and vertical differential elements.

By definition of the unit step function, and referring to Eq. 4.19, the displacement dx of a system at
time t due to a differential force dF (τ ) suddenly applied at time t = τ is given by

dx = dF (τ ) h(t − τ ) (4.21)

Summing over all elemental forces dF (τ ) = F& (τ ) dτ and noting that we must include the effect of
F (0) applied at t = 0 , we obtain

x = F (0) h(t ) + ∫ F& (τ ) h(t − τ ) dτ


t
(4.22)
0

Thus, Eq. 4.22 describes the general displacement response of a system to an arbitrary force F (t ) for
homogeneous initial conditions in a manner equivalent to Eq. 4.18, although this time expressed in
terms of the system’s response to a unit step function h(t − τ ) rather than its response to a unit impulse
g (t − τ ) applied at time t = τ .

172
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

Similarly, the roles of τ and t − τ can be interchanged as done in Eq. 4.20 and schematically shown
in Figure 4.4 for the unit impulse approach yielding an alternate form of Eq. 4.22 given by

x = F (0) h(t ) + ∫ F& (t − τ ) h(t ) dτ


t

0
(4.23)

Example 4.1 Displacement of Particle under Transient Force

Determine the displacement response for t ≥ T of an unconstrained particle of mass m acted upon by
the force shown in Figure 4.6, considering that it starts from rest at time t = 0 . Solve the problem first
by (a) direct application of the elementary impulse-momentum equations, then by (b) Duhamel’s unit
impulse integral approach, and finally by (c) Duhamel’s unit step function integral method.

F (t )

F0

O T 2 T t

Figure 4.6: Force acting on particle.

Solution:

a) In order to determine the displacement of the particle, we must first express v = v(t ) . Thus,
from the basic impulse-momentum equation, Eq. 4.3, we may simply solve the problem successively
for the three time intervals from 0 ≤ t ≤ T 2 , T 2 ≤ t ≤ T , and for t ≥ T .

π
for which F (t ) =
2 F0 t
From Eq. 4.3, for 0 ≤ t ≤ we have
2 T

173
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

dt = m ⋅ (v − v0 ) = m v
t 2 F0 t
∫ 0 T
(a)

F0 t 2
∴v = = x& ; 0≤t ≤T 2 (b)
mt

from which

F0 t 3
x= ; 0≤t ≤T 2 (c)
3 mT

T F T F T2
At time , we have that x& = 0 and x = 0
2 4m 24 m

Similarly, from T 2 ≤ t ≤ T , we can again apply Eq. 4.3, although here it is convenient to measure
T
time t1 from so that 0 ≤ t1 ≤ T 2 applies to this second interval. Accordingly, we have
2

1 t1  t  F T
x& = ∫
m 0
 F0 − 2 F0 1  dt1 + 0
T 4m
(d)

from which

F0  t 
2

x& =  t1 + T − 2 1  ; 0 ≤ t1 ≤ T 2 (e)
m  4 T 

and

F0  1 2 T 
3 2
t
x=  t1 + t1 − 2 1 + T  ; 0 ≤ t1 ≤ T 2 (f)
m  2 4 3 T 24 

T
At t1 = for which t = T , we have
2

174
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

F0 T
x& =
2m
(g)
F T2
x= 0
4m

For values of t ≥ T for which it is convenient to measure time t 2 from T we no longer have a force
acting on the particle. Hence it moves with constant velocity equal to its speed at t = T for which we
may write from Eqs. (g)

F0 T
x& = ; t ≥T (h)
2m

and upon integrating we obtain

F0 T F T2
x= t2 + 0 ; t2 ≥ 0 (i)
2m 4m

or

F0 T  T 
x= t −  ; t ≥T (j)
2m  2 

b) Now, in terms of a unit impulse approach, we may directly apply Eq. 4.18, given by

x = ∫ F (τ ) g (t − τ ) dτ
t
(k)
0

We have previously shown that the response of an unconstrained particle to a unit impulse at t = 0 is

given by g (t ) = t from which it follows that g (t − τ ) = (t − τ ) . Thus, applying Eq. (k) for t ≥ T
1 1
m m
we obtain

2 F0 τ
(t − τ ) dτ + ∫T 2 F0 1 − τ (t − τ ) dτ
T
x=∫
T
2 (l)
mT 2 m  T
0

Integrating Eq. (l) we find

175
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

T T
T
 2 F t τ 2  2  2 F0 τ 3  2 2 F0  τ2 τ2t τ3 
x =  0  −   +  t τ − − +  (m)
 mT 2  0  mT 3  0 m  2 2 T 3T  T
2

or simply

F0 T  T 
x= t −  ; t ≥T (n)
2m  2 

Note that Eq. (n) agrees with our previous results from Eq. (j).

c) In terms of the unit step function integral method, we may use Eq. 4.22; namely

x = F (0) h(t ) + ∫ F& (τ ) h(t − τ ) dτ


t
(o)
0

For the problem at hand, F (0) = 0 and h(t ) is readily found from Eq. 4.19, yielding

h(t ) = ∫ g (t − τ ) dt = (t − τ ) dτ
t 1 t
m ∫0
(p)
0

Upon integrating, we obtain

t2
h(t ) = (q)
2m

from which we may write

h(t − τ ) =
(t − τ )2 (r)
2m

Substituting Eqs. (q) and (r) into Eq. (o), we obtain

 2F  1
(t − τ )2 dτ + ∫T  − 2 F0  1 (t − τ )2 dτ
T
x=∫  0
T
2 (s)
 T  2m
0
2 T  2m

Performing the required integration in Eq. (s), we obtain

176
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

T
T
F  τ3  2 F  τ3 
x = 0  t 2 τ − t τ 2 +  − 0  t 2 τ − t τ 2 +  (t)
mT  3  0 mT  3 T
2

or

F0 T  T 
x= t −  (u)
2m  2 

which again agrees with our previous results.

In determining the response of the system to a unit step function in Eq. (p), we made use of the
previously determined value of its response to a unit impulse. An alternate method for determining
h(t ) follows directly from solving the equation of motion for the particle under the influence of a unit
step function applied at t = 0 . Accordingly, we may write the equation of motion as

m &x& = 1 (v)

Upon integrating this equation subject to homogeneous initial conditions (i.e., x = x& = 0 at t = 0 ) we
obtain

t2
x= = h(t ) (w)
2m

Eq. (w) which describes the displacement response of the particle to a unit step function agrees with
the previously determined value given by Eq. (q).

Example 4.2 Spring-Mass System under Transient Force

Determine the response of the spring-mass system shown in Figure 4.7(a) to the force F (t ) that is
approximated by the series of straight line segments shown in Figure 4.7(b). Consider that the spring
is unstretched for x = 0 and that the system starts from rest at x = x .

177
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

F (t )

α2
α1 − α 2
x F0

m F (t )
k α1

T t

Figure 4.7: Cart modeled as spring-mass system responding to forcing function.

Solution: The differential equation describing the motion of mass m is readily obtained from
Newton’s law, yielding

m &x& + k x = F (t ) (a)

The complete solution of Eq. (a) consists of the homogeneous solution plus a particular solution.

a) Homogeneous Solution

From elementary differential equations, the solution of the homogeneous equation

m &x& + k x = 0 (b)

is given by

 k   k 
x = A sin  ⋅ t  + B cos ⋅ t  (c)
 m   m 

k
where = ω n is the natural frequency of the system.
m

b) Particular Solution

178
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

We may advantageously use the Duhamel integral approach by employing Eq. 4.18 which requires
that we first determine the characteristic response of this system to a unit impulse. Recall that
applying a unit impulse to the system shown in Figure 4.7(a) at t = 0 is equivalent to giving the
1
system an instantaneous initial velocity x& 0 = with no change in position. Thus, using the initial
m
1
conditions of x& 0 = and x0 = 0 at t = 0 in Eq. (c) will yield the displacement response of the
m
system to a unit impulse applied at t = 0 and denoted by g (t ) .

Thus, we have from Eq. (c) at t = 0 that

x0 = 0 = B (d)

Also, differentiating Eq. (c) and evaluating x& at t = 0 yields

1 k
x& 0 = =A = Aω n (e)
m m

Evaluating the constants, we have

1 1
A= = (f)
k m m ωn

and

B=0 (g)

from which we can conclude from Eqs. (c), (f), and (g) that the system’s response to a unit impulse at
t = 0 is given by

g (t ) = sin (ω n t )
1
(h)
m ωn

It follows that the response to a unit impulse applied at time t = τ is given by

179
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

g (t − τ ) = sin (ω n ⋅ (t − τ ))
1
(i)
mωn

The displacement response of mass m from 0 ≤ t ≤ T1 according to Eq. 4.18 becomes

α1 τ
sin (ω n ⋅ (t − τ )) dτ
t
x=∫ (j)
0 m ωn

where we have represented F (τ ) = α 1 τ =


F0
τ . Upon integrating, we obtain
T

α1  
sin (ω n t ) ;
1
x=  t − 0≤t ≤T (k)
k  ωn 

The displacement response for t > T can be obtained by directly applying Eq. 4.18 again. However,
it is also possible to obtain this response by superposition if one simply notes that at t = T the slope of
the forcing function F (t ) suddenly decreases by the amount α 2 . Consequently, it is convenient to
think of the response given by Eq. (k) as continuing on after t = T , but with an additional response
subtracted from this result which is initiated at t = T and continues thereafter. This additional
response must be of the same form as Eq. (k) except that time t1 , which is measured from the
initiation of this change in F (t ) , is now given by t1 = t − T . Thus, we have that the total
displacement response from superposition is given by

α1   α  
sin (ω n t ) − 2  (t − T ) − sin (ω n ⋅ (t − T )) ;
1 1
x=  t − t >T (l)
k  ωn  k  ωn 

Moreover, it is important to note that if one chooses to approximate any applied force F (t ) by n line
segments, each represented by a change in slope α i (positive counterclockwise) beginning at time Ti ,

for this system for example, we would obtain a displacement response for t > T j given by

j
αi  
(t − Ti ) − sin (ω n ⋅ (t − Ti )) ;
1
x=∑ t > Tj (m)
i =1 k  ωn 

180
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

Finally, since the particular solutions obtained by Duhamel’s integral assume homogeneous initial
conditions, we must satisfy the prescribed initial conditions of x = x and x& = 0 at t = 0 by selecting
the constants A and B in the homogeneous solution Eq. (c) as A = 0 and B = x . Thus, the complete
solution to Eq. (a) is given by

α1  
x = x cos(ω n t ) + sin (ω n t ) ;
1
 t − 0≤t ≤T (n)
k  ωn 

and

α1   α  
x = x cos(ω n t ) + sin (ω n t ) − 2 (t − T ) − sin (ω n ⋅ (t − T )) ;
1 1
 t − t ≥T (o)
k  ωn  k  ωn 

It is left as an exercise for the student to verify the conclusions given by Eq. (l) from superposition by
directly applying Eq. 4.18 for t > T .

b) Application of Fourier Series Techniques

Having discussed the necessary integral equations to handle the non-homogeneous portion of linear
differential equations, it is highly desirable here to point out the usefulness of Fourier Series in
connection with the various forms of Duhamel’s integral.

If one encounters a forcing function F (t ) that is periodic, although not harmonic, then one quickly
realizes that the use of Duhamel’s integral without further considerations could lead to an infinite set
of integrals, one corresponding to each discontinuity in F (t ) . A most convenient way of eliminating
this difficulty is through the use of Fourier series. It should be remarked that the following is intended
only to summarize the important properties of Fourier series. For a thorough treatment of the subject
one should consult any of a number of well known textbooks on advanced calculus.

Fourier, in 1822, proved that practically speaking any periodic function can be represented by an
infinite series of sine and cosine terms with appropriate amplitude coefficients. That is, if f (t ) is a
periodic function of period 2 p , then we may represent f (t ) as

181
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S


 nπ t  ∞  nπ t 
f (t ) = a0 + ∑ an cos  + ∑ bn sin  (4.24)
n =1  p  n =1  p 

 nπ t   nπ t 
The sine terms in Eq. 4.24 are referred to as odd functions, since sin −  = − sin  , and are
 p   p 
therefore antisymmetric about the origin of an f (t ) versus t curve. Similarly, the cosine terms are

 nπ t   nπ t 
even functions, since cos  = cos −  , and are symmetric about the origin of an f (t )
 p   p 

versus t curve. The constant term a0 simply shifts the curve of f (t ) versus t uniformly in the
direction of the ordinate. In many cases, this information allows us to determine whether we need
both sine and cosine terms to represent f (t ) .

The coefficients of the series may be determined in the following manner. If we integrate Eq. 4.24
through one period, that is, from q to q + 2 p , we have

q+2 p
2 p a0 = ∫ f (t ) dt
q

or

f (t ) dt
1 q+2 p
2 p ∫q
a0 =

since the sine and cosine terms are all zero when integrated through a full period. Next, if we multiply
 nπ t 
Eq. 4.24 by cos  for a particular value of n and integrate from q to q + 2 p , we find
 p 

1 q+2 p  nπ t 
an = ∫ f (t ) cos  dt (4.26)
p q  p 

 nπ t 
Finally, multiplying Eq. 4.24 by sin  and integrating from q to q + 2 p , we have
 p 

182
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

1 q +2 p  nπ t 
bn = ∫ f (t )sin  dt (4.27)
p q
 p 

Frequently, we encounter q = 0 or q = − p so that the integration takes place over the interval − 0 to
2 p , or − p to p.

In conclusion, with the use of Fourier series for representing any arbitrary periodic forcing function
F (t ) , Duhamel’s integral method is reduced to a single integration.

Example 4.3 Fourier Series of Forcing Function

Determine the Fourier series representation of the periodic forcing function F (t ) shown in Figure 4.8.

F (t )

F0

O P 2P t

Figure 4.8: Periodic triangle wave.

Solution: By a hypothetical extension of F (t ) versus t as indicated in Figure 4.8, we see that F (t ) is


an even function of time and will thus contain no sine terms in its Fourier series expansion.
Accordingly, we find

1 p  t  1 2p  t 
a0 =
1 2p
∫ f (t ) dt = ∫ F0 
 1 − 
 dt + ∫ F0  − 1dt
2p 0 2p 0
 p 2p p
p 

or

183
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

F0
a0 = (a)
2

Similarly,

 nπ t 
f (t ) cos
1 2p
an = ∫
p 0  p 
 dt

1 p  t   nπ t  1 2p  t   nπ t 
an = ∫ F0 1 −  cos
p 0  p  p 
 dt + ∫ F0  − 1 cos
p p p   p 
 dt (b)

Evaluating the integrals, we find

0, when n is even



an =  4 (c)
 n π 2 , when n is odd
F0 2

Therefore, we have

  (2 n − 1)π t  
 ∞
cos  
F (t ) = F0 ⋅  + 2
1 4  p 
2 π ∑
n =1 (2 n − 1)2

(d)
 
 

This value of F (t ) would then be appropriate for use in the superposition integrals for describing the
periodic forcing function shown in Figure 4.8 to a mechanical system.

4.2 Linear Impulse-Momentum Methods for a System of


Particles
In order to develop the principal of linear impulse-momentum for a system of particles, it is
convenient to recall the equation of motion for an arbitrary particle mi of a general system of n
particles given by Eq. 2.15 (referred to Figure 2.5) written as

184
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r n r r
dvi
Fi + ∑ f i , j = mi ; i = 1,2,K , n (4.28)
j =1 dt
j ≠i

r r
where the velocity of mi is written here as vi rather than r&i . Now by multiplying Eq. 4.28 by the

scalar dt, integrating from time t1 to t 2 and summing this result over the system of particles, we
obtain

r 
n
 n r
 n
dv
∑∫  i ∑ i, j  ∑
t2

t2
F + f dt = mi i dt (4.29)
i =1
t1
 j =1  t
i =1 1 dt
 j ≠i 

Performing the required operation in Eq. 4.29, noting that the order of summation and integration may
be interchanged, we find

2
r n
r
FR dt = ∑ mi vi
t2
∫ t1
i =1 1
(4.30)

r n r n n r
where the resultant external force FR = ∑ Fi and ∑∑ f i , j = 0 . Symbolically, we may write Eq.
i =1 i =1 j =1
j ≠i

4.30 as

r n
r
I = ∆ ∑ pi (4.31)
i =1

r r
where I is the resultant impulse acting on the entire system and pi is the linear momentum of each

particle mi .

Equation 4.31 states that for any general system of particles, the total change in linear momentum of
the system of particles during a time interval ∆t is equal to the total external linear impulse exerted on
the system during that time interval.

185
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

In many instances, it is desirable to recast this principal in an alternate form. From the definition of
n
r r
the mass center, we may write for constant mass systems that ∑m vi =1
i i = m vc , which means that the

total linear momentum of the system is analytically equivalent to thinking of the entire mass m as
r
moving with the mass center’s velocity vc . Therefore, Eq. 4.30 takes the form

r
FR dt = m vc 1 = m ⋅ {(vc )2 − (vc )1} = m ∆vc
t2 r 2 r r r
∫ t1
(4.32)

or, symbolically we may write

r r
I = ∆Pc (4.33)

r r
where Pc = m vc .

Thus, Eq. 4.33 states that for any system of particles, the total change in linear momentum of the
system in terms of the motion of its center of mass during a time interval ∆t is equal to the total
external impulse exerted on the system during that time interval. The form of the impulse-momentum
principle given by Eq. 4.33 proves to be most useful in treating a wide variety of problems. Finally,
we should note here that both Eqs. 4.31 and 4.33 are vector equations with, of course, three scalar
component equations:

n
I x = ∆ ∑ ( pi )x = ∆ ( p c )x
i =1
n
I y = ∆ ∑ ( pi ) y = ∆ ( p c ) y (4.34)
i =1
n
I z = ∆ ∑ ( pi ) z = ∆ ( p c )z
i =1

4.3 Conservation of Linear Momentum


A very useful result is obtained from Eq. 4.31 for those situations in which the resultant external force
r r
FR = 0 . For these cases we see that I = 0 and

186
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

n
r
∑m v
i =1
i i = const. (4.35)

Equation 4.35 is called the principal of conservation of linear momentum. Note that if we consider the
r
alternate form of Eq. 4.31, namely Eq. 4.33, then we see that if FR = 0 we may also conclude that

r
m vc = const. (4.36)

Therefore, whenever the total impulse on a system of particles throughout a time interval ∆t is zero,
the velocity of the mass center of the system is constant both in magnitude and direction. Further, we
r
may note that if vc = const. then from Eq. 4.36

r r r
rc = C1 t + C2 (4.37)

r r r
where C1 and C2 are constant vectors. Thus, if C1 = 0 , meaning that the center of mass of the system
r
is initially at rest, and I = 0 , then the center of mass remains at rest.

Example 4.4 Linear Momentum of Man and Boat

A man weighing 150 lb. is standing in the stern of a boat which weighs 100 lb. and is 16 ft. long. If
the man walks to the bow of the boat, how far does the boat move? Neglect the hydrodynamic forces
acting on the boat.

187
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

C.G. of Boat

8'
16'

Figure 4.9: Man walking in boat (initial position).

Solution: It is clear that if we neglect hydrodynamic forces, the external impulse exerted on the
system consisting of the boat and man in the x direction is zero. Thus, since the system was initially at
rest, we may write

mm x& m + mB x& B = 0 (a)

where the subscripts m and B refer to the man and boat respectively. However, from kinematics, we
note that x&m = x& B + x&m / B where x&m / B is the velocity of the man relative to the boat. Consequently,
Eq. (a) becomes

(mm + mB ) x& B + mm x&m / B = 0 (b)

which upon integrating yields

(mm + mB ) xB + mm xm / B = 0 (c)

mm
∴ xB = − x = −9.6 ft (d)
(mm + mB ) m / B

188
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r
An alternative solution is readily available from Eq. 4.37. Recognizing that C1 = 0 , we see that rc is

8(100)
constant. Initially then xc = = 3.2 ft to the right of the man. Thus, after the man has
(150 + 100)
moved to the front of the boat, the center of gravity of the system is clearly 3.2 ft to the left of the man.
Since the position of the system’s mass center must remain fixed in space during the motion, the boat
must have moved to the left a distance of 16 − (3.2 + 3.2) , or 9.6 ft.

4.4 Collision Problems


A very important class of problems which can be studied in an approximate way using linear
momentum concepts is that of collision. If two arbitrary bodies collide in space as shown in Figure
4.10, there is a tendency for them to flatten out along the contact surface shown as the plane of contact
in Figure 4.10. The normal to the plane of contact is called the line of impact and is denoted by the
r
unit vector eN .

y
plane of contact
line of impact
r
en
m1 m2
r
et

Figure 4.10: General collision problem geometry.

According to the impulse-momentum principle for a system of particles given by Eq. 4.31, we may
write that

r r r r r r r
I = ∆p1 + ∆p2 = m1 ⋅ (v1′′ − v1 ) + m2 ⋅ (v2′′ − v2 ) (4.38)

189
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

or

r r t ′′ r r r
m1 v1 + m2 v2 + ∫ FR dt = m1 v1′′ + m2 v2′′ (4.39)
t

r r
where I is the impulse due to the external forces FR acting on the complete system, the double prime
denotes velocities after impact, t denotes the time of initial contact between the masses, and t ′′ the
time at which they terminate contact. The impact time, which measures the time interval of actual
contact between the bodies as they tend to compress together and then separate, is given by ∆t = t ′′ − t
which is ordinarily a very short time interval. Consequently, since the impact forces are internal
forces to the defined system, the integral on the left side of Eq. 4.39 is frequently not a very large term.

r
In particular, if the external forces FR consist only of gravity forces and frictional forces, which are
proportional to gravity with the coefficient of friction being the proportionality constant, the integral
representing the external impulse on the system is frequently negligible compared to the magnitude of
the momentum terms in Eq. 4.39. This occurs since the impact time over which the integration takes
r
place is usually very small and in addition, the resultant of external forces FR is often small compared
to the impact forces acting between the bodies. This is physically reasonable since the impact force
between the bodies is actually the mechanism changing the momentum of the individual masses, and
this impact force acting on each mass is ordinarily very large compared to other active forces.

Therefore, applying the approximation that the external impulse is negligible compared to at least one
of the momentum terms in Eq. 4.39, we reduce the equation to

r r r r
m1 v1 + m2 v2 = m1 v1′′ + m2 v2′′ (4.40)

r r r r
Eq. 4.40 is a statement of conservation of linear momentum of the system where v1 , v2 , v1′′ , and v2′′
refer to the velocities of the mass centers of bodies one and two immediately before and after contact.
It is clear from Eq. 4.40 that in general, there are six scalar unknowns associated with the final
r r r r
velocity vectors v1′′ and v2′′ , and only three scalar equations, given the initial velocities v1 and v2 .

In order that we may proceed in the spirit of demonstrating the basic analytical procedure, without
introducing unnecessary complexities, we will restrict the nature of the problem to a consideration of

190
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

bodies with negligible coefficients of friction on their mutual surface of contact. Consequently, a
consideration of Newton’s second law for each body quickly reveals that all components of velocity
of each body in the plane of contact must remain unchanged. The problem is thus reduced to only two
unknowns, namely v1′′n and v2′′n which are the velocity components along the line of impact.
However, we now only have one independent component equation from Eq. 4.40, namely

m1 v1n + m2 v2 n = m1 v1′′n + m2 v2′′n (4.41)

This analytical approximation is physically reasonable, since the sudden changes in velocity due to
collision are primarily associated with the very large impact force developed between the two bodies
along the line of impact, rather than due to surface frictional effects, as borne out by experimentation.

If it is further assumed that the center of mass of each mass m1 and m2 lies along the line of impact
r
given by eN , then the possibility of impacting rotational motions to either m1 or m2 is eliminated,
since the impact force on each body acts through its center of mass. It is customary to define this
specific problem as central impact, with two subcases described as (a) direct central impact, and (b)
r r
oblique central impact. Direct central impact implies that the initial velocities v1 and v2 are directed
r
along eN . All other cases are referred to as oblique central impact. It is clear that neither their
designation of direct nor oblique add anything analytically new to the problem beyond the concept of
central impact.

Based on the concept of central impact, we may consider a free body diagram of either of the bodies,
taken here as m1 .

r
en
r
c.g. F (t )
m1

Figure 4.11: Free body diagram of m1 .

191
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

Let us then consider the time intervals t to t ′ and t ′ to t ′′ , where t represents the time at the initial
instant of contact, t ′ the time at the instant of the maximum deformation of bodies (i.e., the time at
which they have zero relative velocity) and t ′′ the time at the instant of separation of the bodies. We
may then define the coefficient of restitution for the collision as

t ′′ r
∫ F (t ) dt
t′
e= t′ r (4.42)
∫ F (t ) dt
t

It is clear from Eq. 4.42 that the coefficient of restitution e is an experimental quantity depending on
many parameters, among which the primary ones are the material composition, size, and geometry of
the bodies, their velocities, and the physical parameters of the materials.

F (t )

t t′ t ′′ t

Figure 4.12: Inter-body impact force

If the bodies are perfectly elastic, the force developed during the deformation period t to t ′ and
recovery periods t ′ to t ′′ will be mirror images (see Figure 4.12), since all energy stored in an elastic
deformation is recovered upon separation of the bodies. Thus, e = 1 for a perfectly elastic impact. At
the other end of the spectrum is a perfectly plastic collision, in which there is no rebound, resulting in
e = 0 according to Eq. 4.42. The range of e is thus bounded by

0 ≤ e ≤1 (4.43)

192
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

and practically speaking, e < 1 , although various modern synthetic materials approach the upper
bound. An example, of course, is a super ball.

From Eq. 4.42, we find

m1 ⋅ (v1′′n − v1′n ) v1′′n − v1′n


e= = (4.44)
m1 ⋅ (v1′n − v1n ) v1′n − v1n

where the single prime denotes the velocity at maximum deformation at time t ′ . Similarly, for m2 ,
we have

m2 ⋅ (v′2′n − v2′ n ) v2′′n − v2′ n


e= = (4.45)
m2 ⋅ (v2′ n − v2 n ) v2′ n − v2 n

Combining Eqs. 4.44 and 4.45, and noting that v1′n = v2′ n , we obtain

v′2′n − v1′′n relative velocity of separation


e= = (4.46)
v1n − v2 n relative velocity of approach

Thus, the coefficient of restitution e physically represents the ratio of the components of the relative
velocity of separation of the bodies to their relative velocity of approach along the line of impact.
Therefore, Eqs. 4.41 and 4.46 are sufficient for treating all central impact problems which can be
approximated as frictionless bodies, leading to the following general expressions for the velocities
after impact

v1′′n =
(m1 − e m2 ) v1n + (1 + e) m2 v2 n (4.47)
m1 + m2

v2′′n =
(1 + e ) m1 v1n + (m2 − e m1 ) v2 n (4.48)
m1 + m2

The energy lost to the system during impact is given by

Energy Lost = T2 − T1 =
1
2
[ 2 2

2
] 2
[
m2 (v′2′n ) − (v2 n ) − m2 (v′2′n ) − (v2 n )
1 2
] (4.49)

193
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

For the case of e = 1 , substitution of Eqs. 4.47 and 4.48 into Eq. 4.49 shows that T2 − T1 = 0 . This is
to be expected for a perfectly elastic collision, since all of the energy stored in the bodies during
deformation is recovered upon rebound. For e < 1 , T2 − T1 < 0 since energy is dissipated due to the
permanent deformation of the bodies.

Example 4.5 Impacting Cars

Car 1, weighing 2000 lb., was traveling northward through an intersection when struck broadside by
car 2, weighing 4000 lb. which was traveling eastward. After impact, car 1 skidded to a stop in 30 ft.
on a surface with coefficient of friction µ = 0.6 in the direction shown and car 2 skidded 20 ft. as
shown on a surface for which µ = 0.5 . From the given measurements that can be accurately and
easily made, (a) approximate the velocity of each car before impact, and (b) determine the
approximate coefficient of restitution between the two vehicles for a broadside collision. Would this
approximate analysis serve any useful purpose in a court of law?
y
N µ = 0.6
1
30' Final Positions
Skid of Vehicles
Marks
30°
r
v2 2 2
x
20'
µ = 0.5
1
r
v1
Figure 4.13: Broadside collision of car 2 into car 1.

Solution: By using the work-energy method for each car after impact, we can determine its velocity
immediately after impact. Thus, we have for car 2 during its skidding phase

ωk = ∆T = T f − Ti (a)

194
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

∴ −4000 (0.5)(20 ) = 0 − (v2′′ )2


1 4000
2 32.2

or v2′′ = 25.4 fts . Thus, immediately after impact,

r r
v2′′ = 25.4 i []
ft
s (b)

Similarly, for car 1 we have

− 2000 (0.6 )(30 ) = − (v1′′)2


1 2000
2 32.2

from which v1′′ = 34 fts . Thus,

r r r
v1′′ = 29.5 i + 17 j []
ft
s 1 (c)

for car 1 immediately after impact. Therefore, we can approximate the initial velocity of car 1 as

r r
v1 = 17 j []
ft
s (d)

since its component of velocity along the plane of contact is unchanged for frictionless impact. The
approximate plane of contact and line of impact, considered here to be along the y and x directions
respectively, would be evident from the dents in both cars.

In order to determine the initial velocity of car 2, we can consider that the impact time is very short (of
the same order of magnitude as the length of time in which the loud crash is heard) and apply the
conservation of momentum principle along the line of impact during impact. Thus, referring to Eq.
3.41, we may write

2000
(0) + 4000 (v2 n ) = 2000 (29.5) + 4000 (25.4) (e)
32.2 32.2 32.2 32.2

Solving Eq. (e) for v2 n , we obtain

v2 n = 40.2 ft
s (f)

195
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

from which we conclude that the initial velocity of car 2 was

r r
v2 = 40.2 i []ft
s (g)

Since the coefficient of restitution involves only the components of velocity along the line of impact,
we may determine e for this “experiment” as follows.

relative velocity of separtion 29.5 − 25.4


e= = ≈ 0.1 (h)
relative velocity of approach 40.4 − 0

This value of e could be compared with other experimental values (run under better controlled
conditions) to further check the validity of these approximate results.

4.5 Angular Momentum Methods for a Particle


We have previously operated on Newton’s second law with the vector dot product in Section 3.4
leading to the very important work-energy concepts. Let us now operate on Newton’s second law by
using the vector cross product. In order that we may generalize the results, we will consider the
moment of each side of Newton’s second law taken with respect to an arbitrarily moving point P as
shown in Figure 4.14. This will lead us to the well known angular momentum methods of analytical
mechanics.

r
F
m

r
r ρ
r
r
rP
P
O
y

Figure 4.14: A force acting on a particle causing a moment about O.

196
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r
The particle of mass m is located with respect to the origin of the x, y, z reference frame by r and with
r
respect to the arbitrary point P by ρ . Point P is located with respect to the inertial reference frame by
r
rP . Thus, by taking the moment of Newton’s law with respect to point P, we have

r r
[ r
ρ × F = m &r& ] (4.50)

or

r r r r
ρ × F = ρ × m &r& (4.51)

r
The left side of Eq. 4.51 represents the moment of the resultant force F acting on mass m about point
P and the right hand side may be rewritten as

r
MP =
d r
dt
(r r r
ρ × m r& − ρ& × m r&) (4.52)

r r r
Since ρ& = r& − r&P , we can write Eq. 4.52 in the form

r r& r r
M P = H P + m r&P × r& (4.53)

where

r r r
H P = ρ × m r& (4.54)

r r r
It is seen that m r& is the linear momentum of mass m, so that ρ × m r& is obviously the moment of
r
linear momentum with respect to point P. Thus, H P is called the moment of momentum, or simply
the angular momentum with respect to point P. Eq. 4.53 states that the moment of the resultant force
acting on a particle m about a point P moving in an arbitrary way equals the time rate of change of
moment of momentum with respect to point P plus a correction term which accounts for the fact that
P may be a moving reference point.

Eq. 4.53 reduces to simply

r r&
MP = HP (4.55)

197
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

for the special cases corresponding to:

r
1. r&P = 0 ; point P is a fixed point

r r
2. r&P || r& ; point P and the particle have parallel velocity vectors

r r r
Another important form of Eq. 4.53 may be obtained by noting that since r& = r&P + ρ& , Eq. 4.54 can be
written as

r r r r r
H P = ρ × m ρ& + ρ × m rP

or

r r r r
H P = hP + ρ × m r&P (4.56)

where

r r r
hP = ρ × m ρ& (4.57)

r r
In Eq. 4.57, ρ& is the velocity of m relative to point P, so that m ρ& is a linear momentum vector
r r
relative to P. Consequently, ρ × m ρ& is called the moment of relative linear momentum with respect
r
to point P or simply relative angular momentum hP . Differentiating Eq. 4.56 and substituting into Eq.
4.53, we obtain

( )
r r& r r r r r
M P = hP + ρ × m &r&P + m r&P × r& − ρ&

r r r
Since r& − ρ& = r&P , we have simply

r r& r r
M P = hP + ρ × m &r&P (4.58)

Eq. 4.58 states that the moment of the resultant force acting on a particle of mass m about point P
equals the time rate of change of the relative angular momentum with respect to P plus an additional

198
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

term accounting for the fact that point P may be moving in an arbitrary way. Eq. 4.58 may be
simplified if:

r
1. r&P = const. ; point P moves with a constant velocity vector of which a fixed point is a special
case.

r r r
2. ρ || &r&P ; the acceleration of point P and ρ are parallel.

In this event, we have the statement that moment equals the time rate of change of relative angular
momentum written mathematically as

r r&
M P = hP (4.59)

Finally, we note that if we select point P as the origin O which is a fixed point in the x, y, z coordinate,
we may write from Eq. 4.53 that

r r&
M O = HO (4.60)

Eq. 4.60 states that the moment of the resultant force about the origin of the inertial reference frame is
equal to the time rate of change of the angular momentum H O about O. In order to further clarify
these concepts, we will consider the following example.

Example 4.6 Motion of Pendulum on Cart using Angular Momentum

Shown in Figure 4.15 is a carriage which is moving to the right on a horizontal track with velocity
r r r r
v = q& i and acceleration a = q&& i . Suspended from the carriage by a two-dimensional, frictionless
hinge at point A is a mathematical pendulum of length l and mass m. The carriage is built so that it
r
may rotate relative to its wheel base at a rate ω1 as shown. If at the instant of consideration, the axis
of the frictionless hinge is parallel to the z-axis, determine the equations of motion of the pendulum
for:

r r r
a) a = v = ω1 = 0

199
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r r r
b) ω1 = 0 , but a and v ≠ 0

r r r
c) ω1 , a , v ≠ 0

r
y ω1

r r r r
A v = q& i a = q&& i T
x θ

l
z

mg
θ
m (a)

Figure 4.15: Cart with carriage that rotates relative to wheel base, and (a) a free body diagram of the mass.

Solution:

r r r
a) a = v = ω1 = 0

r r r
If a = v = ω1 = 0 , then point A is a fixed point and we apply Eq. 4.55 directly to the motion of mass m
about point A. Thus, we have

r r&
MA = HA (a)

or

(l sin(θ )ir − l cos(θ ) rj )× (− m g rj )


= [(l sin (θ ) i − l cos(θ ) j )× {m θ& k × (l sin (θ ) i − l cos(θ ) j )}]
d r r r r r (b)
dt

Performing the required operations in Eq. (b), we obtain

r d
− m g l sin (θ ) k =
dt
(r
m l 2 θ& k ) (c)

200
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r&
Differentiating the right side of Eq. (c), where of course k = 0 , we find

m l 2 θ&& + m g l sin (θ ) = 0 (d)

or

θ&& + Ω 2 sin (θ ) = 0 (e)

g
where Ω 2 = .
l

If we restrict attention to small values of θ , then sin (θ ) ≈ θ and Eq. (c) becomes

θ&& + Ω 2 θ = 0 (f)

Ω g l
The frequency of oscillation is thus given by f = = and the period of motion becomes
2π 2π

1 2π
T= = . The solution of Eq. (f) is the same as given in Example 2.3; namely
f g l

θ0
θ = θ 0 cos(Ω t ) + sin (Ω t ) (g)

It is of interest to consider the solution of Eq. (e) if θ is not small. If we note that

dθ& dθ& dθ & dθ&


θ&& = = =θ (h)
dt dθ dt dθ

then we may separate variables in Eq. (e) yielding

θ& dθ = −Ω 2 sin (θ ) dθ (i)

For the initial conditions of θ = θ 0 and θ& = θ&0 , integration of Eq. (i) yields

θ& 2 − θ&0 2 = 2 Ω 2 (cos(θ ) − cos(θ 0 )) (j)

201
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

In order that we may proceed without undue algebra difficulty, let us assume θ&0 = 0 . Then separating
variables once more in Eq. (j) and integrating, we find

t θ dθ
∫ dt = − ∫θ
t0 0
2 Ω (cos(θ ) − cos(θ 0 ))
2
(k)

Note the sign in Eq. (k) has been chosen so that physically the direction of θ& is correct.

Now, since we are primarily interested in the period of motion T, if we select the upper limit of the
integral in Eq. (k) as θ = 0 and noting that the motion is symmetrical about θ = 0 we may conclude
that

dθ θ0 dθ
T = −4 ∫
0
= 4∫ (l)
θ0
2 Ω 2 (cos(θ ) − cos(θ 0 )) 0
2 Ω 2 (cos(θ ) − cos(θ 0 ))

θ 
Using the trigonometry relationship cos(θ ) = 1 − 2 sin 2   , Eq. (l) becomes
2

2 θ0 dθ
Ω ∫0
T= (m)
θ  θ 
sin 2  0  − sin 2  
2 2

Finally, if we make the change of variables

θ  θ 
sin   = sin  0  sin (φ ) = k sin (θ ) (n)
2 2

from which

2 k cos(φ ) dφ
dθ = (o)
1 − k 2 sin 2 (φ )

Eq. (m) becomes

202
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

4 π2 dφ
T= ∫
Ω 0 1 − k 2 sin 2 (φ )
(p)

where 0 < k 21 . The integral in Eq. (p) is called a complete elliptic integral of the first kind.

Occasionally, one also encounters elliptic integrals of the second and third kind, which are
respectively

φ
F (k1 φ )2 = ∫ 1 − k 2 sin 2 (φ ) dφ (q)
0

and

φ dφ
F (k1 φ )3 = ∫ (r)
0 2 2
(
1 − k sin (φ ) ⋅ 1 + a 2 sin 2 (φ ) )
where a 2 ≠ k 2 and 0 < k 2 < 1 . It turns out that several important functions can be expressed in terms
of elliptic integrals of the first, second, and third kinds. These integrals are termed complete if in the
limits of integration, φ = π 2 . The values of these integrals, both complete and incomplete, are
tabulated for a wide range of parameters.

A comparison of the period versus θ 0 is given in Table 4.1 for both the elementary small angle theory

and the elliptic integral solution. Denoting the approximate period based on small angle theory by Ta

F (k , φ )1 Ta which is
2
and the exact period based on the elliptic integral solution by Te , we have Te =
π
shown for various initial angles θ 0 in Table 4.1.

203
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

Table 4.1: Comparison of small angle theory and elliptic


integral solutions.
θ0 F (k , φ )1 Te Ta

10º 1.5738 1.0019

20º 1.5828 1.0076

30º 1.5981 1.0174

40º 1.6200 1.0313

50º 1.6490 1.0498

60º 1.6858 1.0732

70º 1.7313 1.1022

80º 1.7868 1.1375

90º 1.8541 1.1804

It is seen that the exact period Te increases as a function of θ 0 and in all cases is greater than Ta .

However, for initial angles θ 0 less than 50º, the difference is less than 5%.

r r r
b) ω1 = 0 , but a and v ≠ 0

Returning to the statement of the problem, we now consider the equation of motion if point A is
r r r r
moving with the velocity v = q& i and acceleration a = q&&i . We now may apply either Eq. 4.58 for a
moving reference point or Eq. 4.60 for a fixed point O. So that we may emphasize the utility of Eq.
4.58, let us first use a fixed reference point O. Summing moments about point O we have from Eq.
4.60

r r&
M O = HO (s)

for which

204
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

[(q + l sin (θ ))ir − l cos(θ ) rj ]× [− T sin(θ )ir + (T cos(θ ) − m g ) rj ]


= [{(q + l sin (θ )i − l cos(θ ) j }× m ⋅ {(q& + l θ& cos(θ ))i + l θ& sin (θ ) j }]
d r r r r (t)
dt

From Eq. (t), we obtain

r
(T q cos(θ ) − m g q − m g l sin (θ ))k
r (u)
= (m l 2 θ&& + m l q&& cos(θ ) + m q l θ&&sin (θ ) + m q l θ& 2 cos(θ ))k

r r
It is clear that Eq. (u) is not sufficient to solve the problem. We could, of course, use F = m a to
eliminate T. However, rather than take this approach, we note that if we had summed moments about
point A applying Eq. 4.58, then T would not have even entered into the problem, since T passes
through point A. Accordingly, we write from Eq. 4.58

r r& r r
M A = hA = ρ × m &r&A (v)

r
where ρ is a vector from point A to mass m. Therefore, we have

(l sin (θ )i − l cos(θ ) j )× (− m g j )
r r r

= [(l sin (θ )i − l cos(θ ) j )× m ⋅ (l θ& cos(θ )i + l θ& sin (θ ) j )]


d r r r r
(w)
dt
+ (l sin (θ )i − l cos(θ ) j )× m q&&i
r r r

r
Thus, since the equation of motion for θ involves only the k components of Eq. (w) we obtain

r r
(
− m g l sin (θ ) k = m l 2 θ&& + m q&&l cos(θ ) k )
or

q&&
θ&& + Ω 2 sin (θ ) = − cos(θ ) (x)
l

where the acceleration q&& may be considered given as a function of time. Note that if θ is restricted to
small values, then Eq. (x) becomes

205
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

q&&
θ&& + Ω 2 θ = − (y)
l

since a0 = q&& is not constant here.

r r r
c) ω1 , a , v ≠ 0

r r r
For the case of ω1 , a , v ≠ 0 , it is clear from part (b) that the more appropriate approach is to deal
r
with a moving reference point, since the summation of moments, M A , involves only known
quantities. Thus, again we write

r r& r r
M A = hA + ρ × m &r&A (aa)

r
where ρ is the vector from A to mass m. Now

r r r
h A = ρ × m ρ&
(
r r
)
r
{( r
)( r
r
r
= l sin (θ )i − l cos(θ ) j × m ⋅ ω1 j + θ& k × l sin (θ )i − l cos(θ ) j )} (bb)
r r
= m l 2 ω1 sin (θ )cos(θ )i + m l 2 ω1 sin 2 (θ ) j + m l 2 θ& k

r r r r
which describes hA in terms of components along the i , j , k directions attached to the rotating cart.

r&
Applying Eq. 1.55 to determine hA , we have

r& r r r r r
hA = h&x i + h&y j + h&z k + ω × hA (cc)

r r
where ω = ω1 j and h&x , h&y , and h&z are the time rates of change of magnitudes of the components of
r r r r r r
hA = hx i + hy j + hz k as given in Eq. (bb). Further, we note that r&&A = q&& i .

Substituting these values into Eq. (aa) we find

206
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r r r
(l sin (θ )i − l cos (θ ) j )× (− m g j )
[ r
= m l 2 ω& 1 sin (θ )cos (θ ) + m l 2ω1 θ& cos 2 (θ ) − m l 2 ω1 θ& sin 2 (θ ) i ]
[ ] [ ]
r r
+ m l 2 ω& 1 sin 2 (θ ) + 2 m l 2 ω1 θ& sin (θ )cos (θ ) j + m l 2 θ&& k (dd)
r
[ r r
+ ω1 j × m l 2 ω1 sin (θ )cos (θ )i + m l 2 ω1 sin 2 (θ ) j + m l 2 θ& k
r
]
[ r r
+ l sin (θ )i − l cos (θ ) j × m q&& i
r
]
Eq. (dd) yields the following set of scalar equations:

0 = m l 2 ω&1 sin (θ )cos(θ ) + 2 m l 2ω1 θ& cos 2 (θ )


0 = m l 2 ω& sin 2 (θ ) + 2 m l 2 ω θ& sin (θ )cos(θ )
1 1 (ee)
− m g l sin (θ ) = m l 2 θ&& − m l 2 ω1 sin (θ )cos(θ ) + q&& m l cos(θ )
2

which after simplifying becomes

0 = ω&1 sin (θ )cos(θ ) + 2 ω1 θ& cos 2 (θ )

0 = ω&1 sin 2 (θ ) + 2 ω1 θ& sin (θ )cos(θ ) =


d
dt
(
ω1 sin 2 (θ ) ) (ff)

0 = l θ&& − l ω1 sin (θ )cos(θ ) + q&&cos(θ ) + g sin (θ )


2

One may show that only two of Eqs. (ff) are independent, although they are nonlinear and not easily
solved.

However, if we restrict attention now to the case of ω&1 = 0 , then Eqs. (ff) reduce to

0 = 2 ω1 θ& cos 2 (θ )

0=
d
dt
(
ω1 sin 2 (θ ) ) (gg)

0 = l θ&& − l ω1 sin (θ )cos(θ ) + q&& cos(θ ) + g sin (θ )


2

From the first of Eqs. (gg), θ& = 0 for θ ≠ π 2 , and from the second ω1 sin 2 (θ ) = C where C is a
constant. Substituting these quantities into the third equation, we obtain

q&& cos(θ ) + g
ω12 = (hh)
l sin (θ )

207
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

If q&& = 0 we find that the steady state angle θ becomes

 g 
θ = sin −1  
2 
(ii)
ω
 1 
l

4.6 Angular Impulse-Momentum Principle for a Particle


If we restrict attention in prior Section 4.5 to the situation for which the reference point P moves with
r r r r
(
constant velocity; &r&P = 0 , or the acceleration of P is parallel to ρ ; &r& || ρ , then the relative angular )
momentum principle reduces to Eq. 4.59

r
MP =
d r
dt
hP ( ) (4.61)

Multiplying both sides of Eq. (4.61) by dt and integrating from time t1 to t 2 we find


t2

t1
r 2 r
( ) ( ) − (hr )
r
M P dt = ∫ d hP = hP
1 2 P 1 (4.62)

Symbolically, we may write

r
Mˆ P = ∆hP M (4.63)

t2 r r
t1
r
where of course Mˆ P = ∫ M P dt and ∆hP = hP ( ) − (hr ) .
2 P 1 The quantity ∫
t1
t2 r
M dt is defined as the

angular impulse acting through the time interval ∆t = t 2 − t1 . Equation 4.63 states that the change in
relative angular momentum of a particle through a time interval ∆t is equal to the angular impulse
with respect to P delivered to the particle through the same time interval ∆t .

If we turn our attention to Eq. 4.60, written as

( )
r d r
MO = HO (4.64)
dt

208
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

then it follows immediately by a similar procedure

( ) − (H )
t r r r
∫ t1
M O dt = H O 2 O 1 (4.65)

or

r
Mˆ O = ∆H O (4.66)

Equation 4.66 states that the change in absolute angular momentum of a particle through a time
interval ∆t is equal to the angular impulse with respect to O delivered to the particle through the same
time interval ∆t .

From Eq. 4.63, we note that if the angular impulse Mˆ P = 0 , then

r
hP = const. (4.67)

Equation 4.67 is termed the principal of conservation of relative angular momentum. Similarly, if the
angular impulse Mˆ O = 0 , then

r
H O = const. (4.68)

and we have the principal of conservation of absolute angular momentum for a particle.

Finally, we should note that the concepts of the unit step function and unit impulse function may also
be applied to Eqs. 4.63 and 4.66 in a similar way as done in studying linear impulse-momentum
methods in Section 4.1 for a large number of problems.

4.7 Angular Momentum Methods for a System of Particles


Again, let us consider a general system of particles using the previously adopted notation shown in
Figure 3.12 when we discussed work-energy methods for a system of particles in Section 3.4.

209
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

y
r
r fj
r r f j ,i
fi , j mj
fi
mi r
rj
r
r r ρj
ri ρ
r i P ρr
rP r c
rc C (mass center)
x

mn

Figure 4.16: Vector definitions for forces acting on arbitrary masses.

The equation of motion for the mass mi is given by

r n r
r
Fi + ∑ f i , j = mi &r&i (4.69)
j =1
j ≠i

r
If we take the cross product of each side of Eq. 4.69 by ρ i and sum over the system of particles we
have

r n r 
n
 r r& 

i =1
ρ i × Fi + ∑ f i , j = mi ri
 j =1
&

(4.70)
 j ≠i 

or

n
r r n n
r r n
r r
∑ρ
i =1
i × Fi + ∑∑ ρ i × f i , j = ∑ ρ i × mi &r&i
i =1 j =1 i =1
(4.71)
j ≠i

r
for i ≠ j , since f i ,i is not defined.

210
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

The first term on the left hand side of Eq. 4.71 clearly represents the moment of all external forces
r n
r r
about P and we designate it by M P = ∑ ρ i × Fi . The second term on the left hand side of Eq. 4.71
i =1

represents the moment of all internal forces about P. This term will be zero if all internal forces occur
in equal, opposite, and collinear pairs. All known internal forces occur in equal and opposite pairs,
although not necessarily collinear, as is the case with certain electromagnetic force fields. We will
assume here that they are collinear and thus

n n
r r
∑∑ ρ
i =1 j =1
i × f i, j = 0 (4.72)
j ≠i

It is of interest to note that if they are not collinear then one would not have symmetry of the stress
tensor in dynamic elasticity problems or, for that matter, static elasticity problems.

Introducing these simplifications into Eq. 4.71, we have

r n
r r
M P = ∑ ρ i × mi &r&i
i =1

which can be written as

r d n r r n r r
M P =  ∑ ρ i × mi r&i  − ∑ ρ& i × mi r&i (4.73)
dt  i =1  i =1

However, we note from geometry along with the definition of the mass center that

r r r r r
( r r
) r r
n n

∑ ρ& i × mi r&i = ∑ ρ& i × mi ⋅ r&P + ρ& i = m ρ& c × r&P = m r&c × r&P


i =1 i =1
(4.74)

We then may write Eq. 4.73 in the form

r r& r r
M P = H P + m r&P × r&c (4.75)

r
where H P , called the absolute angular momentum of the system, is given by

211
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r n
r r
H P = ∑ ρ i × mi r&i (4.76)
i =1

Equation 4.75 states that for a system of particles, the moment of all external forces acting on the
system with respect to P equals the time rate of change of the absolute angular momentum with
respect to P of tall of the particles of the system plus an additional term which accounts for the motion
of point P.

Eq. 4.75 can be simplified to a more convenient form given by

r r&
MP = HP (4.77)

r
(1) r&P = 0 ; point P is at rest
 r r
(2) rP = rc ; point P is the mass center of the system
if :  r
(3) r&c = 0 ; the center of mass is at rest
(4) r r
r&c || r&P ;
 points P and C have parallel velocity vectors

r
If r&P = 0 for which point P is at rest, then it is often convenient to simply select point P as the origin
r r
(i.e., P and O are coincident, thus ρ i = ri ) and we may write

r r&
M O = HO (4.78)

r n
r r
where H O = ∑ ri × mi r&i . Equation 4.78 thus states that the moment of all external forces with
i =1

respect to a fixed point O, which has been selected as the origin of the inertial, is equal to the time rate
of change of the absolute angular momentum of the system.

An important relationship between absolute and relative angular momentum can be obtained at this
r r r
point by substituting r&i = r&P + ρ& i into Eq. 4.76 to yield

r r r r
( r r
) r r
n n
H P = ∑ ρ i × mi ⋅ r&P + ρ& i = m ρ c × r&P + ∑ ρ i × mi ρ& i
i =1 i =1

or

212
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r r r r
H P = hP + m ρ c × r&P (4.79)

r
Therefore hP , called the relative angular momentum of the system with respect to point P, is defined
by

r n
r r
hP = ∑ ρ i × mi ρ& i (4.80)
i =1

r r
Note that since ρ& i is the velocity of mass mi relative to point P, mi ρ& i is essentially a linear
momentum vector relative to point P. Thus, the name “moment of relative linear momentum”, or
r
simply relative angular momentum hP is associated with Eq. 4.80. This quantity is sometimes also
called apparent angular momentum.

Differentiating Eq. 4.79 and substituting into Eq. 4.75, we obtain

r r& r r r r r r
M P = hP + m ρ& c × r&P + m ρ c × &r& − m r&c × r&P

or

( )
r r& r r r r r
M P = hP + m ρ c × &r& − m r&c − ρ& c × r&P (4.81)

r r r
However, since r&c − ρ& c = r&P , we have

r r& r r
M P = hP + m ρ c × &r&P (4.82)

Thus, Eq. 4.82 relates the moment of all external forces about point P to the time rate of change of
relative angular momentum with respect to P plus an additional term accounting for the fact that point
P may be in motion.

Equation 4.82 reduces to an important simplified form given by

r r&
M P = hP (4.83)

213
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

for the following three cases:

1) &rr& = 0 ; point P has constant velocity of which a fixed point is a special case
P

r
2) ρ c = 0 ; point P is selected as the mass center

r r
3) ρ c || &r&P ; point P has an acceleration vector that is parallel to the vector from P to the mass
center C.

In concluding, there are two other general relationships between the various angular momentum
vectors which will prove useful in our later work. It is left as an exercise for the student to verify that:

r r r r
1) hP = hc + m ρ c × ρ& c (4.84)

r
where hc is the relative angular momentum with respect to the mass center of the system of particles,
and

r r r r
2) H P = hc + m ρ c × r&c (4.85)

Example 4.7 Model of a Spring-Lever System

A model of a spring-restrained lever system is formed by bending a massless slender rod into a rigid
right angle of equal legs as shown in Figure 4.17. Located at points A and B are equal concentrated
masses m. A torsional spring of spring constant K T is located at pivot point O and it is unstressed for
the position θ = 0 . Determine the equation of motion of the lever system and determine its natural
frequency of oscillation for small angles θ .

214
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

y
x
O
kT

l m

l
θ
r

m
A r
er
Figure 4.17: Spring restrained lever system model.

r r&
Solution: By direct application of Eq. 4.78, M O = H O for the system. We have

r r r
− m g l sin (θ ) k − m g ⋅ (l sin (θ ) + l cos(θ )) k − K T θ k

=
d r
dt
( r r r r
( r
l er × m l θ& eθ + (l er + l eθ ) × m ⋅ l θ& eθ − l θ& er )) (a)

=
d
dt
( r r
) r
m l 2 θ& k + 2 m l 2 θ& k = 3 m l 2 θ&& k

Simplifying Eq. (a), we find

g 
θ&& +  (2 sin (θ ) + cos(θ )) + K T 2  θ = 0 (b)
 3l 3ml 

Obviously, Eq. (b) is a nonlinear differential equation whose general solution is unknown. However,
by limiting the analysis to a consideration of small oscillations for which sin (θ ) ≈ θ and cos(θ ) ≈ 1
we have

g 
θ&& +  (2θ + 1) + K T 2  θ = 0 (c)
 3l 3ml 

Furthermore, since the small angle assumption requires θ (in radians) to be small, we may conclude
that for θ << 1 2 we have

215
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

g K  g
θ&& +  + T 2 θ = − (d)
 3l 3 m l  3l

The solution of Eq. (d) is of the form

θ = A sin (Ω t ) + B cos(Ω t ) + C (e)

where

g
g K 3l
Ω2 = + T2 ; C=− (f)
3l 3 m l 2 g KT 
 3l + 
 3ml 2 

Thus, the natural frequency is given by

Ω 1 g K
f = = + T2 (g)
2π 2π 3l 3 m l

Example 4.8 Elastic Mechanism on Rotating Platform

The entire mechanical system shown in Figure 4.18 rotates about the vertical z axis at a constant
r
angular velocity ω . In addition, the uniform bar AB of total mass m and length l is hinged to the rigid
rotating platform at A in such a way that it has an additional rotational degree of freedom described by
the angle θ measured in the rotating SOA plane. Bar AB rotates with the angular velocity θ& and
angular acceleration θ&& relative to the rotating platform and the spring of spring constant K is
unstretched at θ = 0 . Determine the equation of motion describing the rotational motion of bar AB as
it vibrates through small angles θ about pivot point A.

216
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

z r
r
ω =ωk
K B
S
θ θ&,θ&&
a
r
er dS
l
S
O
A r y

R

Figure 4.18: Hinged bar on rotating platform.

Solution: As discussed previously in Example 3.7, the slender rod can be thought of for analytical
purposes as a continuous distribution of infinitesimal particles, each of length ds and mass
m
dm = ds , as shown in Figure 4.18. Thus, using the basic mathematical definition of the integral as
l
the limit of a summation, we may integrate the angular momentum contributions of each elemental
mass in order to calculate the total angular momentum of the rod AB.

Since point A is a moving point, we may conveniently use the relative angular momentum principle
given by Eq. 4.82; namely

r r& r r
M A = hA + m ρ c × &r&A (a)

r
where m is the total mass of the rod, ρ c is a vector from point A to the mass center of the rod given by

r l r
ρ c = er (b)
2

and where the acceleration of point A is given by

r r r r
r&&A = − a ω 2 j = − a ω 2 sin (θ )er − a ω 2 cos(θ )eθ (c)

217
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r
The contribution to the total relative angular momentum hA due to the element Q of length ds is of the
form

r r r
dh A = ρ × ρ& dm (d)

where

ρ& = vQ − v A = {ω × (a j + s er ) + s θ& eθ }− ω × a j
r r r r r r r r r
(e)

or

r r r
ρ& = −ω s sin (θ )i + s θ& eθ (f)

Therefore, we have

r r
{ r r m
dh A = s er × − ω s sin (θ )i + s θ& eθ
l
ds }
(g)
m
l
{ r r
= − ω s 2 sin (θ )eθ − s 2 θ& i ds }

Integrating along the length of the rod, we have

r r m l
l 0
{ r r
h A = ∫ dhA = ∫ − ω s 2 sin (θ )eθ − s 2 θ& i ds }
(h)
r 1 r
= − m l 2 ω sin (θ )eθ − m l 2 θ& i
1
3 3

which in terms of the rotating x, y, z rectangular components becomes

r r 1 r 1 r
h A = − m l 2 θ& i − m l 2 ω sin (2θ ) j + m l 2 ω sin 2 (θ ) k
1
(i)
3 6 3

r r r
Thus, writing the bearing reactions on the rod as M b = M y j + M z k , we have from Eq. (a) that

 1 r r r& m l a ω 2 r
 − m g l sin (θ ) + K l 2
sin (θ ) cos (θ ) i + M b = h A + cos(θ )i (j)
 2  2

218
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

Recalling the method for differentiating a general vector that is described in terms of rectangular
components referred to a rotating frame (section 1.6) we have

 1 r r
 − m g l sin (θ ) + K l 2
sin (θ ) cos (θ )i + M b
 2 
r r 1 r
= − m l 2 θ&&i − m ω l 2 θ& cos(2θ ) j + m ω l 2 θ& sin (2θ ) k
1 1
(k)
3 3 3
r r 1 r
+ ω k × hA + m l a ω 2 cos(θ )i
2

which reduces to

 1 r r
− m g l sin (θ ) + K l sin (θ )cos(θ ) i + M b
2

 2 
 1 r
= − m l 2 θ&& + m l 2 ω 2 sin (2θ ) + m l a ω 2 cos(θ ) i
1 1
(l)
 3 6 2 
r 1 r
− m ω l 2 θ&(1 + cos(2θ )) j + m ω l 2 θ& sin (2θ ) k
1
3 3

Thus, for small angles θ , we obtain the differential equation of motion of the rod AB by equating the
r
i components in Eq. (l) resulting finally in the equation

1 2 &&  2 1 1  1
m l θ +  K l − m g l − m l 2 ω 2 θ = m l a ω 2 (m)
3  2 3  2

whose general solution is given by

θ = A sin (Ω t ) + B cos(Ω t ) + C (n)

where

k 3g
Ω= 3 −ω2 − (o)
m 2l

and

219
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

aω 2
C= (p)
Kl 2
2 − g − lω2
m 3


Notice that the natural frequency of oscillation given by f = is appreciably affected by the

rotational speed ω of the system. As a matter of fact, if the speed ω becomes sufficiently large that
the coefficient multiplying θ in Eq. (m) becomes negative for a given system, then the form of
solution involves hyperbolic sines and cosines written as

θ = A sinh (Ω t ) + B cosh (Ω t ) + C (q)

which becomes unbounded as time increases. Thus, the system tends to diverge from the region of
θ = 0 , flying outward under the excessive action of centrifugal force. The position θ = 0 is termed
an unstable dynamic equilibrium position. Thus, stable oscillations of the bar AB in the neighborhood
of θ = 0 requires that the spring stiffness

1 1 mg
K > mω 2 + (r)
3 2 l

r
Finally, the bearing reaction M b acting on the rod at A is given from Eq. (l) as

r r 1 r
M b = − m l 2 ω θ& ⋅ (1 + cos(2θ )) j + m l 2ω θ& cos(2θ ) k
1
(s)
3 3

where θ and θ& are known from Eq. (n) and its first derivative.

4.8 Angular Impulse Momentum Principle for a System of


Particles
Restricting attention to the special, but very important cases for which the only term on the right side
of the angular momentum principle is a time rate of change of angular momentum term, we recall
Eqs. 4.77, 4.78, and 4.83, summarized below:

220
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

A. Absolute Angular Momentum

r r&
MP = HP (4.77)

r
for (1) r&P = 0
r r
(2) rP = rc

r
(3) r&c = 0

r r
(4) r&c || r&P

B. Fixed Reference Point

r r&
M O = HO (4.78)

C. Relative Angular Momentum

r r&
M P = hP (4.83)

r
for (1) &r&P = 0
r
(2) ρ c = 0

r r
(3) ρ c || &r&P

Thus, upon multiplying each of the above Eqs. 4.77, 4.78, and 4.83 by the differential time dt and
integrating from time t1 to t 2 we obtain the respective principals of angular impulse-momentum for a
system of particles as follows:

( ) − (H )
t2 r r r r
∫ t1
M P dt = ∆H P = H P 2 P 1 (4.86)

( ) ( )
t2 r r r r
∫ t1
M O dt = ∆H O = H O 2 − H O 1
(4.87)

221
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S


t2

t1
r r r
( ) ( )
r
M P dt = ∆hP = hP 2 − hP 1
(4.88)

Each of these cases may be stated in words as the angular impulse on any system of particles equals
the change in angular momentum of the system.

Finally, we note that if the angular impulse is zero (i.e., Mˆ P = Mˆ O = 0 ), then

r
H P = const. (4.89)

or

r
H O = const. (4.90)

or

r
hP = const. (4.91)

Equations 4.89 to 4.91 represent the various principles of conservation of angular momentum for
general systems of particles. It should be emphasized that these are vector equations and, although in
many cases the total angular momentum is not conserved, certain of its components may be
conserved. This turns out to be very useful in the solution to many problems.

We shall see that conservation of angular momentum is a fundamental equation in the stuffy of orbital
motion. In these situations, we are concerned with central force fields due to gravitational attractions
which contribute no moment about the reference body as shown schematically in Figure 4.19.

222
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

r
v2

m2
F
γ m1 m2
F=
r2
m1
r
P r

Orbital path of m2 relative to m1

Figure 4.19: Orbital motion.

Example 4.9 Astronauts Walking in Spinning Space Station

A space station in the form of a torus with interconnecting passages as shown in Figure 4.20 is initially
spinning with an angular velocity of ω0 rad
s about an axis which is perpendicular to the earth’s
surface. The effects of gravity gradient may be neglected. A pair of astronauts who are initially at
point A start to walk through the passages toward point B, always remaining symmetrically opposite
one another. Determine the angular velocity of the space station as a function of their positions along
the passages. The space station may be considered as a thin ring of mass M and radius r, the mass of
each astronaut is m and the mass of the passage is negligible.

223
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

ω0 M
m
r A
B vr D
θ
C
x

r
D vr B
A
m

Figure 4.20: Spinning toroidal space station.

Solution: The resultant external moment acting on the system about its mass center C is zero and
accordingly angular momentum is conserved. Also, since point C is the mass center of the system we
may write from Eq. 4.91 that

r r
(
hc = const. = M r 2 + 2 m r 2 ω0 k ) (a)

For the astronaut at an arbitrary position then we have

r r r
(M r 2
)
+ 2 m r 2 ω0 k = M r 2 ω f k + ρ C / D × 2 m ⋅ (vr + ω f × ρ C / D )
r r r
(b)

r r r
where vr is the velocity of the astronaut relative to the space station. From Eq. (b) using vr = −vr i
r r r r
tan (θ )i +
r
and ρ C / D = j we have
2 2

r r r r
(M r 2
)
+ 2 m r 2 ω0 k = M r 2 ω f k + 2 m r vr k + m r 2 ω f sec 2 (θ ) k (c)

Solving for ω f we find

224
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

ωf =
(M r 2
)
+ 2 m r 2 ω 0 − 2 m r vr
(d)
M r 2 + m r 2 sec 2 (θ )

It is of interest to note that ω f may increase or decrease depending on the relative values of M, m, r,

ω 0 , and vr .

Example 4.10 Projectile Striking a Pendulum

A projectile of mass mb and velocity v strikes the mass center of a ballistic pendulum, which can be
approximately modeled by massless rods and particles of mass 2 m and 5 m as shown.

a) Determine the angular velocity of the system immediately after impact as it begins to rotate
about hinge O.

b) Determine the angle θ max through which the system rotates to its maximum height, relating

the θ max to the initial speed of the projectile.

r l
v
mb θ max

3
4
2m 5m 2m

l l
2 2

Figure 4.21: Projectile impacting a ballistic pendulum.

225
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

Solution:

a) The resultant external moment about point O acting on the entire system comprising both the
ballistic pendulum model and the projectile is zero during impact. Thus, we may write that

r
H O = const. (a)

The initial angular momentum of the system is given by

(H ) = −l rj × m
r  4 r 3 r 4 r
O i b ⋅  v i − v j  = mb l v k (b)
5 5  5

and the final angular momentum immediately after the impact time is given by

(Hr ) = (− l rj )× (5 m + m ){ω k × (− l rj )}+ 2  2l ir − l rj  × 2 m ⋅ ω k ×  2l ir − l rj 


r r
O f b
    

= (− l j )× (5 m + mb )(l ω i ) + 2  i − l j  × 2 m ⋅  l ω i + ω j 
r r  l r r  r l r
2   2 

( )
r r r r
∴ HO f
= (5 m + mb )l 2 ω k + 5 m l 2 ω k = (10 m + mb )l 2 ω k (c)

Therefore, the angular velocity ω immediately after the impact is given by equating Eqs. (b) and (c),
yielding

4 mb v
ω= (d)
5 (10 m + mb )l

b) Using the work-energy methods discussed in Section 3.4, we have during the rotation of the
pendulum from θ = 0 to θ = θ max

ω k = T f − Ti = −Ti (e)

The work done due to gravity equals the product of the total weight of the system times the change in
height of its mass center during rotation which yields

226
C H A P T E R 4 : I M P U L S E - M O M E N T U M M E T H O D S F O R P A R T I C L E S

ω k = −(9 m + mb ) g l ⋅ (1 − cos(θ max )) (f)

The initial kinetic energy Ti is given by

Ti =
1
(5 m + mb )l 2 ω 2 + 2  1 (2 m ) 5 l 2 ω 2 
2 2 4 

or

Ti =
1
(10 m + mb )l 2 ω 2 (g)
2

Therefore, substituting Eqs. (f) and (g) into Eq. (e), we obtain the maximum angle θ max , given by

 1 (10 m + mb ) l ω 2 
θ max = cos 1 −
−1
 (h)
 2 (9 m + mb ) g 

or, in terms of the projectile speed v, we have

 2
8 mb v 2 
θ max = cos −1 1 −  (i)
 25 (9 m + mb )(10 m + mb ) g l 

227
5
Chapter

Kinetics of Rigid Bodies

I n Chapter 2 we considered at some length the dynamical behavior of particles and systems of
particles. Now we will direct our attention to the dynamical behavior of rigid bodies. In the prior
discussions of kinematics we noted that in general a rigid body has six degrees of freedom, three
translational and three rotational, and further that a rigid body can be represented by an infinite set of
infinitesimal particles that remain at fixed distances from one another.

We also noted in our discussions of the kinetics of systems of particles that many principles of
mechanics are conveniently represented in terms of the motion of the mass center of the system. For a
rigid body, the mass center is conveniently a fixed point on the body itself. With these remarks in
mind, we then note that in a treatment of rigid bodies, it will be desirable in many instances to use the
mass center of the rigid body as a reference point and replace the summation procedure for a finite
system of particles with an integration over an infinite system of particles comprising the rigid body.

5.1 Translational Equations of Motion


r
The mass center C of the rigid body shown in Figure 5.1 is located by the radius vector rc with respect
to the inertial X, Y, Z frame with origin O. The mass of the body is m and it is considered to have an
r
absolute angular velocity of Ω as shown. In addition, a set of body axes x, y, z are shown with their
origin at the mass center C.

228
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

y r
Y Ω
vy
r
j r
r r&c r
vz i k vx
C
z r x
rc
O
X

Figure 5.1: An xyz frame attached to a moving body.

From the previous discussions on the application of Newton’s second law to the motion of the mass
center of any general system of particles, the translational equations of motion for the rigid body may
be written in vector form as

r r
F = m &r&c (5.1)

r r
where F is the resultant external force acting on the rigid body and &r&c is the absolute acceleration of
its mass center.

Another very important form of Eq. 5.1 that is frequently used in analytical flight mechanics studies
may be readily obtained by introducing the moving xyz reference frame attached to the moving body
as shown in Figure 5.1. Considering the angular and linear velocity vectors in terms of components
along the x, y, z axes, we may write

r r r r
Ω = Ωx i + Ω y j + Ωz k
r r r r (5.2)
r&c = v x i + v y j + v z k

r
Then, by application of Eq. 1.50 for the time derivative of r&c given by

229
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

r& r
&rr& = drc dr&c r r
c = + Ω × r&c (5.3)
dt R
dt r

where the subscripts R and r refer to derivatives with respect to the inertial and relative frames
respectively. The scalar components of Eq. 5.1 become

Fx = m ⋅ (v&x + Ω y v z − Ω z v y )
Fy = m ⋅ (v& y + Ω z v x − Ω x v z ) (5.4)
Fz = m ⋅ (v&z + Ω x v y − Ω y v x )

The quantities Fx , Fy , Fz are the instantaneous scalar components of external force along the x, y, z

body axes in terms of the instantaneous components of velocities and accelerations. Eqs. 5.4 prove to
be especially useful in treating vehicle motion problems in which the velocity and acceleration
components shown on the right hand side of Eq. 5.4 may be measured by orienting sensors along the
vehicle body axes xyz. Furthermore, the flight loads are ordinarily described in terms of components
along body axes rather than inertial axes.

Furthermore, it was shown in Chapter 2 that for any system of particles, the linear momentum of the
system is given by the product of the system’s mass times the velocity of the mass center. Thus, we
may write the linear momentum of the rigid body shown in Figure 5.1 as simply

r r
P = m r&c (5.5)

5.2 Angular Momentum of a Rigid Body


We have learned that for any system of particles, we may write the moment of all external forces with
respect to an arbitrary point P as

r r& r r
M P = hP + m ρ c × &r&P (5.6)

r n
r r r
where hP = ∑ ρ i × mi ρ& i . For a rigid body, h p is given by
i =1

r r r
hP = ∫ ρ × ρ& dm (5.7)

230
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

where the element of mass dm is given by the product of the local density times the volume of the
element and the integration is performed over the entire body.

r
A rigid body, whose mass center is at C, rotates with an angular velocity of Ω relative to an inertial
frame X, Y, Z as shown in Figure 5.2. Point P is an appropriate reference point for the body and is
r
shown in Figure 5.2 as the origin of an x, y, z reference frame that has an angular velocity ω relative
r r
to the X, Y, Z frame. It is very important to note that Ω and ω , the angular velocities of the rigid
body and of the x, y, z coordinate system respectively, are not required to be equal and point P is not
necessarily on the body itself, although this will often prove convenient.
r
y ωr
Y Ω

r
j r dm
r ρ
r r
i P ρ′ k
C
z r x
rP
O
X

Z
r
Figure 5.2: A rigid body with angular momentum Ω relative to the inertial frame.

If the reference point P is selected in such a way that at any time t each point of the body undergoes a
pure rotation about P (as occurs if P is selected as a fixed point on the rigid body, for example), then
r r r
ρ& = Ω × ρ and Eq. (5.7) becomes

( )
r r r r
hP = ∫ ρ × Ω × ρ dm (5.8)

r
The element of mass dm may be located relative to point P by the radius vector ρ written in terms of
components along the x, y, and z axes as

r r r r
ρ = xi + y j + z k (5.9)

231
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

and the angular velocity vector of the rigid body referred to the same axes becomes

r r r r
Ω = Ωx i + Ω y j + Ωz k (5.10)

Substituting Eqs. (5.9) and (5.10) into Eq. (5.8) and expanding the cross product yields

r r r r
hP = hx i + hy j + hz k (5.11)

where

( )
hx = Ω x ∫ y 2 + z 2 dm − Ω y ∫ x y dm − Ω z ∫ x z dm
(
h y = −Ω x ∫ y x dm + Ω y ∫ x 2 + z 2 dm − Ω z ∫ y z dm) (5.12)
hz = −Ω x ∫ z x dm − Ω y ∫ z y dm + Ω z ∫ x 2 + y 2 dm ( )
In Eq. (5.12) we may define the following terms for convenience

(
I xx = ∫ y 2 + z 2 dm )
I yy = ∫ (x 2
+ z2 )dm (5.13)
I zz = ∫ (x2
+ y2 )dm
and

I xy = I yx = ∫ x y dm

I xz = I zx = ∫ x z dm (5.14)
I yz = I zy = ∫ z y dm

The quantities I xx , I yy , and I zz are referred to as the mass moments of inertia about the subscripted

axes and the quantities I xy , I xz , and I yz are termed the mass products of inertia with respect to the

pair of indicated axes. With these simplifications in notation, Eqs. (5.12) may be compactly written as

232
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

hx = I xx Ω x − I xy Ω y − I xz Ω z
h y = − I yx Ω x + I yy Ω y − I yz Ω z (5.15)
hz = − I zx Ω x − I zy Ω y + I zz Ω z

In matrix form, Eq. (5.15) is given by

[hP ] = [I P ][Ω] (5.16)

where [hP ] and [Ω] are column matrices, or vectors, and I P is a square 3× 3 matrix. In expanded
form, Eq. (5.16) becomes

hx   I xx − I xy − I xz  Ω x 
    
h y  = − I yx I yy − I yz  Ω y  (5.17)
h   − I zx − I zy I zz  Ω z 
 z 

Before proceeding to a discussion of the inertia matrix, we may make a final observation at this time
r
concerning hP . Note that if we select point P as identical with the mass center C, then referring to
r r
Figure 5.1, ρ ′ ≡ ρ . Thus, Eq. (5.16) becomes

[hc ] = [I c ][Ω] (5.18)

where the inertia matrix [I c ] is now computed with respect to distances measured in the x, y, z frame
with the origin at the mass center C.

5.3 Rotational Equations of Motion for Rigid Bodies


Previously it has been shown that for a rigid body, the angular momentum principle is given by

r r& r r
M P = hP + m ρ c × &r&P (5.19)

where

r r r r
hP = hx i + hy j + hz k (5.20)

and

233
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

hx = I xx Ω x − I xy Ω y − I xz Ω z
h y = − I yx Ω x + I yy Ω y − I yz Ω z (5.21)
hz = − I zx Ω x − I zy Ω y + I zz Ω z

r r r r
Recall that in Eq. 5.21 Ω = Ω x i + Ω y j + Ω z k represents the absolute angular velocity of the body

and the I ij are computed with respect to the x, y, z system with its origin at P. The components of the
r
angular momentum vector hP are shown in Figure 5.3 where the coordinate axes x, y, z rotate with the
r
angular velocity ω with respect to the inertial axes X, Y, Z for which

r r r r
ω = ωx i + ω y j + ωz k (5.22)

r
Note that, in general, the absolute angular velocities of the body, Ω , and of the rotating x, y, z
r
reference axes, ω , may be different. If body axes are used in a particular application, we simply take
r r
ω = Ω in the analysis.
r
y ωr
Y Ω

hy C
r
ρc r
hz P hP
&rr& hx
P
z r x
rP
O
X

Z
Figure 5.3: Angular momentum vector components.

r
Referring to Figure 5.3, we may write the time derivative of hP , whose components are referred to the
x, y, z axes as

r& r& r r
hP = hP + ω × hP (5.23)
R r

234
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

where the subscripts R and r refer to derivatives with respect to the inertial and relative frames
respectively. Thus, Eq. 5.19 becomes

r r r r r r r r
M P = h&x i + h&y j + h&z k + ω × hP + m ρ c × &r&P (5.24)

Equation 5.24 may be considered as the general rotational equation of motion for a rigid body. The
primary restriction imposed on Eq. 5.24 is that at time t, the motion of each point of the body relative
to the point P is purely rotational. The most notable example of this is for reference point P selected
as a fixed point on the body itself. If, of course, we want to integrate Eq. 5.24 with respect to time,
then the motion of the body relative to P must be rotational throughout the period of integration.

It is clear from consideration of Eq. 5.24 that, in general, the moments and products of inertia are time
dependent, thus complicating the solution of Eq. 5.24 considerably. However, if we are careful in
selecting the x, y, z rotating frame so that the I ij are invariant in time, the solution is greatly simplified.

Thus, since no physical problems come with coordinate axes already attached, it is only reasonable
that the analyst selects the axes to his advantage, which in this case translates to selecting the x, y, z
axes so that the I ij remain constant in time. Two very important cases of this will be discussed in
r r
detail. Namely, (1) body axes, for which ω ≡ Ω so that the x, y, z coordinates of each element of
mass of the body relative to the rotating axes remain unchanged as the body moves through space, and
r
(2) axes for which the body is allowed to have an angular velocity ω r about axes of radial symmetry
of the body relative to the x, y, z rotating frame. The latter case is particularly important in
gyrodynamics problems involving Euler angles for which the reference axes x, y, z precess and nutate,
allowing the body to spin about its axis of radial symmetry relative to these coordinate axes.

A. Case I: Body Axes

r r
By fixing the x, y, z coordinate axes to the rigid body so that ω ≡ Ω , the moments and products of
inertia remain constant in time, since the coordinates locating each element of mass relative to the x, y,
r
z axes remain unchanged. In this way, the time rates of change of magnitude of the components of hP
along the x, y, z axes, referring to Eqs. 5.21, are given by

235
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

h&x = I xx ω& x − I xy ω& y − I xz ω& z


h&y = − I yx ω& x + I yy ω& y − I yz ω& z (5.25)
h&z = − I zx ω& x − I zy ω& y + I zz ω& z

r r
Moreover, writing the vectors ρ c and &r&P in components along the body axes as

r r r r
ρ c = xc i + y c j + z c k
r r r r (5.26)
r&&P = &x&P i + &y&P j + &z&P k

we may write the last term of Eq. 5.24 as

r r r r r
ρ c × &r&P = m ⋅ ( yc &z&P − z c &y&P )i + m ⋅ ( zc &x&P − xc &z&P ) j + m ⋅ ( xc &y&P − y c &x&P ) k (5.27)

r r
Substituting Eqs. 5.25 and 5.27 into Eq. 5.24 along with ω × hP expanded according to Eqs. 5.20 and
r
5.22, we obtain the scalar components of M P along the body axes as follows:

M x = I xx ω& x + ω y ω z ⋅ (I zz − I yy ) + I xy ⋅ (ω z ω x − ω& y )
( )
− I xz ⋅ (ω x ω y + ω& z ) − I yz ⋅ ω y − ω z + m ⋅ ( yc &z&P − zc &y&P )
2 2

M y = I yy ω& y + ω z ω x ⋅ (I xx − I zz ) + I yz ⋅ (ω x ω y − ω& z )
( 2 2
)
− I yx ⋅ (ω y ω z + ω& x ) − I zx ⋅ ω z − ω x + m ⋅ ( zc &x&P − xc &z&P )
(5.28)

M z = I zz ω& z + ω x ω y ⋅ (I yy − I xx ) + I zx ⋅ (ω y ω z − ω& x )
( )
− I zx ⋅ (ω z ω y + ω& y ) − I xy ⋅ ω x − ω y + m ⋅ ( xc &y&P − yc &x&P )
2 2

It is evident that Eqs. 5.28 constitute a very formidable set of equations to solve, unless possibly the
motion of the body is specified and the driving torques are to be determined. For the inverse problem,
Eqs. 5.28 represent a complex nonlinear set of differential equations whose solution is known for very
few cases.

It is also clear from Eqs. 5.28 that another major simplification will result if, in addition to using body
axes, the x, y, z system is selected to coincide with the principal axes originating from point P. Then
all products of inertia are zero and Eq. 5.28 reduces to

236
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

M x = I xx ω& x + ω y ω z ⋅ (I zz − I yy ) + m ⋅ ( yc &z&P − zc &y&P )


M y = I yy ω& y + ω z ω x ⋅ (I xx − I zz ) + m ⋅ ( zc &x&P − xc &z&P ) (5.29)
M z = I zz ω& z + ω x ω y ⋅ (I yy − I xx ) + m ⋅ ( xc &y&P − yc &x&P )

A final simplification results in Eqs. 5.28 and 5.29 if the reference point P is selected according to any
of the following conditions:

r
1) rP = const. (a fixed point P is a special case)

r
2) ρ c = 0 (P is taken as the mass center) (5.30)

r r
3) ρ c || &r&P

for which

r r
m ρ c × &r&P = 0

Referring to Eq. 5.29, for example, we obtain the well known Euler equations of motion given by

M x = I xx ω& x + ω y ω z ⋅ (I zz − I yy )
M y = I yy ω& y + ω z ω x ⋅ (I xx − I zz ) (5.31)
M z = I zz ω& z + ω x ω y ⋅ (I yy − I xx )

Thus, the Euler equations of motion written in the form given by Eq. 5.31 are subject to the following
restrictions:

1) The x, y, z axes are body axes

2) The x, y, z axes are principle axes

3) The origin of the x, y, z axes either (a) moves with constant velocity, (b) is the mass center or
r
(c) has absolute acceleration that is parallel to ρ c .

Equations 5.31 were named in honor of Euler (1707-1783) who first introduced the concept of
moment of inertia by integrating Newton’s particle dynamics laws for the case of finite bodies.

237
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

We again note that even after considerable simplification, we must still solve a system of nonlinear
differential equations given by Eqs. 5.31 if the force system is given and it is necessary to determine
the resulting motion. It should be emphasized that by a judicious choice of reference point P, it may
r r
be considerably easier to include the correction term m ρ c × &r&P in the rotational equations of motion in
lieu of introducing unnecessary unknown forces into the problem. A good example of this is related
to selecting point P as the pivot point for a pendulum attached to a moving body so that the bearing
forces do not enter into the moment equations.

A final important specialization of Eqs. 5.28 occurs for plane motion, defined as motion for which all
lamina of the body move in parallel planes. If we conventionally designate the plane of motion as the
xy plane, we have that ω x = ω y = zc = z P = 0 and as a result, Eq. 5.28 reduces to

M x = − I xz ⋅ ω& z + I yz ω z
2

M y = − I yz ω& z − I zx ω z
2
(5.32)
M z = I zz ω& z + m ⋅ ( xc &y&P − yc &x&P )

If, in addition the xy plane is a plane of symmetry, then I xz = I yz = 0 from which we obtain

Mx = My = 0
(5.33)
M z = I zz ω& z + m ⋅ ( xc &y&P − yc &x&P )

Furthermore, if point P is selected according to the conditions stated in Eqs. 5.30, Eqs. 5.33 reduce to
the well known form

Mx = My = 0
(5.34)
M z = I zz ω& z

Thus, Eqs. 5.34 are the commonly known equations for plane motion where the moment of inertia Izz
is usually computed with respect to either the mass center of the body or a fixed point in space. If a
fixed point is used as the reference point P, we refer to the motion as pure rotation where in words
“the sum of the external moments about the axis of rotation equals the product of the moment of
inertia about the axis of rotation times the angular acceleration of the body”. If the angular velocity of
the rigid body is zero (i.e., ω z ≡ 0 ) the motion is referred to as translation for which we may state “the

238
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

sum of the external moments about the mass center of the rigid body is zero”. If the rigid body does
not have a symmetrical mass distribution with respect to the plane of motion, or if the reference point
P does not satisfy the conditions laid down in Eqs. 5.30, then the more general form of Eqs. 5.32 must
be used in the analysis.

A very important final observation can be made concerning plane motion of bodies that have
symmetry with respect to the plane of motion. Recalling that the basic premise was to select
coordinate axes so that the moments and products of inertia remained constant in time, it is readily
apparent that Izz satisfies this condition for any set of xy axes through point P shown in Figure 5.4.

y Y

θ
x

P
X
Z , z axes

Figure 5.4: Planar rotation.

This statement is true since the moment of inertia Iz depends simply on the mass distribution relative
to axis z and not on the orientation of the body for plane motion. Thus, Izz is the same angle for any
orientation of the xy body axes (i.e., for any angle θ ). Thus, rather than using body axes x, y, z for the
analysis as required for Euler’s equations, in general, we may simply use the inertial axes X, Y, Z.
Consequently, selecting P according to the conditions of Eqs. 5.30 for convenience, we may write Eq.
5.33 in the form

M X = MY = 0
(5.35)
M Z = I ZZ ω& ZZ

239
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

The use of inertial rather than body axes for these simple plane motion applications will frequently
r r
prove to be very convenient. In addition, the force equation F = m &r&c may be added to yield the
complete set of equations of motion.

4) Case II:

Consider a body of revolution with, for example, the x axis taken as the axis of radial symmetry.
Also, consider the principal body axes 1, 2, 3 with axis 1 coinciding with the x coordinate axis. The
principal body axes 1, 2, 3 are shown in a general position relative to the x, y, z coordinate axes in
r
Figure 5.5 where the angular velocity n represents the motion of the body relative to the coordinate
system. The axes x and 1 always coincide with one another and the y, z and 2, 3 planes are always the
same.

y
r
2 n
φ r
ω

r r
r j Ω
e2 r
ρc
r r r
i e1 C n x,1
rP
k
xc
r
e3
z φ

3
Figure 5.5: Rotation relative to a coordinate system.

r
The angular velocity of the body Ω can be written as

r r r r r
Ω = ω + n = ω + ni 5.36

or in component form

240
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

Ω x = ωx + n
Ωy = ωy 5.37
Ω z = ωz

Moreover, due to the radial symmetry about the x axis, all products of inertia remain constant in time
with respect to the x, y, z axes even though the body rotates relative to these axes.

r
Consequently, for this case, the relative angular momentum vector hP is

r r r r
hP = (ω x + n ) I xx i + ω y I yy j + ω z I zz k 5.38

and the vector from reference point P to the mass center C is

r r
ρ c = xc i 5.39

Consequently, for this case the moment equation given by Eq. 5.24 in component form becomes

M x = I xx ⋅ (ω& x + n& )
M y = I yy ω& y + ω z ⋅ (ω x + n ) I xx − ω x ω z I yy − m xc &z&P 5.40
M z = I zz ω& z − ω y ⋅ (ω x + n ) I xx + ω x ω y I yy + m xc &y&P

Moreover, if point P is selected so that the conditions of Eq. (l) are satisfied, the most important of
which correspond to point P at the mass center C or a fixed point, then Eqs. 5.40 reduce to

M x = I xx ⋅ (ω& x + n& )
M y = I yy ω& y + ω z ⋅ (ω x + n ) I xx − ω x ω z I yy 5.41
M z = I zz ω& z − ω y ⋅ (ω x + n ) I xx + ω x ω y I yy

Obviously, if it is desired to call the axis of radial symmetry the z axis, we may simply perform a
r r
permutation on Eqs. 5.41 replacing x by z, y by x, and z by y where n = n k to obtain

M x = I xx ω& x + ω y ⋅ (ω z + n ) I zz − ω z ω y I xx
M y = I yy ω& y − ω x ⋅ (ω z + n ) I zz + ω z ω x I xx 5.42
M z = I zz ⋅ (ω& z + n& )

241
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

The equations developed here will have wide application in studying the gyrodynamics of tops,
artificial satellites, gear systems, and gyroscopic instruments. It is frequently desirable in such
problems to allow the coordinate system to precess and nutate, with the body allowed to spin relative
to the reference axes x, y, z.

Example 5.1 Force of Rolling Disc constrained by Rod

The thin homogeneous disc of radius r and weight W shown in Figure 5.6 rotates freely about a
slender, massless rod of length l. The rod rotates in a horizontal plane about fixed point O with an
angular velocity N and the disc rolls without slip on a rough circular path producing reaction R.
Determine the necessary speed N for the value of R to equal twice the weight of the disc.

N
l
x

R
Figure 5.6: Wheel rolling on a rod that rotates in the xz plane.

Solution:

r r r&
Method 1: Form the H O vector and then differentiate it to determine M = H O . Let the coordinate
r r r r r
system x, y, z rotate with the rod. Then ωcs = N j and Ω body = N j + n i where n = − Nr l . Therefore

r r r r
HO = I x Ωx i + I y Ω y j + I z Ωz k
1W 2 r 1W 2 W 2 r
= r n i +  r + l  N j
2 g 4 g g 

r
Since the H O vector has constant magnitude,

242
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

r r& r& r r r r r
M O = H O = H O eHr O + ωcs × H O = ωcs × H O
r r
r n N k = l ⋅ (R − W ) k
1W 2
=−
2 g

Thus,

1W 2 2  r N2 
R =W + r N = W ⋅ 1 + 
2 g  2 g 

For R = 2 W ,

2g
N=
r

Method 2: Use Eqs. 5.41 where the disc spins about the x axis relative to the x, y, z frame. In Eqs.
5.41, we have

1W 2 Nl
I xx = r and n = −
2 g r
1W 2 W 2
I yy = I zz = r + l
4 g g
ω& x = ω& y = ω& z = n& = ω z = ω x = 0
ωy = N

Therefore, Eqs. 5.41 become

Mx = 0
My =0

M z = − N ⋅ (0 + n )
1W 2 1W
r = r N2 l
2 g 2 g

Since M z = l ⋅ (R − W ) , from summing external moments about O, we have

 r N2 
R = W ⋅ 1 + 
 2 g 

243
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

Thus, for R = 2 W ,

2g
N=
r

Method 3: Euler’s Eqs. for Principle Body Axes for N constant.

y 2
y
φ r
W e2
N
N Nl r
r n= 3 e3
Ox e1 r x
O z
1 r Nl
Oy φ& = n =
r
Side View R End View R

Figure 5.7: Application of axes to system for use with Euler’s Equations.

Let the origin of the 1, 2, 3 system be at O.

1 1 1
I1 = m r 2 ; I2 = m r 2 + m l 2 ; I3 = m r 2 + m l 2
2 4 4
; ω2 = N cos(φ ) ; ω3 = N sin (φ )
Nl
ω1 = −
r
N& l N2 l
= 0 ; ω& 2 = N& cos(φ ) − N 2 sin (φ ) ; ω& 3 = N& sin (φ ) + cos(φ )
l
ω&1 = −
r r r

Then

244
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

& + Ω Ω ⋅ (I − I ) = 0
M 1 = I1 Ω 1 2 3 3 2

M = I Ω + Ω Ω ⋅ (I − I )
2
&
2 2 3 1 1 2

1    Nl 1 2
=  m r 2 + m l 2  − N 2 sin (φ ) +  − ( N sin (φ )) m r − m l 
l 2

4  r   r  4 
= − m r 2 N 2 sin (φ )
1 l
2 r
= − m N 2 l r sin (φ )
1
2
& + Ω Ω ⋅ (I − I )
M 3 = I3 Ω 3 1 2 2 1

1 2  N l   Nl  1 2
2
=  m r + m l 
2
 cos(φ ) +  − ( N cos(φ )) − m r + m l 
2

4  r   r   4 

= m r 2 N 2 cos(φ )
1 l
2 r
= m N 2 l r cos(φ )
1
2

Finally,

M z = M 3 cos(φ ) − M 2 sin (φ ) = m N 2 l r cos 2 (φ ) + m N 2 l r sin 2 (φ ) = m N 2 l r


1 1 1
2 2 2

Solving for R gives

1
M z = R l −W l = mN2 lr
2

Therefore

1  1 N2 r 
R = W + m N 2 r = W ⋅ 1 + 
2  2 g 

Thus, for R = 2 W ,

2g
N=
r

245
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

5.4 Euler’s Equations of Motion Using Euler Angles


The Euler angles ψ , θ and φ , along with their associated angular velocity component ψ&

(precession), θ& (nutation), and φ& (spin) are shown in Figure 5.8. The axes 1, 2, 3 are considered to be
principal body axes and the axes x, y, z are a set of axes that precess and nutate with the body, but the
body spins relative to them about the z axis. Thus, we may write

r r r
Ω = ω + φ& k 5.43

where

r r r r
ω = θ& i + ψ& sin (θ ) j + ψ& cos(θ ) k 5.44

r
The components of angular momentum hP along the principal body axes are

r r r r
hP = I1 Ω1 e1 + I 2 Ω 2 e2 + I 3 Ω 3 e3 5.45

where

Ω1 = θ& cos(φ ) + ψ& sin (θ )sin (φ )


Ω = −θ& sin (φ ) + ψ& sin (θ ) cos(φ ) r
2 5.46
Ω3 = ψ& cos(θ ) + φ&

These components of angular velocity along with their first derivatives can be sued directly in Euler’s
equations of motion given by

& + Ω Ω ⋅ (I − I )
M 1 = I1 Ω 1 2 3 3 2

M = I Ω + Ω Ω ⋅ (I − I )
2
&
2 2 3 1 1 2 5.47
& + Ω Ω ⋅ (I − I )
M 3 = I3 Ω 3 1 2 2 1

where from Eq. 5.46

& = θ&& cos(φ ) − θ& φ& sin (φ ) + ψ&& sin (θ )sin (φ ) + ψ& θ& cos(θ )sin (φ ) + ψ& φ& sin (θ ) cos(φ )
Ω1
& = −θ&&sin (φ ) − θ& φ& cos(φ ) + ψ&& sin (θ ) cos(φ ) + ψ& θ& cos(θ ) cos(φ ) −ψ& φ& sin (θ )sin (φ )
Ω 5.48
2
& = ψ&& cos(θ ) − ψ& θ& sin (θ ) + φ&&
Ω 3

246
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

Z
ψ& (precession)
θ 2r y
j
3, z
φ
θ
(spin)φ& r
k

P 1 Y
φ
ψ r
i
X θ& (nutation)
x Line of Nodes
Figure 5.8: Euler angle transformations.

Thus, by substituting Eqs. 5.46 and 5.48 into Eqs. 5.47, we obtain Euler’s equations for principal body
axes in terms of Eulerian angles for the reference point P, a fixed point in space or the mass center of
the rigid body shown in Figure 5.8. These resulting equations are extremely complicated and
cumbersome for this general case and are consequently not frequently applied in this form. However,
if the body has an axis of radial symmetry, their form can be greatly simplified by using the analysis
presented in the Section 5.3, Case II. For this case, a body with radial symmetry was allowed to spin
about its axis of symmetry relative to the x, y, z coordinate system, as shown in Figure 5.8.

Denoting the moment of inertia about the spin axis by C and the moment of inertia about an axis
through P perpendicular to the spin axis as A, we may write

r r r r
hP = hx i + h y j + h z k 5.49

where

hx = Aθ&
hy = Aψ& sin (θ ) 5.50
(
hz = C ⋅ ψ& cos(θ ) + φ& )
The external moment about point P is given by

247
C H A P T E R 5 : K I N E T I C S O F R I G I D B O D I E S

r r& r r
M P = hP + m ρ c × &r&P

r r
where ρ c = zc k for this case. This case was worked out in general in the analysis cited above. Thus,
referring to Eq. 5.42 we have, referring to Eqs. 5.44 and 5.48

( )
M x = Aθ&& + ψ& sin (θ ) ⋅ ψ& cos(θ ) + φ& C − Aψ& 2 sin (θ ) cos(θ ) − m zc &y&P
y ( )
M = Aψ&& sin (θ ) + 2 Aθ&ψ& cos(θ ) − C θ& ⋅ ψ& cos(θ ) + φ& + m z &x& c P 5.51

( )
M z = C ⋅ ψ&& cos(θ ) + φ&& − ψ& θ& sin (θ ) = C ψ& cos(θ ) + φ&
d
dt
( )

If point P is selected as the mass center on a fixed point in space, as is usually done, the last term in
the equations for M x and M y is obviously zero.

248
A
Appendix

Moment of Inertia

ry
dm
r rz
r
O
rx y x
z
x
z mass M

Figure A.1: System definition.

By definition, the moment of inertia is

I x = ∫ rx dm = ∫ (y )
+ z 2 dm
2 2

M M

I y = ∫ ry dm = ∫ (x )
+ z 2 dm
2 2
(A.1)
M M

I z = ∫ rz dm = ∫ (x )
+ y 2 dm
2 2

M M

where

r r r r
r = xi + y j + z k (A.2)

and the radii of gyration are defined by

249
A P P E N D I X A : M O M E N T O F I N E R T I A

r r r
rx = y j + z k
r r r
ry = x i + z k (A.3)
r r r
rz = x i + y j

250
B
Appendix

Product of Inertia
By definition, the product of inertia is

I xy = ∫ x y dm
M

I yz = ∫ y z dm (B.1)
M

I zx = ∫ z x dm
M

Note that

a) A set of axes selected in such a manner that the products of inertia are zero are called
“principal axes”.

b) The corresponding moments of inertia are called the “principal moments of inertia”.

c) The principal moments of inertia are the maximum and minimum values for any possible
orientation of the reference axes.

251
C
Chapter

Parallel Axis Theorems


y′
y
dm
r P ( x, y , z )
A(a, b, c ) r′
y′ x′
z′
r x′
R rr
z′ b
O
x
a c

Figure C.1: Parallel set of axes.

The location of dm is given by

x = a + x′
y = b + y′ (C.1)
z = c + z′

Therefore,

Ix = ∫ (y )
+ z 2 dm = ∫ {(b ) ( )}
+ 2 b y ′ + y ′ 2 + c 2 + 2 c z ′ + z ′ 2 dm
2 2

M M
(C.2)
= ∫ (y′ ) (
+ z ′ 2 dm + b 2 + c 2 ) ∫ dm + 2 b ∫ y′ dm + 2 c ∫ z′ dm
2

M M M M

Thus, for A ⇒ G (center of mass), the transfer formulas are

252
A P P E N D I X C : P A R A L L E L A X I S T H E O R E M S

(
I x = I xG′ + b 2 + c 2 M )
I y = I yG′ + (a 2
+ c )M
2
(C.3)
I z = I zG′ + (a
2
+ b )M
2

Similarly, for Products of Inertia,

I xy = ∫ x y dm =
M
∫ (a + x′)(b + y ′)dm
M
(C.4)
= ∫ x′ y ′ dm + a b ∫ dm + b ∫ x′ dm + a ∫ y ′ dm
M M M M

For A ⇒ G (center of mass), the transfer formulas are

I xy = I xG′ yG′ + a b M
I yz = I yG′ zG′ + b c M (C.5)
I zx = I zG′ xG′ + c a M

253
D
Appendix

Transfer of Moments of Inertia for Inclined


Axes

y
dm
y′

r
r ρ
r j x′
j′ r
i′
r O r x
k r i
k′

z z′
Figure D.1: Inclined axes.

The differential mass dm is located by

r r r r r r r
ρ = x i + y j + z k = x′ i ′ + y ′ j ′ + z ′ k ′ (D.1)

The moment of inertia is calculated as

Ix = ∫ (y )
+ z 2 dm = ∫ (ρ )
− x 2 dm = ∫ (x′ )
+ y ′ 2 + z ′ 2 dm − ∫ x 2 dm
2 2 2
(D.2)
M M M M

where

r r r
( r r r
)
x = ρ ⋅ i = x′ i ′ + y ′ j ′ + z ′ k ′ ⋅ i = x′ l xx′ + y ′ l xy′ + z ′ l xz′ (D.3)

254
A P P E N D I X E : S U M M A R Y O F T R A N S F E R F O R M U L A S F O R P R I N C I P A L A X E S

Since l xx′ + l xy′ + l xz′ = 1


2 2 2

Ix = ∫ [(x′ )( )
+ y ′ 2 + z ′ 2 l xx′ + l xy′ + l xz′ − (x′ l xx′ + y ′ l xy′ + z ′ l xz′ ) ]
2 2 2 2 2

I x = l xx′ ∫ (y ′ )
+ z ′ 2 dm + l xy′ ∫ (x′ )
+ z ′ 2 dm + l xz′ ∫ (x′ )
+ y ′ 2 dm
2 2 2 2 2 2

M M M
(D.4)
− 2 l xx′ l xy′ ∫ x′ y ′ dm − 2 l xy′ l xz′ ∫ y ′ z ′ dm − 2 l xx′ l xz′ ∫ x′ z ′ dm
M M M

Therefore,

I x = l xx′ I x′ + l xy′ I y′ + l xz′ I z′ − 2 l xx′ l xy′ I x′y′ − 2 l xy′ l xz′ I y′z′ − 2 l xx′ l xz′ I x′z′
2 2 2
(D.5)

For principal axes, I x′y′ = I y′z′ = I x′z′ = 0 . Thus,

I x = l xx′ I x′ + l xy ′ I y ′ + l xz ′ I z ′
2 2 2

when transferring from principal axes x′ , y ′ , z ′ to any other set x, y, z.

255
E
Appendix

Summary of Transfer Formulas for Principal


Axes
Let axes x′ , y ′ , and z ′ be principal axes which are inclined to the general axes x, y, z. Therefore,

I x = l xx′ I x′ + l xy′ I y′ + l xz′ I z′


2 2 2
(E.1)

I y = l yx′ I x′ + l yy′ I y′ + l yz′ I z′


2 2 2
(E.2)

I z = l zx′ I x′ + l zy′ I y′ + l zz′ I z′


2 2 2
(E.3)

Similarly,

I xy = −l xx′ l yx′ I x′ − l xy′ l yy′ I y′ − l xz′ l yz′ I z′ (E.4)

I yz = −l yx′ l zx′ I x′ − l yy′ l zy′ I y′ − l yz′ l zz′ I z′ (E.5)

I zx = −l zx′ l xx′ I x′ − l zy′ l xy′ I y′ − l zz′ l xz′ I z′ (E.6)

256
F
Chapter

Inertia Tensor
Each of the terms in section (E) above may be written in terms of summation indices as follows:

I αβ = ∑ ∑ lαβ lβα
′ ′ I α ′β ′ (F.1)

where α ′ and β ′ are summed over x′ , y ′ , and z ′ , and α and β refer to any combination of terms
x, y, and z. ( I xx ≡ I x ).

In applying this particular equation, it is necessary to assign a positive sign to the terms in the
summation for which α ′ = β ′ or α = β (i.e., + I x′x′ , + I y′y′ , I z′z′ ) and assign a negative sign to terms

for which α ′ ≠ β ′ or α ≠ β (i.e., − I x′y′ , − I y′z′ , − I z′x′ ). For example, in expanding I xx = I x for

which α = x and β = x

I xx = ∑∑ l xβ ′ l xα ′ I α ′β ′ (F.2)
α′ β′

for α ′ and β ′ summed over x′ , y ′ , and z ′ . Therefore,

[
I xx = ∑ l xx′ l xα ′ I α ′x′ + l xy′ l xα ′ I α ′y′ + l xz′ l xα ′ I z′x′ ]
α′

= l xx′ l xx′ I x′x′ − l xx′ l xy′ I y′x′ − l xx′ l xz′ I z′x′


− l xy′ l xx′ I x′y′ + l xy′ l xy′ I y′y′ − l xy′ l xz′ I z′y′
− l xz′ l xx′ I x′z′ − l xz′ l xy′ I y′z′ + l xz′ l xz′ I z′z′

which reduces to

257
A P P E N D I X F : I N E R T I A T E N S O R

I xx = l xx′ I x′ + l xy′ I y′ + l xz′ I z′ − 2 l xx′ l xy′ I x′y′ − 2 l xx′ l xz′ I x′z′ − 2 l xy′ l xz′ I y′z′
2 2 2
(F.3)

This agrees with the general equation for I x , Eq. D.5. Thus, the inertia tensor is an ordered array of
inertia components written as

+ Ix − I xy − I xz 
 
I = − I yx + Iy − I yz  (F.4)
 − I zx − I zy + I z 

where I αβ = ∑ ∑ lαβ lβα ′ ′ I α ′β ′ , as described above.

258
G
Chapter

Summary of General Inertia Terms for


Inclined Axis
I x = l xx ′ I x ′ + l xy ′ I y ′ + l xz ′ I z ′ − 2 l xx ′ l xy ′ I x ′y ′ − 2 l xx ′ l xz ′ I x ′z ′ − 2 l xy ′ l xz ′ I y ′z ′
2 2 2

I y = l yx ′ I x ′ + l yy ′ I y ′ + l yz ′ I z ′ − 2 l yx ′ l yy ′ I x ′y ′ − 2 l yx ′ l yz ′ I x ′z ′ − 2 l yy ′ l yz ′ I y ′z ′
2 2 2

I z = l zx ′ I x ′ + l zy ′ I y ′ + lzz ′ I z ′ − 2 l zx ′ lzy ′ I x ′y ′ − 2 l zx ′ l zz ′ I x ′z ′ − 2 lzy ′ l zz ′ I y ′z ′


2 2 2

I xy = −l xx′ l yx′ I x′ − l xy′ l yy′ I y′ − l xz′ l yz′ I z′ + (l xx′ l yy′ + l xy′ l yx′ ) I x′y′ + (l xy′ l yz′ + l xz′ l yy′ ) I y′z′ + (l xz′ l yx′ + l xx′ l yz′ ) I z′x′

I yz = −l yx′ l zx′ I x′ − l yy′ l zy′ I y′ − l yz′ l zz′ I z′ + (l yx′ l zy′ + l yy′ l zx′ ) I x′y′ + (l yy′ l zz′ + l yz′ l zy′ ) I y′z′ + (l yz′ l zx′ + l yx′ l zz′ ) I z′x′

I zx = −l zx′ l xx′ I x′ − l zy′ l xy′ I y′ − l zz′ l xz′ I z′ + (l zx′ l xy′ + l zy′ l xx′ ) I x′y′ + (l zy′ l xz′ + l zz′ l xy′ ) I y′z′ + (l zz′ l xx′ + l zx′ l xz′ ) I z′x′

Thus, in order to determine the orientations of the principal axes, it is necessary to set
I xy = I yz = I zx = 0 from above and in addition the orthogonal direction cosines must also satisfy the

following six trigonometric identities:

l xx ′ + l xy ′ + l xz ′ = 1 ; l xx′ l yx′ + l xy′ l yy′ + l xz′ l yz ′ = 0


2 2 2

l yx′ + l yy′ + l yz′ = 1 ; l yx′ l zx′ + l yy′ l zy′ + l yz′ l zz ′ = 0


2 2 2

l zx′ + l zy′ + l zz ′ = 1 ; l zx′ l xx′ + l zy′ l xy′ + l zz′ l xz ′ = 0


2 2 2

259
The simultaneous solution of these nine equations will yield the nine direction cosines
l i , j ; i, j = x, y, z for the principle directions. Then these values can be substituted into the first three

equations on the previous page to determine the principal moments of inertia I x , I y , and I z

260

You might also like