You are on page 1of 40

Chapter 1

Introduction
This chapter is devoted to a brief discussion of the continuum postulate, the types of flow examined in
this text, the fundamental laws governing the motion of fluids, units, physical properties of fluids, and
vectors. The ideas contained in this chapter will appear repeatedly throughout the remainder of the text,
thus one should pay careful attention to the material in this introduction.

1.1 The Continuum Postulate


The objective of this text is to formulate the equations governing the motion of a continuum and to
apply them to the problem of fluid motion. In treating a fluid as a continuum, we assume that functions
such as the velocity, pressure, density, etc., are continuous functions of space and time. From the
molecular perspective, this is not true since fluids are made up of molecules and the density at some point
in space depends on whether that point is occupied by a molecule or not. However, from a less detailed
perspective, it is reasonable to ignore the molecular structure of materials and consider them in terms of
their properties averaged over some region. One can image measuring the density of a fluid using a device
having a volume V containing a mass m. If this volume is so small that it contains only a few molecules at
any given time, the mass m will change dramatically as a function of both the time at which we measure the
mass and the position at which we locate our measuring device. We define the density measured by this
device as
m
= (1-1)
V

and in Figure 1-1 we show a plot of the density as a function of the size of the measuring device, V. The
volume of this device is
V = A3 (1-2)

as indicated in Figure 1-1. There we have also shown a volume, V * , which is large enough so that it
contains thousands of molecules at any instant of time, and the measured density , , will be independent of
the volume of the measuring device when V V * . In order for the volume of the measuring device to
contain thousands of molecules, the length A must be large compared to the average distance between
molecules, and in order for the density at a point to be treated as a continuous function, the length A must
be small compared to the characteristic length of the system under consideration. The average distance
between molecules is related to their mean free path, A mf , and in order that V be large enough to contain
thousands of molecules we require A >> A mf . In order that the size of the measuring device be small
compared to the characteristic length of the system we require A << L where L represents the characteristic
length of the system. This means that the size of our measuring device is constrained by

A mf << A << L (1-3)

This result suggests that the characteristic length for a system must be much, much larger than the mean
free path in order to the continuum postulate to be acceptable. For liquids, the mean free path is on the
order of Angstroms and there is little difficulty in satisfying the constraint given by Eq. 1-3; however, for
gases at atmospheric pressure, the mean free path is on the order to 106 cm and one must pay attention to
the length-scale constraints given by Eq. 1-3.

1
2 Chapter 1

Figure 1-1. Density as a function of volume

At this point it is not obvious what is meant by the characteristic length of a system; however, this
concept will become clear in subsequent chapters. In many cases, a system cannot be characterize in terms
of a single length. A ship would be a classic example in which both the width and the length represent
important characteristic lengths. For the present, one can think of L in Eq. 1-3 as the smallest characteristic
length for any particular system.
The constraint indicated by Eq. 1-3 is overly severe, and this is caused by the fact that representation
given in Figure 1-1 is for the density at some instant of time. One can imagine measuring the density, or
any other function, over a period of time, t , so that the measured value is defined by

t =t

11
= m dt (1-4)
V t
t =0

If one considers the process of measuring the density over a region of space and over a period of time, it
becomes apparent that curve shown in Figure 1-1 will approach the asymptotic value of at smaller values
of V * as t increases.

1.2 Types of Flow


We shall examine several types of flow in this text, and the boundaries that divide them into various
classes are not always clear. However, some distinct differences exist and it will be helpful to discuss
them.

1.2.1 Compressible and incompressible flow


Sometimes fluids are referred to as incompressible with the idea that the density can be treated as a
constant. For many practical problems, one can treat the density of water as a constant in order to obtain
excellent results for the pressure drop and flow rate. However, if one is interested in understanding the
propagation of sound waves caused by snapping shrimp1, one cannot treat the density of ocean water as a
constant. The validity of the assumption of constant density depends on the particular flow under
consideration, thus we will refer to flows (not fluids) for which the density can be approximated as a
constant as incompressible flows. When the assumption of constant density is not valid, we will refer to the
flow as compressible. Very often, the assumption of a constant density fails when the fluid velocity

1
Versluis, M., Schmitz, B., von der Heydt, A. and Lohse, D. 2000, How Snapping Shrimp Snap: Through Cavitating
Bubbles, Science 289, 2114-2117.
Introduction 3

approaches the speed of sound, or the so-called sonic velocity. For air at atmospheric pressure, this is about
335 meters per second while the value for water is considerably higher at about 1430 meters per second.
This indicates that the assumption of incompressible flow is generally more acceptable for liquids than for
gases, and this is indeed the case. Compressible flow is treated in Chapter 14.

1.2.2 Laminar and Turbulent Flow


Laminar flows are characterized by their smooth and regular motion that is caused by the dominance
of viscous forces, while turbulent flows can be recognized by irregular and random motion that is
superimposed on the mean motion of the fluid. However, what is regular or irregular often depends on the
point of view of the observer. These two types of flows can be observed everywhere in the natural
environment, and it is the responsibility of the student to make these observations. An ocean wave crashing
on a coastal rock seems to represent a certain degree of irregularity, while the flow of motor oil out of
bottle and into an engine exhibits a rather regular behavior. The former is most certainly a turbulent flow
while the latter is undeniably laminar.
The transition from laminar to turbulent flow in tubes was first investigated by Osborne Reynolds2,
and a sketch of the apparatus used by Reynolds is illustrated in Figure 1-2. The system consisted of a bell-
mouthed glass tube into which a dye streak was injected with water that entered the tube from a reservoir.

Figure 1-2. Reynolds experimental investigation of the transition from laminar to turbulent flow

Reynolds observed two distinct types of flow. In the first, the dye streak maintained its identity and
remained in the center of the tube, although it spread slowly because of molecular diffusion. In the second
type of flow, the dye streak was soon dispersed throughout the tube when the laminar flow that existed at

2
Reynolds, O. 1883, An Experimental Investigation of the Circumstances which Determine whether the Motion of
Water Shall be Direct or Sinuous and the Law of Resistance in Parallel Channels, Phil. Trans. Roy. Soc. (London)
174A, 935-982.
4 Chapter 1

the entrance of the tube underwent the transition to turbulent flow. Reynolds found that the transition
conditions could be correlated by a dimensionless group which is now known as the Reynolds number
defined as follows:
vz D
Re = (1-5)

The quantities making up the Reynolds number are given by

= density , v z = average velocity in the z -direction

D = tube diameter , = viscosity

Reynolds found that the transition from laminar to turbulent flow begins for values of Re of about 2100
regardless of the specific values of , v z , D and . The complete transition takes place for a Reynolds
number of about 4000, thus the change from laminar to turbulent flow does not occur abruptly when the
Reynolds number is increased to 2100. It is important to understand that these values of Re are for the
transition from laminar to turbulent flow in a straight tube of circular cross section. For other systems,
such as a rectangular conduit, the transition from laminar to turbulent flow will occur for a different range
of Reynolds numbers. Laminar flow is studied in both Chapters 3 and 7, while the more complex turbulent
flow is studied in Chapter 9.

1.2.3 Steady and Unsteady Flow


These two designations are fairly obvious, and we only need to clarify their meaning in the case of
turbulent flow. If a laminar flow is steady, the three components of the velocity, v x , v y and v z , and the
pressure, p, are independent of time. Turbulent flows are essentially unsteady; however, we will refer to a
turbulent flow as steady in the time-averaged sense if the time-averaged components of the velocity and
pressure, v x , v y , v z , and p are independent of time. Our study of the time-averaged equations of
motion appears in Chapter 9.

1.2.4 Uniform Flow


There are a variety of flows for which the velocity profile is independent of the direction of flow. In
Figure 1-3 we have illustrated the laminar flow in the entrance region of a channel. At the entrance to the
channel, the velocity profile is nearly flat except near the solid boundary where it decreases rapidly to zero.
At the walls of the channel, the molecules that make up the fluid undergo collisions both with the wall and
with other molecules. Because of the molecule-wall collisions, the mean motion in the x-direction is
usually small enough so that it can be set equal to zero, and this is the origin of the no-slip condition
illustrated in Figure 1-3. The actual slip that occurs at a solid surface was first determined by Maxwell3,

Figure 1-3. Flow in the entrance region of a channel

3
Maxwell, J.C. 1867, On the dynamical theory of gases, Phil. Trans. Roy. Soc. 157, 49-95.
Introduction 5

and his work indicates that the no-slip condition is an acceptable approximation when the mean free path,
A mf , is small compared to the channel depth. As the fluid moves in the channel, the viscous drag caused
by the walls of the channel retards the fluid near the walls and eventually a uniform velocity profile
develops. The distance, Le , required for the uniform flow to develop is called the entrance length and the
region in which the flow is non-uniform is called the entrance region.

1.2.5 One-Dimensional Flow


By one-dimensional, we mean that the velocity, v, is a function of only one spatial coordinate. One-
dimensional turbulent flow means that the time averaged velocity, v , is a function of only one spatial
coordinate. The laminar flow illustrated in Figure 1-2a is one-dimensional downstream from the entrance
region where v is only a function of r, but in the entrance region the velocity depends on both r and z and
the flow is two-dimensional.

1.3 Newtons Laws and Eulers Laws


In order to develop the governing equations of fluid motion, we need a clear understanding of the laws
of physics that govern that motion. In a previous course in physics, Newtons laws were used to determine
the motion of mass points. These laws can be stated as4
Newton (1642-1727)
I. Every body continues in its state of rest, or of uniform motion straight ahead, unless it
be compelled to change that state by forces impressed upon it.
II. The change of motion is proportional to the motive force impressed, and it takes place
along the right line in which the force is impressed.
III. To an action there is always a contrary and equal reaction; or, the mutual actions of
two bodies upon each other are always directed to contrary parts.
The first two laws can be expressed in mathematical form according to
d
(m v) = f (1-6)
dt
in which f represents the force acting a body having a mass m, and v represents the velocity of the body.
Often a precise definition of v is not given in a discussion of Newtons first and second laws, and we will
return to this matter in the following paragraphs.
Newtons third law for two interacting bodies can be expressed as

f12 = f 21 (1-7)

in which f12 is the force that body #2 exerts on body #1, and f21 is the force that body #1 exerts on
body #2. The most dramatic success of these laws was their use, along with the law of gravitational
attraction, to justify Keplers three empirical laws of planetary motion; however, Newtons three laws are
difficult to apply to the bending of a beam or the deformation of a fluid. In order to analyze the motion and
deformation of fluids, we will make use of the laws of continuum mechanics, attributed to Euler4, which
can be stated as:
Euler (1707-1783)
I. The time rate of change of the momentum of a body equals the force acting on the
body.
II. The time rate of change of the angular momentum of a body equals the torque acting
on the body, where both the torque and the moment are taken with respect to the same
fixed point.

4
Truesdell, C. 1968, Essays in the History of Mechanics, Springer-Verlag, New York
6 Chapter 1

In addition to these two laws, we will accept the Euler cut principle stated as:
Not only do the laws of continuum physics apply to distinct bodies but they also apply to
any arbitrary body that one might imagine as being cut out of a distinct body.
In order to understand how Eulers laws are related to Newtons laws, we need to put Eulers laws in
precise mathematical form. This will allow us to demonstrate that they contain Newtons laws provided
that we restrict ourselves to non-relativistic phenomena. Since Eulers two laws and the Euler cut principle
form the basis for virtually all of engineering mechanics, it is important that we understand these concepts.

1.3.1 Eulers laws


We begin our study of Eulers laws by expressing them in terms of the following two equations

time rate of change


force acting on
of the linear momentum = (1-8)
the body
of a body

time rate of change


torgue acting on
of the angular momentum = (1-9)
of a body the body

to which we add the equation indicating that the mass of a body is a constant.

time rate of change


= 0 (1-10)
of the mass of a body
In Figure 1-4 we have illustrated an arbitrary body that one can image as being cut out of a distinct body.
The volume of this moving, deforming body is designated as Vm (t ) , and the differential volume and surface
elements are identified as dV and dA respectively. The vector force per unit surface area is designated by
the stress vector, t (n ) , which is studied in detail in Chapters 2 and 5. The stress vector represents a contact

Figure 1-4. Moving, deforming body

force that the surroundings exert on the body. The vector force per unit mass is designated by b and it
represents a force acting at a distance that is exerted on the body. For many processes, the force per unit
Introduction 7

mass is equal to the gravitational acceleration, i.e., b = g ; however, this simplification is only valid when
the electrodynamic and electromagnetic forces are negligible.
We begin our analysis of Eq. 1-8 by constructing a mathematical representation of the momentum of
the body. The mass, dm , contained in the differential volume element shown in Figure 1-5 is given by

dm = dV (1-11)

and the momentum (mass times velocity) per unit volume takes the form

v dm = v dV (1-12)

The total momentum of the body is the volume integral of this quantity and we express this as

momentum
=
of the body
Vm (t )
v dV (1-13)

Use of this result in Eq. 1-8 leads to

force acting on

d
v dV = (1-14)
dt the body
Vm (t )

Here we have made use of the classic definition of a time derivative given by

df f (t + t ) f (t )
= lim (1-15)
dt t 0 t

and in terms of the time rate of change of the momentum of a body we have

( v ) dV

( v )t +t dV t
V (t +t )

d Vm (t )
vdV = lim m (1-16)
dt t 0
t
Vm (t )

To complete our mathematical representation of Eq. 1-14, we need to express the force acting on the body
in terms of the body force and the surface force. The first of these is given by

body force acting



on the body
=

Vm ( t )
b dV (1-17)

while the second takes the form

surface force acting



on the body
=

Am (t )
t (n ) dA (1-18)

Use of Eqs. 1-14, 1-17 and 1-18 in the word equation given by Eq.1-8 leads to a precise mathematical
statement of Eulers first law.


d
Euler I: v dV = b dV + t (n ) dA (1-19)
dt
Vm (t ) Vm (t ) Am (t )
8 Chapter 1

It will be left as an exercise for the student to show that the word equation given by Eq. 1-9 can be
expressed as

r v dV r b dV
d
Euler II: = + r t (n ) dA (1-20)
dt
Vm (t ) Vm (t ) Am (t )

To be precise about Eulers two laws, we need to say that the velocity, v, is determined relative to an
inertial frame and that the position vector r is determined relative to some fixed point in an inertial frame.
Inertial frames are discussed in Sec. 1.3.3. In addition to the precise statement of Eulers two laws of
mechanics, we need a similar statement for the principle of conservation of mass. Use of Eq. 1-11 with
Eq. 1-10 leads to the following expression:


d
Mass: dV = 0 (1-21)
dt
Vm (t )

It is important to remember that these three axiomatic statements for linear momentum, angular momentum
and mass apply to any arbitrary body that one imagines as being cut out of a distinct body.
EXAMPLE 1.1. A precise statement of the principle of conservation of mass requires the use of a
volume integral of the density (mass per unit volume). In some cases, the mass of a body can be
represented in terms of a line integral of a density having units of mass per unit length. If
represents the mass per unit length of the wire illustrated in Figure 1.1, the total mass is given by
x=L

mass =
dx
x =0

For the case in which the mass density of the wire is given by

{
= o 1 + }
( x L ) 1 ( x L )2

we can calculate the total mass in terms of the integral given by


x=L

{1 + ( x L ) 1 ( x L ) } dx (5 4) o L
o 2
mass = =
x =0

Figure 1.1. Wire having a non-uniform mass density

If the mass per unit length were uniform, the total mass would be o L .

1.3.2 Eulers laws and Newtons laws


Given Eulers two laws of mechanics and the Euler cut principle, we need to know how they are
related to Newtons three laws. To explore this problem, we consider a body of mass m illustrated in
Figure 1-5, and we locate the center of mass of that body in terms of the position vector defined by
Introduction 9


1
rCM = rdV (1-22)
m
Vm (t )

For a sphere of uniform density, the center of mass would be located at the geometrical center of the
sphere; however, the definition of rCM is completely general and Eq. 1-22 is applicable to any arbitrary
body that is cut out of a distinct body. In our study of the kinematics of deformable media in Chapter 4, we
will show that the velocity of the center of mass is given by (see Problem 4-9)


d rCM 1
v CM = = v dV (1-23)
dt m
Vm (t )

and this indicates that v should be replaced by v CM in Figure 1-5.

Figure 1-5. Motion of a body

As a matter of convenience, we designate the total force acting on the body by

f =

Vm (t )
b dV +

Am (t )
t (n ) dA (1-24)

so that Eq. 1-19 can be represented in the simplified form given by

d
Euler I: ( m v CM ) = f (1-25)
dt

Here we see that Eulers first law for an arbitrary, deforming body contains Newtons first and second laws
provided that the velocity, v, in Eq. 1-6 is interpreted as the velocity of the center of mass, v CM . Feynman
et al.5 describe this situation by saying Newtons law has the peculiar property that if it is right on a
certain scale, then it will be right on a larger scale. Euler most certainly would have agreed.

EXAMPLE 1.2. In the Spring, students are often concerned with the trajectory of water balloons
such the one shown in Figure 1.2. Determination of the trajectory is certainly not a problem in

5
Feynman, R.P., Leighton, R.B. and Sands, M. 1963, The Feynman Lectures on Physics, Sec. 19-1, Addison-Wesley
Pub. Co., New York.
10 Chapter 1

Figure 1.2. Water balloon trajectory

mass point mechanics; however, Eq. 1-25 (to be proved in Chapter 4) represents a useful tool
provided that knowledge of the position of the center of mass is sufficient. If the drag force (see
Chapter 13) of the air on the water balloon can be neglected, the velocity is determined by

dv CM
= g (1a)
dt

IC.1 v CM = uo ( i cos + k sin ) , t=0 (1b)

Here uo represents the launch speed and the angle of inclination of the launch. Integration of
Eq. 1a subject to the initial condition leads to

v CM = g t + uo ( i cos + k sin ) (2)

To locate the position as a function of time, we use Eq. 1-23 to obtain the following initial value
problem for the position of the center of mass:

drCM
= g t + uo ( i cos + k sin ) (3a)
dt

IC.2 rCM = 0, t=0 (3b)

Integration leads to the solution for the position vector given by

rCM = 1 g t 2 + u ( i cos + k sin ) t (4)


2 o

The horizontal and vertical positions are given explicitly by

i rCM = uo cos t , k rCM = 12 g t 2 + uo sin t (5)

From the second of these, we conclude that the arrival time, t1 , of our missile is

t1 = 2uo sin / g (6)

and we can solve for the distance to the target to obtain


2 2
distance = g uo cos sin (7)

This provides the classic result that the maximum target distance is achieved with a launch angle
of 45o ; however, one must remember that this calculation neglects the drag force of the air on the
water balloon.
Introduction 11

While the use of Eq. 1-23 allows us to develop an attractive correspondence between Eqs. 1-6 and
1-19, the situation regarding Newtons third law and Eulers second law is more complex. In order to
extract a simple representation for Eulers second law, we assume that the body is small enough so that r in
Eq. 1-20 can be approximated by rCM . With this simplification, often referred to as the mass-point
approximation, Eulers second law takes the form

d
Euler II: rCM ( m v CM ) = rCM f (1-26)
dt

It will be left as an exercise for the student to show that this form of Eulers second law reduces to Eq. 1-25
for the process illustrated in Figure 1-5. This means that Eulers second law is redundant for this particular
process and the mass-point approximation.
In order to explore the relation of Newtons third law to Eulers first and second laws, we consider the
two-body process illustrated in Figure 1-6 where r1 and r2 represent center-of-mass position vectors. For
the Eulerian Cuts I and II, we use Eq. 1-25 to obtain

d
dt
(
m1 v1 ) = f12 , Cut I (1-27)

d
( m2 v 2 ) = f21 , Cut II (1-28)
dt

where f12 represents the force that body #2 exerts on body #1, and f21 represent the force that body #1
exerts on body #2. Here it is understood that v1 and v 2 represent center-of-mass velocities. Directing our

Figure 1-6. Two-body process

attention to Cut III, we consider the special case in which the two bodies are isolated, thus there are no
external forces acting on the pair. Under these circumstances, we can apply Eq. 1-19 to obtain


d
v dV = 0, Cut III (1-29)
dt
Vm (t )

and this can be expressed as


12 Chapter 1

v dV v dV
d d
+ = 0, Cut III (1-30)
dt dt
VI (t ) VII (t )

Applying the type of analysis indicated by Eqs. 1-22 and 1-23 leads to

d d
( m1v1 ) + ( m2 v 2 ) = 0 , Cut III (1-31)
dt dt

and use of Eqs. 1-27 and 1-28 provides the action-reaction statement of Newtons third law.

Euler I: f12 = f 21 , Cut III (1-32)

Here one should remember that we have not yet made use of Eulers second law, thus we have shown that
Eulers first law contains all three of Newtons laws.
To understand the content of Eulers second law, we apply Eq. 1-20 to Cut III. Provided that r can be
approximated by rCM , this leads to the result

( r2 r1 ) f12 = 0 (1-33)

which is left as an exercise for the student to prove. There are three ways in which this result can be
satisfied, and we list them as
1. ( r2 r1 ) = 0

2. f12 = 0

3. ( r2 r1 ) and f12 are parallel

Since the first two possibilities can not be generally true, we conclude that the interaction force between
two bodies must be parallel to the vector r2 r1 . We express this result as

f12 = 12 ( r2 r1 ) (1-34)

in which 12 is some scalar parameter of the interaction force law. Equation 1-34 indicates that the
interaction force between two bodies must act along the line of centers, i.e., it is a central force.
In this analysis we have shown that Eulers first law contains Newtons three laws, while Eulers
second law provides what is known as the central force law for the case of mass-point mechanics. Given
the power and economy of Eulers laws, one can wonder why Newtons three laws are not discarded in
favor of Eulers two laws. The answers lies in the fact that the central force law, represented by Eq. 1-34,
is a non-relativistic phenomenon. Since forces are propagated at the speed of light, the force that one body
exerts on another cannot lie along the line of centers when the relative velocity between the two bodies
approaches the speed of light. Because of this, physicists prefer to view mechanical phenomena in terms of
Newtons laws and make use of the central force law as a special case which can be discarded when
relativistic phenomena are encountered. Engineers, on the other hand, are rarely involved in relativistic
phenomena and what is a special case for the physicist is the general case for the engineer. Because of
this, engineers uniformly formulate their mechanical problems in terms of Eulers two laws and the Euler
cut principle.

1.3.3 Inertial frames of reference


In our statement of both Newtons laws and Eulers laws, we said nothing about the frame of reference
used to determine the velocity, v. To be correct, we must say that the laws of mechanics (as we have stated
them) are only valid in inertial frames, and this naturally leads to the question, What are inertial frames?
Unfortunately, the answer to this question is that inertial frames are frames in which the laws of mechanics
are valid. This means that inertial frames must be determined by experiment, and the first great experiment
Introduction 13

was the comparison of Newtons laws with Keplers experimental observations of planetary motion. In
that case, Newton found that the stars provided a frame of reference in which his laws, along with the law
of gravitational attraction, were valid. For the types of fluid motion studied in this text, points fixed
relative to the earth will usually provide an acceptable inertial frame6; however, this is not the case for
large-scale atmospheric and oceanographic flows. In those cases, one must make use of the stars to
construct an acceptable inertial frame. Non-inertial frames are discussed in Chapter 15.
Inertial frames play a key role in the theory of special relativity, and while we are primarily concerned
with non-relativistic phenomena, it is useful to be aware of the axioms of special relativity. These can be
stated in the form:
I. Time is homogeneous. This means that there is no preferred origin in time.
II. Space is homogeneous and isotropic. This means that there is no preferred
origin or direction in space.
III. There is no preferred inertial frame. This means that the laws of physics
take the same form in all inertial frames.
IV. The speed of light is the same in all inertial frames.
The first three axioms are employed routinely as a matter of course, while the fourth represents an
astounding affront to our intuition that is difficult to accommodate.

1.4 Units
Under the best of circumstances, engineering is a predictive science; however, engineers are
confronted with complex processes and their predictions must always be compared with quantities that
have been measured experimentally. There is basically only one means of measuring any quantity and that
requires a comparison with a standard. For example, the distance between two points is determined by
counting the number of times that a standard length fits between the two points. Often we call this length a
unit length. The process of evaluating by counting standard units is described by Hurley and Garrod7 who
state:
Since the measurement process is one of counting multiples of some chosen standard, it is
reasonable to ask how many standards we need. If we need a standard for each observable, we
will need a large Bureau of Standards. As a matter of fact, we need only four standards: a
standard of length, a standard of mass, a standard of time, and a standard of electric charge. This
is an extraordinary fact. It means that if one is equipped with a set of these four standards and the
ability to count, one can (in principle) assign a numerical value to any observable, be it distance,
velocity, viscosity, density, pressure, etc.
Here we find that our study of units begins with a great deal of simplicity since we require only four
standards:
MASS
LENGTH
TIME
ELECTRIC CHARGE

Although the concept of a standard is simple, the matter is complicated by the fact that the choice of
standard is arbitrary8.

6
The fact that points fixed relative to the earth do not constitute an inertial frame can be verified by the precession of a
Foucault pendulum (named after the French physicist, Jean Leon Foucault, 1819-1868).
7
Hurley, J.P. and Garrod, C., 1978 Principles of Physics Houghton Mifflin Co., Boston
8
In the Sacramento Bee, November 11, 1999 one finds the headline, Training faulted in loss of $125 million Mars
probe, and in the article that follows one reads, The immediate cause of the spacecrafts Sept. 23 disappearance as it
entered Mars orbit was a failure by a young engineer...to make a simple conversion from English units to metric..
14 Chapter 1

1.4.1 International System of Units


In 1960 a conference was held in Paris to find agreement on a set of standards. From that conference
there arose what are called the SI (Systme International) system of basic units9 which are listed in
Table 1-1. Note that the SI system does not use the electric charge as a standard, but rather the electric
charge per unit time or the electric current. In addition, the SI system of basic units includes three

Table 1-1. SI Basic Units


Quantity Name Symbol Definition
length meter m The meter is the length of the path traveled by
light in vacuum during a time interval of 1/299
792 458 of a second.
mass kilogram kg The kilogram is the unit of mass equal to the
international prototype of the kilogram.
time second s The second is the duration of 9 192 631 770
periods of the radiation corresponding to the
transition between the two hyperfine levels of
the ground state of the cesium 133 atom.
electric current ampere A The ampere is that constant current which, if
maintained in two straight parallel conductors of
infinite length, of negligible circular cross
section, and placed 1 meter apart in vacuum,
would produce between these conductors a force
equal to 2 107 newton per meter of length.
temperature kelvin K The Kelvin, unit of thermodynamic temperature,
is the fraction of 1/273.16 of the thermodynamic
temperature of the triple point of water.
elemental entities mole mol The mole is the amount of substance of a system
which contains as many elementary entities as
there are atoms in 0.012 kilogram of carbon 12.
luminous intensity candela cd The candela is the luminous intensity, in a given
direction, of a source that emits monochromatic
radiation of frequency 540 1012 hertz and that
has a radiant intensity in that direction of 1/683
watt per steradian.

additional units, the thermodynamic temperature, the mole, and the luminous intensity. These three
additional units are not necessary to assign numerical values to any observable, thus their role is somewhat
different than the four fundamental standards identified by Hurley and Garrod. For example, a mole
consists of 6.02209... 1023 entities, such as atoms, molecules, photons, etc., and the basic unit associated
with counting entities is one. Feynman et al.10 emphasize this point with the statement, We use 1 as a unit,
and the chemists use 6 1023 as a unit!. Nevertheless, a mole is a convenient unit for traditional chemical
engineering calculations because one mole is significant while one molecule is not. Energy can be
described in units of kg m2/s2; however, the thermodynamic temperature represents an extremely
convenient unit for the description energy and many engineering calculations would be quite cumbersome
without it. The same comment applies to the luminous intensity which is an observable that can be
assigned a numerical value in terms of the four fundamental standards of length, mass, time and electric
charge. One of the attractive features of the SI system is that alternate units are created as multiples and
submultiples of powers of 10, and these are indicated by prefixes such as giga for 109 , centi for 102 , nano

9
http://physics.nist.gov/cuu/index.html
10
Feynman, R.P., Leighton, R.B. and Sands, M. 1963, The Feynman Lectures on Physics, Sec. 39-5, Addison-Wesley
Pub. Co., New York.
Introduction 15

for 109 , etc. Some of these alternate units are listed in Table 1-2 for the meter. In other systems of units,
multiples of 10 are not necessarily used in the creation of alternate units, and this leads to complications
which in turn lead to errors.

Table 1-2. Alternate Units of Length

1 kilometer (km) = 103meter (m)


1 decimeter (dm) = 10-1m
1 centimeter (cm) = 10-2m
1 millimeter (mm) = 10-3m
1 micrometer (m) = 10-6m
1 nanometer (nm) = 10-9m

1.4.2 Derived units


In addition to using a variety of alternate units for length, time, etc., we make use of many derived
units in the SI system and a few are listed in Table 1-3. Some derived units are sufficiently notorious so
that they are named after famous scientists and represented by specific symbols. For example, the unit of

Table 1-3. Derived S.I. Units


Physical Quantity Unit (Symbol) Definition
force newton (N) kg m/s 2
energy joule (J) kg m 2 /s 2
power watt (W) J/s
electrical potential volt (V) J/(A s)
electric resistance ohm ( ) V/A
frequency hertz (Hz) cycle/second
pressure pascal (Pa) N/m 2
kinematic viscosity stoke (St) cm 2 / s
thermal diffusivity square meter/second m2 / s
molecular diffusivity square meter/second m2 / s

kinematic viscosity is the stoke (St), after the British mathematician, Sir George G. Stokes11, while the
equally important molecular and thermal diffusivities are known only by their generic names and
represented by a variety of symbols. The key point to remember concerning units is that the basic units
represented in Table 1-1 are sufficient to describe all physical phenomena, while the derived standards
illustrated Table 1-2 and the derived units listed in Table 1-3 are used as a matter of convenience.
While the existence of derived units is simply a matter of convenience, this convenience sometimes
leads to confusion. This is particularly true in the case of Newtons second law or Eulers first law which
we repeat here as
d
(m v) = f (1-35)
dt

Let us now think about the use of this result to calculate the force required to accelerate a mass of 12 kg at a
rate of 7 m/s2. From Eq. 1-35 we determine the magnitude of this force to be

f = (12 kg)(7m/s 2 ) = 84 kg m/s 2 (1-36)

11
Rouse, H. and Ince, S. 1957, History of Hydraulics, Dover Publications, Inc., New York
16 Chapter 1

where f is used to represent the magnitude of the vector f. Note that the force is expressed in terms of three
of the four fundamental standards of measure, i.e., mass, length and time. There is no real need to go
beyond Eq. 1-36 in our description of force; however, our intuitive knowledge of force is rather different
from our knowledge of mass, length and time. Consider for example, pushing against a wall with a force
of 84 kg m/s2. This is simply not a satisfactory description of the event. What we want here is a unit that
describes the physical nature of the event, and we obtain this unit by defining a unit of force as

1 newton = (1 kg ) (1 m/s2 ) (1-37)

When pushing against the wall with a force of 84 kg m/s2 we feel comfortable describing the event as

m2 m 2 1 newton
force = 84 kg m/s 2 = 84 kg 2 (1) = 84 kg 2 = 84 newtons (1-38)

s
s kg m s 2

Here we have arranged Eq. 1-37 in the form

newton N
1 = 2
= (1-39)
kg m/s kg m/s 2

and multiplied the quantity 84 kg m/s2 by one in order to effect the change in units. Note that in our
definition of the unit of force we have made use of a one-to-one correspondence represented by Eq. 1-37.
This is a characteristic of the S.I. system and it is certainly one of its attractive features. One must keep in
mind that Eq. 1-37 is nothing more than a definition and if one wished it could be replaced by the alternate
definition given by
1 bernoulli = (18.76 kg ) (1 m/s2 ) (1-40)

However, there is nothing to be gained from this second definition and it is clearly less attractive than that
given by Eq. 1-37.
In the English or standard system of units the one-to-one correspondence is sometimes lost and
confusion results. If we choose our standards of mass, length and time as
mass lbm
length ft
time s
we can define the pound-force as a unit of force according to

1 lbf = (1 lbm ) ( 32.17 ft/s2 ) (1-41)

This definition was chosen so that mass and force would be numerically equivalent when the mass was
acted upon by the earths gravitational field. While this may be a convenience under certain circumstances,
the definition of a unit of force given by Eq. 1-41 is certainly less attractive than that given by Eq. 1-37.
In summary, we note that there are only four standards needed to assign a numerical values to all
observables, and the choice of standard is arbitrary, i.e. the standard of length could be a foot, an inch or a
centimeter. Once the standard is chosen, i.e. the meter is the standard of length, other alternate units can be
constructed such as those listed in Table 1-2. In addition to a variety of alternate units for mass, length and
time, we construct for our own convenience a series of derived units. Some of the derived units for the SI
system are given in Table 1-3. Finally, we find it convenient to tabulate conversion factors for the derived
units for the various different systems of units and some of these are listed in Table 1-4. An interesting
history of the SI system is available from the Bureau International des Poids et Mesures12.

12
http://www.bipm.fr/enus/3_SI/si-history.html
Introduction 17

Table 1-4. Conversion Factors

Length Mass
1 inch (in) = 2.54 centimeter (cm) 1 pound mass (lbm) = 453.6 gram (g)

o 1 g = 10-3 kilogram (kg)


1 angstrom ( A) = 10-8 cm
1 foot (ft) = 0.3048 meter (m) 1 ton (short) = 2000 lbm
1 mile (mi) = 1.609 km 1 ton (long) = 2240 lbm
1 yard (yd) = 0.914 m
Time
Force 1 minute (min) = 60 second (s)
1 newton (N) = 105 dyne 1 hour (hr) = 60 minute (min)
1 dyne = 2.248 10
-6
pound force (lb f ) 1 day = 24 hour (hr)
-2
1 poundal = 3.108 10 lb f
1 lbf = 4.448 N Density

Volume 3
1 lb m /ft = 16.018 kg/m
3

1 in3 = 16.39 cm3 1 lb m /gal(US) = 119.83 kg/m


3

1 ft 3 = 2.83 10-2 m3
1 gallon (US) (gal) = 231 in3 Pressure
1 quart (liquid) (qt) = 0.25 gal (US) 1 atmosphere (atm) = 14.7 lbf /in2
1 barrel = 31.5 gal (US) 2 3
1 lbf /in = 6.89 10 N/m
2

1 gal (US) = 0.003785 m3 1 torr = 133.3 N/m2


1atm = 1.013 106 dyne/cm 2
Area 1 pascal (Pa) = 1 N/m
2

1 in2 = 6.452 cm2


2
1 ft = 9.29 10
2
m
2 Power

1 acre = 4.35 104 ft 2 1 horsepower (hp) = 745.7 watt


1 hp = 42.6 Btu/min
Energy 1 watt = 9.51 10-4 Btu/s
1 calorie (cal) = 4.186 joule 1 ft lbf /sec = 1.356 watt
1 British Thermal Unit (Btu) = 252 cal
1 erg = 10-7 joule Temperature
1 Btu = 1055 watt sec 1 C = 1.8 F
1 ft lbf = 1.356 joule K = C + 273.16
R = F + 459.60
Viscosity F = 1.8C + 32
1 poise = 1 g/cm s
1 poise = 0.10 N s/ m 2 Flow Rate
-2
1 poise = 6.72 10 lb m /ft s 1 gal (US)/min = gpm = 6.309 105 m3 / s
1 centipoise (cp) = 10-3 kg/m s 1 ft 3 / min = 4.719 10 5 m3 / s

Kinematic Viscosity
1 m 2 /s = 3.875 10 4 ft 2 /hr
1 cm 2 /s = 10-4 m 2 /s
1 stokes = 1 cm 2 /s
18 Chapter 1

1.5 Fluid properties


The motion of a fluid depends on the forces that are applied, the density of the fluid, and its mechanical
properties. The more dense fluid is the greater the force required to accelerate it, thus density is an
important fluid property. Similarly, the move viscous a fluid is the greater the force required to accelerate
it, thus viscosity is an important fluid property. The compressibility and the coefficient of thermal
expansion can play important roles in the motion of fluids, thus are important fluid properties.

1.5.1 Density
The density of liquids depends on both temperature and pressure, and it is convenient to characterize
this dependence in terms of the compressibility, , defined as

1
= (1-42)
p T

and the coefficient of thermal expansion, , defined as

1
= (1-43)
T p

Values of the compressibility and the coefficient of thermal expansion are given in Table 1-5 for several
liquids. From Eq. 1-42 we can represent the fractional change in the density of a typical liquid according to
p
p 1 104 (1-44)
atm
and this indicates that a significant change in the density requires a pressure change of one hundred
atmospheres. The influence of the temperature is also quite small and because of this we tend to treat the
density of fluids as constant. However, it is important to remember that small changes can only be
neglected when they give rise to small effects and we will return to this issue repeatedly throughout the text.

Table 1-5*
Approximate Physical Properties of Some Common Liquids

Liquid T, F , lb m / ft 3 , atm-1 , K 1

ethyl alcohol 68 49.2 120 106 1.12 103


benzene 68 54.8 90 106 1.20 103
carbon tetrachloride 68 128 98 106 1.24 103
6
mercury 68 846 3.9 10 0.18 103
SAE 30 oil 100 57.5 50 106 0.90 103
water 60 62.4 52 106 0.21 103

While the density of a liquid can often be treated as a constant, the situation for gases is quite different.
If we accept the ideal gas law

pV = nRT , ideal gas (1-45)

as providing a satisfactory representation of the behavior of a gas, we can express the density of a gas
consisting of a single component according to

*
To be changed to SI units.
Introduction 19

p ( MW )
= , ideal gas (1-46)
RT
Here MW represents the molecular mass, and we can use this result to determine the compressibility and
the coefficient of thermal expansion as

1
= , ideal gas (1-47a)
p

1
= , ideal gas (1-47b)
T
Whenever the temperature or pressure are close to the critical values, i.e., the pressure and temperature at
the critical point, the deviation from the ideal behavior represented by Eq. 1-45 becomes significant and
more realistic equations of state must be employed.

1.5.2 Viscosity
The distinctive mechanical characteristic of a fluid is that it deforms continuously under the action of a
shear stress. Solids undergo a deformation (or strain) under the action of a shear stress; however, this
deformation is independent of time except for the very slow changes that are known as creep. Fluids, on
the other hand, continue to deform as long as a shear stress is applied. The rate of deformation is
determined by the rate of change in the distance between two neighboring points divided by the distance.
Thus the rate of deformation is given by the change in length per unit length per unit time, and the relation
between the applied force and the rate of deformation represents a key mechanical property of a fluid.
For Newtonian fluids there is a simple linear relationship between the applied shear stress (force per
unit area) and the rate of deformation (change of length per unit length per unit time). The coefficient
relating the shear stress to the rate of deformation is called the coefficient of shear viscosity or often simple
the viscosity. The nature of viscosity can be described in terms of an experiment used to measure the
coefficient of shear viscosity, . This experiment consists of forcing one flat plate to move parallel to
another flat plate as illustrated in Figure 1-7. The flow generated by the moving plate is often known as
Couette flow in honor of the French mathematician who first analyzed a class of flows of this type13. If we
let the total force acting on the upper plate be f x and the area of the plate be A, experiments with

Figure 1-7. Velocity distribution between two parallel plates

Newtonian fluids indicate that the force per unit area is directly proportional to the velocity divided by the
channel thickness. Experimental results are suggested in Figure 1-8 and the equation for f x / A is given by

13
Rouse, H. and Ince, S. 1957, History of Hydraulics, Dover Publications, Inc., New York
20 Chapter 1

fx
= ( uo h ) (1-48)
A

Here we note that f x / A is the shear stress ( force / area ) and uo / h is the rate of deformation ( time 1 ).
The proportionality coefficient, , has the units of ( force time / area ) and a variety of representations in
terms of other units are listed in Table 2-4. At this point in time, it may not be obvious that uo / h
represents the change in length per unit length per unit time; however, this will become clear when the
kinematics of deformation is studied in Chapter 6.
The force per unit area, f x / A , represents a force in the x-direction acting on a surface for which y is a
constant. In subsequent chapters, it will become clear that it is convenient to express this force as

fx
= Tyx (1-49)
A

where Tyx is a component of the stress tensor. The nomenclature indicates that this is a force acting on the
y-surface (first subscript) in the x-direction (second subscript) and the double subscript nomenclature is

fx
A

experimental data

uo/h

Figure 1-8. Shear stress versus rate of deformation for a Newtonian fluid

required to properly describe shear stresses. It is also convenient to express uo / h as the velocity gradient,
d v x / dy , so that Eq. 1-49 takes the form

Tyx = ( d v x dy ) (1-50)

This relation between the shear stress, Tyx , and the rate of deformation, d v x / dy , is a special case of
Newtons law of viscosity, and fluids that obey this law are known as Newtonian fluids. The viscosity for a
wide variety of commonly encountered Newtonian fluids is presented as a function of temperature in
Figure 1-9. The fluids that we encounter in practice have viscosities ranging over several orders of
magnitude. For example, common lubricating oils have viscosities in the range of 100 centipoise while
water, gasoline and similar fluids have viscosities on the order of 1 centipoise. Gases, such as air, saturated
steam and hydrogen, have viscosities that are nearly three orders of magnitude less than water.

EXAMPLE 1.3. The type of flow illustrated in Figure 1-7 occurs in many lubricating devices in
which the fluid minimizes the friction caused when one solid surface slides past another. In some
cases this lubricating effect is undesirable, as in the case of an automobile sliding on a wet road
surface, and to see how effective water is as a lubricant, we consider this problem. We want to
determine the drag force exerted on the automobile, and then use this to determine the motion on
Introduction 21

the basis of Eq. 1-25. The drag force exerted on the tires (the moving plate in Figure 1-7) by the
water is given by

f drag = Tyx A (1)

and use of Eqs. 1-48 and 1-49 leads to

f drag = (uo / h) A (2)

Here A represents the contact area between the tires and the water film and h represents the depth
of the water film. We can apply the x-component of Eq. 1-25 to the automobile to obtain

d
(m vx ) = f drag (3)
dt

in which v x represents the rigid body velocity of the automobile in the x-direction. Because of
the no-slip condition, v x is equal to uo and we can use Eq. 2 to express Eq. 3 as

d
( m uo ) = (uo / h) A (4)
dt

We can solve this equation subject to the initial condition

IC. uo = uo , t =0 (5)

to obtain

A
ln uo (t ) uo = t (6)
mh

Given the following representative values

= 0.011 poise , A = 60 cm 2 , m = 2000 lb m , h = 0.001 cm (7)

in conjunction with a velocity change from 60 mph to 5 mph, we determine the time to be

2000 lb
( m )( 0.001 cm ) 5 poise ft s 30.48
t = ln = 3414 s (8)
( 0.011 poise ) 60 cm 2 60 6.72 10-2 lb m ft
( )

Obviously this is not consistent with our previous experience and the use of the flow illustrated in
Figure 1-7 is not acceptable as a model of a tire skidding on a wet pavement. Real pavement is
certainly not flat, in the sense illustrated in Figure 1-7, and tires are designed to minimize the
existence of a continuous liquid film. This comparison between the results of a model and
experiment (your experience) indicates that the model has failed.

1.5.3 Non-Newtonian fluids


Fluids that exhibit a linear relation between the shear stress, Tyx , and the rate of deformation,
d v x / dy , are called Newtonian fluids, while those fluids that exhibit a non-linear relation are generally
referred to as non-Newtonian fluids. The relation between the shear stress and the rate of deformation for
several types of non-Newtonian fluids is illustrated in Figure 1-10, and we will study the flow of non-
Newtonian fluids in Chapter 8. The Bingham plastic represents a curious fluid that behaves as a solid when
the shear stress is less than some critical value. When the shear stress exceeds the critical value, the
Bingham plastic undergoes a rate of deformation that is a linear function of the shear stress. Toothpaste is
a good example of a Bingham plastic.
22 Chapter 1

Figure 1-9. Viscosity of fluids


Introduction 23

Shear thinning fluids have the interesting property that the apparent viscosity decreases with increasing
rate of deformation, i.e., as the rate of deformation is increased the fluid becomes thinner. The apparent
viscosity, app , is determined by the relation

Tyx = app ( d v x dy ) (1-51)

which can be determined by curve shown in Figure 1-10. One can always find a region of linear behavior
at low rates of deformation (think about a Taylor series14 expansion in d v x / dy ), thus many shear thinning

plastic
Bingham

g
nin
ar thin
she

Tyx

ng
ian en
i
on ck
ewt i
N r th
ea
sh

dvx
dy

Figure 1-10. Behavior of non-Newtonian fluids

fluids exhibit what are referred to as lower and upper Newtonian regions. High molecular weight liquids,
such as polymers, often exhibit shear thinning behavior and this represents an important characteristic in
the design of coating devices where shear thinning behavior exacerbates coating film non-uniformities.
In addition to shear thinning fluids, there are fluids which behave in precisely the opposite manner, i.e.,
the apparent viscosity increases with increasing rate of deformation. One convenient example is a
suspension of corn starch in water which has the characteristic that it will flow smoothly when subject to
small rates of deformation and will become almost solid-like when subject to large rates of deformation.
Here we have been very vague since the words small and large have no precise meaning; however, in this
particular example the words small and large can be made precise by performing the experiment with a
suspension of corn starch and water.
Words such as small and large, slow and fast, smooth and rough, etc., have no meaning and
have no place in engineering analysis. In fact, they are dangerous words that should be
avoided since they mean different things to different people. You can give these words
meaning by providing a standard for comparison. For example, if you say that a car is
moving fast on an interstate freeway, your implied standard is 70 mph and you may mean
that the car is moving at 90 mph. On the other hand, if you say that a car is moving fast in a
school zone, your implied standard is 25 mph and you may mean that the car is moving at 35
mph. When someone says that the rate of deformation is large, your immediate response
should be, Large relative to what? To be precise about the behavior of a suspension of corn
starch in water, we need to provide some characteristic rate of deformation that will serve as

14
Stein, S.K. and Barcellos, A. 1992, Calculus and Analytic Geometry, Chapter 11, McGraw-Hill, Inc., New York
24 Chapter 1

a standard. Until we do that, our discussion is vague and imprecise and of little use to an
engineer. As we proceed through this text, we will become more precise and less vague.

1.5.4 Surface tension


Engineers are often confronted with problems of two-phase flow, and in many cases the effect of
surface tension plays a profound role in the behavior of such systems. The molecules at an interface
between two fluids exist in a state different from that of the molecules in the interior of the fluid. At an air-
water interface, for example, the water molecules are (on the average) surrounded by water molecules on
only one side, whereas the interior molecules are completely surrounded. Consequently, the
configurational energy of the surface molecules differs from that of the interior molecules and the surface
exists in a state of tension which is known as the surface or interfacial tension. Surface tension has units of
force per length and is usually denoted by . In Figure 1-11 we have illustrated some liquid drops resting

Figure 1-11. Influence of surface tension on drop configuration

on a solid surface. Surface tension causes drops to become spherical, and in Figure 1-11a the surface
tension effect is large compared to the gravitational effect, and both the small and large drops are nearly
spherical. In Figure 1-11b, we have illustrated a situation is which the surface tension effect is smaller and
the effects of gravity are more apparent. In Chapter 2 we will explore the influence of surface tension for
hydrostatic conditions.
The contact angle between the liquid and the solid also plays a significant role in the behavior of gas-
liquid-solid systems. In Figure 1-12a we have shown the contact angle, , which is the angle between the
solid-liquid interface and the liquid-gas interface. When the angle is greater than 90o, the liquid is
non-wetting and a drop will tend to retain a spherical form as indicated in Figure 1-12b. When the angle is
smaller than 90o, the liquid tends to wet the solid and will spread over the surface as indicated in
Figure 1-12c. Waterproof clothing is made from material that is non-wetting and this causes water to form
droplets, such as the one illustrated in Figure 1-12b, which do not penetrate into the fabric and are removed
by surface flow.

1.5.5 Vapor pressure


Under some conditions, the fluid pressure can become smaller than the vapor pressure, pvap , and
when this happens a vapor phase can form within a liquid phase. The vapor pressure represents the
pressure exerted by a liquid in equilibrium with its vapor at a specific temperature. The temperature at
which the vapor pressure is equal to one atmosphere is known as the boiling point temperature. If the fluid
pressure is greater than the vapor pressure, no vapor will form. When the vapor phase appears in the form
of bubbles, surface tension influences the behavior of the system. When the vapor bubbles move from a
region of low pressure to a region of higher pressure, surface tension will cause the bubbles to collapse and
this is how snapping shrimp make their distinctive sound.

1.6 Vectors
Describing complex physical phenomena in terms of abstract symbols is difficult since the symbols are
not the phenomena, they only represent the phenomena. In addressing this problem, it is wise to use
Introduction 25

Figure 1-12. Influence of contact angle on drop configuration

symbols which contain as much information as possible and which relate this information to the physical
world as closely as possible. For example, the temperature is a scalar and can be described entirely in
terms of the magnitude, thus it is sufficient to represent the temperature by the symbol, T. Since the
velocity is a vector having both magnitude and direction, we choose a notation which symbolizes this fact
and we represent the velocity as v. The use of boldface type to indicate vectors is referred to as Gibbs
notation after Josiah Willard Gibbs (1839-1903). Gibbs is known primarily for his contributions in
thermodynamics; however, he was a physicist with a wide range of interests. His interest in mathematics
led to the development of vector analysis, and his lectures on the subject are the basis of a book by
Wilson15. In the study of mechanics, one must think in terms of vectors for vector analysis is the language
of mechanics.
Our physical notion of a vector is that of a directed line element such as that illustrated in Figure 1-13
where we have shown a position vector locating one body relative to another. The projections of the vector

Figure 1-13. Relative position vector between two particles

15
Wilson, E.B. 1960, Vector Analysis, Dover Press, New York
26 Chapter 1

onto the coordinate axes represent the components of the vector, and in terms of these components we
express r as
r = i rx + j ry + k rz (1-52)

in which i, j and k represent the three unit base vectors. Given any vector a, we can express it in the form
of Eq. 1-52
a = i ax + j a y + k az (1-53)

or we can express it in terms of numerical subscripts as indicated by

a = e1a1 + e2 a2 + e3 a3 (1-54)

Here we have used e1 , e2 and e3 to replace i, j and k, and a1 , a2 , and a3 have been used to replace
ax , a y , and az . The numerical indices used in Eq. 1-54 allow us to represent a vector as

i =3
a = e a
i =1
i i (1-55)

and if we impose the summation convention we can express this result in the compact form given by

Summation Convention: a = ei ai (1-56)

Clearly the rule associated with the summation convention is that repeated indices in any single term (a
monomial) are summed from 1 to 3. Later we will introduce the free index convention which requires that
un-repeated indices (or free indices) represent the numbers 1, 2, and 3.

1.6.1 Scalar product


The dot or scalar product between two vectors can be expressed as

a b = ab cos (1-57)

in which a and b represent the magnitudes of the two vectors, a and b, and represents the angle between
the two vectors. One can also express the scalar product as

a b = ax bx + a y by + a z bz (1-58)

In terms of index notation this result takes the form

a b = a1b1 + a2b2 + a3b3 (1-59)

or in terms of the summation convention, we obtain

a b = ai bi (1-60)

Here one must remember that repeated indices are summed from one to three, and a little thought will
indicate that the particular symbol that one gives to a repeated index is irrelevant provided that it is not the
same as some other index in the equation.

1.6.2 Dummy indices


Repeated indices are often referred to as dummy indices since they have the same characteristic as a
dummy variable of integration. For example, in Eq. 1-56 one can re-label i to be n to obtain

a = ei ai = en an (1-61)

and in Eq. 1-60 one can re-label i to be k in order to express the dot product between two vectors as
Introduction 27

a b = ak bk (1-62)

The process of re-labeling can be used to complete some proofs in a rather simple manner.

1.6.3 Kronecker delta


The Kronecker delta appears often in problems involving vector and tensor manipulation. The dot
product between two unit base vectors can be expressed by

ei e j = ij (1-63)

in which ij is referred to as the Kronecker delta and has the characteristics given by

1, i = j
ij = (1-64)
0, i j
In terms of matrix notation, the elements of the Kronecker delta can be represented in terms of the identity
matrix16 of order 3 which we express as

1 0 0

I3 = 0 1 0 (1-65)
0 0 1

Often the Kronecker delta appears in terms such as ij b j and a little thought will indicate that this can be
written as

ij b j = bi (1-66)

Sometimes this operation is referred to as a contraction.

1.6.4 Cross product


The vector or cross product between two vector is defined by

a b = ab n sin (1-67)

in which represents the angle between the two vectors and n is the unit normal vector to the plane defined
by a and b. The direction of n is determined by the right hand rule that is indicated in Figure 1-14. In
terms of the unit base vectors, the cross product leads to the following results:
ij = k , j k = i, ki = j (1-68a)

ii = 0, j j = 0 , kk = 0 (1-68b)

j i = k , k j = i , ik = j (1-68c)

The structure that we see here suggests the use of the alternator defined by

+1 , ijk an even permutation of 1,2,3



ijk = 0 , any index repeated (1-69)
1 , ijk an odd permutation of 1,2,3

16
Kolman, B. 1997, Introductory Linear Algebra, Chapter 1, Prentice Hall, Upper Saddle River, NJ.
28 Chapter 1

An odd permutation of 1,2,3 is obtained by a single interchanging of two numbers, thus 2,1,3 represents an
odd permutation of 1.2.3. To obtain an even permutation of 1,2,3 we must interchange two numbers twice,

Figure 1-14. Right hand rule for the cross product of two vectors

thus 2,3,1 is an even permutation of 1,2,3. It is useful to remember that a cyclic permutation is an even
permutation. A cyclic permutation of 1,2,3 is achieved by removing the 1 from the first position and
placing it in the last position, thus 2,3,1 is a cyclic permutation of 1,2,3.
We can see that the alternator has the characteristics illustrated by

123 = 1 , 213 = 1 , 122 = 0 (1-70)

and these characteristics appear naturally in the cross products associated with unit base vectors. We
illustrate this with the traditional results
i jk = 1, j i k = 1 , i j j = 0 (1-71)

and note that these three results can also be expressed as

e1 e2 e3 = 1, e2 e1 e3 = 1 , e1 e2 e2 = 0 (1-72)

One can demonstrate that the cross product between any two unit base vectors can be written as
ei e j = ijk ek (1-73)

and this result is useful in the analysis of a variety of vector and tensor operations. The cross product can
also be used to represent the area of parallelograms as we have illustrated in Figure 1-15, and as can be
demonstrated by the application of Eq. 1-67.

1.6.5 Gradient operator


In rectangular coordinates, the gradient operator17 can be expressed as


= i + j + k (1-74)
x y z

The gradient of a scalar is given by

S S S
S = i + j + k (1-75)
x y z

while the divergence of a vector takes the form

17
Stein, S.K. and Barcellos, A. 1992, Calculus and Analytic Geometry, Chapter 14, McGraw-Hill, Inc., New York
Introduction 29

ax a y az
a = + + (1-76)
x y z

Each of these expressions can also be represented by index notation, and that will be left as an exercise for
the student.

1.6.6 Component form


The component form of a vector equation can be developed in terms of Gibbs notation or in terms of
index notation and both representations have certain advantages depending on the use and the user.

Gibbs Notation
In terms of Gibbs notation, we consider a vector equation of the form
a = b (1-77)

and express the two vectors in terms of the base vectors and components to obtain

i ax + j a y + k az = i bx + j by + k bz (1-78)

Forming the dot product with the base vector i leads to

(i i ) ax + (i j) a y + (i k ) az = (i i ) bx + (i j) by + (i k ) bz (1-79)

and from this we conclude that

ax = bx (1-80a)

By the same procedure we can extract

ay = by (1-80b)

az = bz (1-80c)

and we refer to Eqs. 1-80 as the component form of Eq. 1-77.

Figure 1-15. Area of a parallelogram

Index Notation
One can use index notation and the summation convention to express Eq. 1-77 as

ai ei = bje j (1-81)

In order to extract the component form, we form the dot product with the unit base vector ek to obtain

ai ek ei = b j ek e j (1-82)

In terms of the Kronecker delta this result provides


30 Chapter 1

ai ki = b j kj (1-83)

A little thought will indicate that

ai ki = ak , b j kj = bk (1-84)

and this leads to the component form expressed in index notation.

ak = bk (1-85)

On the basis of the free index convention, this result represents three separate equations

Free Index Convention: a1 = b1 , a2 = b2 , a3 = b3 (1-86)

that are equivalent to Eqs. 1-80.

EXAMPLE 1.4. Index notation and component form can be used to quickly prove a variety of
important vector identities. Consider the identity given by

(a b) c = (c a) b (1)

In terms of index notation, the left hand side can be expressed in terms of the following steps
involving the alternator and the Kronecker delta:

(a b) c = (ei ai e j b j ) eA cA
= ( ijk e k ai b j ) eA cA , alternator
(2)
= ( ijk ai b j )e k eA cA
= ( ijk ai b j ) k A cA , Kronecker delta

Making use of the property of the Kronecker delta allows us to express this result in the form

(a b) c = ( ijk ai b j ) k A cA = ijk ai b j ck (3)

A cyclic (even) permutation of the alternator leads to

(a b) c = ijk ai b j ck = kij ai b j ck (4)

and we can rearrange the components of the vectors to obtain the form

(a b) c = kij ai b j ck = kij ck ai b j (5)

On the basis of Eq. 4, we recognize that the last term in Eq. 5 is (c a) b and we have verified
Eq. 1. The advantage of index notation in component form is that complex vector operators can
be carried out quickly and accurately; however, if we wish to reassemble the results in terms of
Gibbs notation, we must be able to recognize certain forms such as kij ck ai b j .

1.6.7 Transformation law


Vectors have the obvious physical significance of a directed line segment, such as we have illustrated
in Figure 1-13. In addition, vectors have definite transformation properties with which we must become
familiar. For example, if we consider a coordinate rotation indicated by x, y, z x , y , z , the system
illustrated in Figure 1-13 takes the form shown in Figure 1-16. In the coordinate system x , y , z , we
express the position vector as r , and we represent the vector in terms of the base vectors and components
according to
Introduction 31

r = i rx + j ry + k rz (1-87)

If we think about two observers, one located in the x, y, z coordinate system and one located in the x , y , z
coordinate system, it should be clear that both observers see precisely the same phenomena. The first

Figure 1-16. Relative position vector in a rotated coordinate system

observer would denote the directed line segment joining the two particles by r while the second observer
(in the barred coordinate system) would denote it by r . Since these two vectors are one and the same, we
require
r = r (1-88)

and note that vectors are invariant to coordinate rotations. While the vector is invariant to coordinate
rotations, the components are certainly not and we need to develop the law that the components must
satisfy under a coordinate rotation.
The transformation properties of the components of a vector are best examined in two dimensions
because the algebraic effort associated with a fully three-dimensional transformation can obscure the
analysis. In Figure 1-17 we have illustrated the vector r and its components in the x, y coordinate system,

Figure 1-17. Two-dimensional transformation


32 Chapter 1

and we have shown the transformed coordinate system, x , y . From the geometrical construction
illustrated in Figure 1-18, we see that the x -component of the vector r is given by

rx = rx cos xx + ry cos xy (1-89)

and a little thought will indicate that the y -component can be expressed as

ry = rx cos yx + ry cos yy (1-90)

The four angles associated with the two-dimensional transformation can be described by
1. xx is the angle between the x -coordinate and the x-coordinate.

2. xy is the angle between the x -coordinate and the y-coordinate.

3. yx is the angle between the y -coordinate and the x-coordinate.

4. yy is the angle between the y -coordinate and the y-coordinate.

and from this it is clear that the subscripts on have a very definite significance. This will be helpful in
constructing a compact formulation of the transformation law for the components of a vector.

Figure 1-18. Components in the x , y coordinate system.

If the type of analysis that led to Eqs. 1-89 and 1-90 is extended to three dimensions, we obtain the
following three representations for the transformed components of the vector r.

rx = rx cos xx + ry cos xy + rz cos xz (1-91a)

ry = rx cos yx + ry cos yy + rz cos yz (1-91b)

rz = rx cos zx + ry cos zy + rz cos zz (1-91c)

The terms cos xx , cos yz , etc. are referred to as the direction cosines for the transformation
x, y, z x , y , z , and the array of nine direction cosines is referred to as the transformation matrix. In
writing Eqs. 1-91 we have simplified the nomenclature by omitting the overbar on the subscripts referring
to the x , y , z coordinate system, thus it is understood that xx is the angle between the x -coordinate and
the x-coordinate. This means that the transformation law represented by Eqs. 1-91 is not a recipe to be
Introduction 33

used for calculation, but instead it is a derivation that must be understood. Obviously Eqs. 1-91 can be
written in matrix form; however, it is convenient to use index notation to obtain a compact form of these
equations. We begin the development of this compact form by expressing Eqs. 1-91 in the form

r1 = r1 cos 11 + r2 cos 12 + r3 cos 13 (1-92a)

r2 = r1 cos 21 + r2 cos 22 + r3 cos 23 (1-92b)

r3 = r1 cos 31 + r2 cos 32 + r3 cos 33 (1-92c)

A more compact version of these three equations can be obtained by the use of the summation convention
which leads to

r1 = r j cos 1 j (1-93a)

r2 = r j cos 2 j (1-93b)

r3 = r j cos 3 j (1-93c)

We can now make use of the free index convention so that these three equations can be expressed as
ri = r j cos ij (1-94)

This is certainly a convenient form of the transformation law for the components of vectors; however, it
requires that one remember the free index convention and the summation convention. Typically, we
represent the direction cosines by

ij = cos ij (1-95)

so that the transformation law that can be expressed as

ri = ij r j (1-96)

It is often convenient to think of ij as the 3 3 matrix that maps the column vector of components in the
x, y, z coordinate system to the column vector of components in the x , y , z coordinate system. In matrix
notation we represent this mapping or transformation explicitly as

r1 11 12 13 r1
r = r (1-97)
2 21 22 23 2
r3 31 32 33 r3

This representation of the transformation law for the components of a vector certainly has considerable
visual appeal while the compact version illustrated by Eq. 1-96 has considerable utility and will be used
throughout the text. When applying Eq. 1-96 one must be careful to interpret ij correctly, and the correct
interpretation is given by Eqs. 1-91 through 1-96. We can summarize this interpretation as

ri = ij r j , requires that
(1-98)
cosine of the angle
ij = cos ij =
between xi and x j
34 Chapter 1

If the transformation is applied in the inverse sense, the interpretation of ij is given by

ri = ij r j , requires that
(1-99)
cosine of the angle
ij = cos ij =
between xi and x j

These results will be especially useful in the development of the general form of Newtons law of viscosity
in Chapter 6.

1.7 Problems

Section 1.1. The continuum postulate


1-1. If the measuring device illustrated in Figure 1-1 is used to measure the mass contained in V according
to

t = t

11
= m dt
V t
t =0

show how the curve of versus V changes as t ranges from zero to infinity.

Section 1.2. Types of flow


1-2. Given that the speed of sound in air at atmospheric pressure is about 335 meters per second, what is
the speed of sound in air at atmospheric pressure in terms of miles per hour? What is the speed of sound
for water in miles per hour?

1-3. If the average volumetric flow rate of water in a pipe of 1 inch diameter is 2 gallons per minute, do you
expect the flow to be laminar or turbulent?

1-4. The entrance length for laminar flow in a tube is given by Le / D = 0.058 Re . What is the entrance
length for the flow described in Problem 1-3?

Section 1.3. Newtons laws and Eulers laws

1-5. Show how Eqs. 1-20 and 1-21 are constructed on the basis of the word equations given by
Eqs. 1-9 and 1-10.

1-6. In Example 1.1, the mass of the non-uniform wire shown in Figure 1.6 was given by
x=L

mass =
dx
x =0
(1)

in which represents the mass per unit length. This quantity is given explicitly by

=
dA
A
(2)

where A represents the cross sectional area of the wire. In terms of cylindrical coordinates, can be
expressed as
Introduction 35

r = ro = 2

=

r =0 =0
d rdr (3)

The density of the wire (mass per unit volume) is given by

= o g ( x) 1 ( r ro )
n
(4)

in which the reference density, o , the parameters, and n, and the function, g ( x) , depend on the

Figure 1.6. Wire having a non-uniform mass density

manufacturing process. In this problem, you are asked to develop a general representation for based on
Eq. 4.

1-7. If the density is a function of position according to,

= o + ( x 12 L1 ) ( z L1 ) + ( y 13 L2 )
2

develop a general expression for the mass contained in the region indicated by

0 x L1 0 y L2 0 z L3

1-8. If the launch speed of the water balloon described in Example 1.2 is 30 m/s, what is the maximum
distance to a target that can be achieved?

1-9. Show that Eulers second law given by Eq. 1-26 reduces to Eulers first law for the process illustrated
in Figure 1-5.

1-10. Show that Eulers second law, subject to the mass-point approximation ( r = rCM ) for both bodies, can
be used to derive Eq. 1-33.

1-11. In Figure 1.11 we have shown an ensemble of mass points that are acted upon by two forces. One is
an external body force while the other is an interactive force between the mass points. Apply Newtons
laws to this system of N mass points in order to derive a linear momentum equation for the ensemble that
has the form
d
( m v ) = mb
dt

Here m is the mass of the ensemble, v is an appropriate velocity to be identified by the student, and b is the
mass-average body force defined by
k=N
mb = m b
k =1
k k
36 Chapter 1

Begin the analysis with Newtons second law in the form


i=N
d
dt
( mk v k ) = mk b k + f
i =1
ki

where f ki represents the force exerted by the i th mass point on the k th mass point. Make use of Newtons
third law

f ki = fik

to complete the development.

rk

Figure 1.11. Mechanics of an ensemble of mass points

1-12. When gas bubbles in a carbonated drink rise to the surface, surface tension effects cause small drops
to be propelled into the air. These are often felt on the tip of the nose and with proper lighting can be
observed visually. If a drop of liquid leaves the surface at a velocity uo , how high will the drop rise? Use
Eulers first law in the form given by Eq. 1-25 and neglect the drag on the drop caused by the surrounding
air. Show that your solution can be expressed in terms of a balance between potential energy per unit
volume and kinetic energy per unit volume given by ghmax = 12 uo2 . The relation between the laws of
mechanics and mechanical energy is examined in Chapter 11.

1-13. Give an example in which the first two axioms of special relativity are routinely employed in the
solution of boundary value problems.

1-14. What is an inertial frame?

1-15 Given Eulers two laws for linear and angular momentum


d
Euler I: v dV = b dV + t (n ) dA
dt
Vm (t ) Vm (t ) Am (t )

r v dV r b dV
d
Euler II: = + r t (n ) dA
dt
Vm (t ) Vm (t ) Am (t )
Introduction 37

prove that the position vector, r, can be determined relative to any arbitrary point in an inertial frame. Do
this by expressing the position vector as r = ro + r , in which ro locates a fixed point, and showing that
Eulers second law is independent of ro .

Section 1.4. Units


1-16. The weight of an object is defined as the force exerted on the object by gravity. If the mass of an
object is 10 lb m , what is its weight on the earth and on a planet where the gravitation acceleration is
10 ft/sec 2 ?

1-17. (a) Express the pressure, p = 12.5 poundals/ft 2 , in terms of both atmospheres and pascals.
(b) Express the kinematic viscosity, = 1.6 104 poise ft 3 / lb m , in square centimeters per second.
(c) Express the volumetric flow rate, Q = 120 gal/min , in cubic feet per second.
(d) An acre-foot of water is the volume of water needed to cover one acre to a depth of one foot. Express
an acre-foot in terms of cubic meters.

Section 1.5. Fluid properties


1-18. What is the percentage change in the density of water for a temperature change of 10 K?

1-19. For an N-component ideal gas mixture we have the following relations

pA V = n A RT , A = 1, 2, ...N

in which p A is the partial pressure of species A and R is the ideal gas law constant. If we sum this result
over all N species we obtain
A= N A= N
pV = nRT , in which p = A=1
pA , and n = n
A =1
A

For an ideal gas mixture, the total density is given by

p MW
=
RT

and in this problem you are asked to derive an expression for the mean molecular mass, MW .

1-20. Show how Eq. 6 is obtained from Eqs. 4 and 5 of Example 1.3.

1-21. In Example 1.3, the influence of wet pavement on the deceleration of an automobile was examined
using the simple model illustrated in Figure 1-7. The result was inconsistent with our experience and we
concluded that solid-solid contact must play an important role in the deceleration process. In California,
where rain is almost unheard of during the summer, there is an abnormal number of automobile accidents
when the first rain of the winter season arrives. One explanation is that drivers have forgotten how slippery
wet roads can be. Another is that the viscous oils which are deposited during the summer are the cause. Is
that possible?

1-22. The mechanical characteristics of a Bingham plastic are illustrated in Figure 1-10 and they can be
described mathematically as follows:

dv x
Tyx = ( o + To dv x dy )
dy
, when Tyx > To
38 Chapter 1

dv x
= 0, when Tyx To
dy

The first of these represents the linear regime in which a Bingham plastic behaves like a Newtonian fluid
while the second equation indicates that a Bingham plastic behaves like a solid if the shear stress is less
than the critical shear stress. If the critical stress, To , for a Bingham plastic is 2.8 N/m 2 and the density is
1390 kg/m3 , what film thickness would result if a glob of this material were applied to a vertical wall as
illustrated in Figure 1.22? Try modeling the glob as a rectangular block of material.

Figure 1.22. Bingham plastic flowing under the action of gravity

1-23. When non-Newtonian fluids are subjected to constant shear stress, such as we have illustrated in
Figure 1-9, one finds a functional relation of the form

Tyx = F ( dv x / dy )

Use a Taylor series expansion (see Problem 1-24) to define the region of linear behavior and identify the
apparent viscosity, app , in terms the function F ( dv x / dy ) .

1-24. A Taylor series expansion is a powerful tool for predicting the value of a function at some position
x = b when information about the function is only available at x = a . If we think about the definite
integral given by
x =b

f (b) = f (a) +
( df
x =a
dx ) dx (1)

we see that we can determine f (b) , given f (a) , only if we know the derivative, df / dx everywhere
between a and b. To see how we can use Eq. 1 to determine f (b) using only information at x = a , we
first make use of the change of variable, or transformation, defined by

= x b (2)

This leads to the relations


Introduction 39

dx = d , df dx = df d (3)

which can be used to express Eq. 1 in the form


=0

f (b) = f (a) +

= a b
( df d ) d (4)

Remember that the rule for differentiating a product

d dV dU
(UV ) = U + V (5)
d d d

is the basis for the technique known as integration by parts where it is employed in the form

dV d dU
U = (UV ) V (6)
d d d

If we make use of the two representations

df
U = , V = (7)
d

Eq. 6 provides the following identity for the derivative of f with respect to :

df d df d df
= (8)
d d d d d

Use this result in Eq. 4 to produce the second term in the Taylor series expansion for f (b) about f (a) .
Repeat this entire procedure to extend the representation for f (b) to obtain the third term in the Taylor
series expansion and thus illustrate how one obtains a representation for f (b) that involves only the
function and its derivatives evaluated at x = a . Keep in mind that the only mathematical tools used in this
derivation are the definition of the definite integral and the rule for differentiating a product18.

1-25. In Problem 1-12, we estimated that the maximum height of a liquid drop propelled into the air above
a carbonated drink was determined by ghmax = 12 uo2 , provided that the drag on the drop was negligible.
The force that causes drops to be ejected from the liquid is due to surface tension. The surface energy
2
(surface force times length) associated with the bubble in the liquid is on the order of Dbubble and this
will be on the order of the kinetic energy of the ejected drop, i.e., 1
2
uo2 Ddrop
3
. Assume that the drop
diameter is bounded by Ddrop < Dbubble , think about the size of bubbles in the typical carbonated drink, and
estimate the height, hmax .

Section 1.6. Vectors


1-26. What is the angle between the vector, a = 10i + 5 j and the x-axis?

1-27. Express the scalar, S = c (a + b) , in index notation.

18
Stein, S.K. and Barcellos, A. 1992, Calculus and Analytic Geometry, Chapters 5 and 3, McGraw-Hill, Inc., New York.
40 Chapter 1

1-28. Given the coordinate rotation shown in Figure 1-18, derive the result given by Eq. 1-90 making use of
the construction illustrated in Figure 1.28.

y
y

yy
x

ry yx yx
x

- yx

1-29. The invariance characteristic represented by Eq. 1-88 requires that the components of a vector and the
base vectors transform in a certain manner. For the coordinate rotation illustrated in Figure 1.28, develop
the transformation law for the base vectors i and j . Begin by representing these vectors as

i = ai + b j, j = ci + d j

Form the scalar product of these representations with the base vectors i and j in order to develop
expressions for a, b, c and d.

1-30. Represent Eqs. 1-74 through 1-76 in index notation.

1-31. Use index notation to show that the divergence of S b is given by S b + b S .

1-32. Use the summation convention and the property of the Kronecker delta to verify, ij b j = bi .

1-33. Use index notation and the alternator, ijk , to prove that

a) ab = ba

b) (a b ) a = 0

c) a (b c) = (a b) c
d) p = 0

1-34. Prove the vector identity (a b) = ( a) b a ( b) using the relation for the cross product
between base vectors given by ei e j = ijk e k .

1-35. Expand the divergence of the vector product, (a b) , to obtain a representation that contains the
gradient operator, , acting separately on both a and b.

You might also like