Fluid K2opt

Fluid Mechanics
Richard Fitzpatrick
Professor of Physics
The University of Texas at Austin
Contents
.............
...............................
...............
...............................
.................
1
7
Overview
1.1
Intended Audience
1.2
Major Sources
...............................
....................
..........................
7
1.3
To Do List
Mathematical
Models
of
Fluid
Motion
2.1
Introduction
9 2.2
..............
...............................
........
..............................
....
...............................
.......
...............................
........
............................
What is a Fluid?
9 2.3
Volume and Surface Forces
10 2.4
General Properties of Stress Tensor
Stress Tensor in a Static Fluid
11 2.5
......................
..........................
...................
.......................
............
...............................
........
...............................
...........
............................
..........
...............................
.............
............................
.....
..........................
.........
.......................
...........................
Stress Tensor in a Moving Fluid
12 2.6
13
2.7
Viscosity
14 2.8
Conservation Laws
15 2.9
Mass Conservation
15 2.10
Convective Time Derivative
16 2.11
Momentum Conservation
16
2.12
18 2.13
Navier-Stokes Equation
Energy Conservation
18
2.14
Equations of Incompressible Fluid Flow
20 2.15
Equations of Compressible Fluid Flow
21 2.16
Dimensionless
Numbers
in
Incompressible
Flow
22 2.17
....
.......................
.
...............................
..............................
.
...............................
..................
...............................
...................
...........................
............
...............................
...................
............................
..........
..........................
..........
.......................
..........
.......................
...
...............................
........
...............................
Dimensionless
Numbers in Compressible
Flow
23 2.18
Fluid
Equations
25
2.19
26
2.20
in
Cartesian
Fluid Equations
Fluid
27 2.21
Equations
Coordinates
in Cylindrical
Coordinates
in Spherical
Coordinates
Exercises
28
31
3.1
Hydrostatics
Introduction
31 3.2
Hydrostatic Pressure
31 3.3
Buoyancy
31
3.4
Equilibrium of Floating Bodies
32 3.5
Vertical Stability of Floating Bodies
33 3.6
Angular Stability of Floating Bodies
34 3.7
Determination
35 3.8
of Metacentric
Height
Energy of a Floating Body
..............
............................
..............
..........................
..
...............................
...........
...............................
..............
...............................
..............
...............................
..................
38 3.9
Curve of Buoyancy
38
3.10
Rotational Hydrostatics
121 2
FLUID MECHANICS
3.1 1
44
3.12
Equilibrium
of
Rotating
Liquid
Body
Maclaurin Spheroids
46 3.13
Jacobi Ellipsoids
...............................
49 3.14
Roche Ellipsoids
51 3.15
Exercises
57
...................
...........................
................
.......................
...................
.......................
..............
...............................
.............
...............................
4
61
4.1
Surface
Tension
Introduction
61 4.2
Young-Laplace Equation
61 4.3
Spherical Interfaces
63 4.4
Capillary Length
63 4.5
Angle of Contact
................
...............................
..............
...............................
.........
............................
..................
...............................
....................
..........................
...
.......................
............
...............................
...
.......................
................
.......................
...............
.......................
..............
...............................
............
64 4.6
Jurins Law
65 4.7
Capillary Curves
66 4.8
Axisymmetric Soap-Bubbles
70
4.9
Exercises
75
5
Incompressible
Inviscid
Fluid
Dynamics
77
5.1
Introduction
77 5.2
Streamlines, Stream Tubes, and Stream Filaments
77 5.3
Bernoullis Theorem
77
5 .4
Vortex Lines, Vortex Tubes, and Vortex Filaments
79 5.5
Circulation and Vorticity
80 5.6
Kelvin Circulation Theorem
80 5.7
Irrotational Flow
81 5.8
Two-Dimensional Flow
............................
.........
..........................
.................................
....
...............................
.................
...
.
...............................
.
...............................
............................ ...
...............................
..................
...............................
....................
..........................
...................
83
5.9
Two-Dimensional Uniform Flow
85 5.10
Two-Dimensional Sources and Sinks
86 5.11
Two-Dimensional Vortex Filaments
87
5.12
Two-Dimensional
Irrotational
Flow
in Cylindrical Coordinates
90 5.13
Inviscid
91 5.14
94 5.15
Flow
Past
Inviscid
Cylindrical
Flow
Past
Inviscid Flow Over
Obstacle
Semi-Infinite
a Semi-Infinite
Wedge
Wedge
95
5.16
97 5.17
Velocity
Potentials
and
Stream
Functions
Exercises
98
101
6.1
2D Potential Flow
Introduction
101 6.2
Complex Functions
......................
.
...............
.......................
...............
......................
.
....................
.......................
....................
.......................
..............
.............................
............
............................
...............
...........................
..................
.............................
101 6.3
Cauchy-Riemann Relations
102 6.4
Complex Velocity Potential
102 6.5
Complex Velocity
103 6.6
Method of Images
104 6.7
Conformal Maps
109 6.8
Complex Line Integrals
113
6.9
Theorem of Blasius
114 6.10
Exercises
118
.................
.............................
.............
.............................
.........
.............................
7
121
Incompressible
7.1
Boundary
Layers
Introduction
121
7.2
No Slip Condition
7.3
Boundary Layer Equations
121
209
CONTENTS
.......
.............................
........
............................
.....
.............................
.......
............................
............
..........................
..............................
..
......................
.
..................
.............................
....................
..........................
...........
.......................
...................
.......................
................
7.4
Self-Similar
125 7.5
Boundary
Boundary Layer
Layers
ona Flat Plate
128
7.6
Wake Downstream of a Flat Plate
Von K arm an Momentum Integral
132 7.7
136
7.8
Boundary Layer Separation
137 7.9
Criterion for Boundary Layer Separation
140 7.10
Approximate
Solutions
of Boundary
Layer Equations
142 7.11
Exercises
147
Incompressible
Aerodynamics
149
8.1
Introduction
149 8.2
Theorem of Kutta and Zhukovskii
149 8.3
Cylindrical Airfoils
151 8.4
Zhukovskiis
Hypothesis
......................
.
......................
.......................
......................
.......................
...............
.......................
...................
.......................
...................
.......................
.................
......................
.
..................
153 8.5
Vortex Sheets
158 8.6
Induced Flow
159 8.7
Three-Dimensional Airfoils
159 8.8
Aerodynamic Forces
162 8.9
Ellipsoidal Airfoils
165 8.10
.............................
Simple Flight Problems
167 8.11
Exercises
168
....................
..........................
..............
.......................
.......
.............................
................
............................
............
9
Incompressible
Viscous
Flow
171
9.1
Introduction
171 9.2
Flow Between Parallel Plates
171 9.3
Flow Down
172 9.4
an Inclined
Plane
Poiseuille Flow
174
9.5
Taylor-Couette Flow
.............................
.....
.............................
..............
............................
...................
...........................
...............
.......................
...
.......................
.......................
..................
.............................
174 9.6
Flow in Slowly-Varying Channels
175 9.7
Lubrication Theory
177
9.8
Stokes Flow
179 9.9
Axisymmetric Stokes Flow
180 9.10
a Solid
Axisymmetric
Stokes Flow Around
Axisymmetric
Stokes Flow In and Around
Sphere
181 9.11
a Fluid
Sphere
185 9.12
Exercises
188
....................
..........................
...............
.............................
.......
.............................
.......
............................
.............
10
Waves
in
Incompressible
Fluids
191
10.1
Introduction
191 10.2
Gravity Waves
19110.3
Gravity Waves inDeep Water
193 10.4
Gravity Waves in Shallow Water
194
10.5
Energy of Gravity Waves
..........................
............
.............................
..................
............................
.........
..........................
..............
.......................
.........
.......................
.....................
.......................
.....................
.......................
.............
.......................
...........
.......................
.........................
.......................
....................
..........................
.................
.......................
195 10.6
Wave Drag
196 10.7
on
Ships
Ship Wakes
198
10.8
Gravity Waves in a Flowing Fluid
202 10.9
Gravity Waves at an Interface
203 10.10
Steady Flow
over a Corrugated
Bottom
205 10.11
Surface Tension
205 10.12
Capillary Waves
206 10.13
Capillary Waves at an Interface
207 10.14
Wind Driven Waves inDeep Water
208 10.15
Exercises
275 4
FLUID MECHANICS
1 1Equilibrium
211
11.1
Introduction
of
Compressible
Fluids
21111.2
Isothermal Atmosphere
211 11.3
..................
.......................
..................
.......................
.................
.......................
.........................
.......................
...................
...........................
...................
.......................
......................
.......................
...........
.......................
...............
.......................
......................
.......................
.......................
.......................
......................
.......................
.........................
.......................
..................
.......................
..................
Adiabatic Atmosphere
212 11.4
Atmospheric Stability
213 11.5
Eddington Solar Model
213 11.6
Exercises
219
A
223
A.1
Vectors
and
Vector
Fields
Introduction
223 A.2
Scalars and Vectors
223 A.3
Vector Algebra
223 A.4
Cartesian Components of a Vector
225 A.5
Coordinate Transformations
226 A.6
Scalar Product
227 A.7
Vector Area
228 A.8
Vector Product
229 A.9
Rotation
231 A.10
Scalar Triple Product
233 A.11
Vector Triple Product
.......................
.....................
.......................
......................
.......................
..................
.......................
.....................
.......................
.................
.......................
.....................
.......................
.........................
.......................
......................
.......................
........................
.......................
...................
.......................
................................
...................
.................
.......................
.........................
.......................
...................
...........................
234 A.12
Vector Calculus
234 A.13
Line Integrals
235 A.14
Vector Line Integrals
237 A.15
Surface Integrals
237 A.16
Vector Surface Integrals
239 A.17
Volume Integrals
239 A.18
Gradient
240 A.19
Grad Operator
243 A.20
Divergence
243 A.21
Laplacian Operator
246 A.22
247
Curl
A.23
Useful Vector Identities
250 A.24
Exercises
250
B
253
B.1
Cartesian
Tensors
Introduction
253 B.2
..............
.......................
.................
.......................
......................
.......................
....................
.......................
.........................
.......................
....................
..........................
..........
.......................
.................
.......................
..................
.......................
.........................
.......................
..........
.............................
............
Tensors and Tensor Notation
253 B.3
Tensor Transformation
255 B.4
Tensor Fields
257 B.5
Isotropic Tensors
259 B.6
Exercises
261
Non-Cartesian
Coordinates
265
Introduction
C.1
265 C.2
Orthogonal Curvilinear Coordinates
265 C.3
Cylindrical Coordinates
268 C.4
Spherical Coordinates
270 C.5
Exercises
272
Calculus
of
Variations
273
D.1
Euler-Lagrange Equation
D.2
Conditional Variation
273
CONTENTS
..........
.............................
...................
.............................
D.3
Multi-Function Variation
276
D.4
Exercises
277
Ellipsoidal
Potential
Theory
6
FLUID MECHANICS
7
Overview
1
Overview
1.1
Intended Audience
This book presents
a single semester course on fluid
mechanics that is intended primarily for advanced

undergraduate
students majoring inphysics. A thorough
understanding of physics at the lower-division level,

including
a basic
working knowledge of the laws of
mechanics, is assumed. It is also taken for granted that

students
are familiar
with the fundamentals of
multi-variate integral and differential calculus, complex

analysis, and ordinary differential
equations. On the other
hand, vector analysis plays such a central role inthe

study of fluid mechanics that
a brief,
comprehensive, review of this subject
but fairly
area
is provided in
Appendix A.Likewise, those aspects of cartesian tensor

theory, orthogonal curvilinear coordinate systems, and the
calculus of variations, that

mechanics
are outlined
are required
inthe study of fluid
inAppendices B,C, and D,
respectively.
1.2
Major Sources
The material appearing inAppendix A is largely based

the authors recollections of a vector analysis
course
on
given
by Dr. Stephen Gull at the University of Cambridge.

Major
sources
for the material appearing inother chapters
and appendices include:
Statics, Including Hydrostatics and the Elements of the
Theory of Elasticity H.Lamb,3rdEdition(Cambridge
University Press, Cambridge UK, 1928).
Hydrodynamics H.Lamb, 6th Edition (Dover, New York

NY, 1945).
Theoretical Aerodynamics L.M. Milne-Thomson, 4th
Edition, Revised and enlarged (Dover, New York NY,
1958). Ellipsoidal Figures of Equilibrium S.
Chandrasekhar (Yale University Press, New Haven CT,
1969). Boundary Layer Theory H.Schlichting, 7th Edition

(McGraw-Hill, New York NY, 1970).
Mathematical Methods for the Physical Sciences

K.F. Riley (Cambridge University Press, Cambridge UK,
1974). Fluid Mechanics L.D. Landau, and E.M. Lifshitz,
2nd Edition (Butterworth-Heinemann,Oxford UK, 1987).
Physical Fluid Dynamics D.J. Tritton, 2nd Edition

(Oxford University Press, Oxford UK, 1988).
Fluid Dynamics for Physicists T.E. Faber,
1st Edition (Cambridge University Press, Cambridge UK,

1995). Schaums Outline of Fluid Dynamics W. Hughes,
and J.Brighton, 3rd Edition (McGraw-Hill, New York NY,
1999). An Introduction to Fluid Dynamics G.K.
Batchelor (Cambridge University Press, Cambridge UK,
2000). Theoretical Hydrodynamics L.M. Milne-Thomson,
5th Edition (Dover, New York NY, 2011).
1.3
To Do List
1.Add chapter
2.Add chapter
on vortex dynamics.
on 3D potential flow.
3.Add appendix
ongroup
velocity and Fourier
transforms.
4.Add chapter
on incompressible
flow inrotating
systems.
5.Add chapter
8
on instabilities.
FLUIDMECHANICS
6.Add chapter
7.Add chapter
8.Add chapter
9.Add chapter
10. Add chapter
on turbulence.
on 1D compressible flow.
on sound waves.
on compressible boundary layers.
on supersonic aerodynamics.
11. Add chapter
on convection.
Mathematical Models ofFluidMotion

9
2 Mathematical Models of Fluid

Motion
2.1
Introduction
we set
In this chapter,
commonly
dynamics
analysis
to
used
of
forth the mathematical
describe
the
models
equilibrium
and
,, .
fluids. Unless stated otherwise, all of the
is performed
Cartesian coordinate
using
system: x1
standard
x2
x3
right-handed
Moreover, the
Einstein summation convention is employed (so repeated
roman
subscripts
are
assumed to be summed from 1

to
3see Appendix B).
2.2
What is a Fluid?
By definition,
rigid material
large stresses, it
a stress
shape,
a
very
solid material is rigid. Now, although
tends to shatter when subjected
can withstand a moderate
to
shear stress (i.e.,
that tends to deform the material by changing its

without
necessarily
indefinite period. To be
changing its volume)
more exact,
when
for
a shear stress
an
is
to a rigid material it deforms slightly, but then
first applied
back
to its original
A plastic
material, such
springs
when the stress
shape
is
relieved.
degree of rigidity. However,
as
clay, also
which it yields is relatively
small, and
exceeded
deforms
the
material
irreversibly, and does
recover
not
possess some
the critical shear stress at
once
this stress is
continuously
and
its original shape when
the stress is relieved.

By definition,
fluid material
to
withstand
any
possesses no
rigidity
small fluid element is unable
tendency
ofan applied shear stress to
at all. In other words,
change its shape. Incidentally, this does not preclude the

possibility
that such
an element may offer

any resistance must
shear stress. However,

of
preventing
occurring,
vanishes
the
which
with
change
implies
the
in shape
resistance
to
be incapable
eventually
from
that the force of resistance
rate of
deformation.
corollary is that the shear stress must be
An obvious
zero
everywhere
a fluid that is inmechanical equilibrium.

Fluids are conventionally classified as either
gases. The most important difference between
inside
liquids
types of fluid lies in their relative compressibility:
gases can
be compressed
Con-sequently,
variations
changes
any motion
is generally
much
more
i.e.,
easily than liquids.
that involves significant

accompanied
or
these two
by
pressure
much
larger
in mass density inthe case of a gas than inthe
case
of
a liquid.
Of course, a macroscopic
a huge
fluid ultimately consists of
number of individual molecules. However, most
practical applications
on
the
intermolecular
typical
circumstances,
properties
of
completely
length-scales
given fluid
continuous
that
are
concerned
far larger than
spacing.
Under
these
suppose that the bulk

are the same as if it were
it is reasonable
are
of fluid mechanics
with behavior
to
in structure. A corollary
assumption is that when, in the following,
of this
we talk about
mean elements
volume elements, we really

are sufficiently small that the bulk fluid properties
(such as mass density, pressure, and velocity)
are
approximately
constant
across them, but are still
sufficiently large that they contain a very great number of
molecules (which implies that we can safely neglect any
infinitesimal
which
statistical
variations
in
continuum
hypothesis
also requires infinitesimal
elements
to
be
much
the
larger
bulk
than
properties).
the
The
volume
molecular
mean-free-path between collisions.
In addition to the continuum hypothesis,

fluid mechanics is premised
1.Fluids
are isotropic
on three major
our
study of
assumptions:
media: i.e., there is no preferred
direction in a fluid.
2. Fluids
are
Newtonian: i.e., there is a linear relationship
between the local shear stress and the local rate of strain,
as
first postulated by Newton. It is also assumed that
there is a linear relationship between the local heat flux

density and the local temperature gradient.
3.Fluids
are classical:
i.e., the macroscopic motion of
ordinary fluids is well-described by Newtonian dynamics,

and both quantum and relativistic effects
can be safely
ignored.
are not
It should be noted that the above assumptions
valid for all fluid types (e.g., certain liquid polymers,
are non-isotropic; thixotropic fluids, such as jelly

or paint, which are non-Newtonian; and quantum fluids,
such as liquid helium, which exhibit non-classical effects on
which
macroscopic
length-scales).
However,
most
practical
10
FLUID MECHANICS
of fluid mechanics involve the equilibrium and motion of
bodies of water
or air, extending over
macroscopic
length-
scales, and situated relatively close to the Earths surface.

Such
bodies
are
very
well-described
Newtonian, classical fluids.
2.3
Volume and Surface Forces
as
isotropic,
Generally speaking, fluids
are acted upon by two
distinct
types of force. The first type is long-range in nature

i.e.,
such
that
it
decreases
relatively
slowly
with
increasing distance between interacting elementsand

capable
fluid. Gravity
force. One
is
a
an obvious example of a long-range
consequence of the relatively slow variation
of completely
penetrating
into the interior of
is
of long-range forces with position is that they act equally
on all of
the fluid contained within
sufficiently
small
volume element. In this situation, the net force acting
on
the
element
volume. For this
becomes
reason,
volume forces. In the
directly
proportional
long-range forces
following,
we
to its
are often called
shall write the total
volume force acting at time t on the fluid contained within
a small volume element of magnitude

a fixed point whose position vector is r,as
dV, centered
on
F(r,t)dV.
(2.1)
The second type offorce is short-range innature, and is
most conveniently modeled
as momentum transport within

a combination
the fluid. Such transport is generally due to
of the mutual forces exerted by contiguous molecules, and
momentum
fluxes caused by relative
molecular
motion.
Suppose that x (r, t) is the net flux density of x-directed
r and
rand time t,
fluid momentum due to short-range forces at position
time t.In other words, suppose that, at position
as a
direct
is flowing
consequence
per
squared
of short-range forces, x-momentum
at the rate of |x |
newton-seconds
second in the direction of vector x
per meter
Consider
an infinitesimal plane surface element, dS = ndS, located

at point r.Here, dS is the area of the element, and nits unit
normal. (See Section A.7.) The fluid which lies on that
side of the element toward which n points is said to lie on its
positive side, and vice versa. The net flux of x-momentum
across the element (in the direction of n) is x dS
newtons, which implies
motion) that the fluid
element experiences
to short-range
fluid
dS in the
on
the positive
on
with the fluid
to Newtons
on the negative
area
third law of motion, the
side. Short-range
are
forces
across
often called
momentum, respectively,
extension
which they act. Let
and
z-
time t. By
surface
negative side is
element,
argument,
above
surface force exerted by the fluid
planar
r and
at position
of
on
dS,
to
proportional
y (r, t) and z (r, t) be the net flux density of
some
a force
x-direction due to interaction

surfac eexwith
perien
the
cesa
fluidf
of the surface element
straightforward
due
the negative
side of the surface experiences
surface
flu id onforces
thepo sitive
because
side.
they
Short
are
-range
directly
the
law of
positive side of the surface
force x dS in the x-direction
interaction
side. According
second
(from Newtons
on the
y-
the
the positive side
on
the fluid
on
net
of
its
= (x
dS, y dS, z dS).
(2.2)
In tensor notation (see Appendix B), the above equation
can be written
=ij
fi
dS
(2.3)
= (x )x
where 11
12
= (x )y
21
= (y )x
etc.
(Note that, since the subscript j is repeated, it is assumed

P
to be summed from 1
to 3.Hence, ij dS
P
j=1,3
dS j .) Here, the ij (r, t)
ij
stresses in the fluid at position

of
per
force
unit
area.
are
is shorthand for
termed the local
r and time t,and have units

the ij
are the
Moreover,
of a second-order tensor (see Appendix B),

as the stress tensor. [This follows because the fi
are the components of a first-order tensor (since all forces
are proper vectors), and the dS i are the components of an
arbitrary first-order tensor (since surface elements are also
proper vectorssee Section A.7and (2.3] holds for
surface elements whose normals point in any direction), so
components
known
application
of the quotient
Equation (2.3) reveals that the

of
the
coordinate
second-order
tensor.]
rule (see
ij
axes as
We can
Section
B.3)
to
transform under rotation

the
components
interpret
ij
(r, t)
of
as
the
i-component
r and
per unit area exerted, at position

a plane surface element normal to the
three diagonal components of ij are
of the force
time t,across
j-direction.
termed
The
normal
the
stresses, since each of them gives
area acting across

a plane surface element parallel to one of the Cartesian
coordinate planes. The six non-diagonal components are
normal component of the force
per
unit
termed shear stresses, since they drive shearing motion in

which parallel layers of fluid slide relative to one another.
11
2.4
General Properties of Stress Tensor
The i-component
element
consisting
of the total force

of
surface S is written
on a
acting
fixed volume
V enclosed
ZV
fi
fluid
by
IS
Fi dV
dS
ij
(2.4)
where
the first
term
on
the
right-hand
side
is
the
integrated volume force acting throughout V, whereas the

second term is the net surface force acting
use of
across
S Making
the tensor divergence theorem (see Section B.4), the
above expression becomes
ZV
fi
ZV
Fi dV
ij
dV.
(2.5)
to
In the limit V 0, it is reasonable

and
suppose
that the Fi
ij /x
approximately
In jtheare
l imit
V 0, it i
element.
In this situation,
constant across
the
s reasonable
to su ppos
on the
equation scale as V.Now,
both contributions
right-hand side of the above
according to Newtonian dynamics, the i-component
net
force
acting
i-component
on
the
element
is
equal
of the
to the
of the rate of change of its linear momentum.
However, in the limit V 0, the linear acceleration
and
mass density of the fluid are both approximately constant

across
the
element.
nd
mass
d ensity
of In
thethis
fluidcase, the rate of change of the
elements linear momentum
also scales as V. In other
words, the net volume force, surface force, and rate of
change
element
of linear
all scale
momentum
as
of
the volume
an
infinitesimal
of the element,
fluid
and
same order of
magnitude as the volume shrinks to zero. We conclude
that the linear equation of motion of an infinitesimal fluid
element places no particular restrictions on the stress
consequently
remain
approximately
the
tensor.
The i-component
of the total torque, taken about the
origin O of the coordinate

element
that consists of
system, acting
on a
fluid
fixed volume V enclosed by
surface S is written [see Equations (A.46) and (B.6)]

ZV
i
IS
xj
ijk
Fk dV
xj
ijk
kl
dS
(2.6)
where the first and second terms
are
to volume
due
[Here,
and
is the third-order
ijk
Equation (B.7).] Making
on the
surface
right-hand side
forces,
permutation
use
respectively.
tensor. See
of the tensor divergence
theorem (see Section B.4), the above expression becomes

ZV
i
ZV
ijk
xj
Fk dV
(x
j kl
)
dV,
ijk
xl
(2.7)
which reduces
to
ZV
i
since
xj
ijk
xi /x
ZV
Fk dV
ijk kj
.
,,
=
ij
ZV
[Here,
dV
ijk
xj
kl
dV,
xl
(2.8)
ij
is the
second-order
identity tensor. See Equation (B.9).] Assuming that point

O lies within the fluid element, and taking the limit V 0
in which within
the Fi
constant
across
ij fluid
and ele
ij men
/x j t,
are
the
andall
ta approximately
king th e lim
we deduce that
on the right-hand side of
the element,
second, and third terms
the first,
the above
equation scale
1/3).
V
Now,
as V
4/3,
V,and V
4/3,
respectively
to Newtonian
according
i-component of the total torque acting

is equal to the i-component
angular
momentum
on the fluid
we
deduce
x
the
element
of the rate of change of its net
about O. Assuming
that the linear
constant
acceleration of the fluid is approximately

the element,
(since
dynamics,
across
that the rate of change of its
as V 4/3 (since the net linear

acceleration scales as V, so the net rate of change of
1/3).
angular momentum scales as x V,and x
netV rate of
Hence,
c hangit
is clear that the rotational equation of motion of a fluid
element,
surrounding
a general point O, becomes
completely
dominated
by the second term on the
angular momentum
scales
right-hand side of (2.8) in the limit that the volume of the
element approaches
zero
(since this term is a factor V
1/3
larger than the other terms). It follows that the second
term must be identically

fluid
element
zero
would acquire
(otherwise
an
an infinitesimal
absurdly
large angular
velocity). This is only possible, for all choices of the

position of point O,and the shape of the element, if
ijk kj
(2.9)
throughout
the fluid. The above relation shows that the
stress tensor must be symmetric: i.e.,

ji
=ij
12
FLUID MECHANICS
components
independent
23
and
,,,,,
follows that the stress tensor only has six
It immediately
(i.e., 11
).
22
33
12
13
Now, it is always possible to choose the orientation
a set of
non-diagonal
of
Cartesian
axes
components
space.
(See
axes,
principal
tensor ij
33
given
that the
symmetric
are all set to zero at a given
second-order tensor field

in
a manner
in such
of
Exercise
B.6.) With reference
the diagonal
components
point
,,
to such
of the stress
so-called principal stresses 11 22

course, in general, the orientation of the
become
say. Of
11
principal axes varies with position. The normal stress
acting across a surface element perpendicular to the first
principal axis corresponds to a tension (or a compression
if
11
for
is negative) in the direction of that axis. Likewise,

22
and
33
tensions,
a
can be regarded as a superposition
Thus, the general state of the fluid, at
particular point in space,
or compressions,
inthree orthogonal directions.
The trace of the stress tensor, ii

is
a scalar,
= 11
and, therefore, independent
+ 22 + 33
of the orientation
of the coordinate
axes. (See
that, irrespective
of the orientation of the principal
Appendix B.) Thus, it follows
axes,
the trace of the stress tensor at a given point is always equal
to the
sum of the principal stresses:
i.e.,
= 11
ii
22
33
(2.11)
Stress Tensor ina Static Fluid
2.5
Consider the surface forces exerted

cubic volume element of
components
constant
stress
the
of
across
a static
on some
infinitesimal
fluid. Suppose that the
tensor
are
approximately
the element. Suppose, further, that the
sides of the cube
are
aligned parallel to the principal
of the local stress tensor. This tensor, which
now
has
axes
zero
non-diagonal
components,
can
be regarded
i.e.,

13
0
0
(2.12)
and
ii
13
0
0
ii
13
ii
as
the
sum
of two tensors:
11
1
ii
3
22
13
0
0
ii
33
1
ii
3
(2.13)
The first of the above tensors is isotropic (see Section
same
B.5), and corresponds to the
area
normal force
per
unit
acting inward (since the sign of ii /3 is invariably
on
negative)
uniform
each face
of the volume
element.
This
compression acts to change the elements volume,
but not its shape, and
can
easily be withstood by the fluid
the
above
within the element.

The
second
of
tensors
represents
the
an isotropic form. The

tensor have zero sum, in
departure of the stress tensor from
components
diagonal
of this
view of (2.11), and thus represent
per
forces
area,
unit
acting
volume element, which
one
pair of opposing
forces
on at
least
forces necessarily
one
by
any
axes.
equal and opposite
opposing
are such that
pair constitute
faces of the
the forces
faces constitute
on at
least
tension, and the
a compression.
Such
tend to change the shape of the volume
element, either elongating
symmetry
on
Moreover,
or compressing
it along
one of its
this tendency cannot be offset
volume force acting
on
the element, since such
forces become arbitrarily small compared to surface forces
in the limit that the elements
surface force scales
as
tends to zerosee
previously defined
any
withstanding
tends to
zero
in the limit
area
of
that the
Section 2.4). Now,
a fluid as a material
zero
force to the net
the volume to the surface
the element, which tends to

volume
volume
the ratio of the net volume
(because
we
have
that is incapable
of
tendency of applied forces to change its
shape. (See Section 2.2.) It follows that if the diagonal
components of the tensor (2. 13) are
,, ,
inside the fluid then it is impossible

point to be at rest. Hence,
stresses,
11
all points
in
22
and
33
the
conclude that the principal
must be equal to one another at

form (2.12) everywhere
fluid. Furthermore,
orientation
components
anywhere
static fluid. This implies that the stress
tensor takes the isotropic

stationary
we
non-zero
for the fluid at that
of
of
the
an
isotropic
in
this is true irrespective
coordinate
tensor
axes,
are
since
of
the
rotationally
invariant. (See Section B.5.)

13
Fluids at rest
it is convenient
are
generally in a state of compression,
to write the stress tensor of
a static
so
fluid in
the form
ij
pij
(2.14)
where
ii /3 is termed
the staticfluidpressure,
a function of r and t.
It follows that, in a stationary fluid, the force per unit area
exerted across a plane surface fluid,
element
the with
f orceunit
per normal
unit area
n
is p n.[See Equation (2.3).] Moreover, this normal force
has the same value for all possible orientations of n.This
well-known resultnamely, that the pressure is the same in
all directions at a given point in a static fluidis known as
Pascals law, and is a direct consequence of the fact that a
fluid element
cannot
withstand
shear
stresses,
or,
alternatively, any tendency of applied forces to change its
and is generally
shape.
Stress Tensor in a Moving Fluid
2.6
We have
seen that in a static
the form
ij
= pij
fluid the stress tensor takes
(2.15)
= ii /3 is the
static pressure: i.e., minus the

any direction. Now, the normal
stress at a given point in a moving fluid generally varies
with direction: i.e., the principal stresses are not equal to
one another.
However,
where
normal stress acting in
we can still define

as ( 11 + 22 + 33 )/3
the
mean
principal
stress
ii /3.
stresses are actually

a coordinate frame aligned with the
principal axes), we can also regard ii /3 as the mean
normal stress. It is convenient to define pressure in a
moving fluid as minus the mean normal stress: i.e.,
Moreover, given that the principal
normal stresses (in
p =
13
ii
(2.16)
we can write the stress tensor in a moving fluid as

sum of an isotropic part, p ij which has the same
form as the stress tensor in a static fluid, and a remaining
non-isotropic part, dij which includes any shear stresses,
and also has diagonal components whose sum is zero. In
Thus,
the
other words,
ij
= pij
+ dij
(2.17)
where
dii
=0.
(2.18)
Moreover, since ij and ij
are both symmetric tensors,
it
follows that dij is also symmetric: i.e.,
ji
=dij
(2.19)
It is clear that the so-called deviatoric stress tensor, dij
is a
consequence
of fluid motion, since it is zero in a static
fluid. Suppose, however, that
we were to view a static

a frame of reference
fluid both in its rest frame and in

moving
at
some constant velocity relative to the rest

we would expect the force distribution
fluid to be the same in both frames of
frame. Now,
within the
reference, since the fluid does not accelerate
in either.
appears stationary
therefore zero, whilst
However, in the first frame, the fluid
stress tensor is
and the deviatoric
in the second it has
a spatially
uniform velocity field and
the deviatoric stress tensor is also
same as
deviatoric stress tensor is

and in
zero
(because it is the
in the rest frame). We, thus, conclude that the
moving
zero
both in
fluid possessing
no
stationary
spatial
fluid
velocity
gradients. This suggests that the deviatoric stress tensor is

driven by velocity gradients within the fluid. Moreover, the
tensor must vanish
as these
gradients vanish.
Let the vi (r,t) be the Cartesian components
fluid velocity
velocity
at point
gradients
and time t. The
of the
various
within the fluid then take the form
vi /x
The
simplest
possible
assumption,
which
is
consistent with the above discussion, is that the components

of the deviatoric stress tensor
are
linear functions of these
velocity gradients: i.e.,
dij
vk
Aijkl
xl
14
FLUID MECHANICS
Here, Aijkl is
quotient
fourth-order tensor (this follows from the
rule because
dij
and vi /x
are
both
proper
second-order tensors). Any fluid in which the deviatoric
stress
tensor
takes
the
Newtonian fluid, since

linear
relationship
above
Newton
between
form
was
is
termed
the first to postulate
stresses
shear
and velocity
gradients.
Now, in
there is
an
no
isotropic fluidthat
preferred
is, a fluid in which
directionwe
would
fourth-order tensor Aijkl to be isotropicthat

form in which all physical distinction
directions
most
is absent. As demonstrated
general
expression
for
an
tensor is
Aijkl
=ij
(2.21)
kl
+ ik
jl
+ il
jk
expect
the
is, to have
between different
in Section B.5, the
isotropic
fourth-order
where , , and
functions
are
arbitrary
can
scalars (which
of position and time). Thus, it follows
be
from
(2.20) and (2.21) that

dij
vk
vi
ij
+
x
xk
v j
xi
(2.22)
However, according to Equation (2.19), dij is a symmetric
tensor, which implies that
dij
=ekk
ij
+ 2eij
=, and
(2.23)
where
eij
vi
v j
+
xi
(2.24)
is called the rate of strain tensor. Finally, according
to
Equation (2.18), dij is a traceless tensor, which yields 3
= 2, and
dij
=2
eij
1
ekk
3
ij
(2.25)
where
= .
We, thus, conclude that the most general
expression for the stress tensor in an isotropic Newtonian
fluid is
!
ij
= pij
+ 2
13
eij
ekk
ij
(2.26)
where p(r,t) and (r,t) are arbitrary scalars.
2.7
Viscosity
The significance
previous
of the parameter
expression
appearing
for the stress tensor,
can
in the
be
seen
from the form taken by the relation (2.25) in the special
case
of simple shearing motion. With v1 /x2

as
non-zero velocity derivative, all of the components
are zero apart from the shear stresses
only
dij
d12
=d21 =
(2.27)
Thus,
v1
x2
the
of
is the constant of proportionality between the rate
per unit area when parallel

over one another. This constant
is generally referred to as viscosity. It is
of shear and the tangential force
plane layers of fluid slide

of proportionality
a matter
of experience
fluid undergoing
oppose
the motion, which implies that
The viscosities
and atmospheric
.0
and 1
the
that the force between layers of
relative sliding motion always tends to
10
viscosity
However,
of dry air and
pressure are
about 1.8
10
kg/(m s),respectively. In neither

exhibit
the
> 0.
pure water at
viscosity
much
variation
of air increases
with
20
kg/(ms)
case does
pressure.
by about
0.3
percent, and that of water decreases by about 3 percent,
per degree Centigrade rise in temperature.

Mathematical
Models
ofFluid
Motion
15
2.8
Conservation Laws
Suppose
of some bulk fluid

mass, momentum, energy) at position r and
other words, suppose that, at time t, an
that (r, t) is the density
property (e.g.,
time
t. In
infinitesimal
position
fluid
element
of volume
r,contains an amount
question. Note, incidentally, that
component of
a vector, or even a
amount
The
total
some
fixed volume V is
of
the
dV, located
at
(r, t)dV of the property in
can be either a scalar, a

component of a tensor.
property
contained
within
ZV
dV,
(2.28)
over
where the integral is taken
all elements of V. Let dS
be an outward directed element of the bounding surface of

V. Suppose
and
so out
per
second
across
r. The
the element,
of V, is v(r,t)dS. Thus, the amount of the fluid
property under consideration
per
element
is located at point
that this element
volume of fluid that flows
across
that is convected
the
second is (r, t)v(r,t) dS. It follows that the
net amount of the property that is convected out of volume

V by fluid flow
across
its bounding surface S is
where
the integral
elements
of S
ZS
dS,
(2.29)
is taken
Suppose,
over
all outward
directed
finally, that the property
question is created within the volume V at the rate S

second. The conservation
in
per
equation for the fluid property
takes the form

d
dt
(2.30)
In other words, the rate of increase in the amount of the
property contained within V is the difference between the

creation rate of the property inside V, and the rate at which
the property is convected out of V by fluid flow. The above
conservation law
can also be written

d
dt
(2.31)
Here,
whereas S
is termed
the flux of the property
out of V,
is called the net generation rate of the property
within
V.
2.9
Mass Conservation
given fluid at point

V, surrounded
by
mass density and velocity of a

r and time t.Consider a fixed volume
a surface S The net mass contained
Let (r, t)and v(r,t)be the
within V is
ZV
dV,
(2.32)
where dV is
an element
of V. Furthermore, the
mass
flux
across
S,and out of V,is [see Equation (2.29)]

ZS
dS,
(2.33)
where dS is
an outward
conservation
requires
mass
directed element of S. Mass
that the rate of increase
contained within V, plus the net
should equal
mass
of the
flux out of V,
zero: i.e.,
dM
dt
(2.34)
[cf., Equation (2.31)]. Here,
no mass
generation
individ-ual molecules
(or
are
we are
assuming that there is
destruction)
effectively
within
V (since
indestructible).
It
follows that
ZV
ZS
dV
dS = 0,
(2.43)
16
since V is non-time-varying. Making
FLUID MECHANICS
use
of the divergence
theorem (see Section A.20), the above equation becomes
ZV "
+ (v)
dV
=0.
(2.36)
However, this result is true irrespective
or location
of the size, shape,
of volume V,which is only possible if
(v)
(2.37)
as the
a direct consequence of
throughout the fluid. The above expression is known

equation of fluid continuity, and is
mass conservation.
2.10
Convective Time Derivative
The quantity
(r, t)/t, appearing
in Equation (2.37),
mass density at
r.Suppose that v(r, t)is the instantaneous
at the same point. It follows that the time
the density, as seen in a frame of reference
represents the time derivative of the fluid

the
fixed point
fluid velocity
derivative
of
which is instantaneously
co-moving with the fluid at point
r,is
(r
+ vt,t + t)
lim
t0
(r,t)
+ v
Dt
(2.38)
where
we
+ v t, t + t) up
have Taylor expanded (r
to
first order in t, and where

D
+ v
Dt
+vi
t
xi
(2.39)
Clearly, the so-called
convective
represents the time derivative
time derivative,
seen
D/Dt,
in the local rest frame
of the fluid.
can
The continuity equation (2.37)
be rewritten in the
form
Dln
Dt
since (v)
volume element
general,
as the element
(A.174)].
V that is co-moving
ZS
Dt
v,
(2.40)
v+v[see
with the fluid. In
seen that
ZV
ZS
v dS =
Consider
is convected by the fluid its volume
changes. In fact, it is easily

DV
Dt
vi
dS
ZV
vi
dV
xi
vdV,
(2.41)
where S is the bounding surface of the element, and
has been made of the divergence
use
theorem. In the limit
constant
that V 0, and v is approximately
element,
across
the
we obtain
1
V
DV
DlnV
Dt
Dt
v.
(2.42)
we
Hence,
velocity at
rate
conclude
a given
increase
of
that the divergence
of the fluid
point in space specifies the fractional

in the
volume
an
of
infinitesimal
co-moving fluid element at that point.
2.11
Momentum Conservation
Consider
a fixed volume
V surrounded by a surface S
The
i-component of the total linear momentum contained within
V is
ZV
Pi
Mathematical Models ofFluid Motion
Moreover, the flux of i-momentum
across
is [see Equation (2.29)]
vi
dV.
17
and out of V,
ZS
i
vi
v j dS
(2.44)
Finally, the i-component of the net force acting
on the fluid
within V is
ZV
fi
IS
Fi dV
ij
dS
(2.45)
on the right-hand

the contributions
side
are
respec-
from volume and surface forces,
tively.
Momentum
conservation
requires
increase of the net i-momentum
that the rate of
of the fluid contained
within V, plus the flux of i-momentum out of V,is equal
to the rate of i-momentum
course,
from
quantity
acting
Newtons
generation
is equal to the i-component
on the
fluid
within V. Of
second law of motion, the latter

of the net force
contained within V. Thus,
we
obtain
= fi
[cf., Equation (2.31)]
dPi
dt
(2.46)
which
can be written
ZV
(vi )
dV
ZV
ZS
vi
v j dS
IS
Fi dV
ij
dS
(2.47)
use of the
since the volume V is non-time-varying. Making
tensor divergence theorem, this becomes

#
ZV "
(vi )
dV
ZV
(vi v j )
ij
Fi +
dV.
(2.48)
However, the above result is valid irrespective of the size,
shape,
or location

(vi v j )
(vi )
+
t
=Fi +
ij
(2.49)
everywhere inside the fluid. Expanding the derivatives, and
rearranging,
we obtain
!
+v j
+
x
v j
vi
!
vi
+
t
vi
+v j
=Fi +
j
ij
(2.50)
Now, in tensor notation, the continuity
equation (2.37) is
written
+ v
t
+
x
v j
0.
(2.51)
So, combining
Equations (2.50) and (2.51),
we
!
vi
vi
+v j
obtain the
following fluid equation of motion,
=Fi +
j
ij
(2.52)
An alternative form of this equation is

Dvi
Fi
Dt
ij
(2.53)
how the net volume and
The above equation describes

surface forces
element
per
unit
mass
on a co-moving
acting
determine its acceleration.
18
fluid
FLUIDMECHANICS
2.12
Navier-Stokes Equation
Equations (2.24), (2.26), and (2.53)

give the equation of motion of
classical fluid: i.e.,
Dvi
Dt
=Fi
vi
+
x
Newtonian,
!#
+
xi
an isotropic,
"
can be combined to
v j
23
xi
xi
!
v j
(2.54)
This equation is generally
equation.
strong
known
Now, in situations
temperature
approximation
gradients
as
the Navier-Stokes
in which there
in the fluid, it is
to treat viscosity
as a
spatially
are no
a good
uniform
quantity,
in
which
case
the
"
Dvi
Dt
When expressed
=Fi
xi
in vector
"
Dt
where
v
t
+ (v)v
use has been made
13
vi
"
2v
13
(v)
2v)i
Note, however, that the above identities
aj
are
a surface
ZV
vi
,
.
only valid in
total energy content of the fluid contained within V is
bi
(2.58)
Energy Conservation
volume V surrounded by
(2.56)
of Equation (2.39). Here,
expression
Cartesian coordinates. (See Appendix C.)
a fixed
(2.55)
= F p +
(2.57)
Consider
xi x
[(a)b]i
2.13
form, the above
becomes
Dv
equation
Navier-Stokes
simplifies somewhat to give
The
ZV
EdV
1
vi vi
dV,
(2.59)
are
the net internal
on the
and kinetic
of the fluid. The
energy
flux
across
is [cf., Equation (2.29)]
side
energies,
Here, E(r, t) is the internal (i.e., thermal)
mass
right-hand
respectively.
energy per
S
unit
and out of V,
ZS
E+
vi vi
v j dS
! #
"
ZV
12
E+
vi vi
dV,
(2.60)
use
where
has
been
theorem. According
made
energy
done on
work
the
tensor
divergence
to the first law of thermodynamics,
the rate of increase of the

the net
of
energy
contained within V,plus
flux out of V, is equal to the net rate of
..
the fluid within V, minus the net heat flux
out of V:i.e.,
dE
dt
Q,
(2.61)
where W is the net rate of work, and Q the net heat flux.
It
can
be
generation
seen
that
Q is the effective
energy
rate within V [cf., Equation (2.31)].
Now, the net rate at which volume and surface forces
do work
on the fluid within V is
ZV
ZV
ZS
vi
Fi dV
vi ij
dS
"vi
(vi
ij
Fi +
dV,
x
(2.62)
where
use
has
been
made
of
tensor
the
divergence
theorem.
19
Generally
speaking, heat flow in fluids is driven by
temperature
gradients.
com-ponents
of the heat flux density at position
Let the qi (r, t) be the Cartesian
t.It follows that the heat flux

dS, located at point
rand time
across a surface element
r, is q dS = qi
temperature of the fluid at position
dS
that there is
components
a
us
time t. Thus,
general temperature gradient takes the form T/xi
assume
Let T(r,t)be the
r and
linear relationship
Let
between
the
of the local heat flux density and the local
temperature gradient: i.e.,
qi
Aij
x
(2.63)
a second-rank tensor
(which can be functions of position and time). Now, in an
isotropic fluid we would expect Aij to be an isotropic
where the Aij
are
the components
of
tensor. (See Section B.5.) However,

second-order isotropic
Hence,
the most general
tensor is simply
multiple of ij
we can write
Aij
ij
(2.64)
where (r, t) is termed the thermal conductivity
of the
fluid. It follows that the most general expression for the
heat flux density in an isotropic fluid is
qi
xi
(2.65)
or, equivalently,
T.
(2.66)
Moreover, it is a matter of experience that heat flows down
temperature gradients: i.e.,
> 0. We conclude
heat flux out of volume V is

Q
ZV
ZS
that the net
dS
xi
dV,
xi
xi
(2.67)
where
use
has
been
made
of
the
tensor
divergence
theorem.
Equations (2.59)(2.62) and (2.67)
give the following
energy
can be combined to
conservation equation:
ZV (
!#
"
E+
12
E+
vi vi
! #)
"
vi vi
dV
"vi
ZV
!#
Fi +
vi ij
dV.
(2.68)
However, this result is valid irrespective of the size, shape,
or location

!#
"
E+
everywhere
vi vi
vi ij
(2.69)
the
Expanding
fluid.
some
the
of
we obtain
derivatives, and rearranging,
inside
E+
Fi +
12
vi vi
=vi
! #
"
12
E+
1
vi vi
=vi
Fi +
vi ij
Dt
(2.70)
where
use
has been made
(2.40). Now, the scalar
of the continuity
product
of
equation
with the fluid
equation of motion (2.53) yields

!
Dvi
vi
Dt
12
vi vi
Dt
=vi Fi
ij
+ vi
x
(2.71)
we get
Combining the previous two equations,
DE
Dt
20
FLUID MECHANICS
Finally, making
conservation
use
of (2.26),
equation
for
an
Dt
Here,
=
!
vi
dij
x
1
2 eij eij
eii
3
e jj =
!
v j
vi
xi
vi
v j
3 xi x
is the rate of heat generation
ij
deduce that the
isotropic
takes the general form
DE
we
!
vi
energy
Newtonian
fluid
"
!#
vi
xi
(2.73)
vi
vi
+
x
per
(2.74)
unit volume
due to
viscosity. When written in vector form, Equation (2.73)

becomes
DE
Dt
(T)
(2.75)
energy per
mass of a co-moving fluid element evolves in time as
a consequence of work done on the element by pressure
as its volume changes, viscous heat generation due to
According to the above equation, the internal
unit
flow shear, and heat conduction.
2.14
Equations of Incompressible Fluid
Flow
interest, the flow of a

as water, is incompressible to a
high degree of accuracy. Now, a fluid is said to be
incompressible when the mass density of a co-moving
volume element does not
change appreciably
as the
element moves through regions of varying pressure. In
other words, for an incompressible fluid, the rate of change
of following the motion is zero: i.e.,
In most
situations
ofgeneral
conventional liquid, such
Dt
=0.
(2.76)
Inthis case, the continuity equation (2.40) reduces to
= 0.
(2.77)
that, as a consequence
of mass
an incompressible fluid must have a
divergence-free,
or solenoidal, velocity field. This
We
conclude
conservation,
immediately
volume of
implies,
a co-moving
from Equation
(2.42),
that
the
fluid element is a constant of the
motion. In most practical
situations,
distribution in an incompressible
the initial density
fluid is uniform in space.
Hence, it follows from (2.76) that the density distribution

remains uniform in space and constant in time. In other
words,
we can generally treat
the density, ,
as a uniform
constant inincompressible fluid flow problems.

Suppose
that the volume force acting
on
the fluid is
conservative in nature (see Section A.18): i.e.,
= ,
(2.78)
energy per unit mass, and

energy per unit volume. Assuming
viscosity is a spatially uniform quantity,
where (r, t) is the potential

the potential
that
the fluid
which is generally
the
case
(unless
there
are strong
variations within the fluid), the Navier-Stokes
temperature
equation for an incompressible fluid reduces to
Dv
Dt
2v,
(2.79)
where
(2.80)
is termed the kinematic viscosity, and has units of meters

t
squared
per second. Roughly speaking, momentum
t
diffuses a distance of order
meters in t seconds as a
consequence
water at 20
of viscosity.
C is about 1.0
The kinematic
6
2/s.
10
viscous momentum diffusion in water is
process.
viscosity
relatively slow
21
The complete set of equations governing incompressible
flow is
v
0,
of
It follows that
(2.81)
Dv
Dt
2v.
(2.82)
are
Here, and
(r, t)
as a
equations
components
namely,
of
unknownsnamely,
components
conservation
regarded
known
as
function.
Equation
Equation
and
we
have
four
the
three
Thus,
(2.81),
plus
(2.82)for
pressure,
the
known constants,
four
p(r, t), plus the three
of the velocity, v(r, t). Note that
equation
is
redundant
in the
an energy
case of
incompressible fluid flow.
2.15
Equations of Compressible Fluid Flow
In many situations of general interest, the flow of

is compressible: i.e., there
are
gases
significant changes in the
mass density as the gas flows from place to place. For the
case of compressible flow, the continuity equation (2.40),
and
the
augmented
as
Navier-Stokes
by the
energy
equation
(2.56),
conservation
must
be
equation (2.75),
as thermodynamic relations that specify the

energy per unit mass, and the temperature in
terms of the density and pressure. For an ideal gas, these
well
internal
relations take the form
cV
T,
(2.83)
MR
(2.84)
= 8.3145JK
molar
is the molar specific heat at constant volume, R
where cV
mass
mol
the molar ideal
(i.e., the
mass
and T the temperature

Mthe
mole corresponds
gas constant, M the

gas molecules),
of 1
mole of
in degrees Kelvin. Incidentally,
to 6.0221
10
24
we
is a
molecules. Here,
have assumed,
correspon
for
ds
the
t osake
6.0 221
of simplicity,
10
that cV
constant. It is also convenient to assume that the
uniform
thermal conductivity, , is a uniform constant. Making

of these
approximations,
and (2.84)
can be combined to give

!
Dp
Dt
use
Equations (2.40), (2.75), (2.83),
p D
Dt
= +
(2.85)
where
cp
cV + R
(2.86)
cV
cV
is the ratio of the molar specific heat at constant
c p to that at constant volume, cV

that c p = cV + R for an ideal gas
pressure,
(Incidentally, the result
is a standard theorem of
thermodynamics.) The ratio of specific heats of dry air at
20
C is 1.40.
The complete set of equations governing compressible
ideal gas flow
are
D
Dt
v,
(2.87)
Dv
Dt
#
"
2v
(v)
p D
Dp
(2.88)
Dt
as
known constants,
is specified in terms of
in Equation (2.74). Here,
regarded
+
R
where the dissipation function

and
Dt
(2.89)
, , M, and R
and (r, t)
as a
are
known
function.
Thus,
we
have
equationsnamely,
five
Equations (2.87) and (2.89), plus the three components of

Equation (2.88)for five unknownsnamely,
(r, t),the
pressure,
the density,
p(r, t),and the three components
of
the velocity, v(r,t).
22
FLUID MECHANICS
2.16
Dimensionless Numbers in
Incompressible Flow
It is helpful to normalize
the equations
of incompressible
=
=
fluid Itflow,
(2.81)(2.82),
in th
the following
manner: ofinc
is helpful
to normalize
e equations
L,
v =v/V0
p/(V
2
0
=(V0 /L)t, =
+gL+V0
Here, L is
typical
fluid
/(gL), and p
/L).
typical spatial variation

velocity,
and
g a
length-scale, V0
typical
gravitational
(assuming that represents a gravitational

energy per unit mass). All barred quantities are
dimensionless, and are designed to be comparable with
acceleration
potential
unity.
The normalized equations of incompressible fluid flow take the form
0,
Dv
Dt
1
+
1
Fr
p
Re
where D/Dt
= /t + v,
and
Re
LV0
Fr
V0
(gL)
1/2
(2.90)
2v
Fr
+
Re
(2.91)
(2.92)
(2.93)
Here, the dimensionless
quantities Re and Fr
are
known
as
the Reynolds number and the Froude number, respectively.

The Reynolds
number is the typical ratio of inertial to
viscous forces within the fluid, whereas the
square
of the
Froude number is the typical ratio of inertial to gravitational

forces. Thus, viscosity is relatively important compared to
inertia
when Re
relatively
and vice
important
versa.
1, and vice
compared
versa.
Likewise, gravity is
to inertia when Fr
Note that, in principal, Re and Fr
are
1,
the
only quantities in Equations (2.90) and (2.91) that
greater
significantly
case
For the
of the Earth,
Re
1.0
(2.94)
Fr
(2.95)
Thus, if L
L(m)V0 (ms
3.2
10
C, located
be
on the
surface
1),
V0 (ms
1m and V0
can
than unity.
of water at 20
10
or smaller
1)/[L(m)]
1ms
1,
1/2.
as is often the case
for terrestrial water dynamics, then the above expressions
sug-gest
that Re
viscous
term
on the
1and Fr O(1). In this situation, the
right-hand side of (2.9 1

)becomes negligible,
and the (unnormalized) incompressible fluid flow equations

reduce to the following inviscid, incompressible, fluid flow
equations,
Dt
case
about 100 times
more
1.0
(2.97)
of lubrication oil at 20
surface of the Earth,
0,
(2.96)
Dv
For the
10
C, located
m2 s 1
viscous than water), and so
on the
(i.e., oilis
Re
1.0
(2.98)
Fr
(2.99)
10
L(m)V0 (ms
3.2
that
Suppose
oil
10
is
1),
V0 (ms
slowly
1)]/[L(m)]
flowing
down
m and
lubrication channel such that L 10

1.
1/2.
ms
lubric ati
It on
follows,
cha nne
from
l
suchthat
the above
L expressions,
10
a narrow
V0 10
that
Re10
1
and Fr 1. In this situation, the inertial term
side of (2.91) becomes negligible,
left-hand
on the
and the
(unnormalized) incompressible fluid flow equations reduce
to the following
inertia-free,
incompressible,
fluid flow
equations,
0,
(2.100)
2v.
23
2.17
Dimensionless Numbers in
Compressible Flow
It is helpful to normalize
the equations
of compressible
ideal
It is
gas
help
flow,
ful(2.87)(2.89),
to norm alize theinequations
the following
of co mpressible
manner:
= L, v = v/V0 t = (V0 /L)t, = /0

= (L/V0 ) , and p = (p p0 )/(0 V
2
2
0
= /(gL),
+ 0 g L + 0
V0 /L). Here, Lis a typical spatial variation length-scale, V0
a typical
fluid velocity, 0
typical
gravitational
represents
Furthermore,
gravitational
p0
are
density, and
(assuming
energy per
pressure at
p
pressure. All
are designed to
to atmospheric
atmospheric
from
dimensionless,
and
gas flow
equations
take the form
v,
(2.102)
Dt
1
+
1
Fr
(2.103)
Re
"
unit mass).
Dt
Dv
ga
that
uniform constant. It follows that
with unity. The normalized
compressible ideal
deviations
barred quantities
be comparable
potential
corresponds
ground level, and is
represents
a typical mass
acceleration
Fr
2v
13
Re
(v)
of
"
+p
p0
Dp

Dt
LV0
+ v,
Fr
and
RePr
p0
+p
1/2
1
+ Re(1+ 1/Fr
2)
V0
(gL)
p0
Re
D
Dt
where D/D t /t
(1+ 1/Fr
Pr
+ 1/Re)
V0
p
p0 /0
M
R0
(2.104)
(2.105)
(2.106)
(2.107)
(2.108)
(2.109)
(2.110)
(2.111)
Here, the dimensionless
known
as the Reynolds
numbers Re, Fr, Pr, and M are

number, Froude number, Prandtl
number, and Mach number, respectively.
The Reynolds
number is the typical ratio of inertial to viscous forces

within
the
gas,
the
square
of the Froude
number
the
typical ratio of inertial to gravitational forces, the Prandtl

number the typical ratio of the momentum
and thermal
diffusion rates, and the Mach number the typical ratio of
gas
flow and
sound propagation
speeds. Thus, thermal
. ,
diffusion is far faster than momentum diffusion when Pr
1, and vice
versa.
Moreover, the
1, an dvice
subsonic
when M
ve
rsa 1,Mo
supersonic
reover
gas
p
flow is termed
the
when
gas
M
flow
1,
is and
term
transonic
M O(1). Note that
when
p0 /0
is the
nd
speed
tranof
sonsound
ic when
in the undisturbed
called the thermal
diffusivity
per
gas a
gas. The quantity H

of the gas, and has units
is
of
meters squared
second. Thus, heat typically diffuses
through the
distance
The
thermal
diffusivity
t meters in t seconds.
pressure
and 20
follows
that heat diffusion
process.
C is about
The
1.5
10
at
air
= 2.1
10
in air is
kinematic
dry air at atmospheric
dry
of
viscosity
pressure
atmospheric
m 2 s 1.
relatively
It
slow
of
and 20
C is about
m2 s 1.
Hence, momentum diffusion in air is also
relatively slow
process.
For the
C,
Re
(2.126)
case
6.7
24
of dry air at atmospheric
10
L(m)V0 (ms
pressure
and 20
1),
FLUID MECHANICS
Fr
3.2
10
(2.113) Pr
(2.114)
V0 (ms
1)/[L(m)]
7.2
2.9
10
1/2,
1,
10
1).
V0 (ms
(2.115)
Thus, if L 1
m and V0
for subsonic
air dynamics
the above expressions
1
ms
1,
as is often
close to the Earths
the
case
surface,
suggest that Re 1, M 1, and
Fr,Pr O(1).
(2.105)
that p0
It immediately
quation
(2.104) ( 2.1
is 05)
dominated
brackets
on
its
follows from Equation
1. However, in this situation, Equation
by
left-hand
the
term in
second
square
can
side. Hence, this equation
only be satisfied if the term in question is small, which

implies that
D
1.
Dt
(2.116)
Equation (2.102) then gives
1
(2.117)
Thus, it is evident that subsonic (i.e., M 1) gas flow is

essentially
incompressible.
The fact that Re 1

implies
that
flow
is inc
also
In
flowsuch
is essent
ially
ompressentially
essi ble. Th inviscid.
e fa ct tha t
Rethe
incompressible
1imp lies
1), the
equations
inviscid limit (in which
(unnormalized)
reduce
to
the
compressible
previously
= 0 and Re
ideal
derived
gas
flow
inviscid,
incompressible, fluid flow equations: i.e.,
v
(2.118)
Dv
Dt
0,
(2.119)
It follows that the equations which
govern
subsonic
gas
are
dynamics close to the surface of the Earth

the
same as those
which
Suppose that L
typically the
over
case
the wing of
govern
essentially
the flow of water.
1m and V0
1,
300ms
as is
for transonic air dynamics (e.g., air flow
fighter jet). In this situation, Equations
(2.105) and
Re,Frn
M,Pr,
w (2.112)(2.115)yield
ing of a fight er j et).I
thi 1
sand
situat
ion p0
the
O(1).
It
follows
right-hand
that
the
two
final
on
terms
sides
can be neglected. Thus,

gas flow equations
set of inviscid, adiabatic, ideal gas,
of Equations (2.103) and (2.104)
the (unnormalized)
compressible
reduce to the following
ideal
flow equations,
D
v,
Dt
(2.120)
Dv
Dt
(2.121)
Dt
0.
(2.122)
In particular, if the initial distribution of p/
is uniform
space, as is often the case, then Equation (2.122)

ensures that the distribution remains uniform as time
progresses. In fact, it can be shown that the entropy per
unit mass of an ideal gas is
in
!
S
cV
ln
M
(2.123)
space is
equivalent to the assumption that the entropy per unit mass
of the gas is a spatial constant. A gas in which this is the
case is termed homentropic. Equation (2.122) ensures that
the entropy of a co-moving gas element is a constant of
the motion in transonic flow. A gas in which this is the
case is termed isentropic. In the homentropic case, the
above compressible gas flow equations simplify somewhat
Hence, the assumption
that p/
is uniform in
to give
D
Dt
(2.124)
Dv
v,
Dt
p
(2.125)
p0
,
.
25
Here, p0 is atmospheric
pressure,
air at atmosphericpressure.
the adiabatic
and 0 is the density of
Equation (2.126)is knownas
gas law, and is a consequence
of the fact that
transonic
thermal
gas
heat
process)
to
takes place far too quickly
dynamics
conduction
have
any
temperature distribution
(which
is
appreciable
within the
gas.
relatively
effect
on
Incidentally,
for
slow
the
a gas
in which thermal diffusion is negligible is generally termed

adiabatic.
2.18
Fluid Equations in Cartesian
Coordinates
Let
us adopt
the conventional Cartesian coordinate system,

to Equation (2.26), the various compo-
x,y,z.According
nents of the stress tensor
are
xx
yy
zz
xy
xz
yz
+ 2
vx
x
p
+ 2
vy
y
,
,
=yx
= zx
=zy
=
=
=
=
=
=
(2.127)
(2.128)
+ 2
vz
,
,
,
z !
vy
vx
+
x
vx
vz
+
z
x !
vy
vz
+
z
where
v is the velocity, p the pressure,
and
(2.129)
(2.130)
(2.131)
(2.132)
the
viscosity. The equations of compressible fluid flow,

(2.87)(2.89)
fluid flow
(from which the equations of incompressible
can easily
be obtained by setting
=0),become
Dt
Dvx
Dt
Dvy
Dt
Dvz
Dt
!
p D
Dt
vx +
vy +
(2.134)
3 y
(2.135)
1 p
,
,
,
Dt
(2.133)
vz +
3
(2.136)
where is the
(2.137)
mass
the
density,
the heat conductivity,
ratio of specific heats,
mass,
M the molar
and R the molar
ideal gas constant. Furthermore,

vy
vx
+
x
(2.138)
(2.139)
+ vy
+vz
+ vx
t
Dt
vz
z 2
,
,
,
(2.140)

!2
!2
vy
vx
2
x
!2
12
vz
vy
vx
y
26
FLUID MECHANICS
!2
!2
12
vx
+
z
In the above,
constants.
2.19
12
vz
Fluid
vy
!2
vz
(2.141)
, and
Equations
are
treated
in
as
uniform
Cylindrical
Coordinates
us
Let
adopt the cylindrical
Making
use
of the results
coordinate
system,
quoted in Section
r, , z.
C.3, the
components
are
of the stress tensor
rr
zz
rz
+ 2
vr
= r
= zr
= z
=
=
=
=
=
=
(2.142)
+ 2
1r
vr
+ 2
vz
(2.143)
(2.144)
1r
vr
vr
1r
vz
vz
(2.145)
(2.146)
+
z
(2.147)
whereas the equations of compressible fluid flow become
Dt
2
Dvr
Dt
vr
Dv
+
Dt
r
Dvz
Dt
!
Dt
p D
Dt
where
(2.148)
1 p
r
2
1 p
r2
1r
vr
vr
r2
3 r
(2.149)
2
v
1 p
r2
vz +
r2
vr
3r
13
(2.150)
(2.151)
!
M
(2.152)
1r (rvr )
1r
vz
(2.153)
Dt
(2.154)
+ vr
+
r
1r
+vz
z 2
(2. 169)
27
!2
1r
vr
2
r
!2
!2
vz
1 1r
vr
+
z
!2
12
vr
vz
+
z
+
r
1
2
1r
+
z
!2
!2
vr
vz
(2.156)
2.20
Fluid
Equations
in
Spherical
Coordinates
Let
us, finally,
, . Making
adopt the spherical coordinate
use
system,
r,
of the results quoted in Section C.4,
the
components of the stress tensor
are
rr
+ 2
+ 2
vr
r
1r
(2.157)
vr
= r
= r
=
=
=
=
=
=
=
(2.158)
+ 2
1r
r sin
1
r sin
1r
(2.159)
(2.160)
(2.161)
cotv
cotv
vr
rsin
vr
vr
(2.162)
whereas the equations of compressible fluid flow become
Dt
2
Dt
vr
Dv
cotv
vr
+
Dt
+ cotv
r
Dt
Dv
v +v
Dvr
Dt
p D
Dt
where
(2.163)
1 p
2vr
vr
r2
r2
2cotv
r2
1
r2
1r
r2
sin
2 cot
2
r2
(2.165)
r2
+
3r
r sin
r sin
(2.164)
2 vr
v
sin
sin
(2.166)
v
r2
sin
r2
sin
vr
2
2 cot
r2
sin
3r sin
(2.167)
!
M
(2.168)
1 (r
vr )
rsin
(sinv )
rsin
(2.169)
28
FLUID MECHANICS
+ vr
t
Dt
+
r
r sin
(2.170)
!
r2
+
r
r2
sin
sin
r2
sin
,
,
(2.171)

vr
!2
r
!2
1r
12
1r
vr
!2
vr
!2
rsin
vr
12
r
12
rsin
1
rsin
r
+
!2
cotv
!2
vr
1r
cotv
r
(2.172)
2.21
Exercises
2.1. Equations (2.66), (2.75), and (2.87)

give the following
energy
compressible fluid:
can be combined to
conservation equation for a non-ideal
DE
where is the
energy per
unit
Dt
= q,
Dt
mass
mass,
density,
the viscous
p the pressure,
energy
E the internal
dissipation rate
per
unit
volume, and
q the heat flux density.

D
We also have
v,
T,
Dt
q
where
v is the fluid velocity,
T the temperature, and
the
thermal conductivity. Now, according to a standard theorem in

thermodynamics,
T dS
= dE
d,
per unit mass.
where S is the entropy
entropy flux density at
a given
Moreover, the
point in the fluid is
s = Sv +
T
where the first term on the right-hand side is due to direct
entropy convection by the fluid, and the second is the entropy flux
density associated with heat conduction.
Derive
an entropy
conservation equation of the form

dS
+S
dt
=S
where S isthe net amount of entropy contained insome

fixed volume V, S the entropy flux out of V,and S the net rate
of
entropy creation within V.Give expressions for S,S

and S
volume
Demonstrate that the entropy creation rate per unit
is
+
T
qq
T
Finally, show that 0, in accordance with the second

law of thermodynamics.
2.2. The Navier-Stokes equation for an incompressible fluid of

uniform
mass
density takes the form
Dv
Dt
where
potential
2v,
v is the fluid velocity, p the pressure,
energy per unit mass, and the (uniform)
viscosity.
the
kinematic
The incompressibility constraint requires that
v = 0.
Finally, the quantity

29
is generally referred to as the fluid vorticity.

Derive the following vorticity evolution equation from the
Navier-Stokes equation:
D
Dt
= ( )v +
2.3. Consider two-dimensional incompressible fluid flow. Let

the velocity field take the form
v =vx (x,y,t)ex + vy (x,y,t)ey
Demonstrate that the equations of incompressible fluid

flow (see Exercise 2.2) can be satisfied by writing
vx
vy
,
,
where
= 2 /x 2
+
t
= 2 ,
and
= 2 .
Here,
is termed
a stream
/y
function, since
2.
Furthermore, the quantity
= 0:i.e., the fluid flow is
everywhere parallel to contours of .
2.4. Consider incompressible irrotational flow: i.e., flow that

satisfies
Dv
Dt
0,
2v,
as well as
v = 0.
v is the fluid velocity, the uniform mass density,
p the pressure, the potential energy per unit mass, and the
Here,
(uniform) kinematic viscosity.
Demonstrate that the above equations

writing
can be satisfied
by
v = ,
where
= 0,
and
+
t
1 2
v +
2
+ = C(t).
Here, C(t) is a spatial constant. This type of flow is known
as potential
a scalar
flow, since the velocity field is derived from
potential.
2.5. The equations of inviscid adiabatic ideal gas flow
are
v,
Dt
Dv
Dt
D
Dt
Here, is the mass density,
0.
v the flow
velocity,
p the
pressure,
the potential
energy per unit mass, and the
(uniform)
ratio of specific heats. Suppose that the pressure and
potential
energy are both time independent:
i.e., p/t
= /t =
0.
Demonstrate that
12
+
1
is a constant of the motion: i.e., DH/Dt

known
as Bernoullis
= 0.This result is
theorem.
30
FLUID MECHANICS
2.6. The equations of inviscid adiabatic non-ideal
D
gas flow are
v,
Dt
Dv
Dt
DE
Dt
Here,
pressure,
2 Dt
0.
is the mass
the potential
density, v the flow velocity, p the

energy per unit mass, and E the
internal
energy per unit mass. Suppose that the pressure

energy are both time independent: i.e., p/t
ernal
potential
=0.
Demonstrate that
and
=/t
1 2
v + E+
2
= 0.This result is
is a constant of the motion: i.e., DH/Dt
a more
general form of Bernoullis theorem.
2.7. Demonstrate that Bernoullis theorem for incompressible,

inviscid fluid flow takes the form DH/Dt
12
= 0,where
+.
31
Hydrostatics
3 Hydrostatics
3.1
Introduction
This chapter discusses hydrostatics, which is the study of

the mechanical equilibrium of incompressible fluids.
3.2
Hydrostatic Pressure
Consider
a body
of water that is stationary
in a reference
frame that is fixed with respect to the Earths surface. In
this
a frame is
z measure vertical
chapter, such
inertial. Let
the region
as
treated
approximately
height, and
suppose that
z>0
0 is occupied by water, and the region
by air. According
to Equation (2.79), the air/water system
remains in mechanical equilibrium (i.e.,
= Dv/Dt = 0)
provided
= p
(3.1)
where
p is the
static fluid
= g z the gravitational
g
the (approximately
pressure, the mass density,

energy per unit mass, and
potential
uniform) acceleration due to gravity.
Now,
(
(z)
z >0
z 0
(3.2)
where 0
is the (approximately
water. Here, the comparatively

has
been neglected.
Since
mass density
mass density of
uniform)
of
small
air
= (z)
immediately follows, from (3.1), that
and
= (z),
it
p = p(z), where
dp
g.
dz
(3.3)
We conclude that constant
pressure
surfaces in a stationary
body of water take the form of horizontal planes. Making
use of (3.2), the above
equation
p(z)
z>0
p0
p0
can be integrated to give
gz
z0
(3.4)
p0
where
10
m2
is atmospheric
According to this expression,

increases
linearly
with
pressure
pressure
at ground
increasing
depth
decreasing
z < 0).liIn fact, given that
wa tez, for
rincreases
and
10
kgm
level.
water
in stationary
(i.e.,
with
2
9.8 ms
3,
we
< 0).
deduce
In fact,
that
giv hydrostatic
en that g 9.8m
pressure
s rises at the rate of 1
2)
5
atmosphere (i.e., 10 Nm
every 10.2m increase indepth
below the surface.
3.3
Buoyancy
the air/water
Consider
section. Let V be
surface
partially
system described
some
that straddles
occupied
the plane
or
air) contained
by
= 0, and
by water, and partially
i-component of the net force acting
water
in the previous
volume, bounded
closed
is thus
by air. The
on the fluid
(i.e. either
within V is written (see Section
2.3)
ZV
ZS
fi
ij
dS
Fi dV,
(3.5)
where
ij
pij
(3.6)
is the stress tensor for a static fluid (see Section 2.5), and
gez
3.15) 32
FLUID MECHANICS
the gravitational
force density. (Recall that the indices 1,
2, and 3 refer to the

fz
f3
x-,y-, and z-axes,
respectively. Thus,
etc.) The first term on the right-handside of (3.5)
represents the net surface force acting
across
whereas
the second term represents the net volume force distributed

throughout
theorem
V. Making
use
of the
tensor
divergence
(see Section B.4), Equations (3.5)(3.7) yield the
following expression for the net force:
= B+W,
(3.8)
where
ZV
Bi
p
dV,
xi
(3.9)
and
Wx
Wy
=0,
ZV
(3.10)
Wz
gdV.
(3.11)
Here, Bis the net surface force, and W the net volume
force.
It follows from Equations (3.4) and (3.9) that
=M0 gez
(3.12)
where M0
=0
V0
Here, V0 is volume of that part of V
which lies below the waterline, and M0 the total mass of
water contained within V.Moreover, from Equations (3.2),

(3.10), and (3.11),
= M0
gez
(3.13)
It
can
seen
be
vertically
that the net surface force, B, is directed
upward, and exactly balances
the net volume
force, W, which is directed vertically downward. Of
course,
W is the weight of the water contained within V. On the

other hand, B,which is generally known
force, is the resultant
surrounding
pressure
as the
buoyancy
of the water immediately
V. We conclude that, in equilibrium, the net
buoyancy
force
across
acting
S exactly
balances
the
weight of the water inside V, so that the total force acting
on the contents
of V is
system inmechanical
zero, as must
be the
the line of action of B (which is vertical)

the center of gravity of
net torque would act

contradict
case
for
equilibrium. We can also deduce that
passes
through
the water inside V. Otherwise,
on the contents
our assumption
of V, which would
that the system is in mechanical
equilibrium.
3.4
Equilibrium of Floating Bodies
Consider the situation described in the previous section.

Suppose that the fluid contained within V is replaced by
partially
submerged
solid
body
corresponds to S. Furthermore,
whose
suppose
outer
surface
that this body is
in mechanical equilibrium with the surrounding fluid (i.e.,

it is stationary, and floating
It follows that the
fluid is unchanged
can
pressure
on the
[since the force balance criterion (3.3)
be integrated to give the
all contiguous
remains
surface of the water).
distribution in the surrounding
pressure
distribution (3.4) at
points in the fluid, provided that the fluid
in mechanical equilibrium]
net surface force acting
across
this is directly related to the

fluid immediately
surrounding
We conclude that the
S is also unchanged (since
pressure
distribution
in the
V), which implies that the
buoyancy
force acting
on the
that acting
previously
on the floating
body is the
same as
water: i.e., the water that
displaced
occupied V.Inother words, from (3.12),
ez
W0
(3.14)
where W0
= M0
g and M0 are
displaced water, respectively.
the weight and
mass
of the
The fact that the buoyancy
force is unchanged also implies that the vertical line of
action of B passes through the center of gravity, H(say), of

the displaced water. Incidentally, H is generally known
as
the center of buoyancy.

A floating body of weight W is acted
upon
by two
forces: namely, its own weight,
= W ez
(3.21)
Hydrostatics
33
and the buoyancy force, B
the surrounding water. Of
passes
= W0 ez due to the pressure of

course, the line of action of W
through the bodys center of gravity, G (say). Now,
zero
zero net
to remain in equilibrium, the body must be subject to

net force and
zero net torque.
force yields W0
The requirement of
= W. In other words, in equilibrium,
weight of the water displaced by

the weight of the body,
a floating
or, alternatively,
magnitude of the buoyancy force acting
the
body is equal to
in equilibrium, the
on a floating
body
is equal to the weight of the displaced water. This famous

result
is
as Archimedes
zero net torque
known
requirement
of
equilibrium,
center
the
principle.
implies
in
and
center
on the same
vertical
of gravity,
of buoyancy, H, of a floating body lie
The
that,
G,
straight-line.
Consider
Let
floating body of
M/V
Archimedes
be
the
bodys
mass M and volume V.

mean mass density.
principle implies that, inequilibrium,

V0
V
s,
(3.16)
where
(3.17)
is termed the bodys specific gravity. (Recall, that V0 is

the submerged
water.)
of
volume,
and 0
the
mass
density
of
We conclude, from (3.16), that the volume fraction
a floating
body that is submerged is equal to the bodys
gravity. Obviously, the specific gravity must be
specific
less than unity, since the submerged
volume
fraction
cannot exceed unity. Infact, if the specific gravity exceeds

unity
then it is impossible
balance
sinks.
the bodys
for the buoyancy
force
to
weight, and the body consequently
Consider
and
It
follows,
the
body
from
submerged
volume
of
gravity
specific
that
Equation
volume
floats
(3. 16),
is V0
in equilibrium.
that
V.
Hence, the volume above the waterline is V1

(1 s) V. Suppose that the
= V V0 =
body is inverted such that its
previously
(1 s)V.
submerged
S uppo se
part
tha t
isthe
raised above the waterline,
versa:
and vice
can
body
i.e., V0
V1
According
to (3.16), the
only remain in equilibrium in this configuration
if its specific gravity changes to
s =
V1
V V0
= 1 s.
(3.18)
We conclude that for every equilibrium configuration of
equilibrium configuration
having the complementary
3.5
s there
an inverted
for a body of the same shape
specific gravity 1
s.
floating body of specific gravity
exists
Vertical Stability of Floating Bodies
Consider
floating
equilibrium, has
downward
force, B
body
a submerged
of
weight
volume V0
which,
in
Thus, the bodys
weight is balanced by the upward buoyancy
=0 V0 g:i.e., 0 V0 g = W.
Let A0 be the
area
cross-sectional
the plane
mean
mean
draft (or
submerged
to define the bodys
depth)
that the body is displaced
Suppose
as
slightly
= V0 /A0
downward,
without rotation, such that its mean draft becomes 0 +1

where |1
the
|
0
vicinity
Assuming that the cross-sectional
of the
= A0 (0
new
W + 0 A0
buoyancy
g1
However,
the
constant,
+1 )
force becomes
(1
weight
Thus, the bodys
is
waterline
submerged volume is V 0
and the
area in
new
the
= V0
= 0
+A0
V0
g=
+ 1 /0 )W.
of
the
body
is
unchanged.
perturbed vertical equation of motion is
written
Wg
dt 2
=W B
W0
1
(3.19)
which reduces to the simple harmonic equation
d
dt 2
g
1
0
.
,
,
of the body at the waterline (i.e., in
= 0). It is convenient
(3.20)
We conclude that when
a floating
body of mean draft 0 is
subject to a small vertical displacement it oscillates about
its equilibrium position at the characteristic frequency
g
0
any
34
FLUID MECHANICS
STABLE
UNSTABLE
Figure 3.1: Stable and unstable configurations for a
floating body.
It follows that such

small
vertical
presupposes
analysis
stable to
body is unconditionally
displacements.
that
course,
Of
the
oscillations
the
above
take
place
sufficiently slowly that the water immediately surrounding
the
body
always
remains
in approximate
hydrostatic
equilibrium.
3.6
Let
Angular Stability of Floating Bodies
us now
angular
only
shall
perpendicular
such
planes
the stability of floating bodies to
investigate
displacements.
consider
planes
body is in
are
For the sake of simplicity,
bodies
that
have
of symmetry.
Suppose
an equilibrium state
= 0 plane
we
mutually
that when
the two symmetry
vertical, and correspond to the
planes. As before, the
two
x = 0 and y = 0
coincides
with the
surface of the water. It follows, from symmetry, that when
the body is in an equilibrium
state its center of gravity, G,
and center of buoyancy, H,both lie on the z-axis.

Suppose
that the body turns through
about
some
goes
through
small angle
the origin.
vertical, straight-line that
= 0, that
horizontal axis, lying in the plane

Let
GH be that, originally
passes
through the bodys center
of gravity, G, and original center of buoyancy, H. Owing
to the altered shape of the volume of displaced water, the

.
center of buoyancy is shifted to some new position H
Let the vertical straight-line
going through H
meet GH
at M. See Figure 3.1. Point M (in the limit 0) is

called the metacenter.
In the disturbed state, the bodys
weight W acts downward through G, and the buoyancy 0

V0
g acts
upward
through
submerged volume, V0
state
(which
M. Let
us assume
that the
is unchanged from the equilibrium
excludes
vertical
oscillations
from
.,
consideration). It follows that the weight and the buoyancy

force
are equal
as can
weight and the buoyancy

magnitude
so that there is no net force on

seen from Figure 3 1 the
force generate a net torque of
and opposite,
the body. However,
be
=W sin. Here, is the length MG: i.e., the
distance between the metacenter and the center of gravity.

This distance is generally known
as
the metacentric height,
and is defined such that it is positive when M lies above G,

and vice
versa.
Moreover,
as is also
clear from Figure 3.1,
when M lies above G the torque acts to reduce , and vice
versa. Inthe former case, the torque

a floating
torque. We conclude that

angular
displacements
horizontal
axis
0 provided
lying
that
its
Since
we
some
z
about
in the plane
metacentric
i.e., provided that the metacenter

gravity.
is known
have
as a righting
body is stable to small
height
is
positive:
lies above the center of
already
floating body is unconditionally
demonstrated
that
stable to small vertical
displacements (and since it is also fairly obvious that such
a body
displacements
is neutrally stable to both horizontal
and angular displacements
about
vertical axis passing
a necessary
a floating
up of arbitrary
through its center of gravity), it follows that

and sufficient
condition
for the stability

body
to a general small perturbation (made
linear and angular
components)
of
is that its metacentric
35
Hydrostatics
horizontal axis.
3.7
Determination of Metacentric Height
Suppose
that
the
floating
body
considered
in the
an equilibrium state. Let A0 be

cross-sectional area at the waterline: i.e., in the plane z
previous section is in
the
= 0. Since
respect to the
the body is assumed
to be symmetric
x =0 and y =0 planes, we have
with
ZA0
ZA0
xdxdy
ZA0
ydxdy
=0,
xydxdy
(3.22)
where the integrals
at z
are
taken
over
the whole cross-section
= 0. Let (x,y) be the bodys draft: i.e., the vertical
distance between the surface of the water and the bodys
lower boundary. It follows, from symmetry, that (x, y)
= lower
(x, y)
b and (x, y) = (x, y). Moreover,
the submerged
volume is
ZA0
V0
ZA0
dxdydz
(x,y)dxdy.
(3.23)
It also follows from symmetry that
ZA0
ZA0
x(x,y)dxdy
=0.
y(x,y)dxdy
(3.24)
Now, the depth of the unperturbed
center of buoyancy
below the surface of the water is

RA0 R
ZA0
zdxdydz
V0
2V0
2(x,y)dxdy
2
0
A0
2V0
(3.25)
where
RA0
1/2
2(x,y)dxdy
A0
(3.26)
Finally,
from
buoyancy lies at
symmetry,
the
of
x =y =0.
that the body
Suppose
center
unperturbed
now turns
through
small
angle about the x-axis. As is easily demonstrated, the

(x, y)
(x, y) y. Hence,
bodys new draft becomes
the
new
submerged volume is
ZA0
V0
ZA0
[(x,y) y]dxdy
= V0
ydxdy
=V0
(3.27)
where
use
has been made of Equations (3.22) and (3.23).
Thus, the submerged

the
case
for
a purely
volume is unchanged,
as
angular displacement. The
should be
new
of the center of buoyancy is

RA0 R
zdxdydz
V0
depth
ZA0
2(x,y)
2y(x,y)+
O(
)]dxdy
2V0
(3.28)
where
use has been made
of (3.24)and (3.25). Thus,
the depth of the center of buoyancy

Moreover,
=h,
is also unchanged.
from symmetry, it is clear that the center of
buoyancy still lies at
x = 0. Finally,
the
new
y-coordinate
of the center of buoyancy is

RA0 R
RA0
ydxdydz
y[(x,y)
y]dxdy
=
V0
x A0
V0
V0
(3.29)
where
use has been made
of (3.24). Here,
RA0
y
x
1/2
dxdy
=
A0
36
(3.38)
FLUID MECHANICS
is the radius of gyration of area A0 about the x-axis.
It follows, from the above analysis, that if the floating
body under consideration
about
the
x-axis
horizontally
perpendicular
then
turns through
its
center
of
small angle
buoyancy
shifts
x
A0 /V0
in the plane
to the axis of rotation. In other words, the
distance
distance HH inFigure3.1 is x A0 /V0
Simple
trigonometry reveals that HH
small). Hence, MH
= HH
the height of the metacenter
/MH
2
=x
relative
. .
(assuming that
Sim
pl
is
A0 /V0
Now, MH
to the
a depth
below the surface of the water (which corresponds
z = 0). Hence,
= x2
zM
A0 /V0
z-coordinates
the z-coordinate
h. Finally, if zG
of the unperturbed
is
center of
buoyancy. However, the center of buoyancy lies
plane
to the
of the metacenter is
and zH
= h
are
the
centers of gravity and
buoyancy, respectively, then

2
zM
A0
+ zH
V0
(3.31)
and the metacentric height,
2
A0
zGH
=zG
zH
V0
(3.32)
where zGH
=zM
zG
becomes
Note that, since x A0 /V0
> 0,the
metacenter always lies above the center of buoyancy.

A simple extension of the above argument reveals that
if the body turns through
a small
then the metacentric height is
angle about the y-axis
2
y
A0
zGH
V0
(3.33)
where
RA0
y
x2
1/2
dxdy
=
A0
(3.34)
is radiusofgyrationofarea A0 aboutthey-axis. Finally,as
is easily demonstrated,if thebodyrotatesabouta
horizontal axis which subtends
an angle
with the x-axis
then
A0
zGH
V0
(3.35)
where
2
2
x
cos 2
2
y
sin
Thus, the minimum value of

It
follows
uncondiamplitude
that
the
tionally
angular
is the lesser of x and y
equilibrium
stable provided
displacements
state
in question
(3.36)
is
it is stable to small
about
horizontal
normal to its two vertical symmetry planes (i.e., the
axes
x =0
y =0 planes).
an example, consider a uniform rectangular block
of specific gravity s floating such that its sides of length a,
b, and c are parallel to the x-, y-, and z-axes, respectively.
Such a block can be thought of as a very crude model of a
ship. The volume of the block is V = a b c. Hence, the
submerged volume is V0 = s V = s a b c. The
cross-sectional area of the block at the waterline (z = 0) is
A0 = a b. It is easily demonstrated that (x, y) = 0 =
V0 /A0 = sc.Thus, the center of buoyancy lies a depth h=
and
As
2
0
A0 /2V0
sc/2 below the surface of the water [see Equation (3.25)].

Moreover, by symmetry, the center of gravity is
c/2 above the bottom surface
located
zH
a depth sc
below the surface of the water. Hence,
= h = s c/2, zG = c/2 s c, and zGH = c (1 s)/2.
Consider
the stability
of the block
to small amplitude
angular displacements about the x-axis. We have

R
2
a height
of the block, which is
a/2
a/2
b/2
b/2
y dxdy
ab
12
(3.37)
Hence, from (3.32), the metacentric height is
(1 s).
12 sc
2
(3.46) Hydrostatics
37
The stability criterion > 0 yields
b
>
s(1 s)
6c 2
0.
(3.39)
Since the maximum value that
that the block
follows
s (1
s)
can
take is 1/4, it
is stable for all specific
gravities
when
c < c0 =
23
b.
(3.40)
On the other hand, if c > c0 then the block is unstable for

intermediate specific gravities such that s
where
q1
c 02 /c 2
<s <s+
s =
2
(3.41)
and is stable otherwise. Assuming that the block is stable,

its angular equation of motion is written
d
I
= W sin W ,
dt 2
(3.42)
where
a/2
b/2
csc
?b
2
(y
gV
a/2
+ 4[(1 s) 3 + s 3]c 2
b/2
+ z 2)dxdydz
sc
(3.43) is the
12g
moment of inertia of the block about the x-axis. Thus, we

obtain the the simple harmonic equation
dt
= 2 ,
(3.44)
where
c 02
c 02 + (8/3)[(1
sc
4 s(1 s)c
s) 3
+ s 3 ]c 2
(3.45)
We conclude
angular
executes
that the block
oscillations
about
the
small amplitude
at the angular
x-axis
. However, this result is only accurate inthe
frequency
are
limit in which the oscillations

the water
surrounding
approximate
the
hydrostatic
sufficiently
block
always
equilibrium.
slow that
remains
For the
case
in
of
rotation about the y-axis, the above analysis is unchanged
except that
b.
The metacentric
height of
a conventional
ship whose
length greatly exceeds its width is typically much less for

rolling
(i.e., rotation
about
horizontal
axis running
along the ships length) than for pitching (i.e., rotation

about
horizontal axis perpendicular to the ships length),
since the radius of gyration for pitching greatly exceeds

that for
with
rolling. As is clear from Equation (3.45),
a relatively
relatively
excessively
of
flooded.
the
ship
versa.
Now,
an
height increases the chances
if the weather
is rough,
or if it is damaged
For this reason, maritime regulatory
shifts,
International
height (for rolling) has
long roll period, and vice

low metacentric
ship capsizing
cargo/ballast
as
small metacentric
Maritime
or
if its
and partially
agencies, such
Organization,
specify
minimum metacentric heights for various different types of
sea-going vessel. A relatively large metacentric height,

the other hand, generally renders
a ship
passengers
the ship executes
and
crew,
because
uncomfortable
period rolls, resulting in large g-forces. Such
on
for
short
forces also
cargo may break loose or shift.

saw earlier, in Section 3.4, that if a body of specific
gravity s floats in vertical equilibrium in a certain position
then a body of the same shape, but of specific gravity 1
s, can float in vertical equilibrium in the inverted position.
We can now demonstrate that these positions are either
position.
both stable, or both unstable, provided the body is of
uniform density. Let V1 and V2 be the volumes that are
increase the risk that
We
above and below the waterline,

position. Let H1 and H2 be the
respectively,
mean centers
in the first
of these two
volumes, and H that of the whole volume. It follows that

H2 is the center of buoyancy in the first position, H1 the
center of buoyancy in the second (inverted) position, and

Hthe center of gravity
inboth positions. Moreover,

!
V1 H1G
=V2
H2 G
V1 V2
V1
+ V2
H1 H2
(3.51) 38
FLUID MECHANICS
where H1G is the distance between points H1 and G, etc.

The metacentric
are
heights in the first and second positions
"A
H1G
V1
V1
V1 + V2
(3.47)
2
H2 G
V2
respectively,
gyration
of the
common
V1 V2
are
area
the
waterline
V1 + V2
=
#
V2
where A and
H1H2
"A
V1 V2
H1H2
(3.48)
and radius of
section, respectively.
Thus,
"A
1 2
which implies that 1
V1 V2
V1 V2
T 0
V1
as
+ V2
3.8
H1H2
T 0, and vice
are
or both unstable.
follows that the first and second positions
stable, both marginally stable,
#2
0,
(3.49)
It
versa.
either both
Energy of a Floating Body
The conditions governing the equilibrium and stability of
a floating
energy.
body
can also
be deduced from the principle of
For the sake of simplicity,
water surface
area
is infinite,
let
so that
us suppose
that the
the immersion of the
body
does not generate
energy
potential
bodys
any
change in the water level. The
of the body itself is W zG
where W is the
weight, and zG the height of its center of gravity, G,
relative to the surface of the water. If the body displaces

volume
of water then this effectively
V0
means
that
a
a
weight 0 V0 of water, whose center of gravity is located
,
.
at the center of buoyancy, H,is removed, and then spread
as an
infinitely
This involves
over
thin film
the surface of the water.
energy
gain of potential
V0 zH
of 0
where zH is the height of Hrelative to the surface of the
water. Now, vertical force balance requires that W
energy
Thus, the potential
anarbitrary
additive
= 0
V0
of the system is W zGH (modulo
constant), where zGH
= zG
zH is the
height of the center of gravity relative to the center of

buoyancy.
an equilibrium
or a maximum of
According to the principles of statics,
state corresponds
the potential
to either
energy.
that
minimum
However, such
stable when the potential

follows
stable
floating body is such
energy
equilibrium
as to
an equilibrium
is minimized.
is only
Thus, it
configuration
of
minimize the height of the
bodys center of gravity relative to its center of buoyancy.
3.9
Curve of Buoyancy
Consider
a floating
body in vertical force balance that is
a horizontal
slowly rotated about
symmetry
vertical
planes.
Let
gravity, G, which necessarily

origin
axis normal to one of its
us
take
the center
of
as
the
lies in this plane,
of
coordinate system that is fixed with respect to the body.

As
illustrated
3.2,
in Figure
as
the
body
rotates,
the locus of its center of buoyancy, H, as
seen in the fixed

reference frame, appears to traces out a curve, AB, in the
plane of symmetry. This curve is known as the curve
of buoyancy. Let r represent the radial distance from the
origin, G, to some point, H, on the curve of buoyancy.
Note that the tangent to the curve of buoyancy is always
orientated horizontally. This follows because,as was shown
inthe previous section, small rotations of a floating body in
vertical force balance cause its center of buoyancy to shift
horizontally,
rather
perpendicular
in vertical height, zGH

the center
distance,
buoyancy
than
vertically,
in
the
plane
to the axis of rotation. Thus, the difference

between the center of gravity and
of buoyancy
p, between
is equal to the
perpendicular
G and the tangent to the
at H. An equilibrium configuration
curve
of
therefore
a maximum or a minimum of p as point

curve of buoyancy. However, the
equilibrium is only stable if p is minimized. Now, if R is
the radius of curvature of the curve of buoyancy then,
according to a standard result in differential calculus,
corresponds
H
moves
to
along the
dr
dp
(3.50)
Writing this result inthe form
r
(3.54) Hydrostatics
39
Rr
p,
r
p
H0
H
Figure 3.2: Curve of buoyancy for a floating body.
it can be seen that maxima and minima of p, which
are the
points
on the curve
= 0, correspond
= 0,and are, thus, coincident with
of buoyancy where p
to the points where r
r.In other words, an equilibrium

to a point of maximum or
minimum r on the curve of buoyancy: i.e., a point at
which GH meets the curve at right-angles. At such a point,
r = p, and the potential energy consequently takes the
value W r.
Let H0 be a point on the curve of buoyancy, and let r0
p0 and R0 be the corresponding values of r,p, and R. For
neighboring points on the curve, we can write
maxima and minima of
corresponds
configuration
?????H0
r r0 =
dr
(p p0 ),
dp
(3.52)
or
r0
R0
(p
p0 ).
r0
(3.53)
It follows that
R0 and r0
p0 has the
are both positive).
> 0)
(i.e., dr/dp
same
sign
as r
r0 (since
[The fact that R0 is positive
follows from the previously
established
result that the metacenter, which is the center of curvature

of the
curve
buoyancy,
of buoyancy, always lies above the center of
implying
that
the
curve
of
buoyancy
is
concave upwards.] Hence, the minima and

r occur simultaneously with those of p.
Consequently,
a stable
equilibrium
configuration
corresponds to a point of minimum r on the curve of
buoyancy: i.e., a minimum in the distance GH between
necessarily
maxima
of
the center of gravity and the center of buoyancy.
We
can use
equilibrium
the above result to determine
configurations
cross-section,
for
and uniform specific
beam
gravity
the stable
square
s, that floats
of
with its length horizontal. In order to achieve this goal,
we must
calculate
configurations
balance. However,
1/2,
since,
ofSection
with
distance
that
GH for all possible
are
in vertical
cases
need only consider
to the
3.6, forevery
configuration with
us
we
according
s = s0 < 1/2
Let
the
of the beam
s<
analysis
stableequilibriumconfiguration
there is
s = 1
force
where
s0
corresponding
stable inverted
> 1/2, and vice versa.
define fixed rectangular
axes, x and y, passing
through the center of the middle section of the beam, and

running parallel to its sides. Let
us start with the case

a side. See Figure
where the waterline PQ is parallel to
3.3. If the length of
a side is 2a then (3.16) yields

AP
BQ
(3
40
= 2a
s.
62)
FLUIDMECHANICS
Oy
x
Q
PP
Figure
3.3:
of square
Beam
cross-section floating with two corners immersed.

Suppose that the beam is turned through
> 0 such
that the waterline
assumes
an angle
the position P Q
in Figure 3.3, but still intersects two opposite sides. The
lengths AP and BQ
satisfy
BQ
AP
2a
tan.
(3.55)
must
match that of the rectangle PABQ in order to
ensure
Moreover,
area
the
that the submerged
of the trapezium
P ABQ
volume remains invariant (otherwise,
the beam would not remain invertical force balance): i.e.,
2
(AP
+ BQ ) a
4 a
s.
(3.56)
It follows that
AP
(3.57) BQ
=
=
a(2
s tan),
a(2 s + tan).
(3.58)
The constraint
that the waterline
sides of the beam implies that AP
intersect
two opposite
> 0,and, hence, that

tan <
s.
(3.59)
The coordinates of the center of buoyancy, H, which is the
mean center of the trapezium P ABQ , are
R
a R a
a
ah(x)
a
a
xdxdy
a
ah(x)
dxdy
(2/3)a
4a
3
2
tan
a
6s
tan,
(3.60)
R
a R a
a
ah(x)
ydxdy
a R a
a
ah(x)
dxdy
=
4a
s(1 s) (1/3)a
4a
tan
= (1 s)a
a
12 s
tan
(3.61)
where
h(x)
= 2 a s + x tan .
(3.69) Hydrostatics
41
Oy
Figure 3.4: Beam of square

cross-section floating with one corner immersed.
=r
Thus, if u
/a
=( x
where
+y 2)/a 2
then
"(1
36 s 2
s)
#2
12 s
(3.63)
tan .
to
corresponds
Now,
minimum
stable
of
equilibrium
with respect
state
to , and,
hence, of u with respect to t.However,

du
d
(3.64)
dt
2u
dt 2
=
=
ht
36 s
1 h3t
36 s 2
,
.
i
2
12 s(1 s) + 2
i
2
12 s(1 s) + 2
(3.65)
The minima and maxima of u occur when du/dt

0,
2u/dt 2
2u/dt 2
d
> 0 and du/dt
0, d
< 0, respectively. It
follows
that the symmetrical position, t
= 0, in which the
are either parallel or perpendicular to the

is always an equilibrium, but is only stable when
sides of the beam

waterline,
s2
s+
16
>
(3.66)
12
i.e., when
s < 1/2
1/
= 0.2113. It is also possible
to
obtain
i.e., whequilibria in asymmetric positions such that t is the
root of
t
12s(1 s) 2.
(3.67)
Such equilibria
only exist for
s > 0.2113, and are
Finally, in order to satisfy the constraint
have
<
s,
which, in combination
(3.59),
stable.
we must
with the above
equation, implies that
s2
s +
1> 0,
(3.68)
or s < 0.25.
Suppose that the constraint (3.59) is not satisfied,
that the immersed portion of the beams cross-section
so
is
triangular. See Figure 3.4. It is clear that

BQ
BP
tan.
42
frame.
FLUID MECHANICS
Moreover, the
area
of the triangle P BQ
in Figure 3.4,
must match that of the rectangle PABQ, in Figure 3.3, in

order
to
ensure
that
the
submerged
volume
remain
invariant: i.e.,
12
BP
BQ
4a
s.
(3.70)
It follows that
BP
(3.71)
BQ
=
=
(8 s/tan)
(8 s tan)
1/2
1/2
a,
a,
(3.72)
or, writing z 2
=tan and =(8/9) s,

2
BP
(3.73) BQ
=
=
3z
a,
3za.
(3.74)
The coordinates of the center of buoyancy, H,which is the
mean center
of triangle P BQ
x
are
a BP /3
= a(1 z 1),
a BQ /3
(3.75) y
= a(1
z),
(3.76)
since the perpendicular
triangle
from
perpendicular
Thus, if u
one
of
mean center
one third of
distance of the
of
its sides
the
is
distance from the side to the opposite vertex.

2
2)/a 2
r 2 /a 2
( x +y
then
1) 2
u
=
(1 z
+ (1 z) 2
du
(3.77)
(z
1)(z
2u
dz
z + 1)
z + 3)
z3
dz
(3.78)
(z
,
,
.
(3.79)
Moreover, the constraint (3.59) yields
z >
32
.
(3.80)
The stable and unstable equilibria correspond to du/dz

2u/dz 2
0, d
> 0 and du/dz, d 2u/dz 2 < 0, respectively. It
z = 1, in which the
or perpendicular
provided < 2/3, or s <
follows that the symmetrical position,
diagonals of the beam
are
either parallel
to the waterline is an equilibrium

1/2, but is only stable when
1/2,
or s > 9/32
= 0.28125.
>
It is also possible to obtain
equilibria inasymmetric positions such that z is the root of
z2
z +
= 0.
(3.81)
2/3
Such equilibria only exist for
< < 1/2, or 1/4 < s
< 9/32, and are stable.

In summary, the stable equilibrium configurations
of a
square cross-section, floating with its length horizontal, are such that the sides are either parallel or
perpendicular to the waterline for s < 0.2113, such that two
corners are immersed but the sides and diagonals are
neither parallel nor perpendicular to the waterline for 0.21
13 < s < 0.25, such that only one corner is immersed but
the sides and diagonals
are neither parallel nor
perpendicular to the wa-terline for 0.25 < s < 0.28125,
and such that the diagonals are either parallel or
perpendicular to waterline for 0.28125 < s < 0.5. For s
> 0.5, the stable configurations are the same as those for
a beam with the complimentary specific gravity 1 s.
beam of
3.10
Rotational Hydrostatics
Consider the equilibrium
is uniformly
rotating
at
of
an incompressible
a fixed
angular
fluid that
velocity
in
some inertial frame of reference. Of course, such a fluid

appears stationary in a non-inertial co-rotating reference
(3.91) Hydrostatics
43
Moreover, according to standard Newtonian dynamics, the

force balance equation for the fluid in the co-rotating frame
takes the form (cf., Section 3.2)
= p
r),
(3.82)

energy per unit mass, and r
a position vector (measured with respect to an origin that
lies on the axis of rotation). The final term on the
where
p is the
static fluid
the gravitational potential
right-hand
side
of
the
represents
above equation
fictitious
centrifugal force
generality,
we can assume
density.
Without
=ez
0 = p
that
the
loss
of
It follows that
),
(3.83)
where
12
(x
+y
2)
(3.84)
is the so-called centrifugal potential. Recall, incidentally,
that is a uniform constant in an incompressible fluid.
As
of
an example,
water,
uniformly
located
rotating
consider the equilibrium
on
about
the
Earths
vertical
angular velocity . It is convenient
of
surface,
body
that
is
axis at the fixed
to adopt cylindrical
z,whose symmetry axis

z increase upward.
= ( 1/2) 2 r 2 Assuming
It follows that = g z and
that the pressure distribution is axisymmetric, so that p =
coordinates (see Section C.3), r,,
coincides with the axis of rotation. Let
p(r,z), the force balance equation, (3.83), reduces to
+
r
(3.85)
0,
0,
(3.86)
or
p
0,
+g
0.
r
(3.87)
p
z
(3.88)
The previous two equations
p(r,z)
(3.89)
=p0 +
12
gz
p0
where
a constant.
is
pressure
constant
Thus,
surfaces in a uniformly rotating body of water take the
paraboloids of revolution about the rotation axis.
form of
Suppose that p0
case,
pressure.
represents atmospheric
In this
= p0 :
pressure
the surface of the water is the locus of p(r, z)
pressure
i.e., it is the constant
surface
whose
matches that of the atmosphere. It follows that the surface

of the water is the paraboloid of revolution
z=
2g
(3.90)
where
is the perpendicular
rotation, and
z =0 the on-axis
distance
from the axis of
height of the surface.
Now, from Section 3.3, it is plain that the buoyancy

force
wholly
on any
acting
or
co-rotating
that which would maintain the

the body
mass
waters curved
represent
respectively,
simply
mass
is limited
and
of
a floating
of the
surface through the body. Let points G and
the
centers
of the body.
the center
rH
case
by the continuation
gravity
of
Of
of gravity
course,
from
the
axis
and
buoyancy,
the latter point is
of the displaced
Suppose that G and Hare located

rG
is
same as
of water displaced by
in relative equilibrium. In the
body, this
solid body, which
partially immersed in the water, is the
of
water.
perpendicular distances
rotation,
respectively.
Finally, let Mbe the
mass
of the body, and M0 the
mass
of the displaced water. It follows that the buoyancy force
has
an upward
horizontal
vertical component
component
M0
M0
rH
g,
an
and
outward
Thus, according
and an out
to
standard Newtonian dynamics, the equation of
horizontal
motion of a general co-rotating body is

2
M( r rG )
44
= M0
rH
FLUID MECHANICS
where
= d/dt. Now, from Archimedes principle, M0
= M for the case of a floating body that is less dense
than water. However, if the body is of uniform density

then rH
> rG as a consequence
waters surface. Hence,
we obtain
r
of the curvature of the
(rH
rG )
<
0.
(3.92)
In other
words,
floating
body
drifts
radially
towards the rotation axis. On the other hand, M0

fully
submerged
.
.
body
that is
more
inward
< M for a
dense than water.
However, if the body is of uniform density then its centers
of gravity
rH
= rG
and buoyancy coincide with
Hence,
r = (M M0 ) 2
we obtain
rG
> 0.
one
another,
so that
(3.93)
In other words,
outward
fully submerged
the
from
rotation
body drifts radially
axis.
The
above
analysis
common observation that objects heavier

than water, such as grains of sand, tend to collect on the
outer side of a bend in a fast flowing river, whilst floating
objects, such as sticks, tend to collect on the inner side.
accounts for the
Equilibrium of a Rotating Liquid Body
3.11
Consider
a self-gravitating
is rotating
uniformly
through its center of
liquid body in outer
about
some
mass. What
fixed
space
axis
that
passing
is the shape of the bodys
bounding surface? This famous theoretical problem had its

origins in investigations of the figure of
as
such
the Earth, that
Maclaurin,
and other celebrated
18th, and 19th centuries.

the
a rotating
undertaken
planet,
by Newton,
Jacobi, Meyer, Liouville, Dirichlet, Dedekind,
Riemann,
treat
were
Earth
calculation,
as a
scientists,
in the 17th,
Incidentally, it is reasonable to
liquid, for the
purpose
of this
because the shear strength of the solid rock
out of which the terrestrial crust is composed is nowhere
near
sufficient
deviate
Earth.
to allow the actual shape of the Earth to
significantly
from that of
hypothetical
liquid
Now, in a co-rotating reference frame, the shape of

self-gravitating, rotating, liquid planet is determined by
competition
between fluid
pressure,
gravity,
and
opposes
fictitious centrifugal force. The latter force
a
a
the
gravity
to the axis of rotation.
in the
plane perpendicular
course,
in the absence of rotation, the planet would be
we
spherical. Thus,
would expect rotation to
planet to expand in the plane perpendicular

axis, and to contract
conserve
cause
For the
whose outer
along the rotation axis (in order to
sake of simplicity,
to
rotating
we
planet
shall
our
restrict
of uniform
density
,,
boundary is ellipsoidal. Now,
an ellipsoid
is
an
ellipse. Let
us
the three-dimensional
generalization
of
adopt the right-handed Cartesian coordinate system x1
the
to the rotation
volume).
investigation
x3
Of
An ellipse whose principal
axes are
x2
aligned along the
x1 and x2 -axes satisfies
x 12
a 12
x 22
a 22
1,
(3.94)
where a1 and a2
Moreover,
are
as is easily
the corresponding
demonstrated,
Z
dA
=a1
a2
principal radii.
(3.95)
Z
x i2
14
dA
a i2
A,
x1 x2 dA
(3.96)
0,
(3.97)
where A is the area, dA an element of A,and the integrals
are taken over the whole interior of the ellipse. Likewise,
an ellipsoid whose principal axes are aligned along the x1
x2
and x3 -axes satisfies
-,
-,
x 12
a 12
x 22
a 22
x 32
a 32
= 1,
(3.98)
1
See Ellipsoidal Figures of Equilibrium, S.Chandrasekhar
(Yale
University Press, New Haven CT, 1969).

Hydrostatics
,,
where a1 a2
Moreover,
and a3
as is easily
45
are
the corresponding
principal radii.
demonstrated,
Z
dV
43
a1 a2 a3
x i2
(3.99)
a i2
V,
x1 x2 dV
(3.100)
dV
x2 x3 dV
= 0,
(3.101)
where V is the volume, dV
integrals
are taken over
an
element
of V, and the
the whole interior of the ellipsoid.
Suppose that the planet is rotating uniformly about the
x3 -axis
at the fixed angular
moment
of inertia about this axis is [cf., Equation (3.100)]
. The planets
velocity
I33
15
M(a 1
+ a 22 ),
(3.102)
where
M is
its
mass.
Thus,
the
angular
planets
momentum is
=I33
1
5
M(a 1
+ a 22 ),
(3.103)
and its rotational kinetic
energy
becomes
12
I33
10
M(a 1
+ a 22 ) 2
(3.104)
to Equations
According
pressure
(3.83) and (3.84), the fluid
distribution within the planet takes the form
"
12
p = p 0
(x 1 + x 2 )
(3.105)
where
is
gravitational
the planet,
and
p 0
Appendix
the
gravitational
potential
energy
a unit test
= M/V the uniform
a constant.
E that
homogeneous
of
the
potential
However,
planetary
the
mass) due to
mass
density,
it is demonstrated
gravitational
self-gravitating
(i.e.,
potential
ellipsoidal
in
inside
body
can
be
written

34
GM
2
i
i=1,3
(3.106)
where G is the gravitational constant, and
du
(3.107)
du
(3.108)
2
(a 1
+ u)
1/2
(a
2
2
+ u)
1/2
,
,
2
(a i
+ u)
2
(a 3
+ u) 1/2.
(3.109)
Thus, we obtain
"
p = p0
!

12
G M 1
2
x1
+
3
G M 2
x 22 +
32
G M 3
x 32
(3.110)
where p0 is the central fluid
pressure.
the planets outer boundary must be

would be
words,
a force
imbalance
across
Now, the
zero,
pressure at
otherwise there
the boundary. In other
we require
"
!
2
x1 +
46
32
G M 2
x2 +
G M 1
3
G M 3
#
2
x3
=p0
FLUID MECHANICS
whenever
2
x2
x 2x 1
2
a1
x2
a 22
x3
a 32
1.
(3.112)
two equations
The previous
can
only be simultaneously
satisfied if
a1 =
(3/2)G M
(3/2)G M
a 2 =3 a 3
(3.113)
we obtain
Rearranging the above expression,
a1 a3
a2
2G
(a 2
a3
udu
)
0
(a
2
2
+ u)(a 32 + u)
Z
2
(a 1
a2
a 12 a 22
)
0
(a
+ u)(a 22 + u)
1
2
a 32
(a
2
3
+ u)
(3.114)
subject to the constraint
du
0,
(3.115)
where
use has been made
Finally, according
potential
energy
of Equation (3.99).
to Section E, the net gravitational
of the planet is
G M
2
0
10
(3.116)
energy
Hence, the bodys total mechanical
= K +U=
10
becomes
M(a 1 + a 2 )
3
GM
2
0
10
(3.117)
3.12
Maclaurin Spheroids
One,
fairly
a1
In other
symmetric
way
obvious,
in which the constraint
(3.115)
words,
can
if the
be satisfied is if a2
planet
is rotationally
about its axis of rotation. Now,
that is rotationally
symmetric
about
an
principal
ellipsoid
axisor,
an ellipsoid with two equal principal

as a spheroid. In fact, if a2 = a1 then the
cross-section of the planets outer boundary in any plane
passing though the x3 -axis is an ellipse of major radius a1
equivalently,
radiiis known
in the direction perpendicular

radius a3
are
inthe direction parallel to the x3 -axis. Here,
assuming
flattened
to the x3 -axis, and minor
that a1
along
>
its axis
a3 :
i.e., that
of rotation.
the planet
The
degree
we
is
of
flattening is conveniently measured by the eccentricity,
e13
(1
2 1/2
a 32 /a 1
)
(3.118)
Thus, if e13
planet
= 0 then
is consequently
no
there is
spherical,
flattening, and the
if e13
whereas
then the flattening is complete, and the planet consequently

collapses to a disk inthe x1-x2 plane.
Let
u = a 12
and
Equation (3.114),
=e
2
13
(1
2G
= a1 in
1. Setting a2
we obtain
/z
1/2
e 13 )
e 13
(1+ )
(1
(1+
z 2 ) 1/2
e 13 )1/2
(1
"Z
3
e 13
dz
2
e 13
) 3/2
2(1
=
2
e13
z2
e 13 )
e13
dz
(1 z 2) 3/2
(3.119)
Performing the integrals, which
2G
3 2e
3
e 13
2
13
are standard,
we find that
(1
e 13 )
1/2
sin
3
e13
2
e 13
(1
e 13 ).
(3.120)
2
See Schaums Mathematical Handbook of Formulas and Tables,
Edition, Murray R.Spiegel, (Mc-Graw Hill, New York NY, 1998).
Hydrostatics
47
e13
e13
2nd
bL
bL
b
E
b
b
b
b
bL
bL
0.00000
0.00000
0.60000
0.05
0.02582
0.01266
0.59980
0.10
0.05168
0.02540
0.59919
0.15
0.07758
0.03830
0.59817
0.20
0.10357
0.05144
0.59672
0.25
0.12967
0.06491
0.59479
0.30
0.15591
0.07882
0.59236
0.35
0.18231
0.09329
0.58936
0.40
0.20889
0.10846
0.58572
0.45
0.23567
0.12450
0.58135
0.50
0.26267
0.57612
0.55
0.28989
0.14163
bL
0.16013
0.60
0.31729
0.18037
0.56233
0.00
0.56986
0.65
0.34484
0.20286
0.55320
0.70
0.37239
0.22834
0.54200
0.75
0.39967
0.25792
0.52800
0.80
0.42612
0.29345
0.51001
0.85
0.45046
0.33833
0.48587
0.90
0.46932
0.39994
0.45107
0.95
0.47045
0.50074
0.39272
0.96
0.46472
0.53194
0.37485
0.97
0.45418
0.57123
0.35273
0.98
0.43475
0.62486
0.32351
0.99
0.39389
0.71209
0.27916
Table 3.1: Properties of the Maclaurin spheroids.
This famous result
was
first obtained by Colin Maclaurin
in 1742. Finally, in order to calculate the potential

(3.116),
we need to evaluate
energy,
a1
Z
e13
a1 e13
dz
(1
d
(1+ )(1
1/2
2
+ e 13
)1/2
sin
a1
e13
e13
(3.121)
= sin .
Let e13
Thus,
= /2
rotational flattening, and
to
0 corresponds
no
to complete flattening.
2/3,
= a0 (cos)
and a3 = a0 (cos)
where
= (3 V/4) is the mean radius. It is
1/2,
b L=
also helpful
to define
= /(2G)
1/2,
2/a0
L/(G M a0 )
and E=
E/(G M to def).
ineb =
Moreover,
a1
= (a1 a2
a3 )
1/3
1/3
a0
1/3
bL
/(2G
The above
) analysis
leads to the following bset of equations
which specify the properties
of the so-called
Maclaurin
spheroids:
cos
sin
"(1
+ 2 cos 2
3 cos
)
sin
(3.122)
6
(cos)
1/6
#1/2
+ 2 cos 2
sin
3 cos
)
sin
(3.123)
(cos)
1/3
10
sin
b
E
"(1
=
#
4 cos
bL
"(1
)
sin
,
,
.
+ 3 cos
(3.124)
These properties
are set out
inTable 3.1.
In the limit, 0, in which the planet is relatively
slowly
rotating
1), andthe
its planet
degree isof
flattening
In the
limit, (i.e.,

0,
in which
relativel
consequently slight, Equations (3.122)(3.124)reduce
to
e13
15
2
6
bL
(3.125)
35
b
E
(3.126)
(3.127)
In other words, in the limit of relatively
when
the
becomes
this
case,
planet
is almost
directly proportional
it is
more
spherical,
slow rotation,
its eccentricity
to its angular velocity. In
conventional to parameterize
angular
velocity in terms of
tum, bL,
48
The
FLUID MECHANICS
a0
g0
32
b
0.2
0.1
0
0
0.10.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
e13
Figure 3.5: Normalized angular velocity squared of a

Maclaurin spheroid (solid) and
(dashed)
where g0
versus
a Jacobi
ellipsoid
the eccentricity e13 inthe x1-x3 plane.
= G M 2/a
2
0
is the
mean
surface gravitational
acceleration.
flattening is
Furthermore,
more
the
ellipticity,
case
= 6.37 10
m, g0
of the Earth (b
1),
= 9.81ms
Thus, it follows
that,
figure would
be
were
rotational
of
2
e 13
a1 a3
Thus, it follows from (3.125) that
Now, for the
degree
expressed in terms of the
conveniently
= 7.27
a0
2
(3.129)
54
m.
(3.130)
10
rad.s
1,
a0
we obtain
291
flattened
ellipticity
its
at the poles, of
54
(3.131)
the Earth homogeneous,
spheroid,
291
233
(3.132)
This result
was
first obtained by Newton. Now, the actual
ellipticity
of
substantially
discrepancy
the
Earth
smaller
is about
than
1 /294, which
Newtons
prediction.
is
The
is due to the fact that the Earth is strongly
inhomogeneous, being much denser at its
core
than in its
outer regions.
Figures
3.5 and
3.6
illustrate
, ,
the variation
of
the
normalized
angular
velocity,
, and
momentum,
Figures
3.5 and
3.6 illust
rateangular
the variatio
nof th
bL,normaliz
Fig
3.5
3 .6
illustra
tegu
the
v ariatio
no
of a Maclaurin
edures
angular
spheroid
veand
locity,
with
,
its
andeccentricity,
an
lar
moe13
mentu
as
bL, normabylized
the
an gular
ve loc ity,b
predicted
Equations
(3.122)(3.124).
from
Figure 3.5, that there is
normalized
angular
velocity
(3.137) Hydrostatics
49
lar
Itand
canangu
be seen,
limit to how large the
of such
spheroid
can
1.2
1.1
1
0.9
0.8
0.7
0.6
L
0.5
0.4
0.3
0.2
0.1
0
0
0.10.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
e13
Figure 3.6: Normalized angularmomentum of a Maclaurin
spheroid (solid) and a Jacobi ellipsoid (dashed)versus the
eccentricity e13 inthe x1-x3 plane.
to
b
limiting value corresponds
0.47399, and
occurs
=
limit
corres
=
0.47399
,an
docc
there
urswhene1
are
two
3 ingvalue
=possible
0.929 95.
Maclaurin
For pondsto
v alue sofb
spheroids,
lyin
gb
one
elow0.4739
with
an
urswh
= 0.92
5. For va
luesofb
ly ing
b elow0.47
eccentricity
9there ene 13less
than99
0.92995,
and
one with
an eccentricity
b
when e13
lim iting
0.92995.
valuecoFor
rrespondst
values ofboblying
0.47
below
399 ,and
0.47399
occ
399there
greater than 0.92995. Note, however, from Figure 3.6, that
,,
despite thethan0
fact that
angular
velocity,
, of a Maclaurin
.92the
9 95.
Note,
howev er,from
Figu re3
varies
than
actthatth
non-monotonic
0.92995.
Note, ho
manner
weoc
ver,ity,b
with
from
the
.6,thatdespi
spheroid
teth in
e fa
e angularvel
, Fig
ofa
bL,
re
eccentricity,
3.6
,th
atdespit
e13
the
ethe
angular
factt
hat
momentum,
the
angul
arvel
increases
oci
ty,b
Maclaurinsph eroid
bL, e13
becoming infinite in the limit
monotonically
ofa Maclaurinspheroid
with e13
1. It follows
that there is
angular momentum of
3.13
a1 (i.e., if the outer boundary
body
is ellipsoidal,
constraint
Z
(3.115)

rather
can only
limit to the
(a
then
the
be satisfied when
+ u)(a 22 + u)
1
2
of the rotating
than spheroidal)
2
a1
a 22
Without loss of generality,
a3
no upper
Maclaurin spheroid.
Jacobi Ellipsoids
If a2
a 32
(a
2
3
du
+ u)
=0.
(3.133)
we can assume
that a1 a2
Let
a2
(3.134) a3
a1 cos,
a1
cos,
(3.135)
where . It follows
that the cross-sections
planets where
outer
boundary
. It in the x1 -x2 and x1-x3
of the
planes
are
ellipses of eccentricities
=
=
e12
(3.136)
e13
(1
a 32 /a 22 )1/2
(1
2
2 1/2
/a 1 )
3
=sin,
= sin,
50
FLUID MECHANICS
respectively. It is also helpful to define
=sin 1(sin/sin).
(3.138)
Let sin
corresponds
= [a /(a +u)] sin . Here, u= 0

to = , and u = to = 0. Equations
(3.115) and (3.114) transform to

"
E(,)
2F(,)+
1
+ (sin tan cos)
cos 2
E(,)
sin tan cos(1
cos
+ sin 2
0,
(3.139)
and
2
"
=2
F(,) E(,)
+
tan sin tan
tan 3 cos 2
cos 2
cosE(,)
tan 2 cos 2
(3.140)
respectively, where
Z
E(,)
(1 sin
sin
1/2
d,
(3.141)
Z
F(,)
=
(1 sin
(3.142)
are special
integrals.
sin
2(cos cos)
) 1/2
functions known
The integral 0
to
as incomplete
elliptic
defined in(3.107), transforms
1/3
F(,).
a0 sin
(3.143)
Finally, making
use of some
of the analysis in the
previous two sections, the normalized angular
momentum, and normalized mechanical
energy,
of the
can be written
6
10
2/3
(cos cos)
3 (cos cos)
1/3
F(,)
b
E
(3.144)
1
+ cos
bL
1
+ cos
planet
20 (cos cos)
sin
2/3
(3.145)
respectively.
Now, the constraint
(3.139) is obviously
the limit 0, since this implies

E(, ), F(, ) . Of
course,
sec-tion.
first
researcher
ellipsoidal
solutions
are
in
to obtain
surprising result that (3
solutions
Jacobi,
known
1834,
the
se,
discussed
t his
was the
in the
very
139) also has non-axisymmetric
characterized
as the
by
>
0. These
Jacobi ellipsoids in his honor.
The properties of the Jacobi ellipsoids,

Equations
in
this limit corresponds to
the axisymmetric
F(, Maclaurin
)
O
spheroids
f cour
previous
satisfied
that 0 and
as determined from
are set
(3.139), (3.140), (3.144), and (3.145),
out in Table 3.2, and illustrated in Figures 3.5 and 3.6.

It can be seen that the
sequence of Jacobi ellipsoids
bifurcates from the
of Maclaurin spheroids when e13
= 0.81267.
sequence
Moreover,
However,
e12
eccentricity,
plane
are no Jacobi ellipsoids

as e13 increases above
there
grows
with
<
e13
in the x1-x2
of the Jacobi ellipsoids
rapidly, approaching
0.81267.
this critical value, the
unity
as
e13 approaches
1, in which a Maclaurin
a disk in the x1-x2 plane, a Jacobi
to a line running along the x1-axis.
unity. Thus, in the limit e13

spheroid
collapses to
ellipsoid
sph eroid
collapses
Note, from
Figures 3.5 and 3.6, that, at fixed e13
the
Jacobi ellipsoids have lower angular velocity and angular
momentum than Maclaurin spheroids (with the

and volume). Furthermore,
as is the case
spheroid, there is a maximum
for
same mass
Maclaurin
a Jacobi
can have (i.e., = 0.43257), but no maximum
angular
maximum
momentum.
ang ular ve loci ty that a Jac obi ellip so id c
maxi
mum
anthe
gular
velocit
y that
am
Jaangu
cobi
an have
Figure
(i.e 3.7
=shows
0.4 325
7), but
mechanical
no maximu
energy
of laellip
the
r
soid
canm.
have
(i.e., =
0.4 3257)
no maximum
momentu
Maclaurin
spheroids
and
Jacobi but
ellipsoids
plottedangu
as laa
mom
entum.
function
of their angular momentum. It can be seen that
the Jacobi ellipsoid with a given angular momentum has a
lower
energy
that
.,
b
ellipsoid
angular velocity
See On Jacobis Figure of Equilibrium for
a Rotating
Mass of Fluid,
G.H. Darwin, Proc. Roy. Soc. London 41, 319 (1886).

4
See Handbook of Mathematical
Functions, M. Abramowitz, and I.A.
Stegun (Dover, New York NY, 1965).
Hydrostatics
51
e12
e13
e12
bL
bL
bL
b
E
e13
b
bL
0.81267
0.43257
0.30375
0.50452
0.05
0.81293
0.43257
0.30375
0.50459
0.10
0.81372
0.43257
0.30375
0.50459
0.15
0.81504
0.43256
0.30377
0.50458
0.20
0.81691
0.43253
0.30380
0.50457
0.25
0.81934
0.43248
0.30388
0.50453
0.30
0.82237
0.43237
0.30402
0.50445
0.35
0.82603
0.43220
0.30427
0.50432
0.40
0.83037
0.43191
0.30468
0.50410
0.45
0.83544
0.43146
0.30532
0.50376
0.50
0.84131
0.43078
0.50326
0.55
0.84808
0.42976
0.30628
bL
0.30772
0.60
0.85585
0.42827
0.30984
0.50138
0.65
0.86480
0.42609
0.31296
0.49975
0.70
0.87510
0.42288
0.31760
0.49734
0.75
0.88705
0.41807
0.32462
0.49372
0.80
0.90102
0.41069
0.33562
0.48814
0.00
0.50250
0.85
0.91761
0.39879
0.35390
0.47908
0.90
0.93778
0.37787
0.38783
0.46295
0.95
0.96340
0.33353
0.46860
0.42782
0.96
0.96950
0.31776
0.50078
0.41499
0.97
0.97605
0.29691
0.54672
0.39771
0.98
0.98317
0.26722
0.62003
0.37241
0.99
0.99101
0.21809
0.76872
0.32842
Table 3.2: Properties of the Jacobi ellipsoids.
that
the
corresponding
Maclaurin
same
the
amountof dissipation
an isolated
fluid system to slowly evolve
volume).
This is significant
generally expect
toward
the equilibrium
subject to
case
(i.e.,
mass, and
because, in the presence of a
(i.e., viscosity), we would
with the
small
spheroid
angular momentum,
spheroid
of
any
global
state with the lowest
energy,
on the system.
For the
constraints
weakly viscous, isolated, rotating, liquid planet,
the relevant constraints

angular momentum
change. Thus,
are
that the
we expect
such
volume, and net
planet to evolve toward the
equilibrium state with the lowest

volume,
mass,
of the system cannot spontaneously
and angular
energy
momentum.
for
This
a given mass,
suggests,
from
Figure 3.7, that at relatively high angular momentum (i.e.,

b L> 0.30375, e13
> 0.81267),
the Jacobi ellipsoid
when
bL
solutions
exist, they
there is
evolve),
are stable equilibrium states (since

no lower energy state to which the system can
whereas the Maclaurin spheroids are unstable.
On the other hand, at relatively low angular
(i.e.,
b L< 0.30375, e13
bL
Jacobi ellipsoid
< 0.81267),
solutions,
momentum
are no
are
is no lower
when there
the Maclaurin
spheroids
stable equilibrium states (again, because there
energy state to which they can evolve). These predictions

are borne by the results of direct stability analysis
performed
on the Maclaurin spheroids and Jacobi
ellipsoids.
In fact, such stability studies demonstrate that
the Maclaurin spheroids

weak dissipation
unstable for e13
the other
for e13
>
are
hand,
are
0.93858,
but
0.93858,
evolving
equilibria (which
are
unstable in the
> 0.81267,
presence
of
and unconditionally
0.95289. The Jacobi ellipsoids,
on
for e13
<
>
unconditionally
unconditionally
toward
lower
are, themselves,
stable
unstable
energy
for
e13
pear
shaped
unstable in the
presence
of weak dissipation).
3.14
Roche Ellipsoids
Consider
a homogeneous
liquid
moon
mass M which
a planet of mass
of
is in a circular orbit of radius R about
Let C, C
and C
moon-planet
system,
respectively.
demonstrated, all three points lie on the
and the distances between

CC
= R and
CC
/(M
according to standard Newtonian
the
the
As
is
easily
same
straight-line,
values
M )] R. Moreover,
dynamics, there exists
inertial frame of reference in which C

the line CC
of
them take the constant
= [M
moon,
mass of
be the center of the
center of the planet, and the center
an
is stationary, and
rotates at the fixed angular velocity , where

G(M + M )
R3
(3.146)
moon and the

common center of
In other words, in the inertial frame, the

planet orbit in a fixed plane about their
mass at the angular
velocity . It is convenient to transform
to a non-inertial reference frame that rotates (with respect

,
to the inertial frame), about an axis passing through C
at
,
the angular velocity . It follows that points C,C
,, ,
appear
stationary
adopt the standard right-handed
x2
e3
x3
= (0, 0, 0), C =
0,
axes
to
Cartesian coordinates, x1
and to choose the coordinate
M )]R,
and C
in this frame. It is also convenient
such that
/(M
(R, 0, 0), and C
([M
+
0).
See Ellipsoidal
Figures
of Equilibrium,
University Press, New Haven CT, 1969).
S. Chandrasekhar
(Yale
52
FLUID MECHANICS
0.6
0.5
0.4
E
0.3
0.2
0.1
0
1
1.2
0.2
0.4
0.6
0.8

L
Figure 3.7: Normalized mechanical
energy
a Maclaurin
versus the
of
spheroid (solid) and a Jacobi ellipsoid (dashed)
normalized angular momentum.
Thus, in the non-inertial

rotation axis
of the
runs
reference
frame, the orbital
parallel to the x3 -axis, and the centers
moon and the planet both lie on the x1-axis.

moon does not rotate (about an axis
Suppose that the
passing
its center of mass) in the non-inertial
through
refer-ence frame. This implies that, in the inertial frame, the
moon appears to rotate
about
an axis
same
and passing through C, at the
orbits
about
synchronous,
and
type
This
ensures
of
that the
parallel to the x3 -axis,

angular velocity
rotation
same
as
it
is termed
hemisphere of the
moon
is always directed toward the planet. Such rotation is
fairly
common
in the Solar System. For instance, the Moon
rotates synchronously
hemisphere is always
in such
a manner
that the
same
visible from the Earth. Synchronous
rotation in the Solar System is
as tidal locking.
Since a synchronously
a consequence
of
process
known
stationary
internal
rotating
in the aforementioned
pressure, p,
is governed
moon
is completely
non-inertial
by
equation of the form [cf., Equation (3.83)]
frame, its
force
balance
= p + (
),
(3.147)
where is the uniform internal
gravitational
potential
gravitational
to
due
mass
density, the
moon,
the
the
potential due to the planet, and

12
!2
x1
M +M
+x 2
(3.148)
the
potential
centrifugal
due
to the
non-inertial frame is rotating (about
x3 -axis and
passing
fact
an axis
that
through point C ) at the angular
velocity [cf., Equation (3.84)]. Suppose that the

much less massive
the
parallel to the
that the planet (i.e., M/M
moon is
1). In this
limit, the centrifugal potential (3.148) reduces to

12
x1
(1/2)
x 12 + (1/2) x 22
(3.157) Hydrostatics
53
where
use has been made
of (3.146).
Suppose that the planet is spherical. It follows that the
potential
by a point
is the same
mass
as that which would
located at C
be generated
In other words,
GM
x1
1 2
x 12 +x 22 + x 32
1/2
R2
GM
1
+
x1
2
2
x1
(1/2) x
2
(1/2)
R2
x 32
(3.150)
where
we have expanded up to second order in x1/R, etc.

32

x1
1
2
x3
(3.151)
where
GM
(3.152)
and any constant terms have been neglected. Thus, the net
force field experienced by the moon due to the combined
action of the fictitious centrifugal force and the
gravitational force field of the planet is

(
=(3 x1
0, x3 ).
(3.153)
The
type
above
of
force
field
is
known
tidal force field, and clearly acts to elongate

along the axis
joining the centers of the
(i.e., the x1-axis), and to
compress
moon
as a
moon
the
and planet
it along the orbital
rotation axis (i.e., the x3 -axis). Moreover, the magnitude of

the tidal force increases
center of the
moon.
linearly
with distance from the
a consequence
The tidal force field is
of the different spatial variation of the centrifugal force and
the planets gravitational
causes
force of attraction. This different
one
away
from the center. As a result of the tidal force field, we
expect the shape of the moon to be distorted from a
sphere. Of course, the moon also generates a tidal force
variation
these two forces, which balance
another at the center of the
moon, to not
balance
field that acts to distort the shape of the planet. However,
we are
assuming
that the tidal distortion of the planet is
much smaller than that of the

earlier statement
As
will
reasonable
be
moon
(which
justifies
that the planet is essentially
demonstrated
provided the
mass
later,
of the
this
assumption
moon
our
spherical).
is
is much less
than that of the planet (assuming that the planet and
moon
have similar densities).

Suppose that the bounding surface of the
ellipsoid
moon
is the
x 12
a 12
x 22
a 22
x 32
= 1,
a 32
(3.154)
where a1 a2 a3
It follows, from Section E,that the
gravitational potential of the moon at an interior point

be written
3
GM
where the integrals i
can
xi
i=1,3
(3.155)
for i 0,3, are defined in
Equations (E.30) and (E.31). Hence, from (3.147) and

(3.151), the
pressure
distribution within the
moon is given
by
"
12
p = p0
3
G M 1
x 12 +
32
G M 2
x 22 +
32
G M 3 +
x 32
(3.156)
pressure. Now, the pressure

on the moons bounding surface, otherwise
where p0 is the central
must be zero
this surface would not be inequilibrium. Thus, inorder to
achieve equilibrium,
we require
"
3
G M 1
x1 +
(3.163)
32
G M 2
54
FLUID MECHANICS
0.05
!
2
x2 +
32
G M 3 +
x3
= p0
0.04
0.03
2
1
3
1
0.02
0.01
0
0
0.10.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
e13
Figure 3.8: Properties of the Roche ellipsoids.
whenever
2
x1
2
a1
x 22
a 22
x 32
a 32
1.
(3.158)
The previous
two equations
can
only be simultaneously
satisfied if
#
"1
"3
a 1 =2 a 2 =
(3/2)G M
+
(3/2)G M
a3
(3.159)
Let a2
= a1 cos and a3 = a1 cos , where . It is

= sin 1(sin/ sin). With the help
also helpful to define

of
some
of the analysis presented in the previous section,
the integrals i
for i 1,3, can be shown to take the form
a1
sin
sin
(3.160)
2
sin
E(,)
3
(3.161)
=
#
"
a 12
F(,) E(,)
3
sin
3
cos
F(,)
cos 2
sin
"
2
2
a1
sin
cos 2
cos
cos sin
E(,)
3
sin
cos
cos 2
cos
,
,
(3.162)
where the incomplete elliptic integrals E(, ) and F(, )
are
defined in Equations (3.141) and (3.142), respectively.
Thus, (3.159) yields
"F(,)(1+
1
sin tan tan
cos
) E(,)
1
+
cos 2
cos 2
sin sin cos cos
Hydrostatics
almost
55
e12
e13
e12
cos 2
e13
0.00
0.00000
0.04
0.04613
0.08
0.09223
0.12
0.13809
0.16
0.18364
0.20
0.22879
0.24
0.27346
0.28
0.31756
0.32
0.36104
0.36
0.40383
0.40
0.44588
0.44
0.48718
0.48
0.52769
Table 3.3:
subject to the constraint
0.00000
0.52
0.56740
0.33440
0.00213
0.56
0.60632
0.38204
0.00852
0.60
0.64445
0.43094
0.01913
0.64
0.68182
0.48027
0.03392
0.68
0.71848
0.52890
0.05282
0.72
0.75446
0.57532
0.07573
0.76
0.78984
0.61729
0.10253
0.80
0.82472
0.65150
0.13308
0.84
0.85923
0.67265
0.16721
0.88
0.89353
0.67151
0.20470
0.92
0.92793
0.62978
0.24528
0.96
0.96294
0.50135
0.28865
1.00
1.00000
0.00000
Properties of the Roche ellipsoids.

"F(,)
cos
2E(,)+ E(,)
cos 2
cos
sin sin cos cos
cos 2
+(3 + cos
cos 2
2E(,)]
(3.164)
"E(,)
2
+ [F(,)
cos
cos 2
2 sin sin cos cos
cos 2
where
and a0
= (a1 a2 a3 )
1/3
is the mean radius of the
The dimensionless parameter
measures
a 03
R
(3.165)
moon.
the strength of
the tidal distortion field, generated by the planet, that acts
on the moon. There
a 0
R3
is an analogous parameter,
(3.166)
where
a 0
is the
mean
radius of the planet, which
the tidal distortion field, generated by the
on
the planet. Now,
former distortion
allowing
we
previously
moon,
assumed
field is much stronger
measures
that acts
that the
than the latter,
us to neglect the tidal distortion of the planet

so to treat it as a sphere. This assumption is
altogether, and
only justified
if
which implies that
where
are
the
M/[(4/3)a
mean
densities
] and
(3.167)
0 3]
M /[(4/3)a
of the
moon
and the planet, respectively.
M M
Assuming
that
are similar, the above condition reduces

Assum
to
or, equivalently, a0 < a 0 In other words,
these densities
neglecting the tidal distortion of the planet, whilst retaining
moon, is generally only reasonable when the

mass of the moon is much less than that of the planet, as
was previously assumed to be the case.
that of the
Equations
(3.163)
ellipsoidal equilibria of
low
mass,
liquid
moon
and (3.164), which
a synchronously
the
due to the tidal force field of the
planet about which it orbits,

in 1850. The properties
describe
rotating, relatively
were
first obtained by Roche
of the so-called Roche ellipsoids
are set out inTable 3.3, and Figures 3.8 and 3.9.
It can be seen, from Table 3.3 and Figure 3.8, that the
eccentricity e12 = sin of a Roche ellipsoid in the x1-x2
plane is almost equal to its eccentricity e13
cular orbit ab
ou ta sph erical planet isabout
56
FLUID MECHANICS
0.07
= sin in the
2.5tim es the
0.06
0.05
0.04
0.03
0.02
0.01
0
0
0.10.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

e13
Figure 3.9: Properties of the Roche ellipsoids.
spheroidal
in shape, being elongated
along the x1-axis
(i.e., the axis joining the centers of the

planet),
the x2
and compressed
-and x3 -axes.
moon
and the
by almost equal amounts along
In the limit
1, in which the tidal
field due to the planet is weak, it is easily shown
distortion
that
2
e 12
2
e 13
15
(3.168)
For the
the
case
Earth,
of the tidal distortion field generated
and
characterized
by M/M
we
= 7.50
obtain
3,
10
and
on
acting
= 0.01230
10
6.
(a1 a3 )/a1
other words,
were
the elongation
the
which
by
is
= 221.29,
= 7.50
and R/a0
It follows that e13
2
5.
e 13
/2 = 2.81
10
In
a homogeneous liquid body,
the Moon
generated
Moon,
by the tidal field of the Earth
would be about 50m.

It
the
can
be
seen,
from Table 3.3 and Figure 3.9, that
parameter
of
eccentricity
attains
maximum
value
as
the
Roche ellipsoid varies from 0 to 1. In
= 0.06757, occurs when e12

= 0.8759. It follows that there is a
fact, this maximum value,
= 0.8594
and
e13
maximum strength of the tidal distortion field, generated

by
planet,
liquid
that
is
consistent
a synchronously
moon in a circular orbit
equilibrium
of
with
rotating,
an
ellipsoidal
homogeneous,
about the planet. It is
plausible that when this maximum strength is exceeded

the
moon
is
tidally
equilibrium condition
disrupted
< 0.06757
by
the
planet.
is equivalent to
The
!1/3
a 0
> 2.455
(3.169)
where
are
M/[(4/3)a
3
0
] and
M /[(4/3)a
3]
mean densities
moon and the planet, respectively. According to the
above expression, there is a minimum orbital radius of a
moon circling a planet. Below this radius, which is called
the Roche radius, the moon is presumably torn apart by
tidal effects. The Roche radius for a syn-chronously
rotating, self-gravitating, liquid moon in a circular orbit
the
of the
Hydrostatics
57
planets radius (assuming that the
moon
and the planet
are
about the same density). Of course, relatively small objects,

such
as
internal
artificial
tensile
satellites,
strength,
which
rather
are
held together
than gravity,
can
by
orbit
inside the Roche radius without being disrupted.
3.15
Exercises
3.1. A hollow vessel floats ina basin. If,as a consequence of
leak, water flows slowly into the vessel, how will the level of the
water inthe basin be affected?

3.2. A hollow spherical shell made
gravity
s >1
has external
up of material of specific
a and b,
and internal radii
respectively.
Demonstrate that the sphere will only float inwater if
!1/3
1s
ba
>
3.3. Show that the equilibrium of

floating with an edge
or corner
a solid of uniform
density
just emerging from the water is
unstable.
3.4. Prove that if a solid of uniform density floats with a flat face
just above the waterline then the equilibrium is stable. 3.5.
Demonstrate that
a uniform
solid cylinder floating with its axis
horizontal is ina stable equilibrium provided that its length

exceeds the breadth of the waterline section. [Hint: The
cylinder is obviously neutrally stable to rotations about its axis,
which
means
that the corresponding metacentric height is
zero.]
a uniform solid cylinder of radius a and height2.

h
can float instable equilibrium, with its axis vertical, if h/a <
If the ratio h/a exceeds this value, prove that the
3.6. Show that
equilibrium is only stable when the specific gravity of the

cylinder lies
outside the range
r1
12
a2
2
h2
3.7. A uniform, thin, hollow cylinder of radius
a and height
his
that h> 2a, prove that the cylinder
open at both ends. Assuming
cannot float upright if its specific gravity lies inthe range
r
1
a2
14
h2
3.8. Show that the cylinder of the preceding exercise
can float
with its axis horizontal provided

3
h
>
sin(s),
2a
s is the specific gravity of the cylinder.

3.9. Prove that any segment of a uniform sphere, made up of a
substance lighter than water, can float instable equilibrium with
where
its plane surface horizontal and immersed.
3.10. A vessel carries
a tank of oil, of specific
gravity
s,
running along its length. Assuming that the surface of the oilis
at sea
level, show that the effect of the oils fluidity
on the rolling
of the vessel is equivalent to a reduction inthe metacentric height
by A
s/V,where V is the displacement of the ship, A the
surface-area of the tank, and
the radius
of gyration of this
area.
In what ratio is the effect diminished when a longitudinal

partition bisects the tank?
3.11. Find the stable equilibrium configurations of a cylinder of

elliptic cross-section, with major and minor radii
respectively, made
up of material
a and b < a,
s,
of specific gravity
which floats with its axis horizontal.

3.12. A cylindrical tank has a circular cross-section of radius
a.
Let the center of gravity of the tank be located
a distance c above
a horizontal
its base. Suppose that the tank is pivoted about
axis passing through its center of gravity, and is then filled with
fluid
up to a depth habove
its base. Demonstrate that the
position inwhich the tanks axis isupright isunstable for all

filling
depths provided
c2 <
a2
2
Show that if c
> (1/2)a
then the upright position is
stable when h lies inthe range
pc
2
a 2 /2.
58
FLUID MECHANICS
3.13. A thin cylindrical vessel of cross-sectional
area A floats
upright, being immersed to a depth h,and contains water to a

depth
k.Show that the work required to pump out the water is 0
Ak(h k)g.
3.14. A sphere of radius
a is just immersed
in water that is
contained in a cylindrical vessel of radius Rwhose axis is
vertical.
Prove that if the sphere is raised just clear of the water then
the waters loss of potential
energy
Wa
is
a2
1
3 R
where W isthe weight of the water originally displaced by

the sphere.
a,weight
3.15. A sphere of radius
rests
on the bottom
W,and specific gravity
s >1,
of a cylindrical vessel of radius Rwhose axis
is vertical, and which contains water to a depth h> 2a.

Show that the work required to lift the sphere out of the vessel is
less than if the water had been absent by
3.16. A lead weight

rotating at
an angular
23
a3
Ws
R2
an amount
in water
is immersed
velocity about
that is steadily
a vertical
axis, the weight
being
suspended
length
l. Prove
vertically
from
downward is stable
g/
or l
>g/ 2
position
on
fixed point
that the position
this axis by
string of
in which the weight
or unstable
depending
hangs
on whether
l
<
respectively. Also, show that if the vertical
is unstable
then there exists
stable
inclined
position in which the string is normal to the surface of equal
pressure
passing though the weight.
3. 17. A thin cylindrical
vessel of radius
and height H is
orientated such that its axis is vertical. Suppose that the vessel is
filled
with liquid of density to
some
height h < Habove the
base, spun about its axis at a steady angular velocity , and the
a steady state. Demonstrate that,

a 2 /g <4hand 2 a 2 /g < 4(H h), the net radial
liquid allowed to attain
provided
thrust
on the vertical
ah
1
+
a2
!2
4gh
3.18. A thin cylindrical

horizontal
walls of the vessel is
vessel
radius
of
with
lid is just filled with liquid of density
plane
, and the
whole rotated
about
vertical axis at
fixed angular velocity . Prove
that the net upward thrust of the fluid
1
4
a
4
on the lid is
2
3.19. A liquid-filled thin spherical vessel of radius
a vertical
a spins
about
diameter at the fixed angular velocity . Assuming that
the liquid co-rotates with the vessel, and that

show that the
pressure on
> g/a,
the wall of the vessel is greatest
depth
g/
thrusts
below
on the lower
the center. Also

and
upper
prove that
are
hemispheres
Mg +
3
M
a,
a,
16
and
14
3
Mg
16
respectively, where Mis the mass of the liquid.
the net normal
3.20. A closed cubic vessel filled with water is rotating about
vertical axis passing through the centers of two opposite sides.

Demonstrate that, as a consequence of the rotation, the net thrust
on a side is increased
by
2,
6
where
a is the length of an edge
of the cube, and the
angular velocity of rotation.
3.21. A closed vessel filled with water is rotating at constant

angular velocity about
a horizontal
axis. Show that, inthe state
of
pressure surfaces inthe

common axis is a height g/ 2
relative equilibrium, the constant
water
are circular
cylinders whose
above the axis of rotation.
a homogeneous, rotating, liquid body of mass

mean radius a0 and angular velocity , whose outer boundary
is a Maclaurin spheroid of eccentricity e.
3.22. Consider
M,
Hydrostatics
59
(a) Demonstrate that
52
(G M/a
3
0
1/2
in the low rotation limit, (G M/a
3
0
1/2
Hence,
e = 0.09262 for the case of a homogeneous body

the same mass and volume as the Earth, which
rotates once every 24 hours.
show that
with
(b) Show that the critical angular velocity at which the

bifurcation to the sequence of Jacobi ellipsoids takes place is
and
= 0.5298(G
occurs
when e
M/a 0 )
1/2,
= 0.81267. Hence, show that, for
the case of a homogeneous body with the same
as the Earth, the bifurcation
mass
and volume
would take place at a
critical rotation period of 2h39m.

(c) Demonstrate that the maximum angular velocity
consistent with a spheroidal shape is
and
= 0.5805(G
occurs
when e
M/a 0 )
1/2,
= 0.92995. Hence, show that, for
the case of a homogeneous body with the same
as the Earth, this maximum
mass
and volume
velocity corresponds to
60
FLUID MECHANICS
Surface Tension
61
4 Surface Tension
4.1
Introduction
As is well-known, small drops of water in air, and small
bubbles of
gas
in water, tend to adopt spherical shapes.
This phenomenon, and a host of other natural phenomena,
can
only be accounted
for
on
the hypothesis
that
an
a
between two different media is associated with
interface
particular
form of
energy
whose magnitude
proportional to the interfacial
is directly
area. To be more exact, if S is
the interfacial
area
the Helmholtz
free energy of the system takes the form S,
then the contribution of the interface to
on
where only depends

composition
of
the temperature
on
the two media
and chemical
either
side of the
interface. It follows, from standard thermodynamics, that
on the system in
an isothermal and
reversible process. However, this work is exactly the same
as that which we would calculate on the assumption that
the interface is in a state of uniform constant tension per
unit length . Thus, can be interpreted as both a free
energy per unit area of the interface, and a surface tension.
This tension is such that a force of magnitude per unit
length is exerted across any line drawn on the interface, in
a direction normal to the line, and tangential to the
S is the work that must
order
to create
the
be performed
via
interface
interface.
Surface
tension
cohesive forces. The
a given
originates
average
isotropic medium
from
intermolecular
energy of a molecule in
possessing an interface with a
free
second medium is independent
of its position, provided
that the molecule does not lie too close to the interface.
However, the free

distance
energy
is modified when the molecules
from the interface becomes less than the

9
the cohesive forces (which is typically 10
range
is
so
macroscopic
small,
the
number
of
range
molecules
system whose free energies
of
m). Since this
are
in
affected by
presence of an interface is directly proportional to the

area. Hence, the contribution of the interface to
the total free energy of the system is also proportionalto the
interfacial area. If only one of the two media in question is
a condensed phase then the parameter is invariably
positive (i.e., such that a reduction in the surface area is
the
interfacial
energetically
molecules of
This
favorable).
a liquid or a solid are
follows
because
subject to
the
an attractive
force from neighboring molecules. However, molecules that
are near to an interface with a gas lack neighbors on one

side, and so experience an unbalanced
cohesive force
directed
toward
the interior
of the liquid/solid.
existence of this force makes it energetically
for the interface to contract (i.e.,

hand, if the interface separates
> 0). On
a liquid
and
The
favorable
the other
a solid, or a
liquid and another liquid, then the sign of cannot be

predicted by this argument. In fact, it is possible for both
signs
of
interfaces.
to
occur at
The surface tension of
= 7.28
10
Nm
1.
liquid/solid
a water/air
and
liquid/liquid
interface at 20
C is
The surface tension at most oil/air

C is
interfaces is much lowertypically,
10
Nm
1.
On the other hand, interfaces between liquid metals and
air
generally
very
have
large
instance, the surface tension of
1
1.
20 C is 4.87
10
Nm
For some pairs
an interface cannot
compression
as
become
as
surface
tensions.
For
mercury/air
interface at
as water
and alcohol,
of liquids, such
generally be observed because it is in
<
(i.e.,
large
an
0). Such
interface
tends to
possible, leading to complete mixing
of the two liquids. In other words, liquids for which
are
immiscible,
miscible.
Finally,
the
liquid/liquid
adsorbed
surface
whereas
those
surface
tension
impurities
tension
deceased
can
interface
in the
for which
at
<
>0
are
or a
presence of
liquid/gas
be affected by the
at the interface. For instance, the
a water/air interface is significantly

presence of adsorbed soap molecules.
at
Impurities that tend to reduce surface tension at interfaces
are termed
4.2
Young-Laplace Equation
Consider
that
surfac-tants.
are
an
interface
in equilibrium
two immiscible
separating
with
one
fluids
another. Let these two
fluids be denoted 1
and 2.Consider
of this interface that is enclosed by
an arbitrary segment S
some closed curve C.
Let t denote
a unit tangent to the curve,
and let
ndenote
unit normal to the interface directed from fluid 1

to
manner.) See Figure 4.1. Suppose that p1 and p2
pressures
of fluids
n in a right-handed
are the
1
and 2,respectively, on either side of S
fluid 2. (Note that C circulates around
Finally, let be the (uniform)
surface
interface.
62
at the
tension
FLUID MECHANICS
21
S
Figure 4.1: Interface between two immiscible fluids.
The net force acting
onZS
S is
IC
IC
(p1 p2 )ndS
ndr,
(4.1)
where dS
=ndS is an element of S,and dr =tdr an
on the right-hand side is

due to the pressure difference across
element of C.Here, the first term

the net normal force
the interface, whereas the second term is the net surface

tension force. Note that body forces play
no role in(4.1),
zero volume. Furthermore,

can be neglected, since both fluids are static.
Now, inequilibrium, the net force acting on S must be
zero: i.e.,
because the interface has
viscous forces
ZS
IC
(p1 p2 )ndS
ndr.
(4.2)
(In fact, the net force would be zero

equilibrium, because the interface
even inthe absence

has zero mass.)
of
Applying Stokes theorem (see Section A.22) to the
curve
C, we find that
IC
ZS
F dr
F dS,
(4.3)
where Fis a general vector field. This theorem
written
IC
can also be
ZS
F tdr
F ndS.
(4.4)
Suppose that F
=g
vector. We obtain
b,where b is an arbitrary constant
IC
ZS
b) tdr
(g
(g
b) ndS.
(4.5)
However, the vector identity (A.179) yields
(g
b)
since bis a constant vector. Hence,

IC
where b (g)
= ( g)b + (b )g,
(4.6)
we get
ZS
n bi
gdr
= b
[(g)
(g j /xi )n
n ( g)n]dS,
(4.7)
Now, since b is also
an
arbitrary vector, the above equation gives

IC
ZS
Taking
Surface Tension
ndr
?(g)
g =n,ZSwe
find that
IC
ed as
63
gdr
n ( g)n
dS.
(4.8)
ZS
[(n)
n ( n)n]dS.
But, (n)
Thus,
=0,because nis a unit vector.
we obtain
IC
2)
n (1/2)(n
ndr
(4.10)
which
ZS
can be combined
?(p1
p2 ) (
(4.11)
n)
n)ndS,
with (4.2) to give

?
ndS
=0.
Finally, given that Sis arbitrary, the above expression

reduces to the pressure balance constraint
= n,
(4.12)
where
known
derived
p
as
= p1
p2
The above relation is generally
the Young-Laplace
by minimizing
equation, and
the free
energy
can
also be
of the interface.
p is the jump in pressure

seen when crossing the interface in the opposite direction
to n. Of course, a plane interface is characterized by
otherahand,
curved interface
generally
on
n t=
o0.
n On
Of the
co urse
p lan eainterface
is cha racter
ized b
(See Section 4.8.) Note that
. ,
=n 0.0.OnIn fact,
has

n
curvature
the
of
interface.
n measures
Thus,
the local
Young-Laplace
mea ncurvatu
equation,
re of the
there is a pressure jump
curved
interface
magnitude
two
between
immiscible
of the jump being proportional
mean
to
according
the
across a
fluids,
the
to the surface
tension.
4.3
Spherical Interfaces
speaking, the equilibrium
Generally
between two immiscible

the
shape of
fluids is determined
an
interface
by solving
force balance equation (3.1) in each fluid, and then
applying
the Young-Laplace
However,
in situations
completely
immersed
droplet in air,
in
or a gas
equation
in which
to the interface.
a mass
second
of
one
fluide.g.,
bubble in waterthe
fluid is
mist
shape of the
interface is fairly obvious. Provided that either the size of
the droplet
or
bubble,
or the
difference in densities
two sides of the interface, is sufficiently

safely ignore the effect of
pressure
pressure
small,
on the
we can
gravity. This implies that the
is uniform in each fluid, and consequently

jump p is constant
over
the interface.
that the
Hence,
mean curvature n of the interface is

also constant. Since a sphere is the only closed surface
which possesses a constant mean curvature, we conclude
from (4. 12), the
that the interface is spherical. This result also follows from
the argument that

minimizes the free
a stable
energy
equilibrium
state is
one
which
subject to the
of the interface,
constraint that the enclosed volume be constant. In other

words, the equilibrium shape of the interface
has the least surface
Suppose
spherical
that
surface
area for a given volume:

the
to
corresponds
interface
is that which
i.e., a sphere.
R, where
is
the
spherical
n = er |r=R
n points away from the
coordinate. (See Section C.4.) It follows that
(Note, for future reference, that
center of curvature of the interface.) Hence, from (C.65),

??????r=R
r2
2
R
(4.13)
The Young-Laplace equation, (4.12), then gives
2
R
(4.14)
Thus, given that p is the
pressure
interface in the opposite direction
seen crossing the

to n, we conclude that
jump
pressure inside a droplet or bubble exceeds that outside

an amount proportional to the surface tension, and
inversely
proportional to the droplet or bubble radius.
This explains why small bubbles are louder that large
ones when they burst at a free surface: e.g., champagne
fizzes louder than beer. Note that soap bubbles in air
the
by
defining the inner and outer extents of
have two interfaces

the
soap
film. Consequently, the net
twice that
4.4
across a
pressure
difference is
single interface.
Capillary Length
Consider
an interface
a
on the surface of
separating the atmosphere from
liquid of uniform density that is at rest
the Earth. Neglecting the density of air compared to that of

the liquid, the pressure in the atmosphere
ve 64
FLUID MECHANICS
can be regarded
Figure 4.2: Interface between
a liquid
(1), a gas (2), and a
solid (3).
pressure in the liquid

as p = p0 g z (see Chapter 3), where p0 is the
pressure of the atmosphere, g the acceleration due to
gravity, and z measures vertical height (relative to the
constant. On the other hand, the
varies
equilibrium
height
of the interface
surface tension). Note that
situation, the Young-Laplace
in the absence
increases
of
upward. In this
equation (4.12) yields
gz
n,
(4.15)
where
nis the normal to the interface
directed from liquid
to air. Now, if R represents the typical radius of curvature

of
the interface
then
the left-hand
equation of
dominates
the interfathe right-hand
and vice
side
of
the above
side whenever
versa. Here,
l,
!1/2
(4.16)
as the capillary length, and takes the value 2.7

3
10
m for pure water at 20 C. We conclude that the effect
of surface tension on the shape of an liquid/air interface is
is known
likely to dominate the effect of gravity when the interfaces

radius of curvature is much less than the capillary length,
and vice
4.5
versa.
Angle of Contact
a liquid/air interface is in contact with a solid,

case for water in a glass tube, or a drop of
resting on a table. Figure 4.2 shows a section
Suppose that
as would
mercury
be the
perpendicular to the edge at which the liquid, 1, the air, 2,
,, ,
per
and the solid, 3, meet. Suppose that the free energies
area at
interfaces are
unit
the
12
liquid/air,
13
liquid/solid,
and 23
and
air/solid
respectively.
If the
boundary between the three media is slightly modified in

the neighborhood
of
as indicated by the dotted

area of contact of the air with the
a small amount r per unit breadth
the edge,
line in the figure, then the

solid is increased by
(perpendicular
to the figure), whereas that of the liquid
with the solid is decreased by r
per
unit breadth, and that
cos per
energy per
of the liquid with the air is decreased by r
unit
breadth. Thus, the net change
unit
in free
breadth is
23
13
12
cos
(4.17)
an equilibrium state is one which minimizes

energy, implying that the above expression is zero
However,
the
free
for
arbitrary (small) r: i.e.,

cos
23
13
12
(4.18)
We conclude that, in equilibrium, the angle of contact, ,
a fixed value that

per unit area at the liquid/air,
between the liquid and the solid takes
depends
os
on the
free energies
(4.20)
Surface Tension
65
air
a
glass tube
liquid
h
air
free surface
z =0
liquid
Figure 4.3: Elevation of liquid level in a capillary tube.
formula could also be obtained from the requirement
the various
balance
surface
one
forces
acting
that
at the edge
another, assuming that it is really appropriate
to interpret 13
the media
tension
as surface tensions when one of

up the interface is a solid.
in Section 4.1 we would generally expect
and 23
making
As explained
12 and 23 to be positive. On the other hand, 13 could be

either
positive
Equation (4.18)
or negative.
can only be
range
23
+ 12 >13 > 23
12
Now, since
|
cos | 1
solved when 13
lies in the
(4.19)
If 13
> 23
+ 12
to the
interface is
not wet the solid
on its
surface. On the other hand, if 13
the angle of contact is 0

where the
free
energy at
< 23
which corresponds
the liquid/solid
small that the liquid completely
out indefinitely
case where the free energy at

so large that the liquid does
at all, but instead breaks up into beads
which corresponds
the liquid/solid
then the angle of contact is 180
until it either
12
then
to the
case
is so
interface
wets the solid, spreading
covers
the whole surface,
or
its thickness reaches molecular dimensions.

The angle of contact between water and glass typically
lies in the
range
25
to 29 , whereas that between

.
mercury
and glass is about 127
4.6
Jurins Law
a situation in which a narrow, cylindrical, glass

a is dipped vertically into a liquid of
, and the liquid level within the tube rises a
h above the free surface as a consequence of
Consider
tube of radius
density
height
surface tension. See Figure 4.3. Suppose that the radius of
the tube is much less than the capillary length. A tube

for which this is the case is generally known
tube. According
to the discussion
shape of the internal
as a capillary
in Section
liquid/air interface within
tube is not significantly

Section 4.3, the interface
4.4, the
a capillary
affected by gravity. Thus, from
is a segment of
a sphere
of radius
R (say). If is the angle of contact of interface with the

glass then simple geometry
(see Figure 4.3) reveals that
a
cos
(4.29) 66
FLUID MECHANICS
Hence, from Equation (4.13), the mean curvature of the
interface is given by
n=
2 cos
(4.21)
where is the associated
surface
sign in the above expression

points
towards
the
center
tension. [The minus
arises from the fact that

of
curvature
the
of
interface, whereas the opposite is true for Equation (4.13).]
Finally, from (4. 15), application
of the Young-Laplace
equation to the interface yields
gh
2 cos
(4.22)
which
can be rearranged to give

2 cos
h
ga
(4.23)
a liquid
This result, which relates the height, h, to which

rises in a capillary tube of radius
tension,
is known
as
a to the
liquids surface
Jurin s law. Note
that
the
assumption that the radius of the tube is much less than the
capillary
height
length is equivalent
of the interface
liquid is much
to the assumption that the
above the free surface
of the
greater than the radius of the tube. This
follows, from (4.16) and (4.23), because
= 2 cos
(4.24)
a2
a limplies that h a.
,
For the case of water at 20
assuming a contact
,
angle of 25
Jurins law yields h(mm)
13.5/a(mm).
Thus, the ordering
mm in a capillary tube
of radius 1
mm, but rises 13.5 cm in a capillary tube
of radius 0.1
mm. Note that in the case of a liquid, such a
mercury, that has an oblique angle of contact with glass, so
that cos < 0, the liquid level in a capillary tube is
Thus,
water rises
height 13.5
depressed below that of the free surface (i.e., h< 0).
4.7
Capillary Curves
Let adopt Cartesian coordinates

such that
increases
the interface of
a liquid
with the atmosphere

where
vertically
on
surface
of density and surface tension
corresponds
of the interface
z = f(x),
z < f(z). Note that
to the surface
the liquid occupies the region
the shape
the Earths
upward. Suppose that
is y-independent.
The unit
normal to the interface (directed from liquid to air) is thus
(z f)
|(z f)|
ez
fx
ex
(1+ f x 2 )1/2
(4.25)
where fx
= df/dx. Hence, the mean curvature
of the
interface is
where fxx
=d
shape
the
of
f/dx
fxx
(1+ f x2 ) 3/2
(4.26)
According to (4.15) and (4.18), the
interface
is governed
by
the
nonlinear
differen-tial equation
fxx
(1+ f x2 )3/2
(4.27)
where the vertical height, f,of the interface is measured
relative to its equilibrium height in the absence of surface
2,
tension. Multiplying the above equation by fx /l
and
integrating with respect to x,we obtain
1
(1+ f x2 ) 1/2
f
C
2l 2
(4.28)
where C is a constant. It follows that
f
C
2l 2
1,
(4.37)
67
Surface Tension
z/l
0
x/l
Figure 4.4: Capillary
curves
for /4 3/4 and (in
= 0.6, 0.7, 0.8, 0.9,

Figure
4.4: Capillary curves
and
order from the top to the bottom)
0.99.
and
Let
where 0 < k < 1, and
Thus, from (4.31) and (4.32),
1
fx
Cf
[1 (C
C
f
2
k2
= 2l (1 k
2/2l
2
/2l 2) 2] 1/2
1,
2
sin
(4.30)
(4.31)
1/2.
(4.32)
and
2l 2
= cos(2),
so the constraint
(4.33)
(4.29) implies that /4 3/4.
Moreover, Equations (4.30) and (4.33) reduce to
fx
dx
tan(2)
df
(4.34)
It follows from (4.32) and (4.34) that
dx
dx
df
df
lk cos(2)
(1 k 2 sin
) 1/2
(4.35)
which

xl
/2
k cos(2)
(1 k 2 sin
d,
) 1/2
(4.36)
assuming that
x =0 when =/2.
Thus, we get
!
xl
68
FLUID MECHANICS
2k
k
F(,k)
2k
E(,k),
0.8
0.7
0.6
0.5
z/l 0.4
0.3
0.2
0.1
0.5
0.5
x/l
Figure 4.5: Liquid/air interface for
a liquid
trapped
between two vertical parallel plates locatedat
contact
where
x = l.
angle of the interface with the plates is
The
=30 .
=
=
E(,k)
F(,k)
(4.38)
E(/2,k)
E(,k),
F(/2,k)
F(,k),
(4.39)
and
Z
E(,k)
(1 k
sin
1/2,
(4.40)
F(,k)
(1 k
sin
1/2,
(4.41)
are types
of incomplete
elliptic integral.
In conclusion,
the interface shape is determined parametrically by

!
xl
2k
=
zl
(4.42)
F(,k)+
2k
(1 k
2k
E(,k),
sin
1/2,
(4.43)
where
/4
Here, the parameter
3/4.
k is restricted
to lie in the
range
< k < 1.
toFigure
lie in the
4.4ran
shows
ge 0 <
the
k<
capillary
1.
curves
predicted
by
(4.42) and (4.43) for various different values of k.Here,
we
have chosen the plus sign in (4.43). However, if the
curves are simply inverted:

x x and z z. In can be seen that all of the
curves shown in the figure are symmetric about x = 0: i.e.,
z z as x x. Consequently, we can use these curves to
determine
the shape of the liquid/air interface which arises
z as
when a liquid is trapped between two flat vertical plates
(made of the same material) that are parallel to one
minus sign is chosen then the
i.e.,
another. Suppose
1
See Handbook of Mathematical Functions, M.Abramowitz,
and I.A.
Stegun (Dover, New York NY, 1965).
Surface Tension
69
that the plates in question lie at
x = d.
Furthermore, let
the angle of contact of the interface with the plates be ,

where
e
expect
,
and
< /2.
Since the angle of contact is acute,
we
the liquid to be drawn upward between the plates,
the
to be
interface
concave
(from
above).
This
corresponds to the positive sign in (4.43). In order for the

interface to meet the plates at the correct angle,
fx
1/tan at
However, if
one
d and fx
= 1/tan
we
at x
require
+d.
of these boundary conditions is satisfied
then, by symmetry,
From Equation
the other is automatically
(4.34)
(choosing
the positive
latter boundary condition yields tan(2 )
satisfied.
sign), the
= 1/ tan at x =
Substituting
to
+d, which is equivalent
= +d when
=
numerically determine the value

shape
interface
= 3/4/2.
=
of k for which x = d. The
is then given by Equations
(4.43), using the aforementioned
range
= 30
/4
+/2
then k
shown
(i.e.,
Consider
(4.42) and
of k, and in the
and the associated
and
angle of
is
interface
we invert this
z z) then we obtain the
to the same plate spacing, but
4.5. Furthermore,
interface which corresponds
an obtuse contact
value
to 3/4 /2. For instance, if of

d k,and
land
= 0.9406,
in Figure
interface
this value of intoxEquation

+d when
(4.42),
we
3 /4
can
if
= 180 30 =150 .
the limit k 1, which is such that the
distance
between
the the
twolimit
plates
much
thanthat
the
Consider
k is1,
whichless
is such
capillary
length. It is easily demonstrated that, at small k,
E(,k)
cos
+ sin
),
F(,k)
(4.44)
+ sin
cos ),
4
(4.45)
where
= /2 . Thus, Equations
reduce to
xl
sin(2 ),
(4.42) and (4.43)
zl
2k
(4.46)
[1 cos(2 )].
2
(4.47)
It follows that the interface
surface
of
y-axis. If
cylinder
is
whose
a segment of the curved

runs parallel to the
axis
the distance between the plates is 2d, and the
contact angle the

is ,distance
then we be
require
tw eex
/2
(which
corresponds
to
n
=the
d when
pl ates
=i3/4
s 2d,
/4
/2).
From
Equation (4.46), this constraint yields

dl
cos.
2
(4.48)
Thus, the height that the liquid rises between the two
platesi.e., h
=z(x = 0) =z( = /2) 2l/kis
given
by
cos
h
gd
(4.49)
a
up between two parallel plates of spacing 2d.
the case k = C = 1, which is such that the
This result is the form taken by Jurins law, (4.23), for

liquid drawn
Consider
distance
leftmost
between
the
plate lie at
two plates
x = 0, and
let
is infinite.
us
Let
completely
the
neglect
the rightmost plate, since it lies at infinity. Suppose that h
= z(x = 0)
surface
is the height of the interface above the free
of the liquid at the point
meets the leftmost
interface with the plate then
= 0. Since C = 1, it
where the interface
plate. If is the angle of contact of the
we require
= 1/ tan at x
fx
follows from (4.28) that
2l2
1 sin,
(4.50)
or
h
2 lsin(/4
/2).
(4.51)
Furthermore, again recalling that C
=1,Equation
(4.30)

Z
x=
z
df
fx
h/2l
=l
z/2l
1
2y
y(1
y 2) 1/2
dy,
(4.52)
#
"
2
See Handbook of Mathematical Functions, M.Abramowitz, and I.A.
(1+ f z 2 ) 1/2
70
FLUIDMECHANICS
1.1
1
0.9
0.8
0.7
0.6
z/l
0.5
0.4
0.3
0.2
0.1
0
0
x/l
Figure 4.6: Liquid/air interface for a liquid in contact with
a vertical
=0.The contact angle of the

=25 .
plate located at x
interface with the plate is
where
we
have chosen the minus sign, and
Making the substitution
xl
"ln
sinu
du
= f/2l.
y =sinu, this becomes

sin
(h/2l)
2
sin
1
+ cosu
#sin
(z/2l)
(h/2l)
+ 2 cosu
sinu
sin
(z/2l)
sinu
(4.53) which reduces to

xl
!
= cosh
2l
cosh
2l
!1/2
l2
z2
!1/2
l2
(4.54)
since cosh
1(z)
ln[z
(z
1)
1/2].
(4.51) and (4.54) specify the shape of

that
meets
an
isolated
vertical
Thus, Equations
a liquid/air interface
x = 0. In
plate at
particular, (4.51) gives the height that the interface climbs
up the plate
(relative to the free surface) due to the action
of surface tension. Note that this height is restricted to lie
in the
range
2 l
h 2 l,irrespective
contact. Figure 4.6 shows

for
= 25 .
an
of the angle of
example interface calculated
4.8
Axisymmetric Soap-Bubbles
Consider
an axisymmetric
soap-bubble whose surface
=f(z) incylindrical
takes the form r
coordinates. See
Sec-tion C.3. The unit normal to the surface is
(r f)
|(r f)|
er
fz
ez
(1+ f z 2 ) 1/2
(4.55)
where fz df/dz. Hence, from (C.39), the
mean curvature
of the surface is given by

#
"
n=
f fz dz
(1+ f z 2 )1/2
Surface Tension
(4.66)
71
The Young-Laplace equation, (4.12), then yields

#
"
f fz
(4.57)
dz
(1+ f z 2 )1/2
where
p0
(4.58)
Here,
is
contributions
the
net
surface
the
from
interfaces. More-over, p0
between
the interior
Equation (4.57)
tension,
internal
and
including
external
= p is the pressure
and the exterior
can be
the
soap/air
difference
of the bubble.
integrated to give
(1+ f z 2 )1/2
+C,
2a
(4.59)
where C is a constant.
,.
z
a fixed radius at its two
Suppose that the bubble occupies the region z1
z2
< z2
where z1
end-points,
z = z1
and
z = z2
achieved by supporting
co-axial rings
energy
E
and has
located at
= z1
This could most easily be
the bubble
on two rigid parallel

z = z2 The net free
can be written
and
required to create the bubble
= S p0 V,
(4.60)
where
S is
area
of the bubble
enclosed volume. The first term
surface, and V the
on the
right-hand side
of the
represents
above expression
overcome
surface
represents
the work
difference,
p0
tension,
the work needed to
whilst
the
term
second
overcome
to
required
the
pressure
between the exterior and the interior of
we
a stable equilibrium state of a mechanical system to
such as to minimize the net free energy, subject to any
the bubble. Now, from the general principles of statics,
expect
be
dynamical
constraints.
It follows
shape of the bubble issuch

z2
as to
that
the
equilibrium
minimize
2
2 f (1+ f z )
1/2
z2
dz p0
z1
dz,
z1
(4.61)
subject to the constraint that the bubble radius, f,be fixed
at
z = z1
and
z = z2
Hence,
we need to find
f(z) that minimizes the integral
the function
z2
L(f, fz )dz,
z1
(4.62)
where
L(f, fz )
2 f (1
+ f z 2 )1/2
p0 f
(4.63)
subject to the constraint that f is fixed at the limits. This
is
a standard
Appendix
problem in the calculus of variations. (See
D.) In fact, since the functional L(f, fz )does
on z,the minimizing
not depend explicitly
function is the
solution of [see Equation (D.14)]

L
L fz
fz
(4.64)
where C
is an arbitrary constant. Thus,
"
(1+ f z 2 )1/2
2a
we obtain
=C
(4.65)
can be rearranged to give Equation (4.59). Hence,

we conclude that application of the Young-Laplace
equation does indeed lead to a bubble shape that minimizes
the net free energy of the soap/air interfaces.
Considerthe case p0 = 0,inwhich there is no pressure
which
differenceacross the surface of the bubble. Inthis situation,
writing C
= b > 0,Equation (4.59) reduces to
=b(1+ f z
1/2.
(4.72)
72
FLUIDMECHANICS
1
0.9
0.8
0.7
0.6
r/c
0.5
0.4
0.3
0.2
0.1
0
0.6 0.4 0.2
0.2
0.4
z/c
Figure 4.7: Radius
soap bubble
of radius
versus
axial distance for a catenoid
supported by two parallel co-axial rings
c located at z = 0.65c.
0.6
Moreover, according to the previous discussion, the bubble
as to minimize the surface

area of the bubble (since the only contribution to the free
energy of the soap/air interfaces is directly proportional to
the bubble area). The above equation can be rearranged to
shape specified by (4.66) is such
give
fz
!1/2
b2
(4.67)
which leads
to
Z
r
z z0 =
df
fz
=
b
df
(f 2/b
= b cosh 1(r/b),
1) 1/2
(4.68)
or
r
b cosh(|z
z0 |/b),
(4.69)
where
z0
is
a constant.
This expression
axisymmetric surface known

Suppose,
for
instance,
generality,
distance
we can
an
as a catenoid.
that
soap bubble is
c that are located a
the
supported by identical rings of radius

perpendicular
describes
2 d apart.
Without
specify that the rings lie at
thus follows, from (4.69), that z0
=0,and
loss
= d.
of
It
b cosh(z/b).
(4.70)
Here, the parameter b must be chosen
so as to satisfy
c = b cosh(d/b).
(4.71)
For example, if d
= 0.65 c then b = 0.6416
c, and
the
resulting bubble shape is illustrated inFigure 4.7.

Let d/c
= and
d/b
= u, in which
case
the above
equation becomes
G(u)
(4.78) Surface Tension
73
1.1
1
0.9
0.8
0.7
0.6
cosh
u=
0.
0.5
0.4
0.3
0.2
0.1
0
z/
Figure 4.8: Radius
soap bubble
versus
axial distance for an unduloid
calculated with k
Now, the functionG(u) attains
=0.95.
a maximum
G(u0 )
value
1
u0
tanhu0
(4.73)
when u0
= sinh 1(1/).
Equation (4.72)
Moreover,
possesses two roots.
if G(u0 ) > 0 then
It turns out that the
root associated with the smaller value of

interface
maximizes
energy, whereas
free energy. Hence,
system
the
u minimizes
the
the
other
root
the
former
root
to
corresponds
stable equilibrium
state, whereas
the
latter corresponds to an unstable equilibrium state. On the

other hand, if G(u0 )< 0 then Equation (4.72)
no roots,
any
implying the absence of
possesses
state.
equilibrium
case G(u0 )= 0 corresponds to u = uc and =

c where uc tanhuc = 1 and c = 1/ sinhuc
It is easily
demonstrated that uc = 1.1997 and c = 0.6627. We
state of a catenoid
conclude that a stable equilibrium
bubble only exists when c which state
corresponds
to d
of a ca tenoid
0.6627 c.If the relative ring spacing d exceeds the critical
The critical
value 0.6627c then the bubble presumably bursts.
case p0
0, in which there is a pressure
across the surface of the bubble. In this
Consider the
difference
situation, writing
2a
(4.74) 2aC
+,
(4.75)
Equation (4.59) becomes
+ ) f
(1+ f z 2 )1/2
(4.76)
which

fz
2) 1/2
(f
2) 1/2
(4.77)
can assume, without loss of generality,

It follows,
We
can assume,
from wit
the above expression, that
Hence, we can write
We
that ||
> ||.
|| f ||.
cos 2 + 2
sin
(4.87) 74
FLUID MECHANICS
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.5
0.5
z/
Figure 4.9: Radius
for a positive
pressure
with k
nodoid
versus axial distance

soap bubble calculated
=0.95.
,
,,
(4.79)
where 0
/2
and 0 < k 1. It follows that

f
(4.80)
=
=
||(1 k
sin
sgn()||(1
1/2
)
2) 1/2
k
(4.81)
and
,
,
dz
1fz df
f+
(4.82)
which
|z|
hE(, k)
||
sgn()
(1
2 1/2
k )
F(, k)
(4.83)
where E(, k) and F(, k) are incomplete elliptic integrals

[see Equations (4.40) and (4.41)]. Here,
that
we have
assumed
=0 when z =0.There are three cases of interest.

In the first case, > 0 and > 0. It follows that
(1
2) 1/2
1for
Inr/
the
first
ca /2
0, and 0.5 /(p0 )
<1
> 0,where
for 1
k
rhE(,k)=
(4.84) |z|
(1 k
+ (1 k
2) 1/2
sin
),
F(,k)
(4.85)
curve
The axisymmetric
of equations
bubble
unduloid
parameterized
as an
is known
always
has positive
(relative to the external pressure): i.e., p0

unduloid
soap
In the
It
follows
by the above pair
unduloid.
an
Note that
pressure
internal
> 0. An example
bubble is illustrated inFigure 4.8
second
that
case,
(1
2 1/4
k )
>
0 and
r/
<
0.
Int he
1
cond
for case,
0 >0 0,
and
and
<
00.
<It/(p0
follows
) that
0.5
(1 for 0
1([1
2)1/2] 1/2/k),
1, where 0
sin
(1 k
and
<k
rhE(,k)=
(4.86)
|z|
Surface Tension
75
(1 k
(1 k
2) 1/2
sin
),
F(,k)
0.6
0.5
0.4
0.3
0.2
0.1
0.1
0.2
0.3
0.4
0.5
z/||
versus axial distance for a negative

soap bubble calculated with k = 0.95.
Figure 4.10: Radius
pressure
nodoid
The axisymmetric
curve
parameterized
by the above pair
of equations is known
as an nodoid.
This particular type
pressure: i.e., p0
soap bubble
of nodoid bubble has positive internal
> 0. An example
positive
pressure
nodoid
is illustrated inFigure 4.9.
In the third case, < 0 and > 0. It follows that

2)1/2|
2 1/4
r/|| | (1 k )
for /2 0
(1 k
(or(1
/2
k
0 ),and 0 > /(p0 ||) 0.5 for 0 < k
1, where
sin
),
(4.88)
i
2
(1 k )
||
hE(,k)
|z|
||(1 k
1/2
F(,k)
(4.89)
curve
a
The axisymmetric
of equations
is again
parameterized
by the above pair
nodoid. However, this particular
type of nodoid bubble has negative internal

p0
<
0. An example
negative
pressure:
pressure
nodoid
i.e.,
soap
bubble is illustrated inFigure 4.10.
4.9
Exercises
4.1. Show that if Nequal spheres of water coalesce
a single spherical
a factor 1/N
drop then the surface

1/3
4.2. A circular cylinder of radius
energy
a,height
so as to form
isdecreased by
h,and specific gravity
s floats
upright inwater. Show that the depth of the base below
the general level of the water surface is

sh +
2
cos,
where isthe surface tension at the air/water interface, and

the contact angle of the interface with the cylinder. 4.3. A film
of water is held between two parallel plates of glass
a small
distance 2d apart. Prove that the apparent attraction

between the plates is
2A cos
+ L sin,
where is the surface tension at the air/water interface,

the angle of contact of the interface with glass, A the area of the
film, and Lthe circumference of the film.
76
FLUID MECHANICS
4.4. Show that if the surface of a sheet of water is slightly

corrugated then the surface
energy
is increased by
!2
dx
per unit breadth
of thecorrugations. Here, xismeasured
horizontally, perpendicular tothe corrugations. Moreover,

denotes the elevation of the surface above the mean level. Finally,
is the surface tension at an air/water interface. If the

corrugations
are sinusoidal,
such that
= a sin(k x),
show that the average increment of the surface
energy per unit

area is(1/4)a 2 k 2
4.5. A mass of liquid, which is held together by surface tension
alone, revolves about a fixed axis at a small angular velocity , so
as to assume a slightly spheroidal shape of mean radius a.Prove
that the ellipticity of the spheroid is
where is the uniform

[If
r+
= (r
mass
a3
density, and the surface tension.
is the maximum radius, and r the minimum radius, then

r ) 1/3, and the ellipticity is defined (r+ r )/r+ .]
mass rotates, in the form of a circular ring of

a and small cross-section, with a constant angular velocity
, about an axis normal to the plane of the ring, and passing
through its center. The mass is held together by surface tension
4.6. A liquid
radius
alone. Show that the section of the ring must be approximately

circular. Demonstrate that
!1/2
2
ac
where is the density, the surface tension, and
c the radius
of the cross-section.
4.7. Two spherical
soap
are made to
gas inthe
to its initial value the radius a
bubbles of radii a1 and a2
coalesce. Show that when the temperature of the

resulting bubble has returned
of the bubble satisfies
p0
a + 4a
=p0 (a
where p0 is the ambient
3
1
+ a 23 )+ 4(a 12 + a 22 ),
pressure,
and the surface tension
of the soap/air interfaces.
4.8. A rigid sphere of radius
a small amount
a rests on a flat
rigid surface, and
of liquid surrounds the contact point, making
lens whose diameter is small compared to
concave-planar
a.
The angle of contact of the liquid/air interface with each of the

solid
surfaces
is
zero,
and
the surface
tension
of
the
interface is . Show that there is an adhesive force of magnitude

4a
acting
on the
sphere. (It is interesting to note that the force
is independent of the volume of liquid.)
4.9. Two small solid bodies
are
floating
on
the surface of
liquid. Show that the effect of surface tension is to make the

objects
approach one another if the liquid/air interface has either
an acute or an obtuse angle of contact with both bodies, and to
make them move away from one another if the interface has
an acute angle of contact with one body, and an obtuse angle of
contact with the other.

Incompressible Inviscid Fluid Dynamics
77
5 Incompressible Inviscid Fluid

Dynamics
5.1
Introduction
This chapter introduces
some
of the fundamental concepts
that arise in the theory of incompressible, inviscid (or, to

be more exact, high Reynolds number) fluid motion.
5.2
Streamlines, Stream Tubes, and Stream
Filaments
A line drawn in a fluid such that its tangent at each point is
a
a given
parallel to the local fluid velocity is called
streamline.
The aggregate of all the streamlines at
instance in
time
constitutes
streamlines
the
instantaneous flow
drawn through each point of
pattern.
closed
The
curve
a stream tube. Finally, a stream filament

as a stream tube whose cross-section is a curve
constitute
defined
is
of
infinitesimal dimensions.
When the flow is unsteady then the configuration
the stream tubes and filaments
changes
of
from time to
time. However, when the flow is steady then the stream
are stationary. In the latter case, a

an actual tube through which the fluid
follows because there can be no flow
tubes and filaments
stream tube acts like

is flowing. This
across
the walls, and into the tube, since the flow is, by
definition, always tangential to these walls. Moreover, the
walls
are
fixed in
space
and time, since the motion is
steady. Thus, the motion of the fluid within the tube would
be unchangedwere the walls replaced by a rigid frictionless

boundary.
a stream
Consider
filament of
an incompressible
fluid
whose motion is steady. Suppose that the cross-sectional
area
of the filament
velocity
is the
is sufficiently
same at
let the cross-section
Moreover,
the direction of this

and v2
which
each point
are the
the
common
Consider
the
the cross-section.
normal to
be everywhere
velocity. Suppose that v1
flow speeds at two points
cross-sectional
respectively.
small that the fluid
on
on the filament at
areas are
and
section of the filament
lying
between these points. Since the fluid is incompressible,
the
same
volume of fluid must
section, in a given time interval,
flow into
as
one
end of the
flows out of the other,
which implies that

v1
=v2
(5.1)
This is the simplest manifestation of the equation of fluid
continuity discussed in Section 2.9. The above result is
equivalent
to the statement
and cross-sectional
filament
Thus,
of
an
area
that the product of the speed

is constant
incompressible
a stream
terminate
unless
infinite.
Leaving
filament
the
within
velocity
this
along
any stream
fluid in steady
such
at that
case out
of
motion.
fluid
point
cannot
becomes
consideration,
it
follows
that
incompressible
stream
fluids
are
in
filaments
steadily
flowing
or terminate
of course, true
either closed loops,
at the boundaries of the fluid. The
same
is,
of streamlines.
5.3
Bernoullis Theorem
In its most general form, Bernoulli s theoremwhich
was
discovered
that,
by Daniel Bernoulli (17001783)states
inthe steady flow of an inviscid fluid, the quantity
(5.2)
is constant along
streamline,
the density, and T the total
where
energy per
The proof is straightforward.
is the
unit
areas
the body of
AB and CD of
,,, ,
the stream filament pictured in Figure 5.1. Let

the values of quantities
mass.
Consider
fluid bounded by the cross-sectional
pressure,
us
denote
at AB and CD by the suffixes 1
Thus, p1 v1 1
S 1 T1 are the
pressure, fluid speed, density, cross-sectional area, and
total energy per unit mass, respectively,
at AB, etc.
Suppose that, after a short time interval t, the body of
and 2, respectively.
fluid has moved such that it occupies the section of the
,
filament bounded by the cross-sections A B and C D
where AA
= v1
t and CC
= v2
t. Since the motion is
steady, the
mass m of the fluid
same as that between
between AB and A B
CD and C D
so
is the
that
m =S 1v1 t1
5.10)
=S
2 v2
t2
78
FLUID MECHANICS
D
D
C
Figure 5.1: Bernoullis theorem.
Let T denote the total
lying between A B
energy
of the section of the fluid
and CD. Thus, the increase in energy
of the fluid body inthe time interval t is
(mT2
T)
(mT1
T)
m(T2
T1).
(5.4)
In the absence of viscous
must
increase
energy
equal the net
dissipation, this
work
done by
energy
the fluid
pressures at AB and CD, which is
p1 S
1 v1
t p2 S 2
v2
=m
p1
p2
(5.5)
Equating expressions (5.4) and (5.5), we find that
T1
p2
+ T1
1
which demonstrates
any two
therefore
points
(5.6)
+T
given
has the
stream
same
filament,
value at
and
is
constant along the filament. Note that Bernoullis
theorem
therefo rehas only been proved for the
motion of
may
that p/
on a
T2
an inviscid
case
of the steady
fluid. However, the fluid in question
or incompressible.
case of an incompressible fluid,
moving in a conservative force-field, the total energy per
unit mass is the sum of the kinetic energy per unit mass, (1
2,
/2) v
and the potential energy per unit mass, , and
either be compressible
For the particular
Bernoullis
theorem thus becomes
12
constant along
streamline.
(5.7)
If
we
focus
on a
particular
1(say), then
streamline,
Bernoullis theorem states that
12
+ = C1
(5.8)
where C1 is a constant characterizing that streamline. If we
consider
a second
streamline, 2 (say), then
12
= C2
(5.9)
where C2 is another constant. It is not generally the
that C1
= C2
case
If,however, the fluid motion is irrotational
then the constant in Bernoullis theorem is the
same
for all
=C
streamlines (see Section 5.7), so that
12
have
79
S2
A
S1
Figure 5.2: A vortex filament.
throughout the fluid.
5.4
Vortex Lines, Vortex Tubes, and Vortex
Filaments
The curl of the velocity field of
a fluid, which
is generally
termed vorticity, is usually represented by the symbol :

i.e.,
v.
(5.11)
A vortex line is
line whose tangent is everywhere
parallel to the local vorticity
vector. The vortex
a closed curve
Finally, a vortex
drawn through each point of

surface of a vortex
vortex
tube
dimensions.
Consider
tube.
whose
cross-section
section
AB of
is
filament
of
a vortex
lines
constitute the
is
infinitesimal
filament.
The
filament is bounded by the curved surface that forms the

filament
wall, as well as two plane surfaces, whose vector
areas are
S1 and S2
(say), which form the ends of the
section at points A and B, respectively. See Figure 5.2. Let
the plane surfaces have outward pointing normals that

parallel (or anti-parallel)
to the vorticity vectors, 1
are
and
at points A and B, respectively. Gausss theorem (see
Section A.20), applied to the section, yields

I
dS =
dV,
(5.12)
where dS is an outward directed surface element, and dV

volume element. However,
(5.13)
v0
[see Equation (A.173)], implying that
dS = 0.
(5.14)
Now,
dS = 0 on the curved
surface of the filament,
since is, by definition, tangential to this surface. Thus,

the only contributions to the surface integral
areas
plane
S1 and S2
come
from the
It follows that
dS = S 2 2
11
=0.
(5.15)
This result
vortex
is essentially
filaments.
magnitude
of
It
an
implies
the vorticity
equation
that
the
of continuity
product
of
and the cross-sectional
for
the
area,
which is termed the vortex intensity, is constant along the
filament. It follows
that
a vortex
filament cannot terminate
in the interior of the fluid. For, if it did, the cross-sectional

of the tube. See 80
FLUID MECHANICS
to vanish, and, therefore, the vorticity, , would have to

become infinite. Thus,
a vortex
filament must either form
closed vortex ring, or must terminate at the fluid
boundary.
a vortex
Since
tube
can be regardedas a bundle
of
vortex filaments whose net intensity is the sum of the

intensities
of the constituent filaments,
we conclude
that the
intensity of a vortex tube remains constant along the tube.
5.5
Circulation and Vorticity
a closed curve
Consider
C situated entirely within
moving fluid. The vector line integral (see Section A.14)

IC
C
v dr,
(5.16)
where dr is
an
element of C, and the integral is taken
curve, is termed the circulation of the

curve. The sense of circulation (e.g., either
around the whole

flow around the
clockwise
or counter-clockwise) is arbitrary.
a surface having the closed curve C for a
and let dS be an element of this surface (see
Let S be
boundary,
Sec-tion A.7) with that direction of the normal which is

related to the chosen
right-hand
sense
of circulation around C by the
circulation rule (see Section A.8). According to
Stokes theorem (see Section A.22),
IC
C
ZS
v dr =
dS.
(5.17)
Thus,
we
conclude
intimately related to
that
one
circulation
and
are
vorticity
another. In fact, according to the
above expression, the circulation of the fluid around loop C

is equal to the net
sum
of the intensities
of the vortex
passing through the loop and piercing the surface
filaments
making a positive, or negative,

sum depending on whether it pierces the
surface in the direction determined by the chosen sense of
circulation around C and the right-hand circulation rule, or
S
(with
filament
contribution to the
in the opposite direction). One important proviso to (5.17)

is that the surface S must lie entirely within the fluid.
5.6
Kelvin Circulation Theorem
According
named
around
to the Kelvin circulation
theorem,
after Lord Kelvin (18241907),
any
independent
co-moving
of
circulation around
time.
in
an
proof
is
loop
The
a given
inviscid
as
fluid
follows.
loop C is defined
C
which is
the circulation
is
The
IC
v dr.
(5.18)
However, for
have dv
loop that is co-moving
we
=d(dr/dt) = d(dr)/dt. Thus,
IC
dC
with the fluid,
IC
dv
dt
dr +
v dv.
dt
(5.19)
However, by definition, dv/dt
=Dv/Dt for a co-moving
loop (see Section 2.10). Moreover, the equation of motion

of
an incompressible
inviscid fluid
can be written
[see
Equation (2.79)]
Dv
Dt
(5.20)
since is a constant. Hence,

IC
dC
dt
since
dv
d(v
2/2
and p/
2/2)
1 2
v
2
= (v 2/2)
dr
0,
(5.21)
dr (see Section A.18),
+ is obviously a single-valued
function.
One corollary of the Kelvin circulation theorem is that
the fluid particles that form the walls of
a vortex
tube at
given instance in time continue to form the walls of
vortex
imagine
tube
at all subsequent
times.
To
prove
this,
closed loop C that is embedded in the wall of
vortex tube but does not circulate around the interior of

Incompressible Inviscid FluidDynamics
81
Figure 5.3: A vortex tube.
Figure 5.3.
The normalcomponentof
the
vorticity over the surface enclosed by C is

zero, since all vorticity vectors are tangential to this surface.
Thus, from (5.17), the circulation around the loop is zero.
By Kelvins circulation theorem, the circulation around the
loop remains
zero as the
tube is convected by the fluid. In
other words, although the surface
as
enclosed by C deforms,
on
can pass through it.
it is convected by the fluid, it always remains
tube wall, since
no vortex
filaments
the
Another corollary of the circulation theorem is that the

intensity
of
a vortex
tube
by the fluid. This
convected
the circulation around the loop C
5.7
constant
remains
can be proved
as
it
is
by considering
pictured inFigure 5.3.
Irrotational Flow
Flow is said to be irrotational when the vorticity has the
zero
magnitude
everywhere. It immediately
follows, from
any arbitrary
an irrotational flow pattern is zero (provided that
loop can be spanned by a surface that lies entirely
Equation (5. 17), that the circulation around

loop in
the
within
the
theorem, if
fluid).
an
Hence,
from
Kelvins
remains irrotational
at all subsequent
seen more
from the equation
inviscid
circulation
inviscid fluid is initially irrotational then it
directly
incompressible
fluid
can
be
of motion of
an
times. This
which,
Equations (2.39)and (2.79), takes the form
according
to
(v
)v
(5.22)
since is a constant. However, from Equation (A.171),
(v
)v
= (v 2/2) v
(5.23)
Thus, we obtain
12
+v
(5.24)
Taking
the
curl of this
equation, and making
use
of the vector identities
0 [see Equation (A. 1

76)]
173)], as well
=0
in
an
as
A 0 [see Equation (A.
the identity (A.179), and the fact that v
incompressible
evolution equation
fluid,
we
obtain
D
Dt
the vorticity
)v.
(5.25)
Thus,
if
consequently,
82
0, initially,
then
=0 at allsubsequent
D/Dt
times.
0, and,
FLUID MECHANICS
Suppose
movable
that O is
point, in
an
fixed point, and P
irrotational
an
fluid. Let
arbitrary
O and P
be joined by two different paths, OAP and OBP (say). It

follows that OAPBO
circulation around
is
such
a closed curve. Now, since the

a curve in an irrotational fluid is
zero, we can write

ZOAP
ZPBO
v dr +
v dr = 0,
(5.26)
which implies that
OAP
ZO
ZOAP
dr
ZOBP
v dr = P
(5.27)
(say). It is clear that P is a scalar function whose value

depends
not
on the position
of P(and the fixed point O), but
on the path taken between
O and P.Thus, if O is the
origin of our coordinate system, and Pan arbitrary point

whose
position vector is r,then

R
we have
effectively
a scalar field
(r) =
v dr.
R
O
a point Q that is sufficiently close to Pthat
the velocity v is constant along PQ. Let be the position
defined
Consider
vector of Q relative to P.It then follows that (see Section

A.18)
Z
= Q
+ P
v dr v .
=
P
(5.28)
The above equation becomes exact in the limit that ||
0. Since Q is arbitrary (provided that it is sufficiently close
to P), the direction of the vector is also arbitrary, which

implies that
v = .
(5.29)
We, thus, conclude
that
if the
motion
irrotational then the associated velocity field

expressed
as
minus the gradient of
a fluid
can always
of
is
be
scalar function of
position, (r). This scalar function is called the velocity
a potential
as potential flow. Note that the velocity potential
is undefined to an arbitrary additive constant.
We have demonstrated
that a velocity
potential
necessarily
exists in a fluid whose velocity field is
irrotational. Conversely, when a velocity potential exists
potential, and flow which is derived from such
is known
the flow is necessarily irrotational. This follows because
[see Equa-tion (A.176)]
= 0.
(5.30)
Incidentally, the fluid velocity at any given point in an

irrotational fluid is normal to the constant- surface that
passes
through that point.
If a flow pattern is both irrotational and incompressible
then we have
v =
(5.31)
and
v = 0.
(5.32)
These two expressions
can be combined to give (see
Section A.21)
=0.
(5.33)
Inother words, the velocity potential in an incompressible
irrotational fluid satisfies Laplaces equation.
According to Equation (5.24), if the flow pattern in
an incompressible
inviscid fluid is also irrotational,
so
that
= 0 and v = ,
then
we can write
p
12 2
+
= 0,
(5.34)
which implies that
12
=C(t),
83
and
A
Figure 5.4: Two-dimensional flow.
where C(t) is uniform in space, but

fact, the time variation of C(t)
can vary in time. In

can be eliminated by adding
the appropriate function of time (but not of space) to the
a procedure
velocity potential, . Note that such
the instantaneous
velocity field
Thus, the above equation
can be rewritten
modify
12
does not
derived from .
=C,
(5.36)
where C is constant in both
space
and time. Expression
(5.36) is a generalization of Bernoullis theorem (see Section
5.3)
that
takes
non-steady
However, this generalization

flow. For the special
flow
into
account.
is only valid for irrotational
case of steady
flow, we get
12
= C,
(5.37)
which demonstrates
constant
in Bernoullis
streamlines.
5.8
that for steady
theorem
(See Section 5.3.)
Two-Dimensional Flow
irrotational
is the
flow the
same on
all
Fluid
motion
is said to be two-dimensional
when the
velocity at every point is parallel to a fixed plane, and is the
same
everywhere
on a given
Cartesian coordinates,
then we
normal to that plane. Thus, in
if the fixed plane is the
can express a general
inthe form
v =vx (x,y,t)ex +vy (x,y,t)ey
x-y
plane
two-dimensional flow pattern
(5.38)
Let A be a fixed point in the
ACP be two
arbitrary
curves,
also in the
x-y plane, and let ABP and

x-y plane, that join A to an
point P. See Figure 5.4. Suppose
that fluid is
nor destroyed in the region, R (say),

bounded by these curves. Since the fluid is incompressible,
which essentially means that its density is uniform and
neither
created
constant, fluid continuity requires that the rate at which the

fluid flows into the region R, from right to left
curve
across
the
ABP, is equal to the rate at which it flows out the of
across the curve ACP. Now,

across a surface is generally termed
the region, from right to left

the rate of fluid flow
the flux. Thus, the flux (per unit length parallel to the
z-axis) from right to left
from right to left
follows
curve
that
the
across
flux
across
ABP is equal to the flux
ACP. Since ACP is arbitrary, it
from
right
to left
across any
joining points A and Pis equal to the flux from right
to left
across
once the base point A has been

on the position of point P,
words, if we denote the flux by
ABP. In fact,
chosen, this flux only depends

and the time t.In other
then it is solely
a function
of the location of Pand the time.
(5 .44)
84
FLUID MECHANICS
P2
P1
Figure 5.5: Two-dimensional flow.
point Phas Cartesian coordinates
x,y,then we can write

= (x, y, t).
(5.39)
The function is known

the existence of
a stream
as the stream function. Moreover,

function is a direct consequence
of the assumed incompressible nature of the flow.

Consider
two points, P1 and P2
in addition to the
fixed point A. See Figure 5.5. Let 1 and 2 be the fluxes

from right to left
..
.
across curves
AP1 and AP2
Now, using
across
across AP1 plus the flux across
P1P2 Thus, the flux across P1P2 from right to left, is 2
1
Now, if P1 and P2 both lie on the same streamline then
the flux across P1P2 is zero, since the local fluid velocity is
similar arguments to those employed above, the flux

AP2
is equal to the flux
directed
everywhere parallel to P1P2
It follows that 1
we conclude that the stream function is

constant along a streamline. The equation of a streamline is
thus = c,where c is an arbitrary constant.
Let P1P2 = s be an infinitesimal arc of a curve that is
sufficiently short that it can be regarded as a straight-line.
The fluid velocity in the vicinity of this arc can be resolved
into components parallel and perpendicular to the arc. The
2
Hence,
component parallel to s contributes
across
the
perpendicular
arc
from
right
to
nothing to the flux
left.
The
component
to s contributes v s to the flux. However,
the flux is equal to 2
Hence,
2 1
(5.40)
In the limit s 0, the perpendicular
to left
across
velocity from right
ds becomes
ds
(5.41)
Thus,
in
infinitesimal
Cartesian
arcs
coordinates,
parallel to the
that
by
considering
x- and y-axes, we
deduce
vx
vy
(5.42)
,
.
(5.43)
These expressions
= ez
(5 .56)
85
Note that when the fluid velocity is written in this form

then
it
immediately
incompressibility
becomes
constraint
satisfied [since (A
and (A.176)]. It is also
clear
that
the
0 is automatically
B) rain
const
0see
t v
Equations
0 i s automa
(A.175)
tic
clear that the stream function is
undefined to an arbitrary additive constant.
The vorticity intwo-dimensional flow takes the form
ez
,
.
(5.45)
where
z
vy
vx
(5.46)
Thus, it follows from Equations (5.42) and (5.43) that
+
y
(5.47)
Moreover,
irrotational
two-dimensional
is
flow
characterized by
0.
(5.48)
When expressed
in terms of cylindrical
coordinates
(see Section C.3), Equation (5.44) yields
= vr (r, , t) er
(r, , t) e
(5.49)
where
vr
1r
Moreover, the vorticity is
=z
ez
(5.50)
(5.51)
where
!
1r
+
r
r2
(5.52)
5.9
Two-Dimensional Uniform Flow
Consider
flow pattern that is
steady two-dimensional
uniform:
i.e., such that the fluid velocity is the

the
x-y plane.
For instance,
suppose
same
everywhere
that the
common
in
fluid
velocity is
= V0
cos
ex +
V0
sin 0
ey
(5.53)
which corresponds
to flow at the uniform speed V0 in a
fixed direction that subtends
(counter-clockwise)
angle
0 with the x-axis. It follows, from Equations (5.42) and
(5.43), that the stream function for steady uniform flow

takes the form
(x, y)
= V0
(sin 0
x cos
y)
(5.54)
When written in terms
of cylindrical
coordinates,
this
becomes
(r, )
V0
sin(
0 ).
(5.55)
Note, from (5.54), that
2/x
= 2/y
=0
Thus, it follows from Equation (5.47) that uniform flow

is irrotational. Hence, according to Section 5.7, such flow
can also
be derived from
a velocity
potential. In fact, it
is easily demonstrated that

(r, )
(5.61) 86
FLUID MECHANICS
V0
cos(
0 ).
Figure 5.6: Streamlines of the flow generated by a line
source
5.10
coincident with the z-axis.
Two-Dimensional Sources and Sinks
Consider
uniform
line
source,
coincident
with the
z-axis, that emits fluid isotropically

Q unit volumes
we expect
the
associated
isotropic, and everywhere
source.
at the steady rate of
per unit length per unit

steady
directed radially
See Figure 5.6. Inother words,
v = vr (r)er
time. By symmetry,
flow
pattern
away
to be
from the
we expect
(5.57)
Consider
cylindrical
z-direction) and radius

In
steady
surface S of unit height (in the
r that
is co-axial with the
state, the rate at which fluid
source.
crosses
this
surface must be equal to the rate at which the section of
the source enclosed by the surface emits fluid. Hence,

ZS
dS
= 2rvr
(r)
=Q,
(5.58)
which implies that
vr (r)
2r
(5.59)
According to Equations (5.50) and (5.51), the stream
function associated with
a line source
coincident with the z-axis is
of strength Q that is
(r,)
2
(5.60)
Note that the streamlines,

from the z-axis,
stream
the
as illustrated
function
= c,are directed
away
in Figure 5.6. Note, also, that
associated
a line source is
cause any particular
with
multivalued. However, this does not

difficulty,
radially
since the stream function is continuous, and its
gradient single-valued.
Note, from Equation (5.60), that /r
2
0. Hence, according to (5.52), z =
=
=
= 0.In other
words, the steady flow pattern associated

with
line
source
is irrotational, and
can, thus,
a uniform
be derived from
velocity potential. In fact, it is easily demonstrated
that
this potential takes the form
lnr.
2
Incompressible
Inviscid Fluid Dynamics
(5.71)
87
(r,)
A uniform line sink, coincident with the z-axis,

which absorbs fluid isotropically at the steady rate of Q
unit volumes
per unit length per unit time has an associated
steady flow pattern
v =
er
2r
(5.62)
whose stream function is

(r,)
Q
.
(5.63)
This flow pattern is also irrotational, and
can be derived
from the velocity potential
(r,)
lnr.
2
(5.64)
source and a line sink of equal

run parallel to the z-axis, and are
located a small distance apart in the x-y plane. Such an
arrangement is known as a dipole or doublet line source.
Suppose that the line source, which is of strength Q, is
located at r = d/2 (where r is a position vector in the x-y
Consider
line
strength, which both
plane), and that the line sink, which is also of strength Q, is
located at
r = d/2.
Q
(r)
= 2
Let the function

Q
tan
2
1(y/x)
(5.65)
be the stream function associated with a line source of

strength Q located at r
0. Thus, Q (r r0 ) is the
a line source of
(r strength
r0
r = r0 Furthermore, the stream function
associated with a line sink of strength Q located at r = r0
is Q (r r0 ). Now, we expect the flow pattern
associated with the combination of a source and a sink to
be the vector sum of the flow patterns generated by the
source and sink taken in isolation. It follows that the overall
stream function is the sum of the stream functions
generated by the source and the sink taken in isolation. In
stream function associated with
Q located at
other words,
(r)
=Q (r d/2) Q (r + d/2) d Q (r),
(5.66)
to first order ind/r. Hence, if d =d(cos0 ex + sin0 ey )

= d[cos( 0 )er sin( 0 )e ],so that the line joining
=the
d[cos(
sink to
the
0 source
sin
subtends a (counter-clockwise) angle
0 with the x-axis, then
(r,)
(5.67)
D sin( 0 )
where D
=Qd is termedthe
strengthof the dipolesource.
Note that the abovestream functionis antisymmetricacross

the line
=0
joining the
source to the sink. It follows
that the associated dipole flow pattern,
D cos( 0 )
vr (r,)
r2
(5.68)
D sin( 0 )
(r,)
r2
(5.69)
across
is symmetric
streamlines
characterized
associated
by
this
> 0 and
dipole
linear
superposition
of
the
pattern
that the flow
2.
A dipole flow pattern is necessarily
shows
flow
= 0. Note
speed ina dipole pattern falls off like 1/r
it is
5.7
line. Figure
with
irrotational since
two irrotational
flow
patterns. The associated velocity potential is

(r,)
cos( 0 )
(5.70)
5.11
Two-Dimensional Vortex Filaments
Consider
a vortex
filament of intensity
that is
coincident with the z-axis. By symmetry,
we expect
the
associated
flow pattern to circulate isotropically around the
filament. See Figure 5.8. Inother words,
v =v (r)e
88
we expect
FLUID MECHANICS
1
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
0.8
1
1
0.5
0.5
x
Figure 5.7: Streamlines of the flow generated by
a dipole
line source coincident with the z-axis and aligned along
the x-axis. The flow is outward along the positive x-axis

and inward along the negative x-axis.
H
vH dr,
any closed curve in the x-y plane is equal to the
netarou
r,
intensity
nd any
ofclo
the
s vortex
ed curve
filaments
in the that
x- ypass
planthrough
e is eq the
ua
curve.
to the net
Consider a circular curve of radius r that is
Now, according to Section 5.5, the circulation,
around
concentric
with the origin. It follows that

I
r
v dr = 2rv
(r)
=,
(5.72)
or
v
(r)
2r
(5.73)
According
to Equations (5.50) and (5.5 1), the stream
function associated with
a vortex
that is coincident with the z-axis is

(r,)
lnr.
2
(5.74)
= c, circulate around the

as illustrated inFigure 5.8.
can be seen, from Equation (5.74), that
Note that the streamlines,

z-axis,
It
(/r)(r/r)
(5.52) that
pattern
= / = 0. Hence, it follows from

= 0. In other words, the flow
= 2
associated
with
irrotational. This is
vortex
straight
is
filament
somewhat surprising result, since
there is a net circulation of the flow around the filament,

and,
according
to
non-zero
implies
Section
vorticity.
5.5,
non-zero
The
circulation
can
a small,
paradox
resolved by supposing that the filament has
be
but
finite, radius. In fact, let the filament have the finite radius
a,
and be such that the vorticity
is uniform inside this
radius, and zero outside: i.e.,

(
z
/a
r a
r >a
(5.75)
Note that the intensity of the filament (i.e., the

product of its vorticity and cross-sectional area) is still .
According to Equation (5.52), and assuming that
=(r),
1r
r
(5.82)
dr
/a
r a
r> a
89
Figure 5.8: Streamlines of the flow
generated by a line vortex coincident with the z-axis.
The solution that is well-behaved

(up to its first derivative) at r
(
(r,)
(/4)(r
2/a 2
at
r = 0, and
continuous
=a, is
1)
r a
r> a
(/2) ln(r/a)
(5.77)
Note that this expression is equivalent to (5.74) (apart from
an unimportant
additive constant) outside the filament, but
differs inside. The associated circulation velocity, v (r)
/r, is
(
v
(r)
(/2)(r/a
2)
r a
r> a
(/2)(1/r)
(5.78)
whereas the circulation, r (r)

(
r
(r)
(r/a)
=2rv (r),is written

r a
r> a
(5.79)
Thus,
we
conclude that the flow pattern associated with
straight vortex filament is irrotational outside the filament,

but has finite vorticity inside the filament. Moreover, the
non-zero
internal
circulation
vorticity
generates
a constant
net
of the flow outside the filament. In the limit in
which the radius of the filament tends to zero, the vorticity

within the filament tends to infinity (insuch
a way that
area of
product of the vorticity and the cross-sectional

filament
the
the
remains constant), and the region of the fluid in
which the vorticity
is
non-zero
becomes
infinitesimal
in
extent.
Let us determined the pressure profile in the vicinity of
a vortex filament of finite radius. Assuming, from symmetry, that p = p(r), Equation (2.149), yields
2
dp
dr
(5.80)
which

Z
dr,
(5.81)
where
is
the
of expression (5.78),
pressure
we obtain
at
infinity.
p (/2)(/2a)
Making
(2
use
r 2 /a 2)
p(r)
p (/2)(/2a)
m )]
(a/r)
(5.92)
90
FLUIDMECHANICS
pressure occurs at
It follows that the minimum

of the vortex (r
the center
=0),and takes the value

?
?2
p0
2a
(5.83)
pressure in a fluid
that a vortex filament
Now, under normal circumstances, the
must remain positive, which implies

of intensity
, embedded
backgroundpressure p
a fluid
a minimum
in
has
and
of density
radius of order
!1/2
amin
(5.84)
Finally,
since
the
flow
pattern outside
vortex filament is irrotational, it

velocity
can
straight
be derived from
potential. In fact, it is easily demonstrated that the
appropriate potential takes the form
(r,)
2
(5.85)
Note that the above potential is multivalued. However, this

does not
cause any
particular difficulty, since the potential
is continuous, and its gradient single-valued.
5.12
Two-Dimensional
Irrotational
Flow in
we have seen, in a two-dimensional flow pattern, we

can automatically satisfy the incompressibility constraint,
v = 0, by expressing the pattern in terms of a
As
stream function. Suppose, however, that, in addition to

being
incompressible, the flow pattern is also irrotational.
Inthis case, Equation (5.47) yields
0.
(5.86)
In cylindrical
coordinates,
since
(r, , t), this
expression implies that (see Section C.3)

!
1r
+
r
=0.
(5.87)
Let
us
search for
separable
Equation (5.87) of the form
steady-state
solution of
(r, )
R(r) ().
(5.88)
It is easily
seen that
R dr
dR
1d
dr
(5.89)
which
can only
be satisfied if
!
dR
dr
R,
dr
d
(5.90)
(5.91)
m2
where
general
is
an
solution
combination
However,
of
arbitrary
Equation
of
constant.
(5.91)
and
exp(im)
assuming
()
and,
is
The
linear
factors
exp(im)
over
all
must be single-valued
in ,
that the flow extends
values, the function
otherwise
(positive)
, ,
hence,
vwould
be
must not
be sinbe
gle -val
follows
that
m can
ued
single-valued
in , otherwi
(which
se
is unphysical).
and he nceIt v
wo ul
dnot
be
only take integer values (and that

rather than
a negative,
of Equation (5.90) is
combination
be
positive,
constant). Now, the general solution
linear combination
factors, except for the special
linear
m 2 must
of
r0
of
rm
r m
it is a
and
case m = 0, when
r factors. Thus,
and ln
the
stream
general
irrotational
function
for
flow (that extends
steady
over
two-dimensional
all values of ) takes
the form
(r,)
from
=0
Xm>0
+ 0 lnr +
(m
r m + m r m)
sin[m(
m )],
91
where m
,,
,
and m
are
recognize the first few terms
arbitrary
on the
constants. We
can
right-hand side of the
zero gradient,
any flow. The term 0
by a vortex filament of
above expression. The constant term 0 has

and, therefore, does not give rise to
ln ris the flow pattern generated
intensity 20
11.) The
of
speed
coincident with the z-axis. (See Section 5.
term 1
1
sin(1 )correspondsto uniformflow
whose
direction
subtends
(counter-clockwise) angle 1 with the minus x-axis. (See

5.9.)
Section
Finally,
the
term 1
sin(
1 )/r
corresponds to a dipole flow pattern. (See Section 5.10.)

The velocity potential associated with the irrotational
stream function (5.92) satisfies [see Equations (5.29) and

(5.44)]
1r
1r
(5.93)
(5.94)
It follows that
(r,)
=0
Xm>0
0
rm
(m
r m)
cos[m( 0 )].
(5.95)
5.13
Inviscid Flow Past
a Cylindrical
Obstacle
Consider
the steady
impenetrable
uniformly
flow
pattern
rigid cylindrical
produced
obstacle
when
is placed
an
a
in
flowing, incompressible, inviscid fluid, with the
cylinder orientated such that its axis is normal to the flow.

For instance,
suppose
that the radius of the cylinder is
and that its axis corresponds

Furthermore,
magnitude V0
let the
to the line
unperturbed
fluid
a,
= y = 0.
velocity
be of
and be directed parallel to the x-axis. Now,
we expect the flow pattern to remain unperturbed very far

away from the cylinder. In other words, we expect v(r, )
farV0
away
ex as
fror/a
m t he
cylin
, which
der. Incorresponds
o the rw ords,
towe
a boundary
e xpect v(
r,)
condition on the stream function of the form (see Section
5.9)
(r,) V0
rsin
as r/a
(5.96)
Given
that
the fluid
velocity
field
large
upstream of the cylinder is irrotational (since

already
seen that
circulation
Section 5.9), it follows from the
theorem
(see Section
velocity field remains irrotational

the cylinder.
have
the flow pattern associated with uniform
flow is irrotationalsee
Kelvin
distance
we
as
5.6) that the
it is convected past
Hence, according to Section 5.8, the stream
function of the flow satisfies Laplaces equation,
= 0.
(5.97)
The appropriate boundary condition at the surface of the
cylinder is simply that the normal fluid velocity there be
zero,
since the fluid must
cylinder,
but
cannot
vr (a,)(1/a)/|r=a
stay in contact
penetrate
its
surface.
with the
Hence,
=0,which implies that

(a, )
0,
(5.98)
since is undetermined to an arbitrary additive constant. It
follows that
we are
searching for the most general solution
of (5.97) that satisfies the boundary conditions (5.96) and
(5.98). Comparison with Equation (5.92) reveals that this

solution takes the form
(r,)
=V0 a
ln
r
a
r
a
ar
sin
(5.99)
where
2aV0
(5.100)
and
is the circulation of the flow around the cylinder.
(Note that the velocity field
possess nonzero
loop that encloses

surface lying
can
be irrotational, but still
circulation around the cylinder, because

the cylinder cannot be spanned by
entirely
such
a loop
from Stokes theorem.)
0, for the sake of definiteness.

92
FLUID MECHANICS
5
4
zero fluid
zero circulation around
Let us assume that
within the fluid. Thus,
vorticity does not necessarily imply
a
a
3
2
1
y/a
0
1
2
3
4
5
5 4 3 2 1 0
x/a
Figure 5.9: Streamlines of the flow generated by a

cylindrical obstacle of radius
a,whose
z-axis, placed inthe uniform flow field

normalized circulation is
Figure
calculated
circulation,
5.95.
=0.
1 1show
for various different

. For
< 2 there
surface of the cylinder
axis
runs along the
v =V0 ex
streamlines
of
The
the
flow
values of the normalized

exist
a pair
of points
on the
zero.
at which the flow speed is
These
are known as stagnation points, and can be located in

as the points at which streamlines
Figures 5.9 and 5.10
of the cylinder at right-angles.
intersect the surface
Now,
the tangential fluid velocity at the surface of the cylinder is

?????r=a
vt
()
=v (a,) =
= V0
+ 2 sin).
(5.101)
The stagnation points correspond to the points at which vt
= 0 (since the normal velocity
is automatically
zero at
the
surface of the cylinder). Thus, the stagnation points lie at

1(/2).
sin
When > 2 the stagnation points coalesce
and move off the surface of the cylinder, as illustrated in

Figure 5.1
1
(the stagnation point corresponds to the point
ce
at which two streamlines cross at right-angles).
The irrotational
form of Bernoullis
theorem, (5.37),
can Th
be e
combined
with
the boundary
condition
v m
V0 (5
as
ir rot ation
al form
of Ber no
ull is theore
r/a , as well as the fact that is constant in the
present case, to give
p = p0 +
?V
?
2
0
(5.102)
where p0 is the constant static fluid
pressure a large
distance from the cylinder. Inparticular, the fluid
on the
= p(a,) = p0
?V
12
?
2
pressure
surface of the cylinder is
vt
=p1 + V
(cos2 2 sin ),
(5.103)
where p1
= p0 (1/2)V
The net force
2
0
(1+
per unit length
2).
exerted
on
the cylinder by the fluid has the Cartesian
components
Fx
Pcosad,
I
(5.104)
Fy
Psinad.
the fact Incompressible Inviscid FluidDynamics

93
5
4
3
2
1
y/a
0
1
2
3
4
5
5 4 3 2 1 0
x/a

a,whose

=1.
Thus, it follows from (5.103) that
axis
runs along
v =V0 ex
The
the
Fx
0,
(5.106) Fy
2V
a =V0
().
(5.107)
Now, the component
exerts
parallel
on an
of the force which
moving fluid
obstacle, placed in its path, in
to that of the unperturbed
direction
flow is usually called
drag. On the other hand, the component
of the force
which the fluid exerts in a direction perpendicular to that

of the unperturbed
above equations
placed in
zero
flow is usually called lift. Hence, the
imply that if
uniformly
cylindrical
obstacle
is
flowing inviscid fluid then there is
drag. On the other
hand,
as
long
as
there is net
circulation
of the flow around the cylinder, the lift is
non-zero.
Now,
lift
circulation
tends
to increase
is generated
because
the fluid
above, and to decrease
it directly
Thus, from Bernoullis
theorem,
(negative)
speed
below, the
the fluid
directly
cylinder.
pressure
is
decreased above, and increased below, the cylinder, giving

rise to a net upward force (i.e., a force inthe +y-direction).
Suppose that the cylinder is placed in a fluid which is
initially at rest, and that the fluids uniform flow velocity,

V0
no
is then
very
slowly ramped
up
(in such
a manner
that
vorticity is induced in the upstream flow at infinity).
Since the flow pattern is initially irrotational, and since the
flow pattern well upstream of the cylinder is assumed to
remain irrotational, the Kelvin circulation theorem indicates
pattern around the cylinder also remains
that
the flow
HC
Consider the time evolution of the circulation,
irrotational.
HC
dr, around
some
fixed
curve
C that lies entirely
within the fluid, and encloses the cylinder. We have

IC
dC
IC
"
dt
1 2
v
2
+v
IC
dr
dr =
dr,
(5.108)
where
use
has been made of (5.24) (with assumed
= z ez in two-dimensional flow,
= dS, where dS is an outward surface element
constant). However,
and dr
of
ez
unit depth (in the z-direction) surface whose normal
lies in the
x-y
plane, and that cuts the
x-y
plane at C. In
other words,
IS
dC
v dS.
dt
(5.109)
We, thus,
circulation
vorticity
conclude
that
the
rate of
change
of
the
around C is equal to minus the flux of the
across
S [assuming that vorticity is convected by
the flow, which follows from (5.25), the fact that

(5.111) 94
= z
FLUID MECHANICS
5
4
3
2
1
y/a
0
1
2
3
4
5
5 4 3 2 1 0
x/a

a,whose

=2.5.
axis
runs along
v =V0 ex
The
the
that /z
= 0 in two-dimensional
have already
seen
flow]. However,
that the flow field surrounding
cylinder
is irrotational (i.e., such that z
that C
is constant
originally,
we
the
= 0).It follows
=0
in time. Moreover, since C
because the fluid surrounding
we deduce
we conclude
the cylinder
was
= 0 at all subsequent
initially at rest,
that C
times. Hence,
that, in an inviscid fluid, if the
zero
zero. It follows, from the above analysis,
that, in such a fluid, zero drag force and zero lift force
are exerted on the cylinder as a consequence of the fluid
flow. This result is known as dAlemberts paradox, after
circulation of the flow around the cylinder is initially
then it remains
the French scientist

(17171783).
it
would
seem, at
approximation
Jean-Baptiste
DAlemberts
first
to neglect
le Rond d Alembert
result is paradoxical because

sight,
to be
viscosity
reasonable
alltogether
in high
we end
a high Reynolds
any force on an
Reynolds number flow. However, if we do this then
up
with the nonsensical
number
prediction that
fluid is incapable
of exerting
obstacle placed inits path.
5.14
Inviscid Flow Past
a Semi-Infinite
Wedge
Consider the situation, illustrated in Figure 5.12, in which
incompressible
Let the cross-section

both
on a impenean angle .
in the x-y plane be
irrotational flow is incident
apex
trable rigid wedge whose
z-independent
of the wedge
and
Furthermore, let the
symmetric
apex
about
of the wedge lie at
upstream flow
Finally, let the
subtends
the
x-axis.
= y = 0.
large distance from the
wedge be parallel to the x-axis.

Since the flow is two-dimensional,
and irrotational,
stream
it
can
that
function
incompressible,
satisfies
in terms of
be represented
Laplaces
equation.
Moreover, in cylindrical coordinates, this equation takes

the form (5.87). The boundary
function
conditions
on the stream
are
(r,/2)
= (r,2
/2)
= (r,) = 0.
(5.110)
The
two
first
normal velocity
boundary
conditions
ensure
that
at the surface of the wedge is
the
zero
The third boundary condition follows from the observation

that, by symmetry, the streamline that meets the
the wedge
apex
of
splits in two, and then flows along its top and
bottom boundaries, combined with well-known result that
is constant
(r,)
ona streamline.
r 1+m
It is easily demonstrated that
sin[(1 + m)(
)]
1
+m
(5.116) Incompressible Inviscid FluidDynamics
95
Figure 5.12: Inviscid flow past
is
a solution
a wedge.
of (5.87). Moreover, this solution satisfies the
boundary conditions provided (1+ m)(1 /2)
= 1,or
(5.112)
Since,
as
equation
is well-known,
correct
problems
(for
conditions)
are
unique,
solution
the
with
solutions
to Laplaces
well-posed
we can be sure
to the problem
boundary
that (5.111) is the
under
investigation.
According to this solution, the tangential velocity
on the
surface of the wedge is given by

vt
(r)
=Ar m
(5.113)
where
0. Note that the tangential velocity
is
zero at
the
m
apex of
the wedge. Since the normal velocity
is also
zero at this point, we conclude that the apex is a stagnation

point
zero of the flow. Figure 5 13 shows the streamlines of the
flow for the case = 1/2.
5.15
Inviscid Flow Over
a Semi-Infinite
Wedge
Consider the situation illustrated in Figure 5.14 in which
an incompressible
irrotational fluid flows
trable rigid wedge whose
apex
over an impenean angle .
subtends
Let the cross-section

both
of the wedge in the
z-independent
and
apex
Furthermore, let the

Finally, let the
symmetric
x-y
about
of the wedge lie at
upstream flow
plane be
the
y-axis.
= y = 0.
large distance from the
wedge be parallel to the x-axis.

Since the flow is two-dimensional,
and irrotational,
stream
function
that
boundary conditions
(r,[3 ]/2 )
can
it
be represented
satisfies
Laplaces
on the stream
= (r,[1
function
]/2 )
incompressible,
in terms of
equation.
The
are
= 0.
(5.114)
These boundary conditions
ensure
that the normal velocity
at the surface of the wedge is zero. It is easily demonstrated

that
(r,)
r 1m
cos[(1
m)( /2)]
(5.115)
is
solution of Laplaces
equation,
this solution satisfies the boundary

that (1 m)(1 /2)
(5.87). Moreover,
conditions
m
96
FLUID MECHANICS
provided
= 1/2, or
=
1
+
3
3
x
Figure 5.13: Streamlines of inviscid incompressible
irrotational flow past
a 90
wedge.
97
3
3
x
Figure 5.15: Streamlines of inviscid
incompressible irrotational flow
where
to
1 .
Laplaces
we can
again be
equation
sure
Since
are
over a 90
the
wedge.
solutions
unique,
that (5.115) is the correct
solution to
the
problem
under
investigation.
solution, the tangential
velocity
vt
this singularity in the flow
slightly
wedge. However,
rounding
is infinite
the
apex.
Figure
Velocity
5.15
=1/2.
streamlines of the flow for the case
Potentials
and
m,
(5.117)
have
seen, a
two-dimensional
shows
velocity
which the flow is everywhere parallel to the
the
Stream
Functions
we
(r)
apex of the wedge.

can be eliminated by
at the
speed,
However,
As
of the
0. Note that the tangential velocity, and hence
the flow
5.16
the surface
wedge is given by
where
to this
According
on
x-y
field in
plane, and
there is no variation along the z-direction, takes the form
= vx (x, y, t) ex
Moreover, if the flow is irrotational

automatically
satisfied by writing
+ vy (x,
then
= ,
y, t) ey
(5.118)
= 0 is
where (x,
y,
t)is termed the velocity potential. (See Section 5.7.) Hence,
vx
(5.119)
vy
(5.120)
,
.
On the other hand, if the flow is incompressible
0 is automatically
=(x,
0 isy,
a t) is termed
= z
then v
, where
the stream function. (See Section 5.8.)
Hence,
vy
(5.121)
vx
98
,
.
FLUIDMECHANICS
Finally, if the flow is both irrotational and incompressible

then Equations (5.119)(5.120)
and (5.121)(5.122) hold
simultaneously, which implies that
It immediately follows, from the previous two expressions, that
xy
yx
or
x 2
Likewise, it can also be shown that
+
y
= 0.
+
y
= 0.
(5.123)
(5.124)
(5.125)
(5.126)
(5.127)
We conclude that, for two-dimensional, irrotational, and
incompressible flow, the velocity potential and the stream
function both satisfy Laplaces equation. Equations
(5.123) and (5.124) also imply that
=0:
(5.128)
i.e., the contours of the velocity potential and the stream
function
5.17
cross at right-angles.
Exercises
5.1. Liquid isled steadily through
a pipeline
that
passes over a
hill of height hinto the valley below, the speed at the crest being
v.Show
cross-sectional
that, by properly adjusting the ratio of the
areas
of the pipe at the crest and inthe valley, the
pressure
may be equalized at these two places.
5.2. For the case of the two-dimensional motion
of an
incompressible fluid, determine the condition that the velocity
components
=
=
vx
vy
a x + by,
c x + dy
satisfy the equation of continuity. Show that the magnitude

of the vorticity is
c b.
5.3. For the case of the two-dimensional motion of an

incompressible fluid, show that
vy
vx
c(a
2c xy,
2
are the velocity components
+x 2
y )
of a possible flow pattern.
Determine the stream function and sketch the streamlines. Prove

that
the motion is irrotational, and find the velocity potential.
5.4. A cylindrical vortex inan incompressible fluid is co-axial

with the z-axis, and such that z takes the constant value for
r a, and
is
zero
for
r > a, where r is a cylindrical
coordinate. Show that
1 dp
dr
a4
where p(r) is the pressure at radius
r inside
the vortex, and
the circulation of the fluid outside the vortex is 2. Deduce that
p(r)
r2
2a 4
+ p0
where p0 is the pressure at the center of the vortex.
99
5.5. Consider the cylindrical vortex discussed inExercise 5.4.
If p(r) is the pressure at radius
r external to the vortex,
demonstrate
that
p(r)
where
2r 2
+ p
is the pressure at infinity.
5.6. Show that the stream function for the cylindrical vortex
discussed inExercises 5.4 and 5.5 is (r)
=(1/2)a
ln(r/a)
for
r > a,and (r) = (1/4)(r 2 a 2) for r a.

a volume V whose boundary is the surface S.
Suppose that V contains an incompressible fluid whose motion
5.7. Consider
is
irrotational. Let the velocity potential be constant
over
S.Prove that has the same constant value throughout V.[Hint:
Consider the identity
(AA)
A + A 2 A.]
5.8. In Exercise 5.7, suppose that, instead of taking
on the boundary,
zero
constant value
everywhere
the normal velocity is
on the boundary.
Show that is constant throughout V.
a liquid the
any small circle of radius r
5.9. Prove that inthe two-dimensional motion of
mean tangential
fluid velocity around
is
r, where 2 is the value of
vy
vx
at the center of the circle. Neglect terms of order

5.10. Show that the equation of continuity for the
two-dimensional motion of an incompressible fluid
r3
can be
written
(rvr )
+
r
= 0,
r, are cylindrical coordinates. Demonstrate that

vr = akr n exp[k(n + 1)] and
n
v = ar exp[k(n + 1)]. Determine the stream function,
show that the fluid speed at any point is
where
this equation is satisfied when
and
1
(n + 1)
+k 2 /r,
where is the stream function at that point (defined such
that
= 0 at r= 0).
5.11. Demonstrate that streamlines
cross at right-angles at a
stagnation point intwo-dimensional, incompressible, irrotational

flow. 5.12. Consider two-dimensional, incompressible, inviscid
flow. Demonstrate that the fluid motion isgoverned by the

following
equations:
+ [,]
where
(1/2)v
v = ez
, [A,B]
+ .
0,
= ez
B, and
= p/ +
5.13. For irrotational, incompressible, inviscid motion in
two-dimensions show that
where
q = |v|.
100
q = q 2q,
FLUID MECHANICS
2D
Potential
Flow
101
6 2D Potential Flow
6.1
Introduction
This chapter
discusses
the
use
of complex
analysis
to
simplify calculations
in-viscid,
in two-dimensional,
irrotational,
incompressible,
fluid
inviscid,
incompressible,
dynamics.
irrotational
Incidentally,
is
flow
usually
referred to a potential flow, since the velocity field
represented
Laplaces
in terms of
equa-tion. In the
can be
velocity potential that satisfies

following, all flow patterns
assumed to be such that the z-coordinate
are
is ignorable. In
other words, the fluid velocity is everywhere parallel to the
= 0. It follows that all line sources and

run parallel to the z-axis. Moreover, all
solid surfaces are of infinite extent along the z-axis, and
have uniform cross-sections. Hence, it is only necessary to
specify the locations of line sources, vortex filaments, and
solid surfaces inthe x-y plane.
x-y plane,
and /z
vortex filaments
6.2
Complex Functions
The complex variable is conventionallywritten
z = x + iy,
(6.1)
where i
represents the
and
y are
corresponding
both
square root
and are
real,
Cartesian
should not be confused

strictly two-dimensional
of minus
identified
coordinates.
with
one.
Here,
with
(Incidentally,
z-coordinate:
the
this is
z
a
discussion.) We can also write
= re
where
r=
px
px
(6.2)
2
+ y 2 and
modulus and argument of
=tan 1(y/x)
z,respectively,
identified with the corresponding
are termed the

but can also be
plane polar coordinates.
Finally, DeMoivres theorem,
= cos + isin ,
(6.3)
implies that
=
=
x
(6.4) y
r cos,
r sin.
(6.5)
can define functions of the complex variable, F(z),

like we would define functions of a real variable. For
We
just
instance,
(6.6)
F(z)
F(z)
z2
1z
,
.
(6.7)
For
given function, F(z),
we can
substitute
= x + iy
and write
F(z)
= (x,
y)
i(x, y),
(6.8)
where
(x,
y) and
functions. Thus, if
(x,
y)
are
real two-dimensional
F(z)
z2
(6.9)
then
F(x
+ iy) =
(x
iy)
= (x
2)
2i
x
y,
(6.10)
giving
x2
(x,y)
(6.11) (x,y)
2 xy.
102
FLUID MECHANICS
6.3
We
the
Cauchy-Riemann Relations
can define the derivative of a complex function in just

same manner that we would define the derivative of a
real function: i.e.,

dF
F(z
+ z)
lim|z|
dz
F(z)
(6.13)
However,
we now
well-behaved
differentiable)
have
function
then
of
the limit
If
F(z)
is
(i.e., finite, single-valued,
plane
in
course, many different

we look at a regular
but if
problem.
it should
direction in the complex

when taking
not
we
matter
complex
and
which
z
are,
approach the point
Equation
possible
from
(6.13).
approach
There
directions,
function, F(z)
=z
(say), then
dF
dz
2z
(6.14)
is perfectly
well-defined,
independent
of the details of how the limit is taken in
and is, therefore,
completely
Equa-tion (6.13).
same
no matter from which direction we approach z,
means that there are some restrictions on the forms of the
The fact that Equation (6.13) has to give the
result,
functions (x, y) and (x, y) in Equation (6.8). Suppose
that
we
approach
z along
the real axis, so that z
= x. We
obtain
dF
dz
(x
lim|x|0
+i
+ x,y) + i(x + x,y) (x,y) i(x,y)
x
(6.15)
we now approach z along

so that z =iy. We get
Suppose that
dF
the imaginary axis,
dz
(x,y
+ y) + i(x,y + y) (x,y) i(x,y)
lim|y|0
= i
+
y
iy
(6.16)
But, if F(z) is a well-behaved function then its derivative
must be well-defined, which implies that the above two
are equivalent.
expressions
(6.17)
This requires that
(6.18)
These expressions
are called
the Cauchy-Riemann
and are, in fact, sufficient to ensure that all

possible
ways
of taking the limit (6.13) give the same
result.
6.4
Note
Complex Velocity Potential

that
Equations
Equations
(6.17)(6.18)
(5.123)(5.124).
and imagi-nary parts of

complex
variable
potential
and
can
stream
are
This suggests
well-behaved
be interpreted
function,
identical
to
that the real
function of the
as
the
respectively,
velocity
of
some
two-dimensional, irrotational, incompressible
suppose
For instance,
F(z)
= V0
flow pattern.
that
z,
(6.19)
where V0 is real. It follows that
(r,)
V0
(6.20) (r,)
rcos,
V0
rsin.
(6.21)
It can be
seen, by
comparison with the analysis of Section
5.9, that the complex velocity potential (6.19) corresponds
to uniform flow of speed V0

Furthermore,
velocity
as
is easily
along the x-axis.

the
complex
potential associated with uniform flow of speed V0
whose direction subtends
with the x-axis is F(z)
2D
directed
demonstrated,
Potential
= V0
Flow
(counter-clockwise)
ze i0
angle 0
103
Suppose that
F(z)
lnz,
2
(6.22)
where Q is real. Since lnz
=lnr +i, it follows that
(6.23)
(r,)
lnr,
2
Q
(r,)
2
(6.24)
to the analysis
Thus, according
to the flow
complex velocity potential (6.22) corresponds
pattern
of
line
source,
of strength Q, located at the
origin. (See Figure 5.6.) As

this result, the
complex
5.10, the
of Section
at the point
strength
Q, located
(Q/2)
ln(z z0 ), where
z0
,.
simple generalization
potential
of
(x0
= x0
line
of
y0 ), is F(z)
+ iy0
Note, from
(6.22), that the complex velocity potential of

is singular at the location of the
of
source,
a line source
source.
Suppose that
F(z)
lnz,
2
(6.25)
where is real. It follows that
(r,)
2
(6.26)
(r,)
lnr.
(6.27)
Thus, according
to the analysis
to the flow
a vortex
pattern
of
origin.
(See
located at the
Figure
complex velocity
5.11, the
of Section
5.8.) Note, from (6.25),
potential of
a vortex
that
the
filament is singular
at the location of the filament.

Suppose, finally, that
F(z)
where V0
z+
+i
ln
(6.28)
a,and , are real. It follows
(r,)
(6.29)
= V0
a2
r+
V0
(r,)
that
V0
a2
a2
sin
cos
2
?
ln
(6.30)
Thus, according
to the analysis
of
Section

inviscid
flow of unperturbed
the x-axis, around

centered
on
an
impenetrable
Here, is the circulation
=0
on the
to uniform
running parallel to
cylinder of radius
a,
(See Figures 5.9, 5.10, and 5.11.)
the origin.
Note that
speed V0
5.13, the
of the flow about the cylinder.
surface of the cylinder (r
= a),
which
ensures that the normal velocity

as must be the case if the cylinder
surface,
6.5
is
zero on
this
is impenetrable.
Complex Velocity
It follows from Equations (5.119), (5.122), and (6.15) that
dF
dz
+i
= vx + ivy
x
(6.31)
Consequently, dF/dz is termed the complex velocity. Note
that
???
???
dF
dz
=v x2
+v y
=v 2
point 104
FLUID MECHANICS
where
v is the flow
A stagnation
speed.
as a point
v, falls to zero.
point is defined
pattern where the flow speed,
in
flow
It follows,
from the previous expression, that
dF
dz
(6.33)
at
a stagnation
point. For instance, the stagnation points
of the flow pattern produced when
a, centered on the origin,
of radius
flow of speed V0
a cylindrical
directed parallel to the x-axis, and the
circulation of the flow around is cylinder is ,

by setting the derivative of the complex
to
zero.
obstacle
is placed in a uniform
It follows that the stagnation
are
found
potential (6.28)
points satisfy the
quadratic equation
dF
= V0
dz
a2
+i
2z
=0.
(6.34)
The solutions
are
z
a
q1
= i
(6.35)
where
= /(4V0
a), with the proviso that |z|/a
since the region |z|/a
<
1is occupied
p1
>
1,
by the cylinder.
stagnation points on the

p1
surface of the cylinder at x/a
2 and y/a
.
p1
On the other hand, if > 1
then there is a single stagnation
Thus,
point
p
if
below
1.
1
then there
are two
the cylinder
at x/a
=0
and y/a
Now, according to Section 5.7, Bernoullis
an steady, irrotational,
p+
12
v
= p0
theorem in
incompressible fluid takes the form
(6.36)
where p0 is a uniform constant, and where gravity (or
other body force) has been neglected. Thus, the
distribution insuch a fluid
any
pressure
can be written
???
12
p0
???
dF
dz
(6.37)
6.6
Method of Images
Let F1(z)
= 1 (x,y) + i1 (x,y) and F2 (z) = 2 (x,y) +
i
2 (x,y)be complex velocity potentials corresponding to
distinct,
two-dimensional,
flow patterns
whose
irrotational,
stream functions
incompressible
are
1 (x, y) and
2 (x,y), respectively. It follows that both stream functions
= 0.
= F1(z) +F2 (z). Writing F3 (z) = 3 (x,
y) + i3 (x,y), it is clear that 3 (x,y) = 1 (x,y) + 2 (x,
=
+
= 0, so 3 also
y). Moreover,
satisfy
Laplaces
equation:
i.e.,
Suppose that F3 (z)
satisfies Laplaces equation. We deduce that two complex
velocity
potentials,
two-dimensional,
patterns,
potential
can
to another
incompressible
below, this idea
can
distinct,
incompressible
to produce
be superposed
that corresponds
irrotational,
to
corresponding
irrotational,
flow
flow
third velocity
two-dimensional,
pattern. As
described
be exploited to determine
the flow
patterns produced by line sources and vortex filaments inthe

vicinity of rigid boundaries.
As
an example,
sources of
two line
consider
situation in which there
are
strength Q located at the points (0, a).
See Figure 6.1. The complex
resulting flow pattern is the

a).
velocity
potential of the
sum of the complex
potentials
of each source taken in isolation. Hence, from Section 6.4,
F(z)
ln(z ia)
2
ln(z
2
+ ia) =
ln(z
+ a 2).
(6.38)
Thus, the stream function of the flow pattern (which is the
imaginary part of the complex potential) is
(x,y)
tan
2
2 xy
x2
y2
+a 2
(6.39)
Note that (x,0)
=0,which implies that there is zero flow
y = 0.Hence, inthe region y >0,we

as that
by a single line source of strength Q,located at
normal to the plane
could interpret the above stream function

generated
(6.44)
2DPotential Flow
105
x
a
Figure 6.1: Two line sources.
(0, a), in the
presence
of
a planar
y = 0.
rigid boundary at
2
This follows because the stream function satisfies

everywhere
in the
=0
y > 0, has the requisite

to a line source of strength Q)
region
singularity (corresponding
at (0,a), and satisfies the physical boundary condition that

velocity be zero at the rigid boundary.
as is well-known, the solutions of Poissons
equation are unique. The streamlines of the resulting flow
pattern are shown in Figure 6.2. Incidentally, we can think
of the two sources in Figure 6.1
as the images of one
the
normal
Moreover,
another
in the boundary
plane. Hence, this method
of
calculation is usually referred to as the method of images.

Now,
the
complex
velocity
associated
with
the
complex velocity potential (6.38) is
dF
dz
z +a2
2
(6.40)
Hence, the flow speed at the boundary is
???
???
dF
(x,0)
dz
|x|/a
a 1
+x 2 /a 2
(6.41)
It follows from (6.37) (and the fact that the flow speed at
excess pressure on the

to the presence of the source, is
infinity is zero) that the
due
boundary,
???
p(x,0)
???
dF
dz
?2
(1+ x 2 /a 2) 2
y=0
x 2 /a 2
(6.42)
Thus, the
acting
excess
on the boundary
Z
Fy
per
force
inthe y-direction is
unit length (in the z-direction)
p(x,0)dx
(1+ 2 ) 2
4a
(6.43)
The fact that the force is positive implies that the boundary
is attracted to the source, and vice

As
second
example,
versa.
consider
the
filaments of
intensities
before, the complex
situation,
are two vortex
in Figure 6.3, in which there
illustrated
and situated at (0, a). As

velocity
potential
of the resulting
flow
Aspattern
before, is
thethe
complex
sum ofvelocity
the complex
potential
potentials
of the reof each
filament taken in isolation. Hence, from Section 6.4,
F(z)
106
=i 2
ln(z ia) i
ln(z
+ ia) = i
z ia
ln
z + ia
FLUID MECHANICS
y/a
x/a
Figure 6.2: Stream lines of the 2D flow pattern due to a
line source at (0,a) inthe presence of a rigid boundary
at
y =0.
y
aa
"
Flow
107
(r/a a/r) sin 2DPotential
y/a
x/a
vortex filament at (0,a) in the presence of a rigid boundary
at
y =0.
Thus, the stream function of the flow pattern is
"
(x,y)
ln
x 2 + (y a) 2
x 2 + (y + a) 2
(6.45)
= 0, which implies that there is zero

flow normal to the plane y = 0. Hence, in the region y >
As before, (x, 0)
0,
we
could interpret the above stream function
generated
by
single
vortex
located at the point (0, a),inthe
as
that
filament
of intensity
presence
of
a planar
rigid
y = 0. The streamlines of the resulting flow

are shown in Figure 6.4. We conclude that a vortex
filament reverses its sense of rotation (i.e., ) when
reflected in a boundary plane.
As a final example, consider the situation, illustrated
in Figure 6.5, in which there is an impenetrable cylinder
of radius a, centered on the origin, and a line source of
strength Q located at (b, 0), where b > a. Consider the
boundary at
pattern
so-called
analog problem, also illustrated in Figure 6.5, in
a source of strength Q,
c < a, and a source of strength Q,
located at the origin. We can think of these two sources as
the images
of the external source in the cylinder.
Moreover, given that the solutions of Poissons equation are
unique, if the analog problem can be adjusted in such a
manner that r = a is a streamline then the flow in the
region r > a will become identical to that in the actual
which the cylinder is replaced by

located at (c, 0), where
problem.
Now,
the complex
velocity
potential
in the
analog problem is simply

Q
ln(z b)
2"(z
2
#
Q
b)(z c)
ln
F(z)
ln(z c) +
lnz
Hence,
writing
re
function takes the form
(6.46)
the
stream
corresponding
(r,)
tan
(r bc/r) sin
"
(r + bc/r) cos (b + c)
(6.47)
Now,
we require
the surface of the cylinder,
r = a, to be a
streamline: i.e., (a, )

constant. This is easily achieved
by setting c
a 2/b. Thus, the stream function becomes
"
(r,)
= 2
tan
(r/a a/r) sin
(r/a
+ a/r) cos
(b/a
+ a/b)
(6.48)
The corresponding
cylinder
are shown
streamlines in the region external to the

inFigure 6.6.
108
FLUID MECHANICS
x
c
Figure 6.5: A line source inthe presence of

impenetrable cylinder.
an
y/a
4
4
x/a
line source at (2a, 0) inthe presence of

of radius
a centered on the origin.
2DPotential Flow
109
a rigid cylinder
x
dz
z0
dz
d
0
Figure 6.7: A conformal
6.7
Conformal Maps
Let
x +
z=
x, and y are
= f(z), where f is a
+ i
and
iy, where , ,
that
Suppose
single-valued,
We
map.
can
z-plane
non-singular,
think of
point (, )
= f(z) as
z = z0
-plane.
We have
z-plane
a map
Let
(i.e.,
function.
from the complex
maps to a
every
point
corresponding
in the complex -plane. Moreover, if f(z) is
well-behaved
unique, and also has

point
and differentiable)
to the complex -plane. In other words,
(x, y) in the complex
indeed
real.
well-behaved
function
in the z-plane
us
then
this
mapping
unique inverse. Suppose
investigate
maps to
the point
how neighboring
is
that the
= 0
points
in the
map.
+ d
(6.49)
=
+ d
+ dz ),
= f(z0 + dz
f(z0
).
(6.50)
Inother words, the points z0

complex z-plane
+ dz
and z0
map to the points
+ dz
+ d
inthe
and 0
|,
inthe complex -plane, respectively. Now, if |dz
+ d
|dz
1
then
d
f (z0 )dz
(6.51) d
f (z0 )dz
(6.52)
where f (z)
=df/dz. Hence,
d
d
dz
dz
(6.53)
Thus, it follows that
|d
|d
|
|
|dz
|dz
|
|
(6.54)
and
arg(d
arg(d
=arg(dz ) arg(dz ).
(6.55)
Now, we
connecting
the point
can think
of dz
neighboring
z = z0
points in the complex z-plane to
Likewise, d
vectors
connecting
complex
-plane.
as infinitesimal vectors
and dz
the
and d
are
infinitesimal
points
in
= z0
the
two
It is clear, from the previous
equations, that, in the vicinity of
the mapping
z-plane to the complex -plane is such
from the complex
that the lengths of dz
same
corresponding
and dz
expand
or contract
by the
factor, and the angle subtended between these two
vectors remains the
same.
See Figure 6.7. This type of
mapping is termed conformal.
110
FLUID MECHANICS
Suppose
that
follows that
F()
and
=2 =
stream
two-dimensional,
where and
(, )
(, ) and (, )
potential
i(, ) is
can
function,
inviscid,
are
= 0. Hence, the functions

as
be interpreted
incompressible
Cartesian
e(x,y)+i
2e
the functions
velocity
2e
some
pattern,
=
= F[(f(z)] =
G(z) is also well-behaved.
= e
e(x,
e(x,y)+i
y) ande(x,y
e ),
(x,
y)
potential
of
flow
coordinates. However, if
e(x,y), where
It follows
e(x,y
that)+i
e(
=x,y)

the velocity
respectively,
f(z), where f(z) is well-behaved, then F()

G(z)
function
ofe the
complex
variable
. Iti
Suppos
tha t F
( ) (,
) + i(,)
well-behaved
and stream
= 0.In other words,
can beeinterpreted as
e
function,
respectively,
the
of
some new,
e( x,y) inviscid,
e
two-dimensional,
incompressible
flow
x and y are Cartesian coordinates. In other

we can use a conformal map to convert a given
pattern, where
words,
two-dimensional,
inviscid,
incompressible
flow
into another, quite different, pattern. Note,
pattern
incidentally,
that a conformal map converts a line source into a line

source of the same strength, and a vortex filament
into a vortex
filament
of the same intensity.
(See Exercise 6.12.)
As an example, consider the conformal
map
ie
z/a
(6.56)
= re i, it is easily demonstrated that x = a ln

r/ and y = a (/ 1
/2). Hence, the positive -axis ( =
0) maps to the line y = a/2, the negative -axis ( = )
maps
to the line y = a/2, and the region > 0 (0
0)
Writing
maps to
points

)
1/2). See
region
< y < a/2. Moreover, the

map to the points z = a (0,1/2
6.8. As we saw in Section 6.6, in the
the region a/2
= (0, 1)
Figure
> 0,the velocity
potential
= i 2
i
+i
F()
ln
(6.57)
corresponds to the flow pattern generated by a vortex
filament of intensity , located at the point
the
presence
of a rigid plane at
?
G(z)
=F(ie z/a) =i
lntanh
= (0,1),in
= 0.Hence,
2a
(6.58)
to the flow pattern generated
corresponds
filament
of intensity
presence
of
two rigid planes at
because the line

point
= (0, 1)
0 corresponds
correspond
, located
= 0 is mapped
by
a vortex
at the origin, in the
= a/2.
to the lines
This follows
y = a/2,
the
=
to
is mapped
a streamline
t o the
then
origin,
the lines
and if
y th
=ea/2
line
also
=
is mapped to the origin, and if the line
to streamlines. The stream function associated
with the above complex velocity potential,

#
"
(x,y)
cosh( xa) cos(y/a)
ln
cosh( x/a)+ cos(y/a)
(6.59)
is shown inFigure 6.9.
As a second example, consider the map
=z
(6.60)
This
maps
the positive -axis to the positive x-axis, the
negative -axis to the positive y-axis, the region
> 0 to
x > 0, y > 0, and the point = (0, 2 a 2) to the

z = (a, a). As we saw in Section 6.6, in the region
the region
point
>0, the velocity
potential
F()
ln(
+ 4a 4),
(6.61)
to the flow pattern generated
corresponds
source
inthe
of strength Q, located at the point
presence
of
by
line
= (0, 2a 2),
a rigid plane at = 0.Thus, the complex
velocity potential
G(z)
=F(z 2) =
ln(z
+ 4a 4),
(6.62)
corresponds to the flow pattern generated by a line source
of strength Q, located at the point
presence
z =(a,a),inthe
of two orthogonal rigid planes at y
= 0 and x =0.
The stream function associated with the above complex

potential,
"
(x,y)
tan
2
4 xy(x
x4
2DPotential Flow
111
6 x 2 y 2 +y 4
2)
+ 4a 4
B
E
A
DE
1
C
Figure 6.8: The conformal
map =ie z/a.
1
0.8
0.6
0.4
0.2
y/a
0
0.2
0.4
0.6
0.8
1
1
0.5
0.5
x/a
vortex filament at the origin in the presence of two

rigid planes at
112
y = a/2.
FLUID MECHANICS
3
y/a
0
0
x/a
line source at (a,a) inthe presence of two rigid planes at
x = 0 and y =0.
is shown inFigure 6.10.

As a final example, consider the map
l
z= +
where l
is real and positive. Writing
2l
(6.65) y
=re
(6.64)
we find that
cosh[ln(r/l)]
cos,
2lsinh[ln(r/l)] sin.
(6.66)
map converts the circle

where a > l,into the ellipse
Thus, the
-plane,
=a
!2
!2
2lcosh[ln(a/l)]
in the
2lsinh[ln(a/l)]
(6.67)
inthe z-plane. Note that the center of the ellipse lies at the
origin, and its major and minor
and the
y-axes,
respectively. As
axes run parallel to the xwe saw in Section 6.4, in
the -plane, the complex velocity potential
V0
a2
(6.68)
corresponds
to uniform flow of unperturbed
speed V0
a circular cylinder
a, centered on the origin. Thus, assuming that a
running parallel to the -axis, around

of radius
> l,in the

flow
to uniform
z-plane the potential corresponds
of unperturbed
speed V0
running
parallel
x-axis (which follows because at large |z| the
map
to the
(6.64)
z = , and so the flow at large distances from the

origin is the same in the complex z- and -planes), around
elliptical cylinder of major radius 2lcosh[ln(a/l)] = a
reduces to
+ l2/a,
aligned along the x-axis, and minor radius 2l

2/a,
l
aligned along the y-axis. The
sinh[ln(a/l)] a
corresponding stream function inthe z-plane is
(x,y)
(6.76)
113
2D
Potential
Flow
= V0
a2
r
sin,
y/l
4
4
x/l
Figure 6. 11:
Stream lines of the 2D flow pattern due to
uniform flow parallel to the x-axis
cylinder.
where
around
an
elliptical
r
(6.70)
l
exp(cosh
tan
x 2/l 2 + y 2/l 2 + 4 +
?[x
2/l 2
+ y 2/l 2
+ 4] 2
p
[p 2 1] 1/2
16
(6.71)
p),
yx
x 2/l 2
?1/2 1/2
(6.72)
Figure 6.1 1
shows the streamlines
calculated for a
6.8
=1.5l.
Complex Line Integrals
Consider the line integral of
complex
closed
of the flow pattern
variable
curve
taken
some
function F(z) of the
(counter-clockwise)
around
C inthe complex plane:

IC
F(z)dz.
(6.73)
Since dz
=dx + idy,and writing F(z) = (x, y) + i(x, y),
where (x,y) and (x,y)
=Jr
+ iJi
(6.74)
we can
IC
(dx dy),
IC
Ji
+ dy).
(dx
(6.75)
also write the above expressions
two-dimensional vector form
Jr
(6.88) 114
FLUID MECHANICS
where dr
it follows that J
Jr
However,
are real functions,
where
in the
IC
dr,
IC
Ji
B dr,
= (dx, dy), A = (, ), and B = (, ). Now,
according to Stokes
IC
ZS
A dr
A)z dS,
IC
B dr
ZS
B)z dS,
where S is the region of the x-y plane enclosed by C.
Hence,
we obtain
!
ZS
Jr
ZS
Ji
dS,
Let
IC
dS.
IC
F(z)dz,
F(z)dz,
(6.77)
theorem (see Section A.22),

(6.78)
(6.79)
(6.80)
(6.81)
(6.82)
(6.83)
where C
is a closed
curve
inthe complex plane that
completely surrounds the smaller curveC. Consider

J
=J J
(6.84)
Writing J
= Jr
+ iJi a direct
generalization
of the
previous analysis reveals that
ZS
Jr
dS,
y
ZS
(6.85)
Ji
where
lying
S is
now
between
Suppose
that
single-valued,
the
F(z)
and
region
curves
the
is
of
follows
components,
and
the
and
that
its
(6.86)
x-y
well-behaved
differentiable)
immediately
dS,
plane
(i.e.,
real
and
respectively,
finite,
throughout
imaginary
satisfy
It
the
relations,
Cauchy-Riemann
(6.17)(6.18),throughout
S.
However, if this is the case
then it is apparent, from the previous two expressions,

that
Jr
well-behaved
Ji
0. In other
throughout S then J
words,
if F(z) is
=J .
some
The circulation of the flow about
closed
curve
in the x-y plane is defined

IC
IC
(vx dx + vy dy)
dF
= Re
dz,
dz
(6.87)
where F(z) is the complex

flow, and
the
use
circulation
the
performed
potential
has been made of Equation (6.31)
can
be evaluated
integral in the complex

from
velocity
previous
around
by performing
z-plane. Moreover,
discussion,
any
loop
this
that
as
integral
can
of the
Thus,
line
is clear
can
be
be continuously
deformed into the loop C whilst still remaining in the fluid,
and not passing
over a singularity
of the complex velocity,
dF/dz.
6.9
Theorem of Blasius
Consider
some
flow pattern in the complex z-plane that is
specified by the complex velocity potential F(z). Let C be
some
closed
curve
in the complex
z-plane.
The fluid
pressure on this curve
is determined from Equation (6.37),
which yields
???
1
p0
???
dF
dz
(6.95) 2DPotential Flow
115
Pdx
dx
Pdy
dy
dl
Pdl
Figure 6.12: Force acting
us
Let
section of a curve.
evaluate the resultant force (per unit length), and
moment
the resultant
fluid
across a short
within the
(per unit length), acting
curve as a consequence
of this
on the
pressure
distribution.
a small element of the curve C, lying

x, y and x + dx, y + dy, which is sufficiently
short that it can be approximated as a straight line. Let P
be the local fluid pressure on the outer (i.e., exterior to the
Consider
between
curve) side
the
of the element. As illustrated in Figure 6.12,
pressure
force
(per
unit
length)
acting inward (i.e., toward the inside

the
element
a component
has
of the curve)
Pdy
in the
across
minus
x-direction, and a component Pdx in the plus y-direction.

Thus, if X and Y
force
(per
unit
are
the components
length)
in the
x-
of the resultant
and
y-directions,
respectively, then
dX
(6.89) dY
=
=
Pdy,
Pdx.
(6.90)
The
pressure
force (per unit length)
element also contributes to
a moment
acting
across
the
(per unit length), M,
acting about the z-axis, where
dM
xdY
ydX
P(xdx
ydy).
(6.91)
x- and y-components of the resultant force (per

unit length) acting on the of the fluid within the curve, as
well as the resultant moment (per unit length) about the
z-axis, are given by
Thus, the
IC
Pdy,
IC
(6.92)
Pdx,
IC
(6.93)
P(xdx
+ ydy),
(6.94)
respectively,
where
the
integrals
are
taken
curve C. Finally, given

on the curve takes the form
(6.88), and that a constant pressure obviously yields
zero force and zero moment, we find that
(counter-clockwise)
that the
pressure
around the
distribution
IC ? ? ?
1
2
6.105) 116
FLUID MECHANICS
???
dF
dy,
dz
Y
M
z = x+iy, and z = xiy,
Now,
where
It follows that
However,
IC ? ? ?
???
dF
dx,
2
12
(6.96)
dz
IC ? ? ?
???
dF
(xdx
+ ydy).
(6.97)
dz
indicates
a complex
conjugate. Hence, d z
IC ? ? ?
X iY
12
???
dF
i
and id
z = dy+idx.
d z.
(6.98)
dz
where dF
= dxidy,
=d+id
???
???
dF
dF d F
d z
dz
and d F
dF d F
dz
= did.
d z
d z
dF
(6.99)
Suppose that
the curveC corresponds to a streamline of the flow, in
that the
curveC
corres
which
case
constant
dF
dF.Hence, on C,
on C.Thus, d
d F,
dz
=0 on C,and so
dF
dF
dz
dF
dF
dz
dF
!
2
dz,
dz
(6.100)
which implies that
IC
X iY
dF
!
2
dz.
dz
(6.101)
This result is known

Now, xdx
as the Blasius
+ ydy =Re(zd
IC ? ? ?
=Re
???
dF
z).Hence,
zd
theorem.
dz
(6.102)
or, making use of an analogous argument to that
employed
above,

=Re
IC
12
dF
!
2
zdz
dz
(6.103)
Note, finally, that Equations (6.101) and (6.103) hold
even
when is not constant
on the curve
C, as long
as
can
be continuously
deformed
without leaving the fluid
into
crossing
a constant- curve
over a singularity of
2.
(dF/dz)
an
As
or
example
of the
use
of the Blasius
theorem,
consider again the situation, discussed in Section 6.6, in

which
a line source
of strength Q is located at (0, a), and
there is a rigid boundary at

complex
y = 0. As we have seen,
y > 0 takes the form
dF
Q
z
the
velocity inthe region
z2 + a 2
dz
(6.104)
Suppose
that
we
evaluate
the Blasius integral, (6.103),
about the contour C shown in Figure 6.13. This contour
runs
along the boundary, and is completed by
in the
upper
demonstrated,
half
in
of
the
the limit
z-plane.
As
a semi-circle
is easily
in which the radius
of the
semi-circle tends to infinity, the contribution of the curved

section of the contour
negligible.
In this
contour contributes
rigid boundary).
corresponds
to
the overall integral
only the straight
to the integral.
section corresponds to
with
to
case,
becomes
section of the
Note that the straight
streamline (since it is coincident

In other words,
the contour C
streamline at all constituent
points that
a finite contribution to the Blasius integral,

ensures that C is a valid contour for the application
make
which
of the
Blasius theorem. In fact, the Blasius integral specifies
the
on the
net force (per unit length) exerted
whole fluid by
the boundary. Note, however, that the contour C can be

,
deformed into the contour C
which takes the form of a
small circle surrounding the
a singularity of (dF/dz) 2.
we can
evaluate
the
source,
without passing
over
See Figure 6.13. Hence,
Blasius
integral
around
without changing its value. Thus,
ez
IC
X iY
12
dF
!
2
dz
dz
(6.111)
2DPotential Flow
117
Figure 6.13: Source in the presence of a rigid boundary.
or
X iY
18
+ ia)(z
Writing
+
ia)
z =ia + e i
(z ia) 2
(z + ia)
dz
(z
"
?2 IC
=i e
dz.
i
(6.106)
d, and taking the limit
0, we find that
X iY
iQ
4a
(6.107)
In other words, the boundary

2/4a)ey
length) F
(Q
exerts
on
force (per unit
the fluid. Hence, the
ce
fluid(perexerts
unit le ngt
an h)F
equal= (Q
and opposite
force
2/4a) ey
( Q
on the boundary. Of course, this
result is
consistent with Equation (6.43). Incidentally, it is easily

demonstrated
from (6.103) that there is
(about the z-axis) exerted
vice
on the boundary
zero moment
by the fluid, and
versa.
Consider
line
source
of strength Q placed (at the
origin) in a uniformly flowing fluid whose velocity is V
= V0 (cos0
sin0 ).From Section 6.4, the complex velocity
potential of the net flow is

Q
F(z)
= 2
lnz V0
ze
i0
(6.108)
The net force (per unit length) acting
(which is calculated
around
by performing
the
large loop that follows streamlines,

the loop to
shrinking
on
source
the Blasius integral

and then
on
small circle centered
the
source) is (see Exercise 6.1)
Q V.
(6.109)
Note that the force acts in the opposite

flow. Thus,
direction
an
external
as the flow, must
force
F,
direction
acting
to the
in the
same
be applied to the source in order
for iton
to remain stationary. In fact, the above result is valid
irecti
even
in
non-uniformly
flowing fluid,
as
long
as
V is
interpreted
source
as
the fluid velocity
(excluding
Finally,
is V
= V0
a vortex filament of
a uniformly flowing
consider
placed at the origin in

velocity
at the location of the
the velocity field of the source itself).
(cos 0
intensity
fluid whose
sin 0 ).From Section 6.4, the
complex velocity potential of the net flow is
F(z)
=i
lnz V0
(6.110)
ze i0
The net force (per unit length) acting
on the filament
(which is calculated by performing the Blasius integral

around
small circle centered
Exercise 6.2)
= V
ez
118
on the filament)
is (see
FLUID MECHANICS
Note that
the force
is directed
direction of the flow (in the

through
90
sense
at right-angles
to the
obtained by rotating V
in the opposite direction to the filaments
direction of rotation). Again, the above result is valid
even
in
a non-uniformly
as the fluid
interpreted
flowing fluid,
velocity
as
long
as
V is
at the location of the
filament (excluding the velocity field of the filament itself).
6.10
Exercises
6.1. Demonstrate that
a line source
of strength Q (running
along the z-axis) situated ina uniform flow of (unperturbed)

velocity
V (lying inthe x-y plane) and density experiences
per unit length

F
a force
= QV.
a vortex
(running along the z-axis) situated ina uniform flow of

(unperturbed)
velocity V (lying inthe x-y plane) and density

experiences
a force per unit length

F
= V
ez
sources of strengths Q and Q ,

located a perpendicular distance r apart, exert a radial force per
/(2r)
unit length Q Q
onone another, the force being
6.3. Show that two parallel line
attractive if QQ
> 0,and repulsive
if Q Q
< 0.
6.4. Show that two parallel vortex filaments of intensities and
a perpendicular
per unit length
located
repulsive if
r apart, exert a radial force

on one another, the force being
distance
/(2r)
> 0,and attractive
if
< 0.
6.5. A vortex filament of intensity runs parallel to, and lies a

perpendicular distance
a from, a rigid planar boundary. Demona net force per unit
strate that the boundary experiences

length
2/(4a)
directed toward the filament.
6.6. Two rigid planar boundaries meet at right-angles. A line
source
of strength Q runs parallel to the line of intersection of
the
planes, and is situated
a perpendicular
distance
a from
per unit
each. Demonstrate that the source is subject to a force
length
2Q
2
3
8a
directed towards the line of intersection of the planes.
6.7. A line source of strength Q is located a distance b from

an impenetrable cylinder of radius a < b (the axis of the
cylinder
being parallel to the source). Demonstrate that the cylinder
experiences
a net force per unit length

Q
2b
(b
a2
2
a2)
directed toward the source.

6.8. A dipole line
source
consists of
a line source
of strength
Q, running parallel to the z-axis, and intersecting the x-y plane at
(d/2)(cos,
that intersects the
sin), and
x-y
a parallel source
plane at (d/2)(cos,
of strength Q
sin). Show that,
in
the limit d 0, and Qd D, the complex velocity
potential of the
source
is
F(z)
Here, De
De
2z
istermed the complex dipole strength.
6.9. A dipole line
a uniformly
source
of complex strength De
is placed in
flowing fluid of speed V0 whose direction of motion
subtends
a (counter-clockwise) angle 0 with the x-axis.

no net force acts on the source, it is subject to
Show that, while
a
moment (per unit length) M
= DV0
sin( 0 )about the
z-axis.
source of complex strength D1 e i1

running along the z-axis, and a second parallel source of complex
i2
strength D2 e
that intersects the x-y plane at (x,0).
Demonstrate that the first source is subject to a moment (per
6.10. Consider
a dipole
line
unit
length) about the z-axis of
M
as well as a force
y-components
D1 D2
2
x2
sin(1 +2 ),
x- and
(per unit length) whose
are
X
D1 D2
cos(1
x3
+2 ),
D1 D2
x3
sin(1 +2 ),
respectively. Show that the second
1
cos(n/2)].
source
issubject to the
2DPotential
Flow
119
6.11. A dipole line source of complex strength De
to, and is located
a perpendicular
distance
runs parallel
a from, a rigid planar
boundary. Show that the boundary experiences
a force per
unit length
2
D
8a
acting toward the source.
a conformal map converts a line source

a line source of the same strength, and a vortex filament into
a vortex filament of the same intensity.
6.13. Consider the conformal map
into
where
z = ic cot(/2),
z = x + iy, = + i, and c isreal and positive.
Show that
c sinh
cosh cos
Demonstrate that
circular
arc of center
c sin
cosh cos
= 0
,
.
where 0 0 ,
maps to a
(0, c cot0 ), and radius c|cosec0 |,that
connects
y > 0.
maps to the continuation of this arc
the points (c, 0), and lies in the region

Demonstrate
that
= 0
inthe region
y < 0. In particular, show that = 0 maps to the region

on the x-axis, whereas = maps to the region |x| < c.
Finally, show that =0 maps to a circle of center (c coth
|x| > c
0), and radius c|cosech|0
6.14. Consider the complex velocity potential

F(z)
where
Here, V0
2ciV0
cot(/n),
z = ic cot(/2).
n,and c are real and positive.
dF
4V0
=
dz
n2
sin (/2)
sin (/n)
Show that
Hence, deduce that the flow at |z| is uniform, parallel
to the x-axis, and of speed V0
(,)
2V0
Show that
sin(2/n)
cosh(2/n) cos(2/n)
Hence, deduce that the streamline

x-axis for |x| > c, but along
a circular arc
= 0 runs along
the
connecting the points
(c,
0) for |x| < c. Furthermore, show that if 1
< n < 2 then
this
arc lies
above the x-axis, and is of maximum height

#
"
=c
cos(n/2)
sin(n/2)
+1
but if 2 < n< 3 then the arc lies below the x-axis, and is
of maximum depth
"
+1
cos(n/2)
=c
|sin(n/2)|
Hence, deduce that if 1

<n<2 then the complex
velocity potential under investigation corresponds to uniform

flow of
speed V0
parallel to a planar boundary that
possesses a
cylindrical bump (whose axis is normal to the flow) of height h
and
width 2c, but if 2 < n<3 then the potential corresponds
to flow parallel to a planar boundary that
possesses a cylindrical
depression of depth d and width 2c. Show, inparticular,
=1then the bump is a half-cylinder,
that if n
and if n
=3 then
the
depression is a half-cylinder. Finally, demonstrate that the
flow speed at the top of the bump (in the case 1
< n< 2), or the
bottom of the depression (in the
v=
case 2 < n< 3) is
2V0
n2
[1 cos(n/2)].
120
FLUID MECHANICS
= cosh(/a)
maps the semi-infinite strip

upper half (y 0) of the
z-plane. Hence, show that the stream function due to
6.15. Show that
a, 0 in the -plane
line
source
rectangular
of strength
onto the
Q placed
at
(0, a/2), in the
region 0
a,
0 bounded by the rigid planes
= 0, and = a, is
(,)
Q sinh(/a) sin(/a)
2[sinh
(/a)
+ cos 2 (/a)]
= 0,
6.16. Show that the complex velocity potential

aV0
F(z)
tanh(a/z)
can be interpreted as that due to uniform flow of speed V0 over a

cylindrical log of radius a lying on the flat bed of a deep stream
(the axis of the log being normal to the flow). Demonstrate

2
that the flow speed at the top of the log is( /4)V0
Finally,
show that the pressure difference between the top and bottom of
mparativelythin laye Incompressible Boundary Layers
121
7 Incompressible Boundary Layers

7.1
Introduction
Previously,
5.13,
in Section
flowing incompressible
inca-pable of exerting
we saw
that
fluid that is modeled
drag force
on a
a uniformly
as inviscid is
rigid stationary
obstacle placed inits path. This result is surprising since, in

practice,
stationary
obstacle
drag when situated in such

the Reynolds
experiences
a fluid, even
significant
in the limit that
number tends to infinity (which corresponds
to the inviscid limit). In this chapter,
we
shall attempt to
reconcile these two results by introducing the concept of

boundary
covers the surface

number
have
layer. This is
an obstacle
of
incompressible
a significant
effect
on
simplicity,
we
shall
Reynolds
stationary
on the flow
number
our
restrict
flows
of
infinite
For the sake of
discussion
layers that form when
fluid
obstacle
to
is assumed
inside the layer, but
the flow outside.
boundary
thin layer that
placed in a high Reynolds
fluidviscosity
effect
negligible
two-dimensional
comparatively
transversely
length
to
the
high
around
and
uniform
cross-section.
7.2
We
No Slip Condition
saw,
in Section 5.13, that when
flows around
a rigid
an
inviscid
fluid
stationary obstacle then the normal
fluid velocity at the surface of the obstacle is required to be
zero. However, in general, the

non-zero. In fact, if the fluid
incompressible
tangential
velocity
velocity
field
and irrotational then it is derivable from
stream function
that
satisfies
Laplaces
is
is both
equation.
(See
a well-known property of Laplaces

we can either specify the solution itself, or its
derivative, on a bounding surface, but we cannot
Section 5.8.) It is
equation that
normal
specify
both these
quantities
simultaneously.
Now, the
constraint
value
zero
zero
of
normal velocity
(say)
on the
to the
is equivalent
that the stream function
requirement
take the constant
surface of the obstacle. Hence, the
normal derivative of the stream function, which determines

velocity, cannot
the tangential
surface, and is,ingeneral,
In reality, all physical fluids
possess
flows
Moreover,
stationary
when
obstacle
viscous
fluid
finite viscosity.
around
rigid
both the normal and the tangential
components of the fluid velocity

the obstacles
at this
also be specified
non-zero.
are
found to be
surface. The additional constraint
tangential
fluid velocity
boundary
is known
as
be
the
zero at
that the
zero at a rigid stationary

no slip condition, and is
ultimately justified via experimental observations.

The concept of
a boundary
layer
was first
introduced
into fluid mechanics by Ludwig Prandtl (18751953)
in
order to account for the modification to the flow pattern of
a high
Reynolds number irrotational fluid necessitated
the imposition of the

impenetrable
no slip condition on the surface
covers
layer
relatively
thin in the direction
an
an
the surface of the obstacle, but is

normal to this surface.
the layer, the flow pattern is the
idealized
by
stationary obstacle. According to Prandtl, the
boundary
Outside
of
inviscid
fluid,
and
is
same as that
thus
of
generally
irrotational. This implies that the normal fluid velocity is
zero on the outer
edge of the layer, where it interfaces
with the irrotational
flow, but, in general, the tangential
velocity
is
non-zero.
However,
requires the tangential velocity

edge of the layer, where
surface. It follows
it interfaces
presence
condition
the inner
with the rigid
a very large normal

across the layer, which
that there is
gradient of the tangential velocity

implies the
no slip
zero on
the
to be
of intense internal vortex
trapped within the layer. Consequently,
filaments
the flow within
we
shall
layer
more
the layer is not irrotational.
In the following,
attempt to make the concept
of
boundary
precise.
7.3
Boundary Layer Equations
Consider
rigid stationary
obstacle
(locally) flat, and corresponds

surface
whose
x-z
the typical
normal thickness
layer thus extends

that occupies
layer,
over
the region
is
assumed
<
to
y < ,
be
<
is
number fluid
7.1. Let be
of the boundary
the region 0 < y
surface
plane. Let this
a high Reynolds
y > 0. See Figure
be in contact with
that occupies the region
the
to the
layer. The
. Now, the fluid
and thus lies outside
both
irrotational
and
(effectively) inviscid. On the other hand, viscosity must be
included in the equation of motion of the fluid within the

layer. The fluid both inside and outside the layer is assumed
to be incompressible.
122
FLUID MECHANICS
irrotational
y
fluid
U(x)
boundary
layer
solid surface
Figure 7.1: A boundary layer.
that the equations
Suppose
of irrotational
flow have
already been solved to determine the fluid velocity outside

the
boundary
layer. This velocity
normal component
(i.e.,
y ).
is
zero at
must be such that its
the outer edge of the layer
On the other hand, the tangential component of
the fluid velocity at the outer edge of the layer, U(x) (say),
is generally
non-zero.
of simplicity,
z-direction,
so
Here,
we are assuming, for the sake

no spatial variation in the
that there is
that
both the irrotational
flow
and the
boundary
we are
layer
are
effectively two-dimensional.
also assuming that
time variation
can
all flows
are
Likewise,
so
steady,
any
that
be neglected. Now, the motion of the
fluid within the boundary layer is governed by the equations

of steady-state,
incompressible,
two-dimensional,
viscous
flow, which take the form (see Section 2.14)

vy
vx
+
x
vx
(7.1)
vx
vy
vx
+
x
+ vy
vx
1 p
vy
(7.2)
+vy
vx
1 p
,
,
vx
+
y
vy
0,
vy
+
y
(7.3)
where is the (constant)
density, and the kinematic
viscosity.
Here,
continuity,
whereas Equations (7.2) and (7.3)
Equation
and y-components
of
the
(7.1)
fluid
is
the
equation
equation
of
are
x-
of
the
motion,
respectively. The boundary conditions at the outer edge of

the layer, where it interfaces with the irrotational fluid, are
vx (x,y)
U(x),
P(x)
pressure at
the outer
(7.4) p(x,y)
(7.5)
as
y/
. Here, P(x) is the fluid
edge of the layer, and
dU
U
(since
vy
= 0, and viscosity
dP
dx
dx
(7.6)
is negligible, just outside the
layer). The boundary conditions at the inner edge of the

layer, where it interfaces with the impenetrable surface,
=
=
vx (x,0)
(7.7) vy (x,0)
are
0,
0.
7.21) Incompressible Boundary Layers
123
course, the first of these constraints

no slip condition.
Let U0 be a typical value of the
Of
velocity,
U(x),
and
let
be
the
corresponds to the
external tangential
typical
variation
length-scale of this quantity. It is reasonable to suppose that

U0 and L are also the characteristic
and variation
of
length-scale
the boundary
variation
Moreover,
layer. Of
length-scale
of
tangential flow velocity
in the x-direction, respectively,
the
course,
layer
is the
in the
typical
y-direction.
/L 1, since the layer is assumed
to be
thin.
is helpful
define the Moreover, /L 1, sinc
of
theItlayer
in theto
y-direction.
normalized variables
X
Y
Vx (X,Y)
Vy (X,Y)
bP(X,Y)
x
L
=
=
,
,
vx
U0
vy
(7.9)
(7.10)
,
,
,
(7.11)
(7.12)
U1
p
p0
where U1 and p0
(7.13)
are constants.
All of these variables
are
designed to be O(1) inside the layer. Equation (7.1) yields

U0
Vx
U1
Vy
+
L
0.
(7.14)
In order for the terms in this equation
another,
we need
to balance
one
U1
U0
L
,
.
(7.15)
In other words, within the layer, continuity
requires the
typical flow velocity in the y-direction, U1
to be much
smaller
than that inthe x-direction, U0
Equation (7.2) gives

U
2
0
Vx
Vx
Vx
2Vx
In order for the
2Vx
+
Y
? "?
U0
bP
p0
Vy
?2
(7.16)
pressure term on the right-hand
side of the
above equation to be of similar magnitude to the advective
terms
on the left-hand
side, we require that
p0
U0
(7.17)
Furthermore,
in order
for
the
viscous
right-hand side to balance the other terms,
term
on
the
we need
U1
U0
1
Re
1/2
(7.18)
where
Re
U0 L
(7.19)
is the Reynolds number of the flow external to the layer.
is
The
the
assumption
R eyn olds
thatnumber
/L of
1
can
thebe
flow
seen
external
to imply
to that
the
Re 1
In other words, the normal
boundary layer separating
an
thickness
of the
irrotational flow pattern from
rigid surface is only much less than the typical variation
length-scale
a rigid
of the pattern when the Reynolds number of
the flow is much greater than unity.

Equation (7.3) yields
!
Vy
1
Vx
Re
+ Vy
Vy
Y
bP
"
1
Re
1
Re
2Vy
2Vy #
+
Y
(7.20)
In the limit Re 1, this reduces to
bP
Y
(7.38)
0.
124
FLUID MECHANICS
Hence, bP
bP(X), where
d bP
dX
b
U
b
dU
dX
b
U(X)
In
other
b
the
pressure
U/U0
words,
is uniform
(7.22)
and use has been made of (7.6).
across
the layer, in the direction
b
normal
to the surface of the obstacle, and is thus the
as that on the outer
same
edge of the layer.
Retaining only O(1) terms,
our
final set of normalized
layer equations becomes
Vx
Vy
Vx
(7.23)
Vx
X
+ Vy
Vy
0,
Y
b
b
U
dU
2Vy
(7.24)
subject
to
the
boundary
conditions
Vx (X, )
b
U(X),
(7.25)
and
Vx (X,0)
(7.26) Vy (X,0)
=
=
0,
0.
(7.27)
In unnormalized form, the above set of layer equations
are written
vy
vx
+
x
vx
vx
+vy
vx
0,
dU
dx
subject to the boundary conditions
vx (x,)
(note that
y =
really
means
= U(x)
y/ ), and
vx (x,0)
vy (x,0)
=
=
0,
0.
(7.28)
2vx
+
y
(7.29)
(7.30)
(7.31)
(7.32)
Now,
Equation
automatically
(7.28)
satisfied
can
be
by
expressing
velocity in terms of a stream function: i.e.,
the
flow
vx
vy
(7.33)
,
.
(7.34)
Inthis case, Equation (7.29) reduces to
+
y
xy
=U
dU
dx
(7.35)
(x,)
U(x),
(7.36)
and
(x,0)
(7.37)
(x,0)
0,
0.
(7.50) Incompressible Boundary Layers
125
To lowest order, the vorticity
ez
is given by
internal to the layer,
(7.39)
whereas the x-component
area
acting
on the
of the viscous force
per
unit
surface of the obstacle is written (see
Sec-tion 2.18)
xy
??????y=0
?????y=0
???y=0
vx
(7.40)
7.4
Self-Similar Boundary Layers
The boundary layer equation, (7.35), takes the form of
a nonlinear
difficult
partial differential equation that is extremely
to
solve
exactly.
However,
considerable
progress can be made if this equation is converted into

an ordi-nary differential equation by demanding that its
solutions be self-similar. Self-similar solutions are such
that, at a given distance, x, along the layer, the tangential
flow profile, vx (x, y), is a scaled version of some
common pro-file: i.e., vx (x, y) = U(x) F[y/(x)], where
(x) is a scale-factor, and F(z) a dimensionless function.
It follows that
=F(z).
(x,y)
= U(x)(x)
f[y/(x)],
where f (z)
Let
us
search for
solution to Equation
self-similar
(7.35) of the general form

"
(x,y)
= U0
xm
f()
m+1
"
#1/2
x m+1
2U0
#1/2
2
(m + 1)U0
f(),
x m1
(7.41)
where
"
(m + 1)U0
x m1
#1/2
y.
2
(7.42)
m1] 1/2,
This implies that (x)
[2/(m + 1)U0 x
and U(x)
m
m
U0 x
Here, U0 and m are constants. Moreover, U0 x
has dimensions
dimensionless.
we find
of velocity,
Transforming
whereas
m,
, and f, are
variables from
x, y
to
x, ,
that
?????y
?????
+
x
"
?????x
(7.43)
=
=
?????x
1 x
(m + 1)U0
x m1
#1/2
?????x
,
.
(7.44)
Hence,
"
xm
U0
(m + 1)U
3
0
x m1
[2m f
x 3m1
+ 1) f + (m
,
1/2
2
U0
[(m
2(m + 1)
#1/2
x m1
U0
+ (m 1) f ],
(m + 1)U
xy
2
0
x 2m1
f
2
where
1) f
=d/d. Thus, Equation (7.35) becomes

],
(7.45)
(7.46)
(7.47)
(7.48)
(7.49)
(m + 1) f
+ (m + 1) f f
2m f
=
U
dU
x 2m1
dx
(7.50)
126
FLUID MECHANICS
Since
the
left-hand
side
of
the
above
equation
is
(non-constant) function of , whilst the right-hand side is a

function
of
variables),
(and
since
the equation
right-hand side takes
and
can
a constant
x are
only
independent
be satisfied
if its
value. In fact, if
1
U
2
0
dU
x 2m1
dx
2m
(7.51)
then
U(x)
U0
xm
(7.52)
(which is consistent with our initial guess), and
+f f
+ (1
2)
0,
(7.53)
where
2m
m+ 1
(7.54)
Expression
(7.53)
is
known
as
the
Falkner-Skan
equation. The solutions to this equation that satisfy the

physical
boundary conditions (7.36)(7.38)are
f(0)
such that
f (0)
0,
(7.55)
and
f
(7.56)
()
=
=
()
1,
0.
(7.57)
(The final condition
corresponds
the vorticity tend to
zero at
to the requirement
that
the edge of the layer.) Note,
from (7.39), (7.42), (7.47), (7.52), (7.55), and (7.56), that
the normally integrated vorticity within the boundary layer
is
Z
dy
U(x).
(7.58)
Furthermore,
x-component
from
(7.40),
(7.47),
of the viscous force
per
and
unit
(7.52),
area
the
acting
on
the surface of the obstacle is

???y=0
xy
12
U
?1/2
Ux
2
(m +
1)
1/2
(0).
(7.59)
It is convenient
to parameterize
this quantity in terms of
skin friction coefficient,

???y=0
xy
cf
(1/2)U
(7.60)
It follows that
cf
(m + 1)
(x)
1/2
[Re(x)]
(0)
1/2
(7.61)
where
Re(x)
U(x) x
(7.62)
is the effective Reynolds number of the flow

edge of the layer at position
Finally, according
x. Hence, c f
to Equation
on the outer
x (m+1)/2.
(x)
(7.41), the width of the
boundary layer is approximately
(x)
[Re(x)]
1/2
(7.63)
which implies that (x)

Incompressible Boundary Layers
x (m1)/2.
127
1.5
1
)
0
(
0.5
Figure 7.2: f
(0) calculated
as a function
of for
solutions of the Falkner-Skan equation.
Note
velocity
that
irrotational
on a
if
0 then the external
= U0
xm
corresponds
tangential
to that of
inviscid flow incident, in a symmetric fashion,
semi-infinite
, where
5.12.)
m >
profile, U(x)
wedge whose
apex
subtends
an angle
= 2m/(m + 1). (See Section 5.14, and Figure
In this
case, U(x) can be interpreted as

a distance x along the surface of
apex (in the direction of the flow).
the
tangential velocity
the
wedge from its
By
analogy,
if
0 then
the external
velocity
profile
corresponds to that of irrotational inviscid flow parallel to
can
can
flat plate (which
semi-infinite
apex
wedge whose
be interpreted
subtends
as
zero
of as a
case, U(x)
a distance x
be thought
angle). In this
the tangential
velocity
along the surface of the plate from its leading edge (in the
direction of the flow). (See Section 7.5.) Finally, if
then the external
irrotational
velocity
inviscid
semi-infinite
flow
wedge whose
), where
= m/(1
profile
over
apex
Figure 5.13.) Inthis case,
m<0
is that of symmetric
the back
surface
an
subtends
of
angle (1
m). (See Section 5.15, and

U(x) can be interpretedas the
tangential
velocity
apex
a distance x along
the surface of the wedge from its
(inthe direction of the flow).

the
Unfortunately,
possesses no
general
equation
relatively
is
Falkner-Skan
analytic
equation,
solutions.
numerically
Falkner-Skan
2m/(m+
equation
7.3, that
for various different
significance. For 0 <
and
via
(0), calculated
1), for the
the boundary conditions

()
In addition, Figure 7.3 shows f
versus ,
Note that, since
7.2
this
solve
of (7.53) that satisfy
(7.55)(7.57).
calculated
function of
as a
numerically
solutions
to
straightforward
numerical methods. Figure 7.2 shows f
(7.53),
However,
as m
values of
, solutions
m.
v
of the
with > 2 have no physical

< 2 it can be seen, from Figures
there
is
single
solution
branch
characterized
by f ()
> 0 and
(0)
> 0. This
branch is
termed the forward flow branch, since it is such that the

tangential velocity,
as
vx ()
f (), is in the
the external tangential velocity [i.e.,
whole layer (i.e., 0 <
< ).
vx
same
direction
across
()]
the
The forward flow branch is
by a positive skin friction coefficient, c f

can also be seen that for < 0 there exists a
characterized
f
(0). It
second solution branch, which is termed the reversed flow

branch, since it is such that the tangential velocity
opposite
direction to the external tangential
(which corresponds
reversed flow branch is characterized
friction coefficient.
are
in
adjacent to the surface
the region of the layer immediately

of the obstacle
is in the
velocity
to
by
0). The
a negative
skin
Note that the reversed flow solutions
probably unphysical, since reversed flow close to the
wall is generally associated with
separation
a phenomenon
known
as
layer
invalidates
the boundary layer orderings. It can be seen that
the two solution branches
0.1989,
Moreover,
which
there
(see
7. 10) which
boundary
merge
corresponds
are no
Section
to
together
solutions
at
= =
0.0905.
to the Falkner-Skan
< or m < m The disappearance

solutions
r-Skan
equat
when
io nm
w ith
becomes
< too negative (i.e., when
equation with
of
the
deceleration of the external flow becomes too large) is also

related to boundary layer separation.
128
FLUID MECHANICS
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
10
Figure 7.3: Solutions of the Falkner-Skan
equation
correspond to forward flow solutions calculated
In
curves
with m =
order from the left to the right, the various solid
4, 1, 1/3, 1/9, 0, 0.05, and 0.0904, respectively.
dashed
4, 1,curve
1/3, 1/
shows
with
7.5
reversed flow solution
The
calculated
m = 0.05.
Boundary Layer
Consider
on a Flat
Plate
flat plate of length L, infinite
width, and
x-z plane, and

whose two edges correspond to x = 0 and x = L.Suppose
that the plate is immersed in a low viscosity fluid whose
unperturbed velocity field is v = U0 ex See Figure 7.4. In
negligible
thickness,
which lies in the
the inviscid limit, the appropriate
the surface of
the plate,
requirement
zero
of
by the unperturbed
vy
boundary
condition at
= 0corresponding
normal velocityis
to the
already satisfied
flow. Hence, the original flow is not
modified by the
presence
of the plate. However, when
take the finite
viscosity
of the fluid into account,
we
an
= 0corresponding
to
additional boundary
the
no
condition,
slip conditionmust
imposition
of
this
vx
be satisfied at the plate. The
additional
constraintsatisf
causes
thin
ied at t he
p
boundary layers, of thickness (x) L, to form above and

below the plate. The fluid flow outside the boundary layers
remains effectively inviscid, whereas that inside the layers
is modified by viscosity. It follows that the flow external
to the layers is unaffected by the

Hence, the tangential
velocity
presence
of the plate.
at the outer edge of the
boundary layers is U(x)
= 0 discussed
(7.52). (Here,
= U0
This corresponds to the
in the previous
we are
sectionsee
case
Equation
assuming that the flow upstream
of
x = L, is unaffected by the
edges presence, and, is, therefore, the same as if the plate
were of infinite length. Of course, the flow downstream of
the edge is modified as a consequence of the finite length
the trailing edge of the plate,
of the plate.)
Making
section
use of the analysis contained in the

m = 0), as well as the fact
previous
(with
that, by
symmetry, the lower boundary layer is the mirror image of

the
upper one,
the tangential velocity
across
profile
the
both layers is written
vx (x,y)
U0 f (),
(7.64)
where
?
?1/2
U0
|y|.
(7.65)
Here, f() is the solution of
f
ckness
129
+f f
=0
boundary layer
plate
U0
x
L
Figure 7.4: Flow
over a flat plate.
that satisfies the boundary conditions

f(0)
=f
(0)
=0,
0.
and
f
Equation (7.66) is known
()
as the Blasius
1, f
()
equation.
wake
(7.67)
(7.68)
(7.69)
It is convenient
to define the so-called displacement
thickness of the upper boundary layer,

Z
(x)
"1
vx (x,y)
dy,
U0
(7.70)
which
can
streamlines
be interpreted
as
the distance through which
just outside the layer
due to the retardation
are
displaced
laterally
of the flow within the layer. (Of
course, the thickness of the lower boundary

same as that of the upper layer.) It follows that
(x)
!1/2
layer is the
[1 f ()]d.
U0
(7.71)
In fact, the numerical
solution of (7.66), subject
to the
boundary conditions (7.67)(7.69),yields
(x)
=1.72
x
U0
!1/2
(7.72)
Hence, the thickness of the boundary
the
square root
as
layer increases
of the distance from the leading edge of
the plate. In particular, the thickness at the trailing edge of

the plate is
(L)
1.72
Re
1/2
(7.73)
where
Re
U0 L
(7.74)
is the appropriate Reynolds number for the interaction of
theflow with the plate. Note that if Re
thickness
as was previously
130
1
thenthe
of the boundary layer is much less than its length,
assumed.
FLUIDMECHANICS
1
0.9
0.8
0.7
0
U
/
x
0.6
0.5
0.4
0.3
0.2
0.1
0
5 4 3 2 1
y /
Figure 7.5: Tangential velocity profile
across
the
boundary
layers located above and below
thickness located at y
=0.
a flat plate
of negligible
The
tangential
velocity
across
profile
the
both
boundary layers, which takes the form
"1.22
vx (x,y)
=U0
#
|y|
(x)
(7.75)
is plotted in Figure 7.5. In addition, the vorticity
across
profile
the layers, which is written
"1.22
(x,y)
= sgn(y)1.22
U0
f
#
|y|
(x)
(x)
(7.76)
is shown in Figure 7.6. Note that the vorticity is negative in
y > 0), positive in the lower

y < 0), and discontinuous across the
plate (which is located at y = 0). Finally, the net viscous
drag force per unit width (along the z-axis) acting on the
the
upper
boundary layer (i.e.,
boundary layer (i.e.,
plate inthe x-direction is

Z
=2
???y=0
xy
dx,
(7.77)
where the factor of 2 is needed to take into account the
presence
of boundary
layers both above and below the
plate. It follows from Equation (7.59) (with
m =0) that
!1/2
D
Z
2
f
!1/2
x 1/2
(0)
=U
dx
=U
2
0
U0
(0).
U0
(7.78)
In fact, the numerical solution of (7.66) yields
=1.33
U
Re
2
0
1/2
=1.33U0
(U0 L)
1/2
(7.79)
The above discussion is premised
that the flow in the
both
steady
assumption
upper
the assumption
(or lower) boundary
and z-independent.
becomes
on
It turns
invalid when the Reynolds
analogy, if Incompressible Boundary Layers
131
layer is
out that this

number
0.6
0.5
0.4
0.3
0.2
)
0.1
U
(
0
0.1
0.2
0.3
0.4
0.5
0.6
5 4 3 2 1 0
y /
Figure 7.6: Vorticity profile
across
the
boundary layers located above and below

of negligible thickness
U0 /, exceeds
located at y
a critical value
a flat plate
=0.
which is about 600. In this
case, small-scale z-dependent disturbances spontaneously

grow to large amplitude, and the layer becomes turbulent.
1/2,
Since x
if the criterion for boundary layer
is not satisfied at the trailing edge of the plate,
turbulence
x = L, then it is not
satisfied anywhere else in the layer.
previous analysis, which neglects turbulence,
Thus, the
remains valid provided U0 (L)/
< 600. According
to
(7.73), this implies that the analysis is valid when 1

Re <
5,
1.210 where
implRe
U0 L/ is the Reynolds numberof
the external flow.

Consider, finally, the situation illustrated in Figure 7.7
in which
an initially
irrotational fluid
passes
between two
flat parallel plates. Let d be the perpendicular
distance
we have seen, the finite viscosity of

the fluid causes boundary layers to form on the inner
surfaces of the upper and lower plates. The flow within
these
layers
possesses non-zero vorticity, and is
between the plates. As
significantly
flow
outside
inviscidthis
affected by viscosity. On the other hand, the
the
layers
is irrotational
(since it can be derived from

Poissons
boundary
and
essentially
type of flow is usually termed potential flow
equation).
a velocity
layers increases
potential satisfying
thickness of the two

1/2,
where x represents
like x
Now,
the
distance, parallel to the flow, measured from the leading

edges of the plates. It follows that,
region of potential
disappears.
See
as x
increases, the
flow shrinks in size, and eventually
Figure
7.7. Assuming
that, prior
to
merging, the two boundary
one
affect
formula
another,
their
(7.72), where U0
layers do not significantly
thickness,
(x),
is given
by
is the speed of the incident
fluid. The region of potential flow thus extends from
= 0 (which corresponds
to x = l,where
(l)
to the leading edge of the plates)
d2
(7.80)
It follows that
l
d
11.8Re,
(7.81)
where
Re
U0 d
(7.82)
Thus, when
passes
an
between
potential
flow
irrotational
high Reynolds
two parallel
extends
number
fluid
plates then the region of
comparatively
long
distance
between the plates, relative to their spacing (i.e., l/d 1).
as we saw
in 132
FLUID MECHANICS
boundary layer
plate
U0
potential flow
Figure 7.7: Flow between two flat parallel plates.
an irrotational
high Reynolds
number fluid
pipe then the fluid remains essentially

has
travelled
considerable
passes
into
distance
along
the
pipe,
compared to its diameter. Obviously, these conclusions
As
are
if the flow becomes turbulent.
modified
7.6
irrotational until it
Wake Downstream of a Flat Plate
we saw
negligible
in the previous
thickness,
section, if
flat plate of
and finite length, is placed in the
path of
parallel
to the plate, then thin boundary
uniform high Reynolds number flow, directed
layers
form
above and below
the plate. Outside the layers, the flow is
irrotational, and essentially inviscid. Inside the layers, the

flow is modified
Downstream
by viscosity, and has

the
of
convected by the flow, and
non-zero
vorticity.
are
merge to form a thin wake. See
plate,
the
boundary
layers
Figure 7.4. Within the wake, the flow is modified

viscosity, and
possesses
the downstream
finite
by
vorticity. Outside the wake,
flow remains irrotational, and effectively
inviscid.
Since there
is
no
solid surface
wake, acting to retard the flow,
we
embedded
in the
would expect the
action of viscosity to cause the velocity within the wake,

long distance downstream
of the plate, to closely match
that of the unperturbed flow. In other words,
we expect
the
fluid velocity within the wake to take the form
vx (x,y)
(7.83) vy (x,y)
U0 u(x,y),
v(x,y),
(7.84)
where
|u|
U0
(7.85)
Assuming that, within the wake,
1x
(7.86)
(7.87)
where
is the wake
thickness,
fluid
continuity
requires that
u.
(7.88)
Now, the flow external to the boundary
wake, is both uniform and essentially
according
to Bernoullis
theorem, the
layers, and the

inviscid.
Hence,
pressure
in this
region is also uniformsee Equation (7.22). However,

(7.98)
133
Section 7.3, there is
the
boundary
uniform
that the
no y-variation
layers.
within the layers.
pressure
pressure across
pressure is
Thus, it is reasonable to assume
It follows
of the
that
the
is also uniform within the wake, since the
wake is formed via the convection of the boundary layers

downstream of the plate. We conclude that
p(x,y) p0
(7.89)
everywhere inthe fluid, where p0 is a constant.

The x-component
of the fluid equation of motion is
written
vx
vx
+ vy
vx
1 p
vx
vx
+
y
(7.90)
Making
use of (7.83)(7.89),the
above expression reduces
to
u
U0
2u
(7.91)
The boundary condition
u(x,)
=0
(7.92)
ensures
that the flow outside the wake remains
unperturbed. Note that Equation (7.91) has the same

mathematical
with
form as a conventional diffusion equation,
x playing
of the diffusion
the role of time, and /U0 playing the role

coefficient. Hence, by analogy with the
standard solution of the diffusion equation,

expect ( x/U0 )1/2.
we would
As can easily be demonstrated, the self-similar solution
to (7.91), subject to the boundary condition (7.92), is
u(x,y)
exp
2
2
(7.93)
where
(x)
=2
!1/2
U0
(7.94)
and Q is a constant. It follows that
udy
= Q,
(7.95)
R
since,
as is well-known,
2)dt
exp(t
takes the form
vx (x,y)
U0
(7.96)
= 1
U0
exp(y
As
as x 1/2.
across the wake, which
expected, the width of the wake scales

The tangential velocity profile
2/
2),
is plotted in Figure 7.8. Inaddition, the vorticity profile
across
the wake, which is written
(x,y)
U0
exp(y
U0
2/
2)
(7.97)
is shown in Figure 7.9. It
can
seen
be
that the profiles
pictured in Figures 7.8 and 7.9
are
out
layer profiles
versions
of the boundary
essentially
smoothed
shown
in
Figures 7.5 and 7.6, respectively.

Suppose
wake
are
that the plate and
enclosed
by
portion of its trailing
cuboid control volume
of unit
to x =
= h to y = h. See Figure 7.10. Here, l
and from
+lLand
h (l),
y = where
hto L
y is=the
h.length
Se e F igure
of the7.
plate,
10. Her
and
depth (in the z-direction) that extends from
+l and from
x = l
(x)
l
the
L width
and
of the wake. Hence, the control volume
plate,
extends
and well
(x) the
upstream
width ofand
the wake.
downstream
Hence, the
of co
the
plate.
Moreover, the volume is much wider than the wake.

Let
us apply
the integral form of the fluid equation of
continuity to the control volume. For
a steady-state,
reduces to (see Section 2.9)
IS
134
dS
this
0,
FLUID MECHANICS
1
0.9
0.8
0.7
U
/
x
0.6
0.5
0.4
0.3
0.2
0.1
0
y /
Figure 7.8: Tangential velocity profile
across
the wake
of a flat plate of negligible thickness located at y
The profile is calculated for Q/(U0 )
=0.5.
=0
0.5
0.4
0.3
0.2
)
0.1
U
(
0.1
0.2
0.3
0.4
0.5
y /
Figure 7.9: Vorticity profile
across
the boundary layers
a flat plate of negligible

y =0.The profile is calculated
above and below
thickness
located at
for Q/(U0 )
0.5.
135
v(x)
y =h
plate
wake
U0 u(y)
U0
y =h
x = l
x =l
v(x)
Figure 7.10: Control volume
surrounding
a flat plate and its trailing
where S is the bounding
surface of the control volume.
The normal fluid velocity is U0
= l, and v(x) at y = h,
(7.98) yields
as
wake.
at
x = l,
U0 u(y) at
indicated in the figure. Hence,
U0 dy
[U0
u(y)]dy+ 2
v(x)dx
=0,
(7.99)
or
u(y)dy
v(x)dx.
(7.100)
u 0 for |y| , and since h , it

a good approximation to replace the limits of integration
on the left-hand side of the above expression by . Thus,
However, given that
is
from Equation (7.95),

Z
u(y)dy
v(x)dx
Q,
(7.101)
where
Q is independent
retardation
presence
of the flow
of the
necessitates
of
inside
plate, which
a small
x.
Note
that
the
slight
the wake, due to the

is parameterized
by Q,
lateral outflow, v(x), in the region of
the fluid external to the wake.

Let
us now
apply the integral form of the x-component
of the fluid equation of motion to the control volume. For
steady-state, this reduces to (see Section 2.11)

ZS
ZS
vx
v dS = Fx
dS
xj
(7.102)
where F x is the net x-directed force exerted
on the
fluid
within the control volume by the plate. It follows, from
Newtons third law of motion, that F x

the viscous drag force
acting
on
the
plate
per
= D,
where D is
unit width (in the z-direction)
in the
x-direction.
Now,
in
an
incompressible fluid (see Section 2.6),
!
ij
= pij
vi
+
x
(7.111) 136
FLUID MECHANICS
Hence, we obtain
Z
2
0
dy +
?U0
u(y)
?2
dy
+2
U0 v(x)dx
l
v j
xi
D 2
u(y)dy,
(7.104)
dl
since the pressure within the fluid is essentially uniform,

tangential fluid velocity at
to be negligible at x
assumed
(7.101),
as well as the fact
y = h is U0 and v is
= l. Making use of Equation
that Q is independent of l,we
get
D
=U0
Q.
(7.105)
Here,
we
any
u. A
have neglected
in the small quantity
terms that
are
comparison
second-order
with Equation
(7.79) reveals that
1.33(U0
L)
1/2
(7.106)
or
U0
=0.664
Lx
?1/2
(7.107)
Hence, from (7.96) and (7.97), the velocity and vorticity
profiles
across
the layer
are
,
.
vx (x,y)
= 1 0.375
U0
Lx
?1/2
exp(y
2/
2),
(7.108)
and
(x,y)
U0
2/
exp(y
1/2.
2),
Finally,
on the
since the above analysis is premised
are
2( x/U0 )
assumption that
1, it is clear that the previous three
|1vx /U0 | |u|/U0

expressions
?1/2
Lx
(7.109)
where (x)
= 0.749
only
valid
when
L (i.e.,
well
downstream of the plate).

The above analysis only holds when the flow within
the wake is non-turbulent.
Let
us assume,
by analogy
with the discussion in the previous section, that this is the
case as
long
(x)/,
remains
approximately
wake
U0
can
the Reynolds number of the wake, U0

less
be written
10
4/Re.
critical
the Reynolds
2Re
1/2
(x/L)
number
that the wake becomes
value
that
is
number of the
1/2,
where
of the external
Re
flow,
turbulent when x/L
>
Hence, the wake is always turbulent sufficiently
downstream
effectively
some
than
600. Since
L/ is the Reynolds
we deduce
far
as
of
assumes
the
plate.
Our
that the wake
analysis,
which
is non-turbulent
in
some
region, immediately downstream of the plate, whose
extent (in x) is large compared with L, is thus only valid

when 1
Re 9
7.7
4.
Von Karm an Momentum Integral
Consider
boundary layer that forms
rigid stationary
length
10
obstacle
and uniform
cross-section)
uniform, transverse, high Reynolds
represent
arc
on the
of arbitrary
surface of
shape (but infinite

placed
in
steady,
number flow. Let
length along the surface, measured (in the
direction of the external flow) from the stagnation point
that forms at the front of the obstacle. (See Figure 7.11.)

Moreover, let
boundary
represent
distance
across
the boundary
normal to the surface. Suppose that the
layer, measured
layer
is
sufficiently
thin
that
it
is
well
as a plane slab inthe immediate vicinity of a

general point on the surface. In this case, writing the
velocity field within the layer in the form v = u(x,y) ex +
v(x, y) ey it is reasonable to model this flow using the
approximated
slab boundary
layer equations
(7.29)]
u
+
x
0,
[see
Equations (7.28) and
(7.110)
+v
dU
U
dx
2u
,
137
subject to the standard boundary conditions
u(x,)
=v(x,0)
(7.112) u(x,0)
U(x),
0.
(7.113)
Here, U(x) is the external tangential fluid velocity at the

edge of the layer. Integrating (7. 111)
making
use of the boundary
across
?????y=0
=
=
=
=
Z
dU
U
dx
dy
y
"(U
dU
u)
dx
+u
(U u)
(U u)
+v
x
the layer,
conditions (7.113), leads to
dy
y
"(U
dU
u)
(U u)
dy
y
"(U
dU
u)
dx
+u
(U u)
+u
(U u)
dx
Z
dU
(U u)dy
dx
+ (U u)
dy
u(U u)dy.
dx
(7.114)
we have integrated the final term on the right-hand

use of Equations (7.110), (7.112),
(7.113). Let us define the displacement thickness of the
Here,
side by parts, making
and
layer [see Equation (7.70)]

Z
1
(x)
?1
dy,
(7.115)
as well as the so-called momentum

Z
2
(x)
?1
dy.
(7.116)
It follows from (7.114) that
thickness
?????y=0
=U
d2
+U
dx
dU
(1
+ 22
).
dx
(7.117)
This
important
momentum
result
is
as
known
integral, and is fundamental
approximation methods commonly

boundary
layer thicknesses
obstacles
von K arm an
to many of the
the
placed
on
employed
to calculate
the surfaces
in high Reynolds
number
of general
flows. (See
Section 7.10.)
7.8
Boundary Layer Separation
we saw in Section 7.5, when a high Reynolds number

passes around a streamlined obstacle, such as a
slender plate that is aligned with the flow, a relatively thin
boundary layer form on the obstacles surface. Here, by
relatively thin, we mean that the typical transverse (to the
As
fluid
flow) thickness of the layer is

the length
and Re
1/2,
L/Re
where L is
of the obstacle (in the direction of the flow),
the
Reynolds
number
of the
external
flow.
Suppose, however, that the obstacle is not streamlined: i.e.,
the surface of the obstacle is not closely aligned with the

streamlines of the unperturbed
the
typically
observed
flow pattern. In this
behavior
case,
is
illustrated in Figure 7.11, which shows the flow pattern of
high
Reynolds
number
irrotational
fluid
around
cylindrical obstacle (whose axis is normal to the direction

of the unperturbed flow). It
can be seen
front of the obstacle. Moreover,
covers
a stagnation
zero, forms in
that
point, at which the flow velocity is locally
thin boundary
the front side of the obstacle. The thickness
layer is smallest at the stagnation
layer
of this
point, and increases
towards the back side of the obstacle. However, at

point
on the back
some
side, the boundary layer separates from
the obstacles surface to form
transverse dimensions
are
itself. This phenomenon
a vortex-filled
wake whose
similar to those of the obstacle

is known
as
boundary
layer
separation.
138
FLUID MECHANICS
separation point
potential flow streamlines
boundary layer
obstacle
stagnation point
wake
Figure 7.11: Boundary layer separation.
Outside the boundary
layer, and the wake, the flow
pattern is irrotational and essentially

Sec-tion 5.13, the tangential
boundary layer (neglecting
inviscid. So, from
flow speed just outside the
any
circulation of the external
flow around the cylinder) is
U()
2 U0
sin,
(7.118)
where U0
is the unperturbed
flow speed, and is
cylindrical
coordinate
such that the stagnation
= 0.Note that the tangential
corresponds to
point
accelerates
along the
flow)
and
defined
on
(i.e., increases
with increasing
surface of the obstacle, in the direction of the
the front side of the obstacle (i.e., 0 /2),
on
decelerates
the
back
side.
Boundary
separation
/2), is
a always
nd de observed
celerates to take place at
the
flow
arc-length,
surface
of
an
layer
a point on
obstacle
where there is deceleration of the external tangential flow.
pressure just outside the

on the surface of the obstacle,
the pressure is uniform across the layer) is
In addition, from Section 5.13, the

boundary layer (and, hence,
since
P()
=p1 +U
2
0
cos2,
(7.119)
a constant. Note that the tangential pressure

as to accelerate the tangential flow on the
front side of the obstaclethis is known as a favorable
pressure gradient. On the other hand, the pressure
gradient is such as to decelerate the flow on the back
sidethis
is known as an adverse pressure gradient.
where p1 is
gradient is such
Boundary layer separation
at
a point on the
is always observed to take place
surface of
an obstacle
where the
pressure
important
physical
gradient is adverse.
Boundary
phenomenon
layer separation
is
an
because it gives rise to
drag force acting
ona non-streamlined
obstacle placed in a
high Reynolds number flow. This is the
pressure
behind
in the comparatively
a non-streamlined
pressure
the
front
side
onthe backside:
i.e.,
P()
(See Figure 7. 12.) In other words, the resultant
force
on
the front
antpressure
opposite to that
the
over the obstacles surface in

It can be seen that the variation
of
the obstaclemirrorsthat
P()
as a consequence of
more exact, in the case
obstacle,
variation
the absence of separation.
on
because the
obstacle, Equation (7. 119) specifies
cylindrical
expected
case
wide wake that forms
separation, is relatively low. To be

of
greatly enhanced
on
side of the obstacle

the back side,
so
pressure
is equal and
that the
pressure
gives rise to
distribution
obstacle.
Figure
distribution
7.12
is modified
layer separation. In this
zero net
illustrates
the
on the
pressure
of boundary
less than that
side of the obstacle. Consequently,
on
how
as a consequence
case, the pressure
separation points is significantly
force
drag acting
between the
on the front
pressure
the resultant
the front side is greater in magnitude
on the back side, giving

on the obstacle. Let D be
than the
oppositely directed force
rise to
significant drag acting
the drag
force
per
unit width (parallel to the axis of the cylinder)
otational)in the absence of boundarylayer
Incompressible
Boundary Layers
139
0
/2
P() p1
3/2
stagnation point
separation points
Figure 7.12: Pressure variation
over surface
of a
cylindrical obstacle in a high Reynolds number flow

both with (dashed curve) and without (solid curve)
boundary layer separation.
parameterize this force in terms of a dimensionless drag

coefficient,
CD
D
U
2
0
(7.120)
where is the fluid density, and
a the
typical transverse
size of the obstacle (in the present example, the radius of

the
cylinder).
The
non-streamlined
number
drag
obstacle
as a
flow,
in
consequence
separation, is generally characterized

of order unity. The exact
strongly
weakly
of
the
obstacle.
Form
area
cross-sectional
Reynolds
boundary
of
a drag
layer
coefficient
obstacle, but only relatively

number
this type of drag
since it depends primarily
high
value of the coefficient depends
on the shape of the

on the Reynolds
Consequently,
by
on a
acts
that
force
placed
on the
drag
the
of
flow.
is termed form drag,
external
scales
shape,
roughly
or form,
as the
(per unit width) of the vortex-filled
wake that forms behind the obstacle.

Boundary
layer separation
is associated
with strong
pressure gradients, or, equivalently, strong flow

decel-eration, on the back side of an obstacle placed in a
high Reynolds number flow. Such gradients can be
adverse
significantly
closely
reduced by streamlining the obstacle: i.e., by
aligning
streamlines of
the surface
thereby
with the unperturbed
can be delayed, or even completely prevented,

of a sufficiently
streamlined
obstacle,
significantly decreasing, or even eliminating, the
separation
on
its back surface
the external flow. Indeed, boundary layer
associated
form
cross-sectional
(essentially,
drag
area
limit that the form drag is reduced

there is still a residual drag acting
boundary
layer
by
viscosity.
(7.79)
even
This
and
the
in the
to a negligible level,
on the obstacle due to
type
called friction drag. As is clear from

Equations
reducing
of the wake). However,
drag
of
comparison
is
of
(7.120),
the drag coefficient

1/2),
associated with friction drag is O(Re
where Re is the
Reynolds number of the flow. Friction drag thus tends to
zero as the Reynolds

The
number tends to infinity.
phenomenon
of
boundary
layer
separation
us to resolve dAlemberts paradox. Recall, from

Sec-tion 5.13, that an idealized fluid that is modeled as
inviscid and irrotational is incapable of exerting a drag
force on a stationary obstacle, despite the fact that very
high Reynolds number, ostensibly irrotational, fluids are
observed to exert significant drag forces on stationary
allows
obstacles.
The
realization
that, in such
resolution
of
the
paradox
fluids, viscosity
lies
can
in the
only
be
neglected (and the flow is consequently only irrotational) in

(7.126)
140
FLUID MECHANICS
case, the region of the fluid in which

a significant role is localized to a thin
layer on the surface of the obstacle, and the
1/2,
friction drag scales as Re
and, therefore,
separation. In this
viscosity
boundary
resultant
plays
disappears
in the
inviscid limit (essentially,
boundary layer shrinks to
zero
the other
boundary
hand, if the
because the
thickness in this limit). On
separates
layer
viscosity is important both in a thin boundary layer

front
of
the
obstacle,
wake
vortex-filled,
and
that
in
forms
wide,
limit. The
presence
low-pressure,
behind
Moreover, the wake does not disappear
then
on the
the
obstacle.
in the
inviscid
of significant fluid vorticity within the
wake invalidates irrotational fluid dynamics. Con-sequently,

the
pressure
on
the
smaller
significantly
back
side
of
the
than that predicted
obstacle
is
by irrotational
pressure force on the

on the back side, and a
on the obstacle. The drag
fluid dynamics. Hence, the resultant

front side is larger than that
significant
drag
is exerted
coefficient associated with this type of drag is generally of
order unity, and
does not tend to
zero as
the Reynolds
number tends to infinity.
7.9
Criterion for Boundary Layer
Separation
As
we
have
(7.110)(7.113)
seen,
the
boundary
layer
equations
generally lead to the conclusion that the
tangential
velocity in
compared
with the normal velocity,
speaking,
this
result
thin boundary
holds
layer, u, is large
v. Mathematically
everywhere
except
in the
immediate
vicinity
of singular
then it follows that the fluid
predominately
can
to the surface of the obstacle, and

from this surface to
effectively
precludes
surface. Hence,
at
a very
points. But, if
moves
only
parallel
move away
limited extent. This restriction
separation
we conclude
the flow
of
that separation
from the
can only occur
point at which the solution of the boundary
layer
equations is singular.
As we approach
a separation
point,
we expect
the flow
to deviate from the boundary layer towards the interior of

the fluid. In other words,
we expect
the normal velocity to
become comparable with the tangential velocity. However,
we
seen
have
that the ratio v/u is of order Re
Equation (7.18)]. Hence,
that
implies
an
an increase
increase
employ
increases by
equations, (7.23)
formally
an
the dimensionless
[see
v to such a degree
a factor Re 1/2. For

we may suppose that v
factor. Indeed, if we
by
sufficiently large Reynolds numbers,

effectively
of
1/2
infinite
form of the boundary
, ,
layer
(7.27), the situation just described
equivalent
to
an
infinite
dimensionless normal velocity, Vy
Let the separation point lie at
value
of
is
the
at the separation point.
x = x0
correspond to the region of the boundary
and let
x < x0
layer upstream
of this point. According to the above discussion,
v(x0
y)
(7.121)
at all
(except, of
course, y
= 0, where
the boundary
conditions at the surface of the obstacle require that
0).It follows that the deriviative v/y is also infinite at
= x0
= 0,
Hence, the equation of continuity, u/x
= ,
+ v/y
or x/u = 0, if x is
y. Let u(x0 y) = u0 (y).
Close to the point of separation, x0 x and u u0 are
small. Thus, we can expand x0 x in powers of u u0 (at
implies that (u/x)x=x0
regarded
as a
function of
fixed y). Since
and
= 0, the
(x/u)u=u0
first term in this
expansion vanishes identically, and we are left with
x0
x = f(y)(u
h(u
u0 )
+O
u0 )
(7.122)
or
u(x,y) u0 (y) + (y)
x0
x,
(7.123)
where
=1/
p
f is some function of y. From the equation
of continuity,
v
(7.124)
(y)
x0
Upon integration, the above expression yields

(y)
v(x,y)
x0
(7.125)
where
(y)
12
(y
)dy
is,the Incompressible Boundary Layers
141
The equation of tangential motion inthe boundary layer,
(7.111), is written
+v
x
=U
dU
2u
+
dx
(7.127)
As is clear from Equation (7.123), the derivative
does not become infinite
as x
x0
The
same
2u/y
is true of
the function UdU/dx, which is determined from the flow
outside the boundary layer. However, both terms
on the
as x
left-hand side of the above expression become infinite
x0
Hence, in the immediate vicinity of the separation
point,
+ v
x
0.
(7.128)
Since u/x
= v/y,
we can
rewrite this equation in the
form
?
= u
+v
u
y
0.
(7.129)
Since
udoes
not, in general, vanish at x
that
?
= x0
we conclude
0.
(7.130)
Inother words, v/u is a function of
x only. Now, from (7.
123) and (7.125),
(y)
x0
=
u0 (y)
+ O(1).
(7.131)
Hence, if this ratio is a function of x alone then (y)
/2)Au0 (y), where A is a constant: i.e.,
=(1
Au0 (y)
v(x,y)
x0
(7.132)
Finally, since (7.126) yields
= 2d/dy =Adu0 /dy, we
obtain
u(x,y) u0 (y) + A
du0
x0
x.
dy
(7.133)
The previous two expressions specify

of
x and y near
of separation,
uand v as functions
the point of separation. Beyond the point
that is for
x >
since
x0
the
the expressions
square roots
are
physically
meaningless,
become
imaginary.
This implies that the solutions of the boundary
layer equations cannot sensibly be continued beyond the

separation
point.
Now, the standard boundary

of the obstacle
require that
conditions at the surface
= v = 0 at
= 0. It,
therefore, follows from Equations (7.132) and (7.133) that
u0 (0)
(7.134)
0,
?????y=0
du0
0.
dy
(7.135)
Thus,
we
obtain the important
tangential velocity,
zero at
result
u,
prediction
the separation point (i.e.,
was originally obtained

we have used to derive
argument
Note
that
if
the
x = x0
and
are
y = 0). This
by Prandtl, although
the
it is due to L.D. Landau.
constant
(7.132) and (7.133) happens to be
x0 and y
that both the
and its first derivative, u/y,
=0, at which the derivative
in
zero
expressions
then the point
x=
no
u/y vanishes, has
and is not a point of separation.

no reason, in general, why A should take
the special value zero. Thus, in practice, a point on the
surface of an obstacle at which u/y = 0 is always a point
particular
properties,
However, there is
of separation.
Incidentally, if there
= x0
> x0
(i.e., if A
In other words,
move away
45)
were no separation at
the point
= 0) then we would have u/y
< 0 for x
u would become negative as we

y being still small. That is,
from the surface,
142
FLUID MECHANICS
fluid beyond the point
x = x0
would
move
tangentially, in
the region of the boundary layer immediately

the surface,
in the direction
opposite
adjacent to
to that of the
external flow: i.e., there would be back-flow
in this
region. In practice, the flow separates from the surface at x
=x0
and the back-flow migrates into the wake.
Note that the dimensionless
are
(7.23)(7.27),
the external
independent
flow (assuming
boundary layer equations,

of the Reynolds number of
that this number
is much
greater than unity). Thus, it follows that the point

surface
independent
location
separation
flow inthe boundary
on the
0 is also
of the Reynolds number. In other words, the
of the
Reynolds number (as long
At
at which u/y
of the obstacle
point
as this
is independent
of the
number is large, and the
layer is non-turbulent).
= 0, the
equation
of tangential
motion in the
boundary layer, (7.111), is written
2u
??????y=0
1 dU
U dx
1 dP
dx
(7.136)
where P(x) is the
pressure
just outside the layer, and use has
been made of (7.6). Now, since
u is positive,
and increases
away
from the surface (upstream of the separation point), it

2u/y 2)y=0
follows that (
> 0 at the separation point itself,
where (u/y)y=0
= 0. Hence,
according
to the above
equation,
!x=x0
dU
<
0,
dx
!x=x0
dP
>
(7.137)
0.
dx
(7.138)
In other words,
we
predict that the external tangential flow
is always decelerating at the separation point, whereas the
pressure
gradient
decelerate
experimental
7.10
the
is always
tangential
adverse
in
flow),
(i.e., such
as to
agreement
with
observations.
Approximate Solutions of Boundary
Layer Equations
The boundary layer equations, (7.110)(7.113),take
the form
u
+
x
dU
+v
x
U
y
dx

u(x,)
u(x,0)
=
=
U(x),
0,
v(x,0)
0.
Furthermore, it follows from (7.140), (7.142), and (7.143) that
0,
(7.139)
2u
(7.140)
(7.141)
(7.142)
(7.143)
2u
??????y=0
= U
dU
dx
(7.144)
can be thought of as an alternative

we saw in Section 7.4, the boundary
The above expression
form of (7.143). As
layer equations
special form U0
can be solved
x m However,
exactly when U(x) takes the

in the general
resort to approximation methods.

Following Pohlhausen, let us assume that
case, we must
u(x,y)
U(x)
(7.
158)
Incompressible
Boundary
f(),
Layers
143
where
= y/(x),
and /x 1/. In particular,
suppose
that
(
f()
a + b + c 2 + d 3 + e 4
=
0
a, b, c,
where
automatically
Moreover,
imply that
>1
d, and
(7.146)
e are constants.
satisfies the boundary condition (7
the boundary
conditions
. .
This expression
14 1)
(7. 142) and (7. 144)
a =0,and
f
(0)
(x),
(7.147)
where
=d/d, and
=
Finally,
let
us assume
that
are continuous at =1:i.e.,
f, f
dU
dx
(7.148)
and
f(1)
1,
=
=
(7.149)
f (1)
(7.150)
(1)
0,
0.
(7.151)
These
constraints
requirements
that
to
corresponds
the velocity,
the
vorticity,
stress tensor, respectively, be continuous

Given
that
0, Equations
reasonable
and viscous
across
the layer.
(7.146),
(7.147),
and
(7.149)(7.151)yield
f()
F()
G()
(7.152)
for 0
1, where
F()
1
(1 )
(1+ ),
G()
(7.153)
(1 )
3.
6
(7.154)
Thus, the tangential velocity profile
function of
single parameter,
across
the layer is
, which is termed
the Pohlhausen parameter. The behavior of this profile is

illustrated
in Figure
7.13. Note
circum-stances, the Pohlhausen
range
f()
12
>
1 for
12. For
some
<
that, under
parameter
> 12, the
profile is such that
1, which is not possible
steady-state solution. On the other hand, for
(0)
profile
stead yisstat
suchethat
solu ftion
close to the wall. As
normal
must lie in the
in
< 12,
the
<O0,nwhich
implies
reversal
th e othe
rhanflow
d,for
<
we
have
seen,
flow
reversal
is
indicative of separation.
to
0, corresponds
(7.152) is only
an
(0)
Indeed, the separation point, f
12.
Note
that
expression
some,
approximation, since it satisfies
but not all, of the boundary conditions satisfied by the true

velocity
profile. For instance,
of (7. 140)
differentiation
y reveals that ( 3u/y 3)y=0

which is not the case for expression (7.152).
with respect to
It follows
from Equations
(7.115),
= 0,
(0)
(7.116),
and
(7.152)(7.154)that
Z
1
(x)
(1 f)d
10
(7.155)
Z
f (1 f)d
37
120
315
(x)
945
9072
(7.156)
Furthermore,
?????y=0
f (0)
?2
(7.157)
Now, the
von K arm an momentum
rearranged to give
integral, (7.1
17), can be
d2
dx
2
2
dU
dx
2+
?????y=0
2
ura t e on
re d u ce
144
FLUIDMECHANICS
1
0.9
0.8
0.7
0.6
f
0.5
0.4
0.3
0.2
0.1
0
0
0.2
0.4
0.6
0.8
Figure 7.13: Pohlhausen velocity profiles for
12 (solid curve) and
= 12 (dashed
curve).
Defining
(x)
2
2
dU
dx
(7.159)
we obtain
dx
dU/dx
=2[F2 () {2 + F1()}] = F(),
(7.160)
where
37
315
945
9072
F1()
!,
1
2
37
120
315
10
F2 ()
!2
945
9072
?????y=0
2
?2
37
315
37
F()
945
315
9072
!"2
116
945
9072
315
+
945
(7.161)
(7.162)
(7.163)
#
120
9072
It is generally
necessary to integrate
(7.164)
Equation (7.158) from
the stagnation point at the front of the obstacle, through the

point of maximum
point
on
stagnation
implies
the back
point
we
that F()
maximum tangential
tangential velocity, to the separation

side
of the obstacle.
have U
0. Furthermore,
velocity
0, which implies that
Now,
= 0 and dU/dx
we have
at the
at the point
dU/dx
= = 0. Finally,
0, which
of
= 0 and U
as we
have
already
seen,
12
at
implies, from (7.161), that

As
was
by the linear
physically
particularly
(where
point, which
first pointed out by Walz, and is illustrated in
Figure 7.14, it is
F()
the separation
= 0.1567.
fairly good approximation to replace

function
0.47
6 for
in the
range.
by theThe
lin ear
approximation
fun ctio n0.47 is
accurate on the front side of the obstacle

> 0). Making use of this approximation,
relevant
(7.170)
145
0.15
0.1
0.05
0.05
Figure 7.14: The function F() (solid curve) and the

linear function 0.47 6 (dashed line).
0.1
to the linear differential equation
dx
which
= 0.47
2
2
assuming that the stagnation point corresponds to x
Recall that the separation point corresponds to x
5(x )dx
= 0.It follows that

x
0.47 dU
U
2
2
dx
0.47
dU
=xs
U (x
dx
)dx
where (xs ) s
(7.165)
(7.166)
0.1567.
Suppose
uniform
flow
that
U(x)
over a
U0
(7.167)
which
corresponds
flat plate. (See Section
to
7.5.) It
follows from Equations (7.166) and (7.167) that

2
(x)
x
where Re
= U0
x/, and
= 0.Moreover,
0.69
Re
1/2
(7.168)
according to
Equations (7.148) and (7.162),
= 0 and 1
Hence, the displacement
of the boundary
width
/2
= 2.55.
layer
becomes
1
(x)
1.75
Re
1/2
(7.169)
This approximate
result
compares very
favorably with the
exact result, (7.73).

Suppose that
x = a and U()
= 2U0
sin , which
corresponds to uniform transverse flow around

cylinder of radius
a. (See
a circular
Section 7.8.) Equation (7.167)
yields
Z
()
=0.47
sin
(7.173) 146
FLUID MECHANICS
0.2
cos
6
sin
0.1
20
40
60
)
(
80
100
Figure 7.15: The function () for flow around

cylinder.
120
a circular
Figure 7.16: Flow
over the back surface
of a semi-infinite
wedge.
Figure 7.15 shows () determined from the above formula.
It can be seen that
=s = 0.1567when
= s
108
other words, the separation point is located 108
In
from the
stagnation point at the front of the cylinder. This suggests

that the low
pressure
as
the cylinder
wide
as
wake behind the cylinder is almost
Suppose, finally, that U
then,
as
= U0
xm
If
is negative
in Figure 7.16, this corresponds
illustrated
uniform flow
itself, and that the associated
form drag is comparatively large.
over the back surface
to
of a semi-infinite wedge
whose angle of dip is
1
+m
(7.171)
(See Section 7.4.) It follows from (7.167) that
0.47m
1
+ 5m
0.47
/2 4
(7.172)
Now,
we
expect boundary layer separation

wedgewhen
where
< s
= 0.1567.
onthe back surface
This corresponds
of the
to
> s
(s )
0.47
+ 4(s
13
147
Hence, boundary
layer separation
can
be prevented
by
making the wedges angle of dip sufficiently shallow: i.e.,

by
streamlining
reducing
the wedge,
the deceleration
which
has the effect
of the flow
back surface. Note that the critical value

0.0125)
of
on the wedges
of m (i.e., ms
=
at
occurs in our approximate =

solution
0.0125is
very
a t whic
similar
hse
to parati
the critical
onoccurs
value in
of m
our
(i.e.,
apprmoximate
= 0.0905)
so lu
tion
at which
is v ery the exact self-similar
solutions described in
can no
be found.
that the
.0905)
Sectiona t
7.4
w hich
thelonger
exa ct self-si
mila This
r solsuggests
uti on s descri
be
absence of self-similar solutions for m < m is related to
which separation
boundarylayer
7.11
separation.
Exercises
7.1. Fluid flows between two non-parallel plane walls, towards
a manner that if x is
a wall from the intersection of the planes
then U(x) = U0 /x,where U0 is a positive constant. Verify that
a solution of the boundary layer equation (7.35) can be
found such that is a function of y/x only. Demonstrate that
the intersection of the planes, insuch
measured along
this
solution yields
"?
u(x,y)
U(x)
where
u= /y,
U0
=F
?1/2
yx
and
= 1,
F2
subject to the boundary conditions F(0)
= 0 and F() = 1.
Verify that
F(z)
= 3 tanh
is a suitable solution of the above differential equation,

where tanh
= 2/3.
7.2. A jet of water issues from a straight
narrow
slit in a wall,
and mixes with the surrounding water, which is at rest. On the

assumption that the motion is non-turbulent
and
two-dimensional, and that the approximations of boundary

layer theory
apply, the stream function satisfies the boundary layer
equation
+
y
xy
= 0.
Here, the symmetry axis of the jet is assumed to run along

the x-direction, whereas the y-direction is perpendicular to this
axis. The velocity of the jet parallel to the symmetry axis is
u(x,y)
where u(x,y)
= u(x,y), and u(x,y) 0 as y . We
expect the momentum flux of the jet parallel to its symmetry axis,
Z
u2
dy,
to be independent of x.
Consider
a self-similar stream function
=0
xp
of the form
F(y/x ).
Demonstrate that the boundary layer equation requires that
p+q
= 1, and that Mis only independent of x when 2 pq = 0.

Hence, deduce that p = 1/3 and q =2/3.
Suppose that
(x,y)
= 6 x
1/3
F(y/x
2/3
).
Demonstrate that F(z) satisfies
subject to the constraints that F
=0,
= F (z), and F (z) 0 as z
+ 2F F + 2F 2
(z)
Show that
F(z)
= tanh(z)
is a suitable solution, and that

M
= 48
148
FLUID MECHANICS
7.3. The growth of
sucking
some
a boundary
of the fluid through
can be inhibited by
a porous wall. Consider
layer
conventional
boundary layer theory. As a consequence of suction, the
modified
to v(x,0)
on the normal
boundary condition
= vs
where
vs
velocity at the wall is
is the (constant) suction
velocity. Demonstrate that, inthe presence of suction, the von
K arm
an
velocity integral becomes
?????y=0
=U
d2
+U
dx
dU
+ 22 )+ Uvs
(1
dx
Suppose that
(
u(x,y)
= U(x)
where
sin(y)
Demonstrate that the displacement
momentum widths of the boundary layer

1
2
/(2)
y > /(2)
=(x).
=
=
are
(/2 1)
(1 /4)
1
1
,,
and
respectively. Hence, deduce that

(/2 1)
= U(1 /4)
d1
Consider
= U0
a boundary
dU
+
dx
layer
+ (/2
dx
!1/2
= (/2 1)
ona flat plate, for which U(x)
Show that, inthe absence of suction,
1)vs
8
4
!1/2
U0
but that inthe presence of suction

1
(/2 1)
vs
Hence, deduce that, for a plate of length L,suction is

capable of significantly reducing the thickness of the boundary
layer
when
vs
U0
where Re
= U0
Re
1/2
L/.
Incompressible Aerodynamics
149
8 Incompressible Aerodynamics
8.1
Introduction
This chapter investigates the forces exerted
on a stationary
obstacle situated in a uniform, high Reynolds number wind,
on
the
assumption
streamlined
boundary
termed
fundamental
apart from
no
the
obstacle
appreciable
is
sufficiently
separation of the
layer from its back surface. Such
an airfoil
an obstacle
to the theory of flight. The flow around
a thin
is
(or aerofoil). Obviously, airfoil theory is
airfoil is essentially
and
that
that there is
irrotational
an
and inviscid everywhere
boundary layer localized to its surface,
thin wake
emitted
by its trailing
edge. (See
Sections 7.5 and 7.6.) It follows that, for the flow external
to the boundary layer and wake, we can write
v = ,
(8.1)
whichautomaticallyensuresthattheflowisirrotational.
Assumingthattheflowis
alsoincompressible,sothatv
0, the velocity potential, , satisfies Laplaces equation: i.e.,
=0.
(8.2)
The appropriate
boundary condition at the surface of the
airfoil is that the normal velocity be zero. In other words,
= 0, where
=
is
unit vector
normal to the
where
n is a velocity
u ni t veat
ctothe
rnormal
surface. n
In
general,0,the
tangential
airfoil
obtained
surface,
by solving
= 0 in
=
the external
airf
oil surfac
e, obtained
region,
subject
to the boundary
by solving
condition
0
nin

the ext ernal
0 on
the surface, is
the
no
non-zero.
Of
course,
this is inconsistent with
slip condition, which demands that the tangential
velocity be zero at the surface. (See Section 7.2.) However,
as
described in the previous chapter, this inconsistency
resolved
tangential
by
velocity
non-zero on
interfaces
the
boundary
layer,
is effectively
the outer
edge
with the irrotational
across
which
discontinuous,
of the
layer
flow), and
is
the
being
(where
zero on
it
the
inner edge (where

discontinuity
it interfaces
in the tangential
with the
velocity
airfoil). The
across
implies the presence of bound vortices covering
the layer
the surface
of the airfoil (see Section 8.7), and also gives rise to
a friction
drag acting
on the
airfoil in the direction of the
external flow. However, the magnitude of this drag scales

as Re 1/2, where Re is the Reynolds number of the wind.
(See Section 7.5.) Hence, such drag becomes negligibly
small
in the
high
Reynolds
we shall assume
following,
that
number
limit.
In the
any form drag, due to the
residual separation of the boundary
layer at the back of
the airfoil, is also negligibly small. Moreover, for the sake

of simplicity,
we
two-dimensional
number
wind
shall initially
situations
transversely
flows
our discussion to
a high Reynolds
around a stationary
restrict
in which
airfoil of infinite length (in the z-direction)
and uniform
cross-section (parallel to the x-y plane).
8.2
Theorem of Kutta and Zhukovskii
Consider
uniform
two-dimensional
wind
(clockwise)
of speed
airfoil that is at rest in
V whose
direction
subtends
a
a
angle with the negative x-axis. It follows that
the wind velocity

ang le iswith
V
the
Vnega
cos
tivexe+
x-axi
V sin
s.Iteyfollo
and
ws
the
th
corresponding complex
velocity
is dF/dz
Section 6.4.) Now, the air velocity
a great
= V e i.
(See
distance from
the airfoil must tend toward this uniform velocity. Thus, for
sufficiently large |z|, we
dF
=V e
Az
dz
z2
can write
(see Section 6.4)
+ .
(8.3)
According to Equation (6.87), the circulation, , of air
about the airfoil is determined by performing the integral
IC
dF
Re
dz
dz
(8.4)
around
loop C that lies just above the airfoil surface.
as
However,
integral
that
discussed
in Section 6.8, the value of this
is unchanged if it is performed around
can
passing
be continuously
through
singularity
the
deformed
airfoil
of the complex
(8.13)
any
loop
onto C, whilst
surface,
or
crossing
velocity, dF/dz (i.e.,
not
a
line
150
FLUID MECHANICS
filament).
Since (in the high Reynolds
which the boundary layer and the wake

there
are no
external
infinitely thin)
sources or z-directed vortex filaments

airfoil, we can evaluate the integral
line
to the
number limit in
are
around
a large
follows that
= Re
circle of radius R, centered
z =Re
and dz
=iRe
hV
Re
i(+)
+ A + O(R
1)
on the origin.
It
=izd. Hence,
=2Im(A),
(8.5)
which implies that
dF
dz
=V e
+i
+
2z
z2
+
(8.6)
at large |z|.
As discussed in Section 6.9, the net force (per unit

length) acting
on the airfoil, L=X ex + Y ey
is
determined by performing the Blasius integral,

IC
1
i
2
dF
!
2
dz
dz
= X iY,
(8.7)
a loop C that lies just above the airfoil surface.

However, as before, the value of the integral isunchanged
around
if
we perform it instead around a large circle of radius

on the origin. Now, far from the airfoil,
centered
R,
dF
!
2
dz
V
2i
+i
V e
V Be
z2
+
z
+ O(z
3).
(8.8)
So, we obtain
X iY
"V
I
12
Re
i(+2)
+i
V e
+ O(R 1)
= ie
V ,
X + iY
(8.9)
=ie
=e
or
i(/2)
V .
(8.10)
In other words, the resultant force (per unit length) acting
on the
airfoil is of magnitude
V , and has the direction
obtained by rotating the wind vector through

in the
sense
force is known
result
as
lift, and is responsible
(8. 10) is known
Zhukovskii,
a right-angle
opposite to that of the circulation. This type of
as
the theorem
for flight. The

of Kutta
and
after the German scientist M.W. Kutta, and
the Russian scientist N.E. Zhukovskii, who discovered
it
independently.
independent
Note that (at fixed circulation)

of the shape
of the airfoil.
the lift is
Furthermore,
to the KuttaZhukovskii theorem, there is
according
drag acting
on the
airfoil (i.e.,
zero
a small
direction of the wind). In reality, there is always

friction drag due to air viscosity,
small form drag due to residual
as
well
as a
(hopefully)
separation of the boundary
layer from the back of the airfoil. There is actually
type of drag, known
as induced
zero
force acting in the
a third
drag, which is discussed in
Section 8.8.
per
As discussed in Section 6.9, the net moment

length (about
M, acting
on
unit
the airfoil is
by performing the integral

IC
Re
12
dF
!
2
determined
the origin),
zdz
dz
(8.11)
around
before,
a loop C that lies just above the airfoil surface. As

we can deform C into a circle of radius R,centered
on the origin, without

Hence, we obtain
changing the value of the integral.
"V
I
M
+i
=Re
V Re
1
i
2
i(+)
V Be
2i(+)
+ O(R
1)
(8.12)
or
M
h2V
=Re
Be
i(/2)
(8.25)
151
8.3
Cylindrical Airfoils
us
For the moment, let

where
circular
= + i.
work in the complex
-plane,
a cylindrical airfoil
radius a, centered on the
Consider
cross-section
of
with
origin,
that is situated in a uniform, high Reynolds number wind

of speed V whose direction subtends
a (clockwise)
angle
with the negative -axis. Let be the circulation
of air
around the airfoil. A slight generalization of the analysis of

Section 6.4 reveals that the appropriate
complex
velocity
potential is
F()
=V
a2
+i
ln
(8.14)
whereas the associated stream function takes the form
(r,)
=V
a2
r
sin(
+ ) +
ln
r
a
where
=re
(8.15)
It follows that
dF
Ve
Va
+i
Comparison with Equation (8.6) (with
e i
2
z = ) reveals
that
(8.16)
a 2 e i.
(8.17)
Hence, Equations (8.10) and (8.13) yield

?cose
(8.18) L
?sine
+ sine
+ cose
(8.19) M
V
V
,,
0,
(8.20)
where V is the wind vector, L the lift vector, and M the
moment of the lift vector about the origin. We conclude

that, for
a cylindrical
airfoil of circular cross-section, the
lift vector is normal to the wind vector, and the line of

action
of
the lift
cross-section
passes
through
the centroid
(since the lift generates
zero moment
of the
about
the origin). See Figure 8.1.

Of
course, a cylindrical
is completely
airfoil of circular cross-section
unrealistic, since its back side (i.e., the side
opposite to that from which the wind is incident) is not

sufficiently
streamlined
to
prevent
boundary
layer
separation.
(See
as
7.) However,
Chapter
described
in
Section 6.7, we can use the conformal transformation

2
= +
(8.21)
to transform
-plane
z-plane.
to
cylinder
cylinder
(Note
that
of circular
cross-section
in the
of elliptical
cross-section
in the
both cross-sections
located at the origin.) Moreover,

elliptical
cross-section
centroids
cylindrical
airfoil of
that is sufficiently
whose major axis subtends

the incident wind
have
a sufficiently
elongated,
direction, constitutes
and
small angle with

realistic airfoil,
since its back side is, for the most part, closely aligned
with the external flow.
An elliptical airfoil of width
c and
thickness
< c,as
shown in Figure 8.2, is obtained when the parameters
are given the following values:

and l
a
+ ),
(c
4
(8.22)
1 2
(c
4
2) 1/2
(8.23)
In this
case,
the
surface
of
the
airfoil
satisfies
the
parametric equations
c
cos,
2
(8.24)
sin.
152
FLUID MECHANICS
Figure 8.1: A cylindrical airfoil of circular
cross-section.
Figure 8.2: A cylindrical airfoil of elliptical cross-section.

153
In particular,
the
correspond to
=0 and =, respectively.
leading
airfoils
and
trailing
edges
According to Equations (8.16) and (8.21), the complex

velocity inthe z-plane is given by
"V
dF
dF d
dz
+i
Va
dz
=ae
Thus, on the airfoil surface, where

"V
dF
dz
=i
(c
l2
(8.26)
(c + )
+ )
we obtain
+ ) +
sin(
e i
( cos
+ ic sin )
(8.27)
A long
way
from the airfoil,
z l2/z, so that
Equation
(8.26) reduces to
dF
Ve
V (l
+i
e i
2z
dz
a 2 e i)
(8.28)
A
comparison
with
Equation
(8.6)
reveals
circulation of air around the airfoil takes the
same
that
the
value, ,
in both the complex - and z-planes. In other words, the

conformal
circulation.
transformation
(8.21)
does not
Note that the transformation
modify the external wind speed
or
modify
the
also does not
direction [since, from
(8.16) and (8.28), dF/d
= dF/dz = V e
at
very
large ||
and |z|]. On the other hand, it is clear that the constant B,
which
takes the value
|
and |z|]
zero
the value
in the complex -plane, takes
in
the
complex
z-plane.
V (l
e i
a 2 e i)
(8.29)
Hence, Equations (8.10) and (8.13) reveal that
(8.31)
(8.30) L
8
V
(c
V ek
2)
,,
sin(2),
(8.32)
where V is the wind vector, L the lift vector, and M the
moment of the lift vector about the origin. Here,

ek
= cos ex
sin
ey
(8.33)
is a unit vector parallel to the incident wind direction, and
is
= ek
unit vector perpendicular
conclude
that,
cross-section,
for
ez
= sin ex
+ cos
ey
(8.34)
to the wind direction. We
cylindrical
airfoil
of
elliptic
the lift vector is normal to the wind vector,
but the line of action of the lift intersects the major axis
of the airfoil
distance
d
14
sin
(c
L cos
(8.35)
in front of the cross-sections
+ )].
/[ V (c
F,bat
centroid, C, where
which the line of action of the lift intersects
airfoils
See Figure 8.2. Incidentally, the point,
major axis is conventionally
the
termed the focus of
the airfoil.
8.4
Zhukovskiis
Hypothesis
on a
in a
on the
can we
According
to the previous analysis, the lift acting
cylindrical
airfoil of elliptic
uniform,
high Reynolds
circulation,
cross-section,
situated
number wind, depends
of air about the airfoil. But, how
determine the value of this circulation?

Figure 8.3 shows the boundary
layer and wake of
streamlined airfoil. The boundary layer, which is localized
on
the surface of the airfoil, has
length in the z-direction
154
FLUID MECHANICS
dS
a vortex
intensity
per unit
equal to U, where U is the
VU
Wake
Boundary Layer
Trailing Edge
Airfoil
Figure 8.3: The boundary layer and wake of a streamlined

airfoil. Only shaded regions
speed immediately
posses non-zero
vorticity.
above the layer. [See Equation (8.64).]
Moreover, the wake is emitted by the airfoils trailing edge,

and subsequently
convected by the external air flow. (See
Section 7.6.) Note that the flow is irrotational everywhere
apart from inside the boundary layer and the wake. Now,
to the analysis
according
change
encloses
the
airfoil
z-directed vorticity
is equal
across
5.13, the rate of
of Section
of the circulation, , around
this
some curve
to minus
the
curve: i.e.,
d
C that
flux
of
dS.
dt
(8.36)
Here,
an
v is the
outward
z-direction)
wind velocity, the wind vorticity, and dS

surface
lying
independent
element
(of
on C. We expect
unit
depth
in
the
the vorticity flux to be
of the size and shape of C, otherwise
the
circulation of the flow, , about the airfoil would not have
unique value. In the limit that C becomes
very
large,
V, where V is the incident wind velocity. Thus,

d
dt
V z
(8.37)
where V is the wind speed, and z
the vortex intensity
per
unit length in the wake (at the point where it
the
curve
C). Here,
we are
assuming
crosses
that the vorticity
within the wake is convected by the flow, giving rise to
across
net flux of vorticity
an extension
assume that
proportional
C.Since the wake is essentially
of the boundary
vortex
its
layer, it is reasonable
intensity
per
unit
length
to
is
to that in the boundary layer at the airfoils
trailing edge, where the wake and boundary layer intersect.
= k U0
In other words, z
velocity immediately
where U0
is the tangential
above the trailing edge of the airfoil,
and k is a constant. It follows that
d
dt
= kV U0
(8.38)
According to Equation (8.27), the tangential velocity just
above the surface of the airfoil is

"
U()
=Re
"V
sin(
#||=a
c sin
(c 2 sin
+ ) +
cos
dF
+ 2 cos 2 ) 1/2
dz
(c
(c
+ )
(c 2 sin
+ )
+ 2 cos 2 ) 1/2
rfoils
155
y/c
2
2
x/c
Figure 8.4: Streamlines around
a slender
cylindrical
airfoil of elliptic cross-section situated ina uniform,
high Reynolds number wind.
The parameters for this calculation
0.1,
=/12, and =0.
are /c =
to
Hence, given that the airfoils trailing edge corresponds
=, we obtain
# ?
"
U0
(c
c +
V sin
+ )
(8.40)
Thus, Equation (8.38) yields
d b
dt
where
= /[ V (c + )], t = t/t0
+ sin,
(8.41)
and t0
= /(k V).
Assumingbthat the circulation of the flow about the airfoil
is initially
zero (i.e.,
=0 at t = 0),the above equation can
be solved to give
?1
exp( t)
sin
(8.42)
b
Clearly,
as
the
normalized
circulation
asymptotes to the constant value

b
sin.
b
(8.43)
The corresponding
circulation is
constant
value of the unnormalized
= V (c
) sin
(8.44)
Note
that, according
to Equation
circulation, , takes the value
(8.40),
when
the
the tangential velocity
zero. In other words,

up around the airfoil is
such as to render its trailing edge a stagnation point of the
flow. This
conclusion
is known
as Zhukovskiis
at the airfoils training edge, U0
the steady-state
circulation
set
is
hypothesis, after its discoverer N.E. Zhukovskii.

Incidentally,
it
should
be
clear,
from the
above
discussion, that the air circulation about the airfoil is only

able to change its value because of the
presence
of the
boundary layer, and the associated wake that trails from

(8.46)
156
FLUID MECHANICS
y/c
2
2
x/c
Figure 8.5: Streamlines around
a slender
cylindrical
airfoil of elliptic cross-section situated ina uniform,
high Reynolds number wind.
The parameters for this calculation
are /c =
0.1,
= /12, and =
trailing edge. This follows because the flow is irrotational

everywhere except within the boundary layer and the wake.
as we
Moreover,
have
seen, a change in circulation is

a net vorticity flux away from
a flux cannot be carried by an
necessarily associated with

the
airfoil, and such
irrotational wind. Thus, in the absence of the boundary

layer and the
wake, the air circulation about the airfoil
zero
would be constrained
to remain
was
in accordance
initially
circulation
zero),
theorem.
from Equation
airfoil, irrespective
be
zero
that it
Kelvin
5.13.)
the
This implies,
lift would act
on
of its shape, and irrespective
incident wind speed

would
(See Section
(8.31), that
(assuming
with
or direction.
impossible.
the
of the
In other words, flight
as
Fortunately,
long
as
the
tangential air velocity at the trailing edge of the airfoil is
non-zero,
a
causes the airfoil
process continues until
the wake that trails behind the airfoil carries
net flux of z-directed vorticity, which

circulation to evolve in time. This
the circulation becomes such that the tangential

the airfoils
trailing
trailing edge is
edge
circulation
is
remains
zero:
stagnation
constant
velocity at
i.e., such that the
point.
(assuming
Thereafter,
that
the
the
wind
speed and direction remain constant). Figures 8.4 and 8.5

show
the streamlines
of the flow
around
slender
cylindrical
airfoil of elliptic cross-section
situated in
uniform, high Reynolds number wind whose direction of
incidence is slightly inclined to the airfoils major axis. In
zero.
as to make the
the first figure, the air circulation about the airfoil is

In the second figure, the circulation is such
trailing edge of the airfoil
According
air
circulation
steady-state
a stagnation
point.
to Equations (8.31) and (8.44), when the

about
value,
=V =V
the
airfoil
the lift acting
has
attained
its
on the airfoil becomes
(c + ) sin.
(8.45)
Note that the lift is positive (i.e., upward) when

when the wind is incident
on the
negative (i.e., downward) when

is incident
on the
> 0 (i.e.,
airfoils bottom surface),
< 0 (i.e., when
airfoils top surface), and
zero
the wind
when
0 (i.e., when the wind is incident parallel to airfoils major

axis). Incidentally,
the
the angle is conventionally
angle of attack. Finally, from Equations
(8.43), the focus of the airfoil is located
1
sin
(c )
4
1
(c )
(8.51) Incompressible Aerodynamics

157
termed
(8.35) and
a distance
180
0
0
10
20
Figure 8.6: The angular locations of the boundary

separation points,
of
calculated
as a function
layer
of the angle
attack, , for a cylindrical airfoil of elliptic cross-section
situated in
uniform, high Reynolds number wind. The
solid,
dashed,
and
short-dashdotted,
curves
long-dashdotted
airfoils
of ellipticity
0.125,
/c
correspond
1.0,
0.5,
to
0.25,
and
respectively.
The trailing edge of the airfoil is located at
=180 .
in front of the centroid of its cross-section.

that the airfoil
distance
very
becomes
asymptotes
to c/4. Thus,
we
c), this
that the
focus
c), this
ofdista
a thin
nce airfoil, which is defined
action of the lift, is located
one quarter
In the limit
conclude
thin (i.e.,
as the point of
way along the
of the
airfoil from its leading edge.
The above analysis is premised

that there is
layer
no
appreciable
on
the assumption
separation of the boundary
from the back of the airfoil, which implies
neglect of form drag. We

reasonable
boundary
can check
the
that this assumption is
by calculating the approximate locations of the

layer separation
Section 7.10. Let
points using the analysis
s represent
of
arc-length along the surface
of the airfoil, measured from the front stagnation point.
Assuming
that,
hypothesis,
the circulation
in
accordance
stagnation point is located at
Equation (8.49).] It follows that
ds
=(dx
+ dy
2) 1/2
with
is such that
12
h()d,
=0
where 0
Zhukovskiis
=
= 2.
this
[See
(8.47)
where
h()
=(c
sin
cos 2 ) 1/2
(8.48)
Moreover, from (8.39) and (8.44), the tangential air speed
just above the surface of the airfoil
U()
=V (c +)
f()
h()
can be written
(8.49)
with
f()
=sin( +) +sin.
(8.50)
Inaddition, it can be shown that
ln
d
158
FLUID MECHANICS
=g(),
vz (y = 0+ )
z
dz
vz (y = 0
Figure 8.7: Side view of a vortex sheet.
where
g()
+ )
cos(
(c
2)
cos sin
h2()
f()
(8.52)
According to the analysis of Section 7.10, the separation

points
are located at =
()
where ( )
= 0.1567,
Z
4()g()
0.47
f
5()
0
and
5(
4(
)
d
(8.53)
Here,
> +
> 0 and 0 >
> .
y-coordinates of the separation points

and
= (/2)
sin
Moreover, the
are x
respectively.
x- and
= (c/2) cos
Figure 8.6 shows
the angular locations of the separation points, calculated
as a function
of the angle of attack, for cylindrical airfoils
of various
been
It
different
re-expressed
can
be
seen
ellipticity,
as an
that
1) the angular distance

large, indicating the
angle
for
roughly proportional
in the
range
presence
a wide
(e.g., /c
bluff airfoil
of
has
0 to 2.)
between the separation
associated form drag (since
other hand, for
/c. (Note that
points is
wake, and
a high
the magnitude of form drag is
to the width of the wake). On the
slender airfoil
(e.g., /c
= 0.125)
the
angular distance between the separation points is much

smaller, indicating the
presence
of
a narrow
wake, and
low associated form drag. Note, however, that, in the latter
case, as the angle of attack

there is an initial gradual
is gradually
increased from
between the separation points, followed
very
zero,
increase in the angular distance

by
an abrupt,
and
a
an
large, increase. We would expect there to be
similar gradual increase in the form drag, followed by

abrupt, and
very
large, increase. The value of the angle of
attack at which this abrupt increase
occurs
critical angle of attack. We conclude

analysis,
which neglects
is termed the
that the previous
form drag, is valid only
for
slender airfoils whose angles of attack do not exceed the

critical value (which is generally only
8.5
a few degrees).
Vortex Sheets
A vortex
sheet is defined
as a
planar
array
of parallel
vortex filaments. Consider
x-z plane,
uniform vortex sheet, lying
in which the vortex filaments
run parallel
to the x-axis. See Figure 8.7. The vorticity
within the
in the
sheet
can
be written
(y)
ex
(8.54)
= x
Here,
length. Let
of
ex is the sheets vortex intensity per unit

vz (y = 0+ )and vz (y = 0 )be the z-component
the fluid velocity
sheet, respectively.
immediately
Consider
the y-z plane that straddles the sheet,
Integration
of
above and below the
small rectangular
v around
as shown
loop in
in the figure.
use
the loop (making
of
the curl theorem) yields

vz
vz (y
= 0+ )
vz (y
0 )
(8.55)
In other words,
tangential flow
a vortex sheet induces a discontinuity in the

across the sheet. The above expression can
easily be generalized to give
v,
(8.56)
where is the sheets vortex intensity
per
unit length,
is a unit vector normal to the sheet, and v is the jump in

tangential
velocity
across
the
direction of n). Furthermore,

that the
sheet
(traveling
it is reasonable
to
in the
assume
above relation holds locally for non-planar
non-uniformvortex sheets.
and
159
8.6
Induced Flow
A vortex filament is necessarily associated with fluid flow

circulating about the filament. Let
us
determine the rela-
tionship between the filament vorticity and the flow field

that it induces. This problem is mathematically
to determining
the magnetic
filament. Inthe latter

0 j
(8.57)
field generated
case, the Maxwell
(8.58)
a current
equation
can be inverted to give the well-known

B(r)
identical
by
1
4
j(r
|r
(r
| 3
Here, j is the current
r )
3r
Biot-Savart law
density,
and B the magnetic
field-strength. By analogy, given that vorticity is related
to fluid velocity via the familiar relation
v,
(8.59)
we can write
v(r)
This expression allows
(r
(r
|r
us to determine
r )
d
| 3
3r .
(8.60)
the flow field
induced by a given vorticity distribution. Inparticular,

for a vortex filament of intensity
reduces to
Z
v(r)
(r
(r
|r
(8.61)
r )
dl
| 3
the above expression
where dlis an element of length along the filament.

Likewise, for a vortex sheet of intensity
we
obtain
v(r)
where dS is an element of
8.7
Let
per unit
(r
(r
|r
length
r | 3
r )
dS
(8.62)
area of the sheet.
Three-Dimensional Airfoils
us now
take into
three-dimensional
Figure 8.8, which shows

of finite
account
airfoils
size, situated
are
the fact
of finite
that
realistic
size. Consider
a top view of a stationary airfoil

a (predominately) horizontal
in
= V ek
In the following, we shall

an airfoil as a wing (although it
actually represents a pair of wings on a standard fixed
wing aircraft). Let us adopt the coordinate system shown
in the figure, which is such that the x-z plane is
wind of velocity
sometimes refer to such
horizontal, the wind is incident predominately

x-direction, and the y-axis points
wing is assumed to lie in the
wingspan,
thickness,
and
(parallel to the
c(z)
plane. Let b be the
of
the
be the
wing
x-y plane). See Figure
the wing is symmetric
so that
x-z
let c(z) and (z)
respectively,
from the
vertically upward. The
width
and
cross-section
8.9. Suppose
about the median plane,
that
= 0,
= c(z) and (z) = (z). It follows that c(z

= 0: i.e., the wing extends from z =
that=
c(z)
=b/2)
>sob/2)
(z >
b/2 to z
= b/2.
up around the wing

a manner as to produce an
upward lift. It follows that the average pressure on the
lower surface of the wing must exceed that on its upper
surface. Consider Figure 8.9, which shows a back view of
the airfoil shown in Figure 8.8. As we go from the
median plane (z = 0) to a wing tip, Y, whether along the
upper or the lower surface of the wing, we must arrive at
the same pressure at Y. It follows that there is a drop in
pressure as we move outward, away from the median
plane, along the wings bottom surface, and a further drop
in pressure as we move inward, toward the median plane,
Suppose that air circulation is set
parallel to the
x-y plane
in such
along the upper surface. Since air is pushed inthe direction

of decreasing
on the
pressure,
wings leading
surface,
deviates
it follows that the air that impinges
edge, and then
sideways
toward
passes over
the
its
median
upper
plane.
passes over the wings lower surface

away from the median plane. See Figure
Likewise, the air that

deviates sideways
8.8.
Now, the air that leaves the trailing edge of the wing
at
some
point Q must have impinged
at the different points P and P

travelled
over
the
wings
on the leading edge

on whether it
or lower surfaces,
depending
upper
160
FLUID MECHANICS
x
airfoil
c(z)
z=b/2
z=0
z = b/2
Figure 8.8: Top view of a three-dimensional airfoil of finite

size.
(z)
Figure 8.9: Back view of a three-dimensional airfoil of finite size,
indication the pressure variation
over its surface.
161
x
Figure 8.10: Top view of the airflow
over
the top (left)
and bottom (right) surfaces of a three-dimensional airfoil

of
finite size.
Moreover,
surface
air that travels
acquires
small
to Q via the wings

sideways
velocity
upper
directed
towards the median plane, whereas that which travels to Q
via the lower surface acquires
away
directed
small sideways velocity
from the median plane. On the other hand,
the air speed at Q must be the

whether the air arrives
because
surface,
of
lower
pressure (which, according to

on the air speed) must be
Thus, we conclude that there is a
theorem, depends
continuous
at Q.
in the direction
trailing edge of
irrespective
upper or
the
Bernoullis
discontinuity
same,
from the wings
a wing. This
of the air emitted
by the
implies that the interface, ,
over the upper

a vortex sheet. (See
between the two streams of air that travel

and lower surfaces of the wing is
Section 8.5.) Of
course,
this vortex sheet constitutes
wake that trails behind the airfoil. Moreover,
expect
generally
incident
= I(z)ek
the
in
the airfoil is such that the tangential

outside
by
per unit length
= vz and vz is tangential velocity

across the wake. [See Equation (8.56).]
we saw previously, the boundary layer that covers
I(z)
discontinuity
As
the
would
can be written
(8.63)
where
to be convected
wind. It follows that the vorticity
the wake
the wake
we
the layer is sharply reduced to
velocity
zero at
U just
the airfoil
surface. Actually, the nature of the substance enclosed by
the surface
is irrelevant
to
our argument,
and nothing is
changed
in
our
analysis
if we suppose that this region

we can replace the airfoil by air
layer by a vortex sheet, S with a
contains air at rest. Thus,
at rest, and the boundary

vortex intensity
per
unit length S
the velocity discontinuity
that is determined
by
U between the air just outside
the boundary layer and that at rest in the region where the
airfoil
was
yields
=n
(8.64)
where
previously
located. In fact, Equation
U,
nis an outward
We conclude
unit normal to the airfoil surface.
that
stationary
in a
a vortex
airfoil situated
uniform wind of constant velocity is equivalent to
sheet
(8.56)
located at the airfoil surface, and
wake that
trails behind the airfoil, the airfoil itself being replaced by

air at rest. The vorticity within S is largely parallel to the
z-axis [since
planesee
nand U are both essentially
to the incident wind direction.

vortex
filaments
See Figure 8.11.
are generally
cannot move off the
within S
filaments (since they

Conversely,
parallel to the
x-y
Equation (8.64)], whereas that in is parallel
the vortex
termed free filaments.
outside S can be written
filaments
termed
airfoil surface).
within
The air velocity
The
bound
are
generally
both inside and
v =V + v + vS
(8.65)
where V is the external wind velocity, v the velocity field

induced by the free vortex filaments that constitute , and
vS the velocity field induced by the bound filaments that
constitute S.
some
Consider
point Pthat lies on S.Let P+ and P be
two neighboring points that
are equidistant
from P,where
P+ lies just outside S,and P lies just inside S,and the

line P P+ is normal to S.We can write
v(P+ )
V + v (P+ )+ vS (P+ ),
(8.66) v(P )
(8.74)
V + v (P )+ vS (P ).
162
FLUID MECHANICS
y
x
z
Figure 8.11: Vortex structure around
a wing.
= U(P), where
However, v(P+ )
U(P) is the tangential
air velocity
on
and
the airfoil surface,
= 0, since the air within S is stationary.

v (P+ )= v (P ) = v (P), since we expect
v(P )
Moreover,
v
just above point P
across
to be continuous
On the other hand,
we
expect the tangential component of vS to be discontinuous
across
S.Let
us define
vS (P)
This quantity
point
can
be identified
P by the bound
contribution
as
vortices
12
[vS (P+ )+ vS (P )].
(8.68)
the velocity induced at
on
excluding
the
from the local bound vortex at P (since this
vortex induces equal and opposite velocities at P+ and P ).

Finally, taking half the
sum of Equations
(8.66) and (8.67),
we obtain
1
U(P)
2
=V + v (P) +vS (P).

(8.69)
8.8
Aerodynamic Forces
The net aerodynamic
force acting
airfoil of finite size can be written
on an three-dimensional

ZS
pndS,
(8.70)
where the integral is taken
over
the surface of the airfoil,
unit normal vector on S,dS is

p is the air pressure. From
Bernoullis
theorem (in an irrotational fluid), we can
2,
p0= (1/2)v
where p0 is a constant pressure.
write p
Since a constant pressure exerts no net force on a closed
surface, we get
S. Here,
an
n is an outward
element
of S, and
ZS
ndS,
(8.71)
where Uis the tangential air velocity just above the surface
of the airfoil. Now,

U
(n
U)
=U
(n
U) U
=U
n,
(8.72)
since
= 0 on the surface. Hence,

1
A=
ZS
(n
U)dS.
(8.73)
Making
use
of Equations
(8.64) and (8.69), the above
expression
can be written
ZS
(V + v
+ vS )
S dS
=L+ D+F,
163
where
ZS
dS,
ZS
(8.75)
(8.76)
dS,
ZS
(8.77)
Here, V, v
and vS
are
vS
dS.
the incident wind velocity, the
velocity induced by the free vortices in the wake, and

the
velocity induced by the bound vortices covering the
surface of the airfoil, respectively. The forces L and D are
called the lift and the induced drag, respectively. (Note, that
L now represents
length.)
a net
force, rather than
force
per
unit
We shall presently demonstrate that the force F is
negligible.
Let
us assume
that
S
ez
(8.78)
i.e., that the bound vortices covering the surface of the
airfoil
run
parallel
correct for
exactly
constant
to the z-axis. This assumption
an
airfoil of infinite
cross-section.
approximation
for
an
Moreover,
it
wingspan
is
is
and
good
airfoil of finite wingspan, provided
the airfoils length greatly exceeds its width (i.e., b c).

Now,
the
incident
velocity
gre atlisywritten
where
wind
dl is
V ek
Moreover,
dS
dldz,
runs parallel to the x-y plane

on the airfoil surface. Now, making use of the
curl theorem, we can easily show that
an
element of length that
whilst lying
IC
z
dl
=(z),
(8.79)
where the closed
curve C is the intersection of the airfoil

z = z,and (z) is the air circulation
surface with the plane
about the airfoil inthis plane. Thus, it follows from

Equation (8.75) that
b/2
=V
(z)dz e
b/2
(8.80)
This expression is the generalization of Equation
(8.31) for a three-dimensionalairfoil of finite size.
As before, the lift is at right-angles to the incident wind
direction.
Let
us make
the further assumptionknown
as the
lifting line approximation (because the lifting action of
the wing is effectively concentrated onto
a line)that
= w(z)e
(8.81)
throughoutS, where w(z)e
is the induced velocity due
to the free vortices in , evaluated at the trailing edge

of the airfoil. Here, the velocity w(z) is called the
downwash velocity. It follows from Equation (8.76) that

Z
b/2
w(z)(z)dz ek
b/2
(8.82)
Note that the induced drag is parallel to the incident
wind direction. The origin of induced drag is as follows.
energy to constantly resupply free vortices to the

as they are swept downstream by the wind (note
that a vortex filament possesses energy by virtue of the
kinetic energy of its induced flow pattern), and this energy
It takes
wake,
is supplied
by the work done in opposing the induced
drag. The drag acting

airfoil with
an
drag) situated
on a well-designed
aerodynamic
in
high Reynolds
implies that the friction
airfoil (i.e.,
shape that minimizes
an
form
number wind (which
drag is negligible)
is generally
dominated by induced drag.
According to Equations (8.62) and (8.77), the force F
is written
ZS
ZS
[S (r
(r
|r
164
FLUID MECHANICS
r )]
r | 3
S (r)
dS dS
OzP
r r
OzP
Figure 8.12: Semi-infinite vortex filament.
We
can
interchange
primed
and
unprimed
variables
without changing the value of the integral. Hence,

ZS
ZS
[S (r)
|r
(r
r)]
r|
S (r
dS
dS.
(8.84)
Taking the half the

obtain
sum
of the previous two equations,

ZS
[S (r
(r
)]
b)
+ [(r r )
|r
(8.85) However, (a
ZS
S (r)
we
S (r)]
S (r
dS dS
| 3
c + (b
c)
a + (c
a)
= 0.
Thus, the above expression yields

ZS
ZS
[S (r
(r
S (r)]
|r
r | 3
But, the assumption (8.78) implies that S (r
0. Hence, F is negligible,
Consider
a closed
as was
bound vortices at
dS dS
(8.86)
S (r)
previously stated.
surface covering the small section
of the airfoil lying between the parallel planes
z = z + dz. The
r )
z = z and
flux of vorticity into the surface due to
z is (z).
The flux of vorticity out of
the surface due to bound vortices at
z + dz is (z + dz).
Finally, the flux of vorticity out of the surface due to the

free vortices inthe part of the wake lying between
dz is I(z) dz. However, the net flux of vorticity

closed
surface is
zero,
since vorticity
z and z +
out of a
is divergence
free.
Hence,
(z)
(z
dz)
I(z) dz,
(8.87)
which implies that

I(z)
d
dz
(8.88)
Finally,
filament
consider
of vortex intensity
straight
semi-infinite
= ex
vortex
that terminates at
the origin, O, as shown in Figure 8.12. Let
us calculate
the
= (0,
=z tan, dl=z sec d, |r r |
= zey Hence, from Equation
induced velocity at Pis v = vy ey where
. ,
flow velocity induced by this filament at the point P

2
0,z).From the diagram l

z sec , and |r r |
(8.61), the
This result allows
vy
/2
cosd
4z
4z
(8.89)
us to
calculate the downwash
velocity,
w(z)s result
vy (z),
induced
at calcu
the trailing
edge
of the
airfoil
by
Thi
allow
sus to
lat e th
edown
wash
veloc
the semi-infinite
in the wake. The vortex
free vortices
intensity in the small section of the wake lying between

and
z + dz is I(z)dz, so we obtain
Z
w(z)
b/2
I(z
)dz
z z
b/2
1
4
d(z
z z
(8.90)
where
use has been made
Incompressible
of (8.88).
Aerodynamics
165
8.9
Ellipsoidal Airfoils
Consider
an
ellipsoidal
airfoil whose
outer
surface
is
specified by the parametric equations
c0
sin cos,
(8.91)
sin sin,
2
b
(8.92)
cos,
2
(8.93)
where 0
wingspan,
c0
maximum
wing
cross-section
and 0
the maximum
thickness.
is elliptical
2. Here, b is the
wing width, and 0

Note
that
the
the
wings
x-y and the x-z

:
i.e., the wingspan
both in the
planes. It is assumed that b > c0
is greater than the wing width, which in turn is much
greater
than
the
wing
thickness.
At
fixed
fixed z), the width and thickness of the airfoil

sin and ()
= 0
= c0
sin, respectively.
Assuming that the two-dimensional
at fixed
(i.e.,
are c()
z,we deduce
result (8.44) holds
that the air circulation about the wing
satisfies
(z)
V c(z) sin
sin,
(8.94)
where
0
V c0
(8.95)
Here, the angle of attack, , is assumed to be small. From
Equations (8.90) and (8.94), the downwash velocity in the
region |z| < b/2 is given by
cos
w()
2b
cos
0
Z
cos
2b
cos
d
cos
cos
(8.96)
Now, the integrand inthe integral
cos
is singular when
(8.97)
= . However,
cos
we can
still obtain
finite value for the integral by taking its principal part: i.e.,
Z
lim
0
cos
cos
+
cos
cos
(8.98)
Physically, this is equivalent to omitting the contribution of
the local free vortex at a given point
edge to the downwash
velocity
which is reasonable because
at its center. Hence,
on the
induced
a vortex
we obtain
airfoils trailing
at that point,
induces
zero
velocity
d
cos
"
sin(1/2)(
"
lim
cos
sin(1/2)(+
ln
sin(1/2)(
+ )
sin(1/2)(
sin
#!
+
sin
0
#!+
ln
"sin(
#!
/2)
lim
ln
sin
sin(
= 0,
+ /2)
which implies that

w()
In the region |z| > b/2,
we can write
2b
= 2z/b, so that
Z
w()
2b
cos
1
2b
(8.99)
(8.100)

||
p
2
166
FLUID MECHANICS
1.2
0.8
0.4
2
/
0
0.4
0.8
(8.101)
1.2
1.6
2
2
z/b)
Figure 8.13: Downwash velocity profile induced at the

trailing edge by an ellipsoidal airfoil.
Hence,
we conclude
that the downwash velocity profile
induced by an ellipsoidal airfoil takes the form

(
w(z)
2b
1
|z|/(z
|z| < b/2

|z| > b/2
2/4) 1/2
(8.102)
This profile is shown in Figure 8.13. It

the downwash
region
velocity
can
be
seen
is uniform and positive
between the wingtips
(i.e., b/a
<z<
that
in the
b/2), but
negative and decaying in the region outside the wingtips.

Hence,
we
conclude
that
as
air
passes over an
airfoil
subject
to
an
upward
velocity component,
the reaction
lift it acquires
which, of
course,
a net downward
consequence of
is a
to the lift. On the other hand, the air
immediately behind and to the sides of the airfoil acquires
net upward velocity component.

acting
on the
behind, and
airfoil is associated
an upwash
airfoil. The existence
the side of
V-formation
In other words, the lift
flying
a downwash
with
of air
behind and to either side of, the

slightly behind and to
of upwash
object
us to
allows
adopted by wild geesea
bird
explain
the
flying in the
upwash of another bird needs to generate less lift in order to
stay in the air, and, consequently, experiences
less induced
drag.
It follows from Equation (8.93) and (8.94) that
Z
b/2
(z)dz
b/2
sin
4
0
b.
(8.103)
Hence, Equation
(8.80), (8.82), and (8.100)
yield
the
following expression for the lift and induced drag acting
onan
ellipsoidal airfoil,
(8.104)
V b0
2
0
,
.
(8.105)
Now, the surface
area of the airfoil inthe x-z plane
is
4
bc0
167
airfoil
T
Figure 8.14: Side view of a fixed wing aircraft in flight.
Moreover, the airfoils aspect-ratio is conventionally

defined
as the length to width ratio for a rectangle
of
length b that has the same
area as the airfoil: i.e.,

A
4 b
c0
(8.107)
It thus follows from Equation (8.95) that
(8.108)
L0
2
L0
(8.109)
where
L0
S.
(8.110)
8.10
Simple Flight Problems
Figure 8.14 shows
a side-view
schematic of
a fixed
wing
aircraft flying in a straight-line at constant speed through
x is a horizontal coordinate, and y

mass of the aircraft
is assumed to be moving with some fixed velocity V that
subtends an angle with the horizontal. Thus, the wind
stationary air. Here,
a vertical
coordinate. The center of
velocity
in the
aircrafts
wings, which
aircrafts
are
rest frame
is V.
Let
the
assumed to be parallel to its
fuselage, be inclined at an angle + to the horizontal. It

that is the angle of attack. The aircraft
follows
subject to four
forces: the thrust, T, developed
is
by its
engine, which is assumed to act parallel to its fuselage;

the lift L, which acts at right-angles
to V; the induced
drag, D, which acts in the opposite direction to V; and

the weight, W,which acts vertically downward.
Vertical force balance yields
T sin(
+ L cos
=W
D sin ,
(8.111)
whereas horizontal force balance gives

T cos( + )
168
=Dcos +Lsin.
FLUID MECHANICS
us assume
Let
that the angles and
are
both small.
According
Equations
and
L0
Le t usto
assum
e th at(8.108)
the an gl
es (8.109),
and aL
reboth
s
2.
2).
and D (2 L0 /A)
Thus, L O() and D O(
Moreover,
L0
it is clear from (8.111) and (8.112) that T
O(
)and W O(). Thus, to lowest order in , Equation
(8.111) yields
L0
(8.113)
whereas Equation (8.112) gives
L0
(8.114)
Expression (8.113) relates the angle of attack to the ratio
weight to its (theoretical)
of the aircrafts
(at
given airspeed).
maximum lift
(8. 114) relates
Expression
the
aircrafts
angle of controlled (i.e., at constant airspeed) ascent to the

thrust
developed
aircraft, such
glider, has
(8. 114) reveals
aircraft,
decentwhich
the value
g=
by its engine.
as a
2 W
A L0
is usually
that
an
Now
zero
unpowered
thrust. For such
the angle
termed the glide angletakes
(8.115)
At fixed airspeed, V, and wing surface
implies that L0 is fixed) this angle
can
making the wing aspect-ratio, A, as large

result
explains
albatrosses)
ones.
an
of controlled
For
accounts
area,
(which
be minimized
as possible.
for the fact that gliders
by
This
(and
have long thin wings, rather than short stubby
powered aircraft, the critical thrust to weight
ratio required to maintain level flight (i.e.,
=0) is
T
g.
W
(8.116)
Hence, this ratio is minimized
angle,
which
generally
explains
why
by minimizing
long-haul
the glide
which
aircraft,
need to minimize fuel consumption, tend to have
long thin wings. Finally,

angle of
attack exceeds
value c
then boundary
back sides of the wings,

level of drag acting
as we saw in Section 8.4, if the

some (generally small) critical
layer separation occurs on the
giving rise to a greatly increased
on the aircraft. In aerodynamics, this

a stall. As is clear from Equations (8.
phenomenon is called
110) and (8.113), the requirement < c is equivalent to

!1/2
V > Vs
W
S c
(8.117)
Inother words,
a stall can be avoided
airspeed above the critical value Vs
by keeping the
which is known
stall speed. Note that the stall speed decreases with

decreasing altitude,
8.11
as the air becomes
Exercises
8.1. Consider the integral
denser.
as the
In ()
=
0
where
nis a non-negative
cos(n
cos
)d
cos
integer. This integral is defined by
its principal value

"Z
In ()
= lim
0
cos(n
cos
)d
cos(n
+
cos
cos
)d
cos
As was demonstrated inSection 8.9,

I0
= 0.
I1
= ,
Show that
and
In+1
+ In1
= 2 cosIn
and hence that
169
8.2. Suppose that
an airfoil of negligible thickness, and

z variation is expressed
wingspan b,has a width whose

parametrically
as
c()
c sin(),
=1,3,5,
for 0 , where
z =
cos.
2
Show that the air circulation about the airfoil takes the
form
()
sin(),
=1,3,5,
where
= V c
Here, is the angle of attack (which
is assumed to be small). Demonstrate that the downwash velocity
at the trailing edge of the airfoil is

X
w()
=1,3,5,
sin()
2b
sin
Hence, show that the lift and induced drag acting

airfoil take the values
V b1
on the
=1,3,5,
respectively. Demonstrate that the drag to lift ratio
can be
written

DL
1
+
=3,5,7,
2
c1
where A is the aspect ratio. Hence, deduce that the airfoil

shape (inthe x-y) plane that minimizes this ratio (at fixed aspect
ratio) is an ellipse (i.e., such that c
8.3. Consider
a plane
= 0 for > 1).
that flies with a constant angle of attack,
and whose thrust is adjusted such that it cancels the induced drag.
The plane is effectively subject to two forces. First, its
weight, W
kvvx
ey
= W ey
and second its lift L
= kvvy
ex +
Here, x and
respectively,
y are horizontal and vertical coordinates,

v is the planes instantaneous velocity, and k is a
positive
constant. Note that the lift isdirected at right angles to

the planes instantaneous direction of motion, and has
magnitude
proportional to the square of its airspeed. Demonstrate
that the planes equations of motion
dvx
vvy
dt
dvy
can be written
vvx
dt
g,
= kg/W is a positive constant
where h
with the
dimensions of length. Show that
12
+ gy
= E,
where E is a constant. Suppose that vx =

pgh(1+u)
and vy =
pghw,
where |u|, |w| 1. Demonstrate that, to first order in
perturbed quantities,
r
du
gh
dt
w,
r
dw
gh
u.
dt
Hence, deduce that if the plane is flying horizontally at
some
2g/v0
speed v0
altitude
2g/v0
and issubject to a small perturbation, then its
oscillates sinusoidally at the angular frequency
=
170
FLUID MECHANICS
Incompressible
Viscous
Flow
171
9 Incompressible Viscous Flow

9.1
Introduction
This chapter investigates

viscosity plays
incompressible
significant
flow in which
role throughout
the bulk of
the fluid. Such flow generally takes place at relatively low

Reynolds
governing
number.
From
incompressible
Section
2. 14, the
viscous
fluid motion
equations
can
be
written
Dv
(9.1)
Dt
v
P
0,
+ 2 v,
(9.2)
where the quantity
=p
(9.3)
which is a combination of the actual fluid
pressure, p, and
energy per unit

pressure. Here,
volume, , is
the gravitational potential
as
known
density,
potential.
9.2
the effective
the fluid viscosity,
is the fluid
and the gravitational
Flow Between Parallel Plates
Consider
steady, two-dimensional,
two parallel
distance
plates
d apart. Let
measuring
distance
transverse
are
that
along
be
viscous flow between
situated
perpendicular
longitudinal
coordinate
the plates, and let
be
coordinate
such that the plates
are
located
at
=0
and
= d.
See Figure 9.1.

Suppose
that there
a uniform
so that
is
gradient inthe x-direction,
effective
dP
dx
pressure
G,
(9.4)
where G is a constant. Here, the quantity G could represent
agradient
inactual fluid pressure, agradient ingravitational

energy (due to an inclination of the plates to the
horizontal), or some combination of the twoit actually
makes no difference to the final result. Suppose that the
potential
fluid velocity profile between the plates takes the form
vx (y) ex
(9.5)
From Section 2.18, this profile automatically
satisfies the
= 0, and
is also such
incompressibility
that Dv/Dt
constraint
=0.Hence, Equation (9.2) reduces to
(9.6)
y =d
vx (y)
P
y =0
uid forms a uniform layer of depth hcovering this surfa 172

FLUID MECHANICS
Figure 9.2: Viscous flow down
or. taking
an inclined
plane.
the x-component,
2vx
d
dy
(9.7)
If the two plates
are stationary
satisfies the no slip constraint,
then the solution that
vx (0) =vx (d) =0,at each
plate is
vx (y)
y(d y).
(9.8)
Thus, steady,
two
two-dimensional,
stationary
parallel
parabolic
velocity
midplane,
y = d/2. The net
the z-direction)
viscous
plates
is
flow
that is symmetric
profile
between
associated
with
about
the
volume flux (per unit width in
of fluid between the plates is

Z
vx dy
0
Gd
12
(9.9)
Note that this flux is directly proportional to the effective
pressure
gradient,
inversely
viscosity, and increases
proportional
as the cube
to the fluid
of the distance between
the plates.
Suppose that the
upper
plate is stationary, but that the
lower plate is moving in the x-direction

speed U. In this
case,
at the constant
no slip boundary condition at

vx (0) = U, and the modified
the
the lower plate becomes
solution to Equation (9.7) is
vx (y)
y(d y) + U
d
d
(9.10)
Hence, the modified velocity profile is
a combination
of
parabolic and linear profiles. This type of flow is known
as
Couette flow. The net volume flux (per unit width) of
fluid between the plates becomes
Gd
12
Ud.
12
(9.11)
9.3
Flow Down
Consider
an Inclined
steady, two-dimensional,
plane that is inclined at
x measure
distance
transverse coordinate
corresponds
to
an angle
along
Plane
viscous flow down
to the horizontal. Let
the plane, and let
be
such that the surface of the plane
= 0. Suppose
that the fluid forms
uniform layer of depth h covering this surface. See Figure
9.2.
Incompressible Viscous Flow
173
The generalized
written
pressure
gradient within the fluid is
dP
dx
= G = g sin ,
(9.12)
where
there
g is the acceleration
is no gradient in
x-direction,
due to gravity. In this

the
actual
pressure
case,
in the
and the flow down the plane is driven entirely
by gravity. As before,
we can write
vx (y) ex
,
,
(9.13)
and Equation (9.2) again reduces to
2vx
d
dy
(9.14)
where
G
= g sin .
(9.15)
Application of the
plane, y
= 0.
slip condition at the surface of the
= 0, yields the standard
However, the appropriate

???y=h
fluid/air interface,
there
no
be
zero
y = h,is
that
???y=h
(i.e., xy
boundary condition
physical constraint
vx (0)
at the
the normal viscous stress
= 0), since
there is nothing
???y=h
above the interface
that
can
exchange
momentum
with
the fluid (assuming that the finite inertia andviscosity of air
are both negligible.)
Hence, from Section 2.18, we get the
boundary
condition
?????y=h
dvx
dy
0.
(9.16)
The solution to Equation (9.14) that satisfies the boundary
conditions is
vx (y)
y(2h y).
2
(9.17)
Thus, the profile is again parabolic. In fact, it is the
as
same
the lower half of the profile obtained when fluid flows
between
two
(stationary)
parallel
plates
situated
perpendicular distance 2h apart.

The net volume flux (per unit width in the z-direction)
of fluid down the plane is
Z
vx dy
0
Gh
g sinh
3
(9.18)
where
use
has been made of Equation (9.15). Here,
/ is the kinematic
viscosity
of the fluid. Thus, given
the rate Q that fluid is poured down the plane, the depth of
the layer covering the plane becomes
!1/3
3Q
g sin
(9.19)
Suppose that the rate at which fluid is poured down the
plane is suddenly increased slightly from Q to Q+Q. We

would expect
an associated
change in depth of the layer
covering the plane from h to h+ h, where
Let the interface
dh
Q.
dQ
(9.20)
between the layers of different
depth
propagate in the x-direction at the constant velocity V.

In a frame of reference that
the volume
co-moves
right and to the left of the interface

Q V (h
with this interface,
fluxes (per unit width) immediately
+ h),
one another.
or
dQ
dh
Gh
to the
Q V h and Q
respectively. However, in
these fluxes must equal
are
steady state,
Hence,
Q
V h,
(9.21)
9Q
!1/3
g sin
(9.22)
As
can
easily
be verified,
this
velocity
is twice
the
maximum fluid velocity inthe layer.
174
FLUID MECHANICS
9.4
Poiseuille Flow
Steady viscous fluid flow driven by
gradient
established
between
an effective pressure
of a long
the two ends
straight
pipe of uniform circular cross-section is generally
known
as
Poiseuille flow,
experimentally
by
that
the
Suppose
Let
us
since
was
it
J. L. M. Poiseuille
pipe
is
of radius
studied
first
in
1838.
a.
adopt cylindrical coordinates whose symmetry
coincides with that of the pipe. Thus,

along the pipe, r
= 0 corresponds
z measures
axis
distance
to the center of the pipe,
and
r= a corresponds to the pipe wall. Suppose
= Gez
that
(9.23)
is the uniform effective

and
v =vz (r)ez
(9.24)
pressure
gradient along the pipe,
the time independentvelocityprofiledrivenbythis
It followsfromSection2.19that
gradient.
= 0 and Dv/Dt = 0.
Hence, (9.2) reduces to
2v
ez
(9.25)
Taking the z-component of this equation,
!
1r
dr
dvz
dr
we obtain
(9.26)
where
use has been made
of (2.155).
The most general solution of the above equation is
vz (r)
r 2 + A lnr + B,
(9.27)
where
A and B
constraints
are
are
arbitrary
constants.
center of the pipe (which implies that A

at the edge of the pipe [i.e., vz (a)
the
no
The physical
that the flow velocity is non-singular
= 0],in accordance
slip condition. Thus, we obtain
at the
= 0),and is zero
with
vz (r)
(a
2).
(9.28)
The volume flux of fluid down the pipe is

Z
2rvz dr
Ga
(9.29)
According
to the above
should be directly
gradient
along
analysis,
proportional
the
pipe.
the
quantity
to the effective
The
accuracy
Q/a
pressure
with
which
experimental observations show that this is indeed the
case
(at relatively low Reynolds number) is strong evidence in

favor of the assumptions
that there is
no
slip at the pipe
walls, and that the flow is non-turbulent. In fact, the result

(9.29), which
is known
experimentally provided
as
Poiseuilles
law, is valid
the Reynolds number of the flow,
= Ua/, remains less than about 6.5 10 3.Here, U

=ReQ/a
= Ua/
is the
remain
means less
ana bout
10 hand, if
flow thspeed.
On6.5
the other
Re
the Reynolds number exceeds the

flow
critical
spee value
d.On
6.5
the
1
oth
0
then
hand,
the
if the
flow
Reynolds
in the
numb
pipe
and
Poiseuilles law breaks down.
9.5
Taylor-Couette Flow
becomes
turbulent,
Consider two thin cylindrical shells with the
same
respectively. Suppose that the annular region r1
filled with fluid of density and viscosity
vertical
axis. Let the inner and outer shells be of radius r1 and r2
r2 is
Let the inner
and outer cylinders rotate at the constant angular velocities
1 and 2
respectively. We wish to determine the steady
flow pattern set
up
within the fluid. Incidentally, this type
of flow is generally known
as Taylor-Couette
flow.
175
It is convenient to adopt cylindrical coordinates,
whose symmetry axis coincides with the
the
two
correspond
shells.
to
Thus,
r = r1
and
the
inner
r = r2
r,, z,
common
axis of
outer
shells
and
respectively.
Suppose
that the flow velocity within the fluid is written
v (r)e
r(r)e
(9.30)
where
(r)
v (r)/r is the angular
Application of the
velocity
no slip condition at the two
profile.
shells leads
to the boundary conditions

(r1 )
(9.31) (r2 )
=
=
1
2
,.
(9.32)
It again follows
Dv/Dt
from Section
= 0. Hence, (9.2) reduces
2.19 that
to
=0
and
(9.33)
Assuming
that P
= 0 within
the fluid, since
any
flow is
driven by the angular rotation of the two shells, rather than

by
pressure
gradients
or gravity,
use
and again making
of
the results quoted in Section 2.19, the above expression

yields
!
1r
dr
dr
r2
=0,
(9.34)
or
r2
dr
dr
0.
(9.35)
The
solution
of
(9.35)
that
satisfies
the
boundary
conditions is

(r)
r2
2
r1
r 22
2
r1
2
r1
r22
r 22
(9.36)
Note that this angular velocity profile is
of the solid body rotation profile
a combination
= constant,
and the
irrotational rotation profile
r 2.
non-zero component
From Section 2. 19, the only
of
the viscous stress tensor within the fluid is
=r d
dr
=r d
dr
(9.37)
Thus, the viscous torque (acting in the -direction)
height (inthe z-direction) exerted
on the inner
per
unit
cylinder is

=2r
2
1
(r1)
= 4
2
r1
r 22
(9.38)
Likewise, the torque
per
unit height exerted
on the outer
cylinder is

2
= 2r
2
2
=4
(r2 )
2
r1
r 22
(9.39)
As expected, these two torques
are
and act to make the two cylinders

angular
rotates
9.6
velocity (in which
case,
equal and opposite,
rotate at the
same
the fluid between them
as a solid body).
Flow inSlowly-Varying Channels
According to Section 9.1, the equations governing steady,
incompressible, viscous fluid flow
(9.40)
(9.53)
(v
are
)v
0,
2v.
P +
176
FLUID MECHANICS
As
we saw in Sections 9.2 and 9.4, for the case of flow

a straight channel of uniform cross-section, v
(v ) v are both identically zero, and the governing
along
and
equations consequently
reduce to the simple relation
(9.42)
Suppose, however, that the cross-section
varies along its length. As

this
variation
is
we
of the channel
shall demonstrate, provided
sufficiently
slow,
the
is
flow
still
approximately described by the above relation.
Consider steady, two-dimensional, viscous flow, that is
predominately
are
in the x-direction, between two plates that
predominately
y-z plane. Let the spacing

vary on some length scale l d.
parallel to the
between the plates, d(x),

Suppose that
G(x)
ex
(9.43)
where
G(x)
also
varies
on
the
same
length
scale.
Assuming
that /x
O(1/l)
and /y
it
O(1/d),
follows from Equation (9.40) that
vy
dl
O
vx
(9.44)
Hence,
2
[(v
)v]x
(9.45)
2v)x
vx
!2
dl
1
+O
,
.
vx
(9.46)
The x-component of Equation (9.41) reduces to

2vx
1
+O
vx
!2
dl
+O
(9.47)
Thus, if
dl
1,
(9.48)
vx
l
(9.49)
i.e.,
if the channel
cross-section
lengththen
varies
narrow,
is sufficiently
slowly
sufficiently
Equation (9.47)
and its
along
its
can be approximated as
vx
(9.50)
This, of
course,
is the
same as
the equation
steady,
two-dimensional,
parallel
plates. See Section 9.2. Assuming that the plates
are
located at
analysis of
viscous
governing
between
flow
y = 0 and y = d(x), and
making
Section 9.2, the appropriate
exactly
use
of the
solution to the
above equation is
vx (x,y)
G(x)
y[d(x) y].
2
(9.51)
The volume
flux (per unit width) of fluid between
the
plates is thus
Z
vx dy
0
G(x)d
3(x)
12
(9.52)
However, for steady incompressible
independent of x,which implies that
flow, this flux must be
G(x)
dP
dx
=12Qd 3(x).
(9.62) Incompressible Viscous Flow

177
Suppose that
a constant
x2
where x2 x1
pressure,
x = x1 and x =
difference in effective
P, is established between the fixed points
=l.Integration
of the above equation
between these two points yields
= 12Qlhd
3i,
(9.54)
where hi
effective
x2
x1
pressure
()dx/l.
Hence, the
volume flux
(per
difference becomes
unit width) of fluid between the plates that is driven by
the
12hd
3i
(9.55)
Moreover, the effective
pressure
gradient at a given point is
G(x)
dP
dx
d 3(x)hd
3i
(9.56)
which allows
point
us to
determine
from Equation
the velocity profile at that
(9.51).
Thus, given
pressure gradient, P/l,

separation, d(x), we can fully specify
and
effective
the
the
average
variable
the flow between the
plates.
Using analogous arguments to those employed above,
but adapting the analysis of Section 9.4, rather than that of
we can easily show that steady viscous flow

a straight pipe of circular cross-section, whose
a varies slowly with distance, z, along the pipe, is
Section 9.1,
down
radius
characterized by
vz (r,z)
G(z) ha 2(z)
4
(9.57)
G(z)
(9.58)
a 4(z)ha
,
,
4i
8ha
4i
(9.59)
Here, Q is the volume flux of fluid down the pipe, P
approximations used to derive the above results
P(z1 ),l z2 z1
and hi
=P(z2 )
are valid provided
z2
z1
()dz/l.
The
al
1,
1.
(9.60)
vz
a2
l
(9.61)
9.7
Lubrication Theory
It is well-known
one another
of
fluid
that two solid bodies
sandwiched
certain circumstances,
within
the
can
slide
over
easily when there is a thin layer
particularly
between
layerthis
them.
large positive
phenomenon
Moreover,
pressure
is
under
develops
exploited
hydraulic bearings, whose aim is to substitute
in
fluid-solid
friction for the much larger friction that acts between solid
are in direct contact
bodies that
up,
the fluid
layer
in
with
hydraulic
one another.
bearings
Once set
offers
great
resistance to being squeezed out, and is often capable of
a useful load.
supporting
Consider the simple two-dimensional
body with
x-z
case
of
solid
plane surface (that is almost parallel to the
plane) gliding steadily
over
another
such body, the
surface of the gliding body being of finite length l

in the
direction
of the motion (the x-direction),
width (in the z-direction).
and of infinite
See Figure 9.3. Experience
shows that the plane surfaces need to be slightly inclined
to
one
another.
Suppose
plane surf
that
ac
es
need
1is
to the
be slightly
angle of
in
transform to a frame of reference in

upper body is stationary. In this frame, the
lower body moves in the x-direction at some fixed speed
U. Suppose that the upper body extends from x = 0 to x
inclination. Let
us
which the
= l,and that the surface of the lower body corresponds to

y = 0. Let d(x) be the thickness (in the y-direction) of the
fluid layer trapped between the bodies, where d(0) = d1 and
d(l) = d2 It follows that
= d1 x,
d(x)
(9.71)
178
FLUIDMECHANICS
y
d1
d2
l
Figure 9.3: The lubrication layer between two planes in
relative motion.
where
d1 d2
l
(9.63)
As discussed in the previous section, provided that
dl
(9.64)
U d
1,
1,
l
(9.65)
the cross-section of the channel between the two bodies is
sufficiently
channel
slowly
can
varying
be treated
as
in the
x-direction
effectively
that the
uniform at each
point along its length. Thus, it follows from Equation
(9.10) that the velocity profile within the channel takes the
form
"
G(x)
vx (x,y)
?d(x)
y
2
d(x)
U
d
(9.66)
where
dp
G(x)
dx
(9.67)
is
the
pressure
gravitational
gradient.
forces
viscous forces.
Here,
with respect
The volume
flux
we
are
to both
per
neglecting
pressure
and
unit width (in the
z-direction) of fluid along the channel is thus

Z
vx dy
G(x)d
3(x)
12
Ud(x).
12
(9.68)
Of
course,
in a steady state, this flux must be independent
of x.Hence,
#
"
dp
dx
= G(x) = 6
U
d 2(x)
2Q
d 3(x)
(9.69)
where d(x)
= d1
x. Integration
of the above equation
yields

p(x) p0
1
U
1
d1
(9.70)
where p0
= p(0). Now, assuming
that the sliding block is
pressure
completely immersed in fluid of uniform ambient
p0
we
would
expect
of the lubricating
the
layer
which implies that p(l)
pressures at
the two ends
to both equal
= p0
equation that
p0
It follows from the above
d1 d2
U
d1
+ d2

(9.80)
179
and
p(x) p0
6U [d1 d(x)][d(x)
d 2(x)(d1
d2 ]
+ d2 )
(9.72)
Note that if d1
p0
> d2
then the
is positive throughout
other
words,
a lubricating
pressure
increment
the layer, and vice

layer sandwiched
that is capable
of supporting
versa.
In
between two
solid bodies in relative motion only generates
pressure,
p(x)
positive
normal load,
when the motion is such
as to drag
pressure
layer. The
increment
has
means
narrower
(by
stresses) fluid from the wider to the
of viscous
end of the
single maximum in
2,
the layer, and its value at this point is of order l

U/d
assuming
that
suggests that
d2 )/d1
(d1
very
is of order
pressures can
large
be set
unity. This
up
inside
thin lubricating layer.

The net normal force (per unit width in the z-direction)
acting
on the lower
Z
fy
plane is
"ln
[p(x) p1 ]dx
(9.73)
6U
!#
d1 d2
d1
d2
d1 + d2
Moreover, the net tangential force (per unit width) acting
on the lower
plane is
!y=0
fx
=
0
(9.74)
"2ln
vx
dx
2U
!#
d1 d2
d1
3
d2
Of course, equal and opposite forces,
fx
fx
Ul
d
2
0
"ln
3
2k 2
1
+k
2k
1
k
d1 + d2
"2ln
(9.75)
fy
fy
1
+k
Ul 1k
3k
1
k
d0
(9.76)
on the upper plane. Here, d0

mean channel width, and k = (d1
act
= (d1
+ d2 )/2 is the
+ d2 ).Note
d2 )/(d1
that if 0 < d2
< d1 then 0 < k < 1. The effective
of
friction,
conventionally
between
fx
fy
fx
f y
sliding
coefficient
bodies
is
as the ratio of the tangential to the

exert onone another. Hence,
defined
normal force that they
two
the
4 d0
H(k),
(9.77)
where
"ln
H(k)
=k
1
+k
#,"ln
3k
1
+k
1
k
2k
1
k
(9.78)
a monotonically decreasing function

of k fun
The
in the
ction
range
H( k)
0<
iska<
m onoton
1. In fact,
ical ly
H(k
decre
asing
0) 3/(4k),
func tio
The function H(k) is
whereas H(k 1) 1. Thus, if k O(1) [i.e., if (d1
d2 )/d1
O(1)]
then
/l) if
1.
other
4k),
whereas
H(k
C1)
O(d0
1. Thus
kIn
O
(1) words,
[i.e., i
f
the
effective
coefficient
of friction
bodies in relative motion that
between
are separated
by
two solid
a thin fluid
layer is independent of the fluid viscosity, and much less
than unity. This result is significant because the coefficient
between two solid bodies in relative motion that
of friction
are
one another is typical of order

presence of a thin lubricating layer does
to a large reduction in the frictional drag
in direct contact with
unity. Hence, the

indeed lead
acting between the bodies.
9.8
Stokes Flow
Steady flow in which the viscous
greatly
fluid
volume
is
exceeds
generally
the
as
known
per unit
inertia
density
Section
force
(see
Reynolds numbers that

the
time
are
independent,
inertia
in the
per
unit
Stokes flow. Since,
definition, the Reynolds number of
ratio of the advective
force density
advective
by
fluid is the typical
volume to the viscous
2.16), Stokes
flow
implies
much less than unity. Now, in
low
Reynolds
number
limit,
Equations (9.1) and (9.2) reduce to
0,
(9.79)
+ 2 v.
(9.93)
180
It follows from these equations that
FLUIDMECHANICS
2v
v)
(9.81)
where
v, and use
has been made of Equation
(A.177). Taking the curl of this expression,
we obtain
0,
(9.82)
which is the governing
use
has
been
made
(A.176), and (A.177).
Flow
Let
r, ,
equation for Stokes flow. Here,
of
Equations
9.9
be standard
(A. 173),
Axisymmetric
spherical
Stokes
coordinates.
Consider
axisymmetric Stokes flow such that
v(r)
= vr (r, )
er +
(r, ) e
(9.83)
According
to Equations
automatically
satisfy
(A.175)
the
and (A.176),
incompressibility
we can
constraint
(9.79) by writing
(9.84)
where (r, ) is
follows that
a stream
function (i.e.,
= 0). It
vr (r,)
(9.85)
Moreover, according to Section C.4, r
(r,)
where
1r (rv )
sin
1
rsin
= 0,and
1r
vr
Hence, given that ||
(r,)
sin
,
.
(9.86)
L()
rsin
(9.87)
sin
(9.88)
= 1/(r sin), we can write

=
v =
L() .
(9.89)
It follows from Equations (A.176) and (A.178) that
[L()].
(9.90)
Hence, by analogy with Equations (9.84) and (9.89), and

making
use of (A.173)
and (A.177),
we obtain
2() .
(9.91)
Equation (9.82) implies that
2()
0,
(9.92)
which is the governing equation for axisymmetric
Stokes
flow. Inaddition, Equations (9.81) and (9.90) yield
9.109)
Viscous
Flow
Axisymmetric
Stokes
Flow
181
9.10
Solid Sphere
Consider
solid sphere of radius
station-ary
fluid
of
density
and
[L()].
Around
that is moving under
at the constant vertical velocity V
gravity
ez
through
viscosity
Here,
is assumed to take the form
gravitational acceleration
gez
Incompressible
g=
Now, provided the typical Reynolds number,

Re
2V
(9.94)
is much less than unity, the flow around the sphere is

example of axisymmetric
to
a frame
and
Stokes flow. Let
us
an
transform
of reference in which the sphere is stationary,
centered
at
the
origin.
Adopting
the
standard
r, , , the surface of the sphere

corresponds to r = a, and the surrounding fluid occupies
the region r > a. By symmetry, the flow field outside the
spherical
coordinates
sphere
is axisymmetric
toroidal component
(i.e.,
= 0), and has no

= 0). The physical boundary
(i.e., /
v
conditions at the surface of the sphere
are
vr (a,)
(9.95)
(a,)
0,
0:
(9.96)
i.e., the normal and tangential fluid velocities
at the surface. A long
way
from the sphere,
fluid velocity to asymptote to v
= V ez
,)
v (r ,)
vr (r
(9.97)
are both zero

we expect the
In other words,
cos,
V sin.
(9.98)
Let
us write
(9.99)
where (r, ) is the stream function. As

previous
section,
axisymmetric
we saw
in the
is
flow
Stokes
characterized by
L
2()
0.
(9.100)
Here, the differential
operator L is specified in Equation
(9.88). The boundary conditions (9.95)(9.98)reduce
to
?????r=a
0,
?????r=a
(9.101)
0,
12
(r ,)
(9.102)
Vr
sin
(9.103)
Equation (9.103) suggests that (r, )
can be written
in
the separable form
(r,)
sin
f(r).
(9.104)
Inthis case,
(9.105)
2 cos f(r)
vr (r,)
r2
(r,)
sin
df
dr
,
,
(9.106)
and Equations (9.100)(9.103)reduce
(9.107)
(9.108)
!2
dr
f(a)
to
2
r2
df
dr
f(r )
0,
0,
?????r=a
1
2
Vr
2
(9
117)
182
FLUIDMECHANICS
z/a
3
3
x/a
Figure 9.4: Contours of the stream function in

the x-z plane for Stokes flow around
Let
us try a test
a solid sphere.
solution to Equation (9.107) of the
form f(r)
=r n.We find that
,,
[n (n 1) 2] [(n 2) (n 3) 2]
which implies
that
=1
= 0,
(9.110)
1 2, 4. Hence, the most
general solution to Equation (9.107) is
f(r)
Ar
+ B r+ C r2 +
Dr
(9.111)
where A, B, C, D are arbitrary constants. However, the
= (1/2) V and D =
(9.108) gives A = (1/4)
boundary condition (9.109) yields C
0, whereas the boundary condition
a3
and B
= (3/4)V
a. Thus, we conclude
f(r)
that
V (r a)
(2r + a)
4r
(9.112)
and the stream function becomes
(r,)
sin
V (r a)
(2r + a)
4r
,
.
(9.113)
See Figure 9.4. From (9.87), the fluid vorticity is
(r,)
L()
rsin
sin
dr
3V
a sin
2r
(9.114)
See Figure 9.5. Moreover, from (9.81),
Hence,
r
P
(9.116)
sin ).
(9.115)
3Va cos
r3
3Va sin
2r
,
,
(9 1
2 4)
183
z/a
3
3
x/a
Figure 9.5: Contours of the vorticity,

plane for Stokes flow around
inthe x-z
a solid sphere.
lines correspond to opposite signs of
Solid/dashed
which
implies
that
the
pressure
effective
within the fluid is
= p0 +
P(r,)
p0
where
is
However, P
an
p(r, )
From
2.20,
of the force
sphere by the fluid
f ()
vr (a,)
g r cos
the
per
9.6.
Figure
See
(a,)
cos
(9.119)
radial
and
area
tangential
exerted
on
the
,
.
!r=a
rr (a,)
3Va
2r
unit
are
fr ()
Now,
2r
(9.118)
constant.
distribution is
= p0
Section
components
(9.120)
3Va cos
= p + , where = gz = grcos . Thus,
pressure
the actual
arbitrary
distribution
1r
+ 2
vr
= v (a,) = 0. Moreover,
!r=a
vr
(9.121)
since
0, it follows from (2.169) that (vr /r)r=a

Equation (9.87) yields (v /r)r=a
fr ()
(9.122)
p(a,)
= p0
=
f ()
= (a,).
+ ga
cos
= 0.Finally,
Hence,
3V
cos,
2a
(a,)
3V
sin.
2a
(9.123)
Thus, the force density at the surface of the sphere is
f()
(9.129) 184
FLUID MECHANICS
3V
2a
ez + (p0 + ga
cos )er
z/a
3
3
x/a
Figure 9.6: Contours of the effective
pressure, P p0
a solid sphere.
inthe x-z plane for Stokes flow around
Solid/dashed lines correspond to opposite signs of P p0
It follows that the net vertical force exerted
on the
sphere
by the fluid is
IS
3V
4a
Fz
dS
+ 2a 2
(p0
2a
ez
+ ga
cos )cos sind,
(9.125) which reduces to
Fz
= 6aV +
a 3 g.
3
(9.126)
By symmetry, the horizontal components

both
on
average to zero.
the right-hand
We
side
can
of the net force
recognize the second term
of the above
equation
as
the
buoyancy force due to the weight of the fluid displaced by

the sphere. (See Chapter 3.) Moreover, the first term
be interpreted
as
the viscous drag acting
on
can
the sphere.
Note that this drag acts in the opposite direction to the

relative motion of the sphere with respect to the fluid, and
its
magnitude
is
directly
proportional
to the relative
velocity.
Vertical force balance requires that
Fz
M g,
(9.127)
where Mis the spheres
mass. In other
words, in a steady
state, the weight of the sphere balances the vertical force

exerted
by
composed
(4/3)a
the
surrounding
of material
of
fluid. If
mean
density
the sphere
is
then M
. Hence, in the frame in which the fluid
a large
distance from the sphere is stationary, the steady vertical

velocity with which the sphere
moves
V
where
/ is
the
Obviously, if the sphere is
if /
> 1)
versa.
a2
(9.128)
kinematic
fluids
more
through the fluid is
viscosity.
dense than the fluid (i.e.,
then it moves downward (i.e., V < 0),and vice
Finally, the typical Reynolds number of the fluid
flow in the vicinity of the sphere is
Re
(9.138)
For the
Re
C,
a3
???
g
2
185
20
2V
we
case
of
a grain of sand
have / 2 and
5) 3,
(a/6
10
where
falling through water at

6
10
m 2/s. Hence,
= 1.0
a
is measured
in meters.
Thus, expression
Re
(a/ 6
(9.128),
1 0 which is strictly speaking only
valid
when
Re
approximately
1, but
turns
which
valid for all Reynolds
out
approximately
water
Hence, Re
m/s. For the
through
air at 20
= 780 and
10
(a/4
meters. Thus, expression

droplets whose radii
are
5) 3,
case
of
less than
water
at
a droplet
of
fall through
have /
falling
pressure, we
10
less than
are
unity, only holds for sand grains whose radii
ess
about
than60
unity,
microns.
on ly hold
Such
s forgrains
to be
numbers
C and atmospheric
5
1.5
10
m 2/s.
where
is measured
1.5in
for water
(9. 128) only holds
less than about 40 microns. Such
droplets fall through air at approximately0.2m/s.
At large values of r/a, Equations (9.105), (9.106), and

(9.112) yield
vr (r,)
(9.130)
V cos
(r,)
32
V cos
ar
+O
ar
3
V sin
V sin
ar
?2
+O
ar
?2
(9.131)
It follows that
[(v
)v]r =
vr
vr
vr
r2
(9.132)
and
( v)r
Hence,
[(v
)v]r
immediate
Now,
our
advective
vicinity
number
is based
as is clear
on
if Re 1it
r
Re
.
,
(9.134)
of the flow in the
the
(9.94).]
assumption
with respect
that
to viscosity.
from the above expression for the ratio
even
if this ratio is
close to the spherein
other words,
of inertia to viscosity within the fluid,
much less than unity
r3
(9.133)
of the sphere. [See Equation
analysis
inertia is negligible
However,
(2v)r
where Re is the Reynolds
Va
vr
inevitably
unity far from the sphere:
clos
becomes
e to t he
much
sphere
greater
i n
than
othe
i.e., for
a/Re. In other
han
words,
unity
inertia
far from
always
thedominates
sphere:
viscosity, and
flow solution therefore
breaks
our
down, at sufficiently
Stokes
large
r/a.
9.11
Axisymmetric Stokes Flow Inand
Around
Suppose
a Fluid Sphere
that the solid sphere discussed
section is replaced by
in the previous
spherical fluid drop of radius
a.
Let
move
the drop
constant velocity V ez
fluid at the
through the surrounding
Obviously, the fluid from which the
drop is composed must be immiscible with the surrounding

fluid. Let
us
transform to
frame of reference in which
the drop is stationary, and centered at the origin. Assuming

that the Reynolds numbers immediately
the drop
are both much
outside and inside
less than unity, and making
use of
the previous analysis, the most general expressions for the
stream function outside and inside the drop

?
=sin
Ar
+Br +C r 2 + Dr 4
are
(9.135)
and
(r,)
=sin
Ar
+ Br +C r + Dr
(9.136)
respectively.
Here, A, B, C, etc.
are
arbitrary
Likewise, the previous analysis also allows
that
constants.
us to
deduce
vr (r,)
(9.137)
2 cos
(r,)
r3
sin
Br
+C + Dr
+
?
r3
Br
+ 2C + 4Dr
(9.142) 186
FLUID MECHANICS
2
z/a
2
2
x/a
Figure 9.7: Contours of the stream function inthe x-z

plane for Stokes flow inand around
Here,
/ 10.
(r,)
(9.139)
=
rr
+ gr
p0
6B
12A
cos
r4
r2
+ 12Dr
r > a, with analogous expressions in the

r < a. Here,
, p0 are the viscosity, density,
ambient pressure of the fluid surrounding the drop. Let
in the region
region
and
r4
+ 6Dr
(r,)
(9.140)
6A
sin
cos
a fluid sphere
, and p0 be the corresponding
that makes
In the region
outside
the drop, the fluid velocity
= V ez at large r/a. This implies

= (1/2)V and D = 0. Furthermore, vr (a,) =
must asymptote to
that C
quantities for the fluid
up the drop.
0i.e., the normal velocity at the drop boundary must be
zero
otherwise,
the drop would
constraint yields A/a
+ B/a+ (1/2)V
change
=0.
shape. This
Inside the drop, the fluid velocity must remain finite
= B = must
0. Furthermore,
velocity
remain fini
=0, which yields C +Da =
as r Inside
0. This
that
A
theimplies
drop, the
fluid
we
e
again require that
0.
Two additional
vr (a,)
physical
at the interface
satisfied
that must
constraints
the two fluids
between
firstly, continuity of tangential velocityi.e.,

v (a+
stressi.e.,
These
2C
secondly,
)and,
(a
constraints
+ 4Da
At
yield
and 6A/a
stage,
this
we
continuity
r (a+
A/a
v (a ,)
of
enough
tangential
).
B/a
= 6Da, respechave
be
are,
tively.
information
to
determine the values of A,B, C, and D. In fact, the stream

functions
outside and inside the drop
the form
(r,)
=
"
1
2
2
V a sin
4
(9.141)
and
ar
2+ 4
can be shown to take
+2
r
a
?2
(r,)
14
!
?
Va
sin
?2
"1
r
a
?2
(9.147) Incompressible Viscous Flow
187
2
z/a
2
2
x/a
Figure 9.8: Contours of the stream function inthe x-z

plane for Stokes flow inand around
Here,
/ 1/10.
respectively. See Figures 9.7 and 9.8.
a fluid sphere
The discontinuity in the radial stress

boundary is
rr
(a+ ,) rr (a ,)
"
= p0
p0
+ ( )ga
Va
cos 3
across
the drop
+ (3/2)
cos.
(9.143)
The final physical constraint that must be satisfied at r
is
rr
(a+ ,) rr (a ,)
=a
(9.144)
where
is the surface
tension of the interface between the
two fluids. (See Section 4.3.) Hence,
we obtain
p0 p0
(9.145)
and
where
,
V
a2
!"
1
3
= /is the kinematic
fluid. The fact that
we have
vertical
as
+ (3/2)
(9.146)
viscosity of the surrounding
at the interface between the
as the drop moves at

V, proves that our previous
long
velocity
been able to completely satisfy
all of the physical constraints
two fluids,
the constant
assumptions
that the interface is spherical, and that the drop
moves
a constant
speed
vertically
through the surrounding
fluid at
were correct. In the limit,

the drop is much more viscous
in which
than the surrounding fluid, we recover Equation (9.128):
i.e., the drop acts like a solid sphere. On the other hand, in
the limit
and , which
is O
appropriate to an air
sphere.
bubble rising through a liquid, we obtain
without changing shape,
188
FLUID MECHANICS
a2
9.12
Exercises
a plane that is inclined

x measure distance along
9.1. Consider viscous fluid flow down
at an angle to the horizontal. Let
the plane (i.e., along the path of steepest decent), and let
be a transverse coordinate such that the surface of the plane
to y
corresponds
= 0, and the free surface
h. Show that within the fluid (i.e., 0
p(y)
y(2h y),
+ g
p0
cos(h y),
where is the kinematic viscosity,

atmospheric
y=
gsin
vx (y)
of the fluid to
y h)
pressure.
9.2. If a viscous fluid flows along
the density, and p0
a cylindrical
is
pipe of circular
cross-section that is inclined at an angle to the horizontal show
that the flow rate is

Q
where
(G + g sin ),
a is the pipe radius,
the fluid viscosity,
the fluid
density, and G the pressure gradient. 9.3. Viscous fluid flows

steadily, parallel to the axis, inthe annular region between two
coaxial cylinders of radii a and
n> 1.Show
na, where
that the volume flux of fluid flow is
a4
where G is the effective
"n
4
(n
1)
lnn
pressure
viscosity. Find the mean flow speed.
gradient, and
the
9.4. Consider viscous flow along
and
that its boundary satisfies
x2
a2
= 1.
b2
Let
v = vz (x,y)ez
Demonstrate that
x 2
where G is the effective
vz
pressure
gradient, and
viscosity. Show that
vz (x,y)
a2
a cylindrical pipe of elliptic

runs parallel to the z-axis,
cross-section. Suppose that the pipe
x2
a +b
2
a2
is a solution of this equation that satisfies the
the fluid
no slip
condition at the boundary. Demonstrate that the flow rate is
Finally, show that
a3
a +b 2
2
a pipe with an elliptic
cross-section has
lower flow rate than an otherwise similar pipe of circular

section that has the same cross-sectional
9.5. Consider
a velocity
field of the form
r (r) sin ,
r,, are spherical coordinates. Demonstrate
v(r)
where
cross-
area.
that
this field satisfies the equations of steady, incompressible, viscous
fluid flow (neglecting advective inertia) with uniform
pressure
(neglecting gravity) provided that

d
dr
Suppose that
dr
a solid sphere
= 0.
of radius
a,centered at the
origin, is rotating about the z-axis, at the uniform angular

velocity
0
ina viscous fluid, of viscosity
infinity. Demonstrate that
(r)
for
r a. Show
= 0
=80 a
189
a solid sphere
viscous fluid of viscosity
r3
,
.
that the torque that the sphere exerts
fluid is
9.6. Consider
a3
that is stationary at
on the
of radius
a moving
at the fixed velocity V
through
=V ez
Let
r,
be spherical coordinates whose origin coincides with the

instantaneous location of the spheres center. Show that, if
inertia
and gravity
are negligible, the fluid velocity, and the radial

a long way from the sphere, are
components of the stress tensor,
ar
32
vr
V cos
ar
3
v
V sin
,
,
a
92
rr
p0
Vcos
r2
0,
on
r,by the
respectively. Hence, deduce that the net force exerted

the fluid lying inside
a large
spherical surface of radius
fluid external to the surface, is

F
= 6aV,
independent of the surface radius.
190
FLUID MECHANICS
Waves
in
Incompressible
Fluids
191
10 Waves inIncompressible Fluids

10.1
Introduction
This chapter investigates low amplitude
waves
propagating
through incompressible fluids.
10.2
Gravity Waves
Consider
a stationary
body of water, of uniform depth d,
located
on the
unperturbed
corresponding
that
small compared
surface
coordinate
Cartesian
surface of the Earth. This body is assumed
to be sufficiently
small
to the Earth that its
is approximately
z measure
planar.
to the aforementioned
surface.
wave propagates
amplitude
Let
the
z=0
vertical height, with
Suppose
horizontally
through the water, and let v(r,t)be the associated velocity

field.
Since water is essentially incompressible,
of motion
are
v
(10.1)
+ (v )v
viscosity,
is the
and
(uniform)
the
density,
(uniform)
gravity. (See Section 2.14.) Let
p(r, t)
0,
2
v,
(10.2)
the (uniform)
acceleration
us write
= p0
p gez +
where
its equations
gz +
due
to
p1 (r, t),
(10.3)
pressure, and p1 the pressure

perturbation due to the wave. Of course, in the absence of
the
wave, the water pressure
a depth h below the surface is p0 + gh.(See Chapter 3.)
where p0
is atmospheric
Substitution into (10.2) yields
+ v,
p1
are
second-order
(10.4)
where
we
have neglected
terms that
small quantities (i.e., terms of order

Let
us
also
approximation
neglect
provided
viscosity,
that
the
in
2).
which
is
wavelength
good
not
is
ridiculously small. [For instance, for gravity waves in water,

2/g) 1/3
5
viscosity is negligible as long as (
5 10
m.] It follows that
p1
(10.5)
Taking the curl of this equation,
we obtain
0,
(10.6)
where
velocity
=
field
Conse-quently,
is the vorticity. We conclude that the
associated
the
with the
previous
wave
equation
is
is irrotational.
automatically
(10.7)
where (r,t) is
velocity potential. (See Section 5.7.)
However, from Equation (10.1), the velocity field is also
divergence-free.
It
follows
that
the
velocity
potential
satisfies Laplaces equation,
0.
(10.8)
Finally, Equations (10.5) and (10.7) yield
p1
(10.9)
We
must
now
be
need to derive the physical constraints
satisfied
at
the
upper
waters
and
that
lower
boundaries. It is assumed that the water is bounded from

below by
solid surface located at
= d.
192
remain
Since the
FLUIDMECHANICS
in contact
physical constraint
with this
surface,
the
appropriate
at the lower boundary is
=0
vz |z=d
(i.e., the normal velocity is zero at the lower boundary),
or
?????z=d
0.
(10.10)
The waters
since it is
upper
a free
boundary is
little
more
surface. Let represent
complicated,
the vertical
displacement of this surface due to the wave. It follows that
?????z=0
= vz |z=0 =
(10.11)
The
appropriate
boundary
physical
discontinuity
pressure, since
across a free
surface tensionsee
(10.3),
pressure
is that the water
atmospheric
at
constraint
there must equal
there cannot
be
a pressure
(in the absence
surface
10.11).
Section
upper
the
of
Accordingly, from
we obtain
p0
g +
p0
p1 |z=0
(10.12)
or
g
p1|z=0
(10.13)
which implies that
?????z=0
=g
?????z=0
p1
,
,
,
(10.14)
whereuse has beenmadeof (10.11).
above expression
can
givethe boundary
The
be combined with Equation (10.9)
condition
??????z=0
?????z=0
t 2
(10.15)
Let
us
search for
wave-like
solution of Equation
(10.8) of the form
(r,t)
F(z)
(10.16)
This solution actually corresponds to
wave
of
wave vector
= kex
cos(t k x).
a propagating
plane
angular frequency , and
amplitude F(z). Substitution into Equation (10.8) yields

2
d F
dz 2
0,
(10.17)
whose independent
Hence,
a general
=Ae
(x,z,t)
solutions
are
exp(+k z) and exp(k z).
solution to (10.8) takes the form
kz
cos(t k x)
+ Be kz
cos(t k x),
(10.18)
where
The
and
boundary
are
condition
satisfied provided that B

he
(x,z,t)
=A
arbitrary
kz
constants.
(10.10)
is
= A exp(2kd),
giving
+ e k(z+2d)
cos(t k x),
(10.19)
The boundary condition (10.15) then yields
?1
A k
2kd
cos(t k x)
=A
?1
+ e 2kd
cos(t k x).
g
(10.20)
which reduces to the dispersion relation
gk
tanh(kd).
(10.21)
The type of
gravity
wave
described in this section is known
as a
wave.
Waves inIncompressible Fluids
193
10.3
Gravity Waves inDeep Water
Consider the so-called deep water limit,

kd 1,
(10.22)
in which the depth, d, of the water greatly exceeds the
wavelength,
gravity
wave
= 2/k,
of the
wave.
In this limit, the
dispersion relation (10.21) reduces to
(gk)
1/2,
(10.23)
x . It follows
waves in deep water
since tanh(x) 1
as
that the phase
velocity of gravity
is
vp
gk
?1/2
(10.24)
Note that this velocity is proportional to the
with
long wavelengths
propagate faster than those with
short wavelengths. Now, the phase velocity,
as the propagation
wave number, k
defined
definite
square root
waves
Hence, deep water gravity
of the wavelength.
velocity of a plane
[and
frequency
vp
= /k, is
wave
with the
given by the
a wave has an infinite

wave of finite spatial
extent, with an approximate wave number k,can be formed
as a linear superposition of plane waves having a range of
different wave numbers centered on k. Such a construct is
known as a wave pulse. As is well-known, wave pulses
propagate at the group velocity,
dispersion
relation (10.23)]. Such
spatial extent. A
more
realistic
vg
d
dk
(10.25)
For the
case
of gravity
relation (10.23) yields
vg
12
gk
?1/2
12
vp
waves
in deep water, the dispersion
(10.26)
In other words, the
group
velocity of such
waves
is half
their phase velocity.
of a particle of water,
r and time t, due to the passage of a
wave. It follows that
Let (r, t) be the displacement

found at position
deep water gravity
v,
(10.27)
where
For
=kex
v(r, t) is
plane
wave
inthelimitkd
the
of
perturbed
wave
velocity.
number
1, Equation(10.19)
(x,z,t)
Ae
kz
yields
cos(t
k x).
(10.28)
Hence, [cf., Equations (10.45)(10.48)]
(x,z,t)
(10.29)
(x,z,t)
(10.30) vx (x,z,t)
(10.31) vz (x,z,t)
(10.32)
and
ae
=
=
=
kz
cos(t k x),
ae kz
ae
ae
sin(t k x),
kz
kz
sin(t k x),
cos(t k x),
p1
=gz
(10.33)
where
use
has been made of Equations (10.7), (10.9), and
(10.27). Here,
at
a is the
waters
the
amplitude of the vertical oscillation
surface.
According to Equations (10.29)(10.32), the passage of the
wave causes a water
particle located
surfaceto executea circularorbit
depth hbelow the
ofradius
ae kh
equilibrium
position. Note that the radius
decreases
exponentially
with
aboutits
of the orbit
increasing
depth.
Furthermore, whenever the particles vertical displacement

attains
horizontally
versa. See
maximum
in the
value
same
the
direction
particle
as
the
is
moving
wave,
and vice
Figure 10.1.
194
FLUIDMECHANICS
z
surface
z =0
Figure 10.1: Motion of water particles associated with a

deep water gravity
Finally, if
we
wave
propagating inthe x-direction.
define h(x, z, t)
= z (x, z,t)
depth of the water particle found at
equilibrium
z as the
a given
point and time then Equations (10.3) and (10.33) yield

p(x,
z, t) =
p0
h(x,
z, t).
(10.34)
pressure at this point and time is

unperturbed
pressure calculated at
In other words, the
the
same as
the
the
equilibrium
10.4
depth of the water particle.
Gravity Waves in Shallow Water
Consider the so-called shallow water limit,
kd
1,
(10.35)
in which the depth, d, of the water is much less than the
wavelength,
gravity
wave
= 2/k,
of the
wave.
In this limit, the
dispersion relation (10.21) reduces to
(gd)
1/2
k,
(10.36)
x as x 0. It follows that
velocities
and
groupvelocities
since
tanh(x)
x as x of gravity waves
since tanh(x)
the phase
in shallow
water alltake the fixed value

vp
vg
(gd)
1/2,
(10.37)
irrespective
of
wave
number. We conclude thatunlike
deep water wavesshallow
water gravity
waves are nonwaves all
dispersive in nature: i.e., waves pulses and plane
propagate at the
same
speed. Note, also, that the velocity
(10.37) increases with increasing water depth.
For
=kex
plane
wave
of
wave
number
in the limit kd 1,Equation (10.19) yields

2
2/2]
(x, z, t)
A [1+ k
(z + d)
cos( t k x).
(10.38)
Hence, Equations (10.7) and (10.27) give [cf., Equations
(10.45)(10.48)]
x
(10.39)
=
=
=
=
(x,z,t)
(x,z,t)
(10.40) vx (x,z,t)
(10.41) vz (x,z,t)
(10.53)
Waves
in
a(kd)
cos(t k x),
a(1+ z/d) sin(t k x),
a(kd)
a(1
sin(t k x)
+ z/d) cos(t
Incompressible
k x).
Fluids
195
Here,
at
the
is again the amplitude of the vertical oscillation
waters
surface.
According
to
the
above
the passage of a shallow water gravity wave

causes a water particle located a depth hbelow the surface
to execute an elliptical orbit, of horizontal radius a/(k d),
and vertical radius a ( 1 h/d), about its equilibrium
expressions,
position.
a/(kd)Note
and
that
vertithecal
orbit
radius
is a(1
greatly
elongated
h/d),abo in
ut its
the
horizontal
decreases
becomes
whenever
direction.
linearly
zero at
Furthermore,
with
its
the bottom (i.e., at h
the particles
vertical
vertical
depth
increasing
radius
such
that
displacement
direction
10.5
attains
maximum value the particle is moving horizontally
same
it
= d). As before,
in the
as the wave, and vice versa.
Energy of Gravity Waves
It is easily demonstrated,
in the
previous
arbitrary
wavenumber
from the analysis contained
sections,
that
k, propagating
wave
gravity
horizontally
of
through
water of depth d,has a phase velocity

"
vp
=(gd)
1/2
#1/2
tanh(kd)
kd
(10.43)
Moreover, the ratio of the group to the phase velocity is
"1
vg
vp
1
2
2kd
sinh(2kd)
(10.44)
Note, that neither
the phase velocity
nor
the
group
velocity of a gravity wave can ever exceed the critical

1/2.
value (gd)
It is also easily demonstrated that the
displacement
gravity
of wavenumber kex
surface amplitude
(10.46)
a,are
vx (x,z,t)
+ d)]
(x,z,t)
with
plane
angular frequency , and
cosh[k(z
(x,z,t)
(10.45)
and velocity fields associated
wave
cos(t k x),
sinh(kd)
sinh[k(z + d)]
sin(t k x),
sinh(kd)
cosh[k(z + d)]
sin(t k x),
sinh(kd)
(10.47)
vz (x,z,t)
+ d)]
sinh[k(z
cos(t k x).
sinh(kd)
(10.48)
mean kinetic energy per

with a gravity wave is defined
Now, the
associated
unit surface
=h
v
d
area
dzi,
where
(x,t)
is the vertical displacement
= a sin(t
Z
hi
d(k x)
()
is an average
k x)
at the surface, and
over a wavelength.
(10.48) that, to second order in a,
Given that hcos
2(t
k x)i
= hsin 2(t
(10.49)
(10.50)
(10.51)
k x)i
= 1/2, it follows
from (10.47) and

K
14
a
cosh[2k(z
sinh
+ d)]
2(kd)
dz
1
ga
gk tanh(kd)
(10.52)
Making
use
of the general dispersion relation (10.21),
we
obtain
1
ga
aincan 196
FLUID MECHANICS
The mean potential
area associated
energy
perturbation
with a gravity
wave
per unit surface
is defined
2.
=h
1
gd
2
=h
1
g(
2
gzdzi
d
2,
(10.54)
which yields
U
2)i
12
gd
gh
2i,
2
(10.55)
or
1
ga
4
2.
(10.56)
Inother words, the mean potential
energy per unit surface

area of a gravity wave is equal to its mean kinetic energy
per unit surface area.
Finally, the mean total energy per unit surface area
associated with a gravity wave is
= K +U=
12
ga
2.
(10.57)
Note that this energy depends
on the wave
amplitude at the
surface, but is independent of the wavelength,
or the water
depth.
10.6
Wave Drag
on Ships
a
over a body of water leaves behind it a train
of gravity waves whose wavefronts are transverse to the
ships direction of motion. Since these waves possess
energy that is carried away from the ship, and eventually
dissipated, this energy must have been produced at the
ships expense. The ship consequently experiences a drag
Under certain circumstances (see the following section),
ship traveling
force, D. Suppose that the ship is moving at the constant
velocity
making
V. We
up
so that they
would
expect
the train to have
the
transverse
waves
matching phase velocity,
a constant phase relation with respect

more exact, we would generally expect
bow to always correspond to a wave maximum
of the pile up of water in front of the bow
by the ships forward motion). The condition v p
maintain
to the ship. To be
the ships
(because
produced
=V,combined
tanh(kd)
kd
with expression (10.43), yields

2
gd
(10.58)
Suppose, for the sake of argument, that the
wave
train is
transverse
of uniform
drawn
downstream
width
w.
Consider
of the ship at right-angles
fixed line
to its path.
The rate at which the length of the train is increasing

line is V. Therefore, the rate at which the
ahead of this
energy of the train is increasing ahead of the line is (1/2)

g a 2 w V, where a is the typical amplitude of the
transverse waves in the train. As is well-known, wave
energy travels at the group velocity, rather than the phase
velocity. Thus, the energy flux per unit width of a
propagating
gravity
wave is simply
E vg
Wave
energy consequently
crosses our fixed line
in the
direction
the
of
ships motion at the rate (1/2)ga
wvg
does work against the drag force, which

the energy of the train in the region ahead
Finally, the ship
goes to increase
of our line, at the
rate DV. Energy conservation thus yields

12
ga
wV
12
ga
(10.59)
However, since V
=v p
wvg
+ DV.
we obtain
"1
ga
2
(10.60)
vg
vp
14
ga
2kd
sinh(2kd)
where
use
has been made of Equation (10.44). Here,
kd is determined
via
implicitly in terms of the ship speed
expression (10.58). Note that this expression cannot
be satisfied when the speed exceeds the critical value

1/2,
since gravity waves cannot propagate at speeds in
(gd)
excess
of this value. In this situation,
no transverse wave
197
up with the ship, and the

waves consequently disappears.
keep
drag associated with such

In fact, we
the above formulae, that when V (gd)
0, and
wave
so
D 0. Note, however,
amplitude,
a, generally
ship
wa vespeed approaches
due to transverse
can see, from
1/2
then kd
that the transverse
increases significantly
as the
the critical value. Hence, the drag
waves
actually peaks strongly at speeds
just below the critical speed, before effectively
falling to
zero as
this speed is exceeded. Consequently,
it usually
requires
a great
deal of propulsion power to force

1/2.
a ship
to travel at speeds faster than (gd)

In the deep water
limit
kd 1, Equation
(10.60)
reduces to
D
14
ga
w.
(10.61)
Note that (at fixed

independent
hence,
independent
actually
rather
amplitude)
wave
amplitude)
of the wavelength
of the ships
misleading.
the drag
acting
In
We
can account
for this
wave
is
train, and,
speed. This result is

fact,
on a
deep water varies significantly
this expression
of the
(at
wave
fixed
ship traveling
through
with the ships
speed
variation by incorporating
the
our analysis. A real ship

moving through water generates a bow wave from its bow,
and a stern wave from its stern. Moreover, the bow wave
tends to have a positive vertical displacement, because
water naturally piles up in front of the bow due to the
forward motion of the ship, whereas the stern wave tends
tends to have a negative vertical displacement, because
finite length of the ship into
water rushes into the void left by the stern. Very roughly
speaking,
suppose
that the vertical
displacement
of the
water surface caused by the ship is of the form

?
(x)
cos
xl
(10.62)
Here, l
is the length of the ship. Moreover, the bow lies
at x = 0 [hence, (0) > 0], and the

= l[hence, (l) < 0]. For the sake of
(instantaneously)
stern at
simplicity,
the upward
bow is assumed
water displacement
due to the
to equal the downward displacement
to the stern. At fixed
bow
wave
displacement,
due
the
amplitude of transverse
g/ V
so
(chosen
gravity
waves
of
wave
that the phase velocity
number k
waves
of the
matches the ships speed, V) produced by the ship is

Z
1l
cos
xl
cos(k x)dx
sin( kl)
kl
kl
+ kl
(10.63)
i.e., the amplitude is proportional to the Fourier coefficient
pattern evaluated for
of the ships vertical displacement
wave
number that matches that of the
(at fixed bow
the transverse
wave displacement)
waves is
wave
the drag produced by
"sin(
a2
train. Hence,
#2
2)
1
1
+ F
F 2
(10.64)
where the dimensionless parameter
F
V
(gl) 1/2
(10.65)
is known
as the Froude
number. See Section 2.16.
Figure 10.2 illustrates the variation of the

with Froude number predicted
we can see, if the Froude
wave
drag
by Equation (10.64). As
number is much less than unity,
which implies that the wavelength
of the
wave
train is
much smaller than the length of the ship, then the drag is
comparatively
small. This is the
case
because the ship is
extremely
waves.
Froude
at driving short wavelength
inefficient
can
It
seen
also be
number
increases,
maximum when Fr
rapidly. Now, Fr
reaching
= Frc = 1/
= Frc
gravity
as
that the drag increases
relatively
the
sharp
= 0.56, and then falls
corresponds to the
case
in which
the length of the ship is equal to half the wavelength of

the
wave
train. In this situation, the bow and stern
waves
a particularly
to a particularly
large
constructively,
interfere
amplitude
wave
wave
drag.
to
leading
train, and, hence,
The
smaller
peaks
visible
correspond to other situations in which
waves
length
constructively.
interfere
of
the
ship
wavelengths of the
displacement,
cannot
= Frc
Froude numbers
one
propulsion
=0.56(gl)
1/2
half
power,
generally
wave drag that occurs

a ship is, therefore, limited to
inthe range 0 < Fr < Frc which implies a
the peak in the
Such
maximum speed of
Vc
when the
and
train.) A heavy ship with large
and limited
overcome
when Fr
to
figure
the bow and stern
(For instance,
corresponds
wave
in the
large
1.75[l(m)]
1/2
m/s
=3.4[l(m)]
1/2
kts.
(10.66)
This characteristic
speed
is sometimes
called the hull
speed. Note that the hull speed increases with the length
of the ship: i.e., long ships have higher hull speeds than
short
198
ones.
FLUID MECHANICS
0.3
0.2
.
.
)
u
a
D
0.1
0
0
0.2
0.4
0.6
0.8
Fr
Figure 10.2: Variation of wave drag with Froude

number for a ship traveling through deep water.
10.7
Ship Wakes
1.2
Let
us now
make
detailed
pattern generated behind
investigation
water, taking into account obliquely
waves,
in addition to transverse
simplicity,
of the wake
a ship as it travels over a body

propagating
waves.
of
gravity
For the sake of
the finite length of the ship is neglected in the
as a
source of gravity waves. Consider Figure 10.3. This
shows a plane gravity wave generated on the surface of
the water by a
moving ship. The water surface
corresponds to the x-y plane. The ship is traveling along
following analysis. In other words, the ship is treated
point
x-direction, at the constant speed
the x-axis, in the negative
V. Suppose that the ships bow is initially at point A

has moved to point A after
type of gravity
wave
a time
and
interval t.Now, the only
that is continuously
excited by the
a constant phase
relation with respect to its bow. In fact, as we have
already mentioned, the bow should always correspond to a
wave maximum. An oblique wavefront associated with
such a wave is shown in the figure. Here, the wavefront
,
,
D
which initially passes through the bow at point A
has moved to CD after a time interval t, such that it again
passes through the bow at point A. Of course, the
passage
of the ship is one that
maintains
wavefront propagates at the phase velocity,
that, in the right-angled
A Eare of lengths V t and v p t,respectively,
vp
It follows
triangle AA E, the sides AA
so that
and
sin
vp
V
(10.67)
This, therefore, is the condition that must be satisfied in
order for an obliquely propagating gravity wave to
maintain
constant phase relation with respect to the ship.

In shallow water, allgravity
same
waves propagate at the
phase velocity:
=(gd)
vp
1/2,
where d is the water depth. Hence, Equation (10.67) yields

"
Note that this equation
can only
=sin
(gd)
1/2
be satisfied when
V > (gd)
1/2.
i.e.,
(10.68)
(10.69)
(10.70)
199
vp t
Vt
Figure 10.3: An oblique plane
wave
generated
the surface of the water by a moving ship.
on
y
D
Figure 10.4: A shallow water wake.
In other words, the ship must be traveling faster than the

1/2.
critical speed (gd)
Moreover, if this is the case then
there is only
one value
of that satisfies Equation (10.69).
This implies the scenario illustrated in Figure 10.4. Here,

the ship is instantaneously
that
it
previously
obliquely, subtending
at A, and the
generatedwhich
a fixed
wave
all
maxima
propagate
angle with the x-axishave
maximum
DAE. In fact, the
was at A
when the ship
wave
maxima
generated
travelled
to B
traveling
over
and C
shallow
whose semi-angle,
Indeed,
as
over
maxima
when the ship
generated
and C
was at
etc. We conclude that
water produces
from
wave
single strong
have travelled to B
, is determined
is apparent
the ship travels
a
wave
to produce
interfered constructively
the
have
ship
V-shaped wake
by the ships speed.
Equation (10.69), the faster
the water, the smaller the angle
becomes. Shallow water wakes

other vessels, and particularly
are especially
dangerous to
destructive of the coastline,
wave energy produced by the ship is

a single large wave maximum. Note,
finally, that the wake contains no transverse waves, since,
as we have already mentioned, such waves cannot keep
up with a ship traveling faster than the critical speed (
because all of the
concentrated
into
1/2.
Let us now discuss the wake generated by a ship
traveling over deep water. Inthis case, the phase
(10.77) 200
FLUID MECHANICS
vg t
Vt
Figure 10.5:
Formation of an interference maximum in a deep water wake.
gravity
waves
is v p
=(g/k) 1/2. Thus, Equation (10.67)
yields
sin
vp
?1/2
g
kV
(10.71)
It follows that indeep water
gravity
value
wave
whose
wave
any
obliquely propagating
number exceeds the critical
k0
(10.72)
can
keep
direction
up
as
with the ship,

propagation
of
long
as
its
is
such that Equation (10.71) is satisfied. Inother words, the

ship continuously
range
of
excites
different
gravity
wave
directions. The wake is essentially

generated by these
waves.
waves
numbers
Now,
and
with
wide
propagation
the interference
pattern
as is well-known, an inter-
ference maximum generated by the superposition of plane
waves
with a range of different wave numbers propagates

group velocity, vg Furthermore, as we have already
seen, the group velocity of deep water gravity waves is
half their phase velocity: i.e., vg = v p /2.
Consider Figure 10.5. The curve APD corresponds to
a particular interference maximum in the wake. Here, A is
the ships instantaneous position. Consider a point P on
this curve. Let x and y be the coordinates of this point,
at the
relative to the ship. Now, the interference maximum at P is
part of the plane wavefront BC emitted

when the ship
subtended
was at point
between
Since interference
this
wavefront
some
and
vg t.Of
the
x-axis.
group velocity,
course, the distance
maxima propagate at the
the distance A P is equal to
time t earlier,
Let be the angle
AA
is equal to V t.
Simple trigonometry
reveals that
Vt
vg t sin,
vg t cos.
Moreover,
dy
= tan,
dx
curve APDi.e.,
vg =v p /2,that
since BC is the tangent to the
y(x)at point P.fact that
the curve
12
X
sin
12
X sin cos,
(10.73)
(10.74)
(10.75)
It follows from Equation (10.71), and the
(10.76)
(10.77)
Waves
201
in
Incompressible
Fluids
0.3
0.2
0.1
0
X
/
0.1
0.2
0.3
A
C
0.1
0.2
0.3
0.4
0.5
0.6
x/X0
Figure 10.6: Locus of an interference maximum in a deep
water wake.
where X()
=V t.The previous
three equations
can be
combined to produce
dy
dx
dy/d
dx/d
(1/2)dX/d sin cos
+ (1/2)X(cos
dX/d[1 (1/2) sin
sin
cos
sin
] X sin cos
(10.78)
which reduces to
dX
tan
(10.79)
This expression
can be solved to give

X
=X0
sin,
(10.80)
where X0 is a constant. Hence, the locus of our

interference maximum is determined parametrically by
1
X0 sin
sin
2
(10.81)
12
X0 sin
cos.
(10.82)
Here, the angle
ranges
curve
from /2 to +/2. The
specified by the above equations is plotted in Figure 10.6.
As usual, A is the instantaneous
can be
10.6.
As
seen
of the transverse
consists
position of the ship. It
that the interference
radial maxima
maximum
maximum
essentially
BCD, and the two
AB and AD. As is easily demonstrated,
point C, which corresponds to

0,lies at x
X0 /2,
8)
y
0. Moreover, the two cusps, B and D, which
8)
1(1/
, lie at x
correspond to tan
19.47
(8/27)
1/2
X0
y = (1/27)
1/2
X0
The complete interference
pattern that constitutes the
X0
many
different wave maximum

curves of the form shown in Figure 10.6, corresponding to
many different values of the parameter X0 However, these
wake is constructed out of
values must be chosen such that the wavelength
the
pattern
wavelength
along
= 2/k0 =
= 0) gravity
the
waves
of
x-axis
corresponds
to the
2/g
2 V
of transverse (i.e.,
whose phase velocity matches the
speed of the ship. This implies that X0
= 2 j0
where j
is a positive integer. A complete deep water wake pattern

is shown in Figure 10.7. Note that this pattern, which is
made
up
of
maxima, fills
wedgewhose
interlocking
transverse
and
radial
wedge-shaped regionknown
as a
semi-angle takes the value tan
1(1/
wave
8)
Kelvin
8)
(10.89)
202
FLUID MECHANICS
6
5
4
3
2
1
0
y
1
2
3
4
5
6
1 0
5
x/0
910 11
Figure 10.7: A deep water wake.
independent
of
the
ships
assumption that the gravity
water
all deep
waves
speed.
waves
our
Finally,
initial
that form the wake
are
is valid provided k0 d 1 which
implies that
(gd)
1/2.
(10.83)
Inother words, the ship must travel at a speed that is much
1/2.
less than the critical speed (gd)
This explains why the
wave
wake contains transverse
10.8
Consider
gravity
flowing horizontally
us
maxima.
Gravity Waves in a Flowing Fluid
wave
traveling through
= V ex
write
v(r,t)
(10.84)
p(r,t)
fluid that is
at the uniform velocity V
Let
V + v1(r,t),
=
p0
gz + p1(r,t),
(10.85)
where
v1 and p1
perturbations,
are
the small velocity
respectively, due to the
and
wave.
pressure
To first order
in small quantities, the fluid equations of motion, ( 1

0.1
)
and (10.2), reduce to
v1
0,
(10.86)
+V
v1
p1
(10.87)
respectively. We
of
in
can also
define the displacement, (r, t),
a fluid particle due to the passage of

a frame co-moving with the fluid, as
the
wave, as seen
+ V =v1
(10.88)
The curl of Equation (10.87) implies that

Hence,
we can write
v1
= , and (10.87) yields
v1
= 0.
+V
p1
( 10. 96)
203
z =d
z= 0
z = d
Figure 10.8: Gravity
waves at an interface
between two immiscible fluids.
Finally, Equation (10.86) gives
(x,z,t)
[A
=kex
0.
(10.90)
The most general traveling

with wavevector k
wave
solution to (10.90),
, is
+ B sinh(kz)]cos(t k x).
and angular frequency
cosh(kz)
(10.91)
It follows from Equation (10.89) that
p1 (x,
= k(V
c) [A cosh(kz)
+ B sinh(kz)]
z,
t)
sin(t k x),
(10.92)
and from Equation (10.88) that

z
sinh(k z)
(x,
z, t, ) =
(V
c)
[A
B cosh(k z)] sin( t k x).
(10.93)
Here,
c =/k
Gravity Waves at an Interface
10.9
Consider
, and
depth d
,
V
situated on top of a layer
layer of fluid of density
uniform horizontal velocity
another
of
is the phase velocity of the wave.
fluid
of density
, depth
d, and uniform
horizontal velocity V. Suppose that the fluids
are
from above
planes. Let
and below
these planes
lie at
unperturbed interface
by rigid horizontal
bounded
zand
= belo
d w
and
b yzri gid
= dho, rizo
andntlet
al
pthe
la
between the two fluids lie at z = 0.
See Figure 10.8.

Consider
wavenumber
a gravity wave
of angular frequency , and
k,propagating through both fluids in the
x-
direction. Let
(x, t)
= 0
sin( t k x)
(10.94)
be the small vertical displacement
wave.
the
potential
In the
lower
so that
the
perturbed
velocity
must be of the form (10.91), with the constants A
vz
= 0 and z (x,0,t)= (x, t).
= (V c)0 / tanh(k d) and B = (V c)
and B chosen such that
It follows
fluid,
of the interface due to
that A
|z=d
(x,z,t)
cosh[k(z
= (V c)0
+ d)]
cos(t k x).
sinh(kd)
(10.95)
upper
In the
fluid, the perturbed velocity potential must
again be of the form (10.91), with the constants A and B

chosen
such
vz |z=d
that
=0
and z (x, 0, t)
(x, t). It follows that A
(V c)0 /tanh(kd
and B
(V
c)0 so that
)]
cosh[k(z d
(x,z,t)
(V c)0
cos(t k x).
)
sinh(kd
204
FLUIDMECHANICS
Here,
c = /k
is the phase velocity of the wave. From
Equations (10.85) and (10.92), the fluid
pressure
just
below the interface is

p(x,0 ,t)
p0
p0 g
g0
+ p1(x,0
,t)
sin(t k x) k(V c)A sin(t k x)

"g
k(V c)
p0
#
0
sin(t k x).
tanh(kd)
(10.97)
Likewise, the fluid
p(x,0+ ,t)
p0
pressure

just above the interface is
+p1(x,0+ ,t)
=
p0
g0
sin(t k x) k(V
c)A sin(t k x)
"
p0
g+
k(V
c)
#
0
tanh(kd )
sin(t k x).
(10.98)
Now, inthe absence of surface tension
at the interface, these two pressure must equal one another:

i.e.,
?p ?z=0+
z=0
=0.
(10.99)
Hence,
we obtain
the dispersion relation
)g
k(V c)
+
tanh(kd)
(V
c)
tanh(kd )
(10.100)
a quadratic equation
c,of the wave. We can see that:
which takes the form of

velocity,
i.If
for the phase
= 0 and V =0 then the dispersion

=c).
relation
reduces to (10.43) (with v p

ii.If the two fluids
are of infinite
depth then the
dispersion relation simplifies to

)g
2
(
k(V c)
+ k
(V
c)
(10.101)
iii.Ingeneral, there
are two values of c that satisfy the

are either both real, or
quadratic equation (10.100). These
form a complex conjugate pair.
iv.The condition for stability is that

alternative is that
c is complex,
c is real. The
which implies that is
also
complex, and, hence, that the perturbation
grows or
decays exponentially intime. Since the complex roots of
a quadratic
equation
occur
incomplex conjugate
pairs, one of the roots always corresponds to an

exponentially
growing mode: i.e., an instability.
v.If both fluids are at rest (i.e., V =V
=0),and of infinite depth, then the dispersion
reduces to
c2 =
g( )
k(
+ )
(10.102)
It follows that the configuration is only stable
when
> :
i.e., when the heavier
fluid is underneath.
As a particularexample, supposethat the lower fluid is
water, and the upperfluid is the atmosphere. Let

1.225 10
to water). Putting V
the fact that
yields
s = / =
be the specific density of air at s.t.p. (relative
= V =0,d
s is small, the dispersion
, and making
use
relation (10.100)
of
"
c (gd)
(1
#1/2
tanh(kd)
1/2
kd
s[1+ tanh(kd)]
(10.103)
Comparing this with (10.43),
presence
we can see that
the
of the atmosphere tends to slightly diminish the
phase velocities of gravity

surface of
a body
waves
propagating
over
the
of water.
205
10.10
a stream
Consider
over
Flow
Steady
Bottom
of water of
horizontal velocity V
= V ex
mean
Corrugated
depth d, and uniform
that flows
over a corrugated
= d + a sin(k x), where a is

much
smaller
than
d. Let
the
of the
free
bottom
whose
elev
ation
iselevation
z = d +a
sin(k
x),surface
where
of the water be z = b sin(k x). We wish to determine the
bottom whose elevation is
a and b.
we expect the velocity
relationship between
Now,
pressure,
and vertical displacement
potential,
perturbed
of the water to be of
the form (10.91), (10.92), and (10.93), respectively, with
=c =
0, since
non-propagating)
we are
looking for
perturbation
driven
stationary
by
the
(i.e.,
static
corrugations in the bottom. The boundary condition at the
bottom is
z
(x, d)
sin(k x),
(10.104)
which yields
[A
sinh(kd)
B cosh(kd)]
a.
(10.105)
At the free surface,
we have
z
(x,0)
sin(k x),
(10.106)
which gives
b
B.
(10.107)
Inaddition,
pressure
gb
balance
sin(k x)
across
the free surface yields
p1(x,0)
=kV A
sin(k x),
(10.108)
which leads to
gb
kV A.
(10.109)
Hence, from (10.105), (10.107), and (10.109),
cosh(kd) (g/kV 2) sinh(kd)

(10.110)
or
b
c = [(g/k) tanh(kd)] 1/2

gravity wave of wave number
where
It
a
cosh(kd)(1
c 2/V
2)
,
,
(10.111)
is the phase velocity of
k. See Equation (10.21).
follows that the peaks and troughs of the free surface
> |c|, and
coincide with those of the bottom when |V|

troughs
ce
co
<
|c|.
If |V|
implying that
are not
the
incide
coincide wwith
iththose
the peaks,
of the
and
bottom
vice when
versa,|V|
when
> |c|,|V|
a
|c| then the ratio b/a becomes infinite,
the oscillations
of small amplitude,
driven by the corrugations

and, therefore,
cannot
be
described by linear theory.
10.11
Surface Tension
As described
energy per
area,
as a
surface
surface tension into
4, there is
, associated
between two immiscible

interpreted
excess
an interface
fluids. The quantity can also be
tension. Let us now incorporate
in Chapter
unit
our analysis.
positive
with
Suppose that the interface
lies at
= (x,
t),
(10.112)
where
||
corresponds
is
small.
Thus,
to the plane
the
unperturbed
= 0. The
interface
unit normal to the
interface is
(z )
|(z )|
(10.113)
It follows that
nx
nz
(10.114)
(10.125)
1.
206
FLUID MECHANICS
Now, the Young-Laplace Equation yields
n,
(10.116)
where
pressure seen crossing

n. See Section
and (10.115), we have
is the jump
in
interface in the opposite direction to
However, from (10.114)
the
4.2.
(10.117)
Hence, Equation (10.116) gives
[p]
z=0+
z=0
(10.118)
This expression is the generalization
of (10.99) that takes
surface tension into account.

Suppose that the interface in question is that between
body of water, of density

atmosphere.
and
and depth d, and the
Let the unperturbed
z = d
occupy the
water lie between
= 0, and let the unperturbed
atmosphere
region z > 0. In the limit in which

upy
atmosphere
th ere gion
is neglected,
z > 0.
the pressure
takes the fixed value
p0
whereas the
the surface of the water is p0 g

the
pressure
perturbation
the density
pressure
+ p1|z=0
to
due
of the
in the atmosphere
the
just below
Here,
p 1 is
wave.
The relation (10.118) yields
g p1 |z=0
= (/z)z=0
x 2
air/water interface.
where is the perturbed
velocity potential of the water. Moreover, from (10.9), p1

(/t).
(10.119)
an
where is the surface tension at

However, /t
Hence, the above expression gives

??????z=0
?????z=0
??????z=0
t 2
z 2 x
(10.120)
This
relation,
which
is
generalization
(10.15), is the condition satisfied at
presence
of non-negligible
boundary
(which
bottom),
condition
already
we obtain
Equation
free surface in the
surface tension. Applying this
to the
satisfies
of
general
the boundary
the dispersion relation
solution,
condition
(10.19)
at the
which is
generalization
tension into account.
10.12
In the
gk +
tanh(kd),
(10.121)
of (10.21)
that takes surface
Capillary Waves
deep
water
limit
k d
1, the dispersion relation (10.121) simplifies to
gk +
(10.122)
It is helpful to introduce the capillary length,

!1/2
capillary
Roughly
wavelength is c
speaking,
(10.123)
(See Section 4.4.) The capillary
interface (See
at s.t.p.
is 4.4.)
2.7 The
10c
Section
length of
3
= 2l =
m.
an
air/water
The
associated
apillary
lengt
1.710
m.
waves whose wavelengths

are much larger than the capillary wavelength, and vice
versa. It is also helpful to introduce the critical phase
surface tension is negligible for
velocity
vc
(2gl)
1/2.
(10.124)
This
critical
the value
velocity
0.23 m/s for
takes
an
air/water
interface
It follows from (10.122)that the phase velocity,
of a surface water
at s.t.p.
= /k,
vp
wave can be written

"
vp
vc
(10.135)
Waves
in
!#1/2
1
2
Incompressible
kl+
1
kl
Fluids
207
Moreover,
velocity,
vg
the ratio of the phase velocity
=d/dk,
to the
group
becomes
"
vg
vp
1
+ 3(kl)
1
+ (kl) 2
(10.126)
In the long wavelength limit c
(i.e., kl 1),
we
obtain
vp
v0
(2kl) 1/2
(10.127)
and
vg
vp
(10.128)
We can identify this type of
wave
wave as the deep water
gravity
discussed in Section 10.3.
In the short wavelength limit c (i.e., kl 1),
we
get
!1/2
vp
kl
vc
(10.129)
and
vg
vp
,
.
(10.130)
This corresponds to a completely
as a capillary wave. Such waves

much
less
(10.129)
than
the
capillary
new type
wave
of
known
have wavelengths that

wavelength.
are
Moreover,
can be rewritten
!1/2
vp
(10.131)
which demonstrates
propagation
of
that gravity
capillary
plays
waveits
seen
tension. Finally, it is easily
surface
no
that the phase
velocity (10.125) attains the minimum value
=c
(i.e., when kl
role in the
place is taken by
= 1).Moreover,
v p = vc
when
vg =
from (10.126),
v p at this wavelength. It follows that the phase velocity
a surface wave propagating over a

never be less than the critical value, vc
of
body of water
can
,
,
,
Capillary Waves at an Interface
10.13
Consider
layer of fluid of density
uniform horizontal velocity V
situated
depth d
and
on top of a layer of
another fluid of density , depth d,and uniform horizontal

velocity
V. Suppose
above and
below
planes be at
that the fluids
are
the two fluids be at
interface between
sin(t
from
z = 0.Suppose
the elevation of the perturbed interface is z
bounded
by rigid horizontal planes. Let these

,
d and z
d
and let the unperturbed
that
=, where =
k x). Finally, let be the surface tension of
the interface. Equations

yield the dispersion
(
(10.97), (10.98), and (10.118)
relation
)g + k
k(V
c) 2
tanh(kd)
k(V
c) 2
tanh(kd )
(10.132)
which
is
generalization
of
the
dispersion
relation
(10.100) that takes surface tension into account. Here,

/k is the phase velocity of
a wave
propagating
c=
along the
interface.
For the
case
are at rest,
in which both fluids
and of
to
infinite depth, the above dispersion relation simplifies
give
(
g+
= (
+ )
c2
(10.133)
Suppose that
s = /
(i.e., the
upper
upper
s<1
is the specific gravity of the
fluid with respect to the lower. In the
case
in which
fluid is lighter than the lower one), it is
helpful to define
"
#1/2
g(1 s)
"2gl
(10.134)
c0
!#1/2
1
+s
(10.144) 208
FLUID MECHANICS
It follows that
c2
c 02
1 1
2
kl
velocity
of
+ kl
(10.136)
Thus,
we
conclude
that the phase
a wave
propagating
achieves
its
minimum
waves
Furthermore,
value,
of
all
between the two fluids
along the interface
c0
when
kl
are
wavelength
able
1.
to
c 2 > 0 for all k).In the

opposite case, in which s > 1(i.e., the upper fluid is
heavier than the lower one), we can redefine the capillary
length as
propagate along the interface (i.e.,
"
#1/2
g(s 1)
(10.137)
The dispersion relation (10.133) then becomes
!
=gl
s+1
1
kl
kl
(10.138)
c 2 < 0 for kl< 1, indicating instability

of the interface for waves whose wavelengths exceed the
critical value c = 2 l.On the other hand, waves whose
wavelengths are less than the critical value are stabilized
It is apparent that
by
surface
experiment
tension.
This
result
is exemplified
in which water is retained
by the
by atmospheric
pressure in an inverted glass whose mouth is closed by a

gauze of fine mesh (the purpose of which is to put an
upper limit on the wavelengths of waves that can exist at
the interface.)
10.14
Wind Driven Waves inDeep Water
Consider the scenario described

Suppose that the lower fluid is
rest, and the
upper
in the previous section.
body of deep water at
fluid is the atmosphere.
Let the air
above the surface of the water move horizontally at the

.
constant velocity V
Suppose that is the density of water,
s = /
the specific gravity of air with respect to water,
and the surface tension at
0, kd
kd
an air/water
interface. With V
the dispersion
relation (10.132)
reduces to
(1 s)g
This expression
= kc
+ k 2
+sk(V
2V
1
+s
which is a quadratic equation for the phase velocity,
=c
1
+s
gk
c,of the wave. Here,

!
c 12 =
c)
1
+s
+
(1
+s)
(10.139)
(10.140)
(10.141)
where c1 is the phase velocity that the
wave
would have in
the absence of the wind. In fact, we can write

!
2
c1
=
c 02
(10.142)
where c
are
= 2 lis the capillary
defined
in
Equations
wavelength, and l
and c0
(10.134)
respectively.
For
attains
given
, the
wavelength,
its maximum
value,
cm
and
wave
when
(10.135),
velocity,
c,
0.
dc/dV
occurs
According to the dispersion relation (10. 140), this
when
V
If the wind has
any
= cm = (1 + s)
According
or
less than
cm
wave velocity, c, becomes

an instability, when
to (10.140), the
complex, indicating
c1
(10.143)
other velocity, greater
then the wave velocity is less than cm
1/2
>
(1+ s)
2
2
c1
=
(1+ s)
2s
+
c
c 02
209
We conclude that if the wind speed exceeds the critical
value
Vc
(1+ s)
c0
s 1/2
= 6.6m/s =12.8kts
then waves whose wavelengths fall within
centered around c
a certain range,
tude.
The two roots of Equation (10.140)
c=
1
+s
are
"c
2
1
sV
#1/2
(1+ s) 2
Moreover, if
< (1+ s 1) 1/2
c1
(10.145)
are unstable
and grow to large ampli-
(10.146)
(10.147)
then these roots have opposite signs. Hence, the waves
can either
travel with the wind, or against it,but travel
faster when they
are moving
with the wind. If V
against the wind. Since c1 has the minimum value c0

follows that
waves
traveling against the wind
exceeds
the value given above then the waves cannot travel
are
it
completely ruled out when

V
> (1+s 1) 1/2
c0
=6.6m/s =12.8kts.
(10.148)
10.15
Exercises
10.1. Find the velocity potential of a standing gravity
wave
in
deep water for which the associated elevation of the free surface
is
z = a cos(t)
cos(k x).
Determine the paths of water particles perturbed by the
wave.
10.2. Deep water fills
a rectangular
tank of length l
and breadth
b.Show that the resonant frequencies of the water inthe tank
(g)
where
1/2
(n l
+ m2
nand mare integers.
1/4
are
You may neglect surface
tension.
a sinusoidal
gravity
with surface elevation
= a cos(t
k x)
wave on deep water
possesses amean momentum per unit surface area
1
a
2
10.4. A seismic
wave passes
along the bed of
anocean of
uniform depth dsuch that the vertical perturbation of the bed
is
a cos[k(x
V t)]. Show that the amplitude of the
consequent gravity
waves at the surface

"
c2
1
where
c is the phase
is
#1
cosh(kd)
velocity of waves of wavenumber k.
10.5. A layer of liquid of density and depth dhas

surface, and lies
over liquid
a free upper
of infinite depth and density
> .
Neglecting surface tension, show that two possible types

of
wave of wavenumber
k,with phase velocities
c2
c
kg,
kg( )
coth(kd)
can propagate
along the layer.
210
FLUID MECHANICS
10.6. Show that, taking surface tension into account,

sinusoidal
wave
of wavenumber k and surface amplitude
a has
amean
kinetic
,
.
energy per unit surface area

K
14
(g
+ k 2 )a 2
a mean potential energy per unit surface area
and
14
(g
+ k 2 )a 2
10.7. Show that inwater of uniform depth d the phase velocity

of surface
waves can only attain a stationary (i.e., maximum or

as a function of wavenumber, k,when
minimum) value
"
where kc
= (g/)
one
stationary
ust
than
capillary
kc
+2kd
Hence, deduce that the phase velocity has j
value (a minimum)
any
for
depth greater
kc
no
stationary values for depths less than that.
waves in deep water, whose group

are half their phase velocities, the group velocities of
waves are 3/2 times their phase velocities. Adapt the
Unlike
velocities
1/2
sinh(2kd)
1/2
4.8mm, but
10.8.
#1/2
sinh(2kd) 2kd
gravity
analysis of Section 10.7 to investigate the generation of capillary
waves
by
a very
small object traveling
water at the constant

surface
corresponds
minus x-direction,
origin.
to the
x-y
the surface of the
that the unperturbed
plane. Let the object travel in the
such that it is instantaneously
Find the present
traveling at
across
speed V. Suppose
position
of
an angle to the objects
waves
that
found at the
were
emitted,
direction of motion, when it
was
located
cos 3
at (X, 0). Show that along

X and
maximum the quantities

takes
a constant
value,
vary
X1
given interference
a manner
in such
(say).
interference maximum is given parametrically
Deduce
that X
that
the
by the equations
x
y
X1 sec
tan
12

32
X1 sec tan.
Sketch this curve, noting that it goes through the points (0.5X1
rding
to the adiabatic gas law,
Equilibrium ofCompressible Fluids
211
11Equilibrium of Compressible
Fluids
11.1
Introduction
we investigate the equilibria of

compressible fluids such as gases. As is the case for an
incompressible fluid (see Chapter 3), a compressible fluid
Inthis chapter,
inmechanical equilibrium must satisfy the force balance

equation
= p + ,
(11.1)

energy per unit mass. In
an ideal gas, the relationship between p and is
determined by the energy conservation equation, (2.89),
which can be written
!
!
D
p
M
p
2
where
is the static fluid
and the gravitational potential
Dt
(11.2)
Here,
is the
ratio
of
conductivity, Mthe molar
constant.
specific
mass,
heats,
the thermal
and R the molar ideal
gas
term has
Note that the viscous heat generation
been omitted from the above equation (since it is zero in
stationary
and
are
gas).
right-hand
dominant
limits,
are
The
limits
in
which
sides
of Equation
termed
the adiabatic
respectively.
In the
isothermal
the
left-
(11.2)
and
isothermal
limit,
in which
thermal transport is comparatively large, so that (11.2) can

2(p/)
only be satisfied when
0, the temperature
(recall that T p/ in an ideal gas) distribution in the
becomes
uniform,
consequently
=constant.
and
the
pressure
and
density
related according to the isothermal
gas
are
gas law,
(11.3)
On the other hand, inthe adiabatic limit, in which thermal
so that (11.2) can only be satisfied

pressure and density are
to the adiabatic gas law,
transport is negligible,
when D/Dt(p/
related according
0,the
=constant.
(11.4)
11.2
Isothermal Atmosphere
The vertical thickness of the atmosphere is only
a few tens
of kilometers, and is, therefore, much less than the radius

of the Earth, which is about 6000 km. Consequently, it is a
good approximation to treat the atmosphere
as a relatively
thin layer, covering the surface of the Earth, in which the
pressure
and density
ground level,
unit
mass
acceleration
z, and
takes
dz
= g.
only functions of altitude above
the form
due to gravity
Equation (11.1) that

dp
are
the gravitational
at
=
z
potential
g z,
where
= 0. It
energy per
g is the
follows
from
(11.5)
Now, in an isothermal atmosphere, in which the
temperature, T,is assumed not to vary with height, the

ideal gas equation of state (2.84) yields [cf., Equation (1
1.3)]
RT
(11.6)
dp
dz
gM
p.
RT
(11.7)
Hence, we obtain
p(z)
= p0
exp(z/H),
212
FLUID MECHANICS
where p0 10
level, and
RT
gM
Nm
is atmospheric
pressure at ground
(11.9)
is
known
as
the
isothermal
scale
atmosphere. Using the values T
g = 9.8 m s
29, and
2,
which
height
= 273 K (0
are
the
of
C), M
typical of the Earths
Mground29,
and
atmosphereC),(at
level),
as well
1
1
8.3 15 J mol
K
we find that H
as
R whi
= 7.99
km.
Equations (11.6) and (11.8) yield
(z)
=0
exp(z/H),
(11.10)
where
= p0 /(gH)
is the atmospheric
According
to Equations
isothermal
atmosphere,
mass
(11.8)
the
density at
and
pressure
(11.10),
and density
both
with increasing altitude, falling to
decrease exponentially
37% of their values at ground level when

only 5% of these values when z
11.3
z=0
in an
=3H.
= H, and to
Adiabatic Atmosphere
In fact, the temperature
of the Earths atmosphere
uniform, but instead decreases
alti-tude. This
effect
is largely
convection currents. When
steadily
due
is not
with increasing
to the action
of
packet of air ascends, under
higher altitudes
generally
diffuse
pressure
leads to
pressure at
of such currents, the diminished
the influence
takes
into
causes
it to expand. Since this expansion
place far
the
packet,
more
rapidly
the work
against
gas, as the packet

in its internal energy, and,
of the surrounding
a reduction
its temperature.
Assuming
that
can
than heat
done
the
expands,
hence, in
the atmosphere
is in
mixed state, whilst remaining in approximate
continually
vertical force balance (such

equilibrium),
a state is
and that the effect
known asa convective
of heat
conduction
is
negligible (because the mixing takes place too rapidly for
we
thermal diffusion to affect the temperature),
expect the adiabatic

accurate
gas
description
atmospheric
law, (11.3).
law, (11.4), to offer

of
the
would
much
relationship
, ,
pressure
more
between
, ,,
and density than the isothermal
gas
b = T/T0
p = p/p0 b = /0 and T
where p0 0
b
and T0 Let
are b p
the
= pressure,
p/p0 ,b =mass
/ 0 density,
and temperature of
Let respectively,
b p = p/p0 ,b =
at ground level. The adiabatic
the atmosphere,
gas law, (11.4), can be combined with the ideal gas
Let b
equation of state, (11.6), to give
=
b
Tb
/(1).
(11.11)
b
bp
The isothermal scale height of the atmosphere is
conveniently
redefined
as [cf., (11.9)]
RT0
p0
gM
g0
(11.12)
Equations (11.5), (11.11), and (11.12) yield
db p
= b p
1/
(11.13)
whereb
= z/H, or, from
(11.11),
d Tb
(11.14)
The above equation
T(z)
=T0
that
the
(11.15)
It
follows
atmosphere
temperature
the rate of [/( 1)](T0 /H) degrees

is known
in
an
adiabatic
decreases linearly with increasing altitude at
as
per meter. This rate
the adiabatic lapse rate of the atmosphere.
Using the values
1.4, T0
=273K, and H=7.99km,
which
are
typical of the Earths atmosphere,
the lapse rate
to be 9.8 Kkm
1.
we
estimate
In reality, the lapse rate
only takes this value in dry air. In moist air, the lapse rate
is considerably
when water
reduced because of the latent heat released
vapor
condenses.
213
Equations (11.11) and (11.15) yield
p(z)
p0
!/(1)
!1/(1)
(11.16)
(z)
,
.
(11.17)
Since /( 1)
pressure
adiabatic
3.5 and 1/( 1)
more
decreases
atmosphere.
equations imply that
upper
boundary
at
rapidly
Moreover,
an adiabatic
= [/(
2.5, it follows that
than
density
the
previous
atmosphere
1)] H
has
in
an
three
a sharp
28 km. At this
pressure, and density all fall to

zero. Of course, above this altitude, the temperature,
pressure, and density remain zero (since they cannot take
negative or imaginary values). In contrast, an isothermal
atmosphere has a diffuse upper boundary in which the
pressure and density never fall to zero, even at extreme
altitude, the temperature,
altitudes. It should be noted that, in reality, the Earths
a sharp upper
gas law does not hold at very
does not have
atmosphere
the adiabatic
11.4
boundary, since
high altitudes.
Atmospheric Stability
Suppose
that
the
non-convecting).
pressure
Consider
and density,
atmosphere
Moreover,
is
static
let p(z)
. ,
respectively,
(i.e.,
and (z)
as functions
be the
of altitude.
packet of air that is in equilibrium
with the
surrounding air at some initial altitude z1 but subsequently
moves to a higher altitude

pressure and density are
(z1 ), respectively.
z2
Thus, the packets initial
p1
at
Now,
p(z1) and 1
the
the packet must adjust its volume in such
pressure
matches
there would
be
that of the surrounding
force
imbalance
boundary. It follows that the packet
z2
is p2
higher
a manner
= p(z2 ). Assuming
that
altitude,
that its
air, otherwise
across the packet

pressure at altitude
the packet moves
on a much faster time-scale than that needed for

heat to diffuse across it (but still a sufficiently
slow
time-scale that it remains in approximate pressure balance
with the surrounding air), we would expect its internal
pressure and density to be related according to the
adiabatic gas law, (11.4). Thus, the packets density at
upward
altitude z2 is 2
= (p2 /p1)
1/
Now, if 2
> (z2
)then
the packet is denser than the surrounding
air. It follows
buoyancy due to the
that the packet weight exceeds the
atmosphere, causing the packet to sink back to its original

altitude. On the otherhand, if 2
less dense than the surrounding
buoyancy
rise to
< (z2
) then the packet is
air. It follows that the
force exceeds the packet weight,
an even
atmosphere
is
[p(z2 )/p(z1 )]
1/
higher
unstable
altitude.
to vertical
(z1 )< (z2 )for any z2
any
z2
to
> z1
?????z2
monotonically
<
?????z1
(11.18)
It follows that the atmosphere
vertical
when
> z1:i.e., when
for
words, the
convection
stable
causing it to
In other
convection
when
p/
is only
is
decreasing function of altitude. As is easily
demonstrated, this stability criterion
can also be written

dln p
<
dln
(11.19)
or, making use of the ideal gas equation
of state,
dlnT
dln
<
1.
(11.20)
Convection
is triggered
in regions
of the atmosphere
where the above stability criterion is violated. However,

such
convection
acts to relax these regions back to
marginally stable state in which p/

adiabatic
11.5
is uniform: i.e., an
equilibrium.
Eddington Solar Model
us investigate the internal structure of the Sun, which is

a self-gravitating sphere of incandescent ionized
gas (consisting mostly of hydrogen). Adopting a spherical
Let
basically
niversity Press, Cambridge UK, 1926).
214
FLUID MECHANICS
coincides with the Suns geometric center, and making the

simplifying (and highly accurate) assumption that the mass
distribution within the Sun is spherically symmetric,
we
find that
dm
dr
= 4r
(11.21)
where m(r) is the total
mass
contained within
a sphere
of
r,centered on the origin, and (r) the mass density

r. Now, as is well-known, the gravitational
acceleration at some radius r in a spherically symmetric
mass distribution is the same as would be obtained were
all the mass located within this radius concentrated at the
center, and the remainder of the mass neglected. In other
radius
at radius
words,
d
Gm
r2
dr
(11.22)
where (r) is the gravitational potential
mass,
energy per
unit
and d/dr the radial gravitational acceleration.
The force balance criterion (11.1) yields

dp
dr
+
dr
=0,
(11.23)
where p(r) is the pressure. The previous three equations

1 d
r2
dr
(11.24)
r2
dp
dr
= 4G.
In order to make
determine
pressure
any
the relationship
progress, we
further
between
and density. Unfortunately,
the Suns
need to
internal
this relationship
is
energy transport, which is a very

process in a star. In fact, a stars energy is
ultimately controlled by
complicated
ultimately
derived from nuclear reactions occurring deep
within its core, the details of which

This
energy
boundary
via
involved.
the outer
combination of convection and radiation.
plays
(Conduction
much less important
an exact
process.) Unfortunately,
transport
are extremely
core to
is then transported from the
requires
an
role in this
calculation of radiative
understanding
of the opacity
of
stellar material, which is an exceptionally difficult subject

Finally,
once
it is radiated
to Eddington,
energy reaches the boundary of the star

away. The following ingenious model, due
1
is appropriate to a star whose internal
the
energy transport is dominated by radiation. This turns out

to be a fairly good approximation for the Sun. The main
model is that it does not require
advantage of Eddingtons
us to
know anything
about stellar nuclear
reactions
or
opacity.
Now, the temperature
large
that
neglected
with respect
other words,
the form
radiation
we must
inside the Sun is sufficiently
pressure
cannot
to conventional
be
completely
gas pressure.
In
write the solar equation of state in
p = pg +pr
11.25)
where
is the gas
pressure
pg =
13
T
(11.27)
the radiation
(modeling the plasma within the Sun as
pressure
and ions), and
(assuming that the radiation within
the Sun is everywhere in local thermodynamic

with
the
mp
(11.26)
an ideal gas of free electrons
pr =
kT
plasma).
Here,
T(r)
is
the
equilibrium
Suns
internal
m p the mass of a
proton, and
the relative molecular mass (i.e., the ratio of
the mean mass of the free particles making up the solar
plasma to that of a proton). Note that the electron mass has
been neglected
with respect to that of a proton.
temperature, k the Boltzmann constant,
Furthermore,
Stefan-Boltzmann
4/c,
where
constant,
is
the
and
the
velocity
of
light
in
a vacuum.
Incidentally, inwriting (11.26), we have expressed M/R in
the equivalent form
mp /k.
A.S. Eddington, The Internal Constitution
University Press, Cambridge UK, 1926).
215
of the Stars (Cambridge
0
0
Figure 11.1: The functions () (solid) and y() (dashed).
Let
pg
(1 )
pr
(11.28)
p,
p,
(11.29)
where the parameter is assumed to be uniform. In other

words,
the ratio
pressure
pressure to
of the radiation
is assumed to be the
same
Sun. This fairly drastic assumption
somewhat
fortuitously
gas
the
everywhere inside the
turns outperhaps,
to lead to approximately
the
correct internal pressure-density relation for the Sun. In

fact, Equa-tions (11.26)(11.29)can
be combined to give
where

It can be
seen, by comparison
the above pressure-density

adiabatic
gas
law with
= 4/3. Note, however,

for
!4
mp
4/3
(11.30)
1/3
(1 ) 4
(11.31)
with Equation (11.4), that
relation takes the form of
an effective
an
ratio of specific heats
that the actual ratio of specific heats
fully ionized hydrogen plasma, in the absence of
= 5/3. Hence, the 4/3 exponent,
appearing
1
.30), is entirely due to the non-negligible
in (1
radiation
radiation, is
pressure
within the Sun.
Let Tc
= T(0), c = (0), and pc = p(0), be the Suns
central temperature, density, and
pressure,
respectively. It
follows from (11.30) that
pc
4/3
K c
(11.32)
and from (11.26) and (11.28) that
Tc
pc m p
(1 ).
c
(11.33)
2
L.Mestel, Phys. Reports 311, 295 (1999).
216
FLUID MECHANICS
)
]
T
(
0
1
g
o
l
0.10.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
m/M
Figure
11.2: Solar temperature
obtainedfrom
versus mass fraction
the Eddington Solar Model (solid) and the
Solar Model (dashed).
Standard
Suppose that
Tc
(11.34)
where
is
dimensionless
function.
According to Equations (11.26), (11.28), and (11.30),
p
(11.35)
pc
(11.36)
,
.
Moreover, it is clear, from the above expressions, that
1
at the center of the Sun, r
= 0, and = 0 at the edge,
= R(say), where the temperature,
density, and
r
pressure are
allassumed to fall to zero. Suppose, finally, that
= a ,
(1
1.37)
where is a dimensionless radial coordinate, and
!1/2
a=
2/3
G c
(11.38)
Thus, the center of the Sun corresponds to
edge to
= 1
=1
= 0, and the
=0,and
(say), where (1 )
a.
(1
1.39)
Equations (11.35)(1 1.38)can be used to transform the

equilibriumrelation(11.24)into the non-dimensionalform
!
1 d
(11.46) Equilibrium ofCompressible Fluids
217
d
d
)
]
m
g
k
[
0
1
g
o
l
1
0
0.10.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
m/M
Figure 11.3: Solar
mass
density
versus mass
fraction
obtained from the Eddington Solar Model (solid) and the

Stan-dard Solar Model (dashed).
Moreover, Equation (11.21)
can be integrated,
with the aid
of Equations (1
1.35), (1
1.37), and (11.40), and the physical
boundary condition m(0)
=0,to give
m
4c
y,
(11.41)
where
y()
d
d
(11.42)
Equation (11.40) is known
(of degree
3), and
can,
as
the Lane-Emden
unfortunately,
only
equation
be solved
The appropriate solution takes the form

1
2/6

+ O( 4) when 1, and must be integrated to
numerically.
= 1
= 0. Figure
where (1 )
related
to
1 function
Note that
obtained
11.1shows (), and the

via numerical
methods.
= 6.897, and y1 =y(1 )=2.018.
According
m(R),
y(),
to Equation (11.41), the solar
mass,
can be written
=4c
a3
y1
(11.43)
which reduces, with the aid of Equations (11.31) and (11
.38), to
(1 )
(11.44)
where
M
M0
(11.45)
and

M0
1/2
!4
1
(G)
(11.51) 218
FLUID MECHANICS
mp
4y1
= 3.586 10
31
kg.
17
16
15
)
]
14
N
[
p
(
13
0
1
g
o
l
12
11
10
0
0.10.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
m/M
Figure 11.4: Solar
pressure versus mass
fraction
obtained from the Eddington Solar Model (solid) and the

Standard
Solar Model (dashed).
Moreover, it is easily demonstrated that
G M
2/3
(4y1 )2/3
(11.47)
to Equations
According
(11.44)
/(1 ), of the radiation
star is
radiative
stellar
in a
mass,
gas pressure
the
in
a strongly increasing function of the

mean molecular weight,
In the case
can be inverted to give
pressure to
M, and
of the Sun, for which
and (11.45), the ratio,
+ O(
1, Equation (11.44)
the approximate solution
6).
(11.48)
Using the observed

and the value
solar
mass
= 1.989
= 0.68 (which represents
10
30
kg,
the best fit to
the Standard Solar Model mentioned below), we find that

4.
6.58
10
In other words, the radiation pressure
inside
the Sun is only
a very
pressure.
This immediately
transport
is
transport.
convection
much less efficient
Indeed,
occurs
in regions
we
would
gas
energy
energy
small fraction of the
implies that radiative

than convective
of the
Sun
expect
the
in which
convective
transport to overwhelm the radiative transport, and
so to
drive
the
adiabatic
convection
local
law
relation
pressure-density
with
an exponent
a minuscule
Equations
(11.31),
(11.32),
(11.39),
GM m p
(1 )
4y1
Rk
(11.49)
R3
4y1
pc
(11.50)
GM
16y
(11.43),
2
1
2.5
1
(
= 1.34
10
= 7.63 10
= 1.24
16
219
regions,
fraction of its mass.
(11.47) yield
Tc
Fortunately,
only takes place in the Suns outer
which contain
an
toward
5/3.
10
and
kgm
Nm
K,
3,
2,
0
1
g
o
3.5
4
0

0.2
0.4
0.6
0.8
m/M
Figure 11.5: Ratio of radiation
pressure to gas pressure
calculated from the Eddington Solar Model (solid) and

the Standard Solar Model (dashed).
where the solar radius R has been given the observed

value 6.960
10
erv
ed v temperature,
alue 6.96 0
central
by the so-called
incorporates
m.
The actual values of the Suns
10
density,
and
Standard
pressure, as
determined
Solar Model (SSM),
which
detailed treatments of nuclear reactions and
are Tc
= 1.58 10 7 K,c = 15.6
3,
16
2,
10
Nm
respectively. It can
10 kg m
and pc 2.38
=
be seen that the values
2.3of
8 Tc10 c and pc obtained from
the Eddington model lie within a factor of two of those
obtained from the much more accurate SSM. Figures
opacity,
4
,,
11.2,
11.3, and
pressure
11.4 show the temperature, density, and
profiles, respectively,
obtained from the SSM
model. The profiles are plotted as

mass fraction, m(r)/M = y()/y1 where
can be seen that there is fairly good
and the Eddington

functions of the
r/a. It
agreement
between
the profiles
calculated
by the two
compares the ratio, /(1

pressure to the gas pressure obtained
models. Finally, Figure 11.5
),
of the radiation
from
/(1
the SSM and the Eddington
model. Recall, that
it is
fundamental assumption of the Eddington model that this
pressure
be seen
ratio is uniform throughout the Sun. In fact, it can

that the
not spatially
pressure
uniform.
ratio calculated by the SSM is
On the other hand, the spatial
variation of the ratio is fairly weak, except close to the

edge of the Sun, where
Eddington
convection
model, thus, becomes
sets in, and the
invalid. We conclude
that, despite its simplicity, the Eddington solar model does
a remarkably
good job of accounting for the Suns internal
structure.
11.6
Exercises
11.1. Prove that the fraction of the whole
mass
of an isothermal
atmosphere which isincluded between the ground and

horizontal
plane of height
z is
e z/H
Evaluate this for z
=H,2H, 3H, respectively.
11.2. If the absolute temperature inthe atmosphere diminishes

upwards according to the law
T
T0
3
4
= 1
http://en.wikipedia.org/wiki/Standard\_Solar\_Model
#1/2
The SSM
data is obtained from http://www.ap.stmarys.ca/
(n1)/n
c
220
4G
FLUID MECHANICS
where
c is a constant,
?1
pressure
show that the
?c/H
p0
varies
as
11.3. If the absolute temperature inthe atmosphere diminishes

upward according to the law
T
1
+ z
T0
where is a constant, show that the pressure varies
p
p0
= exp
12 z
as
11.4. Show that if the absolute temperature, T,inthe atmosphere

is any given function of the altitude,
z,then the vertical
distribu-
tion of pressure inthe atmosphere isgiven by

Z
p
p0
11.5. Show that if the Earth
T0
ln
dz
T
were surrounded by an atmosphere

a distance r from the
of uniform temperature then the pressure

Earths center would be "
p0
where
a2
= exp
a is the Earths
!#
1r
1a
radius.
11.6. Show that if the whole of
space were
occupied by air at
the uniform temperature T then the densities at the surfaces of
the
various planets would be proportional to the
corresponding values of
exp
gMa
RT
where
a is the radius
of the planet, and
g its surface
gravitational acceleration.
11.7. Prove that inan atmosphere arranged inhorizontal strata

the work (per unit mass) required to interchange two thin strata
of
equal
mass
without disturbance of the remaining strata is
1
1
?
1
2
p1
p2
where the suffixes refer to the initial states of the two
strata. Hence, show that for stability the ratio p/
must
increase
upwards.
11.8. A spherically symmetric star is such that m(r) is the mass
r.Show that the stars total gravitational

energy can be written inthe following three
contained within radius

potential
alternative forms:
Z
=
0
Gm
dm
dm = 3
pdV
0
mass, R the radius, (r) the

per unit mass (defined such that
Here, Mis the total

gravitational potential
),
Here, M is the total
mass, R the
p(r) the pressure, and dV
= 4r
dr.
11.9. Suppose that the pressure and density inside

symmetric star
are related
nis termed the polytropic index. Let = c n

mass density. Demonstrate that satisfies
where c isthe central
the
Lane-Emden equation
1 d
where
a spherically
gas law,
according to the polytropic
p = K (1+n)/n,
where
0 as
radi
r = a,
= n
and
"
(n + 1) K
4G
#1/2
(n1)/n
221
Show that the physical solution to the Lane-Emden
equation, which is such that (0)
= 1and (1 )= 0,for some 1
>
0,is
= 1
for n
=0,
sin
for n
= 1,and
(1
+ 2 /3) 1/2
=5.Determine
for n
the ratio of the central density to
the mean density inall three
cases. Finally,
demonstrate that, in
the
general
case, the total gravitational
potential
energy can
be written
GM
=a1
where Mis the total mass, and R
the radius.
11.10. A spherically symmetric star of radius Rhas

density of the form
(r)
= c
(1 r/R).
a mass
Show that the central
mass
density is four times the mean
density. Demonstrate that the central
pc =
pressure
GM
is
where Mis the mass of the star. Finally, show that the
total gravitational potential
energy
of the star can be written
26 G M
35
222
FLUID MECHANICS
Vectors andVector Fields

223
A Vectors and Vector Fields

A.1
Introduction
This Appendix outlines those aspects of vector algebra,
vector calculus, and vector field theory that

the study of fluid dynamics.
A.2
Scalars and Vectors
are helpful in
mass,
Many physical entities (e.g.,
directional
numerical
units).
appropriate
since
by
defined
element,
scalars
can
are
be
entities,
However, there exits

velocity,
second
acceleration,
completely defined when both
as
known
represented
entirely
(expressed
which
scalars
in
no
have
Moreover,
by real numbers
that they obey the laws of ordinary
follows
(e.g.,
Such
are
energy)
magnitude
it
algebra.
class of physical entities

force)
that
a numerical
are
only
magnitude and
in space are specified. Such entities are

as vectors. By definition, a vector obeys
direction
known
as a displacement in space, and

may thus
be represented geometrically by a straight-line, PQ (say),
where
the arrow
indicates
the direction
of the
the
same
algebra
displacement
A. 1
The
(i.e., from point P to point Q). See Figure
magnitude
of the vector is represented
by the
length of the straight-line.
It is conventional
to denote
vectors
by bold-faced
symbols (e.g., a, F) and scalars by non-bold-faced symbols

magnitude of a general vector, a, is
or r,
just
S).
a,Tand
he m
is,
agnitude
by definition,
of a genera
always
l
vgreater
ector,
than or equal to zero. It is convenient to define a vector
with zero magnitudethis
is denoted 0, and has no
direction. Finally, two vectors, a and b,are said to be equal
when their magnitudes and directions are both identical.
(e.g.,
r,
S). The
denoted |a|,
A.3
Vector Algebra
Suppose that the displacements
vectors
a and b,respectively.
PQ and QR represent the
See Figure A.2. It can be
seen
that the result of combining these two displacements

is to
give the net displacement
vector
PR. Hence, if PR represents the
c then we can write

c =a + b.
(A.1)
This defines vector addition. By completing the

parallelogram PQRS,
we can also see that
PR= PQ + QR= PS + SR
(A.2)
same length and direction as

QR,
and,
thus, represents the same vector, b.Likewise, PQ and
SR both represent the vector a.Thus, the above equation is

However,
PS has the
equivalent to
c =a + b =b + a.
(A.3)
Figure A.1: A vector.

224
FLUIDMECHANICS
c =a +b
S
a
b
Figure A.2: Vector addition.
c =a b
c
b
Figure A.3: Vector subtraction.
We conclude that the addition of vectors is commutative. It
can also be shown
that the associative law holds: i.e.,
a +
(b
+ c)
= (a
b)
+ c.
(A.4)
The null vector, 0, is represented

of
zero
of
combining
by
displacement
length and arbitrary direction. Since the result
displacement
such
is the
a displacement with a finite length

same as the latter displacement by
itself, it follows that
a +
= a,
(A.5)
a is a general vector. The negative of a is defined as

that vector which has the same magnitude, but acts in the
opposite direction, and is denoted a. The sum of a and
where
a is thus the null vector: i.e.,
a +
(a)
0.
(A.6)
can also
We
define the difference of two vectors,
a and
b,
as
c
=a
=a
(b).
(A.7)
This definition of vector subtraction is illustrated in Figure

A.3.
If
n > 0 is a
vector whose
na denotes a
same as a, and whose
scalar then the expression
direction
is the
n times that of a. (This definition becomes

nis
is an
n times
integer.)
thatIf of
n isa negative
( T histhen,
defi nition
since n
b
a = |n| (a), it follows that n a is a vector whose
magnitude is |n| times that of a, and whose direction is
magnitude
is
obvious when
(A.16)
225
Vectors
and
Vector
Fields
Figure A.4: A right-handed Cartesian coordinate system.
definitions imply that if n and mare two scalars then
n(ma)
(A.8) (n + m)a
(A.9) n(a
+ b)
nma
=
=
=m(na),
na + ma,
na + nb.
(A.10)
A.4
Cartesian Components of a Vector
Consider a Cartesian coordinate system Oxyz consisting of

an origin, O, and three mutually perpendicular coordinate
axes, Ox, Oy, and Oz. See Figure A.4. Such a system
is said to be right-handed if,when looking along the Oz

direction,
a 90
clockwise rotation about Oz is required to
take Ox into Oy. Otherwise, it is said to be left-handed. It
to always
is conventional
use a
right-handed
system.
It is convenient to define unit vectors,
ex
,, ,
coordinate
ey
and
parallel to Ox, Oy, and Oz, respectively. Incidentally,
ez
a
a vector whose magnitude is unity. The

r, of some general point P whose
coordinates are (x, y,z) is then given by
r = xez + yey + zez
unit vector is
position
Cartesian
vector,
(A.11)
we can get from O to P by moving a

x parallel to Ox, then a distance y parallel to Oy,
then a distance z parallel to Oz. Similarly, if a is an
In other words,
distance
and
arbitrary vector then
where
,,
ax ay
and
az are
= ax
(ax
ay az ).It follows
ez (0, 0,
1, 0), and
that
According
,,
to the three-dimensional
the Pythagorean
(A.12)
termed the Cartesian components
a. It is coventional to write a
ex (1, 0, 0), ey (0,
1). Of course, 0 (0, 0, 0).
of
ex + ay ey + az ez
theorem,
generalization
OP |r|
the distance
given by
of
= r is
qx
r
+y 2
+z 2
(A.13)
By analogy, the magnitude of
a general vector a takes
the
form
If
(ax
,,
+ a y2 + a z2
,,
, ,
,,
x
(A.14)
ay az ) and
b (bx
demonstrated that
a +
qa
a=
b (ax
by
bx
bz )then it is easily
ay +
az +
by
bz ).
(A.15)
Furthermore, if nis a scalar then it is apparent that
na
under 226
FLUID MECHANICS
(nax
nay
naz ).
y
y
Figure A.5: Rotation of the coordinate
axes
about Oz.
A.5
A
Coordinate Transformations
system
coordinate
Cartesian
allows
direction in space to be represented
manner.
such
Unfortunately,
orientations
and
convenient
system also
analysis. After all,
might well choose Cartesian
systems with different origins, and different
coordinate
vector
observers
our
position
very
coordinate
introduces arbitrary elements into
two independent
in a
of the
coordinate
will have different
two coordinate
systems.
In general,
of
However,
same
the
given
in these
direction
and
cases. Hence, the

two sets of components must be related to one another in a
very particular fashion. Actually, since vectors are
magnitude
a are
axes.
sets of components
the
in both
represented by moveable line elements
in space (i.e., in Figure A.2, PQ and SR represent the

vector), it follows that the components of
are not
affected
Cartesian
components
by
coordinate
are
simple shift in
system.
modified
same
a general vector
the origin of a
On the other
hand, the
axes are
when the coordinate
rotated.
Suppose that we transform to a new coordinate system,

z ,
which has the same origin as Oxyz, and is obtained
Ox y
by rotating the coordinate
axes
of Oxyz through
an angle
a general
about Oz. See Figure A.5. Let the coordinates of
point P be (x,
to
According
of
y, z)
in Oxyz and (x
simple
trigonometry,
are
related
xcos
+ ysin,
coordinates
z ) in
Ox y
two
these
one
to
another
via
z .
sets
the
transformation
(A.17) y
(A.18)
xsin
+ ycos,
z.
(A.19)
this displacement
in
is represented
both coordinate
direction
of
systems,
r are
OP. Note that
by the
since
same
symbol,
the magnitude
independent
manifestly
orientation of the coordinate
axes.
of
The coordinates
r,
and
the
of
on the
they must depend in a very specific

(A.17)(A.19)]
which
preserves
manner
the
[i.e., Equations
magnitude
and
direction of r.
components
The
in
transform
(A.17)(A.19):
ax
orientation of the axes: i.e., r (x,

,
, )
z) in Oxyz, and r (x
y
z in Ox y z . However,
do depend
y,
the vector displacement
Consider
(A.20)
=
ay
an
of
vector
general
analogous
manner
i.e.,
ax
cos
+ ay
sin,
ax sin
+ ay
cos,
to
a
Equations
(A.21)
az
az
(A.22)
Moreover,
rotation
are
there
about
,,
rules
transformation
Oy. Equations
for
(A.20)(A.22)
the definition of a vector: i.e., the

ay az )are the components of a vector
constitute
effectively
three quantities (ax

provided
they
that
axes
coordinate
(And
about
under
transform
rotation
about Oz in accordance
(A.20)(A.22).
rotation
similar
and
Ox
also
Ox and Oy). Conversely,
cannot be the components
of
,,
correctly
transform
a vector
the
of
with Equations
(ax
under
ay
az )
if they do not
transform in accordance with Equations (A.20)(A.22).
cos,
rotation
of the
components of
are not
a vector
scalars.
coordinate
(ax
automatically
however,
physical
but which
Scalar Product
that
There
have
are not
are,
both
obviously
We need to check carefully
these quantities
A.7 and A.9).
A.6
and all vectors
velocity, acceleration),
quantities
and direction,
related to displacements.
see whether
vectors,
(e.g.
satisfy Equations (A.20)(A.22).
other
magnitude
227
axes. Thus, the individual

say) are real numbers, but they
Displacement
derived from displacements
Of
A.29)
are
to
really vectors (see Sections
A scalar quantity is invariant under allpossible rotational

transformations.
are
not
The individual components
scalars
transformation.
because
Can
we
they
Suppose that
% b
we were to define
ax
bz
+ ay
bx
of
a vector
under
a scalar out of some

one, or more, vectors?
form
combination of the components
change
of
the percent product,
+ az
= scalar
by
number,
(A.23)
for general vectors
a and
as must
transformation,
be the
a % b invariant
case if it is a
Let
consider
an example.
under
scalar
a
seen that2,
a%b
=
1. Let us now rotate
the0).
coordinate axes througha%b
45
(1, 0,
2 (0, 1, 0 )and b
2,
2, It is easily seen that
2,
about
In the new coordinate system, a (1/
1/
= 1.Oz.
Let
number?
us
b. Is
(0, 1, 0) and b (1, 0, 0).
2,
2,
2,
0) and
b (1/
Suppose that
It is easily
0),giving a%b
1/
Clearly,
a %ate
bsyis
new coordin
stem, a (1/
1/2.
b (
1/
not invariant
1/
under rotational transformation,
so the
above
definition is a bad one.

Consider,
now, the dot product or scalar

ax bx + ay by + az bz =
product:
scalar number.
(A.24)
Let
us rotate
the coordinate
Oz.Let
According
u
to Equations
axes
though degrees about
(A.20)(A.22),
b takes
the form
a b
(ax cos
+ ay
sin)(bx
cos
+ by
sin)
+(ax
sin
+ ay
cos)(bx
sin
+ by
ax
cos)
bx
+ az
+ ay
by
bz
+ az
bz
(A.25)
in the
new
coordinate
system. Thus,
under rotation about Oz. It
can
b is invariant
easily be shown that it is
also invariant
variant
un der rounder
ta tionrotation
ab out Oz.
about
It can
Ox easil
andy be
Oy.
show
We
conclude
that it is that
als o a
a true scalar, and that the above

one. Incidentally, a b is the only
he
simple
abo ve
combination
definitio nisof
a the components of two vectors
which transforms like a scalar. It is readily shown that the
definition
is
b is
good
dot product is commutative and distributive: i.e.,
a b = b a,
a (b + c)
a b + a c.
(A.26)
The
associative
product, because
property
we cannot
is
meaningless
have (a
b)
the
dot
c, since a
b is
for
scalar.
We have shown that the dot product
independent.
But what
this? Well, inthe special
a b is
is the geometric
coordinate
significance
case where a =b,we get

= a x2 + a y2 + az 2 = |a| 2
of
=a
(A.27)
a a is equivalent to the invariance of

a under transformation.
us now investigate the general case. The length
So, the invariance of
the magnitude of vector

Let
squared of AB in the vector triangle shown in Figure A.6 is
(b a)
(b a)
= |a|
|b|
b.
(A.28)
However, according to the cosine rule of trigonometry,
(AB)
se,for
(OA)
(OB)
2 (OA) (OB)
cos
228
FLUID MECHANICS
ba
Figure A.6: A vector triangle.
where (AB) denotes the length of side AB. It follows that
= |a||b| cos.
(A.30)
In this
case,
the invariance of
a b under
transformation
is
equivalentIn tto
the invariance
of of
the
angle
his case,
the invari ance
a b
un de rsubtended
tra nsform
between
|a|
the two vectors. Note that if
= 0, |b| = 0, or
perpendicular.
can easily
cos
the vectors
The angle subtended
= 0 then either
and b
are
mutually
between two vectors
be obtained from the dot product: i.e.,
|a||b|
(A.31)
The work W performed by a constant force Fwhich
moves an object
through
a displacement ris the product
of the magnitude of Ftimes the displacement
inthe
direction of F.If the angle subtended between Fand ris

then
W
= |F|(|r| cos) = F r.
(A.32)
The work dW performed

which
moves
displacement
an
dr in
object
by
a non-constant force f
an infinitesimal
through
time interval dt is dW
Thus, the rate at which the force does work

which is usually
dr/dt,
or
referred to as the
=f
v,
instantaneous velocity.
where
power,
is P
= dr/dt
=f
on the
dr.
object,
= dW/dt = f
is the objects
A.7
,.
Vector Area
we have planar surface of scalar area S

a vector area S whose magnitude is S
Suppose that
can
define
whose
direction
sense
determined
to the plane, in the
is perpendicular
by
We
and
right-hand
circulation
Section A.8) applied to the rim, assuming that
rule (see
a direction
of circulation around the rim is specified. See Figure A.7.
possesses both magnitude and

a true vector? We know that if the
normal to the surface makes an angle x with the x-axis
then the area seen looking along the x-direction is S cos
This
quantity
clearly
direction. But is it
.
,
.
This
(since S x
= cos x
is the
x-component
= ex
where
n is
ex
area seen
the unit normal to the surface).
Similarly, if the normal makes

then the
of
nS
an
angle y with the y-axis
looking along the y-direction is S
This is the y-component of S. If
we limit
surface whose normal is perpendicular
then x
0). If
= /2 y = . It follows
we rotate
= S (cos , sin,
to rotating the normal
about the z-axis by degrees,
= S cos(
to the z-direction
that S
the basis about the z-axis by degrees,
which is equivalent
S x
cos
ourselves to a
=S
so that
to the surface
, then
cos cos
+ S sin
sin
=S x
cos
+S y
sin,
(A.33)
which
is
the
x-component
of
correct transformation
rule for the
a vector. The other components
transform correctly as well. This proves both that a vector

area is a true vector and that the components of a vector
area are the projected areas seen looking down the
coordinate axes.
According to the vector addition theorem, the projected
area of two plane surfaces,

along the x-direction
resultant
of
the
joined together at a line, looking
(say) is the x-component
vector
areas
of
the
(A.38) Vectors andVector Fields
229
Figure A.7: A vector
area.
two
of the
surfaces.
many
joined-up
plane
areas,
down the x-axis, which is the
seen
rim
looking
is the x-component
down
the net area seen looking

same as the area of the outer
the
x-axis,
areas:
of the resultant of all the vector
i.e.,
S
Xi
Si
(A.34)
If
we
approach
limit, by letting the number of plane
facets increase, and their
areas
reduce, then
we
contin-uous surface denoted by the resultant vector
It is clear that the

x-axis is just S x
area
obtain
area
Xi
Si
(A.35)
of the rim seen looking down the
areas of the rim seen

axes. Note that it is
determines the vector area
Similarly, for the
looking down the other coordinate
the rim of the surface that

rather
than the nature of the surface spanning the rim. So,
two different surfaces sharing the
area.
In conclusion, a loop
same
rim both
possess
the same vector
(not all in one plane) has
a vector
area S which is the resultant of the component vector areas

of any surface ending on the loop. The components of S
are the areas of the loop seen looking down the
coordinate axes. As a corollary, a closed surface has S = 0,
since it does not possess a rim.
A.8
Vector Product
We have discovered
components
also
how to construct
of two general vectors,
construct
a vector
which
is
a scalar from the

a and b. Can we
not just a linear
, ,
combination of a and b? Consider the following definition:
b (ax
bx
ay
by
az
bz ).
(A.36)
= (0, 1, 0),
b =
2,
(1, 0,Is
0).
In
this
case,
a
b
=
0.
However,
if
we
rotate
a b a proper
b
2
2, vector?
2 Suppose that a = (0, 1, 0),
2,
the
coordinate
axes
through45
about Oz then a = (1/
= (1,
0,
0
).
In
thi
sca
2,
2,
1/
0),
e,a 2,
b=
0 b
Ho=
we(1/
ver,if we rotate the
1/
0), and a b = (1/2, 1/2, 0). Thus, a bdoes not
1/
transform0),
like
andaab
vector,
= (1/2,
because
1/2,
its 0).
magnitude
Thus, ab
depends
d
on
Is
a proper
the choice of
Consider,
b (ay
vector? Suppose that
axes. So, above definition is a bad one.

now, the cross product or vector product:
bz az by az bx ax bz ax by ay bx ) = c.
Does this rather unlikely
combination
(A.37)
transform
like
vector? Let
angle
new
us try
rotating the coordinate
about Oz using Equations
axes
through
(A.20)(A.22).
an
In the
coordinate system,
cx
(ax
sin
+ ay
(ay bz
cos)bz
az
by )cos
az
(bx
+ (az
=
sin
bx
ax
cx
+ by
cos)
bz )sin
cos
+ cy
sin.
angle): 230
FLUID MECHANICS
thumb
middle finger
index finger
Figure A.8: The right-hand rule for cross products. Here,
is less that 180
Thus, the x-component
can easily
of
b transforms
correctly. It
be shown that the other components
transform
correctly
as
well, and that all components
also transform
a b is
a proper vector. Incidentally, a
b is the only simple
Thus,
combination
a b is aof the components of two vectors that
transforms like a vector (which is non-coplanar with a and
b).The cross product is anticommutative,
correctly under rotation about Ox and Oy. Thus,
= b
a,
(A.39)
distributive,
(b
c)
=a
but is not associative,
cross
The
means
it must
that
magnitude.
product
We
can
and b. Consider
product must
a a
ax (ay
bz
transforms
have
c)
like
+a
(a
b)
b. If this is
zero
which
direction
show that abis perpendicular
a a
az
0.
ax
then the
bz )+ az (ax by
ay
and
to both
be perpendicular to a.Now,
by )+ ay (az bx
c.
(A.41)
a vector,
well-defined
c,
(A.40)
(A.42)
(b
bx )
cross
Therefore,
b is perpendicular
,, .
demonstrated
that
to a. Likewise, it
can be
to b. The vectors
b is perpendicular
a, b,and a b form a right-handed set, like the unit

vectors a,
ex b,
ey and
and
fact,
ey =ez This
defines a
a ez bIn
form
aexright-handed
set, like
unique direction for a
b, which is obtained from a
right-hand rule. See Figure A.8.
us now
Let
(a
b)
(ay bz
evaluate the magnitude of
b. We have
az
by )
+ (az
bx
ax
bz )
+ (ax
ay
by
bx )
(a x
2
y
+ a + a z 2 )(b x2 + by2 + b z 2 ) (ax

=
|a|
|a|
2
2
|b|
|b|
2
2
(a b)
2
|a| |b|
bx
+ ay
by
+ az
bz )
2
2
cos
= |a|
|b|
sin
2
.
(A.43)
Thus,
|a
b|
= |a| |b| sin ,
(A.44)
a and b. Clearly, a
a =e0 for
vector,
since
isdalways
wher
is tany
he a
ng le su
btende
bet we zero
en ainanthis
dbcase.
Cl
Also, if a
b = 0 then either |a| = 0, |b| = 0, or b is
parallel (or antiparallel) to a.
Consider the parallelogram defined by the vectors a
where is the angle subtended between
and
b. See
Figure
A.9. The
scalar
area
of
the
area
area, and is normal to
the plane of the parallelogram, in the sense obtained from
a right-hand circulation rule by rotating a on to b (through
parallelogram
is a bsin . By convention, the vector
has the magnitude
231
of the scalar
i.e., if the fingers of the right-hand circulate inthe

direction of rotation then the thumb of the right-hand
indicates the direction of the vector
area
that
area. So, the vector
is coming out of the page inFigure A.9. It follows
=a
(A.45)
b,
Suppose that
Figure
product
force F is applied at position
A.10. The torque

of the magnitude
the lever
arm
about
r.See
the origin O is the
of the force and the length of
OQ. Thus, the magnitude
of the torque is
|F||r|sin . The direction of the torque is conventionally

defined
is |F||r|as
sin.
the direction of the axis through O about which
the force tries to rotate objects, inthe sense determined by

a right-hand circulation rule. Hence, the torque is out of
the page in Figure A.10.It follows that the vector torque is
given by
=r
(A.46)
F.
The angular momentum,
momentum
p and
a particle of linear
r is simply defined as
l, of
position vector
the moment of its momentum about the origin: i.e.,
=r
p.
(A.47)
A.9
Let
Rotation
us try to define a rotation vector
whose magnitude is
the angle of the rotation, , and whose direction is parallel
to the axis of rotation,

right-hand
circulation
in the
sense
determined
rule. Unfortunately,
by
this is not
a
a
good vector. The problem is that the addition of rotations is
not commutative, whereas vector addition is commuative.

Figure A.11 shows the effect of applying two successive
90
to
rotations,
one
about Ox, and the other about the Oz,
six-sided
standard
die. In the left-hand
case, the
versa in
ends up in
z-rotation is applied before the x-rotation, and vice

the right-hand
case.
two completely
It
can
be
different
z-rotation plus the x-rotation
seen
states.
that the die

In other
words,
the
does not equal the x-rotation
plus the z-rotation. This non-commuting
algebra
cannot
well-defined
quantities.
a
are not vector
by vectors. So, although rotations have
be represented
magnitude
and direction, they
But, this is not quite the end of the story. Suppose that
we take a general vector a and rotate

angle z
axes
This is equivalent
about
the Oz by
(A.20)(A.22),
we have
a small
it about Oz by
to rotating the coordinate

According
to Equations
a +
ez
a,
(A.48)
where
use
has been made of the small angle
approximations sin and cos 1. The above

equation
can easily
be generalized to allow small rotations
about Ox and Oy by x and y
a +
(A.49)
where
respectively. We find that
a,
ex +
ey +
ez
(A.50)
Clearly,
we can
define
rotation vector, , but it only
works for small angle rotations (i.e., sufficiently small that

the
small angle approximations
good).
According
z-rotation plus
to
the
of sine and cosine
above
equation,
are
small
a
S
b
a
232
FLUID MECHANICS
PQ
rsin
z -axis
x-axis
Figure A.10: A torque.
z
x
y
x-axis
z -axis
Figure A.11: Effect of successive rotations about
perpendicular
axes ona six-sided
die.

233
c
b
Figure A.12: A vector parallelepiped.
small
x-rotation
rotations
infinitesimal
velocity,
is (approximately)
applied
in the opposite
rotation
is
a vector
to the
equal
two
order. The fact that

implies
that
angular
lim
t0
(A.51)
must be a vector
as well. Also, if a
is interpreted
as a(t +
t) in Equation (A.49) then it follows that the equation of
motion of
some
a vector
which
precesses
about the origin with
angular velocity is
da
dt
a.
(A.52)
A.10
Scalar Triple Product
a, b, and c. The scalar triple

c. Now, b c is the vector area
of the parallelogram defined by b and c. So, a b
c is the
scalar area ofp ara
this
llel parallelogram
ogram def ine dmultiplied
by b and c.by
So, the
a
component of a in the direction of its normal. It follows
he
that
com
a b
po nen
c istthe
of volume of the parallelepiped defined by
vectors a, b, and c. See Figure A.12. This volume is
independent of how the triple product is formed from a, b,
and c,except that
Consider
three
vectors
product is defined
if
a,
= a
b.
(A.53)
So, the volume
is positive
b, and
right-handed set (i.e., if a lies above the plane of b and

the
sense
determined from
a right-hand
a
c,in
form
circulation rule by
rotating b onto c) and negative if they form
a left-handed
cross
set. The triple product is unchanged if the dot and

product operators
are interchanged,
=a
c.
(A.54)
The triple
product
is also
permutation of a,b,and c,
invariant
=b
under
any
=c
cyclic
b,
(A.55)
but
any anti-cyclic
permutation
causes it to change sign,

a b
c = b a
c.
(A.56)
The scalar triple product is zero if
any two
of
a, b, and c
are parallel, or if a,b,and c are coplanar.

If a, b, and c are non-coplanar
then
any vector rcan be written in terms of them: i.e.,
r= a +
+ c.
(A.57)
Forming
the dot product of this equation with b
then obtain
c, we
= a
c,
(A.58)
so
ab
expressions
parameters , ,
(A.59)a
can be written for and . The
and are uniquely determined provided
i.e., provided that the three vectors are
Analogous
b
c
0:
non-coplanar.
234
FLUIDMECHANICS
A.11
Vector Triple Product
a, b, and c, the vector triple product is

or
defined
three vectors
a (b a,cb,
). The
and c,
brackets
the vector
are important
triple product
because
is d a
(b
c)
(a
b)
c. In fact, it can be demonstrated that
a
(b
c) (a c) b (a b) c
For three vectors
(A.60)
and
(a
b)
(a
c) b
(b
c) a.
(A.61)
Let
us try to prove
The left-hand
proper
side and the right-hand
vectors,
particular
so
if
coordinate
general. Let
the first of the above theorems.
us
take
we can prove
system
side
then it must
convenient
axes
are
this result in
both
one
be true in
such that Ox lies
,, ,,
, ,
. , , ,
, ,
x-y plane. It follows that b

a (ax ay az ).The vector
b
c is directed along Oz: i.e., b
c (0, 0, bx cy ).
Hence, a
(b
c) lies in the x-y plane: i.e., a
(b
c)
(ay bx cy
ax bx cy
0). This is the left-hand side of
along b, and
(bx
lies in the
0, 0), c (cx
cy
0), and
Equation (A.60) in our convenient
evaluate the right-hand side,

and
RHS
= ax
bx
([ax
we
coordinate
need
ac
system. To
= ax
cx + ay cy
It follows that the right-hand side is
cx + ay cy ]bx
0, 0) (ax bx
(ay
cy
bx
cx ax bx cy 0)
cy 0) = LHS,
ax bx
(A.62)
which
A.12
proves
the theorem.
Vector Calculus
Suppose that vector
varies with time,
so
that
= a(t).
The time derivative of the vector is defined
"a(t
da
dt
= lim
+ t)
a(t)
t0
(A.63)
When written out in component form this becomes
, , .
da
dax
dt
dt
day
daz
dt
dt
(A.64)
Suppose that
a is, in fact, the
product of
a scalar
(t)
and another vector b(t). What
now
of
a? We have
which implies that
is the time derivative
Moreover, it is easily demonstrated

and
dax
dt
(bx )
dt
bx +
dbx
dt
dt
(A.65)
da
dt
b+
db
dt
dt
(A.66)
(a
dt
b )=
da
dt
b +a
db
dt
(A.67)
(a
dt
b)
da
dt
that
b +a
db
dt
(A.68)
Hence, it can be seen that the laws of vector differentiation

are analogous to those inconventional calculus.
235
Figure A.13: A line integral.
A.13
Line Integrals
Consider
a two-dimensional
function f(x,y) which is
defined for all x and y.What is meant by the integral of f
along
a given curve
joining the points Pand Q inthe x-y
plane? Well, we first draw out f as a function of length l
along the path. See Figure A.13. The integral is then

simply given by
Z
f(x,y)dl
=Area under the curve,
(A.69)
where dl
pdx
2
+ dy 2.
As an example of this, consider the integral of f(x,y)

2dx
xy 2 between Pand Q along the two routes indicated in
Figure A.14. Along route 1

we have
2dx.
Thus,
x = y,so dl =
xy
dl
2dx
(A.70)
The integration along route 2 gives
Z
??????y=0
xy
dl
xy
??????x=1
xy
dx
dy
0+
dy
(A.71)
Note that the integral depends
on the route
taken between
the initial and final points.

The most
common type of line integral
is that inwhich
the contributions from dx and dy are evaluated separately,

rather that through the path-length element dl:i.e.,
Z
f(x,y)dx
+ g(x,y)dy
(A.72)
As an example of this, consider the integral
hydx
+x 3
P
dy
(A.73)
along the two routes indicated inFigure A.15. Along route
1
we have
x =y + 1
and dx = dy, so
hydx
+x 3
dy
hydy
+ (y + 1) 3
17
4
dy
(A.80)
236
FLUID MECHANICS
y
Q
= (1,1)
=(0,0)
Figure A.14: An example line integral.
y
Q
=(2,1)
=(1,0)
Figure A.15: An example line integral.
Along route 2,
Z
hydx
+x
dy
??????x=1
dy
??????y=1
ydx
7
4
(A.75)
Again, the integral depends
on the path of integration.
Suppose that we have
depend
a line integral which does not

on the path of integration. It follows that
Z
(f dx
+ gdy
= F(Q) F(P)
(A.76)
for some function F.Given F(P) for one point Pinthe x-y
plane, then
Z
=F(P) +
F(Q)
(f dx + gdy )
P
(A.77)
defines F(Q) for allother points inthe plane. We can then
draw
a contour map of F(x,y). The line integral
between
points Pand Q is simply the change inheight inthe
contour
map
Z
between these two points:

Z
(f dx + gdy )
dF(x,y)
= F(Q) F(P).
(A.78)
Thus,
dF(x,y)
= f(x,y)dx
+ g(x,y)dy.
(A.79)
=x
For instance, if F
y then dF =3 x 2
Z
237
+x 3
?3
dy and
hx
x2
(A.84)
ydx
ydx
+x 3
dy
iQP
3
2Q
x
O
Figure A.16: An example vector line integral.
is independent of the path of integration.

It is clear that there
integralthose
are two distinct types of line

on their endpoints
which depend only
and
on the path of integration, and those which depend

both on their endpoints and the integration path. Later on,
we shall learn how to distinguish between these two types
not
(see Section A.18).
A.14
Vector Line Integrals
A vector field is defined
as a set of vectors
associated
with each point in space. For instance, the velocity v(r)

in a moving liquid (e.g., a whirlpool) constitutes
a vector
field. By analogy,
a scalar
field is a set of scalars
associated with each point in space. An example of
a scalar
field is the temperature distribution T(r) in a furnace.
Consider
a general vector
field A(r). Let dr (dx,
dy, dz) be the vector element of line length. Vector line

integrals often arise
Z
as
A dr
(Ax dx + Ay dy + Az dz).
(A.81)
For instance, if A is a force-field then the line integral is

the work done ingoing from Pto Q.
As
an example,
consider the work done by
element of
= r/|r 3|.The
work done is dW = F dr. Take P = (,0,0)
repulsive inverse-square
central field, F
The
and eleme
Q
(a,0,0). The first route considered is along the
x-axis,
so
Z
"
dx
#a
1x
1a
(A.82)
The second route is, firstly, around
a large
circle (r
constant) to the point (a, , 0), and then parallel to the

y-axis. See Figure A.16. Inthe first part, no work is done,
since Fis perpendicular to dr.In the second part,
"
ydy
(a 2 + y 2) 3/2
#0
(y 2
+ a 2) 1/2
1a
(A.83)
Inthis case, the integral is independent of the path.
However, not all vector line integrals
A.15
are path independent.
Surface Integrals
us take a surface S,that is not necessarily co-planar,

and divide it up into (scalar) elements S i Then
Let
Z ZS
f(x,y,z)dS
=lim
S i 0
238
FLUID MECHANICS
y
y2
dy
y1
Xi
f(x,y,z)S
x1
x1
xx
x2
x2
Figure A.17: Decomposition of a surface integral.
is a surface integral. For instance, the volume of water in a

lake of depth D(x,y) is
Z Z
D(x,y)dS.
(A.85)
To evaluate this integral
we must
split the calculation into
two ordinary integrals. The volume in the strip shown in

Figure A.17 is
"Z
x2
D(x,y)dx
dy.
x1
(A.86)
Note that the limits x1 and x2
depend
on y. The
total
volume is the sum over allstrips: i.e.,

Z
"Z
y2
dy
y1
x2 (y)
D(x,y)dx
Z ZS
D(x,y)dxdy.
x1(y)
(A.87)
Of course, the integral
can be evaluated
by taking the strips
the other
way
around: i.e.,
Z
x2
y2
(x)
dx
D(x,y)dy.
x1
y1(x)
(A.88)
Interchanging
the order of integration is
a very
care must
and useful trick. But great
powerful
be taken when
evaluating the limits.

As an example, consider
Z ZS
xy 2
dxdy,
(A.89)
where
S is shown
in Figure
A.18. Suppose
we
that
evaluate the x integral first:

Z
xy
dy
"
1y
2
dx
=y
x2
#1y
dy
(1 y)
dy.
(A.90)
Let
us now
evaluate the y integral:

Z
dy
1
60
(A.91)
We
can
also evaluate
the integral by interchanging
the
order of integration:
Z
1x
2
y dy
xdx
0
x
(1 x)
dx
1
60
239
y
(0,1)
(0,0)
y =x
(1,0)
Figure A.18: An example surface integral.
In some
cases, a surface
integral is just the product of
two separate integrals. For instance,

Z ZS
x2
ydxdy
(A.93)
where S is a unit
Z
This integral
Z
x2
dx
0
square.
ydy
!Z
can be written
x2
dx
ydy
1 12
1
6
(A.94)
since the limits
A.16
are both independent
of the other variable.
Vector Surface Integrals
Surface integrals often
occur
during vector analysis. For
instance, the rate of flow of a liquid of velocity
an
infinitesimal surface of vector
area
rate of flow througha

surface S made
infinitesim
infinitesimal
surfaces is
v through
dS is vdS. The net
up of lots of
Z ZS
v dS = lim
hX
v cosdS
dS0
(A.95)
where is the angle subtended between the normal to the
surface and the flow velocity.
Analogously to line integrals, most surface integrals

depend both on the surface and the rim. But
some
(very
on the rim, and not on

spans it.As an example
important) integrals depend only

the nature of the surface which
of this, consider incompressible fluid flow between two

surfaces S
1 and
which end on the same rim. See Figure
A.23. The volume between the surfaces is constant,
goes
so what
come out, and
in must
Z ZS
Z ZS
v dS =
v dS.
(A.96)
It follows that
Z Z
v dS
(A.97)
depends only
1 and
A.17
on the rim, and not on the form of surfaces
Volume Integrals
A volume integral takes the form
Z Z ZV
f(x,y,z)dV,
A.104) 240
FLUID MECHANICS
where V is
some
volume, and dV
= dxdydz
is
a small
volume
element.
3r,
The
an example
of
written d
As
volume
element
is sometimes
or even d.
a
the center of gravity of

the pyramid has
a square
volume integral, let
us
evaluate
solid pyramid. Suppose that
base of side
a,a height a, and is
composed of material of uniform density. Let the centroid
of the base lie at the origin, and let the apex lie at (0, 0,a).
mass lies on the line joining

apex. In fact, the height of the center of
By symmetry, the center of

the centroid to the
mass is given by
,Z Z Z
Z Z Z
z=
zdV
dV.
(A.99)
The bottom integral is just the volume of the pyramid, and
can be written
Z Z Z
dV
=
0
(a z)
dz
dx
(az)/2
(az)/2
(a
2az
+ z 2 )dz
ha
=
(A.100)
(az)/2
dy
(az)/2
dz
z az 2 + z 3/3
ia0
13
Here,
we haveevaluatedthez-integrallast becausethe
x- andy-integralsarez-dependent. Thetop
limits ofthe
integral takes the form

Z Z Z
zdV
zdz
0
z(a z)
dz
(az)/2
dy
dx
(az)/2
(az)/2
(az)/2
(za
2az
+ z 3 )dz
ha
2
z 2/2 2az 3/3 + z 4/4
ia0
a 4.
12
(A.101)
Thus,
z
13
a =
12
a.
4
(A.102)
Inother words, the center of mass of
a pyramid
lies one
quarter of the way between the centroid of the base and

the
apex.
A.18
Gradient
A one-dimensional
which is defined
function f(x) has
as the slope
gradient
of the tangent to the
df/dx
curve at
x.
We wish to extend this idea to
cover
scalar fields in two
and three dimensions.
Consider
a two-dimensional
scalar field h(x, y), which
is (say) height above sea-level in a hilly region. Let dr(dx,
dy) be
an
element
of horizontal
distance.
Consider
dh/dr, where dh is the change in height after moving
distance dr. This quantity is somewhat
infinitesimal
the one-dimensional
the
direction
gradient, except that dh depends
of dr,
as well as
some point
its magnitude.
an
like
on
In the
P, the slope reduces to
immediate
vicinity of
an inclined
plane. See Figure A.19. The largest value of
dh/dr is straight
any other
up the slope. It is easily
shown that for
direction
!max
dh
dr
dh
cos,
dr
(A.103)
where is the angle shown in Figure A.19. Let
a two-dimensional vector,
grad h,called the gradient of h,
whose magnitude is (dh/dr)max
and whose direction is the
direction of steepest ascent. The
in the above expression
us define
ensures
cos
variation exhibited
that the component
of
gradh in any direction is equal to dh/dr for that direction.
can be
y, and
finding the slope of the tangent to the curve at given x.
The component
of dh/dr in the x-direction
obtained by plotting out the profile of h at constant

then
This quantity is known
respect
to
as the partial
y,
at constant
(h/x)y
x is written
constant x and
there is any
Likewise, the gradient of the profile at constant

(h/y)x
constant
Note that the subscripts denoting
y are
usually
omitted,
derivative of hwith
and is denoted
unless
ambiguity. It follows that in component form
gradh
, .
(A. 112)
241
contours of h(x, y)
high
dr
direction of steepest ascent
low
O
x
Figure A.19: A two-dimensional gradient.
Now, the equation of the tangent plane at P
is
hT (x, y)
h(x0
y0
(x
+ (y
x0 )
= (x0
y0
y0 ).
(A.105)
This has the same local gradients
as h(x, y),so
h
by differentiation
dy
= y y0
plane, so
(A.106)
= x x0
of the above. For small dx
and
the function h is coincident with the tangent
dh
dx +
dy.
(A.107)
But, gradh
(h/x,
h/y) and dr
(dx, dy), so
dh
gradh
dr.
(A.108)
Incidentally, the above equation demonstrates
that grad h
is a proper vector, since the left-hand side is a scalar, and,

according
to the properties
right-hand side is also
a scalar
are both proper vectors
of
the
dot
product,
the
provided that dr and grad h
(dr is an obvious vector, because it
is directly derived from displacements).
now, a three-dimensional temperature

y, z) in (say) a reaction vessel. Let us
grad T, as before, as a vector whose magnitude is
Consider,
distribution T(x,
define
(dT/dr)max
maximum
and whose direction is the direction of the
gradient. This vector is written in component
form
, , .
!
gradT
x
(A.109)
Here,
T/x
one-dimensional
The change
(T/x)y,z
temperature
is
the
gradient
profile at constant
in T in going from point P to
the
of
and
z.
neighbouring
point offset by dr (dx, dy, dz) is
dT
dx +
dy +
dz.
z
(A.110)
In vector form, this becomes
dT
grad T
dr.
(A.111)
Suppose that dT
=0 for some dr.It follows that

dT = grad T
242
FLUID MECHANICS
dr
0.
= constant
gradT
dr
isotherms
Figure A.20: Isotherms.
So, dr is perpendicular to gradT. Since dT
= 0 along
so-called isotherms (i.e., contours of the temperature),
we
conclude that the isotherms (contours)
are everywhere
perpendicular to gradT. See Figure A.20.

It is,of course, possible to integrate dT.For instance,
the line integral of dT between points Pand Q is written
Z
dT
P
(A.113)
gradT
P
dr = T(Q) T(P).
This integral
between
is clearly
P and Q,
so
independent
of the path taken
grad T
dr must
be path
independent.
Consider
a vector
field A(r). In general, the line
integral
R
points.
on
dr depends
R
However,
integral
is
ween
the end
conservative
conservative
path
the path taken between the end
some
for
fields. It
can
field then A
vector
special
independent.
fields
Such
be shown
we have
the
called
that if A is
= grad V for some
V. The proof of this is straightforward.
fields
are
scalar field
Keeping P fixed,
dr
V(Q),
(A.114)
where V(Q) is
independent
well-defined
function, due to the path
nature of the line integral. Consider moving
the position of the end point by
an infinitesimal amount
inthe x-direction. We have

Z
V(Q
Q+dx
+ dx) =V(Q) +
A dr
Q
(A.115)
= V(Q) + Ax
dx.
dx
Hence,
V
=Ax
with analogous relations for the other components of A.It
follows that
A
=gradV.
The force field due to gravity is a good example of a

conservative field. is the work done in traversing
some
path.
If A is conservative then
dr = 0,
(A.116)
(A.117)
R
Now, if A(r) is a force-field then
dr
(A.118)
H
where
corresponds
to the line integral around
a closed
loop. The fact that
zero net
(which
is
why
conservative
A
conservative).
work is done in going around
to the conservation
closed loop is equivalent
are
fields
good example
of
dr
energy
a closed
in going around
energy
non-conservative
H
field is the force field due to friction. Clearly,
system loses
of
called
a frictional
H
so A
cycle,
0.
243
A.19
Grad Operator
It is useful to define the vector operator
, , ,
(A.119)
which is usually
operator acts
until the
on
called the grad
or
del operator.
to its right in
everything
end of the expression
or a
This
expression,
closing bracket
is
reached. For instance,
, , .
grad f
= f
(A.120)
For two scalar fields and ,
grad ( )
grad
grad
(A.121)
can be written more
succinctly
as
( )
(A.122)
that
Suppose
through
By
an
angle
analogy
we
rotate
about
the
axes
coordinate
Oz.
with Equations
(A. 17)(A.
19),
the
old
ones (x , y , z )
coordinates (x, y, z) are related to the new
via
=
=
x
(A.123) y
x
x
y
+y
cos
(A.124)
sin,
sin
cos,
(A.125)
Now,
!y
!y
,z
!y
+
x
,z
,z
.
,
(A.126)
giving
=cos
+ sin
(A.127)
and
x
= cosx
+ siny
,
.
(A.128)
It
can
be
seen,
from Equations
differential operator transforms
(A.20)(A.22),
in
an
analogous
that the
manner
to a vector. This is another proof that f is a good vector.
A.20
Let
Divergence
us start
with a vector field A(r).
HS
Consider
dS over
some
HS
where dS denotes
an outward
closed surface S,
pointing surface HSelement.
This surface integral is usually called the flux of A

out of S
HS
If A represents the velocity of

HS
some
fluid then
dS
is the rate of fluid flow out of S.

If A is constant in space then it is easily demonstrated
that the net flux out of S is zero,
I
A dS
=A
dS
= A S = 0,
(A.129)
area S of a closed surface is zero.

Suppose, now, that A is
not uniform in space.
H
Consider a very small rectangular Hvolume over which A
since the vector
hardly
varies. The contribution to
A dS from the two
faces normal to the x-axis is
Ax (x
+ dx)dydz
surface 244
Ax (x)dydz
Ax
dxdydz
x
Ax
dV,
x
FLUID MECHANICS
y +dy
z + dz
z
y
z
x + dx
Figure A.21: Flux of a vector field out of a small box.
where dV
=dxdydz
is the volume element. See Figure
are analogous contributions from the sides

to the y- and z-axes, so the total of allthe
A.21. There
normal
contributions is
!
A dS
(A.131)
Ay
Ax
+
x
Az
+
y
dV.
The divergence of a vector field is defined
= A =
Ay
Ax
+
x
Az
+
y
(A.132)
Divergence is a good scalar (i.e., it is coordinate
independent),since
it is the dot
of the
vector
Divergence
is aproduct
good scalar
(i.e.,
it
operator with A. The formal definition of
A isH
A = lim
dV0
dS
dV
(A.133)
This definition is independent of the shape of the
infinitesimal volume element.
One of the most important results in vector field
theory is the so-called divergence theorem
or Gauss
theorem. This states that for any volume V surrounded by a

closed surface S,
ZV
IS
dS =
A dV,
(A.134)
where dS is
proof is
very
an
outward pointing volume

element. The
R
straightforward.
We divide
up
the volume
lots of very small cubes, and sum
into
the surfaces. The contributions
Figure
R
us
all of
from the outer surface.
A.22.
We can use Equation (A.131) for each cube

This tells
dS over
from
the interior surfaces
R
cancel
out, leaving just the contribution
See
individually.
R
that the summation is equivalent to
over the whole volume. Thus, the integral of A dS

over the outer surface is equal to the integral of of
A
A dS
over
the whole volume, which proves the divergence theorem.
Now, for a vector field with A = 0, IS
dV
A dS
any closed surface

and S 2 on the same rim,
for
ZS
A dS.
2
(A.136)
See Figure A.23. (Note that the direction of the surface
elements
onS 1 has been reversed
(A.135)
So, for two surfaces,
ZS
dS =
relative to those
on the
closed surface. Hence, the sign of the associated surface

integral is also reversed.) Thus, if
A = 0 then the
245
interior contributions cancel
exterior contributions survive
Figure A.22: The divergence theorem.
S2
S1
rim
Figure A.23: Two surfaces spanning the same rim (right),

and the equivalent closed surface (left).
246
FLUID MECHANICS
Figure A.24: Divergent lines of force.
on
spans
on
integral depends
the rim but not
surface which
it. On the other hand, if
then the integral depends

Consider
is
v. It
since
the nature of the
on both the rim and the surface.
an incompressible
fluid whose velocity field
H
v dS
=
0 for any closed surface,
H
is clear that
what flows into the surface must flow out again.

R
v dV =
R
way in which this is possible
everywhere zero. Thus, the velocity
an incompressible fluid satisfy the
Thus, according to the divergence theorem,
0 for
any
is if
volume. The only
components
is
of
following differential relation:

vy
vx
+
x
vz
+
y
= 0.
(A.137)
a vector field A by
lines of force or field-lines. The direction of a line of force
at any point is the same as the local direction of A.
It is sometimes helpful to represent
The
density
(i.e.,
the
of
lines
numberof lines
perpendicular
crossing
unit
surface
to A) is equal to |A|. For instance, in Figure
A.24, |A| is larger at point 1than at point 2. The number of

lines crossing
surface element dS is A
number of lines leaving
a closed
dS. So, the net
surface is
ZV
IS
dS
dV.
(A.138)
If
= 0 then there
surface. Such
no net flux of lines out of any

a solenoidal vector field.
of a solenoidal vector field is one in
is
field is called
The simplest example
which the lines of force all form closed loops.
A.21
Laplacian Operator
So far we have encountered
, , ,
(A.139)
which is a vector field formed from a scalar field, and
Ay
Ax
Az
(A.140)
a scalar field formed from a vector field. There

are two ways in which we can combine gradient and
We can either
form the vector field
diver-gence.
( A) or the scalar field ().gence. We can eith
which is
Theformeris
notparticularlyinteresting,
() turns
Let
up
but the scalar field
many physical problems, and

is, therefore,
discussion.
but
the sca larworthy
field of
(
) tu rns upin a grea t many ph ysi
in a great
us introduce
the heat flow vector h,which is the
energy per unit area across a surface

perpendicular to the direction of h. In many substances,
heat flows directly down the temperature gradient, so that
we can write
rate of flow of heat
T,
(A.141)
where
HS
flow
is the thermal conductivity. The net rate of heat
h dS out of
some
closed surface S must be equal
HS
to the rate of decrease of heat
energy
in the volume V
enclosed by S.Thus, we have
h dS
= t
IS
cT dV
loop
value:
247
where
c is the specific
theorem that
h = c
heat. It follows from the divergence
(A.143)
Taking the divergence of both sides of Equation (A.141),
and making
(T
use of Equation
=c
(A.143),
we obtain
(A.144)
If is constant then the above equation
(T)
can be written
(A.145)
The scalar field
(T)
takes the form
(T)
2T
2T
+
y
2T

z 2
2T.
(A.146)
Here, the scalar differential operator
+
y
z 2
(A.147)
is called the Laplacian.
The Laplacian is
operator (i.e., it is coordinate independent)

formed
from
combination
of divergence
good scalar
because it is
(another good
scalar operator) and gradient (a good vector operator).
What is the physical significance of the Laplacian? In

2T
2T/x 2.
2T/x 2
reduces to
Now,
is
one dimension,
positive if T(x) is concave (from above) and negative if it

is
convex.
So, if T is less than the average of T in its

2T
is positive, and vice versa.
surroundings
surroundings then
then
In two dimensions,
2T
2T
x 2
2T
+
y
(A.148)
Consider
local minimum
of the temperature.
At the
minimum,
the slope
of T
all directions,
so
Conside
ra local
mincreases
inimu m in
of the
te mp e rature
2T
2T
is positive. Likewise,
is negative at a local
maximum.
Consider,
now, a
steep-sided
valley
in T.
runs parallel to the

2T/y 2
x-axis. At the bottom of the valley
is large and
2T/x 2
positive, whereas
is small and may even be

2T
negative. Thus,
is positive, and this is associated
Suppose that the bottom of the valley
average local value.

us now return to the heat conduction problem:
with T being less than the

Let
(A.149)
It is clear that if
T is positive then T is locally less than
the
It isaverage
clear thvalue,
at if so T/t > 0: i.e., the region heats up.
2T
Likewise, if
is negative then T is locally greater
than the
average
i.e., T/t
makes physical
A.22
Consider
value, and heat flows out of the region:
< 0. Thus,
the above heat conduction equation
sense.
Curl
a vector
field A(r), and
a loop
which lies in one

H
plane. The integral of A around this loop is written

H
dr, where dr is a line element of the loop.

H
conservative
field then A
loops. In general, for
0.
and
a non-conservative
If A is a
dr
field,
H
= 0 for all
H
dr
For
expect
small
H
we
dr to be proportional
Moreover, for
depend
loop
on
a fixed-area
to the
the orientation
area
we expect
loop
of the loop.
H
dr to
of the loop. One particular

(A .156)
248
FLUIDMECHANICS
z +dz
1
y + dy
y
Figure A.25: A vector line integral around
a small rectangular
H
H
dr
= Imax
loop in the y-z plane.
If the loop subtends
optimum orientation then
we expect
an angle with this

= cos . Let us
I Imax
introduce the vector field curlA whose magnitude isH
for the orientation giving Imax
dr
|curlA|
lim
dS
dS0
(A.150)
Here, dS is the
loop. The direction of curl A is perpendicular
area
of the
to the plane
of the loop, when it is in the orientation giving Imax
the sense given by a right-hand circulation rule.
with
Let
us now express
a small
the
of
First, we shall evaluate
components of A.
around
in terms
curlA
dr
rectangle in the y-z plane. See Figure A.25.
The contribution from sides 1

and 3 is
Az (y
+ dy)dz
Az (y)dz
Az
dydz.
y
(A.151)
The contribution from sides 2 and 4 is
Ay (z
+ dz)dy + Ay (z)dy
Ay
dydz.
y
(A.152)
So, the total of allcontributions
I
gives
dr
Ay
Az
dS,
z
(A.153)
where dS
=dydz is the area of the loop.
a non-rectangular (but still small) loop in the

can divide it into rectangular elements, and
form
A dr over all the resultant loops.
The interior
H
contributions
cancel, so we are just left with the
contribution from the outer loop. Also, the area of the outer
loop
the contributi
is the sum of all the areas of the inner loops. We
Consider
H
y-z
plane.
We
H
conclude that
!
A dr
Ay
Az
dS x
(A.154)
a small loop dS = (dS x 0, 0) of any shape in

the y-z plane. Likewise, we can show that if the loop is in
is valid for
the x-z plane then dS
=(0, dS
0) and
!
dr =
Ax
Az
dS
(A.155)
Finally, if the loop is in the
x-y plane then dS =(0, 0, dS
z)
and
!
A
rotation
249
dr =
Ay
Ax
dS z
dS
Figure A.26: Decomposition of a vector
area into its
Cartesian components.
an arbitrary loop of vector area dS = (dS x

can construct this out of three vector areas,
1, 2, and 3, directed in the x-, y-, and z-directions,
respectively, as indicated in Figure A.26. If we form the
Imagine
dS
dS z ).We
line integral
contributions
around
all three
cancel, and
we are
loops
then the
interior
left with the line integral
around the original loop. Thus,

I
dr =
dr1 +
A dr2 +
dr3
(A.157)
giving
dr = curlA dS = |curlA||dS|
cos,
(A.158)
where
curlA
Ay
Az
Ax
Az
Ay
Ax
(A.159)
and is the angle subtended between the directions of curl

and dS. Note that
curlA
(A.160)
A.
This demonstrates
since it is the
that
cross
A is
product
good vector
field,
of the operator (a good
vector operator) and the vector field A.

Consider
angular
a solid
velocity
body rotating about the z-axis. The
is given by
rotation velocity at position
(0, 0, ),
so
the
ris
[see Equation (A.52)]. Let

H
of
us
(A.161)
v on
evaluate
rotation.
x-component
[s ee The
Eq uation
( A .52)]
H
v dr
around a loop in the y-z
H
H
plainly zero. Likewise, the y-component is
H
H
z-component
is
v dr/dS
around some loop in the x-y plane.
Consider
a circular loop.
H
integral
We have
vdr
H
= 2rr
,,
perpendicular
distance
the axis
is
to
L etproportional
u se valuate
with dS
the
v
plane. This is
also
zero.
The
vdr/dS
= r 2. Here, r is the
from the rotation axis. Her

It follows
e,r is
= 2, which is independent of r. So, on the

axis,
( v = (0 0 2 ). Off the axis, at position r0 we
that
that (
v)z
can write
v =
(r r0 )+
r0
(A.162)
same
The first part has the
as the velocity field on the

zero curl, since it is constant.
curl
axis, and the second part has
Thus,
= (0, 0, 2 ) everywhere
allows
form
imagine
v
us to
A(r)
as the
in the body. This
physical picture of
velocity field of
some
A. If
we
fluid then
imagine
A at any given point is equal to twice the local angular
rotation velocity: i.e., 2. Hence,
= 0 everywhere
250
a vector
is said to be irrotational.
field with
FLUID MECHANICS
result of vector field theory is the
Another important
curl theorem
or Stokes
theorem:
IC
ZS
dr
dS,
(A.163)
for
some
theorem
many
a rim C. This
the loop up into
(non-planar) surface S bounded by
can easily
be proved by splitting
small rectangular
loops, and forming
the integral
around all of the resultant loops. All of the contributions

from
the
leaving just
interior
the
loops
cancel,
contribution
from
outer
the
use of Equation (A.158) for each of

we can see that the contribution from
loops,
loops is also equal to the integral of

whole surface. This
One immediate
proves the theorem.

consequence of Stokes
A is incompressible.
Consider
R
dS
rim.
the small
Making
all of the
across
the
theorem is that
any two
surfaces, S
and
share
the same
See
A.23.
It is
at
S 2 Awhich
is incom
pressible
. Crim.
onsi
de raFigure
ny tw o
sur faces,
R
H
clear
from Stokes theorem that
S
RH
for both surfaces. Thus, it follows that

H
for
any
closed
However,
divergence
theorem that
we
dS =
dS is the
surface.
A
H
have
same
dS
=0
from
the
A)dV
0 for any volume. Hence,
A) 0.
(A.164)
So,
So,
A
A is
is a solenoidal
sole noid alfield.
f ield.
We have
0 for
any
seen
that for
a conservative
A is lim
dS0
for some particular

loop.of
It
is clear
However,
the magnitude
A isthen that
a conservative
= .
to
A
H
loop. This is entirely equivalent
However, the magnitude of
dr =
field
= 0 for
dr/dS
field. Inother words,
()
0.
(A.165)
Thus, a conservative field is also
A.23
an irrotational one.
Useful Vector Identities
, ,
Notation:
a, b, c, d are
general vectors; ,
are general
, ,
scalar fields;
Notation:
a, b,
A,c,
Bd
are
aregeneral
generalvector
vectors;
fields;
,
(A
are
)B
ge
2A
2
2
neral
(A Bx
scalar
A fields;
By A
A,
B
Bzare
)and
general
vector
( Ax
fields;
Ay(A
)B
2
Az )(but, only in Cartesian coordinatessee Appendix
C).
(a
(a
(b
b)
b) (c
c)
d)
=
=
=

(a
b)
(c
d)
()
(A
B)
A
(A)
(A
B)
A)
A.24
(c
+ ,
(A
B)
=
=
=
=
=
=
=
=
Exercises
c)b (a b)c,
a)b (c b)a,
(a c)(b d) (a d)(b c),
(a
b d)c (a
b c)d,
(a
(A)
=
=
=
(A.166)
(A.167)
(A.168)
(A.169)
(A.170)

B) + B
)B + (B )A,
A) + (A
(A.171)
(A.172)
0,
A + A ,
B
A A
(A.173)
(A.174)
0,
B,
(A.175)
(A.176)
A) 2A,
A +
A,
( B)A ( A)B + (B )A
(
(A.177)
(A.178)
(A
)B.
(A.179)
A.1. The position vectors of the four points A,

B,C,
and
D are a,
b,3a+2b,
and a3b, respectively. Express AC, DB, BC, and
A.1. Th
CD in terms of
a and b.
251
A.2. Prove the trigonometric law of sines
sina
A
sinb
B
sinc
C
using vector methods. Here, a,b,and c are the three

angles of
a plane
triangle, and A,B,and C the lengths of the
corresponding opposite sides.

A.3. Demonstrate using vectors that the diagonals of
parallelogram bisect
one another.
diagonals
of
a quadrilateral
parallelogram.
bisect
In addition, show that if the
one another
then it is a
A.4. From the inequality
a b = |a||b| cos
|a||b|
deduce the triangle inequality
|a
+ b| |a| + |b|.
a b and the vector
A.5. Find the scalar product
when
(a) a
(b) a
=ex
=ex
+ 3ey
2ey
+ez
ez
,,
= 3ex
= 2ex
.,
product
+ 2ey + ez
+ ey + ez
A.6. Which of the following statements regarding the three
general vectors
c
(b) a
(c) a
(a)
a,b,and care true?

(a
b) = (b
a) c.
(b
c)
(b
c)
= (a b) c.
= (a c)b (a b)c.
= a + bimplies that (a b) d = 0.
c = b c implies that c a c b = c|a b|.
(f) (a
b)
(c
b) = [b (c
a)]b.
(d) d
(e)
A.7. Prove that the length of the shortest straight-line from

point
a to the straight-line
|a
joining points b and c is
b +b
c +c
|b c|
a|
A.8. Identify the following surfaces:
= a,
n = b,
(c) r n = c|r|,
(a) |r|
(b) r
(d) |r (r n)n|
= d.
Here, r is the position vector,
a,b,c,and d are positive
constants, and nis a fixed unit vector.

A.9. Let
a,b,and c be coplanar vectors
related via
+ b + c =0,
where , , and are not all zero. Show that the condition
for the points with position vectors
is
ua, vb, and wc to be colinear
= 0.
A.10. If
p,q, and r are any vectors,
q+r, b
= r+p, and c = p+q

where ,
are
a=
coplanar provided that
are scalars. Show
= 1,
that this condition is
a is perpendicular to p,b to q,and c to r. A.11.

a,b,and c are non-coplanar, and form a
satisfied when
The vectors
and
demonstrate that
non-orthogonal vector base. The vectors A, B,and C, defined by
a b
plus cyclic permutations,
Show that
a = (B
are said to be reciprocal vectors.
C)/(A
C),
plus cyclic permutations.
252
FLUID MECHANICS
A.12. In the notation of the previous question, demonstrate that

the plane passing through points a/, b/, and c/ is normal to
the direction of the vector
= A +B + C.
In addition, show that the perpendicular

1.
plane from the origin is |h|
distance of the
A.13. Evaluate
dr for
A
xex +yey
px
2
+y 2
around the square whose sides
a.
are x = 0,x = a,y = 0,y =
A.14. Consider the following vector field:
= (8 x
A(r)
+3x2 y2
2x
y + 6y, 6).
Is this field conservative? Is it solenoidal? Is it

HC
irrotational? Justify
curve
C is
your answers.
Calculate
dr, where
the
HC
inthe x-y plane, centered
a unit circle
on the origin, and
the direction of integration is clockwise looking down the z-axis.

A.15. Consider the following vector field:
A(r)
=(3 xy
z2
2,
3
y
z 2 + x 2 y,3 x 2
x2
z).
Is this field conservative? Is it solenoidal? Is it irrotational?

Justify
your answers. Calculate

on the origin.
centered
the flux of A out of a unit sphere
A.16. Find the gradients of the following scalar functions of the

position vector
r = (x,y,z):
r,
(a) k
(b) |r|
(c) |r k|
(d) cos(k
n,
r).
Here, k is a fixed vector.

A.17. Find the divergences and curls of the following vector
fields:
(a) k
r,
(b) |r|
r,
n
(c) |r k| (r k),
(d)
a cos(k
r).
Here, k and a are fixed vectors.

2
2
A.18. Calculate
when
f(|r|). Find f if
=0.
Cartesian Tensors
253
BCartesian Tensors
B.1
As
be
Introduction
we saw
in Appendix A, many physical entities
represented
mathematically
as
either
scalars
can
or
on their transformation properties under

axes. However, it turns out that
scalars and vectors are particular types of a more general
class of mathematical constructs called tensors. In fact, a
scalar is a tensor of order zero, and a vector is a tensor of
order one. In fluid mechanics, certain important physical
entities (i.e., stress and rate of strain) are represented
mathematically by tensors of order greater than one. It is
therefore necessary to supplement our investigation of fluid
mechanics with a brief discussion of the mathematics of
tensors. For the sake of simplicity, we shall limit this
vectors, depending
rotation of the coordinate
discussion
expressed
to Cartesian
in such
coordinate
coordinate
systems.
systems
are
Tensors
known
as
Cartesian tensors.
B.2
, ,,
, ,
Tensors and Tensor Notation
x,y, z be
Let the Cartesian coordinates
as the xi
x = x1 y = x2
written
where i
runs from 1
to 3. In other words,
z = x3 Incidentally, in the following, any lowercase

roman subscript (e.g., i,j,k) is assumed to run from 1to
3 We can also write the Cartesian components of a general
vector v as the vi In other words, vx = v1 vy = v2 and vz =
v3
By contrast, a scalar is represented as a variable without
a subscript: e.g., a, . Thus, a scalarwhich is a tensor of
order
zerois
represented
as a variable with zero
subscripts, and a vectorwhich is a tensor of order oneis
represented as a variable with one subscript. It stands to
reason, therefore, that a tensor of order two is represented
as a variable with two subscripts: e.g., aij ij Moreover,
an nth-order tensor is represented as a variable with n
subscripts:
e.g., aijk is a third-ordertensor, and bijkl a
fourth-order tensor. Note that a general nth-order tensor
.
.
and
,.
has 3
independentcomponents.
Now, the components

conveniently
visualized
of
as a
second-order
tensor
are
two-dimensional matrix, just
as the components of a vector are sometimes visualized as

a one-dimensional matrix. However, it is important to
recognize that an nth-order tensor is not simply another
name
for
an n-dimensional
matrix. A matrix is just
ordered set of numbers. A tensor,
set
ordered
components
of
transformation
properties
axes. (See Section
hand, is an
that
specific
have
under rotation of the coordinate
B.3.)
Consider two vectors
ai and bi
an
on the other
a and b that are
represented
as
respectively, in tensor notation. According to
Section A.6, the scalar product of these two vectors takes
the form
= a1 b1 +a2 b2 +a3 b3
(B.1)
= ai bi
can be written more
compactly
as
(B.2)
Here,
we
have
made
use
of
the
Einstein
summation
an expression
containing lower case roman subscripts, any subscript that
appears twice (and only twice) in any term of the
convention,
according
to
which,
in
expression is assumed to be summed from 1

to 3 (unless
stated otherwise). Thus, ai bi
= ai1 b1 + ai2
+ ai3
=a1 b1 +a2 b2 +a3 b3
and aij
an index is
summed it becomes a dummy index, and can be written
as any (unique) symbol: i.e., aij b j and aip b p are
equivalent.
Moreover,
only non-summed,
or free,
bj
b2
b3
Note that when
a tensor expression.
a zeroth-order tensor (because there are no
free indices), and aij b j is a first-order tensor (because
there is only one free index). The process of reducing the
order of a tensor expression by summing indices is known
as contraction. For example, aii is a zeroth-order
indices count toward the order of
Thus, aii is
contraction
when
of the second-order
two
tensors
are
the second-order
are
an outer
Likewise, when two
multiplied together in a manner that involves
contraction then the resulting tensor is called
product:
e.g.,
the first-order
bi
Note, from
tensor aij
is
an
an
inner
inner
tensor aij and the first-order
product of the second-order
tensor
without
product:
tensor ai b j is the outer product of
the two first-order tensors ai and bi
tensors
Incidentally,
together
contraction the resulting tensor is called
e.g.,
tensor aij
multiplied
Equation (B.2), that the scalar
product of two vectors is equivalent to the inner product of

the correspondingfirst-order
254
According
vectors
tensors.
FLUID MECHANICS
to Section A.8, the vector product of two
a and btakes
the form
(a
(B.3) (a
(B.4) (a
b)1
b)2
b)3
=
=
=
in tensor notation. The above expression
a2 b3 a3 b2
a3 b1 a1 b3
,,
a1 b2 a2 b1
(B.5)
can
be written
compactly
as
(a
Here,
b)i
ijk
aj
bk
(B.6)
more
+1
ijk
1
0
if i,j,k is an even permutation of 1,2,3

if i,j,k is an odd permutation of 1,2,3
otherwise
(B.7)
is
as
known
sometimes,
particular,
e.g.,
113
the
third-order permutation
the third-order
that ijk
= 212
one
(or,
tensor). Note, in
Levi-Civita
is zero if
tensor
of its indices is repeated:
= 0. Furthermore,
it follows from (B.7)
that
ijk
jki
kij
kji
jik
ikj
(B.8)
It is helpful to define the second-order

(also known
as the Kroenecker
(
ij
It is easily
seen that
identity tensor
delta tensor),
if i j
otherwise
ij
ii
ik kj
ij
ai b
aki b
ij
ij
aj
,
,, ,
,
ji
=
=
3,
=
=
=
ai
ij
ai bi
aki bi
etc.
The following is a particularly important tensor identity:
ijk
ilm
= jl km
jm kl
(B.9)
(B.10)
(B.11)
(B.12)
(B.13)
(B.14)
(B.15)
(B.16)
Inorder to establish the validity of the above expression,

let us consider the various cases that arise. As is easily
seen, the right-hand side of (B.16) takes the values
+1
if j
(B.17) 1
,,
=land k =m j,
if j = m and k = l
(B.18)
j,
otherwise.
(B.19)
Moreover, in each product
same
on the
left-hand side, i
has the
Thus, for a non-zero

mcan have the same value
as i(because each factor is zero if any of its indices are
repeated). Since a given subscript can only take one of
three values (1, 2, or 3), the only possibilities that
generate non-zero contributions are j = l
and k = m,or j
= m and k = l,excluding j = k = l= m (since each
factor would then have repeated indices, and so be zero).
Thus, the left-hand side reproduces (B.19), as well as the
value
in both
factors.
contribu-tion, none of j,k,l,and
(B.29) Cartesian Tensors
255
in (B.17) and (B.18). The left-hand side also reproduces

the values in (B.17) and (B.18) since if j
then
ijk
= ilm
and the product ijk
ilm
= land k = m
(no summation) is
equal to + 1 whereas if j
ilm
and the product ijk
Here,
use
validity
= m and k = lthen ijk
= iml =
(no summation) is equal to 1
ilm
has been made of Equation
(B.8). Hence, the
of the identity (B.16) has been established.
In order to illustrate the
use
of (B. 16), consider the
vector triple product identity (see Section A.11)
(b
c)
= (a c)b (a b)c.
(B.20)
In tensor notation, the left-hand side of this identity is
written
[a
(b
c)]i
=ijk a j (klm
bl
cm ),
(B.21)
where
use has been made
of Equation (B.6). Employing
Equations (B.8) and (B.16), this becomes
[a
(b
c)]i
?il
=kij
klm
aj
bl
cm =
?
jm im jl
aj
bl
cm
(B.22)
which, with the aid of Equations (B.2) and (B.13), reduces
to
[a
(b
(B.23)
c)]i
=a
cj
bi
aj
b j ci
= [(a c)b (a b)c]i
Thus, we have established the validity of the vector

identity (B.20). Moreover,
our proof
is much
more
rigorous than that given earlier (in Section A.11).
B.3
As
Tensor Transformation
we saw
in Appendix A, scalars and vectors
are
defined
according to their transformation properties under rotation

of the coordinate
axes.
rotation
coordinate
according
of
the
of
a scalar is invarient
axes. On the other
In fact,
under
hand,
to Equations (A.49) and (B.6), the components
general vector
rotation of the coordinate
under an infinitesimal
axes according to
a i = ai + ijk j ak
transform
(B.24)
Here, the ai
are the components of the vector inthe original

a i are the components in the rotated
coordinate system, the
coordinate system, and the latter system is obtained from

the former via
an
a combination
of
an infinitesimal
rotation
1, an
an angle 2 about axis 2,
and an infinitesimal rotation through an angle 3 about
axis 3. These three rotations can take place in any order.
Incidentally, a finite rotation can be built up out of a great
many infinitesimal rotations, so if a vector transforms
properly under an infinitesimal rotation of the coordinate
through
angle
about
infinitesimal rotation through
coordinate
axis
axes
then it will also transform
properly
under
finite
rotation.
Equation (B.24)
can also be written

a
aj
Rij
(B.25)
where
Rij
ij
kij
(B.26)
is
rotation matrix
specific
to the
two
coordinate
between). To first-order in the i

inverted to give
ai
= R ji a j
a tensor,
(which is not
systems
This follows because, to first-order inthe i

Rik R jk
(ik
jk l jk lik
ij
lik
)(
jk m
mjk
m ik mjk
l lij
l lji
since it is
transforms
Equation (B.25)
(B.27)
=ik
it
=ij
can
be
(B.28)
where the dummy index
use
has been relabeled
has been made of Equations
l,and
(B.8), (B.10), and
(B.12). Likewise, it is easily demonstrated that

Rki Rkj
ij
(B .41)
256
FLUID MECHANICS
It can also be shown that, to first-order inthe i

ijk Rli Rmj
Rnk
lmn
(B.30)
This follows because

ijk Rli Rmj
ijk
(li
a ali
)(mj
Rnk
b bmj
)(nk
c cnk
hli
ijk
mj nk
?ali
?i
mj nk
lmn
lmn
+amj
(imn
ial
(ma
li nk
+inl
+ ank
iam
+ ilm
ml na
nl
li mj
ian
+na
+la
lm nm la
ln ma
mn
lmn
(B.31)
Here, there has been much relabeling of dummy indices,
and use has been made of Equations (B.10) and (B.16).
can similarly
It
be shown that
ijk Ril R jm Rkn
lmn
(B.32)
As
direct
generalization
of Equation
second-order tensor transforms under rotation
(B.25),
as
a ij =
Rik
R jl
akl
(B.33)
whereas
a third-order tensor
as
transforms
a ijk
Ril
R jm
almn
Rkn
(B.34)
The
generalization
straight-forward.
zeroth-order
to
For the
higher-order
case
of
tensors
is
scalar, which is
tensor, the transformation rule is particularly
simple: i.e.,
=a
(B.35)
By analogy with Equation (B.27), the inverse transform is

exemplified by
aijk
Rli Rmj
Rnk
a lmn
(B.36)
same order transform

same manner, it immediately follows that two
tensors of the same order whose components are equal in
one particular Cartesian coordinate system will have their
compo-nents equal in all coordinate systems that can be
Incidentally, since all tensors of the
in the
obtained
coordinate
from the original
axes.
system via rotation
of the
Inother words, if
aij
bij
(B.37)
in one particular Cartesian coordinate system then
a ij
ij
(B.38)
same origin
as the original system). Conversely,
it does not make sense to equate tensors of different
order, since such an equation would only be valid in one
par-ticular coordinate system, and so could not have any
physical significance
(since the laws of physics are
in all Cartesian coordinate systems (with the
and system of units
coordinate
It
independent).
can
easily be shown that the outer product of two
tensors transforms
as a tensor
of the appropriate
order.
Thus, if
cijk
ai
b jk
,
,,
(B.39)
and
a i
(B.40) b
oint in
Cartesian
ij
Rij
aj
Rik R jl
bkl
Tensors
257
then
ijk
jk
= Ril
al
R jm Rkn
bmn
= Ril
R jm Rkn
al bmn
Ril R jm Rkn
clmn
(B.42)
which is the correct transformation
rule for
third-order
tensor.
The tensor transformation rule
can be
combined with
the identity (B.29) to show that the scalar product of two
vectors transforms
= Rij a j
as a scalar. Thus,
= Rij Rik a j bk = jk a j
bk
Rik
bi
use
where
bk
=a
= ai
bj
(B.43)
has been made of Equation (B.14). Again, the
above proof is
more
rigorous than that given previously
(in Section A.6). The proof also indicates that the inner
of two tensors
product
appropriate
as a tensor
transforms
of the
order.
The result that both the inner and outer products of
two tensors transform

is known
as the
as tensors
of the appropriate
order
product rule. Closely related to this rule is
the so-called quotient rule, according to which if (say)
cij
aik
b jk
(B.44)
where b jk is
an arbitrary tensor,
and cij
transforms
as a
axes,
then
tensor under all rotations of the coordinate

aik
which
can
be thought of
b jk also transforms
= Ril
R jm
clm
= Ril
as a tensor.
ik
jk
a b
=
R jm
alk bmk
as
the quotient of cij and
The proof is as follows:
= Ril
Ril Rqk
ij
R jm
alk b jq
alk
Rpm Rqk
= Ril
Rkm
pq
alm b
jk
(B.45)
where
use
has been made of the fact that cij
as tensors,
Rearranging, we obtain
transform
as
well
as
Equation
and bij
(B.28).
(a
ik
Ril
alm ) b
Rkm
= 0.
jk
(B.46)
However, the b
ij
are
so
arbitrary,
the above equation
can
only be satisfied, ingeneral, if
a ik
Ril Rkm
alm
(B.47)
which is the correct
transformation
a tensor.
any type of valid
rule for
Incidentally, the quotient rule applies to
tensor product.
The components
ij
identity tensor,
of the second-order
are invariant
axes. This follows because
have the special property
under rotation of the coordinate
ij
Rik
that
R jl
they
kl
Rik
R jk
ij
(B.48)
where
use
components
has been made
of Equation
of the third-order
(B.28).
permutation
also have this special property. This follows because
ijk
Ril
R jm
Rln
lmn
The
tensor, ijk
ijk
(B.49)
use has been made of Equation (B.31).

ijk
transforms
as a proper third-order
where
that
immediately
aj
tensor
implies, from the product rule, that the vector
as a proper vector:
i.e.,
bk is a first-order tensor provided that ai and bi
are
product of two vectors transforms

ijk
The fact
both first-order tensors. This proof is much

that that given earlier (in Section A.8).
more
rigorous
B.4
Tensor Fields
saw
We
in Appendix
,,
scalars associated
where
that
x = (x1
a vector
A that
with
x2
every
x3 )is
scalar field is
point in
a set
space: e.g.,
a position vector.
of
(x),
saw
every
We also
field is a set of vectors associated with
258
FLUID MECHANICS
a
every point
in space: e.g., aij (x). It immediately follows that a scalar
field is a zeroth-order tensor field, and a vector field is a
space: e.g.,
ai (x). It stands
tensor field is
first-order
a set
to
reason,
then, that
of tensors associated with
tensor field.
Most tensor field encountered in physics
are
smoothly
varying and differentiable. Consider the first-order tensor

field
ai (x).
components
The
various
partial
derivatives
of this field with respect
coordinates xi
of
the
to the Cartesian
are written
ai
x
(B.50)
Moreover,
this
set
of
derivatives
transform
as
the
a second-order tensor. In order to

demonstrate this, we need the transformation rule for the xi
which is the same as that for a first-order tensor: i.e.,
components
of
x i
Rij
xj
.
.
(B.51)
Thus,
x
Rij
(B.52)
It is also easily shown that

xi
x
R ji
(B.53)
Now,
a
x
i
j
xk
xk
x
(Ril al )
R jk
xk
= Ril
al
R jk
xk
(B.54)
a second-order
chain rule, as well
which is the correct transformation rule for
tensor. Here,
as
Rij
use
has been made of the
Equation (B.53). [Note, from Equation (B.26), that the
are not
functions
of position.]
above argument, that differentiating
It follows, from the
a tensor
field increases
one: e.g., aij /xk is a third-order tensor. The

only exception to this rule occurs when differentiation and
contraction are combined. Thus, aij /x j is a first-order
tensor, since it only contains a single free index.
The gradient (see Section A. 1
8) of a scalar field is an
example of a first-order tensor field (i.e., a vector field):
its order by
The divergence
(see Section A.20) of
scalar:
xi
(B.55)
a vector
a
as a
field is
tensor field that transforms
contracted second-order
Finally, the curl (see Section A.22) of
()i
ai
xi
(B.56)
a vector field
as a vector
is
contracted fifth-order tensor that transforms
The above definitions
can
a)i
be used to
ak
ijk
(B.57)
prove a
number
of useful results. For instance,
)i
ijk
xk
which follows from symmetry because ikj
2/xk
= 2/x
xk
Likewise,
ijk
xk
0,
(B.58)
= ijk
whereas
a)
ak
ijk
xi
ak
ijk
xi x
0,
259
Cartesian Tensors
which again follows from symmetry. As a final example,
(a
?ijk
b)
aj
=ijk
bk
xi
ak
bi
bk
ai
ijk
bk + ijk
aj
xi
xi
= b (
ijk
bk
a)
b).
(B.60)
According
A.20),
to the divergence
theorem
(see
Section
ZV
IS
ai dS
ai
dV,
xi
(B.61)
where S is
The above
theorem
closed surface surrounding

is easily
the volume V.
to give, for
generalized
example,
ZV
IS
aij dS
aij
dV,
xi
or
IS
aij dS
or even
IS
adS
B.5
which
has
components take the

systems
We
is
called
have
the
same
an
already
special
ZV
aij
dV,
ZV
a
dV.
xi
(B.64)
property
that
. ,
, ,,
,
tensor.
isotropic
two
tensors:
such
tensor, ij
identity
tensor, ijk
third-order permutation
its
value in all Cartesian coordinate
encountered
namely, the second-order
(B.63)
Isotropic Tensors
tensor
(B.62)
Of
course,
and
the
all scalars
are isotropic. Moreover, as is easily demonstrated, there

are no isotropic vectors (other than the null vector). It turns
out that the most general isotropic Cartesian
second-, third-, and fourth-order
kl
+ ik
jl
are scalars.
+ il
Let
jk
are
tensors
ijk
respectively, where ,
us prove
these important results.
The most general second-order

such that
ij
and
of
ij
, , and
isotropic tensor, aij
is
a ij =
for arbitrary
rotations
Rip
of the coordinate
axes.
from Equation (B.24) that, to first-order

inthe i
?mis
asj
a pq
R jq
It follows
?
ais
mjs
aij
(B.65)
0.
(B.66)
However, the i
us
Let
are arbitrary, so we can write
multiply by mik
we obtain
(ii
ks
is
ki
) asj
asj
mis
mjs
ais
0.
(B.67)
With the aid of Equation (B.16),
(ij
ks
is
kj
) ais
= 0,
(B.68)
which reduces to
2 aij
+ a ji
= ass
ij
(B.69)
Interchanging
difference
the labels iand j, and then taking
between the two equations
thus obtained,
the
we
deduce that
aij
a ji
(B.70)
1
This proof is adapted from P.G. Hodge, Jr., American Mathematical
Monthly 68, 793 (1961).

260
FLUID MECHANICS
Hence,
ass
aij
ij
,
.
(B.71)
which implies that
aij
ij
(B.72)
For
case
the
an
of
isotropic
tensor,
third-order
Equation (B.67) generalizes to
,, ,
mis
Multiplying by mit
asjk
mjt
+ a jik + akji
+ a jik + aikj
+ akji + aikj
2aijk
(B.75) 2aijk
,, ,
respectively.
by
i,
jk
ik
mjs
and mkt
= j,and t = k,respectively,
(B.74) 2aijk
mks
0.
(B.73)
and then setting t
=
=
=
assk
ij
assk
ij
asjs
ik
= i,t
+ asjs ik
+ aiss jk
+ aiss jk
,,
,
(B.76)
the above equations
and then setting i i,j
respectively,
aijs
we obtain
However, multiplying
and ij
aisk
= i,and k =
we obtain
respectively, which implies that
Hence,
2aiss
we deduce
+ asis + assi
(B.77) 2asis
(B.78) 2assi
assi
+ aiss + assi
+ aiss + asis
that
+ asis
=
=
, ,
,
assi
asis
+ aiss
+ aiss
(B.79)
aiss
=asis =assi =0.
(B.80)
+ a jik + akji
=
+ a jik + aikj
(B.82) 2aijk + akji + aikj
2aijk
0,
(B.81) 2aijk
=
=
0,
0.
(B.83)
The solution to the above equation must satisfy
a jik
akji
aijk
.
.
(B.84)
This implies, from Equation (B.8), that
aijk
ijk
(B.85)
For
the
case
of
an
isotropic
fourth-order
Equation (B.73) generalizes to

mis
asjkl
mjs
aiskl
mks
aijsl
mls
aijks
tensor,
0.
(B.86)
Multiplying
setting t
2aijkl
il
the above by mit
,,,,
mjt
mkt
mlt
=i,t = j,t =k,and t =l,respectively,
+ a jikl + akjil + aljki
and then
we obtain
ij + asjsl ik + asjks
+ a jikl + aikjl + ailjk
=
asskl ij + aisks jl + aissl jk
+ aikjl + aijlk
= aijss kl + asjsl ik +
(B.89) 2 aijkl + aljki + ailkj +
=
aijss kl + aisks jl + asjks il
asskl
,
,
(B.87) 2 aijkl
+ akjil
(B.88) 2aijkl
aissl
jk
aijlk
(B.90)
respectively. Now, if aijkl is an isotropic fourth-order tensor
an isotropic second-order tensor, which

a multiple of kl
This, and similar
arguments, allows us to deduce that
then asskl is clearly
means
that
is
asskl
(B.91) asjsl
(B.92) asjks
=
=
=
kl
jl
jk
,,
.
Cartesian Tensors
261
Let
us assume,
for the moment, that
aijss
aisks
aissl
=
=
=
Thus, we get
2aijkl
+ a jikl + akjil + aljki
+ a jikl + aikjl + ailkj

+ akjil + aikjl + aijlk
+ aljki + ailkj + aijlk
2aijkl
2aijkl
,,
.
2aijkl
=
=
=
Relations of the form
assij
asisk
asils
(B.94)
(B.95)
(B.96)
ij kl
+ ik jl + il
jk
ij kl
+ ik jl + il
jk
ij kl
+ ik jl + il
jk
ij kl
+ ik jl + il
jk
can
be obtained
,,
,.
(B.97)
(B.98)
(B.99)
(B.100)
aijkl
by subtracting
the
2aijkl
aklij
103) 2ailjk
jl
alkji
Equations
+ (aijlk + aikjl + ailkj )
ik
sum of one pair of

sum of the other pair.
relations justify Equations (B.94)(B.96).
(B.97) and (B.101)
jk
(B.101)
Equations (B.97)(B. 100) from the

These
a jilk
il
ij
(B.102) 2aiklj
jk
+ ij
+ (ailkj + aijlk + aikjl )

ik jl
kl
kl
+ ik jl +il
+ (aikjl + ailkj + aijlk
(B.
il jk
+ ij kl +
(B.104)
The latter two equations

cyclic
permutation
unchanged.
obtained from the first via

l, with iremaining
get
Summing Equations (B.102)(B.104),we

2(aijkl
=( +
are
of j, k, and
+ )(ij
+aiklj +ailjk )+ 3(aijlk + aikjl + ailkj )
+ ik
kl
jl
+ il
jk ).
(B.105) It
follows from symmetry that
aijkl
15
(
+ + )(ij
can
This
be
+ ik
kl
seen
+ aiklj + ailjk
jl
+ il
= aijlk
+ aikjl + ailkj =
jk ).
(B.106)
by swapping the indices k and l

in the
above expression.
Finally, substitution
into Equation (B.
102) yields
aijkl
ij
kl
ik
jl
il
jk
(B.107)
where
=
=
=
..
B.6
B
(B.108)
(B.109)
(4
)/10,
(4 )/10,
(4 )/10.
(B.110)
Exercises
Show
that
general second-order
decomposed into three tensors

aij
where uij
(i.e., uii
= uij
is symmetric
0), vij
+ vij
+ sij
(i.e.,
is isotropic,
u ji
tensor aij
= uij ) and
and sij
only
can
be
traceless
has
three
independent
components.
B.2. Use tensor
to establish
methods
the following
vector
identities:
262
(b) (a
c = a b c = b a c.
b) c = (a c)b (c b)a.
(c) (a
b) (c
(d) (a
b)
(a)
(e)
(a)
(f)
d)
(c
(a)
FLUID MECHANICS
= (a c)(b d) (a d)(b c).

= (a b d)c (a b c)d.
d)
= a + a .
= a +
a.
(g)
(a
= (b )a (a )b + ( b)a ( a)b.
a) = 2a ( a) + 2(a )a.
(i)
(
a) = ( a) a.
Here, [(b )a]i = b ai /x
and ( a)i =
ai
b)
(h) (a
B.3. A quadric surface has
an equation
of the form
a x 12 + b x 22 + c x 32 + 2 f x1 x2 + 2g x1 x3 + 2h x2
x3
= 1.
Show that the coefficients inthe above expression
transform under rotation of the coordinate
axes like the
components of a
symmetric second-order tensor. Hence, demonstrate that
the equation for the surface
can be written
x j = 1,
inthe form
xi Tij
where the Tij
are the components
of the aforementioned
tensor.
B.4. The determinant of a second-order tensor Aij is defined

det(A)
= ijk
Ai1 A
j2
Ak3
(a) Show that
= ijk
det(A)
A1i A2j A3k
is an alternative, and entirely equivalent, definition.
(b) Demonstrate that det(A) is invariant under rotation
axes.
of the coordinate
= Aik
(c) Suppose that Cij
det(C)
Bkj
Show that
= det(A)det(B).
B.5. If
then and x j
Aij x j = xi
are said to be eigenvalues
of the second-order tensor Aij

of Aij
are calculated
and eigenvectors
respectively. The eigenvalues
by solving the related homogeneous matrix
equation
(Aij
ij
)x j
= 0.
Now, it is a standard result inlinear algebra that

equation of the above form only has
det(Aij
a non-trivial
ij
an
solution when
=0.
Demonstrate that the eigenvalues of Aij satisfy the cubic

polynomial
3
tr(A)
where tr(A)
deduce that Aij
+ (A)
det(A)
=Aii and (A) = (Aii A
possesses
jj
three eigenvalues1
Moreover, show that

tr(A)
det(A)
= 0,
,,
Aij A
.,
ji )/2. Hence,
and 3 (say).
+2 +3
1 2 3
B.6. Suppose that Aij is a (real) symmetric second-order tensor:

i.e., A
ji
= Aij
(a) Demonstrate that the eigenvalues of Aij
are all real,
Cartesian
Tensors
263
(b) Show that eigenvectors of Aij corresponding to

different eigenvalues
are orthogonal to one another.
deduce
that the three eigenvectors of Aij
Hence,
are, or can be
chosen to be, mutually orthogonal.
(c) Demonstrate that Aij takes the diagonal form Aij
=i
ij
(no sum) in a Cartesian coordinate system inwhich the

coordinate
axes are each parallel to one of the
eigenvectors.
B.7. In an isotropic elastic medium under stress the

displacement
ui satisfies
ij
ij
12
ui
t 2
uk
cijkl
xl
ul
xk
where ij is the stress tensor (note that
mass
ji
density (which is a uniform constant), and

cijkl
= K ij
kl
+[ik jl +il
jk
=ij ), the
(2/3)ij
the isotropic stiffness tensor. Here, Kand
kl
].
are the bulk
modulus and shear modulus of the medium, respectively. Show
that the divergence and the curl of
uboth satisfy wave
equations. Furthermore, demonstrate that the characteristic
wave
velocities of the divergence and curl waves
are [(K +
264
FLUID MECHANICS
Non-Cartesian
Coordinates
265
CNon-Cartesian Coordinates
Introduction
C.1
In fluid mechanics
non-Cartesian
used to exploit the symmetry

For
example,
it is convenient
to
coordinates
to employ
systems
describe
are
coordinates
often
of particular fluid systems.
cylindrical
possessing
axial
we investigate a particularly
non-Cartesian
coordinates
known as
symmetry. In this Appendix,

class
useful
orthogonal
commonly
mechanics
of
curvilinear
occurring
coordinates.
examples
are cylindrical
of
and spherical
The
this
two
class
most
in fluid
coordinates. (Note,
incidentally, that the Einstein summation convention is not

used inthis Appendix.)
C.2
Let
,,
Orthogonal Curvilinear Coordinates

x1
x2
x3
be
a set
,, ,, ,,
,,,
,
,,
,,
of standard right-handed
Furthermore,
let u1(x1 x2
be three independent
are
Cartesian coordinates.
x3 ), u2 (x1 x2
such that each unique triplet of x1
associated
with
x3 ), u3 (x1 x2 x3 )
functions of these coordinates which
unique triplet of u1
x3 values is
u3
values. It
as an alternative set of
coordinates to distinguish different points in space. Since
the surfaces of constant u1 u2 and u3 are not generally
follows that u1 u2
can
x2
u2
u3
be used
parallel planes, but rather curved surfaces, this type of
, ,
,, ,
coordinate system is termed curvilinear.

1,
1,
Let h1
|u1 |
h2
|u2 |
and h3
follows that e1
a set
==h1|u1u1
e2
==h2|u2
u2
= |u3 1. It
and e3 ==
h3 |u3
u3 are
are
normal to surfaces of
of unit basis vectors which
constant u1 u2
and u3
respectively, at all points in space.
Note, however, that the direction of these basis vectors is
generally
function
of position.
Suppose
where iruns from 1

to 3, are mutually
that the ei
orthogonal at all
points in space: i.e.,
ei
e j =ij
(C.1)
,,
Inthis case, u1 u2
u3
are said to constitute an orthogonal
coordinate system. Suppose, further, that
e1
e2
(C.2)
e3
=1
,,
at allpoints in space, so that u1 u2 u3 also constitute
right-handed coordinate system. It follows that
ei
ej
(C.3)
ek
=ijk
a general vector A,associated

can be written
Finally,
with a particular
point in space,
(C.4)
Xi
Ai ei
where the ei
are
system, and Ai
,,
the local basis vectors of the u1 u2
= ei A is termed
, , , .,
this system.
Consider
two neighboring
coordinates inthe u1 u2
du1 u2
+du2
u3
+ du3
u3
the ith component of A in
points
u3 system
are
,,
in
space
u1 u2
whose
u3 and u1
It is easily shown that the vector
directed from the first to the second of these points takes
the form
dx
(C.5)
du1
du2
e1 +
|u1 |
Xi
du3
e2 +
|u2 |
e3
|u3 |
hi dui ei
,,
Hence, from (C.1), an element of length (squared) inthe
u3 coordinate system is written
dx
dx =
(C.6)
Xi
2
hi du i
Here, the hi which
known
that
as the scale
,,
are generally
are normal to e1
and e3
e2
at a given point in space, take the form dS
266
and
dS
h1 h2 du1 du2
element of volume, at
functions of position,
dVhdu1 du2 du3
are
factors of the system. Elements of area
FLUID MECHANICS
respectively.
a given
Finally,
an
point in space, is written
where
= h1
Note that [see Equation (A.176)]
and
h2
ui
= 0,
ui
h3
(C.7)
(C.8)
u1 u2
0.
The
latter
result
(C.9)
follows
from Equations (A.175) and (A.176)because (h /h)u1
u2
u3
etc. Finally, it is easily demonstrated from (C.1)
and (C.3) that
ui
(C.10) ui
Consider
a scalar
field (u1
chain rule, and the relation ei

Xi
hi
, ,,
uk
Hence, the components
ijk
,.
(C.11)
u2 u3 ).It follows from the
=hi ui
that
Xi
ui
of in the u1
are
hi
ui
ei
ui
system
ij
,,
u2
()i
(C.12)
u3 coordinate
hi
ui
(C.13)
Consider
a vector
field A(u1 ,u2 ,u3 ).We can write
Xi

(Ai
Xi
(hi
ei )
Xi
Ai
hi
ui
Ai ui )
Xi
Ai
Xi
ui
1h
Ai
hi
ui
hi
(C.14)
where
use has been made
(C.10).
coordinate
ofEquations (A. 174), (C.9), and
Thus,thedivergenceofAintheu1 ,u2
system takes the form
A=
Xi
1h
ui
hi
u3
Ai
(C.15)
We can write
Xk
(Ak ek )
Xk
(hk Ak
uk )
Xk
(hk
Xk
(Ak
uk
(hk
uk
Xj,k
uk )
(hk Ak )
uk
use
where
has
been
made
(C.16)
Equations
of
(A.178), (C.8), and (C.12). It follows from (C.11) that
(
Xj,k
ui
u
(C.17)
ei
Xj,k
(hk Ak )
hi
A)i
uk
hi (hk Ak )
ijk
Hence, the components
coordinate system
are
of
A)i
A in the u1
Xj,k
,,
u2
hi (hk Ak )
ijk
(C.
Non-Cartesian
8)
267
Now,
= [see (A.172)], so Equations

2
(C.12)
in the u1
Xi
Coordinates
and (C.15) yield the following expression for
u2 u3 coordinate system:
u3
1h
ui
hi
ui
(C.19)
The
combined
vector identities
(A.171)
and (A.179)
can
to give the following expression for (A
that is valid in a general coordinate system:

(A
)B
be
)B
12
[(A
B)

(A
A)B + ( B)A
B) (
B) B
A)]
(C.20)
Making
use
of Equations
as
(C.13), (C.15), and (C.18),
well as the easily demonstrated results
Xi
A B
Ai Bi
,
,
(C.21)
Xj,k
ijk
Bk
(C.22)
and the tensor identity

(after
a great
expres-sion
for the components
coordinate system:
Xj
[(A
,,
(B.16), Equation (C.20) reduces
deal of tedious algebra) to the following
)B]i =
of (A
Bi
)B
in the u1 u2
B j h
h j u
hi h j
Ai B
ui
hi h j
u3
hi
j
(C.23)
Note, incidentally,
that the commonly
quoted result [(A
=Note,
A incidentally,
Bi is onlythat
valid
the in
co Cartesian
systems (for which h1 = h2 = h3 = 1).
)B]i
Let
us
define the gradient A of
a vector
coordinate
field A
as
the
Letwhose
us def components
tensor
in a Cartesian coordinate system
take the form
Ai
(A)ij
(C.24)
In an orthogonal curvilinear
coordinate system, the above
expression generalizes to
(A)ij
[(e
) A]i
(C.25)
It thus follows from (C.23), and the relation (ei ) j

ij
that
(A)ij
1 Ai
h j u
Xk
j
+ij
hi h j ui
= ei
Ak
ej =
hi
hi hk uk
(C.26)
The
vector
expression
identity
(A.177) yields the following
2A
for
that is valid in a general coordinate
system:
2A
A)
A).
(C.27)
Making
well
as
use
of Equations
(C.15), (C.18), and (C.19),
as
(C.21) and (C.22), and the tensor identity (B.16),
,,
the above equation reduces (after
a great
deal of tedious
algebra) to the following expression for the components of

2A
inthe u1 u2 u3 coordinate system:
(
Xj
2A)i
1 hi
Ai +
hi u
hi
hi
hj
hi h j
j
1
hj
h j ui
ui
hi h j
1 h
ui
ui
Ai hi
hj
ui
ui u

Ai
2
hi h j
268
hj
hi
hi
2hi
!2
2
j
FLUID MECHANICS
er
x
O
Figure C.1: Cylindrical coordinates.
Note, again, that the commonly quoted result (
Ai is only valid in Cartesian coordinate systems (for
which
h1
2A)i
=h2 =h3 =h=1).
C.3
In the cylindrical coordinate system, u1
u3
px
= z,where r=
are
standard
perpendicular
subtended
+ y 2,
px
= tan
coordinates.
Cartesian
= r,u2 = , and
1(y/x),
x,y,z
and
Thus,
is
the
distance from the z-axis, and the angle
between the projection of the radius vector (i.e.,
the vector
the origin to
connecting
space) onto the
x-y plane
general point in
and the x-axis. See Figure C.1.
A general vector A is written

A
where
er
are
= /||,
course,
is easily
(squared)
ez
Az
,
,
,
and
ez
= z/|z|.
the unit basis vectors
mutually orthogonal,
As
er +
(C.29)
= r/|r|,
Figure C. 1 Of
= Ar
= A er
so Ar
demonstrated,
in the cylindrical
er
See
and
ez
etc.
an
element
coordinate
of
length
system takes the
form
dx
dx
dr
+ r2
dz
(C.30)
Hence, comparison
with Equation (C.6) reveals that the
scale factors for this system
are
hr
(C.31) h
(C.32) hz
=
=
=
1,
r,
1.
(C.33)
Thus, surface elements normal to er
(C.54)
divergence,
coordinate
system
and
(r Ar )
1r
drdz,
rdrd,
(C.36)
element takes the form
dV
rdrddz.
Coordinates
(C.13), (C.15), and (C.18),
curl
rddz,
in
the
cylindrical
are written
1r
(C.35) dS z
to Equations
According
gradient,
are written
=
=
=
dS r
Non-Cartesian
269
and ez
(C.34) dS
a volume
respectively, whereas
,,
Az
(C.39)
1r
e +
ez
(C.38)
er +
,
=
1r
Az
Ar
er +
1r Ar
e +
1r
Az
ez
(r A )
(C.40)
respectively. Here, (r) is a general scalar field, and A(r)
general vector field.
to Equation
According
cylindrical
coordinates,
when
(C.19),
the Laplacian
of
expressed
in
scalar field
becomes
!
1r
+
r
+
z
(C.41)
Moreover,
(A
)A
from Equation
(C.23), the components
in the cylindrical coordinate system
are
)A]r
[(A
(C.42)
[(A
)A]
(C.43) [(A
)A]z
Ar
of
Ar A
Az
,
,
.
.
(C.44)
Let
us define
the symmetric gradient tensor
12
hA
(A)
(C.45)
a transpose.
Here, the superscript T denotes

ij element of
some
second-order
correspondingelement of S
the componentsof
system
Ar
is S
ji
Thus, if the
tensor S is S
ij
then the
Accordingto Equation
g in the cylindrical coordinate

A
are
g
( A)rr
g
( A)
g
( A)zz
g
( A)r
g
= ( A)r
g
( A)rz
g
= ( A)zr
g
( A)z
g
= ( A)z
(C.46)
1r A
Az
Ar
(C.47)
(C.48)
12
1r Ar
(C.49)
1 Ar
Az
+
2
(C.50)
1 A
1r Az
+
2
(C.51)
Finally, from Equation (C.28), the components
the cylindrical coordinate system

(
(C.52)
2A)r
2A)
in
2A)z
2 A
Ar
Ar
r2
r2
2 Ar
A
r2
(C.63) 270
FLUID MECHANICS
C.4
2A
are
(C.53)
of
Spherical Coordinates
r2
Az
,
,.
z
r
Figure C.2: Spherical coordinates.
In the spherical coordinate system, u1
u3
tan
px
2
r=
+ y 2 +px
z2
and x, y, z are
Thus, r is the length
= , where
1(y/x),
coordinates.
the angle
= r,u2 = , and
= cos 1(z/r), =
standard
Cartesian
of the radius vector,
between the radius vector and the
subtended
z-axis, and the angle subtended between the projection

of the radius
vector onto the
x-y
plane and the x-axis. See
Figure C.2.
A general vector A is written

A
= Ar
er +
, ,,
(C.55)
where
er
= r/|r|,
= /||,
course, the unit vectors er

so Ar = A er etc.
demonstrated,
an element
See Figure C.2. Of
are
mutually orthogonal,
As
is easily
= /||.
and e
and e
of
length
(squared) inthe spherical coordinate system takes the form
dx
dx = dr 2 + r 2 d 2 + r 2
sin
(C.56)
Hence, comparison
with Equation (C.6) reveals that the
scale factors for this system
are
hr
(C.57) h
=
=
1,
r,
,,
(C.58) h
Thus, surface elements normal to er

dS r
(C.60)
dS
(C.61) dS
(C.59)
are written
and e
r sin.
=
=
r2
r
sindd,
sindrd,
rdrd,
(C.62)
respectively, whereas
a volume
element takes the form
dV
(C.78)
Non-Cartesian
r2
sindrdd.
Coordinates
271
According
gradient,
coordinate
to Equations (C.13), (C.15), and (C.18),
divergence,
system
curl
in
the
er +
1r
e +
r sin
(C.64)
r 2 r
(r
spherical
are written
and
Ar )+
rsin
(sin A )
A
1
rsin
,
,
=
A
(C.65)
"
=
"
rsin
1
r sin
(sinA )
1r
Ar
r#sin
er
(r A ) e
r
"
1r
1r Ar
(r A )
(C.66)
respectively. Here, (r) is a general scalar field, and A(r)
general vector field.

According
spherical
to Equation
coordinates,
the
(C.19),
when
Laplacian
of
expressed
scalar
in
field
becomes
2
r2
r2
+
r
r2
sin
sin
r2
sin
(C.67)
Moreover,
(A
)A
from Equation
(C.23), the components
in the spherical coordinate system
are
2
[(A
)A]r
Ar
Ar A
(C.68)
[(A
)A]
A A +
of
+A 2
cotA
r
,
,
(C.69)
[(A
)A]
Ar A
+ cotA
(C.70)
Now, according to Equation (C.26), the components of
g in the spherical coordinate system

A
are
g
( A)rr
g
( A)
( A)
g
( A)r
g
= ( A)r
g
( A)r
g
= ( A)r
g
( A)
g
= ( A)
Ar
(C.71)
1r A
Ar
(C.72)
Ar
rsin
cot A
(C.73)
12
1r Ar
(C.74)
12
Ar
rsin
(C.75)
12
1r A
rsin
cot A
(C.76)
Finally, from Equation (C.28), the components
of
2A)r
(
2
Ar
r2
r2
2cotA
r2
Ar
r2
272
FLUID MECHANICS
r2
sin
r2
=
A
2A)
sin
(C.77)
in
are
the spherical coordinate system
2Ar
2A
sin
,
,
=
A
A
A
r2
sin
r2
Exercises
C.5
sin
r2
,,
ez
(C.79)
er
,,
e
orthogonal. Do the same for the basis vectors
and e of the spherical coordinate system.
C.2. Use cylindrical coordinates to prove that the volume of
a
2
a and length l
is a l.Demonstrate that
of inertia of a uniform cylinder of mass Mand
right cylinder of radius

the moment
radius
a about
.
.
.
its symmetry axis is (1/2) Ma
C.3. Use spherical coordinates to prove that the volume of
sphere of radius
a is (4/3)a
of
inertia of
an axis passing
a uniform
sphere of
C.5. For what value(s) of
C.6.
mass Mand radius a about
n is
(r
er )= 0,where r is a
a vector field F = Fr (r)er

0. Here, r is a spherical coordinate.
(a) Find
n is (r n er )= 0,where r is a
spherical coordinate?
spherical coordinate?
Demonstrate that the moment
through its center is (2/5) Ma
C.4. For what value(s) of
for
of the cylindrical coordinate system. Verify that the vectors
are mutually
er
sin
C.1. Find the Cartesian components of the basis vectors
and
2cot
Ar
2A)
satisfying
(b) Use the divergence theorem to show that
F=rm
ZV
ZS
dV
m+ 3
m+1
er dS,
where V is a volume enclosed by

(c) Use the above result (for
that the volume of
right
a right cone
a surface S.
m =0) to demonstrate
is one third the volume of the
cylinder having the same base and height.

C.7. The electric field generated by
of moment
a z-directed
electric dipole
p,located at the origin, is
where
p = pez
"
E(r)
r3
40
and
p)er
3(er
r is a spherical
coordinate. Find the
components of E(r) inthe spherical coordinate system.

Calculate
Eand
E.
C.8. Show that the parabolic cylindrical coordinates
defined by the equations
y,z
,,.
are Cartesian
factors hu hv
constant
hz
x = (u 2
v )/2, y
u,v,z,
=uv,z = z,where x,
are orthogonal. Find the scale

are the u= constant and v =
coordinates,
What shapes
surfaces? Write
an expression
for
f inparabolic
cylindrical coordinates.
C.9. Show that the elliptic cylindrical coordinates
defined by the equations
where
x = cosh
cos,
y = sinh
, , z,
sin,
z = z,
x, y,z are Cartesian
, are orthogonal.
are the = constant
,,.
<
coordinates, and 0 ,
Find the scale factors h

and
= constant
hz
What shapes
surfaces? Write
an
Calculus
ofVariations
273
DCalculus of Variations
D.1
It is
Euler-Lagrange Equation
a well-known
fact, first enunciated by Archimedes,
that the shortest distance between two points in

is
straight-line.
demonstrate
However,
suppose
that
we
plane
wish to
Let us
curves, y(x), which run
points, A and B, in a plane, as illustrated
this result
principles.
from first
consider the length, l, of various

between two fixed
inFigure D.1. Now, l

takes the form
Z
[dx
+ dy 2] 1/2 =
[1+ y
2(x)]
1/2
dx,
(D.1)
where
dy/dx. Note that lis
y(x). In mathematics,
functional.
a function
a function
of
of the function
a function
is termed
Figure D.1: Different paths between points A and B.
Now, in order to find the shortest path between points

A and B, we need to minimize the functional l
with respect
to small variations
in the function y(x), subject to the
constraint that the end points, A and B, remain fixed. In

other words,
we need to solve
l
= 0.
(D.2)
The meaning of the above equation is that if y(x) y(x)
+
where y(x)
isabove
small, eq
then
the first-order
Thy(x),
e meaning
of the
uation
is that ifvariation
y( x)
in
l, +
denoted
vanishes.
words,
l
ord
le
+
y(x)
y(x ),l,
where
y (x) isIn
sm other
all, then
the fir
st2).
O(y
obviously
The particular
yields
function y(x) for which l
an extremum
of l
(i.e., either
or a minimum). Hopefully, inthe case under

it yields a minimum of l.
Consider a general functional of the form
Z
F(y,y
a
(D.3)
x)dx,
=0
a maximum
consideration,
are
where the end points of the integration

that y(x) y(x)
where
+ y(x).
fixed. Suppose
The first-order variation in I

is
written
Z
F
y
dx,
274
FLUID MECHANICS
where y
=d(y)/dx.
Setting I to zero, we obtain

Z
F
y
dx
0.
(D.5)
This equation
must be satisfied
for all possible
small
perturbations y(x).
Integrating
the second term in the integrand
above equation by parts,

Z
"
"
!#
of the
we get
ydx
dx
#ba
F
y
=0.
(D.6)
are fixed then y = 0 at x = a and x

= b.Hence, the last term on the left-hand side of the above
equation is zero. Thus, we obtain
Now, if the end points
"
!#
dx
ydx
=0.
(D.7)
The
above
equation
must
perturbations y(x). The only

is for the
expression
integral to be
zero.
be
satisfied
way
in which this is possible
enclosed in
square
for
brackets in the
Hence, the functional
extremum value whenever
all small
I
attains
an
dx
0.
(D.8)
as the Euler-Lagrange equation.

Let us consider some special cases. Suppose that
does not explicitly depend on y. It follows that F/y
This condition is known
0. Hence, the Euler-Lagrange equation (D.8) simplifies to

F
y
const.
(D.9)
suppose
that F does not depend explicitly

,
Multiplying Equation (D.8) by y we obtain
Next,
on x.
dx
0.
(D.10)
However,
y
dx
F
y
+y
dx
(D.11)
Thus, we get
dx
=y
+y
(D.12)
an
Now, if F is not
right-hand
side
of
explicit
the
function
above
of
equation
then the
is the
total
we obtain
derivative of F, namely dF/dx. Hence,
y
dx
F
y
dF
dx
(D.13)
which yields
F
y
const.
(D.14)
case
to Equation (D.1) and (D.3). Hence, F
is not
an explicit
under consideration,
we have
Returning to the
+ y 2, according
p1
function of y,so Equation (D.9) yields

F
y
y
p1
+ y 2
c,
(D.15)
where
c is a constant.
So,
const.
c2
(D.16)
Of
course, y
= constant
Thus, the shortest distance between

plane is indeed
a straight-line.
two fixed points in a
is the equation of
a straight-line.
Calculus ofVariations
275
D.2
Conditional Variation
we wish to find the function

or minimizes the functional
Suppose that
maximizes
Z
y(x) which
F(y,y
x)dx,
(D.17)
subject to the constraint that the value of

Z
G(y,y
x)dx
(D.18)
remains constant. We
can
achieve
our
goal by finding
an
new functional K = I
+ J, where (x)
an undetermined function. We know that J = 0, since
the value of J is fixed, so if K = 0 then I = 0 as well.
In other words, finding an extremum of K is equivalent to
finding an extremum
of I. Application
of the
extremum of the
is
Euler-Lagrange
equation yields
"
dx
[G]
dx
+
y
[G]
0.
(D.19)
In
principle,
constraint
the
above
(D. 18), yields
equation,
the functions
Incidentally, is generally termed

Fand G have
no explicit
together
(x)
a Lagrange
x-dependence
with
the
and y(x).
multiplier. If
then is usually
constant.
As an example, consider the following famous problem.
Suppose that a uniform chain of fixed length l
is suspended
by its ends from two equal-height
a distance a apart,
where
fixed points which
a < l.What
are
is the equilibrium
configuration of the chain?

Suppose that the chain has the uniform density
length . Let the
x- and y-axes
per unit
be horizontal and vertical,
respectively, and let the two ends of the chain lie at (a/2,
0). The equilibrium
configuration
of the chain is specified
by the function y(x), for a/2
+a/2, where y(x)
ed
is the vertical distance of the chain below its end points
at horizontal position
x. Of course,
y(a/2)
= y(+a/2) = 0.
According to standard Newtonian dynamics, the stable

equilibrium state of
potential
energy
a conservative
the systems
which minimizes
= g
energy.
Now, the
of the chain is written
dynamical system is one
potential
yds
a/2
= g
y[1+ y
2]
1/2
dx,
a/2
(D.20)
where ds
pdx
2
the chain, and

Hence,
we
variations
+ dy 2
is an element of length along

pdx
is the acceleration due to gravity.
need to minimize
U with respect
in y(x). However, the variations
to small
in y(x) must
conserve the fixed length of the chain. Hence,

our minimization procedure is subject to the constraint that
be such as to
ds
a/2
[1+ y
2]
1/2
dx
a/2
(D.21)
remains constant.
It follows, from the above discussion, that
minimize the functional
we need to
a/2
=U+l =
(gy
+ )[1 + y 2] 1/2
dx,
a/2
(D.22)
where is an, as yet, undeterminedconstant. Since the

integrand inthe functional does not depend explicitly
we
y
on x,
have from Equation (D.14) that
(gy
+ )[1 + y 2] 1/2
(gy
+ )[1 + y 2] 1/2
=k,
(D.23)
where k is a constant. This expression reduces to
?
2
?2
1,
276
FLUID MECHANICS
where
= /k, and h = k/g.
Let
+
h
coshz.
(D.25)
Making this substitution, Equation (D.24) yields
dz
dx
1.
(D.26)
Hence,
+ c,
(D.27)
where
c is a constant.
It follows from Equation (D.25) that
y(x)
h [
cosh(x/h
c)].
(D.28)
The
above
solution
contains
three
undetermined
,
constants, h,
and c. We can eliminate two of these
constants
y(a/2)
by
application
of
the
boundary
conditions
= 0. This yields
cosh(a/2
c)
0.
(D.29)
Hence,
c =0,and
y(x)
= cosh(a/2h). It follows
= h [cosh(a/2 h)
that
cosh(x/h)].
(D.30)
The final unknown
constant,
h, is determined
via the
application of the constraint (D.21). Thus,

Z
a/2
[1
a/2
+ y 2] 1/2
dx
a/2
=2h sinh(a/2h).
cosh(x/h)dx
a/2
(D.31)
Hence, the equilibrium configuration of the chain is given
by the
curve
as a catenary
(D.30), which is known
(from
the Latin for chain), where the parameter hsatisfies

?
2h
D.3
sinh
2h
(D.32)
Multi-Function Variation
Suppose
that
we
wish
or
to maximize
minimize
functional
Z
F(y1,y2
,,yF
,y ,y ,,y
the
x)dx.
(D.33)
Here,
the integrand
independent
straightforward
functions
F is
now a
yi (x), for
extension
functional
1, F. A
fairly
in Section D. 1
of the analysis
yields F separate Euler-Lagrange
of the F
equations,
!
dx
F
i
0,
yi
(D.34)
for i
1,F, which determine the F functions yi (x). If F
does not explicitly depend
on the function
yk then the kth
Euler-Lagrange equation simplifies to

F
y
const.
(D.35)
Likewise, if F does not explicitly depend

Euler-Lagrange
on x
then all F
equations simplify to
const,
(D.36)
for i 1,F.
Calculus ofVariations
277
D.4
Exercises
D.1. Find the extremal
curves y = y(x) of the following
constrained optimization problems, using the method of
Lagrange
multipliers:
R 1 hy
(a)
R
i
2
+x 2
R 1
dx,such that
(b)
dx,such that y(0)
= y() =0,and
dx
= 2.
dx
= 2.
(c)
R 1
0
ydx, such that y(0)
= y(1) = 1,and
p1
+ y 2
dx
=2/3.
D.2. Suppose Pand Q are two points lying inthe x-y plane,
which is orientated vertically such that Pis above Q. Imagine
there is
a thin, flexible
wire connecting the two points and lying
entirely inthe x-y plane. A frictionless bead travels down the
wire,
impelled by gravity alone. Show that the shape of the
wire that results inthe bead reaching the point Q inthe least
amount
of time is a cycloid, which takes the parametric form
x()
y()
=
=
k ( sin ),
k (1 cos ),
where k is a constant.
D.3. Find the
curve
y(x), in the interval 0
x p, which
length and maximizes

Z
ydx.
0
278
FLUID MECHANICS
Ellipsoidal
Potential
Theory
279
EEllipsoidal Potential Theory
is of
,.
Let
us
x2
adopt the right-handed Cartesian coordinate system x1
x3
Consider
homogeneous
ellipsoidal body whose
outer boundary satisfies

2
x1
2
a1
x 22
2
a2
x 32
a 32
1.
(E.1)
Let
us
potential
x2
calculate
energy
gravitational
a unit test
potential
mass) at
some
(i.e.,
the
point P (x1
x3 ) lying within this body.

Consider
the
of
the
mass
the contribution
contained within
to the potential at P from
a double cone,
whose
apex
is
P, and which is terminated in both directions at the bodys
cone subtends a
a volume element is written dV = r 2
drd, where r measures displacement from P along the
axis of
the cone. Thus, from standard
classical
outer boundary. See Figure E. 1 If the
solid angle d then
gravitational theory, the contribution to the potential takes
the form
Z
dV
dV,
r
(r)
(E.2)
where
= |PQ|,
= |PR|,
and is the constant
mass
we obtain
density of the ellipsoid. Hence,
d
Z
rdr +
rdr
= 1G(r
+ r 2)d.
(E.3)
The net potential at P is obtained by integrating
over
all
solid angle, and dividing the result by two to adjust for

double counting. This yields
From Figure
E. 1
1
G
4
(r
+ r 2)d.
(E.4)
the position
vector of point Q,
relative to the origin, O, is
where
x = (x1
,,
x2
=x +r
n,
(E.5)
x3 )is the position vector of point P,
and na unit vector pointing from Pto Q. Likewise, the

position vector of point R is
=x+r
n.
(E.6)
However, Q and R both lie
on the
bodys outer boundary.
It follows, from (E.1), that
and r
are the two roots of
xi
+ rni
!2
ai
i=1,3
1,
(E.7)
which reduces to the quadratic
A
= 0,
r2 + B r + C
(E.8)
where
ni2
i=1,3
(E.9)
xi ni
2
i=1,3
(E.10)
=
i=1,3
a i2
2
i
a i2
a i2
,
,
1.
(E. 19 )
280
FLUIDMECHANICS
r
O
Figure E.1: Calculation of ellipsoidal gravitational

potential.
Now, according to standard polynomial equation theory,
+r
= B/A,
and
r r
= C/A. Thus,
r 2 + r 2
=(r
+r
2r
B
A
2
2
CA
(E.12)
and (E.4) becomes

1
G
2
?Pi=1,3
?2
Pi=1,3
xi ni /a i
?Pi=1,3
?2
ni2 /a i2
Pi=1,3
x i2 /a i2
ni2 /a i2
d.
(E.13)

Pi,j=1,3
can also be written
1
G
2
xi
xj
ni
n j /(a i2 a 2j )
?Pi=1,3
?2
ni2 /a i2
Pi=1,3
Pi=1,3
x i2 /a i2
d.
ni2 /a i2
(E.14)
However, the cross terms (i.e., i j)integrate to zero by
symmetry, and
we are left with

Pi=1,3
12
G
x i2 ni2 /a i4
?Pi=1,3
?2
ni2 /a i2
Pi=1,3
Pi=1,3
x i2 /a i2
ni2 /a i2
d.
(E.15)
Let
It follows that
Thus, (E.15)
can be written
where
I
d
Pi=1,3
ni2 /a i2
(E.16)
1 J
?Pi=1,3
ai ai
Ai
i=1,3
Ai
12
1 J
a i2
ai ai
(E.17)
d.
?2
ni2 /a i2
2n i /a i
x i2
(E.18)
(E.19)
281
Ellipsoidal Potential Theory
At this stage, it is convenient
to adopt the spherical
angular coordinates, and (see Section C.4), in terms
of which
cos,
(sin
sin
sin,
cos),
(E.20)
and d
=sindd.
We find, from (E.16), that

Z
/2
=8
0
sin
sin
a 22
Let t
/2
sin
cos
sind
a1
cos 2
d.
a 32
(E.21)
=tan. It follows that

Z
/2
=8
sind
0
dt
a + bt 2
/2
=4
0
sind
(ab)
1/2
(E.22)
where
sin
(E.23)
2
a1
sin
cos 2
a 32
cos 2
a2
a3
,
.
(E.24)
Hence,
we obtain
J
Z
=4a1
a2
/2
sin
a3
(a
2
1
+ a 32
sec 2
tan 2 ) 1/2 (a
d
2
2
+ a 32
tan 2 ) 1/2
(E.25)
Let
u=a 32
tan . It follows that
=2a1
du
a2 a3
(E.26)
where
(a 1
u)
1/2
(a 2
u)
1/2
(a 3
u)
1/2.
(E.27)
Now, from (E.19), (E.26), and (E.27),
Ai
Z
2a1 a2 a3
ai ai
= 2a1
2a1 a2 a3
a2 a3
du
2
(a i
+ u)
du
ai
ai
(E.28)
Thus, from (E.18), (E.26), and (E.28),
du
2a1 a2 a3
a i2
du
34
GM
xi
i=1,3
(E.29)
where
Z
(E.30)
=
0
du
du
2
(a i
+ u)
,
.
(E.31)
Here, M
mass
= V and V = (4/3) a1 a2
a3
are
the bodys
and volume, respectively.
The total gravitational potential
energy
of the body is
written
Z
dV,
2
282
FLUIDMECHANICS
where the integral
over
is taken
all interior
points. It
follows from (E.29) that

3
GM
15
ai
i=1,3
(E.33)
In writing the above,
use
demonstrated result
x i2
dV
2
du
has been made of the easily
= (1/5)a
=
2
i
V.Now,
a i2
i=1,3
(a
2
i
+ u)
(E.34)
so
X
i
a i2
i=1,3
a i2
=
Z
"2
d
=
0
du
i=1,3
2
(a i
+ u)
du
du
= 0
(E.35)
Hence, we obtain
U
GM
10
(E.36)

Fluid K2opt

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Fluid K2opt

Caricato da

Copyright:

Formati disponibili

Fluid Mechanics

The University of Texas at Austin

Volume and Surface Forces

General Properties of Stress Tensor

Stress Tensor in a Static Fluid

Convective Time Derivative

Equations of Incompressible Fluid Flow

Equations of Compressible Fluid Flow

Equilibrium of Floating Bodies

Vertical Stability of Floating Bodies

Angular Stability of Floating Bodies

Energy of a Floating Body

Streamlines, Stream Tubes, and Stream Filaments

Vortex Lines, Vortex Tubes, and Vortex Filaments

Circulation and Vorticity

Kelvin Circulation Theorem

Two-Dimensional Uniform Flow

Two-Dimensional Sources and Sinks

Two-Dimensional Vortex Filaments

Inviscid Flow Over

Complex Velocity Potential

Complex Line Integrals

Boundary Layer Equations

ona Flat Plate

Wake Downstream of a Flat Plate

Von K arm an Momentum Integral

Boundary Layer Separation

Criterion for Boundary Layer Separation

Theorem of Kutta and Zhukovskii

Flow Between Parallel Plates

Flow in Slowly-Varying Channels

Axisymmetric Stokes Flow

Stokes Flow Around

Stokes Flow In and Around

Gravity Waves inDeep Water

Gravity Waves in Shallow Water

Energy of Gravity Waves

Gravity Waves in a Flowing Fluid

Gravity Waves at an Interface

Capillary Waves at an Interface

Wind Driven Waves inDeep Water

Eddington Solar Model

Scalars and Vectors

Cartesian Components of a Vector

Scalar Triple Product

Vector Triple Product

Vector Line Integrals

Vector Surface Integrals

Useful Vector Identities

Orthogonal Curvilinear Coordinates

This book presents

a single semester course on fluid

mechanics that is intended primarily for advanced

students majoring inphysics. A thorough

understanding of physics at the lower-division level,

working knowledge of the laws of

mechanics, is assumed. It is also taken for granted that

with the fundamentals of

multi-variate integral and differential calculus, complex

equations. On the other

hand, vector analysis plays such a central role inthe

comprehensive, review of this subject