Chapter 1

Introduction
The characteristic that distinguishes a fluid from a solid is its inability to resist deformation
under an applied shear stress (a tangential force per unit area). For example, if one were to
impose a shear stress on a solid block of steel , the block would not begin to change shape until
an extreme amount of stress has been applied. But if we apply a shear stress to a fluid element,
for example of water, we observe that no matter how small the stress, the fluid element deforms.
Furthermore, the more stress that is applied, the more the fluid element will deform. This
provides us with a characterizing feature of liquids (and gasesfluids, in general) that
distinguishes them from other forms of matter, and we can thus give a formal definition.
Definition 1.1 : A fluid is any substance that deforms continuously when subjected to a shear
stress, no matter how small.

We will later need to be able to study forces acting on fluid elements, and shear stress will be
very important in deriving the equations of fluid motion, and later in the calculation of drag due
to flow over submerged objects and flow resistance in pipes and ducts. Since shear stress is a
(tangential) force per unit area, we can express this for a finite area A as
F
=
A
where F is the tangential force applied over the area A. This is the average shear stress acting on
the finite area A. But in the derivation and analysis of the differential equations describing fluid
motion it is often necessary to consider shear stress at a point. The natural way to define this is
F
= lim
A0 A
Fluid Properties
Our next task is to consider various properties of fluids which to some extent permit us to
distinguish one fluid from another, and they allow us to make estimates of physical behavior of
any specific fluid. There are two main classes of properties to consider: transport properties and
(other, general) physical properties. We will begin by considering three basic transport
properties, namely viscosity, thermal conductivity and mass diffusivity; but we will not in these
1

lectures place much emphasis on the latter two of these because they will not be needed for the
single-phase, single-component, incompressible flows to be treated herein. Following this, we
will study basic physical properties such as density and pressure, both of which might also be
viewed as thermodynamic properties, especially in the context of fluids. Finally, we will briefly
consider surface tension.
Viscosity
Our intuitive notion of viscosity should correspond to something like internal friction. Viscous
fluids tend to be gooey or sticky, indicating that fluid parcels do not slide past one another, or
past solid surfaces, very readily (but in a fluid they do always slide). This can be an indication of
some degree of internal molecular order, or possibly other effects on molecular scales; but in any
case it implies a resistance to shear stresses. These observations lead us to the following
definition.
Definition 1.2: Viscosity is that fluid property by virtue of which a fluid offers resistance to
shear stresses.
Newtons Law of Viscosity. For a given rate of angular deformation of a fluid, shear stress is
directly proportional to viscosity.
Flow Between Two Horizontal Plates with One in Motion We consider flow between two
horizontal parallel flat plates spaced a distance h apart, as depicted in Figure below. We apply a
tangential force F to the upper plate sufficient to move it at a constant velocity U in the x
direction, and study the resulting fluid motion between the plates.

Referring back to above figure, we see that the consequence of this is deformation of the region
abcd to a new shape ab*c*d in a unit of time. Experiments show that the force needed to produce
motion of the upper plate with constant speed U is proportional to the area of this plate, and to
the speed U; furthermore, it is inversely proportional to the spacing
2

between the plates, h. Thus,

F =

(1)

AU
h

with being the constant of proportionality. Note that this inverse proportionality with distance
h between the plates further reflects physical viscous behavior arising from zero velocity at the
lower plate. In particular, the upper plate motion acts, through viscosity, to attempt to drag the
lower plate; but this effect diminishes with distance h between the plates, so the force needed to
move the upper plate decreases. Now recall that the (average) shear stress is defined by = F/A,
so the above equation becomes
U
=
h
It is also clear that if u is the velocity at any point in the fluid, as depicted in the velocity
profile of the above figure, then
du U
=
dy h
Hence, we expect that Newtons law of viscosity can be written in the more general (differential)
form at any point of the fluid as
du
=
(2)
dy
Non-Newtonian Fluids
Note that not all fluids satisfy Newtons law of viscosity given in Eq. (2). In particular, in some
fluids this simple linear relation must be replaced by a more complicated description. Some
common examples are ketchup, various paints and polymers, blood and numerous others.
It is common for the shear stress of non-Newtonian fluids to be expressed in terms of an
empirical relation of the form
n

du
(3)
=
dy
where the exponent n is called the flow behavior index, and K is termed the consistency index.
This representation is called a power law, and fluids whose shear stress can be accurately
represented in this way are often called power-law fluids.
Density
Definition 1.3 The density of a fluid (or any other form of matter) is the amount of mass per
unit volume.

Pressure
In general, fluids exert both normal and tangential forces on surfaces with which they are in
contact (e.g., surfaces of containers and surfaces of adjacent fluid elements). We have already
seen in our discussions of viscosity that tangential forces arise from shear stresses, which in turn
are caused by relative motion of layers of fluid. Pressure is the name given normal forces per
unit area; i.e.,
Definition 1.4 : Pressure is a normal force per unit area in a fluid.
Classification of Flow Phenomena
Steady and unsteady flows
One of the most important, and often easiest to recognize, distinctions is that associated with
steady and unsteady flow. In the most general case all flow properties depend on time; for
example the functional dependence of pressure at any point (x, y, z) at any instant might be
denoted p(x, y, z, t).
This suggests the following:
Definition 1.5 If all properties of a flow are independent of time, then the flow is steady;
otherwise, it is unsteady.
Uniform and non-uniform flows
We often encounter situations in which a significant simplification can be had if we are able to
make an assumption of uniform flow. We begin by giving a precise definition of this useful
concept, and we then provide some examples of uniform and non-uniform flows.
Definition 1.6: A uniform flow is one in which all velocity vectors are identical (in both
direction and magnitude) at every point of the flow for any given instant of time. Flows for
which this is not true are said to be non-uniform.
Viscous and inviscid flows
It is a physical fact that all fluids possess the property of viscosity which we have already treated
in some detail. But in some flow situations it turns out that the forces on fluid elements that arise
from viscosity are small compared with other forces. Understanding such cases will be easier
after we have before us the complete equations of fluid motion from which we will be able to
identify appropriate terms and estimate their sizes. But for now it is sufficient to consider a case
in which viscosity is small (such as a gas flow at low temperature); hence, the shear stress will be
reasonably small (recall Eq. (2)) and, in turn, the corresponding shear forces will be small.
Assume further that pressure forces are large by comparison with the shear forces. In this
situation it might be appropriate to treat the flow as in viscid and ignore the effects of viscosity.
On the other hand, in situations where viscous effects are important, they must not be neglected,
and the flow is said to be viscous. In practice, it can be the case that the region of a flow field in

which viscous forces are substantial may be very small, implying that they contribute little to
integrated forces; such flows might be treated as in viscid.
Compressible flow and Incompressible
A compressible flow is a flow in which the fluid density varies significantly within the flow
field. Therefore, (x, y, z) must now be treated as a field variable rather than simply a constant.
Incompressible fluids are a mathematical idealization based on the approximation that the
density is constant independent of the pressure of the fluid. Real fluids cannot be incompressible,
because the speed of sound of an incompressible fluid would be infinite, which violates special
relativity. Incompressible flow is nevertheless a very useful approximation in many cases in
particular if the fluid velocity it is less than the speed of sound, that is if the Mach number is less
than one.
Laminar or Streamline Flow,is a well-ordered flow and is characterized by the smooth sliding
of adjacent fluid layers (or lamina) over one another, with mixing between layers occurring only
on a molecular level.
Turbulent Flow, is an erratic (irregular) flow and is characterized by the transfer of small
packets of fluid particles between layers. Thus it is accompanied by fluctuations in velocity.

Chapter 2
The Equations of Fluid Motion
In the study of fluid motion there are two main approaches to describing what is happening. The
first, known as the Lagrangian viewpoint, involves watching the trajectory of each individual
fluid parcel as it moves from some initial location, often described as placing a coordinate
system on each fluid parcel and riding on that parcel as it travels through the fluid. At each
instant in time the fluid particle(s) being studied will have a different set of coordinates within
some global coordinate system, but each particle will be associated with a specific initial set of
coordinates. The alternative is the Eulerian description. This corresponds to a coordinate system
fixed in space, and in which fluid properties are studied as functions of time as the flow passes
fixed spatial locations

Lagrangian view of fluid motion

Eulerian view of fluid motion

The substantial derivative

The disadvantage of using an Eulerian reference frame, especially in the context of deriving
the equations of fluid motion, is the difficulty of obtaining acceleration at a point. In the case of
the Lagrangian formulation, heuristically, one need only attach an accelerometer to a fluid parcel
and record the results. But when taking measurements at a single (or a few, possibly, widelyspaced) points as in the Eulerian approach, it is more difficult to produce a formula for
acceleration (and for velocity as well) for fluid elements in their direction of motionwhich is
what is needed to apply Newtons second law. With respect to the mathematical treatment of the
equations of motion, this difficulty is overcome by expressing accelerations in an Eulerian
reference frame in terms of those in a Lagrangian system. This can be done via a particular
differential operator known as the substantial (or material) derivative which can be derived by
enforcing an equivalence of motion in the two types of reference frames. We first state the

formal definition of this operator, after which we will consider some of the physical and
mathematical details.
Definition 2.1 The substantial (material) derivative of any fluid property f(x, y, z, t) in a flow
field with velocity vector U = (u,v,w) is given by
Df f
f
f
f
= +u +v + w
x
y
z
Dt t
f
=
+ U .f
t

(4)

Hence , the substantial (material) derivative operator is

D

= + (U .)
Dt
t

For a general function f which might represent any arbitrary fluid property, we can write (for a
fluid parcel)
f(x, y, z, t) = f(x(t), y(t), z(t), t)
if we recall that in a Lagrangian system the spatial coordinates of fluid particles are functions of
timeso, any property associated with that fluid particle would also, in general, change with
time. Now differentiate f with respect to t using the chain rule:

df f dx f dy f dz f dt
=
+
+
+
dt x dt y dt z dt x dt
f
f
f
f
= +u +v + w
t
x
y
z

(5)

which is identical to Eq. (4) except for notation on the left-hand side.
Review of Pertinent Vector Calculus
In this section we will briefly review the parts of vector calculus that will be needed for deriving
the equations of fluid motion. There are two main theorems from which essentially everything
else we will need can be derived: Gausss theorem and the general transport theorem. The first of
these is usually encountered in elementary physics classes, but we will provide a fairly detailed
(but non-rigorous) treatment here. The second is rather obscure and occurs mainly only in fluid
dynamics; but it is very important, and it is directly related to an elementary integration formula
due to Leibnitz.
The Divergence Theorem
In one space dimension this is precisely the fundamental theorem of calculus:
7

)dx
f ( x=

F (b) f (a )

where F(x) is a function such that F '( x) = f ( x) .

Theorem 2.1
(Gauss, or divergence) For any smooth vector field F over a region R in R3 with a smooth
boundary S
(5)
= .FdV
F .ndA
S

Application of Gausss Theorem to a Scalar Function

Suppose we consider a vector field F that can be expressed as
F = bf

where f is a scalar function, and b is an arbitrary (but nonzero) constant vector. Then we can
correctly write Eq. (5) as
= . fb()dV
( fb).ndA
S

Now we apply product-rule differentiation to the divergence in the left-hand side integral to
obtain
.( fb) = b.f + f .b
= b.f

We can now write

= b.fdV
( fb).ndA
S

= b.fdV
( fb).ndA
S

0
b. ( fndA fdV =
S
R

= fdV
fndA
S

(6)

Transport theorems
As was mentioned earlier, the main uses of transport theorems come in deriving the equations of
fluid dynamics, and other transport phenomena. We will follow an approach in this section
similar to that emloyed in the discussion of Gausss theorem; namely, we will first consider a 1D example on the real line R to introduce the basic notions, and then extend these in a mainly
heuristic way to multi-dimensional cases such as we will need in the present lectures. Thus, we
will begin with an integration formula on the real line known as Leibnitzs formula; we then
generalize this to 3-D (but it will work in any number of dimensions), and we conclude by
providing a special case known as the Reynolds transport theorem. The basic problem being
treated in all these cases is moving differentiation of an integral into the integrand when the
limits of integration depend on the parameter with respect to which differentiation is being
performed.

Leibnitzs Formula
The simplest application of this type is Leibnitzs formula which, as we have already noted,
applies on the real line. This takes the form
b (t )
b (t )
d
f
b
a
f ( x, t )d t=
d x+
f ( x, t )
f ( x, t )

dt a (t )
x
t
t
a (t )
It is of interest to consider the qualitative implications of this formula because they will be the
same for all other integration formulas introduced in this section. Namely, Leibnitzs formula
implies that the time-rate of change of a quantity f in a region (interval in this case) [a(t), b(t)]
that is changing in time is the rate of change of f within the region plus the net amount of f
crossing the boundary of the region due to movement of the boundary.

The General Transport Theorem

Just as was true for the fundamental theorem of calculus, Leibnitzs formula possesses higher
dimensional analogues. In particular, in three dimensions we have the following.
Theorem 2.2
(General Transport Theorem). Let F be a smooth vector (or scalar) field on a region R(t) whose
boundary is S(t), and let W be the velocity field of the time-dependent movement of S(t). Then
d
=
FdV
dt
R (t )

t
R (t )

dV +

FW .ndA

(7)

S (t )

The formula Eq. (7) is called the general transport theorem, and certain specific cases of it will
be widely used in the sequel.

Before going on to this, however, it is worthwhile to again check whether the indicated vector
operations are consistent from term to term in this equation. It should be clear, independent of
the order of F, that the first term on the right-hand side is consistent with that on the left-hand
side since they differ only with respect to where (inside, or outside, the integral) differentiation
by a scalar parameter is performed. It is the second term on the right that must be examined with
more care. First, consider the case when F is a vector field, say of dimension three for
definiteness. Then the first integral on the right is acting also on a 3-D vector field. We will
assume that the velocity field W is also 3D (although this is not strictly necessary), and that the
region R(t) is spatially 3D. Then the question is What is the order of the integrand of this
second integral? But it is easy to see that the dot product of W and n leads to a scalar, a
numbermaybe different at each point of S(t), but nevertheless, a scalar; and a scalar times a
vector is a vector. Hence, we obtain the correct order.
Now suppose F is replaced with a scalar function F. Then clearly the left-hand side and the first
term on the right-hand side of Eq. (7) are scalars. But now this is also true of the second integral
because, as before, the dot product of W and n produces a scalar, and this now multiplies the
scalar F. In fact, it should be clear that F could be a matrix (often called a tensor in fluid
dynamics), and the formula would still work.
Reynolds Transport Theorem
The most widely-encountered corollary of the general transport theorem, at least in
fluiddynamics, is the following.
Theorem 2.3 Let be any smooth vector (or scalar) field, and suppose R(t) is a fluid element
with surface S(t) traveling at the flow velocity U. Then
D

(8)
=
dV dV + U .ndA

Dt R (t )
t
R (t )
S (t )
This formula is called the Reynolds transport theorem, and just as was the case for F in the
general transport theorem, can have any order. But we will deal mainly with the scalar case in
the sequel. There are two important things to notice in comparing this special case with the
general formula. The first is that the integral on the left-hand side is now being differentiated
with respect to the substantial derivative, and the second is that the velocity field in the second
integral on the right is now U, which must be the velocity of an arbitrary fluid parcel. These
restrictions will play a crucial role in later derivations.
The Continuity Equation
We start by considering a fixed mass m of fluid contained in an arbitrary region R(t). As we have
already hinted, we can identify this region with a fluid element, but in some cases we will choose
to associate this with a macroscopic domain. In either case, the boundary S(t) of R(t) can in
general move with time. Any such region is often termed a system, especially in thermodynamics
contexts, and it might be either open or closed. From our point of view it is only important that it
have fixed mass; it does not matter whether it is the same mass at all timesonly that the
amount is the same.
10

It is convenient for our purposes to relate the mass of the system to the density of the fluid
comprising it via
(9)
m = dV
R (t )

We emphasize that R(t) and may both change with time, but they must do so in a way that
leaves m unchanged if we are to have conservation of mass. An example of this might be a
balloon filled with hot air surrounded by cooler air. As heat is transferred from the balloon to its
surroundings, the temperature of the air inside the balloon will decrease, and the density will
increase (equation of state for a perfect gas). At the same time the size of the balloon will shrink,
corresponding to a change in R(t). But the mass of air inside the balloon remains constantat
least if there are no leaks.
We can express this mathematically as
dm d
= =
dV 0
dt dt
R (t )

(10)

Applying the general transport theorem we obtain

dV + U .ndA =
0

t
R (t )
S (t )

(11)

where U is the velocity of the fluid element.

Using Gausss Theorem, we get

dV + .( U )dV =
0

t
R (t )
R (t )

+ .( U ) dV = 0

R (t )
We now recall that the region R(t) was arbitrary (i.e., it can be made arbitrarily smallwithin
the confines of the continuum hypothesis), and this implies that the integrand must be zero
everywhere within R(t). If this were not so (e.g., the integral is zero because there are positive
and negative contributions that cancel), we could subdivide R(t) into smaller regions over which
the integral was either positive or negative, and hence violating the fact that it is actually zero.
Thus, we conclude that

+ .( U ) = 0
t
This called continuity equation.

(12)

However, in certain cases it is useful to simplify further. For an incompressible fluid, density is
constant. Setting the derivative of the density equal to zero and dividing through by a constant
, we obtain the simplest form of the equation
.U =
0
(13)
The Navier-Stokes Equation
11

Newtons second Law of motion states that the time rate of change of linear momentum is equal
to the sum of the forces acting on it.
dp
(14)
= F
dt
Suppose the fluid element R(t) with surface S(t) travels at velocity U. Then, the linear
momentum of the system is
p = UdV
R (t )

dp d
=
UdV
dt dt
R (t )
Applying General Transport Theorem, we have
d
( U )
=
UdV
dV + UU .ndA

dt R (t )
t
R (t )
S (t )
( U )
dV + .( UU )dV
t
R (t )
R (t )

=
=

( U )

+ U (.U ) + (U .)U dV
t

R (t )

t U +
R (t )

(15)

+ U (.U ) + (U .)U dV
t

+ .( U ) +
+ (U .)U dV

R (t )
The first bracket in the fifth line of Eq. (15) vanishes by virtue of Eq. (12), the continuity
equation. The second bracket represents the material derivative of the velocity, which is the
acceleration.
=

U t

Hence ,
d
DU
dV
UdV =

dt R (t )
Dt
R (t )

(16)

Consequently, Eq.(14) can be written as

R (t )

DU
dV = F
Dt

(17)

Sum of forces

12

We next consider sum of forces acting on fluid element . There are two main types of forces to
analyze:
(i) body forces acting on the entire region R(t),
Fb = fdV
R (t )

where f is the body force per unit mass.

(ii) surface forces acting only on the surface s(t) of R(t),
Fs =

.ndA

s (t )

w here is the stress tensor constitute of the surface stress associated with pressure(P) and
the viscous stress tensor due to the viscous nature of the fluid.
Hence Eq.(17) can be written as

R (t )

DU
dV= Fb + Fs
Dt

fdV + .ndA
R (t )

S (t )

Let us look at the following figure to determine .

13

( 19)

 P + xx

yx
=
zx

xy
P + yy
zy

xz
yz

P + zz

1 0 0 xx xy xz

=
P 0 1 0 + yx yy yz
0 0 1 zx zy zz

=
PI +

(20)

where is the viscous tensor.

Hence,

.n =
Pn + .n
(21)
The minus sign is because the pressure force acts toward the area dA and thus is opposite to the
unit normal.
Hence Eq.(19) becomes
14


R (t )

DU
dV
=
Dt

fdV + PndA + .ndA

R (t )

S (t )

(22)

S (t )

We can denote by
( x )
( y)
=

=
( z )

xx xy xz i

yx yy yz j
zx zy zz k

= ( xx i + xy j + xx k ) + ( yx i + yy j + yz k ) + ( zx i + zy j + zz k )

Note that the first subscript on the components tells us the direction in which the area faces and
the second subscript gives us the direction of the force components on that face.

xx , yy , zz are referred to as normal stresses and xy , xz , yx , yz , zx , zy are referred as shear

stresses.
Using Eq.(6), we have

PndA= PdV

s (t )

(23)

R (t )

Applying Divergence Theorem,

.ndA= . dV

s (t )

(24)

R (t )

From Eq.(22), Eq.(23), and Eq.(24), we have

R (t )

DU
dV
=
Dt

[ f P + . ] dV
R (t )

This implies that

DU
= f P + .
Dt

(25)

Eq.(25) is called Navier-Stokes equation (momentum equation). The NavierStokes equations

were named for the French engineer and scientist Claude Louis Marie Henri Navier and the
English mathematical physicist George Gabriel Stokes who independently derived the equation.
We are now prepared to calculate appearing in Eq. (25), thus completing our derivation of
the NS. equations.
15

 yx zx xy yy zy xz yz zz

. xx +
=
+
+
+
+
+
i +
j +
k
y
z x
y
z x
y
z
x
Eq.(25), is a vector form of Navier-Stokes equation which can also be expressed as
U

+ (U .)U = f P + .
t

In Cartesian coordinates the Navier-stokes equation can be expressed as

u
u
u
u
P xx yx zx
+ u + v + w = fx
+
+
+

x
y
z
x x
y
z
t

v
v
v
v
P xy yy zy
+ u + v + w = fy
+
+
+

x
y
z
y x
y
z
t

(26)

w
w
w
w
P xz yz zz
+u
+v
+ w = fz
+
+
+

x
y
z
z x
y
z
t

Navier-Stokes Equations for Newtonian Fluids

In the late seventeenth century Isaac Newton stated that shear stress in a fluid is proportional to
the time-rate-of-strain, i.e. velocity gradients. Such fluids are called Newtonian fluids. (Fluids in
which is not proportional to the velocity gradients are non-Newtonian fluids; blood flow is one
example.) In virtually all practical aerodynamic problems, the fluid can be assumed to be
Newtonian. For such fluids, Stokes, in 1845, obtained:
u
xx =.U + 2
x
v
yy =.U + 2
y
w
zz =.U + 2
z
v u
xy = yx = +
x y
w u
+
x z

xz = zx =

16

xy = yx = +
y y

(27)

yz = zy = +
z y
where is the molecular viscosity coefficient and is the bulk viscosity coefficient. Stokes made
the hypothesis that
2
=
3
which is frequently used but which has still not been definitely confirmed to the present day.
Substituting Eq. (27) into Eq. (26), we obtain the complete NavierStokes
equations in conservation form:

u
u
u
u
P
u u v
+ u + v + w = f x
+ .U + 2 + + +
x
y
z
x x
x y y x
t

w u
+
z x z

v
v
v
v
P v u
v
+ u + v + w = f y
+ + + .U + 2
x
y
z
y x x y y
y
t

w v
+

z y z

w
w
w
w
P u w v w
+u
+v
+ w = fz
+
+ + +
x
y
z
z x z x y z y
t

The continuity equation dictates that .U =

w
.U + 2

z
z
u v w
+ +
= 0 for incompressible flow.
x y z

Hence,

17

(28)


u u v w u
+
.U + 2 + + +
x
x y y x z x z
2u 2u 2 v 2 w 2u
= 2 2 + 2 +
+
+

y
yx zx z 2
x
2u 2u 2u
u v w
= 2 + 2 + 2 + + +

y
z
x x y z
x
2u 2u 2u
= 2 + 2 + 2
y
z
x

v u
v w v
+
+ + .U + 2 +
x x y y
y z y z
2 v 2u
2v 2 w 2v
= 2 +
+2 2 +
+

y
zy z 2
x xy
2v 2v 2v
u v w
= 2 + 2 + 2 + + +

z
y x y z
x y
2v 2v 2v
= 2 + 2 + 2
z
x y
u w v w
w
+ + + + .U + 2

x z x y z y z
z
2u 2 w 2 v 2 w 2 w
2w
=
+ 2 +
+
+
+2 2
yz y 2 yz
z
xz x
2w 2w 2w
u v w
= 2 + 2 + 2 + + +

z
z
z x y z
x
2w 2w 2w
= 2 + 2 + 2
z
z
x
Hence, we have
2u 2u 2u
u
u
u
u
P
+ 2 + 2 + 2
+ u + v + w = fx
x
y
z
x
y
z
t
x

2v 2v 2v
v
v
v
v
P
+ u + v + w = fy
+ 2 + 2 + 2
x
y
z
y
t
x y z

2w 2w 2w
w
w
w
w
P
+u
+v
+ w = fz
+ 2 + 2 + 2
x
y
z
z
z
z
t
x

18

(29)

In vector form, the Navier-Stokes Equation for incompressible Newtonian fluid is

U

+ (U .)U = f P + 2U
t

(30)

2
2
2
+
+
where =
is the Laplacian operator.
x 2 y 2 z 2
2

Dimensionless form of Governing Equations

The goal of such dimensional analysis is to identify the set of dimensionless parameters
associated with a given physical situation (in the present case, fluid flow represented by the N.
S. equations) which completely characterizes behavior of the system (i.e., solutions to the
equations). The first step in this process is identification of independent and dependent variables,
and parameters, that fully describe the system. Once this has been done, we introduce typical
values of independent and dependent variables in such a way as to render the system
dimensionless. Then, usually after some rearrangement of the equations, the dimensionless
parameters that characterize solutions will be evident, and it is these that must be matched
between flows about two geometrically similar objects to guarantee dynamic similarity.

Nondimensionalization of a governing equation is accomplished by dividing every dependent

and independent variable in the equation by an appropriate combination of characteristic
dimensions, thereby making each variable dimensionless. A characteristic dimension is a
x
x = , physical dimension that is in some way characteristic of the flow field under
L0
investigation. Common examples of characteristic dimensions include a characteristic length
scale L, usually derived from the geometry; a characteristic velocity scale U, usually defined as
the average fluid velocity; a characteristic pressure P; and a characteristic time scale T.
We will illustrate the process of obtaining nondimensional governing equations for the case of a
constant density, constant viscosity, flow of a Newtonian fluid.
The technique to make the equations dimensionless consists of introducing a simple change of
variables,

=
0
where is the original variable with dimensions, 0 is a dimensional constant (in particular, a
reference or characteristic value of the variable) and consequently, is a dimensionless
variable.
Next, we introduce the dimensionless independent variables

19

f
x
y
z
t
, f ' = , g is gravitational acceleration
, y' = , z' = , t'=
g
L0
L0
T0
L0
and the dimensionless dependent variables
v
u
w
P
, w' =
u'=
, v' =
, P' =
U0
U0
U0
P0
When applied to a derivative, this change of variables yields
(0 )

t '
=
= =
0 = 0
0
t
t
t
t t T0 t
(0 )

xi 0
=
= =
0 =
0
xi
xi
xi
xi xi L0 xi
x' =

. '
L0
Solving for the dimensional variables in terms of their dimensionless counterparts and
substituting the result in continuity equation, we find
.U =
0

Hence, . =

U0
.U ' =
0
L0
U
Dividing by 0 , we get the dimensionless form of the continuity equation
L0
.U =
0
Similarly, Navier-Stokes Equation becomes

(31)

+ (U .)U = f P + 2U
t

2
U U U 0

P
U
f 0 f 0 P + 20 2U
+
(U .)U =
0
L0
L0
T0 t L0

Dividing both sides of the equation by

U 02
L0

, we the dimensionless form of NavierStokes

equation as
L0 U
L g
P
2
)U 0 2 f 0 2 P +
+ (U .=

U
T
t
U
U
U
L

0
0 0
0 0
0

(32)

We see that there are four dimensionless numbers in dimensionless form of Navier-Stokes:
L0
L00
=
= St , Strouhal Number
U 0T0
U0

20

L0 g
U0
related to Froude Number Fr =
2
U0
L0 g
P0
= Eu, Euler Number
U 02

U 0 L0

related to Reynolds number, R e =

U 0 L

With the introduction of these dimensionless numbers, the NavierStokes equations in

dimensionless form is

( St )

U
1
1 2
)U 2 f ( Eu ) P +
+ (U .
=
U
t
Fr
Re

21

(32)