II.

Governing Equations
In aerodynamics, or fluid mechanics, there are six properties of the flow an
engineer is usually interested in pressure p, density , the three velocity components
(u,v,w), and the temperature T. For gases and mixtures of gases (e.g. air) the equation of
state links p, and T by:
Equation of State:
p = RT
(2.1)
We need to come up five additional equations linking the 6 properties. These five
equations are PDEs and turn out to be:
a)
b)
c)
d)
e)

Conservation of Mass or Continuity

Conservation of u- momentum
Conservation of v-momentum
Conservation of w-momentum
Conservation of energy

These equations may be derived using a Lagrangean approach, or an Eulerian

approach.
In the Lagrangean approach, we follow a fixed set of fluid particles (e.g. a cloud,
a tornado, tip vortices from an aircraft) and write down equations governing their motion.
This is somewhat like tracking satellites and missiles in space, using equations to
describe their position in space and the forces acting on them.
In the Eulerian approach, we look at a (usually) fixed or (sometimes) moving
volume in space surrounded by permeable boundaries. We develop equations describing
what happens to the fluid inside the control volume as new fluid enters and old fluid
particles leave. Eulerian approach is the preferred approach in most fluid dynamics
applications. This is what we will follow in our derivations.
Conservation of Mass (Continuity):
The conservation of mass stems from the principle that mass can not be created or
destroyed inside the control volume. Obviously, we are situations (e.g. nuclear reactions)
involving the conversion of mass into energy.
Let V be a control volume, a balloon like shape in space. We will assume that it, and its
surface S remain fixed in space. The surface is permeable so that fluid can freely enter in
and leave. The continuity equation says
The time rate of change of Mass within the control volume V =
Rate at which mass enters V through the boundary S

We can assume that the control volume V is made of several infinitely small
(infinitesimal) volume elements dV. The mass of the fluid inside each of these elements is
dV, where is the fluid density. The density is free to change from point to point, from
one sub-element dV to another within V. Thus,
Total mass within V =

dV
V

In the above equation, the three integral signs simply indicate that we are doing a threedimensional integration, or volume integration. The subscript V says that this integration
takes place inside V. This subscript may seem redundant at this time, but may be used to
distinguish three or more control volumes V1, V2, V3 etc. from one another in some
problems.
Then,
Time rate of change of mass within the control volume V =

d
dt

dV
V

In calculus, the order of integration or differentiation may be interchanged so that

these operations do not interfere with each other. Since our control volume is fixed in
space, the limits of the volume integral are not functions of time, and there is no
interaction between the two operations. Thus,
Time rate of change of mass within the control volume V =

t dV
V

(2.2)
Notice that we are using the partial derivative inside the integral since is a
function of (x,y,z,t) and we are only interested in its variation inside each sub-element dV
with respect to time, while the spatial location (x,y,z) of the sub-element dV remains
fixed.
We next turn our attention to the amount of fluid that enters V through the surface
S. For this purpose, we assume that the surface S is made of many quilt-like infinitesimal
patches dS. At the center of each element is a unit normal vector (i.e. a vector of length

unity, normal to the surface) n . By common convention, this normal is pointing away
from the surface dS. The normal component of fluid velocity
pointing towards the control

volume (entering the control volume) is given by V n . Notice the negative sign. We
are interested in the component of velocity pointing towards the control volume, not
away from it.
We can compute the rate at which mass enters the control volume through dS as
the product of density times normal velocity times area. Thus,

Rate at which mass enters the control voume through dS = - V ndS

Then,

Rate at which mass enters the control voume through the entire surface S = - V ndS
S

(2.3)

In (2.3) we are doing two-dimensional integration over a surface, hence the two integrals.
The little circle over the integral signs tells us that our surface is a closed surface. The
subscript S just reminds us that the integration is over the surface S.
Equating the relation (2.2), which is the left-hand side of continuity with the
relation (2.3), which represents the right hand side of continuity equation, we get:

t dV
V

V ndS
S

Bringing the expression on right hand side to the left, we get the conservation of mass
equation in integral form:

t
V

dV V ndS 0
S

(2.4)
Note that equation (2.4) holds for steady or unsteady, viscous or inviscid, compressible or
incompressible, two-dimensional or three-dimensional flow. It may be used in all speed
regimes from flow around a submarine in water to hypersonic flow over the space shuttle
during reentry.
Divergence Theorem:
From Calculus, one can derive the following equation (which hasnothing directly
to do with fluid mechanics) called the divergence theorem. Let F be any threedimensional vector, and is a general function of (x,y,z,and t). Then, divergence theorem
applied to a control volume V surrounded by a surface S states:

F ndS FdV
S

(2.5)

The inverted triangle is called a del operator. It is a collection of derivatives in the

x-, y- and z- directions arranged to form a vector: In Cartesian coordinate system this
operator is given by:

=i
j
k
x
y
z

In a cylindrical coordinate system this operator is given by

1
ez
er
e
z
r
r
When this operator operates on a function g(x,y,z,t) the result is called the gradient of g,
or simply del g. The gradient of a scalar in a Cartesian coordinate system is:

g
g g
g = i
j
k
x
y
z

Notice that the operation produces a vector.

Since the deloperator is a vector, it can
act onother vectors giving a dot product
or

cross product. The dot product between and F is called a divergence of F , and the

cross product between and F is called a curl of F . These operations are performed
as follows:

Let
Then,

Divergence

Curl of

F =

F1i

of

F =

F2

F3 k

F =

F =

F
x

F1

F2

F
y

z
F3

Assignment: We have shown these operations in a Cartesian coordinate system. Read the
relevant sections of Chapter II to find out how the divergence and curl operations are
carried out in the cylindrical coordinate system.
We return back to continuity equation (2.4) and the divergence theorem (2.5) from our
diversion into the land of vector calculus, a mystery for some and a nightmare for many.
Applying equation (2.5) to the surface integral in the continuity equation (2.4) we get:

t dV V dV
V

or,

dV 0

Consider the above volume integral. It must hold for any arbitrarily shaped
control volume V, at any instance in time for all flows. The only way this can be true is if
the integrand is zero. Therefore,

V 0
t

(2.6)
Equation (2.6) is called the differential form (PDE form) of continuity equation.
For steady flows, the time derivative vanishes. For incompressible flows, r is a constant.
Thus, continuity equation for incompressible flows becomes:

V 0
or ,
u
x

v
y

w
z

(2.7)

F
z

II. Governing Equations (Continued)

Derivation of the Momentum Equations
In this section we will derive the conservation of u-momentum equation, and
extend the resulting form to the conservation of v- and w- momentum equations in a
straightforward manner.
The u- momentum equation is an extension of Newtons law which states that
the rate of change of momentum of a particle (i.e. a system with fixed mass) is equal to
the force acting on the particle. In our work, we are dealing with open systems, i.e.
control volumes, whose mass may change with time as a result of fluid entering or
leaving the control volume. We therefore say
The rate of change of u-momentum of particles within a control volume = Forces along
the x-Direction acting on the control volume + Net rate at which u-momentum enters the
control volume through the boundaries.
As before, we consider a control volume V, surrounded by stationary boundaries

S. The outward-facing unit normal at any point on the surface S is called n .

Let dV be an infinitesimal volume within V. Then,
Mass with dV = dV
u - Momentum within dV = udV
u - Momentum within the entire control volume V =

udV
V

d
Rate of Change of u - momentum within V =
dt

udV
V

Interchanging the integration and differentiation operations for a fixed control volume,
Rate of change of u - momentum within V =

u
dV
t

See the previous handout titled Chapter II to find out what the triple integrals
mean, and why we could interchange the time derivative and spatial integration
operations.
We have now derived an expression for the left hand side of the conservation of umomentum equation. We now turn our attention to the right hand side, which consists of
two contributions X- component of forces acting on V, rate at which u- momentum
enters V through S. Let us consider these two issues one at a time.
Forces Acting on V: The forces may be divided into two broad categories body forces
and surface forces.
Body Forces: These are forces acting on every fluid particle within V. Examples of body
forces include gravity, electrostatic forces, and magnetic forces. Let us assume the

symbol a X represents the x- component of all these effects acting on the fluid particles
within V. The quantity a X may vary with x,y,z or t. Then,
Body force along the x - direction acting on all paticles wthin dV = a x dV
Body force along the x - direction acting on all paticles wthin V =

dV

Surface Pressure Forces: The control volume V is surrounded by the surface S. Pressure
from surrounding fluid (or solid) acts on this surface S. Pressure forces are always
directed towards the fluid within, and is always normal to the surface. The pressure

forces acting on S may therefore be written as pndS . Notice the negative sign. It is
S

there because the normal vector n is pointing outwards, whereas the pressure forces are
acting inwards. The x- component of these pressure forces is found by performing a dot

product of this expression with i .

Thus,

X component of Surface Forces acting on V= pn i dS

Surface Viscous forces: The surrounding fluid can exert an additional type of force on the
surface S, called a viscous force. This force may have a component normal to the
surface S, and components tangential to the surface. We will study the effects of viscosity
in more detail later. For now, we will call this contribution Fx-Viscous.
We can finally turn our attention to the rate at which the u- momentum is brought
into V though the surface S. Let dS be an infinitesimal element on S. Then rate at which

u-momentum enters the control volume through dS is uV ndS . Summing up

contributions over the entire surface S, we get:

Rate at which u - momentum enters V through S = - uV ndS
S

We have independently found all the terms on the left and right hand side of the umomentum equation. Bringing over all the terms, except the viscous forces and the bodyforces, to the left-hand side, we get:

dV pi uV ndS
t
S

x Body

dV Fx Viscous

(2.8)
The above equation is the conservation of u- momentum equation in integral
form.

We can similarly derive the v- and w- momentum equations. Or simply replace all

places where u appears with v, i with j , etc. Then we get

Conservation of v- momentum Equation in Integral Form:

dV pj vV ndS
t
S

y Body

dV Fy Viscous

(2.9)
Conservation of w- momentum Equation in Integral Form:

dV pk wV ndS
t
S

z Body

dV Fz Viscous

(2.10)
In this course, we will henceforth neglect the viscous forces acting on the fluid.
Needless to say, neglecting this effect is an approximation, and all the resulting solutions
only approximate real flows. Viscous forces are small except in regions of high velocity
gradient. In many flows, velocity varies slowly. Neglecting viscous effects turns out to be
an acceptable compromise.
We will also neglect body forces such as gravity, electrical and magnetic effects.
These effects are small in many applications compared to the rest of the terms in the
momentum equations. There are applications such as formation of weather systems where
gravity does play a part. In ionized flows, the ions are attracted and deflected by electrical
and magnetic forces. This is what causes the Aurora Borealis phenomenon! We neglect
these forces because they are small compared to other forces acting on the control
volume.
The effect of these approximations is that the right hand sides of the integral
forms of u- v- and w- momentum equations (2.8, 2.9, 2.10) disappear, leaving us only
with the left-hand side terms.

Chapter 2 - Continued...
PDE form of Momentum Equations
In the previous handout, we derived the integral form of the momentum
equations. In this section, we look at turning these integral expressions into PDEs.
We start with the u- momentum equation in integral form:


u
dV

u
V

n
dS

p
i

S
S ndS 0
t
V

(2.11)
Divergence theorem states that any surface integral can be converted into a volume
integral using:

A ndS AdV
S

(2.12)
where A is any vector. Using this relationship to the surface integrals in the u-momentum
equation, we get

uV n dS uV dV

pi n dS

pi dV

(2.13)
The u-momentum equation thus becomes:

u
uV pi dV 0
t

(2.14)
The only way this volume integral will be equal to zero for all arbitrary control volumes
at all time levels is if the integrand is zero. We therefore get:

u
uV pi 0
t

(2.15)
Let us do some of the dot products involving the 'Del' operator and the vectors inside the
parentheses.

uV u ui vj wk u 2 i uvj uwk

uV i j k u 2 i uvj uwk u 2 uv uw
x
y
z
x y z
p
pi i j k pi
x
x y z

(2.16)

With these operations, equation (2.15) becomes:

u u 2 uv uw p 0
t
x
y
z
x

(2.17)
Equation (2.17) is the PDE form of the u- momentum equation, while equation (2.11) in
the integral form.
We can similarly convert the v- and w- momentum equations into PDE forms. Sparing
you the details, here are the final forms:
v- Momentum equation in PDE form:

v uv v 2 vw p 0
t
x
y
z
y

(2.18)
w- Momentum Equation in PDE form:

w uw vw w 2 p 0
t
x
y
z
z

(2.19)
You need not memorize these equations. You should, however, be able to recognize them
and identify them if you are given the expressions.
Equations (2.17) - (2.19) may be simplified further. Consider the various terms in
equation (2.17). The derivatives are products of several terms. We can use the product
rule. That is,


u u u
t
t
t
2
u
uu
u
u

u
u
x
x
x
x
uv
u
v
v
u
y
y
y
uw
u
w
w
u
z
z
z

(2.20)

Using these expansions (2.20) in equation (2.17), we get:

u
u v w
u
u
u
p
u
v
w
u

0
x
y
z
x
x
y
z
t
t

(2.21)
The term inside the square bracket is the continuity equation and vanishes. The umomentum equation in PDE form thus becomes:
u

u
v
w
0
u-Momentum Equation:
x
y
z
x
t

(2.22)
By a similar process, the v- and w- momentum equations become:
v

u
v
w
0
v-Momentum Equation:
x
y
z
y
t

(2.23)
w

u
v
w
0
w-Momentum equation:
x
y
z
z
t

(2.24)
Note that equations (2.22)-(2.24) may be written as:
Du p

0
Dt
x
Dv p

0
Dt
y
Dw p

0
Dt
z
where,

u
v
w
Dt
t
x
y
z

The operator D/Dt is called the "substantial derivative" or "material derivative".

As discussed in the text (section 2.10) this represents derivative of the property (u,v,w,

etc.) of a fluid particle with respect to time, as the fluid particle moves in space. You
should remember and recognize equation (2.22, 2.23, 2.24).
The physical significance of D/Dt may be explained as follows. Consider a fluid
particle at a point A in space, given by the coordinates (x,y,z). A short time later this
B
(x+x,y+y,z+z,t+t)

A
(x,y,z,t)

particle has moved to a new location B given by the coordinates (x+x, y+y, z+z).
Then, the changes in a property, say u, between A and B may be written as:

u ( x x, y y ), z z , t t ) u ( x, y, z , t )

u
u
u
u
t
x
y
z
t
x
y
z

u u x u y u z
t

x t
y t
z t
t
u u
u
u

u
v
w t
x
y
z
t

Thus,
Du

Dt

Limit
t 0

u B u A u u
u
u

u
v
w
t
x
y
z
t