Fluids Dynamics Notes

Fluid Mechanics MECH1001 1
University College London Dr. V. Daz

FLUID MECHANICS

MECH1001

Dr. Vanessa Daz

Department of Mechanical Engineering
University College London



Chapter 1 Introduction to Fluid Mechanics

What is fluid mechanics?
Why do we have to study it?
Why should we want to study it?
How does it relate to subject areas with which we are already familiar?

I. Introduction

Fluid mechanics is concerned with understanding, predicting and controlling the
behaviour of a fluid.
Since we live in a dense gas atmosphere on a planet mostly covered by liquid, a
rudimentary grasp of fluid mechanics is part of everyday life.
1. Examples
Fluid mechanics = important field of the engineering with many practical and exciting
applications:
- Municipal water, sewage and electrical systems: heavy dependence on fluid
machinery (pumps, steam turbines, valves, piping)
- Aerodynamics: aircrafts, automobiles, ships, spacecrafts involve interaction with
fluids (outside, inside engine, hydraulic control system, etc.)

Learning about fluid mechanics allows us to better understand our bodies and many
interesting features of the environment.
- The heart and lungs, for example, are wonderfully designed pumps that operate
intermittently rather than steadily.
o The heart moves blood efficiently through the branching network of arteries,
capillaries and veins.
o The lungs cycle air through the branching pulmonary passages, thereby
keeping the cells of our body alive and functioning.
- The environment is another source of complex and interesting fluid mechanics.
o Prediction of weather, hurricanes and tornadoes
o Spread and control of air and water pollution
o Flow of rivers, streams; movement of groundwater, jet stream and ocean
currents, tidal flows in estuaries.
o Lava flows of volcanoes


o Movements of molten rock within the earth
- Beyond the Earth: stellar processes and interstellar events: striking examples of
fluids in motion on a grand scale.
- Curving flight of a tennis, golf or rugby balls, different pitches in baseball.

2. Statics/Dynamics
Fluid mechanics is historically divided into two branches: fluid statics and fluid dynamics
- Fluid statics, or hydrostatics, is concerned with the behaviour of a fluid at rest or
nearly so. (picture of a dam)
- Fluid dynamics involves the study of a fluid in motion (picture of a rocket)

Engineering: ability to create and solve mathematical models of physical systems.
Fluid mechanics is seen as a challenging subject, because the underlying mathematical
model appears to be complex and difficult to apply.
- The governing equation of fluid statics, the hydrostatic equation, is actually
simple solved to find the pressure distribution in a fluid
- However, the governing equation of fluid dynamics, Navier-Stokes equation is
not easy to solve. Since Leonardo da Vinci, scientists have developed various
approaches to solve specific fluid flow problems. The common theme is to simplify
the mathematical or experimental model to describe the flow without sacrificing the
relevant physical phenomena.

Ex: approximation for low speed air-flow: neglect the air compressibility.
Accurate for speeds as high as 400 km/h!
[Do you consider 400 km/h a low speed flow? It is in fluid mechanics.]

3. Art of Fluid Mechanics
The art of fluid mechanics is developed primarily through experience. It consists in
knowing when it is safe to neglect the effects of physical phenomena that are judged to
have little impact on the flow
If we neglect certain physical phenomena, we can drop the corresponding terms in the
governing equations
We decrease the difficulty in obtaining a solution.

Ex of simplification: Fluid statics = simplification of fluid dynamics.

- Fluid mechanics problem


- Appropriate mathematical model
- Resolution of model to find solution
o Analytical solution exists (rare): representation of the flow variables as
functions of space and time
o Use of computational methods to solve the equations of fluid motion (finite
difference, finite element, finite volume, etc.). Need to discretize the
equations
- Experimental methods: empirical results (used extensively in the design of piping
systems, pumps, turbines, engines

II. Definition of a fluid

Definition: A fluid is a substance that deforms continuously under the action of a shear
(tangential) stress.
A F / =
F = tangential force [N]
= shear stress [N/m
2
= Pa]
A = surface area [m
2
]

Solid: Changes shape under the application of shear stress. However, a solid quickly
returns to rest and retains the new shape as long as the stress is maintained.

For a shear stress below the elastic limit, the shear stress, , is related to the shear strain,
= x / y, by the shear modulus of the solid, G:
G =
Definition of shear strain: displacement in the direction of the applied force, x,
normalized by the height of the solid, y.
y x = /

Ex: G
steel
= 12 10
6
psi; G
aluminium
= 3.8 10
6
psi
y

= F / A
Figure 1.1.


For similar stresses applied, the displacement in aluminium block
would be more than three times as larges as that in the steel block.

Fluid: Cannot withstand shear stress: continues to deform.

Fluid resists being sheared, but its underlying molecular structure does not allow it to
prevent the resulting deformation the fluid is set in motion.
A fluid exhibits a relationship between shear stress and the shear strain rate d / dt.
For many common liquids and all gases, the relationship is linear, the dynamic viscosity,
, being the constant of proportionality
dy
du
dt
d
= = (Newtons law of viscosity)

The shear rate and velocity gradient are equal d/dt = d/dt (x/y) = du/dy.
A fluid that follows this equation is called a Newtonian fluid.

III. Types of fluids

Fluids comprise the liquid and gas (or vapour) phases of the physical forms in which
matter exists.

Liquids: takes the shape of its container, but retains a constant volume. The liquid can
form a free surface in a container whose volume is greater than the liquid volume.

Liquids are treated as incompressible: the intermolecular spacing in a liquid is
essentially constant.

y
= F / A
u(y)
U
Figure 1.2.
Figure 1.3.


Gases: takes the shape of the container and expands to fill the region completely.
Gases are compressible: the distance between molecules in a gas is much greater and
highly variable.

Gases can behave as incompressible when in motion if the flow speed is much less than
the speed of sound in the gas (a ~ 330 m/s in air at sea-level).
It corresponds to very small Mach numbers
a
u
M = << 1

IV. Methods of description
1. Continuum hypothesis
Continuum hypothesis: The study of fluid mechanics considers the fluid to be a
continuous medium, i.e. the molecular structure of the fluid is ignored.
Ex: To predict the fluid velocity distribution created by stirring a glass of
water with a spoon, one does not model the water molecules as interacting
hard spheres: they are on the order of 10
26
molecules!

The underlying molecular structure is replaced by a limited set of fluid properties,
defined at each point in the fluid at every instant.
Mathematically speaking, the continuum hypothesis allows the use of differential
calculus in the modelling and solution of fluid problems. Each fluid property, such as
density, is considered to be a continuous function of position and time. (There may be
discontinuous jumps in a property value at fluid-fluid or fluid-solid interfaces.)

Limit: the assumption is not valid anymore if the length scale of a physical phenomenon
is of the same order as molecular dimensions.
Ex: In a gas, the largest molecular dimension is the mean free path
(inversely proportional to the density): ~ 10
-7
m or 0.1 m.
Hypothesis is not valid for a flow through a 0.2 m pore.

2. System and control volume
To construct a theory of fluid mechanics, it is necessary to represent the basic laws
describing the forces and motion:
- Equation of conservation of mass (mass = constant)
- Equation of conservation of momentum (Newtons second law of motion)


+ = =
surf
body F F F a m
r

Definition: The momentum is u m p
r r
=

In Thermodynamics, the focus is energy:
- Equation of conservation of energy

The first step in solving a problem is to define the system that is being analyzed.

Definition: A system is defined as a fixed, identifiable quantity of mass. The system
boundaries separate the system from the surroundings. The boundaries of the system may
be fixed or movable, but no mass crosses the system boundaries.

The system is used in Mechanics, but in Fluid Mechanics we are concerned with the flow
of fluids through devices (turbines, pipes, etc.)
Difficult to isolate a fixed identifiable quantity of mass
More convenient to look at a volume in space through which fluid flows = Control
volume.

Definition: A control volume (CV) is an arbitrary volume in space through which fluid
flows. The geometric boundary of the control volume is called the control surface (CS).

3. Lagrangian description
In the Lagrangian description, the fundamental entity is a fluid particle. We assume that
the fluid consists of a continuous distribution of small discrete particles of fluid, each of
Cylinder
Piston
M
Gas
System
boundary

Figure 1.4.
Control surface
Control volume

Figure 1.5.


which has a fixed mass but is otherwise shapeless. The particles are tracked to arrive at
the motion of the bulk fluid.

Ex: Imagine standing on a bridge and spraying small droplets of a red dye
in a river. The particles follow the movement of the water. As your eyes
follow the droplets downstream, you are tracking a number of marked
fluid particles.
2
2
dt
r d
m
dt
u d
m a m F = = =

The Lagrangian description is typically used in Solid Mechanics.

4. Eulerian description
In the Eulerian description, we describe physical space by means of a coordinate
system that serves as a backdrop for fluid motion.

Ex: Instead of using dye to mark the movement of fluid particles, suppose
we use a fixed array of flow meters to measure a simultaneous set of fluid
velocity vectors at different points in the river. We focus at a certain
location, where we observe the velocity.

The fluid pressure at a certain point at a certain instant corresponds to the pressure
associated with the fluid particle that coincides with that spatial point at that time.

The Eulerian approach constitutes a field description of the fluid (seen in calculus). Each
property of the fluid depends on time and space: velocity and acceleration fields
(vectors), pressure, density and temperature fields (scalars)

Pressure sensor
Figure 1.6.
12.345


V. Dimensions and unit systems
1. [MLtT] systems
There are many fluid properties, many with distinctive units. An engineer must be
capable of knowing their dimensions.

Definition: A dimension is a physical variable used to specify some characteristic of a
system.
Ex: Mass [M], length [L], time [t], temperature [T].

Definition: A unit is a particular amount of a physical quantity or dimension.
Ex: length measured in meter (m)

A base dimension set is [MLtT]: enough to describe most problems in fluid mechanics
(e.g. not electrical or magnetic fields).

2. Unit systems
Systme International dUnits (SI): the base units for mass, length, time and temperature
are respectively the kilogram (kg), meter (m), second (s) and Kelvin (K).

Other unit for forces is the Newton: 1 N = 1 kg.m/s
2

Each unit is specifically defined (e.g. a meter is defined as 1,650,763.73 wavelengths of
the radiation corresponding to the transition between the electronic energy levels 2p
10
and
5d
5
of the krypton-86 atom in a vacuum).


Chapter 2 Fluid Properties

To address the interesting flow problems in biology, the environment, mechanical
devices, transportation systems, athletic equipment, etc, we must agree on a common
language and set of symbols to use in defining fundamental fluid properties and their role
in fluid mechanics.
The density, pressure, temperature and viscosity of a fluid are flow properties: their value
depends on the nature of the fluid and type of flow. Remember that all fluid properties
are variables (not constant): they are functions of position and time.

Note carefully that in fluid mechanics a few symbols are used to represent more than one
property. This is unavoidable, but usually creates no confusion because the context of a
problem makes it clear to which property the symbol refers.

I. Mass, weight and density

- Definition: The mass of an object is defined as a measure of its resistance to
acceleration.
For an object of fixed mass, the relationship between acceleration, a , force, F , and
mass, m, is given by Newtons second law:
a m F =
- Definition: The weight of an object, W
r
, is the magnitude of the force acting on the
object due to Earths gravity field, g
r
. Weight defined by Newtons equation
g m W
r
r
=

- Mass and weight depend on the amount of material (fluid in a swimming pool or
teacup).
Definition: Properties of a fluid that depend on the amount of fluid in a system are called
extensive properties.
If an extensive property is divided by the total mass in a system, the result is an intensive
property, i.e. property per unit mass.

In Fluid Mechanics, where the total mass of fluid is usually not relevant, we use
intensive properties defined per unit volume.


- Definition: The density of a fluid, , is determined by dividing the mass by its volume,
V.
V
M
= with ) , , , ( t z y x =
To determine the density at point C, we select a small volume V of mass m. The
density at a point is then

We will see that in a liquid, density normally remains nearly constant throughout the
region of flow. However, in gases, density often varies significantly.

Heavy = higher density
Ex: Mercury is heavier than water, water is heavier than oil, seawater is
heavier than fresh water, cold air is heavier than hot air.

II. Temperature and other thermal properties

The temperature T of a fluid is a thermodynamic state variable that provides a measure of
the internal energy of the fluid.
The density, pressure and temperature of a fluid are related to one another by an equation
of state. For many gases, the state equation is well approximated by the perfect gas law.
RT
P
=

III. Viscosity

Where do stresses come from?
For a solid, stresses develop when the material is elastically deformed or strained.
For a fluid, shear stresses arise due to viscous flow.
Solids are elastic, fluids are viscous

Definition: The viscosity of a fluid is defined as the constant of proportionality between
the shear stress and the transverse velocity gradient (rate of deformation).
C
m,
V

x
z
y
V
m
V
0
lim
=


The viscosity, , is defined as the dynamic viscosity.
A different way to express the viscosity is to divide it by the density. This normalized
form is called the kinematic viscosity,
=
An important consequence of the existence of shear viscosity is a loss of energy when
the fluid is sheared = FRICTION (loss due to viscous dissipation)
The general action of viscosity is to convert useful energy into heat (increase in T).

23 . 1 =
air
kg/m
3
,
5
10 73 . 1

=
air
Ns/m
2
,
5
10 4 . 1

=
air
m
2
/s
1000 =
water
kg/m
3
,
3
10
=
water
Ns/m
2
,
6
10
=
water
m
2
/s

IV. Pressure
1. Surface forces
All fluids are composed of energetic molecules in motion. When these molecules collide
with a surface, they exert a normal and tangential force on the surface due to the change
in momentum of colliding molecules.

surf
F Axial stress (shear) = tensile stress (only if fluid is moving)
Normal stress (pressure) = compressing stress (fluids at rest or moving)

Importance of surface forces (see Newtons law):

+ = =
surf body
F F F a M

Definition: The pressure on a planar surface is defined as the compressive normal
stress. The normal force
N
F
r
to a surface of surface area A is

= =
A A
N
dA n p A pd F
r
r r

p

n
r



- The pressure p is an absolute pressure, which may take any positive value. It is
defined and measured in reference to a perfect vacuum.
- In a vacuum, there are no fluid molecules to exert a force on a surface the
absolute pressure in a perfect gas is zero.
- However, most measurements are made in reference to the ambient pressure
(atmospheric pressure). This measurement is referred to the gage pressure.
The standard atmospheric pressure at sea level is 1atm = 101,300 Pa.
p
abs
= p
gage
+ p
atm

2. Is pressure directional?
Pascals law:
At a point, the pressure intensity is independent of the direction of the surface used to
define it. Pressure is a scalar quantity.

Proof: Lets consider a small wedge around a point of interest. The wedge is small
enough to neglect its weight.

Applying Newtons second law to the small wedge of fluid at rest:
Fluid at rest no acceleration
- Forces along x: 0 cos sin 5 . 0 cos sin 5 . 0
2
5
2
4
= L p L p
5 4
p p =
System pressure
Perfect vacuum pressure
Atmospheric
pressure
Gage pressure
Absolute
pressure
Instruments measure gage pressure
Figure 2.1.
C
C
B
B
D
A
B
B
A
L sin

L L cos

L
p
2

p
1

p
3

y
x
z

Figure 2.2.
p
4

p
5



- Forces along y: 0 sin sin
2
3
2
2
= L p L p
3 2
p p =
- Forces along z: 0 cos cos
2
3
2
1
= L p L p
3 1
p p =
Therefore, the pressure is not directional.



Chapter 3 Fluid Statics

For a fluid at rest, the shear stress is zero only normal stress is present (i.e. pressure).
In this chapter, we will study fluid statics, also called hydrostatics:
Hydro: water; static: at rest

I. The basic equation of fluid statics

The objective is to obtain an equation to compute the pressure field in a static fluid.
From everyday experience, we can deduce that pressure is function of depth (increases
with depth).
Ex: The deeper you swim or dive, the greater the water pressure.
The pressurized cabin of an aircraft only maintains a certain minimum
pressure at altitude. Since the pressure is much less than at sea level, the
ears may pop during the pressure changes at takeoff and landing.

- Lets apply Newtons second law to an infinitesimal cylinder of fluid of length x and
surface area A. The only forces applied to the element are the pressure forces at x and
x+x
0 A x
x
p
p A p = |
\
|
+
Since 0 x and 0 A , and since it is similar in the y-direction:
0
x
p
=
; 0
y
p
=

- Lets consider an infinitesimal fluid element in the z-direction. Net force equilibrium:
p

z+z

A
W

z

z
p

A
x

x
x+x
p + p x

x
p + p z

z
Figure 3.1.


0 z A g A z
z
p
p A p = |
\
|
+
g
z
p
=

Conclusions:
For any fluid at rest in Earths gravity field, there is no pressure gradient in the x- and y-
directions:
p is independent of x and y.
p = f (z): isobars are horizontal in a constant gravitational field.
The pressure distribution satisfies
g z
dz
dp
) ( =
where ) (z is the density distribution.

This is a first-order ordinary differential equation, so we must specify one boundary
condition: typically a known value, p
0
, for the pressure at a certain elevation z
0
.
The solution to this equation is obtained by separating variables and integrating from z
0
to
z:
' ) ' ( '
0 0

=
z
z
p
p
gdz z dp
' ) ' ( ) (
0
0

=
z
z
gdz z p z p
Since the density occurs inside the integral, the nature of the fluid involved in a
hydrostatics problem is important: one needs to know the function ) (z .

II. Constant density fluids in equilibrium

For most liquids, the density is a constant. The hydrostatic pressure distribution in a
constant density fluid is:
) ( ) (
0 0
z z g p z p =
Defining the height of fluid, z z h =
0

gh p p + =
0



Implications of the result: The pressure distribution does not depend on the shape of the
container. Isobars are horizontal in a fluid at rest
p
a
= p
b
= p
c
= p
d

p
A
= p
B
= p
C
( p
D
)

III. Manometry

The direct proportionality between gage pressure and h for liquids enables pressure to be
visualized in terms of a column of liquid above the point of interest whether or not it is
a free surface.
1. Barometer
This is the principle of the mercury filled device long used to report the barometric
pressure in Earths atmosphere.
Mercurys vapour pressure is so low that
gH p
H g p p H z p
atm
atm V
=
= = = ) 0 ( ) (

water
a b c d
mercury
A B C D
Figure 3.3.
Free surface
z
h
x
z0
p0
Figure 3.2.
pV ~ 0
patm
Hg
H
z
Figure 3.4.


By measuring the height of the column, we measure the absolute pressure in the
atmosphere.
Pressures in fluids can be expressed in mm of Hg (cardiac pressure) or
mm of H
2
O. 1 atm = 762 mmHg = 33.95 ft H
2
O

2. U-tube manometer
gH p p
gh p p
B atm B
A A
+ =
+ =

Since
B A
p p =
gage A B atm
p gh gH p p = =
If A is a gas, then gH p
B gage
= , since
A
<<
B

For pressure differences caused for example by an orifice plate in a pipe, the same
principle applies (p
A
= p
B
):
gH gh p gH gh p
B A A A
+ + = + +
2 1

gH p gH p
B A
+ = +
2 1

gH p p
A B
) (
2 1
=

H
Fluid B
B

Fluid A
A

p
h
p p
atm

p
B

p
A

Figure 3.5.
Fluid B
B

Fluid A
A

p
1

h
H
p
2

p
A

p
B

Figure 3.6.


3. Piezometers (for fluids only!)
A piezometer is a small tube which is open to the atmosphere and of sufficient length as
to accommodate a free surface:
g a h p g a h p ) ( ) (
2 2 1 1
+ = +
g h p g h p
2 2 1 1
=
g h h p p ) (
2 1 2 1
=

4. Reservoir manometers
One limb of the manometer is enlarged to reduce the meniscus on that limb to make
readings simpler.
Suppose the area of the reservoir is A
R
and the area of the pipe is A
P
. When a pressure is
applied, the level in the reservoir will drop by
R P
A hA / , h being the height in the pipe,
which is negligible (A
R
~ 400A
P
).

5. Other manometers
- Inclined reservoir manometers
The precision of a pressure measurement with a U-tube manometer depends to some
extent on the type of liquid used and the effects of surface tension in causing the
meniscus to form in the tube at the free surface. The precision can be greatly improved by
using an inclined manometer

- Inverted U-tube manometers

- Other methods to measure the pressure?

Research for yourself

p
1
p
2

h
2

h
1

Figure 3.7.


IV. Variable density fluids in equilibrium

The general condition of equilibrium for a fluid is
g z
dz
dp
) ( =
Typically for a gas, the density varies with the temperature. For a perfect gas
RT
p
=
g
RT
p
dz
dp
=

We know that p = f(z), but is the temperature also a function of the altitude?

1. Isothermal gas
If T = constant = T
0
, the equation becomes
dz
RT
g
p
dp
0
=

=
z
z
p
p
dz
RT
g
p
dp
0 0
0

) ( ln ln
0
0
0
z z
RT
g
p p =
(

=
0
0
0
) (
exp
RT
z z g
p
p

2. Linearly decreasing temperature
If T = T
0
- z (constant temperature lapse rate), the equation becomes
z
, T, p
O, TO, pO
Figure 3.8.


z T
dz
R
g
p
dp

=
0

=
z
z
p
p
z T
dz
R
g
p
dp
0 0
0

[ ] ) ln( ) ln( ln ln
0 0 0 0
z T z T
R
g
p p
=
(
=
0 0
0
0
ln ln
z T
z T
R
g
p
p
R
g
z T
z T
p
p
(
=
0 0
0
0

3. Limits for small altitudes
Lets assume z
0
= 0. For a small altitude z, both results provide the same value for the
pressure distribution.
Isothermal:
... 1 ... 1 exp
0
0
0
0
0 0
+ = +
(
p
gz
RT
gz
RT
gz
p
p
z

...
! 3 ! 2
1
3 2
0
+ +
z z
z e
z
z
: series expansion
Constant temperature lapse rate:
... 1 ... 1 1
0
0
0
0
0 0
+ = +
(
p
gz
RT
gz
T
z
p
p
z
R
g

...
! 3
) 2 )( 1 (
! 2
) 1 (
1 ) 1 (
3 2
0
+

z n n n z n n
nz z
z
n
: series expansion

Thus, for small heights of fluid, gz p p
0 0
= , which is the result we derived for
constant density fluids. But what is meant by small? Lets consider the atmosphere:
Small is actually quite large. The atmosphere can be predicted to behave
like a constant density fluid up to about 300 m, using any practical
assumption regarding the temperature variation.



The lower atmosphere was measured to have a lapse rate about 0.0065 K/m up to about
11 km. From 11 km to 20 km, the temperature is constant about -55.5C and beyond that,
the temperature rises again.

In dealing with gases on a length scale of only a few meters, the elevation term in the
expression for p is negligible (gz ~ 1.23 9.81m/s
2
1m ~ 12 Pa << 1 atm). In these
cases, the pressure can be taken as constant p ~ p
0
.

V. Hydrostatic force on submerged surfaces

Now that we have determined the manner in which the pressure varies in a static fluid,
we can determine the resultant force acting on a submerged surface. We must specify
- The magnitude of the force.
- The direction of the force.
- The line of action of the force.
T
0

T
z (km)

50

20
11
troposphere
stratosphere
Figure 3.10.
mesosphere
Figure 3.9.


Applications:
Design of dams, sluice gates, etc.

General approach:
- Draw the system
- Define coordinate axis (choose a suitable origin)
- Define an elementary surface area in the system: dA
- Assess the elementary forces and moments acting on the element: F d
r
and M d
r

A d p r F d r M d
A pd F d
r
r
r
r
r
r r
) ( = =
=

The moment is the cross product of the moment arm and the force.
The vector area element dA n A d
r
r
= has a magnitude dA and is directed along the outward
normal unit vector n
r
.

- Integrate over the surface to find the total force and moment.

= =
= =
A
A
A d p r M d M
A d p F d F
r
r
r r
r r r
) (
) (

Example: Consider a rectangular gate of width b and height a hinged about its upper
edge and set in the wall of a tank as shown

Elementary force on the gate
z z atm atm z
e dxdy y h g e dxdy p y h g p e dF dF F d
r r r
r
) ( ] ) ( [ ) (
2 1
+ = + + = =
Elementary moment on the gate
x x
e ydxdy y h g e dM M d
r r
r
) ( + = =

F F = = ? ?
d d = = ? ?
Figure 3.11.


Total force on the gate
[ ] [ ]
z z
a
b
z
a b
gate
e a h gab e y hy x g e dy y h dx g F d F
r r r
r r
) 2 / ( 2 / ) (
0
2
0
0 0
+ = + = + = =

Total moment on the gate
[ ] [ ]
x x
a
b
x
a b
gate
e a h b ga e y hy x g e ydy y h dx g M d M
r r r
r r
) 3 / 2 / ( 3 / 2 / ) (
2
0
3 2
0
0 0
+ = + = + = =

Centre of force: The distribution of pressure is equivalent to a unique force acting at a
point at y
' Fy M =
2 /
) 3 / 2 / (
'
a h
a h a
F
M
y
+
+
= =
For a very shallow immersion (very small h), y tends towards 3a/2.
For a very deep immersion (very large h), y tends towards a/2.

O
p
atm

y

z

h

p
atm

dF
1

dF
2

y

x

dy
b

a

dx
Figure 3.12.


VI. Buoyancy and Archimedes principle

Archimedes principle states that a buoyancy force, F
buoyancy
, acts at the centre of mass
in the direction opposite to that of the gravitational force (weight), W, and has a
magnitude equal to the weight of the displaced fluid.

The increase in hydrostatic pressure with depth in a fluid creates a net force on an
immersed object. The net vertical force acting on an object due to hydrostatic pressure is
called the buoyancy force (up thrust).

The net force on the submerged object (in the z-direction) is
F
net
= F
buoyancy
W

gV ghA A p p F
buoyancy
= = = ) (
1 2

Ex: The hot air balloon: The buoyancy force acts on a heated volume of fluid immersed
in the same fluid at a colder temperature. It is therefore
(
cold
-
hot
) g V

Be careful: What is the buoyancy in a box submerged in a fluid, resting on the bottom of
the tank?

VII. Stability of immersed and floating bodies
1. Immersed bodies (e.g. single driver submersible)
An angular displacement sets up a restoring couple on the vessel of magnitude mgd.
Thus the craft is stable provided G is below B.
Figure 3.14.
W
z
p1
h
R
W z
p1
p2
h
Figure 3.13.


2. Floating vessels

An angular displacement causes B to move to B, the centroid of the new displaced
volume.

- M is called the metacentre. It is the point through which the buoyancy force
appears to act when a vessel is heeled over through small angles.
- GM is referred to as the metacentric height.

The restoring couple on the vessel is now mg GM sin.
Thus the craft is stable provided M is above G.

(Cross-section of
submerged craft)
mg
F
B
B
G
d
In equilibrium
F
B
= mg
B
G
mg
F
B

Pressure
hull
Propulsion and
life support
Figure 3.15.

G
In equilibrium
F
B
= mg
B
G
F
B
mg
B
F
B
mg
M
GM sin
Figure 3.16.


VIII. Equilibrium of accelerated masses
1. Velocity field
In the Eulerian description, the fluid is not thought of as consisting of fluid particles,
but rather as an indivisible, continuous material that moves with a velocity u
r
under the
action of external and internal forces.
The fundamental variable is the velocity vector, which is function of space and time.

Cartesian coordinates:
z y x
e t z y x w e t z y x v e t z y x u u
t z y x u u
r r r r
r r
) , , , ( ) , , , ( ) , , , (
) , , , (
+ + =
=

Cylindrical coordinates:
z r r
e t z r w e t z r u e t z r u u
t z r u u
r r r r
r r
) , , , ( ) , , , ( ) , , , (
) , , , (

+ + =
=

2. Constant density fluid in rectilinear acceleration
The structural design of the fuel tanks of a liquid-fuelled rocket must take into account
the pressure forces developed during acceleration. Similar considerations apply to
storage tanks for fuel and lubricants in automobiles and aircraft, where acceleration may
cause the free surface of a liquid to tilt (the inlet of a pump may run dry momentarily).

u
r

u
x
r
u
y
z
r

w
O
O
Cartesian coordinates Cylindrical coordinates
O
u
r
u
z
Figure 3.17.

O
O
u
r
x
y
u
w
v
a
x
a
x
Fluid-filled
container
Free surface
Figure 3.18.


Although it is unusual to maintain a constant rectilinear acceleration for any length of
time, the hydrostatic pressure distribution caused by the acceleration is established almost
instantly and adjusts itself rapidly to a changing acceleration. We, thus, assume that the
liquid is at rest with respect to its container and focus on calculating the shape of the free
surface.

Assumption: no relative motion between fluid particles (i.e. no flow within fluid mass).

Equilibrium conditions on infinitesimal elements (similar to section I) for a fluid
under x-acceleration:
- Lets apply Newtons second law in the x-direction (applied to an infinitesimal
cylinder of fluid of length x and surface area A.
The only forces applied to the element are the pressure forces at x and x+x:
xA a A x
x
p
p A p
x
= |
\
|
+
x
a
x
p
=

- Net equilibrium in the y-direction:
0 =
y
p

- Net force equilibrium in the z-direction:
0 z A g A z
z
p
p A p = |
\
|
+
g
z
p
=

Therefore p = f(x,z).
gdz dx a dz
z
p
dx
x
p
dp
x
=
=
The pressure varies linearly in each direction.
An isobar is defined as a line of constant pressure: dp = 0. Therefore the isobars have the
following expression
g
a
dx
dz
x
const p
=
|
\
|
=

The slope of the isobars are


|
|
\
|
=
g
a
x
arctan

3. Constant density fluid in rigid rotation
When a container first begins to rotate, the fluid tends to remain at rest, but a thin layer of
fluid near the wall is dragged around with the wall as a result of the effects of viscosity
and the no-slip condition.
As the container continues to rotate, however, more and more fluid gradually begins to
rotate.
After a sufficient time, viscous effects cause the entire volume of fluid inside the
container to rotate along with it as a rigid body.
When a rigid body rotation is achieved in a fluid, all shear stresses vanish and there is no
deformation of fluid elements. The fluid is at rest relative to the rotating cylinder with a
hydrostatic pressure distribution.

Equilibrium conditions on infinitesimal elements (similar to section I):
- For equilibrium of a small element of fluid in the r-direction, rotating at a constant
angular velocity, there must be a pressure gradient to produce the centripetal
acceleration u
2
/r towards the centre.
rA
r
u
A r
r
p
p A p
2
=
|
\
|
+
Fluid-filled
container

r
u

O
(PLAN VIEW)
z
r

O
Figure 3.20.
a
x
a
x
Isobars

Figure 3.19.


2
2
= =
r
r
u
r
p

- Net force equilibrium in the z-direction:
g
z
p
=

Therefore p = f(r,z).
gdz dr r dz
z
p
dr
r
p
dp =
=
2

1
2 2
2
1
C gz r p + =
It can be deduced that isobars are paraboloids:
3
2
2
C r C z + =

Lets consider the case of a free surface:

We know that p = p
atm
at the free surface. Let z
0
be the level at the centre.
1 0
C gz p
atm
+ =
0 1
gz p C
atm
+ =
) (
2
1
0
2 2
z z g r p p
atm
+ =
At the free surface, the equation of the paraboloid is
g
r
z z
2
2 2
0
+ =

r
z
O
r
z
p
atm
Isobars
z
0
O
Figure 3.21.


Chapter 4 Incompressible inviscid flow

In Chapter 5, we will develop the basic equations in integral form for a control volume.
Integral equations are useful when we are interested in the gross behaviour of a flow field
and its effect on various devices. However, the integral approach does not enable us to
obtain detailed point-by-point knowledge of the flow field. To do so, we must apply the
equations of fluid motion in differential form.
Next year, you will see the derivation of the differential equations for the conservation
of mass and Newtons second law of motion (momentum conservation).

The difficulty comes when trying to solve analytically these equations. Only in a few
simple configurations, can an analytical solution be found. Otherwise, the equations need
to be solved numerically (computational fluid dynamics, CFD).

In this chapter, we shall derive these differential equations along one streamline in the
simplified case of an inviscid fluid flowing steadily. Although truly inviscid fluids do
not exist, many flow problems (mainly in aerodynamics) can be successfully analyzed
with the approximation that = 0.

I. Representation and description of movement
1. Classification of flows
Unsteady 3D flow: Flow dependent on space (e.g. x, y, z) and time t.
) , , , ( t z y x u u
r r
=
Steady flow: Independent of time: 0 / = t
) , , ( z y x u u
r r
=
2D flow: Independent of one space coordinate (e.g. y: 0 / = y )
) , , ( t z x u u
r r
=
1D flow: Independent of 2 space coordinates (e.g. y, z: 0 / = y and 0 / = z )
) , ( t x u u
r r
=


2. Definitions
Streamlines: Line that is tangent to the fluid velocity vector at any point (tangent dy/dx =
v/u) No flow across a streamline
Streamlines may change from instant to instant in an unsteady flow.

Pathlines: Trajectory of an individual fluid particle (coincident with streamline in a
steady flow)

Streamtubes: Imaginary tube (not necessary circular) in a flow whose walls consist of
streamlines. Flow enters and leaves along through the ends.

3. Fluid acceleration (substantial derivative)
The Eulerian acceleration ) , , , ( t z y x a
r
at a point ) , , , ( z y x at time t must be identical to the
acceleration of the fluid particle located at his point.
Thus we can calculate the acceleration by following the motion of the particle

) , , , ( t z y x u u
old
r r
=
) , , , ( t t z z y y x x u u
new
+ + + + =
r r

x+x,y+y,z+z
x,y,z
Figure 4.3.
x
u
1
u
2
x
y
z
Figure 4.1.
y
x
Figure 4.2.


t
t z y x u t t z z y y x x u
t
u u
a
t
old new
t

) , , , ( ) , , , (
lim lim
0 0
r r r r
r + + + +
=
=

Lets calculate the first component of the acceleration a
x

t
a
t
x
) , , , ( ) , , , (
lim
0
+ + + +
=

We use a Taylor series expansion to write ) , , , ( t t z z y y x x u + + + + as a function of
the nearby value ) , , , ( t z y x u
... ) , , , ( ) , , , ( +
|
|
\
|
+ = + + + + t
t
u
z
z
u
y
y
u
x
x
u
...
) , , , ( ) , , , (
+
|
|
\
|
=
+ + + +
t
u
t
z
z
u
t
y
y
u
t
x
x
u
t

Taking the limit provides the x-component of the acceleration (similarly for a
y
, a
z
)
t
w
z
w
w
y
w
v
x
w
u a
t
v
z
v
w
y
v
v
x
v
u a
t
u
z
u
w
y
u
v
x
u
u a
z
y
x
=

It is customary in Fluid Mechanics to use the symbol D(.)/Dt to represent the total
derivative
( ) ( )
{
( ) ( ) ( )
4 4 4 4 3 4 4 4 4 2 1
on accelerati convective
on accelerati temporal
z
w
y
v
x
u
t Dt
D

II. Bernoullis equation
1. Derivation of Bernoullis equation
Assumptions: (1) Steady flow ( 0 / = t )
(2) Incompressible flow ( = const)
(3) Viscous forces (friction) neglected
Forces therefore arise from pressure and gravity only (no shear forces)



Control volume:
Lets choose a small element of fluid along one streamline (i.e. we examine one stream
filament).
Let s be the distance along the streamline.

Conservation of momentum (Newtons second law) along the streamline
( ) s F F s
Dt
u m D
weight pressure
r
r r
r
r
+ =
) (

Since Newtons second law is applied along the streamline, we need to project the forces
along the streamline (only the first order terms will be retained retained):
- Pressure at s: pA s A pd
A
=
r
r

- Pressure at s+s: ... ) )( ( ) ( + = + + = +
+
A p p A pA A A p p s A d p p
A A

r
r

- Pressure on the lateral sides: ... ) ( ) ( + = + = +
A p A p k p s A d p k p
lat
A

r
r

- Weight: ... ... cos + = + =
gAdz s gA dV s g
CV

r r

Therefore, ( ) gAdz A p A p p A pA pA s F F
weight pressure
+ + = + ) ( ) ( ) (
r
r r

At first order, the mass of the small element is s A . Using the expression of the
acceleration, one gets
s
U
sU A
s
U
U
t
U
s A
Dt
DU
s A
Dt
s u D
s A s
Dt
u m D
=
|
\
|
= =
=
r r
r
r
) (

where U is the velocity magnitude.
z gA p A
s
U
sU A =

Dividing by s A
s
stagnation point (U = 0)

U
s
A
p
A + A
p + p
z = s cos
s A g
Figure 4.4.


0 = + +
s
z
g
s
p
s
U
U

In the limit 0 s
0 = + +
ds
dz
g
ds
dp
ds
dU
U along a streamline
This is Eulers equation for steady flow along a streamline. It relates p, U, z and along a
streamline.

For a constant density fluid, we can rearrange this equation (
ds
dU
ds
dU
U
2
2
1
= )
( ) 0
2
2
1
= + + gz U p
ds
d

Integrating along the streamline, we get
const gz U p = + +
2
2
1
along a streamline
Or in terms of pressure heads
const z
g
U
g
p
= + +
2
2
along a streamline

This result, in either form, is called the Bernoullis equation.
(Note that if U = 0, this equation reduces to the hydrostatic equation).

Along a streamline, pressure reduces with vertical distance z similarly to hydrostatic
variation.
Along a streamline, as the flow speed U increases, the pressure p reduces

If we were able to measure p and p
0
at a point in the flow, we would have a means of
measuring the local velocity in a steady flow system.

2. Understanding Bernoullis equation
a. Total pressure
Lets consider the frictionless flow in a duct past a vertical object:
{
const gz U p = + +
3 2 1
pressure dynamic
2
2
1
pressure static
along a streamline
At a stagnation point within this horizontal flow (where U = 0), the pressure must
increase to a value


4 4 3 4 4 2 1
pressure stagnation pressure, total
2
2
1
0
U p p + = on a given streamline
P
0
is referred to as the total or stagnation pressure of the streamline.

For a gas, the elevation term is usually negligible. Bernoullis equation reduces
0
2
2
1
p const U p = = +
For a liquid, Bernoullis equation can be expressed in the same form
*
0
2
2
1
*
p const U p = = +
Where gz p p + =
*
is the piezometric pressure

b. Constant
Is the constant the same for all streamlines?

- Yes, if the upstream conditions are such that each streamline effectively starts from or
passes through the same conditions of U and p + gz (possibly case 1).

- No, if the upstream conditions are such that each streamline starts or passes through
different conditions of U and p + gz (possibly case 2).
With so drastic conditions (steady, frictionless and incompressible flow), is Bernoullis
equation of practical value?

- Yes. For streamlines in a steady flow away from close proximity to any objects in a
stream, the behaviour is well modelled by Bernoullis equation.

III. Applications of Bernoullis equation

General approach:
- Draw streamline linking points of interest
Plan view
y
x
s
Side View
s
8
s
7
s
6

s
1
Case 1 Case 2
x
Figure 4.5.
z


- Label pressures and velocities
- Apply Bernoullis equation

1. Pressure and velocity measurements
- Stagnation pressure can be measured by means of a Pitot tube (a).
Static pressure can be measured by means of a static pressure hole on a surface aligned
parallel to the flow (b).

- Velocity can be measured by means of a Pitot-static tube.
Static holes are designed and positioned in the Pitot tube to measure the free-stream
value of pressure
p .
A manometer is arranged to measure pressure holes difference:
2
2
1
0
= U p p
=
p p
U
0
2

2. Pressure coefficients
The pressure distribution on bodies can be non-dimensionalised as a pressure coefficient.
manometer
0
, 0 p p U = =
(a)
U p flow in Conditions ,
p
manometer
(b)
Figure 4.6.
0
p
U
p
0
p
p
p
Figure 4.7.


We can define the local pressure coefficient
2
2
1
=
U
p p
C
P

If Bernoullis equation applies, we can write it along the streamline up to the point of
measurement
2
2
1
2
2
1
U p U p + = +

(for a fluid, replace p by p*)
So that
2
2
1
=
U
U
C
P

Pressure distribution around the airfoil (c = length of body = chord of an airfoil)

- At the stagnation point, U = 0, p = p
.
- When C
P
= 0: U = U
.
- When C
P
> 0: U < U
, p > p
.
- When C
P
< 0: U > U
, p < p
.

Example of a polar plot of pressure distribution around a cylindrical body (e.g. tall
chimney or leg of offshore platform), in a real flow (not inviscid):

U
p
x
to manometer
p U
Figure 4.8.
Pressure distribution on top
surface
1.0
0.0
-0.4
1.0
x/c
c
p
Figure 4.9.
U
c
p
+ve
c
p
-ve
Figure 4.10.


IV. Conservation of mass (Continuity)
1. Continuity equation
Conservation of mass requires the mass contained in a system, m, to be constant:
0 =
system
Dt
Dm
with

= =
CV system
dV dm m
0 =
CV
dV
Dt
D

The direct evaluation of the time derivative of this integral over the system is often
difficult, because the system is moving, its boundary is changing shape and the integrand
is also changing in time.
The Reynolds transport theorem allows us to avoid the above difficulties and evaluate
the time rate of change in mass

0 ) (
CS gh flux throu Mass CV in on accumulati mass of Rate
= +

4 43 4 42 1
r r
43 42 1
CS CV
dA n u dV
t

r
r
r
.

The principal of conservation of mass applied to a control volume states that

For steady flows, the rate of accumulation is zero (no time-dependence):

CV
CS
n
r

n
r

n
r

n
r

Figure 4.11.
Mass flux
IN
= Mass flux
OUT
Mass flux
IN
= + Mass flux
OUT
Rate of mass
accumulation in CV


2. One-dimensional steady flow
Steady: 0 / = t
1D flow:
x
e x u u
r r
) ( =
= constant

0 ) 0 ( ) ( ) ( ) ( ) ( ) (
1 2 1
2 1
= + + =

lateral
A
x
A
x x
A
x x
CS
dA n e dA e e U dA e e U dA n u
r r r r r r r r

const Q A U A U = = =
2 2 1 1

Q is the volume flow rate:
=
A
A d u Q
r
r

We see that in a divergent, the velocity decreases:
1 2 1 1 2
/ U A A U U =

3. Two-dimensional unidirectional steady flow
Steady: 0 / = t
2D unidirectional flow:
x
e y x u u
r r
) , ( =
= constant

The velocity is not constant along the inner and outer boundaries. Lets introduce the
mean velocities
1 1
1
1
1
A
Q
A d u
A
U
A
= =
r
r
,
2 2
2
2
1
A
Q
A d u
A
U
A
= =
r
r

CV
A
1
A
2
CS
n
r

n
r

n
r

n
r

1
U
2
U Figure 4.13.
CV
A
1
A
2
CS
n
r

n
r

n
r

n
r

1
U
2
U
Figure 4.12.


We can see that, for a steady flow, the flow rate is constant and that
2 1 1 2
/ A A U U =

V. Flow measurements
1. Liquid flow through a sharp-edged orifice
When the flow goes through a vena contracta, it forms a round jet. We assume the
velocity to be uniform across the jet. Inside the jet, the pressure is equal to the
atmospheric pressure (no effect of surface tension).

We assume that the tank cross-section area is much greater than the area of the hole.

We apply Bernoullis equation along a streamline between points 1 and 2 in the flow.
*
0
2
2
1
*
p const U p = = +
2
2
2 2
1
2 1
2
1 2
1
1
gz U p gz U p + + = + +
atm
p p p = =
2 1
, 0
1
= U
0 2 1
2
2
2 ) ( 2 gh z z g U = =
0 2
2gh U = Torricellis equation

This velocity is the theoretical or ideal discharge rate. If we assume the velocity to be
constant along the radius, the ideal flow rate is
0 2 0
a U Q =

But the area of vena contracta
c
a is smaller than the orifice area
0
a . We can define a
coefficient of contraction
2
p
a
q
2
free jet
2
1
z
1
z
2
Datum
z = 0
p
a
p
a
h
o
Figure 4.14.


0
/ a a c
c c
=
Also, the actual jet velocity '
2
U is less than the theoretical velocity due to viscous
effects. So we can define a coefficient of velocity
2 2
/ ' U U c
v
=
and a coefficient of discharge
c v
c
d
c c
a U
a U
Q
Q
c = = =
0 2
2
0
'

The actual discharge is therefore
c
a U Q '
2
=
0 0
2gh a c
d
=
For small orifices and low viscosity fluids, empirical values of the coefficients may be
found:
98 . 0 ~
v
c , 65 . 0 ~
c
c , 62 . 0 ~
d
c

Note: The theory and result apply not only to flows through holes in the bottom of
containers, but also to holes in the side (if the size of the hole is small compared to the
depth).
Similarly, the theory and result can also be applied to flow into vessels with a hole
submerged in a fluid, or between vessels containing the same liquid, but at different
heads.

2. Gas efflux from a large vessel through shaped outlet or hose

There is no contraction of jet, but friction losses greater than with a sharp orifice.
For the equilibrium of the jet boundary,
a
p p =
Applying Bernoullis equation along a streamline gives
2
2
1
0
U p p
a
+ = .
Thus
) ( 2
0 a
p p
U

=
p
0
p
a
Free jet.
(Streamlines parallel)
p, q
Figure 4.15.


3. Venturi meter
Assumptions: One dimensional flow, steady, negligible friction.

Goal: find Q
The ideal volume flow rate is
0
Q
The actual flow rate is
0
Q c Q
d
= , where 98 . 0 ~
d
c for low viscosity fluids.

Continuity equation:
2 2 1 1 0
U A U A Q = =
Bernoullis equation:
2
2 2
1
2
2
1 2
1
1
U p U p + = +
When we rearrange and combine:
|
|
\
|
=
|
|
\
|
=
2
1
2
2
2
2 2
1
2
2
2
1
2
2 2
1
2 1
1 1
A
A
U
U
U
U p p

The ideal flow rate is
( )
2
1
2
2
2 1
2 2 2 0
/ 1
) ( 2
A A
p p
A U A Q
= =

The actual flow rate is
( )
2
1
2
2
2 1
2
/ 1
) ( 2
A A
p p
A c Q
d

- If the venturi is inclined, the same equation is obtained with
1
p and
2
p being
interpreted as piezometric pressures gz p + .
- Application: measurement of flow rate in a pipe system.
- It can also be expressed in terms of head difference.
g
gz p gz p
h
) ( ) (
2 2 1 1
+ +
=
( )
2
1
2
2
2
/ 1
2
A A
h g
A c Q
d
=
To Manometer
p
1
p
2
Q
Q
U
1
U
2
2
1
A
2
A
1
Figure 4.16.


Chapter 5 Conservation equations in integral form

The differential form of the conservation equations is called Navier Stokes equations
(derived next year).
This year, we will consider the conversation equations in integral form.

The conservation equations are:
- Conservation of mass (scalar equation 1 equation)
- Conservation of momentum (vector equation 1 equation in 1D, 2 equations in 2D,
3 equations in 3D)

I. Conservation of mass (Continuity equation)

Lets consider a system of mass m.
Conservation of mass requires the mass to be constant:
0 =
system
Dt
Dm
with
=
CV
dV m
0 =
CV
dV
Dt
D

The direct evaluation of the time derivative of this integral over the system is often
difficult, because the system is moving, its boundary is changing shape and the integrand
is also changing in time.
The Reynolds transport theorem allows us to avoid the above difficulties and evaluate
the time rate of change in mass
CV
CS
n
r

n
r

n
r

n
r

Figure 5.1.
A
1
A
2
A
lat
F
visc

F
visc



0 ) (
CS gh flux throu Mass CV in on accumulati mass of Rate
= +

4 43 4 42 1
r r
43 42 1
CS CV
dA n u dV
t

r
r
r
.

For the control volume of Figure 5.1
0 ) ( ) ( ) (
0
2 1
= + + +
=

4 43 4 42 1
r r r r r r
lat
A A A CV
dA n u dA n u dA n u dV
t

So the principal of conservation of mass applied to a control volume states that

II. Conservation of momentum
1. Equation of conservation of momentum
We need to apply Newtons second law to a control volume. But similarly to the
derivation of the mass conservation equation, we shall begin with the formulation for a
system and use the Reynolds transport theorem to go from the system to the control
volume formulation.

Newtons second law for a moving system

+ = =
surf body
F F F
Dt
u m D
r r r
r
) (

The linear momentum is given by

= =
CV system
dV u dm u u m
r r r

The system and control volume formulations are related using Reynolds transport
theorem

+
= =
CS CV CV
dA n u u dV
t
u
dV u
Dt
D
Dt
u m D
) )( (
) (
) (
) ( r r r
r
r
r

+ = +
surf body
CS CV
F F dA n u u dV
t
u
) )( (
) ( r r r
r

Mass flux
IN
= + Mass flux
OUT
Rate of mass
accumulation in CV


Body forces: The body force act on the control volume.
The principal body force that we will encounter is the weight
=
CV
dV g W
r
r

Other body forces include the electrical and magnetic forces, but these will rarely be
present in any of the problems that we will study.

Surface forces: The surface forces only act on the control surface.
As we have seen, the surface forces are the pressure forces and the viscous forces

+ =
CS CS
surf
dA n dA n p F
r t r
r

where p is the pressure and
t
the viscous stress.

3 2 1
r t
43 42 1
r
43 42 1
r
4 4 3 4 4 2 1
r r r
4 43 4 42 1
r
force Viscous force Pressure force nal Gravitatio
force Inertial
CS gh flux throu Momentum
CV in momentum of
on accumulati of Rate
) )( (
) (

+ + = +
=
CS CS CV CS CV
dA n dA n p dV g dA n u u dV
t
u

This is a vector equation. In 3D, it contains 3 equations. In Cartesian coordinates:
+ + = = +
+ + = = +
+ + = = +

CS CS CV
z
CS CV
CS CS CV
y
CS CV
CS CS CV
x
CS CV
dA k n dA k n p dV k g F dA n u w dV
t
w
dA j n dA j n p dV j g F dA n u v dV
t
v
dA i n dA i n p dV i g F dA n u u dV
t
u
)
( )
( )
( ) )( (
) (
)
( )
( )
( ) )( (
) (
)
( )
( )
( ) )( (
) (
r t r r r r
r t r r r r
r t r r r r

2. Applications: impacts of jets on surfaces
a. Stationary normal plate
Water from a stationary nozzle strikes a flat plate as shown below. Assuming the water is
directed normally to a plate and flows along the plate, determine the horizontal force on
the support.
The pressure is the atmospheric pressure.


- Choice of control volume (see figure on the left)

- Surface and body forces acting on the control volume
- Weight
- Reaction of the plate on the fluid (includes the friction force)
- Pressure forces
We denote R
x
, R
y
the components of the reaction force of the support on the control
volume. The force of the fluid on the support is equal and opposite to (R
x
, R
y
).

- Assumptions: (1) Steady flow ( 0 / = t )
(3) Uniform flow at each section
(4) Friction neglected
(5) Gravity neglected: the velocities are high and the plate is close
enough to the jet for the gravity effects to be neglected

- Goal: we are looking for the horizontal force of the jet on the plate (F
x
= - R
x
)

Mass conservation equation:
0 =
CS
dA n u
r r

0
0
* 2
1 2
= + +
=

43 42 1
r r r r r r
lat
A A A
dA n u dA n u dA n u
2 2 1 1
2 A U A U Q = =

x-momentum conservation equation:
x
CS A CS
R dA i n p dA n u u dA n u u + = =

) ( ) (
1
r r r r r

CV
Force on plate U
1

A
1

U
2
A
2

y
x
CV
Force on plate U
1

A
1

U
2
A
2

Alternative for CV
Figure 5.2.


The pressure force is uniformly acting on the control surface, so they cancel out.
1 1
2
1 1 1 1
1 1
) )( ( ) )( ( U m A U dA U U dA n u u R
A A
x
&
r r
= = = =

The horizontal force on the support is
1
2
1
A U F
x
= .
An alternative for the control volume is to choose one following the jet (stream tube). See
figure on the right.

b. Normal plate moving with horizontal velocity c (c < U
1
)
We place ourselves in the reference frame of the moving plate. In this reference frame,
the incoming jet velocity is U
1
c.

The horizontal force on the support is
1
2
1
) ( A c U F
x
= .

c. Stationary inclined plate.
Lets define the coordinates system as shown in the figure. The control volume is chosen
as the stream tube around the jet.
We neglect the weight and friction: the y-resultant of the force of the plate acting on the
jet is zero.

The X-component of the incoming velocity is U
1
cos.
The horizontal force on the support is cos cos
1 1
2
1
U m A U F
X
& = = .

d. Stationary curved vane tangential at contact
The plate is arranged so that the jet is tangential at its left end.
For the control volume, we choose the stream tube containing the jet along the plate

2 2 1 1
A U A U Q = =
CV
Force on plate
U
1

A
1

Y
X

Figure 5.3.


Since A
1
= A
2
, U
1
= U
2
.

Momentum conservation equations:

sin sin ) (
) 1 (cos cos ) (
1
2
1 2
2
2
1
2
1 2
2
2 1
2
1
A U A U A d u v R
A U A U A U A d u u R
CS
y
CS
x
= = =
= + = =
r
r
r
r

The force on the plate is ) cos 1 (
1
2
1
= A U F
x
, sin
1
2
1
A U F
y
= .

3. Flows in duct contractions and nozzles
a. Straight contraction
Water flows steadily through a straight contraction (convergent). Determine the force
acting on the contraction.
The pressure is the atmospheric pressure.

- Control volume (see figure)

- Surface and body forces acting on the control volume
- Weight W
- Reaction of the contraction on the fluid R
x
, R
y

- Pressure forces
- Assumptions: (1) Steady flow ( 0 / = t )
(3) Uniform flow at each section
(4) Friction neglected

y
x
A
2
U
1
A
1
Force on plate

U
2
Figure 5.4.
U
1
A
1
p
1
U
2
A
2
p
2
y
x
Force on contraction

A
lat
Figure 5.5.
2 1
A A


(5) Neglect weight of contraction and water in contraction

- Goal: we are looking for the force of the water on the contraction (F
x
= - R
x
, F
y
= - R
y
)

0 =
CS
A d u
r
r

2 2 1 1
A U A U Q = =
Since A
1
> A
2
, then U
1
< U
2
: the velocity increases inside the contraction.

y-momentum conservation equation:
{ y
CS CS
o
R j A pd A d u v + =
=
=

43 42 1
r r
r
0
) (
0 =
y
R
x-momentum conservation equation:
x
CS CS
R i A pd A d u u + =

) (
r r
r

x
A
lat
A A A A
R i A d p dA p dA p dA U U dA U U
lat
+ = +

) ( ) (
2 1 2 1
2 1 2 2 1 1
r

2 2 1 1 1 2 1 1
2 2 1 1 1 2 1 1
2 1 2 2 1 1 2
2
2 1
2
1
) (
) ( ) ( ) (
) (
A p A p U U A U R
A p p A p p U U A U R
A A p A p A p A U A U R
gage gage x
atm atm x
atm x
+ =
+ =
+ + + =

The force on the contraction is
2 2 1 1 1 2 1 1
) ( A p A p U U A U F
gage gage x
+ = , 0 =
y
F .

b. Nozzle
The force acting on the contraction is
1 1 1 2 1 1
) ( A p U U A U F
gage x
+ = , 0 =
y
F .

U
1 A
1
p
1
U
2
A
2
P
atm
y
x
Force on contraction

A
lat
Figure 5.6.


c. Jet reaction
The propulsive force is
1
2
1
A U F
x
= .

Force on
body
p
atm
Force on
fluid
cv
A, U

Figure 5.7.


Chapter 6 Incompressible viscous flow
Dimensional analysis and similitude

I. Effects of viscosity
So far in this course, we have been looking at idealized flows that have no viscosity,
where adjacent particles can slip past each other without any frictional forces being
generated. But if we consider the molecular level, we know that there is transfer of
molecules between layers due to thermal agitation, which leads to a momentum transfer.

Molecules with higher momentum from level 2 will move into a layer with lower
momentum (level 1). They will mix and increase the overall momentum of layer 1.
Likewise, momentum from 1 will mix into layer 2 and will lower the relative velocity
between the two layers.

Viscosity arises due to this thermal agitation of particles and due to the intermolecular
forces of attraction/repulsion (Van der Waals forces).

Gases:
Particles thermal agitation is the dominant cause of viscosity since the intermolecular
forces of attraction/repulsion are very small in gases with molecules being so far apart.
Since thermal agitation increases with temperature, viscosity increases with temperature.
Analysis based on kinetic theory predicts
T
Experiments correlate quite well with the theoretical prediction.
T S
T b
/ 1+
=
For air, b = 1.458 10
-6
kg/msK
1/2
, S = 110.4 K

u
Layer 1
u + u
Layer 2
Figure 6.1.


Viscosity is, however, relatively independent of pressure. Increased pressure may
increase the frequency and intensity of collisions of molecules, but it does not appear to
significantly affect the momentum exchange between layers.

Liquids:
Liquids have closer packed molecules than gases. Intermolecular forces are much
more significant, and they are modified under the action of shear. The relationships are
not clear and are still subject of research.
Liquid viscosities drastically decrease with temperature (see graph of viscosity as a
function of temperature). The dependence on absolute temperature may be represented by
the empirical equation
C T
B
Ae

=
Viscosity of liquids is again essentially independent of pressure.

Data are given in Appendix A of the book in the tables of properties.

II. Boundary layers
1. No-slip boundary condition
Lets introduce the important concept of boundary layers in real fluid flows.

The boundary layer forms as a consequence of the no-slip boundary condition at a wall:
at the wall, the fluid has the same velocity as the wall.
U
wall
U
U
0 =
wall
U
wall
U
Figure 6.2.


wall x x
u u
wall
r r
r r
=
=

The boundary layer is the layer of fluid close to the wall, where the fluid adapts its
velocity from the outer flow velocity to the wall velocity. The boundary layer thickness,
, increases with distance along the wall.

2. Shear stresses and wall shear stresses
A shear force (F) or shear stress appears in the fluid because of the gradient of
velocity.
Let du be the relative velocity between layers dy apart.
For a Newtonian fluid, that has a dynamic viscosity , the shear stress developed
between layers in a laminar flow is given by
dy
du
=

The wall shear stress on a surface is given by
0
0
=
=
y
dy
du

3. Entrance length; fully developed flow
When the flow enters inside a tube, a boundary layer builds up progressively.
0

x
u+du
u
Figure 6.4.
y
y+dy
y
y
u = f(y)
0
/
= y
dy du
u
y
U
No slip at surface
0
Shear stress at wall due to

velocity gradient at surface
Figure 6.3.


When the flow is fully developed, the velocity profile does not depend on x
0 =
x
u

The fully developed flow in a circular tube is a parabola.

The value of the entrance length is given by semi-empirical correlations, such as
) Re 06 . 0 ( 06 . 0
D
e
UD
D
L
= =

4. Viscous diffusion
Lets consider the example of an impulsive start of a plate

The thickness of the boundary layer increases with time due to viscous diffusion.
The length scale of viscous diffusion varies with the kinematic viscosity and time t
t L ~

III. Laminar and turbulent flows
1. Flow transition inside a circular tube
a. Reynolds experiments
Sir Osborn Reynolds (1842-1912) conducted a very large set of experiments looking at
the flow inside tubes. He looked at the flow behaviour inside a tube using a thin fillet of
dye injected locally into the flow and being convected by it.
Figure 6.6. U
t
0
= 0
U
t
1
> t
0

U
t
2
> t
1

Fluid
at
rest
Fully developed flow
Figure 6.5.
x
Entrance length L
e


He consistently changed the size of the tube (diameter D and length L), as well as its
rugosity. He also changed the velocity of the incoming flow.

Sir Reynolds showed that
As ,
crit
U
As D,
crit
U
As ,
crit
U
After Sir Reynolds death, scientists found out that a unique parameter determined flow
transition (i.e. transition always occurs at the same critical value of the parameter). This
parameter is called the Reynolds number.

UD UD
= = Re
The Reynolds number is a non-dimensionalized number.
- For Re < Re
critical
, the fluid is laminar, i.e. the flow behaves nicely and layers can be
envisaged to slide over each other with friction between.
Dye
Long glass tube Control
valve
Figure 6.7.
U < U
critical
: Laminar flow no mixing.
Velocity
increasing
Random motions of particles in all
directions superimposed on main flow
Transition point
Figure 6.8.
For a given tube (diameter D) and a certain fluid (viscosity , density ):
U = U
critical
: Transitional flow begins to waver.
U > U
critical
: Turbulent flow high mixing


- For Re > Re
critical
, the flow is unstable and undergoes a transition to turbulent
behaviour.

From empirical results, the critical Reynolds number of the flow in a circular tube is
2300 Re =
critical

The value of the critical Reynolds number is universal. Transition always occurs at the
same critical value, no matter which fluid (liquid or gas) or which tube geometry is used.

b. Turbulence
Flow in a tube exhibits a transition when the value of the Reynolds number is higher
than a critical value. Above this value, any perturbation placed in the flow grows in
amplitude and the flow becomes turbulent.
Turbulence is characterized by a random, chaotic motion of fluid particles. Mixing is
enhanced, as well as any dissipation mechanism (viscous, heat, etc.). The flow is made
out of large vortex structures that carry high levels of energy.
Below the critical value of the Reynolds number, the flow is laminar. Perturbations
cannot grow and therefore die out.

c. Friction factor
One can define the friction factor as the ratio of the wall shear stress to the dynamic
pressure in the pipe flow (based on the mean velocity)
2
2
1
0
U
f
= (non-dimensional)

The graph indicates that the friction factor is a function of the Reynolds number and of
the pipe roughness.
Experimental
verification:
10
2
10
3
10
4
10
5
0.001
0.01
0.1
f
2300 = Re
D

crit
Increasing pipe
roughness
Laminar
Smooth pipe
results (solid
line)
(Note log-log scale)
Turbulent
Figure 6.9.
D
Re


The value of the critical Reynolds number depends on the rugosity of the tube and the
importance of perturbations in the flow.
The flow will transition to turbulence at a lower Reynolds number when roughness is
increased.

d. Effect of flow transition on pressure loss in a tube due to friction
Friction of a flow in a tube gives rise to a loss in pressure. The pressure at the end of the
tube is lower than the pressure at the entrance. For a fully developed flow, the pressure
loss in a tube due to friction can be measured as follow

p
d
U
Valve
varying
U or Q
Entrance length (long)
Fully developed flow
(Conditions independent of x)
x
L
Measuring length
Figure 6.10.
Pressure
gradient
L
p
D U
= Re
Increasing U
Decreasing U
Turbulent
Laminar
n
U
L
p
(n 1.7 to 2.0)
Lower Re
critical

Higher Re
critical

(Can be increased to some extent by eliminating vibration)
U
L
p
For a given
pipe/fluid
combination
Figure 6.11.


2. Flow transition of external flows
Flow transition is not only true in a confined tube. Flow also becomes turbulent when
flowing around objects.

Remark: for each type of flow, it is critical to find the appropriate characteristic length
and velocity. In general, for a characteristic velocity U and characteristic length L, the
Reynolds number is

UL UL
= = Re

a. Flow over a flat plate
For the flow over a flat plate, the characteristic length is the distance along the plate. The
Reynolds number is therefore
x U
x
= Re
The critical Reynolds number is
Re
x

critical
=
5 5
10 5 10 3

b. Separation of boundary layer in a decelerating flow
Flow separation may occur within the boundary layer, when the outer fluid is
decelerating.
Separation is caused by the presence of slowly moving fluid at the boundary (effect of
viscosity). Separation is caused by the presence of a positive (adverse) pressure
gradient, which results from a reduction of velocity in the boundary layer
Separation can only occur when an adverse pressure gradient exists.

Flow over a flat plate with zero or negative pressure gradient will never separate before
reaching the end of the plate.

Laminar boundary
layer
x
Turbulent
boundary
layer
Figure 6.12.


Examples of separation:
Plate: separation at a salient edge

Airfoil: loss of lift at large angles of attack

c. Flow around a cylinder
For the flow outside a cylinder, one should use the diameter D, when calculating the
Reynolds number, and not the radius R.
UD
D
= Re
A wake is created behind the cylinder for Re > 40 and a vortex pair is shed. Alternate
vortex shedding occurs for Re ~ 140. The wake becomes turbulent for
5
critical
10 3 Re =
D

Separation points
y
U(x)
u(x,y)
U increasing, so p decreasing
(Bernoulli)
x
Decelerating outer flow
0 >
dx
dp
0 >
dx
dU
0 <
dx
dp Favourable
pressure
gradient
Adverse
pressure
gradient
0 <
dx
dU
Separation
streamline
Flow reversal
region
Outer flow
Boundary
Layer
Accelerating outer flow
U decreasing, so p increasing
(Bernoulli)
Figure 6.13.
Separation points


Turbulent boundary layers last longer before separating: boundary layer separation is
delayed because of greater mixing due to the turbulence and steeper velocity gradients
which are less prone to separate.

Remark 1: for a sphere, the critical Reynolds number is Re
crit
= 10
5
.

Remark 2: In the interests of reducing pressure drag or form drag, turbulence in the
boundary layer may be deliberately stimulated e.g. dimples on golf balls.

Separation of turbulent B.L.
(Re
d
> ~ 3 10
5
)
Separation of laminar B.L.
(Re
d
< 10
5
)
Separation points
Figure 6.15.
Re ~ 40
Re >
5
10 3
Figure 6.14.
Transition to a turbulent wake
Re ~ 1
Von Karman alley
Re ~ 140
A symmetric vortex is shed.
Remains stable in the wake
Alternate vortex shedding
Vortices convected by flow:


IV. Dimensional analysis Similarity

Lets consider the flow around an object (the prototype). For purpose of simulation, a
model has to be build at a smaller size. How should the flow around the model be, so that
the two geometrically flows behave in a same dynamic fashion, i.e. are dynamically
similar?
1. Levels of similarity
The levels of similarity between a model and a prototype are
- Geometric similarity: It requires that the ratio of corresponding dimensions between
model and prototype are the same for ALL features of the system (proportional
transformation).
- Kinematic similarity: It requires that the ratio of corresponding velocities and
accelerations between model and prototype are the same for ALL features of the
systems.
- Dynamic similarity: It requires that the ratio of corresponding forces between model
and prototype are the same for ALL features of the systems.
For any hope of dynamic similarity in a fluid flow system, you must first, at least, have
geometric and kinematic similarities of the boundaries in the system.

2. Example of similarity between two airfoils
a. Finding the non-dimensional numbers
Consider two steady incompressible flows past two geometrically similar bodies totally
immersed. The bodies have different scales and the fluids are different.

Consider (scaled) fluid elements at corresponding (scaled) locations P in the flows. The
forces acting on the elements here are considered to arise only from the pressure
distribution and the action of viscosity in each case.
For dynamic similarity we require similarity of the force polygons (here triangles).
P
U
p
Prototype
P
U
m
Model
Figure 6.16.


The reaction to the resultant force [= mass acceleration] is called the inertia force.
This is a measure of the reluctance of the element to being accelerated (because of its
mass).
{
4 4 4 3 4 4 4 2 1
ons accelerati inertial
acc
temp
z
y
x
z
w
w
y
w
v
x
w
u
t
w
a
z
v
w
y
v
v
x
v
u
t
v
a
z
u
w
y
u
v
x
u
u
t
u
a

.
.
=

For dynamic similarity, the ratio of the magnitudes of any two of the three forces
involved must be the same for the model and prototype (The directions of the forces are
covered by the geometric and kinematic similarity requirements.)
prototype model

|
|
\
|
=
|
|
\
|
force viscous
force inertial
force viscous
force inertial

prototype model

|
|
\
|
=
|
|
\
|
force pressure
force inertial
force pressure
force inertial

b. Expression of the non-dimensional numbers
To express the non-dimensional numbers, one needs to make the list of the all the
characteristic parameters on which the flow depends on

Characteristic dimensions of
- mass: ~ L
3

U

L
,
Figure 6.18.
Pressure
force
Viscous force
Resultant = mass acceleration
= Inertial force
Pressure
force
Viscous
force
Resultant = mass acceleration
= Inertial force
Figure 6.17.


- acceleration: ~
L
U
L
U
U
2
=
- inertial force:
2 2
2
3
~ U L
L
U
L force inertial =

Characteristic dimensions of
- shear stress ( dy du / = ): ~
L
U

- area: ~ L
2

- viscous force: L U L
L
U
force viscous =
2
~
Therefore:
Re ~

2 2
= = =

L U L U
L U
L U
force viscous
force inertial

For dynamic similarity, the Reynolds must be identical for both flows.

c. Interpretation of the Reynolds number
The Reynolds number is the ratio of inertial force to the viscous force.
For Re << 1, the flow is dominated by viscous forces.
For Re >> 1, the flow is dominated by inertial forces. Viscous forces only play a role in
the boundary layer.

Prandtls concept of the boundary layer in high Reynolds number flows is that although
viscous forces can be considered small, their effects are not negligible close to the body.
The flow can be divided into two regions:
Outer flow behaves as inviscid.

(very small velocity gradients)
u
y
U
No slip at surface
0
Shear stress at wall due to velocity

gradient at surface
Very thin boundary layer,
where viscous effects are
confined
Figure 6.19.


- The outer region, where the flow behaves like an inviscid fluid
- The inner region, or boundary layer, where viscous effects dominate

3. Dimensional analysis: Pi-Buckingham theorem
a. Dimensional analysis
Lets take the example of the friction in a pipe under a fully developed flow. There are
only 5 parameters governing the problem (it does not depend on the tube length, since the
flow is fully developed):
0
, d,U , ,

Reynolds raw results must have filled several hundred notebooks!

But we have seen that the friction depends on the values of the other parameters and that
the results are not independent. The parameters are therefore linked by a relationship:
) , , , (
0
U d =
The underlying physical dependence between variables in any system can be determined
by Dimensional Analysis where Buckinghams Pi Theorem is employed.
Dimensional analysis is the preliminary analysis necessary in any experimental design.

b. Buckinghams Pi Theorem
Buckinghams Pi theorem:
Let n physical quantities V
1
, , V
n
of a given system be functionally related.
Let r be the number of fundamental magnitudes required to express the dimensional
formulae of the n variables (mass M, length L, time T).
0
U
d
Variations with
Varia-
tions
with
Figure 6.20.


Then the problem is governed by (n r) dimensionless groups
r n
,..., , ,
3 2 1
,
refereed to as Pi groups (denoting product).
The functional relationship between the n variables can be reduced to a functional
relationship between n - r dimensionless variables.
) ,..., , (
3 2 1 n
V V V V = ) ,..., , (
3 2 1 r n
=

Method for forming Pi groups:
- Write down the dimensional formulae of the n variables in the list. The n variables
depend on r dimensions (M, L, T).
[F
visc
] = MLT
-2

[ d ] = L
[U ] = LT
-1

[ ] = ML
-3

[ ] = ML
-1
T
-1

n = 5, r = 3, so there are n r = 2 dimensionless groups.

- Choose r variables from the list to form a recurring set.
Method to choose the recurring set:
o Each of the fundamental magnitudes must be contained at least once in this
chosen set, but the variables chosen must not be capable of forming a
dimensionless group by themselves.
o The variable of prime interest should not appear in the chosen set.

Choice: d , U ,
(M, L, T) appear in this set, but they cannot form a Pi group on their own, because M and
T appear only once. Do not choose
0
, since it is the dependent variable.

- Express the fundamental magnitudes in terms of the recurring set.
M = [ d
3
]
L = [ d ]
T = [ d U
-1
]

- Re-cast the dimensional formulae of the remaining n r variables in the list.
[F
visc
] = MLT
-2
= [( d
3
) ( d ) ( d U
-1
)
-2
] = [ d
2
U
2
]
[ ] = ML
-1
T
-1
= [( d
3
) ( d )
-1
( d U
-1
)
-1
] = [ d U ]



- Form n r dimensionless groups.
The 2 dimensionless groups are based on the remaining variables, non-dimensionalized
by a combination of the set of recurring variables:

c b a
visc
d U
F
=
1
,
f e d
d U
=
2
.
The values of the exponents a, b, c, d, e, f are found using the previous step.
2 2
1
d U
F
visc
= (a = 1, b = 2, c = 2)
d U
=
2
(d = 1, e = 1, f = 1)
We recognize the friction coefficient f
2 2
d U
F
visc
= and the Reynolds number
d U
= Re .
Hence, from the Pi-Buckingham theorem, one can conclude that:
) (
2 1
= or (Re) = f

V. Dimensional analysis for external flows
1. Force coefficients
Pressure coefficient:
2
2
1
=
U
p p
C
P

Friction factor:
2
2
1
U
f
=

Reynolds number:

Ud Ud
= = Re (ratio of inertial to viscous forces)

The typical forces of interest for bodies immersed in a flow are the DRAG and LIFT.
- Drag: component of force on a body acting parallel to the direction of relative
motion.
- Lift: component of force on a body acting perpendicular to the fluid motion

U
p
0
U


Drag coefficient:
A U
D
C
D
2
2
1

Lift coefficient:
A U
L
C
L
2
2
1

The surface area A is a representative surface area of the body, chosen by convention.
- Sphere

- Finite cylinder / bluff bodies

- Slender bodies (e.g. aerofoil)

See document for comparison of C
D
for bodies of different shapes.

2. High Reynolds number flows around bodies
High Reynolds numbers implies that the magnitude of the inertial forces is much greater
than the magnitude of the viscous forces in a system. But it does not imply that viscous
forces do not play a role.
Lets compare the theory for zero viscosity with experiments for high Re flow past a
submerged cylinder:
c
b
U
cb A = ,
c U
= Re
d
b
U
bd A = ,
d U
= Re
L
D
LIFT
( to U

)
DRAG
( // to U

)
U
d
4
2
d
A

= ,
d U
= Re


Similarly, we can compare the theory for zero viscosity flows with experiments at high
Reynolds number flows past a submerged aerofoil:

3. Contributions to drag and lift
The pressure p acting on the object is determined by Bernoulli outside the very thin
boundary layer. When the boundary layer is not separated, the pressure is transmitted
unchanged across it to act on the surface.
The local pressure force pdA has a component in the stream direction. The integration of
this component over the whole body surface gives the drag due to the pressure
distribution or form drag.
p
p dA
dA
0
dA
Figure 6.23.
High velocity
Low pressure
Full pressure
recovery
Separation of
boundary layer
WAKE REGION
Low pressure and large
scale eddies
INVISCID THEORY EXPERIMENT
No drag predicted: D = 0
Large drag
observed
Figure 6.21.
No lift predicted (L = 0)
No drag predicted (D = 0)
Confirms predicted lift L
Drag observed is very small
INVISCID THEORY
Can be made to predict lift L
But still no drag predicted (D = 0)
L
D
Narrow wake
L
EXPERIMENT
Figure 6.22.


The same element of area dA experiences a shear stress
0
due to the velocity gradient
normal to the surface. The shear force,
0
dA, has a component in the stream direction,
which when integrated over the body surface gives the drag due to skin friction.
The total drag, also known as Profile Drag, is
Total drag = Form drag + Skin friction

Flow separation leads to large form drag, which can be reduced through streamlining:

The integration of the pressure force component perpendicular to the stream direction
over the whole body surface gives the lift force.

Analytical expressions of the drag and lift forces are not possible to get in practice. The
drag and lift coefficients therefore need to be evaluated experimentally.

4. Non negligible effects of gravity force
In circumstances where gravity forces are important, achievement of dynamic similarity
requires satisfying a further condition.
In fluids, gravity forces are important when a liquid free surface is present and when it is
disturbed to produce waves as the flow passes by a floating or submerged body.
In the dimensional analysis, we must include a further variable, the gravitational
acceleration, g.

Example: drag force on a ship

The drag force on a ship depends on 5 variables
U

Figure 6.25.
Airfoils
Skin friction dominant
Low form drag below stall condition
C
D
~ 0.01 at high Re
Bluff Bodies
Form drag dominant
Skin friction insignificant
C
D
~ 1.0 at high Re
Figure 6.24.


) , , , , ( g U D l
=

We thus expect 6 3 = 3 groups. Following through the analysis as before with the
additional term, and using the same recurring set we can show that
] [ ] [ ] [ ] [
2 1 2 2 2

= = = U U LT g l l l , thus
l g
U
2
3
=

The flow depends on an additional dimensional number, called the Froude number
l g
U
Fr

=
The Froude number is proportional to the square root of the ratio of the inertial force to
the gravitational force:
2
2
3
2 2

Fr
g
U
g
U
force Gravity
force Inertial
= =

l l
l

Hence, from the Pi-Buckingham theorem, one finds
) , (
3 2 1
= ) (Re, Fr C
D
=

Model testers dilemma:
Alas for ship model testers, it is found that whilst having to accept the limitation of
testing on earth (g is fixed), there is no practical fluid that can be used to achieve equality
of Reynolds number and Froude number simultaneously!
All model testers can do is to satisfy the Fr requirement (similarity of wave patterns)
whilst using as big a scale as possible in order to get Re as high as possible.
Even for systems without free surfaces, it can sometimes turn out to be impossible to get
Re equal for both the model and the prototype system (e.g. limitation of scale possible
and/or air speed possible within available wind tunnel facilities).
Where Re
m
= Re
p
cannot be achieved, efforts are made to achieve at least fully turbulent
conditions in the boundary layer by introducing some artificial roughness to trip / mix
the early laminar boundary layer flow near the leading edge of the body. We then rely on
C
D
, etc, being insensitive to actual Re.

U

Trip wires or studs Figure 6.26.


VI. Vortex motion
1. Vortex generation

Possible effects of vortices:
- Vortex shedding can cause vibrations. If the frequency of shedding matches a
natural vibration frequency of the structure then structural resonance can occur
(e.g. Tacoma Bridge).
- Cavitation can occur at the tips of the blades of propellers or impellers in liquids as
a result of local boiling of the liquid in the low pressure centre of the trailing
vortex from the tips. Vapour bubbles are formed which are unstable and can
collapse with very high impact loads when they pass into higher pressure regions.

2. Flow induced by a vortex
All the above phenomena occur at high Reynolds numbers. Away from any boundaries,
the flow can be therefore assumed to be inviscid.

Let us examine the equilibrium of such a particle of fluid in a steady flow system
passing along a curved portion of a streamline. Let the radius of curvature of the
streamline at a given point be r and the local velocity be u.
Assumptions:
- Steady flow
Aircraft wing
Trailing vortex
from wing tip
Vortex street
Propeller vortex system
(Tip and hub vortices)
Figure 6.27.
O
centre of
curvature
r
u
Figure 6.28.


- Incompressible
- Inviscid
- Uniform velocity u
Lets apply Newtons 2
nd
law in the direction normal to the streamline to the particle.
For equilibrium of the small element of fluid shown, there must be a pressure gradient to
produce the centripetal acceleration u
2
/r towards the centre.
A p A r
r
p
p
r
u
r A
|
\
|
+ =
2

r
p
r
u
=
2

Remark:
If the streamlines are parallel [let the radius of curvature r ], then the (piezometric)
static pressure must be constant across the flow. Thus, there is no pressure gradient
perpendicular to the streamlines.
As r tends to infinity (parallel streamlines), we have

3. Free vortex
A free vortex is generated by pressure differences contrary to a forced vortex, which
is generated by an external torque.
To examine this phenomenon we have to postulate a flow having concentric circular
streamlines for which the pressure p
0
is the same for all the streamlines. So
) (r u u =
A r
r
p
p |
\
|
+
O
r
r
A p
u
Figure 6.29.
O at
r =
Figure 6.30.
0
2
= =

r
U
r
p

u = U
(z)

p = p(z)


The equation of conservation of momentum is
r
p
r
u
=
2

Combining both equations, one gets an equation for the velocity field
r
u
u
r
u
=
2

0 = +
r
u
r
u

Since ) (r u u =
0 = +
r
u
dr
du

0 = +
r
dr
u
du

= + r u ln ln constant
r
K
u = , where K = constant

The pressure distribution can be found using the conservation of momentum
3
2 2
r
K
r
u
dr
dp
r
p
= = =

2
2
0
2
2
2
1
2
1
r
K
p C
r
K
p = + =
since the stagnation pressure p
0
is the same for all the streamlines.

r
u
Figure 6.31.
u
r
p
r
p
0
2
2
2
2
1
2
1
r
K
u =

Figure 6.32.
If p
0
is equal to the same constant for all streamlines
then 0 )
2
1
(
2 0
= +
u p
r r
p

0 =
r
u
u
r
p



The model predicts infinite velocity and negative (!?!) infinite pressure at the centre,
neither of which are physically possible.

However, such a flow exists at least in the region away from the centre of the vortex. It is
known as a FREE VORTEX and the constant K is known as the vortex strength. The
constant is linked to the circulation of the vortex
2
= K
Near the centre, the model breaks down because of the infinite values of the velocity and
pressure.

4. Rankine vortex (or combined vortex)
Recognising the physical impossibilities of the free vortex in real fluids leads us to
postulate a combined model, which has the free vortex behaviour away from the core, but
forced vortex behaviour within the core. This model is known as the Rankine vortex or
combined vortex.

Pressure distribution in the forced vortex, where the velocity field is = r r u ) ( :
2
2
= = =
r
r
u
dr
dp
r
p

C
p r C r p + = + =
2 2 2 2
2
1
'
2
1

where
C
p is the pressure at the centre of the vortex (r = 0).

Lets assume that the vortex core has a radius b. Velocity and pressure must be matched
at r = b with the values for the free vortex
Velocity match:
b r
free forced
u u
=
=

Forced vortex
core
r Free vortex
behaviour
u = r
u = K/r
Figure 6.33.


= b
b
K

=
K
b
Pressure match:
b r
free forced
p p
=
=
2
2
0
2
1
r
K
p p
free
=
C forced
p r p + =
2 2
2
1

C
p b
b
K
p + =
2 2
2
2
0
2
1
2
1

2 2
2
2
0
2
1
2
1
b
b
K
p p
C
=

The Rankine (or combined) vortex model is found to correlate well with measurements
within free vortices in real fluids.

b
Forced
r b
r
u
p
p
0

p
b
p
c

u
b

Free Forced Free
Figure 6.34.


Chapter 7 Internal incompressible viscous flow

I. Fully developed laminar flows between infinite parallel plates
1. Simple Couette flow

Control volume: Infinitesimal CV of volume dxdydz.

Assumptions:
- Incompressible fluid
- Steady flow ( 0 / = t )
- Infinite plates in the x- and z-directions: flow independent of x and z ( 0 / = x ,
0 / = z ) ) ( y u u =
- Uniform pressure no pressure force
- No gravitational effects

Steady flow no temporal acceleration
Flow independent of x no momentum flux through CS
There is no net force on the element: F
surf
= 0
0 ) ( ) ( = + dxdz dxdz dy
dy
d

0 = dxdydz
dy
d

0 =
dy
d
, since 0 dxdydz

b
u(y)
U
Figure 7.1.
dxdz dy
dy
d
) (

+
dxdz
dy
y
x


A const = =
A
dy
du
= dy
A
du
= B y
A
u + =
.

Boundary conditions: 0 =
=o y
u B = 0
U u
b y
=
=

b
U
A

=
b
y
U y u = ) ( (linear velocity distribution)
NB: The shear stress throughout the flow is equal to the wall shear stress:
b
U
= =
0

2. Poiseuille flow between two infinite plates

Control volume: Infinitesimal CV of volume dxdydz.

Assumptions:
- Infinite plates in the x- and z-directions: flow independent of x and z ( 0 / = x ,
0 / = z ) ) ( y u u =
- Flow driven by a constant negative pressure gradient G
dx
dp
= < 0

surf
= 0
0 ) ( ) ( ) ( ) ( = + + + dxdz dxdz dy
dy
d
dydz dx
dx
dp
p dydz p

b
u(y)
Figure 7.2.
y
x
dxdz dy
dy
d
) (

+
dxdz
dydz dx
dx
dp
p ) ( +
pdydz


G
dx
dp
dy
d
= =

G
dy
du
dy
d
= ) (
A Gy
dy
du
+ =

A
y
G
dy
du
+ =
B y
A
y
G
u + + =

2
2

Boundary conditions: 0 =
=o y
u B = 0
0 =
=b y
u
2
Gb
A =
=
(
=
2
2 2
2
2 2
2 2
) (
b
y
b
y b
dx
dp
b
y
b
y Gb
y u

(parabolic velocity distribution)

Shear stress distribution
The shear stress distribution is given by
(
=
(
= + =
b
y
b
dx
dp
b
y
Gb
Gb
Gy
2
1
2
1
2

Volume flow rate
The volume flow rate is given by
=
A
dA n u Q
r r

The flux occurs through a rectangle of width w in the z-direction and height b in the y-
direction
b
b b
b
y
b
y
w
b
dx
dp
dy
b
y
b
y
w
b
dx
dp
uwdy Q
0
2
3 2 2
0
2
2 2
0
3 2 2 2
(
=
(
= =

w b
dx
dp
Q
3
12
1
= , or
3
12
1
b
dx
dp
w
Q
=
Mean velocity
The average velocity is given by


2
12
1
b
dx
dp
wb
Q
A
Q
U
= = =
Location of maximum velocity
The velocity is maximum where 0 / = dy du , which occurs at 2 / b y = .
The maximum velocity is
U b
dx
dp
U
2
3
8
1
2
max
= =

Flow rate as a function of pressure drop
Since the pressure gradient is constant, the pressure varies linearly with x
L
p
dx
dp
=
L
pb
w
Q
12
3
=

II. Fully developed laminar flow in a circular pipe

As a final example of fully developed laminar flow, let us consider fully developed
laminar flow in a pipe.

Control volume: Infinitesimal CV of volume rdrdx 2 .

Assumptions:
- Axial symmetry of the geometry and flow ( 0 / = )
- Infinite tube in the x-direction and fully developed flow: flow independent of x
( 0 / = x ) ) ( y u u =
- Flow driven by a constant negative pressure gradient G
dx
dp
= < 0
R
u(r)
Figure 7.3.
x
dx dr r dr
dr
d
) ( 2 ) ( + +

r
r
rdx 2
rdr dx
x
p
p 2 ) (
+ rdr p 2



surf
= 0
0 2 ) ( ) ( 2 ) ( 2 ) ( 2 ) ( = + + + + rdx dx dr r dr
dr
d
rdr dx
dx
dp
p rdr p

Keeping only the leading terms
0 2 2 2 = + + rdrdx
dr
d
drdx rdrdx
dx
dp

Dividing by rdrdx 2
dr
r d
r dr
d
r dx
dp ) ( 1
= + =
G
dx
dp
dy
r d
r
= =
) ( 1
(= const)
Since
dr
du
=
Gr
dr
du
r
dr
d
= ) (
A r
G
dr
du
r + =
2
2

r
A
r
G
dr
du
+ =
2

B r A r
G
u + + = ln
4
2

Boundary conditions:
o r
u
=
is finite A = 0
0 =
=R r
u
4
2
GR
B =
=
(
=
2
2
2
2 2
1 1
4
) (
R
r
U
R
r GR
r u
(parabolic velocity distribution)

where
4
2
GR
U =



Shear stress distribution
The shear stress distribution is given by
r
R
U
r
G
dr
du
2
2
2
= = =
Volume flow rate
The flux occurs through a annular region of cross-sectional area rdr 2
R
R R
A
R
r r R
dx
dp
rdr
R
r R
dx
dp
rdr u dA n u Q
0
2
4 2 2
0
2
2 2
0
4 2 2
1
4
2 2
(
=
(
= = =

r r

2
4
2 8
R
U
dx
dp R
Q
= =
Average velocity
The average velocity is given by
2 8
2
2
U
dx
dp R
R
Q
A
Q
U = = = =

Location of maximum velocity
The velocity is maximum where 0 / = dr du , which occurs at 0 = r .
The maximum velocity is
U U
dx
dp R
U = = =
2
1
4
2
max

Flow rate as a function of pressure drop
Since the pressure gradient is constant, the pressure varies linearly with x
L
p
dx
dp
=
L
pR
Q
8
4
= Hagen-Poiseuille formula for a steady laminar flow in a pipe



III. Extensions to Bernoullis equation
1. Bernoullis equation interpreted as an energy equation
We recall that Bernoullis equation was derived from Newtons second law applied to a
particle of fluid passing along a streamline. Bernoullis equation is only valid for an
inviscid, incompressible fluid flowing steadily.
When applied between two points 1 and 2, the equation may be written in the following
forms:
Energy/vol: const gz U p gz U p = + + = + +
2
2
2 1 1
2
1 1
2
1
2
1

Heads: const z
g
U
g
p
z
g
U
g
p
= + + = + +
2
2
2 2
1
2
1 1
2 2

Energy/mass: const gz
U p
gz
U p
= + + = + +
2
2
2 2
1
2
1 1
2 2

Bernoulli is an energy equation derived for frictionless (i.e. inviscid), incompressible
conditions with no energy input or extraction between 1 and 2.
It is a special form of the more general steady flow energy equation (SFEE). The latter
applies to real fluids with viscosity and compressibility, and can also account for heat and
work transfer to the fluid.
In this course, we shall only look at pipe losses due to friction, bends, valves, etc. and
work/energy input to an incompressible flow, and see how to represent these effects
simply by introducing additional terms to the Bernoulli equation.

2. Useful power output or power input

- Turbine with power output P and mass flow rate m& .
{
LOSS
ENERGY
OUTPUT INPUT
m
P
gz
U p
gz
U p
&
4 4 3 4 4 2 1 4 4 3 4 4 2 1
+ + + = + +
2
2
2 2
1
2
1 1
2 2

- Pump with power input P and mass flow rate m& .
P
m&
1
2
P
m&
1
2
Turbine Pump
Figure 7.4.


{
4 4 3 4 4 2 1
&
4 4 3 4 4 2 1
OUTPUT
GAIN
ENERGY
INPUT
gz
U p
m
P
gz
U p
2
2
2 2
1
2
1 1
2 2
+ + = + + +

These equations represent forms of the steady flow energy equation for incompressible
flows with no heat transfer.

3. Pipe losses
Let us represent a loss in useful energy in going from state 1 to state 2. Energy losses
can be due to friction, valves, bends, sectional changes, etc. In terms of heads and
energy/volume:
Heads:
{
LOSS
l
OUTPUT INPUT
h z
g
U
g
p
z
g
U
g
p
+ + + = + +
4 4 3 4 4 2 1 4 43 4 42 1
2
2
2 2
1
2
1 1
2 2

Energy/mass:
3 2 1
&
4 4 3 4 4 2 1 4 4 3 4 4 2 1
LOSS
OUTPUT INPUT
m
L
gz
U p
gz
U p
+ + + = + +
2
2
2 2
1
2
1 1
2 2

We define a head loss coefficient, k, such that the head loss is
g
U
k h
l
2
2
=

IV. Head losses due to friction
1. Friction losses
h
f
is the loss of total head due to viscosity effects (friction) and L
f
is the rate of energy
loss due to friction ) (
f f
gh m L & = .

Example 1: Straight horizontal pipe.

g
p
1
g
p
2
1
z
1 2
z z =
f
h
1 2
Figure 7.5.
U


2 1
u u = and
2 1
z z = , thus
g
p
g
p p
h
f

=
2 1

Example 2: pump discharging through a pipe with friction
f
gh gz
u p
gz
u p
m
P
gz
u p
+ + + = + + = + + +
3
2
3 3
2
2
2 2
1
2
1 1
2 2 2 &

2. Expression of head loss due to friction
The head loss due to friction is
g
p
h
f
=
To find an expression of the pressure difference, we can use the result derived when
looking at equilibrium of an infinitesimal control volume in laminar or turbulent flow
under a constant pressure gradient
L
p
dx
dp
dr
r d
r
= =
) ( 1

We found that the wall shear stress is
D
L
p
R
L
p
=
=
4
1
2
1
0

By definition of the friction coefficient
2
0
2
1
U f = , so
D
L
p
U f

=
4
1
2
1
2

Therefore
g
U
D
fL
g
p
h
f
2
4
2
=
Darcys formula
An expression of the head loss due to friction can be worked out using the following
formulae for the friction factor f for laminar and turbulent flows.
Figure 7.6.
P
m&
1
2
3


3. Friction factor in laminar flow
The laminar flow profile in a tube is parabolic.
The wall shear stress is given by
R
U
R
U
R
G
dr
du
R r

4 2
2
0
= = = =
=

The friction factor is
D U R U U R
U
U
f
16 8 1 4
2
2
1
2
2
1
0
= = = =
D
f
Re
16
= for laminar flow only
4. Friction factor in turbulent flow
The friction factor is mainly found experimentally: empirical relations are necessary.
In a turbulent flow, the time-averaged velocity is much flatter than laminar parabolic.

The velocity profile is approximated by power laws
n
R
r
U
u
/ 1
1
(
= , n ~ 5 10
One often takes n = 7, so the velocity profile is known as the one-seventh power law.

Several empirical relationships exist for turbulent friction factors for smooth pipes.
For example: Blasius equation
25 . 0
) (Re 079 . 0

=
D
f , for
5 3
10 Re 10 3
d

Or better
[ ]
2
6 . 2 ) ln(Re 52 . 1

=
D
f , for
9
10 ~ Re
d

r
R
Laminar profile
U
Turbulent profile
Figure 7.7.


V. Other pipe losses
Let us represent a loss in useful energy due to changes in cross-sections or bends in going
from state 1 to state 2:
Heads:
{
LOSS
l
OUTPUT INPUT
h z
g
u
g
p
z
g
u
g
p
+ + + = + +
4 43 4 42 1 4 43 4 42 1
2
2
2 2
1
2
1 1
2 2
with
g
U
k h
l
2
2
=
1. Abrupt contraction
An abrupt contraction occurs for example at the inlet at pipe and causes an entry loss.
From empirical results at relatively high Reynolds numbers, k = 0.5.

This k value corresponds to the flow entry from a reservoir (or large tank, or large pipe)
into a much smaller pipe. In the case of a pipe-to-pipe contraction, the value of k reduces
to zero as the diameter ratio reduces from to 1.0.
The entry loss can be alleviated by streamlining the entrance to the pipe (Figure 7.9 left),
or else it can be aggravated by miss-alignment (Figure 7.9 right).

Figure 7.9.
k ~ 1.0 k ~ 0.0
Flow
separation
Energy dissipation in
eddies
Figure 7.8.
k ~ 0.5


2. Abrupt enlargement

This problem is amenable to an approximate analysis. Assuming steady incompressible
flow at high Re with uniform flows at inlet and outlet, it can be shown from continuity
and momentum that the loss in total head from 1 to 2 is
2
1
2
2
1
2
2 1
1
2 2
) (
|
|
\
|
=
=
U
U
g
U
g
U U
h
l

From continuity
2 2 1 1
A U A U = , so
43 42 1
k
l
A
A
g
U
h
=
|
|
\
|
=
2
2
1
2
1
1
2

The corresponding loss in stagnation pressure is
2
2
1
2
1 0
1
2
1
|
|
\
|
= =
A
A
U p p
l

The previous results for a general abrupt enlargement can be manipulated to give the
submerged exit loss from a pipe discharging into a reservoir or large tank.
If we allow
2
A , then the exit loss is given by
1
U
2
U
p
2
A
2
p
1
A
1
Figure 7.10.
U U =
1
0
2
= U
The energy in the jet in the pipe
is dissipated in the turbulence at
the outlet
Figure 7.11.


g
U
h
l
2
2
1
= i.e. k = 1.0

An enlargement loss can be considerably reduced if the change in section is arranged to
be gradual, so producing what is called a diffuser. Diffusers aim at increasing the static
pressure in a flow by avoiding separation of the boundary layer from the wall.

Without separation of the boundary layer, there is negligible loss in stagnation pressure
(There is still some from pure friction effects).
= + = +
2
2 2
2
1 1
2
1
2
1
U p U p constant
From continuity,
2 1
U U > and therefore
2 1
p p < . An adverse pressure gradient like this
acting on slowly moving fluid in the boundary layer is what can cause boundary layer
separation. A well-designed diffuser reduces the pressure gradient to a level that the fluid
can sustain without separating.

3. Pipe bends and other fittings
Pipe bends and fittings lead to a head loss and are characterized by a head loss factor k
0 .

VI. Applications of SFEE with losses
1. Submerged outlet to reservoir
If the areas of the reservoirs/tanks are large compared with the pipe, then the velocities at
1 & 2 will be negligible and the flow can be regarded as essentially steady for a
considerable length of time.
Thus, we can apply the Steady Flow Energy Equation (SFEE) to the flow to work out the
mean flow velocity in the pipe, whilst the reservoir/tank levels are as shown.
p
1
2
U
1
U
A
1
p
2
A
2 Figure 7.12.


As the levels change, then the mean flow velocity will change, but we can use this
concept of quasi steady flow to get ) (t U as a function of t and then, for example, find
out how long it would take for the levels to equalise.

Looking at this snapshot of the system, lets find the mean velocity using the SFEE
equation
+ + = +
losses
a a
h z
g
p
z
g
p
2 1

=
losses
h z z
2 1

{
g
U
B
g
U
g
U
D
L
f
g
U
z z h
outlet
submerged
loss Exit
loss Friction loss Entry
2 2 2
4
2
5 . 0
2

2

2

2
2 1
= + + = =
43 42 1 3 2 1
with
(
+ =
D
L
f B 4 5 . 1
B
gh
U
2
=

2. Discharge to atmosphere

The SFEE equation is
+ + + = +
losses
a a
h z
g
U
g
p
z
g
p
2
2
1
2

1
2
p
a
p
a
d
z
1
z
2
L
U
h
Figure 7.13.
d
L
U
z
1
z
2
1
2
p
a
p
a
U
h
Datum
Figure 7.14.


+ =
losses
h
g
U
z z
2
2
2 1

g
U
B
g
U
D
L
f
g
U
g
U
z z h
loss Friction loss Entry
2 2
4
2
5 . 0
2
2

2

2 2
2 1
= + + = =
43 42 1 3 2 1
with
(
+ =
D
L
f B 4 5 . 1
( )
B
gh
B
z z g
U
2 2
2 1
=
=
It leads to the same expression for U , but notice that the value of h is different.

3. Quasi-steady pipe flow
The concept was introduced earlier. This refers to situations in which a flow is unsteady
(i.e. changes with time) but where the accelerations and forces producing accelerations
are so small that the SFEE can be applied with sufficient accuracy at any instant in time
during the flow.

Example: The draining of a tank or reservoir discharging to atmosphere
- Tank cross-sectional area A
t
is very large compared to the pipe area A
p

- h varies with time, but slowly
- Always draw system at an intermediate time t

SFEE between points 1 and 2:
+ + = +
losses
a a
h
g
U
g
p
h
g
p
2
2
2

p
a
1
2
L
h
Datum
p
a
U
Figure 7.17.
(constant diameter pipe)


+ =
losses
h
g
U
h
2
2
2

All the losses are proportional to g U 2 /
2
2
, so we can write
g
U
B h
2
2
=
where B is a coefficient that takes account of the dynamic head plus all the losses through
friction, bends, valves, expansions, etc.

Thus the quasi steady velocity at time t is
B t gh t U / ) ( 2 ) ( =
After a small time interval dt the level in the tank falls by a distance -dh. (Note the minus
sign as by convention dh is the increase in h occurring in time dt.)
Therefore, by continuity:
dt A U dh A
p t
= ) (
U
dh
A
A
dt
p
t
= , where B t gh t U / ) ( 2 ) ( =
The time to achieve level h
1
from an initial level h
0
at t = 0:
[ ] [ ]
5 . 0
1
5 . 0
0
5 . 0
5 . 0
1
2
2
2 2
1
0
1
0
h h
g
B
A
A
h
g
B
A
A
h
dh
g
B
A
A
t
p
t
h
h
p
t
h
h p
t
= = =

4. Conditions for pipes in parallel

Continuity:
B A
Q Q Q + =
SFEE:
1
2
z
1
z
2
1
U
2
U
Q
Q
A
Q
B
Q
Figure 7.15.


+ + + = + +
losses
h z
g
U
g
p
z
g
U
g
p
2
2
2 2
1
2
1 1
2 2

Fluid suffers the same head loss regardless of whether it goes through pipe A or pipe B
since the total head (or pressure) must be unique at 1 & 2.
Thus ( ) ( )
B l A l
h h = (Q
A
and Q
B
adjust themselves to satisfy this condition.)

5. Pipe bends and other fittings

SFEE between points 1 and 2:
+ + + = + +
losses
h z
g
U
g
p
z
g
U
g
p
2
2
2 2
1
2
1 1
2 2

( )
(
+ + + + + + = +
g
U
k k k
g
U
D
L
f z
g
p
z
g
p
2
...
2
4
2
3 2 1
2
2
2
1
1

Then rearrange to get U in terms of everything else, etc.

k
1
bend
k
2
bend
k
3
valve
Total pipe length, L,
Constant diameter d
U U U = =
2 1
d
1
U
2
1
z
1
z
2
2
U
Figure 7.16.


Chapter 8 External incompressible viscous flow

I. Flows past bodies
1. Flow around an airfoil
Steady air flow around an infinite airfoil. Lets assume first the flow to be incompressible
and frictionless.
Control volume taken around one blade.

Integral form of mass conservation equation
2 2 1 1
A u A u m = = &
Since
2 1
A A =
2 1
u u =
Since the flow is steady, incompressible and frictionless, one can apply Bernoullis
equation (the change of pressure due to gravity is negligible in air)
2
2 2
2
1 1
2
1
2
1
U p U p + = +
) (
2
1
) (
2
1
2
1
2
2
2
1
2
2 2 1
v v U U p p = =

Integral form of momentum conservation equation
Forces acting on the Control Volume:
- Pressure forces
- Reaction of airfoil R
r

- Gravity neglected
R dA n p dA n u u
CS CS
r
r r r r
+ =

) (

The x-momentum conservation equation reads
x
CS CS
R dA i n p dA n u u + =

. ) (
r r r



0 ) ( ) ( ) ( ) ( ) (
1 2 2 2 2 1 1 1
2 1
= = + = + =

u u m u A u u A u dA n u u dA n u u LHS
A A
&
r r r r

1 2 1 2 2 1 1
) (
. A p p A p A p dA i n p R
CS
x
= + = =
r

The x-momentum conservation equation reads
y
CS CS
R dA j n p dA n u v + =

. ) (
r r r

) ( ) ( ) ( ) ( ) (
1 2 2 2 2 1 1 1
2 1
v v m v A u v A u dA n u v dA n u v LHS
A A
= + = + =

&
r r r r

) (
. ) (
1 2
0
1 2
v v m dA j n p v v m R
CS
y
= + =
=
&
43 42 1
r
&

So the force acting on the airfoil per unit height is
b p p F
x
) (
2 1
= ; 0 ) (
1 2
< = v v m F
y
& Maybe it has to be negative to produce a lift
force (in the upwards direction).
Should Fx have the inverse sign? So far it seems to be positive (p1>p2). But are we
expecting a drag force in the x direction?

Check math. Problem of sign with v
y
x

b CV
F
y

Forces on
one blade
1
2
F
x

U
1
U
2
u
1

v
1

u
2

v
2

p
1

p
2

CV
Top view
z
x y
A
1
A
2

Fluids Dynamics Notes

Uploaded by

Copyright:

Available Formats

Fluids Dynamics Notes

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fluids Dynamics Notes

Uploaded by

Copyright:

Available Formats

Fluid Mechanics MECH1001 1

University College London Dr. V. Daz

= = (Newtons law of viscosity)

Shear stress at wall due to

Outer flow behaves as inviscid.

Shear stress at wall due to velocity

= and the Reynolds number

(parabolic velocity distribution)

= Hagen-Poiseuille formula for a steady laminar flow in a pipe

You might also like