FLUID DYNAMICS

for
Engineering Undergraduates

PART 1

Thusitha Sugathapala
Fluid Mechanics for Engineering Undergraduates - I 1

PREFACE

This book is written according to the syllabus of Level 2 Mechanical Engineering

specialization of four-year B.Sc. Engineering Honours Degree Programme at the
Faculty of Engineering, University of Moratuwa, Sri Lanka. Fluid Mechanics,
which includes static, kinematics and dynamics aspects of fluid systems, is a core
and fundamental subject area of Mechanical Engineering undergraduate degree
programme, usually extending over all the four levels of academic progressions.
Therefore, this module covers an intermediate level of fluid mechanics, with
emphasis on kinematics and dynamics aspects of fluid in motion. It is expected that
the readers have gone through a typical Fluid Mechanics at Level 1 of an
Engineering Degree programme, including fluid static and basic knowledge on fluid
in motion.

The five chapters in this book cover an introduction to fluid dynamics, governing
equations, ideal flow, viscous flow and dimensional analysis. The chapter on
governing equations presents the basic conservation laws applicable for fluids in
motion, typical boundary conditions and constitutive relations. It also highlights the
basic fluid flow analysis techniques. The ideal flow section introduces the basic
concepts of stream function, velocity potential and flow singularity modelling. The
viscous flow section includes three key areas: laminar flow of Newtonian fluids
through bounded systems including introduction to Moody chart, laminar flow of
non-Newtonian fluids through bounded systems and external flow around solid
bodies. Finally, the Chapter Five presents the concept of dimensional analysis and
similarity theory and covers a range of engineering applications including flow
around fully submerged bodies, free surface flows including motion of ships, and
distorted modelling.

Each section also contains number of exercises for the better understanding of the
fundamentals covered and also to illustrate their applications. A separate solution
manual will be prepared to facilitate the learning process of the students.

AGT Sugathapala
Thermo-Fluids & Energy Group
Department of Mechanical Engineering
University of Moratuwa
Sri Lanka.

September 2009.
 CONTENTS

PREFACE i

CHAPTER ONE: INTRODUCTION 01

1.1 Definition 01
1.2 Properties of Fluid in Motion 01
1.3 Classification of Fluids 02
1.4 Classification of Flows 03
1.5 Behaviour of real fluids: Laminar and Turbulent flow 03
1.6 Internal and External Flow Configurations 06
1.7 Basic flow analysis techniques 07
1.8 Equation of Motion 08

CHAPTER TWO: GOVERNING EQUATIONS IN FLUID MECHANICS 09

2.1 Continuum Approach 09
2.2 Methods of Description of Fluid Motion 10
2.3 Fluid Kinematics 10
2.3.1 Flow Lines 10
2.3.2 Translation, Rotation and Rate of Deformation 13
2.4 Fundamental Physical Laws of Fluid Mechanics 15
2.5 Basic Flow Analysis Techniques 15
2.6 Conservation of Mass - The Continuity Equation 16
2.7 Conservation of Momentum - The Equation of Motion 17
2.7.1 Conservation of Linear Momentum 17
2.7.2 Conservation of Angular Momentum 19
2.8 Conservation of Energy - The Energy Equation 20
2.9 Constitutive Relations 21
2.10 The Navier-Stokes Equation 22
2.11 The Bernoulli's Equation 22

CHAPTER THREE: IDEAL FLUID FLOW 29

3.1 Introduction 29
3.2 Governing Equations 29
3.3 Velocity Potential 31
3.4 Stream Function 32
3.5 Flow Nets 33
3.6 Flow Singularity Modelling 34
3.6.1 Introduction 34
3.6.2 Basic Flow Patterns 34
3.6.3 Combinations of Basic Flow Patterns 36

CHAPTER FOUR: VISCOUS FLOW 41

4.1 Introduction 41
4.2 Viscous Flow through Bounded Systems 42
Fluid Mechanics for Engineering Undergraduates - I 1

4.2.1 Applications 42
4.2.2 Entry Length and Fully Developed Flow 43
4.3 Laminar Flow through Simple Bounded Systems:
Theoretical Analysis 44
4.3.1 Formulation of the Problem 44
4.3.2 Flow in a Circular Pipe 45
4.3.3 Flow between Parallel Plates 46
4.3.4 Flow over a Plane (with a Free Surface) 47
4.4 Energy Loss in Pipe Flow: Darcy Formula 47
4.5 Laminar Flow of Non-Newtonian Fluids through Bounded Systems 53
4.5.1 Introduction 53
4.5.2 Classification of Non-Newtonian Fluids 53
4.5.3 Flow through a Circular Pipe 55
4.6 External Flow: Flow around Solid Bodies 60
4.6.1 Introduction 60
4.6.2 Flow Regions around a Body 60
4.6.3 Types of Forces: Drag and Lift 62
4.6.4 Types of Body Shapes: Bluff Bodies & Streamlined Bodies 62
4.6.5 Pressure and Force Coefficients 63

CHAPTER FIVE: DIMENSIONAL ANALYSIS AND SIMILARITY THEORY 68

5.1 Introduction 68
5.2 Fundamental and Derived Units and Dimensions 68
5.3 Dimensional Homogeneity 69
5.4 Methods of Dimensional Analysis 69
5.4.1 Rayleigh's Method 69
5.4.2 Buckingham's  - theorem 70
5.5 Physical Significance of Non-dimensional Groups 71
5.6 Theory of Physical Similarity and Model Testing 73
5.6.1 Physical Similarity 73
5.6.2 Model testing of flows without free surfaces 75
5.6.3 Model testing of free surface flows 76
5.7 Distorted Models 77
Fluid Mechanics for Engineering Undergraduates - I 1

CHAPTER ONE

INTRODUCTION
1.1 Definition

Fluid Mechanics is a branch of general mechanics which deals with static,

kinematic and dynamic aspects of fluids.

Fluid Static is the study of fluid at rest. Fluid Kinematics is the study of the motion
of fluid particles without consideration of the characteristics of the particles or the
influence of forces upon their motion. Fluid Dynamics is the study of the motion
with taking into consideration of the forces associated with the flow.

1.2 Properties of Fluids in Motion

The following properties or parameters of fluids and flow characteristics are of

general importance in the study of fluid mechanics:

 Density , Viscosity , pressure p, Bulk modulus K } Scalars

 Velocity (linear/angular) v 
Acceleration a, force  Vectors (tensor of order 1)
 Stress (direct/shear)   Vectors (tensor of order 2)
Strain (direct/shear) , 

Some of the above parameters could be identified as fluid properties such as

density, viscosity and bulk modulus and some as flow properties such as velocity. In
addition to the fluid properties and flow characteristics, general fluid flow
problems are governed by geometrical properties or parameters, such as shape, size
and surface roughness of solid boundaries. For example, In the case of flow through
orifice meter, diameters of the pipe and the orifice are geometrical properties of the
flow. Another way of categorizing the above parameters is as independent
parameters and dependent parameters. In the case of flow through a pipe fluid
properties and geometry are independent parameters and flow rate is a dependent
parameter.

 Viscosity:

One of the main characteristics of a real fluid is the resistance to flow due to
molecular friction between fluid particles. The property of the fluid which offers
this internal resistance is called the viscosity of the fluid. Newton (1687) postulated
that for a straight and parallel motion of a given fluid, the shear stress between two
adjacent layers is proportional to the velocity gradient in a direction perpendicular
to the layers.
Fluid Mechanics for Engineering Undergraduates - I 2

Consider a fluid moving with a velocity u parallel to the x-axis of a rectangular co-
ordinate system x-y. According to the Newton’s law the shear stress is proportional
to the velocity gradient in the transverse direction, that is
du du
  or    .
dy dy
This equation is known as Newton’s law of viscosity and any fluid which obeys this
law is referred to as Newtonian fluid. The coefficient of proportionality  is known
as the coefficient of viscosity or the dynamic viscosity of the fluid. The kinematic
viscosity  is defined as  = /. The units of  and  are Ns/m2 and m2/s
respectively. Note that the above expression is very similar to stress-strain
relationship of a solid material given by  = E.

It can be shown that for a straight and parallel flow the velocity gradient is equal to
the rate of change of shear strain. Hence, the Newton’s law of viscosity describes a
fundamental constitutive relationship of fluid in motion as shear stress is
proportional to the shear rate.

The above form of Newton’s law of viscosity is strictly applicable for straight and
parallel flow. Flow around a bend and flow through a divergent or convergent duct
section are some applications for which more general form of stress-strain
relationship is to be derived.

Exercise 1.1
Shear strain of a fluid in motion is defined as the change in 90 angle of a fluid
element. Show that, for a fluid in straight and parallel motion, the velocity gradient
is equal to the rate of change of shear strain.

1.3 Classification of Fluids

Fluids may be classified by considering their various properties.

(a) By considering the density:

 Incompressible fluids (negligible variation in density during the flow ),

 Compressible fluids ( varies significantly during the flow).

(b) By considering the viscosity:

 Inviscid fluids (or frictionless or non-viscous fluids)  = 0,

 Viscous fluids (real fluids): Viscous fluids could be further categorized
according the variability of their viscosities as
- Newtonian fluids (obeys Newton’s law = du/dy, where  is a constant)
e.g. Water, air.
- Non-Newtonian fluids (the viscosity is dependent on the velocity gradient)
e.g. Paints, clay, liquid foods, biological fluids, plastics etc.
Fluid Mechanics for Engineering Undergraduates - I 3

1.4 Classification of Flows:

In general the flow parameters such as velocity, pressure and density, which
describe the behaviour of a fluid, are not constant in a particular set of
circumstances. They may vary from one point to another or from one instant of time
to another. That is u  u(x, y, z ,t); p  p(x, y, z, t) etc.

Therefore fluid flows may be classified by considering their time variations or

spatial variations.

(a) Steady and Unsteady Flow

If the flow parameters at different points in the region do not vary with time, the
flow is said to be steady. The time dependent flow is termed unsteady flow.

e.g. Steady flow: u  u(x, y, z); p  p(x, y, z); etc.

Unsteady flow: u  u(x, y, z, t); p  p(x, y, z, t); etc.

(b) Uniform and Non-uniform Flow

A flow region is called uniform if the flow parameters do not vary from point to
point over the region at a given instant of time. In non-uniform flow these
parameters are functions of the position.

e.g. Uniform flow: u  uo(t); p  po(t) etc.

Non-uniform flow: u  u(x); p  p(x) etc.

(c) One, Two or Three-Dimensional Flow

In general fluid flows are three-dimensional (3-D) in the sense that the flow
parameters vary in all three co-ordinate directions (and time). Considerable
simplifications in analysis may often be achieved by selecting the co-ordinates
directions so that appreciable variation of the parameters (such as velocity, pressure
etc.) occurs in only two directions or even in one direction. Such types of flow are
called two-dimensional (2-D) or one-dimensional (1-D) flow, respectively.

e.g. 3-D flow: u  u(x, y, z, t); p  p(x, y, z, t); etc.

2-D flow: u  u(x, y, z, t); p  p(x, y, t); etc.
1-D flow: u  u(x, t); p  p(x, t); etc.

1.5 Behaviour of real fluids: Laminar and Turbulent flow

Real fluid flow can be of two kinds. At relatively low velocities (i.e. low inertial
forces), the influence of viscosity is predominant such that the fluid particles move
in a layer (or laminae), not necessarily in a straight line, without disturbing the fluid
Fluid Mechanics for Engineering Undergraduates - I 4

particles moving in adjacent layers. This type of flow usually called laminar flow
(or stream flow or viscous flow). With increasing velocity (i.e. increasing inertial
forces) the flow changes from laminar to turbulent. In turbulent flow fluid particles
move in disorderly manner with continuous and rapid mixing with adjacent fluid
particles (see Figure 1.1).

Velocity
VA
 A 
VB VA
B

VB

Time

Figure 1.1: Laminar flow

Laminar flow is deterministic in nature. The influence of viscosity is predominant in

laminar flow to damp out any disturbance which could cause the flow to become
turbulent. Laminar flow of Newtonian fluid is governed by Newton’s law of
viscosity.

At higher velocities, the influence of viscosity is not sufficient to damp out any
disturbance that may set in the flow (as example due to irregularities on a solid
surface) and the flow becomes turbulent. Turbulent flow is random in nature.
Turbulent flow can be described only in terms of its mean properties and
fluctuations of properties. Turbulent set up much greater shear stresses causing
higher energy losses. Newton’s law of viscosity is not applicable for turbulent flow.

Velocity


VA VB VA
B
 A
VB

Time

Figure 1.2: Turbulent flow

Fluid Mechanics for Engineering Undergraduates - I 5

In a fluid in motion, internal resistance is generated through two different

fundamental mechanisms: inter-molecular forces and momentum exchange between
fluid layers. In laminar flow, both these mechanisms exist in molecular level and
momentum exchange due to random motion of molecules is negligible and the
viscosity entirely depends on inter-molecular forces (i.e. molecular friction).
However, in turbulent flow random motion of fluid particles results in large-scale
mixing and the momentum exchange between fluid elements surpasses the
molecular effects resulting much higher frictional losses (i.e turbulent viscosity is
much higher than laminar viscosity). The shear stress in turbulent flow can be
related to fluctuations in velocity components.

The experiments of Reynolds (1883) were the first to illustrate the difference
between laminar and turbulent flow. In these experiments, water was allowed to
flow through a glass pipe. A thin stream of dye was injected and its motion through
the pipe was observed. At low velocities, the dye moved in a straight line along the
tube indicating laminar flow. As the velocity was increased, the dye become thinner
and began unsteady wave like motion. A further increase in the velocity caused the
dye to break up into segments indicating turbulent flow.

Dye injection

Air
Control valve

Water
Visual test section To drain

 Low flow-rate Laminar

 Medium flow-rate Transition

 High flow-rate Turbulent

Figure 1.3: Reynold’s experiment for laminar and turbulent flows

These observations show that the relative magnitude of inertial and viscous forces
determines the type of flow. Let the characteristic values for length and velocity are
l and u respectively.

Inertial force Fi = mass  acceleration = density  volume  acceleration

 u  u
Fi  l 3 u   l 3 u   l 2 u 2
 y  l
Fluid Mechanics for Engineering Undergraduates - I 6

Viscous force Fs = shear stress  area =   A

 u  u
Fs    l 2    l 2  ul
 y  l
Then
Fi u 2 l 2 ul
 
Fs ul 

This ratio is a fundamental characteristic of the flow and is know as Reynolds

Number (Re). ul ul
Re   .
 
e.g. (i) Flow around a sphere of diameter d moving with velocity U0, Re=U0d/..

(ii) For the flow through a circular pipe of diameter d and mean flow velocity
U, Re = Ud/. For this case when
- Re < 2000, flow is laminar,
- Re > 4000, flow is turbulent,
- 2000 < Re < 4000, flow is transition (between laminar and turbulent).

Note that the above specific values are applicable only for the flow through a
straight circular pipe only, and not applicable for non-circular pipes / ducts or for a
curved pipe even with a circular section.

1.6 Internal and External Flow Configurations

Most of the fluid flow problems are associated with solid boundaries, and can be
divided into two groups.

(a) Internal Flow:

Internal flow is that which is surrounded partly or entirely by rigid boundaries. Flow
through pipes and ducts, and flow in rivers and channels are few examples.

Flow U

(i) Flow though conduits (ii) Flow in open channels

Figure 1.4: Internal flow

Note: In these situations the volume of the fluid involved is limited and problems
are governed by the volumetric flow rates. Usually solid boundaries are at
rest and the fluid is in motion.
Fluid Mechanics for Engineering Undergraduates - I 7

(b) External Flow:

External flow is that in which the fluid flow takes place outside solid boundaries /
objects. Flow around buildings and structures, flow around vehicles and other
moving bodies, flying of birds are some examples. Usually the volume of fluid
associated with the flow is unlimited and the volumetric flow rates are not involved
in these problems. Pressure distribution and related forces are the important
parameters in the problems.

Wind

(i) Flow around building and structures (ii) Flow around moving bodies

Figure 1.5: Internal flow

1.7 Basic Flow Analysis Techniques

Fluid static problems basically governed only by the density of the fluid and the
vertical height below the free surface. But most fluid flow problems require the
analysis of an arbitrary state of variable fluid motion defined by the geometry, the
boundary conditions, and the laws of mechanics. There are three basic approaches
to attack a fluid flow problem.

(a) Large-scale analysis based on control volume technique,

(b) Small-scale analysis based on fluid element technique,
(c) Experimental analysis based on dimensional analysis similarity theory.

In the control volume technique, a fluid volume of finite size appropriate for the
flow configuration is selected. The governing laws are applied for the fluid passing
through this control volume (which may be fixed or moving). Details of the flow are
usually ignored and therefore this method provides approximate but useful
quantitative information on the overall effects of the flow.

For some problems integral analysis does not provide a solution and/or more detail
is needed. Fluid element technique, which attempts to describe the fluid flow at
each point in the flow field, may provide the solution in such situations. In this
method the governing laws are applied to an infinitesimal fluid element with a fixed
mass. The results yield the basic differential equations of fluid flow. These
equations are then integrated subject to the boundary conditions of a particular
problem.
Fluid Mechanics for Engineering Undergraduates - I 8

Rotor
Fluid Element
of mass m

dy
Control Volume 
(x, y)
dx

Control Surface

(a) Flow around a wind turbine rotor (b) Free surface flow over a plane

Figure 1.6: Control volume technique and fluid element technique

The above two techniques are theoretical methods and usually can provide
satisfactory solutions for only for simple flow configurations, unless numerical
techniques with computer tools are used. However, practical fluid flow systems
very often cannot be formulated mathematically because the phenomenon to be
investigated is too complex to be described by a satisfactory model. In these cases,
experimental methods of investigation have to be used. Even for theoretical
modeling, the final validation of the results is done through laboratory and/or field
testing. More details of these techniques are discussed in the next sections.

Exercise 1.1 There are three basic techniques of analysis of a fluid flow problem:

(i) Large-scale analysis based on control volume technique,

(ii) Small-scale analysis based on fluid element technique,
(iii) Dimensional Analysis based on experimental technique.

Explain clearly the above three approaches of solving fluid flow

problems stating one practical application in each.

1.8 Equation of Motion

The general equation of motion of fluid is the Navier-Stokes Equation. This

equation indicates that the fluid flow around a body is governed by four kinds of
forces: inertial force, viscous force, pressure gradient force and body forces (such as
gravity). For inviscid and incompressible fluid (i.e. ideal fluid), the Navier-Stokes
equation can be integrated to obtain the Bernoulli's equation:
p V2
  z  Constant , along a streamline , or
g 2 g
1
p  V 2  Constant, neglecting gravity effects.
2
Fluid Mechanics for Engineering Undergraduates - I 9

CHAPTER TWO

GOVERNING EQUATIONS
IN FLUID MECHANICS
2.1 Continuum Approach

As any other material, fluids consist of molecules. Since molecules move randomly,
the properties of a fluid, such as velocity, density, etc., have a violently non-uniform
distribution when the fluid is viewed on molecular scale (microscopic scale). In fact,
it is impossible to measure the positions and velocities of individual molecules.
However, in general, fluid mechanics is concerned with behavior of matter in a
larger scale (macroscopic scale) where the scale of interest is much larger than the
distance between molecules. Then the molecular structure may be ignored and the
fluid is assumed as a continuous matter where the physical and dynamical properties
of the fluid vary continuously from point to point. These macroscopic properties are
defined by averaging over a small volume of fluid, known as fluid particle, which
consists of a large number of molecules but still be effectively at a point with
respect to the flow as a whole (see Figure 1.1). Thus the fluid is considered to be
consists of continuous aggregate of fluid particles. This concept is known as
continuum hypothesis, and is the common approach used to derive the governing
equations of fluid in motion.
Measured
Property
Variation due to Variations associated with
Molecular fluctuations spatial distribution of the
property

Local value of
the property

10-10 10-7 10-4 10-1

Particle Size a (m)

Figure 2.1: Variation of the average value of a property with side a of a

cuboid particle in air.

The continuum hypothesis implies that at each point there is a fluid particle and the
region of flow consists of a continuous aggregate of such particles, each having a
Fluid Mechanics for Engineering Undergraduates - I 10

certain velocity, density, pressure etc. and that in general the values of these
properties are continuous functions of position and time. Note that the properties of
a fluid which are due to molecular effects such as viscosity, heat conductivity, etc.,
enter the governing equations as empirical parameters obtained by experiments.

2.2 Methods of Description of Fluid in Motion

There are two different forms of representation for flow field in fluid mechanics:
Lagrangian and Eulerian. In the Lagrangian method, which is more appropriate to
solid mechanics, the physical variables are described for a given fluid particle as it
traverses the flow. The fluid particle is identified by its position in the flow field,
say ro = (xo, yo, zo), at some arbitrary time (usually t = 0). Then the value of a
physical variable, say V, is written in functional form as
V  V ( xo , yo , z o , t ).
The Eulerian method, which is more appropriate to fluid mechanics, leads to
expressions that give physical variables at each point in the flow field as a function
of the coordinates of the point r = (x, y, z) and time t. Then the functional form is

V  V ( x, y, z, t) , where x, y, z and t are all independent variables.

Thus, the Lagrangian method traces each fluid particle and describes the conditions
of that particle, whereas the Eulerian method simply describes the conditions at a
point without reference to which particle occupies the point. Both these methods
attempt to describe the fluid flow at each point in the flow field and therefore
resulting equations are generally in differential form. Therefore it is important to
establish the relation between the derivatives of the two methods, which is given in
the next paragraph.

Fixed Mass
(Particle) u = u (ro, t) u = u (r, t)
t=t t=t

t=0 r = r(ro, t) r

ro
Fixed Rigid
Rigid Surface Body
Body

(a) Lagrangian Representation (b) Eulerian Representation

Figure 2.2: The two methods of description of flow field around a rigid body
Fluid Mechanics for Engineering Undergraduates - I 11

Let V represents any physical variable of the fluid and its value is changed by V
during a short time duration t. Therefore in the Lagrangian method of description
the rate of change of V is given by V/t. Whereas in the Eulerian method of
description V is considered to be a function of the independent variables x, y, z and
t, and the change in V during the time t may be calculated from differential
calculus as
V V V V
V  t  x  y  z .
t x y z
At the limit t  0 above equation can be rearranged to obtain
DV V V V V
 u v w ,
Dt t x y z
where u = (u, v, w) is the velocity vector and DV/Dt = t  0 {V/t}. Above
equation may be rewritten as
DV V V V
  ( u. )V   uk .
Dt t t xk
The term DV/Dt is called the material derivative (or total derivative) which
represents the total change in V as observed in Lagrangian framework (Lagrangian
derivative). The entire right-hand side represents the total change in V in Eulerian
framework (Eulerian derivative) in which the two terms are called the local rate of
change and convective rate of change, respectively. Note that the material derivative
of the position vector gives the velocity and the material derivative of the velocity
vector gives the acceleration.

Exercise 2.1 A two dimensional velocity distribution (in Eulerian method) is given
by u = x/(1+t) and v = y/(1+2t). Derive expressions for

(a) the position, velocity and acceleration in Lagrangian form,

(b) the local, convective and total acceleration in Eulerian form.

2.3 Fluid Kinematics

2.3.1 Flow Lines

Fluid kinematics is basically deals with pattern of fluid in motion, in which motion
of the fluid is analyses without considering the forces associated with it. Usually
understanding of fluid kinematics is very useful especially for solving more
complicate fluid flow problems. Analysis of ventilation inside a building and
movement of air pollutants emitted from a chimney of a power plant are two
examples.

Frequently flow visualization plays important part in understanding complex fluid

flow problems. There are three basic types of flow lines which are used for flow-
visualizations. These are:
Fluid Mechanics for Engineering Undergraduates - I 12

(a) Streamline: A line in space which is tangent to the velocity vector at a given
instant of time. As illustrated in Figure 2.3(a), the velocity of
each fluid particle is tangent to the streamline it belongs to, and
the set of streamlines represents instantaneous picture of the
directions of the velocity vector field of the flow.
(b) Streak line: The locus of all particles which have earlier passed through a
prescribed point. Figure 2.3(b) shows two streak lines: one
containing the particles P1, P2 and P3, which have gone through
the point A at earlier instances of time, and the other containing
the particles P4, P5 and P6, which have gone through the point B
at earlier instances of time.
(c) Path line: The actual path traversed by a given fluid particle. The two path
lines shown in Figure 2.3(c) corresponds to the particles P1 and
P2, which traverse in the flow region while the time passes
through t0, t1 to t2.
t=t0 P2 t=t0
U2 U6 P 1 U0
U5
P2 P6 P1 V0 P
U1 U4 t=t1
2
P4 P4 P5 t=t 1
U3 U4 U3
P1 U6 P3 V1
U5 U1
U2
P3 P6 B
P5 U1 P2 t=t2 P1 P2
V t=t2
t = t* = Fixed A t = t* = Fixed U2 2
P1
t > t1 > t 0
(a) Streamlines (b) Streak lines (c) Path lines
Figure 2.3: Flow lines.

For a general flow field, these lines are all different, but for steady flow they
become identical. The streamline is convenient to calculate mathematically, while
streak line and path line are easier to generate experimentally. Note that a streak line
can be visualized by injecting a neutrally buoyant coloured fluid in to the flow at a
fixed point. A path line can be found by a time exposure of a single neutrally
buoyant marked particle moving through the flow. For instance, smoke emitted
from a chimney shows streak lines while trajectory of a bird during flying is a path
line.

The streamline pattern for a known velocity distribution u(x, y, z) = (u, v, w) is

given by the pair of differential equations
dx dy dz
  .
u v w
The streamline passing through the point (x0, y0, z0) can be determined by
integrating above set of equations, for constant time t. Path line is defined by time
integration of the relation between velocity and displacement, given by
Fluid Mechanics for Engineering Undergraduates - I 13

dx dy dz
 u,  v and  w.
dt dt dt
with the initial condition (x0, y0, z0 , t0). Elimination of time leads to the equation of
the path line. The integrated results of above equation may be used to obtain the
equation of streak line. Here the integration constants have to be evaluated as
functions of a time parameter  (say) so that the path lines pass through (x0, y0, z0)
for a sequence of times  < t. The elimination of  gives the required expression.

Exercise 2.2 Derive the equations of the streamline, path line and streak line
which pass through the point (1, 1, 0) at time t = 0 in the flow field
with the velocity distribution u = x(1+2t), v = y and w = 0.

2.3.2 Translation, Rotation and Rate of Deformation

In general motion of a fluid element may consists of a translation, a rotation, a

dilatation (change in size) and a deformation (change in shape – both linear and
angular), as shown in Figure 2.4. In general, the change in shape (both linear and
angular) of the element, that is the rate of shear deformation  has a complicated
form. For incompressible flow the rate of linear (or normal) strain is given by u/x,
v/y and w/z in the three directions of Cartesian coordinates. The rate of shear
strain (which represents the change of 90 angle) in 2-D flow is given by ½(v/x+
u/y).

Pure Translation Pure Rotation Linear D eformation

Angular Deformation Dilatation

Figure 2.4: Basic modes of fluid motion

Mathematical expressions for these modes of motion may be obtained by analysing

the movement of a fluid element. Consider two neighboring fluid particles in the
element. Let the velocity of the particle at a point r is u(r, t) and the velocity of the
Fluid Mechanics for Engineering Undergraduates - I 14

particle at a neighboring point r + dr is u(r + dr, t). The velocity of the second
particle can be expanded as a Taylor series about the point r as
u u u u
u(r  dr )  u(r )  du  u(r )  dx  dy  dz  u i  i dx j ,
x y z x j
neglecting higher order terms. In this expression, the term ui represents the
translation of the fluid element. Consider the last term of the above expression. It
can be rewritten in the form
u i 1  u u j   u
 and d ij  1  u i  j

dx j   ij dx j  d ij dx j , where  ij   i  .
x j 2  x j xi 
 2  x j xi 


It can be shown that the anti-symmetric term ij represents the rate of rotation
(angular speed) and the symmetric term dij represents the rate of deformation (or
strain) of the element. Note that the term dij represents the angular deformation (i.e.
shear strain) when i  j which results in change in shape only. However, when i = j
the term dij (i.e. direct strains) represents the change in shape (linear deformation) as
well as change in size (dilatation). The mean value of the rate of normal strain
represents the change in size. That is
1 1  u v w  1
Dilatation  d ii       .u
3 3  x y z  3

Therefore the change in shape (both linear and angular) of the element, that is the
rate of shear deformation  is given by
1  u u j  1 u k
 ij   i    ij ,
2  x j xi  3 x k
where ij is the Kronecker delta. Note that for incompressible flow, .u = 0
(continuity equation) and the dilatation is zero. The angular speed of the fluid is
represented by the tensor ij which in vector form is
1 1
  u  ,
2 2
where  is defined as the vorticity. If the vorticity is zero everywhere in the flow
region, the flow is said to be irrotational. That is for irrotational flow   u  0 .

Note: The circulation  around a closed curve in the flow is defined by the line
integral of velocity as C u
   u .dl.
C
dl

Now the Stokes theorem may be used to convert the line integral to a
surface integral as  
 u .dl    u.dS   .dS ,
C S S
where S is the surface bounded by the curve C. Then the circulation per unit
area is the vorticity. Therefore, the circulation is zero in irrotational flow.
Fluid Mechanics for Engineering Undergraduates - I 15

Exercise 2.3 Derive an expression for the angular velocity of the fluid for the
following two dimensional flows and comments on the results:
(a) Simple shear flow with the velocity distribution u=(ay, 0), where
a is a constant.
(b) Free-vortex flow (i.e. circular flow with velocity inversely
proportional to the radius).
(c) Forced-vortex flow (i.e. circular flow with velocity proportional
to the radius).
Also calculate, for the cases (b) and (c) above, the circulations around
circular contours of radii R and 2R and comment on the results.

2.4 Fundamental Physical Laws of Fluid Mechanics

There are three basic conservation laws that apply to the analysis of fluid flows
independent of the nature of the fluid under consideration:

(a) Conservation of mass (Continuity)

(b) Conservation of Momentum (Newton's second law of motion)
(c) Conservation of Energy (First law of thermodynamics)

In addition the flow must satisfy constitutive relations (e.g. Newton's law of
viscosity, Fourier's law of heat conduction), thermodynamic state relations (e.g. p =
p(, T)) and appropriate boundary conditions.

In general mechanics, basic physical laws are applied to a system, which consists of
a fixed quantity of mass. The resulting equations apply to either solid or fluid
systems. They are more appropriate for solid mechanics, where the same system is
followed throughout. However majority of fluid flow systems consist of well-
defined region through which the fluid flow takes place and therefore the fluid mass
occupying the region changes from instant to instant. Then the more convenient
method of analysis is to apply the basic physical laws to a specific finite region
fixed or moving in space through which fluid flows. Such region is known as a
control volume.

2.5 Basic Flow Analysis Techniques

As briefed in Section 1.7, there are three basic techniques of analysis of a fluid flow
problem: Large scale analysis (based on control volume technique), Small-scale
analysis (based on fluid element technique) and Experimental analysis (based on
dimensional analysis and similarity theory).

In large scale analysis analysis, a control volume of finite size appropriate for the
flow configuration is selected (see Figure 1.5(a)). Mass, momentum and energy are
allowed to pass across the control surface and the conservation laws are applied to
the fluid. Details of the flow are usually ignored and therefore this method (system
Fluid Mechanics for Engineering Undergraduates - I 16

approach) provides approximate but useful quantitative information on the overall

effects of the flow (eg. Forces acting on pipe fittings).

For some problems integral analysis does not provide a solution and more details
are needed. Differential analysis, which attempts to describe the fluid flow at each
point in the flow field may provide the solution in such situations. In this method,
the physical laws are applied to an infinitesimal fluid element (i.e. constant mass
system). The result yields the basic differential equations of fluid flow. These
equations are then integrated subject to the boundary conditions of a particular
problem. Exact analytical solutions are possible only for very simple geometrical
shapes and boundary conditions (eg. the velocity distribution of laminar flow
through a circular pipe). The alternative method of solution is the numerical
integration on a computer. Even computational analysis often fail to provide an
accurate simulation, because of either inadequate storage or inability to model the
finely detailed flow structure characteristics of irregular geometrical shapes or
turbulent flow patterns (eg. wind flow around a building).

Most of the practical fluid flow systems require the analysis of an arbitrary state of
variable fluid motion defined by the geometry, the boundary conditions and the
laws of mechanics. Very often the problem cannot be formulated mathematically
because the phenomenon to be investigated is too complex to be described by a
satisfactory model. In these cases, experimental methods of investigation is often
the best way to establish the essential physical features of the problem. Dimensional
analysis and similarity theory determine the conditions under which the model
experiments are to be carried out so that the experimental data gives an accurate
estimate of the prototype (full-scale system). Some of the areas of application of
model testing are aerodynamics of building and structures, aerodynamics of aircraft,
aerodynamics of ground vehicles, wind turbines, wave energy generators,
hydrodynamics of ships, river and harbor modeling.

2.6 Conservation of Mass - The Continuity Equation

Usually the principle of conservation of mass is not mentioned in constant mass

systems. However, in control volume systems mass should be explicitly conserved.
Consider a fixed, closed surface S in the fluid containing volume V, as shown in
Figure 2.5.
dS
S
u
V

Figure 2.5: A control volume for the derivation of continuity equation

Fluid Mechanics for Engineering Undergraduates - I 17

The conservation of mass implies that the rate of increase of mass in the volume V
is equal to the net rate of flow of mass into V through the surface S. The mass of
fluid containing in V is given by m0 = V dV, where  is the fluid density. The net
rate of flow of mass into V through the surface S is given by min = -S u.dS, where u
is the velocity vector. Therefore, the mathematical statement of conservation of
mass can be represented by
d   
dV   u.dS 
  
dV  u.dS  0. 
dt  V 
 S V
t S

In many practical applications, flow field is bounded by solid surfaces and the flow
crosses the control surface only a number of inlets and outlets where the flow
properties across the cross sections are nearly uniform (i.e. one-dimensional). In
such cases the surface integral of the above equation may be reduced to a simple

 
u i S i out  u i S i in
form as

u.dS 
S i i

This expression is very useful in integral analysis which provides an approximate

but quick solution to many fluid flow problems. This equation, given in integral
form, can be simplified further by using the divergence theorem to obtain a
differential equation, as follows:
    
 t dV   u.dS
V S

V
 t dV   .udV
V V
 .udV  0.

  t
Since V is an arbitrary volume, above expression is true only if the integrand is zero.

 .u  0.
That is
t
This is known as the continuity equation. For an incompressible fluid  = constant,
and the continuity equation reduces to .u  0.

If the control volume has only a number of one-dimensional inlets and outlets, for
steady flow the continuity equation simplifies to
  u S 
i
i ni i out    u S 
i
i ni i in or m in  m out .

2.7 Conservation of Momentum - The Equation of Motion

2.7.1 Conservation of Linear Momentum - The Equation of Motion

The conservation of linear momentum (or momentum equation) is obtained from

Newton's Second Law. This states that, when considering a given mass (i.e.
Lagrangian frame) the rate of change of momentum of a fluid mass is equal to the
net external force acting on the mass. The external forces include body forces (such
as gravitational or electromagnetic forces) and surface forces (such as pressure or
viscous forces). Let f represents the resultant body force per unit mass and P
represents the resultant surface force per unit area. Then the net external body force
Fluid Mechanics for Engineering Undergraduates - I 18

on a mass of volume V will be V fdV and the net external surface force acting on
the surface S containing V will be S PdS, in which P contains nine components of
stress ij (i,j = 1,2,3 representing x, y and z directions) as shown in Figure 2.6.

22 ij = Stress in the i

direction on a surface
12
normal to the j axis
32
x2 21
23
13 11
33 31
x1
x3

Figure 2.6: Notation for stresses

Figure 1.5: Notation for stresses
Therefore the application of conservation of momentum leads to

F
D
Dt V 
udV    fdV    PdS.
V S

As in the previous section, it can be shown that for a control volume with one-
dimensional approximation the momentum equation reduces to

F   m u 
i
i out   m u 
i
i in    u S u 
i
i ni i i out    u S u 
i
i ni i i in ,

under steady flow condition, where un represents the velocity component normal to
the surface S.

Exercise 2.4 A jet of water with a uniform velocity U and cross sectional area A
strikes an inclined fixed plate and breaks in to two jets of equal
velocity V = U, as shown in the figure. Determine the force acting
on the plate and the ratio of the flow rate of the two jets (i.e. Q1/Q2)
deflected by the plate. Interpret the above condition of equal
velocity V = U, and discuss the validity and effects of this condition
in real flow. . Q1

, A, U


Q2
Fluid Mechanics for Engineering Undergraduates - I 19

2.7.2 Conservation of Angular Momentum

This states that, when considering a given mass the rate of change of angular
momentum of a fluid mass is equal to the net external moment acting on the mass. It
can be shown that the application of the conservation of angular momentum to a
fluid element leads to the well-known result
 ij   ji when i  j.
Therefore there is no differential equation of angular momentum. Note that when i 
j, ij represents the shear stresses, hence shear stresses are symmetrical.

However, control volume analysis can be applied to the angular momentum relation
to obtain an expression, which has direct application to many important fluid flow
problems involving torque and moments (e.g. turbo-machinery).

2.7.3 Thermodynamic and Mechanical Pressure

It is useful to identify certain properties of surface stress system before proceeding

to the next conservation equation. When a fluid is at rest, the stress is hydrostatic.
That is   0
ij when i  j and    p,
ii

where the pressure p is defined as the hydrostatic pressure or thermodynamic

pressure (since the fluid is in equilibrium). However in a moving fluid 11  22 
33 and it is difficult to define a pressure in terms of thermodynamic relations, as
fluid elements are not in exact thermodynamic equilibrium. Therefore the pressure
at a point in a moving fluid may be defined by the mean normal stress, which has a
mechanical significance, as
p    11   22   33 .
1
3
This is known as the hydrodynamic pressure or mechanical pressure. Note that this
is a purely mechanical definition of pressure and the precise connection of this with
the thermodynamic pressure in a moving fluid is not simple. This pressure includes
hydrostatic plus a component which is induced by the stresses which results from
the motion of fluid and therefore, in general, will be different from the
thermodynamic pressure. Note that the thermodynamic pressure in a moving fluid is
different from the hydrostatic pressure. The mechanical pressure is a measure of the
translational energy of the molecules only. The thermodynamic pressure is a
measure of total energy, which includes translational mode of energy as well as
vibrational and rotational modes. Now the surface stress tensor ij may be rewrite in
the form    p   ,
ij ij ij

where ij depends on the motion of the fluid only and is called the shear-stress
tensor. Note that the first term of the right-hand side of this equation tends to change
the size of a fluid element and the second term tends to change its shape. Therefore
ij can be related to the internal (molecular) friction of the fluid.
Fluid Mechanics for Engineering Undergraduates - I 20

The general stress distribution and its separation into pressure and shear
components could be represented in matrix form too as

  11  12  13   p 0 0    11  12  13 
     
σ    21  22  23    0  p 0    21  22  23 .
  33   0 0  p   31  32  33 
 31  32 

Also note that the term ij comprised of not only shear stress components (i.e. six
components ij, i  j) but also direct stress components (i.e. three diagonal
components ij, i = j). The former components lead to change in angular shape of the
fluid element, whereas the latter components result lead to change in linear shape
(see Figure 2.3 for these two modes of deformations). This explanation is further
illustrated in Figure 2.6 for the corresponding two dimensional case, as most
students are on the (mis)understanding that the change in shape is due to shear stress
only and direct stress contributes to change in size only.

Consider a two dimensional fluid element subjected to a stress distribution of 11 =

- 6 kN/m2, 22 = - 2 kN/m2 and 12 = 21 = 2 kN/m2, as shown in Figure 2.7 (a).
Now, the total stress distribution could be rearranged as linear summation of three
fundamental modes (i.e. superposition theorem for linear systems) shown in Figure
2.7 (b), (c) and (d), representing dilatation (i.e. change in size), linear deformation
and angular deformation, respectively.

-6
-4 -2
2 2
-2 -2 -4 -4 2 2
 + +2 2

2 2
-4 -2
-6

(a) (b) (c) (d)

Figure 2.7: Representation of a stress distribution by its fundamental modes.

2.8 Conservation of Energy - The Energy Equation

An expression for the conservation of energy applied to a fluid in motion may be

developed from the first law of thermodynamics. Since a fluid flow necessarily
occurs under thermodynamic non-equilibrium conditions, some reformation is
needed. This is achieved by considering the instantaneous energy of the fluid to
consist two parts, intrinsic (or internal) energy and kinetic energy. Then the
modified form of the first law thermodynamics states that the rate of change of the
total energy (sum of intrinsic or internal and kinetic) of a fluid system (or element)
Fluid Mechanics for Engineering Undergraduates - I 21

as it flows is equal to the sum of the rate of work done on the fluid by external
forces (including gravity) and the rate of heat added to the fluid.

The total energy containing in a volume V is given by V (e + 1/2u.u)dV where e

represents the internal energy per unit mass and u is the velocity vector. The total
work done by surface stresses P is S u.PdS and by body forces is V u.fdV, where f
is the body force per unit mass. The net rate of heat leaving the fluid volume is
given by S q.dS, where q is the conductive heat flux leaving the control volume.
Then the conservation of energy may be written in analytical form as
 e  12 u.udV 
D

Dt V S
 
u.PdS  u.fdV  q.dS
V S

This expression includes both mechanical energy (i.e. work done by mechanical
forces) and thermal energy in the fluid. The terms which represent these two energy
forms may be distinguished by rearranging the above expression with the use of
exact expressions for surface forces (pressure p and viscous ) and body forces.
Then the momentum equation (which consists of the mechanical forces F) could be
used to reduce the energy equation to the equation of conservation of thermal
energy, which indicates that the rate of change of internal energy is equal to the rate
of conversion of mechanical energy into thermal energy due to surface stresses plus
the rate of heat transfer by conduction.

Note that for incompressible and inviscid fluid, there is no conversion of

mechanical energy to thermal energy. In this case the rate of work done by the
mechanical forces leads to the change in kinetic energy only and the rate of change
of internal energy is due to the rate of heat transfer by conduction only.

2.9 Constitutive Relations

The basic conservation laws represent five scalar differential equations (continuity
equation, momentum equations in three directions and energy equation). In order to
obtain a complete set of equations, the stresses and the heat-flux must be further
specified. These relations are provided by constitutive relations, in which the stress
is related to deformation (or strain) and the heat-flux is related to the temperature
gradients.

Newton's Law of Viscosity

This states that for Laminar flow of Newtonian fluids, the shear stress is
proportional to the rate of shear strain. For a simple shear flow where the velocity
distribution is given by u = u(y), v = w = 0, the Newton's Law of viscosity becomes
du
  
dy
where  is the coefficient of viscosity (or dynamic viscosity) of the fluid (see also
Section 1.3).
Fluid Mechanics for Engineering Undergraduates - I 22

Fourier's Law of Heat Conduction

This states that the heat-flux by conduction is proportional to the negative

temperature gradient, that is T
qi   k ,
xi
where k is the thermal conductivity of the fluid.

2.10 The Navier-Stokes Equation

Application of momentum equation for a fluid element leads to a differential

equation relating the acceleration of the fluid element to the forces (or stresses)
acting on the element, which include pressure, shear stress and the body force
(effect due to gravity). The Newton's law of viscosity can be used to replace the
shear stress component of this equation by velocity gradients, which yields the
Navier-Stokes equation of motion. This in vector notation
u  
 u.u  f  p    2 u   .u .
Du 1
  
Dt t  3 
where f is the body force per unit mass,  is the vector operator and 2=. is the
Laplace operator. Note that this is a vector equation and therefore represents three
scalar equations in the three directions of the coordinate axes.

For incompressible fluid, the continuity equation is .u = 0 and above expression
reduces to
Du
  f  p   2 u .
Dt
For inviscid flow
Du
  f  p .
Dt
This equation is known as Eular equation.

2.11 The Bernoulli's Equation

Along a streamline, the equation of motion for inviscid flow under steady
conditions becomes dV dZ dp
V  g  ,
dS dS dS
where V is the fluid velocity (which is tangent to the streamline) and S in the
distance along the streamline. This equation (which is one dimensional form of the
Eular Equation) may be integrated along a streamline for an incompressible flow to
obtain the Bernoulli's equation as
p V2
  Z  C, where C is a constant .
g 2 g
Fluid Mechanics for Engineering Undergraduates - I 23

Note that the terms in the Bernoulli's equation represent different form of
mechanical energy in the fluid. These terms represent pressure energy per unit
weight, kinetic energy per unit weight and potential energy per unit weight. The
Bernoulli's constant C therefore represents the total energy per unit weight and also
known as total head H of the fluid (as the unit of the quantity is m). The Bernoulli's
Equation is applicable only for inviscid and incompressible flow under steady state
condition. Inviscid and incompressible flow is referred to as ideal flow.

There are many applications of Bernoulli's Equation (together with one dimensional
momentum equation), which are of practical importance. These include forces on
pipe fittings, fluid machinery (e.g. pumps, fans, hydro turbines, wind turbines), fluid
dynamic forces on solid bodies (e.g. civil engineering structures, aircraft, ground
vehicles), etc.

Exercise 2.5

Figure shows the flow control arrangement of a circular air duct consisting of a
conical valve of 90 included angle. Diameter of the duct is 0.8 m. At a particular
valve setting the conical air jet discharging to the atmosphere is 50 mm thick at 1.0
m diameter. The corresponding upstream static pressure within the duct is 0.5
kN/m2 (gauge pressure). Neglecting all frictional effects and assuming that the jet
has a uniform velocity, determine

(i) the velocity of the jet at the discharge,

(ii) the total head of the air in the duct,
(iii) the axial force exerted on the cone.

Assume that the airflow is incompressible and that the air has a density of 1.2
kg/m3.
50 mm

Flow
Direction
0.8 m 45 1.0 m
diameter diameter
Fluid Mechanics for Engineering Undergraduates - I 24

Exercise 2.6

Figure shows a horizontal, straight nozzle of diameter 20 mm located at the end of a

pipe of diameter 40 mm. Water flows through the system at a rate of 2 l/s, and
discharges to the atmosphere. The jet of water from the nozzle strikes a curved vane
and is deflected through 180 in the horizontal plane, as indicated in the figure.
Stating your assumptions clearly calculate

(a) the force acting on the nozzle,

(b) the force acting on the vane for the each of the following cases:
(i) the vane is stationary
(ii) the vane is moved in the direction of the jet at a velocity equal to two-
third of the jet velocity.

Pipe Nozzle Vane

Jet

Exercise 2.7

Figure shows a pipe fitting containing a long radius elbow and a nozzle connected at
the end of a pipe. The system lies on a vertical plane. Water of density 1000 kg/m3
enters in a horizontal direction with an absolute pressure of 150 kN/m2 at Section-1
and discharges to the atmosphere at Section-3 flowing vertically downwards. The
Section-2 and Section-3 lie 0.3 m and 0.4 m below the axis of the pipe, respectively.
The diameter of the elbow between Section-1 and Section-2 is 75 mm and the outlet
diameter of the nozzle at Section-3 is 25 mm. The volume of the water in the pipe
fitting between Section-1 and Section-3 is 0.005 m3. Stating your assumptions
clearly, determine

(i) the pressure at Section-2

(ii) the flow rate through the pipe fitting
(iii) the magnitude and direction of the force on the pipe fitting due to the flow.
1

Flow

0.3 m
1

2 2
0.1 m
3 3
Fluid Mechanics for Engineering Undergraduates - I 25

Exercise 2.8

Figure shows a jet pump designed to extracts 3 l/s of water from an open tank T,
which is situated 2 m below the center line of the pump. The diameters of the outer
pipe BC and the inner pipe AB of the pump are 50 mm and 25 mm, respectively.
The pump discharges to atmosphere at section C.

Calculate the pressure at section B and the required velocity of water in the pipe
AB. It may be assumed that the both supply pipes are loss-free.

A B C

Exercise 2.9

The flow of air in a horizontal 2m2m square duct is obstructed by a streamlined

support of width 1m, as shown in the figure. The support extends the full 2m height
of the duct. The pressure and the velocity of the flow at Section A, well upstream of
the support are 10 mm Water Gauge and 5 m/s, respectively. Neglecting shear
forces, calculate
(i) the pressure at Section B
(ii) the force on the support
(iii) pressure at Section C, well downstream of the support where the velocity may
be taken as uniform
(iv) total rate of energy dissipation between Sections A and C.

A B C

Flow
1m 2m

A B C
Fluid Mechanics for Engineering Undergraduates - I 26

Exercise 2.10

Wind turbines make use of kinetic energy of the wind to generate electrical energy.
The rotor of a turbine converts wind energy to mechanical energy through
aerodynamic forces and drives the electric generator. A simple fluid dynamic theory
could be developed to predict the energy extracted by a wind turbine rotor without
going into details of the rotor design. Figure shows one such flow model based on
control volume technique, which is referred to as Rankine-Froude axial momentum
theory. This model is based on the idealized one dimensional, incompressible,
inviscid flow and the flow is assumed to be entirely axial with no rotational motion.
Governing parameters of the flow problem are given in the figure. Take density of
air as .

(i) Apply the conservation of linear momentum to derive an expression for the
axial trust FA acting on the rotor in terms of density of air , rotor area A and
wind velocities U0, U and U1 (refer Figure Q4(b) for the notations).

(ii) Apply the conservation of energy to derive an expression for the axial trust FA
acting on the rotor in terms of density of air , rotor area A and wind
velocities U0 and U1.

(iii) Use the results in (i) and (ii) above to derive an expression for the wind
velocity U through the rotor in terms of U0 and U1.

(iv) The energy conversion efficiency of the rotor is expressed by the power
coefficient of the rotor CP, which is defined as the power extracted by the
wind rotor divided by wind power available through the rotor area A. Derive
an expression for CP as a function of x = U1/U0. Show that the optimum value
of x is 1/3 and determine the corresponding maximum value of CP. Discuss
the limitations of this result in real applications.
Rotor U0 - Wind velocity
U0, U0, p0 U - Wind velocity through the
p0 rotor
U1 - Final wake velocity
p0 - Atmospheric pressure
pu pd U1, p0 pu - Pressure just upstream of the
rotor
 pd - Pressure just downstream of
A U the rotor
Control  - Rotor speed
Volum A - Rotor area
e
U0, p0
Control Surface
Fluid Mechanics for Engineering Undergraduates - I 27

Exercise 2.11

The figure shows a schematic diagram of a simplified version of ideal flow around a
ceiling fan at steady state. The area of the rotor is A and the velocity of the flow
through the rotor is V, which can be considered as axial and uniform. The final
velocity of the air jet generated by the fan is U0 and the area of the jet is A0. The
resultant force acting on the ceiling wall due to the flow is F0 and the density of air
is ρ.

(i) Apply the conservation of energy to show that the axial thrust acting on the
rotor is given by .

(ii) Select an appropriate control volume and apply the conservation of linear
momentum to derive an expression for the axial thrust FA in terms of ρ, A, V,
U0, and F0.

(iii) Use the results in (i) and (ii) above together with continuity equation to derive
an expression for the area ratio A0/A in terms of ρ, A ,U0 and F0. Show that in
the case of a fan operating in free space without a ceiling, the above area ratio
becomes ½.

Sketch the flow pattern corresponding to real flow around a ceiling fan to show how
it differs from the simplified flow pattern shown in the figure.
Fluid Mechanics for Engineering Undergraduates - I 28

Exercise 2.12

Figure shows a pipe junction containing one inlet pipe of diameter 100 mm and two
outlet pipes of diameters 75 mm and 50 mm, one inline with the inlet pipe and the
other perpendicular it. Water of density 1000 kg/m3 flows into the junction at
Section-(1) at a flow rate of 0.15 m3/s. The gauge pressures at the Section-(1) and
Section-(2) are 100 kN/m2 and 75 kN/m2, respectively. The energy losses in the
flow and the gravity effects can be neglected.

(a) Determine
(i) the flow rates leaving the junction at Section-(2) and Section-(3),
(ii) the pressure at Section-(3),
(iii) the magnitude and direction of the force acting on the junction due to the
flow.

(b) It is given that the orientation of the 75 mm diameter outlet pipe could vary
within the limits 45    135.
(i) Comment on the effects of the angle  on the results in part (a) above.
(ii) Determine the values of the angle  which make the resultant force a
maximum and a minimum.

Section-(3)

75 mm

Section-(1)
Section-(2)

100 mm
Flow =90 50 mm
Fluid Mechanics for Engineering Undergraduates - I 29

CHAPTER THREE

IDIAL FLUID FLOW

3.1 Introduction

The theoretical development of fluid dynamics started in eighteenth-century, which

is usually referred to as classical hydrodynamics through which the motion of fluids
is sought to specify by mathematical relations. In order to simplify these relations,
certain approximations had to be used. The main assumption incorporated in the
analysis was ‘ideal fluid’, in which fluid is considered to be inviscid and
incompressible.

Although, all real fluids are viscous and also possess some degree of
compressibility, there are many practical applications of flow around solid bodies
where the fluids (both liquids and gases) behave as almost ideal in certain region of
the flow. For example, flow region outside the boundary layer and the wake (outer
region) of a flow around a body at high Reynolds numbers has negligible viscous
effects. Further, if the velocity of the flow is not comparable with the sound
velocity, the compressible effects are negligible. Therefore, under these
circumstances, the behaviour of the outer flow region could be predicted by ideal
flow approximations. In particular, in the case of flow around streamlined bodies
(such as aerofoils under normal operating conditions), the boundary layers and the
wake region are very thin and almost the entire flow is ideal. Therefore, ideal flow
theories have their main applications related to aerofoil configurations.

For example, consider the flow around a moving vehicle illustrated in Figure 3.1.
Except of the thing viscous layer adjacent to the body surface (i.e. boundary layer)
and the viscous wake region behind the body (that comprised of eddies), the entire
outer region could be considered as inviscid as the effects of viscosity are
insignificant and pressure field dominates the behaviour. Further, as the velocity
involved is much less than the sound velocity, even air behaves as an
incompressible fluid.

Viscous Boundary Layer

Invisid Outer Flow Viscous Wake

Figure 3.1: Flow field around a ground vehicle.

Fluid Mechanics for Engineering Undergraduates - I 30

Note that, for ideal flow under steady state, the pressure distribution in the flow is
given by the Bernoulli’s Equation p V2
  z  Constant,
g 2g

along a streamline, with usual notations. Therefore, the basic approach in the
theoretical modelling of ideal fluid flow is to predict the velocity distribution (i.e.
fluid kinematics). Then the pressure distribution on solid surface and the resulting
forces can be derived by the use of Bernoulli’s Equation.

In this chapter, only steady two dimensional flow of ideal fluid and its application to
few simple configurations are considered. The governing parameters in such
situation are the two velocity components, which are functions of only the position
of the point defined by its coordinates. In Cartesian coordinate system the velocity
components are denoted by u and v in x-direction and y-direction, respectively, and
both are functions of x and y only.

Exercise 3.1 Explain why the outer flow region of typical flow configurations
associated with wind flow around buildings & structures and flow
around moving bodies such as sport balls and ground vehicles could
be considered as ideal flow.

3.2 Governing Equations

The equations which must be satisfied by ideal fluid flows are Euler’s Equation of
Motion and Continuity Equation, given in vector form as
Dq q 1
  q.q  f  p ,
Dt t 
q  0,

where q is the velocity vector. Euler’s equation can be rearranged to obtain a simple
kinematic equation involving only the velocity q.

Let the ideal fluid flow is subjected to a conservative body force with scalar
potential , i.e. f = . Taking the curl of Euler’s equation, together with the vector
identity q.q = (½q.q) - qq,
  q 
   q    q   0.
t
In the above expression, the term q represents vorticity . The second term of the
above expression can be simplified further [using the vector identity (q) =
(.)q – (q.) + (.)q – (.q) ] to obtain
  D
 q.     . q,
t Dt
Fluid Mechanics for Engineering Undergraduates - I 31

where the continuity equation .q = 0 is used. For two-dimensional flow the
vorticity vector  is perpendicular to the velocity q, and right hand side of the
equation becomes zero. Then
D
 0,
Dt
which is known as Vorticity Equation. That is, for an ideal fluid, the vorticity
remains constant provided that the body forces are conservative. Further, since
circulation  per unit area is the vorticity, circulation around any closed circuit that
moves with the fluid remains constant. This theorem is known as Kelvin’s
Circulation Theorem. Now, if the flow has zero vorticity or circulation (i.e.
irrotational flow) at some state, then the flow is always irrotational thereafter. In
particular, if a motion of ideal fluid is started from rest (for example by motion of a
body), then the flow is always irrotational since it is obviously irrotational at the
start. Thus  = curl q = 0.

Therefore the governing equations of motion becomes, in vector form

q  0 and   q  0,
or in Cartesian coordinates u v u v
  0 and   0.
x y y x
Exercise 3.2 Express the vector equations .q = 0 and q = 0 in cylindrical polar
coordinates.

The above two equations, together with appropriate boundary conditions, are
sufficient to determine the velocity field. Then Bernoulli’s equation (or the Euler’s
equation) can be used to predict the pressure field. The most common boundary
condition used in flow around solid bodies is that the fluid velocity normal to the
surface at any point on the body is equal to the velocity of the body in that direction.
That is on the solid surface
q.n  U.n,
where U is the velocity vector of the body and n is the unit vector normal to the
surface. In addition, if the fluid is of infinite extent, then the behaviour of the fluid
at infinity must be prescribed (for example at infinity, flow is uniform or at rest).

Exercise 3.3 Figure shows a two dimensional flow

around a cylindrical structural component U R
of radius R, in a wind of velocity U.
Identify appropriate conditions to be used
in the analysis of the flow field.

3.3 Velocity Potential

For irrotational flow, as q = 0, there exists a scalar function  such that q =  .
The function  is known as velocity potential (or potential function). That is
Fluid Mechanics for Engineering Undergraduates - I 32

velocity potential provides an alternative means of expressing the velocity

components in irrotational flow. For example:
  
In cartesian coordinates u , v  and w  .
x y z
 1  
In cylindrical polar coordinates u r  , u  and u z  .
r r  z
Note: Consider the line integral from a reference point A to a general point P,

P P

 q.ds   s ds   PA  P   A .
A A
where s is the distance along the line. Therefore the value of the integral
depends only on the position of P relative to A and independent of the path.
Hence velocity potential  is a property of irrotational flows. Therefore
irrotational flow is also known as potential flow.

Now, the above expression for  can be substituted in the continuity equation to
obtain    2  0.

Thus velocity potential satisfies Laplace’s equation and is consequently a harmonic

function. Therefore, the introduction of velocity potential reduces the equations of
motion of ideal fluids (continuity equation and Euler’s equation) to a well-known
simple partial deferential equation. However, in principle, there are infinite
solutions to Laplace’s equations, and the appropriate solution for a given flow is
determined by the boundary conditions. In particular the condition on solid
boundaries defined previously simplifies, with usual notations, to

 U.n
n

3.4 Stream Function

In the previous section, the velocity potential  was defined such that the condition
for irrotationality was automatically satisfied. Similarly, if the flow is two
dimensional, another scalar function may be defined such that the continuity
equation is automatically satisfied. That is:
 
In cartesian coordinates u  and v   .
y x

1  
In polar coordinates ur  and u   .
r  r
The function  is known as stream function, since the streamlines of the flow are in
fact the set of curves defined by  = constant. In order to prove this, consider
 
d  dx  dy   vdx  udy.
x y
Fluid Mechanics for Engineering Undergraduates - I 33

Then, on constant  , d = 0 and dx dy

 ,
u v
which is the differential equation of streamlines.

Note: Consider the integral along a line with normal vector n, which defines the
rate of flow of fluid cross the line per unit depth as
B B B

 q.nds   ui  vj. dyi  dxj   udy  vdx   BA   B  A .

A A A
That is, the change in the stream function between two points represents the
flow rate per unit depth across any line joining the points.

Now, the above expression for  can be substituted in the irrotational condition q
= 0, to obtain u v   2  2 
    2  2     2  0.
y x  y x 
Thus stream function satisfies two dimensional Laplace’s equation and is
consequently a harmonic function. Since there is no flow through solid boundaries,
they are streamlines and therefore the boundary condition becomes  = constant on
solid surfaces.

3.5 Flow Nets

For any two dimensional irrotational flow of ideal fluids, both velocity potential 
and stream function  satisfy Laplace’s equation and are harmonic functions.
Further, these two functions are related through velocity components by
   
 and   ,
x y y x
which are known as Cauchy-Riemann Conditions. Further, above two expressions
may be combined to obtain    
  .  0.
x x y y
That is, the equi-potential lines  = constant and streamlines  = constant intersect
each other orthogonally. Thus equi-potential lines and streamlines together form a
grid of quadrilaterals having 90 corners, which is known as a flow net and
represents the flow pattern. Note that flow nets represent actual pattern of flow only
if the flow is everywhere irrotational. However, there are many practical
applications where the flow is rotational in certain regions, especially due to flow
separation and formation of wakes. If the edge of the separation regions can be
defined, flow nets may still be used to indicate the flow outside these regions by
treating the edge of these regions as streamlines. For example, the flow field around
a moving vehicle considered in Section 3.1 is reproduced in Figure 3.2, showing the
flow net in the inviscid outer region. Solid lines represent streamlines and dash lines
represent equi-potential lines.
Fluid Mechanics for Engineering Undergraduates - I 34

Figure 3.2: Flow net in the outer inviscid flow region of a moving vehicle.

3.6 Flow Singularity Modelling

3.6.1 Introduction

Prediction of the characteristics of flow around solid bodies is one the most
important but very complex problem in fluid mechanics. There are many analytical,
numerical and experimental methods (both direct and indirect) developed to study
these problems. One such indirect method is flow singularity modelling, in which
the effects of solid boundary on the outer flow are modelled by imaginary types of
flow generators known as flow singularities. In particular, this indirect approach
could be used to analyse various problems in ideal flow. In this section, few
examples of simple flow patterns, including fundamental flow singularities, and also
combination of them which leads to solutions of practical interest, are investigated.

3.6.2 Basic Flow Patterns

(a) Uniform Rectilinear Flow

Consider a uniform flow with velocity U at an angle  to the x-axis. Then the
velocity components are given by u = Ucos and v = Usin.
Therefore the stream function is given by
   
  
d   dx  dy    vdx  udy

 x y 
 U   sin  dx  cos  dy   U cos  y  sin  x   C.

The integration constant C may take any convenient value. For example, let at the
origin (0,0),  = 0, then C = 0. Note that, in this case on the streamline y = tan x, 
= 0. Similarly, the expression for velocity potential  can be obtain as
  U cos x  sin y .
Fluid Mechanics for Engineering Undergraduates - I 35

(b) Sources and Sinks

A point is called a source if the motion of a fluid consists of symmetrical radial

outward flow in all directions from that point. A source of negative strength, that is
inward radial flow, is called a sink. The velocity of the flow at the source or sink
becomes infinite, and therefore becomes a singular point. Therefore these flow
generators are known as flow singularities. The strengths of these flow singularities
are defined as the total volume flow rates generated by them. Sources and sinks in
two dimensions are known as line sources and line sinks. The strength of a line
source (or sink) is the total volume flow rate per unit length. The velocity
components of the flow field of a line source are given by, in polar coordinates, ur =
Q/2r and u = 0. Then the stream function is given by
     Q 
   dr  d    ru r d     ,
 r    2 
where the constant of integration is taken as zero (i.e.  = 0 at  = 0). Similarly, it
can be shown that the velocity potential is given by
    Q r
    dr  d    u r dr     ln .
 r    2   C 
The corresponding expressions for line sink are given by replacing the strength by
“-Q”.

(c) Irrotational Vortex

Free (or irrotational) vortex is another flow singularity, in which the fluid particles
move in concentric circles where the tangential velocity is inversely proportional to
the radius. That is, ur = 0 and u = A/r, where A is a constant. It can be shown that
the angular velocity  (or vorticity ) of the flow is zero (i.e. irrotational), except at
the center of the vortex (which is a singular point). Further circulation  around any
closed curve, where the vortex is located outside the curve, is zero. However,
circulation around any curve inclosing the vortex becomes a constant, given by

  u dl  2A, indicating that the total strength (which is defined as the
circulation  around it) of the vortex is concentrated at its center.

As in the case of a line source, the stream function and the velocity potential of the
flow field generated by an irrotational line vortex are given by
   r   
     ln   and     , respective ly.
 2   C   2 

Exercise 3.4 Sketch the streamlines and equi-potential lines of the following flow
fields: (i) Rectilinear flow, (ii) Line source, (iii) Line sink and (iv)
Irrotational vortex.
Fluid Mechanics for Engineering Undergraduates - I 36

(d) Doublets

A combination of a source and a sink of equal strengths Q at a small distance S

apart, where in the limit Q is taken infinitely large and S infinitely small such that
the product QS remains finite, is called a doublet of strength  = QS. The axis of
the doublet is defined as the line taken in the sense from –Q to +Q. The expressions
for stream function and velocity potential of a doublet can be derived from the
superposition of the flow field generated by the source and the sink at the above
limits.

Exercise 3.5 Show that the stream function and velocity potential of a doublet of
strength  located at the origin of the coordinates with its axis along
positive x direction are given by
   sin     cos 
    and     .
 2  r  2  r
Also sketch the patterns of streamlines and equi-potential lines.

3.6.3 Combinations of Basic Flow Patterns

(a) Rectilinear Flow and Line Source

Consider a uniform flow of velocity U parallel to the x-axis, around a line source of
strength Q located at the origin of the coordinates. The resulting flow pattern can be
obtained by the superposition of flow fields generated by the rectilinear flow and the
source flow. For example, the stream function of the combined flow is given by
Q  y Q
  Uy    tan 1    Ur sin     ,
 2   x  2 

Source Flow
S O x

Figure 3.3: Combination of a uniform flow and a source flow

Fluid Mechanics for Engineering Undergraduates - I 37

The streamline pattern given by the above expression is shown in Figure 3.3. The
shaded area represents the region occupied by the source flow. At the stagnation
point S, the velocity of the flow becomes zero. The streamlines SA and SB divide
the source flow and the flow originated from the rectilinear flow, and therefore there
is no interaction between the two flows. This indicates that, if the shaded area is
occupied by a solid body, the pattern of streamlines outside the body is identical to
the above pattern. Therefore, the source could be considered as a hypothetical
device for modeling the flow around the body of above shape (which is known as
Rankine half body).

Exercise 3.6 For the flow field shown in Figure 3.1, determine the value of the
stream function on (i) +ve x-axis (ii) SA and (iii) SB. Also derive an
equation for the contour of the Rankine half body and estimate the
asymptotic width of the body.

(b) Source and Sink of equal strength in a Uniform Rectilinear Flow

Consider a uniform flow of velocity U parallel to the x-axis, around a source and a
sink of strength Q and –Q, a distance 2l apart, located on the x-axis symmetrically.
The resulting streamline pattern is shown in Figure 3.4. As in the case of rectilinear
flow around a source, the flow pattern outside the shaded area (which is occupied
by the flow from the source and the sink) could be considered as ideal flow around a
solid body of the same shape (which is known as Rankine Oval or Rankine Full
Body).

y
y

S1 Source/Sink Flow S2 x
+Q O -Q

Figure 3.4: Combination of a uniform flow, source flow and sink flow.

Exercise 3.7 Derive expressions for the stream function and velocity potential of
the above flow field and estimate the width and height of the Rankine
Oval.
Fluid Mechanics for Engineering Undergraduates - I 38

(c) Doublet in a Uniform Rectilinear Flow

The pattern of flow corresponding to this configuration could be deduced from the
above flow field by moving the source and the sink infinitely closer together but
maintaining the product of the strength and the distance a constant. Then the
Rankine Oval becomes a circle, representing two dimensional ideal flow around a
circular cylinder, which is one of the important flow configuration in practice. The
stream function and the velocity potential of the flow are given by

   sin    R 2 
  Ur sin      Ur sin  1    
 2  r   r  

   cos    R 2  
and   Ur cos      Ur cos  1    , where R2  .
 2  r   r   2U

The expression for the stream function shows that on r = R,  = 0 and therefore a
streamline, which represents the circle that separates the outer flow from the doublet
flow.

S1 S2 x
µ O

Figure 3.5: Combination of a uniform flow and a doublet

Exercise 3.8 Derive an expression for the velocity distribution on the above circle.
Use Bernoulli’s equation to predict the pressure distribution acting on
a circular cylinder in a uniform ideal flow and compare it with that
corresponds to real flow at relatively high Reynolds number.
Comment on the resultant force acting on the cylinder at ideal flow
conditions.
Fluid Mechanics for Engineering Undergraduates - I 39

Exercise 3.9 Figure shows an ideal flow configuration consisting of two sources of
strength Q each at the points (-h, 0) and (0, 0), a sink of strength –2Q
at the point (h, 0) and a uniform flow of velocity U. If Q = 2hU,
(i) write down an expression for the velocity distribution on the x-
axis,
(ii) sketch, without any calculation, the streamline pattern of the
flow field. y

U
+Q +Q -2Q x
h h

Exercise 3.10 Figure shows an ideal flow configuration consisting of a source of

strength 2Q at the origin, a sink of strength –Q at the point (h, 0)
and a uniform flow of velocity U. If Q = 2hU,
(i) write down an expression for the velocity distribution on the
x-axis,
(ii) determine the position(s) of the stagnation point(s), if any,
(iii) sketch the streamline pattern of the flow field.
y

U x
+2Q -Q
h

Exercise 3.11 The ideal flow configuration shown in the figure consists of equal
and opposite vortices of strength , a distance 2c apart, placed
symmetrically in a uniform flow of velocity U.
(i) Write down an expression for the velocity distribution on the
x-axis,
(ii) Determine the position(s) of the stagnation point(s), if any,
(iii) Sketch the streamline pattern of the flow field.
y

-
U c
x

c
+
Fluid Mechanics for Engineering Undergraduates - I 40

Exercise 3.12 A mathematical model that represents flow field of an eddy in a

river consists of a line vortex of strength  and a line sink of
strength –Q, both located at the center of the eddy. Sketch the
streamline pattern of the flow field.

Exercise 3.13 A circular cylindrical chimney of diameter 1.5 m is exposed to a

steady wind flow of uniform velocity 8.0 m/s. It is required to
model this configuration under ideal flow condition. Determine the
strength of the doublet needed to represent the flow. Also estimate
the location and magnitude of the maximum velocity in the flow
filed.

(d) Doublet and Irrotational Vortex in a Uniform Rectilinear Flow

An irrotational vortex may be added to the uniform rectilinear flow around the
doublet to obtain a useful flow configuration, which represents flow around a
rotating cylinder in a uniform flow. Note that, since the streamlines of the vortex are
concentric circles they do not intersect with the cylinder surface. Although the flow
around the cylinder without the vortex is symmetric, the addition of the vortex
increases the flow velocity on one side of the cylinder and reduces it on the other
side. This leads to non-symmetric pressure distribution on the cylinder and gives a
transverse (or lift) force.

In real fluid flow, the circulation around any rigid body could be generated by
rotating the body. The movement of the solid surface imparts a circulation on the
fluid through the viscous action (no slip condition). The circulation generated by the
body motion is proportional to its angular velocity. If there is a relative flow
velocity in addition to body’s rotation, the body is acted upon a lift force. This
phenomenon is known as Magnus Effect.

In the case of flow around aerofoil sections (which are non-rotating), the generation
and maintenance of circulation are due to different phenomenon, which will be
discussed in detailed later. In brief, the flow around the aerofoil adjust itself to
remove flow singularities (infinite velocities) thereby generating (and maintaining)
a circulation around the body, which leads to a lift force.

Exercise 3.14 A uniform flow of velocity U takes place outside a circular cylinder
of radius R with a circulation  around it.
(a) Derive an expression for the velocity distribution on the
cylinder surface and determine the position of the stagnation
point(s).
(b) Sketch the streamline patterns for the following cases: (i)  <
4RU, (i)  = 4RU and (iii)  > 4RU.
(c) Derive expressions for the lift and drag forces per unit
length of the cylinder.
Fluid Mechanics for Engineering Undergraduates - I 41

CHAPTER FOUR

VISCOUS FLOW
4.1 Introduction

Almost all branches of engineering somehow involved with flows of liquids and
gases through ducts, pipes, channels or various types of apparatus. These include
mechanical, chemical, civil, aeronautical, bio-chemical & pharmaceutical, nuclear,
petroleum, hydraulic, agricultural and space engineering plants, apparatus and
systems. In some applications most part of the fluid flow systems are consist of very
long pipes, ducts or channels, for examples oil pipelines, gas lines, water lines,
steam pipes, air ducts, rivers, etc. In other cases the ducting are relatively short but
there are fittings and branches in the form of valves, bends, junctions, barriers, etc.
Examples of such systems include air ducts of complex ventilation systems;
industrial drying systems; flue gas ducting of chemical & other factories, boilers &
furnaces, nuclear reactors. In some cases, fluid flows with free surfaces, such as
flow in open channels and flow of water in down pipes.

In fluid flow systems, the motion is usually created by a difference in pressure

and/or gravity effects. In addition, moving solid boundaries could also create a flow,
for example in lubrication of bearings. As the fluid flows through the system, there
are resistances to the flow, which results in energy loss in the fluid. The knowledge
of the resistance to flow is an essential requirement, primarily to determine the
pumping (or energy) requirements for any apparatus or eventually for the entire
plant involved in the fluid flow system. In fact, thorough knowledge of the subject
is required to design fluid flow systems for optimum operational performances.

There are two kinds of energy losses: frictional losses and local losses. The fluid
frictional losses are due to the viscosity (both laminar and turbulent) of real fluids in
motion. The resistance to flow in a straight pipe of uniform cross section is purely
due to this effect. For example, consider a steady flow through a pipe, as shown in
Figure 4.1.

A B

U Flow U

A B
Figure 4.1: Fluid Friction Losses
Fluid Mechanics for Engineering Undergraduates - I 42

Due to the viscous effects the total energy of fluid at Section B is less than that of
Section A. In this case, the kinetic energy and the elevational energy of the fluid
remain constant and therefore the loss of energy is through reduction in pressure
energy. That is, the fluid pressure at Section B is less than that of Section A (i.e. pA -
pB = p > 0). In fact, the flow through the pipe is created by this pressure difference.

The local losses are caused by number of factors such as local disturbances of the
flow, separation of flow from the walls and formation of large eddies or vortices at
places where the geometrical configurations of the flow system change, merging /
branching of flow or fluid streams meet obstructions. All of these factors contribute
to the exchange of momentum between the moving fluid particles (i.e. turbulent
friction), thus enhancing energy dissipation (see Figure 4.2). Therefore, there is a
loss of energy across the component and if the inlet and outlet areas are the same,
energy loss is represented by a loss in pressure across the component (i.e. pA - pB =
p > 0).
Eddy
Zone B
Eddy
A Zones
B

B
Eddy
Zone
A B
A A
(a) Flow through a Pipe Bend (b) Flow through a Pipe
Orifice
Figure 4.2: Local Losses in Pipe Fittings

Note that the molecular and turbulent viscosity of the moving fluid irreversibly
converts the mechanical energy (the work of the resistance forces) into heat.
Therefore, the term fluid resistance represents the irreversible loss of total energy of
the system. The main objective of this section is to discuss the basic features of
viscous flow through bounded systems. In particular, theoretical analyses of laminar
flow through simple bounded systems (such as pipes and channels) are presented.

4.2 Viscous Flow through Bounded Systems

4.2.1 Applications

Most of the fluid flow systems with engineering interests are turbulent where the
fluid particles move in disorderly manner with continuous and rapid mixing with
Fluid Mechanics for Engineering Undergraduates - I 43

adjacent fluid particles. However, there are very important practical applications
where the flow is laminar (or viscous). In this type of flow, fluid particles move in
layers (not necessarily in straight lines) without disturbing the fluid particles in
adjacent layers as the resistance due to molecular friction (i.e. influence of
viscosity) is predominant to damp out any disturbance which could cause mixing.
Consequently laminar flow is deterministic (or predictable) in nature.

For a given flow to be laminar, the viscous forces should have a predominant effect
(i.e. low Reynolds number). This occurs when the flow configuration is such that
the velocities are relatively low and/or the flow is effectively bounded and
controlled by solid surfaces (i.e. the flow area is relatively low). Some of the
practical applications of laminar flow with engineering interests are:

 Flow through bounded systems (such as pipes, pipe fittings, ducts, channels) at
relatively low flow rates where the fluid may be Newtonian (water, some oils
etc.) or Non-Newtonian (paints, liquid foods, biological fluids, molten plastics
etc.)
 Leakage flow
 Lubrication (flow in bearings)
 Flow in hydraulic dampers (dash-pot, the device used in shock-absorbers to
damp mechanical vibrations).
 Flow through packed bed of solid particles (ground water flow, flow through
filters, sand beds etc.)

4.2.2 Entry Length and Fully Developed Flow

Laminar Inviscid Laminar Turbulent
Boundary Center Boundary Boundary
Layer Flow Layer Layer

Fully Developed Fully Developed

Entrance Length Le Flow Region
Entrance Length Le Flow Region
(Laminar) (Turbulent)

(a) Laminar Fully Developed Flow (b) Turbulent Fully Developed Flow
Figure 4.3: Development of the flow in the entrance to a duct
Figure 4.3: Development of the flow in the entrance to a duct
When a fluid enters a pipe with uniform velocity, the fluid layers adjacent to the
wall are slowed down due to wall shear stresses and a boundary layer forms at the
entrance. In fact, the fluid particles immediately adjacent to the wall are at rest, as
they get trapped within the pockets of valleys created by roughness (at micro-level)
Fluid Mechanics for Engineering Undergraduates - I 44

of the solid surfaces (which is known as no slip condition). As the fluid passes in to
the pipe, more and more fluid particles get retarded and the thickness of the
boundary layer increases. At some distance downstream, boundary layers from
opposite walls meet at the center, after which the conditions of the flow remain
constant. This flow region is known as fully developed flow and the region before
the boundary layers meet is known as entry length. If the flow in the boundary
layers is laminar when they meet, then laminar flow exists in the entire region of
fully developed flow. Whereas if the transition has already taken place before they
meet, turbulent flow will persist in the region of fully developed flow. Figure 4.3
shows the development of the flow in the entrance to a conduit with smooth surface.

The entry length Le of the flow through a smooth-walled pipe is a function of the
Reynolds number Re of the flow. In the case of flow through a circular pipe
following expressions can be used to estimate Le: Laminar Flow - Le/d  0.06Re;
Turbulent Flow - Le/d  4.4Re1/6

Exercise 4.1 Water ( = 103 kg/m3,  = 0.8710-3 Ns/m2) flows through a 2.5cm
diameter circular pipe. Water enters to the pipe, without any
disturbance, at a uniform velocity. Determine the entry length, if the
flow rate is (i) 100 l/h, (ii) 750 l/h.

4.3 Laminar Flow through Simple Bounded Systems: Theoretical Analysis

4.3.1 Formulation of the Problem

In this section laminar flow of Newtonian fluid through simple bounded systems are
analyzed. The main objectives of the analysis are to predict the shear stress and
velocity distributions across the flow and to determine the flow rate.

Main Assumptions:
 Steady and incompressible flow  Newtonian fluid  Laminar flow
 Fully developed flow  Pressure across the flow is uniform

Governing Equations:
 Equation of motion applied to a fluid element
 Newton’s law of viscosity (constitutive relation)

Boundary Conditions:
 No slip at solid boundaries  No shear forces on free-surfaces

Method of Analysis:
The basic steps of the analysis may be summarized as follows:
 Select a suitable axis system (one axis along the flow and one axis across the
flow) and sketch a fluid element at a general point.
 Indicate the elementary forces/stresses acting on the fluid element.
Fluid Mechanics for Engineering Undergraduates - I 45

Note: There are three different types of forces:

1. Surface forces (pressure and shear) due to surrounding fluid
2. Forces due to solid boundaries in or adjacent to the element
3. Body forces (such as gravity, centrifugal)
 Apply the equation of motion in the direction of flow. In general, the resulting
equation may be integrated to obtain an expression for the shear stress
distribution.
 Use Newton’s law of viscosity to replace the shear stress by rate of shear
(velocity gradient). The integration of this equation together with appropriate
boundary condition(s) provides the required velocity distribution. Now the
rate of flow can be readily determined.

4.3.2 Flow in a Circular Pipe


p

p+dp
R
x r

x

dx
mg
Figure 4.4: Flow through a circular pipe

For steady flow, the equation of motion of the element becomes

   
p r 2   p  dp  r 2   2rdx  r 2 dxg sin   0.
After some algebraic this simplifies to
r  dp  r  dp dz  r  dp * 
    g sin      g    ,
2  dx  2  dx dx  2  dx 
where p* = p+gz is the piezometric pressure.

For the laminar flow of a Newtonian fluid the constitutive relation gives
du   dp * 
    r .
dr 2  dx 
This can be integrated to obtain an expression for the velocity distribution as
1  dp *  2
u  r  c .
4  dx 
Fluid Mechanics for Engineering Undergraduates - I 46

The integration constant c can be evaluated through the no-slip condition u = 0 on

the pipe wall r = R, and then the velocity distribution becomes
R 2  dp *    r  
2
u   1    .
4  dx    R  

Note: This is a parabolic distribution. The velocity is positive only if the

piezometric pressure gradient is negative (i.e. adverse pressure gradient).

Exercise 4.2 Sketch the shear stress and velocity distributions across the pipe and
show that the maximum velocity is two times the mean velocity of the
flow.

4.3.3 Flow between Parallel Plates

y Moving Plate
U
+d

dy p p+dp

y  x

x dx

Fixed Plate

Figure 4.5: Flow between parallel plates

Consider flow between two horizontal plates, a distance h apart, with one plate is
fixed and the other is moving with a velocity U, under a constant pressure gradient.
For steady flow, the equation of motion of the fluid element reduces to d /dy =
dp/dx. After integration this becomes  = (dp/dx)y + c1, where c1 is a constant. For
the laminar flow of Newtonian fluid  = (du/dy), and therefore the velocity
distribution is given by 1  dp  2 1
u   y  c1 y  c2 .
2  dx  
Constants c1 and c2 can be determined by the boundary conditions on the plates: u =
0 on y = 0 & u = U on y = h. Then the shear stress and velocity distributions are
given by
 U  1  dp   y  1  dp 
        (h  2 y) and u  U       y (h  y )
 h  2  dx   h  2  dx 
Fluid Mechanics for Engineering Undergraduates - I 47

Exercise 4.3 Derive an expression for the flow rate per unit width and sketch the
shear stress and velocity distributions across the flow for the
following cases:
(i) dp/dx = 2U/h2, (ii) dp/dx = 6U/h2.

4.3.4 Flow over a Plane (with a Free Surface)

y
Free Surface

y

x
dx x
mg 

Figure 4.6: Flow over a plane

For steady flow, the equation of motion of the fluid element becomes
-dx + (h-y)dxgsin = 0.
Therefore the shear stress distribution is given by
 = (h-y)gsin.
For lamina flow of Newtonian fluid  =  (du/dy), and substitution of this in the
above equation and integrating with respect to y yields
u = (gsin/2)(2h-y)y + c.
On the plane y=0, u=0 and therefore c=0.

Hence the velocity distribution becomes u = (gsin/2)(2h-y)y.

Exercise 4.4 For the free surface flow over a plane, sketch the shear stress and
velocity distributions across the flow and show that the maximum
velocity is 1.5 times the mean velocity of the flow.

4.4 Energy Loss in Pipe Flow: Darcy Formula

Figure 4.7 shows a fully developed flow of a steady and incompressible flow
through a straight conduit with uniform cross-section. Consider a control volume
enclosing the flow inside the pipe. Under the above conditions, the rate of change of
Fluid Mechanics for Engineering Undergraduates - I 48

momentum is zero and also there is no change in total kinetic energy of the fluid as
it flows through the pipe.

Section (1) L Section (2) Control

Volume
w
Weight mg = ALg
Shear Force = PLw
p1 where A - Flow Area
p2 P - Wetted Perimeter
 w - Wall Shear Stress
w
z1 z2 z1 - z2 = L sin
mg

Datum Level

Figure 4.7: Control volume analysis of steady fully developed flow in a straight
duct

Then the energy loss per unit weight of the fluid is given by
p1  p2
h f  H1  H 2   z1  z2 ,
g
and application of momentum equation gives (p1 - p2)A + gALsin - wPL = 0, or
p1  p2  w PL
 z1  z2   .
g gA
Now, the skin-friction coefficient or friction factor is defined by f = w/(1/2V2),
where V is the average flow velocity. For a circular pipe of diameter d, A = d2/4
and P = d. Then, by combining the energy equation and momentum equation, the
head loss due to friction may be written as
L V 2  L V 2  V 2 
h f  4 f        K f  
d  2g  d  2g   2g 
This is known as Darcy-Weisbach equation. The parameter f is known as Fanning or
Darcy friction factor,  is known as Moody friction factor and Kf is the coefficient
of fluid resistance. For non circular cross-sections, diameter d could be replaced by
the equivalent diameter defined by de = 4A/P.

For laminar flow the friction factor f is a function of Reynolds number only. In
turbulent flow f is a function of both Reynolds number and the relative roughness of
the pipe wall. The variation of f with Reynolds number and relative roughness is
usually represented graphically by "Moody chart", shown in Figure 4.8. Note that in
some literature the notation f is used for the parameter , which is 4f in the present
notations.
Fluid Mechanics for Engineering Undergraduates - I 49

f


Re
Figure 4.8: Moody chart for pipe friction.

There are four distinct regions in the chart. First is the laminar flow region (Re<
2000), in which f = 16/Re. The second region is known as the critical zone (2000 <
Re < 4000), which represents the transition of the flow conditions from laminar to
turbulent. Frictional losses cannot be predicted in this region. The third region
corresponds to turbulent motion of the fluid where f is function of both Re and ,
and referred to as the transition zone (i.e. transition zone between viscous region
and fully rough pipes). For turbulent flow through smooth pipes f =0.079Re-0.25, if
4000 < Re < 105. Other cases the variation of f represents some complicated
behaviour. The fourth region, known as fully turbulent region, represents flow
through rough pipes at sufficiently high Reynolds numbers, where f becomes
function of  only.

Exercise 4.5 Use the results derived in Section 4.3 to show that the friction factor
for laminar flow through circular pipe is given by f = 16/Re.

The coefficient of fluid resistance Kf due to pipe friction is linearly proportional to

the pipe length and inversely proportional to its diameter. In fluid flow systems with
lengthy pipes, the main contribution of the energy loss is due to the pipe resistance
and therefore known as major loss. The local losses due to fittings and equipment
are relatively small and therefore known as minor losses. The energy loss
coefficient due to local losses, denoted by K, usually taken as constants, except for
valves for which it depends on the amount of opening. However, these values are in
general functions of the Reynolds number of the flow, size of the fitting and the
Fluid Mechanics for Engineering Undergraduates - I 50

relative roughness of the surface. Table 4.1 gives K values of pipe fittings
commonly encountered in fluid flow systems.

Table 4.1: Energy loss coefficients of pipe fittings

Type of Fitting K Type of Fitting K
90 Standard Elbow 0.9 Gate Valve Fully open 0.17
45 Standard Elbow 0.45 3/4 open 0.9
90 Smooth Bend 0.3 1/2 open 4.5
180 Bend 1.5 1/4 open 24.0
Standard T (side outlet) 1.8 Globe Valve Fully open 9.0
Coupling/Union 0.04 3/4 open 13.0
Pipe Entrance (sharp) 0.5 1/2 open 36.0
Pipe Exit 1.0 1/4 open 112.0

Exercise 4.6

Water flows at a mean velocity of 0.5 m/s through a pipe system consists of 2.5 cm
diameter circular cast iron pipes of total length 20 m connected through three 90
Standard Elbows and a globe valve (1/2 open). Calculate the total rate of energy
loss in the system. You may take  = 103 kg/m3,  = 0.8910-3 Ns/m2 for water.

Exercise 4.7

A viscous liquid of viscosity  and density  flows down a vertical cylindrical shaft
of radius a. The film thickness is uniform and constant with outer radius ka, where
the constant k>1.
(i) Derive expressions for the shear stress and velocity distribution across the
film.
(ii) Show that, when k = 2 the maximum and mean velocity of the flow are umax 
0.636ga2/ and umean  0.473ga2/, respectively.

Exercise 4.8

The vertical shaft of radius a shown in the figure rotates at an angular velocity  in
a bearing of length H. The thrust at the lower end of the shaft is transferred by a flat
disc of radius b to a flat housing. The space between the shaft and bearing and
between disc and housing is filled with oil of viscosity . In each case the film
thickness is h (« a, H). H

Assuming that the velocity

gradient is constant across the
thickness of each oil film, show Oil Film 2a
that the power absorbed in fluid Oil Film
friction is 2(b4 + 4Ha3)/2h.
2b
Fluid Mechanics for Engineering Undergraduates - I 51

Exercise 4.9

(a) Show that for fully developed laminar flow of a fluid of viscosity  between
horizontal parallel plates a distance h apart, the pressure gradient dp/dx is
related to the mean velocity um by the expression dp/dx = -12um/h2.

(b) The figure shows a flange pipe joint in a di

h
pipe of internal diameter di containing a
viscous fluid of viscosity  at gauge
pressure po. The flange has an outer po
diameter do and is imperfectly tightened so
that there is a narrow gap of thickness h.
Show that the volumetric flow of the leak do
is h3po/[6loge(do/di)].

Exercise 4.10

Water is discharged from a reservoir through a pipeline of diameter 5 cm, effective

surface roughness 0.5 mm and length 85 m into a second reservoir. The difference
of the water levels in the two reservoirs is 10 m. The pipeline consists of a gate
valve (1/2 Open), a 90 Standard Elbow and a 90 Smooth Bend. Estimate the
velocity of the flow through the pipe. You may take for water  = 103 kg/m3,  =
0.8910-3 Ns/m2.

Exercise 4.11

Water supply system consists of an overhead tank, a distribution tank at the ground
level and a hydraulically smooth pipeline of diameter 5 cm connecting the two
tanks. The difference of the water levels in the two tanks is 20 m and the total length
of the pipeline is 90 m. The pipeline consists of a Gate Valve, four 90 Standard
Elbows, one 180 Bend and fifteen Couplings. At a particular opening of the gate
valve, the flow rate is found to be 5 l/s. What is the opening position of the gate
valve? Also estimate the maximum flow rate possible through the pipeline. Use the
information given in Appendix to estimate the energy losses. You may take  = 103
kg/m3,  = 0.910-3 Ns/m2, for water.

Exercise 4.12

Figure shows a simplified version of a fluid coupling for

power transmission between input and output shafts,
comprising of two flat discs of diameter d, separated by a
distance h (<< d) apart and filled with an oil of viscosity . 1 2
Rotational speeds of the input and output shafts are 1 and
2, respectively. Derive an expression for the torque
available at the output shaft in terms of given parameters.
h
Fluid Mechanics for Engineering Undergraduates - I 52

Exercise 4.13

(a) A viscous liquid of viscosity  and density  flows between two horizontal
parallel plates, a distance h apart, one plate is stationary and other moves with
a uniform velocity U. Derive expressions for shear stress and velocity
distributions of the flow for the case of zero pressure gradient. State your
assumptions clearly.

(b) The figure shows a rotary viscometer comprised of two concentric cylinders
with the gap between them filled with a Newtonian fluid. The inner and outer
cylinders have radii a and b respectively and the length of the inner cylinder
immersed in the fluid is h. The thickness of the fluid layer under the
immersed end face of the inner cylinder is t. The outer cylinder is rotated at a
constant angular velocity  and the inner cylinder is suspended on an elastic
wire. The torque T acting on the inner cylinder due to viscous action causes
the wire to twist until a steady state is reached. Assuming that the thicknesses
of the fluid layers are much smaller than the other dimensions, derive an
expression for the viscosity of the fluid in terms of the parameters given
above.

Wire

Fluid

h
b


Fluid Mechanics for Engineering Undergraduates - I 53

4.5 Laminar Flow of Non-Newtonian Fluids through Bounded Systems

4.5.1 Introduction

In laminar flow of Newtonian fluids, the shear stress is linearly proportional to the
rate of shear strain, i.e.  = du/dy, where the constant of proportionality  is a
property of the fluid referred to as dynamic viscosity of the fluid. Examples of
Newtonian fluids include pure gases, mixture of gases, pure liquids of low-
molecular-weight (i.e. non-polymeric), etc.

However, a fairly large category of fluids in industrial applications exhibit more

complex behaviour and the viscosity is a function of rate of shear strain. These
liquids are referred to as Non-Newtonian. Examples of Newtonian fluids include
biological fluids (blood, milk, saliva, etc.), liquid foods, some medicines, paints,
many emulsions, greases, concrete mixes, mud, polymers (molten or solution), tooth
paste, etc.

4.5.2 Classification of Non-Newtonian Fluids

Non-Newtonian fluids can be classified by their time-independent behaviour or

time-dependent behaviour. In addition, they can be classified by their visco-
elasticity.

(a) Time-Independent Behaviour

Figure 4.9 shows the basic shear diagram for several classes of fluids under steady
flow conditions.

Generalized
Shear Plastics
Stress Shear-Thinning
(Pseudo Plastic) Newtonian
Bingham
y Plastics Fluids

Shear-Thickening
(Dilatent)

Rate of Shear Strain

Figure 4.9: Shear diagram of Non-Newtonian fluids

Fluid Mechanics for Engineering Undergraduates - I 54

 Pseudo-Plastic Fluids

Pseudo-plastics are those in which the viscosity decreases with rate of shear, but the
material deforms as soon as a shearing stress is applied. Examples of this type of
fluids include gelatin, emulsions, most polymers (molten or solutions), liquid
cement, milk, blood.

 Dilatent Fluids

Dilatents show a viscosity increase with increasing rate of shear. Material deforms
as soon as a shearing stress is applied. These fluids are less commonly encountered
than pseudo-plastics. Some examples are quick sand, wet bench sand, concentrated
solution of sugar in water.

Both pseudoplastic and dilatents follow power-law behaviour and therefore known
as power-law fluids. The constitutive relation can be given in the form
n
du
  K   K
n
,
dy
where n - flow behaviour index and K - fluid consistency index. In the case of
theoretical investigations, it is more convenient to express the rate of shear in terms
of shear stress as
du
    ,
m

dy
where m = 1/n - flow exponent and  = (1/K)m - fluidity.

 Bingham Plastic Fluids

For Bingham plastic fluids the shear stress - rate of shear strain relationship is linear
but does not go through the origin. That is, the fluid does not flow until the shear
stress exceeds the yield stress value y. This relation can be expressed
mathematically as
du
   y   p   y  p when    y
dy
du
   0 when    y
dy
where p is the plastic viscosity.

Note that, in practice this expression represents an ideal material. Experimental

observations indicate that the plastic viscosity p is not constant. Hence the curve
denoted as generalized plastic is more representative of actual behaviour. Examples
of fluids with yield stress include tooth paste, butter, potter's clay, paints, ketchup,
drilling mud, printing ink.
Fluid Mechanics for Engineering Undergraduates - I 55

(b) Time-Dependent Behaviour

Viscosity can also show complex time-dependent behaviour. The most important of
which is thixotropy. A fluid is said to be thixotropic, if the viscosity decreases with
the time duration of the shear stress but slowly recover when the shear is removed.
The opposite behaviour is termed rheopectic.

For example, printing ink which has been resting unmoved for a while will not pour
easily, but a recently shaken bottle of printing ink pours easily.

(c) Viscoelastic Behaviour

Some non-Newtonian fluids exhibit both viscous behaviour of fluids and elastic
behaviour of solids. These materials are referred to as viscoelastic. Some examples
are: saliva, flour dough, molten polymers, blood, polymeric solutions.

4.5.3 Flow through a Circular Pipe

As derived in Section 3.3, the equation of motion of the flow, which describe the
variation of shear stress with radial position, is given by
r  dp * 
    , where 0  r  R and p * is the piezometri c pressure.
2  dx 
Now, the velocity distribution of the flow can be determined by using the
appropriate constitutive relation for a given non-Newtonian fluid. As given in the
previous section the exact form of the - relation depends on the sign of .
Therefore it is essential to identify the sign of the shear stress  before solving the
governing equation. In the present application, in order to have a positive flow, the
pressure gradient should be negative and therefore, according to the above equation,
  0.

(a) Power-Law Fluids

      .
du
The constitutive relation is given by, for   0,
m

dr
m
Then substitution for  yields
du  1 dp *  m
     r .
dr  2 dx 
This expression can be integrated, together with the boundary condition u=0 at r=R,
  1 dp *  m1 m1
 
m
to obtain
u     R r .
m  1  2 dx 
The mean velocity of the flow is then given by

u mean 
 udA 
  1 dp * 
  R
m
m 1  m 1 
  u max .
A m  3  2 dx   m 3
Fluid Mechanics for Engineering Undergraduates - I 56

The velocity distribution can be rewritten in the form

m3  r  
m 1
u
 1    .
umean m  1   R  

This velocity distribution for three different types of fluids, pseudo-plastic (n=1/m <
1), Newtonian (n = 1) and dilatent (n > 1) are presented in Figure 4.10. The
Newtonian fluid shows the parabolic shape, pseudo-plastic fluid shows flat-nose-
like behaviour and dilatent shows sharp-nose-like behaviour.

n
du
  K
n=1 dr
Flow u

n>1
n<1

Figure 4.10: Velocity profiles of flow through a circular pipe for different
types of fluids

(b) Bingham Plastic Fluids

The constitutive relation for Bingham plastic fluids is given by, for   0,
1  r  dp *  
du

1
p

  y    
 p  2  dx 
  y when    y
dr 
du
 0 when    y
dr
Let at r = r0, || = y. Then r0 = 2y/(-dp*/dx) and the velocity gradient can be
represented by
du y  r  du
 1   when r  r0 and  0 when r  r0
dr  p  r0  dr
When r < r0 , u = constant = u0 (say), which is the velocity of the core. When r > r0 ,
the velocity distribution is given by the integration of the above expression, together
with the boundary condition u = 0 at r = R as
y  
u  R  r  R  r  1 when r  r0 .
p  2r0 
Now, an expression for the velocity of the core u0 can be obtained by substituting r
= r0 as  y R  r0 2  dp * 
u0  , where r0  2 y   .
 p 2r0  dx 
Fluid Mechanics for Engineering Undergraduates - I 57

The velocity profile of the flow is presented graphically in Figure 4.11.

u0
Flow r0 u

Figure 4.11: Velocity profiles of flow through a circular pipe for a Bingham
plastic fluid

Exercise 4.13

Show that, for the flow of Bingham plastic fluid through circular pipe, the mean
velocity is given by u 3  2   2 y
mean
 ,  ,
u0 6 0
where u0 is the maximum velocity, y is the yield stress and 0 is the wall shear
stress.

Exercise 4.14

A Bingham plastic fluid flows between horizontal surfaces separated by a distance

2h under the influence of a pressure gradient dp/dl. The fluid has a yield stress y
and a viscosity . Derive an expression for

(i) the velocity distribution, and

(ii) the volumetric flow rate per unit width.

Exercise 4.15

A Bingham plastic fluid of viscosity 0.05 Ns/m2 and a yield shear stress 0.6 N/m2
flows along a 15 mm diameter tube of length 3.0 m. Evaluate the minimum pressure
drop pmin for flow to occur. If the pressure drop is 2pmin, determine

(i) the radius of the solid core

(ii) the velocity of the solid core
(iii) the volumetric flow rate.

Sketch the velocity profile.

Fluid Mechanics for Engineering Undergraduates - I 58

Exercise 4.16

A cylindrical viscometer comprises two concentric cylinders with the annular space
between the cylinders filled with a non-Newtonian fluid. The outer cylinder is
rotated at a steady angular velocity  and the inner cylinder is suspended on a wire.
Viscous action on the inner cylinder causes the wire to twist until a steady torque T
is developed to restrain the cylinder from further rotation. The fluid viscosity is
known to be on the form  = k(d/dt)n, in which  is the fluid shear stress,  is the
shear strain and k is a constant. The inner and outer cylinders have radii R1 and R2
respectively and the length of the inner cylinder immersed in the fluid is h.

Show that, under steady conditions and ignoring viscous action on the immersed
end face of the inner cylinder, n T   1 1 
1/ n
     2 / n  2 / n .
2  2hk   R1 R2 
Describe how this relation may be used to obtain values for k and n for the fluid.
d du d
It may be assumed without proof that    r ,
dt dr dr
where  is the angular velocity of the fluid at radius r.

Exercise 4.17

A non-Newtonian fluid having the constitutive equation  = k   , where  is

n
(a)
the shear stress,  is the rate of shear strain and k and n are constants, flows
between horizontal parallel plates a distance h apart, under the influence of an
adverse pressure gradient dp/dl. 1 n 1
 n  1  dp  n  h  n
um          .
 2n  1  k  dl   2 
(b) The figure shows a flange pipe joint in a pipe of internal radius a containing
the non-Newtonian fluid given in (a) above, at gauge pressure p0. The flange
has an outer radius b and is imperfectly tightened so that there is a narrow gap
of uniform thickness h. Derive an expression for the volumetric flow of the
leak in terms of h, a, b, p0, k and n.

2a
h

p = p0 p=0

2b
Fluid Mechanics for Engineering Undergraduates - I 59

Exercise 4.18

A Bingham plastic fluid of density  = 790 kg/m3, viscosity p = 0.09 Ns/m2 and
yield stress y = 20 N/m2 flows over an inclined flat boundary with a free surface,
under the influence of gravity. Depth of the flow is 10 mm. Determine the minimum
angle of inclination of the boundary (to the horizontal) for the flow to occur. If the
angle of inclination is two times that of the minimum value calculated above,

(i) derive an expression for the velocity distribution across the flow and sketch
the variation indicating salient features/values,

(ii) determine the mean velocity.

Exercise 4.19

A non-Newtonian fluid has the constitutive equation for shear stress  = k(du/dy)n,
where du/dy is the velocity gradient and k and n are constants. It flows through a
circular pipe of radius R under the influence of a pressure gradient (-dp/dl).

(i) Derive the following expression for the flow rate Q.

(ii) Following table gives experimental test results undertaken using a pipe of
radius 10 mm and length 200 mm. Determine the values of the constants k
and n for the fluid.

Pressure Drop p (N/m2) 100 200 400 800

Flow Rate Q (10-6 m3/s) 2.71 6.15 12.0 27.3

Exercise 4.20

Figure shows two stationary horizontal flat circular discs of outer diameter D, which
are placed at a uniform distance h (<< D) apart. The lower disc is connected to a
circular pipe of diameter d through which a non-Newtonian fluid having the
constitutive equation  = k   , where  is the shear stress,  is the rate of shear
n

strain and k and n are constants, is pumped into the space between the discs. The
fluid flows radially outwards to the surroundings at atmospheric pressure. If the
mean velocity of the flow through the pipe is U, determine the pressure in the
supply pipe.
D

d
Fluid Mechanics for Engineering Undergraduates - I 60

4.6 External Flow: Flow around Solid Bodies

4.6.1 Introduction

Another important application of fluid dynamics where viscous effects play a

leading role in determining the nature of the flow field is external flow around to
solid bodies. There are many such problems of great practical significance in
several engineering fields. Some examples are: motion of aircraft, ground vehicles,
submarines, flow around particulate solids, wind flow around buildings and
structures, wind and wave energy extraction devices, etc. It is particularly important
for engineers to have a qualitative and quantitative understanding of the force
exerted on the body by the fluid. This requires accurate modelling and prediction of
the flow field around the body. However, the complexity of the flow structure
around solid bodies makes such a modelling/prediction a very difficult task. For
example, even the fundamental case of two dimensional, uniform, steady flow
around a fixed cylindrical body is considered, the flow field is characterized by the
formation of a turbulent wake due to boundary layer separation (especially at high
Reynolds numbers). In a more general case the problem is further complicated by
the shape of the body, three dimensional effects, response of the body due to its
flexibility (oscillations), multi-body configurations, etc.

Contributions to the resultant force acting on a body arise from the normal stress (or
pressure) and tangential stress (or shear) distributions on the surface. However, at
high Reynolds numbers the contribution of the pressure on the resultant force
dominates and the contribution of the shear force (due to viscous effects) could be
neglected. Therefore, determination of fluid dynamic force usually requires
prediction of the pressure distribution on the body. However the nature of the flow
around the body (such as boundary layer separation) is primarily determined by the
viscosity of the fluid and therefore cannot be neglected.

There are many analytical and computational models developed to analyze fluid
flow around solid bodies. However all these models have some degree of
empiricism, as there is no theory which could describe the behaviour of turbulent
flow completely. The main objective of this section is not to discuss the details of
these flow models but to present the fundamental features of flow around solid
bodies and their practical significance.

4.6.2 Flow Regions around a Body

When a real (viscous) fluid flow takes place around a body, large shear stresses are
setup near the body surface and fluid particles immediately adjacent to the solid
surface are at rest. Therefore there is a region of flow next to the body in which the
fluid velocity increases from zero at the surface to the full value of the main stream.
This region of flow is known as the boundary layer. At high Reynolds numbers this
region of flow represents a very thin fluid layer. On the downstream side of the
body, the boundary layer no longer follows the surface, but separates and rolls up to
Fluid Mechanics for Engineering Undergraduates - I 61

form a viscous, unsteady flow region with large-scale eddies (or vortices) behind
the body. This region of flow is referred to as the wake. Therefore the flow field
could be treated as consisting of three distinct regions (see Figure 4.12):

(a) The boundary layer immediately adjacent to the body where the effect of
viscosity is important
(b) The wake behind the body
(c) The region outside the boundary layer and the wake in which the flow may be
treated as inviscid (frictionless). This region is known as the inviscid outer
flow and the ideal fluid flow theories (e.g. Bernoulli's equation) are applicable
in this region.

Boundary Layer Inviscid Outer Flow

Free
Stream
S1
Far
O Wake
Near
Wake
S2

Boundary
Layer
Inviscid Outer Flow
O - Stagnation Point
S1, S2 - Separation Points

Figure 4.12: Flow regions around a body

Note:
(i) The wake consists of two regions: near wake and far wake. Near wake, a
region immediately behind the body, contains a relatively stagnant fluid. In
contrast, the flow in the far wake is highly turbulent and therefore results in
high energy losses. Consequently the pressure in the near wake is
considerably low and the surface of the body exposed to the wake is
subjected to a uniform "wake under-pressure".
(ii) Flow in the boundary layer and the wake is viscous and therefore
Bernoulli's equation is not applicable.
(iii) At high Reynolds numbers the boundary layer thickness is very small
compared with the dimensions of the body. Consequently the pressure at the
outer edge of the boundary layer (which could be determined by Bernoulli's
equation) is transmitted without much change to the body surface. Thus the
pressure distribution on the upstream surface S1OS2 is determined by the
nature of the inviscid outer flow. The pressure distribution on the surface
exposed to the wake may be estimated through the experimental
observation that the pressure in this region is uniform and therefore equal to
the pressure at the point S1 and S2.
Fluid Mechanics for Engineering Undergraduates - I 62

S1
Positive O
Flow pressure Negative
pressure
S2

Figure 4.13: Pressure distribution on a body

4.6.3 Types of Forces: Drag and Lift

Both the pressure and shear stress distributions on the surface of a solid body in a
fluid flow contribute to the resulting force acting on the body. The component of the
resultant force in the direction of the relative velocity of the fluid with respect to the
body is known as the Drag Force FD (or Drag D). The force component
perpendicular to the relative velocity is called the Lift Force L (or Lift L).

Lift L

Relative
Velocity
Drag D

Figure 4.14: Drag and lift

The contributions on the drag force from the pressure distribution and shear stress
distribution on the body surface are referred to as the Pressure Drag (Dp) and the
skin friction Drag (Df), respectively. Usually, the lift is entirely due to the pressure
distribution and the contribution from shear stress distribution is negligible.

4.6.4 Types of Body Shapes: Bluff Bodies and Streamlined Bodies

A body with extensive boundary layer separation and followed by formation of a

wake with large-scale eddies (or vortices) is referred to as a Bluff Body. For
example circular cylinder, rectangular body, sphere are some of the bluff body
shapes commonly encountered in practice.
Fluid Mechanics for Engineering Undergraduates - I 63

A body is termed a streamlined body if the shape of the body is such that its surface
elements are almost parallel to the oncoming flow and hence the flow separation
occurs well towards the rear part of the body which results in formation of a very
small wake region. Aerofoils fall under this category.

(a) Bluff Body (b) Streamlined Body

Figure 4.15: Bluff body and streamlined body

In the case of a bluff body, the surface area exposed to the wake is large.
Consequently, the drag is relatively large and is mainly due to the pressure drag (see
Figure 4.2 for pressure distribution of a typical bluff body). The drag of a
streamlined body is usually very small and is mainly due to the skin friction drag.
However, the pressure distributions on the upper and lower surfaces are different, as
shown in Figure 4.16, and result in a high lift force.

-ve pressure

Free Stream
(zero pressure)

+ve pressure

Figure 4.16: Pressure distribution on a streamlined body

4.6.5 Pressure and Force Coefficients

In the analysis of dynamic parameters such as pressure distribution and forces

associated with flow around solid bodies it is useful to express these quantities in
non-dimensional form so that the results represent more general configurations.
Fluid Mechanics for Engineering Undergraduates - I 64

 Pressure Coefficient (Cp)

Consider a steady flow around a solid body. Let the velocity and pressure of the
fluid far upstream of the body are p0 and U0, respectively. The pressure field of the
outer inviscid flow around the body is governed by the Bernoulli's theorem, given
by p V2 p0 U 02
   , neglecting variation in z.
g 2 g g 2 g
Thus the change in pressure is given by p  p  p0 
1
2
 
 U 02  V 2 .

This indicates that the pressure variation in the flow around the body is essentially a
consequence of the velocity variation. Note that the term 1/2V2 has the dimension
of pressure and therefore known as dynamic pressure. Now the dynamic pressure of
the oncoming free-stream could be used to non-dimensionalized the pressure change
as  V2 
p  p0
Cp   1  2 ,
1  U0 
U 02
2
where Cp is known as pressure coefficient, defined by the ratio between the change
in pressure and dynamic pressure of the free-stream.

 Force Coefficients: Drag Coefficient (CD) and Lift Coefficient (CL)

The drag force acting on a bluff body is mainly due to the pressure distribution on
the surface. Therefore it may be expected that the total drag will be of the order p
times the projected area A of the body (see Figure 4.17). Based on this, it is
common practice to introduce a non-dimensional form of the drag, the drag
coefficient as FD
C  .
1 2
D
 U 0  A
 2 

Projected Plane normal to U0

Area - A

U0
U0

Area - S

Figure 4.17: Reference areas for flow around bluff and streamlined bodies
Fluid Mechanics for Engineering Undergraduates - I 65

In the case of flow around bluff bodies, if the flow configuration is symmetric the
mean lift force is zero. For non-symmetric configurations, there is a contribution of
the pressure distribution on the lift force and therefore the lift coefficient is also
defined by the expression (where CD and FD are replaced by CL and FL,
respectively.

However for streamlined bodies, the above definition is not appropriate as the drag
force is mainly due to the shear stress distribution and therefore proportional to the
wetted surface area of the body. Consequently, the drag coefficient for streamlined
bodies is defined by FD
CD  ,
1 2
  U 0  S
2 
where S is the area of the body projected parallel to the flow, which is proportional
to the wetted surface area (approximately 50%).

The lift force on a streamlined body is due to the difference in pressures on upper
and lower surfaces and therefore proportional to the area S defined above. Hence,
the above expression is used to define the lift coefficient, in which CD and FD are
replaced by CL and FL, respectively.

CD = 1.98 CD = 1.18

CD = 2.0 CD = 1.0 – 1.2

CD = 1.7
CD = 2.2

CD = 1.4 CD = 0.4

CD = 1.3
CD = 1.1

CD = 2.0
CD = 0.7

2-D Shapes 3-D Shapes

Figure 4.18: Drag coefficients of typical 2-D and 3-D bluff bodies.
Fluid Mechanics for Engineering Undergraduates - I 66

Figure 4.18 presents typical values for drag coefficients of 2-D and 3-D bluff body
shapes at relatively high Reynolds numbers.

Note that the force coefficients are in general function of Reynolds number and
surface roughness. However, in a wide range of Reynolds number these coefficients
remain approximately constants. For example, the force coefficients of some typical
body shapes are: Circular cylinder CD=1.1, Rectangular cylinder CD=2.0, Sphere
CD=0.4, Aerofoil CD=0.02, CL=1.2.

Exercise 4.21

The figure shows a steady, two dimensional,

incompressible flow around a rigid flat plate of

negligible thickness at a high Reynolds number.
Sketch the patterns of flow for the cases of  =
0, 3, 45 and 90, where  is the angle of Flow
attack. Comment on the relative magnitudes of
the forces acting on the plate for these cases.

Exercise 4.22

The figure shows a cylindrical building of height H and diameter D located on a flat
open land. The steady wind velocity U at the site varies as a function of vertical
height above the ground z, given by the expression U = A(z)n, where A and n are
constants. Assuming that the flow is two dimensional and the drag coefficient CD of
the flow around the body is a constant, derive expressions for the resultant drag
force and the overturning moment acting on the building due to the wind.

Wind
U
H
Building
z

Ground
Fluid Mechanics for Engineering Undergraduates - I 67

Exercise 4.23

A fluid of density and viscosity  flows around a rectangular cylinder of width w

and length l, as shown in the figure. The velocity of the flow in the far-field is U and
makes an angle  to the length side of the cylinder. The flow could be considered as
steady, two dimensional and incompressible.

(a) Sketch the streamline pattern of the flow around the body for each of the
following cases:
(i) For relatively high Reynolds number flow at  = 0, l  0.
(ii) For ideal flow at  = 0, l  0.
(iii) For relatively high Reynolds number flow at  = 0, l >> w.

(b) Sketch, without any calculation, the variation of the drag coefficient CD of a
square shaped body (i.e. l = w) as the angle of attack  increases from 0 to
180 for each of the following cases:
(i) For relatively high Reynolds number flow.
(ii) For ideal flow.
Give approximate values of CD at specific points such as  = 0, 90 and
180.

w
U


l

Exercise 4.24

The figure shows two dimensional, steady and incompressible fluid flow past a
cylinder with half-circular cross-section. The velocity of the flow in the far-field
makes an angle  to the flat side of the cylinder. Sketch the streamline pattern of the
flow around the body for each of the following cases:
(i) For relatively high Reynolds number flow at  = 0,  = 90, and  = 120.
(ii) For ideal flow at  = 0,  = 90, and  = 120.

In each of the six cases above, comment on the relative magnitudes of lift and drag
forces. Among these, identify the configuration that gives the maximum drag force
and the configuration that gives the maximum lift force.

Uniform Flow

Fluid Mechanics for Engineering Undergraduates - I 68

CHAPTER FIVE

DIMENSIONAL ANALYSIS AND

SIMILARITY THEORY
5.1 Introduction

Problems in Mechanics reduce to the determination of certain function and

characteristic parameters. The relevant laws of nature and geometrical relations are
represented as functional equations. In purely theoretical investigations these
equations are used to establish the general qualitative properties of motion and to
calculate unknown physical quantities using mathematical techniques. However,
very often the problem cannot be formulated mathematically because the
mechanical phenomenon to be investigated is too complex and the governing
equations are unknown. Preliminary analysis of such problems is made possible by
Dimensional Analyse and similarity theory.

Dimensional analysis is a mathematical technique which deals with the dimensions

of the physical quantities involved in the phenomenon and thereby establish a form
of the relation between them. The qualitative solution obtained by dimensional
analysis can be converted into quantitative result using experimental data or
mathematical equations of motion.

5.2 Fundamental and Derived Units and Dimensions

The fundamental units and dimensions in SI are presented in Table 5.1.

Table 5.1: Fundamental units and dimensions

Quantity Units Dimension
Mass kilogram (kg) M
Length meter (m) L
Time second (s) T
Temperature degree Celsius (C) 

The expressions of the derived units of physical quantities in terms of the

fundamental units are called their dimensions.

Examples:
[Force] = MLT-2 [Stress] = ML-1T-2 [Velocity] = LT-1
[] = ML -3
[Angle] = M0L0T0 [] = ML-1T-1
[Angular velocity] = T-1 [Specific Heat Capacity] = L2T-2-1
Fluid Mechanics for Engineering Undergraduates - I 69

5.3 Dimensional Homogeneity

An equation is called dimensionally homogeneous if the dimensions of all the terms

in that equation are identical. Foe example consider the following equation
u 1 p  2u
  2 .
t  x y
The dimensions of the each term of the above equation are the same and given by

 u   1 p    2u 
   2   LT
-2
 t    x 
     y 

5.4 Methods of Dimensional Analysis

5.4.1 Rayleigh's Method

In this method, the functional relationship of some physical quantities is expressed

in the form of an exponential equation. This expression is then rearranged and
simplify by using the principle of dimensional homogeneity.

Consider a general relationship between several variables in the functional form as

Y = (X1, X2, X3, .........) =  Ai X1ai X2bi X3ci .... ,
where Y - dependent quantity
X1, X2, X3, .. - independent quantities
Ai, ai, bi, ci, .. - dimensionless constants

Then find relationships between ai, bi, ci, ..., in order to satisfy the condition
[Y] [X1a1 X2b1 X3c1 ...] [X1a2 X2b2 X3c2...]  ....
Substitution of the above results in the functional relationship leads to simplification
of the general form, in which the variables (including the dependent variable) could
be arranged in the form of dimensionless groups.

Example: Fluid motion in pipes

The pressure drop along a pipe per unit length is known to depend on the diameter-
d, coefficient of viscosity-, density-, the mean velocity-u and the effective surface
roughness k. Derive an expression for the pressure drop per unit length p/l in terms
of these quantities.
Let p/l = (d, u, , , k) =  Ai di ui i i ki ....... (1)
Substituting dimensions of each quantity in Equation (1) it can be shown that, for
dimensional homogeneity
Fluid Mechanics for Engineering Undergraduates - I 70

 i = 1 - i , i = 2 - i and i = -1 - i - i .
By substituting these results in Equation (1) and after some simplifications

p  u 2   ud   k  i
i
 u 2 
l
 
 d  i
 
 Ai  

   d 
    d  Re,  ,




where Re = ud/ = Reynolds number and  = k/d = relative roughness. Then the
head loss is given by p  l  u   l  u 
2 2
hf      Re,    4 f   ,
g  d  g   d  2 g 
where f = /2 is called the friction factor.

Now the problem reduces to determination of the function f(Re,). Experimental

data shows that the function f(Re,) has two main branches one of which
corresponds to the laminar and the other to the turbulent modes of motion (see
Moody chart given in Figure 4.8).

Laminar Region: Re< 2000 f = 16/Re

Turbulent Region: (i) Smooth pipe (=0) 3000 < R e<105 : f = 0.079Re-1/4
(ii) Rough pipes 4000 < R e< 107, 0< <0.01 :
f = 0.0014[1+100(0.02+ 1/Re)1/3]

5.4.2 Buckingham's  - theorem

When a large number of variables are involved in the problem, the Rayleigh's
method of analysis becomes difficult and the Buckingham's method may be
conveniently used.

According to this theorem, if there are n variables (including the dependent

variable) in a functional relationship, and if these variables contain m fundamental
dimensions, they may be grouped into n-m independent dimensionless terms
(referred to as  - terms).

Mathematical Representation:
Let Y    X 1 , X 2 , ..........., X n 1 
If there are m fundamental dimensions, then according to Buckingham's  - theorem
  ,  1 ,  2 , ...........,  n m 1   constant

Basic Steps of the Method:

(1) Write the general equation of the functional relationship using the given data.

(2) Select m variables to use as repeating variables from the independent

variables.
Fluid Mechanics for Engineering Undergraduates - I 71

Note. (i) Repeating variables should contain all the fundamental

dimensions among them.
(ii) They should be quantities which are likely to have a substantial
effect on the dependent variable.
(iii) In general the repeating variables selected are , V and l , that is
 representing the fluid property, V representing the flow
characteristic and l representing the geometrical characteristic of
the body or the boundary.

(3) Combine the repeating variables, in exponential form, with the remaining
variables taking one at a time to form the independent dimensionless groups
(A simpler method of deriving dimensionless groups involves representing
the fundamental dimensions in terms of dimensional equation of repeating
variables).

(4) Write the functional relation in the required form,

n-m-1)

Note. Since the  - groups are dimensionless, the functional relationship between
them can be rearrange by the process of division, multiplication or by
raising to any power, provided that the total number of  - terms is
contained in the resulting groups. As an example, expression
4) could be rewritten as 3).

Exercise 5.1 Show that the force F acting on a solid body moving in a fluid could
be expressed by  Vl V k l 
F  l 2V 2  , , , ,  ,
  gl l V 
 
where V, , l and k are the linear velocity, angular velocity, the
characteristic length and the surface roughness of the body,  and 
are the density and viscosity of the fluid and  is the angle of attack.
Simplify the above expression for the following cases:
(i) Spherical sport ball, (ii) Aircraft wing, and (iii) Ship.

5.5 Physical Significance of Non-dimensional Groups

Non-dimensional groups represent the ratios of two similar physical properties of

the system. They are useful means of defining the conditions which exist in the
system and indicating which properties are of importance. As an example, consider
the non-dimensional group  = F/V2l2 obtained in the previous problem. Since F is
the total force on the body, the term V2l2 should also present a force existing in the
system. This term can be rewritten as
2
l 
 V 2 l 2 =    l 2 =  l 3  2   Mass  Acceleration = Inertial Force Fi
l 
t  t 
Fluid Mechanics for Engineering Undergraduates - I 72

Thus F/V2l2 represents the ratio total force/inertial force. In general, the resultant
force is a combination of different types of forces and the ratio of each force to the
inertial force indicates the relative importance of the force type in a particular flow
system. Table 5.2 lists some of the non-dimensional groups that are encountered in
fluid mechanics problems with engineering applications. Many of these groups have
alternative definitions that may depend on the particular problem.

Table 5.2: Non-dimensional groups

Name Symbol Formula Meaning Application
Bingham No. Bm yl/(U) (Yield/Viscous) Stresses Bingham Plastic flow
Cavitation No. Q (p-pv)/(½U2) Excess over vapour Cavitation
pressure/dynamic
pressure
Cauchy Ca U2/K (Inertial/Compressibility) Compressible flow
Number forces
Dean Number De UR3/2/(r1/2) (Centrifugal/Viscous) Flow in curved pipes
Force
Force CF F/(½U2A) Resultant Force/Dynamic External Flows
Coefficient Pressure Force
Eckert Ec U2/(cpT) (Kinetic/Thermal) Energy Heat transfer in
Number flowing fluid
Ekman Ek /l2 Turbulent/Rotational Rotating flow
Number terms
Euler Number Eu p/U2 (Pressure/Kinetic) Energy General flow
Friction Factor f or w/(½U2) Wall stress/Momentum Flow in ducts
Cf flux
Froude Fr U/(gl)1/2 (Inertial/Gravity) Forces Free Surface Flow
Number
Grashof Gr Tg l32/2 Heat transfer/Heat by Natural Convection
Number friction
Head CP P/(Q 2l2) (Out put/Rotational) Rotating Machinery
Coefficient Power
Mach Number M U/c Flow speed/Sound speed Compressible flow
Nusselt Nu hl/k Heat transfer/Diffusivity Heat transfer
Number
Peclet Number Pe cpUl/k=Ul/Ð Advection/Diffusion Mass or heat transport
Prandtl Pr cp/k Dissipation/Conduction Heat Conduction
Number
Relative  k/d Absolute roughness/ Turbulent flow in
Roughness Characteristics length ducts
Reynolds No. Re Ul/ (Inertial/Viscous) Terms Viscous flow
Specific Heat  cp/cv Enthalpy/Internal Energy Compressible flow
Ratio
Strouhal No. St l/U Time scale of flow/ Flow around bluff
Oscillating period bodies, Vortices
Weber We U2l/ Inertial/Surface Tension Free surface flow
Number
Fluid Mechanics for Engineering Undergraduates - I 73

5.6 Theory of Physical Similarity and Model Testing

5.6.1 Physical Similarity

Dimensional analysis and similarity theory play a very important part in

experiments and calculations in Fluid Mechanics, through testing of models. Wind
tunnel testing on air-crafts; ground vehicles and civil engineering structures; towing
tests on ships and submarines; river and harbour tests are some examples of model
testing.

The basic idea of modelling is that the information required about the nature of
effects and various quantities related to the original problem can be derived from
the results of experiments with models. Dimensional analysis and similarity theory
determine the conditions under which the model experiments are to be carried out
and identify the key parameters involved in the problem.

In general, two or more systems are said to be physically similar in respect to

certain specified physical quantities when the ratios of corresponding magnitudes of
these quantities between the systems are everywhere the same. Usually comparisons
are made between the actual system (prototype) and its scale model (or model). This
is called model testing. Model tests usually depend on three basic types of physical
similarity: Geometric, Kinematic and Dynamic similarities.

(a) Geometric similarity

This is similarity of shape. The ratio of any length in one system to the
corresponding length in the other system is the same throughout. This ratio is called
scale factor. For example, Figure 5.1 shows prototype car and its 1:4 scale
geometrically similar model.

Prototype Model
Figure 5.1: Geometrically similar systems (scale factor 1:4)

Usually, it is impossible to achieve exact geometrical similarity between the

prototype system and its scaled-down model constructed for laboratory testing.
Followings are among the reasons for the departure from true geometric similarity
in models:
(i) Failure to scale the roughness of the solid boundaries.
(ii) There can be confined boundaries of the model experiment which are
absent in the prototype system.
Fluid Mechanics for Engineering Undergraduates - I 74

(b) Kinematic similarity

This is similarity of motion, which implies both geometric similarity and similarity
of time intervals. Therefore velocities and accelerations must also be similar. When
two systems are kinematically similar, the flow patterns (streamlines, path lines and
streak lines) are geometrically similar.

Prototype Model

Figure 5.2: Kinematically similar systems

(c) Dynamic similarity

This is similarity of forces. Then the magnitudes of forces at similarly located points
in each system are in a fixed ratio at corresponding times. Dynamic similarity
requires both kinematic similarity and geometric similarity.

Consider a fluid particle in the flow system. In general, the particle is subjected to
forces due to pressure (Fp), friction (Ff), gravity (Fg), etc., which results in inertial
force (Fi). Then for similarity of forces
[ F f ]P [ F p ]P [ F g ]P [ Fi ]P
= = = = ......   F
[ F f ]M [ F p ]M [ F g ]M [ Fi ]M
or replacing F in terms of [Fi]P/[Fi]M
[ Fi ]P [ Fi ]M [ Fi ]P [ Fi ]M
= , = , etc.
[ F f ]P [ F f ]M [ F g ]P [ F g ]M

Above expressions implies that between dynamically similar systems, the non-
dimensional groups such as Reynolds number, Froude number, etc. are the same.
That is ReP = ReM, FrP = FrM , etc.

Therefore, in order to achieve complete similarity, the non-dimensional groups or

parameters derived through dimensional analysis should be kept the same for the
prototype system and the model. This requirement defines the conditions under
which the model testing should be carried out and how the test results related to the
prototype values.
Fluid Mechanics for Engineering Undergraduates - I 75

5.6.2 Model testing of flows without free surfaces

Majority of the fluid flow problems could be considered under the category of flow
without free surfaces. These include both internal and external flows such as
(i) Flow around submerged bodies (aircrafts, submarines, structures etc.)
(ii) Flow through bounded systems (pipe flow, building ventilation, etc.)

In most of the practical applications under this category, the predominant factors are
Reynolds number Re and March number M. Therefore the expression for the
dependant parameter can be reduced to  = (Re, M). If the flow velocity is small
enough to neglect compressibility effects (ie M < 0.3),  = (Re). Therefore the
predominant factor that must be kept constant between the model and the prototype
is Re.

If the model is smaller than the prototype, in order to satisfy the condition of
identical values for Re, either the velocity must be increased (which is usually
restricted in practice) or the kinematic viscosity of the fluid must be altered
substantially. In modelling a motion of a body in air either a water tunnel or a
compressed air wind tunnel could be used.

Note that another parameter of importance for this category of flow, though not
related to a force, is the surface roughness k. The non-dimensional group associated
with this is the relative roughness  = k/d. Relative roughness could alter the
behaviour of flow and thereby could contribute to the change in both viscous and
pressure forces and therefore becomes an important parameter to be considered for
dynamic similarity. Another important parameter for the case of flow around bodies
is the orientation of the body to the oncoming flow (i.e. angle of attack). For
example the friction factor of the flow through a pipe depends on Reynolds number
and relative roughness f = f(Re, ), lift coefficient of an aerofoil depends on angle of
attack and Reynolds number CL = f(, Re), and drag coefficient of a sphere depends
on Reynolds number and relative roughness CD = f(Re, ).

Exercise 5.2 A fluid of density  and viscosity  flows around a rectangular

cylinder of width w and length l, as shown in the figure. The velocity
of the flow in the far-field is U and makes an angle  to the length
side of the cylinder. The flow could be considered as steady, two
dimensional and incompressible.

Use Buckingham- theorem to derive an

expression for the resultant force acting
w on the body in terms of given parameters
U in the form of non-dimensional groups.
 Interpret the result, indicating physical
l meaning/significance of each non-
dimensional group.
Fluid Mechanics for Engineering Undergraduates - I 76

5.6.3 Model testing of free surface flows

Free surface flows could occur in both internal and external flows categories where
the fluid is a liquid. The most common examples include
(i) Open channel flow, flow over weirs and notches etc.
(ii) Motion of partially submerged body (Ship motion etc.)

In free surface flows, usually gravity effects are predominant (especially for open
channel flow and flow over weirs and notches etc.) and therefore  = (Fr). For the
similarity FrP = FrM (such a model is called Froudian model). However for the case
of motion of partially submerged bodies, such as ship motion, both gravity effects
(wave resistance) and frictional effects (shear force) are predominant. Therefore,
compared with most of the other applications, experimental testing of ship motion is
relatively complex as two non-dimensional parameters have to be kept constant
between the prototype and the model. In order to handle this issue, a special
methodology has been developed, as described below.

The total drag force acting of a ship can be written, with usually notations, as
 Vl V 
F
    ,    Re, Fr .
u l
2 2   gl 
 
Then to have complete dynamic similarity, ReP = ReM and FrP = FrM. That is the
condition on Reynolds number results [ul/]M = [ul/]P or if an identical fluid is
used uM/uP = lP/lM. In contrast, the condition of Froude number leads to [u/gl]M =
[u/gl]P or uM/uP = (lM/lP). According to the two relations, it can be seen that
complete modelling of the phenomenon is impossible.

The determination of ship drag using model experiments is based on the practical
possibility of decomposing the drag into two components:
F = Ff + FW = Skin friction drag + Wave drag
= u2l2 f[Re] + u2l2 w[Fr]

Compared with the skin friction drag Ff, the wave drag FW is more difficult to
analyse. Usually Ff for the model and prototype are found by analytical techniques
based on semi-empirical formulae. Then wave drag FW can be determined by model
analysis with the condition FrP = FrM or [FW/u2l2]M = [FW/u2l2]P .

The basic steps of this analysis can be summarized as follows:

Step 1. Measure [F]M
Step 2. Calculate [Ff]M using semi-empirical formula
Step 3. Evaluate [Fw]M = [F]M - [Ff]M
Step 4. Calculate [Fw]P using the similarity law
Step 5. Calculate [Ff]P using semi-empirical formula
Step 6. Evaluate [F]P = [Fw]P + [Ff]P
Fluid Mechanics for Engineering Undergraduates - I 77

Exercise 5.3 A model of a ship of length 2.5 m and wetted surface area of 1.15 m2
is tested in fresh water and the measured total resistance at a speed of
2.5 m/s is 18.0 N.
(i) Determine the speed of a similar ship of length 62.5m.
(ii) Estimate the total resistance and the effective power of the ship
in seawater at the above speed using the following formula with
usual notation for the skin friction coefficient:
S V 2 = 0.075 log10 Re  - 22 .
1
C f = Df
2
For fresh water  = 1000 kg/m3 ,  = 1.1410-6 m2/s.
For sea water  = 1025 kg/m3 ,  = 1.1910-6 m2/s.

5.7 Distorted Models

Flow in a river or harbour cannot be modelled in the laboratory with the same
geometrical scale for vertical and horizontal dimensions, since the depth of the flow
is comparatively smaller than the other dimensions. In these situations if the depth is
modelled using the same scale as length and breadth (i.e. undistorted model)
(i) the depths involved will be very small so that construction of model will be
difficult,
(ii) flow regime in the model may be different from the prototype.

Thus different scales for vertical and horizontal dimensions have to be used. This
type of models is called distorted models. For example, typical scale factors used
are, for vertical scale: z = 1/50 - 1/100 and for horizontal scale: x = y = 1/250 -
1/500. In this type of flow, it is expected that the flow type will be the same at
corresponding points if the Froude Number Fr is the same.

Exercise 5.4 A model of a river is built to a horizontal scale of Xr and a vertical

scale of Yr. If at any location the depth of flow of the prototype river
cross-section is denoted by Yp and the width by Xp, derive the
following, assuming that the model and the prototype have the same
bed slope.
(i) Average flow velocity scale ratio: Vr = Yr
(ii) Average volumetric flow rate scale ratio: Qr = XrYr3/2
(iii) Manning's coefficient scale ratio:
2/3
 X p      1 2/3 1/2
nr = Y 1/6 1  2  1 + 2 Y r  Y p  ; where V = R So
Y p    X r  X p 
r
  n
(iv) Time scale ratio: Tr = Xr/Yr1/2 ,
(v) Horizontal acceleration scale ratio: Hr = Yr/Xr ,
(vi) Horizontal force scale ratio: FHr = XrYr2
(vii) Vertical acceleration scale ratio: Vr = Yr2/Xr2 ,
(viii) Vertical force scale ratio: FVr = Yr3
(xi) Scale ratio of fall velocity of silt: Wr = Yr3/2/Xr .
Fluid Mechanics for Engineering Undergraduates - I 78

Exercise 5.5

It can be shown that, in general, the force F acting on a solid body moving in a fluid
is given by  Vl V k l 
F  l 2V 2  , , , ,  ,
  gl l V 
 
where V, , l and k are the linear velocity, angular velocity, the characteristic length
and the surface roughness of the body,  and  are the density and viscosity of the
fluid, g is the acceleration due to gravity and  is the angle of attack.

(a) Show that the above expression satisfies the conditions specified in
Buckingham -theorem. Indicating the reasons clearly, simplify this
expression for the case of ship motion.

(b) Use the result in (i) to show that the complete modeling of the phenomenon of
fluid dynamics of ship motion is not possible with scale down models.
Explain clearly how this difficulty is overcome in model testing of ships.

(c) A 1/25th scale model of a cargo ship is tested in a fresh water towing tank.
The speed of the prototype ship in seawater is 8.0 m/s. The length of the
model is 5.0 m and its wetted surface area is 4.8 m2.

(i) Estimate the speed of the model.

(ii) If the measured total resistance of the model is 32.0 N, calculate the
total resistance of the ship at the speed of 8 m/s.

You may use the following data:

For fresh water, density = 1000 kg/m3 and dynamic viscosity = 1.1410-3
Ns/m2.
For seawater, density = 1025 kg/m3 and dynamic viscosity = 1.2210-3 Ns/m2.

Skin frictional drag Ff = 160.65fSV1.825, where Ff is in N, wetted surface area S

is in m2 and velocity V is in m/s and f is the skin friction coefficient. For the
model f = 0.011 and for the ship f = 0.009.

Exercise 5.6

Derive an expression for the volume flow rate of a liquid (of density , viscosity 
and surface tension ) over a V-notch of given angle . Experiments show that for
water flowing over a 60 V-notch, a useful practical formula is Q = 0.762 h2.47 in SI
units, where h is the head over the notch. What limitation would you expect on the
validity of this formula? Determine the head over a similar notch when a liquid
with a kinematic viscosity 8 times that of water flows over it at the rate of 20
liters/s.
Fluid Mechanics for Engineering Undergraduates - I 79

Exercise 5.7

The pressure rise p, generated by a fan of given geometry depends on the impeller
diameter D, its rotational speed N, the fluid density , viscosity , bulk modulus K
and volumetric flow rate through the fan Q. Show that the relations between these
variables may be expressed by
p  Q N D2  N 2 D2 
=   , , 
 N 2 D2  ND
3
 K 
A 0.4m diameter fan running at 970 rpm is tested when the air temperature is 10oC
and the pressure 772mm Hg and the following data are observed: Q = 0.7 m3/s; p
= 25mm of water; shaft power=250W. Find the corresponding values of Q, p, and
shaft power of a 1m diameter geometrically similar fan, running at 500 rpm when
the air temperature is 16oC and the pressure is 760mm Hg. Assume that the fan
efficiency is unchanged.

Exercise 5.8

An orifice of diameter d is used to measure the volume rate of flow Q of a fluid

having density  and coefficient of viscosity  along a pipe of internal diameter D.
The difference in pressure across the orifice is p. Show by dimensional analysis
that the volume rate of flow is given by the expression
p
Q = Cd 2 ,

in which the coefficient C depends only on the value of the criterion (/1/2dp1/2)
for all geometrically similar pipe arrangements in which d/D is constant.

A plate with a 90 mm diameter orifice was fitted into a 250 mm diameter pipeline to
measure the rate of water and the difference in pressure head across the orifice was
820 mm of mercury.

If the flow of air in a 100 mm diameter pipe is to be dynamically similar calculate

(i) the difference in pressure across the orifice in N/m2,
(ii) the ratio of the volume rates of flow.

For water =1000 kg/m3, =1.1410-6 m2/s; and for air =1.28 kg/m3, =14.810-6
m2/s.