Fluid Mechanics Student Notebook

Department of Chemical Engineering and Biotechnology
Department of
Fluid Mechanics
CET I 2017-2018
Dr Mark Williamson
Name: College : email:
University of Cambridge ☐ Department of Chemical Engineering and Biotechnology ☐ CET1 Fluid Mechanics ☐ version 13/09/2017 07:38:59 ☐ 1
Acknowledgement
These lecture notes are based, in part, on

material previously presented in this course
Department
by Prof. Markus of
Kraft and others.
Mark Williamson
Cambridge
September 2014
1. Introduction
This is a course of 16 lectures covering the statics (stationary) and dynamics (moving) behaviour of fluids. Related courses:
Part I: Fluid Mechanics Laboratory explores experimentally some of the material covered in these lectures.
Part IIA: Fluid Mechanics II covers more advanced topics – equations of motion, compressible flow and 2-phase flow
Part IIB: Computational Fluid Dynamics – use of ANSYS CFX to model complex flow fields
Why do Chemical Engineers need fluid mechanics?
In Chemical Engineering, fluid mechanics is important in almost all unit operations. We use our understanding of fluid mechanics
to design optimal equipment and processes. Today, the analytical methods and calculations presented here are incorporated into
sophisticated computer modelling software, capable of predicting the fluid flow field and heat transfer in complex geometries.
Single-phase flow in tubes, pipes and ducts Mixing of fluids in vessels
1. Introduction
Separation processes Multi-phase flow in pipes and ducts
1. Introduction
Fluid mechanics is also used extensively in other disciplines:
Mechanical Engineering - pumping and power generation, aeronautics, turbine

design, wind forces on buildings
Applied Mathematics and Physics – behaviour of stars and galaxies, modelling
of solid suspensions, polymer rheology
Geography – modelling of rivers, oceans, the earth’s atmosphere, contaminant
dispersion (plumes, slicks). Weather forecasting.
Geology – prediction of magma/rock flows, flow of fluids through porous solids
Medical Sciences - blood flow, breathing
1. Introduction
The aims of this course in Fluid Mechanics are:
i. To introduce the relevant physical parameters and basic

equations for the steady flow of ideal and Newtonian
fluids
ii. To predict the behaviour of systems with simple
geometries, and with practical applications to
engineering (using the equations developed in (i) )
iii. To analyse laminar and turbulent flow in pipes and pipe
networks so that the operational characteristics of such
systems can be predicted
iv. To introduce the method of dimensional analysis for

modelling of systems with fluid flow, including the
prediction of the performance characteristics of centrifugal
pumps
v. To analyse and predict the interaction between pumps and
pipe networks
vi. To analyse the relationship between flow and pressure

drop in a packed bed, and the requirements for
fluidisation of the particles in a bed
vii. To predict the performance characteristics of packed

beds and filters (using the equations developed in (vi) )
1. Introduction
Chemical Engineering Chemical Engineering Fluid Mechanics and

Fluid Mechanics
Vol. 1, J.M. Coulson and Vol. 2, J.M. Coulson Transfer Processes, J.M.
Suggested reference texts J.J. Richardson, and J.J. Richardson, Kay and R.M. Nedderman,
Pergamon Press 5th Ed. Pergamon Press 4th Cambridge University
1996 Ed. 1997 Press, 1985
1. Introduction 1, 2
Fluid statics. Physical properties, pressure and shear force. Surface
tension, contact angle, capillary effect. Laminar and turbulent flow
2. Flow of ideal fluids 2, 6.3 3
Continuity equation. Bernoulli's equation. Velocity and flow
measurement. Momentum equation. Unsteady flows
3. Laminar flow 3.5, 3.6 5.9
Flow down a vertical plate. Flow in a cylindrical pipe. Flow between
parallel plates. Stokes' Law.
4. Turbulent flow 1.4, 1.5, 3.3 - 3.5 8.1, 8.2, 15.1, 15.2
Force on a sphere, friction factors. Energy losses. Boundary layers.
5. Pumps 6.1 - 6.3 10.1 - 10.4
Introduction to pump types. Dimensional analysis. Pump and
pipeline characteristics. Cavitation and NPSH.
6. Flow through packed beds 4.1, 4.2, 6.1 - 6.3 21.3 - 21.5,
Phase volume fractions, superficial and interstitial velocities. Darcy’s
Law and Ergun equation. Filtration. Introduction to fluidisation. 21.6, 21.7
7. Liquid mixing 7
Types of mixing, mixing mechanisms (laminar/turbulent), flow
patterns and scale-up for stirred vessels, mixing power requirement
Two (highly related) recent books are also very useful:

Fluid Mechanics; fundamentals and applications, Y.A. Cengel and J.M. Cimbala, McGraw Hill (2006). The full text.
Essentials of Fluid Mechanics; fundamentals and applications, J.M. Cimbala and Y.A. Cengel, McGraw Hill (2008) A shorter version of the above, but adequate.
1. Introduction
Thursday, November 11, 1999 So….

We will be working in ‘strict’
Orbiter loss blamed on 'silly mistakes' SI units whenever possible.
By BBC News Online Science Editor Dr David Whitehouse
This will help to check the
…the $125m Mars Climate Orbiter spacecraft validity of the various energy,
should have gone into orbit around Mars on 30 momentum and mass
September, but instead burnt-up in the planet's balances that we will be
atmosphere…… developing.
Instead of passing about 150 km above the Martian
To do these checks quickly
atmosphere before entering orbit, the spacecraft and effectively under
actually passed about 60 km above the surface. This examination conditions you
was far too close, and the spacecraft burnt up due must know, (without
to friction with the atmosphere. thinking!):
Mission controllers said that the "root

cause" of the loss of the spacecraft was
the failure to convert English
measurement units into metric units in
the navigation software.
A navigation team at the Jet Propulsion
Laboratory used the metric system of
Newtons and meters in its calculations,
while Lockheed Martin Astronautics in
Denver, which designed and built the
spacecraft, provided crucial data in the NO CALCULATIONS IN THIS COURSE
English system of inches, feet and pounds JUSTIFY A RESULT WITH MORE THAN 3
SIGNIFICANT FIGURES – OFTEN ONLY 2!!!!
1. Introduction
What are fluids?

static equilibrium:
Matter normally exists in one of three states: gas, liquid or solid depending on the ….when all the particles of a
temperature (T) and pressure (P). system are at rest, and the total
Gases and liquids both flow when a force is applied to them and are classified together force on each particle is
as fluids. permanently zero
A fluid has no shape other than the vessel in which it is contained, whereas a solid has a shearing forces:
definite shape . More formally: ……unaligned forces pushing one
part of a body in one direction,
"The distinguishing feature of a fluid is its inability to resist
and another part the body in the
shearing forces while remaining in static equilibrium”
opposite direction
Gases – fundamental properties For air we will assume:
• Gases are compressible and will fill any space that is available to them.
• Small bubbles of gas dispersed in a liquid have a spherical shape when at rest, but
distort significantly when moving
• A thermodynamic equation of state can be used to predict the volume and density of
a gas as a function of the absolute temperature and absolute pressure.
• In this course we shall assume that gases are "perfect" (also known as "ideal"), and
that the relationship between P, V and T is given to sufficient accuracy by:
P = absolute pressure (N/m2 or Pa)
………………………………………………………………………………………………. (1.1) V = absolute volume of gas (m3)
n = number of moles of gas (mol)
N 3  J 
m = mol  K R = Universal gas constant 8.31 (J mol-1K-1)
m2  mol ⋅ K  T = absolute temperature (K)
1. Introduction
…………………………………………………………………………………………………………. (1.2) Compressibility of water

5.1 × 10-10 Pa-1
kg  (kg / mol)⋅ (N / m 2 )  Deep ocean (4 km) 400 bar 2 %
=  increase in density
m3  (J / (mol ⋅ K)⋅ K 
Effect of temperature:
ρ = the mass density of the gas (kg/m3) 0 °C 1000 kg/m3
Μ = molecular mass of gas (kg/mol) 100 °C 960 kg/m3 4% decrease
Liquids – fundamental properties
Attractive forces between molecules in
a liquid much stronger than those in a
• Compared to gases, liquids are relative incompressible
gas….
and a certain mass of liquid will have a certain volume
…..but not strong enough to prevent
(which varies only slightly with T and P)
‘flow’ …..
whereas in a solid the attractive forces
• In this lecture course we shall always assume that
are strong enough to prevent flow.
liquids are incompressible, i.e. their density does not
vary with temperature or pressure.
• A liquid will exhibit a free surface if a larger space is

available than that required to contain its volume. Molecules at the surface of a liquid
experience a net force towards the
• Liquids are often found in a dispersed form as droplets. bulk.
These droplets can be in a spray (gas as the continuous The resulting (small) pressure means
phase) or an emulsion (immiscible liquid as the that liquid surfaces contract to give the
continuous phase). In both cases the droplets are minimum possible area for a given
spherical when stationary. volume – which is a sphere.
1. Introduction
Forces and physical properties in fluid mechanics
Pressure, P, is defined as the normal force on a surface, per unit area
…………………………………………………………………………………………………….……. (1.3)
N N
 2 = 2
m  m
N/m2 is also abbreviated as Pa (Pascals)

Other units of pressure commonly used in the chemical engineering literature:
• 1.0 bar = 100 000 Pa
• 1.0 pounds per square inch (psi) = 6895 Pa
• 1.0 mm of mercury (mm Hg) = 133.3 Pa
A useful correlation:
5.256
Atmospheric pressure results from the weight of the earth’s atmosphere on an object  44331 − z 
P = 0.1 
Sea level (el. 0 m) 101.3 kPa  11881 
Johannesburg (el. 1753 m) 82.0 kPa P = atmospheric pressure (kPa)
Mt. Everest (el. 8848 m) 31.4 kPa z = altitude above sea level (m)
1. Introduction
The earth’s atmosphere is an example of the variation of pressure, P,

with elevation (or height = h) in a stagnant fluid: P = f(h)
It is a result of the earth’s gravitational pull (or force, F ) on the gas
molecules in the atmosphere
Consider a small element of the atmosphere, of cross-sectional area A
(m2):
Gravitational force
elevation area
P+dP F=m.g
A
h+dh P
h
h=0 Surface of the earth
1. Introduction
For liquids, the situation is less complex, since the liquid density is not a Force balance on the small element in the
significant function of height within the liquid. Consider liquid in a tank: ‘h’ direction:
Force on Force on Weight
Variations in lower = upper + of small
Patm atmospheric pressure surface surface element
don’t change this
P+dP
analysis, since the
upper and lower
h+dh surfaces of the small
h P
element are equally
affected.
h=0
The pressure of a gas in a vessel can be measured using a U-tube manometer:
PA = PA' (A-A’ same height in liquid)

Ptop
Pgas = Patm + hL ρ L g (B’ – A’)
Patm
h  Pgas − Patm  = hL ρ L g
Pgas
H = 3.0 m
B B’
The gauge pressure in the vessel is
hL  Pgas − Patm  kPa(g)
A A’
h=0
Pbottom The absolute pressure in the vessel is
Pgas = [ Patm + hL ρ L g] kPa(abs)
1. Introduction
The same principle as the manometer can be used to measure the The level controller will use the signal from the
level of a liquid in a vessel, such as a flash pot. The level must be differential pressure sensor to regulate the flow
measured and controlled so that only gas leaves the top and only of liquid out of the vessel, by adjusting a control
liquid leaves the bottom. valve in the liquid line, or the speed of a pump
that is transferring the liquid away from the
Gas vessel.
The system can operate satisfactorily even when
the pressure in the gas phase is changing, since
gas pressure = Pgas both sides of the sensor are equally affected by
this pressure.
Gas + liquid
Differential
pressure
h=hL transmitter
Signal is
h=0 ∆PE proportional to
liquid level in
vessel
liquid
1. Introduction
Surface tension
In Chemical Engineering applications we frequently encounter liquids

in dispersed forms - as drops, sprays or entrained mists.
Liquid droplets are spherical, consistent with the principle that the
surface energy will attain the minimum possible value for a given
volume of liquid.
The surface tension force (σs) acts parallel to the droplet or bubble
surface. It is usually expressed as force/unit length (N/m).
A spherical interface implies a pressure difference ∆P
Spray drier Liquid/gas
(e.g. for milk powder) separator
1. Introduction
Capillary effect Selected values

This is phenomenon of the rise (or fall) or a liquid within a small-diameter open-ended for σs (N/m)
tube that is inserted into a liquid. It is caused by the differing magnitudes of cohesive
forces between like molecules (e.g. water-water) and adhesive forces between unlike
molecules (e.g. water-glass). Consider a force balance on the element of water shown:
These guys also know all about surface tension….
Φ is the contact angle or wetting angle.
Φ < 90° liquid wets the surface Φ > 90° liquid does not wet the surface
Φ (mercury/glass/air)= 130 °
Φ (water, organic liquids/glass/air)∼0 °
1. Introduction
Archimedean up-thrust
Archimedes' principle states that the upward buoyant force that is exerted on a body
immersed in a fluid, whether fully or partially submerged, is equal to the weight of the
fluid that the body displaces. (remember that weight is a force!)
Consider a solid object (of volume V, density ρs ) immersed in a fluid of density ρL :

287-212 BC
Fu < Fd
Fu = upthrust force
Fd = downward force
Fu = Fd Fu = upthrust force
h H
Fd = downward force
1. Introduction
Brick thrown out of a boat
H1 H2
What is relationship between H1 and H2?

Temperature required inside a hot air balloon
Mballoon=350 kg Vballoon = 2800 m3
5 passengers, averaging 70 kg each
Drifting at 1000 m above sea level

with ambient temperature of 9°C
1. Introduction
Shear stress and viscosity

Shear stress, τ , is defined as
τ = tangential force on a surface, per unit area of that surface
Shear stresses can be sustained in a fluid in motion, but not one at rest.
Consider two parallel plates, the lower plate stationary and the upper plate moving as
shown with velocity uu and a fluid filling the region between them:
y area A
moving force F
upper plate
y=a uu
uu= 3 mm/min
y=0
x
stationary
laminar flow - horizontal lower plate
layers of fluid slipping
past each other τ
τ
Newton’s law of fluid friction defines µ, the absolute viscosity:
du
τ =µ ………………………………………………………………………………….……. (1.4)
dy
N Ns  m / s  We will use the convention that τ (the shear stress on a fluid)
=   is positive for the section of fluid below a shear plane.
m2 m2  m 
1. Introduction
The no-slip boundary condition …….. a closer look Surface trace of a typical 2B polish finish on
stainless steel flat plate (Ra = 0.00020 mm)
Nature of Roughness
Material
Material [mm] mm
Steel pipe drawn, new 0.02 - 0.1
galvanized, new 0.15 + 0.0002
used, cleaned 0.15 - 0.2 0
lightly corroded 0.1 - 0.4 - 0.0002
severely corroded 0.4 - 3
light scaling 1 - 1.5
heavy scaling 1.5 - 4
Sheet steel Smooth, unpolished 0.07
Glass, lead, 0.0001 -
polished 0 1 2 3 4 5 6 7 8 9 10
copper, brass 0.0015
Distance across surface (mm)
Conclusions:
Water molecule 0.0002 • The bulk flow of the fluid doesn’t come into significant
mm contact with the wall – there are pools of essentially
0.1 nm stagnant fluid molecules trapped at the wall.
= 0.0000001 mm • First layer of fluid has effectively zero velocity
0.7 mm • Heat transfer normal and very close to the wall is by
conduction only – no convection
2000 water • Mass transfer close to the wall is by diffusion only – no
molecules deep! bulk transport
1. Introduction
Newtonian and non-Newtonian fluids

shear
Bingham
thickening
plastic
τ (quicksand)
(toothpaste)
Newtonian
shear stress
(water)
shear
thinning
critical (paint)
value of τc
must be du
rate of strain or γ&
exceeded
before flow dy
begins
• Most liquids of simple chemical structure behave in a Newtonian

manner
• Liquids with high molecular weight and complex molecular
structure are found not to obey Newton's law and are called
non-Newtonian fluids.
• Liquids which contain dispersions of particulates are often non-
Newtonian.
1. Introduction
Consider the general steady-state case for an element of Newtonian

fluid, of unit depth into the paper, flowing between two flat plates:
uu
dx
τ+δτ dy
2a τ
y
x
stationary
P P+δP plate
At steady state conditions, the forces on the element must balance, so:
∂2u ∂P
µ 2= ………………………………………………………….……. (1.5)
∂y ∂x
Integrating twice, noting that P≠f(y):
1 2 dP
µ ⋅u = y + c1 ⋅ y + c2
2 dx
For a stationary lower plate and upper plate moving at uu:
u= +
(
uu y y 2 − 2 ya dP ) ………………………………….……. (1.6)
2a 2µ dx
2nd term results from pressure gradient

1st term results from moving upper plate
1. Introduction
Now consider the case when both plates are stationary:
u
τ τ+δ dy
2a
dx x
P P+δP
∂2u ∂P
µ 2= ………………………………………………………….……. (1.5)
∂y ∂x
Integrating again, noting that P≠f(y) gives a parabolic velocity profile:
y 2 dP u=
(a 2
)
− y 2  dP 
−  …………….……. (1.7)
u= + c2
2µ dx 2µ  dx 
Q = 2∫ u(1⋅ dy) = 2∫
a
(a 2
)
− y2  dP  1  dP  2 y3  2a3  dP 
 − dy =  − a y −  = − 
a
0 0
2µ  dx  µ  dx  3 0 3µ  dx 
a 2  dP 
umax = − 
2µ  dx 
1. Introduction
Look at velocity profiles within the fluid

u=
( )
uu y y 2 − 2 ya dP
+  y  2a dP  y   y  
2 2
u = uu   +  − 
 2a  µ dx  2a   2a  
(equation 1.6) for some special cases: 2a 2µ dx
laminar flow of a fluid flowing

P P between two stationary plates,
y with a pressure driving force left
to right
x
uu uu

P P P P between a stationary lower
y plate and a moving upper plate,
with a pressure driving force left
x
to right
uu

P P between a stationary lower
plate and a moving upper plate,
with a pressure driving force
right to left
uu

P P between a stationary lower
plate and a moving upper plate,
with no pressure driving force “
Couette flow”
1. Introduction
Viscosity….some useful information

• it depends on the relative motion of adjacent fluid layers i.e. the velocity gradient within the fluid.
• it is related to the "stickiness" of the fluid and determines how well the fluid resists the shear stress.
• it gives rise to energy loss in a fluid (heating)
• viscosity of a gas is about 1% that of a liquid of comparable molecular mass
• viscosity of gases
increases with increasing temperature
is independent of pressure unless close to critical point
• viscosity of liquids
decreases with increasing temperature
(hot treacle flows faster than treacle kept in the fridge)
increases with increasing molecular weight
hexane (M= 86 kg/kmol) µ = 2.94 x 10-4 Pa s
dodecane (M= 170 kg/kmol) µ = 1.34 x 10-3 Pa s
increases rapidly with increasing interactions between molecules, e.g. hydrogen bonding
glycerol (M=92 kg/kmol) µ = 1.41 Pa s
hexane (M=86 kg/kmol) µ = 2.94 x 10-4 Pa s
is independent of pressure.
Some useful viscosity values…..
• air (at 20 °C and 1 atm) µ = 1.81 x 10-5 Pa s
• water (at 20 °C and 1 atm) µ = 1.00 x 10-3 Pa s
Further values are given in the Department of Chemical Engineering Data Book.
All questions, be they on examples papers or Tripos papers, assume that you have access to the information in the Data Book.
You will not be specifically told to find information in there…….
1. Introduction
Measurement of viscosity
Consider the concentric cylinder device, length L …..

a1
If (a2-a1) << a1 then treat as flow between parallel plates
Ω
dP P≠f(y) since L is too small a2
uu = Ωa1 =0
dx for gravity (weight) to be
a significant factor
uu y (y 2 − 2 ya ) dP 2πa1
u= + ……………. (1.6)
2a 2µ dx
Ω⋅ a1 ⋅ y y
2a = (a2 − a1 ) u=
( a2 − a1 ) L
∂u µ ⋅ Ω⋅ a1
τ =µ ….….……. (1.4) τ=
∂y a2 − a1 x
3
2π ⋅ µ ⋅ Ω⋅ L ⋅ ( a1 )
T (measured torque on shaft) = [ 2π a1L ⋅ τ ] ⋅ a1 =
( a2 − a1 )
1. Introduction
Laminar and turbulent flow

In the late 19th Century, experiments investigating the dependence of the
pressure drop on the flow rate of a liquid through a cylindrical pipe gave
conflicting results.
mean um
velocity Pin Pout
of flow ∆P = Pin-Pout
Professor Osborne Reynolds resolved

the problem in 1883……
This figure is taken from his paper:
"An Experimental Investigation of the

Circumstances which Determine
whether the Motion of Water shall be
Direct or Sinuous, and of the Law of
Resistance in Parallel Channels”
Philosophical Transactions of the Royal Osborne Reynolds

Society of London 174, 935-82, (1883). (1842 – 1912)
Fellow
Queens’ College
Reynolds' apparatus consisted of a pipe with a 'trumpet' entry to minimize Cambridge
disturbance to the flow, into which a dye tracer was injected.
His results showed clearly that there are two types of flow, as described on
another page from his paper…..
1. Introduction
1. Introduction
Transition laminar↔ turbulent flow occurs near a fixed critical For flow round a sphere:
value of a dimensionless group, the Reynolds number: Recrit ≈ 100 000
ρud u = characteristic velocity u = approach velocity
Re = d = sphere diameter
µ d = characteristic dimension
(kg / m3 )(m / s)(m)
Ns / m 2
Recritical depends on: Modern traffic

For flow in a pipe
vibrations now cause
• the system geometry Recrit ≈ 2200
Reynolds' original
• surface smoothness u = um
apparatus to give a
• vibration d = pipe diameter
lower value of Recrit
than that found by
Re is actually a dimensionless ratio of forces: Reynolds himself.
  π d 2 
ρ um   um
ρ um d   4  m& ⋅ u
Re = = = = inertial forces
µ 
 um  π d 2
τ ⋅A viscous forces
 µ  
 d  4 
• For low values of Re, the viscous forces are sufficiently large to suppress
fluctuations and keep the fluid in-line
• For large values of Re, the inertial forces are greater that the viscous forces and
the latter cannot prevent random and rapid fluctuations occurring within the
fluid
• URL for web resources: http://web.mit.edu/hml/ncfmf.html
1. Introduction
Laminar and turbulent flow regimes in pipes have different velocity profiles:
Laminar and turbulent flow
2.5
2 laminar
1.5
u /u m
1
0.5 turbulent
0
0 0.2 0.4 0.6 0.8 1
r /a
A velocity profiles takes time to develop fully – a pipe has an entrance region:
2. Flow of ideal fluids
Steady flow of a low viscosity incompressible fluid

Consider fluid flowing through a pipe of constant cross-sectional area = A
(not necessarily cylindrical):
u=um
element of fluid:
cross-sectional area = dA
volume = u.dA
Summing over all elements gives the total volumetric flow rate, Q (m3/s):
∫ u ⋅ dA ⋅ dt
Q= A
= ∫ u ⋅ dA = u ∫ dA = u ⋅ A
dt A A
With the same assumptions, consider fluid flowing through a pipe of

varying cross-sectional area (again not necessarily cylindrical):
u1 u2
1
2
Bernoulli's equation Conservation of mass within the stream

Conservation of energy in fluid flow, applicable to: tube dictates that the incoming and
• steady flows (velocity at any fixed position is independent of time) outgoing masses are the same – dm.
• low-viscosity fluid (effects of viscosity negligible)

• constant density fluid
A stream line is a fluid path. No net flow takes place across a streamline – i.e.
the net velocity vector is a tangent to a stream line along its whole length.
A stream tube is a tube of arbitrary and variable cross section that is entirely
bounded by streamlines - essentially within the bulk flowing fluid it is an
imaginary pipe which has no flow through its walls
P2
u2 bulk fluid flow
P1
h2
u1
h1
datum, h=0
In an incremental period of time, the fluid that is initially entirely contained within
the control surface moves to a new position slightly to the right in the diagram.
• some of the fluid on the right will have emerged from the stream tube
• additional fluid will have entered the stream tube from the left
The Bernoulli equation strictly applies along streamlines, and does not The Bernoulli equation is often written in terms
take account of losses. of ‘heads’:
In practice, some of the energy accounted for by the Bernoulli
equation is converted to heat. ‘EL’ is the lost energy per unit mass of fluid
flow (J/kg).
∆P 1 E
+
ρg 2 g
( )
⋅ ∆ u 2 + ∆h + L = 0
g
…………………………. (2.3) laminar flow: viscous dissipation
friction with the pipe wall
turbulent flow: viscous dissipation
friction with the pipe wall
pressure velocity potential lost rotational flow (eddies)
head head head head
A pump that is transferring a fluid in a
Common applications of Bernoulli's equation – no. 1 chemical plant must deliver sufficient energy
to the fluid account for all of these ‘heads’.
Small hole in the side of a tank, with fluid level initially Hi above the hole
Compare some arbitrary point in the vessel (1) with fluid passing
through the hole (2):
p1 = patm+(Hi-h) ρ g
hole u1 ≈ 0
1
diameter (pseudo steady state
Hi d assumption)
h h1 =h
2 u p2 = patm
u2 = u
h2 =0
D
Area of hole is greater than minimum area of fluid jet at (3),

the vena contracta
1 2
d dj
How long will flow through the hole continue?

Consider a mass balance on the system:
πD2
volume of fluid above the hole : V = Hi (t )
4
Hi
time t
Common applications of Bernoulli's equation – no. 2 Since there is no velocity component normal to the
direction of fluid flow, u1=0
Use of a Pitot tube to measure gas velocity in a duct
The difference in height between the two sensing
positions is negligible – so no hydrostatic pressure
Patm difference
u
no flow inside
Pitot tube measured using a
∆P digital
manometer
Common applications of Bernoulli's equation – no. 3

Use of an orifice plate to measure velocity in a pipe:
d1 do d2
1 2 3
uniform vena uniform
flow contracta flow
plate with circular

1 2 hole diameter do
machined to have
sharp edges
other designs for slurries
This page intentionally left blank for additional notes
Common applications of Bernoulli's equation – no. 4

Use of a venturi meter to measure velocity in a duct
ideal
Static 90% pressure
pressure recovery
actual
distance
The momentum equation
Consider the force exerted by a fluid on a duct of varying

diameter:
1 Patm
force on wall 2
R
R
force on fluid
Patm A2 ρ2 P2 u2
A1 ρ1 P1 u1
Newton’s 2nd law: “force = rate of change of momentum”

applied to region inside the duct between (1) and (2).
Example:
Water discharges to atmosphere at a volumetric flow rate
of 0.007 m3/s through the pipe shown below. What is the
net force on the pipe?
d1 = 4 cm
h=8 cm
d2 = 2 cm
60°
calculate the impulse functions:

2
π ( 0.04)
F1 = A1 ( P1 + ρ u ) =
2
1 1 ( 233000 +1000 × 5.572 ) = 331N
4
2
F2 = A2 ( P2 + ρ2 u22 ) =
π ( 0.02 )
4
( 0 +1000 × 22.32 ) =156 N
these vectors must be added in the usual way to give the

resultant force, R:
force on wall
y R -F2
θ 60°
x F1
Common applications of the momentum equation – no. 1

pressure recovery downstream of an orifice plate:
1 2 3
vena uniform
contracta flow
1 2 3
What is the force on an orifice plate for water flowing at 2 m/s in

a 25 mm diameter pipe, and 6 m/s at the vena contracta ?
vena
uniform
1 2 contracta flow 3
Some kinetic energy is converted to heat. How significant is the
temperature rise? Use Bernoulli’s equation (with losses) from (2)
to (3) ∆P 1 E
ρg 2g
( )
+ ⋅ ∆ u 2 + ∆h + L = 0
g
1 2 P −P
EL =
2
( u2 − u32 ) + 2 3
ρ
1 2 1
EL =
2
( u2 − u32 ) + u1 ( u1 − u2 ) =
2
( u2 − u1 )
2
1 1 2
( u2 − u1 ) = ( 6 − 2 )
2
EL = = 8.0 J / kg
2 2
Temperature rise ∆T is given by:
8.0
EL = c p ⋅ ∆T with cp = 4180 J/kg.K ∆T = = 0.002 K
4180
J J
= ⋅K
kg kg ⋅ K
Common applications of the momentum equation – no. 2
Sudden pipe expansion
1
uniform
2 3
flow
• No vena contracta in this flow geometry, hence assume P2 = P1
• No impact of fluid on any part of the pipe, so R=0
A sudden pipe contraction can be analyzed in in a similar way to

an orifice plate
3. Laminar flow
Laminar (adjective)
• consisting of laminae (singular: lamina).
• of a flow - taking place along constant streamlines; not turbulent.
Lamina (noun)
• a thin layer, plate, or scale of sedimentary rock, organic tissue, or other material.
3. Laminar flow
Development of parabolic velocity profile for a circular pipe

consider a small cylinder of fluid:
τ r
P r P 2a
x
dx
P P+dP
P ⋅ π r 2 +τ ⋅ (2π r ⋅ dx) = (P + dP) ⋅π r 2
2πr ⋅τ ⋅ dx = πr 2 ⋅ dP
r dP
τ= ⋅ OK, since both τ and dP/dx are negative
2 dx
using equation 1.4 to eliminate τ :
du r dP
µ = ⋅
dr 2 dx
integrating:
u r
dP r 1  dP  2 2
µ ∫ du = ∫ ⋅ dr u= −  a − r
4µ  dx 
( )
0
dx a 2
3. Laminar flow
maximum velocity occurs a2  dP  D2  dP 

at center line (r=0) umax=  −  = − 
4µ  dx  16µ  dx 
the volumetric flow rate can be calculated as follows:
 1  dP 
) ⋅ (2πr ⋅ dr)
a a
Q = ∫ u ⋅ dA = ∫   −  a2 − r 2
4µ  dx 
(
0 0 
a
π  dP 
∫ [(a r − r )]⋅ dr
2 3
Q = − 
2µ  dx  0
a
π  dP  a2r 2 r 4  πa4  dP  πD4  dP  Hagen-Poiseuille
Q =  −  −  = −  = −  equation
2µ  dx  2 4 0 8µ  dx  128µ  dx 
the mean velocity can be calculated as follows:
Q D2  dP  u=
umax
u= = − 
( )
πD2 4 32µ  dx  2
for a pipe of length ‘L’ the pressure drop can be calculated in terms
of the mean velocity as follows:
 dP  ∆P 128µQ 128µ LQ
−  = = ∆P =
 dx  L πD4 π D4
32µ L ⋅ u ∆P ∝ u
and replacing Q with the average velocity: ∆P =
D2
3. Laminar flow
Flow of a viscous liquid down a vertical plate

consider a small element of fluid (unit depth into page):
fluid reservoir
free fluid
entrance surface
region u(x,y)
Practical application: wetted-

u(y) wall column for measuring mass
y
x transfer coefficients
steady
state
laminar
region
u≠f(x)
dp/dy=0
y
y=h
3. Laminar flow
Rotating cylinder viscometer – another look….

View looking down into apparatus, which measures
the torque T required to turn the cylinder of length
L at various angular velocities Ω :
a1
Ω
a2
compare this result to the earlier
torque balance at radius r (a1<r<a2) ….. analysis, in which the cylinders were
T + τ ⋅ [ 2π rL ] .r = 0 a1 < r < a2 −T approximated as flat plates:
τ=
2π r 2 L
force x distance T  2 ( a2 − a1 ) 
µ=  
4πΩL  a13 
rotational version of constitutive equation (rω is the local angular velocity):
 dω  16
τ = µ ⋅ γ& = µ  r  not an easy derivation…… 14
 dr 
Error (% over-read)
12
eliminating τ and using the no-slip condition at both cylindrical 10
walls: 8
6
Ω
− T a1 dr T 1 1 4
∫ dω =
0
2πµL ∫a2 r 3
µ=  − 
4πΩL  a12 a22  2
0
0.00 0.02 0.04 0.06 0.08 0.10
for Newtonian fluids µ ≠ f(T, Ω) (a2-a1)/a1
4. Turbulent flow
Flow around a sphere

This situation can described using dimensional analysis (DA).
The most common use of DA in fluid mechanics is the testing of models,
which are made to be geometrically similar to the full-scale object:
• scale models of cars, aircraft, buildings in wind tunnels (to avoid the cost
of a full-scale prototype and and perhaps an impracticably large wind
tunnel)
• scale models of ships, yachts, oil rigs in test ponds (same reasons)
• objects tested in a different fluid to that which they will encounter in
service, for convenience of experimental set-up (e.g. testing of flow of air
around an object instead of molten salt)
Consider the flow of air or water around a sphere, and the associated drag
force, F .
Dimensionless analysis suggests :
F  ρuD 
= f  
ρu2 D 2  µ 
The convention is to define a drag coefficient, cD
F
cD =
 πD  1 2 
2
  ρu 
 4  2 
For Re > 2 we can determine the functional relationship cD = f(Re)
experimentally – often by measuring ut , the terminal velocity of the sphere
falling through a liquid.
4. Turbulent flow
Drag coefficient for a smooth sphere 1 Re = 0.1

1000
Attached flow
1 (Stokes flow) or
steady separated
100
flow at slightly
higher Re
10 Re = 57
2 Separated unsteady
flow, laminar
CD
2
boundary layer
1
3 upstream of
separation, vortex
4 street downstream
0.1
Re = 15,000
3
Separated unsteady
flow, laminar
0.01
boundary layer
1.00E-01 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1.00E+07
upstream of
Re
separation,
turbulent wake
Re < 2 cD =24/Re downstream
4 Re = 800,000
2 < Re < 103 cD =(24/Re) × (1 + 0.15 Re0.687)
Post-critical
separated flow,
103 < Re < 105 cD = 0.44
with turbulent
boundary layer
Re > 105 cD = 0.1
4. Turbulent flow
The terminal velocity of a sphere can also be predicted using another

dimensionless group, ND which is defined at terminal velocity conditions:
2
2 4 gD(ρ S − ρ F )  ρ F ut D  4 gD3 ρ F (ρ S − ρ F ) 4
N D = cD(t ) Re (t) ND =   = = Ga
3ρ F ut2  µ  3 3µ 2 3
The Galileo number (Ga) is a sometimes referred to as the Archimedes number (Ar).
Example #1
Find the terminal velocity in water of a 1 mm diameter glass
sphere (density 2650 kg/m3) at ambient conditions.
Drag coefficient for a smooth sphere
1000
cD
100
10
CD
0.1
0.01
1.00E-01 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1.00E+07
Re
4. Turbulent flow
Example #2
A well-struck golf ball (43 mm diameter) travels
at 50 m/s initially, then decelerates through most
of its flight to land at a final velocity of 15 m/s.
15 m/s
50 m/s
Consider a laboratory bench-top experiment that will predict the

drag force (Fair) on a golf ball travelling through air at 50 m/s, using
laboratory tests carried out on the golf ball moving through water.
4. Turbulent flow
Flow in pipes Force balance on element of fluid in the pipe:

τw
Consideration of incompressible flow in a smooth pipe suggests y
that the following parameters are relevant: P+ΔP u l P
x
∆P = P1 − P2 = f (u, L, D, ρ , µ )
M L M M 
2
= f  , L, L, 3 , 
LT T L LT 
P1 u P2 D
L
Dimensional analysis with (6-3) = 3 groups and choosing one
group as the Reynolds number (Re), since this is a flow system:
∆P  ρuD L 
= f  , 
ρu2  µ D
Experimental results show that ∆P α L/D, so we can simplify

this to:
∆P L  ρuD 
= ⋅ f 
ρu 2 D  µ 
Experimental data is often plotted as a Moody diagram –
which covers both turbulent and laminar flow regimes.
4. Turbulent flow
In real pipework systems, the wall roughness

affects the pressure drop for turbulent flow . The
rougher the pipe, the more friction that occurs –
refer to page 18 for typical values.
The effect of surface roughness is predicted by
the following correlation:
  ε 1000 000  
1/3

c f = 0.00138 1+ 20 000 +  
  D Re 
Approximately how much pressure drop can we
expect along a pipeline with turbulent flow?
The laminar region of this diagram can be predicted theoretically,
using the Hagen-Poiseuille equation:
32µ L ⋅ u 4lτ w τ
∆P = together with these identities: ∆P = cf = w
D2 D 1 2
ρu
16 2
and eliminating τw and ∆P: c f =
Re
The transition between laminar and turbulent flow causes a
discontinuity at Re = 2,000 – 3,000. The amount of friction
experienced increases markedly when the transition to turbulent
flow occurs. The turbulent region of this diagram must be
measured experimentally.
For smooth pipes, with 2200 < Re < 500 000 , the Blasius correlation
is used:  1 
0.25
c f = 0.079  
 Re 
4. Turbulent flow
Example #3
Flow of water down a smooth 6” pipe from a large reservoir,
discharging to atmosphere
H1 = 50 m
reservoir
D=150 mm
H2 = 0 m
L=3 km u
discharge
4. Turbulent flow
Losses in pipe fittings

Pipe fittings (such as bends and T-junctions) cause loss of
pressure in fluid flowing through them. So do control
valves, non-return valves, orifice plates and
expansion/contractions in a pipe.
The diameter of chemical plant pipework is usually
determined by economic considerations – a trade-off
between capital cost and subsequent operating cost. Small
pipes are less expensive to install, but incur higher
pressure drop costs (e.g. pump or compressor electricity
consumption) for the life of the plant.
Typical installations have u≈2 m/s for liquids and u≈30 m/s
for gases, both of which dictate turbulent flow.
Losses associated with various fitting and valves have been

measured and tabulated.
They are often reported as either ‘K’ (number of velocity
heads) or ‘equivalent pipe length’.
4. Turbulent flow
Example #4
Find the flow of water than can be expected through a
pipe of length 20 m and diameter 0.01 m containing 8
standard right-angle bends and a ¾-open gate valve, all
operating with a head difference across the system of 8 m.
Entry to the pipe is smooth (no losses), but discharge is
into a large tank.
Using Fanning friction equation:
4. Turbulent flow
Pipe networks
Chemical plants often have complex, lengthy pipework
systems for utilities (e.g. cooling water, steam + condensate
return, pump gland seal water, instrument air), which are
provided at a central location and then distributed throughout
an entire manufacturing site.
Example #5
Consider a garden hosepipe (diameter D) lying on a horizontal
lawn, connected to a tap with mains water pressure ∆Hsupply ,
and L1 L2 L3 >> 50 D :
L2 C
L1 u2
∆Hsupply
A u1 B
u3
L3
D
4. Turbulent flow
Example #6
Consider the flow from two large reservoirs at different elevations
with L1 = L2 = L3 = 200 m, D = 0.03 m and cf = 0.005 (constant) :
A
B
u1
u2
HA = 30 m L1
L2 HB = 20 m
u3
D L3 C HC
The first issue to resolve is the direction of flow in pipe B-C . What
value of the height HB would result in exactly no flow through pipe
B-C? Compare this to the actual height of 20 m.
• No flow in B-C implies u1 = u3 and since L1 = L3
∆HL(AC) = ∆HL(CD)
• Bernoulli’s equation A – D (ignoring KE terms) with PA=PD=0
∆HL(AC) + ∆HL(CD) = 2∆HL(CD) = 30 m hence ∆HL(CD) = 15 m
• Bernoulli’s equation B-C (since no friction losses at no flow)
PC = (HB-HC)ρg
• Bernoulli’s equation C-D
PC = (∆HL(CD) -HC)ρg
Solve these equations for HB = 15m ﬂow is B→C
4. Turbulent flow
L1 u12 PC
H C = H A − 4c f − β = 30 − 6.8u12 mass balance at C :
D 2g ρ g u1 =1.27m / s
L 4 × 0.005× 200 u3 = u1 + u2
2
L2 u PC 4c f = = 6.8
H C = H B − 4c f 2
− 2gD 0.03× 2 × 9.81 β = 20 − 6.8u22 u2 = 0.36 m / s
D 2g ρ g
 P  30 − β + 20 − β = β u3 =1.63m / s
4c f L3u32 PC β =  HC + C  2
HC = −  ρg  β = 6.8u 3
2g D ρ g β =19.1m ( by iteration)
Example #7
A ring main
L1 = L2 = L3 = 100 m, D = 0.04 m and cf = 0.005 (constant) :
3 L/s
Q B 3-Q
5 L/s A C 1 L/s
5-Q D
4-Q
1 L/s
5. Pumps
Overview
There are many different designs of pump, each of which is used because it is
well-suited to a particular duty.
Considerations are:
• flow-rate (high or low)
• pressure increase (high or low)
• NPSH available (limited, not an issue)
• fluid/slurry to be pumped (corrosive, abrasive, toxic, flammable, expensive)
• operating temperature and pressure
• control response requirement (fast, not an issue)
• duty cycle (continuous, intermittent)
Pumps are broadly divided into two categories, which have very different
operating characteristics:
• positive displacement
• centrifugal
Positive displacement pumps must be used with care. If there is any situation
in which such a pump could operate against a closed discharge (e.g. if a
control valve in the discharge line closes fully) the pump will be destroyed
almost immediately.
A centrifugal pump can tolerate this condition for a period of time (perhaps
several minutes), until the fluid trapped in the pump casing boils due to the
work input from the pump.
Positive displacement pumps can, if required, generate much higher
pressures than can be achieved with centrifugal pumps, and are also the only
choice for vacuum pumps (e.g. as needed for an evaporator).
5. Pumps
Typical centrifugal pump/blower

installations
5. Pumps
Centrifugal pumps
Fluid/slurry enters the centre of a rotating impeller and is accelerated by
centrifugal force. It then passes to the volute chamber where it decelerates,
and its kinetic energy is converted into pressure at the pump discharge.
Centrifugal pumps are generally driven by an AC motor.
Dimensional analysis can be applied to centrifugal pumps that are of the
same shape, i.e. geometrically similar.
Parameters expected to be relevant:
∆P pressure rise through pump N/m2
Q volumetric flow rate m3/s
D impeller diameter m
N rotational speed of impeller revs/s i.e. s-1 ∆P = f ( D, N,Q, ρ, µ )
ρ fluid density kg/m3
µ fluid viscosity Ns/m2
6 parameters, 3 dimensions so (6-3) = 3 dimensionless groups required

Choose (N×D) as characteristic velocity (m/s)
∆P Hg
∆P  Q ρ ND 2  CH = head coefficient CH = ≡ 2 2
= f ,  ρ (N D ) (N D )
2 2
ρ ( N 2 D 2 )  ND 3 µ  Q
CQ = flow coefficient CQ =
CH = f (CQ , Re) ND3
Performance tests show that there is no dependence on Re when Re > 105

Most operating conditions of practical interest have Re > 105, so we can
simplify the relationship to: CH = f (CQ )
5. Pumps
Experimentally determined
The power required to drive a centrifugal pump (input power) can be characterised pump characteristics
using an additional dimensionless group, CP – the power coefficient:
P
P  Q  ρ ( N 3D5 )
CP = = f   power
ρ ( N 3 D 5 )  ND3 
This relationship does not include mechanical losses (such as bearing friction), or
the electrical efficiency of the motor.
 Q 
The power transferred to the fluid (PPV = rate of PV work) is always less than the  
 ND 3 
∆P
actual power required to drive the pump, P. head
ρ ( N 2 D2 )
PPV = ∆P × Q = ρ gHQ
Hence an efficiency can be defined:
output power PPV ρ gHQ CH CQ

η= = = =  Q 
input power P P CP  
 ND3 
The selection of design of centrifugal pump type with the highest efficiency
efficiency for a particular duty is made possible by consideration of a η
further dimensionless group, the specific speed , NS :
 Q 
1/2
1/2  3
NS =
(C )
Q
=
 ND 
=
NQ1/2
 Q 
 gH 
3/4 3/4 3/4
( CP )  2 2
( gH )  
 ND 3 
N D 
5. Pumps
Centrifugal pump types Umkomaas, KwaZulu Natal

South Africa
March 1984
NS 0.06 - 0.20 NS 0.20 – 0.60 NS 0.60 – 1.00

Example – centrifugal pump characterisation tests
The following test data were collected for pump ‘A’ (D=0.15 m) running
at a speed of N = 1450 rpm. The test fluid was water (ρ = 1000 kg/m3
µ = 0.001 Pa s).
Q (litre/s) 0.000 2.83 6.08 7.99 9.75 10.77
H (m) 7.086 7.19 6.87 6.19 5.45 4.86
P
P (kW)
(kW) -- 0.408
0.408 0.533
0.533 0.607
0.607 0.677
0.677 0.695
0.695
CCQQ - 0.035 0.075 0.098 0.120 0.132
CCHH 5.29 5.37 5.13 4.62 4.07 3.63
CPP - 0.380 0.497 0.566 0.632 0.648
NS - 0.053 0.080 0.099 0.121 0.138
η (%) - 49 77 80 77 74
5. Pumps
8 90 8 90
7 80 7 80
6 70 6 70
60 60
efficiency, η
efficiency, η
5 5
H, (m)
50 50
H,(m)
4 4
40 40
3 3
30 30
2 20 2 20
1 10 1 10 NS(η max)=0.097
0 0 0 0
0 2 4 6 8 10 12 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
Q (litre/s)
NS
Pump ‘B’ (D=0.20 m) is geometrically similar to pump A, and will be
used at 1750 rpm to transfer oil ( ρ = 800 kg/m3 µ= 0.004 Pa s). 20 4.5
Plot the predicted performance curves for this pump. 18 4
16 3.5
Input power, P (kW)

14
3
QB N B DB3 1750 × 0.2 3
Haed, H (m)
12
CQ( pumpA) = CQ( pumpB) = = = 2.86 2.5
QA N A DA3 1450 × 0.153 10
2
8
H B N B2 DB2 1750 2 × 0.2 2 6
1.5
CH ( pumpA) = CH ( pumpB) = = = 2.59 1
H A N A2 DA2 1450 2 × 0.152 4
2 0.5
3 5 3 5
PB ρ B N D 800 ×1750 × 0.2 0 0
CP( pumpA) = CP( pumpB) = = B B
= 5.93 0 5 10 15 20 25 30 35
3
A
5
PA ρ A N D 1000 ×14503 × 0.155
A Flow rate, Q litres/s
5. Pumps
Interaction between centrifugal pumps and pipework systems

This can be predicted using the know performance characteristics
of the pump and the pipework system. For example:
h = 10
(m)
pipe diameter D = 0.08 (m)
pipe equivalent length
(including all fittings, bends
25 7
oil etc.)
LE = 30 (m)
pump ‘B’ 20 6
Input power, P (kW)

If the pump performance characteristic is available as an algebraic
Haed, H (m)
15 5
correlation, then a graphical method can be avoided.
Two strategies to control a fluid flow rate:
10 4
a) adjust a valve on the discharge line (change the system
characteristic)
b) vary the speed of the pump / compressor (change the pump 5 3
characteristic)
0 2
H (m) H (m) 0 5 10 15 20 25 30 35
Flow rate, Q litres/s
Q (m 3 / s) Q (m 3 / s)
5. Pumps
Bends 2” NB sch 10 ID 54.8 mm 2.0 m/s
12 m 37 m
LIC LT
3m 105
2 105
L 1000× 2
LV
[
∆Ppipework = 1000 (3 + 2) 9.81+ 4 0.079Re−0.25 ]0.0548
E
2
Water 105
2m
90°C 1/2 open ∆Ppipework = (29 + 19) + (8 + 7 + 23) + 8 + 5 + 2 = 101kPa
psat= 70 kPa Elevation Pipe Control Bends Outlet to
13 m change friction valve header tank El. 1753 m
125 207
100 182
75 157
Pipeline
50 132 pressure
Pipeline (kPa abs)
pressure 25 107
(kPag)
82
-25 57
-50 32
-75 7
(1000)(2.0)(0.0538) = 108000 LE ρu 2 LV LT
Re = ∆Ppipework = hρg + 4c f 105 105
0.001 D 2
LE = 13 + 12 + 37 + [225+ (4 × 38) + 50] × 0.0548 LIC

105
5. Pumps
Net positive suction head (NPSH) considerations when pumping liquids

Bernoulli’s equation tells us that as a liquid is accelerated by a pump’s
impeller, its velocity increases and therefore its pressure must decrease.
It is therefore possible for the local pressure within a pump casing to fall to
below the vapour pressure of the liquid being pumped (psat), causing the
liquid to boil.
Cavitation vapour bubbles are formed, which collapse again a fraction of a
second later as they reach the region of higher pressure in the pump outlet
(volute). The collapse of the bubbles is highly undesirable as it causes noise,
loss of efficiency, and pitting / erosion of the impeller blades - eventually
leading to blade breakage.
The Net Positive Suction Head (NPSH) is defined as the difference between a
pump's inlet stagnation pressure and the vapour pressure of the liquid
(which is a strong function of temperature). It is usually expressed in m.
The stagnation pressure head is the notional head that the fluid would have
if it were brought to rest. It is calculated from the Bernoulli equation, hence
for a particular set of operating conditions, the available NPSH is given by:
 Pinlet uinlet2  psat
NPSH( avail) =  + −
 ρg 2g  ρg
The minimum NPSH required to avoid cavitation is a characteristic of a
particular pump design, not the fluid it is pumping. Pump manufacturers
provide NPSH(req) data as part of the performance characteristics of a pump.
5. Pumps
Pump priming
Consider a swimming pool pump that can generate
Before start-up, a liquid pump and its associated
∆Pmax= 3 bar against a closed outlet.
pipework may contain air.
Either:
∆P
• The pump must be capable of initially pumping the
ρ ( N 2 D2 )
air out of the system, and replacing it with liquid
• The pump must be primed by manually filling the
whole system with liquid before the pump is
3 bar
started.
When priming is not possible, or inconvenient a self-  Q 
  Sand
priming pump design must be used. Swimming pool  ND3 
filter
pumps are usually self-priming.
0.3 m
6. Flow through packed beds, filters and fluidized beds
Packed beds
Packed beds consist of vessel partially filled with a tightly packed bed of uniformly-sized
particles or packing-pieces.
• Particles are usually spherical or cylindrical, depending on the application, and can range
from 1-20 mm in diameter
• Packing-pieces come in a range of designs (Pall rings, Raschig rings, Intalox saddles etc.)
and range from 10 - 80 mm in diameter
Packed beds can be used in a wide variety of applications:

• Adsorption columns – water removal, air separation, (bio)chemical purifications, kitchen
hob extraction hoods
• Catalytic reactions – hydrocracking of crude oils, ammonia synthesis, water gas shift for
syngas enhancement, converters in car exhausts
• Ion exchange – water purification
An important characteristic of a packed bed is ε its void fraction (also called voidage or
porosity).
If the particles (or packing pieces) are arranged randomly, and are small compared to the
diameter of the vessel, we can assume that every cross-section through the height, l, of the
packing has the same area, Af , available for fluid flow:
Pressure drop is an important consideration when designing a packed bed. Ancillary equipment
(pumps, compressors) must be specified to provide the required ∆P, and the energy input
associated with this may significantly affect the operating cost of the process.
In 1846, Henry Philibert Gaspard Darcy (a municipal water supply engineer in Dijon, France)
studied the flow of water trickling through packed filter beds (i.e. low values of Re). He devised
Darcy's Law, an empirical equation that relates the superficial velocity of flow through a bed (u) to
the pressure gradient through the bed (dP/dl) and the viscosity of the fluid (µ) :
K dP
u= ⋅ K = permeability of packed bed
µ dl
µ ul
hence it is predicted that: ∆P = In case of non-
K
spherical particles:
The Carman-Kozeny equation is a correlation (i.e. based on experimental results) that describes
laminar flow through a packed bed of particles, diameter d :
volume Vp
µul (1 − ε ) 2
dP µu (1 − ε ) 2
surface area Ap,
∆P = 180 ⋅ = 180 2 ⋅
d2 ε3 dl d ε3
d 2ε 3 use the volume-
hence K in Darcy’s law is given by : K =
180(1 − ε )
2
surface diameter:
Vp
The physical properties of gases will change as they flow through a packed bed. As the pressure, P d vs = 6
Ap
decreases along the length of a packed bed, the volume of the gas must increase, and hence the
gas velocity increases along the bed. Changes in temperature along the bed (perhaps due to a heat
of adsorption or reaction) must also be accounted for.
1.0E+08
Sabri Ergun (1952) fitted experimental data in both laminar and

1.0E+07
turbulent flow through packed beds of spheres of diameter d with :
2
∆P µ u (1− ε ) ρ u2 (1− ε ) 1.0E+06
=150 2 ⋅ 3 +1.75 ⋅ 3
L d ε d ε ∆pρd 3  ε 
3
1.0E+05
 
laminar term turbulent term µ 2L 1− ε 
1.0E+04
In dimensionless form : 1.0E+03 CK

Ergun
∆P  ρd 3  ε  
3 2
ρud  ρud  1.0E+02
    = 150 + 1.75 
L  µ 2  1 − ε   (1− ε )µ  (1 − ε )µ  1.0E+01
1.0E-01 1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04
ρud
and defining a modified Reynolds number Re’ : Re' = Re'
3
(1− ε ) µ
∆Pρd  ε 
3
 = 150(Re') + 1.75(Re')
2
2 
µ L  1− ε 
Example:
Water flows under gravity through a 0.5 m deep bed of spherical
sand particles of diameter 0.7 mm and void fraction ε = 0.40.
Estimate the water superficial velocity through the bed when the
liquid surface is 1.0 m above the upper surface of the bed.
1
1.0 m
0.5 m
h=0 2
Re is a dimensionless ratio of forces: =
L= characteristic dimension
Consider a packed bed of void faction ɛ that is 1.0 m high, with cross sectional area 1.0 m2
containing n particles of diameter d m with fluid approaching at velocity u m/s for which
we define a modified Reynolds number Re’:
• for calculation of the inertial forces (rate of change of momentum) the appropriate
velocity (v) to use in this ratio should be the interstitial velocity, ui
• for calculation of the viscous forces due to shear stress at the fluid/particle interfaces
the appropriate area (A) to use in this ratio should be the interfacial area Ai
• The appropriate characteristic dimension, L, should be the hydraulic diameter DH

based on the fluid volume Vvoids and a wetted perimeter that is the interfacial area Ai
By convention the factor of 2/3 is dropped:
Example:
Isothermal catalytic reactor removing traces of
pollutants from combustion gases
d = 5 mm
L = 3.0 m ε = 0.40
u = 0.5 m/s
µ = 0.001 Pa.s
Pin = 2 bar(abs)
M = 30 g/mol
T = 300 °C
P u
2.0
0.8
1.25 0.5
z z
0 3.0 0 3.0
Filtration
In industrial batch filtration, a slurry flows towards an array of fine cloth panels which are
supported by perforated steel plates. The cloth is selected for a particular duty to allow the
liquid to pass through, whilst retaining the solids. The depth of the retained solids increases
with time, until eventually ∆P becomes prohibitively high. The filter must then be taken off-
line to remove the filter cake, after which it is reassembled - ready to be used again.
t=0 t=t
slurry εs V/A
L L-(V/A)-l filter cake εc
l
filtrate (pure liquid)
filter cloth volume V
Area A
A filter press is usually supplied with slurry by a centrifugal pump, or by gravity from a
header tank.
The control system can be configured in one of three ways:
• the flow rate of incoming slurry is measured (perhaps by using an orifice plate to
measure the flow of outgoing filtrate) and kept at a constant value. As the filter cake
increases in thickness the pump must work harder to overcome the increasing pressure
drop. This can be achieved by gradually increasing the speed of the pump. When the
maximum speed of the pump is reached, the filtration cycle is terminated.
• a constant pressure is applied to the filter, for example from a header tank. The filtration
cycle is terminated when the flow rate of filtrate falls below a certain value.
• the filter is supplied by centrifugal pump operating at a fixed speed. As the pressure drop
increases across the filter cake, the flow rate reduces in accordance with the pump
characteristic curve. The filtration cycle is terminated when the pressure drop reaches
the capability of the pump, or when the filtrate flow rate falls below a certain value.
∆P ∝ u⋅ (l + lm ) lm = “effective length” of filter medium (equivalent depth of cake)
The filter medium itself (the cloth) has a pressure drop across it which must be taken into
account.
Filter cakes are often compressible, meaning that the cake void fraction reduces with
time, as the pressure drop across the cake increases.
l expected at constant flow rate
total blockage due

actual to compressed cake
t
Filter cartridge calculation assuming laminar flow, Carman-Kozeny equation:

Old engine oil is left in a drum to drain through dQ
a filter by gravity. How long will it be until the
level of engine oil in the drum falls to the top of
the filter cartridge?
D
ε = 0.3
ro= 35 mm
ri = 20 mm
d = 0.04 mm
D = 0.4 m
h = 10 cm
H0 = 1.0 m
µ= 0.06 Pa.s
ro r
H0 i ρ = 800 kg/m3
h
z
Batch plate filtration at constant pressure – optimum cycle time
cake
slurry
filtrate
1. slurry is fed to the filter

tF
for an extended period…….
time-averaged production rate of cake: …..then, as quickly as possible……
2. the filter cake is washed
(optional)
3. the filter cake is dried by blowing
to find maximum value of W, differentiate with respect to t and set = 0: tD with air (optional)
4. the filter is dismantled, the cake
removed, and the filter re-
assembled
5. The filtrate is initially recycled
tF
∆P(kPa)
t=0
Q (m 3 / s)
Fluidized beds
Fluidization is the result of an upward flow of fluid through a bed of particles
which is supported on a flow distributor plate and has an unconstrained upper
surface.
As u increases the pressure drop across bed
due to friction, ∆P , increases (Carman-Kozeny
particles or the Ergun equation).
h This results in an increasing upward drag force
∆P on the particles.
When u reaches umf the minimum fluidizing
velocity, the particles become suspended in
the fluid, and the bed is said to be fluidized.
u
h As u is increased further:
• the bed starts to expand, i.e. h increases
• the bed voidage ε increases.
These increases continue until the height of
the bed reaches that of its container (vessel)
and thereafter particles are swept out of the
u bed in the fluid flow; this is called elutriation.
u mf
The behaviour of a fluidised bed can be investigated by gradually increasing the

fluid velocity, u, until the bed becomes fluidised, then gradually reducing it.
1. insufficient flow and pressure drop to fluidize the bed
2. fluidization occurs
3. the bed expands – ε and h increase with u
4. the bed settles at εmf and hmf
umf can be estimated using the Ergun equation.

ρF umf d Gas-solid systems
First define Remf Re mf =
µ • Reaction between the gas phase and the particles.
Particles swirl around in the bed creating excellent
at minimum fluidising conditions: ∆Pmf = (1− ε ) ⋅ ( ρS − ρ F ) ghmf
mixing and good heat/mass transfer. Examples are
catalytic reactors, combustion systems (including
Ergun equation at same conditions:
pulverised coal particles for power stations).
∆Pmf µumf (1 − ε )2 ρF umf 2 (1 − ε ) • Drying of particulate solids, such as sewage pellets.
= 150 2 ⋅ + 1.75 ⋅ 3 hence:
hmf d ε3 d ε
Liquid-solid systems
(1 − ε ) ⋅ (ρS − ρF )ghmf µumf (1 − ε ) 2 2
ρF umf (1 − ε ) • Segregation / separation of particles by size
= 150 ⋅ + 1.75 ⋅ 3 (classification – big particles migrate to bottom of
hmf d2 ε3 d ε
bed, small migrate to top).
ρ F ( ρS − ρ F ) gd 3
=150
(1− εmf ) Re +1.75 1 Re2 • Bio-separation - expanded beds (with minimal
mf mf
µ2 ε mf3 ε mf3 mixing) allow adsorption of biomolecules in the
presence of broken cell material, eliminating need for
LHS is the Galileo number, Ga Assuming εmf = 0.4
feedstock centrifugation/ filtration)
(a typical value)
• Waste water treatment.
(
Re mf = 27.8 1+ 5.52 ×10−5 Ga −1 )
27.3Re 2mf +1406 Re mf − Ga = 0
Correlations based on experimental data are often used:

Wen and Yu: (
Re mf = 33.7 1+ 3.59 ×10 −5 Ga −1 )
Leva: Re mf = 0.0011⋅Ga 0.94
A vertical cylindrical reactor vessel contains a bed of

spherical catalyst pellets, diameter d = 0.50 mm,
density ρS = 1800 kg/m3. In pilot plant trials, when the
dilute aqueous solution of reactants enters the base
of the vessel at 20°C, incipient fluidization occurs at a
velocity u = 0.0016 m/s and bed height l =2.0 m.
What is the void fraction at incipient fluidisation?

What is the pressure drop across the reactor vessel?
particles
h
∆P
There are several types of behaviour exhibited by fluidized beds when u > umf
Liquid-solid systems
These usually expand uniformly - this is particulate fluidization.
Richardson and Zaki correlated their experimental results:
u n ut = terminal velocity of a single particle in a large

=ε volume of fluid
ut u = superficial velocity of fluid
ε = bed voidage
n = exponent that is determined experimentally
ρ F ut d with d = diameter of a sphere with same
and defining Ret =
µ volume as a particle
d
λ= particle-to-vessel diameter ratio
D
Ret < 0.2 n = 4.65+19.5λ

0.2 < Ret <1 n = [ 4.35+17.5λ ] Re−0.03
t
1 < Ret < 200 n = [ 4.45+18.0 λ ] Re−0.1

t
200 < Ret < 500 n = 4.45Re−0.1

t
500 < Ret n = 2.39

Gas-solid systems
The behaviour of these systems for u > umf depends on
the characteristics of the particles.
• Particulate fluidisation (as described for liquid-solid
systems) is observed in very few situations.
• In aggregative or bubbling fluidization, bubbles
(particle-free zones) formed at the gas distributor rise
through the bed whilst coalescing, as seen with gas
bubbles rising through a liquid. There is good mixing
between gas and particles.
• In slugging fluidization the bubbles become very large
and are of comparable size to the diameter of the bed.
• In channelling fluidization, often observes with
fine/cohesive particles, gas passes through channels
and cracks in the bed.
• In jetting fluidisation, long vertical jets of gas move up
through the bed.
• In spouting fluidization, often observe when the
particles are large, the gas may pass through the bed in
a central channel.
Each of these flow regimes may have negative
consequences for chemical reactions and/or heat transfer
in the bed, as a result of gas trapped in the bubbles
having little contact with the solid particles. “Bypassing"
of gas trapped in the bubbles is the subject of
considerable continued research.
Geldart proposed a
method of classification
for air fluidised beds,
based on the particle
diameter and the
density difference
between the solid and
fluid phase.
7. Liquid Mixing
Mixing is one of the most common unit operations in chemical

engineering. Reasons for mixing (which improves homogeneity)
include:
• Consistency of product quality
• Improvement of heat transfer (overall rate, avoidance of ‘cold
spots’ and ‘hot spots’ which may affect the product quality)
• Improvement of reaction rates, by increased contact between
reagents
• Dissolution of solids or gases
Single phase liquid mixing

e.g. blending of petroleum fractions of differing viscosities to
make specialist lubricants and fuels
Immiscible liquid mixing
One phase disperses as small droplets in the other. Liquid/Liquid
extraction is a process consisting of successive mixing and settling
stages, where a large interfacial area between phases is
important to promote rapid mass transfer.
Very fine droplets can form semi-stable emulsions, as used in
brewing and pharmaceutical processes.
Gas-liquid mixing
e.g. aerobic fermentation, waste water treatment
Solid-liquid mixing
e.g. slurries used extensively in the mining industry. Important to
promote reaction and prevent settling of solids.
7. Liquid Mixing
Mechanisms of mixing Laminar mixing

Usually associated with viscous liquids. (µ > 10 Ns/m2)
Two main requirements: Inertial forces imparted by the impeller dissipate
quickly, so the impeller must be designed to cover most
• Bulk, or convective flow to eliminate stagnant regions of the vessel cross-section in order to generate
(dead spots) sufficient bulk motion
Turbulent mixing
• One or more zones of intensive or high-shear mixing in
which inhomogeneities are broken down. Usually associated with viscous liquids. (µ < 10 Ns/m2)
The inertia imparted by the impeller is sufficient to cause
the liquid to circulate throughout the vessel and return to
Both processes consume energy, which is dissipated as heat.
the impeller.
Scale-up of stirred vessels
Often required, to migrate pilot plant scale trials to production N
P = pitch of agitator blade
scale. Important to achieve: W= width of agitator paddle
Geometric similarity exists if all major dimensional ratios are
consistent, e.g.:
Kinematic similarity exists if velocity ratios at all locations are H

consistent, and the flow patterns are alike
Dynamic similarity exists if the ratios of all forces at all WB D ZA
locations are consistent. Inertial, gravitational, viscous and
surface tension forces may all be contributing factors. DT
7. Liquid Mixing
It can be shown that a) achieving dynamic and geometric

Reynolds number
similarity ensures kinematic similarity b) the following
dimensionless numbers should be the same for similar systems:
Froude number
Mixing power consumption in stirred vessels
Low viscosity systems typically consist of a vertical cylindrical
vessel with H/DT ratio in the range 1.5-2, D/DT ratio ~0.3 and Weber number
10<N<25 Hz (note not RPM, but RPs !!)
Dimensional analysis using relevant parameters suggests: Power number
The most common form used in practice is a power law :

Where K’, b and c must be determined experimentally.
Note that this correlation does not account for losses in bearings
and/or gearbox, and for electrical motor efficiency. Fr can be neglected
unless there is gross
vortexing, or Re>300:
For Re <10, b is found experimentally to

be -1, so the shaft power, P, is given by:
7. Liquid Mixing
Example: First check dimensional ratios to confirm that it is valid

An aqueous solution of density 1350 kg/m3 and to use empirical data from chart on previous page:
viscosity 0.03 Ns/m2 is stirred by a mixer of 0.5 m empirical data this example
impeller diameter in a cylindrical pilot plant vessel of
diameter 2.4 m. The impeller is located 0.5 m above
the floor of the tank and the liquid depth is 2.1 m. It is
found that a rotational speed of 3 Hz is required to
achieve satisfactory mixing. What shaft power is
required?
What should the rotor speed be in order to achieve the

same degree of mixing in a geometrically similar full-
scale tank of 3X the diameter? What is the expected
power requirement?
In practice, one of two criteria are used for scale-up
calculations:
• Constant impeller tip speed (for miscible liquids,
when suspended solids are present, or if there is
heat transfer to a coil or jacket)
• Constant power input per unit volume (immiscible
liquids, emulsions and gas/liquid systems)
7. Liquid Mixing
Mixing rate – time required
Definition of ‘homogeneous’ or ‘well-enough mixed’ depends concentration concentration

c σ variance
on the particular process requirements.
Typically homogeneity is assessed by introducing a dye or c∞

tracer material at time t=0, then taking samples for analysis at
subsequent times. Concentration variance is then plotted σ t arget
against elapsed time. 0 t 0 t
trequired
Dimensionless analysis suggests:
ρND 2 DN 2
N ⋅ t required = f( , , dimensional ratios)
µ g
Mixing equipment
The wide range of mixing system configurations that are

commercially available reflects the many different process
requirements encountered in industry.
In this course we will review briefly only mechanical agitation.

Static mixers and extruders are described in the recommended
texts for this course.
7. Liquid Mixing
Mixing vessels are typically vertical cylindrical tanks, up to 10m in diameter,

with H/DT~1. Mixer motor and gearbox mounted on structural support
beams across top of tank.
Some gas/liquid systems have multiple impellers and H/DT~3.
Baffles are used to prevent gross vortexing, which is detrimental to mixing
efficiency.
Impeller design is mainly determined by liquid viscosity:
Propeller
Turbine
Increasing
viscosity
Paddle
Anchor
Helical ribbon
Helical screw
Course Summary
Introduction to fluid Solid/fluid
Chemical systems systems
Engineering
Dimensional Conservation Force Archimedes Force Conservation

analysis laws balances Principle balances laws
Mass Energy Momentum No shear Mass

(continuity (Bernoulli (momentum With shear (continuity
equation) equation) equation) - viscosity equation)
Laminar flow Flow Flow Flow in Flow

vs. between down a cylindrical around a
turbulent flow flat plates flat plate pipes sphere
Hagen-Poiseuille Fanning Blasius Carmen Ergun Richardson

equation equation correlation Kozeny correlation Zaki
correlation correlation
Mixers Centrifugal Pipework Orifice Pitot Venturi DP
pumps systems plate tube meter cell Filters Packed Fluidized
beds beds
Pump / pipework Flow Level
interactions measurement measurement

Fluid Mechanics Student Notebook

Uploaded by

Copyright:

Available Formats

Fluid Mechanics Student Notebook

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fluid Mechanics Student Notebook

Uploaded by

Copyright:

Available Formats

Department of Chemical Engineering and Biotechnology

These lecture notes are based, in part, on

Single-phase flow in tubes, pipes and ducts Mixing of fluids in vessels

Separation processes Multi-phase flow in pipes and ducts

Fluid mechanics is also used extensively in other disciplines:

Mechanical Engineering - pumping and power generation, aeronautics, turbine

The aims of this course in Fluid Mechanics are:

i. To introduce the relevant physical parameters and basic

iv. To introduce the method of dimensional analysis for

vi. To analyse the relationship between flow and pressure

vii. To predict the performance characteristics of packed

Chemical Engineering Chemical Engineering Fluid Mechanics and

Two (highly related) recent books are also very useful:

Thursday, November 11, 1999 So….

Mission controllers said that the "root

What are fluids?

…………………………………………………………………………………………………………. (1.2) Compressibility of water

• A liquid will exhibit a free surface if a larger space is

Forces and physical properties in fluid mechanics

Pressure, P, is defined as the normal force on a surface, per unit area

N/m2 is also abbreviated as Pa (Pascals)

The earth’s atmosphere is an example of the variation of pressure, P,

h=0 Surface of the earth

The pressure of a gas in a vessel can be measured using a U-tube manometer:

PA = PA' (A-A’ same height in liquid)

In Chemical Engineering applications we frequently encounter liquids

Capillary effect Selected values

These guys also know all about surface tension….

Φ is the contact angle or wetting angle.

Consider a solid object (of volume V, density ρs ) immersed in a fluid of density ρL :

Brick thrown out of a boat

What is relationship between H1 and H2?

Mballoon=350 kg Vballoon = 2800 m3

5 passengers, averaging 70 kg each

Drifting at 1000 m above sea level

Shear stress and viscosity

Newtonian and non-Newtonian fluids

• Most liquids of simple chemical structure behave in a Newtonian

Consider the general steady-state case for an element of Newtonian

2nd term results from pressure gradient

Now consider the case when both plates are stationary:

Look at velocity profiles within the fluid

laminar flow of a fluid flowing

laminar flow of a fluid flowing

laminar flow of a fluid flowing

laminar flow of a fluid flowing

Viscosity….some useful information

Consider the concentric cylinder device, length L …..

Laminar and turbulent flow

Professor Osborne Reynolds resolved

"An Experimental Investigation of the

Philosophical Transactions of the Royal Osborne Reynolds

Recritical depends on: Modern traffic

Steady flow of a low viscosity incompressible fluid

With the same assumptions, consider fluid flowing through a pipe of

Bernoulli's equation Conservation of mass within the stream

• low-viscosity fluid (effects of viscosity negligible)

Area of hole is greater than minimum area of fluid jet at (3),

How long will flow through the hole continue?