Sedimentation

A STUDY OF CHANNELLING BEHAVIOUR IN
BATCH SEDIMENTATION
A thesis submitted in fulfillment of the requirements for the

degree of Doctor of Philosophy
NILUFER KURT
B.Eng., M.Eng
School of Civil and Chemical Engineering

RMIT University
March 2006
DECLARATION
I certify that expect where due acknowledgement has been made, the work is that of the
author alone; the work has not been submitted previously, in whole or in part, to qualify for
any other academic award; the content of the thesis is the result of work which has been
carried out since the official commencement date of the approved research program; and any
editorial work, paid or unpaid, carried by a third party is acknowledged.
Nilufer KURT
20.03.2006
ii
ACKNOWLEDGMENTS
This thesis was funded in Scholarship for Women Engineers by RMIT University.
I would especially like to thank Dr. Anthony Stickland who helped me find a way in this
complex research topic with his navigational skills on the latest related theories and
experimental work.
I wish to thank Dr. John O’Donnell who was my supervisor for his able guidance, advice and
encouragement. I am grateful for his support and dedication in helping me in every possible
manner. I really admire my supervisor dedication towards the cause of the students who work
under him. I fell honoured to have worked with you.
I gratefully appreciate for his technical help and friendship of Senior Lecturer Mr. Phred
Petersen, Scientific and Industrial Photography. I would like to thank Dr. Sumanta Raha, Dr.
Deeptangshu Chaudhary, Dr. Ertan Moustafa, Khokan Kanti Majumder, Dr. Ertan Yesilnacar,
Nicola Hanitzsch and Ayze Koch for their help, support and encouragement at the different
stages of my study.
I would like to thank, my husband, Mr. Akilan Krishnamoorthy, who has been an inspiration
to me throughout this journey.
Also I would like to thank my fellowship postgraduate friends who were always there for me
during my study.
Finally, I wish to thank my family for their never-ending support and patience.
iii
KEYWORDS
Sedimentation, batch settling, hindered settling function, flocculated suspension, channelling,
stepped pressure filtration, sedimentation modes, floc, concentration gradients, compressible
sediment.
iv
PUBLICATIONS
Bayrak, N. (1996) Valuation of Zonguldak Fine Coals by Column Flotation Methods, Master
Thesis, Zonguldak Karaelmas University, Turkey, pp.99.
Toroğlu, I and I. Buzkan and N. Bayrak, (1996) “Concentration of waste gypsum by column
flotation”, 6th International Mineral Processing Symposium, Sept. 24-26, Kusadasi, Turkey,
pp.421.
Bayrak, N. and I. Toroğlu (1998) “Recovery of Armutcuk fine coal by column flotation unit”,
7th International Mineral Processing Symposium, Sept. 15-17, Istanbul, Turkey, pp.1.
Sütçü, H., N. Bayrak and I. Toroğlu (2001) “Yüksek küllü bitümlü toz atık kömürlerden temiz
kömür kazanımı”, II. Çevre ve Enerji Kongresi, Kasım 15-17, Istanbul, Turkey.
Bayrak, N., J. O’Donnell and I. Toroğlu (2002) “Recovery of fine coal by column flotation”,
9th APCChE Congress and CHEMECA 2002, Sept. 29-Oct 3, Christchurch, New Zealand,
pp.107.
Bayrak, N and J. O’Donnell (2004) “Principles of channelling in batch sedimentation”,

Proceedings of the Xth International Mineral Processing Symposium, Oct. 5-7, Cesme,
Turkey, pp. 587
O’Donnell, J. and N. Bayrak (2004) “Channelling in batch sedimentation of calcium

carbonate slurries” Chemeca 2004 Conference, Sept. 26-29, Sydney, Australia, pp:177
v
ABSTRACT
Batch sedimentation is a method that enables us to understand the mechanism of compaction
and compression of sedimenting slurry. However, batch settling behaviour is a very complex
phenomenon that is not easily described fully by a mathematical model. This causes
unrealistically large empirical calculations when the thickener size estimations are required.
Channelling, reverse concentration gradients and the initial concentration of the slurry have
large effects on batch settling. Existing procedures do not provide clear relationships
involving these three significant variables. In this study, batch sedimentation phenomena are
examined in detail and possible explanations are given to clarify the complex behaviour using
recent theories. Modern research has shown that channelling is an unwanted formation
because channels can change the concentration at the bottom and top of the bed by carrying a
great amount of flocs upwards.
Batch sedimentation tests were performed using flocculated slurry of Calcium Carbonate at
various initial concentrations such as 250 g/l, 500 g/l, 750 g/l and 1000 g/l to observe
channelling and reverse concentration gradients. Flux plots for the batch system reveal
behaviour which can be attributed to the upward flow of solids. In addition, photographic
methods were used to observe settling processes, channelling mechanisms and flocs in the
channels.
One of the purposes of this work was to examine the phenomenological solid-liquid
separation theory of Buscall and White (1987), which employs the material properties of the
local volume fraction, compressive yield stress Py (φ ) and hindered settling function R (φ ) to
identify the material behaviour in batch sedimentation. Stepped-pressure filtration and batch
vi
settling tests were used to measure the material characteristics for the flocculated CaCO3
suspension.
Experimental data were demonstrated using Height versus Time and Height versus
Concentration graphs and displayed the possible region of reverse concentration gradients and
channelling in the settling bed. Mathematical predictions adopted from Usher (2002) were
performed employing material characteristics of the material and graphical documentations
were presented. The results of mathematical predictions were compared to the experimental
results and the modes of sedimentation explained by Lester et al. (2005).
Fundamental theoretical models and experimental observations highlight that the main driving
force for channelling is the high-pressure gradient at the bottom of the bed and the most
important factors that cause channelling are high initial concentration of slurry and settling
time. The predictions also show that the material and flocculant used for the batch settling
tests demonstrate important effect on the settling process. The knowledge and information
gained from this study is valuable to maximize the thickening process.
vii
Table of Contents
DECLARATION .....................................................................................................ii
ACKNOWLEDGMENTS...................................................................................... iii
KEYWORDS..........................................................................................................iv
PUBLICATIONS.....................................................................................................v
ABSTRACT ...........................................................................................................vi
NOMENCLATURE............................................................................................xviii
1 CHAPTER 1 .............................................................................................. 1
1.1 INTRODUCTION.........................................................................................1
1.2 RESEARCH OBJECTIVES ..........................................................................4
1.3 STRUCTURE OF THESIS ...........................................................................5
2 CHAPTER 2 .............................................................................................. 8
2.1 LITERATURE REVIEW ..............................................................................8
2.2 BATCH SETTLING ...................................................................................10
2.3 FUNDAMENTAL THEORIES OF THICKENING AND BATCH
SETTLING ............................................................................................................14
2.3.1 Stokes Law ................................................................................................... 15
2.3.2 Coe and Clevenger........................................................................................ 15
2.3.3 Comings ....................................................................................................... 16
2.3.4 Kynch ........................................................................................................... 16
2.3.5 Talmage and Fitch ........................................................................................ 17
2.3.6 Recently Developed Theories ....................................................................... 17
2.3.7 Material Properties....................................................................................... 18
2.3.7.1 Compressive Yield Stress Py (φ ) .............................................................................18
viii
2.3.7.2 The Gel Point φ g .....................................................................................................19
2.3.7.3 Hindered Settling Function, R(φ ) ...........................................................................19
2.3.7.4 Solids Diffusivity, D (φ ) ..........................................................................................20
2.4 GOVERNING EQUATION ........................................................................20
2.4.1 Boundary Conditions .................................................................................... 22
2.4.2 Modes of Sedimentation by Burger ............................................................... 30
2.5 CHANNELLING IN BATCH SEDIMENTATION ....................................33
2.5.1 Possible Causes of Channelling..................................................................... 33
2.5.2 Mechanism of Channelling ........................................................................... 34
2.6 VOLCANO FORMATION .........................................................................40
2.7 REVERSE CONCENTRATION GRADIENTS ..........................................41
3 CHAPTER 3 ............................................................................................ 43
3.1 EXPERIMENTAL SETUP .........................................................................43
3.1.1 The Settling Tube ......................................................................................... 44
3.1.2 Mobile Platform............................................................................................ 45
3.1.3 Gamma (γ) Radiation Densitometer and Radiation Detector.......................... 45
3.1.4 Measurement of Radiation and Rate Meter.................................................... 48
3.1.5 Data Collection............................................................................................. 49
3.2 SUSPENSION MATERIAL........................................................................52
3.2.1 Sample Preparation ....................................................................................... 52
3.2.2 Determining Concentration ........................................................................... 52
3.2.3 Solid Particle Properties................................................................................ 52
3.2.4 Flocculation and Flocculant .......................................................................... 53
3.2.5 Calibrations .................................................................................................. 55
3.3 EXPERIMENTAL PROCEDURES ............................................................55
ix
3.4 STEPPED PRESSURE FILTRATION........................................................57
3.4.1 Stepped Pressure Filtration Technique .......................................................... 58
3.4.2 Apparatus and Measurements ....................................................................... 58
4 CHAPTER 4 ............................................................................................ 63
4.1 PHOTOGRAPHIC STUDY ........................................................................63
4.2 PURPOSE OF WORK ................................................................................64
4.3 STREAK PHOTOGRAPHY .......................................................................65
4.3.1 Equipment for Streak Photography ............................................................... 65
4.3.2 Camera and Film .......................................................................................... 66
4.3.3 Bracket Housing ........................................................................................... 68
4.3.4 Timing Mark................................................................................................. 68
4.3.5 Power Supply ............................................................................................... 70
4.4 OBSERVATIONS OF STREAK PHOTOGRAPHY...................................70
4.5 HIGH SPEED VIDEOGRAPHY.................................................................73
4.5.1 Equipment for High Speed Videography....................................................... 74
4.5.1.1 The High Speed Camera ..........................................................................................74
4.5.1.2 Lens..........................................................................................................................75
4.5.1.3 Lights .......................................................................................................................75
4.5.1.4 Camera Control Software........................................................................................75
4.5.2 Measurement Process.................................................................................... 76
4.6 VISUAL OBSERVATIONS .......................................................................79
5 CHAPTER 5 ............................................................................................ 83
5.1 RESULTS AND DISCUSSION ..................................................................83
5.2 OBSERVATIONS ON SETTLING BED....................................................84
5.3 CHANNELING IN SETTLING BED .........................................................85
x
5.4 VOLCANOES ............................................................................................89
5.5 H-T AND H-C GRAPHS ............................................................................92
5.6 THE DESCENT PATHS........................................................................... 101
5.7 EVALUATION OF PHOTOGRAPHIC WORK ....................................... 106
5.8 MATERIAL CHARACTERISATION RESULTS .................................... 107
5.8.1 Stepped-pressure filtration results ............................................................... 107
5.8.2 Batch Settling Tests Results........................................................................ 110
5.9 MATHEMATICAL MODELLING RESULTS......................................... 112
6 CHAPTER 6 .......................................................................................... 123
6.1 CONCLUSIONS AND FURTHER WORK .............................................. 123
6.2 CONCLUSIONS....................................................................................... 123
6.2.1 Channelling Behaviour ............................................................................... 124
6.2.2 Reverse Concentration Gradients ................................................................ 125
6.2.3 Material Characteristics .............................................................................. 125
6.2.4 Photographic Technique ............................................................................. 125
6.2.5 Mathematical Modelling Predictions........................................................... 126
6.3 RECOMMENDATIONS FOR FURTHER WORK................................... 127
6.3.1 Advanced High Speed Videography............................................................ 127
6.3.2 Flocculant and Material Effects................................................................... 127
6.3.3 Using Mixing Tank..................................................................................... 127
REFERENCES .................................................................................................... 128
A APPENDIX A – THE EXPERIMENTAL DATA................................ A-1
B APPENDIX B - THE CONCENTRATION LOCI CURVE DATA ... B-1
C APPENDIX C - THE DATA OF DESCENT PATH GRAPHS .......... C-1
xi
D APPENDIX D - STREAK PHOTOGRAPHY READINGS AND THE
HIGH SPEED CAMERA PHOTOS ........................................................... D-1
E APPENDIX E - DRY AND WET PARTICLE SIZE ANALYSIS...... E-1
F APPENDIX F - HEIGHT AND CONCENTRATION CALIBRATIONS
F-1
G APPENDIX G - CACO3 BATCH SETTLING TESTS .......................G-1
xii
List of Figures
Figure 2.1 - Batch settling process ....................................................................................... 13
Figure 2.2 - Typical compressive yield stress, Py (φ ) and hindered settling function, R (φ ) .... 19
Figure 2.3 - Consolidating batch settling bed........................................................................ 23
φg
Figure 2.4 - Typical flux plot depicting gel point , analysis limits φ min , φ max , inflection
point φ inf and mode ½ boundary φ1 .................................................................................... 25
Figure 2.5- Sedimentation dynamics for settling mode 1 ..................................................... 26
Figure 2.6- Sedimentation dynamics for settling mode 2a ................................................... 27
Figure 2.7- Sedimentation dynamics for settling mode 2b ................................................... 28
Figure 2.8- Sedimentation dynamics for settling mode 3a ................................................... 28
Figure 2.9- Sedimentation dynamics for settling mode 3b ................................................... 29
Figure 2.10- Settling plot for the batch sedimentation of an ideal suspensions: the positions of
the lines (a) between clear water and suspension; (b) suspension and sediment; (c) clear
water-sediment .................................................................................................................... 31
Figure 2.11 - Onset of channel formation ............................................................................. 36
Figure 2.12 - Joining channels.............................................................................................. 39
Figure 2.13 - An appearance of the volcano formation on the settling bed surface................ 40
Figure 2.14 - Reserve concentration gradient profile in different times................................. 42
Figure 3.1 – Laboratory Size Batch Sedimentation Unit ....................................................... 46
Figure 3.2 - Operating gamma rays through a settling column.............................................. 48
Figure 3.3 - Mobile platform controller unit and the countrate meter.................................... 49
Figure 3.4 - The datataker unit used for recording data from experiments........................... 50
Figure 3.5 - Datataker software program-Delogger............................................................... 51
Figure 3.6 - DeLogger program used for collecting .............................................................. 51
xiii
Figure 3.7 - The schematic diagram (a) and photographs (b) of the stepped pressure
unit in The University of Melbourne .................................................................................... 60
Figure 3.8 - Example of t versus V 2 results for the stepped-pressure compressibility test..... 61
Figure 3.9- Example of t versus V 2 results for the stepped-pressure permeability test.......... 62
Figure 4.1 - Camera set up for streak photography ............................................................... 67
Figure 4.2 – Diagrammatic explanation of streak photography process ................................ 67
Figure 4.3 - The bracket holds the camera in place under the gears ...................................... 68
Figure 4.4 - Motor and two gear sets rewinding the camera.................................................. 69
Figure 4.5 - The LED in the field of view of the camera slit ................................................. 69
Figure 4.6 - The 240v-12v converter, variable power supply ................................................ 70
Figure 4.7 - Formation of streak photography....................................................................... 71
Figure 4.8 - Example of LED light timing module on the base of the image ......................... 72
Figure 4.9 - Two different layers can be seen from the photograph....................................... 72
Figure 4.10 - The curve taken from the Figure 4.9 is shown ................................................. 72
Figure 4.11 - Phantom 4.1M Camera.................................................................................... 75
Figure 4.12 – Setting and recording process with Phantom 4.1............................................. 76
Figure 4.13 - Measurements dialog box on the screen .......................................................... 77
Figure 4.14 - The view cine info on the screen ..................................................................... 78
Figure 4.15 - Distance and speed measurement .................................................................... 78
Figure 4.16 - Particle 1 in three different positions in the channel ........................................ 80
Figure 4.17 - Calibrating of photographs.............................................................................. 81
Figure 5.1 - Channels on the wall of the settling tube (current study, run 1).......................... 85
Figure 5.2 - Collapse of a channel in the settling bed (current study, run 4 and run 5) .......... 87
Figure 5.3 – The Established major channel on the settling bed wall .................................... 88
Figure 5.4 – A minor channel on the settling bed wall (current study, run 7) ........................ 88
Figure 5.5 – The local channel on the settling bed wall (current study, run 8)....................... 89
xiv
Figure 5.6 – Volcano formations at the surface of the settling bed........................................ 90
Figure 5.7 – An erupting volcano on the surface of the settling bed...................................... 90
Figure 5.8 – Top view of volcano formation on the surface of settling bed ........................... 91
Figure 5.9 - Concentration in settling bed as a function of time for an initial concentration of
250 g/l ................................................................................................................................. 93
500 g/l ................................................................................................................................. 95
750 g/l ................................................................................................................................. 96
1000 g/l................................................................................................................................ 97
Figure 5.13 - Height vs. concentration plot showing reverse concentration gradients for an
initial concentration of 250 g/l ............................................................................................. 98
initial concentration of 500 g/l ............................................................................................. 99
initial concentration of 750g/l ............................................................................................ 100
initial concentration of 1000 g/l.......................................................................................... 101
Figure 5.17 - Calculation of pulp density solids by using descent paths .............................. 102
Figure 5.18 - The effect of solids upflow on descent path at steady settling condition at initial
concentration 250 g/l.......................................................................................................... 103
xv
concentration 1000 g/l........................................................................................................ 105
Figure 5.22 - Comparison of streak photographic measurements with the experimental results
(Initial concentration 500 g/l)............................................................................................. 106
Figure 5.23- Time versus (Specific Volume)2 results from the compressibility................... 108
Figure 5.24- Time versus (Specific Volume)2 results from the permeability ....................... 109
Figure 5.25- Height versus time results for batch settling tests ........................................... 111
Figure 5.26- Compressive yield stress results ..................................................................... 111
Figure 5.27- Hindered settling function results................................................................... 112
Figure 5.28- Mathematical predictions of height versus time for an initial concentration of 250
g/l ...................................................................................................................................... 113
g/l ...................................................................................................................................... 115
g/l ...................................................................................................................................... 116
Figure 5.31- Mathematical predictions of height versus time for an initial concentration of
1000 g/l.............................................................................................................................. 117
Figure 5.32- Mathematical predictions of concentration distribution with time for an initial
concentration of 250g/l ...................................................................................................... 119
concentration of 500 g/l ..................................................................................................... 120
concentration of 750 g/l ..................................................................................................... 121
concentration of 1000 g/l ................................................................................................... 122
xvi
List of Tables
Table 1-1- Various industrial applications of solid/liquid separation by use of thickeners....... 2
Table 3-1 - Collimator slit dimensions ................................................................................. 48
Table 3-2 - Properties of calcium carbonate used in the research .......................................... 53
Table 5-1- Compressibility results from filtration testing.................................................... 108
Table 5-2- Hindered settling function results from pressure filtration testing...................... 110
xvii
NOMENCLATURE
Latin Alphabet
c Refers column wall (-)
d Distance in the medium (m)
dc Thickness of the wall (m)
df Inside diameter of the column (m)
D (φ ) Solid diffusivity (m2 s-1)
f Refers to the suspension (-)
f (φ ) Scaled compressive yield stress (-)
fbk (φ ) Kynch batch flux density function (-)
g Gravitational acceleration (9.8 m s-2)
h(t) Interface height (sediment) (cm)
h’(t) Sediment/supernatant interface height (cm)
h0 Initial suspension interface height (sediment) (cm)
hc Critical suspension height (sediment) (cm)
I Intensity of the beam (photons/m2-sec)
I0 Incident intensity of the beam (photons/ m2-sec)
l Refers to the liquid (-)
L Initial concentration height (m)
pf, pp Scaled particle and fluid pressure (-)
Py (φ ) Compressive yield stress (Pa)
q(t) Bulk flow (kg/m3)
R (φ ) Hindered settling function (Pa s m-2)
s Refers to the solids (-)
t Time (s)
t Time for batch sedimentation (min)
uf Liquid velocity (m s-1)
up Particle settling velocity (m s-1)
xviii
V Specific filtrate volume (m3)
x Sediment bed height (m)
xc Characteristic position (m)
Greek alphabet
ε Voidage (-)
ρf ,ρp Liquid density and solid density (kg m-3)
∆ρ Density difference between solids and liquid phases (kg m-3)
φ Solids volume fraction (-)
φ0 Initial solids volume fraction (-)
φg Gel point solids volume fraction (-)
φ* Solids concentration at h( τ 1+ ) (-)
φ∞ Equilibrium solids volume fraction (-)
φ max Maximum solids volume fraction (-)
φ min Minimum solids volume fraction (-)
φ inf Inflection point for f( φ ) (-)
λ linear attenuation coefficient of the medium (-)
ρ Suspension density (kg m-3)
σ Shock propagation velocity (m s-1)
τ0 Time associated with decay of initialization mechanics (s)
τ1 Onset time of nonlinearity in h(t) (s)
τ max Time associated with onset of nonlinearity in h(t) (s)
xix
1 CHAPTER 1
1.1 INTRODUCTION
Sedimentation is a process in which particles or aggregates are separated by gravity settling
(thickening process) into clear liquid and slurry of higher solids content. The mechanism of
sedimentation deals with how solids settle from slurries during batch sedimentation. Solids
suspended in the fluid settle under their own weight until an equilibrium state has been
achieved. It is one of the most important large-scale processes used in industries such as
wastewater treatment, and various mining and chemical industries. The residue produced from
these industries contains different compounds separated through sedimentation by the
addition of flocculant. Sedimentation processes are utilized in food, pulp and paper, plastic,
cement and many other process industries. A thickener, a standard vessel filled with a
suspension in batch operation, is a sedimentation unit that operates according to the principles
of gravity settling and has advantages such as simplicity in design, economy of operation, the
capacity to handle extremely large flow volumes and versatility. The clear liquid is decanted
and the sediment removed periodically after settling. Generally, the actual thickening units are
not studied as most of the research into sedimentation is done in a laboratory environment
rather than in a more complex pilot plant. The various industrial applications of thickeners are
given in Table 1-1. Sedimentation is especially useful for efficient recovery of water from
high concentrations of tailings in solid dewatering operations. Improvement in the
dewaterability of the suspensions and more efficient operation of process equipment has a
critical role to maximize the recovery and the final solids concentration.
1
Table 1-1- Various industrial applications of solid/liquid separation by use of thickeners
(modified from Wakeman 1981; Cheremisinoff 1995)
Industry Applications
Mining and minerals

Alumina Red mud and hydrate
Iron ore Hydroseparation, concentrate, tailings
Copper/molybdenum ore Concentrate, tailings
Nickel/cobalt Concentrate, tailings
Potash Tailings
Phosphates Slimes, phosphoric acid clarification
Uranium Yellow cake
Zinc, gold, lead Tailings
Magnesia Concentration
Coal Refuse treatment
Steel Mill wastes
Flue gas desulfurization Scrubber water
Sand Slimes
Water treatment Sedimentation of grit and aggregates
Sewage treatment Sedimentation of silt
Chemical processing Separation of precipitated or crystallized solids
Plastics Separation of polymer particles from wash water
Food Clarification of preserves
Brewing Sedimentation of yeast skimming
Sugar Clarification of cane juice
Cement Thickening of cement rock slurry
Paper and board Effluent treatment
Paint and pigments Thickening of precipitated solids and
comminuted raw materials
Effective dewatering is vitally important to the cost and efficiency of many processes such as
the transportation and mixing of high concentrations of slurries (Glanister and Abbott 1989).
For the mineral and mining industries especially, the expectation is to optimize the process
conditions for the recovery of suspensions which has suspended contaminants.
Many factors affect the settling phenomenon such as the size and shape of particles, size
distribution, solid-liquid density differences, concentration of slurry, surface properties such
as zeta potential, chemical content, chemical additives and the suspension medium, etc
2
(Osborne 1990) and (Harris et al. 1975). Due to the interaction of these factors, the solids may
undergo particulate (individual) or hindered (aggregate moving) settling.
Since gravity thickeners are designed for thickening of sludges, hindered and compression
settling is usually the governing criteria for design of thickeners. If the solid in a suspension is
mainly composed of individual particles and particle size is only a few micrometres in
diameter, the gravity settling rate would be very low and possibly too low for practical
operations (McCabe et al. 1980). In general, these individual particles are clustered together
into flocs by physical and chemical means. Flocs begin to move about individually at quite
high speeds and in more or less random directions, after the initial period in which very little
movement takes place while flocs are forming.
Forbes (1912) proposed certain standards for conducting batch settling tests which should
permit operators to better predict thickener performance, and manufacturers to meet settling
requirements of equipment on thickener performance. When information on the settlement of
a suspension is needed for the design of continuous gravity thickeners, batch sedimentation
tests are often performed. The main purpose of conducting a batch thickening test is to
produce data in order to predict the performance of continuous thickening tanks (Kos 1978).
Because it usually points to the solids dewatering capability of the process, the shape of the
concentration profile, plotted by using data from batch sedimentation experiments, is
especially of interest to the designer and operator of the thickener. This information is used to
investigate sedimentation for process optimization, control and knowledge (Holdich and Butt
1995). Because this method is usually the first process of the sludge handling and disposal
scheme, it affects the size, the performance and the economy of the subsequent processes. The
scope of this project is to improve the thickening productivity by identifying and analyzing
some unwanted behaviour called channelling in the batch sedimentation process.
3
The problematic issue arises at the bottom of a thickener which is the compression zone
where channels occur, requiring specific operating conditions. The boundary conditions at
these compression zones are not well known. Fitch (1993) stated the first problem as, “we do
not know how to handle the effects of channelling”, when the concentration profile in a batch
settling test were formed after the settling was completed. It is of interest to investigate what
would happen if the operating conditions or methods were changed according to thickener
unit predictions. At this point, the data from batch sedimentation experiments are to be the
guide in terms of understanding the action at the problem area. Factors such as pressure drop
in the settling bed, porosity profiles around the channel, initial concentrations, and the density
of the solid and liquid, make the determination of channelling complicated. These variables
have influence on the batch sedimentation process in the settling bed. The most important
point is that the channels change the bed concentration and cause other events such as reverse
concentration gradients in the settling bed. In the batch sedimentation process, wherever flux
curves are plotted, channelling (also described as short circuiting of fluid through the higher
concentrations) during compression causes the real concentration value in the bed to be higher
than the normal value. This leads to overestimation of the thickener capacity (Svarovsky
1990).
It is important to know how the systems would respond to changes in the variables mentioned
above and what their effect on the settling process is. Such designs are based on performing
batch test with sludges over a range of concentrations that provide a valuable framework for
the simulation, design and control of thickeners in typical commercial practice.
1.2 RESEARCH OBJECTIVES
This study endeavors to extend the understanding on previous investigations by studying the
phenomenon of channelling in the batch sedimentation process using experimental data and
4
mathematical dewatering model proposed by Buscall and White (Buscall and White 1987).
The objectives of this research are:
1. Reviewing of fundamental theories on thickening and batch settling,
2. Observing the mechanism of settling, channels and volcano formations by applying
photographic techniques,
3. Understanding the nature and causes of channelling and determining boundaries, also
giving extensive information about when and how channels occur, carrying out
investigations to determine the key parameters and the variables that affect the
formation of channels in a settling bed,
4. Analyzing the relationship between channelling behaviour and reverse concentration
gradients,
5. Evaluating the effect of concentration on channelling phenomena by testing the
settling behaviour at different concentrations,
6. Measuring the material properties using batch settling and pressure filtration
techniques, which through comparisons of experimental results with model
predictions, gives indications of the effects of channeling phenomena.
A major contribution to the scientific and the industry community from this project is to
determine optimized operating conditions for batch sedimentation and indirectly for
thickeners as well. Results of these investigations should provide perspective about how these
properties affect the batch sedimentation process.
1.3 STRUCTURE OF THESIS
This thesis is oriented towards the understanding of batch sedimentation of particulate
suspensions. Particular attention is given to channelling phenomena and reverse concentration
gradients which occur in the settling bed.
5
In Chapter 1, the interest of the research is emphasized by the review of industrial issues and
expectations. The research objectives and structure of thesis are placed in this chapter.
Chapter 2 addresses general information of suspensions, solid/liquid separation and batch
sedimentation. The aim here is to provide general information about the key terms often used
in the subsequent chapters. Widely used and well known fundamental theories on thickening
and batch sedimentation from the 19th century to the present are then summarized. The
material properties are given, followed by governing equations and modes of sedimentation.
Then, explanations on channelling are presented, focusing on the conditions and mechanism
of channelling, the characteristics and boundaries of channels, and the different forms of
channels. In addition, volcano formation and reverse concentration gradients and how they
relate to batch sedimentation process are also explained.
In Chapter 3, the experimental equipment, apparatus, flocculant and material used for the
experimental investigation are introduced and the experimental process is described. In
addition, the stepped pressure filtration technique used for determining material properties is
explained.
In Chapter 4, a description of the photographic techniques used, such as streak photography
and high speed videography with equipment are given. Some of the previous studies that lead
to these photographic methods are given, and comparisons are made regarding the results.
Examples of images of flocs in the channels are also visually documented.
In Chapter 5, the detailed results gained after observations from the experimental and
photographic study are given. In addition, the material properties of suspension employing
stepped pressure filtration information are presented. The outcomes from the mathematical
6
theory using material properties are compared against the experimental data.
In Chapter 6, overall conclusions and major outcomes are drawn from the study. The results
are evaluated against the research objectives. Furthermore, recommendations for further work
are given.
7
2 CHAPTER 2
2.1 LITERATURE REVIEW
Suspensions are mixtures of two immiscible phases such as oil in water (emulsions) or solids
in water (particulate suspensions or sludge). Examples of particulate suspensions include
mineral tailings, water and wastewater treatment sludges, and pigments. Suspensions are used
everywhere in our everyday life, especially in the minerals industries in numerous process
operations. Factors that have an influence on the suspension behaviour include solids
concentration, particle size distribution, liquid density, liquid viscosity, liquid chemistry, the
surface charge on the solid particles, interparticle forces, chemical additives such as
flocculants and coagulants, and other influences. Suspensions in liquids can be separated by
simple filtration, as by pouring the suspension through a sieve or filter.
Solid-liquid separation is used in sedimentation, centrifugation and filtration operations.
Dewatering operations are useful in that they allow efficient recovery of water and at the same
time, produce tailings, the desired product, at a high concentration of solids. Effective
dewatering is vitally important to the cost and efficiency of many processes (Glanister and
Abbott 1989). In saturated dewatering processes, the important forces applied to suspended
solids are gravitational force in gravity thickening, applied pressure in pressure filtration,
centrifugal force and shear force.
The understanding of the physical and chemical behaviour of suspended solutions, through
fundamental theories and techniques for the measurement of sludge properties, helps
engineers to identify, predict and optimize process conditions, evaluate alternative solid-liquid
separation processes, open new fields of application and improve the process quality.
8
The results from laboratory tests are used as a basis to determine the design parameters for
industrial devices.
Provided that gravity overcomes interparticle repulsive forces, particulate suspensions settle
due to phase density differences. Sedimentation is a process of separating the heavier particles
in suspended solids from the solution. Solid matter settles on the bottom of the vessel through
batch or continuous thickening, and the clear water placed on the upper part of the vessel. The
degree of separation depends on the size and the nature of suspensions, the fluid viscosity, and
the retention time. Equipment size and effectiveness of gravity separation depends on the
solids settling velocity, which is a function of the particles size, density difference, fluid
viscosity, and particle concentration (hindered settling). Gravity separation is also used for
classification where particles of different sizes are separated.
In this chapter, a description of important terms is given, followed by a history of the theories
for sedimentation of flocculated suspensions. Based on the phenomenological model of
Buscall and White (1987), the relevant fundamental material properties to determine the
behaviour of the solids during batch sedimentation processes are, the compressive yield stress,
Py (φ ) and the hindered settling function, R (φ ) , where φ is the solid volume fraction.
Thickening, filtration and centrifugation models based on fundamental theories provide means
of optimizing process equipment. The fundamental material properties are used to establish
the mathematical equations to describe behavior of the material in any dewatering device. It is
significant to provide useful experimental and modeling tools in order to help the operators
make direct comparison of additives and sludge types, and help them to make accurate
operational, design and control decisions. This assists in minimizing the flocculant usage and
optimizes the performance of thickeners, filters and centrifuges, thus reducing the cost
associated with the sludges.
9
Batch settling tests are generally used to investigate the gravitational settling behavior of
suspensions. However, some unexpected activities occur during batch sedimentation, such as,
reverse concentration gradients, channelling, volcano formation and wall effects. Under
certain conditions some sludges release liquid in streams or short-circuits which are large but
few in number. In transparent settling columns, these streams can be first seen as small
channels along the outside of the thickening sludge which cause miniature volcanoes at the
solid-liquid interface. This flow of water in channels can cause particles to move from their
original location towards the upper regions of the settling bed. This migration process causes
a higher concentration above a lower concentration in the settling bed, and is called a reverse
concentration gradient. Reverse concentration gradients change the settling bed behavior.
Previous studies of channelling, volcanoes and reverse concentration gradients in settling beds
are investigated in the literature review.
2.2 BATCH SETTLING
Batch settling tests generally involve placing a suspension in a vertical cylinder and recording
the height of the interface between the settling solids and the clear-liquid over time. The
factors affecting the settling behavior of particulate suspension include (Harris et al. 1975;
Osborne 1990) :
a. Particle size distribution, shape, density, mineralogical and chemical properties,

b. Initial suspension height and concentration,
c. Pressure gradients,
d. Floc and floc aggregate size,
e. Rheological properties of suspension,
f. Solid concentration,
g. Chemical conditioning, flocculation, heating/cooling etc,
h. The type of containing vessel; size, shape, wall effects etc,
10
Batch settling tests are applied for non-flocculated suspensions (ideal suspensions) and
flocculated suspensions (non-ideal suspensions). In this study, all experimental tests are based
on flocculated suspensions. Although the actual mechanisms of bridging flocculation are not
the subject of this study, brief information is given about flocculation in Chapter 3. More
information can be reviewed in detail Glover et al. (2000), Owen et al. (2002) and Swift et al.
(2004).
Buscall and White (1987) introduced the rheological properties of concentrated suspensions
and applied it to the settling under gravity of a flocculated suspension in a closed-bottom
vessel. The flocs form a connected aggregate structure after an electrolyte or polymer
flocculant is added to the suspension, and these flocs fall more quickly than single solid
particles because of their larger mass-to-surface-area ratio. While flocculation speeds up the
settling process, the consolidated bed forms an open network of flocs whose average volume
fraction is lower than that achieved by an unflocculated sedimenting suspension.
The most important issue in this process is the concentration profile of the solids, because the
profile displayed in batch settling reflects settling regimes which are related to concentrations
of the solids. For instance, at very low concentrations, the flocs settle more or less
independently of each other. As equilibrium is reached between gravitational and drag forces,
the particles reaches their terminal velocity which affects the flow of liquid past a neighboring
floc.
The sedimentation of flocculated suspensions is divided into three concentration regions.
These are; zone settling, intermediate concentration region (channeling or phase settling
region), and compression. Therefore, the bottom-lying flocs are compressed by the weight of
the others that settle upon them. Coe and Clevenger (1916) developed the first clear
11
fundamental picture about the principles of batch-hindered sedimentation and compression in
a continuous thickener. The mechanism of sedimentation may be best described by
observation of what occurs during a batch settling test as solids settle from a slurry in a
vertical glass cylinder. Four zones appear as shown in Figure 2.1. As soon as the process
starts, all particles begin to settle and are assumed to rapidly approach the terminal velocities
under hindered settling conditions. Highly concentrated suspensions display interference of
individual particle settling as a result of this phenomenon. It usually causes the descent of the
particulate mass to occur at a more or less uniform rate.
Hindered settling (or thickening) is a term used to describe behavior at higher concentrations
where sedimentation rates are largely related to concentration rather than to particle size
(Rushton et al. 1996). Particulate motion is said to be hindered because settling takes place by
a massive subsidence rather than by individual particle fall.
When the suspension is placed in a vertical glass cylinder, and the solid percentage is above a
minimum value, a distinct line forms near the upper surface as seen in Figure 2.1. The first
zone, Zone A, is a region of clear liquid. Zone B (the Hindered Settling Zone) is a uniform
concentration zone that has the same concentration and distribution as the initial
concentration. Zone C (the Transition Zone) is the boundary between the hindered settling
and the compression zones. It is a region of non-uniform concentration. Zone D (the
Compression Zone), consists of networked solids. Channels occur through the compression
zone. The fluid that is entrapped during the settling process in Zone D is forced out as it
compresses.
12
A Clear liquid C Variable concentration
B Initial concentration D Settled bed
Figure 2.1 - Batch settling process
The heights of the various zones change with time as shown. When the slurry being fed per
unit time to a continuous thickener is equal to the rate of sludge and clear liquid removal, they
reach a steady state condition and each zone height will be constant. Particles with a narrow
size range settle with the same velocity. The sediment height increases as the particles settle
from zone C. As the upper interface approaches the sludge build up on the bottom of the
container, the slurry appears as a more uniform sludge in Zone D. The uniformity of the slurry
depends on the material used for the sedimentation. The settling zone B disappears followed
by slow compaction of solids in zone D.
13
2.3 FUNDAMENTAL THEORIES OF THICKENING AND
BATCH SETTLING
Batch sedimentation has been used in the laboratory for investigating the settling
characteristics of suspensions since the 19th century. Some widely used and well known
theories are presented here.
Classical design procedures for sizing gravity thickeners have been developed in parallel to
filtration theory. It began with Stokes (1851) who only considered a single particle behavior
in the settling process. Coe and Clevenger (1916), Kynch (1952), Talmage and Fitch (1955)
ignored the role of the compressing sediment and employed the method of flux theory, which
is based upon the rate of sedimentation being uniquely dependent upon the concentration of
solids. Comings (1940-1950) drew attention to the importance of local concentration for the
thickening process.
Fitch (1983) further investigated Kynch’s theory and modified the procedures to determine
the characteristics line on a batch settling flux curve. Font (1988, 1991) reviewed these works
and discussed the compression zone effects in batch sedimentation. During the 1980’s, a
group of researches, the likes of Concha and Bustos (1985), Buscall and White (1987),
Auzerais et al. (1988), Landman et al. (1988), Bascur (1989) and Davis and Russel (1989),
published papers well recognized by the international scientific community based on the
phenomenological behavior of particulate networks.
Furthermore, the scientists required mathematical solutions for solving specific problems
because there was no sufficient numerical method that existed for solving the initial boundary
value problems of batch and continuous thickening.
14
Buscall and White (1987) developed a mathematical theory of the dewatering of flocculated
suspensions by using the concept of a flocculated particle network structure. In 1998, Burger
started to work on sedimentation and thickening theory and added his cooperation on the topic
of mathematical analysis of sedimentation models.
2.3.1 Stokes Law
The terminal settling velocity of a single spherical particle in laminar flow was first derived
by Stokes (1851). When the solids concentration is low, particles settle independently, a
phenomenon called free settling. This type of particle settling can be explained by Stokes
Law. The characteristics of the particle and the properties of the fluid both have an influence
on the settling velocity of the particle, while that particle settles independently from others in
the settling column. Stoke’s Law predicts velocities proportional to the square of the particle
diameters. However, when the concentration of solids is high, the distance between particles
becomes so small that the particles come into contact with one another.
As a result, the settling process occurs differently and Stoke’s Law fails to describe the
settling of particles beyond infinite dilution (Renko 1998). In practical cases, the drag on the
particles is dependent on turbulence and shape of the particle as well as the presence of other
particles, for which Stoke's Law cannot be applied.
2.3.2 Coe and Clevenger
Coe and Clevenger (1916) were the first scientists to use the observed batch settling data in a
laboratory column for the design of an industrial thickener. They made a basic assumption
that the settling velocity of solids is a function of concentration only and discussed the solids
flux density. However, in reality the solid volume fraction varies with depth in the
compression zone. The four zones discussed previously were first defined by Coe and
Clevenger (1916). They also used different materials and reported their behavior. Although
there are still complexities due to compressible behavior, their equation continues to be the
15
most common method of thickener design to date (Concha and Burger 2002).
2.3.3 Comings
The first paper to recognize the importance of local concentration and sediment composition
for the thickening process was published by Comings in 1940 and was called “Thickening
Calcium Carbonate Slurries”. He and his co-workers showed different zones in a continuous
thickener: the clarification zone at the top, the settling zone underneath, and the upper
compression zone at the bottom. Two findings shed some light on thickening process from
their study. First, Comings and his co-workers stated that the concentration in the settling
zone is nearly constant for a thickener in the steady state condition. The concentration in the
settling zone depends on the rate at which the solids are fed into the thickener, and not on the
concentration of the feed suspension.
Secondly, increasing or decreasing the sediment depth could adjust the underflow
concentration for the same feed rate (Concha and Burger 2002), which is indicative of the role
of compression in thickeners.
2.3.4 Kynch
Kynch’s theory had great influence in the development of sedimentation theory. He developed
a method to find the flux-concentration relationship in a single batch test. Kynch’s theory
predicts the sedimentation behavior of highly concentrated suspensions of mono-dispersed
rigid spherical particles. Yet, Kynch (1952) ignored the details of forces on the particles and
the wall effects when they settle. The assumption of his theory is that the settling velocity of a
particle depends only on the local concentration of suspension at any point in the suspension.
His theory is applicable to only certain materials and does not apply to the compressible zone.
He also proposed a graphical method for estimating permeability. In Kynch’s theory, the
settling curve shows that the solids concentration layers travel upwards at uniform velocity
and the velocities are constant for each concentration.

16
2.3.5 Talmage and Fitch
Talmage and Fitch (1955) introduced the settling plot and affirmed that one settling plot
contained all the information needed to design a thickener. Two important pieces of
information were used when they were plotting the graph – the solid flux that can be passed
through the thickener and the initial slurry height and concentration in a batch settling test.
The slope from the water-suspension interface (height (h) versus time (t) diagram) gave the
settling rate of the suspension, the slope at different times represented the settling velocity at
different concentrations.
2.3.6 Recently Developed Theories
There are many researchers who have studied the batch settling process since 1980. The few
new theories from Bustos, Concha and Burger; Landman and White; Buscall and White who
used material properties and came up with mathematical equations for batch sedimentation
are introduced.
Buscall and White (1987) established the gel point, compressibility and permeability as the
fundamental physical properties that determine the dewaterability of flocculated suspensions.
They also discussed the rheological properties of flocculated suspensions, such as, the
compressible yield stress, Py (φ ) , and the hindered settling function R (φ ) , of concentrated
suspensions, and described the estimation of the yield stress from a batch centrifuge
experiment. A new experimental technique for calculating Py (φ ) commenced after their
theorical work.
Landman and White (1997) developed the technique to determine R (φ ) and Py (φ ) from
pressure filtration experiments. The technique takes into account both the compressibility and
permeability of the filter cake during formation and consolidation.
17
A valuable framework for the simulation of batch and continuous thickening is provided with
a series of papers on the phenomenological theory of sedimentation of flocculated
suspensions by various researches (Concha et al. 1996; Burger and Concha 1998; Burger et
al. 1999; Bustos et al.1999). The numerical methods have a built-in property to reproduce the
appropriate discontinuities (Burger et al. 2000a and Burger et al. 2001). Burger et al. (2000c)
and Garrido et al. (2000) show the application of the method to several batch sedimentation
processes.
2.3.7 Material Properties
The material properties stated by new theorists are introduced for further explanations of the
sedimentation process. These material properties can be used to predict dewatering behavior
in batch sedimentation, (Howells et al. 1990; Landman and White 1994).
2.3.7.1 Compressive Yield Stress Py (φ )
The compressive yield stress is a material characteristic parameter which is a function of φ . In
a flocculated suspension under pressure, it is a direct measure of the particle pressure, or of
the strength of the bonds between particles (Landman and White 1994). When a compressive
pressure is applied, φ may be viewed as the solids concentration at which the material will
settle (Usher 2002). The compressive yield stress for the suspension needs to be determined in
order to characterize the compressional rheology of concentrated flocculated suspensions
(Green et al. 1996). The relationship between the compressive yield stress, Py and solid
volume fraction, φ, is given in Figure 2.2.a, and the relationship between the hindered settling
function, R and the solid volume fraction, φ, is shown in Figure 2.2.b.
18
Hindered Settling Function, R (φ )
Compressive Yield Stress, Py (φ )
φg φg
Solids Volume Fraction, φ (v/v) Solids Volume Fraction, φ (v/v)

(a) (b)
Figure 2.2 - Typical compressive yield stress, Py (φ ) and hindered settling function, R (φ )
(Stickland 2005)
2.3.7.2 The Gel Point φ g
The gel point, φ g , is a critical concentration in the network where flocs or particles come into
contact to form a continuous network structure for flocculated materials. The gel point is the
lowest solids volume fraction at which particles or flocs form a self-supporting network, such
that, the material at the top of a bed has no compressive forces and is at φ g . As a result, the
gel point may alternatively be defined as the solids concentration at which the compressive
yield stress becomes zero.
2.3.7.3 Hindered Settling Function, R(φ )
The hindered settling function, R(φ ) , quantifies the interphase drag of colloidal suspensions
for all solids volume fractions, φ (Lester et al. 2005). It is also defined as the hydrodynamic
resistance to flow of liquid through a particulate suspension. The hydrodynamic interactions
between particles hinder the settling velocity and R(φ ) increases as the concentration
increases.
19
This factor is useful in characterizing the consolidation rate of the suspension at all
concentrations. The usefulness of the hindered settling function, R(φ ) , is that it has sound
fundamental basis and can be used together with the compressive yield stress, Py (φ ) , to
completely model solid-liquid separation in the design of all types of consolidation equipment
(Green et al. 1998).
2.3.7.4 Solids Diffusivity, D(φ )
Solid diffusivity, D (φ ) , is used to describe the combined effects of compressibility and
permeability. It is a measure of the overall dewaterability of the suspension and is related to
Py (φ ) and R(φ ) :
dPy (φ ) (1 − φ )2
D(φ ) = 2-1
dφ R (φ )
A suspension with a higher D(φ ) is more dewaterable due to the inverse relationship between
D(φ ) and R(φ ) , accounting for the effects of Py (φ ) .
2.4 GOVERNING EQUATION
Landman and White (1994) reviewed the rheological model developed by Buscall and White
(1987). The conservation of momentum equations for the solid and liquid phases balances the
network pressure, hydrodynamic (drag), hydrostatic, and acceleration (gravitational or
centrifugal) forces acting upon a volume difference of solids, and are given by Equations 2.2
and 2.3 respectively:
∂p f ∂p p
φR(φ )(u p − u f ) − φ − − φρ p g = 0 2-2
∂x ∂x
20
∂p f ∂p p
φR(φ )(u p − u f ) + (1 − φ ) + + (1 − φ ) ρ f g = 0 2-3
∂x ∂x
where g is the gravitational acceleration, u p − u f is the local velocity of the particle relative
to the fluid, ρ p and ρ f are the particle and fluid densities respectively, and pp and pf are the
∂p f
local particle and fluid pressures. is eliminated from Equations 2.2 and 2.3 to give;
∂x
φ ∂p p
R(φ )(u p − u f ) − − φ∆ρg = 0 2-4
1−φ ∂x
The density difference between the solid and liquid phases is given as ∆ρ . The first term in
Equation 2.4 represents the drag due to the fluid-solid interaction, the second term
corresponds to the strength of the particulate network in response to an applied pressure, and
the third term represents the gravitational force. The conservation of mass equations for the
particle and fluid phases are given by Equations 2.5 and 2.6, respectively, assuming
incompressibility of both phases;
∂φ ∂ (φu p )
= 2-5
∂t ∂x
∂(1 − φ ) ∂ ((1 − φ )u f )
= 2-6
∂t ∂x
The overall conservation of mass is given by adding Equations 2.5 and 2.6;
∂
∂x
[
φu p + (1 − φ )u f = 0 ] 2-7
Equation 2.7 is integrated and equated to the bulk flow, q(t), which is constant with respect to
x.
21
φu p + (1 − φ )u f = q (t ) 2-8
The bulk flow, q(t), is zero for batch sedimentation and equal to the specific flow rate of
filtrate, dV/dt, for filtration or continuous thickening. The solids velocity relative to the fluid
velocity can be written rearranging Equation 2.8 with q(t) = 0:
up
up −uf = 2-9
1−φ
Using Equation 2.9 to eliminate u f in Equation 2.4 gives:
φu p ∂p p
R(φ ) − − φ∆ρg = 0 2-10
(1 − φ ) 2
∂x
Buscall and White (1987) stated that if the viscous drainage of fluid is rate determining rather
than the breaking and/or re-forming of interparticle bonds, and the applied pressure is
constant or increasing, then p p = Py (φ ) . Equation 2.10 is re-written with p p = Py (φ ) and
combined with Equation 2.5 to eliminate u p , giving:
∂φ ∂  ∂φ 
=  D(φ ) + f bk (φ ) 2-11
∂t ∂x  ∂x 
where f bk (φ ) is the batch flux density function (Burger and Concha 1998), and is given by:
f bk (φ ) = φ∆ρg
(1 − φ )2 2-12
R(φ )
2.4.1 Boundary Conditions
Batch settling concerns the sedimentation and consolidation of a suspension at uniform initial
solids concentration, φ 0 , and initial suspension height, h0 , under gravity in a vertical vessel. A
22
suspension with an initial solids concentration below the gel point ( φ 0 < φ g ) is considered,
which means there is initially no network between the particles. An illustration of a typical
batch settling process for φ 0 < φ g is shown with changes of parameters of the three zones in
the settling column in Figure 2.3.
Initially, the solids settle to form a clear layer of supernatant ( φ =0) at the top of the column,
while the consolidating bed ( φ ≥ φ g ) is at the bottom. Depending on the sedimentation type, a
transition zone of unnetworked suspension ( φ 0 < φ < φ g ) can also form between the falling
zone ( φ = φ 0 ) and the compression zone ( ∂φ < 0, φ g < φ < φ (0, t ) ).

∂x
Initial Solids Height h0
Supernatant ( φ = 0 )
Solids Height h(t)
Free Fall Zone ( φ = φ 0 )

Critical Height xc(t)
Compression Zone ( φ > φ g )
Vessel Bottom x=0
Figure 2.3 - Consolidating batch settling bed
In the free fall zone, D(φ ) = 0, therefore Equation 2.11 becomes
∂φ ∂f bk (φ )
= 2-13
∂t ∂x
23
The solution to this equation depends on the functional form of fbk( φ ).
Lester et al. (2005) showed the possible analytic solution of batch settling experiments for
φ 0 < φ g by using the entropy condition (see Equation 2.14) and estimating R(φ ) . Their
analysis is limited to flux curves f (φ ) , given in Figure 2.4, containing at most, one inflection
point and their method for the analytical solution only requires information of φ o , φ g and ∆ρ .
The assumptions are that:
• Batch settling is a one-dimensional process;
• The suspension is initially unnetworked ( φ 0 < φ g );
• φ 0 and φ g are known; and
• There is irreversible compression (Buscall and White 1987)
The entropy condition given by Equation 2.14 is written as;
( )
f (φ ) − f φ −
≤ σ φ(+
, φ)−
≤
( )
f φ + − f (φ )
for φ − < φ < φ + 2-14
φ −φ− φ+ −φ
( )
where σ φ + , φ − is given as the shock propagation velocity between the two states φ + and
φ − . φ 0 is used to determine the different modes of sedimentation. However, this analysis is
limited to flux curves f (φ ) containing at most one inflection point (de Kretser et al. 2001;
Landman et al. 1999). These curves display three different sedimentation modes, as
characterized by the nature of the φ 0 → φ g transition at the sediment-suspension interface.
The settling modes are:
• Mode 1: shock
• Mode 2: contact discontinuity: (a) no origin fan, (b) origin fan

24
• Mode 3: fan: (a) origin fan, (b) no origin fan
Settling mode 1 occurs if a flux curve does not contain an inflection point φinf or φinf ≥ φ g . In
order to determine the settling mode in φinf ≤ φ g , the initial solid volume fraction φ 0 is
specified with respect to the concentrations φ1 , φ inf , φ min , φ max which can be summarized as :
φ1 is the mode ½ boundary solids volume fraction, φinf is an inflection point for f (φ ) , φ min is a
minimum analytic solids volume fraction, φ max is a maximum analytic solids volume fraction.
φg φ
Figure 2.4 - Typical flux plot depicting gel point , analysis limits φ min , max , inflection point
φ inf and mode ½ boundary φ1 (Lester et al. 2005)
Settling mode = { 0 < φ ≤ φ1 mode 1
{ φ1 < φ ≤ φ min mode 2a
{ φ min < φ < φinf mode 2b
{ φinf < φ ≤ φ max mode 3a
{ φ max < φ < φ g mode 3b
25
The settling mode 1 is characterized by linear behavior of h(t ) up to t = τ 1 = τ max as depicted
in Figure 2.5, where τ 1 is the onset time of nonlinearity in h(t ) and τ max is the time
associated with onset of nonlinearity in h(t ) . In this settling mode, a shock exists between the
sediment and consolidation region, where φ steps from φ 0 to φ g . The velocity of the
sediment/supernatant interface height ( h ' (t ) ) is constant for t < τ 1 and may be discontinuous
at t = τ 1 .
Figure 2.5- Sedimentation dynamics for settling mode 1 (Lester et al. 2005)
The graphical explanation is given in Figure 2.6 for settling mode 2a. A transition from linear
to nonlinear behavior in h(t ) takes place at time τ 1 = τ max and here h ' (t ) is discontinuous.
This arises due to a contact discontinuity and associated fan originating tangentially from the
consolidating bed, where φ steps from φ 0 to φ * , and φ * is the solids concentration at h(τ 1 ) .
26
Figure 2.6- Sedimentation dynamics for settling mode 2a (Lester et al. 2005)
In Figure 2.7 the graphical explanation is given for the settling mode 2b. Two types of
discontinuity exist. A contact discontinuity ( φ steps from φ 0 to φ * ) spreads from the origin,
generating a discontinuity in h ' (t ) at τ 1 . Where φ increases from φ * to φ max , the fan can be
seen which radiates from the origin and intercepts the sediment/supernatant interface at τ max
and thereafter, a second fan takes place tangentially from the consolidating bed.
According to Figure 2.8, h(t ) changes from linear to nonlinear behavior at τ 1 , even though in
this case h ' (t ) is continuous for settling mode 3a. The fan radiating from the origin covers the
transition from φ 0 to φ max and the corresponding characteristics intercept h(t ) at τ 1 and τ max ,
respectively. A second fan emanates tangentially from the consolidating bed.
27
Figure 2.7- Sedimentation dynamics for settling mode 2b (Lester et al. 2005)
Figure 2.8- Sedimentation dynamics for settling mode 3a (Lester et al. 2005)
28
Although now the fan only emanates tangentially from the consolidating bed, h(t ) is fluent at
the transition from linear to nonlinear behavior at τ 1 = τ max in the settling mode 3b shown in
Figure 2.9.
Figure 2.9- Sedimentation dynamics for settling mode 3b (Lester et al. 2005)
Unexpected phenomena such as channel formation can be seen during the batch settling
process, resulting in S-shaped settling curves Equation 2-11 does not cover the whole
dynamics of batch settling, especially at early times. The assumption has been made by Lester
et al. (2005), that the initialization (unexpected phenomena) mechanisms do not have
influence beyond τ 0 when these effects become negligible. Therefore there is no effect on the
result except for temporal delays in the sedimentation.
Batch settling tests involve a range of solid concentrations from very dilute up to φ ∞ , however
they are simple experiments to perform. Estimation of the flux curve function from
experimental batch settling requires knowledge of φ g , φ 0 and h0 ∆ρ .
29
When combined with pressure filtration data, this gives complete characterization over all
relevant φ data.
2.4.2 Modes of Sedimentation by Burger
It is possible to describe the different types of Kynch Sedimentation Process (KSP) for an
ideal suspension and these are called Modes of Sedimentation (MS) in the literature. The
Kynch theory of batch sedimentation dwells on a single conservation law with constant initial
and boundary conditions. In certain circumstances, Kynch (1952) formulated a mathematical
treatment of the details of batch settling and found that the complete solution could be
determined only if the possibility of discontinuities or shock waves were admitted. The shock
causes different cases which appear for typical flux curves found in the literature (Rhee et al.
1986), and presents one point of inflection given in the paper published by Auzerais et al.
(1988). They constructed the complete solution of the sedimentation problem for an initially
uniform suspension of any volume fraction φ 0 , together with appropriate combinations of
Kynch theory and shock conditions (Concha and Bustos 1991).
In the late eighties, Bustos and Concha (1988) showed the mathematical solutions
concentrated on the initial and boundary value problems with different initial conditions. They
used three field variables for their mathematical model of an ideal suspension in an ideal
thickener as the volume fraction of solids φ (x, t ) , the solid flux density f ( x, t ) and the volume
average velocity q (x, t ) . The position of the water-suspension interface, and the top layer of
the sediment with time in batch sedimentation of an ideal suspension, is shown in Figure 2.10
as a settling plot. Concha and Bustos (1991) observed that there is not only a shock wave but
also a constant increase of concentration which will fill the wedge in the settling plot from the
bottom to top of the settling bed.
30
Z
(a)
(c)
zc
(b)
t1 tc t
Figure 2.10- Settling plot for the batch sedimentation of an ideal suspensions: the positions of the
lines (a) between clear water and suspension; (b) suspension and sediment; (c) clear water-
sediment (Concha and Bustos 1991)
This is called a rarefaction wave which is the fan with slope increasing with z. They described
batch sedimentation as the suspension having a homogeneous initial concentration φ 0 , and
that at the bottom of the settling column there is a continuous but extremely rapid increase of
concentration from φ 0 to the final concentration φ ∞ . They introduced the solution of the batch
sedimentation problem using the method of characteristics and the theory developed by
Ballou (1970) for two inflection points in which zones of constant concentration are separated
by shocks, rarefaction waves or combinations of these. There are five different modes of
sedimentation reported for flux density functions with at most two inflection points. These
are:
MS-I: a shock
MS-II: a contact discontinuity and a rarefaction wave
MS-III: a rarefaction wave
MS-IV: a contact discontinuity, a rarefaction wave and another contact discontinuity
MS-V: a rarefaction wave and a contact discontinuity
31
After that, Bustos et al. (1999) presented two new modes of sedimentation and stated that the
supernate-suspension interface is not a sharp shock but a rarefaction wave. All seven modes
are determined by the constitutive equation of the flux-density function and the initial
concentration φ 0 of the suspension. Recently, flux density functions with one or two inflection
points and constant initial concentration have a maximum of seven MS. The seven different
behaviors were observed that result in seven different settling plots characterized by two
regions of constant concentration φ 0 and φ ∞ , while only considering the regions below the
water-suspension interface in a settling plot. Modes of sedimentation are described (Bustos et
al. 1999) based on two definitions:
Definition 1: If φ (x, t ) is an entropy solution of problem, the pair (φ , f b (φ )) establishes a KSP.
The KSP is completely defined when a constitutive equation is postulated for f b (φ ) (Bustos
and Concha 1988).
Definition 2: Modes of Sedimentation (MS) are the different possible types of KSP. The
seven types of concentration profiles existing in a sedimentation column for small times
t 0 > 0 when proceeding downwards from x = L to x = 0 .
Each mode of sedimentation is introduced by an explanation and figures by Concha and
Burger (2002). In this analysis, the Kynch batch flux density function is allowed to have two
inflection points. As a result, the bulk suspension is separated from the supernate by a
rarefaction wave or concentration gradient. This analysis has greater versatility than that of
Lester et al. (2005), for the settling suspension. Although the modes of sedimentation by
Concha and Burger (2002) give a wide range of explanations on sedimentation, unfortunately,
they do not consider the compressible nature of the sediment itself.
32
None of the sedimentation modes are adequate to describe the physical solution of batch
sedimentation behavior of compressible materials.
2.5 CHANNELLING IN BATCH SEDIMENTATION
Channelling is the formation of macroscopic pathways in the sediment which may short-
circuit normal consolidation processes by means of upward movement of fluid carrying
particles towards the top of the settling bed. Channelling is a process involved in many
practical applications. Nevertheless, it is an unwanted and unexpected behavior during batch
sedimentation. Although it is possible to find information and examples in the literature on
this issue, the role of channelling in batch sedimentation and its relevance in thickening theory
is still not well understood. Moreover, possible reasons for channelling and the effect of
channelling in gravity thickening are still unknown subjects being studied by researchers.
2.5.1 Possible Causes of Channelling
In the literature, a few common reasons have been postulated by different researchers to
describe possible causes of channelling. In an actual thickening operation, channelling is
dependent on a large number of operational variables in addition to the nature of the particular
sludge involved. This project concentrates on the channelling characteristics of the slurry
itself and ignores the effects of operational variables such as stirring or amount of flocculant
added. From the literature, it is possible to gather some clues for probable causes of
channelling:
a. The first possible cause mentioned was the breaking of the solid matrix by impurities
and air bubbles (Glasrud et al. 1993) that break flow paths through the sludge bed. A
layer of sludge that has been bombarded by such material would have punctures which
might provide seed tunnels for later channels.
b. Instabilities, mentioned by Tory and Shannon (1965), could be of the type formed in
33
fluidized beds. In fluidization work with small particles (<30 microns effective
diameter), bed expansion is very small, the solid mixing range is very low, and some
inhomogeneities take the shape of channels.
c. The possibility of shear failure of the compression region (Vesilind 1968).
d. Effect of stirring studied by researchers such as Comings et al. (1954); Dixon (1979);
Holdich and Butt (1995) and Park et al. (1983).
e. Some authors have studied the cylinder wall effect. Dixon (1979) presented the
possibility that channelling is the result of wall effects.
f. A few references in the literature propose that the degree of channelling and/or size of
the channels are perhaps responsive to changes in the flocculant nature of a given
suspension (DeBoer 1990). More information can be found in the literature such as
Chandler (1983), Dell and Keleghan (1970), Dell and Kaynar (1968) and Chakravarti
and Dell (1970).
g. The size, shape and density of the solid particles are factors that affected channelling
(Leva 1959).
2.5.2 Mechanism of Channelling
Channels are first noticed near the side of the column in the lower initial solid concentration
and easily seen with the naked eye as cracks on the wall of settling bed. Water is allowed to
escape the settling slurry at a high rate during channelling which occurs during the batch
thickening of concentrated slurries. Holdich and Butt (1997) identified this zone as a very “ill-
defined region” where the suspension displays the properties of a connected network.
However the channelling may also be verified by the zones of constant concentration, and
variable concentrations owing to the propagation of concentration characteristics and
sediment, similar to the description of zone settling. Coe and Clevenger (1916) recognized the
existence of channels in the compression zone of their batch test in the early twentieth
century. Scott (1968a) stated that channelling occurs with slurries of intermediate solids
34
concentrations for his material. He assumed that the growth of channels is the primary reason
for the induction period. When the settling flocs almost touch each other in the vessel,
channelling occurs. This explanation could be considered reasonable since the hydrodynamic
resistance increases rapidly when the water path between flocs becomes quite narrow (Chen et
al. 1996). Fitch (1979) stated that when the slurry entered into the thickeners, the particles
started to settle, initially at low concentrations. After the slurry became concentrated it
exhibited short-circuiting or channelling. In the batch sedimentation process, after initial
mixing, the particles appear to be individual solid particles which mean they are not
agglomerated. However, a short time later, the particles come together and form flocs. Scott
(1968c) indicated that the diameters of channels in intermediate slurries, developed during the
induction period, were of the same order of magnitude as the diameter of the flocs.
Interestingly, the channels were wide enough so the solid particles could be seen traveling
upward in the channel.
In general, there are two types of flocs that form depending on concentration values; small
flocs are formed in low concentration and large flocs are formed in high concentration
slurries. O’Donnell (1972) describes low concentration slurries as being <200 g/l of Calcium
Carbonate (CaCO3). On the other hand, Jones (1986) mentioned that no channels occurred in
the experiment with calcium carbonate if the concentration value is less than 325 g/L. In the
literature, high concentration values are given as > 400 g/l CaCO3 (O’Donnell 1972), and 500
g/l CaCO3 (Vesilind and Jones 1993) for batch settling. Scott (1968a) mentioned that he did
not observe channels in batch tests at higher concentrations. A visualization of the formation
of channels is shown in Figure 2.11.
35
a. Separation of large flocs b. Breaking of flocs
(Formation of weaknesses)
c. Flow paths formation d. Consolidation of walls
e. Solidifies walls f. A well established channel
Figure 2.11 - Onset of channel formation (O'Donnell 1972)
36
At the same time as the slurry stops all movements at low concentration, the flocs become
interconnected in order to start the settling of the bed as one individual floc. When this
happens flocs of a much larger size are formed and the connection between the flocs appears
to reach the bottom of the bed.
It has been supporting the weight of the solid to some extent and this behavior is called
compaction. The floc structure is not consistent here as it contains many unequal shapes and
fractures. At this point, the beginning of the formation of channels has already become
evident as small streams appear, showing significant movement between the flocs. The large
flocs are carried upward by established channels and the consolidated region forms very
quickly. For a low concentration slurry, the channels are approximately similar in diameter
and also appear to have approximately the same distribution throughout the bed.
Scott (1968c) observed that the channel diameters in intermediate slurries are of the same
order of magnitude as the diameter of the floc during the induction period. In this stage,
settling proceeds at a decreasing rate as the liquid is released and continues until the channels
close.
The paths of the liquid strings are still apparent even after the walls of the channels have
completely collapsed against each other (Harris et al. 1975). When the solid particles settle,
the high-pressure forces the liquid in the compression zone to move through the solid region
towards the surface. The best indication of this is cracks appearing in the bed. Font (1991)
stated that after the initial period, the presence of the channels inside the sediment was
observed experimentally as particles settle in the bed. When channelling appears in the early
part of compression, its behavior is governed by major parameters such as the settling velocity
and the resistance of each phase to compression for the particular concentration at which the
37
channelling occurs (Jones 1986). Somasundaran (1981) indicated that air bubbles act to
change the physical features of the slurry such that water seepage becomes possible through
more channels. The rate at which the channelling zone travels upwards increased with
increasing dosage of flocculant (Holdich and Butt 1995). Channelling occurs initially in the
top segment of the suspension and moves downwards slowly. The channels are formed by
rising air bubbles and when degassed, the settling rates are reduced significantly.
Coad and Ives (1981) affirmed that there are three different channel forms. The first one is
called “Major channels” or “wormholes” (Ives 1986) offering the least resistance to flow and
operating at the highest velocity that also allows little deposition. The bed structure around
these channels is mostly compact. The second type is a minor channel that operates at average
velocity in the cracks and helps to deposit the solid effectively. Because of the deposition, the
minor channel size is reduced. The last type is a simple flowpath. Sometimes it is possible to
see a horizontal channel and joined channels as well. It has been noted that the channels
formed at high initial concentrations were several centimeters long in each experiment. Figure
2.12 depicts the joining of channels.
Knoer (1983) indicated that the existence of the channels ought not to take away from the use
of bench scale experimental data for continuous thickener scale-up, as long as the appearances
of the channels in the sediment are fairly reproducible. No channels were present when the
slurry was separated continuously in the same suspension used during batch sedimentation,
although dilution of the upper region was reported (Kos 1978). Tory and Shannon (1965)
described the channels as “instabilities” and described the appearance of channels and
interface volcanoes in calcium carbonate sludges.
38
Figure 2.12 - Joining channels (O’Donnell 1972)
It is possible to recognize the channels by two means. First, is the actual appearance of the
channels whether they are vertical channels, diagonal channels, or volcanoes at the solid-
liquid interface. Second, is the doubly concave flux curve, which is the result of abnormally
high interface settling velocities. Since these two phenomena have frequently been observed
to have a strong correlation during thickening research, a causative relationship between the
two has generally been accepted (Jones 1986). Holdich and Butt (1996) reported that
channelling during batch sedimentation led to a considerable increase in the settling rate of
the suspension/supernatant interface, and hence, an increase in the batch settling flux.
According to Vesilind and Jones (1990) channelling during the settlement of calcium
carbonate suspensions shows that:
a. Channels only appear at intermediate concentrations,

b. The upper region of the suspensions which is near the supernatant interface is
diluted,
c. The velocity of channel zone was constant,
d. It was possible to see the vertical channels formed just above the compression
zone,
e. The zone of channelling appeared higher in the suspension when the initial
concentration was higher.
39
The settling velocity increases in the region of the thickening slurry containing the channels
moving upward towards the descending interface. Scott (1968) however, indicated that
channels did not need to be appearing on the surface of the sediment suspension to change the
rate of decent of the settling interface. When the channels become visible on the surface of the
settling bed, they appear as a crater of a volcano.
2.6 VOLCANO FORMATION
When the channels finally pass upwards through the bed and meet the surface due to the non-
uniform structure (Chen et al. 1996), volcano formations are observed on top of the bed
indicating the mouth of each channel (Figure 2.13). Channels pass solid out of the mouth of
the crater to fall back to the surface and build up the walls. Only individual particles are being
carried upwards, with the occasional small floc at this stage of the process.
Figure 2.13 - An appearance of the volcano formation on the settling bed surface
(Tonmukayakul 1998)
40
The volcanoes can form cones of considerable size once the consolidated region has reached
the surface.
During ideal gravity thickening, liquid flows upward through aggregates and around
individual particles and flocs in uniform seepage across the horizontal cross section of the
thickening device. However, tests have shown that some sludge will release liquid in streams
which are few in number although large, when compared to uniform seepage flow paths.
When viewed in transparent settling columns miniature volcanoes are noticeable at the solid-
liquid interface and small streams flow along the outside of the sludge (Dixon 1982).
2.7 REVERSE CONCENTRATION GRADIENTS
Reverse concentration gradients have been termed as inverse concentration gradients or local
concentration gradients by Tory and Shannon (1965), O’Donnell (1972) and Fitch (1983).
Reverse concentration gradients can be described as the existence of higher concentrations
above lower concentrations, caused by solid flowing upwards from the settled bed at the
bottom of the settling column (Figure 2.14).
Shannon et al. (1964) reported “a rising concentration gradient with its intersection with the
liquid-slurry interface that accounts for the nonlinearity of settling curves”. Tory and Shannon
(1965) affirmed that the concentration gradients may occur for a very narrow range of
concentrations. The necessities of this condition for their material are: firstly, the final
packing particulate suspension void ratio is less than 0.42 and secondly, the settling of
particles is in the low Reynolds number range (for laminar flow).
At the beginning of the settling tests, compression forces are very small with respect to
gravitational force. Tory and Pickard (1982) observed that a concentration gradient formed
41
below the interface and this gradient expanded with time but they did not measure its position.
They also indicated that a certain length of time is required to establish a stable concentration
gradient from the interface to the constant concentration region. The highest initial
concentration examined by Gaudin and Fuerstenau (1962) was 16.6 weight percent solids of
CaCO3 slurries. The existence of reverse concentration gradients generally occurs near this
percentage. An imbalance in the floc internal forces and other forces produced by rising fluid
caused “a roof” or “bridge” to form at the top of channels. Fitch (1966) attributed the
formation of reverse concentration gradients to solids movement opposite to the average
direction and air bubbles.
120
30 min 90 min
100
150 min 200 min
Sediment bed height (cm)
80
220 min
60
40
20
0
0 200 400 600 800 1000 1200
Concentration (g/L)
Figure 2.14 - Reserve concentration gradient profile in different times

(Tonmukayakul 1998)
The solids had yield strength, which was enough to hold up the roof of the bubbles. The flux
of solids out of the region may be greater than the flux into it. The greatest flux would occur
at the lowest concentration when the reverse concentration gradient range is beyond the hump
of the flux curve, which would be at the floor of the bubble.
42
3 CHAPTER 3
3.1 EXPERIMENTAL SETUP
It is not possible to simulate all phenomena occurring in a continuous thickener through the
use of one experimental pilot plant. Various investigators employed equipment for batch
sedimentation that differed significantly in size, design and in the control of the process.
Additional accessories and apparatus have also been used to observe different effects on the
thickening process. However, there are some important limiting factors determining the actual
size of thickening units and their equipment, such as laboratory space, cost of research,
volumes of handled sludges and investigation period. In the case of batch sedimentation, the
material properties of the suspension and the settling time also affect the settling process. For
instance the sludge with high permeability and low compressibility settles to a lower height.
On the other hand, the sludge with low permeability and high compressibility will settle in a
short period. This means that the test condition is very important in predicting the settling
rates in thickeners. SVI is the percentage of solids that settle after a certain amount of time in
batch settling tests (Kopp and Dichtl 2001). However, SVI has limited application in
predicting settling rates in thickeners at the test conditions, and thus not considered in
practical applications.
A new method developed at the University of Melbourne which contains an experimental
method and analysis techniques were developed by de Krester et al. (2001) and Lester et al.
(2005). The method is used to determine the fundamental compressional rheology parameters
of Py (φ ) , φ g , R (φ ) and D (φ ) for a range of sludges. Since these properties vary by large
orders of magnitude with concentration, they were measured over the entire operational
volume fraction range within the device to be modeled.
43
This chapter explains the laboratory, batch settling devices and measurements, software used
for recording data, information about material used, particle size analysis, sample preparation,
flocculants, determination of sediment concentration, and the technique of stepped pressure
filtration, as well as related issues such as laboratory conditions.
Batch settling is probably the simplest way to separate a flocculated suspension into
concentrated sediment and a clear liquid with the intention that the sets of experiments are
organized to determine the occurrence of channelling behavior for range of initial solids
concentrations.
This chapter outlines the laboratory techniques of batch settling and discusses general issues
related to the equipment used. The method of experimental work is first introduced. The
apparatus consisted of a settling tube, mobile platform, position potentiometer, the gamma
radiation densitometer and a radiation detector. Secondly, a description of the datataker and
associated software, material used and flocculant is provided. Finally, calculations for
calibrations have been summarized in terms of experimental procedure. Experimental
apparatus was designed to measure concentration within settling beds without disturbing the
settling process. The monitoring equipment moves at a constant rate upwards and downwards
along the settling tube to capture accurate data during the continuous settling process.
3.1.1 The Settling Tube
The major item of equipment used in the study is the laboratory size settling tube, shown in
Figure 3.1 as No 1. It is a long, vertical, cylindrical Perspex tube.
The overall height of the tube is approximately 180 cm and has an inside diameter (ID) of
15.24 cm. It is sealed at the bottom by a P.V.C. plastic plug covered by a layer of epoxy resin.
The hollow tube is held on a 5cm-thick Perspex plate on the top of a 5mm-thick aluminium
plate which is fixed to a sliding unit. The support structure was bolted to the laboratory floor.
44
To enable visual inspection, the tube has been made from clear Perspex, which will also allow
efficient passage of γ- radiation through it. The manual mixer has been located just above the
tube for mixing of the slurry. A diagram of the laboratory apparatus is shown in Figure 3.1.
3.1.2 Mobile Platform
The mobile platform, shown in Figure 3.1 as no 3, provided a secure base for the radiation
source and detector to operate. The platform traveled continuously over the height of the
sediment column with a constant speed of 10 cm/min. Measuring the concentration of slurry
at any depth in the column, the passage of the platform along the settling tube allowed the
radiation unit to beam through the settling bed. The mobile platform was operated by level
limit and by-pass switches located at both the top and the bottom of the column for changing
the direction. Fail-safe switches were fitted to the rig to switch off the power supply to halt
the movement of the platform in case the limit switches failed to activate.
3.1.3 Gamma (γ) Radiation Densitometer and Radiation Detector
In order to study the concentration distribution changes during batch settling, a method is
required for the continuous measurement of density over the height of the sedimentation tube
without disturbing the slurry.
In the literature, there exist a number of transducer systems available for measuring density
over the height of the sedimentation using numbers such as radiation (x-rays, γ-rays) and
60
optical and ultrasonic transducers. For instance, Sawai et al. (1990) used Co gamma rays
and electron beams to measure the sedimentation and dewatering of sewage sludge.
45
1
4
2
6
3
1 Settling Tube
2 Gamma Radiation Source
3 Moving Platform
4 Count Rate Meter
5 Platform Controller
6 Potentiometer
Figure 3.1 – Laboratory Size Batch Sedimentation Unit
46
Different materials possess different reduction factors for radiation beams. For our experiment
a Caesium-137 γ-radiation source shown in Figure 3.1 as no 2 was used. The absorption of
monochromatic photons passing through a uniform medium can be described by Lambert’s
Law:
I = I 0 e − λd 3-1
Where I is the intensity of the beam (photons/m2-sec), I 0 is the incident intensity of the beam
(photons/ m2-sec), d is the distance in the medium (m) and λ is the linear attenuation
coefficient of the medium (-). When gamma rays are passed through a settling column, they
penetrate the liquid, the solids and the wall of the Plexiglas column as shown in Figure 3.1.
Equation 3.1 can be rewritten as:
− ( λ c d c + λ f d f + λs d s )
I = I 0e 3-2
with
λ f = ε s λ s + (1 − ε s )λl 3-3
Where dc is the thickness of the wall (m), df is the inside diameter of the column (m), c refers
the column wall, f refers to the suspension, l refers to the liquid and s refers to the solids. The
schematic description of gamma rays passing through settling column is given in Figure 3.2.
In the experimental set up, the radiation beam was directed through the settling tube to a slit
25 mm wide and 5 mm high cut in a lead block which contained the detector. A thallium
activated sodium iodide crystal approximately 25 mm diameter and an ECKO scintillation
detector N618A detected and amplified the radiation beam. The electronic pulses gathered
from the detector tube were transferred to the counter and SCA (Single Channel Analyser).
47
The background radiation was eliminated with the ‘Baseline and Window’ by the SCA
(Tonmukayakul 1998).
Channels
Detector
Gamma Source
Figure 3.2 - Operating gamma rays through a settling column
3.1.4 Measurement of Radiation and Rate Meter
During the experiments, the radiation signal is converted into an electronic signal and
integrated by a CANBERRA (lin/log) rate-meter model 1481L giving a time constant of 2
seconds. The equipment offers the user a choice of the average number of pulses per second
being received at the input, and either a linear or logarithmic presentation via a front panel
meter. The collimator’s slit dimensions are given in Table 3-1 and the rate meter is shown in
Figure 3.1 as no 4.
Table 3-1 - Collimator slit dimensions
Sections Dimensions (cm)

Length of Slit 1.3
Width of Slit 2.5
Height of Slit 0.5
48
The negligible background count rate at working conditions was measured as approximately
20 pulses per second. It was eliminated using the threshold discriminator. The minimum and
maximum count rates observed in experiments were approximately 13000 and 18000 pulses
per second, respectively. For improved accuracy, a circuit was used to deduct a base value
and amplify the remainder of the signal since the concentration signal from the rate meter
included a very large base value. The Count Rate Meter and Mobile Platform Controller Unit
are shown in Figure 3.3.
4
2 1 3
1 Start Up Button
2 Start Down Button
3 Stop Button
4 Platform Controller
5 Count Rate Meter
Figure 3.3 - Mobile platform controller unit and the countrate meter
3.1.5 Data Collection
In this investigation, a Datataker Model 50 (Figure 3.4) was used to record the data. “The
DT50 features 5 to 10 analogue channels depending on sensor type, five digital input
49
channels, 3 high speed counters and a sampling speed of 25-70 samples per second”
(www.datataker.com, 2004). The data sent from the data taker is stored in the PC utilising the
appropriate software. The datataker was connected to the computer by communication cable
(RS 232).
1 RS 232 Port
2 Channels
Figure 3.4 - The datataker unit used for recording data from experiments
Voltages were measured every second by the datataker as signals and processed by the data
logger. The software was configured according to the required data. Figure 3.5 shows a block
diagram of the software program and the Delogger software interface is shown in Figure 3.6.
As shown on the software and the datataker, the sensors for measuring height and
concentration were connected to channel numbers 1 and 2 which were analogue channels
enabling the voltage to be measured.
50
Figure 3.5 - Datataker software program-Delogger
Figure 3.6 - DeLogger program used for collecting
51
3.2 SUSPENSION MATERIAL
3.2.1 Sample Preparation
In order to keep settling characteristics, the samples for experimental study were treated with
care prior to experimental testing. Particular issues of importance are listed below:
• All experimental material were kept in the proper container
• The experimental material was kept at dry and shadow condition at room temperature
• The initial concentration was determined from oven-drying
3.2.2 Determining Concentration
The solid concentration for the suspension was determined by taking samples from the slurry.
First, two jars were weighed as empty by using an analytical balance and the values were
noted. Then the slurry was well mixed for 20 minutes and then the samples were taken from
the mixed slurry. The two jars full of samples were weighed and noted. Then the samples
were filtered from the filter and the liquid was removed. The final cakes were placed in the
jars and oven dried overnight at 90ºC. The samples were left to cool down the next day and
reweighed. Water and material amount were determined, and then using simple ratio
calculation, the concentration of the suspension in the settling column was calculated.
3.2.3 Solid Particle Properties
In order to describe the material behavior sufficiently for solid particle properties that have
been used for the experiments, it was required to have certain physical properties such as
good Gamma Ray absorption, long-term stability or immobility and easy disposal of waste. In
this research, due to the presence of channelling and the adequate physical properties, calcium
carbonate was chosen as a research material for all experimental studies. The calcium
carbonate sample which has a mean particle diameter of 26 µm and a density is 2590 kg.m-3
was supplied as dry solids in 25 kg bags by Omya Australia Pty. Ltd.
52
The various properties of the sludge measured by the manufacture are listed in Table 3-2.
Table 3-2 - Properties of calcium carbonate used in the research

(Omya Australia Pty. Ltd. 2001)
TYPICAL CHEMICAL ANALYSIS Typical

Calcium Carbonate CaCO3 % ≥ 95.0
Magnesium carbonate MgCO3 % ≤ 4.0
Acid Insolubles % ≤ 1.5
TYPICAL PHYSICAL PROPERTIES
Brightness (DRY) % 91.0
Oil absorption g/100 g 11
DOP absorption g/100 g 23
Specific gravity 2.7
Moisture content % 0.2
Bulk density, loose g/cc 1.0
tapped g/cc 1.8
Sizing Analysis
Max particle size microns 220
Mean Particle Diameter microns 26
Residue on 250 µm screen % 0.2
Particle size distributions were determined using the Malvern Mastersizer X and the
distribution of wet and dry particles with more detailed results are given in Appendix E. From
the wet particle analysis, it can be seen that 80 % of the material has particles under 30
microns in size. The dry particle size analysis shows that 80 % of the material is less than 80
microns in size. This is because when the particles are dry, they generally stick to each other
by electrostatic attraction. Larger particle size is usually obtained from the dry particle size
analysis.
3.2.4 Flocculation and Flocculant
Flocculation is used to improve the settling and filtration rates of a particulate suspension by
changing the nature of the suspension without significantly compromising compressibility. It
helps to reduce the fine particulates in the overflow or filtrate. The reduction of fine
particulates could cause two results; first, a higher output solid concentration; second,
53
increased throughput to be achieved in a dewatering operation such as filtration or thickening.
The mechanism of flocculation can be described as a combination of small particles to form
larger agglomerates called flocs that have enhanced settling rates at low solids concentrations,
and a more permeable suspension network structure at higher solids concentration.
The flocculants are mainly chemically additive which have different effects on the
flocculation process according to the dose, chemical nature and the molecular weight of
flocculants which all have significant influence on floc structure.
The objective for adding flocculants is to encourage floc formation and thus enhance settling
and compression rates. However, choosing the right type and amount of additives for the
sludge processing by solid-liquid separation is an important problem for the industrial
applications. The selection of the flocculant and required dose is often based on
measurements of settling rates.
Flocculants cause a dispersed system to coagulate and form flocs. They are used widely in
mineral processing and in wastewater treatment. The purpose of using of flocculant is to
improve the sedimentation rate. The different characteristics of flocculants are vital for the
settling effect required on a system. Akers (1976) proposed a reaction scheme for polymer
flocculation. Onal et al. (2003) stated that Superfloc was very effective as a flocculant in the
sedimentation of clays. Weir and Moody (2002) researched the importance of flocculant
choice to achieve efficient solid/liquid separation.
The flocculant used in this study was Teepol, a commonly liquid detergent, made by Shell
Chemicals Ltd, and it is regarded as a ‘household’ detergent and also as a surface-active
agent. The ingredients of Teepol are ionic and non-ionic surfactants that are also
biodegradable. It is one of the optimum flocculating agents for calcium carbonate slurries, at a
54
specific concentration, and this was also supported and emphasized following the work of
O’Donnell (1972). It is a surface-active agent containing long chain hydrocarbons and is
biodegradable. The main criteria for choosing Teepol was its long life, low cost and
recyclability.
3.2.5 Calibrations
The readings collected from the experimental unit were in signal forms. In order to work with
numerical values, it was necessary to convert the signal values to numerical values. Two
curves were prepared for the calibrations, one for height and the other for concentration. The
detailed information is given in Appendix F.
3.3 EXPERIMENTAL PROCEDURES
All batch settling experiments were performed in the transparent cylinder settling tube
introduced earlier in order to enable visual identification of channel formation.
The first step for the experiment was to calculate the required slurry concentration. Based on
the volume of the settling tube, the required amount of CaCO3 was determined and placed in
the settling tube. The tube was then filled with distilled water up to the uppermost point to
which mobile platform could travel. During the filling process, the end of the filling hose was
positioned approximately 2 cm above the top part of the column to avoid of turbulence in the
column. The rising slurry in the column eventually reached the requested level of the column.
The platform moves at 10 cm/min (chosen as the optimum speed for data collection). The
slurry is then mixed well for 20 min with a long vaned Perspex rod. The mixing was carried
out for approximately 20 min in order to guarantee that all the solids were completely in
suspension. The bottom of the tank was checked with the rod to insure that the solid was
totally being suspended rather that being sat at the bottom of the column. At the end of the
55
mixing, two samples were taken from the settling tube for analysis of concentration. They
were first weighed and then dried at 90ºC overnight in the oven. The dry samples were
weighed again the next day.
The second step was to add the flocculant. The optimum type and amount of flocculants was
decided through separate experimentation. Four 1000 ml capacity settling cylinders were
prepared with slurry concentrations of 250 g/l, 500g/l, 750 g/l and 1000 g/l.
Different amounts of Teepol (25ml/l, 50 ml/l, and 100 ml/l) were added to the cylinders and
stirred for 5 minutes. It was observed that after almost 1 hour all the particles had settled
when 50 ml/l of Teepol was used. The experiments showed that increasing the Teepol
concentration beyond 50 ml/l did not affect the settling characteristics of the slurry.
The third step was to set up the software (Delogger) and load it on the computer in terms of
program, chart, text and form (Figure 3.5). The datataker was switched on and the program
was connected to the datataker. The datataker measured both the height and concentration
signals every second and logged them into the computer.
The fourth step was stirring the slurry to remove all the caked solids from the bottom. The
stirring was achieved by a mixer in the form of a long perspex rod with attached vanes. Using
this mixer the slurry was agitated by hand for 20 min. Tests had shown that increased stirring
time did not influence the settling in any way. Also a color chemical which has no effect on
the process was added to create a contrast between the slurry and the settling bed. At
completion of mixing the stirrer was withdrawn from the top of the settling tube and clamped
a few centimetres above the slurry surface to drain. The recorder chart was switched on one
minute before the experiment began. After this, the mobile platform was started. The
56
movement of the platform was mostly initiated in the upward direction, but not always. The
datataker, the software and the mobile platform was left running until the there was no
detectable change in sludge height, which took approximately 14 hours. Using the control
program on the PC, the data from the experiment set was logged and analyzed to provide run-
time information on the progress of experiments.
3.4 STEPPED PRESSURE FILTRATION
A new stepped pressure filtration technique (de Krester et al. 2001b; Usher et al. 2001) was
developed at The University of Melbourne. The stepped pressure filtration is employed to
determine the compressive yield stress, Py (φ ) , and hindered settling function, R(φ ) . The data
are determined as a function of the solids volume fraction, φ , for flocculated suspensions over
a number of hours. Standard constant pressure filtration requires five or more individual
filtration tests to characterize a suspension across a range of solids concentrations. However,
the characterization of flocculated suspensions is an experimental process which often takes
days to perform. A new stepped pressure filtration technique has been developed to reduce
characterization time.
Using only two filtration tests and a wide range of pressures from 5 kPa to 300 kPa, the
stepped pressure filtration test can characterize suspensions. The base plate and filtration
cylinder are water jacketed allowing thermal control from 10-90ºC. Nevertheless, there is the
potential to build a filtration test rig with different equipment that operates at lower pressure,
at higher pressures or at higher temperatures (Usher 2002). Characterisation can be
determined by using only a small amount of sample which usually takes of the order of 1-8
hours depending on the permeability of the suspension. In order to keep the right of accuracy
of the technique, the computer and semi-automation method is used.
57
The positive aspect of the filtration test rig is that it can be disassembled for transportation,
enabling use onsite. The limitation of this technique is that it is mechanically complex, the
equipment has been expensive to develop and the results are only produced for high solids
concentrations.
3.4.1 Stepped Pressure Filtration Technique
A different form of the classical filtration test is the step technique which makes efficient use
of time in terms of equipment use and labor. The main difference is that the new method uses
only one stepped pressure compressibility filtration test and one truncated stepped pressure
permeability filtration test to determine Py (φ ) and R(φ ) . Five or more solids volume fractions
substantially decreases the time required for sample characterization. It is possible that using
the data from these same filtration tests, the solids diffusivity, D (φ ) , can be determined at
these same solids concentrations. The compressibility is measured at a range of pressures in
the first test.
In the compressibility test, the process starts with the lowest pressure applied until the filter
cake stops exuding liquid and then the pressure is stepped and continuous. The compressive
yield stress Py (φ ) for a number of solids volume fractions is determined using data obtained.
In the permeability test, the process starts when the first pressure is applied until a specified
gradient is stable, and then the pressure is incremented. In order to calculate the hindered
settling function, R (φ ) , the traditional calculation is used from the gradient of a t/V versus V
plot for a single applied pressure (where t is the time of filtration and V is the specific volume
of filtrate).
3.4.2 Apparatus and Measurements
The piston driven filtration test rig is a laboratory scale computer controlled filter press that
was used for the rapid characterization of the dewaterability of suspensions. The schematic
58
diagram and photograph are shown in Figure 3.7 (de Krester et al. 2001). The key
components of the piston driven pressure filtration test rig included:
• a water jacketed filtration cylinder and base plate

• a permeable sintered disk
• a pneumatic cylinder
• a pressure controller
• a linear encoder
• a pressure transducer
• a 100 psi air supply
• an electronic data acquisition system
• a personal computer
The suspension sample between an air driven piston above and a permeable membrane below
is contained by the filtration cylinder. The water jacket is facilitated by temperature controlled
testing and the filtration cylinder and base assembly were adapted to incorporate with it. The
filtration rig employed a pneumatic cylinder to apply pressure to the suspension sample
contained in the filtration cylinder. The liquid is forced out through a filter membrane at the
cylinder base. Stepped Pressure tests are required for the computer control of the pressure
stepping and it was facilitated by a Bronkhorst EL Press pressure controller, which controls
the pressure in the pneumatic cylinder. Pressure range can be changed from 5 kPa to 300 kPa
which can be tested by the pressure control system. The upper temperature limit for the
pressure transducer is 110ºC but the reliable temperature is 90ºC. Using the control program
on the PC, the data from the pressure transducer and linear encoder was logged and analyzed
to provide run-time information on the progress of experiments and automatic control. The
necessary information for use in the subsequent determination of the dewatering parameters is
generated as a final data output file in the form of Excel spreadsheet at the end of each test.
59
Linear
Encoder Pressure
Controller
P
Air Line
Double Ended
Pneumatic Cylinder
Bleed
Line
Stainless
Steel Piston
Pressure
Stainless Steel Transducer
Filtration Cylinder
Membrane
Beaker PC
Electronic Balance
a. The schematic diagram of pressure filtration test rig
b. Photographs of the stepped pressure unit
Figure 3.7 - The schematic diagram (a) and photographs (b) of the stepped pressure
unit in The University of Melbourne
60
Normal batch filtration at the lowest test pressure was allowed to proceed until the piston
movement dropped below a very low rate in the compressibility test for determining Py (φ ∞ )
shown in Figure 3.8. First, the pressure was increased to the next suggested value and then
kept constant until equilibrium was again reached.
Second, by generating a series of equilibrium heights, the experiment is repeated over a range
of pressures, which were converted to φ ∞ values. In order to give a complete Py (φ ) curve, the
results were combined with measurements of the gel point. As a conclusion, the overall
operating time was reduced by loading fewer samples.
Figure 3.8 - Example of t versus V 2 results for the stepped-pressure compressibility test
Using normal batch filtration and determining Py (φ ) , the lowest test pressure was allowed to
proceed until the piston movement dropped below a very low rate in the compressibility test.
The time at each pressure was slowly increased such that the pressure was stepped when the
61
slope of t versus V2 was stable in the permeability test given in Figure 3.9.
The average slope was taken by the control software using defined number of preceding data.
The stepped-pressure compressibility test took four to five hours as indicated by the time
scale while the permeability test took one to two hours for typical water treatment sludge.
However, the two tests were easily performed in a day.
Figure 3.9- Example of t versus V 2 results for the stepped-pressure permeability test
The file in the form of an Excel spreadsheet was generated at the conclusion of each test with
a final data output, providing the necessary information for use in the subsequent
determination of the experiment parameters.
62
4 CHAPTER 4
4.1 PHOTOGRAPHIC STUDY
Although many researchers have observed channelling during sedimentation, the influence of
channelling on the settling bed has not been clarified completely, because it is difficult to take
the required measurements without disturbing the system. Auzerais et al. (1988) stated that to
examine the concentration or consolidation of stable and flocculated suspensions of silica
particles without disturbing the process required a sophisticated technique.
Graphical methods have been applied to batch experiments to simulate changes in the settling
bed during the process (Concha and Burger 2002). Many reports have been published were
based on observations (Dell and Kaynar 1968; Leenvaar and Rebhun 1983; Allen and Uhlherr
1989; Glasrud et al. 1993; Curran et al. 2003). They explain channelling with settling height
versus time curves that are S-shaped (O’Donnell 1972; Kos 1978). However, no direct visual
evidence was presented. Most of the studies did not mention the dimensions of the channels
(Glasrud et al. 1993). For this reason, it was thought that it would be helpful to collect more
information and help to describe the behaviour. The advanced visualization methods such as
using photography and video systems are being looked at to gather more information about
channelling, floc structure during batch settling, measurement of settling heights, effect of air
bubbles, flocs sizes and the flocs velocities in the channels. In addition, these parameters were
observed by recording the conditions in which channels form and propagate, and the
channelling influence on the structure of settling bed.
The earliest photographic observations were made by Dell and Kaynar (1968). Allen and
Uhlherr (1989) used visual observations by means of hand tracing original photographs that
63
are transformed into dotted patterns, where contrast is represented by different concentrations
of dots. Recently, Curran et al. (2003), Novales et al. (2003) and Glasrud et al. (1993) have
used photographic techniques and video systems for their studies.
4.2 PURPOSE OF WORK
In this chapter two photographic investigations are employed. First, the settling phenomenon
was observed with a manual camera over long periods of time, which is called the time-lapse
method. The aim of this study is to record information continuously and accurately so that
quantitative data can be obtained. Using different photographic methods, the settling bed
movement and settling heights can be recorded. The information collected from direct visual
observations has been compared with data collected from other experiments indicating it is a
very accurate method. Streak photography will be introduced in section 4.3. It can be
described as a method of recording all information/movement in front of the lens. As the
height of the settling bed changes, it is continuously recorded through a slit aperture onto a
moving film. A similar method is used for finish line photographs at racetracks. Time lapse,
in contrast, may capture one photograph every hour for a number of hours resulting in a
moving sequence. However, this is an intermittent recording method with most of the event
itself unrecorded on film. If the object changes between photos, this will not be detected by
the time-lapse method.
The second photographic technique is high speed videography (Section 4.5). The aim of this
part of the photographic study is to observe the settling phenomena, channelling, volcano
formation, measure floc sizes and floc velocities in the channels, and also to determine the
width and length of the channels. These photographic observations will record the information
in detail. The video record enables events in the channels to be observed over the whole
period of settling.
64
The video record can be re-played for detailed inspections such as freezing the image (or
framing) and comparing with other real-time image frames. By using this method it is
possible to observe channelling and floc behaviour at many positions.
4.3 STREAK PHOTOGRAPHY
‘Streak’ photography is a method which uses transportation of the image or the film or
another recording surface across a slit in the camera. The function of streak photography is to
limit the field of view to a single slit or to a line in front of the lens.
A visual element of the photograph displays time parallel to the direction of film transport. It
also produces a photograph with a length proportional to time so that the final photo is a
record of what has occurred in front of the camera over time. This type of photography is a
mixture of both motion photography and regular still photography. However, the images are
usually hard to recognize as a visual element of the photographed scene because of it is being
‘streaked’. Streak cameras are an adaptation of strip cameras. The image of an object in strip
photography usually moves at the same speed and direction relative to the film transport. In
our case, the direction of motion or change is perpendicular to the direction of film transport
and the motion of any point is recorded as an angular streak on film. The process of recording
sedimentation occurs over 7–10 hours. During this process, the slurry surface falls slowly.
Streak photography enables a continuous, real-time record of time vs. height along the length
and height of the film. The image size will be determined by the rate of transport per hour of
the film and the vertical height of the sediment tube (Perspex).
4.3.1 Equipment for Streak Photography
The equipment used for this process comprised a 35 mm camera, winding mechanism, timing
device, a light source, power supply and some accessories. To photograph the sedimentation
65
process, a system has to be reasonably transportable and versatile to accommodate different
situations. A slit was placed directly in front of the film plane for the camera and the film was
transported at a known and constant rate past the slit. The lighting for this project has to show
the tube enough to give a good image. It has to be positioned to avoid reflections from the
Perspex tube which would hinder the view of the slurry through the glass.
4.3.2 Camera and Film
The camera itself is a manual camera with a rewind handle and depressible clutch. The
process of streak photography involves film usually longer than a single frame moving past
the slit. Therefore the film is rolled into the camera, and then re-rolled back into the canister at
a constant rate. The camera should be put into a light tight bag to advance the film to its end.
The film of choice was Kodak Professional Technical pan, which has variable contrast,
medium film speed, and can be processed in a variety of chemicals. The slit is placed inside
the back of the camera using two pieces of card. The card has to be placed accurately between
the two guide rails above and below the frame ensuring that the film does not rub the tape or
card, as it may produce small slivers of dust on the film. Once the film has been rolled onto
the camera in a light tight bag, the film is re-rolled past the slit with the shutters open. In order
to transport the film at a constant rate over a long period of time, a motor has to be used in
conjunction with the camera. The motor has to be adapted to turn the rewind knob of the
camera at an adjustable rate of turns per hour. The motors are required to operate under a high
torque situation. This is so that the motors will be capable of transporting the film under stress
at a constant rate. The camera set up for streak photography and schematic diagram are given
in Figure 4.1 and Figure 4.2 respectively.
66
35 mm
camera
Film Transport
Mechanism
Figure 4.1 - Camera set up for streak photography
TOP VIEW
Sedimentation
Cylinder
Camera Lens
Slit
Film Movement
Figure 4.2 – Diagrammatic explanation of streak photography process
67
A diagrammatic explanation given in Figure 4.2 shows that the slit limits the image on film to
1 mm wide, and motor of film “paints” sequential images and slit into a continuous
photographic record.
4.3.3 Bracket Housing
The camera needs to be placed on a bracket that holds the motors above the camera and
affixed to the tripod. By using an ‘L’ shaped bracket, the camera can be fixed to the
horizontal, motors to the vertical, and the tripod bolted underneath the camera. The mount has
to be quite rigid and strong or else the motors may be moved instead of the rewind knob. The
film transport mechanism is shown in Figure 4.3 and Figure 4.4.
Coupling to
Gearbox2 rewind crank
Motor
Gearbox1
Figure 4.3 - The bracket holds the camera in place under the gears
4.3.4 Timing Mark
The rate of film movement is reliant on the rotational speed of the take up spool. This can be
set at a constant and known rate. However as the film is taken up into the film canister past
the slit, the diameter of the take up spool increases, thus increasing the speed at which the film
is pulled past the slit. Accounting for this effect, a light marker or timer will have to be
introduced to make a record of time on the film. This can be done internally. Here a LED
marker which appears at the bottom of the photographic record (light emitting diode) is used
68
which is positioned in the view of the slit as a photographed object (Figure 4.5). This was
activated by a micro-switch that is located on the motor turning the rewind knob. It is left on
for approximately 3 minutes to expose a line into the base of the streak image.
Gearbox2
Coupled (very slow)

Motor (Fast)
to rewind
crank of
Gearbox1
(Slower)
Figure 4.4 - Motor and two gear sets rewinding the camera
LED
Figure 4.5 - The LED in the field of view of the camera slit
69
4.3.5 Power Supply
The power supply (Figure 4.6) provides variable voltage at a constant current. The speed of
the motor driving the film transport is directly related to the voltage supplied. The output can
be set anywhere between 0–12 volts. It was set initially at a starting voltage with which turned
the first geared motor 330 times to a single rotation of the second gear in one hour. The gears
also have to operate over an extended period of time and to maintain a constant rate.
Figure 4.6 - The 240v-12v converter, variable power supply
4.4 OBSERVATIONS OF STREAK PHOTOGRAPHY
In our process, both the film and developer combination worked fairly well, producing
photographs which could be scanned into a computer. These showed the sedimentation of the
solid on a time vs. height scale (Figure 4.7). By scanning these strips of the streaked images,
the image brightness, contrast and relative levels can be manipulated in different software
programs such Photoshop or Imaging to accentuate the differences between the solid and
liquid. Once the sedimentation separation process is visible, the curve can be traced by using
the line and pen tool in Photoshop. From the base of the strip the markings which were made
from the timing LED are visible, lighting up every hour for a few minutes. Each marker can
then be labeled successively as 1 hour, 2 hours, 3 hours, 4 hours, etc. The settling process
70
yielded photographic results which were similar. All photos taken by the streak photography
technique are shown from Figure 4.8 to Figure 4.10.
Streak/strip photography can indeed be used as a tool to photograph moving objects,
integrating the individual slit images into a continuous record that illustrates visual changes
with the passage of time. The process was also quite easy to set up, run, and develop results.
This simple method of photography became very useful in many time and motion studies.
The graphical description about method of streak photography is given in Figure 4.7.
T=1 T=2 T=3 T=4 T=5
T=1
T=2
T=3
T=4
T=5
Figure 4.7 - Formation of streak photography
71
Figure 4.8 - Example of LED light timing module on the base of the image
Figure 4.9 - Two different layers can be seen from the photograph
Figure 4.10 - The curve taken from the Figure 4.9 is shown
When the vessel is full, the time is shown as T=1. Then the settling process in the vessel starts
slowly and the time changes as T=2, T=3, T=4, T=5 and so on. The changes occurring in the
vessel as a result of settling process are recorded constantly by camera. When the settling
process is completed, the camera also finishes to taking photos. After processing whole film
the movement in the vessel can be seen from the processed film (in Figure 4.8, Figure 4.9 and
Figure 4.10). Data from streak photography is compared with the experimental batch
sedimentation results and the accuracy of this process is shown in Chapter 5.
72
4.5 HIGH SPEED VIDEOGRAPHY
The main purpose of using high speed videography is to obtain information, by using a
camera-video recording system and special software, about structures that are very small but
move quickly such as channels, volcanoes and particle flocs, the settling bed behaviour and
how they change during the settling process. Glasrud et al. (1993) used a time lapse videotape
technique to observe the settling behaviour of iron oxide suspensions. They investigated
causes of channelling, cracking and collapse phenomena by using the system for side and top
views.
There are a few special reasons to apply photographic methods on batch sedimentation.
Firstly, the settling bed behaviour can be observed with this system over the total settling
range without disturbing the settling bed. Secondly, the work can concentrate on channels.
Establishment, growth and disappearance of channels have been investigated to understand
the nature of the channels in a settling bed. Information has been collected such as the length
and width of channels. Somasundaran (1981) studied the effect of air bubbles and coarse
particles on settling systems and channelling but did not mention the dimensions of the
channels. Thirdly, volcano formation can be viewed from the surface of the settling front. The
onset of volcanoes has been monitored and a novel collapse phenomenon has been detected.
The relationship between channels and volcanoes has been identified. In addition, the
distribution of volcanoes has been accounted. Lastly, the particle aggregates or particle flocs
in the channels can be observed. Positions of the flocs and floc velocities have been recorded
by using special software. Li and Ganczarczyk (1987) applied a multi-exposure photographic
method for the combined measurement of the settling velocity and the size of activated sludge
flocs. They also affirmed that the smallest flocs measured in their experiments were about 50
microns.
73
The relationship between individual floc-settling velocity and floc size was examined to
determine whether it was linear or a fractional power function. However, measuring floc size
was always a problem for researchers. Hermanowicz and Ganczarczyk (1983) made accurate
determinations of settling velocity by employing a stroboscope to trigger a series of exposures
on a single photograph at pre-set time intervals. They separately measured the particle size by
microscopic observation.
On the other hand, Leenvaar and Rebhun (1983) concentrated their efforts on the
photographic technique and produced negatives with sufficiently sharp floc images for
accurate determination of floc size. However, their technique was developed only for floc size
measurements and was not designed for measuring the velocity of the floc.
4.5.1 Equipment for High Speed Videography
In this part of the photographic study, a high speed video camera (Phantom V4.1M) attached
to a lap top computer was able to record the vision. A high quality lens was used with the
camera which allowed the image to be magnified.
4.5.1.1 The High Speed Camera
The Phantom 4.1M Camera is used for highly specialized applications because it has special
features with an impressive set of capabilities such as a high memory capacity to store the
enormous amount of image information captured from the sensor and an interface system that
allows for data downloads and streaming. The reasons for choosing this camera are that the
Phantom V4.1M (Figure 4.11) offers high light sensitivity (ISO 1200) and high resolution
(512 X 512 pixels). This allowed for recordings to be made with minimum lighting
requirements. The subject’s small size and fine structure required that minimum software
interpolation was carried out, and hence the choice was made to use a Monochrome camera
giving better apparent resolution.
74
Figure 4.11 - Phantom 4.1M Camera
4.5.1.2 Lens
The lens used was a 65mm Macro-Nikkor, a dedicated close-up lens made by Nikon
Instruments. PN11 and PK13 are extension rings used for close-up photography. They are
simply used to extend the lens- to- image- plane distance to increase magnification. In
addition, there was a Nikon C-mount adapter to mount the lens and extension ring
combination to the camera. This added a slight extension to the lens system. Final
magnification was a result of the total lens extension.
4.5.1.3 Lights
Lighting was provided by an Olympus TGHM microscope illuminator. This unit contains an
8U/5A Tungsten lamp that provides constant illumination of approximately 3200K at full
operating voltage. The lamp can be focused to a very small area for very intense illumination.
4.5.1.4 Camera Control Software
It is possible to modify the image captured by using the Phantom 4.1M extensively. For this
purpose the camera control software package allows manipulation of the images.
75
The software is easy to use and contains various menus. The accessibility of these menus
varies based on the operational mode determined by the user. A cine or movie file may
contain hundreds of pictures, while an image file contains a single picture extracted from a
cine file. Cine Files use the file extension .cin for saving movies. Recording (Figure 4.12) can
be in both color and monochrome and in a non-compressed format. Phantom allows saving
selected sequences of any length or the entire movie. Images may also be saved as single
images in other formats for example GIF, LEAD, LEAD-JFIF, LEAD-JTIF, OS/2, PCX,
TGA, or TIF. AVI files use the file extension .avi. They can be viewed with the standard
Windows Media Player.
Figure 4.12 – Setting and recording process with Phantom 4.1
4.5.2 Measurement Process
It is possible to measure different properties by using the Phantom 4.1M camera and software.
The measurement menu can perform measurements on an opened single image file as well as
76
a cine file, while the measure button in the playback window is opened. From this window the
set Unit window offers various options such as distance unit, speed unit, angular speed unit,
angle unit and acceleration unit. When the dialog box appears on the screen (Figure 4.13),
units of measurement for distance, speed, acceleration, angle and angular speed are selected.
Figure 4.13 - Measurements dialog box on the screen
Metric (SI) or US units may be chosen. Scaling option can be selected from the pop-up
selection window, by selecting Measurement>Scaling in the Phantom Camera Control main
screen to set the scale for a single image file. It is also possible to follow the information on
active view shown in Figure 4.14 to view recent information.
The gauge length is described as the dimension of any object or space in the image and used
for calibration. It is recommended to select a larger object in the image for the gauge process.
When the dialog box opens, Gauge Length is entered (Figure 4.15). The gauge length must be
entered in the same units of measurement selected using the Units command.
77
Figure 4.14 - The view cine info on the screen
moving particle distance measurement measured speed of floc

from first point to second point
Figure 4.15 - Distance and speed measurement

78
All these options help the operator to measure physical characteristics of images in any unit
that is required. It makes the process easier, faster and cheaper by using the one system.
4.6 VISUAL OBSERVATIONS
The camera and lights were set up besides the settling column in order to record the images
clearly. The lens used with the video camera allowed the images to be magnified from any
required distance.
The major types of channels or “wormholes” (see Chapter 2) were chosen for recording due to
the size of the channels and because this type of channel lets the flocs move into the channels
without difficulty. After the channels were established, the image was magnified and focused
such that the channels and the flocs in the channel appeared as large and clear as possible on
the monitor. Sharp focus depends on such factors as the distance focused on, lens used and
aperture adopted. Exposure time will affect the amount of image blur due to motion of the
flocs. The very short distance between the camera lens and objects such as flocs results in a
very shallow depth of field. The movements of the flocs in the channels were recorded on
video. Frames were uploaded to a laptop and saved as black and white, high resolution images
and as cine files. When video was being recorded, the control software of the camera allowed
modifications to the picture on the monitor such as brightness, contrast, selection for higher
rates and reduced image size. The same process was repeated for 10 different channels and
flocs. This resulted in a series of images of a floc at different positions along the channels.
Three different photos of each floc in three different positions in the channels are shown in
Figure 4.16. Rotation of the particles shows different shapes of images in successive frames.
For a given combination of lens and extension tube, magnification remains constant for each
combination.
79
a. Position 1
b. Position 2
c. Position 3
Figure 4.16 - Particle 1 in three different positions in the channel
80
Video images were calibrated by photographing a vernier caliper set to just less than the width
of the video frame. Other floc images on the photographs are shown in Appendix D.
In this study, high magnification video-photography was used. The recordings provided
valuable information on behaviour of the settling bed, formation of channelling, movement of
flocs in the channels and floc structures. In order to get accurate information about the
dimensions of channels, it is important that the image transferred to the monitor is as large
and clear as possible. After the image was focused properly, the length and width of channels
were determined by using the gauge length (Figure 4.17) determined for the lens and
extension tube combination that was used to capture the image.
Gauge Length
Figure 4.17 - Calibrating of photographs
Some channels in the photos were wide and very well established, so they were able to
transfer a lot of liquid as well as flocs. Each floc from each image was sized and their
81
velocities determined from the video. To find out each floc velocity, the images of the flocs
were observed in two positions on the video play back.
The distance between these two images and the velocity of the floc is automatically measured
by the control software. Also by observing the position of a particular floc in images 10
frames apart, the distance traveled and the velocity can be found. At the same time the
diameter of the channels was measured at different points. As a result, the physical properties
of flocs and channels were demonstrated at various times. The information is collected from
the photographic work.
82
5 CHAPTER 5
5.1 RESULTS AND DISCUSSION
In this chapter the results from experimental, photographic and mathematical work are
presented and analyzed. The information on channels and the effect of reverse concentration
gradients are discussed. Detailed graphical demonstrations are performed by using various
types of plots. Measurement of the material properties using batch settling and pressure
filtration techniques are outlined through comparisons of experimental results with model
predictions, thereby giving indications of the effects of channeling phenomena.
As mentioned in Chapter 3, the equipment measures concentration at different heights and
times during settling. In this study, initial concentration values were in the range of 250 g/l to
1000 g/l. The height, time and concentration values recorded and the H (height) vs. T (time)
and H vs. C (concentration) graphs are plotted. The activities in the settling bed according to
initial concentration are explained. The places where the reverse concentration and channels
occur are pointed out by graphical values. The effect of solids upflow on descent path at
steady settling condition at different initial concentrations is shown.
From observations of settling beds in Chapter 3 and from considerations of the theories in
Chapter 2, it can be stated that channelling and reverse concentration gradients have
influenced the settling bed. Slurries with higher concentration demonstrate channelling and
reverse concentration in the settling bed to a much larger extent. Also, the initial
concentration has a significant effect on batch settling.
83
At the beginning of the experiments the suspension is placed into the settling column. It is
homogeneous and well dispersed. The particles then start to settle down as individual flocs or
particles, as in this early stage of free settling, they are not hindered by other particles or flocs.
However, as the particles or flocs approach each other, the settling of each particle will be
affected by the surrounding particles, leading to a process of hindered settling. From this point
of time, the settling particles can be seen as a single body undergoing compaction.
5.2 OBSERVATIONS ON SETTLING BED
In free settling, the flocs begin to move individually at quite high speeds and in unsystematic
directions after the induction period. It can be observed that the amount of solid is similar
traveling downwards and upwards. After a while, at the bottom of the container a
consolidated region is formed. This region can be seen with the naked eye. On top of this
level is a region where the flocs became interconnected and fell as one floc right from the start
of the experiment. At the very top region, it was observed that flocs flowed upwards as a
result of the fluid flow from lower regions. When the liquid forces its way upward, it has
preferred paths so that only a part of the bed appears to be supporting the upward flow of
fluid. When the liquid flowing upward breaks the crevices between the larger interconnected
flocs and passes between them, it picks up smaller solid particles in its way. Finally, once the
crevices are formed there is a great deal of bunching and breaking of flocs that act as a firmly
established passageway for the liquid flow. As the concentration increases, the initial frantic
movement slows down and the solid particles come close together. After less than an hour it
can be observed that streams carrying liquid and solid move upward within the bed from the
upper part of the consolidated region.
84
5.3 CHANNELING IN SETTLING BED
Observations show that until particles actually break from the wall and are carried upwards,
the general solid movement in the body of the slurry is downwards and sideways towards the
channels. The bottom of the channels persists and continues to carry solid upwards from all
levels of the sediment tube after the consolidated region has passed.
The characteristic of flocculent suspensions is that the solid particles are not separated
completely. They are in the form of flocs. The coarse particles settle early to form a
consolidated compact bed at the bottom although the upper particles create a force upon the
settled particles at the lower level. This behaviour creates a higher concentration at the bottom
of the settling bed. The liquid at the bottom of the bed is squeezed by the weight of the
compressed bed. Visible movement on the surface is observed within 1 hour of the beginning
of the sedimentation process. After about two hours, channels start to appear on the wall of
the settling bed (Figure 5.1). In order to clarify the process of channelling, a few of the steps
are explained.
Channels
Figure 5.1 - Channels on the wall of the settling tube (current study, run 1)
85
The liquid at bottom level of the suspension flows upward, establishes channels in the upper
level, and causes dilution at the top of the bed. Due to this phenomenon the suspension at the
top settles faster. Furthermore, two other phenomenon could appear while settling proceeds.
Firstly, the liquid and small flocs can rise after water bubbles start to form channels through
the bed. Bubbles rise all the way to the settling interface and liquid can flow through them.
Secondly, the rising water occupies the space the pores left.
These pores enlarge as the liquid flows through and after a while they become established as
channels. The pressure of liquid in the channels tends to keep the channels open. From
photographic observations (see Chapter 4), a range of information is documented. Channels
usually started to form from the bottom of the settling bed and continued all along the bed.
However some channels began to build up from the middle of the settling bed. Also some can
be seen very close to the surface of the bed. Different channels were seen to have different
diameters, and again for individual channels the diameters were not constant and varied with
height and time. It was observed that the channels were not perfectly vertical, and the
presence of some channels moving diagonally was also noted. As the particles blocked the
way, channels could not grow any longer. They would twist and turn in varying directions.
Channels have different average cross section and length. The diameter of the channels was
found to be between 0.3–2 mm and the length of the channels was 2 - 20 cm depending on the
initial concentration value. If the settling bed has a high concentration value the length of the
channel is quite long, up to 20 cm, otherwise the channel is as short as 2 cm. The flows
through large channels have a higher velocity than the flow through narrow channels because
the pressure drop per cm of the bed length must be the same regardless of the channel length
under consideration.
86
The channels are not only blocked simply by the walls closing in to reduce the gap, but also
blocked by larger particles. As soon as a blockage occurs in this way solid passing up the
channel adds to the blockage and fully closes the channel (Figure 5.2 a and b).
a. Channel just starts collapsing by b. Fully closed channel no more

blocking flocs liquid flow
Figure 5.2 - Collapse of a channel in the settling bed (current study, run 4 and run 5)
The first type of channel is called a major channel and the floc clusters can be seen with the
naked eye in these types of channels (Figure 5.3). The bed structure around these channels is
mostly compact. The second type is a minor channel (Figure 5.4) that operates at average
velocity in the cracks and help to deposit the solid effectively. Because of the deposition, the
minor channel size is reduced.
The last type of channel is a simple flow path which is named here as a “local channel”
(Figure 5.5) with no important effect on the solid particles. They lead the flow to the major
channels. They are formed only from the minor channels, after the minor channels are
blocked or clogged by flocs. Under this condition the only possible flow is the slow seepage
of water through the deposit.
87
Major
Channel
Figure 5.3 – The Established major channel on the settling bed wall
(current study, run 6)
Minor
Channel
Figure 5.4 – A minor channel on the settling bed wall (current study, run 7)
88
Major channel
Local channel
Figure 5.5 – The local channel on the settling bed wall (current study, run 8)
Established channels are large and stable. These can be easily seen through the experimental
column. Generally, the orientation of the channels is vertical but sometimes two or more
channels could combine to form one channel. Many branches could form but they are
generally not stable due to insufficient flow.
In this study, channels have been observed in 500 g/l, 750 g/l and 1000 g/l initial
concentrations. Channels formed in low concentration beds are not very stable. However at
high concentrations channels are strong and much more stable. No channels were observed in
250 g/l initial concentration. Though it is hard to determine the numbers of the channels (as
volcanoes on the surface) in the settling bed, images from photographic study indicate that
their number could be approximately 3-4 per cm2.
5.4 VOLCANOES
The volcanoes, as they appear when viewed in transparent settling columns on the top of the
settling bed are shown in Figure 5.6. Besides the apparent channelling that exists in the sludge
sediment, the formation of volcanoes is also observed at the upper surface or
sediment/supernatant interface which is an important characteristic. Enormous amounts of
particles through the surface of volcanoes.

89
a. Volcano on surface o f the bed b. Volcano on the wall
Figure 5.6 – Volcano formations at the surface of the settling bed
(current study, runs 10 & 11)
Particles are deposited on top of the bed (Figure 5.7) which changes the bed concentration.
Volcanoes on the slurry surface are formed as a result of the channeled section having moved
upward through the interface and vanishing afterwards.
Miniature
Volcano
Figure 5.7 – An erupting volcano on the surface of the settling bed

As mentioned in Chapter 2, the volcanoes are interesting formations to investigate. The top
view photographs of the volcano formation (Figure 5.8) were taken after 8 hours of settling.
The photos show that the surface of the settling bed is covered by many holes that seem like
volcanoes. Volcano formation has been observed after 2.5-3 hours indicating that solid was
90
carried from the bottom to the top for more than 3 hours causing reverse concentration
gradients in the settling bed.
Figure 5.8 – Top view of volcano formation on the surface of settling bed
Causes of channelling have been explained previously in Chapter 2. Because of channelling,
the velocity of some flocs is expected to increase in order to fill the void left by liquid that can
flow upward more freely through the channel (with less resistance). However, as the liquid
also drags some solid with it, the average settling process will be slowed down, and as a
result, the average floc velocity will probably decrease.
Observations indicate that reverse concentration gradients occur when the settling process
starts with higher initial concentrations. However, there is a limit to such high initial
concentrations, and if the experiment has too high an initial concentration, that is in excess of
1000 g/l or so (O’Donnell 1972), there is no time between the start of settling and the
formation of reverse concentration gradients.

91
5.5 H-T AND H-C GRAPHS
The data from the batch sedimentation experiments can be plotted as H vs. T (Height against
Time) or H vs. C (Height against Concentration) curves. Due to solids upflow, S shaped
concentration curves show the formation of reverse concentration gradients. Even for low
initial concentration experiments, the concentration loci show a tendency to form S shapes.
The formation of an S shaped curves occurred from the early stages.
All values for four different concentrations (250 g/l, 500 g/l, 750 g/l and 1000 g/l) from
experimental work are given in Appendix A (in the CD-ROM). The values of H versus T, and
H versus C are shown in Appendix B.
By plotting the concentration loci curve, it is possible to show the formation, persistence and
decay of concentration gradients. Auzerais (1989) affirmed that the concentration gradients
form at the bottom of the container and propagate upward to meet the zone of initial
concentration. When the settling solids in lower layers move upwards through channels in the
bed, reverse concentration gradients occur. They are caused by liquid finding shortcuts
through the sediment bed. The liquid carries settled particles upwards to be deposited near the
top of the compaction region. This region is just below the initial concentration. As the solids
settle under these conditions the result is a reverse concentration gradient and the
concentration loci curves present S-shaped profiles. As this period continues, the solid tends
to spread over several vertical layers of particles. After a certain period of time, the
concentration of solid in the upper region of the bed increases while it decreases in the middle
levels.
The main purpose of plotting the concentration profiles (H-C) is to show the distribution of
solids in the sedimentation column at specific settling times.

92
The profile is drawn as a series of curves of height versus slurry concentration with time as
parameter. Reverse concentration gradients usually occur at the beginning of batch settling,
when a high initial concentration is used. This means channel formation is established at a
very early stage in the settling. Figure 5.9 shows the experimental results for an initial
concentration of 250 g/l. Results show that the activities in the settling tube are totally
different from the high initial concentration experiments.
Figure 5.9 - Concentration in settling bed as a function of time for an initial concentration of 250
g/l
During mixing, particles are randomly moving in the slurry and, due to low concentrations,
particles are not touching each other. When particles fall, they encounter slight jostling at the
bottom and can change position and settle in a tightly packed pattern. In the early period of
the settling process at low initial concentrations, the flocs mostly fall under free settling. As
they are not hindered by neighboring flocs, they are likely to attain the maximum possible
velocity which is the terminal velocity. Although the total volume of solids going down
93
should be equal to the volume of liquid going upwards, due to the overall low porosity, the
upward liquid velocity will be low.
Figure 5.9 shows that the different concentrations build the layers at a very early stage of the
process but they are not stable. Some layers start to form from the beginning of the process
but then quickly disappear (for example, the 500 and 750 g/l curves merge with the 1000 g/l
curve). This type of instability could be observed for about 50 min, and thereafter stable
formations of the concentration layers were observed.
Finally, near the end of the experiment, few concentration layers could be seen but no
channels or reverse concentration gradients were seen to occur, and if the initial concentration
was less than 200-250 g/l, no channels occurred. At the end of the settling, the bed reached a
stable state with a more or less constant concentration distribution, as can be seen from Figure
5.9 where the constant-concentration lines became almost horizontal.
Figure 5.10 shows the behaviour of a 500 g/l initial concentration slurry. Unlike the previous
case (Figure 5.9) with an initial concentration of 250 g/l where no reverse concentration was
observed, the H-T graph of 500 g/l initial concentration clearly indicates the occurrence and
position of reverse concentration gradients in the 1200 and 1300 g/l concentration curves. The
shape of the constant-concentration lines of 1200 g/l and 1300 g/l show a particular S-shape
behaviour indicating the same values of concentrations at different heights at the same time.
The positions of the reverse concentration region, as can be seen from Figure 5.10, was
between 30-50 cm from the bottom of the bed, and this phenomenon occurred between 100
and 200 minutes after the experiment started. As mentioned before, two types of settling
occur during batch sedimentation with respect to the initial concentration. In this experiment
the initial concentration is higher than 300-350 g/l so the effective settling type is different
94
from the previous experiment.
500 g/l
At the very beginning of the experiment, up to around 60 minutes after the start, hindered
settling behaviour can be clearly seen at the top part of the settling column, down to a height
of about 50 cm. However, below that height, no settling was observed.
It suggests that as the slurry is allowed to settle, the concentration decreases rapidly at
relatively higher heights and hence the solid particles have more and more space to move
around and they can settle quicker. On the other hand, the lower region of the settling bed
becomes more concentrated and the individual particles start to come in contact with each
other, forming a consolidated or network structure. The particles can no longer move
individually but collectively they act as a compacting bed.
95
In the next stage of the settling process, some form of dispersion or classification is seen to
occur near the surface of the formed bed. After approximately 100 minutes (Figure 5.10) a
large amount of solid starts to move from inside the settling bed to the top. The distribution of
concentrations displayed an irregular behaviour between 100 minutes and 420 minutes. At the
end of the experiment, the entire bed is seen to attain a constant concentration distribution
with only a small gradient, and from Figure 5.10 an average value of 1500 g/l can be
estimated.
Figure 5.11 shows the behaviour for an initial concentration value of 750 g/l. The settling
process takes longer than the 500 g/l initial concentration experiment (approximately 700
minutes). The first changes in the settling process can be seen after 90 minutes. The 900g/l,
1000 g/l and 1100 g/l lines show similar behaviour.
750 g/l
96
The differences start to appear in the 1200 g/l concentration line at the very bottom of the bed
in the first 30 minutes. The changes in concentration are clearly visible within 120 minutes. In
100 minutes the S-shaped curves start to appear showing reverse concentration gradients.
Such reverse concentration gradients are found to take place at the middle of the bed where
concentrations ranged from 1300 g/l to 1500 g/l. There is no such behaviour near the top of
the bed where the concentration is less than 1200 g/l. The concentration layers start to change
at 1300 g/l concentration. The H-T graph shows S-type curves in the 1400 g/l and 1500 g/l
concentrations where the reverse concentration gradients are present. At the highest
concentration (1600 g/l), the bottom of the settling bed reaches the steady state condition and
the concentration is fixed.
Figure 5.12 shows the results when a high initial concentration (1000 g/l) is used. Hindered
settling can clearly be seen at the beginning of the settling process.
1000 g/l
97
Initially, before the concentration reaches 850g/l, solid particles move randomly in the settling
tube, such that there are some dispersion effects. For the first 80 minutes, they demonstrate
hindered settling in the suspension at the top of the bed. Channels form and behave like tubes
in the bed and carry a large amount of solids upward. After 1400 g/l, the concentration layers
in the settling bed start to change. Reverse concentrations appear in the bed. For these
experiments, the settling bed becomes compact around 1450g/l. Therefore, the reverse
concentration gradients and channels are seen at 1400 g/l, 1500 g/l and 1600 g/l. For
concentrations higher than 1600g/l, the plot would be a straight line.
In Figure 5.13, the H-C graph has been plotted at a low initial concentration of 250 g/l.
According to this graph, settling occupies some time in the settling tube prior to 42 minutes.
The initial concentration value line lies at the left of the plot. The time distributions show
stable behaviour until 300 g/l.
Figure 5.13 - Height vs. concentration plot showing reverse concentration gradients for an initial
concentration of 250 g/l
98
Most of the lines show a movement between 800 g/l and 1300 g/l. There is an unusual
behaviour in the case of the 600 g/l and 1400 g/l concentrations. It is thought that this
abnormal behaviour can be explained by bubbling in the settling bed.
The H-C graph in Figure 5.14 demonstrates almost constant behaviour at times t ≥ 125
minutes, at the height of 60 cm, until the concentration is approximately 1300 g/l. It can be
seen clearly that the concentration changes start at the early stage of experiment (seen in the t
= 83 min curve). At later times, the S-shaped curve can be seen between 1300 g/l and 1400 g/l
concentrations indicating channelling behaviour.
At 167 minutes, the plot shows significant changes in concentration indicating a considerable
movement of the flocs, which suggests that the solid bed is unstable. In addition, the H-C
graph indicates reverse concentration gradients between 1200 g/l and 1500 g/l concentration
values. When the solid bed reaches 375 minutes, all material is in the bed, but a reverse
99
concentration gradient is still observed
The heights versus concentration results for 750 g/l are shown in Figure 5.15. The top of the
bed is clearly seen by the discontinuities at about 1400 g/l. Small reverse concentration
gradients arise at this concentration for times between 292 and 417 minutes, indicating that
this phenomenon occurs at the top of the bed.
concentration of 750g/l
Figure 5.16 shows the 1000 g/l results. Due to high concentration, there is limited space for
settling and greater interaction between the particles, which results in the slurry behaving as
one whole floc indicated by the similarity of the plots. At 42 minutes, the settling material is
at about 925 g/l; at 83 and 125 minutes, there is a sharp decline at about 850 g/l – these results
are indicative that the settling material is also diluting. At 208 minutes and greater, Figure
100
5.16 shows S-shaped curves for concentrations greater than 1400 g/l (the start of the
compression zone), indicating reverse concentration gradients in the bed.
concentration of 1000 g/l.
5.6 THE DESCENT PATHS
The descent path or filtration rate curve is shown on a plot of height against time. The purpose
of creating a descent path curve is to demonstrate the movement of particles (Figure 5.17) (a)
and (b)) within the settling bed. This method was introduced by Gaudin and Fuerstenau
(1962) and is a convenient method to describe and evaluate other published results.
Anomalies on the graphs were prepared by multiplying the time values at any point with the
percentage value of the solids. On the filtration rate curve, at any point in time, the mass
above the curve remains in a constant ratio to the total mass of slurry such that, for any profile
of height against concentration, the area of graph above the curve is constant.
101
Figure 5.17 - Calculation of pulp density solids by using descent paths
(modified from O’Donnell 1972)
The area criterion was described more fully by Gaudin and Fuerstenau (1962). The descent
path curves are shown in Figure 5.18 through Figure 5.21. The values appearing on the graphs
are calculated from the tables in Appendix C. It is possible to calculate the settling flux from
the slopes of descent paths after multiplying them by the respective concentrations existing at
that point.
For the plots of the descent paths, a similar procedure as used for the plotting of concentration
loci was used, namely, joining the calculated points by smooth lines. It is evident that the
locus of particles, as defined by the area criterion, is not always downwards with respect to
time. The plots show that at some time, the solid is flowing upwards within the bed. It is
interesting that the shape of the surface contour on these descent path plots and that of the
concentration loci reflects more of the behaviour beneath the solid rather than the surface
concentrations (O’Donnell 1972). The values of descent path graphs for four different
concentrations are shown in Appendix C.
102
Figure 5.18 depicts the descent path when the initial concentration is 250 g/l. The presence of
anomalies can be seen in the first 15 minutes and continues until 50 minutes on the graph.
Because of some turbulence in the flow, the flocs at this time travel upwards and downwards.
Also, the contours show a dip, up to time of about 30 minutes. During experimentation, the
free settling process was observed in this early period. After free settling is completed, there is
no further movement in the settling bed as indicated by the fact that the curves become
straight lines at later times. Similar behaviour has also been shown in Figure 5.19.
concentration 250 g/l
Comparing Figure 5.18 and Figure 5.19, it can be seen that the plots show different outcomes.
There are relatively wider spaces between the contours because of the higher initial
concentration. The movements in the settling bed start early. The turbulence of flow in free
settling zone might be the reason for this.
103
The unusual activity starts at 40 minutes and persists until 125 minutes after which time there
is no further anomaly observed in the settling bed. When the initial concentration increases,
the influence of the high concentration (initial concentration 750 g/l) on the descent paths can
be seen in Figure 5.20. The plots started at around 80 minutes and continued until 160
minutes. The plots that are above 30% of total solid show early onset of non-linearity which
increases as the concentration increases. This indicates that channelling occurs between 80
minutes and 150 minutes. Comparing Figure 5.20 with Figure 5.19, when the initial
concentration is 500g/l the non-linear behaviour becomes apparent.
Figure 5.21 shows descent paths for the highest initial concentrations used. The distribution of
the plots shows the smoothest behaviours of all initial concentrations. The plots display slow
changes around 80 minutes continuing until 100 minutes. This is where channelling take
place. This movement begins around 50% then goes on up to 100%. After 100 minutes the
plots show slow horizontal movement due to the effect of high concentration.
104
105
5.7 EVALUATION OF PHOTOGRAPHIC WORK
The batch sedimentation process was observed using streak photography without disturbing
the settling bed. Information about the methods used for streak photography was given in
Chapter 4. The aim here was to collect online data by applying the photographic method and
then to compare the data with the plotted experimental results. Figure 5.22 shows the
measurements from streak photography superimposed on the experimental result graph (for an
initial concentration of 500 g/l) indicating a close match between the visually observed data
and the experimentally measured and calculated data. Readings of time versus height values
from streak photographic observation can be seen in Appendix D.
Photographic result
120 0 g/l 700 g/l

1000 g/l 1200 g/l
100 1400 g/l 1500 g/l
1300 g/l Photo results
80
Height (cm)
60
40
20
0
0 50 100 150 200 250 300 350 400 450 500 550 600
Time (min)
Figure 5.22 - Comparison of streak photographic measurements with the experimental results
(Initial concentration 500 g/l)
106
5.8 MATERIAL CHARACTERISATION RESULTS
The compressive yield stress, Py(φ), and hindered settling function, R(φ), data were
determined as a function of the solids volume fraction, φ, for the flocculated calcium
carbonate suspension. The compressibility and permeability measurements were used to
understand the material behaviour for a flocculated suspension.
In this study, the sample of calcium carbonate flocculated with Teepol was characterised in
terms of its compressibility and permeability using a combination of stepped-pressure
filtration and batch settling testing. The application and detailed information on stepped
filtration method was given in Section 3.4 in Chapter 3.
5.8.1 Stepped-pressure filtration results
The solids density was calculated gravimetrically using a density bottle to be 2.681 g/cm3.
Two tests were performed using the filtration apparatus. The compressibility test results are
presented in Figure 5.23, which shows that, at each pressure, the test proceeds until
equilibrium is reached.
The final concentration at the last pressure was determined by oven drying. The volume
fraction at each pressure is calculated from the volume of filtrate exuded at equilibrium,
giving the Py(φ) curve for high φ. The values are given in Table 5-1, which shows that the
material is considerably incompressible, given that the volume fraction only changes from
0.610 to 0.637 between 5 and 200 kPa.
107
600000
200 kPa
500000
100 kPa
400000
Time, t (s)
50 kPa
300000
200000
20 kPa
100000
10 kPa
5 kPa
0
0.0E+00 5.0E-06 1.0E-05 1.5E-05 2.0E-05
(Specific Volume)2, V 2 (m2)
Figure 5.23- Time versus (Specific Volume)2 results from the compressibility
stepped-pressure test
Table 5-1- Compressibility results from filtration testing
∆P φ∞
(kPa) (v/v)
5.28 0.6103
10.7 0.6143
20.5 0.6194
51.0 0.6242
100.4 0.6288
200.6 0.6367
108
400
350
300
20 kPa
250
Time,t (s)
200
10 kPa
150
100
50 5 kPa
0
0 1E-05 2E-05 3E-05 4E-05 5E-05 6E-05
(Specific Volume)2, V 2 (m2)
Figure 5.24- Time versus (Specific Volume)2 results from the permeability
stepped-pressure test
The permeability test results are presented in Figure 5.24. The test proceeds at each pressure
until linearity in t versus V2 is achieved, and then steps to the next pressure. The results show
that the material went into cake compression (non-linearity in t vs V2) at 20 kPa, and therefore
the results are only valid up to this pressure. The slope of the plot (β2) at each pressure (∆P)
was used to calculate R(φ) for high φ using the method of Landman et al. (1999), given by
Equation 5-1. The values are given in Table Table 5-2.
2  1 1
R(φ ) =  − (1 − φ )2
2 
5-1
dβ  φ 0 φ 
d∆ P
109
Table 5-2- Hindered settling function results from pressure filtration testing
φ R(φ)
(v/v) (Pas/m2)
0.6042 5.21x109
0.6102 5.49x109
0.6160 5.74x109
5.8.2 Batch Settling Tests Results
Batch settling testing was used to give a volume fraction permeability and compressibility
data. Four batch settling tests (Appendix G) were performed in 500 ml measuring cylinders,
with initial concentrations of 229, 421, 586 and 728 g/l (0.0853, 0.1572, 0.2186 and 0.2717
v/v) respectively and then the suspension allowed to settle. The interfacial height was
measured as a function of time until the interface stopped falling and the results are presented
in Figure 5.25. The height versus time results were analyzed in conjunction with the pressure
filtration results using software developed at the University of Melbourne based on the work
of Lester et al. (2005). The inputs to the software were the initial solids concentration, the
initial height, the transient height, Py(φ) and R(φ) at high φ. The software converts the batch
settling results to give Py(φ) and R(φ) at low volume fractions.
No analysis was possible for the 750 g/L results since the final average volume fraction was
greater than the compressive yield stress and is likely to be an anomaly. Also presented in
Figure 5.25 are the curve fits from the analysis program, as predicted by the material property
outputs. They show very good agreement with the raw experimental data.
110
0.35
229 g/L
0.3 229 g/L fit
421 g/L
0.25 421 g/L fit
Height, h (t ) (m)
586g/L
0.2
728 g/L
728 g/L fit
0.15
0.1
0.05
0
0 1000 2000 3000 4000
Time, t (s)
Figure 5.25- Height versus time results for batch settling tests
The compressive yield stress results are shown in Figure 5.26. The average gel point is
calculated as 0.505 ± 0.067 v/v (1354 ± 180 g/L).
1000
Compressive Yield Stress, Py (f ) (kPa)
Filtration
229 g/L
421 g/L
100
728 g/L
10
0.1
0.5 0.55 0.6 0.65
Solids Volume Fraction, f (v/v)
Figure 5.26- Compressive yield stress results
111
The results in Figure 5.27 show that the permeability changes by up to two orders of
magnitude over the volume fraction range in question. The combination of Py(φ) and R(φ)
results allows the predictive modelling of the behaviour of the material under any saturated
dewatering process, such as filtration, centrifugation and thickening. Model predictions for
batch sedimentation are presented next.

-
1.E+10
Hindered Settling Function, R (f ) (Pasm
Filtration
229 g/L
421 g/L
728 g/L
1.E+09
)
2
1.E+08
1.E+07
0 0.2 0.4 0.6 0.8
Solids Volume Fraction, f (v/v)
Figure 5.27- Hindered settling function results
5.9 MATHEMATICAL MODELLING RESULTS
The material characterisation results were used in mathematical modelling adopted from
Usher (2002) to predict the suspension behaviour in sedimentation. The theory is developed
in Chapter 2. The inputs to the model were Py(φ), R(φ ) , h0 and φ 0 , and the results are
presented as Height (cm) versus Time (min) and Height (cm) versus Concentration (g/L)
plots.
112
The Height versus Time graphs show three main solutions to the governing equation, such as
constant concentration, fan and compression in the settling bed. The plots correspond to
sedimentation mode 3a mentioned in Lester et al. paper (2005).
The reverse concentration gradients and channelling behaviour build up in the fan region.
However, the mathematical model corresponds only to a one-dimensional study and so it is
not possible to observe either the reverse concentration gradients or channelling on Height
versus Time graphs in Figure 5.29, Figure 5.30 and Figure 5.31.
Figure 5.28 depicts the profile of mathematical modelling in the settling bed as a function of
time for an initial concentration of 250 g/l.
Constant concentration
Fan Compression
g/l
113
The first part of the plot shows the constant concentration where φ = φ 0 . The settling plot
demonstrates changes in h ' (t ) from linear to non-linear behaviour when h(t) is 38 cm and τ 1
is 45 minutes. After this point, a fan arises from the origin which covers the transition from
φ 0 to φ max . The second fan, where the concentration is 1250 g/l, occurs tangentially from the
consolidating bed. Finally the settling bed reaches the compression point where φ ≥ φ g
Comparison of Figure 5.9 and Figure 5.28 shows similarity in terms of maintaining different
regions in the settling bed. In Figure 5.9 the plots reflect a constant concentration region at the
beginning, a continuous fan region and lastly a compression region. The fan region
establishes at a height of 38 cm and around 48 minutes after the experiment starts in Figure
5.9. When the time reaches 160 minutes, the fan region disappears and the bed goes to
compression which can be seen from both the graphs in Figure 5.9 and Figure 5.28. The final
measured height was 40 cm, whereas the final predicted height was 22.5 cm, suggesting that
either the characteristics for the material are slightly too compressible, or that the experiment
has some random error in the measurement. Overall, the observations indicate the similarity in
the settling bed between two graphs.
Figure 5.29 shows the concentration profile in the settling bed as a function of time for an
initial concentration of 500 g/l. The first part of the plot depicts the constant concentration
where φ = φ 0 . h ' (t ) changes from linear to non-linear behaviour when h(t) is 56 cm and τ 1 is
80 minutes. After this point the fan where φ 0 < φ < φ max arises from the origin which covers
the transition from φ 0 to φ max . The second fan occurs where concentration is 1500 g/l
tangentially from the consolidating bed. Finally the settling bed starts compressing around
1600 g/l where φ ≥ φ g .
114
Constant concentration
Fan
Compression
g/l
When Figure 5.10 and Figure 5.29 are compared, the constant concentration region, the fan
region and the compression region are shown in the settling bed. The fan region begins at a
height of 57 cm and around 60 minutes (slightly quicker than the predictions) and it
disappears in 400 minutes. The settling bed goes into compression at the latest stage of
settling – the final measured height was 55 cm, and the final predicted height was 41.5 cm.
The experimental data and mathematical predictions show similar behaviour in the settling
bed between two graphs.
Figure 5.30 indicates the profile of mathematical modelling in settling bed as a function of
time for an initial concentration of 750 g/l. The material characteristics used for this were the
421 g/l results. The first part of the plot depicts the constant concentration where φ = φ 0 . The
115
settling plot demonstrates changes in h ' (t ) from linear to non-linear behaviour at 75 cm and
148 minutes, where the first fan ( φ 0 < φ < φ max ) arises from the origin. A second fan exists
between 1400 and 1500 g/l, originating tangentially from the compressing bed.
Constant
concentration
Fan
Compression
g/l
The fans are very close to each other because of the high initial concentration. Finally, the bed
( φ ≥ φ g ) continues compressing after all material has settled.
Figure 5.11 and Figure 5.30 are compared and the three regions are identified in both graphs.
The fan region establishes at 77 cm height and around 120 minutes (again, slightly quicker
than the predicted results) and it finishes at 550 minutes. The settling bed goes into
compression at the latest stage of settling – the final measured height was 66 cm, and the final
predicted height was 61 cm. The experimental data and mathematical prediction of the
116
material characterisation results point out the similarity in the settling bed between two
graphs.
Figure 5.31 indicates the profile of mathematical modelling in settling bed as a function of
time for an initial concentration of 1000 g/l. The first part of the plot depicts the constant
concentration where φ = φ 0 . h ' (t ) changes from linear to non-linear behaviour at around 103
cm and 198 minutes, which is the starting point of fan behaviour where φ 0 < φ < φ max . The
compressing bed occurs after 1350 g/l where φ ≥ φ g .
Constant
concentration
Fan
Compression
g/l
Figure 5.12 and Figure 5.31 maintain the same regions in the settling bed: constant
concentration, fan and compression. In both figures, the fan region establishes at a height of
117
105 cm at around 120 minutes and slowly disappears by the end of the settling. The last stage
of settling bed is compression: the experimental results show a final height of around 68 cm,
while the predicted results show a height of 85 cm. This suggests that either the Py(φ) is too
low, the experimental results have been disturbed (for instance, because of vibrations in the
laboratory) or that channelling causes the bed to compact more than predicted. In general, the
mathematical predictions using the material characterisation results and the experimental data
show similar behaviour. The Height versus Concentration predictions demonstrate the
concentration changes with time in the settling bed. None of the predictions are expected to
show the S shape curves that indicate reverse concentration gradients since it is a one-
dimensional model.
Figure 5.32 indicates the low concentration profile of mathematical modelling in settling bed
for an initial concentration of 250 g/l. Beyond 42 minutes, settling has mostly finished. The
concentration in the bed gradually increases.
When Figure 5.13 and Figure 5.32 are compared, both figures indicate changes between 400
g/l and 1300 g/l concentration ranges and between 0-25 cm heights. The gel point is
determined around 1280 g/l in both graphs. Although there are noticeable changes in
concentration profile as time goes on, there is no reverse concentration or channelling
behaviour taking place in the settling bed due to the low initial concentration. Both graphs
show similar behaviour during the settling process.
Figure 5.33 indicates the concentration profile for an initial concentration of 500 g/l. The
forms of the flux curves show semi-linear behaviour until 500 g/L, representing settling. The
concentration changes between 500 g/l and 1500 g/l represent the fan behaviour. The gel
point appears at the latest stage (around 1500 g/l) of settling.
118
Initial concentration
Gel
point
concentration of 250g/l
Comparison of Figure 5.14 and Figure 5.33 indicate the noticeable changes in the fan area,
especially between 1200 g/l and 1400 g/l concentration and between 0-50 cm heights. This
area corresponds to the region of reverse concentration gradients due to channelling in Figure
5.14. The gel point is determined around 1280 g/l in both graphs. Again, the two graphs show
similar behaviour during the settling process.
Figure 5.34 shows the concentration profile predictions for an initial concentration of 750 g/l.
As a result of the high initial concentration, the settling takes longer than the previous
example. As for the previous results, the distribution shows settling, fan and bed regions. The
fan region starts after 750 g/l and exists until and the gel point at 1500 g/l.
119
Initial
concentration Gel
point
The variety of changes on the time plots corresponding to reverse concentration gradients can
be seen between 1100 g/l and 1500 g/l in Figure 5.15 and Figure 5.34, again in the fan region.
The gel point is verified around 1500 g/l in both graphs. The settling process shows similar
behaviour in both graphs.
Figure 5.35 shows the predictions for the concentration profile for an initial concentration of
1000 g/l. The concentration profiles demonstrate linear behaviour up until 1000 g/l, indicating
the settling bed. Between 1000g/l and 1300 g/l, noticeable changes can be seen as the fan
region develops. The gel point appears at 1300 g/l, beyond which the compression grows. The
bed reaches a constant distribution by 800 minutes.
120
Initial
concentration
Gel
Point
Figure 5.35 shows interesting behaviour at low concentrations. The slight drop in height at
low concentrations in this figure is due to the numerical solution of the governing equation.
The same formation is seen at 850 g/l in Figure 5.16, but is an experimental result. There are
noticeable concentration changes between 1000 g/l and 1300 g/l which correspond to reverse
concentration gradients in Figure 5.16, indicating that this phenomenon occurs in the fan
region. The settling demonstrates a high height value due to the very high initial concentration
than the other experiments. This can be seen in both graphs. The gel point is determined to be
around 1300 g/l in both graphs. Overall, the two graphs show similar behaviour during the
settling process.
121
Initial
concentration
Gel
Point
122
6 CHAPTER 6
6.1 CONCLUSIONS AND FURTHER WORK
This work investigated the phenomena of reverse concentration gradients, channelling and
volcano formation in batch sedimentation. These effects can have large impacts on settling
behaviour, and are not accounted for in theories of sedimentation. Experiments were
performed using a laboratory-scale sedimentation unit in which the local concentration was
measured as a function of height and time using a gamma-ray source and a detector. A
photographic study was performed in conjunction with the experiments. The dewatering
characterisation of the material was performed using stepped-pressure filtration and batch
settling experiments, and the results used to predict batch sedimentation which was then
compared with the experimental results.
The scope of this project was to improve knowledge of thickening by understanding
behaviour such as channelling in the batch sedimentation process and highlighting the
importance of material characterisation testing. A greater understanding of the principles
involved will impact on better thickener designs and therefore reduce the costs of recovery
processes.
6.2 CONCLUSIONS
In this study, theoretical and experimental work on channelling behaviour in batch
sedimentation has been presented. Graphical interpretations show the changes in the settling
bed at different concentrations. Numerical computations allow description of the relationship
between the experimental study and the mathematical model predictions. General conclusions
123
are given here from the basis of observations of the settling bed and from considerations of
the theories and results of the previous chapters.
6.2.1 Channelling Behaviour
The role of channels during batch settling is examined in order to understand their importance
in batch sedimentation (see Chapter 2 and Chapter 5 for further details). Channelling in the
settling bed can affect the performance of thickeners but is not accounted for in current
theories. Inspections showed that channels need an appreciable time to form and usually
vanish towards the end of the batch test. Channels formed that were from 2 to 20 cm in length
and 0.5 to 3 mm wide. Channelling occurred in the fan region and at the top of the
compressing bed. Three types of channels (major, minor and local) were observed by the
photographic study.
Part of this study was to demonstrate that channels of liquid carry solid particles to the upper
reaches of the sedimenting bed. The major cause of channelling is the high-pressure gradient
at the bottom of the bed and water bubbles in the settling bed. The fluid rising from the
compression zone causes cracks through the bed. The most important factors that cause
channelling are high initial concentration (up to 1000 g/l) and settling time. No channelling
was observed in the 250 g/l experiment, confirming that there is a lower limit in the initial
concentration for CaCO3 suspensions below which no channels are formed. It has been seen
in this study that channelling can occur only when the initial concentration is more than or
equal to 500 g/l. This can be important information for design and operation of batch settlers.
The second important factor was found to be time. Even when the initial concentration was
high, channelling occurred after 2 hours.
This critical time, however, depended on the initial concentration. Higher initial concentration
caused channelling to start earlier, bringing the critical time to 1.5 hours.
124
6.2.2 Reverse Concentration Gradients
Reverse concentration gradients are a result of channelling in batch sedimentation. Once
channelling starts, the liquid flowing upward in the channels carries the solids with it, and
there is a net upward transportation of solid in the channelling region of the bed.
As with the channelling, no reverse concentration gradients were observed for the 250g/l
experiment. At very high initial concentration, the reverse concentration gradients started
almost from the very beginning of settling. The significance of reverse concentration
gradients is discussed in Chapter 2 and graphs reflecting this phenomenon are shown in
Chapter 5. The literature review in this research showed that the reverse concentration
gradients have been usually ignored and not counted for the mathematical modelling by the
researchers.
6.2.3 Material Characteristics
Stepped-pressure filtration and batch settling tests were used to examine the material
characteristic parameters such as Py( φ ) and R( φ ). The filtration tests showed that the
suspension used for the experimental work was not very compressible even under very high
pressure (200 kPa) applied. The average gel point was calculated as 0.505 ± 0.067 v/v (1354 ±
180 g/l). The permeability changes by two orders of magnitude over the volume fraction
range in question. The combination of Py(φ) and R(φ) results were used for the predictive
modelling of the behaviour of the material.
6.2.4 Photographic Technique
Examining the flocculated suspensions without disrupting the process required a complex
technique. Applying photography in scientific research on batch sedimentation has a very
short history in the literature. The photographic techniques used in this study successfully
demonstrated different perspectives of collecting data, and constituted a significant
125
achievement. According to literature, various photographic techniques have been used for
different materials such as Kaolin, Calcium Oxide and Ferric Oxide for different purposes,
such as measuring settling velocity of particles and floc size. However, this study was unique
in using a high-speed camera for tracking flocs and observing the channels, volcanoes and the
flocs in the channels. In addition, the movements of the settling bed from the beginning of the
settling process, and establishment and eruption stages of volcanoes were recorded as direct
evidence to support this study. The settling flux curve was created from the streak
photography and compared with real experimental data. The excellent similarity of two
graphs was shown in Chapter 5.
6.2.5 Mathematical Modelling Predictions
The agreement between mathematical predictions of suspension behaviour and experimental
data (see Section 5.9) showed close similarity. The flux curves prepared with mathematical
predictions were compared with sedimentation modes introduced by Lester et al. (2005). The
results plotted from mathematical modelling predictions indicate the regions of constant
concentration, fan and compression in four different initial concentrations. Comparing with
experimental results showed that channelling usually occurred in the fan region. The
differences between predictions and observations were due to the combination of
experimental variation, model deficiencies and the effects of channels.
Overall, the variety of methods used in this study to present information on batch
sedimentation outlines valuable information in terms of understanding flocculated suspension
settling and the reasons for the establishment of channels and reverse concentration gradients.
The understanding of such variables has crucial importance to increase the productivity of
thickening process.
126
6.3 RECOMMENDATIONS FOR FURTHER WORK
The following are recommended as subjects for future studies regarding batch sedimentation
phenomena and channelling behaviour.
6.3.1 Advanced High Speed Videography
As an extension to the current study, further photographic work can be done by using a high
speed camera. A problem in this work was changing the position of the camera and lights to
follow the formation of channels, such that the process took a long time to capture useful
photos. A circular rig could be set up around the experiment with the camera and lights set up
on the rig to give more flexibility to the operator and allow more images to be taken at any
position and any time.
6.3.2 Flocculant and Material Effects
Different types and doses of flocculants such as polymer types (Zetag, Magnafloc, Superfloc),
detergents (Teepol) and electrostatic flocculating agents can be used with various materials,
for example Kaolin, Calcium Oxide, Ferric Oxide, Limestone and Sandstone in batch
sedimentation. Flocculation affects the settling time in batch sedimentation tests. Factors
which can be used as a basis to identify the best flocculant include; settling time, chemical
behaviour, compatibility and recyclability. By carrying out a study of various flocculants and
their effects, it can be shown which material, type and amount of flocculant is the most
suitable.
6.3.3 Using Mixing Tank
A useful and practical set up would be to add a tank and mixer unit next to the settling
column. The idea here would be to mix the material and liquid in the tank during flocculation
and then transfer to the settling column using a low-shear pump. Thus efficient and uniform
mixing would be ensured.
127
REFERENCES
1. Akers, R. J. (1976) "Flocculation, Industrial Research Fellowship Report" Powder

Technology 15 (2): pp.294.
2. Allen T. and T.H.P. Uhlherr (1989) Nonhomogeneous sedimentation in viscoelastic

fluids. Journal of Rheology 33 (4): pp.627.
3. Auzerais, F.M., R. Jackson and W.B. Russel (1988) "The resolution of shocks and the
effects of compressible sediments in transient settling" Journal of Fluid Mechanics 195:
pp.437.
4. Auzerais, F.M. (1989) The transient settling of stable and flocculated dispersions, PhD
Thesis, Department of Chemical Engineering, Princeton University: pp.193.
5. Ballou, D.P. (1970) “Solutions to non-linear hyperbolic Cauchy problems without

convexity conditions”, Transactions of the American Mathematical Society, 152:pp.441.
6. Bascur, O. (1989) A unified solid-liquid separation framework. First Regional

Meeting of the American Filtration Society. Houston, Texas: pp.11.
7. Bürger, R. and F. Concha (1998) "Mathematical model and numerical simulation of

the settling of flocculated suspensions" International Journal of Multiphase Flow 24: pp.1005.
8. Bürger, R., M.C. Bustos and F. Concha (1999) "Settling velocities of particulate
systems : Phenomenological theory of sedimentation process: numerical simulation of the
transient behavior of flocculated suspensions in an ideal batch or continuous thickener"
International Journal of Mineral Process 55: pp.267.
9. Bürger, R., F. Concha and F.M. Tiller (2000) "Applications of the phenomenological
theory of several published experimental cases of sedimentation processes" Chemical
Engineering Journal 80: pp.105.
10. Bürger, R., F. Concha, K.K. Fjelde and K.H Karlsen (2000a) “Numerical simulation
of the settling of polydisperse suspensions of spheres”, Powder Technology, 133:pp.30.
128
11. Bürger, R., F. Concha and K.H. Karlsen (2001) "Phenomenological model of filtration
processes: 1. Cake formation and expression." Chemical Engineering Science 56 (15):
pp.4537.
12. Bürger, R., S. Evje, and K.H. Karlsen (2000c) "On strongly degenerate convection-
diffusion problems modeling sedimentation-consolidation processes" Journal of Mathematical
Analysis and Applications 247: pp.517.
13. Buscall, R. and L.R. White (1987) "The consolidation of concentrated suspensions.
Part 1.The theory of sedimentation" Journal of the Chemical Society, Faraday Transactions.
1(83): pp.873.
14. Bustos, M.C. and F. Concha (1988) "Simulation of Batch Sedimentation with
Compression." AIChE Journal 34(5): pp.859.
15. Bustos, M.C., F. Concha, R. Burger and E.M. Tory (1999) Sedimentation and
Thickening: Phenomenological Foundation and Mathematical Theory, Kluwer Academic
Publishers, Dordrecht, Netherlands .
16. Chakravarti, A. and C. C. Dell (1970) "The behavior of highly flocculated suspensions
in batch test" Powder Technology 3: pp.287.
17. Chandler, J. L. (1983) "Dewatering by deep thickeners without rakes" Filtration and
Separation 20: pp.104.
18. Chen, G. W., L. I. Chang, et al. (1996) "Regimes for zone settling of waste activated
sludges" Water Research 30(8): pp.1844.
19. Cheremisinoff, N. P. (1995) “Gravity separation” Solids/Liquids Separation,

Technomic Publishing Company: pp.58-82.
20. Coad, M. A. and K. J. Ives (1981) Deposition and flow patterns in the pose of deep
bed filters. World Congress of Chemical Engineering, Montreal.
129
21. Coe, H. S. and G. H. Clevenger (1916) "Methods for determining the capacities of
slime-settling tanks" AIME Transactions 55: pp.356.
22. Comings, E. W. (1940) "Thickening Calcium Carbonate Slurries." Industrial &

Engineering Chemistry 32 (5): pp.667.
23. Comings, E. W., C. E. Pruiss and C. DeBord (1954) "Continuous Settling and
Thickening" Industrial & Engineering Chemistry 46(6): pp.1164.
24. Concha, F. and M.C. Bustos (1985) Theory of sedimentation of flocculated fine
particles. Proceedings of the Engineering Foundation conference on Flocculation,
Sedimentation and Consolidation. B. M. Moudgil and P. Somasundaran. New York, AIChE
Journal: pp.275.
25. Concha, F. and M.C. Bustos (1991) "Settling velocities of particulate systems, 6.
Kynch sedimentation process: batch settling." International Journal of Mineral Processing 32:
pp.193.
26. Concha, F., M. C. Bustos and A. Barrientos (1996) Phenological theory of

sedimentation. sedimentation. Sedimentation of Small Particles in a Viscous Fluid. Ed.E. M.
Tory, Computational Mechanics Publications, Southampton, UK: pp: 51-96.
27. Concha, F. and R. Bürger (2002) "A century of research in sedimentation and
thickening" Chemical Engineering Science 20: pp.38.
28. Curran, J.K, G.T. Hill and G.T. Milligan (2003) “Time variation of floc properties in a
settling column” Journal of Sea Research 49: pp.1.
29. Davis, K. E. and W.B. Russel (1989) "An asymptotic description of transient settling
and ultrafiltration of colloidal dispersions" Physics of Fluids A 1: pp.82.
30. de Krester, G.R., P.S. Usher, et al. (2001) "Rapid filtration measurement of dewatering
design and optimization parameters" AIChE Journal 47(8): pp.1758.
130
31. DeBoer, D.E. (1990) The role of suspension characteristics in continuous gravity
thickening. Department of Civil and Construction Engineering. Iowa, Iowa State University:
pp.366.
32. Dell, C.C. and M.B. Kaynar (1968) "Channelling in flocculated suspensions"
Filtration & Separation 5(4): pp.323.
33. Dell, C.C. and H. T.W. Keleghan (1970) "Compressibility of flocculated clay
sediments" Filtration & Separation 7: pp.419.
34. Dixon, D.C. (1979) Theory of gravity thickening. Progress in Filtration and
Separation. R. J. Wakeman (Ed), Elsevier Press.
35. Dixon, D.C. (1982) "Compression effects in batch settling tests" Journal of the
Environmental Engineering Division, (ASCE) 108 (EE4): pp.499.
36. Fitch, B. (1966) "A mechanism of sedimentation" Industrial and Engineering

Chemistry Research Fundamentals 5(1): pp.129.
37. Fitch, B. (1979) "Sedimentation of flocculant suspensions: State of the art" AIChE
Journal 25 (6): pp.913.
38. Fitch, B. (1983) "Kynch theory and compression zones" AIChE Journal 29 (6):
pp.940.
39. Fitch, B. (1993) "Thickening theories-an analysis" AIChE Journal 39 (1): pp.27.
40. Font, R. (1988) "Compression zone effect in batch sedimentation" AIChE Journal 34
(2): pp.229.
41. Font, R. (1991) "Analysis of the batch sedimentation test." Chemical Engineering
Science 46 (10): pp.2473.
42. Forbes, D. L. H. (1912) "The settling of mill slimes" Engineering and Mining
131
43. Garrido, P., R. Burger and F. Concha (2000) "Settling velocities of particulate
systems:11. Comparison of the phenomenological sedimentation-consolidation model with
published experimental results" International Journal of Mineral Process 60: pp.213.
44. Gaudin, A.M. and M.C. Fuerstenau (1962) "Experimental and Mathematical model of
thickening" Transaction of the American Institute of Mining, Metallurgical and Petroleum
Engineers 223: pp.122.
45. Glanister, D. J. and M. T. Abbott (1989) Dewatering Technology and Practice

Conference, Brisbane, Australia.
46. Glasrud, G. G., R. C. Navarrete, et al. (1993). "Settling Behaviors of Iron Oxide
Suspensions" AIChE Journal 39(4): pp.560.
47. Glover, S.M., Y. Yan, et al. (2000) "Bridging flocculation studied by light scattering
and settling" Chemical Engineering Journal 80 (1-3): pp.3.
48. Green, M.D., M. Eberl, et al. (1996) "Compressive yield stress of flocculated
suspensions: determination via experiment" AIChE Journal 42(8): pp.2308.
49. Green, M. D., K.A. Landman, et al. (1998) "Pressure filtration technique for complete
characterization of consolidating suspensions" Industrial & Engineering Chemistry Research
37: pp.4152.
50. Harris, C.C., P. Somasundaran, et al. (1975) "Sedimentation of Compressible

Materials: Analysis of Batch Sedimentation Curve" Powder Technology 11: pp.75.
51. Hermanowicz S.W. and J.J. Ganczarczyk (1983) “Some fluidization characteristics of
biological beds” Biotechnology and Bioengineering 25: pp.1321.
52. Holdich, R. G. and G. Butt (1995) "An experimental study of channelling and solid
concentration during the batch sedimentation of calcite suspensions" Transaction of the
Institution of Chemical Engineers 73 (Part A): pp.833.
132
53. Holdich, R. G. and G. Butt (1996). "Compression and channelling in gravity
sedimentation systems" Minerals Engineering 9: pp.115.
54. Holdich, R. G. and G. Butt (1997) "Solid/liquid separation by sedimentation" Journal

of Process Mechanical Engineering 211(1-Part E): pp.43.
55. Howells, I., K.A. Landman, et al. (1990) "Time-dependent batch settling of
flocculated suspensions" Applied Mathematical Modeling 14 (2): pp.77.
56. Ives, K. J. (1986) Developments in deep bed filtration. World Congress III of
Chemical Engineering, 3,Tokyo.
57. Jones, G. N. (1986) Channelling during gravity thickening. PhD Thesis, Department
of Civil and Environmental Engineering. Durham, N.C., Duke University: pp.156.
58. Knoer, P. (1983) Compression zone requirements of gravity thickeners. Master Thesis,
Department of Chemical Engineering. Houston, University of Houston: pp.179.
59. Kopp, J. and N. Dichtl (2001) "Influence of the free water content on the
dewaterability of sewage sludges" Water Science and Technology 44(10): pp.177.
60. Kos, P. (1978) Gravity thickening of sludges. PhD Thesis, Department of Civil
Engineering. Massachusetts, University of Massachusetts: pp.306.
61. Kynch, G. J. (1952) "A theory of sedimentation" Transaction of Faraday Society 48:
pp.166.
62. Landman, K.A., L.R. White and R. Buscall (1988) “The continuous flow gravity
thickener: Steady state behaviour” AIChE J., 34(2):pp.239.
63. Landman, K.A. and L.R. White (1994) "Solid/liquid separation of flocculated
suspensions" Advances in Colloid and Interface Science 51: pp.175.
64. Landman, K.A. and L.R. White (1997). "Predicting filtration time and maximizing
throughput in a pressure filter" AIChE Journal 43 (12): pp.3147.
133
65. Landman, K.A., J.M. Stankovich, et al. (1999) "Measurement of the filtration
diffusivity D( ) of a flocculated suspensions" AIChE Journal 45(9): pp.1875.
66. Leenvaar, J.and M. Rebhun (1983) “Strenght of ferric hydroxide flocs” Water
Research 17(8): pp.895.
67. Lester, D. R., S. P. Usher and P. J. Scales (2005) "Estimation of the hindered settling
function R( ) from batch-settling tests" AIChE Journal 51: pp.1158.
68. Leva, M. (1959) "Fluidization" Journal of the Franklin Institute 268(4): pp.320.
69. Li, D.H. and J.J. Ganczarczyk (1987) “Stroboscopic determination of settling velocity,
size and porosity of activated sludge floc” Water Research 21(3): pp.257.
70. McCabe, L. W., C. J. Smith, et al. (1980) Unit Operations of Chemical

Engineering, 6th ed, September 2000, pp.1132.
71. Novales B., A.S. Papineaus and M.A.V. Axelos (2003) “Characterization of emulsions
and suspensions by video image analysis” Colloids and Surfaces A 221: pp.81.
72. O'Donnell, A. J. (1972) Compression of slurries. PhD Thesis, School of Chemical

Engineering. Sydney, University of New South Wales: pp.265.
73. Onal, G., M. Ozer and F. Aslan (2003) "Sedimentation of clay in ultrasonic medium"
Minerals Engineering 16 (2): pp.129.
74. Osborne, G.D. (1990) Gravity thickening. Solid-liquid Separation. Ed. L. Svarovsky,
Butterworth & Co.pp.132.
75. Owen, A.T., P.D. Fawell, et al. (2002). "The impact of polyacrylamide flocculant
solution age on flocculation performance" International Journal of Mineral Processing 67:
pp.123.
134
76. Park, K.H., J.R.G Andrews. and P.H.T. Uhlherr (1983) Batch sedimentation and
thickener behavior. PACHEC- SEOUL.
77. Renko, E.K. (1998) “Modelling hindered batch settling part II: A model for computing
solids profile of calcium carbonate slurry”, Water SA, 24(10):pp.331.
78. Rhee, H.K., R. Aris and N.R Amundson (1986) First order partial differential
equations, Theory and Application of Single Equation, 1, Prentice-Hall, pp. 350.
79. Rushton, A., A.S. Ward, et al. (1996) Solid-Liquid Filtration and Separation
Technology, Weinheim, Germany, VCH.
80. Sawai, T., M. Yamazaki, et al. (1990) "Improvement of sedimentation and dewatering
of municipal sludge by radiation" Radiation Physics and Chemistry 35(1-3): pp.465.
81. Scott, K. J. (1968a) "Thickening of calcium carbonate slurries" Industrial &

Engineering Chemistry Fundamentals 7: pp.484.
82. Scott, K. J. (1968c) "Theory of thickening: factors affecting settling rate of solids in
flocculated pulps" Transactions of the Institution of Mining and Metallurgy 77: pp.85.
83. Shannon, P.T., R.D. Dehaas, et al. (1964) "Batch and continuous thickening"
Industrial & Engineering Chemistry Fundamentals 3 (3): pp.250.
84. Stickland, D.A (2005) Solid-Liquid Separation in the Water and Wastewater
Industries. PhD Thesis, Department of Chemical Engineering. Melbourne, The University of
Melbourne, pp.496.
85. Somasundaran, P. (1981) Thickening or dewatering of slow-settling mineral

suspensions. 13th International Mineral Processing Congress, Warsaw, Amsterdam, Elsevier.
86. Stokes, G.G. (1851) "On the effect of the internal friction of fluids on the motion of
pendulums" Transactions of the Cambridge Philosophical Society 9: pp.8.
135
87. Svarovsky, L. (1990) Solid-Liquid Separation, London, Butterworths.
88. Swift, J.D., K. Simic, et al. (2004) "A study of the polymer flocculation reaction in a
linear pipe with a focused beam reflectance measurement probe" International Journal of
Mineral Processing 73 (2-4): pp.103.
89. Talmage, W.P. and E.B. Fitch (1955) "Determining Thickener Unit Areas" Industrial
Engineering Chemistry 47 (1): pp.38.
90. Tonmukayakul, N. (1998) The effect of moving bed on rate of settling in batch
sedimentation. Master Thesis, Department of Chemical and Metallurgical Engineering.
Melbourne, RMIT: pp.135.
91. Tory, E.M. and D.K. Pickard (1982) "Extensions and refinements of a Markow Model
of sedimentation" Journal of Mathematical Analysis and Applications 86 (2): pp.442.
92. Tory, E.M. and P. Shannon (1965) "Reappraisal of the concept of settling in
compression" Industrial & Engineering Chemistry Fundamentals 4(2): pp.194.
93. Usher, S.P. (2002) Suspension dewatering: Characterization and optimization. PhD
Thesis, Department of Chemical Engineering. Melbourne, The University of Melbourne:
pp.326.
94. Usher, S.P., R.G. de Kretser and P.J Scales (2001). "Validation of a new filtration
technique for dewaterability characterization." AIChE Journal 47(7): pp.1561.
95. Vesilind, P.A. (1968) "Design of prototype thickeners from batch settling tests" Water
and Sewage Works pp.302.
96. Vesilind, P.A. and N.G. Jones (1990) "A reexamination of the batch-thickening curve"
Research Journal of Water Pollution Control Federation 62: pp.887.
97. Vesilind, P.A. and N.G. Jones (1993) "Channelling in batch thickening" Water
Science Technology 28 (1): pp.59.
136
98. Wakeman, R. J. (1981) "Thickening and Filtration: A Review and Evaluation of
Recent Research" Transaction Institution of Chemical Engineering 59: pp.260.
99. Weir, S. and M.G. Moody (2002) "The importance of flocculant choice with
consideration to mixing energy to achieve efficient solid/liquid separation" Minerals
Engineering 16: pp.109.
137
A APPENDIX A – THE
EXPERIMENTAL DATA
The experimental values for four different initial concentrations are
contained in the CD.
A-1
B APPENDIX B - THE
CONCENTRATION LOCI CURVE DATA
Height-Time (H vs. T) and Height-Concentration (H vs.C) Graphs
Values
B-1
The Concentration Loci Curve Data
Height-Time (H vs. T) Graphs
Height versus Time graphs data are given corresponding to initial concentrations.
Initial Concentration 250 g/l
Conc 200 g/l Conc 250 g/l Conc 500 g/l Conc 750 g/l Conc 1000 g/l
Time Height Time Height Time Height Time Height Time Height
(min) (cm) (min) (cm) (min) (cm) (min) (cm) (min) (cm)
0 128.0 0 00.00 0 00.00 0 00.00 0 00.00
49 49.34 42 41.34 42 34.90 71 37.76 43 30.81
64 48.53 49 45.35 48 37.36 88 38.27 94 35.92
185 45.06 49 46.57 71 39.81 94 39.92 111 35.81
276 41.44 64 48.53 88 40.42 110 39.91 185 39.40
322 40.07 94 41.97 202 38.23
120 41.50 208 39.46
231 39.66
293 38.54
B-2
Conc 1250 g/l Conc 1500 g/l
Time Height Time Height
(min) (cm) (min) (cm)
0 00.00 0 00.00
43 28.77 44 18.18
66 30.51 69 19.20
88 32.25 92 19.30
116 33.87 115 19.80
139 34.28 138 19.70
162 34.89 158 18.78
185 35.40 181 18.38
202 34.24 203 17.52
230 35.67 229 19.05
248 35.57 249 18.85
276 36.49 272 18.44
293 36.59 295 17.93
317 17.89
B-3
Clear Liquid Conc 700 g/l Conc 1000 g/l Conc 1200 g/l
Time Height Time Height Time Height Time Height
(min) (cm) (min) (cm) (min) (cm) (min) (cm)
0 128 1 13.16 21 20.13 24 19.15
33 116.38 21 23.96 24 22.93 47 25.46
60 98.73 42 36.72 43 31.25 70 35.14
75 92.30 48 40.79 48 35.38 88 39.21
84 81.55 64 48.64 65 42.18 117 42.28
97 77.49 72 56.87 86 53.36 110 48.08
108 66.65 86 59.76 109 58.59 132 51.04
154 63.66 164 61.43 200 57.88 155 54.93
200 59.66 187 60.1 278 57.11 187 55.50
233 59.60 346 56.40 200 54.99
278 58.84 428 55.07 232 56.16
369 57.67 529 55.46 301 55.91
392 57.80 611 54.58 337 54.65
460 57.46 346 55.24
497 56.34 360 54.71
506 57.16 369 55.37
519 56.42 437 55.07
588 56.15 497 54.02
643 56.95 620 54.30
679 55.38
B-4
Conc 1300 g/l Conc 1500 g/l
Time Height Time Height
(min) (cm) (min) (cm)
0 5.8 23 7.14
24 15.42 67 13.58
66 20.66 135 15.63
134 25.82 183 17.9
184 30.47 229 19.86
185 33.53 272 20.44
133 39.23 317 20.84
156 42.63 389 25.4
201 46.23 412 27.35
246 48.87 453 28.37
315 50.51 504 30.33
383 51.38 527 31.59
460 51.59
551 53.04
588 52.64
B-5
Clear Liquid Conc 900 g/l Conc 1000 g/l Conc 1100 g/l Conc 1200 g/l
0 128 0 5.49 0 4.84 0 3.53 0 2.87
60 97.99 21 22.70 21 20.14 21 14.37 22 5.94
130 78.13 25 26.24 43 33.82 43 28.96 44 20.81
280 74.04 42 36.24 72 49.67 87 50.28 66 32.25
427 71.58 49 39.93 108 65.09 119 64.17 88 40.85
553 68.68 64 47.25 166 74.42 165 72.81 118 51.96
633 66.11 86 57.27 233 67.03
770 66.09 108 68.47 542 68.40
130 76.47
B-6
Conc 1300 g/l Conc 1400 g/l Conc 1500 g/l Conc 1600 g/l
Time Height Time Height Time Height Time Height
(min) (cm) (min) (cm) (min) (cm) (min) (cm)

0 22
3.4 3.29 23 3 22 1.28
90 115 2.18
10.88 8.75 114 5.47 68
134 229 2.98
34.23 13.39 204 7.86 136
178 340 3.41
50.9 22.05 318 8.54 183
291 293 3.68
58.98 35.99 434 12.89 250
405 406 3.84
64.92 50.94 367 25.06 365
542 506 4.23
67.28 55.30 436 40.63 433
598 3.86
63.19 566 50.92 501
689 3.74
61.49 688 53.96 616
3.69
775
B-7
Clear Liquid Conc 850 g/l Conc 950 g/l Conc 1050 g/l Conc 1150 g/l
0 128.00 10 13.18 1 9.98 0 6.76 0 5.43

21 16.15
11 124.43 9 27.93 21 21.3 21 19.38
43 27.62
35 119.09 8 40.44 25 26.42 43 31.98
65 36.62
58 113.50 16 50.65 49 39.25 65 40.88 87 45.76
132 59.25
82 105.79 18 81.43 72 50.51 72 47.01
154 63.01
105 101.40 19 97.05 96 58.47 96 55.68
177 68.33
122 95.38 25 103.63 131 68.39 131 64.46 199 69.87
222 70.07
144 89.24 42 103.32 153 73.13 153 71.49
268 69.84
166 83.77 64 91.95 176 77.75 176 74.46
336 69.39
212 78.07 72 85.06 198 77.07 221 75.02 427 68.01
519 67.16
244 74.42 86 60.94 257 75.13 267 72.63
587 66.97
290 72.14 108 73.87 325 73.25 336 69.91
678 66.77
336 70.54 153 80.75 393 70.64 404 68.11 701 67.03
427 69.13 166 82.23 462 70.27 473 67.92
610 69.12 530 69.59 541 67.29
678 67.80 611 68.53 655 67.21
701 67.57 747 67.18
826 70.81 815 67.53
B-8
Conc 1250 g/l Conc 1350 g/l Conc 1450 g/l Conc 1500 g/l Conc 1600 g/l
0 4.76 0 4.15 0 3.43 23 4.57 0 2.75
22 9.06 23 6.58 46 5.91 69 5.91 91 5.07
44 22.06 46 9.16 69 9.87 137 8.15 228 6.91
66 28.27 69 13.76 115 14.13 183 13.54 272 13.49
94 40.88 89 23.27 161 19.1 252 20.94 298 17.47
110 45.09 111 28.83 207 23.58 295 25.2 343 19.65
132 52.24 134 34.18 226 25.78 208 31.76 364 20.47
155 57.2 163 43.07 185 31.82 299 39.22 317 31.72
187 60.83 201 52.15 247 39.84 338 43.43 407 33.83
222 63.91 291 61.2 292 44.79 391 51.08 453 37.19
268 64.79 337 62.67 337 52.75 474 57.05 521 41.09
336 67.19 405 64.18 451 62.48 565 57.61 566 45.97
405 67.57 473 66.23 507 63.41 643 59.62 619 50.38
496 66.92 541 66.19 552 63.46 712 60.02 665 54.31
564 66.81 633 66.06 554 63.77 757 61.89 725 55.5
655 66.61 724 65.32 712 63.55
724 66.48 792 65.86 757 63.6
815 66.96
B-9
Height-Concentration (H vs. C) Graphs
The Height vs. Concentration graphs data are given corresponding to initial concentrations.
42 min 83 min 125 min 167 min 208 min
Height Conc Height Conc Height Conc Height Conc Height Conc
(cm) (g/l) (cm) (g/l) (cm) (g/l) (cm) (g/l) (cm) (g/l)
110 100 102 100 106 100 104 100 102 100
46.86 200 49.59 200 51.88 200 54.16 200 45.51 200
31.44 500 39.7 500 42.67 500 43.61 500 40 500
25.89 750 37.67 750 37.5 750 36.64 750 35.77 750
26 1000 27.45 1000 35.34 1000 33.98 1000 35 1000
28.79 1250 30.03 1250 33.83 1250 34.87 1250 35.37 1250
27.24 1500 19.18 1500 19.81 1500 18.8 1500 17.48 1500
B-10
250 min 292 min
Height Conc Height Conc
(cm) (g/l) (cm) (g/l)
100 100 101 100
46.33 200 41.52 200
38 500 41 500
34.91 750 40 750
36 1000 39.96 1000
35.56 1250 36.48 1250
18.85 1500 18.48 1500
B-11
128 500 90 0 70 0 62 0 60 0
43 710 63 700 58 700 59 700 58 700
35.6 800 53 1000 57 1000 57 1000 58 1000
33.7 900 43 1200 50 1200 53 1200 57 1200
31.4 1000 24 1300 25 1300 42 1300 50 1300
28.39 1100 21 1400 24 1400 35 1350 30 1400
24 1200 17 1500 18 1500 30 1300 22 1500
20 1300 22 1400
18 1400 12 1500
13 1500
B-12
250 min 292 min 333 min 375 min
Height Conc Height Conc Height Conc Height Conc
(cm) (g/l) (cm) (g/l) (cm) (g/l) (cm) (g/l)
60 0 56 0 56 0 56 0
56 700 56 700 55 700 56 700
56 1000 55 1000 55 1000 54 1000
55 1200 54 1200 54 1200 57 1200
50 1300 50 1300 50 1300 55 1300
33 1400 38 1400 42 1400 54 1400
25 1500 18 1500 41 1400 51 1400
38 1400 48 1350
30 1500 43 1350
20 1500
B-13
128 0 85 0 76.00 0 74 0 72 0
89.82 750 81.22 750 74.96 750 72.56 750 71.23 750
63.69 900 75.92 900 74.3 900 71.52 900 70.59 900
54.93 1000 72.88 1000 73.76 1000 71.61 1000 70.52 1000
51.27 1100 69.27 1100 72.18 1100 70.92 1100 69.96 1100
44.68 1200 63.03 1200 69.37 1200 69.32 1200 68.41 1200
36.63 1300 48.46 1300 62 1300 63.07 1300 59.72 1300
7.8 1400 14.34 1400 37.8 1400 43.79 1400 21.86 1400
3.83 1500 4.58 1500 19.3 1400 21.09 1400 8.29 1500
2.5 1600 2.9 1600 12.89 1400 6.74 1500 3.67 1600
6.13 1500 2.57 1600
2.87 1600
B-14
292 min 333 min 375 min 417 min
Height Conc Height Conc Height Conc Height Conc
(cm) (g/l) (cm) (g/l) (cm) (g/l) (cm) (g/l)
70 0 69 0 68 0 68 0
69.71 750 68.66 750 67.87 750 67.27 750
69.3 900 67.8 900 67.3 900 66.52 900
68.88 1000 68.55 1000 67.83 1000 67.35 1000
69.12 1100 67.38 1100 67.68 1100 66.15 1100
67.67 1200 66.78 1200 66.33 1200 68.16 1200
63.66 1300 65.66 1300 66.32 1300 67.43 1300
54.94 1400 54.57 1400 56.91 1400 61.7 1400
35.4 1400 34.39 1400 48.72 1400 34 1400
24.28 1400 25.18 1400 39.88 1400 21.63 1500
14.99 1500 15.76 1500 28.33 1400 4.11 1600
2.42 1600 2.27 1600 16.3 1500
2.96 1600
B-15
125 0 92 0 90 0 82 0 75 0
105 850 85 850 85 850 80 850 74 850
42 950 68 850 80 850 75 950 67 850
27 1150 63 950 80 850 65 1150 62 950
20 1250 50 1150 75 950 55 1250 55 1150
10 1350 40 1250 65 1150 40 1350 52 1250
5 1450 28 1350 50 1250 15 1450 46 1350
3 1500 10 1450 32 1350 8 1500 42 1450
2 1600 5 1500 15 1450 4 1600 40 1500
3 1600 5 1500 5 1600
3 1600
B-16
250 min 292 min
Height Conc Height Conc
(cm) (g/l) (cm) (g/l)
76 0 77 0
77 850 75 850
73 950 72 850
68 1150 68 950
62 1250 62 1150
51 1350 56.4 1250
42 1450 43 1350
38 1500 37 1450
30 1600 34 1500
10 16 1600
B-17
C APPENDIX C - THE DATA OF
DESCENT PATH GRAPHS
C-1
The Descent Path Curve
The descent path curves data are given corresponding to initial concentrations.
Initial concentration 250 g/l
Time Height
(min) (cm) 90% 80% 70% 60% 50% 40% 30% 20% 10%
0 128 115.2 102.40 89.60 76.80 64.00 51.20 38.40 25.60 12.80
1 93.53 0 74.83 65.47 56.12 46.77 37.41 28.06 18.71 9.35
26 44.94 84.18 35.95 31.46 26.96 22.47 17.97 13.48 8.99 4.49
41 46.88 40.44 37.51 32.82 28.13 23.44 18.75 14.07 9.38 4.69
49 49.34 42.20 39.48 34.54 29.61 24.67 19.74 14.80 9.87 4.93
64 48.53 44.41 38.82 33.97 29.12 24.27 19.41 14.56 9.71 4.85
185 45.06 43.68 36.05 31.54 27.04 22.53 18.02 13.52 9.01 4.51
276 41.44 40.55 33.15 29.01 24.86 20.72 16.58 12.43 8.29 4.14
322 40.07 37.30 32.06 28.05 24.04 20.04 16.03 12.02 8.01 4.01
36.06
C-2
Time Height
(min) (cm) 90% 80% 70% 60% 50% 40% 30% 20% 10%
0 128.00 115.20 102.40 89.60 76.80 64.00 51.20 38.40 25.60 12.80
35 114.95 103.45 91.96 80.46 68.97 57.47 45.98 34.48 22.99 11.49
54 103.16 92.84 82.53 72.21 61.90 51.58 41.26 30.95 20.63 10.32
76 93.30 83.97 74.64 65.31 55.98 46.65 37.32 27.99 18.66 9.33
97 77.49 69.74 61.99 54.25 46.50 38.75 31.00 23.25 15.50 7.75
119 65.63 59.07 52.50 45.94 39.38 32.81 26.25 19.69 13.13 6.56
154 63.66 57.29 50.93 44.56 38.19 31.83 25.46 19.10 12.73 6.37
278 58.84 52.95 47.07 41.19 35.30 29.42 23.53 17.65 11.77 5.88
383 57.23 51.51 45.78 40.06 34.34 28.62 22.89 17.17 11.45 5.72
392 57.80 52.02 46.24 40.46 34.68 28.90 23.12 17.34 11.56 5.78
451 56.98 51.28 45.59 39.89 34.19 28.49 22.79 17.09 11.40 5.70
483 57.13 51.42 45.70 39.99 34.28 28.56 22.85 17.14 11.43 5.71
506 57.16 51.44 45.72 40.01 34.29 28.58 22.86 17.15 11.43 5.72
519 56.42 50.78 45.14 39.50 33.85 28.21 22.57 16.93 11.28 5.64
588 56.15 50.54 44.92 39.31 33.69 28.08 22.46 16.85 11.23 5.62
611 56.32 50.69 45.05 39.42 33.79 28.16 22.53 16.90 11.26 5.63
643 56.95 51.26 45.56 39.87 34.17 28.48 22.78 17.09 11.39 5.70
665 56.66 50.99 45.33 39.66 34.00 28.33 22.66 17.00 11.33 5.67
C-3
Time Height
(min) (cm) 90% 80% 70% 60% 50% 40% 30% 20% 10%
0 128 115.20 102.40 89.60 76.80 64.00 51.20 38.40 25.60 12.80
34 114.21 102.79 91.37 79.95 68.53 57.11 45.68 34.26 22.84 11.42
52 103.63 93.27 82.90 72.54 62.18 51.82 41.45 31.09 20.73 10.36
65 93.78 84.40 75.02 65.65 56.27 46.89 37.51 28.13 18.76 9.38
88 77.67 69.90 62.14 54.37 46.60 38.84 31.07 23.30 15.53 7.77
130 75.4 67.86 60.32 52.78 45.24 37.70 30.16 22.62 15.08 7.54
153 69.86 62.87 55.89 48.90 41.92 34.93 27.94 20.96 13.97 6.99
221 67.08 60.37 53.66 46.96 40.25 33.54 26.83 20.12 13.42 6.71
292 66.02 59.42 52.82 46.21 39.61 33.01 26.41 19.81 13.20 6.60
375 65.82 59.24 52.66 46.07 39.49 32.91 26.33 19.75 13.16 6.58
435 65.51 58.96 52.41 45.86 39.31 32.76 26.20 19.65 13.10 6.55
495 65.18 58.66 52.14 45.63 39.11 32.59 26.07 19.55 13.04 6.52
C-4
Time Height
(min) (cm) 90% 80% 70% 60% 50% 40% 30% 20% 10%
11 128.00 115.2 102.40 89.60 76.80 64.00 51.20 38.40 25.60 12.80
35 119.09 0 95.27 83.36 71.45 59.55 47.64 35.73 23.82 11.91
58 113.50 107.1 90.80 79.45 68.10 56.75 45.40 34.05 22.70 11.35
82 105.79 8 84.63 74.05 63.47 52.90 42.32 31.74 21.16 10.58
105 101.40 102.1 81.12 70.98 60.84 50.70 40.56 30.42 20.28 10.14
122 95.38 5 76.30 66.76 57.23 47.69 38.15 28.61 19.08 9.54
144 89.24 95.21 71.39 62.47 53.55 44.62 35.70 26.77 17.85 8.92
166 83.77 91.26 67.01 58.64 50.26 41.88 33.51 25.13 16.75 8.38
212 78.07 85.84 62.46 54.65 46.84 39.04 31.23 23.42 15.61 7.81
244 74.42 80.32 59.54 52.10 44.65 37.21 29.77 22.33 14.88 7.44
290 72.14 75.39 57.71 50.50 43.28 36.07 28.85 21.64 14.43 7.21
336 70.54 70.26 56.43 49.38 42.32 35.27 28.21 21.16 14.11 7.05
427 69.13 66.98 55.30 48.39 41.48 34.56 27.65 20.74 13.83 6.91
610 69.12 64.92 55.30 48.38 41.47 34.56 27.65 20.74 13.82 6.91
678 67.80 63.48 54.24 47.46 40.68 33.90 27.12 20.34 13.56 6.78
62.22
62.21
61.02
C-5
D APPENDIX D - STREAK
PHOTOGRAPHY READINGS AND THE
HIGH SPEED CAMERA PHOTOS
D-1
Table D-1 - Readings of (time vs. height) from streak photographic observation
Time (min) Height (cm)
1.30 130.27
2.90 128.84
4.33 127.9
6.25 126.81
9.58 125.03
13.20 122.81
17.98 120.1
22.05 117.87
27.97 114.52
37.47 108.94
44.38 104.78
56.45 97.12
59.43 95.3
60.08 95
63.40 94.5
70.03 90.2
81.64 81.4
84.10 80.3
98.26 75.3
104.84 70.9
110.10 66.9
112.96 65
120.08 63.7
123.02 63.05
125.35 63
D-2
131.51 63.2
141.33 62.3
152.63 61.5
174.97 60.6
180.08 59.8
213.43 58.7
222.56 57.4
240.08 57.1
D-3
a. Position 1
b. Position 2
c. Position 3
Figure D-1 - Particle 2 in three different positions in the channel
D-4
a. Position 1
b. Position 2
c. Position 3
D-5
a. Position 1
b. Position 2
c. Position 3
D-6
a. Position 1
b. Position 2
c. Position 3
D-7
a. Position 1
b. Position 2
c. Position 3
D-8
a. Position 1
b. Position 2
c. Position 3

D-9
a. Position 1
b. Position 2
c. Position 3
D-10
a. Position 1
b. Position 2
c. Position 3
D-11
a. Position 1
b. Position 2
c. Position 3
Figure D-9- Particle 10 in three different positions in the channel
D-12
E APPENDIX E - DRY AND WET
PARTICLE SIZE ANALYSIS
E-1
Volume (%)
10 100
90
80
70
60
50
40
30
20
10
0 0
0.1 1.0 10.0 100.0
Particle Diameter (µm.)
Figure E-1 - Wet particle size distribution of CaCO3
Table E-1 - Analysis results of wet particle size
E-2
Figure E-2 - Dry particle size distribution of CaCO3
Table E-2 - Analysis results of dry particle size
E-3
F APPENDIX F - HEIGHT AND
CONCENTRATION CALIBRATIONS
F-1
Calibrations
Height calibration
For this calibration, the top and bottom of the settling tube was marked then the height values
were noted against signal values in millivolts. Figure F.1 shows the signal values and
corresponding heights as solid points.
140
120
y = 8.5969x - 552.46
100 2
R = 0.9963
Height (cm)
80
60
40
20
0
64 66 68 70 72 74 76 78 80
Reading from datataker (mV)
Figure F-1 - Height calibration curve
The solid line represents the linear trendline through these points. The resultant linear
equation is:
h = 8.5969 z − 552.46
where z is the reading from the datataker in millivolts and h is the height in cm. The R2 was
found to be 0.9963. The error is estimated ±2 mm.
F-2
Concentration calibration
A second calibration graph was prepared for determining the concentration values. For this
calibration, samples of slurries with different concentrations were prepared in the batch
settling column. A mixer was employed to maintain homogeneity of the mixed slurry while
the voltage signals from the data taker were recorded. In addition, the samples were vacuum-
dried and weighed to determine the concentrations. To form a relationship between
concentration and voltage signal, these values were plotted on the graph (Figure 3.2) and
fitted with a polynomial equation. The fitted 3rd order polynomial equation, which is shown
below, were used for calibrations of concentrations. The R2 was found to be 0.9989. The error
is estimated ±1 g/l.
1250
Solid concentration (g/l)
3 2
y = -6.7924x + 441.1x - 9608.7x + 70877
1000 2
R = 0.9989
750
500
250
0
18 19 20 21 22 23 24 25
Signal from datataker (mV)
Figure F-2 - Concentration calibration curve
F-3
G APPENDIX G - CACO3 BATCH
SETTLING TESTS
G-1
CaCO3 batch settling tests
Density 2.681 g/cm3
250 g/L (=228.68 g/L) 500 g/L (=421.41 g/L) 1000 g/L (728.35 g/L)
Time Height Time Height Time Height
(min) (ml) (min) (ml) (min) (ml)
0 531 0 581 0 314
0.5 485 1 525 1 308
1 470 2 496 1.5 305
1.5 447 2.5 481 2 303
2 423 3 466 2.5 300
2.5 402 3.5 449 3 296
3 377 4 437 3.5 293
3.5 355 4.5 420 4 290
4 332 5 408 4.5 287
4.5 310 6 380 5 284
5 287 6.5 369 5.5 280
5.5 265 7 357 6 276
6 244 7.5 343 6.5 275
7 200 8 332 7 273
7.5 182 8.5 321 8 265
8 168 9 311 9 258
8.5 160 10 293 10 251
9 153 10.5 284 11 245
9.5 148 11 277 12 238.5
10 143 11.5 270 13 231
10.5 139 12.5 258 14 225
11 135 13 250 15 217
11.5 132 14 240 16.5 208
G-2
12 129.5 15 230 17.5 200
12.5 126 16 221 18.5 195
13 123 17 214 19.5 189
13.5 120.5 18 209 21 183
14 118 19 201 22 181
15 114 20 197 23 180
16 110 21 192 24 178
17 105.5 22 188 25 176
18 102 23 182 26 175
19 100.5 24 180 27 174.5
20 99.5 25 179.5 32 171
22 98 26 178 34 170
23 97 27 177.5 36 169.5
24 96 28 176 43 168
25 95.5 29 174 79 164
26 95 30 172 112 164
27 95 31 171
34 92 32 170.5
43 90.1 34 170
58 88.5 35 170
71 88 39 169.5
79 87.7 47 168
139 87.6 48 166
186 87.5 57 162
G-3

Sedimentation

Uploaded by

Copyright:

Available Formats

Sedimentation

Uploaded by

Document Information

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

Sedimentation

Uploaded by

Copyright:

Available Formats

What is the thesis about?

What is the thesis about?

What challenges are discussed in the thesis?

What challenges are discussed in the thesis?

A STUDY OF CHANNELLING BEHAVIOUR IN

A thesis submitted in fulfillment of the requirements for the

School of Civil and Chemical Engineering

editorial work, paid or unpaid, carried by a third party is acknowledged.

under him. I fell honoured to have worked with you.

to me throughout this journey.

Sedimentation, batch settling, hindered settling function, flocculated suspension, channelling,

stepped pressure filtration, sedimentation modes, floc, concentration gradients, compressible

Bayrak, N and J. O’Donnell (2004) “Principles of channelling in batch sedimentation”,

O’Donnell, J. and N. Bayrak (2004) “Channelling in batch sedimentation of calcium

Batch sedimentation is a method that enables us to understand the mechanism of compaction

great amount of flocs upwards.

performed employing material characteristics of the material and graphical documentations

results and the modes of sedimentation explained by Lester et al. (2005).

gained from this study is valuable to maximize the thickening process.

1.2 RESEARCH OBJECTIVES ..........................................................................4

1.3 STRUCTURE OF THESIS ...........................................................................5

2.1 LITERATURE REVIEW ..............................................................................8

2.2 BATCH SETTLING ...................................................................................10

2.3 FUNDAMENTAL THEORIES OF THICKENING AND BATCH

2.3.1 Stokes Law ................................................................................................... 15

2.3.2 Coe and Clevenger........................................................................................ 15

2.3.3 Comings ....................................................................................................... 16

2.3.4 Kynch ........................................................................................................... 16

2.3.5 Talmage and Fitch ........................................................................................ 17

2.3.6 Recently Developed Theories ....................................................................... 17

2.3.7 Material Properties....................................................................................... 18

2.3.7.1 Compressive Yield Stress Py (φ ) .............................................................................18

2.3.7.3 Hindered Settling Function, R(φ ) ...........................................................................19

2.3.7.4 Solids Diffusivity, D (φ ) ..........................................................................................20

2.4 GOVERNING EQUATION ........................................................................20

2.4.1 Boundary Conditions .................................................................................... 22

2.4.2 Modes of Sedimentation by Burger ............................................................... 30

2.5 CHANNELLING IN BATCH SEDIMENTATION ....................................33

2.5.1 Possible Causes of Channelling..................................................................... 33

2.5.2 Mechanism of Channelling ........................................................................... 34

2.6 VOLCANO FORMATION .........................................................................40

2.7 REVERSE CONCENTRATION GRADIENTS ..........................................41

3.1 EXPERIMENTAL SETUP .........................................................................43

3.1.1 The Settling Tube ......................................................................................... 44

3.1.2 Mobile Platform............................................................................................ 45

3.1.3 Gamma (γ) Radiation Densitometer and Radiation Detector.......................... 45

3.1.4 Measurement of Radiation and Rate Meter.................................................... 48

3.1.5 Data Collection............................................................................................. 49

3.2 SUSPENSION MATERIAL........................................................................52

3.2.1 Sample Preparation ....................................................................................... 52

3.2.2 Determining Concentration ........................................................................... 52

3.2.3 Solid Particle Properties................................................................................ 52

3.2.4 Flocculation and Flocculant .......................................................................... 53

3.2.5 Calibrations .................................................................................................. 55

3.3 EXPERIMENTAL PROCEDURES ............................................................55

3.4.1 Stepped Pressure Filtration Technique .......................................................... 58

3.4.2 Apparatus and Measurements ....................................................................... 58

4.1 PHOTOGRAPHIC STUDY ........................................................................63

4.2 PURPOSE OF WORK ................................................................................64

4.3 STREAK PHOTOGRAPHY .......................................................................65