Rheological Properties of Lihir Ore

Rheological properties of Lihir ore
Peter William Legge
Bachelor of Engineering (Hons)
A thesis submitted for the degree of Master of Philosophy at
The University of Queensland in 2019
Sustainable Minerals Institute (SMI)
Julius Kruttschnitt Mineral Research Centre (JKMRC)

Abstract
The Papuan New Guinian gold mine Lihir has a rheology problem
within its grinding circuit. Highly viscous slurry pools in the semi -
autogenous grinding (SAG) mill which reduces the power draw and
causes extra wear and tear. This thesis aims to characterise the
rheology and the sedimentation rate of Lihir ore slurry.
Lihir has a complicated mineralogy and is currently defined by

nine alteration domains. Of these nine domains, seven domains had
samples characterised in this thesis. Most tests were performed o n
slurry created from the provided crush ore sieved to less than 1 mm.
Other tests were performed after grinding the sample for five minutes.
The ores had a large natural variability in particle size and particle
size distribution. Furthermore, the slurrie s’ chemistry was not altered
but the pH was monitored during the course of this investigation and
found to vary greatly between different samples.
The slurry rheology was measured through two different

rotational rheology techniques which are referred to in this thesis as
coaxial cylinder rheometry and helical flow rheometry. Helical flow
rheometry is a modified cylinder rheometry technique where the slurry
flows through a measuring chamber allowing the prevention of
sedimentation. The helical flow rheomet ry was found to be
unnecessary for the high concentration slurry because the suspensions
were found to be non or very slow settling which can be more easily
and accurately characterised by the traditional coaxial c ylinder
method. Due to the wide gap configuration a novel data analysis
procedure was used to fit the raw results to the Casson Model. A total
of 105 rheograms were produced.
It was found that Lihir ore slurry has a complicated shearing
profile; this shearing profile is best described as consisting of a time
independent and time dependent components. All tests (with the
ii
exception of one Newtonian sample) indicated time independ ent shear
thinning with yield stress. Time independency, thixotropy and anti -
thixotropy were observed; the precise behaviour was dependent on
both the sample and sample concentration.
The tests showed that concentration and particle size make a

major difference to the slurry rheology; a higher concentration and a
smaller particle size increased the yield stress and apparent viscosity.
There was a great deal of variety between samples within a domain and
between the different domains. Argillic and Lower Epit hermal samples
were found to be much more viscous than the other domains. Using
QXRD (Quantitative X-ray Diffraction) analysis, individual samples
had their rheology linked to amorphous silicates, kaolinite content,
and total clay composition. There was no obvious correlation between
these variables and the apparent viscosity at 50 s -1 so a multilinear
regression analysis was performed.
The regression analysis also found a poor correlation with the

aforementioned variables. However, additional analysis was performed
using the logarithm of the apparent viscosity which indicated a
complicated relationship between the variables. Concentration and
fines content were strongly linked with an increase in the logarithm of
the apparent viscosity. A low to intermedia te pH was found to peak the
viscosity. The effects of the mineral composition are complicated and
require further investigation. Given that all of these effects were
interrelated it is recommended that further experimental work and
analysis be conducted before any plant suggestions are made.
A sedimentation investigation was performed simultaneously to

see if sedimentation could be used as a proxy for viscosity. It was
found that sedimentation rate was linked with a lower viscosity and
with larger particle sizes. Certain slurries were found to swell making
sedimentation analysis compli cated. Another observation which
warrants further investigation is sedimentation was sometimes
accompanied by a time delay.
iii
Declaration by author
This thesis is composed of my original work, and contains no
material previously published or written by anot her person except
where due reference has been made in the text. I have clearly stated
the contribution by others to jointly -authored works that I have
included in my thesis.
I have clearly stated the contribution of others to my thesis as a

whole, including statistical assistance, survey design, data analysis,
significant technical procedures, professional editorial advice,
financial support and any other original research wo rk used or
reported in my thesis. The content of my thesis is the result of work I
have carried out since the commencement of my higher degree by
research candidature and does not include a substantial part of work
that has been submitted to qualify for the award of any other degree or
diploma in any university or other tertiary instit ution. I have clearly
stated which parts of my thesis, if any, have been submitted to qualify
for another award.
I acknowledge that an electronic copy of my thesis must be

lodged with the University Library and, subject to the policy and
procedures of The University of Queensland, the thesis be made
available for research and study in accordance with the Copyright Act
1968 unless a period of embargo has been approved by the De an of the
Graduate School.
I acknowledge that copyright of all material contain ed in my

thesis resides with the copyright holder(s) of that material. Where
appropriate I have obtained copyright permission from the copyright
iv
holder to reproduce material i n this thesis and have sought permission
from co-authors for any jointly authored works included in the thesis.
v
Publications included in this thesis
No publications included.
Submitted manuscripts included in this

thesis
No manuscripts submitted for publication.
Other publications during candidature

No other publications.
Contributions by others to the thesis

This thesis contains QXRD analysis performed by the Queensland
University of Technology.
Experimental work was assisted by Luyan da “Loui” Ncube and

Kaitlyn Porter. Kaitlyn helped with almost every RM180 rheological
test. Luyanda also helped with some of the RM180 rheological tests
and some of the helical flow rheomety tests. Both Luyanda and Kaitlyn
aided me in splitting the samples in preparation for the rheological
tests. Furthermore, Luyanda was of huge assistance in other s ample
preparation tasks such as assisting with sieving and grinding the Lihir
ore.
Ding Sheng Kaw assisted me in analysing some of the

sedimentation videos and sieving some of the samples for sizing which
Luyanda and I hadn’t yet completed.
Tim Napier-Munn aided with the statistical analysis.
Michael “Mick” Kilmartin assisted in designing and constructing

the testing rig.
vi
My supervisors, Frank Shi and Yongjun Peng, conceived this
project and helped with experimental design and the interpretation of
experimental data.
Statement of parts of the thesis submitted to

qualify for the award of another degree
No works submitted towards another degree have been included
in this thesis.
Research Involving Human or Animal

Subjects
No animal or human subjects were involved in this research.
vii
Acknowledgements
Numerous people have helped me in one way or another through
this project:
• Newcrest, for their generous financial support and

supplying the material for this project. In particular I
would like to thank David Seaman, Luke Vollert and Karyn
Gardner who took the time to meet with me and explain the
context for this project and the geology of Lihir
• My advisory team: Frank Shi and Yongjun Peng for their

guidance and academic support through out this project.
• Loui Ncube and Kaitlyn Porter through their help and

support through a very difficult time. Without you I
wouldn’t have gotten half the work done that I managed to
achieve.
• The pilot plant workers and technicians, in particular David

Garcia and Michael “Mick” Kilmartin. David ensured that I
had the space that I needed to conduct my work whilst Mick
managed to come up with any lab gear, I needed whether
finding it in storage or making it from scratch!
• Ryan Anderson who ensured that the In dooroopilly mine

site was a safe place to work and who kindly took the time
to assist me with my risk assessment writing.
• All of my fellow students throughout SMI, in particular The

Jackals cheers and good luck with your projects! Of special
note I would like to acknowledge the people who I shared
an office with:
o Maruf
o John
viii
o Daniel
o Raoni
o Ditend
And my student guide when I first came to the SMI -JKMRC

Juanjo.
• Alice Clark and Neville Pint for their stewardship of the

Sustainable Minerals Institute and the J ulius Kruttschnitt
mineral research centre and support of the student body.
• The administrative support at the JKMRC, in particular

Sherrin Brundle, Melinda Winton, and Khushbu Srivastava
• Pat Griffin from rheological solutions for assisting me with

all HAAKE related matters
• Ashish Kumar from MEP instruments/Anton Paar who was

very helpful with all of the Anton Paar related queries
• Fred Koshir for all the technical support in regard to the

RM180
• Karen Holtham for her assistance in library matters
• Nigel Tiong for helping me with all my IT needs (which

ended up being many!)
• SMI Postgrad, in particular Tess Dobinson for assisting the

student body organise the 2017 SMI HDR conference
• My milestone review panel, Tim Napier -Munn, Elaine

Wightman, and Kym Runge for your feedback, assistance
and advice on how to achieve my Masters. Tim was of great
assistance in the statistical component of this thesis whilst
Elaine was critical in helping me navigate the universities
milestone procedures. Kym handled the relationship with
Newcrest and ensured that the project had sufficient
funding
ix
• Ding for assisting me with the sedimentation analysis
• My family and friends who supported me throughout this

project
• And especially my wife Kirsti, who supported me

throughout this entire project including the move to
Queensland. Without you I wouldn’t have been able to do
it!
x
Financial support
This project was generously supported by Newcrest Mining
Limited.
Keywords
Rheology, sedimentation, viscosity, mining, gold ore, & clay
Australian and New Zealand Standard

Research Classifications (ANZSRC)
ANZSRC code: 090408, Rheology 100%
Fields of Research (FoR) Classification

FoR code: 0914 Resources Engineering and Extractive Metallurgy
xi
Dedication
I dedicate with love this thesis to my wife Kirsti. Your love and
kindness supported me throughout this very challenging chapter in my
life; you were vital to me staying the course and completing this
degree. Now it’s your turn!
xii
Contents
Abstract .............................................................................................................................................................. ii
Declaration by author ........................................................................................................................................ iv
Publications included in this thesis ................................................................................................................... vi
Submitted manuscripts included in this thesis.................................................................................................. vi
Other publications during candidature ............................................................................................................. vi
Contributions by others to the thesis ................................................................................................................ vi
Statement of parts of the thesis submitted to qualify for the award of another degree ................................ vii
Research Involving Human or Animal Subjects ................................................................................................ vii
Acknowledgements ......................................................................................................................................... viii
Financial support ............................................................................................................................................... xi
Keywords ........................................................................................................................................................... xi
Australian and New Zealand Standard Research Classifications (ANZSRC)....................................................... xi
Fields of Research (FoR) Classification .............................................................................................................. xi
Dedication......................................................................................................................................................... xii
Contents .......................................................................................................................................................... xiii
Table of figures .............................................................................................................................................. xviii
Table of tables ............................................................................................................................................... xxiii
List of abbreviations ....................................................................................................................................... xxv
Chapter 1 Introduction ................................................................................................................................... 1
1.1 Background ........................................................................................................................................ 1
1.2 Aims and objectives ........................................................................................................................... 2
1.3 Statement of Sustainability ............................................................................................................... 3
1.4 Thesis structure overview.................................................................................................................. 3
Chapter 2 Literature Review .......................................................................................................................... 5
2.1 Lihir ore .............................................................................................................................................. 5
2.2 Background on rheology theory ........................................................................................................ 8
2.2.1 Time independent behaviour .................................................................................................... 9
2.2.2 Time dependent behaviour ..................................................................................................... 10
2.3 Effect of solids concentration on viscosity ...................................................................................... 13
2.4 Rheology dependency ..................................................................................................................... 16
2.4.1 Solids concentration ................................................................................................................ 17
xiii
2.4.2 Particle size distribution .......................................................................................................... 20
2.4.3 Chemical environment ............................................................................................................ 21
2.4.4 Temperature ............................................................................................................................ 24
2.5 Flow models and the importance of using them............................................................................. 25
2.6 Rheometer selection ....................................................................................................................... 26
2.6.1 Parallel Plate ............................................................................................................................ 26
2.6.2 Cone and plate......................................................................................................................... 27
2.6.3 Coaxial cylinder or ‘cup and bob’ ............................................................................................ 28
2.6.4 Vane ......................................................................................................................................... 29
2.6.5 Capillary rheometer ................................................................................................................. 31
2.6.6 Summary .................................................................................................................................. 32
2.7 Coaxial cylinder shear rate calculations .......................................................................................... 33
2.7.1 Narrow gap .............................................................................................................................. 33
2.7.2 Infinite sea ............................................................................................................................... 35
2.7.3 Krieger approximate solution .................................................................................................. 36
2.7.4 Estelle et al Bingham approximation ....................................................................................... 36
2.8 Measuring the rheology of settling slurries..................................................................................... 37
2.8.1 Elongated coaxial cylinders ..................................................................................................... 39
2.8.2 Debex ....................................................................................................................................... 40
2.8.3 Akroyd-Nguyen helical flow rheometer .................................................................................. 42
2.9 Clay mineralogy ............................................................................................................................... 44
2.9.1 Clay .......................................................................................................................................... 45
2.9.2 Lihir Clays ................................................................................................................................. 49
2.9.3 Rheology measurements of slurries with high clay contents .................................................. 53
2.9.4 Time dependency in clay suspensions..................................................................................... 54
2.9.5 Accurate sizing of clay particles ............................................................................................... 58
2.10 Rheology’s effect on mineral processing performance ................................................................... 59
2.10.1 Grinding circuit ........................................................................................................................ 59
2.10.2 Flotation................................................................................................................................... 62
2.10.3 Other mineral processes.......................................................................................................... 64
2.11 Sedimentation of mineral ores ........................................................................................................ 64
2.12 Predicting Lihir ore slurry rheology ................................................................................................. 66
2.13 Gaps in the literature....................................................................................................................... 68
Chapter 3 Testing rig and rheometer commissioning .................................................................................. 70
xiv
3.1 Design considerations common to both helical flow rheometry and coaxial flow rheometry ....... 71
3.1.1 Sample stability ....................................................................................................................... 72
3.1.2 Predicting if a flow curve is producible with a given rheometer............................................. 73
3.2 Helical flow rheometer commissioning ........................................................................................... 75
3.2.1 Measuring chamber................................................................................................................. 77
3.2.2 Pump........................................................................................................................................ 79
3.2.3 Sump ........................................................................................................................................ 80
3.2.4 Testing rig operation ............................................................................................................... 81
3.3 Rheometer selection ....................................................................................................................... 82
3.3.1 Selection of the HAAKE Viscometer ........................................................................................ 82
3.3.2 Alternative rheometers ........................................................................................................... 82
3.3.3 Anton Paar ............................................................................................................................... 83
3.4 RM180 co-axial cup commissioning ................................................................................................ 84
3.4.1 Time dependency .................................................................................................................... 85
3.5 Summary of design considerations ................................................................................................. 87
Chapter 4 Experimental methodology ......................................................................................................... 89
4.1 Data entry ........................................................................................................................................ 89
4.2 Materials .......................................................................................................................................... 89
4.3 Particle sizing ................................................................................................................................... 90
4.4 Sample preparation ......................................................................................................................... 91
4.4.1 Splitting .................................................................................................................................... 92
4.4.2 Grinding procedure ................................................................................................................. 93
4.5 Sedimentation tests......................................................................................................................... 94
4.6 Measuring pH, dissolved oxygen, and temperature ....................................................................... 95
4.7 HAAKE experimental procedure ...................................................................................................... 95
4.7.1 Premixing the slurry and loading the helical flow rheometer ................................................. 96
4.7.2 Rheology test ........................................................................................................................... 97
4.7.3 Flow rate .................................................................................................................................. 97
4.7.4 Adjusting sample concentration .............................................................................................. 97
4.8 RM180 experimental procedure ..................................................................................................... 98
4.8.1 Mixing the sample ................................................................................................................... 99
4.8.2 Loading and unloading the rheometer .................................................................................... 99
4.8.3 Selecting a bob ...................................................................................................................... 100
4.8.4 Yield stress tests .................................................................................................................... 102
xv
4.8.5 Rheology test: Ramping procedure ....................................................................................... 103
4.8.6 Benefits and limitations ......................................................................................................... 103
4.9 Data reduction ............................................................................................................................... 104
4.9.1 Coaxial cylinder ...................................................................................................................... 104
4.9.2 Modified Estelle method ....................................................................................................... 105
4.9.3 Turbulence ............................................................................................................................. 114
4.9.4 Casson approximation ........................................................................................................... 114
4.10 Summary........................................................................................................................................ 116
Chapter 5 Test results................................................................................................................................. 118
5.1 RM180 calibration and validation of the final testing procedure ................................................. 118
5.2 Confirmation that the master sizing is comparable to sieving ...................................................... 121
5.3 Particle size and size distribution .................................................................................................. 122
5.4 Summary of the rheology and sedimentation data ...................................................................... 128
5.5 Summary of test results................................................................................................................. 132
Chapter 6 Analysis of the slurry rheology .................................................................................................. 134
6.1 Time dependency .......................................................................................................................... 134
6.1.1 Advanced argillic .................................................................................................................... 139
6.1.2 Argillic .................................................................................................................................... 139
6.1.3 Upper epithermal .................................................................................................................. 139
6.1.4 Silica breccia .......................................................................................................................... 140
6.1.5 Lower epithermal .................................................................................................................. 141
6.1.6 Outer biotite .......................................................................................................................... 142
6.1.7 Inner biotite ........................................................................................................................... 142
6.1.8 Comparison to other materials ............................................................................................. 143
6.2 Solids concentration effect on yield stress and viscosity .............................................................. 144
6.3 Grinding effect ............................................................................................................................... 146
6.4 Domain .......................................................................................................................................... 153
6.5 The dependence on ore characteristics ........................................................................................ 160
6.6 Multilinear regression ................................................................................................................... 163
6.7 Summary........................................................................................................................................ 172
Chapter 7 Analysis of slurry stability .......................................................................................................... 174
7.1 Concentration effect...................................................................................................................... 174
7.2 Alteration domain.......................................................................................................................... 175
7.3 Particle size and particle size distribution ..................................................................................... 176
xvi
7.4 Swelling.......................................................................................................................................... 178
7.5 Sedimentation rate and time delay ............................................................................................... 180
7.6 Apparent viscosity and sedimentation rate .................................................................................. 181
7.7 Slurry stability and time-dependency ........................................................................................... 184
Chapter 8 Conclusion and further work ..................................................................................................... 185
8.1 Summary of equipment design ..................................................................................................... 185
8.2 Summary of major research findings ............................................................................................ 186
8.3 Future work ................................................................................................................................... 189
References ..................................................................................................................................................... 192
Appendixes Overview .................................................................................................................................... 197
Appendix A Sequence list .......................................................................................................................... 198
Appendix B Tables of rheology and sedimentation data .......................................................................... 213
A.1: Advanced argillic............................................................................................................................ 213
A.2: Argillic ............................................................................................................................................ 214
A.3: Upper epithermal samples ............................................................................................................ 220
A.4: Silica Breccia .................................................................................................................................. 229
A.5: Lower Epithermal .......................................................................................................................... 234
A.6: Outer biotite .................................................................................................................................. 240
A.7: Inner biotite ................................................................................................................................... 252
Appendix C Statistical results .................................................................................................................... 258
A.8: Regression analysis of apparent viscosity vs pH, concentration, passing 9.86, passing 38, passing
53, amorphous silicate, kaolinite and clay content ................................................................................... 258
A.9: Regression analysis of the logarithm of apparent viscosity vs pH, concentration, passing 9.86,
passing 38, passing 53, amorphous silicate, kaolinite and clay content ................................................... 262
Appendix D Sizing data .............................................................................................................................. 267
Appendix E Data processing example including raw data ........................................................................ 275
A.10: Estelle method ........................................................................................................................... 281
A.11: Casson approximation ............................................................................................................... 294
A.12: Summary .................................................................................................................................... 302
A.13: Raw data for SEQ233 ................................................................................................................. 304
Appendix F Sedimentation curves ............................................................................................................ 323
Appendix G Flow curves ............................................................................................................................ 324
xvii
Table of figures
Figure 2-1: Schematic approximation of variation in Lihir mineralogy as
taken from Gardner (2016). The new domains currently used by Newcrest
may be seen on the lower horizontal axis whilst the older domain
characterisation is on the upper horizontal axis. The y -axis is the weight
p e r c e n t a g e ( w t % ) ................................................................................................................................... 6
Figure 2-2: Pyrite value classification diagram. No chemical or textural
s p a t i a l d i s t i n c t i o n e s t a b l i s h e d . T a k e n f r o m ( G a r d n e r , 2 0 1 6 ) . ................................ 8
Figure 2-3: Flow curves for time-independent slurries. Taken from
( K l i m p e l , 1 9 8 4 ) . ..................................................................................................................................... 9
Figure 2-4: The loop and step tests for determining time -dependent
r h e o l o g i c a l b e h a v i o u r . T a k e n f r o m ( T a d r o s , 2 0 1 0 ) ...................................................... 13
Figure 2-5 : Mendoza (2017) plot of relative viscosity; the predicted
values are represented by the lines and the experimental values by the
s y m b o l s ....................................................................................................................................................... 15
F i g u r e 2 - 6 : T h i x o t r o p y o f i r o n m i n e t a i l i n g s l u r r y ( J e o n g e t a l . , 2 0 1 5 ) ....... 17
F i g u r e 2 - 7 : S h i ( 1 9 9 5 ) s i m u l a t e d f l o w c u r v e s o f H e l l y e r g r i n d i n g s l u r r y . .. 19
Figure 2-8: Flow curves of discharge slurries of Emu Gold ball mills. As
can be seen (a) is dilatant with low yiel d stress whilst (b) is Pseudoplastic
w i t h h i g h y i e l d s t r e s s . T a k e n f r o m S h i ( 2 0 0 2 ) . .............................................................. 21
Figure 2-9: Effect of low molecular weight polyacrylic acid on the shear
yield stress-pH behaviour for a zirconia suspension at 57% (wt%) (Boger,
2 0 0 9 ) ............................................................................................................................................................ 23
Figure 2-10: Zirconia suspension with additive of polyacrylic at 57% with
u n a d j u s t e d p H a n d 8 8 . 7 % w i t h a d j u s t e d p H ( B o g e r , 2 0 0 9 ) . ................................... 24
F i g u r e 2 - 1 1 : P a r a l l e l p l a t e ( A n o n , 2 0 1 6 ) ............................................................................ 26
F i g u r e 2 - 1 2 : C o n e a n d p l a t e ( A n o n , 2 0 1 6 ) ......................................................................... 27
Figure 2-13:DIN 53019 / ISO 3219 coaxial cylinder geometry (Anon, 2016)
......................................................................................................................................................................... 28
F i g u r e 2 - 1 4 : V a n e G e o m e t r i e s ( A n o n , 2 0 1 6 ) .................................................................... 30
F i g u r e 2 - 1 5 : C a p i l l a r y r h e o m e t e r s c h e m a t i c ( A k r o y d , 2 0 0 4 ) .................................. 32
F i g u r e 2 - 1 6 : C o a x i a l c y l i n d e r i m a g e t a k e n f r o m I S O ( 1 9 9 5 ) ................................ 35
Figure 2-17: A schematic showing particle settling zones formed in a
c o n v e n t i o n a l c u p a n d b o b r h e o m e t e r ( K l e i n e t a l . , 1 9 9 5 ) ........................................ 38
Figure 2-18: Rheological measurements of a fly ash slurry using a standard
b o b a n d c u p g e o m e t r y ( A k r o y d , 2 0 0 4 ) . ................................................................................. 39
Figure 2-19: The Klein et al (1995) solution to measuring the rheology of
s e t t l i n g s l u r r i e s . .................................................................................................................................. 40
Figure 2-20: A schematic of a testing rig built on the same principles as
t h e D e b e x r h e o m e t e r ( S h i , 2 0 1 6 ) .............................................................................................. 41
F i g u r e 2 - 2 1 : A p r e v i o u s v i s c o s i t y t e s t i n g r i g u s e d b y S h i ( 2 0 1 6 ) ..................... 41
Figure 2-22: A schematic used as the base of the design of the Akroyd
( 2 0 0 4 ) h e l i c a l f l o w r h e o m e t e r .................................................................................................... 43
xviii
Figure 2-23 Error in shear rate between flow rheometer data analysed using
helical flow and Couette flow theory for a 1 wt% CMC solution (taken fr om
A k r o y d ( 2 0 0 4 ) ) ...................................................................................................................................... 44
Figure 2-24: Apparent viscosity of quartz and amorphous silica
suspensions at 100s-1 as a function of solids concentration (Chen et al.,
2 0 1 7 ) ............................................................................................................................................................ 45
Figure 2-25: Diagram showing a clay mineral layer (A); a particle (B); an
aggregate (c); and an assembly of aggregates (D) (taken from Bergaya and
L a g a l y ( 2 0 1 3 ) ) . ..................................................................................................................................... 47
Figure 2-26:edge/face and face/face contacts that link the particles to a
h o u s e o f c a r d s ( a ) o r a b a n d - l i k e ( b ) s t r u c t u r e ( L a g a l y , 1 9 8 9 ) . ......................... 48
Figure 2-27:A classification of phyllosilicate clay minerals (Ndlovu, 2014)
......................................................................................................................................................................... 49
Figure 2-28: Viscosity ternary plots in volumetric fractions at ph7. Note,
that the plots are a different solid concentration; the solids are at 15% and
30% volumetric concentration for the graphs on the left and right
r e s p e c t i v e l y . ........................................................................................................................................... 52
Figure 2-29: Time dependent viscosity of nuclear waste slurries showing
thixotropic, anti-thixotropic and time independent behaviour under
d i f f e r e n t s h e a r r a t e s ( S m i t h e t a l . , 2 0 0 0 ) . ......................................................................... 54
Figure 2-30: Viscosity against time for 1997 ppm partially hydrolysed
p o l y a c r y l a m i d e i n w a t e r / g l y c e r i n e ( B u i t e n h u i s a n d S p r i n g e r , 2 0 0 3 ) ............... 55
Figure 2-31: Flow curve for a thixotropic yield -stress fluid experiencing
bifurcation contrasted against an ideal yield stress fluid (Coussot et al.,
2 0 0 2 ) . .......................................................................................................................................................... 56
Figure 2-32: shear rate versus time for a 4.5% bentonite suspension after
20s of rest following an intense 60s of 26 Pa pre -shear using a vane in cup
g e o m e t r y . .................................................................................................................................................. 57
Figure 2-33: Comparison of different sizing techniques (Napier -Munn et
a l . , 1 9 9 6 ) ................................................................................................................................................... 59
F i g u r e 2 - 3 4 : T h e f l o t a t i o n p r o c e s s . T a k e n f r o m L i ( 2 0 1 6 ) ..................................... 63
Figure 2-35: Ferrosilicon sedimentation rate at a variety of densities (Shi,
2 0 1 6 ) ............................................................................................................................................................ 66
F i g u r e 3 - 1 : T h e f i r s t i t e r a t i o n o f t h e n e w t e s t i n g r i g ................................................ 76
F i g u r e 3 - 2 : T h e c u r r e n t t e s t i n g r i g ......................................................................................... 77
Figure 3-3: Cumulative percent passing stated size in various processes of
L i h i r ' s g r i n d i n g c i r c u i t ( S h i , 2 0 1 7 a ) ...................................................................................... 78
Figure 3-4:HAAKE rheometer running a calibration test. Immersion tube
w a s u s e d f o r a n a r r o w g a p g e o m e t r y . ..................................................................................... 79
Figure 4-1: Particle size distribution curves of NTS050 after grinding for a
d i f f e r e n t i a l a m o u n t o f t i m e . ......................................................................................................... 94
F i g u r e 4 - 2 : S m a l l b o b a n d l a r g e b o b u s e d i n R M 1 8 0 r h e o m e t e r . ........................ 101
F i g u r e 4 - 3 : A n g u l a r s p e e d v s t o r q u e f o r S E Q 2 3 3 . ...................................................... 107
Figure 4-4: Comparison between the Casson approximation and the Estelle
method using a 6th order polynomial fit of Ω(M). The data was from
S E Q 2 3 3 . ................................................................................................................................................... 112
xix
Figure 4-5: Fitting a Casson model to the Estelle approximation of SEQ233
u s i n g a 6 t h o r d e r p o l y n o m i a l f i t o f Ω ( M ) . ......................................................................... 113
Figure 5-1: Calibration of the RM180 using 486 mPa.s Newtonian oil (prior
t o c o r r e c t i o n f a c t o r s ) . .................................................................................................................... 119
Figure 5-2: NTS100 53% (v/v) Comparison of raw results for the larg e bob
....................................................................................................................................................................... 120
F i g u r e 5 - 3 : R e p e a t o f N T S 0 5 4 u s i n g f r e s h s a m p l e s ................................................... 121
Figure 5-4: Comparison of the Malvern MasterSizer to wet sieving bel ow
3 8 μ m . ....................................................................................................................................................... 122
F i g u r e 5 - 5 : O u t e r b i o t i t e c u m u l a t i v e p a s s i n g s i z e ....................................................... 123
F i g u r e 5 - 6 : N T S 1 1 0 a t 4 6 % ( v / v ) a f t e r 5 m i n u t e s o f g r i n d i n g . .......................... 132
F i g u r e 6 - 1 : N T S 0 5 4 a t 5 7 % s o l i d v o l u m e t r i c c o n c e n t r a t i o n . .............................. 135
F i g u r e 6 - 2 : N T S 1 0 8 a t 4 6 % s o l i d v o l u m e t r i c c o n c e n t r a t i o n . .............................. 135
Figure 6-3: Raw data of the hysteresis loop conducted on NTS108
( S e q u e n c e 2 3 1 w i t h a s o l i d v o l u m e t r i c c o n c e n t r a t i o n o f 4 6 % ) .......................... 136
Figure 6-4: The raw data of NTS108 (SEQ231) at a solid concentration of
4 6 % ( v / v ) b e i n g s h e a r e d b y t h e s m a l l b o b a t 9 9 9 r o t a t i o n s p e r m i n u t e . ...... 137
Figure 6-5:The change in the “yield stress” of an argillic sa mple (NTS109
at 53% concentration; SEQ400) prior to the first, second, and third loop
t e s t . ............................................................................................................................................................. 138
Figure 6-6: The increase in yield stress over time of NTS111 (after
g r i n d i n g a n d a t 4 6 % s o l i d v o l u m e t r i c c o n c e n t r a t i o n ) . ............................................. 140
Figure 6-7: Apparent viscosity (at a shear rate of 50 s -1) in relation to
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r a r g i l l i c s a m p l e N T S 1 0 8 . ............................... 144
Figure 6-8: Flow curves of NTS008 (upper epithermal) at 53% and 46%
s o l i d s c o n c e n t r a t i o n ( v / v ) . .......................................................................................................... 145
Figure 6-9: The grinding effect on the relationship of slurry viscosity with
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r o u t e r b i o t i t e ( N T S 0 5 4 ) .................................. 148
Figure 6-10: The grinding effect on the relationship of slurry viscosity
w i t h s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r i n n e r b i o t i t e ( N T S 0 9 7 ) ...................... 149
w i t h s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r a r g i l l i c ( N T S 1 0 8 ) .................................. 150
w i t h s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r l o w e r e p i t h e r m a l N T S 1 1 0 ................ 151
w i t h s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r u p p e r e p i t h e r m a l ( N T S 1 1 1 ) ............ 152
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r a d v a n c e d a r g i l l i c s a m p l e N T S 0 3 7 . ........ 153
Figure 6-15: Apparent viscosity (at a shear rate of 50 s-1) in relation to
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r a r g i l l i c s a m p l e s .................................................. 154
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r u p p e r e p i t h e r m a l s a m p l e s ............................ 155
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r s i l i c a b r e c c i a s a m p l e s .................................... 156
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r u p p e r l o w e r e p i t h e r m a l ................................. 156
xx
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r i n n e r b i o t i t e s a m p l e s . .................................... 157
s o l i d v o l u m e t r i c c o n c e n t r a t i o n f o r o u t e r b i o t i t e s a m p l e s ...................................... 158
Figure 6-21: Slurry viscosities of various samples at a constant 46% solid
v o l u m e t r i c c o n c e n t r a t i o n . ............................................................................................................ 159
F i g u r e 6 - 2 2 : S l u r r y v i s c o s i t y i n r e l a t i o n t o p H f o r a l l t e s t e d s a m p l e s ........ 162
Figure 6-23: Minitab graph of the multilinear regression results against the
f i t t e d v a l u e s .......................................................................................................................................... 164
Figure 6-24: Minitab results of the fitted multilinear results against the
m e a s u r e d a p p a r e n t v i s c o s i t y a t 5 0 s - 1 . ................................................................................ 164
Figure 6-25: The predicted vs measured natural logarithms of the apparent
v i s c o s i t y a t 5 0 s - 1 . .............................................................................................................................. 165
Figure 6-26: Minitab’s four in one chart showing the residuals relationship
w i t h t h e l o g a r i t h m o f t h e a p p a r e n t v i s c o s i t y . ................................................................ 166
Figure 6-27: Surface plot of the logarithm of apparent viscosity at 50s -1 vs
c o n c e n t r a t i o n a n d p H . .................................................................................................................... 168
Figure 6-28: Surface plot of the logarithm of vi scosity at 50 s-1 vs the
p a r t i c l e s i z e s s m a l l e r t h a n 9 . 8 6 µ m a n d t h e a m o r p h o u s s i l i c a c o n t e n t . ....... 169
Figure 6-29: Surface plot of the logarithm of viscosity at 50 s-1 vs the clay
c o n t e n t a n d t h e p H ............................................................................................................................ 169
Figure 6-30: Surface plot of the logarithm of viscosity at 50 s -1 vs the
p a r t i c l e s i z e s s m a l l e r t h a n 9 . 8 6 µ m a n d t h e p H ............................................................ 170
Figure 6-31: Surface plot of the logarithm of viscosity at 50 s-1 vs the
K a o l i n i t e c o n t e n t a n d t h e p H ..................................................................................................... 171
Figure 7-1: Average sedimentation rate in relation to solid volumetric
c o n c e n t r a t i o n f o r N T S 0 0 8 ............................................................................................................ 175
Figure 7-2:Comparison of highest concentrations when sedimentation above
2 m L / h o u r w a s o b s e r v e d ............................................................................................................ 176
Figure 7-3: Sedimentation rate in relation to solid volumetric
c o n c e n t r a t i o n f o r a r g i l l i c o r e s a m p l e s . .............................................................................. 177
Figure 7-4:An example of a swelling sample, during sedimentation
m e a s u r e m e n t o v e r 3 . 4 h o u r s ....................................................................................................... 178
Figure 7-5:The volume of the mud-line over time in the sedimentation
m e a s u r e m e n t f o r N T S 0 0 8 a t 5 5 % s o l i d v o l u m e t r i c c o n c e n t r a t i o n . ................... 180
Figure 7-6: Time delay in the sedimentation process for NTS008 sample
after one, twelve and twenty-four minutes after starting the sedimentation
e x p e r i m e n t . ............................................................................................................................................ 181
Figure 7-7: Slurry viscosity in relation to sedimentation rate for the entire
o r e s a m p l e s ............................................................................................................................................ 182
Figure 7-8: Linear regression on the data of slurry viscosity in relation to
s e d i m e n t a t i o n r a t e f o r t h e e n t i r e o r e s a m p l e s ................................................................ 183
F i g u r e D - 1 : A d v a n c e d a r g i l l i c c u m u l a t i v e p a s s i n g s i z e ........................................... 271
F i g u r e D - 2 : A r g i l l i c c u m u l a t i v e p a s s i n g s i z e ................................................................ 272
F i g u r e D - 3 : U p p e r e p i t h e r m a l c u m u l a t i v e p a s s i n g s i z e ............................................ 272
F i g u r e D - 4 : S i l i c a b r e c c i a c u m u l a t i v e p a s s i n g s i z e .................................................... 273
xxi
F i g u r e D - 5 : U p p e r e p i t h e r m a l c u m u l a t i v e p a s s i n g s i z e ........................................... 273
F i g u r e D - 6 : I n n e r b i o t i t e c u m u l a t i v e p a s s i n g s i z e ....................................................... 275
Figure E-1: A comparison of the predicted vs measured angula r velocity
w h e r e 𝛺𝑃 i s t h e p r e d i c t e d a n g u l a r v e l o c i t y a n d 𝛺𝐸 i s t h e e x p e r i m e n t a l
a n g u l a r v e l o c i t y . ................................................................................................................................ 301
Figure E-2: A comparison of the different rheograms produced by the
d i f f e r e n t m e t h o d s . ............................................................................................................................. 303
xxii
Table of tables
T a b l e 2 - 1 : T h e a v e r a g e g o l d a n d s u l p h u r c o n t e n t o f t h e L i h i r o r e ' t y p e s ' ..... 7
T a b l e 3 - 1 : T a b l e o f l i q u i d l e v e l t o t a n k d i a m e t e r r a t i o s ( J a m e s , 2 0 1 5 ) ....... 81
T a b l e 4 - 1 : M a s s t a r g e t s w h e n s p l i t t i n g ................................................................................ 92
T a b l e 4 - 2 : B o b g e o m e t r i e s .......................................................................................................... 102
Table 4-3: Calculations of shear rate using Estelle et al method and fitting
t h e a n g u l a r s p e e d t o t o r q u e t h r o u g h t h e u s e o f a q u a d r a t i c . ................................ 108
Table 4-4: Table of coefficients of Ω(M) for different orders of polynomial
f i t s f o r S E Q 2 3 3 . ................................................................................................................................. 110
Table 4-5: Comparisons of the derivatives of angular speed with respect to
t o r q u e ( dΩdM) b e t w e e n t h e d i f f e r e n t o r d e r s o f a n g u l a r s p e e d ( Ω ( M ) ) f o r
S E Q 2 3 3 . ................................................................................................................................................... 110
Table 4-6: Comparisons of the shear rates estimated with different orders
of Ω(M) for SEQ233. Some data is excluded as it was found to be in the
T a y l o r r e g i o n . ...................................................................................................................................... 111
Table 4-7: Comparison of the parameters of the Casson curve. The directly
fitted curve refers to the parameters found using the method as described
in section 4.9.4. The fitted to Estelle column refers to the fit as seen in
F i g u r e 4 - 5 . ............................................................................................................................................. 113
Table 5-1: A table showing the correction factors used for calculating
s h e a r s t r e s s ............................................................................................................................................ 119
Table 5-2: Size distribution on as provided material sieved to less than 1
m m ............................................................................................................................................................... 125
Table 5-3: The size distribution of the material after being sieved to less
t h a n 1 m m a n d u n d e r g o i n g a f u r t h e r 5 m i n u t e s o f g r i n d i n g ................................. 127
T a b l e 5 - 4 : N T S 0 4 7 r e s u l t s .......................................................................................................... 128
Table 5-5: Rheology and sedimentation charact eristic parameters measured
a t t h e h i g h e s t s o l i d c o n c e n t r a t i o n a c h i e v e d . .................................................................. 130
Table 5-6: Rheology and sedimentation characteristic parameters measured
a t t h e h i g h e s t s o l i d c o n c e n t r a t i o n a c h i e v e d . .................................................................. 131
T a b l e 6 - 1 : P a r t i c l e s i z e i n f o r m a t i o n f o r t h e g r o u n d s a m p l e s .............................. 147
T a b l e 6 - 2 : C l a y i n f o r m a t i o n ........................................................................................................ 161
Table 7-1: Summary of swelling behaviour amongst the different domains
....................................................................................................................................................................... 179
T a b l e A - 1 : S e q u e n c e l i s t .............................................................................................................. 198
T a b l e B - 1 : N T S 0 3 7 r e s u l t s .......................................................................................................... 214
T a b l e B - 5 : N T S 1 0 8 G r i n d f o r 5 m i n u t e s ............................................................................ 218
T a b l e B - 6 : N T S 1 0 9 ........................................................................................................................... 219
T a b l e B - 7 : N T S 2 3 8 ........................................................................................................................... 220
xxiii
T a b l e B - 9 : N T S 1 1 1 a s p r o v i d e d r e s u l t s .............................................................................. 224
T a b l e B - 1 0 : N T S 1 1 1 a f t e r g r i n d i n g f o r 5 m i n u t e s ...................................................... 225
Table B-11: Solids concentration and swelling details for NTS111 after
g r i n d i n g f o r 5 m i n u t e s ................................................................................................................... 226
T a b l e B - 1 2 : N T S 2 3 9 r e s u l t s ....................................................................................................... 227
T a b l e B - 1 3 : S o l i d s c o n c e n t r a t i o n a n d s w e l l i n g d e t a i l s f o r N T S 2 3 9 ............... 228
T a b l e B - 1 4 : N T S 0 1 0 r e s u l t s ......................................................................................................... 230
T a b l e B - 2 1 : N T S 1 1 0 a f t e r 5 m i n u t e s g r i n d i n g ............................................................... 237
f i v e m i n u t e s o f g r i n d i n g ............................................................................................................... 238
T a b l e B - 2 5 : N T S 0 4 4 g r o u n d f o r 5 m i n u t e s r e s u l t s ..................................................... 242
f i v e m i n u t e s o f g r i n d i n g ............................................................................................................... 243
T a b l e B - 3 1 : N T S 0 5 4 a f t e r g r i n d i n g f o r 5 m i n u t e s r e s u l t s ..................................... 249
Table B-32: Solids concentration and swelling details for NTS054 after 5
m i n u t e s g r i n d i n g ................................................................................................................................ 250
T a b l e B - 3 5 : N T S S 0 5 5 r e s u l t s .................................................................................................... 254
T a b l e B - 3 7 : N T S 0 9 7 a f t e r g r i n d i n g f o r 5 m i n u t e s ...................................................... 256
Table B-38: Solids concentration and swelling details for NTS097 after 5
m i n u t e s g r i n d i n g ................................................................................................................................ 257
Table E-1: Measuring system characteristics of SEQ233. SEQ233 used the
large bob and the RM180 rheometer in the wide gap configuration
d e s c r i b e d e l s e w h e r e i n t h i s t h e s i s . ........................................................................................ 276
T a b l e E - 2 : M e a s u r i n g s y s t e m p r o p e r t i e s ........................................................................... 276
Table E-3 Averaged raw data and the calculat ed shear stress and angular
v e l o c i t y r e s u l t s . ................................................................................................................................. 277
Table E-4: Data used for calculating shear rate. From Table E -3, the last 5
rpm rate is treated as the equilibrium yield stress. At 500 rpm, the torque
exceeded the manufacturers specification, so the data point is omitted from
f u r t h e r a n a l y s i s . ................................................................................................................................. 279
xxiv
Table E-5 : The components required for the Estelle approximation.
H i g h l i g h t e d i n r e d i s t h e p h y s i c a l l y i m p o s s i b l e r e s u l t o f a n e g a t i v e 𝑑𝛺𝑑𝑀.
....................................................................................................................................................................... 284
Table E-6: SEQ233 angular speed vs torque fits. As can be seen they are
split into three parts labelled a, b, and c referring to three separate parts
o f t h e c u r v e s h o w n i n F i g u r e 4 - 3 . .......................................................................................... 285
T a b l e E - 7 : T h e s h e a r r a t e c a l c u l a t i o n u s i n g t h e E s t e l l e m e t h o d . ..................... 286
Table E-8: Calculating the Taylor number with the Estell e method. The red
l i n e s i n d i c a t e c e l l s w i t h T a y l o r v o r t i c e s ........................................................................... 289
T a b l e E - 9 : F i n a l r e s u l t f r o m t h e E s t e l l e c a l c u l a t i o n ................................................ 292
T a b l e E - 1 0 : T h e p a r a m e t e r s o f t h e C a i s s o n f l u i d m o d e l ......................................... 296
T a b l e E - 1 1 : S t e p s 2 a n d 3 o f t h e C a s s o n a n a l y s i s ...................................................... 296
Table E-12: The Taylor number calculation technique. Note, for this
example, the shearing is always partial due to the high yield stress. Under
t h e f u l l y s h e a r e d c o n d i t i o n t h e d i s t a n c e w o u l d b e 0 . 0 1 m .................................... 298
T a b l e E - 1 3 : T h e p a r a m e t e r s o f t h e C a s s o n f l u i d m o d e l ........................................... 299
T a b l e E - 1 4 : S t e p s 2 a n d 3 o f t h e C a s s o n a n a l y s i s ...................................................... 299
T a b l e E - 1 5 : T a y l o r d a t a ................................................................................................................ 300
T a b l e E - 1 6 : F i n a l r e s u l t s o f t h e C a s s o n m e t h o d .......................................................... 301
List of abbreviations
F = correction factor
K = Herschel − Bulkey consistency index
K B = bob constant parameter
K V = vane constant parameter
L = length of the bob or vane
M = torque
r = radius
r1 = rb = radius of the bob
r2 = rc or ry
rc = radius of the cup
ry = radius of the sheared region
xxv
R = linear correlation coefficient
R2 = coefficient of determination
t = t statistic for the two sided Student t − distribution
Ta = Taylor number
α = ratio of the radius of cup to the radius of bob
γ̇ = shear rate
η = viscosity
ηapparent = apparent viscosity
η0 = carrier fluid viscosity
ηr = relative fluid viscosity
ηB = Binghman viscosity
η∞ = infinite viscosity
τ = shear stress
τ0 = yield stress
τ1 = shear stress at the surface of the bob
ϕ = solids concentration
Ω = angular speed
xxvi
Chapter 1
Introduction
1.1 Background
Comminution operations have a high energy intensity. It has been
reported that comminution processes consume approximately 1.8% of
the world’s generated electrical energy; the comminution of gold and
copper ores alone consume about 0.2% (CEEC, 2016). Improving the
efficiency of mills clearly has the potential to reduce the energy cost,
operating cost and environmental impact of a mine. The rheolo gical
characteristics of wet ground ores are known to affect the energy
efficiency, throughput, and total recovery of minerals; it is believed
these effects are significant yet they are still not well understood
(Shi, 1995).
This project is a part of a larger project investigating the effect

of rheology on Lihir’s grinding circuits. From previous research and
experience it was hypothesised that the rheology of Lihir’s slurry
affects the SAG mill performance and that the rheology is dependent
on the feed ore.
The goal of the sponsors, Newcrest, is to increase the efficiency

and throughput of their Lihir operation. Lihir has mult iple rheological
problems with their grinding circuit: wear and tear caused by highly
viscous slurry, blocked mill greats, and the formation of a slurry pool
within their semi-autogenous grinding mill (SAG mill). At the time of
writing, Newcrest have successfully increased the throughput of the
downstream mineral processes and the upstream processes such as
comminution and flotation are becoming bottlenecks.
1
Previous research by Klimpel (1984) has found four major factors
that affect slurry rheology:
1. Slurry volume concentration
2. Fines content
3. Slurry temperature
4. Chemical environment
The significance of improving comminution is only going to

grow. The mineral industry are exhausting high quality sources of ore
and are beginning to mine lower quality grades (Cruz et al., 2013;
Schubert, 2008). The source of Lihir’s rheology problem is linked to
its complicated geology. Lihir ore is known to contain high clay
content; as clay content has been known to increase the yield stress,
increase apparent viscosity and introduce time dependency it is
suspected of being the cause of the rheology problems. Most of the
gold is located within pyrite and pyrite is associated with high clay
content (Gardner, 2016). The clays of interest in this study include
kaolinite, illite/smectite, and montmorillonite. A more detailed
discussion of the geology of Lihir is provided in section 2.1.
1.2 Aims and objectives

This Masters of Philosophy forms a part of a broader project
“Rheology study of Lihir grinding circuits”. The aims and objectives
of this thesis are:
1. to characterise the rheology and the sedimentatio n rate of

Lihir ore slurry;
2. to understand how the rheology and sedimentation rate v ary

with solids concentration, particle size and size
distribution, and the ore type;
3. and to investigate the correlations bet ween the rheology

and sedimentation rate.
2
1.3 Statement of Sustainability
This research was undertaken whilst studying at the Julius
Kruttschnitt Mineral Research Centre (JKMRC) which is a part of the
University of Queensland’s Sustainable Minerals Institute ( SMI). As
such, the research undertaken was done under the broad goal of
sustainability.
This research lays the foundations for a rh eological model of the

Lihir ore slurry. The plan is for a future student to further this work
and develop a model and to then use this model to optimise Lihir plant
operations (in particular, the grinding circuit). Optimising the
grinding circuit will improve throughput, decrease waste and water
consumption, hence improving the sustainability of Lihir both
economically and environmentally.
This thesis will demonstrate how to characterise rheology and

stability of coarse unstable mineral slurries with high clay contents.
Others will be able to emulate this work in order to improve their
mine’s grinding circuits. If other mine sit es can improve their
grinding circuits everyone will benefit as mining will become cheaper
and less environmentally destructive; as stated earlier comminution is
estimated to account for half of the energy used in mining.
Furthermore, it will lead to impro ved processing of high clay content
minerals thus making more potential sites feasible.
All of these benefits will not involve any new technology, rather,
they stem from improved management. The research allows a better
scientific understanding of ore and how to process it. As such, these
improvements should be rapidly implementable.
1.4 Thesis structure overview

The thesis contains nine chapters. Chapter 1 and Chapter 2 give
an overview of the background to this work and the surrounding
literature.
3
Chapter 3 explains the commissioning of the rheology
experimental equipment. This includes discussion explaining
rheometer selection, selecting an appropriating coaxial gap, and the
design considerations behind commissioning the “testing rig”
component of the helical flow rheometer. Of the preceding tasks, the
commissioning of the testing rig component of the flow rheometer was
the most important contribution in this chapter. The testing rig is the
component of the flow rheometer which circulates and mixes slurry
preventing it from settling. It consists of several pieces of equipment
that had to be either designed and constructed at the JKMRC pilot
plant or altered to handle slurry flow such as a sump, a pump, a
measuring chamber, a mixing impeller and the associate d piping.
There was a catastrophic failure in the first choice an air-bearing

HAAKE Viscotester IQ rheometer. It could not be repla ced in time for
the completion of the experimental program and an alternative
rheometer, the Mettler RM180 was selected. Expe rimental work was
re-designed with a traditional coaxial cylinder geometry due to a
combination of sample and equipment limitations .
Chapter 4 is an overview of the experimental program of the

HAAKE flow rheometer and the RM180 coaxial cylinder tests.
Included is a discussion of the data processing techniques.
Chapter 5 is an overview of the results whilst chapter 6 and 7 are

the analytic selections of the thesis. Chapter 6 focuses on the Lihir
ore slurries rheology, whilst chapter 7 on the sedimentation, swelling,
and the link between sedimentat ion and rheology.
Chapter 8 concludes the thesis and recommends future work.
4
Chapter 2
Literature Review
As this project is fundamentally about understanding the
rheology and the stability of Lihir ores, and how the effe cts of the
underlying properties interact, the literature review covers:
• the geology of Lihir;
• a basic over-view of rheology;
• the properties of clay and how they interact with rheology;
• rheology within the mineral processing industry;
• the selection of an appropriate rheology apparatus.
2.1 Lihir ore

Lihir ore is a mixture of Argillic, epithermal and porphyry
bodies. It contains a mixture of clay, pyrite, K-feldspar, quartz,
Anhydrite, carbonate, biotite, Na -feldspar and chlorite (Gardner,
2016). Newcrest characterises clay into three groups:
1. Kaolinite
2. Illite/Smectite
3. Montmorillonite
Newcrest divides the ore samples into nine domains as seen in

Figure 2-1. Seven of these domains are regularly processed and their
rheological properties are of great interest. The upper argillic domain
rarely mineralises and therefore isn’t processed. They don’t proc ess
distal chlorite.
5
Figure 2-1: Schematic approximation of variation in Lihir
mineralogy as taken from Gardner (2016). The new domains currently
used by Newcrest may be seen on the lower horizontal axis whilst the
older domain characterisation is on the upper horizontal axis. The y -
axis is the weight percentage (wt%)
Historically, Newcrest had a three-domain system; these domains

were Argillic, epithermal, and po rphyry. The highest gold grade is in
the Epithermal grade ore as can be seen in Table 2-1.
6
Table 2-1: The average gold and sulphur content of the Lihir ore
'types'
Gold grade and sulphur content across Lihir zones
Zone Gold grade Total Sulphur
Argillic 2.1 g/t 8.0%
Epithermal 2.4 g/t 6.1%
Anhydrite 2.0 g/t 3.2 %
Over 90% of gold is located within pyrite. Newcrest has

characterised pyrite through a ‘pyrite value’ classification as can be
seen in Figure 2-2. High value pyrite is present in similar amounts in
the key material types being processed.
7
Figure 2-2: Pyrite value classification diagram. No chemical or
textural spatial distinction established. Taken from (Gardner, 2016).
2.2 Background on rheology theory

Within continuum dynamics when a force is applied to matter it
will either undergo a reversible deformation known as elasticity or an
irreversible deformation known as flow. Elasticity is associated with
solids whilst flow is associated with fluids, howeve r, this relationship
doesn’t always hold; rheology is the subsection of continuum
dynamics studying the behaviour of liquid -like solids and solid-like
liquids. This project is concerned with the rheological behaviour of
Lihir ore slurries rather than with the dry rocks and thus will ap proach
the subject from the perspective of solid-like fluids. As explained by
Shi (1995) the elastic effects of mineral slurries are not significant in
8
most processes and it is only necessary to consider the viscous
effects.
Conveniently, rheology may be broken into time independent and

dependent effects.
2.2.1 Time independent behaviour

Cheng et al. (2014) defined viscosity as a rheological property of
liquids that presents itself when the velocity varies between different
layers of liquid. The simplest and most classical way of expressing
viscous systems is the infinite parallel plane model. A tangential force
(shear stress, denoted by τ) is applied thus resulting in the planes
sliding over each other; the change in velocity be tween plates is
k n o w n a s t h e s h e a r r a t e ( 𝛾̇ ) a n d v i s c o s i t y ( η ) i s t h e r e s i s t a n c e t o t h e
deformation from the force. The usual co -efficient of viscosity is
apparent viscosity which is defined as shear stress divided by shear
rate as seen in equation 2.1
𝜏
𝜂𝑎𝑝𝑝𝑎𝑟𝑒𝑛𝑡 = 𝛾̇ 2-1
Figure 2-3: Flow curves for time-independent slurries. Taken

from (Klimpel, 1984).
9
As is commonly known there are several types of time
independent flow as shown in Figure 2-3. Under Newtonian flow, the
fluid has a constant viscosity and flow initiates under even very small
forces. A fluid which undergoes shear ind uced thickening (i.e. grows
more viscous the larger the force tha t is applied) is called dilatant
whilst shear induced thinni ng (becomes less viscous when a greater
force is applied) is called pseudoplastic. Examples of these
phenomenon include (but by no means are limited to): water is a
Newtonian fluid, concentrated s and suspensions are dilatant (Barnes,
1989), and polymers are typically pseudoplastic (Rapp, 2017).
Some fluids require a certain shear stress before flow occurs;

this stress is known as the yield stress. Below the yield stress the flow
regime is highly complicated as it consists of a combination of elastic
and viscous flows (Barnes, 1999). Historically it was believed that
yield stress was a physical property in which no flow would occur
until sufficient stress was applied, however, with superior rheometers
it has become clear that many substance s creep given sufficient time.
Barnes (1999) proposes that “everything flows” and yield st ress is just
an extremely viscous Newtonian plateau. Despite this, yield stress is
clearly a highly useful property for engineering purposes as it
describes numerous practical features of non -Newtonian fluids.
It is believed that pseudoplasticity is cause d by a breakdown in

structure with increasing shear rates. Dilatancy is caused by volume
changes; suspensions which have reached th eir maximum packing
density undergoing shear must increase their volume to undergo shear
(Hallbom, 2008)
2.2.2 Time dependent behaviour

As alluded to before, rheological behaviour of fluids is typically
separated into time dependent or time independent components. It is
generally accepted that mineral slurry rheology is time independent
(Shi, 1995); however, this may not be assumed for Lihir or e slurry as
10
it contains clay minerals which often display thixotropic effects (see
section 2.9). Tadros (2010) described three types of time dependent
flow:
1. Thixotropic
2. Anti-thixotropic
3. Rheopectic
Thixotropy refers to time-dependent reversible decrease in

viscosity. Certain fluids, such as some paints, undergo a decrease in
viscosity as the system is sheared and return to their previous
viscosity upon shearing ceasing. Anti -thixotropy (sometimes known as
negative thixotropy) is time dependent shear thickening. Certain
fluids, such as corn starch suspended in milk, undergo an increase in
viscosity over time as it is sheared. Like thixotropic systems, anti -
thixotropic systems are reversible and as the shearing ce ases the
viscosity decreases over time to its original value (Tadros, 2010).
Anti-thixotropy is often confused with rheopexy; it is not

difficult to find rheopexy as being incorrectly described as time
dependent shear thickening. Rheopexy is actually the increase in
thixotropic recovery rate when an oscillation is applied to the system.
Rheopexy is unlikely to be occur and it isn’t likely to be relevant
(Tadros, 2010).
There are two common methods using a rotational rheometer for

determining time-dependent rheological behaviour: the step test and
the loop test as seen in Figure 2-4. The loop test is the more
commonly used method; it is wh ere the shear rate is increased
continuously and linearly to a maximum value before being decreased
back to rest. This test is very good at qualitatively identifying time
dependent rheological behaviour but struggles with quantifying the
effect because the relaxation process is decoupled from the strain thus
potentially preventing the material from recovering (Tadros, 2010).
11
The step change test is where the slurry is subjected to a
constant high shear rate. The application of the shear rate needs to be
sudden, ideally instantaneous. The sheer stress is recorded over a
period of time; eventually the shear stress reaches an equilibrium.
After the equilibrium is reached, the shear rate is suddenly dropped to
a lower value (not zero). The shear stress should suddenly drop and
eventually re-build over time to an equilibrium level of stress
(Tadros, 2010).
Viscoelastic effects are often inter -related with time-dependent

rheological behaviour. As such, Tadros (2010) recommends
viscoelastic measurements be used when accurately measurin g
thixotropy.
12
Figure 2-4: The loop and step tests for determining time -
dependent rheological behaviour. Taken from (Tadros, 2010)
Another method of evaluating time dependency was conducted by

Nguyen and Boger as described in the review article Boger (2009).
Nguyen and Boger observed partially irreversible time dependency in
red mud tailings. Using a vane rheometer, they demonstrated that
different sample treatments resulted in a change in the yield stress.
The longer they ran an impeller the less solid the mud b ecame; leaving
the mud undisturbed it slowly begun rebuilding its structure but never
fully recovered.
2.3 Effect of solids concentration on viscosity

The prediction of viscosity as a function of suspension prope rties
began with Einstein (1906); (Mendoza, 2017).
𝜂(𝜙) = 𝜂0 (1 + [𝜂]𝜙) 2-2
13
w h e r e η(ϕ) i s t h e v i s c o s i t y a s a f u n c t i o n o f t h e p a r t i c l e s v o l u m e
f r a c t i o n , η0 = l i q u i d p h a s e v i s c o s i t y , [η] = t h e i n t r i n s i c v i s c o s i t y ( a
single particle property that depends on factors such as shape and
p o r o s i t y ) a n d ϕ = solids volume fraction. F o r h a r d s p h e r e s , [η]= 2 . 5 , t h e
equation becomes:
𝜂(𝜙)
𝜂𝑟 = = 1 + 2.5𝜙 2-3
𝜂0
w h e r e ηr = relative viscosity. A s t h e e q u a t i o n a s s u m e s n o i n t e r a c t i o n s
between the suspended particles, it only holds when the slurry is
extremely dilute. Numerous authors have expanded the equation as a
power series:
𝜂𝑟 = 1 + 𝑘1 𝜙 + 𝑘2 𝜙 2 + 𝑘3 𝜙 3 + ⋯ 2-4
Which may be fitted against experimental data. A limitation of

this approach is that it only describes suspensions of monomodal
particles; clearly, in mining the suspended particles are mixtures of
particles, which as Mendoza (2017) states have “different shapes,
sizes, porosities, electric charge or other physical and chemical
properties that may be important in predicting the viscosity of the
suspension”. Mendoza (2017) developed a procedure for predicting the
effective viscosity of multicomponent suspensions with excellent
results when matched to experimental data as seen in Figure 2-5.
14
Figure 2-5 : Mendoza (2017) plot of relative viscosity; the
predicted values are represented by the lines and the experimental
values by the symbols
A problem with this work is that they only predict the relative
viscosity (suspension viscosity divided by liquid phase viscosity). As
stated by Shi (1995), in the mineral grinding process, knowledge of
the rheological behaviour of the slurry over a range of shear rates is
more important than knowing the viscosity value of slurry at one shear
rate. The published relationships cannot be used to predict rheological
flow curves of slurry from the measured slurry properties.
Shi found that although some mathematical models have been

developed to describe the relationship between shear s tress and shear
rate these models take into account the overall, macroscopi c behaviour
and cannot be used for the prediction of rheological behaviour from
slurry properties (Shi, 1995). Shi developed a model based upon easily
measured slurry properties; it is described in more detail in section
2.10.1.
A development in this area has b een achieved by Shewan and

Stokes (2015) who used the combination of the Farr and Groot (2009)
15
sphere packing formula and the Quemada (1977) model to analytically
predict the viscosity of a polydisperse su spension of hard spheres.
2.4 Rheology dependency

As discussed, the rheological characteristics are heavily
dependent on solids concentration and particle size distribution.
Klimpel (1984) found that in addition for mineral slurries the rheology
depends on particle size, the chemical environment, and temperature.
It is also well known that rheology is dependent on the mineralogy of
the ore (Becker et al., 2013; Burdukova, 2008; Farrokhpay et al.,
2016; Ndlovu, 2014; Ndlovu et al., 2014); it is hypothesised that clays
in particular are causing the rheology problems at Lihir. The rheology
of clay particles is discussed in detail in section 2.9.
The difficulties with mineral rheology are severa l fold as

described by Shi (2016). Firstly, slurries are often non -Newtonian.
Secondly, the rheological characteristics are o ften difficult to measure
and the tests themselves are time consuming and difficult to perform.
Thirdly, there are multiple components which affect the slurry
rheology and so models need to be developed t o predict the rheology
of untested components. Fourt hly, once the rheology of an ore is
established, it is difficult to link the rheological parameters to plant
performance.
It is common knowledge that for most mineral slurries the time

dependent rheological effects are negligible. This is not true for all
ores (as can be seen by Figure 2-6 (Jeong et al., 2015)) and it has not
been established for Lihir ores. According to Klimpel (1984) the
strongest influences on rheology are the tim e independent variables:
• solids concentration;
• particle size and particle size distribution;
• temperature;
• chemical environment.
16
Figure 2-6: Thixotropy of iron mine tailing slurry (Jeong et al.,
2015)
In addition, it has also been shown that the miner alogy and the
particle shapes also effect rheology (Mueller et al., 2010; Ndlovu,
2014).
2.4.1 Solids concentration

Shi (1995) states solids concentration is the most important
factor determining the viscosity of slurry. As already described in
section 2.3 viscosity increases with an increase i n solids
concentration. In the dilute region, the viscosit y increases linearly
with an increase in concentration; as the concentration approaches the
maximum attainable concentration the slurry viscosity rapidly
increases.
Concentrated mineral slurries not only display an increase in

viscosity but also report changes in the flow curve. Several
researchers have reported the flow curve as being Newtonian, then
dilatant, then pseudoplastic (with or without yield) as soli d
concentration increases (Klimpel, 1984; Shabalala et al., 2011; Shi,
17
1995; Shi and Napier-Munn, 2002). Other researchers, such as (Yue
and Klein, 2004) did not find dilatancy; Klimpel’s results are
explainable by sedimentation whilst Shi, Napier-Munn and Shabalala’s
results due to turbulence effects. Sedimentation within the bob and
cup system acts as a drag on the bob increasing the applied torque. It
is well known that turbulence causes an increase in the measured
torque due to inertia. Despite these propositions, no one has fully
demonstrated that dilatancy can occur at low concentrations. Figure
2-7 shows the simulated flow curves of the Hellyer grinding slurry
using a MV-model and a complicated turbulence correction procedure.
18
Figure 2-7: Shi (1995) simulated flow curves of Hellyer grinding
slurry.
At high enough concentrations the suspension will transition

from being pseudoplastic to shear thickening. This shear thickening
effect is caused by the maximum solids concentration being
approached causing the particles to jam (Barnes, 1989; Bender and
Wagner, 1996; Bertrand et al., 2002; Hoffman, 1998; Maranzano and
Wagner, 2001; Wagner and Brady, 2009).
19
2.4.2 Particle size distribution
It has been reported that the particle size affects viscosity and
that decreasing particle size increases viscosity (Kawatra, 1988; Shi,
1995). The hypothesis is that decreased particle size increases the
chances for collisions. However, Shi found that some others reported
the opposite; considering that slurry rheology is extremely complex
and is affected by many factors it is possible that secondary effects
overpowered the expected effect of increased viscosity from decreased
particle size. Any tests must control for secondar y effects to be
scientifically accurate. The link between thermodynami c effects and
particle size are negligible at mineral processing sizes.
As described in section 2.3, particle size distribution effects

viscosity. A wider size distribution allows a greater maximum packing
thus allowing an increase in the maximum solid concentration.
Viscosity increases as the solid concentration approaches maximum
packing.
Shi (1995) believed that a good understanding of the size and the
distribution was key and that the commonly used percent passin g size
was insufficient to describe the particle and size distribution
characteristics. Shi developed a semi-empirical model to predict slurry
rheology which incorporates particle size and distribution; it is
described in section 2.10.1,
20
Figure 2-8: Flow curves of discharge slurries of Emu Gold ball
mills. As can be seen (a) is dilatant with low yield stress whilst (b) is
Pseudoplastic with high yield stress. Taken from Shi (2002).
2.4.3 Chemical environment

The chemical environment has a significant effect on both the
slurry viscosity and rheological character. In a grinding circuit the
aim of grinding aid is to:
1) reduce viscosity;
2) increase the pseudoplasticity;
21
3) and reduce the yield point.
Lihir uses no grinding aids (Gardner and Vollert, 2017).

However, the pH was likely effect both rheology and the fluid
stability; as such it was monitored but not controlled during the
experiments. The dramatic effect that both dispersant (polyacrylic
acid) and the pH can have on the rheology of s uspensions can be seen
in Figure 2-9 and Figure 2-10. As seen in Figure 2-9, there is a pH
which maximises the yield stress, however, as additional dispersant is
added not only does the yield stress decrease but th e pH of the
maximum yield stress shifts to become more acidic. Furthermore, as
seen in Figure 2-10, if the lowest yield stress materia l
(PAA/zirconia=0.01) has its pH ad justed to a value of 8.9 and its
concentration increased to 88.7% (weight percentage) its viscosity is
still lower than the unadjusted pH at a concent ration of 57%(Boger,
2009).
22
Figure 2-9: Effect of low molecular weight polyacrylic acid on
the shear yield stress-pH behaviour for a zirconia suspension at 57%
(wt%) (Boger, 2009)
23
Figure 2-10: Zirconia suspension with additive of polyacrylic at
57% with unadjusted pH and 88.7% with adjusted pH (Boger, 2009).
2.4.4 Temperature
Increasing the temperature decreases the apparent viscosity and
the yield point. Often, performing the rheology tests leads to an
increase in temperature which needs to be corrected. Napier-Munn and
Shi corrected temperature by assuming that the changes in the
24
viscosity of the slurry reflect the changes in the eff ect on temperature
(Shi, 1995); a problem with this is that if the viscosity stems from
mechanical interactions between particles the dependence on
temperature may be small.
Kaolinite clay viscosity and slurry st ability increase with

elevated temperature (Lin et al., 2016). This means that the
temperature correction methods performed by Shi and Napier -Munn
may not be applicable in ores that have clay content. It is noted
elsewhere in this thesis that non -Newtonian flow often depend on
chemical processes such as flocculation. As chemical processes are
temperature dependent it is questionable if the correction methods are
ever applicable.
2.5 Flow models and the importan ce of using

them
There are many models which describe non -Newtonian flow
(Akroyd, 2004; Cruz and Peng, 2016). The most famous are the purely
empirical Bingham and Herschel-Bulkley models. Both models are
often used to calculate yield stress thro ugh data fitting rheograms,
B i n g h a m : 𝜏 = 𝜏0 + 𝜂𝐵 𝛾̇ 2-5
w h e r e τ i s t h e s h e a r s t r e s s , 𝛾̇ i s t h e s h e a r r a t e , 𝜏0 i s d e f i n e d a s t h e
y i e l d s t r e s s a n d 𝜂𝐵 t h e B i n g h m a n v i s c o s i t y .
H e r s c h e l - B u l k e y : τ = τ0 + Kγ̇ n 2-6
where K and n are the consistency and flow indexes respectively.
The issue with using these models to descr ibe slurry rheology
was well expressed by Hallbom (2008) who demonstrated that through
the use of different curve fitting function s dramatically different flow
models were developed. The contradictions were a major concern as
the different flow properties lead to completely different engineering
decisions. He suggested the use of the Casson m odel (2.7) or other
similar models within the same family such as given by 2.8:
25
√𝜏 = √𝜏0 + √𝜂∞ . 𝛾̇ 2-7
𝜏 𝑘 = 𝜏0𝑘 + (𝜂∞ . 𝛾̇ )𝑘 2-8
2.6 Rheometer selection

Viscometers are devices used to measure viscosity; rheometers
are a special type of viscometer which are able to measure the
viscosities of fluids across a variety of shears. This is important if
there is the possibility of non-Newtonian behaviour (May, 2015).
Rotational rheometers operate by rotating a part of the device

and measuring the resulting torque. Rotational rheometers come in
several measuring geometries, the component of the instrument tha t is
in direct contact with the sample, such as parallel plate , cone and
plate, vane, and co-axial cylinder. In contrast, tube rheometers are
based on pipe flow (Akroyd, 2004; Fisher et al., 2007). Due to the
short timeframe, and previous experience at JKMRC, it was decided in
the planning of this project betw een the supervisory team and the
sponsors that the project would use rotational rheometers wi th the co-
axial cylinder geometry.
2.6.1 Parallel Plate
Figure 2-11: Parallel plate (Anon, 2016)
The parallel plate geometry, as seen in Figure 2-11, consists (as

it name implies) of two parallel plates separated by an adjustable gap.
26
The slurry is placed between the plates; one plate remains stationary
whilst the other is rotated. The rotated disk has either a set speed or
an applied torque; in the first case a torque reading may be gained b y
measuring the resistance to motion and in t he latter case the speed can
be measured (Anon, 2016).
The chief advantage of the parallel plate, regarding measu ring

mineral slurries rheology, is that they h ave an adjustable gap. This
allows wide-gap measurements to be conducted (Shah, 2007).
However, parallel plates are not appropriate for settling slurries
(Akroyd, 2004; Boger, 2009).
The disadvantages of parallel plates are that the shear rate varies
across the gap. Hence, an average shear rate must be determined by
varying the gap width. This task is considerably more difficult in non -
Newtonian fluids. Furthermore, sample filling is a non-trivial process
(Anon, 2016).
2.6.2 Cone and plate
Figure 2-12: Cone and plate (Anon, 2016)
The cone and plate geometry are similar to the parallel plate
geometry. The fluid is placed between the cone and the plate; the cone
is attached to the rotor. Like the parallel plate geometry, sample
filling the cone and plate is a non-trivial process. Unlike the parallel
plate geometry the shear rate is constant throughout the sample (Anon,
2016).
27
The chief advantage of the cone and plate are that it has a
consistent shear rate throughout the sample (Anon, 2016). They are
also quick and easy to clean (Anon, 2016).
A disadvantage of the cone and plate is that particles may be

ground at the centre. To avoid this problem the cone is often truncated
Furthermore, the axial gap setting is crucial for ac hieving correct
results (an accuracy in the range of 1 μm is essential) (Anon, 2016).
Like the parallel plate the cone and plate is inappropriate for
measuring settling slurries (Akroyd, 2004; Boger, 2009).
2.6.3 Coaxial cylinder or ‘cup and bob’

The coaxial cylinder geometry consists of two cylinders , one
held within the other as seen in Figure 2-13and are often referred
colloquially as cup and bob rheometers. There are two types of coaxial
cylinder geometries: Couette and Searle. In a Couette geometry the
inner cylinder (the bob) remai ns still and the outer cylinder rotates
(the cup). In Searle, the opposite process happens. Both are commonly
used; there is a difference in the calculations but the pr ocess from the
user’s perspective is almost identical (Van Wazer, 1963).
Figure 2-13:DIN 53019 / ISO 3219 coaxial cylinder geometry

(Anon, 2016)
28
As discussed by Krieger and Elrod (1953), the shear rate is
dependent upon the fluid model and may be calculated by:
𝛾̇ = 𝑓(𝜏1 ) 2-9
r γ̇ 1 τ f(τ)
Ω = ∫r 2 r dr = − 2 ∫τ 2 dτ 2-10
1 1 τ
where Ω = angular speed. The shear stress at the surfac e of the bob
is calculated by:
𝑀
𝜏1 = 2𝜋𝑟 2 𝐿 2-11
1
The co-axial cylinder geometry was used in this project and is

discussed in more detail in section 2.7. The geometry has several
advantages including:
• the large sample volume above and below the cylindrical

part of the rotor make it relatively insensitive to sample
filling (Anon, 2016);
• the cone at the bottom of the rotor makes it easier to insert

the rotor into a higher viscosity samples (Anon, 2016);
• the mathematics describing the shearing is well known

(section 2.7);
• and it can be modified to handle settling slurries (section

2.8).
The disadvantages include:
• cleaning can be time consuming (Anon, 2016);
• relatively high shears can occur as the bob is lowered into

the slurry causing thixotropic breakdown (Fisher et al.,
2007).
2.6.4 Vane
The vane method is where a spindle is attached to a cylindrical
shaft and lowered into the slurry. It has two variables, shear rate and
29
torque; setting one variable allows the other to be measured. Vane
rheology is most often used for finding the yield stress (Barnes and
Nguyen, 2001).
Nguyen and Boger (1983) introduced the vane method for finding
the yield point to minerals processing from soil science. The vane
method is simple to use, allows the direct measurement of yield stress
and avoids wall slip. The yield point is found by increasing the stress
applied to a system over time. As the yield point is reached the vane
begins to rotate.
Vane rheology is not often used for the development of flow

curves. The issue is that it is difficult to know the flow conditions and
thus a flow curve cannot be calculated (Fisher et al., 2007). However,
if the flow conditions could be indirectly gained from the testing rig
geometry it may prove to be a more accurate method of building flow
curves. Some researchers have found that the vane acts as a cylinder
when the fluid either has yield stress or is highly pseudop lastic
(n<0.5) (Barnes and Nguyen, 2001; Barnes and Carnali, 1990; Fisher
et al., 2007). Under such situations the vane rheometer effectively
becomes a low slip co-axial cylinder geometry which can be entered
into the sample with little to no disturbance. It then can use the
coaxial cylinder shear rate calculation s to generate a rheogram as
described in section 2.7.
Figure 2-14: Vane Geometries (Anon, 2016)
30
As written earlier, Shi and Zheng (2003) used a vane geometry to
measure the rheology of a flotation circuit. Shi and Zheng (2003)
argued that vane rheology:
• minimises the disturbance of the slurry upon entry;
• is superior for coarse particles as it avoids the clogging

problem;
• avoids wall slip effects.
The superior ability to handle coarse particles idealises it for

mineral slurry rheology. Li et al. (2015) further developed the use of
vane rheology for measuring the rheology o f flotation circuits and
overcame Shi and Zheng (2003) issue with horizontal froth flow
through the addition of a cylinder and the applied Fisher’s “vane in a
cup” calculations (Fisher et al., 2007; Li, 2016).
Li et al. (2015) vane and cup system can estimate shear stress by
the equation 2-12.
𝑀 𝑀
𝜏=𝐾 = 𝐿 2 2-12
𝑣 2𝜋𝑟 3 ( + )
𝑟 3
where shear stress (τ), vane constant parameter (K v), torque (M),
radius (R), length of the vane (L). To estimate shear depends if the
material is partially or fully sheared. If partially sheared, the
infinite sea solution could be used (see section 2.7.2) and if fully
sheared a modified Krieger solution (see section 2.7.3). To decide
whether the system is partially or fully sheared:
𝜏
𝑟𝑦 = 𝑟1 . √𝜏 2-13
0
w h e r e t h e s h e a r e d r e g i o n i s 𝑟𝑦 .
2.6.5 Capillary rheometer

The capillary rheometer is another common type of rheometer.
The rheometer consists of a reservoir and a capillary tube. Pressure is
31
applied to the reservoir causing it to flow through the capillary tube;
viscosity is determined from the pressure drop and fluid flow rate
(Akroyd, 2004). Previous researchers have found the capillary
(Nguyen and Boger, 1985) to be effective at determining the rheology
of mineral suspensions, however, the capillary rheometer has two
major drawbacks when testing mineral suspensions: it requires a large
amount of sample and a large capillary (Fisher et al., 2007). Due to
particle migration effects, it is not suitable for rapidly settling
slurries (Akroyd, 2004).
Figure 2-15:Capillary rheometer schematic (Akroyd, 2004)
2.6.6 Summary
Four geometries of rotational rheometers were discussed in the
previous subsections (2.6.1 through 2.6.4). Upon review, the selection
of the coaxial cylinder geometry was sound as
• the coaxial cylinder method can be modified to take into

account settling suspensions (this is described in detail in
section 2.8);
• the cone and plate geometries are unsuitable for wide gap
geometries;
32
• the parallel plate was more difficult to modify for settling
suspensions;
• and the vane rheometer could only be used for producing

flow curves in non-settling highly plastic slurries.
However, for such slurries it is possible that the vane
rheometer would have been more suitable.
2.7 Coaxial cylinder shear rate calculations

There are several considerations when using a co-axial cylinder
geometry. The fluid may slip either at the bob’s surface or at the wall.
Wall slip effects can be reduced through a wider gap between the two
cylinders (ISO, 1995; Van Wazer, 1963). However, as the shear stress
and shear rate are not uniformly distributed throughout the fluid
annulus, the process of measuring the rheology becomes significantly
more difficult when the fluid model is unknown (Krieger and Elrod,
1953).
There are three common solutions to this problem:
1. a narrow gap geometry is used, and the wall slip, and non-
Newtonian effects assumed negligible;
2. the data is analysed assuming a fluid model;
3. or approximate methods can be used to estimate the effects.
2.7.1 Narrow gap

Shear rate in Newtonian fluids are dependent upon the angular
speed of the bob and the geometry of the system. Hence, for the
purposes of calculating the shear rate it is a common practice to
assume that the sample shears as a Newtonian fluid. In the stand ard,
narrow gap geometry, this works reasonably well with the calculated
shear rate being within 20% of the true shear rate for many fluids
(DIN53019-1). The size of the error is dependent upon the gap
between the inner and outer cylinders, and also on the fluid
characteristics.
33
The Newtonian shear rate is given by
2Ω.𝑟 2
𝛾̇ = 𝑟 2 −𝑟12 2-14
2 1
Under narrow gap the shear rate may be approximated as being

Newtonian. The DIN53019-1 (2008) standard specifies flow zones for
measuring the flow behaviour of fluids (both Newtonian and Non -
Newtonian) in rotational viscometers by describing a number of
instrument geometries, the flow behaviour being described by the
shear viscosity, the shear viscosity function, the viscosity curve or the
flow curve. According to DIN53019-1 (2008) the standard only applies
if
a) There is no slippage between the fluid and the boundary

surfaces i.e. the fluid adheres to the walls
b) The accelerating forces in the fluid remai n so low that the

flow pattern is determined solely by laminar, stationary
layer flow.
c) Negligible temperature variation (both across the fluid and

time)
d) The flow zones are described by boundary surfaces, each of

which comprises a lateral cylinder surface, a lateral cone
surface or a circular surface (simple boundary surfaces). In
practice, composite boundary surfaces are also used.
e) Any boundary surfaces other than these simple ones have

no effect on the flow zone
f) Any circularity defects in the boundary su rfaces or errors

in their axial alignment shall be negligible.
The standard requires that the size of any solid particles in the
sample shall not be greater than one fifth of the narrowest point in the
measurement gap.
34
Figure 2-16: Coaxial cylinder image taken from ISO (1995)
2.7.2 Infinite sea

In an incompressible yield stress fluid, the fluid within the gap
will either undergo complete shearing or be partially sheared between
the two cylinders. Naturally, this allows the equation:
r γ̇ 1 τ f(τ)
Ω = ∫r 2 r dr = − 2 ∫τ 2 dτ 2-10
1 1 τ
35
w h e r e a n g u l a r s p e e d ( Ω) , o u t e r r a d i u s ( r2 ) , i n n e r r a d i u s ( r1 ) , s h e a r r a t e
( 𝛾̇ ) , s h e a r s t r e s s ( τ) , s h e a r s t r e s s a t t h e o u t e r r a d i u s ( τ2 ) , s h e a r s t r e s s
a t t h e i n n e r r a d i u s ( τ1 ) , a n d t h e f l u i d m o d e l ( f(τ)) t o b e s o l v e d w i t h o u t
any additional knowledge of the fluid model (Krieger, 1952). As
explained by Nguyen and Boger (1987), the shear rate is given by:
Ω
𝛾̇ = 2 n 2-15
𝑑𝑀
𝑛= 2-16
𝑑𝛺
where torque (M), which is equivalent to:

𝑑𝛺
𝛾̇ = 2 𝛺 2-17
𝑑𝑀
2.7.3 Krieger approximate solution

Krieger and Elrod (1953) developed an approximate solution is
often given as:
Ω (𝑚.𝑙𝑛𝜀)2 (𝑙𝑛𝜀)2 𝑑𝑚
𝛾̇ = 𝑙𝑛𝜀 . (1 + 𝑚. 𝑙𝑛𝛼 + + . 𝑑𝑙𝑛(𝜏)) 2-18
3 3
w h e r e s h e a r r a t e ( 𝛾̇ ) , a n g u l a r s p e e d ( Ω) , c u p t o b o b r a d i u s r a t i o
(α), yield stress (τ0) and
𝑑𝑙𝑛Ω
𝑚 = 𝑑𝑙𝑛(𝜏) 2-19
2.7.4 Estelle et al Bingham approximation

Estelle et al (2008) re-arranged the Bingham equation into the
rheometer’s measured properties and the derivative of angular speed
vs torque. It was selected as the preferred method and is discussed
more in the methodology (see section 4.9). The shear varies
throughout a Couette system with larger shearing forces experienced
near the inner cylinder and small forces experienced towards the outer
cylinder. In a yield stress fluid this may result in it becoming only
partially sheared; the shear stress is sufficient near the inner cylinder
to cause flow but is insufficient at the outer cylinder. Hence, two
36
alternative shear rate calculations are performed : a fully sheared and a
partially sheared condition.
𝑑Ω
𝛾̇ = 2𝑀 (𝑑𝑀) , 𝑖𝑓 τ2 < τ0 2-20
𝑑Ω 𝑑Ω
𝑀( ) Ω−𝑀( )
𝑑𝑀 𝑑𝑀
𝛾̇ = 2 𝑟2
− 𝑟1 , 𝑖𝑓 τ2 > τ0 2-21
1−( 12 ) ln( )
𝑟2
𝑟2
w h e r e s h e a r r a t e ( 𝛾̇ ) , t o r q u e ( M ) , a n g u l a r s p e e d ( Ω) , y i e l d s t r e s s
( τ 0 ) , o u t e r r a d i u s ( r2 ) , i n n e r r a d i u s ( r1 ) , s h e a r s t r e s s a t t h e o u t e r
r a d i u s ( τ2 ) , s h e a r s t r e s s a t t h e i n n e r r a d i u s ( τ1 ) .
The shear rate is then determined from the maximum of those to

equations. It was shown that this estimate produces very low error on
a variety of synthetic fluids (Estellé et al., 2008).
2.8 Measuring the rheology of settling slurries

Many researchers have attempted to measure the rheology of
settling mineral slurries as slurry viscosity affects the energy
consumption and design factors of mineral processing units. As a
suspension settles forms a concentration gradient between a solids
depleted supernatant and a consolidated solid zone (see Figure 2-17).
The settling nature of the slurries make measurement difficult as

it can give erroneously too low or too high viscosities. If the solids
concentration decreases around the bob then the results of the
rheological tests will incorrectly show that the substance is less
viscous then it really is and that it is thixotropic (see Figure 2-18).
Likewise, if the sediment builds up it can impede the movement of the
bob giving falsely high shear stress readings (Klein et al., 1995). One
method to overcome the sedimentation problem is too elongate a
37
coaxial cylinder; another is the use of helical flow rheometers.
Figure 2-17: A schematic showing particle sett ling zones formed

in a conventional cup and bob rheometer (Klein et al., 1995)
38
Figure 2-18: Rheological measurements of a fly ash slurry using
a standard bob and cup geometry (Akroyd, 2004).
2.8.1 Elongated coaxial cylinders

Klein et al (1995) developed a modified coaxial cylinder system
to allow the rheological properties to be measured sedimentation. The
measuring device was an elongated double cup system. The bob is
positioned with the constant density zone of the settling suspension
for the duration of the measurement as seen in Figure 2-19. The
authors also note that the used vertical grooves to reduce wall slip
errors and a narrow gap to allow the Newtonian shear rate
approximation to be used with minimal error. The confirmed the
39
Newtonian solution produces minimum error by comparing the results
of the Newtonian calculation with the Krieger.
Figure 2-19: The Klein et al (1995) solution to measuring the

rheology of settling slurries.
2.8.2 Debex
Debex style rheometers have been used in several projects at the
JKMRC (Reeves, 1990; Shi, 1995, 2016). Slurry is pumped into a
measuring chamber where the viscosity could be measured as a Couette
flow. The pressure head was maintained by the overflow.
40
Figure 2-20: A schematic of a testing rig built on the same
principles as the Debex rheometer (Shi, 2016)
Some researchers have used modified recirculating versions

based off the Debex basic des ign as seen in Figure 2-21.
Figure 2-21: A previous viscosity testing rig used by Shi (2016)
The major problems with these rheometers were temperature

effects weren’t controlled (though correction procedures were used)
41
and axial effects weren’t considered in the data analysis. Another
acknowledged difficulty was turbulence effects; complica ted
statistical work was conducted to overcome these issues (Shi, 1995).
Furthermore, these rheometers were designed to consu me a large
volume of slurry required (greater than 20L). I was unable to find any
detailed discussion of the commissioning of such rheometers nor their
start-up within the literature.
2.8.3 Akroyd-Nguyen helical flow rheometer

Nguyen et al. (1999) developed of a helical flow rheometer
which, like the Debex, combined the Couette f low with the axial.
However, unlike the Debex style rheometers, the fluid flows upwards
into the measuring chamber in a helical fashion; hence its name the
helical flow rheometer. Careful attention was applied to the analysis
as it effectively involved asp ects of both rotational and capillary
rheometers.
Akroyd (2004) investigated the helical flow rheometer against

other standard laboratory equipment and found good agreement.
Particle migration, although it could not be directly observed, was
found to be minimal or non-existent presumably due to the short
residence time.
A significant contribution by Akroyd and Nguyen was the

inclusion of both Couette and axial flow components in the Taylor
vortices analysis. The stability criteria for Newtonian helical flow was
well known; they developed a modified Taylor number using the
power-law model to predict the onset of Taylor vortice s for non-
Newtonian fluids.
Extensive discussion is provided by Akroyd (2004) on the

commissioning of his rheometer. However, the major drawback in his
design is the extreme sensitivity to pulsed flow as the fluid flows
upwards into the measuring chamber. Like the Debex, I was unable to
42
find any discussion in the start-up of the helical flow rheometer.
Akroyd (2004) also faced difficulties with re-circulating his slurry.
Figure 2-22: A schematic used as the base of the design of the

Akroyd (2004) helical flow rheometer
Akroyd (2004) calculated large errors in the shear rate when

using the coaxial flow assumption for analysing the data generated by
his helical flow rheometer. The error was strongly associated with the
slurry pressure and the rotational speed of the bob as seen in Figure
2-23. Clearly, the larger the pressure and the smaller the rotational
speed result in larger errors. A problem with generalising this result
for the Debex style rheometer is the pressures involved in this
experiment are unlikely to be relevant; the Debex rheometer axial fl ow
through the measuring chamber flows under gravitational force rat her
than pumping. Assuming that the density of the slurry is twice that of
water and that the measuring chamber is 10 cm long than the
hydrostatic pressure would be less than 2000 Pa which is half the
smallest pressure measured by Akroyd (2004).
43
Figure 2-23 Error in shear rate between flow rheometer data
analysed using helical flow and Couet te flow theory for a 1 wt% CMC
solution (taken from Akroyd (2004))
2.9 Clay mineralogy

Mineralogy effects the rheological pro perties of minerals (Boger,
2009; Burdukova, 2008). Of the components of Lihir it was
hypothesised that the minerals of most interest for their rheological
effects were clay minerals. Also of interest were amorphous silica
(non-crystalline silica (SiO2)) as previous research by Chen et al.
(2017) found that that amorphous silica had a significant effect on the
viscosity in comparison to quartz (a crystalline silicate); this effect
grew more pronounced with increasing concentratio n as seen in Figure
2-24.
44
Figure 2-24: Apparent viscosity of quartz and amorph ous silica
suspensions at 100s-1 as a function of solids concentration (Chen et
al., 2017)
2.9.1 Clay
Clays are well known as a complicating factor in mineral
processing caused by numerous properties of clay such as clay
agglomerating and swelling. In additional, clay particles have been
shown to increase the viscosity and the y ield stress (Cruz and Peng,
2016).
‘Clay minerals’ refer to a large quantity of different substances;

according to Bergaya and Lagaly (2013) clays can be either
phyllosilicates or non-phyllosilicates and have no size connotation.
The structure of phyllosilicates is based on a tetrahedral (T) and
octahedral (O) sheets that may condense either into a 1:1 or 2:1
proportion to form anisotropic TO or TOT layer. These layers then
45
join together to form particles, which can aggregate into larger
assemblies as shown in Figure 2-25.
46
Figure 2-25: Diagram showing a clay mineral layer (A); a
particle (B); an aggregate (c); and an assembly of aggregates (D)
(taken from Bergaya and Lagaly (2013)).
47
Figure 2-26:edge/face and face/face contacts that link the
particles to a house of cards (a) or a band-like (b) structure (Lagaly,
1989).
Clay agglomerates due to a variety of causes such as the presence

of inorganic salts (such as mineral dispersions). The agglomeration is
explained by the well-known DLVO theory which describes there being
competing electrostatic repulsion between diffuse ionic layers around
the particles and the van der Waals attractions. Thus a change in
chemical composition may have a major effect on clay agglomerates
forming, and the agglomerates stability (Bergaya and Lagaly, 2013).
Lagaly (1989) found that the flow of kaolin and bentonite suspensions
was determined by the formation of house of cards (positive
edge/negative face) and band like structures (negative face/negative
face). Both form the house of cards structure in an acidic environment
but band like structures in a basic environment.
The swelling effect depends on the particular kind of clay (not

all clays swell), the chemical environment and the temperature. For
accurate rheology, constant volume ne eds to be maintained thus this
needs to be taken into account (Bergaya and Lagaly, 2013; Cruz et al.,
2015).
48
2.9.2 Lihir Clays
The Lihir domain classification system includes clay and as
stated in section 2.1 consists of kaolinite, illite/smectite, and
montmorillonite. It is likely that, from the most to the least complex
effects on rheology is likely to be (in descending order from the most
complex to the least): smectites (including montmorillonite);
kaolinite; then illite (Farrokhpay et al., 2016; Ndlovu, 2014). The
structure of these clays is given in Figure 2-27.
Figure 2-27:A classification of phyllosilicate clay minerals

(Ndlovu, 2014)
49
2.9.2.1 Kaolinite
Kaolinite is a 1:1 structure (Bergaya and Lagaly, 2013; Bridges,
2014; Chakraborty, 2013; Cruz et al., 2015). Kaolinite does not swell
due to its fixed 1:1 silicon atom to aluminium atoms held together by
oxygen and has little isomorphic substitution and a low cation
exchange capacity (Bridges, 2014).
Lin et al. (2016) found that the yield stress of Kaolinite

suspensions increases with increasing temperatur e. The yield stress
difference is more obvious at higher concentrations and it is believed
to be due to electrostatic potential difference.
Previous researchers have found cases where kaolinite was

strongly time dependent (Lagaly, 1989; Smith et al., 2000). The time
dependent behaviour was noted to depend on the shear rate, the pH’s
and salt content can result in either thixotropic or anti-thixotropic
behaviour (Lagaly, 1989; Smith et al., 2000).
2.9.2.2 Illite/Smectite
Illite and smectite are often found mixed. Both have a 2:1 -layer
structure. Smectites are known for having a profuse hydrational
behaviour and are known to swell whilst illites are non -expanding
(Bergaya and Lagaly, 2013). Previous research indicates that illite has
little effect compared to other clays on the rheology (Farrokhpay et
al., 2016).
Bentonite is a clay which belongs to the smectite group and its

effect on rheology has been actively re searched. As explained by Cruz
et al. (2015), when a bentonite slurry has a high concentration of Ca 2+
cations, the Na+ cations in the interlayer space may be exchange d such
that it is no longer possible for hydration and swelling to occur.
However, the presence of gypsum and lime can increase water
absorption and swelling pressure of bentonite; it is believed tha t the
interactions of Ca2+ cations and SO42− anions play a role on this
effect. Cruz et al. (2015) showed that bentonite becomes pseudo -
plastic from 10% (wt%).
50
2.9.2.3 Montmorillonite
Montmorillonite is a variety of smectite. It has a 2:1 layer
structure, a cation exchange capacity and the ability to swell. Smith et
al. (2000) referenced a case where montmorillonite had both
thixotropic and anti-thixotropic depending on the shear rate history.
2.9.2.4 Mixtures of different clays
Merrill et al. (2017) compared the slurry rheology of different
compositions of five minerals; three minerals were clay mine rals (two
bentonites, primarily montmorillonite from different sources and a
kaolin), quartz and white mica. Up to three minerals out of five were
combined allowing the production of ternary plots comparing how
different mineral combinations effect the rh eology. It was found that
blending these ores led to synergistic effects; the rheology was lower
in mixtures then in either components pure state (see Figure 2-28).
Merrill et al. (2017) proposed several theories to explain these effects
such as the changes in the microscopic structure or that the presence
of multiple minerals result in competitive ion adsorption generating
non-DLVO interactions. The knowledge of these synergetic effects
could lead to beneficial decision making including blendin g ores to
reduce yield stress and viscosity.
51
Figure 2-28: Viscosity ternary plots in volumetric fractions at
ph7. Note, that the plots are a different solid concentration ; the solids
are at 15% and 30% volumetric concentration for the graphs on the left
and right respectively.
2.9.2.5 Summary of Lihir clays
It is likely that, from the most to the least complex effects on
rheology is likely to be (in descending order from the most complex to
the least): smectites (including montmorillonite); kaolinite; then illite
(Brandenburg and Lagaly, 1988; Farrokhpay et al., 2016; Ndlovu,
2014). Both smectites and kaolinite are known to display time
dependency (Lagaly, 1989; Smith et al., 2000); this time dependency
is highly complex consisting of both thixotropy and anti -thixotropy.
The rheological effects caused by clay particles are highly

dependent on their surface chemistry. This surface chemistry can be
manipulated to radically change the rheological properties (see section
2.4.3).
The blending of different clay particles can have synergetic

effects on the rheology. As Lihir is a real, highly complicated ore
body many clay minerals will be naturally mixed. From Figure 2-1 it
appears as that argillic material will be more problematic then
epithermal or porphyry material. Within the epithermal group upper
52
epithermal would likely be the most viscous whilst silica breccia and
lower epithermal samples would behave similarly.
2.9.3 Rheology measurements of slurrie s with

high clay content s
The presence of clay minerals within flotation slurries negatively
affects concentrate grade and recovery. Clay particles form loose
aggregates which mechanically entrain precious metals reducing
concentrate grade. These aggregates form cross-linked clay network
structures, such as the house-of-cards, and lead to high pulp viscosity
thus reducing recovery (Cruz and Peng, 2016). The Bingham,
Herschel-Bulkley and Casson models are often used to calculate yield
stress through data fitting rheograms.
Both of these models are of limited use in slurries with high clay
contents in regard to flotation performance. The models do not give,
or in fact incorporate, information on particle associations as they are
focused on yield stress and apparent viscosity; metrics which fail to
predict floatation performance in slurries wi th high clay contents
(Cruz and Peng, 2016). This is likely due to time dependent effects
such as thixotropy stemming from the clay particle agglomerates.
53
Figure 2-29: Time dependent viscosity of nuclear waste slurries
showing thixotropic, anti-thixotropic and time independent behaviour
under different shear rates (Smith et al., 2000).
Smith et al. (2000) investigated the rheological properties of

nuclear waste slurries finding a complicated profile with anti-
thixotropic, thixotropic, and pseudoplasticity. As seen in Figure 2-29,
at low shear rates, time dependent shear thickening occurs and at
higher shears time dependent shear thin ning; Smith et al explain this
as being caused by changes in the microstructure. At low shears
particles aggregate due to hypercoagulation whilst at higher shears
particle agglomerates break.
2.9.4 Time dependency in clay suspensions

Complex time dependency has been noted in other mineral
suspensions (Barnes, 1997; Boger, 2009; Brandenburg and Lagaly,
1988; Coussot et al., 2002; Lagaly, 1989; Larson and Wei, 2019;
Mewis and Wagner, 2009; Nguyen and Boger, 1985). Larson and Wei
(2019) reviewed time dependency and found that real thixo tropic
54
materials typically show a yield stress while anti -thixotropic material
are often flowable fluids at low stress that becom e more viscous or
even solid-like during and after flow.
Figure 2-30: Viscosity against time for 1997 ppm partially

hydrolysed polyacrylamide in water/glycerine (Buitenhuis and
Springer, 2003)
A typical example of rheological behaviour is given in Figure

2-30. The figure depicts the constant shearing at a constant stress of
2.74 Pa of a partially hydrolysed polyacrylamide solution in a
water/glycerine mixture. The fluid has no yield stress, and thus begins
to flow under the initiation of s hear; after reaching a critical shear
stress, it thickens until it reaches a second, higher viscosity. The
molecular origins of the flow-induced build-up of the network
structure are unknown. The gradual nature of the build -up
distinguishes the flow from t he sudden build increase in viscosity
caused by jamming.
Other researchers have described clay suspensions which could

undergo both time dependent shear thinning and shear thickening
55
behaviour (Brandenburg and Lagaly, 1988; Coussot et al., 2002;
Lagaly, 1989). Coussot et al. (2002) defined these flow curves as
being “viscosity bifurcation”; unlike ideal yield stress fluids they stop
flowing abruptly below a critical stress and start flowing at a hi gh
velocity beyond a critical stress which in addition increase with the
time of preliminary rest. Hence, for such fluids there is no
(homogeneous) steady state flow at a shear rate below a critical value
as there is bifurcation of the behaviour towards ei ther stoppage or
rapid shear. This effect is illustrated in Figure 2-31.
Figure 2-31: Flow curve for a thixotropic yield -stress fluid

experiencing bifurcation contrasted against an ideal yield stress fluid
(Coussot et al., 2002).
Coussot et al. (2002) used an industrial grade bentonite to create

a model suspension. The bentonite was mixed w ith distilled water at
solids concentration of 4.5% with no chemical additives. The
suspension was first agitated continuously for about three hours to
ensure complete homogenization and then left to rest for at least 24
hours to allow both hydration and d ispersion of the bentonite
particles. Prior to the rheological tests the suspension was gently
56
stirred for one hour. All rheological tests were performed using a vane
in cup geometry with a wide gap; the apparent shear rate was
calculated by assuming homogeneous shear within the gap. A parallel -
plate geometry with two different gap spacings (1 and 2 mm ) that was
roughed with sand paper on both sides was also used to co obtain data
for comparison with the vane in cup results; its apparent shear rate
was calculated from the relative velocity between the two disks at the
periphery for the second geometry. Surprising, these two techniques
gave similar results. The results of the vane in cup are shown in
Figure 2-32; they clearly indicate that with bentonite suspension there
can be the aforementioned viscosity bifurcation effect.
Figure 2-32: shear rate versus time for a 4.5% be ntonite

suspension after 20s of rest following an intense 60s of 26 Pa pre-
shear using a vane in cup geometry.
57
2.9.5 Accurate sizing of clay particles
Clay science is a multidisciplinary field in which the scientists
have a wide variety of backgrounds including geology, mineralogy,
chemistry, physics, and biology reflecting the wide variety of interest
in its applications. This has incidentally led to several different
definitions of clay; and while there is no generally agreed upper limit
to the particle size of clay it is a key parameter in all definiti ons
(Bergaya and Lagaly, 2013).
Bergaya and Lagaly (2013) gave a series of different maximum

sizes for clay ranging from 1 µm to 4 µm. Cruz et al investigated the
interaction of clay minerals with gypsum and its e ffects on copper-
gold flotation and in doing s o measured using a MasterSizer Microplus
(Malvern Instruments) both kaolinite and bentonite as having a P80
(80% passing rate) of 14 µm (Cruz et al., 2015). This is significantly
larger than the largest size given by Bergaya and Lagaly (2013); Cruz
et al performed an XRD analysis on their kaolinite sample finding that
their clays were only 63 to 85 % clay. This indicates either:
1. The non-clay component significantly increased the particle

size.
2. The MasterSizer is incapable of fully distinguishing cla y

particles and clay agglomerates.
3. The non-spherical nature of clay makes it inappropriate for

spherical sizing techniques.
Of the three it is certain that the clay’s non -spherical shape

prevents it from being sized by laser diffraction techniques (Napier-
Munn et al., 1996; Otunniyi, 2013; Trottier et al., 2010; Wills and
Napier-Munn, 2006). Other common mineral sizing techniques such as
sieving and cyclosizing face the same problem with non -spherical
sizing. Furthermore, sieving is difficult to perform on particles
58
smaller than 38 μm and cyclosizing is not appropriate for mixtures
with large differences in particle weights (Napier-Munn et al., 1996).
Care should be taken when comparing the results from different sizing
techniques as seen in Figure 2-33.
Figure 2-33: Comparison of different sizing techniques (Napier-

Munn et al., 1996)
2.10 Rheology’s effect on mineral processing

performance
2.10.1 Grinding circuit
Shi (1995) recommended future work be undertaken into
incorporating rheology into the modelling of ball mill operations. In
particular, he recommended linking the slurry’s effect on power dr aw
with the particle breakage rate. Currently, there is a gap in the
literature on recent work on the breakage rate but power draw has
continued to be developed by Morrell who has successfully developed
a power draw model which incorporates the slurry phas e in autogenous
(AG), semi-autogenous (SAG) and ball mills. The model accurately
describes the effect that slurry can have on AG and SAG mills.
Morrell doesn’t provide his model, however, in his paper he describes
59
its structure. Other researchers have inve stigated rheology’s effects in
the grinding circuit; a particular notable example is Yue and Klein
(2004) who showed that fine production is hindered by the emergence
of yield stress.
2.10.1.1 Shi’s MV model
Shi developed a semi-empirical model for predicting rheology
from easily measured slurry properties. It differs from other viscosity
models as it predicts machine output readings of a particular
viscometer at various bobbin rotational speeds; these machine output
readings and bobbin rotational speeds can then be converted to shear
stress and shear rate allowing the rheological flow curves to be built.
The slurry properties incorporated into the model are
• solids volume fraction;
• particle size and size distribution;
• and slurry temperature.
The solids volume fraction is determined from the multiplication

of the mass fraction and the specific gravity. The particle and size and
size distribution effects are expressed thr ough the size at which 80%
(P80) and 20% (P20) of the particles pass. Temperature is assumed to
only effect the viscosity based on the changes made to the carrying
fluid.
The model is given by

𝜙
𝑀𝑉 = 𝑀𝑉0 + 𝑘4 . 𝐻𝐷. 𝛺 + 𝑘5 . 𝑃80 . 𝛺 2 + 𝑘6 . 𝛺 3 2-22
where MV is the viscometer reading, k x are fitted model parameters,

Ω is rotations per minute, and 𝜙 is solids volume fraction, HD is
hydrodynamic interactions, and
𝜙
𝐻𝐷 = 𝑒𝑥𝑝 (1−𝜙∗(𝑃20)𝑘2 .𝑒𝑥𝑝 (𝑘 ) 2-23
3 .𝑇)
𝑀𝑉0 = 𝑘1 . (𝐻𝐷 − 1) 2-24
60
The advantage of Shi’s MV model, is that it allowed rheology to
be inferred without measurement. This is highly useful when
measurement is physically difficult to perform such as inside a
grinding mill. Its faults were th e inability to translate the results to
other rheometer, it involved complicated data reduction procedures,
and as can be seen involves many parameters.
2.10.1.2 Breakage and mass transport
Slurry rheology affects milling energy efficiency as it influences
both the breakage behaviour and the mass transpo rt mechanism (Shi
and Napier-Munn, 2002). Shi and Napier-Munn (2002) developed a
macroscopic term they named the Grinding index which represented
the overall breakage of particles passing through a grindi ng mill. It is
not based on a size by size breakage rate nor does it take into account
the mill throughput and power consumption and hence, it cannot be
said that increasing grinding index necessarily increases the grinding
efficiency. Despite these limitations, the grinding index allows certain
interpretations on grinding to be developed.
𝑆 −𝑆
𝐷 𝐹
𝐺𝑟𝑖𝑛𝑑𝑖𝑛𝑔 𝐼𝑛𝑑𝑒𝑥 = (100−𝑆 ) ∗ 100 2-25
𝐹
where
𝑆𝐷 = 𝑝𝑒𝑟𝑐𝑒𝑛𝑡 𝑝𝑎𝑠𝑠𝑖𝑛𝑔 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑠𝑖𝑧𝑒 (38 𝜇𝑚) 𝑖𝑛 𝑚𝑖𝑙𝑙 𝑑𝑖𝑠𝑐ℎ𝑎𝑟𝑔𝑒
𝑆𝐹 = 𝑝𝑒𝑟𝑐𝑒𝑛𝑡 𝑝𝑎𝑠𝑠𝑖𝑛𝑔 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑠𝑖𝑧𝑒 (38 𝜇𝑚) 𝑖𝑛 𝑚𝑖𝑙𝑙 𝑓𝑒𝑒𝑑
Shi and Napier-Munn recommended the development of grinding

control based upon slurry rheology (Shi and Napier-Munn, 2002). They
gave several suggestions:
• if a slurry contains a relatively high content of fines and

low solids concentration, a dilatant nature is likely to
develop.
• In grinding a dilatant slurry with a low yield stress, the

Grinding Index is increased by raising the slurr y density.
61
• If a slurry has a high yield stress further increasing slurry
density will significantly increase the apparent viscosity
and decrease the grinding index.
• Slurries with relatively few fines and a moderate solids

concentration will have a low yiel d pseudoplastic flow.
• A higher Grinding index can be achieved in low yield

pseudoplastics by increasing the fines in the mill feed as
this increases the particle packing efficiency.
As explained in section 2.4.1, it is possible that the reported

dilatant nature was due to the testing work being carried out under
turbulent conditions. However, much of the general, qualitative advice
should still hold.
2.10.2 Flotation
Flotation is a popular separation technique that uses differences
in surface properties to separate valuable minerals from gangue as
shown in Figure 2-34. A ground slurry is diluted and treated with
reagents making either the minerals or the gangue hydrophobic befor e
being fed into the flotation cell. The flotation cell aerates the slurry;
the hydrophobic particles attach to the bubbles surfaces and are
transported to the pulp surface. Froth, three phase bubbles, forms a
zone at the top of the pulp; the froth is tran sported over the
concentrate lauder. Not all of the froth makes it to the concentrate;
some of the bubbles coalescence or burst and the particles they were
carrying return to the pulp zone.
62
Figure 2-34: The flotation process. Taken from Li (2016)
The critical factor in flotation recovery is the froth. The

rheology of the froth affects the froth mobility and stability (Shi and
Zheng, 2003), which ultimately means it affects the froth recovery.
When a froth has greater viscosity it has a longer residence time and
an increase chance of froth drop back or collapse. Li (2016) observed
that the change in froth viscosity was a function of the change in the
bubble size and the percentage of the bubble surface covered by
particles. The smaller the bubbles and the greater the particle
coverage of the surface area of the bubbles led to a higher viscosity.
Shabalala et al. (2011) found that hydrodynamics of a slurry in a

flotation cell are influenced by the cell characteristics, impeller
properties, and the slurry properties such as density and rheology.
Cruz (2015) found that adding clay minerals to a fl otation cell
increases pulp viscosity and causes higher yield stress.
Froth’s rheology is difficult to measure. Froth’s are unstable 3

phase mixtures, and are often smaller than coaxial cylinders, hence
using coaxial cylinders can be destructive thus unrel iable. Shi and
Zheng (2003) instead of coaxial cylinders a vane measuring geometry
(see section 2.6.4) as:
63
• It minimises the disturbance when lowered into froth
compared to other measuring systems.
• There is no measuring gap thus avoiding the problem with

the coaxial cylinders system.
• Wall slip effects can be avoided.
This technique failed to take into account horizontal froth flow.

The horizontal froth flow interferes with the rheology measurement
masking changes in the rheology. Li et al. (2015) used the vane in cup
geometry to overcome the horizontal froth flow which is discussed in
detail in section 2.6.4.
2.10.3 Other mineral processes

Rheology is interesting in other mineral processes such as
autoclave, hydrocyclone, and tailings disposal (Boger, 2009).
However, this is beyond the scope of this project and has not been
reviewed in detail.
2.11 Sedimentation of mineral ores

As is well known, sedimentation occurs in suspensions where the
solids have greater density than the carrying fluid. Due to drag, the
solids will settle at a reduced rate i n a viscous fluid than a solid
falling through a vacuum. The equation for the free settling velocity
of a sphere in the dilute region is given by equation 2 -26.
𝑑2 (𝜌𝑠 −𝜌𝑓 )𝑔
𝑈∞ = 2-26
18𝜂
w h e r e 𝑈∞ i s f r e e s e t t l i n g v e l o c i t y , d i s t h e d i a m e t e r o f t h e
settling solid, ρ is the density, g is gravity and the subscripts s and f
refer to solid and fluid respectively. There are several issues with the
use of equation 2-26 in the sedimentation of mineral slurries; of these
three are particularly relevant for this study. Firstly, mineral particles
are neither spherical nor are they of a uniform size. Secondly, mineral
slurries often display yield stress; m aterials with sufficient yield
64
stress (Ovarlez et al., 2012) do not settle. Thirdly, mineral slurries
and suspension have high concentrations; at hig h concentrations of
solids hindered rather than free settling occurs (Lee, 1989).
As such, clearly sedimentation is linked to rheology in three

important ways. Firstly, if sedimentation occurs, the accuracy of the
rheological tests will be impacted. Secondly, sedimentation does not
occur in materials with sufficient yield stress (Ovarlez et al., 2012).
Thirdly, a higher viscosity results in a lower particles sedimentation
rate. Despite these three links, little research has been performed in
the minerals area on the link between sedimentation and rheological
properties. What research has been performed has found that mineral
suspensions are fast settling and that the sedimentation rate is linear
as seen in Figure 2-35.
Sedimentation is important to several downstream processes such

as dewatering and flotation (Lee, 1989). As it is linked with both
rheological characteristics (both yield stress and plastic viscosity) and
solid concentration which in turn are linked (see section 2.4.1) it can
be difficult to decouple these effects.
65
Figure 2-35: Ferrosilicon sedimentation rate at a variety of
densities (Shi, 2016)
2.12 Predicting Lihir ore slurry rheology

From the literature, it would appear that Lihir ore slurries would
become more viscous at higher concentrations (see section 2.3 for
more details). A smaller particle size and particle size distribution
will also lead to an increase in v iscosity due to packing effects
(2.4.2).
A higher temperature decreases the viscosity of the carrying

fluid (water) but has been observed to increase the yield stress of
Kaolinite materials (see section 2.9.2; (Lin et al., 2016) ).This makes
it impossible to predict the temperature effect on the rheology.
The slurry rheology will be dependent on the chemical

environment. The naturally different pH’s may result in different
rheological behaviour.
66
The clays contained in Lihir are known to swell (see section
2.1). For reliable rheology it is required to know the volume
concentration; however, the point of this project is to apply this
knowledge to a real grinding circuit (see sections 1.1 through 1.3 for
more details). As such the slurries should be well mixed before
beginning the experiments but not mixed for so long that they are not
materially similar to what would be occurring in the grinding circuit.
Swelling, if it occurs, would cause a decrease in the water volume and
an increase in the clay size thus leading to an increase in the solid
volumetric concentration. If this occurs during a rheology experiment,
then the effects of increasing solids concentration will be observed.
At sufficient concentrations mineral slurries display non -

Newtonian behaviour (see section 2.4). Both mineral and clay
suspensions at sufficient concentrations have been observed to contain
yield stress, and shear thinning behaviour. Mineral slurries also have
been observed to contain shear thickening behaviour both at very high
concentrations and very low concentrations. The high concentration
shear thickening behaviour effects are generally accepted; there are
reasons to doubt the low concentration results as being due to
instrumental effects. On the weight of the literature it appears that the
slurries will:
• be Newtonian at very low concentrations;
• become non-Newtonian at higher concentrations with the

emergence of yield stress. The yield stress will grow with
further increases in concentration. The literature on the
shearing profile is contradictory wit h some authors noting:
o shear thinning behaviour;
o shear thickening behaviour followed by shear thinning at

a further increase in concentration. Several literature
articles cast doubt on this observed behaviour being due
to instrument effects; on the weight of the evidence I
67
agree with the sceptics and believe that the material wi ll
not transition from Newtonian to shear thickening to
shear thinning.
• At very high concentrations, the material will become shear

thickening due to the maximum packing being appro ached
(see section 2.3).
Time dependency has been observed in yield stress materials

(Barnes, 1997; Barnes and Carnali, 1990; Boger, 2009; Nguyen and
Boger, 1992). Under yield stress tests the material has a gradual
increase in stress before reaching a peak and declining to a new lower
equilibrium. This time reduction in yield stress has been observed to
be partially reversible and partially irreversi ble; materials therefore
needs to be handled identically before a test to make them comparable
(see section 2.2.2).
Complicated shear dependent tim e dependency has also been

observed in clay materials (see section 2.9.3). Both shear thinning and
shear thickening behaviour have been observed in the same material
under different conditions.
2.13 Gaps in the literature

From the literature, it would appear that Lihir ore slurries at
sufficient concentrations would display both yield stress and shear
thinning behaviour (see section 2.12). However, it is not clear what
plastic model would best describe the slurries, let alone the
parameters of the model (see section 2.5). Furthermore, it is not clear
at what concentrations the yield stress would become sufficient for
them to become non-settling (or at least settle very slow slowly).
Flow rheology has been successfully used to produce flow

curves, however, there are gaps in the literature about commissioning
a helical flow rheometer, and the practicalities of loading and the
start-up of the rheometer. In summary, the identified gaps in the
literature to the aims of this project are:
68
• how to build and operate a helical flow rheometer allowing
the accurate measurement and production of flow curves for
a coarse unstable slurry with a high clay content and
limited sample;
• the rheological and stability (i.e. sedimentation rate)

characteristics of Lihir ore slurries
69
Chapter 3
Testing rig and rheometer
commissioning
At the beginning of this project it was believed by the
supervisory team that the Lihir ore slurry would be fast settling even
at the high concentrations (see section 2.11) as such the project
required the commissioning of a helical flow rheometer. There were
sample availability constraints which prevented the testing of the
sedimentation rate until after the helical flow rheometer was
commissioned. The initial validation experiments were performed on
outer biotite samples which were found to be fast settling. However,
through the course of this project it was discovered that many of the
samples were in fact slow settling, especially at high concentratio ns.
The core rheometer unit within the helical flow rheometer

malfunctioned preventing further use. This catastrophic failure
sparked a re-design of the experimental methodology; instead of using
a helical flow rheometer a more traditional coaxial cylinde r geometry
was used. There were several downsides to this design change but the
most significant is that only high concentrations could b e investigated
as the ore becomes faster settling at low concentrations. Higher rates
of sedimentation reduce the relia bility of the results as discussed in
section 2.8.
This chapter explains the commissioning o f both the helical flow

rheometer and the coaxial rheometer used in this project. This
includes discussing the design considerations and some of the
difficulties which emerged throughout this project. A summary is
70
provided at the end of this chapter (secti on 3.5) reviewing the
different considerations and also the changes in the experimental plan.
3.1 Design conside rations common to both helical

flow rheometry and coaxial flow rheometry
Measurement of slurry rheology is one of the impediments to the
study of rheological effects on grinding. As explained by Shi (2016),
industrial scale milling (and the settling rates of unstable slurries)
favours the use of viscometers which can be:
• used online;
• in an industrial environment;
• and with continuous flow.
The viscometer also must be able to gain accurate readings.

Rotational viscometers work by reading torque and rotational speed
and then converting them into the more interesting shear stress and
shear rate. For a given geom etry, torque is dependent upon shear rate.
Shear rate affects the flow region; laminar flow occurs at a lower
shear rate compared to turbulent conditions . Thus, the greater the
torque readings sensitivity the lower the minimum shear rate that can
be tested. Hence, there is a minimum sensitivity which would allow
the tests to occur, else the tests couldn’t be performed under laminar
conditions.
As described earlier, the ore at Lihir ore is complicated being

mostly K-feldspar mixed with varying quantities of clay. Shi
researched ferrosilicon rheology (Shi 2016) and identified three
challenges:
1. Designing a suitable measuring cup: The material contained

coarse aspects (in his case as high as 1.4 mm). The DIN
53019-1 and ISO 3219 (1995) require a gap to particle ratio
of 1:5 but it is generally considered best practice to have a
1:10 gap (Shi, 1995).
71
2. Maintaining laminar flow: laminar flow must be maintained
to produce true flow curves. Hence, all data with a Taylor
number greater than 41.3 should be excluded (the point
where Taylor vortices begin to occur). This is explained in
more detail in section 3.1.2.
3. Constant medium density: The final challenge was to

maintain constant medium density throughout the test
despite the material being fast settling.
The points are to some extent in contradiction of each other. To

overcome the first challenge a wide gap measuring system needs to be
used. Wide gap measuring systems however transi tion to turbulent
flow at lower shear rates than narrow gap systems. Finally, to
maintain a constant fluid density for a fast settling system a flow
helical rheometer will need to be employed. Helical flow rheometers
are more likely to cause turbulence and calculating the turbulence is
more complicated than in a coaxial cylinder system (Akroyd 2004).
3.1.1 Sample stability

Rheological tests rely on sample stability. If sedi mentation
occurs the measured apparent viscosity will be affected as the fluid
becomes non-homogenous. This can cause an apparent time dependent
behaviour through two distinct mechanisms:
1. The drop in the slurry concentration causes a drop in the

concentration thus leading to lower measured viscosities;
2. A sediment bed builds which the bob can get caught on.
Whether flow is turbulent or laminar depends largely on the

location of the fluid within the grinding circuit. Under turbulent
conditions sedimentation is likely to either be prevented or reversed.
However, most rheometers require the sample to be tested under
laminar conditions. Thus, any data gained from a turbulent system will
need to be corrected. Correction methods are not ideal and will create
72
unwanted complexity in an already complicated project . To prove that
the correction method is accurate would require performing the tests
with a control.
3.1.2 Predicting if a flow curve is producible

with a given rheometer
Before purchasing a rheometer or designing and experiment it is
important to consider whether the device is capable of measuring the
fluid or suspensions rheology. This consists of two tests:
1. Estimating the torque (assuming a given viscosity)
2. Estimating if the fluid would be within the laminar regi on.
For the early purposes of commissioning equipment before

conducting experiments it was deemed reasonable to assume that the
Taylor number would be representative even with the introduction of
an additional axial flow. As such, if the torque is within the
instruments measurable range and the Taylor number is below 41.3
then it can be assumed that the flow curve will be producible.
3.1.2.1 Predicting the Torque
There are two authoritative standards on rheology; DIN53019 -1
(2008) and the ISO 3219 (1995) [28]. DIN53019-1 (2008) [29]
describes the process in which you can estimate the tor que. A problem
with both standards is that they rely on a mathematical trick; rather
than estimate the torque at the edge of the inner cylinder where it is
measured, the estimated values are taken to be at some point in the
ring gap between the two cylinders. The estimated values are
considered to be “representative”. It is generally estimated that the
representative shear is the arithmetic mean of the shear stresses at the
outer and inner cylinders.
Each rheometer will need to be calibrated by its manufac turer to

conform to the previously mentioned standards. For the Viscotester
IQ, the manual gives the shear stress as seen in equation 3 -1:
73
𝜏 = 𝐴 . 𝑀𝑑 3-1
where Md is the applied torque, τ is the shear stre ss, and A is the
g e o m e t r i c f a c t o r . T h e s h e a r r a t e ( 𝛾̇ ) c a n b e g i v e n b y :
𝛾̇ = 𝐵 . 𝛺 3-2
where B is another geometric factor and Ω is the angular speed.

Apparent viscosity is by definition:
𝜏
hence, it is possible to estimate the torque given a particular

viscosity (of a Newtonian fluid) and rpm if the geometric factors are
known. For commissioning purposes, two previously tested slurries
can be used to estimate a viscosity range for the as yet untested
slurries. The shear rate, the s hear stress and the torque can be found
through the above equations (equations 3 -1, 3-2, and 2-1). This torque
can be compared to the minimum and maximum torque supplied by the
manufacturer. In summary:
1 . S e l e c t Ω and η
2 . W o r k o u t 𝛾̇ b y 𝛾̇ = 𝑀 . Ω
3 . W o r k o u t τ b y 𝜏 = 𝜂 . 𝛾̇
𝜏
4 . W o r k o u t M d b y 𝑀𝑑 = 𝐴
5. Compare the torque with the manufacturer’s minimum and

maximum torque
This of course, relies on having an accurate estimate of the upper

and lower range of the sample to be tested.
3.1.2.2 Finding the Taylor number
The Taylor number is a dimensionless number tha t characterises
a fluid undergoing centrifugal flow. According to Shi (2016);
Schlichting defined three flow regimes:
𝑇𝑎 < 41.3 ∶ 𝑙𝑎𝑚𝑖𝑛𝑎𝑟 𝑓𝑙𝑜𝑤 3-3
74
41.3 < 𝑇𝑎 < 400 ∶ 𝑙𝑎𝑚𝑖𝑛𝑎𝑟 𝑓𝑙𝑜𝑤 𝑤𝑖𝑡ℎ 𝑇𝑎𝑦𝑙𝑜𝑟 𝑣𝑜𝑟𝑡𝑖𝑐𝑒𝑠 3-4
𝑇𝑎 > 400 ∶ 𝑡𝑢𝑟𝑏𝑢𝑙𝑒𝑛𝑡 𝑓𝑙𝑜𝑤 3-5
where Ta is the Taylor number. There are multiple forms of the

Taylor number one of which is given by:
(𝑅𝑐 −𝑅𝑏 )∗𝑈𝑏 ∗𝜌 𝑅𝑐 −𝑅𝑏

𝑇𝑎 = ∗√ 3-6
𝜂 𝑅𝑏
where RC is the radius of the cup, Rb is the radius of the bob, Ub is

the peripheral velocity of the bob, and ρ is the density. The
geometry is given, and the peripheral velocity of the bob can be
estimated in a similar method as was used for finding the shear
rate when predicting the torque. Newtonian fluids can be used to
estimate the viscosity.
3.2 Helical flow rheometer commissioning

As made clear in the introduction to this chapter, a helical flow
rheometer was commissioned. The goal of the helical flow rheometer
was to allow the testing of fast settling ore slurries.
A helical flow rheometer consists of a rotational rheometer

combined with a testing rig. Similar testing rigs have been previously
used at the JKMRC as can be seen in Figure 2-21. Such testing rigs
work by circulating the slurry. Outside the mea suring chamber the
flow is turbulent thus keeping the slurry homogeneous. Inside the
measuring chamber the flow is laminar allowing rheological analysis.
A new rig was required as
1. there was limited sample;
2. temperature control was required;
3. the selected rheometer, the HAAKE Viscotester IQ (see

section 3.3) may only be used in an airconditioned room.
The other rheometers in previous use had been conducted
within the Indooroopilly Experimental Mine Site’s Pilot
Plant.
75
As seen in Figure 3-1, a new testing rig was built for this
project. It could be placed in an air -conditioned laboratory, and the
jacketed pipe allowed slurry temperature control. Optimisation
immediately begun as seen in Figure 3-2. A new smaller sump was
designed which allowed less sample to be used in line with the limited
availability whilst also making the cooling pipe redundant.
Figure 3-1: The first iteration of the new testing rig
76
Figure 3-2: The current testing rig
3.2.1 Measuring chamber

As stated earlier, a 1:10 particle to gap ratio is generally
preferred in the literature and DIN 53019-1 and ISO 3219 (1995)
require at least a 1:5 gap. It is generally accepted that large particles
are unlikely to affect viscosity, however, it is believed a large mass
fraction of Lihir ore slurries within the grinding circuit are large
particles. It is unscientific to exclude such a large part of the sample.
The benefits of having a smaller gap are:
77
1. uses less sample;
2. Newtonian shear rate approximations ca n be used;
3. reduces the chances of turbulence effects.
Referring to Figure 3-3, it is clear that taking a maximum

particle size of 1 mm would allow most of the particle size
distribution of the SAG trommel undersize (approximately 80%) and
the ball mill discharge (approximately 90%). Likewise, all though it
would only be approximately 19% of the SAG feed, most particles
above 1 mm wouldn’t probably be a pa rt of the slurry pool. Hence, the
as provided ore (particle size <3.35 mm) is more related to the
secondary grinding circuit i.e. the ball mill.
Figure 3-3: Cumulative percent passing stated size in various

processes of Lihir's grinding circuit (Shi, 2017a)
A 1mm max particle size requires a 10 mm gap. It was decided to

build this as an integrated part of the device. If a smaller gap was
78
desired, immersion tubes could be attached to t he rheometer as seen in
Figure 3-4.
Figure 3-4:HAAKE rheometer running a calibration test.

Immersion tube was used for a narrow gap geometry.
3.2.2 Pump
The pump had to handle solid and abrasive particles, and ideally
be non-pulsed. Unfortunately, the most suitable pump available to the
project at the JKMRC pilot plant was peristaltic. It prevented
sedimentation and could handle highly viscous samples, yet it pulsed;
as such, an overflow (rather than the underflow) needed to be used as
to maintain a near constant head within the measuring chamber as
possible.
Pumping these particles on a laboratory scale is unusual and

there are not many suitable commercial ly available pumps. Akroyd
(2004) gives an excellent description of pump selection and alteration;
it is recommended reading if designing a helical flow rheometer. He
recommended a slurry centrifugal pump as it provides non-pulsed
79
flow, is erosion resistant, and can handle solid particles. He was
unable to source such a pump (at a suitable flow rate) and described
how to alter a helical rotor pump to handle solid particles. He ruled
out a peristaltic pumps as he used an underflow rheometer (see section
2.8.3) which require non-pulsed flow.
3.2.3 Sump
The sump’s purpose is to collect and hold the slurry, and to
ensure that it is well mixed. Mixer tank design is still largely an art
with rules of thumb for design. There are two major considerations in
sump design:
1. How much volume will be required for the experiments;
2. And liquid level to diameter ratio as seen i n Table 3-1.
The sump was designed through a combination of
• a mass balance;
• and assumptions that:
o there will be insignificant volume change of the ore

when adding to the water to form the slurry i.e. no
swelling or dissolving. The particles will become
suspended;
o specific gravity of ore is eq ual to 2.7;
o the restriction that the upper volume fraction would be

0.80
80
Table 3-1 : Table of liquid level to tank diameter ratios (James,
2015)
Minimum ideal Okay Max
Liquid 0.6 0.8 1 1.4

level to
tank
diameter
ratio
A final point, to ensure the best mixing the impeller is placed off
centre which reduces the vortex (Couper, 2010). Considering the small
sump size, and that the intention is to agitate slurry baffles were
believed to be inappropriate.
3.2.4 Testing rig operation

The testing rig proved to be difficult to operate. Loading often
resulted in some sample overflowing or pump blockages. Losing
sample made the mass balance calculations difficult and delayed the
experiments. Pump blockages often could only be resolved with the
catastrophic loss of a large amount of sample and a couple of days of
work.
Lihir ore slurries have highly variable rheology as described in

later sections. An under appreciation of this fact led to a large amount
of time being wasted in operating the testing rig. Only a couple of
sample types were used for proof of concept exper imentation; and
loading was therefore optimised given these samples rheological
characteristics at their particular concentration level. This was
reasonable due to the constraint of the project however it led to the
aforementioned failures. The same loadin g solids concentration will
not work with different slurry samp les. The maximum slurry
concentration that the testing rig can handle without overflowing
varies greatly between different domains. Likewise, the sedimentation
81
rate is vastly different allowing a lower or higher concentration to be
immediately added.
3.3 Rheometer selection

3.3.1 Selection of the HAAKE Viscometer
The HAAKE Viscotester IQ was selected as the most suitable
rheometer. It is easily transported and is capable of measuring both
large and very small torque readings. Furthermore, it was also highly
price competitive.
The HAAKE Viscotester IQ rheometer is both highly accurate and

versatile. Unfortunately, the rheometer we had acquired
catastrophically failed during the early stages of the experimental
program. Time was spent trying to fix the device through
communicating with the supplier and factory. However, it could not be
fixed at the university and needed to be returned to the supplier. The
supplier was unable to repair the rheometer and hence organised a
replacement.
3.3.2 Alternative rheometers

The cause of the device failure was not clear and so alternative
experimental plans were investigated. Several alternative rheometers
located at the University of Queensland were identified, however, the
most appropriate rheometers (when considering technical
specifications, access issues, and sample transportation) were both
located at the JKMRC, an Anton Paar DSR301 and a Mettler RM180.
The Anton Paar is a high quality air bearing rheometer, which is

able to measure 10 times lower shear stresses than the HAA KE (Anon,
2006). However, it is fixed to the wall and would not be appropriate to
use with the testing rig. The RM180, like the HAAKE, is a portable
rheometer. However, being mechanical rather than air bearing, it
limits the detail of the sample at the lower she ar stress ranges.
82
At high concentrations the sample became effectively non -
settling; using co-axial cylinder rheology would thus be quicker,
simpler, and use less sample. On the other hand, it would not allow
the generation of data for the fast settling s lurries. On balance, it was
decided to use the Anton Paar.
3.3.3 Anton Paar

The available cups that the JKMRC had available for the Anton
Paar were not suitable given the large particles. A new cup was
commissioned and used for some proof of concept experiments. The
new cup had the following geometry:
• 25 mm internal radius;
• A wider radius outside of the measuring chamber to help

minimise overloading;
• elongated, being 15 mm long. The elongation was inspired

by Klein et al (1995) (see section 2.8) who used elongated
double gap coaxial cup rheometers for investigating
mineral slurries.
The Anton Paar was working very well; then, qui te unfortunately
the computer with the rheology software failed. The IT department
advised that its failure was purel y caused by age. The Anton Paar is
second hand and came with the now broken computer. After
investigating and contacting the suppliers, it was determined that
JKMRC was never supplied with the software installation separate to
the computer; and that it was only licenced for a very old version of
the software. The project funding was inadequate to buy an upgrade to
the software, however, the supplier managed through internal contacts
to track down a copy in Germany. The day after a new computer was
set up the power supplier controlling the rheometer failed and there
were no funds available for repairs.
83
No data from the Anton Paar experiments w as recovered from the
computer. The experiments performed were purely proof of concept
and to familiarise myself with the equipment. However, they
demonstrated that the co-axial rheology experiments could be
conducted on the high concentration samples and that it was quicker
and easier than using the testing rig. Thus, when developing the
experimental program for the RM1 80, a choice was presented between
using the testing rig or using the cup developed for the Anton Paar.
This will be discussed in more detail in the next section.
3.4 RM180 co-axial cup commissioning

Two issues were immediately present when beginning work on t he
RM180. Firstly, as described in the previous section, there was the
choice between using the testing rig or using the commissioned cup.
The other issue was that the RM180 has a much smaller range of
possible torque readings than either the Anton Paar or the HAAKE.
The choice between using the cup or testing rig was resolved
through a simple proof of concept experiment. A sample’s rheology
was measured at a series of concentrations proceeding from a high
concentration to lower concentrations. The high co ncentrations were
slow settling and the rheology was easy to measure. Once
sedimentation became an issue the testing rig was used; the one
significant alteration from previous experiments being that the RM180
was used in place of the HAAKE.
All of the data was either too low for the torque to be measured
or the mixing was turbulent; sometimes both. Hence, the experimental
plan described in section Chapter 4.
The second immediate issue was to deal with the much smaller
range of possible torque readings. A larger bob allows a smaller shear
stress to be measured for the same torque as seen by
𝑀
𝜏1 = 2𝜋𝑟 2 𝐿 2-11
1
84
Hence, by using two different bobs a larger range of materials
could be measured. A larger bob where increased sensitivity was
required (less viscous, less solid samples), and a smaller bob where
fluid instability or exceeding the maximum torque were possible
issues (such as in more viscous or solid materials).
3.4.1 Time dependency

From the beginning of the experiments time dependency was
investigated using loop tests (see section 2.2.2). Although time
dependency was not an issue when the testing rig was used, it was
noticed when using a cup and bob system. The early proof of concept
experiments using both the Anton Paar and the RM180 indicated some
time dependent shear thickening. Apparent time dependent shear
thickening isn’t necessarily due to the aging of the material, it can
also appear due to:
• Sedimentation: As the material settles the apparent

viscosity will normally drop. However, a thickened
sediment may form, in which cas e, the bob may give
artificially high readings.
• Drying: As the slurry dehydrates the solid s concentration

increases. Higher solids concentrations are more viscous
(see section 2.4.1).
• Statistical variation
• Chemical reaction: The chemistry of the sample may be

unstable. What may appear to be time dependent shear
thickening may actually be the chemistry changing with
time. The sample would have become thicker even without
the shearing from the bob.
• Instrument effects: False readings are known to occur if

ramping occurs too suddenly.
85
Before determining whether shear thickening occurs,
sedimentation, drying, experimental error, chemical instability and
instrument effects must be ruled out. There were samples tha t
appeared to have characteristic time dependent shear thickening in
which sedimentation could be ruled out. Furthermore, the pH was
stable, and the error bars between the two curves didn’t overlap.
Drying however, was a possible issue in some of these ear ly proof of
concept experiments. After the proof of concepts were completed, the
drying issue was addressed by ensuring the samples were tightly
covered with foil until required. However, before starting on the
experimental program, the Anton Paar DSR301 b ecame unavailable as
described above.
The early RM180 proof of concept experiments didn’t indicate

(outside of standard deviation) time dependent shear thickening.
Hence, the experimental procedure was modelled on procedure used
with the HAAKE and testing rig:
1. A slurry sample was prepared at 75% solid s concentration

(wt%) in a beaker.
2. Rheological measurements were performed on a subsample

taken from the beaker.
3. The subsample after the rheological measurement was added

back into the beaker.
4. Some sample was taken for sedimentation tests.
5. Water was added to reduce the concentration.
6. Repeat steps 2 to 5 until the slurry reaches 50% solid s

concentration.
No control was placed on how long the sample was being mixed.
Several different samples were selected including ar gillic, outer
biotite and distil chlorite to understand how different domains may
behave differently. Sample NTS045 (an outer biotite sampl e) indicated
86
statistically significant levels of time dependent shear thickening. The
ramping procedure described in section 4.8.4 was used eliminating the
possibility of instrument artefacts. The pH was stable, the sample was
slow settling and the slurry density underwent negligible change. This
left no possibility but time dependent shear thickening. The longer the
sample was mixed, the thicker it became. Hence, mixing times were
made consistent. It was expected that the container’s size, shape, and
the volume of sample could also affect the thickness. As such, these
were also standardised as described in section 4.8.1. The previous
results prior to these chambers were also treated as proof of concept.
The aforementioned solid concentrations (75% through 50%

(wt%)) were selected as they are similar to the concentrations used
inside the SAG (70% to 75%) and the ball mill (70%) (Shi, 2017b).
3.5 Summary of design considerations

This project used rotational rheometers. Two designs of
rotational rheometers were used in this project: a c oaxial cylinder
rheometer and a helical flow rheometer (a modified coaxial cylinder
rheometer with axial flow).
The helical flow rheometer was commissioned because it was

assumed at the beginning of the project that the slurry would be fast
settling. This assumption was not tested until after the helical flow
rheometer was commissioned due to sample limitations. Low
concentration slurries are fast settling as expected but at high
concentrations the suspensions become non or very slowly settling.
The key design considerations in creating the helical flow

rheometer were:
• designing a suitable measuring chamber;
• selecting a suitable pump;
• deciding on a flow direction;
• designing a suitable sump;
87
• and experimenting with operating the testing rig.
A combination of catastrophic equipment failures, sample

limitations and the discovery that not all the slurry was fast settling
led to a shift in the experimental plan. Instead of using a helica l flow
rheometer, the tests would be performed using a mechanical bearing
RM180 coaxial cylinder rheometer. The experimental procedure of the
helical flow rheometer and the coaxial cylinder rheometer are
explained in more detail in sections 4.7 and 4.8 respectively. This had
the advantage of allowing the generation of key results presented in
chapters 5 through 7, however, it prevented the measurement of the
rheology of low concentration samples.
A wide gap geometry with an elongated cup was used in both the
rheometer designs. Calculations for the flow curves within a wide gap
coaxial cylinder geometry were presented in the literature review (see
section 2.7). Additional information on how it was applied to this
project is provided in section 4.9. The important considerations are:
• accurate volume measurement;
• accurate placement of the bob in the middle of the vessel;
• the gap is kept wide enough to prevent jamming;
• an appropriate fluid model is selected for data analysis;
• minimum sedimentation occurs;
• ensuring data is within the laminar region.
88
Chapter 4
Experimental methodology
This section provides an overview of the final methodology for
both the flow rheometry conducted with the HAAKE Viscotester IQ
and also the concentric cylinder rheometry with the Mettler RM180.
Both methods had much in common and discussing them in separate
chapters would be highly repetitive. As discussed in section 3.3, a
limited number of tests were able to be performed due to the failure of
the HAAKE rheometer before completing the testing program.
Given the difficulty of analysing flow rheology data, the small

quantity of samples and the issues with time dependency the results
were treated as qualitative. The method is presented here to give
future researchers an understanding of the process of conducting flow
rheology tests.
4.1 Data entry

All samples provided by Newcrest had been given an identif ier of
“NTS###”. This identifier will be used for the remainder of the
discussion of the thesis. In addition to the NTS description, this thesis
will refer to the tests using the RM180 as being SEQ### where the
number refers to a unique rheology program. A ppendix A contains a
list linking the NTS information alongside the concentration of the
slurry and the SEQ###.
4.2 Materials
Lihir ore was provided by Newcrest. The received sample was
delivered in two-kilogram bags previously crushed and sieved to less
than 3.35 mm. All ore had been previously stored by Newcrest in
89
Brisbane before being transferred to this project. No site work was
conducted, and the samples were of varying ages.
The water used in the experiments was identified as a possible

variable in the slurries’ rheology. Both Lihir and Brisbane tap water
would likely contain impurities which could affect the chemistry.
Since it is not practical to use the water collected from Lihir
operation circuit, three alternative options were available:
1. synthetic plant water;
2. de-ionised water;
3. Brisbane tap water.
De-ionised water was selected as the best option for the

following reasons:
1. it is impossible to say if synthetic plant water is more

representative of Lihir water (for rheology purposes) than
DI water without extensive testing;
2. producing the synthetic plant water would be time

consuming and expensive;
3. de-ionised water is readily available within the laboratory;
4.3 Particle sizing

A split was taken from the samples used in the rheological
experiments to allow sizing.
Particles were sized using two methods:
1. Sieving;
2. Malvern MasterSizer.
Sieving operates by separating particles by size. All material was

wet sieved to 38 μm. Both the plus and minus 38 μm particles were
then filter pressed and dried.
90
The dried plus 38 μm particles were sized via the JKMRC
rotatory screens (“Rotap”) following an internal protocol modelled on
AS 3638 – 1993. In the dry sieving procedure, material is placed on a
stack of seven screens which are then vibrated for 10 minutes. As
more than seven size distributions were required, the undersized was
collected and sieved for a further 10 minutes.
The dried minus 38 μm particles were sized through further wet

sieving or the use of a Malvern MasterSizer. The results of the sizing
can be found in Appendix D and section 5.3. As discussed in section
2.9.3 both laser diffraction and sievi ng are spherical sizing
techniques. In general, the most accurate method for sizing non -
spherical particles is microscopic image analysis as suc h techniques
are able to distinguish particles not just by diameter but also shape
(Otunniyi, 2013; Trottier et al., 2010). However, the Malvern
MasterSizing was determined to be appropriate for this stud y: it has
been found that laser diffraction gives similar results to microscopic
image analysis (Kursun, 2009); it is easy and quick to perform, the
results (due to the consist methodology) are comparable, and any
additional data would have been beyond the scope of this study.
4.4 Sample preparation

As described earlier, the samples were received in two -kilogram
bags previously crushed to less than 3.35 mm. The RM180 tests
required approximately 4 kg whilst the flow rheometry tests required 5
kg (including addition for sizing; refer to section 4.3). Hence, the
bags were combined and sieved to less than one millimetre.
All rheology and sedimentation tests were considered destructive

due to the potential for ageing effects. Furthermore, the volume of the
sample being mixed was kept as consistent as possible; for the final
methodology within the RM180 tests it was decided to target 600 mL
as this allowed sufficient sample for both the rheology and
sedimentation tests. Hence, each test required its own particular
91
quantity of ore based off the solids concentration of the test as seen in
section 4.4.1.
For the flow rheology tests, exact volume of slurry being added
to the testing rig couldn’t be full y controlled due to the difficulties
with loading (see section 3.2.4 for more details).
4.4.1 Splitting
The sample was split using a combination of rotatory splitters,
riffle splitters, and spooning. On consultation with the advisory team
the rotatory splitters were preferred as they are les s likely to adjust
the sample size distribution. Riffle splitters were used in conjunction
with the rotatory splitters, to approach the approximate mass for a
concentration (see Table 4-1).
Table 4-1: Mass targets when splitting
Target solids concentration Mass (grams)

(wt%)
75% 844.25
70% 744.48
65% 655.15
60% 574.70
55% 501.87
Na (split taken for 200 to 300

accumulative size fraction
measurement)
Since a precise volume was targeted and considering the riffle

splitter and rotatory splitters were only capable of dividing the sample
into halves and eighths respectively; a table spoon was used to get the
sample to the correct weight if the difference b etween the mass and
the target mass was less than twenty grams. All sample was weighed
using analytical scales.
92
4.4.2 Grinding procedure
Particle size and particle distribution are both critical factors
that affect slurry rheology and sedimentation rate (see se ction 2.4.2).
To allow the results to be reproducible, the standard JKTech bond ball
procedure was followed.
Six NTS materials were ground and subsequently h ad their

rheology tested on the same day. They are:
• NTS044 (outer biotite)
• NTS054 (outer biotite)
• NTS097 (inner biotite)
• NTS108 (argillic)
• NTS110 (lower epithermal)
• NTS111 (upper epithermal)
All samples were ground for 5 minutes. A split for sizing was
taken.
These samples represent 5 of the 9 domains at L ihir. The domains

not included either were not of interest to Newcrest or were provided
in insufficient quantity for these tests.
A calibration curve had been prepared by Mr. Ncube (see Figure

4-1) which showed that the Lihir ore significantly changes its size
distribution after 5 minutes of grinding. The material used for
creating this calibration curve was a distal chlorite sample from Lihir
which was not required for the rheology tests. The goal of grinding
the material was to see how the rheology changed after grinding.
Future work should target a particular grind size for the rheology
tests.
93
Figure 4-1: Particle size distribution curves of NTS050 after
grinding for a differential amount of time.
4.5 Sedimentation tests

As mentioned in section 2.12, sedimentation is linked to
rheology in three important ways. Firstly, i f sedimentation occurs,
most rheological tests cannot be performed. Secondly, sedimentation
does not occur in materials with sufficient yield stress (Ovarlez et al.,
2012). Thirdly, a higher viscosity results in a lower particles
sedimentation rate. Despite these three links, little research has been
performed in the minerals area on the link between sedimentation and
rheological properties.
Sedimentation tests were performed using clear glass measuring

tubes (250 mL) and a high definition video camera. For the flow
rheology, the material was taken from the underflow of the measuring
chamber. If necessary, the cylinders were inverted to ensure
homogeneity.
94
For the final coaxial cylinder rheology, the slurry was taken
from the measuring beaker, and poured into the measuring cylinder. A
target of 200 mL was chosen to make results easier to compare. Not a ll
sub-samples had a sedimentation test performed as some were clearly
pastes and it would have been impossible to accurately fill the
measuring cylinder.
4.6 Measuring pH, dissolved oxygen, and

temperature
The pH and the dissolved oxygen were measured via la boratory
pH and oxygen meter. Temperature was measured via temperature
probes attached to the rheometers. Neither pH nor the dissolved
oxygen was altered; they were only monitored. Temperature within the
helical flow rheometer was controlled through a cool ing jacket
surrounding the sump.
The temperature measuring probe was an integrated part of the

HAAKE viscometer. The probe was placed outside of the measuring
chamber, thus, preventing an interference with the rheology
measurement. Unfortunately, the tem perature probe was attached to the
RM180 in such a way that it couldn’t be removed from the measuring
chamber. Although it wasn’t observed to have an effect it may have
led to less accurate results.
4.7 HAAKE experimental procedure

The experiments consisted of several steps:
1. mixing the slurry and loading the slurry into the testing rig
(4.7.1);
2. performing the rheology test;
3. measuring the flow rate (4.7.2) and removing some sample

for further analysis:
a. sedimentation and slurry density tests (section 4.5)
b. pH and oxygen tests (section 4.6)
95
4. adjusting the slurries concentration through r emoving
slurry and adding water (section 4.7.4;
5. steps 2 through 5 were repeated for each concentration;
6. and cleaning the testing rig and rheometer.
4.7.1 Premixing the slurry and loading the

helical flow rheometer
It was identified that the mixing impeller, although strong
enough to stir already prepared slurry, wasn’t strong enough to mix
the dry sample with water. An alternative mixing impeller was
employed; however, it was too large to fit within the sump. Thus, the
slurry had to be at least partially prepared outside of the testing rig.
One technique was to prepare the sample outside of the testing

rig at the desired concentration. This sometimes worked; however, it
was inconsistent due to the varying viscosity level s between samples;
adding too viscous of a sample could cause the pump to fail from
being overloaded. Another potential problem was the viscosity was too
great to allow the slurry to discharg e at a sufficient rate resulting in
an overflow from the top of t he measuring chamber.
The other technique was to partially prepare the slurry below the
desired concentration, begin pumping and add sample to the testing
rig. This was highly time consuming, increased the chances of spills,
and could cause the pump to bl ock as a lower concentration results in
a higher sedimentation rate.
It was hypothesised that starting the pump with flowing water,

and gradually adding slurry could prevent the pump from blo cking or
the measuring chamber from overflowing. This was never t ested
primarily due to sample restrictions. There was insufficient sample;
the concentrations we were targeting required us to greatly minimise
the water within the system. A higher quantity of water would require
more sample.
96
In hindsight, the difficulty with loading the helical flow
rheometer was due to the very large differences in the slurries
rheologies between different NTS numbers. It is recommended that
before future flow rheometry is conducted that a coaxial cylinder
rheology and sedimentation pro gram be conducted. From those results,
the lowest viscosity slow settling sample concentration should be
selected to be added to the testing rig. The testing pump can then be
started with some water already being inside to prevent it from
overloading with the addition of the slurry.
4.7.2 Rheology test

The slurry was ramped from a slow speed to a high speed. Three
loops were conducted per sample; the loop tests give a qualitative
indication of time dependency. A continuous ramp was used for these
tests rather than a step ramp as described in section 4.8.4 as an
oversight. For future work it is recommended that step ramps be
conducted as a continuous ramp can introduce inertia effects.
4.7.3 Flow rate

Flow rate through the measuring chamber was measured using a
stop watch and filling a measuring vessel. As the sample size of the
experiments was limited (due to a combination of limited sample and
the time-consuming nature of sample preparation) only a small
quantity of sample could be removed from the testing rig at a time
which made measuring the flow rate highly error prone. Ideally, it
would have been measured using a flow meter, but no suitable flow
meters were available for this project.
The flow rate was controlled through adjusting the pump speed
and opening and adjusting the size of the valve openings.
4.7.4 Adjusting sample concentration

Solids concentration has a large influence on the rheology of
slurries and pastes (see section 2.3). One of the ultimate aims of this
97
project was optimising Lihir’s grinding circuit. Grinding circuits are
volume constrained; hence, increasing the solid s concentration should
theoretically increase the throughput at the same volume flow rates.
However, an increase in viscosity or yield stress would reduce the
grinding efficiency and thus reduce the throug hput.
Given the limitations of the project (sample, time, etc), the

slurry wasn’t replaced when investigating the different concentrations
but rather had its solids concentration varied through removing slurry
and adding water. The original concentration was the highest
concentration to be investigated, each additional concentration was
lower than the previous. The removed slurry was used for
sedimentation experiments (section 4.5), the measuring of the pH and
dissolved oxygen levels (section 4.6), and the slurry density. From the
slurry density the solids concentration could be estimated by assuming
a specific gravity of the ore. This approach is somewhat limited by the
specific gravity assumption; Lihir ore contains clays that swell. The
sample removed for sedimentation purposes was filter pressed and
weighed allowing a more precise concentration to be determined after
completing the experiment. However, this was also limited as
sometimes the same sample would give different specific gravities for
different concentrations; the different s pecific gravities didn’t appear
to follow a trend and could increase and decrease from reducing the
slurry concentration.
Lihir SAG operations operate at solid concentrations ranging

from 70-75% (wt%), and ball mill at around 70% (wt%) which
corresponds to 46% to 53% (v/v) (Shi, 2017b).
4.8 RM180 experimental procedure

The experiments consisted of mixing sample, loading it into the
cup, performing the rheology and sedimentatio n tests, and cleaning the
equipment. When sample wasn’t being used, it was covered in
aluminium foil to prevent drying.
98
4.8.1 Mixing the sample
Time dependency was identified as an issue. As such mixing was
always conducted in a 1 L plastic beaker for 4 minutes using a drill
with an attached paint stirrer. The procedure operated as follows:
1. Weigh and zero the beaker
2. Add the ore to the beaker in a fume cabinet (to avoid the
dust)
3. Weigh the ore and calculate the required water to reach the
concentration.
4. Add the de-ionised water using a combination of pouring

and a pipette
5. Record the weight of the water
6. Mix the sample for 4 minutes using a phone as a timer
7. Clean the drill bit
8. Record the mass (using analytic scales) and volume of the

beaker (the beaker has volume m arkings) containing the
sample. Sometimes the sample could adhere to the sides of
the beaker; in such cases a spatula was used to scrape the
sample in level with the rest of the slurry or paste.
Step 8 was only used if densities couldn’t be calculated by other

means such as through the sedimentation tests.
4.8.2 Loading and unloading the rheometer

Sample was carefully loaded into the cup. Depending on the
sample it could be pourable or a thick paste. A spatula was used for
the thicker samples. The slurry was fil led to the desired height; the
cup was lightly tapped to ensure a level surface and eliminate the
chances of air bubbles.
99
The bob was lowered into the sample. If required, the cup was
lightly tapped again to re-level the surface disturbed by the bob’s
entry. The sample was allowed to rest for 30 s.
The bob and cup were always cleaned and dried when changing
the concentration or sample. An air-compressor using plant air was
used to ensure that both were completely dry.
In case of sedimentation within the m ixing beaker, the slurry was

lightly stirred by hand using a spatula.
4.8.3 Selecting a bob

Two different bobs were used with the RM180, which, hence
forth will be referred to as the “large bob” and the “small bob”. Both
were made at the JKMRC, and the characte ristics were modelled on the
previously manufacturer supplied bobs but were grooved instead of
smooth. They are shown in Figure 4-2.
100
Figure 4-2:Small bob and large bob used in RM180 rheometer.
The small bob was used for more viscous slurries whilst the large
bob was used for the less viscous slurries. If the large bob was used
on the more viscous slurries, it would overload; if the small bob was
used on the less viscous slurries it would generate too small torque
readings to generate useful data. They had distinct geometries; the
large bob had a large conical section on its upper side to prevent
sediment being caught whilst the small bob kept the classical structure
described in DIN53019-1. The parameters of the bobs are described in
Table 4-2.
101
Table 4-2: Bob geometries
Bob
Small Large
Radius 7 mm 15 mm
Height (excluding 21 mm 45 mm
conical sections)
Radial gap1 18 mm 10 mm
4.8.4 Yield stress tests

Yield stress may be found either by extrapolation of the
rheological data or via direct measurement. Extrapolation obviously
relies on assuming a flow model; it has been found that assuming the
wrong model can lead to significant error (Boger, 2009; Nguyen and
Boger, 1992). The best method for direct measurement is the direct
yield stress measurement using a vane geometry as described i n the
literature review. Unfortunately, a vane wasn’t available for the
RM180; hence, bobs were used instead. Using vanes are superior to
cylinders as they are less likely to disturb the sample when entering
the slurry.
The key considerations when measurin g yield stress are:
• not to disturb the fluid when entering the slurry;
• that the experiment is carried out under shear stress

controlled or shear rate-controlled conditions;
• that the cylinder/vane moves as slowly as possible.
The maximum measured torque occurs at the yield stress. The

fluid is assumed to move between the blades of the vane as a solid
1 Using the cup described in section 3.4; the cup has a radius of 25 mm and its measuring chamber is 15
mm long
102
allowing the equations for concentric cylinders to be used for the
calculation of the yield stress.
The yield stress directly measured through this technique i s

recorded in Appendix G but is not used in the analysis. Instead, best
fits of the yield stress were produced as described in section 4.9.
4.8.5 Rheology test: Ramping procedure

The standard test consists of an rpm “step ramp”. A step ramp is
where the shear rate or the shear stress is kept constant and then
suddenly changed (“ramped”) to either a higher level (“ramp up”) or
to a lower level (“ramp down”). As a wide gap geometry was being
used, the Newtonian shear rate approximatio n would be highly
inaccurate (see section 2.6.3). The RM180 is a strain-controlled
device allowing only the rpm, not the shear stress, to be controlled.
To prevent instrument artefacts, each step was held for at least ten
seconds.
The different sub-samples didn’t receive the same ramp profile.

As some samples are highly viscous, ramping too high would cause the
rheometer to overload and fai l. If a sample was less viscous, and at
risk of sedimentation, the lower rpm’ s may be skipped, or the loop
performed in the reverse order. The ramp profile for each sample is
presented in Appendix G.
4.8.6 Benefits and limitations

The RM180 coaxial cylinder method was quicker and easier to
perform than the HAAKE with the testing rig. Th e limitation of the
RM180 was that it could only measure highly concentrated samples
whilst the testing rig allowed the measurement of dilute samples. The
RM180 also had less range than the HAAKE rheometer, making it
necessary to use two bobs. A clear adva ntage of the coaxial cylinder
geometry over the helical flow rheometer is the much simpler data
reduction procedure.
103
4.9 Data reduction
As discussed in the previous sections, the experimental tests for
the helical flow rheometer and the coaxial cylinder were l argely
similar. Essentially, the helical flow rheometer is a coaxial cylinder
rheometer with axial flow. The solution to axial flow has been solved
for the Newtonian solution but it’s still an active area of research for
the helical flow rheometer (Alharbi, 2016). As described in the
introduction of this chapter the results for the flow r heology were
treated qualitatively; they were only evaluated ignoring the axial
component.
4.9.1 Coaxial cylinder

A worked example to the data reduction procedure is provided in
Appendix E.
The following data was collected from the rheometer: the

temperature of the sample, the rotations per minute of the bob and the
torque required to rotate the slurry. The manufacturer advises that the
RM180 can measure torques between 0.25 and 7.5 mN.m; torque
readings from outside of this range were excluded from the data
analysis for being unreliable. Shear stress is directly proportional to
the torque and the dimensions of the bob whilst shear rate is
dependent upon the angular speed, the rheometer’s geometry and the
fluid characteristics. As stated earlier in the thesis th e shear stress
equation can, theoretically, be directly found from the experimental
data:
𝑀
𝜏1 = 2-11
2𝜋𝑟12 𝐿
However, in practice the instrument often needs to be calibrated

due to both experimental effects (such as end and wall effects) and to
simply ensure that the rheometer is of suitable quality. If calibrated
against a Newtonian oil, and the calibration reports a linear
correlation, the bob’s shear stress may be calibrated by equation 4 -1
104
where F is the correction factor. The calibration results for the RM180
are described in Table 5-1 found on page 119.
𝑀
𝜏1 = 2𝜋𝑟 2 𝐿𝐹 4-1
1
In a perfectly calibrated rheometer, the correction factor is

equivalent to taking the “effective height” of the bob. The effective
height of the bob is the height once end effects are considered. There
are numerous methods to estimate these effects, howev er, as they are
captured by the correction factor they weren’t pursued. The advantage
of the correction factor approach is that it combines the rheometer
calibration and instrument effects (wall effects, end effects, etc.) into
a single step. This was cons idered sufficient given the research
constraints.
The shear rate, unlike the shear stress, depends on the fluid
model. Assuming the wrong fluid model results in error; the error
scales with a wider gap. As stated in the literature review there are
three popular methods for approximately the shear rate:
1. the Newtonian assumption
2. the Krieger
3. the Estelle
As explained below, the Estelle method was selected as the most

appropriate. However, an alternative procedure was developed and is
referred to in section 4.9.4.
4.9.2 Modified Estelle method

As discussed in the literature review the Newtonian assumption
leads to large errors in wide gap geometries. Krieger showed that the
shear rate can be solved analytically if the bob is effectively in an
infinite sea:
𝑑Ω
𝛾̇ = 2𝑀 (𝑑𝑀) , 𝑖𝑓 τ2 < τ0 2-20
105
Krieger (and Elrod) then went on to produce an approximate
solution to the shear rate equation:
Ω (𝑚.𝑙𝑛𝜀)2 (𝑙𝑛𝜀)2 𝑑𝑚
𝛾̇ = 𝑙𝑛𝜀 . (1 + 𝑚. 𝑙𝑛𝛼 + + . 𝑑𝑙𝑛(𝜏)) 2-18
3 3
Estellé et al. (2008) showed that the Bingham approximation

gives excellent results and presented formulas which required no prior
assumptions about the fluid characteristics; only the rheometer’s
dimensions and the raw readings are required.
𝑑Ω
𝛾̇ = 2𝑀 (𝑑𝑀) , 𝑖𝑓 τ2 < τ0 2-20
𝑑Ω 𝑑Ω
𝑀( ) Ω−𝑀( )
𝑑𝑀 𝑑𝑀
𝛾̇ = 2 𝑟 2 − 𝑟1 , 𝑖𝑓 τ2 > τ0 2-21
1−( 12 ) ln( )
𝑟2
𝑟2
The shear rate is then assumed to be w hichever is greater

between the partially and fully sheared conditions. Estelle et al
reported that this equation produced less error than Krieger’s. An
additional benefit is that it only required finding the first der ivative
of raw data rather than the Krieger method which typically requires
two.
Estellé et al. (2008) noise reduction technique was not followed;

instead (as described above) the rheology was measured as a step ramp
(see section 4.8.4). Each step was held for at least ten seconds and ten
readings were taken. These readings were then averaged, thus
minimising both the noise and instrument effects.
𝑑Ω
Finding proved to be quite tedious and time consuming, as
𝑑𝑀
fitting a polynomial equation with all the data produced implausible

𝑑Ω
results. An example is seen in Figure 4-3 and Table 4-3, at rpm = 5 𝑑𝑀
is apparently negative for both the quad ratic and cubic fit, hence,
producing a negative shear rate and thus a negative apparent viscosity.
This is clearly implausible; it could be avoided by splitting the graph
into multiple sections, however, no general procedure for this
106
technique could be found and could result in the data processing
becoming subjective.
Angular momentum vs Torque (SEQ233)

60
50
W = 5E+06M2 - 25072M + 32.566
40 R² = 0.9993
W (s-1)
30
20
10
0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007
M (N.m)
Figure 4-3: Angular speed vs torque for SEQ233.
107
Table 4-3: Calculations of shear rate us ing Estelle et al method
and fitting the angular speed to torque through the use of a quadratic.
Shear rate
dW/dM Partial Fully Max
-1178.7 -5.7 -2.3 -2.3
1170.6 6.2 5.7 6.2
4752.3 28.8 20.9 28.8
6812.3 44.2 31.9 44.2
8425.7 57.4 41.7 57.4
9613.9 67.8 49.9 67.8
10714.1 78.0 57.8 78.0
11596.1 86.5 64.8 86.5
12805.9 98.7 74.0 98.7
13559.6 106.6 80.7 106.6
14335.0 114.9 87.6 114.9
17304.9 149.2 117.8 149.2
19904.1 182.1 147.3 182.1
22243.8 214.1 176.2 214.1
24274.2 243.7 203.8 243.7
26389.0 276.3 233.0 276.3
28412.8 309.2 262.5 309.2
30272.3 340.9 291.2 340.9
32193.6 375.1 321.4 375.1
Estellé et al. (2008)report averaging the derivative over 5 data

points. Nguyen and Boger (1983) recommends seven when reporting on
108
a similar technique. Both techniques narrow the data rang e; it isn’t
obvious either why either would necessarily produce a smooth curve.
Higher order polynomials could also be used. However, there was

a concern that this may have led to the overfitting of the curves.
Using SEQ233 as an example, to produce reliab le results throughout
the flow curve required a 6th order polynomial. There was no a priori
reason why a 6th order polynomial would truly fit the data well and
not be merely the result of overfitting.
The coefficients for the fits of SEQ233 for different orders of

dΩ
Ω(M) are given in Table 4-4. Table 4-5 and Table 4-6 show the and
dM
𝛾̇ r e s p e c t i v e l y . A t Ω = 0 . 5 2 t h e s h e a r r a t e s f o r t h e 2 n d a n d 3 r d o r d e r f i t s
are negative; hence, the fitting procedure was clearly inappropriate.
dΩ
For the 4th and 5th polynomials is overestimated at Ω=0.52 as the
dM
values are greater than the val ues for Ω=1.05 which is clearly
infeasible; this results in an overestim ate of the shear rate at Ω=0.52.
Both the 6th and 7th polynomials provide similar results and the Ω v M
curves seemed reasonable.
109
Table 4-4: Table of coefficients of Ω(M) for different orders of
polynomial fits for SEQ233.
Coefficients
Order a7 a6 a5 a4 a3 a2 a1 a0
2nd - - - - - 5185976 -27070.9 36.10419
3rd - - - - 25679982 4884793 -25933.4 34.72377
4th - - - -5.4E+11 8.55E+09 -4.5E+07 99272.69 -80.428
5th - - 7.77E+13 -2.1E+12 2.06E+10 -9.1E+07 185715 -143.855
6th - 1.16E+17 -2.8E+15 2.72E+13 -1.4E+11 3.7E+08 -525830 304.8464
7th 1.71E+19 -3.8E+17 3.23E+15 -1.3E+13 2.34E+10 -1176249 -52065.5 49.98781
Table 4-5: Comparisons of the derivatives of angular speed with

dΩ
r e s p e c t t o t o r q u e ( dM) b e t w e e n t h e d i f f e r e n t o r d e r s o f a n g u l a r s p e e d
(Ω(M)) for SEQ233.
Order of W(M)
W M t1 2nd 3rd 4th 5th 6th 7th
dW/dM
0.52 0.002431 30.2 -1860.8 -1732.2 2455.9 3174.7 1118.5 1432.7
1.05 0.00267 33.2 618.0 696.5 2325.0 2424.3 2428.0 2369.1
2.09 0.003034 37.8 4397.2 4416.3 3928.7 3760.4 4181.6 4158.2
3.14 0.003244 40.4 6570.7 6564.9 5627.7 5503.8 5640.5 5670.1
4.19 0.003408 42.4 8273.0 8252.5 7257.3 7206.8 7118.0 7159.4
5.24 0.003529 43.9 9526.8 9498.0 8588.7 8596.1 8390.5 8424.2
6.28 0.00364 45.3 10687.6 10653.3 9898.5 9954.1 9691.5 9709.7
7.33 0.00373 46.4 11618.2 11580.7 10989.6 11076.3 10805.2 10808.2
8.38 0.003853 47.9 12894.7 12855.0 12527.3 12642.2 12409.0 12391.7
9.42 0.00393 48.9 13690.0 13650.1 13499.8 13622.5 13437.1 13409.6
10.47 0.004009 49.9 14508.1 14469.0 14504.6 14626.8 14505.1 14470.2
15.71 0.004311 53.6 17641.7 17614.2 18280.3 18326.0 18487.6 18462.5
20.94 0.004575 56.9 20384.2 20378.4 21282.3 21198.1 21482.4 21502.5
26.18 0.004813 59.9 22852.9 22875.9 23536.8 23352.6 23526.2 23576.5
31.42 0.00502 62.5 24995.2 25050.3 25009.5 24822.5 24740.9 24777.9
36.65 0.005235 65.1 27226.5 27322.0 25937.1 25916.9 25546.7 25524.2
41.89 0.005441 67.7 29361.9 29502.7 26128.7 26527.5 26124.9 26044.5
47.12 0.00563 70.1 - - 25606.8 26709.9 26850.9 26818.7
52.36 0.005826 72.5 - - - - - -
110
Table 4-6: Comparisons of the shear rates estimated with
different orders of Ω(M) for SEQ233. Some data is excluded as it was
found to be in the Taylor region.
Order of W(M)
W M t1 2nd 3rd 4th 5th 6th 7th
Shear rate
0.52 0.002431 30.2 -4.3 -3.9 11.9 15.4 5.4 7.0
1.05 0.00267 33.2 4.0 4.2 12.4 12.9 13.0 12.6
2.09 0.003034 37.8 26.7 26.8 23.8 22.8 25.4 25.2
3.14 0.003244 40.4 42.6 42.6 36.5 35.7 36.6 36.8
4.19 0.003408 42.4 56.4 56.2 49.5 49.1 48.5 48.8
5.24 0.003529 43.9 67.2 67.0 60.6 60.7 59.2 59.5
6.28 0.00364 45.3 77.8 77.6 72.1 72.5 70.6 70.7
7.33 0.00373 46.4 86.7 86.4 82.0 82.6 80.6 80.6
8.38 0.003853 47.9 99.4 99.1 96.5 97.4 95.6 95.5
9.42 0.00393 48.9 107.6 107.3 106.1 107.1 105.6 105.4
10.47 0.004009 49.9 116.3 116.0 116.3 117.3 116.3 116.0
15.71 0.004311 53.6 152.1 151.9 157.6 158.0 159.4 159.2
20.94 0.004575 56.9 186.5 186.5 194.7 194.0 196.6 196.8
26.18 0.004813 59.9 220.0 220.2 226.6 224.8 226.5 227.0
31.42 0.00502 62.5 250.9 251.5 251.1 249.2 248.4 248.8
36.65 0.005235 65.1 285.1 286.1 271.6 271.4 267.5 267.2
41.89 0.005441 67.7 319.5 321.0 284.3 288.7 284.3 283.4
47.12 0.00563 70.1 - - 288.3 300.8 302.3 302.0
52.36 0.005826 72.5 - - - - - -
111
Figure 4-4: Comparison between the Casson approximation and
the Estelle method using a 6th order polynomial fit of Ω(M). The data
was from SEQ233.
As seen in Figure 4-4, the flow curves produced by the Casson

approximation (see section 4.9.4) produces shear rates which are
reasonably comparable to the Estelle method using a 6 th order
polynomial fit. The Casson model can be fi tted to this Estelle
approximation; doing so produces a good fit (R 2=0.99) as seen in
Figure 4-5. Table 4-7 compares the parameters from the direc tly fitted
values (Figure 4-4) to the Estelle fitted values ( Figure 4-5) and are
similar. Essentially, the Casson approximation produces a similar
quality of fit, if not better, using less parameters. For t his reason
(amongst others discussed in section 4.9.4) the Casson model was
chosen as the primary method for determining the flow curves.
112
Figure 4-5: Fitting a Casson model to the Estelle approximation
of SEQ233 using a 6th order polynomial fit of Ω(M).
Table 4-7: Comparison of the parameters of the Casson curve.

The directly fitted curve refers to the parameters found using the
method as described in section 4.9.4. The fitted to Estelle column
refers to the fit as seen in Figure 4-5.
Parameter Directly fitted Fitted to Estelle
Infinite viscosity 29.5 32.1
(mPa.s)
Yield Stress 27.1 26.5
(Pa)
113
4.9.3 Turbulence
Rheology must be performed under laminar conditions. For each
shear rate the Taylor number was found using the local apparent
viscosity and checked to see if in range using:

𝑇𝑎 = ∗√ 3-6
𝜂 𝑅𝑏
The obvious limitation in this method is that Taylor generat ed

his results using Newtonian fluids and may not be appropriate for non-
Newtonian fluids. However, given the lack of a suitable alternative
metric or the equipment to directly check if the flow is actually
lamina, the Taylor number is the most suitable te chnique available.
Any data within the turbulent region mu st be excluded.
4.9.4 Casson approximation

Another technique to finding the shear rate is to directly fit the
raw data to a fluid model. The quality of the fit can then be evaluated
by using the fluid model parameters to predict the angular speed from
a given torque reading. This was performed for the Casson equation as
seen in equations 4.2 through 4 -4.
2
√τ−√τ0
( )
r γ̇ 1 τ f(τ) 1 τ2 √η∞ 1 τ 2√τ0 τ
Ω = ∫r 2 r dr = − 2 ∫τ 2 dτ = − 2 ∫τ dτ = − 2η ∫τ 2 (1 − + τ0 ) dτ 4-2
1 1 τ 1 τ ∞ 1 √τ
I f τ2 = τ0 :
1 τ
Ω = 2η (τ − 4√τ0 τ + 3τ0 + τ0 ln (τ )) 4-3
∞ 0
1
I f τ2 > τ0 , t h e n τ2 = 𝛼2 τ1 , h e n c e
1 1 1
Ω = 2η ( (1 − α2)τ1 − 4(1 − α)√τ0 τ1 + 2τ0 ln(α)) 4-4
∞
114
The R2 value of the predicted angular speed against the actual
can then be found through the use of the excel “CORRELL” function.
Excel solver was used to fit the parameters to minimise the sum of
squares whilst setting the yield stress to be at least 0 but also be
beneath the minimum measured stress. A third and final re quirement
was that the infinity viscosity was set to be greater than 1 mPa.s as
this is the viscosity of the carrying fluid (water).
The advantages of this technique are that it is quick to perfo rm,

automatable, and consistent. Different curves can then al so be
prepared and can be objectively assessed. If the flui d is well described
by the Casson model, then it also saves time fitting the parameters to
the produced shear rate data since it has alre ady been performed
against the angular speed and torque.
Given that the numbers found in the Casson equation could be

affected by data points within the Taylor range, the Casson
approximation involved an iterative procedure where:
1. The data points outside the manufacturer’s specifications

were excluded.
2. The Casson curve fluid parameters were generated using

excel solver minimising
∑(ΩPredicted − ΩMeasured )2 4-5
3. The Taylor number was calculated for each shear stress
4. If the Casson fluid was modelled over any data point which
included Taylor numbers, those points were excluded, and
the process was re-iterated from step 2.
R2 cannot be, by its nature, tested statistically. However, R can

through using a t-test to test the null hypothesis that it is zero (i.e. a
2-sided t-test with n-2 degrees of freedom). From the t-test the p-
value can be evaluated using Microsoft Excels inbuilt functions
(Napier-Munn, 2017).
115
𝑅√𝑛−2
𝑡= 4-6
√1−𝑅 2
4.10 Summary
Two type of rheological experiments were performed: the HAAKE
experimental procedure i.e. the helical flow rheometer (see section
4.7) and the RM180 experimental procedure i.e. the coaxial cylinder
method (see section 4.8). The HAAKE procedure is presented here to
give future readers an understanding of conducting flow rheometry
tests; the results presented in chapter 5 and discussed in chapters 6
through 7 are based off the RM180 experimental procedure.
The experimental program of the results as presented in chapter 5

can be summarised in 4 steps:
1. Materials was received and prepared for rheological

testing. Each sample had a unique identifier consisting of
the letters NTS followed by three numbers (section 4.1).
The samples were split into subsamples allowing
rheological tests and sizing (section 4.4.1). If necessary,
the subsamples were ground (see section 4.4.2).
2. The subsamples were mixed with water (see section 4.8.1)

and loaded into the RM180 rheometer cup (see section
4.8.2). A bob was selected (see section 4.8.3) and a yield
stress test (see section 4.8.4) was performed followed by a
ramp up, hold, and ramp down loop (see section 4.8.5).
During this test the excess sample had its pH and dissolved
oxygen measured (see section 4.6).
3. After the rheology test was completed, a sedimenta tion test

was performed on the same sample that had been used in the
rheology test (see section 4.5). The rheometer was cleaned
and step 2 was completed for the next subsample.
The data was analysed as describ ed in section 4.9. It was

found that the Casson approximation gave a good fit for
116
most of the data. This processed data is presented in
chapter 5 and discussed in chapters 6 and 7.
117
Chapter 5
Test results
This section provides an overview of both the slurry rheology
and stability (i.e. sedimentation and swelling) results. Analysis is
largely left to the following three chapters; the chief exception to this
is section 5.1 which gives an overview of the calibration of the RM180
and the validation of the final testing procedure. As was first
referenced in section 4.1, the samples each have an NTS number whilst
each rheology test is described by a unique SEQ number.
5.1 RM180 calibration and validation of the final

testing procedure
There were large variations in the rheology of different samples.
Two different bobs were required: these will hence forth be referred to
the “large bob” and the “small bob”. The large bob would overlo ad on
the more viscous/solid samples whilst the small would be unable to
accurately measure the shear stress for the less viscous/solid. At the
beginning of the project, it was impossible to predict which bob was
more appropriate and numerous repeats had t o be conducted.
Another issue was determining whether a result was reliable.

Numerous tests had to be discarded because either the torque readings
were insufficient or due to turbulence effects.
118
Figure 5-1: Calibration of the RM180 using 486 mPa.s Newtonian
oil (prior to correction factors).
It was found that the results from using the two bobs were
largely comperable. Both bobs indicated Newtonian behaviour when
shearing calibration oil as can be seen in Figure 5-1. The calibration
oil was passed its manufacturer reconmended use by date; however,
the oil was found to have a viscosity of 486 mPa.s through using a
recently callibrated HAAKE viscotester IQ. As can be seen in Table
5-1, correction factors for calculating the shear stress were used to
allow the viscosity to be accurately determined by the RM180.
Table 5-1: A table showing the correction factors used for

calculating shear stress
Viscosity prior to Correction Viscosity post

Bob test (Pa.s) factor correction (Pa.s)
Small 0.659765948 1.357543163 0.486
Large 0.613974391 1.263321848 0.486
As stated earlier (refer to secti on 4.8.3), the correction factor

method is equivelent to an effective height calculation (assuming that
the rheometer is perfectly tuned). The effective height of the large
119
bob was found to be 57 mm which is similar to previous found values
for the bob in a different cup system (Shi, 2016). However, the
effective height for the small bob was surprisingly large; it was 28.5
mm even though the combined height of the cylindrical and conical
section is only 25 mm. The callibration was peformed identically to
the rheology tests and both bobs correctly gave Newtonian solutions;
hence, it is clear that the rheometer gave reliable results post
correction. The correction factor would have captured any cause of a
scalar discrepency including the effective heights due to end effects,
the rheometer being poorly tuned, excetera.
However, an issue was idetified in using the large bob: at higher

RPM’s it gave inconsistent readings as can be seen in Figure 5-2; the
raw results indicate that the RM180 is inconsistent at higher RPM’s.
SEQ203 and SEQ204 refer to the rheology sequence. Both used the
same sample as a repeat.
Figure 5-2: NTS100 53% (v/v) Comparison of raw results for the
large bob
120
To validate that the sample was consistent, and that the
rheometer was reliable, a test was performed where slurry was made
on separate days using different subsamples from the same NTS
number. Each slurry underwent identical sample preparation,
rheological conditions and gave near identical answers as can be seen
in Figure 5-3.
Figure 5-3: Repeat of NTS054 using fresh samples
5.2 Confirmation that the master sizing is

comparable to sieving
The Malvern MasterSizer results were compared to fine wet
sieving as can be seen in Figure 5-4. The sample, NTS050, belongs to
the distal chlorite domain and was given to the project to be used as
dummy sample for the commissioning and confirmation testing. The
results clearly show that using the Malvern, despite the caution
expressed in the literature review, is sufficient for determining Lihir
121
ore particle sizes beneath 38 μm as it gives similar results to wet
sieving.
Figure 5-4: Comparison of the Malvern MasterSizer to wet

sieving below 38 μm.
As discussed in sections 2.9.5 and 4.3 the Malvern MasterSizer is

a laser diffraction technique. Neither laser diffraction or wet sieving
are capable of taking particle shape into account; the results assume
that the particles are spherical, and the size is the characteristic
diameter. Future work should consider alternative techniques such as
microscopic image analysis which were viewed as being beyond the
scope of this study.
5.3 Particle size and size distribution

All ore samples, as was described in sections 4.1 & 4.3, was
provided as crushed and sieved to minus 3.35 mm. It was then further
sieved to less than 1.00 mm; at this stage no efforts were made to
further change the size distribution. It was determined from the
122
beginning of the project that we should be investig ating the slurry
under a polydisperse condition as the slurry within the grinding
circuits are likely to be polydisperse. The particle size dis tribution of
outer biotite is presented in Figure 5-5; other similar graphs are
presented in the Appendix D.
Figure 5-5:Outer biotite cumulative passing size
123
Table 5-2 summarises the data by provided the cumulative size
fractions at three fine contents (9.86, 38, and 53 µm) and a series of
passing sizes (P20, P50, and P80). The full results are provided in
both tabulated and graphical forms in Appendix D. Table 5-2 shows
that there is a large variation in the ore particle size and particle size
distribution.
124
Table 5-2: Size distribution on as provided mat erial sieved to
less than 1 mm
NTS Alteration Percentage passing Passing size (µm)

domain
-9.86 -38 -53 P20 P50 P80
(µm) (µm) (µm)
NTS008 Upper 19 29 33 12 154 534

epithermal
NTS010 Silica 16 25 29 18 210 593

Breccia
NTS019 Inner 11 19 23 41 213 614

biotite
NTS027 Outer 11 21 25 36 215 633

biotite
NTS037 Advanced 22 41 45 8 73 351

Argillic
NTS038 Argillic 28 43 47 5 66 398
NTS044 Outer 14 21 25 31 230 642

biotite
NTS045 Outer 17 25 30 16 177 574

biotite
NTS046 Outer 21 31 35 9 134 524

biotite
NTS047 Lower 20 30 34 10 139 536

Epithermal
NTS049 Argillic 12 23 27 27 200 572
NTS054 Outer 13 20 24 36 235 640

biotite
125
NTS Alteration Percentage passing Passing size (µm)
domain
-9.86 -38 -53 P20 P50 P80
(µm) (µm) (µm)
NTS055 Inner 9 17 22 47 198 596

biotite
NTS097 Inner 18 30 35 12 127 478

biotite
NTS100 Outer 9 18 23 43 196 560

biotite
NTS106 Silica 12 21 25 32 269 688

Breccia
NTS108 Argillic 17 28 32 13 153 511
NTS109 Argillic 23 35 39 7 110 414
NTS110 Lower 23 36 41 7 99 437

Epithermal
NTS111 Upper 10 19 22 41 278 636

epithermal
NTS128 Silica 11 20 23 40 255 662

Breccia
NTS238 Argillic 23 35 38 8 116 463
NTS239 Upper 17 26 31 15 165 579

epithermal
NTS240 Lower 26 43 46 6 80 469

Epithermal
To further investigate the particle size effect, some samples were

ground for 5 minutes in order to reduce the mean particle size and
narrow the size distribution. As stated in section 4.4.2, these samples
represent 5 of the 9 domains at Lihir. The domains not included either
126
were not of interest to Newcrest or were provided in insufficient
quantity for these tests.
Table 5-3: The size distribution of the material after being

sieved to less than 1 mm and undergoing a further 5 minutes of
grinding
NTS Alteration domain Percent passing size Percent size (µm)
-9.86 -38 -53 P20 P50 P80
(µm) (µm) (µm)
NTS044 Outer biotite 25 40 45 6 67 159
NTS054 Outer biotite 22 39 44 8 69 161
NTS097 Inner biotite 29 49 57 5 40 100
NTS108 Argillic 34 54 55 4 28 122
NTS110 Lower Epithermal 39 62 69 3 19 77
NTS111 Upper epithermal 21 37 43 9 73 170
127
5.4 Summary of the rheology and sedimentation
data
This is a summary of the major results. Further data can be found
in the Appendix B, F, and G.
As referred to in section 6.1, the sample underwent clear time

dependent shear thickening. However, as described in section 2.2.2,
the loop test is a qualitative test for determining time dependency. To
make them comparable, the change in the ar ea under the rheogram
(using the trapezoid method) is reported as the time dependency.
The concentration, pH, apparent viscosity at 50 s-1 Casson fitting

variables (yield stress, infinite viscosity, and the R 2 of the fit), time
dependency, sedimentation rate and swelling have been tabulated for
each sample; an example of such a table is Table 5-4 which
summarises the results for NTS047 (a lower epithermal sample).
Similar tables for the other samples are provided in Appendix B. The
apparent viscosity of 50 s-1 was selected as it intermediate viscosity
that is within the shear rate range estimated to occur in b all mills (Shi
and Napier-Munn, 1999).
Table 5-4: NTS047 results
128
Table 5-5 and Table 5-6 show respectively the results for the
highest concentration achieved and the highest concentration achieved
whilst observing sedimentation. From these tables it is clear that
silica breccia, inner biotite and outer biotite settle at a higher
concentrations and are less viscous. It is also clear that there was a
large variance within and between the domains. This will be discussed
further in chapters 6 and 7.
129
Table 5-5: Rheology and sedimentation characteristic parameters
measured at the highest solid concentration achieved.
Domain NTS Sequence Solids pH Apparent Yield Infinite Sedimentation
number (SEQ) volumetric viscosity stress viscosity rate
concentration (mPa.s) (Pa (mPa.s) (mL/hour)
at 50 s-1
Advanced NTS037 158 52 249
Argillic 1.38 2.6 73 NA
Argillic NTS038 394 53 1.83 7192 168.5 716 NA
NTS049 215 53 2.91 1727 46.6 122 NA
NTS108 231 46 2.26 1795 40.5 193 NA
NTS108 247 42 3.17 3364 NA
grind 91.5 232
NTS109 399 53% 2.14 9889 265.9 703 NA
NTS238 293 46% 2.17 1804 49.7 120 NA
Upper NTS008 193 53% 6.51 2160 61.903 127.54 NA
epithermal NTS111 346 58% 1.96 2838 115.1 28 NA
NTS111 3014 NA
grind 180 53% 2.24 96.8 119
NTS239 361 53% 4.04 3190 60.6 470 NA
Silica NTS010 351 53% 2.02 842 24.4 48.19 0
breccia NTS106 338 53% 2.74 892 27.0 44 0.9
NTS128 263 53% 5.76 510 15.7 24 2.2
Lower NTS047 329 53% 2.9 3902 110.2 241 NA
epithermal NTS110 305 46% 3.43 4792 109.9 499 NA
NTS110 4406
grind 298 41% 4.75 122.1 288 NA
NTS240 289 59% 2.95 3900 138.9 95 NA
Outer NTS027 384 53% 7.3 451 5.9 108.11 2.0
biotite NTS044 1449 NA
ground 270 46% 6.05 44.3 69.16
NTS045 377 53% 5.27 2078 47.9 214.2 NA
NTS046 245 53% 5.57 4728 139.9 251.84 NA
NTS054 375 57% 5.8 1852 44.3 176 NA
NTS054 NA
grind 323 53% 6.00 3313 97.8 178
NTS100 203 53% 6.77 682 10.2 140 1.8
Inner NTS019 374 53% 6.92 539 12.0 59.43 9
biotite NTS055 278 53% 6.42 515 3.4 211 4.0
NTS097 238 53% 4.21 1637 35.9 186 NA
NTS097 NA
grind 313 46% 5.1 1514 40.7 108
130
Table 5-6: Rheology and sedimentation characteristic parameters
measured at the highest solid concentration achieved.
Domain NTS Sequence Solids pH Apparent Yield Infinite Sedimentation
number (SEQ) volumetric viscosity stress viscosity rate
concentration (mPa.s) (mPa.s) (mL/hour)
at 50 s-1
Advanced NTS037 123
Argillic 168 46% 1.33 0.0 123 1.8
Argillic NTS038 389 31% 2.03 7 5 2.6 2 2.7
NTS049 216 46% 2.75 5 1 8 10.8 65 0.5
NTS108 235 36% 2.31 2 1 9 6.7 11 1.2
NTS108 1.1
grind 251 27% 3.22 104 3.4 4
NTS109 397 36% 2.24 318 10.3 12 1.1
NTS238 297 42% 2.26 1076 37.3 30 0.8
Upper NTS008 196 41% 2.64 168 4.152 14.82 1.5
epithermal NTS111 341 53% 2 784 22.5 46 0.3
NTS111 1.7
grind 183 41% 2.31 99 2.7 7
NTS239 356 41% 4.06 368 10.4 23 0.8
Silica NTS010 352 46% 2.03 322 6.5 43.53 0.8
breccia NTS106 338 53% 2.74 892 27.0 44 0.9
NTS128 263 53% 5.76 510 15.7 24 2.2
Lower NTS047 332 46% 2.91 1059 30.0 64 0.2
epithermal NTS110 307 31% 3.39 181 6.0 6 0.8
NTS110
grind 300 31% 4.57 4 9 4 19.0 7 1.3
NTS240 284 51% 2.91 6 1 7 16.5 45 0.8
Outer NTS027 384 53% 7.3 4 5 1 5.9 108.11 2.0
biotite NTS044 0.81
ground 269 42% 5.94 726 25.5 19.20
NTS045 379 46% 5.39 705 13.8 98.8 2.6
NTS046 258 41% 3.86 317 9.0 19.36 0.25
NTS054 364 53% 5.47 828 18.5 91 1.8
NTS054 0.8
grind 324 46% 6.04 815 22.6 53
NTS100 203 53% 6.77 682 10.2 140 1.8
Inner NTS019 374 53% 6.92 539 12.0 59.43 9
biotite NTS055 278 53% 6.42 515 3.4 211 4.0
NTS097 239 46% 4.21 461 13.0 29 1.6
NTS097 5.3
grind 311 38% 5.09 2 2 2 7.0 9
The differences between the maximum solid concentration found
in
131
Table 5-5 was due to the largest differences in viscosities.
Certain samples it was not possible to go to a higher concentration;
for example, NTS110 after 5 minutes of grinding was so solid that it
could maintain its own structure at 46% solid s as seen in Figure 5-6.
Clearly it had become a thick paste and was no longer a slurry.
Figure 5-6: NTS110 at 46% (v/v) after 5 minutes of grinding.
5.5 Summary of test results

In section 5.1 it was shown that the RM180 gives consistent
results. The calibration procedure was also explained. Section 5.2
showed that the Malvern MasterSizer gave similar results to wet
sieving at beneath 38 μm. Section 5.3 summarised the particle size
distribution of Lihir ores whilst section 5.4 gave an overview of which
samples are highly viscous and the maximum solid s concentration in
which the slurry was observed to begin settling. The full results can
be found in the appendixes:
132
• Appendix A connects the sequence list to the NTS### and
its solid concentration;
• Appendix B provides the summarises of the rheology and

sedimentation for each NTS###. It is sub -divided into
subsections for each domain;
• Appendix D provides the particle sizing data ;
• Appendix F and Appendix G are in separate pdf files. They

contain the sedimentation curves and the flow curves
respectively.
133
Chapter 6
Analysis of the slurry
rheology
As seen in chapter 5, several different samples were tested at a
variety of different concentrations. The difference in the results was
dramatic; many samples failed to produce flow curves whilst other
samples were essentially pastes.
6.1 Time dependency

Time dependency was sample dependent. It appears that some
samples are entirely time independent, such as NTS054 (see Figure
6-1), whilst other samples are highly time dependent, such as sample
NTS108 (please see Figure 6-2). Discerning time dependency is
difficult as some of the samples settled during the tests; as discussed
elsewhere sedimentation can lead to an apparent time dependent shear
thinning (due to a decrease in solids concentration) and shear
thickening (if the bob gets caught in the sediment). Settling wil l be
discussed more in chapter 7.
134
Figure 6-1: NTS054 at 57% solid volumetric concentration.
Figure 6-2: NTS108 at 46% solid volumetric concentration.
135
Figure 6-2 shows the flow curve of a highly time dependent
sample. The area under the ramp down curve is twice the size of the
area under the ramp up. This time dependent shear thickening makes
the analysis highly difficult; shear rate is dependent on the change in
angular speed with regards to shear stress as seen in
r γ̇ 1 τ f(τ)
Ω = ∫r 2 r dr = − 2 ∫τ 2 dτ 2-10
1 1 τ
For the data reduction, the shear rate was fo und for the ramp up
using a similar method to Estellé et al. (2008); the shear rate was then
linked with RPM for the ramp down. This simplified the comparisons,
however, using a relevant time dependent model may yield a more
accurate evaluation of the shear rate. This isn’t a simple task and it is
not clear which time dependent model would be appropriate. As such,
it is likely that the fluid is even more viscous; this is an area
requiring future work. The raw data illustrates the scale of time
dependency as shown by Figure 6-3 and Figure 6-4.
Figure 6-3: Raw data of the hysteresis loop conducted on NTS108

(Sequence 231 with a solid volumetric concentration of 46%)
136
Figure 6-4: The raw data of NTS108 (SEQ231) at a solid
concentration of 46% (v/v) being sheared by the small bob at 999
rotations per minute.
The “yield tests” (section 4.8.4) underwent the characteristic

time dependent changes as described by Nguyen and Boger (1983). As
described in section 4.8; the standard procedure was to perform a
single low RPM cycle followed by a single loop (ramp up, hold and
ramp down). To investigate the possibility of the thixotropy in the
“yield stress” tests the standard p rocedure was altered for a highly
viscous shear thickening sample NTS109 at 53% solids concentration
(v/v) (SEQ400). This test had additional loop tests performed; it
showed that the suspension continued to be thixotropic at low speeds
as seen in Figure 6-5.
137
Figure 6-5:The change in the “yield stress” of an argillic sample
(NTS109 at 53% concentration; SEQ400) prior to the first, second, and
third loop test.
Each domain (with the exception of advanced argillic)

demonstrated a combination of both thixotropy and anti-thixotropy.
This analysis is complicated by the fact that some samples settled and
by the nature of the rheology tests; loop tests (see section 4.8.5) can
only give qualitative information on thixotropy, to allow the
thixotropy to be compared, the change in the area under the rheogram
(using the trapezoid method) is used to describe time dependency
(section 5.4). Hence, the numbers given describing the thickening and
thinning in the following paragraphs is merely qualitative and not
quantitative. Each sample which displayed anti -thixotropy only
displayed it at high rotational speeds (as seen in Figure 6-3 and Figure
6-4), at low rotational speeds thixotropy was observed as seen in
Figure 6-5.
138
6.1.1 Advanced argillic
Only one advanced argillic sample was experimented on: NTS037.
It was displayed no time dependency in any loop tests. Repeats using
the same sample were performed and the viscosity underwent no
change. The pH was extremely low: 1.38 which may have dispersed the
slurry.
6.1.2 Argillic
The argillic clay samples displayed anti -thixotropic, thixotropic
and time independent properties. There were five different argillic
samples NTS038, NTS049, NTS108 (both as provided and ground),
NTS109, and NTS238; each samp le had rather different particle sizes
and particle size distributions. With the exception of NT S049, all
samples were anti-thixotropic at the highest concentration (the exact
concentration varied due to the highly thick nature of the slurry); the
second highest concentration (46%) of NTS049 thickened with time
like the other samples. The size of the anti-thixotropy becomes
smaller as the concentration drops; transitioning to time independent
(noted in every sample) and in NTS108 (as provided) becoming
thixotropic (thins by 17%) at 31% concentration. This thixotropic
behaviour at 31% in NTS108 may hav e been due to sedimentation as
the ground NTS108 was time independent at similar concentrations
(30% and 27%). The sedimentation rate was however low (3.4
mL/hour) and the unground material was significantly more acidic (pH
2.26-2.35 vs 3.17 to 3.22); as discussed in section 6.1.4 this may be a
genuine effect caused by the change in the pH.
6.1.3 Upper epithermal

There were three different upper epithermal samples NTS008,
NTS111 (both as provided and ground), and NTS239. Upper epithermal
samples underwent anti-thixotropic thickening at 46% concentration.
However, at 53% concentration, the size of the thickening effect
decreased with NTS239 becoming less anti-thixotropic, NTS008
139
became time independent and NTS111 (both as provided and ground)
became thixotropic. NTS111 (as provided) was also tested at a higher
concentration of 58%; it was even more thixotropic than at 53%
concentration. NTS111(after grinding) was investigated as shown in
Figure 6-6; the material was clearly anti -thixotropic.
Figure 6-6: The increase in yield stress over time of NTS111

(after grinding and at 46% solid volumetric concentration).
It is not clear why the material displayed greater anti -thixotropy

at lower concentrations. It is possible that th e structure was still
building after mixing.
6.1.4 Silica breccia

Silica breccia underwent a complicated time dependency. There
were three different silica breccia samples NTS010, NTS106 and
NTS128; all three had similar particle sizes and size distributions.
However, NTS010 and NTS106 both were highly acidic (approximately
140
pH 2) whilst NTS128 was mildly acidic (approximately pH 5.7). The
two highly acidic samples underwent the same trend at 53%
concentration: they were thixotropic, thinning between 7 to 10%.
NTS010 and NTS106 thickened by 29% and 46% at the lower
concentration of 46% whilst sedimentation remained low at 0.8
mL/hour and 2/6 mL/hour respectively. There are two possibilities that
explain why the time dependency changed with the change in solids
concentration. Firstly, as the thickening samples were undergoing
sedimentation, it is possible that a thickened mud had formed around
the bob leading to increased torque readings. Secondly, the
suspensions breakage and rejuvenation mechanics are interlinked with
the concentration.
The other silica breccia sample, NTS128, d id not display any

thixotropic effects. At 53% concentration, the sample thickened by 5%
and underwent some sedimentation (sedimentation rate of 2.2
mL/hour). At 46% concentration it thicke ns by 8% but settles much
faster at 8.9 mL/hour thus limiting the v alue of the time dependency
analysis at this concentration. Given that NTS128 had a similar
particle size distribution, it seems unlikely that the shear thickening
seen in NTS010 and NTS106 at 46% is due to sedimentation as it
settled at a similar rate at 53% concentration for NTS128. An apparent
anti-thixotropy can appear in settling samples due to the bob being
caught on the thickened sediment (Klein et al., 1995); as stated in the
previous sentence this seems unlikely as it settled at a similar rate at
a higher concentration but underwent less thickening. If the bob was
being caught it would be more likely that the higher concentration
sample would give the same if not higher readings. Thus, it is likely
that the silica breccia’s time dependency is dependent o n the
concentration and the pH.
6.1.5 Lower epithermal

There were three different lower epithermal samples NTS047,
NTS110 (both as provided and ground), and NTS240. Like the time -
141
independent characteristics of the lower epithermal samples, there was
a great degree of variance in the time dependence within the lower
epithermal samples. For NTS110 it was anti-thixotropic between 46%
and 36% (v/v); the higher the concentration the gre ater the effect. At
these concentrations NTS110 was non -settling; clearly the anti-
thixotropy was dependent on changes in the suspensions structure
rather than an artefact of the device. At lower solid concentrations,
31% and 27% (v/v), the sample transiti oned to become time
independent and thixotropic respectfully. However, this coincided with
the onset of sedimentation so the thixotropy could possibl y be an
artefact of the device (Klein et al., 1995).
6.1.6 Outer biotite

Outer biotite samples time dependencies were highly variable. At
elevated concentrations (>53%) the samples were time independent. At
lower concentrations they displayed anti -thixotropy but outer biotite
was typically faster settling then other suspensions so this could have
been due to device artefacts (Klein et al., 1995).
6.1.7 Inner biotite

Inner biotite samples settled even at high concentrations making
the analysis of whether there were anti-thixotropic effects or if there
were just sedimentation effects impossible to answer. Of the samples
NTS019, NTS055 and NTS097 (as provided and ground) only the
ground NTS097 had results where there was no significant
sedimentation. It was noted that the samples were fast settling and
alternative ramping procedure (ramping from a high rpm to a low
rather than from a low to a high) was employe d for NTS019 at 53%
solids concentration (v/v; SEQ374); the nature of the suspension was
unusual in that it thickened around the bob but was fluid above and
beneath it.
142
From analysing NTS097 (after grinding) it appears that inner
biotite is anti-thixotropic at elevated concentrations; the time
dependency becomes less significant as the concentration decreases.
6.1.8 Comparison to other materials

As stated in section 2.9.4 complex time dependency has been
noted in other mineral suspensions (Barnes, 1997; Boger, 2009;
Brandenburg and Lagaly, 1988; Coussot et al., 2002; Lagaly, 1989;
Larson and Wei, 2019; Mewis and Wagner, 2009; Nguyen and Boger,
1985). Larson and Wei (2019) reviewed time dependency and found
that real thixotropic materials typically show a yield stress while anti -
thixotropic material are often flowable fluids at low stress that
become more viscous or even solid-like during and after flow as seen
in Figure 2-30. Clearly, the anti-thixotropic curves produced in this
study are different to the fluid described in Figure 2-30. Significantly,
all Lihir ore suspensions (with the exception of NTS037 at 46% solids
concentration (v/v)) have yield stress. Furthermore, as seen in Figure
6-6, the suspensions which display anti-thixotropic effects also
display thixotropic behaviour at low shear rates.
This viscosity bifurcation was distinct from the bifurcation

observed by Brandenburg and Lagaly (1988); Lagaly (1989) and
Coussot et al. (2002). These researchers found that the bifurcation was
towards either stoppage or rapid shear as illustr ated in Figure 2-31; or
in other words the apparent anti -thixotropic effects occur at low shear
rates, and the thixotropic effects occur at high shear r ates. In this
study the dual thixotropic and anti -thixotropic behaviour in Lihi r ore
suspensions was the opposite to what was observed in these previous
studies; the viscosity grew thinner at a low shear rate and thicker at a
high shear rate. It is likely that this complicated rheological
behaviour of Lihir ore slurry, similarly to t he studies done by
Brandenburg and Lagaly (1988); Lagaly (1989), is also dependent on
the pH, salt concentration and the temperature; these shou ld be
investigated in the future.
143
6.2 Solids concentration effect on yield stress and
viscosity
Higher solids concentrations, as expected, resulted in an
increased yield stress and apparent viscosity across all tested samples;
Figure 6-7 is a diagram demonstrating the effect of an increase in
concentration on viscosity. Similar diagrams are displayed overlapped
with other samples that share a domain in section 6.4.
Figure 6-7: Apparent viscosity (at a shear rate of 50 s -1) in

relation to solid volumetric concentration for argillic sample NTS108.
The scale of the effect that concentration has on the rheological

characteristics is made abundantly clear when super-imposing the
various flow curves onto a single diagram as in Figure 6-8. Figure 6-8
shows that as the solid volumetric concentration in creases both the
yield stress and apparent viscosity also increase.
144
Figure 6-8: Flow curves of NTS008 (upper epithermal) at 53%
and 46% solids concentration (v/v).
The effect is similar for the apparent yi eld stress. Every “ramp
up” curve in this thesis was well described by the Casson model (see
section 5.4) with the exception of NTS037 at 46% (v/v) which was
Newtonian. This indicates that the concentrations used in this study
remain beneath the critical concentration to cause time independent
shear thickening (Barnes, 1989; Bender and Wagner, 1996; Bertrand et
al., 2002; Hoffman, 1998; Maranzano and Wagner, 2001; Wagner and
Brady, 2009). Barnes (1989) argues that any suspension will
experience shear thickening at a high enough shear rate; Shi and
Napier-Munn (1999) estimated that the shear rate within a ball mill
ranged between 13 and 730 s-1 which is above the highest shear range
achieved in this study. Taking this upper limit into consideration, it is
possible that the suspensions could be shear thickening under the
conditions inside the ball mill. However, Shi and Napier-Munn (1999)
calculated this upper limit from the shear stress applied to a ball
being 110 Pa, assuming a viscosity and that the fluid underwent
145
Newtonian flow. The viscosities of the suspensions investigated in
this study are much higher and are non -Newtonian, as such, it is
unlikely that any sample at the concentrations investigated in this
study would be shear thickening at the shear rates within the ball mill.
This requires further investigation.
6.3 Grinding effect

The reduction in particle size and particle size distribution was
expected to increase the viscosity (see section 2.4.2); as expected
grinding did increase the viscosity and yield stress of the samples.
The more solid-like nature of the slurries prevente d an exact
comparison of the samples across the total same solid concentration
range. Table 6-1 provides the particle size information of both the as
provided and ground samples. Comparisons between the apparent
viscosity of the sample as provided and after grinding are provided in
Figure 6-9 through Figure 6-13.
146
Table 6-1: Particle size information for the ground samples
NTS Alteration Percent passing size Size (µm)

domain
-9.86 -38 -53 P20 P50 P80
(µm) (µm) (µm)
Outer 14 21 25 31 230 642

NTS044 biotite
NTS044 grind 5 Outer 25 40 45 6 67 159

min biotite
Outer 13 20 24 36 235 640

NTS054 biotite
NTS054 grind 5 Outer 22 39 44 8 69 161

min biotite
Inner 18 30 35 12 127 478

NTS097 biotite
NTS097 grind 5 Inner 29 49 57 5 40 100

min biotite
NTS108 Argillic 17 28 32 13 153 511
NTS108 grind 5 34 54 55 4 28 122

min Argillic
Lower 23 36 41 7 99 437
NTS110 Epithermal
NTS110 grind 5 Lower 39 62 69 3 19 77

min Epithermal
Upper 10 19 22 41 278 636

NTS111 epithermal
NTS111 grind 5 Upper 21 37 43 9 73 170

min epithermal
147
Figure 6-9: The grinding effect on the relationship of slurry
viscosity with solid volumetric concentration for outer biotite
(NTS054)
Figure 6-9 indicates that NTS054 (an outer biotite sample)

undergoes a large increase in the viscosity with a reduction in particle
size. The effect appears to grow stronger with an increase in
concentration; this should be investigated in further work.
148
viscosity with solid volumetric concentration for inner biotite
(NTS097)
NTS097 shows a large increase in viscosity with a decrease in

particle size as seen in Figure 6-10. It appears as if the curve was
shifted leftwards by the grinding. This implies that the relative
difference in viscosity may be scalar. This should be investigated in
further work.
149
Figure 6-11: The grinding effect on the relation ship of slurry
viscosity with solid volumetric concentration for argillic (NTS108)
NTS108 is an argillic sample which is highly viscous as seen in

Figure 6-11. The sample was viscous was high even prior to grinding;
not only did grinding increase the viscosity but it appears to have
changed the relationship between solids concentration and apparent
viscosity. It appears that the rate of increase in viscosity as solids
concentration increases is higher for the sample a fter grinding; as
mentioned in relation to NTS054 this should be investigated in further
work.
150
viscosity with solid volumetric concentration f or lower epithermal
NTS110
Figure 6-12 shows that NTS110 also increased its viscosity

significantly after grinding. As seen in Figure 5-6 (in section 5.4) by
46% solids concentration (v/v) NTS110 was no longer a slurry but a
thickened paste with characteristics more similar to wet conc rete then
mineral slurry. At that concentration, a slump test failed to produce a
result as the yield stress was great enough that no measurable slump
occurred (Boger, 2009; Pashias et al., 1996).
151
viscosity with solid volumetric concentration for upper epithermal
(NTS111)
Figure 6-13 indicates that the increase in viscosity after grinding

is dependent on concentration. NTS111 was significantly less viscous
then either NTS108 (argillic) or NTS110 (lower epithermal) but was
more viscous then NTS054 (outer biotite) and NTS097 (inner biotite);
yet it indicated the smallest effect that grinding had on ap parent
viscosity at 50 s-1. This should be investigated in future work.
Shi and Napier Munn (1999), estimated that the shear rate inside
a ball mill ranged between 13 and 730 s -1. The upper limit was
estimated from calculating a shear stress and then, ass uming a
particular apparent viscosity, working backwards to find the shear
rate. The shear stress they calculated was 110 Pa; several of the flow
curves generated as a part of this project had yield stresses within this
range. Such a slurry would not flow as a liquid but rather move as a
solid; this requires further investigation but may be the reason behind
the slurry pooling.
152
6.4 Domain
As discussed in section 2.1, the ore is characterised by Newcrest
into nine domains. Eight of these domains are processed by the mill;
unfortunately, due to sample limitations only seven of these domains
were investigated and diagrams comparing different samples can be
seen below in Figure 6-14 through Figure 6-20. At Lihir SAG
operations operate at solid concentrations ranging from 70 -75% (wt%),
and ball mill at around 70% (wt%); this corresponds to 46 -53% (v/v)
(Shi, 2017b).

relation to solid volumetric concentration for advanced argillic sample
NTS037.
Only one sample was used to characterise advance argillic

(NTS037). As seen in Figure 6-14 it had unusually small viscosity.
NTS037 was an unusual sample, at 36% became a Newtonian fluid. No
153
other sample tested displayed a lack of yield stress. It was also
extremely acidic with a pH of 1.32.

relation to solid volumetric concentration for argillic samples
As seen in Figure 6-15, NTS049 gave significantly different

results from the other argillic samples illustrating the large variability
within domains. The rheometer was at its torque limits for NTS038 and
NTS109 preventing the measurement at higher concentrations, hence,
further work is required.
154
relation to solid volumetric concentration for upper epitherma l
samples
155
Figure 6-17: Apparent viscosity (at a shear rate of 50 s-1) in
relation to solid volumetric concentration for silica breccia samples

relation to solid volumetric concentration for upper l ower epithermal
156
relation to solid volumetric concentration for inner biotite samples.
As seen in Figure 6-19, inner biotite has a low viscosity. The

relatively few data points shown weren’t because of a lack of
additional tests; the RM180 torque rea dings were too low at lower
concentrations to produce reliable results.
157
relation to solid volumetric concentration for outer biotite samples
Additional tests to what can be seen in the above graphs were

also performed. However, characterising the viscosity is error prone:
the rheometer either gave readings which were below the
specifications of the manufacturer or at higher speeds found to be in
the turbulent region.
The effect that alteration domain is best seen in Figure 6-21. The
figure compares the various as provided samples at 46% solids
concentration; as is clearly seen certain domains have a larger
apparent viscosity than others but there is significant variance within
each domain. It was not deduced why the Lower Epithermal domain
had such a large outlier. This concentration is at the upper range of
what is used in the SAG mill at the concentration used within the ball
mill (Shi, 2017b).
158
Figure 6-21: Slurry viscosities of various samples at a constant
46% solid volumetric concentration.
As stated in section 6.2 and seen by Figure 6-14 through Figure

6-20, each sample’s viscosity underwent significant increase in
viscosity with an increase in concentration. This was expected (see
section 2.4.1); it was hoped that this could be used for determining a
159
critical solid’s concentration for each domain all owing the operator
which to keep the slurry beneath. However, as can be seen in Figure
6-14 through Figure 6-20 there was large variance within each domain
thus preventing the recommendation of such a solids concentration.
Additional research is necessary before a critical concentration can be
recommended; in particular future research should establish:
• the underlying causes of the rheological variance within the

domains;
• and the rheological characteristics that should be

controlled such as a maximum yield stress or viscosity.
6.5 The dependence on ore chara cteristics

As can be seen in section 6.4, Lihir ore slurry apparent viscosity
depends on its alteration domain. However, there is a large quantity o f
variance within each domain between the samples; it was hypothesised
that the rheology may be dependent on either its clay content or the
amorphous silicate levels.
As can be seen in Table 6-2, the chemical composition of Lihir

ore is highly variable. The table does not indicate that any particular
clay will result in a highly viscous slurry or in an elevated yield
stress. Some highly viscous samples, such as NTS109, have only trace
levels of clay. From previous research it is clear that clay content has
a significant impact on rheology; and it is clear that certain domains
are more likely to have rheological challenges.
The table does not explain the different rheological behaviour

exhibited by different samples from the same alteration domain. The
upper epithermal samples had highly variable rheologies, yet none had
above a minor quantity of any particular clay. Likewise , within the
silica breccia domain clay content was highly variable, yet the
rheology was almost identical.
160
Table 6-2:Clay information
Clay information
NTS Alteration domain
QXRD_ClayMica QXRD_ClaySmectite QXRD_ClayChlorite QXRD_ClayKaolinite QXRD_ClayVermiculite QXRD_ClayCorrensite
NTS008 Upper epithermal trace
NTS010 Silica Breccia minor
NTS019 Inner biotite major minor
NTS027 Outer biotite abundant major minor major
NTS037 Advanced Argillic minor minor
NTS038 Argillic minor minor
NTS044 Outer biotite minor minor minor
NTS045 Outer biotite minor trace minor
NTS047 Lower epithermal minor minor
NTS049 Argillic minor minor minor minor
NTS055 Inner biotite minor trace trace
NTS097 Inner biotite major minor minor
NTS100 Outer biotite major minor minor
NTS106 Silica Breccia minor trace
NTS108 Argillic minor major abundant
NTS109 Argillic trace trace trace
NTS110 Lower epithermal minor minor major minor
NTS111 Upper epithermal minor minor
NTS128 Silica Breccia major minor
NTS238 Argillic major minor abundant
NTS239 Upper epithermal minor trace minor
NTS240 Lower epithermal abundant minor
161
This does not mean that there is no dependenc e on clay content.
The pH, particle size and particl e size distribution were all variable.
Rheology is a physical chemistry discipline; and it is well known that
rheology of slurries is often dependent on the chemical environment
(see section 2.4.3). In this project there was no control on the pH of
the slurries; it was merely monitored. The pH varied widely between
the samples, though, observing it in isolation it does not appear to
have affected the rheology as can be seen by Figure 6-22.
Figure 6-22: Slurry viscosity in relation to pH for all tested

samples
To investigate further a multilinear regression analysis was

performed. Given the variability in the particle size, size distribution,
and pH; it was thought that the amorphous silicates, kaolinite content,
and total clay composition effect on rheology may have been hidden
by the interaction of these other factors.
162
6.6 Multilinear regression
Due to the complicated nature of the response it was decided to
perform a multilinear regression analysis in order to link the
regression to:
1. the solids concentration;
2. the pH;
3. the particle size distribution;
4. and the geology.
As seen in Appendix D, there was a large quantity of data

surrounding the particle size distribution. Given the desire to limit the
number of terms particle sizes of 9.8, 38, and 53 µm were selected as
these terms are generally associated with the fines content in mineral
ore slurries. Further analysis should be perf ormed using other parts of
the size distribution.
The geological parameters most likely to affect the rheology

were identified as being the amorphous silicates content, the kaolinite
and the total clay content. Further investigations should be conducted
using alternative aspects of the geology.
Using Minitab, a stepwise regression was performed and can be

found in Appendix C. The regression gave a reasonable fit as can be
seen in Figure 6-23; however, it is clear that the variance is dependent
upon the scale of viscosity as seen in Figure 6-24.
163
Figure 6-23: Minitab graph of the multilinear regression results
against the fitted values
Figure 6-24: Minitab results of the fitted multilinear results

against the measured apparent viscosity at 50 s -1.
164
Emeritus Prof. Napier-Munn assisted with the regression analysis
and suggested using the natural log of the apparent viscosity at 50 s -1.
It greatly improved the regression results as seen in Figure 6-25 and
Figure 6-26.
Figure 6-25: The predicted vs measured natural logarithms of the

apparent viscosity at 50s-1.
165
Figure 6-26: Minitab’s four in one chart showing the residuals
relationship with the logarithm of the apparent viscosity.
166
The fitted equation of the viscosity logarithm was:
ln(η) = -13.25 - 0.405 pH + 15.71 Concentration - 6-1

64.0 Passing 9.86 - 154.1 Passing 38
+ 197.0 Passing 53 - 0.0935 amorphous -
0.383 QXRD_Kaolinite + 0.0353 Clay
- 0.0580 pH*pH -
1631 Passing 9.86*Passing 9.86 -
692 Passing 53*Passing 53
+ 0.002086 amorphous*amorphous -
10.12 pH*Passing 9.86
+ 8.19 pH*Passing 38
+ 0.00421 pH*amorphous -
0.0432 pH*QXRD_Kaolinite
+ 10.96 Concentration*Passing 53
+ 163 Passing 9.86*Passing 38
+ 2.505 Passing 9.86*QXRD_Kaolinite
The regression indicates that the viscosity is strongly correlated

to the concentration and particle sizes with larger concentrations a nd
smaller particle sizes (see passing 53 term) leading to higher
viscosities.
The relationship with pH is quite complic ated. Figure 6-27,

Figure 6-28, and Figure 6-29 indicate that a low to intermediate pH is
associated with higher viscosity whilst Figure 6-30 indicates that an
intermediate to high pH is associated with a higher viscosity.
167
Figure 6-27: Surface plot of the logarithm of apparent viscosity
at 50s-1 vs concentration and pH.
168
Figure 6-28: Surface plot of the logarithm of viscosit y at 50 s-1
vs the particle sizes smaller than 9.86 µm and the amorphous silica
content.
Figure 6-29: Surface plot of the logarithm of viscosity at 50 s -1

vs the clay content and the pH
169
vs the particle sizes smaller than 9.86 µm and the pH
The exact relationship with pH and the viscosity is clearly

complicated. It is likely that an intermediate pH leads to the highest
viscosity given the previous results; this is further supported by
Figure 6-31 which clearly shows that a low to intermediate pH results
in a high viscosity in low Kaolinite slurries whilst a very low pH
dramatically increases the viscosity in high Kaolinite samples.
170
vs the Kaolinite content and the pH
Further work is possible and needed in this area. The fitted graph
of the logarithm of viscosity shows sta tistical significance and a high
R2 value when made dependent on the aforementioned variables. More
work could be performed varying the particle size distribution
variables (i.e. taking other values than -9.86, -38, -53 µm) and
selecting alternative geological components. In addition, alternative
rheological parameters could, and should, be investigated including
the yield stress and apparent viscosities at different shear r ates.
Although no clear link between the mineralogy and the slurry

rheology has been found they are obviously related. In this study
neither the pH nor the particle size was standardised. It is known that
the pH, the particle size and size distribution ef fect mineral slurry
rheology (Becker et al., 2013; Boger, 2009; Burdukova, 2008; Coussot
et al., 2002; Cruz and Peng, 2016; Farrokhpay et al., 2016; Ndlovu,
2014; Ndlovu et al., 2014). Future work should standardise both the
171
particle size and the pH to allow the mineralogy to be explored with
greater precision.
6.7 Summary
It was hypothesised that concentration and part icle size would
increase the apparent viscosity and yield stress. These hypothesises
were correct as seen in sections 6.2, 6.3 and 6.6; an additional
unexpected effect was that concentration was linked to time
dependency (see section 6.1). A recommended maximum soli ds
concentration cannot be made because the high viscosity and yield
stresses occur at vastly different concentration for different samples
(see section 6.4). Currently, all that can be recomm ended is that the
solids concentration be reduced immediately by adding water if
undesirable rheological behaviour is observed. It is clear that
(ignoring the time dependency) the suspensions at each concentration
studied were plastic and well described by the Casson model. As such,
if a suitable chemical dispersant can be found the viscosity of the
problematic samples can be greatly reduced.
It was also been shown that there is a link between the rheology
and the geology. As can be clearly seen in section 6.4 certain samples
(particularly argillic and lower epithermal) are substantially more
viscous than other samples. However, this study found no clear culprit
for the rheological problems in the QXRD results. A multilinear
regression found that there was a link between the minerology and the
rheology, but it was complicated as it was linked with the logarithm of
viscosity rather than directly with the viscosity itself, and the
mineralogy heavily interacted with the particle size an d the pH.
Neither the particle size nor pH was standardised in this study; this
study was more interested on the rheology of the slurry depended on
the ore as it was provided rather than how it would have been if it had
been altered. This is because at site they currently do not use any
grinding aids and the goal of this study was to understand what the
problems currently are. Certain samples are naturally finer, and their
172
pH’s varied from being extremely acidic to mildly bas ic. The particle
size of certain samples was varied by grinding and found that this
always increased the viscosity, but it was ground for a set amount of
time rather than aiming for a particular particle size. Future
experimental work should control for par ticle size, particle size
distribution and pH as this may reveal more clearly what in the
mineralogy is affected the rheology.
The time dependency was complicated. As expected, all samples

displayed time dependent shear thinning during the initial yield s tress
tests. However, some samples displayed time dependent shear
thickening over the course of the ramp particularly at high shear rates.
This was the opposite of what was observed in the literature; in
previous studies with time dependent shear thickenin g and shear
thinning behaviour the suspensions thickened at the lower shear
stresses and thinned at higher shear stresses.
173
Chapter 7
Analysis of slurry stability
The sedimentation rate was predicted to, and was found, to
depend on the solids concentration, the alteration domain, and the
particle size. In addition, it was found that some sedimentation tests
had a time delay before sedimentation began.
Slurry didn’t just settle; it was also observed to swell. In some

cases, the slurry swelled and maintained a hig her level. However, in
other cases the swelling effect was followed by a decrease under
sedimentation.
NTS110 (a Lower Epithermal sample) is one example of a

swelling slurry. At 41% solids concentration (v/v) it exhibited the
simple swelling described above; the mudline increased from 204 mL
to 207 mL in a 250 ml measuring cylinder. In addition, a further 1 mL
of clear liquid collected on the surface of the mud giving the slurry a
total increase in volume of 4 mL. This occurred over two hours and
thirteen minutes inside a measuring cyli nder. At the lower
concentrations, the slurry exhibited a more complicated behaviour as
shown in section 5.4.
A goal of this project was to investigate the link between

sedimentation and rheology. In particular , it was hypothesised that
sedimentation rate could be used as a proxy for rheology.
All of the sedimentation data is included in Appendix F and

section 5.4.
7.1 Concentration effect

Increasing the concentration decreases the sedimentation rate; an
example is given in Figure 7-1. The rates are presented as mL/hour to
174
ease interpretation. This was seen in all of the samples. As discussed
in section 2.11, this could be due to either the increase in yield stress,
viscosity or hindered settling or a combination of the three.
Figure 7-1: Average sedimentation rate in relation to solid

volumetric concentration for NTS008
7.2 Alteration domain

Alteration domain had an impact on the slurry stability. The
lower clay samples, such as the outer biotite, would settle far more
readily than the higher clay content. However, additional analysis is
required and a non-rheological link between alteration domain and
sedimentation rate has not been ruled out.
Unfortunately, no sedimentation results were produced for each

sample at a single concentration. This was because the highest
concentration in which significant sedimentation rate (>2 mL/hour)
occurred is vastly different between the different samples as seen in
Figure 7-2. Future tests should use a helical flow rheometer to test the
175
rheology of the samples at a low concentration of 20% (v/v) and more
sedimentation data be developed.
Figure 7-2:Comparison of highest concentrations when

sedimentation above 2 mL / hour was observed
7.3 Particle size and part icle size distribution

Grinding significantly slowed the slurries’ sedimentation rate as
shown in Figure 7-3. Sedimentation occurs due to a difference in
density, whilst the rate of sedimentation is dependent upon the
particle size and the viscous forces. A higher viscosity will decrease
sedimentation, and smaller particles settle slower than larger ones. In
addition, these slurries, as descri bed in chapter 5, exhibit yield stress.
As smaller particles have smaller buoyancy and s ettling forces, the
yield stress is less likely to be overcome.
176
Figure 7-3: Sedimentation rate in relation to solid vo lumetric
concentration for argillic ore samples.
As can be seen in Figure 7-3, higher concentrations significantly

reduce the sedimentation rate as do smaller particle sizes. All samples
are classified as argillic and therefore would have similar geological
properties.
177
7.4 Swelling
Figure 7-4:An example of a swelling sample, during

sedimentation measurement over 3.4 hours
As can be clearly seen in Figure 7-4, the mudline increases from

202 mL to 204 mL over 3.4 hours.
Swelling was observed in:
• All of the ground samples except NTS108
• All of the non (or slow) settling samples that indicated

time dependency at high shear rates.
As such, it is likely that the samples are still undergoing a

chemical restructuring after several minutes of extremely high
shearing. Such processes may affect the mill operation since, time
effects are by definition dynamic whilst, at the time of writing, the
industry’s current process control standards are static. This effect
should be explored further both in a deeper analysis of the results
produced during the course of this thesis and through additional
experimental work.
Before additional analysis is conducted the level of experimental

errors needs to be established. As no repeats were pe rformed on the
swelling samples the repeatability is unknown.
Table 7-1 summarises the swelling behaviour of the different

samples.
178
Table 7-1: Summary of swelling behaviour amongst the different
domains
Domain Swelling behaviour
Advanced argillic The advanced argillic sample NTS037

did not swell.
Argillic Only NTS109 at 41% solids

concentration (v/v) swelled. NTS109
was the most viscous argillic sample.
Upper epithermal All upper epithermal samples swelled.
Silica breccia Two out of three samples swelled

(NTS010 and NTS106 swelled; NTS128
did not). Of these three samples, the
least viscous was the only sample not
to swell.
Lower epithermal Lower epithermal sample NTS110

swelled significantly both ground and
unground. NTS047 also swelled whilst
NTS240 was non-swelling. NTS110 was
the most viscous lower epithermal
sample and NTS240 the least.
Outer biotite The ground outer biotites (NTS044

and NTS054) and NTS046 swelled;
these were the most viscous outer
biotite samples.
Inner biotite Inner biotite was non-swelling.
179
7.5 Sedimentation rate and time delay
Several samples experience a time delay in the sedimentation.
This is an interesting observation , considering that some of the
samples which demonstrated this effect had very low torque readings
indicating a small viscosity. NTS008 at 55% showed, as can be clearly
seen in Figure 7-5, a large time delay before sedimentation occurre d;
however, the RM180 was incapable of generating reliable flow curves
on this sample. Further work is required to investigate this
relationship using more precise rheological devices; for this work I
offer the hypothesise that the time delay was caused by a high yield
stress. This effect was not noticed in the previous rheological studies
conducted in the previous studies conducted at the University of
Queensland’s experimental mine site (Shi, 2016).
Figure 7-5:The volume of the mud-line over time in the

sedimentation measurement for NTS008 at 55% solid volumetric
concentration.
180
There is a clear delay in the sedimentation where the mudline
drops by 2mL over the first 12 minutes, and then drops another 8 mL
over the next 12 minutes.
Figure 7-6: Time delay in the sedimentation process for NTS0 08

sample after one, twelve and twenty-four minutes after starting the
sedimentation experiment.
As can be seen in Figure 7-6, there is a clear time delay in the

sedimentation process.
7.6 Apparent viscosity and sedime ntation rate

Higher apparent viscosities significantly slow the rate of
sedimentation. All fast settlin g samples (samples which settled more
than 1 mL in 10 minutes) had low apparent viscosities. This was
expected; a higher viscosity is known to slow the s edimentation rate.
Although the link between viscosity and sedimentation isn’t novel,
this data will allow the development of a proxy which can be used on
site to quickly evaluate qualitatively the viscosity of the sample.
Figure 7-7 and Figure 7-8 demonstrate the strong link between
sedimentation rate and viscosity. Excluded from both figures are:
• Swelling samples
• Non-settling samples
• Samples in which the viscosity could not be accurately

measured
181
Figure 7-7: Slurry viscosity in relation to sedimentation rate for
the entire ore samples
Figure 7-7 indicates the strong dependence that sedimentation

has on viscosity.
182
Figure 7-8: Linear regression on the data of slurry viscosity in
relation to sedimentation rate for the entire ore samples
Both Figure 7-7 and Figure 7-8 show the entire ore body rather
than any particular domain. Figure 7-8 indicates a possible proxy
model for viscosity and sedimentation rate. There was lar ge variance
not just between the domains but also within the domains. Further test
work should be performed at low concentrations using a helical flow
rheometer to generate extra data.
183
Despite the need for further experimental work additional
analysis can be performed. It is known that sedimentation and the
viscosity are strongly linked with yield stress (sections 2.2.1, 2.5,
and 2.11), the concentration (sections 6.2 and 7.1) and the particle
size (sections 6.3 and 7.3). It also is not clear what shear rate should
be chosen as the representative for evaluating the apparent viscosity.
Apparent viscosities at other shear rates should be investigated; the
apparent viscosity at a shear rate of 50 s -1 was selected more due to
the limitations of the rheometer not being able to accurately determine
the shear stress for low viscosity tests.
Additional work should see if the sedimentation correlates well

with any available hindered settling equations.
7.7 Slurry stability and time -dependency

The samples which indicated significant time dependent shear
thickening were all linked to the slurry stability. As expected, some
fast settling slurries indicated an increase in s hear stress after an
initial decrease. This was due to the sample beginning to settle, which
reduced the slurry concentration and thus the viscosity. Eventually,
the settling caused the development of a thickened core around the
base of the cup which the bob would catch on leading to erroneously
high shear stress readings.
The other, less expected effect was slurry swelling leading to an

increase in viscosity. Many of the highly viscous time dependent
slurries underwent swelling. The use of anti -swelling agents on ore
which contains swelling clays may m ake a dramatic difference on mill
performance and should be investigated as future work.
Another point of interest is that even low clay material, such as

NTS110, underwent swelling. It was expected that the high clay
materials could swell; it was unexpect ed that the low clay content
could swell.
184
Chapter 8
Conclusion and further
work
This chapter provides a summary of the research findings, the
contributions to knowledge and makes recommendations for future
work.
8.1 Summary of equipment design

To measure the rheology of settling slurries coaxial cylinders
can be modified with the addition of axial flow via a sump-pump
testing rig. These modified rheometers have been referred throughout
this thesis as helical flow rheometers.
A helical flow rheometer was designed and commissioned; it used

significantly smaller volume of sample (4 L) compared to previous
testing rigs (20 L) used at the JKMRC pilot plant. The smaller sample
volume also made temperature control possible, a major flaw in the
previous constructions. Unfortunately, it was under-used, due to the
availability of some ore samples and the core rhe ometer unit (HAAKE
Viscotester IQ) malfunctioning. Hence, the project developed towards
performing tests using a Mettler RM180 fitted with a coaxial cup
geometry to keep consistent measurements.
Despite the reduced results from the testing rig, experience with
its design and operation have been acquired. It has been found that:
1. the successful operation of a helical flow rheometer is

critically dependent on the concentration of the suspension.
185
In the case of Lihir ore slurry three phases of the
suspension were observed (with increasing concentration):
a. a fast-settling low concentration slurry;
b. a slow settling high concentration slurry;
c. a very slow settling, high concentra tion paste which was

more solid with liquid-like behaviour than a liquid with
solid-like behaviour.
Sedimentation, yield stresses and high viscosity can overload the

pump or block the pipes causing the testing rig component to fail. The
ideal method of loading the slurry into the testing rig is to add a
slurry which isn’t so viscous that the p ump overloads whilst
simultaneously being slow settling, so it doesn’t block the pipes. The
intermediate option (b) was the most successful in this project.
2. Preliminary rheological and sedimentation investigations

should be conducted using a coaxial cylinde r geometry
prior to designing the flow rheology tests. These tests will
allow the identification of the ideal concentration to load
the testing rig. Furthermore, if t he slurry is sufficiently
slow settling the results can be used as an alternative to
performing the helical flow tests.
8.2 Summary of major research findings

In this study, 24 mining domain samples collected from Lihir
operation had their rheology and sed imentation investigated. The
rheological study was carried out using a Mettler RM180 rheom eter
with a wide-gap concentric cylinder geometry. The sedimentation
study was performed using a 250 mL measuring cylinder with a video
camera to measure particle sedimentation rates, using the same
slurries immediately following the rheological measuremen ts. A Total
of 105 rheograms were produced on fresh Lihir ore suspensions and an
186
additional 12 results of duplicated tests are presented on re -used
samples. The major findings are summarised below:
1. The apparent viscosity at a shear rate of 50 s -1 and the

yield stress were found to increase with higher solids
concentration for any given sample. However, a high yield
stress emerges at vastly different levels for differen t
samples; for example, NTS110 at a concentration of 36%
(v/v) has a similar yield stress to NTS240 at 51% (v/v; 15.4
vs 16.5 Pa respectively). From the results, all sample
clearly indicated yield stress at sufficient concentration
along with time-independent shear thinning behaviour. The
only result with no measured yield stress was NTS037 at
46% solids concentration (v/v).
2. A smaller particle size and more narrow size distribution

increases the viscosity and the yield stress, and reduces the
sedimentation rate of the sample. The particle size of the
provided samples varied widely. It is likely th at this is
strongly linked to the complicated observed effects. The
particle size of certain samples was also varied by grinding
for a set amount of time; this alwa ys increased the
viscosity.
3. There is a link between rheology and geology. The rheology

and sedimentation data all show that these were linked to
the ore domain. However, significant variation in the
rheological characteristics was observed within the
domains. Argillic samples are more likely to be highly
viscous; however, the most viscous sample was a particular
Lower Epithermal sample NTS110. In this study each
sample was analysed by QXRD which indicated the ore’s
geological composition including its clay content.
Statistical investigation indicated that there was no clear
dependence of apparent viscosity on the total clay,
187
kaolinite or amorphous silicates grades. This does not mean
that there was no dependence; future work will be required
to investigate this further.
4. There was a complicated time dependent relationship in the

samples. All samples which could produce measurable shear
stress at 5 RPM displayed time dependent shear thinning at
low shear rates, however, some samples (particularly the
highly viscous) displayed time dependent shear thickening
over the course of the ramp. This anti -thixotropy was more
common in highly viscous samples.
5. The sedimentation rate increased with a decrease in slurry

concentration and an increase in particle size. It wa s
clearly dependent on the alteration domain. The rate of
sedimentation is slower at higher concentrations and higher
viscosities. All samples exhibited slow or non -settling at
above 75% solids concentration (wt%). This study was
unable to fully separate out the viscosity effects and the
other likely reasons for reduced sedimentation rate at
elevated concentrations. No proxy model was developed
whereby one could estimate the viscosity from the
sedimentation rate, although a linear regression showing
the correlation between viscosity and sedimentation rate
was produced. Also noted was the n on-linear behaviour in
the sedimentation rate of some samples with respect to
time.
6. Slurry pH may exert an impact on slurry rheology and

sedimentation. It was observed that the pH values of the
slurries were highly variable between samples. It was found
that an intermediate pH (approximately 4) led to an
increase in the apparent viscosity at a shear rate of 50 s -1.
Due to the fact that the pH was merely monitored but not
188
controlled in this experimental program, the effect of pH
on slurry rheology and sedimentation is not conclusive.
7. It was found that the rheology of many of the samples is

well described by the Casson model. As such, the Casson
parameters could be fitted to th e raw data without an
intermediate shear rate calculation step and the shear rate
could then be found through the Casson equation.
8. It was observed that within domains that samples which

swelled were more viscous. The evidence of this swelling
isn’t conclusive and should be investigated in future work.
9. No plant operation suggestions can b e made at this stage.
8.3 Future work

Future work is required to further develop the findings of this
project into building full rheological and sedimentation models of
Lihir ore suspensions. These models could then be applied at the mine
site plant to optimise operations.
The rheological and sedimentation models could be built through

a combination of further analysis and experimental work. The
additional analytical work coul d be based off the current data set and
include:
1. additional analysis of particle si ze effects (vary the

passing size);
2. investigating different components and combinations of

chemicals from the QXRD table; although no clear link
between the clay or amorphous silicate content was found
perhaps after further analysis the ore components whi ch
control the rheology may be discovered. It should however
be noted that these effects may have been hidden by the
lack of standardisation of pH, particle size, and parti cle
189
size distribution and that the exact effects may only be
revealed after further experiments in which these factors
are controlled.
3. The link between sedimentation and the concentration,

rheological characteristics, particle size and mineralogy;
To build this model further test work using the techniques used
in this thesis will need to be conducted:
1. more NTS numbers should be evaluated;
2. different NTS numbers should be blended and their

rheology and sedimentation rates investigated;
3. detailed work should be performed to clarify how pH

affects the rheology and sedimentation rate;
4. the sedimentation and rheology link should be further

investigated by going to lower concentrations using helical
flow rheometry; this may allow the creation of a
sedimentation rate proxy for rheological behaviour.
5. dispersants should be investigated. It has been fou nd that

dispersants alongside pH modifiers can significantly reduce
the viscosity of mineralogical suspensions (see section
2.4.3);
6. temperature effects should be investigated. The samples of

these tests were at room temperature in an air -conditioned
laboratory. Lihir the ambient temperature is often at 40OC;
the temperature may be even higher inside the mill.
7. Particle sizes and distributions should be controlled;
8. and repeats on the swelling samples to determine the

variance in the swelling rate. If the swelling rates found
through the techniques described in this thesis are reliable,
it should be modelled and linked with to the apparent
viscosity, time dependency, and particle size.
190
Other validation work:
1. additional test work should be done to check that de -

ionised water will have a largely similar effect to site
water (see section 4.1 where the assumption is made). The
age of the samples could not be controlled in this project.
Its effect, if there is an effect, should be investigated in
future site-based projects.
2. The Casson model approximation was a consistent, accurate

and automatable method for producing the shear rate
information. More work to demonstrate its utility should be
performed in the similar manner to the other authors
(Estellé et al., 2008; Krieger, 1952; Nguyen and Boger,
1987) where the error in the method should be evaluated
from synthetic rheology data.
Newcrest has performed various characterisations on the

mineralogy of Lihir. This characterisation d ata, together with the
rheology and sedimentation models, should be analysed to develop into
a mine schedule to optimise the metallurgy and economic benefits for
Lihir mine site optimisation. The site-based optimisations could
include ore blending, solids concentration control and the addition of
grinding aids. Implementing this control methodology could be
conducted in conjunction with a PhD program.
The use of slurry rheology within mineral processing is a

worthwhile investment. It offers the opportunity to optimise various
processes leading to triple bottom line returns. The data collected in
this study is a significant first step towards optimising Lihir grinding
operation; we now know what sample is likely to demonstrate highly
viscous (or even solid-like) behaviour and the approximate solid
concentration when this will occur. Furthermore, we now know some
sample will thicken with time and plant operations should con sider
this effect. Future work should focus on furthering our understanding
and finding techniques to reduce unwanted rheological characteristics.
191
References
Akroyd, T.J., 2004. Continuous Flow Rheometry for Settling Slurries, In
School of Chemical Engineering. The University of Adelaide, p. 2004.
Alharbi, F.M., 2016. Helical Flow of Yield Stress Fluids. RMIT University,
College of Science Engineering and Health.
Anon, 2005. xlxtrfun Extra Functions for Microsoft Excel. Advanced
Systems Design and Development, p. Download location of XlXtrFun:
http://www.xlxtrfun.com/XlXtrFun/XlXtrFun.htm.
Anon, 2006. DSR301 Measuring Head Instruction Manual. Anton Paar.
Anon, 2016. HAAKE Viscotester iQ Reference Manual. Thermo Scientific.
Barnes, H., Nguyen, Q., Rotating vane rheometry - a review. Journal of
Non-Newtonian Fluid Mechanics (Netherlands), 2001, 98(1), 1-14.
Barnes, H.A., Shear‐Thickening (“Dilatancy”) in Suspensions of
Nonaggregating Solid Particles Dispersed in Newtonian Liquids. Journal of
Rheology, 1989, 33(2), 329-366.
Barnes, H.A., Thixotropy—a review. Journal of Non-Newtonian Fluid
Mechanics, 1997, 70(1), 1-33.
Barnes, H.A., The yield stress—a review or ‘παντα ρει’—everything flows?
Journal of Non-Newtonian Fluid Mechanics, 1999, 81(1), 133-178.
Barnes, H.A., Carnali, J.O., The vane‐in‐cup as a novel rheometer
geometry for shear thinning and thixotropic materials. Journal of
Rheology, 1990, 34(6), 841-866.
Becker, M., Yorath, G., Ndlovu, B., Harris, M., Deglon, D., Franzidis, J.P.,
A rheological investigation of the behaviour of two Southern African
platinum ores. Minerals Engineering, 2013, 49, 92-97.
Bender, J., Wagner, N.J., Reversible shear thickening in monodisperse and
bidisperse colloidal dispersions. Journal of Rheology, 1996, 40(5), 899-
916.
Bergaya, F., Lagaly, G., Handbook of Clay Science. 2nd ed.. edn. 2013,
Elsevier Science, Burlington.
Bertrand, E., Bibette, J., Schmitt, V., From shear thickening to shear -
induced jamming. Physical review. E, Statistical, nonlinear, and soft
matter physics, 2002, 66(6 Pt 1), 060401-060401.
Boger, D.V., Rheology and the Minerals Industry. Mineral Processing and
Extractive Metallurgy Review, 2009, 20(1), 1-25.
Brandenburg, U., Lagaly, G., Rheological properties of sodium
montmorillonite dispersions. Applied Clay Science, 1988, 3(3), 263-279.
Bridges, E.M., 2014. Kaolinite, In Encyclopedia of Environmental Change,
ed. Matthews, J.A., pp. 600-600.
Buitenhuis, Springer, Negative thixotrop y of polymer solutions. 2. A
systematic study of the time-dependent viscosity of partially hydrolyzed
polyacrylamide. Kolloid-Zeitschrift und Zeitschrift für Polymere., 2003,
281(3), 260-266.
Burdukova, E.B., M.; Ndlovu, B; Mokgethi, B.; Deglon, D., 2008.
Relationship between Slurry Rheology and its Mineralogical Content, In
24th International Minerals Processing Congress , Beijing, China.
192
CEEC, 2016. CEEC Factsheet - Energy Consumed in Comminution, Version 3
ed. Coalition for Energy Efficient Comminution
Chakraborty, A.K., Phase Transformation of Kaolinite Clay. 2013,
Springer, Dordrecht.
Chen, X., Hadde, E., Liu, S., Peng, Y., The effect of amorphous silica on
pulp rheology and copper flotation. Minerals Engineering, 2017, 113, 41-
46.
Cheng, J., Gröbner, J., Hort, N., Kainer, K.U., Schmid-Fetzer, R.,
Measurement and calculation of the viscosity of metals —a review of the
current status and developing trends. Measurement Science and
Technology, 2014, 25(6), 062001.
Couper, J.R., Chemical process equipment selection and design / James R.
Couper ... [et al.]. Rev. 2nd ed.. edn. 2010, Amsterdam ; Boston :
Elsevier/Butterworth-Heinemann, Amsterdam ; Boston.
Coussot, P., Nguyen, Q.D., Huynh, H.T., Bonn, D., Viscosity bifurcation in
thixotropic, yielding fluids. Jo urnal of Rheology, 2002, 46(3), 573-589.
Cruz, N., Peng, Y., Rheology measurements for flotation slurries with high
clay contents – A critical review. Minerals Engineering, 2016, 98, 137-150.
Cruz, N., Peng, Y., Farrokhpay, S., Bradshaw, D., Interactions o f clay
minerals in copper–gold flotation: Part 1 – Rheological properties of clay
mineral suspensions in the presence of flota tion reagents. Minerals
Engineering, 2013, 50-51, 30-37.
Cruz, N., Peng, Y., Wightman, E., Xu, N., The interaction of clay mineral s
with gypsum and its effects on copper –gold flotation. Minerals
Engineering, 2015, 77, 121-130.
DIN, 2008. DIN 53019-1, In Viscometry-Measurement of viscosities and
flow curves by means of rotational viscometers -Part 1: Principles and
geometry of measuring system, p. 22.
Einstein, A., Eine neue Bestimmung der Moleküldimensionen. Annalen der
Physik, 1906, 324(2), 289-306.
Estellé, P., Lanos, C., Perrot, A., Processing the Couette viscometry data
using a Bingham approximation in shear rate calculation. Journ al of Non-
Newtonian Fluid Mechanics, 2008, 154(1), 31-38.
Farr, R.S., Groot, R.D., Close packing density of polydisperse hard
spheres. 2009.
Farrokhpay, S., Ndlovu, B., Bradshaw, D., Behaviour of swelling clays
versus non-swelling clays in flotation. Minerals Engineering, 2016, 96-97,
59-66.
Fisher, D.T., Clayton, S.A., Boger, D.V., Scales, P.J., The bucket rheometer
for shear stress-shear rate measurement of industrial suspensions. Journal
of Rheology, 2007, 51(5), 821-831.
Gardner, K., Geology & ODK for L ihir Metallurgists. 2016.
Gardner, K., Vollert, L., 2017. ed. Legge, P.W., A meeting between Peter
Legge and Karyn Gardner and Luke Vollert ed, NA.
Hallbom, D.J., 2008. Pipe flow of homogeneous slurry. University of
British Columbia.
Hoffman, R., Explanations for the cause of shear thickening in
concentrated colloidal suspensions. J. Rheol., 1998, 42(1), 111-123.
ISO, 1995. ISO 3219.
James, B., 2015. Mixing 101: Optimal Tank Design.
193
Jeong, S.W., Locat, J., Torrance, J.K., Leroueil, S., Thixotropic and anti -
thixotropic behaviors of fine-grained soils in various flocculated systems.
Engineering Geology, 2015, 196, 119-125.
Kawatra, S.K., Eisele, T.C., Rheological effects in grinding ciruits. Int. J.
Miner. Process., 1988(22), 251-259.
Klein, B., Laskowski, J.S., Partridge, S.J., A new viscometer for rheological
measurements on settling suspensions. Journal of Rheology, 1995, 39(5),
827-840.
Klimpel, R.R., The influence of material breakage properties and
associated slurry rheology on breakage rates in the wet g rinding of
coal/ores in tumbling media mills. Particulate Science and Technology,
1984, 2(2), 147-156.
Krieger, I.M., Direct Determination of the Flow Curves of Non -Newtonian
Fluids. Journal of Applied Physics, 1952, 23(147(1952)).
Krieger, I.M., Elrod, H., Direct Determination of the flow curves of non-
Newtonian fluids II. Shearing rate in the concentric cylinder viscometer.
Journal of Applied Physics, 1953, 24(134).
Kursun, I., Particle size and shape characteristics of kemerburgaz quartz
sands obtained by sieving, laser diffraction, and digital image processing
methods. Mineral Processing and Extractive Metallurgy Review, 2009,
30(4), 346-360.
Lagaly, G., Principles of flow of kaolin and bentonite dispersions. Applied
Clay Science, 1989, 4(2), 105-123.
Larson, R.G., Wei, Y., A review of thixotropy and its rheological modeling.
J. Rheol., 2019, 63(3), 477-501.
Lee, C.-H., 1989. Modeling of batch hindered settling, ed. Luckie, P.T.
ProQuest Dissertations Publishing.
Li, C., 2016. An investigation of flotatio n froth rheology. The University of
Queensland, Sustainable Minerals Institute.
Li, C., Farrokhpay, S., Shi, F., Runge, K., A novel approach to measure
froth rheology in flotation. Minerals Engineering, 2015, 71, 89-96.
Lin, Y., Cheah, L.K.-J., Phan-Thien, N., Khoo, B.C., Effect of temperature
on rheological behavior of kaolinite and bentonite suspensions. Colloids
and Surfaces A: Physicochemical and Engineering Aspects, 2016, 506, 1-5.
Maranzano, B.J., Wagner, N.J., The effects of interparticle interaction s
and particle size on reversible shear thickening: Hard-sphere colloidal
dispersions. Journal of Rheology, 2001, 45(5), 1205-1222.
May, M., 2015. Vicometers Vs Rheometers. Lab manager.com.
Mendoza, C., A simple semiempirical model for the effective viscos ity of
multicomponent suspensions. Rheologica Acta, 2017, 56(5), 487-499.
Merrill, J., Voisin, L., Montenegro, V., Ihle, C.F., McFarlane, A., Slurry
rheology prediction based on hyperspectral characterization models for
minerals quantification. Minerals Engineering, 2017, 109, 126-134.
Mewis, J., Wagner, N.J., Thixotropy. Advances in Colloid and Interface
Science, 2009, 147(C), 214-227.
Mueller, S., Llewellin, E.W., Mader, H.M., The rheology of suspensions of
solid particles. Proceedings of the Royal Societ y A, 2010, 466(2116), 1201-
1228.
194
Napier-Munn, T.J., 2017. An introduction to comparative statistics and
experimental design for minerals engineers. The University of Queensland
(Course Notes).
Napier-Munn, T.J., Morrell, S., Morrison, R.D., Kojovic, T., Mineral
Comminution Circuits Their Operation and Optimisation . 1996.
Ndlovu, B., Forbes, E., Farrokhpay, S., Becker, M., Bradshaw, D., Deglon,
D., A preliminary rheological classification of phyllosilicate group
minerals. Minerals Engineering, 2014, 55, 190-200.
Ndlovu, B., Forbes, E., Farrokhpay, S., Becker, M., Bradshaw. D., Deglon,
D., , A preliminary rheological classification of phyl losilicate group
minerals. Miner. Eng., 2014, 55(190-200).
Nguyen, Q., Boger, D., Characterization of yield stress fluids with
concentric cylinder viscometers. An International Journal of Rheology,
1987, 26(6), 508-515.
Nguyen, Q.D., Boger, D.V., Yield Stress Measurement for Concentrated
Suspensions. Journal of Rheology, 1983, 27(4), 321-349.
Nguyen, Q.D., Boger, D.V., Thixotropic behaviour of concentrated bauxite
residue suspensions. J. Rheol., 1985, 29(null), 335.
Nguyen, Q.D., Boger, D.V., 1992. Measurin g the Flow Properties of Yield
Stress Fluids, In Annu. Rev. Fluid. Mech., pp. 47-88.
Nguyen, Q.D., Devasagayam, C., Bown, D.J., Development of an on-line
flow rheometer for mineral slurries. Mineral Processing and Extractive
Metallurgy Review, 1999, 20(1), 75-91.
Otunniyi, I.O., Exploiting Image Processing and Analysis Versus Other
Techniques for Composite Fines Particle Size and Shape Distribution
Analysis. Particulate Science and Technology, 2013, 31(6), 585-589.
Ovarlez, G., Bertrand, F., Coussot, P., Xavier, C., Shear-induced
sedimentation in yield stress fluids. Journal of Non -Newtonian Fluid
Mechanics, 2012, 177-178, 19-28.
Pashias, K., Boger, D.V., Summers, J., Glenister, D.J., A fifty cent
rheometer for yield stress measurement. J. Rheol, 1996, 40(null), 1179.
Quemada, D., Rheology of concentrated disperse systems and minimum
energy dissipation principle. An Internation al Journal of Rheology, 1977,
16(1), 82-94.
Rapp, B.E., Microfluidics : modeling mechanics and mathematics . First
edition.. edn. 2017, William Andrew, Amsterdam, Netherlands.
Reeves, T.J., On-Line Viscosity Measurement Under Industrial Conditions.
Coal Preparation, 1990, 8(3-4), 135-144.
Schubert, H., On the optimization of hydrodynamics in fine particle
flotation. Miner. Eng., 2008, 21(12-14), 930-936.
Shabalala, N.Z.P., Harris, M., Leal Filho, L.S., Deglon, D.A., Effect of slurry
rheology on gas dispersion in a pilot-scale mechanical flotation cell.
Minerals Engineering, 2011, 24(13), 1448-1453.
Shah, A., 2007. Rheology of shear thicken ing mineral slurries, ed. Shah, A.
Shewan, H.M., Stokes, J.R., Analytically predicting the viscosity of hard
sphere suspensions from the particle size distribution. Journal of Non -
Newtonian Fluid Mechanics, 2015, 222, 72-81.
Shi, F., 1995. Slurry Rheology And Its Effects On Grinding. University of
Queensland.
195
Shi, F., Determination of ferrosilicon medium rheology and stabilit y.
Minerals Engineering, 2016, 98, 60-70.
Shi, F.N., 2017a. Lihir grinding circuit survey, ed. Legge, P.W., Personal
communication ed.
Shi, F.N., 2017b. The solid concentration within Lihir grinding circuits.
Shi, F.N., Napier-Munn, T.J., Estimation of shear rates inside a ball mill.
International Journal of Mineral Processing, 1999, 57(3), 167-183.
Shi, F.N., Napier-Munn, T.J., Effects of slurry rheology on industrial
grinding performance. Int. J. Miner. Process., 2002, 65, p125-140.
Shi, F.N., Zheng, X.F., The rheology of flotation froths. International
Journal of Mineral Processing, 2003, 69(1), 115-128.
Smith, P.A., Rector, D.R., Shekarriz, A., Microstructural and Rheological
Characterization of Colloidal Aggregates of Nuclear Waste Slurries.
Mineral Processing and Extractive Metallurgy Review, 2000, 20(1), 311-
324.
Tadros, T.F., Rheology of dispersions principles and applications. 2010,
Wiley-VCH, Weinheim.
Trottier, R., Dhodapkar, S., Wood, S., Particle sizing across the CPI: this
survey of modern measurement technologies demonstrates how selection
criteria vary by application.(Solids Processing). Chemi cal Engineering,
2010, 117(4), 59.
Van Wazer, J.R., Viscosity and flow measurement : a laboratory handbook
of rheology. 1963, Interscience Publishers, New York.
Wagner, N.J., Brady, J.F., Shear thickening in colloidal dispersions.
Physics Today, 2009, 62(10), 27-32.
Wills, B.A., Napier-Munn, T.J., Wills' Mineral processing technology : an
introduction to the practical aspects of ore treatment and mineral
recovery. 7th ed. / Barry Wills, Tim Napier -Munn.. edn. 2006, Butterworth-
Heinemann, Oxford.
Yue, J., Klein, B., Influence of rheology on the performance of horizontal
stirred mills. Minerals Engineering, 2004, 17(11), 1169-1177.
196
Appendixes Overview
• Appendix A connects the sequence list to the NTS###,
whether it used the large or small bob and its solid
concentration;
• Appendix B provides the summarises of the rheology and

sedimentation for each NTS###. It is sub -divided into
subsections for each domain;
• Appendix C contains the statistical results;
• Appendix D provides the particle sizing data;
• Appendix E is a data processing exampl e;
• Appendix F and Appendix G are in separate pdf files. They

contain the sedimentation curves and the flow curves
respectively.
197
Appendix A
Sequence list
The sequence list is provided below. The tables in Appendix F
and Appendix G are in the sequence order.
Table A-1: Sequence list
Solid Solid
Sequence concentration concentration

(SEQ) Domain NTS Bob (wt%) (v/v)
Advanced
158 Argillic NTS037 Large 74.8% 52.3%
Advanced
Advanced
Advanced
Advanced
NTS111
after 5
minutes
Upper of
180 epithermal grinding Small 75.0% 52.6%
Upper NTS111
181 epithermal after 5 Small 70.0% 46.4%
198
Solid Solid

minutes
of
grinding
NTS111
after 5
minutes
Upper of
182 epithermal grinding Large 70.0% 46.4%
NTS111
after 5
minutes
Upper of
NTS111
after 5
minutes
Upper of
NTS111
after 5
minutes
Upper of
NTS111
after 5
Upper minutes
189 epithermal of Large 70.0% 46.4%
199
Solid Solid

grinding
and an
additional
30
minutes
of mixing
Upper
193 epithermal NTS008 Small 75.0% 52.6%
Upper
195 epithermal NTS008 Large 70.0% 46.4%
Upper
Upper
Upper
Outer
203 biotite NTS100 Large 75.0% 52.6%
Outer
Outer
Outer
215 Argillic NTS049 Small 75.0% 52.6%
200
Solid Solid

Outer
224 biotite NTS054 Small 74.9% 52.4%
Outer
Outer
Outer
Outer
NTS108
after
additional
4 minutes
236 Argillic of mixing Large 54.9% 31.0%
238 Inner NTS097 Small 75.0% 52.6%
201
Solid Solid

biotite
Inner
Inner
Inner
Inner
Outer
NTS108
after 5
minutes
of
247 Argillic grinding Small 65.9% 41.8%
NTS108
after 5
minutes
of
248 Argillic grinding Small 62.0% 37.7%
NTS108
after 5
minutes
of
249 Argillic grinding Large 58.0% 33.8%
202
Solid Solid

NTS108
after 5
minutes
of
NTS108
after 5
minutes
of
NTS108
after 5
minutes
of
Na Argillic grinding Na 39.0% 19.1%
Outer
Outer
Outer
Outer
Silica
263 Breccia NTS128 Small 75.0% 52.6%
265 Silica NTS128 Large 70.0% 46.4%
203
Solid Solid

Breccia
Silica
267 Breccia NTS128 Large 65.0% 40.8%
NTS044
after 5
minutes
Outer of
269 biotite grinding Large 66.0% 41.8%
NTS044
after 5
minutes
Outer of
270 biotite grinding Small 70.0% 46.4%
NTS044
after 5
minutes
Outer of
NTS044
after 5
minutes
Outer of
NTS044
after 5
Outer minutes
273 biotite of Large 51.3% 28.0%
204
Solid Solid

grinding
Inner
Inner
Inner
Lower
284 Epithermal NTS240 Small 73.5% 50.7%
Lower
288 Epithermal NTS240 Large 69.4% 45.6%
Lower
Lower NTS110 5
298 Epithermal min Small 65.0% 40.8%
Lower NTS110 5
299 Epithermal min Small 60.0% 35.7%
Lower NTS110 5
300 Epithermal min Large 55.0% 31.2%
205
Solid Solid

Lower NTS110 5
Lower NTS110 5
Lower
Lower
Lower
Lower
Lower
NTS097
after 5
minutes
Inner of
NTS097
after 5
minutes
Inner of
312 Inner NTS097 Large 58.0% 33.8%
206
Solid Solid

biotite after 5
minutes
of
grinding
NTS097
after 5
minutes
Inner of
NTS097
after 5
minutes
Inner of
NTS054
after 5
minutes
Outer of
NTS054
after 5
minutes
Outer of
NTS054
Outer after 5
321 biotite minutes Large 56.6% 32.6%
207
Solid Solid

of
grinding
NTS054
after 5
minutes
Outer of
NTS054
after 5
minutes
Outer of
NTS054
after 5
minutes
Outer of
Lower
Lower
Lower
Lower
335 Lower NTS047 Large 55.0% 31.2%
208
Solid Solid

Epithermal
Silica
Silica
Silica
Upper
Upper
Upper
Upper
Upper
Silica
Silica
Silica
Silica
209
Solid Solid

Upper
Upper
Upper
Upper
Upper
Outer
Outer
Outer
Outer
Inner
Inner
Outer
377 Outer NTS045 Small 75.0% 52.6%
210
Solid Solid

biotite
Outer
Outer
Outer
Outer
Outer
Outer
211
Solid Solid

212
Appendix B
Tables of rheology and sedimentation
data
A.1: Advanced argillic
Table B-1 shows the results for NTS037. SEQ numbers 158, 168 and 170 were used in subsequent
analysis. SEQ159 and SEQ160 were conducted on the s ame sample to investigate possible time effects.
As can be seen there was little to no variation in their rheology between t hese tests; each also
reported very little change in the area underneath their curves between the ramp up and ramp down
leading to the conclusion that advanced argillic is time independent. SEQ170 is presented here,
though, its rheological results are qual itative as the torque readings were all beneath the
manufacturers minimum torque reading or Taylor vortices were calculated. The val ue was therefore
calculated from the torque readings below the manufacturer’s recommendation. SEQ170 was not used
in any further rheology analysis because of the qualitative nature of the data; it is presented here
purely qualitatively.
213
Table B-1: NTS037 results
SEQ Solids pH Apparent R2 Yield Infinite Time Sedimentation Swelling

number concentration viscosity stress viscosity dependency rate
at 50 s-1 (mL/hour)
(v/v) (Pa) (mPa.s) (% change
(mPa.s) in area
under
curve)
158 52% 1.38 249 0.996 2.6 73 1% NA NA
159 52% 1.38 231 0.996 2.9 57 0% NA NA
160 52% 1.38 236 0.997 2.6 67 2% NA NA
168 46% 1.33 123 0.970 0.0 123 -4% 1.8 NA
170 38% 1.42 46 0.912 0.0 46 Na 7.8 NA
A.2: Argillic
Table B-2 shows the results for NTS038; all sequences were used in subsequent rheological
analysis. SEQ389 gave a relatively low R2 value, however, this could be converted into a p-value of
1.5E-06 v a l u e i n d i c a t i n g a v e r y h i g h p r o b a b i l i t y t h a t t h e f l o w c u r v e w a s r e a s o n a b l y a c c u r a t e .
214

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
3942 53% 1.83 7192 0.999 168.5 716 Na NA NA
387 46% 1.76 1616 0.997 38.8 153 18% NA NA
388 41% 1.93 627 1.000 18.0 367 6% 0 None
395 36% 1.99 213 0.998 6.1 13 2% 0 None
389 31% 2.03 75 0.932 2.6 2 -12% 2.7 None
NTS049 is described by Table B-3; only SEQ215, SEQ216, and SEQ217 could be used for
subsequent rheological analysis. The apparent viscosity reported for SEQ218 and SEQ221 are both
qualitative, based off using the unrecommende d range of the rheometer. They are listed here for
demonstration purposes and are not used in any subse quent analysis. Using the Casson approximation
2 Overloaded, and the machine shopped early. As there was no down curve
215
on these low viscosity equations indicate that NTS049 becomes Newtonian by 31% solids
concentration

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
215 53% 2.91 1727 0.998 46.6 122 -11% NA NA
216 46% 2.75 518 0.998 10.8 65 15% 0.5 None
217 41% 2.95 166 0.997 3.7 19 10% 3.0 None
2183 36% 2.95 59 0.6 22 Na 25.8 None
2214 31% 5.47 43 0 43 Na 49.2 None
3 Very interesting curve. Time delay followed by a high speed slowing back down. Took sedimentation rate as being for the first hour
4 Settling slows down significantly after 30 minutes
216
NTS108 as provided results are described in Table B-4 and its ground results Table B-5. As can
be seen the samples are slow settling and non -swelling. Although SEQ236 and SEQ251 indicate a low
R2, the associated p-values are 0.0012 and 5.2E-07 respectively indicating a reasonabl y close fit for
rheological purposes. They are subsequently used in later analysis. NTS108 after five minutes of
grinding has a 19% concentration which was used to evaluate the sedimentation rate.

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
231 46% 2.26 1795 0.999 40.5 193 29% NA
232 41% 2.24 634 0.999 20.2 26 16% 0.0 None
2335 41% 2.24 825 0.999 27.1 30 11% 0.0 None
235 36% 2.31 219 0.998 6.7 11 11% 1.2 None
236 31% 2.35 78 0.766 1.7 9 -17% 3.4 None
5
SEQ233 is the same sample as SEQ232 except with an additional 4 minutes of mixing.
217
Table B-5: NTS108 Grind for 5 minutes

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
247 42% 3.17 3364 0.995 91.5 232 39% NA
248 38% 3.16 1120 0.997 32.9 61 32% NA Na
249 34% 3.19 570 1.000 21.7 9 12% NA Na
250 30% 3.22 222 0.997 7.8 6 4% NA Na
251 27% 3.22 104 0.927 3.4 4 -1% 1.1 Na
Na 19% 3.27 Na Na Na Na Na 15.3 None
Table B-6 and Table B-7 contain NTS109 and NTS238 samples respectively. There was some
indication of swelling in NTS109 at 41% concentration; note, the mudline did not move yet 1 mL of
clear liquid formed on top of the mud layer. SEQ400 used the same sample as SEQ399; it was tested as
a side experiment to see how viscosity would change over several loop tests. The results are reported
in section 6.1 but is otherwise not used in any other test.
218
Table B-6: NTS109

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
399 53% 2.14 9889 0.995 265.9 703 Na NA NA
4006 53% 2.14 10139 0.987 235.1 1032 8% NA NA
393 46% 2.26 2189 1.000 64.1 121 20% NA NA
1 mL of
396 41% 2.18 900 1.000 30.1 30 11% 0.0 liquid
397 36% 2.24 318 0.999 10.3 12 10% 1.1 none
398 31% 2.26 137 0.993 5.1 3 -2% 3.5 none
6 Not used in regression analysis. It used the same sample as 399, it was a side experiment to see how the shear stress changed with additional loops.
219
Table B-7: NTS238

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
293 46% 2.17 1804 0.998 49.7 120 12% NA NA
297 42% 2.26 1076 0.999 37.3 30 8% 0.8 None
294 39% 2.23 568 0.999 18.8 20 12% 2.7 None
295 36% 2.25 297 0.999 9.6 11 5% 4.6 None
296 31% 2.28 98 0.986 3.3 3 0% 6.1 None
A.3: Upper epithermal samples

Table B-8 reports on the results of NTS008; two concentrations (36% and 31%; or SEQ197 and
SEQ198) had viscosities which were too low to be included in an y analysis. SEQ198 the sample had
such a low viscosity that the qualitative estimates of viscosity led to the conclusion that even at very
low rpms Taylor vortices were occurring. SEQ196 indicated a complicated time dependency; the total
volume during the sedimentation test increased by 2 mL, though, the mud layer did slightly decrease.
220
Confusingly, significant time dependency was indicated at 41% and 46% but not at 53%. Sedimentation
did not significantly occur, so it appears to be a physical property of t he suspensions. Further work is
warranted.
221

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
193 53% 6.51 2160 1.000 61.903 127.54 2% NA NA
195 46% 2.63 695 0.990 11.409 126.94 19% NA NA
196 41% 2.64 168 0.994 4.152 14.82 22% 1.5 Yes
1977 36% 2.67 68 0.568 1.807 4.86 22% 8 None
1988 31% 2.72 <44 Na Na Na Na 21 None
Table B-9 and Table B-10 show the as provided and ground for five minutes results for NTS111.
The swelling information is complicated for the ground sample and is displayed in Table B-11. As can
7 As can be seen, NTS008 at 60% gives an unreliably low R2 reading
8 All the data points were below the manual minimum torque; fitting the data to the Casson model
(despite the torque being out of range) found that the
substance was best described as Newtonian and that the entire region was beyond the maximum Taylor number
222
be seen grinding the samples significantly increases the swelling rate. SEQ182 and SEQ189 are
discussed in section 6.2; they use the same sample a s SEQ181 and are therefore not used in any other
subsequent rheology analysis. SEQ343, although it has a poor R2, its results have a low p -value
indicating suitability for further use. SEQ345, SEQ185, and SEQ187 are not used in any rheological
analysis. Unlike many of the other samples, NTS111 indicates time dependent shear thinning at high
concentrations whilst becoming shear thickening at lower concentrations. The a pparent shear
thickening behaviour coincides with the onset of sedimentation. As discussed elsewhere,
sedimentation can cause an apparent shear thickening behaviour as the bob can get caught on the
sediment.
223
Table B-9: NTS111 as provided results

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
346 58% 1.96 2838 0.982 115.1 28 -18% NA NA
1 mL
total
increase
in
341 53% 2 784 0.993 22.5 46 -15% 0.3 volume
342 46% 2.05 371 0.995 1.9 170 34% 5.0 None
3439 41% 2.09 99 0.609 1.1 27 76% 16.0 None
345 36% 2.1 23 Na 17.0 None
9 p value=0.008
224
Table B-10: NTS111 after grinding for 5 minutes

number concentration viscosity stress viscosity dependency rate (mL/hour)
at 50 s-1
(v/v) (Pa) (mPa.s) (% change in
(mPa.s) area under
curve)
180 53% 2.24 3014 0.988 96.8 119 -10% NA Yes
181 46% 2.33 441 0.998 8.7 61 9% 0.0 Yes
182 46% 2.33 552 1.000 14.7 41 14% 0.0 Yes
189 46% 2.33 759 0.999 22.3 41 11% 0.0 Yes
183 41% 2.31 99 0.982 2.7 7 20% 1.7 Yes
185 36% 2.32 39 0.982 0.0 56 Na 8.0 None
1 8 7 10 31% Na 15.8 None
10 Torque readings were too low to estimate
225
Table B-11: Solids concentration and swelling details for NTS111 after grinding for 5 minutes
Solids Swelling
concentration
(v/v)
53% NA
46% Mudd swelled by 2 mL and 1 mL of liquid formed on top
Settled without the mud layer increasi ng. However, the total volume increased by
41% 1 mL
Settled without the mud level increasing. However, the total volume increased by
36% 1 mL
31% No swelling
Table B-12 shows the results for NTS239; the swelling is complicated, so it is described in more
detail in Table B-13. As is seen large levels of swelling can occ ur and swelling can clearly occur
during sedimentation.
226

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
361 53% 4.04 3190 0.987 60.6 470 7% NA NA
355 46% 3.99 1018 0.991 28.6 63 9% 0.0 Yes
356 41% 4.06 368 0.999 10.4 23 11% 0.8 Yes
357 36% 4.07 144 0.978 4.4 7 4% 4.1 Yes
3 5 8 11 31% 4.17 46 Na 25.0 Yes
11 Not reliable
227
Table B-13: Solids concentration and swelling details for NTS239
Solids Swelling
concentration
(v/v)
53% NA
46% Swelled by 2 ml, mud settled by 2 mL
41% Sedimentation occurred but total volume increased by 3 mL
36% Sedimentation occurred; total volume swelled by approximately 2 mL
31% Sedimentation occurred; total volume swelled by approximately 2 mL
228
A.4: Silica Breccia
Table B-14 through Table B-17 show the results for the silica breccia samples (NTS010, NTS106
and NTS128). All were low viscosity slurries which were difficult to measure. SEQ354, SEQ340 and
SEQ267 are qualitative results and are not used for any rheological analysis because either all of their
torque readings were beneath the manu facturers minimum torque reading or Taylor vortices were
calculated.
229
Table B-14:NTS010 results

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
351 53% 2.02 842 0.993 24.4 48.19 -7% 0 Yes
352 46% 2.03 322 0.999 6.5 43.53 29% 0.8 Yes
353 41% 2.11 108 0.982 2.4 11.99 37% 1.0 None
354 36% 2.14 54 0.997 0 54 Na 3.5 None
230
Solids Swelling
concentration
(v/v)
53% Slurry did not settle. 2 mL of liquid formed on top
46% Slurry settled by 1 mL. 2 mL of liquid formed on top
41% None
36% None
231

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
338 53% 2.74 892 0.997 27.0 44 -10% 0.9 1 mL
339 46% 2.71 379 0.996 4.0 111 46% 2.6 none
340 41% 2.72 164 0.997 0 164 Na 5.1 None
232

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
263 53% 5.76 510 0.952 15.7 24 5% 2.2 None
8% None
265 46% 5.82 271 0.998 1.7 113 8.9
267 41% 5.67 84 0.956 0.1 66 Na 15.0 None
233
A.5: Lower Epithermal
Lower Epithermal samples, seen in Table B-18 through Table B-23, were of relatively high
viscosities.
Table B-18 :NTS047 results

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
329 53% 2.9 3902 0.999 110.2 241 9% NA NA
1 mL of
total
332 46% 2.91 1059 0.998 30.0 64 1% 0.2 volume
333 41% 2.97 400 0.999 10.9 27 17% 1.3 None
334 36% 3.03 180 0.992 4.3 17 6% 1.7 None
335 31% 3.02 69 0.876 2.5 1 -10% 6.0 None
234

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
305 46% 3.43 4792 0.992 109.9 499 31% NA NA
304 41% 3.36 1219 0.995 33.3 83 16% 0 Yes
306 36% 3.33 459 0.999 15.4 15 15% 0.0 Yes
307 31% 3.39 181 0.996 6.0 6 4% 0.8 Yes
308 27% 3.42 68 0.544 1.6 7 -10% 6.7 None
235
Solids Swelling
concentration
(v/v)
46% NA
41% Mud swells by 3 mL and an additional 1 mL liquid layer forms on top
36% Mud swells by 2 mL and an additional 2 mL liquid layer forms on top
Slurry settles by 2 mL but the total volume swells by 2 mL due to a liquid layer
31% forming on top
27% None
The sedimentation rates for SEQ298 and SEQ299 (see Table B-21) are for after the slurry had
finished swelling.
236
Table B-21: NTS110 after 5 minutes grinding

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
N A 12 46% NA NA NA NA NA NA NA NA
298 41% 4.75 4406 0.984 122.1 288 24% NA NA
299 36% 4.79 1093 0.960 29.8 75 31% 0 Yes
300 31% 4.57 494 0.998 19.0 7 10% 1.3 Yes
301 27% 4.51 182 0.993 6.6 4 12% 0.6 Yes
302 2 3 % 13 4.69 79 0.896 2.2 5 8% 4.1 Yes
12 No rheology tests were performed as the suspension was too solid to feasibly load the rheometer
13 p value = 3.3E-5
237
Table B-22: Solids concentration and swelling details for NTS110 after five minutes of grinding
Solids Swelling
concentration
(v/v)
46% NA
41% NA
Swell by 3 mL over 224 minutes. An additional layer of liquid formed on top.

Left over night, and it settled to 203, being an increase of 2 mL of mud and 2 mL
36% of a liquid
Swelled by 3 mL before settling. The total volume increase by 4 mL including th e

31% liquid layer formed on top.
Swelled by 2 mL before settling. The total volume increase by 4 mL including the

27% liquid layer formed on top.
23% The mud settled but the total volume increased by 2 mL
238

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
289 59% 2.95 3900 0.980 138.9 95 -9% NA NA
284 51% 2.91 617 0.981 16.5 45 1% 0.8 None
288 46% 2.83 335 0.993 2.4 129 47% 5.1 None
239
A.6: Outer biotite
Outer biotite samples were quite variant as seen in Table B-24 to Table B-33. Of these samples:
• NTS027, NTS045, NTS054 (as provided) and NTS100 underwent no swelling;
• NTSO44 (after grinding for 5 minutes), NTS046, and NTS054 after grinding swelled.
• SEQ273, SEQ382, SEQ260, SEQ368, SEQ227, and SEQ228, SEQ321 and SEQ211 are not
used in any rheology analysis
240

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
384 53% 7.3 451 0.995 5.9 108.11 27% 2.0 None
385 46% 8.09 100 0.981 2.6 7.53 21% 26 None
241
Table B-25: NTS044 ground for 5 minutes results

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
270 46% 6.05 1449 0.963 44.3 69.16 0% NA NA
269 42% 5.94 726 0.995 25.5 19.20 8% 0.81 Yes
271 38% 6.01 240 0.996 8.1 7.48 11% 1.95 Yes
272 34% 6.14 127 0.993 4.2 4.55 3% 3.16 Yes
273 28% 6.31 51 0.900 0.9 9.09 Na 14.5 Yes
242
Table B-26: Solids concentration and swelling details for NTS044 after five minutes of grinding
Solids Swelling
concentration
(v/v)
46% NA
42% Total volume swelled by 2 mL though the slurry settled.
28% The mudline settled and the total volume increased by app roximately 4 mL
243

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
377 53% 5.27 2078 0.991 47.9 214.2 2% NA Na
379 46% 5.39 705 0.984 13.8 98.8 3% 2.6 None
380 41% 5.44 266 0.997 6.5 24.1 26% 2.1 None
3 8 1 14 36% 5.71 86 0.896 3.3 1.3 -13% 8 None
3 8 2 15 31% 5.47 40 0.997 Na Na Na 27 None
14 p-value = 3.2x10-6.
15 Approximate value. Not used in any rheology analysis
244

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
245 53% 5.57 4728 0.998 139.9 251.84 14% NA NA
Yes
257 46% 3.85 826 0.993 23.3 50.94 10% 0
258 41% 3.86 317 0.999 9.0 19.36 19% 0.25 Yes
259 36% 4.05 111 0.980 3.5 4.43 -5% 2.5 Yes
2 6 0 16 31% 5.47 43 0.997 NA 20.5 None
16 Had to exclude a couple of outliers to generate the approximate viscosity. The shear stresses are outside of the rheometers range. The data is not used in
any rheology analysis
245
Solids Swelling
concentration
(v/v)
53% NA
Slurry rises by 1 mL over 24 minutes and 30 seconds (2.4 mL / hour). The slurry
doesn’t settle but an additional 5 mL liquid layer forms on top of the mudline
46%
Liquid layer forms on top without the mud level increasing. The mudline
41% decreases after a very long time
36% The total volume swells by 1 mL; The mudline settles
31% None
Table B-30 contains the results for NTS054. As can be seen in the table, the SEQ numbers are
quite different at the same concentration. This is because the tests were performed on different dates
and from independently splitting the sample. Beneath 36% concentration (v/v) no useful flow curves
can be generated from the available data. At 41% the R 2 were relatively low, as such the p -values were
246
checked and found to be 5.4E-06 for SEQ226 and 5.4E-07 for SEQ367. As such, the data is viewed as
reliable.
247

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
375 57% 5.8 1852 0.997 44.3 176 0% NA Na
364 53% 5.47 828 0.999 18.5 91 0% 1.8 None
224 52% 5.55 872 0.999 23.2 64 -7% NA Na
225 46% 5.61 344 0.995 4.5 81 33% 2.8 None
366 46% 5.68 364 0.999 6.4 61 37% 2.8 None
226 41% 5.71 80 0.884 2.9 2 11% 21.3 None
367 41% 5.74 85 0.927 3.2 2 6% 21.1 None
368 36% 5.79 Na Na Na Na Na 21.5 None
227 36% 5.75 Na Na Na Na Na 38.7 None
228 31% 5.76 Na Na Na Na Na 38.6 None
248
Table B-31: NTS054 after grinding for 5 minutes results

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
323 53% 6.00 3313 0.947 97.8 178 1% NA NA
324 46% 6.04 815 0.984 22.6 53 9% 0.8 Yes
3 2 5 17 46% 6.04 1070 0.972 28.0 82 Na 0.8 Yes
319 42% 5.89 330 0.997 10.5 13 15% 1.3 Yes
320 38% 6.15 117 0.985 3.3 7 7% 2.7 Yes
3 2 1 18 31% 6.25 49 0.859 1.2 5 Na 31.8 Yes
17 Not used as 324 was preferred as it was the older test. It is noteworthy the considerable differences in apparent viscosities.
18 Not reliable and thus not used in later analysis. Very complicated sedimentation
249
Table B-32: Solids concentration and swelling details for NTS054 after 5 minutes grinding
Solids Swelling
concentration
(v/v)
53% NA
Settled without mud layer increasing. However, the total volume increased by 4
46% mL
46% mL
42% mL
38% mL
31% mL
250

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
203 53% 6.77 682 0.997 10.2 140 Na 1.8 None
204 53% 6.77 728 0.988 6.3 249 4% 1.8 None
205 46% 6.78 95 0.953 1.4 19 -38% 18.9 None
2 1 1 19 41% 7.09 33 Na Na Na Na 108.1 None
19 Slows down rapidly after 13 minutes and 19 s. Took the sedimentation rate off those initial high speeds
251
A.7: Inner biotite
Table B-34 shows the results for NTS019; it was a fast settling slurry. The 46% concentrati on
was tested first and found to be fast settling. The high degree of time dependency is likely to be
because the sample settled around the bob during the test. Surprising, the sample was fluid underneath
the bob, and was relatively fluid above. An alterna tive ramping procedure was used for SEQ 374 were
the sample was ramped from 800 rpm to 10 without a loop test.

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
374 53% 6.92 539.259 0.971 12.0 59.43 Na 9 None
372 46% 6.15 77.465 1.000 0.5 33.47 32% 19.5 None
NTS055 is shown in Table B-35; like the other inner biotite samples NTS055 has low viscosity
and a high settling rate. Both SEQ279 and SEQ281 underwent sedimentation which began with a linear
252
profile before slowing down. S EQ279 began to slow after 16 minutes after having dropped 26 mL.
SEQ281 cannot be used for rheology analysis.
253
Table B-35: NTSS055 results

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
278 53% 6.42 515 0.992 3.4 211 25% 4.0 None
2 7 9 20 46% 6.56 99 0.896 0.7 38 -22% 22.7 None
2 8 1 21 41% 5.76 37 0 Na 33.2 None
Table B-36, Table B-37, and Table B-38 show the results of NTS097 before and after grinding.
Only three SEQs need to be dismi ssed: SEQ241, SEQ243 and SEQ317. The as provided NTS097 did not
swell; after grinding for five minutes significant swelling occurred and is described in Table B-38.
20 Corresponds to a p value of 0.0012
21 Unreliable. Took the Newtonian viscosity; all data was clearly in the turbulent range.
254

at 50 s-1 (mL/hour)
(mPa.s) in area
under
curve)
238 53% 4.21 1637 0.998 35.9 186 16% NA NA
239 46% 4.21 461 0.999 13.0 29 18% 1.6 None
240 41% 4.31 134 0.990 3.7 9 15% 3.3 None
241 35% 4.21 61 0.711 0.9 13 Na 17.3 None
255
Table B-37: NTS097 after grinding for 5 minut es

at 50 s-1 (Pa) (mL/hour)
(v/v) (mPa.s) (% change
(mPa.s) in area
under
curve)
313 46% 5.1 1514 0.998 40.7 108 12% NA Yes
310 42% 4.98 687 0.997 24.6 16 6% 0 . 0 22 Yes
311 38% 5.09 222 0.998 7.0 9 4% 5.3 Yes
312 34% 5.2 97 0.967 2.7 6 9% 5.5 Yes
317 30% 0.393899 49 0.906 1.2 5 Na 8.0 Yes
22
No sedimentation was observed via the camera which recorded several hours of footage. However the slurry had settled over night by 3 mL
256
Table B-38: Solids concentration and swelling details for NTS097 after 5 minutes grinding
Solids Swelling
concentration
(v/v)
46% NA
42% M u d s e t t l e s b u t 3 m L o f l i q u i d f o r m s o n t o p 23
38% Mud settles but 1 mL of liquid forms on top
23 No sedimentation was observed via the camera. However the slurry had settled over night by 3 mL
257
Appendix C
Statistical results
A.8: Regression analysis of apparent viscosity vs
pH, concentration, passing 9.86, passing 38,
passing 53, amorphous silicate, kaolinite and clay
content
In the below regression analysis, the follo wing are used to
describe the variables:
• η = apparent viscosity at 50 mPa.s
• concentration= solids concentration on a volumetric basis
• passing ## = amount of solids beneath ## in diameter
• amorphous= amorphous silicate fraction as determined by

QXRD
• QXRD_Kaolinite= kaolinite fraction as determined by

QXRD
• Clay = total clay fraction as determined by QXRD
Method
Rows 5
unused
Stepwise Selection of Terms
α to enter = 0.15, α to remove = 0.15

The stepwise procedure added terms during the procedure in order to
maintain a hierarchical model at each step.
258
Analysis of Variance
F- P-
Source DF Adj SS Adj MS Value Value
Regression 15 128.299 8.5532 22.35 0.000
Concentration 1 13.591 13.5909 35.51 0.000
Passing 9.86 1 11.625 11.6251 30.37 0.000
Passing 38 1 6.353 6.3535 16.60 0.000
Passing 53 1 3.230 3.2300 8.44 0.005
amorphous 1 0.115 0.1146 0.30 0.586
QXRD_Kaolinite 1 0.253 0.2535 0.66 0.418
Clay 1 0.009 0.0090 0.02 0.879
Concentration* 1 16.154 16.1543 42.21 0.000
Concentration
Passing 9.86*Passing 1 12.684 12.6840 33.14 0.000
9.86
amorphous*amorphous 1 2.307 2.3067 6.03 0.016
Concentration*Passing 1 6.540 6.5401 17.09 0.000
9.86
Concentration*Passing 1 3.288 3.2881 8.59 0.004
38
Concentration* 1 0.832 0.8324 2.17 0.144
QXRD_Kaolinite
Passing 9.86*Passing 1 11.142 11.1416 29.11 0.000
38
amorphous*Clay 1 3.121 3.1207 8.15 0.005
Error 83 31.768 0.3827
Total 98 160.067
Model Summary
R R- R-
S -sq sq(adj) sq(pred)
0.6 8 76. 68.1

18667 0.15% 57% 7%
259
Coefficients
T- P-
Term Coef SE Coef Value Value VIF
Constant 19.05 4.53 4.20 0.000
Concentration -91.6 15.4 -5.96 0.000 366.74
Passing 9.86 -320.3 58.1 -5.51 0.000 4962.6
4
Passing 38 152.4 37.4 4.07 0.000 4820.2
1
Passing 53 22.80 7.85 2.90 0.005 232.29
amorphous -0.0135 0.0247 -0.55 0.586 40.50
QXRD_Kaolinite -0.118 0.146 -0.81 0.418 76.56
Clay 0.0049 0.0321 0.15 0.879 13.92
Concentration*Concentratio 94.5 14.5 6.50 0.000 235.24
n
Passing 9.86*Passing 9.86 874 152 5.76 0.000 7515.3
5
amorphous*amorphous 0.00121 0.00049 2.45 0.016 23.97
3 4
Concentration*Passing 9.86 407.4 98.6 4.13 0.000 1828.0
3
Concentration*Passing 38 -178.0 60.7 -2.93 0.004 1734.3
7
Concentration*QXRD_Kaolini 0.505 0.343 1.47 0.144 60.24
te
Passing 9.86*Passing 38 -548 102 -5.40 0.000 8380.2
8
amorphous*Clay - 0.00183 -2.86 0.005 11.78
0.00522
260
Regression Equation
η = 19.05 - 91.6 Concentration - 320.3 Passing 9.86 C-1
+ 152.4 Passing 38 + 22.80 Passing 53
- 0.0135 amorphous - 0.118 QXRD_Kaolinite
+ 0.0049 Clay
+ 94.5 Concentration*Concentration
+ 874 Passing 9.86*Passing 9.86
+ 0.001213 amorphous*amorphous
+ 407.4 Concentration*Passing 9.86
- 178.0 Concentration*Passing 38
+ 0.505 Concentration*QXRD_Kaolinite
- 548 Passing 9.86*Passing 38 -
0.00522 amorphous*Clay
Fits and Diagnostics for Unusual Observations
Obs vis Fit Resid Std Resid
15 1.616 2.809 -1.193 -2.05 R
18 7.192 5.224 1.968 3.69 R
53 4.792 2.753 2.038 3.54 R
93 0.451 2.141 -1.690 -2.91 R
100 3.900 4.984 -1.084 -2.48 R X
R Large residual
X Unusual X
261
A.9: Regression analysis of the loga rithm of
apparent viscosity vs pH, concentration, passing
9.86, passing 38, passing 53, amorphous silicate,
kaolinite and clay content
• ln(𝜂)= l o g a r i t h m o f t h e a p p a r e n t v i s c o s i t y a t 5 0 m P a . s
• concentration= solids concentration on a volumetric basis
• passing ## = amount of solids beneath ## in diameter
• amorphous= amorphous silicate fraction as determined by

QXRD
• QXRD_Kaolinite= kaolinite fraction as determined by

QXRD
• Clay = total clay fraction as determined by QXRD
Method
Rows 5
unused
Stepwise Selection of Terms
α to enter = 0.15, α to remove = 0.15

The stepwise procedure added terms during the procedure in order to
maintain a hierarchical
model at each step.
262
Analysis of Variance
Source DF Adj SS Adj MS F-Value P-Value

Regression 20 138.252 6.91258 122.06 0.000
pH 1 0.153 0.15287 2.70 0.104
Concentration 1 6.780 6.77961 119.71 0.000
Passing 9.86 1 0.432 0.43176 7.62 0.007
Passing 38 1 1.288 1.28815 22.75 0.000
Passing 53 1 2.174 2.17390 38.39 0.000
amorphous 1 2.437 2.43716 43.03 0.000
QXRD_Kaolinite 1 0.773 0.77343 13.66 0.000
Clay 1 0.296 0.29551 5.22 0.025
pH*pH 1 0.560 0.56047 9.90 0.002
Passing 9.86*Passing 1 1.281 1.28133 22.62 0.000
9.86
Passing 53*Passing 53 1 1.984 1.98364 35.03 0.000
1 2.607 2.60732 46.04 0.000
amorphous*amorphous
pH*Passing 9.86 1 1.134 1.13377 20.02 0.000
pH*Passing 38 1 1.820 1.81964 32.13 0.000
pH*amorphous 1 0.398 0.39791 7.03 0.010
pH*QXRD_Kaolinite 1 0.424 0.42409 7.49 0.008
1 0.545 0.54518 9.63 0.003
Concentration*Passing
53
Passing 9.86*Passing 1 0.120 0.11971 2.11 0.150
38
Passing 9.86*Passing 1 1.766 1.76582 31.18 0.000
53
Passing 1 2.051 2.05107 36.22 0.000
9.86*QXRD_Kaolinite
Error 78 4.417 0.05663
Total 98 142.669
263
Model Summary
S R-sq R-sq(adj) R-sq(pred)
0.237978 96.90% 96.11% 94.92%
Coefficients
T- P-
Constant -13.25 1.27 -10.42 0.000
pH -0.405 0.246 -1.64 0.104 278.85
Concentration 15.71 1.44 10.94 0.000 21.63
Passing 9.86 -64.0 23.2 -2.76 0.007 5331.14
Passing 38 -154.1 32.3 -4.77 0.000 24305.1
9
Passing 53 197.0 31.8 6.20 0.000 25770.2
7
amorphous -0.0935 0.0143 -6.56 0.000 90.92
QXRD_Kaolinite -0.383 0.104 -3.70 0.000 262.70
Clay 0.0353 0.0154 2.28 0.025 21.85
pH*pH -0.0580 0.0184 -3.15 0.002 115.42
Passing 9.86*Passing -1631 343 -4.76 0.000 258911.
9.86 74
Passing 53*Passing 53 -692 117 -5.92 0.000 264604.
53
amorphous*amorphous 0.00208 0.00030 6.79 0.000 62.78
6 8
pH*Passing 9.86 -10.12 2.26 -4.47 0.000 1572.95
pH*Passing 38 8.19 1.45 5.67 0.000 1662.72
pH*amorphous 0.00421 0.00159 2.65 0.010 30.32
pH*QXRD_Kaolinite -0.0432 0.0158 -2.74 0.008 52.32
Concentration*Passing 10.96 3.53 3.10 0.003 43.94
53
264
T- P-
Passing 9.86*Passing 38 163 112 1.45 0.150 68599.1
8
Passing 9.86*Passing 53 1945 348 5.58 0.000 781356.
22
Passing 2.505 0.416 6.02 0.000 341.15
9.86*QXRD_Kaolinite
Regression Equation
ln(𝜂) = - 1 3 . 2 5 - 0 . 4 0 5 p H + 1 5 . 7 1 C o n c e n t r a t i o n - 6-1
64.0 Passing 9.86 - 154.1 Passing 38
+ 197.0 Passing 53 - 0.0935 amorphous -
0.383 QXRD_Kaolinite + 0.0353 Clay
- 0.0580 pH*pH -
1631 Passing 9.86*Passing 9.86 -
692 Passing 53*Passing 53
+ 0.002086 amorphous*amorphous -
10.12 pH*Passing 9.86 + 8.19 pH*Passing 38
+ 0.00421 pH*amorphous -
0.0432 pH*QXRD_Kaolinite
+ 10.96 Concentration*Passing 53
+ 2.505 Passing 9.86*QXRD_Kaolinite
265
Fits and Diagnostics for Unusual Observations
Obs LN VIS Fit Resid Std Resid
10 -2.352 -1.918 -0.434 -2.19 R
23 1.553 1.030 0.523 2.67 R
31 -0.191 0.510 -0.701 -3.15 R
32 -1.148 -0.588 -0.561 -2.51 R
33 -2.200 -1.634 -0.566 -2.57 R
74 1.103 0.600 0.503 2.38 R
76 -2.311 -1.803 -0.508 -2.38 R
R Large residual
266
Appendix D
Sizing data
The following pages contain tables recording the size
distribution data for each sample. Following these tables, there are
several figures which give the size distribution in graphical form.
267
268
269
270
Figure D-1:Advanced argillic cumulative passing si ze
271
Figure D-2: Argillic cumulative passing size
Figure D-3:Upper epithermal cumulative passing size
272
Figure D-4:Silica breccia cumulative passing size
Figure D-5: Upper epithermal cumulative passing size
273
Figure 5-5:Outer biotite cumulative passing size
274
Figure D-6:Inner biotite cumulative passing size
Appendix E
Data processing example
including raw data
An example of the data processing of SEQ233 is outlined here;
the raw data can be found in section A.13:. As stated earlier two
different data reduction procedures were used: the Estelle and Casson
approximation methods. Both methods contain some common early data
processing steps; the first of which is defining the measuring systems
characteristics and properties as in Table E-1 and Table E-2.
275
Table E-1: Measuring system characteristics of SEQ233. SEQ233
used the large bob and the RM180 rheometer in the wide gap
configuration described elsewhere in this thesis.
Radius of the bob 𝑟𝑏 = 𝑟1 0.015

(m)
Radius of the cup (m) 𝑟𝑐 0.025
Height of the bob (m) h 0.045
𝑘𝑔 Ρ 1871.7
Density ( )
𝑚3
Correction factor F 1.263321848
Table E-2: Measuring system properties
Formula Value
Α 𝑟𝑐 1.666666667
=
𝑟𝑏
Radial gap (m) = 𝑟𝑐 − 𝑟1 0.01
𝐾𝑏 = 2π. 𝑟12 . ℎ. 𝐹 8 . 0 3 6 9 1 E - 0 5
Table E-3 shows the averaging of the raw data and the shear
stress and angular speed calculations. Th e raw data is averaged at each
RPM from 10 data points as this helps smooth the results. If the RPM
was being held constant for longer than 10 s, then multiple aver aged
points are presented.
Shear stress at the surface of the bob is calculated from the

torque and the geometry of the bob. The shear stress at the cup is
lower than the surface, and, assuming that the gap is fully sheared and
under laminar conditions, is equal to the shear stress at the bob
divided by the radial gap squared. The angular velocit y is just the
RPM multiplied by a constant.
𝑀
𝜏1 = 2𝜋𝑟 2 𝐿𝐹 4-1
1
276
r γ̇ 1 τ f(τ)
Ω = ∫r 2 dr = − ∫τ 2 dτ 2-10
1 r 2 1 τ
Table E-3 Averaged raw data and the calculated shear stress and
angular velocity results.
RPM M (N.m) τ𝑏 ( P a ) τ𝑐 ( P a ) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠

Ω ( )
(Averaged s
(Averaged
machine machine output)
output)
M
=
2πr22 LF
M M
= =
2πr12 LF 2πr12 𝛼 2 LF
M τ1 2π
= = = . 𝑅𝑃𝑀
𝐾𝑏 𝛼2 60
5 0.001502784 =18.7 =6.7 0.52
5 0.0021797 27.1 9.8 0.52
5 0.00228851 28.5 10.3 0.52
5 0.00233974 29.1 10.5 0.52
5 0.00238873 29.7 10.7 0.52
5 0.00243061 30.2 10.9 0.52
10 0.0026696 33.2 12.0 1.05
20 0.00303396 37.8 13.6 2.09
30 0.00324352 40.4 14.5 3.14
40 0.00340765 42.4 15.3 4.19
50 0.00352853 43.9 15.8 5.24
60 0.00364045 45.3 16.3 6.28
70 0.00373017 46.4 16.7 7.33
80 0.00385324 47.9 17.3 8.38
90 0.00392992 48.9 17.6 9.42
100 0.0040088 49.9 18.0 10.47
150 0.00431092 53.6 19.3 15.71
200 0.00457533 56.9 20.5 20.94
250 0.00481335 59.9 21.6 26.18
300 0.0050199 62.5 22.5 31.42
277
RPM M (N.m) τ𝑏 ( P a ) τ𝑐 ( P a ) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠
Ω ( )
(Averaged s
(Averaged
machine machine output)
output)
350 0.00523503 65.1 23.4 36.65
400 0.00544091 67.7 24.4 41.89
450 0.00563007 70.1 25.2 47.12
500 0.00582553 72.5 26.1 52.36
500 0.0059172 73.6 26.5 52.36
500 0.00599362 74.6 26.8 52.36
500 0.00604968 75.3 27.1 52.36
500 0.00614162 76.4 27.5 52.36
450 0.00602603 75.0 27.0 47.12
400 0.00588598 73.2 26.4 41.89
350 0.00572696 71.3 25.7 36.65
300 0.0055601 69.2 24.9 31.42
250 0.0053547 66.6 24.0 26.18
200 0.00511641 63.7 22.9 20.94
150 0.00483761 60.2 21.7 15.71
100 0.00447027 55.6 20.0 10.47
90 0.00440289 54.8 19.7 9.42
80 0.00433052 53.9 19.4 8.38
70 0.00425943 53.0 19.1 7.33
60 0.00414214 51.5 18.6 6.28
50 0.00401673 50.0 18.0 5.24
40 0.00388232 48.3 17.4 4.19
30 0.00371478 46.2 16.6 3.14
20 0.00351036 43.7 15.7 2.09
10 0.00318076 39.6 14.2 1.05
278
Table E-3 contains data points that need to be excluded; the
reduced data set is seen in Table E-4. There are two reasons why we
exclude data:
1. The torque range of the rheometer, according to the

manufacturer’s specifications, is between 0.00025 and
0.0075 Nm. Any reading outside of this range needs to be
removed before undergoing further analysis. For SEQ233
all the data was within the appropriate torque range.
2. To calculate the shear rate, we need to remove the effects

of time dependency. An assumption behind the
mathematical equations governing coaxial cylinder rheology
is that all time derivatives in the equations of continuity
and motion are zero; in practice many materials however
display time dependency. During the initial shearing of a
yield stress fluid its shear stress will peak before declining
to an equilibrium. As such, only the f inal 5 rpm ramp
averaged data is taken as being representative of the fluid
properties. The fluid may also be time dependently shear
thinning or shear thickening o ver the hysteresis loop; as
such, the first ramp up is assumed to represent the shearing
profile for the hold, ramp down and all subsequent loops.
Table E-4: Data used for calculating shear rate. From Table E-3,
the last 5 rpm rate is treated as the equilibrium yield stress. At 500
rpm, the torque exceeded the manufacturers specification, so the data
point is omitted from further analysis.
𝑟𝑎𝑑𝑖𝑎𝑛𝑠
τ𝑏 ( P a ) Ω ( )
s
18.7 0.52
27.1 0.52
28.5 0.52
279
τ𝑏 ( P a ) Ω ( )
s
29.1 0.52
29.7 0.52
30.2 0.52
33.2 1.05
37.8 2.09
40.4 3.14
42.4 4.19
43.9 5.24
45.3 6.28
46.4 7.33
47.9 8.38
48.9 9.42
49.9 10.47
53.6 15.71
56.9 20.94
59.9 26.18
62.5 31.42
65.1 36.65
67.7 41.89
70.1 47.12
72.5 52.36
73.6 52.36
74.6 52.36
280
τ𝑏 ( P a ) Ω ( )
s
75.3 52.36
76.4 52.36
75.0 47.12
73.2 41.89
71.3 36.65
69.2 31.42
66.6 26.18
63.7 20.94
60.2 15.71
55.6 10.47
54.8 9.42
53.9 8.38
53.0 7.33
51.5 6.28
50.0 5.24
48.3 4.19
46.2 3.14
43.7 2.09
39.6 1.05
A.10: Estelle method

The Estelle method assumes that the fluid has a Bingham
shearing profile. As such there are two distinct flow conditions: the
281
partially and fully sheared as seen respectively in the following two
equations:
𝑑Ω
𝛾̇ = 2𝑀 (𝑑𝑀) , 𝑖𝑓 τ2 < τ0 2-20
𝑑Ω 𝑑Ω
𝑀( ) Ω−𝑀( )
𝑑𝑀 𝑑𝑀
𝛾̇ = 2 𝑟2
− 𝑟1 , 𝑖𝑓 τ2 > τ0 2-21
1−( 12) ln( )
𝑟2
𝑟2
All of the properties above are either directly relatable to the

machine readings (torque and angular speed) or to the system set up
(radius of the cup and of the bob) with the exception of yield stress
𝑑𝛺
a n d t h e d e r i v a t i v e o f a n g u l a r s p e e d a g a i n s t t o r q u e ( 𝑑𝑀). T h e y i e l d
stress problem is avoided by the assumption that the true shear rate
occurs under the condition which maximises the energy dispensation,
i.e.
𝑑Ω 𝑑Ω
𝑑𝛺 𝑀( ) Ω−𝑀( )
𝑑𝑀 𝑑𝑀
𝛾̇ = max (2 𝛺 ;2 𝑟 2 − 𝑟b ) E-1
𝑑𝑀 ln( )
1−( b ) 𝑟𝑐
𝑟2c
hence, the only difficulty is finding the derivative of angular

speed against torque. This was attempted by fitting polynomial
equations to the angular velocity vs torque curve and taking there
derivative. The polynomial equations were fitted using the Microsoft
excels inbuilt LINEST function which is based off the least squares
method. Curves were tried with both 2 nd and 3rd order polynomials
which achieved very high R2 values yet would often give implausible
𝑑𝛺
results such as a negative . An example of such a curve is seen in
𝑑𝑀
Figure 4-3; the R2 value is almost 1 yet as can be seen in Table E-5
there was a negative derivati ve of angular speed against torque. As
described in section 4.9.2 taking higher order fits could overcome
these problems but at the expense of possibly overfitting the data.
282
Figure 4-3: Angular speed vs torque for SEQ233.
283
Table E-5 : The components required for the Estelle
approximation. Highlighted in red is the physically impossible result
𝑑𝛺
of a negative .
𝑑𝑀
𝑟𝑎𝑑𝑖𝑎𝑛𝑠 𝑑Ω
τ𝑏 ( P a ) Ω ( ) = 10𝐸 +07 𝑀 − 25072
s 𝑑𝑀
30.2 0.52 -1178.71138
33.2 1.05 1170.580736
37.8 2.09 4752.270662
40.4 3.14 6812.263364
42.4 4.19 8425.675285
43.9 5.24 9613.936012
45.3 6.28 10714.11917
46.4 7.33 11596.07444
47.9 8.38 12805.86305
48.9 9.42 13559.634
49.9 10.47 14335.03114
53.6 15.71 17304.89655
56.9 20.94 19904.06944
59.9 26.18 22243.82637
62.5 31.42 24274.23052
65.1 36.65 26388.97679
67.7 41.89 28412.79478
70.1 47.12 30272.25374
72.5 52.36 32193.64224
284
This could be avoided by splitting the curve into multiple
components. Table E-6 shows how SEQ233 was split into three parts:
• Ω between 0.52 and 5.24;
• Ω between 6.28 and 10.47;
• Ω between 15.71 and 52.36.
This division was somewhat arbitrary (a clear problem with the

method) and is based off a graphical assessment of Figure 4-3. The
parameters (also seen in Table E-6) a2, a1, and a0 are the coefficients
seen in the equation below. Table E-7 shows the calculation of the
shear rate:
𝛺 = 𝑎2 𝜏 2 + 𝑎1 𝜏 + 𝑎0 E-2
Table E-6: SEQ233 angular speed vs torque fits. As can be seen

they are split into three parts labelled a, b, and c referring to three
separate parts of the curve shown in Figure 4-3.
Part 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 Coefficients

Ω ( )
of the s
curve min max a2 a1 a0
a 0.52 5.24 3301777.416 -15531.07369 18.84169308
b 6.28 10.47 4376192.75 -22253.52857 29.30284736
c 15.71 52.36 3239489.889 -8578.04977 -7.390359854
285
Table E-7: The shear rate calculation using the Estelle method.
𝛾̇
𝑟𝑎𝑑𝑖𝑎𝑛𝑠 𝑑Ω
τ𝑏 ( P a ) Ω ( )
s Part of the curve 𝑑𝑀 Partial Fully Max
30.2 0.52 520 2.5 2.5 2.5
33.2 1.05 2098 11.2 8.6 11.2
37.8 2.09 4504 27.3 20.1 27.3
40.4 3.14 5888 38.2 28.4 38.2
42.4 4.19 6972 47.5 35.9 47.5
43.9 5.24 a 7770 54.8 42.3 54.8
45.3 6.28 9609 70.0 53.1 70.0
46.4 7.33 10394 77.5 59.6 77.5
47.9 8.38 11472 88.4 68.0 88.4
48.9 9.42 12143 95.4 74.2 95.4
49.9 10.47 b 12833 102.9 80.6 102.9
53.6 15.71 c 19352 166.9 128.1 166.9
286
56.9 20.94 21065 192.8 153.5 192.8
59.9 26.18 22608 217.6 178.3 217.6
62.5 31.42 23946 240.4 201.8 240.4
65.1 36.65 25340 265.3 226.6 265.3
67.7 41.89 26673 290.3 251.4 290.3
70.1 47.12 27899 314.1 275.6 314.1
72.5 52.36 29165 339.8 300.8 339.8
287
The final steps are calculating the apparent viscosity and
trimming the data to exclude shear rates which have a Taylor number
above 41.3.
𝜏

𝑇𝑎 = ∗√ 3-6
𝜂 𝑅𝑏
w h e r e U𝑏 = Ω. 𝑟𝑏
Since this method doesn’t involve the calculation of the shearing

radius, the Taylor number was calculated assuming that the entire
annulus was sheared:
(R𝑐 −Rb )∗Ub ∗ρ Rc −Rb

Ta = ∗√
η Rb
Table E-8 contains a worked example continuing with the results

from SEQ233.
288
Table E-8: Calculating the Taylor number with the Estelle method. The red lines indicate cells
with Taylor vortices
τ𝑏 ( P a ) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 𝛾̇ 𝜂𝑎𝑝𝑝𝑎𝑟𝑒𝑛𝑡 Ub Taylor fully sheared

Ω ( )
s
30.2 0.52 2.5 11.97 0.008 0.0
33.2 1.05 11.2 2.97 0.016 0.1
37.8 2.09 27.3 1.38 0.031 0.3
40.4 3.14 38.2 1.06 0.047 0.6
42.4 4.19 47.5 0.89 0.063 0.9
43.9 5.24 54.8 0.80 0.079 1.3
45.3 6.28 70.0 0.65 0.094 2.0
46.4 7.33 77.5 0.60 0.110 2.5
47.9 8.38 88.4 0.54 0.126 3.1
48.9 9.42 95.4 0.51 0.141 3.7
49.9 10.47 102.9 0.48 0.157 4.3
289
τ𝑏 ( P a ) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 𝛾̇ 𝜂𝑎𝑝𝑝𝑎𝑟𝑒𝑛𝑡 Ub Taylor fully sheared
Ω ( )
s
53.6 15.71 166.9 0.32 0.236 9.8
56.9 20.94 192.8 0.30 0.314 14.3
59.9 26.18 217.6 0.28 0.393 19.1
62.5 31.42 240.4 0.26 0.471 24.3
65.1 36.65 265.3 0.25 0.550 30.0
67.7 41.89 290.3 0.23 0.628 36.1
70.1 47.12 314.1 0.22 0.707 42.5
72.5 52.36 339.8 0.21 0.785 49.3
290
The final results of the Estelle approximation are presented
in Table E-9. The shear rates for the original yield stress test and
the hysteresis loop are included and were assumed to be equal to
the shear rates at the equivalent rpm. This allows the exploration
of any time dependency.
291
Table E-9: Final result from the Estelle calculation
τ𝑏 ( P a ) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 𝛾̇ 𝜂𝑎𝑝𝑝𝑎𝑟𝑒𝑛𝑡
RPM M Ω ( )
s
5 0.001503 18.70 0.52 2.5 7.40
5 0.00218 27.12 0.52 2.5 10.74
5 0.002289 28.48 0.52 2.5 11.27
5 0.00234 29.11 0.52 2.5 11.53
5 0.002389 29.72 0.52 2.5 11.77
5 0.002431 30.24 0.52 2.5 11.97
10 0.00267 33.22 1.05 11.2 2.97
20 0.003034 37.75 2.09 27.3 1.38
30 0.003244 40.36 3.14 38.2 1.06
40 0.003408 42.40 4.19 47.5 0.89
50 0.003529 43.90 5.24 54.8 0.80
60 0.00364 45.30 6.28 70.0 0.65
70 0.00373 46.41 7.33 77.5 0.60
80 0.003853 47.94 8.38 88.4 0.54
90 0.00393 48.90 9.42 95.4 0.51
100 0.004009 49.88 10.47 102.9 0.48
150 0.004311 53.64 15.71 166.9 0.32
200 0.004575 56.93 20.94 192.8 0.30
250 0.004813 59.89 26.18 217.6 0.28
300 0.00502 62.46 31.42 240.4 0.26
350 0.005235 65.14 36.65 265.3 0.25
400 0.005441 67.70 41.89 290.3 0.23
292
τ𝑏 ( P a ) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 𝛾̇ 𝜂𝑎𝑝𝑝𝑎𝑟𝑒𝑛𝑡
Ω ( )
RPM M s
400 0.005886 73.24 41.89 290.3 0.25
350 0.005727 71.26 36.65 265.3 0.27
300 0.00556 69.18 31.42 240.4 0.29
250 0.005355 66.63 26.18 217.6 0.31
200 0.005116 63.66 20.94 192.8 0.33
150 0.004838 60.19 15.71 166.9 0.36
100 0.00447 55.62 10.47 102.9 0.54
90 0.004403 54.78 9.42 95.4 0.57
80 0.004331 53.88 8.38 88.4 0.61
70 0.004259 53.00 7.33 77.5 0.68
60 0.004142 51.54 6.28 70.0 0.74
50 0.004017 49.98 5.24 54.8 0.91
40 0.003882 48.31 4.19 47.5 1.02
30 0.003715 46.22 3.14 38.2 1.21
20 0.00351 43.68 2.09 27.3 1.60
10 0.003181 39.58 1.05 11.2 3.53
293
To compare the relative viscosities of different samples, a shear
rate of 50 s-1 was selected. The exact selection of 50 s -1 was due to a
combination of
• most samples torque readings were sufficient to find a

reading of 50 s-1;
• most samples had laminar conditions at 50 s -1;
• and it is an intermediate shear rate based on previous

estimates inside a tumbling ball mill
The apparent viscosity was calculated through interpolation

using the excel addon xlxtrfun (Anon, 2005). Only the ramp up data
was selected to avoid any time effects. For seq233 the apparent
viscosity calculated from the Estelle method was 0.858 Pas.
A.11: Casson approximation

The Casson model (as seen below) is often used to describe the
rheograms of mineral slurries and pastes. As it is an analytical model,
the angular speed can be directly predicted from the shear stres s and
fluid parameters.
√𝜏 = √𝜏0 + √𝜂∞ . 𝛾̇ 2-7

2
√τ−√τ0
( )
r γ̇ 1 τ f(τ) 1 τ2 √η∞ 1 τ 2√τ0 τ
Ω = ∫r 2 r dr = − 2 ∫τ 2 dτ = − 2 ∫τ dτ = − 2η ∫τ 2 (1 − + τ0 ) dτ 4-2
1 1 τ 1 τ ∞ 1 √τ
Therefore, the partial and fully sheared conditions are:

1 τ
Ω = 2η (τ − 4√τ0 τ + 3τ0 + τ0 ln (τ )) 4-3
∞ 0
1 1 1
Ω = 2η ( (1 − α2)τ1 − 4(1 − α)√τ0 τ1 + 2τ0 ln(α)) 4-4
∞
As such, the fluid can be modelled by the Casson fluid using an

iterative procedure:
1. Set the yield stress and the infinite viscosity of the fluid
294
2. Calculate whether, at a given shear stress, th e fluid is fully
or partially yielded. This is done by calculating the r adial
distance (d) that the fluid would be sheared if the bob was
placed in an infinite medium.
τ1
𝑑 = 𝑟1 . √
τ0
In this example, the radial gap = 0.01 m. Hence:
• 𝑃𝑎𝑟𝑡𝑖𝑎𝑙𝑙𝑦 𝑠ℎ𝑒𝑎𝑟𝑒𝑑 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛: 𝑑 < 0.01 𝑚
• 𝐹𝑢𝑙𝑙𝑦 𝑠ℎ𝑒𝑎𝑟𝑒𝑑 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛: 𝑑 > 0.01 𝑚
3. The angular velocity of the bob is calculated un der both the

partial and fully sheared conditions. The appropriate
condition is selected based off the result from step 2 as the
predicted angular velocity.
4. The sum of residual squared of the measured against

predicted angular velocity is calculated. Using Excel’s
inbuilt solver, the sum of the residuals squared is
minimised by varying the yield stress and i nfinite
viscosity. The boundaries on the solution are:
a . 𝜂∞ > 0.001; a s t h e v i s c o s i t y o f a s u s p e n s i o n c a n n o t b e
smaller than the carrying fluid.
b . 𝜏𝑦 ≥ 0; a s n e g a t i v e y i e l d s t r e s s i s p h y s i c a l l y i m p o s s i b l e .
c . 𝜏𝑦 < 𝐸𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚 𝑠𝑡𝑟𝑒𝑠𝑠 𝑎𝑡 5 𝑟𝑝𝑚; a s t h e f l u i d i s b e i n g s h e a r e d

at 5 rpm so it must yield at a lower stress.
Excel solver will iterate steps 1 through 3 automatically until

the sum of least squares is minimised.
5. Calculate the shear rate and apparent viscosities from the

shear stress. Check if there are any Taylor vortices. If so,
exclude them from the data analysis and repeat step 4.
295
6. Calculate the R2 value using Excel’s CORREL fu nction. The
CORREL function only calculates R, so it will need to be
squared.
An example of step 1 through 4 is shown below in Table E-10 and

Table E-11.
Table E-10: The parameters of the Caisson fluid model
I n f i n i t e v i s c o s i t y ( 𝜂∞ ; 𝑃𝑎. 𝑠) 0.0304
Y i e l d s t r e s s ( 𝜏𝑦 ; 𝑃𝑎) 26.88
Table E-11: Steps 2 and 3 of the Casson analy sis
τ𝑏 ( P a ) Yielding 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 (𝛺𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑,𝑖

𝛺( )
distance s − 𝛺𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑,𝑖 )2
Measured Partial Fully Predicted

30.2 0.01591 10.9 0.063 19.741 0.063 0.212
33.2 0.01667 12.0 0.378 15.012 0.378 0.448
37.8 0.01777 13.6 1.642 10.780 1.642 0.205
40.4 0.01838 14.5 2.888 9.757 2.888 0.064
42.4 0.01884 15.3 4.153 9.592 4.153 0.001
43.9 0.01917 15.8 5.252 9.803 5.252 0.000
45.3 0.01947 16.3 6.399 10.236 6.399 0.013
46.4 0.01971 16.7 7.407 10.740 7.407 0.006
47.9 0.02003 17.3 8.922 11.650 8.922 0.297
48.9 0.02023 17.6 9.943 12.339 9.943 0.268
49.9 0.02043 18.0 11.053 13.143 11.053 0.338
53.6 0.02119 19.3 15.881 17.063 15.881 0.030
56.9 0.02183 20.5 20.841 21.508 20.841 0.011
59.9 0.02239 21.6 25.882 26.247 25.882 0.089
62.5 0.02286 22.5 30.689 30.886 30.689 0.528
65.1 0.02335 23.4 36.114 36.203 36.114 0.290
67.7 0.02380 24.4 41.694 41.728 41.694 0.038
296
τ𝑏 ( P a ) Yielding 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 (𝛺𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑,𝑖
𝛺( )
distance s − 𝛺𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑,𝑖 )2
Measured Partial Fully Predicted

70.1 0.02421 25.2 47.150 47.159 47.150 0.001
72.5 0.02463 26.1 53.109 53.110 53.109 0.561
Table E-12 contains the shear rate, apparent viscosity, shearing
distance and Taylor number calculations. The shear rate is found
through re-arranging the Casson equation to:
1 2
𝛾̇ = 𝜂 . (√τ1 − √τ0 ) E-3
∞
Unlike the Estelle calculation, the Casson approximation

contains a yield parameter (see Table E-10). As such the Taylor
number can be calculated by:
𝑑∗Ub ∗ρ d
Ta = ∗ √𝑟 E-4
η b
w h e r e U𝑏 = Ω. 𝑟𝑏
As can be seen in Table E-12 the Taylor number is above 41.3

when the shear stress equals 72.5.
297
Table E-12: The Taylor number calculation technique. Note, for this example, the shearing is
always partial due to the high y ield stress. Under the fully sheared condition the distance would be
0.01 m
τ𝑏 (Pa) Ub 1 η d=Shearing distance (m) Taylor number
𝛾̇ ( )
s
30.2 0.008 3.3 9.29 0.0009 0.0
33.2 0.016 11.0 3.02 0.0017 0.0
37.8 0.031 30.3 1.25 0.0028 0.0
40.4 0.047 44.9 0.90 0.0034 0.1
42.4 0.063 57.9 0.73 0.0038 0.3
43.9 0.079 68.4 0.64 0.0042 0.4
45.3 0.094 78.6 0.58 0.0045 0.7
46.4 0.110 87.2 0.53 0.0047 0.9
47.9 0.126 99.6 0.48 0.0050 1.2
48.9 0.141 107.6 0.45 0.0052 1.6
49.9 0.157 116.1 0.43 0.0054 2.0
53.6 0.236 150.6 0.36 0.0062 4.3
56.9 0.314 183.4 0.31 0.0068 7.7
59.9 0.393 214.8 0.28 0.0074 12.0
62.5 0.471 243.3 0.26 0.0079 17.2
65.1 0.550 274.2 0.24 0.0083 23.7
67.7 0.628 304.9 0.22 0.0088 31.3
70.1 0.707 334.0 0.21 0.0092 39.9
72.5 0.785 364.8 0.20 0.0096 50.1
298
Since there were possible vortices, the procedure was
iterated. The results are shown in Table E-13, Table E-13 and
Table E-14.
Table E-13: The parameters of the Casson fluid model
I n f i n i t e v i s c o s i t y ( 𝜂∞ ; 𝑃𝑎. 𝑠) 0.0295
Y i e l d s t r e s s ( 𝜏𝑦 ; 𝑃𝑎) 27.10
Table E-14: Steps 2 and 3 of the Casson analysis

τ𝑏 (Pa) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 Yielding 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 (𝛺𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑,𝑖
𝛺𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑 ( ) 𝛺 ( )
s distance s − 𝛺𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑,𝑖 )2
Partial Fully Predicted
30.2 10.9 0.01585 0.053 20.948 0.053 0.222
33.2 12.0 0.01661 0.348 15.935 0.348 0.489
37.8 13.6 0.01770 1.580 11.366 1.580 0.264
40.4 14.5 0.01831 2.813 10.196 2.813 0.108
42.4 15.3 0.01876 4.071 9.938 4.071 0.014
43.9 15.8 0.01909 5.169 10.091 5.169 0.005
45.3 16.3 0.01939 6.315 10.477 6.315 0.001
46.4 16.7 0.01963 7.326 10.949 7.326 0.000
47.9 17.3 0.01995 8.846 11.822 8.846 0.220
48.9 17.6 0.02015 9.872 12.493 9.872 0.200
49.9 18.0 0.02035 10.989 13.281 10.989 0.267
53.6 19.3 0.02110 15.853 17.167 15.853 0.021
56.9 20.5 0.02174 20.863 21.615 20.863 0.007
59.9 21.6 0.02230 25.962 26.382 25.962 0.048
62.5 22.5 0.02277 30.829 31.061 30.829 0.344
65.1 23.4 0.02326 36.327 36.437 36.327 0.105
67.7 24.4 0.02371 41.988 42.032 41.988 0.010
70.1 25.2 0.02412 47.526 47.540 47.526 0.162
299
Table E-15: Taylor data
τ𝑏 (Pa) Ub 1 η Shearing distance (d) Taylor number
𝛾̇ ( )
s
30.2 0.008 2.9 10.36 0.0008 0.0
33.2 0.016 10.5 3.15 0.0016 0.0
37.8 0.031 29.8 1.27 0.0027 0.0
40.4 0.047 44.6 0.91 0.0033 0.1
42.4 0.063 57.8 0.73 0.0038 0.3
43.9 0.079 68.3 0.64 0.0041 0.4
45.3 0.094 78.7 0.58 0.0044 0.6
46.4 0.110 87.5 0.53 0.0046 0.9
47.9 0.126 100.0 0.48 0.0050 1.2
48.9 0.141 108.2 0.45 0.0051 1.5
49.9 0.157 116.8 0.43 0.0054 1.9
53.6 0.236 152.0 0.35 0.0061 4.3
56.9 0.314 185.4 0.31 0.0067 7.6
59.9 0.393 217.4 0.28 0.0073 11.9
62.5 0.471 246.5 0.25 0.0078 17.1
65.1 0.550 278.0 0.23 0.0083 23.6
67.7 0.628 309.4 0.22 0.0087 31.3
70.1 0.707 339.1 0.21 0.0091 39.9
300
As can be seen in Table E-15 no other shear stresses contain
Taylor numbers above 41.3.
The final step was step 6, where the R2 value was found using
Excel’s inbuilt CORREL function. A plot of the graph can be seen in
Figure E-1.
Figure E-1: A comparison of the predicted vs measured angular

v e l o c i t y w h e r e 𝛺𝑃 i s t h e p r e d i c t e d a n g u l a r v e l o c i t y a n d 𝛺𝐸 i s t h e
experimental angular velocity.
The final results of the Casson approximation are presented in

Table E-6. Like the Estelle method, the shear rates for the original
yield stress test and the hysteresis loop are included and were assumed
to be equal to the shear rates at the equivalent rpm. This allows the
exploration of any time dependency.
Table E-16: Final results of the Casson method

τ𝑏 (Pa) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 1 𝜂𝑎𝑝𝑝𝑎𝑟𝑒𝑛𝑡 (𝑃𝑎𝑠)
RPM M (Nm) 𝛺 ( ) 𝛾̇ ( )
s s
5 0.001503 18.70 0.52 2.9 6.403534
5 0.00218 27.12 0.52 2.9 9.287951
5 0.002289 28.48 0.52 2.9 9.751602
5 0.00234 29.11 0.52 2.9 9.969899
5 0.002389 29.72 0.52 2.9 10.17865
5 0.002431 30.24 0.52 2.9 10.36
10 0.00267 33.22 1.05 10.5 3.15
20 0.003034 37.75 2.09 29.8 1.27
301
τ𝑏 (Pa) 𝑟𝑎𝑑𝑖𝑎𝑛𝑠 1 𝜂𝑎𝑝𝑝𝑎𝑟𝑒𝑛𝑡 (𝑃𝑎𝑠)
RPM M (Nm) 𝛺 ( ) 𝛾̇ ( )
s s
30 0.003244 40.36 3.14 44.6 0.91
40 0.003408 42.40 4.19 57.8 0.73
50 0.003529 43.90 5.24 68.3 0.64
60 0.00364 45.30 6.28 78.7 0.58
70 0.00373 46.41 7.33 87.5 0.53
80 0.003853 47.94 8.38 100.0 0.48
90 0.00393 48.90 9.42 108.2 0.45
100 0.004009 49.88 10.47 116.8 0.43
150 0.004311 53.64 15.71 152.0 0.35
200 0.004575 56.93 20.94 185.4 0.31
250 0.004813 59.89 26.18 217.4 0.28
300 0.00502 62.46 31.42 246.5 0.25
350 0.005235 65.14 36.65 278.0 0.23
400 0.005441 67.70 41.89 309.4 0.22
450 0.00563 70.05 47.12 339.1 0.21
450 0.006026 74.98 47.12 278.0 0.269671
400 0.005886 73.24 41.89 246.5 0.297146
350 0.005727 71.26 36.65 217.4 0.32784
300 0.00556 69.18 31.42 185.4 0.373199
250 0.005355 66.63 26.18 152.0 0.438425
200 0.005116 63.66 20.94 116.8 0.545114
150 0.004838 60.19 15.71 108.2 0.556478
100 0.00447 55.62 10.47 100.0 0.556071
90 0.004403 54.78 9.42 87.5 0.626312
80 0.004331 53.88 8.38 78.7 0.684434
70 0.004259 52.99838 7.33 68.32603 0.775669
60 0.004142 51.53899 6.28 57.75453 0.89238
50 0.004017 49.97856 5.24 44.56505 1.121474
40 0.003882 48.30615 4.19 29.82636 1.619579
30 0.003715 46.22152 3.14 10.533 4.388258
20 0.00351 43.678 2.09 2.920034 14.95804
10 0.003181 39.57692 1.05 10.533 3.757422
A.12: Summary
The data processing procedure was outlined using SEQ233 as an
example. A comparison between the different calculation techniques is
provided in Figure E-2; as can be seen both the Estelle and Casson
methods produce similar answers. The Newtonian approximation is
also included; its equation is given by:
302
2Ω. 𝑟𝑐2
𝛾̇ =
𝑟𝑐2 − 𝑟𝑏2
Clearly, the Newtonian approximation should be avoided a s it

significantly underestimates the shear rate and thus overestimates the
viscosity. The advantages of the Casson method over the Est elle:
• its automatable;
• consistent;
• verifiable;
• and, if the fluid is well described by the Casson model,

then it also saves time fitting the Casson parameter to the
rheogram as they have already been generated from the raw
data.
Likewise, the Estelle method should be used instead of the

Casson in situations where the Casson provides a poor fit.
Figure E-2: A comparison of the different rheograms produced by

the different methods.
303
A.13: Raw data for SEQ233
Raw results for SEQ233
RPM M (Nm) Temperature (OC)
5 0.00018773 21.044
5 0.00073941 21.036
5 0.0012379 21.036
5 0.0014994 21.036
5 0.0016697 21.021
5 0.0017566 21.044
5 0.001924 21.021
5 0.0019663 21.044
5 0.0020058 21.029
5 0.002041 21.014
5 0.0021316 21.021
5 0.0021201 21.029
5 0.0021344 21.029
5 0.0020899 21.014
5 0.0021366 20.991
5 0.0022314 20.999
5 0.0021998 21.021
5 0.0022221 20.999
5 0.0022817 21.014
5 0.0022494 21.029
5 0.002268 21.006
5 0.0022041 21.006
304
5 0.0022156 20.999
5 0.0023262 21.029
5 0.0023176 21.014
5 0.0022343 21.014
5 0.0022889 20.999
5 0.0023291 20.999
5 0.0023262 21.014
5 0.0023751 20.969
5 0.0023176 20.984
5 0.0022946 20.984
5 0.0023025 20.999
5 0.0022846 20.976
5 0.0022882 20.991
5 0.002462 20.984
5 0.002406 20.999
5 0.0023765 20.969
5 0.0023399 20.961
5 0.0023255 20.954
5 0.0023787 20.991
5 0.0024246 20.969
5 0.0023621 20.961
5 0.0023277 20.976
5 0.0023736 20.976
305
5 0.0024096 20.947
5 0.0024081 20.969
5 0.0024139 20.961
5 0.0023751 20.969
5 0.0024139 20.976
5 0.0024879 20.947
5 0.0025202 20.954
5 0.0023801 20.954
5 0.0023851 20.932
5 0.0024699 20.961
5 0.0024354 20.969
5 0.0023765 20.939
5 0.002406 20.961
5 0.0023952 20.947
5 0.0024498 20.961
10 0.0025863 20.917
10 0.0026006 20.947
10 0.0026552 20.947
10 0.0026725 20.917
10 0.0026006 20.947
10 0.0026983 20.887
10 0.0026955 20.917
10 0.002661 20.917
306
10 0.0027673 20.917
10 0.0027587 20.947
20 0.0029791 20.917
20 0.0029532 20.947
20 0.0030653 20.887
20 0.0030021 20.887
20 0.0030365 20.887
20 0.0030538 20.917
20 0.003071 20.917
20 0.003071 20.947
20 0.0030624 20.917
20 0.0030452 20.917
30 0.0032441 20.887
30 0.0032326 20.887
30 0.0032613 20.917
30 0.0032527 20.917
30 0.003224 20.887
30 0.0032613 20.917
30 0.0032182 20.976
30 0.0032786 20.917
30 0.0032125 20.917
30 0.0032499 20.887
40 0.0034255 20.917
307
40 0.0033737 20.887
40 0.0033623 20.917
40 0.0034341 20.887
40 0.0033766 20.917
40 0.0034513 20.857
40 0.0033737 20.857
40 0.0034255 20.887
40 0.003437 20.857
40 0.0034168 20.917
50 0.0035481 20.887
50 0.0035193 20.917
50 0.0035107 20.917
50 0.0035193 20.917
50 0.0034992 20.887
50 0.0035481 20.917
50 0.003528 20.887
50 0.0035538 20.887
50 0.0035136 20.827
50 0.0035452 20.857
60 0.0036356 20.857
60 0.0036212 20.857
60 0.0036298 20.887
60 0.0036356 20.887
308
60 0.0036586 20.857
60 0.00367 20.857
60 0.0036499 20.857
60 0.0036183 20.887
60 0.0036212 20.857
60 0.0036643 20.917
70 0.0035733 20.827
70 0.0037773 20.887
70 0.0037227 20.887
70 0.0037313 20.887
70 0.0037543 20.857
70 0.0037256 20.857
70 0.0037801 20.857
70 0.0037457 20.857
70 0.0037342 20.827
70 0.0037572 20.857
80 0.0038785 20.857
80 0.0038239 20.857
80 0.0038757 20.827
80 0.0038785 20.857
80 0.0038699 20.857
80 0.0038757 20.827
80 0.0038354 20.827
309
80 0.0038412 20.827
80 0.0038239 20.857
80 0.0038297 20.827
90 0.0039308 20.857
90 0.0039423 20.857
90 0.0039222 20.827
90 0.0039394 20.827
90 0.0039365 20.827
90 0.0039193 20.827
90 0.0039365 20.857
90 0.0039279 20.857
90 0.0039078 20.857
90 0.0039365 20.827
100 0.0040232 20.857
100 0.0040088 20.857
100 0.0040059 20.857
100 0.0040088 20.857
100 0.0039887 20.857
100 0.0039944 20.857
100 0.0040289 20.857
100 0.0040203 20.857
100 0.0039973 20.857
100 0.0040117 20.827
310
150 0.0042917 20.797
150 0.004306 20.827
150 0.0043204 20.827
150 0.0043032 20.827
150 0.004329 20.797
150 0.0043003 20.797
150 0.0043089 20.797
150 0.004329 20.857
150 0.0043147 20.827
150 0.004306 20.797
200 0.0046227 20.797
200 0.0045796 20.827
200 0.004571 20.767
200 0.0045739 20.797
200 0.0045682 20.797
200 0.0045682 20.797
200 0.0045595 20.797
200 0.0045624 20.767
200 0.0045825 20.797
200 0.0045653 20.797
250 0.0048702 20.857
250 0.0048185 20.797
250 0.0048099 20.857
311
250 0.00483 20.827
250 0.0048042 20.797
250 0.0047812 20.827
250 0.004807 20.797
250 0.0047783 20.797
250 0.0048128 20.797
250 0.0048214 20.827
300 0.0050084 20.827
300 0.0050199 20.857
300 0.0050055 20.857
300 0.005017 20.797
300 0.0050228 20.797
300 0.0050199 20.797
300 0.0050084 20.827
300 0.0050314 20.797
300 0.0050285 20.827
300 0.0050372 20.767
350 0.0052615 20.797
350 0.0052327 20.827
350 0.0052097 20.767
350 0.0052155 20.797
350 0.00525 20.797
350 0.005204 20.797
312
350 0.0052299 20.797
350 0.0052672 20.797
350 0.0052385 20.767
350 0.0052413 20.797
400 0.0054599 20.827
400 0.005411 20.797
400 0.0054283 20.797
400 0.0054369 20.827
400 0.0054455 20.797
400 0.0054426 20.797
400 0.0054283 20.827
400 0.005457 20.797
400 0.0054541 20.827
400 0.0054455 20.857
450 0.0056496 20.797
450 0.005618 20.797
450 0.0056352 20.857
450 0.0056266 20.797
450 0.0056266 20.797
450 0.0056209 20.797
450 0.0056209 20.797
450 0.0056324 20.827
450 0.0056324 20.857
313
450 0.0056381 20.827
500 0.0058594 20.797
500 0.0058163 20.857
500 0.0058106 20.767
500 0.0058163 20.827
500 0.005825 20.797
500 0.0058221 20.797
500 0.0058049 20.857
500 0.0058163 20.797
500 0.0058422 20.857
500 0.0058422 20.827
500 0.0058824 20.797
500 0.0058997 20.797
500 0.0059025 20.797
500 0.0059083 20.827
500 0.0059198 20.797
500 0.005914 20.827
500 0.0059342 20.797
500 0.0059457 20.827
500 0.0059255 20.827
500 0.0059399 20.827
500 0.0059715 20.857
500 0.0059686 20.797
314
500 0.0059686 20.857
500 0.0059916 20.827
500 0.005983 20.827
500 0.0060232 20.827
500 0.0059945 20.827
500 0.0059916 20.797
500 0.0060146 20.797
500 0.006029 20.797
500 0.0060232 20.857
500 0.0060146 20.827
500 0.0060491 20.827
500 0.006052 20.827
500 0.0060405 20.797
500 0.006052 20.887
500 0.0060635 20.827
500 0.0060692 20.857
500 0.0060692 20.857
500 0.0060635 20.827
500 0.0061094 20.827
500 0.0061238 20.827
500 0.0061152 20.797
500 0.0061382 20.827
500 0.0061382 20.797
315
500 0.0061525 20.797
500 0.0061554 20.827
500 0.0061583 20.797
500 0.0061554 20.827
500 0.0061698 20.827
450 0.0059312 20.827
450 0.0059915 20.827
450 0.0060117 20.857
450 0.0060203 20.857
450 0.0060203 20.797
450 0.0060318 20.857
450 0.0060605 20.857
450 0.0060519 20.857
450 0.0060691 20.857
450 0.006072 20.827
400 0.0058047 20.887
400 0.0058736 20.797
400 0.0058708 20.797
400 0.0058765 20.857
400 0.0058851 20.797
400 0.0058937 20.857
400 0.0058966 20.857
400 0.0059081 20.797
316
400 0.0059196 20.857
400 0.0059311 20.827
350 0.0056752 20.827
350 0.0056752 20.827
350 0.0057241 20.857
350 0.0057155 20.797
350 0.0057155 20.827
350 0.0057471 20.827
350 0.0057787 20.857
350 0.0057528 20.797
350 0.0057327 20.797
350 0.0057528 20.857
300 0.0055515 20.857
300 0.0055515 20.797
300 0.0055285 20.827
300 0.0055314 20.857
300 0.0055659 20.857
300 0.0055457 20.827
300 0.0055917 20.857
300 0.0055773 20.827
300 0.0055773 20.797
300 0.0055802 20.857
250 0.0053041 20.827
317
250 0.0053501 20.797
250 0.0053329 20.827
250 0.0053645 20.827
250 0.0053271 20.827
250 0.0053443 20.827
250 0.0053386 20.797
250 0.0053731 20.827
250 0.0054076 20.797
250 0.0054047 20.827
200 0.0050624 20.797
200 0.0051055 20.797
200 0.0050825 20.797
200 0.0050997 20.827
200 0.0051198 20.827
200 0.0051285 20.827
200 0.0051285 20.797
200 0.00514 20.827
200 0.0051486 20.767
200 0.0051486 20.827
150 0.0047572 20.827
150 0.0048232 20.797
150 0.0048405 20.767
150 0.0048405 20.797
318
150 0.004829 20.767
150 0.0048347 20.767
150 0.004852 20.797
150 0.0048577 20.767
150 0.0048865 20.797
150 0.0048548 20.767
100 0.0044398 20.767
100 0.0044542 20.827
100 0.0044599 20.797
100 0.0044858 20.767
100 0.0044714 20.797
100 0.0044599 20.797
100 0.0044973 20.767
100 0.0044628 20.797
100 0.0044973 20.797
100 0.0044743 20.767
90 0.0043963 20.797
90 0.004402 20.797
90 0.0043762 20.797
90 0.0043963 20.797
90 0.0043934 20.797
90 0.004402 20.797
90 0.0043991 20.767
319
90 0.0044078 20.767
90 0.004448 20.797
90 0.0044078 20.797
80 0.0043182 20.797
80 0.0043153 20.767
80 0.00439 20.797
80 0.0043239 20.797
80 0.0043411 20.797
80 0.0043153 20.797
80 0.0043124 20.767
80 0.0043498 20.797
80 0.0043268 20.797
80 0.0043124 20.797
70 0.0042054 20.767
70 0.0042284 20.797
70 0.0042456 20.767
70 0.0043002 20.767
70 0.0042514 20.797
70 0.0042974 20.767
70 0.0042571 20.737
70 0.004306 20.797
70 0.0042744 20.827
70 0.0042284 20.767
320
60 0.0041298 20.767
60 0.0041585 20.797
60 0.0041413 20.737
60 0.0041413 20.767
60 0.0041269 20.767
60 0.0041355 20.797
60 0.0041499 20.767
60 0.0041413 20.767
60 0.0041154 20.767
60 0.0041815 20.797
50 0.003982 20.737
50 0.0039992 20.767
50 0.0040337 20.767
50 0.0040107 20.797
50 0.0040279 20.797
50 0.0039992 20.737
50 0.0040193 20.767
50 0.0040251 20.767
50 0.0040423 20.797
50 0.0040279 20.797
40 0.0039139 20.737
40 0.0038593 20.797
40 0.0038306 20.767
321
40 0.0039111 20.737
40 0.003845 20.737
40 0.0038737 20.797
40 0.0038507 20.797
40 0.0039513 20.767
40 0.0039053 20.797
40 0.0038823 20.737
30 0.0037326 20.767
30 0.0037584 20.767
30 0.0036924 20.767
30 0.003724 20.767
30 0.0036694 20.737
30 0.0037125 20.797
30 0.0036751 20.767
30 0.0037671 20.737
30 0.003701 20.767
30 0.0037153 20.797
20 0.0034618 20.737
20 0.0034963 20.797
20 0.0035595 20.708
20 0.0035279 20.767
20 0.0035336 20.767
20 0.0034589 20.767
322
20 0.0035221 20.797
20 0.0035279 20.797
20 0.0035308 20.737
20 0.0034848 20.797
10 0.0031839 20.797
10 0.0031351 20.767
10 0.003204 20.797
10 0.0031696 20.737
10 0.0032155 20.737
10 0.0031954 20.737
10 0.0031552 20.737
10 0.0031839 20.737
10 0.0032127 20.767
10 0.0031523 20.767
Appendix F
Sedimentation curves
Please see attached document labelled
s4464068_MPhil_thesis_supplementary1.pdf
323
Appendix G
Flow curves
Please see attached document labelled
s4464068_MPhil_thesis_supplementary2.pdf
324

Rheological Properties of Lihir Ore

Uploaded by

Copyright:

Available Formats

Rheological Properties of Lihir Ore

Uploaded by

Document Information

Original Title

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

Rheological Properties of Lihir Ore

Uploaded by

Copyright:

Available Formats

Rheological properties of Lihir ore

Peter William Legge

Bachelor of Engineering (Hons)

A thesis submitted for the degree of Master of Philosophy at

The University of Queensland in 2019

Sustainable Minerals Institute (SMI)

Julius Kruttschnitt Mineral Research Centre (JKMRC)

Lihir has a complicated mineralogy and is currently defined by

The slurry rheology was measured through two different

The tests showed that concentration and particle size make a

The regression analysis also found a poor correlation with the

A sedimentation investigation was performed simultaneously to

I have clearly stated the contribution of others to my thesis as a

I acknowledge that an electronic copy of my thesis must be

I acknowledge that copyright of all material contain ed in my

Submitted manuscripts included in this

Other publications during candidature

Contributions by others to the thesis

Experimental work was assisted by Luyan da “Loui” Ncube and

Ding Sheng Kaw assisted me in analysing some of the

Tim Napier-Munn aided with the statistical analysis.

Michael “Mick” Kilmartin assisted in designing and constructing

Statement of parts of the thesis submitted to

Research Involving Human or Animal

• Newcrest, for their generous financial support and

• My advisory team: Frank Shi and Yongjun Peng for their

• Loui Ncube and Kaitlyn Porter through their help and

• The pilot plant workers and technicians, in particular David

• Ryan Anderson who ensured that the In dooroopilly mine

• All of my fellow students throughout SMI, in particular The

And my student guide when I first came to the SMI -JKMRC

• Alice Clark and Neville Pint for their stewardship of the

• The administrative support at the JKMRC, in particular

• Pat Griffin from rheological solutions for assisting me with

• Ashish Kumar from MEP instruments/Anton Paar who was

• Fred Koshir for all the technical support in regard to the

• Karen Holtham for her assistance in library matters

• Nigel Tiong for helping me with all my IT needs (which

• SMI Postgrad, in particular Tess Dobinson for assisting the

• My milestone review panel, Tim Napier -Munn, Elaine

• My family and friends who supported me throughout this

• And especially my wife Kirsti, who supported me

Australian and New Zealand Standard

Fields of Research (FoR) Classification

K = Herschel − Bulkey consistency index

K B = bob constant parameter

K V = vane constant parameter

L = length of the bob or vane

r1 = rb = radius of the bob

rc = radius of the cup

ry = radius of the sheared region

t = t statistic for the two sided Student t − distribution

α = ratio of the radius of cup to the radius of bob

ηapparent = apparent viscosity

η0 = carrier fluid viscosity

ηr = relative fluid viscosity

τ1 = shear stress at the surface of the bob

This project is a part of a larger project investigating the effect

The goal of the sponsors, Newcrest, is to increase the efficiency